diff --git a/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/checkpoints/best_model.safetensors b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/checkpoints/best_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c2bb001d61d70fa6b682eb94361690332795d828 --- /dev/null +++ b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/checkpoints/best_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:630d55bc258d986e0de4b04d9223f418855bb94c2d310c13ae808912553d6c7f +size 9138230516 diff --git a/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_500_model.safetensors b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_500_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..263418863e2f17f088d6bc8264ad4cf9e199569e --- /dev/null +++ b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_500_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ad5f732f19bbfda68d6686e39e875f07957d08980bf6325dea7f85eccb4abc70 +size 9138230516 diff --git a/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/config.full.yaml b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..b1cf9a095a6bb761434f78d991e6fe39b0577fb2 --- /dev/null +++ b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/config.full.yaml @@ -0,0 +1,191 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 7 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_2 + data_mix: deadly_corridor_train__bridge + eval_data_mix: deadly_corridor_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + converted_name: deadly_corridor_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: deadly_corridor_train + mixed_converted_name: deadly_corridor_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 500 + num_warmup_steps: 100 + save_interval: 50 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: You are playing Deadly Corridor in VizDoom. Choose actions from + MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, ATTACK. + deadly: + action_layout: multibinary_7 + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_2/deadly_corridor_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 50 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 2 + - 4 + - 6 + - 8 + num_episodes: 5 + max_steps_per_episode: 3600 + task: deadly_corridor + deadly_corridor_loss_type: null + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: deadly_corridor +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: true + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_deadly_corridor_fix_latency_2 +output_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2 +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/config.yaml b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..0f3f2e255702f884e3fe9670d750f0069b9947cc --- /dev/null +++ b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/config.yaml @@ -0,0 +1,92 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: true + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: deadly_corridor_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: deadly_corridor_train__bridge__val + include_state: true + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 7 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2 +rl_games: + env_eval: + deadly: + action_layout: multibinary_7 + distributed_mode: none + enabled: true + fixed_episode_seeds: true + frameskip: 4 + image_size: 224 + latency: + prompt_map_path: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_2/deadly_corridor_train__bridge/latency_prompt_map.json + latency_seed_stride: 0 + mid_train: + enabled: true + interval_steps: 50 + latencies: + - 2 + max_steps_per_episode: 3600 + num_episodes: 5 + seed: 42 + task_description: You are playing Deadly Corridor in VizDoom. Choose actions from + MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, ATTACK. + task_seed_stride: 0 + vectorized: + enabled: false + model_alias: openvla + task: deadly_corridor +run_id: openvla_deadly_corridor_fix_latency_2 +run_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_interval: 100 + eval_num_batches: 20 + freeze_modules: '' + gradient_accumulation_steps: 16 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 500 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + pretrained_checkpoint: null + reload_modules: null + save_interval: 50 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/dataset_statistics.json b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..71b10d0cdac7c030ee86cbf61147c45735a75f98 --- /dev/null +++ b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.9800952076911926, + 0.00043271310278214514, + 0.04197317361831665, + 0.853310227394104, + 0.33102554082870483, + 0.3102552890777588, + 0.8277801871299744 + ], + "std": [ + 0.13967378437519073, + 0.02079719677567482, + 0.2005293071269989, + 0.3537939190864563, + 0.47058752179145813, + 0.4625938832759857, + 0.3775743842124939 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 0.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 2311, + "num_trajectories": 90 + } +} \ No newline at end of file diff --git a/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/dataset_statistics_eval.json b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..2033cfd707777ca6f779897cbd5ed01ea8b5f2f2 --- /dev/null +++ b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.9674796462059021, + 0.0, + 0.05284553021192551, + 0.8211382031440735, + 0.3211382031440735, + 0.32926830649375916, + 0.8373983502388 + ], + "std": [ + 0.17737756669521332, + 0.0, + 0.22372478246688843, + 0.38323649764060974, + 0.4669141173362732, + 0.46994760632514954, + 0.36900219321250916 + ], + "max": [ + 1.0, + 0.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 0.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 246, + "num_trajectories": 10 + } +} \ No newline at end of file diff --git a/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/hydra/.hydra/config.yaml b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..38920500915668e11bc55616a8252c63931ad3bd --- /dev/null +++ b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/hydra/.hydra/config.yaml @@ -0,0 +1,189 @@ +framework: + qwenvl: + base_vlm: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 7 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_2 + data_mix: deadly_corridor_train__bridge + eval_data_mix: deadly_corridor_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: deadly_corridor_train + mixed_converted_name: deadly_corridor_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 500 + num_warmup_steps: 100 + save_interval: 50 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: You are playing Deadly Corridor in VizDoom. Choose actions from + MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, ATTACK. + deadly: + action_layout: multibinary_7 + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_2/deadly_corridor_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 50 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 2 + - 4 + - 6 + - 8 + num_episodes: 5 + max_steps_per_episode: 3600 + task: deadly_corridor + deadly_corridor_loss_type: null + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: deadly_corridor +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: true + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_deadly_corridor_fix_latency_2 +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/hydra/.hydra/hydra.yaml b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..c5831f46e9390876e8ec08abffe79d8094921d5d --- /dev/null +++ b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/hydra/.hydra/hydra.yaml @@ -0,0 +1,252 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=deadly_corridor + - init=bridge + - mode=single + - run_id=openvla_deadly_corridor_fix_latency_2 + - run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=true + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=500 + - trainer.num_warmup_steps=100 + - trainer.save_interval=50 + - trainer.eval_interval=100 + - trainer.eval_num_batches=20 + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=16 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9, 0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=7 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_2 + - datasets.vla_data.data_mix=deadly_corridor_train__bridge + - datasets.vla_data.eval_data_mix=deadly_corridor_train__bridge__val + - framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=deadly_corridor + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=4 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - rl_games.env_eval.task_description='You are playing Deadly Corridor in VizDoom. + Choose actions from MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, + TURN_RIGHT, ATTACK.' + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_2/deadly_corridor_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=true + - rl_games.env_eval.mid_train.interval_steps=50 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[2] + - rl_games.env_eval.post_train.enabled=true + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,2,4,6,8] + - rl_games.env_eval.deadly.action_layout=multibinary_7 + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=true,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=deadly_corridor_train__bridge,datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_2,datasets.vla_data.eval_data_mix=deadly_corridor_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.video_backend=torchvision_av,env=deadly_corridor,framework.action_model.action_dim=7,framework.action_model.action_env_dim=7,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.deadly.action_layout=multibinary_7,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=4,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_2/deadly_corridor_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=true,rl_games.env_eval.mid_train.interval_steps=50,rl_games.env_eval.mid_train.latencies=[2],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=true,rl_games.env_eval.post_train.latencies=[0,2,4,6,8],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_description='You + are playing Deadly Corridor in VizDoom. Choose actions from MOVE_FORWARD, MOVE_BACKWARD, + MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, ATTACK.',rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=deadly_corridor,run_id=openvla_deadly_corridor_fix_latency_2,run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=20,trainer.gradient_accumulation_steps=16,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=500,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9, + 0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=50,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.2 + version_base: '1.1' + cwd: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: deadly_corridor + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/hydra/.hydra/overrides.yaml b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..d2e3f77641e7cf4c0a15864a91c0da373c3f6227 --- /dev/null +++ b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/hydra/.hydra/overrides.yaml @@ -0,0 +1,89 @@ +- model=openvla +- env=deadly_corridor +- init=bridge +- mode=single +- run_id=openvla_deadly_corridor_fix_latency_2 +- run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=true +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=500 +- trainer.num_warmup_steps=100 +- trainer.save_interval=50 +- trainer.eval_interval=100 +- trainer.eval_num_batches=20 +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=16 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9, 0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=7 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_2 +- datasets.vla_data.data_mix=deadly_corridor_train__bridge +- datasets.vla_data.eval_data_mix=deadly_corridor_train__bridge__val +- framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=deadly_corridor +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=4 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- rl_games.env_eval.task_description='You are playing Deadly Corridor in VizDoom. + Choose actions from MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, + TURN_RIGHT, ATTACK.' +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_2/deadly_corridor_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=true +- rl_games.env_eval.mid_train.interval_steps=50 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[2] +- rl_games.env_eval.post_train.enabled=true +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,2,4,6,8] +- rl_games.env_eval.deadly.action_layout=multibinary_7 diff --git a/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/hydra/train_starvla_hydra.log b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..3eb97fe23b6286bc7e467bcaa2714055378ee7d8 --- /dev/null +++ b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/hydra/train_starvla_hydra.log @@ -0,0 +1,722 @@ +[2026-06-10 15:19:09,407][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-10 15:19:09,418][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-10 15:20:15,783][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-10 15:20:15,821][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `deadly_corridor_train__bridge` +[2026-06-10 15:20:24,616][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-10 15:20:24,616][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-10 15:20:24,616][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-10 15:20:24,616][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-10 15:20:24,616][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-10 15:20:24,617][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-10 15:20:24,617][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-10 15:20:24,617][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-10 15:20:24,617][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-10 15:20:24,617][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-10 15:20:24,617][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-10 15:20:24,617][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-10 15:20:24,617][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-10 15:20:24,617][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-10 15:20:24,617][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-10 15:20:24,617][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-10 15:20:24,617][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-10 15:20:24,618][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-10 15:20:24,618][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-10 15:20:24,619][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-10 15:20:24,619][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-10 15:20:24,619][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-10 15:20:24,620][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-10 15:20:24,620][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-10 15:20:24,620][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-10 15:20:24,620][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-10 15:20:24,620][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-10 15:20:24,620][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-10 15:20:24,620][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-10 15:20:24,620][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-10 15:20:24,620][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-10 15:20:24,633][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `deadly_corridor_train__bridge__val` +[2026-06-10 15:20:24,658][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/config.full.yaml +[2026-06-10 15:20:24,664][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/config.yaml +[2026-06-10 15:20:24,664][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-10 15:20:29,040][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-10 15:20:29,040][starVLA.training.train_starvla][INFO] - Total optimization steps = 100 +[2026-06-10 15:20:29,040][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-10 15:20:29,040][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-10 15:20:29,040][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-10 15:21:27,918][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-10 15:21:27,930][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-10 15:22:17,129][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-10 15:22:17,218][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `deadly_corridor_train__bridge` +[2026-06-10 15:22:26,011][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-10 15:22:26,012][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-10 15:22:26,012][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-10 15:22:26,012][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-10 15:22:26,012][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-10 15:22:26,012][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-10 15:22:26,012][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-10 15:22:26,012][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-10 15:22:26,013][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-10 15:22:26,013][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-10 15:22:26,013][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-10 15:22:26,013][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-10 15:22:26,013][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-10 15:22:26,013][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-10 15:22:26,014][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-10 15:22:26,014][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-10 15:22:26,014][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-10 15:22:26,014][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-10 15:22:26,014][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-10 15:22:26,014][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-10 15:22:26,015][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-10 15:22:26,015][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-10 15:22:26,015][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-10 15:22:26,015][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-10 15:22:26,015][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-10 15:22:26,015][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-10 15:22:26,015][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-10 15:22:26,015][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-10 15:22:26,015][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-10 15:22:26,016][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-10 15:22:26,016][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-10 15:22:26,107][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `deadly_corridor_train__bridge__val` +[2026-06-10 15:22:26,220][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/config.full.yaml +[2026-06-10 15:22:26,229][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/config.yaml +[2026-06-10 15:22:26,229][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-10 15:22:37,820][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-10 15:22:37,820][starVLA.training.train_starvla][INFO] - Total optimization steps = 500 +[2026-06-10 15:22:37,821][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-10 15:22:37,821][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-10 15:22:37,898][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-10 15:23:16,813][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 1.97444748878479, 'train/grad_norm_pre_clip': 4.268056392669678, 'timing/data': 0.014521025121212006, 'timing/model': 1.2673113644123077, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.01}) +[2026-06-10 15:23:42,505][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 1.9743142127990723, 'train/grad_norm_pre_clip': 4.274668216705322, 'timing/data': 0.015417300164699554, 'timing/model': 0.9913996420800686, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.03}) +[2026-06-10 15:24:03,801][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 1.9627574682235718, 'train/grad_norm_pre_clip': 3.9627294540405273, 'timing/data': 0.02552858367562294, 'timing/model': 0.9675509221851826, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.04}) +[2026-06-10 15:24:21,885][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 1.9022002220153809, 'train/grad_norm_pre_clip': 4.062123775482178, 'timing/data': 0.030992604792118073, 'timing/model': 1.074785016477108, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.05}) +[2026-06-10 15:24:30,662][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 1.8614400625228882, 'train/grad_norm_pre_clip': 4.193300247192383, 'timing/data': 0.13956258445978165, 'timing/model': 0.5560094453394413, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.07}) +[2026-06-10 15:24:52,723][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 1.7844524383544922, 'train/grad_norm_pre_clip': 3.970360517501831, 'timing/data': 0.029600009322166443, 'timing/model': 1.3408335335552692, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.08}) +[2026-06-10 15:25:17,646][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 1.6866049766540527, 'train/grad_norm_pre_clip': 4.045138835906982, 'timing/data': 0.02000102773308754, 'timing/model': 0.9531933292746544, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.1}) +[2026-06-10 15:25:31,994][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 1.5734776258468628, 'train/grad_norm_pre_clip': 3.9338581562042236, 'timing/data': 0.017177827656269073, 'timing/model': 0.942395456135273, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.11}) +[2026-06-10 15:25:47,119][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 1.4849880933761597, 'train/grad_norm_pre_clip': 3.5523059368133545, 'timing/data': 0.021661844104528427, 'timing/model': 1.0069093853235245, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.12}) +[2026-06-10 15:25:55,516][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 1.309391975402832, 'train/grad_norm_pre_clip': 3.631101608276367, 'timing/data': 0.0903129018843174, 'timing/model': 0.46406661719083786, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.14}) +[2026-06-10 15:26:13,993][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 1.1854125261306763, 'train/grad_norm_pre_clip': 3.259222984313965, 'timing/data': 0.03366134315729141, 'timing/model': 0.9378977678716183, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.15}) +[2026-06-10 15:26:29,282][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 1.0097609758377075, 'train/grad_norm_pre_clip': 3.1587021350860596, 'timing/data': 0.0404953770339489, 'timing/model': 0.8201900981366634, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.16}) +[2026-06-10 15:26:47,896][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.9403414726257324, 'train/grad_norm_pre_clip': 2.555083990097046, 'timing/data': 0.037351712584495544, 'timing/model': 0.9360312223434448, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.18}) +[2026-06-10 15:27:07,917][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.7047943472862244, 'train/grad_norm_pre_clip': 2.513965129852295, 'timing/data': 0.04138984531164169, 'timing/model': 3.339410610496998, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.19}) +[2026-06-10 15:27:18,875][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.5632044672966003, 'train/grad_norm_pre_clip': 2.145667791366577, 'timing/data': 0.09232072159647942, 'timing/model': 0.35871826112270355, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.21}) +[2026-06-10 15:27:36,471][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.6637392044067383, 'train/grad_norm_pre_clip': 1.3268460035324097, 'timing/data': 0.037711165845394135, 'timing/model': 0.9302815534174442, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.22}) +[2026-06-10 15:27:54,566][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.3290349543094635, 'train/grad_norm_pre_clip': 1.3981032371520996, 'timing/data': 0.037346016615629196, 'timing/model': 1.5461004339158535, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.23}) +[2026-06-10 15:28:19,904][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.3378920555114746, 'train/grad_norm_pre_clip': 0.9160902500152588, 'timing/data': 0.08998304978013039, 'timing/model': 6.58543099462986, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.25}) +[2026-06-10 15:28:35,977][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.1691640168428421, 'train/grad_norm_pre_clip': 0.7870073914527893, 'timing/data': 0.07213690131902695, 'timing/model': 0.9561226442456245, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.26}) +[2026-06-10 15:28:44,301][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.11560604721307755, 'train/grad_norm_pre_clip': 0.5486017465591431, 'timing/data': 0.0880313366651535, 'timing/model': 0.33067982643842697, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.27}) +[2026-06-10 15:29:01,908][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.07748236507177353, 'train/grad_norm_pre_clip': 0.37336716055870056, 'timing/data': 0.030386071652173996, 'timing/model': 0.9544416703283787, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.29}) +[2026-06-10 15:29:17,000][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.05049315467476845, 'train/grad_norm_pre_clip': 0.24665790796279907, 'timing/data': 0.10034680366516113, 'timing/model': 1.4864055663347244, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.3}) +[2026-06-10 15:29:32,141][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.1947031319141388, 'train/grad_norm_pre_clip': 0.13665245473384857, 'timing/data': 0.027427103370428085, 'timing/model': 0.9769338890910149, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.32}) +[2026-06-10 15:29:51,024][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.021820001304149628, 'train/grad_norm_pre_clip': 0.10838501155376434, 'timing/data': 0.025201015174388885, 'timing/model': 0.9676813259720802, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.33}) +[2026-06-10 15:29:59,614][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.014447391964495182, 'train/grad_norm_pre_clip': 0.07234737277030945, 'timing/data': 0.08897757902741432, 'timing/model': 0.3515680655837059, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.34}) +[2026-06-10 15:30:22,306][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.009675266221165657, 'train/grad_norm_pre_clip': 0.04861605912446976, 'timing/data': 0.03857757896184921, 'timing/model': 0.8441285192966461, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.36}) +[2026-06-10 15:30:41,146][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.006619219668209553, 'train/grad_norm_pre_clip': 0.03338249400258064, 'timing/data': 0.03758486360311508, 'timing/model': 3.550745192915201, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.37}) +[2026-06-10 15:31:00,540][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.45120832324028015, 'train/grad_norm_pre_clip': 0.39679068326950073, 'timing/data': 0.04353082925081253, 'timing/model': 0.9662880264222622, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.38}) +[2026-06-10 15:31:16,684][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.2333398312330246, 'train/grad_norm_pre_clip': 0.1990753710269928, 'timing/data': 0.0372893363237381, 'timing/model': 0.9183487184345722, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.4}) +[2026-06-10 15:31:25,141][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.002543021459132433, 'train/grad_norm_pre_clip': 0.012948674149811268, 'timing/data': 0.08837451413273811, 'timing/model': 0.3587854281067848, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.41}) +[2026-06-10 15:31:45,726][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.24542011320590973, 'train/grad_norm_pre_clip': 0.20353659987449646, 'timing/data': 0.03140260651707649, 'timing/model': 0.9719989597797394, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.42}) +[2026-06-10 15:32:04,841][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.2550507187843323, 'train/grad_norm_pre_clip': 0.20085451006889343, 'timing/data': 0.03315722197294235, 'timing/model': 1.9923718608915806, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.44}) +[2026-06-10 15:32:24,473][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.2470867931842804, 'train/grad_norm_pre_clip': 0.19884298741817474, 'timing/data': 0.029391877353191376, 'timing/model': 1.2024827003479004, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.45}) +[2026-06-10 15:32:39,897][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.0012857719557359815, 'train/grad_norm_pre_clip': 0.006612436845898628, 'timing/data': 0.03979630023241043, 'timing/model': 0.9188624694943428, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.47}) +[2026-06-10 15:32:47,727][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.00115333148278296, 'train/grad_norm_pre_clip': 0.0058737811632454395, 'timing/data': 0.08445878326892853, 'timing/model': 0.36036287620663643, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.48}) +[2026-06-10 15:33:12,828][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.24877409636974335, 'train/grad_norm_pre_clip': 0.20157866179943085, 'timing/data': 0.033850666135549545, 'timing/model': 0.9759048521518707, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.49}) +[2026-06-10 15:33:27,871][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.0010419375030323863, 'train/grad_norm_pre_clip': 0.005388024728745222, 'timing/data': 0.028890322893857956, 'timing/model': 0.870535247027874, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.51}) +[2026-06-10 15:33:42,675][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.2690544128417969, 'train/grad_norm_pre_clip': 0.20507904887199402, 'timing/data': 0.032763462513685226, 'timing/model': 0.9554961510002613, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.52}) +[2026-06-10 15:33:57,335][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.0010475716553628445, 'train/grad_norm_pre_clip': 0.005457029212266207, 'timing/data': 0.03438679128885269, 'timing/model': 0.9510710313916206, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.53}) +[2026-06-10 15:34:05,046][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.0010803542099893093, 'train/grad_norm_pre_clip': 0.0057122111320495605, 'timing/data': 0.09319417923688889, 'timing/model': 0.35944200307130814, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.55}) +[2026-06-10 15:34:22,032][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.001066327327862382, 'train/grad_norm_pre_clip': 0.005621247459203005, 'timing/data': 0.014822516590356827, 'timing/model': 0.9878028929233551, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.56}) +[2026-06-10 15:34:37,332][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.0010890050325542688, 'train/grad_norm_pre_clip': 0.005764280911535025, 'timing/data': 0.01012573391199112, 'timing/model': 1.0090578608214855, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.58}) +[2026-06-10 15:34:54,905][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.46670737862586975, 'train/grad_norm_pre_clip': 0.4051949381828308, 'timing/data': 0.0007939152419567108, 'timing/model': 0.9528146535158157, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.59}) +[2026-06-10 15:35:10,880][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.22680304944515228, 'train/grad_norm_pre_clip': 0.20036137104034424, 'timing/data': 0.0005180798470973969, 'timing/model': 0.9306638091802597, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.6}) +[2026-06-10 15:35:19,032][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.0015499457949772477, 'train/grad_norm_pre_clip': 0.00865671131759882, 'timing/data': 0.24128110706806183, 'timing/model': 0.38647959008812904, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.62}) +[2026-06-10 15:35:37,638][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.0018641134956851602, 'train/grad_norm_pre_clip': 0.010417983867228031, 'timing/data': 0.016286339610815048, 'timing/model': 0.9419924542307854, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.63}) +[2026-06-10 15:35:52,809][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.5272336602210999, 'train/grad_norm_pre_clip': 0.35131993889808655, 'timing/data': 0.011872518807649612, 'timing/model': 0.9766516238451004, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.64}) +[2026-06-10 15:36:10,574][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.00294292694889009, 'train/grad_norm_pre_clip': 0.016835803166031837, 'timing/data': 0.01920359581708908, 'timing/model': 1.0057260319590569, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.66}) +[2026-06-10 15:36:26,304][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.0037002565804868937, 'train/grad_norm_pre_clip': 0.02131272479891777, 'timing/data': 0.017692435532808304, 'timing/model': 1.060898069292307, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.67}) +[2026-06-10 15:36:35,898][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpwqxsn8uy/test.c -o /tmp/tmpwqxsn8uy/test.o +[2026-06-10 15:36:35,911][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpwqxsn8uy/test.o -laio -o /tmp/tmpwqxsn8uy/a.out +[2026-06-10 15:36:36,539][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmp5uu9dqy1/test.c -o /tmp/tmp5uu9dqy1/test.o +[2026-06-10 15:36:36,553][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmp5uu9dqy1/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp5uu9dqy1/a.out +[2026-06-10 15:36:36,588][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpp9owoyui/test.c -o /tmp/tmpp9owoyui/test.o +[2026-06-10 15:36:36,604][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpp9owoyui/test.o -laio -o /tmp/tmpp9owoyui/a.out +[2026-06-10 15:36:49,870][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/best_state +[2026-06-10 15:36:49,875][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 15:36:58,951][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/best_state/model.safetensors +[2026-06-10 15:37:18,235][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/best_state/optimizer.bin +[2026-06-10 15:37:18,239][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/best_state/sampler.bin +[2026-06-10 15:37:18,240][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/best_state/sampler_1.bin +[2026-06-10 15:37:18,244][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/best_state/random_states_0.pkl +[2026-06-10 15:37:18,269][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-10 15:37:18,269][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 50 +[2026-06-10 15:37:18,272][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.004467309918254614, 'train/grad_norm_pre_clip': 0.025899028405547142, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 257.1217102050781, 'rl_games_eval/mid_train/mean_length': 19.8, 'rl_games_eval/mid_train/std_reward': 254.82196354391502, 'rl_games_eval/mid_train/std_length': 9.410632284814874, 'rl_games_eval/mid_train/macro_mean_reward': 257.1217102050781, 'rl_games_eval/mid_train/macro_mean_length': 19.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_reward': 257.1217102050781, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_length': 19.8, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_reward': 254.82196354391502, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_length': 9.410632284814874, 'rl_games_eval/mid_train/deadly_corridor__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 257.1217102050781, 'checkpoint/best_score': 257.1217102050781, 'checkpoint/best_step': 50.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.08757447451353073, 'timing/model': 0.4681384563446045, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.68}) +[2026-06-10 15:37:18,272][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_50_state +[2026-06-10 15:37:18,280][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 15:37:27,440][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_50_state/model.safetensors +[2026-06-10 15:37:43,066][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_50_state/optimizer.bin +[2026-06-10 15:37:43,067][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_50_state/sampler.bin +[2026-06-10 15:37:43,067][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_50_state/sampler_1.bin +[2026-06-10 15:37:43,070][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_50_state/random_states_0.pkl +[2026-06-10 15:37:43,075][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 50 +[2026-06-10 15:37:43,075][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 15:37:43,088][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 15:38:04,608][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.005381704308092594, 'train/grad_norm_pre_clip': 0.03115997649729252, 'timing/data': 0.03922787308692932, 'timing/model': 1.681797094643116, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.7}) +[2026-06-10 15:38:25,329][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.006233999039977789, 'train/grad_norm_pre_clip': 0.0361686535179615, 'timing/data': 0.03386421874165535, 'timing/model': 1.0254963301122189, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.71}) +[2026-06-10 15:38:41,088][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.0069456822238862514, 'train/grad_norm_pre_clip': 0.04028130695223808, 'timing/data': 0.02811519056558609, 'timing/model': 0.968523770570755, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.73}) +[2026-06-10 15:38:57,846][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.007588651496917009, 'train/grad_norm_pre_clip': 0.04406380280852318, 'timing/data': 0.032421376556158066, 'timing/model': 0.9449989944696426, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.74}) +[2026-06-10 15:39:05,815][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.008144685998558998, 'train/grad_norm_pre_clip': 0.047300249338150024, 'timing/data': 0.09213335812091827, 'timing/model': 0.5009769052267075, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.75}) +[2026-06-10 15:39:26,916][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.4946633577346802, 'train/grad_norm_pre_clip': 0.4127407968044281, 'timing/data': 0.008396565914154053, 'timing/model': 0.9883642420172691, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.77}) +[2026-06-10 15:39:44,990][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.44645172357559204, 'train/grad_norm_pre_clip': 0.3196694850921631, 'timing/data': 0.007937777787446976, 'timing/model': 0.9891372099518776, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.78}) +[2026-06-10 15:40:06,781][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.20656761527061462, 'train/grad_norm_pre_clip': 0.17459283769130707, 'timing/data': 0.018323231488466263, 'timing/model': 1.117833312600851, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.79}) +[2026-06-10 15:40:26,118][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.22866858541965485, 'train/grad_norm_pre_clip': 0.17532441020011902, 'timing/data': 0.013638082891702652, 'timing/model': 0.9589397795498371, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.81}) +[2026-06-10 15:40:34,416][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.0177469439804554, 'train/grad_norm_pre_clip': 0.10009825974702835, 'timing/data': 0.08951735869050026, 'timing/model': 0.6934634484350681, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.82}) +[2026-06-10 15:40:57,458][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.02016899734735489, 'train/grad_norm_pre_clip': 0.11210630089044571, 'timing/data': 0.03248879685997963, 'timing/model': 2.000125329941511, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.84}) +[2026-06-10 15:41:19,881][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.021637864410877228, 'train/grad_norm_pre_clip': 0.11887708306312561, 'timing/data': 0.037002261728048325, 'timing/model': 0.9320480264723301, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.85}) +[2026-06-10 15:41:36,404][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.020672213286161423, 'train/grad_norm_pre_clip': 0.11192070692777634, 'timing/data': 0.031304631382226944, 'timing/model': 0.9214237742125988, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.86}) +[2026-06-10 15:41:52,088][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.3200724720954895, 'train/grad_norm_pre_clip': 0.2617736756801605, 'timing/data': 0.02698393166065216, 'timing/model': 0.908314973115921, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.88}) +[2026-06-10 15:42:01,982][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.020922930911183357, 'train/grad_norm_pre_clip': 0.11171909421682358, 'timing/data': 0.13280876353383064, 'timing/model': 0.3693968206644058, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.89}) +[2026-06-10 15:42:25,699][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.1725761592388153, 'train/grad_norm_pre_clip': 0.13618852198123932, 'timing/data': 0.19635208323597908, 'timing/model': 2.997271489351988, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.9}) +[2026-06-10 15:42:41,259][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.3233376443386078, 'train/grad_norm_pre_clip': 0.45769020915031433, 'timing/data': 0.03661217540502548, 'timing/model': 1.018345095217228, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.92}) +[2026-06-10 15:43:00,604][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.16373658180236816, 'train/grad_norm_pre_clip': 0.20601323246955872, 'timing/data': 0.05065447837114334, 'timing/model': 3.1920444406569004, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.93}) +[2026-06-10 15:43:17,472][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.037274088710546494, 'train/grad_norm_pre_clip': 0.2023499608039856, 'timing/data': 0.047608450055122375, 'timing/model': 1.497961312532425, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.95}) +[2026-06-10 15:43:28,679][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.040965307503938675, 'train/grad_norm_pre_clip': 0.23082280158996582, 'timing/data': 0.09190371632575989, 'timing/model': 0.5456654652953148, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.96}) +[2026-06-10 15:43:50,892][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.15796241164207458, 'train/grad_norm_pre_clip': 0.17177096009254456, 'timing/data': 0.020821131765842438, 'timing/model': 0.9716170690953732, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.97}) +[2026-06-10 15:44:07,103][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.42568695545196533, 'train/grad_norm_pre_clip': 0.46632513403892517, 'timing/data': 0.016508545726537704, 'timing/model': 1.1040709055960178, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.99}) +[2026-06-10 15:44:24,080][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.3002372086048126, 'train/grad_norm_pre_clip': 0.22647881507873535, 'timing/data': 0.012640342116355896, 'timing/model': 0.953368715941906, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 1.0}) +[2026-06-10 15:44:43,394][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.327877402305603, 'train/grad_norm_pre_clip': 0.27178719639778137, 'timing/data': 0.01737476885318756, 'timing/model': 0.9650998748838902, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 1.01}) +[2026-06-10 15:44:51,178][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.06978661566972733, 'train/grad_norm_pre_clip': 0.40291091799736023, 'timing/data': 0.07360208407044411, 'timing/model': 0.35247503966093063, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 1.03}) +[2026-06-10 15:45:10,029][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 0.0536937452852726, 'train/grad_norm_pre_clip': 0.3084329068660736, 'timing/data': 0.023793216794729233, 'timing/model': 1.3907920271158218, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 1.04}) +[2026-06-10 15:45:29,981][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.03805062919855118, 'train/grad_norm_pre_clip': 0.21825255453586578, 'timing/data': 0.02374335750937462, 'timing/model': 1.9313336908817291, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 1.05}) +[2026-06-10 15:45:47,971][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.15089870989322662, 'train/grad_norm_pre_clip': 0.24367155134677887, 'timing/data': 0.01937428116798401, 'timing/model': 1.0579011738300323, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 1.07}) +[2026-06-10 15:46:04,655][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.1447756588459015, 'train/grad_norm_pre_clip': 0.1907799392938614, 'timing/data': 0.02607167512178421, 'timing/model': 0.9883635453879833, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 1.08}) +[2026-06-10 15:46:16,078][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.018487369641661644, 'train/grad_norm_pre_clip': 0.1044689491391182, 'timing/data': 0.07308975607156754, 'timing/model': 0.7763701044023037, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 1.1}) +[2026-06-10 15:46:38,481][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.014106720685958862, 'train/grad_norm_pre_clip': 0.08114352822303772, 'timing/data': 0.027469810098409653, 'timing/model': 0.9013592973351479, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 1.11}) +[2026-06-10 15:46:53,576][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.16191227734088898, 'train/grad_norm_pre_clip': 0.23082922399044037, 'timing/data': 0.022065773606300354, 'timing/model': 0.9962559305131435, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 1.12}) +[2026-06-10 15:47:11,082][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.007930872030556202, 'train/grad_norm_pre_clip': 0.04625186696648598, 'timing/data': 0.0006031356751918793, 'timing/model': 0.9626811295747757, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 1.14}) +[2026-06-10 15:47:28,303][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.006117596756666899, 'train/grad_norm_pre_clip': 0.034375421702861786, 'timing/data': 0.0005069784820079803, 'timing/model': 0.9006800875067711, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 1.15}) +[2026-06-10 15:47:36,903][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.005304380785673857, 'train/grad_norm_pre_clip': 0.0435841903090477, 'timing/data': 0.07472135126590729, 'timing/model': 0.35654500126838684, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 1.16}) +[2026-06-10 15:48:05,385][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.22263701260089874, 'train/grad_norm_pre_clip': 0.212443009018898, 'timing/data': 0.02491360902786255, 'timing/model': 1.2971761710941792, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 1.18}) +[2026-06-10 15:48:23,821][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.48861435055732727, 'train/grad_norm_pre_clip': 0.37975677847862244, 'timing/data': 0.016285110265016556, 'timing/model': 0.9618513844907284, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 1.19}) +[2026-06-10 15:48:38,881][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.005324311554431915, 'train/grad_norm_pre_clip': 0.03649388998746872, 'timing/data': 0.022880088537931442, 'timing/model': 0.7995105050504208, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 1.21}) +[2026-06-10 15:48:57,336][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.00657261535525322, 'train/grad_norm_pre_clip': 0.04413581266999245, 'timing/data': 0.025751236826181412, 'timing/model': 0.9234252274036407, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 1.22}) +[2026-06-10 15:49:05,197][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.007562023587524891, 'train/grad_norm_pre_clip': 0.069313645362854, 'timing/data': 0.07446328178048134, 'timing/model': 0.36020294576883316, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 1.23}) +[2026-06-10 15:49:28,194][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.1558494120836258, 'train/grad_norm_pre_clip': 0.22370968759059906, 'timing/data': 0.11344525218009949, 'timing/model': 1.755329366773367, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 1.25}) +[2026-06-10 15:49:51,329][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.226729154586792, 'train/grad_norm_pre_clip': 0.263700395822525, 'timing/data': 0.027608491480350494, 'timing/model': 1.1978408582508564, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 1.26}) +[2026-06-10 15:50:07,734][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.015016401186585426, 'train/grad_norm_pre_clip': 0.12138697504997253, 'timing/data': 0.022547181695699692, 'timing/model': 0.9477734379470348, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 1.27}) +[2026-06-10 15:50:25,143][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.1855359971523285, 'train/grad_norm_pre_clip': 0.2403213232755661, 'timing/data': 0.031010430306196213, 'timing/model': 0.9667121954262257, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 1.29}) +[2026-06-10 15:50:34,382][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.010294774547219276, 'train/grad_norm_pre_clip': 0.07881255447864532, 'timing/data': 0.07708361372351646, 'timing/model': 0.43895405530929565, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 1.3}) +[2026-06-10 15:50:56,525][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.019269295036792755, 'train/grad_norm_pre_clip': 0.17266924679279327, 'timing/data': 0.01959129422903061, 'timing/model': 0.9768020249903202, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 1.32}) +[2026-06-10 15:51:12,113][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.014054132625460625, 'train/grad_norm_pre_clip': 0.09397834539413452, 'timing/data': 0.019441694021224976, 'timing/model': 1.0739290602505207, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 1.33}) +[2026-06-10 15:51:28,756][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.014399031177163124, 'train/grad_norm_pre_clip': 0.10447793453931808, 'timing/data': 0.01931571587920189, 'timing/model': 1.561876229941845, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 1.34}) +[2026-06-10 15:51:47,759][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.17038172483444214, 'train/grad_norm_pre_clip': 0.5528930425643921, 'timing/data': 0.015978768467903137, 'timing/model': 1.2118704542517662, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 1.36}) +[2026-06-10 15:52:15,128][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 100 +[2026-06-10 15:52:15,130][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.016063403338193893, 'train/grad_norm_pre_clip': 0.10611450672149658, 'eval/loss': 0.13268058001995087, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 177.52330017089844, 'rl_games_eval/mid_train/mean_length': 16.0, 'rl_games_eval/mid_train/std_reward': 99.05187072760963, 'rl_games_eval/mid_train/std_length': 4.09878030638384, 'rl_games_eval/mid_train/macro_mean_reward': 177.52330017089844, 'rl_games_eval/mid_train/macro_mean_length': 16.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_reward': 177.52330017089844, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_length': 16.0, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_reward': 99.05187072760963, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_length': 4.09878030638384, 'rl_games_eval/mid_train/deadly_corridor__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 177.52330017089844, 'checkpoint/best_score': 257.1217102050781, 'checkpoint/best_step': 50.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0711691826581955, 'timing/model': 0.36381779983639717, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 1.37}) +[2026-06-10 15:52:15,131][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_100_state +[2026-06-10 15:52:15,136][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 15:52:23,321][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_100_state/model.safetensors +[2026-06-10 15:52:50,711][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_100_state/optimizer.bin +[2026-06-10 15:52:50,713][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_100_state/sampler.bin +[2026-06-10 15:52:50,713][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_100_state/sampler_1.bin +[2026-06-10 15:52:50,715][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-10 15:52:50,721][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-10 15:52:50,721][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 15:52:50,736][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 15:53:18,714][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.012413701973855495, 'train/grad_norm_pre_clip': 0.07521940767765045, 'timing/data': 0.02200167626142502, 'timing/model': 1.3133537024259567, 'learning_rate/qwen_vl_interface': 9.999853498812753e-06, 'learning_rate/action_model': 9.999853498812753e-05, 'epoch': 1.38}) +[2026-06-10 15:53:41,309][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.009903749451041222, 'train/grad_norm_pre_clip': 0.059719931334257126, 'timing/data': 0.025993939489126205, 'timing/model': 1.4450918398797512, 'learning_rate/qwen_vl_interface': 9.999414004287889e-06, 'learning_rate/action_model': 9.999414004287888e-05, 'epoch': 1.4}) +[2026-06-10 15:54:02,328][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.00794594082981348, 'train/grad_norm_pre_clip': 0.04970991611480713, 'timing/data': 0.026407048106193542, 'timing/model': 1.1662334874272346, 'learning_rate/qwen_vl_interface': 9.998681543535504e-06, 'learning_rate/action_model': 9.998681543535502e-05, 'epoch': 1.41}) +[2026-06-10 15:54:34,074][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.20540902018547058, 'train/grad_norm_pre_clip': 0.22446772456169128, 'timing/data': 0.019100435078144073, 'timing/model': 1.2115152589976788, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 1.42}) +[2026-06-10 15:54:48,410][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.00619363272562623, 'train/grad_norm_pre_clip': 0.03895643725991249, 'timing/data': 0.24434442445635796, 'timing/model': 0.9623892679810524, 'learning_rate/qwen_vl_interface': 9.996337922143435e-06, 'learning_rate/action_model': 9.996337922143436e-05, 'epoch': 1.44}) +[2026-06-10 15:55:13,817][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.007012228947132826, 'train/grad_norm_pre_clip': 0.044541049748659134, 'timing/data': 0.03275586664676666, 'timing/model': 3.541928607970476, 'learning_rate/qwen_vl_interface': 9.994726906069359e-06, 'learning_rate/action_model': 9.994726906069358e-05, 'epoch': 1.45}) +[2026-06-10 15:55:39,479][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.19185732305049896, 'train/grad_norm_pre_clip': 0.7710397839546204, 'timing/data': 0.022633593529462814, 'timing/model': 1.1923152543604374, 'learning_rate/qwen_vl_interface': 9.992823212890057e-06, 'learning_rate/action_model': 9.992823212890057e-05, 'epoch': 1.47}) +[2026-06-10 15:56:00,407][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.36285629868507385, 'train/grad_norm_pre_clip': 0.5055741667747498, 'timing/data': 0.020121589303016663, 'timing/model': 1.1269811131060123, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 1.48}) +[2026-06-10 15:56:22,429][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.010523502714931965, 'train/grad_norm_pre_clip': 0.0683899074792862, 'timing/data': 0.024124979972839355, 'timing/model': 1.5193144865334034, 'learning_rate/qwen_vl_interface': 9.988138282977284e-06, 'learning_rate/action_model': 9.988138282977283e-05, 'epoch': 1.49}) +[2026-06-10 15:56:35,344][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.023233210667967796, 'train/grad_norm_pre_clip': 0.24683193862438202, 'timing/data': 0.09901111572980881, 'timing/model': 0.4333251155912876, 'learning_rate/qwen_vl_interface': 9.985357335232359e-06, 'learning_rate/action_model': 9.98535733523236e-05, 'epoch': 1.51}) +[2026-06-10 15:56:56,940][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.02323797531425953, 'train/grad_norm_pre_clip': 0.2165258228778839, 'timing/data': 0.016789797693490982, 'timing/model': 1.2198389917612076, 'learning_rate/qwen_vl_interface': 9.982284288341471e-06, 'learning_rate/action_model': 9.982284288341471e-05, 'epoch': 1.52}) +[2026-06-10 15:57:19,625][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.1506246030330658, 'train/grad_norm_pre_clip': 0.22135894000530243, 'timing/data': 0.010251674801111221, 'timing/model': 1.288187451660633, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 1.53}) +[2026-06-10 15:57:40,204][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.025595488026738167, 'train/grad_norm_pre_clip': 0.15611791610717773, 'timing/data': 0.013314608484506607, 'timing/model': 1.3673527166247368, 'learning_rate/qwen_vl_interface': 9.975262673368201e-06, 'learning_rate/action_model': 9.9752626733682e-05, 'epoch': 1.55}) +[2026-06-10 15:58:01,697][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.01960364356637001, 'train/grad_norm_pre_clip': 0.14231246709823608, 'timing/data': 0.012962397187948227, 'timing/model': 1.3073198571801186, 'learning_rate/qwen_vl_interface': 9.971314538412104e-06, 'learning_rate/action_model': 9.971314538412104e-05, 'epoch': 1.56}) +[2026-06-10 15:58:12,227][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.01629207842051983, 'train/grad_norm_pre_clip': 0.17056278884410858, 'timing/data': 0.1130155436694622, 'timing/model': 0.4731622226536274, 'learning_rate/qwen_vl_interface': 9.967075170535901e-06, 'learning_rate/action_model': 9.967075170535901e-05, 'epoch': 1.58}) +[2026-06-10 15:58:42,265][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.008752290159463882, 'train/grad_norm_pre_clip': 0.06108246371150017, 'timing/data': 0.00890066847205162, 'timing/model': 5.831273376941681, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 1.59}) +[2026-06-10 15:59:08,697][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.1473693698644638, 'train/grad_norm_pre_clip': 0.283253937959671, 'timing/data': 0.011504452675580978, 'timing/model': 1.3088728860020638, 'learning_rate/qwen_vl_interface': 9.95772379998838e-06, 'learning_rate/action_model': 9.957723799988381e-05, 'epoch': 1.6}) +[2026-06-10 15:59:28,133][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.005280571058392525, 'train/grad_norm_pre_clip': 0.03817056491971016, 'timing/data': 0.008483462035655975, 'timing/model': 1.1971771158277988, 'learning_rate/qwen_vl_interface': 9.95261237415365e-06, 'learning_rate/action_model': 9.952612374153649e-05, 'epoch': 1.62}) +[2026-06-10 15:59:47,716][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.0038910643197596073, 'train/grad_norm_pre_clip': 0.02426910400390625, 'timing/data': 0.009871985763311386, 'timing/model': 1.1529134400188923, 'learning_rate/qwen_vl_interface': 9.9472108690364e-06, 'learning_rate/action_model': 9.947210869036399e-05, 'epoch': 1.63}) +[2026-06-10 15:59:57,980][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.003884564619511366, 'train/grad_norm_pre_clip': 0.02880481816828251, 'timing/data': 0.08930604159832001, 'timing/model': 0.45303141698241234, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 1.64}) +[2026-06-10 16:00:20,593][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.5460001230239868, 'train/grad_norm_pre_clip': 0.6502519249916077, 'timing/data': 0.018129266798496246, 'timing/model': 1.1844494082033634, 'learning_rate/qwen_vl_interface': 9.935538971588355e-06, 'learning_rate/action_model': 9.935538971588353e-05, 'epoch': 1.66}) +[2026-06-10 16:00:46,963][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.17055504024028778, 'train/grad_norm_pre_clip': 0.2648599445819855, 'timing/data': 0.01667475327849388, 'timing/model': 1.413946732878685, 'learning_rate/qwen_vl_interface': 9.929269299235177e-06, 'learning_rate/action_model': 9.929269299235176e-05, 'epoch': 1.67}) +[2026-06-10 16:01:07,911][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.42158329486846924, 'train/grad_norm_pre_clip': 0.4405122697353363, 'timing/data': 0.015283279120922089, 'timing/model': 1.3892500065267086, 'learning_rate/qwen_vl_interface': 9.922710987510297e-06, 'learning_rate/action_model': 9.922710987510296e-05, 'epoch': 1.68}) +[2026-06-10 16:01:30,982][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.27696385979652405, 'train/grad_norm_pre_clip': 0.43869706988334656, 'timing/data': 0.024819478392601013, 'timing/model': 1.1186934746801853, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 1.7}) +[2026-06-10 16:01:46,837][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.017787061631679535, 'train/grad_norm_pre_clip': 0.1283324658870697, 'timing/data': 0.09426369518041611, 'timing/model': 0.9224562458693981, 'learning_rate/qwen_vl_interface': 9.908730081915346e-06, 'learning_rate/action_model': 9.908730081915345e-05, 'epoch': 1.71}) +[2026-06-10 16:02:06,986][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.36682334542274475, 'train/grad_norm_pre_clip': 0.6441139578819275, 'timing/data': 0.03582886606454849, 'timing/model': 1.5917059145867825, 'learning_rate/qwen_vl_interface': 9.901308350453388e-06, 'learning_rate/action_model': 9.901308350453388e-05, 'epoch': 1.73}) +[2026-06-10 16:02:26,784][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.04601176455616951, 'train/grad_norm_pre_clip': 0.2955017685890198, 'timing/data': 0.0427645705640316, 'timing/model': 1.1984298042953014, 'learning_rate/qwen_vl_interface': 9.893599704382757e-06, 'learning_rate/action_model': 9.893599704382757e-05, 'epoch': 1.74}) +[2026-06-10 16:02:47,674][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.1530558317899704, 'train/grad_norm_pre_clip': 0.17749102413654327, 'timing/data': 0.0384148433804512, 'timing/model': 2.0732519887387753, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 1.75}) +[2026-06-10 16:03:07,265][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.05876196548342705, 'train/grad_norm_pre_clip': 0.30736932158470154, 'timing/data': 0.05663498118519783, 'timing/model': 1.540573038160801, 'learning_rate/qwen_vl_interface': 9.877323588106785e-06, 'learning_rate/action_model': 9.877323588106785e-05, 'epoch': 1.77}) +[2026-06-10 16:03:19,394][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.02514059655368328, 'train/grad_norm_pre_clip': 0.1756489872932434, 'timing/data': 0.06885478273034096, 'timing/model': 0.49520938470959663, 'learning_rate/qwen_vl_interface': 9.868757121888965e-06, 'learning_rate/action_model': 9.868757121888964e-05, 'epoch': 1.78}) +[2026-06-10 16:03:45,400][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.026817595586180687, 'train/grad_norm_pre_clip': 0.16392932832241058, 'timing/data': 0.01273958757519722, 'timing/model': 1.3632354587316513, 'learning_rate/qwen_vl_interface': 9.859905748975577e-06, 'learning_rate/action_model': 9.859905748975578e-05, 'epoch': 1.79}) +[2026-06-10 16:04:07,029][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.18062442541122437, 'train/grad_norm_pre_clip': 0.22836844623088837, 'timing/data': 0.02415987104177475, 'timing/model': 1.2305647656321526, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 1.81}) +[2026-06-10 16:04:27,809][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.01153993047773838, 'train/grad_norm_pre_clip': 0.5476459264755249, 'timing/data': 0.01807619258761406, 'timing/model': 1.2437100186944008, 'learning_rate/qwen_vl_interface': 9.84135048458031e-06, 'learning_rate/action_model': 9.84135048458031e-05, 'epoch': 1.82}) +[2026-06-10 16:04:48,022][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.2227020412683487, 'train/grad_norm_pre_clip': 1.1358047723770142, 'timing/data': 0.03130003437399864, 'timing/model': 1.1609914302825928, 'learning_rate/qwen_vl_interface': 9.83164773767454e-06, 'learning_rate/action_model': 9.83164773767454e-05, 'epoch': 1.84}) +[2026-06-10 16:04:58,040][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.011033539660274982, 'train/grad_norm_pre_clip': 0.08329262584447861, 'timing/data': 0.08628806099295616, 'timing/model': 0.6628074310719967, 'learning_rate/qwen_vl_interface': 9.821662373154824e-06, 'learning_rate/action_model': 9.821662373154825e-05, 'epoch': 1.85}) +[2026-06-10 16:05:24,613][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.008918865583837032, 'train/grad_norm_pre_clip': 0.060389887541532516, 'timing/data': 0.04555502533912659, 'timing/model': 1.2667676247656345, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 1.86}) +[2026-06-10 16:05:48,505][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.18928936123847961, 'train/grad_norm_pre_clip': 0.18340922892093658, 'timing/data': 0.03057761862874031, 'timing/model': 1.188332386314869, 'learning_rate/qwen_vl_interface': 9.800846272446017e-06, 'learning_rate/action_model': 9.800846272446018e-05, 'epoch': 1.88}) +[2026-06-10 16:06:07,663][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.005752375815063715, 'train/grad_norm_pre_clip': 0.048428718000650406, 'timing/data': 0.0289970301091671, 'timing/model': 1.4803376346826553, 'learning_rate/qwen_vl_interface': 9.790016820292068e-06, 'learning_rate/action_model': 9.790016820292069e-05, 'epoch': 1.89}) +[2026-06-10 16:06:30,411][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.1970745474100113, 'train/grad_norm_pre_clip': 0.26571375131607056, 'timing/data': 0.037440378218889236, 'timing/model': 1.1276446878910065, 'learning_rate/qwen_vl_interface': 9.778907318515255e-06, 'learning_rate/action_model': 9.778907318515255e-05, 'epoch': 1.9}) +[2026-06-10 16:06:40,575][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.007368098944425583, 'train/grad_norm_pre_clip': 0.06931032985448837, 'timing/data': 0.14091192185878754, 'timing/model': 0.625000324100256, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 1.92}) +[2026-06-10 16:07:01,089][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.19168540835380554, 'train/grad_norm_pre_clip': 0.37962448596954346, 'timing/data': 0.0006670355796813965, 'timing/model': 1.157172530889511, 'learning_rate/qwen_vl_interface': 9.755850924471148e-06, 'learning_rate/action_model': 9.755850924471147e-05, 'epoch': 1.93}) +[2026-06-10 16:07:22,398][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.003533719340339303, 'train/grad_norm_pre_clip': 0.04313391447067261, 'timing/data': 0.0007835961878299713, 'timing/model': 1.3479439541697502, 'learning_rate/qwen_vl_interface': 9.74390545443084e-06, 'learning_rate/action_model': 9.74390545443084e-05, 'epoch': 1.95}) +[2026-06-10 16:07:38,067][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.007080843672156334, 'train/grad_norm_pre_clip': 0.0719887763261795, 'timing/data': 0.0005538873374462128, 'timing/model': 0.9253846369683743, 'learning_rate/qwen_vl_interface': 9.73168277913392e-06, 'learning_rate/action_model': 9.73168277913392e-05, 'epoch': 1.96}) +[2026-06-10 16:07:52,663][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.16909006237983704, 'train/grad_norm_pre_clip': 0.20407941937446594, 'timing/data': 0.0005225911736488342, 'timing/model': 0.918618518859148, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 1.97}) +[2026-06-10 16:08:00,819][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.0048301066271960735, 'train/grad_norm_pre_clip': 0.06160793825984001, 'timing/data': 0.07424857839941978, 'timing/model': 0.366643525660038, 'learning_rate/qwen_vl_interface': 9.7064088456318e-06, 'learning_rate/action_model': 9.7064088456318e-05, 'epoch': 1.99}) +[2026-06-10 16:08:17,133][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.009028390981256962, 'train/grad_norm_pre_clip': 0.0834236666560173, 'timing/data': 0.0164174847304821, 'timing/model': 0.892997033894062, 'learning_rate/qwen_vl_interface': 9.693359146441872e-06, 'learning_rate/action_model': 9.693359146441872e-05, 'epoch': 2.0}) +[2026-06-10 16:08:33,923][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.15553750097751617, 'train/grad_norm_pre_clip': 0.2458219975233078, 'timing/data': 0.01780734211206436, 'timing/model': 1.014733463525772, 'learning_rate/qwen_vl_interface': 9.680035359929698e-06, 'learning_rate/action_model': 9.680035359929697e-05, 'epoch': 2.01}) +[2026-06-10 16:08:48,578][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.13068893551826477, 'train/grad_norm_pre_clip': 0.328093945980072, 'timing/data': 0.019108138978481293, 'timing/model': 0.8945554383099079, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 2.03}) +[2026-06-10 16:09:03,050][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.006350755225867033, 'train/grad_norm_pre_clip': 0.13419128954410553, 'timing/data': 0.015261363238096237, 'timing/model': 1.0585183277726173, 'learning_rate/qwen_vl_interface': 9.652568829290574e-06, 'learning_rate/action_model': 9.652568829290573e-05, 'epoch': 2.04}) +[2026-06-10 16:09:20,424][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 150 +[2026-06-10 16:09:20,426][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.026569943875074387, 'train/grad_norm_pre_clip': 0.2056470662355423, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/mean_length': 16.8, 'rl_games_eval/mid_train/std_reward': 164.1876467304262, 'rl_games_eval/mid_train/std_length': 5.114684741017769, 'rl_games_eval/mid_train/macro_mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/macro_mean_length': 16.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_length': 16.8, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_reward': 164.1876467304262, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_length': 5.114684741017769, 'rl_games_eval/mid_train/deadly_corridor__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 235.99388122558594, 'checkpoint/best_score': 257.1217102050781, 'checkpoint/best_step': 50.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.08039886504411697, 'timing/model': 0.3832947500050068, 'learning_rate/qwen_vl_interface': 9.638427779428613e-06, 'learning_rate/action_model': 9.638427779428613e-05, 'epoch': 2.05}) +[2026-06-10 16:09:20,427][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_150_state +[2026-06-10 16:09:20,432][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 16:09:27,065][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_150_state/model.safetensors +[2026-06-10 16:09:38,013][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_150_state/optimizer.bin +[2026-06-10 16:09:38,015][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_150_state/sampler.bin +[2026-06-10 16:09:38,015][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_150_state/sampler_1.bin +[2026-06-10 16:09:38,018][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_150_state/random_states_0.pkl +[2026-06-10 16:09:38,028][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 150 +[2026-06-10 16:09:38,028][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 16:09:38,044][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 16:10:02,417][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.007714506704360247, 'train/grad_norm_pre_clip': 0.2208782136440277, 'timing/data': 0.027252696454524994, 'timing/model': 1.1492266245186329, 'learning_rate/qwen_vl_interface': 9.624016030669876e-06, 'learning_rate/action_model': 9.624016030669876e-05, 'epoch': 2.07}) +[2026-06-10 16:10:21,373][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.010085075162351131, 'train/grad_norm_pre_clip': 0.1527489870786667, 'timing/data': 0.024976588785648346, 'timing/model': 1.252115674316883, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 2.08}) +[2026-06-10 16:10:40,304][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.17493261396884918, 'train/grad_norm_pre_clip': 1.4504605531692505, 'timing/data': 0.022269047796726227, 'timing/model': 1.4917309023439884, 'learning_rate/qwen_vl_interface': 9.594384009043415e-06, 'learning_rate/action_model': 9.594384009043416e-05, 'epoch': 2.1}) +[2026-06-10 16:10:59,539][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.15261311829090118, 'train/grad_norm_pre_clip': 1.4611446857452393, 'timing/data': 0.01966606080532074, 'timing/model': 1.3475856557488441, 'learning_rate/qwen_vl_interface': 9.579165564018366e-06, 'learning_rate/action_model': 9.579165564018366e-05, 'epoch': 2.11}) +[2026-06-10 16:11:12,626][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 1.2498867511749268, 'train/grad_norm_pre_clip': 2.3784162998199463, 'timing/data': 0.17739572748541832, 'timing/model': 1.3080599009990692, 'learning_rate/qwen_vl_interface': 9.563680075669138e-06, 'learning_rate/action_model': 9.563680075669137e-05, 'epoch': 2.12}) +[2026-06-10 16:11:38,372][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.018136216327548027, 'train/grad_norm_pre_clip': 0.5541039109230042, 'timing/data': 0.042140185832977295, 'timing/model': 1.1993047297000885, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 2.14}) +[2026-06-10 16:11:57,431][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.006999579723924398, 'train/grad_norm_pre_clip': 0.2197972983121872, 'timing/data': 0.05143669620156288, 'timing/model': 1.1867255941033363, 'learning_rate/qwen_vl_interface': 9.531911806283167e-06, 'learning_rate/action_model': 9.531911806283167e-05, 'epoch': 2.15}) +[2026-06-10 16:12:16,785][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.027178285643458366, 'train/grad_norm_pre_clip': 0.43800604343414307, 'timing/data': 0.04628331586718559, 'timing/model': 1.2601720243692398, 'learning_rate/qwen_vl_interface': 9.515630984862924e-06, 'learning_rate/action_model': 9.515630984862924e-05, 'epoch': 2.16}) +[2026-06-10 16:12:36,259][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.024554505944252014, 'train/grad_norm_pre_clip': 0.27881839871406555, 'timing/data': 0.03438105061650276, 'timing/model': 1.143825277686119, 'learning_rate/qwen_vl_interface': 9.49908703923062e-06, 'learning_rate/action_model': 9.49908703923062e-05, 'epoch': 2.18}) +[2026-06-10 16:12:48,863][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.015462146140635014, 'train/grad_norm_pre_clip': 0.18710778653621674, 'timing/data': 0.16089780256152153, 'timing/model': 0.4276486597955227, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 2.19}) +[2026-06-10 16:13:11,587][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.1714976578950882, 'train/grad_norm_pre_clip': 0.3048895299434662, 'timing/data': 0.048064298927783966, 'timing/model': 1.1694794595241547, 'learning_rate/qwen_vl_interface': 9.465213873531596e-06, 'learning_rate/action_model': 9.465213873531596e-05, 'epoch': 2.21}) +[2026-06-10 16:13:30,502][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.3966646194458008, 'train/grad_norm_pre_clip': 0.49821287393569946, 'timing/data': 0.043719664216041565, 'timing/model': 1.3323410525918007, 'learning_rate/qwen_vl_interface': 9.447886742921295e-06, 'learning_rate/action_model': 9.447886742921294e-05, 'epoch': 2.22}) +[2026-06-10 16:13:51,113][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.13942433893680573, 'train/grad_norm_pre_clip': 0.3411158323287964, 'timing/data': 0.037936754524707794, 'timing/model': 1.2049748748540878, 'learning_rate/qwen_vl_interface': 9.430300666882878e-06, 'learning_rate/action_model': 9.430300666882877e-05, 'epoch': 2.23}) +[2026-06-10 16:14:12,711][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.26355940103530884, 'train/grad_norm_pre_clip': 0.2513013482093811, 'timing/data': 0.05000976473093033, 'timing/model': 1.152851264923811, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 2.25}) +[2026-06-10 16:14:22,654][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.04038158059120178, 'train/grad_norm_pre_clip': 0.2543351352214813, 'timing/data': 0.07886937633156776, 'timing/model': 0.450802281498909, 'learning_rate/qwen_vl_interface': 9.394356033595787e-06, 'learning_rate/action_model': 9.394356033595787e-05, 'epoch': 2.26}) +[2026-06-10 16:14:45,135][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.20297189056873322, 'train/grad_norm_pre_clip': 0.7282475233078003, 'timing/data': 0.02238555997610092, 'timing/model': 1.3445090353488922, 'learning_rate/qwen_vl_interface': 9.37599969358141e-06, 'learning_rate/action_model': 9.375999693581409e-05, 'epoch': 2.27}) +[2026-06-10 16:15:08,483][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.04151805117726326, 'train/grad_norm_pre_clip': 0.24517521262168884, 'timing/data': 0.015359140932559967, 'timing/model': 1.1600130759179592, 'learning_rate/qwen_vl_interface': 9.357388842470736e-06, 'learning_rate/action_model': 9.357388842470737e-05, 'epoch': 2.29}) +[2026-06-10 16:15:37,291][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.028692910447716713, 'train/grad_norm_pre_clip': 0.1674327254295349, 'timing/data': 0.017536666244268417, 'timing/model': 1.1110703237354755, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 2.3}) +[2026-06-10 16:15:55,834][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.2741128206253052, 'train/grad_norm_pre_clip': 0.37430423498153687, 'timing/data': 0.017165370285511017, 'timing/model': 1.3824826888740063, 'learning_rate/qwen_vl_interface': 9.319408214608987e-06, 'learning_rate/action_model': 9.319408214608987e-05, 'epoch': 2.32}) +[2026-06-10 16:16:05,270][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.017407193779945374, 'train/grad_norm_pre_clip': 0.2127074897289276, 'timing/data': 0.08687800541520119, 'timing/model': 0.47208065539598465, 'learning_rate/qwen_vl_interface': 9.300040780681938e-06, 'learning_rate/action_model': 9.300040780681938e-05, 'epoch': 2.33}) +[2026-06-10 16:16:25,542][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.020536283031105995, 'train/grad_norm_pre_clip': 0.12238182872533798, 'timing/data': 0.02320254221558571, 'timing/model': 1.1260779313743114, 'learning_rate/qwen_vl_interface': 9.280423521162143e-06, 'learning_rate/action_model': 9.280423521162143e-05, 'epoch': 2.34}) +[2026-06-10 16:16:44,712][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.012212392874062061, 'train/grad_norm_pre_clip': 0.07407024502754211, 'timing/data': 0.038261640816926956, 'timing/model': 1.039051365107298, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 2.36}) +[2026-06-10 16:17:04,971][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.010125139728188515, 'train/grad_norm_pre_clip': 0.06144147366285324, 'timing/data': 0.03023134544491768, 'timing/model': 1.355154786258936, 'learning_rate/qwen_vl_interface': 9.240444381019974e-06, 'learning_rate/action_model': 9.240444381019974e-05, 'epoch': 2.37}) +[2026-06-10 16:17:23,312][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.006967976223677397, 'train/grad_norm_pre_clip': 0.04218485951423645, 'timing/data': 0.021521691232919693, 'timing/model': 1.1804757416248322, 'learning_rate/qwen_vl_interface': 9.220084966499285e-06, 'learning_rate/action_model': 9.220084966499286e-05, 'epoch': 2.38}) +[2026-06-10 16:17:34,486][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.0030935879331082106, 'train/grad_norm_pre_clip': 0.018510015681385994, 'timing/data': 0.08675698190927505, 'timing/model': 0.4542442783713341, 'learning_rate/qwen_vl_interface': 9.199480658437092e-06, 'learning_rate/action_model': 9.199480658437091e-05, 'epoch': 2.4}) +[2026-06-10 16:17:58,559][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.18626153469085693, 'train/grad_norm_pre_clip': 0.352193683385849, 'timing/data': 0.03455072641372681, 'timing/model': 2.105614088475704, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 2.41}) +[2026-06-10 16:18:17,616][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.003989160992205143, 'train/grad_norm_pre_clip': 0.026682354509830475, 'timing/data': 0.030678048729896545, 'timing/model': 1.1537837348878384, 'learning_rate/qwen_vl_interface': 9.157542460599085e-06, 'learning_rate/action_model': 9.157542460599085e-05, 'epoch': 2.42}) +[2026-06-10 16:18:44,621][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.0037373004015535116, 'train/grad_norm_pre_clip': 0.024218935519456863, 'timing/data': 0.11631385236978531, 'timing/model': 1.8143652267754078, 'learning_rate/qwen_vl_interface': 9.136211157768862e-06, 'learning_rate/action_model': 9.136211157768861e-05, 'epoch': 2.44}) +[2026-06-10 16:19:04,073][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.003393865190446377, 'train/grad_norm_pre_clip': 0.026479477062821388, 'timing/data': 0.035416554659605026, 'timing/model': 1.16181118786335, 'learning_rate/qwen_vl_interface': 9.11464013512874e-06, 'learning_rate/action_model': 9.11464013512874e-05, 'epoch': 2.45}) +[2026-06-10 16:19:13,993][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.0011307692620903254, 'train/grad_norm_pre_clip': 0.0068806749768555164, 'timing/data': 0.08865086361765862, 'timing/model': 0.5905025936663151, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 2.47}) +[2026-06-10 16:19:39,137][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.20866841077804565, 'train/grad_norm_pre_clip': 0.4857955873012543, 'timing/data': 0.018298093229532242, 'timing/model': 1.1368848457932472, 'learning_rate/qwen_vl_interface': 9.070784267532901e-06, 'learning_rate/action_model': 9.070784267532901e-05, 'epoch': 2.48}) +[2026-06-10 16:19:57,681][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.22120918333530426, 'train/grad_norm_pre_clip': 0.4926007390022278, 'timing/data': 0.014683928340673447, 'timing/model': 1.3939652256667614, 'learning_rate/qwen_vl_interface': 9.048502127813682e-06, 'learning_rate/action_model': 9.048502127813682e-05, 'epoch': 2.49}) +[2026-06-10 16:20:18,788][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.13751333951950073, 'train/grad_norm_pre_clip': 0.29317474365234375, 'timing/data': 0.022031348198652267, 'timing/model': 1.2032672204077244, 'learning_rate/qwen_vl_interface': 9.025985678590673e-06, 'learning_rate/action_model': 9.025985678590673e-05, 'epoch': 2.51}) +[2026-06-10 16:20:38,180][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.006033599842339754, 'train/grad_norm_pre_clip': 0.05219314619898796, 'timing/data': 0.019391220062971115, 'timing/model': 1.1912530325353146, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 2.52}) +[2026-06-10 16:20:48,685][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.008196750655770302, 'train/grad_norm_pre_clip': 0.13362865149974823, 'timing/data': 0.0997188426554203, 'timing/model': 0.454966027289629, 'learning_rate/qwen_vl_interface': 8.98025542168354e-06, 'learning_rate/action_model': 8.98025542168354e-05, 'epoch': 2.53}) +[2026-06-10 16:21:12,290][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.0143018439412117, 'train/grad_norm_pre_clip': 0.3408452868461609, 'timing/data': 0.008228879421949387, 'timing/model': 1.1610708646476269, 'learning_rate/qwen_vl_interface': 8.957044434857067e-06, 'learning_rate/action_model': 8.957044434857067e-05, 'epoch': 2.55}) +[2026-06-10 16:21:30,640][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.009974702261388302, 'train/grad_norm_pre_clip': 0.07438816130161285, 'timing/data': 0.012419208884239197, 'timing/model': 1.1209888607263565, 'learning_rate/qwen_vl_interface': 8.933604780068114e-06, 'learning_rate/action_model': 8.933604780068114e-05, 'epoch': 2.56}) +[2026-06-10 16:21:48,873][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.011535904370248318, 'train/grad_norm_pre_clip': 0.10423265397548676, 'timing/data': 0.016953513026237488, 'timing/model': 1.128194522112608, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 2.58}) +[2026-06-10 16:22:07,091][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.008090760558843613, 'train/grad_norm_pre_clip': 0.1021760106086731, 'timing/data': 0.008595272898674011, 'timing/model': 1.2071071118116379, 'learning_rate/qwen_vl_interface': 8.88604526409217e-06, 'learning_rate/action_model': 8.886045264092169e-05, 'epoch': 2.59}) +[2026-06-10 16:22:17,007][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.004879170563071966, 'train/grad_norm_pre_clip': 0.10041502863168716, 'timing/data': 0.09255144000053406, 'timing/model': 0.6728614717721939, 'learning_rate/qwen_vl_interface': 8.861928336600148e-06, 'learning_rate/action_model': 8.861928336600148e-05, 'epoch': 2.6}) +[2026-06-10 16:22:35,532][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.010540720075368881, 'train/grad_norm_pre_clip': 0.08409912884235382, 'timing/data': 0.03427110239863396, 'timing/model': 1.122438870370388, 'learning_rate/qwen_vl_interface': 8.837588608354623e-06, 'learning_rate/action_model': 8.837588608354624e-05, 'epoch': 2.62}) +[2026-06-10 16:22:55,111][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.008485097438097, 'train/grad_norm_pre_clip': 0.07879206538200378, 'timing/data': 0.036554377526044846, 'timing/model': 1.139834925532341, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 2.63}) +[2026-06-10 16:23:17,646][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.005844308529049158, 'train/grad_norm_pre_clip': 0.05089464783668518, 'timing/data': 0.03840041160583496, 'timing/model': 1.3453505262732506, 'learning_rate/qwen_vl_interface': 8.788246768810206e-06, 'learning_rate/action_model': 8.788246768810206e-05, 'epoch': 2.64}) +[2026-06-10 16:23:38,431][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.0056090946309268475, 'train/grad_norm_pre_clip': 0.06669917702674866, 'timing/data': 0.03655622899532318, 'timing/model': 1.1521492078900337, 'learning_rate/qwen_vl_interface': 8.763247701148398e-06, 'learning_rate/action_model': 8.763247701148397e-05, 'epoch': 2.66}) +[2026-06-10 16:23:47,987][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.003332731081172824, 'train/grad_norm_pre_clip': 0.043730102479457855, 'timing/data': 0.07317293807864189, 'timing/model': 0.46036266908049583, 'learning_rate/qwen_vl_interface': 8.738031919819508e-06, 'learning_rate/action_model': 8.738031919819508e-05, 'epoch': 2.67}) +[2026-06-10 16:24:09,817][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.14123345911502838, 'train/grad_norm_pre_clip': 0.6394193172454834, 'timing/data': 0.010512787848711014, 'timing/model': 1.2287147156894207, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 2.68}) +[2026-06-10 16:24:29,408][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.280281662940979, 'train/grad_norm_pre_clip': 0.3440137803554535, 'timing/data': 0.012015238404273987, 'timing/model': 1.4483543708920479, 'learning_rate/qwen_vl_interface': 8.686956451145135e-06, 'learning_rate/action_model': 8.686956451145134e-05, 'epoch': 2.7}) +[2026-06-10 16:24:47,849][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.0029701772145926952, 'train/grad_norm_pre_clip': 0.023520009592175484, 'timing/data': 0.006200313568115234, 'timing/model': 1.151880994439125, 'learning_rate/qwen_vl_interface': 8.661099914375148e-06, 'learning_rate/action_model': 8.661099914375147e-05, 'epoch': 2.71}) +[2026-06-10 16:25:13,331][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.13135980069637299, 'train/grad_norm_pre_clip': 0.7937515377998352, 'timing/data': 0.007037028670310974, 'timing/model': 1.130488358438015, 'learning_rate/qwen_vl_interface': 8.635032964894726e-06, 'learning_rate/action_model': 8.635032964894726e-05, 'epoch': 2.73}) +[2026-06-10 16:25:37,337][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 200 +[2026-06-10 16:25:37,338][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.7632240056991577, 'train/grad_norm_pre_clip': 1.5237421989440918, 'eval/loss': 0.11490369588136673, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/mean_length': 16.8, 'rl_games_eval/mid_train/std_reward': 164.1876467304262, 'rl_games_eval/mid_train/std_length': 5.114684741017769, 'rl_games_eval/mid_train/macro_mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/macro_mean_length': 16.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_length': 16.8, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_reward': 164.1876467304262, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_length': 5.114684741017769, 'rl_games_eval/mid_train/deadly_corridor__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 235.99388122558594, 'checkpoint/best_score': 257.1217102050781, 'checkpoint/best_step': 50.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.07264747470617294, 'timing/model': 0.44432028010487556, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 2.74}) +[2026-06-10 16:25:37,340][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_200_state +[2026-06-10 16:25:37,345][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 16:25:43,893][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_200_state/model.safetensors +[2026-06-10 16:25:55,306][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_200_state/optimizer.bin +[2026-06-10 16:25:55,307][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_200_state/sampler.bin +[2026-06-10 16:25:55,308][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_200_state/sampler_1.bin +[2026-06-10 16:25:55,310][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-10 16:25:55,328][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-10 16:25:55,329][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 16:25:55,342][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 16:26:17,863][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.006453613750636578, 'train/grad_norm_pre_clip': 0.04980089142918587, 'timing/data': 0.032101575285196304, 'timing/model': 1.1409121192991734, 'learning_rate/qwen_vl_interface': 8.582274272411566e-06, 'learning_rate/action_model': 8.582274272411566e-05, 'epoch': 2.75}) +[2026-06-10 16:26:35,990][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.017767636105418205, 'train/grad_norm_pre_clip': 0.13350577652454376, 'timing/data': 0.026578586548566818, 'timing/model': 1.1591313853859901, 'learning_rate/qwen_vl_interface': 8.555585783813493e-06, 'learning_rate/action_model': 8.555585783813493e-05, 'epoch': 2.77}) +[2026-06-10 16:26:54,149][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.19900229573249817, 'train/grad_norm_pre_clip': 0.3432444930076599, 'timing/data': 0.02662493661046028, 'timing/model': 1.1380437724292278, 'learning_rate/qwen_vl_interface': 8.528693391113577e-06, 'learning_rate/action_model': 8.528693391113577e-05, 'epoch': 2.78}) +[2026-06-10 16:27:13,022][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.02135656215250492, 'train/grad_norm_pre_clip': 0.2243938148021698, 'timing/data': 0.030536912381649017, 'timing/model': 1.299464575946331, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 2.79}) +[2026-06-10 16:27:22,531][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.06392215937376022, 'train/grad_norm_pre_clip': 0.512492299079895, 'timing/data': 0.08436140790581703, 'timing/model': 0.4621829204261303, 'learning_rate/qwen_vl_interface': 8.474303541281474e-06, 'learning_rate/action_model': 8.474303541281474e-05, 'epoch': 2.81}) +[2026-06-10 16:27:42,665][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.1766376793384552, 'train/grad_norm_pre_clip': 0.7083221077919006, 'timing/data': 0.03809168189764023, 'timing/model': 1.3073893710970879, 'learning_rate/qwen_vl_interface': 8.446809439171424e-06, 'learning_rate/action_model': 8.446809439171423e-05, 'epoch': 2.82}) +[2026-06-10 16:28:01,268][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.10928723961114883, 'train/grad_norm_pre_clip': 0.2210579812526703, 'timing/data': 0.03067239746451378, 'timing/model': 1.3107307516038418, 'learning_rate/qwen_vl_interface': 8.41911814279685e-06, 'learning_rate/action_model': 8.41911814279685e-05, 'epoch': 2.84}) +[2026-06-10 16:28:22,363][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.009652046486735344, 'train/grad_norm_pre_clip': 0.1122189462184906, 'timing/data': 0.036200832575559616, 'timing/model': 1.1295768804848194, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 2.85}) +[2026-06-10 16:28:42,126][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.02152906358242035, 'train/grad_norm_pre_clip': 0.160444974899292, 'timing/data': 0.03161481395363808, 'timing/model': 1.093492567539215, 'learning_rate/qwen_vl_interface': 8.363150811831022e-06, 'learning_rate/action_model': 8.36315081183102e-05, 'epoch': 2.86}) +[2026-06-10 16:28:51,801][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.013655503280460835, 'train/grad_norm_pre_clip': 0.164088174700737, 'timing/data': 0.07642970234155655, 'timing/model': 0.4528978057205677, 'learning_rate/qwen_vl_interface': 8.334878229568372e-06, 'learning_rate/action_model': 8.334878229568372e-05, 'epoch': 2.88}) +[2026-06-10 16:29:12,647][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.004883279092609882, 'train/grad_norm_pre_clip': 0.05257940664887428, 'timing/data': 0.03182714805006981, 'timing/model': 1.1778731048107147, 'learning_rate/qwen_vl_interface': 8.306415357485451e-06, 'learning_rate/action_model': 8.306415357485451e-05, 'epoch': 2.89}) +[2026-06-10 16:29:33,198][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.002974172355607152, 'train/grad_norm_pre_clip': 0.03154733404517174, 'timing/data': 0.042860303074121475, 'timing/model': 1.1473451890051365, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 2.9}) +[2026-06-10 16:29:52,103][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.006398382596671581, 'train/grad_norm_pre_clip': 0.08357887715101242, 'timing/data': 0.04455624148249626, 'timing/model': 1.1366027556359768, 'learning_rate/qwen_vl_interface': 8.248925778384545e-06, 'learning_rate/action_model': 8.248925778384545e-05, 'epoch': 2.92}) +[2026-06-10 16:30:19,752][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.3017650246620178, 'train/grad_norm_pre_clip': 0.7587185502052307, 'timing/data': 0.04417116567492485, 'timing/model': 1.1631984561681747, 'learning_rate/qwen_vl_interface': 8.219902617594601e-06, 'learning_rate/action_model': 8.219902617594602e-05, 'epoch': 2.93}) +[2026-06-10 16:30:29,163][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.001610336359590292, 'train/grad_norm_pre_clip': 0.027314335107803345, 'timing/data': 0.07842365652322769, 'timing/model': 0.44161245599389076, 'learning_rate/qwen_vl_interface': 8.190696259221713e-06, 'learning_rate/action_model': 8.190696259221713e-05, 'epoch': 2.95}) +[2026-06-10 16:30:49,702][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.6498162150382996, 'train/grad_norm_pre_clip': 1.011536955833435, 'timing/data': 0.024749089032411575, 'timing/model': 1.2597039937973022, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 2.96}) +[2026-06-10 16:31:09,791][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.0010494531597942114, 'train/grad_norm_pre_clip': 0.009554476477205753, 'timing/data': 0.0241534486413002, 'timing/model': 1.3026268891990185, 'learning_rate/qwen_vl_interface': 8.131741167259499e-06, 'learning_rate/action_model': 8.131741167259499e-05, 'epoch': 2.97}) +[2026-06-10 16:31:31,075][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.2330479919910431, 'train/grad_norm_pre_clip': 0.48817065358161926, 'timing/data': 0.027590088546276093, 'timing/model': 1.1392971463501453, 'learning_rate/qwen_vl_interface': 8.10199607029795e-06, 'learning_rate/action_model': 8.101996070297951e-05, 'epoch': 2.99}) +[2026-06-10 16:31:50,654][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.003334547160193324, 'train/grad_norm_pre_clip': 0.06761573255062103, 'timing/data': 0.024899523705244064, 'timing/model': 1.166015736758709, 'learning_rate/qwen_vl_interface': 8.072075048784685e-06, 'learning_rate/action_model': 8.072075048784683e-05, 'epoch': 3.0}) +[2026-06-10 16:32:01,001][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.010990443639457226, 'train/grad_norm_pre_clip': 0.3071775436401367, 'timing/data': 0.08052332326769829, 'timing/model': 0.4435633681714535, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 3.01}) +[2026-06-10 16:32:25,560][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.14198580384254456, 'train/grad_norm_pre_clip': 1.2755950689315796, 'timing/data': 0.02540026605129242, 'timing/model': 1.181239489465952, 'learning_rate/qwen_vl_interface': 8.011712625519198e-06, 'learning_rate/action_model': 8.011712625519196e-05, 'epoch': 3.03}) +[2026-06-10 16:32:44,566][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.008212253451347351, 'train/grad_norm_pre_clip': 0.06617137044668198, 'timing/data': 0.01677824929356575, 'timing/model': 1.1341507397592068, 'learning_rate/qwen_vl_interface': 7.981274947205574e-06, 'learning_rate/action_model': 7.981274947205573e-05, 'epoch': 3.04}) +[2026-06-10 16:33:02,709][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.28315433859825134, 'train/grad_norm_pre_clip': 0.7135084271430969, 'timing/data': 0.027403347194194794, 'timing/model': 1.1424373425543308, 'learning_rate/qwen_vl_interface': 7.950668790987754e-06, 'learning_rate/action_model': 7.950668790987755e-05, 'epoch': 3.05}) +[2026-06-10 16:33:21,405][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.030489442870020866, 'train/grad_norm_pre_clip': 0.31543004512786865, 'timing/data': 0.02194589376449585, 'timing/model': 1.2733984664082527, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 3.07}) +[2026-06-10 16:33:31,699][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.02228865586221218, 'train/grad_norm_pre_clip': 0.25421977043151855, 'timing/data': 0.08138654008507729, 'timing/model': 0.5503714866936207, 'learning_rate/qwen_vl_interface': 7.888958606843111e-06, 'learning_rate/action_model': 7.888958606843111e-05, 'epoch': 3.08}) +[2026-06-10 16:33:56,229][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.12840060889720917, 'train/grad_norm_pre_clip': 0.5731278657913208, 'timing/data': 0.022827237844467163, 'timing/model': 1.532789658755064, 'learning_rate/qwen_vl_interface': 7.857858385491127e-06, 'learning_rate/action_model': 7.857858385491127e-05, 'epoch': 3.1}) +[2026-06-10 16:34:16,315][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.15223370492458344, 'train/grad_norm_pre_clip': 0.43948400020599365, 'timing/data': 0.020561479032039642, 'timing/model': 1.1389730796217918, 'learning_rate/qwen_vl_interface': 7.826597299149815e-06, 'learning_rate/action_model': 7.826597299149815e-05, 'epoch': 3.11}) +[2026-06-10 16:34:34,832][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.034643761813640594, 'train/grad_norm_pre_clip': 1.0302256345748901, 'timing/data': 0.01612825319170952, 'timing/model': 1.114300835877657, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 3.12}) +[2026-06-10 16:34:53,319][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.03816164657473564, 'train/grad_norm_pre_clip': 0.2939571440219879, 'timing/data': 0.021740715950727463, 'timing/model': 1.3799742572009563, 'learning_rate/qwen_vl_interface': 7.763600254627406e-06, 'learning_rate/action_model': 7.763600254627405e-05, 'epoch': 3.14}) +[2026-06-10 16:35:02,921][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.018687952309846878, 'train/grad_norm_pre_clip': 0.2528648376464844, 'timing/data': 0.08418058604001999, 'timing/model': 0.46749312430620193, 'learning_rate/qwen_vl_interface': 7.731868182400758e-06, 'learning_rate/action_model': 7.731868182400758e-05, 'epoch': 3.15}) +[2026-06-10 16:35:23,459][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.012598274275660515, 'train/grad_norm_pre_clip': 0.25973740220069885, 'timing/data': 0.020071059465408325, 'timing/model': 1.32053904235363, 'learning_rate/qwen_vl_interface': 7.699983016853976e-06, 'learning_rate/action_model': 7.699983016853976e-05, 'epoch': 3.16}) +[2026-06-10 16:35:43,096][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.008790161460638046, 'train/grad_norm_pre_clip': 0.2447543889284134, 'timing/data': 0.018006552010774612, 'timing/model': 1.1352419294416904, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 3.18}) +[2026-06-10 16:36:03,650][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.15793238580226898, 'train/grad_norm_pre_clip': 3.7019600868225098, 'timing/data': 0.026163116097450256, 'timing/model': 1.2260773926973343, 'learning_rate/qwen_vl_interface': 7.635761282431021e-06, 'learning_rate/action_model': 7.635761282431019e-05, 'epoch': 3.19}) +[2026-06-10 16:36:23,451][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.00735175795853138, 'train/grad_norm_pre_clip': 0.6141111254692078, 'timing/data': 0.022166796028614044, 'timing/model': 1.7174817584455013, 'learning_rate/qwen_vl_interface': 7.603428675053936e-06, 'learning_rate/action_model': 7.603428675053937e-05, 'epoch': 3.21}) +[2026-06-10 16:36:33,927][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.0015571096446365118, 'train/grad_norm_pre_clip': 0.0159609392285347, 'timing/data': 0.10051709786057472, 'timing/model': 0.4533451609313488, 'learning_rate/qwen_vl_interface': 7.570950897110537e-06, 'learning_rate/action_model': 7.570950897110536e-05, 'epoch': 3.22}) +[2026-06-10 16:36:54,705][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.004749901592731476, 'train/grad_norm_pre_clip': 0.12905432283878326, 'timing/data': 0.03350871428847313, 'timing/model': 1.0543750412762165, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 3.23}) +[2026-06-10 16:37:14,876][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.24723854660987854, 'train/grad_norm_pre_clip': 0.8759916424751282, 'timing/data': 0.03096163272857666, 'timing/model': 1.4443059898912907, 'learning_rate/qwen_vl_interface': 7.505567851885325e-06, 'learning_rate/action_model': 7.505567851885325e-05, 'epoch': 3.25}) +[2026-06-10 16:37:33,145][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.0007128264987841249, 'train/grad_norm_pre_clip': 0.00705982930958271, 'timing/data': 0.030392568558454514, 'timing/model': 1.068273637443781, 'learning_rate/qwen_vl_interface': 7.472666617737725e-06, 'learning_rate/action_model': 7.472666617737725e-05, 'epoch': 3.26}) +[2026-06-10 16:37:51,700][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.21693487465381622, 'train/grad_norm_pre_clip': 0.4677640199661255, 'timing/data': 0.01984410732984543, 'timing/model': 1.1218982078135014, 'learning_rate/qwen_vl_interface': 7.439628279043447e-06, 'learning_rate/action_model': 7.439628279043447e-05, 'epoch': 3.27}) +[2026-06-10 16:38:03,142][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.003121444024145603, 'train/grad_norm_pre_clip': 0.14974847435951233, 'timing/data': 0.08852509781718254, 'timing/model': 0.5064493902027607, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 3.29}) +[2026-06-10 16:38:26,611][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.0008585065952502191, 'train/grad_norm_pre_clip': 0.025551918894052505, 'timing/data': 0.029104482382535934, 'timing/model': 1.126476638019085, 'learning_rate/qwen_vl_interface': 7.373148448187827e-06, 'learning_rate/action_model': 7.373148448187827e-05, 'epoch': 3.3}) +[2026-06-10 16:38:44,863][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.0011561234714463353, 'train/grad_norm_pre_clip': 0.015109756030142307, 'timing/data': 0.046196382492780685, 'timing/model': 1.1314065046608448, 'learning_rate/qwen_vl_interface': 7.339711056815598e-06, 'learning_rate/action_model': 7.339711056815598e-05, 'epoch': 3.32}) +[2026-06-10 16:39:04,906][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.000785299576818943, 'train/grad_norm_pre_clip': 0.005570724606513977, 'timing/data': 0.03900236263871193, 'timing/model': 1.1165284290909767, 'learning_rate/qwen_vl_interface': 7.306144762221968e-06, 'learning_rate/action_model': 7.306144762221968e-05, 'epoch': 3.33}) +[2026-06-10 16:39:26,023][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.0010877404129132628, 'train/grad_norm_pre_clip': 0.016242612153291702, 'timing/data': 0.03862135857343674, 'timing/model': 1.1725678741931915, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 3.34}) +[2026-06-10 16:39:35,768][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.00047740264562889934, 'train/grad_norm_pre_clip': 0.004037020727992058, 'timing/data': 0.08749917894601822, 'timing/model': 0.45140037313103676, 'learning_rate/qwen_vl_interface': 7.238633753302784e-06, 'learning_rate/action_model': 7.238633753302784e-05, 'epoch': 3.36}) +[2026-06-10 16:39:57,002][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.0016457941383123398, 'train/grad_norm_pre_clip': 0.07973943650722504, 'timing/data': 0.03948983550071716, 'timing/model': 1.1978941597044468, 'learning_rate/qwen_vl_interface': 7.204693203374268e-06, 'learning_rate/action_model': 7.204693203374267e-05, 'epoch': 3.37}) +[2026-06-10 16:40:16,956][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.17135439813137054, 'train/grad_norm_pre_clip': 1.653503656387329, 'timing/data': 0.02572290599346161, 'timing/model': 1.1779216565191746, 'learning_rate/qwen_vl_interface': 7.17063207876154e-06, 'learning_rate/action_model': 7.17063207876154e-05, 'epoch': 3.38}) +[2026-06-10 16:40:35,327][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.0016040243208408356, 'train/grad_norm_pre_clip': 0.014436905272305012, 'timing/data': 0.021829873323440552, 'timing/model': 1.1917116343975067, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 3.4}) +[2026-06-10 16:40:54,897][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.1007888987660408, 'train/grad_norm_pre_clip': 0.7329632639884949, 'timing/data': 0.03579455986618996, 'timing/model': 1.136199925094843, 'learning_rate/qwen_vl_interface': 7.1021565169938955e-06, 'learning_rate/action_model': 7.102156516993895e-05, 'epoch': 3.41}) +[2026-06-10 16:41:11,092][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 250 +[2026-06-10 16:41:11,094][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.0014681542525067925, 'train/grad_norm_pre_clip': 0.04419873654842377, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/mean_length': 16.8, 'rl_games_eval/mid_train/std_reward': 164.1876467304262, 'rl_games_eval/mid_train/std_length': 5.114684741017769, 'rl_games_eval/mid_train/macro_mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/macro_mean_length': 16.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_length': 16.8, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_reward': 164.1876467304262, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_length': 5.114684741017769, 'rl_games_eval/mid_train/deadly_corridor__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 235.99388122558594, 'checkpoint/best_score': 257.1217102050781, 'checkpoint/best_step': 50.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.12004859745502472, 'timing/model': 0.6325753666460514, 'learning_rate/qwen_vl_interface': 7.067746303734178e-06, 'learning_rate/action_model': 7.067746303734177e-05, 'epoch': 3.42}) +[2026-06-10 16:41:11,095][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_250_state +[2026-06-10 16:41:11,099][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 16:41:18,736][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_250_state/model.safetensors +[2026-06-10 16:41:31,892][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_250_state/optimizer.bin +[2026-06-10 16:41:31,894][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_250_state/sampler.bin +[2026-06-10 16:41:31,895][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_250_state/sampler_1.bin +[2026-06-10 16:41:31,897][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_250_state/random_states_0.pkl +[2026-06-10 16:41:31,927][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 250 +[2026-06-10 16:41:31,927][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 16:41:31,970][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 16:41:55,121][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.006695177871733904, 'train/grad_norm_pre_clip': 0.08009526878595352, 'timing/data': 0.038582295179367065, 'timing/model': 1.1557668894529343, 'learning_rate/qwen_vl_interface': 7.033223963320096e-06, 'learning_rate/action_model': 7.033223963320095e-05, 'epoch': 3.44}) +[2026-06-10 16:42:13,600][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.010801995173096657, 'train/grad_norm_pre_clip': 0.1590433418750763, 'timing/data': 0.02614116668701172, 'timing/model': 1.150921929627657, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 3.45}) +[2026-06-10 16:42:32,396][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.01567087136209011, 'train/grad_norm_pre_clip': 0.2806206941604614, 'timing/data': 0.02627711370587349, 'timing/model': 1.250446017831564, 'learning_rate/qwen_vl_interface': 6.963851425816299e-06, 'learning_rate/action_model': 6.963851425816299e-05, 'epoch': 3.47}) +[2026-06-10 16:42:50,890][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.12289048731327057, 'train/grad_norm_pre_clip': 0.8844390511512756, 'timing/data': 0.02731424942612648, 'timing/model': 1.1859724409878254, 'learning_rate/qwen_vl_interface': 6.929005507951473e-06, 'learning_rate/action_model': 6.929005507951473e-05, 'epoch': 3.48}) +[2026-06-10 16:43:00,603][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.0026299075689166784, 'train/grad_norm_pre_clip': 0.054543618112802505, 'timing/data': 0.06781893596053123, 'timing/model': 0.4423501640558243, 'learning_rate/qwen_vl_interface': 6.894056021118092e-06, 'learning_rate/action_model': 6.894056021118091e-05, 'epoch': 3.49}) +[2026-06-10 16:43:20,929][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.12364804744720459, 'train/grad_norm_pre_clip': 0.7284449338912964, 'timing/data': 0.00927816703915596, 'timing/model': 1.1910122856497765, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 3.51}) +[2026-06-10 16:43:39,234][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.029750436544418335, 'train/grad_norm_pre_clip': 0.5693946480751038, 'timing/data': 0.009802889078855515, 'timing/model': 1.1504965424537659, 'learning_rate/qwen_vl_interface': 6.823854970197212e-06, 'learning_rate/action_model': 6.823854970197213e-05, 'epoch': 3.52}) +[2026-06-10 16:44:01,117][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.2124958485364914, 'train/grad_norm_pre_clip': 0.326730877161026, 'timing/data': 0.010490603744983673, 'timing/model': 1.24815733730793, 'learning_rate/qwen_vl_interface': 6.788607736441211e-06, 'learning_rate/action_model': 6.78860773644121e-05, 'epoch': 3.53}) +[2026-06-10 16:44:20,115][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.08891825377941132, 'train/grad_norm_pre_clip': 0.6207981109619141, 'timing/data': 0.011811256408691406, 'timing/model': 1.3033621795475483, 'learning_rate/qwen_vl_interface': 6.753265594112535e-06, 'learning_rate/action_model': 6.753265594112534e-05, 'epoch': 3.55}) +[2026-06-10 16:44:31,127][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.0019557420164346695, 'train/grad_norm_pre_clip': 0.03970443457365036, 'timing/data': 0.07843432202935219, 'timing/model': 0.5825708322227001, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 3.56}) +[2026-06-10 16:44:57,135][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.003526887856423855, 'train/grad_norm_pre_clip': 0.14475949108600616, 'timing/data': 0.010932032018899918, 'timing/model': 1.4009960852563381, 'learning_rate/qwen_vl_interface': 6.682305309736355e-06, 'learning_rate/action_model': 6.682305309736354e-05, 'epoch': 3.58}) +[2026-06-10 16:45:16,690][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.021447183564305305, 'train/grad_norm_pre_clip': 0.26949840784072876, 'timing/data': 0.011827684938907623, 'timing/model': 1.2505773454904556, 'learning_rate/qwen_vl_interface': 6.646691544853444e-06, 'learning_rate/action_model': 6.646691544853444e-05, 'epoch': 3.59}) +[2026-06-10 16:45:35,296][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.0008019344531930983, 'train/grad_norm_pre_clip': 0.0070761628448963165, 'timing/data': 0.011340931057929993, 'timing/model': 1.1966325715184212, 'learning_rate/qwen_vl_interface': 6.610991625457042e-06, 'learning_rate/action_model': 6.610991625457043e-05, 'epoch': 3.6}) +[2026-06-10 16:45:55,616][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.16823238134384155, 'train/grad_norm_pre_clip': 0.476230651140213, 'timing/data': 0.01740725338459015, 'timing/model': 1.4122758656740189, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 3.62}) +[2026-06-10 16:46:07,897][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.032304033637046814, 'train/grad_norm_pre_clip': 2.505490303039551, 'timing/data': 0.10276360437273979, 'timing/model': 0.6222924627363682, 'learning_rate/qwen_vl_interface': 6.539342136859104e-06, 'learning_rate/action_model': 6.539342136859104e-05, 'epoch': 3.63}) +[2026-06-10 16:46:30,635][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.2537839114665985, 'train/grad_norm_pre_clip': 0.4095171391963959, 'timing/data': 0.027109496295452118, 'timing/model': 1.1833229064941406, 'learning_rate/qwen_vl_interface': 6.503396987335522e-06, 'learning_rate/action_model': 6.503396987335521e-05, 'epoch': 3.64}) +[2026-06-10 16:46:57,104][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.0008524323347955942, 'train/grad_norm_pre_clip': 0.03395368903875351, 'timing/data': 0.025109514594078064, 'timing/model': 1.2478658966720104, 'learning_rate/qwen_vl_interface': 6.4673745223817355e-06, 'learning_rate/action_model': 6.467374522381734e-05, 'epoch': 3.66}) +[2026-06-10 16:47:17,409][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.012345461174845695, 'train/grad_norm_pre_clip': 0.29728785157203674, 'timing/data': 0.028645876795053482, 'timing/model': 1.3374711833894253, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 3.67}) +[2026-06-10 16:47:36,117][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.0013586831046268344, 'train/grad_norm_pre_clip': 0.018412142992019653, 'timing/data': 0.03091634064912796, 'timing/model': 1.1604314371943474, 'learning_rate/qwen_vl_interface': 6.395106538956934e-06, 'learning_rate/action_model': 6.395106538956934e-05, 'epoch': 3.68}) +[2026-06-10 16:47:46,045][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.0011219108710065484, 'train/grad_norm_pre_clip': 0.015945320948958397, 'timing/data': 0.07577906548976898, 'timing/model': 0.4325004331767559, 'learning_rate/qwen_vl_interface': 6.3588654783155525e-06, 'learning_rate/action_model': 6.358865478315553e-05, 'epoch': 3.7}) +[2026-06-10 16:48:06,867][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.0918525829911232, 'train/grad_norm_pre_clip': 0.28223496675491333, 'timing/data': 0.023868467658758163, 'timing/model': 1.2540221884846687, 'learning_rate/qwen_vl_interface': 6.322556017628244e-06, 'learning_rate/action_model': 6.322556017628244e-05, 'epoch': 3.71}) +[2026-06-10 16:48:25,474][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.011309930123388767, 'train/grad_norm_pre_clip': 0.17295584082603455, 'timing/data': 0.03233364224433899, 'timing/model': 1.33509935811162, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 3.73}) +[2026-06-10 16:48:47,133][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.0025921501219272614, 'train/grad_norm_pre_clip': 0.022336818277835846, 'timing/data': 0.020839013159275055, 'timing/model': 1.1669117622077465, 'learning_rate/qwen_vl_interface': 6.249740859151202e-06, 'learning_rate/action_model': 6.249740859151202e-05, 'epoch': 3.74}) +[2026-06-10 16:49:09,186][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.08010698109865189, 'train/grad_norm_pre_clip': 3.943284511566162, 'timing/data': 0.053651727735996246, 'timing/model': 1.3710197769105434, 'learning_rate/qwen_vl_interface': 6.213239652943435e-06, 'learning_rate/action_model': 6.213239652943435e-05, 'epoch': 3.75}) +[2026-06-10 16:49:22,020][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.001907464349642396, 'train/grad_norm_pre_clip': 0.016864275559782982, 'timing/data': 0.07286980748176575, 'timing/model': 0.45895394310355186, 'learning_rate/qwen_vl_interface': 6.176679029576611e-06, 'learning_rate/action_model': 6.17667902957661e-05, 'epoch': 3.77}) +[2026-06-10 16:49:48,235][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.19002127647399902, 'train/grad_norm_pre_clip': 4.203022480010986, 'timing/data': 0.020914874970912933, 'timing/model': 1.2504131868481636, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 3.78}) +[2026-06-10 16:50:08,308][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.006383041851222515, 'train/grad_norm_pre_clip': 0.12316037714481354, 'timing/data': 0.02997102588415146, 'timing/model': 1.13329616189003, 'learning_rate/qwen_vl_interface': 6.103388555817666e-06, 'learning_rate/action_model': 6.1033885558176665e-05, 'epoch': 3.79}) +[2026-06-10 16:50:27,234][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.007620502728968859, 'train/grad_norm_pre_clip': 0.45578622817993164, 'timing/data': 0.026193324476480484, 'timing/model': 1.4616937898099422, 'learning_rate/qwen_vl_interface': 6.066663226327195e-06, 'learning_rate/action_model': 6.066663226327195e-05, 'epoch': 3.81}) +[2026-06-10 16:50:49,416][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.002668800065293908, 'train/grad_norm_pre_clip': 0.019426319748163223, 'timing/data': 0.034611284732818604, 'timing/model': 1.5900836363434792, 'learning_rate/qwen_vl_interface': 6.0298875212021e-06, 'learning_rate/action_model': 6.0298875212021e-05, 'epoch': 3.82}) +[2026-06-10 16:50:59,937][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.8157626986503601, 'train/grad_norm_pre_clip': 3.309123992919922, 'timing/data': 0.07227405905723572, 'timing/model': 0.43432097136974335, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 3.84}) +[2026-06-10 16:51:26,427][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.012052354402840137, 'train/grad_norm_pre_clip': 0.33675748109817505, 'timing/data': 0.010572846978902817, 'timing/model': 1.2132440246641636, 'learning_rate/qwen_vl_interface': 5.9561940610103865e-06, 'learning_rate/action_model': 5.956194061010386e-05, 'epoch': 3.85}) +[2026-06-10 16:51:46,936][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.00515668885782361, 'train/grad_norm_pre_clip': 0.1226777657866478, 'timing/data': 0.013784728944301605, 'timing/model': 1.168183345347643, 'learning_rate/qwen_vl_interface': 5.919280851703518e-06, 'learning_rate/action_model': 5.919280851703518e-05, 'epoch': 3.86}) +[2026-06-10 16:52:05,633][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.057577285915613174, 'train/grad_norm_pre_clip': 0.3656483292579651, 'timing/data': 0.011241264641284943, 'timing/model': 1.261338297277689, 'learning_rate/qwen_vl_interface': 5.8823263580011245e-06, 'learning_rate/action_model': 5.882326358001125e-05, 'epoch': 3.88}) +[2026-06-10 16:52:25,565][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.01993294432759285, 'train/grad_norm_pre_clip': 0.67426997423172, 'timing/data': 0.012611161917448044, 'timing/model': 1.1362369060516357, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 3.89}) +[2026-06-10 16:52:35,308][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.0023990552872419357, 'train/grad_norm_pre_clip': 0.026596689596772194, 'timing/data': 0.07382884621620178, 'timing/model': 0.5810355879366398, 'learning_rate/qwen_vl_interface': 5.80830263792473e-06, 'learning_rate/action_model': 5.80830263792473e-05, 'epoch': 3.9}) +[2026-06-10 16:52:56,725][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.18462993204593658, 'train/grad_norm_pre_clip': 0.4460226595401764, 'timing/data': 0.020776957273483276, 'timing/model': 1.2017074041068554, 'learning_rate/qwen_vl_interface': 5.771237977682465e-06, 'learning_rate/action_model': 5.771237977682465e-05, 'epoch': 3.92}) +[2026-06-10 16:53:15,703][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.041890282183885574, 'train/grad_norm_pre_clip': 0.3077322542667389, 'timing/data': 0.02438037097454071, 'timing/model': 1.0914086997509003, 'learning_rate/qwen_vl_interface': 5.734141165026489e-06, 'learning_rate/action_model': 5.7341411650264886e-05, 'epoch': 3.93}) +[2026-06-10 16:53:38,202][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.13180087506771088, 'train/grad_norm_pre_clip': 1.6217331886291504, 'timing/data': 0.02086116001009941, 'timing/model': 1.4551070146262646, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 3.95}) +[2026-06-10 16:53:57,718][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.22724102437496185, 'train/grad_norm_pre_clip': 0.7761465311050415, 'timing/data': 0.017599549144506454, 'timing/model': 1.1683428548276424, 'learning_rate/qwen_vl_interface': 5.659860237540138e-06, 'learning_rate/action_model': 5.659860237540138e-05, 'epoch': 3.96}) +[2026-06-10 16:54:08,028][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.011975686997175217, 'train/grad_norm_pre_clip': 0.2806258797645569, 'timing/data': 0.07616503909230232, 'timing/model': 0.4595099650323391, 'learning_rate/qwen_vl_interface': 5.622680704707265e-06, 'learning_rate/action_model': 5.622680704707265e-05, 'epoch': 3.97}) +[2026-06-10 16:54:37,281][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.11336372792720795, 'train/grad_norm_pre_clip': 1.5471341609954834, 'timing/data': 0.023605503141880035, 'timing/model': 1.2005775719881058, 'learning_rate/qwen_vl_interface': 5.585478183173043e-06, 'learning_rate/action_model': 5.585478183173043e-05, 'epoch': 3.99}) +[2026-06-10 16:54:59,403][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.07368048280477524, 'train/grad_norm_pre_clip': 2.281010389328003, 'timing/data': 0.02209959179162979, 'timing/model': 3.5166054032742977, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 4.0}) +[2026-06-10 16:55:24,082][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.15104150772094727, 'train/grad_norm_pre_clip': 2.2987449169158936, 'timing/data': 0.019475501030683517, 'timing/model': 1.198299903422594, 'learning_rate/qwen_vl_interface': 5.511013354584118e-06, 'learning_rate/action_model': 5.511013354584118e-05, 'epoch': 4.01}) +[2026-06-10 16:55:42,795][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.008992649614810944, 'train/grad_norm_pre_clip': 0.2116873562335968, 'timing/data': 0.021781206130981445, 'timing/model': 1.1403323113918304, 'learning_rate/qwen_vl_interface': 5.473755640870802e-06, 'learning_rate/action_model': 5.473755640870802e-05, 'epoch': 4.03}) +[2026-06-10 16:55:52,540][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.0791960135102272, 'train/grad_norm_pre_clip': 0.659469723701477, 'timing/data': 0.07610192894935608, 'timing/model': 0.43697264418005943, 'learning_rate/qwen_vl_interface': 5.436484124855577e-06, 'learning_rate/action_model': 5.4364841248555764e-05, 'epoch': 4.04}) +[2026-06-10 16:56:17,014][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.12823614478111267, 'train/grad_norm_pre_clip': 0.842124879360199, 'timing/data': 0.02992982417345047, 'timing/model': 1.1806981079280376, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 4.05}) +[2026-06-10 16:56:36,182][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.1049652174115181, 'train/grad_norm_pre_clip': 0.49719128012657166, 'timing/data': 0.0331929586827755, 'timing/model': 1.1602152176201344, 'learning_rate/qwen_vl_interface': 5.36190888295965e-06, 'learning_rate/action_model': 5.36190888295965e-05, 'epoch': 4.07}) +[2026-06-10 16:56:58,864][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.02997410111129284, 'train/grad_norm_pre_clip': 0.5518897771835327, 'timing/data': 0.029326733201742172, 'timing/model': 1.1685457229614258, 'learning_rate/qwen_vl_interface': 5.324609757231148e-06, 'learning_rate/action_model': 5.324609757231148e-05, 'epoch': 4.08}) +[2026-06-10 16:57:17,620][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.021024130284786224, 'train/grad_norm_pre_clip': 0.3933468163013458, 'timing/data': 0.03152778744697571, 'timing/model': 1.167479958385229, 'learning_rate/qwen_vl_interface': 5.287306029221379e-06, 'learning_rate/action_model': 5.287306029221378e-05, 'epoch': 4.1}) +[2026-06-10 16:57:44,829][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 300 +[2026-06-10 16:57:44,831][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.003656119341030717, 'train/grad_norm_pre_clip': 0.10944332182407379, 'eval/loss': 0.08058981597423553, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/mean_length': 16.8, 'rl_games_eval/mid_train/std_reward': 164.1876467304262, 'rl_games_eval/mid_train/std_length': 5.114684741017769, 'rl_games_eval/mid_train/macro_mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/macro_mean_length': 16.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_length': 16.8, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_reward': 164.1876467304262, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_length': 5.114684741017769, 'rl_games_eval/mid_train/deadly_corridor__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 235.99388122558594, 'checkpoint/best_score': 257.1217102050781, 'checkpoint/best_step': 50.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0771879144012928, 'timing/model': 0.5701720677316189, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 4.11}) +[2026-06-10 16:57:44,832][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_300_state +[2026-06-10 16:57:44,838][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 16:57:51,219][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_300_state/model.safetensors +[2026-06-10 16:58:01,829][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_300_state/optimizer.bin +[2026-06-10 16:58:01,830][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_300_state/sampler.bin +[2026-06-10 16:58:01,830][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_300_state/sampler_1.bin +[2026-06-10 16:58:01,832][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-10 16:58:01,840][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-10 16:58:01,840][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 16:58:01,854][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 16:58:29,279][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.10177973657846451, 'train/grad_norm_pre_clip': 0.3958897292613983, 'timing/data': 0.03529861569404602, 'timing/model': 1.547797441482544, 'learning_rate/qwen_vl_interface': 5.212693970778623e-06, 'learning_rate/action_model': 5.2126939707786224e-05, 'epoch': 4.12}) +[2026-06-10 16:58:47,922][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.02689732238650322, 'train/grad_norm_pre_clip': 0.21768000721931458, 'timing/data': 0.03237360715866089, 'timing/model': 1.2841318622231483, 'learning_rate/qwen_vl_interface': 5.175390242768853e-06, 'learning_rate/action_model': 5.175390242768853e-05, 'epoch': 4.14}) +[2026-06-10 16:59:11,354][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.001547530060634017, 'train/grad_norm_pre_clip': 0.017126865684986115, 'timing/data': 0.03128284960985184, 'timing/model': 1.1437837518751621, 'learning_rate/qwen_vl_interface': 5.1380911170403525e-06, 'learning_rate/action_model': 5.1380911170403525e-05, 'epoch': 4.15}) +[2026-06-10 16:59:30,661][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.21422366797924042, 'train/grad_norm_pre_clip': 0.7301243543624878, 'timing/data': 0.0362425222992897, 'timing/model': 2.1210304722189903, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 4.16}) +[2026-06-10 16:59:43,356][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.0016529994318261743, 'train/grad_norm_pre_clip': 0.053309597074985504, 'timing/data': 0.12238961830735207, 'timing/model': 0.6245225816965103, 'learning_rate/qwen_vl_interface': 5.063515875144426e-06, 'learning_rate/action_model': 5.0635158751444254e-05, 'epoch': 4.18}) +[2026-06-10 17:00:03,828][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.02879544161260128, 'train/grad_norm_pre_clip': 3.3744208812713623, 'timing/data': 0.020881637930870056, 'timing/model': 1.142689362168312, 'learning_rate/qwen_vl_interface': 5.026244359129198e-06, 'learning_rate/action_model': 5.0262443591291975e-05, 'epoch': 4.19}) +[2026-06-10 17:00:25,133][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.21394149959087372, 'train/grad_norm_pre_clip': 0.46014150977134705, 'timing/data': 0.028954744338989258, 'timing/model': 1.1187137179076672, 'learning_rate/qwen_vl_interface': 4.988986645415884e-06, 'learning_rate/action_model': 4.988986645415884e-05, 'epoch': 4.21}) +[2026-06-10 17:00:43,515][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.009729091078042984, 'train/grad_norm_pre_clip': 2.230489730834961, 'timing/data': 0.02330305054783821, 'timing/model': 1.3361230082809925, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 4.22}) +[2026-06-10 17:01:01,935][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.0015984821366146207, 'train/grad_norm_pre_clip': 0.017928913235664368, 'timing/data': 0.029124025255441666, 'timing/model': 1.5763628333806992, 'learning_rate/qwen_vl_interface': 4.91452181682696e-06, 'learning_rate/action_model': 4.9145218168269594e-05, 'epoch': 4.23}) +[2026-06-10 17:01:11,364][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.545932412147522, 'train/grad_norm_pre_clip': 1.485346794128418, 'timing/data': 0.08917322754859924, 'timing/model': 0.38186390697956085, 'learning_rate/qwen_vl_interface': 4.877319295292737e-06, 'learning_rate/action_model': 4.8773192952927374e-05, 'epoch': 4.25}) +[2026-06-10 17:01:33,630][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.017148617655038834, 'train/grad_norm_pre_clip': 0.704447329044342, 'timing/data': 0.036513809114694595, 'timing/model': 1.2201076559722424, 'learning_rate/qwen_vl_interface': 4.840139762459865e-06, 'learning_rate/action_model': 4.8401397624598645e-05, 'epoch': 4.26}) +[2026-06-10 17:01:52,527][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.08015638589859009, 'train/grad_norm_pre_clip': 0.17703908681869507, 'timing/data': 0.023069094866514206, 'timing/model': 1.150646559894085, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 4.27}) +[2026-06-10 17:02:13,014][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.11223083734512329, 'train/grad_norm_pre_clip': 3.564905881881714, 'timing/data': 0.03269079327583313, 'timing/model': 1.138362918049097, 'learning_rate/qwen_vl_interface': 4.7658588349735135e-06, 'learning_rate/action_model': 4.765858834973513e-05, 'epoch': 4.29}) +[2026-06-10 17:02:31,327][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.037215568125247955, 'train/grad_norm_pre_clip': 18.869279861450195, 'timing/data': 0.032815683633089066, 'timing/model': 1.1504939123988152, 'learning_rate/qwen_vl_interface': 4.728762022317535e-06, 'learning_rate/action_model': 4.728762022317535e-05, 'epoch': 4.3}) +[2026-06-10 17:02:41,068][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.004413109738379717, 'train/grad_norm_pre_clip': 0.08042816072702408, 'timing/data': 0.12497086450457573, 'timing/model': 0.4633653201162815, 'learning_rate/qwen_vl_interface': 4.691697362075272e-06, 'learning_rate/action_model': 4.691697362075272e-05, 'epoch': 4.32}) +[2026-06-10 17:03:03,179][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.01948958821594715, 'train/grad_norm_pre_clip': 0.13646721839904785, 'timing/data': 0.032655682414770126, 'timing/model': 1.1459049880504608, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 4.33}) +[2026-06-10 17:03:25,808][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.014087243936955929, 'train/grad_norm_pre_clip': 0.0938696563243866, 'timing/data': 0.042426399886608124, 'timing/model': 1.372735258191824, 'learning_rate/qwen_vl_interface': 4.617673641998877e-06, 'learning_rate/action_model': 4.617673641998877e-05, 'epoch': 4.34}) +[2026-06-10 17:03:44,416][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.00857694074511528, 'train/grad_norm_pre_clip': 0.5054372549057007, 'timing/data': 0.0503796748816967, 'timing/model': 1.1267123222351074, 'learning_rate/qwen_vl_interface': 4.580719148296482e-06, 'learning_rate/action_model': 4.580719148296482e-05, 'epoch': 4.36}) +[2026-06-10 17:04:02,497][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.022520272061228752, 'train/grad_norm_pre_clip': 2.948458194732666, 'timing/data': 0.033455315977334976, 'timing/model': 1.1890139319002628, 'learning_rate/qwen_vl_interface': 4.543805938989616e-06, 'learning_rate/action_model': 4.543805938989616e-05, 'epoch': 4.37}) +[2026-06-10 17:04:12,053][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.0005413353210315108, 'train/grad_norm_pre_clip': 0.011675619520246983, 'timing/data': 0.07973336428403854, 'timing/model': 0.43896960839629173, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 4.38}) +[2026-06-10 17:04:32,907][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.005343874450773001, 'train/grad_norm_pre_clip': 0.19736343622207642, 'timing/data': 0.022474218159914017, 'timing/model': 1.4582949839532375, 'learning_rate/qwen_vl_interface': 4.470112478797902e-06, 'learning_rate/action_model': 4.470112478797901e-05, 'epoch': 4.4}) +[2026-06-10 17:04:52,170][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.003812342882156372, 'train/grad_norm_pre_clip': 0.15846344828605652, 'timing/data': 0.026922903954982758, 'timing/model': 1.139097087085247, 'learning_rate/qwen_vl_interface': 4.433336773672804e-06, 'learning_rate/action_model': 4.4333367736728035e-05, 'epoch': 4.41}) +[2026-06-10 17:05:10,043][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.0019422663608565927, 'train/grad_norm_pre_clip': 0.0376257561147213, 'timing/data': 0.02069772407412529, 'timing/model': 1.1237933300435543, 'learning_rate/qwen_vl_interface': 4.396611444182336e-06, 'learning_rate/action_model': 4.3966114441823353e-05, 'epoch': 4.42}) +[2026-06-10 17:05:29,573][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.0029323850758373737, 'train/grad_norm_pre_clip': 0.15122796595096588, 'timing/data': 0.013935968279838562, 'timing/model': 1.1545045487582684, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 4.44}) +[2026-06-10 17:05:39,412][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.0050612702034413815, 'train/grad_norm_pre_clip': 0.20228788256645203, 'timing/data': 0.07962802797555923, 'timing/model': 0.6186561286449432, 'learning_rate/qwen_vl_interface': 4.323320970423391e-06, 'learning_rate/action_model': 4.3233209704233915e-05, 'epoch': 4.45}) +[2026-06-10 17:06:01,356][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.21682055294513702, 'train/grad_norm_pre_clip': 1.1441246271133423, 'timing/data': 0.00991305336356163, 'timing/model': 1.1452848426997662, 'learning_rate/qwen_vl_interface': 4.286760347056567e-06, 'learning_rate/action_model': 4.286760347056567e-05, 'epoch': 4.47}) +[2026-06-10 17:06:19,440][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.0011392062297090888, 'train/grad_norm_pre_clip': 0.05859602615237236, 'timing/data': 0.008457086980342865, 'timing/model': 1.1465333998203278, 'learning_rate/qwen_vl_interface': 4.250259140848798e-06, 'learning_rate/action_model': 4.2502591408487985e-05, 'epoch': 4.48}) +[2026-06-10 17:06:38,055][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.003901012474671006, 'train/grad_norm_pre_clip': 1.2074326276779175, 'timing/data': 0.013810582458972931, 'timing/model': 1.2864212952554226, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 4.49}) +[2026-06-10 17:06:55,780][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.004227076191455126, 'train/grad_norm_pre_clip': 0.09797821193933487, 'timing/data': 0.011351872235536575, 'timing/model': 1.1457763947546482, 'learning_rate/qwen_vl_interface': 4.177443982371757e-06, 'learning_rate/action_model': 4.177443982371758e-05, 'epoch': 4.51}) +[2026-06-10 17:07:06,079][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.03440019115805626, 'train/grad_norm_pre_clip': 0.2568061351776123, 'timing/data': 0.07293614745140076, 'timing/model': 0.45993082597851753, 'learning_rate/qwen_vl_interface': 4.14113452168445e-06, 'learning_rate/action_model': 4.1411345216844504e-05, 'epoch': 4.52}) +[2026-06-10 17:07:27,537][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.0004974791663698852, 'train/grad_norm_pre_clip': 0.006560095120221376, 'timing/data': 0.044925544410943985, 'timing/model': 1.3200318962335587, 'learning_rate/qwen_vl_interface': 4.104893461043067e-06, 'learning_rate/action_model': 4.1048934610430664e-05, 'epoch': 4.53}) +[2026-06-10 17:07:47,129][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.0017487402074038982, 'train/grad_norm_pre_clip': 0.06104402616620064, 'timing/data': 0.06848260015249252, 'timing/model': 1.3770633563399315, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 4.55}) +[2026-06-10 17:08:05,704][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.003001976292580366, 'train/grad_norm_pre_clip': 1.7210251092910767, 'timing/data': 0.03202224522829056, 'timing/model': 1.6668934635818005, 'learning_rate/qwen_vl_interface': 4.032625477618266e-06, 'learning_rate/action_model': 4.0326254776182654e-05, 'epoch': 4.56}) +[2026-06-10 17:08:24,519][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.24769598245620728, 'train/grad_norm_pre_clip': 1.7751421928405762, 'timing/data': 0.0335644856095314, 'timing/model': 1.1442159377038479, 'learning_rate/qwen_vl_interface': 3.99660301266448e-06, 'learning_rate/action_model': 3.99660301266448e-05, 'epoch': 4.58}) +[2026-06-10 17:08:37,393][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.0007859821780584753, 'train/grad_norm_pre_clip': 0.015800856053829193, 'timing/data': 0.07519354298710823, 'timing/model': 1.534657672047615, 'learning_rate/qwen_vl_interface': 3.960657863140897e-06, 'learning_rate/action_model': 3.9606578631408975e-05, 'epoch': 4.59}) +[2026-06-10 17:09:03,999][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.10036350041627884, 'train/grad_norm_pre_clip': 1.4574710130691528, 'timing/data': 0.019368082284927368, 'timing/model': 1.108277566730976, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 4.6}) +[2026-06-10 17:09:22,218][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.2768130302429199, 'train/grad_norm_pre_clip': 1.205014944076538, 'timing/data': 0.019403912127017975, 'timing/model': 1.0834719836711884, 'learning_rate/qwen_vl_interface': 3.889008374542957e-06, 'learning_rate/action_model': 3.889008374542957e-05, 'epoch': 4.62}) +[2026-06-10 17:09:40,977][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.0028549458365887403, 'train/grad_norm_pre_clip': 0.1991640329360962, 'timing/data': 0.0212249718606472, 'timing/model': 1.153727114200592, 'learning_rate/qwen_vl_interface': 3.853308455146557e-06, 'learning_rate/action_model': 3.8533084551465566e-05, 'epoch': 4.63}) +[2026-06-10 17:10:02,695][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.004221201874315739, 'train/grad_norm_pre_clip': 0.10959210991859436, 'timing/data': 0.0009036846458911896, 'timing/model': 1.341573428362608, 'learning_rate/qwen_vl_interface': 3.817694690263646e-06, 'learning_rate/action_model': 3.8176946902636455e-05, 'epoch': 4.64}) +[2026-06-10 17:10:13,231][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.0010900854831561446, 'train/grad_norm_pre_clip': 0.020202752202749252, 'timing/data': 0.1321585737168789, 'timing/model': 0.4745515286922455, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 4.66}) +[2026-06-10 17:10:35,285][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.0022118305787444115, 'train/grad_norm_pre_clip': 0.04437751695513725, 'timing/data': 0.020276032388210297, 'timing/model': 1.310206662863493, 'learning_rate/qwen_vl_interface': 3.746734405887466e-06, 'learning_rate/action_model': 3.746734405887466e-05, 'epoch': 4.67}) +[2026-06-10 17:10:53,820][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.002949399407953024, 'train/grad_norm_pre_clip': 0.04632997512817383, 'timing/data': 0.020790468901395798, 'timing/model': 1.2780439928174019, 'learning_rate/qwen_vl_interface': 3.711392263558791e-06, 'learning_rate/action_model': 3.7113922635587913e-05, 'epoch': 4.68}) +[2026-06-10 17:11:13,003][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.006952777970582247, 'train/grad_norm_pre_clip': 0.16188763082027435, 'timing/data': 0.024642180651426315, 'timing/model': 1.1411104016005993, 'learning_rate/qwen_vl_interface': 3.6761450298027874e-06, 'learning_rate/action_model': 3.676145029802787e-05, 'epoch': 4.7}) +[2026-06-10 17:11:31,586][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.0031142039224505424, 'train/grad_norm_pre_clip': 0.037931278347969055, 'timing/data': 0.009083028882741928, 'timing/model': 1.1597589403390884, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 4.71}) +[2026-06-10 17:11:41,662][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.5646161437034607, 'train/grad_norm_pre_clip': 2.7876007556915283, 'timing/data': 0.07992688938975334, 'timing/model': 0.6126087121665478, 'learning_rate/qwen_vl_interface': 3.605943978881907e-06, 'learning_rate/action_model': 3.605943978881907e-05, 'epoch': 4.73}) +[2026-06-10 17:12:02,269][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.006126905791461468, 'train/grad_norm_pre_clip': 0.1560860574245453, 'timing/data': 0.010066505521535873, 'timing/model': 1.1088379807770252, 'learning_rate/qwen_vl_interface': 3.5709944920485296e-06, 'learning_rate/action_model': 3.5709944920485295e-05, 'epoch': 4.74}) +[2026-06-10 17:12:20,215][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.008436430245637894, 'train/grad_norm_pre_clip': 0.11259154230356216, 'timing/data': 0.010525871068239212, 'timing/model': 1.1385106854140759, 'learning_rate/qwen_vl_interface': 3.5361485741837017e-06, 'learning_rate/action_model': 3.5361485741837014e-05, 'epoch': 4.75}) +[2026-06-10 17:12:38,515][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.003934036009013653, 'train/grad_norm_pre_clip': 0.0573342889547348, 'timing/data': 0.012758035212755203, 'timing/model': 1.3036229312419891, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 4.77}) +[2026-06-10 17:13:02,479][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.011807848699390888, 'train/grad_norm_pre_clip': 0.41622301936149597, 'timing/data': 0.008680827915668488, 'timing/model': 1.1998745203018188, 'learning_rate/qwen_vl_interface': 3.4667760366799048e-06, 'learning_rate/action_model': 3.4667760366799043e-05, 'epoch': 4.78}) +[2026-06-10 17:13:21,707][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 350 +[2026-06-10 17:13:21,709][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.0003314051718916744, 'train/grad_norm_pre_clip': 0.0038265837356448174, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/mean_length': 16.8, 'rl_games_eval/mid_train/std_reward': 164.1876467304262, 'rl_games_eval/mid_train/std_length': 5.114684741017769, 'rl_games_eval/mid_train/macro_mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/macro_mean_length': 16.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_length': 16.8, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_reward': 164.1876467304262, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_length': 5.114684741017769, 'rl_games_eval/mid_train/deadly_corridor__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 235.99388122558594, 'checkpoint/best_score': 257.1217102050781, 'checkpoint/best_step': 50.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.07807713747024536, 'timing/model': 0.46809452027082443, 'learning_rate/qwen_vl_interface': 3.432253696265824e-06, 'learning_rate/action_model': 3.432253696265824e-05, 'epoch': 4.79}) +[2026-06-10 17:13:21,710][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_350_state +[2026-06-10 17:13:21,722][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 17:13:29,846][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_350_state/model.safetensors +[2026-06-10 17:13:42,155][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_350_state/optimizer.bin +[2026-06-10 17:13:42,157][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_350_state/sampler.bin +[2026-06-10 17:13:42,157][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_350_state/sampler_1.bin +[2026-06-10 17:13:42,160][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_350_state/random_states_0.pkl +[2026-06-10 17:13:42,178][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 350 +[2026-06-10 17:13:42,187][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 17:13:42,201][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 17:14:02,693][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.006745277438312769, 'train/grad_norm_pre_clip': 0.12405840307474136, 'timing/data': 0.033187802881002426, 'timing/model': 1.1515924036502838, 'learning_rate/qwen_vl_interface': 3.397843483006107e-06, 'learning_rate/action_model': 3.397843483006107e-05, 'epoch': 4.81}) +[2026-06-10 17:14:21,302][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.13034772872924805, 'train/grad_norm_pre_clip': 0.4325312674045563, 'timing/data': 0.03278889134526253, 'timing/model': 1.3230720534920692, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 4.82}) +[2026-06-10 17:14:47,163][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.005690073594450951, 'train/grad_norm_pre_clip': 0.10391110926866531, 'timing/data': 0.02576635777950287, 'timing/model': 1.1415014378726482, 'learning_rate/qwen_vl_interface': 3.3293679212384616e-06, 'learning_rate/action_model': 3.329367921238461e-05, 'epoch': 4.84}) +[2026-06-10 17:15:08,667][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.0186433345079422, 'train/grad_norm_pre_clip': 1.7160277366638184, 'timing/data': 0.030643489211797714, 'timing/model': 2.4063851423561573, 'learning_rate/qwen_vl_interface': 3.2953067966257334e-06, 'learning_rate/action_model': 3.295306796625733e-05, 'epoch': 4.85}) +[2026-06-10 17:15:23,151][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.0021216084714978933, 'train/grad_norm_pre_clip': 0.04081645980477333, 'timing/data': 0.08708129450678825, 'timing/model': 0.5203545093536377, 'learning_rate/qwen_vl_interface': 3.2613662466972178e-06, 'learning_rate/action_model': 3.2613662466972176e-05, 'epoch': 4.86}) +[2026-06-10 17:15:50,702][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.004631354007869959, 'train/grad_norm_pre_clip': 0.09052956104278564, 'timing/data': 0.03758472204208374, 'timing/model': 2.4888655953109264, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 4.88}) +[2026-06-10 17:16:09,290][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.0020897898357361555, 'train/grad_norm_pre_clip': 0.04352600499987602, 'timing/data': 0.03229598328471184, 'timing/model': 1.1291849613189697, 'learning_rate/qwen_vl_interface': 3.193855237778036e-06, 'learning_rate/action_model': 3.1938552377780355e-05, 'epoch': 4.89}) +[2026-06-10 17:16:32,180][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.005328015424311161, 'train/grad_norm_pre_clip': 0.103033147752285, 'timing/data': 0.030326463282108307, 'timing/model': 1.1374192610383034, 'learning_rate/qwen_vl_interface': 3.160288943184403e-06, 'learning_rate/action_model': 3.160288943184403e-05, 'epoch': 4.9}) +[2026-06-10 17:16:52,577][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.005326277110725641, 'train/grad_norm_pre_clip': 0.2117765098810196, 'timing/data': 0.02385846897959709, 'timing/model': 1.3442540392279625, 'learning_rate/qwen_vl_interface': 3.126851551812175e-06, 'learning_rate/action_model': 3.126851551812175e-05, 'epoch': 4.92}) +[2026-06-10 17:17:02,168][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.0007468366529792547, 'train/grad_norm_pre_clip': 0.0362941212952137, 'timing/data': 0.08474500104784966, 'timing/model': 0.47877873852849007, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 4.93}) +[2026-06-10 17:17:22,214][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.11451976746320724, 'train/grad_norm_pre_clip': 3.6464767456054688, 'timing/data': 0.021411243826150894, 'timing/model': 1.1310797929763794, 'learning_rate/qwen_vl_interface': 3.0603717209565546e-06, 'learning_rate/action_model': 3.0603717209565546e-05, 'epoch': 4.95}) +[2026-06-10 17:17:45,103][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.23070518672466278, 'train/grad_norm_pre_clip': 2.097810983657837, 'timing/data': 0.010873369872570038, 'timing/model': 1.1942318454384804, 'learning_rate/qwen_vl_interface': 3.027333382262277e-06, 'learning_rate/action_model': 3.027333382262277e-05, 'epoch': 4.96}) +[2026-06-10 17:18:03,161][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.11169973760843277, 'train/grad_norm_pre_clip': 0.694015383720398, 'timing/data': 0.01577349752187729, 'timing/model': 1.2808407731354237, 'learning_rate/qwen_vl_interface': 2.9944321481146766e-06, 'learning_rate/action_model': 2.9944321481146764e-05, 'epoch': 4.97}) +[2026-06-10 17:18:23,011][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.0025524727534502745, 'train/grad_norm_pre_clip': 0.04810638725757599, 'timing/data': 0.02005373314023018, 'timing/model': 1.1537460573017597, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 4.99}) +[2026-06-10 17:18:32,611][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.740702211856842, 'train/grad_norm_pre_clip': 12.222283363342285, 'timing/data': 0.08577073737978935, 'timing/model': 0.5130280591547489, 'learning_rate/qwen_vl_interface': 2.9290491028894658e-06, 'learning_rate/action_model': 2.9290491028894658e-05, 'epoch': 5.0}) +[2026-06-10 17:18:53,461][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.00736267352476716, 'train/grad_norm_pre_clip': 0.48278576135635376, 'timing/data': 0.037883881479501724, 'timing/model': 1.1488555632531643, 'learning_rate/qwen_vl_interface': 2.8965713249460643e-06, 'learning_rate/action_model': 2.896571324946064e-05, 'epoch': 5.01}) +[2026-06-10 17:19:12,070][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.08461727946996689, 'train/grad_norm_pre_clip': 0.19332516193389893, 'timing/data': 0.03804365172982216, 'timing/model': 1.14420498162508, 'learning_rate/qwen_vl_interface': 2.8642387175689806e-06, 'learning_rate/action_model': 2.8642387175689805e-05, 'epoch': 5.03}) +[2026-06-10 17:19:32,119][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.14655210077762604, 'train/grad_norm_pre_clip': 2.5713956356048584, 'timing/data': 0.031070593744516373, 'timing/model': 1.146899588406086, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 5.04}) +[2026-06-10 17:19:50,426][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.001075333566404879, 'train/grad_norm_pre_clip': 0.020987818017601967, 'timing/data': 0.037320226430892944, 'timing/model': 1.1559634357690811, 'learning_rate/qwen_vl_interface': 2.800016983146025e-06, 'learning_rate/action_model': 2.800016983146025e-05, 'epoch': 5.05}) +[2026-06-10 17:20:00,215][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.001401193207129836, 'train/grad_norm_pre_clip': 0.0338464081287384, 'timing/data': 0.10488026961684227, 'timing/model': 0.43404170498251915, 'learning_rate/qwen_vl_interface': 2.768131817599243e-06, 'learning_rate/action_model': 2.768131817599243e-05, 'epoch': 5.07}) +[2026-06-10 17:20:23,418][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.009176474995911121, 'train/grad_norm_pre_clip': 0.15895429253578186, 'timing/data': 0.015425369143486023, 'timing/model': 1.6341349110007286, 'learning_rate/qwen_vl_interface': 2.7363997453725954e-06, 'learning_rate/action_model': 2.7363997453725954e-05, 'epoch': 5.08}) +[2026-06-10 17:20:45,941][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.01079951599240303, 'train/grad_norm_pre_clip': 0.10717032849788666, 'timing/data': 0.023691337555646896, 'timing/model': 1.358057476580143, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 5.1}) +[2026-06-10 17:21:05,644][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.04651612415909767, 'train/grad_norm_pre_clip': 0.37098291516304016, 'timing/data': 0.020280878990888596, 'timing/model': 1.264789868146181, 'learning_rate/qwen_vl_interface': 2.6734027008501866e-06, 'learning_rate/action_model': 2.6734027008501865e-05, 'epoch': 5.11}) +[2026-06-10 17:21:25,135][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.001208989298902452, 'train/grad_norm_pre_clip': 0.026533057913184166, 'timing/data': 0.01774825155735016, 'timing/model': 1.1738344952464104, 'learning_rate/qwen_vl_interface': 2.6421416145088737e-06, 'learning_rate/action_model': 2.6421416145088736e-05, 'epoch': 5.12}) +[2026-06-10 17:21:34,646][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.0006428968044929206, 'train/grad_norm_pre_clip': 0.023491762578487396, 'timing/data': 0.09482944011688232, 'timing/model': 0.45623713359236717, 'learning_rate/qwen_vl_interface': 2.6110413931568906e-06, 'learning_rate/action_model': 2.6110413931568904e-05, 'epoch': 5.14}) +[2026-06-10 17:21:57,179][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.0028511211276054382, 'train/grad_norm_pre_clip': 0.04793951287865639, 'timing/data': 0.010689772665500641, 'timing/model': 1.1643886268138885, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 5.15}) +[2026-06-10 17:22:16,367][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.003655640408396721, 'train/grad_norm_pre_clip': 0.08600723743438721, 'timing/data': 0.007408790290355682, 'timing/model': 1.1466291658580303, 'learning_rate/qwen_vl_interface': 2.5493312090122463e-06, 'learning_rate/action_model': 2.549331209012246e-05, 'epoch': 5.16}) +[2026-06-10 17:22:37,820][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.0036421678960323334, 'train/grad_norm_pre_clip': 0.0926261842250824, 'timing/data': 0.014042846858501434, 'timing/model': 1.1737516298890114, 'learning_rate/qwen_vl_interface': 2.5187250527944283e-06, 'learning_rate/action_model': 2.5187250527944284e-05, 'epoch': 5.18}) +[2026-06-10 17:22:56,903][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.05564868822693825, 'train/grad_norm_pre_clip': 0.4736616909503937, 'timing/data': 0.010925807058811188, 'timing/model': 1.3169365152716637, 'learning_rate/qwen_vl_interface': 2.488287374480805e-06, 'learning_rate/action_model': 2.4882873744808046e-05, 'epoch': 5.19}) +[2026-06-10 17:23:06,351][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.0003755529469344765, 'train/grad_norm_pre_clip': 0.00978775043040514, 'timing/data': 0.12717941775918007, 'timing/model': 0.43385525420308113, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 5.21}) +[2026-06-10 17:23:26,004][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.002312382683157921, 'train/grad_norm_pre_clip': 0.050278495997190475, 'timing/data': 0.038561467081308365, 'timing/model': 1.158308144658804, 'learning_rate/qwen_vl_interface': 2.427924951215317e-06, 'learning_rate/action_model': 2.427924951215317e-05, 'epoch': 5.22}) +[2026-06-10 17:23:45,219][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.2097882628440857, 'train/grad_norm_pre_clip': 3.709625005722046, 'timing/data': 0.049480047076940536, 'timing/model': 1.143435776233673, 'learning_rate/qwen_vl_interface': 2.3980039297020513e-06, 'learning_rate/action_model': 2.3980039297020513e-05, 'epoch': 5.23}) +[2026-06-10 17:24:03,554][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.006492982618510723, 'train/grad_norm_pre_clip': 0.03570236638188362, 'timing/data': 0.04243329539895058, 'timing/model': 1.1344840005040169, 'learning_rate/qwen_vl_interface': 2.3682588327405025e-06, 'learning_rate/action_model': 2.3682588327405024e-05, 'epoch': 5.25}) +[2026-06-10 17:24:22,149][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.04297707974910736, 'train/grad_norm_pre_clip': 0.444909930229187, 'timing/data': 0.05500131472945213, 'timing/model': 1.1192528195679188, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 5.26}) +[2026-06-10 17:24:35,660][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.00361258490011096, 'train/grad_norm_pre_clip': 0.09947101771831512, 'timing/data': 0.08120061457157135, 'timing/model': 0.4773022308945656, 'learning_rate/qwen_vl_interface': 2.3093037407782883e-06, 'learning_rate/action_model': 2.309303740778288e-05, 'epoch': 5.27}) +[2026-06-10 17:24:55,670][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.1465008705854416, 'train/grad_norm_pre_clip': 3.191211462020874, 'timing/data': 0.04443806782364845, 'timing/model': 1.1546256057918072, 'learning_rate/qwen_vl_interface': 2.280097382405402e-06, 'learning_rate/action_model': 2.2800973824054018e-05, 'epoch': 5.29}) +[2026-06-10 17:25:14,862][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.0005670961691066623, 'train/grad_norm_pre_clip': 0.03442396968603134, 'timing/data': 0.06427860260009766, 'timing/model': 1.2341524958610535, 'learning_rate/qwen_vl_interface': 2.2510742216154557e-06, 'learning_rate/action_model': 2.2510742216154557e-05, 'epoch': 5.3}) +[2026-06-10 17:25:33,718][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.003515024436637759, 'train/grad_norm_pre_clip': 0.2288234680891037, 'timing/data': 0.04070858284831047, 'timing/model': 1.1451375372707844, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 5.32}) +[2026-06-10 17:25:51,871][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.1518888920545578, 'train/grad_norm_pre_clip': 12.86481761932373, 'timing/data': 0.03211585059762001, 'timing/model': 1.1450343243777752, 'learning_rate/qwen_vl_interface': 2.19358464251455e-06, 'learning_rate/action_model': 2.19358464251455e-05, 'epoch': 5.33}) +[2026-06-10 17:26:09,590][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.0011635953560471535, 'train/grad_norm_pre_clip': 0.035674069076776505, 'timing/data': 0.073708925396204, 'timing/model': 0.511808056384325, 'learning_rate/qwen_vl_interface': 2.1651217704316283e-06, 'learning_rate/action_model': 2.165121770431628e-05, 'epoch': 5.34}) +[2026-06-10 17:26:29,326][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.009390388615429401, 'train/grad_norm_pre_clip': 0.25118234753608704, 'timing/data': 0.01788289099931717, 'timing/model': 1.1278258711099625, 'learning_rate/qwen_vl_interface': 2.1368491881689794e-06, 'learning_rate/action_model': 2.136849188168979e-05, 'epoch': 5.36}) +[2026-06-10 17:26:47,456][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.11909816414117813, 'train/grad_norm_pre_clip': 0.8067814707756042, 'timing/data': 0.01834067702293396, 'timing/model': 1.3591648526489735, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 5.37}) +[2026-06-10 17:27:09,680][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.04002907872200012, 'train/grad_norm_pre_clip': 1.8693127632141113, 'timing/data': 0.008635886013507843, 'timing/model': 1.128145795315504, 'learning_rate/qwen_vl_interface': 2.0808818572031507e-06, 'learning_rate/action_model': 2.0808818572031506e-05, 'epoch': 5.38}) +[2026-06-10 17:27:27,598][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.009675257839262486, 'train/grad_norm_pre_clip': 0.10747041553258896, 'timing/data': 0.009884435683488846, 'timing/model': 1.1346336342394352, 'learning_rate/qwen_vl_interface': 2.053190560828578e-06, 'learning_rate/action_model': 2.0531905608285777e-05, 'epoch': 5.4}) +[2026-06-10 17:27:37,257][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.0004957925993949175, 'train/grad_norm_pre_clip': 0.009244117885828018, 'timing/data': 0.07761260122060776, 'timing/model': 0.45449792593717575, 'learning_rate/qwen_vl_interface': 2.0256964587185266e-06, 'learning_rate/action_model': 2.0256964587185266e-05, 'epoch': 5.41}) +[2026-06-10 17:27:57,615][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.003805730026215315, 'train/grad_norm_pre_clip': 0.16775256395339966, 'timing/data': 0.006060473620891571, 'timing/model': 1.1794226989150047, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 5.42}) +[2026-06-10 17:28:16,138][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.034760940819978714, 'train/grad_norm_pre_clip': 4.996750354766846, 'timing/data': 0.008144848048686981, 'timing/model': 1.22105722874403, 'learning_rate/qwen_vl_interface': 1.9713066088864233e-06, 'learning_rate/action_model': 1.9713066088864232e-05, 'epoch': 5.44}) +[2026-06-10 17:28:42,237][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.03296845778822899, 'train/grad_norm_pre_clip': 0.6531102061271667, 'timing/data': 0.010922815650701523, 'timing/model': 3.4164648316800594, 'learning_rate/qwen_vl_interface': 1.944414216186507e-06, 'learning_rate/action_model': 1.944414216186507e-05, 'epoch': 5.45}) +[2026-06-10 17:29:06,297][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.06664564460515976, 'train/grad_norm_pre_clip': 4.631814479827881, 'timing/data': 0.009005237370729446, 'timing/model': 1.1572584807872772, 'learning_rate/qwen_vl_interface': 1.917725727588435e-06, 'learning_rate/action_model': 1.9177257275884353e-05, 'epoch': 5.47}) +[2026-06-10 17:29:34,223][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 400 +[2026-06-10 17:29:34,225][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.0003949541714973748, 'train/grad_norm_pre_clip': 0.014560997486114502, 'eval/loss': 0.0730557069182396, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/mean_length': 16.8, 'rl_games_eval/mid_train/std_reward': 164.1876467304262, 'rl_games_eval/mid_train/std_length': 5.114684741017769, 'rl_games_eval/mid_train/macro_mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/macro_mean_length': 16.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_length': 16.8, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_reward': 164.1876467304262, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_length': 5.114684741017769, 'rl_games_eval/mid_train/deadly_corridor__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 235.99388122558594, 'checkpoint/best_score': 257.1217102050781, 'checkpoint/best_step': 50.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0811612494289875, 'timing/model': 0.5177338942885399, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 5.48}) +[2026-06-10 17:29:34,227][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_400_state +[2026-06-10 17:29:34,232][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 17:29:40,694][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_400_state/model.safetensors +[2026-06-10 17:29:52,154][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_400_state/optimizer.bin +[2026-06-10 17:29:52,156][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_400_state/sampler.bin +[2026-06-10 17:29:52,156][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_400_state/sampler_1.bin +[2026-06-10 17:29:52,159][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-10 17:29:52,166][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-10 17:29:52,166][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 17:29:52,179][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 17:30:18,511][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.002980007091537118, 'train/grad_norm_pre_clip': 0.1375935971736908, 'timing/data': 0.029730066657066345, 'timing/model': 1.5801033042371273, 'learning_rate/qwen_vl_interface': 1.8649670351052746e-06, 'learning_rate/action_model': 1.8649670351052744e-05, 'epoch': 5.49}) +[2026-06-10 17:30:36,653][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.0005477344966493547, 'train/grad_norm_pre_clip': 0.018470345064997673, 'timing/data': 0.030951619148254395, 'timing/model': 1.193251110613346, 'learning_rate/qwen_vl_interface': 1.8389000856248537e-06, 'learning_rate/action_model': 1.8389000856248537e-05, 'epoch': 5.51}) +[2026-06-10 17:30:55,113][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.27133694291114807, 'train/grad_norm_pre_clip': 3.3738553524017334, 'timing/data': 0.027700889855623245, 'timing/model': 1.1506315842270851, 'learning_rate/qwen_vl_interface': 1.8130435488548661e-06, 'learning_rate/action_model': 1.8130435488548662e-05, 'epoch': 5.52}) +[2026-06-10 17:31:13,446][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.003748109797015786, 'train/grad_norm_pre_clip': 0.2406531423330307, 'timing/data': 0.02917986735701561, 'timing/model': 1.123241126537323, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 5.53}) +[2026-06-10 17:31:22,804][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 7.017230382189155e-05, 'train/grad_norm_pre_clip': 0.0020273765549063683, 'timing/data': 0.10106294229626656, 'timing/model': 0.5767959095537663, 'learning_rate/qwen_vl_interface': 1.7619680801804947e-06, 'learning_rate/action_model': 1.7619680801804947e-05, 'epoch': 5.55}) +[2026-06-10 17:31:43,933][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.0011566670145839453, 'train/grad_norm_pre_clip': 0.04641791805624962, 'timing/data': 0.02242453768849373, 'timing/model': 1.139579188078642, 'learning_rate/qwen_vl_interface': 1.7367522988516035e-06, 'learning_rate/action_model': 1.7367522988516035e-05, 'epoch': 5.56}) +[2026-06-10 17:32:02,216][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.051861923187971115, 'train/grad_norm_pre_clip': 0.751457691192627, 'timing/data': 0.021249286830425262, 'timing/model': 1.1173281222581863, 'learning_rate/qwen_vl_interface': 1.7117532311897967e-06, 'learning_rate/action_model': 1.7117532311897966e-05, 'epoch': 5.58}) +[2026-06-10 17:32:21,898][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.00025208850274793804, 'train/grad_norm_pre_clip': 0.009091631509363651, 'timing/data': 0.026702404022216797, 'timing/model': 1.322202518582344, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 5.59}) +[2026-06-10 17:32:40,319][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.0027127156499773264, 'train/grad_norm_pre_clip': 0.1969139128923416, 'timing/data': 0.0282820425927639, 'timing/model': 1.1786633618175983, 'learning_rate/qwen_vl_interface': 1.6624113916453782e-06, 'learning_rate/action_model': 1.6624113916453783e-05, 'epoch': 5.6}) +[2026-06-10 17:32:52,864][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.00048691019765101373, 'train/grad_norm_pre_clip': 0.022269802168011665, 'timing/data': 0.22288885340094566, 'timing/model': 1.0567027553915977, 'learning_rate/qwen_vl_interface': 1.6380716633998532e-06, 'learning_rate/action_model': 1.638071663399853e-05, 'epoch': 5.62}) +[2026-06-10 17:33:20,614][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.013096357695758343, 'train/grad_norm_pre_clip': 0.3937464952468872, 'timing/data': 0.012522093951702118, 'timing/model': 1.1554820574820042, 'learning_rate/qwen_vl_interface': 1.613954735907832e-06, 'learning_rate/action_model': 1.613954735907832e-05, 'epoch': 5.63}) +[2026-06-10 17:33:38,942][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.0008172732777893543, 'train/grad_norm_pre_clip': 0.03531686216592789, 'timing/data': 0.023644793778657913, 'timing/model': 1.4140349887311459, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 5.64}) +[2026-06-10 17:33:57,709][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.06423916667699814, 'train/grad_norm_pre_clip': 0.17423079907894135, 'timing/data': 0.019304480403661728, 'timing/model': 1.2140695415437222, 'learning_rate/qwen_vl_interface': 1.5663952199318869e-06, 'learning_rate/action_model': 1.5663952199318867e-05, 'epoch': 5.66}) +[2026-06-10 17:34:16,127][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.0005578064592555165, 'train/grad_norm_pre_clip': 0.014648426324129105, 'timing/data': 0.01564117893576622, 'timing/model': 1.2114934735000134, 'learning_rate/qwen_vl_interface': 1.5429555651429334e-06, 'learning_rate/action_model': 1.5429555651429335e-05, 'epoch': 5.67}) +[2026-06-10 17:34:25,656][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.000288504408672452, 'train/grad_norm_pre_clip': 0.006989678833633661, 'timing/data': 0.09216342493891716, 'timing/model': 0.4476019814610481, 'learning_rate/qwen_vl_interface': 1.5197445783164627e-06, 'learning_rate/action_model': 1.5197445783164626e-05, 'epoch': 5.68}) +[2026-06-10 17:34:47,532][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.05930487439036369, 'train/grad_norm_pre_clip': 0.16575674712657928, 'timing/data': 0.010137844830751419, 'timing/model': 1.8169100731611252, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 5.7}) +[2026-06-10 17:35:06,531][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.0008109479094855487, 'train/grad_norm_pre_clip': 0.024965234100818634, 'timing/data': 0.012977376580238342, 'timing/model': 1.153664406388998, 'learning_rate/qwen_vl_interface': 1.4740143214093276e-06, 'learning_rate/action_model': 1.4740143214093274e-05, 'epoch': 5.71}) +[2026-06-10 17:35:25,083][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.00015825631271582097, 'train/grad_norm_pre_clip': 0.004358032718300819, 'timing/data': 0.007336288690567017, 'timing/model': 1.1375712789595127, 'learning_rate/qwen_vl_interface': 1.4514978721863198e-06, 'learning_rate/action_model': 1.4514978721863198e-05, 'epoch': 5.73}) +[2026-06-10 17:35:54,797][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.026962419971823692, 'train/grad_norm_pre_clip': 3.7946274280548096, 'timing/data': 0.009460385888814926, 'timing/model': 1.3363619521260262, 'learning_rate/qwen_vl_interface': 1.4292157324670988e-06, 'learning_rate/action_model': 1.4292157324670987e-05, 'epoch': 5.74}) +[2026-06-10 17:36:04,985][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 3.2712741813156754e-05, 'train/grad_norm_pre_clip': 0.0011274436255916953, 'timing/data': 0.11187722533941269, 'timing/model': 0.5259189829230309, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 5.75}) +[2026-06-10 17:36:26,631][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.005922350566834211, 'train/grad_norm_pre_clip': 0.46828362345695496, 'timing/data': 0.010010316967964172, 'timing/model': 1.1364568062126637, 'learning_rate/qwen_vl_interface': 1.3853598648712621e-06, 'learning_rate/action_model': 1.385359864871262e-05, 'epoch': 5.77}) +[2026-06-10 17:36:45,250][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.12189614027738571, 'train/grad_norm_pre_clip': 2.6041269302368164, 'timing/data': 0.015286259353160858, 'timing/model': 1.1576414555311203, 'learning_rate/qwen_vl_interface': 1.3637888422311379e-06, 'learning_rate/action_model': 1.3637888422311379e-05, 'epoch': 5.78}) +[2026-06-10 17:37:03,729][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.0012603774666786194, 'train/grad_norm_pre_clip': 0.08190369606018066, 'timing/data': 0.007507592439651489, 'timing/model': 1.154530793428421, 'learning_rate/qwen_vl_interface': 1.3424575394009165e-06, 'learning_rate/action_model': 1.3424575394009165e-05, 'epoch': 5.79}) +[2026-06-10 17:37:21,746][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.0005672627012245357, 'train/grad_norm_pre_clip': 0.029996175318956375, 'timing/data': 0.00815567746758461, 'timing/model': 1.1408240646123886, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 5.81}) +[2026-06-10 17:37:31,083][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 5.334963498171419e-05, 'train/grad_norm_pre_clip': 0.002168468665331602, 'timing/data': 0.08671915158629417, 'timing/model': 0.4300752729177475, 'learning_rate/qwen_vl_interface': 1.3005193415629098e-06, 'learning_rate/action_model': 1.3005193415629097e-05, 'epoch': 5.82}) +[2026-06-10 17:37:50,698][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.0001995646016439423, 'train/grad_norm_pre_clip': 0.00879363901913166, 'timing/data': 0.029636472463607788, 'timing/model': 1.1063507683575153, 'learning_rate/qwen_vl_interface': 1.2799150335007174e-06, 'learning_rate/action_model': 1.2799150335007173e-05, 'epoch': 5.84}) +[2026-06-10 17:38:08,942][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.0005119366687722504, 'train/grad_norm_pre_clip': 0.022711805999279022, 'timing/data': 0.041882142424583435, 'timing/model': 1.1430509686470032, 'learning_rate/qwen_vl_interface': 1.259555618980026e-06, 'learning_rate/action_model': 1.259555618980026e-05, 'epoch': 5.85}) +[2026-06-10 17:38:27,158][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.00029848350095562637, 'train/grad_norm_pre_clip': 0.007558637298643589, 'timing/data': 0.025810834020376205, 'timing/model': 1.265890371054411, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 5.86}) +[2026-06-10 17:38:48,032][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.28302526473999023, 'train/grad_norm_pre_clip': 12.49207878112793, 'timing/data': 0.028407033532857895, 'timing/model': 1.1472839638590813, 'learning_rate/qwen_vl_interface': 1.219576478837858e-06, 'learning_rate/action_model': 1.219576478837858e-05, 'epoch': 5.88}) +[2026-06-10 17:38:57,649][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 5.294143920764327e-05, 'train/grad_norm_pre_clip': 0.0014901375398039818, 'timing/data': 0.08507714793086052, 'timing/model': 0.5324139557778835, 'learning_rate/qwen_vl_interface': 1.1999592193180623e-06, 'learning_rate/action_model': 1.1999592193180623e-05, 'epoch': 5.89}) +[2026-06-10 17:39:16,682][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.00014322728384286165, 'train/grad_norm_pre_clip': 0.005553013179451227, 'timing/data': 0.010647814720869064, 'timing/model': 1.1490683257579803, 'learning_rate/qwen_vl_interface': 1.180591785391015e-06, 'learning_rate/action_model': 1.1805917853910148e-05, 'epoch': 5.9}) +[2026-06-10 17:39:35,611][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.00017064413987100124, 'train/grad_norm_pre_clip': 0.009617057628929615, 'timing/data': 0.008129261434078217, 'timing/model': 1.1537154093384743, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 5.92}) +[2026-06-10 17:39:55,706][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.34102049469947815, 'train/grad_norm_pre_clip': 1.4021962881088257, 'timing/data': 0.011926360428333282, 'timing/model': 1.2342868521809578, 'learning_rate/qwen_vl_interface': 1.1426111575292639e-06, 'learning_rate/action_model': 1.1426111575292639e-05, 'epoch': 5.93}) +[2026-06-10 17:40:16,137][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 8.54197860462591e-05, 'train/grad_norm_pre_clip': 0.003859087824821472, 'timing/data': 0.007435105741024017, 'timing/model': 1.2819210216403008, 'learning_rate/qwen_vl_interface': 1.1240003064185923e-06, 'learning_rate/action_model': 1.1240003064185923e-05, 'epoch': 5.95}) +[2026-06-10 17:40:26,073][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.0002849932061508298, 'train/grad_norm_pre_clip': 0.010527054779231548, 'timing/data': 0.15045693144202232, 'timing/model': 0.4379640370607376, 'learning_rate/qwen_vl_interface': 1.1056439664042135e-06, 'learning_rate/action_model': 1.1056439664042134e-05, 'epoch': 5.96}) +[2026-06-10 17:40:46,346][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.043540775775909424, 'train/grad_norm_pre_clip': 1.8206816911697388, 'timing/data': 0.008487619459629059, 'timing/model': 1.7733174376189709, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 5.97}) +[2026-06-10 17:41:07,243][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.0003648281854111701, 'train/grad_norm_pre_clip': 0.016400404274463654, 'timing/data': 0.011380769312381744, 'timing/model': 1.125558152794838, 'learning_rate/qwen_vl_interface': 1.0696993331171232e-06, 'learning_rate/action_model': 1.0696993331171231e-05, 'epoch': 5.99}) +[2026-06-10 17:41:25,446][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.0002823495597112924, 'train/grad_norm_pre_clip': 0.022294051945209503, 'timing/data': 0.010041248053312302, 'timing/model': 1.1533149033784866, 'learning_rate/qwen_vl_interface': 1.052113257078706e-06, 'learning_rate/action_model': 1.0521132570787059e-05, 'epoch': 6.0}) +[2026-06-10 17:41:48,147][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.15593819320201874, 'train/grad_norm_pre_clip': 10.84642505645752, 'timing/data': 0.013436291366815567, 'timing/model': 1.1307047568261623, 'learning_rate/qwen_vl_interface': 1.034786126468405e-06, 'learning_rate/action_model': 1.034786126468405e-05, 'epoch': 6.01}) +[2026-06-10 17:41:58,184][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.0001655200612731278, 'train/grad_norm_pre_clip': 0.007135049905627966, 'timing/data': 0.09590969234704971, 'timing/model': 0.44274063408374786, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 6.03}) +[2026-06-10 17:42:22,201][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 5.695725485566072e-05, 'train/grad_norm_pre_clip': 0.0012522306060418487, 'timing/data': 0.02399466559290886, 'timing/model': 1.366932451725006, 'learning_rate/qwen_vl_interface': 1.0009129607693812e-06, 'learning_rate/action_model': 1.0009129607693813e-05, 'epoch': 6.04}) +[2026-06-10 17:42:41,276][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.00040764198638498783, 'train/grad_norm_pre_clip': 0.017647435888648033, 'timing/data': 0.020457357168197632, 'timing/model': 1.162821240723133, 'learning_rate/qwen_vl_interface': 9.843690151370762e-07, 'learning_rate/action_model': 9.843690151370762e-06, 'epoch': 6.05}) +[2026-06-10 17:42:59,866][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.01995602808892727, 'train/grad_norm_pre_clip': 0.11594890803098679, 'timing/data': 0.024411551654338837, 'timing/model': 1.1192798502743244, 'learning_rate/qwen_vl_interface': 9.680881937168323e-07, 'learning_rate/action_model': 9.680881937168324e-06, 'epoch': 6.07}) +[2026-06-10 17:43:20,899][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.022450342774391174, 'train/grad_norm_pre_clip': 0.21130409836769104, 'timing/data': 0.016052186489105225, 'timing/model': 1.3445488698780537, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 6.08}) +[2026-06-10 17:43:31,232][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 9.496096026850864e-05, 'train/grad_norm_pre_clip': 0.00460277684032917, 'timing/data': 0.09849521890282631, 'timing/model': 1.4216489717364311, 'learning_rate/qwen_vl_interface': 9.363199243308635e-07, 'learning_rate/action_model': 9.363199243308635e-06, 'epoch': 6.1}) +[2026-06-10 17:43:55,697][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.00011317402822896838, 'train/grad_norm_pre_clip': 0.004155631177127361, 'timing/data': 0.03625142201781273, 'timing/model': 1.1471311450004578, 'learning_rate/qwen_vl_interface': 9.208344359816349e-07, 'learning_rate/action_model': 9.208344359816348e-06, 'epoch': 6.11}) +[2026-06-10 17:44:15,020][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.1430189311504364, 'train/grad_norm_pre_clip': 6.853608131408691, 'timing/data': 0.0326203890144825, 'timing/model': 1.1411217749118805, 'learning_rate/qwen_vl_interface': 9.056159909565854e-07, 'learning_rate/action_model': 9.056159909565854e-06, 'epoch': 6.12}) +[2026-06-10 17:44:32,888][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.12236468493938446, 'train/grad_norm_pre_clip': 8.876688957214355, 'timing/data': 0.03606203570961952, 'timing/model': 1.1771134436130524, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 6.14}) +[2026-06-10 17:44:51,747][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.0235920250415802, 'train/grad_norm_pre_clip': 0.6541281938552856, 'timing/data': 0.08491901308298111, 'timing/model': 1.137705821543932, 'learning_rate/qwen_vl_interface': 8.759839693301256e-07, 'learning_rate/action_model': 8.759839693301255e-06, 'epoch': 6.15}) +[2026-06-10 17:45:08,982][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 450 +[2026-06-10 17:45:08,984][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 8.843469549901783e-05, 'train/grad_norm_pre_clip': 0.004030302166938782, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/mean_length': 16.8, 'rl_games_eval/mid_train/std_reward': 164.1876467304262, 'rl_games_eval/mid_train/std_length': 5.114684741017769, 'rl_games_eval/mid_train/macro_mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/macro_mean_length': 16.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_length': 16.8, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_reward': 164.1876467304262, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_length': 5.114684741017769, 'rl_games_eval/mid_train/deadly_corridor__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 235.99388122558594, 'checkpoint/best_score': 257.1217102050781, 'checkpoint/best_step': 50.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.07416069135069847, 'timing/model': 0.7084418050944805, 'learning_rate/qwen_vl_interface': 8.615722205713881e-07, 'learning_rate/action_model': 8.61572220571388e-06, 'epoch': 6.16}) +[2026-06-10 17:45:08,993][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_450_state +[2026-06-10 17:45:08,998][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 17:45:15,501][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_450_state/model.safetensors +[2026-06-10 17:45:26,352][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_450_state/optimizer.bin +[2026-06-10 17:45:26,353][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_450_state/sampler.bin +[2026-06-10 17:45:26,353][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_450_state/sampler_1.bin +[2026-06-10 17:45:26,356][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_450_state/random_states_0.pkl +[2026-06-10 17:45:26,365][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 450 +[2026-06-10 17:45:26,365][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 17:45:26,378][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 17:46:01,202][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.0668807104229927, 'train/grad_norm_pre_clip': 0.2951046824455261, 'timing/data': 0.03937048465013504, 'timing/model': 1.268674824386835, 'learning_rate/qwen_vl_interface': 8.474311707094262e-07, 'learning_rate/action_model': 8.474311707094263e-06, 'epoch': 6.18}) +[2026-06-10 17:46:20,257][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.01824377104640007, 'train/grad_norm_pre_clip': 0.09333423525094986, 'timing/data': 0.03834632411599159, 'timing/model': 1.1393312029540539, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 6.19}) +[2026-06-10 17:46:38,526][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.11964753270149231, 'train/grad_norm_pre_clip': 10.14875602722168, 'timing/data': 0.031071096658706665, 'timing/model': 1.1594778932631016, 'learning_rate/qwen_vl_interface': 8.199646400703035e-07, 'learning_rate/action_model': 8.199646400703035e-06, 'epoch': 6.21}) +[2026-06-10 17:46:57,648][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.003546525491401553, 'train/grad_norm_pre_clip': 0.492775559425354, 'timing/data': 0.03428119793534279, 'timing/model': 1.1227402426302433, 'learning_rate/qwen_vl_interface': 8.066408535581304e-07, 'learning_rate/action_model': 8.066408535581304e-06, 'epoch': 6.22}) +[2026-06-10 17:47:08,819][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.004025709815323353, 'train/grad_norm_pre_clip': 0.18098366260528564, 'timing/data': 0.13666266575455666, 'timing/model': 0.5581258051097393, 'learning_rate/qwen_vl_interface': 7.935911543682008e-07, 'learning_rate/action_model': 7.935911543682007e-06, 'epoch': 6.23}) +[2026-06-10 17:47:31,953][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.061709437519311905, 'train/grad_norm_pre_clip': 0.27866795659065247, 'timing/data': 0.028309404850006104, 'timing/model': 1.232435766607523, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 6.25}) +[2026-06-10 17:47:50,591][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.001115121878683567, 'train/grad_norm_pre_clip': 0.04814448207616806, 'timing/data': 0.036589737981557846, 'timing/model': 1.3177353702485561, 'learning_rate/qwen_vl_interface': 7.683172208660809e-07, 'learning_rate/action_model': 7.68317220866081e-06, 'epoch': 6.26}) +[2026-06-10 17:48:08,730][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.24754849076271057, 'train/grad_norm_pre_clip': 2.1413562297821045, 'timing/data': 0.023567289113998413, 'timing/model': 1.1399752981960773, 'learning_rate/qwen_vl_interface': 7.560945455691598e-07, 'learning_rate/action_model': 7.560945455691598e-06, 'epoch': 6.27}) +[2026-06-10 17:48:28,000][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.04436331242322922, 'train/grad_norm_pre_clip': 0.1459244191646576, 'timing/data': 0.0320759192109108, 'timing/model': 1.2310237884521484, 'learning_rate/qwen_vl_interface': 7.441490755288531e-07, 'learning_rate/action_model': 7.441490755288531e-06, 'epoch': 6.29}) +[2026-06-10 17:48:37,664][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 9.73306960077025e-05, 'train/grad_norm_pre_clip': 0.0021695513278245926, 'timing/data': 0.08590452745556831, 'timing/model': 0.4557858482003212, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 6.3}) +[2026-06-10 17:48:58,200][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.10446324944496155, 'train/grad_norm_pre_clip': 1.4956353902816772, 'timing/data': 0.018312331289052963, 'timing/model': 1.3165921345353127, 'learning_rate/qwen_vl_interface': 7.210926814847443e-07, 'learning_rate/action_model': 7.210926814847444e-06, 'epoch': 6.32}) +[2026-06-10 17:49:17,622][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.0007199136889539659, 'train/grad_norm_pre_clip': 0.05299428105354309, 'timing/data': 0.024904485791921616, 'timing/model': 1.1546661257743835, 'learning_rate/qwen_vl_interface': 7.09983179707932e-07, 'learning_rate/action_model': 7.099831797079319e-06, 'epoch': 6.33}) +[2026-06-10 17:49:39,836][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.07636832445859909, 'train/grad_norm_pre_clip': 0.24499519169330597, 'timing/data': 0.03015150874853134, 'timing/model': 1.1572212763130665, 'learning_rate/qwen_vl_interface': 6.991537275539842e-07, 'learning_rate/action_model': 6.991537275539842e-06, 'epoch': 6.34}) +[2026-06-10 17:50:02,897][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.0010181036777794361, 'train/grad_norm_pre_clip': 0.1821737140417099, 'timing/data': 0.018945448100566864, 'timing/model': 1.3517905995249748, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 6.36}) +[2026-06-10 17:50:12,593][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 9.564666834194213e-05, 'train/grad_norm_pre_clip': 0.0045265075750648975, 'timing/data': 0.0948081836104393, 'timing/model': 0.44791433215141296, 'learning_rate/qwen_vl_interface': 6.78337626845176e-07, 'learning_rate/action_model': 6.783376268451759e-06, 'epoch': 6.37}) +[2026-06-10 17:50:32,793][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.0009922549361363053, 'train/grad_norm_pre_clip': 0.05334904417395592, 'timing/data': 0.018136274069547653, 'timing/model': 1.3359144367277622, 'learning_rate/qwen_vl_interface': 6.683522623254591e-07, 'learning_rate/action_model': 6.683522623254591e-06, 'epoch': 6.38}) +[2026-06-10 17:50:52,356][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.08573219180107117, 'train/grad_norm_pre_clip': 2.0601346492767334, 'timing/data': 0.019422106444835663, 'timing/model': 1.302593644708395, 'learning_rate/qwen_vl_interface': 6.586495154196905e-07, 'learning_rate/action_model': 6.586495154196905e-06, 'epoch': 6.4}) +[2026-06-10 17:51:10,482][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.0018550842069089413, 'train/grad_norm_pre_clip': 0.20604261755943298, 'timing/data': 0.020913049578666687, 'timing/model': 1.1931048035621643, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 6.41}) +[2026-06-10 17:51:28,868][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.3748895823955536, 'train/grad_norm_pre_clip': 1.9782261848449707, 'timing/data': 0.023339860141277313, 'timing/model': 1.279220201075077, 'learning_rate/qwen_vl_interface': 6.400942510244234e-07, 'learning_rate/action_model': 6.400942510244233e-06, 'epoch': 6.42}) +[2026-06-10 17:51:40,187][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.0005198487197048962, 'train/grad_norm_pre_clip': 0.03480864688754082, 'timing/data': 0.08598412573337555, 'timing/model': 0.43515876308083534, 'learning_rate/qwen_vl_interface': 6.312428781110364e-07, 'learning_rate/action_model': 6.312428781110363e-06, 'epoch': 6.44}) +[2026-06-10 17:52:01,425][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.0017043424304574728, 'train/grad_norm_pre_clip': 0.08124461024999619, 'timing/data': 0.009864185005426407, 'timing/model': 1.1734507270157337, 'learning_rate/qwen_vl_interface': 6.226764118932156e-07, 'learning_rate/action_model': 6.226764118932156e-06, 'epoch': 6.45}) +[2026-06-10 17:52:32,172][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.0013063744409009814, 'train/grad_norm_pre_clip': 0.03257964551448822, 'timing/data': 0.0184616819024086, 'timing/model': 1.1964656449854374, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 6.47}) +[2026-06-10 17:52:53,034][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.0003928066580556333, 'train/grad_norm_pre_clip': 0.026546385139226913, 'timing/data': 0.008282743394374847, 'timing/model': 1.1575458087027073, 'learning_rate/qwen_vl_interface': 6.064002956172447e-07, 'learning_rate/action_model': 6.064002956172447e-06, 'epoch': 6.48}) +[2026-06-10 17:53:11,725][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.0007167403819039464, 'train/grad_norm_pre_clip': 0.042781807482242584, 'timing/data': 0.016046788543462753, 'timing/model': 1.5134005434811115, 'learning_rate/qwen_vl_interface': 5.986916495466121e-07, 'learning_rate/action_model': 5.986916495466122e-06, 'epoch': 6.49}) +[2026-06-10 17:53:23,046][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 6.412772199837491e-05, 'train/grad_norm_pre_clip': 0.003318044124171138, 'timing/data': 0.22013963758945465, 'timing/model': 1.0235981829464436, 'learning_rate/qwen_vl_interface': 5.912699180846554e-07, 'learning_rate/action_model': 5.9126991808465545e-06, 'epoch': 6.51}) +[2026-06-10 17:53:42,553][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.0003116678853984922, 'train/grad_norm_pre_clip': 0.009520363993942738, 'timing/data': 0.014437556266784668, 'timing/model': 1.1801817417144775, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 6.52}) +[2026-06-10 17:54:03,231][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.2751065492630005, 'train/grad_norm_pre_clip': 1.1725027561187744, 'timing/data': 0.017406780272722244, 'timing/model': 2.9004920534789562, 'learning_rate/qwen_vl_interface': 5.772890124897043e-07, 'learning_rate/action_model': 5.772890124897042e-06, 'epoch': 6.53}) +[2026-06-10 17:54:23,095][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.040159568190574646, 'train/grad_norm_pre_clip': 1.2871460914611816, 'timing/data': 0.01799539104104042, 'timing/model': 1.1242426559329033, 'learning_rate/qwen_vl_interface': 5.707307007648241e-07, 'learning_rate/action_model': 5.707307007648241e-06, 'epoch': 6.55}) +[2026-06-10 17:54:40,996][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.000740261806640774, 'train/grad_norm_pre_clip': 0.03437098488211632, 'timing/data': 0.015882112085819244, 'timing/model': 1.1073269844055176, 'learning_rate/qwen_vl_interface': 5.644610284116467e-07, 'learning_rate/action_model': 5.644610284116467e-06, 'epoch': 6.56}) +[2026-06-10 17:54:50,537][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.23350423574447632, 'train/grad_norm_pre_clip': 1.691593885421753, 'timing/data': 0.08622850477695465, 'timing/model': 0.4527740105986595, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 6.58}) +[2026-06-10 17:55:12,138][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.0010059723863378167, 'train/grad_norm_pre_clip': 0.029648497700691223, 'timing/data': 0.02563430368900299, 'timing/model': 1.149127520620823, 'learning_rate/qwen_vl_interface': 5.527891309636035e-07, 'learning_rate/action_model': 5.527891309636035e-06, 'epoch': 6.59}) +[2026-06-10 17:55:30,958][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.0010159699013456702, 'train/grad_norm_pre_clip': 0.04243822023272514, 'timing/data': 0.0289805568754673, 'timing/model': 1.2007917687296867, 'learning_rate/qwen_vl_interface': 5.473876258463517e-07, 'learning_rate/action_model': 5.473876258463516e-06, 'epoch': 6.6}) +[2026-06-10 17:55:51,461][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.05303728207945824, 'train/grad_norm_pre_clip': 0.26799479126930237, 'timing/data': 0.035878222435712814, 'timing/model': 1.3342799916863441, 'learning_rate/qwen_vl_interface': 5.422762000116201e-07, 'learning_rate/action_model': 5.4227620001162e-06, 'epoch': 6.62}) +[2026-06-10 17:56:13,733][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.020127825438976288, 'train/grad_norm_pre_clip': 0.11581555008888245, 'timing/data': 0.027495581656694412, 'timing/model': 3.177072923630476, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 6.63}) +[2026-06-10 17:56:23,333][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.002990393666550517, 'train/grad_norm_pre_clip': 0.37363144755363464, 'timing/data': 0.0881655216217041, 'timing/model': 0.4330289289355278, 'learning_rate/qwen_vl_interface': 5.329248294641001e-07, 'learning_rate/action_model': 5.329248294641001e-06, 'epoch': 6.64}) +[2026-06-10 17:56:44,460][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.0003442430461291224, 'train/grad_norm_pre_clip': 0.019002895802259445, 'timing/data': 0.04906374216079712, 'timing/model': 1.295610886067152, 'learning_rate/qwen_vl_interface': 5.286854615878969e-07, 'learning_rate/action_model': 5.286854615878969e-06, 'epoch': 6.66}) +[2026-06-10 17:57:05,698][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.02034675143659115, 'train/grad_norm_pre_clip': 0.11095807701349258, 'timing/data': 0.033453937619924545, 'timing/model': 1.9665419980883598, 'learning_rate/qwen_vl_interface': 5.247373266317992e-07, 'learning_rate/action_model': 5.2473732663179916e-06, 'epoch': 6.67}) +[2026-06-10 17:57:24,262][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.15434502065181732, 'train/grad_norm_pre_clip': 12.073455810546875, 'timing/data': 0.04734406992793083, 'timing/model': 1.1506593599915504, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 6.68}) +[2026-06-10 17:57:44,635][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.005116784013807774, 'train/grad_norm_pre_clip': 0.5134771466255188, 'timing/data': 0.041006579995155334, 'timing/model': 1.1469655334949493, 'learning_rate/qwen_vl_interface': 5.17715711658529e-07, 'learning_rate/action_model': 5.17715711658529e-06, 'epoch': 6.7}) +[2026-06-10 17:57:55,063][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.0005014890339225531, 'train/grad_norm_pre_clip': 0.019961684942245483, 'timing/data': 0.08760533854365349, 'timing/model': 0.4575648084282875, 'learning_rate/qwen_vl_interface': 5.146426647676421e-07, 'learning_rate/action_model': 5.146426647676421e-06, 'epoch': 6.71}) +[2026-06-10 17:58:20,871][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.04879506304860115, 'train/grad_norm_pre_clip': 0.1980522722005844, 'timing/data': 0.019471272826194763, 'timing/model': 1.132724866271019, 'learning_rate/qwen_vl_interface': 5.118617170227173e-07, 'learning_rate/action_model': 5.118617170227172e-06, 'epoch': 6.73}) +[2026-06-10 17:58:40,065][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.002362103434279561, 'train/grad_norm_pre_clip': 0.07708486169576645, 'timing/data': 0.0234205424785614, 'timing/model': 1.268509428948164, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 6.74}) +[2026-06-10 17:59:00,244][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.03533938527107239, 'train/grad_norm_pre_clip': 1.7669757604599, 'timing/data': 0.019246693700551987, 'timing/model': 1.328523639589548, 'learning_rate/qwen_vl_interface': 5.071767871099448e-07, 'learning_rate/action_model': 5.071767871099448e-06, 'epoch': 6.75}) +[2026-06-10 17:59:18,971][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.04609562084078789, 'train/grad_norm_pre_clip': 0.1817934811115265, 'timing/data': 0.028461474925279617, 'timing/model': 1.2150029353797436, 'learning_rate/qwen_vl_interface': 5.052730939306429e-07, 'learning_rate/action_model': 5.052730939306429e-06, 'epoch': 6.77}) +[2026-06-10 17:59:30,104][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.0001978241343749687, 'train/grad_norm_pre_clip': 0.009969300590455532, 'timing/data': 0.10022183880209923, 'timing/model': 1.0216223262250423, 'learning_rate/qwen_vl_interface': 5.036620778565662e-07, 'learning_rate/action_model': 5.036620778565662e-06, 'epoch': 6.78}) +[2026-06-10 17:59:50,627][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.000659876735880971, 'train/grad_norm_pre_clip': 0.023877214640378952, 'timing/data': 0.03252991661429405, 'timing/model': 1.1963950246572495, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 6.79}) +[2026-06-10 18:00:09,276][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.0457756370306015, 'train/grad_norm_pre_clip': 0.23531390726566315, 'timing/data': 0.03448925539851189, 'timing/model': 1.1372043564915657, 'learning_rate/qwen_vl_interface': 5.013184564644973e-07, 'learning_rate/action_model': 5.013184564644973e-06, 'epoch': 6.81}) +[2026-06-10 18:00:27,395][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.001349646714515984, 'train/grad_norm_pre_clip': 0.055385977029800415, 'timing/data': 0.03086160123348236, 'timing/model': 1.1436088606715202, 'learning_rate/qwen_vl_interface': 5.005859957121123e-07, 'learning_rate/action_model': 5.0058599571211224e-06, 'epoch': 6.82}) +[2026-06-10 18:00:45,615][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.04615437611937523, 'train/grad_norm_pre_clip': 0.14579395949840546, 'timing/data': 0.0377093143761158, 'timing/model': 1.2855401001870632, 'learning_rate/qwen_vl_interface': 5.001465011872491e-07, 'learning_rate/action_model': 5.001465011872491e-06, 'epoch': 6.84}) +[2026-06-10 18:01:06,497][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-10 18:01:06,498][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.6169770359992981, 'train/grad_norm_pre_clip': 24.90024757385254, 'eval/loss': 0.07888815551996231, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/mean_length': 16.8, 'rl_games_eval/mid_train/std_reward': 164.1876467304262, 'rl_games_eval/mid_train/std_length': 5.114684741017769, 'rl_games_eval/mid_train/macro_mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/macro_mean_length': 16.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_reward': 235.99388122558594, 'rl_games_eval/mid_train/deadly_corridor__latency_2/mean_length': 16.8, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_reward': 164.1876467304262, 'rl_games_eval/mid_train/deadly_corridor__latency_2/std_length': 5.114684741017769, 'rl_games_eval/mid_train/deadly_corridor__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 235.99388122558594, 'checkpoint/best_score': 257.1217102050781, 'checkpoint/best_step': 50.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.08684037253260612, 'timing/model': 0.44645436853170395, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 6.85}) +[2026-06-10 18:01:06,500][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_500_state +[2026-06-10 18:01:06,505][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 18:01:14,912][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_500_state/model.safetensors +[2026-06-10 18:01:25,691][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_500_state/optimizer.bin +[2026-06-10 18:01:25,692][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_500_state/sampler.bin +[2026-06-10 18:01:25,693][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_500_state/sampler_1.bin +[2026-06-10 18:01:25,695][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-10 18:01:25,709][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-10 18:01:25,709][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 18:01:25,724][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 18:01:25,724][starVLA.training.train_starvla][INFO] - Training complete. Best checkpoint step=50 score=257.1217102050781 path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/best_state +[2026-06-10 18:01:25,725][accelerate.accelerator][INFO] - Loading states from /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_2/checkpoints/best_state +[2026-06-10 18:01:52,206][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-10 18:02:43,271][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-10 18:02:43,273][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-10 18:02:43,273][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-10 18:02:43,319][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-10 18:02:44,212][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-10 18:03:31,425][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-10 18:03:31,703][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/summary.jsonl b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..d6b4d703cde264eca50aad49c46affff9660ce5b --- /dev/null +++ b/latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/summary.jsonl @@ -0,0 +1,10 @@ +{"steps": 50} +{"steps": 100} +{"steps": 150} +{"steps": 200} +{"steps": 250} +{"steps": 300} +{"steps": 350} +{"steps": 400} +{"steps": 450} +{"steps": 500} diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/checkpoints/best_model.safetensors b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/checkpoints/best_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ec4e8e17588737fcb8a8de1540de0efeffc395b1 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/checkpoints/best_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:adb853d16890002501675633ddefc2126fd3f102a63cf6ece7a5e56d0c8c662c +size 9138230516 diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/checkpoints/steps_2000_model.safetensors b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/checkpoints/steps_2000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ec4e8e17588737fcb8a8de1540de0efeffc395b1 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/checkpoints/steps_2000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:adb853d16890002501675633ddefc2126fd3f102a63cf6ece7a5e56d0c8c662c +size 9138230516 diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/config.full.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..b22beb2860b5c6baadbdf46edcf5bb36c2bac1b7 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/config.full.yaml @@ -0,0 +1,190 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 6 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/demon_attack_fix_latency_2_small + data_mix: demon_attack_train__bridge + eval_data_mix: demon_attack_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + converted_name: demon_attack_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: demon_attack_train + mixed_converted_name: demon_attack_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 2000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: true + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2_small/checkpoints/steps_600_state + resume_step: 600 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + demon_attack: + noop_max: 30 + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/starVLA/data/demon_attack_fix_latency_2_small/demon_attack_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: demon_attack + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: demon_attack +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: true + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_demon_attack_fix_latency_2_small +output_dir: /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2_small +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/config.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..41220e1ab7d3b8f1137461f984ecb1d716c929da --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/config.yaml @@ -0,0 +1,92 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: true + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: demon_attack_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: demon_attack_train__bridge__val + include_state: true + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 6 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2_small +rl_games: + env_eval: + demon_attack: + noop_max: 30 + distributed_mode: none + enabled: true + fixed_episode_seeds: true + frameskip: 4 + image_size: 224 + latency: + prompt_map_path: /workspace/starVLA/data/demon_attack_fix_latency_2_small/demon_attack_train__bridge/latency_prompt_map.json + latency_seed_stride: 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 2 + max_steps_per_episode: 3600 + num_episodes: 5 + seed: 42 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + task_seed_stride: 0 + vectorized: + enabled: false + model_alias: openvla + task: demon_attack +run_id: openvla_demon_attack_fix_latency_2_small +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_interval: 100 + eval_num_batches: 20 + freeze_modules: '' + gradient_accumulation_steps: 16 + gradient_clipping: 1.0 + is_resume: true + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 10 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 2000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2_small/checkpoints/steps_600_state + resume_step: 600 + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/dataset_statistics.json b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..d51725efd5d585983ef196c148e39517f35db7af --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.12531481683254242, + 0.3168888986110687, + 0.1617870330810547, + 0.12757407128810883, + 0.1149444431066513, + 0.15349073708057404, + 0.0 + ], + "std": [ + 0.33097583055496216, + 0.46518969535827637, + 0.36836159229278564, + 0.33368170261383057, + 0.31906843185424805, + 0.36036959290504456, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 108000, + "num_trajectories": 30 + } +} \ No newline at end of file diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/dataset_statistics_eval.json b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..2884e483e6beda330a7563a367dbff8ae015327c --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.125083327293396, + 0.3141111135482788, + 0.16361111402511597, + 0.12720833718776703, + 0.11493055522441864, + 0.15505555272102356, + 0.0 + ], + "std": [ + 0.3307972252368927, + 0.4641607403755188, + 0.36990460753440857, + 0.33313921093940735, + 0.31889820098876953, + 0.3619128167629242, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 72000, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/hydra/.hydra/config.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..1f819928e6ab27fb4900c19033e5ae11a799e5c1 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/hydra/.hydra/config.yaml @@ -0,0 +1,188 @@ +framework: + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 6 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/demon_attack_fix_latency_2 + data_mix: demon_attack_train__bridge + eval_data_mix: demon_attack_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: demon_attack_train + mixed_converted_name: demon_attack_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 2000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: true + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_600_state + resume_step: 600 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + demon_attack: + noop_max: 30 + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/starVLA/data/demon_attack_fix_latency_2/demon_attack_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: demon_attack + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: demon_attack +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: true + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_demon_attack_fix_latency_2 +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/hydra/.hydra/hydra.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..afb4429b3e11f51dec61d3ab79888ea8597fa6a6 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/hydra/.hydra/hydra.yaml @@ -0,0 +1,251 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=demon_attack + - init=bridge + - mode=single + - run_id=openvla_demon_attack_fix_latency_2 + - run_root_dir=/workspace/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=true + - checkpoint.save_pt_file=false + - trainer.is_resume=true + - trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_600_state + - trainer.resume_step=600 + - trainer.max_train_steps=2000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=20 + - trainer.logging_frequency=10 + - trainer.gradient_accumulation_steps=16 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9, 0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=6 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/starVLA/data/demon_attack_fix_latency_2 + - datasets.vla_data.data_mix=demon_attack_train__bridge + - datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=demon_attack + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=4 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - 'rl_games.env_eval.task_description=''You are playing Demon Attack from a single + game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, + LEFTFIRE.''' + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/demon_attack_fix_latency_2/demon_attack_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=true + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[2] + - rl_games.env_eval.post_train.enabled=true + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: 'checkpoint.local.keep_last_n=1,checkpoint.save_best_model=true,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=demon_attack_train__bridge,datasets.vla_data.data_root_dir=/workspace/starVLA/data/demon_attack_fix_latency_2,datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.video_backend=torchvision_av,env=demon_attack,framework.action_model.action_dim=7,framework.action_model.action_env_dim=6,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=4,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/demon_attack_fix_latency_2/demon_attack_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=true,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[2],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=true,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_description=''You + are playing Demon Attack from a single game image. Choose exactly one action + from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.'',rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=demon_attack,run_id=openvla_demon_attack_fix_latency_2,run_root_dir=/workspace/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=20,trainer.gradient_accumulation_steps=16,trainer.gradient_clipping=1.0,trainer.is_resume=true,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=10,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=2000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9, + 0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_600_state,trainer.resume_step=600,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games' + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.2 + version_base: '1.1' + cwd: /workspace/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: demon_attack + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/hydra/.hydra/overrides.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..2116f32b5de690ffdb9811fe392f1cc52591ee0e --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/hydra/.hydra/overrides.yaml @@ -0,0 +1,88 @@ +- model=openvla +- env=demon_attack +- init=bridge +- mode=single +- run_id=openvla_demon_attack_fix_latency_2 +- run_root_dir=/workspace/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=true +- checkpoint.save_pt_file=false +- trainer.is_resume=true +- trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_600_state +- trainer.resume_step=600 +- trainer.max_train_steps=2000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=20 +- trainer.logging_frequency=10 +- trainer.gradient_accumulation_steps=16 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9, 0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=6 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/starVLA/data/demon_attack_fix_latency_2 +- datasets.vla_data.data_mix=demon_attack_train__bridge +- datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=demon_attack +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=4 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- 'rl_games.env_eval.task_description=''You are playing Demon Attack from a single + game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, + LEFTFIRE.''' +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/demon_attack_fix_latency_2/demon_attack_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=true +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[2] +- rl_games.env_eval.post_train.enabled=true +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/hydra/train_starvla_hydra.log b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..ecde356a617c62947cb6c36a23841a246edcf72e --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/hydra/train_starvla_hydra.log @@ -0,0 +1,663 @@ +[2026-06-07 07:15:01,306][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-07 07:15:01,363][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-07 07:15:02,612][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `demon_attack_train__bridge` +[2026-06-07 07:15:03,312][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-07 07:15:03,313][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-07 07:15:03,313][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-07 07:15:03,313][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-07 07:15:03,313][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-07 07:15:03,314][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-07 07:15:03,314][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-07 07:15:03,314][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-07 07:15:03,314][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-07 07:15:03,314][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-07 07:15:03,315][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-07 07:15:03,315][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-07 07:15:03,315][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-07 07:15:03,315][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-07 07:15:03,315][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-07 07:15:03,315][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-07 07:15:03,316][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-07 07:15:03,316][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-07 07:15:03,316][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-07 07:15:03,317][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-07 07:15:03,317][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-07 07:15:03,317][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-07 07:15:03,317][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-07 07:15:03,317][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-07 07:15:03,317][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-07 07:15:03,318][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-07 07:15:03,318][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-07 07:15:03,318][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-07 07:15:03,318][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-07 07:15:03,318][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-07 07:15:03,319][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-07 07:15:03,354][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `demon_attack_train__bridge__val` +[2026-06-07 07:15:03,386][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/config.full.yaml +[2026-06-07 07:15:03,392][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/config.yaml +[2026-06-07 07:15:11,656][starVLA.training.train_starvla][INFO] - Loaded pretrained checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt, steps: 0 +[2026-06-07 07:15:14,417][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-07 07:15:14,417][starVLA.training.train_starvla][INFO] - Total optimization steps = 2000 +[2026-06-07 07:15:14,417][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-07 07:15:14,417][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-07 07:15:14,417][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-07 07:20:37,848][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 1.6900874376296997, 'train/grad_norm_pre_clip': 0.7091546654701233, 'timing/data': 0.013445502147078514, 'timing/model': 2.0647044461220503, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-07 07:25:59,249][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 1.6803603172302246, 'train/grad_norm_pre_clip': 0.8270317912101746, 'timing/data': 0.007665406912565231, 'timing/model': 2.0748111885041, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-07 07:31:20,250][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 1.7036755084991455, 'train/grad_norm_pre_clip': 1.1068987846374512, 'timing/data': 0.007902853190898895, 'timing/model': 2.0778655912727118, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-07 07:36:41,548][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 1.7473605871200562, 'train/grad_norm_pre_clip': 0.8823908567428589, 'timing/data': 0.007451947778463364, 'timing/model': 2.0750297755002975, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.01}) +[2026-06-07 07:42:02,646][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 1.7572671175003052, 'train/grad_norm_pre_clip': 0.9268008470535278, 'timing/data': 0.008645052090287209, 'timing/model': 2.072696039453149, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.01}) +[2026-06-07 07:47:23,845][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 1.6156092882156372, 'train/grad_norm_pre_clip': 0.9839194416999817, 'timing/data': 0.006576148793101311, 'timing/model': 2.0803442522883415, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.02}) +[2026-06-07 07:52:44,942][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 1.6682250499725342, 'train/grad_norm_pre_clip': 2.836000680923462, 'timing/data': 0.007316950708627701, 'timing/model': 2.0784263275563717, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.02}) +[2026-06-07 07:58:05,842][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 1.7530035972595215, 'train/grad_norm_pre_clip': 1.9407379627227783, 'timing/data': 0.007310418412089348, 'timing/model': 2.063944324851036, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.02}) +[2026-06-07 08:03:27,036][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 1.6238561868667603, 'train/grad_norm_pre_clip': 1.6749229431152344, 'timing/data': 0.007433578372001648, 'timing/model': 2.055103560909629, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.03}) +[2026-06-07 08:09:06,212][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp9l1vta9d/test.c -o /tmp/tmp9l1vta9d/test.o +[2026-06-07 08:09:06,241][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp9l1vta9d/test.o -laio -o /tmp/tmp9l1vta9d/a.out +[2026-06-07 08:09:06,861][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpgflmsw53/test.c -o /tmp/tmpgflmsw53/test.o +[2026-06-07 08:09:06,891][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpgflmsw53/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpgflmsw53/a.out +[2026-06-07 08:11:15,910][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state +[2026-06-07 08:11:15,915][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-07 08:11:22,677][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/model.safetensors +[2026-06-07 08:11:30,034][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/optimizer.bin +[2026-06-07 08:11:30,035][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler.bin +[2026-06-07 08:11:30,035][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler_1.bin +[2026-06-07 08:11:30,036][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/random_states_0.pkl +[2026-06-07 08:11:30,052][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-07 08:11:30,052][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 100 +[2026-06-07 08:11:30,053][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 1.554612398147583, 'train/grad_norm_pre_clip': 3.6693525314331055, 'eval/loss': 1.7458257675170898, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 36.0, 'rl_games_eval/mid_train/mean_length': 355.2, 'rl_games_eval/mid_train/std_reward': 4.898979485566356, 'rl_games_eval/mid_train/std_length': 39.44312360855818, 'rl_games_eval/mid_train/macro_mean_reward': 36.0, 'rl_games_eval/mid_train/macro_mean_length': 355.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 36.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 355.2, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 4.898979485566356, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 39.44312360855818, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 36.0, 'checkpoint/best_score': 36.0, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.007787572219967842, 'timing/model': 2.0744138825684786, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.03}) +[2026-06-07 08:11:30,054][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_100_state +[2026-06-07 08:11:30,058][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-07 08:11:35,263][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_100_state/model.safetensors +[2026-06-07 08:11:42,673][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_100_state/optimizer.bin +[2026-06-07 08:11:42,674][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_100_state/sampler.bin +[2026-06-07 08:11:42,674][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_100_state/sampler_1.bin +[2026-06-07 08:11:42,675][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-07 08:11:42,678][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-07 08:11:42,678][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-07 08:11:42,690][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-07 08:18:03,172][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 1.5400313138961792, 'train/grad_norm_pre_clip': 1.6952404975891113, 'timing/data': 0.006287729367613792, 'timing/model': 2.4665151480585337, 'learning_rate/qwen_vl_interface': 9.999350698714228e-06, 'learning_rate/action_model': 9.999350698714228e-05, 'epoch': 0.03}) +[2026-06-07 08:24:22,677][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 1.48056161403656, 'train/grad_norm_pre_clip': 2.784954309463501, 'timing/data': 0.005710635334253311, 'timing/model': 2.462334608659148, 'learning_rate/qwen_vl_interface': 9.9974029723694e-06, 'learning_rate/action_model': 9.9974029723694e-05, 'epoch': 0.04}) +[2026-06-07 08:30:41,988][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 1.6136187314987183, 'train/grad_norm_pre_clip': 3.2916452884674072, 'timing/data': 0.005518937483429909, 'timing/model': 2.397388705983758, 'learning_rate/qwen_vl_interface': 9.994157353454447e-06, 'learning_rate/action_model': 9.994157353454448e-05, 'epoch': 0.04}) +[2026-06-07 08:37:01,096][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 1.4385911226272583, 'train/grad_norm_pre_clip': 3.4915096759796143, 'timing/data': 0.0059457216411828995, 'timing/model': 2.40259150788188, 'learning_rate/qwen_vl_interface': 9.98961472928918e-06, 'learning_rate/action_model': 9.98961472928918e-05, 'epoch': 0.04}) +[2026-06-07 08:43:21,279][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 1.4543087482452393, 'train/grad_norm_pre_clip': 6.326903343200684, 'timing/data': 0.005830803886055946, 'timing/model': 2.4777157120406628, 'learning_rate/qwen_vl_interface': 9.983776341781682e-06, 'learning_rate/action_model': 9.983776341781682e-05, 'epoch': 0.04}) +[2026-06-07 09:29:19,588][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-07 09:29:19,791][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-07 09:29:21,074][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `demon_attack_train__bridge` +[2026-06-07 09:29:21,891][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-07 09:29:21,892][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-07 09:29:21,892][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-07 09:29:21,892][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-07 09:29:21,893][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-07 09:29:21,893][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-07 09:29:21,893][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-07 09:29:21,893][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-07 09:29:21,893][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-07 09:29:21,894][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-07 09:29:21,894][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-07 09:29:21,894][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-07 09:29:21,894][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-07 09:29:21,895][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-07 09:29:21,895][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-07 09:29:21,895][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-07 09:29:21,896][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-07 09:29:21,896][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-07 09:29:21,896][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-07 09:29:21,897][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-07 09:29:21,897][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-07 09:29:21,897][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-07 09:29:21,898][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-07 09:29:21,898][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-07 09:29:21,898][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-07 09:29:21,899][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-07 09:29:21,899][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-07 09:29:21,899][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-07 09:29:21,900][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-07 09:29:21,900][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-07 09:29:21,900][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-07 09:29:21,944][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `demon_attack_train__bridge__val` +[2026-06-07 09:29:21,981][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/config.full.yaml +[2026-06-07 09:29:21,987][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/config.yaml +[2026-06-07 09:29:21,987][starVLA.training.train_starvla][INFO] - Loaded best checkpoint metadata: step=100 score=36.0 +[2026-06-07 09:29:21,988][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_100_state, steps: 100 +[2026-06-07 09:29:23,394][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_100_state +[2026-06-07 09:29:24,519][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-07 09:29:35,059][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-07 09:29:35,059][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-07 09:29:35,059][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-07 09:29:35,062][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-07 09:29:35,953][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-07 09:29:37,481][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-07 09:29:37,482][starVLA.training.train_starvla][INFO] - Total optimization steps = 2000 +[2026-06-07 09:29:37,482][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-07 09:29:37,483][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-07 09:29:37,483][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-07 09:35:04,356][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 1.8543710708618164, 'train/grad_norm_pre_clip': 3.2952487468719482, 'timing/data': 0.006262082606554031, 'timing/model': 2.076169192790985, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.03}) +[2026-06-07 09:40:25,856][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 1.6127259731292725, 'train/grad_norm_pre_clip': 3.5264153480529785, 'timing/data': 0.006098775193095207, 'timing/model': 2.078546851873398, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.04}) +[2026-06-07 09:45:46,954][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 1.5892964601516724, 'train/grad_norm_pre_clip': 2.3840339183807373, 'timing/data': 0.0061683934181928635, 'timing/model': 2.0735276360064745, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.04}) +[2026-06-07 09:51:07,852][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 1.5594425201416016, 'train/grad_norm_pre_clip': 3.202446937561035, 'timing/data': 0.0062471237033605576, 'timing/model': 2.071339910849929, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.04}) +[2026-06-07 09:56:29,155][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 1.7088286876678467, 'train/grad_norm_pre_clip': 3.023923635482788, 'timing/data': 0.00618961825966835, 'timing/model': 2.077190414071083, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.04}) +[2026-06-07 10:01:50,456][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 1.385920763015747, 'train/grad_norm_pre_clip': 2.74164080619812, 'timing/data': 0.006111029535531998, 'timing/model': 2.0780180003494024, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.05}) +[2026-06-07 10:07:11,553][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 1.3745038509368896, 'train/grad_norm_pre_clip': 2.6708290576934814, 'timing/data': 0.006019903346896172, 'timing/model': 2.0787313412874937, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.05}) +[2026-06-07 10:12:32,855][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 1.5666024684906006, 'train/grad_norm_pre_clip': 3.24245023727417, 'timing/data': 0.00877237319946289, 'timing/model': 2.0741499848663807, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.05}) +[2026-06-07 10:17:53,653][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 1.4092310667037964, 'train/grad_norm_pre_clip': 3.1214733123779297, 'timing/data': 0.006036611273884773, 'timing/model': 2.0778745114803314, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.06}) +[2026-06-07 10:29:29,301][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state +[2026-06-07 10:29:29,308][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-07 10:29:35,159][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/model.safetensors +[2026-06-07 10:29:43,453][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/optimizer.bin +[2026-06-07 10:29:43,454][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler.bin +[2026-06-07 10:29:43,454][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler_1.bin +[2026-06-07 10:29:43,456][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/random_states_0.pkl +[2026-06-07 10:29:43,473][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-07 10:29:43,473][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 200 +[2026-06-07 10:29:43,474][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 1.565604567527771, 'train/grad_norm_pre_clip': 6.555597305297852, 'eval/loss': 1.5901659727096558, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 180.0, 'rl_games_eval/mid_train/mean_length': 1050.0, 'rl_games_eval/mid_train/std_reward': 50.0, 'rl_games_eval/mid_train/std_length': 239.0213379596056, 'rl_games_eval/mid_train/macro_mean_reward': 180.0, 'rl_games_eval/mid_train/macro_mean_length': 1050.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 180.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 1050.0, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 50.0, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 239.0213379596056, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 180.0, 'checkpoint/best_score': 180.0, 'checkpoint/best_step': 200.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.006028896197676659, 'timing/model': 2.075489778071642, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.06}) +[2026-06-07 10:29:43,474][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_200_state +[2026-06-07 10:29:43,479][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-07 10:29:48,197][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_200_state/model.safetensors +[2026-06-07 10:29:56,358][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_200_state/optimizer.bin +[2026-06-07 10:29:56,359][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_200_state/sampler.bin +[2026-06-07 10:29:56,359][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_200_state/sampler_1.bin +[2026-06-07 10:29:56,360][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-07 10:29:56,363][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-07 10:29:56,363][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-07 10:29:56,376][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-07 10:36:42,247][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 1.5061947107315063, 'train/grad_norm_pre_clip': 3.3651771545410156, 'timing/data': 0.006023585796356201, 'timing/model': 2.5844460539519787, 'learning_rate/qwen_vl_interface': 9.999350698714228e-06, 'learning_rate/action_model': 9.999350698714228e-05, 'epoch': 0.06}) +[2026-06-07 10:43:29,143][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 1.3934903144836426, 'train/grad_norm_pre_clip': 2.979198932647705, 'timing/data': 0.005450040102005005, 'timing/model': 2.5750251337885857, 'learning_rate/qwen_vl_interface': 9.9974029723694e-06, 'learning_rate/action_model': 9.9974029723694e-05, 'epoch': 0.07}) +[2026-06-07 10:50:16,047][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 1.5902225971221924, 'train/grad_norm_pre_clip': 3.4807868003845215, 'timing/data': 0.0057854074984788895, 'timing/model': 2.57532175257802, 'learning_rate/qwen_vl_interface': 9.994157353454447e-06, 'learning_rate/action_model': 9.994157353454448e-05, 'epoch': 0.07}) +[2026-06-07 10:57:02,951][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 1.32206392288208, 'train/grad_norm_pre_clip': 4.716767311096191, 'timing/data': 0.0058371759951114655, 'timing/model': 2.5820524394512177, 'learning_rate/qwen_vl_interface': 9.98961472928918e-06, 'learning_rate/action_model': 9.98961472928918e-05, 'epoch': 0.07}) +[2026-06-07 11:03:50,444][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 1.2612545490264893, 'train/grad_norm_pre_clip': 4.509425163269043, 'timing/data': 0.006209947168827057, 'timing/model': 2.772025739774108, 'learning_rate/qwen_vl_interface': 9.983776341781682e-06, 'learning_rate/action_model': 9.983776341781682e-05, 'epoch': 0.07}) +[2026-06-07 11:10:36,351][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 1.0635374784469604, 'train/grad_norm_pre_clip': 3.0580520629882812, 'timing/data': 0.006030391901731491, 'timing/model': 2.589461298659444, 'learning_rate/qwen_vl_interface': 9.976643787088806e-06, 'learning_rate/action_model': 9.976643787088805e-05, 'epoch': 0.08}) +[2026-06-07 11:17:24,150][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 1.4641655683517456, 'train/grad_norm_pre_clip': 2.767932415008545, 'timing/data': 0.005857832729816437, 'timing/model': 2.566812079399824, 'learning_rate/qwen_vl_interface': 9.968219015179783e-06, 'learning_rate/action_model': 9.968219015179783e-05, 'epoch': 0.08}) +[2026-06-07 11:24:11,338][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 1.519821286201477, 'train/grad_norm_pre_clip': 3.592302083969116, 'timing/data': 0.005873983725905418, 'timing/model': 2.5801806710660458, 'learning_rate/qwen_vl_interface': 9.958504329303132e-06, 'learning_rate/action_model': 9.958504329303133e-05, 'epoch': 0.08}) +[2026-06-07 11:30:58,638][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 1.2651046514511108, 'train/grad_norm_pre_clip': 2.5724141597747803, 'timing/data': 0.0057195089757442474, 'timing/model': 2.7715201172977686, 'learning_rate/qwen_vl_interface': 9.947502385356977e-06, 'learning_rate/action_model': 9.947502385356976e-05, 'epoch': 0.09}) +[2026-06-07 11:39:50,775][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 300 +[2026-06-07 11:39:50,776][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 1.2938605546951294, 'train/grad_norm_pre_clip': 2.2695536613464355, 'eval/loss': 1.4117716550827026, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 74.0, 'rl_games_eval/mid_train/mean_length': 315.8, 'rl_games_eval/mid_train/std_reward': 28.0, 'rl_games_eval/mid_train/std_length': 62.681416703836554, 'rl_games_eval/mid_train/macro_mean_reward': 74.0, 'rl_games_eval/mid_train/macro_mean_length': 315.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 74.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 315.8, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 28.0, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 62.681416703836554, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 74.0, 'checkpoint/best_score': 180.0, 'checkpoint/best_step': 200.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.005943005904555321, 'timing/model': 2.5886406488716602, 'learning_rate/qwen_vl_interface': 9.935216191162932e-06, 'learning_rate/action_model': 9.935216191162932e-05, 'epoch': 0.09}) +[2026-06-07 11:39:50,777][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_300_state +[2026-06-07 11:39:50,782][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-07 11:39:57,163][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_300_state/model.safetensors +[2026-06-07 11:40:05,481][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_300_state/optimizer.bin +[2026-06-07 11:40:05,482][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_300_state/sampler.bin +[2026-06-07 11:40:05,482][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_300_state/sampler_1.bin +[2026-06-07 11:40:05,483][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-07 11:40:05,618][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-07 11:40:05,619][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-07 11:40:05,632][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-07 11:46:53,749][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 1.7349348068237305, 'train/grad_norm_pre_clip': 3.911410093307495, 'timing/data': 0.005588117986917496, 'timing/model': 2.559780441224575, 'learning_rate/qwen_vl_interface': 9.921649105643817e-06, 'learning_rate/action_model': 9.921649105643817e-05, 'epoch': 0.09}) +[2026-06-07 11:53:45,046][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 1.3821769952774048, 'train/grad_norm_pre_clip': 2.3163046836853027, 'timing/data': 0.0005154833197593689, 'timing/model': 2.589668219909072, 'learning_rate/qwen_vl_interface': 9.906804837905345e-06, 'learning_rate/action_model': 9.906804837905344e-05, 'epoch': 0.09}) +[2026-06-07 12:00:30,863][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 1.3274093866348267, 'train/grad_norm_pre_clip': 3.720726728439331, 'timing/data': 0.000529533252120018, 'timing/model': 2.6073569748550653, 'learning_rate/qwen_vl_interface': 9.890687446222098e-06, 'learning_rate/action_model': 9.890687446222098e-05, 'epoch': 0.1}) +[2026-06-07 12:07:16,645][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 1.4418412446975708, 'train/grad_norm_pre_clip': 3.247678279876709, 'timing/data': 0.0005364641547203064, 'timing/model': 2.5849837716668844, 'learning_rate/qwen_vl_interface': 9.873301336928028e-06, 'learning_rate/action_model': 9.873301336928028e-05, 'epoch': 0.1}) +[2026-06-07 12:14:02,229][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 1.9160956144332886, 'train/grad_norm_pre_clip': 7.47042179107666, 'timing/data': 0.0005087144672870636, 'timing/model': 2.569401064887643, 'learning_rate/qwen_vl_interface': 9.854651263211819e-06, 'learning_rate/action_model': 9.85465126321182e-05, 'epoch': 0.1}) +[2026-06-07 12:20:47,871][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 1.5078513622283936, 'train/grad_norm_pre_clip': 4.087432861328125, 'timing/data': 0.0005321148782968521, 'timing/model': 2.5191153194755316, 'learning_rate/qwen_vl_interface': 9.834742323817407e-06, 'learning_rate/action_model': 9.834742323817407e-05, 'epoch': 0.11}) +[2026-06-07 12:27:34,046][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 1.479500412940979, 'train/grad_norm_pre_clip': 3.1461801528930664, 'timing/data': 0.0005217045545578003, 'timing/model': 2.580815799534321, 'learning_rate/qwen_vl_interface': 9.81357996165003e-06, 'learning_rate/action_model': 9.81357996165003e-05, 'epoch': 0.11}) +[2026-06-07 12:34:20,237][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 1.2682108879089355, 'train/grad_norm_pre_clip': 2.099487066268921, 'timing/data': 0.0005088355392217636, 'timing/model': 2.771605519577861, 'learning_rate/qwen_vl_interface': 9.7911699622882e-06, 'learning_rate/action_model': 9.791169962288199e-05, 'epoch': 0.11}) +[2026-06-07 12:41:05,747][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 1.2863385677337646, 'train/grad_norm_pre_clip': 2.923218011856079, 'timing/data': 0.000526653602719307, 'timing/model': 2.593752834945917, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.12}) +[2026-06-07 13:00:13,099][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state +[2026-06-07 13:00:13,106][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-07 13:00:18,927][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/model.safetensors +[2026-06-07 13:00:27,037][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/optimizer.bin +[2026-06-07 13:00:27,038][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler.bin +[2026-06-07 13:00:27,038][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler_1.bin +[2026-06-07 13:00:27,039][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/random_states_0.pkl +[2026-06-07 13:00:27,054][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-07 13:00:27,054][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 400 +[2026-06-07 13:00:27,055][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 1.4796538352966309, 'train/grad_norm_pre_clip': 3.0603644847869873, 'eval/loss': 1.2227153778076172, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 2116.0, 'rl_games_eval/mid_train/mean_length': 2146.8, 'rl_games_eval/mid_train/std_reward': 1917.7236505815952, 'rl_games_eval/mid_train/std_length': 1271.8212767523587, 'rl_games_eval/mid_train/macro_mean_reward': 2116.0, 'rl_games_eval/mid_train/macro_mean_length': 2146.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 2116.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 2146.8, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 1917.7236505815952, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 1271.8212767523587, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 2116.0, 'checkpoint/best_score': 2116.0, 'checkpoint/best_step': 400.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.0005164928734302521, 'timing/model': 2.5816364120692015, 'learning_rate/qwen_vl_interface': 9.742631898078014e-06, 'learning_rate/action_model': 9.742631898078014e-05, 'epoch': 0.12}) +[2026-06-07 13:00:27,055][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_400_state +[2026-06-07 13:00:27,062][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-07 13:00:31,610][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_400_state/model.safetensors +[2026-06-07 13:00:39,781][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_400_state/optimizer.bin +[2026-06-07 13:00:39,782][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_400_state/sampler.bin +[2026-06-07 13:00:39,782][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_400_state/sampler_1.bin +[2026-06-07 13:00:39,783][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-07 13:00:39,912][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-07 13:00:39,912][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-07 13:00:39,926][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-07 13:07:26,736][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 1.1639021635055542, 'train/grad_norm_pre_clip': 2.3904974460601807, 'timing/data': 0.0005197431892156601, 'timing/model': 2.7674606312066317, 'learning_rate/qwen_vl_interface': 9.716517103051767e-06, 'learning_rate/action_model': 9.716517103051767e-05, 'epoch': 0.12}) +[2026-06-07 13:14:13,151][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 1.164440631866455, 'train/grad_norm_pre_clip': 2.611636161804199, 'timing/data': 0.0005188267678022385, 'timing/model': 2.5796876829117537, 'learning_rate/qwen_vl_interface': 9.68918120684744e-06, 'learning_rate/action_model': 9.68918120684744e-05, 'epoch': 0.12}) +[2026-06-07 13:20:59,738][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 1.2117929458618164, 'train/grad_norm_pre_clip': 3.140529155731201, 'timing/data': 0.0005074162036180496, 'timing/model': 2.772029832005501, 'learning_rate/qwen_vl_interface': 9.660631682826109e-06, 'learning_rate/action_model': 9.660631682826109e-05, 'epoch': 0.13}) +[2026-06-07 13:27:46,853][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 1.3575948476791382, 'train/grad_norm_pre_clip': 3.7867393493652344, 'timing/data': 0.0005092956125736237, 'timing/model': 2.5914175044745207, 'learning_rate/qwen_vl_interface': 9.630876336142578e-06, 'learning_rate/action_model': 9.630876336142577e-05, 'epoch': 0.13}) +[2026-06-07 13:34:34,406][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 1.1217519044876099, 'train/grad_norm_pre_clip': 2.2530369758605957, 'timing/data': 0.006252571940422058, 'timing/model': 2.8150722309947014, 'learning_rate/qwen_vl_interface': 9.599923301611523e-06, 'learning_rate/action_model': 9.599923301611523e-05, 'epoch': 0.13}) +[2026-06-07 13:41:19,846][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 1.2168798446655273, 'train/grad_norm_pre_clip': 3.308729887008667, 'timing/data': 0.005807716399431229, 'timing/model': 2.558481879532337, 'learning_rate/qwen_vl_interface': 9.567781041483523e-06, 'learning_rate/action_model': 9.567781041483523e-05, 'epoch': 0.14}) +[2026-06-07 13:48:05,532][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 1.5551680326461792, 'train/grad_norm_pre_clip': 3.6676812171936035, 'timing/data': 0.01075056567788124, 'timing/model': 2.7579600270837545, 'learning_rate/qwen_vl_interface': 9.534458343131556e-06, 'learning_rate/action_model': 9.534458343131556e-05, 'epoch': 0.14}) +[2026-06-07 13:54:51,450][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 1.2705992460250854, 'train/grad_norm_pre_clip': 2.8817100524902344, 'timing/data': 0.010361263528466225, 'timing/model': 2.5710525028407574, 'learning_rate/qwen_vl_interface': 9.499964316648628e-06, 'learning_rate/action_model': 9.499964316648627e-05, 'epoch': 0.14}) +[2026-06-07 14:01:37,640][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 1.4522151947021484, 'train/grad_norm_pre_clip': 2.41363787651062, 'timing/data': 0.010770643129944801, 'timing/model': 2.7659148517996073, 'learning_rate/qwen_vl_interface': 9.464308392357156e-06, 'learning_rate/action_model': 9.464308392357156e-05, 'epoch': 0.15}) +[2026-06-07 14:28:34,050][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state +[2026-06-07 14:28:34,057][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-07 14:28:39,732][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/model.safetensors +[2026-06-07 14:28:47,803][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/optimizer.bin +[2026-06-07 14:28:47,804][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler.bin +[2026-06-07 14:28:47,804][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler_1.bin +[2026-06-07 14:28:47,805][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/random_states_0.pkl +[2026-06-07 14:28:47,819][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-07 14:28:47,820][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-07 14:28:47,821][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 1.252345323562622, 'train/grad_norm_pre_clip': 2.778843402862549, 'eval/loss': 1.1983155012130737, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 4199.0, 'rl_games_eval/mid_train/mean_length': 3520.2, 'rl_games_eval/mid_train/std_reward': 282.0, 'rl_games_eval/mid_train/std_length': 106.6478316704095, 'rl_games_eval/mid_train/macro_mean_reward': 4199.0, 'rl_games_eval/mid_train/macro_mean_length': 3520.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 4199.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 3520.2, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 282.0, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 106.6478316704095, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 4199.0, 'checkpoint/best_score': 4199.0, 'checkpoint/best_step': 500.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.010613750666379929, 'timing/model': 2.5678577665239573, 'learning_rate/qwen_vl_interface': 9.427500318230823e-06, 'learning_rate/action_model': 9.427500318230823e-05, 'epoch': 0.15}) +[2026-06-07 14:28:47,821][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_500_state +[2026-06-07 14:28:47,825][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-07 14:28:52,567][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_500_state/model.safetensors +[2026-06-07 14:29:00,720][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_500_state/optimizer.bin +[2026-06-07 14:29:00,720][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_500_state/sampler.bin +[2026-06-07 14:29:00,721][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_500_state/sampler_1.bin +[2026-06-07 14:29:00,722][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-07 14:29:00,857][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-07 14:29:00,858][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-07 14:29:00,871][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-07 14:35:46,951][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 1.4852486848831177, 'train/grad_norm_pre_clip': 3.8823018074035645, 'timing/data': 0.010710032656788826, 'timing/model': 2.5818015187978745, 'learning_rate/qwen_vl_interface': 9.389550157229568e-06, 'learning_rate/action_model': 9.389550157229567e-05, 'epoch': 0.15}) +[2026-06-07 14:42:33,256][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 1.206125259399414, 'train/grad_norm_pre_clip': 4.759215831756592, 'timing/data': 0.0060976725071668625, 'timing/model': 2.5880944076925516, 'learning_rate/qwen_vl_interface': 9.350468284548478e-06, 'learning_rate/action_model': 9.350468284548476e-05, 'epoch': 0.15}) +[2026-06-07 14:49:24,735][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 1.3764821290969849, 'train/grad_norm_pre_clip': 4.566978454589844, 'timing/data': 0.010297155007719994, 'timing/model': 2.57461198233068, 'learning_rate/qwen_vl_interface': 9.31026538478131e-06, 'learning_rate/action_model': 9.310265384781311e-05, 'epoch': 0.16}) +[2026-06-07 14:56:11,035][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 1.3781639337539673, 'train/grad_norm_pre_clip': 2.6718320846557617, 'timing/data': 0.009367777034640312, 'timing/model': 2.775119610130787, 'learning_rate/qwen_vl_interface': 9.268952448999444e-06, 'learning_rate/action_model': 9.268952448999442e-05, 'epoch': 0.16}) +[2026-06-07 15:02:57,643][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 1.4944634437561035, 'train/grad_norm_pre_clip': 3.214156150817871, 'timing/data': 0.010721467435359955, 'timing/model': 2.5659136679023504, 'learning_rate/qwen_vl_interface': 9.22654077174701e-06, 'learning_rate/action_model': 9.226540771747011e-05, 'epoch': 0.16}) +[2026-06-07 15:09:44,549][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 1.1980315446853638, 'train/grad_norm_pre_clip': 3.0173237323760986, 'timing/data': 0.005704909563064575, 'timing/model': 2.5723236594349146, 'learning_rate/qwen_vl_interface': 9.183041947953097e-06, 'learning_rate/action_model': 9.183041947953097e-05, 'epoch': 0.17}) +[2026-06-07 15:16:31,836][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 1.3432116508483887, 'train/grad_norm_pre_clip': 2.798677921295166, 'timing/data': 0.0057713668793439865, 'timing/model': 2.763002432882786, 'learning_rate/qwen_vl_interface': 9.13846786976179e-06, 'learning_rate/action_model': 9.13846786976179e-05, 'epoch': 0.17}) +[2026-06-07 15:23:18,950][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 1.257025122642517, 'train/grad_norm_pre_clip': 3.5184385776519775, 'timing/data': 0.005656147375702858, 'timing/model': 2.586982125416398, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.17}) +[2026-06-07 15:30:05,854][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 1.1500110626220703, 'train/grad_norm_pre_clip': 2.7995529174804688, 'timing/data': 0.005922973155975342, 'timing/model': 2.5854353960603476, 'learning_rate/qwen_vl_interface': 9.046142985250894e-06, 'learning_rate/action_model': 9.046142985250893e-05, 'epoch': 0.17}) +[2026-06-07 15:39:13,376][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 600 +[2026-06-07 15:39:13,377][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 1.143204927444458, 'train/grad_norm_pre_clip': 2.6561119556427, 'eval/loss': 1.1654857397079468, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 102.0, 'rl_games_eval/mid_train/mean_length': 361.4, 'rl_games_eval/mid_train/std_reward': 89.7552226892675, 'rl_games_eval/mid_train/std_length': 253.3026648103016, 'rl_games_eval/mid_train/macro_mean_reward': 102.0, 'rl_games_eval/mid_train/macro_mean_length': 361.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 102.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 361.4, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 89.7552226892675, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 253.3026648103016, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 102.0, 'checkpoint/best_score': 4199.0, 'checkpoint/best_step': 500.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.005986994132399559, 'timing/model': 2.772297903895378, 'learning_rate/qwen_vl_interface': 8.998417419632872e-06, 'learning_rate/action_model': 8.99841741963287e-05, 'epoch': 0.18}) +[2026-06-07 15:39:13,378][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_600_state +[2026-06-07 15:39:13,383][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-07 15:39:19,262][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_600_state/model.safetensors +[2026-06-07 15:39:27,446][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_600_state/optimizer.bin +[2026-06-07 15:39:27,447][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_600_state/sampler.bin +[2026-06-07 15:39:27,447][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_600_state/sampler_1.bin +[2026-06-07 15:39:27,448][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_600_state/random_states_0.pkl +[2026-06-07 15:39:27,572][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600 +[2026-06-07 15:39:27,572][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-07 15:39:27,586][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-07 15:46:13,939][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 1.0347563028335571, 'train/grad_norm_pre_clip': 5.202309608459473, 'timing/data': 0.005799347534775734, 'timing/model': 2.7661149743944407, 'learning_rate/qwen_vl_interface': 8.949667074120035e-06, 'learning_rate/action_model': 8.949667074120035e-05, 'epoch': 0.18}) +[2026-06-07 15:53:00,137][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 1.0652070045471191, 'train/grad_norm_pre_clip': 2.7185983657836914, 'timing/data': 0.005716485902667046, 'timing/model': 2.579190745949745, 'learning_rate/qwen_vl_interface': 8.899905276570082e-06, 'learning_rate/action_model': 8.899905276570082e-05, 'epoch': 0.18}) +[2026-06-07 15:59:47,738][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 1.0377368927001953, 'train/grad_norm_pre_clip': 3.0369625091552734, 'timing/data': 0.00569278746843338, 'timing/model': 2.7690334487706423, 'learning_rate/qwen_vl_interface': 8.84914563136159e-06, 'learning_rate/action_model': 8.849145631361591e-05, 'epoch': 0.19}) +[2026-06-07 16:06:34,049][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 1.0238913297653198, 'train/grad_norm_pre_clip': 2.390184164047241, 'timing/data': 0.005958385765552521, 'timing/model': 2.5765671394765377, 'learning_rate/qwen_vl_interface': 8.797402015674731e-06, 'learning_rate/action_model': 8.797402015674732e-05, 'epoch': 0.19}) +[2026-06-07 16:13:21,249][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 1.024673581123352, 'train/grad_norm_pre_clip': 2.6085164546966553, 'timing/data': 0.005772195756435394, 'timing/model': 2.5862742997705936, 'learning_rate/qwen_vl_interface': 8.744688575697376e-06, 'learning_rate/action_model': 8.744688575697377e-05, 'epoch': 0.19}) +[2026-06-08 07:02:40,362][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-08 07:02:40,374][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-08 07:02:41,627][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `demon_attack_train__bridge` +[2026-06-08 07:02:42,334][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-08 07:02:42,335][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-08 07:02:42,335][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-08 07:02:42,335][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-08 07:02:42,335][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-08 07:02:42,335][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-08 07:02:42,336][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-08 07:02:42,336][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-08 07:02:42,336][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-08 07:02:42,336][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-08 07:02:42,337][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-08 07:02:42,337][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-08 07:02:42,337][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-08 07:02:42,338][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-08 07:02:42,338][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-08 07:02:42,338][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-08 07:02:42,338][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-08 07:02:42,338][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-08 07:02:42,338][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-08 07:02:42,339][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-08 07:02:42,339][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-08 07:02:42,339][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-08 07:02:42,340][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-08 07:02:42,341][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-08 07:02:42,341][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-08 07:02:42,341][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-08 07:02:42,341][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-08 07:02:42,341][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-08 07:02:42,342][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-08 07:02:42,342][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-08 07:02:42,342][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-08 07:02:42,377][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `demon_attack_train__bridge__val` +[2026-06-08 07:02:42,406][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/config.full.yaml +[2026-06-08 07:02:42,413][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/config.yaml +[2026-06-08 07:02:42,413][starVLA.training.train_starvla][INFO] - Loaded best checkpoint metadata: step=500 score=4199.0 +[2026-06-08 07:02:42,413][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_600_state, steps: 600 +[2026-06-08 07:02:43,810][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_600_state +[2026-06-08 07:02:44,897][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-08 07:02:54,578][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-08 07:02:54,579][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-08 07:02:54,579][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-08 07:02:54,581][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-08 07:02:55,532][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-08 07:02:57,105][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-08 07:02:57,106][starVLA.training.train_starvla][INFO] - Total optimization steps = 2000 +[2026-06-08 07:02:57,106][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-08 07:02:57,107][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-08 07:02:57,107][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-08 07:08:08,675][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 1.420785665512085, 'train/grad_norm_pre_clip': 2.8695552349090576, 'timing/data': 0.015960225835442543, 'timing/model': 1.964878935366869, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.18}) +[2026-06-08 07:13:13,975][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 1.1499197483062744, 'train/grad_norm_pre_clip': 2.3909568786621094, 'timing/data': 0.015721529722213745, 'timing/model': 1.969408867880702, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.18}) +[2026-06-08 07:18:19,275][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 1.1300479173660278, 'train/grad_norm_pre_clip': 3.258448362350464, 'timing/data': 0.015741556882858276, 'timing/model': 1.965002678334713, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.19}) +[2026-06-08 07:23:24,477][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 1.0284771919250488, 'train/grad_norm_pre_clip': 3.639848470687866, 'timing/data': 0.016583120450377464, 'timing/model': 1.967682046815753, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.19}) +[2026-06-08 07:28:30,175][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 1.243815541267395, 'train/grad_norm_pre_clip': 2.8993003368377686, 'timing/data': 0.01161062903702259, 'timing/model': 1.9686437007039785, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.19}) +[2026-06-08 07:33:35,177][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 1.0965391397476196, 'train/grad_norm_pre_clip': 2.828278064727783, 'timing/data': 0.006442766636610031, 'timing/model': 1.9788195341825485, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.2}) +[2026-06-08 07:38:40,174][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 1.22694993019104, 'train/grad_norm_pre_clip': 3.7055747509002686, 'timing/data': 0.006169548258185387, 'timing/model': 1.9769129119813442, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.2}) +[2026-06-08 07:43:45,474][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 1.2176710367202759, 'train/grad_norm_pre_clip': 2.4701030254364014, 'timing/data': 0.005775395780801773, 'timing/model': 1.977739680558443, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.2}) +[2026-06-08 07:48:50,873][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.960399866104126, 'train/grad_norm_pre_clip': 2.1588268280029297, 'timing/data': 0.010917982086539268, 'timing/model': 1.9737035259604454, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.2}) +[2026-06-08 07:54:14,975][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp5cyvc68q/test.c -o /tmp/tmp5cyvc68q/test.o +[2026-06-08 07:54:15,004][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp5cyvc68q/test.o -laio -o /tmp/tmp5cyvc68q/a.out +[2026-06-08 07:54:15,433][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpus1r5bao/test.c -o /tmp/tmpus1r5bao/test.o +[2026-06-08 07:54:15,462][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpus1r5bao/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpus1r5bao/a.out +[2026-06-08 08:03:52,174][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 700 +[2026-06-08 08:03:52,175][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 1.0270600318908691, 'train/grad_norm_pre_clip': 3.6088926792144775, 'eval/loss': 1.1295944452285767, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 1400.0, 'rl_games_eval/mid_train/mean_length': 1612.6, 'rl_games_eval/mid_train/std_reward': 560.0, 'rl_games_eval/mid_train/std_length': 321.81584796277514, 'rl_games_eval/mid_train/macro_mean_reward': 1400.0, 'rl_games_eval/mid_train/macro_mean_length': 1612.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 1400.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 1612.6, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 560.0, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 321.81584796277514, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 1400.0, 'checkpoint/best_score': 4199.0, 'checkpoint/best_step': 500.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0063071828335523605, 'timing/model': 1.9811745211482048, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.21}) +[2026-06-08 08:03:52,176][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_700_state +[2026-06-08 08:03:52,180][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 08:03:58,251][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_700_state/model.safetensors +[2026-06-08 08:04:06,440][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_700_state/optimizer.bin +[2026-06-08 08:04:06,441][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_700_state/sampler.bin +[2026-06-08 08:04:06,441][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_700_state/sampler_1.bin +[2026-06-08 08:04:06,443][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_700_state/random_states_0.pkl +[2026-06-08 08:04:06,446][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700 +[2026-06-08 08:04:06,446][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 08:04:06,461][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 08:10:23,727][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 1.0295491218566895, 'train/grad_norm_pre_clip': 2.353490114212036, 'timing/data': 0.005850158631801605, 'timing/model': 2.4002269692718983, 'learning_rate/qwen_vl_interface': 9.999350698714228e-06, 'learning_rate/action_model': 9.999350698714228e-05, 'epoch': 0.21}) +[2026-06-08 08:16:41,504][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 1.0849347114562988, 'train/grad_norm_pre_clip': 4.363910675048828, 'timing/data': 0.005778739228844643, 'timing/model': 2.4610884711146355, 'learning_rate/qwen_vl_interface': 9.9974029723694e-06, 'learning_rate/action_model': 9.9974029723694e-05, 'epoch': 0.21}) +[2026-06-08 08:22:58,619][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 1.3987473249435425, 'train/grad_norm_pre_clip': 4.419915199279785, 'timing/data': 0.005917303264141083, 'timing/model': 2.460196003317833, 'learning_rate/qwen_vl_interface': 9.994157353454447e-06, 'learning_rate/action_model': 9.994157353454448e-05, 'epoch': 0.22}) +[2026-06-08 08:29:16,127][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.9841600656509399, 'train/grad_norm_pre_clip': 4.716740131378174, 'timing/data': 0.005986358970403671, 'timing/model': 2.401150621473789, 'learning_rate/qwen_vl_interface': 9.98961472928918e-06, 'learning_rate/action_model': 9.98961472928918e-05, 'epoch': 0.22}) +[2026-06-08 08:35:33,035][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.7222082614898682, 'train/grad_norm_pre_clip': 2.306041717529297, 'timing/data': 0.005698537454009056, 'timing/model': 2.397034166380763, 'learning_rate/qwen_vl_interface': 9.983776341781682e-06, 'learning_rate/action_model': 9.983776341781682e-05, 'epoch': 0.22}) +[2026-06-08 08:41:50,320][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.7258343696594238, 'train/grad_norm_pre_clip': 2.5289158821105957, 'timing/data': 0.005892306566238403, 'timing/model': 2.4617094565182924, 'learning_rate/qwen_vl_interface': 9.976643787088806e-06, 'learning_rate/action_model': 9.976643787088805e-05, 'epoch': 0.23}) +[2026-06-08 08:48:07,435][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 1.0538042783737183, 'train/grad_norm_pre_clip': 3.9133951663970947, 'timing/data': 0.006211994215846062, 'timing/model': 2.393337035551667, 'learning_rate/qwen_vl_interface': 9.968219015179783e-06, 'learning_rate/action_model': 9.968219015179783e-05, 'epoch': 0.23}) +[2026-06-08 08:54:24,622][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 1.3494739532470703, 'train/grad_norm_pre_clip': 4.405797481536865, 'timing/data': 0.0058113690465688705, 'timing/model': 2.4615849275141954, 'learning_rate/qwen_vl_interface': 9.958504329303132e-06, 'learning_rate/action_model': 9.958504329303133e-05, 'epoch': 0.23}) +[2026-06-08 09:00:41,436][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 1.0229706764221191, 'train/grad_norm_pre_clip': 4.955576419830322, 'timing/data': 0.005239175632596016, 'timing/model': 2.3966915607452393, 'learning_rate/qwen_vl_interface': 9.947502385356977e-06, 'learning_rate/action_model': 9.947502385356976e-05, 'epoch': 0.23}) +[2026-06-08 09:21:12,429][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 800 +[2026-06-08 09:21:12,430][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.9407396912574768, 'train/grad_norm_pre_clip': 2.4502463340759277, 'eval/loss': 1.113853096961975, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 3094.0, 'rl_games_eval/mid_train/mean_length': 2344.4, 'rl_games_eval/mid_train/std_reward': 2460.094307135399, 'rl_games_eval/mid_train/std_length': 1537.7896605192793, 'rl_games_eval/mid_train/macro_mean_reward': 3094.0, 'rl_games_eval/mid_train/macro_mean_length': 2344.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 3094.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 2344.4, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 2460.094307135399, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 1537.7896605192793, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 3094.0, 'checkpoint/best_score': 4199.0, 'checkpoint/best_step': 500.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00602116622030735, 'timing/model': 2.457622244954109, 'learning_rate/qwen_vl_interface': 9.935216191162932e-06, 'learning_rate/action_model': 9.935216191162932e-05, 'epoch': 0.24}) +[2026-06-08 09:21:12,431][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_800_state +[2026-06-08 09:21:12,437][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 09:21:18,987][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_800_state/model.safetensors +[2026-06-08 09:21:27,236][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_800_state/optimizer.bin +[2026-06-08 09:21:27,238][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_800_state/sampler.bin +[2026-06-08 09:21:27,238][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_800_state/sampler_1.bin +[2026-06-08 09:21:27,240][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_800_state/random_states_0.pkl +[2026-06-08 09:21:27,384][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800 +[2026-06-08 09:21:27,384][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 09:21:27,401][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 09:27:43,837][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 1.299156904220581, 'train/grad_norm_pre_clip': 4.076747417449951, 'timing/data': 0.00596633180975914, 'timing/model': 2.3979320116341114, 'learning_rate/qwen_vl_interface': 9.921649105643817e-06, 'learning_rate/action_model': 9.921649105643817e-05, 'epoch': 0.24}) +[2026-06-08 09:34:06,916][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 1.1827017068862915, 'train/grad_norm_pre_clip': 4.754055500030518, 'timing/data': 0.0005475040525197983, 'timing/model': 2.39188002794981, 'learning_rate/qwen_vl_interface': 9.906804837905345e-06, 'learning_rate/action_model': 9.906804837905344e-05, 'epoch': 0.24}) +[2026-06-08 09:40:24,334][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.8995064496994019, 'train/grad_norm_pre_clip': 3.30954909324646, 'timing/data': 0.0005342047661542892, 'timing/model': 2.3908971305936575, 'learning_rate/qwen_vl_interface': 9.890687446222098e-06, 'learning_rate/action_model': 9.890687446222098e-05, 'epoch': 0.25}) +[2026-06-08 09:46:41,616][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 1.3701348304748535, 'train/grad_norm_pre_clip': 3.946835517883301, 'timing/data': 0.0006955508142709732, 'timing/model': 2.4852495826780796, 'learning_rate/qwen_vl_interface': 9.873301336928028e-06, 'learning_rate/action_model': 9.873301336928028e-05, 'epoch': 0.25}) +[2026-06-08 09:52:58,322][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 1.591193675994873, 'train/grad_norm_pre_clip': 6.2772417068481445, 'timing/data': 0.0005408022552728653, 'timing/model': 2.3973457850515842, 'learning_rate/qwen_vl_interface': 9.854651263211819e-06, 'learning_rate/action_model': 9.85465126321182e-05, 'epoch': 0.25}) +[2026-06-08 09:59:15,127][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 1.2465392351150513, 'train/grad_norm_pre_clip': 4.804559707641602, 'timing/data': 0.0005575045943260193, 'timing/model': 2.392449790611863, 'learning_rate/qwen_vl_interface': 9.834742323817407e-06, 'learning_rate/action_model': 9.834742323817407e-05, 'epoch': 0.25}) +[2026-06-08 10:05:31,737][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 1.1705182790756226, 'train/grad_norm_pre_clip': 4.531994342803955, 'timing/data': 0.0006188936531543732, 'timing/model': 2.4620881006121635, 'learning_rate/qwen_vl_interface': 9.81357996165003e-06, 'learning_rate/action_model': 9.81357996165003e-05, 'epoch': 0.26}) +[2026-06-08 10:11:48,033][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 1.0571683645248413, 'train/grad_norm_pre_clip': 2.966623544692993, 'timing/data': 0.0005451645702123642, 'timing/model': 2.3976377453655005, 'learning_rate/qwen_vl_interface': 9.7911699622882e-06, 'learning_rate/action_model': 9.791169962288199e-05, 'epoch': 0.26}) +[2026-06-08 10:18:04,132][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.9346114993095398, 'train/grad_norm_pre_clip': 4.443077564239502, 'timing/data': 0.0005320832133293152, 'timing/model': 2.3938405234366655, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.26}) +[2026-06-08 10:41:07,419][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 900 +[2026-06-08 10:41:07,420][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 1.214288353919983, 'train/grad_norm_pre_clip': 3.4552643299102783, 'eval/loss': 1.0450712442398071, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 3511.0, 'rl_games_eval/mid_train/mean_length': 2783.2, 'rl_games_eval/mid_train/std_reward': 2121.787925312047, 'rl_games_eval/mid_train/std_length': 1126.5201995525867, 'rl_games_eval/mid_train/macro_mean_reward': 3511.0, 'rl_games_eval/mid_train/macro_mean_length': 2783.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 3511.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 2783.2, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 2121.787925312047, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 1126.5201995525867, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 3511.0, 'checkpoint/best_score': 4199.0, 'checkpoint/best_step': 500.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005379114300012589, 'timing/model': 2.4667833875864744, 'learning_rate/qwen_vl_interface': 9.742631898078014e-06, 'learning_rate/action_model': 9.742631898078014e-05, 'epoch': 0.27}) +[2026-06-08 10:41:07,420][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_900_state +[2026-06-08 10:41:07,425][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 10:41:13,321][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_900_state/model.safetensors +[2026-06-08 10:41:21,461][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_900_state/optimizer.bin +[2026-06-08 10:41:21,461][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_900_state/sampler.bin +[2026-06-08 10:41:21,461][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_900_state/sampler_1.bin +[2026-06-08 10:41:21,462][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_900_state/random_states_0.pkl +[2026-06-08 10:41:21,583][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-06-08 10:41:21,584][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 10:41:21,598][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 10:47:39,822][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.9228300452232361, 'train/grad_norm_pre_clip': 3.264188766479492, 'timing/data': 0.0005331747233867645, 'timing/model': 2.4300544932484627, 'learning_rate/qwen_vl_interface': 9.716517103051767e-06, 'learning_rate/action_model': 9.716517103051767e-05, 'epoch': 0.27}) +[2026-06-08 10:53:57,333][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 1.0301495790481567, 'train/grad_norm_pre_clip': 4.15120267868042, 'timing/data': 0.0005621928721666336, 'timing/model': 2.397552475333214, 'learning_rate/qwen_vl_interface': 9.68918120684744e-06, 'learning_rate/action_model': 9.68918120684744e-05, 'epoch': 0.27}) +[2026-06-08 11:00:14,531][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.9994879364967346, 'train/grad_norm_pre_clip': 2.8673532009124756, 'timing/data': 0.0005490239709615707, 'timing/model': 2.394269237294793, 'learning_rate/qwen_vl_interface': 9.660631682826109e-06, 'learning_rate/action_model': 9.660631682826109e-05, 'epoch': 0.28}) +[2026-06-08 11:06:30,827][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.9144805669784546, 'train/grad_norm_pre_clip': 4.228509426116943, 'timing/data': 0.0005381237715482712, 'timing/model': 2.3932620640844107, 'learning_rate/qwen_vl_interface': 9.630876336142578e-06, 'learning_rate/action_model': 9.630876336142577e-05, 'epoch': 0.28}) +[2026-06-08 11:12:46,934][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.8507940769195557, 'train/grad_norm_pre_clip': 5.005660057067871, 'timing/data': 0.0005496349185705185, 'timing/model': 2.399266691878438, 'learning_rate/qwen_vl_interface': 9.599923301611523e-06, 'learning_rate/action_model': 9.599923301611523e-05, 'epoch': 0.28}) +[2026-06-08 11:19:04,128][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 1.059356451034546, 'train/grad_norm_pre_clip': 3.986670732498169, 'timing/data': 0.0006291009485721588, 'timing/model': 2.3923126198351383, 'learning_rate/qwen_vl_interface': 9.567781041483523e-06, 'learning_rate/action_model': 9.567781041483523e-05, 'epoch': 0.28}) +[2026-06-08 11:25:21,333][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 1.3806349039077759, 'train/grad_norm_pre_clip': 5.543453216552734, 'timing/data': 0.0005384627729654312, 'timing/model': 2.3994794115424156, 'learning_rate/qwen_vl_interface': 9.534458343131556e-06, 'learning_rate/action_model': 9.534458343131556e-05, 'epoch': 0.29}) +[2026-06-08 11:31:38,922][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 1.0659602880477905, 'train/grad_norm_pre_clip': 3.5213422775268555, 'timing/data': 0.0005356352776288986, 'timing/model': 2.393177930265665, 'learning_rate/qwen_vl_interface': 9.499964316648628e-06, 'learning_rate/action_model': 9.499964316648627e-05, 'epoch': 0.29}) +[2026-06-08 11:37:54,619][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 1.3414167165756226, 'train/grad_norm_pre_clip': 6.713064670562744, 'timing/data': 0.0005318131297826767, 'timing/model': 2.467078670859337, 'learning_rate/qwen_vl_interface': 9.464308392357156e-06, 'learning_rate/action_model': 9.464308392357156e-05, 'epoch': 0.29}) +[2026-06-08 12:05:08,274][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state +[2026-06-08 12:05:08,281][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 12:05:13,825][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/model.safetensors +[2026-06-08 12:05:21,916][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/optimizer.bin +[2026-06-08 12:05:21,917][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler.bin +[2026-06-08 12:05:21,917][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler_1.bin +[2026-06-08 12:05:21,919][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/random_states_0.pkl +[2026-06-08 12:05:21,936][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 12:05:21,936][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-06-08 12:05:21,938][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.9910812377929688, 'train/grad_norm_pre_clip': 3.65725040435791, 'eval/loss': 1.0873119831085205, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 4286.0, 'rl_games_eval/mid_train/mean_length': 3415.0, 'rl_games_eval/mid_train/std_reward': 701.2160865239758, 'rl_games_eval/mid_train/std_length': 190.0820875306245, 'rl_games_eval/mid_train/macro_mean_reward': 4286.0, 'rl_games_eval/mid_train/macro_mean_length': 3415.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 4286.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 3415.0, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 701.2160865239758, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 190.0820875306245, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 4286.0, 'checkpoint/best_score': 4286.0, 'checkpoint/best_step': 1000.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.0005354136228561401, 'timing/model': 2.3911345209926367, 'learning_rate/qwen_vl_interface': 9.427500318230823e-06, 'learning_rate/action_model': 9.427500318230823e-05, 'epoch': 0.3}) +[2026-06-08 12:05:21,938][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1000_state +[2026-06-08 12:05:21,943][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 12:05:26,362][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1000_state/model.safetensors +[2026-06-08 12:05:34,566][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1000_state/optimizer.bin +[2026-06-08 12:05:34,567][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1000_state/sampler.bin +[2026-06-08 12:05:34,567][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-08 12:05:34,569][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-08 12:05:34,711][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-08 12:05:34,711][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 12:05:34,726][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 12:11:52,218][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 1.3711518049240112, 'train/grad_norm_pre_clip': 4.279980659484863, 'timing/data': 0.0005485154688358307, 'timing/model': 2.3908892571926117, 'learning_rate/qwen_vl_interface': 9.389550157229568e-06, 'learning_rate/action_model': 9.389550157229567e-05, 'epoch': 0.3}) +[2026-06-08 12:18:10,227][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 1.144603967666626, 'train/grad_norm_pre_clip': 5.90156364440918, 'timing/data': 0.0006529446691274643, 'timing/model': 2.3943443559110165, 'learning_rate/qwen_vl_interface': 9.350468284548478e-06, 'learning_rate/action_model': 9.350468284548476e-05, 'epoch': 0.3}) +[2026-06-08 12:24:33,409][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.9691801071166992, 'train/grad_norm_pre_clip': 4.518621444702148, 'timing/data': 0.000527113676071167, 'timing/model': 2.4612740725278854, 'learning_rate/qwen_vl_interface': 9.31026538478131e-06, 'learning_rate/action_model': 9.310265384781311e-05, 'epoch': 0.31}) +[2026-06-08 12:30:50,031][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 1.1329736709594727, 'train/grad_norm_pre_clip': 3.8002240657806396, 'timing/data': 0.0005397554486989975, 'timing/model': 2.3910610005259514, 'learning_rate/qwen_vl_interface': 9.268952448999444e-06, 'learning_rate/action_model': 9.268952448999442e-05, 'epoch': 0.31}) +[2026-06-08 12:37:07,312][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 1.1674929857254028, 'train/grad_norm_pre_clip': 3.9416048526763916, 'timing/data': 0.000539751723408699, 'timing/model': 2.464493567124009, 'learning_rate/qwen_vl_interface': 9.22654077174701e-06, 'learning_rate/action_model': 9.226540771747011e-05, 'epoch': 0.31}) +[2026-06-08 12:43:26,220][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.7569307088851929, 'train/grad_norm_pre_clip': 3.0707919597625732, 'timing/data': 0.0005191043019294739, 'timing/model': 2.4682109728455544, 'learning_rate/qwen_vl_interface': 9.183041947953097e-06, 'learning_rate/action_model': 9.183041947953097e-05, 'epoch': 0.31}) +[2026-06-08 12:49:44,613][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.9160227179527283, 'train/grad_norm_pre_clip': 3.9439423084259033, 'timing/data': 0.0005275756120681763, 'timing/model': 2.464166671037674, 'learning_rate/qwen_vl_interface': 9.13846786976179e-06, 'learning_rate/action_model': 9.13846786976179e-05, 'epoch': 0.32}) +[2026-06-08 12:56:03,160][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.9098962545394897, 'train/grad_norm_pre_clip': 4.375267028808594, 'timing/data': 0.0005406439304351807, 'timing/model': 2.5233623310923576, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.32}) +[2026-06-08 13:02:21,623][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.8103939294815063, 'train/grad_norm_pre_clip': 5.143364429473877, 'timing/data': 0.0005201753228902817, 'timing/model': 2.3913236502557993, 'learning_rate/qwen_vl_interface': 9.046142985250894e-06, 'learning_rate/action_model': 9.046142985250893e-05, 'epoch': 0.32}) +[2026-06-08 13:26:52,348][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state +[2026-06-08 13:26:52,356][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 13:26:57,832][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/model.safetensors +[2026-06-08 13:27:05,817][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/optimizer.bin +[2026-06-08 13:27:05,818][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler.bin +[2026-06-08 13:27:05,819][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler_1.bin +[2026-06-08 13:27:05,820][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/random_states_0.pkl +[2026-06-08 13:27:05,839][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 13:27:05,839][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1100 +[2026-06-08 13:27:05,841][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 1.0413190126419067, 'train/grad_norm_pre_clip': 3.591252326965332, 'eval/loss': 1.0032492876052856, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 4436.0, 'rl_games_eval/mid_train/mean_length': 2968.0, 'rl_games_eval/mid_train/std_reward': 2101.747844057417, 'rl_games_eval/mid_train/std_length': 774.0390171044352, 'rl_games_eval/mid_train/macro_mean_reward': 4436.0, 'rl_games_eval/mid_train/macro_mean_length': 2968.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 4436.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 2968.0, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 2101.747844057417, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 774.0390171044352, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 4436.0, 'checkpoint/best_score': 4436.0, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.0005311351269483566, 'timing/model': 2.3935585021972656, 'learning_rate/qwen_vl_interface': 8.998417419632872e-06, 'learning_rate/action_model': 8.99841741963287e-05, 'epoch': 0.33}) +[2026-06-08 13:27:05,841][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1100_state +[2026-06-08 13:27:05,847][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 13:27:11,064][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1100_state/model.safetensors +[2026-06-08 13:27:19,295][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1100_state/optimizer.bin +[2026-06-08 13:27:19,296][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1100_state/sampler.bin +[2026-06-08 13:27:19,297][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1100_state/sampler_1.bin +[2026-06-08 13:27:19,298][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1100_state/random_states_0.pkl +[2026-06-08 13:27:19,430][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-06-08 13:27:19,431][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 13:27:19,446][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 13:33:37,434][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 1.0978609323501587, 'train/grad_norm_pre_clip': 3.3093481063842773, 'timing/data': 0.0005511846393346786, 'timing/model': 2.3945086020976305, 'learning_rate/qwen_vl_interface': 8.949667074120035e-06, 'learning_rate/action_model': 8.949667074120035e-05, 'epoch': 0.33}) +[2026-06-08 13:39:55,524][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.8589965105056763, 'train/grad_norm_pre_clip': 3.001570701599121, 'timing/data': 0.0005728509277105331, 'timing/model': 2.404070422053337, 'learning_rate/qwen_vl_interface': 8.899905276570082e-06, 'learning_rate/action_model': 8.899905276570082e-05, 'epoch': 0.33}) +[2026-06-08 13:46:14,632][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.8371606469154358, 'train/grad_norm_pre_clip': 3.339404582977295, 'timing/data': 0.0005662031471729279, 'timing/model': 2.396814275532961, 'learning_rate/qwen_vl_interface': 8.84914563136159e-06, 'learning_rate/action_model': 8.849145631361591e-05, 'epoch': 0.33}) +[2026-06-08 13:52:33,228][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.811881959438324, 'train/grad_norm_pre_clip': 3.3308804035186768, 'timing/data': 0.0005858521908521652, 'timing/model': 2.404475811868906, 'learning_rate/qwen_vl_interface': 8.797402015674731e-06, 'learning_rate/action_model': 8.797402015674732e-05, 'epoch': 0.34}) +[2026-06-08 13:58:52,328][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 1.0381462574005127, 'train/grad_norm_pre_clip': 4.400218963623047, 'timing/data': 0.0005486235022544861, 'timing/model': 2.3945932500064373, 'learning_rate/qwen_vl_interface': 8.744688575697376e-06, 'learning_rate/action_model': 8.744688575697377e-05, 'epoch': 0.34}) +[2026-06-08 14:05:11,040][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.9417327642440796, 'train/grad_norm_pre_clip': 3.6548211574554443, 'timing/data': 0.0005509145557880402, 'timing/model': 2.4017536845058203, 'learning_rate/qwen_vl_interface': 8.691019722757675e-06, 'learning_rate/action_model': 8.691019722757674e-05, 'epoch': 0.34}) +[2026-06-08 14:11:30,032][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 1.0419505834579468, 'train/grad_norm_pre_clip': 3.935539722442627, 'timing/data': 0.000564301386475563, 'timing/model': 2.3968076705932617, 'learning_rate/qwen_vl_interface': 8.636410129384137e-06, 'learning_rate/action_model': 8.636410129384137e-05, 'epoch': 0.35}) +[2026-06-08 14:17:48,637][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.8424609303474426, 'train/grad_norm_pre_clip': 3.7355785369873047, 'timing/data': 0.0006769709289073944, 'timing/model': 2.397487925365567, 'learning_rate/qwen_vl_interface': 8.580874725294321e-06, 'learning_rate/action_model': 8.580874725294321e-05, 'epoch': 0.35}) +[2026-06-08 14:24:07,439][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 1.0235060453414917, 'train/grad_norm_pre_clip': 5.691439151763916, 'timing/data': 0.0006650928407907486, 'timing/model': 2.3984890412539244, 'learning_rate/qwen_vl_interface': 8.524428693313189e-06, 'learning_rate/action_model': 8.524428693313188e-05, 'epoch': 0.35}) +[2026-06-08 14:48:09,442][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state +[2026-06-08 14:48:09,448][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 14:48:15,769][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/model.safetensors +[2026-06-08 14:48:23,796][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/optimizer.bin +[2026-06-08 14:48:23,797][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler.bin +[2026-06-08 14:48:23,797][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler_1.bin +[2026-06-08 14:48:23,798][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/random_states_0.pkl +[2026-06-08 14:48:23,814][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 14:48:23,815][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1200 +[2026-06-08 14:48:23,816][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 1.3663575649261475, 'train/grad_norm_pre_clip': 3.5234622955322266, 'eval/loss': 0.9726598858833313, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 4574.0, 'rl_games_eval/mid_train/mean_length': 2910.8, 'rl_games_eval/mid_train/std_reward': 2443.019443230037, 'rl_games_eval/mid_train/std_length': 1287.0579474134022, 'rl_games_eval/mid_train/macro_mean_reward': 4574.0, 'rl_games_eval/mid_train/macro_mean_length': 2910.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 4574.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 2910.8, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 2443.019443230037, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 1287.0579474134022, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 4574.0, 'checkpoint/best_score': 4574.0, 'checkpoint/best_step': 1200.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.0005479529500007629, 'timing/model': 2.3934664484113455, 'learning_rate/qwen_vl_interface': 8.467087465222272e-06, 'learning_rate/action_model': 8.46708746522227e-05, 'epoch': 0.36}) +[2026-06-08 14:48:23,816][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1200_state +[2026-06-08 14:48:23,821][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 14:48:29,266][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1200_state/model.safetensors +[2026-06-08 14:48:37,918][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1200_state/optimizer.bin +[2026-06-08 14:48:37,919][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1200_state/sampler.bin +[2026-06-08 14:48:37,919][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1200_state/sampler_1.bin +[2026-06-08 14:48:37,920][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1200_state/random_states_0.pkl +[2026-06-08 14:48:38,012][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-06-08 14:48:38,013][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 14:48:38,027][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 14:54:56,632][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 1.076026439666748, 'train/grad_norm_pre_clip': 3.629709005355835, 'timing/data': 0.0005295220762491226, 'timing/model': 2.3928860798478127, 'learning_rate/qwen_vl_interface': 8.408866717540778e-06, 'learning_rate/action_model': 8.408866717540779e-05, 'epoch': 0.36}) +[2026-06-08 15:01:15,440][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.8805854916572571, 'train/grad_norm_pre_clip': 3.20639967918396, 'timing/data': 0.0005307849496603012, 'timing/model': 2.4013203643262386, 'learning_rate/qwen_vl_interface': 8.349782367239801e-06, 'learning_rate/action_model': 8.349782367239802e-05, 'epoch': 0.36}) +[2026-06-08 15:07:33,932][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 1.054509162902832, 'train/grad_norm_pre_clip': 5.330147743225098, 'timing/data': 0.0005524232983589172, 'timing/model': 2.3930026702582836, 'learning_rate/qwen_vl_interface': 8.289850567390754e-06, 'learning_rate/action_model': 8.289850567390755e-05, 'epoch': 0.36}) +[2026-06-08 15:13:55,932][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.8537613749504089, 'train/grad_norm_pre_clip': 3.585017681121826, 'timing/data': 0.0005316045135259628, 'timing/model': 2.3904387559741735, 'learning_rate/qwen_vl_interface': 8.229087702749302e-06, 'learning_rate/action_model': 8.229087702749301e-05, 'epoch': 0.37}) +[2026-06-08 15:20:12,828][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.8868990540504456, 'train/grad_norm_pre_clip': 4.004549980163574, 'timing/data': 0.0005488339811563492, 'timing/model': 2.3953779619187117, 'learning_rate/qwen_vl_interface': 8.167510385275923e-06, 'learning_rate/action_model': 8.167510385275923e-05, 'epoch': 0.37}) +[2026-06-08 15:26:29,837][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 1.1557300090789795, 'train/grad_norm_pre_clip': 4.920704364776611, 'timing/data': 0.0005434732884168625, 'timing/model': 2.3975362051278353, 'learning_rate/qwen_vl_interface': 8.10513544959437e-06, 'learning_rate/action_model': 8.10513544959437e-05, 'epoch': 0.37}) +[2026-06-08 15:32:47,518][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 1.0210579633712769, 'train/grad_norm_pre_clip': 3.5761446952819824, 'timing/data': 0.0005415435880422592, 'timing/model': 2.393512723967433, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.38}) +[2026-06-08 15:39:04,329][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.973196268081665, 'train/grad_norm_pre_clip': 3.5056347846984863, 'timing/data': 0.0005291532725095749, 'timing/model': 2.3983834236860275, 'learning_rate/qwen_vl_interface': 7.978061147743983e-06, 'learning_rate/action_model': 7.978061147743982e-05, 'epoch': 0.38}) +[2026-06-08 15:45:21,012][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 1.0033162832260132, 'train/grad_norm_pre_clip': 7.542726993560791, 'timing/data': 0.0005296133458614349, 'timing/model': 2.3895579110831022, 'learning_rate/qwen_vl_interface': 7.913396522420447e-06, 'learning_rate/action_model': 7.913396522420447e-05, 'epoch': 0.38}) +[2026-06-08 16:10:48,781][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1300 +[2026-06-08 16:10:48,782][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 1.1756011247634888, 'train/grad_norm_pre_clip': 3.9284379482269287, 'eval/loss': 0.9057267308235168, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 3784.0, 'rl_games_eval/mid_train/mean_length': 3167.2, 'rl_games_eval/mid_train/std_reward': 2198.295703493959, 'rl_games_eval/mid_train/std_length': 368.882311855692, 'rl_games_eval/mid_train/macro_mean_reward': 3784.0, 'rl_games_eval/mid_train/macro_mean_length': 3167.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 3784.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 3167.2, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 2198.295703493959, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 368.882311855692, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 3784.0, 'checkpoint/best_score': 4574.0, 'checkpoint/best_step': 1200.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005380157381296158, 'timing/model': 2.389451075345278, 'learning_rate/qwen_vl_interface': 7.84800375108153e-06, 'learning_rate/action_model': 7.84800375108153e-05, 'epoch': 0.39}) +[2026-06-08 16:10:48,782][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1300_state +[2026-06-08 16:10:48,787][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 16:10:54,716][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1300_state/model.safetensors +[2026-06-08 16:11:02,832][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1300_state/optimizer.bin +[2026-06-08 16:11:02,833][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1300_state/sampler.bin +[2026-06-08 16:11:02,833][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1300_state/sampler_1.bin +[2026-06-08 16:11:02,835][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1300_state/random_states_0.pkl +[2026-06-08 16:11:02,977][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-06-08 16:11:02,978][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 16:11:02,993][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 16:17:20,716][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.9165563583374023, 'train/grad_norm_pre_clip': 7.153371810913086, 'timing/data': 0.0006073042750358582, 'timing/model': 2.3909861110150814, 'learning_rate/qwen_vl_interface': 7.781900711457971e-06, 'learning_rate/action_model': 7.781900711457971e-05, 'epoch': 0.39}) +[2026-06-08 16:23:38,415][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 1.0691813230514526, 'train/grad_norm_pre_clip': 3.9372100830078125, 'timing/data': 0.0005554836243391037, 'timing/model': 2.3914110865443945, 'learning_rate/qwen_vl_interface': 7.715105475460773e-06, 'learning_rate/action_model': 7.715105475460773e-05, 'epoch': 0.39}) +[2026-06-08 16:29:56,134][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 1.004744052886963, 'train/grad_norm_pre_clip': 9.743778228759766, 'timing/data': 0.0005558952689170837, 'timing/model': 2.393678853288293, 'learning_rate/qwen_vl_interface': 7.647636304240509e-06, 'learning_rate/action_model': 7.647636304240509e-05, 'epoch': 0.39}) +[2026-06-08 16:36:15,339][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 1.3211554288864136, 'train/grad_norm_pre_clip': 5.022520065307617, 'timing/data': 0.0005535148084163666, 'timing/model': 2.402736274525523, 'learning_rate/qwen_vl_interface': 7.579511643194914e-06, 'learning_rate/action_model': 7.579511643194914e-05, 'epoch': 0.4}) +[2026-06-08 16:42:33,434][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.7104201912879944, 'train/grad_norm_pre_clip': 3.3323731422424316, 'timing/data': 0.00607709214091301, 'timing/model': 2.4073367808014154, 'learning_rate/qwen_vl_interface': 7.510750116926101e-06, 'learning_rate/action_model': 7.510750116926101e-05, 'epoch': 0.4}) +[2026-06-08 16:48:52,435][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.8317937850952148, 'train/grad_norm_pre_clip': 7.390615463256836, 'timing/data': 0.005795003846287727, 'timing/model': 2.4032735768705606, 'learning_rate/qwen_vl_interface': 7.441370524148768e-06, 'learning_rate/action_model': 7.441370524148768e-05, 'epoch': 0.4}) +[2026-06-08 16:55:12,719][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 1.150464415550232, 'train/grad_norm_pre_clip': 5.173668384552002, 'timing/data': 0.005909815430641174, 'timing/model': 2.4743913542479277, 'learning_rate/qwen_vl_interface': 7.371391832550833e-06, 'learning_rate/action_model': 7.371391832550832e-05, 'epoch': 0.41}) +[2026-06-08 17:01:30,017][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 1.0232315063476562, 'train/grad_norm_pre_clip': 4.20911979675293, 'timing/data': 0.00592893548309803, 'timing/model': 2.437984334304929, 'learning_rate/qwen_vl_interface': 7.300833173607842e-06, 'learning_rate/action_model': 7.300833173607842e-05, 'epoch': 0.41}) +[2026-06-08 17:07:47,415][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 1.008752703666687, 'train/grad_norm_pre_clip': 4.105093002319336, 'timing/data': 0.006131824105978012, 'timing/model': 2.476142890751362, 'learning_rate/qwen_vl_interface': 7.22971383735263e-06, 'learning_rate/action_model': 7.229713837352629e-05, 'epoch': 0.41}) +[2026-06-08 17:36:03,991][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state +[2026-06-08 17:36:03,997][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 17:36:10,060][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/model.safetensors +[2026-06-08 17:36:18,368][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/optimizer.bin +[2026-06-08 17:36:18,370][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler.bin +[2026-06-08 17:36:18,370][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler_1.bin +[2026-06-08 17:36:18,372][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/random_states_0.pkl +[2026-06-08 17:36:18,390][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 17:36:18,390][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1400 +[2026-06-08 17:36:18,392][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.8162170648574829, 'train/grad_norm_pre_clip': 3.553481101989746, 'eval/loss': 0.8715826272964478, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 5496.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 12.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 5496.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 5496.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 12.0, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 5496.0, 'checkpoint/best_score': 5496.0, 'checkpoint/best_step': 1400.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.005774276331067085, 'timing/model': 2.4613329861313105, 'learning_rate/qwen_vl_interface': 7.158053267101607e-06, 'learning_rate/action_model': 7.158053267101606e-05, 'epoch': 0.41}) +[2026-06-08 17:36:18,392][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1400_state +[2026-06-08 17:36:18,398][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 17:36:23,006][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1400_state/model.safetensors +[2026-06-08 17:36:31,323][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1400_state/optimizer.bin +[2026-06-08 17:36:31,324][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1400_state/sampler.bin +[2026-06-08 17:36:31,325][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1400_state/sampler_1.bin +[2026-06-08 17:36:31,326][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1400_state/random_states_0.pkl +[2026-06-08 17:36:31,458][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-06-08 17:36:31,459][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 17:36:31,474][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 17:42:52,337][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 1.0435634851455688, 'train/grad_norm_pre_clip': 4.6974945068359375, 'timing/data': 0.005695687606930733, 'timing/model': 2.7011631540954113, 'learning_rate/qwen_vl_interface': 7.0858710541391485e-06, 'learning_rate/action_model': 7.085871054139148e-05, 'epoch': 0.42}) +[2026-06-08 17:49:11,119][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.9570038914680481, 'train/grad_norm_pre_clip': 4.6061201095581055, 'timing/data': 0.005844036117196083, 'timing/model': 2.4727474935352802, 'learning_rate/qwen_vl_interface': 7.013186932361549e-06, 'learning_rate/action_model': 7.013186932361549e-05, 'epoch': 0.42}) +[2026-06-08 17:55:30,284][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.7042329907417297, 'train/grad_norm_pre_clip': 3.6245641708374023, 'timing/data': 0.0060750339180231094, 'timing/model': 2.4363371524959803, 'learning_rate/qwen_vl_interface': 6.9400207728819544e-06, 'learning_rate/action_model': 6.940020772881954e-05, 'epoch': 0.42}) +[2026-06-08 18:01:49,528][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 1.2044817209243774, 'train/grad_norm_pre_clip': 4.947127342224121, 'timing/data': 0.005988495424389839, 'timing/model': 2.5699320547282696, 'learning_rate/qwen_vl_interface': 6.866392578597806e-06, 'learning_rate/action_model': 6.866392578597806e-05, 'epoch': 0.43}) +[2026-06-08 18:08:12,834][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.7446308732032776, 'train/grad_norm_pre_clip': 5.667764186859131, 'timing/data': 0.010560935363173485, 'timing/model': 2.3951097913086414, 'learning_rate/qwen_vl_interface': 6.792322478722248e-06, 'learning_rate/action_model': 6.792322478722247e-05, 'epoch': 0.43}) +[2026-06-08 18:14:33,526][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.78305983543396, 'train/grad_norm_pre_clip': 4.060763835906982, 'timing/data': 0.010894391685724258, 'timing/model': 2.470203261822462, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.43}) +[2026-06-08 18:20:54,023][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.5995436906814575, 'train/grad_norm_pre_clip': 9.995461463928223, 'timing/data': 0.010855620726943016, 'timing/model': 2.3928668312728405, 'learning_rate/qwen_vl_interface': 6.642937677576209e-06, 'learning_rate/action_model': 6.642937677576209e-05, 'epoch': 0.44}) +[2026-06-08 18:27:13,422][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.9301248788833618, 'train/grad_norm_pre_clip': 4.335921287536621, 'timing/data': 0.010696271434426308, 'timing/model': 2.46710486151278, 'learning_rate/qwen_vl_interface': 6.5676638166187625e-06, 'learning_rate/action_model': 6.567663816618762e-05, 'epoch': 0.44}) +[2026-06-08 18:33:33,914][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.7987260818481445, 'train/grad_norm_pre_clip': 22.191387176513672, 'timing/data': 0.010481994599103928, 'timing/model': 2.4650435894727707, 'learning_rate/qwen_vl_interface': 6.492029719530645e-06, 'learning_rate/action_model': 6.492029719530645e-05, 'epoch': 0.44}) +[2026-06-08 19:00:35,065][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-08 19:00:35,066][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 1.072854995727539, 'train/grad_norm_pre_clip': 6.152418613433838, 'eval/loss': 0.8813081979751587, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 4806.0, 'rl_games_eval/mid_train/mean_length': 3407.6, 'rl_games_eval/mid_train/std_reward': 461.6752105105926, 'rl_games_eval/mid_train/std_length': 221.3355823178912, 'rl_games_eval/mid_train/macro_mean_reward': 4806.0, 'rl_games_eval/mid_train/macro_mean_length': 3407.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 4806.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 3407.6, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 461.6752105105926, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 221.3355823178912, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 4806.0, 'checkpoint/best_score': 5496.0, 'checkpoint/best_step': 1400.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.010659011080861092, 'timing/model': 2.49168124422431, 'learning_rate/qwen_vl_interface': 6.416056063918798e-06, 'learning_rate/action_model': 6.416056063918797e-05, 'epoch': 0.44}) +[2026-06-08 19:00:35,067][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1500_state +[2026-06-08 19:00:35,071][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 19:00:41,005][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1500_state/model.safetensors +[2026-06-08 19:00:49,346][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1500_state/optimizer.bin +[2026-06-08 19:00:49,347][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1500_state/sampler.bin +[2026-06-08 19:00:49,347][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-08 19:00:49,348][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-08 19:00:49,483][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-08 19:00:49,483][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 19:00:49,498][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 19:07:09,316][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 1.052375316619873, 'train/grad_norm_pre_clip': 8.48251724243164, 'timing/data': 0.010761389508843422, 'timing/model': 2.4706217516213655, 'learning_rate/qwen_vl_interface': 6.339763620222071e-06, 'learning_rate/action_model': 6.339763620222071e-05, 'epoch': 0.45}) +[2026-06-08 19:13:29,919][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.6695024967193604, 'train/grad_norm_pre_clip': 7.108556270599365, 'timing/data': 0.011503882706165314, 'timing/model': 2.45621645078063, 'learning_rate/qwen_vl_interface': 6.263173246032799e-06, 'learning_rate/action_model': 6.263173246032799e-05, 'epoch': 0.45}) +[2026-06-08 19:19:49,036][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.6780129075050354, 'train/grad_norm_pre_clip': 3.150485038757324, 'timing/data': 0.006596546620130539, 'timing/model': 2.3939798194915056, 'learning_rate/qwen_vl_interface': 6.1863058803945315e-06, 'learning_rate/action_model': 6.186305880394532e-05, 'epoch': 0.45}) +[2026-06-08 19:26:09,002][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.873222827911377, 'train/grad_norm_pre_clip': 4.003431797027588, 'timing/data': 0.0007062647491693497, 'timing/model': 2.45702600851655, 'learning_rate/qwen_vl_interface': 6.109182538077524e-06, 'learning_rate/action_model': 6.109182538077523e-05, 'epoch': 0.46}) +[2026-06-08 19:32:32,532][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.9779145121574402, 'train/grad_norm_pre_clip': 7.597695350646973, 'timing/data': 0.0008270610123872757, 'timing/model': 2.48423308134079, 'learning_rate/qwen_vl_interface': 6.0318243038334865e-06, 'learning_rate/action_model': 6.031824303833487e-05, 'epoch': 0.46}) +[2026-06-08 19:38:57,343][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 1.1588259935379028, 'train/grad_norm_pre_clip': 5.784881114959717, 'timing/data': 0.0007184017449617386, 'timing/model': 2.4958517346531153, 'learning_rate/qwen_vl_interface': 5.95425232663125e-06, 'learning_rate/action_model': 5.95425232663125e-05, 'epoch': 0.46}) +[2026-06-08 19:45:20,610][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.9702133536338806, 'train/grad_norm_pre_clip': 4.162489891052246, 'timing/data': 0.0005522575229406357, 'timing/model': 2.762457098811865, 'learning_rate/qwen_vl_interface': 5.876487813874829e-06, 'learning_rate/action_model': 5.876487813874829e-05, 'epoch': 0.47}) +[2026-06-08 19:51:42,529][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.8728047609329224, 'train/grad_norm_pre_clip': 9.385937690734863, 'timing/data': 0.0066744983196258545, 'timing/model': 2.4731709212064743, 'learning_rate/qwen_vl_interface': 5.798552025605536e-06, 'learning_rate/action_model': 5.798552025605536e-05, 'epoch': 0.47}) +[2026-06-08 19:58:03,139][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.9007447361946106, 'train/grad_norm_pre_clip': 5.227511882781982, 'timing/data': 0.005950724706053734, 'timing/model': 2.4006207063794136, 'learning_rate/qwen_vl_interface': 5.7204662686897e-06, 'learning_rate/action_model': 5.720466268689699e-05, 'epoch': 0.47}) +[2026-06-08 20:16:20,521][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1600 +[2026-06-08 20:16:20,523][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.528522789478302, 'train/grad_norm_pre_clip': 2.9918181896209717, 'eval/loss': 0.8004674911499023, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 1608.0, 'rl_games_eval/mid_train/mean_length': 1931.6, 'rl_games_eval/mid_train/std_reward': 867.119368945245, 'rl_games_eval/mid_train/std_length': 473.91881161228446, 'rl_games_eval/mid_train/macro_mean_reward': 1608.0, 'rl_games_eval/mid_train/macro_mean_length': 1931.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 1608.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 1931.6, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 867.119368945245, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 473.91881161228446, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 1608.0, 'checkpoint/best_score': 5496.0, 'checkpoint/best_step': 1400.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.006558174267411232, 'timing/model': 2.554684905335307, 'learning_rate/qwen_vl_interface': 5.64225189099358e-06, 'learning_rate/action_model': 5.6422518909935796e-05, 'epoch': 0.47}) +[2026-06-08 20:16:20,523][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1600_state +[2026-06-08 20:16:20,528][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 20:16:26,416][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1600_state/model.safetensors +[2026-06-08 20:16:34,855][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1600_state/optimizer.bin +[2026-06-08 20:16:34,858][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1600_state/sampler.bin +[2026-06-08 20:16:34,858][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1600_state/sampler_1.bin +[2026-06-08 20:16:34,861][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1600_state/random_states_0.pkl +[2026-06-08 20:16:34,987][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1600 +[2026-06-08 20:16:34,988][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 20:16:35,006][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 20:22:56,341][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.9133950471878052, 'train/grad_norm_pre_clip': 5.875555992126465, 'timing/data': 0.006003024056553841, 'timing/model': 2.400258008390665, 'learning_rate/qwen_vl_interface': 5.563930275547072e-06, 'learning_rate/action_model': 5.5639302755470725e-05, 'epoch': 0.48}) +[2026-06-08 20:29:15,023][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.9572523832321167, 'train/grad_norm_pre_clip': 8.764750480651855, 'timing/data': 0.0062290821224451065, 'timing/model': 2.463423216715455, 'learning_rate/qwen_vl_interface': 5.485522834697806e-06, 'learning_rate/action_model': 5.485522834697806e-05, 'epoch': 0.48}) +[2026-06-08 20:35:33,721][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.9575082063674927, 'train/grad_norm_pre_clip': 4.775829792022705, 'timing/data': 0.006056604906916618, 'timing/model': 2.469166051596403, 'learning_rate/qwen_vl_interface': 5.407051004257214e-06, 'learning_rate/action_model': 5.4070510042572144e-05, 'epoch': 0.48}) +[2026-06-08 20:41:51,915][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.7338576912879944, 'train/grad_norm_pre_clip': 4.37357234954834, 'timing/data': 0.00635630264878273, 'timing/model': 2.4624903108924627, 'learning_rate/qwen_vl_interface': 5.3285362376402035e-06, 'learning_rate/action_model': 5.328536237640204e-05, 'epoch': 0.49}) +[2026-06-08 20:48:09,915][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.8567306995391846, 'train/grad_norm_pre_clip': 4.956024169921875, 'timing/data': 0.005986757576465607, 'timing/model': 2.4625067114830017, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.49}) +[2026-06-08 20:54:32,617][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.7542180418968201, 'train/grad_norm_pre_clip': 5.5670576095581055, 'timing/data': 0.011032219976186752, 'timing/model': 2.4597755074501038, 'learning_rate/qwen_vl_interface': 5.1714637623597976e-06, 'learning_rate/action_model': 5.171463762359797e-05, 'epoch': 0.49}) +[2026-06-08 21:00:51,236][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 1.1606101989746094, 'train/grad_norm_pre_clip': 6.633658409118652, 'timing/data': 0.005835065618157387, 'timing/model': 2.3967254012823105, 'learning_rate/qwen_vl_interface': 5.092948995742786e-06, 'learning_rate/action_model': 5.092948995742785e-05, 'epoch': 0.49}) +[2026-06-08 21:07:09,341][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.7052096128463745, 'train/grad_norm_pre_clip': 4.545790195465088, 'timing/data': 0.005700923502445221, 'timing/model': 2.398073084652424, 'learning_rate/qwen_vl_interface': 5.014477165302194e-06, 'learning_rate/action_model': 5.0144771653021937e-05, 'epoch': 0.5}) +[2026-06-08 21:13:27,440][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.786749541759491, 'train/grad_norm_pre_clip': 6.512526988983154, 'timing/data': 0.006543146446347237, 'timing/model': 2.3988296892493963, 'learning_rate/qwen_vl_interface': 4.936069724452928e-06, 'learning_rate/action_model': 4.936069724452928e-05, 'epoch': 0.5}) +[2026-06-08 21:31:15,149][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1700 +[2026-06-08 21:31:15,150][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.7130063772201538, 'train/grad_norm_pre_clip': 5.7182769775390625, 'eval/loss': 0.7828163504600525, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 2620.0, 'rl_games_eval/mid_train/mean_length': 1893.6, 'rl_games_eval/mid_train/std_reward': 2106.561178793533, 'rl_games_eval/mid_train/std_length': 1301.1709495681187, 'rl_games_eval/mid_train/macro_mean_reward': 2620.0, 'rl_games_eval/mid_train/macro_mean_length': 1893.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 2620.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 1893.6, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 2106.561178793533, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 1301.1709495681187, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 2620.0, 'checkpoint/best_score': 5496.0, 'checkpoint/best_step': 1400.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.006155218929052353, 'timing/model': 2.4007179718464613, 'learning_rate/qwen_vl_interface': 4.857748109006422e-06, 'learning_rate/action_model': 4.857748109006422e-05, 'epoch': 0.5}) +[2026-06-08 21:31:15,150][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1700_state +[2026-06-08 21:31:15,155][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 21:31:20,786][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1700_state/model.safetensors +[2026-06-08 21:31:29,142][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1700_state/optimizer.bin +[2026-06-08 21:31:29,142][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1700_state/sampler.bin +[2026-06-08 21:31:29,143][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1700_state/sampler_1.bin +[2026-06-08 21:31:29,144][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1700_state/random_states_0.pkl +[2026-06-08 21:31:29,278][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1700 +[2026-06-08 21:31:29,279][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 21:31:29,294][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 21:37:47,145][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 1.098874568939209, 'train/grad_norm_pre_clip': 6.796520233154297, 'timing/data': 0.005939474329352379, 'timing/model': 2.4064925108104944, 'learning_rate/qwen_vl_interface': 4.779533731310302e-06, 'learning_rate/action_model': 4.779533731310302e-05, 'epoch': 0.51}) +[2026-06-08 21:44:05,924][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.7092232704162598, 'train/grad_norm_pre_clip': 5.104571342468262, 'timing/data': 0.005855906754732132, 'timing/model': 2.4656847789883614, 'learning_rate/qwen_vl_interface': 4.701447974394466e-06, 'learning_rate/action_model': 4.7014479743944665e-05, 'epoch': 0.51}) +[2026-06-08 21:50:23,935][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.9136937260627747, 'train/grad_norm_pre_clip': 12.339966773986816, 'timing/data': 0.0061491709202528, 'timing/model': 2.3971591852605343, 'learning_rate/qwen_vl_interface': 4.623512186125173e-06, 'learning_rate/action_model': 4.623512186125172e-05, 'epoch': 0.51}) +[2026-06-08 21:56:44,353][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.6721765398979187, 'train/grad_norm_pre_clip': 4.293489933013916, 'timing/data': 0.006685715168714523, 'timing/model': 2.4914204981178045, 'learning_rate/qwen_vl_interface': 4.54574767336875e-06, 'learning_rate/action_model': 4.54574767336875e-05, 'epoch': 0.52}) +[2026-06-08 22:03:02,031][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.7582390308380127, 'train/grad_norm_pre_clip': 6.078740119934082, 'timing/data': 0.006182452663779259, 'timing/model': 2.398981999605894, 'learning_rate/qwen_vl_interface': 4.468175696166513e-06, 'learning_rate/action_model': 4.468175696166513e-05, 'epoch': 0.52}) +[2026-06-08 22:09:20,218][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.791888952255249, 'train/grad_norm_pre_clip': 5.098238945007324, 'timing/data': 0.010744253173470497, 'timing/model': 2.4728929083794355, 'learning_rate/qwen_vl_interface': 4.39081746192248e-06, 'learning_rate/action_model': 4.3908174619224794e-05, 'epoch': 0.52}) +[2026-06-08 22:15:37,828][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.5618426203727722, 'train/grad_norm_pre_clip': 4.282317638397217, 'timing/data': 0.010429136455059052, 'timing/model': 2.3870572820305824, 'learning_rate/qwen_vl_interface': 4.3136941196054696e-06, 'learning_rate/action_model': 4.313694119605469e-05, 'epoch': 0.52}) +[2026-06-08 22:21:54,127][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.8646587133407593, 'train/grad_norm_pre_clip': 6.646336078643799, 'timing/data': 0.010791292414069176, 'timing/model': 2.475179823115468, 'learning_rate/qwen_vl_interface': 4.236826753967203e-06, 'learning_rate/action_model': 4.236826753967203e-05, 'epoch': 0.53}) +[2026-06-08 22:28:11,316][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 1.044505000114441, 'train/grad_norm_pre_clip': 6.462123394012451, 'timing/data': 0.011125016957521439, 'timing/model': 2.4547731317579746, 'learning_rate/qwen_vl_interface': 4.160236379777931e-06, 'learning_rate/action_model': 4.16023637977793e-05, 'epoch': 0.53}) +[2026-06-08 22:53:29,031][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1800 +[2026-06-08 22:53:29,034][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.5241185426712036, 'train/grad_norm_pre_clip': 3.957439661026001, 'eval/loss': 0.7693632245063782, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 3798.0, 'rl_games_eval/mid_train/mean_length': 3143.4, 'rl_games_eval/mid_train/std_reward': 1200.9188149079853, 'rl_games_eval/mid_train/std_length': 576.1682393190378, 'rl_games_eval/mid_train/macro_mean_reward': 3798.0, 'rl_games_eval/mid_train/macro_mean_length': 3143.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 3798.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 3143.4, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 1200.9188149079853, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 576.1682393190378, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 3798.0, 'checkpoint/best_score': 5496.0, 'checkpoint/best_step': 1400.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.01613239198923111, 'timing/model': 2.4692840557545424, 'learning_rate/qwen_vl_interface': 4.083943936081204e-06, 'learning_rate/action_model': 4.083943936081204e-05, 'epoch': 0.53}) +[2026-06-08 22:53:29,035][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1800_state +[2026-06-08 22:53:29,041][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 22:53:34,955][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1800_state/model.safetensors +[2026-06-08 22:53:42,978][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1800_state/optimizer.bin +[2026-06-08 22:53:42,979][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1800_state/sampler.bin +[2026-06-08 22:53:42,979][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1800_state/sampler_1.bin +[2026-06-08 22:53:42,981][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1800_state/random_states_0.pkl +[2026-06-08 22:53:43,116][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1800 +[2026-06-08 22:53:43,117][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 22:53:43,133][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 23:00:03,120][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.9390072822570801, 'train/grad_norm_pre_clip': 5.695897102355957, 'timing/data': 0.01563863269984722, 'timing/model': 2.461304387077689, 'learning_rate/qwen_vl_interface': 4.007970280469356e-06, 'learning_rate/action_model': 4.007970280469356e-05, 'epoch': 0.54}) +[2026-06-08 23:06:22,022][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.8493680953979492, 'train/grad_norm_pre_clip': 6.298111438751221, 'timing/data': 0.022172313183546066, 'timing/model': 2.4476267863065004, 'learning_rate/qwen_vl_interface': 3.9323361833812395e-06, 'learning_rate/action_model': 3.932336183381239e-05, 'epoch': 0.54}) +[2026-06-08 23:12:41,212][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.7798306345939636, 'train/grad_norm_pre_clip': 5.959247589111328, 'timing/data': 0.009968698024749756, 'timing/model': 2.462166301906109, 'learning_rate/qwen_vl_interface': 3.857062322423793e-06, 'learning_rate/action_model': 3.857062322423793e-05, 'epoch': 0.54}) +[2026-06-08 23:19:00,946][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.5702759027481079, 'train/grad_norm_pre_clip': 5.3749518394470215, 'timing/data': 0.006609383970499039, 'timing/model': 2.404436206445098, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 0.55}) +[2026-06-08 23:25:21,818][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.6544257402420044, 'train/grad_norm_pre_clip': 16.18918228149414, 'timing/data': 0.006726324558258057, 'timing/model': 2.4614794719964266, 'learning_rate/qwen_vl_interface': 3.7076775212777537e-06, 'learning_rate/action_model': 3.707677521277754e-05, 'epoch': 0.55}) +[2026-06-08 23:31:41,818][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.7465526461601257, 'train/grad_norm_pre_clip': 32.18019485473633, 'timing/data': 0.006233958527445793, 'timing/model': 2.4609410539269447, 'learning_rate/qwen_vl_interface': 3.6336074214021953e-06, 'learning_rate/action_model': 3.633607421402195e-05, 'epoch': 0.55}) +[2026-06-08 23:38:06,023][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.8127292394638062, 'train/grad_norm_pre_clip': 8.20327377319336, 'timing/data': 0.0005789045244455338, 'timing/model': 2.4644482042640448, 'learning_rate/qwen_vl_interface': 3.5599792271180463e-06, 'learning_rate/action_model': 3.559979227118046e-05, 'epoch': 0.55}) +[2026-06-08 23:44:25,725][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.505500316619873, 'train/grad_norm_pre_clip': 4.479865550994873, 'timing/data': 0.0005591847002506256, 'timing/model': 2.392201481387019, 'learning_rate/qwen_vl_interface': 3.4868130676384505e-06, 'learning_rate/action_model': 3.48681306763845e-05, 'epoch': 0.56}) +[2026-06-08 23:50:44,916][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 1.0153372287750244, 'train/grad_norm_pre_clip': 6.107088565826416, 'timing/data': 0.0005846433341503143, 'timing/model': 2.468984918668866, 'learning_rate/qwen_vl_interface': 3.414128945860853e-06, 'learning_rate/action_model': 3.414128945860853e-05, 'epoch': 0.56}) +[2026-06-09 00:17:22,619][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1900 +[2026-06-09 00:17:22,620][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.655678927898407, 'train/grad_norm_pre_clip': 5.759228706359863, 'eval/loss': 0.7352083921432495, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 4976.0, 'rl_games_eval/mid_train/mean_length': 3329.4, 'rl_games_eval/mid_train/std_reward': 1263.3067719283388, 'rl_games_eval/mid_train/std_length': 369.18320655197743, 'rl_games_eval/mid_train/macro_mean_reward': 4976.0, 'rl_games_eval/mid_train/macro_mean_length': 3329.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 4976.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 3329.4, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 1263.3067719283388, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 369.18320655197743, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 4976.0, 'checkpoint/best_score': 5496.0, 'checkpoint/best_step': 1400.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006577838212251663, 'timing/model': 2.46534619666636, 'learning_rate/qwen_vl_interface': 3.341946732898396e-06, 'learning_rate/action_model': 3.341946732898396e-05, 'epoch': 0.56}) +[2026-06-09 00:17:22,621][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1900_state +[2026-06-09 00:17:22,625][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-09 00:17:28,437][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1900_state/model.safetensors +[2026-06-09 00:17:36,783][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1900_state/optimizer.bin +[2026-06-09 00:17:36,784][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1900_state/sampler.bin +[2026-06-09 00:17:36,784][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1900_state/sampler_1.bin +[2026-06-09 00:17:36,785][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_1900_state/random_states_0.pkl +[2026-06-09 00:17:36,906][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1900 +[2026-06-09 00:17:36,907][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-09 00:17:36,923][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-09 00:23:55,844][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.5264529585838318, 'train/grad_norm_pre_clip': 4.635226249694824, 'timing/data': 0.0005946308374404907, 'timing/model': 2.402101593092084, 'learning_rate/qwen_vl_interface': 3.2702861626473714e-06, 'learning_rate/action_model': 3.2702861626473714e-05, 'epoch': 0.57}) +[2026-06-09 00:30:15,717][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.7736521363258362, 'train/grad_norm_pre_clip': 6.123677730560303, 'timing/data': 0.0006190519779920578, 'timing/model': 2.471772100776434, 'learning_rate/qwen_vl_interface': 3.1991668263921593e-06, 'learning_rate/action_model': 3.1991668263921595e-05, 'epoch': 0.57}) +[2026-06-09 00:36:37,129][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 1.093765377998352, 'train/grad_norm_pre_clip': 6.225223064422607, 'timing/data': 0.0005974750965833664, 'timing/model': 2.392565356567502, 'learning_rate/qwen_vl_interface': 3.1286081674491694e-06, 'learning_rate/action_model': 3.128608167449169e-05, 'epoch': 0.57}) +[2026-06-09 00:42:55,829][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.5632399320602417, 'train/grad_norm_pre_clip': 8.504582405090332, 'timing/data': 0.0006141532212495804, 'timing/model': 2.4664688911288977, 'learning_rate/qwen_vl_interface': 3.0586294758512335e-06, 'learning_rate/action_model': 3.058629475851233e-05, 'epoch': 0.57}) +[2026-06-09 00:49:13,433][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.5913549661636353, 'train/grad_norm_pre_clip': 6.6967997550964355, 'timing/data': 0.000582084059715271, 'timing/model': 2.3932312168180943, 'learning_rate/qwen_vl_interface': 2.989249883073901e-06, 'learning_rate/action_model': 2.9892498830739008e-05, 'epoch': 0.58}) +[2026-06-09 00:55:30,922][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.8385488986968994, 'train/grad_norm_pre_clip': 7.041893005371094, 'timing/data': 0.0005848333239555359, 'timing/model': 2.4686663560569286, 'learning_rate/qwen_vl_interface': 2.920488356805088e-06, 'learning_rate/action_model': 2.9204883568050878e-05, 'epoch': 0.58}) +[2026-06-09 01:01:50,719][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.5191760659217834, 'train/grad_norm_pre_clip': 4.8316850662231445, 'timing/data': 0.0006393417716026306, 'timing/model': 2.4767008684575558, 'learning_rate/qwen_vl_interface': 2.8523636957594935e-06, 'learning_rate/action_model': 2.8523636957594936e-05, 'epoch': 0.58}) +[2026-06-09 01:08:12,246][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.5343713760375977, 'train/grad_norm_pre_clip': 6.782827377319336, 'timing/data': 0.0009045079350471497, 'timing/model': 2.4047962054610252, 'learning_rate/qwen_vl_interface': 2.7848945245392295e-06, 'learning_rate/action_model': 2.7848945245392292e-05, 'epoch': 0.59}) +[2026-06-09 01:14:32,831][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.8761717081069946, 'train/grad_norm_pre_clip': 6.957479953765869, 'timing/data': 0.0005657635629177094, 'timing/model': 2.39332714676857, 'learning_rate/qwen_vl_interface': 2.718099288542031e-06, 'learning_rate/action_model': 2.718099288542031e-05, 'epoch': 0.59}) +[2026-06-09 01:42:24,568][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state +[2026-06-09 01:42:24,574][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-09 01:42:30,251][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/model.safetensors +[2026-06-09 01:42:38,519][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/optimizer.bin +[2026-06-09 01:42:38,519][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler.bin +[2026-06-09 01:42:38,519][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/sampler_1.bin +[2026-06-09 01:42:38,520][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state/random_states_0.pkl +[2026-06-09 01:42:38,536][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-09 01:42:38,537][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2000 +[2026-06-09 01:42:38,538][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.7989387512207031, 'train/grad_norm_pre_clip': 4.844052314758301, 'eval/loss': 0.7062148451805115, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 6030.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 745.0905985180594, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 6030.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_reward': 6030.0, 'rl_games_eval/mid_train/demon_attack__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_2/std_reward': 745.0905985180594, 'rl_games_eval/mid_train/demon_attack__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 6030.0, 'checkpoint/best_score': 6030.0, 'checkpoint/best_step': 2000.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.000535985454916954, 'timing/model': 2.3941799569875, 'learning_rate/qwen_vl_interface': 2.651996248918473e-06, 'learning_rate/action_model': 2.651996248918473e-05, 'epoch': 0.59}) +[2026-06-09 01:42:38,538][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_2000_state +[2026-06-09 01:42:38,543][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-09 01:42:43,086][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_2000_state/model.safetensors +[2026-06-09 01:42:51,344][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_2000_state/optimizer.bin +[2026-06-09 01:42:51,345][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_2000_state/sampler.bin +[2026-06-09 01:42:51,345][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-09 01:42:51,346][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-09 01:42:51,468][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-09 01:42:51,468][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-09 01:42:51,483][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-09 01:42:51,483][starVLA.training.train_starvla][INFO] - Training complete. Best checkpoint step=2000 score=6030.0 path=/workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state +[2026-06-09 01:42:51,483][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_2/checkpoints/best_state +[2026-06-09 01:42:52,502][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-09 01:43:00,954][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-09 01:43:00,954][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-09 01:43:00,955][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-09 01:43:00,966][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-09 01:43:01,936][accelerate.accelerator][INFO] - Loading in 0 custom states diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/summary.jsonl b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..e6d8d889008dec36af7e37c22a4f0aa0263a91f9 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_2_small/summary.jsonl @@ -0,0 +1,20 @@ +{"steps": 100} +{"steps": 200} +{"steps": 300} +{"steps": 400} +{"steps": 500} +{"steps": 600} +{"steps": 700} +{"steps": 800} +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500} +{"steps": 1600} +{"steps": 1700} +{"steps": 1800} +{"steps": 1900} +{"steps": 2000} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/checkpoints/steps_5000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/checkpoints/steps_5000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4a4d113179e50cdceba76c3423042c55bc106ca1 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/checkpoints/steps_5000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cff93451f827160bf275aa8bc0bcab83cbd71645ef0605b2e8869bd3691a3bb6 +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..344d078acb0201afbb2fcbbc7e6eba0a6f64e78e --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/config.full.yaml @@ -0,0 +1,198 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_fix_latency_1_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: null + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_1_200ep +output_dir: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..bea64c0d260b7eb1bdfc206d7e4451c641151b34 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/config.yaml @@ -0,0 +1,71 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep +rl_games: + env_eval: + enabled: true + mid_train: + enabled: false + interval_steps: 100 + task: flappy +run_id: flappy_fix_latency_1_200ep +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_action_classification: true + eval_interval: 100 + eval_num_batches: 100 + freeze_modules: '' + gradient_accumulation_steps: 4 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + pretrained_checkpoint: null + reload_modules: null + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..bdf8b3324aa996a468133627e1c3690eff3690ce --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.6979423761367798, + 0.3020575940608978, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4588429629802704, + 0.4588429629802704, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 297483, + "num_trajectories": 180 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..88b7fb21af76becf6ed0ffdc38ba13902347d226 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8606944680213928, + 0.13930556178092957, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.3463059663772583, + 0.3463059663772583, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 72000, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/hydra/.hydra/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..603dd7a31006ff5ab1a03244ee0b45d979c1092f --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/hydra/.hydra/config.yaml @@ -0,0 +1,196 @@ +framework: + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_fix_latency_1_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + config_name: null + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: null + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_1_200ep +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/hydra/.hydra/hydra.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..dde20157ae69fc7c87f41e1cb6fb731ac372ff14 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/hydra/.hydra/hydra.yaml @@ -0,0 +1,252 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=single + - run_id=flappy_fix_latency_1_200ep + - run_root_dir=/workspace/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=false + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=5000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=100 + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=4 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9, 0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=2 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.shuffle=true + - datasets.vla_data.action_balance.enabled=false + - datasets.vla_data.action_balance.strategy=balanced_epoch + - datasets.vla_data.action_balance.action_key=action_id + - datasets.vla_data.action_balance.target_flap_fraction=0.3 + - datasets.vla_data.action_balance.noop_id=0 + - datasets.vla_data.action_balance.flap_id=1 + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1_200ep + - datasets.vla_data.data_mix=flappy_train__bridge + - datasets.vla_data.eval_data_mix=flappy_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=flappy + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=1 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.mid_train.enabled=false + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[0] + - rl_games.env_eval.post_train.enabled=false + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=false,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_balance.action_key=action_id,datasets.vla_data.action_balance.enabled=false,datasets.vla_data.action_balance.flap_id=1,datasets.vla_data.action_balance.noop_id=0,datasets.vla_data.action_balance.strategy=balanced_epoch,datasets.vla_data.action_balance.target_flap_fraction=0.3,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=flappy_train__bridge,datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1_200ep,datasets.vla_data.eval_data_mix=flappy_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.shuffle=true,datasets.vla_data.video_backend=torchvision_av,env=flappy,framework.action_model.action_dim=7,framework.action_model.action_env_dim=2,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=1,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=false,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[0],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=false,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=flappy,run_id=flappy_fix_latency_1_200ep,run_root_dir=/workspace/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=100,trainer.gradient_accumulation_steps=4,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=5000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9, + 0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.3 + version_base: '1.1' + cwd: /workspace/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/hydra/.hydra/overrides.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..3b93d3643bb6359d2e2f5636d3eddb055c2315b3 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/hydra/.hydra/overrides.yaml @@ -0,0 +1,91 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=single +- run_id=flappy_fix_latency_1_200ep +- run_root_dir=/workspace/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=false +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=5000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=100 +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=4 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9, 0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=2 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.shuffle=true +- datasets.vla_data.action_balance.enabled=false +- datasets.vla_data.action_balance.strategy=balanced_epoch +- datasets.vla_data.action_balance.action_key=action_id +- datasets.vla_data.action_balance.target_flap_fraction=0.3 +- datasets.vla_data.action_balance.noop_id=0 +- datasets.vla_data.action_balance.flap_id=1 +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1_200ep +- datasets.vla_data.data_mix=flappy_train__bridge +- datasets.vla_data.eval_data_mix=flappy_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=flappy +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=1 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.mid_train.enabled=false +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[0] +- rl_games.env_eval.post_train.enabled=false +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/hydra/train_starvla_hydra.log b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..42856b208af5bfa06c587a1876fd0d0acc21f1ec --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/hydra/train_starvla_hydra.log @@ -0,0 +1,5552 @@ +[2026-06-15 17:11:23,389][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-15 17:11:23,401][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-15 17:11:32,046][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-15 17:11:32,083][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-15 17:11:32,744][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-15 17:11:32,744][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-15 17:11:32,744][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-15 17:11:32,744][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-15 17:11:32,744][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-15 17:11:32,744][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-15 17:11:32,744][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-15 17:11:32,744][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-15 17:11:32,744][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-15 17:11:32,744][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-15 17:11:32,745][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-15 17:11:32,745][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-15 17:11:32,745][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-15 17:11:32,745][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-15 17:11:32,745][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-15 17:11:32,745][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-15 17:11:32,745][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-15 17:11:32,745][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-15 17:11:32,745][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-15 17:11:32,745][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-15 17:11:32,745][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-15 17:11:32,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-15 17:11:32,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-15 17:11:32,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-15 17:11:32,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-15 17:11:32,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-15 17:11:32,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-15 17:11:32,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-15 17:11:32,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-15 17:11:32,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-15 17:11:32,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-15 17:11:32,782][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-15 17:11:32,811][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/config.full.yaml +[2026-06-15 17:11:32,817][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/config.yaml +[2026-06-15 17:11:32,818][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-15 17:11:58,942][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-15 17:11:58,943][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-06-15 17:11:58,943][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-15 17:11:58,943][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 4 +[2026-06-15 17:11:58,943][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-15 17:12:24,477][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 0.6951658725738525, 'train/grad_norm_pre_clip': 4.08105993270874, 'timing/data': 0.0206312210066244, 'timing/model': 1.2377813389757648, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-15 17:12:29,236][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 0.6962764263153076, 'train/grad_norm_pre_clip': 7.312942028045654, 'timing/data': 0.019866556976921856, 'timing/model': 1.2480296040303074, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-15 17:12:34,037][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 0.6920099258422852, 'train/grad_norm_pre_clip': 2.2723686695098877, 'timing/data': 0.022392341983504593, 'timing/model': 1.2465286389924586, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-06-15 17:12:38,837][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 0.677629292011261, 'train/grad_norm_pre_clip': 5.142302513122559, 'timing/data': 0.02086617797613144, 'timing/model': 1.2496429519960657, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-06-15 17:12:43,637][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 0.6822172403335571, 'train/grad_norm_pre_clip': 1.4956973791122437, 'timing/data': 0.011849095986690372, 'timing/model': 1.2472688530106097, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-06-15 17:12:48,535][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 0.6534649729728699, 'train/grad_norm_pre_clip': 3.1045994758605957, 'timing/data': 0.01252083299914375, 'timing/model': 1.2575192509684712, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-06-15 17:12:53,341][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 0.6756784319877625, 'train/grad_norm_pre_clip': 0.25623956322669983, 'timing/data': 0.012579309986904263, 'timing/model': 1.2576215640292503, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-06-15 17:12:58,244][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.5778442621231079, 'train/grad_norm_pre_clip': 3.811169147491455, 'timing/data': 0.012610857025720179, 'timing/model': 1.2532831460121088, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-06-15 17:13:03,142][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 0.5998625159263611, 'train/grad_norm_pre_clip': 1.259131908416748, 'timing/data': 0.011707007011864334, 'timing/model': 1.2585246760281734, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-06-15 17:13:08,047][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.5949894189834595, 'train/grad_norm_pre_clip': 0.22332780063152313, 'timing/data': 0.015090633009094745, 'timing/model': 1.2529266870114952, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-15 17:13:13,119][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.7373162508010864, 'train/grad_norm_pre_clip': 4.0895233154296875, 'timing/data': 0.007627286948263645, 'timing/model': 1.4298648170079105, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-06-15 17:13:17,949][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.6417015194892883, 'train/grad_norm_pre_clip': 2.3291893005371094, 'timing/data': 0.007910253014415503, 'timing/model': 1.2584896530024707, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.01}) +[2026-06-15 17:13:22,848][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.6945158243179321, 'train/grad_norm_pre_clip': 3.524616241455078, 'timing/data': 0.007933276996482164, 'timing/model': 1.2562269170302898, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.01}) +[2026-06-15 17:13:27,749][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.6123600602149963, 'train/grad_norm_pre_clip': 2.0367414951324463, 'timing/data': 0.007749437005259097, 'timing/model': 1.2644494429696351, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.01}) +[2026-06-15 17:13:32,648][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.6109489798545837, 'train/grad_norm_pre_clip': 1.979401707649231, 'timing/data': 0.007575201045256108, 'timing/model': 1.2509594759903848, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.01}) +[2026-06-15 17:13:37,548][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.5662912130355835, 'train/grad_norm_pre_clip': 0.8786501884460449, 'timing/data': 0.007446872012224048, 'timing/model': 1.2599335659760982, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 17:13:42,448][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.6351573467254639, 'train/grad_norm_pre_clip': 1.9793405532836914, 'timing/data': 0.005687955010216683, 'timing/model': 1.257187980983872, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 17:13:47,351][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.6582469940185547, 'train/grad_norm_pre_clip': 1.9978207349777222, 'timing/data': 0.007682597963139415, 'timing/model': 1.2626618230133317, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.01}) +[2026-06-15 17:13:52,249][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.5697274804115295, 'train/grad_norm_pre_clip': 1.481632947921753, 'timing/data': 0.007092660991474986, 'timing/model': 1.2627233659732156, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.01}) +[2026-06-15 17:13:57,150][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.620878279209137, 'train/grad_norm_pre_clip': 0.6497197151184082, 'timing/data': 0.008083018008619547, 'timing/model': 1.2555469739600085, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-15 17:14:02,050][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.6423026323318481, 'train/grad_norm_pre_clip': 0.36552292108535767, 'timing/data': 0.0076694689923897386, 'timing/model': 1.2630936150089838, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.01}) +[2026-06-15 17:14:06,948][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.5439112782478333, 'train/grad_norm_pre_clip': 5.528648853302002, 'timing/data': 0.007709468016400933, 'timing/model': 1.2631973499665037, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 17:14:11,848][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.6422163844108582, 'train/grad_norm_pre_clip': 0.7119241952896118, 'timing/data': 0.005999143002554774, 'timing/model': 1.2639523330144584, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 17:14:16,749][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.6239731311798096, 'train/grad_norm_pre_clip': 1.1275532245635986, 'timing/data': 0.006089325994253159, 'timing/model': 1.2599301100126468, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.01}) +[2026-06-15 17:14:21,651][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.5298264622688293, 'train/grad_norm_pre_clip': 3.6327288150787354, 'timing/data': 0.00753376999637112, 'timing/model': 1.2561565609648824, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.01}) +[2026-06-15 17:14:26,553][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.6207600235939026, 'train/grad_norm_pre_clip': 0.8086363077163696, 'timing/data': 0.0056092089507728815, 'timing/model': 1.255430165969301, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.01}) +[2026-06-15 17:14:31,450][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.4008227586746216, 'train/grad_norm_pre_clip': 4.35903263092041, 'timing/data': 0.007657793001271784, 'timing/model': 1.2603907230077311, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.01}) +[2026-06-15 17:14:36,347][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.7498447299003601, 'train/grad_norm_pre_clip': 4.719263076782227, 'timing/data': 0.0077724510338157415, 'timing/model': 1.25116134097334, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 17:14:41,245][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.7688010931015015, 'train/grad_norm_pre_clip': 5.133347034454346, 'timing/data': 0.007628318970091641, 'timing/model': 1.24973725999007, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.01}) +[2026-06-15 17:14:46,145][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.6063246130943298, 'train/grad_norm_pre_clip': 1.990403652191162, 'timing/data': 0.007636901980731636, 'timing/model': 1.2564420730341226, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-15 17:14:51,050][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.634807288646698, 'train/grad_norm_pre_clip': 2.441073417663574, 'timing/data': 0.0076803750125691295, 'timing/model': 1.2562280799611472, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.01}) +[2026-06-15 17:14:55,945][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.6206977963447571, 'train/grad_norm_pre_clip': 1.7327450513839722, 'timing/data': 0.007502935011871159, 'timing/model': 1.2441659609903581, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.01}) +[2026-06-15 17:15:00,875][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.6426975727081299, 'train/grad_norm_pre_clip': 1.3748831748962402, 'timing/data': 0.005674054962582886, 'timing/model': 1.173486074025277, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.01}) +[2026-06-15 17:15:05,377][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.66142737865448, 'train/grad_norm_pre_clip': 1.0493576526641846, 'timing/data': 0.007648900034837425, 'timing/model': 1.1596011350047775, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.01}) +[2026-06-15 17:15:10,250][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.6026657819747925, 'train/grad_norm_pre_clip': 7.361914157867432, 'timing/data': 0.007724019989836961, 'timing/model': 1.2613355289795436, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.02}) +[2026-06-15 17:15:15,155][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.6810007095336914, 'train/grad_norm_pre_clip': 2.950390577316284, 'timing/data': 0.005789974995423108, 'timing/model': 1.2626880600000732, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.02}) +[2026-06-15 17:15:20,043][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.6818114519119263, 'train/grad_norm_pre_clip': 9.14316463470459, 'timing/data': 0.007664021977689117, 'timing/model': 1.2716609290218912, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.02}) +[2026-06-15 17:15:25,050][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.6530902981758118, 'train/grad_norm_pre_clip': 8.485137939453125, 'timing/data': 0.00561896403087303, 'timing/model': 1.2435674429871142, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.02}) +[2026-06-15 17:15:29,952][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.7050058841705322, 'train/grad_norm_pre_clip': 2.909726619720459, 'timing/data': 0.005964491982012987, 'timing/model': 1.2554849529988132, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 17:15:34,853][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.6296271085739136, 'train/grad_norm_pre_clip': 2.8057379722595215, 'timing/data': 0.0060243799816817045, 'timing/model': 1.2555089099914767, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.02}) +[2026-06-15 17:15:39,756][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.6461753249168396, 'train/grad_norm_pre_clip': 0.7184538841247559, 'timing/data': 0.007709408993832767, 'timing/model': 1.2622064920142293, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.02}) +[2026-06-15 17:15:44,652][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.6449429392814636, 'train/grad_norm_pre_clip': 1.2030322551727295, 'timing/data': 0.007604884973261505, 'timing/model': 1.256529844016768, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.02}) +[2026-06-15 17:15:49,552][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.553926944732666, 'train/grad_norm_pre_clip': 1.9469562768936157, 'timing/data': 0.007694776984862983, 'timing/model': 1.2561623259680346, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.02}) +[2026-06-15 17:15:54,450][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.5893182158470154, 'train/grad_norm_pre_clip': 2.7702136039733887, 'timing/data': 0.007496375008486211, 'timing/model': 1.2633178610121831, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 17:15:59,356][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.4463704228401184, 'train/grad_norm_pre_clip': 2.240161180496216, 'timing/data': 0.005964792973827571, 'timing/model': 1.2649852460017428, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.02}) +[2026-06-15 17:16:04,246][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.5046498775482178, 'train/grad_norm_pre_clip': 4.4680891036987305, 'timing/data': 0.007559448014944792, 'timing/model': 1.2472589070093818, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.02}) +[2026-06-15 17:16:09,150][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.37726977467536926, 'train/grad_norm_pre_clip': 2.6144354343414307, 'timing/data': 0.0077576799667440355, 'timing/model': 1.2554871200118214, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.02}) +[2026-06-15 17:16:14,050][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.6605536937713623, 'train/grad_norm_pre_clip': 11.4744873046875, 'timing/data': 0.007887009996920824, 'timing/model': 1.2528817399870604, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.02}) +[2026-06-15 17:16:18,951][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.5679894685745239, 'train/grad_norm_pre_clip': 9.4162015914917, 'timing/data': 0.007377620961051434, 'timing/model': 1.2594025220023468, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.02}) +[2026-06-15 17:16:23,856][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.6216248869895935, 'train/grad_norm_pre_clip': 5.887674331665039, 'timing/data': 0.007295018993318081, 'timing/model': 1.2628595780115575, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.02}) +[2026-06-15 17:16:28,686][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.4769483506679535, 'train/grad_norm_pre_clip': 9.241266250610352, 'timing/data': 0.0076093110255897045, 'timing/model': 1.1867549400194548, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 17:16:33,454][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.62701815366745, 'train/grad_norm_pre_clip': 4.134958744049072, 'timing/data': 0.007764019013848156, 'timing/model': 1.261627278989181, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.02}) +[2026-06-15 17:16:38,348][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.5650136470794678, 'train/grad_norm_pre_clip': 6.652028560638428, 'timing/data': 0.007715518004260957, 'timing/model': 1.2544092149473727, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.02}) +[2026-06-15 17:16:43,256][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.5363719463348389, 'train/grad_norm_pre_clip': 6.156933307647705, 'timing/data': 0.0074085459928028286, 'timing/model': 1.258310162986163, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.02}) +[2026-06-15 17:16:48,150][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.5268742442131042, 'train/grad_norm_pre_clip': 13.73447036743164, 'timing/data': 0.007599637028761208, 'timing/model': 1.2549398179980926, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.02}) +[2026-06-15 17:16:53,059][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.8434702157974243, 'train/grad_norm_pre_clip': 36.28550720214844, 'timing/data': 0.007570764981210232, 'timing/model': 1.2716217549750581, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 17:16:57,956][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.665726363658905, 'train/grad_norm_pre_clip': 21.095502853393555, 'timing/data': 0.00757134601008147, 'timing/model': 1.2563462409889325, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.02}) +[2026-06-15 17:17:02,856][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.5093168020248413, 'train/grad_norm_pre_clip': 25.61714744567871, 'timing/data': 0.005939995986409485, 'timing/model': 1.2633475189795718, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.02}) +[2026-06-15 17:17:07,756][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.5321093797683716, 'train/grad_norm_pre_clip': 14.095230102539062, 'timing/data': 0.005833539995364845, 'timing/model': 1.2584332230035216, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.03}) +[2026-06-15 17:17:12,655][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.5350135564804077, 'train/grad_norm_pre_clip': 12.231548309326172, 'timing/data': 0.00581804703688249, 'timing/model': 1.2571433980483562, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.03}) +[2026-06-15 17:17:17,555][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.5511114597320557, 'train/grad_norm_pre_clip': 15.186236381530762, 'timing/data': 0.007400313974358141, 'timing/model': 1.2560765700181946, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.03}) +[2026-06-15 17:17:22,450][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.34807032346725464, 'train/grad_norm_pre_clip': 13.976232528686523, 'timing/data': 0.007480703003238887, 'timing/model': 1.2521851080236956, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.03}) +[2026-06-15 17:17:27,353][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.4221065044403076, 'train/grad_norm_pre_clip': 5.043248653411865, 'timing/data': 0.007334235997404903, 'timing/model': 1.2545197730069049, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.03}) +[2026-06-15 17:17:32,406][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.28886377811431885, 'train/grad_norm_pre_clip': 12.117141723632812, 'timing/data': 0.007422887021675706, 'timing/model': 1.4087044339976273, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.03}) +[2026-06-15 17:17:37,249][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.5222750306129456, 'train/grad_norm_pre_clip': 13.106603622436523, 'timing/data': 0.007446391973644495, 'timing/model': 1.2505615950212814, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.03}) +[2026-06-15 17:17:42,152][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.6974601149559021, 'train/grad_norm_pre_clip': 22.030471801757812, 'timing/data': 0.00790489703649655, 'timing/model': 1.2544800900504924, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.03}) +[2026-06-15 17:17:47,058][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.5790579915046692, 'train/grad_norm_pre_clip': 11.075520515441895, 'timing/data': 0.005772780976258218, 'timing/model': 1.2600197980063967, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.03}) +[2026-06-15 17:17:51,955][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.5068415403366089, 'train/grad_norm_pre_clip': 6.8147053718566895, 'timing/data': 0.005823175015393645, 'timing/model': 1.2582276939647272, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.03}) +[2026-06-15 17:17:56,856][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.5463736653327942, 'train/grad_norm_pre_clip': 13.472175598144531, 'timing/data': 0.005882792000193149, 'timing/model': 1.2565385609632358, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.03}) +[2026-06-15 17:18:01,749][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.589496374130249, 'train/grad_norm_pre_clip': 14.350689888000488, 'timing/data': 0.007571775990072638, 'timing/model': 1.2489322749897838, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.03}) +[2026-06-15 17:18:06,660][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.5808697938919067, 'train/grad_norm_pre_clip': 16.37445068359375, 'timing/data': 0.007688137993682176, 'timing/model': 1.2679105100105517, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.03}) +[2026-06-15 17:18:11,557][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.46173083782196045, 'train/grad_norm_pre_clip': 13.321398735046387, 'timing/data': 0.0077030190150253475, 'timing/model': 1.2542633889825083, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.03}) +[2026-06-15 17:18:16,459][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.5671804547309875, 'train/grad_norm_pre_clip': 14.455787658691406, 'timing/data': 0.005936731991823763, 'timing/model': 1.258738161996007, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.03}) +[2026-06-15 17:18:21,358][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.6608521938323975, 'train/grad_norm_pre_clip': 27.276235580444336, 'timing/data': 0.007839610043447465, 'timing/model': 1.2540323020075448, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.03}) +[2026-06-15 17:18:26,258][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.8883955478668213, 'train/grad_norm_pre_clip': 54.90785217285156, 'timing/data': 0.007246227003633976, 'timing/model': 1.2552478010184132, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.03}) +[2026-06-15 17:18:31,158][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 0.6532884836196899, 'train/grad_norm_pre_clip': 30.164745330810547, 'timing/data': 0.005827962013427168, 'timing/model': 1.2589949539978988, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.03}) +[2026-06-15 17:18:36,058][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.6102802753448486, 'train/grad_norm_pre_clip': 23.19230842590332, 'timing/data': 0.007926136953756213, 'timing/model': 1.2568110930151306, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.03}) +[2026-06-15 17:18:40,955][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.47927913069725037, 'train/grad_norm_pre_clip': 21.6875057220459, 'timing/data': 0.007150706951506436, 'timing/model': 1.260660847008694, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.03}) +[2026-06-15 17:18:45,855][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.455230176448822, 'train/grad_norm_pre_clip': 13.074633598327637, 'timing/data': 0.005970261001493782, 'timing/model': 1.2610009179916233, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.03}) +[2026-06-15 17:18:50,757][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.48010435700416565, 'train/grad_norm_pre_clip': 20.803415298461914, 'timing/data': 0.006017731036990881, 'timing/model': 1.2610123860067688, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.03}) +[2026-06-15 17:18:55,657][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.4718311131000519, 'train/grad_norm_pre_clip': 13.879246711730957, 'timing/data': 0.007731321034952998, 'timing/model': 1.2536843300331384, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.03}) +[2026-06-15 17:19:00,550][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.3804079294204712, 'train/grad_norm_pre_clip': 9.353201866149902, 'timing/data': 0.0074691150221042335, 'timing/model': 1.2564420760027133, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.04}) +[2026-06-15 17:19:05,450][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.3792285621166229, 'train/grad_norm_pre_clip': 8.825414657592773, 'timing/data': 0.00732622598297894, 'timing/model': 1.2548382289824076, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.04}) +[2026-06-15 17:19:10,372][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.42425042390823364, 'train/grad_norm_pre_clip': 10.587881088256836, 'timing/data': 0.0005204859771765769, 'timing/model': 1.2758967210538685, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.04}) +[2026-06-15 17:19:15,154][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.357700377702713, 'train/grad_norm_pre_clip': 6.110969066619873, 'timing/data': 0.000572777003981173, 'timing/model': 1.2614260019618087, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.04}) +[2026-06-15 17:19:20,059][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.44369885325431824, 'train/grad_norm_pre_clip': 4.48214054107666, 'timing/data': 0.0005766720278188586, 'timing/model': 1.2653934319969267, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.04}) +[2026-06-15 17:19:24,956][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.3638988435268402, 'train/grad_norm_pre_clip': 12.612748146057129, 'timing/data': 0.0005082109710201621, 'timing/model': 1.26455184398219, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.04}) +[2026-06-15 17:19:29,859][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.3933295011520386, 'train/grad_norm_pre_clip': 9.149380683898926, 'timing/data': 0.0005163229652680457, 'timing/model': 1.2653027799678966, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.04}) +[2026-06-15 17:19:34,752][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.48850753903388977, 'train/grad_norm_pre_clip': 9.065019607543945, 'timing/data': 0.0005837520002387464, 'timing/model': 1.264186738000717, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.04}) +[2026-06-15 17:19:39,658][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.5328279137611389, 'train/grad_norm_pre_clip': 7.820563793182373, 'timing/data': 0.0005894610076211393, 'timing/model': 1.264039127039723, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.04}) +[2026-06-15 17:19:44,554][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.5058289170265198, 'train/grad_norm_pre_clip': 15.353269577026367, 'timing/data': 0.0005833119503222406, 'timing/model': 1.2438661089981906, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.04}) +[2026-06-15 17:19:49,455][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.2933756709098816, 'train/grad_norm_pre_clip': 3.8636491298675537, 'timing/data': 0.000518636021297425, 'timing/model': 1.2632201150408946, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.04}) +[2026-06-15 17:19:54,350][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.4024823307991028, 'train/grad_norm_pre_clip': 10.706547737121582, 'timing/data': 0.0005197179852984846, 'timing/model': 1.2570616059820168, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.04}) +[2026-06-15 17:19:59,261][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.3681020140647888, 'train/grad_norm_pre_clip': 7.712454795837402, 'timing/data': 0.0004974050098098814, 'timing/model': 1.2653225800022483, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.04}) +[2026-06-15 17:20:04,157][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.41213446855545044, 'train/grad_norm_pre_clip': 12.043767929077148, 'timing/data': 0.0005247950321063399, 'timing/model': 1.2616342440014705, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.04}) +[2026-06-15 17:20:09,058][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.5246142148971558, 'train/grad_norm_pre_clip': 16.752059936523438, 'timing/data': 0.0005172049859538674, 'timing/model': 1.2605491830036044, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.04}) +[2026-06-15 17:20:13,961][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.6760421395301819, 'train/grad_norm_pre_clip': 17.5834903717041, 'timing/data': 0.000507569988258183, 'timing/model': 1.2691014499869198, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.04}) +[2026-06-15 17:20:18,850][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.4741785228252411, 'train/grad_norm_pre_clip': 18.27480697631836, 'timing/data': 0.0005080109694972634, 'timing/model': 1.2561500719748437, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.04}) +[2026-06-15 17:20:23,760][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.7407244443893433, 'train/grad_norm_pre_clip': 29.94851303100586, 'timing/data': 0.0005042540142312646, 'timing/model': 1.2639352249680087, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.04}) +[2026-06-15 17:22:53,338][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp4avj56in/test.c -o /tmp/tmp4avj56in/test.o +[2026-06-15 17:22:53,366][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp4avj56in/test.o -laio -o /tmp/tmp4avj56in/a.out +[2026-06-15 17:22:53,796][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpodokai7l/test.c -o /tmp/tmpodokai7l/test.o +[2026-06-15 17:22:53,822][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpodokai7l/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpodokai7l/a.out +[2026-06-15 17:22:53,877][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp3ceczbdk/test.c -o /tmp/tmp3ceczbdk/test.o +[2026-06-15 17:22:53,894][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp3ceczbdk/test.o -laio -o /tmp/tmp3ceczbdk/a.out +[2026-06-15 17:30:05,953][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.4387386739253998, 'train/grad_norm_pre_clip': 13.60667896270752, 'eval/loss': 0.20527732372283936, 'eval/flap_precision': 0.7665847665847666, 'eval/flap_recall': 0.625250501002004, 'eval/flap_f1': 0.6887417218543046, 'eval/flap_rate_model': 0.11305555555555556, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.6887417218543046, 'timing/data': 0.0005044459830969572, 'timing/model': 1.2646956630051136, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.04}) +[2026-06-15 17:30:05,953][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_100_state +[2026-06-15 17:30:05,957][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 17:30:11,395][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_100_state/model.safetensors +[2026-06-15 17:30:19,406][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_100_state/optimizer.bin +[2026-06-15 17:30:19,406][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_100_state/sampler.bin +[2026-06-15 17:30:19,407][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_100_state/sampler_1.bin +[2026-06-15 17:30:19,408][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-15 17:30:19,410][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-15 17:30:19,410][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 17:30:19,419][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 17:30:25,429][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.35171231627464294, 'train/grad_norm_pre_clip': 7.162493705749512, 'timing/data': 0.0005163309979252517, 'timing/model': 1.4855329240090214, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 0.04}) +[2026-06-15 17:30:31,409][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.4234701991081238, 'train/grad_norm_pre_clip': 8.564221382141113, 'timing/data': 0.000518293003551662, 'timing/model': 1.5136401370400563, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 0.04}) +[2026-06-15 17:30:37,270][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.5176264047622681, 'train/grad_norm_pre_clip': 11.203487396240234, 'timing/data': 0.0005269570392556489, 'timing/model': 1.6937688420293853, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 0.04}) +[2026-06-15 17:30:42,942][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.4288291931152344, 'train/grad_norm_pre_clip': 13.027825355529785, 'timing/data': 0.0005190650117583573, 'timing/model': 1.4546432019560598, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 0.04}) +[2026-06-15 17:30:48,845][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.32663586735725403, 'train/grad_norm_pre_clip': 8.500844955444336, 'timing/data': 0.0005191449890844524, 'timing/model': 1.4646266829804517, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 0.05}) +[2026-06-15 17:30:54,747][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.38048484921455383, 'train/grad_norm_pre_clip': 13.302903175354004, 'timing/data': 0.0005244720377959311, 'timing/model': 1.4595212049898691, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 0.05}) +[2026-06-15 17:31:00,715][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.2640857398509979, 'train/grad_norm_pre_clip': 15.082655906677246, 'timing/data': 0.0005246030050329864, 'timing/model': 1.5014435379998758, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 0.05}) +[2026-06-15 17:31:06,675][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.41320979595184326, 'train/grad_norm_pre_clip': 32.933692932128906, 'timing/data': 0.000523531052749604, 'timing/model': 1.6930183219956234, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 0.05}) +[2026-06-15 17:31:12,451][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.3863512873649597, 'train/grad_norm_pre_clip': 11.166521072387695, 'timing/data': 0.0005209679948166013, 'timing/model': 1.5679126169998199, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 0.05}) +[2026-06-15 17:31:18,148][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.47862714529037476, 'train/grad_norm_pre_clip': 19.336322784423828, 'timing/data': 0.0005448629963211715, 'timing/model': 1.4636049970285967, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.05}) +[2026-06-15 17:31:24,127][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.365055114030838, 'train/grad_norm_pre_clip': 10.432880401611328, 'timing/data': 0.000499946007039398, 'timing/model': 1.5413149209925905, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 0.05}) +[2026-06-15 17:31:30,151][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.34604379534721375, 'train/grad_norm_pre_clip': 8.444631576538086, 'timing/data': 0.0005281689809635282, 'timing/model': 1.4828141030156985, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 0.05}) +[2026-06-15 17:31:36,449][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.2560995817184448, 'train/grad_norm_pre_clip': 9.34968376159668, 'timing/data': 0.0005243230261839926, 'timing/model': 1.5025278070243075, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 0.05}) +[2026-06-15 17:31:42,377][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.216714009642601, 'train/grad_norm_pre_clip': 7.028524875640869, 'timing/data': 0.0005166910123080015, 'timing/model': 1.6859171140240505, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 0.05}) +[2026-06-15 17:31:48,365][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.39367738366127014, 'train/grad_norm_pre_clip': 23.820175170898438, 'timing/data': 0.0005250630201771855, 'timing/model': 1.563361150037963, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 0.05}) +[2026-06-15 17:31:54,053][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.5783948302268982, 'train/grad_norm_pre_clip': 42.19896697998047, 'timing/data': 0.0005330750136636198, 'timing/model': 1.464767609955743, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 0.05}) +[2026-06-15 17:31:59,946][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.17569877207279205, 'train/grad_norm_pre_clip': 14.820755004882812, 'timing/data': 0.0005206270143389702, 'timing/model': 1.456243649998214, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 0.05}) +[2026-06-15 17:32:05,850][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.3972640335559845, 'train/grad_norm_pre_clip': 11.628570556640625, 'timing/data': 0.000527898024301976, 'timing/model': 1.4513807030161843, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 0.05}) +[2026-06-15 17:32:11,828][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.4616866111755371, 'train/grad_norm_pre_clip': 21.14017677307129, 'timing/data': 0.0005334860179573298, 'timing/model': 1.5004876360180788, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 0.05}) +[2026-06-15 17:32:17,860][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.5277553200721741, 'train/grad_norm_pre_clip': 11.362005233764648, 'timing/data': 0.0005305319791659713, 'timing/model': 1.77381250198232, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.05}) +[2026-06-15 17:32:23,551][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.27288907766342163, 'train/grad_norm_pre_clip': 6.632918357849121, 'timing/data': 0.0005016889772377908, 'timing/model': 1.4630826709908433, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 0.05}) +[2026-06-15 17:32:29,453][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.2575198709964752, 'train/grad_norm_pre_clip': 7.271175861358643, 'timing/data': 0.000529951008502394, 'timing/model': 1.4639077139436267, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 0.05}) +[2026-06-15 17:32:35,353][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.2575788199901581, 'train/grad_norm_pre_clip': 9.846747398376465, 'timing/data': 0.0005528749898076057, 'timing/model': 1.463427304988727, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 0.05}) +[2026-06-15 17:32:41,253][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.40108567476272583, 'train/grad_norm_pre_clip': 11.637887954711914, 'timing/data': 0.0005358390044420958, 'timing/model': 1.4615205739974044, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 0.05}) +[2026-06-15 17:32:47,156][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.3302546441555023, 'train/grad_norm_pre_clip': 14.841827392578125, 'timing/data': 0.0005263659986667335, 'timing/model': 1.4335374959628098, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 0.05}) +[2026-06-15 17:32:53,083][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.2606334388256073, 'train/grad_norm_pre_clip': 13.825631141662598, 'timing/data': 0.0005280280020087957, 'timing/model': 1.6939834909862839, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 0.05}) +[2026-06-15 17:32:58,818][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.4998006820678711, 'train/grad_norm_pre_clip': 7.342131614685059, 'timing/data': 0.0005350279971025884, 'timing/model': 1.5285843799938448, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 0.05}) +[2026-06-15 17:33:04,753][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.44483399391174316, 'train/grad_norm_pre_clip': 9.312597274780273, 'timing/data': 0.0005220689927227795, 'timing/model': 1.4635628179530613, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 0.06}) +[2026-06-15 17:33:10,655][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.41205018758773804, 'train/grad_norm_pre_clip': 15.140864372253418, 'timing/data': 0.0005310720298439264, 'timing/model': 1.4600017930497415, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 0.06}) +[2026-06-15 17:33:16,840][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.3532228469848633, 'train/grad_norm_pre_clip': 3.5160980224609375, 'timing/data': 0.000582897977437824, 'timing/model': 1.5535710760159418, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.06}) +[2026-06-15 17:33:22,762][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.4030255973339081, 'train/grad_norm_pre_clip': 12.430974960327148, 'timing/data': 0.0005293589783832431, 'timing/model': 1.4397344500175677, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 0.06}) +[2026-06-15 17:33:28,758][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.2996552586555481, 'train/grad_norm_pre_clip': 9.859963417053223, 'timing/data': 0.0005173119716346264, 'timing/model': 1.7674511640216224, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 0.06}) +[2026-06-15 17:33:34,601][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.41981932520866394, 'train/grad_norm_pre_clip': 25.915008544921875, 'timing/data': 0.000557231018319726, 'timing/model': 1.6103723509586416, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 0.06}) +[2026-06-15 17:33:40,256][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.3463205099105835, 'train/grad_norm_pre_clip': 25.63372802734375, 'timing/data': 0.0005266360240057111, 'timing/model': 1.4620515719871037, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 0.06}) +[2026-06-15 17:33:46,157][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.3238694369792938, 'train/grad_norm_pre_clip': 13.679970741271973, 'timing/data': 0.0005311929853633046, 'timing/model': 1.4649329669773579, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.06}) +[2026-06-15 17:33:52,160][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.30470266938209534, 'train/grad_norm_pre_clip': 9.104803085327148, 'timing/data': 0.0005428490112535655, 'timing/model': 1.454831239010673, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 0.06}) +[2026-06-15 17:33:58,059][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.23589123785495758, 'train/grad_norm_pre_clip': 13.422826766967773, 'timing/data': 0.0005224300548434258, 'timing/model': 1.4351074530277401, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 0.06}) +[2026-06-15 17:34:03,980][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.25785738229751587, 'train/grad_norm_pre_clip': 10.999791145324707, 'timing/data': 0.0005362600204534829, 'timing/model': 1.6866555039887317, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 0.06}) +[2026-06-15 17:34:09,892][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.1715146005153656, 'train/grad_norm_pre_clip': 18.802858352661133, 'timing/data': 0.0005293100257404149, 'timing/model': 1.6923857619985938, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 0.06}) +[2026-06-15 17:34:15,557][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 1.2768523693084717, 'train/grad_norm_pre_clip': 77.22891998291016, 'timing/data': 0.0005232699913904071, 'timing/model': 1.4601061749854125, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.06}) +[2026-06-15 17:34:21,455][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.6672025322914124, 'train/grad_norm_pre_clip': 44.88473892211914, 'timing/data': 0.0005365110118873417, 'timing/model': 1.4622839949443005, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 0.06}) +[2026-06-15 17:34:27,457][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 1.0620081424713135, 'train/grad_norm_pre_clip': 34.95909118652344, 'timing/data': 0.000539214990567416, 'timing/model': 1.4859612659784034, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 0.06}) +[2026-06-15 17:34:33,357][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.9552986025810242, 'train/grad_norm_pre_clip': 31.778940200805664, 'timing/data': 0.000524642993696034, 'timing/model': 1.4358047029818408, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 0.06}) +[2026-06-15 17:34:39,283][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 1.3068485260009766, 'train/grad_norm_pre_clip': 41.436893463134766, 'timing/data': 0.0005230910028330982, 'timing/model': 1.6887360549881123, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 0.06}) +[2026-06-15 17:34:45,134][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.6134541034698486, 'train/grad_norm_pre_clip': 22.46466064453125, 'timing/data': 0.000525434035807848, 'timing/model': 1.6423512960318476, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 0.06}) +[2026-06-15 17:34:50,856][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.47972047328948975, 'train/grad_norm_pre_clip': 12.656312942504883, 'timing/data': 0.0005344070377759635, 'timing/model': 1.463990671967622, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 0.06}) +[2026-06-15 17:34:56,756][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.5617272257804871, 'train/grad_norm_pre_clip': 18.000822067260742, 'timing/data': 0.0005383120151236653, 'timing/model': 1.46338959003333, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 0.06}) +[2026-06-15 17:35:02,756][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.4467156231403351, 'train/grad_norm_pre_clip': 10.391773223876953, 'timing/data': 0.00053469801787287, 'timing/model': 1.46184916497441, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 0.06}) +[2026-06-15 17:35:08,656][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.5696952939033508, 'train/grad_norm_pre_clip': 11.54337215423584, 'timing/data': 0.0005370809813030064, 'timing/model': 1.438276043976657, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.06}) +[2026-06-15 17:35:14,555][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.365139365196228, 'train/grad_norm_pre_clip': 10.348468780517578, 'timing/data': 0.0005307120154611766, 'timing/model': 1.4319533070083708, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.06}) +[2026-06-15 17:35:20,501][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.6107245087623596, 'train/grad_norm_pre_clip': 21.980722427368164, 'timing/data': 0.0005309819825924933, 'timing/model': 1.7099805559846573, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 0.06}) +[2026-06-15 17:35:26,230][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.5959978699684143, 'train/grad_norm_pre_clip': 16.486770629882812, 'timing/data': 0.0005244530038908124, 'timing/model': 1.5385836799978279, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 0.07}) +[2026-06-15 17:35:32,159][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.3075938820838928, 'train/grad_norm_pre_clip': 4.878628253936768, 'timing/data': 0.0005371709703467786, 'timing/model': 1.4685967109980993, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 0.07}) +[2026-06-15 17:35:38,155][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.4539995789527893, 'train/grad_norm_pre_clip': 4.8216071128845215, 'timing/data': 0.0005319339688867331, 'timing/model': 1.4788534499821253, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 0.07}) +[2026-06-15 17:35:44,060][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.31299707293510437, 'train/grad_norm_pre_clip': 9.400148391723633, 'timing/data': 0.0005250330432318151, 'timing/model': 1.435012554982677, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 0.07}) +[2026-06-15 17:35:49,984][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.4126104712486267, 'train/grad_norm_pre_clip': 7.146132469177246, 'timing/data': 0.0005385130061767995, 'timing/model': 1.691248502000235, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 0.07}) +[2026-06-15 17:35:55,659][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.5334564447402954, 'train/grad_norm_pre_clip': 10.534945487976074, 'timing/data': 0.0005242819897830486, 'timing/model': 1.4680173080414534, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 0.07}) +[2026-06-15 17:36:01,657][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.2539101243019104, 'train/grad_norm_pre_clip': 4.046664237976074, 'timing/data': 0.0005280369659885764, 'timing/model': 1.4641087370109744, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 0.07}) +[2026-06-15 17:36:07,457][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.24251113831996918, 'train/grad_norm_pre_clip': 3.3537395000457764, 'timing/data': 0.0005243420018814504, 'timing/model': 1.4655487629934214, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 0.07}) +[2026-06-15 17:36:13,356][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.1328323632478714, 'train/grad_norm_pre_clip': 5.350983619689941, 'timing/data': 0.0005440309760160744, 'timing/model': 1.444039395020809, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.07}) +[2026-06-15 17:36:19,316][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.9000504612922668, 'train/grad_norm_pre_clip': 42.43674850463867, 'timing/data': 0.0005259949830360711, 'timing/model': 1.5456567640067078, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 0.07}) +[2026-06-15 17:36:25,287][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.3982720673084259, 'train/grad_norm_pre_clip': 13.992897987365723, 'timing/data': 0.000522890011779964, 'timing/model': 1.6956956200301647, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 0.07}) +[2026-06-15 17:36:31,053][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.34699705243110657, 'train/grad_norm_pre_clip': 23.78548240661621, 'timing/data': 0.0005334160523489118, 'timing/model': 1.55814591498347, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 0.07}) +[2026-06-15 17:36:36,956][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.3443668484687805, 'train/grad_norm_pre_clip': 11.628033638000488, 'timing/data': 0.0005262549966573715, 'timing/model': 1.4634502129629254, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 0.07}) +[2026-06-15 17:36:43,020][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.3866678476333618, 'train/grad_norm_pre_clip': 5.3297119140625, 'timing/data': 0.0005303900106810033, 'timing/model': 1.5552669760072604, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 0.07}) +[2026-06-15 17:36:49,056][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.42991673946380615, 'train/grad_norm_pre_clip': 12.003682136535645, 'timing/data': 0.000528198026586324, 'timing/model': 1.4839189240010455, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 0.07}) +[2026-06-15 17:36:54,957][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.6153129935264587, 'train/grad_norm_pre_clip': 30.044334411621094, 'timing/data': 0.0005334760062396526, 'timing/model': 1.4638362600235268, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 0.07}) +[2026-06-15 17:37:00,891][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.5788055062294006, 'train/grad_norm_pre_clip': 20.44801902770996, 'timing/data': 0.0005273359711281955, 'timing/model': 1.6988643820513971, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 0.07}) +[2026-06-15 17:37:06,555][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.5781440138816833, 'train/grad_norm_pre_clip': 30.169588088989258, 'timing/data': 0.0005286579835228622, 'timing/model': 1.463580056966748, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 0.07}) +[2026-06-15 17:37:12,454][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.390794962644577, 'train/grad_norm_pre_clip': 20.108448028564453, 'timing/data': 0.0005193339893594384, 'timing/model': 1.4632571479887702, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.07}) +[2026-06-15 17:37:18,455][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.2388269007205963, 'train/grad_norm_pre_clip': 6.726767539978027, 'timing/data': 0.0005231709801591933, 'timing/model': 1.4820249690092169, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 0.07}) +[2026-06-15 17:37:24,354][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.8172577023506165, 'train/grad_norm_pre_clip': 18.24703598022461, 'timing/data': 0.0005218680016696453, 'timing/model': 1.4640004070242867, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 0.07}) +[2026-06-15 17:37:30,332][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.3665038049221039, 'train/grad_norm_pre_clip': 12.881196022033691, 'timing/data': 0.0005187529604882002, 'timing/model': 1.5026589750195853, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 0.07}) +[2026-06-15 17:37:36,289][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.3726038634777069, 'train/grad_norm_pre_clip': 8.367372512817383, 'timing/data': 0.000510932004544884, 'timing/model': 1.6985908510396257, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 0.07}) +[2026-06-15 17:37:41,959][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 1.0182054042816162, 'train/grad_norm_pre_clip': 23.192228317260742, 'timing/data': 0.0005408459692262113, 'timing/model': 1.4682616459904239, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 0.08}) +[2026-06-15 17:37:47,858][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.32801496982574463, 'train/grad_norm_pre_clip': 3.212665557861328, 'timing/data': 0.0006118100136518478, 'timing/model': 1.4664131609606557, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 0.08}) +[2026-06-15 17:37:53,758][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.40914908051490784, 'train/grad_norm_pre_clip': 9.373019218444824, 'timing/data': 0.0005271059926599264, 'timing/model': 1.4658205069717951, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 0.08}) +[2026-06-15 17:37:59,657][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.3926566541194916, 'train/grad_norm_pre_clip': 5.598566055297852, 'timing/data': 0.0005965490126982331, 'timing/model': 1.4646394459996372, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 0.08}) +[2026-06-15 17:38:05,628][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.5620604157447815, 'train/grad_norm_pre_clip': 8.958230018615723, 'timing/data': 0.0005272859707474709, 'timing/model': 1.5002618129947223, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 0.08}) +[2026-06-15 17:38:11,586][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.36681804060935974, 'train/grad_norm_pre_clip': 12.825590133666992, 'timing/data': 0.0005254040006548166, 'timing/model': 1.6942293769679964, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.08}) +[2026-06-15 17:38:17,256][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.25476449728012085, 'train/grad_norm_pre_clip': 13.177349090576172, 'timing/data': 0.0005226299981586635, 'timing/model': 1.4624680650304072, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 0.08}) +[2026-06-15 17:38:23,152][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.493782103061676, 'train/grad_norm_pre_clip': 23.30182647705078, 'timing/data': 0.0005255939904600382, 'timing/model': 1.4600743200280704, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 0.08}) +[2026-06-15 17:38:29,154][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.4035727083683014, 'train/grad_norm_pre_clip': 14.256827354431152, 'timing/data': 0.0005300799966789782, 'timing/model': 1.4622548799961805, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 0.08}) +[2026-06-15 17:38:35,055][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.21444982290267944, 'train/grad_norm_pre_clip': 6.431421279907227, 'timing/data': 0.000522670045029372, 'timing/model': 1.4623898470308632, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 0.08}) +[2026-06-15 17:38:40,959][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.3671781122684479, 'train/grad_norm_pre_clip': 14.084024429321289, 'timing/data': 0.0005316020105965436, 'timing/model': 1.4372113249846734, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 0.08}) +[2026-06-15 17:38:46,885][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.3864145874977112, 'train/grad_norm_pre_clip': 6.9688568115234375, 'timing/data': 0.0005307710380293429, 'timing/model': 1.6920437830267474, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 0.08}) +[2026-06-15 17:38:52,556][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.19996017217636108, 'train/grad_norm_pre_clip': 11.69876480102539, 'timing/data': 0.0005190550000406802, 'timing/model': 1.4653501129942015, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 0.08}) +[2026-06-15 17:38:58,455][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.3557039797306061, 'train/grad_norm_pre_clip': 17.53162956237793, 'timing/data': 0.0005304209771566093, 'timing/model': 1.461614252999425, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 0.08}) +[2026-06-15 17:39:04,352][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.37613150477409363, 'train/grad_norm_pre_clip': 14.631318092346191, 'timing/data': 0.0005212980322539806, 'timing/model': 1.448519472964108, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 0.08}) +[2026-06-15 17:39:10,352][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.43887338042259216, 'train/grad_norm_pre_clip': 16.413101196289062, 'timing/data': 0.0005136359832249582, 'timing/model': 1.4725868359673768, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.08}) +[2026-06-15 17:39:16,351][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.2372962385416031, 'train/grad_norm_pre_clip': 9.30416202545166, 'timing/data': 0.00054209801601246, 'timing/model': 1.490247828012798, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 0.08}) +[2026-06-15 17:39:22,391][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.5130438804626465, 'train/grad_norm_pre_clip': 10.594612121582031, 'timing/data': 0.0005217579891905189, 'timing/model': 1.7223228699876927, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 0.08}) +[2026-06-15 17:39:28,201][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.17437992990016937, 'train/grad_norm_pre_clip': 7.888798236846924, 'timing/data': 0.0005273770075291395, 'timing/model': 1.61047045502346, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 0.08}) +[2026-06-15 17:39:33,855][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.2679053246974945, 'train/grad_norm_pre_clip': 10.704357147216797, 'timing/data': 0.0005217279540374875, 'timing/model': 1.463208189001307, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 0.08}) +[2026-06-15 17:39:39,757][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.32355788350105286, 'train/grad_norm_pre_clip': 10.528194427490234, 'timing/data': 0.0005197850405238569, 'timing/model': 1.4655040599755011, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 0.08}) +[2026-06-15 17:39:45,756][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.3041543960571289, 'train/grad_norm_pre_clip': 12.571370124816895, 'timing/data': 0.0005255540017969906, 'timing/model': 1.480414931953419, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 0.08}) +[2026-06-15 17:39:51,735][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.3671739101409912, 'train/grad_norm_pre_clip': 7.006789684295654, 'timing/data': 0.0005289990222081542, 'timing/model': 1.5073053810046986, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 0.08}) +[2026-06-15 17:39:57,677][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.3610836863517761, 'train/grad_norm_pre_clip': 6.600467205047607, 'timing/data': 0.0005198750295676291, 'timing/model': 1.6847664120141417, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.09}) +[2026-06-15 17:40:03,585][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.42150434851646423, 'train/grad_norm_pre_clip': 13.742530822753906, 'timing/data': 0.0006589600234292448, 'timing/model': 1.6958367789629847, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 0.09}) +[2026-06-15 17:49:17,244][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.2498415857553482, 'train/grad_norm_pre_clip': 6.763615608215332, 'eval/loss': 0.16297538578510284, 'eval/flap_precision': 0.697151424287856, 'eval/flap_recall': 0.9318637274549099, 'eval/flap_f1': 0.7975986277873072, 'eval/flap_rate_model': 0.1852777777777778, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.7975986277873073, 'timing/data': 0.000528267992194742, 'timing/model': 1.4610128219937906, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.09}) +[2026-06-15 17:49:17,245][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_200_state +[2026-06-15 17:49:17,260][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 17:49:22,592][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_200_state/model.safetensors +[2026-06-15 17:49:29,634][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_200_state/optimizer.bin +[2026-06-15 17:49:29,635][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_200_state/sampler.bin +[2026-06-15 17:49:29,635][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_200_state/sampler_1.bin +[2026-06-15 17:49:29,636][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-15 17:49:29,889][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-15 17:49:29,889][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 17:49:29,899][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 17:49:35,539][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.23591259121894836, 'train/grad_norm_pre_clip': 6.015803337097168, 'timing/data': 0.0008009790326468647, 'timing/model': 1.4664103320101276, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 0.09}) +[2026-06-15 17:49:41,536][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.23832257091999054, 'train/grad_norm_pre_clip': 8.794732093811035, 'timing/data': 0.0005926230223849416, 'timing/model': 1.4621021480415948, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 0.09}) +[2026-06-15 17:49:47,632][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.4726751446723938, 'train/grad_norm_pre_clip': 15.78302001953125, 'timing/data': 0.0005496389931067824, 'timing/model': 1.5578820980153978, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 0.09}) +[2026-06-15 17:49:53,551][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.571739912033081, 'train/grad_norm_pre_clip': 8.802082061767578, 'timing/data': 0.0005438799853436649, 'timing/model': 1.4744939000229351, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 0.09}) +[2026-06-15 17:49:59,445][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.2747284770011902, 'train/grad_norm_pre_clip': 12.401674270629883, 'timing/data': 0.000545342976693064, 'timing/model': 1.4289082430186681, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.09}) +[2026-06-15 17:50:05,376][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.18066830933094025, 'train/grad_norm_pre_clip': 5.711649417877197, 'timing/data': 0.0005332150030881166, 'timing/model': 1.6869633339811116, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 0.09}) +[2026-06-15 17:50:11,044][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.43177056312561035, 'train/grad_norm_pre_clip': 10.9130859375, 'timing/data': 0.0005391230224631727, 'timing/model': 1.4641950340010226, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 0.09}) +[2026-06-15 17:50:17,050][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.4601197838783264, 'train/grad_norm_pre_clip': 9.908650398254395, 'timing/data': 0.0005441100220195949, 'timing/model': 1.4455421270104125, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 0.09}) +[2026-06-15 17:50:22,850][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.47829821705818176, 'train/grad_norm_pre_clip': 20.909297943115234, 'timing/data': 0.00052763702115044, 'timing/model': 1.4655941679957323, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 0.09}) +[2026-06-15 17:50:28,747][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.323748916387558, 'train/grad_norm_pre_clip': 13.563909530639648, 'timing/data': 0.000529990007635206, 'timing/model': 1.4658123459666967, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.09}) +[2026-06-15 17:50:34,649][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.4998129904270172, 'train/grad_norm_pre_clip': 13.50572395324707, 'timing/data': 0.0005264150095172226, 'timing/model': 1.4807150379638188, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 0.09}) +[2026-06-15 17:50:40,592][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.18961676955223083, 'train/grad_norm_pre_clip': 3.808475971221924, 'timing/data': 0.0005206159548833966, 'timing/model': 1.7047783650341444, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 0.09}) +[2026-06-15 17:50:46,248][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.31575849652290344, 'train/grad_norm_pre_clip': 8.122234344482422, 'timing/data': 0.0005246829823590815, 'timing/model': 1.4639497570460662, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 0.09}) +[2026-06-15 17:50:52,353][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.22434142231941223, 'train/grad_norm_pre_clip': 10.653972625732422, 'timing/data': 0.0005185029585845768, 'timing/model': 1.474510197993368, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 0.09}) +[2026-06-15 17:50:58,350][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.3031803071498871, 'train/grad_norm_pre_clip': 8.297098159790039, 'timing/data': 0.0005272759590297937, 'timing/model': 1.4689376929891296, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 0.09}) +[2026-06-15 17:51:04,253][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.19485503435134888, 'train/grad_norm_pre_clip': 6.992532253265381, 'timing/data': 0.0005261339829303324, 'timing/model': 1.4625385119579732, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 0.09}) +[2026-06-15 17:51:10,149][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.1515757143497467, 'train/grad_norm_pre_clip': 9.297269821166992, 'timing/data': 0.0005307309911586344, 'timing/model': 1.4638064559549093, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 0.09}) +[2026-06-15 17:51:16,083][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.396111398935318, 'train/grad_norm_pre_clip': 18.78374481201172, 'timing/data': 0.0005165199982002378, 'timing/model': 1.6906038660090417, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 0.09}) +[2026-06-15 17:51:21,750][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.20944757759571075, 'train/grad_norm_pre_clip': 3.3772640228271484, 'timing/data': 0.0005376309854909778, 'timing/model': 1.4657354400260374, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 0.09}) +[2026-06-15 17:51:27,757][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.49923717975616455, 'train/grad_norm_pre_clip': 31.4157657623291, 'timing/data': 0.0005278769531287253, 'timing/model': 1.4672933209803887, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.09}) +[2026-06-15 17:51:33,649][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.37645354866981506, 'train/grad_norm_pre_clip': 13.852182388305664, 'timing/data': 0.0005352479638531804, 'timing/model': 1.4404647719929926, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 0.1}) +[2026-06-15 17:51:39,554][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.35802754759788513, 'train/grad_norm_pre_clip': 6.943861961364746, 'timing/data': 0.0005192540120333433, 'timing/model': 1.4628862690296955, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 0.1}) +[2026-06-15 17:51:45,456][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.3034248352050781, 'train/grad_norm_pre_clip': 11.733972549438477, 'timing/data': 0.000522548973094672, 'timing/model': 1.4358865389949642, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 0.1}) +[2026-06-15 17:51:51,479][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.36635705828666687, 'train/grad_norm_pre_clip': 9.351752281188965, 'timing/data': 0.0005258439923636615, 'timing/model': 1.6896794379572384, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 0.1}) +[2026-06-15 17:51:57,156][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.3041277229785919, 'train/grad_norm_pre_clip': 13.12665843963623, 'timing/data': 0.0005214880220592022, 'timing/model': 1.463908727047965, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 0.1}) +[2026-06-15 17:52:03,056][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.23159459233283997, 'train/grad_norm_pre_clip': 18.073335647583008, 'timing/data': 0.0005251929978840053, 'timing/model': 1.465640373993665, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 0.1}) +[2026-06-15 17:52:08,952][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.3149440884590149, 'train/grad_norm_pre_clip': 14.301188468933105, 'timing/data': 0.0005277470336295664, 'timing/model': 1.4513853110256605, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 0.1}) +[2026-06-15 17:52:14,856][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.44561320543289185, 'train/grad_norm_pre_clip': 10.640317916870117, 'timing/data': 0.0005356680485419929, 'timing/model': 1.4535347819910385, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 0.1}) +[2026-06-15 17:52:20,839][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.37838536500930786, 'train/grad_norm_pre_clip': 8.913191795349121, 'timing/data': 0.000537051004357636, 'timing/model': 1.502507857978344, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 0.1}) +[2026-06-15 17:52:26,780][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.3007740378379822, 'train/grad_norm_pre_clip': 7.865523338317871, 'timing/data': 0.0005197050049901009, 'timing/model': 1.6918795619858429, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.1}) +[2026-06-15 17:52:32,451][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.30173274874687195, 'train/grad_norm_pre_clip': 3.9786641597747803, 'timing/data': 0.0005001559620723128, 'timing/model': 1.462532759993337, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 0.1}) +[2026-06-15 17:52:38,452][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.3432613015174866, 'train/grad_norm_pre_clip': 5.5042500495910645, 'timing/data': 0.0005126850446686149, 'timing/model': 1.4666073939879425, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 0.1}) +[2026-06-15 17:52:44,352][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.4337058961391449, 'train/grad_norm_pre_clip': 4.388476848602295, 'timing/data': 0.0005233900155872107, 'timing/model': 1.4640625969623215, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 0.1}) +[2026-06-15 17:52:50,251][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.2558327317237854, 'train/grad_norm_pre_clip': 7.641529560089111, 'timing/data': 0.0005312510184012353, 'timing/model': 1.4642832889803685, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 0.1}) +[2026-06-15 17:52:56,253][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.25984686613082886, 'train/grad_norm_pre_clip': 8.272327423095703, 'timing/data': 0.0005284969811327755, 'timing/model': 1.4916614210233092, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 0.1}) +[2026-06-15 17:53:02,183][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.2758599519729614, 'train/grad_norm_pre_clip': 4.83419942855835, 'timing/data': 0.0005277669988572598, 'timing/model': 1.6921148429973982, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 0.1}) +[2026-06-15 17:53:07,855][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.5005112290382385, 'train/grad_norm_pre_clip': 15.112062454223633, 'timing/data': 0.000524681992828846, 'timing/model': 1.4620961649925448, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 0.1}) +[2026-06-15 17:53:13,855][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.3713752329349518, 'train/grad_norm_pre_clip': 6.429142951965332, 'timing/data': 0.0005225390195846558, 'timing/model': 1.4623982469784096, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 0.1}) +[2026-06-15 17:53:19,755][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.17647838592529297, 'train/grad_norm_pre_clip': 4.738633632659912, 'timing/data': 0.0005300600314512849, 'timing/model': 1.463462902000174, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 0.1}) +[2026-06-15 17:53:25,751][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.29118871688842773, 'train/grad_norm_pre_clip': 6.877262592315674, 'timing/data': 0.0005261590122245252, 'timing/model': 1.4497289460268803, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.1}) +[2026-06-15 17:53:31,755][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.33982977271080017, 'train/grad_norm_pre_clip': 11.10471248626709, 'timing/data': 0.000522963993716985, 'timing/model': 1.506788222992327, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 0.1}) +[2026-06-15 17:53:37,682][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.07578681409358978, 'train/grad_norm_pre_clip': 3.8441450595855713, 'timing/data': 0.0005161550361663103, 'timing/model': 1.6873219539993443, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 0.1}) +[2026-06-15 17:53:43,501][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.17159508168697357, 'train/grad_norm_pre_clip': 5.293442249298096, 'timing/data': 0.0005210509989410639, 'timing/model': 1.610201767005492, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 0.1}) +[2026-06-15 17:53:49,157][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.15384133160114288, 'train/grad_norm_pre_clip': 6.127810001373291, 'timing/data': 0.000528491975273937, 'timing/model': 1.4619996349792928, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 0.1}) +[2026-06-15 17:53:55,053][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.3188580274581909, 'train/grad_norm_pre_clip': 12.002764701843262, 'timing/data': 0.0005251569673418999, 'timing/model': 1.4591180300340056, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 0.11}) +[2026-06-15 17:54:01,057][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.20643846690654755, 'train/grad_norm_pre_clip': 14.642441749572754, 'timing/data': 0.0005214930279180408, 'timing/model': 1.4656905069714412, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 0.11}) +[2026-06-15 17:54:06,955][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.2973684072494507, 'train/grad_norm_pre_clip': 10.068460464477539, 'timing/data': 0.0004955439944751561, 'timing/model': 1.4352660960284993, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.11}) +[2026-06-15 17:54:12,878][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.2884235978126526, 'train/grad_norm_pre_clip': 12.535284042358398, 'timing/data': 0.0005613910034298897, 'timing/model': 1.6865854649804533, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 0.11}) +[2026-06-15 17:54:18,785][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.32346636056900024, 'train/grad_norm_pre_clip': 11.800985336303711, 'timing/data': 0.000534771999809891, 'timing/model': 1.6920788119896315, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 0.11}) +[2026-06-15 17:54:24,456][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.6105082035064697, 'train/grad_norm_pre_clip': 7.898355960845947, 'timing/data': 0.0005279019824229181, 'timing/model': 1.4648309540352784, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.11}) +[2026-06-15 17:54:30,357][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.2320222705602646, 'train/grad_norm_pre_clip': 12.16608715057373, 'timing/data': 0.0005181070300750434, 'timing/model': 1.4670956129557453, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 0.11}) +[2026-06-15 17:54:36,351][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.31288403272628784, 'train/grad_norm_pre_clip': 7.494228363037109, 'timing/data': 0.000523224996868521, 'timing/model': 1.4815607370110229, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 0.11}) +[2026-06-15 17:54:42,258][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.20638804137706757, 'train/grad_norm_pre_clip': 11.62428092956543, 'timing/data': 0.0004939420032314956, 'timing/model': 1.432730763044674, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 0.11}) +[2026-06-15 17:54:48,187][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.2476262003183365, 'train/grad_norm_pre_clip': 6.270979881286621, 'timing/data': 0.000495164014864713, 'timing/model': 1.6932793899904937, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 0.11}) +[2026-06-15 17:54:53,855][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.27038395404815674, 'train/grad_norm_pre_clip': 3.403218984603882, 'timing/data': 0.0004935520119033754, 'timing/model': 1.46408861404052, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 0.11}) +[2026-06-15 17:54:59,854][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.4678939878940582, 'train/grad_norm_pre_clip': 5.017664909362793, 'timing/data': 0.0005050189793109894, 'timing/model': 1.4508313559927046, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 0.11}) +[2026-06-15 17:55:05,856][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.3845134973526001, 'train/grad_norm_pre_clip': 11.45323657989502, 'timing/data': 0.000497969042044133, 'timing/model': 1.4575219939579256, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 0.11}) +[2026-06-15 17:55:11,836][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.29741576313972473, 'train/grad_norm_pre_clip': 4.719603061676025, 'timing/data': 0.0005033159977756441, 'timing/model': 1.5053349999943748, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 0.11}) +[2026-06-15 17:55:17,747][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.3837515413761139, 'train/grad_norm_pre_clip': 4.177221298217773, 'timing/data': 0.000509105040691793, 'timing/model': 1.5049235700280406, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 0.11}) +[2026-06-15 17:55:23,640][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.22765208780765533, 'train/grad_norm_pre_clip': 4.544206619262695, 'timing/data': 0.0005311769782565534, 'timing/model': 1.645581824996043, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.11}) +[2026-06-15 17:55:29,355][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.19729585945606232, 'train/grad_norm_pre_clip': 3.2316081523895264, 'timing/data': 0.0005271219997666776, 'timing/model': 1.4647053370135836, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 0.11}) +[2026-06-15 17:55:35,354][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.18319320678710938, 'train/grad_norm_pre_clip': 6.185336589813232, 'timing/data': 0.0005024849670007825, 'timing/model': 1.462360000004992, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 0.11}) +[2026-06-15 17:55:41,357][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.23367789387702942, 'train/grad_norm_pre_clip': 5.025805473327637, 'timing/data': 0.0005285629886202514, 'timing/model': 1.4682053470169194, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 0.11}) +[2026-06-15 17:55:47,258][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.3115123212337494, 'train/grad_norm_pre_clip': 6.5892415046691895, 'timing/data': 0.0005301659693941474, 'timing/model': 1.439690797997173, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 0.11}) +[2026-06-15 17:55:53,155][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.1839659959077835, 'train/grad_norm_pre_clip': 5.424461841583252, 'timing/data': 0.0004977890057489276, 'timing/model': 1.4351025890209712, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 0.11}) +[2026-06-15 17:55:59,040][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.26846832036972046, 'train/grad_norm_pre_clip': 7.751742839813232, 'timing/data': 0.0005095659871585667, 'timing/model': 1.650613389967475, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 0.11}) +[2026-06-15 17:56:05,056][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.3111496865749359, 'train/grad_norm_pre_clip': 14.377303123474121, 'timing/data': 0.0005748510011471808, 'timing/model': 1.450912619009614, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 0.11}) +[2026-06-15 17:56:11,056][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.17744384706020355, 'train/grad_norm_pre_clip': 9.814691543579102, 'timing/data': 0.0005746509996242821, 'timing/model': 1.4658178010140546, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 0.12}) +[2026-06-15 17:56:17,054][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.37007609009742737, 'train/grad_norm_pre_clip': 7.649496555328369, 'timing/data': 0.0005569850327447057, 'timing/model': 1.45347262598807, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 0.12}) +[2026-06-15 17:56:22,958][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.33392778038978577, 'train/grad_norm_pre_clip': 10.14686393737793, 'timing/data': 0.0005647559883072972, 'timing/model': 1.433647152967751, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.12}) +[2026-06-15 17:56:28,887][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.053039345890283585, 'train/grad_norm_pre_clip': 3.429781198501587, 'timing/data': 0.0005654380074702203, 'timing/model': 1.6978959160041995, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 0.12}) +[2026-06-15 17:56:34,556][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.2822284996509552, 'train/grad_norm_pre_clip': 16.080190658569336, 'timing/data': 0.0004976980271749198, 'timing/model': 1.4642680779797956, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 0.12}) +[2026-06-15 17:56:40,456][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.23837365210056305, 'train/grad_norm_pre_clip': 6.228557109832764, 'timing/data': 0.0004956760094501078, 'timing/model': 1.4362734490423463, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 0.12}) +[2026-06-15 17:56:46,357][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.1743210405111313, 'train/grad_norm_pre_clip': 3.6124677658081055, 'timing/data': 0.0005682620103470981, 'timing/model': 1.4610290099517442, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 0.12}) +[2026-06-15 17:56:52,255][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.19553494453430176, 'train/grad_norm_pre_clip': 11.304529190063477, 'timing/data': 0.0004920699866488576, 'timing/model': 1.4661122980178334, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.12}) +[2026-06-15 17:56:58,156][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.3457222580909729, 'train/grad_norm_pre_clip': 11.504671096801758, 'timing/data': 0.0005706250085495412, 'timing/model': 1.4869418639573269, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 0.12}) +[2026-06-15 17:57:04,088][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.4149683117866516, 'train/grad_norm_pre_clip': 14.34716510772705, 'timing/data': 0.0005202119937166572, 'timing/model': 1.694850251020398, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 0.12}) +[2026-06-15 17:57:09,757][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.24822616577148438, 'train/grad_norm_pre_clip': 7.964492321014404, 'timing/data': 0.0005238570156507194, 'timing/model': 1.4674283519852906, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 0.12}) +[2026-06-15 17:57:15,659][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.20689502358436584, 'train/grad_norm_pre_clip': 10.208166122436523, 'timing/data': 0.0005180189618840814, 'timing/model': 1.466783138050232, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 0.12}) +[2026-06-15 17:57:21,556][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.22801446914672852, 'train/grad_norm_pre_clip': 4.372491359710693, 'timing/data': 0.0005061720148660243, 'timing/model': 1.4591543650021777, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.12}) +[2026-06-15 17:57:27,553][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.34701478481292725, 'train/grad_norm_pre_clip': 9.45402717590332, 'timing/data': 0.0005004230188205838, 'timing/model': 1.4521954319789074, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 0.12}) +[2026-06-15 17:57:33,454][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.2967982590198517, 'train/grad_norm_pre_clip': 10.332290649414062, 'timing/data': 0.000497529050335288, 'timing/model': 1.4676751839579083, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 0.12}) +[2026-06-15 17:57:39,382][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.07671204209327698, 'train/grad_norm_pre_clip': 2.5596072673797607, 'timing/data': 0.0004978890065103769, 'timing/model': 1.691666303027887, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 0.12}) +[2026-06-15 17:57:45,054][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.13427674770355225, 'train/grad_norm_pre_clip': 6.7159104347229, 'timing/data': 0.0004978890065103769, 'timing/model': 1.4616432840120979, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 0.12}) +[2026-06-15 17:57:50,956][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.18320666253566742, 'train/grad_norm_pre_clip': 8.61048698425293, 'timing/data': 0.0004970580339431763, 'timing/model': 1.4662965649622492, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 0.12}) +[2026-06-15 17:57:57,026][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.34337154030799866, 'train/grad_norm_pre_clip': 12.895950317382812, 'timing/data': 0.000506411015521735, 'timing/model': 1.5320381420315243, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 0.12}) +[2026-06-15 17:58:02,954][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.4360072612762451, 'train/grad_norm_pre_clip': 16.381315231323242, 'timing/data': 0.0005215040291659534, 'timing/model': 1.4615753059624694, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 0.12}) +[2026-06-15 17:58:08,935][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.4497884511947632, 'train/grad_norm_pre_clip': 12.493048667907715, 'timing/data': 0.0005050000036135316, 'timing/model': 1.4985670360038057, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 0.12}) +[2026-06-15 17:58:14,883][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.12184571474790573, 'train/grad_norm_pre_clip': 6.157914638519287, 'timing/data': 0.0005077039822936058, 'timing/model': 1.6930840330314822, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 0.12}) +[2026-06-15 17:58:20,554][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.17553868889808655, 'train/grad_norm_pre_clip': 8.32660961151123, 'timing/data': 0.0005041289841756225, 'timing/model': 1.464318825979717, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.12}) +[2026-06-15 17:58:26,455][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.2335256189107895, 'train/grad_norm_pre_clip': 8.815762519836426, 'timing/data': 0.0004968779976479709, 'timing/model': 1.4614506309735589, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 0.13}) +[2026-06-15 17:58:32,353][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.19876785576343536, 'train/grad_norm_pre_clip': 9.483070373535156, 'timing/data': 0.000487354991491884, 'timing/model': 1.4631445889826864, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 0.13}) +[2026-06-15 17:58:38,255][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.2527049779891968, 'train/grad_norm_pre_clip': 6.461085796356201, 'timing/data': 0.0004968779976479709, 'timing/model': 1.465805173967965, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 0.13}) +[2026-06-15 17:58:44,153][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.29517945647239685, 'train/grad_norm_pre_clip': 10.902589797973633, 'timing/data': 0.0005015850183553994, 'timing/model': 1.4315389010007493, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 0.13}) +[2026-06-15 17:58:50,181][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.23548972606658936, 'train/grad_norm_pre_clip': 7.129393577575684, 'timing/data': 0.0005026169819757342, 'timing/model': 1.6905288259731606, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 0.13}) +[2026-06-15 17:58:55,853][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.151797816157341, 'train/grad_norm_pre_clip': 3.7462358474731445, 'timing/data': 0.0005091759958304465, 'timing/model': 1.4597674050019123, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.13}) +[2026-06-15 17:59:01,753][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.2693662643432617, 'train/grad_norm_pre_clip': 4.34935998916626, 'timing/data': 0.00049034901894629, 'timing/model': 1.4645571159780957, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 0.13}) +[2026-06-15 17:59:07,654][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.13079562783241272, 'train/grad_norm_pre_clip': 5.860286235809326, 'timing/data': 0.0005042689736001194, 'timing/model': 1.464118083997164, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 0.13}) +[2026-06-15 17:59:13,554][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.4243068993091583, 'train/grad_norm_pre_clip': 11.787215232849121, 'timing/data': 0.0005340829957276583, 'timing/model': 1.4619058519601822, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 0.13}) +[2026-06-15 18:08:26,754][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.1420062929391861, 'train/grad_norm_pre_clip': 3.496986150741577, 'eval/loss': 0.13170699775218964, 'eval/flap_precision': 0.7657807308970099, 'eval/flap_recall': 0.9238476953907816, 'eval/flap_f1': 0.8374205267938237, 'eval/flap_rate_model': 0.16722222222222222, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.8374205267938237, 'timing/data': 0.0005315589951351285, 'timing/model': 1.5033076200052164, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.13}) +[2026-06-15 18:08:26,754][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_300_state +[2026-06-15 18:08:26,758][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 18:08:32,263][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_300_state/model.safetensors +[2026-06-15 18:08:40,246][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_300_state/optimizer.bin +[2026-06-15 18:08:40,247][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_300_state/sampler.bin +[2026-06-15 18:08:40,247][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_300_state/sampler_1.bin +[2026-06-15 18:08:40,248][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-15 18:08:40,355][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-15 18:08:40,356][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 18:08:40,366][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 18:08:46,329][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.31591352820396423, 'train/grad_norm_pre_clip': 7.791259765625, 'timing/data': 0.0005228680092841387, 'timing/model': 1.4879655049880967, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 0.13}) +[2026-06-15 18:08:52,252][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.41339051723480225, 'train/grad_norm_pre_clip': 15.087465286254883, 'timing/data': 0.0005381309892982244, 'timing/model': 1.4321103119873442, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 0.13}) +[2026-06-15 18:08:58,226][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.19156479835510254, 'train/grad_norm_pre_clip': 8.660073280334473, 'timing/data': 0.0005201539606787264, 'timing/model': 1.7398248339886777, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 0.13}) +[2026-06-15 18:09:03,948][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.1871267408132553, 'train/grad_norm_pre_clip': 4.9225311279296875, 'timing/data': 0.0005203250329941511, 'timing/model': 1.468011249031406, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 0.13}) +[2026-06-15 18:09:09,845][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.18346209824085236, 'train/grad_norm_pre_clip': 5.808328628540039, 'timing/data': 0.0005270050023682415, 'timing/model': 1.4647826820146292, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 0.13}) +[2026-06-15 18:09:15,846][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.1542109102010727, 'train/grad_norm_pre_clip': 6.676932334899902, 'timing/data': 0.0005310900160111487, 'timing/model': 1.483307390997652, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 0.13}) +[2026-06-15 18:09:21,755][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.15397129952907562, 'train/grad_norm_pre_clip': 4.572085380554199, 'timing/data': 0.000534134975168854, 'timing/model': 1.4433291149907745, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 0.13}) +[2026-06-15 18:09:27,646][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.3331241309642792, 'train/grad_norm_pre_clip': 9.746270179748535, 'timing/data': 0.000504321011248976, 'timing/model': 1.4359526520129293, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 0.13}) +[2026-06-15 18:09:33,609][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.17587421834468842, 'train/grad_norm_pre_clip': 7.9075608253479, 'timing/data': 0.0005162690067663789, 'timing/model': 1.623698509996757, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 0.13}) +[2026-06-15 18:09:39,249][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.2670045495033264, 'train/grad_norm_pre_clip': 19.01724624633789, 'timing/data': 0.0005255020223557949, 'timing/model': 1.4653059010161087, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.13}) +[2026-06-15 18:09:45,143][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.2579787075519562, 'train/grad_norm_pre_clip': 14.970861434936523, 'timing/data': 0.0005251110414974391, 'timing/model': 1.4510290600010194, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 0.13}) +[2026-06-15 18:09:51,044][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.2437315434217453, 'train/grad_norm_pre_clip': 7.6310200691223145, 'timing/data': 0.0005039399839006364, 'timing/model': 1.4650119650177658, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 0.13}) +[2026-06-15 18:09:56,950][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.076019786298275, 'train/grad_norm_pre_clip': 6.376834392547607, 'timing/data': 0.0004952679737471044, 'timing/model': 1.4244570070295595, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 0.13}) +[2026-06-15 18:10:02,880][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.3455018699169159, 'train/grad_norm_pre_clip': 8.194520950317383, 'timing/data': 0.0004993939655832946, 'timing/model': 1.697091143054422, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 0.14}) +[2026-06-15 18:10:08,547][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.2464127093553543, 'train/grad_norm_pre_clip': 15.032241821289062, 'timing/data': 0.0004940159851685166, 'timing/model': 1.4535358719876967, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 0.14}) +[2026-06-15 18:10:14,449][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.2580069899559021, 'train/grad_norm_pre_clip': 10.125402450561523, 'timing/data': 0.0005007360014133155, 'timing/model': 1.4663523940253071, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 0.14}) +[2026-06-15 18:10:20,452][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.10759329050779343, 'train/grad_norm_pre_clip': 8.79139518737793, 'timing/data': 0.0004918129998259246, 'timing/model': 1.471243591979146, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 0.14}) +[2026-06-15 18:10:26,352][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.22651921212673187, 'train/grad_norm_pre_clip': 11.681504249572754, 'timing/data': 0.0005085970042273402, 'timing/model': 1.4334915680228733, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 0.14}) +[2026-06-15 18:10:32,235][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.2530069649219513, 'train/grad_norm_pre_clip': 5.196437835693359, 'timing/data': 0.000499835005030036, 'timing/model': 1.5011428939760663, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 0.14}) +[2026-06-15 18:10:38,131][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.2660009562969208, 'train/grad_norm_pre_clip': 12.71674919128418, 'timing/data': 0.0004907510010525584, 'timing/model': 1.6400133509887382, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.14}) +[2026-06-15 18:10:43,852][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.2989848852157593, 'train/grad_norm_pre_clip': 9.743359565734863, 'timing/data': 0.0005032690241932869, 'timing/model': 1.4629666760447435, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 0.14}) +[2026-06-15 18:10:49,751][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.5398572087287903, 'train/grad_norm_pre_clip': 27.746158599853516, 'timing/data': 0.0005057629896327853, 'timing/model': 1.4640809330157936, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 0.14}) +[2026-06-15 18:10:55,750][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 1.2021280527114868, 'train/grad_norm_pre_clip': 45.7065315246582, 'timing/data': 0.0004976119962520897, 'timing/model': 1.4940054810140282, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 0.14}) +[2026-06-15 18:11:01,653][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.46140727400779724, 'train/grad_norm_pre_clip': 16.770822525024414, 'timing/data': 0.0005017980001866817, 'timing/model': 1.4390820530243218, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 0.14}) +[2026-06-15 18:11:07,552][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.5960312485694885, 'train/grad_norm_pre_clip': 22.56309700012207, 'timing/data': 0.0004972399910911918, 'timing/model': 1.4294132459908724, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 0.14}) +[2026-06-15 18:11:13,441][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.2273799180984497, 'train/grad_norm_pre_clip': 5.699766159057617, 'timing/data': 0.0004963690298609436, 'timing/model': 1.6499443089705892, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 0.14}) +[2026-06-15 18:11:19,152][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.20473052561283112, 'train/grad_norm_pre_clip': 4.924313068389893, 'timing/data': 0.0004964199615642428, 'timing/model': 1.4647389589808881, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 0.14}) +[2026-06-15 18:11:25,051][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.43775805830955505, 'train/grad_norm_pre_clip': 7.702662467956543, 'timing/data': 0.0005016769864596426, 'timing/model': 1.4662260630284436, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 0.14}) +[2026-06-15 18:11:31,053][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.27875959873199463, 'train/grad_norm_pre_clip': 5.987722396850586, 'timing/data': 0.0005107909673824906, 'timing/model': 1.463440538966097, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 0.14}) +[2026-06-15 18:11:36,957][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.3264852464199066, 'train/grad_norm_pre_clip': 8.326645851135254, 'timing/data': 0.0004961289814673364, 'timing/model': 1.4305848209769465, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.14}) +[2026-06-15 18:11:42,882][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.4830048084259033, 'train/grad_norm_pre_clip': 8.024864196777344, 'timing/data': 0.0005061239935457706, 'timing/model': 1.6940210500033572, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 0.14}) +[2026-06-15 18:11:48,550][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.4262961745262146, 'train/grad_norm_pre_clip': 8.156071662902832, 'timing/data': 0.000502637994941324, 'timing/model': 1.4645842249738052, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 0.14}) +[2026-06-15 18:11:54,452][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.2643074095249176, 'train/grad_norm_pre_clip': 4.920836925506592, 'timing/data': 0.0005054930225014687, 'timing/model': 1.4659082590369508, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 0.14}) +[2026-06-15 18:12:00,450][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.34835079312324524, 'train/grad_norm_pre_clip': 7.5998005867004395, 'timing/data': 0.0004952779854647815, 'timing/model': 1.4809656629804522, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 0.14}) +[2026-06-15 18:12:06,353][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.3403501510620117, 'train/grad_norm_pre_clip': 6.9061994552612305, 'timing/data': 0.0005036399816162884, 'timing/model': 1.4459582190029323, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 0.14}) +[2026-06-15 18:12:12,252][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.29382097721099854, 'train/grad_norm_pre_clip': 7.305133819580078, 'timing/data': 0.0004987929714843631, 'timing/model': 1.4785986599745229, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 0.14}) +[2026-06-15 18:12:18,186][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.18975631892681122, 'train/grad_norm_pre_clip': 3.641596794128418, 'timing/data': 0.0005001250538043678, 'timing/model': 1.698453076009173, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 0.15}) +[2026-06-15 18:12:23,855][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.22597500681877136, 'train/grad_norm_pre_clip': 5.721365451812744, 'timing/data': 0.0004942569648846984, 'timing/model': 1.4623647979460657, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 0.15}) +[2026-06-15 18:12:29,754][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.3442351818084717, 'train/grad_norm_pre_clip': 10.422280311584473, 'timing/data': 0.000491021026391536, 'timing/model': 1.4637648329953663, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 0.15}) +[2026-06-15 18:12:35,753][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.1799723207950592, 'train/grad_norm_pre_clip': 7.677374839782715, 'timing/data': 0.0005087870522402227, 'timing/model': 1.4557110539753921, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.15}) +[2026-06-15 18:12:41,754][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.2797696590423584, 'train/grad_norm_pre_clip': 16.488603591918945, 'timing/data': 0.0005020879907533526, 'timing/model': 1.454504397988785, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 0.15}) +[2026-06-15 18:12:47,652][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.2466675341129303, 'train/grad_norm_pre_clip': 10.043400764465332, 'timing/data': 0.0004948170389980078, 'timing/model': 1.4583470820216462, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 0.15}) +[2026-06-15 18:12:53,585][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.140895277261734, 'train/grad_norm_pre_clip': 8.524628639221191, 'timing/data': 0.000505962991155684, 'timing/model': 1.696632796956692, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 0.15}) +[2026-06-15 18:12:59,254][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.2571994960308075, 'train/grad_norm_pre_clip': 19.29912567138672, 'timing/data': 0.00048785697435960174, 'timing/model': 1.460646221996285, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 0.15}) +[2026-06-15 18:13:05,154][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.6838192343711853, 'train/grad_norm_pre_clip': 27.161283493041992, 'timing/data': 0.0004948069690726697, 'timing/model': 1.4654363060253672, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.15}) +[2026-06-15 18:13:11,154][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.2941613495349884, 'train/grad_norm_pre_clip': 7.745155334472656, 'timing/data': 0.0004965700209140778, 'timing/model': 1.4533794639864936, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 0.15}) +[2026-06-15 18:13:17,052][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.21273985505104065, 'train/grad_norm_pre_clip': 6.941059589385986, 'timing/data': 0.0004887279937975109, 'timing/model': 1.4642620019731112, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 0.15}) +[2026-06-15 18:13:22,955][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.2954219579696655, 'train/grad_norm_pre_clip': 12.452411651611328, 'timing/data': 0.0005042810225859284, 'timing/model': 1.430860310967546, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 0.15}) +[2026-06-15 18:13:28,884][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.14554154872894287, 'train/grad_norm_pre_clip': 8.668268203735352, 'timing/data': 0.0004924039822071791, 'timing/model': 1.6904828859842382, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 0.15}) +[2026-06-15 18:13:34,654][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.26925763487815857, 'train/grad_norm_pre_clip': 4.011838912963867, 'timing/data': 0.0005024679703637958, 'timing/model': 1.4495146939880215, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.15}) +[2026-06-15 18:13:40,552][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.257782906293869, 'train/grad_norm_pre_clip': 7.302401065826416, 'timing/data': 0.0004977919743396342, 'timing/model': 1.4638234869926237, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 0.15}) +[2026-06-15 18:13:46,454][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.3556677997112274, 'train/grad_norm_pre_clip': 17.172691345214844, 'timing/data': 0.000505432952195406, 'timing/model': 1.4651992639992386, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 0.15}) +[2026-06-15 18:13:52,354][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.21201656758785248, 'train/grad_norm_pre_clip': 4.581627368927002, 'timing/data': 0.000503199000377208, 'timing/model': 1.45415377500467, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 0.15}) +[2026-06-15 18:13:58,255][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.2527702748775482, 'train/grad_norm_pre_clip': 9.672004699707031, 'timing/data': 0.0005043710116297007, 'timing/model': 1.4312574989744462, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 0.15}) +[2026-06-15 18:14:04,185][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.31048867106437683, 'train/grad_norm_pre_clip': 8.685932159423828, 'timing/data': 0.0005073560168966651, 'timing/model': 1.6968877490144223, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 0.15}) +[2026-06-15 18:14:09,953][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.2996683120727539, 'train/grad_norm_pre_clip': 5.763308048248291, 'timing/data': 0.0004888279945589602, 'timing/model': 1.4623016609693877, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 0.15}) +[2026-06-15 18:14:15,853][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.2722516655921936, 'train/grad_norm_pre_clip': 9.821022033691406, 'timing/data': 0.0004999149823561311, 'timing/model': 1.4630442159832455, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 0.15}) +[2026-06-15 18:14:21,755][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.13516302406787872, 'train/grad_norm_pre_clip': 8.147789001464844, 'timing/data': 0.0005008959560655057, 'timing/model': 1.4650688259862363, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 0.15}) +[2026-06-15 18:14:27,653][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.12784357368946075, 'train/grad_norm_pre_clip': 4.108422756195068, 'timing/data': 0.0004999050288461149, 'timing/model': 1.463302153046243, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 0.15}) +[2026-06-15 18:14:33,553][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.1946180760860443, 'train/grad_norm_pre_clip': 6.958036422729492, 'timing/data': 0.0004920429782941937, 'timing/model': 1.4308372020022944, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.15}) +[2026-06-15 18:14:39,483][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.25763580203056335, 'train/grad_norm_pre_clip': 7.263786792755127, 'timing/data': 0.0005014869966544211, 'timing/model': 1.6905169899691828, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 0.16}) +[2026-06-15 18:14:45,153][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.4102827310562134, 'train/grad_norm_pre_clip': 10.889495849609375, 'timing/data': 0.0005037799710407853, 'timing/model': 1.462777909997385, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 0.16}) +[2026-06-15 18:14:51,054][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.2715096175670624, 'train/grad_norm_pre_clip': 8.6893949508667, 'timing/data': 0.000497782020829618, 'timing/model': 1.4652771269902587, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 0.16}) +[2026-06-15 18:14:57,052][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.2120441198348999, 'train/grad_norm_pre_clip': 6.379589080810547, 'timing/data': 0.000492393970489502, 'timing/model': 1.5032356190495193, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 0.16}) +[2026-06-15 18:15:03,049][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.2216271460056305, 'train/grad_norm_pre_clip': 6.9727983474731445, 'timing/data': 0.0004984429688192904, 'timing/model': 1.4771762610180303, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 0.16}) +[2026-06-15 18:15:08,954][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.40260079503059387, 'train/grad_norm_pre_clip': 12.530710220336914, 'timing/data': 0.0005729919648729265, 'timing/model': 1.4356583720073104, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 0.16}) +[2026-06-15 18:15:14,882][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.3163125514984131, 'train/grad_norm_pre_clip': 9.771805763244629, 'timing/data': 0.0005588109488599002, 'timing/model': 1.690967349044513, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 0.16}) +[2026-06-15 18:15:20,703][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.24917608499526978, 'train/grad_norm_pre_clip': 8.965376853942871, 'timing/data': 0.0005689859972335398, 'timing/model': 1.6085871219984256, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 0.16}) +[2026-06-15 18:15:26,355][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.1688929796218872, 'train/grad_norm_pre_clip': 3.5597496032714844, 'timing/data': 0.0005591109511442482, 'timing/model': 1.4652641309658065, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 0.16}) +[2026-06-15 18:15:32,255][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.15571212768554688, 'train/grad_norm_pre_clip': 4.425814628601074, 'timing/data': 0.0005448909942060709, 'timing/model': 1.4595015840022825, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.16}) +[2026-06-15 18:15:38,262][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.3300199806690216, 'train/grad_norm_pre_clip': 10.710206031799316, 'timing/data': 0.0005177810089662671, 'timing/model': 1.4569090449949726, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 0.16}) +[2026-06-15 18:15:44,162][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.4157140851020813, 'train/grad_norm_pre_clip': 19.911102294921875, 'timing/data': 0.0005189130315557122, 'timing/model': 1.4363059440511279, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 0.16}) +[2026-06-15 18:15:50,085][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.40613406896591187, 'train/grad_norm_pre_clip': 12.44504165649414, 'timing/data': 0.0005193530232645571, 'timing/model': 1.6940774510148913, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 0.16}) +[2026-06-15 18:15:55,982][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.12197382748126984, 'train/grad_norm_pre_clip': 5.698019027709961, 'timing/data': 0.0005233589909039438, 'timing/model': 1.6917267990065739, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 0.16}) +[2026-06-15 18:16:01,657][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.23739181458950043, 'train/grad_norm_pre_clip': 6.365739345550537, 'timing/data': 0.0005237589939497411, 'timing/model': 1.4630400909809396, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 0.16}) +[2026-06-15 18:16:07,558][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.15083548426628113, 'train/grad_norm_pre_clip': 6.73546028137207, 'timing/data': 0.000531370984390378, 'timing/model': 1.466682396014221, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 0.16}) +[2026-06-15 18:16:13,563][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.1974700391292572, 'train/grad_norm_pre_clip': 5.798819065093994, 'timing/data': 0.0005213459953665733, 'timing/model': 1.4850097360322252, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 0.16}) +[2026-06-15 18:16:19,541][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.17276522517204285, 'train/grad_norm_pre_clip': 7.165640354156494, 'timing/data': 0.0005277059972286224, 'timing/model': 1.5098206750117242, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 0.16}) +[2026-06-15 18:16:25,501][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.17820380628108978, 'train/grad_norm_pre_clip': 8.455331802368164, 'timing/data': 0.0005236490396782756, 'timing/model': 1.7101864560390823, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 0.16}) +[2026-06-15 18:16:31,157][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.20524248480796814, 'train/grad_norm_pre_clip': 10.094897270202637, 'timing/data': 0.0005162389716133475, 'timing/model': 1.463771059003193, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.16}) +[2026-06-15 18:16:37,159][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.2352508306503296, 'train/grad_norm_pre_clip': 12.534608840942383, 'timing/data': 0.0005177209968678653, 'timing/model': 1.4663172319997102, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 0.16}) +[2026-06-15 18:16:43,154][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.2038632184267044, 'train/grad_norm_pre_clip': 19.302143096923828, 'timing/data': 0.00053264299640432, 'timing/model': 1.479799719993025, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 0.16}) +[2026-06-15 18:16:49,056][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.09303999692201614, 'train/grad_norm_pre_clip': 4.799651622772217, 'timing/data': 0.0005117419641464949, 'timing/model': 1.4397343790042214, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 0.16}) +[2026-06-15 18:16:54,941][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.1143488958477974, 'train/grad_norm_pre_clip': 8.329341888427734, 'timing/data': 0.0005234690033830702, 'timing/model': 1.5058964109630324, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 0.17}) +[2026-06-15 18:17:00,842][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.1931706815958023, 'train/grad_norm_pre_clip': 9.848405838012695, 'timing/data': 0.0005654810229316354, 'timing/model': 1.6527559999958612, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 0.17}) +[2026-06-15 18:17:06,556][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.32653531432151794, 'train/grad_norm_pre_clip': 12.02812671661377, 'timing/data': 0.0005239200545474887, 'timing/model': 1.4653301420039497, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 0.17}) +[2026-06-15 18:17:12,454][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.0664411261677742, 'train/grad_norm_pre_clip': 6.330013275146484, 'timing/data': 0.0005274550057947636, 'timing/model': 1.4653120849980041, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 0.17}) +[2026-06-15 18:17:18,453][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.6740595698356628, 'train/grad_norm_pre_clip': 17.767126083374023, 'timing/data': 0.0004939259961247444, 'timing/model': 1.4593444329802878, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 0.17}) +[2026-06-15 18:17:24,352][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.2225339710712433, 'train/grad_norm_pre_clip': 7.684620380401611, 'timing/data': 0.0004980419762432575, 'timing/model': 1.4371484160074033, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 0.17}) +[2026-06-15 18:17:30,254][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.23291350901126862, 'train/grad_norm_pre_clip': 11.960391998291016, 'timing/data': 0.0006229350110515952, 'timing/model': 1.436003971961327, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.17}) +[2026-06-15 18:17:36,141][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.3071877062320709, 'train/grad_norm_pre_clip': 9.188663482666016, 'timing/data': 0.0005757960025221109, 'timing/model': 1.6512155790114775, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 0.17}) +[2026-06-15 18:17:41,851][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.36285653710365295, 'train/grad_norm_pre_clip': 11.5173978805542, 'timing/data': 0.0005009459564462304, 'timing/model': 1.46184618701227, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 0.17}) +[2026-06-15 18:17:47,752][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.1334591656923294, 'train/grad_norm_pre_clip': 5.200084209442139, 'timing/data': 0.0005079659749753773, 'timing/model': 1.4635363499983214, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 0.17}) +[2026-06-15 18:17:53,851][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.1897232085466385, 'train/grad_norm_pre_clip': 7.697657585144043, 'timing/data': 0.0005192030221223831, 'timing/model': 1.4657868779613636, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.17}) +[2026-06-15 18:17:59,752][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.2182510793209076, 'train/grad_norm_pre_clip': 6.588620662689209, 'timing/data': 0.0005226589855737984, 'timing/model': 1.431902197015006, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 0.17}) +[2026-06-15 18:18:05,685][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.12957468628883362, 'train/grad_norm_pre_clip': 9.710958480834961, 'timing/data': 0.0005260729813016951, 'timing/model': 1.6916685350006446, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 0.17}) +[2026-06-15 18:18:11,354][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.25050103664398193, 'train/grad_norm_pre_clip': 10.721552848815918, 'timing/data': 0.0005218780133873224, 'timing/model': 1.4630496100289747, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 0.17}) +[2026-06-15 18:18:17,254][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.21592171490192413, 'train/grad_norm_pre_clip': 7.890695571899414, 'timing/data': 0.0005343250231817365, 'timing/model': 1.463901662034914, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 0.17}) +[2026-06-15 18:18:23,055][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.4313732385635376, 'train/grad_norm_pre_clip': 17.64318084716797, 'timing/data': 0.0005190829979255795, 'timing/model': 1.4638684310484678, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 0.17}) +[2026-06-15 18:27:31,652][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.2940811812877655, 'train/grad_norm_pre_clip': 8.680411338806152, 'eval/loss': 0.1261361539363861, 'eval/flap_precision': 0.8571428571428571, 'eval/flap_recall': 0.8296593186372746, 'eval/flap_f1': 0.8431771894093686, 'eval/flap_rate_model': 0.13416666666666666, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.8431771894093685, 'timing/data': 0.0005026290309615433, 'timing/model': 1.4648817079723813, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.17}) +[2026-06-15 18:27:31,653][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_400_state +[2026-06-15 18:27:31,657][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 18:27:37,592][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_400_state/model.safetensors +[2026-06-15 18:27:45,752][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_400_state/optimizer.bin +[2026-06-15 18:27:45,753][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_400_state/sampler.bin +[2026-06-15 18:27:45,753][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_400_state/sampler_1.bin +[2026-06-15 18:27:45,754][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-15 18:27:45,868][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-15 18:27:45,869][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 18:27:45,879][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 18:27:51,837][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.3277009427547455, 'train/grad_norm_pre_clip': 6.144942283630371, 'timing/data': 0.0005291730049066246, 'timing/model': 1.4637320839683525, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 0.17}) +[2026-06-15 18:27:57,743][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.3114754557609558, 'train/grad_norm_pre_clip': 6.263035297393799, 'timing/data': 0.0005296129966154695, 'timing/model': 1.4333117789938115, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 0.17}) +[2026-06-15 18:28:03,543][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.15054355561733246, 'train/grad_norm_pre_clip': 2.9297127723693848, 'timing/data': 0.0005059479735791683, 'timing/model': 1.4588505200226791, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 0.17}) +[2026-06-15 18:28:09,482][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.2739294469356537, 'train/grad_norm_pre_clip': 7.382421970367432, 'timing/data': 0.0005057469825260341, 'timing/model': 1.7040742079843767, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 0.17}) +[2026-06-15 18:28:15,244][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.34171414375305176, 'train/grad_norm_pre_clip': 12.117141723632812, 'timing/data': 0.0005099939880892634, 'timing/model': 1.4643026099656709, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 0.17}) +[2026-06-15 18:28:21,247][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.2330656349658966, 'train/grad_norm_pre_clip': 5.427123546600342, 'timing/data': 0.0005017720395699143, 'timing/model': 1.4542824840173125, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 0.17}) +[2026-06-15 18:28:27,248][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.36266687512397766, 'train/grad_norm_pre_clip': 6.4132466316223145, 'timing/data': 0.0005021719844080508, 'timing/model': 1.469083046016749, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 0.18}) +[2026-06-15 18:28:33,245][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.1541290581226349, 'train/grad_norm_pre_clip': 5.096348285675049, 'timing/data': 0.0004990380257368088, 'timing/model': 1.4592714000027627, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 0.18}) +[2026-06-15 18:28:39,319][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.0873216837644577, 'train/grad_norm_pre_clip': 2.636990547180176, 'timing/data': 0.0005193770048208535, 'timing/model': 1.4901290020206943, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 0.18}) +[2026-06-15 18:28:45,186][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.24770858883857727, 'train/grad_norm_pre_clip': 11.062164306640625, 'timing/data': 0.0004914560122415423, 'timing/model': 1.6927337700035423, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.18}) +[2026-06-15 18:28:50,952][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.38067853450775146, 'train/grad_norm_pre_clip': 15.029480934143066, 'timing/data': 0.0005313149886205792, 'timing/model': 1.5612052700016648, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 0.18}) +[2026-06-15 18:28:57,024][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.3805497884750366, 'train/grad_norm_pre_clip': 17.326412200927734, 'timing/data': 0.0004977959906682372, 'timing/model': 1.5213777809985913, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 0.18}) +[2026-06-15 18:29:02,954][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.07583115249872208, 'train/grad_norm_pre_clip': 2.5681283473968506, 'timing/data': 0.000500358990393579, 'timing/model': 1.4436748770531267, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 0.18}) +[2026-06-15 18:29:08,954][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.10900451987981796, 'train/grad_norm_pre_clip': 5.856906414031982, 'timing/data': 0.000501219998113811, 'timing/model': 1.4638768220320344, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 0.18}) +[2026-06-15 18:29:14,858][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.17899182438850403, 'train/grad_norm_pre_clip': 8.049206733703613, 'timing/data': 0.0005240339669398963, 'timing/model': 1.4346938299713656, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.18}) +[2026-06-15 18:29:20,790][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.2098003327846527, 'train/grad_norm_pre_clip': 7.230748176574707, 'timing/data': 0.000524083967320621, 'timing/model': 1.697458129026927, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 0.18}) +[2026-06-15 18:29:26,478][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.053266752511262894, 'train/grad_norm_pre_clip': 3.840092182159424, 'timing/data': 0.0005365629913285375, 'timing/model': 1.4895385580020957, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 0.18}) +[2026-06-15 18:29:32,456][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.08311893790960312, 'train/grad_norm_pre_clip': 7.020906925201416, 'timing/data': 0.0005133880185894668, 'timing/model': 1.4643890639999881, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 0.18}) +[2026-06-15 18:29:38,353][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.08144409954547882, 'train/grad_norm_pre_clip': 7.287850379943848, 'timing/data': 0.0005225120112299919, 'timing/model': 1.4641293520107865, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 0.18}) +[2026-06-15 18:29:44,256][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.1848309189081192, 'train/grad_norm_pre_clip': 11.899209976196289, 'timing/data': 0.0005254359566606581, 'timing/model': 1.4677776750177145, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.18}) +[2026-06-15 18:29:50,154][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.02044071815907955, 'train/grad_norm_pre_clip': 2.6581473350524902, 'timing/data': 0.0005192770040594041, 'timing/model': 1.4335600070189685, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 0.18}) +[2026-06-15 18:29:56,081][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.15310657024383545, 'train/grad_norm_pre_clip': 16.953187942504883, 'timing/data': 0.000524875009432435, 'timing/model': 1.6947179169510491, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 0.18}) +[2026-06-15 18:30:01,757][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.36627262830734253, 'train/grad_norm_pre_clip': 9.203780174255371, 'timing/data': 0.0004922570078633726, 'timing/model': 1.4665850719902664, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 0.18}) +[2026-06-15 18:30:07,657][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.4014357924461365, 'train/grad_norm_pre_clip': 36.513038635253906, 'timing/data': 0.0004890820127911866, 'timing/model': 1.4657297049998306, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 0.18}) +[2026-06-15 18:30:13,657][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.05531768500804901, 'train/grad_norm_pre_clip': 10.02175235748291, 'timing/data': 0.0004958230420015752, 'timing/model': 1.4680339050246403, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 0.18}) +[2026-06-15 18:30:19,657][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.18794117867946625, 'train/grad_norm_pre_clip': 17.50621223449707, 'timing/data': 0.0005037939990870655, 'timing/model': 1.456336960021872, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 0.18}) +[2026-06-15 18:30:25,556][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.04364260658621788, 'train/grad_norm_pre_clip': 8.51283073425293, 'timing/data': 0.0004938690108247101, 'timing/model': 1.4358397120377049, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 0.18}) +[2026-06-15 18:30:31,485][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.2921939492225647, 'train/grad_norm_pre_clip': 15.593080520629883, 'timing/data': 0.0004911349969916046, 'timing/model': 1.693530689983163, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 0.18}) +[2026-06-15 18:30:37,402][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.3110564947128296, 'train/grad_norm_pre_clip': 19.312856674194336, 'timing/data': 0.0005034329951740801, 'timing/model': 1.6057166639948264, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 0.18}) +[2026-06-15 18:30:43,158][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.21458649635314941, 'train/grad_norm_pre_clip': 15.55097484588623, 'timing/data': 0.0005617190035991371, 'timing/model': 1.4495297109824605, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.19}) +[2026-06-15 18:30:49,158][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.12242785096168518, 'train/grad_norm_pre_clip': 9.075058937072754, 'timing/data': 0.0005627209902741015, 'timing/model': 1.4644303749664687, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 0.19}) +[2026-06-15 18:30:55,156][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.5658861994743347, 'train/grad_norm_pre_clip': 30.928483963012695, 'timing/data': 0.0005669869715347886, 'timing/model': 1.4718954600393772, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 0.19}) +[2026-06-15 18:31:01,255][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.2574993968009949, 'train/grad_norm_pre_clip': 16.77032470703125, 'timing/data': 0.0005082600400783122, 'timing/model': 1.5140840570093133, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 0.19}) +[2026-06-15 18:31:07,185][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.20829522609710693, 'train/grad_norm_pre_clip': 13.86091423034668, 'timing/data': 0.0005279799806885421, 'timing/model': 1.6909912769915536, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 0.19}) +[2026-06-15 18:31:13,091][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.29022732377052307, 'train/grad_norm_pre_clip': 11.620641708374023, 'timing/data': 0.0005345589597709477, 'timing/model': 1.6972309849807061, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 0.19}) +[2026-06-15 18:31:18,854][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.13714060187339783, 'train/grad_norm_pre_clip': 7.730504989624023, 'timing/data': 0.0005133780068717897, 'timing/model': 1.463737586978823, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 0.19}) +[2026-06-15 18:31:24,756][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.25309720635414124, 'train/grad_norm_pre_clip': 22.08057975769043, 'timing/data': 0.0005112640210427344, 'timing/model': 1.4643427599803545, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 0.19}) +[2026-06-15 18:31:30,756][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.36757057905197144, 'train/grad_norm_pre_clip': 7.758730888366699, 'timing/data': 0.0005175639525987208, 'timing/model': 1.4830215809633955, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 0.19}) +[2026-06-15 18:31:36,661][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.145704984664917, 'train/grad_norm_pre_clip': 5.053610324859619, 'timing/data': 0.0005218599690124393, 'timing/model': 1.4357504789950326, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 0.19}) +[2026-06-15 18:31:42,589][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.12321723997592926, 'train/grad_norm_pre_clip': 4.913635730743408, 'timing/data': 0.0005195660050958395, 'timing/model': 1.6989267289754935, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.19}) +[2026-06-15 18:31:48,254][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.30836835503578186, 'train/grad_norm_pre_clip': 16.908187866210938, 'timing/data': 0.000507378950715065, 'timing/model': 1.4648623479879461, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 0.19}) +[2026-06-15 18:31:54,155][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.3624584972858429, 'train/grad_norm_pre_clip': 23.962387084960938, 'timing/data': 0.0004974540206603706, 'timing/model': 1.4656994990073144, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 0.19}) +[2026-06-15 18:32:00,159][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.15140895545482635, 'train/grad_norm_pre_clip': 6.135745525360107, 'timing/data': 0.0005230619572103024, 'timing/model': 1.4797129739890806, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.19}) +[2026-06-15 18:32:06,060][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.13919012248516083, 'train/grad_norm_pre_clip': 4.185288906097412, 'timing/data': 0.0005237830337136984, 'timing/model': 1.4380923690041527, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 0.19}) +[2026-06-15 18:32:11,949][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.36257147789001465, 'train/grad_norm_pre_clip': 9.026586532592773, 'timing/data': 0.0005235629505477846, 'timing/model': 1.5084226239705458, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 0.19}) +[2026-06-15 18:32:17,933][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.29445910453796387, 'train/grad_norm_pre_clip': 12.808389663696289, 'timing/data': 0.0005015799542888999, 'timing/model': 1.6442819199874066, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 0.19}) +[2026-06-15 18:32:23,660][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.2320428043603897, 'train/grad_norm_pre_clip': 5.8595871925354, 'timing/data': 0.0005257250159047544, 'timing/model': 1.4659335290198214, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 0.19}) +[2026-06-15 18:32:29,660][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.20048697292804718, 'train/grad_norm_pre_clip': 9.444868087768555, 'timing/data': 0.0005349200218915939, 'timing/model': 1.4412031169631518, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 0.19}) +[2026-06-15 18:32:35,658][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.21921391785144806, 'train/grad_norm_pre_clip': 6.015780448913574, 'timing/data': 0.0005238120211288333, 'timing/model': 1.475333477021195, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 0.19}) +[2026-06-15 18:32:41,559][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.1520780473947525, 'train/grad_norm_pre_clip': 3.9231958389282227, 'timing/data': 0.0005260060424916446, 'timing/model': 1.4361227040062658, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.19}) +[2026-06-15 18:32:47,465][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.2598208785057068, 'train/grad_norm_pre_clip': 9.34598159790039, 'timing/data': 0.0005167730269022286, 'timing/model': 1.4386720380280167, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 0.19}) +[2026-06-15 18:32:53,446][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.21676458418369293, 'train/grad_norm_pre_clip': 6.005530834197998, 'timing/data': 0.0005211289972066879, 'timing/model': 1.6781806359649636, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 0.19}) +[2026-06-15 18:32:59,159][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.1301484853029251, 'train/grad_norm_pre_clip': 3.9165194034576416, 'timing/data': 0.0005153300007805228, 'timing/model': 1.4663616609759629, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 0.19}) +[2026-06-15 18:33:05,057][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.23985394835472107, 'train/grad_norm_pre_clip': 9.408863067626953, 'timing/data': 0.0005149099742993712, 'timing/model': 1.4635760569944978, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 0.2}) +[2026-06-15 18:33:11,057][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.3079468607902527, 'train/grad_norm_pre_clip': 9.688623428344727, 'timing/data': 0.0005221000174060464, 'timing/model': 1.48567925702082, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 0.2}) +[2026-06-15 18:33:17,039][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.15035706758499146, 'train/grad_norm_pre_clip': 5.084706783294678, 'timing/data': 0.000518594984896481, 'timing/model': 1.5089111240231432, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 0.2}) +[2026-06-15 18:33:22,987][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.1301233172416687, 'train/grad_norm_pre_clip': 7.209721088409424, 'timing/data': 0.0005167630151845515, 'timing/model': 1.6932171660009772, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 0.2}) +[2026-06-15 18:33:28,655][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.1947423219680786, 'train/grad_norm_pre_clip': 7.207759857177734, 'timing/data': 0.0005145189934410155, 'timing/model': 1.4631561619462445, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 0.2}) +[2026-06-15 18:33:34,654][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.2224133312702179, 'train/grad_norm_pre_clip': 11.78286361694336, 'timing/data': 0.0005281390040181577, 'timing/model': 1.4650885990122333, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 0.2}) +[2026-06-15 18:33:40,554][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.22089038789272308, 'train/grad_norm_pre_clip': 21.923255920410156, 'timing/data': 0.0005163320456631482, 'timing/model': 1.4521512850187719, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.2}) +[2026-06-15 18:33:46,468][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.26709046959877014, 'train/grad_norm_pre_clip': 11.514654159545898, 'timing/data': 0.0005205179913900793, 'timing/model': 1.478657921019476, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 0.2}) +[2026-06-15 18:33:52,357][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.2949390113353729, 'train/grad_norm_pre_clip': 6.954136371612549, 'timing/data': 0.0005155800026841462, 'timing/model': 1.4848853070288897, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 0.2}) +[2026-06-15 18:33:58,372][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.18289335072040558, 'train/grad_norm_pre_clip': 5.405186176300049, 'timing/data': 0.0005099729751236737, 'timing/model': 1.7369745819596574, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 0.2}) +[2026-06-15 18:34:04,057][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.3141597807407379, 'train/grad_norm_pre_clip': 15.038722038269043, 'timing/data': 0.0004987160209566355, 'timing/model': 1.4643639129935764, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 0.2}) +[2026-06-15 18:34:09,958][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.35563087463378906, 'train/grad_norm_pre_clip': 16.145660400390625, 'timing/data': 0.0004917159676551819, 'timing/model': 1.4666821549762972, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 0.2}) +[2026-06-15 18:34:15,958][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.10522794723510742, 'train/grad_norm_pre_clip': 3.6958651542663574, 'timing/data': 0.0005213189870119095, 'timing/model': 1.488783962035086, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 0.2}) +[2026-06-15 18:34:22,001][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.18484020233154297, 'train/grad_norm_pre_clip': 8.183025360107422, 'timing/data': 0.0005168920033611357, 'timing/model': 1.5158160789869726, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 0.2}) +[2026-06-15 18:34:27,952][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.5979277491569519, 'train/grad_norm_pre_clip': 21.847341537475586, 'timing/data': 0.0005214089760556817, 'timing/model': 1.5117685010191053, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 0.2}) +[2026-06-15 18:34:33,889][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.1871231496334076, 'train/grad_norm_pre_clip': 6.810000419616699, 'timing/data': 0.0005110040074214339, 'timing/model': 1.6951714010210708, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 0.2}) +[2026-06-15 18:34:39,658][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.2981853783130646, 'train/grad_norm_pre_clip': 11.073128700256348, 'timing/data': 0.0005190949887037277, 'timing/model': 1.4572926770197228, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.2}) +[2026-06-15 18:34:45,558][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.27622029185295105, 'train/grad_norm_pre_clip': 10.461894035339355, 'timing/data': 0.0005222799954935908, 'timing/model': 1.4642519959597848, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 0.2}) +[2026-06-15 18:34:51,557][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.3798443377017975, 'train/grad_norm_pre_clip': 13.903825759887695, 'timing/data': 0.0005186740309000015, 'timing/model': 1.4861822379752994, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 0.2}) +[2026-06-15 18:34:57,488][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.2755192220211029, 'train/grad_norm_pre_clip': 10.850444793701172, 'timing/data': 0.0005443220143206418, 'timing/model': 1.480773670016788, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 0.2}) +[2026-06-15 18:35:03,557][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.2266993224620819, 'train/grad_norm_pre_clip': 4.759668827056885, 'timing/data': 0.000514708983246237, 'timing/model': 1.4955848550307564, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 0.2}) +[2026-06-15 18:35:09,487][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.11063722521066666, 'train/grad_norm_pre_clip': 2.8511366844177246, 'timing/data': 0.0005382740055210888, 'timing/model': 1.692023399984464, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 0.2}) +[2026-06-15 18:35:15,158][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.20690183341503143, 'train/grad_norm_pre_clip': 6.202858924865723, 'timing/data': 0.0005118449917063117, 'timing/model': 1.464062235027086, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 0.2}) +[2026-06-15 18:35:21,158][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.1652139574289322, 'train/grad_norm_pre_clip': 6.051533222198486, 'timing/data': 0.0005083190044388175, 'timing/model': 1.4640022260136902, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 0.21}) +[2026-06-15 18:35:27,054][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.2293471097946167, 'train/grad_norm_pre_clip': 5.104027271270752, 'timing/data': 0.0005117849796079099, 'timing/model': 1.464762668008916, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 0.21}) +[2026-06-15 18:35:32,957][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.09162182360887527, 'train/grad_norm_pre_clip': 4.33484411239624, 'timing/data': 0.0005240020109340549, 'timing/model': 1.4677774230367504, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 0.21}) +[2026-06-15 18:35:38,859][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.13108615577220917, 'train/grad_norm_pre_clip': 6.195688247680664, 'timing/data': 0.0005158000276423991, 'timing/model': 1.4358830439741723, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.21}) +[2026-06-15 18:35:44,787][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.330167680978775, 'train/grad_norm_pre_clip': 11.652766227722168, 'timing/data': 0.0005147489719092846, 'timing/model': 1.6958237020298839, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 0.21}) +[2026-06-15 18:35:50,459][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.37332749366760254, 'train/grad_norm_pre_clip': 16.67987823486328, 'timing/data': 0.000522290007211268, 'timing/model': 1.465794826974161, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 0.21}) +[2026-06-15 18:35:56,457][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.10074887424707413, 'train/grad_norm_pre_clip': 3.9434032440185547, 'timing/data': 0.0005916219670325518, 'timing/model': 1.451505494012963, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 0.21}) +[2026-06-15 18:36:02,458][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.45957839488983154, 'train/grad_norm_pre_clip': 13.34226131439209, 'timing/data': 0.0005210879608057439, 'timing/model': 1.4639301030547358, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 0.21}) +[2026-06-15 18:36:08,434][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.4115293622016907, 'train/grad_norm_pre_clip': 15.392444610595703, 'timing/data': 0.0005075879744254053, 'timing/model': 1.5414520140038803, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.21}) +[2026-06-15 18:36:14,452][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.4244942367076874, 'train/grad_norm_pre_clip': 13.695935249328613, 'timing/data': 0.0005181239685043693, 'timing/model': 1.5027212629793212, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 0.21}) +[2026-06-15 18:36:20,383][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.2919541001319885, 'train/grad_norm_pre_clip': 16.044095993041992, 'timing/data': 0.0005156899569556117, 'timing/model': 1.6898396299802698, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 0.21}) +[2026-06-15 18:36:26,064][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.2609957158565521, 'train/grad_norm_pre_clip': 14.725752830505371, 'timing/data': 0.0005143479793332517, 'timing/model': 1.4726133039803244, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 0.21}) +[2026-06-15 18:36:31,959][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.24178345501422882, 'train/grad_norm_pre_clip': 10.133747100830078, 'timing/data': 0.0005204380140639842, 'timing/model': 1.4672129200189374, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 0.21}) +[2026-06-15 18:36:37,859][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.5101400017738342, 'train/grad_norm_pre_clip': 13.231823921203613, 'timing/data': 0.0005218389560468495, 'timing/model': 1.464025547960773, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 0.21}) +[2026-06-15 18:36:43,872][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.17115294933319092, 'train/grad_norm_pre_clip': 7.5537190437316895, 'timing/data': 0.0005327959661372006, 'timing/model': 1.4775380510254763, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 0.21}) +[2026-06-15 18:36:49,762][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.2685355246067047, 'train/grad_norm_pre_clip': 8.482300758361816, 'timing/data': 0.0005234419950284064, 'timing/model': 1.4360866929637268, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.21}) +[2026-06-15 18:36:55,684][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.1332709789276123, 'train/grad_norm_pre_clip': 5.931707382202148, 'timing/data': 0.0005586729967035353, 'timing/model': 1.6901706000207923, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 0.21}) +[2026-06-15 18:37:01,504][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.3428592085838318, 'train/grad_norm_pre_clip': 5.180180072784424, 'timing/data': 0.0005188239738345146, 'timing/model': 1.6103763290448114, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 0.21}) +[2026-06-15 18:37:07,159][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.2722505033016205, 'train/grad_norm_pre_clip': 7.2466301918029785, 'timing/data': 0.0005172429955564439, 'timing/model': 1.4635256339679472, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 0.21}) +[2026-06-15 18:37:13,059][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.17661140859127045, 'train/grad_norm_pre_clip': 4.172900676727295, 'timing/data': 0.0005134060047566891, 'timing/model': 1.4635679459897801, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 0.21}) +[2026-06-15 18:37:19,060][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.25777506828308105, 'train/grad_norm_pre_clip': 5.836801528930664, 'timing/data': 0.0005177130224183202, 'timing/model': 1.44882187404437, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 0.21}) +[2026-06-15 18:37:25,064][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.3628036379814148, 'train/grad_norm_pre_clip': 5.937369346618652, 'timing/data': 0.0005200770101509988, 'timing/model': 1.4375108639942482, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 0.21}) +[2026-06-15 18:37:30,985][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.1508326381444931, 'train/grad_norm_pre_clip': 3.4400715827941895, 'timing/data': 0.0005146590410731733, 'timing/model': 1.6899130249512382, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 0.21}) +[2026-06-15 18:46:46,161][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.45345431566238403, 'train/grad_norm_pre_clip': 8.185179710388184, 'eval/loss': 0.10957035422325134, 'eval/flap_precision': 0.8704453441295547, 'eval/flap_recall': 0.8617234468937875, 'eval/flap_f1': 0.866062437059416, 'eval/flap_rate_model': 0.13722222222222222, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.8660624370594161, 'timing/data': 0.0005195050034672022, 'timing/model': 1.6985123220365494, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 0.22}) +[2026-06-15 18:46:46,162][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_500_state +[2026-06-15 18:46:46,167][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 18:46:52,146][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_500_state/model.safetensors +[2026-06-15 18:47:06,328][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_500_state/optimizer.bin +[2026-06-15 18:47:06,329][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_500_state/sampler.bin +[2026-06-15 18:47:06,330][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_500_state/sampler_1.bin +[2026-06-15 18:47:06,331][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-15 18:47:06,443][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-15 18:47:06,443][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 18:47:06,453][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 18:47:12,527][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.2248946577310562, 'train/grad_norm_pre_clip': 5.791399955749512, 'timing/data': 0.0005318130133673549, 'timing/model': 1.7480306409997866, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 0.22}) +[2026-06-15 18:47:18,345][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.2046668827533722, 'train/grad_norm_pre_clip': 6.289888381958008, 'timing/data': 0.0006900759763084352, 'timing/model': 1.4608462190371938, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 0.22}) +[2026-06-15 18:47:24,416][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.36163803935050964, 'train/grad_norm_pre_clip': 8.232626914978027, 'timing/data': 0.0005938850226812065, 'timing/model': 1.520349862985313, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 0.22}) +[2026-06-15 18:47:30,450][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.28590577840805054, 'train/grad_norm_pre_clip': 8.482001304626465, 'timing/data': 0.0005245819920673966, 'timing/model': 1.4573929699836299, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 0.22}) +[2026-06-15 18:47:36,455][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.2439047247171402, 'train/grad_norm_pre_clip': 6.670882701873779, 'timing/data': 0.0005217579891905189, 'timing/model': 1.4394546099938452, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 0.22}) +[2026-06-15 18:47:42,352][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.31138622760772705, 'train/grad_norm_pre_clip': 5.495937347412109, 'timing/data': 0.0005108320037834346, 'timing/model': 1.4824618399725296, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 0.22}) +[2026-06-15 18:47:48,478][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.18688669800758362, 'train/grad_norm_pre_clip': 5.703086853027344, 'timing/data': 0.0005873150075785816, 'timing/model': 1.7032977410126477, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 0.22}) +[2026-06-15 18:47:54,157][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.37351182103157043, 'train/grad_norm_pre_clip': 8.032737731933594, 'timing/data': 0.000529429002199322, 'timing/model': 1.4656954850070179, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 0.22}) +[2026-06-15 18:48:00,256][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.3962462842464447, 'train/grad_norm_pre_clip': 17.350263595581055, 'timing/data': 0.0005229890230111778, 'timing/model': 1.4739455580129288, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 0.22}) +[2026-06-15 18:48:06,258][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.24499765038490295, 'train/grad_norm_pre_clip': 6.057784080505371, 'timing/data': 0.0005346370162442327, 'timing/model': 1.4494568970403634, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 0.22}) +[2026-06-15 18:48:12,253][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.23653379082679749, 'train/grad_norm_pre_clip': 5.742826461791992, 'timing/data': 0.0005309719708748162, 'timing/model': 1.4312020229990594, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 0.22}) +[2026-06-15 18:48:18,251][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.2708662152290344, 'train/grad_norm_pre_clip': 9.746490478515625, 'timing/data': 0.000601025007199496, 'timing/model': 1.4754544550087303, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 0.22}) +[2026-06-15 18:48:24,196][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.11189154535531998, 'train/grad_norm_pre_clip': 4.420711040496826, 'timing/data': 0.0005153879756107926, 'timing/model': 1.7031583079951815, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 0.22}) +[2026-06-15 18:48:29,958][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.2707229554653168, 'train/grad_norm_pre_clip': 9.403581619262695, 'timing/data': 0.0005904189893044531, 'timing/model': 1.4632162089692429, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 0.22}) +[2026-06-15 18:48:35,859][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.39741724729537964, 'train/grad_norm_pre_clip': 16.5506591796875, 'timing/data': 0.0005255939904600382, 'timing/model': 1.4689183110021986, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 0.22}) +[2026-06-15 18:48:41,857][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.6956472992897034, 'train/grad_norm_pre_clip': 34.37263870239258, 'timing/data': 0.0005386329721659422, 'timing/model': 1.455191834946163, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 0.22}) +[2026-06-15 18:48:47,936][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.25926128029823303, 'train/grad_norm_pre_clip': 11.474034309387207, 'timing/data': 0.0006444089813157916, 'timing/model': 1.5113436019746587, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 0.22}) +[2026-06-15 18:48:53,857][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.4048463702201843, 'train/grad_norm_pre_clip': 16.913211822509766, 'timing/data': 0.0005297499592415988, 'timing/model': 1.4610223569907248, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 0.22}) +[2026-06-15 18:49:00,072][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.11590167135000229, 'train/grad_norm_pre_clip': 3.670745372772217, 'timing/data': 0.0005992419901303947, 'timing/model': 1.7689715120359324, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 0.22}) +[2026-06-15 18:49:06,121][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.1953534185886383, 'train/grad_norm_pre_clip': 6.592992782592773, 'timing/data': 0.0005706800147891045, 'timing/model': 1.5541881970129907, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.22}) +[2026-06-15 18:49:12,236][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.19056817889213562, 'train/grad_norm_pre_clip': 11.890558242797852, 'timing/data': 0.0005608649807982147, 'timing/model': 1.5299474040511996, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 0.22}) +[2026-06-15 18:49:18,537][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.3027401268482208, 'train/grad_norm_pre_clip': 7.359457969665527, 'timing/data': 0.0007740400033071637, 'timing/model': 1.5395339939859696, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 0.22}) +[2026-06-15 18:49:24,559][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.5512816905975342, 'train/grad_norm_pre_clip': 10.95800495147705, 'timing/data': 0.0005299000185914338, 'timing/model': 1.4663074649870396, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 0.23}) +[2026-06-15 18:49:30,661][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.5516865253448486, 'train/grad_norm_pre_clip': 15.23140811920166, 'timing/data': 0.0005194050027057528, 'timing/model': 1.5128194999997504, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 0.23}) +[2026-06-15 18:49:36,712][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.5322534441947937, 'train/grad_norm_pre_clip': 13.990472793579102, 'timing/data': 0.0005219079903326929, 'timing/model': 1.7644369239569642, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 0.23}) +[2026-06-15 18:49:42,357][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.42364826798439026, 'train/grad_norm_pre_clip': 11.601703643798828, 'timing/data': 0.000522670045029372, 'timing/model': 1.462766945012845, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 0.23}) +[2026-06-15 18:49:48,273][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.2829773724079132, 'train/grad_norm_pre_clip': 8.675633430480957, 'timing/data': 0.0005224589840508997, 'timing/model': 1.4786137819755822, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 0.23}) +[2026-06-15 18:49:54,176][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.12074302136898041, 'train/grad_norm_pre_clip': 3.144054651260376, 'timing/data': 0.0005326440441422164, 'timing/model': 1.4825068230275065, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 0.23}) +[2026-06-15 18:50:00,157][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.22539769113063812, 'train/grad_norm_pre_clip': 9.185832023620605, 'timing/data': 0.0005338860210031271, 'timing/model': 1.4856341170379892, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 0.23}) +[2026-06-15 18:50:06,102][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.3537503480911255, 'train/grad_norm_pre_clip': 12.84663200378418, 'timing/data': 0.0005305709782987833, 'timing/model': 1.474770106957294, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 0.23}) +[2026-06-15 18:50:12,011][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.2567859888076782, 'train/grad_norm_pre_clip': 6.581827163696289, 'timing/data': 0.000531331985257566, 'timing/model': 1.7166040820302442, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 0.23}) +[2026-06-15 18:50:17,779][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.41476523876190186, 'train/grad_norm_pre_clip': 12.271411895751953, 'timing/data': 0.0005290789995342493, 'timing/model': 1.4657130180276, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 0.23}) +[2026-06-15 18:50:23,773][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.157518208026886, 'train/grad_norm_pre_clip': 5.417212009429932, 'timing/data': 0.0008154900278896093, 'timing/model': 1.4695853280136362, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 0.23}) +[2026-06-15 18:50:29,835][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.31344619393348694, 'train/grad_norm_pre_clip': 10.984142303466797, 'timing/data': 0.000543220026884228, 'timing/model': 1.5115503589622676, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 0.23}) +[2026-06-15 18:50:35,859][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.15495362877845764, 'train/grad_norm_pre_clip': 5.0648369789123535, 'timing/data': 0.0005235309945419431, 'timing/model': 1.4464476350112818, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 0.23}) +[2026-06-15 18:50:41,956][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.1871616244316101, 'train/grad_norm_pre_clip': 8.811775207519531, 'timing/data': 0.0005274760187603533, 'timing/model': 1.5031133230077103, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 0.23}) +[2026-06-15 18:50:47,889][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.13048440217971802, 'train/grad_norm_pre_clip': 6.632662773132324, 'timing/data': 0.0005103209987282753, 'timing/model': 1.699683246028144, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 0.23}) +[2026-06-15 18:50:53,657][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.3284544348716736, 'train/grad_norm_pre_clip': 14.360599517822266, 'timing/data': 0.0005193039542064071, 'timing/model': 1.4797831759788096, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 0.23}) +[2026-06-15 18:50:59,555][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.52430260181427, 'train/grad_norm_pre_clip': 22.387178421020508, 'timing/data': 0.0005267149535939097, 'timing/model': 1.4624230060144328, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 0.23}) +[2026-06-15 18:51:05,555][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.24714872241020203, 'train/grad_norm_pre_clip': 5.401111125946045, 'timing/data': 0.0005193150136619806, 'timing/model': 1.4644303829991259, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 0.23}) +[2026-06-15 18:51:11,557][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.09079547971487045, 'train/grad_norm_pre_clip': 3.462002992630005, 'timing/data': 0.0005322539946064353, 'timing/model': 1.4474102429812774, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.23}) +[2026-06-15 18:51:17,460][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.2730214297771454, 'train/grad_norm_pre_clip': 9.807943344116211, 'timing/data': 0.0005296799936331809, 'timing/model': 1.4382572729955427, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 0.23}) +[2026-06-15 18:51:23,495][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.15822117030620575, 'train/grad_norm_pre_clip': 2.9117367267608643, 'timing/data': 0.0006121810292825103, 'timing/model': 1.6915954370051622, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 0.23}) +[2026-06-15 18:51:29,412][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.1059773862361908, 'train/grad_norm_pre_clip': 3.9801063537597656, 'timing/data': 0.00053439597832039, 'timing/model': 1.6173747549764812, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 0.23}) +[2026-06-15 18:51:35,157][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.1904090791940689, 'train/grad_norm_pre_clip': 6.579398155212402, 'timing/data': 0.0005154790123924613, 'timing/model': 1.4998191219638102, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 0.23}) +[2026-06-15 18:51:41,259][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.29868176579475403, 'train/grad_norm_pre_clip': 9.706780433654785, 'timing/data': 0.0005324939847923815, 'timing/model': 1.4630574469920248, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 0.23}) +[2026-06-15 18:51:47,260][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.454245388507843, 'train/grad_norm_pre_clip': 17.714628219604492, 'timing/data': 0.0005550669739022851, 'timing/model': 1.4644854000071064, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 0.24}) +[2026-06-15 18:51:53,361][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.3946448266506195, 'train/grad_norm_pre_clip': 21.39072036743164, 'timing/data': 0.00055906199850142, 'timing/model': 1.5116626499802805, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 0.24}) +[2026-06-15 18:51:59,531][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.3481844961643219, 'train/grad_norm_pre_clip': 10.713083267211914, 'timing/data': 0.0005786319961771369, 'timing/model': 1.830949772032909, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 0.24}) +[2026-06-15 18:52:05,716][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.11993106454610825, 'train/grad_norm_pre_clip': 6.311375141143799, 'timing/data': 0.0005229400121606886, 'timing/model': 1.798733155010268, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 0.24}) +[2026-06-15 18:52:11,476][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.16447234153747559, 'train/grad_norm_pre_clip': 9.916411399841309, 'timing/data': 0.0005458130035549402, 'timing/model': 1.457602962036617, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 0.24}) +[2026-06-15 18:52:17,658][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.1577770709991455, 'train/grad_norm_pre_clip': 4.898775577545166, 'timing/data': 0.0005518220132216811, 'timing/model': 1.5480147950001992, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 0.24}) +[2026-06-15 18:52:23,876][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.20201244950294495, 'train/grad_norm_pre_clip': 6.110115051269531, 'timing/data': 0.0005345770041458309, 'timing/model': 1.5603008099715225, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 0.24}) +[2026-06-15 18:52:30,061][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.20773422718048096, 'train/grad_norm_pre_clip': 7.62800931930542, 'timing/data': 0.0005340459756553173, 'timing/model': 1.561315105005633, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 0.24}) +[2026-06-15 18:52:36,428][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.152090921998024, 'train/grad_norm_pre_clip': 4.564596176147461, 'timing/data': 0.000545573013368994, 'timing/model': 1.882179518986959, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.24}) +[2026-06-15 18:52:42,406][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.13098953664302826, 'train/grad_norm_pre_clip': 6.486236572265625, 'timing/data': 0.0005329140112735331, 'timing/model': 1.503949008998461, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 0.24}) +[2026-06-15 18:52:48,688][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.09967659413814545, 'train/grad_norm_pre_clip': 4.589882850646973, 'timing/data': 0.0007329589570872486, 'timing/model': 1.4755063500488177, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 0.24}) +[2026-06-15 18:52:54,795][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.23491278290748596, 'train/grad_norm_pre_clip': 6.484775066375732, 'timing/data': 0.000746968958992511, 'timing/model': 1.5161500329850242, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 0.24}) +[2026-06-15 18:53:01,172][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.22225935757160187, 'train/grad_norm_pre_clip': 5.952833652496338, 'timing/data': 0.0009482059977017343, 'timing/model': 1.620819317002315, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 0.24}) +[2026-06-15 18:53:07,467][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.26926010847091675, 'train/grad_norm_pre_clip': 12.009941101074219, 'timing/data': 0.0005576710100285709, 'timing/model': 1.6171060840133578, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 0.24}) +[2026-06-15 18:53:13,680][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.24483290314674377, 'train/grad_norm_pre_clip': 7.1504716873168945, 'timing/data': 0.000589778006542474, 'timing/model': 1.77098297100747, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 0.24}) +[2026-06-15 18:53:19,486][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.18569070100784302, 'train/grad_norm_pre_clip': 11.788887977600098, 'timing/data': 0.000529509037733078, 'timing/model': 1.4965717610321008, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 0.24}) +[2026-06-15 18:53:25,780][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.09908643364906311, 'train/grad_norm_pre_clip': 5.761070728302002, 'timing/data': 0.0008732149726711214, 'timing/model': 1.5512343249865808, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 0.24}) +[2026-06-15 18:53:31,996][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.23987917602062225, 'train/grad_norm_pre_clip': 15.4793062210083, 'timing/data': 0.000931070011574775, 'timing/model': 1.4546665929956362, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 0.24}) +[2026-06-15 18:53:38,331][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.23813055455684662, 'train/grad_norm_pre_clip': 8.942785263061523, 'timing/data': 0.0012936739949509501, 'timing/model': 1.5968133629648946, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 0.24}) +[2026-06-15 18:53:44,584][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.14347673952579498, 'train/grad_norm_pre_clip': 10.797744750976562, 'timing/data': 0.000535928993485868, 'timing/model': 1.489335885969922, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 0.24}) +[2026-06-15 18:53:50,750][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.2940954566001892, 'train/grad_norm_pre_clip': 9.743387222290039, 'timing/data': 0.0005259040044620633, 'timing/model': 1.7440137030207552, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 0.24}) +[2026-06-15 18:53:56,782][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.09700153023004532, 'train/grad_norm_pre_clip': 5.325364112854004, 'timing/data': 0.0006067740032449365, 'timing/model': 1.5754778429982252, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 0.24}) +[2026-06-15 18:54:02,925][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.13206487894058228, 'train/grad_norm_pre_clip': 8.251532554626465, 'timing/data': 0.0008314939914271235, 'timing/model': 1.4818425989942625, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 0.24}) +[2026-06-15 18:54:09,075][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.3730800151824951, 'train/grad_norm_pre_clip': 11.374744415283203, 'timing/data': 0.0005264039500616491, 'timing/model': 1.5712457720073871, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 0.25}) +[2026-06-15 18:54:15,269][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.3087484836578369, 'train/grad_norm_pre_clip': 11.419913291931152, 'timing/data': 0.0005151689983904362, 'timing/model': 1.5557393439812586, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 0.25}) +[2026-06-15 18:54:21,376][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.14741475880146027, 'train/grad_norm_pre_clip': 13.972905158996582, 'timing/data': 0.0005342260119505227, 'timing/model': 1.7715468109818175, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 0.25}) +[2026-06-15 18:54:27,165][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.10578937083482742, 'train/grad_norm_pre_clip': 6.396186828613281, 'timing/data': 0.0007738790009170771, 'timing/model': 1.4727107270155102, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 0.25}) +[2026-06-15 18:54:33,368][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.3594666123390198, 'train/grad_norm_pre_clip': 10.073654174804688, 'timing/data': 0.0006438270211219788, 'timing/model': 1.5193173029692844, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 0.25}) +[2026-06-15 18:54:39,400][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.2034367322921753, 'train/grad_norm_pre_clip': 10.56064510345459, 'timing/data': 0.0005263449857011437, 'timing/model': 1.504607999988366, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 0.25}) +[2026-06-15 18:54:45,565][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.5207766890525818, 'train/grad_norm_pre_clip': 15.64610767364502, 'timing/data': 0.0007803780026733875, 'timing/model': 1.5232698600157164, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 0.25}) +[2026-06-15 18:54:51,784][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.1797213852405548, 'train/grad_norm_pre_clip': 5.243558883666992, 'timing/data': 0.0005227089859545231, 'timing/model': 1.4965308819664642, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 0.25}) +[2026-06-15 18:54:57,918][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.16092748939990997, 'train/grad_norm_pre_clip': 124.8389663696289, 'timing/data': 0.0005365600227378309, 'timing/model': 1.8034218630054966, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 0.25}) +[2026-06-15 18:55:03,768][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.18187226355075836, 'train/grad_norm_pre_clip': 9.404285430908203, 'timing/data': 0.0005368099664337933, 'timing/model': 1.4382209869800135, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 0.25}) +[2026-06-15 18:55:09,924][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.22901009023189545, 'train/grad_norm_pre_clip': 8.815191268920898, 'timing/data': 0.0008717019809409976, 'timing/model': 1.4753734879777767, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 0.25}) +[2026-06-15 18:55:16,100][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.27158311009407043, 'train/grad_norm_pre_clip': 11.558918952941895, 'timing/data': 0.0005344059900380671, 'timing/model': 1.5261209990130737, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 0.25}) +[2026-06-15 18:55:22,282][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.22486305236816406, 'train/grad_norm_pre_clip': 5.774979591369629, 'timing/data': 0.0007101949886418879, 'timing/model': 1.524132548016496, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 0.25}) +[2026-06-15 18:55:28,526][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.14885960519313812, 'train/grad_norm_pre_clip': 5.32034969329834, 'timing/data': 0.0012197350151836872, 'timing/model': 1.4907359019853175, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 0.25}) +[2026-06-15 18:55:34,760][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.08409196883440018, 'train/grad_norm_pre_clip': 2.642965793609619, 'timing/data': 0.000545492977835238, 'timing/model': 1.8418832590105012, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 0.25}) +[2026-06-15 18:55:40,776][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.3799314498901367, 'train/grad_norm_pre_clip': 10.511392593383789, 'timing/data': 0.0007683309959247708, 'timing/model': 1.5306134800193831, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 0.25}) +[2026-06-15 18:55:47,188][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.2599545419216156, 'train/grad_norm_pre_clip': 16.213077545166016, 'timing/data': 0.0008118749829009175, 'timing/model': 1.5434540919959545, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 0.25}) +[2026-06-15 18:55:53,354][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.1632050722837448, 'train/grad_norm_pre_clip': 6.350218772888184, 'timing/data': 0.0007412100094370544, 'timing/model': 1.5256506060250103, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 0.25}) +[2026-06-15 18:55:59,469][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.19440798461437225, 'train/grad_norm_pre_clip': 8.335334777832031, 'timing/data': 0.0009098690352402627, 'timing/model': 1.533422649023123, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 0.25}) +[2026-06-15 18:56:05,737][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.15380050241947174, 'train/grad_norm_pre_clip': 6.393424987792969, 'timing/data': 0.0005150579963810742, 'timing/model': 1.527349065989256, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 0.25}) +[2026-06-15 18:56:11,879][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.1343046873807907, 'train/grad_norm_pre_clip': 10.67108154296875, 'timing/data': 0.0005940140108577907, 'timing/model': 1.7801823589834385, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.25}) +[2026-06-15 18:56:17,688][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.11674146354198456, 'train/grad_norm_pre_clip': 5.266050815582275, 'timing/data': 0.0005238110315985978, 'timing/model': 1.4888190429774113, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 0.25}) +[2026-06-15 18:56:23,836][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.14659932255744934, 'train/grad_norm_pre_clip': 12.535045623779297, 'timing/data': 0.0006089669768698514, 'timing/model': 1.4842297449940816, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 0.25}) +[2026-06-15 18:56:29,964][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.3602968454360962, 'train/grad_norm_pre_clip': 22.815622329711914, 'timing/data': 0.0006428859778679907, 'timing/model': 1.4557984889834188, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 0.26}) +[2026-06-15 18:56:35,967][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.0819714143872261, 'train/grad_norm_pre_clip': 6.493619441986084, 'timing/data': 0.0005214670090936124, 'timing/model': 1.4583137960289605, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 0.26}) +[2026-06-15 18:56:42,092][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.549225926399231, 'train/grad_norm_pre_clip': 27.66193389892578, 'timing/data': 0.0008778920164331794, 'timing/model': 1.5595580899971537, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 0.26}) +[2026-06-15 18:56:48,187][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.5583018064498901, 'train/grad_norm_pre_clip': 13.662806510925293, 'timing/data': 0.0010939600178971887, 'timing/model': 1.7559323610039428, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 0.26}) +[2026-06-15 18:56:54,005][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.22625049948692322, 'train/grad_norm_pre_clip': 8.016304016113281, 'timing/data': 0.0005181719898246229, 'timing/model': 1.5321912160143256, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 0.26}) +[2026-06-15 18:57:00,292][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.20118403434753418, 'train/grad_norm_pre_clip': 10.234664916992188, 'timing/data': 0.000945231004152447, 'timing/model': 1.5704902150318958, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 0.26}) +[2026-06-15 18:57:06,441][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.19473476707935333, 'train/grad_norm_pre_clip': 10.412915229797363, 'timing/data': 0.0006998099852353334, 'timing/model': 1.5472929920069873, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 0.26}) +[2026-06-15 19:05:08,458][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.08952536433935165, 'train/grad_norm_pre_clip': 4.202521324157715, 'eval/loss': 0.1674201488494873, 'eval/flap_precision': 0.6995645863570392, 'eval/flap_recall': 0.9659318637274549, 'eval/flap_f1': 0.8114478114478114, 'eval/flap_rate_model': 0.1913888888888889, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.8114478114478114, 'timing/data': 0.0007606289582327008, 'timing/model': 1.467492459982168, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 0.26}) +[2026-06-15 19:05:08,459][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_600_state +[2026-06-15 19:05:08,464][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 19:05:14,298][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_600_state/model.safetensors +[2026-06-15 19:05:27,982][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_600_state/optimizer.bin +[2026-06-15 19:05:27,983][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_600_state/sampler.bin +[2026-06-15 19:05:27,983][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_600_state/sampler_1.bin +[2026-06-15 19:05:27,985][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_600_state/random_states_0.pkl +[2026-06-15 19:05:28,099][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600 +[2026-06-15 19:05:28,100][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 19:05:28,110][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 19:05:34,149][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.2166091501712799, 'train/grad_norm_pre_clip': 11.661829948425293, 'timing/data': 0.0006221229559741914, 'timing/model': 1.5721440779743716, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 0.26}) +[2026-06-15 19:05:40,253][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.4265964925289154, 'train/grad_norm_pre_clip': 15.093677520751953, 'timing/data': 0.0006686800043098629, 'timing/model': 1.4875638980302028, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 0.26}) +[2026-06-15 19:05:46,334][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.5015653371810913, 'train/grad_norm_pre_clip': 13.421253204345703, 'timing/data': 0.0006632519653066993, 'timing/model': 1.5526310759596527, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 0.26}) +[2026-06-15 19:05:52,500][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.13680773973464966, 'train/grad_norm_pre_clip': 6.1360039710998535, 'timing/data': 0.0005578880081884563, 'timing/model': 1.7189779990003444, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 0.26}) +[2026-06-15 19:05:58,252][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.14455370604991913, 'train/grad_norm_pre_clip': 8.287705421447754, 'timing/data': 0.0005882229888811707, 'timing/model': 1.465691510005854, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 0.26}) +[2026-06-15 19:06:04,252][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.26079899072647095, 'train/grad_norm_pre_clip': 9.359878540039062, 'timing/data': 0.000544487964361906, 'timing/model': 1.4645943870418705, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 0.26}) +[2026-06-15 19:06:10,256][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.13444675505161285, 'train/grad_norm_pre_clip': 4.413346767425537, 'timing/data': 0.0005364359822124243, 'timing/model': 1.467838728975039, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 0.26}) +[2026-06-15 19:06:16,170][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.18094167113304138, 'train/grad_norm_pre_clip': 7.025485515594482, 'timing/data': 0.0005475730285979807, 'timing/model': 1.442494033020921, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 0.26}) +[2026-06-15 19:06:22,086][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.08271825313568115, 'train/grad_norm_pre_clip': 4.685123920440674, 'timing/data': 0.0005847970023751259, 'timing/model': 1.695127355051227, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 0.26}) +[2026-06-15 19:06:27,945][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.15029583871364594, 'train/grad_norm_pre_clip': 8.250981330871582, 'timing/data': 0.000570046016946435, 'timing/model': 1.649001738987863, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 0.26}) +[2026-06-15 19:06:33,758][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.36109817028045654, 'train/grad_norm_pre_clip': 9.431844711303711, 'timing/data': 0.0005272430134937167, 'timing/model': 1.4768091919831932, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 0.26}) +[2026-06-15 19:06:39,759][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.21711280941963196, 'train/grad_norm_pre_clip': 7.532894611358643, 'timing/data': 0.0005260609905235469, 'timing/model': 1.4634605559986085, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 0.26}) +[2026-06-15 19:06:45,760][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.2970274090766907, 'train/grad_norm_pre_clip': 11.561627388000488, 'timing/data': 0.0005606420454569161, 'timing/model': 1.457366466987878, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 0.26}) +[2026-06-15 19:06:51,833][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.11454208940267563, 'train/grad_norm_pre_clip': 5.510223388671875, 'timing/data': 0.0005252999835647643, 'timing/model': 1.5080734920338728, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 0.26}) +[2026-06-15 19:06:57,824][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.22168564796447754, 'train/grad_norm_pre_clip': 12.791400909423828, 'timing/data': 0.0005298369796946645, 'timing/model': 1.4955908230040222, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 0.26}) +[2026-06-15 19:07:03,752][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.09057798981666565, 'train/grad_norm_pre_clip': 4.73565673828125, 'timing/data': 0.0005269829998724163, 'timing/model': 1.6550500050070696, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 0.27}) +[2026-06-15 19:07:09,459][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.1751963347196579, 'train/grad_norm_pre_clip': 5.858436584472656, 'timing/data': 0.0005243690102361143, 'timing/model': 1.4632918409770355, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 0.27}) +[2026-06-15 19:07:15,461][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.2079790234565735, 'train/grad_norm_pre_clip': 5.011987209320068, 'timing/data': 0.0005346740363165736, 'timing/model': 1.4642251470359042, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 0.27}) +[2026-06-15 19:07:21,461][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.12664636969566345, 'train/grad_norm_pre_clip': 5.015481472015381, 'timing/data': 0.0005216749850660563, 'timing/model': 1.473547078028787, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 0.27}) +[2026-06-15 19:07:27,442][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.18619370460510254, 'train/grad_norm_pre_clip': 7.695035934448242, 'timing/data': 0.0005152560188435018, 'timing/model': 1.5095489980303682, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 0.27}) +[2026-06-15 19:07:33,398][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.12736910581588745, 'train/grad_norm_pre_clip': 4.169371128082275, 'timing/data': 0.0005184909678064287, 'timing/model': 1.693773327046074, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 0.27}) +[2026-06-15 19:07:39,062][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.05865102261304855, 'train/grad_norm_pre_clip': 4.011172294616699, 'timing/data': 0.0005212250398471951, 'timing/model': 1.4634391909930855, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 0.27}) +[2026-06-15 19:07:45,060][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.2994603216648102, 'train/grad_norm_pre_clip': 14.997320175170898, 'timing/data': 0.0005302170175127685, 'timing/model': 1.4603936250205152, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 0.27}) +[2026-06-15 19:07:51,077][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.6726669669151306, 'train/grad_norm_pre_clip': 22.41623878479004, 'timing/data': 0.0005289060063660145, 'timing/model': 1.4856473899562843, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 0.27}) +[2026-06-15 19:07:56,963][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.40711739659309387, 'train/grad_norm_pre_clip': 15.505852699279785, 'timing/data': 0.0005269030225463212, 'timing/model': 1.4663025100016966, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.27}) +[2026-06-15 19:08:02,864][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.15633828938007355, 'train/grad_norm_pre_clip': 11.470479011535645, 'timing/data': 0.0005452600307762623, 'timing/model': 1.4885475789778866, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 0.27}) +[2026-06-15 19:08:08,897][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.1565050631761551, 'train/grad_norm_pre_clip': 3.7708301544189453, 'timing/data': 0.0005408829892985523, 'timing/model': 1.6941203370224684, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 0.27}) +[2026-06-15 19:08:14,666][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.18276208639144897, 'train/grad_norm_pre_clip': 3.779026985168457, 'timing/data': 0.0008130730129778385, 'timing/model': 1.4631546439486556, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 0.27}) +[2026-06-15 19:08:20,847][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.2892056107521057, 'train/grad_norm_pre_clip': 12.027405738830566, 'timing/data': 0.0005326510290615261, 'timing/model': 1.5614616110105999, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 0.27}) +[2026-06-15 19:08:27,061][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.4396395683288574, 'train/grad_norm_pre_clip': 8.209883689880371, 'timing/data': 0.0005418150103650987, 'timing/model': 1.546647558978293, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 0.27}) +[2026-06-15 19:08:33,065][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.14279571175575256, 'train/grad_norm_pre_clip': 6.645758152008057, 'timing/data': 0.000519382010679692, 'timing/model': 1.483613324991893, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 0.27}) +[2026-06-15 19:08:38,964][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.1804240494966507, 'train/grad_norm_pre_clip': 8.110663414001465, 'timing/data': 0.0006484519690275192, 'timing/model': 1.4651376419933513, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 0.27}) +[2026-06-15 19:08:45,028][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.14506517350673676, 'train/grad_norm_pre_clip': 3.7956676483154297, 'timing/data': 0.0005259619792923331, 'timing/model': 1.7778327550040558, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 0.27}) +[2026-06-15 19:08:50,865][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.34113970398902893, 'train/grad_norm_pre_clip': 5.336684226989746, 'timing/data': 0.0005338640185073018, 'timing/model': 1.4666725089773536, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 0.27}) +[2026-06-15 19:08:56,865][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.21666476130485535, 'train/grad_norm_pre_clip': 7.674048900604248, 'timing/data': 0.0005438579828478396, 'timing/model': 1.4440801609889604, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 0.27}) +[2026-06-15 19:09:02,964][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.20761752128601074, 'train/grad_norm_pre_clip': 7.446877479553223, 'timing/data': 0.00053899094928056, 'timing/model': 1.4888387820101343, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 0.27}) +[2026-06-15 19:09:09,240][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.20528948307037354, 'train/grad_norm_pre_clip': 8.14453411102295, 'timing/data': 0.0005514299846254289, 'timing/model': 1.565235178975854, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 0.27}) +[2026-06-15 19:09:15,639][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.20820656418800354, 'train/grad_norm_pre_clip': 4.732523441314697, 'timing/data': 0.0005325020174495876, 'timing/model': 1.5565757820149884, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 0.27}) +[2026-06-15 19:09:21,715][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.31502917408943176, 'train/grad_norm_pre_clip': 5.251490116119385, 'timing/data': 0.0005430770106613636, 'timing/model': 1.7143837759504095, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.27}) +[2026-06-15 19:09:27,468][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.15185542404651642, 'train/grad_norm_pre_clip': 5.532016754150391, 'timing/data': 0.000594070996157825, 'timing/model': 1.4667198059614748, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 0.28}) +[2026-06-15 19:09:33,564][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.16349782049655914, 'train/grad_norm_pre_clip': 5.240922927856445, 'timing/data': 0.0005469730240292847, 'timing/model': 1.4631945560104214, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 0.28}) +[2026-06-15 19:09:39,564][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.29178667068481445, 'train/grad_norm_pre_clip': 4.39312744140625, 'timing/data': 0.0005300279590301216, 'timing/model': 1.45898934698198, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 0.28}) +[2026-06-15 19:09:45,573][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.14561307430267334, 'train/grad_norm_pre_clip': 7.066459655761719, 'timing/data': 0.0005341239739209414, 'timing/model': 1.499212467984762, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 0.28}) +[2026-06-15 19:09:51,760][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.47109273076057434, 'train/grad_norm_pre_clip': 15.24574089050293, 'timing/data': 0.0005254909629002213, 'timing/model': 1.4732190989889205, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 0.28}) +[2026-06-15 19:09:57,713][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.16452839970588684, 'train/grad_norm_pre_clip': 5.484654426574707, 'timing/data': 0.0005284660146571696, 'timing/model': 1.712661047000438, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 0.28}) +[2026-06-15 19:10:03,463][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.16760164499282837, 'train/grad_norm_pre_clip': 6.634957790374756, 'timing/data': 0.0006680010119453073, 'timing/model': 1.464705013029743, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 0.28}) +[2026-06-15 19:10:09,561][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.24881403148174286, 'train/grad_norm_pre_clip': 6.08549690246582, 'timing/data': 0.0005010060267522931, 'timing/model': 1.4913738510222174, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 0.28}) +[2026-06-15 19:10:15,568][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.0911550372838974, 'train/grad_norm_pre_clip': 3.3456594944000244, 'timing/data': 0.0004895080346614122, 'timing/model': 1.446392312005628, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 0.28}) +[2026-06-15 19:10:21,563][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.14586074650287628, 'train/grad_norm_pre_clip': 6.560056686401367, 'timing/data': 0.00051496597006917, 'timing/model': 1.465156416001264, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 0.28}) +[2026-06-15 19:10:27,564][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.13908201456069946, 'train/grad_norm_pre_clip': 3.7356557846069336, 'timing/data': 0.0005086060264147818, 'timing/model': 1.4635173890274018, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 0.28}) +[2026-06-15 19:10:33,679][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.1325749307870865, 'train/grad_norm_pre_clip': 6.444411754608154, 'timing/data': 0.0005188009818084538, 'timing/model': 1.740616263996344, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 0.28}) +[2026-06-15 19:10:39,460][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.2173822820186615, 'train/grad_norm_pre_clip': 5.106549263000488, 'timing/data': 0.0005178800201974809, 'timing/model': 1.5006375989760272, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 0.28}) +[2026-06-15 19:10:45,560][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.13259367644786835, 'train/grad_norm_pre_clip': 7.805902004241943, 'timing/data': 0.000601884035859257, 'timing/model': 1.5127881349762902, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 0.28}) +[2026-06-15 19:10:51,562][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.06836046278476715, 'train/grad_norm_pre_clip': 7.241716384887695, 'timing/data': 0.0005963450530543923, 'timing/model': 1.4636454689898528, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 0.28}) +[2026-06-15 19:10:57,506][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.14657698571681976, 'train/grad_norm_pre_clip': 3.5306029319763184, 'timing/data': 0.0005888449959456921, 'timing/model': 1.4999333099694923, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 0.28}) +[2026-06-15 19:11:03,361][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.11423642188310623, 'train/grad_norm_pre_clip': 14.874011993408203, 'timing/data': 0.0005942229763604701, 'timing/model': 1.4850853969692253, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 0.28}) +[2026-06-15 19:11:09,459][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.13152866065502167, 'train/grad_norm_pre_clip': 6.994372844696045, 'timing/data': 0.0005293069989420474, 'timing/model': 1.762117243022658, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 0.28}) +[2026-06-15 19:11:15,160][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.24603630602359772, 'train/grad_norm_pre_clip': 6.608251571655273, 'timing/data': 0.000516418949700892, 'timing/model': 1.464721461001318, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 0.28}) +[2026-06-15 19:11:21,167][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.1420472264289856, 'train/grad_norm_pre_clip': 8.827028274536133, 'timing/data': 0.0005252709961496294, 'timing/model': 1.4708399000228383, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 0.28}) +[2026-06-15 19:11:27,164][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.3578997254371643, 'train/grad_norm_pre_clip': 9.931268692016602, 'timing/data': 0.0007044239901006222, 'timing/model': 1.444714367040433, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.28}) +[2026-06-15 19:11:33,171][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.043869033455848694, 'train/grad_norm_pre_clip': 6.840756893157959, 'timing/data': 0.0005385109689086676, 'timing/model': 1.445282926026266, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 0.28}) +[2026-06-15 19:11:39,160][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.12512145936489105, 'train/grad_norm_pre_clip': 4.4407877922058105, 'timing/data': 0.000650756002869457, 'timing/model': 1.5017651789821684, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 0.28}) +[2026-06-15 19:11:45,093][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.15045568346977234, 'train/grad_norm_pre_clip': 8.693232536315918, 'timing/data': 0.0005922300042584538, 'timing/model': 1.6962397249881178, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 0.29}) +[2026-06-15 19:11:50,861][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.1495596021413803, 'train/grad_norm_pre_clip': 5.7811479568481445, 'timing/data': 0.0005218959995545447, 'timing/model': 1.4674830500152893, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 0.29}) +[2026-06-15 19:11:56,923][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.33614659309387207, 'train/grad_norm_pre_clip': 12.18111801147461, 'timing/data': 0.0005586699699051678, 'timing/model': 1.5331072499975562, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 0.29}) +[2026-06-15 19:12:02,856][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.26720547676086426, 'train/grad_norm_pre_clip': 43.89712142944336, 'timing/data': 0.0004857840249314904, 'timing/model': 1.4622751079732552, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 0.29}) +[2026-06-15 19:12:08,759][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.2427881509065628, 'train/grad_norm_pre_clip': 7.2187066078186035, 'timing/data': 0.0005570980138145387, 'timing/model': 1.4642235110513866, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 0.29}) +[2026-06-15 19:12:14,665][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.22697630524635315, 'train/grad_norm_pre_clip': 5.739564895629883, 'timing/data': 0.0005165889742784202, 'timing/model': 1.4374839380034246, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 0.29}) +[2026-06-15 19:12:20,719][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.18570983409881592, 'train/grad_norm_pre_clip': 10.749007225036621, 'timing/data': 0.0004851620178669691, 'timing/model': 1.696563542005606, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 0.29}) +[2026-06-15 19:12:26,560][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.16976980865001678, 'train/grad_norm_pre_clip': 8.173555374145508, 'timing/data': 0.0006177370087243617, 'timing/model': 1.4635317689971998, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 0.29}) +[2026-06-15 19:12:32,559][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.2133248895406723, 'train/grad_norm_pre_clip': 6.755066394805908, 'timing/data': 0.0005044009885750711, 'timing/model': 1.4660690349992365, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 0.29}) +[2026-06-15 19:12:38,357][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.189858078956604, 'train/grad_norm_pre_clip': 2.777730941772461, 'timing/data': 0.000541234971024096, 'timing/model': 1.4649604359874502, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 0.29}) +[2026-06-15 19:12:44,262][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.11554858088493347, 'train/grad_norm_pre_clip': 4.289089679718018, 'timing/data': 0.000512954022269696, 'timing/model': 1.4666726600262336, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 0.29}) +[2026-06-15 19:12:50,139][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.17250388860702515, 'train/grad_norm_pre_clip': 4.606449127197266, 'timing/data': 0.0005133129889145494, 'timing/model': 1.5122948749922216, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 0.29}) +[2026-06-15 19:12:56,228][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.12425270676612854, 'train/grad_norm_pre_clip': 3.8498828411102295, 'timing/data': 0.0005179099971428514, 'timing/model': 1.631610714015551, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 0.29}) +[2026-06-15 19:13:02,062][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.1720293015241623, 'train/grad_norm_pre_clip': 4.466509819030762, 'timing/data': 0.0005050119943916798, 'timing/model': 1.5130326789803803, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 0.29}) +[2026-06-15 19:13:08,062][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.12192554771900177, 'train/grad_norm_pre_clip': 10.228453636169434, 'timing/data': 0.0005188109935261309, 'timing/model': 1.4664907670230605, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 0.29}) +[2026-06-15 19:13:14,167][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.2790387272834778, 'train/grad_norm_pre_clip': 29.194644927978516, 'timing/data': 0.0005105390446260571, 'timing/model': 1.473955538996961, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 0.29}) +[2026-06-15 19:13:20,155][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.21660202741622925, 'train/grad_norm_pre_clip': 6.672609329223633, 'timing/data': 0.0005134940147399902, 'timing/model': 1.515686254017055, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 0.29}) +[2026-06-15 19:13:26,195][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.29221048951148987, 'train/grad_norm_pre_clip': 8.991905212402344, 'timing/data': 0.0005099990521557629, 'timing/model': 1.6980729890055954, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 0.29}) +[2026-06-15 19:13:32,145][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.2656550407409668, 'train/grad_norm_pre_clip': 7.58045768737793, 'timing/data': 0.0005359870265237987, 'timing/model': 1.67383908998454, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 0.29}) +[2026-06-15 19:13:37,859][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.29040682315826416, 'train/grad_norm_pre_clip': 10.104290008544922, 'timing/data': 0.0005207740468904376, 'timing/model': 1.4637772900168784, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 0.29}) +[2026-06-15 19:13:43,865][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.43467384576797485, 'train/grad_norm_pre_clip': 15.805578231811523, 'timing/data': 0.0005338040064089, 'timing/model': 1.462115166010335, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 0.29}) +[2026-06-15 19:13:49,856][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.304095596075058, 'train/grad_norm_pre_clip': 10.066729545593262, 'timing/data': 0.0005148359923623502, 'timing/model': 1.4761676110210828, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 0.29}) +[2026-06-15 19:13:55,843][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.32495763897895813, 'train/grad_norm_pre_clip': 8.289261817932129, 'timing/data': 0.0005095379892736673, 'timing/model': 1.5043974689906463, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 0.29}) +[2026-06-15 19:14:01,945][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.20381951332092285, 'train/grad_norm_pre_clip': 9.291945457458496, 'timing/data': 0.0005157279665581882, 'timing/model': 1.76341960899299, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 0.3}) +[2026-06-15 19:14:07,974][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.16906344890594482, 'train/grad_norm_pre_clip': 5.215861797332764, 'timing/data': 0.00055334297940135, 'timing/model': 1.6132174810045399, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 0.3}) +[2026-06-15 19:14:14,030][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.14045435190200806, 'train/grad_norm_pre_clip': 4.261584758758545, 'timing/data': 0.000529768003616482, 'timing/model': 1.5640337800141424, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.3}) +[2026-06-15 19:14:20,118][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.18026091158390045, 'train/grad_norm_pre_clip': 4.42814302444458, 'timing/data': 0.0005080160335637629, 'timing/model': 1.5230685880524106, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 0.3}) +[2026-06-15 19:14:26,209][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.2312266230583191, 'train/grad_norm_pre_clip': 4.160666465759277, 'timing/data': 0.0005239300080575049, 'timing/model': 1.5316959859919734, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 0.3}) +[2026-06-15 19:14:32,280][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.3630019724369049, 'train/grad_norm_pre_clip': 6.013134479522705, 'timing/data': 0.0005164379836060107, 'timing/model': 1.4965303889475763, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 0.3}) +[2026-06-15 19:14:38,389][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.1851552426815033, 'train/grad_norm_pre_clip': 3.1739754676818848, 'timing/data': 0.0007643129792995751, 'timing/model': 1.6931370070087723, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 0.3}) +[2026-06-15 19:14:44,060][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.298518568277359, 'train/grad_norm_pre_clip': 6.569944858551025, 'timing/data': 0.0006499749724753201, 'timing/model': 1.4632512850221246, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 0.3}) +[2026-06-15 19:14:50,060][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.27424415946006775, 'train/grad_norm_pre_clip': 5.679843425750732, 'timing/data': 0.0005131340003572404, 'timing/model': 1.4638524220208637, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 0.3}) +[2026-06-15 19:14:56,088][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.4085004925727844, 'train/grad_norm_pre_clip': 7.305747985839844, 'timing/data': 0.000522518006619066, 'timing/model': 1.5206939480267465, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.3}) +[2026-06-15 19:15:01,957][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.18277084827423096, 'train/grad_norm_pre_clip': 5.56427526473999, 'timing/data': 0.0005652300314977765, 'timing/model': 1.4662217579898424, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 0.3}) +[2026-06-15 19:15:07,958][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.2711847722530365, 'train/grad_norm_pre_clip': 6.518951892852783, 'timing/data': 0.0005057030357420444, 'timing/model': 1.4527532560168765, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 0.3}) +[2026-06-15 19:15:13,887][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.21899715065956116, 'train/grad_norm_pre_clip': 6.51202917098999, 'timing/data': 0.0005541139980778098, 'timing/model': 1.691282539977692, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 0.3}) +[2026-06-15 19:15:19,560][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.045610461384058, 'train/grad_norm_pre_clip': 1.9275903701782227, 'timing/data': 0.0005010960157960653, 'timing/model': 1.468262862006668, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 0.3}) +[2026-06-15 19:24:28,856][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.19830232858657837, 'train/grad_norm_pre_clip': 9.41087818145752, 'eval/loss': 0.1077994704246521, 'eval/flap_precision': 0.8423005565862709, 'eval/flap_recall': 0.9098196392785571, 'eval/flap_f1': 0.8747591522157997, 'eval/flap_rate_model': 0.14972222222222223, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.8747591522157997, 'timing/data': 0.0005022480036132038, 'timing/model': 1.4616705200169235, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 0.3}) +[2026-06-15 19:24:28,857][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_700_state +[2026-06-15 19:24:28,861][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 19:24:34,300][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_700_state/model.safetensors +[2026-06-15 19:24:47,656][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_700_state/optimizer.bin +[2026-06-15 19:24:47,656][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_700_state/sampler.bin +[2026-06-15 19:24:47,657][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_700_state/sampler_1.bin +[2026-06-15 19:24:47,658][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_700_state/random_states_0.pkl +[2026-06-15 19:24:47,766][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700 +[2026-06-15 19:24:47,766][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 19:24:47,776][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 19:24:53,437][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.24116578698158264, 'train/grad_norm_pre_clip': 10.347452163696289, 'timing/data': 0.000534434977453202, 'timing/model': 1.46171643800335, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 0.3}) +[2026-06-15 19:24:59,438][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.12642237544059753, 'train/grad_norm_pre_clip': 14.182554244995117, 'timing/data': 0.0005129739874973893, 'timing/model': 1.4639131629955955, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 0.3}) +[2026-06-15 19:25:05,241][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.14650209248065948, 'train/grad_norm_pre_clip': 13.260672569274902, 'timing/data': 0.0005162190063856542, 'timing/model': 1.4651411550003104, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 0.3}) +[2026-06-15 19:25:11,142][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.25816065073013306, 'train/grad_norm_pre_clip': 12.16744613647461, 'timing/data': 0.0005223169573582709, 'timing/model': 1.4643819860066287, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 0.3}) +[2026-06-15 19:25:17,144][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.27893662452697754, 'train/grad_norm_pre_clip': 13.798347473144531, 'timing/data': 0.0005222279578447342, 'timing/model': 1.4619007949950173, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 0.3}) +[2026-06-15 19:25:23,083][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.22840873897075653, 'train/grad_norm_pre_clip': 7.148141384124756, 'timing/data': 0.0005147160263732076, 'timing/model': 1.6986076809698716, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 0.3}) +[2026-06-15 19:25:28,750][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.14190378785133362, 'train/grad_norm_pre_clip': 8.63552188873291, 'timing/data': 0.0005239300080575049, 'timing/model': 1.4636779599823058, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 0.3}) +[2026-06-15 19:25:34,752][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.09237803518772125, 'train/grad_norm_pre_clip': 3.669682264328003, 'timing/data': 0.0005102900322526693, 'timing/model': 1.4684415689553134, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 0.3}) +[2026-06-15 19:25:40,653][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.07646068185567856, 'train/grad_norm_pre_clip': 4.394596576690674, 'timing/data': 0.0005122629809193313, 'timing/model': 1.4652062709792517, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 0.31}) +[2026-06-15 19:25:46,550][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.14877066016197205, 'train/grad_norm_pre_clip': 13.429574012756348, 'timing/data': 0.0005122930160723627, 'timing/model': 1.4642594549804926, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 0.31}) +[2026-06-15 19:25:52,452][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.11780522763729095, 'train/grad_norm_pre_clip': 4.8314642906188965, 'timing/data': 0.0005138550186529756, 'timing/model': 1.4304003829602152, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 0.31}) +[2026-06-15 19:25:58,389][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.24592076241970062, 'train/grad_norm_pre_clip': 13.250862121582031, 'timing/data': 0.0005042709526605904, 'timing/model': 1.7023582870024256, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 0.31}) +[2026-06-15 19:26:04,053][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.1707613170146942, 'train/grad_norm_pre_clip': 7.046619892120361, 'timing/data': 0.0005095790256746113, 'timing/model': 1.4651972919818945, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 0.31}) +[2026-06-15 19:26:10,054][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.10926523059606552, 'train/grad_norm_pre_clip': 8.752640724182129, 'timing/data': 0.0005197640275582671, 'timing/model': 1.4661951949819922, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 0.31}) +[2026-06-15 19:26:15,851][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.18615853786468506, 'train/grad_norm_pre_clip': 8.465020179748535, 'timing/data': 0.0005123530281707644, 'timing/model': 1.4659708010149188, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 0.31}) +[2026-06-15 19:26:21,652][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.142754465341568, 'train/grad_norm_pre_clip': 7.84899377822876, 'timing/data': 0.000511190970428288, 'timing/model': 1.4667118359939195, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 0.31}) +[2026-06-15 19:26:27,450][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.29360058903694153, 'train/grad_norm_pre_clip': 12.189184188842773, 'timing/data': 0.0005160280270501971, 'timing/model': 1.4371174810221419, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 0.31}) +[2026-06-15 19:26:33,310][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.10487286746501923, 'train/grad_norm_pre_clip': 6.662900447845459, 'timing/data': 0.0005170799558982253, 'timing/model': 1.6244556320016272, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 0.31}) +[2026-06-15 19:26:38,952][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.08768099546432495, 'train/grad_norm_pre_clip': 2.5200769901275635, 'timing/data': 0.00051112100481987, 'timing/model': 1.4652401600033045, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 0.31}) +[2026-06-15 19:26:44,851][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.14506326615810394, 'train/grad_norm_pre_clip': 5.129317283630371, 'timing/data': 0.0005130440113134682, 'timing/model': 1.4651237649959512, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 0.31}) +[2026-06-15 19:26:50,840][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.3637716472148895, 'train/grad_norm_pre_clip': 10.663288116455078, 'timing/data': 0.0005103300209157169, 'timing/model': 1.5510162830469199, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 0.31}) +[2026-06-15 19:26:56,752][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.13056771457195282, 'train/grad_norm_pre_clip': 5.288717746734619, 'timing/data': 0.0005220780149102211, 'timing/model': 1.4328046400332823, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 0.31}) +[2026-06-15 19:27:02,681][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.3276984691619873, 'train/grad_norm_pre_clip': 14.71088695526123, 'timing/data': 0.0005159690044820309, 'timing/model': 1.692246507969685, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 0.31}) +[2026-06-15 19:27:08,535][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.09525343030691147, 'train/grad_norm_pre_clip': 4.32056999206543, 'timing/data': 0.000517470994964242, 'timing/model': 1.6470654729637317, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 0.31}) +[2026-06-15 19:27:14,255][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.09308046102523804, 'train/grad_norm_pre_clip': 3.375845432281494, 'timing/data': 0.0005214670090936124, 'timing/model': 1.4663837240077555, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 0.31}) +[2026-06-15 19:27:20,255][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.18740108609199524, 'train/grad_norm_pre_clip': 5.101644515991211, 'timing/data': 0.0005815550102852285, 'timing/model': 1.4662332480074838, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 0.31}) +[2026-06-15 19:27:26,153][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.20410564541816711, 'train/grad_norm_pre_clip': 6.762373924255371, 'timing/data': 0.0005161190056242049, 'timing/model': 1.4608430410153233, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 0.31}) +[2026-06-15 19:27:32,058][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.2439573109149933, 'train/grad_norm_pre_clip': 6.447679042816162, 'timing/data': 0.0005174599937163293, 'timing/model': 1.430943045008462, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 0.31}) +[2026-06-15 19:27:37,982][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.19986388087272644, 'train/grad_norm_pre_clip': 10.115184783935547, 'timing/data': 0.0005129440105520189, 'timing/model': 1.6909330439521, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 0.31}) +[2026-06-15 19:27:43,759][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.1398591548204422, 'train/grad_norm_pre_clip': 8.476012229919434, 'timing/data': 0.0005152170197106898, 'timing/model': 1.5667067069443874, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.31}) +[2026-06-15 19:27:49,456][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.21382208168506622, 'train/grad_norm_pre_clip': 7.789460182189941, 'timing/data': 0.0005605240003205836, 'timing/model': 1.465717525978107, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 0.31}) +[2026-06-15 19:27:55,633][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.08839715272188187, 'train/grad_norm_pre_clip': 3.4639406204223633, 'timing/data': 0.0005229079979471862, 'timing/model': 1.5596639750292525, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 0.31}) +[2026-06-15 19:28:01,657][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.2060536891222, 'train/grad_norm_pre_clip': 19.22196388244629, 'timing/data': 0.0005193940014578402, 'timing/model': 1.4769555350067094, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 0.32}) +[2026-06-15 19:28:07,614][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.2736692726612091, 'train/grad_norm_pre_clip': 9.067057609558105, 'timing/data': 0.0005185419577173889, 'timing/model': 1.4878500750055537, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 0.32}) +[2026-06-15 19:28:13,488][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.25591331720352173, 'train/grad_norm_pre_clip': 5.350110054016113, 'timing/data': 0.00052225700346753, 'timing/model': 1.7008571660262533, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 0.32}) +[2026-06-15 19:28:19,153][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.07963084429502487, 'train/grad_norm_pre_clip': 4.1864213943481445, 'timing/data': 0.0004979019868187606, 'timing/model': 1.4621807709918357, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 0.32}) +[2026-06-15 19:28:25,223][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.2226608246564865, 'train/grad_norm_pre_clip': 4.563947677612305, 'timing/data': 0.0004926839610561728, 'timing/model': 1.5347538049682043, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.32}) +[2026-06-15 19:28:31,154][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.092313252389431, 'train/grad_norm_pre_clip': 6.375314712524414, 'timing/data': 0.000509848992805928, 'timing/model': 1.4638619220349938, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 0.32}) +[2026-06-15 19:28:37,056][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.5820580124855042, 'train/grad_norm_pre_clip': 48.380638122558594, 'timing/data': 0.0005159279680810869, 'timing/model': 1.466980601020623, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 0.32}) +[2026-06-15 19:28:42,954][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.5363665223121643, 'train/grad_norm_pre_clip': 43.33097457885742, 'timing/data': 0.0004949980066157877, 'timing/model': 1.4327780089806765, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 0.32}) +[2026-06-15 19:28:48,882][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.14820538461208344, 'train/grad_norm_pre_clip': 4.549400329589844, 'timing/data': 0.0004952879971824586, 'timing/model': 1.6945010350318626, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 0.32}) +[2026-06-15 19:28:54,553][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.1675349324941635, 'train/grad_norm_pre_clip': 7.750378131866455, 'timing/data': 0.0004845020011998713, 'timing/model': 1.4663830460049212, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 0.32}) +[2026-06-15 19:29:00,555][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.0672304555773735, 'train/grad_norm_pre_clip': 12.442081451416016, 'timing/data': 0.0005173400277271867, 'timing/model': 1.463889745005872, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 0.32}) +[2026-06-15 19:29:06,350][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.15945415198802948, 'train/grad_norm_pre_clip': 10.83951187133789, 'timing/data': 0.0005075160297565162, 'timing/model': 1.460451174993068, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 0.32}) +[2026-06-15 19:29:12,153][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.1994251161813736, 'train/grad_norm_pre_clip': 6.710169315338135, 'timing/data': 0.0005040699616074562, 'timing/model': 1.4686801949865185, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 0.32}) +[2026-06-15 19:29:17,957][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.23534071445465088, 'train/grad_norm_pre_clip': 7.521369457244873, 'timing/data': 0.0005055329529568553, 'timing/model': 1.4458792619989254, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 0.32}) +[2026-06-15 19:29:23,808][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.10326817631721497, 'train/grad_norm_pre_clip': 10.600239753723145, 'timing/data': 0.0005203550099395216, 'timing/model': 1.6229596610064618, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 0.32}) +[2026-06-15 19:29:29,447][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.13290128111839294, 'train/grad_norm_pre_clip': 4.461176872253418, 'timing/data': 0.0005080770351924002, 'timing/model': 1.46055262995651, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 0.32}) +[2026-06-15 19:29:35,353][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.10575941205024719, 'train/grad_norm_pre_clip': 10.78095531463623, 'timing/data': 0.0005228790105320513, 'timing/model': 1.4643585889716633, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 0.32}) +[2026-06-15 19:29:41,252][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.06366318464279175, 'train/grad_norm_pre_clip': 4.119124412536621, 'timing/data': 0.0005178920109756291, 'timing/model': 1.4624840979813598, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 0.32}) +[2026-06-15 19:29:47,238][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.26272153854370117, 'train/grad_norm_pre_clip': 7.92027473449707, 'timing/data': 0.0005158879794180393, 'timing/model': 1.502417572948616, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 0.32}) +[2026-06-15 19:29:53,183][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.16797339916229248, 'train/grad_norm_pre_clip': 5.7042694091796875, 'timing/data': 0.0005389319849200547, 'timing/model': 1.6893032770021819, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 0.32}) +[2026-06-15 19:29:59,039][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.08658067137002945, 'train/grad_norm_pre_clip': 8.452718734741211, 'timing/data': 0.0005257820012047887, 'timing/model': 1.6413219689857215, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 0.32}) +[2026-06-15 19:30:04,757][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.4266919493675232, 'train/grad_norm_pre_clip': 10.586669921875, 'timing/data': 0.0005262040067464113, 'timing/model': 1.4611468879738823, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 0.32}) +[2026-06-15 19:30:10,758][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.1414254903793335, 'train/grad_norm_pre_clip': 13.54299259185791, 'timing/data': 0.0005184420151636004, 'timing/model': 1.4668840349768288, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 0.32}) +[2026-06-15 19:30:16,754][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.2124142199754715, 'train/grad_norm_pre_clip': 4.865269660949707, 'timing/data': 0.0007643930148333311, 'timing/model': 1.483302625012584, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 0.33}) +[2026-06-15 19:30:22,661][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.07724034041166306, 'train/grad_norm_pre_clip': 4.443365573883057, 'timing/data': 0.0005220170132815838, 'timing/model': 1.4340622349991463, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 0.33}) +[2026-06-15 19:30:28,590][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.05957404524087906, 'train/grad_norm_pre_clip': 2.2504448890686035, 'timing/data': 0.000523298978805542, 'timing/model': 1.7016549410182051, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 0.33}) +[2026-06-15 19:30:34,390][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.4284091591835022, 'train/grad_norm_pre_clip': 19.40889549255371, 'timing/data': 0.0005792920128442347, 'timing/model': 1.609723505040165, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 0.33}) +[2026-06-15 19:30:40,052][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.2581911087036133, 'train/grad_norm_pre_clip': 5.337838172912598, 'timing/data': 0.0004960700171068311, 'timing/model': 1.4679374109837227, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 0.33}) +[2026-06-15 19:30:46,052][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.25185805559158325, 'train/grad_norm_pre_clip': 9.604567527770996, 'timing/data': 0.0005060740513727069, 'timing/model': 1.464493892970495, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 0.33}) +[2026-06-15 19:30:51,952][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.06808018684387207, 'train/grad_norm_pre_clip': 4.749769687652588, 'timing/data': 0.0004898400511592627, 'timing/model': 1.4651753079961054, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 0.33}) +[2026-06-15 19:30:57,852][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.1434352695941925, 'train/grad_norm_pre_clip': 10.615585327148438, 'timing/data': 0.0004916730104014277, 'timing/model': 1.4293931819847785, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 0.33}) +[2026-06-15 19:31:03,907][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.09257642924785614, 'train/grad_norm_pre_clip': 5.038510322570801, 'timing/data': 0.0004937360063195229, 'timing/model': 1.7408436840050854, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 0.33}) +[2026-06-15 19:31:09,554][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.23880797624588013, 'train/grad_norm_pre_clip': 9.30353832244873, 'timing/data': 0.0004852830315940082, 'timing/model': 1.4555514120147564, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.33}) +[2026-06-15 19:31:15,554][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.1298278570175171, 'train/grad_norm_pre_clip': 6.54444694519043, 'timing/data': 0.0004909220151603222, 'timing/model': 1.4667359570157714, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 0.33}) +[2026-06-15 19:31:21,453][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.06058647483587265, 'train/grad_norm_pre_clip': 10.621500968933105, 'timing/data': 0.0004933049785904586, 'timing/model': 1.4676147320424207, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 0.33}) +[2026-06-15 19:31:27,354][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.22861801087856293, 'train/grad_norm_pre_clip': 9.496930122375488, 'timing/data': 0.0004917830228805542, 'timing/model': 1.4674589859787375, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 0.33}) +[2026-06-15 19:31:33,253][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.31205570697784424, 'train/grad_norm_pre_clip': 12.58615493774414, 'timing/data': 0.0005002859979867935, 'timing/model': 1.4305640020174906, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 0.33}) +[2026-06-15 19:31:39,186][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.04051853343844414, 'train/grad_norm_pre_clip': 2.413337230682373, 'timing/data': 0.0004958590143360198, 'timing/model': 1.6954470219789073, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 0.33}) +[2026-06-15 19:31:44,956][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.2087084800004959, 'train/grad_norm_pre_clip': 10.992802619934082, 'timing/data': 0.0004886980168521404, 'timing/model': 1.4668204569607042, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 0.33}) +[2026-06-15 19:31:50,956][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.16587260365486145, 'train/grad_norm_pre_clip': 6.912711143493652, 'timing/data': 0.000511601974721998, 'timing/model': 1.465247658023145, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 0.33}) +[2026-06-15 19:31:56,753][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.19372904300689697, 'train/grad_norm_pre_clip': 15.078107833862305, 'timing/data': 0.0005092790233902633, 'timing/model': 1.4665546619798988, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 0.33}) +[2026-06-15 19:32:02,553][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.2963334619998932, 'train/grad_norm_pre_clip': 8.515145301818848, 'timing/data': 0.0005260329926386476, 'timing/model': 1.4647798669757321, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 0.33}) +[2026-06-15 19:32:08,358][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.35803911089897156, 'train/grad_norm_pre_clip': 12.4424467086792, 'timing/data': 0.0005200940067879856, 'timing/model': 1.4400388189824298, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 0.33}) +[2026-06-15 19:32:14,217][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.19667287170886993, 'train/grad_norm_pre_clip': 10.421446800231934, 'timing/data': 0.00075018301140517, 'timing/model': 1.6236014469759539, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 0.33}) +[2026-06-15 19:32:19,858][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.1389879733324051, 'train/grad_norm_pre_clip': 6.6530442237854, 'timing/data': 0.0005276559968478978, 'timing/model': 1.4643729469971731, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 0.33}) +[2026-06-15 19:32:25,757][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.2151545137166977, 'train/grad_norm_pre_clip': 6.285572528839111, 'timing/data': 0.000505362986586988, 'timing/model': 1.466126047016587, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 0.33}) +[2026-06-15 19:32:31,658][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.11638801544904709, 'train/grad_norm_pre_clip': 3.936763048171997, 'timing/data': 0.0005087580066174269, 'timing/model': 1.4624087050324306, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 0.34}) +[2026-06-15 19:32:37,558][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.19762387871742249, 'train/grad_norm_pre_clip': 4.621673107147217, 'timing/data': 0.0005204849876463413, 'timing/model': 1.4372262780088931, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 0.34}) +[2026-06-15 19:32:43,481][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.2179013341665268, 'train/grad_norm_pre_clip': 7.980936527252197, 'timing/data': 0.0005222279578447342, 'timing/model': 1.686740901961457, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 0.34}) +[2026-06-15 19:32:49,337][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.3456518352031708, 'train/grad_norm_pre_clip': 10.334129333496094, 'timing/data': 0.0005072649801149964, 'timing/model': 1.6446909290389158, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 0.34}) +[2026-06-15 19:32:55,060][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.13432125747203827, 'train/grad_norm_pre_clip': 3.8795831203460693, 'timing/data': 0.0005052830092608929, 'timing/model': 1.466604144021403, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 0.34}) +[2026-06-15 19:33:01,059][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.1665678322315216, 'train/grad_norm_pre_clip': 6.882910251617432, 'timing/data': 0.0005295380251482129, 'timing/model': 1.4652197019895539, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 0.34}) +[2026-06-15 19:33:07,059][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.0923076644539833, 'train/grad_norm_pre_clip': 2.664746046066284, 'timing/data': 0.0005237199948169291, 'timing/model': 1.4643170170020312, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 0.34}) +[2026-06-15 19:33:13,029][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.25098666548728943, 'train/grad_norm_pre_clip': 17.17135238647461, 'timing/data': 0.0005260129692032933, 'timing/model': 1.501558126998134, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.34}) +[2026-06-15 19:33:18,988][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.19204995036125183, 'train/grad_norm_pre_clip': 8.267446517944336, 'timing/data': 0.0005177310085855424, 'timing/model': 1.6945555990096182, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 0.34}) +[2026-06-15 19:33:24,760][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.172562375664711, 'train/grad_norm_pre_clip': 5.587918758392334, 'timing/data': 0.0005630480009131134, 'timing/model': 1.566830067022238, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 0.34}) +[2026-06-15 19:33:30,460][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.05201328918337822, 'train/grad_norm_pre_clip': 4.374748706817627, 'timing/data': 0.0005277559976093471, 'timing/model': 1.46626976900734, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 0.34}) +[2026-06-15 19:33:36,456][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.21122530102729797, 'train/grad_norm_pre_clip': 10.23863697052002, 'timing/data': 0.0006150039844214916, 'timing/model': 1.4574101570178755, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 0.34}) +[2026-06-15 19:33:42,457][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.2737065851688385, 'train/grad_norm_pre_clip': 14.174981117248535, 'timing/data': 0.0005545250023715198, 'timing/model': 1.463173151947558, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 0.34}) +[2026-06-15 19:33:48,543][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.1426532119512558, 'train/grad_norm_pre_clip': 10.063220024108887, 'timing/data': 0.0005362580413930118, 'timing/model': 1.5087390490225516, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 0.34}) +[2026-06-15 19:33:54,519][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.07529378682374954, 'train/grad_norm_pre_clip': 5.368988037109375, 'timing/data': 0.0005572590162046254, 'timing/model': 1.7245363999973051, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 0.34}) +[2026-06-15 19:34:00,159][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.08564934134483337, 'train/grad_norm_pre_clip': 3.4146385192871094, 'timing/data': 0.0005398329813033342, 'timing/model': 1.4551703830366023, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 0.34}) +[2026-06-15 19:34:06,156][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.1695447862148285, 'train/grad_norm_pre_clip': 11.174505233764648, 'timing/data': 0.0005136349936947227, 'timing/model': 1.4646186259924434, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 0.34}) +[2026-06-15 19:34:12,056][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.011631454341113567, 'train/grad_norm_pre_clip': 1.0954965353012085, 'timing/data': 0.0005170800141058862, 'timing/model': 1.465079846966546, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 0.34}) +[2026-06-15 19:34:17,949][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.0386197604238987, 'train/grad_norm_pre_clip': 6.692592144012451, 'timing/data': 0.0005084770382381976, 'timing/model': 1.4625855519552715, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 0.34}) +[2026-06-15 19:34:23,860][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.2594355046749115, 'train/grad_norm_pre_clip': 10.88076400756836, 'timing/data': 0.0005172400269657373, 'timing/model': 1.432416422991082, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 0.34}) +[2026-06-15 19:34:29,784][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.4094392955303192, 'train/grad_norm_pre_clip': 20.294597625732422, 'timing/data': 0.0005182819440960884, 'timing/model': 1.6957330830045976, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 0.34}) +[2026-06-15 19:42:58,549][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.5191944241523743, 'train/grad_norm_pre_clip': 20.133045196533203, 'eval/loss': 0.13508853316307068, 'eval/flap_precision': 0.8852813852813853, 'eval/flap_recall': 0.8196392785571143, 'eval/flap_f1': 0.8511966701352757, 'eval/flap_rate_model': 0.12833333333333333, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.8511966701352756, 'timing/data': 0.0005127229960635304, 'timing/model': 1.462347938970197, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.34}) +[2026-06-15 19:42:58,549][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_800_state +[2026-06-15 19:42:58,554][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 19:43:04,384][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_800_state/model.safetensors +[2026-06-15 19:43:18,317][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_800_state/optimizer.bin +[2026-06-15 19:43:18,318][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_800_state/sampler.bin +[2026-06-15 19:43:18,318][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_800_state/sampler_1.bin +[2026-06-15 19:43:18,319][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_800_state/random_states_0.pkl +[2026-06-15 19:43:18,433][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800 +[2026-06-15 19:43:18,434][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 19:43:18,444][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 19:43:24,494][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.22763314843177795, 'train/grad_norm_pre_clip': 16.92005157470703, 'timing/data': 0.0005122640286572278, 'timing/model': 1.720661224971991, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 0.34}) +[2026-06-15 19:43:30,228][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.0846758633852005, 'train/grad_norm_pre_clip': 8.512738227844238, 'timing/data': 0.0005139670101925731, 'timing/model': 1.549388215993531, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 0.35}) +[2026-06-15 19:43:36,148][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.06645333021879196, 'train/grad_norm_pre_clip': 7.861963272094727, 'timing/data': 0.0005076379748061299, 'timing/model': 1.4686285789939575, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 0.35}) +[2026-06-15 19:43:42,050][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.42654916644096375, 'train/grad_norm_pre_clip': 11.823800086975098, 'timing/data': 0.0005299199838191271, 'timing/model': 1.4688096409663558, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 0.35}) +[2026-06-15 19:43:47,957][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.279386430978775, 'train/grad_norm_pre_clip': 18.257278442382812, 'timing/data': 0.0005170409567654133, 'timing/model': 1.465952548023779, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 0.35}) +[2026-06-15 19:43:53,898][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.3600950539112091, 'train/grad_norm_pre_clip': 29.63985824584961, 'timing/data': 0.0005085180164314806, 'timing/model': 1.4741201070137322, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 0.35}) +[2026-06-15 19:43:59,798][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.3368683457374573, 'train/grad_norm_pre_clip': 29.941268920898438, 'timing/data': 0.0005279369652271271, 'timing/model': 1.7088126610033214, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 0.35}) +[2026-06-15 19:44:05,462][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.042495325207710266, 'train/grad_norm_pre_clip': 3.581731081008911, 'timing/data': 0.0005246620276011527, 'timing/model': 1.4345257489476353, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 0.35}) +[2026-06-15 19:44:11,452][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.22637765109539032, 'train/grad_norm_pre_clip': 17.27581787109375, 'timing/data': 0.000573564029764384, 'timing/model': 1.4617890589870512, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 0.35}) +[2026-06-15 19:44:17,359][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.1744011640548706, 'train/grad_norm_pre_clip': 19.818098068237305, 'timing/data': 0.0005742250359617174, 'timing/model': 1.4620198549819179, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 0.35}) +[2026-06-15 19:44:23,261][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.0655369982123375, 'train/grad_norm_pre_clip': 4.7066969871521, 'timing/data': 0.0005173520185053349, 'timing/model': 1.4390093279653229, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 0.35}) +[2026-06-15 19:44:29,236][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.27724048495292664, 'train/grad_norm_pre_clip': 12.119721412658691, 'timing/data': 0.0005216780118644238, 'timing/model': 1.5069889859878458, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 0.35}) +[2026-06-15 19:44:35,151][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.12325144559144974, 'train/grad_norm_pre_clip': 8.113519668579102, 'timing/data': 0.0005142269656062126, 'timing/model': 1.6534674200229347, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 0.35}) +[2026-06-15 19:44:40,863][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.1455385833978653, 'train/grad_norm_pre_clip': 9.02787971496582, 'timing/data': 0.0005231800023466349, 'timing/model': 1.4647132270038128, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 0.35}) +[2026-06-15 19:44:46,866][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.1634770780801773, 'train/grad_norm_pre_clip': 11.739557266235352, 'timing/data': 0.0005070560146123171, 'timing/model': 1.467646031989716, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 0.35}) +[2026-06-15 19:44:52,862][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.08944831043481827, 'train/grad_norm_pre_clip': 3.6590628623962402, 'timing/data': 0.0005141469882801175, 'timing/model': 1.4845070089795627, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 0.35}) +[2026-06-15 19:44:58,855][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.23483891785144806, 'train/grad_norm_pre_clip': 9.386383056640625, 'timing/data': 0.0005206660134717822, 'timing/model': 1.5109775409800932, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 0.35}) +[2026-06-15 19:45:04,767][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.17234376072883606, 'train/grad_norm_pre_clip': 4.969449043273926, 'timing/data': 0.000519865017849952, 'timing/model': 1.4448163490160368, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 0.35}) +[2026-06-15 19:45:10,676][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.20501655340194702, 'train/grad_norm_pre_clip': 16.299617767333984, 'timing/data': 0.0005129550117999315, 'timing/model': 1.6751572049688548, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 0.35}) +[2026-06-15 19:45:16,364][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.18175546824932098, 'train/grad_norm_pre_clip': 6.621756076812744, 'timing/data': 0.0005932939820922911, 'timing/model': 1.4671694790013134, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 0.35}) +[2026-06-15 19:45:22,365][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.16085165739059448, 'train/grad_norm_pre_clip': 3.7469916343688965, 'timing/data': 0.000506034994032234, 'timing/model': 1.4664525130065158, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 0.35}) +[2026-06-15 19:45:28,364][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.16159315407276154, 'train/grad_norm_pre_clip': 12.1106595993042, 'timing/data': 0.0005058550159446895, 'timing/model': 1.4840899120317772, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 0.35}) +[2026-06-15 19:45:34,344][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.42317143082618713, 'train/grad_norm_pre_clip': 11.345542907714844, 'timing/data': 0.0005061249830760062, 'timing/model': 1.5154588740551844, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 0.35}) +[2026-06-15 19:45:40,226][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.31380313634872437, 'train/grad_norm_pre_clip': 13.955245971679688, 'timing/data': 0.0005237709847278893, 'timing/model': 1.5011958289542235, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 0.35}) +[2026-06-15 19:45:46,304][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.2437593936920166, 'train/grad_norm_pre_clip': 8.019153594970703, 'timing/data': 0.000527706986758858, 'timing/model': 1.7290952629991807, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 0.35}) +[2026-06-15 19:45:52,062][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.1580144464969635, 'train/grad_norm_pre_clip': 4.954633712768555, 'timing/data': 0.0005098910187371075, 'timing/model': 1.459135498967953, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 0.36}) +[2026-06-15 19:45:58,222][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.10085681825876236, 'train/grad_norm_pre_clip': 3.7056617736816406, 'timing/data': 0.0005196949932724237, 'timing/model': 1.549166882992722, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 0.36}) +[2026-06-15 19:46:04,300][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.09323512017726898, 'train/grad_norm_pre_clip': 4.886906147003174, 'timing/data': 0.000509890029206872, 'timing/model': 1.5181838190183043, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 0.36}) +[2026-06-15 19:46:10,417][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.25049063563346863, 'train/grad_norm_pre_clip': 10.415197372436523, 'timing/data': 0.0005168510251678526, 'timing/model': 1.529329493001569, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 0.36}) +[2026-06-15 19:46:16,557][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.36117786169052124, 'train/grad_norm_pre_clip': 39.020442962646484, 'timing/data': 0.000517993001267314, 'timing/model': 1.5135959850158542, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 0.36}) +[2026-06-15 19:46:22,599][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.13699039816856384, 'train/grad_norm_pre_clip': 14.14293098449707, 'timing/data': 0.0005047429585829377, 'timing/model': 1.7012630870449357, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 0.36}) +[2026-06-15 19:46:28,259][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.5509292483329773, 'train/grad_norm_pre_clip': 18.4818058013916, 'timing/data': 0.0004984640399925411, 'timing/model': 1.464356045005843, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 0.36}) +[2026-06-15 19:46:34,260][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.20730960369110107, 'train/grad_norm_pre_clip': 7.996971607208252, 'timing/data': 0.00048656697617843747, 'timing/model': 1.4666382250143215, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 0.36}) +[2026-06-15 19:46:40,260][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.12446959316730499, 'train/grad_norm_pre_clip': 5.981166362762451, 'timing/data': 0.0005015790229663253, 'timing/model': 1.5032289600349031, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 0.36}) +[2026-06-15 19:46:46,166][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.19285117089748383, 'train/grad_norm_pre_clip': 9.134937286376953, 'timing/data': 0.0005192539538256824, 'timing/model': 1.4399307310231961, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.36}) +[2026-06-15 19:46:52,087][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.1405307799577713, 'train/grad_norm_pre_clip': 4.260213851928711, 'timing/data': 0.0005162000306881964, 'timing/model': 1.691268220951315, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 0.36}) +[2026-06-15 19:46:57,910][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.2871743142604828, 'train/grad_norm_pre_clip': 8.990671157836914, 'timing/data': 0.0005145770264789462, 'timing/model': 1.615471568016801, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 0.36}) +[2026-06-15 19:47:03,561][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.06940539181232452, 'train/grad_norm_pre_clip': 4.545845985412598, 'timing/data': 0.0005130450008437037, 'timing/model': 1.4631289110402577, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 0.36}) +[2026-06-15 19:47:09,461][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.25377583503723145, 'train/grad_norm_pre_clip': 11.54942512512207, 'timing/data': 0.0005143170128576458, 'timing/model': 1.4637977079837583, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 0.36}) +[2026-06-15 19:47:15,467][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.031995128840208054, 'train/grad_norm_pre_clip': 1.415804386138916, 'timing/data': 0.0005068059545010328, 'timing/model': 1.473253617004957, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.36}) +[2026-06-15 19:47:21,565][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.10029434412717819, 'train/grad_norm_pre_clip': 4.746496200561523, 'timing/data': 0.0005203050095587969, 'timing/model': 1.5287743899971247, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 0.36}) +[2026-06-15 19:47:27,511][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.23230358958244324, 'train/grad_norm_pre_clip': 9.75153923034668, 'timing/data': 0.000584619992878288, 'timing/model': 1.7161332260002382, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 0.36}) +[2026-06-15 19:47:33,259][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.3666665852069855, 'train/grad_norm_pre_clip': 9.737468719482422, 'timing/data': 0.0005104310112074018, 'timing/model': 1.4601608189987019, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 0.36}) +[2026-06-15 19:47:39,258][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.027466561645269394, 'train/grad_norm_pre_clip': 1.8409886360168457, 'timing/data': 0.0005199050065129995, 'timing/model': 1.458114231005311, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 0.36}) +[2026-06-15 19:47:45,168][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.10224372148513794, 'train/grad_norm_pre_clip': 1.7519547939300537, 'timing/data': 0.0005622570170089602, 'timing/model': 1.4486986580304801, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 0.36}) +[2026-06-15 19:47:51,159][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.062004074454307556, 'train/grad_norm_pre_clip': 8.567293167114258, 'timing/data': 0.00047962600365281105, 'timing/model': 1.5367132080136798, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 0.36}) +[2026-06-15 19:47:57,064][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.17751894891262054, 'train/grad_norm_pre_clip': 8.65960693359375, 'timing/data': 0.0005045330035500228, 'timing/model': 1.4393223479855806, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 0.36}) +[2026-06-15 19:48:03,057][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.29713380336761475, 'train/grad_norm_pre_clip': 23.6118106842041, 'timing/data': 0.0005063550197519362, 'timing/model': 1.6582972100004554, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 0.36}) +[2026-06-15 19:48:08,762][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.025917043909430504, 'train/grad_norm_pre_clip': 3.5383284091949463, 'timing/data': 0.0005063950084149837, 'timing/model': 1.4690659589832649, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 0.37}) +[2026-06-15 19:48:14,766][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.2790153920650482, 'train/grad_norm_pre_clip': 8.580915451049805, 'timing/data': 0.0004959000507369637, 'timing/model': 1.467005581012927, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.37}) +[2026-06-15 19:48:20,767][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.12463324517011642, 'train/grad_norm_pre_clip': 9.780356407165527, 'timing/data': 0.000517581996973604, 'timing/model': 1.4624721549917012, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 0.37}) +[2026-06-15 19:48:26,847][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.21746192872524261, 'train/grad_norm_pre_clip': 6.472532749176025, 'timing/data': 0.0005027399747632444, 'timing/model': 1.5259872670285404, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 0.37}) +[2026-06-15 19:48:32,861][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.20181749761104584, 'train/grad_norm_pre_clip': 11.829925537109375, 'timing/data': 0.000516951025929302, 'timing/model': 1.4370730679947883, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 0.37}) +[2026-06-15 19:48:38,776][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.18623656034469604, 'train/grad_norm_pre_clip': 11.579309463500977, 'timing/data': 0.0005148469936102629, 'timing/model': 1.6854276009835303, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 0.37}) +[2026-06-15 19:48:44,465][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.03285236284136772, 'train/grad_norm_pre_clip': 2.3704729080200195, 'timing/data': 0.0005939650000073016, 'timing/model': 1.4664268149645068, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 0.37}) +[2026-06-15 19:48:50,464][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.07694002240896225, 'train/grad_norm_pre_clip': 6.426998138427734, 'timing/data': 0.0005173509707674384, 'timing/model': 1.4655136780347675, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 0.37}) +[2026-06-15 19:48:56,464][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.5606614351272583, 'train/grad_norm_pre_clip': 13.09443187713623, 'timing/data': 0.0005004570120945573, 'timing/model': 1.460152253974229, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 0.37}) +[2026-06-15 19:49:02,367][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.06722420454025269, 'train/grad_norm_pre_clip': 4.263628005981445, 'timing/data': 0.0005461639957502484, 'timing/model': 1.4386840289807878, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 0.37}) +[2026-06-15 19:49:08,161][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.25399670004844666, 'train/grad_norm_pre_clip': 8.708049774169922, 'timing/data': 0.0005085280281491578, 'timing/model': 1.4605928949895315, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 0.37}) +[2026-06-15 19:49:14,098][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.16726328432559967, 'train/grad_norm_pre_clip': 3.059060573577881, 'timing/data': 0.0005232400144450366, 'timing/model': 1.6983267070027068, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.37}) +[2026-06-15 19:49:19,766][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.1278459131717682, 'train/grad_norm_pre_clip': 8.584290504455566, 'timing/data': 0.0005142770241945982, 'timing/model': 1.4670936950133182, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 0.37}) +[2026-06-15 19:49:25,763][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.21458888053894043, 'train/grad_norm_pre_clip': 11.054941177368164, 'timing/data': 0.000539403990842402, 'timing/model': 1.4670074949972332, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 0.37}) +[2026-06-15 19:49:31,758][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.18017780780792236, 'train/grad_norm_pre_clip': 8.394450187683105, 'timing/data': 0.0007636439986526966, 'timing/model': 1.4550098270410672, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 0.37}) +[2026-06-15 19:49:37,668][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.1829042285680771, 'train/grad_norm_pre_clip': 8.313130378723145, 'timing/data': 0.0005220780149102211, 'timing/model': 1.440702872001566, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 0.37}) +[2026-06-15 19:49:43,565][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.40895989537239075, 'train/grad_norm_pre_clip': 11.93154525756836, 'timing/data': 0.0005199050065129995, 'timing/model': 1.4433424129965715, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 0.37}) +[2026-06-15 19:49:49,491][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.08018898963928223, 'train/grad_norm_pre_clip': 4.713376522064209, 'timing/data': 0.0005616769776679575, 'timing/model': 1.6983566809794866, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 0.37}) +[2026-06-15 19:49:55,168][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.11765710264444351, 'train/grad_norm_pre_clip': 6.4281768798828125, 'timing/data': 0.0005205560009926558, 'timing/model': 1.471395097032655, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 0.37}) +[2026-06-15 19:50:01,342][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.1952630579471588, 'train/grad_norm_pre_clip': 5.756136417388916, 'timing/data': 0.0010647359886206686, 'timing/model': 1.5702854749979451, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 0.37}) +[2026-06-15 19:50:07,517][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.2508113384246826, 'train/grad_norm_pre_clip': 9.072457313537598, 'timing/data': 0.0006093159900046885, 'timing/model': 1.5216936810174957, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 0.37}) +[2026-06-15 19:50:13,538][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.17387640476226807, 'train/grad_norm_pre_clip': 2.7526230812072754, 'timing/data': 0.0005775400204584002, 'timing/model': 1.5086190710426308, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.37}) +[2026-06-15 19:50:19,522][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.06881366670131683, 'train/grad_norm_pre_clip': 4.087835311889648, 'timing/data': 0.0004956689663231373, 'timing/model': 1.7275543410214595, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 0.37}) +[2026-06-15 19:50:25,409][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.1124420240521431, 'train/grad_norm_pre_clip': 5.949719429016113, 'timing/data': 0.00048580497968941927, 'timing/model': 1.6114985519670881, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 0.38}) +[2026-06-15 19:50:31,064][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.20839910209178925, 'train/grad_norm_pre_clip': 6.661923408508301, 'timing/data': 0.000503640971146524, 'timing/model': 1.460271917982027, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 0.38}) +[2026-06-15 19:50:36,965][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.16599532961845398, 'train/grad_norm_pre_clip': 4.909924507141113, 'timing/data': 0.0005262939957901835, 'timing/model': 1.466068034991622, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 0.38}) +[2026-06-15 19:50:43,064][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.1862725466489792, 'train/grad_norm_pre_clip': 7.237696647644043, 'timing/data': 0.0004951690207235515, 'timing/model': 1.4503881509881467, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 0.38}) +[2026-06-15 19:50:49,044][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.05969490855932236, 'train/grad_norm_pre_clip': 2.7051150798797607, 'timing/data': 0.0005100510315969586, 'timing/model': 1.5113580700126477, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 0.38}) +[2026-06-15 19:50:55,095][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.18092221021652222, 'train/grad_norm_pre_clip': 6.080633163452148, 'timing/data': 0.0004901509964838624, 'timing/model': 1.6998310890048742, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 0.38}) +[2026-06-15 19:51:00,760][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.29731109738349915, 'train/grad_norm_pre_clip': 15.376656532287598, 'timing/data': 0.00048217899166047573, 'timing/model': 1.4643917920184322, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 0.38}) +[2026-06-15 19:51:06,760][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.08696742355823517, 'train/grad_norm_pre_clip': 4.085290431976318, 'timing/data': 0.0005161790177226067, 'timing/model': 1.4493567500030622, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 0.38}) +[2026-06-15 19:51:12,763][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.1306038498878479, 'train/grad_norm_pre_clip': 5.032175540924072, 'timing/data': 0.0005233199917711318, 'timing/model': 1.4521036410005763, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 0.38}) +[2026-06-15 19:51:18,762][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.21696853637695312, 'train/grad_norm_pre_clip': 6.359750270843506, 'timing/data': 0.0005055939545854926, 'timing/model': 1.4642303430009633, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 0.38}) +[2026-06-15 19:51:24,767][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.16370731592178345, 'train/grad_norm_pre_clip': 5.215304374694824, 'timing/data': 0.0005080369883216918, 'timing/model': 1.4405600050231442, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 0.38}) +[2026-06-15 19:51:30,649][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.15931403636932373, 'train/grad_norm_pre_clip': 6.241617202758789, 'timing/data': 0.0005086889723315835, 'timing/model': 1.650142024969682, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 0.38}) +[2026-06-15 19:51:36,363][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.25788241624832153, 'train/grad_norm_pre_clip': 11.541491508483887, 'timing/data': 0.0005091089988127351, 'timing/model': 1.4645237229997292, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.38}) +[2026-06-15 19:51:42,461][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.1811823695898056, 'train/grad_norm_pre_clip': 14.077272415161133, 'timing/data': 0.0005216869758442044, 'timing/model': 1.4645919929607771, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 0.38}) +[2026-06-15 19:51:48,461][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.18673360347747803, 'train/grad_norm_pre_clip': 4.743409156799316, 'timing/data': 0.000509740028064698, 'timing/model': 1.4841783600277267, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 0.38}) +[2026-06-15 19:51:54,409][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.2456914782524109, 'train/grad_norm_pre_clip': 17.299999237060547, 'timing/data': 0.0007532989839091897, 'timing/model': 1.4836661759763956, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 0.38}) +[2026-06-15 19:52:00,260][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.3186725974082947, 'train/grad_norm_pre_clip': 17.893030166625977, 'timing/data': 0.00050279003335163, 'timing/model': 1.4398842509835958, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 0.38}) +[2026-06-15 19:52:06,167][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.21400302648544312, 'train/grad_norm_pre_clip': 11.796903610229492, 'timing/data': 0.0005081479903310537, 'timing/model': 1.673670444986783, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 0.38}) +[2026-06-15 19:52:11,858][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.33252307772636414, 'train/grad_norm_pre_clip': 14.189878463745117, 'timing/data': 0.0005154079990461469, 'timing/model': 1.464571543969214, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 0.38}) +[2026-06-15 19:52:17,860][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.08750171959400177, 'train/grad_norm_pre_clip': 2.543056011199951, 'timing/data': 0.0005166300106793642, 'timing/model': 1.4665946909808554, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 0.38}) +[2026-06-15 19:52:23,758][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.15469075739383698, 'train/grad_norm_pre_clip': 5.244675159454346, 'timing/data': 0.0005171910161152482, 'timing/model': 1.4614117669989355, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 0.38}) +[2026-06-15 19:52:29,660][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.15265275537967682, 'train/grad_norm_pre_clip': 4.70999002456665, 'timing/data': 0.0005047520389780402, 'timing/model': 1.4392189159989357, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 0.38}) +[2026-06-15 19:52:35,459][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.11030793935060501, 'train/grad_norm_pre_clip': 4.119363307952881, 'timing/data': 0.0005069960025139153, 'timing/model': 1.4632018420379609, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 0.38}) +[2026-06-15 19:52:41,489][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.08371942490339279, 'train/grad_norm_pre_clip': 2.7368788719177246, 'timing/data': 0.000506484997458756, 'timing/model': 1.6963739339844324, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 0.39}) +[2026-06-15 19:52:47,161][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.0632716715335846, 'train/grad_norm_pre_clip': 2.4710891246795654, 'timing/data': 0.0004931860021315515, 'timing/model': 1.4667397989542224, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 0.39}) +[2026-06-15 19:52:53,157][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.17104323208332062, 'train/grad_norm_pre_clip': 4.510529518127441, 'timing/data': 0.0004883190267719328, 'timing/model': 1.463078344007954, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 0.39}) +[2026-06-15 19:52:59,056][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.27303317189216614, 'train/grad_norm_pre_clip': 7.614484786987305, 'timing/data': 0.0004954289761371911, 'timing/model': 1.4581036970485002, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 0.39}) +[2026-06-15 19:53:05,033][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.18353411555290222, 'train/grad_norm_pre_clip': 8.287703514099121, 'timing/data': 0.0005064649740234017, 'timing/model': 1.5029865599935874, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 0.39}) +[2026-06-15 20:02:19,353][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.23149868845939636, 'train/grad_norm_pre_clip': 11.265621185302734, 'eval/loss': 0.09871678799390793, 'eval/flap_precision': 0.9030927835051547, 'eval/flap_recall': 0.8777555110220441, 'eval/flap_f1': 0.8902439024390244, 'eval/flap_rate_model': 0.13472222222222222, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.8902439024390244, 'timing/data': 0.00048806797713041306, 'timing/model': 1.4375909710070118, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 0.39}) +[2026-06-15 20:02:19,353][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_900_state +[2026-06-15 20:02:19,358][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:02:24,858][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_900_state/model.safetensors +[2026-06-15 20:02:38,332][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_900_state/optimizer.bin +[2026-06-15 20:02:38,333][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_900_state/sampler.bin +[2026-06-15 20:02:38,333][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_900_state/sampler_1.bin +[2026-06-15 20:02:38,334][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_900_state/random_states_0.pkl +[2026-06-15 20:02:38,446][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-06-15 20:02:38,447][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:02:38,456][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:02:44,436][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.05405091866850853, 'train/grad_norm_pre_clip': 3.8054943084716797, 'timing/data': 0.00052524299826473, 'timing/model': 1.4596182450186461, 'learning_rate/qwen_vl_interface': 9.387268552285024e-06, 'learning_rate/action_model': 9.387268552285025e-05, 'epoch': 0.39}) +[2026-06-15 20:02:50,370][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.19032539427280426, 'train/grad_norm_pre_clip': 7.688075065612793, 'timing/data': 0.0005357179907150567, 'timing/model': 1.6821484749671072, 'learning_rate/qwen_vl_interface': 9.385771559309125e-06, 'learning_rate/action_model': 9.385771559309125e-05, 'epoch': 0.39}) +[2026-06-15 20:02:56,273][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.09547258168458939, 'train/grad_norm_pre_clip': 5.050271987915039, 'timing/data': 0.0007422619964927435, 'timing/model': 1.6963026900193654, 'learning_rate/qwen_vl_interface': 9.38427286627376e-06, 'learning_rate/action_model': 9.38427286627376e-05, 'epoch': 0.39}) +[2026-06-15 20:03:01,945][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.41560807824134827, 'train/grad_norm_pre_clip': 11.210565567016602, 'timing/data': 0.0005239900201559067, 'timing/model': 1.4673731430084445, 'learning_rate/qwen_vl_interface': 9.382772473794985e-06, 'learning_rate/action_model': 9.382772473794985e-05, 'epoch': 0.39}) +[2026-06-15 20:03:08,135][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.11997874826192856, 'train/grad_norm_pre_clip': 8.892370223999023, 'timing/data': 0.0005651109968312085, 'timing/model': 1.572563610970974, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.39}) +[2026-06-15 20:03:14,051][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.15612637996673584, 'train/grad_norm_pre_clip': 10.773463249206543, 'timing/data': 0.0005186329944990575, 'timing/model': 1.469524341984652, 'learning_rate/qwen_vl_interface': 9.37976659297492e-06, 'learning_rate/action_model': 9.379766592974921e-05, 'epoch': 0.39}) +[2026-06-15 20:03:19,947][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.2659393548965454, 'train/grad_norm_pre_clip': 11.232690811157227, 'timing/data': 0.0005216869758442044, 'timing/model': 1.4329926960053854, 'learning_rate/qwen_vl_interface': 9.378261105869237e-06, 'learning_rate/action_model': 9.378261105869237e-05, 'epoch': 0.39}) +[2026-06-15 20:03:25,887][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.25555238127708435, 'train/grad_norm_pre_clip': 7.631953716278076, 'timing/data': 0.0005025489954277873, 'timing/model': 1.7052257899777032, 'learning_rate/qwen_vl_interface': 9.376753921791348e-06, 'learning_rate/action_model': 9.376753921791348e-05, 'epoch': 0.39}) +[2026-06-15 20:03:31,708][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.12674638628959656, 'train/grad_norm_pre_clip': 2.8642776012420654, 'timing/data': 0.0005222079926170409, 'timing/model': 1.6235641639796086, 'learning_rate/qwen_vl_interface': 9.375245041360805e-06, 'learning_rate/action_model': 9.375245041360804e-05, 'epoch': 0.39}) +[2026-06-15 20:03:37,348][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.19367878139019012, 'train/grad_norm_pre_clip': 9.583410263061523, 'timing/data': 0.0005160089931450784, 'timing/model': 1.463502935017459, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 0.39}) +[2026-06-15 20:03:43,248][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.07325989007949829, 'train/grad_norm_pre_clip': 3.447784900665283, 'timing/data': 0.0005188729846850038, 'timing/model': 1.4485157190356404, 'learning_rate/qwen_vl_interface': 9.372222193923421e-06, 'learning_rate/action_model': 9.372222193923421e-05, 'epoch': 0.39}) +[2026-06-15 20:03:49,150][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.14668375253677368, 'train/grad_norm_pre_clip': 7.280818939208984, 'timing/data': 0.0005164589965716004, 'timing/model': 1.4629940110025927, 'learning_rate/qwen_vl_interface': 9.370708228159158e-06, 'learning_rate/action_model': 9.370708228159159e-05, 'epoch': 0.39}) +[2026-06-15 20:03:55,233][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.25852400064468384, 'train/grad_norm_pre_clip': 9.537590026855469, 'timing/data': 0.0005158990388736129, 'timing/model': 1.4985922279884107, 'learning_rate/qwen_vl_interface': 9.369192568527398e-06, 'learning_rate/action_model': 9.369192568527398e-05, 'epoch': 0.39}) +[2026-06-15 20:04:01,183][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.20842258632183075, 'train/grad_norm_pre_clip': 8.568495750427246, 'timing/data': 0.0005194339901208878, 'timing/model': 1.6961289620376192, 'learning_rate/qwen_vl_interface': 9.367675215651168e-06, 'learning_rate/action_model': 9.367675215651168e-05, 'epoch': 0.39}) +[2026-06-15 20:04:06,851][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.14515183866024017, 'train/grad_norm_pre_clip': 7.519071102142334, 'timing/data': 0.0005820670048706234, 'timing/model': 1.4640324920183048, 'learning_rate/qwen_vl_interface': 9.366156170154194e-06, 'learning_rate/action_model': 9.366156170154194e-05, 'epoch': 0.39}) +[2026-06-15 20:04:12,932][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.2751675546169281, 'train/grad_norm_pre_clip': 6.748793601989746, 'timing/data': 0.0005155379767529666, 'timing/model': 1.543833152973093, 'learning_rate/qwen_vl_interface': 9.3646354326609e-06, 'learning_rate/action_model': 9.364635432660899e-05, 'epoch': 0.39}) +[2026-06-15 20:04:18,852][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.16281777620315552, 'train/grad_norm_pre_clip': 7.535518646240234, 'timing/data': 0.0005193440010771155, 'timing/model': 1.4568579050246626, 'learning_rate/qwen_vl_interface': 9.3631130037964e-06, 'learning_rate/action_model': 9.363113003796401e-05, 'epoch': 0.39}) +[2026-06-15 20:04:24,930][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.19689950346946716, 'train/grad_norm_pre_clip': 4.569006443023682, 'timing/data': 0.0005106809549033642, 'timing/model': 1.5985113520291634, 'learning_rate/qwen_vl_interface': 9.361588884186513e-06, 'learning_rate/action_model': 9.361588884186513e-05, 'epoch': 0.39}) +[2026-06-15 20:04:30,751][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.4342476427555084, 'train/grad_norm_pre_clip': 8.618387222290039, 'timing/data': 0.0005156989791430533, 'timing/model': 1.437577772012446, 'learning_rate/qwen_vl_interface': 9.360063074457744e-06, 'learning_rate/action_model': 9.360063074457743e-05, 'epoch': 0.4}) +[2026-06-15 20:04:36,640][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.08758722245693207, 'train/grad_norm_pre_clip': 2.4989922046661377, 'timing/data': 0.0005173810059204698, 'timing/model': 1.652079376974143, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 0.4}) +[2026-06-15 20:04:42,453][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.11379533261060715, 'train/grad_norm_pre_clip': 3.0375871658325195, 'timing/data': 0.000554656027816236, 'timing/model': 1.485651736962609, 'learning_rate/qwen_vl_interface': 9.357006387153065e-06, 'learning_rate/action_model': 9.357006387153065e-05, 'epoch': 0.4}) +[2026-06-15 20:04:48,452][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.06691966950893402, 'train/grad_norm_pre_clip': 1.8078711032867432, 'timing/data': 0.00051194301340729, 'timing/model': 1.4627376410062425, 'learning_rate/qwen_vl_interface': 9.355475510833643e-06, 'learning_rate/action_model': 9.355475510833643e-05, 'epoch': 0.4}) +[2026-06-15 20:04:54,354][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.12844839692115784, 'train/grad_norm_pre_clip': 3.272204875946045, 'timing/data': 0.0005258740275166929, 'timing/model': 1.4624674720107578, 'learning_rate/qwen_vl_interface': 9.353942946908318e-06, 'learning_rate/action_model': 9.353942946908317e-05, 'epoch': 0.4}) +[2026-06-15 20:05:00,257][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.1291232407093048, 'train/grad_norm_pre_clip': 3.9912235736846924, 'timing/data': 0.0005108120385557413, 'timing/model': 1.4336520369979553, 'learning_rate/qwen_vl_interface': 9.352408696007065e-06, 'learning_rate/action_model': 9.352408696007065e-05, 'epoch': 0.4}) +[2026-06-15 20:05:06,155][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.07379479706287384, 'train/grad_norm_pre_clip': 2.124537706375122, 'timing/data': 0.000519684050232172, 'timing/model': 1.4381262060487643, 'learning_rate/qwen_vl_interface': 9.350872758760558e-06, 'learning_rate/action_model': 9.350872758760559e-05, 'epoch': 0.4}) +[2026-06-15 20:05:12,069][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.09951416403055191, 'train/grad_norm_pre_clip': 3.6133785247802734, 'timing/data': 0.0005145869799889624, 'timing/model': 1.6791811340372078, 'learning_rate/qwen_vl_interface': 9.349335135800164e-06, 'learning_rate/action_model': 9.349335135800163e-05, 'epoch': 0.4}) +[2026-06-15 20:05:17,755][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.1087641790509224, 'train/grad_norm_pre_clip': 7.00615119934082, 'timing/data': 0.0005238099838607013, 'timing/model': 1.4661224790033884, 'learning_rate/qwen_vl_interface': 9.347795827757942e-06, 'learning_rate/action_model': 9.347795827757941e-05, 'epoch': 0.4}) +[2026-06-15 20:05:23,756][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.28599056601524353, 'train/grad_norm_pre_clip': 6.181658744812012, 'timing/data': 0.0005146869807504117, 'timing/model': 1.4671411660383455, 'learning_rate/qwen_vl_interface': 9.346254835266643e-06, 'learning_rate/action_model': 9.346254835266642e-05, 'epoch': 0.4}) +[2026-06-15 20:05:29,653][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.23785381019115448, 'train/grad_norm_pre_clip': 14.745841026306152, 'timing/data': 0.000519865017849952, 'timing/model': 1.4653724529780447, 'learning_rate/qwen_vl_interface': 9.344712158959709e-06, 'learning_rate/action_model': 9.344712158959709e-05, 'epoch': 0.4}) +[2026-06-15 20:05:35,553][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.32194170355796814, 'train/grad_norm_pre_clip': 7.235892295837402, 'timing/data': 0.0005120140267536044, 'timing/model': 1.432041235966608, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 0.4}) +[2026-06-15 20:05:41,354][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.270336389541626, 'train/grad_norm_pre_clip': 12.927411079406738, 'timing/data': 0.0005175109836272895, 'timing/model': 1.4617773609934375, 'learning_rate/qwen_vl_interface': 9.341621757436177e-06, 'learning_rate/action_model': 9.341621757436177e-05, 'epoch': 0.4}) +[2026-06-15 20:05:47,288][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.2000131756067276, 'train/grad_norm_pre_clip': 11.892243385314941, 'timing/data': 0.0005023890407755971, 'timing/model': 1.687803563021589, 'learning_rate/qwen_vl_interface': 9.340074033489924e-06, 'learning_rate/action_model': 9.340074033489923e-05, 'epoch': 0.4}) +[2026-06-15 20:05:52,952][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.18948888778686523, 'train/grad_norm_pre_clip': 13.60561466217041, 'timing/data': 0.0004925049724988639, 'timing/model': 1.4650028109899722, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.4}) +[2026-06-15 20:05:59,053][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.16715186834335327, 'train/grad_norm_pre_clip': 8.74058723449707, 'timing/data': 0.0004965309635736048, 'timing/model': 1.4687379130045883, 'learning_rate/qwen_vl_interface': 9.336973542409505e-06, 'learning_rate/action_model': 9.336973542409504e-05, 'epoch': 0.4}) +[2026-06-15 20:06:04,952][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.10410763323307037, 'train/grad_norm_pre_clip': 3.4272966384887695, 'timing/data': 0.0004983230028301477, 'timing/model': 1.465237723023165, 'learning_rate/qwen_vl_interface': 9.33542077654983e-06, 'learning_rate/action_model': 9.33542077654983e-05, 'epoch': 0.4}) +[2026-06-15 20:06:11,058][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.3232218623161316, 'train/grad_norm_pre_clip': 16.804752349853516, 'timing/data': 0.0004942470113746822, 'timing/model': 1.4383362340158783, 'learning_rate/qwen_vl_interface': 9.333866331327995e-06, 'learning_rate/action_model': 9.333866331327995e-05, 'epoch': 0.4}) +[2026-06-15 20:06:16,954][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.2168314903974533, 'train/grad_norm_pre_clip': 11.968058586120605, 'timing/data': 0.0004890589625574648, 'timing/model': 1.442289257014636, 'learning_rate/qwen_vl_interface': 9.332310207382972e-06, 'learning_rate/action_model': 9.332310207382972e-05, 'epoch': 0.4}) +[2026-06-15 20:06:22,992][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.1993655264377594, 'train/grad_norm_pre_clip': 15.909320831298828, 'timing/data': 0.0004898710176348686, 'timing/model': 1.7028493889956735, 'learning_rate/qwen_vl_interface': 9.330752405354426e-06, 'learning_rate/action_model': 9.330752405354426e-05, 'epoch': 0.4}) +[2026-06-15 20:06:28,652][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.09224458783864975, 'train/grad_norm_pre_clip': 4.648028373718262, 'timing/data': 0.0004885380039922893, 'timing/model': 1.4652766429935582, 'learning_rate/qwen_vl_interface': 9.32919292588271e-06, 'learning_rate/action_model': 9.32919292588271e-05, 'epoch': 0.4}) +[2026-06-15 20:06:34,654][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.18668420612812042, 'train/grad_norm_pre_clip': 7.644501686096191, 'timing/data': 0.0004954389878548682, 'timing/model': 1.4640025169937871, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.4}) +[2026-06-15 20:06:40,555][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.33072665333747864, 'train/grad_norm_pre_clip': 9.687135696411133, 'timing/data': 0.0004843619535677135, 'timing/model': 1.463701923028566, 'learning_rate/qwen_vl_interface': 9.326068937174627e-06, 'learning_rate/action_model': 9.326068937174626e-05, 'epoch': 0.4}) +[2026-06-15 20:06:46,458][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.1411510407924652, 'train/grad_norm_pre_clip': 2.9608395099639893, 'timing/data': 0.0005037910304963589, 'timing/model': 1.4330887160031125, 'learning_rate/qwen_vl_interface': 9.324504429222412e-06, 'learning_rate/action_model': 9.324504429222412e-05, 'epoch': 0.41}) +[2026-06-15 20:06:52,389][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.3410741686820984, 'train/grad_norm_pre_clip': 7.784010410308838, 'timing/data': 0.0005041020340286195, 'timing/model': 1.6976621189969592, 'learning_rate/qwen_vl_interface': 9.322938246395336e-06, 'learning_rate/action_model': 9.322938246395337e-05, 'epoch': 0.41}) +[2026-06-15 20:06:58,209][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.20730403065681458, 'train/grad_norm_pre_clip': 5.474737167358398, 'timing/data': 0.0005244220374152064, 'timing/model': 1.615213661978487, 'learning_rate/qwen_vl_interface': 9.321370389337196e-06, 'learning_rate/action_model': 9.321370389337196e-05, 'epoch': 0.41}) +[2026-06-15 20:07:03,855][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.18188484013080597, 'train/grad_norm_pre_clip': 4.5199785232543945, 'timing/data': 0.0005580100114457309, 'timing/model': 1.465280067990534, 'learning_rate/qwen_vl_interface': 9.319800858692478e-06, 'learning_rate/action_model': 9.319800858692476e-05, 'epoch': 0.41}) +[2026-06-15 20:07:09,755][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.14891645312309265, 'train/grad_norm_pre_clip': 3.4814023971557617, 'timing/data': 0.0005205860361456871, 'timing/model': 1.465263001038693, 'learning_rate/qwen_vl_interface': 9.318229655106357e-06, 'learning_rate/action_model': 9.318229655106356e-05, 'epoch': 0.41}) +[2026-06-15 20:07:15,657][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.1752299666404724, 'train/grad_norm_pre_clip': 3.7696099281311035, 'timing/data': 0.0005042619886808097, 'timing/model': 1.4589856370002963, 'learning_rate/qwen_vl_interface': 9.316656779224693e-06, 'learning_rate/action_model': 9.316656779224693e-05, 'epoch': 0.41}) +[2026-06-15 20:07:21,637][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.1429044008255005, 'train/grad_norm_pre_clip': 3.8471555709838867, 'timing/data': 0.00051911297487095, 'timing/model': 1.5073714909958653, 'learning_rate/qwen_vl_interface': 9.31508223169404e-06, 'learning_rate/action_model': 9.31508223169404e-05, 'epoch': 0.41}) +[2026-06-15 20:07:27,592][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.16703104972839355, 'train/grad_norm_pre_clip': 3.151979446411133, 'timing/data': 0.0005301089840941131, 'timing/model': 1.6951196180307306, 'learning_rate/qwen_vl_interface': 9.313506013161633e-06, 'learning_rate/action_model': 9.313506013161633e-05, 'epoch': 0.41}) +[2026-06-15 20:07:33,253][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.2051718384027481, 'train/grad_norm_pre_clip': 2.9540653228759766, 'timing/data': 0.0005108810146339238, 'timing/model': 1.462191207043361, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 0.41}) +[2026-06-15 20:07:39,254][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.10990878194570541, 'train/grad_norm_pre_clip': 2.652055025100708, 'timing/data': 0.0005160190048627555, 'timing/model': 1.4646041640080512, 'learning_rate/qwen_vl_interface': 9.310348565683938e-06, 'learning_rate/action_model': 9.310348565683939e-05, 'epoch': 0.41}) +[2026-06-15 20:07:45,159][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.1840815246105194, 'train/grad_norm_pre_clip': 2.5336849689483643, 'timing/data': 0.0005295190494507551, 'timing/model': 1.4702490349882282, 'learning_rate/qwen_vl_interface': 9.30876733803656e-06, 'learning_rate/action_model': 9.308767338036559e-05, 'epoch': 0.41}) +[2026-06-15 20:07:51,054][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.1455203741788864, 'train/grad_norm_pre_clip': 5.576025485992432, 'timing/data': 0.0005133249796926975, 'timing/model': 1.4348499579937197, 'learning_rate/qwen_vl_interface': 9.30718444198324e-06, 'learning_rate/action_model': 9.307184441983239e-05, 'epoch': 0.41}) +[2026-06-15 20:07:56,952][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.07196776568889618, 'train/grad_norm_pre_clip': 2.656081438064575, 'timing/data': 0.0005021290271542966, 'timing/model': 1.4358884639805183, 'learning_rate/qwen_vl_interface': 9.30559987817465e-06, 'learning_rate/action_model': 9.30559987817465e-05, 'epoch': 0.41}) +[2026-06-15 20:08:02,851][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.22266453504562378, 'train/grad_norm_pre_clip': 10.682494163513184, 'timing/data': 0.00048664602218195796, 'timing/model': 1.6633597759646364, 'learning_rate/qwen_vl_interface': 9.304013647262143e-06, 'learning_rate/action_model': 9.304013647262143e-05, 'epoch': 0.41}) +[2026-06-15 20:08:08,555][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.035267505794763565, 'train/grad_norm_pre_clip': 1.3277664184570312, 'timing/data': 0.0005091080092824996, 'timing/model': 1.464502235990949, 'learning_rate/qwen_vl_interface': 9.302425749897761e-06, 'learning_rate/action_model': 9.302425749897761e-05, 'epoch': 0.41}) +[2026-06-15 20:08:14,552][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.1640150249004364, 'train/grad_norm_pre_clip': 4.960310935974121, 'timing/data': 0.0005198550061322749, 'timing/model': 1.463113481993787, 'learning_rate/qwen_vl_interface': 9.300836186734225e-06, 'learning_rate/action_model': 9.300836186734226e-05, 'epoch': 0.41}) +[2026-06-15 20:08:20,455][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.33973515033721924, 'train/grad_norm_pre_clip': 7.241103172302246, 'timing/data': 0.0006374369841068983, 'timing/model': 1.4639415390556678, 'learning_rate/qwen_vl_interface': 9.299244958424947e-06, 'learning_rate/action_model': 9.299244958424946e-05, 'epoch': 0.41}) +[2026-06-15 20:08:26,354][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.3476804196834564, 'train/grad_norm_pre_clip': 6.276562213897705, 'timing/data': 0.0005088680190965533, 'timing/model': 1.4373059020144865, 'learning_rate/qwen_vl_interface': 9.29765206562402e-06, 'learning_rate/action_model': 9.297652065624018e-05, 'epoch': 0.41}) +[2026-06-15 20:08:32,257][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.12544657289981842, 'train/grad_norm_pre_clip': 6.639837741851807, 'timing/data': 0.0005181620363146067, 'timing/model': 1.4422628810280003, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 0.41}) +[2026-06-15 20:08:38,166][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.20820294320583344, 'train/grad_norm_pre_clip': 12.737178802490234, 'timing/data': 0.0005041819531470537, 'timing/model': 1.6737265639821999, 'learning_rate/qwen_vl_interface': 9.294461289167013e-06, 'learning_rate/action_model': 9.294461289167014e-05, 'epoch': 0.41}) +[2026-06-15 20:08:43,854][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.3705348074436188, 'train/grad_norm_pre_clip': 19.71045684814453, 'timing/data': 0.0004952980089001358, 'timing/model': 1.4656437269877642, 'learning_rate/qwen_vl_interface': 9.292863406822543e-06, 'learning_rate/action_model': 9.292863406822542e-05, 'epoch': 0.41}) +[2026-06-15 20:08:49,850][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.3360830545425415, 'train/grad_norm_pre_clip': 13.948684692382812, 'timing/data': 0.0005734239821322262, 'timing/model': 1.4512058510445058, 'learning_rate/qwen_vl_interface': 9.291263862609638e-06, 'learning_rate/action_model': 9.291263862609638e-05, 'epoch': 0.41}) +[2026-06-15 20:08:55,848][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.2814544141292572, 'train/grad_norm_pre_clip': 11.603987693786621, 'timing/data': 0.0005552470101974905, 'timing/model': 1.4631326110102236, 'learning_rate/qwen_vl_interface': 9.28966265718581e-06, 'learning_rate/action_model': 9.28966265718581e-05, 'epoch': 0.41}) +[2026-06-15 20:09:01,751][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.12151124328374863, 'train/grad_norm_pre_clip': 2.6763648986816406, 'timing/data': 0.0005704489885829389, 'timing/model': 1.4357466400251724, 'learning_rate/qwen_vl_interface': 9.288059791209256e-06, 'learning_rate/action_model': 9.288059791209256e-05, 'epoch': 0.42}) +[2026-06-15 20:09:07,552][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.06476522982120514, 'train/grad_norm_pre_clip': 7.979267120361328, 'timing/data': 0.0005002360558137298, 'timing/model': 1.4637066629948094, 'learning_rate/qwen_vl_interface': 9.286455265338853e-06, 'learning_rate/action_model': 9.286455265338853e-05, 'epoch': 0.42}) +[2026-06-15 20:09:13,483][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.2831689417362213, 'train/grad_norm_pre_clip': 12.874783515930176, 'timing/data': 0.0005129339988343418, 'timing/model': 1.6940849669626914, 'learning_rate/qwen_vl_interface': 9.284849080234161e-06, 'learning_rate/action_model': 9.28484908023416e-05, 'epoch': 0.42}) +[2026-06-15 20:09:19,153][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.13322840631008148, 'train/grad_norm_pre_clip': 7.141074180603027, 'timing/data': 0.0005149469943717122, 'timing/model': 1.4640235360129736, 'learning_rate/qwen_vl_interface': 9.283241236555421e-06, 'learning_rate/action_model': 9.283241236555421e-05, 'epoch': 0.42}) +[2026-06-15 20:09:25,153][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.410785049200058, 'train/grad_norm_pre_clip': 12.6067533493042, 'timing/data': 0.00051369599532336, 'timing/model': 1.4639181100064889, 'learning_rate/qwen_vl_interface': 9.281631734963557e-06, 'learning_rate/action_model': 9.281631734963557e-05, 'epoch': 0.42}) +[2026-06-15 20:09:31,057][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.12537997961044312, 'train/grad_norm_pre_clip': 13.326355934143066, 'timing/data': 0.0005084670265205204, 'timing/model': 1.4640405980171636, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 0.42}) +[2026-06-15 20:09:36,957][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.23000459372997284, 'train/grad_norm_pre_clip': 7.638133525848389, 'timing/data': 0.0005127639742568135, 'timing/model': 1.4371135339606553, 'learning_rate/qwen_vl_interface': 9.278407760687562e-06, 'learning_rate/action_model': 9.278407760687562e-05, 'epoch': 0.42}) +[2026-06-15 20:09:42,856][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.2741856276988983, 'train/grad_norm_pre_clip': 7.35741662979126, 'timing/data': 0.0004979619989171624, 'timing/model': 1.4421289690071717, 'learning_rate/qwen_vl_interface': 9.276793289328686e-06, 'learning_rate/action_model': 9.276793289328685e-05, 'epoch': 0.42}) +[2026-06-15 20:09:48,785][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.12936313450336456, 'train/grad_norm_pre_clip': 5.61792516708374, 'timing/data': 0.0005150579963810742, 'timing/model': 1.6808611190062948, 'learning_rate/qwen_vl_interface': 9.27517716270719e-06, 'learning_rate/action_model': 9.275177162707191e-05, 'epoch': 0.42}) +[2026-06-15 20:09:54,456][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.1946343183517456, 'train/grad_norm_pre_clip': 3.6134448051452637, 'timing/data': 0.0005077930400148034, 'timing/model': 1.452765302034095, 'learning_rate/qwen_vl_interface': 9.273559381487409e-06, 'learning_rate/action_model': 9.273559381487409e-05, 'epoch': 0.42}) +[2026-06-15 20:10:00,457][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.23430635035037994, 'train/grad_norm_pre_clip': 4.519742488861084, 'timing/data': 0.0005122990114614367, 'timing/model': 1.4673467809916474, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.42}) +[2026-06-15 20:10:06,351][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.07862057536840439, 'train/grad_norm_pre_clip': 4.334444046020508, 'timing/data': 0.0005222430336289108, 'timing/model': 1.4614721510442905, 'learning_rate/qwen_vl_interface': 9.2703188579137e-06, 'learning_rate/action_model': 9.2703188579137e-05, 'epoch': 0.42}) +[2026-06-15 20:10:12,317][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.16687849164009094, 'train/grad_norm_pre_clip': 4.618782997131348, 'timing/data': 0.0005014130147174001, 'timing/model': 1.4863003519712947, 'learning_rate/qwen_vl_interface': 9.26869611689183e-06, 'learning_rate/action_model': 9.26869611689183e-05, 'epoch': 0.42}) +[2026-06-15 20:10:18,183][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.31011441349983215, 'train/grad_norm_pre_clip': 5.896462917327881, 'timing/data': 0.0005139110144227743, 'timing/model': 1.6886478060041554, 'learning_rate/qwen_vl_interface': 9.267071723935784e-06, 'learning_rate/action_model': 9.267071723935784e-05, 'epoch': 0.42}) +[2026-06-15 20:10:24,001][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.2475070357322693, 'train/grad_norm_pre_clip': 10.11884880065918, 'timing/data': 0.0004960849764756858, 'timing/model': 1.6113685159943998, 'learning_rate/qwen_vl_interface': 9.26544567971329e-06, 'learning_rate/action_model': 9.26544567971329e-05, 'epoch': 0.42}) +[2026-06-15 20:10:29,654][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.12931658327579498, 'train/grad_norm_pre_clip': 2.8518900871276855, 'timing/data': 0.0004928999696858227, 'timing/model': 1.4638697790214792, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 0.42}) +[2026-06-15 20:10:35,554][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.17771592736244202, 'train/grad_norm_pre_clip': 5.850144863128662, 'timing/data': 0.0004943719832226634, 'timing/model': 1.4635111810057424, 'learning_rate/qwen_vl_interface': 9.262188640143262e-06, 'learning_rate/action_model': 9.262188640143262e-05, 'epoch': 0.42}) +[2026-06-15 20:10:41,455][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.20048214495182037, 'train/grad_norm_pre_clip': 6.389613151550293, 'timing/data': 0.0005051679909229279, 'timing/model': 1.4663027409696952, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.42}) +[2026-06-15 20:10:47,355][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.36885449290275574, 'train/grad_norm_pre_clip': 12.09506607055664, 'timing/data': 0.0005608510109595954, 'timing/model': 1.4329545720247552, 'learning_rate/qwen_vl_interface': 9.258925003537127e-06, 'learning_rate/action_model': 9.258925003537126e-05, 'epoch': 0.42}) +[2026-06-15 20:10:53,286][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.19692905247211456, 'train/grad_norm_pre_clip': 6.418405532836914, 'timing/data': 0.00048627000069245696, 'timing/model': 1.696253481961321, 'learning_rate/qwen_vl_interface': 9.257290713022041e-06, 'learning_rate/action_model': 9.257290713022041e-05, 'epoch': 0.42}) +[2026-06-15 20:10:58,955][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.0937630832195282, 'train/grad_norm_pre_clip': 7.050942420959473, 'timing/data': 0.0005417520296759903, 'timing/model': 1.4665011210017838, 'learning_rate/qwen_vl_interface': 9.255654775261112e-06, 'learning_rate/action_model': 9.255654775261112e-05, 'epoch': 0.42}) +[2026-06-15 20:11:04,955][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.04528801515698433, 'train/grad_norm_pre_clip': 2.0586941242218018, 'timing/data': 0.0005468589952215552, 'timing/model': 1.464957563963253, 'learning_rate/qwen_vl_interface': 9.254017190926813e-06, 'learning_rate/action_model': 9.254017190926813e-05, 'epoch': 0.42}) +[2026-06-15 20:11:10,855][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.15898671746253967, 'train/grad_norm_pre_clip': 3.620159387588501, 'timing/data': 0.000520802044775337, 'timing/model': 1.4647629889659584, 'learning_rate/qwen_vl_interface': 9.25237796069229e-06, 'learning_rate/action_model': 9.25237796069229e-05, 'epoch': 0.42}) +[2026-06-15 20:11:16,752][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.23809608817100525, 'train/grad_norm_pre_clip': 4.776183128356934, 'timing/data': 0.0005170449730940163, 'timing/model': 1.4394026389927603, 'learning_rate/qwen_vl_interface': 9.250737085231371e-06, 'learning_rate/action_model': 9.250737085231372e-05, 'epoch': 0.43}) +[2026-06-15 20:11:22,655][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.11847731471061707, 'train/grad_norm_pre_clip': 4.428424835205078, 'timing/data': 0.0005099849659018219, 'timing/model': 1.4417722669895738, 'learning_rate/qwen_vl_interface': 9.24909456521856e-06, 'learning_rate/action_model': 9.249094565218559e-05, 'epoch': 0.43}) +[2026-06-15 20:11:28,541][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.366795152425766, 'train/grad_norm_pre_clip': 6.135815143585205, 'timing/data': 0.0005085730226710439, 'timing/model': 1.6489092750125565, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 0.43}) +[2026-06-15 20:11:34,255][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.2420749068260193, 'train/grad_norm_pre_clip': 6.947270393371582, 'timing/data': 0.0005042160046286881, 'timing/model': 1.4641629079706036, 'learning_rate/qwen_vl_interface': 9.245804594238637e-06, 'learning_rate/action_model': 9.245804594238636e-05, 'epoch': 0.43}) +[2026-06-15 20:11:40,354][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.1598639041185379, 'train/grad_norm_pre_clip': 10.225171089172363, 'timing/data': 0.0005189679795876145, 'timing/model': 1.4914458160055801, 'learning_rate/qwen_vl_interface': 9.244157144623909e-06, 'learning_rate/action_model': 9.244157144623909e-05, 'epoch': 0.43}) +[2026-06-15 20:11:46,256][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.1987479031085968, 'train/grad_norm_pre_clip': 9.755160331726074, 'timing/data': 0.0005155929829925299, 'timing/model': 1.465215968026314, 'learning_rate/qwen_vl_interface': 9.242508053162051e-06, 'learning_rate/action_model': 9.242508053162051e-05, 'epoch': 0.43}) +[2026-06-15 20:11:52,156][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.29060015082359314, 'train/grad_norm_pre_clip': 8.86298656463623, 'timing/data': 0.0005167950293980539, 'timing/model': 1.4396301400265656, 'learning_rate/qwen_vl_interface': 9.240857320530944e-06, 'learning_rate/action_model': 9.240857320530943e-05, 'epoch': 0.43}) +[2026-06-15 20:11:58,158][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.08524908870458603, 'train/grad_norm_pre_clip': 3.6824088096618652, 'timing/data': 0.0005185069749131799, 'timing/model': 1.4338176359888166, 'learning_rate/qwen_vl_interface': 9.23920494740914e-06, 'learning_rate/action_model': 9.23920494740914e-05, 'epoch': 0.43}) +[2026-06-15 20:12:04,062][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.2815276086330414, 'train/grad_norm_pre_clip': 8.631816864013672, 'timing/data': 0.0005080309929326177, 'timing/model': 1.6736800229991786, 'learning_rate/qwen_vl_interface': 9.237550934475868e-06, 'learning_rate/action_model': 9.237550934475867e-05, 'epoch': 0.43}) +[2026-06-15 20:12:09,752][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.0735652893781662, 'train/grad_norm_pre_clip': 3.9595887660980225, 'timing/data': 0.000506969983689487, 'timing/model': 1.4641465790336952, 'learning_rate/qwen_vl_interface': 9.235895282411029e-06, 'learning_rate/action_model': 9.235895282411029e-05, 'epoch': 0.43}) +[2026-06-15 20:12:15,754][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.11716997623443604, 'train/grad_norm_pre_clip': 4.556501388549805, 'timing/data': 0.0005164939793758094, 'timing/model': 1.4651453359983861, 'learning_rate/qwen_vl_interface': 9.2342379918952e-06, 'learning_rate/action_model': 9.2342379918952e-05, 'epoch': 0.43}) +[2026-06-15 20:12:21,654][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.09109088778495789, 'train/grad_norm_pre_clip': 1.7819578647613525, 'timing/data': 0.00048756098840385675, 'timing/model': 1.4635512589593418, 'learning_rate/qwen_vl_interface': 9.232579063609631e-06, 'learning_rate/action_model': 9.232579063609631e-05, 'epoch': 0.43}) +[2026-06-15 20:21:33,151][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.2645851969718933, 'train/grad_norm_pre_clip': 12.676923751831055, 'eval/loss': 0.08977678418159485, 'eval/flap_precision': 0.8822355289421158, 'eval/flap_recall': 0.8857715430861723, 'eval/flap_f1': 0.884, 'eval/flap_rate_model': 0.13916666666666666, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.884, 'timing/data': 0.0005069599719718099, 'timing/model': 1.4355229499633424, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 0.43}) +[2026-06-15 20:21:33,152][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1000_state +[2026-06-15 20:21:33,156][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:21:38,654][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1000_state/model.safetensors +[2026-06-15 20:21:46,722][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1000_state/optimizer.bin +[2026-06-15 20:21:46,722][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1000_state/sampler.bin +[2026-06-15 20:21:46,722][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-15 20:21:46,723][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-15 20:21:46,840][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-15 20:21:46,841][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:21:46,851][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:21:52,733][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.21795214712619781, 'train/grad_norm_pre_clip': 7.362494945526123, 'timing/data': 0.000518894987180829, 'timing/model': 1.464567819959484, 'learning_rate/qwen_vl_interface': 9.229256296457635e-06, 'learning_rate/action_model': 9.229256296457634e-05, 'epoch': 0.43}) +[2026-06-15 20:21:58,637][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.21489351987838745, 'train/grad_norm_pre_clip': 11.119274139404297, 'timing/data': 0.0005083290161564946, 'timing/model': 1.4323489269590937, 'learning_rate/qwen_vl_interface': 9.22759245895707e-06, 'learning_rate/action_model': 9.227592458957071e-05, 'epoch': 0.43}) +[2026-06-15 20:22:04,588][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.14132429659366608, 'train/grad_norm_pre_clip': 6.580606460571289, 'timing/data': 0.0006306099821813405, 'timing/model': 1.7119523599976674, 'learning_rate/qwen_vl_interface': 9.225926986418494e-06, 'learning_rate/action_model': 9.225926986418494e-05, 'epoch': 0.43}) +[2026-06-15 20:22:10,336][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.024493655189871788, 'train/grad_norm_pre_clip': 1.4729644060134888, 'timing/data': 0.0005464549758471549, 'timing/model': 1.4666415369720198, 'learning_rate/qwen_vl_interface': 9.224259879526519e-06, 'learning_rate/action_model': 9.224259879526519e-05, 'epoch': 0.43}) +[2026-06-15 20:22:16,339][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.12860596179962158, 'train/grad_norm_pre_clip': 5.761250019073486, 'timing/data': 0.0005354189779609442, 'timing/model': 1.4641516039846465, 'learning_rate/qwen_vl_interface': 9.222591138966427e-06, 'learning_rate/action_model': 9.222591138966426e-05, 'epoch': 0.43}) +[2026-06-15 20:22:22,444][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.1961587816476822, 'train/grad_norm_pre_clip': 9.182181358337402, 'timing/data': 0.0005600159638561308, 'timing/model': 1.4850996269960888, 'learning_rate/qwen_vl_interface': 9.220920765424175e-06, 'learning_rate/action_model': 9.220920765424174e-05, 'epoch': 0.43}) +[2026-06-15 20:22:28,342][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.4523051977157593, 'train/grad_norm_pre_clip': 15.920502662658691, 'timing/data': 0.0005169420037418604, 'timing/model': 1.4647576590068638, 'learning_rate/qwen_vl_interface': 9.219248759586392e-06, 'learning_rate/action_model': 9.219248759586391e-05, 'epoch': 0.43}) +[2026-06-15 20:22:34,245][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.10900288820266724, 'train/grad_norm_pre_clip': 5.160008907318115, 'timing/data': 0.0005116650136187673, 'timing/model': 1.4314933120040223, 'learning_rate/qwen_vl_interface': 9.217575122140375e-06, 'learning_rate/action_model': 9.217575122140375e-05, 'epoch': 0.43}) +[2026-06-15 20:22:40,178][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.24881163239479065, 'train/grad_norm_pre_clip': 7.730912208557129, 'timing/data': 0.0005127059994265437, 'timing/model': 1.692576740984805, 'learning_rate/qwen_vl_interface': 9.215899853774094e-06, 'learning_rate/action_model': 9.215899853774093e-05, 'epoch': 0.43}) +[2026-06-15 20:22:45,846][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.1737741231918335, 'train/grad_norm_pre_clip': 5.868114471435547, 'timing/data': 0.0005296109593473375, 'timing/model': 1.462538304971531, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.43}) +[2026-06-15 20:22:51,745][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.3612056076526642, 'train/grad_norm_pre_clip': 7.0711283683776855, 'timing/data': 0.0005069070030003786, 'timing/model': 1.4633087940164842, 'learning_rate/qwen_vl_interface': 9.212544427035967e-06, 'learning_rate/action_model': 9.212544427035967e-05, 'epoch': 0.44}) +[2026-06-15 20:22:57,548][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.17913807928562164, 'train/grad_norm_pre_clip': 3.4019253253936768, 'timing/data': 0.0005123850423842669, 'timing/model': 1.4680739819887094, 'learning_rate/qwen_vl_interface': 9.21086427004341e-06, 'learning_rate/action_model': 9.210864270043411e-05, 'epoch': 0.44}) +[2026-06-15 20:23:03,446][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.263004332780838, 'train/grad_norm_pre_clip': 7.737522602081299, 'timing/data': 0.0005122750299051404, 'timing/model': 1.4634064349811524, 'learning_rate/qwen_vl_interface': 9.209182484889168e-06, 'learning_rate/action_model': 9.209182484889167e-05, 'epoch': 0.44}) +[2026-06-15 20:23:09,431][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.12053775042295456, 'train/grad_norm_pre_clip': 7.460049629211426, 'timing/data': 0.0005180829903110862, 'timing/model': 1.5011624229955487, 'learning_rate/qwen_vl_interface': 9.207499072264557e-06, 'learning_rate/action_model': 9.207499072264558e-05, 'epoch': 0.44}) +[2026-06-15 20:23:15,379][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.2032490372657776, 'train/grad_norm_pre_clip': 8.071569442749023, 'timing/data': 0.0005114839877933264, 'timing/model': 1.6910789019893855, 'learning_rate/qwen_vl_interface': 9.205814032861566e-06, 'learning_rate/action_model': 9.205814032861566e-05, 'epoch': 0.44}) +[2026-06-15 20:23:21,277][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.23139524459838867, 'train/grad_norm_pre_clip': 4.7552995681762695, 'timing/data': 0.0005128360353410244, 'timing/model': 1.6884203430381604, 'learning_rate/qwen_vl_interface': 9.204127367372849e-06, 'learning_rate/action_model': 9.204127367372849e-05, 'epoch': 0.44}) +[2026-06-15 20:23:26,954][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.2703090310096741, 'train/grad_norm_pre_clip': 6.745294570922852, 'timing/data': 0.0005026409635320306, 'timing/model': 1.4650005150469951, 'learning_rate/qwen_vl_interface': 9.202439076491732e-06, 'learning_rate/action_model': 9.202439076491732e-05, 'epoch': 0.44}) +[2026-06-15 20:23:32,847][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.05739513039588928, 'train/grad_norm_pre_clip': 1.9164406061172485, 'timing/data': 0.0005043029668740928, 'timing/model': 1.4655431309947744, 'learning_rate/qwen_vl_interface': 9.200749160912208e-06, 'learning_rate/action_model': 9.200749160912208e-05, 'epoch': 0.44}) +[2026-06-15 20:23:38,748][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.1292579621076584, 'train/grad_norm_pre_clip': 3.0005764961242676, 'timing/data': 0.00048057903768494725, 'timing/model': 1.4652996879885904, 'learning_rate/qwen_vl_interface': 9.199057621328937e-06, 'learning_rate/action_model': 9.199057621328935e-05, 'epoch': 0.44}) +[2026-06-15 20:23:44,653][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.12240567058324814, 'train/grad_norm_pre_clip': 3.691113233566284, 'timing/data': 0.0004855950246565044, 'timing/model': 1.43237004900584, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 0.44}) +[2026-06-15 20:23:50,582][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.21049003303050995, 'train/grad_norm_pre_clip': 7.1834821701049805, 'timing/data': 0.0005106719909235835, 'timing/model': 1.6846116550150327, 'learning_rate/qwen_vl_interface': 9.19566967293313e-06, 'learning_rate/action_model': 9.19566967293313e-05, 'epoch': 0.44}) +[2026-06-15 20:23:56,250][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.31375694274902344, 'train/grad_norm_pre_clip': 10.11105728149414, 'timing/data': 0.0005063170101493597, 'timing/model': 1.4521856810315512, 'learning_rate/qwen_vl_interface': 9.193973265513253e-06, 'learning_rate/action_model': 9.193973265513253e-05, 'epoch': 0.44}) +[2026-06-15 20:24:02,250][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.2480648308992386, 'train/grad_norm_pre_clip': 9.919572830200195, 'timing/data': 0.0005086499731987715, 'timing/model': 1.4529765350162052, 'learning_rate/qwen_vl_interface': 9.192275236874946e-06, 'learning_rate/action_model': 9.192275236874946e-05, 'epoch': 0.44}) +[2026-06-15 20:24:08,255][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.2231041043996811, 'train/grad_norm_pre_clip': 7.731393337249756, 'timing/data': 0.0005172829842194915, 'timing/model': 1.464460298011545, 'learning_rate/qwen_vl_interface': 9.190575587716199e-06, 'learning_rate/action_model': 9.190575587716199e-05, 'epoch': 0.44}) +[2026-06-15 20:24:14,254][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.06531400233507156, 'train/grad_norm_pre_clip': 2.8660173416137695, 'timing/data': 0.000506637035869062, 'timing/model': 1.4686794940498658, 'learning_rate/qwen_vl_interface': 9.188874318735677e-06, 'learning_rate/action_model': 9.188874318735676e-05, 'epoch': 0.44}) +[2026-06-15 20:24:20,154][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.25828468799591064, 'train/grad_norm_pre_clip': 10.788354873657227, 'timing/data': 0.0005639209994114935, 'timing/model': 1.4692845239769667, 'learning_rate/qwen_vl_interface': 9.187171430632707e-06, 'learning_rate/action_model': 9.187171430632707e-05, 'epoch': 0.44}) +[2026-06-15 20:24:26,080][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.28543198108673096, 'train/grad_norm_pre_clip': 16.553260803222656, 'timing/data': 0.0005031019682064652, 'timing/model': 1.6919804259669036, 'learning_rate/qwen_vl_interface': 9.185466924107283e-06, 'learning_rate/action_model': 9.185466924107282e-05, 'epoch': 0.44}) +[2026-06-15 20:24:31,750][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.12502296268939972, 'train/grad_norm_pre_clip': 6.9671759605407715, 'timing/data': 0.0005041129770688713, 'timing/model': 1.4631270589889027, 'learning_rate/qwen_vl_interface': 9.18376079986006e-06, 'learning_rate/action_model': 9.183760799860061e-05, 'epoch': 0.44}) +[2026-06-15 20:24:37,750][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.36447256803512573, 'train/grad_norm_pre_clip': 13.639579772949219, 'timing/data': 0.0005045129801146686, 'timing/model': 1.463845348043833, 'learning_rate/qwen_vl_interface': 9.182053058592365e-06, 'learning_rate/action_model': 9.182053058592365e-05, 'epoch': 0.44}) +[2026-06-15 20:24:43,675][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.3230043649673462, 'train/grad_norm_pre_clip': 17.30243682861328, 'timing/data': 0.000510070996824652, 'timing/model': 1.449169489031192, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 0.44}) +[2026-06-15 20:24:49,558][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.20131449401378632, 'train/grad_norm_pre_clip': 7.453866004943848, 'timing/data': 0.000507908989675343, 'timing/model': 1.473296465992462, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.44}) +[2026-06-15 20:24:55,535][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.13450349867343903, 'train/grad_norm_pre_clip': 2.5217912197113037, 'timing/data': 0.0005149689968675375, 'timing/model': 1.4983830460114405, 'learning_rate/qwen_vl_interface': 9.176920139689638e-06, 'learning_rate/action_model': 9.176920139689637e-05, 'epoch': 0.44}) +[2026-06-15 20:25:01,481][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.06856676936149597, 'train/grad_norm_pre_clip': 4.373998641967773, 'timing/data': 0.0004832019913010299, 'timing/model': 1.6940608060103841, 'learning_rate/qwen_vl_interface': 9.175205937366568e-06, 'learning_rate/action_model': 9.175205937366569e-05, 'epoch': 0.44}) +[2026-06-15 20:25:07,152][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.07585753500461578, 'train/grad_norm_pre_clip': 4.567632675170898, 'timing/data': 0.0004906130488961935, 'timing/model': 1.4633267379831523, 'learning_rate/qwen_vl_interface': 9.173490121539608e-06, 'learning_rate/action_model': 9.173490121539608e-05, 'epoch': 0.44}) +[2026-06-15 20:25:13,154][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.20697008073329926, 'train/grad_norm_pre_clip': 8.52636432647705, 'timing/data': 0.0004938580095767975, 'timing/model': 1.4657189949648455, 'learning_rate/qwen_vl_interface': 9.171772692914062e-06, 'learning_rate/action_model': 9.171772692914061e-05, 'epoch': 0.45}) +[2026-06-15 20:25:18,951][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.0554288849234581, 'train/grad_norm_pre_clip': 2.0524253845214844, 'timing/data': 0.0004966220003552735, 'timing/model': 1.4625728820101358, 'learning_rate/qwen_vl_interface': 9.1700536521959e-06, 'learning_rate/action_model': 9.170053652195899e-05, 'epoch': 0.45}) +[2026-06-15 20:25:24,752][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.29068875312805176, 'train/grad_norm_pre_clip': 10.1690092086792, 'timing/data': 0.0004964909749105573, 'timing/model': 1.4655717440182343, 'learning_rate/qwen_vl_interface': 9.168333000091754e-06, 'learning_rate/action_model': 9.168333000091754e-05, 'epoch': 0.45}) +[2026-06-15 20:25:30,653][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.2754569947719574, 'train/grad_norm_pre_clip': 9.887957572937012, 'timing/data': 0.0004896210157312453, 'timing/model': 1.44173713499913, 'learning_rate/qwen_vl_interface': 9.16661073730892e-06, 'learning_rate/action_model': 9.166610737308919e-05, 'epoch': 0.45}) +[2026-06-15 20:25:36,515][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.354894757270813, 'train/grad_norm_pre_clip': 12.418274879455566, 'timing/data': 0.0005109619814902544, 'timing/model': 1.6286799449590035, 'learning_rate/qwen_vl_interface': 9.164886864555355e-06, 'learning_rate/action_model': 9.164886864555355e-05, 'epoch': 0.45}) +[2026-06-15 20:25:42,148][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.19905884563922882, 'train/grad_norm_pre_clip': 7.242917060852051, 'timing/data': 0.0005099609843455255, 'timing/model': 1.4628594780224375, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 0.45}) +[2026-06-15 20:25:48,051][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.1079837828874588, 'train/grad_norm_pre_clip': 5.70793342590332, 'timing/data': 0.0005093200015835464, 'timing/model': 1.4635587119846605, 'learning_rate/qwen_vl_interface': 9.16143429197117e-06, 'learning_rate/action_model': 9.16143429197117e-05, 'epoch': 0.45}) +[2026-06-15 20:25:53,953][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.07465364038944244, 'train/grad_norm_pre_clip': 3.6664845943450928, 'timing/data': 0.0005117440014146268, 'timing/model': 1.4634732700069435, 'learning_rate/qwen_vl_interface': 9.159705593559773e-06, 'learning_rate/action_model': 9.159705593559773e-05, 'epoch': 0.45}) +[2026-06-15 20:25:59,852][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.18555928766727448, 'train/grad_norm_pre_clip': 6.580631732940674, 'timing/data': 0.0005136369727551937, 'timing/model': 1.430663770006504, 'learning_rate/qwen_vl_interface': 9.157975288016089e-06, 'learning_rate/action_model': 9.157975288016088e-05, 'epoch': 0.45}) +[2026-06-15 20:26:05,784][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.15578237175941467, 'train/grad_norm_pre_clip': 6.674189567565918, 'timing/data': 0.0005221689934842288, 'timing/model': 1.695096282986924, 'learning_rate/qwen_vl_interface': 9.156243376051381e-06, 'learning_rate/action_model': 9.156243376051381e-05, 'epoch': 0.45}) +[2026-06-15 20:26:11,635][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.28610193729400635, 'train/grad_norm_pre_clip': 11.644529342651367, 'timing/data': 0.0005020289681851864, 'timing/model': 1.648279091983568, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.45}) +[2026-06-15 20:26:17,351][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.24518802762031555, 'train/grad_norm_pre_clip': 8.045516014099121, 'timing/data': 0.0005206170026212931, 'timing/model': 1.4641769870067947, 'learning_rate/qwen_vl_interface': 9.152774735707248e-06, 'learning_rate/action_model': 9.152774735707247e-05, 'epoch': 0.45}) +[2026-06-15 20:26:23,352][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.16046884655952454, 'train/grad_norm_pre_clip': 11.441789627075195, 'timing/data': 0.0005133869708515704, 'timing/model': 1.4664890780113637, 'learning_rate/qwen_vl_interface': 9.15103800875365e-06, 'learning_rate/action_model': 9.15103800875365e-05, 'epoch': 0.45}) +[2026-06-15 20:26:29,352][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.19487237930297852, 'train/grad_norm_pre_clip': 6.673714637756348, 'timing/data': 0.000528378994204104, 'timing/model': 1.463519788987469, 'learning_rate/qwen_vl_interface': 9.149299678230681e-06, 'learning_rate/action_model': 9.149299678230681e-05, 'epoch': 0.45}) +[2026-06-15 20:26:35,256][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.11934154480695724, 'train/grad_norm_pre_clip': 5.948238849639893, 'timing/data': 0.0005175619735382497, 'timing/model': 1.4311511139967479, 'learning_rate/qwen_vl_interface': 9.147559744852905e-06, 'learning_rate/action_model': 9.147559744852905e-05, 'epoch': 0.45}) +[2026-06-15 20:26:41,187][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.3040357232093811, 'train/grad_norm_pre_clip': 12.017173767089844, 'timing/data': 0.0005124749732203782, 'timing/model': 1.7001711709890515, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 0.45}) +[2026-06-15 20:26:46,959][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.25390857458114624, 'train/grad_norm_pre_clip': 9.979665756225586, 'timing/data': 0.0005194750265218318, 'timing/model': 1.5704323869431391, 'learning_rate/qwen_vl_interface': 9.14407507239447e-06, 'learning_rate/action_model': 9.14407507239447e-05, 'epoch': 0.45}) +[2026-06-15 20:26:52,652][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.19515442848205566, 'train/grad_norm_pre_clip': 7.275787830352783, 'timing/data': 0.0005147680058144033, 'timing/model': 1.464266455033794, 'learning_rate/qwen_vl_interface': 9.142330334746227e-06, 'learning_rate/action_model': 9.142330334746227e-05, 'epoch': 0.45}) +[2026-06-15 20:26:58,651][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.2155080884695053, 'train/grad_norm_pre_clip': 6.49404764175415, 'timing/data': 0.0005107729812152684, 'timing/model': 1.4649143170099705, 'learning_rate/qwen_vl_interface': 9.14058399710801e-06, 'learning_rate/action_model': 9.14058399710801e-05, 'epoch': 0.45}) +[2026-06-15 20:27:04,551][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.20786353945732117, 'train/grad_norm_pre_clip': 6.592855930328369, 'timing/data': 0.0005164309986867011, 'timing/model': 1.4604082360165194, 'learning_rate/qwen_vl_interface': 9.138836060197673e-06, 'learning_rate/action_model': 9.138836060197672e-05, 'epoch': 0.45}) +[2026-06-15 20:27:10,458][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.19889338314533234, 'train/grad_norm_pre_clip': 3.886763572692871, 'timing/data': 0.0005137959960848093, 'timing/model': 1.4325917020323686, 'learning_rate/qwen_vl_interface': 9.137086524733725e-06, 'learning_rate/action_model': 9.137086524733724e-05, 'epoch': 0.45}) +[2026-06-15 20:27:16,391][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.14947988092899323, 'train/grad_norm_pre_clip': 8.317529678344727, 'timing/data': 0.0005101920105516911, 'timing/model': 1.6998985219979659, 'learning_rate/qwen_vl_interface': 9.135335391435336e-06, 'learning_rate/action_model': 9.135335391435335e-05, 'epoch': 0.45}) +[2026-06-15 20:27:22,056][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.06669344753026962, 'train/grad_norm_pre_clip': 1.8505823612213135, 'timing/data': 0.0005169520154595375, 'timing/model': 1.4644753329921514, 'learning_rate/qwen_vl_interface': 9.13358266102233e-06, 'learning_rate/action_model': 9.133582661022328e-05, 'epoch': 0.45}) +[2026-06-15 20:27:28,051][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.3859645128250122, 'train/grad_norm_pre_clip': 6.233681678771973, 'timing/data': 0.000516280997544527, 'timing/model': 1.4617191000143066, 'learning_rate/qwen_vl_interface': 9.131828334215185e-06, 'learning_rate/action_model': 9.131828334215185e-05, 'epoch': 0.46}) +[2026-06-15 20:27:33,955][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.18714265525341034, 'train/grad_norm_pre_clip': 4.9103899002075195, 'timing/data': 0.0005186640191823244, 'timing/model': 1.464798144006636, 'learning_rate/qwen_vl_interface': 9.130072411735043e-06, 'learning_rate/action_model': 9.130072411735043e-05, 'epoch': 0.46}) +[2026-06-15 20:27:39,853][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.2982044816017151, 'train/grad_norm_pre_clip': 9.60634708404541, 'timing/data': 0.0005249829846434295, 'timing/model': 1.4642818130087107, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 0.46}) +[2026-06-15 20:27:45,833][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.2830328345298767, 'train/grad_norm_pre_clip': 5.692394256591797, 'timing/data': 0.0005124449962750077, 'timing/model': 1.5026456490159035, 'learning_rate/qwen_vl_interface': 9.126555782643592e-06, 'learning_rate/action_model': 9.126555782643591e-05, 'epoch': 0.46}) +[2026-06-15 20:27:51,788][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.3126532733440399, 'train/grad_norm_pre_clip': 6.212480545043945, 'timing/data': 0.0005191750242374837, 'timing/model': 1.698646709031891, 'learning_rate/qwen_vl_interface': 9.124795077477835e-06, 'learning_rate/action_model': 9.124795077477834e-05, 'epoch': 0.46}) +[2026-06-15 20:27:57,454][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.07906140387058258, 'train/grad_norm_pre_clip': 3.4440131187438965, 'timing/data': 0.0005154989776201546, 'timing/model': 1.464549244963564, 'learning_rate/qwen_vl_interface': 9.123032779530186e-06, 'learning_rate/action_model': 9.123032779530187e-05, 'epoch': 0.46}) +[2026-06-15 20:28:03,455][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.27438247203826904, 'train/grad_norm_pre_clip': 5.573906898498535, 'timing/data': 0.0005067659658379853, 'timing/model': 1.466049384966027, 'learning_rate/qwen_vl_interface': 9.12126888952506e-06, 'learning_rate/action_model': 9.12126888952506e-05, 'epoch': 0.46}) +[2026-06-15 20:28:09,255][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.16597828269004822, 'train/grad_norm_pre_clip': 6.607961177825928, 'timing/data': 0.0005122150178067386, 'timing/model': 1.4641888340120204, 'learning_rate/qwen_vl_interface': 9.119503408187521e-06, 'learning_rate/action_model': 9.119503408187522e-05, 'epoch': 0.46}) +[2026-06-15 20:28:15,056][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.13778448104858398, 'train/grad_norm_pre_clip': 5.957864284515381, 'timing/data': 0.0005191850359551609, 'timing/model': 1.4661450170096941, 'learning_rate/qwen_vl_interface': 9.117736336243299e-06, 'learning_rate/action_model': 9.117736336243298e-05, 'epoch': 0.46}) +[2026-06-15 20:28:20,955][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.0918693020939827, 'train/grad_norm_pre_clip': 3.8635833263397217, 'timing/data': 0.0005161899607628584, 'timing/model': 1.4618943459936418, 'learning_rate/qwen_vl_interface': 9.115967674418765e-06, 'learning_rate/action_model': 9.115967674418765e-05, 'epoch': 0.46}) +[2026-06-15 20:28:26,817][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.1312803030014038, 'train/grad_norm_pre_clip': 5.968211650848389, 'timing/data': 0.0005153089878149331, 'timing/model': 1.6243253919528797, 'learning_rate/qwen_vl_interface': 9.114197423440949e-06, 'learning_rate/action_model': 9.11419742344095e-05, 'epoch': 0.46}) +[2026-06-15 20:28:32,455][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.12817910313606262, 'train/grad_norm_pre_clip': 7.13939094543457, 'timing/data': 0.0005009680171497166, 'timing/model': 1.4638791119796224, 'learning_rate/qwen_vl_interface': 9.112425584037536e-06, 'learning_rate/action_model': 9.112425584037536e-05, 'epoch': 0.46}) +[2026-06-15 20:28:38,351][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.20780088007450104, 'train/grad_norm_pre_clip': 8.606578826904297, 'timing/data': 0.0005156199913471937, 'timing/model': 1.4606299469596706, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 0.46}) +[2026-06-15 20:28:44,250][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.11633741110563278, 'train/grad_norm_pre_clip': 4.923062324523926, 'timing/data': 0.0005057749804109335, 'timing/model': 1.4640551540069282, 'learning_rate/qwen_vl_interface': 9.108877142867918e-06, 'learning_rate/action_model': 9.108877142867918e-05, 'epoch': 0.46}) +[2026-06-15 20:28:50,288][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.10073655843734741, 'train/grad_norm_pre_clip': 4.720112323760986, 'timing/data': 0.0005033009801991284, 'timing/model': 1.4330600390094332, 'learning_rate/qwen_vl_interface': 9.107100542560342e-06, 'learning_rate/action_model': 9.107100542560342e-05, 'epoch': 0.46}) +[2026-06-15 20:28:56,278][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.10968230664730072, 'train/grad_norm_pre_clip': 4.233614444732666, 'timing/data': 0.000504322990309447, 'timing/model': 1.6909417269635014, 'learning_rate/qwen_vl_interface': 9.105322356744429e-06, 'learning_rate/action_model': 9.105322356744429e-05, 'epoch': 0.46}) +[2026-06-15 20:29:02,127][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.09135209769010544, 'train/grad_norm_pre_clip': 8.067380905151367, 'timing/data': 0.0004878690233454108, 'timing/model': 1.642684612015728, 'learning_rate/qwen_vl_interface': 9.103542586151121e-06, 'learning_rate/action_model': 9.10354258615112e-05, 'epoch': 0.46}) +[2026-06-15 20:29:07,851][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.017971254885196686, 'train/grad_norm_pre_clip': 1.2822108268737793, 'timing/data': 0.000490012054797262, 'timing/model': 1.4633563330280595, 'learning_rate/qwen_vl_interface': 9.101761231512017e-06, 'learning_rate/action_model': 9.101761231512017e-05, 'epoch': 0.46}) +[2026-06-15 20:29:13,853][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.13327732682228088, 'train/grad_norm_pre_clip': 8.166769027709961, 'timing/data': 0.0004908629925921559, 'timing/model': 1.4652089240262285, 'learning_rate/qwen_vl_interface': 9.099978293559365e-06, 'learning_rate/action_model': 9.099978293559366e-05, 'epoch': 0.46}) +[2026-06-15 20:29:19,752][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.3345184326171875, 'train/grad_norm_pre_clip': 6.104869842529297, 'timing/data': 0.0004873779835179448, 'timing/model': 1.4645153539604507, 'learning_rate/qwen_vl_interface': 9.098193773026065e-06, 'learning_rate/action_model': 9.098193773026064e-05, 'epoch': 0.46}) +[2026-06-15 20:29:25,656][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.11552508175373077, 'train/grad_norm_pre_clip': 5.902817249298096, 'timing/data': 0.0005135160172358155, 'timing/model': 1.4338550149695948, 'learning_rate/qwen_vl_interface': 9.096407670645663e-06, 'learning_rate/action_model': 9.096407670645662e-05, 'epoch': 0.46}) +[2026-06-15 20:29:31,583][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.22162887454032898, 'train/grad_norm_pre_clip': 14.901158332824707, 'timing/data': 0.0005152589874342084, 'timing/model': 1.6941660650190897, 'learning_rate/qwen_vl_interface': 9.094619987152357e-06, 'learning_rate/action_model': 9.094619987152358e-05, 'epoch': 0.46}) +[2026-06-15 20:29:37,357][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.17838583886623383, 'train/grad_norm_pre_clip': 8.635778427124023, 'timing/data': 0.0005080690025351942, 'timing/model': 1.5681638419628143, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.46}) +[2026-06-15 20:29:43,054][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.07678526639938354, 'train/grad_norm_pre_clip': 5.898282527923584, 'timing/data': 0.0005183439934626222, 'timing/model': 1.4648021300090477, 'learning_rate/qwen_vl_interface': 9.09103987976709e-06, 'learning_rate/action_model': 9.09103987976709e-05, 'epoch': 0.47}) +[2026-06-15 20:29:49,053][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.10604994744062424, 'train/grad_norm_pre_clip': 4.548569202423096, 'timing/data': 0.0005065059522166848, 'timing/model': 1.4637847450212575, 'learning_rate/qwen_vl_interface': 9.089247457346774e-06, 'learning_rate/action_model': 9.089247457346773e-05, 'epoch': 0.47}) +[2026-06-15 20:29:54,949][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.2075716108083725, 'train/grad_norm_pre_clip': 7.627572059631348, 'timing/data': 0.0004949489957652986, 'timing/model': 1.4645304570440203, 'learning_rate/qwen_vl_interface': 9.087453456756849e-06, 'learning_rate/action_model': 9.087453456756849e-05, 'epoch': 0.47}) +[2026-06-15 20:30:00,853][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.08946486562490463, 'train/grad_norm_pre_clip': 9.350488662719727, 'timing/data': 0.0005600450094789267, 'timing/model': 1.430103464052081, 'learning_rate/qwen_vl_interface': 9.085657878734759e-06, 'learning_rate/action_model': 9.085657878734758e-05, 'epoch': 0.47}) +[2026-06-15 20:30:06,783][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.32651785016059875, 'train/grad_norm_pre_clip': 13.350373268127441, 'timing/data': 0.0005563200102187693, 'timing/model': 1.697169796039816, 'learning_rate/qwen_vl_interface': 9.083860724018603e-06, 'learning_rate/action_model': 9.083860724018602e-05, 'epoch': 0.47}) +[2026-06-15 20:30:12,454][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.07760679721832275, 'train/grad_norm_pre_clip': 4.380886077880859, 'timing/data': 0.0004890199634246528, 'timing/model': 1.4528817429672927, 'learning_rate/qwen_vl_interface': 9.08206199334712e-06, 'learning_rate/action_model': 9.082061993347119e-05, 'epoch': 0.47}) +[2026-06-15 20:30:18,353][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.008817470632493496, 'train/grad_norm_pre_clip': 1.3119362592697144, 'timing/data': 0.0004989050212316215, 'timing/model': 1.464689148007892, 'learning_rate/qwen_vl_interface': 9.0802616874597e-06, 'learning_rate/action_model': 9.0802616874597e-05, 'epoch': 0.47}) +[2026-06-15 20:30:24,255][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.19193024933338165, 'train/grad_norm_pre_clip': 8.666509628295898, 'timing/data': 0.0005215290002524853, 'timing/model': 1.4672906839987263, 'learning_rate/qwen_vl_interface': 9.078459807096385e-06, 'learning_rate/action_model': 9.078459807096385e-05, 'epoch': 0.47}) +[2026-06-15 20:30:30,154][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.2993379831314087, 'train/grad_norm_pre_clip': 19.35368537902832, 'timing/data': 0.0005117039545439184, 'timing/model': 1.46627689601155, 'learning_rate/qwen_vl_interface': 9.076656352997854e-06, 'learning_rate/action_model': 9.076656352997853e-05, 'epoch': 0.47}) +[2026-06-15 20:30:36,052][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.07877817749977112, 'train/grad_norm_pre_clip': 5.9092583656311035, 'timing/data': 0.0005087190074846148, 'timing/model': 1.4361028699786402, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 0.47}) +[2026-06-15 20:30:41,987][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.19450780749320984, 'train/grad_norm_pre_clip': 7.336245536804199, 'timing/data': 0.0005454329657368362, 'timing/model': 1.700269375985954, 'learning_rate/qwen_vl_interface': 9.073044726561128e-06, 'learning_rate/action_model': 9.073044726561129e-05, 'epoch': 0.47}) +[2026-06-15 20:30:47,648][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.3084132671356201, 'train/grad_norm_pre_clip': 11.862747192382812, 'timing/data': 0.0005097410175949335, 'timing/model': 1.4631870620069094, 'learning_rate/qwen_vl_interface': 9.071236555707536e-06, 'learning_rate/action_model': 9.071236555707537e-05, 'epoch': 0.47}) +[2026-06-15 20:30:53,655][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.11388860642910004, 'train/grad_norm_pre_clip': 6.53602409362793, 'timing/data': 0.0005101009737700224, 'timing/model': 1.4660087029915303, 'learning_rate/qwen_vl_interface': 9.069426814087937e-06, 'learning_rate/action_model': 9.069426814087937e-05, 'epoch': 0.47}) +[2026-06-15 20:30:59,456][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.056268516927957535, 'train/grad_norm_pre_clip': 3.3533058166503906, 'timing/data': 0.0005056340014562011, 'timing/model': 1.4677736969897524, 'learning_rate/qwen_vl_interface': 9.067615502446247e-06, 'learning_rate/action_model': 9.067615502446246e-05, 'epoch': 0.47}) +[2026-06-15 20:31:05,342][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.11079302430152893, 'train/grad_norm_pre_clip': 6.196831226348877, 'timing/data': 0.0005077079986222088, 'timing/model': 1.4490909680025652, 'learning_rate/qwen_vl_interface': 9.06580262152703e-06, 'learning_rate/action_model': 9.065802621527028e-05, 'epoch': 0.47}) +[2026-06-15 20:31:11,227][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.039635781198740005, 'train/grad_norm_pre_clip': 3.2462425231933594, 'timing/data': 0.0005088599864393473, 'timing/model': 1.511975358997006, 'learning_rate/qwen_vl_interface': 9.063988172075486e-06, 'learning_rate/action_model': 9.063988172075486e-05, 'epoch': 0.47}) +[2026-06-15 20:31:17,114][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.15053288638591766, 'train/grad_norm_pre_clip': 4.669069290161133, 'timing/data': 0.000502929964568466, 'timing/model': 1.6182165929931216, 'learning_rate/qwen_vl_interface': 9.062172154837475e-06, 'learning_rate/action_model': 9.062172154837475e-05, 'epoch': 0.47}) +[2026-06-15 20:31:22,938][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.14891648292541504, 'train/grad_norm_pre_clip': 5.406376838684082, 'timing/data': 0.0005013790214434266, 'timing/model': 1.547137325978838, 'learning_rate/qwen_vl_interface': 9.060354570559487e-06, 'learning_rate/action_model': 9.060354570559487e-05, 'epoch': 0.47}) +[2026-06-15 20:31:28,849][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.051373135298490524, 'train/grad_norm_pre_clip': 2.9780704975128174, 'timing/data': 0.0005089089972898364, 'timing/model': 1.4554604650475085, 'learning_rate/qwen_vl_interface': 9.058535419988667e-06, 'learning_rate/action_model': 9.058535419988667e-05, 'epoch': 0.47}) +[2026-06-15 20:40:25,360][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.09255202859640121, 'train/grad_norm_pre_clip': 5.097239971160889, 'eval/loss': 0.09270778298377991, 'eval/flap_precision': 0.8224137931034483, 'eval/flap_recall': 0.9559118236472945, 'eval/flap_f1': 0.8841519925857275, 'eval/flap_rate_model': 0.16111111111111112, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.8841519925857275, 'timing/data': 0.0005065660225227475, 'timing/model': 1.4702309349668212, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 0.47}) +[2026-06-15 20:40:25,360][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1100_state +[2026-06-15 20:40:25,365][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:40:30,998][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1100_state/model.safetensors +[2026-06-15 20:40:39,104][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1100_state/optimizer.bin +[2026-06-15 20:40:39,105][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1100_state/sampler.bin +[2026-06-15 20:40:39,105][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1100_state/sampler_1.bin +[2026-06-15 20:40:39,106][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1100_state/random_states_0.pkl +[2026-06-15 20:40:39,218][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-06-15 20:40:39,219][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:40:39,229][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:40:45,138][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.20312099158763885, 'train/grad_norm_pre_clip': 5.312642574310303, 'timing/data': 0.0005177620332688093, 'timing/model': 1.4823126609553583, 'learning_rate/qwen_vl_interface': 9.0548924229603e-06, 'learning_rate/action_model': 9.054892422960301e-05, 'epoch': 0.47}) +[2026-06-15 20:40:51,106][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.40774255990982056, 'train/grad_norm_pre_clip': 15.032599449157715, 'timing/data': 0.0004955389886163175, 'timing/model': 1.5544485269929282, 'learning_rate/qwen_vl_interface': 9.053068578000256e-06, 'learning_rate/action_model': 9.053068578000256e-05, 'epoch': 0.47}) +[2026-06-15 20:40:57,139][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.04027564078569412, 'train/grad_norm_pre_clip': 2.556972026824951, 'timing/data': 0.0005005769780836999, 'timing/model': 1.4981404339778237, 'learning_rate/qwen_vl_interface': 9.051243169742372e-06, 'learning_rate/action_model': 9.051243169742372e-05, 'epoch': 0.47}) +[2026-06-15 20:41:03,132][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.05423438549041748, 'train/grad_norm_pre_clip': 5.75779914855957, 'timing/data': 0.0005100509733892977, 'timing/model': 1.654866372991819, 'learning_rate/qwen_vl_interface': 9.049416198937008e-06, 'learning_rate/action_model': 9.049416198937007e-05, 'epoch': 0.48}) +[2026-06-15 20:41:08,847][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.0663006380200386, 'train/grad_norm_pre_clip': 5.975722312927246, 'timing/data': 0.0004975119954906404, 'timing/model': 1.465287409024313, 'learning_rate/qwen_vl_interface': 9.047587666335159e-06, 'learning_rate/action_model': 9.047587666335159e-05, 'epoch': 0.48}) +[2026-06-15 20:41:14,848][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.1227455884218216, 'train/grad_norm_pre_clip': 7.221618175506592, 'timing/data': 0.0004950989969074726, 'timing/model': 1.4652456429903395, 'learning_rate/qwen_vl_interface': 9.045757572688469e-06, 'learning_rate/action_model': 9.04575757268847e-05, 'epoch': 0.48}) +[2026-06-15 20:41:20,749][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.310481458902359, 'train/grad_norm_pre_clip': 10.835014343261719, 'timing/data': 0.000515229010488838, 'timing/model': 1.4644920899881981, 'learning_rate/qwen_vl_interface': 9.043925918749218e-06, 'learning_rate/action_model': 9.043925918749218e-05, 'epoch': 0.48}) +[2026-06-15 20:41:26,652][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.16022886335849762, 'train/grad_norm_pre_clip': 14.27799129486084, 'timing/data': 0.0005853009643033147, 'timing/model': 1.4375926040229388, 'learning_rate/qwen_vl_interface': 9.042092705270333e-06, 'learning_rate/action_model': 9.042092705270333e-05, 'epoch': 0.48}) +[2026-06-15 20:41:32,551][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.19016793370246887, 'train/grad_norm_pre_clip': 16.05937385559082, 'timing/data': 0.0005167610361240804, 'timing/model': 1.4319784049876034, 'learning_rate/qwen_vl_interface': 9.040257933005375e-06, 'learning_rate/action_model': 9.040257933005374e-05, 'epoch': 0.48}) +[2026-06-15 20:41:38,466][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.15870417654514313, 'train/grad_norm_pre_clip': 4.366441249847412, 'timing/data': 0.0005210370291024446, 'timing/model': 1.6776449690223671, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 0.48}) +[2026-06-15 20:41:44,152][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.09439177811145782, 'train/grad_norm_pre_clip': 5.4490132331848145, 'timing/data': 0.0005130750359967351, 'timing/model': 1.4651701920083724, 'learning_rate/qwen_vl_interface': 9.036583715134709e-06, 'learning_rate/action_model': 9.036583715134708e-05, 'epoch': 0.48}) +[2026-06-15 20:41:50,151][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.07865575700998306, 'train/grad_norm_pre_clip': 5.611357688903809, 'timing/data': 0.0005288489628583193, 'timing/model': 1.464205407013651, 'learning_rate/qwen_vl_interface': 9.034744271039332e-06, 'learning_rate/action_model': 9.034744271039332e-05, 'epoch': 0.48}) +[2026-06-15 20:41:56,054][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.06265515089035034, 'train/grad_norm_pre_clip': 3.5884945392608643, 'timing/data': 0.0005050239851698279, 'timing/model': 1.4651549409609288, 'learning_rate/qwen_vl_interface': 9.03290327117855e-06, 'learning_rate/action_model': 9.032903271178549e-05, 'epoch': 0.48}) +[2026-06-15 20:42:01,954][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.1473025232553482, 'train/grad_norm_pre_clip': 6.767032623291016, 'timing/data': 0.0005209569935686886, 'timing/model': 1.431285945000127, 'learning_rate/qwen_vl_interface': 9.031060716309122e-06, 'learning_rate/action_model': 9.031060716309121e-05, 'epoch': 0.48}) +[2026-06-15 20:42:07,753][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.07782305032014847, 'train/grad_norm_pre_clip': 3.266597270965576, 'timing/data': 0.0005457429797388613, 'timing/model': 1.4636038690223359, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.48}) +[2026-06-15 20:42:13,682][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.6487935185432434, 'train/grad_norm_pre_clip': 24.94326400756836, 'timing/data': 0.0005125449970364571, 'timing/model': 1.6952809460344724, 'learning_rate/qwen_vl_interface': 9.027370944574604e-06, 'learning_rate/action_model': 9.027370944574604e-05, 'epoch': 0.48}) +[2026-06-15 20:42:19,434][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.3645605444908142, 'train/grad_norm_pre_clip': 21.97935676574707, 'timing/data': 0.0005174919497221708, 'timing/model': 1.5459994329721667, 'learning_rate/qwen_vl_interface': 9.025523729226236e-06, 'learning_rate/action_model': 9.025523729226236e-05, 'epoch': 0.48}) +[2026-06-15 20:42:25,454][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.2591090500354767, 'train/grad_norm_pre_clip': 12.021661758422852, 'timing/data': 0.0005140160210430622, 'timing/model': 1.4669188580010086, 'learning_rate/qwen_vl_interface': 9.023674961902677e-06, 'learning_rate/action_model': 9.023674961902676e-05, 'epoch': 0.48}) +[2026-06-15 20:42:31,356][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.18809527158737183, 'train/grad_norm_pre_clip': 3.1754653453826904, 'timing/data': 0.0005202959873713553, 'timing/model': 1.4657552720163949, 'learning_rate/qwen_vl_interface': 9.021824643363883e-06, 'learning_rate/action_model': 9.021824643363883e-05, 'epoch': 0.48}) +[2026-06-15 20:42:37,254][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.36109715700149536, 'train/grad_norm_pre_clip': 9.551737785339355, 'timing/data': 0.0004917040350846946, 'timing/model': 1.437119330978021, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 0.48}) +[2026-06-15 20:42:43,154][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.14788611233234406, 'train/grad_norm_pre_clip': 8.256553649902344, 'timing/data': 0.000495128973852843, 'timing/model': 1.4411366929998621, 'learning_rate/qwen_vl_interface': 9.018119355683623e-06, 'learning_rate/action_model': 9.018119355683623e-05, 'epoch': 0.48}) +[2026-06-15 20:42:49,088][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.14474111795425415, 'train/grad_norm_pre_clip': 7.463897705078125, 'timing/data': 0.00048657599836587906, 'timing/model': 1.7007591449655592, 'learning_rate/qwen_vl_interface': 9.016264388065258e-06, 'learning_rate/action_model': 9.016264388065258e-05, 'epoch': 0.48}) +[2026-06-15 20:42:54,752][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.0478207990527153, 'train/grad_norm_pre_clip': 1.938218116760254, 'timing/data': 0.0005008170264773071, 'timing/model': 1.4641900929855183, 'learning_rate/qwen_vl_interface': 9.014407872277869e-06, 'learning_rate/action_model': 9.014407872277869e-05, 'epoch': 0.48}) +[2026-06-15 20:43:00,753][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.20242804288864136, 'train/grad_norm_pre_clip': 5.553758144378662, 'timing/data': 0.0004884999943897128, 'timing/model': 1.4644444540026598, 'learning_rate/qwen_vl_interface': 9.012549809084595e-06, 'learning_rate/action_model': 9.012549809084595e-05, 'epoch': 0.48}) +[2026-06-15 20:43:06,654][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.0548957958817482, 'train/grad_norm_pre_clip': 2.7646923065185547, 'timing/data': 0.00048372201854363084, 'timing/model': 1.466647555003874, 'learning_rate/qwen_vl_interface': 9.01069019924922e-06, 'learning_rate/action_model': 9.010690199249219e-05, 'epoch': 0.48}) +[2026-06-15 20:43:12,554][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.14698585867881775, 'train/grad_norm_pre_clip': 7.427480697631836, 'timing/data': 0.0004960399819537997, 'timing/model': 1.4334650710225105, 'learning_rate/qwen_vl_interface': 9.008829043536157e-06, 'learning_rate/action_model': 9.008829043536156e-05, 'epoch': 0.48}) +[2026-06-15 20:43:18,482][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.3488757014274597, 'train/grad_norm_pre_clip': 12.599249839782715, 'timing/data': 0.0004892399883829057, 'timing/model': 1.6960425209836103, 'learning_rate/qwen_vl_interface': 9.006966342710458e-06, 'learning_rate/action_model': 9.006966342710457e-05, 'epoch': 0.48}) +[2026-06-15 20:43:24,297][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.14743594825267792, 'train/grad_norm_pre_clip': 7.787285804748535, 'timing/data': 0.0005054540233686566, 'timing/model': 1.608025012013968, 'learning_rate/qwen_vl_interface': 9.005102097537809e-06, 'learning_rate/action_model': 9.005102097537809e-05, 'epoch': 0.49}) +[2026-06-15 20:43:29,953][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.324375718832016, 'train/grad_norm_pre_clip': 11.473100662231445, 'timing/data': 0.0005224290071055293, 'timing/model': 1.463565583981108, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.49}) +[2026-06-15 20:43:35,852][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.18775060772895813, 'train/grad_norm_pre_clip': 13.050530433654785, 'timing/data': 0.0005187439965084195, 'timing/model': 1.46654228301486, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 0.49}) +[2026-06-15 20:43:41,861][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.08944369107484818, 'train/grad_norm_pre_clip': 4.783924579620361, 'timing/data': 0.0005051529733464122, 'timing/model': 1.4778697949950583, 'learning_rate/qwen_vl_interface': 8.999500103604538e-06, 'learning_rate/action_model': 8.999500103604536e-05, 'epoch': 0.49}) +[2026-06-15 20:43:47,766][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.15468792617321014, 'train/grad_norm_pre_clip': 10.864092826843262, 'timing/data': 0.0005041920230723917, 'timing/model': 1.4450448590214364, 'learning_rate/qwen_vl_interface': 8.997629688713639e-06, 'learning_rate/action_model': 8.997629688713638e-05, 'epoch': 0.49}) +[2026-06-15 20:43:53,725][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.1459282785654068, 'train/grad_norm_pre_clip': 7.27055025100708, 'timing/data': 0.000506896001752466, 'timing/model': 1.7365317050134763, 'learning_rate/qwen_vl_interface': 8.995757733313733e-06, 'learning_rate/action_model': 8.995757733313732e-05, 'epoch': 0.49}) +[2026-06-15 20:43:59,456][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.06590276956558228, 'train/grad_norm_pre_clip': 7.35024356842041, 'timing/data': 0.0006872520316392183, 'timing/model': 1.4675813390058465, 'learning_rate/qwen_vl_interface': 8.993884238174314e-06, 'learning_rate/action_model': 8.993884238174314e-05, 'epoch': 0.49}) +[2026-06-15 20:44:05,455][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.11836957931518555, 'train/grad_norm_pre_clip': 5.384055137634277, 'timing/data': 0.0007416099542751908, 'timing/model': 1.4629472509841435, 'learning_rate/qwen_vl_interface': 8.992009204065505e-06, 'learning_rate/action_model': 8.992009204065505e-05, 'epoch': 0.49}) +[2026-06-15 20:44:11,357][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.14221301674842834, 'train/grad_norm_pre_clip': 6.602344036102295, 'timing/data': 0.0006041209562681615, 'timing/model': 1.4632123989867978, 'learning_rate/qwen_vl_interface': 8.990132631758058e-06, 'learning_rate/action_model': 8.990132631758059e-05, 'epoch': 0.49}) +[2026-06-15 20:44:17,326][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.177664116024971, 'train/grad_norm_pre_clip': 10.804038047790527, 'timing/data': 0.0006953539559617639, 'timing/model': 1.4962817130144686, 'learning_rate/qwen_vl_interface': 8.988254522023364e-06, 'learning_rate/action_model': 8.988254522023363e-05, 'epoch': 0.49}) +[2026-06-15 20:44:23,257][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.3490293323993683, 'train/grad_norm_pre_clip': 11.724882125854492, 'timing/data': 0.0005842509563080966, 'timing/model': 1.440385071968194, 'learning_rate/qwen_vl_interface': 8.986374875633441e-06, 'learning_rate/action_model': 8.98637487563344e-05, 'epoch': 0.49}) +[2026-06-15 20:44:29,149][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.2897578477859497, 'train/grad_norm_pre_clip': 17.680604934692383, 'timing/data': 0.0006032889941707253, 'timing/model': 1.659783057984896, 'learning_rate/qwen_vl_interface': 8.984493693360944e-06, 'learning_rate/action_model': 8.984493693360943e-05, 'epoch': 0.49}) +[2026-06-15 20:44:34,856][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.2369210124015808, 'train/grad_norm_pre_clip': 12.80828857421875, 'timing/data': 0.0004973450559191406, 'timing/model': 1.4658928470453247, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 0.49}) +[2026-06-15 20:44:40,855][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.4217180907726288, 'train/grad_norm_pre_clip': 11.406600952148438, 'timing/data': 0.0005818990175612271, 'timing/model': 1.4638005719752982, 'learning_rate/qwen_vl_interface': 8.98072672426198e-06, 'learning_rate/action_model': 8.980726724261981e-05, 'epoch': 0.49}) +[2026-06-15 20:44:46,755][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.2300093173980713, 'train/grad_norm_pre_clip': 8.60206127166748, 'timing/data': 0.0006002049776725471, 'timing/model': 1.4599564459640533, 'learning_rate/qwen_vl_interface': 8.978840938983977e-06, 'learning_rate/action_model': 8.978840938983977e-05, 'epoch': 0.49}) +[2026-06-15 20:44:52,656][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.5987566113471985, 'train/grad_norm_pre_clip': 10.491324424743652, 'timing/data': 0.0005662349867634475, 'timing/model': 1.4410038969945163, 'learning_rate/qwen_vl_interface': 8.976953620920313e-06, 'learning_rate/action_model': 8.976953620920312e-05, 'epoch': 0.49}) +[2026-06-15 20:44:58,558][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.19261161983013153, 'train/grad_norm_pre_clip': 5.334681034088135, 'timing/data': 0.0005624800105579197, 'timing/model': 1.4421770400367677, 'learning_rate/qwen_vl_interface': 8.975064770846794e-06, 'learning_rate/action_model': 8.975064770846794e-05, 'epoch': 0.49}) +[2026-06-15 20:45:04,465][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.17928412556648254, 'train/grad_norm_pre_clip': 6.615068435668945, 'timing/data': 0.0005374040338210762, 'timing/model': 1.6748743059579283, 'learning_rate/qwen_vl_interface': 8.973174389539856e-06, 'learning_rate/action_model': 8.973174389539855e-05, 'epoch': 0.49}) +[2026-06-15 20:45:10,158][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.02767110802233219, 'train/grad_norm_pre_clip': 1.1025053262710571, 'timing/data': 0.0005151110235601664, 'timing/model': 1.464838677959051, 'learning_rate/qwen_vl_interface': 8.971282477776563e-06, 'learning_rate/action_model': 8.971282477776563e-05, 'epoch': 0.49}) +[2026-06-15 20:45:16,155][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.0882655456662178, 'train/grad_norm_pre_clip': 5.711727142333984, 'timing/data': 0.00048238300951197743, 'timing/model': 1.463162022002507, 'learning_rate/qwen_vl_interface': 8.969389036334609e-06, 'learning_rate/action_model': 8.969389036334609e-05, 'epoch': 0.49}) +[2026-06-15 20:45:22,060][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.29064255952835083, 'train/grad_norm_pre_clip': 8.295031547546387, 'timing/data': 0.0004859079490415752, 'timing/model': 1.471128426026553, 'learning_rate/qwen_vl_interface': 8.967494065992315e-06, 'learning_rate/action_model': 8.967494065992315e-05, 'epoch': 0.49}) +[2026-06-15 20:45:27,954][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.1745920181274414, 'train/grad_norm_pre_clip': 5.948382377624512, 'timing/data': 0.0005068400059826672, 'timing/model': 1.4367337290314026, 'learning_rate/qwen_vl_interface': 8.965597567528634e-06, 'learning_rate/action_model': 8.965597567528633e-05, 'epoch': 0.49}) +[2026-06-15 20:45:33,756][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.16840849816799164, 'train/grad_norm_pre_clip': 6.584384918212891, 'timing/data': 0.0005129079800099134, 'timing/model': 1.4654231949825771, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.49}) +[2026-06-15 20:45:39,685][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.1556238979101181, 'train/grad_norm_pre_clip': 5.537458419799805, 'timing/data': 0.0005210600211285055, 'timing/model': 1.6949070690316148, 'learning_rate/qwen_vl_interface': 8.961799989356048e-06, 'learning_rate/action_model': 8.961799989356048e-05, 'epoch': 0.5}) +[2026-06-15 20:45:45,354][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.14870141446590424, 'train/grad_norm_pre_clip': 4.308233261108398, 'timing/data': 0.0004910060088150203, 'timing/model': 1.4662555990507826, 'learning_rate/qwen_vl_interface': 8.959898911208185e-06, 'learning_rate/action_model': 8.959898911208185e-05, 'epoch': 0.5}) +[2026-06-15 20:45:51,355][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.29553237557411194, 'train/grad_norm_pre_clip': 6.146665096282959, 'timing/data': 0.0004911170108243823, 'timing/model': 1.4645510379923508, 'learning_rate/qwen_vl_interface': 8.957996308061015e-06, 'learning_rate/action_model': 8.957996308061014e-05, 'epoch': 0.5}) +[2026-06-15 20:45:57,255][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.1100701317191124, 'train/grad_norm_pre_clip': 2.7762250900268555, 'timing/data': 0.000491878017783165, 'timing/model': 1.4596270920010284, 'learning_rate/qwen_vl_interface': 8.956092180696626e-06, 'learning_rate/action_model': 8.956092180696625e-05, 'epoch': 0.5}) +[2026-06-15 20:46:03,156][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.10559622198343277, 'train/grad_norm_pre_clip': 4.665041923522949, 'timing/data': 0.0004912969889119267, 'timing/model': 1.4395312950364314, 'learning_rate/qwen_vl_interface': 8.95418652989773e-06, 'learning_rate/action_model': 8.95418652989773e-05, 'epoch': 0.5}) +[2026-06-15 20:46:09,054][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.04707550257444382, 'train/grad_norm_pre_clip': 2.072390079498291, 'timing/data': 0.0004809820093214512, 'timing/model': 1.4423483909922652, 'learning_rate/qwen_vl_interface': 8.952279356447675e-06, 'learning_rate/action_model': 8.952279356447675e-05, 'epoch': 0.5}) +[2026-06-15 20:46:14,986][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.2928411662578583, 'train/grad_norm_pre_clip': 5.227128982543945, 'timing/data': 0.00048825202975422144, 'timing/model': 1.6955606890260242, 'learning_rate/qwen_vl_interface': 8.950370661130421e-06, 'learning_rate/action_model': 8.950370661130422e-05, 'epoch': 0.5}) +[2026-06-15 20:46:20,655][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.1487114131450653, 'train/grad_norm_pre_clip': 4.957897663116455, 'timing/data': 0.0005002999678254128, 'timing/model': 1.4656889599864371, 'learning_rate/qwen_vl_interface': 8.948460444730566e-06, 'learning_rate/action_model': 8.948460444730566e-05, 'epoch': 0.5}) +[2026-06-15 20:46:26,654][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.11597109586000443, 'train/grad_norm_pre_clip': 9.674339294433594, 'timing/data': 0.0004885930102318525, 'timing/model': 1.464887854992412, 'learning_rate/qwen_vl_interface': 8.946548708033324e-06, 'learning_rate/action_model': 8.946548708033323e-05, 'epoch': 0.5}) +[2026-06-15 20:46:32,577][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.2691170573234558, 'train/grad_norm_pre_clip': 12.003500938415527, 'timing/data': 0.0005729760159738362, 'timing/model': 1.4887251749751158, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 0.5}) +[2026-06-15 20:46:38,529][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.304821640253067, 'train/grad_norm_pre_clip': 10.060284614562988, 'timing/data': 0.0005247869994491339, 'timing/model': 1.4989670750219375, 'learning_rate/qwen_vl_interface': 8.942720676890677e-06, 'learning_rate/action_model': 8.942720676890677e-05, 'epoch': 0.5}) +[2026-06-15 20:46:44,490][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.15656809508800507, 'train/grad_norm_pre_clip': 8.152040481567383, 'timing/data': 0.0005161329754628241, 'timing/model': 1.7077524490305223, 'learning_rate/qwen_vl_interface': 8.940804384018833e-06, 'learning_rate/action_model': 8.940804384018832e-05, 'epoch': 0.5}) +[2026-06-15 20:46:50,294][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.4074735641479492, 'train/grad_norm_pre_clip': 13.353228569030762, 'timing/data': 0.0005532479844987392, 'timing/model': 1.613320456002839, 'learning_rate/qwen_vl_interface': 8.93888657399672e-06, 'learning_rate/action_model': 8.93888657399672e-05, 'epoch': 0.5}) +[2026-06-15 20:46:55,948][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.19385962188243866, 'train/grad_norm_pre_clip': 5.428444862365723, 'timing/data': 0.0005160439759492874, 'timing/model': 1.4630780710140243, 'learning_rate/qwen_vl_interface': 8.936967247612678e-06, 'learning_rate/action_model': 8.936967247612678e-05, 'epoch': 0.5}) +[2026-06-15 20:47:01,849][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.19133169949054718, 'train/grad_norm_pre_clip': 7.723018646240234, 'timing/data': 0.0005100150010548532, 'timing/model': 1.4645039250026457, 'learning_rate/qwen_vl_interface': 8.93504640565567e-06, 'learning_rate/action_model': 8.93504640565567e-05, 'epoch': 0.5}) +[2026-06-15 20:47:07,848][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.04695206135511398, 'train/grad_norm_pre_clip': 2.1702070236206055, 'timing/data': 0.000508352997712791, 'timing/model': 1.4651417970308103, 'learning_rate/qwen_vl_interface': 8.933124048915278e-06, 'learning_rate/action_model': 8.933124048915278e-05, 'epoch': 0.5}) +[2026-06-15 20:47:13,843][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.22661013901233673, 'train/grad_norm_pre_clip': 7.288968086242676, 'timing/data': 0.0005112370126880705, 'timing/model': 1.4989674180396833, 'learning_rate/qwen_vl_interface': 8.931200178181715e-06, 'learning_rate/action_model': 8.931200178181715e-05, 'epoch': 0.5}) +[2026-06-15 20:47:19,780][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.1452995240688324, 'train/grad_norm_pre_clip': 5.7178568840026855, 'timing/data': 0.0005146620096638799, 'timing/model': 1.6909101700293832, 'learning_rate/qwen_vl_interface': 8.929274794245812e-06, 'learning_rate/action_model': 8.929274794245812e-05, 'epoch': 0.5}) +[2026-06-15 20:47:25,450][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.10607826709747314, 'train/grad_norm_pre_clip': 7.180850982666016, 'timing/data': 0.0005178970168344676, 'timing/model': 1.4669599990011193, 'learning_rate/qwen_vl_interface': 8.927347897899018e-06, 'learning_rate/action_model': 8.927347897899016e-05, 'epoch': 0.5}) +[2026-06-15 20:47:31,450][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.169955775141716, 'train/grad_norm_pre_clip': 8.060786247253418, 'timing/data': 0.0005095750093460083, 'timing/model': 1.4660831810324453, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 0.5}) +[2026-06-15 20:47:37,351][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.2763932943344116, 'train/grad_norm_pre_clip': 5.4336137771606445, 'timing/data': 0.0005104060401208699, 'timing/model': 1.4646922270185314, 'learning_rate/qwen_vl_interface': 8.923489571141679e-06, 'learning_rate/action_model': 8.923489571141679e-05, 'epoch': 0.5}) +[2026-06-15 20:47:43,251][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.1889038383960724, 'train/grad_norm_pre_clip': 13.459301948547363, 'timing/data': 0.0005057689850218594, 'timing/model': 1.4403965239762329, 'learning_rate/qwen_vl_interface': 8.921558142317147e-06, 'learning_rate/action_model': 8.921558142317147e-05, 'epoch': 0.5}) +[2026-06-15 20:47:49,150][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.19171582162380219, 'train/grad_norm_pre_clip': 5.363498210906982, 'timing/data': 0.0005153929814696312, 'timing/model': 1.4412512150011025, 'learning_rate/qwen_vl_interface': 8.91962520425375e-06, 'learning_rate/action_model': 8.91962520425375e-05, 'epoch': 0.5}) +[2026-06-15 20:47:55,037][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.17169776558876038, 'train/grad_norm_pre_clip': 8.188268661499023, 'timing/data': 0.0004967559943906963, 'timing/model': 1.6504348259768449, 'learning_rate/qwen_vl_interface': 8.917690757746044e-06, 'learning_rate/action_model': 8.917690757746044e-05, 'epoch': 0.51}) +[2026-06-15 20:48:00,750][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.06004784628748894, 'train/grad_norm_pre_clip': 8.293742179870605, 'timing/data': 0.0005111070349812508, 'timing/model': 1.465156449994538, 'learning_rate/qwen_vl_interface': 8.915754803589208e-06, 'learning_rate/action_model': 8.915754803589209e-05, 'epoch': 0.51}) +[2026-06-15 20:48:06,751][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.13060368597507477, 'train/grad_norm_pre_clip': 5.488373756408691, 'timing/data': 0.0005184279871173203, 'timing/model': 1.4666701959795319, 'learning_rate/qwen_vl_interface': 8.913817342579042e-06, 'learning_rate/action_model': 8.913817342579042e-05, 'epoch': 0.51}) +[2026-06-15 20:48:12,650][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.033423297107219696, 'train/grad_norm_pre_clip': 3.1421897411346436, 'timing/data': 0.0005135709652677178, 'timing/model': 1.4663588010007516, 'learning_rate/qwen_vl_interface': 8.91187837551196e-06, 'learning_rate/action_model': 8.911878375511961e-05, 'epoch': 0.51}) +[2026-06-15 20:48:18,550][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.23502299189567566, 'train/grad_norm_pre_clip': 8.765205383300781, 'timing/data': 0.0005200699670240283, 'timing/model': 1.4379681079881266, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.51}) +[2026-06-15 20:48:24,450][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.13822509348392487, 'train/grad_norm_pre_clip': 8.108872413635254, 'timing/data': 0.000520872010383755, 'timing/model': 1.4382286430336535, 'learning_rate/qwen_vl_interface': 8.907995926395814e-06, 'learning_rate/action_model': 8.907995926395815e-05, 'epoch': 0.51}) +[2026-06-15 20:48:30,358][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.22214460372924805, 'train/grad_norm_pre_clip': 11.089797973632812, 'timing/data': 0.000519449997227639, 'timing/model': 1.672684699005913, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 0.51}) +[2026-06-15 20:48:36,054][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.05023543909192085, 'train/grad_norm_pre_clip': 1.7607841491699219, 'timing/data': 0.0005122990114614367, 'timing/model': 1.465266769984737, 'learning_rate/qwen_vl_interface': 8.904107462624484e-06, 'learning_rate/action_model': 8.904107462624483e-05, 'epoch': 0.51}) +[2026-06-15 20:48:42,053][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.07338450849056244, 'train/grad_norm_pre_clip': 4.170619010925293, 'timing/data': 0.000514072016812861, 'timing/model': 1.4658957729698159, 'learning_rate/qwen_vl_interface': 8.902160977240739e-06, 'learning_rate/action_model': 8.902160977240739e-05, 'epoch': 0.51}) +[2026-06-15 20:48:47,951][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.213945671916008, 'train/grad_norm_pre_clip': 10.753129005432129, 'timing/data': 0.0005182890454307199, 'timing/model': 1.461000763985794, 'learning_rate/qwen_vl_interface': 8.90021299059157e-06, 'learning_rate/action_model': 8.90021299059157e-05, 'epoch': 0.51}) +[2026-06-15 20:48:53,852][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.3811469078063965, 'train/grad_norm_pre_clip': 9.38513469696045, 'timing/data': 0.0005171570228412747, 'timing/model': 1.4401219839928672, 'learning_rate/qwen_vl_interface': 8.898263503477722e-06, 'learning_rate/action_model': 8.898263503477721e-05, 'epoch': 0.51}) +[2026-06-15 20:48:59,652][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.270521879196167, 'train/grad_norm_pre_clip': 9.94520092010498, 'timing/data': 0.0005059909890405834, 'timing/model': 1.4662766070105135, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.51}) +[2026-06-15 20:49:05,583][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.2093086689710617, 'train/grad_norm_pre_clip': 5.769506931304932, 'timing/data': 0.0005111380014568567, 'timing/model': 1.6952345999889076, 'learning_rate/qwen_vl_interface': 8.894360031062043e-06, 'learning_rate/action_model': 8.894360031062043e-05, 'epoch': 0.51}) +[2026-06-15 20:49:11,252][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.05721423402428627, 'train/grad_norm_pre_clip': 8.37504768371582, 'timing/data': 0.0005098460242152214, 'timing/model': 1.463202856015414, 'learning_rate/qwen_vl_interface': 8.892406047364783e-06, 'learning_rate/action_model': 8.892406047364784e-05, 'epoch': 0.51}) +[2026-06-15 20:49:17,248][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.18731190264225006, 'train/grad_norm_pre_clip': 8.529901504516602, 'timing/data': 0.0005122700240463018, 'timing/model': 1.4617408409831114, 'learning_rate/qwen_vl_interface': 8.890450566411984e-06, 'learning_rate/action_model': 8.890450566411984e-05, 'epoch': 0.51}) +[2026-06-15 20:49:23,152][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.19474585354328156, 'train/grad_norm_pre_clip': 7.9536824226379395, 'timing/data': 0.0005084340227767825, 'timing/model': 1.4645212789764628, 'learning_rate/qwen_vl_interface': 8.888493589007467e-06, 'learning_rate/action_model': 8.888493589007468e-05, 'epoch': 0.51}) +[2026-06-15 20:49:29,053][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.3232074975967407, 'train/grad_norm_pre_clip': 8.460880279541016, 'timing/data': 0.0005122490110807121, 'timing/model': 1.441368922998663, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 0.51}) +[2026-06-15 20:49:34,952][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.13344809412956238, 'train/grad_norm_pre_clip': 6.457731246948242, 'timing/data': 0.0005269509856589139, 'timing/model': 1.4418453559628688, 'learning_rate/qwen_vl_interface': 8.884575148061662e-06, 'learning_rate/action_model': 8.88457514806166e-05, 'epoch': 0.51}) +[2026-06-15 20:49:40,886][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.16321882605552673, 'train/grad_norm_pre_clip': 11.051694869995117, 'timing/data': 0.000508543977048248, 'timing/model': 1.6964806080213748, 'learning_rate/qwen_vl_interface': 8.882613686131095e-06, 'learning_rate/action_model': 8.882613686131093e-05, 'epoch': 0.51}) +[2026-06-15 20:49:46,556][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.15914791822433472, 'train/grad_norm_pre_clip': 5.857616424560547, 'timing/data': 0.0005255290307104588, 'timing/model': 1.4696363850380294, 'learning_rate/qwen_vl_interface': 8.880650730970254e-06, 'learning_rate/action_model': 8.880650730970254e-05, 'epoch': 0.51}) +[2026-06-15 20:49:52,553][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.08814214169979095, 'train/grad_norm_pre_clip': 2.7600815296173096, 'timing/data': 0.0005018740193918347, 'timing/model': 1.466684210987296, 'learning_rate/qwen_vl_interface': 8.878686283386042e-06, 'learning_rate/action_model': 8.878686283386042e-05, 'epoch': 0.51}) +[2026-06-15 20:49:58,455][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.11748655885457993, 'train/grad_norm_pre_clip': 3.1739258766174316, 'timing/data': 0.0005093759973533452, 'timing/model': 1.467535822011996, 'learning_rate/qwen_vl_interface': 8.876720344185967e-06, 'learning_rate/action_model': 8.876720344185967e-05, 'epoch': 0.51}) +[2026-06-15 20:50:04,351][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.3833806812763214, 'train/grad_norm_pre_clip': 15.107132911682129, 'timing/data': 0.0005083139985799789, 'timing/model': 1.4363432970130816, 'learning_rate/qwen_vl_interface': 8.874752914178149e-06, 'learning_rate/action_model': 8.87475291417815e-05, 'epoch': 0.51}) +[2026-06-15 20:50:10,280][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.09561140090227127, 'train/grad_norm_pre_clip': 4.385097503662109, 'timing/data': 0.0005693039856851101, 'timing/model': 1.6945795390056446, 'learning_rate/qwen_vl_interface': 8.87278399417133e-06, 'learning_rate/action_model': 8.872783994171329e-05, 'epoch': 0.52}) +[2026-06-15 20:50:16,097][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.11734239012002945, 'train/grad_norm_pre_clip': 5.2024359703063965, 'timing/data': 0.0005686819786205888, 'timing/model': 1.6116670140181668, 'learning_rate/qwen_vl_interface': 8.870813584974852e-06, 'learning_rate/action_model': 8.870813584974852e-05, 'epoch': 0.52}) +[2026-06-15 20:50:21,752][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.1704719066619873, 'train/grad_norm_pre_clip': 7.785717964172363, 'timing/data': 0.0005764439702033997, 'timing/model': 1.4653569559450261, 'learning_rate/qwen_vl_interface': 8.868841687398683e-06, 'learning_rate/action_model': 8.868841687398683e-05, 'epoch': 0.52}) +[2026-06-15 20:59:32,145][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.06599494814872742, 'train/grad_norm_pre_clip': 4.267703533172607, 'eval/loss': 0.09126544743776321, 'eval/flap_precision': 0.9301204819277108, 'eval/flap_recall': 0.7735470941883767, 'eval/flap_f1': 0.8446389496717723, 'eval/flap_rate_model': 0.11527777777777778, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.8446389496717723, 'timing/data': 0.0005057699745520949, 'timing/model': 1.4661618520040065, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 0.52}) +[2026-06-15 20:59:32,145][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1200_state +[2026-06-15 20:59:32,150][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:59:37,615][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1200_state/model.safetensors +[2026-06-15 20:59:45,650][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1200_state/optimizer.bin +[2026-06-15 20:59:45,651][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1200_state/sampler.bin +[2026-06-15 20:59:45,651][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1200_state/sampler_1.bin +[2026-06-15 20:59:45,652][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1200_state/random_states_0.pkl +[2026-06-15 20:59:45,753][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-06-15 20:59:45,753][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:59:45,763][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:59:51,954][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.19752387702465057, 'train/grad_norm_pre_clip': 7.865567207336426, 'timing/data': 0.000516387983225286, 'timing/model': 1.4845148660242558, 'learning_rate/qwen_vl_interface': 8.864893430350159e-06, 'learning_rate/action_model': 8.864893430350159e-05, 'epoch': 0.52}) +[2026-06-15 20:59:57,937][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.23078563809394836, 'train/grad_norm_pre_clip': 6.155376434326172, 'timing/data': 0.0005530920461751521, 'timing/model': 1.463373321981635, 'learning_rate/qwen_vl_interface': 8.862917072500785e-06, 'learning_rate/action_model': 8.862917072500784e-05, 'epoch': 0.52}) +[2026-06-15 21:00:03,947][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.12117324024438858, 'train/grad_norm_pre_clip': 5.29799222946167, 'timing/data': 0.0005649799713864923, 'timing/model': 1.4694504200015217, 'learning_rate/qwen_vl_interface': 8.860939229517674e-06, 'learning_rate/action_model': 8.860939229517674e-05, 'epoch': 0.52}) +[2026-06-15 21:00:09,926][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.2675747275352478, 'train/grad_norm_pre_clip': 7.142160415649414, 'timing/data': 0.0005683549679815769, 'timing/model': 1.492946813988965, 'learning_rate/qwen_vl_interface': 8.858959902213844e-06, 'learning_rate/action_model': 8.858959902213845e-05, 'epoch': 0.52}) +[2026-06-15 21:00:15,940][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.076405368745327, 'train/grad_norm_pre_clip': 2.5825634002685547, 'timing/data': 0.0004927630070596933, 'timing/model': 1.6609438859741203, 'learning_rate/qwen_vl_interface': 8.856979091402921e-06, 'learning_rate/action_model': 8.85697909140292e-05, 'epoch': 0.52}) +[2026-06-15 21:00:21,646][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.0944773405790329, 'train/grad_norm_pre_clip': 4.134598255157471, 'timing/data': 0.0005096790264360607, 'timing/model': 1.4681673889863305, 'learning_rate/qwen_vl_interface': 8.854996797899139e-06, 'learning_rate/action_model': 8.854996797899139e-05, 'epoch': 0.52}) +[2026-06-15 21:00:27,645][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.15944017469882965, 'train/grad_norm_pre_clip': 7.204824447631836, 'timing/data': 0.0005480649997480214, 'timing/model': 1.4641043869778514, 'learning_rate/qwen_vl_interface': 8.853013022517348e-06, 'learning_rate/action_model': 8.853013022517348e-05, 'epoch': 0.52}) +[2026-06-15 21:00:33,544][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.2437773048877716, 'train/grad_norm_pre_clip': 3.7860801219940186, 'timing/data': 0.0005067539750598371, 'timing/model': 1.464612114999909, 'learning_rate/qwen_vl_interface': 8.851027766073002e-06, 'learning_rate/action_model': 8.851027766073001e-05, 'epoch': 0.52}) +[2026-06-15 21:00:39,566][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.10885521024465561, 'train/grad_norm_pre_clip': 3.6855063438415527, 'timing/data': 0.0005335229798220098, 'timing/model': 1.4455852990504354, 'learning_rate/qwen_vl_interface': 8.84904102938216e-06, 'learning_rate/action_model': 8.84904102938216e-05, 'epoch': 0.52}) +[2026-06-15 21:00:45,547][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.28379538655281067, 'train/grad_norm_pre_clip': 7.818203926086426, 'timing/data': 0.0005219070008024573, 'timing/model': 1.458436411980074, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 0.52}) +[2026-06-15 21:00:51,592][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.177717924118042, 'train/grad_norm_pre_clip': 3.4848990440368652, 'timing/data': 0.0005082869902253151, 'timing/model': 1.71115818497492, 'learning_rate/qwen_vl_interface': 8.845063118528303e-06, 'learning_rate/action_model': 8.845063118528304e-05, 'epoch': 0.52}) +[2026-06-15 21:00:57,246][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.10410204529762268, 'train/grad_norm_pre_clip': 4.011404991149902, 'timing/data': 0.0005089669721201062, 'timing/model': 1.4641670300043188, 'learning_rate/qwen_vl_interface': 8.843071946000455e-06, 'learning_rate/action_model': 8.843071946000456e-05, 'epoch': 0.52}) +[2026-06-15 21:01:03,147][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.058902308344841, 'train/grad_norm_pre_clip': 2.724954843521118, 'timing/data': 0.0005061340052634478, 'timing/model': 1.4647029669722542, 'learning_rate/qwen_vl_interface': 8.841079296496451e-06, 'learning_rate/action_model': 8.84107929649645e-05, 'epoch': 0.52}) +[2026-06-15 21:01:09,050][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.33659863471984863, 'train/grad_norm_pre_clip': 9.035014152526855, 'timing/data': 0.0005199930164963007, 'timing/model': 1.4646661630249582, 'learning_rate/qwen_vl_interface': 8.839085170835396e-06, 'learning_rate/action_model': 8.839085170835395e-05, 'epoch': 0.52}) +[2026-06-15 21:01:14,952][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.0935618132352829, 'train/grad_norm_pre_clip': 3.4925780296325684, 'timing/data': 0.0005123520386405289, 'timing/model': 1.4362361740204506, 'learning_rate/qwen_vl_interface': 8.837089569837e-06, 'learning_rate/action_model': 8.837089569836998e-05, 'epoch': 0.52}) +[2026-06-15 21:01:20,752][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.20272043347358704, 'train/grad_norm_pre_clip': 5.758121967315674, 'timing/data': 0.0005041100084781647, 'timing/model': 1.4570660220342688, 'learning_rate/qwen_vl_interface': 8.835092494321575e-06, 'learning_rate/action_model': 8.835092494321575e-05, 'epoch': 0.52}) +[2026-06-15 21:01:26,685][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.11201317608356476, 'train/grad_norm_pre_clip': 3.6401195526123047, 'timing/data': 0.0005088869947940111, 'timing/model': 1.6952721070265397, 'learning_rate/qwen_vl_interface': 8.833093945110048e-06, 'learning_rate/action_model': 8.833093945110047e-05, 'epoch': 0.52}) +[2026-06-15 21:01:32,352][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.18273718655109406, 'train/grad_norm_pre_clip': 8.297009468078613, 'timing/data': 0.0005079759866930544, 'timing/model': 1.4654232129687443, 'learning_rate/qwen_vl_interface': 8.831093923023946e-06, 'learning_rate/action_model': 8.831093923023945e-05, 'epoch': 0.52}) +[2026-06-15 21:01:38,352][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.21343575417995453, 'train/grad_norm_pre_clip': 9.266456604003906, 'timing/data': 0.0005129630444571376, 'timing/model': 1.466246580996085, 'learning_rate/qwen_vl_interface': 8.8290924288854e-06, 'learning_rate/action_model': 8.8290924288854e-05, 'epoch': 0.52}) +[2026-06-15 21:01:44,253][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.20357948541641235, 'train/grad_norm_pre_clip': 24.840059280395508, 'timing/data': 0.0005076050292700529, 'timing/model': 1.4648785300087184, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.52}) +[2026-06-15 21:01:50,153][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.11786133050918579, 'train/grad_norm_pre_clip': 6.13158655166626, 'timing/data': 0.0005152469966560602, 'timing/model': 1.4360718029784039, 'learning_rate/qwen_vl_interface': 8.825085027742545e-06, 'learning_rate/action_model': 8.825085027742544e-05, 'epoch': 0.53}) +[2026-06-15 21:01:56,154][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.29954516887664795, 'train/grad_norm_pre_clip': 7.093754291534424, 'timing/data': 0.0005073950160294771, 'timing/model': 1.440144031017553, 'learning_rate/qwen_vl_interface': 8.823079122385523e-06, 'learning_rate/action_model': 8.823079122385523e-05, 'epoch': 0.53}) +[2026-06-15 21:02:02,299][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.200673446059227, 'train/grad_norm_pre_clip': 13.70781135559082, 'timing/data': 0.0005226479843258858, 'timing/model': 1.7082389279967174, 'learning_rate/qwen_vl_interface': 8.821071748270644e-06, 'learning_rate/action_model': 8.821071748270644e-05, 'epoch': 0.53}) +[2026-06-15 21:02:07,954][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.19567376375198364, 'train/grad_norm_pre_clip': 8.289589881896973, 'timing/data': 0.0005316209862940013, 'timing/model': 1.466814064013306, 'learning_rate/qwen_vl_interface': 8.81906290622306e-06, 'learning_rate/action_model': 8.81906290622306e-05, 'epoch': 0.53}) +[2026-06-15 21:02:13,953][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.13542693853378296, 'train/grad_norm_pre_clip': 8.042275428771973, 'timing/data': 0.0005147959454916418, 'timing/model': 1.4623722480027936, 'learning_rate/qwen_vl_interface': 8.81705259706853e-06, 'learning_rate/action_model': 8.817052597068529e-05, 'epoch': 0.53}) +[2026-06-15 21:02:19,853][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.1777820885181427, 'train/grad_norm_pre_clip': 3.9856531620025635, 'timing/data': 0.0005141160218045115, 'timing/model': 1.4616562030278146, 'learning_rate/qwen_vl_interface': 8.815040821633416e-06, 'learning_rate/action_model': 8.815040821633416e-05, 'epoch': 0.53}) +[2026-06-15 21:02:25,755][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.09776203334331512, 'train/grad_norm_pre_clip': 2.5020105838775635, 'timing/data': 0.000509848992805928, 'timing/model': 1.43310680805007, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.53}) +[2026-06-15 21:02:31,685][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.0754154846072197, 'train/grad_norm_pre_clip': 3.8166096210479736, 'timing/data': 0.000515798048581928, 'timing/model': 1.6959754819981754, 'learning_rate/qwen_vl_interface': 8.8110128752299e-06, 'learning_rate/action_model': 8.8110128752299e-05, 'epoch': 0.53}) +[2026-06-15 21:02:37,503][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.19045527279376984, 'train/grad_norm_pre_clip': 8.01594066619873, 'timing/data': 0.0005123430164530873, 'timing/model': 1.6120143120060675, 'learning_rate/qwen_vl_interface': 8.808996705917235e-06, 'learning_rate/action_model': 8.808996705917235e-05, 'epoch': 0.53}) +[2026-06-15 21:02:43,154][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.09345906227827072, 'train/grad_norm_pre_clip': 4.0887651443481445, 'timing/data': 0.0005118820117786527, 'timing/model': 1.4645488940295763, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 0.53}) +[2026-06-15 21:02:49,054][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.20233304798603058, 'train/grad_norm_pre_clip': 7.192732810974121, 'timing/data': 0.000513323990162462, 'timing/model': 1.4636219369713217, 'learning_rate/qwen_vl_interface': 8.804959979213945e-06, 'learning_rate/action_model': 8.804959979213945e-05, 'epoch': 0.53}) +[2026-06-15 21:02:55,055][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.0390530526638031, 'train/grad_norm_pre_clip': 3.0160152912139893, 'timing/data': 0.000508357014041394, 'timing/model': 1.4848179159453139, 'learning_rate/qwen_vl_interface': 8.802939423482667e-06, 'learning_rate/action_model': 8.802939423482666e-05, 'epoch': 0.53}) +[2026-06-15 21:03:00,960][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.10911915451288223, 'train/grad_norm_pre_clip': 9.473626136779785, 'timing/data': 0.0005147660267539322, 'timing/model': 1.43599796196213, 'learning_rate/qwen_vl_interface': 8.800917407272198e-06, 'learning_rate/action_model': 8.800917407272198e-05, 'epoch': 0.53}) +[2026-06-15 21:03:06,889][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.11765970289707184, 'train/grad_norm_pre_clip': 5.643414497375488, 'timing/data': 0.0005148359923623502, 'timing/model': 1.694753388990648, 'learning_rate/qwen_vl_interface': 8.798893931413711e-06, 'learning_rate/action_model': 8.79889393141371e-05, 'epoch': 0.53}) +[2026-06-15 21:03:12,556][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.06198493763804436, 'train/grad_norm_pre_clip': 4.331801414489746, 'timing/data': 0.0005052119959145784, 'timing/model': 1.4638371710316278, 'learning_rate/qwen_vl_interface': 8.796868996738985e-06, 'learning_rate/action_model': 8.796868996738984e-05, 'epoch': 0.53}) +[2026-06-15 21:03:18,557][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.09193871915340424, 'train/grad_norm_pre_clip': 6.831202507019043, 'timing/data': 0.0005077859968878329, 'timing/model': 1.4649572109919973, 'learning_rate/qwen_vl_interface': 8.794842604080389e-06, 'learning_rate/action_model': 8.794842604080389e-05, 'epoch': 0.53}) +[2026-06-15 21:03:24,555][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.14108361303806305, 'train/grad_norm_pre_clip': 7.136399745941162, 'timing/data': 0.0005095190135762095, 'timing/model': 1.485375516989734, 'learning_rate/qwen_vl_interface': 8.7928147542709e-06, 'learning_rate/action_model': 8.7928147542709e-05, 'epoch': 0.53}) +[2026-06-15 21:03:30,454][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.1415550261735916, 'train/grad_norm_pre_clip': 10.147250175476074, 'timing/data': 0.0004984720144420862, 'timing/model': 1.4360786579782143, 'learning_rate/qwen_vl_interface': 8.790785448144088e-06, 'learning_rate/action_model': 8.790785448144088e-05, 'epoch': 0.53}) +[2026-06-15 21:03:36,356][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.07456934452056885, 'train/grad_norm_pre_clip': 7.9660491943359375, 'timing/data': 0.0005094280350022018, 'timing/model': 1.4418460320448503, 'learning_rate/qwen_vl_interface': 8.788754686534126e-06, 'learning_rate/action_model': 8.788754686534127e-05, 'epoch': 0.53}) +[2026-06-15 21:03:42,253][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.186409592628479, 'train/grad_norm_pre_clip': 10.08871841430664, 'timing/data': 0.0005093479994684458, 'timing/model': 1.6666294179740362, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 0.53}) +[2026-06-15 21:03:47,958][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.2885589003562927, 'train/grad_norm_pre_clip': 11.910476684570312, 'timing/data': 0.0005085670272819698, 'timing/model': 1.4682296010432765, 'learning_rate/qwen_vl_interface': 8.784688800204427e-06, 'learning_rate/action_model': 8.784688800204426e-05, 'epoch': 0.53}) +[2026-06-15 21:03:53,957][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.12867048382759094, 'train/grad_norm_pre_clip': 10.49707317352295, 'timing/data': 0.000515307008754462, 'timing/model': 1.4654146779794246, 'learning_rate/qwen_vl_interface': 8.782653677156022e-06, 'learning_rate/action_model': 8.782653677156022e-05, 'epoch': 0.53}) +[2026-06-15 21:03:59,855][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.13612888753414154, 'train/grad_norm_pre_clip': 12.49224853515625, 'timing/data': 0.000502857961691916, 'timing/model': 1.4644970529479906, 'learning_rate/qwen_vl_interface': 8.780617101967129e-06, 'learning_rate/action_model': 8.780617101967129e-05, 'epoch': 0.53}) +[2026-06-15 21:04:05,757][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.12098820507526398, 'train/grad_norm_pre_clip': 11.425773620605469, 'timing/data': 0.0005149960052222013, 'timing/model': 1.440774024988059, 'learning_rate/qwen_vl_interface': 8.77857907547491e-06, 'learning_rate/action_model': 8.77857907547491e-05, 'epoch': 0.54}) +[2026-06-15 21:04:11,658][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.19452476501464844, 'train/grad_norm_pre_clip': 15.82983684539795, 'timing/data': 0.0005116320098750293, 'timing/model': 1.4409403200261295, 'learning_rate/qwen_vl_interface': 8.776539598517119e-06, 'learning_rate/action_model': 8.776539598517118e-05, 'epoch': 0.54}) +[2026-06-15 21:04:17,573][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.18647271394729614, 'train/grad_norm_pre_clip': 15.307990074157715, 'timing/data': 0.0005050119943916798, 'timing/model': 1.6818872300209478, 'learning_rate/qwen_vl_interface': 8.774498671932108e-06, 'learning_rate/action_model': 8.774498671932107e-05, 'epoch': 0.54}) +[2026-06-15 21:04:23,264][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.2102476954460144, 'train/grad_norm_pre_clip': 8.897424697875977, 'timing/data': 0.0005169500363990664, 'timing/model': 1.4741985090076923, 'learning_rate/qwen_vl_interface': 8.772456296558824e-06, 'learning_rate/action_model': 8.772456296558824e-05, 'epoch': 0.54}) +[2026-06-15 21:04:29,256][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.036899384111166, 'train/grad_norm_pre_clip': 4.548092842102051, 'timing/data': 0.0005027379957027733, 'timing/model': 1.4657549950061366, 'learning_rate/qwen_vl_interface': 8.77041247323681e-06, 'learning_rate/action_model': 8.770412473236811e-05, 'epoch': 0.54}) +[2026-06-15 21:04:35,164][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.03976212069392204, 'train/grad_norm_pre_clip': 2.7222039699554443, 'timing/data': 0.0005027389852330089, 'timing/model': 1.4737682989798486, 'learning_rate/qwen_vl_interface': 8.76836720280621e-06, 'learning_rate/action_model': 8.768367202806209e-05, 'epoch': 0.54}) +[2026-06-15 21:04:41,059][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.3091683089733124, 'train/grad_norm_pre_clip': 14.90107536315918, 'timing/data': 0.0005074459477327764, 'timing/model': 1.441975861962419, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 0.54}) +[2026-06-15 21:04:46,856][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.24022868275642395, 'train/grad_norm_pre_clip': 12.96549129486084, 'timing/data': 0.000517209991812706, 'timing/model': 1.4692951070028357, 'learning_rate/qwen_vl_interface': 8.76427232398276e-06, 'learning_rate/action_model': 8.76427232398276e-05, 'epoch': 0.54}) +[2026-06-15 21:04:52,786][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.42200759053230286, 'train/grad_norm_pre_clip': 11.941844940185547, 'timing/data': 0.0005089379847049713, 'timing/model': 1.6939887970220298, 'learning_rate/qwen_vl_interface': 8.762222717273167e-06, 'learning_rate/action_model': 8.762222717273167e-05, 'epoch': 0.54}) +[2026-06-15 21:04:58,455][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.4262515902519226, 'train/grad_norm_pre_clip': 23.014198303222656, 'timing/data': 0.0005035190260969102, 'timing/model': 1.463602922973223, 'learning_rate/qwen_vl_interface': 8.76017166682148e-06, 'learning_rate/action_model': 8.760171666821481e-05, 'epoch': 0.54}) +[2026-06-15 21:05:04,455][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.09331756085157394, 'train/grad_norm_pre_clip': 11.941125869750977, 'timing/data': 0.000620862003415823, 'timing/model': 1.4647465129964985, 'learning_rate/qwen_vl_interface': 8.758119173470815e-06, 'learning_rate/action_model': 8.758119173470815e-05, 'epoch': 0.54}) +[2026-06-15 21:05:10,356][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.06335064023733139, 'train/grad_norm_pre_clip': 4.2093095779418945, 'timing/data': 0.0005166089977137744, 'timing/model': 1.4628831939771771, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 0.54}) +[2026-06-15 21:05:16,257][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.10297912359237671, 'train/grad_norm_pre_clip': 2.632129192352295, 'timing/data': 0.0005062739946879447, 'timing/model': 1.4409011279931292, 'learning_rate/qwen_vl_interface': 8.75400986144794e-06, 'learning_rate/action_model': 8.75400986144794e-05, 'epoch': 0.54}) +[2026-06-15 21:05:22,155][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.12333594262599945, 'train/grad_norm_pre_clip': 8.073530197143555, 'timing/data': 0.0005202239844948053, 'timing/model': 1.4416640769923106, 'learning_rate/qwen_vl_interface': 8.751953044464913e-06, 'learning_rate/action_model': 8.751953044464913e-05, 'epoch': 0.54}) +[2026-06-15 21:05:28,193][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.0719524547457695, 'train/grad_norm_pre_clip': 3.8040449619293213, 'timing/data': 0.0005225079949013889, 'timing/model': 1.7005767280352302, 'learning_rate/qwen_vl_interface': 8.749894787961272e-06, 'learning_rate/action_model': 8.749894787961271e-05, 'epoch': 0.54}) +[2026-06-15 21:05:33,857][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.3918973207473755, 'train/grad_norm_pre_clip': 23.801347732543945, 'timing/data': 0.0005233989795669913, 'timing/model': 1.461807677987963, 'learning_rate/qwen_vl_interface': 8.747835092783086e-06, 'learning_rate/action_model': 8.747835092783085e-05, 'epoch': 0.54}) +[2026-06-15 21:05:39,857][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.18548905849456787, 'train/grad_norm_pre_clip': 8.850293159484863, 'timing/data': 0.0005151769728399813, 'timing/model': 1.4645694339997135, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 0.54}) +[2026-06-15 21:05:45,855][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.11711462587118149, 'train/grad_norm_pre_clip': 6.626526832580566, 'timing/data': 0.0005107909673824906, 'timing/model': 1.489844956959132, 'learning_rate/qwen_vl_interface': 8.743711389790321e-06, 'learning_rate/action_model': 8.74371138979032e-05, 'epoch': 0.54}) +[2026-06-15 21:05:51,757][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.20638778805732727, 'train/grad_norm_pre_clip': 13.709356307983398, 'timing/data': 0.0005149960052222013, 'timing/model': 1.4342230929760262, 'learning_rate/qwen_vl_interface': 8.741647383670839e-06, 'learning_rate/action_model': 8.741647383670839e-05, 'epoch': 0.54}) +[2026-06-15 21:05:57,693][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.02983018569648266, 'train/grad_norm_pre_clip': 1.581061601638794, 'timing/data': 0.0005194530240260065, 'timing/model': 1.698075626976788, 'learning_rate/qwen_vl_interface': 8.739581942267011e-06, 'learning_rate/action_model': 8.739581942267011e-05, 'epoch': 0.54}) +[2026-06-15 21:06:03,511][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.28505226969718933, 'train/grad_norm_pre_clip': 17.039005279541016, 'timing/data': 0.0005271240370348096, 'timing/model': 1.6207510859821923, 'learning_rate/qwen_vl_interface': 8.737515066427861e-06, 'learning_rate/action_model': 8.73751506642786e-05, 'epoch': 0.54}) +[2026-06-15 21:06:09,151][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.018952948972582817, 'train/grad_norm_pre_clip': 1.5580267906188965, 'timing/data': 0.000513062987010926, 'timing/model': 1.462013243988622, 'learning_rate/qwen_vl_interface': 8.735446757003e-06, 'learning_rate/action_model': 8.735446757003e-05, 'epoch': 0.54}) +[2026-06-15 21:06:15,052][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.29693251848220825, 'train/grad_norm_pre_clip': 15.240242004394531, 'timing/data': 0.0005185020272620022, 'timing/model': 1.4659458189853467, 'learning_rate/qwen_vl_interface': 8.733377014842635e-06, 'learning_rate/action_model': 8.733377014842635e-05, 'epoch': 0.54}) +[2026-06-15 21:06:20,956][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.19744043052196503, 'train/grad_norm_pre_clip': 7.194958686828613, 'timing/data': 0.00051071000052616, 'timing/model': 1.4595460720011033, 'learning_rate/qwen_vl_interface': 8.731305840797557e-06, 'learning_rate/action_model': 8.731305840797558e-05, 'epoch': 0.55}) +[2026-06-15 21:06:26,929][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.22953228652477264, 'train/grad_norm_pre_clip': 4.721710681915283, 'timing/data': 0.00052826595492661, 'timing/model': 1.50119194900617, 'learning_rate/qwen_vl_interface': 8.72923323571915e-06, 'learning_rate/action_model': 8.72923323571915e-05, 'epoch': 0.55}) +[2026-06-15 21:06:32,888][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.5257294178009033, 'train/grad_norm_pre_clip': 16.931289672851562, 'timing/data': 0.0005049919709563255, 'timing/model': 1.6945283399545588, 'learning_rate/qwen_vl_interface': 8.727159200459381e-06, 'learning_rate/action_model': 8.727159200459381e-05, 'epoch': 0.55}) +[2026-06-15 21:06:38,554][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.5645097494125366, 'train/grad_norm_pre_clip': 14.407973289489746, 'timing/data': 0.0005181719898246229, 'timing/model': 1.463936701999046, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 0.55}) +[2026-06-15 21:06:44,555][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.18377618491649628, 'train/grad_norm_pre_clip': 4.155176639556885, 'timing/data': 0.0005176099948585033, 'timing/model': 1.4645942940260284, 'learning_rate/qwen_vl_interface': 8.723006842806575e-06, 'learning_rate/action_model': 8.723006842806576e-05, 'epoch': 0.55}) +[2026-06-15 21:06:50,556][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.041279714554548264, 'train/grad_norm_pre_clip': 4.074300289154053, 'timing/data': 0.0005184509791433811, 'timing/model': 1.4792289540055208, 'learning_rate/qwen_vl_interface': 8.720928522120419e-06, 'learning_rate/action_model': 8.720928522120418e-05, 'epoch': 0.55}) +[2026-06-15 21:06:56,454][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.07730553299188614, 'train/grad_norm_pre_clip': 5.523318290710449, 'timing/data': 0.0005035899812355638, 'timing/model': 1.4389246280188672, 'learning_rate/qwen_vl_interface': 8.718848774666655e-06, 'learning_rate/action_model': 8.718848774666655e-05, 'epoch': 0.55}) +[2026-06-15 21:07:02,353][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.3764383792877197, 'train/grad_norm_pre_clip': 15.007264137268066, 'timing/data': 0.0005083369906060398, 'timing/model': 1.43390963395359, 'learning_rate/qwen_vl_interface': 8.716767601300189e-06, 'learning_rate/action_model': 8.716767601300188e-05, 'epoch': 0.55}) +[2026-06-15 21:07:08,244][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.11055488139390945, 'train/grad_norm_pre_clip': 8.3591947555542, 'timing/data': 0.000515687046572566, 'timing/model': 1.6559838310349733, 'learning_rate/qwen_vl_interface': 8.714685002876514e-06, 'learning_rate/action_model': 8.714685002876513e-05, 'epoch': 0.55}) +[2026-06-15 21:07:14,056][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.10495121031999588, 'train/grad_norm_pre_clip': 8.286371231079102, 'timing/data': 0.0005048710154369473, 'timing/model': 1.466623060987331, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.55}) +[2026-06-15 21:07:20,055][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.2556515336036682, 'train/grad_norm_pre_clip': 16.46088409423828, 'timing/data': 0.000503199000377208, 'timing/model': 1.4670690019847825, 'learning_rate/qwen_vl_interface': 8.710515534282428e-06, 'learning_rate/action_model': 8.710515534282428e-05, 'epoch': 0.55}) +[2026-06-15 21:07:25,951][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.20068436861038208, 'train/grad_norm_pre_clip': 18.599790573120117, 'timing/data': 0.0005190530209802091, 'timing/model': 1.4553111899876967, 'learning_rate/qwen_vl_interface': 8.708428665825931e-06, 'learning_rate/action_model': 8.708428665825931e-05, 'epoch': 0.55}) +[2026-06-15 21:07:31,864][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.2139807939529419, 'train/grad_norm_pre_clip': 5.158731937408447, 'timing/data': 0.0005114420200698078, 'timing/model': 1.448341351991985, 'learning_rate/qwen_vl_interface': 8.706340375740045e-06, 'learning_rate/action_model': 8.706340375740045e-05, 'epoch': 0.55}) +[2026-06-15 21:07:37,770][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.05153840407729149, 'train/grad_norm_pre_clip': 2.579526662826538, 'timing/data': 0.0005116419633850455, 'timing/model': 1.456003075989429, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 0.55}) +[2026-06-15 21:07:43,666][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.08092188835144043, 'train/grad_norm_pre_clip': 3.2655301094055176, 'timing/data': 0.0005251009715721011, 'timing/model': 1.6792473039822653, 'learning_rate/qwen_vl_interface': 8.70215953411436e-06, 'learning_rate/action_model': 8.70215953411436e-05, 'epoch': 0.55}) +[2026-06-15 21:07:49,351][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.1720755398273468, 'train/grad_norm_pre_clip': 6.8429036140441895, 'timing/data': 0.0005289770197123289, 'timing/model': 1.4636714209918864, 'learning_rate/qwen_vl_interface': 8.70006698429315e-06, 'learning_rate/action_model': 8.700066984293149e-05, 'epoch': 0.55}) +[2026-06-15 21:07:55,356][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.4831823706626892, 'train/grad_norm_pre_clip': 30.872638702392578, 'timing/data': 0.0005340250208973885, 'timing/model': 1.4641765370033681, 'learning_rate/qwen_vl_interface': 8.697973016279722e-06, 'learning_rate/action_model': 8.697973016279722e-05, 'epoch': 0.55}) +[2026-06-15 21:08:01,353][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.11393748223781586, 'train/grad_norm_pre_clip': 5.331543922424316, 'timing/data': 0.0005051830084994435, 'timing/model': 1.4659797109779902, 'learning_rate/qwen_vl_interface': 8.695877630934827e-06, 'learning_rate/action_model': 8.695877630934826e-05, 'epoch': 0.55}) +[2026-06-15 21:08:07,343][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.1093079075217247, 'train/grad_norm_pre_clip': 2.9396309852600098, 'timing/data': 0.0005201739841140807, 'timing/model': 1.507766589988023, 'learning_rate/qwen_vl_interface': 8.693780829119801e-06, 'learning_rate/action_model': 8.693780829119801e-05, 'epoch': 0.55}) +[2026-06-15 21:08:13,257][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.31402599811553955, 'train/grad_norm_pre_clip': 7.514249324798584, 'timing/data': 0.0005214670090936124, 'timing/model': 1.4581442839698866, 'learning_rate/qwen_vl_interface': 8.69168261169656e-06, 'learning_rate/action_model': 8.691682611696559e-05, 'epoch': 0.55}) +[2026-06-15 21:08:19,193][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.24061356484889984, 'train/grad_norm_pre_clip': 6.525157451629639, 'timing/data': 0.0005209860391914845, 'timing/model': 1.7029667439637706, 'learning_rate/qwen_vl_interface': 8.689582979527599e-06, 'learning_rate/action_model': 8.689582979527599e-05, 'epoch': 0.55}) +[2026-06-15 21:08:24,858][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.14257864654064178, 'train/grad_norm_pre_clip': 3.173042058944702, 'timing/data': 0.000522598042152822, 'timing/model': 1.4636361799784936, 'learning_rate/qwen_vl_interface': 8.687481933476e-06, 'learning_rate/action_model': 8.687481933476e-05, 'epoch': 0.55}) +[2026-06-15 21:08:30,854][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.13670988380908966, 'train/grad_norm_pre_clip': 3.7056145668029785, 'timing/data': 0.0005196839920245111, 'timing/model': 1.463253199995961, 'learning_rate/qwen_vl_interface': 8.685379474405423e-06, 'learning_rate/action_model': 8.685379474405423e-05, 'epoch': 0.55}) +[2026-06-15 21:08:36,855][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.0926188975572586, 'train/grad_norm_pre_clip': 2.4967150688171387, 'timing/data': 0.000525251030921936, 'timing/model': 1.4826982470112853, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.56}) +[2026-06-15 21:08:42,755][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.08268605917692184, 'train/grad_norm_pre_clip': 4.447399616241455, 'timing/data': 0.0005271250265650451, 'timing/model': 1.4331591909867711, 'learning_rate/qwen_vl_interface': 8.681170320664883e-06, 'learning_rate/action_model': 8.681170320664883e-05, 'epoch': 0.56}) +[2026-06-15 21:08:48,653][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.24559034407138824, 'train/grad_norm_pre_clip': 7.914453983306885, 'timing/data': 0.0005193529650568962, 'timing/model': 1.43987736600684, 'learning_rate/qwen_vl_interface': 8.679063627725143e-06, 'learning_rate/action_model': 8.679063627725144e-05, 'epoch': 0.56}) +[2026-06-15 21:08:54,593][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.3163313865661621, 'train/grad_norm_pre_clip': 7.007115840911865, 'timing/data': 0.0005352460430003703, 'timing/model': 1.705133076000493, 'learning_rate/qwen_vl_interface': 8.676955525226876e-06, 'learning_rate/action_model': 8.676955525226875e-05, 'epoch': 0.56}) +[2026-06-15 21:09:00,255][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.30031436681747437, 'train/grad_norm_pre_clip': 3.856717109680176, 'timing/data': 0.0005169999785721302, 'timing/model': 1.4669697839999571, 'learning_rate/qwen_vl_interface': 8.674846014036636e-06, 'learning_rate/action_model': 8.674846014036636e-05, 'epoch': 0.56}) +[2026-06-15 21:09:06,416][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.1661836802959442, 'train/grad_norm_pre_clip': 3.449739933013916, 'timing/data': 0.0005238890298642218, 'timing/model': 1.5387213900103234, 'learning_rate/qwen_vl_interface': 8.672735095021569e-06, 'learning_rate/action_model': 8.672735095021569e-05, 'epoch': 0.56}) +[2026-06-15 21:09:12,255][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.04148199036717415, 'train/grad_norm_pre_clip': 2.3493456840515137, 'timing/data': 0.0007663259748369455, 'timing/model': 1.4652559199603274, 'learning_rate/qwen_vl_interface': 8.670622769049393e-06, 'learning_rate/action_model': 8.670622769049392e-05, 'epoch': 0.56}) +[2026-06-15 21:09:18,165][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.15599112212657928, 'train/grad_norm_pre_clip': 5.001701831817627, 'timing/data': 0.00050686503527686, 'timing/model': 1.4464262090041302, 'learning_rate/qwen_vl_interface': 8.668509036988406e-06, 'learning_rate/action_model': 8.668509036988405e-05, 'epoch': 0.56}) +[2026-06-15 21:09:24,107][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.09247131645679474, 'train/grad_norm_pre_clip': 3.172970771789551, 'timing/data': 0.0005084769800305367, 'timing/model': 1.716722265991848, 'learning_rate/qwen_vl_interface': 8.66639389970748e-06, 'learning_rate/action_model': 8.666393899707479e-05, 'epoch': 0.56}) +[2026-06-15 21:09:29,911][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.12069695442914963, 'train/grad_norm_pre_clip': 3.4170150756835938, 'timing/data': 0.0005072149797342718, 'timing/model': 1.6189723360002972, 'learning_rate/qwen_vl_interface': 8.66427735807607e-06, 'learning_rate/action_model': 8.664277358076069e-05, 'epoch': 0.56}) +[2026-06-15 21:18:46,455][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.31298425793647766, 'train/grad_norm_pre_clip': 12.128934860229492, 'eval/loss': 0.09974082559347153, 'eval/flap_precision': 0.8550458715596331, 'eval/flap_recall': 0.9338677354709419, 'eval/flap_f1': 0.89272030651341, 'eval/flap_rate_model': 0.15138888888888888, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.89272030651341, 'timing/data': 0.0005246310029178858, 'timing/model': 1.4656527580227703, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 0.56}) +[2026-06-15 21:18:46,455][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1300_state +[2026-06-15 21:18:46,459][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 21:18:51,992][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1300_state/model.safetensors +[2026-06-15 21:19:05,529][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1300_state/optimizer.bin +[2026-06-15 21:19:05,529][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1300_state/sampler.bin +[2026-06-15 21:19:05,530][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1300_state/sampler_1.bin +[2026-06-15 21:19:05,531][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1300_state/random_states_0.pkl +[2026-06-15 21:19:05,645][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-06-15 21:19:05,646][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 21:19:05,655][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 21:19:11,441][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.2305811047554016, 'train/grad_norm_pre_clip': 9.293928146362305, 'timing/data': 0.0005083630094304681, 'timing/model': 1.464158151007723, 'learning_rate/qwen_vl_interface': 8.660040065242499e-06, 'learning_rate/action_model': 8.660040065242498e-05, 'epoch': 0.56}) +[2026-06-15 21:19:17,444][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.24614781141281128, 'train/grad_norm_pre_clip': 6.923752307891846, 'timing/data': 0.0005493639619089663, 'timing/model': 1.4643695389968343, 'learning_rate/qwen_vl_interface': 8.657919315782127e-06, 'learning_rate/action_model': 8.657919315782127e-05, 'epoch': 0.56}) +[2026-06-15 21:19:23,448][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.13944844901561737, 'train/grad_norm_pre_clip': 5.73020601272583, 'timing/data': 0.0007389039965346456, 'timing/model': 1.4481954719522037, 'learning_rate/qwen_vl_interface': 8.655797165454853e-06, 'learning_rate/action_model': 8.655797165454853e-05, 'epoch': 0.56}) +[2026-06-15 21:19:29,351][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.27193689346313477, 'train/grad_norm_pre_clip': 15.368566513061523, 'timing/data': 0.0004983879625797272, 'timing/model': 1.4674687089864165, 'learning_rate/qwen_vl_interface': 8.653673615133013e-06, 'learning_rate/action_model': 8.653673615133013e-05, 'epoch': 0.56}) +[2026-06-15 21:19:35,248][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.23087449371814728, 'train/grad_norm_pre_clip': 9.909317016601562, 'timing/data': 0.0005100260023027658, 'timing/model': 1.4727857370162383, 'learning_rate/qwen_vl_interface': 8.651548665689519e-06, 'learning_rate/action_model': 8.651548665689518e-05, 'epoch': 0.56}) +[2026-06-15 21:19:41,347][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.2802577316761017, 'train/grad_norm_pre_clip': 5.126028060913086, 'timing/data': 0.0005278210155665874, 'timing/model': 1.7560931199695915, 'learning_rate/qwen_vl_interface': 8.649422317997855e-06, 'learning_rate/action_model': 8.649422317997856e-05, 'epoch': 0.56}) +[2026-06-15 21:19:47,256][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.0858016237616539, 'train/grad_norm_pre_clip': 2.8450100421905518, 'timing/data': 0.0005040359683334827, 'timing/model': 1.4645898059825413, 'learning_rate/qwen_vl_interface': 8.647294572932082e-06, 'learning_rate/action_model': 8.647294572932082e-05, 'epoch': 0.56}) +[2026-06-15 21:19:53,255][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.13327200710773468, 'train/grad_norm_pre_clip': 3.1707100868225098, 'timing/data': 0.0004893349832855165, 'timing/model': 1.4608671510359272, 'learning_rate/qwen_vl_interface': 8.64516543136684e-06, 'learning_rate/action_model': 8.64516543136684e-05, 'epoch': 0.56}) +[2026-06-15 21:19:59,158][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.14274579286575317, 'train/grad_norm_pre_clip': 4.039685249328613, 'timing/data': 0.0005132699734531343, 'timing/model': 1.460113665962126, 'learning_rate/qwen_vl_interface': 8.643034894177335e-06, 'learning_rate/action_model': 8.643034894177335e-05, 'epoch': 0.56}) +[2026-06-15 21:20:05,252][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.11866635084152222, 'train/grad_norm_pre_clip': 1.8555923700332642, 'timing/data': 0.0005054980283603072, 'timing/model': 1.5097909030155279, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 0.56}) +[2026-06-15 21:20:11,185][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.16941611468791962, 'train/grad_norm_pre_clip': 3.08129620552063, 'timing/data': 0.000522583955898881, 'timing/model': 1.6894071889691986, 'learning_rate/qwen_vl_interface': 8.638769636429245e-06, 'learning_rate/action_model': 8.638769636429245e-05, 'epoch': 0.56}) +[2026-06-15 21:20:17,075][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.17582349479198456, 'train/grad_norm_pre_clip': 3.2980611324310303, 'timing/data': 0.0005340109928511083, 'timing/model': 1.5773236520471983, 'learning_rate/qwen_vl_interface': 8.636634917623946e-06, 'learning_rate/action_model': 8.636634917623944e-05, 'epoch': 0.56}) +[2026-06-15 21:20:22,857][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.11644557863473892, 'train/grad_norm_pre_clip': 3.9708805084228516, 'timing/data': 0.000517586013302207, 'timing/model': 1.46773104602471, 'learning_rate/qwen_vl_interface': 8.634498806700955e-06, 'learning_rate/action_model': 8.634498806700954e-05, 'epoch': 0.56}) +[2026-06-15 21:20:28,760][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.11704512685537338, 'train/grad_norm_pre_clip': 4.604834079742432, 'timing/data': 0.0005010910099372268, 'timing/model': 1.4659197219880298, 'learning_rate/qwen_vl_interface': 8.63236130453835e-06, 'learning_rate/action_model': 8.63236130453835e-05, 'epoch': 0.57}) +[2026-06-15 21:20:34,858][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.22274507582187653, 'train/grad_norm_pre_clip': 6.031199932098389, 'timing/data': 0.0005153520032763481, 'timing/model': 1.4979825370246544, 'learning_rate/qwen_vl_interface': 8.630222412014777e-06, 'learning_rate/action_model': 8.630222412014776e-05, 'epoch': 0.57}) +[2026-06-15 21:20:40,934][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.22464585304260254, 'train/grad_norm_pre_clip': 6.958667278289795, 'timing/data': 0.0004981669480912387, 'timing/model': 1.5460079290205613, 'learning_rate/qwen_vl_interface': 8.628082130009451e-06, 'learning_rate/action_model': 8.628082130009451e-05, 'epoch': 0.57}) +[2026-06-15 21:20:46,907][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.14751139283180237, 'train/grad_norm_pre_clip': 6.184738636016846, 'timing/data': 0.0007251129718497396, 'timing/model': 1.7091514169587754, 'learning_rate/qwen_vl_interface': 8.62594045940216e-06, 'learning_rate/action_model': 8.62594045940216e-05, 'epoch': 0.57}) +[2026-06-15 21:20:52,556][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.05710156261920929, 'train/grad_norm_pre_clip': 2.9823460578918457, 'timing/data': 0.0005283320206217468, 'timing/model': 1.461819465039298, 'learning_rate/qwen_vl_interface': 8.623797401073272e-06, 'learning_rate/action_model': 8.623797401073271e-05, 'epoch': 0.57}) +[2026-06-15 21:20:58,557][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.22611761093139648, 'train/grad_norm_pre_clip': 4.779149532318115, 'timing/data': 0.0005019329837523401, 'timing/model': 1.4677376850158907, 'learning_rate/qwen_vl_interface': 8.62165295590371e-06, 'learning_rate/action_model': 8.621652955903709e-05, 'epoch': 0.57}) +[2026-06-15 21:21:04,456][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.17943952977657318, 'train/grad_norm_pre_clip': 8.467421531677246, 'timing/data': 0.0005219019949436188, 'timing/model': 1.4603290199884214, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 0.57}) +[2026-06-15 21:21:10,461][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.124027781188488, 'train/grad_norm_pre_clip': 4.229003429412842, 'timing/data': 0.000524125003721565, 'timing/model': 1.4659255759906955, 'learning_rate/qwen_vl_interface': 8.61735990856914e-06, 'learning_rate/action_model': 8.617359908569139e-05, 'epoch': 0.57}) +[2026-06-15 21:21:16,363][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.2947065234184265, 'train/grad_norm_pre_clip': 5.250763416290283, 'timing/data': 0.0005165640031918883, 'timing/model': 1.4400161890080199, 'learning_rate/qwen_vl_interface': 8.615211308168842e-06, 'learning_rate/action_model': 8.61521130816884e-05, 'epoch': 0.57}) +[2026-06-15 21:21:22,478][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.10182671993970871, 'train/grad_norm_pre_clip': 4.726232528686523, 'timing/data': 0.0005167439812794328, 'timing/model': 1.742087453021668, 'learning_rate/qwen_vl_interface': 8.613061324457289e-06, 'learning_rate/action_model': 8.613061324457288e-05, 'epoch': 0.57}) +[2026-06-15 21:21:28,204][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.4692324995994568, 'train/grad_norm_pre_clip': 7.144114971160889, 'timing/data': 0.0005160740111023188, 'timing/model': 1.503398656961508, 'learning_rate/qwen_vl_interface': 8.610909958318262e-06, 'learning_rate/action_model': 8.610909958318261e-05, 'epoch': 0.57}) +[2026-06-15 21:21:34,213][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.028254695236682892, 'train/grad_norm_pre_clip': 2.7241106033325195, 'timing/data': 0.0005161240114830434, 'timing/model': 1.5134758700150996, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.57}) +[2026-06-15 21:21:40,163][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.25546780228614807, 'train/grad_norm_pre_clip': 15.8612699508667, 'timing/data': 0.0005023229750804603, 'timing/model': 1.4999311690335162, 'learning_rate/qwen_vl_interface': 8.606603082295724e-06, 'learning_rate/action_model': 8.606603082295724e-05, 'epoch': 0.57}) +[2026-06-15 21:21:46,167][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.3439572751522064, 'train/grad_norm_pre_clip': 10.845296859741211, 'timing/data': 0.0005236139986664057, 'timing/model': 1.441984303004574, 'learning_rate/qwen_vl_interface': 8.60444757418261e-06, 'learning_rate/action_model': 8.604447574182609e-05, 'epoch': 0.57}) +[2026-06-15 21:21:52,248][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.1715601533651352, 'train/grad_norm_pre_clip': 8.0087251663208, 'timing/data': 0.0005085120210424066, 'timing/model': 1.539861810975708, 'learning_rate/qwen_vl_interface': 8.602290687182805e-06, 'learning_rate/action_model': 8.602290687182806e-05, 'epoch': 0.57}) +[2026-06-15 21:21:58,258][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.0266389399766922, 'train/grad_norm_pre_clip': 3.2222185134887695, 'timing/data': 0.0005022819968871772, 'timing/model': 1.6752305059926584, 'learning_rate/qwen_vl_interface': 8.600132422182929e-06, 'learning_rate/action_model': 8.600132422182928e-05, 'epoch': 0.57}) +[2026-06-15 21:22:04,060][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.2720736861228943, 'train/grad_norm_pre_clip': 6.446859359741211, 'timing/data': 0.0004995280178263783, 'timing/model': 1.4632485249894671, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 0.57}) +[2026-06-15 21:22:10,163][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.2907087504863739, 'train/grad_norm_pre_clip': 12.415250778198242, 'timing/data': 0.000492178020067513, 'timing/model': 1.466440521995537, 'learning_rate/qwen_vl_interface': 8.595811761732239e-06, 'learning_rate/action_model': 8.595811761732238e-05, 'epoch': 0.57}) +[2026-06-15 21:22:16,262][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.04183454439043999, 'train/grad_norm_pre_clip': 3.6133031845092773, 'timing/data': 0.0005172649980522692, 'timing/model': 1.47530111100059, 'learning_rate/qwen_vl_interface': 8.593649368057489e-06, 'learning_rate/action_model': 8.59364936805749e-05, 'epoch': 0.57}) +[2026-06-15 21:22:22,159][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.3937489092350006, 'train/grad_norm_pre_clip': 15.271225929260254, 'timing/data': 0.0005090719787403941, 'timing/model': 1.438795824011322, 'learning_rate/qwen_vl_interface': 8.591485599934786e-06, 'learning_rate/action_model': 8.591485599934785e-05, 'epoch': 0.57}) +[2026-06-15 21:22:28,059][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.2341029793024063, 'train/grad_norm_pre_clip': 9.883072853088379, 'timing/data': 0.0005066490266472101, 'timing/model': 1.4397108849952929, 'learning_rate/qwen_vl_interface': 8.58932045825357e-06, 'learning_rate/action_model': 8.589320458253569e-05, 'epoch': 0.57}) +[2026-06-15 21:22:33,970][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.15585027635097504, 'train/grad_norm_pre_clip': 6.860710144042969, 'timing/data': 0.0005073100328445435, 'timing/model': 1.6709085399634205, 'learning_rate/qwen_vl_interface': 8.58715394390385e-06, 'learning_rate/action_model': 8.58715394390385e-05, 'epoch': 0.57}) +[2026-06-15 21:22:39,660][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.12973052263259888, 'train/grad_norm_pre_clip': 3.039425849914551, 'timing/data': 0.0005067090387456119, 'timing/model': 1.460860238992609, 'learning_rate/qwen_vl_interface': 8.5849860577762e-06, 'learning_rate/action_model': 8.5849860577762e-05, 'epoch': 0.57}) +[2026-06-15 21:22:45,701][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.09070612490177155, 'train/grad_norm_pre_clip': 2.853367805480957, 'timing/data': 0.0007027589599601924, 'timing/model': 1.5032637940021232, 'learning_rate/qwen_vl_interface': 8.582816800761756e-06, 'learning_rate/action_model': 8.582816800761755e-05, 'epoch': 0.58}) +[2026-06-15 21:22:51,845][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.12051664292812347, 'train/grad_norm_pre_clip': 2.49448561668396, 'timing/data': 0.0005669380188919604, 'timing/model': 1.5339258880121633, 'learning_rate/qwen_vl_interface': 8.580646173752214e-06, 'learning_rate/action_model': 8.580646173752213e-05, 'epoch': 0.58}) +[2026-06-15 21:22:57,826][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.07315350323915482, 'train/grad_norm_pre_clip': 2.7665109634399414, 'timing/data': 0.000588360009714961, 'timing/model': 1.5005385670228861, 'learning_rate/qwen_vl_interface': 8.578474177639841e-06, 'learning_rate/action_model': 8.578474177639842e-05, 'epoch': 0.58}) +[2026-06-15 21:23:03,667][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.31184500455856323, 'train/grad_norm_pre_clip': 8.171365737915039, 'timing/data': 0.0005330979474820197, 'timing/model': 1.4700363469892181, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 0.58}) +[2026-06-15 21:23:09,590][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.2261466681957245, 'train/grad_norm_pre_clip': 4.469647407531738, 'timing/data': 0.0005075799999758601, 'timing/model': 1.685524164000526, 'learning_rate/qwen_vl_interface': 8.574126081678463e-06, 'learning_rate/action_model': 8.574126081678462e-05, 'epoch': 0.58}) +[2026-06-15 21:23:15,261][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.22201724350452423, 'train/grad_norm_pre_clip': 6.110485076904297, 'timing/data': 0.0005148709751665592, 'timing/model': 1.4625860569649376, 'learning_rate/qwen_vl_interface': 8.571949983616796e-06, 'learning_rate/action_model': 8.571949983616795e-05, 'epoch': 0.58}) +[2026-06-15 21:23:21,261][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.04360567405819893, 'train/grad_norm_pre_clip': 2.317139148712158, 'timing/data': 0.0005093530053272843, 'timing/model': 1.4587049559922889, 'learning_rate/qwen_vl_interface': 8.569772520026971e-06, 'learning_rate/action_model': 8.569772520026971e-05, 'epoch': 0.58}) +[2026-06-15 21:23:27,259][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.3558593988418579, 'train/grad_norm_pre_clip': 8.179919242858887, 'timing/data': 0.0005214599659666419, 'timing/model': 1.485766279976815, 'learning_rate/qwen_vl_interface': 8.567593691804061e-06, 'learning_rate/action_model': 8.567593691804061e-05, 'epoch': 0.58}) +[2026-06-15 21:23:33,162][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.16710025072097778, 'train/grad_norm_pre_clip': 3.9633991718292236, 'timing/data': 0.0005158029962331057, 'timing/model': 1.4402303220122121, 'learning_rate/qwen_vl_interface': 8.565413499843703e-06, 'learning_rate/action_model': 8.565413499843703e-05, 'epoch': 0.58}) +[2026-06-15 21:23:39,063][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.23209728300571442, 'train/grad_norm_pre_clip': 5.074759483337402, 'timing/data': 0.0005111059872433543, 'timing/model': 1.4406421420280822, 'learning_rate/qwen_vl_interface': 8.563231945042087e-06, 'learning_rate/action_model': 8.563231945042088e-05, 'epoch': 0.58}) +[2026-06-15 21:23:44,992][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.3734724819660187, 'train/grad_norm_pre_clip': 9.76190185546875, 'timing/data': 0.0005142699810676277, 'timing/model': 1.6958984340308234, 'learning_rate/qwen_vl_interface': 8.56104902829597e-06, 'learning_rate/action_model': 8.56104902829597e-05, 'epoch': 0.58}) +[2026-06-15 21:23:50,660][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.12878356873989105, 'train/grad_norm_pre_clip': 5.107663631439209, 'timing/data': 0.0004986359854228795, 'timing/model': 1.4658129349700175, 'learning_rate/qwen_vl_interface': 8.558864750502669e-06, 'learning_rate/action_model': 8.558864750502668e-05, 'epoch': 0.58}) +[2026-06-15 21:23:56,658][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.16807681322097778, 'train/grad_norm_pre_clip': 5.120228290557861, 'timing/data': 0.0005025920108892024, 'timing/model': 1.462699145951774, 'learning_rate/qwen_vl_interface': 8.556679112560055e-06, 'learning_rate/action_model': 8.556679112560054e-05, 'epoch': 0.58}) +[2026-06-15 21:24:02,659][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.09068308770656586, 'train/grad_norm_pre_clip': 4.329320430755615, 'timing/data': 0.0005153219681233168, 'timing/model': 1.4973640980315395, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 0.58}) +[2026-06-15 21:24:08,767][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.22752629220485687, 'train/grad_norm_pre_clip': 7.615449905395508, 'timing/data': 0.0005117059918120503, 'timing/model': 1.446452799020335, 'learning_rate/qwen_vl_interface': 8.55230375982118e-06, 'learning_rate/action_model': 8.55230375982118e-05, 'epoch': 0.58}) +[2026-06-15 21:24:14,768][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.22007110714912415, 'train/grad_norm_pre_clip': 8.334146499633789, 'timing/data': 0.0005101329879835248, 'timing/model': 1.7740946839912795, 'learning_rate/qwen_vl_interface': 8.550114046823463e-06, 'learning_rate/action_model': 8.550114046823463e-05, 'epoch': 0.58}) +[2026-06-15 21:24:20,846][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.2367100715637207, 'train/grad_norm_pre_clip': 7.630064010620117, 'timing/data': 0.0008267510565929115, 'timing/model': 1.7294684210210107, 'learning_rate/qwen_vl_interface': 8.547922977273518e-06, 'learning_rate/action_model': 8.547922977273518e-05, 'epoch': 0.58}) +[2026-06-15 21:24:26,561][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.16480585932731628, 'train/grad_norm_pre_clip': 6.274632930755615, 'timing/data': 0.0005355810280889273, 'timing/model': 1.4591697119758464, 'learning_rate/qwen_vl_interface': 8.54573055207201e-06, 'learning_rate/action_model': 8.545730552072008e-05, 'epoch': 0.58}) +[2026-06-15 21:24:32,640][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.08050177246332169, 'train/grad_norm_pre_clip': 2.62481951713562, 'timing/data': 0.0005144700407981873, 'timing/model': 1.5442932570003904, 'learning_rate/qwen_vl_interface': 8.543536772120161e-06, 'learning_rate/action_model': 8.54353677212016e-05, 'epoch': 0.58}) +[2026-06-15 21:24:38,849][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.06053834408521652, 'train/grad_norm_pre_clip': 2.9371204376220703, 'timing/data': 0.0005122169968672097, 'timing/model': 1.5616755609516986, 'learning_rate/qwen_vl_interface': 8.541341638319754e-06, 'learning_rate/action_model': 8.541341638319754e-05, 'epoch': 0.58}) +[2026-06-15 21:24:44,837][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.2283041626214981, 'train/grad_norm_pre_clip': 7.930202484130859, 'timing/data': 0.0005222809850238264, 'timing/model': 1.5057801130460575, 'learning_rate/qwen_vl_interface': 8.539145151573126e-06, 'learning_rate/action_model': 8.539145151573125e-05, 'epoch': 0.58}) +[2026-06-15 21:24:50,981][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.15432380139827728, 'train/grad_norm_pre_clip': 7.473763942718506, 'timing/data': 0.0005140990251675248, 'timing/model': 1.7856894230353646, 'learning_rate/qwen_vl_interface': 8.536947312783167e-06, 'learning_rate/action_model': 8.536947312783166e-05, 'epoch': 0.58}) +[2026-06-15 21:24:56,818][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.26122599840164185, 'train/grad_norm_pre_clip': 8.422123908996582, 'timing/data': 0.0005182750173844397, 'timing/model': 1.545399377006106, 'learning_rate/qwen_vl_interface': 8.534748122853326e-06, 'learning_rate/action_model': 8.534748122853325e-05, 'epoch': 0.58}) +[2026-06-15 21:25:02,761][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.20208683609962463, 'train/grad_norm_pre_clip': 9.737159729003906, 'timing/data': 0.0005210789968259633, 'timing/model': 1.463392732955981, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.59}) +[2026-06-15 21:25:08,806][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.18910083174705505, 'train/grad_norm_pre_clip': 5.831225395202637, 'timing/data': 0.0005203280015848577, 'timing/model': 1.5300636070314795, 'learning_rate/qwen_vl_interface': 8.530345693190573e-06, 'learning_rate/action_model': 8.530345693190572e-05, 'epoch': 0.59}) +[2026-06-15 21:25:14,761][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.1095329001545906, 'train/grad_norm_pre_clip': 5.957254886627197, 'timing/data': 0.0005074189975857735, 'timing/model': 1.4830166450119577, 'learning_rate/qwen_vl_interface': 8.528142455267328e-06, 'learning_rate/action_model': 8.528142455267329e-05, 'epoch': 0.59}) +[2026-06-15 21:25:20,719][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.2719772458076477, 'train/grad_norm_pre_clip': 12.960137367248535, 'timing/data': 0.000515161023940891, 'timing/model': 1.4913801180082373, 'learning_rate/qwen_vl_interface': 8.52593786982355e-06, 'learning_rate/action_model': 8.525937869823551e-05, 'epoch': 0.59}) +[2026-06-15 21:25:26,751][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.08076800405979156, 'train/grad_norm_pre_clip': 4.567761421203613, 'timing/data': 0.0005120059940963984, 'timing/model': 1.671836651978083, 'learning_rate/qwen_vl_interface': 8.523731937765456e-06, 'learning_rate/action_model': 8.523731937765456e-05, 'epoch': 0.59}) +[2026-06-15 21:25:32,575][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.22518227994441986, 'train/grad_norm_pre_clip': 8.978431701660156, 'timing/data': 0.0005119459819979966, 'timing/model': 1.479505515017081, 'learning_rate/qwen_vl_interface': 8.521524659999823e-06, 'learning_rate/action_model': 8.521524659999822e-05, 'epoch': 0.59}) +[2026-06-15 21:25:38,563][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.08890406787395477, 'train/grad_norm_pre_clip': 7.84865140914917, 'timing/data': 0.0005076790112070739, 'timing/model': 1.4657905909698457, 'learning_rate/qwen_vl_interface': 8.519316037433976e-06, 'learning_rate/action_model': 8.519316037433977e-05, 'epoch': 0.59}) +[2026-06-15 21:25:44,663][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.24292494356632233, 'train/grad_norm_pre_clip': 8.581999778747559, 'timing/data': 0.0005072889616712928, 'timing/model': 1.4875404589693062, 'learning_rate/qwen_vl_interface': 8.517106070975801e-06, 'learning_rate/action_model': 8.5171060709758e-05, 'epoch': 0.59}) +[2026-06-15 21:25:50,640][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.13164693117141724, 'train/grad_norm_pre_clip': 4.470120429992676, 'timing/data': 0.0005387960118241608, 'timing/model': 1.5098264379776083, 'learning_rate/qwen_vl_interface': 8.514894761533729e-06, 'learning_rate/action_model': 8.514894761533728e-05, 'epoch': 0.59}) +[2026-06-15 21:25:56,725][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.13334479928016663, 'train/grad_norm_pre_clip': 4.84459114074707, 'timing/data': 0.0005021719844080508, 'timing/model': 1.5551652580033988, 'learning_rate/qwen_vl_interface': 8.512682110016747e-06, 'learning_rate/action_model': 8.512682110016747e-05, 'epoch': 0.59}) +[2026-06-15 21:26:02,867][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.2130553275346756, 'train/grad_norm_pre_clip': 10.707695007324219, 'timing/data': 0.0005090120248496532, 'timing/model': 1.726064950984437, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 0.59}) +[2026-06-15 21:26:08,626][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.10648152977228165, 'train/grad_norm_pre_clip': 4.93372917175293, 'timing/data': 0.0005137079861015081, 'timing/model': 1.5271792680141516, 'learning_rate/qwen_vl_interface': 8.50825278439675e-06, 'learning_rate/action_model': 8.508252784396749e-05, 'epoch': 0.59}) +[2026-06-15 21:26:14,661][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.13171109557151794, 'train/grad_norm_pre_clip': 5.954367637634277, 'timing/data': 0.0005129270139150321, 'timing/model': 1.464456938963849, 'learning_rate/qwen_vl_interface': 8.506036112114465e-06, 'learning_rate/action_model': 8.506036112114464e-05, 'epoch': 0.59}) +[2026-06-15 21:26:20,819][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.2937365472316742, 'train/grad_norm_pre_clip': 10.454684257507324, 'timing/data': 0.0005196569836698472, 'timing/model': 1.5249983359826729, 'learning_rate/qwen_vl_interface': 8.503818101398725e-06, 'learning_rate/action_model': 8.503818101398725e-05, 'epoch': 0.59}) +[2026-06-15 21:26:26,868][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.365038126707077, 'train/grad_norm_pre_clip': 9.313658714294434, 'timing/data': 0.0005283599602989852, 'timing/model': 1.4602174150059, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.59}) +[2026-06-15 21:26:32,669][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.21940922737121582, 'train/grad_norm_pre_clip': 6.034578323364258, 'timing/data': 0.0005195069825276732, 'timing/model': 1.467002886987757, 'learning_rate/qwen_vl_interface': 8.499378068314395e-06, 'learning_rate/action_model': 8.499378068314395e-05, 'epoch': 0.59}) +[2026-06-15 21:26:38,596][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.10226171463727951, 'train/grad_norm_pre_clip': 3.9006452560424805, 'timing/data': 0.0005151900113560259, 'timing/model': 1.6976083790068515, 'learning_rate/qwen_vl_interface': 8.497156047770932e-06, 'learning_rate/action_model': 8.497156047770933e-05, 'epoch': 0.59}) +[2026-06-15 21:26:44,341][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.15057973563671112, 'train/grad_norm_pre_clip': 4.034644603729248, 'timing/data': 0.0005216200370341539, 'timing/model': 1.5431808030116372, 'learning_rate/qwen_vl_interface': 8.494932692444276e-06, 'learning_rate/action_model': 8.494932692444276e-05, 'epoch': 0.59}) +[2026-06-15 21:26:50,458][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.2267170399427414, 'train/grad_norm_pre_clip': 5.803964614868164, 'timing/data': 0.0005269470275379717, 'timing/model': 1.4621432719868608, 'learning_rate/qwen_vl_interface': 8.492708003248362e-06, 'learning_rate/action_model': 8.492708003248362e-05, 'epoch': 0.59}) +[2026-06-15 21:26:56,542][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.2742386758327484, 'train/grad_norm_pre_clip': 12.924320220947266, 'timing/data': 0.0005166829796507955, 'timing/model': 1.5458491149474867, 'learning_rate/qwen_vl_interface': 8.490481981097674e-06, 'learning_rate/action_model': 8.490481981097674e-05, 'epoch': 0.59}) +[2026-06-15 21:27:02,728][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.15081757307052612, 'train/grad_norm_pre_clip': 5.503726482391357, 'timing/data': 0.0005353399901650846, 'timing/model': 1.5027782139950432, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 0.59}) +[2026-06-15 21:27:08,740][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.13498175144195557, 'train/grad_norm_pre_clip': 11.997513771057129, 'timing/data': 0.000517514010425657, 'timing/model': 1.5166446749935858, 'learning_rate/qwen_vl_interface': 8.486025941592664e-06, 'learning_rate/action_model': 8.486025941592664e-05, 'epoch': 0.59}) +[2026-06-15 21:27:14,777][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.12799681723117828, 'train/grad_norm_pre_clip': 7.911293983459473, 'timing/data': 0.0005199769511818886, 'timing/model': 1.782316847005859, 'learning_rate/qwen_vl_interface': 8.483795926070054e-06, 'learning_rate/action_model': 8.483795926070052e-05, 'epoch': 0.59}) +[2026-06-15 21:27:20,543][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.1646319180727005, 'train/grad_norm_pre_clip': 5.898673057556152, 'timing/data': 0.0005252049886621535, 'timing/model': 1.5106271610129625, 'learning_rate/qwen_vl_interface': 8.481564581256087e-06, 'learning_rate/action_model': 8.481564581256087e-05, 'epoch': 0.6}) +[2026-06-15 21:27:26,693][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.15370434522628784, 'train/grad_norm_pre_clip': 6.657328128814697, 'timing/data': 0.0005133369704708457, 'timing/model': 1.5328463860205375, 'learning_rate/qwen_vl_interface': 8.479331908067987e-06, 'learning_rate/action_model': 8.479331908067987e-05, 'epoch': 0.6}) +[2026-06-15 21:27:32,677][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.19947031140327454, 'train/grad_norm_pre_clip': 4.22598934173584, 'timing/data': 0.0005173530080355704, 'timing/model': 1.4889058659900911, 'learning_rate/qwen_vl_interface': 8.477097907423524e-06, 'learning_rate/action_model': 8.477097907423524e-05, 'epoch': 0.6}) +[2026-06-15 21:27:38,639][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.10359222441911697, 'train/grad_norm_pre_clip': 5.195022106170654, 'timing/data': 0.0005156099796295166, 'timing/model': 1.5108386499923654, 'learning_rate/qwen_vl_interface': 8.474862580241006e-06, 'learning_rate/action_model': 8.474862580241006e-05, 'epoch': 0.6}) +[2026-06-15 21:27:44,706][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.1256653070449829, 'train/grad_norm_pre_clip': 4.721030235290527, 'timing/data': 0.0007843669736757874, 'timing/model': 1.7374973580008373, 'learning_rate/qwen_vl_interface': 8.472625927439299e-06, 'learning_rate/action_model': 8.472625927439299e-05, 'epoch': 0.6}) +[2026-06-15 21:27:50,676][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.1781730055809021, 'train/grad_norm_pre_clip': 5.29360818862915, 'timing/data': 0.0005210490198805928, 'timing/model': 1.6793815130367875, 'learning_rate/qwen_vl_interface': 8.4703879499378e-06, 'learning_rate/action_model': 8.4703879499378e-05, 'epoch': 0.6}) +[2026-06-15 21:27:56,602][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.11047609895467758, 'train/grad_norm_pre_clip': 3.8218564987182617, 'timing/data': 0.0005069680046290159, 'timing/model': 1.5501256710267626, 'learning_rate/qwen_vl_interface': 8.46814864865646e-06, 'learning_rate/action_model': 8.46814864865646e-05, 'epoch': 0.6}) +[2026-06-15 21:28:02,719][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.055647607892751694, 'train/grad_norm_pre_clip': 2.772972345352173, 'timing/data': 0.0005155510152690113, 'timing/model': 1.5680171689600684, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 0.6}) +[2026-06-15 21:28:08,762][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.07245667278766632, 'train/grad_norm_pre_clip': 3.3735945224761963, 'timing/data': 0.000507458986248821, 'timing/model': 1.4870105859590694, 'learning_rate/qwen_vl_interface': 8.463666078436769e-06, 'learning_rate/action_model': 8.463666078436769e-05, 'epoch': 0.6}) +[2026-06-15 21:28:14,963][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.13958975672721863, 'train/grad_norm_pre_clip': 4.1528096199035645, 'timing/data': 0.0005149600328877568, 'timing/model': 1.4805402950150892, 'learning_rate/qwen_vl_interface': 8.46142281134103e-06, 'learning_rate/action_model': 8.461422811341031e-05, 'epoch': 0.6}) +[2026-06-15 21:28:21,029][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.0456230528652668, 'train/grad_norm_pre_clip': 2.282015800476074, 'timing/data': 0.0005148489726707339, 'timing/model': 1.7622401370317675, 'learning_rate/qwen_vl_interface': 8.459178224150683e-06, 'learning_rate/action_model': 8.459178224150683e-05, 'epoch': 0.6}) +[2026-06-15 21:28:26,844][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.02937997505068779, 'train/grad_norm_pre_clip': 2.839155673980713, 'timing/data': 0.0005185850313864648, 'timing/model': 1.5141078180167824, 'learning_rate/qwen_vl_interface': 8.456932317788389e-06, 'learning_rate/action_model': 8.456932317788389e-05, 'epoch': 0.6}) +[2026-06-15 21:28:32,948][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.083028644323349, 'train/grad_norm_pre_clip': 3.445002555847168, 'timing/data': 0.0005144790047779679, 'timing/model': 1.5112961159902625, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 0.6}) +[2026-06-15 21:28:38,963][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.06337717920541763, 'train/grad_norm_pre_clip': 5.118017196655273, 'timing/data': 0.0005066270241513848, 'timing/model': 1.4614486980135553, 'learning_rate/qwen_vl_interface': 8.452436551241328e-06, 'learning_rate/action_model': 8.452436551241328e-05, 'epoch': 0.6}) +[2026-06-15 21:28:45,129][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.10888942331075668, 'train/grad_norm_pre_clip': 8.589497566223145, 'timing/data': 0.0005108730401843786, 'timing/model': 1.5037616129848175, 'learning_rate/qwen_vl_interface': 8.450186692904605e-06, 'learning_rate/action_model': 8.450186692904606e-05, 'epoch': 0.6}) +[2026-06-15 21:28:51,144][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.26164957880973816, 'train/grad_norm_pre_clip': 12.58349895477295, 'timing/data': 0.0005215999553911388, 'timing/model': 1.5114052270073444, 'learning_rate/qwen_vl_interface': 8.447935519092014e-06, 'learning_rate/action_model': 8.447935519092014e-05, 'epoch': 0.6}) +[2026-06-15 21:28:57,170][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.01794891245663166, 'train/grad_norm_pre_clip': 2.3683109283447266, 'timing/data': 0.0005060869734734297, 'timing/model': 1.7094068720471114, 'learning_rate/qwen_vl_interface': 8.445683030728925e-06, 'learning_rate/action_model': 8.445683030728925e-05, 'epoch': 0.6}) +[2026-06-15 21:38:16,849][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.19753773510456085, 'train/grad_norm_pre_clip': 10.39860725402832, 'eval/loss': 0.08111919462680817, 'eval/flap_precision': 0.8895348837209303, 'eval/flap_recall': 0.9198396793587175, 'eval/flap_f1': 0.9044334975369458, 'eval/flap_rate_model': 0.14333333333333334, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9044334975369459, 'timing/data': 0.0005084199947305024, 'timing/model': 1.4504158039926551, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 0.6}) +[2026-06-15 21:38:16,849][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1400_state +[2026-06-15 21:38:16,856][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 21:38:22,411][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1400_state/model.safetensors +[2026-06-15 21:38:36,035][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1400_state/optimizer.bin +[2026-06-15 21:38:36,036][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1400_state/sampler.bin +[2026-06-15 21:38:36,036][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1400_state/sampler_1.bin +[2026-06-15 21:38:36,037][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1400_state/random_states_0.pkl +[2026-06-15 21:38:36,148][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-06-15 21:38:36,148][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 21:38:36,158][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 21:38:41,971][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.02589082159101963, 'train/grad_norm_pre_clip': 3.3305933475494385, 'timing/data': 0.0005338469636626542, 'timing/model': 1.4879219779977575, 'learning_rate/qwen_vl_interface': 8.441174114055457e-06, 'learning_rate/action_model': 8.441174114055456e-05, 'epoch': 0.6}) +[2026-06-15 21:38:47,950][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.17321594059467316, 'train/grad_norm_pre_clip': 11.508252143859863, 'timing/data': 0.0006405930034816265, 'timing/model': 1.467529101006221, 'learning_rate/qwen_vl_interface': 8.43891768759852e-06, 'learning_rate/action_model': 8.43891768759852e-05, 'epoch': 0.6}) +[2026-06-15 21:38:53,954][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.05770716071128845, 'train/grad_norm_pre_clip': 6.42905330657959, 'timing/data': 0.0005075180088169873, 'timing/model': 1.4591327370144427, 'learning_rate/qwen_vl_interface': 8.436659950297978e-06, 'learning_rate/action_model': 8.436659950297977e-05, 'epoch': 0.6}) +[2026-06-15 21:38:59,951][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.018916161730885506, 'train/grad_norm_pre_clip': 2.781926393508911, 'timing/data': 0.0005198550061322749, 'timing/model': 1.479797973006498, 'learning_rate/qwen_vl_interface': 8.4344009030819e-06, 'learning_rate/action_model': 8.4344009030819e-05, 'epoch': 0.6}) +[2026-06-15 21:39:05,855][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.17887012660503387, 'train/grad_norm_pre_clip': 10.101609230041504, 'timing/data': 0.000512315018568188, 'timing/model': 1.43446102400776, 'learning_rate/qwen_vl_interface': 8.432140546878899e-06, 'learning_rate/action_model': 8.432140546878898e-05, 'epoch': 0.6}) +[2026-06-15 21:39:11,796][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.1151118278503418, 'train/grad_norm_pre_clip': 12.454093933105469, 'timing/data': 0.00051492900820449, 'timing/model': 1.7052694350131787, 'learning_rate/qwen_vl_interface': 8.429878882618116e-06, 'learning_rate/action_model': 8.429878882618116e-05, 'epoch': 0.6}) +[2026-06-15 21:39:17,684][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.2304813265800476, 'train/grad_norm_pre_clip': 16.956212997436523, 'timing/data': 0.0005197749705985188, 'timing/model': 1.6737305499846116, 'learning_rate/qwen_vl_interface': 8.42761591122924e-06, 'learning_rate/action_model': 8.427615911229241e-05, 'epoch': 0.61}) +[2026-06-15 21:39:23,464][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.1006951853632927, 'train/grad_norm_pre_clip': 20.38609504699707, 'timing/data': 0.00053469801787287, 'timing/model': 1.513522261986509, 'learning_rate/qwen_vl_interface': 8.425351633642491e-06, 'learning_rate/action_model': 8.425351633642492e-05, 'epoch': 0.61}) +[2026-06-15 21:39:29,559][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.09712019562721252, 'train/grad_norm_pre_clip': 8.33989429473877, 'timing/data': 0.0005164209869690239, 'timing/model': 1.4991447580396198, 'learning_rate/qwen_vl_interface': 8.423086050788632e-06, 'learning_rate/action_model': 8.423086050788631e-05, 'epoch': 0.61}) +[2026-06-15 21:39:35,560][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.3079027235507965, 'train/grad_norm_pre_clip': 15.716187477111816, 'timing/data': 0.0005254840361885726, 'timing/model': 1.4874326110002585, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 0.61}) +[2026-06-15 21:39:41,548][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.3784419596195221, 'train/grad_norm_pre_clip': 21.191465377807617, 'timing/data': 0.0005144579918123782, 'timing/model': 1.5195272820419632, 'learning_rate/qwen_vl_interface': 8.418550973005289e-06, 'learning_rate/action_model': 8.418550973005289e-05, 'epoch': 0.61}) +[2026-06-15 21:39:47,495][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.4383769929409027, 'train/grad_norm_pre_clip': 18.450355529785156, 'timing/data': 0.0005214280099608004, 'timing/model': 1.6744901980273426, 'learning_rate/qwen_vl_interface': 8.416281479940005e-06, 'learning_rate/action_model': 8.416281479940005e-05, 'epoch': 0.61}) +[2026-06-15 21:39:53,318][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.17384487390518188, 'train/grad_norm_pre_clip': 16.419607162475586, 'timing/data': 0.0005130959907546639, 'timing/model': 1.6166642460157163, 'learning_rate/qwen_vl_interface': 8.414010685336008e-06, 'learning_rate/action_model': 8.414010685336008e-05, 'epoch': 0.61}) +[2026-06-15 21:39:58,962][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.38084763288497925, 'train/grad_norm_pre_clip': 20.096391677856445, 'timing/data': 0.0006547850207425654, 'timing/model': 1.4625430239830166, 'learning_rate/qwen_vl_interface': 8.411738590126733e-06, 'learning_rate/action_model': 8.411738590126732e-05, 'epoch': 0.61}) +[2026-06-15 21:40:04,858][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.22406382858753204, 'train/grad_norm_pre_clip': 8.837306022644043, 'timing/data': 0.0005175420083105564, 'timing/model': 1.4614079300081357, 'learning_rate/qwen_vl_interface': 8.409465195246152e-06, 'learning_rate/action_model': 8.409465195246152e-05, 'epoch': 0.61}) +[2026-06-15 21:40:10,872][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.06961514800786972, 'train/grad_norm_pre_clip': 8.990923881530762, 'timing/data': 0.0005201360327191651, 'timing/model': 1.492904839979019, 'learning_rate/qwen_vl_interface': 8.407190501628772e-06, 'learning_rate/action_model': 8.407190501628772e-05, 'epoch': 0.61}) +[2026-06-15 21:40:16,956][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.09508516639471054, 'train/grad_norm_pre_clip': 8.644815444946289, 'timing/data': 0.00050588499289006, 'timing/model': 1.5121280809980817, 'learning_rate/qwen_vl_interface': 8.404914510209635e-06, 'learning_rate/action_model': 8.404914510209635e-05, 'epoch': 0.61}) +[2026-06-15 21:40:23,101][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.33935773372650146, 'train/grad_norm_pre_clip': 14.494483947753906, 'timing/data': 0.0005120040150359273, 'timing/model': 1.727696146001108, 'learning_rate/qwen_vl_interface': 8.402637221924315e-06, 'learning_rate/action_model': 8.402637221924314e-05, 'epoch': 0.61}) +[2026-06-15 21:40:28,774][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.22893279790878296, 'train/grad_norm_pre_clip': 13.848945617675781, 'timing/data': 0.0007468499825336039, 'timing/model': 1.472882328031119, 'learning_rate/qwen_vl_interface': 8.400358637708919e-06, 'learning_rate/action_model': 8.400358637708918e-05, 'epoch': 0.61}) +[2026-06-15 21:40:34,763][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.22308078408241272, 'train/grad_norm_pre_clip': 13.102355003356934, 'timing/data': 0.0005121240392327309, 'timing/model': 1.4663301049731672, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 0.61}) +[2026-06-15 21:40:40,763][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.23887738585472107, 'train/grad_norm_pre_clip': 12.191731452941895, 'timing/data': 0.0005139970453456044, 'timing/model': 1.4756967289722525, 'learning_rate/qwen_vl_interface': 8.395797585234987e-06, 'learning_rate/action_model': 8.395797585234988e-05, 'epoch': 0.61}) +[2026-06-15 21:40:46,750][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.2313225418329239, 'train/grad_norm_pre_clip': 9.651698112487793, 'timing/data': 0.0005158499698154628, 'timing/model': 1.5156444950262085, 'learning_rate/qwen_vl_interface': 8.39351511885133e-06, 'learning_rate/action_model': 8.393515118851329e-05, 'epoch': 0.61}) +[2026-06-15 21:40:52,775][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.267005980014801, 'train/grad_norm_pre_clip': 12.43492603302002, 'timing/data': 0.0005082989810034633, 'timing/model': 1.452274644980207, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.61}) +[2026-06-15 21:40:58,780][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.26409491896629333, 'train/grad_norm_pre_clip': 10.844098091125488, 'timing/data': 0.0007065010140649974, 'timing/model': 1.6765737109817564, 'learning_rate/qwen_vl_interface': 8.388946310481807e-06, 'learning_rate/action_model': 8.388946310481807e-05, 'epoch': 0.61}) +[2026-06-15 21:41:04,462][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.08975402265787125, 'train/grad_norm_pre_clip': 3.682323932647705, 'timing/data': 0.000512395054101944, 'timing/model': 1.4649694390245713, 'learning_rate/qwen_vl_interface': 8.386659970374007e-06, 'learning_rate/action_model': 8.386659970374006e-05, 'epoch': 0.61}) +[2026-06-15 21:41:10,632][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.1372067928314209, 'train/grad_norm_pre_clip': 6.392395973205566, 'timing/data': 0.0004999469965696335, 'timing/model': 1.5611409660195932, 'learning_rate/qwen_vl_interface': 8.384372340903775e-06, 'learning_rate/action_model': 8.384372340903775e-05, 'epoch': 0.61}) +[2026-06-15 21:41:16,562][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.03958079591393471, 'train/grad_norm_pre_clip': 1.5427454710006714, 'timing/data': 0.0004989349981769919, 'timing/model': 1.4615961879608221, 'learning_rate/qwen_vl_interface': 8.382083423011472e-06, 'learning_rate/action_model': 8.382083423011471e-05, 'epoch': 0.61}) +[2026-06-15 21:41:22,467][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.08929487317800522, 'train/grad_norm_pre_clip': 5.2307448387146, 'timing/data': 0.0005117440014146268, 'timing/model': 1.4411519349669106, 'learning_rate/qwen_vl_interface': 8.379793217637979e-06, 'learning_rate/action_model': 8.379793217637979e-05, 'epoch': 0.61}) +[2026-06-15 21:41:28,444][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.06772180646657944, 'train/grad_norm_pre_clip': 3.48578143119812, 'timing/data': 0.0005080580012872815, 'timing/model': 1.5148545259726234, 'learning_rate/qwen_vl_interface': 8.377501725724719e-06, 'learning_rate/action_model': 8.377501725724719e-05, 'epoch': 0.61}) +[2026-06-15 21:41:34,454][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.08941105008125305, 'train/grad_norm_pre_clip': 5.553213596343994, 'timing/data': 0.0005104920128360391, 'timing/model': 1.7156997079728171, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.62}) +[2026-06-15 21:41:40,162][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.25155824422836304, 'train/grad_norm_pre_clip': 8.209331512451172, 'timing/data': 0.0005057750386185944, 'timing/model': 1.4674666890059598, 'learning_rate/qwen_vl_interface': 8.372914886047203e-06, 'learning_rate/action_model': 8.372914886047203e-05, 'epoch': 0.62}) +[2026-06-15 21:41:46,159][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.08853083848953247, 'train/grad_norm_pre_clip': 5.220266342163086, 'timing/data': 0.0005295099690556526, 'timing/model': 1.4428378940210678, 'learning_rate/qwen_vl_interface': 8.370619540168422e-06, 'learning_rate/action_model': 8.370619540168423e-05, 'epoch': 0.62}) +[2026-06-15 21:41:52,161][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.14696671068668365, 'train/grad_norm_pre_clip': 8.086164474487305, 'timing/data': 0.0005238520097918808, 'timing/model': 1.4672300679958425, 'learning_rate/qwen_vl_interface': 8.368322911520826e-06, 'learning_rate/action_model': 8.368322911520826e-05, 'epoch': 0.62}) +[2026-06-15 21:41:58,156][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.3327096402645111, 'train/grad_norm_pre_clip': 9.48233699798584, 'timing/data': 0.0005176630220375955, 'timing/model': 1.4829867939697579, 'learning_rate/qwen_vl_interface': 8.36602500104847e-06, 'learning_rate/action_model': 8.36602500104847e-05, 'epoch': 0.62}) +[2026-06-15 21:42:03,962][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.07930894196033478, 'train/grad_norm_pre_clip': 4.028172016143799, 'timing/data': 0.000516009982675314, 'timing/model': 1.4647611409891397, 'learning_rate/qwen_vl_interface': 8.36372580969594e-06, 'learning_rate/action_model': 8.36372580969594e-05, 'epoch': 0.62}) +[2026-06-15 21:42:09,890][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.0714220181107521, 'train/grad_norm_pre_clip': 2.720010995864868, 'timing/data': 0.0005190049996599555, 'timing/model': 1.695774728024844, 'learning_rate/qwen_vl_interface': 8.361425338408344e-06, 'learning_rate/action_model': 8.361425338408343e-05, 'epoch': 0.62}) +[2026-06-15 21:42:15,558][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.3789941072463989, 'train/grad_norm_pre_clip': 9.430033683776855, 'timing/data': 0.0005189939984120429, 'timing/model': 1.464705098012928, 'learning_rate/qwen_vl_interface': 8.359123588131319e-06, 'learning_rate/action_model': 8.359123588131319e-05, 'epoch': 0.62}) +[2026-06-15 21:42:21,558][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.04319338500499725, 'train/grad_norm_pre_clip': 3.592644214630127, 'timing/data': 0.0005162700545042753, 'timing/model': 1.4663429300417192, 'learning_rate/qwen_vl_interface': 8.35682055981103e-06, 'learning_rate/action_model': 8.35682055981103e-05, 'epoch': 0.62}) +[2026-06-15 21:42:27,562][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.285429447889328, 'train/grad_norm_pre_clip': 12.726144790649414, 'timing/data': 0.0005135469837114215, 'timing/model': 1.5017859489889815, 'learning_rate/qwen_vl_interface': 8.354516254394163e-06, 'learning_rate/action_model': 8.354516254394162e-05, 'epoch': 0.62}) +[2026-06-15 21:42:33,456][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.20462167263031006, 'train/grad_norm_pre_clip': 7.922441482543945, 'timing/data': 0.0004926649853587151, 'timing/model': 1.4373099110089242, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 0.62}) +[2026-06-15 21:42:39,457][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.08996691554784775, 'train/grad_norm_pre_clip': 2.4334118366241455, 'timing/data': 0.0004930959548801184, 'timing/model': 1.4422687939950265, 'learning_rate/qwen_vl_interface': 8.349903816060074e-06, 'learning_rate/action_model': 8.349903816060075e-05, 'epoch': 0.62}) +[2026-06-15 21:42:45,393][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.04684709012508392, 'train/grad_norm_pre_clip': 3.5572731494903564, 'timing/data': 0.000519334978889674, 'timing/model': 1.6888492890284397, 'learning_rate/qwen_vl_interface': 8.34759568503885e-06, 'learning_rate/action_model': 8.347595685038851e-05, 'epoch': 0.62}) +[2026-06-15 21:42:51,062][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.15869206190109253, 'train/grad_norm_pre_clip': 5.98848295211792, 'timing/data': 0.0005135569954290986, 'timing/model': 1.4656616779975593, 'learning_rate/qwen_vl_interface': 8.345286280713049e-06, 'learning_rate/action_model': 8.345286280713048e-05, 'epoch': 0.62}) +[2026-06-15 21:42:57,062][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.13276909291744232, 'train/grad_norm_pre_clip': 8.074655532836914, 'timing/data': 0.0005148680065758526, 'timing/model': 1.4653026399901137, 'learning_rate/qwen_vl_interface': 8.342975604031974e-06, 'learning_rate/action_model': 8.342975604031974e-05, 'epoch': 0.62}) +[2026-06-15 21:43:03,058][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.2123112678527832, 'train/grad_norm_pre_clip': 7.2200517654418945, 'timing/data': 0.0005001160316169262, 'timing/model': 1.4858630200033076, 'learning_rate/qwen_vl_interface': 8.340663655945462e-06, 'learning_rate/action_model': 8.340663655945462e-05, 'epoch': 0.62}) +[2026-06-15 21:43:08,964][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.17931272089481354, 'train/grad_norm_pre_clip': 8.808990478515625, 'timing/data': 0.0005162610323168337, 'timing/model': 1.440287019009702, 'learning_rate/qwen_vl_interface': 8.338350437403866e-06, 'learning_rate/action_model': 8.338350437403865e-05, 'epoch': 0.62}) +[2026-06-15 21:43:14,890][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.483760267496109, 'train/grad_norm_pre_clip': 19.03623390197754, 'timing/data': 0.0005204270128160715, 'timing/model': 1.6939170790137723, 'learning_rate/qwen_vl_interface': 8.336035949358063e-06, 'learning_rate/action_model': 8.336035949358062e-05, 'epoch': 0.62}) +[2026-06-15 21:43:20,711][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.14633330702781677, 'train/grad_norm_pre_clip': 6.114869594573975, 'timing/data': 0.000517281994689256, 'timing/model': 1.6119112960295752, 'learning_rate/qwen_vl_interface': 8.333720192759448e-06, 'learning_rate/action_model': 8.333720192759448e-05, 'epoch': 0.62}) +[2026-06-15 21:43:26,363][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.1535617560148239, 'train/grad_norm_pre_clip': 4.534506797790527, 'timing/data': 0.0005207670037634671, 'timing/model': 1.4666079059825279, 'learning_rate/qwen_vl_interface': 8.331403168559946e-06, 'learning_rate/action_model': 8.331403168559945e-05, 'epoch': 0.62}) +[2026-06-15 21:43:32,278][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.09293346852064133, 'train/grad_norm_pre_clip': 4.988860607147217, 'timing/data': 0.0005080580012872815, 'timing/model': 1.4755608519772068, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 0.62}) +[2026-06-15 21:43:38,261][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.19092580676078796, 'train/grad_norm_pre_clip': 6.865053176879883, 'timing/data': 0.0005155500257387757, 'timing/model': 1.4897832780261524, 'learning_rate/qwen_vl_interface': 8.32676532116856e-06, 'learning_rate/action_model': 8.326765321168559e-05, 'epoch': 0.62}) +[2026-06-15 21:43:44,248][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.16356532275676727, 'train/grad_norm_pre_clip': 6.643462657928467, 'timing/data': 0.0005170409567654133, 'timing/model': 1.5135830200160854, 'learning_rate/qwen_vl_interface': 8.324444499883119e-06, 'learning_rate/action_model': 8.324444499883119e-05, 'epoch': 0.62}) +[2026-06-15 21:43:50,201][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.0490855947136879, 'train/grad_norm_pre_clip': 2.3772075176239014, 'timing/data': 0.0005303809884935617, 'timing/model': 1.7045732110273093, 'learning_rate/qwen_vl_interface': 8.322122414809678e-06, 'learning_rate/action_model': 8.322122414809677e-05, 'epoch': 0.63}) +[2026-06-15 21:43:55,962][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.18984092772006989, 'train/grad_norm_pre_clip': 8.837274551391602, 'timing/data': 0.0005046129808761179, 'timing/model': 1.489817263034638, 'learning_rate/qwen_vl_interface': 8.319799066902755e-06, 'learning_rate/action_model': 8.319799066902755e-05, 'epoch': 0.63}) +[2026-06-15 21:44:01,959][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.05796215310692787, 'train/grad_norm_pre_clip': 3.96770977973938, 'timing/data': 0.0005089699989184737, 'timing/model': 1.4670642559649423, 'learning_rate/qwen_vl_interface': 8.317474457117393e-06, 'learning_rate/action_model': 8.317474457117392e-05, 'epoch': 0.63}) +[2026-06-15 21:44:07,958][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.25350165367126465, 'train/grad_norm_pre_clip': 10.883611679077148, 'timing/data': 0.000503200979437679, 'timing/model': 1.4669603350339457, 'learning_rate/qwen_vl_interface': 8.315148586409148e-06, 'learning_rate/action_model': 8.315148586409148e-05, 'epoch': 0.63}) +[2026-06-15 21:44:13,860][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.17081233859062195, 'train/grad_norm_pre_clip': 7.582523345947266, 'timing/data': 0.00050985103007406, 'timing/model': 1.438033988000825, 'learning_rate/qwen_vl_interface': 8.312821455734104e-06, 'learning_rate/action_model': 8.312821455734103e-05, 'epoch': 0.63}) +[2026-06-15 21:44:19,763][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.12144874781370163, 'train/grad_norm_pre_clip': 5.884231090545654, 'timing/data': 0.000513326027430594, 'timing/model': 1.4398978149984032, 'learning_rate/qwen_vl_interface': 8.31049306604885e-06, 'learning_rate/action_model': 8.310493066048849e-05, 'epoch': 0.63}) +[2026-06-15 21:44:25,647][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.18316462635993958, 'train/grad_norm_pre_clip': 4.227227687835693, 'timing/data': 0.0005002760444767773, 'timing/model': 1.6483128909603693, 'learning_rate/qwen_vl_interface': 8.3081634183105e-06, 'learning_rate/action_model': 8.308163418310501e-05, 'epoch': 0.63}) +[2026-06-15 21:44:31,357][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.29770922660827637, 'train/grad_norm_pre_clip': 10.29211711883545, 'timing/data': 0.0005008470034226775, 'timing/model': 1.4625053820200264, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 0.63}) +[2026-06-15 21:44:37,359][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.1705268919467926, 'train/grad_norm_pre_clip': 6.825005054473877, 'timing/data': 0.0005260739708319306, 'timing/model': 1.4665044830180705, 'learning_rate/qwen_vl_interface': 8.303500352505557e-06, 'learning_rate/action_model': 8.303500352505557e-05, 'epoch': 0.63}) +[2026-06-15 21:44:43,364][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.08608800917863846, 'train/grad_norm_pre_clip': 4.6259684562683105, 'timing/data': 0.0005057750386185944, 'timing/model': 1.4655053750029765, 'learning_rate/qwen_vl_interface': 8.301166936355773e-06, 'learning_rate/action_model': 8.301166936355773e-05, 'epoch': 0.63}) +[2026-06-15 21:44:49,353][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.0556560754776001, 'train/grad_norm_pre_clip': 2.19252347946167, 'timing/data': 0.0006074439734220505, 'timing/model': 1.514845066005364, 'learning_rate/qwen_vl_interface': 8.298832265986514e-06, 'learning_rate/action_model': 8.298832265986513e-05, 'epoch': 0.63}) +[2026-06-15 21:44:55,334][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.19617140293121338, 'train/grad_norm_pre_clip': 6.009607315063477, 'timing/data': 0.0005034719943068922, 'timing/model': 1.5082198019954376, 'learning_rate/qwen_vl_interface': 8.296496342357475e-06, 'learning_rate/action_model': 8.296496342357474e-05, 'epoch': 0.63}) +[2026-06-15 21:45:01,265][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.18166311085224152, 'train/grad_norm_pre_clip': 7.5700249671936035, 'timing/data': 0.000505554024130106, 'timing/model': 1.6657914289971814, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 0.63}) +[2026-06-15 21:45:06,957][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.2002829611301422, 'train/grad_norm_pre_clip': 4.582210540771484, 'timing/data': 0.0005540960119105875, 'timing/model': 1.4632975239655934, 'learning_rate/qwen_vl_interface': 8.29182073916141e-06, 'learning_rate/action_model': 8.29182073916141e-05, 'epoch': 0.63}) +[2026-06-15 21:45:12,957][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.18117201328277588, 'train/grad_norm_pre_clip': 11.577061653137207, 'timing/data': 0.0005515230004675686, 'timing/model': 1.4638628879911266, 'learning_rate/qwen_vl_interface': 8.28948106151635e-06, 'learning_rate/action_model': 8.28948106151635e-05, 'epoch': 0.63}) +[2026-06-15 21:45:18,857][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.23895207047462463, 'train/grad_norm_pre_clip': 12.263269424438477, 'timing/data': 0.0004909330164082348, 'timing/model': 1.4595628390088677, 'learning_rate/qwen_vl_interface': 8.287140134455435e-06, 'learning_rate/action_model': 8.287140134455435e-05, 'epoch': 0.63}) +[2026-06-15 21:45:24,762][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.1494060754776001, 'train/grad_norm_pre_clip': 10.064180374145508, 'timing/data': 0.00061511603416875, 'timing/model': 1.4401071509928443, 'learning_rate/qwen_vl_interface': 8.284797958940934e-06, 'learning_rate/action_model': 8.284797958940935e-05, 'epoch': 0.63}) +[2026-06-15 21:45:30,566][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.14678193628787994, 'train/grad_norm_pre_clip': 6.427920818328857, 'timing/data': 0.0004917440237477422, 'timing/model': 1.463821115961764, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 0.63}) +[2026-06-15 21:45:36,495][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.17314626276493073, 'train/grad_norm_pre_clip': 5.060735702514648, 'timing/data': 0.0004887200193479657, 'timing/model': 1.6985109720262699, 'learning_rate/qwen_vl_interface': 8.280109866402805e-06, 'learning_rate/action_model': 8.280109866402806e-05, 'epoch': 0.63}) +[2026-06-15 21:45:42,160][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.14952337741851807, 'train/grad_norm_pre_clip': 11.810705184936523, 'timing/data': 0.0004971620510332286, 'timing/model': 1.4638243869994767, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.63}) +[2026-06-15 21:45:48,162][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.19493237137794495, 'train/grad_norm_pre_clip': 9.827413558959961, 'timing/data': 0.0004817990120500326, 'timing/model': 1.4640916740172543, 'learning_rate/qwen_vl_interface': 8.275416791610356e-06, 'learning_rate/action_model': 8.275416791610355e-05, 'epoch': 0.63}) +[2026-06-15 21:45:54,194][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.04378000646829605, 'train/grad_norm_pre_clip': 1.6963512897491455, 'timing/data': 0.0005692890263162553, 'timing/model': 1.4584190030000173, 'learning_rate/qwen_vl_interface': 8.27306838827987e-06, 'learning_rate/action_model': 8.273068388279871e-05, 'epoch': 0.63}) +[2026-06-15 21:46:00,141][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.07157672941684723, 'train/grad_norm_pre_clip': 4.983488082885742, 'timing/data': 0.0004914539749734104, 'timing/model': 1.5121406160178594, 'learning_rate/qwen_vl_interface': 8.270718742280162e-06, 'learning_rate/action_model': 8.270718742280162e-05, 'epoch': 0.63}) +[2026-06-15 21:46:06,144][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.08870184421539307, 'train/grad_norm_pre_clip': 3.5257720947265625, 'timing/data': 0.0005214479751884937, 'timing/model': 1.5136006290558726, 'learning_rate/qwen_vl_interface': 8.268367854577084e-06, 'learning_rate/action_model': 8.268367854577082e-05, 'epoch': 0.64}) +[2026-06-15 21:46:12,094][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.11432633548974991, 'train/grad_norm_pre_clip': 5.4580254554748535, 'timing/data': 0.0004982330137863755, 'timing/model': 1.699616143014282, 'learning_rate/qwen_vl_interface': 8.266015726136991e-06, 'learning_rate/action_model': 8.266015726136991e-05, 'epoch': 0.64}) +[2026-06-15 21:46:17,764][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.15950480103492737, 'train/grad_norm_pre_clip': 4.846052646636963, 'timing/data': 0.0005003259866498411, 'timing/model': 1.4617439130088314, 'learning_rate/qwen_vl_interface': 8.26366235792676e-06, 'learning_rate/action_model': 8.26366235792676e-05, 'epoch': 0.64}) +[2026-06-15 21:46:23,765][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.09828601777553558, 'train/grad_norm_pre_clip': 3.7973053455352783, 'timing/data': 0.0005114530213177204, 'timing/model': 1.465214989031665, 'learning_rate/qwen_vl_interface': 8.261307750913769e-06, 'learning_rate/action_model': 8.261307750913769e-05, 'epoch': 0.64}) +[2026-06-15 21:46:29,764][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.09020591527223587, 'train/grad_norm_pre_clip': 4.975818157196045, 'timing/data': 0.0005083390278741717, 'timing/model': 1.4852916559902951, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 0.64}) +[2026-06-15 21:46:35,764][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.3897039592266083, 'train/grad_norm_pre_clip': 8.090230941772461, 'timing/data': 0.000517281994689256, 'timing/model': 1.4784116190276109, 'learning_rate/qwen_vl_interface': 8.256594824351577e-06, 'learning_rate/action_model': 8.256594824351576e-05, 'epoch': 0.64}) +[2026-06-15 21:46:41,692][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.09016634523868561, 'train/grad_norm_pre_clip': 4.003823757171631, 'timing/data': 0.0005122039583511651, 'timing/model': 1.6939189179684035, 'learning_rate/qwen_vl_interface': 8.254236506739683e-06, 'learning_rate/action_model': 8.254236506739681e-05, 'epoch': 0.64}) +[2026-06-15 21:46:47,708][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.021848618984222412, 'train/grad_norm_pre_clip': 0.9854744076728821, 'timing/data': 0.0006474340334534645, 'timing/model': 1.65432247298304, 'learning_rate/qwen_vl_interface': 8.25187695419964e-06, 'learning_rate/action_model': 8.25187695419964e-05, 'epoch': 0.64}) +[2026-06-15 21:46:53,466][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.08496276289224625, 'train/grad_norm_pre_clip': 3.5171220302581787, 'timing/data': 0.000515048042871058, 'timing/model': 1.4615608570165932, 'learning_rate/qwen_vl_interface': 8.249516167701373e-06, 'learning_rate/action_model': 8.249516167701374e-05, 'epoch': 0.64}) +[2026-06-15 21:46:59,363][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.17975708842277527, 'train/grad_norm_pre_clip': 10.778558731079102, 'timing/data': 0.0005280280020087957, 'timing/model': 1.4633414230193011, 'learning_rate/qwen_vl_interface': 8.247154148215311e-06, 'learning_rate/action_model': 8.24715414821531e-05, 'epoch': 0.64}) +[2026-06-15 21:47:05,364][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.20478874444961548, 'train/grad_norm_pre_clip': 10.647156715393066, 'timing/data': 0.0005066560115665197, 'timing/model': 1.4658814769936725, 'learning_rate/qwen_vl_interface': 8.24479089671239e-06, 'learning_rate/action_model': 8.24479089671239e-05, 'epoch': 0.64}) +[2026-06-15 21:47:11,444][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.24674323201179504, 'train/grad_norm_pre_clip': 9.876275062561035, 'timing/data': 0.0005184930050745606, 'timing/model': 1.510687519039493, 'learning_rate/qwen_vl_interface': 8.242426414164056e-06, 'learning_rate/action_model': 8.242426414164056e-05, 'epoch': 0.64}) +[2026-06-15 21:47:17,398][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.16351374983787537, 'train/grad_norm_pre_clip': 7.114494800567627, 'timing/data': 0.0005164510221220553, 'timing/model': 1.69484825496329, 'learning_rate/qwen_vl_interface': 8.240060701542257e-06, 'learning_rate/action_model': 8.240060701542257e-05, 'epoch': 0.64}) +[2026-06-15 21:47:23,164][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.08622828125953674, 'train/grad_norm_pre_clip': 8.539752960205078, 'timing/data': 0.0005268159438855946, 'timing/model': 1.4901430360041559, 'learning_rate/qwen_vl_interface': 8.237693759819447e-06, 'learning_rate/action_model': 8.237693759819448e-05, 'epoch': 0.64}) +[2026-06-15 21:47:29,261][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.16967537999153137, 'train/grad_norm_pre_clip': 12.879321098327637, 'timing/data': 0.0005111120408400893, 'timing/model': 1.504324883047957, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 0.64}) +[2026-06-15 21:47:35,285][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.24215109646320343, 'train/grad_norm_pre_clip': 10.13333797454834, 'timing/data': 0.0005871939938515425, 'timing/model': 1.4941111069638282, 'learning_rate/qwen_vl_interface': 8.232956192963143e-06, 'learning_rate/action_model': 8.232956192963144e-05, 'epoch': 0.64}) +[2026-06-15 21:47:41,234][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.0062582120299339294, 'train/grad_norm_pre_clip': 0.5700933933258057, 'timing/data': 0.0005899080424569547, 'timing/model': 1.5084910260047764, 'learning_rate/qwen_vl_interface': 8.230585569777086e-06, 'learning_rate/action_model': 8.230585569777085e-05, 'epoch': 0.64}) +[2026-06-15 21:47:47,200][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.16172149777412415, 'train/grad_norm_pre_clip': 11.109339714050293, 'timing/data': 0.0005900280084460974, 'timing/model': 1.4750033099553548, 'learning_rate/qwen_vl_interface': 8.228213721384885e-06, 'learning_rate/action_model': 8.228213721384885e-05, 'epoch': 0.64}) +[2026-06-15 21:47:53,052][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.10351881384849548, 'train/grad_norm_pre_clip': 6.323953151702881, 'timing/data': 0.0005909199826419353, 'timing/model': 1.6512214369722642, 'learning_rate/qwen_vl_interface': 8.225840648761523e-06, 'learning_rate/action_model': 8.225840648761523e-05, 'epoch': 0.64}) +[2026-06-15 21:47:58,864][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.13834016025066376, 'train/grad_norm_pre_clip': 12.70750904083252, 'timing/data': 0.0005867439904250205, 'timing/model': 1.4842689939541742, 'learning_rate/qwen_vl_interface': 8.223466352882476e-06, 'learning_rate/action_model': 8.223466352882475e-05, 'epoch': 0.64}) +[2026-06-15 21:48:04,944][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.0649854764342308, 'train/grad_norm_pre_clip': 11.31578540802002, 'timing/data': 0.0005138670094311237, 'timing/model': 1.5114720540004782, 'learning_rate/qwen_vl_interface': 8.22109083472373e-06, 'learning_rate/action_model': 8.22109083472373e-05, 'epoch': 0.64}) +[2026-06-15 21:48:11,030][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.2878776490688324, 'train/grad_norm_pre_clip': 10.466197967529297, 'timing/data': 0.000588655995670706, 'timing/model': 1.5478003610041924, 'learning_rate/qwen_vl_interface': 8.218714095261769e-06, 'learning_rate/action_model': 8.218714095261768e-05, 'epoch': 0.64}) +[2026-06-15 21:48:17,042][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.15278424322605133, 'train/grad_norm_pre_clip': 14.218022346496582, 'timing/data': 0.0005064159631729126, 'timing/model': 1.5129515719600022, 'learning_rate/qwen_vl_interface': 8.21633613547358e-06, 'learning_rate/action_model': 8.21633613547358e-05, 'epoch': 0.64}) +[2026-06-15 21:48:23,163][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.26198476552963257, 'train/grad_norm_pre_clip': 8.537485122680664, 'timing/data': 0.0005136569961905479, 'timing/model': 1.507264212006703, 'learning_rate/qwen_vl_interface': 8.213956956336656e-06, 'learning_rate/action_model': 8.213956956336656e-05, 'epoch': 0.64}) +[2026-06-15 21:57:28,651][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.18974284827709198, 'train/grad_norm_pre_clip': 7.5185065269470215, 'eval/loss': 0.07963841408491135, 'eval/flap_precision': 0.9092783505154639, 'eval/flap_recall': 0.8837675350701403, 'eval/flap_f1': 0.896341463414634, 'eval/flap_rate_model': 0.13472222222222222, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.896341463414634, 'timing/data': 0.0005085989832878113, 'timing/model': 1.7263793099555187, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.65}) +[2026-06-15 21:57:28,652][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1500_state +[2026-06-15 21:57:28,657][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 21:57:34,234][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1500_state/model.safetensors +[2026-06-15 21:57:42,427][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1500_state/optimizer.bin +[2026-06-15 21:57:42,427][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1500_state/sampler.bin +[2026-06-15 21:57:42,427][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-15 21:57:42,428][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-15 21:57:42,539][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-15 21:57:42,539][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 21:57:42,549][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 21:57:48,498][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.08759818971157074, 'train/grad_norm_pre_clip': 4.645554065704346, 'timing/data': 0.0005264719948172569, 'timing/model': 1.729918789002113, 'learning_rate/qwen_vl_interface': 8.209194943929059e-06, 'learning_rate/action_model': 8.20919494392906e-05, 'epoch': 0.65}) +[2026-06-15 21:57:54,134][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.21183079481124878, 'train/grad_norm_pre_clip': 9.887167930603027, 'timing/data': 0.0005134429666213691, 'timing/model': 1.467781144019682, 'learning_rate/qwen_vl_interface': 8.20681211261587e-06, 'learning_rate/action_model': 8.20681211261587e-05, 'epoch': 0.65}) +[2026-06-15 21:58:00,033][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.14805100858211517, 'train/grad_norm_pre_clip': 9.285271644592285, 'timing/data': 0.0005706659867428243, 'timing/model': 1.461308091005776, 'learning_rate/qwen_vl_interface': 8.204428065868911e-06, 'learning_rate/action_model': 8.204428065868911e-05, 'epoch': 0.65}) +[2026-06-15 21:58:05,941][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.0845029354095459, 'train/grad_norm_pre_clip': 18.931148529052734, 'timing/data': 0.0005180200096219778, 'timing/model': 1.4673549099825323, 'learning_rate/qwen_vl_interface': 8.202042804668173e-06, 'learning_rate/action_model': 8.202042804668172e-05, 'epoch': 0.65}) +[2026-06-15 21:58:11,942][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.6243852972984314, 'train/grad_norm_pre_clip': 21.237306594848633, 'timing/data': 0.0005238380399532616, 'timing/model': 1.4615947809652425, 'learning_rate/qwen_vl_interface': 8.199656329994145e-06, 'learning_rate/action_model': 8.199656329994145e-05, 'epoch': 0.65}) +[2026-06-15 21:58:18,043][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.08563926070928574, 'train/grad_norm_pre_clip': 9.65987777709961, 'timing/data': 0.0005316790193319321, 'timing/model': 1.4799226169707254, 'learning_rate/qwen_vl_interface': 8.19726864282782e-06, 'learning_rate/action_model': 8.197268642827818e-05, 'epoch': 0.65}) +[2026-06-15 21:58:23,980][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.0368526391685009, 'train/grad_norm_pre_clip': 1.9505815505981445, 'timing/data': 0.0005135730025358498, 'timing/model': 1.6966514029772952, 'learning_rate/qwen_vl_interface': 8.194879744150682e-06, 'learning_rate/action_model': 8.194879744150681e-05, 'epoch': 0.65}) +[2026-06-15 21:58:29,846][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.08225245028734207, 'train/grad_norm_pre_clip': 6.088641166687012, 'timing/data': 0.0005202119937166572, 'timing/model': 1.4711673649726436, 'learning_rate/qwen_vl_interface': 8.192489634944718e-06, 'learning_rate/action_model': 8.192489634944718e-05, 'epoch': 0.65}) +[2026-06-15 21:58:35,848][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.10897078365087509, 'train/grad_norm_pre_clip': 5.195736885070801, 'timing/data': 0.0005157359992153943, 'timing/model': 1.4395033700275235, 'learning_rate/qwen_vl_interface': 8.190098316192416e-06, 'learning_rate/action_model': 8.190098316192416e-05, 'epoch': 0.65}) +[2026-06-15 21:58:41,753][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.193415567278862, 'train/grad_norm_pre_clip': 8.243889808654785, 'timing/data': 0.0005177990533411503, 'timing/model': 1.4533984729787335, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 0.65}) +[2026-06-15 21:58:47,832][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.08334068953990936, 'train/grad_norm_pre_clip': 5.163661003112793, 'timing/data': 0.0005059009999968112, 'timing/model': 1.557208193000406, 'learning_rate/qwen_vl_interface': 8.185312053981204e-06, 'learning_rate/action_model': 8.185312053981204e-05, 'epoch': 0.65}) +[2026-06-15 21:58:53,845][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.16630105674266815, 'train/grad_norm_pre_clip': 6.661147117614746, 'timing/data': 0.0005289260298013687, 'timing/model': 1.5023635429679416, 'learning_rate/qwen_vl_interface': 8.182917112489747e-06, 'learning_rate/action_model': 8.182917112489746e-05, 'epoch': 0.65}) +[2026-06-15 21:58:59,843][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.033839259296655655, 'train/grad_norm_pre_clip': 2.3596889972686768, 'timing/data': 0.0005181089509278536, 'timing/model': 1.7447097430122085, 'learning_rate/qwen_vl_interface': 8.18052096538685e-06, 'learning_rate/action_model': 8.18052096538685e-05, 'epoch': 0.65}) +[2026-06-15 21:59:05,549][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.0436348132789135, 'train/grad_norm_pre_clip': 3.4556384086608887, 'timing/data': 0.0005092469509691, 'timing/model': 1.4653747439733706, 'learning_rate/qwen_vl_interface': 8.178123613657483e-06, 'learning_rate/action_model': 8.178123613657482e-05, 'epoch': 0.65}) +[2026-06-15 21:59:11,550][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.062056802213191986, 'train/grad_norm_pre_clip': 4.84609842300415, 'timing/data': 0.0005082149873487651, 'timing/model': 1.4648532650317065, 'learning_rate/qwen_vl_interface': 8.175725058287102e-06, 'learning_rate/action_model': 8.175725058287102e-05, 'epoch': 0.65}) +[2026-06-15 21:59:17,455][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.049667470157146454, 'train/grad_norm_pre_clip': 3.57375431060791, 'timing/data': 0.0005027070292271674, 'timing/model': 1.465697548002936, 'learning_rate/qwen_vl_interface': 8.173325300261664e-06, 'learning_rate/action_model': 8.173325300261663e-05, 'epoch': 0.65}) +[2026-06-15 21:59:23,354][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.10070730000734329, 'train/grad_norm_pre_clip': 8.487021446228027, 'timing/data': 0.0005075749941170216, 'timing/model': 1.4360917679732665, 'learning_rate/qwen_vl_interface': 8.170924340567617e-06, 'learning_rate/action_model': 8.170924340567616e-05, 'epoch': 0.65}) +[2026-06-15 21:59:29,240][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.16519705951213837, 'train/grad_norm_pre_clip': 12.202493667602539, 'timing/data': 0.0005122619913890958, 'timing/model': 1.5049839630373754, 'learning_rate/qwen_vl_interface': 8.168522180191909e-06, 'learning_rate/action_model': 8.168522180191909e-05, 'epoch': 0.65}) +[2026-06-15 21:59:35,132][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.15914088487625122, 'train/grad_norm_pre_clip': 14.091652870178223, 'timing/data': 0.0005149449571035802, 'timing/model': 1.6428481789771467, 'learning_rate/qwen_vl_interface': 8.166118820121972e-06, 'learning_rate/action_model': 8.166118820121973e-05, 'epoch': 0.65}) +[2026-06-15 21:59:40,855][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.08267521113157272, 'train/grad_norm_pre_clip': 2.5005500316619873, 'timing/data': 0.0005046700243838131, 'timing/model': 1.4653781820088625, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 0.65}) +[2026-06-15 21:59:46,853][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.23992955684661865, 'train/grad_norm_pre_clip': 20.073040008544922, 'timing/data': 0.0005248600500635803, 'timing/model': 1.46360816701781, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.65}) +[2026-06-15 21:59:52,859][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.3003537356853485, 'train/grad_norm_pre_clip': 11.735688209533691, 'timing/data': 0.000522146001458168, 'timing/model': 1.4705114630050957, 'learning_rate/qwen_vl_interface': 8.158901551628579e-06, 'learning_rate/action_model': 8.158901551628579e-05, 'epoch': 0.65}) +[2026-06-15 21:59:58,848][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.1279280036687851, 'train/grad_norm_pre_clip': 12.612403869628906, 'timing/data': 0.0005244190106168389, 'timing/model': 1.5166969440178946, 'learning_rate/qwen_vl_interface': 8.156493402665971e-06, 'learning_rate/action_model': 8.15649340266597e-05, 'epoch': 0.66}) +[2026-06-15 22:00:04,785][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.18198484182357788, 'train/grad_norm_pre_clip': 11.03460693359375, 'timing/data': 0.0005112100043334067, 'timing/model': 1.6931402059854008, 'learning_rate/qwen_vl_interface': 8.154084058953713e-06, 'learning_rate/action_model': 8.154084058953713e-05, 'epoch': 0.66}) +[2026-06-15 22:00:10,555][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.12866899371147156, 'train/grad_norm_pre_clip': 6.92122220993042, 'timing/data': 0.0005192219978198409, 'timing/model': 1.5667681189952418, 'learning_rate/qwen_vl_interface': 8.151673521482193e-06, 'learning_rate/action_model': 8.151673521482194e-05, 'epoch': 0.66}) +[2026-06-15 22:00:16,258][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.07683375477790833, 'train/grad_norm_pre_clip': 7.021897315979004, 'timing/data': 0.0005242089973762631, 'timing/model': 1.465836481016595, 'learning_rate/qwen_vl_interface': 8.149261791242297e-06, 'learning_rate/action_model': 8.149261791242297e-05, 'epoch': 0.66}) +[2026-06-15 22:00:22,254][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.19984941184520721, 'train/grad_norm_pre_clip': 10.442756652832031, 'timing/data': 0.0005184400361031294, 'timing/model': 1.4632054069661535, 'learning_rate/qwen_vl_interface': 8.146848869225391e-06, 'learning_rate/action_model': 8.146848869225391e-05, 'epoch': 0.66}) +[2026-06-15 22:00:28,158][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.05860692635178566, 'train/grad_norm_pre_clip': 6.622519493103027, 'timing/data': 0.00048782600788399577, 'timing/model': 1.4654176970361732, 'learning_rate/qwen_vl_interface': 8.14443475642334e-06, 'learning_rate/action_model': 8.144434756423339e-05, 'epoch': 0.66}) +[2026-06-15 22:00:34,053][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.037333983927965164, 'train/grad_norm_pre_clip': 3.67140793800354, 'timing/data': 0.0004913409939035773, 'timing/model': 1.4317685689893551, 'learning_rate/qwen_vl_interface': 8.142019453828493e-06, 'learning_rate/action_model': 8.142019453828493e-05, 'epoch': 0.66}) +[2026-06-15 22:00:40,084][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.1906077116727829, 'train/grad_norm_pre_clip': 7.102936744689941, 'timing/data': 0.0004889669944532216, 'timing/model': 1.724992869014386, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 0.66}) +[2026-06-15 22:00:45,753][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.2962303161621094, 'train/grad_norm_pre_clip': 8.677419662475586, 'timing/data': 0.0005203839973546565, 'timing/model': 1.4624774700496346, 'learning_rate/qwen_vl_interface': 8.137185283232255e-06, 'learning_rate/action_model': 8.137185283232256e-05, 'epoch': 0.66}) +[2026-06-15 22:00:51,752][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.022328542545437813, 'train/grad_norm_pre_clip': 8.765157699584961, 'timing/data': 0.0005089269834570587, 'timing/model': 1.4605127300019376, 'learning_rate/qwen_vl_interface': 8.134766417218014e-06, 'learning_rate/action_model': 8.134766417218012e-05, 'epoch': 0.66}) +[2026-06-15 22:00:57,658][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.1586744785308838, 'train/grad_norm_pre_clip': 8.043597221374512, 'timing/data': 0.0005908670136705041, 'timing/model': 1.467570076987613, 'learning_rate/qwen_vl_interface': 8.132346365385262e-06, 'learning_rate/action_model': 8.132346365385261e-05, 'epoch': 0.66}) +[2026-06-15 22:01:03,556][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.18070575594902039, 'train/grad_norm_pre_clip': 11.644599914550781, 'timing/data': 0.0005156369879841805, 'timing/model': 1.465816073003225, 'learning_rate/qwen_vl_interface': 8.129925128728796e-06, 'learning_rate/action_model': 8.129925128728796e-05, 'epoch': 0.66}) +[2026-06-15 22:01:09,553][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.10245514661073685, 'train/grad_norm_pre_clip': 8.413847923278809, 'timing/data': 0.0005031489999964833, 'timing/model': 1.4298682200023904, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 0.66}) +[2026-06-15 22:01:15,486][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.10177943855524063, 'train/grad_norm_pre_clip': 8.92121696472168, 'timing/data': 0.0005433769547380507, 'timing/model': 1.696834352973383, 'learning_rate/qwen_vl_interface': 8.125079104926324e-06, 'learning_rate/action_model': 8.125079104926323e-05, 'epoch': 0.66}) +[2026-06-15 22:01:21,155][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.022850334644317627, 'train/grad_norm_pre_clip': 1.448726773262024, 'timing/data': 0.0005126320174895227, 'timing/model': 1.4676770780351944, 'learning_rate/qwen_vl_interface': 8.122654319772332e-06, 'learning_rate/action_model': 8.122654319772332e-05, 'epoch': 0.66}) +[2026-06-15 22:01:27,155][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.1338445395231247, 'train/grad_norm_pre_clip': 9.34116268157959, 'timing/data': 0.0005099279806017876, 'timing/model': 1.4499666549963877, 'learning_rate/qwen_vl_interface': 8.120228353778661e-06, 'learning_rate/action_model': 8.12022835377866e-05, 'epoch': 0.66}) +[2026-06-15 22:01:32,954][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.07511700689792633, 'train/grad_norm_pre_clip': 8.896342277526855, 'timing/data': 0.0005812530289404094, 'timing/model': 1.4635834830114618, 'learning_rate/qwen_vl_interface': 8.117801207942528e-06, 'learning_rate/action_model': 8.117801207942528e-05, 'epoch': 0.66}) +[2026-06-15 22:01:38,755][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.15546941757202148, 'train/grad_norm_pre_clip': 11.284272193908691, 'timing/data': 0.0005716090090572834, 'timing/model': 1.4583439800189808, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 0.66}) +[2026-06-15 22:01:44,554][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.0955514907836914, 'train/grad_norm_pre_clip': 8.395332336425781, 'timing/data': 0.0005757350008934736, 'timing/model': 1.4346385819953866, 'learning_rate/qwen_vl_interface': 8.11294338073421e-06, 'learning_rate/action_model': 8.11294338073421e-05, 'epoch': 0.66}) +[2026-06-15 22:01:50,416][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.046294864267110825, 'train/grad_norm_pre_clip': 4.383398532867432, 'timing/data': 0.0005817530327476561, 'timing/model': 1.6263070029672235, 'learning_rate/qwen_vl_interface': 8.110512701358892e-06, 'learning_rate/action_model': 8.110512701358891e-05, 'epoch': 0.66}) +[2026-06-15 22:01:56,056][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.025330306962132454, 'train/grad_norm_pre_clip': 1.2410004138946533, 'timing/data': 0.000521365029271692, 'timing/model': 1.463996215024963, 'learning_rate/qwen_vl_interface': 8.108080846134852e-06, 'learning_rate/action_model': 8.108080846134852e-05, 'epoch': 0.66}) +[2026-06-15 22:02:01,952][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.0173800028860569, 'train/grad_norm_pre_clip': 2.1482064723968506, 'timing/data': 0.0005077949608676136, 'timing/model': 1.4595005890005268, 'learning_rate/qwen_vl_interface': 8.105647816061734e-06, 'learning_rate/action_model': 8.105647816061735e-05, 'epoch': 0.66}) +[2026-06-15 22:02:07,855][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.18509967625141144, 'train/grad_norm_pre_clip': 5.111687660217285, 'timing/data': 0.0005115409730933607, 'timing/model': 1.456500538974069, 'learning_rate/qwen_vl_interface': 8.103213612139667e-06, 'learning_rate/action_model': 8.103213612139667e-05, 'epoch': 0.66}) +[2026-06-15 22:02:13,749][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.2894918620586395, 'train/grad_norm_pre_clip': 9.282058715820312, 'timing/data': 0.0005079560214653611, 'timing/model': 1.4376805459614843, 'learning_rate/qwen_vl_interface': 8.100778235369262e-06, 'learning_rate/action_model': 8.10077823536926e-05, 'epoch': 0.67}) +[2026-06-15 22:02:19,678][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.5748761892318726, 'train/grad_norm_pre_clip': 28.520231246948242, 'timing/data': 0.00050895701861009, 'timing/model': 1.6902107300120406, 'learning_rate/qwen_vl_interface': 8.098341686751605e-06, 'learning_rate/action_model': 8.098341686751605e-05, 'epoch': 0.67}) +[2026-06-15 22:02:25,529][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.06137832626700401, 'train/grad_norm_pre_clip': 8.410871505737305, 'timing/data': 0.0005169590003788471, 'timing/model': 1.6436384730041027, 'learning_rate/qwen_vl_interface': 8.095903967288273e-06, 'learning_rate/action_model': 8.095903967288272e-05, 'epoch': 0.67}) +[2026-06-15 22:02:31,254][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.11624468863010406, 'train/grad_norm_pre_clip': 8.368053436279297, 'timing/data': 0.0005201839958317578, 'timing/model': 1.4652576170046814, 'learning_rate/qwen_vl_interface': 8.093465077981318e-06, 'learning_rate/action_model': 8.093465077981318e-05, 'epoch': 0.67}) +[2026-06-15 22:02:37,253][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.29481786489486694, 'train/grad_norm_pre_clip': 9.43668270111084, 'timing/data': 0.0005102590075694025, 'timing/model': 1.4669937470462173, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 0.67}) +[2026-06-15 22:02:43,249][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.18337324261665344, 'train/grad_norm_pre_clip': 15.320221900939941, 'timing/data': 0.0005020470125600696, 'timing/model': 1.4650618510204367, 'learning_rate/qwen_vl_interface': 8.088583793847165e-06, 'learning_rate/action_model': 8.088583793847164e-05, 'epoch': 0.67}) +[2026-06-15 22:02:49,150][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.03410651907324791, 'train/grad_norm_pre_clip': 5.858914852142334, 'timing/data': 0.0005147760384716094, 'timing/model': 1.4353594039566815, 'learning_rate/qwen_vl_interface': 8.086141401026478e-06, 'learning_rate/action_model': 8.086141401026478e-05, 'epoch': 0.67}) +[2026-06-15 22:02:55,076][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.36433207988739014, 'train/grad_norm_pre_clip': 11.964661598205566, 'timing/data': 0.0005140949506312609, 'timing/model': 1.6902818370144814, 'learning_rate/qwen_vl_interface': 8.083697842375193e-06, 'learning_rate/action_model': 8.083697842375193e-05, 'epoch': 0.67}) +[2026-06-15 22:03:01,057][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.029383786022663116, 'train/grad_norm_pre_clip': 6.803391456604004, 'timing/data': 0.0005110299680382013, 'timing/model': 1.5695240719942376, 'learning_rate/qwen_vl_interface': 8.081253118897762e-06, 'learning_rate/action_model': 8.081253118897761e-05, 'epoch': 0.67}) +[2026-06-15 22:03:06,754][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.41727912425994873, 'train/grad_norm_pre_clip': 14.921417236328125, 'timing/data': 0.0005076759844087064, 'timing/model': 1.4653195799910463, 'learning_rate/qwen_vl_interface': 8.07880723159912e-06, 'learning_rate/action_model': 8.07880723159912e-05, 'epoch': 0.67}) +[2026-06-15 22:03:12,756][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.14833757281303406, 'train/grad_norm_pre_clip': 7.829457759857178, 'timing/data': 0.0005086460150778294, 'timing/model': 1.4650621989858337, 'learning_rate/qwen_vl_interface': 8.076360181484676e-06, 'learning_rate/action_model': 8.076360181484676e-05, 'epoch': 0.67}) +[2026-06-15 22:03:18,654][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.10416436195373535, 'train/grad_norm_pre_clip': 6.295219898223877, 'timing/data': 0.000511271005962044, 'timing/model': 1.4613505870220251, 'learning_rate/qwen_vl_interface': 8.073911969560326e-06, 'learning_rate/action_model': 8.073911969560326e-05, 'epoch': 0.67}) +[2026-06-15 22:03:24,559][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.34226030111312866, 'train/grad_norm_pre_clip': 10.677153587341309, 'timing/data': 0.0005120209534652531, 'timing/model': 1.43359143397538, 'learning_rate/qwen_vl_interface': 8.071462596832432e-06, 'learning_rate/action_model': 8.071462596832432e-05, 'epoch': 0.67}) +[2026-06-15 22:03:30,485][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.20879140496253967, 'train/grad_norm_pre_clip': 7.598793983459473, 'timing/data': 0.0005127530312165618, 'timing/model': 1.6904113089549355, 'learning_rate/qwen_vl_interface': 8.06901206430784e-06, 'learning_rate/action_model': 8.06901206430784e-05, 'epoch': 0.67}) +[2026-06-15 22:03:36,156][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.08262251317501068, 'train/grad_norm_pre_clip': 6.634962558746338, 'timing/data': 0.0005072149797342718, 'timing/model': 1.466088906978257, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 0.67}) +[2026-06-15 22:03:42,158][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.11025857925415039, 'train/grad_norm_pre_clip': 5.647039890289307, 'timing/data': 0.0005163380410522223, 'timing/model': 1.4635271499864757, 'learning_rate/qwen_vl_interface': 8.064107523898327e-06, 'learning_rate/action_model': 8.064107523898326e-05, 'epoch': 0.67}) +[2026-06-15 22:03:48,057][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.029635729268193245, 'train/grad_norm_pre_clip': 2.3803398609161377, 'timing/data': 0.0005143359885551035, 'timing/model': 1.4630037340102717, 'learning_rate/qwen_vl_interface': 8.061653518029476e-06, 'learning_rate/action_model': 8.061653518029476e-05, 'epoch': 0.67}) +[2026-06-15 22:03:53,955][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.06513982266187668, 'train/grad_norm_pre_clip': 6.32677698135376, 'timing/data': 0.0005122920265421271, 'timing/model': 1.4647602069890127, 'learning_rate/qwen_vl_interface': 8.059198356396069e-06, 'learning_rate/action_model': 8.059198356396068e-05, 'epoch': 0.67}) +[2026-06-15 22:03:59,858][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.23490628600120544, 'train/grad_norm_pre_clip': 9.575915336608887, 'timing/data': 0.000513323990162462, 'timing/model': 1.434777860995382, 'learning_rate/qwen_vl_interface': 8.05674204000733e-06, 'learning_rate/action_model': 8.05674204000733e-05, 'epoch': 0.67}) +[2026-06-15 22:04:05,782][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.07394518703222275, 'train/grad_norm_pre_clip': 5.800490856170654, 'timing/data': 0.0005152969970367849, 'timing/model': 1.6931533640017733, 'learning_rate/qwen_vl_interface': 8.05428456987296e-06, 'learning_rate/action_model': 8.05428456987296e-05, 'epoch': 0.67}) +[2026-06-15 22:04:11,452][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.26201122999191284, 'train/grad_norm_pre_clip': 11.016521453857422, 'timing/data': 0.0005149259814061224, 'timing/model': 1.4618126439745538, 'learning_rate/qwen_vl_interface': 8.05182594700313e-06, 'learning_rate/action_model': 8.05182594700313e-05, 'epoch': 0.67}) +[2026-06-15 22:04:17,454][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.14194455742835999, 'train/grad_norm_pre_clip': 6.123376369476318, 'timing/data': 0.000513965031132102, 'timing/model': 1.4684970409725793, 'learning_rate/qwen_vl_interface': 8.049366172408484e-06, 'learning_rate/action_model': 8.049366172408485e-05, 'epoch': 0.67}) +[2026-06-15 22:04:23,334][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.05821866914629936, 'train/grad_norm_pre_clip': 6.993646621704102, 'timing/data': 0.0005093379877507687, 'timing/model': 1.5402273219660856, 'learning_rate/qwen_vl_interface': 8.04690524710015e-06, 'learning_rate/action_model': 8.046905247100151e-05, 'epoch': 0.67}) +[2026-06-15 22:04:29,388][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.17561176419258118, 'train/grad_norm_pre_clip': 9.241463661193848, 'timing/data': 0.0008199540316127241, 'timing/model': 1.5146379049983807, 'learning_rate/qwen_vl_interface': 8.044443172089715e-06, 'learning_rate/action_model': 8.044443172089714e-05, 'epoch': 0.68}) +[2026-06-15 22:04:35,341][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.10664207488298416, 'train/grad_norm_pre_clip': 4.636128902435303, 'timing/data': 0.0005039600073359907, 'timing/model': 1.5296985350432806, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.68}) +[2026-06-15 22:04:41,468][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.13544052839279175, 'train/grad_norm_pre_clip': 7.776520729064941, 'timing/data': 0.0005143450107425451, 'timing/model': 1.6774591660359874, 'learning_rate/qwen_vl_interface': 8.039515577011287e-06, 'learning_rate/action_model': 8.039515577011288e-05, 'epoch': 0.68}) +[2026-06-15 22:04:47,244][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.24775436520576477, 'train/grad_norm_pre_clip': 9.108743667602539, 'timing/data': 0.0005253719864413142, 'timing/model': 1.55565052194288, 'learning_rate/qwen_vl_interface': 8.037050058968841e-06, 'learning_rate/action_model': 8.03705005896884e-05, 'epoch': 0.68}) +[2026-06-15 22:04:53,270][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.046372588723897934, 'train/grad_norm_pre_clip': 2.3180651664733887, 'timing/data': 0.0005345550016500056, 'timing/model': 1.4791671720449813, 'learning_rate/qwen_vl_interface': 8.034583395275392e-06, 'learning_rate/action_model': 8.034583395275392e-05, 'epoch': 0.68}) +[2026-06-15 22:04:59,451][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.06939657032489777, 'train/grad_norm_pre_clip': 3.8036997318267822, 'timing/data': 0.0005789299611933529, 'timing/model': 1.5486130590434186, 'learning_rate/qwen_vl_interface': 8.032115586944893e-06, 'learning_rate/action_model': 8.032115586944892e-05, 'epoch': 0.68}) +[2026-06-15 22:05:05,612][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.16987326741218567, 'train/grad_norm_pre_clip': 6.502492904663086, 'timing/data': 0.0005261219921521842, 'timing/model': 1.5606761379749514, 'learning_rate/qwen_vl_interface': 8.029646634991762e-06, 'learning_rate/action_model': 8.029646634991762e-05, 'epoch': 0.68}) +[2026-06-15 22:05:11,556][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.18534812331199646, 'train/grad_norm_pre_clip': 3.007124423980713, 'timing/data': 0.0005175899714231491, 'timing/model': 1.764022059040144, 'learning_rate/qwen_vl_interface': 8.0271765404309e-06, 'learning_rate/action_model': 8.0271765404309e-05, 'epoch': 0.68}) +[2026-06-15 22:05:17,533][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.14581429958343506, 'train/grad_norm_pre_clip': 5.528903961181641, 'timing/data': 0.0005249310052022338, 'timing/model': 1.641129697964061, 'learning_rate/qwen_vl_interface': 8.024705304277664e-06, 'learning_rate/action_model': 8.024705304277664e-05, 'epoch': 0.68}) +[2026-06-15 22:05:23,257][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.22584569454193115, 'train/grad_norm_pre_clip': 4.748701572418213, 'timing/data': 0.0005107299657538533, 'timing/model': 1.468642612977419, 'learning_rate/qwen_vl_interface': 8.022232927547888e-06, 'learning_rate/action_model': 8.022232927547886e-05, 'epoch': 0.68}) +[2026-06-15 22:05:29,426][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.026806917041540146, 'train/grad_norm_pre_clip': 1.7283178567886353, 'timing/data': 0.000521626032423228, 'timing/model': 1.5567882540053688, 'learning_rate/qwen_vl_interface': 8.01975941125787e-06, 'learning_rate/action_model': 8.01975941125787e-05, 'epoch': 0.68}) +[2026-06-15 22:05:35,441][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.3720926344394684, 'train/grad_norm_pre_clip': 10.014867782592773, 'timing/data': 0.0005208849906921387, 'timing/model': 1.5428271350101568, 'learning_rate/qwen_vl_interface': 8.017284756424381e-06, 'learning_rate/action_model': 8.017284756424381e-05, 'epoch': 0.68}) +[2026-06-15 22:05:41,459][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.22578322887420654, 'train/grad_norm_pre_clip': 6.3619890213012695, 'timing/data': 0.0005212050164118409, 'timing/model': 1.479217335989233, 'learning_rate/qwen_vl_interface': 8.01480896406466e-06, 'learning_rate/action_model': 8.01480896406466e-05, 'epoch': 0.68}) +[2026-06-15 22:05:47,488][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.21320560574531555, 'train/grad_norm_pre_clip': 7.468632698059082, 'timing/data': 0.0005711890407837927, 'timing/model': 1.7313409320195206, 'learning_rate/qwen_vl_interface': 8.01233203519641e-06, 'learning_rate/action_model': 8.012332035196409e-05, 'epoch': 0.68}) +[2026-06-15 22:05:53,364][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.23403006792068481, 'train/grad_norm_pre_clip': 7.623688220977783, 'timing/data': 0.000568735005799681, 'timing/model': 1.6118793489877135, 'learning_rate/qwen_vl_interface': 8.009853970837801e-06, 'learning_rate/action_model': 8.009853970837801e-05, 'epoch': 0.68}) +[2026-06-15 22:05:59,155][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.19825129210948944, 'train/grad_norm_pre_clip': 10.634698867797852, 'timing/data': 0.0005636989953927696, 'timing/model': 1.4642150899744593, 'learning_rate/qwen_vl_interface': 8.007374772007474e-06, 'learning_rate/action_model': 8.007374772007474e-05, 'epoch': 0.68}) +[2026-06-15 22:06:05,155][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.22405895590782166, 'train/grad_norm_pre_clip': 8.256486892700195, 'timing/data': 0.0005676739965565503, 'timing/model': 1.4630592259927653, 'learning_rate/qwen_vl_interface': 8.004894439724533e-06, 'learning_rate/action_model': 8.004894439724533e-05, 'epoch': 0.68}) +[2026-06-15 22:06:11,056][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.1902303695678711, 'train/grad_norm_pre_clip': 11.3259916305542, 'timing/data': 0.0005060029798187315, 'timing/model': 1.4631540279951878, 'learning_rate/qwen_vl_interface': 8.002412975008552e-06, 'learning_rate/action_model': 8.002412975008552e-05, 'epoch': 0.68}) +[2026-06-15 22:06:17,237][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.5762539505958557, 'train/grad_norm_pre_clip': 22.11231231689453, 'timing/data': 0.0005690960097126663, 'timing/model': 1.4995015480089933, 'learning_rate/qwen_vl_interface': 7.999930378879562e-06, 'learning_rate/action_model': 7.999930378879561e-05, 'epoch': 0.68}) +[2026-06-15 22:06:23,283][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.11910060793161392, 'train/grad_norm_pre_clip': 6.637042045593262, 'timing/data': 0.0005545940366573632, 'timing/model': 1.695109382038936, 'learning_rate/qwen_vl_interface': 7.997446652358067e-06, 'learning_rate/action_model': 7.997446652358068e-05, 'epoch': 0.68}) +[2026-06-15 22:06:28,988][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.1292688101530075, 'train/grad_norm_pre_clip': 5.844550132751465, 'timing/data': 0.0005350660067051649, 'timing/model': 1.4931630619685166, 'learning_rate/qwen_vl_interface': 7.994961796465035e-06, 'learning_rate/action_model': 7.994961796465035e-05, 'epoch': 0.68}) +[2026-06-15 22:06:35,282][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.18490883708000183, 'train/grad_norm_pre_clip': 8.494781494140625, 'timing/data': 0.0005192129756323993, 'timing/model': 1.501766384986695, 'learning_rate/qwen_vl_interface': 7.992475812221894e-06, 'learning_rate/action_model': 7.992475812221894e-05, 'epoch': 0.68}) +[2026-06-15 22:06:41,336][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.08761920034885406, 'train/grad_norm_pre_clip': 4.561844825744629, 'timing/data': 0.000532081990968436, 'timing/model': 1.5401528120273724, 'learning_rate/qwen_vl_interface': 7.989988700650538e-06, 'learning_rate/action_model': 7.989988700650538e-05, 'epoch': 0.68}) +[2026-06-15 22:06:47,433][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.14121749997138977, 'train/grad_norm_pre_clip': 5.080204486846924, 'timing/data': 0.0004979419754818082, 'timing/model': 1.5458582780556753, 'learning_rate/qwen_vl_interface': 7.987500462773326e-06, 'learning_rate/action_model': 7.987500462773326e-05, 'epoch': 0.69}) +[2026-06-15 22:06:53,579][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.0990249365568161, 'train/grad_norm_pre_clip': 4.952712535858154, 'timing/data': 0.0005531630013138056, 'timing/model': 1.4547236880171113, 'learning_rate/qwen_vl_interface': 7.985011099613078e-06, 'learning_rate/action_model': 7.985011099613077e-05, 'epoch': 0.69}) +[2026-06-15 22:06:59,678][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.17206740379333496, 'train/grad_norm_pre_clip': 5.839359283447266, 'timing/data': 0.000511051039211452, 'timing/model': 1.7804317790432833, 'learning_rate/qwen_vl_interface': 7.98252061219308e-06, 'learning_rate/action_model': 7.982520612193078e-05, 'epoch': 0.69}) +[2026-06-15 22:07:05,560][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.1784495711326599, 'train/grad_norm_pre_clip': 9.498353958129883, 'timing/data': 0.0008196649723686278, 'timing/model': 1.4809502310235985, 'learning_rate/qwen_vl_interface': 7.980029001537073e-06, 'learning_rate/action_model': 7.980029001537072e-05, 'epoch': 0.69}) +[2026-06-15 22:07:11,656][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.289670467376709, 'train/grad_norm_pre_clip': 8.58372974395752, 'timing/data': 0.000511571008246392, 'timing/model': 1.455245624994859, 'learning_rate/qwen_vl_interface': 7.977536268669264e-06, 'learning_rate/action_model': 7.977536268669264e-05, 'epoch': 0.69}) +[2026-06-15 22:07:17,450][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.2456643283367157, 'train/grad_norm_pre_clip': 3.5427966117858887, 'timing/data': 0.0005092889769002795, 'timing/model': 1.4663736030342989, 'learning_rate/qwen_vl_interface': 7.975042414614324e-06, 'learning_rate/action_model': 7.975042414614324e-05, 'epoch': 0.69}) +[2026-06-15 22:07:23,351][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.08702029287815094, 'train/grad_norm_pre_clip': 3.205507755279541, 'timing/data': 0.0005188820068724453, 'timing/model': 1.464595619996544, 'learning_rate/qwen_vl_interface': 7.972547440397382e-06, 'learning_rate/action_model': 7.972547440397383e-05, 'epoch': 0.69}) +[2026-06-15 22:07:29,398][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.14416655898094177, 'train/grad_norm_pre_clip': 6.2092604637146, 'timing/data': 0.0005090879858471453, 'timing/model': 1.5568749539670534, 'learning_rate/qwen_vl_interface': 7.970051347044026e-06, 'learning_rate/action_model': 7.970051347044025e-05, 'epoch': 0.69}) +[2026-06-15 22:16:31,755][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.08334749191999435, 'train/grad_norm_pre_clip': 4.378327369689941, 'eval/loss': 0.08142215758562088, 'eval/flap_precision': 0.848987108655617, 'eval/flap_recall': 0.9238476953907816, 'eval/flap_f1': 0.8848368522072937, 'eval/flap_rate_model': 0.15083333333333335, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.8848368522072937, 'timing/data': 0.0005197640275582671, 'timing/model': 1.7030636810231954, 'learning_rate/qwen_vl_interface': 7.967554135580307e-06, 'learning_rate/action_model': 7.967554135580307e-05, 'epoch': 0.69}) +[2026-06-15 22:16:31,756][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1600_state +[2026-06-15 22:16:31,760][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 22:16:37,272][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1600_state/model.safetensors +[2026-06-15 22:16:50,779][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1600_state/optimizer.bin +[2026-06-15 22:16:50,779][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1600_state/sampler.bin +[2026-06-15 22:16:50,779][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1600_state/sampler_1.bin +[2026-06-15 22:16:50,780][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1600_state/random_states_0.pkl +[2026-06-15 22:16:50,887][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1600 +[2026-06-15 22:16:50,887][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 22:16:50,897][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 22:16:56,908][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.12672537565231323, 'train/grad_norm_pre_clip': 5.471191883087158, 'timing/data': 0.0005388720310293138, 'timing/model': 1.7292487259837799, 'learning_rate/qwen_vl_interface': 7.965055807032734e-06, 'learning_rate/action_model': 7.965055807032735e-05, 'epoch': 0.69}) +[2026-06-15 22:17:02,643][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.198572039604187, 'train/grad_norm_pre_clip': 4.630706310272217, 'timing/data': 0.0005126039613969624, 'timing/model': 1.4885949649615213, 'learning_rate/qwen_vl_interface': 7.962556362428275e-06, 'learning_rate/action_model': 7.962556362428274e-05, 'epoch': 0.69}) +[2026-06-15 22:17:08,540][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.17162059247493744, 'train/grad_norm_pre_clip': 5.2840046882629395, 'timing/data': 0.0005290079861879349, 'timing/model': 1.4654650239972398, 'learning_rate/qwen_vl_interface': 7.960055802794358e-06, 'learning_rate/action_model': 7.960055802794357e-05, 'epoch': 0.69}) +[2026-06-15 22:17:14,754][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.02709011547267437, 'train/grad_norm_pre_clip': 1.255751371383667, 'timing/data': 0.0005067249876447022, 'timing/model': 1.4648569679702632, 'learning_rate/qwen_vl_interface': 7.957554129158866e-06, 'learning_rate/action_model': 7.957554129158866e-05, 'epoch': 0.69}) +[2026-06-15 22:17:20,759][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.19493450224399567, 'train/grad_norm_pre_clip': 6.858506679534912, 'timing/data': 0.0005319420015439391, 'timing/model': 1.489506223006174, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 0.69}) +[2026-06-15 22:17:26,731][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.25931939482688904, 'train/grad_norm_pre_clip': 16.319814682006836, 'timing/data': 0.0005355280009098351, 'timing/model': 1.5029478769865818, 'learning_rate/qwen_vl_interface': 7.952547443997e-06, 'learning_rate/action_model': 7.952547443997e-05, 'epoch': 0.69}) +[2026-06-15 22:17:32,680][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.33245566487312317, 'train/grad_norm_pre_clip': 11.911285400390625, 'timing/data': 0.0005138859851285815, 'timing/model': 1.6954041630378924, 'learning_rate/qwen_vl_interface': 7.950042434528681e-06, 'learning_rate/action_model': 7.950042434528682e-05, 'epoch': 0.69}) +[2026-06-15 22:17:38,353][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.1711522489786148, 'train/grad_norm_pre_clip': 16.453718185424805, 'timing/data': 0.0004849830293096602, 'timing/model': 1.4690676350146532, 'learning_rate/qwen_vl_interface': 7.94753631517491e-06, 'learning_rate/action_model': 7.947536315174909e-05, 'epoch': 0.69}) +[2026-06-15 22:17:44,416][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.05593349039554596, 'train/grad_norm_pre_clip': 2.9198145866394043, 'timing/data': 0.0005016979994252324, 'timing/model': 1.5243925270042382, 'learning_rate/qwen_vl_interface': 7.945029086965854e-06, 'learning_rate/action_model': 7.945029086965853e-05, 'epoch': 0.69}) +[2026-06-15 22:17:50,354][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.02622852846980095, 'train/grad_norm_pre_clip': 1.7547680139541626, 'timing/data': 0.0005240110331214964, 'timing/model': 1.466171943990048, 'learning_rate/qwen_vl_interface': 7.942520750932138e-06, 'learning_rate/action_model': 7.942520750932138e-05, 'epoch': 0.69}) +[2026-06-15 22:17:56,260][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.0541713610291481, 'train/grad_norm_pre_clip': 4.527308940887451, 'timing/data': 0.000512354017701, 'timing/model': 1.4587076759780757, 'learning_rate/qwen_vl_interface': 7.940011308104848e-06, 'learning_rate/action_model': 7.940011308104848e-05, 'epoch': 0.69}) +[2026-06-15 22:18:02,161][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.08227532356977463, 'train/grad_norm_pre_clip': 3.311587333679199, 'timing/data': 0.0005133950035087764, 'timing/model': 1.4600727320066653, 'learning_rate/qwen_vl_interface': 7.93750075951552e-06, 'learning_rate/action_model': 7.93750075951552e-05, 'epoch': 0.69}) +[2026-06-15 22:18:08,180][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.027808940038084984, 'train/grad_norm_pre_clip': 2.0148024559020996, 'timing/data': 0.000503390037920326, 'timing/model': 1.7818532820092514, 'learning_rate/qwen_vl_interface': 7.934989106196145e-06, 'learning_rate/action_model': 7.934989106196146e-05, 'epoch': 0.69}) +[2026-06-15 22:18:13,862][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.12835045158863068, 'train/grad_norm_pre_clip': 3.5068750381469727, 'timing/data': 0.0005223880289122462, 'timing/model': 1.4590202699764632, 'learning_rate/qwen_vl_interface': 7.93247634917917e-06, 'learning_rate/action_model': 7.93247634917917e-05, 'epoch': 0.69}) +[2026-06-15 22:18:19,860][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.22341716289520264, 'train/grad_norm_pre_clip': 10.594430923461914, 'timing/data': 0.0005089179612696171, 'timing/model': 1.458766785042826, 'learning_rate/qwen_vl_interface': 7.929962489497494e-06, 'learning_rate/action_model': 7.929962489497493e-05, 'epoch': 0.69}) +[2026-06-15 22:18:25,968][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.44257745146751404, 'train/grad_norm_pre_clip': 15.611451148986816, 'timing/data': 0.0006520689930766821, 'timing/model': 1.4627744559547864, 'learning_rate/qwen_vl_interface': 7.927447528184469e-06, 'learning_rate/action_model': 7.927447528184468e-05, 'epoch': 0.7}) +[2026-06-15 22:18:31,968][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.3265419900417328, 'train/grad_norm_pre_clip': 11.82542610168457, 'timing/data': 0.000507826975081116, 'timing/model': 1.5098172510042787, 'learning_rate/qwen_vl_interface': 7.924931466273901e-06, 'learning_rate/action_model': 7.924931466273902e-05, 'epoch': 0.7}) +[2026-06-15 22:18:38,066][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.21408958733081818, 'train/grad_norm_pre_clip': 10.356285095214844, 'timing/data': 0.0005079569527879357, 'timing/model': 1.513955664995592, 'learning_rate/qwen_vl_interface': 7.92241430480005e-06, 'learning_rate/action_model': 7.922414304800049e-05, 'epoch': 0.7}) +[2026-06-15 22:18:43,991][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.053499724715948105, 'train/grad_norm_pre_clip': 8.12206745147705, 'timing/data': 0.0005136450054123998, 'timing/model': 1.696383255010005, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.7}) +[2026-06-15 22:18:49,665][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.1461682766675949, 'train/grad_norm_pre_clip': 9.898648262023926, 'timing/data': 0.0006573970313183963, 'timing/model': 1.463510839967057, 'learning_rate/qwen_vl_interface': 7.91737668730178e-06, 'learning_rate/action_model': 7.91737668730178e-05, 'epoch': 0.7}) +[2026-06-15 22:18:55,662][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.05648655444383621, 'train/grad_norm_pre_clip': 8.56603717803955, 'timing/data': 0.0005065950099378824, 'timing/model': 1.4646814789739437, 'learning_rate/qwen_vl_interface': 7.91485623334814e-06, 'learning_rate/action_model': 7.91485623334814e-05, 'epoch': 0.7}) +[2026-06-15 22:19:01,461][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.2246149629354477, 'train/grad_norm_pre_clip': 12.975135803222656, 'timing/data': 0.0005111119826324284, 'timing/model': 1.462353462993633, 'learning_rate/qwen_vl_interface': 7.91233468397276e-06, 'learning_rate/action_model': 7.912334683972761e-05, 'epoch': 0.7}) +[2026-06-15 22:19:07,449][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.4312485456466675, 'train/grad_norm_pre_clip': 10.402071952819824, 'timing/data': 0.0005553870578296483, 'timing/model': 1.552192452014424, 'learning_rate/qwen_vl_interface': 7.90981204021216e-06, 'learning_rate/action_model': 7.90981204021216e-05, 'epoch': 0.7}) +[2026-06-15 22:19:13,442][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.08766841143369675, 'train/grad_norm_pre_clip': 12.208306312561035, 'timing/data': 0.00047787302173674107, 'timing/model': 1.5119406479643658, 'learning_rate/qwen_vl_interface': 7.907288303103298e-06, 'learning_rate/action_model': 7.907288303103298e-05, 'epoch': 0.7}) +[2026-06-15 22:19:19,493][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.06159818917512894, 'train/grad_norm_pre_clip': 3.861224412918091, 'timing/data': 0.0004920639912597835, 'timing/model': 1.6966159979929216, 'learning_rate/qwen_vl_interface': 7.90476347368359e-06, 'learning_rate/action_model': 7.90476347368359e-05, 'epoch': 0.7}) +[2026-06-15 22:19:25,233][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.038028161972761154, 'train/grad_norm_pre_clip': 4.466080665588379, 'timing/data': 0.0005145769682712853, 'timing/model': 1.534133116947487, 'learning_rate/qwen_vl_interface': 7.902237552990897e-06, 'learning_rate/action_model': 7.902237552990897e-05, 'epoch': 0.7}) +[2026-06-15 22:19:31,263][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.09426367282867432, 'train/grad_norm_pre_clip': 11.066967964172363, 'timing/data': 0.0005197140271775424, 'timing/model': 1.4658464150270447, 'learning_rate/qwen_vl_interface': 7.89971054206353e-06, 'learning_rate/action_model': 7.89971054206353e-05, 'epoch': 0.7}) +[2026-06-15 22:19:37,264][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.19174987077713013, 'train/grad_norm_pre_clip': 10.101262092590332, 'timing/data': 0.0005095590022392571, 'timing/model': 1.486184349982068, 'learning_rate/qwen_vl_interface': 7.897182441940247e-06, 'learning_rate/action_model': 7.897182441940247e-05, 'epoch': 0.7}) +[2026-06-15 22:19:43,243][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.10038059204816818, 'train/grad_norm_pre_clip': 6.629166126251221, 'timing/data': 0.0005103910225443542, 'timing/model': 1.5153847779729404, 'learning_rate/qwen_vl_interface': 7.894653253660256e-06, 'learning_rate/action_model': 7.894653253660256e-05, 'epoch': 0.7}) +[2026-06-15 22:19:49,162][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.20284156501293182, 'train/grad_norm_pre_clip': 25.52018928527832, 'timing/data': 0.0005166199989616871, 'timing/model': 1.513594091986306, 'learning_rate/qwen_vl_interface': 7.89212297826321e-06, 'learning_rate/action_model': 7.892122978263209e-05, 'epoch': 0.7}) +[2026-06-15 22:19:55,040][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.3306105136871338, 'train/grad_norm_pre_clip': 11.535900115966797, 'timing/data': 0.000501957954838872, 'timing/model': 1.6369698100024834, 'learning_rate/qwen_vl_interface': 7.889591616789206e-06, 'learning_rate/action_model': 7.889591616789206e-05, 'epoch': 0.7}) +[2026-06-15 22:20:00,758][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.21674498915672302, 'train/grad_norm_pre_clip': 10.211318969726562, 'timing/data': 0.0005964270094409585, 'timing/model': 1.462984579964541, 'learning_rate/qwen_vl_interface': 7.887059170278799e-06, 'learning_rate/action_model': 7.887059170278797e-05, 'epoch': 0.7}) +[2026-06-15 22:20:06,859][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.05257643759250641, 'train/grad_norm_pre_clip': 1.9575248956680298, 'timing/data': 0.0005033899797126651, 'timing/model': 1.4605220850207843, 'learning_rate/qwen_vl_interface': 7.884525639772974e-06, 'learning_rate/action_model': 7.884525639772974e-05, 'epoch': 0.7}) +[2026-06-15 22:20:12,762][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.12081770598888397, 'train/grad_norm_pre_clip': 6.013581275939941, 'timing/data': 0.0005206959904171526, 'timing/model': 1.4616129259811714, 'learning_rate/qwen_vl_interface': 7.881991026313176e-06, 'learning_rate/action_model': 7.881991026313175e-05, 'epoch': 0.7}) +[2026-06-15 22:20:18,753][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.2280513346195221, 'train/grad_norm_pre_clip': 9.678220748901367, 'timing/data': 0.0005287679960019886, 'timing/model': 1.518303279997781, 'learning_rate/qwen_vl_interface': 7.879455330941283e-06, 'learning_rate/action_model': 7.879455330941282e-05, 'epoch': 0.7}) +[2026-06-15 22:20:24,729][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.07286322861909866, 'train/grad_norm_pre_clip': 7.361813068389893, 'timing/data': 0.0005104700103402138, 'timing/model': 1.727364840975497, 'learning_rate/qwen_vl_interface': 7.876918554699627e-06, 'learning_rate/action_model': 7.876918554699627e-05, 'epoch': 0.7}) +[2026-06-15 22:20:30,681][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.07106804102659225, 'train/grad_norm_pre_clip': 4.802302360534668, 'timing/data': 0.0005314419977366924, 'timing/model': 1.5817577689886093, 'learning_rate/qwen_vl_interface': 7.874380698630982e-06, 'learning_rate/action_model': 7.874380698630982e-05, 'epoch': 0.7}) +[2026-06-15 22:20:36,364][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.3415633738040924, 'train/grad_norm_pre_clip': 7.178119659423828, 'timing/data': 0.0005167800118215382, 'timing/model': 1.4629557650187053, 'learning_rate/qwen_vl_interface': 7.871841763778562e-06, 'learning_rate/action_model': 7.871841763778563e-05, 'epoch': 0.7}) +[2026-06-15 22:20:42,364][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.1651081144809723, 'train/grad_norm_pre_clip': 10.63940143585205, 'timing/data': 0.0007859249599277973, 'timing/model': 1.4618359739542939, 'learning_rate/qwen_vl_interface': 7.86930175118603e-06, 'learning_rate/action_model': 7.86930175118603e-05, 'epoch': 0.71}) +[2026-06-15 22:20:48,363][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.20058616995811462, 'train/grad_norm_pre_clip': 11.944476127624512, 'timing/data': 0.0005181109881959856, 'timing/model': 1.478578081005253, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.71}) +[2026-06-15 22:20:54,349][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.06403753161430359, 'train/grad_norm_pre_clip': 7.04630708694458, 'timing/data': 0.0005127750337123871, 'timing/model': 1.5152971469797194, 'learning_rate/qwen_vl_interface': 7.86421849695748e-06, 'learning_rate/action_model': 7.864218496957481e-05, 'epoch': 0.71}) +[2026-06-15 22:21:00,298][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.18050600588321686, 'train/grad_norm_pre_clip': 6.777245998382568, 'timing/data': 0.0005144869792275131, 'timing/model': 1.7004879880114459, 'learning_rate/qwen_vl_interface': 7.861675257410998e-06, 'learning_rate/action_model': 7.861675257410997e-05, 'epoch': 0.71}) +[2026-06-15 22:21:05,965][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.09262386709451675, 'train/grad_norm_pre_clip': 3.57193660736084, 'timing/data': 0.0005108609911985695, 'timing/model': 1.4652202539728023, 'learning_rate/qwen_vl_interface': 7.859130944303468e-06, 'learning_rate/action_model': 7.859130944303467e-05, 'epoch': 0.71}) +[2026-06-15 22:21:11,963][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.17074552178382874, 'train/grad_norm_pre_clip': 6.162480354309082, 'timing/data': 0.0005057439557276666, 'timing/model': 1.4654374399688095, 'learning_rate/qwen_vl_interface': 7.856585558680761e-06, 'learning_rate/action_model': 7.856585558680761e-05, 'epoch': 0.71}) +[2026-06-15 22:21:17,864][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.21706336736679077, 'train/grad_norm_pre_clip': 7.784573554992676, 'timing/data': 0.0005340250208973885, 'timing/model': 1.4624994749901816, 'learning_rate/qwen_vl_interface': 7.854039101589191e-06, 'learning_rate/action_model': 7.854039101589191e-05, 'epoch': 0.71}) +[2026-06-15 22:21:23,759][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.12129122763872147, 'train/grad_norm_pre_clip': 3.7142739295959473, 'timing/data': 0.0004919639904983342, 'timing/model': 1.4689134519430809, 'learning_rate/qwen_vl_interface': 7.851491574075509e-06, 'learning_rate/action_model': 7.851491574075509e-05, 'epoch': 0.71}) +[2026-06-15 22:21:29,667][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.13982149958610535, 'train/grad_norm_pre_clip': 7.828459739685059, 'timing/data': 0.0005044309655204415, 'timing/model': 1.4398130209883675, 'learning_rate/qwen_vl_interface': 7.848942977186909e-06, 'learning_rate/action_model': 7.848942977186908e-05, 'epoch': 0.71}) +[2026-06-15 22:21:35,773][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.2334129363298416, 'train/grad_norm_pre_clip': 8.21603012084961, 'timing/data': 0.00048502301797270775, 'timing/model': 1.7805415960028768, 'learning_rate/qwen_vl_interface': 7.84639331197102e-06, 'learning_rate/action_model': 7.84639331197102e-05, 'epoch': 0.71}) +[2026-06-15 22:21:41,463][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.07830260694026947, 'train/grad_norm_pre_clip': 2.726879835128784, 'timing/data': 0.0004950279835611582, 'timing/model': 1.4598607759689912, 'learning_rate/qwen_vl_interface': 7.843842579475918e-06, 'learning_rate/action_model': 7.843842579475918e-05, 'epoch': 0.71}) +[2026-06-15 22:21:47,467][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.19269070029258728, 'train/grad_norm_pre_clip': 6.870146751403809, 'timing/data': 0.0005158690037205815, 'timing/model': 1.467447948991321, 'learning_rate/qwen_vl_interface': 7.841290780750107e-06, 'learning_rate/action_model': 7.841290780750107e-05, 'epoch': 0.71}) +[2026-06-15 22:21:53,339][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.251646488904953, 'train/grad_norm_pre_clip': 8.461447715759277, 'timing/data': 0.0005106910248287022, 'timing/model': 1.539507259032689, 'learning_rate/qwen_vl_interface': 7.838737916842538e-06, 'learning_rate/action_model': 7.838737916842538e-05, 'epoch': 0.71}) +[2026-06-15 22:21:59,164][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.09712539613246918, 'train/grad_norm_pre_clip': 8.568002700805664, 'timing/data': 0.0005138659616932273, 'timing/model': 1.4640124770230614, 'learning_rate/qwen_vl_interface': 7.836183988802596e-06, 'learning_rate/action_model': 7.836183988802595e-05, 'epoch': 0.71}) +[2026-06-15 22:22:05,043][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.1706399917602539, 'train/grad_norm_pre_clip': 5.207048416137695, 'timing/data': 0.0005062749842181802, 'timing/model': 1.5172524800291285, 'learning_rate/qwen_vl_interface': 7.833628997680107e-06, 'learning_rate/action_model': 7.833628997680107e-05, 'epoch': 0.71}) +[2026-06-15 22:22:10,990][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.16042092442512512, 'train/grad_norm_pre_clip': 17.237525939941406, 'timing/data': 0.00051882304251194, 'timing/model': 1.6888239189866, 'learning_rate/qwen_vl_interface': 7.831072944525329e-06, 'learning_rate/action_model': 7.831072944525328e-05, 'epoch': 0.71}) +[2026-06-15 22:22:16,665][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.14776352047920227, 'train/grad_norm_pre_clip': 4.0725555419921875, 'timing/data': 0.0005163699970580637, 'timing/model': 1.466269071039278, 'learning_rate/qwen_vl_interface': 7.828515830388956e-06, 'learning_rate/action_model': 7.828515830388956e-05, 'epoch': 0.71}) +[2026-06-15 22:22:22,597][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.07197026163339615, 'train/grad_norm_pre_clip': 3.8072547912597656, 'timing/data': 0.0005119629786349833, 'timing/model': 1.4962625459884293, 'learning_rate/qwen_vl_interface': 7.825957656322129e-06, 'learning_rate/action_model': 7.825957656322128e-05, 'epoch': 0.71}) +[2026-06-15 22:22:28,564][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.1864033043384552, 'train/grad_norm_pre_clip': 4.514517307281494, 'timing/data': 0.000509629026055336, 'timing/model': 1.4495263670105487, 'learning_rate/qwen_vl_interface': 7.823398423376406e-06, 'learning_rate/action_model': 7.823398423376406e-05, 'epoch': 0.71}) +[2026-06-15 22:22:34,551][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.24766628444194794, 'train/grad_norm_pre_clip': 9.536755561828613, 'timing/data': 0.0005315010203048587, 'timing/model': 1.5207738239550963, 'learning_rate/qwen_vl_interface': 7.820838132603801e-06, 'learning_rate/action_model': 7.8208381326038e-05, 'epoch': 0.71}) +[2026-06-15 22:22:40,487][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.20499606430530548, 'train/grad_norm_pre_clip': 8.472954750061035, 'timing/data': 0.0005582509911619127, 'timing/model': 1.687857725948561, 'learning_rate/qwen_vl_interface': 7.818276785056748e-06, 'learning_rate/action_model': 7.818276785056749e-05, 'epoch': 0.71}) +[2026-06-15 22:22:46,415][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.0812799409031868, 'train/grad_norm_pre_clip': 9.418535232543945, 'timing/data': 0.0005063950084149837, 'timing/model': 1.7160739380051382, 'learning_rate/qwen_vl_interface': 7.815714381788123e-06, 'learning_rate/action_model': 7.815714381788122e-05, 'epoch': 0.71}) +[2026-06-15 22:22:52,165][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.16129297018051147, 'train/grad_norm_pre_clip': 4.396832466125488, 'timing/data': 0.0006022260058671236, 'timing/model': 1.4651403090101667, 'learning_rate/qwen_vl_interface': 7.813150923851227e-06, 'learning_rate/action_model': 7.813150923851227e-05, 'epoch': 0.71}) +[2026-06-15 22:22:58,163][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.2186633199453354, 'train/grad_norm_pre_clip': 7.0479350090026855, 'timing/data': 0.000579461979214102, 'timing/model': 1.4633923629880883, 'learning_rate/qwen_vl_interface': 7.810586412299807e-06, 'learning_rate/action_model': 7.810586412299806e-05, 'epoch': 0.72}) +[2026-06-15 22:23:04,247][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.03462892770767212, 'train/grad_norm_pre_clip': 1.5128042697906494, 'timing/data': 0.0005631580133922398, 'timing/model': 1.569876613037195, 'learning_rate/qwen_vl_interface': 7.808020848188034e-06, 'learning_rate/action_model': 7.808020848188034e-05, 'epoch': 0.72}) +[2026-06-15 22:23:10,244][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.1483413577079773, 'train/grad_norm_pre_clip': 7.227765083312988, 'timing/data': 0.0005202050087973475, 'timing/model': 1.5149456119979732, 'learning_rate/qwen_vl_interface': 7.805454232570517e-06, 'learning_rate/action_model': 7.805454232570517e-05, 'epoch': 0.72}) +[2026-06-15 22:23:16,270][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.1899951994419098, 'train/grad_norm_pre_clip': 6.718771934509277, 'timing/data': 0.000542767986189574, 'timing/model': 1.7709291399805807, 'learning_rate/qwen_vl_interface': 7.80288656650229e-06, 'learning_rate/action_model': 7.80288656650229e-05, 'epoch': 0.72}) +[2026-06-15 22:23:22,058][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.19857460260391235, 'train/grad_norm_pre_clip': 13.180035591125488, 'timing/data': 0.0005143469898030162, 'timing/model': 1.5627391210291535, 'learning_rate/qwen_vl_interface': 7.800317851038829e-06, 'learning_rate/action_model': 7.800317851038828e-05, 'epoch': 0.72}) +[2026-06-15 22:23:27,761][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.11209127306938171, 'train/grad_norm_pre_clip': 3.189866542816162, 'timing/data': 0.000517920998390764, 'timing/model': 1.462352266011294, 'learning_rate/qwen_vl_interface': 7.797748087236033e-06, 'learning_rate/action_model': 7.797748087236033e-05, 'epoch': 0.72}) +[2026-06-15 22:23:33,769][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.25726157426834106, 'train/grad_norm_pre_clip': 8.10109806060791, 'timing/data': 0.0005182320019230247, 'timing/model': 1.4573490120237693, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.72}) +[2026-06-15 22:23:39,764][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.10918018966913223, 'train/grad_norm_pre_clip': 3.621285915374756, 'timing/data': 0.0005143970483914018, 'timing/model': 1.4843541629961692, 'learning_rate/qwen_vl_interface': 7.792605418838196e-06, 'learning_rate/action_model': 7.792605418838197e-05, 'epoch': 0.72}) +[2026-06-15 22:23:45,740][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.10735374689102173, 'train/grad_norm_pre_clip': 4.577950477600098, 'timing/data': 0.000509879959281534, 'timing/model': 1.5098984150099568, 'learning_rate/qwen_vl_interface': 7.790032516357116e-06, 'learning_rate/action_model': 7.790032516357117e-05, 'epoch': 0.72}) +[2026-06-15 22:23:51,764][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.22820405662059784, 'train/grad_norm_pre_clip': 12.753594398498535, 'timing/data': 0.0005145869799889624, 'timing/model': 1.7661743190255947, 'learning_rate/qwen_vl_interface': 7.787458569764613e-06, 'learning_rate/action_model': 7.787458569764612e-05, 'epoch': 0.72}) +[2026-06-15 22:23:57,464][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.13674324750900269, 'train/grad_norm_pre_clip': 7.008415699005127, 'timing/data': 0.0005167089984752238, 'timing/model': 1.4650772340246476, 'learning_rate/qwen_vl_interface': 7.784883580118738e-06, 'learning_rate/action_model': 7.784883580118738e-05, 'epoch': 0.72}) +[2026-06-15 22:24:03,464][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.08482465893030167, 'train/grad_norm_pre_clip': 4.181641578674316, 'timing/data': 0.0005113619845360518, 'timing/model': 1.4657115310546942, 'learning_rate/qwen_vl_interface': 7.782307548477975e-06, 'learning_rate/action_model': 7.782307548477975e-05, 'epoch': 0.72}) +[2026-06-15 22:24:09,477][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.15481796860694885, 'train/grad_norm_pre_clip': 4.529543399810791, 'timing/data': 0.0005308809923008084, 'timing/model': 1.4718079220037907, 'learning_rate/qwen_vl_interface': 7.779730475901233e-06, 'learning_rate/action_model': 7.779730475901232e-05, 'epoch': 0.72}) +[2026-06-15 22:24:15,363][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.1635977029800415, 'train/grad_norm_pre_clip': 5.9816179275512695, 'timing/data': 0.0005203249747864902, 'timing/model': 1.4648387149791233, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 0.72}) +[2026-06-15 22:24:21,343][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.19110463559627533, 'train/grad_norm_pre_clip': 8.569199562072754, 'timing/data': 0.00048572401283308864, 'timing/model': 1.5135381430154666, 'learning_rate/qwen_vl_interface': 7.774573212177587e-06, 'learning_rate/action_model': 7.774573212177588e-05, 'epoch': 0.72}) +[2026-06-15 22:24:27,470][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.13041585683822632, 'train/grad_norm_pre_clip': 7.261539936065674, 'timing/data': 0.0004896899918094277, 'timing/model': 1.773235802014824, 'learning_rate/qwen_vl_interface': 7.77199302315064e-06, 'learning_rate/action_model': 7.77199302315064e-05, 'epoch': 0.72}) +[2026-06-15 22:24:33,364][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.25162792205810547, 'train/grad_norm_pre_clip': 6.473409175872803, 'timing/data': 0.0005153879756107926, 'timing/model': 1.46051919501042, 'learning_rate/qwen_vl_interface': 7.769411797427622e-06, 'learning_rate/action_model': 7.769411797427623e-05, 'epoch': 0.72}) +[2026-06-15 22:24:39,369][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.1286046802997589, 'train/grad_norm_pre_clip': 11.836627006530762, 'timing/data': 0.000510671001393348, 'timing/model': 1.4745517970295623, 'learning_rate/qwen_vl_interface': 7.766829536069584e-06, 'learning_rate/action_model': 7.766829536069584e-05, 'epoch': 0.72}) +[2026-06-15 22:24:45,278][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.08419915288686752, 'train/grad_norm_pre_clip': 3.382962703704834, 'timing/data': 0.0005799729842692614, 'timing/model': 1.5786186760524288, 'learning_rate/qwen_vl_interface': 7.764246240137991e-06, 'learning_rate/action_model': 7.764246240137991e-05, 'epoch': 0.72}) +[2026-06-15 22:24:51,168][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.1415978968143463, 'train/grad_norm_pre_clip': 7.1814446449279785, 'timing/data': 0.0006753439665772021, 'timing/model': 1.450734484998975, 'learning_rate/qwen_vl_interface': 7.761661910694741e-06, 'learning_rate/action_model': 7.761661910694741e-05, 'epoch': 0.72}) +[2026-06-15 22:24:57,163][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.05431819334626198, 'train/grad_norm_pre_clip': 2.0632970333099365, 'timing/data': 0.0005931620253250003, 'timing/model': 1.4388207269948907, 'learning_rate/qwen_vl_interface': 7.759076548802153e-06, 'learning_rate/action_model': 7.759076548802152e-05, 'epoch': 0.72}) +[2026-06-15 22:25:03,142][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.114221952855587, 'train/grad_norm_pre_clip': 6.991847991943359, 'timing/data': 0.0005120530258864164, 'timing/model': 1.7373394159949385, 'learning_rate/qwen_vl_interface': 7.75649015552297e-06, 'learning_rate/action_model': 7.75649015552297e-05, 'epoch': 0.72}) +[2026-06-15 22:25:08,863][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.3071928322315216, 'train/grad_norm_pre_clip': 14.25333023071289, 'timing/data': 0.0005254030111245811, 'timing/model': 1.4639013809501193, 'learning_rate/qwen_vl_interface': 7.753902731920364e-06, 'learning_rate/action_model': 7.753902731920364e-05, 'epoch': 0.72}) +[2026-06-15 22:25:14,864][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.12738507986068726, 'train/grad_norm_pre_clip': 3.966766357421875, 'timing/data': 0.0005016580107621849, 'timing/model': 1.4636715759988874, 'learning_rate/qwen_vl_interface': 7.751314279057925e-06, 'learning_rate/action_model': 7.751314279057923e-05, 'epoch': 0.73}) +[2026-06-15 22:25:20,865][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.1395905315876007, 'train/grad_norm_pre_clip': 5.002435684204102, 'timing/data': 0.0005061649717390537, 'timing/model': 1.471894794027321, 'learning_rate/qwen_vl_interface': 7.74872479799967e-06, 'learning_rate/action_model': 7.74872479799967e-05, 'epoch': 0.73}) +[2026-06-15 22:25:26,847][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.08916543424129486, 'train/grad_norm_pre_clip': 3.840315818786621, 'timing/data': 0.0008736349991522729, 'timing/model': 1.510544481978286, 'learning_rate/qwen_vl_interface': 7.746134289810034e-06, 'learning_rate/action_model': 7.746134289810033e-05, 'epoch': 0.73}) +[2026-06-15 22:25:32,760][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.021288612857460976, 'train/grad_norm_pre_clip': 1.419521689414978, 'timing/data': 0.000511862977873534, 'timing/model': 1.5089556089951657, 'learning_rate/qwen_vl_interface': 7.74354275555388e-06, 'learning_rate/action_model': 7.74354275555388e-05, 'epoch': 0.73}) +[2026-06-15 22:25:38,740][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.13432002067565918, 'train/grad_norm_pre_clip': 5.301513195037842, 'timing/data': 0.00048559397691860795, 'timing/model': 1.6619831490097567, 'learning_rate/qwen_vl_interface': 7.740950196296488e-06, 'learning_rate/action_model': 7.740950196296488e-05, 'epoch': 0.73}) +[2026-06-15 22:25:44,563][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.11661971360445023, 'train/grad_norm_pre_clip': 6.1010260581970215, 'timing/data': 0.0005004859995096922, 'timing/model': 1.466055179014802, 'learning_rate/qwen_vl_interface': 7.738356613103562e-06, 'learning_rate/action_model': 7.73835661310356e-05, 'epoch': 0.73}) +[2026-06-15 22:25:50,563][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.07256793975830078, 'train/grad_norm_pre_clip': 7.413430690765381, 'timing/data': 0.0005181719898246229, 'timing/model': 1.4643427039845847, 'learning_rate/qwen_vl_interface': 7.735762007041226e-06, 'learning_rate/action_model': 7.735762007041226e-05, 'epoch': 0.73}) +[2026-06-15 22:25:56,562][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.18744127452373505, 'train/grad_norm_pre_clip': 5.600135326385498, 'timing/data': 0.000687330961227417, 'timing/model': 1.4833672979730181, 'learning_rate/qwen_vl_interface': 7.733166379176023e-06, 'learning_rate/action_model': 7.733166379176023e-05, 'epoch': 0.73}) +[2026-06-15 22:26:02,640][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.1697990596294403, 'train/grad_norm_pre_clip': 8.570124626159668, 'timing/data': 0.0007368940277956426, 'timing/model': 1.5954594230279326, 'learning_rate/qwen_vl_interface': 7.730569730574921e-06, 'learning_rate/action_model': 7.73056973057492e-05, 'epoch': 0.73}) +[2026-06-15 22:26:08,717][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.06969630718231201, 'train/grad_norm_pre_clip': 2.625422239303589, 'timing/data': 0.0005841689999215305, 'timing/model': 1.7193782829563133, 'learning_rate/qwen_vl_interface': 7.727972062305299e-06, 'learning_rate/action_model': 7.727972062305299e-05, 'epoch': 0.73}) +[2026-06-15 22:26:14,462][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.07858194410800934, 'train/grad_norm_pre_clip': 5.044356346130371, 'timing/data': 0.0004890490090474486, 'timing/model': 1.5679763680091128, 'learning_rate/qwen_vl_interface': 7.725373375434965e-06, 'learning_rate/action_model': 7.725373375434965e-05, 'epoch': 0.73}) +[2026-06-15 22:26:20,160][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.16733458638191223, 'train/grad_norm_pre_clip': 6.240115165710449, 'timing/data': 0.0005066759767942131, 'timing/model': 1.4674688570084982, 'learning_rate/qwen_vl_interface': 7.722773671032141e-06, 'learning_rate/action_model': 7.722773671032142e-05, 'epoch': 0.73}) +[2026-06-15 22:26:26,159][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.1855451464653015, 'train/grad_norm_pre_clip': 8.383015632629395, 'timing/data': 0.0004941360093653202, 'timing/model': 1.4657635689945891, 'learning_rate/qwen_vl_interface': 7.720172950165466e-06, 'learning_rate/action_model': 7.720172950165465e-05, 'epoch': 0.73}) +[2026-06-15 22:26:32,159][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.12443338334560394, 'train/grad_norm_pre_clip': 7.765529155731201, 'timing/data': 0.0005013979971408844, 'timing/model': 1.4677854250185192, 'learning_rate/qwen_vl_interface': 7.717571213903996e-06, 'learning_rate/action_model': 7.717571213903996e-05, 'epoch': 0.73}) +[2026-06-15 22:26:38,167][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.22398912906646729, 'train/grad_norm_pre_clip': 12.168200492858887, 'timing/data': 0.0005765380337834358, 'timing/model': 1.5004924859968014, 'learning_rate/qwen_vl_interface': 7.714968463317207e-06, 'learning_rate/action_model': 7.714968463317207e-05, 'epoch': 0.73}) +[2026-06-15 22:35:53,252][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.08721166104078293, 'train/grad_norm_pre_clip': 5.533425807952881, 'eval/loss': 0.07989400625228882, 'eval/flap_precision': 0.8816568047337278, 'eval/flap_recall': 0.8957915831663327, 'eval/flap_f1': 0.8886679920477136, 'eval/flap_rate_model': 0.14083333333333334, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.8886679920477136, 'timing/data': 0.0005787420086562634, 'timing/model': 1.7049975729896687, 'learning_rate/qwen_vl_interface': 7.712364699474996e-06, 'learning_rate/action_model': 7.712364699474996e-05, 'epoch': 0.73}) +[2026-06-15 22:35:53,253][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1700_state +[2026-06-15 22:35:53,257][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 22:35:58,756][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1700_state/model.safetensors +[2026-06-15 22:36:12,114][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1700_state/optimizer.bin +[2026-06-15 22:36:12,115][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1700_state/sampler.bin +[2026-06-15 22:36:12,115][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1700_state/sampler_1.bin +[2026-06-15 22:36:12,116][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1700_state/random_states_0.pkl +[2026-06-15 22:36:12,228][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1700 +[2026-06-15 22:36:12,229][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 22:36:12,239][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 22:36:18,154][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.12064565718173981, 'train/grad_norm_pre_clip': 6.34467077255249, 'timing/data': 0.0005279770120978355, 'timing/model': 1.4692568159662187, 'learning_rate/qwen_vl_interface': 7.709759923447667e-06, 'learning_rate/action_model': 7.709759923447665e-05, 'epoch': 0.73}) +[2026-06-15 22:36:24,102][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.09481452405452728, 'train/grad_norm_pre_clip': 9.19659423828125, 'timing/data': 0.0004836319712921977, 'timing/model': 1.7250245459727012, 'learning_rate/qwen_vl_interface': 7.707154136305945e-06, 'learning_rate/action_model': 7.707154136305945e-05, 'epoch': 0.73}) +[2026-06-15 22:36:29,748][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.2721743583679199, 'train/grad_norm_pre_clip': 7.444473743438721, 'timing/data': 0.0005113219958730042, 'timing/model': 1.4687756330240518, 'learning_rate/qwen_vl_interface': 7.704547339120974e-06, 'learning_rate/action_model': 7.704547339120974e-05, 'epoch': 0.73}) +[2026-06-15 22:36:35,746][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.3206387460231781, 'train/grad_norm_pre_clip': 16.178287506103516, 'timing/data': 0.0005228790105320513, 'timing/model': 1.4615186239825562, 'learning_rate/qwen_vl_interface': 7.70193953296431e-06, 'learning_rate/action_model': 7.70193953296431e-05, 'epoch': 0.73}) +[2026-06-15 22:36:41,843][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.04950151965022087, 'train/grad_norm_pre_clip': 4.376704216003418, 'timing/data': 0.0005282569909468293, 'timing/model': 1.5755550900357775, 'learning_rate/qwen_vl_interface': 7.69933071890792e-06, 'learning_rate/action_model': 7.699330718907919e-05, 'epoch': 0.73}) +[2026-06-15 22:36:47,845][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.13426831364631653, 'train/grad_norm_pre_clip': 8.368971824645996, 'timing/data': 0.0004883679794147611, 'timing/model': 1.4461217069765553, 'learning_rate/qwen_vl_interface': 7.696720898024188e-06, 'learning_rate/action_model': 7.69672089802419e-05, 'epoch': 0.73}) +[2026-06-15 22:36:53,792][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.11217963695526123, 'train/grad_norm_pre_clip': 6.355627536773682, 'timing/data': 0.0005284170038066804, 'timing/model': 1.4644342350075021, 'learning_rate/qwen_vl_interface': 7.69411007138592e-06, 'learning_rate/action_model': 7.694110071385919e-05, 'epoch': 0.73}) +[2026-06-15 22:36:59,699][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.29039502143859863, 'train/grad_norm_pre_clip': 21.242929458618164, 'timing/data': 0.0005799029604531825, 'timing/model': 1.7016856609843671, 'learning_rate/qwen_vl_interface': 7.691498240066322e-06, 'learning_rate/action_model': 7.691498240066322e-05, 'epoch': 0.73}) +[2026-06-15 22:37:05,356][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.12207051366567612, 'train/grad_norm_pre_clip': 5.627114295959473, 'timing/data': 0.0007339799776673317, 'timing/model': 1.4662722540087998, 'learning_rate/qwen_vl_interface': 7.688885405139022e-06, 'learning_rate/action_model': 7.688885405139021e-05, 'epoch': 0.74}) +[2026-06-15 22:37:11,356][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.18258073925971985, 'train/grad_norm_pre_clip': 7.473114013671875, 'timing/data': 0.0005647409707307816, 'timing/model': 1.4626226020045578, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.74}) +[2026-06-15 22:37:17,646][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.016904441639780998, 'train/grad_norm_pre_clip': 1.134649395942688, 'timing/data': 0.0005251119728200138, 'timing/model': 1.4708432329935022, 'learning_rate/qwen_vl_interface': 7.683656728757873e-06, 'learning_rate/action_model': 7.683656728757872e-05, 'epoch': 0.74}) +[2026-06-15 22:37:23,556][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.057824861258268356, 'train/grad_norm_pre_clip': 5.5208048820495605, 'timing/data': 0.0005103810108266771, 'timing/model': 1.4346205909969285, 'learning_rate/qwen_vl_interface': 7.681040889453337e-06, 'learning_rate/action_model': 7.681040889453336e-05, 'epoch': 0.74}) +[2026-06-15 22:37:29,782][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.04886560142040253, 'train/grad_norm_pre_clip': 5.629032135009766, 'timing/data': 0.0005016479990445077, 'timing/model': 1.7172407399630174, 'learning_rate/qwen_vl_interface': 7.678424050839722e-06, 'learning_rate/action_model': 7.67842405083972e-05, 'epoch': 0.74}) +[2026-06-15 22:37:35,682][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.20222154259681702, 'train/grad_norm_pre_clip': 12.145526885986328, 'timing/data': 0.0005016179638914764, 'timing/model': 1.690818406001199, 'learning_rate/qwen_vl_interface': 7.675806213992706e-06, 'learning_rate/action_model': 7.675806213992707e-05, 'epoch': 0.74}) +[2026-06-15 22:37:41,358][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.18624374270439148, 'train/grad_norm_pre_clip': 19.11646270751953, 'timing/data': 0.0005154179525561631, 'timing/model': 1.4640630790381692, 'learning_rate/qwen_vl_interface': 7.673187379988388e-06, 'learning_rate/action_model': 7.673187379988388e-05, 'epoch': 0.74}) +[2026-06-15 22:37:47,359][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.12515679001808167, 'train/grad_norm_pre_clip': 8.29446029663086, 'timing/data': 0.0005004870472475886, 'timing/model': 1.464914028008934, 'learning_rate/qwen_vl_interface': 7.67056754990327e-06, 'learning_rate/action_model': 7.670567549903269e-05, 'epoch': 0.74}) +[2026-06-15 22:37:53,359][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.14929711818695068, 'train/grad_norm_pre_clip': 12.968935012817383, 'timing/data': 0.000518032000400126, 'timing/model': 1.4592737180064432, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.74}) +[2026-06-15 22:37:59,344][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.2964462339878082, 'train/grad_norm_pre_clip': 7.637232780456543, 'timing/data': 0.0005111219943501055, 'timing/model': 1.5115170640056022, 'learning_rate/qwen_vl_interface': 7.665324905798694e-06, 'learning_rate/action_model': 7.665324905798695e-05, 'epoch': 0.74}) +[2026-06-15 22:38:05,295][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.09656813740730286, 'train/grad_norm_pre_clip': 6.8800153732299805, 'timing/data': 0.0005171999800950289, 'timing/model': 1.6992647779989056, 'learning_rate/qwen_vl_interface': 7.66270209393429e-06, 'learning_rate/action_model': 7.66270209393429e-05, 'epoch': 0.74}) +[2026-06-15 22:38:11,250][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.20347630977630615, 'train/grad_norm_pre_clip': 10.622408866882324, 'timing/data': 0.0005095499800518155, 'timing/model': 1.505480679043103, 'learning_rate/qwen_vl_interface': 7.66007829029919e-06, 'learning_rate/action_model': 7.660078290299191e-05, 'epoch': 0.74}) +[2026-06-15 22:38:17,362][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.11418354511260986, 'train/grad_norm_pre_clip': 8.261406898498535, 'timing/data': 0.0005057940143160522, 'timing/model': 1.4803767949924804, 'learning_rate/qwen_vl_interface': 7.657453495971942e-06, 'learning_rate/action_model': 7.657453495971943e-05, 'epoch': 0.74}) +[2026-06-15 22:38:23,261][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.10369248688220978, 'train/grad_norm_pre_clip': 7.257206916809082, 'timing/data': 0.0005001850076951087, 'timing/model': 1.4592507120105438, 'learning_rate/qwen_vl_interface': 7.6548277120315e-06, 'learning_rate/action_model': 7.6548277120315e-05, 'epoch': 0.74}) +[2026-06-15 22:38:29,241][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.1378060281276703, 'train/grad_norm_pre_clip': 7.015644073486328, 'timing/data': 0.0005129239871166646, 'timing/model': 1.5092178389895707, 'learning_rate/qwen_vl_interface': 7.652200939557225e-06, 'learning_rate/action_model': 7.652200939557225e-05, 'epoch': 0.74}) +[2026-06-15 22:38:35,261][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.13703452050685883, 'train/grad_norm_pre_clip': 7.539955139160156, 'timing/data': 0.0005121940048411489, 'timing/model': 1.439576675009448, 'learning_rate/qwen_vl_interface': 7.649573179628881e-06, 'learning_rate/action_model': 7.649573179628881e-05, 'epoch': 0.74}) +[2026-06-15 22:38:41,252][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.055422671139240265, 'train/grad_norm_pre_clip': 5.87097692489624, 'timing/data': 0.0005085780285298824, 'timing/model': 1.6963228639797308, 'learning_rate/qwen_vl_interface': 7.646944433326645e-06, 'learning_rate/action_model': 7.646944433326643e-05, 'epoch': 0.74}) +[2026-06-15 22:38:46,962][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.050589289516210556, 'train/grad_norm_pre_clip': 4.56116247177124, 'timing/data': 0.000491462997160852, 'timing/model': 1.441546238027513, 'learning_rate/qwen_vl_interface': 7.64431470173109e-06, 'learning_rate/action_model': 7.64431470173109e-05, 'epoch': 0.74}) +[2026-06-15 22:38:53,252][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.037536896765232086, 'train/grad_norm_pre_clip': 3.5258731842041016, 'timing/data': 0.0005099900299683213, 'timing/model': 1.468576310959179, 'learning_rate/qwen_vl_interface': 7.641683985923203e-06, 'learning_rate/action_model': 7.641683985923203e-05, 'epoch': 0.74}) +[2026-06-15 22:38:59,259][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.07460954785346985, 'train/grad_norm_pre_clip': 2.845863103866577, 'timing/data': 0.0004988830187357962, 'timing/model': 1.4875725849997252, 'learning_rate/qwen_vl_interface': 7.63905228698437e-06, 'learning_rate/action_model': 7.63905228698437e-05, 'epoch': 0.74}) +[2026-06-15 22:39:05,244][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.1766156107187271, 'train/grad_norm_pre_clip': 9.758810997009277, 'timing/data': 0.000563770008739084, 'timing/model': 1.5168992269900627, 'learning_rate/qwen_vl_interface': 7.636419605996385e-06, 'learning_rate/action_model': 7.636419605996384e-05, 'epoch': 0.74}) +[2026-06-15 22:39:11,240][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.10823486745357513, 'train/grad_norm_pre_clip': 5.606267929077148, 'timing/data': 0.0005107210017740726, 'timing/model': 1.50936173601076, 'learning_rate/qwen_vl_interface': 7.633785944041444e-06, 'learning_rate/action_model': 7.633785944041444e-05, 'epoch': 0.74}) +[2026-06-15 22:39:17,332][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.041529908776283264, 'train/grad_norm_pre_clip': 3.032320737838745, 'timing/data': 0.0005058140377514064, 'timing/model': 1.7351084199617617, 'learning_rate/qwen_vl_interface': 7.631151302202145e-06, 'learning_rate/action_model': 7.631151302202144e-05, 'epoch': 0.74}) +[2026-06-15 22:39:23,060][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.17646248638629913, 'train/grad_norm_pre_clip': 10.092509269714355, 'timing/data': 0.0005266750231385231, 'timing/model': 1.4662503600120544, 'learning_rate/qwen_vl_interface': 7.628515681561488e-06, 'learning_rate/action_model': 7.628515681561487e-05, 'epoch': 0.75}) +[2026-06-15 22:39:29,059][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.09397610276937485, 'train/grad_norm_pre_clip': 5.728977203369141, 'timing/data': 0.0005125440075062215, 'timing/model': 1.4660878980066627, 'learning_rate/qwen_vl_interface': 7.62587908320288e-06, 'learning_rate/action_model': 7.62587908320288e-05, 'epoch': 0.75}) +[2026-06-15 22:39:35,064][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.2498794049024582, 'train/grad_norm_pre_clip': 13.310644149780273, 'timing/data': 0.0005088190082460642, 'timing/model': 1.4800989580107853, 'learning_rate/qwen_vl_interface': 7.6232415082101255e-06, 'learning_rate/action_model': 7.623241508210124e-05, 'epoch': 0.75}) +[2026-06-15 22:39:41,046][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.06337611377239227, 'train/grad_norm_pre_clip': 2.4179532527923584, 'timing/data': 0.000516579020768404, 'timing/model': 1.5143022639676929, 'learning_rate/qwen_vl_interface': 7.620602957667432e-06, 'learning_rate/action_model': 7.620602957667432e-05, 'epoch': 0.75}) +[2026-06-15 22:39:46,880][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.24217817187309265, 'train/grad_norm_pre_clip': 12.854207992553711, 'timing/data': 0.0005161589942872524, 'timing/model': 1.4745754480245523, 'learning_rate/qwen_vl_interface': 7.617963432659408e-06, 'learning_rate/action_model': 7.617963432659408e-05, 'epoch': 0.75}) +[2026-06-15 22:39:52,822][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.10586768388748169, 'train/grad_norm_pre_clip': 8.229632377624512, 'timing/data': 0.0005136050167493522, 'timing/model': 1.7253164199646562, 'learning_rate/qwen_vl_interface': 7.615322934271063e-06, 'learning_rate/action_model': 7.615322934271064e-05, 'epoch': 0.75}) +[2026-06-15 22:39:58,562][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.20867034792900085, 'train/grad_norm_pre_clip': 13.702735900878906, 'timing/data': 0.00048404198605567217, 'timing/model': 1.4589082169695757, 'learning_rate/qwen_vl_interface': 7.6126814635878055e-06, 'learning_rate/action_model': 7.612681463587805e-05, 'epoch': 0.75}) +[2026-06-15 22:40:04,636][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.18299700319766998, 'train/grad_norm_pre_clip': 5.2981743812561035, 'timing/data': 0.0005089890328235924, 'timing/model': 1.5335976229980588, 'learning_rate/qwen_vl_interface': 7.6100390216954445e-06, 'learning_rate/action_model': 7.610039021695445e-05, 'epoch': 0.75}) +[2026-06-15 22:40:10,806][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.01469492819160223, 'train/grad_norm_pre_clip': 1.352605938911438, 'timing/data': 0.00048552401131018996, 'timing/model': 1.4822023320011795, 'learning_rate/qwen_vl_interface': 7.607395609680189e-06, 'learning_rate/action_model': 7.60739560968019e-05, 'epoch': 0.75}) +[2026-06-15 22:40:16,860][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.02007683925330639, 'train/grad_norm_pre_clip': 1.4853090047836304, 'timing/data': 0.0007443149806931615, 'timing/model': 1.5072117010131478, 'learning_rate/qwen_vl_interface': 7.604751228628644e-06, 'learning_rate/action_model': 7.604751228628644e-05, 'epoch': 0.75}) +[2026-06-15 22:40:22,855][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.44777053594589233, 'train/grad_norm_pre_clip': 10.081178665161133, 'timing/data': 0.0005000760429538786, 'timing/model': 1.5115463170222938, 'learning_rate/qwen_vl_interface': 7.602105879627814e-06, 'learning_rate/action_model': 7.602105879627814e-05, 'epoch': 0.75}) +[2026-06-15 22:40:28,894][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.16074547171592712, 'train/grad_norm_pre_clip': 7.134818077087402, 'timing/data': 0.0004950889851897955, 'timing/model': 1.6989005859941244, 'learning_rate/qwen_vl_interface': 7.599459563765107e-06, 'learning_rate/action_model': 7.599459563765106e-05, 'epoch': 0.75}) +[2026-06-15 22:40:34,558][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.25948092341423035, 'train/grad_norm_pre_clip': 19.737245559692383, 'timing/data': 0.0005263240309432149, 'timing/model': 1.4629231430008076, 'learning_rate/qwen_vl_interface': 7.596812282128319e-06, 'learning_rate/action_model': 7.596812282128319e-05, 'epoch': 0.75}) +[2026-06-15 22:40:40,562][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.04108527675271034, 'train/grad_norm_pre_clip': 4.388396263122559, 'timing/data': 0.000514326966367662, 'timing/model': 1.4683025740087032, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 0.75}) +[2026-06-15 22:40:46,462][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.135140061378479, 'train/grad_norm_pre_clip': 7.25820779800415, 'timing/data': 0.0005183120374567807, 'timing/model': 1.4591959850513376, 'learning_rate/qwen_vl_interface': 7.59151482588569e-06, 'learning_rate/action_model': 7.59151482588569e-05, 'epoch': 0.75}) +[2026-06-15 22:40:52,565][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.21015037596225739, 'train/grad_norm_pre_clip': 13.207464218139648, 'timing/data': 0.0005159690044820309, 'timing/model': 1.4797558399732225, 'learning_rate/qwen_vl_interface': 7.588864653457433e-06, 'learning_rate/action_model': 7.588864653457432e-05, 'epoch': 0.75}) +[2026-06-15 22:40:58,598][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.13597752153873444, 'train/grad_norm_pre_clip': 9.044843673706055, 'timing/data': 0.0005304710357449949, 'timing/model': 1.7026746490155347, 'learning_rate/qwen_vl_interface': 7.5862135196102634e-06, 'learning_rate/action_model': 7.586213519610263e-05, 'epoch': 0.75}) +[2026-06-15 22:41:04,632][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.21587330102920532, 'train/grad_norm_pre_clip': 10.804439544677734, 'timing/data': 0.000516159983817488, 'timing/model': 1.6320934030227363, 'learning_rate/qwen_vl_interface': 7.583561425433962e-06, 'learning_rate/action_model': 7.583561425433962e-05, 'epoch': 0.75}) +[2026-06-15 22:41:10,367][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.15060070157051086, 'train/grad_norm_pre_clip': 9.957735061645508, 'timing/data': 0.0005116329994052649, 'timing/model': 1.4655430759885348, 'learning_rate/qwen_vl_interface': 7.580908372018706e-06, 'learning_rate/action_model': 7.580908372018705e-05, 'epoch': 0.75}) +[2026-06-15 22:41:16,365][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.26148149371147156, 'train/grad_norm_pre_clip': 12.684917449951172, 'timing/data': 0.0005835889605805278, 'timing/model': 1.4639652539626695, 'learning_rate/qwen_vl_interface': 7.5782543604550655e-06, 'learning_rate/action_model': 7.578254360455064e-05, 'epoch': 0.75}) +[2026-06-15 22:41:22,408][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.25056618452072144, 'train/grad_norm_pre_clip': 6.033421039581299, 'timing/data': 0.000518401968292892, 'timing/model': 1.5234391349949874, 'learning_rate/qwen_vl_interface': 7.575599391834e-06, 'learning_rate/action_model': 7.575599391834e-05, 'epoch': 0.75}) +[2026-06-15 22:41:28,349][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.13480044901371002, 'train/grad_norm_pre_clip': 5.019632339477539, 'timing/data': 0.0005202850443311036, 'timing/model': 1.5106753260479309, 'learning_rate/qwen_vl_interface': 7.572943467246871e-06, 'learning_rate/action_model': 7.57294346724687e-05, 'epoch': 0.75}) +[2026-06-15 22:41:34,441][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.10159093886613846, 'train/grad_norm_pre_clip': 6.877683162689209, 'timing/data': 0.0005186229827813804, 'timing/model': 1.7609484050190076, 'learning_rate/qwen_vl_interface': 7.570286587785428e-06, 'learning_rate/action_model': 7.570286587785428e-05, 'epoch': 0.75}) +[2026-06-15 22:41:40,367][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.031468555331230164, 'train/grad_norm_pre_clip': 1.8859095573425293, 'timing/data': 0.0005122340517118573, 'timing/model': 1.4908907279605046, 'learning_rate/qwen_vl_interface': 7.567628754541815e-06, 'learning_rate/action_model': 7.567628754541814e-05, 'epoch': 0.76}) +[2026-06-15 22:41:46,426][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.10583463311195374, 'train/grad_norm_pre_clip': 7.18409538269043, 'timing/data': 0.0005178920109756291, 'timing/model': 1.5281205649953336, 'learning_rate/qwen_vl_interface': 7.564969968608565e-06, 'learning_rate/action_model': 7.564969968608565e-05, 'epoch': 0.76}) +[2026-06-15 22:41:52,264][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.319265216588974, 'train/grad_norm_pre_clip': 11.785111427307129, 'timing/data': 0.000580844993237406, 'timing/model': 1.4497873469954357, 'learning_rate/qwen_vl_interface': 7.562310231078605e-06, 'learning_rate/action_model': 7.562310231078604e-05, 'epoch': 0.76}) +[2026-06-15 22:41:58,265][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.02582698129117489, 'train/grad_norm_pre_clip': 2.3575878143310547, 'timing/data': 0.0006826149765402079, 'timing/model': 1.4461984960362315, 'learning_rate/qwen_vl_interface': 7.559649543045252e-06, 'learning_rate/action_model': 7.559649543045252e-05, 'epoch': 0.76}) +[2026-06-15 22:42:04,362][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.11874601989984512, 'train/grad_norm_pre_clip': 5.118750095367432, 'timing/data': 0.0005101409624330699, 'timing/model': 1.4544675389770418, 'learning_rate/qwen_vl_interface': 7.556987905602215e-06, 'learning_rate/action_model': 7.556987905602215e-05, 'epoch': 0.76}) +[2026-06-15 22:42:10,352][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.08810591697692871, 'train/grad_norm_pre_clip': 7.740545749664307, 'timing/data': 0.00048528402112424374, 'timing/model': 1.6635412980103865, 'learning_rate/qwen_vl_interface': 7.554325319843593e-06, 'learning_rate/action_model': 7.554325319843592e-05, 'epoch': 0.76}) +[2026-06-15 22:42:16,164][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.24723637104034424, 'train/grad_norm_pre_clip': 9.575350761413574, 'timing/data': 0.0005169209907762706, 'timing/model': 1.4922817050246522, 'learning_rate/qwen_vl_interface': 7.551661786863874e-06, 'learning_rate/action_model': 7.551661786863874e-05, 'epoch': 0.76}) +[2026-06-15 22:42:22,165][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.1780831217765808, 'train/grad_norm_pre_clip': 8.860578536987305, 'timing/data': 0.0005162990419194102, 'timing/model': 1.4669653209857643, 'learning_rate/qwen_vl_interface': 7.548997307757936e-06, 'learning_rate/action_model': 7.548997307757935e-05, 'epoch': 0.76}) +[2026-06-15 22:42:28,165][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.08780089765787125, 'train/grad_norm_pre_clip': 6.650823593139648, 'timing/data': 0.0005017090006731451, 'timing/model': 1.4813809060142376, 'learning_rate/qwen_vl_interface': 7.546331883621044e-06, 'learning_rate/action_model': 7.546331883621044e-05, 'epoch': 0.76}) +[2026-06-15 22:42:34,145][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.09749982506036758, 'train/grad_norm_pre_clip': 7.643004417419434, 'timing/data': 0.0005119530251249671, 'timing/model': 1.5140768539858982, 'learning_rate/qwen_vl_interface': 7.543665515548857e-06, 'learning_rate/action_model': 7.543665515548856e-05, 'epoch': 0.76}) +[2026-06-15 22:42:40,066][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.14690211415290833, 'train/grad_norm_pre_clip': 7.692779541015625, 'timing/data': 0.0005120130372233689, 'timing/model': 1.4409996790345758, 'learning_rate/qwen_vl_interface': 7.540998204637413e-06, 'learning_rate/action_model': 7.540998204637413e-05, 'epoch': 0.76}) +[2026-06-15 22:42:45,978][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.1865953952074051, 'train/grad_norm_pre_clip': 6.345273017883301, 'timing/data': 0.0005097899702377617, 'timing/model': 1.6772441260400228, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.76}) +[2026-06-15 22:42:51,664][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.013331678695976734, 'train/grad_norm_pre_clip': 0.7282708883285522, 'timing/data': 0.0004898809711448848, 'timing/model': 1.465961848036386, 'learning_rate/qwen_vl_interface': 7.535660758682877e-06, 'learning_rate/action_model': 7.535660758682876e-05, 'epoch': 0.76}) +[2026-06-15 22:42:57,666][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.18097059428691864, 'train/grad_norm_pre_clip': 6.534189701080322, 'timing/data': 0.0005854210467077792, 'timing/model': 1.460164617979899, 'learning_rate/qwen_vl_interface': 7.532990625833806e-06, 'learning_rate/action_model': 7.532990625833805e-05, 'epoch': 0.76}) +[2026-06-15 22:43:03,664][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.24506574869155884, 'train/grad_norm_pre_clip': 12.1589937210083, 'timing/data': 0.0005164400208741426, 'timing/model': 1.478575090994127, 'learning_rate/qwen_vl_interface': 7.5303195545335225e-06, 'learning_rate/action_model': 7.530319554533522e-05, 'epoch': 0.76}) +[2026-06-15 22:43:09,643][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.16855081915855408, 'train/grad_norm_pre_clip': 4.341705799102783, 'timing/data': 0.0005082680145278573, 'timing/model': 1.511408129998017, 'learning_rate/qwen_vl_interface': 7.527647545880007e-06, 'learning_rate/action_model': 7.527647545880007e-05, 'epoch': 0.76}) +[2026-06-15 22:43:15,466][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.03459538519382477, 'train/grad_norm_pre_clip': 3.327404260635376, 'timing/data': 0.0005150180077180266, 'timing/model': 1.4602357490221038, 'learning_rate/qwen_vl_interface': 7.52497460097162e-06, 'learning_rate/action_model': 7.52497460097162e-05, 'epoch': 0.76}) +[2026-06-15 22:43:21,498][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.05992565676569939, 'train/grad_norm_pre_clip': 2.549978494644165, 'timing/data': 0.0005174310063011944, 'timing/model': 1.700166208029259, 'learning_rate/qwen_vl_interface': 7.522300720907105e-06, 'learning_rate/action_model': 7.522300720907105e-05, 'epoch': 0.76}) +[2026-06-15 22:43:27,163][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.10622546076774597, 'train/grad_norm_pre_clip': 5.57914400100708, 'timing/data': 0.0005114520317874849, 'timing/model': 1.4675640569766983, 'learning_rate/qwen_vl_interface': 7.519625906785601e-06, 'learning_rate/action_model': 7.5196259067856e-05, 'epoch': 0.76}) +[2026-06-15 22:43:33,165][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.26282232999801636, 'train/grad_norm_pre_clip': 12.617842674255371, 'timing/data': 0.0005064050201326609, 'timing/model': 1.4670318120042793, 'learning_rate/qwen_vl_interface': 7.516950159706614e-06, 'learning_rate/action_model': 7.516950159706614e-05, 'epoch': 0.76}) +[2026-06-15 22:43:39,164][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.13459566235542297, 'train/grad_norm_pre_clip': 6.054872035980225, 'timing/data': 0.0005207060021348298, 'timing/model': 1.4843267740216106, 'learning_rate/qwen_vl_interface': 7.514273480770047e-06, 'learning_rate/action_model': 7.514273480770047e-05, 'epoch': 0.76}) +[2026-06-15 22:43:45,060][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.23889276385307312, 'train/grad_norm_pre_clip': 10.28719425201416, 'timing/data': 0.0005152179510332644, 'timing/model': 1.439955428999383, 'learning_rate/qwen_vl_interface': 7.511595871076181e-06, 'learning_rate/action_model': 7.511595871076181e-05, 'epoch': 0.76}) +[2026-06-15 22:43:50,961][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.08564263582229614, 'train/grad_norm_pre_clip': 5.567554473876953, 'timing/data': 0.0005190939991734922, 'timing/model': 1.4414810619782656, 'learning_rate/qwen_vl_interface': 7.50891733172568e-06, 'learning_rate/action_model': 7.50891733172568e-05, 'epoch': 0.76}) +[2026-06-15 22:43:56,895][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.06392567604780197, 'train/grad_norm_pre_clip': 3.1904327869415283, 'timing/data': 0.0005017980001866817, 'timing/model': 1.6987924089771695, 'learning_rate/qwen_vl_interface': 7.506237863819593e-06, 'learning_rate/action_model': 7.506237863819593e-05, 'epoch': 0.77}) +[2026-06-15 22:44:02,561][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.06524214148521423, 'train/grad_norm_pre_clip': 3.0238006114959717, 'timing/data': 0.0005123339942656457, 'timing/model': 1.4617773980135098, 'learning_rate/qwen_vl_interface': 7.503557468459345e-06, 'learning_rate/action_model': 7.503557468459344e-05, 'epoch': 0.77}) +[2026-06-15 22:44:08,560][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.03993600234389305, 'train/grad_norm_pre_clip': 3.9959070682525635, 'timing/data': 0.0005083280266262591, 'timing/model': 1.4611517559969798, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.77}) +[2026-06-15 22:44:14,560][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.2970674932003021, 'train/grad_norm_pre_clip': 5.86197566986084, 'timing/data': 0.0005079469992779195, 'timing/model': 1.4809026300208643, 'learning_rate/qwen_vl_interface': 7.498193899783983e-06, 'learning_rate/action_model': 7.498193899783983e-05, 'epoch': 0.77}) +[2026-06-15 22:44:20,559][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.14491483569145203, 'train/grad_norm_pre_clip': 6.8149027824401855, 'timing/data': 0.000585180998314172, 'timing/model': 1.4945052199764177, 'learning_rate/qwen_vl_interface': 7.495510728673631e-06, 'learning_rate/action_model': 7.495510728673632e-05, 'epoch': 0.77}) +[2026-06-15 22:44:26,486][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.20271547138690948, 'train/grad_norm_pre_clip': 12.096701622009277, 'timing/data': 0.000567134004086256, 'timing/model': 1.692043345014099, 'learning_rate/qwen_vl_interface': 7.492826634518642e-06, 'learning_rate/action_model': 7.492826634518641e-05, 'epoch': 0.77}) +[2026-06-15 22:44:32,410][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.06559092551469803, 'train/grad_norm_pre_clip': 4.612383842468262, 'timing/data': 0.00047951500164344907, 'timing/model': 1.6296796899987385, 'learning_rate/qwen_vl_interface': 7.4901416184223404e-06, 'learning_rate/action_model': 7.49014161842234e-05, 'epoch': 0.77}) +[2026-06-15 22:44:38,064][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.07759822905063629, 'train/grad_norm_pre_clip': 7.196615695953369, 'timing/data': 0.0005595729453489184, 'timing/model': 1.4679281290154904, 'learning_rate/qwen_vl_interface': 7.487455681488437e-06, 'learning_rate/action_model': 7.487455681488436e-05, 'epoch': 0.77}) +[2026-06-15 22:44:43,963][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.38031789660453796, 'train/grad_norm_pre_clip': 16.885526657104492, 'timing/data': 0.0005641600000672042, 'timing/model': 1.465880025003571, 'learning_rate/qwen_vl_interface': 7.484768824821018e-06, 'learning_rate/action_model': 7.484768824821018e-05, 'epoch': 0.77}) +[2026-06-15 22:44:49,961][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.14262528717517853, 'train/grad_norm_pre_clip': 8.51921272277832, 'timing/data': 0.0005461130058392882, 'timing/model': 1.4656746319960803, 'learning_rate/qwen_vl_interface': 7.482081049524551e-06, 'learning_rate/action_model': 7.482081049524551e-05, 'epoch': 0.77}) +[2026-06-15 22:44:55,946][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.03168980032205582, 'train/grad_norm_pre_clip': 1.7164833545684814, 'timing/data': 0.0005574900424107909, 'timing/model': 1.5153408839832991, 'learning_rate/qwen_vl_interface': 7.479392356703878e-06, 'learning_rate/action_model': 7.479392356703878e-05, 'epoch': 0.77}) +[2026-06-15 22:45:01,899][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.21903803944587708, 'train/grad_norm_pre_clip': 8.996806144714355, 'timing/data': 0.0005133049562573433, 'timing/model': 1.698364796990063, 'learning_rate/qwen_vl_interface': 7.476702747464219e-06, 'learning_rate/action_model': 7.476702747464219e-05, 'epoch': 0.77}) +[2026-06-15 22:45:07,573][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.2351180464029312, 'train/grad_norm_pre_clip': 13.769905090332031, 'timing/data': 0.0005179219879209995, 'timing/model': 1.4735859659849666, 'learning_rate/qwen_vl_interface': 7.474012222911169e-06, 'learning_rate/action_model': 7.474012222911169e-05, 'epoch': 0.77}) +[2026-06-15 22:45:13,665][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.2216798961162567, 'train/grad_norm_pre_clip': 5.77844762802124, 'timing/data': 0.0005225379718467593, 'timing/model': 1.4663049050141126, 'learning_rate/qwen_vl_interface': 7.471320784150703e-06, 'learning_rate/action_model': 7.471320784150703e-05, 'epoch': 0.77}) +[2026-06-15 22:45:19,663][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.058277226984500885, 'train/grad_norm_pre_clip': 5.992888450622559, 'timing/data': 0.000535598024725914, 'timing/model': 1.484271674999036, 'learning_rate/qwen_vl_interface': 7.468628432289169e-06, 'learning_rate/action_model': 7.468628432289168e-05, 'epoch': 0.77}) +[2026-06-15 22:45:25,642][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.10008763521909714, 'train/grad_norm_pre_clip': 4.743524551391602, 'timing/data': 0.0005267540109343827, 'timing/model': 1.5132670779712498, 'learning_rate/qwen_vl_interface': 7.465935168433292e-06, 'learning_rate/action_model': 7.465935168433292e-05, 'epoch': 0.77}) +[2026-06-15 22:45:31,665][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.3563746511936188, 'train/grad_norm_pre_clip': 8.148106575012207, 'timing/data': 0.0005158990388736129, 'timing/model': 1.455290939018596, 'learning_rate/qwen_vl_interface': 7.463240993690168e-06, 'learning_rate/action_model': 7.463240993690168e-05, 'epoch': 0.77}) +[2026-06-15 22:45:37,590][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.20327091217041016, 'train/grad_norm_pre_clip': 7.8509840965271, 'timing/data': 0.000522548973094672, 'timing/model': 1.6964840129949152, 'learning_rate/qwen_vl_interface': 7.460545909167276e-06, 'learning_rate/action_model': 7.460545909167275e-05, 'epoch': 0.77}) +[2026-06-15 22:45:43,261][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.2711798846721649, 'train/grad_norm_pre_clip': 9.565622329711914, 'timing/data': 0.0005866630235686898, 'timing/model': 1.4643905239645392, 'learning_rate/qwen_vl_interface': 7.457849915972459e-06, 'learning_rate/action_model': 7.457849915972458e-05, 'epoch': 0.77}) +[2026-06-15 22:45:49,262][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.09120700508356094, 'train/grad_norm_pre_clip': 3.8098232746124268, 'timing/data': 0.000520234985742718, 'timing/model': 1.4658882159856148, 'learning_rate/qwen_vl_interface': 7.455153015213941e-06, 'learning_rate/action_model': 7.45515301521394e-05, 'epoch': 0.77}) +[2026-06-15 22:45:55,259][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.025967629626393318, 'train/grad_norm_pre_clip': 1.1308265924453735, 'timing/data': 0.0005418169894255698, 'timing/model': 1.4904573379899375, 'learning_rate/qwen_vl_interface': 7.452455208000311e-06, 'learning_rate/action_model': 7.452455208000311e-05, 'epoch': 0.77}) +[2026-06-15 22:46:01,161][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.2098575234413147, 'train/grad_norm_pre_clip': 7.602613925933838, 'timing/data': 0.0004911330179311335, 'timing/model': 1.4394781829905696, 'learning_rate/qwen_vl_interface': 7.449756495440541e-06, 'learning_rate/action_model': 7.44975649544054e-05, 'epoch': 0.77}) +[2026-06-15 22:55:10,355][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.1193377822637558, 'train/grad_norm_pre_clip': 5.8819804191589355, 'eval/loss': 0.06728699803352356, 'eval/flap_precision': 0.9090909090909091, 'eval/flap_recall': 0.9018036072144289, 'eval/flap_f1': 0.9054325955734407, 'eval/flap_rate_model': 0.1375, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9054325955734407, 'timing/data': 0.0004999659722670913, 'timing/model': 1.4400195889757015, 'learning_rate/qwen_vl_interface': 7.447056878643969e-06, 'learning_rate/action_model': 7.447056878643968e-05, 'epoch': 0.77}) +[2026-06-15 22:55:10,356][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1800_state +[2026-06-15 22:55:10,361][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 22:55:15,892][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1800_state/model.safetensors +[2026-06-15 22:55:29,430][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1800_state/optimizer.bin +[2026-06-15 22:55:29,431][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1800_state/sampler.bin +[2026-06-15 22:55:29,431][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1800_state/sampler_1.bin +[2026-06-15 22:55:29,432][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1800_state/random_states_0.pkl +[2026-06-15 22:55:29,544][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1800 +[2026-06-15 22:55:29,544][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 22:55:29,554][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 22:55:35,536][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.10838012397289276, 'train/grad_norm_pre_clip': 4.766684055328369, 'timing/data': 0.0005239509628154337, 'timing/model': 1.4893343820003793, 'learning_rate/qwen_vl_interface': 7.444356358720303e-06, 'learning_rate/action_model': 7.444356358720302e-05, 'epoch': 0.77}) +[2026-06-15 22:55:41,462][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.25620290637016296, 'train/grad_norm_pre_clip': 9.770694732666016, 'timing/data': 0.0005037310183979571, 'timing/model': 1.6876332570100203, 'learning_rate/qwen_vl_interface': 7.4416549367796255e-06, 'learning_rate/action_model': 7.441654936779625e-05, 'epoch': 0.78}) +[2026-06-15 22:55:47,284][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.10279322415590286, 'train/grad_norm_pre_clip': 5.514211654663086, 'timing/data': 0.0005401850212365389, 'timing/model': 1.6083087279694155, 'learning_rate/qwen_vl_interface': 7.4389526139323895e-06, 'learning_rate/action_model': 7.438952613932389e-05, 'epoch': 0.78}) +[2026-06-15 22:55:53,021][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.4135683476924896, 'train/grad_norm_pre_clip': 24.67351722717285, 'timing/data': 0.0005463340203277767, 'timing/model': 1.540817615983542, 'learning_rate/qwen_vl_interface': 7.436249391289419e-06, 'learning_rate/action_model': 7.436249391289418e-05, 'epoch': 0.78}) +[2026-06-15 22:55:59,039][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.20513412356376648, 'train/grad_norm_pre_clip': 5.945409297943115, 'timing/data': 0.0005080080009065568, 'timing/model': 1.5580936080077663, 'learning_rate/qwen_vl_interface': 7.433545269961906e-06, 'learning_rate/action_model': 7.433545269961906e-05, 'epoch': 0.78}) +[2026-06-15 22:56:04,948][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.060097791254520416, 'train/grad_norm_pre_clip': 2.7912745475769043, 'timing/data': 0.0005364089738577604, 'timing/model': 1.4654065819922835, 'learning_rate/qwen_vl_interface': 7.430840251061411e-06, 'learning_rate/action_model': 7.430840251061411e-05, 'epoch': 0.78}) +[2026-06-15 22:56:10,945][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.04604718089103699, 'train/grad_norm_pre_clip': 3.8208625316619873, 'timing/data': 0.0006196119938977063, 'timing/model': 1.4502640360151418, 'learning_rate/qwen_vl_interface': 7.428134335699866e-06, 'learning_rate/action_model': 7.428134335699865e-05, 'epoch': 0.78}) +[2026-06-15 22:56:16,881][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.48855769634246826, 'train/grad_norm_pre_clip': 10.957143783569336, 'timing/data': 0.0005042520351707935, 'timing/model': 1.6983716030372307, 'learning_rate/qwen_vl_interface': 7.425427524989572e-06, 'learning_rate/action_model': 7.425427524989572e-05, 'epoch': 0.78}) +[2026-06-15 22:56:22,647][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.09387397021055222, 'train/grad_norm_pre_clip': 8.129560470581055, 'timing/data': 0.0005211469833739102, 'timing/model': 1.466801372996997, 'learning_rate/qwen_vl_interface': 7.422719820043194e-06, 'learning_rate/action_model': 7.422719820043193e-05, 'epoch': 0.78}) +[2026-06-15 22:56:28,648][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.19031493365764618, 'train/grad_norm_pre_clip': 10.253494262695312, 'timing/data': 0.0005287079839035869, 'timing/model': 1.4678543219924904, 'learning_rate/qwen_vl_interface': 7.420011221973771e-06, 'learning_rate/action_model': 7.420011221973771e-05, 'epoch': 0.78}) +[2026-06-15 22:56:34,547][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.1997235119342804, 'train/grad_norm_pre_clip': 10.387307167053223, 'timing/data': 0.00051462696865201, 'timing/model': 1.4639831569511443, 'learning_rate/qwen_vl_interface': 7.417301731894702e-06, 'learning_rate/action_model': 7.417301731894701e-05, 'epoch': 0.78}) +[2026-06-15 22:56:40,443][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.04958281293511391, 'train/grad_norm_pre_clip': 2.460233211517334, 'timing/data': 0.000515487976372242, 'timing/model': 1.462323796004057, 'learning_rate/qwen_vl_interface': 7.414591350919756e-06, 'learning_rate/action_model': 7.414591350919756e-05, 'epoch': 0.78}) +[2026-06-15 22:56:46,347][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.10373315215110779, 'train/grad_norm_pre_clip': 10.463236808776855, 'timing/data': 0.0005078570102341473, 'timing/model': 1.4293408450321294, 'learning_rate/qwen_vl_interface': 7.4118800801630715e-06, 'learning_rate/action_model': 7.411880080163072e-05, 'epoch': 0.78}) +[2026-06-15 22:56:52,278][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.09774579852819443, 'train/grad_norm_pre_clip': 6.280360221862793, 'timing/data': 0.0004941970109939575, 'timing/model': 1.6962570699979551, 'learning_rate/qwen_vl_interface': 7.409167920739146e-06, 'learning_rate/action_model': 7.409167920739147e-05, 'epoch': 0.78}) +[2026-06-15 22:56:57,948][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.20642805099487305, 'train/grad_norm_pre_clip': 11.871915817260742, 'timing/data': 0.000484332034830004, 'timing/model': 1.4641182139748707, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.78}) +[2026-06-15 22:57:03,852][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.11580280214548111, 'train/grad_norm_pre_clip': 6.365407943725586, 'timing/data': 0.00048819801304489374, 'timing/model': 1.465066338016186, 'learning_rate/qwen_vl_interface': 7.403740940349408e-06, 'learning_rate/action_model': 7.403740940349408e-05, 'epoch': 0.78}) +[2026-06-15 22:57:09,754][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.14460286498069763, 'train/grad_norm_pre_clip': 8.1779203414917, 'timing/data': 0.0004877980099990964, 'timing/model': 1.4673326119664125, 'learning_rate/qwen_vl_interface': 7.401026121614421e-06, 'learning_rate/action_model': 7.40102612161442e-05, 'epoch': 0.78}) +[2026-06-15 22:57:15,553][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.1915382742881775, 'train/grad_norm_pre_clip': 15.042999267578125, 'timing/data': 0.0004976720083504915, 'timing/model': 1.4649556850199588, 'learning_rate/qwen_vl_interface': 7.398310418673846e-06, 'learning_rate/action_model': 7.398310418673846e-05, 'epoch': 0.78}) +[2026-06-15 22:57:21,354][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.13403020799160004, 'train/grad_norm_pre_clip': 4.701270580291748, 'timing/data': 0.0004949780413880944, 'timing/model': 1.4426364329992794, 'learning_rate/qwen_vl_interface': 7.395593832644007e-06, 'learning_rate/action_model': 7.395593832644007e-05, 'epoch': 0.78}) +[2026-06-15 22:57:27,211][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.09056153148412704, 'train/grad_norm_pre_clip': 3.9685471057891846, 'timing/data': 0.0004814279964193702, 'timing/model': 1.6218565829913132, 'learning_rate/qwen_vl_interface': 7.39287636464159e-06, 'learning_rate/action_model': 7.39287636464159e-05, 'epoch': 0.78}) +[2026-06-15 22:57:32,853][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.1457330584526062, 'train/grad_norm_pre_clip': 13.764936447143555, 'timing/data': 0.0004834210267290473, 'timing/model': 1.4639992879820056, 'learning_rate/qwen_vl_interface': 7.3901580157836425e-06, 'learning_rate/action_model': 7.390158015783643e-05, 'epoch': 0.78}) +[2026-06-15 22:57:38,756][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.06302164494991302, 'train/grad_norm_pre_clip': 3.0168025493621826, 'timing/data': 0.0004959990037605166, 'timing/model': 1.4671964200097136, 'learning_rate/qwen_vl_interface': 7.387438787187576e-06, 'learning_rate/action_model': 7.387438787187576e-05, 'epoch': 0.78}) +[2026-06-15 22:57:44,655][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.3257954716682434, 'train/grad_norm_pre_clip': 12.18483829498291, 'timing/data': 0.0004927249974571168, 'timing/model': 1.4666945419739932, 'learning_rate/qwen_vl_interface': 7.3847186799711625e-06, 'learning_rate/action_model': 7.384718679971162e-05, 'epoch': 0.78}) +[2026-06-15 22:57:50,656][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.14559412002563477, 'train/grad_norm_pre_clip': 7.941246509552002, 'timing/data': 0.0005059140385128558, 'timing/model': 1.4343112250207923, 'learning_rate/qwen_vl_interface': 7.381997695252535e-06, 'learning_rate/action_model': 7.381997695252535e-05, 'epoch': 0.78}) +[2026-06-15 22:57:56,580][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.07468722015619278, 'train/grad_norm_pre_clip': 4.120753288269043, 'timing/data': 0.0005167099880054593, 'timing/model': 1.6770371040329337, 'learning_rate/qwen_vl_interface': 7.379275834150186e-06, 'learning_rate/action_model': 7.379275834150186e-05, 'epoch': 0.79}) +[2026-06-15 22:58:02,431][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.10745561122894287, 'train/grad_norm_pre_clip': 3.6829473972320557, 'timing/data': 0.0005209060036577284, 'timing/model': 1.6430039900005795, 'learning_rate/qwen_vl_interface': 7.376553097782972e-06, 'learning_rate/action_model': 7.376553097782972e-05, 'epoch': 0.79}) +[2026-06-15 22:58:08,151][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.2890065312385559, 'train/grad_norm_pre_clip': 12.713582038879395, 'timing/data': 0.000509359990246594, 'timing/model': 1.4644302200176753, 'learning_rate/qwen_vl_interface': 7.373829487270109e-06, 'learning_rate/action_model': 7.373829487270109e-05, 'epoch': 0.79}) +[2026-06-15 22:58:14,146][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.08767655491828918, 'train/grad_norm_pre_clip': 4.309113502502441, 'timing/data': 0.0005018289666622877, 'timing/model': 1.5458193870144896, 'learning_rate/qwen_vl_interface': 7.371105003731165e-06, 'learning_rate/action_model': 7.371105003731164e-05, 'epoch': 0.79}) +[2026-06-15 22:58:20,245][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.2613561451435089, 'train/grad_norm_pre_clip': 8.009117126464844, 'timing/data': 0.0005174609832465649, 'timing/model': 1.4791035949601792, 'learning_rate/qwen_vl_interface': 7.3683796482860755e-06, 'learning_rate/action_model': 7.368379648286075e-05, 'epoch': 0.79}) +[2026-06-15 22:58:26,263][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.16773176193237305, 'train/grad_norm_pre_clip': 7.267613410949707, 'timing/data': 0.0005911890184506774, 'timing/model': 1.4292647910187952, 'learning_rate/qwen_vl_interface': 7.365653422055131e-06, 'learning_rate/action_model': 7.365653422055131e-05, 'epoch': 0.79}) +[2026-06-15 22:58:32,279][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.13494284451007843, 'train/grad_norm_pre_clip': 7.695278167724609, 'timing/data': 0.0007248870097100735, 'timing/model': 1.7150327049894258, 'learning_rate/qwen_vl_interface': 7.3629263261589816e-06, 'learning_rate/action_model': 7.362926326158981e-05, 'epoch': 0.79}) +[2026-06-15 22:58:38,052][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.18712486326694489, 'train/grad_norm_pre_clip': 7.063964366912842, 'timing/data': 0.0005043220007792115, 'timing/model': 1.5643643660005182, 'learning_rate/qwen_vl_interface': 7.36019836171863e-06, 'learning_rate/action_model': 7.360198361718629e-05, 'epoch': 0.79}) +[2026-06-15 22:58:43,754][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.020741688087582588, 'train/grad_norm_pre_clip': 1.6295671463012695, 'timing/data': 0.0004948179703205824, 'timing/model': 1.4667210289626382, 'learning_rate/qwen_vl_interface': 7.357469529855441e-06, 'learning_rate/action_model': 7.35746952985544e-05, 'epoch': 0.79}) +[2026-06-15 22:58:49,754][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.1029864102602005, 'train/grad_norm_pre_clip': 4.884558200836182, 'timing/data': 0.0006166069651953876, 'timing/model': 1.465245077968575, 'learning_rate/qwen_vl_interface': 7.354739831691134e-06, 'learning_rate/action_model': 7.354739831691135e-05, 'epoch': 0.79}) +[2026-06-15 22:58:55,653][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.14333367347717285, 'train/grad_norm_pre_clip': 11.444684982299805, 'timing/data': 0.0004964699619449675, 'timing/model': 1.4594967719749548, 'learning_rate/qwen_vl_interface': 7.352009268347786e-06, 'learning_rate/action_model': 7.352009268347786e-05, 'epoch': 0.79}) +[2026-06-15 22:59:01,553][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.08259470015764236, 'train/grad_norm_pre_clip': 3.2588584423065186, 'timing/data': 0.0004896610043942928, 'timing/model': 1.4320398510317318, 'learning_rate/qwen_vl_interface': 7.349277840947828e-06, 'learning_rate/action_model': 7.349277840947828e-05, 'epoch': 0.79}) +[2026-06-15 22:59:07,482][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.09068793058395386, 'train/grad_norm_pre_clip': 4.088081359863281, 'timing/data': 0.00048556504771113396, 'timing/model': 1.6964355259551667, 'learning_rate/qwen_vl_interface': 7.346545550614046e-06, 'learning_rate/action_model': 7.346545550614046e-05, 'epoch': 0.79}) +[2026-06-15 22:59:13,151][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.0482618547976017, 'train/grad_norm_pre_clip': 4.729430675506592, 'timing/data': 0.0004900009953416884, 'timing/model': 1.4501781330327503, 'learning_rate/qwen_vl_interface': 7.343812398469582e-06, 'learning_rate/action_model': 7.343812398469581e-05, 'epoch': 0.79}) +[2026-06-15 22:59:19,154][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.03879764676094055, 'train/grad_norm_pre_clip': 1.2674647569656372, 'timing/data': 0.0005197739810682833, 'timing/model': 1.4626820760313421, 'learning_rate/qwen_vl_interface': 7.34107838563793e-06, 'learning_rate/action_model': 7.341078385637929e-05, 'epoch': 0.79}) +[2026-06-15 22:59:25,053][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.08140324056148529, 'train/grad_norm_pre_clip': 5.528182029724121, 'timing/data': 0.0005393540486693382, 'timing/model': 1.4652430139831267, 'learning_rate/qwen_vl_interface': 7.338343513242941e-06, 'learning_rate/action_model': 7.338343513242941e-05, 'epoch': 0.79}) +[2026-06-15 22:59:30,953][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.08844691514968872, 'train/grad_norm_pre_clip': 3.9032647609710693, 'timing/data': 0.0005099100526422262, 'timing/model': 1.4643119130050763, 'learning_rate/qwen_vl_interface': 7.335607782408818e-06, 'learning_rate/action_model': 7.335607782408818e-05, 'epoch': 0.79}) +[2026-06-15 22:59:36,856][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.03358374908566475, 'train/grad_norm_pre_clip': 2.819046974182129, 'timing/data': 0.0005152279627509415, 'timing/model': 1.4335286599816754, 'learning_rate/qwen_vl_interface': 7.3328711942601145e-06, 'learning_rate/action_model': 7.332871194260115e-05, 'epoch': 0.79}) +[2026-06-15 22:59:42,784][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.02186243049800396, 'train/grad_norm_pre_clip': 1.7065385580062866, 'timing/data': 0.0005121830035932362, 'timing/model': 1.6943653299822472, 'learning_rate/qwen_vl_interface': 7.33013374992174e-06, 'learning_rate/action_model': 7.33013374992174e-05, 'epoch': 0.79}) +[2026-06-15 22:59:48,456][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.258318156003952, 'train/grad_norm_pre_clip': 11.958093643188477, 'timing/data': 0.0005052229971624911, 'timing/model': 1.464244884031359, 'learning_rate/qwen_vl_interface': 7.327395450518956e-06, 'learning_rate/action_model': 7.327395450518956e-05, 'epoch': 0.79}) +[2026-06-15 22:59:54,455][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.31769269704818726, 'train/grad_norm_pre_clip': 19.194440841674805, 'timing/data': 0.0005177219863981009, 'timing/model': 1.4667257639812306, 'learning_rate/qwen_vl_interface': 7.324656297177372e-06, 'learning_rate/action_model': 7.324656297177371e-05, 'epoch': 0.79}) +[2026-06-15 23:00:00,255][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.07250626385211945, 'train/grad_norm_pre_clip': 2.3719677925109863, 'timing/data': 0.0005142170120961964, 'timing/model': 1.4651126809767447, 'learning_rate/qwen_vl_interface': 7.32191629102295e-06, 'learning_rate/action_model': 7.32191629102295e-05, 'epoch': 0.79}) +[2026-06-15 23:00:06,156][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.00441511208191514, 'train/grad_norm_pre_clip': 1.361453652381897, 'timing/data': 0.0005073060165159404, 'timing/model': 1.4652749900124036, 'learning_rate/qwen_vl_interface': 7.319175433182004e-06, 'learning_rate/action_model': 7.319175433182004e-05, 'epoch': 0.79}) +[2026-06-15 23:00:12,060][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.11417548358440399, 'train/grad_norm_pre_clip': 5.2322096824646, 'timing/data': 0.0005079369875602424, 'timing/model': 1.4356409670435824, 'learning_rate/qwen_vl_interface': 7.316433724781198e-06, 'learning_rate/action_model': 7.316433724781197e-05, 'epoch': 0.8}) +[2026-06-15 23:00:17,985][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.041454512625932693, 'train/grad_norm_pre_clip': 5.143363952636719, 'timing/data': 0.000500246009323746, 'timing/model': 1.6952613170142286, 'learning_rate/qwen_vl_interface': 7.313691166947543e-06, 'learning_rate/action_model': 7.313691166947543e-05, 'epoch': 0.8}) +[2026-06-15 23:00:23,654][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.14808648824691772, 'train/grad_norm_pre_clip': 12.426390647888184, 'timing/data': 0.0004886789829470217, 'timing/model': 1.4649854130111635, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.8}) +[2026-06-15 23:00:29,653][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.08582773804664612, 'train/grad_norm_pre_clip': 17.092004776000977, 'timing/data': 0.0004854639992117882, 'timing/model': 1.4652553590130992, 'learning_rate/qwen_vl_interface': 7.308203507491485e-06, 'learning_rate/action_model': 7.308203507491484e-05, 'epoch': 0.8}) +[2026-06-15 23:00:35,555][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.0849015861749649, 'train/grad_norm_pre_clip': 6.5182952880859375, 'timing/data': 0.00048648600932210684, 'timing/model': 1.4674463569535874, 'learning_rate/qwen_vl_interface': 7.3054584081248515e-06, 'learning_rate/action_model': 7.305458408124852e-05, 'epoch': 0.8}) +[2026-06-15 23:00:41,455][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.0466916486620903, 'train/grad_norm_pre_clip': 2.8380775451660156, 'timing/data': 0.0005148479831404984, 'timing/model': 1.4350893689552322, 'learning_rate/qwen_vl_interface': 7.302712463836907e-06, 'learning_rate/action_model': 7.302712463836907e-05, 'epoch': 0.8}) +[2026-06-15 23:00:47,348][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.14534512162208557, 'train/grad_norm_pre_clip': 8.661383628845215, 'timing/data': 0.0005209970404393971, 'timing/model': 1.5085208009695634, 'learning_rate/qwen_vl_interface': 7.2999656757564086e-06, 'learning_rate/action_model': 7.299965675756409e-05, 'epoch': 0.8}) +[2026-06-15 23:00:53,235][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.2573722004890442, 'train/grad_norm_pre_clip': 9.4132719039917, 'timing/data': 0.0005184439942240715, 'timing/model': 1.6428502789931372, 'learning_rate/qwen_vl_interface': 7.297218045012453e-06, 'learning_rate/action_model': 7.297218045012453e-05, 'epoch': 0.8}) +[2026-06-15 23:00:58,955][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.2765395939350128, 'train/grad_norm_pre_clip': 39.35463333129883, 'timing/data': 0.000517801963724196, 'timing/model': 1.46546447603032, 'learning_rate/qwen_vl_interface': 7.29446957273449e-06, 'learning_rate/action_model': 7.294469572734489e-05, 'epoch': 0.8}) +[2026-06-15 23:01:05,055][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.19776833057403564, 'train/grad_norm_pre_clip': 6.44352388381958, 'timing/data': 0.0005125639727339149, 'timing/model': 1.46462317899568, 'learning_rate/qwen_vl_interface': 7.2917202600523104e-06, 'learning_rate/action_model': 7.29172026005231e-05, 'epoch': 0.8}) +[2026-06-15 23:01:11,056][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.05871511250734329, 'train/grad_norm_pre_clip': 7.118752479553223, 'timing/data': 0.0005199049483053386, 'timing/model': 1.4666875409893692, 'learning_rate/qwen_vl_interface': 7.2889701080960565e-06, 'learning_rate/action_model': 7.288970108096056e-05, 'epoch': 0.8}) +[2026-06-15 23:01:16,959][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.13514553010463715, 'train/grad_norm_pre_clip': 7.163771629333496, 'timing/data': 0.0005077369860373437, 'timing/model': 1.434538542991504, 'learning_rate/qwen_vl_interface': 7.286219117996209e-06, 'learning_rate/action_model': 7.286219117996207e-05, 'epoch': 0.8}) +[2026-06-15 23:01:22,883][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.11103016138076782, 'train/grad_norm_pre_clip': 10.634186744689941, 'timing/data': 0.0005153079982846975, 'timing/model': 1.6946066439850256, 'learning_rate/qwen_vl_interface': 7.283467290883598e-06, 'learning_rate/action_model': 7.283467290883598e-05, 'epoch': 0.8}) +[2026-06-15 23:01:28,659][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.06767264008522034, 'train/grad_norm_pre_clip': 3.148970603942871, 'timing/data': 0.0005146180046722293, 'timing/model': 1.5683447949704714, 'learning_rate/qwen_vl_interface': 7.2807146278893925e-06, 'learning_rate/action_model': 7.280714627889393e-05, 'epoch': 0.8}) +[2026-06-15 23:01:34,357][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.010229798965156078, 'train/grad_norm_pre_clip': 1.1390132904052734, 'timing/data': 0.0005244720377959311, 'timing/model': 1.4660185169777833, 'learning_rate/qwen_vl_interface': 7.277961130145111e-06, 'learning_rate/action_model': 7.277961130145111e-05, 'epoch': 0.8}) +[2026-06-15 23:01:40,355][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.13996049761772156, 'train/grad_norm_pre_clip': 10.725786209106445, 'timing/data': 0.0005217479774728417, 'timing/model': 1.4638053919770755, 'learning_rate/qwen_vl_interface': 7.2752067987826116e-06, 'learning_rate/action_model': 7.275206798782611e-05, 'epoch': 0.8}) +[2026-06-15 23:01:46,254][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.22486840188503265, 'train/grad_norm_pre_clip': 14.785486221313477, 'timing/data': 0.0005298500182107091, 'timing/model': 1.4522357729729265, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.8}) +[2026-06-15 23:01:52,161][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.12268068641424179, 'train/grad_norm_pre_clip': 9.118185043334961, 'timing/data': 0.0005366799887269735, 'timing/model': 1.4331941950367764, 'learning_rate/qwen_vl_interface': 7.269695639732108e-06, 'learning_rate/action_model': 7.269695639732108e-05, 'epoch': 0.8}) +[2026-06-15 23:01:58,132][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.034710854291915894, 'train/grad_norm_pre_clip': 2.8425254821777344, 'timing/data': 0.0005095499800518155, 'timing/model': 1.7258254869957455, 'learning_rate/qwen_vl_interface': 7.266938814309533e-06, 'learning_rate/action_model': 7.266938814309533e-05, 'epoch': 0.8}) +[2026-06-15 23:02:03,852][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.12322665750980377, 'train/grad_norm_pre_clip': 8.709176063537598, 'timing/data': 0.0005161900189705193, 'timing/model': 1.464229331992101, 'learning_rate/qwen_vl_interface': 7.264181159799598e-06, 'learning_rate/action_model': 7.264181159799597e-05, 'epoch': 0.8}) +[2026-06-15 23:02:09,752][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.11970290541648865, 'train/grad_norm_pre_clip': 6.914144515991211, 'timing/data': 0.0005094599910080433, 'timing/model': 1.4658578229718842, 'learning_rate/qwen_vl_interface': 7.261422677335871e-06, 'learning_rate/action_model': 7.261422677335871e-05, 'epoch': 0.8}) +[2026-06-15 23:02:15,651][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.07791941612958908, 'train/grad_norm_pre_clip': 14.254037857055664, 'timing/data': 0.0005123240407556295, 'timing/model': 1.4654810320353135, 'learning_rate/qwen_vl_interface': 7.258663368052259e-06, 'learning_rate/action_model': 7.258663368052259e-05, 'epoch': 0.8}) +[2026-06-15 23:02:21,548][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.1179356798529625, 'train/grad_norm_pre_clip': 12.064162254333496, 'timing/data': 0.0005130160134285688, 'timing/model': 1.4645625040284358, 'learning_rate/qwen_vl_interface': 7.2559032330830105e-06, 'learning_rate/action_model': 7.25590323308301e-05, 'epoch': 0.8}) +[2026-06-15 23:02:27,451][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.24671532213687897, 'train/grad_norm_pre_clip': 8.37466049194336, 'timing/data': 0.0005186330527067184, 'timing/model': 1.4290177039802074, 'learning_rate/qwen_vl_interface': 7.253142273562712e-06, 'learning_rate/action_model': 7.253142273562712e-05, 'epoch': 0.81}) +[2026-06-15 23:02:33,388][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.07039833068847656, 'train/grad_norm_pre_clip': 8.672544479370117, 'timing/data': 0.0005406960262916982, 'timing/model': 1.699606662034057, 'learning_rate/qwen_vl_interface': 7.25038049062629e-06, 'learning_rate/action_model': 7.250380490626289e-05, 'epoch': 0.81}) +[2026-06-15 23:02:39,156][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.066470205783844, 'train/grad_norm_pre_clip': 4.573671340942383, 'timing/data': 0.0005216680001467466, 'timing/model': 1.4636025069630705, 'learning_rate/qwen_vl_interface': 7.24761788540901e-06, 'learning_rate/action_model': 7.24761788540901e-05, 'epoch': 0.81}) +[2026-06-15 23:02:45,152][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.1543608158826828, 'train/grad_norm_pre_clip': 7.055873870849609, 'timing/data': 0.0005187939968891442, 'timing/model': 1.443378550990019, 'learning_rate/qwen_vl_interface': 7.244854459046473e-06, 'learning_rate/action_model': 7.244854459046473e-05, 'epoch': 0.81}) +[2026-06-15 23:02:50,952][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.07664434611797333, 'train/grad_norm_pre_clip': 10.221846580505371, 'timing/data': 0.0005089189507998526, 'timing/model': 1.4614932120312005, 'learning_rate/qwen_vl_interface': 7.24209021267462e-06, 'learning_rate/action_model': 7.24209021267462e-05, 'epoch': 0.81}) +[2026-06-15 23:02:56,753][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.1373264342546463, 'train/grad_norm_pre_clip': 10.89693832397461, 'timing/data': 0.0005135259707458317, 'timing/model': 1.4653716969769448, 'learning_rate/qwen_vl_interface': 7.23932514742973e-06, 'learning_rate/action_model': 7.23932514742973e-05, 'epoch': 0.81}) +[2026-06-15 23:03:02,957][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.12315256148576736, 'train/grad_norm_pre_clip': 7.421051502227783, 'timing/data': 0.0005024499841965735, 'timing/model': 1.4380472299526446, 'learning_rate/qwen_vl_interface': 7.236559264448413e-06, 'learning_rate/action_model': 7.236559264448412e-05, 'epoch': 0.81}) +[2026-06-15 23:03:08,822][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.09579689055681229, 'train/grad_norm_pre_clip': 9.658050537109375, 'timing/data': 0.0005473559722304344, 'timing/model': 1.6276744070346467, 'learning_rate/qwen_vl_interface': 7.233792564867623e-06, 'learning_rate/action_model': 7.233792564867623e-05, 'epoch': 0.81}) +[2026-06-15 23:03:14,556][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.03601526468992233, 'train/grad_norm_pre_clip': 1.8426547050476074, 'timing/data': 0.000631138973403722, 'timing/model': 1.4622576299589127, 'learning_rate/qwen_vl_interface': 7.231025049824644e-06, 'learning_rate/action_model': 7.231025049824643e-05, 'epoch': 0.81}) +[2026-06-15 23:03:20,654][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.31643030047416687, 'train/grad_norm_pre_clip': 17.10999870300293, 'timing/data': 0.0005242420011200011, 'timing/model': 1.483925680979155, 'learning_rate/qwen_vl_interface': 7.228256720457097e-06, 'learning_rate/action_model': 7.228256720457096e-05, 'epoch': 0.81}) +[2026-06-15 23:03:26,731][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.06376731395721436, 'train/grad_norm_pre_clip': 4.6006760597229, 'timing/data': 0.0005013380432501435, 'timing/model': 1.537063653988298, 'learning_rate/qwen_vl_interface': 7.225487577902937e-06, 'learning_rate/action_model': 7.225487577902936e-05, 'epoch': 0.81}) +[2026-06-15 23:03:32,724][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.023730603978037834, 'train/grad_norm_pre_clip': 4.804285526275635, 'timing/data': 0.0005034109926782548, 'timing/model': 1.496066874009557, 'learning_rate/qwen_vl_interface': 7.222717623300455e-06, 'learning_rate/action_model': 7.222717623300455e-05, 'epoch': 0.81}) +[2026-06-15 23:03:38,709][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.2553846538066864, 'train/grad_norm_pre_clip': 7.6393280029296875, 'timing/data': 0.0005092399660497904, 'timing/model': 1.71734017401468, 'learning_rate/qwen_vl_interface': 7.219946857788273e-06, 'learning_rate/action_model': 7.219946857788272e-05, 'epoch': 0.81}) +[2026-06-15 23:03:44,539][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.018284600228071213, 'train/grad_norm_pre_clip': 2.329653739929199, 'timing/data': 0.0005083380383439362, 'timing/model': 1.646355633973144, 'learning_rate/qwen_vl_interface': 7.21717528250535e-06, 'learning_rate/action_model': 7.21717528250535e-05, 'epoch': 0.81}) +[2026-06-15 23:03:50,257][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.20466327667236328, 'train/grad_norm_pre_clip': 10.579876899719238, 'timing/data': 0.0005130759673193097, 'timing/model': 1.4565276049543172, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 0.81}) +[2026-06-15 23:03:56,256][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.24386899173259735, 'train/grad_norm_pre_clip': 9.027144432067871, 'timing/data': 0.0005047530285082757, 'timing/model': 1.4642876950092614, 'learning_rate/qwen_vl_interface': 7.211629707184768e-06, 'learning_rate/action_model': 7.211629707184768e-05, 'epoch': 0.81}) +[2026-06-15 23:04:02,257][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.06229344755411148, 'train/grad_norm_pre_clip': 5.793797492980957, 'timing/data': 0.0005152779631316662, 'timing/model': 1.4595996119896881, 'learning_rate/qwen_vl_interface': 7.2088557094266885e-06, 'learning_rate/action_model': 7.208855709426688e-05, 'epoch': 0.81}) +[2026-06-15 23:04:08,159][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.10694572329521179, 'train/grad_norm_pre_clip': 6.667070388793945, 'timing/data': 0.0005093900253996253, 'timing/model': 1.4357523079961538, 'learning_rate/qwen_vl_interface': 7.2060809064570195e-06, 'learning_rate/action_model': 7.20608090645702e-05, 'epoch': 0.81}) +[2026-06-15 23:04:14,086][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.04143056273460388, 'train/grad_norm_pre_clip': 3.6204869747161865, 'timing/data': 0.0005150079960003495, 'timing/model': 1.6945076870033517, 'learning_rate/qwen_vl_interface': 7.203305299416375e-06, 'learning_rate/action_model': 7.203305299416375e-05, 'epoch': 0.81}) +[2026-06-15 23:04:19,858][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.22958995401859283, 'train/grad_norm_pre_clip': 5.199850559234619, 'timing/data': 0.0005147180054336786, 'timing/model': 1.5658518960117362, 'learning_rate/qwen_vl_interface': 7.200528889445707e-06, 'learning_rate/action_model': 7.200528889445706e-05, 'epoch': 0.81}) +[2026-06-15 23:04:25,556][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.04212042689323425, 'train/grad_norm_pre_clip': 3.3025870323181152, 'timing/data': 0.0005236510187387466, 'timing/model': 1.4644822590053082, 'learning_rate/qwen_vl_interface': 7.19775167768629e-06, 'learning_rate/action_model': 7.19775167768629e-05, 'epoch': 0.81}) +[2026-06-15 23:04:31,656][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.17928528785705566, 'train/grad_norm_pre_clip': 4.66728401184082, 'timing/data': 0.00052367098396644, 'timing/model': 1.4638923889724538, 'learning_rate/qwen_vl_interface': 7.194973665279731e-06, 'learning_rate/action_model': 7.19497366527973e-05, 'epoch': 0.81}) +[2026-06-15 23:04:37,556][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.17630153894424438, 'train/grad_norm_pre_clip': 13.733049392700195, 'timing/data': 0.0005139370332472026, 'timing/model': 1.4590077169705182, 'learning_rate/qwen_vl_interface': 7.1921948533679666e-06, 'learning_rate/action_model': 7.192194853367967e-05, 'epoch': 0.81}) +[2026-06-15 23:04:43,458][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.043627023696899414, 'train/grad_norm_pre_clip': 2.8945226669311523, 'timing/data': 0.0005058639799244702, 'timing/model': 1.4339914250303991, 'learning_rate/qwen_vl_interface': 7.1894152430932605e-06, 'learning_rate/action_model': 7.189415243093261e-05, 'epoch': 0.81}) +[2026-06-15 23:04:49,485][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.1218520924448967, 'train/grad_norm_pre_clip': 6.285579204559326, 'timing/data': 0.000520795991178602, 'timing/model': 1.694682952016592, 'learning_rate/qwen_vl_interface': 7.186634835598205e-06, 'learning_rate/action_model': 7.186634835598205e-05, 'epoch': 0.82}) +[2026-06-15 23:04:55,157][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.013552774675190449, 'train/grad_norm_pre_clip': 1.196661353111267, 'timing/data': 0.0005241910112090409, 'timing/model': 1.46364694199292, 'learning_rate/qwen_vl_interface': 7.183853632025724e-06, 'learning_rate/action_model': 7.183853632025723e-05, 'epoch': 0.82}) +[2026-06-15 23:05:01,150][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.07124806940555573, 'train/grad_norm_pre_clip': 3.7593719959259033, 'timing/data': 0.0005335850291885436, 'timing/model': 1.4588757869787514, 'learning_rate/qwen_vl_interface': 7.181071633519062e-06, 'learning_rate/action_model': 7.181071633519062e-05, 'epoch': 0.82}) +[2026-06-15 23:05:07,053][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.16363734006881714, 'train/grad_norm_pre_clip': 4.506092548370361, 'timing/data': 0.0005065349978394806, 'timing/model': 1.4662026700098068, 'learning_rate/qwen_vl_interface': 7.178288841221795e-06, 'learning_rate/action_model': 7.178288841221794e-05, 'epoch': 0.82}) +[2026-06-15 23:05:13,122][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.44871261715888977, 'train/grad_norm_pre_clip': 13.78542709350586, 'timing/data': 0.0004901720094494522, 'timing/model': 1.5546778480056673, 'learning_rate/qwen_vl_interface': 7.1755052562778215e-06, 'learning_rate/action_model': 7.175505256277822e-05, 'epoch': 0.82}) +[2026-06-15 23:14:31,757][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.0066234958358109, 'train/grad_norm_pre_clip': 0.7488064169883728, 'eval/loss': 0.07391320914030075, 'eval/flap_precision': 0.9281183932346723, 'eval/flap_recall': 0.8797595190380761, 'eval/flap_f1': 0.9032921810699588, 'eval/flap_rate_model': 0.1313888888888889, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9032921810699589, 'timing/data': 0.000498262990731746, 'timing/model': 1.4321623069699854, 'learning_rate/qwen_vl_interface': 7.1727208798313704e-06, 'learning_rate/action_model': 7.172720879831371e-05, 'epoch': 0.82}) +[2026-06-15 23:14:31,757][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1900_state +[2026-06-15 23:14:31,762][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:14:37,184][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1900_state/model.safetensors +[2026-06-15 23:14:45,235][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1900_state/optimizer.bin +[2026-06-15 23:14:45,236][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1900_state/sampler.bin +[2026-06-15 23:14:45,236][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1900_state/sampler_1.bin +[2026-06-15 23:14:45,237][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_1900_state/random_states_0.pkl +[2026-06-15 23:14:45,345][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1900 +[2026-06-15 23:14:45,345][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:14:45,355][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:14:51,242][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.041440173983573914, 'train/grad_norm_pre_clip': 6.2391180992126465, 'timing/data': 0.0005974189843982458, 'timing/model': 1.47347482299665, 'learning_rate/qwen_vl_interface': 7.169935713026994e-06, 'learning_rate/action_model': 7.169935713026994e-05, 'epoch': 0.82}) +[2026-06-15 23:14:57,147][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.022320952266454697, 'train/grad_norm_pre_clip': 3.354846477508545, 'timing/data': 0.0005794729804620147, 'timing/model': 1.4297294029965997, 'learning_rate/qwen_vl_interface': 7.1671497570095684e-06, 'learning_rate/action_model': 7.167149757009568e-05, 'epoch': 0.82}) +[2026-06-15 23:15:03,079][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.4336683452129364, 'train/grad_norm_pre_clip': 8.526515007019043, 'timing/data': 0.0005007759900763631, 'timing/model': 1.707828996994067, 'learning_rate/qwen_vl_interface': 7.164363012924295e-06, 'learning_rate/action_model': 7.164363012924294e-05, 'epoch': 0.82}) +[2026-06-15 23:15:08,835][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.23214714229106903, 'train/grad_norm_pre_clip': 11.004281997680664, 'timing/data': 0.0006103480118326843, 'timing/model': 1.4641794259659946, 'learning_rate/qwen_vl_interface': 7.161575481916699e-06, 'learning_rate/action_model': 7.1615754819167e-05, 'epoch': 0.82}) +[2026-06-15 23:15:14,849][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.10598799586296082, 'train/grad_norm_pre_clip': 8.534106254577637, 'timing/data': 0.0005452720215544105, 'timing/model': 1.46055218501715, 'learning_rate/qwen_vl_interface': 7.158787165132627e-06, 'learning_rate/action_model': 7.158787165132626e-05, 'epoch': 0.82}) +[2026-06-15 23:15:20,952][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.18831013143062592, 'train/grad_norm_pre_clip': 12.482385635375977, 'timing/data': 0.0005613259854726493, 'timing/model': 1.4632209279807284, 'learning_rate/qwen_vl_interface': 7.1559980637182535e-06, 'learning_rate/action_model': 7.155998063718254e-05, 'epoch': 0.82}) +[2026-06-15 23:15:26,951][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.2155381739139557, 'train/grad_norm_pre_clip': 16.23797607421875, 'timing/data': 0.0005277570453472435, 'timing/model': 1.470629054994788, 'learning_rate/qwen_vl_interface': 7.153208178820071e-06, 'learning_rate/action_model': 7.153208178820072e-05, 'epoch': 0.82}) +[2026-06-15 23:15:32,934][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.09602658450603485, 'train/grad_norm_pre_clip': 7.349881649017334, 'timing/data': 0.0005346660036593676, 'timing/model': 1.4984363759867847, 'learning_rate/qwen_vl_interface': 7.150417511584898e-06, 'learning_rate/action_model': 7.150417511584898e-05, 'epoch': 0.82}) +[2026-06-15 23:15:38,885][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.10495931655168533, 'train/grad_norm_pre_clip': 5.871097564697266, 'timing/data': 0.0005344669916667044, 'timing/model': 1.6845618590014055, 'learning_rate/qwen_vl_interface': 7.147626063159872e-06, 'learning_rate/action_model': 7.147626063159872e-05, 'epoch': 0.82}) +[2026-06-15 23:15:44,653][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.06710994243621826, 'train/grad_norm_pre_clip': 1.8595821857452393, 'timing/data': 0.0005392640014179051, 'timing/model': 1.4667363080079667, 'learning_rate/qwen_vl_interface': 7.14483383469245e-06, 'learning_rate/action_model': 7.144833834692449e-05, 'epoch': 0.82}) +[2026-06-15 23:15:50,651][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.2119307965040207, 'train/grad_norm_pre_clip': 10.894570350646973, 'timing/data': 0.0005120229907333851, 'timing/model': 1.4648077070014551, 'learning_rate/qwen_vl_interface': 7.142040827330411e-06, 'learning_rate/action_model': 7.142040827330411e-05, 'epoch': 0.82}) +[2026-06-15 23:15:56,651][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.2100456953048706, 'train/grad_norm_pre_clip': 12.276880264282227, 'timing/data': 0.0005110920174047351, 'timing/model': 1.4703265230054967, 'learning_rate/qwen_vl_interface': 7.139247042221859e-06, 'learning_rate/action_model': 7.139247042221858e-05, 'epoch': 0.82}) +[2026-06-15 23:16:02,553][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.05762519687414169, 'train/grad_norm_pre_clip': 3.2417359352111816, 'timing/data': 0.0005130249774083495, 'timing/model': 1.4322207139921375, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.82}) +[2026-06-15 23:16:08,452][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.12466638535261154, 'train/grad_norm_pre_clip': 6.226924419403076, 'timing/data': 0.0005230390233919024, 'timing/model': 1.4398195410030894, 'learning_rate/qwen_vl_interface': 7.133657143359203e-06, 'learning_rate/action_model': 7.133657143359202e-05, 'epoch': 0.82}) +[2026-06-15 23:16:14,343][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.14045751094818115, 'train/grad_norm_pre_clip': 9.273033142089844, 'timing/data': 0.0005102099967189133, 'timing/model': 1.6558335790177807, 'learning_rate/qwen_vl_interface': 7.130861031902897e-06, 'learning_rate/action_model': 7.130861031902897e-05, 'epoch': 0.82}) +[2026-06-15 23:16:20,052][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.07708597183227539, 'train/grad_norm_pre_clip': 3.0839264392852783, 'timing/data': 0.0005044820136390626, 'timing/model': 1.4634846150293015, 'learning_rate/qwen_vl_interface': 7.128064147295667e-06, 'learning_rate/action_model': 7.128064147295667e-05, 'epoch': 0.82}) +[2026-06-15 23:16:26,052][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.08930398523807526, 'train/grad_norm_pre_clip': 3.305475950241089, 'timing/data': 0.0005182820023037493, 'timing/model': 1.4645046819932759, 'learning_rate/qwen_vl_interface': 7.125266490687206e-06, 'learning_rate/action_model': 7.125266490687206e-05, 'epoch': 0.82}) +[2026-06-15 23:16:32,054][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.18058830499649048, 'train/grad_norm_pre_clip': 5.823142051696777, 'timing/data': 0.0005210070521570742, 'timing/model': 1.4613456120132469, 'learning_rate/qwen_vl_interface': 7.1224680632275265e-06, 'learning_rate/action_model': 7.122468063227527e-05, 'epoch': 0.83}) +[2026-06-15 23:16:37,952][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.11333262920379639, 'train/grad_norm_pre_clip': 7.2499003410339355, 'timing/data': 0.0005112820072099566, 'timing/model': 1.4332593840081245, 'learning_rate/qwen_vl_interface': 7.1196688660669555e-06, 'learning_rate/action_model': 7.119668866066955e-05, 'epoch': 0.83}) +[2026-06-15 23:16:43,853][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.10070683807134628, 'train/grad_norm_pre_clip': 10.648215293884277, 'timing/data': 0.0005082180141471326, 'timing/model': 1.4368502449942753, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.83}) +[2026-06-15 23:16:49,764][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.178909569978714, 'train/grad_norm_pre_clip': 5.912815093994141, 'timing/data': 0.0005198050057515502, 'timing/model': 1.674155886983499, 'learning_rate/qwen_vl_interface': 7.114068167246033e-06, 'learning_rate/action_model': 7.114068167246033e-05, 'epoch': 0.83}) +[2026-06-15 23:16:55,457][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.1718418300151825, 'train/grad_norm_pre_clip': 6.05015754699707, 'timing/data': 0.0005375100299715996, 'timing/model': 1.4660058970330283, 'learning_rate/qwen_vl_interface': 7.111266667887916e-06, 'learning_rate/action_model': 7.111266667887915e-05, 'epoch': 0.83}) +[2026-06-15 23:17:01,456][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.17678305506706238, 'train/grad_norm_pre_clip': 6.546557903289795, 'timing/data': 0.0005037109949626029, 'timing/model': 1.466956262011081, 'learning_rate/qwen_vl_interface': 7.108464403433378e-06, 'learning_rate/action_model': 7.108464403433378e-05, 'epoch': 0.83}) +[2026-06-15 23:17:07,353][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.22213049232959747, 'train/grad_norm_pre_clip': 9.92753791809082, 'timing/data': 0.0005286380182951689, 'timing/model': 1.4597913479665294, 'learning_rate/qwen_vl_interface': 7.105661375034324e-06, 'learning_rate/action_model': 7.105661375034324e-05, 'epoch': 0.83}) +[2026-06-15 23:17:13,256][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.11504820734262466, 'train/grad_norm_pre_clip': 6.770294666290283, 'timing/data': 0.0005123239825479686, 'timing/model': 1.4339312670053914, 'learning_rate/qwen_vl_interface': 7.102857583842973e-06, 'learning_rate/action_model': 7.102857583842974e-05, 'epoch': 0.83}) +[2026-06-15 23:17:19,059][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.16629159450531006, 'train/grad_norm_pre_clip': 6.956529140472412, 'timing/data': 0.0005210860399529338, 'timing/model': 1.4672583189676516, 'learning_rate/qwen_vl_interface': 7.100053031011859e-06, 'learning_rate/action_model': 7.100053031011859e-05, 'epoch': 0.83}) +[2026-06-15 23:17:24,986][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.14283637702465057, 'train/grad_norm_pre_clip': 6.146427631378174, 'timing/data': 0.0005253120325505733, 'timing/model': 1.6956506739952601, 'learning_rate/qwen_vl_interface': 7.097247717693827e-06, 'learning_rate/action_model': 7.097247717693827e-05, 'epoch': 0.83}) +[2026-06-15 23:17:30,724][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.1574333906173706, 'train/grad_norm_pre_clip': 4.309828758239746, 'timing/data': 0.0004939470090903342, 'timing/model': 1.5346857309923507, 'learning_rate/qwen_vl_interface': 7.094441645042033e-06, 'learning_rate/action_model': 7.094441645042032e-05, 'epoch': 0.83}) +[2026-06-15 23:17:36,661][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.09464696794748306, 'train/grad_norm_pre_clip': 6.377182483673096, 'timing/data': 0.0006642179796472192, 'timing/model': 1.468682786973659, 'learning_rate/qwen_vl_interface': 7.09163481420995e-06, 'learning_rate/action_model': 7.09163481420995e-05, 'epoch': 0.83}) +[2026-06-15 23:17:42,555][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.17086505889892578, 'train/grad_norm_pre_clip': 7.645333766937256, 'timing/data': 0.0005046029691584408, 'timing/model': 1.4661647290340625, 'learning_rate/qwen_vl_interface': 7.0888272263513595e-06, 'learning_rate/action_model': 7.088827226351359e-05, 'epoch': 0.83}) +[2026-06-15 23:17:48,533][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.11310631036758423, 'train/grad_norm_pre_clip': 7.429628849029541, 'timing/data': 0.0005149479839019477, 'timing/model': 1.5043253430048935, 'learning_rate/qwen_vl_interface': 7.086018882620354e-06, 'learning_rate/action_model': 7.086018882620354e-05, 'epoch': 0.83}) +[2026-06-15 23:17:54,454][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.34346431493759155, 'train/grad_norm_pre_clip': 6.9415435791015625, 'timing/data': 0.000499385001603514, 'timing/model': 1.4378460280131549, 'learning_rate/qwen_vl_interface': 7.083209784171341e-06, 'learning_rate/action_model': 7.083209784171341e-05, 'epoch': 0.83}) +[2026-06-15 23:18:00,389][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.1510673612356186, 'train/grad_norm_pre_clip': 3.7855870723724365, 'timing/data': 0.0004944279789924622, 'timing/model': 1.7012080329586752, 'learning_rate/qwen_vl_interface': 7.080399932159029e-06, 'learning_rate/action_model': 7.080399932159029e-05, 'epoch': 0.83}) +[2026-06-15 23:18:06,055][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.09323527663946152, 'train/grad_norm_pre_clip': 2.7219741344451904, 'timing/data': 0.00048515398520976305, 'timing/model': 1.4661154469940811, 'learning_rate/qwen_vl_interface': 7.077589327738448e-06, 'learning_rate/action_model': 7.077589327738447e-05, 'epoch': 0.83}) +[2026-06-15 23:18:12,055][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.05220940336585045, 'train/grad_norm_pre_clip': 4.51272439956665, 'timing/data': 0.0005020989920012653, 'timing/model': 1.4660205050022341, 'learning_rate/qwen_vl_interface': 7.0747779720649265e-06, 'learning_rate/action_model': 7.074777972064927e-05, 'epoch': 0.83}) +[2026-06-15 23:18:17,955][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.1158488467335701, 'train/grad_norm_pre_clip': 7.074548244476318, 'timing/data': 0.000493076047860086, 'timing/model': 1.4630343689932488, 'learning_rate/qwen_vl_interface': 7.071965866294108e-06, 'learning_rate/action_model': 7.071965866294108e-05, 'epoch': 0.83}) +[2026-06-15 23:18:23,856][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.10411013662815094, 'train/grad_norm_pre_clip': 6.5510029792785645, 'timing/data': 0.0005775699974037707, 'timing/model': 1.4331599429715425, 'learning_rate/qwen_vl_interface': 7.0691530115819454e-06, 'learning_rate/action_model': 7.069153011581945e-05, 'epoch': 0.83}) +[2026-06-15 23:18:29,785][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.1924559623003006, 'train/grad_norm_pre_clip': 5.395853042602539, 'timing/data': 0.0004915429744869471, 'timing/model': 1.6949867859948426, 'learning_rate/qwen_vl_interface': 7.066339409084694e-06, 'learning_rate/action_model': 7.066339409084694e-05, 'epoch': 0.83}) +[2026-06-15 23:18:35,605][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.1691054254770279, 'train/grad_norm_pre_clip': 16.72879981994629, 'timing/data': 0.0004926449619233608, 'timing/model': 1.6154794939793646, 'learning_rate/qwen_vl_interface': 7.0635250599589215e-06, 'learning_rate/action_model': 7.06352505995892e-05, 'epoch': 0.83}) +[2026-06-15 23:18:41,257][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.0442642979323864, 'train/grad_norm_pre_clip': 4.131332874298096, 'timing/data': 0.0005084080039523542, 'timing/model': 1.4639795510447584, 'learning_rate/qwen_vl_interface': 7.060709965361499e-06, 'learning_rate/action_model': 7.060709965361499e-05, 'epoch': 0.83}) +[2026-06-15 23:18:47,157][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.19068020582199097, 'train/grad_norm_pre_clip': 3.9713993072509766, 'timing/data': 0.0004984529805369675, 'timing/model': 1.4657904829946347, 'learning_rate/qwen_vl_interface': 7.057894126449606e-06, 'learning_rate/action_model': 7.057894126449606e-05, 'epoch': 0.84}) +[2026-06-15 23:18:53,056][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.10439490526914597, 'train/grad_norm_pre_clip': 4.907460689544678, 'timing/data': 0.0004966200212948024, 'timing/model': 1.4628674690029584, 'learning_rate/qwen_vl_interface': 7.055077544380728e-06, 'learning_rate/action_model': 7.055077544380728e-05, 'epoch': 0.84}) +[2026-06-15 23:18:58,956][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.029523903504014015, 'train/grad_norm_pre_clip': 1.842395544052124, 'timing/data': 0.0005085080047138035, 'timing/model': 1.4306557850213721, 'learning_rate/qwen_vl_interface': 7.0522602203126554e-06, 'learning_rate/action_model': 7.052260220312656e-05, 'epoch': 0.84}) +[2026-06-15 23:19:04,890][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.0291210375726223, 'train/grad_norm_pre_clip': 1.7615787982940674, 'timing/data': 0.0004973920295014977, 'timing/model': 1.7023492430453189, 'learning_rate/qwen_vl_interface': 7.0494421554034845e-06, 'learning_rate/action_model': 7.049442155403484e-05, 'epoch': 0.84}) +[2026-06-15 23:19:10,555][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.049473557621240616, 'train/grad_norm_pre_clip': 4.06982421875, 'timing/data': 0.0005046119913458824, 'timing/model': 1.4655208580079488, 'learning_rate/qwen_vl_interface': 7.046623350811612e-06, 'learning_rate/action_model': 7.046623350811612e-05, 'epoch': 0.84}) +[2026-06-15 23:19:16,555][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.2643783390522003, 'train/grad_norm_pre_clip': 9.043839454650879, 'timing/data': 0.0004950980073772371, 'timing/model': 1.46688840695424, 'learning_rate/qwen_vl_interface': 7.043803807695747e-06, 'learning_rate/action_model': 7.043803807695747e-05, 'epoch': 0.84}) +[2026-06-15 23:19:22,455][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.11416154354810715, 'train/grad_norm_pre_clip': 8.71442699432373, 'timing/data': 0.000494348001666367, 'timing/model': 1.4667093410389498, 'learning_rate/qwen_vl_interface': 7.040983527214893e-06, 'learning_rate/action_model': 7.040983527214893e-05, 'epoch': 0.84}) +[2026-06-15 23:19:28,354][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.05918854475021362, 'train/grad_norm_pre_clip': 3.9732139110565186, 'timing/data': 0.0004962899838574231, 'timing/model': 1.4360024809720926, 'learning_rate/qwen_vl_interface': 7.038162510528363e-06, 'learning_rate/action_model': 7.038162510528362e-05, 'epoch': 0.84}) +[2026-06-15 23:19:34,252][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.06496623158454895, 'train/grad_norm_pre_clip': 3.0890088081359863, 'timing/data': 0.0004977020435035229, 'timing/model': 1.4340848969877698, 'learning_rate/qwen_vl_interface': 7.03534075879577e-06, 'learning_rate/action_model': 7.03534075879577e-05, 'epoch': 0.84}) +[2026-06-15 23:19:40,144][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.01898549310863018, 'train/grad_norm_pre_clip': 4.9840087890625, 'timing/data': 0.000492364983074367, 'timing/model': 1.6580806269776076, 'learning_rate/qwen_vl_interface': 7.032518273177028e-06, 'learning_rate/action_model': 7.032518273177028e-05, 'epoch': 0.84}) +[2026-06-15 23:19:45,852][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.1557779610157013, 'train/grad_norm_pre_clip': 16.789518356323242, 'timing/data': 0.0004951180308125913, 'timing/model': 1.464327719004359, 'learning_rate/qwen_vl_interface': 7.029695054832357e-06, 'learning_rate/action_model': 7.029695054832357e-05, 'epoch': 0.84}) +[2026-06-15 23:19:51,851][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.02168906107544899, 'train/grad_norm_pre_clip': 2.6474978923797607, 'timing/data': 0.0004974119947291911, 'timing/model': 1.4639154709875584, 'learning_rate/qwen_vl_interface': 7.026871104922273e-06, 'learning_rate/action_model': 7.026871104922273e-05, 'epoch': 0.84}) +[2026-06-15 23:19:57,752][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.06735332310199738, 'train/grad_norm_pre_clip': 8.310230255126953, 'timing/data': 0.00048746803076937795, 'timing/model': 1.463723475986626, 'learning_rate/qwen_vl_interface': 7.024046424607596e-06, 'learning_rate/action_model': 7.024046424607595e-05, 'epoch': 0.84}) +[2026-06-15 23:20:03,654][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.03258965536952019, 'train/grad_norm_pre_clip': 2.440478801727295, 'timing/data': 0.0005589729989878833, 'timing/model': 1.430615012999624, 'learning_rate/qwen_vl_interface': 7.021221015049445e-06, 'learning_rate/action_model': 7.021221015049446e-05, 'epoch': 0.84}) +[2026-06-15 23:20:09,553][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.06228870898485184, 'train/grad_norm_pre_clip': 9.122265815734863, 'timing/data': 0.0005061950068920851, 'timing/model': 1.4309667479828931, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 0.84}) +[2026-06-15 23:20:15,465][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.023956557735800743, 'train/grad_norm_pre_clip': 3.6311540603637695, 'timing/data': 0.0004863949725404382, 'timing/model': 1.6767779319779947, 'learning_rate/qwen_vl_interface': 7.015568012848699e-06, 'learning_rate/action_model': 7.015568012848699e-05, 'epoch': 0.84}) +[2026-06-15 23:20:21,152][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.18391132354736328, 'train/grad_norm_pre_clip': 20.60208511352539, 'timing/data': 0.0004981730016879737, 'timing/model': 1.437953845015727, 'learning_rate/qwen_vl_interface': 7.01274042252984e-06, 'learning_rate/action_model': 7.01274042252984e-05, 'epoch': 0.84}) +[2026-06-15 23:20:27,152][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.13107243180274963, 'train/grad_norm_pre_clip': 9.432882308959961, 'timing/data': 0.0005042620468884706, 'timing/model': 1.4659325070097111, 'learning_rate/qwen_vl_interface': 7.009912107614978e-06, 'learning_rate/action_model': 7.009912107614978e-05, 'epoch': 0.84}) +[2026-06-15 23:20:33,054][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.21590298414230347, 'train/grad_norm_pre_clip': 17.57750129699707, 'timing/data': 0.0005001160316169262, 'timing/model': 1.4642705000005662, 'learning_rate/qwen_vl_interface': 7.0070830692667244e-06, 'learning_rate/action_model': 7.007083069266724e-05, 'epoch': 0.84}) +[2026-06-15 23:20:38,951][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.06605208665132523, 'train/grad_norm_pre_clip': 10.319265365600586, 'timing/data': 0.0004991249879822135, 'timing/model': 1.4339738030103035, 'learning_rate/qwen_vl_interface': 7.004253308647993e-06, 'learning_rate/action_model': 7.004253308647994e-05, 'epoch': 0.84}) +[2026-06-15 23:20:44,752][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.10370141267776489, 'train/grad_norm_pre_clip': 10.14881706237793, 'timing/data': 0.0004939970094710588, 'timing/model': 1.4597062910324894, 'learning_rate/qwen_vl_interface': 7.001422826921991e-06, 'learning_rate/action_model': 7.00142282692199e-05, 'epoch': 0.84}) +[2026-06-15 23:20:50,687][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.21123230457305908, 'train/grad_norm_pre_clip': 20.150554656982422, 'timing/data': 0.0005047119921073318, 'timing/model': 1.699458719987888, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 0.84}) +[2026-06-15 23:20:56,354][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.009055335074663162, 'train/grad_norm_pre_clip': 1.4006177186965942, 'timing/data': 0.0005069460021331906, 'timing/model': 1.4653543980093673, 'learning_rate/qwen_vl_interface': 6.995759704802485e-06, 'learning_rate/action_model': 6.995759704802485e-05, 'epoch': 0.84}) +[2026-06-15 23:21:02,425][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.24609561264514923, 'train/grad_norm_pre_clip': 18.58981704711914, 'timing/data': 0.0005159889697097242, 'timing/model': 1.5363559380057268, 'learning_rate/qwen_vl_interface': 6.992927066736876e-06, 'learning_rate/action_model': 6.992927066736875e-05, 'epoch': 0.85}) +[2026-06-15 23:21:08,449][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.05908549576997757, 'train/grad_norm_pre_clip': 10.969670295715332, 'timing/data': 0.0005857510259374976, 'timing/model': 1.4806475639925338, 'learning_rate/qwen_vl_interface': 6.990093712219787e-06, 'learning_rate/action_model': 6.990093712219787e-05, 'epoch': 0.85}) +[2026-06-15 23:21:14,455][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.03613538667559624, 'train/grad_norm_pre_clip': 5.71354341506958, 'timing/data': 0.0005700589972548187, 'timing/model': 1.4434703550068662, 'learning_rate/qwen_vl_interface': 6.9872596424159e-06, 'learning_rate/action_model': 6.9872596424159e-05, 'epoch': 0.85}) +[2026-06-15 23:21:20,355][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.5339752435684204, 'train/grad_norm_pre_clip': 26.8876895904541, 'timing/data': 0.0005718409665860236, 'timing/model': 1.4369847520138137, 'learning_rate/qwen_vl_interface': 6.9844248584902e-06, 'learning_rate/action_model': 6.984424858490199e-05, 'epoch': 0.85}) +[2026-06-15 23:21:26,293][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.03557128086686134, 'train/grad_norm_pre_clip': 2.420574903488159, 'timing/data': 0.0005769690033048391, 'timing/model': 1.703567713033408, 'learning_rate/qwen_vl_interface': 6.981589361607952e-06, 'learning_rate/action_model': 6.981589361607953e-05, 'epoch': 0.85}) +[2026-06-15 23:21:32,051][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.10561959445476532, 'train/grad_norm_pre_clip': 10.184701919555664, 'timing/data': 0.0005734439473599195, 'timing/model': 1.4659389070002362, 'learning_rate/qwen_vl_interface': 6.9787531529347274e-06, 'learning_rate/action_model': 6.978753152934727e-05, 'epoch': 0.85}) +[2026-06-15 23:21:38,052][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.5873355269432068, 'train/grad_norm_pre_clip': 20.111135482788086, 'timing/data': 0.0005309009575285017, 'timing/model': 1.4653506799950264, 'learning_rate/qwen_vl_interface': 6.97591623363638e-06, 'learning_rate/action_model': 6.97591623363638e-05, 'epoch': 0.85}) +[2026-06-15 23:21:43,956][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.0488281287252903, 'train/grad_norm_pre_clip': 10.575255393981934, 'timing/data': 0.0005027800216339529, 'timing/model': 1.464868368988391, 'learning_rate/qwen_vl_interface': 6.9730786048790644e-06, 'learning_rate/action_model': 6.973078604879064e-05, 'epoch': 0.85}) +[2026-06-15 23:21:49,853][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.4036136865615845, 'train/grad_norm_pre_clip': 17.80606460571289, 'timing/data': 0.0005004159756936133, 'timing/model': 1.4335851109935902, 'learning_rate/qwen_vl_interface': 6.970240267829221e-06, 'learning_rate/action_model': 6.970240267829222e-05, 'epoch': 0.85}) +[2026-06-15 23:21:55,779][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.09783726185560226, 'train/grad_norm_pre_clip': 14.741003036499023, 'timing/data': 0.0007149819866754115, 'timing/model': 1.6912238229997456, 'learning_rate/qwen_vl_interface': 6.967401223653582e-06, 'learning_rate/action_model': 6.967401223653581e-05, 'epoch': 0.85}) +[2026-06-15 23:22:01,599][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.24436312913894653, 'train/grad_norm_pre_clip': 25.61796760559082, 'timing/data': 0.0005228790105320513, 'timing/model': 1.6109052820247598, 'learning_rate/qwen_vl_interface': 6.9645614735191724e-06, 'learning_rate/action_model': 6.964561473519172e-05, 'epoch': 0.85}) +[2026-06-15 23:22:07,345][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.11113570630550385, 'train/grad_norm_pre_clip': 4.871145248413086, 'timing/data': 0.0005196050042286515, 'timing/model': 1.4584178130025975, 'learning_rate/qwen_vl_interface': 6.961721018593305e-06, 'learning_rate/action_model': 6.961721018593305e-05, 'epoch': 0.85}) +[2026-06-15 23:22:13,253][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.26180925965309143, 'train/grad_norm_pre_clip': 9.839875221252441, 'timing/data': 0.0005206759669817984, 'timing/model': 1.4643387920223176, 'learning_rate/qwen_vl_interface': 6.958879860043585e-06, 'learning_rate/action_model': 6.958879860043585e-05, 'epoch': 0.85}) +[2026-06-15 23:22:19,253][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.11428532749414444, 'train/grad_norm_pre_clip': 18.016063690185547, 'timing/data': 0.0005173020181246102, 'timing/model': 1.465557373012416, 'learning_rate/qwen_vl_interface': 6.9560379990379025e-06, 'learning_rate/action_model': 6.956037999037902e-05, 'epoch': 0.85}) +[2026-06-15 23:22:25,157][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.0428459569811821, 'train/grad_norm_pre_clip': 7.9122138023376465, 'timing/data': 0.00051795196486637, 'timing/model': 1.4317483010236174, 'learning_rate/qwen_vl_interface': 6.953195436744442e-06, 'learning_rate/action_model': 6.953195436744442e-05, 'epoch': 0.85}) +[2026-06-15 23:22:31,136][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.04588843137025833, 'train/grad_norm_pre_clip': 5.079633712768555, 'timing/data': 0.0005143670132383704, 'timing/model': 1.7332695719669573, 'learning_rate/qwen_vl_interface': 6.950352174331676e-06, 'learning_rate/action_model': 6.950352174331675e-05, 'epoch': 0.85}) +[2026-06-15 23:22:36,852][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.19514231383800507, 'train/grad_norm_pre_clip': 12.79419994354248, 'timing/data': 0.0004982730024494231, 'timing/model': 1.4618376360158436, 'learning_rate/qwen_vl_interface': 6.947508212968356e-06, 'learning_rate/action_model': 6.947508212968356e-05, 'epoch': 0.85}) +[2026-06-15 23:22:42,851][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.01846499741077423, 'train/grad_norm_pre_clip': 0.9611750841140747, 'timing/data': 0.0005095999804325402, 'timing/model': 1.4647837899974547, 'learning_rate/qwen_vl_interface': 6.9446635538235295e-06, 'learning_rate/action_model': 6.944663553823529e-05, 'epoch': 0.85}) +[2026-06-15 23:22:48,753][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.26032063364982605, 'train/grad_norm_pre_clip': 11.267240524291992, 'timing/data': 0.00048054702347144485, 'timing/model': 1.4632532259565778, 'learning_rate/qwen_vl_interface': 6.94181819806653e-06, 'learning_rate/action_model': 6.94181819806653e-05, 'epoch': 0.85}) +[2026-06-15 23:22:54,656][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.06036011502146721, 'train/grad_norm_pre_clip': 5.852465629577637, 'timing/data': 0.0005092889769002795, 'timing/model': 1.43842105101794, 'learning_rate/qwen_vl_interface': 6.938972146866975e-06, 'learning_rate/action_model': 6.938972146866974e-05, 'epoch': 0.85}) +[2026-06-15 23:23:00,555][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.24187688529491425, 'train/grad_norm_pre_clip': 11.24302864074707, 'timing/data': 0.0005126739852130413, 'timing/model': 1.4392355980235152, 'learning_rate/qwen_vl_interface': 6.936125401394769e-06, 'learning_rate/action_model': 6.936125401394768e-05, 'epoch': 0.85}) +[2026-06-15 23:23:06,543][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.08618132025003433, 'train/grad_norm_pre_clip': 12.314024925231934, 'timing/data': 0.0005085979937575758, 'timing/model': 1.652313285972923, 'learning_rate/qwen_vl_interface': 6.933277962820101e-06, 'learning_rate/action_model': 6.9332779628201e-05, 'epoch': 0.85}) +[2026-06-15 23:23:12,254][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.06584508717060089, 'train/grad_norm_pre_clip': 2.698854923248291, 'timing/data': 0.0004980330122634768, 'timing/model': 1.4663987390231341, 'learning_rate/qwen_vl_interface': 6.930429832313444e-06, 'learning_rate/action_model': 6.930429832313444e-05, 'epoch': 0.85}) +[2026-06-15 23:23:18,250][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.23891332745552063, 'train/grad_norm_pre_clip': 7.282867431640625, 'timing/data': 0.0004889190313406289, 'timing/model': 1.4612470479914919, 'learning_rate/qwen_vl_interface': 6.92758101104556e-06, 'learning_rate/action_model': 6.92758101104556e-05, 'epoch': 0.85}) +[2026-06-15 23:23:24,151][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.13635562360286713, 'train/grad_norm_pre_clip': 10.525192260742188, 'timing/data': 0.00048764701932668686, 'timing/model': 1.4671852329629473, 'learning_rate/qwen_vl_interface': 6.9247315001874885e-06, 'learning_rate/action_model': 6.924731500187488e-05, 'epoch': 0.86}) +[2026-06-15 23:23:30,151][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.145400270819664, 'train/grad_norm_pre_clip': 11.388591766357422, 'timing/data': 0.000493075989652425, 'timing/model': 1.433516796969343, 'learning_rate/qwen_vl_interface': 6.9218813009105605e-06, 'learning_rate/action_model': 6.92188130091056e-05, 'epoch': 0.86}) +[2026-06-15 23:23:36,054][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.2286754846572876, 'train/grad_norm_pre_clip': 6.107729911804199, 'timing/data': 0.0005162699962966144, 'timing/model': 1.4380298160249367, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.86}) +[2026-06-15 23:23:41,966][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.12780052423477173, 'train/grad_norm_pre_clip': 10.315749168395996, 'timing/data': 0.0005132350488565862, 'timing/model': 1.6743323919945396, 'learning_rate/qwen_vl_interface': 6.916178841786839e-06, 'learning_rate/action_model': 6.916178841786838e-05, 'epoch': 0.86}) +[2026-06-15 23:23:47,654][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.12396152317523956, 'train/grad_norm_pre_clip': 9.381396293640137, 'timing/data': 0.0004887490067631006, 'timing/model': 1.4531027890043333, 'learning_rate/qwen_vl_interface': 6.913326584284115e-06, 'learning_rate/action_model': 6.913326584284115e-05, 'epoch': 0.86}) +[2026-06-15 23:23:53,653][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.04402051866054535, 'train/grad_norm_pre_clip': 3.177488327026367, 'timing/data': 0.0004972509923391044, 'timing/model': 1.4647867580060847, 'learning_rate/qwen_vl_interface': 6.910473643050662e-06, 'learning_rate/action_model': 6.910473643050662e-05, 'epoch': 0.86}) +[2026-06-15 23:23:59,654][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.14795181155204773, 'train/grad_norm_pre_clip': 5.541773319244385, 'timing/data': 0.0005098700057715178, 'timing/model': 1.4655882770312019, 'learning_rate/qwen_vl_interface': 6.907620019259215e-06, 'learning_rate/action_model': 6.907620019259216e-05, 'epoch': 0.86}) +[2026-06-15 23:24:05,625][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.058281734585762024, 'train/grad_norm_pre_clip': 4.6278862953186035, 'timing/data': 0.0005056440131738782, 'timing/model': 1.4952353130211122, 'learning_rate/qwen_vl_interface': 6.9047657140827935e-06, 'learning_rate/action_model': 6.904765714082794e-05, 'epoch': 0.86}) +[2026-06-15 23:24:11,453][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.1854502260684967, 'train/grad_norm_pre_clip': 6.296106338500977, 'timing/data': 0.0005076169618405402, 'timing/model': 1.4626094209961593, 'learning_rate/qwen_vl_interface': 6.901910728694694e-06, 'learning_rate/action_model': 6.901910728694694e-05, 'epoch': 0.86}) +[2026-06-15 23:24:17,389][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.4677742123603821, 'train/grad_norm_pre_clip': 11.101644515991211, 'timing/data': 0.000517620996106416, 'timing/model': 1.7011431999853812, 'learning_rate/qwen_vl_interface': 6.899055064268489e-06, 'learning_rate/action_model': 6.899055064268489e-05, 'epoch': 0.86}) +[2026-06-15 23:24:23,164][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.13375554978847504, 'train/grad_norm_pre_clip': 3.7628672122955322, 'timing/data': 0.0004995650378987193, 'timing/model': 1.4900861330097541, 'learning_rate/qwen_vl_interface': 6.896198721978038e-06, 'learning_rate/action_model': 6.896198721978038e-05, 'epoch': 0.86}) +[2026-06-15 23:24:29,170][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.12056150287389755, 'train/grad_norm_pre_clip': 8.557661056518555, 'timing/data': 0.0004918840131722391, 'timing/model': 1.4851641510031186, 'learning_rate/qwen_vl_interface': 6.893341702997475e-06, 'learning_rate/action_model': 6.893341702997475e-05, 'epoch': 0.86}) +[2026-06-15 23:33:18,238][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.0332706980407238, 'train/grad_norm_pre_clip': 2.5135176181793213, 'eval/loss': 0.0725640282034874, 'eval/flap_precision': 0.9432314410480349, 'eval/flap_recall': 0.8657314629258517, 'eval/flap_f1': 0.90282131661442, 'eval/flap_rate_model': 0.1272222222222222, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9028213166144199, 'timing/data': 0.0004928450216539204, 'timing/model': 1.4744887939887121, 'learning_rate/qwen_vl_interface': 6.890484008501211e-06, 'learning_rate/action_model': 6.890484008501211e-05, 'epoch': 0.86}) +[2026-06-15 23:33:18,238][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2000_state +[2026-06-15 23:33:18,242][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:33:23,603][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2000_state/model.safetensors +[2026-06-15 23:33:31,578][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2000_state/optimizer.bin +[2026-06-15 23:33:31,579][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2000_state/sampler.bin +[2026-06-15 23:33:31,579][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-15 23:33:31,580][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-15 23:33:31,688][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-15 23:33:31,688][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:33:31,698][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:33:37,534][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.0293482206761837, 'train/grad_norm_pre_clip': 2.2143075466156006, 'timing/data': 0.0005273660062812269, 'timing/model': 1.4671852940227836, 'learning_rate/qwen_vl_interface': 6.887625639663941e-06, 'learning_rate/action_model': 6.88762563966394e-05, 'epoch': 0.86}) +[2026-06-15 23:33:43,434][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.09684084355831146, 'train/grad_norm_pre_clip': 4.5794901847839355, 'timing/data': 0.0005255939904600382, 'timing/model': 1.4651769590564072, 'learning_rate/qwen_vl_interface': 6.884766597660626e-06, 'learning_rate/action_model': 6.884766597660626e-05, 'epoch': 0.86}) +[2026-06-15 23:33:49,337][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.2251603752374649, 'train/grad_norm_pre_clip': 8.207741737365723, 'timing/data': 0.0005092889769002795, 'timing/model': 1.4308082069619559, 'learning_rate/qwen_vl_interface': 6.881906883666515e-06, 'learning_rate/action_model': 6.881906883666515e-05, 'epoch': 0.86}) +[2026-06-15 23:33:55,267][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.17696456611156464, 'train/grad_norm_pre_clip': 7.695741653442383, 'timing/data': 0.0005173310055397451, 'timing/model': 1.6856414930080064, 'learning_rate/qwen_vl_interface': 6.879046498857125e-06, 'learning_rate/action_model': 6.879046498857124e-05, 'epoch': 0.86}) +[2026-06-15 23:34:01,041][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.216094970703125, 'train/grad_norm_pre_clip': 7.212331771850586, 'timing/data': 0.0005186529597267509, 'timing/model': 1.448267120984383, 'learning_rate/qwen_vl_interface': 6.876185444408255e-06, 'learning_rate/action_model': 6.876185444408255e-05, 'epoch': 0.86}) +[2026-06-15 23:34:07,348][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.19271570444107056, 'train/grad_norm_pre_clip': 8.24751091003418, 'timing/data': 0.0005117029650136828, 'timing/model': 1.432267813012004, 'learning_rate/qwen_vl_interface': 6.873323721495976e-06, 'learning_rate/action_model': 6.873323721495976e-05, 'epoch': 0.86}) +[2026-06-15 23:34:13,145][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.19505830109119415, 'train/grad_norm_pre_clip': 8.017861366271973, 'timing/data': 0.0005204449989832938, 'timing/model': 1.4620893459650688, 'learning_rate/qwen_vl_interface': 6.870461331296632e-06, 'learning_rate/action_model': 6.870461331296632e-05, 'epoch': 0.86}) +[2026-06-15 23:34:19,049][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.05084102600812912, 'train/grad_norm_pre_clip': 3.9431324005126953, 'timing/data': 0.0005056240479461849, 'timing/model': 1.4588842900120653, 'learning_rate/qwen_vl_interface': 6.8675982749868455e-06, 'learning_rate/action_model': 6.867598274986845e-05, 'epoch': 0.86}) +[2026-06-15 23:34:24,949][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.05678631737828255, 'train/grad_norm_pre_clip': 3.7267162799835205, 'timing/data': 0.0005163190071471035, 'timing/model': 1.43621837295359, 'learning_rate/qwen_vl_interface': 6.864734553743509e-06, 'learning_rate/action_model': 6.86473455374351e-05, 'epoch': 0.86}) +[2026-06-15 23:34:30,885][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.08849921077489853, 'train/grad_norm_pre_clip': 5.686526775360107, 'timing/data': 0.000517170992679894, 'timing/model': 1.6977919960045256, 'learning_rate/qwen_vl_interface': 6.861870168743793e-06, 'learning_rate/action_model': 6.861870168743793e-05, 'epoch': 0.86}) +[2026-06-15 23:34:36,548][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.07880529016256332, 'train/grad_norm_pre_clip': 6.417171478271484, 'timing/data': 0.0005009369924664497, 'timing/model': 1.4619414159678854, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 0.87}) +[2026-06-15 23:34:42,549][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.13904251158237457, 'train/grad_norm_pre_clip': 5.7422871589660645, 'timing/data': 0.00048692699056118727, 'timing/model': 1.463906618009787, 'learning_rate/qwen_vl_interface': 6.85613941218525e-06, 'learning_rate/action_model': 6.85613941218525e-05, 'epoch': 0.87}) +[2026-06-15 23:34:48,450][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.16528593003749847, 'train/grad_norm_pre_clip': 6.5714945793151855, 'timing/data': 0.0005004260456189513, 'timing/model': 1.4636237950180657, 'learning_rate/qwen_vl_interface': 6.8532730429821215e-06, 'learning_rate/action_model': 6.853273042982122e-05, 'epoch': 0.87}) +[2026-06-15 23:34:54,249][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.19714666903018951, 'train/grad_norm_pre_clip': 6.927742958068848, 'timing/data': 0.0005011869943700731, 'timing/model': 1.4582134379888885, 'learning_rate/qwen_vl_interface': 6.850406014734005e-06, 'learning_rate/action_model': 6.850406014734005e-05, 'epoch': 0.87}) +[2026-06-15 23:35:00,152][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.18078280985355377, 'train/grad_norm_pre_clip': 7.316200256347656, 'timing/data': 0.0004990749876014888, 'timing/model': 1.4899278599768877, 'learning_rate/qwen_vl_interface': 6.8475383286194295e-06, 'learning_rate/action_model': 6.84753832861943e-05, 'epoch': 0.87}) +[2026-06-15 23:35:06,050][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.0907161608338356, 'train/grad_norm_pre_clip': 4.64927864074707, 'timing/data': 0.0005001759855076671, 'timing/model': 1.6630286290310323, 'learning_rate/qwen_vl_interface': 6.844669985817192e-06, 'learning_rate/action_model': 6.844669985817192e-05, 'epoch': 0.87}) +[2026-06-15 23:35:11,747][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.0928637832403183, 'train/grad_norm_pre_clip': 6.9351396560668945, 'timing/data': 0.0004962700186297297, 'timing/model': 1.4621288380003534, 'learning_rate/qwen_vl_interface': 6.8418009875063594e-06, 'learning_rate/action_model': 6.84180098750636e-05, 'epoch': 0.87}) +[2026-06-15 23:35:17,748][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.1992853730916977, 'train/grad_norm_pre_clip': 5.424145698547363, 'timing/data': 0.0004973819595761597, 'timing/model': 1.4649750660173595, 'learning_rate/qwen_vl_interface': 6.838931334866267e-06, 'learning_rate/action_model': 6.838931334866266e-05, 'epoch': 0.87}) +[2026-06-15 23:35:23,648][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.07650374621152878, 'train/grad_norm_pre_clip': 4.58626127243042, 'timing/data': 0.0005017489893361926, 'timing/model': 1.4644016409874894, 'learning_rate/qwen_vl_interface': 6.836061029076523e-06, 'learning_rate/action_model': 6.836061029076522e-05, 'epoch': 0.87}) +[2026-06-15 23:35:29,558][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.1434164196252823, 'train/grad_norm_pre_clip': 6.440880298614502, 'timing/data': 0.0005205060006119311, 'timing/model': 1.4305721650016494, 'learning_rate/qwen_vl_interface': 6.833190071317e-06, 'learning_rate/action_model': 6.833190071316999e-05, 'epoch': 0.87}) +[2026-06-15 23:35:35,480][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.35067319869995117, 'train/grad_norm_pre_clip': 10.017083168029785, 'timing/data': 0.0005171810043975711, 'timing/model': 1.6910536370123737, 'learning_rate/qwen_vl_interface': 6.830318462767841e-06, 'learning_rate/action_model': 6.830318462767841e-05, 'epoch': 0.87}) +[2026-06-15 23:35:41,248][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.05490913242101669, 'train/grad_norm_pre_clip': 2.869295835494995, 'timing/data': 0.0005286279483698308, 'timing/model': 1.5634387040045112, 'learning_rate/qwen_vl_interface': 6.827446204609456e-06, 'learning_rate/action_model': 6.827446204609456e-05, 'epoch': 0.87}) +[2026-06-15 23:35:46,951][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.14412912726402283, 'train/grad_norm_pre_clip': 5.955697059631348, 'timing/data': 0.0005100109847262502, 'timing/model': 1.4652288859942928, 'learning_rate/qwen_vl_interface': 6.824573298022522e-06, 'learning_rate/action_model': 6.824573298022522e-05, 'epoch': 0.87}) +[2026-06-15 23:35:52,851][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.29663699865341187, 'train/grad_norm_pre_clip': 11.61303997039795, 'timing/data': 0.000508358993101865, 'timing/model': 1.464166602003388, 'learning_rate/qwen_vl_interface': 6.8216997441879805e-06, 'learning_rate/action_model': 6.821699744187981e-05, 'epoch': 0.87}) +[2026-06-15 23:35:58,750][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.1247372031211853, 'train/grad_norm_pre_clip': 4.478562831878662, 'timing/data': 0.0005075270310044289, 'timing/model': 1.464294378994964, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 0.87}) +[2026-06-15 23:36:04,657][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.1340254247188568, 'train/grad_norm_pre_clip': 5.169595241546631, 'timing/data': 0.0005111029604449868, 'timing/model': 1.4327638490358368, 'learning_rate/qwen_vl_interface': 6.815950699501186e-06, 'learning_rate/action_model': 6.815950699501186e-05, 'epoch': 0.87}) +[2026-06-15 23:36:10,585][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.1885477602481842, 'train/grad_norm_pre_clip': 7.65651798248291, 'timing/data': 0.0005290789995342493, 'timing/model': 1.695638373028487, 'learning_rate/qwen_vl_interface': 6.813075211012148e-06, 'learning_rate/action_model': 6.813075211012147e-05, 'epoch': 0.87}) +[2026-06-15 23:36:16,255][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.05235693231225014, 'train/grad_norm_pre_clip': 6.698172092437744, 'timing/data': 0.0005091190105304122, 'timing/model': 1.4660218710196204, 'learning_rate/qwen_vl_interface': 6.81019908000193e-06, 'learning_rate/action_model': 6.81019908000193e-05, 'epoch': 0.87}) +[2026-06-15 23:36:22,372][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.11197363585233688, 'train/grad_norm_pre_clip': 8.584614753723145, 'timing/data': 0.0006818940164521337, 'timing/model': 1.484097347012721, 'learning_rate/qwen_vl_interface': 6.807322307652806e-06, 'learning_rate/action_model': 6.807322307652806e-05, 'epoch': 0.87}) +[2026-06-15 23:36:28,375][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.1089530885219574, 'train/grad_norm_pre_clip': 7.56932258605957, 'timing/data': 0.0008266560034826398, 'timing/model': 1.4668511120253243, 'learning_rate/qwen_vl_interface': 6.804444895147306e-06, 'learning_rate/action_model': 6.804444895147306e-05, 'epoch': 0.87}) +[2026-06-15 23:36:34,571][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.03662315011024475, 'train/grad_norm_pre_clip': 2.1614089012145996, 'timing/data': 0.0009701679809950292, 'timing/model': 1.4637564069707878, 'learning_rate/qwen_vl_interface': 6.8015668436682255e-06, 'learning_rate/action_model': 6.801566843668226e-05, 'epoch': 0.87}) +[2026-06-15 23:36:40,770][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.13066387176513672, 'train/grad_norm_pre_clip': 4.728222370147705, 'timing/data': 0.0008620190201327205, 'timing/model': 1.4520959740038961, 'learning_rate/qwen_vl_interface': 6.7986881543986245e-06, 'learning_rate/action_model': 6.798688154398624e-05, 'epoch': 0.87}) +[2026-06-15 23:36:46,975][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.11958633363246918, 'train/grad_norm_pre_clip': 3.9394493103027344, 'timing/data': 0.0010191900073550642, 'timing/model': 1.867637632007245, 'learning_rate/qwen_vl_interface': 6.795808828521822e-06, 'learning_rate/action_model': 6.795808828521821e-05, 'epoch': 0.87}) +[2026-06-15 23:36:52,774][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.19001013040542603, 'train/grad_norm_pre_clip': 6.697893142700195, 'timing/data': 0.0008305020164698362, 'timing/model': 1.4701014870079234, 'learning_rate/qwen_vl_interface': 6.792928867221401e-06, 'learning_rate/action_model': 6.792928867221401e-05, 'epoch': 0.88}) +[2026-06-15 23:36:58,972][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.13864782452583313, 'train/grad_norm_pre_clip': 8.169107437133789, 'timing/data': 0.0009509299998171628, 'timing/model': 1.4652211939683184, 'learning_rate/qwen_vl_interface': 6.790048271681202e-06, 'learning_rate/action_model': 6.790048271681202e-05, 'epoch': 0.88}) +[2026-06-15 23:37:04,970][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.13931842148303986, 'train/grad_norm_pre_clip': 4.164287090301514, 'timing/data': 0.000881978019606322, 'timing/model': 1.4638988979859278, 'learning_rate/qwen_vl_interface': 6.7871670430853335e-06, 'learning_rate/action_model': 6.787167043085333e-05, 'epoch': 0.88}) +[2026-06-15 23:37:11,072][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.141699880361557, 'train/grad_norm_pre_clip': 11.48022174835205, 'timing/data': 0.0008194660185836256, 'timing/model': 1.4654806169564836, 'learning_rate/qwen_vl_interface': 6.784285182618156e-06, 'learning_rate/action_model': 6.784285182618156e-05, 'epoch': 0.88}) +[2026-06-15 23:37:17,172][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.09336014091968536, 'train/grad_norm_pre_clip': 5.643789768218994, 'timing/data': 0.0008238030131906271, 'timing/model': 1.469346828002017, 'learning_rate/qwen_vl_interface': 6.781402691464297e-06, 'learning_rate/action_model': 6.781402691464297e-05, 'epoch': 0.88}) +[2026-06-15 23:37:23,360][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.28933554887771606, 'train/grad_norm_pre_clip': 16.01653289794922, 'timing/data': 0.0008322750218212605, 'timing/model': 1.8521418839809485, 'learning_rate/qwen_vl_interface': 6.778519570808636e-06, 'learning_rate/action_model': 6.778519570808636e-05, 'epoch': 0.88}) +[2026-06-15 23:37:29,408][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.07734192162752151, 'train/grad_norm_pre_clip': 4.0162858963012695, 'timing/data': 0.0008085300214588642, 'timing/model': 1.6960190079989843, 'learning_rate/qwen_vl_interface': 6.775635821836317e-06, 'learning_rate/action_model': 6.775635821836317e-05, 'epoch': 0.88}) +[2026-06-15 23:37:35,271][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.06487064808607101, 'train/grad_norm_pre_clip': 5.65215539932251, 'timing/data': 0.0008555100066587329, 'timing/model': 1.4793804789660498, 'learning_rate/qwen_vl_interface': 6.77275144573274e-06, 'learning_rate/action_model': 6.77275144573274e-05, 'epoch': 0.88}) +[2026-06-15 23:37:41,371][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.3342011570930481, 'train/grad_norm_pre_clip': 7.030769348144531, 'timing/data': 0.0006980070029385388, 'timing/model': 1.4641994159901515, 'learning_rate/qwen_vl_interface': 6.769866443683564e-06, 'learning_rate/action_model': 6.769866443683563e-05, 'epoch': 0.88}) +[2026-06-15 23:37:47,572][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.14506939053535461, 'train/grad_norm_pre_clip': 5.655924320220947, 'timing/data': 0.0007039259653538465, 'timing/model': 1.4650562590104528, 'learning_rate/qwen_vl_interface': 6.766980816874701e-06, 'learning_rate/action_model': 6.7669808168747e-05, 'epoch': 0.88}) +[2026-06-15 23:37:53,772][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.2755555510520935, 'train/grad_norm_pre_clip': 7.023799419403076, 'timing/data': 0.000964139006100595, 'timing/model': 1.4655544629786164, 'learning_rate/qwen_vl_interface': 6.764094566492327e-06, 'learning_rate/action_model': 6.764094566492328e-05, 'epoch': 0.88}) +[2026-06-15 23:38:00,063][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.1138436570763588, 'train/grad_norm_pre_clip': 7.379590034484863, 'timing/data': 0.0008063970017246902, 'timing/model': 1.855710178031586, 'learning_rate/qwen_vl_interface': 6.7612076937228685e-06, 'learning_rate/action_model': 6.761207693722869e-05, 'epoch': 0.88}) +[2026-06-15 23:38:05,871][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.12692220509052277, 'train/grad_norm_pre_clip': 3.3080029487609863, 'timing/data': 0.0009539040038362145, 'timing/model': 1.4644948469940573, 'learning_rate/qwen_vl_interface': 6.7583201997530126e-06, 'learning_rate/action_model': 6.758320199753012e-05, 'epoch': 0.88}) +[2026-06-15 23:38:12,058][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.14423291385173798, 'train/grad_norm_pre_clip': 6.439336776733398, 'timing/data': 0.0008247739751823246, 'timing/model': 1.551972676999867, 'learning_rate/qwen_vl_interface': 6.755432085769693e-06, 'learning_rate/action_model': 6.755432085769694e-05, 'epoch': 0.88}) +[2026-06-15 23:38:18,069][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.08842387795448303, 'train/grad_norm_pre_clip': 3.068640947341919, 'timing/data': 0.0008249940001405776, 'timing/model': 1.4620272890315391, 'learning_rate/qwen_vl_interface': 6.75254335296011e-06, 'learning_rate/action_model': 6.75254335296011e-05, 'epoch': 0.88}) +[2026-06-15 23:38:24,169][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.0846816897392273, 'train/grad_norm_pre_clip': 5.113087177276611, 'timing/data': 0.000951350957620889, 'timing/model': 1.4588862780365162, 'learning_rate/qwen_vl_interface': 6.749654002511712e-06, 'learning_rate/action_model': 6.749654002511712e-05, 'epoch': 0.88}) +[2026-06-15 23:38:30,269][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.03687269613146782, 'train/grad_norm_pre_clip': 2.0667452812194824, 'timing/data': 0.0006859700079075992, 'timing/model': 1.4905626800027676, 'learning_rate/qwen_vl_interface': 6.7467640356122e-06, 'learning_rate/action_model': 6.746764035612199e-05, 'epoch': 0.88}) +[2026-06-15 23:38:36,361][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.13400761783123016, 'train/grad_norm_pre_clip': 7.268151760101318, 'timing/data': 0.0008260260219685733, 'timing/model': 1.858520862995647, 'learning_rate/qwen_vl_interface': 6.74387345344953e-06, 'learning_rate/action_model': 6.74387345344953e-05, 'epoch': 0.88}) +[2026-06-15 23:38:42,070][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.09879287332296371, 'train/grad_norm_pre_clip': 5.116000175476074, 'timing/data': 0.00082180998288095, 'timing/model': 1.4613638509763405, 'learning_rate/qwen_vl_interface': 6.740982257211911e-06, 'learning_rate/action_model': 6.740982257211911e-05, 'epoch': 0.88}) +[2026-06-15 23:38:48,170][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.11289530992507935, 'train/grad_norm_pre_clip': 3.8807334899902344, 'timing/data': 0.000952892005443573, 'timing/model': 1.4649198059923947, 'learning_rate/qwen_vl_interface': 6.738090448087806e-06, 'learning_rate/action_model': 6.738090448087806e-05, 'epoch': 0.88}) +[2026-06-15 23:38:54,369][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.023043200373649597, 'train/grad_norm_pre_clip': 1.3517193794250488, 'timing/data': 0.0009820449631661177, 'timing/model': 1.4626008360064588, 'learning_rate/qwen_vl_interface': 6.735198027265928e-06, 'learning_rate/action_model': 6.735198027265928e-05, 'epoch': 0.88}) +[2026-06-15 23:39:00,574][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.23060189187526703, 'train/grad_norm_pre_clip': 11.374176025390625, 'timing/data': 0.0008313030120916665, 'timing/model': 1.4669921210152097, 'learning_rate/qwen_vl_interface': 6.7323049959352424e-06, 'learning_rate/action_model': 6.732304995935243e-05, 'epoch': 0.88}) +[2026-06-15 23:39:06,774][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.20967935025691986, 'train/grad_norm_pre_clip': 7.458044528961182, 'timing/data': 0.0008243529591709375, 'timing/model': 1.4613885160069913, 'learning_rate/qwen_vl_interface': 6.729411355284962e-06, 'learning_rate/action_model': 6.729411355284963e-05, 'epoch': 0.88}) +[2026-06-15 23:39:12,969][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.13215821981430054, 'train/grad_norm_pre_clip': 7.217673301696777, 'timing/data': 0.0008043330162763596, 'timing/model': 1.8607043490046635, 'learning_rate/qwen_vl_interface': 6.726517106504559e-06, 'learning_rate/action_model': 6.726517106504558e-05, 'epoch': 0.89}) +[2026-06-15 23:39:18,774][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.13081000745296478, 'train/grad_norm_pre_clip': 9.76158332824707, 'timing/data': 0.0008170920191332698, 'timing/model': 1.4633222779957578, 'learning_rate/qwen_vl_interface': 6.723622250783745e-06, 'learning_rate/action_model': 6.723622250783744e-05, 'epoch': 0.89}) +[2026-06-15 23:39:24,873][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.11526276171207428, 'train/grad_norm_pre_clip': 8.735249519348145, 'timing/data': 0.0006639069761149585, 'timing/model': 1.4585309600224718, 'learning_rate/qwen_vl_interface': 6.720726789312487e-06, 'learning_rate/action_model': 6.720726789312487e-05, 'epoch': 0.89}) +[2026-06-15 23:39:30,873][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.22511379420757294, 'train/grad_norm_pre_clip': 13.523544311523438, 'timing/data': 0.0008185349870473146, 'timing/model': 1.4631178330164403, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.89}) +[2026-06-15 23:39:36,973][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.1764785200357437, 'train/grad_norm_pre_clip': 9.264083862304688, 'timing/data': 0.0008202469907701015, 'timing/model': 1.4645817399723455, 'learning_rate/qwen_vl_interface': 6.714934053879749e-06, 'learning_rate/action_model': 6.71493405387975e-05, 'epoch': 0.89}) +[2026-06-15 23:39:43,077][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.025995703414082527, 'train/grad_norm_pre_clip': 3.683654308319092, 'timing/data': 0.0008713319548405707, 'timing/model': 1.4517439299961552, 'learning_rate/qwen_vl_interface': 6.712036782299442e-06, 'learning_rate/action_model': 6.712036782299443e-05, 'epoch': 0.89}) +[2026-06-15 23:39:49,205][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.058803752064704895, 'train/grad_norm_pre_clip': 6.277528285980225, 'timing/data': 0.0008055350044742227, 'timing/model': 1.7972663849941455, 'learning_rate/qwen_vl_interface': 6.709138909731041e-06, 'learning_rate/action_model': 6.709138909731041e-05, 'epoch': 0.89}) +[2026-06-15 23:39:54,973][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.2059239149093628, 'train/grad_norm_pre_clip': 13.388545036315918, 'timing/data': 0.0008194459951482713, 'timing/model': 1.4631708440138027, 'learning_rate/qwen_vl_interface': 6.706240437365751e-06, 'learning_rate/action_model': 6.70624043736575e-05, 'epoch': 0.89}) +[2026-06-15 23:40:01,072][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.1299462765455246, 'train/grad_norm_pre_clip': 7.955031394958496, 'timing/data': 0.000839676009491086, 'timing/model': 1.4631967949680984, 'learning_rate/qwen_vl_interface': 6.703341366395023e-06, 'learning_rate/action_model': 6.703341366395023e-05, 'epoch': 0.89}) +[2026-06-15 23:40:07,269][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.19792868196964264, 'train/grad_norm_pre_clip': 12.708612442016602, 'timing/data': 0.0008313140133395791, 'timing/model': 1.4642706390004605, 'learning_rate/qwen_vl_interface': 6.700441698010558e-06, 'learning_rate/action_model': 6.700441698010558e-05, 'epoch': 0.89}) +[2026-06-15 23:40:13,472][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.11696706712245941, 'train/grad_norm_pre_clip': 10.575288772583008, 'timing/data': 0.0006800609990023077, 'timing/model': 1.4750919479993172, 'learning_rate/qwen_vl_interface': 6.697541433404299e-06, 'learning_rate/action_model': 6.697541433404299e-05, 'epoch': 0.89}) +[2026-06-15 23:40:19,661][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.25680065155029297, 'train/grad_norm_pre_clip': 7.431396484375, 'timing/data': 0.0006778980023227632, 'timing/model': 1.8516445110435598, 'learning_rate/qwen_vl_interface': 6.6946405737684325e-06, 'learning_rate/action_model': 6.694640573768433e-05, 'epoch': 0.89}) +[2026-06-15 23:40:25,817][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.24251078069210052, 'train/grad_norm_pre_clip': 12.596872329711914, 'timing/data': 0.0006811119965277612, 'timing/model': 1.8093652020324953, 'learning_rate/qwen_vl_interface': 6.691739120295397e-06, 'learning_rate/action_model': 6.691739120295397e-05, 'epoch': 0.89}) +[2026-06-15 23:40:31,573][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.07113433629274368, 'train/grad_norm_pre_clip': 7.130914688110352, 'timing/data': 0.0009710289887152612, 'timing/model': 1.460332827991806, 'learning_rate/qwen_vl_interface': 6.688837074177868e-06, 'learning_rate/action_model': 6.688837074177868e-05, 'epoch': 0.89}) +[2026-06-15 23:40:37,673][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.0389956571161747, 'train/grad_norm_pre_clip': 6.75609016418457, 'timing/data': 0.0008339670021086931, 'timing/model': 1.4654190670116805, 'learning_rate/qwen_vl_interface': 6.68593443660877e-06, 'learning_rate/action_model': 6.685934436608769e-05, 'epoch': 0.89}) +[2026-06-15 23:40:43,874][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.07927710562944412, 'train/grad_norm_pre_clip': 12.20193099975586, 'timing/data': 0.0008275780128315091, 'timing/model': 1.4651282430277206, 'learning_rate/qwen_vl_interface': 6.6830312087812625e-06, 'learning_rate/action_model': 6.683031208781262e-05, 'epoch': 0.89}) +[2026-06-15 23:40:50,068][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.08613931387662888, 'train/grad_norm_pre_clip': 6.327871799468994, 'timing/data': 0.0006778679671697319, 'timing/model': 1.4653351610177197, 'learning_rate/qwen_vl_interface': 6.680127391888754e-06, 'learning_rate/action_model': 6.680127391888754e-05, 'epoch': 0.89}) +[2026-06-15 23:40:56,255][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.038661520928144455, 'train/grad_norm_pre_clip': 3.502145290374756, 'timing/data': 0.0009449900244362652, 'timing/model': 1.8446998540312052, 'learning_rate/qwen_vl_interface': 6.677222987124897e-06, 'learning_rate/action_model': 6.677222987124897e-05, 'epoch': 0.89}) +[2026-06-15 23:41:02,340][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.20121411979198456, 'train/grad_norm_pre_clip': 7.702558517456055, 'timing/data': 0.0008119449485093355, 'timing/model': 1.7342064499971457, 'learning_rate/qwen_vl_interface': 6.67431799568358e-06, 'learning_rate/action_model': 6.67431799568358e-05, 'epoch': 0.89}) +[2026-06-15 23:41:08,158][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.08619558811187744, 'train/grad_norm_pre_clip': 5.129056453704834, 'timing/data': 0.0005255439900793135, 'timing/model': 1.4567940440028906, 'learning_rate/qwen_vl_interface': 6.671412418758939e-06, 'learning_rate/action_model': 6.671412418758938e-05, 'epoch': 0.89}) +[2026-06-15 23:41:14,271][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.06787414848804474, 'train/grad_norm_pre_clip': 9.263138771057129, 'timing/data': 0.0008083099965006113, 'timing/model': 1.4635455540264957, 'learning_rate/qwen_vl_interface': 6.668506257545341e-06, 'learning_rate/action_model': 6.668506257545341e-05, 'epoch': 0.89}) +[2026-06-15 23:41:20,472][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.1018487960100174, 'train/grad_norm_pre_clip': 10.472016334533691, 'timing/data': 0.00081165402662009, 'timing/model': 1.4647520540165715, 'learning_rate/qwen_vl_interface': 6.665599513237403e-06, 'learning_rate/action_model': 6.665599513237402e-05, 'epoch': 0.89}) +[2026-06-15 23:41:26,672][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.16551831364631653, 'train/grad_norm_pre_clip': 5.275552272796631, 'timing/data': 0.0009512600372545421, 'timing/model': 1.4767330899485387, 'learning_rate/qwen_vl_interface': 6.662692187029978e-06, 'learning_rate/action_model': 6.662692187029978e-05, 'epoch': 0.89}) +[2026-06-15 23:41:32,850][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.14340347051620483, 'train/grad_norm_pre_clip': 11.127890586853027, 'timing/data': 0.0008073379867710173, 'timing/model': 1.84185276000062, 'learning_rate/qwen_vl_interface': 6.659784280118156e-06, 'learning_rate/action_model': 6.659784280118156e-05, 'epoch': 0.89}) +[2026-06-15 23:41:38,668][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.20716625452041626, 'train/grad_norm_pre_clip': 7.096310138702393, 'timing/data': 0.0006722900434397161, 'timing/model': 1.4604052909999155, 'learning_rate/qwen_vl_interface': 6.65687579369727e-06, 'learning_rate/action_model': 6.656875793697271e-05, 'epoch': 0.9}) +[2026-06-15 23:41:44,772][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.19991633296012878, 'train/grad_norm_pre_clip': 8.652884483337402, 'timing/data': 0.0006736110080964863, 'timing/model': 1.4641339020454325, 'learning_rate/qwen_vl_interface': 6.653966728962885e-06, 'learning_rate/action_model': 6.653966728962885e-05, 'epoch': 0.9}) +[2026-06-15 23:41:50,870][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.026121504604816437, 'train/grad_norm_pre_clip': 2.1817705631256104, 'timing/data': 0.0008056560182012618, 'timing/model': 1.4627924829837866, 'learning_rate/qwen_vl_interface': 6.651057087110813e-06, 'learning_rate/action_model': 6.651057087110813e-05, 'epoch': 0.9}) +[2026-06-15 23:41:56,953][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.23632606863975525, 'train/grad_norm_pre_clip': 10.955337524414062, 'timing/data': 0.0009228879935108125, 'timing/model': 1.507418128021527, 'learning_rate/qwen_vl_interface': 6.6481468693370924e-06, 'learning_rate/action_model': 6.648146869337091e-05, 'epoch': 0.9}) +[2026-06-15 23:42:03,018][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.04647951200604439, 'train/grad_norm_pre_clip': 4.17061185836792, 'timing/data': 0.0006113999988883734, 'timing/model': 1.5436878379550762, 'learning_rate/qwen_vl_interface': 6.645236076838009e-06, 'learning_rate/action_model': 6.645236076838009e-05, 'epoch': 0.9}) +[2026-06-15 23:42:09,073][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.08986598253250122, 'train/grad_norm_pre_clip': 15.159908294677734, 'timing/data': 0.0006769459578208625, 'timing/model': 1.7849108849768527, 'learning_rate/qwen_vl_interface': 6.642324710810075e-06, 'learning_rate/action_model': 6.642324710810076e-05, 'epoch': 0.9}) +[2026-06-15 23:42:14,756][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.02190200239419937, 'train/grad_norm_pre_clip': 1.6080560684204102, 'timing/data': 0.0005061250412836671, 'timing/model': 1.466091487032827, 'learning_rate/qwen_vl_interface': 6.639412772450047e-06, 'learning_rate/action_model': 6.639412772450047e-05, 'epoch': 0.9}) +[2026-06-15 23:42:20,753][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.07306730002164841, 'train/grad_norm_pre_clip': 5.683276176452637, 'timing/data': 0.0005380319780670106, 'timing/model': 1.4656710529816337, 'learning_rate/qwen_vl_interface': 6.6365002629549104e-06, 'learning_rate/action_model': 6.63650026295491e-05, 'epoch': 0.9}) +[2026-06-15 23:42:26,671][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.20661088824272156, 'train/grad_norm_pre_clip': 19.49849510192871, 'timing/data': 0.0005271859699860215, 'timing/model': 1.4710125990095548, 'learning_rate/qwen_vl_interface': 6.633587183521887e-06, 'learning_rate/action_model': 6.633587183521887e-05, 'epoch': 0.9}) +[2026-06-15 23:42:32,455][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.03797508031129837, 'train/grad_norm_pre_clip': 6.32890510559082, 'timing/data': 0.0005344469682313502, 'timing/model': 1.4610281009809114, 'learning_rate/qwen_vl_interface': 6.6306735353484365e-06, 'learning_rate/action_model': 6.630673535348436e-05, 'epoch': 0.9}) +[2026-06-15 23:42:38,342][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.18077433109283447, 'train/grad_norm_pre_clip': 7.320516109466553, 'timing/data': 0.0005417169886641204, 'timing/model': 1.500815206032712, 'learning_rate/qwen_vl_interface': 6.627759319632246e-06, 'learning_rate/action_model': 6.627759319632247e-05, 'epoch': 0.9}) +[2026-06-15 23:42:44,266][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.014769970439374447, 'train/grad_norm_pre_clip': 1.2901606559753418, 'timing/data': 0.0005528629990294576, 'timing/model': 1.6756534890155308, 'learning_rate/qwen_vl_interface': 6.624844537571245e-06, 'learning_rate/action_model': 6.624844537571245e-05, 'epoch': 0.9}) +[2026-06-15 23:42:50,057][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.24348431825637817, 'train/grad_norm_pre_clip': 12.175047874450684, 'timing/data': 0.0005333949811756611, 'timing/model': 1.4660882339812815, 'learning_rate/qwen_vl_interface': 6.621929190363585e-06, 'learning_rate/action_model': 6.621929190363585e-05, 'epoch': 0.9}) +[2026-06-15 23:42:56,056][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.1081705093383789, 'train/grad_norm_pre_clip': 8.423645973205566, 'timing/data': 0.0007610699976794422, 'timing/model': 1.4560546739958227, 'learning_rate/qwen_vl_interface': 6.619013279207657e-06, 'learning_rate/action_model': 6.619013279207656e-05, 'epoch': 0.9}) +[2026-06-15 23:43:02,154][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.3025360405445099, 'train/grad_norm_pre_clip': 9.167068481445312, 'timing/data': 0.0006125209620222449, 'timing/model': 1.4696233269642107, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 0.9}) +[2026-06-15 23:43:08,362][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.4297321140766144, 'train/grad_norm_pre_clip': 12.0744047164917, 'timing/data': 0.0005521830171346664, 'timing/model': 1.6023434670059942, 'learning_rate/qwen_vl_interface': 6.6131797698457135e-06, 'learning_rate/action_model': 6.613179769845713e-05, 'epoch': 0.9}) +[2026-06-15 23:43:14,332][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.11627131700515747, 'train/grad_norm_pre_clip': 5.57893705368042, 'timing/data': 0.0006372469943016768, 'timing/model': 1.7352346089901403, 'learning_rate/qwen_vl_interface': 6.610262174037633e-06, 'learning_rate/action_model': 6.610262174037633e-05, 'epoch': 0.9}) +[2026-06-15 23:43:20,282][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.17157630622386932, 'train/grad_norm_pre_clip': 9.785833358764648, 'timing/data': 0.0007384470081888139, 'timing/model': 1.6595061860280111, 'learning_rate/qwen_vl_interface': 6.607344019077155e-06, 'learning_rate/action_model': 6.607344019077156e-05, 'epoch': 0.9}) +[2026-06-15 23:43:25,952][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.12221148610115051, 'train/grad_norm_pre_clip': 5.449608325958252, 'timing/data': 0.0005642800242640078, 'timing/model': 1.4636523720109835, 'learning_rate/qwen_vl_interface': 6.604425306163822e-06, 'learning_rate/action_model': 6.604425306163822e-05, 'epoch': 0.9}) +[2026-06-15 23:52:19,452][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.04317281395196915, 'train/grad_norm_pre_clip': 5.303985118865967, 'eval/loss': 0.05985938757658005, 'eval/flap_precision': 0.903921568627451, 'eval/flap_recall': 0.9238476953907816, 'eval/flap_f1': 0.9137760158572845, 'eval/flap_rate_model': 0.14166666666666666, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9137760158572845, 'timing/data': 0.0005361189832910895, 'timing/model': 1.4582460410310887, 'learning_rate/qwen_vl_interface': 6.6015060364974045e-06, 'learning_rate/action_model': 6.601506036497404e-05, 'epoch': 0.9}) +[2026-06-15 23:52:19,453][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2100_state +[2026-06-15 23:52:19,457][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:52:24,951][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2100_state/model.safetensors +[2026-06-15 23:52:38,402][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2100_state/optimizer.bin +[2026-06-15 23:52:38,403][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2100_state/sampler.bin +[2026-06-15 23:52:38,403][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2100_state/sampler_1.bin +[2026-06-15 23:52:38,405][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2100_state/random_states_0.pkl +[2026-06-15 23:52:38,526][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2100 +[2026-06-15 23:52:38,526][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:52:38,537][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:52:44,635][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.25119346380233765, 'train/grad_norm_pre_clip': 7.371922016143799, 'timing/data': 0.0005453519988805056, 'timing/model': 1.4647688560071401, 'learning_rate/qwen_vl_interface': 6.598586211277907e-06, 'learning_rate/action_model': 6.598586211277907e-05, 'epoch': 0.9}) +[2026-06-15 23:52:50,539][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.051110729575157166, 'train/grad_norm_pre_clip': 3.4946391582489014, 'timing/data': 0.0005220579914748669, 'timing/model': 1.449426751001738, 'learning_rate/qwen_vl_interface': 6.59566583170556e-06, 'learning_rate/action_model': 6.59566583170556e-05, 'epoch': 0.9}) +[2026-06-15 23:52:56,542][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.051691874861717224, 'train/grad_norm_pre_clip': 4.694790363311768, 'timing/data': 0.0005368699785321951, 'timing/model': 1.4681404039729387, 'learning_rate/qwen_vl_interface': 6.59274489898082e-06, 'learning_rate/action_model': 6.592744898980819e-05, 'epoch': 0.9}) +[2026-06-15 23:53:02,605][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.041659265756607056, 'train/grad_norm_pre_clip': 4.10847806930542, 'timing/data': 0.000501257018186152, 'timing/model': 1.5354552479693666, 'learning_rate/qwen_vl_interface': 6.5898234143043716e-06, 'learning_rate/action_model': 6.589823414304371e-05, 'epoch': 0.91}) +[2026-06-15 23:53:08,502][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.09830722212791443, 'train/grad_norm_pre_clip': 8.503173828125, 'timing/data': 0.0005014380440115929, 'timing/model': 1.723649142018985, 'learning_rate/qwen_vl_interface': 6.5869013788771264e-06, 'learning_rate/action_model': 6.586901378877126e-05, 'epoch': 0.91}) +[2026-06-15 23:53:14,248][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.0684538334608078, 'train/grad_norm_pre_clip': 6.462264537811279, 'timing/data': 0.0005016080103814602, 'timing/model': 1.4688823010073975, 'learning_rate/qwen_vl_interface': 6.583978793900223e-06, 'learning_rate/action_model': 6.583978793900223e-05, 'epoch': 0.91}) +[2026-06-15 23:53:20,247][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.20751769840717316, 'train/grad_norm_pre_clip': 11.671976089477539, 'timing/data': 0.0005055440124124289, 'timing/model': 1.465172361000441, 'learning_rate/qwen_vl_interface': 6.581055660575028e-06, 'learning_rate/action_model': 6.581055660575027e-05, 'epoch': 0.91}) +[2026-06-15 23:53:26,246][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.24599169194698334, 'train/grad_norm_pre_clip': 9.340108871459961, 'timing/data': 0.0005027799634262919, 'timing/model': 1.4778140200069174, 'learning_rate/qwen_vl_interface': 6.578131980103128e-06, 'learning_rate/action_model': 6.578131980103128e-05, 'epoch': 0.91}) +[2026-06-15 23:53:32,145][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.01586354710161686, 'train/grad_norm_pre_clip': 1.3294798135757446, 'timing/data': 0.0005034710047766566, 'timing/model': 1.463625492004212, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 0.91}) +[2026-06-15 23:53:38,143][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.2930402159690857, 'train/grad_norm_pre_clip': 12.21494197845459, 'timing/data': 0.0005092589999549091, 'timing/model': 1.4995006670360453, 'learning_rate/qwen_vl_interface': 6.572282982526702e-06, 'learning_rate/action_model': 6.572282982526702e-05, 'epoch': 0.91}) +[2026-06-15 23:53:44,103][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.09940546751022339, 'train/grad_norm_pre_clip': 10.999737739562988, 'timing/data': 0.0005120940040796995, 'timing/model': 1.7218859479762614, 'learning_rate/qwen_vl_interface': 6.5693576678264755e-06, 'learning_rate/action_model': 6.569357667826476e-05, 'epoch': 0.91}) +[2026-06-15 23:53:49,748][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.09343713521957397, 'train/grad_norm_pre_clip': 6.7657294273376465, 'timing/data': 0.0005046430160291493, 'timing/model': 1.465779228019528, 'learning_rate/qwen_vl_interface': 6.566431810788148e-06, 'learning_rate/action_model': 6.566431810788148e-05, 'epoch': 0.91}) +[2026-06-15 23:53:55,751][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.057442557066679, 'train/grad_norm_pre_clip': 3.9388768672943115, 'timing/data': 0.0005614859983325005, 'timing/model': 1.465977607993409, 'learning_rate/qwen_vl_interface': 6.5635054126144306e-06, 'learning_rate/action_model': 6.56350541261443e-05, 'epoch': 0.91}) +[2026-06-15 23:54:01,651][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.0820997804403305, 'train/grad_norm_pre_clip': 5.368826866149902, 'timing/data': 0.0006366070010699332, 'timing/model': 1.4650653949938715, 'learning_rate/qwen_vl_interface': 6.560578474508254e-06, 'learning_rate/action_model': 6.560578474508253e-05, 'epoch': 0.91}) +[2026-06-15 23:54:07,452][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.06875742226839066, 'train/grad_norm_pre_clip': 3.4527339935302734, 'timing/data': 0.0005337349721230567, 'timing/model': 1.4651287460001186, 'learning_rate/qwen_vl_interface': 6.5576509976727685e-06, 'learning_rate/action_model': 6.557650997672769e-05, 'epoch': 0.91}) +[2026-06-15 23:54:13,348][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.17947755753993988, 'train/grad_norm_pre_clip': 13.895176887512207, 'timing/data': 0.000522479007486254, 'timing/model': 1.5064527489594184, 'learning_rate/qwen_vl_interface': 6.554722983311354e-06, 'learning_rate/action_model': 6.554722983311355e-05, 'epoch': 0.91}) +[2026-06-15 23:54:19,241][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.06403184682130814, 'train/grad_norm_pre_clip': 5.382143020629883, 'timing/data': 0.0005202459869906306, 'timing/model': 1.6512301879702136, 'learning_rate/qwen_vl_interface': 6.551794432627605e-06, 'learning_rate/action_model': 6.551794432627605e-05, 'epoch': 0.91}) +[2026-06-15 23:54:24,957][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.09613275527954102, 'train/grad_norm_pre_clip': 9.635814666748047, 'timing/data': 0.0005236609722487628, 'timing/model': 1.4715428699855693, 'learning_rate/qwen_vl_interface': 6.548865346825338e-06, 'learning_rate/action_model': 6.548865346825338e-05, 'epoch': 0.91}) +[2026-06-15 23:54:30,953][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.19323976337909698, 'train/grad_norm_pre_clip': 25.490432739257812, 'timing/data': 0.0005262439954094589, 'timing/model': 1.4643034979817457, 'learning_rate/qwen_vl_interface': 6.545935727108589e-06, 'learning_rate/action_model': 6.545935727108588e-05, 'epoch': 0.91}) +[2026-06-15 23:54:36,956][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.01038599293678999, 'train/grad_norm_pre_clip': 0.5895979404449463, 'timing/data': 0.0005292889545671642, 'timing/model': 1.4677006150013767, 'learning_rate/qwen_vl_interface': 6.543005574681613e-06, 'learning_rate/action_model': 6.543005574681613e-05, 'epoch': 0.91}) +[2026-06-15 23:54:42,953][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.08385887742042542, 'train/grad_norm_pre_clip': 6.653653144836426, 'timing/data': 0.0005339159979484975, 'timing/model': 1.5061296470230445, 'learning_rate/qwen_vl_interface': 6.5400748907488886e-06, 'learning_rate/action_model': 6.540074890748888e-05, 'epoch': 0.91}) +[2026-06-15 23:54:48,906][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.06542332470417023, 'train/grad_norm_pre_clip': 3.666930913925171, 'timing/data': 0.0005291479756124318, 'timing/model': 1.7158471760340035, 'learning_rate/qwen_vl_interface': 6.537143676515105e-06, 'learning_rate/action_model': 6.537143676515105e-05, 'epoch': 0.91}) +[2026-06-15 23:54:54,760][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.07136557251214981, 'train/grad_norm_pre_clip': 9.785356521606445, 'timing/data': 0.0005295690498314798, 'timing/model': 1.6696129980264232, 'learning_rate/qwen_vl_interface': 6.534211933185176e-06, 'learning_rate/action_model': 6.534211933185176e-05, 'epoch': 0.91}) +[2026-06-15 23:55:00,450][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.30347228050231934, 'train/grad_norm_pre_clip': 8.718704223632812, 'timing/data': 0.0005228689988143742, 'timing/model': 1.4614675490302034, 'learning_rate/qwen_vl_interface': 6.531279661964231e-06, 'learning_rate/action_model': 6.531279661964231e-05, 'epoch': 0.91}) +[2026-06-15 23:55:06,449][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.3142426908016205, 'train/grad_norm_pre_clip': 9.404972076416016, 'timing/data': 0.0005300000193528831, 'timing/model': 1.4545817769831046, 'learning_rate/qwen_vl_interface': 6.528346864057616e-06, 'learning_rate/action_model': 6.528346864057615e-05, 'epoch': 0.91}) +[2026-06-15 23:55:12,451][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.12461066246032715, 'train/grad_norm_pre_clip': 7.579461097717285, 'timing/data': 0.0005242409533821046, 'timing/model': 1.457193238020409, 'learning_rate/qwen_vl_interface': 6.5254135406708906e-06, 'learning_rate/action_model': 6.52541354067089e-05, 'epoch': 0.91}) +[2026-06-15 23:55:18,454][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.07693549245595932, 'train/grad_norm_pre_clip': 6.9745917320251465, 'timing/data': 0.0005342859658412635, 'timing/model': 1.4408141570165753, 'learning_rate/qwen_vl_interface': 6.522479693009836e-06, 'learning_rate/action_model': 6.522479693009837e-05, 'epoch': 0.92}) +[2026-06-15 23:55:24,408][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.008007154799997807, 'train/grad_norm_pre_clip': 0.9239175319671631, 'timing/data': 0.0005635989946313202, 'timing/model': 1.7179484209627844, 'learning_rate/qwen_vl_interface': 6.519545322280446e-06, 'learning_rate/action_model': 6.519545322280446e-05, 'epoch': 0.92}) +[2026-06-15 23:55:30,183][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.04428587481379509, 'train/grad_norm_pre_clip': 7.151437759399414, 'timing/data': 0.0005223379703238606, 'timing/model': 1.5894410589826293, 'learning_rate/qwen_vl_interface': 6.5166104296889275e-06, 'learning_rate/action_model': 6.516610429688927e-05, 'epoch': 0.92}) +[2026-06-15 23:55:35,859][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.059872664511203766, 'train/grad_norm_pre_clip': 8.491048812866211, 'timing/data': 0.0005199249717406929, 'timing/model': 1.4714243020280264, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.92}) +[2026-06-15 23:55:41,867][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.25993239879608154, 'train/grad_norm_pre_clip': 11.263370513916016, 'timing/data': 0.0005315819871611893, 'timing/model': 1.4767311699688435, 'learning_rate/qwen_vl_interface': 6.510739083745422e-06, 'learning_rate/action_model': 6.510739083745422e-05, 'epoch': 0.92}) +[2026-06-15 23:55:47,869][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.06264863908290863, 'train/grad_norm_pre_clip': 4.072386741638184, 'timing/data': 0.0005131749785505235, 'timing/model': 1.4797587529756129, 'learning_rate/qwen_vl_interface': 6.507802632806919e-06, 'learning_rate/action_model': 6.507802632806919e-05, 'epoch': 0.92}) +[2026-06-15 23:55:53,852][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.08092468231916428, 'train/grad_norm_pre_clip': 5.29508638381958, 'timing/data': 0.0005551170324906707, 'timing/model': 1.505480208026711, 'learning_rate/qwen_vl_interface': 6.5048656648332646e-06, 'learning_rate/action_model': 6.504865664833264e-05, 'epoch': 0.92}) +[2026-06-15 23:55:59,842][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.01037443708628416, 'train/grad_norm_pre_clip': 1.3813632726669312, 'timing/data': 0.0005060450057499111, 'timing/model': 1.7544079080107622, 'learning_rate/qwen_vl_interface': 6.501928181031737e-06, 'learning_rate/action_model': 6.501928181031737e-05, 'epoch': 0.92}) +[2026-06-15 23:56:05,562][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.01811564527451992, 'train/grad_norm_pre_clip': 1.9596023559570312, 'timing/data': 0.0005081979907117784, 'timing/model': 1.4763396300259046, 'learning_rate/qwen_vl_interface': 6.498990182609823e-06, 'learning_rate/action_model': 6.498990182609823e-05, 'epoch': 0.92}) +[2026-06-15 23:56:11,553][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.18700440227985382, 'train/grad_norm_pre_clip': 11.216784477233887, 'timing/data': 0.0005297790048643947, 'timing/model': 1.4672020149882883, 'learning_rate/qwen_vl_interface': 6.4960516707752255e-06, 'learning_rate/action_model': 6.496051670775226e-05, 'epoch': 0.92}) +[2026-06-15 23:56:17,456][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.10814446955919266, 'train/grad_norm_pre_clip': 7.365284442901611, 'timing/data': 0.0005292390123941004, 'timing/model': 1.4662383350078017, 'learning_rate/qwen_vl_interface': 6.493112646735851e-06, 'learning_rate/action_model': 6.493112646735851e-05, 'epoch': 0.92}) +[2026-06-15 23:56:23,472][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.15978673100471497, 'train/grad_norm_pre_clip': 14.477869033813477, 'timing/data': 0.0005341260111890733, 'timing/model': 1.4821184489992447, 'learning_rate/qwen_vl_interface': 6.490173111699824e-06, 'learning_rate/action_model': 6.490173111699825e-05, 'epoch': 0.92}) +[2026-06-15 23:56:29,460][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.07285914570093155, 'train/grad_norm_pre_clip': 5.367877006530762, 'timing/data': 0.0005255930009298027, 'timing/model': 1.4925564769655466, 'learning_rate/qwen_vl_interface': 6.487233066875479e-06, 'learning_rate/action_model': 6.487233066875479e-05, 'epoch': 0.92}) +[2026-06-15 23:56:35,596][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.11010480672121048, 'train/grad_norm_pre_clip': 7.730643272399902, 'timing/data': 0.0005272559938021004, 'timing/model': 1.7735307579860091, 'learning_rate/qwen_vl_interface': 6.484292513471352e-06, 'learning_rate/action_model': 6.484292513471352e-05, 'epoch': 0.92}) +[2026-06-15 23:56:41,362][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.24419231712818146, 'train/grad_norm_pre_clip': 6.685782432556152, 'timing/data': 0.000541756977327168, 'timing/model': 1.4675383850117214, 'learning_rate/qwen_vl_interface': 6.4813514526961975e-06, 'learning_rate/action_model': 6.481351452696198e-05, 'epoch': 0.92}) +[2026-06-15 23:56:47,354][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.09844871610403061, 'train/grad_norm_pre_clip': 5.796412467956543, 'timing/data': 0.0005295189912430942, 'timing/model': 1.4621749119833112, 'learning_rate/qwen_vl_interface': 6.478409885758972e-06, 'learning_rate/action_model': 6.478409885758971e-05, 'epoch': 0.92}) +[2026-06-15 23:56:53,260][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.0354854054749012, 'train/grad_norm_pre_clip': 3.075914144515991, 'timing/data': 0.0005404659896157682, 'timing/model': 1.4677352019934915, 'learning_rate/qwen_vl_interface': 6.475467813868844e-06, 'learning_rate/action_model': 6.475467813868844e-05, 'epoch': 0.92}) +[2026-06-15 23:56:59,058][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.14706361293792725, 'train/grad_norm_pre_clip': 4.680354118347168, 'timing/data': 0.0005257640150375664, 'timing/model': 1.4612945650005713, 'learning_rate/qwen_vl_interface': 6.472525238235187e-06, 'learning_rate/action_model': 6.472525238235186e-05, 'epoch': 0.92}) +[2026-06-15 23:57:04,941][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.16727688908576965, 'train/grad_norm_pre_clip': 8.548666954040527, 'timing/data': 0.0005278559983707964, 'timing/model': 1.5017190130311064, 'learning_rate/qwen_vl_interface': 6.4695821600675834e-06, 'learning_rate/action_model': 6.469582160067583e-05, 'epoch': 0.92}) +[2026-06-15 23:57:10,837][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.10021626204252243, 'train/grad_norm_pre_clip': 5.524069786071777, 'timing/data': 0.0005188829964026809, 'timing/model': 1.6475804750225507, 'learning_rate/qwen_vl_interface': 6.466638580575824e-06, 'learning_rate/action_model': 6.466638580575824e-05, 'epoch': 0.92}) +[2026-06-15 23:57:16,556][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.035451386123895645, 'train/grad_norm_pre_clip': 3.120959997177124, 'timing/data': 0.0005090889753773808, 'timing/model': 1.4528584689833224, 'learning_rate/qwen_vl_interface': 6.463694500969898e-06, 'learning_rate/action_model': 6.463694500969898e-05, 'epoch': 0.92}) +[2026-06-15 23:57:22,563][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.2408105432987213, 'train/grad_norm_pre_clip': 14.423113822937012, 'timing/data': 0.000599712016992271, 'timing/model': 1.4730205640080385, 'learning_rate/qwen_vl_interface': 6.460749922460008e-06, 'learning_rate/action_model': 6.460749922460008e-05, 'epoch': 0.92}) +[2026-06-15 23:57:28,553][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.1864006221294403, 'train/grad_norm_pre_clip': 11.822338104248047, 'timing/data': 0.0005352570442482829, 'timing/model': 1.473343177989591, 'learning_rate/qwen_vl_interface': 6.457804846256559e-06, 'learning_rate/action_model': 6.457804846256559e-05, 'epoch': 0.92}) +[2026-06-15 23:57:34,551][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.13369722664356232, 'train/grad_norm_pre_clip': 7.009040355682373, 'timing/data': 0.0005233999690972269, 'timing/model': 1.4867161140427925, 'learning_rate/qwen_vl_interface': 6.45485927357016e-06, 'learning_rate/action_model': 6.454859273570159e-05, 'epoch': 0.93}) +[2026-06-15 23:57:40,502][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.03145642951130867, 'train/grad_norm_pre_clip': 1.9135867357254028, 'timing/data': 0.0005275459843687713, 'timing/model': 1.7117880209698342, 'learning_rate/qwen_vl_interface': 6.451913205611626e-06, 'learning_rate/action_model': 6.451913205611626e-05, 'epoch': 0.93}) +[2026-06-15 23:57:46,356][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.07447108626365662, 'train/grad_norm_pre_clip': 9.422454833984375, 'timing/data': 0.000544792041182518, 'timing/model': 1.6637237509712577, 'learning_rate/qwen_vl_interface': 6.448966643591974e-06, 'learning_rate/action_model': 6.448966643591974e-05, 'epoch': 0.93}) +[2026-06-15 23:57:52,054][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.07044992595911026, 'train/grad_norm_pre_clip': 3.293358564376831, 'timing/data': 0.0005363189848139882, 'timing/model': 1.4613500389968976, 'learning_rate/qwen_vl_interface': 6.446019588722423e-06, 'learning_rate/action_model': 6.446019588722422e-05, 'epoch': 0.93}) +[2026-06-15 23:57:58,068][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.18535035848617554, 'train/grad_norm_pre_clip': 8.211686134338379, 'timing/data': 0.0005298900068737566, 'timing/model': 1.4766592240193859, 'learning_rate/qwen_vl_interface': 6.443072042214398e-06, 'learning_rate/action_model': 6.443072042214397e-05, 'epoch': 0.93}) +[2026-06-15 23:58:04,060][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.18290501832962036, 'train/grad_norm_pre_clip': 6.929291248321533, 'timing/data': 0.0005241109756752849, 'timing/model': 1.4783292799838819, 'learning_rate/qwen_vl_interface': 6.440124005279521e-06, 'learning_rate/action_model': 6.440124005279521e-05, 'epoch': 0.93}) +[2026-06-15 23:58:10,044][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.17590060830116272, 'train/grad_norm_pre_clip': 13.118880271911621, 'timing/data': 0.0005224380292929709, 'timing/model': 1.5000440980074927, 'learning_rate/qwen_vl_interface': 6.43717547912962e-06, 'learning_rate/action_model': 6.437175479129621e-05, 'epoch': 0.93}) +[2026-06-15 23:58:16,003][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.2529366612434387, 'train/grad_norm_pre_clip': 11.491703033447266, 'timing/data': 0.0005257640150375664, 'timing/model': 1.7132348450249992, 'learning_rate/qwen_vl_interface': 6.434226464976725e-06, 'learning_rate/action_model': 6.434226464976724e-05, 'epoch': 0.93}) +[2026-06-15 23:58:21,777][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.058731984347105026, 'train/grad_norm_pre_clip': 2.4834482669830322, 'timing/data': 0.0005267750238999724, 'timing/model': 1.5864687570137903, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 0.93}) +[2026-06-15 23:58:27,452][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.09102637320756912, 'train/grad_norm_pre_clip': 5.90140438079834, 'timing/data': 0.0005301809869706631, 'timing/model': 1.4618081279913895, 'learning_rate/qwen_vl_interface': 6.428326977511059e-06, 'learning_rate/action_model': 6.428326977511058e-05, 'epoch': 0.93}) +[2026-06-15 23:58:33,453][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.04038984328508377, 'train/grad_norm_pre_clip': 2.8444199562072754, 'timing/data': 0.0005089289625175297, 'timing/model': 1.46936285297852, 'learning_rate/qwen_vl_interface': 6.425376506623344e-06, 'learning_rate/action_model': 6.425376506623344e-05, 'epoch': 0.93}) +[2026-06-15 23:58:39,453][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.12115979194641113, 'train/grad_norm_pre_clip': 8.143672943115234, 'timing/data': 0.0005022480036132038, 'timing/model': 1.4475325419916771, 'learning_rate/qwen_vl_interface': 6.422425552582747e-06, 'learning_rate/action_model': 6.422425552582747e-05, 'epoch': 0.93}) +[2026-06-15 23:58:45,453][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.08882685005664825, 'train/grad_norm_pre_clip': 4.970510482788086, 'timing/data': 0.0005097900284454226, 'timing/model': 1.5080567239783704, 'learning_rate/qwen_vl_interface': 6.41947411660229e-06, 'learning_rate/action_model': 6.41947411660229e-05, 'epoch': 0.93}) +[2026-06-15 23:58:51,409][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.1453281193971634, 'train/grad_norm_pre_clip': 9.008910179138184, 'timing/data': 0.0005225189961493015, 'timing/model': 1.7174328359542415, 'learning_rate/qwen_vl_interface': 6.4165221998951995e-06, 'learning_rate/action_model': 6.4165221998952e-05, 'epoch': 0.93}) +[2026-06-15 23:58:57,055][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.19532060623168945, 'train/grad_norm_pre_clip': 6.956859588623047, 'timing/data': 0.0005150180077180266, 'timing/model': 1.4630469069816172, 'learning_rate/qwen_vl_interface': 6.413569803674896e-06, 'learning_rate/action_model': 6.413569803674896e-05, 'epoch': 0.93}) +[2026-06-15 23:59:03,055][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.23227708041667938, 'train/grad_norm_pre_clip': 10.016534805297852, 'timing/data': 0.0005258640157990158, 'timing/model': 1.46369450900238, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 0.93}) +[2026-06-15 23:59:08,955][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.10150113701820374, 'train/grad_norm_pre_clip': 4.948971271514893, 'timing/data': 0.0005279459874145687, 'timing/model': 1.4643347609671764, 'learning_rate/qwen_vl_interface': 6.40766357754932e-06, 'learning_rate/action_model': 6.407663577549319e-05, 'epoch': 0.93}) +[2026-06-15 23:59:15,033][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.10686028003692627, 'train/grad_norm_pre_clip': 9.792607307434082, 'timing/data': 0.0005299799959175289, 'timing/model': 1.5786899260128848, 'learning_rate/qwen_vl_interface': 6.404709750071875e-06, 'learning_rate/action_model': 6.404709750071874e-05, 'epoch': 0.93}) +[2026-06-15 23:59:21,055][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.1621198058128357, 'train/grad_norm_pre_clip': 4.472515106201172, 'timing/data': 0.0005270960391499102, 'timing/model': 1.506615858990699, 'learning_rate/qwen_vl_interface': 6.401755447936869e-06, 'learning_rate/action_model': 6.401755447936869e-05, 'epoch': 0.93}) +[2026-06-15 23:59:27,007][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.11322766542434692, 'train/grad_norm_pre_clip': 4.318731784820557, 'timing/data': 0.0005021379911340773, 'timing/model': 1.7173479259945452, 'learning_rate/qwen_vl_interface': 6.3988006723587036e-06, 'learning_rate/action_model': 6.398800672358703e-05, 'epoch': 0.93}) +[2026-06-15 23:59:32,662][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.17296309769153595, 'train/grad_norm_pre_clip': 8.823460578918457, 'timing/data': 0.0004980929661542177, 'timing/model': 1.47354097396601, 'learning_rate/qwen_vl_interface': 6.395845424551975e-06, 'learning_rate/action_model': 6.395845424551975e-05, 'epoch': 0.93}) +[2026-06-15 23:59:38,654][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.09731413424015045, 'train/grad_norm_pre_clip': 7.01479434967041, 'timing/data': 0.0005342260119505227, 'timing/model': 1.4600135899963789, 'learning_rate/qwen_vl_interface': 6.392889705731475e-06, 'learning_rate/action_model': 6.392889705731476e-05, 'epoch': 0.93}) +[2026-06-15 23:59:44,555][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.27957502007484436, 'train/grad_norm_pre_clip': 9.914447784423828, 'timing/data': 0.0005269250250421464, 'timing/model': 1.462188079021871, 'learning_rate/qwen_vl_interface': 6.389933517112188e-06, 'learning_rate/action_model': 6.389933517112187e-05, 'epoch': 0.93}) +[2026-06-15 23:59:50,367][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.04824436083436012, 'train/grad_norm_pre_clip': 3.226759672164917, 'timing/data': 0.000535067985765636, 'timing/model': 1.4587898280005902, 'learning_rate/qwen_vl_interface': 6.386976859909293e-06, 'learning_rate/action_model': 6.386976859909292e-05, 'epoch': 0.93}) +[2026-06-15 23:59:56,255][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.1202712133526802, 'train/grad_norm_pre_clip': 6.788760185241699, 'timing/data': 0.0005915400106459856, 'timing/model': 1.4900447329855524, 'learning_rate/qwen_vl_interface': 6.384019735338156e-06, 'learning_rate/action_model': 6.384019735338156e-05, 'epoch': 0.94}) +[2026-06-16 00:00:02,138][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.24915580451488495, 'train/grad_norm_pre_clip': 8.20395565032959, 'timing/data': 0.0005670840037055314, 'timing/model': 1.6467947339988314, 'learning_rate/qwen_vl_interface': 6.381062144614342e-06, 'learning_rate/action_model': 6.381062144614342e-05, 'epoch': 0.94}) +[2026-06-16 00:00:07,855][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.0871264785528183, 'train/grad_norm_pre_clip': 6.9543843269348145, 'timing/data': 0.00051328499102965, 'timing/model': 1.4645539170014672, 'learning_rate/qwen_vl_interface': 6.378104088953605e-06, 'learning_rate/action_model': 6.378104088953605e-05, 'epoch': 0.94}) +[2026-06-16 00:00:13,856][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.08402389287948608, 'train/grad_norm_pre_clip': 5.620567321777344, 'timing/data': 0.0005063749849796295, 'timing/model': 1.466089995985385, 'learning_rate/qwen_vl_interface': 6.375145569571889e-06, 'learning_rate/action_model': 6.375145569571889e-05, 'epoch': 0.94}) +[2026-06-16 00:00:19,854][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.12089148908853531, 'train/grad_norm_pre_clip': 5.916298866271973, 'timing/data': 0.0004955589538440108, 'timing/model': 1.4647174359997734, 'learning_rate/qwen_vl_interface': 6.372186587685329e-06, 'learning_rate/action_model': 6.372186587685329e-05, 'epoch': 0.94}) +[2026-06-16 00:00:25,843][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.11540432274341583, 'train/grad_norm_pre_clip': 5.932507038116455, 'timing/data': 0.00052349001634866, 'timing/model': 1.4994938930030912, 'learning_rate/qwen_vl_interface': 6.369227144510252e-06, 'learning_rate/action_model': 6.369227144510252e-05, 'epoch': 0.94}) +[2026-06-16 00:00:31,801][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.05151534825563431, 'train/grad_norm_pre_clip': 1.9937803745269775, 'timing/data': 0.0005205560009926558, 'timing/model': 1.7109725279733539, 'learning_rate/qwen_vl_interface': 6.3662672412631736e-06, 'learning_rate/action_model': 6.366267241263173e-05, 'epoch': 0.94}) +[2026-06-16 00:00:37,658][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.14458177983760834, 'train/grad_norm_pre_clip': 6.3066558837890625, 'timing/data': 0.0005252929986454546, 'timing/model': 1.6686374680139124, 'learning_rate/qwen_vl_interface': 6.3633068791607954e-06, 'learning_rate/action_model': 6.363306879160795e-05, 'epoch': 0.94}) +[2026-06-16 00:00:43,351][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.12186925113201141, 'train/grad_norm_pre_clip': 5.265474796295166, 'timing/data': 0.0007456570165231824, 'timing/model': 1.462304533983115, 'learning_rate/qwen_vl_interface': 6.360346059420014e-06, 'learning_rate/action_model': 6.360346059420014e-05, 'epoch': 0.94}) +[2026-06-16 00:00:49,353][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.18936491012573242, 'train/grad_norm_pre_clip': 5.6882171630859375, 'timing/data': 0.0005258630262687802, 'timing/model': 1.4626663950039074, 'learning_rate/qwen_vl_interface': 6.357384783257909e-06, 'learning_rate/action_model': 6.357384783257908e-05, 'epoch': 0.94}) +[2026-06-16 00:00:55,353][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.09657766669988632, 'train/grad_norm_pre_clip': 11.258196830749512, 'timing/data': 0.0005238709854893386, 'timing/model': 1.463796577998437, 'learning_rate/qwen_vl_interface': 6.354423051891748e-06, 'learning_rate/action_model': 6.354423051891749e-05, 'epoch': 0.94}) +[2026-06-16 00:01:01,352][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.0625227838754654, 'train/grad_norm_pre_clip': 3.6483142375946045, 'timing/data': 0.0005266150110401213, 'timing/model': 1.4904555610264651, 'learning_rate/qwen_vl_interface': 6.35146086653899e-06, 'learning_rate/action_model': 6.35146086653899e-05, 'epoch': 0.94}) +[2026-06-16 00:01:07,407][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.10367152839899063, 'train/grad_norm_pre_clip': 5.286365509033203, 'timing/data': 0.0005129040218889713, 'timing/model': 1.7144577809958719, 'learning_rate/qwen_vl_interface': 6.348498228417273e-06, 'learning_rate/action_model': 6.348498228417273e-05, 'epoch': 0.94}) +[2026-06-16 00:01:13,182][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.06222676485776901, 'train/grad_norm_pre_clip': 4.940730094909668, 'timing/data': 0.0005618269788101315, 'timing/model': 1.5906644010101445, 'learning_rate/qwen_vl_interface': 6.34553513874443e-06, 'learning_rate/action_model': 6.34553513874443e-05, 'epoch': 0.94}) +[2026-06-16 00:01:18,857][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.051448363810777664, 'train/grad_norm_pre_clip': 3.0679285526275635, 'timing/data': 0.0005197040154598653, 'timing/model': 1.464780582988169, 'learning_rate/qwen_vl_interface': 6.342571598738473e-06, 'learning_rate/action_model': 6.342571598738472e-05, 'epoch': 0.94}) +[2026-06-16 00:01:24,855][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.16070012748241425, 'train/grad_norm_pre_clip': 9.162353515625, 'timing/data': 0.0005005960119888186, 'timing/model': 1.4675491669913754, 'learning_rate/qwen_vl_interface': 6.339607609617603e-06, 'learning_rate/action_model': 6.339607609617603e-05, 'epoch': 0.94}) +[2026-06-16 00:01:30,857][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.11504766345024109, 'train/grad_norm_pre_clip': 11.006453514099121, 'timing/data': 0.0005255829892121255, 'timing/model': 1.4678132149856538, 'learning_rate/qwen_vl_interface': 6.336643172600202e-06, 'learning_rate/action_model': 6.336643172600202e-05, 'epoch': 0.94}) +[2026-06-16 00:01:36,853][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.039623603224754333, 'train/grad_norm_pre_clip': 2.873196601867676, 'timing/data': 0.0005330240237526596, 'timing/model': 1.5040167710394599, 'learning_rate/qwen_vl_interface': 6.333678288904838e-06, 'learning_rate/action_model': 6.333678288904837e-05, 'epoch': 0.94}) +[2026-06-16 00:01:42,812][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.051050350069999695, 'train/grad_norm_pre_clip': 3.189286231994629, 'timing/data': 0.0005366089753806591, 'timing/model': 1.7222288419725373, 'learning_rate/qwen_vl_interface': 6.330712959750264e-06, 'learning_rate/action_model': 6.330712959750263e-05, 'epoch': 0.94}) +[2026-06-16 00:01:48,456][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.2605242133140564, 'train/grad_norm_pre_clip': 7.880209445953369, 'timing/data': 0.0005234609707258642, 'timing/model': 1.4663974600262009, 'learning_rate/qwen_vl_interface': 6.3277471863554145e-06, 'learning_rate/action_model': 6.327747186355414e-05, 'epoch': 0.94}) +[2026-06-16 00:01:54,454][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.09886281192302704, 'train/grad_norm_pre_clip': 6.3847808837890625, 'timing/data': 0.0005289479740895331, 'timing/model': 1.4646237489650957, 'learning_rate/qwen_vl_interface': 6.324780969939406e-06, 'learning_rate/action_model': 6.324780969939405e-05, 'epoch': 0.94}) +[2026-06-16 00:02:00,354][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.17161309719085693, 'train/grad_norm_pre_clip': 6.662405014038086, 'timing/data': 0.0005409459699876606, 'timing/model': 1.4613246750086546, 'learning_rate/qwen_vl_interface': 6.32181431172154e-06, 'learning_rate/action_model': 6.32181431172154e-05, 'epoch': 0.94}) +[2026-06-16 00:02:06,255][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.009546545334160328, 'train/grad_norm_pre_clip': 1.2283921241760254, 'timing/data': 0.0005297689931467175, 'timing/model': 1.4643436770420521, 'learning_rate/qwen_vl_interface': 6.318847212921295e-06, 'learning_rate/action_model': 6.318847212921295e-05, 'epoch': 0.94}) +[2026-06-16 00:02:12,256][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.0346996895968914, 'train/grad_norm_pre_clip': 2.2357091903686523, 'timing/data': 0.0005273959832265973, 'timing/model': 1.502778275986202, 'learning_rate/qwen_vl_interface': 6.3158796747583364e-06, 'learning_rate/action_model': 6.315879674758336e-05, 'epoch': 0.95}) +[2026-06-16 00:02:18,211][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.1509152203798294, 'train/grad_norm_pre_clip': 6.053441047668457, 'timing/data': 0.0005296589806675911, 'timing/model': 1.7170423869974911, 'learning_rate/qwen_vl_interface': 6.312911698452503e-06, 'learning_rate/action_model': 6.312911698452502e-05, 'epoch': 0.95}) +[2026-06-16 00:02:23,856][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.06481903046369553, 'train/grad_norm_pre_clip': 3.51796555519104, 'timing/data': 0.0005177719867788255, 'timing/model': 1.465325560013298, 'learning_rate/qwen_vl_interface': 6.309943285223824e-06, 'learning_rate/action_model': 6.309943285223824e-05, 'epoch': 0.95}) +[2026-06-16 00:11:41,752][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.12267304211854935, 'train/grad_norm_pre_clip': 8.793564796447754, 'eval/loss': 0.06686650216579437, 'eval/flap_precision': 0.8836772983114447, 'eval/flap_recall': 0.9438877755511023, 'eval/flap_f1': 0.9127906976744186, 'eval/flap_rate_model': 0.14805555555555555, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9127906976744187, 'timing/data': 0.0005230890237726271, 'timing/model': 1.4800115769612603, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.95}) +[2026-06-16 00:11:41,753][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2200_state +[2026-06-16 00:11:41,757][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 00:11:47,319][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2200_state/model.safetensors +[2026-06-16 00:11:55,545][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2200_state/optimizer.bin +[2026-06-16 00:11:55,546][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2200_state/sampler.bin +[2026-06-16 00:11:55,546][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2200_state/sampler_1.bin +[2026-06-16 00:11:55,547][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2200_state/random_states_0.pkl +[2026-06-16 00:11:55,657][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2200 +[2026-06-16 00:11:55,657][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 00:11:55,667][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 00:12:01,338][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.21464292705059052, 'train/grad_norm_pre_clip': 7.714463233947754, 'timing/data': 0.0005527730099856853, 'timing/model': 1.4439198080217466, 'learning_rate/qwen_vl_interface': 6.304005152878902e-06, 'learning_rate/action_model': 6.304005152878901e-05, 'epoch': 0.95}) +[2026-06-16 00:12:07,338][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.22309456765651703, 'train/grad_norm_pre_clip': 45.67424392700195, 'timing/data': 0.0005292589776217937, 'timing/model': 1.4560130119789392, 'learning_rate/qwen_vl_interface': 6.301035436203602e-06, 'learning_rate/action_model': 6.301035436203602e-05, 'epoch': 0.95}) +[2026-06-16 00:12:13,341][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.24391113221645355, 'train/grad_norm_pre_clip': 24.22358512878418, 'timing/data': 0.0006076639983803034, 'timing/model': 1.4651770760538056, 'learning_rate/qwen_vl_interface': 6.298065287487337e-06, 'learning_rate/action_model': 6.298065287487337e-05, 'epoch': 0.95}) +[2026-06-16 00:12:19,447][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.03477364778518677, 'train/grad_norm_pre_clip': 4.797867298126221, 'timing/data': 0.0008036619983613491, 'timing/model': 1.4572064370149747, 'learning_rate/qwen_vl_interface': 6.295094707951019e-06, 'learning_rate/action_model': 6.295094707951019e-05, 'epoch': 0.95}) +[2026-06-16 00:12:25,748][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.1582333892583847, 'train/grad_norm_pre_clip': 6.322177410125732, 'timing/data': 0.0006586089730262756, 'timing/model': 1.477933912014123, 'learning_rate/qwen_vl_interface': 6.292123698815745e-06, 'learning_rate/action_model': 6.292123698815745e-05, 'epoch': 0.95}) +[2026-06-16 00:12:31,807][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.20507878065109253, 'train/grad_norm_pre_clip': 7.457394599914551, 'timing/data': 0.0008201369782909751, 'timing/model': 1.7340632310369983, 'learning_rate/qwen_vl_interface': 6.289152261302783e-06, 'learning_rate/action_model': 6.289152261302784e-05, 'epoch': 0.95}) +[2026-06-16 00:12:37,455][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.04672868177294731, 'train/grad_norm_pre_clip': 4.5456390380859375, 'timing/data': 0.000534506980329752, 'timing/model': 1.4681944539770484, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 0.95}) +[2026-06-16 00:12:43,548][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.05408705025911331, 'train/grad_norm_pre_clip': 7.216098308563232, 'timing/data': 0.0005694880383089185, 'timing/model': 1.4644251750432886, 'learning_rate/qwen_vl_interface': 6.283208106029753e-06, 'learning_rate/action_model': 6.283208106029752e-05, 'epoch': 0.95}) +[2026-06-16 00:12:49,648][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.1884901225566864, 'train/grad_norm_pre_clip': 16.9887752532959, 'timing/data': 0.0005242509650997818, 'timing/model': 1.4857634780346416, 'learning_rate/qwen_vl_interface': 6.280235390713102e-06, 'learning_rate/action_model': 6.280235390713102e-05, 'epoch': 0.95}) +[2026-06-16 00:12:55,552][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.09297756850719452, 'train/grad_norm_pre_clip': 9.624138832092285, 'timing/data': 0.0005178020219318569, 'timing/model': 1.466027902031783, 'learning_rate/qwen_vl_interface': 6.277262251905597e-06, 'learning_rate/action_model': 6.277262251905597e-05, 'epoch': 0.95}) +[2026-06-16 00:13:01,552][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.11130859702825546, 'train/grad_norm_pre_clip': 18.967193603515625, 'timing/data': 0.00053346500499174, 'timing/model': 1.4443569499999285, 'learning_rate/qwen_vl_interface': 6.274288690829385e-06, 'learning_rate/action_model': 6.274288690829384e-05, 'epoch': 0.95}) +[2026-06-16 00:13:07,506][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.019304055720567703, 'train/grad_norm_pre_clip': 1.758422613143921, 'timing/data': 0.0005429589655250311, 'timing/model': 1.7189415009925142, 'learning_rate/qwen_vl_interface': 6.271314708706781e-06, 'learning_rate/action_model': 6.271314708706782e-05, 'epoch': 0.95}) +[2026-06-16 00:13:13,156][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.32930856943130493, 'train/grad_norm_pre_clip': 18.608057022094727, 'timing/data': 0.00054967898176983, 'timing/model': 1.4649988919845782, 'learning_rate/qwen_vl_interface': 6.26834030676028e-06, 'learning_rate/action_model': 6.26834030676028e-05, 'epoch': 0.95}) +[2026-06-16 00:13:19,154][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.010461523197591305, 'train/grad_norm_pre_clip': 1.4472249746322632, 'timing/data': 0.0005416970234364271, 'timing/model': 1.465078042994719, 'learning_rate/qwen_vl_interface': 6.265365486212545e-06, 'learning_rate/action_model': 6.265365486212545e-05, 'epoch': 0.95}) +[2026-06-16 00:13:25,154][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.016925973817706108, 'train/grad_norm_pre_clip': 1.4219584465026855, 'timing/data': 0.0005624369950965047, 'timing/model': 1.4662350590224378, 'learning_rate/qwen_vl_interface': 6.26239024828641e-06, 'learning_rate/action_model': 6.26239024828641e-05, 'epoch': 0.95}) +[2026-06-16 00:13:31,055][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.033561043441295624, 'train/grad_norm_pre_clip': 3.23190975189209, 'timing/data': 0.0004985330160707235, 'timing/model': 1.4653317350312136, 'learning_rate/qwen_vl_interface': 6.259414594204888e-06, 'learning_rate/action_model': 6.259414594204888e-05, 'epoch': 0.95}) +[2026-06-16 00:13:37,049][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.20511889457702637, 'train/grad_norm_pre_clip': 12.20140266418457, 'timing/data': 0.0005106510361656547, 'timing/model': 1.5085524159949273, 'learning_rate/qwen_vl_interface': 6.256438525191155e-06, 'learning_rate/action_model': 6.256438525191156e-05, 'epoch': 0.95}) +[2026-06-16 00:13:42,996][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.17283716797828674, 'train/grad_norm_pre_clip': 25.273887634277344, 'timing/data': 0.0005228790105320513, 'timing/model': 1.7101364400004968, 'learning_rate/qwen_vl_interface': 6.253462042468561e-06, 'learning_rate/action_model': 6.25346204246856e-05, 'epoch': 0.95}) +[2026-06-16 00:13:48,653][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.26158407330513, 'train/grad_norm_pre_clip': 8.790696144104004, 'timing/data': 0.0005166800110600889, 'timing/model': 1.4638228279654868, 'learning_rate/qwen_vl_interface': 6.2504851472606246e-06, 'learning_rate/action_model': 6.250485147260625e-05, 'epoch': 0.95}) +[2026-06-16 00:13:54,656][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.07416960597038269, 'train/grad_norm_pre_clip': 3.2378273010253906, 'timing/data': 0.0005259640165604651, 'timing/model': 1.4684955660486594, 'learning_rate/qwen_vl_interface': 6.247507840791035e-06, 'learning_rate/action_model': 6.247507840791035e-05, 'epoch': 0.96}) +[2026-06-16 00:14:00,456][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.2195410430431366, 'train/grad_norm_pre_clip': 12.400108337402344, 'timing/data': 0.0005312520079314709, 'timing/model': 1.4653746589901857, 'learning_rate/qwen_vl_interface': 6.244530124283651e-06, 'learning_rate/action_model': 6.24453012428365e-05, 'epoch': 0.96}) +[2026-06-16 00:14:06,257][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.18806669116020203, 'train/grad_norm_pre_clip': 10.82563304901123, 'timing/data': 0.0005248920060694218, 'timing/model': 1.4628370899590664, 'learning_rate/qwen_vl_interface': 6.2415519989625016e-06, 'learning_rate/action_model': 6.241551998962501e-05, 'epoch': 0.96}) +[2026-06-16 00:14:12,138][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.18599548935890198, 'train/grad_norm_pre_clip': 11.049965858459473, 'timing/data': 0.0005269449902698398, 'timing/model': 1.5055566750233993, 'learning_rate/qwen_vl_interface': 6.238573466051781e-06, 'learning_rate/action_model': 6.23857346605178e-05, 'epoch': 0.96}) +[2026-06-16 00:14:18,130][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.14941799640655518, 'train/grad_norm_pre_clip': 14.627655982971191, 'timing/data': 0.0005294489674270153, 'timing/model': 1.6383557579829358, 'learning_rate/qwen_vl_interface': 6.2355945267758465e-06, 'learning_rate/action_model': 6.235594526775846e-05, 'epoch': 0.96}) +[2026-06-16 00:14:23,868][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.17252381145954132, 'train/grad_norm_pre_clip': 8.69190788269043, 'timing/data': 0.0005326540558598936, 'timing/model': 1.4766397660132498, 'learning_rate/qwen_vl_interface': 6.232615182359231e-06, 'learning_rate/action_model': 6.232615182359231e-05, 'epoch': 0.96}) +[2026-06-16 00:14:29,755][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.10483279824256897, 'train/grad_norm_pre_clip': 7.791424751281738, 'timing/data': 0.0005320720374584198, 'timing/model': 1.4647509510396048, 'learning_rate/qwen_vl_interface': 6.229635434026634e-06, 'learning_rate/action_model': 6.229635434026634e-05, 'epoch': 0.96}) +[2026-06-16 00:14:35,757][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.10616563260555267, 'train/grad_norm_pre_clip': 13.42431640625, 'timing/data': 0.0005075969966128469, 'timing/model': 1.466134480026085, 'learning_rate/qwen_vl_interface': 6.226655283002912e-06, 'learning_rate/action_model': 6.226655283002911e-05, 'epoch': 0.96}) +[2026-06-16 00:14:41,740][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.24863825738430023, 'train/grad_norm_pre_clip': 20.64542007446289, 'timing/data': 0.0005071659688837826, 'timing/model': 1.5029972799820825, 'learning_rate/qwen_vl_interface': 6.223674730513096e-06, 'learning_rate/action_model': 6.223674730513096e-05, 'epoch': 0.96}) +[2026-06-16 00:14:47,698][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.21871721744537354, 'train/grad_norm_pre_clip': 20.913652420043945, 'timing/data': 0.0005088490433990955, 'timing/model': 1.7060027900151908, 'learning_rate/qwen_vl_interface': 6.220693777782375e-06, 'learning_rate/action_model': 6.220693777782375e-05, 'epoch': 0.96}) +[2026-06-16 00:14:53,553][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.17526774108409882, 'train/grad_norm_pre_clip': 6.350588321685791, 'timing/data': 0.0005087380413897336, 'timing/model': 1.6612060170155019, 'learning_rate/qwen_vl_interface': 6.21771242603611e-06, 'learning_rate/action_model': 6.21771242603611e-05, 'epoch': 0.96}) +[2026-06-16 00:14:59,256][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.07417531311511993, 'train/grad_norm_pre_clip': 6.644535064697266, 'timing/data': 0.0004939970094710588, 'timing/model': 1.4655093850451522, 'learning_rate/qwen_vl_interface': 6.21473067649982e-06, 'learning_rate/action_model': 6.21473067649982e-05, 'epoch': 0.96}) +[2026-06-16 00:15:05,255][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.11354503035545349, 'train/grad_norm_pre_clip': 8.80631160736084, 'timing/data': 0.000502929964568466, 'timing/model': 1.463970665994566, 'learning_rate/qwen_vl_interface': 6.211748530399189e-06, 'learning_rate/action_model': 6.211748530399189e-05, 'epoch': 0.96}) +[2026-06-16 00:15:11,326][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.15708231925964355, 'train/grad_norm_pre_clip': 8.768549919128418, 'timing/data': 0.0005045930156484246, 'timing/model': 1.5404407589812763, 'learning_rate/qwen_vl_interface': 6.208765988960067e-06, 'learning_rate/action_model': 6.208765988960066e-05, 'epoch': 0.96}) +[2026-06-16 00:15:17,362][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.17376859486103058, 'train/grad_norm_pre_clip': 9.291070938110352, 'timing/data': 0.0006094769923947752, 'timing/model': 1.5153625019593164, 'learning_rate/qwen_vl_interface': 6.2057830534084615e-06, 'learning_rate/action_model': 6.205783053408461e-05, 'epoch': 0.96}) +[2026-06-16 00:15:23,344][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.23109610378742218, 'train/grad_norm_pre_clip': 11.53671932220459, 'timing/data': 0.0006180899799801409, 'timing/model': 1.7525369680370204, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 0.96}) +[2026-06-16 00:15:29,171][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.10178456455469131, 'train/grad_norm_pre_clip': 5.041715145111084, 'timing/data': 0.0005280569894239306, 'timing/model': 1.5795872160233557, 'learning_rate/qwen_vl_interface': 6.199816004872655e-06, 'learning_rate/action_model': 6.199816004872655e-05, 'epoch': 0.96}) +[2026-06-16 00:15:34,855][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.22767676413059235, 'train/grad_norm_pre_clip': 8.334772109985352, 'timing/data': 0.000525693001691252, 'timing/model': 1.464754375978373, 'learning_rate/qwen_vl_interface': 6.196831894341281e-06, 'learning_rate/action_model': 6.19683189434128e-05, 'epoch': 0.96}) +[2026-06-16 00:15:40,855][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.039599087089300156, 'train/grad_norm_pre_clip': 3.0193536281585693, 'timing/data': 0.0005221680039539933, 'timing/model': 1.4559730209875852, 'learning_rate/qwen_vl_interface': 6.1938473946030795e-06, 'learning_rate/action_model': 6.193847394603079e-05, 'epoch': 0.96}) +[2026-06-16 00:15:46,869][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.11949217319488525, 'train/grad_norm_pre_clip': 9.538009643554688, 'timing/data': 0.0006073139957152307, 'timing/model': 1.489228375023231, 'learning_rate/qwen_vl_interface': 6.1908625068848695e-06, 'learning_rate/action_model': 6.19086250688487e-05, 'epoch': 0.96}) +[2026-06-16 00:15:52,861][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.20609210431575775, 'train/grad_norm_pre_clip': 5.1779069900512695, 'timing/data': 0.0005412559839896858, 'timing/model': 1.5211327510187402, 'learning_rate/qwen_vl_interface': 6.187877232413619e-06, 'learning_rate/action_model': 6.18787723241362e-05, 'epoch': 0.96}) +[2026-06-16 00:15:58,840][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.2227705419063568, 'train/grad_norm_pre_clip': 11.96114444732666, 'timing/data': 0.0006265119882300496, 'timing/model': 1.7513589189620689, 'learning_rate/qwen_vl_interface': 6.184891572416467e-06, 'learning_rate/action_model': 6.184891572416466e-05, 'epoch': 0.96}) +[2026-06-16 00:16:04,551][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.134028360247612, 'train/grad_norm_pre_clip': 3.8878629207611084, 'timing/data': 0.0004991840105503798, 'timing/model': 1.4633033479913138, 'learning_rate/qwen_vl_interface': 6.181905528120703e-06, 'learning_rate/action_model': 6.181905528120703e-05, 'epoch': 0.96}) +[2026-06-16 00:16:10,552][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.1702772080898285, 'train/grad_norm_pre_clip': 8.618616104125977, 'timing/data': 0.0005334240267984569, 'timing/model': 1.4631367459660396, 'learning_rate/qwen_vl_interface': 6.178919100753778e-06, 'learning_rate/action_model': 6.178919100753778e-05, 'epoch': 0.97}) +[2026-06-16 00:16:16,452][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.23282736539840698, 'train/grad_norm_pre_clip': 9.060361862182617, 'timing/data': 0.0005288379616104066, 'timing/model': 1.4632781769614667, 'learning_rate/qwen_vl_interface': 6.175932291543299e-06, 'learning_rate/action_model': 6.175932291543298e-05, 'epoch': 0.97}) +[2026-06-16 00:16:22,357][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.04192778095602989, 'train/grad_norm_pre_clip': 2.127338409423828, 'timing/data': 0.0004909420385956764, 'timing/model': 1.4667122919927351, 'learning_rate/qwen_vl_interface': 6.172945101717029e-06, 'learning_rate/action_model': 6.17294510171703e-05, 'epoch': 0.97}) +[2026-06-16 00:16:28,352][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.01606328971683979, 'train/grad_norm_pre_clip': 0.8875933885574341, 'timing/data': 0.0005268949898891151, 'timing/model': 1.5121040290105157, 'learning_rate/qwen_vl_interface': 6.169957532502892e-06, 'learning_rate/action_model': 6.169957532502892e-05, 'epoch': 0.97}) +[2026-06-16 00:16:34,309][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.15287449955940247, 'train/grad_norm_pre_clip': 6.336199760437012, 'timing/data': 0.0005309410043992102, 'timing/model': 1.7045492409961298, 'learning_rate/qwen_vl_interface': 6.166969585128962e-06, 'learning_rate/action_model': 6.166969585128962e-05, 'epoch': 0.97}) +[2026-06-16 00:16:39,951][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.08384779095649719, 'train/grad_norm_pre_clip': 2.807673931121826, 'timing/data': 0.000508768018335104, 'timing/model': 1.4641221400233917, 'learning_rate/qwen_vl_interface': 6.163981260823473e-06, 'learning_rate/action_model': 6.163981260823473e-05, 'epoch': 0.97}) +[2026-06-16 00:16:45,952][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.1468426138162613, 'train/grad_norm_pre_clip': 6.486851692199707, 'timing/data': 0.0005993120139464736, 'timing/model': 1.464870605035685, 'learning_rate/qwen_vl_interface': 6.16099256081481e-06, 'learning_rate/action_model': 6.16099256081481e-05, 'epoch': 0.97}) +[2026-06-16 00:16:51,750][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.04200487956404686, 'train/grad_norm_pre_clip': 5.198221206665039, 'timing/data': 0.000542809022590518, 'timing/model': 1.4606899779755622, 'learning_rate/qwen_vl_interface': 6.15800348633152e-06, 'learning_rate/action_model': 6.15800348633152e-05, 'epoch': 0.97}) +[2026-06-16 00:16:57,554][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.06911104172468185, 'train/grad_norm_pre_clip': 6.762024402618408, 'timing/data': 0.0005700379842892289, 'timing/model': 1.466161300020758, 'learning_rate/qwen_vl_interface': 6.155014038602293e-06, 'learning_rate/action_model': 6.155014038602292e-05, 'epoch': 0.97}) +[2026-06-16 00:17:03,401][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.07821764796972275, 'train/grad_norm_pre_clip': 6.322823524475098, 'timing/data': 0.000566272996366024, 'timing/model': 1.4729424369870685, 'learning_rate/qwen_vl_interface': 6.152024218855979e-06, 'learning_rate/action_model': 6.15202421885598e-05, 'epoch': 0.97}) +[2026-06-16 00:17:09,227][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.30637118220329285, 'train/grad_norm_pre_clip': 12.909440994262695, 'timing/data': 0.0006616439786739647, 'timing/model': 1.638516051985789, 'learning_rate/qwen_vl_interface': 6.149034028321583e-06, 'learning_rate/action_model': 6.149034028321583e-05, 'epoch': 0.97}) +[2026-06-16 00:17:14,953][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.2162204384803772, 'train/grad_norm_pre_clip': 11.198237419128418, 'timing/data': 0.0005328640108928084, 'timing/model': 1.4625206919736229, 'learning_rate/qwen_vl_interface': 6.1460434682282566e-06, 'learning_rate/action_model': 6.146043468228256e-05, 'epoch': 0.97}) +[2026-06-16 00:17:20,851][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.06626340746879578, 'train/grad_norm_pre_clip': 3.9910573959350586, 'timing/data': 0.0005417169886641204, 'timing/model': 1.4612074629985727, 'learning_rate/qwen_vl_interface': 6.143052539805308e-06, 'learning_rate/action_model': 6.143052539805308e-05, 'epoch': 0.97}) +[2026-06-16 00:17:26,853][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.12370719760656357, 'train/grad_norm_pre_clip': 11.694432258605957, 'timing/data': 0.0005190839874558151, 'timing/model': 1.4627017739694566, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 0.97}) +[2026-06-16 00:17:32,875][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.13548985123634338, 'train/grad_norm_pre_clip': 7.268789291381836, 'timing/data': 0.0005333450390025973, 'timing/model': 1.5206393579719588, 'learning_rate/qwen_vl_interface': 6.137069582888521e-06, 'learning_rate/action_model': 6.137069582888521e-05, 'epoch': 0.97}) +[2026-06-16 00:17:38,810][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.04725239425897598, 'train/grad_norm_pre_clip': 5.921642303466797, 'timing/data': 0.0005933130159974098, 'timing/model': 1.7033896219800226, 'learning_rate/qwen_vl_interface': 6.13407755685405e-06, 'learning_rate/action_model': 6.13407755685405e-05, 'epoch': 0.97}) +[2026-06-16 00:17:44,646][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.0640852078795433, 'train/grad_norm_pre_clip': 2.047170639038086, 'timing/data': 0.0007838329765945673, 'timing/model': 1.6431112199788913, 'learning_rate/qwen_vl_interface': 6.131085167408692e-06, 'learning_rate/action_model': 6.131085167408692e-05, 'epoch': 0.97}) +[2026-06-16 00:17:50,353][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.14038023352622986, 'train/grad_norm_pre_clip': 20.785398483276367, 'timing/data': 0.0005160889704711735, 'timing/model': 1.4640572080388665, 'learning_rate/qwen_vl_interface': 6.128092415782502e-06, 'learning_rate/action_model': 6.128092415782501e-05, 'epoch': 0.97}) +[2026-06-16 00:17:56,357][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.217996746301651, 'train/grad_norm_pre_clip': 8.695809364318848, 'timing/data': 0.0006018750136718154, 'timing/model': 1.4695642919978127, 'learning_rate/qwen_vl_interface': 6.125099303205688e-06, 'learning_rate/action_model': 6.125099303205688e-05, 'epoch': 0.97}) +[2026-06-16 00:18:02,351][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.21712790429592133, 'train/grad_norm_pre_clip': 8.613247871398926, 'timing/data': 0.0005274760187603533, 'timing/model': 1.4686434600153007, 'learning_rate/qwen_vl_interface': 6.122105830908607e-06, 'learning_rate/action_model': 6.122105830908606e-05, 'epoch': 0.97}) +[2026-06-16 00:18:08,335][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.05486099421977997, 'train/grad_norm_pre_clip': 3.929053544998169, 'timing/data': 0.0005000949604436755, 'timing/model': 1.501985971990507, 'learning_rate/qwen_vl_interface': 6.11911200012176e-06, 'learning_rate/action_model': 6.11911200012176e-05, 'epoch': 0.97}) +[2026-06-16 00:18:14,297][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.14570270478725433, 'train/grad_norm_pre_clip': 9.316011428833008, 'timing/data': 0.0005032109911553562, 'timing/model': 1.706842137966305, 'learning_rate/qwen_vl_interface': 6.116117812075799e-06, 'learning_rate/action_model': 6.116117812075799e-05, 'epoch': 0.97}) +[2026-06-16 00:18:20,074][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.05206245929002762, 'train/grad_norm_pre_clip': 3.972456216812134, 'timing/data': 0.0005308610270731151, 'timing/model': 1.5724676940008067, 'learning_rate/qwen_vl_interface': 6.113123268001524e-06, 'learning_rate/action_model': 6.113123268001524e-05, 'epoch': 0.97}) +[2026-06-16 00:18:25,755][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.045857541263103485, 'train/grad_norm_pre_clip': 3.63107967376709, 'timing/data': 0.0005262739723548293, 'timing/model': 1.4663983470527455, 'learning_rate/qwen_vl_interface': 6.110128369129873e-06, 'learning_rate/action_model': 6.110128369129873e-05, 'epoch': 0.98}) +[2026-06-16 00:18:31,755][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.1185908168554306, 'train/grad_norm_pre_clip': 5.325497627258301, 'timing/data': 0.0005223590414971113, 'timing/model': 1.4661669389461167, 'learning_rate/qwen_vl_interface': 6.107133116691943e-06, 'learning_rate/action_model': 6.107133116691942e-05, 'epoch': 0.98}) +[2026-06-16 00:18:38,054][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.050919052213430405, 'train/grad_norm_pre_clip': 4.314103603363037, 'timing/data': 0.0005293089780025184, 'timing/model': 1.4372328169993125, 'learning_rate/qwen_vl_interface': 6.104137511918963e-06, 'learning_rate/action_model': 6.104137511918963e-05, 'epoch': 0.98}) +[2026-06-16 00:18:44,048][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.022608965635299683, 'train/grad_norm_pre_clip': 1.5266985893249512, 'timing/data': 0.0005183820030651987, 'timing/model': 1.5077255310025066, 'learning_rate/qwen_vl_interface': 6.1011415560423155e-06, 'learning_rate/action_model': 6.101141556042316e-05, 'epoch': 0.98}) +[2026-06-16 00:18:50,009][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.07844001799821854, 'train/grad_norm_pre_clip': 4.304502964019775, 'timing/data': 0.0005038609961047769, 'timing/model': 1.7223100200062618, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.98}) +[2026-06-16 00:18:55,653][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.09692450612783432, 'train/grad_norm_pre_clip': 11.794503211975098, 'timing/data': 0.0005190939991734922, 'timing/model': 1.463600003044121, 'learning_rate/qwen_vl_interface': 6.09514859590426e-06, 'learning_rate/action_model': 6.09514859590426e-05, 'epoch': 0.98}) +[2026-06-16 00:19:01,656][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.05566869676113129, 'train/grad_norm_pre_clip': 4.5732221603393555, 'timing/data': 0.0005263640196062624, 'timing/model': 1.4647327910060994, 'learning_rate/qwen_vl_interface': 6.09215159410633e-06, 'learning_rate/action_model': 6.09215159410633e-05, 'epoch': 0.98}) +[2026-06-16 00:19:07,553][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.22689004242420197, 'train/grad_norm_pre_clip': 12.843826293945312, 'timing/data': 0.0005922410055063665, 'timing/model': 1.464107353007421, 'learning_rate/qwen_vl_interface': 6.089154246131692e-06, 'learning_rate/action_model': 6.089154246131692e-05, 'epoch': 0.98}) +[2026-06-16 00:19:13,463][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.04889027401804924, 'train/grad_norm_pre_clip': 4.880247592926025, 'timing/data': 0.0004967309650965035, 'timing/model': 1.4717076069791801, 'learning_rate/qwen_vl_interface': 6.08615655321244e-06, 'learning_rate/action_model': 6.08615655321244e-05, 'epoch': 0.98}) +[2026-06-16 00:19:19,452][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.21582114696502686, 'train/grad_norm_pre_clip': 10.07778549194336, 'timing/data': 0.0006131720147095621, 'timing/model': 1.5108455349691212, 'learning_rate/qwen_vl_interface': 6.083158516580814e-06, 'learning_rate/action_model': 6.083158516580814e-05, 'epoch': 0.98}) +[2026-06-16 00:19:25,397][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.06767485290765762, 'train/grad_norm_pre_clip': 6.933276653289795, 'timing/data': 0.0005199849838390946, 'timing/model': 1.7038969779969193, 'learning_rate/qwen_vl_interface': 6.080160137469192e-06, 'learning_rate/action_model': 6.0801601374691916e-05, 'epoch': 0.98}) +[2026-06-16 00:19:31,055][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.21888116002082825, 'train/grad_norm_pre_clip': 14.135056495666504, 'timing/data': 0.0005182719905860722, 'timing/model': 1.4651578270131722, 'learning_rate/qwen_vl_interface': 6.077161417110097e-06, 'learning_rate/action_model': 6.0771614171100965e-05, 'epoch': 0.98}) +[2026-06-16 00:19:37,054][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.2636873126029968, 'train/grad_norm_pre_clip': 11.071456909179688, 'timing/data': 0.0007264890009537339, 'timing/model': 1.4661238310509361, 'learning_rate/qwen_vl_interface': 6.0741623567361864e-06, 'learning_rate/action_model': 6.074162356736186e-05, 'epoch': 0.98}) +[2026-06-16 00:19:42,856][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.06886295974254608, 'train/grad_norm_pre_clip': 7.544243812561035, 'timing/data': 0.0005786409601569176, 'timing/model': 1.4669124229694717, 'learning_rate/qwen_vl_interface': 6.071162957580262e-06, 'learning_rate/action_model': 6.071162957580262e-05, 'epoch': 0.98}) +[2026-06-16 00:19:48,651][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.10221757739782333, 'train/grad_norm_pre_clip': 8.959202766418457, 'timing/data': 0.0005174420075491071, 'timing/model': 1.4544615779886954, 'learning_rate/qwen_vl_interface': 6.068163220875266e-06, 'learning_rate/action_model': 6.068163220875266e-05, 'epoch': 0.98}) +[2026-06-16 00:19:54,456][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.07040177285671234, 'train/grad_norm_pre_clip': 6.997915267944336, 'timing/data': 0.0004991139867343009, 'timing/model': 1.431696470011957, 'learning_rate/qwen_vl_interface': 6.065163147854273e-06, 'learning_rate/action_model': 6.065163147854273e-05, 'epoch': 0.98}) +[2026-06-16 00:20:00,324][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.11638669669628143, 'train/grad_norm_pre_clip': 10.999093055725098, 'timing/data': 0.0005410159938037395, 'timing/model': 1.6353923039860092, 'learning_rate/qwen_vl_interface': 6.062162739750502e-06, 'learning_rate/action_model': 6.062162739750502e-05, 'epoch': 0.98}) +[2026-06-16 00:20:06,057][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.11533249914646149, 'train/grad_norm_pre_clip': 5.364002227783203, 'timing/data': 0.0005209269584156573, 'timing/model': 1.4641626550001092, 'learning_rate/qwen_vl_interface': 6.059161997797306e-06, 'learning_rate/action_model': 6.0591619977973056e-05, 'epoch': 0.98}) +[2026-06-16 00:20:11,956][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.08696065098047256, 'train/grad_norm_pre_clip': 16.216779708862305, 'timing/data': 0.0005122429574839771, 'timing/model': 1.4635766250430606, 'learning_rate/qwen_vl_interface': 6.056160923228178e-06, 'learning_rate/action_model': 6.056160923228178e-05, 'epoch': 0.98}) +[2026-06-16 00:20:17,953][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.04555269330739975, 'train/grad_norm_pre_clip': 4.343619346618652, 'timing/data': 0.0005249219830147922, 'timing/model': 1.4566211650380865, 'learning_rate/qwen_vl_interface': 6.0531595172767455e-06, 'learning_rate/action_model': 6.0531595172767455e-05, 'epoch': 0.98}) +[2026-06-16 00:20:23,855][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.02582077495753765, 'train/grad_norm_pre_clip': 2.6325361728668213, 'timing/data': 0.0005201649619266391, 'timing/model': 1.429059160000179, 'learning_rate/qwen_vl_interface': 6.050157781176774e-06, 'learning_rate/action_model': 6.050157781176774e-05, 'epoch': 0.98}) +[2026-06-16 00:20:29,791][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.07461925595998764, 'train/grad_norm_pre_clip': 5.1358323097229, 'timing/data': 0.0004938970087096095, 'timing/model': 1.7043057049741037, 'learning_rate/qwen_vl_interface': 6.047155716162164e-06, 'learning_rate/action_model': 6.047155716162164e-05, 'epoch': 0.98}) +[2026-06-16 00:20:35,648][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.010508879087865353, 'train/grad_norm_pre_clip': 0.9820321798324585, 'timing/data': 0.0005220780149102211, 'timing/model': 1.6558975519728847, 'learning_rate/qwen_vl_interface': 6.044153323466948e-06, 'learning_rate/action_model': 6.044153323466948e-05, 'epoch': 0.98}) +[2026-06-16 00:20:41,357][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.07014618813991547, 'train/grad_norm_pre_clip': 5.598041534423828, 'timing/data': 0.0005169909563846886, 'timing/model': 1.4655420869821683, 'learning_rate/qwen_vl_interface': 6.0411506043253e-06, 'learning_rate/action_model': 6.0411506043253e-05, 'epoch': 0.98}) +[2026-06-16 00:20:47,453][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.2973547577857971, 'train/grad_norm_pre_clip': 14.249641418457031, 'timing/data': 0.0005179920117370784, 'timing/model': 1.4630118960049003, 'learning_rate/qwen_vl_interface': 6.038147559971523e-06, 'learning_rate/action_model': 6.038147559971523e-05, 'epoch': 0.99}) +[2026-06-16 00:20:53,460][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.037853218615055084, 'train/grad_norm_pre_clip': 7.047621250152588, 'timing/data': 0.0005292990244925022, 'timing/model': 1.4885129040339962, 'learning_rate/qwen_vl_interface': 6.035144191640055e-06, 'learning_rate/action_model': 6.035144191640055e-05, 'epoch': 0.99}) +[2026-06-16 00:20:59,383][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.02383577451109886, 'train/grad_norm_pre_clip': 4.869434356689453, 'timing/data': 0.000508729019202292, 'timing/model': 1.4542493289918639, 'learning_rate/qwen_vl_interface': 6.032140500565467e-06, 'learning_rate/action_model': 6.0321405005654665e-05, 'epoch': 0.99}) +[2026-06-16 00:21:05,289][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.18701091408729553, 'train/grad_norm_pre_clip': 26.20323371887207, 'timing/data': 0.0005007770378142595, 'timing/model': 1.7014925690018572, 'learning_rate/qwen_vl_interface': 6.029136487982463e-06, 'learning_rate/action_model': 6.029136487982463e-05, 'epoch': 0.99}) +[2026-06-16 00:21:11,059][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.05984314903616905, 'train/grad_norm_pre_clip': 6.289840221405029, 'timing/data': 0.0005016479990445077, 'timing/model': 1.5599150320049375, 'learning_rate/qwen_vl_interface': 6.02613215512588e-06, 'learning_rate/action_model': 6.0261321551258796e-05, 'epoch': 0.99}) +[2026-06-16 00:21:16,751][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.3625571131706238, 'train/grad_norm_pre_clip': 20.000896453857422, 'timing/data': 0.0005031710024923086, 'timing/model': 1.4631020879605785, 'learning_rate/qwen_vl_interface': 6.023127503230684e-06, 'learning_rate/action_model': 6.023127503230683e-05, 'epoch': 0.99}) +[2026-06-16 00:21:22,754][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.23696479201316833, 'train/grad_norm_pre_clip': 15.684576034545898, 'timing/data': 0.0005219380254857242, 'timing/model': 1.4640729180537164, 'learning_rate/qwen_vl_interface': 6.020122533531976e-06, 'learning_rate/action_model': 6.0201225335319755e-05, 'epoch': 0.99}) +[2026-06-16 00:21:28,652][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.02292952686548233, 'train/grad_norm_pre_clip': 1.847697377204895, 'timing/data': 0.0005465740105137229, 'timing/model': 1.4539014759939164, 'learning_rate/qwen_vl_interface': 6.017117247264983e-06, 'learning_rate/action_model': 6.017117247264983e-05, 'epoch': 0.99}) +[2026-06-16 00:21:34,625][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.20168536901474, 'train/grad_norm_pre_clip': 9.876060485839844, 'timing/data': 0.0005326140089891851, 'timing/model': 1.4974081219988875, 'learning_rate/qwen_vl_interface': 6.014111645665069e-06, 'learning_rate/action_model': 6.0141116456650684e-05, 'epoch': 0.99}) +[2026-06-16 00:21:40,589][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.2543812692165375, 'train/grad_norm_pre_clip': 17.36271095275879, 'timing/data': 0.0005189729854464531, 'timing/model': 1.6998763129813597, 'learning_rate/qwen_vl_interface': 6.011105729967719e-06, 'learning_rate/action_model': 6.0111057299677184e-05, 'epoch': 0.99}) +[2026-06-16 00:31:05,554][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.005226734559983015, 'train/grad_norm_pre_clip': 0.8625870943069458, 'eval/loss': 0.06703925877809525, 'eval/flap_precision': 0.9015444015444015, 'eval/flap_recall': 0.935871743486974, 'eval/flap_f1': 0.9183874139626352, 'eval/flap_rate_model': 0.1438888888888889, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9183874139626352, 'timing/data': 0.0005225890199653804, 'timing/model': 1.4645965049858205, 'learning_rate/qwen_vl_interface': 6.008099501408552e-06, 'learning_rate/action_model': 6.008099501408552e-05, 'epoch': 0.99}) +[2026-06-16 00:31:05,554][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2300_state +[2026-06-16 00:31:05,559][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 00:31:11,196][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2300_state/model.safetensors +[2026-06-16 00:31:24,891][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2300_state/optimizer.bin +[2026-06-16 00:31:24,892][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2300_state/sampler.bin +[2026-06-16 00:31:24,892][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2300_state/sampler_1.bin +[2026-06-16 00:31:24,893][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2300_state/random_states_0.pkl +[2026-06-16 00:31:24,998][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2300 +[2026-06-16 00:31:24,998][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 00:31:25,008][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 00:31:30,920][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.010788158513605595, 'train/grad_norm_pre_clip': 1.574410080909729, 'timing/data': 0.0005224689957685769, 'timing/model': 1.7449731249944307, 'learning_rate/qwen_vl_interface': 6.0050929612233165e-06, 'learning_rate/action_model': 6.005092961223316e-05, 'epoch': 0.99}) +[2026-06-16 00:31:36,544][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.006059693638235331, 'train/grad_norm_pre_clip': 1.404695749282837, 'timing/data': 0.0005195250269025564, 'timing/model': 1.4672001220169477, 'learning_rate/qwen_vl_interface': 6.002086110647888e-06, 'learning_rate/action_model': 6.002086110647888e-05, 'epoch': 0.99}) +[2026-06-16 00:31:42,543][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.10308841615915298, 'train/grad_norm_pre_clip': 8.844024658203125, 'timing/data': 0.0005134450038895011, 'timing/model': 1.4642518349573947, 'learning_rate/qwen_vl_interface': 5.999078950918267e-06, 'learning_rate/action_model': 5.999078950918267e-05, 'epoch': 0.99}) +[2026-06-16 00:31:48,551][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.29035985469818115, 'train/grad_norm_pre_clip': 21.328575134277344, 'timing/data': 0.0005240209866315126, 'timing/model': 1.4652962620020844, 'learning_rate/qwen_vl_interface': 5.996071483270585e-06, 'learning_rate/action_model': 5.996071483270585e-05, 'epoch': 0.99}) +[2026-06-16 00:31:54,545][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.06353136152029037, 'train/grad_norm_pre_clip': 5.673633098602295, 'timing/data': 0.0005799930077046156, 'timing/model': 1.4587402369943447, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.99}) +[2026-06-16 00:32:00,533][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.32891154289245605, 'train/grad_norm_pre_clip': 20.147586822509766, 'timing/data': 0.0006164569640532136, 'timing/model': 1.5024618220049888, 'learning_rate/qwen_vl_interface': 5.990055629166187e-06, 'learning_rate/action_model': 5.990055629166187e-05, 'epoch': 0.99}) +[2026-06-16 00:32:06,567][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.02937999553978443, 'train/grad_norm_pre_clip': 7.833500385284424, 'timing/data': 0.0005558180273510516, 'timing/model': 1.779261793999467, 'learning_rate/qwen_vl_interface': 5.987047245182363e-06, 'learning_rate/action_model': 5.987047245182362e-05, 'epoch': 0.99}) +[2026-06-16 00:32:12,455][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.23166772723197937, 'train/grad_norm_pre_clip': 13.962175369262695, 'timing/data': 0.0006548530072905123, 'timing/model': 1.4810300219687633, 'learning_rate/qwen_vl_interface': 5.984038558226255e-06, 'learning_rate/action_model': 5.9840385582262546e-05, 'epoch': 0.99}) +[2026-06-16 00:32:18,458][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.5201386213302612, 'train/grad_norm_pre_clip': 32.5049934387207, 'timing/data': 0.000635245000012219, 'timing/model': 1.4693318619974889, 'learning_rate/qwen_vl_interface': 5.981029569534625e-06, 'learning_rate/action_model': 5.981029569534625e-05, 'epoch': 0.99}) +[2026-06-16 00:32:24,358][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.07943478971719742, 'train/grad_norm_pre_clip': 4.574386119842529, 'timing/data': 0.000590297975577414, 'timing/model': 1.4648636690108106, 'learning_rate/qwen_vl_interface': 5.978020280344351e-06, 'learning_rate/action_model': 5.9780202803443506e-05, 'epoch': 0.99}) +[2026-06-16 00:32:30,380][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.10272327810525894, 'train/grad_norm_pre_clip': 16.42928695678711, 'timing/data': 0.0006137730088084936, 'timing/model': 1.5135403930325992, 'learning_rate/qwen_vl_interface': 5.97501069189244e-06, 'learning_rate/action_model': 5.97501069189244e-05, 'epoch': 0.99}) +[2026-06-16 00:32:36,340][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.3427891433238983, 'train/grad_norm_pre_clip': 14.090019226074219, 'timing/data': 0.0006203730008564889, 'timing/model': 1.5067608889658004, 'learning_rate/qwen_vl_interface': 5.97200080541602e-06, 'learning_rate/action_model': 5.97200080541602e-05, 'epoch': 0.99}) +[2026-06-16 00:32:42,296][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.10778312385082245, 'train/grad_norm_pre_clip': 6.588142395019531, 'timing/data': 0.0005200449959374964, 'timing/model': 1.7037538789445534, 'learning_rate/qwen_vl_interface': 5.968990622152341e-06, 'learning_rate/action_model': 5.9689906221523404e-05, 'epoch': 1.0}) +[2026-06-16 00:32:47,958][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.03386847302317619, 'train/grad_norm_pre_clip': 4.213405132293701, 'timing/data': 0.0005202259635552764, 'timing/model': 1.4641991430544294, 'learning_rate/qwen_vl_interface': 5.965980143338774e-06, 'learning_rate/action_model': 5.965980143338774e-05, 'epoch': 1.0}) +[2026-06-16 00:32:53,958][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.07322284579277039, 'train/grad_norm_pre_clip': 10.59589958190918, 'timing/data': 0.0005291089764796197, 'timing/model': 1.4644795560161583, 'learning_rate/qwen_vl_interface': 5.962969370212819e-06, 'learning_rate/action_model': 5.9629693702128185e-05, 'epoch': 1.0}) +[2026-06-16 00:32:59,957][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.06058310717344284, 'train/grad_norm_pre_clip': 4.992276668548584, 'timing/data': 0.0005174920079298317, 'timing/model': 1.483563733054325, 'learning_rate/qwen_vl_interface': 5.959958304012085e-06, 'learning_rate/action_model': 5.9599583040120845e-05, 'epoch': 1.0}) +[2026-06-16 00:33:05,857][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.06213586777448654, 'train/grad_norm_pre_clip': 6.588716506958008, 'timing/data': 0.000515379011631012, 'timing/model': 1.4460181759786792, 'learning_rate/qwen_vl_interface': 5.956946945974309e-06, 'learning_rate/action_model': 5.9569469459743085e-05, 'epoch': 1.0}) +[2026-06-16 00:33:12,065][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.10944268107414246, 'train/grad_norm_pre_clip': 7.276947975158691, 'timing/data': 0.000604950007982552, 'timing/model': 1.634254943986889, 'learning_rate/qwen_vl_interface': 5.953935297337347e-06, 'learning_rate/action_model': 5.953935297337347e-05, 'epoch': 1.0}) +[2026-06-16 00:33:17,955][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.16800004243850708, 'train/grad_norm_pre_clip': 10.690047264099121, 'timing/data': 0.0006098379963077605, 'timing/model': 1.6637051619472913, 'learning_rate/qwen_vl_interface': 5.950923359339175e-06, 'learning_rate/action_model': 5.950923359339174e-05, 'epoch': 1.0}) +[2026-06-16 00:33:23,756][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.06979838013648987, 'train/grad_norm_pre_clip': 10.103259086608887, 'timing/data': 0.0005315520102158189, 'timing/model': 1.4614549279795028, 'learning_rate/qwen_vl_interface': 5.947911133217886e-06, 'learning_rate/action_model': 5.947911133217886e-05, 'epoch': 1.0}) +[2026-06-16 00:33:29,756][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.06103694811463356, 'train/grad_norm_pre_clip': 4.178344249725342, 'timing/data': 0.0005285170045681298, 'timing/model': 1.4657287660520524, 'learning_rate/qwen_vl_interface': 5.944898620211693e-06, 'learning_rate/action_model': 5.944898620211693e-05, 'epoch': 1.0}) +[2026-06-16 00:33:35,761][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.33588531613349915, 'train/grad_norm_pre_clip': 16.26631736755371, 'timing/data': 0.0005310510168783367, 'timing/model': 1.479202838963829, 'learning_rate/qwen_vl_interface': 5.941885821558925e-06, 'learning_rate/action_model': 5.9418858215589245e-05, 'epoch': 1.0}) +[2026-06-16 00:33:41,732][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.16719935834407806, 'train/grad_norm_pre_clip': 8.226049423217773, 'timing/data': 0.0005113920196890831, 'timing/model': 1.5037123099900782, 'learning_rate/qwen_vl_interface': 5.938872738498031e-06, 'learning_rate/action_model': 5.93887273849803e-05, 'epoch': 1.0}) +[2026-06-16 00:33:47,741][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.11761672794818878, 'train/grad_norm_pre_clip': 11.497505187988281, 'timing/data': 0.000512423983309418, 'timing/model': 1.5095235090120696, 'learning_rate/qwen_vl_interface': 5.935859372267575e-06, 'learning_rate/action_model': 5.935859372267575e-05, 'epoch': 1.0}) +[2026-06-16 00:33:53,759][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.21279948949813843, 'train/grad_norm_pre_clip': 15.31335163116455, 'timing/data': 0.007707259035669267, 'timing/model': 1.4556163999950513, 'learning_rate/qwen_vl_interface': 5.9328457241062396e-06, 'learning_rate/action_model': 5.9328457241062394e-05, 'epoch': 1.0}) +[2026-06-16 00:33:59,763][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.15161138772964478, 'train/grad_norm_pre_clip': 11.509614944458008, 'timing/data': 0.007391995983198285, 'timing/model': 1.46235141903162, 'learning_rate/qwen_vl_interface': 5.929831795252819e-06, 'learning_rate/action_model': 5.9298317952528184e-05, 'epoch': 1.0}) +[2026-06-16 00:34:05,661][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.0476241298019886, 'train/grad_norm_pre_clip': 4.954720973968506, 'timing/data': 0.006765723985154182, 'timing/model': 1.4594545470317826, 'learning_rate/qwen_vl_interface': 5.926817586946227e-06, 'learning_rate/action_model': 5.926817586946227e-05, 'epoch': 1.0}) +[2026-06-16 00:34:11,661][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.047084446996450424, 'train/grad_norm_pre_clip': 3.6922714710235596, 'timing/data': 0.0066710939863696694, 'timing/model': 1.46885993599426, 'learning_rate/qwen_vl_interface': 5.923803100425491e-06, 'learning_rate/action_model': 5.92380310042549e-05, 'epoch': 1.0}) +[2026-06-16 00:34:17,676][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.013421139679849148, 'train/grad_norm_pre_clip': 3.8442015647888184, 'timing/data': 0.006982783961575478, 'timing/model': 1.5326608609757386, 'learning_rate/qwen_vl_interface': 5.92078833692975e-06, 'learning_rate/action_model': 5.92078833692975e-05, 'epoch': 1.0}) +[2026-06-16 00:34:23,588][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.04863889142870903, 'train/grad_norm_pre_clip': 2.2638397216796875, 'timing/data': 0.006887613038998097, 'timing/model': 1.6895572409848683, 'learning_rate/qwen_vl_interface': 5.917773297698264e-06, 'learning_rate/action_model': 5.9177732976982636e-05, 'epoch': 1.0}) +[2026-06-16 00:34:29,477][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.1382194459438324, 'train/grad_norm_pre_clip': 13.805866241455078, 'timing/data': 0.0064466140465810895, 'timing/model': 1.6791793429874815, 'learning_rate/qwen_vl_interface': 5.9147579839704e-06, 'learning_rate/action_model': 5.9147579839704e-05, 'epoch': 1.0}) +[2026-06-16 00:34:35,161][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.1007516011595726, 'train/grad_norm_pre_clip': 5.299160003662109, 'timing/data': 0.007213601958937943, 'timing/model': 1.4581871050177142, 'learning_rate/qwen_vl_interface': 5.911742396985637e-06, 'learning_rate/action_model': 5.9117423969856364e-05, 'epoch': 1.0}) +[2026-06-16 00:34:41,067][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.3228911757469177, 'train/grad_norm_pre_clip': 15.101585388183594, 'timing/data': 0.007153294049203396, 'timing/model': 1.4645536220050417, 'learning_rate/qwen_vl_interface': 5.908726537983571e-06, 'learning_rate/action_model': 5.908726537983571e-05, 'epoch': 1.0}) +[2026-06-16 00:34:47,057][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.13654035329818726, 'train/grad_norm_pre_clip': 8.994227409362793, 'timing/data': 0.006743291043676436, 'timing/model': 1.4735959160025232, 'learning_rate/qwen_vl_interface': 5.9057104082039085e-06, 'learning_rate/action_model': 5.905710408203908e-05, 'epoch': 1.0}) +[2026-06-16 00:34:53,048][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.11221050471067429, 'train/grad_norm_pre_clip': 10.525768280029297, 'timing/data': 0.00679352501174435, 'timing/model': 1.5106664040358737, 'learning_rate/qwen_vl_interface': 5.902694008886466e-06, 'learning_rate/action_model': 5.902694008886466e-05, 'epoch': 1.0}) +[2026-06-16 00:34:59,013][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.2426348328590393, 'train/grad_norm_pre_clip': 9.296083450317383, 'timing/data': 0.007404153002426028, 'timing/model': 1.7108419879805297, 'learning_rate/qwen_vl_interface': 5.8996773412711694e-06, 'learning_rate/action_model': 5.899677341271169e-05, 'epoch': 1.01}) +[2026-06-16 00:35:04,901][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.12129327654838562, 'train/grad_norm_pre_clip': 10.462933540344238, 'timing/data': 0.006585459050256759, 'timing/model': 1.6966750160208903, 'learning_rate/qwen_vl_interface': 5.896660406598059e-06, 'learning_rate/action_model': 5.896660406598059e-05, 'epoch': 1.01}) +[2026-06-16 00:35:10,560][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.015546487644314766, 'train/grad_norm_pre_clip': 1.470941424369812, 'timing/data': 0.00679679901804775, 'timing/model': 1.4588838020572439, 'learning_rate/qwen_vl_interface': 5.893643206107283e-06, 'learning_rate/action_model': 5.893643206107282e-05, 'epoch': 1.01}) +[2026-06-16 00:35:16,560][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.13164712488651276, 'train/grad_norm_pre_clip': 8.632591247558594, 'timing/data': 0.006546351010911167, 'timing/model': 1.4619545200257562, 'learning_rate/qwen_vl_interface': 5.890625741039096e-06, 'learning_rate/action_model': 5.890625741039095e-05, 'epoch': 1.01}) +[2026-06-16 00:35:22,560][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.18456198275089264, 'train/grad_norm_pre_clip': 8.866795539855957, 'timing/data': 0.006603215006180108, 'timing/model': 1.477831149008125, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 1.01}) +[2026-06-16 00:35:28,548][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.09091709554195404, 'train/grad_norm_pre_clip': 7.735429763793945, 'timing/data': 0.0064420870039612055, 'timing/model': 1.5089984430233017, 'learning_rate/qwen_vl_interface': 5.884590022132064e-06, 'learning_rate/action_model': 5.884590022132064e-05, 'epoch': 1.01}) +[2026-06-16 00:35:34,493][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.10708066076040268, 'train/grad_norm_pre_clip': 12.49078369140625, 'timing/data': 0.0073870979831553996, 'timing/model': 1.6867777860024944, 'learning_rate/qwen_vl_interface': 5.881571770774278e-06, 'learning_rate/action_model': 5.881571770774278e-05, 'epoch': 1.01}) +[2026-06-16 00:35:40,344][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.2486610859632492, 'train/grad_norm_pre_clip': 20.579504013061523, 'timing/data': 0.007194013975095004, 'timing/model': 1.6395332399988547, 'learning_rate/qwen_vl_interface': 5.878553259801192e-06, 'learning_rate/action_model': 5.878553259801192e-05, 'epoch': 1.01}) +[2026-06-16 00:35:46,063][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.04053007811307907, 'train/grad_norm_pre_clip': 2.7017898559570312, 'timing/data': 0.006852902995888144, 'timing/model': 1.4564049969776534, 'learning_rate/qwen_vl_interface': 5.8755344904536025e-06, 'learning_rate/action_model': 5.8755344904536025e-05, 'epoch': 1.01}) +[2026-06-16 00:35:52,063][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.04463353753089905, 'train/grad_norm_pre_clip': 3.2506728172302246, 'timing/data': 0.006862767040729523, 'timing/model': 1.4591811319696717, 'learning_rate/qwen_vl_interface': 5.872515463972415e-06, 'learning_rate/action_model': 5.8725154639724146e-05, 'epoch': 1.01}) +[2026-06-16 00:35:58,070][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.23931410908699036, 'train/grad_norm_pre_clip': 10.96973991394043, 'timing/data': 0.006825441960245371, 'timing/model': 1.4652400159975514, 'learning_rate/qwen_vl_interface': 5.869496181598633e-06, 'learning_rate/action_model': 5.869496181598633e-05, 'epoch': 1.01}) +[2026-06-16 00:36:04,062][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.02935103327035904, 'train/grad_norm_pre_clip': 2.0241050720214844, 'timing/data': 0.006579059001524001, 'timing/model': 1.5155261289910413, 'learning_rate/qwen_vl_interface': 5.866476644573371e-06, 'learning_rate/action_model': 5.866476644573371e-05, 'epoch': 1.01}) +[2026-06-16 00:36:10,148][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.16118435561656952, 'train/grad_norm_pre_clip': 8.19054889678955, 'timing/data': 0.0065278440015390515, 'timing/model': 1.5098880939767696, 'learning_rate/qwen_vl_interface': 5.863456854137848e-06, 'learning_rate/action_model': 5.8634568541378474e-05, 'epoch': 1.01}) +[2026-06-16 00:36:16,164][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.005972592625766993, 'train/grad_norm_pre_clip': 1.2322371006011963, 'timing/data': 0.008105975983198732, 'timing/model': 1.7525015539722517, 'learning_rate/qwen_vl_interface': 5.860436811533383e-06, 'learning_rate/action_model': 5.860436811533383e-05, 'epoch': 1.01}) +[2026-06-16 00:36:21,862][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.06537268310785294, 'train/grad_norm_pre_clip': 3.838071346282959, 'timing/data': 0.006924347020685673, 'timing/model': 1.4575309249921702, 'learning_rate/qwen_vl_interface': 5.857416518001404e-06, 'learning_rate/action_model': 5.857416518001404e-05, 'epoch': 1.01}) +[2026-06-16 00:36:27,869][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.0712161511182785, 'train/grad_norm_pre_clip': 6.641641139984131, 'timing/data': 0.006504218967165798, 'timing/model': 1.4681479059509002, 'learning_rate/qwen_vl_interface': 5.854395974783439e-06, 'learning_rate/action_model': 5.8543959747834386e-05, 'epoch': 1.01}) +[2026-06-16 00:36:33,861][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.22332093119621277, 'train/grad_norm_pre_clip': 11.420291900634766, 'timing/data': 0.006785162957385182, 'timing/model': 1.4768975990009494, 'learning_rate/qwen_vl_interface': 5.851375183121119e-06, 'learning_rate/action_model': 5.8513751831211186e-05, 'epoch': 1.01}) +[2026-06-16 00:36:39,854][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.17080044746398926, 'train/grad_norm_pre_clip': 6.736972332000732, 'timing/data': 0.006532189960125834, 'timing/model': 1.5120932499994524, 'learning_rate/qwen_vl_interface': 5.848354144256175e-06, 'learning_rate/action_model': 5.8483541442561754e-05, 'epoch': 1.01}) +[2026-06-16 00:36:45,818][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.12591199576854706, 'train/grad_norm_pre_clip': 6.089268684387207, 'timing/data': 0.006759105017408729, 'timing/model': 1.7102847440401092, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 1.01}) +[2026-06-16 00:36:51,563][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.1814277470111847, 'train/grad_norm_pre_clip': 19.2852840423584, 'timing/data': 0.006631125987041742, 'timing/model': 1.4585409429855645, 'learning_rate/qwen_vl_interface': 5.842311329885866e-06, 'learning_rate/action_model': 5.842311329885866e-05, 'epoch': 1.01}) +[2026-06-16 00:36:57,563][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.2655601501464844, 'train/grad_norm_pre_clip': 18.96275520324707, 'timing/data': 0.006927831971552223, 'timing/model': 1.45868096902268, 'learning_rate/qwen_vl_interface': 5.839289556864472e-06, 'learning_rate/action_model': 5.8392895568644714e-05, 'epoch': 1.01}) +[2026-06-16 00:37:03,382][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.19559116661548615, 'train/grad_norm_pre_clip': 9.5732421875, 'timing/data': 0.0069569459883496165, 'timing/model': 1.4765300199505873, 'learning_rate/qwen_vl_interface': 5.836267541608398e-06, 'learning_rate/action_model': 5.836267541608398e-05, 'epoch': 1.01}) +[2026-06-16 00:37:09,264][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.11380565166473389, 'train/grad_norm_pre_clip': 6.994560241699219, 'timing/data': 0.006754416972398758, 'timing/model': 1.4597968510352075, 'learning_rate/qwen_vl_interface': 5.833245285359887e-06, 'learning_rate/action_model': 5.833245285359886e-05, 'epoch': 1.01}) +[2026-06-16 00:37:15,164][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.01719795912504196, 'train/grad_norm_pre_clip': 2.131624460220337, 'timing/data': 0.006916765996720642, 'timing/model': 1.4725081730284728, 'learning_rate/qwen_vl_interface': 5.830222789361268e-06, 'learning_rate/action_model': 5.8302227893612684e-05, 'epoch': 1.02}) +[2026-06-16 00:37:21,100][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.027846256271004677, 'train/grad_norm_pre_clip': 4.551137447357178, 'timing/data': 0.006752864981535822, 'timing/model': 1.6931672630016692, 'learning_rate/qwen_vl_interface': 5.8272000548549776e-06, 'learning_rate/action_model': 5.827200054854978e-05, 'epoch': 1.02}) +[2026-06-16 00:37:26,765][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.24666553735733032, 'train/grad_norm_pre_clip': 8.352692604064941, 'timing/data': 0.0067752780159935355, 'timing/model': 1.451616587000899, 'learning_rate/qwen_vl_interface': 5.824177083083547e-06, 'learning_rate/action_model': 5.8241770830835475e-05, 'epoch': 1.02}) +[2026-06-16 00:37:32,762][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.013841605745255947, 'train/grad_norm_pre_clip': 1.6467151641845703, 'timing/data': 0.006561393034644425, 'timing/model': 1.4574706809944473, 'learning_rate/qwen_vl_interface': 5.8211538752896056e-06, 'learning_rate/action_model': 5.821153875289606e-05, 'epoch': 1.02}) +[2026-06-16 00:37:38,775][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.12669780850410461, 'train/grad_norm_pre_clip': 10.043184280395508, 'timing/data': 0.006387426983565092, 'timing/model': 1.4726788450498134, 'learning_rate/qwen_vl_interface': 5.8181304327158814e-06, 'learning_rate/action_model': 5.8181304327158814e-05, 'epoch': 1.02}) +[2026-06-16 00:37:44,673][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.08368922024965286, 'train/grad_norm_pre_clip': 6.796473503112793, 'timing/data': 0.006438912998419255, 'timing/model': 1.473371664993465, 'learning_rate/qwen_vl_interface': 5.8151067566051965e-06, 'learning_rate/action_model': 5.815106756605196e-05, 'epoch': 1.02}) +[2026-06-16 00:37:50,661][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.17462606728076935, 'train/grad_norm_pre_clip': 11.261214256286621, 'timing/data': 0.0068900169571861625, 'timing/model': 1.466714089969173, 'learning_rate/qwen_vl_interface': 5.81208284820047e-06, 'learning_rate/action_model': 5.81208284820047e-05, 'epoch': 1.02}) +[2026-06-16 00:37:56,680][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.15078933537006378, 'train/grad_norm_pre_clip': 6.574200630187988, 'timing/data': 0.006897887971717864, 'timing/model': 1.7450428860029206, 'learning_rate/qwen_vl_interface': 5.809058708744716e-06, 'learning_rate/action_model': 5.809058708744716e-05, 'epoch': 1.02}) +[2026-06-16 00:38:02,662][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.04527350887656212, 'train/grad_norm_pre_clip': 6.21980094909668, 'timing/data': 0.006974461022764444, 'timing/model': 1.468096173950471, 'learning_rate/qwen_vl_interface': 5.8060343394810445e-06, 'learning_rate/action_model': 5.806034339481045e-05, 'epoch': 1.02}) +[2026-06-16 00:38:08,761][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.14373284578323364, 'train/grad_norm_pre_clip': 9.042753219604492, 'timing/data': 0.007966289995238185, 'timing/model': 1.5156920590088703, 'learning_rate/qwen_vl_interface': 5.803009741652659e-06, 'learning_rate/action_model': 5.8030097416526585e-05, 'epoch': 1.02}) +[2026-06-16 00:38:14,761][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.1171620562672615, 'train/grad_norm_pre_clip': 7.1622419357299805, 'timing/data': 0.0066280519822612405, 'timing/model': 1.4740963430376723, 'learning_rate/qwen_vl_interface': 5.799984916502859e-06, 'learning_rate/action_model': 5.799984916502859e-05, 'epoch': 1.02}) +[2026-06-16 00:38:20,746][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.11050427705049515, 'train/grad_norm_pre_clip': 4.944655418395996, 'timing/data': 0.007017004012595862, 'timing/model': 1.5095624410314485, 'learning_rate/qwen_vl_interface': 5.7969598652750334e-06, 'learning_rate/action_model': 5.796959865275033e-05, 'epoch': 1.02}) +[2026-06-16 00:38:26,562][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.28964850306510925, 'train/grad_norm_pre_clip': 17.0262508392334, 'timing/data': 0.007682531955651939, 'timing/model': 1.4535954410093836, 'learning_rate/qwen_vl_interface': 5.793934589212666e-06, 'learning_rate/action_model': 5.793934589212666e-05, 'epoch': 1.02}) +[2026-06-16 00:38:32,599][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.05069565027952194, 'train/grad_norm_pre_clip': 2.7301175594329834, 'timing/data': 0.006744001992046833, 'timing/model': 1.694651398982387, 'learning_rate/qwen_vl_interface': 5.7909090895593375e-06, 'learning_rate/action_model': 5.790909089559338e-05, 'epoch': 1.02}) +[2026-06-16 00:38:38,363][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.059425290673971176, 'train/grad_norm_pre_clip': 6.080070495605469, 'timing/data': 0.006759115029126406, 'timing/model': 1.4584155830089003, 'learning_rate/qwen_vl_interface': 5.7878833675587135e-06, 'learning_rate/action_model': 5.787883367558713e-05, 'epoch': 1.02}) +[2026-06-16 00:38:44,363][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.3929485082626343, 'train/grad_norm_pre_clip': 15.592123031616211, 'timing/data': 0.007036512019112706, 'timing/model': 1.4584192290203646, 'learning_rate/qwen_vl_interface': 5.784857424454556e-06, 'learning_rate/action_model': 5.7848574244545563e-05, 'epoch': 1.02}) +[2026-06-16 00:38:50,360][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.26787981390953064, 'train/grad_norm_pre_clip': 15.593317985534668, 'timing/data': 0.006801967974752188, 'timing/model': 1.4742286279797554, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 1.02}) +[2026-06-16 00:38:56,287][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.16517125070095062, 'train/grad_norm_pre_clip': 9.193831443786621, 'timing/data': 0.007378865964710712, 'timing/model': 1.5048125239554793, 'learning_rate/qwen_vl_interface': 5.778804879911129e-06, 'learning_rate/action_model': 5.778804879911128e-05, 'epoch': 1.02}) +[2026-06-16 00:39:02,161][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.028366325423121452, 'train/grad_norm_pre_clip': 3.634350538253784, 'timing/data': 0.007527424953877926, 'timing/model': 1.4429512720089406, 'learning_rate/qwen_vl_interface': 5.775778280959832e-06, 'learning_rate/action_model': 5.7757782809598315e-05, 'epoch': 1.02}) +[2026-06-16 00:39:08,098][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.08764772117137909, 'train/grad_norm_pre_clip': 5.504275321960449, 'timing/data': 0.006605147966183722, 'timing/model': 1.6966061069979332, 'learning_rate/qwen_vl_interface': 5.7727514658809435e-06, 'learning_rate/action_model': 5.772751465880944e-05, 'epoch': 1.02}) +[2026-06-16 00:39:13,787][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.0684894546866417, 'train/grad_norm_pre_clip': 5.146440029144287, 'timing/data': 0.006618806975893676, 'timing/model': 1.4774768379866146, 'learning_rate/qwen_vl_interface': 5.769724435918676e-06, 'learning_rate/action_model': 5.7697244359186755e-05, 'epoch': 1.02}) +[2026-06-16 00:39:19,764][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.10286474972963333, 'train/grad_norm_pre_clip': 12.545450210571289, 'timing/data': 0.0067841309937648475, 'timing/model': 1.4604559059953317, 'learning_rate/qwen_vl_interface': 5.766697192317323e-06, 'learning_rate/action_model': 5.766697192317323e-05, 'epoch': 1.02}) +[2026-06-16 00:39:25,756][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.03536918759346008, 'train/grad_norm_pre_clip': 3.477372646331787, 'timing/data': 0.007056532020214945, 'timing/model': 1.4719989889999852, 'learning_rate/qwen_vl_interface': 5.763669736321271e-06, 'learning_rate/action_model': 5.763669736321271e-05, 'epoch': 1.02}) +[2026-06-16 00:39:31,658][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.013926642015576363, 'train/grad_norm_pre_clip': 0.7485732436180115, 'timing/data': 0.006388738984242082, 'timing/model': 1.4592176949954592, 'learning_rate/qwen_vl_interface': 5.760642069174994e-06, 'learning_rate/action_model': 5.760642069174994e-05, 'epoch': 1.02}) +[2026-06-16 00:39:37,652][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.1256483793258667, 'train/grad_norm_pre_clip': 9.41950798034668, 'timing/data': 0.006629484007135034, 'timing/model': 1.5118599550332874, 'learning_rate/qwen_vl_interface': 5.757614192123051e-06, 'learning_rate/action_model': 5.757614192123051e-05, 'epoch': 1.03}) +[2026-06-16 00:39:43,595][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.13389970362186432, 'train/grad_norm_pre_clip': 8.114344596862793, 'timing/data': 0.006395119009539485, 'timing/model': 1.697695410985034, 'learning_rate/qwen_vl_interface': 5.754586106410089e-06, 'learning_rate/action_model': 5.754586106410088e-05, 'epoch': 1.03}) +[2026-06-16 00:39:49,259][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.21978849172592163, 'train/grad_norm_pre_clip': 9.20335578918457, 'timing/data': 0.006717453012242913, 'timing/model': 1.4578523669624701, 'learning_rate/qwen_vl_interface': 5.751557813280838e-06, 'learning_rate/action_model': 5.751557813280838e-05, 'epoch': 1.03}) +[2026-06-16 00:39:55,259][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.03115709312260151, 'train/grad_norm_pre_clip': 2.6302058696746826, 'timing/data': 0.00661817699437961, 'timing/model': 1.4582498819800094, 'learning_rate/qwen_vl_interface': 5.7485293139801146e-06, 'learning_rate/action_model': 5.7485293139801146e-05, 'epoch': 1.03}) +[2026-06-16 00:40:01,258][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.2949948012828827, 'train/grad_norm_pre_clip': 14.441198348999023, 'timing/data': 0.006863237009383738, 'timing/model': 1.4598901550052688, 'learning_rate/qwen_vl_interface': 5.745500609752821e-06, 'learning_rate/action_model': 5.745500609752821e-05, 'epoch': 1.03}) +[2026-06-16 00:40:07,160][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.11209381371736526, 'train/grad_norm_pre_clip': 4.548725605010986, 'timing/data': 0.006704975035972893, 'timing/model': 1.459757000033278, 'learning_rate/qwen_vl_interface': 5.742471701843943e-06, 'learning_rate/action_model': 5.742471701843943e-05, 'epoch': 1.03}) +[2026-06-16 00:40:13,157][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.02326815575361252, 'train/grad_norm_pre_clip': 2.2510557174682617, 'timing/data': 0.00642754603177309, 'timing/model': 1.5135000279988162, 'learning_rate/qwen_vl_interface': 5.739442591498553e-06, 'learning_rate/action_model': 5.739442591498553e-05, 'epoch': 1.03}) +[2026-06-16 00:40:19,096][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.2274407595396042, 'train/grad_norm_pre_clip': 9.843092918395996, 'timing/data': 0.006817188987042755, 'timing/model': 1.6929284690413624, 'learning_rate/qwen_vl_interface': 5.7364132799617985e-06, 'learning_rate/action_model': 5.7364132799617984e-05, 'epoch': 1.03}) +[2026-06-16 00:40:24,760][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.02363760955631733, 'train/grad_norm_pre_clip': 3.7381746768951416, 'timing/data': 0.006630945019423962, 'timing/model': 1.460903107014019, 'learning_rate/qwen_vl_interface': 5.733383768478923e-06, 'learning_rate/action_model': 5.7333837684789225e-05, 'epoch': 1.03}) +[2026-06-16 00:40:30,840][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.07985785603523254, 'train/grad_norm_pre_clip': 4.524807453155518, 'timing/data': 0.006643183005508035, 'timing/model': 1.5376235200092196, 'learning_rate/qwen_vl_interface': 5.730354058295234e-06, 'learning_rate/action_model': 5.730354058295234e-05, 'epoch': 1.03}) +[2026-06-16 00:40:36,837][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.08732984960079193, 'train/grad_norm_pre_clip': 3.1787607669830322, 'timing/data': 0.006786543992348015, 'timing/model': 1.537005931022577, 'learning_rate/qwen_vl_interface': 5.72732415065614e-06, 'learning_rate/action_model': 5.727324150656139e-05, 'epoch': 1.03}) +[2026-06-16 00:40:42,863][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.10801494866609573, 'train/grad_norm_pre_clip': 13.748927116394043, 'timing/data': 0.0067414980148896575, 'timing/model': 1.4589827120071277, 'learning_rate/qwen_vl_interface': 5.724294046807115e-06, 'learning_rate/action_model': 5.724294046807115e-05, 'epoch': 1.03}) +[2026-06-16 00:40:49,035][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.22443488240242004, 'train/grad_norm_pre_clip': 13.765809059143066, 'timing/data': 0.006799213006161153, 'timing/model': 1.5122683470253833, 'learning_rate/qwen_vl_interface': 5.7212637479937234e-06, 'learning_rate/action_model': 5.721263747993724e-05, 'epoch': 1.03}) +[2026-06-16 00:40:55,096][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.07241488993167877, 'train/grad_norm_pre_clip': 9.832403182983398, 'timing/data': 0.0066916150390170515, 'timing/model': 1.6913504189578816, 'learning_rate/qwen_vl_interface': 5.718233255461607e-06, 'learning_rate/action_model': 5.718233255461607e-05, 'epoch': 1.03}) +[2026-06-16 00:41:00,916][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.06578763574361801, 'train/grad_norm_pre_clip': 3.3265466690063477, 'timing/data': 0.006753875990398228, 'timing/model': 1.6155493480036967, 'learning_rate/qwen_vl_interface': 5.715202570456486e-06, 'learning_rate/action_model': 5.715202570456486e-05, 'epoch': 1.03}) +[2026-06-16 00:41:06,663][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.10370912402868271, 'train/grad_norm_pre_clip': 9.4514799118042, 'timing/data': 0.007035590999294072, 'timing/model': 1.4596011219546199, 'learning_rate/qwen_vl_interface': 5.7121716942241615e-06, 'learning_rate/action_model': 5.712171694224161e-05, 'epoch': 1.03}) +[2026-06-16 00:41:12,564][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.0319063626229763, 'train/grad_norm_pre_clip': 4.238169193267822, 'timing/data': 0.006769198982510716, 'timing/model': 1.458454534993507, 'learning_rate/qwen_vl_interface': 5.709140628010509e-06, 'learning_rate/action_model': 5.7091406280105085e-05, 'epoch': 1.03}) +[2026-06-16 00:50:21,748][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.05403739586472511, 'train/grad_norm_pre_clip': 4.138063430786133, 'eval/loss': 0.0898507684469223, 'eval/flap_precision': 0.8738574040219378, 'eval/flap_recall': 0.9579158316633266, 'eval/flap_f1': 0.9139579349904398, 'eval/flap_rate_model': 0.15194444444444444, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9139579349904396, 'timing/data': 0.00681128102587536, 'timing/model': 1.4728014229913242, 'learning_rate/qwen_vl_interface': 5.706109373061489e-06, 'learning_rate/action_model': 5.706109373061489e-05, 'epoch': 1.03}) +[2026-06-16 00:50:21,749][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2400_state +[2026-06-16 00:50:21,753][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 00:50:27,173][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2400_state/model.safetensors +[2026-06-16 00:50:40,370][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2400_state/optimizer.bin +[2026-06-16 00:50:40,370][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2400_state/sampler.bin +[2026-06-16 00:50:40,370][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2400_state/sampler_1.bin +[2026-06-16 00:50:40,371][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2400_state/random_states_0.pkl +[2026-06-16 00:50:40,482][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2400 +[2026-06-16 00:50:40,482][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 00:50:40,492][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 00:50:46,541][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.14835689961910248, 'train/grad_norm_pre_clip': 11.44637680053711, 'timing/data': 0.0066106090089306235, 'timing/model': 1.4404088690062054, 'learning_rate/qwen_vl_interface': 5.703077930623134e-06, 'learning_rate/action_model': 5.7030779306231344e-05, 'epoch': 1.03}) +[2026-06-16 00:50:52,440][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.029150163754820824, 'train/grad_norm_pre_clip': 3.2762198448181152, 'timing/data': 0.006433517963159829, 'timing/model': 1.4530643390025944, 'learning_rate/qwen_vl_interface': 5.7000463019415565e-06, 'learning_rate/action_model': 5.7000463019415563e-05, 'epoch': 1.03}) +[2026-06-16 00:50:58,349][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.13575512170791626, 'train/grad_norm_pre_clip': 5.203769207000732, 'timing/data': 0.006639802013523877, 'timing/model': 1.425760869984515, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 1.03}) +[2026-06-16 00:51:04,385][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.12681418657302856, 'train/grad_norm_pre_clip': 5.605724334716797, 'timing/data': 0.006459026015363634, 'timing/model': 1.7190496760304086, 'learning_rate/qwen_vl_interface': 5.693982490833561e-06, 'learning_rate/action_model': 5.693982490833561e-05, 'epoch': 1.03}) +[2026-06-16 00:51:10,246][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.05935254693031311, 'train/grad_norm_pre_clip': 14.158343315124512, 'timing/data': 0.007747962023131549, 'timing/model': 1.4908568109967746, 'learning_rate/qwen_vl_interface': 5.690950310899745e-06, 'learning_rate/action_model': 5.6909503108997456e-05, 'epoch': 1.03}) +[2026-06-16 00:51:16,419][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.09790226072072983, 'train/grad_norm_pre_clip': 7.068813323974609, 'timing/data': 0.006726168969180435, 'timing/model': 1.5506787080084905, 'learning_rate/qwen_vl_interface': 5.687917949707913e-06, 'learning_rate/action_model': 5.687917949707913e-05, 'epoch': 1.04}) +[2026-06-16 00:51:22,560][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.07054845243692398, 'train/grad_norm_pre_clip': 3.599468231201172, 'timing/data': 0.006636628007981926, 'timing/model': 1.5692466520122252, 'learning_rate/qwen_vl_interface': 5.684885408504555e-06, 'learning_rate/action_model': 5.684885408504554e-05, 'epoch': 1.04}) +[2026-06-16 00:51:28,821][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.04260806366801262, 'train/grad_norm_pre_clip': 8.303656578063965, 'timing/data': 0.00652543397154659, 'timing/model': 1.5255398949957453, 'learning_rate/qwen_vl_interface': 5.681852688536231e-06, 'learning_rate/action_model': 5.681852688536231e-05, 'epoch': 1.04}) +[2026-06-16 00:51:35,015][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.08585694432258606, 'train/grad_norm_pre_clip': 11.679584503173828, 'timing/data': 0.006507327954750508, 'timing/model': 1.7481881140265614, 'learning_rate/qwen_vl_interface': 5.67881979104958e-06, 'learning_rate/action_model': 5.67881979104958e-05, 'epoch': 1.04}) +[2026-06-16 00:51:40,771][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.02327752485871315, 'train/grad_norm_pre_clip': 1.4605947732925415, 'timing/data': 0.00674240302760154, 'timing/model': 1.4753848679829389, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 1.04}) +[2026-06-16 00:51:46,816][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.2664344012737274, 'train/grad_norm_pre_clip': 16.4688720703125, 'timing/data': 0.005802211002446711, 'timing/model': 1.5216875890037045, 'learning_rate/qwen_vl_interface': 5.672753468508204e-06, 'learning_rate/action_model': 5.672753468508204e-05, 'epoch': 1.04}) +[2026-06-16 00:51:52,853][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.15898774564266205, 'train/grad_norm_pre_clip': 11.363752365112305, 'timing/data': 0.006261367001570761, 'timing/model': 1.4956924499711022, 'learning_rate/qwen_vl_interface': 5.6697200459471165e-06, 'learning_rate/action_model': 5.669720045947117e-05, 'epoch': 1.04}) +[2026-06-16 00:51:58,996][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.186824232339859, 'train/grad_norm_pre_clip': 10.113109588623047, 'timing/data': 0.006239212991204113, 'timing/model': 1.524832887982484, 'learning_rate/qwen_vl_interface': 5.666686450854973e-06, 'learning_rate/action_model': 5.666686450854973e-05, 'epoch': 1.04}) +[2026-06-16 00:52:05,055][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.14392226934432983, 'train/grad_norm_pre_clip': 9.55829906463623, 'timing/data': 0.006261235976126045, 'timing/model': 1.4747515940107405, 'learning_rate/qwen_vl_interface': 5.663652684478768e-06, 'learning_rate/action_model': 5.6636526844787676e-05, 'epoch': 1.04}) +[2026-06-16 00:52:11,207][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.057952482253313065, 'train/grad_norm_pre_clip': 9.74856948852539, 'timing/data': 0.006381543003953993, 'timing/model': 1.7169386519817635, 'learning_rate/qwen_vl_interface': 5.660618748065569e-06, 'learning_rate/action_model': 5.6606187480655695e-05, 'epoch': 1.04}) +[2026-06-16 00:52:17,084][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.2979784905910492, 'train/grad_norm_pre_clip': 18.326936721801758, 'timing/data': 0.006536290980875492, 'timing/model': 1.6867645689635538, 'learning_rate/qwen_vl_interface': 5.657584642862514e-06, 'learning_rate/action_model': 5.657584642862513e-05, 'epoch': 1.04}) +[2026-06-16 00:52:22,757][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.3144437372684479, 'train/grad_norm_pre_clip': 18.369043350219727, 'timing/data': 0.006814358988776803, 'timing/model': 1.4588633669773117, 'learning_rate/qwen_vl_interface': 5.654550370116808e-06, 'learning_rate/action_model': 5.654550370116808e-05, 'epoch': 1.04}) +[2026-06-16 00:52:28,735][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.3070719242095947, 'train/grad_norm_pre_clip': 16.60655975341797, 'timing/data': 0.006631380005273968, 'timing/model': 1.5095821829745546, 'learning_rate/qwen_vl_interface': 5.651515931075726e-06, 'learning_rate/action_model': 5.651515931075726e-05, 'epoch': 1.04}) +[2026-06-16 00:52:34,657][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.09652384370565414, 'train/grad_norm_pre_clip': 7.9582672119140625, 'timing/data': 0.006633523968048394, 'timing/model': 1.456262114981655, 'learning_rate/qwen_vl_interface': 5.6484813269866126e-06, 'learning_rate/action_model': 5.648481326986612e-05, 'epoch': 1.04}) +[2026-06-16 00:52:40,658][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.010528107173740864, 'train/grad_norm_pre_clip': 0.9423392415046692, 'timing/data': 0.0073261120123788714, 'timing/model': 1.4776330679887906, 'learning_rate/qwen_vl_interface': 5.645446559096875e-06, 'learning_rate/action_model': 5.645446559096875e-05, 'epoch': 1.04}) +[2026-06-16 00:52:46,592][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.08263901621103287, 'train/grad_norm_pre_clip': 5.300095081329346, 'timing/data': 0.00670531002106145, 'timing/model': 1.693146075005643, 'learning_rate/qwen_vl_interface': 5.6424116286539965e-06, 'learning_rate/action_model': 5.642411628653996e-05, 'epoch': 1.04}) +[2026-06-16 00:52:52,260][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.2890840768814087, 'train/grad_norm_pre_clip': 15.591392517089844, 'timing/data': 0.006511054001748562, 'timing/model': 1.4443487990065478, 'learning_rate/qwen_vl_interface': 5.639376536905517e-06, 'learning_rate/action_model': 5.6393765369055175e-05, 'epoch': 1.04}) +[2026-06-16 00:52:58,257][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.16736678779125214, 'train/grad_norm_pre_clip': 7.562113285064697, 'timing/data': 0.006144315004348755, 'timing/model': 1.5121770030236803, 'learning_rate/qwen_vl_interface': 5.636341285099055e-06, 'learning_rate/action_model': 5.636341285099055e-05, 'epoch': 1.04}) +[2026-06-16 00:53:04,066][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.2117937207221985, 'train/grad_norm_pre_clip': 11.058015823364258, 'timing/data': 0.006606244016438723, 'timing/model': 1.4678392689675093, 'learning_rate/qwen_vl_interface': 5.6333058744822835e-06, 'learning_rate/action_model': 5.633305874482283e-05, 'epoch': 1.04}) +[2026-06-16 00:53:09,959][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.15020106732845306, 'train/grad_norm_pre_clip': 5.422211647033691, 'timing/data': 0.006889229989610612, 'timing/model': 1.4660803320002742, 'learning_rate/qwen_vl_interface': 5.630270306302944e-06, 'learning_rate/action_model': 5.6302703063029436e-05, 'epoch': 1.04}) +[2026-06-16 00:53:15,958][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.1348942518234253, 'train/grad_norm_pre_clip': 5.75770378112793, 'timing/data': 0.006230571016203612, 'timing/model': 1.5055130450055003, 'learning_rate/qwen_vl_interface': 5.627234581808848e-06, 'learning_rate/action_model': 5.627234581808848e-05, 'epoch': 1.04}) +[2026-06-16 00:53:21,904][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.08897840976715088, 'train/grad_norm_pre_clip': 4.704929828643799, 'timing/data': 0.006364097993355244, 'timing/model': 1.697660171019379, 'learning_rate/qwen_vl_interface': 5.624198702247863e-06, 'learning_rate/action_model': 5.624198702247863e-05, 'epoch': 1.04}) +[2026-06-16 00:53:27,560][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.07989192008972168, 'train/grad_norm_pre_clip': 8.330503463745117, 'timing/data': 0.006183793011587113, 'timing/model': 1.460583400039468, 'learning_rate/qwen_vl_interface': 5.621162668867925e-06, 'learning_rate/action_model': 5.621162668867925e-05, 'epoch': 1.04}) +[2026-06-16 00:53:33,566][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.23392537236213684, 'train/grad_norm_pre_clip': 8.41282844543457, 'timing/data': 0.006832045968621969, 'timing/model': 1.4645119350170717, 'learning_rate/qwen_vl_interface': 5.618126482917037e-06, 'learning_rate/action_model': 5.618126482917036e-05, 'epoch': 1.05}) +[2026-06-16 00:53:39,463][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.12467221915721893, 'train/grad_norm_pre_clip': 6.74173641204834, 'timing/data': 0.006185586040373892, 'timing/model': 1.4504428880172782, 'learning_rate/qwen_vl_interface': 5.6150901456432546e-06, 'learning_rate/action_model': 5.615090145643255e-05, 'epoch': 1.05}) +[2026-06-16 00:53:45,461][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.10268805176019669, 'train/grad_norm_pre_clip': 4.734820365905762, 'timing/data': 0.005909679981414229, 'timing/model': 1.4858593050157651, 'learning_rate/qwen_vl_interface': 5.612053658294705e-06, 'learning_rate/action_model': 5.6120536582947045e-05, 'epoch': 1.05}) +[2026-06-16 00:53:51,360][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.056075569242239, 'train/grad_norm_pre_clip': 3.4576897621154785, 'timing/data': 0.006593214988242835, 'timing/model': 1.4708454469800927, 'learning_rate/qwen_vl_interface': 5.609017022119572e-06, 'learning_rate/action_model': 5.609017022119573e-05, 'epoch': 1.05}) +[2026-06-16 00:53:57,295][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.03837515041232109, 'train/grad_norm_pre_clip': 2.2294812202453613, 'timing/data': 0.00638570002047345, 'timing/model': 1.6928812109981664, 'learning_rate/qwen_vl_interface': 5.605980238366104e-06, 'learning_rate/action_model': 5.605980238366103e-05, 'epoch': 1.05}) +[2026-06-16 00:54:03,136][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.3333463668823242, 'train/grad_norm_pre_clip': 16.423660278320312, 'timing/data': 0.0065966000547632575, 'timing/model': 1.5348630200023763, 'learning_rate/qwen_vl_interface': 5.602943308282606e-06, 'learning_rate/action_model': 5.602943308282605e-05, 'epoch': 1.05}) +[2026-06-16 00:54:09,162][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.07917985320091248, 'train/grad_norm_pre_clip': 6.405837535858154, 'timing/data': 0.0065636319923214614, 'timing/model': 1.460338253993541, 'learning_rate/qwen_vl_interface': 5.599906233117445e-06, 'learning_rate/action_model': 5.5999062331174446e-05, 'epoch': 1.05}) +[2026-06-16 00:54:15,059][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.22182273864746094, 'train/grad_norm_pre_clip': 12.541540145874023, 'timing/data': 0.006661395018454641, 'timing/model': 1.4554083929979242, 'learning_rate/qwen_vl_interface': 5.596869014119048e-06, 'learning_rate/action_model': 5.596869014119048e-05, 'epoch': 1.05}) +[2026-06-16 00:54:21,048][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.20295028388500214, 'train/grad_norm_pre_clip': 10.340099334716797, 'timing/data': 0.006374504009727389, 'timing/model': 1.5181723959976807, 'learning_rate/qwen_vl_interface': 5.593831652535902e-06, 'learning_rate/action_model': 5.593831652535902e-05, 'epoch': 1.05}) +[2026-06-16 00:54:27,048][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.19664716720581055, 'train/grad_norm_pre_clip': 6.872040271759033, 'timing/data': 0.00574938504723832, 'timing/model': 1.4442419409751892, 'learning_rate/qwen_vl_interface': 5.590794149616552e-06, 'learning_rate/action_model': 5.590794149616552e-05, 'epoch': 1.05}) +[2026-06-16 00:54:32,992][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.12328562885522842, 'train/grad_norm_pre_clip': 6.288573265075684, 'timing/data': 0.007040762982796878, 'timing/model': 1.7000308849965222, 'learning_rate/qwen_vl_interface': 5.587756506609601e-06, 'learning_rate/action_model': 5.5877565066096006e-05, 'epoch': 1.05}) +[2026-06-16 00:54:38,664][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.08101836591959, 'train/grad_norm_pre_clip': 5.855666160583496, 'timing/data': 0.006598132953513414, 'timing/model': 1.4630461799679324, 'learning_rate/qwen_vl_interface': 5.584718724763707e-06, 'learning_rate/action_model': 5.584718724763707e-05, 'epoch': 1.05}) +[2026-06-16 00:54:44,661][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.15317277610301971, 'train/grad_norm_pre_clip': 16.232423782348633, 'timing/data': 0.006598163046874106, 'timing/model': 1.4593384260078892, 'learning_rate/qwen_vl_interface': 5.581680805327588e-06, 'learning_rate/action_model': 5.5816808053275885e-05, 'epoch': 1.05}) +[2026-06-16 00:54:50,562][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.10120157152414322, 'train/grad_norm_pre_clip': 3.947627544403076, 'timing/data': 0.006981507001910359, 'timing/model': 1.457939186017029, 'learning_rate/qwen_vl_interface': 5.578642749550019e-06, 'learning_rate/action_model': 5.578642749550019e-05, 'epoch': 1.05}) +[2026-06-16 00:54:56,463][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.13925738632678986, 'train/grad_norm_pre_clip': 6.892665386199951, 'timing/data': 0.006468812993261963, 'timing/model': 1.462072291993536, 'learning_rate/qwen_vl_interface': 5.575604558679831e-06, 'learning_rate/action_model': 5.57560455867983e-05, 'epoch': 1.05}) +[2026-06-16 00:55:02,455][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.05228766053915024, 'train/grad_norm_pre_clip': 4.332056045532227, 'timing/data': 0.00674688199069351, 'timing/model': 1.5116959129809402, 'learning_rate/qwen_vl_interface': 5.5725662339659085e-06, 'learning_rate/action_model': 5.572566233965908e-05, 'epoch': 1.05}) +[2026-06-16 00:55:08,396][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.014451297000050545, 'train/grad_norm_pre_clip': 0.8477168679237366, 'timing/data': 0.00636419904185459, 'timing/model': 1.689020100980997, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 1.05}) +[2026-06-16 00:55:14,063][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.07519883662462234, 'train/grad_norm_pre_clip': 4.058475494384766, 'timing/data': 0.006223491975106299, 'timing/model': 1.4595907559851184, 'learning_rate/qwen_vl_interface': 5.566489188002671e-06, 'learning_rate/action_model': 5.56648918800267e-05, 'epoch': 1.05}) +[2026-06-16 00:55:20,061][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.18338139355182648, 'train/grad_norm_pre_clip': 12.510442733764648, 'timing/data': 0.006068683986086398, 'timing/model': 1.4455414039548486, 'learning_rate/qwen_vl_interface': 5.563450469251403e-06, 'learning_rate/action_model': 5.563450469251402e-05, 'epoch': 1.05}) +[2026-06-16 00:55:25,965][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.05551852658390999, 'train/grad_norm_pre_clip': 5.817993640899658, 'timing/data': 0.006400482030585408, 'timing/model': 1.462471637991257, 'learning_rate/qwen_vl_interface': 5.560411621652488e-06, 'learning_rate/action_model': 5.560411621652488e-05, 'epoch': 1.05}) +[2026-06-16 00:55:32,147][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.028201395645737648, 'train/grad_norm_pre_clip': 2.4171319007873535, 'timing/data': 0.010219779040198773, 'timing/model': 1.7406912020524032, 'learning_rate/qwen_vl_interface': 5.557372646455081e-06, 'learning_rate/action_model': 5.55737264645508e-05, 'epoch': 1.05}) +[2026-06-16 00:55:38,141][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.0603090301156044, 'train/grad_norm_pre_clip': 5.209765911102295, 'timing/data': 0.007166228024289012, 'timing/model': 1.503623530967161, 'learning_rate/qwen_vl_interface': 5.554333544908388e-06, 'learning_rate/action_model': 5.5543335449083875e-05, 'epoch': 1.05}) +[2026-06-16 00:55:44,193][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.10144681483507156, 'train/grad_norm_pre_clip': 8.671905517578125, 'timing/data': 0.0061367739690467715, 'timing/model': 1.6927189939888194, 'learning_rate/qwen_vl_interface': 5.551294318261671e-06, 'learning_rate/action_model': 5.5512943182616705e-05, 'epoch': 1.05}) +[2026-06-16 00:55:49,958][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.07107431441545486, 'train/grad_norm_pre_clip': 5.19338846206665, 'timing/data': 0.00677902897587046, 'timing/model': 1.4576205649646, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 1.06}) +[2026-06-16 00:55:55,958][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.022415408864617348, 'train/grad_norm_pre_clip': 2.0320041179656982, 'timing/data': 0.0066035810159519315, 'timing/model': 1.4590054690488614, 'learning_rate/qwen_vl_interface': 5.545215494665458e-06, 'learning_rate/action_model': 5.545215494665458e-05, 'epoch': 1.06}) +[2026-06-16 00:56:02,034][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.22301901876926422, 'train/grad_norm_pre_clip': 16.986459732055664, 'timing/data': 0.006669647991657257, 'timing/model': 1.5288181810174137, 'learning_rate/qwen_vl_interface': 5.5421759002147376e-06, 'learning_rate/action_model': 5.542175900214738e-05, 'epoch': 1.06}) +[2026-06-16 00:56:08,057][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.13506221771240234, 'train/grad_norm_pre_clip': 5.5492072105407715, 'timing/data': 0.006639694969635457, 'timing/model': 1.4593102419748902, 'learning_rate/qwen_vl_interface': 5.53913618566154e-06, 'learning_rate/action_model': 5.53913618566154e-05, 'epoch': 1.06}) +[2026-06-16 00:56:14,101][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.10354039818048477, 'train/grad_norm_pre_clip': 5.43106746673584, 'timing/data': 0.006657620018813759, 'timing/model': 1.5156714729964733, 'learning_rate/qwen_vl_interface': 5.536096352255381e-06, 'learning_rate/action_model': 5.536096352255381e-05, 'epoch': 1.06}) +[2026-06-16 00:56:19,987][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.06779846549034119, 'train/grad_norm_pre_clip': 4.361670017242432, 'timing/data': 0.007173189020249993, 'timing/model': 1.684116643969901, 'learning_rate/qwen_vl_interface': 5.533056401245817e-06, 'learning_rate/action_model': 5.5330564012458164e-05, 'epoch': 1.06}) +[2026-06-16 00:56:25,827][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.10033845156431198, 'train/grad_norm_pre_clip': 4.7569990158081055, 'timing/data': 0.0065507129766047, 'timing/model': 1.5289426230010577, 'learning_rate/qwen_vl_interface': 5.530016333882456e-06, 'learning_rate/action_model': 5.5300163338824565e-05, 'epoch': 1.06}) +[2026-06-16 00:56:31,761][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.13893966376781464, 'train/grad_norm_pre_clip': 7.681190013885498, 'timing/data': 0.007233797980006784, 'timing/model': 1.4607101979781874, 'learning_rate/qwen_vl_interface': 5.52697615141496e-06, 'learning_rate/action_model': 5.52697615141496e-05, 'epoch': 1.06}) +[2026-06-16 00:56:37,832][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.03599787876009941, 'train/grad_norm_pre_clip': 4.8546671867370605, 'timing/data': 0.007023920014034957, 'timing/model': 1.5711263340199366, 'learning_rate/qwen_vl_interface': 5.52393585509303e-06, 'learning_rate/action_model': 5.52393585509303e-05, 'epoch': 1.06}) +[2026-06-16 00:56:44,007][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.18228203058242798, 'train/grad_norm_pre_clip': 15.021671295166016, 'timing/data': 0.006618954008445144, 'timing/model': 1.544258679030463, 'learning_rate/qwen_vl_interface': 5.520895446166418e-06, 'learning_rate/action_model': 5.5208954461664175e-05, 'epoch': 1.06}) +[2026-06-16 00:56:49,862][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.010619865730404854, 'train/grad_norm_pre_clip': 1.057447910308838, 'timing/data': 0.0064228459959849715, 'timing/model': 1.436308894015383, 'learning_rate/qwen_vl_interface': 5.517854925884919e-06, 'learning_rate/action_model': 5.517854925884919e-05, 'epoch': 1.06}) +[2026-06-16 00:56:55,929][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.06440278887748718, 'train/grad_norm_pre_clip': 4.214620113372803, 'timing/data': 0.006823244038969278, 'timing/model': 1.771433740039356, 'learning_rate/qwen_vl_interface': 5.514814295498379e-06, 'learning_rate/action_model': 5.514814295498378e-05, 'epoch': 1.06}) +[2026-06-16 00:57:01,923][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.11527533084154129, 'train/grad_norm_pre_clip': 5.362910270690918, 'timing/data': 0.0073776799836196005, 'timing/model': 1.6868608740041964, 'learning_rate/qwen_vl_interface': 5.5117735562566846e-06, 'learning_rate/action_model': 5.5117735562566844e-05, 'epoch': 1.06}) +[2026-06-16 00:57:07,760][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.10401882976293564, 'train/grad_norm_pre_clip': 7.893272399902344, 'timing/data': 0.007170194992795587, 'timing/model': 1.459806839004159, 'learning_rate/qwen_vl_interface': 5.5087327094097685e-06, 'learning_rate/action_model': 5.508732709409769e-05, 'epoch': 1.06}) +[2026-06-16 00:57:13,667][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.1500600278377533, 'train/grad_norm_pre_clip': 12.170604705810547, 'timing/data': 0.00784724997356534, 'timing/model': 1.4638809370226227, 'learning_rate/qwen_vl_interface': 5.505691756207609e-06, 'learning_rate/action_model': 5.505691756207609e-05, 'epoch': 1.06}) +[2026-06-16 00:57:19,660][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.11913926154375076, 'train/grad_norm_pre_clip': 6.253121376037598, 'timing/data': 0.007157435989938676, 'timing/model': 1.475954529014416, 'learning_rate/qwen_vl_interface': 5.502650697900228e-06, 'learning_rate/action_model': 5.502650697900228e-05, 'epoch': 1.06}) +[2026-06-16 00:57:25,641][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.11137105524539948, 'train/grad_norm_pre_clip': 13.101937294006348, 'timing/data': 0.006985472980886698, 'timing/model': 1.5070499050198123, 'learning_rate/qwen_vl_interface': 5.499609535737688e-06, 'learning_rate/action_model': 5.4996095357376875e-05, 'epoch': 1.06}) +[2026-06-16 00:57:31,588][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.17944186925888062, 'train/grad_norm_pre_clip': 10.853551864624023, 'timing/data': 0.006767712999135256, 'timing/model': 1.6864802300115116, 'learning_rate/qwen_vl_interface': 5.496568270970096e-06, 'learning_rate/action_model': 5.496568270970096e-05, 'epoch': 1.06}) +[2026-06-16 00:57:37,487][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.01650453731417656, 'train/grad_norm_pre_clip': 1.0058406591415405, 'timing/data': 0.006578835018444806, 'timing/model': 1.6810743960086256, 'learning_rate/qwen_vl_interface': 5.4935269048476015e-06, 'learning_rate/action_model': 5.493526904847601e-05, 'epoch': 1.06}) +[2026-06-16 00:57:43,358][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.10547120869159698, 'train/grad_norm_pre_clip': 11.420698165893555, 'timing/data': 0.007433161023072898, 'timing/model': 1.6020159859908745, 'learning_rate/qwen_vl_interface': 5.490485438620397e-06, 'learning_rate/action_model': 5.4904854386203965e-05, 'epoch': 1.06}) +[2026-06-16 00:57:49,352][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.09394607692956924, 'train/grad_norm_pre_clip': 11.121979713439941, 'timing/data': 0.006750998029019684, 'timing/model': 1.514483572973404, 'learning_rate/qwen_vl_interface': 5.487443873538712e-06, 'learning_rate/action_model': 5.4874438735387115e-05, 'epoch': 1.06}) +[2026-06-16 00:57:55,370][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.030500492081046104, 'train/grad_norm_pre_clip': 3.5248758792877197, 'timing/data': 0.007046512968372554, 'timing/model': 1.480029508995358, 'learning_rate/qwen_vl_interface': 5.484402210852819e-06, 'learning_rate/action_model': 5.484402210852818e-05, 'epoch': 1.06}) +[2026-06-16 00:58:01,345][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.047677457332611084, 'train/grad_norm_pre_clip': 13.81227970123291, 'timing/data': 0.006557312968652695, 'timing/model': 1.512093730038032, 'learning_rate/qwen_vl_interface': 5.48136045181303e-06, 'learning_rate/action_model': 5.48136045181303e-05, 'epoch': 1.06}) +[2026-06-16 00:58:07,387][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.032244566828012466, 'train/grad_norm_pre_clip': 2.5392215251922607, 'timing/data': 0.0064308870350942016, 'timing/model': 1.6967646250268444, 'learning_rate/qwen_vl_interface': 5.478318597669699e-06, 'learning_rate/action_model': 5.4783185976696995e-05, 'epoch': 1.06}) +[2026-06-16 00:58:13,239][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.017610490322113037, 'train/grad_norm_pre_clip': 2.9523911476135254, 'timing/data': 0.006617792998440564, 'timing/model': 1.6282802330097184, 'learning_rate/qwen_vl_interface': 5.475276649673218e-06, 'learning_rate/action_model': 5.4752766496732174e-05, 'epoch': 1.07}) +[2026-06-16 00:58:18,958][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.08505336195230484, 'train/grad_norm_pre_clip': 7.285443305969238, 'timing/data': 0.006502803007606417, 'timing/model': 1.4488025159807876, 'learning_rate/qwen_vl_interface': 5.472234609074014e-06, 'learning_rate/action_model': 5.472234609074014e-05, 'epoch': 1.07}) +[2026-06-16 00:58:24,965][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.3818737864494324, 'train/grad_norm_pre_clip': 38.098819732666016, 'timing/data': 0.006760212010703981, 'timing/model': 1.4599414140102454, 'learning_rate/qwen_vl_interface': 5.46919247712256e-06, 'learning_rate/action_model': 5.46919247712256e-05, 'epoch': 1.07}) +[2026-06-16 00:58:30,960][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.26732978224754333, 'train/grad_norm_pre_clip': 11.337531089782715, 'timing/data': 0.0068914450239390135, 'timing/model': 1.4581612610490993, 'learning_rate/qwen_vl_interface': 5.466150255069356e-06, 'learning_rate/action_model': 5.4661502550693555e-05, 'epoch': 1.07}) +[2026-06-16 00:58:37,135][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.12898819148540497, 'train/grad_norm_pre_clip': 7.439620018005371, 'timing/data': 0.006830344966147095, 'timing/model': 1.5285745720029809, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 1.07}) +[2026-06-16 00:58:43,143][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.08979827165603638, 'train/grad_norm_pre_clip': 7.88215446472168, 'timing/data': 0.006650911993347108, 'timing/model': 1.5117515869787894, 'learning_rate/qwen_vl_interface': 5.46006554565991e-06, 'learning_rate/action_model': 5.4600655456599105e-05, 'epoch': 1.07}) +[2026-06-16 00:58:49,252][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.22845782339572906, 'train/grad_norm_pre_clip': 9.723862648010254, 'timing/data': 0.006559327011927962, 'timing/model': 1.6485225599608384, 'learning_rate/qwen_vl_interface': 5.457023060804865e-06, 'learning_rate/action_model': 5.457023060804864e-05, 'epoch': 1.07}) +[2026-06-16 00:58:55,143][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.06905126571655273, 'train/grad_norm_pre_clip': 6.92655611038208, 'timing/data': 0.006550683989189565, 'timing/model': 1.5176006989786401, 'learning_rate/qwen_vl_interface': 5.453980490850459e-06, 'learning_rate/action_model': 5.453980490850459e-05, 'epoch': 1.07}) +[2026-06-16 00:59:01,340][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.06401341408491135, 'train/grad_norm_pre_clip': 5.003530025482178, 'timing/data': 0.007168021984398365, 'timing/model': 1.5667819860391319, 'learning_rate/qwen_vl_interface': 5.4509378370473785e-06, 'learning_rate/action_model': 5.450937837047378e-05, 'epoch': 1.07}) +[2026-06-16 00:59:07,390][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.1318405121564865, 'train/grad_norm_pre_clip': 8.858052253723145, 'timing/data': 0.006660134997218847, 'timing/model': 1.5065515009919181, 'learning_rate/qwen_vl_interface': 5.4478951006463424e-06, 'learning_rate/action_model': 5.447895100646342e-05, 'epoch': 1.07}) +[2026-06-16 00:59:13,463][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.08730698376893997, 'train/grad_norm_pre_clip': 9.858670234680176, 'timing/data': 0.006656328972894698, 'timing/model': 1.444984518981073, 'learning_rate/qwen_vl_interface': 5.444852282898106e-06, 'learning_rate/action_model': 5.444852282898106e-05, 'epoch': 1.07}) +[2026-06-16 00:59:19,494][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.2779921591281891, 'train/grad_norm_pre_clip': 15.486898422241211, 'timing/data': 0.006860569992568344, 'timing/model': 1.7051066979765892, 'learning_rate/qwen_vl_interface': 5.441809385053456e-06, 'learning_rate/action_model': 5.4418093850534555e-05, 'epoch': 1.07}) +[2026-06-16 00:59:25,160][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.04858291149139404, 'train/grad_norm_pre_clip': 8.816163063049316, 'timing/data': 0.007327867031563073, 'timing/model': 1.458686719997786, 'learning_rate/qwen_vl_interface': 5.438766408363214e-06, 'learning_rate/action_model': 5.438766408363214e-05, 'epoch': 1.07}) +[2026-06-16 00:59:31,251][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.06448867172002792, 'train/grad_norm_pre_clip': 9.77851390838623, 'timing/data': 0.0064948120270855725, 'timing/model': 1.519021169980988, 'learning_rate/qwen_vl_interface': 5.43572335407823e-06, 'learning_rate/action_model': 5.43572335407823e-05, 'epoch': 1.07}) +[2026-06-16 00:59:37,076][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.04740230366587639, 'train/grad_norm_pre_clip': 3.3111422061920166, 'timing/data': 0.006907940027303994, 'timing/model': 1.476749996014405, 'learning_rate/qwen_vl_interface': 5.432680223449392e-06, 'learning_rate/action_model': 5.4326802234493924e-05, 'epoch': 1.07}) +[2026-06-16 00:59:43,070][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.38904091715812683, 'train/grad_norm_pre_clip': 16.769906997680664, 'timing/data': 0.006693353992886841, 'timing/model': 1.478198461001739, 'learning_rate/qwen_vl_interface': 5.429637017727614e-06, 'learning_rate/action_model': 5.4296370177276144e-05, 'epoch': 1.07}) +[2026-06-16 00:59:49,161][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.29466763138771057, 'train/grad_norm_pre_clip': 13.554903984069824, 'timing/data': 0.015496179985348135, 'timing/model': 1.4625491990009323, 'learning_rate/qwen_vl_interface': 5.426593738163845e-06, 'learning_rate/action_model': 5.4265937381638446e-05, 'epoch': 1.07}) +[2026-06-16 00:59:55,117][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.01590803638100624, 'train/grad_norm_pre_clip': 4.614767551422119, 'timing/data': 0.00850254500983283, 'timing/model': 1.7041977640474215, 'learning_rate/qwen_vl_interface': 5.42355038600906e-06, 'learning_rate/action_model': 5.42355038600906e-05, 'epoch': 1.07}) +[2026-06-16 01:00:00,858][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.1027749627828598, 'train/grad_norm_pre_clip': 5.226675510406494, 'timing/data': 0.013373218011111021, 'timing/model': 1.4382412880077027, 'learning_rate/qwen_vl_interface': 5.420506962514267e-06, 'learning_rate/action_model': 5.420506962514267e-05, 'epoch': 1.07}) +[2026-06-16 01:00:06,862][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.023207364603877068, 'train/grad_norm_pre_clip': 2.7396819591522217, 'timing/data': 0.01066659699426964, 'timing/model': 1.462076057039667, 'learning_rate/qwen_vl_interface': 5.4174634689305015e-06, 'learning_rate/action_model': 5.417463468930501e-05, 'epoch': 1.07}) +[2026-06-16 01:00:12,761][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.2270360291004181, 'train/grad_norm_pre_clip': 19.169679641723633, 'timing/data': 0.006682067003566772, 'timing/model': 1.4628754880395718, 'learning_rate/qwen_vl_interface': 5.41441990650883e-06, 'learning_rate/action_model': 5.4144199065088295e-05, 'epoch': 1.07}) +[2026-06-16 01:00:18,659][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.05064057931303978, 'train/grad_norm_pre_clip': 7.653543472290039, 'timing/data': 0.01313319499604404, 'timing/model': 1.4514868420083076, 'learning_rate/qwen_vl_interface': 5.411376276500345e-06, 'learning_rate/action_model': 5.411376276500345e-05, 'epoch': 1.07}) +[2026-06-16 01:00:24,560][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.16600903868675232, 'train/grad_norm_pre_clip': 16.677289962768555, 'timing/data': 0.007084520009811968, 'timing/model': 1.4687985579948872, 'learning_rate/qwen_vl_interface': 5.408332580156168e-06, 'learning_rate/action_model': 5.408332580156168e-05, 'epoch': 1.07}) +[2026-06-16 01:00:30,518][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.08823119103908539, 'train/grad_norm_pre_clip': 5.010271072387695, 'timing/data': 0.007595470990054309, 'timing/model': 1.711207785003353, 'learning_rate/qwen_vl_interface': 5.40528881872745e-06, 'learning_rate/action_model': 5.40528881872745e-05, 'epoch': 1.08}) +[2026-06-16 01:09:42,854][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.13273555040359497, 'train/grad_norm_pre_clip': 11.35216999053955, 'eval/loss': 0.0652199313044548, 'eval/flap_precision': 0.9297520661157025, 'eval/flap_recall': 0.9018036072144289, 'eval/flap_f1': 0.9155645981688707, 'eval/flap_rate_model': 0.13444444444444445, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9155645981688706, 'timing/data': 0.013928545988164842, 'timing/model': 1.4546070540091023, 'learning_rate/qwen_vl_interface': 5.402244993465364e-06, 'learning_rate/action_model': 5.402244993465364e-05, 'epoch': 1.08}) +[2026-06-16 01:09:42,855][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2500_state +[2026-06-16 01:09:42,859][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 01:09:48,245][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2500_state/model.safetensors +[2026-06-16 01:10:01,390][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2500_state/optimizer.bin +[2026-06-16 01:10:01,390][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2500_state/sampler.bin +[2026-06-16 01:10:01,391][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-16 01:10:01,391][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-16 01:10:01,495][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-16 01:10:01,496][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 01:10:01,505][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 01:10:07,611][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.15716552734375, 'train/grad_norm_pre_clip': 8.462987899780273, 'timing/data': 0.00688145199092105, 'timing/model': 1.7462491180049255, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 1.08}) +[2026-06-16 01:10:13,548][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.08287031203508377, 'train/grad_norm_pre_clip': 8.665895462036133, 'timing/data': 0.006663391017355025, 'timing/model': 1.5924110759515315, 'learning_rate/qwen_vl_interface': 5.396157156445918e-06, 'learning_rate/action_model': 5.396157156445918e-05, 'epoch': 1.08}) +[2026-06-16 01:10:19,680][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.07265880703926086, 'train/grad_norm_pre_clip': 3.5859439373016357, 'timing/data': 0.007269192021340132, 'timing/model': 1.491595806030091, 'learning_rate/qwen_vl_interface': 5.39311314719104e-06, 'learning_rate/action_model': 5.39311314719104e-05, 'epoch': 1.08}) +[2026-06-16 01:10:25,752][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.01474502868950367, 'train/grad_norm_pre_clip': 3.435293197631836, 'timing/data': 0.006968119996599853, 'timing/model': 1.4532734889653511, 'learning_rate/qwen_vl_interface': 5.390069079107755e-06, 'learning_rate/action_model': 5.390069079107755e-05, 'epoch': 1.08}) +[2026-06-16 01:10:31,552][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.061038434505462646, 'train/grad_norm_pre_clip': 5.045755386352539, 'timing/data': 0.006032333010807633, 'timing/model': 1.455354299978353, 'learning_rate/qwen_vl_interface': 5.387024953447361e-06, 'learning_rate/action_model': 5.387024953447361e-05, 'epoch': 1.08}) +[2026-06-16 01:10:37,536][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.23629344999790192, 'train/grad_norm_pre_clip': 12.490086555480957, 'timing/data': 0.006842985982075334, 'timing/model': 1.5070206489763223, 'learning_rate/qwen_vl_interface': 5.383980771461184e-06, 'learning_rate/action_model': 5.3839807714611846e-05, 'epoch': 1.08}) +[2026-06-16 01:10:43,879][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.31241416931152344, 'train/grad_norm_pre_clip': 10.744996070861816, 'timing/data': 0.006845659983810037, 'timing/model': 1.7957358779967763, 'learning_rate/qwen_vl_interface': 5.380936534400572e-06, 'learning_rate/action_model': 5.3809365344005726e-05, 'epoch': 1.08}) +[2026-06-16 01:10:49,585][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.09482681751251221, 'train/grad_norm_pre_clip': 12.973960876464844, 'timing/data': 0.0064233680022880435, 'timing/model': 1.4844513140269555, 'learning_rate/qwen_vl_interface': 5.377892243516899e-06, 'learning_rate/action_model': 5.3778922435168985e-05, 'epoch': 1.08}) +[2026-06-16 01:10:55,731][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.3116006851196289, 'train/grad_norm_pre_clip': 13.098783493041992, 'timing/data': 0.006601781002245843, 'timing/model': 1.5607468010275625, 'learning_rate/qwen_vl_interface': 5.3748479000615525e-06, 'learning_rate/action_model': 5.374847900061553e-05, 'epoch': 1.08}) +[2026-06-16 01:11:01,864][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.23258952796459198, 'train/grad_norm_pre_clip': 31.660476684570312, 'timing/data': 0.0066670270171016455, 'timing/model': 1.511513883015141, 'learning_rate/qwen_vl_interface': 5.371803505285954e-06, 'learning_rate/action_model': 5.371803505285954e-05, 'epoch': 1.08}) +[2026-06-16 01:11:08,135][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.01636752486228943, 'train/grad_norm_pre_clip': 1.116317868232727, 'timing/data': 0.0062332189991138875, 'timing/model': 1.5615330900182016, 'learning_rate/qwen_vl_interface': 5.368759060441534e-06, 'learning_rate/action_model': 5.368759060441534e-05, 'epoch': 1.08}) +[2026-06-16 01:11:14,433][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.1874151974916458, 'train/grad_norm_pre_clip': 7.917879581451416, 'timing/data': 0.006849415018223226, 'timing/model': 1.5211959679727443, 'learning_rate/qwen_vl_interface': 5.365714566779749e-06, 'learning_rate/action_model': 5.3657145667797484e-05, 'epoch': 1.08}) +[2026-06-16 01:11:20,485][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.17540490627288818, 'train/grad_norm_pre_clip': 8.9367094039917, 'timing/data': 0.006577815976925194, 'timing/model': 1.6811270259786397, 'learning_rate/qwen_vl_interface': 5.362670025552077e-06, 'learning_rate/action_model': 5.362670025552077e-05, 'epoch': 1.08}) +[2026-06-16 01:11:26,157][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.11289419233798981, 'train/grad_norm_pre_clip': 7.5111541748046875, 'timing/data': 0.007272417016793042, 'timing/model': 1.4575997150386684, 'learning_rate/qwen_vl_interface': 5.359625438010013e-06, 'learning_rate/action_model': 5.359625438010013e-05, 'epoch': 1.08}) +[2026-06-16 01:11:32,192][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.16228115558624268, 'train/grad_norm_pre_clip': 9.831684112548828, 'timing/data': 0.006555371975991875, 'timing/model': 1.4478027049917728, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 1.08}) +[2026-06-16 01:11:38,162][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.11235640943050385, 'train/grad_norm_pre_clip': 6.25247049331665, 'timing/data': 0.006746022962033749, 'timing/model': 1.459165220032446, 'learning_rate/qwen_vl_interface': 5.353536128988791e-06, 'learning_rate/action_model': 5.353536128988791e-05, 'epoch': 1.08}) +[2026-06-16 01:11:44,242][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.16074614226818085, 'train/grad_norm_pre_clip': 8.1899995803833, 'timing/data': 0.007037110975943506, 'timing/model': 1.569007522019092, 'learning_rate/qwen_vl_interface': 5.350491410012717e-06, 'learning_rate/action_model': 5.350491410012717e-05, 'epoch': 1.08}) +[2026-06-16 01:11:50,348][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.08887658268213272, 'train/grad_norm_pre_clip': 4.561344146728516, 'timing/data': 0.006363980995956808, 'timing/model': 1.511518069950398, 'learning_rate/qwen_vl_interface': 5.347446649728419e-06, 'learning_rate/action_model': 5.347446649728418e-05, 'epoch': 1.08}) +[2026-06-16 01:11:56,539][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.05938968434929848, 'train/grad_norm_pre_clip': 2.6994922161102295, 'timing/data': 0.008641541004180908, 'timing/model': 1.8295201209839433, 'learning_rate/qwen_vl_interface': 5.344401849387484e-06, 'learning_rate/action_model': 5.344401849387484e-05, 'epoch': 1.08}) +[2026-06-16 01:12:02,577][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.16283175349235535, 'train/grad_norm_pre_clip': 6.638066291809082, 'timing/data': 0.006486230995506048, 'timing/model': 1.5004762200405821, 'learning_rate/qwen_vl_interface': 5.3413570102415144e-06, 'learning_rate/action_model': 5.3413570102415144e-05, 'epoch': 1.08}) +[2026-06-16 01:12:08,832][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.07718835771083832, 'train/grad_norm_pre_clip': 3.8409197330474854, 'timing/data': 0.00696557603077963, 'timing/model': 1.5564166060066782, 'learning_rate/qwen_vl_interface': 5.338312133542131e-06, 'learning_rate/action_model': 5.3383121335421305e-05, 'epoch': 1.08}) +[2026-06-16 01:12:14,939][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.10323840379714966, 'train/grad_norm_pre_clip': 11.15818977355957, 'timing/data': 0.006721696991007775, 'timing/model': 1.56030583800748, 'learning_rate/qwen_vl_interface': 5.335267220540963e-06, 'learning_rate/action_model': 5.335267220540963e-05, 'epoch': 1.09}) +[2026-06-16 01:12:21,130][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.07584748417139053, 'train/grad_norm_pre_clip': 9.000347137451172, 'timing/data': 0.007124921015929431, 'timing/model': 1.5536789320176467, 'learning_rate/qwen_vl_interface': 5.332222272489662e-06, 'learning_rate/action_model': 5.332222272489662e-05, 'epoch': 1.09}) +[2026-06-16 01:12:27,342][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.09813136607408524, 'train/grad_norm_pre_clip': 6.334081172943115, 'timing/data': 0.0076966710039414465, 'timing/model': 1.5177544910111465, 'learning_rate/qwen_vl_interface': 5.3291772906398895e-06, 'learning_rate/action_model': 5.32917729063989e-05, 'epoch': 1.09}) +[2026-06-16 01:12:33,677][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.03810955956578255, 'train/grad_norm_pre_clip': 2.2443926334381104, 'timing/data': 0.006677121971733868, 'timing/model': 1.7959796729846857, 'learning_rate/qwen_vl_interface': 5.326132276243327e-06, 'learning_rate/action_model': 5.3261322762433266e-05, 'epoch': 1.09}) +[2026-06-16 01:12:39,891][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.20745332539081573, 'train/grad_norm_pre_clip': 11.702427864074707, 'timing/data': 0.006851949030533433, 'timing/model': 1.7337126840138808, 'learning_rate/qwen_vl_interface': 5.323087230551659e-06, 'learning_rate/action_model': 5.323087230551659e-05, 'epoch': 1.09}) +[2026-06-16 01:12:45,940][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.11372647434473038, 'train/grad_norm_pre_clip': 5.134994983673096, 'timing/data': 0.006502234959043562, 'timing/model': 1.5579618260380812, 'learning_rate/qwen_vl_interface': 5.3200421548165916e-06, 'learning_rate/action_model': 5.3200421548165914e-05, 'epoch': 1.09}) +[2026-06-16 01:12:52,236][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.12089735269546509, 'train/grad_norm_pre_clip': 9.543606758117676, 'timing/data': 0.006731250963639468, 'timing/model': 1.550229322980158, 'learning_rate/qwen_vl_interface': 5.3169970502898405e-06, 'learning_rate/action_model': 5.31699705028984e-05, 'epoch': 1.09}) +[2026-06-16 01:12:58,542][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.14820034801959991, 'train/grad_norm_pre_clip': 8.667498588562012, 'timing/data': 0.0065949910203926265, 'timing/model': 1.5708890429814346, 'learning_rate/qwen_vl_interface': 5.31395191822313e-06, 'learning_rate/action_model': 5.31395191822313e-05, 'epoch': 1.09}) +[2026-06-16 01:13:04,842][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.2578718960285187, 'train/grad_norm_pre_clip': 11.63605785369873, 'timing/data': 0.007133211998734623, 'timing/model': 1.5295798900187947, 'learning_rate/qwen_vl_interface': 5.310906759868205e-06, 'learning_rate/action_model': 5.310906759868205e-05, 'epoch': 1.09}) +[2026-06-16 01:13:11,162][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.11205333471298218, 'train/grad_norm_pre_clip': 5.566657066345215, 'timing/data': 0.007179411011748016, 'timing/model': 1.7747046169824898, 'learning_rate/qwen_vl_interface': 5.30786157647681e-06, 'learning_rate/action_model': 5.30786157647681e-05, 'epoch': 1.09}) +[2026-06-16 01:13:17,469][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.09018094837665558, 'train/grad_norm_pre_clip': 10.366448402404785, 'timing/data': 0.006764460005797446, 'timing/model': 1.7842987299663946, 'learning_rate/qwen_vl_interface': 5.304816369300709e-06, 'learning_rate/action_model': 5.304816369300709e-05, 'epoch': 1.09}) +[2026-06-16 01:13:23,533][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.10701466351747513, 'train/grad_norm_pre_clip': 5.352607727050781, 'timing/data': 0.0065442860359326005, 'timing/model': 1.5605854609748349, 'learning_rate/qwen_vl_interface': 5.301771139591666e-06, 'learning_rate/action_model': 5.301771139591666e-05, 'epoch': 1.09}) +[2026-06-16 01:13:29,837][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.12602989375591278, 'train/grad_norm_pre_clip': 12.240280151367188, 'timing/data': 0.006913229008205235, 'timing/model': 1.554144497960806, 'learning_rate/qwen_vl_interface': 5.298725888601465e-06, 'learning_rate/action_model': 5.298725888601465e-05, 'epoch': 1.09}) +[2026-06-16 01:13:36,142][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.4166618585586548, 'train/grad_norm_pre_clip': 19.41886329650879, 'timing/data': 0.007214663026388735, 'timing/model': 1.556301241973415, 'learning_rate/qwen_vl_interface': 5.295680617581891e-06, 'learning_rate/action_model': 5.2956806175818904e-05, 'epoch': 1.09}) +[2026-06-16 01:13:42,364][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.09799020737409592, 'train/grad_norm_pre_clip': 6.770829677581787, 'timing/data': 0.009144150011707097, 'timing/model': 1.445519058033824, 'learning_rate/qwen_vl_interface': 5.2926353277847395e-06, 'learning_rate/action_model': 5.292635327784739e-05, 'epoch': 1.09}) +[2026-06-16 01:13:48,390][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.11391005665063858, 'train/grad_norm_pre_clip': 8.424259185791016, 'timing/data': 0.006575020961463451, 'timing/model': 1.68542134197196, 'learning_rate/qwen_vl_interface': 5.2895900204618175e-06, 'learning_rate/action_model': 5.289590020461818e-05, 'epoch': 1.09}) +[2026-06-16 01:13:54,554][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.04377290606498718, 'train/grad_norm_pre_clip': 5.479077339172363, 'timing/data': 0.006787463964428753, 'timing/model': 1.6861463399836794, 'learning_rate/qwen_vl_interface': 5.286544696864933e-06, 'learning_rate/action_model': 5.2865446968649324e-05, 'epoch': 1.09}) +[2026-06-16 01:14:00,623][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.0203892569988966, 'train/grad_norm_pre_clip': 2.3304202556610107, 'timing/data': 0.0068707269965671, 'timing/model': 1.5507270469679497, 'learning_rate/qwen_vl_interface': 5.283499358245904e-06, 'learning_rate/action_model': 5.2834993582459034e-05, 'epoch': 1.09}) +[2026-06-16 01:14:06,737][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.020824529230594635, 'train/grad_norm_pre_clip': 4.641877174377441, 'timing/data': 0.006512359017506242, 'timing/model': 1.538546449970454, 'learning_rate/qwen_vl_interface': 5.280454005856553e-06, 'learning_rate/action_model': 5.2804540058565534e-05, 'epoch': 1.09}) +[2026-06-16 01:14:13,032][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.08865445107221603, 'train/grad_norm_pre_clip': 4.7759552001953125, 'timing/data': 0.00663508998695761, 'timing/model': 1.5483015610370785, 'learning_rate/qwen_vl_interface': 5.277408640948711e-06, 'learning_rate/action_model': 5.277408640948711e-05, 'epoch': 1.09}) +[2026-06-16 01:14:19,265][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.06438565254211426, 'train/grad_norm_pre_clip': 3.844119071960449, 'timing/data': 0.008829688013065606, 'timing/model': 1.5526491570053622, 'learning_rate/qwen_vl_interface': 5.274363264774213e-06, 'learning_rate/action_model': 5.274363264774212e-05, 'epoch': 1.09}) +[2026-06-16 01:14:25,423][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.3071404695510864, 'train/grad_norm_pre_clip': 23.222068786621094, 'timing/data': 0.006687596964184195, 'timing/model': 1.510763305996079, 'learning_rate/qwen_vl_interface': 5.271317878584897e-06, 'learning_rate/action_model': 5.2713178785848974e-05, 'epoch': 1.09}) +[2026-06-16 01:14:31,668][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.16136638820171356, 'train/grad_norm_pre_clip': 14.100303649902344, 'timing/data': 0.006793052016291767, 'timing/model': 1.7897625810001045, 'learning_rate/qwen_vl_interface': 5.268272483632606e-06, 'learning_rate/action_model': 5.268272483632606e-05, 'epoch': 1.09}) +[2026-06-16 01:14:37,638][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.1372535228729248, 'train/grad_norm_pre_clip': 13.01241397857666, 'timing/data': 0.00685670500388369, 'timing/model': 1.5326479279901832, 'learning_rate/qwen_vl_interface': 5.265227081169187e-06, 'learning_rate/action_model': 5.265227081169186e-05, 'epoch': 1.1}) +[2026-06-16 01:14:43,940][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.12380269914865494, 'train/grad_norm_pre_clip': 8.908493995666504, 'timing/data': 0.007410169986542314, 'timing/model': 1.5546198299853131, 'learning_rate/qwen_vl_interface': 5.2621816724464895e-06, 'learning_rate/action_model': 5.262181672446489e-05, 'epoch': 1.1}) +[2026-06-16 01:14:50,246][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.11064255982637405, 'train/grad_norm_pre_clip': 6.636218547821045, 'timing/data': 0.006884115980938077, 'timing/model': 1.5691839330247603, 'learning_rate/qwen_vl_interface': 5.259136258716367e-06, 'learning_rate/action_model': 5.2591362587163675e-05, 'epoch': 1.1}) +[2026-06-16 01:14:56,544][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.01450760941952467, 'train/grad_norm_pre_clip': 2.0394320487976074, 'timing/data': 0.006729970045853406, 'timing/model': 1.5292098030331545, 'learning_rate/qwen_vl_interface': 5.256090841230675e-06, 'learning_rate/action_model': 5.256090841230674e-05, 'epoch': 1.1}) +[2026-06-16 01:15:02,780][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.040768153965473175, 'train/grad_norm_pre_clip': 9.409220695495605, 'timing/data': 0.006781275034882128, 'timing/model': 1.7740272899973206, 'learning_rate/qwen_vl_interface': 5.253045421241266e-06, 'learning_rate/action_model': 5.253045421241266e-05, 'epoch': 1.1}) +[2026-06-16 01:15:08,665][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.07046292722225189, 'train/grad_norm_pre_clip': 4.254405498504639, 'timing/data': 0.00677502597682178, 'timing/model': 1.4607182380277663, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 1.1}) +[2026-06-16 01:15:14,984][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.08482035249471664, 'train/grad_norm_pre_clip': 6.610997200012207, 'timing/data': 0.008526362013071775, 'timing/model': 1.4567749080015346, 'learning_rate/qwen_vl_interface': 5.246954578758735e-06, 'learning_rate/action_model': 5.246954578758735e-05, 'epoch': 1.1}) +[2026-06-16 01:15:21,072][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.11796354502439499, 'train/grad_norm_pre_clip': 8.773700714111328, 'timing/data': 0.008505960984621197, 'timing/model': 1.4543223909568042, 'learning_rate/qwen_vl_interface': 5.243909158769327e-06, 'learning_rate/action_model': 5.2439091587693264e-05, 'epoch': 1.1}) +[2026-06-16 01:15:27,275][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.01269467268139124, 'train/grad_norm_pre_clip': 1.1828304529190063, 'timing/data': 0.008973420015536249, 'timing/model': 1.4584690619958565, 'learning_rate/qwen_vl_interface': 5.240863741283634e-06, 'learning_rate/action_model': 5.2408637412836336e-05, 'epoch': 1.1}) +[2026-06-16 01:15:33,549][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.050543081015348434, 'train/grad_norm_pre_clip': 3.818410873413086, 'timing/data': 0.008589955978095531, 'timing/model': 1.5849358800332993, 'learning_rate/qwen_vl_interface': 5.2378183275535125e-06, 'learning_rate/action_model': 5.237818327553512e-05, 'epoch': 1.1}) +[2026-06-16 01:15:40,072][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.0060602473095059395, 'train/grad_norm_pre_clip': 1.5054153203964233, 'timing/data': 0.008751804009079933, 'timing/model': 1.9762436770251952, 'learning_rate/qwen_vl_interface': 5.234772918830815e-06, 'learning_rate/action_model': 5.234772918830815e-05, 'epoch': 1.1}) +[2026-06-16 01:15:46,152][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.2532866597175598, 'train/grad_norm_pre_clip': 22.43857765197754, 'timing/data': 0.008870218007359654, 'timing/model': 1.5601510439882986, 'learning_rate/qwen_vl_interface': 5.231727516367395e-06, 'learning_rate/action_model': 5.231727516367395e-05, 'epoch': 1.1}) +[2026-06-16 01:15:52,638][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.2126404345035553, 'train/grad_norm_pre_clip': 11.519814491271973, 'timing/data': 0.008765142993070185, 'timing/model': 1.5452866190462373, 'learning_rate/qwen_vl_interface': 5.228682121415104e-06, 'learning_rate/action_model': 5.228682121415104e-05, 'epoch': 1.1}) +[2026-06-16 01:15:58,971][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.2239474654197693, 'train/grad_norm_pre_clip': 15.60983657836914, 'timing/data': 0.008797009999398142, 'timing/model': 1.4738810950075276, 'learning_rate/qwen_vl_interface': 5.225636735225789e-06, 'learning_rate/action_model': 5.225636735225788e-05, 'epoch': 1.1}) +[2026-06-16 01:16:05,146][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.1795937865972519, 'train/grad_norm_pre_clip': 10.292941093444824, 'timing/data': 0.009190698969177902, 'timing/model': 1.5573694340419024, 'learning_rate/qwen_vl_interface': 5.22259135905129e-06, 'learning_rate/action_model': 5.2225913590512895e-05, 'epoch': 1.1}) +[2026-06-16 01:16:11,652][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.10365284234285355, 'train/grad_norm_pre_clip': 10.279329299926758, 'timing/data': 0.008866611984558403, 'timing/model': 1.5916514529963024, 'learning_rate/qwen_vl_interface': 5.219545994143448e-06, 'learning_rate/action_model': 5.219545994143448e-05, 'epoch': 1.1}) +[2026-06-16 01:16:18,163][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.09449884295463562, 'train/grad_norm_pre_clip': 9.287342071533203, 'timing/data': 0.008817950962111354, 'timing/model': 1.969019345997367, 'learning_rate/qwen_vl_interface': 5.216500641754097e-06, 'learning_rate/action_model': 5.216500641754097e-05, 'epoch': 1.1}) +[2026-06-16 01:16:24,251][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.05296005308628082, 'train/grad_norm_pre_clip': 3.9156270027160645, 'timing/data': 0.009041259007062763, 'timing/model': 1.552449358976446, 'learning_rate/qwen_vl_interface': 5.213455303135067e-06, 'learning_rate/action_model': 5.2134553031350673e-05, 'epoch': 1.1}) +[2026-06-16 01:16:30,746][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.24959267675876617, 'train/grad_norm_pre_clip': 16.907503128051758, 'timing/data': 0.008862968010362238, 'timing/model': 1.5528922519879416, 'learning_rate/qwen_vl_interface': 5.2104099795381836e-06, 'learning_rate/action_model': 5.210409979538183e-05, 'epoch': 1.1}) +[2026-06-16 01:16:37,252][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.15612009167671204, 'train/grad_norm_pre_clip': 12.101252555847168, 'timing/data': 0.009130559978075325, 'timing/model': 1.5584812029846944, 'learning_rate/qwen_vl_interface': 5.207364672215261e-06, 'learning_rate/action_model': 5.2073646722152604e-05, 'epoch': 1.1}) +[2026-06-16 01:16:43,749][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.07766761630773544, 'train/grad_norm_pre_clip': 13.122232437133789, 'timing/data': 0.009400678041856736, 'timing/model': 1.5715127030271105, 'learning_rate/qwen_vl_interface': 5.2043193824181095e-06, 'learning_rate/action_model': 5.2043193824181094e-05, 'epoch': 1.1}) +[2026-06-16 01:16:50,248][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.30021631717681885, 'train/grad_norm_pre_clip': 12.646016120910645, 'timing/data': 0.009135247964877635, 'timing/model': 1.5791817450080998, 'learning_rate/qwen_vl_interface': 5.201274111398535e-06, 'learning_rate/action_model': 5.2012741113985356e-05, 'epoch': 1.1}) +[2026-06-16 01:16:56,752][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.01503351703286171, 'train/grad_norm_pre_clip': 2.0925097465515137, 'timing/data': 0.00893443200038746, 'timing/model': 1.9591818210319616, 'learning_rate/qwen_vl_interface': 5.198228860408334e-06, 'learning_rate/action_model': 5.198228860408334e-05, 'epoch': 1.1}) +[2026-06-16 01:17:02,849][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.039855923503637314, 'train/grad_norm_pre_clip': 3.9391837120056152, 'timing/data': 0.009117971989326179, 'timing/model': 1.5520276530296542, 'learning_rate/qwen_vl_interface': 5.195183630699292e-06, 'learning_rate/action_model': 5.195183630699292e-05, 'epoch': 1.1}) +[2026-06-16 01:17:09,355][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.09595701843500137, 'train/grad_norm_pre_clip': 8.146574974060059, 'timing/data': 0.008828871010337025, 'timing/model': 1.5617651670472696, 'learning_rate/qwen_vl_interface': 5.192138423523189e-06, 'learning_rate/action_model': 5.1921384235231895e-05, 'epoch': 1.11}) +[2026-06-16 01:17:15,849][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.20672635734081268, 'train/grad_norm_pre_clip': 9.991349220275879, 'timing/data': 0.008997989993076771, 'timing/model': 1.5547598529956304, 'learning_rate/qwen_vl_interface': 5.189093240131795e-06, 'learning_rate/action_model': 5.189093240131795e-05, 'epoch': 1.11}) +[2026-06-16 01:17:21,803][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.08460468053817749, 'train/grad_norm_pre_clip': 13.503816604614258, 'timing/data': 0.006596425024326891, 'timing/model': 1.5030232719727792, 'learning_rate/qwen_vl_interface': 5.186048081776869e-06, 'learning_rate/action_model': 5.1860480817768685e-05, 'epoch': 1.11}) +[2026-06-16 01:17:27,760][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.07080437988042831, 'train/grad_norm_pre_clip': 8.216264724731445, 'timing/data': 0.006368037953507155, 'timing/model': 1.5106564330053516, 'learning_rate/qwen_vl_interface': 5.183002949710161e-06, 'learning_rate/action_model': 5.18300294971016e-05, 'epoch': 1.11}) +[2026-06-16 01:17:33,807][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.15733934938907623, 'train/grad_norm_pre_clip': 15.633031845092773, 'timing/data': 0.007202539010904729, 'timing/model': 1.7447207699879073, 'learning_rate/qwen_vl_interface': 5.179957845183411e-06, 'learning_rate/action_model': 5.179957845183411e-05, 'epoch': 1.11}) +[2026-06-16 01:17:39,463][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.16100455820560455, 'train/grad_norm_pre_clip': 15.019826889038086, 'timing/data': 0.0073553340043872595, 'timing/model': 1.4583074679831043, 'learning_rate/qwen_vl_interface': 5.176912769448343e-06, 'learning_rate/action_model': 5.176912769448343e-05, 'epoch': 1.11}) +[2026-06-16 01:17:45,584][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.017107482999563217, 'train/grad_norm_pre_clip': 2.769662380218506, 'timing/data': 0.006577474996447563, 'timing/model': 1.4839744920027442, 'learning_rate/qwen_vl_interface': 5.173867723756676e-06, 'learning_rate/action_model': 5.173867723756676e-05, 'epoch': 1.11}) +[2026-06-16 01:17:51,578][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.23257604241371155, 'train/grad_norm_pre_clip': 10.20258617401123, 'timing/data': 0.007335604052059352, 'timing/model': 1.5128462130087428, 'learning_rate/qwen_vl_interface': 5.170822709360112e-06, 'learning_rate/action_model': 5.170822709360111e-05, 'epoch': 1.11}) +[2026-06-16 01:17:57,662][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.0257320124655962, 'train/grad_norm_pre_clip': 3.9089441299438477, 'timing/data': 0.00785820494638756, 'timing/model': 1.459928721014876, 'learning_rate/qwen_vl_interface': 5.16777772751034e-06, 'learning_rate/action_model': 5.16777772751034e-05, 'epoch': 1.11}) +[2026-06-16 01:18:03,765][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.07697455585002899, 'train/grad_norm_pre_clip': 7.165934085845947, 'timing/data': 0.0066118050017394125, 'timing/model': 1.5473568200250156, 'learning_rate/qwen_vl_interface': 5.16473277945904e-06, 'learning_rate/action_model': 5.1647327794590396e-05, 'epoch': 1.11}) +[2026-06-16 01:18:10,068][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.10109104961156845, 'train/grad_norm_pre_clip': 4.158632755279541, 'timing/data': 0.006718864024151117, 'timing/model': 1.7880671580205671, 'learning_rate/qwen_vl_interface': 5.161687866457872e-06, 'learning_rate/action_model': 5.161687866457871e-05, 'epoch': 1.11}) +[2026-06-16 01:18:16,130][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.2180030643939972, 'train/grad_norm_pre_clip': 8.279830932617188, 'timing/data': 0.0063884350238367915, 'timing/model': 1.552442669053562, 'learning_rate/qwen_vl_interface': 5.158642989758487e-06, 'learning_rate/action_model': 5.158642989758487e-05, 'epoch': 1.11}) +[2026-06-16 01:18:22,367][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.06072027236223221, 'train/grad_norm_pre_clip': 6.312272548675537, 'timing/data': 0.006971385038923472, 'timing/model': 1.6006172599736601, 'learning_rate/qwen_vl_interface': 5.155598150612519e-06, 'learning_rate/action_model': 5.1555981506125185e-05, 'epoch': 1.11}) +[2026-06-16 01:18:28,858][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.014919624663889408, 'train/grad_norm_pre_clip': 1.6933658123016357, 'timing/data': 0.006584252987522632, 'timing/model': 1.4640263940091245, 'learning_rate/qwen_vl_interface': 5.152553350271584e-06, 'learning_rate/action_model': 5.152553350271584e-05, 'epoch': 1.11}) +[2026-06-16 01:18:34,854][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.09458620101213455, 'train/grad_norm_pre_clip': 11.198721885681152, 'timing/data': 0.0067098080180585384, 'timing/model': 1.4586930419900455, 'learning_rate/qwen_vl_interface': 5.149508589987286e-06, 'learning_rate/action_model': 5.1495085899872855e-05, 'epoch': 1.11}) +[2026-06-16 01:18:41,133][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.12986668944358826, 'train/grad_norm_pre_clip': 8.571418762207031, 'timing/data': 0.00904429197544232, 'timing/model': 1.5596415749751031, 'learning_rate/qwen_vl_interface': 5.14646387101121e-06, 'learning_rate/action_model': 5.14646387101121e-05, 'epoch': 1.11}) +[2026-06-16 01:18:47,208][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.07521260529756546, 'train/grad_norm_pre_clip': 5.6096649169921875, 'timing/data': 0.006661065970547497, 'timing/model': 1.7065669870353304, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 1.11}) +[2026-06-16 01:18:53,028][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.21500158309936523, 'train/grad_norm_pre_clip': 13.873761177062988, 'timing/data': 0.008982430037576705, 'timing/model': 1.5898471120162867, 'learning_rate/qwen_vl_interface': 5.140374561989988e-06, 'learning_rate/action_model': 5.140374561989988e-05, 'epoch': 1.11}) +[2026-06-16 01:18:59,270][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.07734420150518417, 'train/grad_norm_pre_clip': 8.175381660461426, 'timing/data': 0.007309641980100423, 'timing/model': 1.6141719229635783, 'learning_rate/qwen_vl_interface': 5.137329974447924e-06, 'learning_rate/action_model': 5.137329974447924e-05, 'epoch': 1.11}) +[2026-06-16 01:19:05,530][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.042981769889593124, 'train/grad_norm_pre_clip': 3.676227331161499, 'timing/data': 0.006711549998726696, 'timing/model': 1.5523467999882996, 'learning_rate/qwen_vl_interface': 5.1342854332202516e-06, 'learning_rate/action_model': 5.1342854332202514e-05, 'epoch': 1.11}) +[2026-06-16 01:19:11,725][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.30057913064956665, 'train/grad_norm_pre_clip': 17.331579208374023, 'timing/data': 0.006974086980335414, 'timing/model': 1.546076385944616, 'learning_rate/qwen_vl_interface': 5.131240939558467e-06, 'learning_rate/action_model': 5.131240939558467e-05, 'epoch': 1.11}) +[2026-06-16 01:19:17,936][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.16207069158554077, 'train/grad_norm_pre_clip': 12.05344295501709, 'timing/data': 0.006651771022006869, 'timing/model': 1.5376911250059493, 'learning_rate/qwen_vl_interface': 5.128196494714048e-06, 'learning_rate/action_model': 5.128196494714047e-05, 'epoch': 1.11}) +[2026-06-16 01:19:24,198][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.17304784059524536, 'train/grad_norm_pre_clip': 6.864980220794678, 'timing/data': 0.0073174419812858105, 'timing/model': 1.7531349509954453, 'learning_rate/qwen_vl_interface': 5.125152099938448e-06, 'learning_rate/action_model': 5.125152099938447e-05, 'epoch': 1.11}) +[2026-06-16 01:19:30,385][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.01570681668817997, 'train/grad_norm_pre_clip': 1.3302546739578247, 'timing/data': 0.006793600041419268, 'timing/model': 1.7004302369896322, 'learning_rate/qwen_vl_interface': 5.122107756483104e-06, 'learning_rate/action_model': 5.122107756483103e-05, 'epoch': 1.12}) +[2026-06-16 01:19:36,436][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.20662780106067657, 'train/grad_norm_pre_clip': 10.7078857421875, 'timing/data': 0.00686447398038581, 'timing/model': 1.5553491509635933, 'learning_rate/qwen_vl_interface': 5.119063465599429e-06, 'learning_rate/action_model': 5.1190634655994285e-05, 'epoch': 1.12}) +[2026-06-16 01:19:42,394][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.3004815876483917, 'train/grad_norm_pre_clip': 7.999438762664795, 'timing/data': 0.006605751987081021, 'timing/model': 1.4896930229733698, 'learning_rate/qwen_vl_interface': 5.116019228538818e-06, 'learning_rate/action_model': 5.116019228538818e-05, 'epoch': 1.12}) +[2026-06-16 01:19:48,441][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.1929771602153778, 'train/grad_norm_pre_clip': 8.123510360717773, 'timing/data': 0.008852255006786436, 'timing/model': 1.543137894012034, 'learning_rate/qwen_vl_interface': 5.112975046552641e-06, 'learning_rate/action_model': 5.1129750465526404e-05, 'epoch': 1.12}) +[2026-06-16 01:19:54,547][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.1257011592388153, 'train/grad_norm_pre_clip': 6.793917655944824, 'timing/data': 0.006631410040426999, 'timing/model': 1.5098873510141857, 'learning_rate/qwen_vl_interface': 5.109930920892246e-06, 'learning_rate/action_model': 5.1099309208922464e-05, 'epoch': 1.12}) +[2026-06-16 01:20:00,500][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.16524259746074677, 'train/grad_norm_pre_clip': 7.2551398277282715, 'timing/data': 0.006583048030734062, 'timing/model': 1.6942400660482235, 'learning_rate/qwen_vl_interface': 5.10688685280896e-06, 'learning_rate/action_model': 5.10688685280896e-05, 'epoch': 1.12}) +[2026-06-16 01:20:06,400][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.08640163391828537, 'train/grad_norm_pre_clip': 5.302855491638184, 'timing/data': 0.00671529199462384, 'timing/model': 1.6939022470032796, 'learning_rate/qwen_vl_interface': 5.103842843554082e-06, 'learning_rate/action_model': 5.103842843554082e-05, 'epoch': 1.12}) +[2026-06-16 01:20:12,057][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.16509579122066498, 'train/grad_norm_pre_clip': 6.325407028198242, 'timing/data': 0.006591579993255436, 'timing/model': 1.4559735500370152, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 1.12}) +[2026-06-16 01:29:20,248][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.058027297258377075, 'train/grad_norm_pre_clip': 7.107080936431885, 'eval/loss': 0.05393025651574135, 'eval/flap_precision': 0.9333333333333333, 'eval/flap_recall': 0.9258517034068137, 'eval/flap_f1': 0.9295774647887324, 'eval/flap_rate_model': 0.1375, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9295774647887325, 'timing/data': 0.006522528012283146, 'timing/model': 1.5584709349786863, 'learning_rate/qwen_vl_interface': 5.097755006534638e-06, 'learning_rate/action_model': 5.097755006534638e-05, 'epoch': 1.12}) +[2026-06-16 01:29:20,248][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2600_state +[2026-06-16 01:29:20,253][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 01:29:25,605][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2600_state/model.safetensors +[2026-06-16 01:29:38,810][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2600_state/optimizer.bin +[2026-06-16 01:29:38,811][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2600_state/sampler.bin +[2026-06-16 01:29:38,811][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2600_state/sampler_1.bin +[2026-06-16 01:29:38,812][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2600_state/random_states_0.pkl +[2026-06-16 01:29:38,918][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2600 +[2026-06-16 01:29:38,919][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 01:29:38,928][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 01:29:45,038][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.10427817702293396, 'train/grad_norm_pre_clip': 10.060615539550781, 'timing/data': 0.006192467990331352, 'timing/model': 1.4493721619946882, 'learning_rate/qwen_vl_interface': 5.094711181272551e-06, 'learning_rate/action_model': 5.094711181272551e-05, 'epoch': 1.12}) +[2026-06-16 01:29:50,853][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.13976933062076569, 'train/grad_norm_pre_clip': 9.260554313659668, 'timing/data': 0.006993708957452327, 'timing/model': 1.446286061021965, 'learning_rate/qwen_vl_interface': 5.091667419843833e-06, 'learning_rate/action_model': 5.091667419843833e-05, 'epoch': 1.12}) +[2026-06-16 01:29:56,753][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.10970915108919144, 'train/grad_norm_pre_clip': 6.232311248779297, 'timing/data': 0.006902173045091331, 'timing/model': 1.4545892709866166, 'learning_rate/qwen_vl_interface': 5.088623723499657e-06, 'learning_rate/action_model': 5.088623723499657e-05, 'epoch': 1.12}) +[2026-06-16 01:30:02,761][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.06947144120931625, 'train/grad_norm_pre_clip': 5.924242973327637, 'timing/data': 0.007678447000216693, 'timing/model': 1.4727996399742551, 'learning_rate/qwen_vl_interface': 5.085580093491171e-06, 'learning_rate/action_model': 5.085580093491171e-05, 'epoch': 1.12}) +[2026-06-16 01:30:08,767][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.21872979402542114, 'train/grad_norm_pre_clip': 11.297836303710938, 'timing/data': 0.006754825997631997, 'timing/model': 1.6723405569791794, 'learning_rate/qwen_vl_interface': 5.0825365310695e-06, 'learning_rate/action_model': 5.082536531069499e-05, 'epoch': 1.12}) +[2026-06-16 01:30:14,757][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.13132771849632263, 'train/grad_norm_pre_clip': 11.267753601074219, 'timing/data': 0.009554137999657542, 'timing/model': 1.4868250789586455, 'learning_rate/qwen_vl_interface': 5.079493037485735e-06, 'learning_rate/action_model': 5.0794930374857344e-05, 'epoch': 1.12}) +[2026-06-16 01:30:20,857][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.07376101613044739, 'train/grad_norm_pre_clip': 5.143588542938232, 'timing/data': 0.006809215992689133, 'timing/model': 1.4646202659932896, 'learning_rate/qwen_vl_interface': 5.076449613990941e-06, 'learning_rate/action_model': 5.076449613990941e-05, 'epoch': 1.12}) +[2026-06-16 01:30:26,853][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.04002154618501663, 'train/grad_norm_pre_clip': 3.724355459213257, 'timing/data': 0.006594971055164933, 'timing/model': 1.4664455490419641, 'learning_rate/qwen_vl_interface': 5.073406261836157e-06, 'learning_rate/action_model': 5.0734062618361565e-05, 'epoch': 1.12}) +[2026-06-16 01:30:32,839][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.08990857750177383, 'train/grad_norm_pre_clip': 8.615485191345215, 'timing/data': 0.0069630329962819815, 'timing/model': 1.5073127779760398, 'learning_rate/qwen_vl_interface': 5.070362982272386e-06, 'learning_rate/action_model': 5.070362982272385e-05, 'epoch': 1.12}) +[2026-06-16 01:30:38,985][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.007438682019710541, 'train/grad_norm_pre_clip': 0.6612432599067688, 'timing/data': 0.008495107991620898, 'timing/model': 1.7874932319973595, 'learning_rate/qwen_vl_interface': 5.067319776550607e-06, 'learning_rate/action_model': 5.067319776550607e-05, 'epoch': 1.12}) +[2026-06-16 01:30:44,856][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.05222146585583687, 'train/grad_norm_pre_clip': 4.041845798492432, 'timing/data': 0.0067020689602941275, 'timing/model': 1.4805588829913177, 'learning_rate/qwen_vl_interface': 5.064276645921768e-06, 'learning_rate/action_model': 5.064276645921768e-05, 'epoch': 1.12}) +[2026-06-16 01:30:50,905][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.11018788069486618, 'train/grad_norm_pre_clip': 21.73236083984375, 'timing/data': 0.006768335995730013, 'timing/model': 1.5075209069764242, 'learning_rate/qwen_vl_interface': 5.061233591636787e-06, 'learning_rate/action_model': 5.061233591636787e-05, 'epoch': 1.12}) +[2026-06-16 01:30:56,832][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.05525599792599678, 'train/grad_norm_pre_clip': 5.880061626434326, 'timing/data': 0.006753563997335732, 'timing/model': 1.5347105260007083, 'learning_rate/qwen_vl_interface': 5.058190614946544e-06, 'learning_rate/action_model': 5.058190614946544e-05, 'epoch': 1.12}) +[2026-06-16 01:31:02,758][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.06912948191165924, 'train/grad_norm_pre_clip': 9.483917236328125, 'timing/data': 0.0072737999726086855, 'timing/model': 1.4577604089863598, 'learning_rate/qwen_vl_interface': 5.055147717101895e-06, 'learning_rate/action_model': 5.055147717101894e-05, 'epoch': 1.12}) +[2026-06-16 01:31:08,757][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.055408500134944916, 'train/grad_norm_pre_clip': 6.127666473388672, 'timing/data': 0.0068922279751859605, 'timing/model': 1.4774408519733697, 'learning_rate/qwen_vl_interface': 5.052104899353658e-06, 'learning_rate/action_model': 5.0521048993536577e-05, 'epoch': 1.13}) +[2026-06-16 01:31:14,754][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.04403457045555115, 'train/grad_norm_pre_clip': 6.409405708312988, 'timing/data': 0.0067571490071713924, 'timing/model': 1.7538383199716918, 'learning_rate/qwen_vl_interface': 5.049062162952622e-06, 'learning_rate/action_model': 5.0490621629526214e-05, 'epoch': 1.13}) +[2026-06-16 01:31:20,468][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.03223462402820587, 'train/grad_norm_pre_clip': 8.441489219665527, 'timing/data': 0.007396030006930232, 'timing/model': 1.467129501979798, 'learning_rate/qwen_vl_interface': 5.046019509149541e-06, 'learning_rate/action_model': 5.046019509149541e-05, 'epoch': 1.13}) +[2026-06-16 01:31:26,456][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.07500134408473969, 'train/grad_norm_pre_clip': 9.767120361328125, 'timing/data': 0.007221633044537157, 'timing/model': 1.4379232650389895, 'learning_rate/qwen_vl_interface': 5.042976939195135e-06, 'learning_rate/action_model': 5.0429769391951355e-05, 'epoch': 1.13}) +[2026-06-16 01:31:32,530][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.17438273131847382, 'train/grad_norm_pre_clip': 11.538453102111816, 'timing/data': 0.006725322047714144, 'timing/model': 1.551836080965586, 'learning_rate/qwen_vl_interface': 5.039934454340091e-06, 'learning_rate/action_model': 5.0399344543400907e-05, 'epoch': 1.13}) +[2026-06-16 01:31:38,458][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.0889350026845932, 'train/grad_norm_pre_clip': 6.901652812957764, 'timing/data': 0.006801514013204724, 'timing/model': 1.4552483370061964, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 1.13}) +[2026-06-16 01:31:44,356][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.027532516047358513, 'train/grad_norm_pre_clip': 2.2315895557403564, 'timing/data': 0.006914580997545272, 'timing/model': 1.4665432529873215, 'learning_rate/qwen_vl_interface': 5.033849744930647e-06, 'learning_rate/action_model': 5.033849744930646e-05, 'epoch': 1.13}) +[2026-06-16 01:31:50,399][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.23044468462467194, 'train/grad_norm_pre_clip': 16.41974449157715, 'timing/data': 0.007453083002474159, 'timing/model': 1.691134711029008, 'learning_rate/qwen_vl_interface': 5.030807522877443e-06, 'learning_rate/action_model': 5.030807522877443e-05, 'epoch': 1.13}) +[2026-06-16 01:31:56,174][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.02815294824540615, 'train/grad_norm_pre_clip': 4.952312469482422, 'timing/data': 0.006404699990525842, 'timing/model': 1.471749911957886, 'learning_rate/qwen_vl_interface': 5.027765390925987e-06, 'learning_rate/action_model': 5.027765390925987e-05, 'epoch': 1.13}) +[2026-06-16 01:32:02,259][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.047453105449676514, 'train/grad_norm_pre_clip': 5.4708638191223145, 'timing/data': 0.007049528998322785, 'timing/model': 1.4805012040305883, 'learning_rate/qwen_vl_interface': 5.024723350326783e-06, 'learning_rate/action_model': 5.024723350326783e-05, 'epoch': 1.13}) +[2026-06-16 01:32:08,557][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.03819089010357857, 'train/grad_norm_pre_clip': 3.78920316696167, 'timing/data': 0.007211006013676524, 'timing/model': 1.454618192045018, 'learning_rate/qwen_vl_interface': 5.021681402330303e-06, 'learning_rate/action_model': 5.021681402330303e-05, 'epoch': 1.13}) +[2026-06-16 01:32:14,540][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.03600997105240822, 'train/grad_norm_pre_clip': 3.6050829887390137, 'timing/data': 0.006802415009588003, 'timing/model': 1.5068291089846753, 'learning_rate/qwen_vl_interface': 5.018639548186972e-06, 'learning_rate/action_model': 5.0186395481869716e-05, 'epoch': 1.13}) +[2026-06-16 01:32:20,517][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.2138930708169937, 'train/grad_norm_pre_clip': 12.761910438537598, 'timing/data': 0.00653400004375726, 'timing/model': 1.5034868349903263, 'learning_rate/qwen_vl_interface': 5.015597789147184e-06, 'learning_rate/action_model': 5.015597789147184e-05, 'epoch': 1.13}) +[2026-06-16 01:32:26,603][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.14858752489089966, 'train/grad_norm_pre_clip': 5.949779033660889, 'timing/data': 0.006361095991451293, 'timing/model': 1.7011656899703667, 'learning_rate/qwen_vl_interface': 5.012556126461291e-06, 'learning_rate/action_model': 5.012556126461291e-05, 'epoch': 1.13}) +[2026-06-16 01:32:32,350][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.18871812522411346, 'train/grad_norm_pre_clip': 13.343165397644043, 'timing/data': 0.0059931830037385225, 'timing/model': 1.472044188005384, 'learning_rate/qwen_vl_interface': 5.0095145613796044e-06, 'learning_rate/action_model': 5.0095145613796046e-05, 'epoch': 1.13}) +[2026-06-16 01:32:38,467][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.0694074034690857, 'train/grad_norm_pre_clip': 15.253649711608887, 'timing/data': 0.007684934011194855, 'timing/model': 1.4743722220300697, 'learning_rate/qwen_vl_interface': 5.0064730951524e-06, 'learning_rate/action_model': 5.0064730951524e-05, 'epoch': 1.13}) +[2026-06-16 01:32:44,457][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.10375439375638962, 'train/grad_norm_pre_clip': 8.483579635620117, 'timing/data': 0.006496794987469912, 'timing/model': 1.463963245972991, 'learning_rate/qwen_vl_interface': 5.003431729029905e-06, 'learning_rate/action_model': 5.003431729029905e-05, 'epoch': 1.13}) +[2026-06-16 01:32:50,458][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.11521264165639877, 'train/grad_norm_pre_clip': 13.174068450927734, 'timing/data': 0.006710549001581967, 'timing/model': 1.4590796240372583, 'learning_rate/qwen_vl_interface': 5.000390464262314e-06, 'learning_rate/action_model': 5.0003904642623136e-05, 'epoch': 1.13}) +[2026-06-16 01:32:56,452][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.013576513156294823, 'train/grad_norm_pre_clip': 1.3766560554504395, 'timing/data': 0.0068840450257994235, 'timing/model': 1.5108917320030741, 'learning_rate/qwen_vl_interface': 4.997349302099774e-06, 'learning_rate/action_model': 4.997349302099773e-05, 'epoch': 1.13}) +[2026-06-16 01:33:02,512][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.12107651680707932, 'train/grad_norm_pre_clip': 12.649822235107422, 'timing/data': 0.00670860696118325, 'timing/model': 1.7709939390188083, 'learning_rate/qwen_vl_interface': 4.994308243792391e-06, 'learning_rate/action_model': 4.994308243792391e-05, 'epoch': 1.13}) +[2026-06-16 01:33:08,362][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.13233919441699982, 'train/grad_norm_pre_clip': 9.167855262756348, 'timing/data': 0.006686654000077397, 'timing/model': 1.4649804970249534, 'learning_rate/qwen_vl_interface': 4.991267290590233e-06, 'learning_rate/action_model': 4.991267290590233e-05, 'epoch': 1.13}) +[2026-06-16 01:33:14,362][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.128020241856575, 'train/grad_norm_pre_clip': 10.899002075195312, 'timing/data': 0.006634696968831122, 'timing/model': 1.4617896279669367, 'learning_rate/qwen_vl_interface': 4.988226443743317e-06, 'learning_rate/action_model': 4.9882264437433174e-05, 'epoch': 1.13}) +[2026-06-16 01:33:20,257][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.2684805989265442, 'train/grad_norm_pre_clip': 6.795154571533203, 'timing/data': 0.007325815036892891, 'timing/model': 1.4580225209938362, 'learning_rate/qwen_vl_interface': 4.985185704501623e-06, 'learning_rate/action_model': 4.985185704501623e-05, 'epoch': 1.13}) +[2026-06-16 01:33:26,162][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.18143197894096375, 'train/grad_norm_pre_clip': 22.0772705078125, 'timing/data': 0.0064512380049563944, 'timing/model': 1.4597541890107095, 'learning_rate/qwen_vl_interface': 4.982145074115082e-06, 'learning_rate/action_model': 4.982145074115082e-05, 'epoch': 1.14}) +[2026-06-16 01:33:32,147][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.1053931713104248, 'train/grad_norm_pre_clip': 7.673044204711914, 'timing/data': 0.0064454489620402455, 'timing/model': 1.508596423023846, 'learning_rate/qwen_vl_interface': 4.979104553833584e-06, 'learning_rate/action_model': 4.979104553833584e-05, 'epoch': 1.14}) +[2026-06-16 01:33:38,094][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.06358711421489716, 'train/grad_norm_pre_clip': 6.710102081298828, 'timing/data': 0.006719661992974579, 'timing/model': 1.6914369760197587, 'learning_rate/qwen_vl_interface': 4.9760641449069724e-06, 'learning_rate/action_model': 4.976064144906972e-05, 'epoch': 1.14}) +[2026-06-16 01:33:43,760][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.14107219874858856, 'train/grad_norm_pre_clip': 13.192326545715332, 'timing/data': 0.006722736987285316, 'timing/model': 1.4571571740088984, 'learning_rate/qwen_vl_interface': 4.973023848585042e-06, 'learning_rate/action_model': 4.973023848585042e-05, 'epoch': 1.14}) +[2026-06-16 01:33:49,868][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.02429129183292389, 'train/grad_norm_pre_clip': 3.846264362335205, 'timing/data': 0.007479421037714928, 'timing/model': 1.469081167015247, 'learning_rate/qwen_vl_interface': 4.969983666117545e-06, 'learning_rate/action_model': 4.9699836661175446e-05, 'epoch': 1.14}) +[2026-06-16 01:33:56,069][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.05441643297672272, 'train/grad_norm_pre_clip': 6.780842304229736, 'timing/data': 0.006536822998896241, 'timing/model': 1.4666880629956722, 'learning_rate/qwen_vl_interface': 4.966943598754185e-06, 'learning_rate/action_model': 4.9669435987541854e-05, 'epoch': 1.14}) +[2026-06-16 01:34:01,963][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.0607367679476738, 'train/grad_norm_pre_clip': 7.093538761138916, 'timing/data': 0.006954737997148186, 'timing/model': 1.4585886300192215, 'learning_rate/qwen_vl_interface': 4.963903647744621e-06, 'learning_rate/action_model': 4.96390364774462e-05, 'epoch': 1.14}) +[2026-06-16 01:34:07,958][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.05456746369600296, 'train/grad_norm_pre_clip': 4.840396881103516, 'timing/data': 0.006414151983335614, 'timing/model': 1.5191655870294198, 'learning_rate/qwen_vl_interface': 4.9608638143384604e-06, 'learning_rate/action_model': 4.960863814338461e-05, 'epoch': 1.14}) +[2026-06-16 01:34:13,894][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.028033843263983727, 'train/grad_norm_pre_clip': 2.8601906299591064, 'timing/data': 0.007243924017529935, 'timing/model': 1.6905332030146383, 'learning_rate/qwen_vl_interface': 4.9578240997852636e-06, 'learning_rate/action_model': 4.957824099785264e-05, 'epoch': 1.14}) +[2026-06-16 01:34:19,561][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.13134019076824188, 'train/grad_norm_pre_clip': 12.647405624389648, 'timing/data': 0.006765038997400552, 'timing/model': 1.465263436024543, 'learning_rate/qwen_vl_interface': 4.954784505334544e-06, 'learning_rate/action_model': 4.9547845053345434e-05, 'epoch': 1.14}) +[2026-06-16 01:34:25,461][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.19971877336502075, 'train/grad_norm_pre_clip': 18.276897430419922, 'timing/data': 0.0067224359954707325, 'timing/model': 1.4611004670150578, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 1.14}) +[2026-06-16 01:34:31,362][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.1892639845609665, 'train/grad_norm_pre_clip': 9.471864700317383, 'timing/data': 0.00668520200997591, 'timing/model': 1.4596316159586422, 'learning_rate/qwen_vl_interface': 4.94870568173833e-06, 'learning_rate/action_model': 4.9487056817383307e-05, 'epoch': 1.14}) +[2026-06-16 01:34:37,359][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.09388963878154755, 'train/grad_norm_pre_clip': 6.605303764343262, 'timing/data': 0.007703660987317562, 'timing/model': 1.4633936650352553, 'learning_rate/qwen_vl_interface': 4.945666455091614e-06, 'learning_rate/action_model': 4.9456664550916136e-05, 'epoch': 1.14}) +[2026-06-16 01:34:43,437][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.062130220234394073, 'train/grad_norm_pre_clip': 5.601552963256836, 'timing/data': 0.006243570998776704, 'timing/model': 1.5070423500146717, 'learning_rate/qwen_vl_interface': 4.942627353544921e-06, 'learning_rate/action_model': 4.942627353544921e-05, 'epoch': 1.14}) +[2026-06-16 01:34:49,380][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.14001566171646118, 'train/grad_norm_pre_clip': 8.967977523803711, 'timing/data': 0.006220437993761152, 'timing/model': 1.6815961200045422, 'learning_rate/qwen_vl_interface': 4.939588378347513e-06, 'learning_rate/action_model': 4.9395883783475125e-05, 'epoch': 1.14}) +[2026-06-16 01:34:55,204][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.1569792479276657, 'train/grad_norm_pre_clip': 10.396214485168457, 'timing/data': 0.0067313690087758005, 'timing/model': 1.6082849789527245, 'learning_rate/qwen_vl_interface': 4.936549530748597e-06, 'learning_rate/action_model': 4.9365495307485963e-05, 'epoch': 1.14}) +[2026-06-16 01:35:00,857][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.25004157423973083, 'train/grad_norm_pre_clip': 25.121217727661133, 'timing/data': 0.006625382986385375, 'timing/model': 1.4593816309934482, 'learning_rate/qwen_vl_interface': 4.9335108119973286e-06, 'learning_rate/action_model': 4.933510811997328e-05, 'epoch': 1.14}) +[2026-06-16 01:35:06,755][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.05401577427983284, 'train/grad_norm_pre_clip': 5.807580471038818, 'timing/data': 0.006814711960032582, 'timing/model': 1.4466240920010023, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 1.14}) +[2026-06-16 01:35:12,657][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.06620876491069794, 'train/grad_norm_pre_clip': 4.093469142913818, 'timing/data': 0.007400023052468896, 'timing/model': 1.454754881968256, 'learning_rate/qwen_vl_interface': 4.927433766034093e-06, 'learning_rate/action_model': 4.9274337660340934e-05, 'epoch': 1.14}) +[2026-06-16 01:35:18,734][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.029517602175474167, 'train/grad_norm_pre_clip': 3.7149362564086914, 'timing/data': 0.0065701419953256845, 'timing/model': 1.5029076870414428, 'learning_rate/qwen_vl_interface': 4.92439544132017e-06, 'learning_rate/action_model': 4.92439544132017e-05, 'epoch': 1.14}) +[2026-06-16 01:35:24,717][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.18888704478740692, 'train/grad_norm_pre_clip': 10.497221946716309, 'timing/data': 0.006975118012633175, 'timing/model': 1.7181644800002687, 'learning_rate/qwen_vl_interface': 4.92135725044998e-06, 'learning_rate/action_model': 4.92135725044998e-05, 'epoch': 1.14}) +[2026-06-16 01:35:30,690][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.10174465924501419, 'train/grad_norm_pre_clip': 13.95305347442627, 'timing/data': 0.007266417029313743, 'timing/model': 1.6891094090533443, 'learning_rate/qwen_vl_interface': 4.918319194672412e-06, 'learning_rate/action_model': 4.918319194672412e-05, 'epoch': 1.14}) +[2026-06-16 01:35:36,359][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.1276540458202362, 'train/grad_norm_pre_clip': 7.360565185546875, 'timing/data': 0.0072647539782337844, 'timing/model': 1.4572589740273543, 'learning_rate/qwen_vl_interface': 4.9152812752362935e-06, 'learning_rate/action_model': 4.915281275236293e-05, 'epoch': 1.14}) +[2026-06-16 01:35:42,256][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.052605319768190384, 'train/grad_norm_pre_clip': 4.403881072998047, 'timing/data': 0.007098889036569744, 'timing/model': 1.4601089310017414, 'learning_rate/qwen_vl_interface': 4.9122434933904e-06, 'learning_rate/action_model': 4.9122434933904e-05, 'epoch': 1.14}) +[2026-06-16 01:35:48,258][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.20388729870319366, 'train/grad_norm_pre_clip': 15.135809898376465, 'timing/data': 0.006821401009801775, 'timing/model': 1.475164155010134, 'learning_rate/qwen_vl_interface': 4.909205850383449e-06, 'learning_rate/action_model': 4.9092058503834485e-05, 'epoch': 1.15}) +[2026-06-16 01:35:54,367][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.16734391450881958, 'train/grad_norm_pre_clip': 13.169346809387207, 'timing/data': 0.005976918037049472, 'timing/model': 1.4443247190210968, 'learning_rate/qwen_vl_interface': 4.906168347464098e-06, 'learning_rate/action_model': 4.906168347464098e-05, 'epoch': 1.15}) +[2026-06-16 01:36:00,278][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.13988758623600006, 'train/grad_norm_pre_clip': 7.189727306365967, 'timing/data': 0.005933112988714129, 'timing/model': 1.6900546409888193, 'learning_rate/qwen_vl_interface': 4.903130985880955e-06, 'learning_rate/action_model': 4.903130985880954e-05, 'epoch': 1.15}) +[2026-06-16 01:36:06,228][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.038884565234184265, 'train/grad_norm_pre_clip': 3.8054211139678955, 'timing/data': 0.006639113009441644, 'timing/model': 1.6429095340427011, 'learning_rate/qwen_vl_interface': 4.900093766882558e-06, 'learning_rate/action_model': 4.900093766882558e-05, 'epoch': 1.15}) +[2026-06-16 01:36:12,059][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.21311379969120026, 'train/grad_norm_pre_clip': 25.6590633392334, 'timing/data': 0.006785938981920481, 'timing/model': 1.4591407150146551, 'learning_rate/qwen_vl_interface': 4.897056691717397e-06, 'learning_rate/action_model': 4.8970566917173966e-05, 'epoch': 1.15}) +[2026-06-16 01:36:18,169][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.006102593615651131, 'train/grad_norm_pre_clip': 0.6071325540542603, 'timing/data': 0.00776695302920416, 'timing/model': 1.5663885570247658, 'learning_rate/qwen_vl_interface': 4.894019761633898e-06, 'learning_rate/action_model': 4.894019761633898e-05, 'epoch': 1.15}) +[2026-06-16 01:36:24,162][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.14095786213874817, 'train/grad_norm_pre_clip': 7.4297261238098145, 'timing/data': 0.007233888027258217, 'timing/model': 1.488334843015764, 'learning_rate/qwen_vl_interface': 4.8909829778804304e-06, 'learning_rate/action_model': 4.89098297788043e-05, 'epoch': 1.15}) +[2026-06-16 01:36:30,148][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.06574930250644684, 'train/grad_norm_pre_clip': 5.150010108947754, 'timing/data': 0.006514940003398806, 'timing/model': 1.5150881150038913, 'learning_rate/qwen_vl_interface': 4.887946341705298e-06, 'learning_rate/action_model': 4.887946341705297e-05, 'epoch': 1.15}) +[2026-06-16 01:36:36,146][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.09017981588840485, 'train/grad_norm_pre_clip': 7.590855121612549, 'timing/data': 0.006476042966824025, 'timing/model': 1.5071389169897884, 'learning_rate/qwen_vl_interface': 4.8849098543567465e-06, 'learning_rate/action_model': 4.884909854356747e-05, 'epoch': 1.15}) +[2026-06-16 01:36:42,127][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.07597813755273819, 'train/grad_norm_pre_clip': 8.674020767211914, 'timing/data': 0.006636548030655831, 'timing/model': 1.7239351099706255, 'learning_rate/qwen_vl_interface': 4.881873517082965e-06, 'learning_rate/action_model': 4.8818735170829655e-05, 'epoch': 1.15}) +[2026-06-16 01:36:47,858][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.04118328168988228, 'train/grad_norm_pre_clip': 5.158273220062256, 'timing/data': 0.006528299010824412, 'timing/model': 1.453703295032028, 'learning_rate/qwen_vl_interface': 4.878837331132076e-06, 'learning_rate/action_model': 4.8788373311320754e-05, 'epoch': 1.15}) +[2026-06-16 01:36:54,016][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.2935030460357666, 'train/grad_norm_pre_clip': 63.971588134765625, 'timing/data': 0.006635957979597151, 'timing/model': 1.5164788120309822, 'learning_rate/qwen_vl_interface': 4.87580129775214e-06, 'learning_rate/action_model': 4.87580129775214e-05, 'epoch': 1.15}) +[2026-06-16 01:37:00,064][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.2023874819278717, 'train/grad_norm_pre_clip': 10.282915115356445, 'timing/data': 0.0066518610110506415, 'timing/model': 1.4843649929971434, 'learning_rate/qwen_vl_interface': 4.872765418191156e-06, 'learning_rate/action_model': 4.872765418191155e-05, 'epoch': 1.15}) +[2026-06-16 01:37:06,039][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.06475324928760529, 'train/grad_norm_pre_clip': 3.0554025173187256, 'timing/data': 0.006656498007941991, 'timing/model': 1.502886018017307, 'learning_rate/qwen_vl_interface': 4.8697296936970575e-06, 'learning_rate/action_model': 4.869729693697057e-05, 'epoch': 1.15}) +[2026-06-16 01:37:12,053][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.05540410429239273, 'train/grad_norm_pre_clip': 7.001133918762207, 'timing/data': 0.006454401009250432, 'timing/model': 1.7526564740110189, 'learning_rate/qwen_vl_interface': 4.8666941255177185e-06, 'learning_rate/action_model': 4.8666941255177185e-05, 'epoch': 1.15}) +[2026-06-16 01:37:17,900][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.1255885809659958, 'train/grad_norm_pre_clip': 8.512505531311035, 'timing/data': 0.006541678973007947, 'timing/model': 1.5044603910064325, 'learning_rate/qwen_vl_interface': 4.863658714900945e-06, 'learning_rate/action_model': 4.863658714900945e-05, 'epoch': 1.15}) +[2026-06-16 01:37:23,868][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.1436503380537033, 'train/grad_norm_pre_clip': 12.417159080505371, 'timing/data': 0.006375514029059559, 'timing/model': 1.4647333369939588, 'learning_rate/qwen_vl_interface': 4.860623463094483e-06, 'learning_rate/action_model': 4.860623463094483e-05, 'epoch': 1.15}) +[2026-06-16 01:37:29,659][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.06746947020292282, 'train/grad_norm_pre_clip': 3.2830801010131836, 'timing/data': 0.0069808950065635145, 'timing/model': 1.4579008030123077, 'learning_rate/qwen_vl_interface': 4.857588371346006e-06, 'learning_rate/action_model': 4.8575883713460057e-05, 'epoch': 1.15}) +[2026-06-16 01:37:35,562][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.032726190984249115, 'train/grad_norm_pre_clip': 6.570186138153076, 'timing/data': 0.006564882991369814, 'timing/model': 1.460219610016793, 'learning_rate/qwen_vl_interface': 4.854553440903127e-06, 'learning_rate/action_model': 4.854553440903126e-05, 'epoch': 1.15}) +[2026-06-16 01:37:41,462][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.03893042355775833, 'train/grad_norm_pre_clip': 3.568654775619507, 'timing/data': 0.006712790986057371, 'timing/model': 1.4723963770084083, 'learning_rate/qwen_vl_interface': 4.85151867301339e-06, 'learning_rate/action_model': 4.85151867301339e-05, 'epoch': 1.15}) +[2026-06-16 01:37:47,399][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.11638443917036057, 'train/grad_norm_pre_clip': 11.835325241088867, 'timing/data': 0.00717985798837617, 'timing/model': 1.6963768849964254, 'learning_rate/qwen_vl_interface': 4.848484068924274e-06, 'learning_rate/action_model': 4.848484068924275e-05, 'epoch': 1.15}) +[2026-06-16 01:37:53,070][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.13877363502979279, 'train/grad_norm_pre_clip': 10.789588928222656, 'timing/data': 0.0065870959660969675, 'timing/model': 1.465037231973838, 'learning_rate/qwen_vl_interface': 4.845449629883193e-06, 'learning_rate/action_model': 4.8454496298831934e-05, 'epoch': 1.15}) +[2026-06-16 01:37:59,060][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.036627255380153656, 'train/grad_norm_pre_clip': 5.249696731567383, 'timing/data': 0.006672601972240955, 'timing/model': 1.457908532989677, 'learning_rate/qwen_vl_interface': 4.842415357137487e-06, 'learning_rate/action_model': 4.842415357137487e-05, 'epoch': 1.15}) +[2026-06-16 01:38:05,058][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.06878402084112167, 'train/grad_norm_pre_clip': 4.5409979820251465, 'timing/data': 0.006599384010769427, 'timing/model': 1.4830237330170348, 'learning_rate/qwen_vl_interface': 4.839381251934432e-06, 'learning_rate/action_model': 4.8393812519344316e-05, 'epoch': 1.16}) +[2026-06-16 01:38:11,070][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.3184865713119507, 'train/grad_norm_pre_clip': 11.937400817871094, 'timing/data': 0.006983107945416123, 'timing/model': 1.4778624700265937, 'learning_rate/qwen_vl_interface': 4.836347315521233e-06, 'learning_rate/action_model': 4.836347315521233e-05, 'epoch': 1.16}) +[2026-06-16 01:38:16,957][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.048006121069192886, 'train/grad_norm_pre_clip': 4.034717559814453, 'timing/data': 0.0070115199778229, 'timing/model': 1.4384141589980572, 'learning_rate/qwen_vl_interface': 4.833313549145029e-06, 'learning_rate/action_model': 4.833313549145029e-05, 'epoch': 1.16}) +[2026-06-16 01:38:22,996][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.08080531656742096, 'train/grad_norm_pre_clip': 8.537899017333984, 'timing/data': 0.007063988014124334, 'timing/model': 1.7917753590154462, 'learning_rate/qwen_vl_interface': 4.830279954052884e-06, 'learning_rate/action_model': 4.8302799540528836e-05, 'epoch': 1.16}) +[2026-06-16 01:38:28,661][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.038350798189640045, 'train/grad_norm_pre_clip': 2.4591245651245117, 'timing/data': 0.006594627047888935, 'timing/model': 1.46171578502981, 'learning_rate/qwen_vl_interface': 4.827246531491797e-06, 'learning_rate/action_model': 4.827246531491797e-05, 'epoch': 1.16}) +[2026-06-16 01:38:34,656][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.07094412297010422, 'train/grad_norm_pre_clip': 8.420598983764648, 'timing/data': 0.006729805958457291, 'timing/model': 1.4562250470044091, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 1.16}) +[2026-06-16 01:38:40,557][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.1203911155462265, 'train/grad_norm_pre_clip': 19.021394729614258, 'timing/data': 0.007293855014722794, 'timing/model': 1.4548756549484096, 'learning_rate/qwen_vl_interface': 4.821180208950422e-06, 'learning_rate/action_model': 4.821180208950421e-05, 'epoch': 1.16}) +[2026-06-16 01:38:46,639][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.01601695828139782, 'train/grad_norm_pre_clip': 1.8825212717056274, 'timing/data': 0.006869441014714539, 'timing/model': 1.5103279490140267, 'learning_rate/qwen_vl_interface': 4.81814731146377e-06, 'learning_rate/action_model': 4.8181473114637695e-05, 'epoch': 1.16}) +[2026-06-16 01:38:52,558][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.5779247879981995, 'train/grad_norm_pre_clip': 26.66914939880371, 'timing/data': 0.006585363007616252, 'timing/model': 1.4508895369945094, 'learning_rate/qwen_vl_interface': 4.8151145914954465e-06, 'learning_rate/action_model': 4.815114591495446e-05, 'epoch': 1.16}) +[2026-06-16 01:38:58,491][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.09805484861135483, 'train/grad_norm_pre_clip': 9.9376802444458, 'timing/data': 0.007114021049346775, 'timing/model': 1.6897666600416414, 'learning_rate/qwen_vl_interface': 4.812082050292088e-06, 'learning_rate/action_model': 4.8120820502920876e-05, 'epoch': 1.16}) +[2026-06-16 01:39:04,157][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.2674483060836792, 'train/grad_norm_pre_clip': 11.499406814575195, 'timing/data': 0.006432397000025958, 'timing/model': 1.4587401199969463, 'learning_rate/qwen_vl_interface': 4.809049689100256e-06, 'learning_rate/action_model': 4.8090496891002555e-05, 'epoch': 1.16}) +[2026-06-16 01:39:10,157][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.013506874442100525, 'train/grad_norm_pre_clip': 3.0103538036346436, 'timing/data': 0.006682515959255397, 'timing/model': 1.4593669569585472, 'learning_rate/qwen_vl_interface': 4.80601750916644e-06, 'learning_rate/action_model': 4.8060175091664404e-05, 'epoch': 1.16}) +[2026-06-16 01:39:16,059][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.2723263204097748, 'train/grad_norm_pre_clip': 5.174902439117432, 'timing/data': 0.00629327300703153, 'timing/model': 1.4572556240018457, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 1.16}) +[2026-06-16 01:39:21,957][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.09147727489471436, 'train/grad_norm_pre_clip': 10.98494815826416, 'timing/data': 0.006596359016839415, 'timing/model': 1.461263288045302, 'learning_rate/qwen_vl_interface': 4.799953698058444e-06, 'learning_rate/action_model': 4.7999536980584434e-05, 'epoch': 1.16}) +[2026-06-16 01:39:27,944][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.11785580217838287, 'train/grad_norm_pre_clip': 13.453795433044434, 'timing/data': 0.0068405079655349255, 'timing/model': 1.5089668750297278, 'learning_rate/qwen_vl_interface': 4.796922069376867e-06, 'learning_rate/action_model': 4.796922069376867e-05, 'epoch': 1.16}) +[2026-06-16 01:48:31,748][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.04659020155668259, 'train/grad_norm_pre_clip': 4.582798957824707, 'eval/loss': 0.05729096382856369, 'eval/flap_precision': 0.9174852652259332, 'eval/flap_recall': 0.935871743486974, 'eval/flap_f1': 0.9265873015873016, 'eval/flap_rate_model': 0.1413888888888889, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9265873015873016, 'timing/data': 0.007828331959899515, 'timing/model': 1.6905424330034293, 'learning_rate/qwen_vl_interface': 4.793890626938511e-06, 'learning_rate/action_model': 4.793890626938511e-05, 'epoch': 1.16}) +[2026-06-16 01:48:31,749][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2700_state +[2026-06-16 01:48:31,753][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 01:48:37,100][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2700_state/model.safetensors +[2026-06-16 01:48:50,104][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2700_state/optimizer.bin +[2026-06-16 01:48:50,104][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2700_state/sampler.bin +[2026-06-16 01:48:50,104][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2700_state/sampler_1.bin +[2026-06-16 01:48:50,105][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2700_state/random_states_0.pkl +[2026-06-16 01:48:50,211][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2700 +[2026-06-16 01:48:50,212][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 01:48:50,222][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 01:48:56,040][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.035466138273477554, 'train/grad_norm_pre_clip': 3.081508159637451, 'timing/data': 0.006679830024950206, 'timing/model': 1.4250960340141319, 'learning_rate/qwen_vl_interface': 4.7908593719894914e-06, 'learning_rate/action_model': 4.790859371989491e-05, 'epoch': 1.16}) +[2026-06-16 01:49:02,079][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.2101828157901764, 'train/grad_norm_pre_clip': 9.147680282592773, 'timing/data': 0.006513856002129614, 'timing/model': 1.7983476080116816, 'learning_rate/qwen_vl_interface': 4.78782830577584e-06, 'learning_rate/action_model': 4.7878283057758396e-05, 'epoch': 1.16}) +[2026-06-16 01:49:07,745][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.3584098219871521, 'train/grad_norm_pre_clip': 14.875750541687012, 'timing/data': 0.006509479018859565, 'timing/model': 1.4579680559691042, 'learning_rate/qwen_vl_interface': 4.784797429543513e-06, 'learning_rate/action_model': 4.784797429543513e-05, 'epoch': 1.16}) +[2026-06-16 01:49:13,738][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.2038925588130951, 'train/grad_norm_pre_clip': 13.687625885009766, 'timing/data': 0.01022956398082897, 'timing/model': 1.541799616010394, 'learning_rate/qwen_vl_interface': 4.781766744538393e-06, 'learning_rate/action_model': 4.781766744538393e-05, 'epoch': 1.16}) +[2026-06-16 01:49:19,851][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.11537422239780426, 'train/grad_norm_pre_clip': 7.969146728515625, 'timing/data': 0.0071860760217532516, 'timing/model': 1.451993410999421, 'learning_rate/qwen_vl_interface': 4.778736252006277e-06, 'learning_rate/action_model': 4.7787362520062765e-05, 'epoch': 1.16}) +[2026-06-16 01:49:25,847][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.08586477488279343, 'train/grad_norm_pre_clip': 8.59160041809082, 'timing/data': 0.006647612957749516, 'timing/model': 1.425278133014217, 'learning_rate/qwen_vl_interface': 4.775705953192885e-06, 'learning_rate/action_model': 4.7757059531928856e-05, 'epoch': 1.16}) +[2026-06-16 01:49:31,836][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.2988140881061554, 'train/grad_norm_pre_clip': 13.77868366241455, 'timing/data': 0.006806866964325309, 'timing/model': 1.7362488390062936, 'learning_rate/qwen_vl_interface': 4.7726758493438615e-06, 'learning_rate/action_model': 4.7726758493438613e-05, 'epoch': 1.16}) +[2026-06-16 01:49:37,798][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.03735850378870964, 'train/grad_norm_pre_clip': 3.790423631668091, 'timing/data': 0.00652775599155575, 'timing/model': 1.700415733968839, 'learning_rate/qwen_vl_interface': 4.769645941704766e-06, 'learning_rate/action_model': 4.769645941704766e-05, 'epoch': 1.17}) +[2026-06-16 01:49:43,555][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.07567804306745529, 'train/grad_norm_pre_clip': 8.107094764709473, 'timing/data': 0.007028724008705467, 'timing/model': 1.4580142349586822, 'learning_rate/qwen_vl_interface': 4.766616231521081e-06, 'learning_rate/action_model': 4.7666162315210806e-05, 'epoch': 1.17}) +[2026-06-16 01:49:49,555][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.31482604146003723, 'train/grad_norm_pre_clip': 12.892755508422852, 'timing/data': 0.006571892008651048, 'timing/model': 1.459120940999128, 'learning_rate/qwen_vl_interface': 4.763586720038202e-06, 'learning_rate/action_model': 4.7635867200382014e-05, 'epoch': 1.17}) +[2026-06-16 01:49:55,455][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.07859876751899719, 'train/grad_norm_pre_clip': 7.350769519805908, 'timing/data': 0.007049584994092584, 'timing/model': 1.4565939710009843, 'learning_rate/qwen_vl_interface': 4.76055740850145e-06, 'learning_rate/action_model': 4.76055740850145e-05, 'epoch': 1.17}) +[2026-06-16 01:50:01,359][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.10072991997003555, 'train/grad_norm_pre_clip': 9.520519256591797, 'timing/data': 0.006855178042314947, 'timing/model': 1.4341020690044388, 'learning_rate/qwen_vl_interface': 4.757528298156058e-06, 'learning_rate/action_model': 4.757528298156058e-05, 'epoch': 1.17}) +[2026-06-16 01:50:07,362][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.07425346970558167, 'train/grad_norm_pre_clip': 8.573452949523926, 'timing/data': 0.006605950999073684, 'timing/model': 1.5254664730164222, 'learning_rate/qwen_vl_interface': 4.754499390247181e-06, 'learning_rate/action_model': 4.754499390247181e-05, 'epoch': 1.17}) +[2026-06-16 01:50:13,297][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.18933545053005219, 'train/grad_norm_pre_clip': 13.423300743103027, 'timing/data': 0.006590028002392501, 'timing/model': 1.6055741010350175, 'learning_rate/qwen_vl_interface': 4.751470686019887e-06, 'learning_rate/action_model': 4.751470686019888e-05, 'epoch': 1.17}) +[2026-06-16 01:50:18,954][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.0780223160982132, 'train/grad_norm_pre_clip': 5.29866361618042, 'timing/data': 0.007258291007019579, 'timing/model': 1.458684300945606, 'learning_rate/qwen_vl_interface': 4.7484421867191645e-06, 'learning_rate/action_model': 4.748442186719165e-05, 'epoch': 1.17}) +[2026-06-16 01:50:24,857][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.033937714993953705, 'train/grad_norm_pre_clip': 1.8019529581069946, 'timing/data': 0.006379537982866168, 'timing/model': 1.4608870120136999, 'learning_rate/qwen_vl_interface': 4.745413893589913e-06, 'learning_rate/action_model': 4.7454138935899125e-05, 'epoch': 1.17}) +[2026-06-16 01:50:30,756][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.15529541671276093, 'train/grad_norm_pre_clip': 10.216399192810059, 'timing/data': 0.0067727669957093894, 'timing/model': 1.459241489006672, 'learning_rate/qwen_vl_interface': 4.7423858078769506e-06, 'learning_rate/action_model': 4.742385807876951e-05, 'epoch': 1.17}) +[2026-06-16 01:50:36,661][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.07340557128190994, 'train/grad_norm_pre_clip': 4.581710338592529, 'timing/data': 0.006517481000628322, 'timing/model': 1.4378656600019895, 'learning_rate/qwen_vl_interface': 4.739357930825008e-06, 'learning_rate/action_model': 4.739357930825008e-05, 'epoch': 1.17}) +[2026-06-16 01:50:42,593][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.17170946300029755, 'train/grad_norm_pre_clip': 7.303827285766602, 'timing/data': 0.006522167997900397, 'timing/model': 1.693128721031826, 'learning_rate/qwen_vl_interface': 4.736330263678731e-06, 'learning_rate/action_model': 4.736330263678731e-05, 'epoch': 1.17}) +[2026-06-16 01:50:48,487][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.14635691046714783, 'train/grad_norm_pre_clip': 6.9571404457092285, 'timing/data': 0.0065265740267932415, 'timing/model': 1.6867612189962529, 'learning_rate/qwen_vl_interface': 4.733302807682679e-06, 'learning_rate/action_model': 4.733302807682679e-05, 'epoch': 1.17}) +[2026-06-16 01:50:54,157][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.2814807891845703, 'train/grad_norm_pre_clip': 13.388980865478516, 'timing/data': 0.006635625031776726, 'timing/model': 1.4543051889631897, 'learning_rate/qwen_vl_interface': 4.7302755640813255e-06, 'learning_rate/action_model': 4.7302755640813256e-05, 'epoch': 1.17}) +[2026-06-16 01:51:00,244][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.014579891227185726, 'train/grad_norm_pre_clip': 1.7258177995681763, 'timing/data': 0.00660739402519539, 'timing/model': 1.5049404720193706, 'learning_rate/qwen_vl_interface': 4.727248534119058e-06, 'learning_rate/action_model': 4.7272485341190574e-05, 'epoch': 1.17}) +[2026-06-16 01:51:06,361][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.007995771244168282, 'train/grad_norm_pre_clip': 1.0555938482284546, 'timing/data': 0.007186716014984995, 'timing/model': 1.4608993199653924, 'learning_rate/qwen_vl_interface': 4.72422171904017e-06, 'learning_rate/action_model': 4.7242217190401697e-05, 'epoch': 1.17}) +[2026-06-16 01:51:12,334][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.2307974249124527, 'train/grad_norm_pre_clip': 10.429801940917969, 'timing/data': 0.006572562037035823, 'timing/model': 1.502270229975693, 'learning_rate/qwen_vl_interface': 4.721195120088873e-06, 'learning_rate/action_model': 4.721195120088873e-05, 'epoch': 1.17}) +[2026-06-16 01:51:18,288][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.2887289226055145, 'train/grad_norm_pre_clip': 8.445475578308105, 'timing/data': 0.006660270970314741, 'timing/model': 1.6936676310142502, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 1.17}) +[2026-06-16 01:51:23,958][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.04128185287117958, 'train/grad_norm_pre_clip': 3.3584811687469482, 'timing/data': 0.006706218991894275, 'timing/model': 1.4590629759477451, 'learning_rate/qwen_vl_interface': 4.715142575545445e-06, 'learning_rate/action_model': 4.7151425755454454e-05, 'epoch': 1.17}) +[2026-06-16 01:51:29,858][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.18577800691127777, 'train/grad_norm_pre_clip': 4.463001251220703, 'timing/data': 0.006440537981688976, 'timing/model': 1.461403760011308, 'learning_rate/qwen_vl_interface': 4.712116632441287e-06, 'learning_rate/action_model': 4.7121166324412866e-05, 'epoch': 1.17}) +[2026-06-16 01:51:35,776][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.18516471982002258, 'train/grad_norm_pre_clip': 13.9668550491333, 'timing/data': 0.006853194965515286, 'timing/model': 1.477546414011158, 'learning_rate/qwen_vl_interface': 4.709090910440664e-06, 'learning_rate/action_model': 4.7090909104406634e-05, 'epoch': 1.17}) +[2026-06-16 01:51:41,673][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.2379062920808792, 'train/grad_norm_pre_clip': 12.185783386230469, 'timing/data': 0.0066977060050703585, 'timing/model': 1.4507723029819317, 'learning_rate/qwen_vl_interface': 4.7060654107873335e-06, 'learning_rate/action_model': 4.7060654107873335e-05, 'epoch': 1.17}) +[2026-06-16 01:51:47,551][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.12925904989242554, 'train/grad_norm_pre_clip': 6.034627914428711, 'timing/data': 0.006971038004849106, 'timing/model': 1.5080555499880575, 'learning_rate/qwen_vl_interface': 4.7030401347249685e-06, 'learning_rate/action_model': 4.703040134724968e-05, 'epoch': 1.17}) +[2026-06-16 01:51:53,450][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.1287498027086258, 'train/grad_norm_pre_clip': 6.7428975105285645, 'timing/data': 0.006681272003334016, 'timing/model': 1.6511993200401776, 'learning_rate/qwen_vl_interface': 4.700015083497143e-06, 'learning_rate/action_model': 4.700015083497142e-05, 'epoch': 1.18}) +[2026-06-16 01:51:59,160][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.12346287816762924, 'train/grad_norm_pre_clip': 4.5061116218566895, 'timing/data': 0.007378137961495668, 'timing/model': 1.4535093889571726, 'learning_rate/qwen_vl_interface': 4.696990258347342e-06, 'learning_rate/action_model': 4.696990258347342e-05, 'epoch': 1.18}) +[2026-06-16 01:52:05,161][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.1793939173221588, 'train/grad_norm_pre_clip': 11.788686752319336, 'timing/data': 0.006481268035713583, 'timing/model': 1.4661107530118898, 'learning_rate/qwen_vl_interface': 4.693965660518957e-06, 'learning_rate/action_model': 4.693965660518957e-05, 'epoch': 1.18}) +[2026-06-16 01:52:11,059][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.1174561157822609, 'train/grad_norm_pre_clip': 3.223541259765625, 'timing/data': 0.006508908001706004, 'timing/model': 1.461839034978766, 'learning_rate/qwen_vl_interface': 4.690941291255285e-06, 'learning_rate/action_model': 4.6909412912552844e-05, 'epoch': 1.18}) +[2026-06-16 01:52:16,966][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.08957656472921371, 'train/grad_norm_pre_clip': 10.545816421508789, 'timing/data': 0.0068787430063821375, 'timing/model': 1.438955990015529, 'learning_rate/qwen_vl_interface': 4.68791715179953e-06, 'learning_rate/action_model': 4.68791715179953e-05, 'epoch': 1.18}) +[2026-06-16 01:52:22,940][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.17046873271465302, 'train/grad_norm_pre_clip': 7.841940402984619, 'timing/data': 0.00661577598657459, 'timing/model': 1.5078708049841225, 'learning_rate/qwen_vl_interface': 4.684893243394804e-06, 'learning_rate/action_model': 4.6848932433948036e-05, 'epoch': 1.18}) +[2026-06-16 01:52:28,859][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.18448001146316528, 'train/grad_norm_pre_clip': 9.641891479492188, 'timing/data': 0.006640702951699495, 'timing/model': 1.6605384879512712, 'learning_rate/qwen_vl_interface': 4.68186956728412e-06, 'learning_rate/action_model': 4.68186956728412e-05, 'epoch': 1.18}) +[2026-06-16 01:52:34,562][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.23256780207157135, 'train/grad_norm_pre_clip': 11.487604141235352, 'timing/data': 0.007103753974661231, 'timing/model': 1.4644570450182073, 'learning_rate/qwen_vl_interface': 4.678846124710395e-06, 'learning_rate/action_model': 4.678846124710394e-05, 'epoch': 1.18}) +[2026-06-16 01:52:40,557][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.1395421028137207, 'train/grad_norm_pre_clip': 6.103368282318115, 'timing/data': 0.006699588964693248, 'timing/model': 1.462360970035661, 'learning_rate/qwen_vl_interface': 4.675822916916454e-06, 'learning_rate/action_model': 4.6758229169164536e-05, 'epoch': 1.18}) +[2026-06-16 01:52:46,573][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.037938982248306274, 'train/grad_norm_pre_clip': 3.098588228225708, 'timing/data': 0.007072417996823788, 'timing/model': 1.5037588310078718, 'learning_rate/qwen_vl_interface': 4.6727999451450236e-06, 'learning_rate/action_model': 4.672799945145023e-05, 'epoch': 1.18}) +[2026-06-16 01:52:52,539][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.03505907207727432, 'train/grad_norm_pre_clip': 2.175926685333252, 'timing/data': 0.006741260993294418, 'timing/model': 1.5067983329645358, 'learning_rate/qwen_vl_interface': 4.669777210638733e-06, 'learning_rate/action_model': 4.669777210638733e-05, 'epoch': 1.18}) +[2026-06-16 01:52:58,490][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.029179057106375694, 'train/grad_norm_pre_clip': 2.25358510017395, 'timing/data': 0.007508719980251044, 'timing/model': 1.6916364679927938, 'learning_rate/qwen_vl_interface': 4.666754714640115e-06, 'learning_rate/action_model': 4.666754714640115e-05, 'epoch': 1.18}) +[2026-06-16 01:53:04,171][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.07442676275968552, 'train/grad_norm_pre_clip': 5.155036926269531, 'timing/data': 0.008127099019475281, 'timing/model': 1.47207692603115, 'learning_rate/qwen_vl_interface': 4.663732458391602e-06, 'learning_rate/action_model': 4.663732458391601e-05, 'epoch': 1.18}) +[2026-06-16 01:53:10,155][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.05976729840040207, 'train/grad_norm_pre_clip': 3.6491739749908447, 'timing/data': 0.005949906015302986, 'timing/model': 1.4544417940196581, 'learning_rate/qwen_vl_interface': 4.6607104431355295e-06, 'learning_rate/action_model': 4.66071044313553e-05, 'epoch': 1.18}) +[2026-06-16 01:53:16,039][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.053499963134527206, 'train/grad_norm_pre_clip': 4.526166915893555, 'timing/data': 0.007102351984940469, 'timing/model': 1.540485773992259, 'learning_rate/qwen_vl_interface': 4.657688670114135e-06, 'learning_rate/action_model': 4.657688670114135e-05, 'epoch': 1.18}) +[2026-06-16 01:53:21,959][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.08347100764513016, 'train/grad_norm_pre_clip': 7.299714088439941, 'timing/data': 0.007177811989095062, 'timing/model': 1.4600700849550776, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 1.18}) +[2026-06-16 01:53:27,956][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.03289385512471199, 'train/grad_norm_pre_clip': 2.2989866733551025, 'timing/data': 0.006780789000913501, 'timing/model': 1.4718060540035367, 'learning_rate/qwen_vl_interface': 4.651645855743825e-06, 'learning_rate/action_model': 4.651645855743825e-05, 'epoch': 1.18}) +[2026-06-16 01:53:33,913][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.06412855535745621, 'train/grad_norm_pre_clip': 4.784324645996094, 'timing/data': 0.007026921026408672, 'timing/model': 1.70822649798356, 'learning_rate/qwen_vl_interface': 4.648624816878883e-06, 'learning_rate/action_model': 4.6486248168788825e-05, 'epoch': 1.18}) +[2026-06-16 01:53:39,670][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.14554618299007416, 'train/grad_norm_pre_clip': 10.454535484313965, 'timing/data': 0.008912394987419248, 'timing/model': 1.481779070047196, 'learning_rate/qwen_vl_interface': 4.6456040252165615e-06, 'learning_rate/action_model': 4.645604025216562e-05, 'epoch': 1.18}) +[2026-06-16 01:53:45,763][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.1117718294262886, 'train/grad_norm_pre_clip': 12.168146133422852, 'timing/data': 0.007379740010946989, 'timing/model': 1.4777911499841139, 'learning_rate/qwen_vl_interface': 4.642583481998596e-06, 'learning_rate/action_model': 4.642583481998596e-05, 'epoch': 1.18}) +[2026-06-16 01:53:51,661][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.0722135677933693, 'train/grad_norm_pre_clip': 8.997823715209961, 'timing/data': 0.0070595889701507986, 'timing/model': 1.4550253780325875, 'learning_rate/qwen_vl_interface': 4.639563188466616e-06, 'learning_rate/action_model': 4.6395631884666165e-05, 'epoch': 1.18}) +[2026-06-16 01:53:57,763][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.4122694134712219, 'train/grad_norm_pre_clip': 9.9341402053833, 'timing/data': 0.006952591007575393, 'timing/model': 1.4585728169768117, 'learning_rate/qwen_vl_interface': 4.636543145862152e-06, 'learning_rate/action_model': 4.6365431458621523e-05, 'epoch': 1.18}) +[2026-06-16 01:54:03,664][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.08027712255716324, 'train/grad_norm_pre_clip': 9.071745872497559, 'timing/data': 0.0071584240067750216, 'timing/model': 1.4723881380050443, 'learning_rate/qwen_vl_interface': 4.633523355426629e-06, 'learning_rate/action_model': 4.6335233554266295e-05, 'epoch': 1.18}) +[2026-06-16 01:54:09,609][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.02122291922569275, 'train/grad_norm_pre_clip': 3.1780877113342285, 'timing/data': 0.006934554025065154, 'timing/model': 1.7049916929681785, 'learning_rate/qwen_vl_interface': 4.6305038184013675e-06, 'learning_rate/action_model': 4.6305038184013675e-05, 'epoch': 1.18}) +[2026-06-16 01:54:15,262][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.13066594302654266, 'train/grad_norm_pre_clip': 11.241622924804688, 'timing/data': 0.007116853026673198, 'timing/model': 1.45785370800877, 'learning_rate/qwen_vl_interface': 4.627484536027589e-06, 'learning_rate/action_model': 4.627484536027588e-05, 'epoch': 1.19}) +[2026-06-16 01:54:21,262][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.10279872268438339, 'train/grad_norm_pre_clip': 6.399387359619141, 'timing/data': 0.0068075580056756735, 'timing/model': 1.4602926359511912, 'learning_rate/qwen_vl_interface': 4.624465509546399e-06, 'learning_rate/action_model': 4.6244655095463986e-05, 'epoch': 1.19}) +[2026-06-16 01:54:27,264][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.17421355843544006, 'train/grad_norm_pre_clip': 12.967561721801758, 'timing/data': 0.006962315994314849, 'timing/model': 1.4877523150062189, 'learning_rate/qwen_vl_interface': 4.6214467401988106e-06, 'learning_rate/action_model': 4.621446740198811e-05, 'epoch': 1.19}) +[2026-06-16 01:54:33,245][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.1699661910533905, 'train/grad_norm_pre_clip': 13.225321769714355, 'timing/data': 0.006676124990917742, 'timing/model': 1.5133121980470605, 'learning_rate/qwen_vl_interface': 4.6184282292257245e-06, 'learning_rate/action_model': 4.618428229225725e-05, 'epoch': 1.19}) +[2026-06-16 01:54:39,167][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.015789886936545372, 'train/grad_norm_pre_clip': 1.7151228189468384, 'timing/data': 0.007010395987890661, 'timing/model': 1.444750783033669, 'learning_rate/qwen_vl_interface': 4.615409977867938e-06, 'learning_rate/action_model': 4.615409977867937e-05, 'epoch': 1.19}) +[2026-06-16 01:54:45,096][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.013701858930289745, 'train/grad_norm_pre_clip': 1.2025617361068726, 'timing/data': 0.007054871995933354, 'timing/model': 1.6899528980138712, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 1.19}) +[2026-06-16 01:54:50,758][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.18920835852622986, 'train/grad_norm_pre_clip': 15.58239459991455, 'timing/data': 0.0065290380152873695, 'timing/model': 1.4538537680055015, 'learning_rate/qwen_vl_interface': 4.609374258960908e-06, 'learning_rate/action_model': 4.6093742589609074e-05, 'epoch': 1.19}) +[2026-06-16 01:54:56,867][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.3584156930446625, 'train/grad_norm_pre_clip': 11.526604652404785, 'timing/data': 0.006834506988525391, 'timing/model': 1.4804592300206423, 'learning_rate/qwen_vl_interface': 4.60635679389272e-06, 'learning_rate/action_model': 4.60635679389272e-05, 'epoch': 1.19}) +[2026-06-16 01:55:02,966][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.05181890353560448, 'train/grad_norm_pre_clip': 5.305007457733154, 'timing/data': 0.006793606968130916, 'timing/model': 1.5010709820198826, 'learning_rate/qwen_vl_interface': 4.603339593401943e-06, 'learning_rate/action_model': 4.603339593401942e-05, 'epoch': 1.19}) +[2026-06-16 01:55:08,868][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.27871426939964294, 'train/grad_norm_pre_clip': 16.79753875732422, 'timing/data': 0.006863380025606602, 'timing/model': 1.4619801650405861, 'learning_rate/qwen_vl_interface': 4.600322658728832e-06, 'learning_rate/action_model': 4.600322658728832e-05, 'epoch': 1.19}) +[2026-06-16 01:55:14,976][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.18453961610794067, 'train/grad_norm_pre_clip': 15.843099594116211, 'timing/data': 0.009153830003924668, 'timing/model': 1.6325590260094032, 'learning_rate/qwen_vl_interface': 4.597305991113537e-06, 'learning_rate/action_model': 4.597305991113537e-05, 'epoch': 1.19}) +[2026-06-16 01:55:20,898][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.06566300243139267, 'train/grad_norm_pre_clip': 14.787493705749512, 'timing/data': 0.006626792019233108, 'timing/model': 1.6893872130312957, 'learning_rate/qwen_vl_interface': 4.5942895917960935e-06, 'learning_rate/action_model': 4.5942895917960935e-05, 'epoch': 1.19}) +[2026-06-16 01:55:26,590][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.039653439074754715, 'train/grad_norm_pre_clip': 3.63214111328125, 'timing/data': 0.007134969986509532, 'timing/model': 1.4792829869547859, 'learning_rate/qwen_vl_interface': 4.59127346201643e-06, 'learning_rate/action_model': 4.59127346201643e-05, 'epoch': 1.19}) +[2026-06-16 01:55:32,642][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.13491316139698029, 'train/grad_norm_pre_clip': 29.49590301513672, 'timing/data': 0.0065539240022189915, 'timing/model': 1.5126270549953915, 'learning_rate/qwen_vl_interface': 4.588257603014364e-06, 'learning_rate/action_model': 4.588257603014364e-05, 'epoch': 1.19}) +[2026-06-16 01:55:38,567][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.006252617575228214, 'train/grad_norm_pre_clip': 1.7385047674179077, 'timing/data': 0.00670387496938929, 'timing/model': 1.4576576619874686, 'learning_rate/qwen_vl_interface': 4.585242016029602e-06, 'learning_rate/action_model': 4.5852420160296015e-05, 'epoch': 1.19}) +[2026-06-16 01:55:44,582][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.02551470696926117, 'train/grad_norm_pre_clip': 4.840194225311279, 'timing/data': 0.0068186940043233335, 'timing/model': 1.461385307018645, 'learning_rate/qwen_vl_interface': 4.582226702301737e-06, 'learning_rate/action_model': 4.5822267023017375e-05, 'epoch': 1.19}) +[2026-06-16 01:55:50,765][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.42524901032447815, 'train/grad_norm_pre_clip': 19.840200424194336, 'timing/data': 0.006872703030239791, 'timing/model': 1.5176155139924958, 'learning_rate/qwen_vl_interface': 4.57921166307025e-06, 'learning_rate/action_model': 4.57921166307025e-05, 'epoch': 1.19}) +[2026-06-16 01:55:56,702][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.40515077114105225, 'train/grad_norm_pre_clip': 20.26374626159668, 'timing/data': 0.0070435560191981494, 'timing/model': 1.690731116977986, 'learning_rate/qwen_vl_interface': 4.576196899574511e-06, 'learning_rate/action_model': 4.576196899574511e-05, 'epoch': 1.19}) +[2026-06-16 01:56:02,369][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.07032635807991028, 'train/grad_norm_pre_clip': 7.205132484436035, 'timing/data': 0.007036364986561239, 'timing/model': 1.457114169956185, 'learning_rate/qwen_vl_interface': 4.573182413053775e-06, 'learning_rate/action_model': 4.5731824130537746e-05, 'epoch': 1.19}) +[2026-06-16 01:56:08,369][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.27710357308387756, 'train/grad_norm_pre_clip': 24.52092933654785, 'timing/data': 0.007160256966017187, 'timing/model': 1.4599951400305144, 'learning_rate/qwen_vl_interface': 4.570168204747183e-06, 'learning_rate/action_model': 4.570168204747183e-05, 'epoch': 1.19}) +[2026-06-16 01:56:14,467][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.02230445295572281, 'train/grad_norm_pre_clip': 3.2887980937957764, 'timing/data': 0.0065949150011874735, 'timing/model': 1.4582707619993016, 'learning_rate/qwen_vl_interface': 4.567154275893762e-06, 'learning_rate/action_model': 4.5671542758937624e-05, 'epoch': 1.19}) +[2026-06-16 01:56:20,367][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.012601234018802643, 'train/grad_norm_pre_clip': 1.4761277437210083, 'timing/data': 0.006797052046749741, 'timing/model': 1.457343367044814, 'learning_rate/qwen_vl_interface': 4.564140627732426e-06, 'learning_rate/action_model': 4.564140627732426e-05, 'epoch': 1.19}) +[2026-06-16 01:56:26,367][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.038851991295814514, 'train/grad_norm_pre_clip': 4.683666229248047, 'timing/data': 0.006624788977205753, 'timing/model': 1.5211064190370962, 'learning_rate/qwen_vl_interface': 4.56112726150197e-06, 'learning_rate/action_model': 4.56112726150197e-05, 'epoch': 1.19}) +[2026-06-16 01:56:32,303][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.1010313481092453, 'train/grad_norm_pre_clip': 8.94746208190918, 'timing/data': 0.00681550899753347, 'timing/model': 1.6928997949580662, 'learning_rate/qwen_vl_interface': 4.558114178441076e-06, 'learning_rate/action_model': 4.558114178441076e-05, 'epoch': 1.2}) +[2026-06-16 01:56:37,969][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.049347665160894394, 'train/grad_norm_pre_clip': 7.188389301300049, 'timing/data': 0.006799466034863144, 'timing/model': 1.4594263799954206, 'learning_rate/qwen_vl_interface': 4.555101379788308e-06, 'learning_rate/action_model': 4.555101379788308e-05, 'epoch': 1.2}) +[2026-06-16 01:56:43,972][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.11601325869560242, 'train/grad_norm_pre_clip': 4.961357593536377, 'timing/data': 0.006725315994117409, 'timing/model': 1.4643208329798654, 'learning_rate/qwen_vl_interface': 4.552088866782115e-06, 'learning_rate/action_model': 4.552088866782115e-05, 'epoch': 1.2}) +[2026-06-16 01:56:49,969][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.056333642452955246, 'train/grad_norm_pre_clip': 7.104182720184326, 'timing/data': 0.017570536001585424, 'timing/model': 1.476630386954639, 'learning_rate/qwen_vl_interface': 4.5490766406608265e-06, 'learning_rate/action_model': 4.549076640660827e-05, 'epoch': 1.2}) +[2026-06-16 01:56:56,238][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.2315048724412918, 'train/grad_norm_pre_clip': 6.627114772796631, 'timing/data': 0.007934295979794115, 'timing/model': 1.7330811040010303, 'learning_rate/qwen_vl_interface': 4.546064702662655e-06, 'learning_rate/action_model': 4.546064702662655e-05, 'epoch': 1.2}) +[2026-06-16 01:57:02,275][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.004948961548507214, 'train/grad_norm_pre_clip': 0.5010130405426025, 'timing/data': 0.007156561012379825, 'timing/model': 1.5322148129926063, 'learning_rate/qwen_vl_interface': 4.543053054025693e-06, 'learning_rate/action_model': 4.5430530540256926e-05, 'epoch': 1.2}) +[2026-06-16 01:57:08,333][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.07560974359512329, 'train/grad_norm_pre_clip': 6.989691734313965, 'timing/data': 0.007671980012673885, 'timing/model': 1.7302871320280246, 'learning_rate/qwen_vl_interface': 4.540041695987917e-06, 'learning_rate/action_model': 4.540041695987917e-05, 'epoch': 1.2}) +[2026-06-16 01:57:14,399][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.06708935648202896, 'train/grad_norm_pre_clip': 8.889269828796387, 'timing/data': 0.007354683009907603, 'timing/model': 1.7023433989961632, 'learning_rate/qwen_vl_interface': 4.537030629787182e-06, 'learning_rate/action_model': 4.5370306297871826e-05, 'epoch': 1.2}) +[2026-06-16 01:57:20,256][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.09602934122085571, 'train/grad_norm_pre_clip': 10.727710723876953, 'timing/data': 0.006642394990194589, 'timing/model': 1.4352974210050888, 'learning_rate/qwen_vl_interface': 4.534019856661226e-06, 'learning_rate/action_model': 4.534019856661226e-05, 'epoch': 1.2}) +[2026-06-16 01:57:26,226][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.2290954887866974, 'train/grad_norm_pre_clip': 15.088663101196289, 'timing/data': 0.0065362880122847855, 'timing/model': 1.5274277270073071, 'learning_rate/qwen_vl_interface': 4.531009377847661e-06, 'learning_rate/action_model': 4.531009377847661e-05, 'epoch': 1.2}) +[2026-06-16 01:57:32,265][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.14806868135929108, 'train/grad_norm_pre_clip': 11.7061767578125, 'timing/data': 0.006650797033216804, 'timing/model': 1.471267431974411, 'learning_rate/qwen_vl_interface': 4.527999194583982e-06, 'learning_rate/action_model': 4.527999194583981e-05, 'epoch': 1.2}) +[2026-06-16 01:57:38,258][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.25853005051612854, 'train/grad_norm_pre_clip': 8.753658294677734, 'timing/data': 0.0067913440288975835, 'timing/model': 1.49497809400782, 'learning_rate/qwen_vl_interface': 4.52498930810756e-06, 'learning_rate/action_model': 4.52498930810756e-05, 'epoch': 1.2}) +[2026-06-16 01:57:44,196][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.051120128482580185, 'train/grad_norm_pre_clip': 6.790086269378662, 'timing/data': 0.006533124018460512, 'timing/model': 1.6906398650025949, 'learning_rate/qwen_vl_interface': 4.521979719655648e-06, 'learning_rate/action_model': 4.5219797196556484e-05, 'epoch': 1.2}) +[2026-06-16 01:57:50,086][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.05726223438978195, 'train/grad_norm_pre_clip': 13.32571792602539, 'timing/data': 0.00663430301938206, 'timing/model': 1.6862722969963215, 'learning_rate/qwen_vl_interface': 4.518970430465377e-06, 'learning_rate/action_model': 4.518970430465376e-05, 'epoch': 1.2}) +[2026-06-16 01:57:55,756][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.06545447558164597, 'train/grad_norm_pre_clip': 10.921631813049316, 'timing/data': 0.0065833679982461035, 'timing/model': 1.459269265003968, 'learning_rate/qwen_vl_interface': 4.515961441773745e-06, 'learning_rate/action_model': 4.515961441773745e-05, 'epoch': 1.2}) +[2026-06-16 01:58:01,656][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.3126036822795868, 'train/grad_norm_pre_clip': 16.97085189819336, 'timing/data': 0.006697576027363539, 'timing/model': 1.4596211119787768, 'learning_rate/qwen_vl_interface': 4.5129527548176385e-06, 'learning_rate/action_model': 4.512952754817638e-05, 'epoch': 1.2}) +[2026-06-16 01:58:07,554][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.051832959055900574, 'train/grad_norm_pre_clip': 4.846492290496826, 'timing/data': 0.0067780750105157495, 'timing/model': 1.4571955620194785, 'learning_rate/qwen_vl_interface': 4.509944370833813e-06, 'learning_rate/action_model': 4.509944370833813e-05, 'epoch': 1.2}) +[2026-06-16 01:58:13,561][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.21907369792461395, 'train/grad_norm_pre_clip': 14.825118064880371, 'timing/data': 0.006948725960683078, 'timing/model': 1.5283230119966902, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 1.2}) +[2026-06-16 01:58:19,589][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.038216885179281235, 'train/grad_norm_pre_clip': 8.03956127166748, 'timing/data': 0.006600793974939734, 'timing/model': 1.6928853269782849, 'learning_rate/qwen_vl_interface': 4.503928516729416e-06, 'learning_rate/action_model': 4.503928516729416e-05, 'epoch': 1.2}) +[2026-06-16 01:58:25,359][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.2523016929626465, 'train/grad_norm_pre_clip': 18.015762329101562, 'timing/data': 0.0068405159981921315, 'timing/model': 1.4588268520310521, 'learning_rate/qwen_vl_interface': 4.500921049081734e-06, 'learning_rate/action_model': 4.500921049081734e-05, 'epoch': 1.2}) +[2026-06-16 01:58:31,531][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.10002405941486359, 'train/grad_norm_pre_clip': 5.02443265914917, 'timing/data': 0.008122613013256341, 'timing/model': 1.5444097639992833, 'learning_rate/qwen_vl_interface': 4.497913889352113e-06, 'learning_rate/action_model': 4.4979138893521126e-05, 'epoch': 1.2}) +[2026-06-16 01:58:37,557][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.14292900264263153, 'train/grad_norm_pre_clip': 5.637599468231201, 'timing/data': 0.0070652169524692, 'timing/model': 1.4616971149807796, 'learning_rate/qwen_vl_interface': 4.494907038776684e-06, 'learning_rate/action_model': 4.4949070387766836e-05, 'epoch': 1.2}) +[2026-06-16 02:07:42,551][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.1960277408361435, 'train/grad_norm_pre_clip': 9.529767036437988, 'eval/loss': 0.05242224410176277, 'eval/flap_precision': 0.9458333333333333, 'eval/flap_recall': 0.9098196392785571, 'eval/flap_f1': 0.9274770173646578, 'eval/flap_rate_model': 0.13333333333333333, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9274770173646578, 'timing/data': 0.0068968599662184715, 'timing/model': 1.4350644330261275, 'learning_rate/qwen_vl_interface': 4.49190049859145e-06, 'learning_rate/action_model': 4.49190049859145e-05, 'epoch': 1.2}) +[2026-06-16 02:07:42,552][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2800_state +[2026-06-16 02:07:42,556][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:07:48,239][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2800_state/model.safetensors +[2026-06-16 02:08:01,875][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2800_state/optimizer.bin +[2026-06-16 02:08:01,875][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2800_state/sampler.bin +[2026-06-16 02:08:01,875][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2800_state/sampler_1.bin +[2026-06-16 02:08:01,876][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2800_state/random_states_0.pkl +[2026-06-16 02:08:01,988][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2800 +[2026-06-16 02:08:01,989][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:08:01,999][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:08:08,146][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.22960297763347626, 'train/grad_norm_pre_clip': 14.989761352539062, 'timing/data': 0.006744584999978542, 'timing/model': 1.4771907559479587, 'learning_rate/qwen_vl_interface': 4.488894270032284e-06, 'learning_rate/action_model': 4.4888942700322834e-05, 'epoch': 1.21}) +[2026-06-16 02:08:14,180][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.03491593897342682, 'train/grad_norm_pre_clip': 2.8231139183044434, 'timing/data': 0.006551511003635824, 'timing/model': 1.4742325720144436, 'learning_rate/qwen_vl_interface': 4.485888354334933e-06, 'learning_rate/action_model': 4.4858883543349334e-05, 'epoch': 1.21}) +[2026-06-16 02:08:20,287][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.10273132473230362, 'train/grad_norm_pre_clip': 8.796781539916992, 'timing/data': 0.00643790396861732, 'timing/model': 1.6954508789931424, 'learning_rate/qwen_vl_interface': 4.482882752735017e-06, 'learning_rate/action_model': 4.482882752735017e-05, 'epoch': 1.21}) +[2026-06-16 02:08:26,257][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.11901140958070755, 'train/grad_norm_pre_clip': 5.716528415679932, 'timing/data': 0.006574324972461909, 'timing/model': 1.6564117399975657, 'learning_rate/qwen_vl_interface': 4.479877466468027e-06, 'learning_rate/action_model': 4.479877466468026e-05, 'epoch': 1.21}) +[2026-06-16 02:08:32,058][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.12286657840013504, 'train/grad_norm_pre_clip': 11.616394996643066, 'timing/data': 0.0067529669613577425, 'timing/model': 1.4595550239901058, 'learning_rate/qwen_vl_interface': 4.476872496769318e-06, 'learning_rate/action_model': 4.476872496769318e-05, 'epoch': 1.21}) +[2026-06-16 02:08:38,060][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.07274816930294037, 'train/grad_norm_pre_clip': 4.643102645874023, 'timing/data': 0.006505492958240211, 'timing/model': 1.4605409140349366, 'learning_rate/qwen_vl_interface': 4.473867844874122e-06, 'learning_rate/action_model': 4.473867844874122e-05, 'epoch': 1.21}) +[2026-06-16 02:08:44,161][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.03525814786553383, 'train/grad_norm_pre_clip': 3.086559772491455, 'timing/data': 0.006691496993880719, 'timing/model': 1.4598956750123762, 'learning_rate/qwen_vl_interface': 4.470863512017539e-06, 'learning_rate/action_model': 4.470863512017539e-05, 'epoch': 1.21}) +[2026-06-16 02:08:50,319][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.2780378460884094, 'train/grad_norm_pre_clip': 12.36790657043457, 'timing/data': 0.006707599968649447, 'timing/model': 1.574554982013069, 'learning_rate/qwen_vl_interface': 4.4678594994345344e-06, 'learning_rate/action_model': 4.467859499434534e-05, 'epoch': 1.21}) +[2026-06-16 02:08:56,557][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.15185761451721191, 'train/grad_norm_pre_clip': 9.067157745361328, 'timing/data': 0.007404606032650918, 'timing/model': 1.5414963549701497, 'learning_rate/qwen_vl_interface': 4.4648558083599475e-06, 'learning_rate/action_model': 4.464855808359947e-05, 'epoch': 1.21}) +[2026-06-16 02:09:02,753][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.09923318028450012, 'train/grad_norm_pre_clip': 6.668928623199463, 'timing/data': 0.006819494999945164, 'timing/model': 1.6832823010045104, 'learning_rate/qwen_vl_interface': 4.461852440028479e-06, 'learning_rate/action_model': 4.461852440028479e-05, 'epoch': 1.21}) +[2026-06-16 02:09:08,740][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.12036070227622986, 'train/grad_norm_pre_clip': 6.187668800354004, 'timing/data': 0.006914263998623937, 'timing/model': 1.538503630028572, 'learning_rate/qwen_vl_interface': 4.458849395674702e-06, 'learning_rate/action_model': 4.458849395674701e-05, 'epoch': 1.21}) +[2026-06-16 02:09:15,048][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.13306687772274017, 'train/grad_norm_pre_clip': 6.9867730140686035, 'timing/data': 0.006663746025878936, 'timing/model': 1.5713730289717205, 'learning_rate/qwen_vl_interface': 4.455846676533052e-06, 'learning_rate/action_model': 4.4558466765330525e-05, 'epoch': 1.21}) +[2026-06-16 02:09:21,237][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.07857256382703781, 'train/grad_norm_pre_clip': 3.6948540210723877, 'timing/data': 0.0069003350217826664, 'timing/model': 1.5224004329647869, 'learning_rate/qwen_vl_interface': 4.452844283837838e-06, 'learning_rate/action_model': 4.452844283837838e-05, 'epoch': 1.21}) +[2026-06-16 02:09:27,391][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.343155175447464, 'train/grad_norm_pre_clip': 9.331598281860352, 'timing/data': 0.006722873018588871, 'timing/model': 1.4813270819722675, 'learning_rate/qwen_vl_interface': 4.449842218823228e-06, 'learning_rate/action_model': 4.449842218823228e-05, 'epoch': 1.21}) +[2026-06-16 02:09:33,531][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.057404857128858566, 'train/grad_norm_pre_clip': 2.833801746368408, 'timing/data': 0.006821397983003408, 'timing/model': 1.7270072940154932, 'learning_rate/qwen_vl_interface': 4.446840482723256e-06, 'learning_rate/action_model': 4.446840482723255e-05, 'epoch': 1.21}) +[2026-06-16 02:09:39,537][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.05907130986452103, 'train/grad_norm_pre_clip': 3.3542871475219727, 'timing/data': 0.006499244016595185, 'timing/model': 1.536845539987553, 'learning_rate/qwen_vl_interface': 4.443839076771823e-06, 'learning_rate/action_model': 4.443839076771823e-05, 'epoch': 1.21}) +[2026-06-16 02:09:45,829][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.11926876753568649, 'train/grad_norm_pre_clip': 10.76960563659668, 'timing/data': 0.007169199991039932, 'timing/model': 1.5479836100130342, 'learning_rate/qwen_vl_interface': 4.4408380022026945e-06, 'learning_rate/action_model': 4.440838002202694e-05, 'epoch': 1.21}) +[2026-06-16 02:09:51,859][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.1329154074192047, 'train/grad_norm_pre_clip': 9.668947219848633, 'timing/data': 0.0066721480106934905, 'timing/model': 1.4983717830036767, 'learning_rate/qwen_vl_interface': 4.437837260249499e-06, 'learning_rate/action_model': 4.437837260249499e-05, 'epoch': 1.21}) +[2026-06-16 02:09:57,887][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.2074097990989685, 'train/grad_norm_pre_clip': 7.116100311279297, 'timing/data': 0.006361340987496078, 'timing/model': 1.49433819798287, 'learning_rate/qwen_vl_interface': 4.434836852145728e-06, 'learning_rate/action_model': 4.434836852145728e-05, 'epoch': 1.21}) +[2026-06-16 02:10:03,862][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.06914981454610825, 'train/grad_norm_pre_clip': 4.473045349121094, 'timing/data': 0.006530709972139448, 'timing/model': 1.5084788859821856, 'learning_rate/qwen_vl_interface': 4.431836779124736e-06, 'learning_rate/action_model': 4.431836779124735e-05, 'epoch': 1.21}) +[2026-06-16 02:10:10,154][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.2770428955554962, 'train/grad_norm_pre_clip': 8.081863403320312, 'timing/data': 0.00685785204404965, 'timing/model': 1.8039691179874353, 'learning_rate/qwen_vl_interface': 4.428837042419738e-06, 'learning_rate/action_model': 4.428837042419738e-05, 'epoch': 1.21}) +[2026-06-16 02:10:16,061][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.1457538604736328, 'train/grad_norm_pre_clip': 3.4165878295898438, 'timing/data': 0.006370514049194753, 'timing/model': 1.4573201530147344, 'learning_rate/qwen_vl_interface': 4.425837643263815e-06, 'learning_rate/action_model': 4.425837643263814e-05, 'epoch': 1.21}) +[2026-06-16 02:10:22,093][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.048091597855091095, 'train/grad_norm_pre_clip': 2.4438774585723877, 'timing/data': 0.006285999959800392, 'timing/model': 1.488417084037792, 'learning_rate/qwen_vl_interface': 4.4228385828899045e-06, 'learning_rate/action_model': 4.4228385828899046e-05, 'epoch': 1.21}) +[2026-06-16 02:10:28,164][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.0839269608259201, 'train/grad_norm_pre_clip': 4.362938404083252, 'timing/data': 0.006139816017821431, 'timing/model': 1.4816745920106769, 'learning_rate/qwen_vl_interface': 4.419839862530809e-06, 'learning_rate/action_model': 4.419839862530809e-05, 'epoch': 1.22}) +[2026-06-16 02:10:34,362][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.09097287058830261, 'train/grad_norm_pre_clip': 6.033321380615234, 'timing/data': 0.006732057023327798, 'timing/model': 1.5008089089533314, 'learning_rate/qwen_vl_interface': 4.416841483419188e-06, 'learning_rate/action_model': 4.416841483419188e-05, 'epoch': 1.22}) +[2026-06-16 02:10:40,535][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.011861943639814854, 'train/grad_norm_pre_clip': 1.058061957359314, 'timing/data': 0.006653221033047885, 'timing/model': 1.5665598459891044, 'learning_rate/qwen_vl_interface': 4.413843446787561e-06, 'learning_rate/action_model': 4.413843446787561e-05, 'epoch': 1.22}) +[2026-06-16 02:10:46,768][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.028870543465018272, 'train/grad_norm_pre_clip': 1.5673625469207764, 'timing/data': 0.006762681994587183, 'timing/model': 1.7685284540057182, 'learning_rate/qwen_vl_interface': 4.41084575386831e-06, 'learning_rate/action_model': 4.4108457538683093e-05, 'epoch': 1.22}) +[2026-06-16 02:10:52,736][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.02805597521364689, 'train/grad_norm_pre_clip': 1.9812982082366943, 'timing/data': 0.006562447000760585, 'timing/model': 1.5318958350108005, 'learning_rate/qwen_vl_interface': 4.40784840589367e-06, 'learning_rate/action_model': 4.4078484058936706e-05, 'epoch': 1.22}) +[2026-06-16 02:10:58,993][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.045952435582876205, 'train/grad_norm_pre_clip': 2.3622984886169434, 'timing/data': 0.006866804964374751, 'timing/model': 1.521355754986871, 'learning_rate/qwen_vl_interface': 4.404851404095742e-06, 'learning_rate/action_model': 4.404851404095742e-05, 'epoch': 1.22}) +[2026-06-16 02:11:05,226][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.09409916400909424, 'train/grad_norm_pre_clip': 7.703947067260742, 'timing/data': 0.006599231041036546, 'timing/model': 1.5515138740302064, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 1.22}) +[2026-06-16 02:11:11,395][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.08646611124277115, 'train/grad_norm_pre_clip': 4.942317485809326, 'timing/data': 0.0066395309986546636, 'timing/model': 1.4900466519757174, 'learning_rate/qwen_vl_interface': 4.398858443957686e-06, 'learning_rate/action_model': 4.398858443957686e-05, 'epoch': 1.22}) +[2026-06-16 02:11:17,521][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.24892349541187286, 'train/grad_norm_pre_clip': 7.871582984924316, 'timing/data': 0.006375272001605481, 'timing/model': 1.558671637030784, 'learning_rate/qwen_vl_interface': 4.395862488081038e-06, 'learning_rate/action_model': 4.395862488081038e-05, 'epoch': 1.22}) +[2026-06-16 02:11:23,756][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.12809604406356812, 'train/grad_norm_pre_clip': 8.380989074707031, 'timing/data': 0.006758395989891142, 'timing/model': 1.797214096004609, 'learning_rate/qwen_vl_interface': 4.392866883308059e-06, 'learning_rate/action_model': 4.392866883308059e-05, 'epoch': 1.22}) +[2026-06-16 02:11:29,866][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.06551273912191391, 'train/grad_norm_pre_clip': 7.778074741363525, 'timing/data': 0.006313291029073298, 'timing/model': 1.5959631229634397, 'learning_rate/qwen_vl_interface': 4.389871630870127e-06, 'learning_rate/action_model': 4.389871630870127e-05, 'epoch': 1.22}) +[2026-06-16 02:11:36,129][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.08508456498384476, 'train/grad_norm_pre_clip': 11.985647201538086, 'timing/data': 0.00616347003960982, 'timing/model': 1.5617041289806366, 'learning_rate/qwen_vl_interface': 4.386876731998478e-06, 'learning_rate/action_model': 4.386876731998478e-05, 'epoch': 1.22}) +[2026-06-16 02:11:42,264][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.010687187314033508, 'train/grad_norm_pre_clip': 1.6260790824890137, 'timing/data': 0.006475218979176134, 'timing/model': 1.4799897909979336, 'learning_rate/qwen_vl_interface': 4.3838821879242006e-06, 'learning_rate/action_model': 4.3838821879242e-05, 'epoch': 1.22}) +[2026-06-16 02:11:48,276][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.3636457622051239, 'train/grad_norm_pre_clip': 15.666483879089355, 'timing/data': 0.006739517964888364, 'timing/model': 1.5394924730062485, 'learning_rate/qwen_vl_interface': 4.38088799987824e-06, 'learning_rate/action_model': 4.38088799987824e-05, 'epoch': 1.22}) +[2026-06-16 02:11:54,342][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.06428512185811996, 'train/grad_norm_pre_clip': 5.859727382659912, 'timing/data': 0.007270939997397363, 'timing/model': 1.5124666439951397, 'learning_rate/qwen_vl_interface': 4.3778941690913935e-06, 'learning_rate/action_model': 4.3778941690913934e-05, 'epoch': 1.22}) +[2026-06-16 02:12:00,297][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.11309387534856796, 'train/grad_norm_pre_clip': 18.639991760253906, 'timing/data': 0.006394649972207844, 'timing/model': 1.6922964829718694, 'learning_rate/qwen_vl_interface': 4.374900696794311e-06, 'learning_rate/action_model': 4.374900696794311e-05, 'epoch': 1.22}) +[2026-06-16 02:12:06,155][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.015013417229056358, 'train/grad_norm_pre_clip': 1.3067562580108643, 'timing/data': 0.006171211018227041, 'timing/model': 1.5181466409703717, 'learning_rate/qwen_vl_interface': 4.371907584217499e-06, 'learning_rate/action_model': 4.371907584217498e-05, 'epoch': 1.22}) +[2026-06-16 02:12:12,469][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.14986267685890198, 'train/grad_norm_pre_clip': 16.773530960083008, 'timing/data': 0.006699769990518689, 'timing/model': 1.6192902399925515, 'learning_rate/qwen_vl_interface': 4.368914832591309e-06, 'learning_rate/action_model': 4.368914832591309e-05, 'epoch': 1.22}) +[2026-06-16 02:12:18,633][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.20417194068431854, 'train/grad_norm_pre_clip': 70.3316879272461, 'timing/data': 0.007942477997858077, 'timing/model': 1.5208933430258185, 'learning_rate/qwen_vl_interface': 4.365922443145949e-06, 'learning_rate/action_model': 4.365922443145949e-05, 'epoch': 1.22}) +[2026-06-16 02:12:24,829][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.029470739886164665, 'train/grad_norm_pre_clip': 2.9734244346618652, 'timing/data': 0.006919412000570446, 'timing/model': 1.523558223969303, 'learning_rate/qwen_vl_interface': 4.362930417111479e-06, 'learning_rate/action_model': 4.362930417111479e-05, 'epoch': 1.22}) +[2026-06-16 02:12:31,026][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.4272174537181854, 'train/grad_norm_pre_clip': 18.93204116821289, 'timing/data': 0.006921955966390669, 'timing/model': 1.5031651950557716, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 1.22}) +[2026-06-16 02:12:37,031][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.009341404773294926, 'train/grad_norm_pre_clip': 1.999179720878601, 'timing/data': 0.007667622994631529, 'timing/model': 1.7198297689901665, 'learning_rate/qwen_vl_interface': 4.3569474601946934e-06, 'learning_rate/action_model': 4.356947460194694e-05, 'epoch': 1.22}) +[2026-06-16 02:12:42,767][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.018610753118991852, 'train/grad_norm_pre_clip': 2.4086661338806152, 'timing/data': 0.006625288981013, 'timing/model': 1.4558700680499896, 'learning_rate/qwen_vl_interface': 4.3539565317717446e-06, 'learning_rate/action_model': 4.353956531771745e-05, 'epoch': 1.22}) +[2026-06-16 02:12:48,770][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.03461349010467529, 'train/grad_norm_pre_clip': 9.100341796875, 'timing/data': 0.007106417964678258, 'timing/model': 1.4561530440114439, 'learning_rate/qwen_vl_interface': 4.350965971678419e-06, 'learning_rate/action_model': 4.3509659716784185e-05, 'epoch': 1.22}) +[2026-06-16 02:12:54,767][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.04466361179947853, 'train/grad_norm_pre_clip': 5.122725486755371, 'timing/data': 0.006826816010288894, 'timing/model': 1.4574554199934937, 'learning_rate/qwen_vl_interface': 4.3479757811440225e-06, 'learning_rate/action_model': 4.347975781144022e-05, 'epoch': 1.23}) +[2026-06-16 02:13:00,843][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.02989877015352249, 'train/grad_norm_pre_clip': 3.909470319747925, 'timing/data': 0.006734941038303077, 'timing/model': 1.5803341229911894, 'learning_rate/qwen_vl_interface': 4.34498596139771e-06, 'learning_rate/action_model': 4.34498596139771e-05, 'epoch': 1.23}) +[2026-06-16 02:13:07,156][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.0861169844865799, 'train/grad_norm_pre_clip': 10.520262718200684, 'timing/data': 0.0066151840146631, 'timing/model': 1.5197095799958333, 'learning_rate/qwen_vl_interface': 4.341996513668483e-06, 'learning_rate/action_model': 4.3419965136684824e-05, 'epoch': 1.23}) +[2026-06-16 02:13:13,423][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.3253476619720459, 'train/grad_norm_pre_clip': 18.380754470825195, 'timing/data': 0.007073199027217925, 'timing/model': 1.7705483850440942, 'learning_rate/qwen_vl_interface': 4.33900743918519e-06, 'learning_rate/action_model': 4.33900743918519e-05, 'epoch': 1.23}) +[2026-06-16 02:13:19,312][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.09506435692310333, 'train/grad_norm_pre_clip': 13.304747581481934, 'timing/data': 0.006778394978027791, 'timing/model': 1.5086292859632522, 'learning_rate/qwen_vl_interface': 4.336018739176529e-06, 'learning_rate/action_model': 4.33601873917653e-05, 'epoch': 1.23}) +[2026-06-16 02:13:25,572][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.13921555876731873, 'train/grad_norm_pre_clip': 10.377494812011719, 'timing/data': 0.0066000319784507155, 'timing/model': 1.5873999709729105, 'learning_rate/qwen_vl_interface': 4.333030414871041e-06, 'learning_rate/action_model': 4.33303041487104e-05, 'epoch': 1.23}) +[2026-06-16 02:13:31,563][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.17776252329349518, 'train/grad_norm_pre_clip': 12.490021705627441, 'timing/data': 0.0068245219881646335, 'timing/model': 1.4429717260063626, 'learning_rate/qwen_vl_interface': 4.330042467497111e-06, 'learning_rate/action_model': 4.330042467497111e-05, 'epoch': 1.23}) +[2026-06-16 02:13:37,731][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.23710747063159943, 'train/grad_norm_pre_clip': 8.569461822509766, 'timing/data': 0.0068843800108879805, 'timing/model': 1.5594590640394017, 'learning_rate/qwen_vl_interface': 4.327054898282973e-06, 'learning_rate/action_model': 4.3270548982829727e-05, 'epoch': 1.23}) +[2026-06-16 02:13:43,951][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.06368393450975418, 'train/grad_norm_pre_clip': 4.874399185180664, 'timing/data': 0.006698888028040528, 'timing/model': 1.5116415779921226, 'learning_rate/qwen_vl_interface': 4.324067708456703e-06, 'learning_rate/action_model': 4.3240677084567036e-05, 'epoch': 1.23}) +[2026-06-16 02:13:49,965][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.21714532375335693, 'train/grad_norm_pre_clip': 10.619051933288574, 'timing/data': 0.006599451007787138, 'timing/model': 1.762410629016813, 'learning_rate/qwen_vl_interface': 4.321080899246225e-06, 'learning_rate/action_model': 4.321080899246225e-05, 'epoch': 1.23}) +[2026-06-16 02:13:56,219][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.2587178647518158, 'train/grad_norm_pre_clip': 10.341652870178223, 'timing/data': 0.0068074470036663115, 'timing/model': 1.760049053991679, 'learning_rate/qwen_vl_interface': 4.318094471879299e-06, 'learning_rate/action_model': 4.3180944718792985e-05, 'epoch': 1.23}) +[2026-06-16 02:14:02,234][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.09689730405807495, 'train/grad_norm_pre_clip': 7.179844379425049, 'timing/data': 0.006507807003799826, 'timing/model': 1.5626543359830976, 'learning_rate/qwen_vl_interface': 4.315108427583535e-06, 'learning_rate/action_model': 4.315108427583535e-05, 'epoch': 1.23}) +[2026-06-16 02:14:08,532][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.02462548017501831, 'train/grad_norm_pre_clip': 5.55293607711792, 'timing/data': 0.006880764965899289, 'timing/model': 1.5607433629920706, 'learning_rate/qwen_vl_interface': 4.312122767586381e-06, 'learning_rate/action_model': 4.3121227675863815e-05, 'epoch': 1.23}) +[2026-06-16 02:14:14,728][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.008819733746349812, 'train/grad_norm_pre_clip': 1.0194323062896729, 'timing/data': 0.006801798997912556, 'timing/model': 1.5518913179985248, 'learning_rate/qwen_vl_interface': 4.3091374931151324e-06, 'learning_rate/action_model': 4.309137493115132e-05, 'epoch': 1.23}) +[2026-06-16 02:14:20,861][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.07441361993551254, 'train/grad_norm_pre_clip': 15.32919692993164, 'timing/data': 0.0068732850486412644, 'timing/model': 1.4950467679882422, 'learning_rate/qwen_vl_interface': 4.306152605396921e-06, 'learning_rate/action_model': 4.306152605396921e-05, 'epoch': 1.23}) +[2026-06-16 02:14:26,801][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.024751463904976845, 'train/grad_norm_pre_clip': 3.5845324993133545, 'timing/data': 0.006479625008068979, 'timing/model': 1.6955726430169307, 'learning_rate/qwen_vl_interface': 4.30316810565872e-06, 'learning_rate/action_model': 4.3031681056587206e-05, 'epoch': 1.23}) +[2026-06-16 02:14:32,880][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.020990554243326187, 'train/grad_norm_pre_clip': 3.101585626602173, 'timing/data': 0.007050065032672137, 'timing/model': 1.813257852976676, 'learning_rate/qwen_vl_interface': 4.300183995127347e-06, 'learning_rate/action_model': 4.300183995127346e-05, 'epoch': 1.23}) +[2026-06-16 02:14:38,762][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.2094656229019165, 'train/grad_norm_pre_clip': 9.426695823669434, 'timing/data': 0.006247672950848937, 'timing/model': 1.4938773759640753, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 1.23}) +[2026-06-16 02:14:44,998][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.2934488356113434, 'train/grad_norm_pre_clip': 6.239920616149902, 'timing/data': 0.006701141013763845, 'timing/model': 1.5216536799562164, 'learning_rate/qwen_vl_interface': 4.29421694659154e-06, 'learning_rate/action_model': 4.294216946591539e-05, 'epoch': 1.23}) +[2026-06-16 02:14:51,236][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.0294007770717144, 'train/grad_norm_pre_clip': 4.8852996826171875, 'timing/data': 0.006905130983795971, 'timing/model': 1.5316402160096914, 'learning_rate/qwen_vl_interface': 4.2912340110399345e-06, 'learning_rate/action_model': 4.291234011039934e-05, 'epoch': 1.23}) +[2026-06-16 02:14:57,532][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.024285122752189636, 'train/grad_norm_pre_clip': 3.441959857940674, 'timing/data': 0.006658889004029334, 'timing/model': 1.5226494150119834, 'learning_rate/qwen_vl_interface': 4.2882514696008125e-06, 'learning_rate/action_model': 4.2882514696008125e-05, 'epoch': 1.23}) +[2026-06-16 02:15:03,673][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.033317115157842636, 'train/grad_norm_pre_clip': 5.379231929779053, 'timing/data': 0.006509178027044982, 'timing/model': 1.7750307589885779, 'learning_rate/qwen_vl_interface': 4.2852693235001815e-06, 'learning_rate/action_model': 4.2852693235001815e-05, 'epoch': 1.23}) +[2026-06-16 02:15:09,652][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.23578819632530212, 'train/grad_norm_pre_clip': 9.85019302368164, 'timing/data': 0.007350365980528295, 'timing/model': 1.514536708011292, 'learning_rate/qwen_vl_interface': 4.2822875739638915e-06, 'learning_rate/action_model': 4.282287573963891e-05, 'epoch': 1.23}) +[2026-06-16 02:15:15,763][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.02419177070260048, 'train/grad_norm_pre_clip': 5.400768756866455, 'timing/data': 0.00656675297068432, 'timing/model': 1.453067360969726, 'learning_rate/qwen_vl_interface': 4.279306222217625e-06, 'learning_rate/action_model': 4.279306222217626e-05, 'epoch': 1.24}) +[2026-06-16 02:15:21,765][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.007268955931067467, 'train/grad_norm_pre_clip': 0.7000276446342468, 'timing/data': 0.006931911048013717, 'timing/model': 1.4626720459782518, 'learning_rate/qwen_vl_interface': 4.276325269486906e-06, 'learning_rate/action_model': 4.276325269486906e-05, 'epoch': 1.24}) +[2026-06-16 02:15:27,887][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.14271479845046997, 'train/grad_norm_pre_clip': 11.560585975646973, 'timing/data': 0.006703424034640193, 'timing/model': 1.5330656040459871, 'learning_rate/qwen_vl_interface': 4.273344716997089e-06, 'learning_rate/action_model': 4.273344716997089e-05, 'epoch': 1.24}) +[2026-06-16 02:15:33,814][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.012607593089342117, 'train/grad_norm_pre_clip': 1.9629329442977905, 'timing/data': 0.0068000469473190606, 'timing/model': 1.5464754470158368, 'learning_rate/qwen_vl_interface': 4.270364565973368e-06, 'learning_rate/action_model': 4.270364565973367e-05, 'epoch': 1.24}) +[2026-06-16 02:15:39,762][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.08660665899515152, 'train/grad_norm_pre_clip': 12.145340919494629, 'timing/data': 0.0069984290166758, 'timing/model': 1.6537167190108448, 'learning_rate/qwen_vl_interface': 4.267384817640768e-06, 'learning_rate/action_model': 4.2673848176407685e-05, 'epoch': 1.24}) +[2026-06-16 02:15:45,642][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.03629685938358307, 'train/grad_norm_pre_clip': 5.210392475128174, 'timing/data': 0.006807367026340216, 'timing/model': 1.5313394530094229, 'learning_rate/qwen_vl_interface': 4.264405473224155e-06, 'learning_rate/action_model': 4.2644054732241546e-05, 'epoch': 1.24}) +[2026-06-16 02:15:51,938][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.007129644509404898, 'train/grad_norm_pre_clip': 1.0349023342132568, 'timing/data': 0.006929466966539621, 'timing/model': 1.5562730939709581, 'learning_rate/qwen_vl_interface': 4.261426533948221e-06, 'learning_rate/action_model': 4.2614265339482213e-05, 'epoch': 1.24}) +[2026-06-16 02:15:58,062][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.21440263092517853, 'train/grad_norm_pre_clip': 7.45123291015625, 'timing/data': 0.006525853008497506, 'timing/model': 1.4583255489706062, 'learning_rate/qwen_vl_interface': 4.258448001037499e-06, 'learning_rate/action_model': 4.2584480010374985e-05, 'epoch': 1.24}) +[2026-06-16 02:16:04,131][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.014429749920964241, 'train/grad_norm_pre_clip': 1.8153672218322754, 'timing/data': 0.006338538019917905, 'timing/model': 1.5749273750116117, 'learning_rate/qwen_vl_interface': 4.255469875716348e-06, 'learning_rate/action_model': 4.255469875716348e-05, 'epoch': 1.24}) +[2026-06-16 02:16:10,156][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.03348511457443237, 'train/grad_norm_pre_clip': 4.508343696594238, 'timing/data': 0.006551111000590026, 'timing/model': 1.5056638300302438, 'learning_rate/qwen_vl_interface': 4.252492159208966e-06, 'learning_rate/action_model': 4.252492159208966e-05, 'epoch': 1.24}) +[2026-06-16 02:16:16,272][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.005321076605468988, 'train/grad_norm_pre_clip': 0.684412956237793, 'timing/data': 0.0067445250460878015, 'timing/model': 1.7973514850018546, 'learning_rate/qwen_vl_interface': 4.2495148527393766e-06, 'learning_rate/action_model': 4.2495148527393766e-05, 'epoch': 1.24}) +[2026-06-16 02:16:22,086][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.25194546580314636, 'train/grad_norm_pre_clip': 9.495550155639648, 'timing/data': 0.007002073980402201, 'timing/model': 1.5130649210186675, 'learning_rate/qwen_vl_interface': 4.246537957531441e-06, 'learning_rate/action_model': 4.2465379575314404e-05, 'epoch': 1.24}) +[2026-06-16 02:16:28,196][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.1052185595035553, 'train/grad_norm_pre_clip': 22.129194259643555, 'timing/data': 0.006940582999959588, 'timing/model': 1.490682472998742, 'learning_rate/qwen_vl_interface': 4.243561474808846e-06, 'learning_rate/action_model': 4.2435614748088456e-05, 'epoch': 1.24}) +[2026-06-16 02:16:34,318][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.18177396059036255, 'train/grad_norm_pre_clip': 11.718546867370605, 'timing/data': 0.006818813970312476, 'timing/model': 1.5445903519866988, 'learning_rate/qwen_vl_interface': 4.240585405795112e-06, 'learning_rate/action_model': 4.240585405795112e-05, 'epoch': 1.24}) +[2026-06-16 02:16:40,462][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.0933220162987709, 'train/grad_norm_pre_clip': 15.212701797485352, 'timing/data': 0.006983006023801863, 'timing/model': 1.4615470690187067, 'learning_rate/qwen_vl_interface': 4.237609751713589e-06, 'learning_rate/action_model': 4.2376097517135896e-05, 'epoch': 1.24}) +[2026-06-16 02:16:46,500][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.28679829835891724, 'train/grad_norm_pre_clip': 15.273730278015137, 'timing/data': 0.0065065440139733255, 'timing/model': 1.788206306984648, 'learning_rate/qwen_vl_interface': 4.234634513787455e-06, 'learning_rate/action_model': 4.2346345137874556e-05, 'epoch': 1.24}) +[2026-06-16 02:16:52,464][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.14810612797737122, 'train/grad_norm_pre_clip': 12.532383918762207, 'timing/data': 0.00675096403574571, 'timing/model': 1.4829241170082241, 'learning_rate/qwen_vl_interface': 4.2316596932397195e-06, 'learning_rate/action_model': 4.2316596932397195e-05, 'epoch': 1.24}) +[2026-06-16 02:16:58,633][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.012660393491387367, 'train/grad_norm_pre_clip': 5.307772159576416, 'timing/data': 0.006829279009252787, 'timing/model': 1.5554048359626904, 'learning_rate/qwen_vl_interface': 4.22868529129322e-06, 'learning_rate/action_model': 4.22868529129322e-05, 'epoch': 1.24}) +[2026-06-16 02:17:04,739][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.13090616464614868, 'train/grad_norm_pre_clip': 14.066364288330078, 'timing/data': 0.006481738004367799, 'timing/model': 1.5316168810240924, 'learning_rate/qwen_vl_interface': 4.2257113091706165e-06, 'learning_rate/action_model': 4.2257113091706165e-05, 'epoch': 1.24}) +[2026-06-16 02:17:10,940][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.13115915656089783, 'train/grad_norm_pre_clip': 13.477984428405762, 'timing/data': 0.006842670030891895, 'timing/model': 1.5319385950569995, 'learning_rate/qwen_vl_interface': 4.222737748094404e-06, 'learning_rate/action_model': 4.222737748094404e-05, 'epoch': 1.24}) +[2026-06-16 02:17:17,140][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.02905949391424656, 'train/grad_norm_pre_clip': 3.567017078399658, 'timing/data': 0.006420217046979815, 'timing/model': 1.5758652450167574, 'learning_rate/qwen_vl_interface': 4.219764609286901e-06, 'learning_rate/action_model': 4.219764609286901e-05, 'epoch': 1.24}) +[2026-06-16 02:17:23,300][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.12313654273748398, 'train/grad_norm_pre_clip': 16.147499084472656, 'timing/data': 0.006275535037275404, 'timing/model': 1.7490475429804064, 'learning_rate/qwen_vl_interface': 4.21679189397025e-06, 'learning_rate/action_model': 4.21679189397025e-05, 'epoch': 1.24}) +[2026-06-16 02:17:29,266][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.015300513245165348, 'train/grad_norm_pre_clip': 2.1086268424987793, 'timing/data': 0.00685929402243346, 'timing/model': 1.4842785390210338, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 1.24}) +[2026-06-16 02:17:35,473][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.10694318264722824, 'train/grad_norm_pre_clip': 2.801567554473877, 'timing/data': 0.00596572901122272, 'timing/model': 1.4797237869934179, 'learning_rate/qwen_vl_interface': 4.210847738697219e-06, 'learning_rate/action_model': 4.210847738697218e-05, 'epoch': 1.25}) +[2026-06-16 02:17:41,661][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.21509577333927155, 'train/grad_norm_pre_clip': 19.259241104125977, 'timing/data': 0.007156501000281423, 'timing/model': 1.448372172017116, 'learning_rate/qwen_vl_interface': 4.207876301184256e-06, 'learning_rate/action_model': 4.207876301184256e-05, 'epoch': 1.25}) +[2026-06-16 02:17:47,562][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.0032919920049607754, 'train/grad_norm_pre_clip': 0.5958111882209778, 'timing/data': 0.006991018017288297, 'timing/model': 1.447863113018684, 'learning_rate/qwen_vl_interface': 4.2049052920489825e-06, 'learning_rate/action_model': 4.2049052920489825e-05, 'epoch': 1.25}) +[2026-06-16 02:17:53,585][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.05590884014964104, 'train/grad_norm_pre_clip': 7.12399959564209, 'timing/data': 0.00688391097355634, 'timing/model': 1.4670009150286205, 'learning_rate/qwen_vl_interface': 4.201934712512666e-06, 'learning_rate/action_model': 4.2019347125126656e-05, 'epoch': 1.25}) +[2026-06-16 02:17:59,875][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.0532718226313591, 'train/grad_norm_pre_clip': 6.7430291175842285, 'timing/data': 0.006987052038311958, 'timing/model': 1.7948397480067797, 'learning_rate/qwen_vl_interface': 4.1989645637964e-06, 'learning_rate/action_model': 4.1989645637964e-05, 'epoch': 1.25}) +[2026-06-16 02:18:05,834][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.04306735843420029, 'train/grad_norm_pre_clip': 4.047186851501465, 'timing/data': 0.007092245970852673, 'timing/model': 1.5447703669779003, 'learning_rate/qwen_vl_interface': 4.195994847121101e-06, 'learning_rate/action_model': 4.195994847121101e-05, 'epoch': 1.25}) +[2026-06-16 02:27:14,850][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.23553600907325745, 'train/grad_norm_pre_clip': 15.082789421081543, 'eval/loss': 0.05848442763090134, 'eval/flap_precision': 0.9200779727095516, 'eval/flap_recall': 0.9458917835671342, 'eval/flap_f1': 0.9328063241106719, 'eval/flap_rate_model': 0.1425, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9328063241106719, 'timing/data': 0.006729462998919189, 'timing/model': 1.5271782989730127, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 1.25}) +[2026-06-16 02:27:14,851][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2900_state +[2026-06-16 02:27:14,855][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:27:20,554][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2900_state/model.safetensors +[2026-06-16 02:27:34,371][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2900_state/optimizer.bin +[2026-06-16 02:27:34,372][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2900_state/sampler.bin +[2026-06-16 02:27:34,372][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2900_state/sampler_1.bin +[2026-06-16 02:27:34,373][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_2900_state/random_states_0.pkl +[2026-06-16 02:27:34,482][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2900 +[2026-06-16 02:27:34,482][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:27:34,492][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:27:40,268][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.2820078134536743, 'train/grad_norm_pre_clip': 19.237136840820312, 'timing/data': 0.007220387982670218, 'timing/model': 1.4870441169478, 'learning_rate/qwen_vl_interface': 4.19005671477618e-06, 'learning_rate/action_model': 4.1900567147761794e-05, 'epoch': 1.25}) +[2026-06-16 02:27:46,176][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.13159644603729248, 'train/grad_norm_pre_clip': 14.829428672790527, 'timing/data': 0.006784118013456464, 'timing/model': 1.4926902099978179, 'learning_rate/qwen_vl_interface': 4.1870883015474975e-06, 'learning_rate/action_model': 4.187088301547497e-05, 'epoch': 1.25}) +[2026-06-16 02:27:52,071][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.011412364430725574, 'train/grad_norm_pre_clip': 1.3120481967926025, 'timing/data': 0.008682113024406135, 'timing/model': 1.4779536160058342, 'learning_rate/qwen_vl_interface': 4.1841203252416655e-06, 'learning_rate/action_model': 4.1841203252416655e-05, 'epoch': 1.25}) +[2026-06-16 02:27:57,967][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.0989702045917511, 'train/grad_norm_pre_clip': 6.151796340942383, 'timing/data': 0.00711469299858436, 'timing/model': 1.4429587639751844, 'learning_rate/qwen_vl_interface': 4.181152787078705e-06, 'learning_rate/action_model': 4.1811527870787056e-05, 'epoch': 1.25}) +[2026-06-16 02:28:03,876][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.20080672204494476, 'train/grad_norm_pre_clip': 10.64090347290039, 'timing/data': 0.006596421997528523, 'timing/model': 1.6811291669728234, 'learning_rate/qwen_vl_interface': 4.178185688278462e-06, 'learning_rate/action_model': 4.1781856882784616e-05, 'epoch': 1.25}) +[2026-06-16 02:28:09,665][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.07207459211349487, 'train/grad_norm_pre_clip': 13.605877876281738, 'timing/data': 0.006630732968915254, 'timing/model': 1.5586124859983101, 'learning_rate/qwen_vl_interface': 4.175219030060595e-06, 'learning_rate/action_model': 4.175219030060595e-05, 'epoch': 1.25}) +[2026-06-16 02:28:15,353][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.015329094603657722, 'train/grad_norm_pre_clip': 1.3096060752868652, 'timing/data': 0.006808143982198089, 'timing/model': 1.458389651030302, 'learning_rate/qwen_vl_interface': 4.1722528136445875e-06, 'learning_rate/action_model': 4.172252813644588e-05, 'epoch': 1.25}) +[2026-06-16 02:28:21,254][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.13004545867443085, 'train/grad_norm_pre_clip': 8.879716873168945, 'timing/data': 0.007006354979239404, 'timing/model': 1.4590087109827437, 'learning_rate/qwen_vl_interface': 4.169287040249738e-06, 'learning_rate/action_model': 4.169287040249738e-05, 'epoch': 1.25}) +[2026-06-16 02:28:27,154][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.02328657917678356, 'train/grad_norm_pre_clip': 2.9370481967926025, 'timing/data': 0.0068552030134014785, 'timing/model': 1.4582819049828686, 'learning_rate/qwen_vl_interface': 4.166321711095164e-06, 'learning_rate/action_model': 4.166321711095164e-05, 'epoch': 1.25}) +[2026-06-16 02:28:33,129][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.011412378400564194, 'train/grad_norm_pre_clip': 4.718047618865967, 'timing/data': 0.0062931079883128405, 'timing/model': 1.5012405199813657, 'learning_rate/qwen_vl_interface': 4.1633568273998e-06, 'learning_rate/action_model': 4.1633568273998e-05, 'epoch': 1.25}) +[2026-06-16 02:28:39,187][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.2687556743621826, 'train/grad_norm_pre_clip': 10.331733703613281, 'timing/data': 0.006710530957207084, 'timing/model': 1.6896912599913776, 'learning_rate/qwen_vl_interface': 4.1603923903823995e-06, 'learning_rate/action_model': 4.160392390382399e-05, 'epoch': 1.25}) +[2026-06-16 02:28:44,860][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.013361134566366673, 'train/grad_norm_pre_clip': 2.6428475379943848, 'timing/data': 0.006626538001000881, 'timing/model': 1.4592168970266357, 'learning_rate/qwen_vl_interface': 4.1574284012615276e-06, 'learning_rate/action_model': 4.157428401261528e-05, 'epoch': 1.25}) +[2026-06-16 02:28:50,760][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.10871534049510956, 'train/grad_norm_pre_clip': 9.696736335754395, 'timing/data': 0.006502155971247703, 'timing/model': 1.4585736720473506, 'learning_rate/qwen_vl_interface': 4.154464861255571e-06, 'learning_rate/action_model': 4.154464861255571e-05, 'epoch': 1.25}) +[2026-06-16 02:28:56,760][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.12143579870462418, 'train/grad_norm_pre_clip': 19.576858520507812, 'timing/data': 0.006565788993611932, 'timing/model': 1.4573835129849613, 'learning_rate/qwen_vl_interface': 4.151501771582728e-06, 'learning_rate/action_model': 4.151501771582728e-05, 'epoch': 1.25}) +[2026-06-16 02:29:02,662][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.15555556118488312, 'train/grad_norm_pre_clip': 6.66038179397583, 'timing/data': 0.006604537018574774, 'timing/model': 1.4527792309527285, 'learning_rate/qwen_vl_interface': 4.148539133461012e-06, 'learning_rate/action_model': 4.1485391334610116e-05, 'epoch': 1.25}) +[2026-06-16 02:29:08,646][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.03859328106045723, 'train/grad_norm_pre_clip': 6.659477710723877, 'timing/data': 0.006809207960031927, 'timing/model': 1.512160484038759, 'learning_rate/qwen_vl_interface': 4.145576948108254e-06, 'learning_rate/action_model': 4.145576948108253e-05, 'epoch': 1.25}) +[2026-06-16 02:29:14,593][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.19195380806922913, 'train/grad_norm_pre_clip': 17.85529327392578, 'timing/data': 0.007262283004820347, 'timing/model': 1.6884684389806353, 'learning_rate/qwen_vl_interface': 4.142615216742093e-06, 'learning_rate/action_model': 4.142615216742093e-05, 'epoch': 1.26}) +[2026-06-16 02:29:20,271][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.3774932622909546, 'train/grad_norm_pre_clip': 13.046894073486328, 'timing/data': 0.006636404024902731, 'timing/model': 1.4672133029671386, 'learning_rate/qwen_vl_interface': 4.139653940579988e-06, 'learning_rate/action_model': 4.1396539405799874e-05, 'epoch': 1.26}) +[2026-06-16 02:29:26,426][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.04549732804298401, 'train/grad_norm_pre_clip': 8.38574504852295, 'timing/data': 0.0065856100409291685, 'timing/model': 1.524933342996519, 'learning_rate/qwen_vl_interface': 4.136693120839206e-06, 'learning_rate/action_model': 4.136693120839205e-05, 'epoch': 1.26}) +[2026-06-16 02:29:32,452][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.010811894200742245, 'train/grad_norm_pre_clip': 0.8321487903594971, 'timing/data': 0.006878710992168635, 'timing/model': 1.5134038620162755, 'learning_rate/qwen_vl_interface': 4.133732758736828e-06, 'learning_rate/action_model': 4.133732758736828e-05, 'epoch': 1.26}) +[2026-06-16 02:29:38,360][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.036881983280181885, 'train/grad_norm_pre_clip': 8.30296516418457, 'timing/data': 0.0066845149849541485, 'timing/model': 1.4425073640304618, 'learning_rate/qwen_vl_interface': 4.13077285548975e-06, 'learning_rate/action_model': 4.13077285548975e-05, 'epoch': 1.26}) +[2026-06-16 02:29:44,365][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.03820715472102165, 'train/grad_norm_pre_clip': 4.627062797546387, 'timing/data': 0.006839824025519192, 'timing/model': 1.5353697889950126, 'learning_rate/qwen_vl_interface': 4.127813412314672e-06, 'learning_rate/action_model': 4.1278134123146725e-05, 'epoch': 1.26}) +[2026-06-16 02:29:50,286][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.1871533989906311, 'train/grad_norm_pre_clip': 12.251622200012207, 'timing/data': 0.006575095001608133, 'timing/model': 1.6849661680171266, 'learning_rate/qwen_vl_interface': 4.124854430428113e-06, 'learning_rate/action_model': 4.1248544304281126e-05, 'epoch': 1.26}) +[2026-06-16 02:29:56,209][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.2048262655735016, 'train/grad_norm_pre_clip': 11.903512001037598, 'timing/data': 0.0066311879782006145, 'timing/model': 1.6523988239932805, 'learning_rate/qwen_vl_interface': 4.121895911046397e-06, 'learning_rate/action_model': 4.121895911046396e-05, 'epoch': 1.26}) +[2026-06-16 02:30:01,864][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.019506379961967468, 'train/grad_norm_pre_clip': 2.1575064659118652, 'timing/data': 0.006621412991080433, 'timing/model': 1.4582910819444805, 'learning_rate/qwen_vl_interface': 4.118937855385659e-06, 'learning_rate/action_model': 4.1189378553856584e-05, 'epoch': 1.26}) +[2026-06-16 02:30:07,759][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.028732554987072945, 'train/grad_norm_pre_clip': 7.5774407386779785, 'timing/data': 0.006762501026969403, 'timing/model': 1.4625532630016096, 'learning_rate/qwen_vl_interface': 4.115980264661844e-06, 'learning_rate/action_model': 4.115980264661844e-05, 'epoch': 1.26}) +[2026-06-16 02:30:13,660][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.17404596507549286, 'train/grad_norm_pre_clip': 9.095565795898438, 'timing/data': 0.006519665010273457, 'timing/model': 1.45392019697465, 'learning_rate/qwen_vl_interface': 4.113023140090709e-06, 'learning_rate/action_model': 4.113023140090709e-05, 'epoch': 1.26}) +[2026-06-16 02:30:19,636][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.025385601446032524, 'train/grad_norm_pre_clip': 3.880842447280884, 'timing/data': 0.006574294995516539, 'timing/model': 1.5069173800293356, 'learning_rate/qwen_vl_interface': 4.110066482887812e-06, 'learning_rate/action_model': 4.110066482887812e-05, 'epoch': 1.26}) +[2026-06-16 02:30:25,611][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.0779469683766365, 'train/grad_norm_pre_clip': 10.243436813354492, 'timing/data': 0.006371458002831787, 'timing/model': 1.7099010680103675, 'learning_rate/qwen_vl_interface': 4.107110294268525e-06, 'learning_rate/action_model': 4.107110294268525e-05, 'epoch': 1.26}) +[2026-06-16 02:30:31,666][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.012885899282991886, 'train/grad_norm_pre_clip': 1.885268211364746, 'timing/data': 0.006489620951469988, 'timing/model': 1.7615160589921288, 'learning_rate/qwen_vl_interface': 4.104154575448026e-06, 'learning_rate/action_model': 4.104154575448025e-05, 'epoch': 1.26}) +[2026-06-16 02:30:37,375][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.12900693714618683, 'train/grad_norm_pre_clip': 12.127345085144043, 'timing/data': 0.007230849994812161, 'timing/model': 1.4702875560033135, 'learning_rate/qwen_vl_interface': 4.101199327641297e-06, 'learning_rate/action_model': 4.101199327641297e-05, 'epoch': 1.26}) +[2026-06-16 02:30:43,264][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.029887771233916283, 'train/grad_norm_pre_clip': 7.422342300415039, 'timing/data': 0.007536439050454646, 'timing/model': 1.4581315830000676, 'learning_rate/qwen_vl_interface': 4.098244552063133e-06, 'learning_rate/action_model': 4.098244552063132e-05, 'epoch': 1.26}) +[2026-06-16 02:30:49,263][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.15918609499931335, 'train/grad_norm_pre_clip': 15.337615013122559, 'timing/data': 0.007136040949262679, 'timing/model': 1.45881137996912, 'learning_rate/qwen_vl_interface': 4.0952902499281254e-06, 'learning_rate/action_model': 4.095290249928125e-05, 'epoch': 1.26}) +[2026-06-16 02:30:55,450][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.038222234696149826, 'train/grad_norm_pre_clip': 6.874237537384033, 'timing/data': 0.006871954014059156, 'timing/model': 1.5128586819628254, 'learning_rate/qwen_vl_interface': 4.092336422450681e-06, 'learning_rate/action_model': 4.0923364224506804e-05, 'epoch': 1.26}) +[2026-06-16 02:31:01,495][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.13163717091083527, 'train/grad_norm_pre_clip': 13.08872127532959, 'timing/data': 0.006714190996717662, 'timing/model': 1.692494787974283, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 1.26}) +[2026-06-16 02:31:07,361][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.1529393196105957, 'train/grad_norm_pre_clip': 8.052508354187012, 'timing/data': 0.00665420398581773, 'timing/model': 1.6543526379973628, 'learning_rate/qwen_vl_interface': 4.086430196325106e-06, 'learning_rate/action_model': 4.086430196325106e-05, 'epoch': 1.26}) +[2026-06-16 02:31:13,068][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.07428950816392899, 'train/grad_norm_pre_clip': 8.480090141296387, 'timing/data': 0.0067316279746592045, 'timing/model': 1.4610738649498671, 'learning_rate/qwen_vl_interface': 4.083477800104802e-06, 'learning_rate/action_model': 4.083477800104802e-05, 'epoch': 1.26}) +[2026-06-16 02:31:19,061][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.11388526856899261, 'train/grad_norm_pre_clip': 16.14935874938965, 'timing/data': 0.006492677028290927, 'timing/model': 1.4578180089592934, 'learning_rate/qwen_vl_interface': 4.080525883397711e-06, 'learning_rate/action_model': 4.080525883397711e-05, 'epoch': 1.26}) +[2026-06-16 02:31:25,059][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.018612034618854523, 'train/grad_norm_pre_clip': 4.661375999450684, 'timing/data': 0.006633453944232315, 'timing/model': 1.4610852350015193, 'learning_rate/qwen_vl_interface': 4.077574447417256e-06, 'learning_rate/action_model': 4.077574447417256e-05, 'epoch': 1.26}) +[2026-06-16 02:31:31,040][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.25666293501853943, 'train/grad_norm_pre_clip': 15.930784225463867, 'timing/data': 0.0068078499753028154, 'timing/model': 1.5109318710165098, 'learning_rate/qwen_vl_interface': 4.074623493376657e-06, 'learning_rate/action_model': 4.074623493376657e-05, 'epoch': 1.27}) +[2026-06-16 02:31:37,055][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.15458890795707703, 'train/grad_norm_pre_clip': 7.681396007537842, 'timing/data': 0.006473317975178361, 'timing/model': 1.5155269780079834, 'learning_rate/qwen_vl_interface': 4.071673022488944e-06, 'learning_rate/action_model': 4.071673022488944e-05, 'epoch': 1.27}) +[2026-06-16 02:31:42,945][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.08915944397449493, 'train/grad_norm_pre_clip': 6.61943244934082, 'timing/data': 0.006475361995398998, 'timing/model': 1.6301594440010376, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 1.27}) +[2026-06-16 02:31:48,661][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.2110111266374588, 'train/grad_norm_pre_clip': 14.116762161254883, 'timing/data': 0.00677779596298933, 'timing/model': 1.4586623859941028, 'learning_rate/qwen_vl_interface': 4.065773535023277e-06, 'learning_rate/action_model': 4.0657735350232766e-05, 'epoch': 1.27}) +[2026-06-16 02:31:54,663][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.14650391042232513, 'train/grad_norm_pre_clip': 14.52690601348877, 'timing/data': 0.006596649996936321, 'timing/model': 1.4596806970075704, 'learning_rate/qwen_vl_interface': 4.062824520870382e-06, 'learning_rate/action_model': 4.062824520870381e-05, 'epoch': 1.27}) +[2026-06-16 02:32:00,661][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.05555504560470581, 'train/grad_norm_pre_clip': 7.833553791046143, 'timing/data': 0.006580587010830641, 'timing/model': 1.47219037899049, 'learning_rate/qwen_vl_interface': 4.059875994720481e-06, 'learning_rate/action_model': 4.059875994720481e-05, 'epoch': 1.27}) +[2026-06-16 02:32:06,646][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.06684750318527222, 'train/grad_norm_pre_clip': 6.765859603881836, 'timing/data': 0.006354935991112143, 'timing/model': 1.509985524986405, 'learning_rate/qwen_vl_interface': 4.056927957785605e-06, 'learning_rate/action_model': 4.056927957785605e-05, 'epoch': 1.27}) +[2026-06-16 02:32:12,594][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.11281119287014008, 'train/grad_norm_pre_clip': 9.76571273803711, 'timing/data': 0.006581268971785903, 'timing/model': 1.688455485040322, 'learning_rate/qwen_vl_interface': 4.0539804112775785e-06, 'learning_rate/action_model': 4.053980411277578e-05, 'epoch': 1.27}) +[2026-06-16 02:32:18,264][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.05136600509285927, 'train/grad_norm_pre_clip': 9.139517784118652, 'timing/data': 0.006813479994889349, 'timing/model': 1.4597948970040306, 'learning_rate/qwen_vl_interface': 4.051033356408027e-06, 'learning_rate/action_model': 4.0510333564080264e-05, 'epoch': 1.27}) +[2026-06-16 02:32:24,263][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.038113974034786224, 'train/grad_norm_pre_clip': 4.278504848480225, 'timing/data': 0.006487641017884016, 'timing/model': 1.4613481519627385, 'learning_rate/qwen_vl_interface': 4.048086794388375e-06, 'learning_rate/action_model': 4.048086794388375e-05, 'epoch': 1.27}) +[2026-06-16 02:32:30,079][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.05682536959648132, 'train/grad_norm_pre_clip': 8.113051414489746, 'timing/data': 0.007202561013400555, 'timing/model': 1.4744829399860464, 'learning_rate/qwen_vl_interface': 4.045140726429841e-06, 'learning_rate/action_model': 4.0451407264298406e-05, 'epoch': 1.27}) +[2026-06-16 02:32:35,962][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.09192200750112534, 'train/grad_norm_pre_clip': 5.2781548500061035, 'timing/data': 0.006569802004378289, 'timing/model': 1.4483264830196276, 'learning_rate/qwen_vl_interface': 4.042195153743443e-06, 'learning_rate/action_model': 4.042195153743443e-05, 'epoch': 1.27}) +[2026-06-16 02:32:41,858][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.08055977523326874, 'train/grad_norm_pre_clip': 9.562955856323242, 'timing/data': 0.007401273993309587, 'timing/model': 1.4781058160006069, 'learning_rate/qwen_vl_interface': 4.039250077539993e-06, 'learning_rate/action_model': 4.039250077539993e-05, 'epoch': 1.27}) +[2026-06-16 02:32:47,793][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.2517414689064026, 'train/grad_norm_pre_clip': 13.71639633178711, 'timing/data': 0.006545365962665528, 'timing/model': 1.6917912029894069, 'learning_rate/qwen_vl_interface': 4.036305499030104e-06, 'learning_rate/action_model': 4.0363054990301036e-05, 'epoch': 1.27}) +[2026-06-16 02:32:53,460][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.1997932344675064, 'train/grad_norm_pre_clip': 7.569875717163086, 'timing/data': 0.00631726102437824, 'timing/model': 1.4604314590105787, 'learning_rate/qwen_vl_interface': 4.033361419424178e-06, 'learning_rate/action_model': 4.033361419424178e-05, 'epoch': 1.27}) +[2026-06-16 02:32:59,462][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.002225700067356229, 'train/grad_norm_pre_clip': 0.44277408719062805, 'timing/data': 0.006720303965266794, 'timing/model': 1.4596786930342205, 'learning_rate/qwen_vl_interface': 4.030417839932418e-06, 'learning_rate/action_model': 4.030417839932417e-05, 'epoch': 1.27}) +[2026-06-16 02:33:05,460][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.1187942698597908, 'train/grad_norm_pre_clip': 16.62866973876953, 'timing/data': 0.0066033920156769454, 'timing/model': 1.4718000729917549, 'learning_rate/qwen_vl_interface': 4.027474761764815e-06, 'learning_rate/action_model': 4.027474761764814e-05, 'epoch': 1.27}) +[2026-06-16 02:33:11,371][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.0763569250702858, 'train/grad_norm_pre_clip': 7.065917015075684, 'timing/data': 0.00609933101804927, 'timing/model': 1.4729077100055292, 'learning_rate/qwen_vl_interface': 4.024532186131158e-06, 'learning_rate/action_model': 4.0245321861311575e-05, 'epoch': 1.27}) +[2026-06-16 02:33:17,266][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.10496656596660614, 'train/grad_norm_pre_clip': 6.431698799133301, 'timing/data': 0.006391872011590749, 'timing/model': 1.4775018690270372, 'learning_rate/qwen_vl_interface': 4.021590114241029e-06, 'learning_rate/action_model': 4.021590114241029e-05, 'epoch': 1.27}) +[2026-06-16 02:33:23,230][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.1552075892686844, 'train/grad_norm_pre_clip': 8.794133186340332, 'timing/data': 0.0069069089950062335, 'timing/model': 1.7277264389558695, 'learning_rate/qwen_vl_interface': 4.018648547303804e-06, 'learning_rate/action_model': 4.0186485473038036e-05, 'epoch': 1.27}) +[2026-06-16 02:33:28,960][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.044316455721855164, 'train/grad_norm_pre_clip': 5.756691932678223, 'timing/data': 0.006820803042501211, 'timing/model': 1.4585777799948119, 'learning_rate/qwen_vl_interface': 4.015707486528649e-06, 'learning_rate/action_model': 4.015707486528649e-05, 'epoch': 1.27}) +[2026-06-16 02:33:34,959][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.013446612283587456, 'train/grad_norm_pre_clip': 5.515598297119141, 'timing/data': 0.007099412032403052, 'timing/model': 1.4588508299784735, 'learning_rate/qwen_vl_interface': 4.012766933124523e-06, 'learning_rate/action_model': 4.012766933124523e-05, 'epoch': 1.27}) +[2026-06-16 02:33:40,860][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.048491109162569046, 'train/grad_norm_pre_clip': 3.8359851837158203, 'timing/data': 0.007754964986816049, 'timing/model': 1.4576965419691987, 'learning_rate/qwen_vl_interface': 4.009826888300176e-06, 'learning_rate/action_model': 4.0098268883001755e-05, 'epoch': 1.27}) +[2026-06-16 02:33:46,757][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.03232444077730179, 'train/grad_norm_pre_clip': 2.9959142208099365, 'timing/data': 0.0058484639739617705, 'timing/model': 1.4270874420180917, 'learning_rate/qwen_vl_interface': 4.00688735326415e-06, 'learning_rate/action_model': 4.0068873532641495e-05, 'epoch': 1.27}) +[2026-06-16 02:33:52,658][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.0161124337464571, 'train/grad_norm_pre_clip': 1.4832144975662231, 'timing/data': 0.00663351797265932, 'timing/model': 1.4581705129821785, 'learning_rate/qwen_vl_interface': 4.0039483292247765e-06, 'learning_rate/action_model': 4.003948329224776e-05, 'epoch': 1.28}) +[2026-06-16 02:33:58,590][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.07074485719203949, 'train/grad_norm_pre_clip': 13.570778846740723, 'timing/data': 0.006572799000423402, 'timing/model': 1.6917683170177042, 'learning_rate/qwen_vl_interface': 4.001009817390178e-06, 'learning_rate/action_model': 4.001009817390178e-05, 'epoch': 1.28}) +[2026-06-16 02:34:04,259][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.12440277636051178, 'train/grad_norm_pre_clip': 9.502346992492676, 'timing/data': 0.006502434960566461, 'timing/model': 1.4571315689827316, 'learning_rate/qwen_vl_interface': 3.998071818968264e-06, 'learning_rate/action_model': 3.998071818968264e-05, 'epoch': 1.28}) +[2026-06-16 02:34:10,262][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.08225897699594498, 'train/grad_norm_pre_clip': 7.323578834533691, 'timing/data': 0.006546621036250144, 'timing/model': 1.4585425780387595, 'learning_rate/qwen_vl_interface': 3.995134335166736e-06, 'learning_rate/action_model': 3.995134335166736e-05, 'epoch': 1.28}) +[2026-06-16 02:34:16,357][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.06744309514760971, 'train/grad_norm_pre_clip': 9.469444274902344, 'timing/data': 0.007296893047168851, 'timing/model': 1.4784723509801552, 'learning_rate/qwen_vl_interface': 3.992197367193083e-06, 'learning_rate/action_model': 3.992197367193082e-05, 'epoch': 1.28}) +[2026-06-16 02:34:22,259][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.10313752293586731, 'train/grad_norm_pre_clip': 7.911699295043945, 'timing/data': 0.00681198004167527, 'timing/model': 1.4591297730221413, 'learning_rate/qwen_vl_interface': 3.98926091625458e-06, 'learning_rate/action_model': 3.9892609162545795e-05, 'epoch': 1.28}) +[2026-06-16 02:34:28,245][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.1468077152967453, 'train/grad_norm_pre_clip': 13.805635452270508, 'timing/data': 0.006634449993725866, 'timing/model': 1.5099610569886863, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 1.28}) +[2026-06-16 02:34:34,189][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.14592312276363373, 'train/grad_norm_pre_clip': 13.193747520446777, 'timing/data': 0.007004844024777412, 'timing/model': 1.6869393699453212, 'learning_rate/qwen_vl_interface': 3.983389570311073e-06, 'learning_rate/action_model': 3.9833895703110724e-05, 'epoch': 1.28}) +[2026-06-16 02:34:39,859][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.04701090231537819, 'train/grad_norm_pre_clip': 8.468321800231934, 'timing/data': 0.006410010042600334, 'timing/model': 1.4585582370054908, 'learning_rate/qwen_vl_interface': 3.9804546777195546e-06, 'learning_rate/action_model': 3.980454677719554e-05, 'epoch': 1.28}) +[2026-06-16 02:34:45,765][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.1008346900343895, 'train/grad_norm_pre_clip': 13.122174263000488, 'timing/data': 0.0071950440178625286, 'timing/model': 1.4620030150399543, 'learning_rate/qwen_vl_interface': 3.977520306990164e-06, 'learning_rate/action_model': 3.977520306990163e-05, 'epoch': 1.28}) +[2026-06-16 02:34:51,659][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.0851995050907135, 'train/grad_norm_pre_clip': 6.889392375946045, 'timing/data': 0.0073231629794463515, 'timing/model': 1.4563583129784092, 'learning_rate/qwen_vl_interface': 3.974586459329109e-06, 'learning_rate/action_model': 3.9745864593291085e-05, 'epoch': 1.28}) +[2026-06-16 02:34:57,564][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.0033122708555310965, 'train/grad_norm_pre_clip': 0.8071526885032654, 'timing/data': 0.006567182019352913, 'timing/model': 1.4657591829891317, 'learning_rate/qwen_vl_interface': 3.9716531359423855e-06, 'learning_rate/action_model': 3.971653135942385e-05, 'epoch': 1.28}) +[2026-06-16 02:35:03,539][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.14740407466888428, 'train/grad_norm_pre_clip': 15.004324913024902, 'timing/data': 0.006344063964206725, 'timing/model': 1.507349899969995, 'learning_rate/qwen_vl_interface': 3.968720338035769e-06, 'learning_rate/action_model': 3.968720338035769e-05, 'epoch': 1.28}) +[2026-06-16 02:35:09,487][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.01711767166852951, 'train/grad_norm_pre_clip': 1.895813226699829, 'timing/data': 0.0060388449928723276, 'timing/model': 1.6883929219911806, 'learning_rate/qwen_vl_interface': 3.965788066814824e-06, 'learning_rate/action_model': 3.965788066814823e-05, 'epoch': 1.28}) +[2026-06-16 02:35:15,160][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.02880602516233921, 'train/grad_norm_pre_clip': 5.968738079071045, 'timing/data': 0.006299978995230049, 'timing/model': 1.460037882963661, 'learning_rate/qwen_vl_interface': 3.962856323484895e-06, 'learning_rate/action_model': 3.9628563234848945e-05, 'epoch': 1.28}) +[2026-06-16 02:35:21,060][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.13680115342140198, 'train/grad_norm_pre_clip': 11.19085693359375, 'timing/data': 0.006418053002562374, 'timing/model': 1.4613296150346287, 'learning_rate/qwen_vl_interface': 3.959925109251112e-06, 'learning_rate/action_model': 3.959925109251112e-05, 'epoch': 1.28}) +[2026-06-16 02:35:27,058][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.25109001994132996, 'train/grad_norm_pre_clip': 16.812744140625, 'timing/data': 0.006910086027346551, 'timing/model': 1.458462182956282, 'learning_rate/qwen_vl_interface': 3.956994425318387e-06, 'learning_rate/action_model': 3.956994425318387e-05, 'epoch': 1.28}) +[2026-06-16 02:35:32,960][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.07840969413518906, 'train/grad_norm_pre_clip': 8.413073539733887, 'timing/data': 0.007000387995503843, 'timing/model': 1.4594108100282028, 'learning_rate/qwen_vl_interface': 3.9540642728914115e-06, 'learning_rate/action_model': 3.9540642728914115e-05, 'epoch': 1.28}) +[2026-06-16 02:35:38,942][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.1919860541820526, 'train/grad_norm_pre_clip': 22.700796127319336, 'timing/data': 0.007036170980427414, 'timing/model': 1.5081749809905887, 'learning_rate/qwen_vl_interface': 3.951134653174664e-06, 'learning_rate/action_model': 3.951134653174663e-05, 'epoch': 1.28}) +[2026-06-16 02:35:44,891][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.22689288854599, 'train/grad_norm_pre_clip': 9.279552459716797, 'timing/data': 0.006227223027963191, 'timing/model': 1.6923630710225552, 'learning_rate/qwen_vl_interface': 3.948205567372397e-06, 'learning_rate/action_model': 3.948205567372397e-05, 'epoch': 1.28}) +[2026-06-16 02:35:50,571][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.09192883223295212, 'train/grad_norm_pre_clip': 12.99006462097168, 'timing/data': 0.0064231110154651105, 'timing/model': 1.4667615389917046, 'learning_rate/qwen_vl_interface': 3.945277016688647e-06, 'learning_rate/action_model': 3.945277016688647e-05, 'epoch': 1.28}) +[2026-06-16 02:35:56,462][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.006358516868203878, 'train/grad_norm_pre_clip': 0.9946568012237549, 'timing/data': 0.006918077997397631, 'timing/model': 1.4584439149475656, 'learning_rate/qwen_vl_interface': 3.9423490023272326e-06, 'learning_rate/action_model': 3.9423490023272324e-05, 'epoch': 1.28}) +[2026-06-16 02:36:02,362][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.16821236908435822, 'train/grad_norm_pre_clip': 13.867431640625, 'timing/data': 0.0075304280035197735, 'timing/model': 1.4579989219782874, 'learning_rate/qwen_vl_interface': 3.939421525491748e-06, 'learning_rate/action_model': 3.9394215254917483e-05, 'epoch': 1.28}) +[2026-06-16 02:36:08,360][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.13633275032043457, 'train/grad_norm_pre_clip': 9.545594215393066, 'timing/data': 0.006627390976063907, 'timing/model': 1.471117744978983, 'learning_rate/qwen_vl_interface': 3.9364945873855705e-06, 'learning_rate/action_model': 3.93649458738557e-05, 'epoch': 1.29}) +[2026-06-16 02:36:14,360][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.03316797316074371, 'train/grad_norm_pre_clip': 11.408292770385742, 'timing/data': 0.006920220970641822, 'timing/model': 1.5024391120532528, 'learning_rate/qwen_vl_interface': 3.9335681892118536e-06, 'learning_rate/action_model': 3.933568189211854e-05, 'epoch': 1.29}) +[2026-06-16 02:36:20,287][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.12122013419866562, 'train/grad_norm_pre_clip': 13.322641372680664, 'timing/data': 0.006585640017874539, 'timing/model': 1.6891058289911598, 'learning_rate/qwen_vl_interface': 3.930642332173527e-06, 'learning_rate/action_model': 3.930642332173527e-05, 'epoch': 1.29}) +[2026-06-16 02:36:26,125][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.19966015219688416, 'train/grad_norm_pre_clip': 11.200325965881348, 'timing/data': 0.0069939299719408154, 'timing/model': 1.5992749860160984, 'learning_rate/qwen_vl_interface': 3.927717017473301e-06, 'learning_rate/action_model': 3.9277170174733006e-05, 'epoch': 1.29}) +[2026-06-16 02:36:31,863][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.30383676290512085, 'train/grad_norm_pre_clip': 15.359068870544434, 'timing/data': 0.006662402010988444, 'timing/model': 1.4618292429950088, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 1.29}) +[2026-06-16 02:36:37,760][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.12560594081878662, 'train/grad_norm_pre_clip': 15.500994682312012, 'timing/data': 0.006678896956145763, 'timing/model': 1.4580996509757824, 'learning_rate/qwen_vl_interface': 3.921868019896873e-06, 'learning_rate/action_model': 3.921868019896873e-05, 'epoch': 1.29}) +[2026-06-16 02:36:43,860][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.2336958944797516, 'train/grad_norm_pre_clip': 12.506952285766602, 'timing/data': 0.006271697988267988, 'timing/model': 1.4934821480419487, 'learning_rate/qwen_vl_interface': 3.918944339424974e-06, 'learning_rate/action_model': 3.9189443394249736e-05, 'epoch': 1.29}) +[2026-06-16 02:36:49,864][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.05648694559931755, 'train/grad_norm_pre_clip': 7.476441860198975, 'timing/data': 0.006634782999753952, 'timing/model': 1.5332717640558258, 'learning_rate/qwen_vl_interface': 3.9160212060997774e-06, 'learning_rate/action_model': 3.9160212060997774e-05, 'epoch': 1.29}) +[2026-06-16 02:36:55,789][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.2616174817085266, 'train/grad_norm_pre_clip': 20.651817321777344, 'timing/data': 0.006238100002519786, 'timing/model': 1.6857339479611255, 'learning_rate/qwen_vl_interface': 3.913098621122875e-06, 'learning_rate/action_model': 3.913098621122875e-05, 'epoch': 1.29}) +[2026-06-16 02:37:01,790][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.16473908722400665, 'train/grad_norm_pre_clip': 9.524604797363281, 'timing/data': 0.005930948012974113, 'timing/model': 1.7393147189868614, 'learning_rate/qwen_vl_interface': 3.91017658569563e-06, 'learning_rate/action_model': 3.91017658569563e-05, 'epoch': 1.29}) +[2026-06-16 02:37:07,562][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.1315797120332718, 'train/grad_norm_pre_clip': 12.008530616760254, 'timing/data': 0.00640537595609203, 'timing/model': 1.4619885730207898, 'learning_rate/qwen_vl_interface': 3.907255101019181e-06, 'learning_rate/action_model': 3.907255101019181e-05, 'epoch': 1.29}) +[2026-06-16 02:37:13,542][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.05971594899892807, 'train/grad_norm_pre_clip': 4.348094463348389, 'timing/data': 0.006518401030916721, 'timing/model': 1.508974878990557, 'learning_rate/qwen_vl_interface': 3.904334168294442e-06, 'learning_rate/action_model': 3.9043341682944415e-05, 'epoch': 1.29}) +[2026-06-16 02:37:19,570][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.11120627075433731, 'train/grad_norm_pre_clip': 7.912327289581299, 'timing/data': 0.006226502009667456, 'timing/model': 1.4760055239894427, 'learning_rate/qwen_vl_interface': 3.901413788722093e-06, 'learning_rate/action_model': 3.901413788722093e-05, 'epoch': 1.29}) +[2026-06-16 02:46:28,849][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.13950751721858978, 'train/grad_norm_pre_clip': 7.880918025970459, 'eval/loss': 0.050546206533908844, 'eval/flap_precision': 0.9474789915966386, 'eval/flap_recall': 0.9038076152304609, 'eval/flap_f1': 0.925128205128205, 'eval/flap_rate_model': 0.1322222222222222, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.925128205128205, 'timing/data': 0.006428109016269445, 'timing/model': 1.510134881013073, 'learning_rate/qwen_vl_interface': 3.8984939635025966e-06, 'learning_rate/action_model': 3.898493963502596e-05, 'epoch': 1.29}) +[2026-06-16 02:46:28,850][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3000_state +[2026-06-16 02:46:28,854][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:46:34,540][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3000_state/model.safetensors +[2026-06-16 02:46:48,122][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3000_state/optimizer.bin +[2026-06-16 02:46:48,122][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3000_state/sampler.bin +[2026-06-16 02:46:48,123][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-16 02:46:48,124][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-16 02:46:48,240][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-16 02:46:48,240][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:46:48,250][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:46:54,349][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.1002589762210846, 'train/grad_norm_pre_clip': 5.560507297515869, 'timing/data': 0.006872157042380422, 'timing/model': 1.48566824500449, 'learning_rate/qwen_vl_interface': 3.895574693836181e-06, 'learning_rate/action_model': 3.8955746938361804e-05, 'epoch': 1.29}) +[2026-06-16 02:47:00,556][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.07915593683719635, 'train/grad_norm_pre_clip': 3.328035831451416, 'timing/data': 0.006920857995282859, 'timing/model': 1.518656326981727, 'learning_rate/qwen_vl_interface': 3.892655980922847e-06, 'learning_rate/action_model': 3.8926559809228466e-05, 'epoch': 1.29}) +[2026-06-16 02:47:06,485][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.12069230526685715, 'train/grad_norm_pre_clip': 7.175553321838379, 'timing/data': 0.00672275701072067, 'timing/model': 1.692125149013009, 'learning_rate/qwen_vl_interface': 3.889737825962368e-06, 'learning_rate/action_model': 3.889737825962368e-05, 'epoch': 1.29}) +[2026-06-16 02:47:12,379][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.10251182317733765, 'train/grad_norm_pre_clip': 12.133139610290527, 'timing/data': 0.006689509027637541, 'timing/model': 1.6847389430040494, 'learning_rate/qwen_vl_interface': 3.886820230154288e-06, 'learning_rate/action_model': 3.886820230154287e-05, 'epoch': 1.29}) +[2026-06-16 02:47:18,058][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.026870597153902054, 'train/grad_norm_pre_clip': 2.6852173805236816, 'timing/data': 0.00718959397636354, 'timing/model': 1.4509980599978007, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 1.29}) +[2026-06-16 02:47:24,056][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.038972292095422745, 'train/grad_norm_pre_clip': 3.7375729084014893, 'timing/data': 0.006908990966621786, 'timing/model': 1.4583764679846354, 'learning_rate/qwen_vl_interface': 3.880986720792344e-06, 'learning_rate/action_model': 3.880986720792344e-05, 'epoch': 1.29}) +[2026-06-16 02:47:30,072][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.09628098458051682, 'train/grad_norm_pre_clip': 6.076244831085205, 'timing/data': 0.006716738978866488, 'timing/model': 1.491890330973547, 'learning_rate/qwen_vl_interface': 3.878070809636416e-06, 'learning_rate/action_model': 3.878070809636416e-05, 'epoch': 1.29}) +[2026-06-16 02:47:36,046][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.24516595900058746, 'train/grad_norm_pre_clip': 5.99215841293335, 'timing/data': 0.006531646009534597, 'timing/model': 1.5091139170108363, 'learning_rate/qwen_vl_interface': 3.875155462428758e-06, 'learning_rate/action_model': 3.875155462428758e-05, 'epoch': 1.29}) +[2026-06-16 02:47:41,988][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.2077033370733261, 'train/grad_norm_pre_clip': 9.459328651428223, 'timing/data': 0.006522662995848805, 'timing/model': 1.6826972319977358, 'learning_rate/qwen_vl_interface': 3.872240680367754e-06, 'learning_rate/action_model': 3.872240680367754e-05, 'epoch': 1.29}) +[2026-06-16 02:47:47,706][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.11528100073337555, 'train/grad_norm_pre_clip': 14.735600471496582, 'timing/data': 0.006385651999153197, 'timing/model': 1.5042095360113308, 'learning_rate/qwen_vl_interface': 3.869326464651566e-06, 'learning_rate/action_model': 3.869326464651566e-05, 'epoch': 1.3}) +[2026-06-16 02:47:53,841][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.04639098048210144, 'train/grad_norm_pre_clip': 4.596045017242432, 'timing/data': 0.006780562980566174, 'timing/model': 1.5741757710347883, 'learning_rate/qwen_vl_interface': 3.866412816478114e-06, 'learning_rate/action_model': 3.8664128164781144e-05, 'epoch': 1.3}) +[2026-06-16 02:47:59,693][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.2142970710992813, 'train/grad_norm_pre_clip': 12.479480743408203, 'timing/data': 0.006389115995261818, 'timing/model': 1.4892049290356226, 'learning_rate/qwen_vl_interface': 3.8634997370450915e-06, 'learning_rate/action_model': 3.8634997370450914e-05, 'epoch': 1.3}) +[2026-06-16 02:48:05,561][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.04285559803247452, 'train/grad_norm_pre_clip': 2.7594709396362305, 'timing/data': 0.006717530020978302, 'timing/model': 1.465645726013463, 'learning_rate/qwen_vl_interface': 3.860587227549956e-06, 'learning_rate/action_model': 3.860587227549955e-05, 'epoch': 1.3}) +[2026-06-16 02:48:11,462][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.05390196293592453, 'train/grad_norm_pre_clip': 4.2725911140441895, 'timing/data': 0.006774713983759284, 'timing/model': 1.4797370950109325, 'learning_rate/qwen_vl_interface': 3.857675289189926e-06, 'learning_rate/action_model': 3.857675289189926e-05, 'epoch': 1.3}) +[2026-06-16 02:48:17,394][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.13899771869182587, 'train/grad_norm_pre_clip': 5.673895835876465, 'timing/data': 0.006863504997454584, 'timing/model': 1.6891350859659724, 'learning_rate/qwen_vl_interface': 3.854763923161993e-06, 'learning_rate/action_model': 3.854763923161992e-05, 'epoch': 1.3}) +[2026-06-16 02:48:23,064][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.04473383352160454, 'train/grad_norm_pre_clip': 5.255086421966553, 'timing/data': 0.0066359000047668815, 'timing/model': 1.459075586986728, 'learning_rate/qwen_vl_interface': 3.851853130662908e-06, 'learning_rate/action_model': 3.8518531306629077e-05, 'epoch': 1.3}) +[2026-06-16 02:48:29,066][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.0906856432557106, 'train/grad_norm_pre_clip': 4.585334777832031, 'timing/data': 0.006733864021953195, 'timing/model': 1.457969025010243, 'learning_rate/qwen_vl_interface': 3.8489429128891884e-06, 'learning_rate/action_model': 3.848942912889188e-05, 'epoch': 1.3}) +[2026-06-16 02:48:35,462][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.06474538892507553, 'train/grad_norm_pre_clip': 3.0733683109283447, 'timing/data': 0.006975039024837315, 'timing/model': 1.4655606260057539, 'learning_rate/qwen_vl_interface': 3.846033271037114e-06, 'learning_rate/action_model': 3.8460332710371136e-05, 'epoch': 1.3}) +[2026-06-16 02:48:41,628][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.0971149131655693, 'train/grad_norm_pre_clip': 7.750946998596191, 'timing/data': 0.0064068519859574735, 'timing/model': 1.5606214090366848, 'learning_rate/qwen_vl_interface': 3.843124206302732e-06, 'learning_rate/action_model': 3.8431242063027316e-05, 'epoch': 1.3}) +[2026-06-16 02:48:47,570][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.035360198467969894, 'train/grad_norm_pre_clip': 3.209048271179199, 'timing/data': 0.0074376489501446486, 'timing/model': 1.4740890079992823, 'learning_rate/qwen_vl_interface': 3.840215719881844e-06, 'learning_rate/action_model': 3.8402157198818445e-05, 'epoch': 1.3}) +[2026-06-16 02:48:53,663][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.2201908379793167, 'train/grad_norm_pre_clip': 11.81391716003418, 'timing/data': 0.006839148001745343, 'timing/model': 1.752301089989487, 'learning_rate/qwen_vl_interface': 3.837307812970022e-06, 'learning_rate/action_model': 3.837307812970022e-05, 'epoch': 1.3}) +[2026-06-16 02:48:59,720][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.05186283588409424, 'train/grad_norm_pre_clip': 7.988397121429443, 'timing/data': 0.006667095993179828, 'timing/model': 1.5385648749652319, 'learning_rate/qwen_vl_interface': 3.834400486762596e-06, 'learning_rate/action_model': 3.834400486762596e-05, 'epoch': 1.3}) +[2026-06-16 02:49:05,844][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.046155694872140884, 'train/grad_norm_pre_clip': 4.385798454284668, 'timing/data': 0.00708208599826321, 'timing/model': 1.5547779150074348, 'learning_rate/qwen_vl_interface': 3.8314937424546585e-06, 'learning_rate/action_model': 3.8314937424546584e-05, 'epoch': 1.3}) +[2026-06-16 02:49:11,991][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.17144986987113953, 'train/grad_norm_pre_clip': 9.833946228027344, 'timing/data': 0.006571345031261444, 'timing/model': 1.4845961299724877, 'learning_rate/qwen_vl_interface': 3.828587581241064e-06, 'learning_rate/action_model': 3.8285875812410635e-05, 'epoch': 1.3}) +[2026-06-16 02:49:18,240][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.045975588262081146, 'train/grad_norm_pre_clip': 6.690943241119385, 'timing/data': 0.006208601000253111, 'timing/model': 1.5502094560069963, 'learning_rate/qwen_vl_interface': 3.825682004316419e-06, 'learning_rate/action_model': 3.825682004316419e-05, 'epoch': 1.3}) +[2026-06-16 02:49:24,538][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.10199746489524841, 'train/grad_norm_pre_clip': 8.610960006713867, 'timing/data': 0.006409937981516123, 'timing/model': 1.567756358999759, 'learning_rate/qwen_vl_interface': 3.822777012875104e-06, 'learning_rate/action_model': 3.822777012875103e-05, 'epoch': 1.3}) +[2026-06-16 02:49:30,858][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.018057778477668762, 'train/grad_norm_pre_clip': 1.7304422855377197, 'timing/data': 0.006513350002933294, 'timing/model': 1.7888680460164323, 'learning_rate/qwen_vl_interface': 3.819872608111248e-06, 'learning_rate/action_model': 3.819872608111248e-05, 'epoch': 1.3}) +[2026-06-16 02:49:36,827][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.13828524947166443, 'train/grad_norm_pre_clip': 7.113588333129883, 'timing/data': 0.00696117797633633, 'timing/model': 1.520147857023403, 'learning_rate/qwen_vl_interface': 3.816968791218741e-06, 'learning_rate/action_model': 3.816968791218741e-05, 'epoch': 1.3}) +[2026-06-16 02:49:43,024][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.23376500606536865, 'train/grad_norm_pre_clip': 12.586304664611816, 'timing/data': 0.006939565995708108, 'timing/model': 1.5458635830436833, 'learning_rate/qwen_vl_interface': 3.8140655633912337e-06, 'learning_rate/action_model': 3.814065563391234e-05, 'epoch': 1.3}) +[2026-06-16 02:49:49,156][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.15106099843978882, 'train/grad_norm_pre_clip': 6.2760396003723145, 'timing/data': 0.006538275978527963, 'timing/model': 1.5138911140384153, 'learning_rate/qwen_vl_interface': 3.811162925822133e-06, 'learning_rate/action_model': 3.811162925822132e-05, 'epoch': 1.3}) +[2026-06-16 02:49:55,335][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.035915542393922806, 'train/grad_norm_pre_clip': 2.728041887283325, 'timing/data': 0.006813461019191891, 'timing/model': 1.5537062679650262, 'learning_rate/qwen_vl_interface': 3.8082608797046045e-06, 'learning_rate/action_model': 3.8082608797046045e-05, 'epoch': 1.3}) +[2026-06-16 02:50:01,362][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.19094422459602356, 'train/grad_norm_pre_clip': 14.08218765258789, 'timing/data': 0.0065768229542300105, 'timing/model': 1.518160512961913, 'learning_rate/qwen_vl_interface': 3.8053594262315687e-06, 'learning_rate/action_model': 3.805359426231568e-05, 'epoch': 1.3}) +[2026-06-16 02:50:07,554][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.05276567488908768, 'train/grad_norm_pre_clip': 4.589264392852783, 'timing/data': 0.0065790360094979405, 'timing/model': 1.7832758619915694, 'learning_rate/qwen_vl_interface': 3.802458566595704e-06, 'learning_rate/action_model': 3.802458566595704e-05, 'epoch': 1.31}) +[2026-06-16 02:50:13,542][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.06900264322757721, 'train/grad_norm_pre_clip': 6.948300838470459, 'timing/data': 0.006876835017465055, 'timing/model': 1.5569870959734544, 'learning_rate/qwen_vl_interface': 3.7995583019894444e-06, 'learning_rate/action_model': 3.7995583019894445e-05, 'epoch': 1.31}) +[2026-06-16 02:50:19,837][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.028665592893958092, 'train/grad_norm_pre_clip': 9.918168067932129, 'timing/data': 0.0068936789757572114, 'timing/model': 1.555012055032421, 'learning_rate/qwen_vl_interface': 3.7966586336049782e-06, 'learning_rate/action_model': 3.796658633604978e-05, 'epoch': 1.31}) +[2026-06-16 02:50:26,040][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.022384630516171455, 'train/grad_norm_pre_clip': 2.157266139984131, 'timing/data': 0.006817576999310404, 'timing/model': 1.555464214994572, 'learning_rate/qwen_vl_interface': 3.7937595626342516e-06, 'learning_rate/action_model': 3.793759562634252e-05, 'epoch': 1.31}) +[2026-06-16 02:50:32,228][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.014010505750775337, 'train/grad_norm_pre_clip': 1.4306467771530151, 'timing/data': 0.006795043998863548, 'timing/model': 1.562312311027199, 'learning_rate/qwen_vl_interface': 3.79086109026896e-06, 'learning_rate/action_model': 3.7908610902689604e-05, 'epoch': 1.31}) +[2026-06-16 02:50:38,361][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.3349185287952423, 'train/grad_norm_pre_clip': 20.108152389526367, 'timing/data': 0.006761735014151782, 'timing/model': 1.478387900046073, 'learning_rate/qwen_vl_interface': 3.7879632177005584e-06, 'learning_rate/action_model': 3.787963217700558e-05, 'epoch': 1.31}) +[2026-06-16 02:50:44,294][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.13829442858695984, 'train/grad_norm_pre_clip': 10.681804656982422, 'timing/data': 0.0073398249805904925, 'timing/model': 1.686679519014433, 'learning_rate/qwen_vl_interface': 3.7850659461202523e-06, 'learning_rate/action_model': 3.785065946120252e-05, 'epoch': 1.31}) +[2026-06-16 02:50:49,966][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.10595833510160446, 'train/grad_norm_pre_clip': 11.070941925048828, 'timing/data': 0.006758500006981194, 'timing/model': 1.4593641179963015, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 1.31}) +[2026-06-16 02:50:55,966][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.2708151340484619, 'train/grad_norm_pre_clip': 14.341726303100586, 'timing/data': 0.0071467210073024035, 'timing/model': 1.4608845980255865, 'learning_rate/qwen_vl_interface': 3.779273210687514e-06, 'learning_rate/action_model': 3.779273210687514e-05, 'epoch': 1.31}) +[2026-06-16 02:51:01,967][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.039192162454128265, 'train/grad_norm_pre_clip': 6.3381667137146, 'timing/data': 0.0068729890044778585, 'timing/model': 1.450069534999784, 'learning_rate/qwen_vl_interface': 3.776377749216257e-06, 'learning_rate/action_model': 3.776377749216257e-05, 'epoch': 1.31}) +[2026-06-16 02:51:07,867][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.02989456057548523, 'train/grad_norm_pre_clip': 3.8438198566436768, 'timing/data': 0.007114433974493295, 'timing/model': 1.4573262719786726, 'learning_rate/qwen_vl_interface': 3.773482893495442e-06, 'learning_rate/action_model': 3.773482893495442e-05, 'epoch': 1.31}) +[2026-06-16 02:51:13,857][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.19043371081352234, 'train/grad_norm_pre_clip': 9.719189643859863, 'timing/data': 0.006811357976403087, 'timing/model': 1.517140570038464, 'learning_rate/qwen_vl_interface': 3.7705886447150375e-06, 'learning_rate/action_model': 3.7705886447150376e-05, 'epoch': 1.31}) +[2026-06-16 02:51:19,795][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.2734486162662506, 'train/grad_norm_pre_clip': 14.68623161315918, 'timing/data': 0.007445349998306483, 'timing/model': 1.6874547830084339, 'learning_rate/qwen_vl_interface': 3.767695004064759e-06, 'learning_rate/action_model': 3.767695004064759e-05, 'epoch': 1.31}) +[2026-06-16 02:51:25,466][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.08037488907575607, 'train/grad_norm_pre_clip': 10.015066146850586, 'timing/data': 0.006433351954910904, 'timing/model': 1.4588791760033928, 'learning_rate/qwen_vl_interface': 3.7648019727340725e-06, 'learning_rate/action_model': 3.764801972734072e-05, 'epoch': 1.31}) +[2026-06-16 02:51:31,465][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.11596298217773438, 'train/grad_norm_pre_clip': 8.00074291229248, 'timing/data': 0.006815264001488686, 'timing/model': 1.459310573991388, 'learning_rate/qwen_vl_interface': 3.761909551912196e-06, 'learning_rate/action_model': 3.761909551912196e-05, 'epoch': 1.31}) +[2026-06-16 02:51:37,364][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.0527450330555439, 'train/grad_norm_pre_clip': 9.333477973937988, 'timing/data': 0.006443816993851215, 'timing/model': 1.4588179619750008, 'learning_rate/qwen_vl_interface': 3.75901774278809e-06, 'learning_rate/action_model': 3.7590177427880904e-05, 'epoch': 1.31}) +[2026-06-16 02:51:43,373][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.12788283824920654, 'train/grad_norm_pre_clip': 10.484485626220703, 'timing/data': 0.0071244590217247605, 'timing/model': 1.4815872250474058, 'learning_rate/qwen_vl_interface': 3.756126546550472e-06, 'learning_rate/action_model': 3.756126546550472e-05, 'epoch': 1.31}) +[2026-06-16 02:51:49,356][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.19818751513957977, 'train/grad_norm_pre_clip': 11.819120407104492, 'timing/data': 0.0074829760123975575, 'timing/model': 1.514970789954532, 'learning_rate/qwen_vl_interface': 3.753235964387802e-06, 'learning_rate/action_model': 3.753235964387802e-05, 'epoch': 1.31}) +[2026-06-16 02:51:55,387][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.08831503242254257, 'train/grad_norm_pre_clip': 6.586131572723389, 'timing/data': 0.006584805028978735, 'timing/model': 1.6985570259857923, 'learning_rate/qwen_vl_interface': 3.750345997488289e-06, 'learning_rate/action_model': 3.7503459974882886e-05, 'epoch': 1.31}) +[2026-06-16 02:52:01,210][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.02101290039718151, 'train/grad_norm_pre_clip': 3.5206351280212402, 'timing/data': 0.006735336035490036, 'timing/model': 1.602451116021257, 'learning_rate/qwen_vl_interface': 3.7474566470398903e-06, 'learning_rate/action_model': 3.74745664703989e-05, 'epoch': 1.31}) +[2026-06-16 02:52:06,867][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.20634295046329498, 'train/grad_norm_pre_clip': 17.25523567199707, 'timing/data': 0.006654536991845816, 'timing/model': 1.4610414790222421, 'learning_rate/qwen_vl_interface': 3.744567914230307e-06, 'learning_rate/action_model': 3.744567914230307e-05, 'epoch': 1.31}) +[2026-06-16 02:52:12,770][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.004060762003064156, 'train/grad_norm_pre_clip': 0.41435810923576355, 'timing/data': 0.006637673010118306, 'timing/model': 1.462570371979382, 'learning_rate/qwen_vl_interface': 3.7416798002469902e-06, 'learning_rate/action_model': 3.74167980024699e-05, 'epoch': 1.31}) +[2026-06-16 02:52:18,763][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.07217983156442642, 'train/grad_norm_pre_clip': 4.593227386474609, 'timing/data': 0.006843015027698129, 'timing/model': 1.4747815149603412, 'learning_rate/qwen_vl_interface': 3.7387923062771313e-06, 'learning_rate/action_model': 3.738792306277131e-05, 'epoch': 1.31}) +[2026-06-16 02:52:24,757][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.007840322330594063, 'train/grad_norm_pre_clip': 2.5685157775878906, 'timing/data': 0.007312044966965914, 'timing/model': 1.5137820320087485, 'learning_rate/qwen_vl_interface': 3.7359054335076732e-06, 'learning_rate/action_model': 3.7359054335076734e-05, 'epoch': 1.31}) +[2026-06-16 02:52:30,691][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.11822065711021423, 'train/grad_norm_pre_clip': 8.367401123046875, 'timing/data': 0.006733002956025302, 'timing/model': 1.68601490103174, 'learning_rate/qwen_vl_interface': 3.7330191831253006e-06, 'learning_rate/action_model': 3.7330191831253e-05, 'epoch': 1.32}) +[2026-06-16 02:52:36,596][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.029237525537610054, 'train/grad_norm_pre_clip': 2.9217655658721924, 'timing/data': 0.006598324980586767, 'timing/model': 1.688907338946592, 'learning_rate/qwen_vl_interface': 3.7301335563164382e-06, 'learning_rate/action_model': 3.730133556316438e-05, 'epoch': 1.32}) +[2026-06-16 02:52:42,266][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.0519399456679821, 'train/grad_norm_pre_clip': 6.138342380523682, 'timing/data': 0.006657162040937692, 'timing/model': 1.4591509469901212, 'learning_rate/qwen_vl_interface': 3.727248554267261e-06, 'learning_rate/action_model': 3.727248554267261e-05, 'epoch': 1.32}) +[2026-06-16 02:52:48,298][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.1524578332901001, 'train/grad_norm_pre_clip': 11.373164176940918, 'timing/data': 0.006946817971765995, 'timing/model': 1.5242150000412948, 'learning_rate/qwen_vl_interface': 3.7243641781636838e-06, 'learning_rate/action_model': 3.7243641781636836e-05, 'epoch': 1.32}) +[2026-06-16 02:52:54,456][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.04086115211248398, 'train/grad_norm_pre_clip': 4.578466892242432, 'timing/data': 0.00718613900244236, 'timing/model': 1.559258065011818, 'learning_rate/qwen_vl_interface': 3.721480429191364e-06, 'learning_rate/action_model': 3.721480429191364e-05, 'epoch': 1.32}) +[2026-06-16 02:53:00,607][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.01844131015241146, 'train/grad_norm_pre_clip': 2.3418822288513184, 'timing/data': 0.006566468044184148, 'timing/model': 1.4937324809725396, 'learning_rate/qwen_vl_interface': 3.7185973085357045e-06, 'learning_rate/action_model': 3.718597308535705e-05, 'epoch': 1.32}) +[2026-06-16 02:53:06,490][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.053557123988866806, 'train/grad_norm_pre_clip': 9.392285346984863, 'timing/data': 0.006639024999458343, 'timing/model': 1.6831699369940907, 'learning_rate/qwen_vl_interface': 3.7157148173818443e-06, 'learning_rate/action_model': 3.715714817381844e-05, 'epoch': 1.32}) +[2026-06-16 02:53:12,516][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.16786602139472961, 'train/grad_norm_pre_clip': 11.645069122314453, 'timing/data': 0.007346856989897788, 'timing/model': 1.733508039964363, 'learning_rate/qwen_vl_interface': 3.7128329569146672e-06, 'learning_rate/action_model': 3.7128329569146666e-05, 'epoch': 1.32}) +[2026-06-16 02:53:18,539][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.16245707869529724, 'train/grad_norm_pre_clip': 14.531381607055664, 'timing/data': 0.006813600950408727, 'timing/model': 1.5506180290249176, 'learning_rate/qwen_vl_interface': 3.7099517283187977e-06, 'learning_rate/action_model': 3.7099517283187976e-05, 'epoch': 1.32}) +[2026-06-16 02:53:24,663][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.008482515811920166, 'train/grad_norm_pre_clip': 1.1969573497772217, 'timing/data': 0.006661217019427568, 'timing/model': 1.4592687109834515, 'learning_rate/qwen_vl_interface': 3.7070711327786e-06, 'learning_rate/action_model': 3.7070711327785997e-05, 'epoch': 1.32}) +[2026-06-16 02:53:30,766][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.11674914509057999, 'train/grad_norm_pre_clip': 9.127352714538574, 'timing/data': 0.007635932008270174, 'timing/model': 1.574276082974393, 'learning_rate/qwen_vl_interface': 3.704191171478179e-06, 'learning_rate/action_model': 3.704191171478179e-05, 'epoch': 1.32}) +[2026-06-16 02:53:36,746][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.01978161372244358, 'train/grad_norm_pre_clip': 6.62130880355835, 'timing/data': 0.0063429989968426526, 'timing/model': 1.514809699030593, 'learning_rate/qwen_vl_interface': 3.701311845601376e-06, 'learning_rate/action_model': 3.701311845601376e-05, 'epoch': 1.32}) +[2026-06-16 02:53:42,921][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.10310397297143936, 'train/grad_norm_pre_clip': 10.02934455871582, 'timing/data': 0.006470787979196757, 'timing/model': 1.5572795399930328, 'learning_rate/qwen_vl_interface': 3.6984331563317744e-06, 'learning_rate/action_model': 3.698433156331774e-05, 'epoch': 1.32}) +[2026-06-16 02:53:49,052][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.21679776906967163, 'train/grad_norm_pre_clip': 13.33178997039795, 'timing/data': 0.006712542963214219, 'timing/model': 1.6693128069746308, 'learning_rate/qwen_vl_interface': 3.695555104852695e-06, 'learning_rate/action_model': 3.6955551048526946e-05, 'epoch': 1.32}) +[2026-06-16 02:53:54,969][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.1479163020849228, 'train/grad_norm_pre_clip': 18.244726181030273, 'timing/data': 0.006784788973163813, 'timing/model': 1.4587005579960532, 'learning_rate/qwen_vl_interface': 3.6926776923471948e-06, 'learning_rate/action_model': 3.6926776923471946e-05, 'epoch': 1.32}) +[2026-06-16 02:54:01,167][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.03422153741121292, 'train/grad_norm_pre_clip': 4.3076491355896, 'timing/data': 0.006389759015291929, 'timing/model': 1.531898446031846, 'learning_rate/qwen_vl_interface': 3.6898009199980715e-06, 'learning_rate/action_model': 3.6898009199980714e-05, 'epoch': 1.32}) +[2026-06-16 02:54:07,164][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.02132246270775795, 'train/grad_norm_pre_clip': 2.7634804248809814, 'timing/data': 0.00672692502848804, 'timing/model': 1.475384032994043, 'learning_rate/qwen_vl_interface': 3.686924788987855e-06, 'learning_rate/action_model': 3.686924788987855e-05, 'epoch': 1.32}) +[2026-06-16 02:54:13,222][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.10088826715946198, 'train/grad_norm_pre_clip': 8.592426300048828, 'timing/data': 0.006793271983042359, 'timing/model': 1.5744864419684745, 'learning_rate/qwen_vl_interface': 3.6840493004988155e-06, 'learning_rate/action_model': 3.684049300498815e-05, 'epoch': 1.32}) +[2026-06-16 02:54:19,468][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.13965970277786255, 'train/grad_norm_pre_clip': 18.461456298828125, 'timing/data': 0.006396809010766447, 'timing/model': 1.8006522750365548, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 1.32}) +[2026-06-16 02:54:25,538][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.48992758989334106, 'train/grad_norm_pre_clip': 20.803165435791016, 'timing/data': 0.006723880011122674, 'timing/model': 1.5781391700147651, 'learning_rate/qwen_vl_interface': 3.6783002558120206e-06, 'learning_rate/action_model': 3.67830025581202e-05, 'epoch': 1.32}) +[2026-06-16 02:54:31,833][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.05869920924305916, 'train/grad_norm_pre_clip': 7.75632381439209, 'timing/data': 0.006949722010176629, 'timing/model': 1.5579369280021638, 'learning_rate/qwen_vl_interface': 3.6754267019774804e-06, 'learning_rate/action_model': 3.67542670197748e-05, 'epoch': 1.32}) +[2026-06-16 02:54:37,940][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.07291214913129807, 'train/grad_norm_pre_clip': 10.514762878417969, 'timing/data': 0.007115715998224914, 'timing/model': 1.5642466510180384, 'learning_rate/qwen_vl_interface': 3.6725537953905453e-06, 'learning_rate/action_model': 3.672553795390545e-05, 'epoch': 1.32}) +[2026-06-16 02:54:44,134][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.07894200086593628, 'train/grad_norm_pre_clip': 6.848736763000488, 'timing/data': 0.006555452011525631, 'timing/model': 1.5595423039630987, 'learning_rate/qwen_vl_interface': 3.6696815372321603e-06, 'learning_rate/action_model': 3.6696815372321604e-05, 'epoch': 1.32}) +[2026-06-16 02:54:50,434][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.24452877044677734, 'train/grad_norm_pre_clip': 17.663991928100586, 'timing/data': 0.006846801028586924, 'timing/model': 1.5699997390038334, 'learning_rate/qwen_vl_interface': 3.6668099286830023e-06, 'learning_rate/action_model': 3.666809928683002e-05, 'epoch': 1.33}) +[2026-06-16 02:54:56,602][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.13926127552986145, 'train/grad_norm_pre_clip': 15.409167289733887, 'timing/data': 0.0066616080002859235, 'timing/model': 1.69442249200074, 'learning_rate/qwen_vl_interface': 3.66393897092348e-06, 'learning_rate/action_model': 3.66393897092348e-05, 'epoch': 1.33}) +[2026-06-16 02:55:02,365][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.13745862245559692, 'train/grad_norm_pre_clip': 11.71837329864502, 'timing/data': 0.006777659000363201, 'timing/model': 1.4593770409701392, 'learning_rate/qwen_vl_interface': 3.661068665133735e-06, 'learning_rate/action_model': 3.661068665133735e-05, 'epoch': 1.33}) +[2026-06-16 02:55:08,366][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.2995043396949768, 'train/grad_norm_pre_clip': 16.109533309936523, 'timing/data': 0.007142616028431803, 'timing/model': 1.456569774018135, 'learning_rate/qwen_vl_interface': 3.6581990124936434e-06, 'learning_rate/action_model': 3.6581990124936434e-05, 'epoch': 1.33}) +[2026-06-16 02:55:14,366][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.0493604838848114, 'train/grad_norm_pre_clip': 6.526252746582031, 'timing/data': 0.006653585995081812, 'timing/model': 1.4558733220328577, 'learning_rate/qwen_vl_interface': 3.65533001418281e-06, 'learning_rate/action_model': 3.655330014182809e-05, 'epoch': 1.33}) +[2026-06-16 02:55:20,365][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.0992807000875473, 'train/grad_norm_pre_clip': 7.916057109832764, 'timing/data': 0.0069316449807956815, 'timing/model': 1.479192694008816, 'learning_rate/qwen_vl_interface': 3.652461671380572e-06, 'learning_rate/action_model': 3.652461671380572e-05, 'epoch': 1.33}) +[2026-06-16 02:55:26,265][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.02127298340201378, 'train/grad_norm_pre_clip': 3.0627169609069824, 'timing/data': 0.006793962966185063, 'timing/model': 1.4630801160237752, 'learning_rate/qwen_vl_interface': 3.6495939852659966e-06, 'learning_rate/action_model': 3.649593985265997e-05, 'epoch': 1.33}) +[2026-06-16 02:55:32,200][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.08710770308971405, 'train/grad_norm_pre_clip': 6.337512493133545, 'timing/data': 0.006926507980097085, 'timing/model': 1.6923367499839514, 'learning_rate/qwen_vl_interface': 3.646726957017881e-06, 'learning_rate/action_model': 3.646726957017881e-05, 'epoch': 1.33}) +[2026-06-16 02:55:37,966][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.08330467343330383, 'train/grad_norm_pre_clip': 9.62525463104248, 'timing/data': 0.006453312002122402, 'timing/model': 1.5043190880096518, 'learning_rate/qwen_vl_interface': 3.643860587814752e-06, 'learning_rate/action_model': 3.643860587814752e-05, 'epoch': 1.33}) +[2026-06-16 02:55:44,237][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.04134340584278107, 'train/grad_norm_pre_clip': 6.095707416534424, 'timing/data': 0.006516825989820063, 'timing/model': 1.5536819589906372, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 1.33}) +[2026-06-16 02:55:50,541][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.32234540581703186, 'train/grad_norm_pre_clip': 23.922178268432617, 'timing/data': 0.007259928970597684, 'timing/model': 1.5754365000175312, 'learning_rate/qwen_vl_interface': 3.6381298312562095e-06, 'learning_rate/action_model': 3.6381298312562094e-05, 'epoch': 1.33}) +[2026-06-16 02:55:56,782][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.034856561571359634, 'train/grad_norm_pre_clip': 4.478333473205566, 'timing/data': 0.00689152698032558, 'timing/model': 1.5431095320382155, 'learning_rate/qwen_vl_interface': 3.635265446256493e-06, 'learning_rate/action_model': 3.6352654462564926e-05, 'epoch': 1.33}) +[2026-06-16 02:56:02,671][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.09392447024583817, 'train/grad_norm_pre_clip': 10.67803955078125, 'timing/data': 0.006693015049677342, 'timing/model': 1.4498879559687339, 'learning_rate/qwen_vl_interface': 3.6324017250131565e-06, 'learning_rate/action_model': 3.6324017250131567e-05, 'epoch': 1.33}) +[2026-06-16 02:56:08,710][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.02398192696273327, 'train/grad_norm_pre_clip': 3.0942728519439697, 'timing/data': 0.007928933016955853, 'timing/model': 1.7332128990092315, 'learning_rate/qwen_vl_interface': 3.6295386687033696e-06, 'learning_rate/action_model': 3.6295386687033695e-05, 'epoch': 1.33}) +[2026-06-16 02:56:14,472][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.021538235247135162, 'train/grad_norm_pre_clip': 5.703918933868408, 'timing/data': 0.007689880963880569, 'timing/model': 1.4547252950142138, 'learning_rate/qwen_vl_interface': 3.6266762785040253e-06, 'learning_rate/action_model': 3.626676278504025e-05, 'epoch': 1.33}) +[2026-06-16 02:56:20,464][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.17859211564064026, 'train/grad_norm_pre_clip': 10.64051342010498, 'timing/data': 0.006664893007837236, 'timing/model': 1.4587056820164435, 'learning_rate/qwen_vl_interface': 3.6238145555917464e-06, 'learning_rate/action_model': 3.6238145555917465e-05, 'epoch': 1.33}) +[2026-06-16 02:56:26,364][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.05173352733254433, 'train/grad_norm_pre_clip': 6.338646411895752, 'timing/data': 0.006547791010234505, 'timing/model': 1.457372160977684, 'learning_rate/qwen_vl_interface': 3.6209535011428757e-06, 'learning_rate/action_model': 3.620953501142876e-05, 'epoch': 1.33}) +[2026-06-16 02:56:32,263][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.05697336420416832, 'train/grad_norm_pre_clip': 9.976432800292969, 'timing/data': 0.006437439005821943, 'timing/model': 1.4574128439999186, 'learning_rate/qwen_vl_interface': 3.6180931163334863e-06, 'learning_rate/action_model': 3.6180931163334863e-05, 'epoch': 1.33}) +[2026-06-16 02:56:38,273][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.1654849648475647, 'train/grad_norm_pre_clip': 9.825791358947754, 'timing/data': 0.006579728971701115, 'timing/model': 1.5368021189933643, 'learning_rate/qwen_vl_interface': 3.6152334023393747e-06, 'learning_rate/action_model': 3.6152334023393745e-05, 'epoch': 1.33}) +[2026-06-16 02:56:44,193][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.10837041586637497, 'train/grad_norm_pre_clip': 10.030208587646484, 'timing/data': 0.006499780050944537, 'timing/model': 1.6872037879657, 'learning_rate/qwen_vl_interface': 3.6123743603360613e-06, 'learning_rate/action_model': 3.612374360336061e-05, 'epoch': 1.33}) +[2026-06-16 03:05:56,648][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.07687470316886902, 'train/grad_norm_pre_clip': 8.188011169433594, 'eval/loss': 0.05162375420331955, 'eval/flap_precision': 0.9520833333333333, 'eval/flap_recall': 0.9158316633266533, 'eval/flap_f1': 0.933605720122574, 'eval/flap_rate_model': 0.13333333333333333, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.933605720122574, 'timing/data': 0.006777508999221027, 'timing/model': 1.4619486409937963, 'learning_rate/qwen_vl_interface': 3.6095159914987893e-06, 'learning_rate/action_model': 3.6095159914987894e-05, 'epoch': 1.33}) +[2026-06-16 03:05:56,649][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3100_state +[2026-06-16 03:05:56,653][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 03:06:02,310][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3100_state/model.safetensors +[2026-06-16 03:06:15,874][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3100_state/optimizer.bin +[2026-06-16 03:06:15,875][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3100_state/sampler.bin +[2026-06-16 03:06:15,875][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3100_state/sampler_1.bin +[2026-06-16 03:06:15,876][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3100_state/random_states_0.pkl +[2026-06-16 03:06:15,990][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3100 +[2026-06-16 03:06:15,990][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 03:06:16,000][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 03:06:22,062][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.007775350008159876, 'train/grad_norm_pre_clip': 0.6243131160736084, 'timing/data': 0.006605583999771625, 'timing/model': 1.780835304001812, 'learning_rate/qwen_vl_interface': 3.6066582970025266e-06, 'learning_rate/action_model': 3.6066582970025264e-05, 'epoch': 1.33}) +[2026-06-16 03:06:28,031][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.1088472455739975, 'train/grad_norm_pre_clip': 9.361335754394531, 'timing/data': 0.0068257569801062346, 'timing/model': 1.4987634849967435, 'learning_rate/qwen_vl_interface': 3.603801278021964e-06, 'learning_rate/action_model': 3.6038012780219635e-05, 'epoch': 1.33}) +[2026-06-16 03:06:33,954][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.057903967797756195, 'train/grad_norm_pre_clip': 6.332563400268555, 'timing/data': 0.0065247740130871534, 'timing/model': 1.4573864720296115, 'learning_rate/qwen_vl_interface': 3.6009449357315122e-06, 'learning_rate/action_model': 3.6009449357315125e-05, 'epoch': 1.34}) +[2026-06-16 03:06:40,254][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.018265988677740097, 'train/grad_norm_pre_clip': 3.7589943408966064, 'timing/data': 0.007252956973388791, 'timing/model': 1.5850659299758263, 'learning_rate/qwen_vl_interface': 3.5980892713053084e-06, 'learning_rate/action_model': 3.598089271305308e-05, 'epoch': 1.34}) +[2026-06-16 03:06:46,280][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.058528926223516464, 'train/grad_norm_pre_clip': 13.04607105255127, 'timing/data': 0.006990670983213931, 'timing/model': 1.477901753038168, 'learning_rate/qwen_vl_interface': 3.5952342859172063e-06, 'learning_rate/action_model': 3.595234285917206e-05, 'epoch': 1.34}) +[2026-06-16 03:06:52,262][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.009204707108438015, 'train/grad_norm_pre_clip': 1.2614116668701172, 'timing/data': 0.006639992992859334, 'timing/model': 1.4349740489851683, 'learning_rate/qwen_vl_interface': 3.592379980740785e-06, 'learning_rate/action_model': 3.5923799807407845e-05, 'epoch': 1.34}) +[2026-06-16 03:06:58,258][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.011423895135521889, 'train/grad_norm_pre_clip': 1.4825623035430908, 'timing/data': 0.006585283030290157, 'timing/model': 1.7586164560052566, 'learning_rate/qwen_vl_interface': 3.5895263569493384e-06, 'learning_rate/action_model': 3.589526356949338e-05, 'epoch': 1.34}) +[2026-06-16 03:07:03,957][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.1390973925590515, 'train/grad_norm_pre_clip': 8.349206924438477, 'timing/data': 0.006958792975638062, 'timing/model': 1.4581910680280998, 'learning_rate/qwen_vl_interface': 3.586673415715885e-06, 'learning_rate/action_model': 3.586673415715885e-05, 'epoch': 1.34}) +[2026-06-16 03:07:10,006][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.03274175524711609, 'train/grad_norm_pre_clip': 6.064845085144043, 'timing/data': 0.0066733019775711, 'timing/model': 1.4895471449708566, 'learning_rate/qwen_vl_interface': 3.5838211582131603e-06, 'learning_rate/action_model': 3.58382115821316e-05, 'epoch': 1.34}) +[2026-06-16 03:07:15,861][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.0234619602560997, 'train/grad_norm_pre_clip': 2.0906314849853516, 'timing/data': 0.006400940997991711, 'timing/model': 1.4611662300303578, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 1.34}) +[2026-06-16 03:07:21,762][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.051547821611166, 'train/grad_norm_pre_clip': 12.604591369628906, 'timing/data': 0.006561436981428415, 'timing/model': 1.4592623939970508, 'learning_rate/qwen_vl_interface': 3.5781186990894423e-06, 'learning_rate/action_model': 3.578118699089442e-05, 'epoch': 1.34}) +[2026-06-16 03:07:27,865][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.004111988935619593, 'train/grad_norm_pre_clip': 2.0132625102996826, 'timing/data': 0.006909110001288354, 'timing/model': 1.4596735700033605, 'learning_rate/qwen_vl_interface': 3.5752684998125113e-06, 'learning_rate/action_model': 3.5752684998125116e-05, 'epoch': 1.34}) +[2026-06-16 03:07:34,166][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.12216609716415405, 'train/grad_norm_pre_clip': 10.047592163085938, 'timing/data': 0.0067032360238954425, 'timing/model': 1.793529653979931, 'learning_rate/qwen_vl_interface': 3.5724189889544407e-06, 'learning_rate/action_model': 3.572418988954441e-05, 'epoch': 1.34}) +[2026-06-16 03:07:40,238][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.1544933021068573, 'train/grad_norm_pre_clip': 14.34073257446289, 'timing/data': 0.006352288997732103, 'timing/model': 1.5644997759955004, 'learning_rate/qwen_vl_interface': 3.569570167686556e-06, 'learning_rate/action_model': 3.569570167686556e-05, 'epoch': 1.34}) +[2026-06-16 03:07:46,532][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.060285162180662155, 'train/grad_norm_pre_clip': 8.714079856872559, 'timing/data': 0.006779117044061422, 'timing/model': 1.549588505004067, 'learning_rate/qwen_vl_interface': 3.5667220371798995e-06, 'learning_rate/action_model': 3.566722037179899e-05, 'epoch': 1.34}) +[2026-06-16 03:07:52,663][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.04896281659603119, 'train/grad_norm_pre_clip': 5.470011234283447, 'timing/data': 0.007026731967926025, 'timing/model': 1.479062173049897, 'learning_rate/qwen_vl_interface': 3.563874598605232e-06, 'learning_rate/action_model': 3.5638745986052316e-05, 'epoch': 1.34}) +[2026-06-16 03:07:58,666][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.07801183313131332, 'train/grad_norm_pre_clip': 15.164887428283691, 'timing/data': 0.007042776036541909, 'timing/model': 1.4783173279720359, 'learning_rate/qwen_vl_interface': 3.561027853133025e-06, 'learning_rate/action_model': 3.561027853133025e-05, 'epoch': 1.34}) +[2026-06-16 03:08:04,657][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.2563496232032776, 'train/grad_norm_pre_clip': 10.09846305847168, 'timing/data': 0.006715133029501885, 'timing/model': 1.5185023899539374, 'learning_rate/qwen_vl_interface': 3.558181801933472e-06, 'learning_rate/action_model': 3.5581818019334714e-05, 'epoch': 1.34}) +[2026-06-16 03:08:10,591][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.036282699555158615, 'train/grad_norm_pre_clip': 6.4523539543151855, 'timing/data': 0.0066614539828151464, 'timing/model': 1.6762119340128265, 'learning_rate/qwen_vl_interface': 3.555336446176473e-06, 'learning_rate/action_model': 3.5553364461764726e-05, 'epoch': 1.34}) +[2026-06-16 03:08:16,493][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.03617047145962715, 'train/grad_norm_pre_clip': 7.310108661651611, 'timing/data': 0.006482819968368858, 'timing/model': 1.6888256140518934, 'learning_rate/qwen_vl_interface': 3.552491787031647e-06, 'learning_rate/action_model': 3.552491787031647e-05, 'epoch': 1.34}) +[2026-06-16 03:08:22,165][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.09277018904685974, 'train/grad_norm_pre_clip': 7.830437660217285, 'timing/data': 0.006730695022270083, 'timing/model': 1.4588846620172262, 'learning_rate/qwen_vl_interface': 3.5496478256683277e-06, 'learning_rate/action_model': 3.549647825668327e-05, 'epoch': 1.34}) +[2026-06-16 03:08:28,067][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.24103842675685883, 'train/grad_norm_pre_clip': 25.42873191833496, 'timing/data': 0.006682834995444864, 'timing/model': 1.4584984380053356, 'learning_rate/qwen_vl_interface': 3.5468045632555585e-06, 'learning_rate/action_model': 3.546804563255559e-05, 'epoch': 1.34}) +[2026-06-16 03:08:34,149][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.09343545138835907, 'train/grad_norm_pre_clip': 7.282669544219971, 'timing/data': 0.006534015003126115, 'timing/model': 1.514885985001456, 'learning_rate/qwen_vl_interface': 3.5439620009620995e-06, 'learning_rate/action_model': 3.543962000962099e-05, 'epoch': 1.34}) +[2026-06-16 03:08:40,439][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.14826826751232147, 'train/grad_norm_pre_clip': 17.901447296142578, 'timing/data': 0.006910981028340757, 'timing/model': 1.5346230110153556, 'learning_rate/qwen_vl_interface': 3.541120139956419e-06, 'learning_rate/action_model': 3.5411201399564184e-05, 'epoch': 1.34}) +[2026-06-16 03:08:46,760][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.005339223425835371, 'train/grad_norm_pre_clip': 1.0186430215835571, 'timing/data': 0.006962937011849135, 'timing/model': 1.7934843190014362, 'learning_rate/qwen_vl_interface': 3.538278981406697e-06, 'learning_rate/action_model': 3.5382789814066966e-05, 'epoch': 1.34}) +[2026-06-16 03:08:52,836][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.07279157638549805, 'train/grad_norm_pre_clip': 8.24306583404541, 'timing/data': 0.006911872013006359, 'timing/model': 1.5572931690257974, 'learning_rate/qwen_vl_interface': 3.5354385264808295e-06, 'learning_rate/action_model': 3.535438526480829e-05, 'epoch': 1.35}) +[2026-06-16 03:08:59,034][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.2312009632587433, 'train/grad_norm_pre_clip': 7.930309295654297, 'timing/data': 0.006590628006961197, 'timing/model': 1.557349043956492, 'learning_rate/qwen_vl_interface': 3.5325987763464196e-06, 'learning_rate/action_model': 3.5325987763464194e-05, 'epoch': 1.35}) +[2026-06-16 03:09:05,327][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.011821148917078972, 'train/grad_norm_pre_clip': 2.6088554859161377, 'timing/data': 0.006656554993242025, 'timing/model': 1.5415504249976948, 'learning_rate/qwen_vl_interface': 3.529759732170781e-06, 'learning_rate/action_model': 3.529759732170781e-05, 'epoch': 1.35}) +[2026-06-16 03:09:11,547][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.06551480293273926, 'train/grad_norm_pre_clip': 7.111921787261963, 'timing/data': 0.006865422998089343, 'timing/model': 1.537528613989707, 'learning_rate/qwen_vl_interface': 3.5269213951209375e-06, 'learning_rate/action_model': 3.5269213951209374e-05, 'epoch': 1.35}) +[2026-06-16 03:09:17,465][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.4140859842300415, 'train/grad_norm_pre_clip': 11.740147590637207, 'timing/data': 0.006403881998267025, 'timing/model': 1.4511264309985563, 'learning_rate/qwen_vl_interface': 3.5240837663636207e-06, 'learning_rate/action_model': 3.5240837663636204e-05, 'epoch': 1.35}) +[2026-06-16 03:09:23,583][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.0764104500412941, 'train/grad_norm_pre_clip': 16.654319763183594, 'timing/data': 0.006535195978358388, 'timing/model': 1.6678077180404216, 'learning_rate/qwen_vl_interface': 3.5212468470652745e-06, 'learning_rate/action_model': 3.521246847065274e-05, 'epoch': 1.35}) +[2026-06-16 03:09:29,566][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.040821533650159836, 'train/grad_norm_pre_clip': 8.253434181213379, 'timing/data': 0.006540373025927693, 'timing/model': 1.5856344049680047, 'learning_rate/qwen_vl_interface': 3.5184106383920487e-06, 'learning_rate/action_model': 3.5184106383920486e-05, 'epoch': 1.35}) +[2026-06-16 03:09:35,685][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.02198963053524494, 'train/grad_norm_pre_clip': 4.734913349151611, 'timing/data': 0.0066131409839726985, 'timing/model': 1.4811247859615833, 'learning_rate/qwen_vl_interface': 3.5155751415098022e-06, 'learning_rate/action_model': 3.515575141509802e-05, 'epoch': 1.35}) +[2026-06-16 03:09:41,669][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.06966940313577652, 'train/grad_norm_pre_clip': 8.721932411193848, 'timing/data': 0.006844831979833543, 'timing/model': 1.4482332360348664, 'learning_rate/qwen_vl_interface': 3.5127403575841e-06, 'learning_rate/action_model': 3.5127403575841e-05, 'epoch': 1.35}) +[2026-06-16 03:09:47,762][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.003952003084123135, 'train/grad_norm_pre_clip': 0.6893328428268433, 'timing/data': 0.006382229970768094, 'timing/model': 1.504424276005011, 'learning_rate/qwen_vl_interface': 3.509906287780215e-06, 'learning_rate/action_model': 3.5099062877802144e-05, 'epoch': 1.35}) +[2026-06-16 03:09:53,848][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.0920904129743576, 'train/grad_norm_pre_clip': 6.553824424743652, 'timing/data': 0.0066879509831778705, 'timing/model': 1.5143026119912975, 'learning_rate/qwen_vl_interface': 3.5070729332631247e-06, 'learning_rate/action_model': 3.507072933263125e-05, 'epoch': 1.35}) +[2026-06-16 03:09:59,950][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.051908332854509354, 'train/grad_norm_pre_clip': 10.81234359741211, 'timing/data': 0.006405054009519517, 'timing/model': 1.769228197983466, 'learning_rate/qwen_vl_interface': 3.5042402951975167e-06, 'learning_rate/action_model': 3.5042402951975167e-05, 'epoch': 1.35}) +[2026-06-16 03:10:05,889][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.17957273125648499, 'train/grad_norm_pre_clip': 10.769152641296387, 'timing/data': 0.006690363981761038, 'timing/model': 1.4842250159708783, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 1.35}) +[2026-06-16 03:10:11,939][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.10885004699230194, 'train/grad_norm_pre_clip': 12.436038970947266, 'timing/data': 0.006854436011053622, 'timing/model': 1.5329722209717147, 'learning_rate/qwen_vl_interface': 3.498577173078011e-06, 'learning_rate/action_model': 3.498577173078011e-05, 'epoch': 1.35}) +[2026-06-16 03:10:18,028][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.3136598467826843, 'train/grad_norm_pre_clip': 22.073881149291992, 'timing/data': 0.007333451008889824, 'timing/model': 1.5426449639489874, 'learning_rate/qwen_vl_interface': 3.495746691352008e-06, 'learning_rate/action_model': 3.495746691352008e-05, 'epoch': 1.35}) +[2026-06-16 03:10:24,162][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.1098652333021164, 'train/grad_norm_pre_clip': 9.692811965942383, 'timing/data': 0.006706847983878106, 'timing/model': 1.4635714319883846, 'learning_rate/qwen_vl_interface': 3.4929169307332763e-06, 'learning_rate/action_model': 3.492916930733276e-05, 'epoch': 1.35}) +[2026-06-16 03:10:30,157][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.08510291576385498, 'train/grad_norm_pre_clip': 12.973139762878418, 'timing/data': 0.006650775030720979, 'timing/model': 1.7522563809761778, 'learning_rate/qwen_vl_interface': 3.4900878923850234e-06, 'learning_rate/action_model': 3.490087892385023e-05, 'epoch': 1.35}) +[2026-06-16 03:10:35,964][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.22578711807727814, 'train/grad_norm_pre_clip': 15.275157928466797, 'timing/data': 0.0063635120168328285, 'timing/model': 1.459317867003847, 'learning_rate/qwen_vl_interface': 3.4872595774701607e-06, 'learning_rate/action_model': 3.487259577470161e-05, 'epoch': 1.35}) +[2026-06-16 03:10:41,973][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.3650585114955902, 'train/grad_norm_pre_clip': 32.244712829589844, 'timing/data': 0.006908675015438348, 'timing/model': 1.4666273709735833, 'learning_rate/qwen_vl_interface': 3.4844319871513014e-06, 'learning_rate/action_model': 3.484431987151301e-05, 'epoch': 1.35}) +[2026-06-16 03:10:47,967][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.3033682703971863, 'train/grad_norm_pre_clip': 24.881484985351562, 'timing/data': 0.02280627703294158, 'timing/model': 1.447916975012049, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 1.35}) +[2026-06-16 03:10:54,034][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.35971200466156006, 'train/grad_norm_pre_clip': 25.11143684387207, 'timing/data': 0.006665186025202274, 'timing/model': 1.5679104829905555, 'learning_rate/qwen_vl_interface': 3.478778984950556e-06, 'learning_rate/action_model': 3.478778984950556e-05, 'epoch': 1.35}) +[2026-06-16 03:11:00,005][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.1661740094423294, 'train/grad_norm_pre_clip': 14.782210350036621, 'timing/data': 0.0065810829983092844, 'timing/model': 1.5208920470322482, 'learning_rate/qwen_vl_interface': 3.475953575392405e-06, 'learning_rate/action_model': 3.475953575392405e-05, 'epoch': 1.35}) +[2026-06-16 03:11:05,908][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.17875055968761444, 'train/grad_norm_pre_clip': 12.538698196411133, 'timing/data': 0.007303285994566977, 'timing/model': 1.695410059997812, 'learning_rate/qwen_vl_interface': 3.473128895077728e-06, 'learning_rate/action_model': 3.473128895077728e-05, 'epoch': 1.35}) +[2026-06-16 03:11:11,827][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.07103365659713745, 'train/grad_norm_pre_clip': 12.596372604370117, 'timing/data': 0.006692195020150393, 'timing/model': 1.5614838960464112, 'learning_rate/qwen_vl_interface': 3.4703049451676453e-06, 'learning_rate/action_model': 3.4703049451676454e-05, 'epoch': 1.35}) +[2026-06-16 03:11:17,767][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.06071128696203232, 'train/grad_norm_pre_clip': 15.569642066955566, 'timing/data': 0.0064790420001372695, 'timing/model': 1.4587903319625184, 'learning_rate/qwen_vl_interface': 3.4674817268229732e-06, 'learning_rate/action_model': 3.467481726822973e-05, 'epoch': 1.36}) +[2026-06-16 03:11:23,864][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.20572960376739502, 'train/grad_norm_pre_clip': 10.8820161819458, 'timing/data': 0.007942325959447771, 'timing/model': 1.519162680953741, 'learning_rate/qwen_vl_interface': 3.464659241204232e-06, 'learning_rate/action_model': 3.4646592412042314e-05, 'epoch': 1.36}) +[2026-06-16 03:11:29,767][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.12968087196350098, 'train/grad_norm_pre_clip': 7.037480354309082, 'timing/data': 0.006393665971700102, 'timing/model': 1.4593625160050578, 'learning_rate/qwen_vl_interface': 3.4618374894716377e-06, 'learning_rate/action_model': 3.461837489471638e-05, 'epoch': 1.36}) +[2026-06-16 03:11:35,860][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.08181553333997726, 'train/grad_norm_pre_clip': 5.331235408782959, 'timing/data': 0.006401626975275576, 'timing/model': 1.5316816970007494, 'learning_rate/qwen_vl_interface': 3.459016472785107e-06, 'learning_rate/action_model': 3.459016472785107e-05, 'epoch': 1.36}) +[2026-06-16 03:11:41,931][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.12246034294366837, 'train/grad_norm_pre_clip': 8.694355010986328, 'timing/data': 0.007916157017461956, 'timing/model': 1.7204938780050725, 'learning_rate/qwen_vl_interface': 3.456196192304254e-06, 'learning_rate/action_model': 3.456196192304254e-05, 'epoch': 1.36}) +[2026-06-16 03:11:47,767][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.18816886842250824, 'train/grad_norm_pre_clip': 13.721710205078125, 'timing/data': 0.006800403993111104, 'timing/model': 1.4457040280103683, 'learning_rate/qwen_vl_interface': 3.4533766491883887e-06, 'learning_rate/action_model': 3.453376649188388e-05, 'epoch': 1.36}) +[2026-06-16 03:11:53,921][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.05625901743769646, 'train/grad_norm_pre_clip': 6.39069938659668, 'timing/data': 0.007398995978292078, 'timing/model': 1.561002689995803, 'learning_rate/qwen_vl_interface': 3.450557844596517e-06, 'learning_rate/action_model': 3.450557844596517e-05, 'epoch': 1.36}) +[2026-06-16 03:12:00,005][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.16264450550079346, 'train/grad_norm_pre_clip': 10.724031448364258, 'timing/data': 0.0068429470411501825, 'timing/model': 1.5164605480385944, 'learning_rate/qwen_vl_interface': 3.447739779687345e-06, 'learning_rate/action_model': 3.4477397796873446e-05, 'epoch': 1.36}) +[2026-06-16 03:12:06,051][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.076954685151577, 'train/grad_norm_pre_clip': 4.03131628036499, 'timing/data': 0.007230357034131885, 'timing/model': 1.5128161840257235, 'learning_rate/qwen_vl_interface': 3.4449224556192722e-06, 'learning_rate/action_model': 3.444922455619272e-05, 'epoch': 1.36}) +[2026-06-16 03:12:12,162][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.05951086804270744, 'train/grad_norm_pre_clip': 9.933981895446777, 'timing/data': 0.0067913009552285075, 'timing/model': 1.5632779430015944, 'learning_rate/qwen_vl_interface': 3.442105873550395e-06, 'learning_rate/action_model': 3.442105873550395e-05, 'epoch': 1.36}) +[2026-06-16 03:12:18,299][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.09332256019115448, 'train/grad_norm_pre_clip': 16.192138671875, 'timing/data': 0.006726945983245969, 'timing/model': 1.6930613879812881, 'learning_rate/qwen_vl_interface': 3.4392900346385022e-06, 'learning_rate/action_model': 3.4392900346385024e-05, 'epoch': 1.36}) +[2026-06-16 03:12:24,165][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.1038888618350029, 'train/grad_norm_pre_clip': 5.103661060333252, 'timing/data': 0.006604515016078949, 'timing/model': 1.4942037629662082, 'learning_rate/qwen_vl_interface': 3.4364749400410796e-06, 'learning_rate/action_model': 3.43647494004108e-05, 'epoch': 1.36}) +[2026-06-16 03:12:30,270][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.055513713508844376, 'train/grad_norm_pre_clip': 4.167762279510498, 'timing/data': 0.006744240992702544, 'timing/model': 1.4998570299940184, 'learning_rate/qwen_vl_interface': 3.4336605909153066e-06, 'learning_rate/action_model': 3.433660590915306e-05, 'epoch': 1.36}) +[2026-06-16 03:12:36,268][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.030937325209379196, 'train/grad_norm_pre_clip': 3.109442710876465, 'timing/data': 0.006704183004330844, 'timing/model': 1.4487296470324509, 'learning_rate/qwen_vl_interface': 3.430846988418057e-06, 'learning_rate/action_model': 3.430846988418057e-05, 'epoch': 1.36}) +[2026-06-16 03:12:42,285][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.24305459856987, 'train/grad_norm_pre_clip': 9.094293594360352, 'timing/data': 0.006868163007311523, 'timing/model': 1.476681039959658, 'learning_rate/qwen_vl_interface': 3.428034133705893e-06, 'learning_rate/action_model': 3.428034133705893e-05, 'epoch': 1.36}) +[2026-06-16 03:12:48,447][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.03265518322587013, 'train/grad_norm_pre_clip': 2.4353959560394287, 'timing/data': 0.007275192998349667, 'timing/model': 1.588506084983237, 'learning_rate/qwen_vl_interface': 3.4252220279350763e-06, 'learning_rate/action_model': 3.425222027935076e-05, 'epoch': 1.36}) +[2026-06-16 03:12:54,589][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.059442415833473206, 'train/grad_norm_pre_clip': 5.672324180603027, 'timing/data': 0.006560890993569046, 'timing/model': 1.7287496030330658, 'learning_rate/qwen_vl_interface': 3.4224106722615547e-06, 'learning_rate/action_model': 3.422410672261554e-05, 'epoch': 1.36}) +[2026-06-16 03:13:00,366][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.05162954702973366, 'train/grad_norm_pre_clip': 4.400082588195801, 'timing/data': 0.0076900540152564645, 'timing/model': 1.4613773670280352, 'learning_rate/qwen_vl_interface': 3.419600067840973e-06, 'learning_rate/action_model': 3.419600067840973e-05, 'epoch': 1.36}) +[2026-06-16 03:13:06,369][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.1148320883512497, 'train/grad_norm_pre_clip': 12.072563171386719, 'timing/data': 0.00689008500194177, 'timing/model': 1.4653921630233526, 'learning_rate/qwen_vl_interface': 3.4167902158286627e-06, 'learning_rate/action_model': 3.4167902158286626e-05, 'epoch': 1.36}) +[2026-06-16 03:13:12,266][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.0542009212076664, 'train/grad_norm_pre_clip': 4.847143173217773, 'timing/data': 0.00660057901404798, 'timing/model': 1.4539201370207593, 'learning_rate/qwen_vl_interface': 3.4139811173796468e-06, 'learning_rate/action_model': 3.413981117379647e-05, 'epoch': 1.36}) +[2026-06-16 03:13:18,266][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.014281071722507477, 'train/grad_norm_pre_clip': 1.3017617464065552, 'timing/data': 0.006847443000879139, 'timing/model': 1.45716851396719, 'learning_rate/qwen_vl_interface': 3.411172773648642e-06, 'learning_rate/action_model': 3.411172773648642e-05, 'epoch': 1.36}) +[2026-06-16 03:13:24,252][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.020697399973869324, 'train/grad_norm_pre_clip': 7.565295219421387, 'timing/data': 0.0069165549939498305, 'timing/model': 1.5141209780122153, 'learning_rate/qwen_vl_interface': 3.408365185790052e-06, 'learning_rate/action_model': 3.4083651857900515e-05, 'epoch': 1.36}) +[2026-06-16 03:13:30,295][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.04748983681201935, 'train/grad_norm_pre_clip': 4.462156295776367, 'timing/data': 0.006742097029928118, 'timing/model': 1.6869884569896385, 'learning_rate/qwen_vl_interface': 3.40555835495797e-06, 'learning_rate/action_model': 3.40555835495797e-05, 'epoch': 1.36}) +[2026-06-16 03:13:36,074][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.03626300394535065, 'train/grad_norm_pre_clip': 3.348416328430176, 'timing/data': 0.006813341984525323, 'timing/model': 1.5076554989791475, 'learning_rate/qwen_vl_interface': 3.4027522823061763e-06, 'learning_rate/action_model': 3.4027522823061765e-05, 'epoch': 1.37}) +[2026-06-16 03:13:42,064][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.0237627774477005, 'train/grad_norm_pre_clip': 3.767315626144409, 'timing/data': 0.006851718993857503, 'timing/model': 1.4555019320105202, 'learning_rate/qwen_vl_interface': 3.399946968988143e-06, 'learning_rate/action_model': 3.399946968988143e-05, 'epoch': 1.37}) +[2026-06-16 03:13:48,078][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.16957174241542816, 'train/grad_norm_pre_clip': 11.044503211975098, 'timing/data': 0.007016921998001635, 'timing/model': 1.4615329079679213, 'learning_rate/qwen_vl_interface': 3.397142416157028e-06, 'learning_rate/action_model': 3.397142416157028e-05, 'epoch': 1.37}) +[2026-06-16 03:13:53,994][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.10826727747917175, 'train/grad_norm_pre_clip': 9.49484634399414, 'timing/data': 0.006697532022371888, 'timing/model': 1.4866332419915125, 'learning_rate/qwen_vl_interface': 3.3943386249656766e-06, 'learning_rate/action_model': 3.394338624965677e-05, 'epoch': 1.37}) +[2026-06-16 03:13:59,948][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.08081892132759094, 'train/grad_norm_pre_clip': 6.6664042472839355, 'timing/data': 0.00668442202731967, 'timing/model': 1.5102029209956527, 'learning_rate/qwen_vl_interface': 3.391535596566624e-06, 'learning_rate/action_model': 3.391535596566624e-05, 'epoch': 1.37}) +[2026-06-16 03:14:05,993][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.13316205143928528, 'train/grad_norm_pre_clip': 8.798514366149902, 'timing/data': 0.007029951026197523, 'timing/model': 1.7348700489965267, 'learning_rate/qwen_vl_interface': 3.388733332112086e-06, 'learning_rate/action_model': 3.3887333321120856e-05, 'epoch': 1.37}) +[2026-06-16 03:14:11,708][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.024449270218610764, 'train/grad_norm_pre_clip': 3.410306215286255, 'timing/data': 0.006773333007004112, 'timing/model': 1.5012377829989418, 'learning_rate/qwen_vl_interface': 3.3859318327539693e-06, 'learning_rate/action_model': 3.385931832753969e-05, 'epoch': 1.37}) +[2026-06-16 03:14:17,665][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.17265839874744415, 'train/grad_norm_pre_clip': 19.20758819580078, 'timing/data': 0.0065134200267493725, 'timing/model': 1.4575748729985207, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 1.37}) +[2026-06-16 03:14:23,771][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.1256692260503769, 'train/grad_norm_pre_clip': 9.356849670410156, 'timing/data': 0.008900795015506446, 'timing/model': 1.5628694989718497, 'learning_rate/qwen_vl_interface': 3.3803311339330447e-06, 'learning_rate/action_model': 3.3803311339330444e-05, 'epoch': 1.37}) +[2026-06-16 03:14:30,043][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.10188604891300201, 'train/grad_norm_pre_clip': 16.742475509643555, 'timing/data': 0.006937054044101387, 'timing/model': 1.5607658170047216, 'learning_rate/qwen_vl_interface': 3.3775319367724746e-06, 'learning_rate/action_model': 3.3775319367724745e-05, 'epoch': 1.37}) +[2026-06-16 03:14:36,344][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.1965314745903015, 'train/grad_norm_pre_clip': 13.788763999938965, 'timing/data': 0.006913178949616849, 'timing/model': 1.5318418169626966, 'learning_rate/qwen_vl_interface': 3.374733509312795e-06, 'learning_rate/action_model': 3.374733509312795e-05, 'epoch': 1.37}) +[2026-06-16 03:14:42,673][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.012920294888317585, 'train/grad_norm_pre_clip': 1.2117727994918823, 'timing/data': 0.006786321988329291, 'timing/model': 1.7959397859522142, 'learning_rate/qwen_vl_interface': 3.3719358527043345e-06, 'learning_rate/action_model': 3.371935852704334e-05, 'epoch': 1.37}) +[2026-06-16 03:14:48,894][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.006533865351229906, 'train/grad_norm_pre_clip': 1.2328733205795288, 'timing/data': 0.006988878943957388, 'timing/model': 1.709280351002235, 'learning_rate/qwen_vl_interface': 3.3691389680971047e-06, 'learning_rate/action_model': 3.3691389680971044e-05, 'epoch': 1.37}) +[2026-06-16 03:14:54,941][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.057958751916885376, 'train/grad_norm_pre_clip': 9.764688491821289, 'timing/data': 0.006786091020330787, 'timing/model': 1.5546791179804131, 'learning_rate/qwen_vl_interface': 3.3663428566407985e-06, 'learning_rate/action_model': 3.366342856640798e-05, 'epoch': 1.37}) +[2026-06-16 03:15:01,240][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.11058012396097183, 'train/grad_norm_pre_clip': 14.208802223205566, 'timing/data': 0.007300689001567662, 'timing/model': 1.557614887948148, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 1.37}) +[2026-06-16 03:15:07,534][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.005254607647657394, 'train/grad_norm_pre_clip': 0.8849268555641174, 'timing/data': 0.0065929670236073434, 'timing/model': 1.5514229470281862, 'learning_rate/qwen_vl_interface': 3.3607529577781433e-06, 'learning_rate/action_model': 3.360752957778143e-05, 'epoch': 1.37}) +[2026-06-16 03:15:13,820][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.06756958365440369, 'train/grad_norm_pre_clip': 5.30554723739624, 'timing/data': 0.006736838025972247, 'timing/model': 1.5231589309987612, 'learning_rate/qwen_vl_interface': 3.3579591726695906e-06, 'learning_rate/action_model': 3.35795917266959e-05, 'epoch': 1.37}) +[2026-06-16 03:15:20,067][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.14487600326538086, 'train/grad_norm_pre_clip': 24.603809356689453, 'timing/data': 0.006815063010435551, 'timing/model': 1.7839402959798463, 'learning_rate/qwen_vl_interface': 3.3551661653075522e-06, 'learning_rate/action_model': 3.355166165307552e-05, 'epoch': 1.37}) +[2026-06-16 03:15:26,367][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.19734597206115723, 'train/grad_norm_pre_clip': 13.571993827819824, 'timing/data': 0.006574739993084222, 'timing/model': 1.7880447139614262, 'learning_rate/qwen_vl_interface': 3.35237393684013e-06, 'learning_rate/action_model': 3.35237393684013e-05, 'epoch': 1.37}) +[2026-06-16 03:15:32,347][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.011796162463724613, 'train/grad_norm_pre_clip': 2.4966824054718018, 'timing/data': 0.006716939038597047, 'timing/model': 1.5118516570073552, 'learning_rate/qwen_vl_interface': 3.349582488415103e-06, 'learning_rate/action_model': 3.3495824884151026e-05, 'epoch': 1.37}) +[2026-06-16 03:15:38,264][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.1232614815235138, 'train/grad_norm_pre_clip': 13.148597717285156, 'timing/data': 0.006465098005719483, 'timing/model': 1.4556329079787247, 'learning_rate/qwen_vl_interface': 3.3467918211799294e-06, 'learning_rate/action_model': 3.3467918211799295e-05, 'epoch': 1.37}) +[2026-06-16 03:15:44,430][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.010077781975269318, 'train/grad_norm_pre_clip': 1.1219778060913086, 'timing/data': 0.006322438013739884, 'timing/model': 1.5489406560081989, 'learning_rate/qwen_vl_interface': 3.3440019362817476e-06, 'learning_rate/action_model': 3.344001936281747e-05, 'epoch': 1.37}) +[2026-06-16 03:15:50,636][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.21755588054656982, 'train/grad_norm_pre_clip': 14.038508415222168, 'timing/data': 0.006359803024679422, 'timing/model': 1.5317439869977534, 'learning_rate/qwen_vl_interface': 3.3412128348673745e-06, 'learning_rate/action_model': 3.341212834867374e-05, 'epoch': 1.37}) +[2026-06-16 03:15:56,795][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.06179334968328476, 'train/grad_norm_pre_clip': 8.609614372253418, 'timing/data': 0.0067066740011796355, 'timing/model': 1.7024006770225242, 'learning_rate/qwen_vl_interface': 3.338424518083303e-06, 'learning_rate/action_model': 3.338424518083303e-05, 'epoch': 1.38}) +[2026-06-16 03:16:02,530][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.22672826051712036, 'train/grad_norm_pre_clip': 8.748187065124512, 'timing/data': 0.006802543997764587, 'timing/model': 1.5194904239615425, 'learning_rate/qwen_vl_interface': 3.3356369870757056e-06, 'learning_rate/action_model': 3.335636987075705e-05, 'epoch': 1.38}) +[2026-06-16 03:16:08,663][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.009617534466087818, 'train/grad_norm_pre_clip': 1.5629466772079468, 'timing/data': 0.0070891460054554045, 'timing/model': 1.4915798830334097, 'learning_rate/qwen_vl_interface': 3.3328502429904327e-06, 'learning_rate/action_model': 3.3328502429904325e-05, 'epoch': 1.38}) +[2026-06-16 03:16:14,764][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.09056631475687027, 'train/grad_norm_pre_clip': 16.786502838134766, 'timing/data': 0.006877685023937374, 'timing/model': 1.4762069499702193, 'learning_rate/qwen_vl_interface': 3.3300642869730066e-06, 'learning_rate/action_model': 3.330064286973006e-05, 'epoch': 1.38}) +[2026-06-16 03:25:16,048][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.07447361201047897, 'train/grad_norm_pre_clip': 10.38082218170166, 'eval/loss': 0.05728205665946007, 'eval/flap_precision': 0.9302788844621513, 'eval/flap_recall': 0.935871743486974, 'eval/flap_f1': 0.9330669330669331, 'eval/flap_rate_model': 0.13944444444444445, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9330669330669332, 'timing/data': 0.006509453000035137, 'timing/model': 1.5187195250182413, 'learning_rate/qwen_vl_interface': 3.3272791201686294e-06, 'learning_rate/action_model': 3.327279120168629e-05, 'epoch': 1.38}) +[2026-06-16 03:25:16,049][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3200_state +[2026-06-16 03:25:16,053][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 03:25:21,708][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3200_state/model.safetensors +[2026-06-16 03:25:35,241][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3200_state/optimizer.bin +[2026-06-16 03:25:35,241][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3200_state/sampler.bin +[2026-06-16 03:25:35,242][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3200_state/sampler_1.bin +[2026-06-16 03:25:35,243][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3200_state/random_states_0.pkl +[2026-06-16 03:25:35,352][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3200 +[2026-06-16 03:25:35,353][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 03:25:35,362][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 03:25:41,354][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.07304523140192032, 'train/grad_norm_pre_clip': 8.176656723022461, 'timing/data': 0.007146858028136194, 'timing/model': 1.4725321090081707, 'learning_rate/qwen_vl_interface': 3.324494743722179e-06, 'learning_rate/action_model': 3.3244947437221785e-05, 'epoch': 1.38}) +[2026-06-16 03:25:47,430][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.03584706038236618, 'train/grad_norm_pre_clip': 10.805767059326172, 'timing/data': 0.007795383979100734, 'timing/model': 1.494605571962893, 'learning_rate/qwen_vl_interface': 3.321711158778205e-06, 'learning_rate/action_model': 3.321711158778205e-05, 'epoch': 1.38}) +[2026-06-16 03:25:53,379][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.026423422619700432, 'train/grad_norm_pre_clip': 3.9849038124084473, 'timing/data': 0.00661646801745519, 'timing/model': 1.6917161989840679, 'learning_rate/qwen_vl_interface': 3.3189283664809393e-06, 'learning_rate/action_model': 3.318928366480939e-05, 'epoch': 1.38}) +[2026-06-16 03:25:59,060][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.1077251061797142, 'train/grad_norm_pre_clip': 11.96465015411377, 'timing/data': 0.0073560060118325055, 'timing/model': 1.467657123983372, 'learning_rate/qwen_vl_interface': 3.316146367974276e-06, 'learning_rate/action_model': 3.316146367974276e-05, 'epoch': 1.38}) +[2026-06-16 03:26:05,143][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.06318750977516174, 'train/grad_norm_pre_clip': 9.566032409667969, 'timing/data': 0.01110495300963521, 'timing/model': 1.5585203989758156, 'learning_rate/qwen_vl_interface': 3.3133651644017945e-06, 'learning_rate/action_model': 3.3133651644017944e-05, 'epoch': 1.38}) +[2026-06-16 03:26:11,250][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.018027545884251595, 'train/grad_norm_pre_clip': 1.6101309061050415, 'timing/data': 0.006631549971643835, 'timing/model': 1.4807567569660023, 'learning_rate/qwen_vl_interface': 3.31058475690674e-06, 'learning_rate/action_model': 3.31058475690674e-05, 'epoch': 1.38}) +[2026-06-16 03:26:17,054][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.15735302865505219, 'train/grad_norm_pre_clip': 19.810890197753906, 'timing/data': 0.006568336975760758, 'timing/model': 1.4546939550200477, 'learning_rate/qwen_vl_interface': 3.3078051466320337e-06, 'learning_rate/action_model': 3.307805146632034e-05, 'epoch': 1.38}) +[2026-06-16 03:26:23,051][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.008043835870921612, 'train/grad_norm_pre_clip': 0.9329987168312073, 'timing/data': 0.0067124690394848585, 'timing/model': 1.4883238070178777, 'learning_rate/qwen_vl_interface': 3.3050263347202712e-06, 'learning_rate/action_model': 3.305026334720271e-05, 'epoch': 1.38}) +[2026-06-16 03:26:29,054][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.08778633922338486, 'train/grad_norm_pre_clip': 13.697916030883789, 'timing/data': 0.0064514350378885865, 'timing/model': 1.6585625630104914, 'learning_rate/qwen_vl_interface': 3.302248322313712e-06, 'learning_rate/action_model': 3.302248322313712e-05, 'epoch': 1.38}) +[2026-06-16 03:26:34,769][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.02599005587399006, 'train/grad_norm_pre_clip': 6.349453926086426, 'timing/data': 0.0068867450463585556, 'timing/model': 1.4707109680166468, 'learning_rate/qwen_vl_interface': 3.299471110554295e-06, 'learning_rate/action_model': 3.299471110554295e-05, 'epoch': 1.38}) +[2026-06-16 03:26:40,862][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.009866587817668915, 'train/grad_norm_pre_clip': 1.2076865434646606, 'timing/data': 0.007171464036218822, 'timing/model': 1.5238248349633068, 'learning_rate/qwen_vl_interface': 3.296694700583626e-06, 'learning_rate/action_model': 3.296694700583626e-05, 'epoch': 1.38}) +[2026-06-16 03:26:46,871][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.1812341958284378, 'train/grad_norm_pre_clip': 20.758052825927734, 'timing/data': 0.0070007030153647065, 'timing/model': 1.4845841000205837, 'learning_rate/qwen_vl_interface': 3.2939190935429833e-06, 'learning_rate/action_model': 3.293919093542983e-05, 'epoch': 1.38}) +[2026-06-16 03:26:52,855][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.020122407004237175, 'train/grad_norm_pre_clip': 2.9111177921295166, 'timing/data': 0.007054872985463589, 'timing/model': 1.5197362509788945, 'learning_rate/qwen_vl_interface': 3.2911442905733126e-06, 'learning_rate/action_model': 3.291144290573312e-05, 'epoch': 1.38}) +[2026-06-16 03:26:58,771][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.37940576672554016, 'train/grad_norm_pre_clip': 20.35982894897461, 'timing/data': 0.007486917951609939, 'timing/model': 1.4909897709731013, 'learning_rate/qwen_vl_interface': 3.2883702928152323e-06, 'learning_rate/action_model': 3.288370292815232e-05, 'epoch': 1.38}) +[2026-06-16 03:27:04,728][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.05722487345337868, 'train/grad_norm_pre_clip': 20.93061065673828, 'timing/data': 0.006783042976167053, 'timing/model': 1.727830906980671, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 1.38}) +[2026-06-16 03:27:10,477][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.12496406584978104, 'train/grad_norm_pre_clip': 9.26767349243164, 'timing/data': 0.006625590962357819, 'timing/model': 1.476725812011864, 'learning_rate/qwen_vl_interface': 3.2828247174946528e-06, 'learning_rate/action_model': 3.2828247174946524e-05, 'epoch': 1.38}) +[2026-06-16 03:27:16,561][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.14793530106544495, 'train/grad_norm_pre_clip': 14.580249786376953, 'timing/data': 0.006565593008417636, 'timing/model': 1.5033300739596598, 'learning_rate/qwen_vl_interface': 3.280053142211729e-06, 'learning_rate/action_model': 3.280053142211729e-05, 'epoch': 1.38}) +[2026-06-16 03:27:22,561][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.0344800166785717, 'train/grad_norm_pre_clip': 8.662226676940918, 'timing/data': 0.00720756797818467, 'timing/model': 1.454982627008576, 'learning_rate/qwen_vl_interface': 3.2772823766995475e-06, 'learning_rate/action_model': 3.277282376699548e-05, 'epoch': 1.38}) +[2026-06-16 03:27:28,659][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.06166190654039383, 'train/grad_norm_pre_clip': 23.301870346069336, 'timing/data': 0.006556939973961562, 'timing/model': 1.4497681190259755, 'learning_rate/qwen_vl_interface': 3.2745124220970646e-06, 'learning_rate/action_model': 3.274512422097065e-05, 'epoch': 1.39}) +[2026-06-16 03:27:34,561][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.022429604083299637, 'train/grad_norm_pre_clip': 9.570321083068848, 'timing/data': 0.006494018016383052, 'timing/model': 1.4721978369634598, 'learning_rate/qwen_vl_interface': 3.2717432795429057e-06, 'learning_rate/action_model': 3.271743279542905e-05, 'epoch': 1.39}) +[2026-06-16 03:27:40,491][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.015224488452076912, 'train/grad_norm_pre_clip': 4.179454803466797, 'timing/data': 0.006728172011207789, 'timing/model': 1.6889223440084606, 'learning_rate/qwen_vl_interface': 3.2689749501753578e-06, 'learning_rate/action_model': 3.2689749501753574e-05, 'epoch': 1.39}) +[2026-06-16 03:27:46,164][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.0049830237403512, 'train/grad_norm_pre_clip': 1.2843430042266846, 'timing/data': 0.0065190140157938, 'timing/model': 1.4605604140087962, 'learning_rate/qwen_vl_interface': 3.2662074351323774e-06, 'learning_rate/action_model': 3.266207435132378e-05, 'epoch': 1.39}) +[2026-06-16 03:27:52,158][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.2626887857913971, 'train/grad_norm_pre_clip': 27.12296485900879, 'timing/data': 0.006275905994698405, 'timing/model': 1.4622135140234604, 'learning_rate/qwen_vl_interface': 3.2634407355515875e-06, 'learning_rate/action_model': 3.2634407355515875e-05, 'epoch': 1.39}) +[2026-06-16 03:27:58,057][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.2153792679309845, 'train/grad_norm_pre_clip': 19.469911575317383, 'timing/data': 0.0068173829931765795, 'timing/model': 1.4528871760121547, 'learning_rate/qwen_vl_interface': 3.2606748525702715e-06, 'learning_rate/action_model': 3.260674852570271e-05, 'epoch': 1.39}) +[2026-06-16 03:28:03,965][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.11852280050516129, 'train/grad_norm_pre_clip': 8.720419883728027, 'timing/data': 0.007586735999211669, 'timing/model': 1.4378501930041239, 'learning_rate/qwen_vl_interface': 3.2579097873253803e-06, 'learning_rate/action_model': 3.25790978732538e-05, 'epoch': 1.39}) +[2026-06-16 03:28:09,858][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.16123326122760773, 'train/grad_norm_pre_clip': 15.20077133178711, 'timing/data': 0.006436051975470036, 'timing/model': 1.4581365270423703, 'learning_rate/qwen_vl_interface': 3.2551455409535274e-06, 'learning_rate/action_model': 3.255145540953528e-05, 'epoch': 1.39}) +[2026-06-16 03:28:15,794][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.13495810329914093, 'train/grad_norm_pre_clip': 16.572614669799805, 'timing/data': 0.0062630269676446915, 'timing/model': 1.6936365330475383, 'learning_rate/qwen_vl_interface': 3.2523821145909907e-06, 'learning_rate/action_model': 3.252382114590991e-05, 'epoch': 1.39}) +[2026-06-16 03:28:21,461][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.09820462018251419, 'train/grad_norm_pre_clip': 13.071649551391602, 'timing/data': 0.006471875007264316, 'timing/model': 1.4573306550155394, 'learning_rate/qwen_vl_interface': 3.249619509373712e-06, 'learning_rate/action_model': 3.249619509373712e-05, 'epoch': 1.39}) +[2026-06-16 03:28:27,462][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.12699465453624725, 'train/grad_norm_pre_clip': 8.751837730407715, 'timing/data': 0.006624659989029169, 'timing/model': 1.4597203089506365, 'learning_rate/qwen_vl_interface': 3.2468577264372905e-06, 'learning_rate/action_model': 3.24685772643729e-05, 'epoch': 1.39}) +[2026-06-16 03:28:33,361][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.14875979721546173, 'train/grad_norm_pre_clip': 10.841358184814453, 'timing/data': 0.006832435028627515, 'timing/model': 1.4543983140029013, 'learning_rate/qwen_vl_interface': 3.244096766916992e-06, 'learning_rate/action_model': 3.244096766916992e-05, 'epoch': 1.39}) +[2026-06-16 03:28:39,261][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.03546435758471489, 'train/grad_norm_pre_clip': 11.046538352966309, 'timing/data': 0.0068518740008585155, 'timing/model': 1.459126872010529, 'learning_rate/qwen_vl_interface': 3.241336631947743e-06, 'learning_rate/action_model': 3.2413366319477427e-05, 'epoch': 1.39}) +[2026-06-16 03:28:45,246][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.19162757694721222, 'train/grad_norm_pre_clip': 10.210087776184082, 'timing/data': 0.006698368000797927, 'timing/model': 1.5090178140089847, 'learning_rate/qwen_vl_interface': 3.238577322664131e-06, 'learning_rate/action_model': 3.238577322664131e-05, 'epoch': 1.39}) +[2026-06-16 03:28:51,194][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.24843838810920715, 'train/grad_norm_pre_clip': 11.95373249053955, 'timing/data': 0.006927945010829717, 'timing/model': 1.6905852549825795, 'learning_rate/qwen_vl_interface': 3.2358188402004026e-06, 'learning_rate/action_model': 3.2358188402004024e-05, 'epoch': 1.39}) +[2026-06-16 03:28:56,862][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.37687233090400696, 'train/grad_norm_pre_clip': 21.1594181060791, 'timing/data': 0.006659440987277776, 'timing/model': 1.459481005964335, 'learning_rate/qwen_vl_interface': 3.2330611856904675e-06, 'learning_rate/action_model': 3.233061185690468e-05, 'epoch': 1.39}) +[2026-06-16 03:29:02,863][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.031701840460300446, 'train/grad_norm_pre_clip': 5.804576873779297, 'timing/data': 0.006658519967459142, 'timing/model': 1.4605504570063204, 'learning_rate/qwen_vl_interface': 3.2303043602678928e-06, 'learning_rate/action_model': 3.230304360267893e-05, 'epoch': 1.39}) +[2026-06-16 03:29:08,760][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.22504711151123047, 'train/grad_norm_pre_clip': 17.84080696105957, 'timing/data': 0.006691107002552599, 'timing/model': 1.4566531089949422, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 1.39}) +[2026-06-16 03:29:14,763][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.09776832163333893, 'train/grad_norm_pre_clip': 8.111066818237305, 'timing/data': 0.006658900005277246, 'timing/model': 1.44996259000618, 'learning_rate/qwen_vl_interface': 3.2247932012173887e-06, 'learning_rate/action_model': 3.2247932012173884e-05, 'epoch': 1.39}) +[2026-06-16 03:29:20,761][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.01920616254210472, 'train/grad_norm_pre_clip': 4.223708629608154, 'timing/data': 0.00633755698800087, 'timing/model': 1.462506789015606, 'learning_rate/qwen_vl_interface': 3.2220388698548897e-06, 'learning_rate/action_model': 3.22203886985489e-05, 'epoch': 1.39}) +[2026-06-16 03:29:26,694][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.1066071018576622, 'train/grad_norm_pre_clip': 11.415491104125977, 'timing/data': 0.00743570300983265, 'timing/model': 1.6904849709826522, 'learning_rate/qwen_vl_interface': 3.21928537211061e-06, 'learning_rate/action_model': 3.21928537211061e-05, 'epoch': 1.39}) +[2026-06-16 03:29:32,360][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.04842180386185646, 'train/grad_norm_pre_clip': 17.66421890258789, 'timing/data': 0.0067043970339000225, 'timing/model': 1.4542262880131602, 'learning_rate/qwen_vl_interface': 3.216532709116405e-06, 'learning_rate/action_model': 3.216532709116405e-05, 'epoch': 1.39}) +[2026-06-16 03:29:38,361][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.08472396433353424, 'train/grad_norm_pre_clip': 7.599329471588135, 'timing/data': 0.006346540991216898, 'timing/model': 1.4601792969624512, 'learning_rate/qwen_vl_interface': 3.2137808820037925e-06, 'learning_rate/action_model': 3.2137808820037924e-05, 'epoch': 1.39}) +[2026-06-16 03:29:44,420][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.25394076108932495, 'train/grad_norm_pre_clip': 13.716756820678711, 'timing/data': 0.006566914031282067, 'timing/model': 1.5179083039984107, 'learning_rate/qwen_vl_interface': 3.211029891903945e-06, 'learning_rate/action_model': 3.211029891903945e-05, 'epoch': 1.39}) +[2026-06-16 03:29:50,362][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.01573644019663334, 'train/grad_norm_pre_clip': 1.479572057723999, 'timing/data': 0.006872994999866933, 'timing/model': 1.4595900049898773, 'learning_rate/qwen_vl_interface': 3.2082797399476894e-06, 'learning_rate/action_model': 3.2082797399476895e-05, 'epoch': 1.4}) +[2026-06-16 03:29:56,339][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.011031640693545341, 'train/grad_norm_pre_clip': 1.3983924388885498, 'timing/data': 0.006377685989718884, 'timing/model': 1.5079191989498213, 'learning_rate/qwen_vl_interface': 3.2055304272655114e-06, 'learning_rate/action_model': 3.205530427265511e-05, 'epoch': 1.4}) +[2026-06-16 03:30:02,376][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.1288830190896988, 'train/grad_norm_pre_clip': 10.753252029418945, 'timing/data': 0.006610537995584309, 'timing/model': 1.744231063988991, 'learning_rate/qwen_vl_interface': 3.2027819549875477e-06, 'learning_rate/action_model': 3.202781954987548e-05, 'epoch': 1.4}) +[2026-06-16 03:30:08,062][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.11432536691427231, 'train/grad_norm_pre_clip': 9.842716217041016, 'timing/data': 0.006665119028184563, 'timing/model': 1.4580253870226443, 'learning_rate/qwen_vl_interface': 3.2000343242435926e-06, 'learning_rate/action_model': 3.2000343242435924e-05, 'epoch': 1.4}) +[2026-06-16 03:30:14,083][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.005845424719154835, 'train/grad_norm_pre_clip': 0.8473210334777832, 'timing/data': 0.006515489018056542, 'timing/model': 1.5318090420332737, 'learning_rate/qwen_vl_interface': 3.1972875361630923e-06, 'learning_rate/action_model': 3.1972875361630923e-05, 'epoch': 1.4}) +[2026-06-16 03:30:19,960][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.056856073439121246, 'train/grad_norm_pre_clip': 4.03678035736084, 'timing/data': 0.00669129699235782, 'timing/model': 1.4583878759876825, 'learning_rate/qwen_vl_interface': 3.1945415918751484e-06, 'learning_rate/action_model': 3.194541591875148e-05, 'epoch': 1.4}) +[2026-06-16 03:30:25,973][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.15703114867210388, 'train/grad_norm_pre_clip': 9.709720611572266, 'timing/data': 0.0065684959990903735, 'timing/model': 1.471987047989387, 'learning_rate/qwen_vl_interface': 3.191796492508516e-06, 'learning_rate/action_model': 3.191796492508516e-05, 'epoch': 1.4}) +[2026-06-16 03:30:31,937][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.01735408417880535, 'train/grad_norm_pre_clip': 4.3491010665893555, 'timing/data': 0.006787889986298978, 'timing/model': 1.5057193989632651, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 1.4}) +[2026-06-16 03:30:37,884][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.0772867500782013, 'train/grad_norm_pre_clip': 6.094006061553955, 'timing/data': 0.006680332007817924, 'timing/model': 1.681497196026612, 'learning_rate/qwen_vl_interface': 3.1863088330524577e-06, 'learning_rate/action_model': 3.186308833052458e-05, 'epoch': 1.4}) +[2026-06-16 03:30:43,711][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.19532504677772522, 'train/grad_norm_pre_clip': 27.29961395263672, 'timing/data': 0.006786798010580242, 'timing/model': 1.6093843769631349, 'learning_rate/qwen_vl_interface': 3.183566275218803e-06, 'learning_rate/action_model': 3.1835662752188026e-05, 'epoch': 1.4}) +[2026-06-16 03:30:49,361][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.1463490128517151, 'train/grad_norm_pre_clip': 10.121315002441406, 'timing/data': 0.006597058963961899, 'timing/model': 1.4595399019890465, 'learning_rate/qwen_vl_interface': 3.180824566817996e-06, 'learning_rate/action_model': 3.1808245668179956e-05, 'epoch': 1.4}) +[2026-06-16 03:30:55,260][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.07285531610250473, 'train/grad_norm_pre_clip': 9.420151710510254, 'timing/data': 0.0066445890115574, 'timing/model': 1.4581996270571835, 'learning_rate/qwen_vl_interface': 3.1780837089770524e-06, 'learning_rate/action_model': 3.1780837089770524e-05, 'epoch': 1.4}) +[2026-06-16 03:31:01,370][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.037176087498664856, 'train/grad_norm_pre_clip': 8.825193405151367, 'timing/data': 0.006748871994204819, 'timing/model': 1.585080532997381, 'learning_rate/qwen_vl_interface': 3.17534370282263e-06, 'learning_rate/action_model': 3.1753437028226303e-05, 'epoch': 1.4}) +[2026-06-16 03:31:07,497][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.09662551432847977, 'train/grad_norm_pre_clip': 13.160565376281738, 'timing/data': 0.006702914019115269, 'timing/model': 1.5232063919538632, 'learning_rate/qwen_vl_interface': 3.172604549481046e-06, 'learning_rate/action_model': 3.1726045494810455e-05, 'epoch': 1.4}) +[2026-06-16 03:31:13,493][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.09713634848594666, 'train/grad_norm_pre_clip': 14.199605941772461, 'timing/data': 0.00670910399639979, 'timing/model': 1.6898636060068384, 'learning_rate/qwen_vl_interface': 3.169866250078261e-06, 'learning_rate/action_model': 3.169866250078261e-05, 'epoch': 1.4}) +[2026-06-16 03:31:19,404][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.06798044592142105, 'train/grad_norm_pre_clip': 8.054804801940918, 'timing/data': 0.006495730020105839, 'timing/model': 1.701496572990436, 'learning_rate/qwen_vl_interface': 3.167128805739887e-06, 'learning_rate/action_model': 3.1671288057398866e-05, 'epoch': 1.4}) +[2026-06-16 03:31:25,262][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.008336707949638367, 'train/grad_norm_pre_clip': 2.0095434188842773, 'timing/data': 0.006779186020139605, 'timing/model': 1.4899387280456722, 'learning_rate/qwen_vl_interface': 3.1643922175911847e-06, 'learning_rate/action_model': 3.164392217591185e-05, 'epoch': 1.4}) +[2026-06-16 03:31:31,267][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.06302329152822495, 'train/grad_norm_pre_clip': 11.158586502075195, 'timing/data': 0.006613283010665327, 'timing/model': 1.4564472679630853, 'learning_rate/qwen_vl_interface': 3.1616564867570603e-06, 'learning_rate/action_model': 3.16165648675706e-05, 'epoch': 1.4}) +[2026-06-16 03:31:37,263][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.015373947098851204, 'train/grad_norm_pre_clip': 2.2471985816955566, 'timing/data': 0.0065252629574388266, 'timing/model': 1.4509165870258585, 'learning_rate/qwen_vl_interface': 3.1589216143620733e-06, 'learning_rate/action_model': 3.1589216143620734e-05, 'epoch': 1.4}) +[2026-06-16 03:31:43,243][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.05312122777104378, 'train/grad_norm_pre_clip': 11.345602989196777, 'timing/data': 0.006588997028302401, 'timing/model': 1.5089209480211139, 'learning_rate/qwen_vl_interface': 3.1561876015304215e-06, 'learning_rate/action_model': 3.156187601530421e-05, 'epoch': 1.4}) +[2026-06-16 03:31:49,195][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.12228496372699738, 'train/grad_norm_pre_clip': 12.335859298706055, 'timing/data': 0.00741977000143379, 'timing/model': 1.6774256689823233, 'learning_rate/qwen_vl_interface': 3.1534544493859563e-06, 'learning_rate/action_model': 3.1534544493859566e-05, 'epoch': 1.4}) +[2026-06-16 03:31:55,036][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.09704189747571945, 'train/grad_norm_pre_clip': 7.539076805114746, 'timing/data': 0.006631088035646826, 'timing/model': 1.5787516119889915, 'learning_rate/qwen_vl_interface': 3.150722159052174e-06, 'learning_rate/action_model': 3.150722159052174e-05, 'epoch': 1.4}) +[2026-06-16 03:32:01,175][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.1548047810792923, 'train/grad_norm_pre_clip': 8.093564987182617, 'timing/data': 0.006657748017460108, 'timing/model': 1.5367052920046262, 'learning_rate/qwen_vl_interface': 3.147990731652215e-06, 'learning_rate/action_model': 3.147990731652215e-05, 'epoch': 1.4}) +[2026-06-16 03:32:07,162][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.029061730951070786, 'train/grad_norm_pre_clip': 4.784608840942383, 'timing/data': 0.006822941009886563, 'timing/model': 1.486227252986282, 'learning_rate/qwen_vl_interface': 3.1452601683088673e-06, 'learning_rate/action_model': 3.145260168308867e-05, 'epoch': 1.41}) +[2026-06-16 03:32:13,262][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.11381025612354279, 'train/grad_norm_pre_clip': 6.921633243560791, 'timing/data': 0.006910108961164951, 'timing/model': 1.467067585035693, 'learning_rate/qwen_vl_interface': 3.1425304701445608e-06, 'learning_rate/action_model': 3.1425304701445605e-05, 'epoch': 1.41}) +[2026-06-16 03:32:19,280][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.20959380269050598, 'train/grad_norm_pre_clip': 15.156364440917969, 'timing/data': 0.006829791993368417, 'timing/model': 1.5299707560334355, 'learning_rate/qwen_vl_interface': 3.139801638281372e-06, 'learning_rate/action_model': 3.139801638281372e-05, 'epoch': 1.41}) +[2026-06-16 03:32:25,206][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.016310149803757668, 'train/grad_norm_pre_clip': 4.90530252456665, 'timing/data': 0.006642535037826747, 'timing/model': 1.7023026979877613, 'learning_rate/qwen_vl_interface': 3.13707367384102e-06, 'learning_rate/action_model': 3.13707367384102e-05, 'epoch': 1.41}) +[2026-06-16 03:32:31,038][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.0065917521715164185, 'train/grad_norm_pre_clip': 1.099039912223816, 'timing/data': 0.006594935024622828, 'timing/model': 1.533695733989589, 'learning_rate/qwen_vl_interface': 3.134346577944869e-06, 'learning_rate/action_model': 3.134346577944869e-05, 'epoch': 1.41}) +[2026-06-16 03:32:37,061][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.08032910525798798, 'train/grad_norm_pre_clip': 10.896617889404297, 'timing/data': 0.006445626029744744, 'timing/model': 1.4590285979793407, 'learning_rate/qwen_vl_interface': 3.1316203517139243e-06, 'learning_rate/action_model': 3.131620351713924e-05, 'epoch': 1.41}) +[2026-06-16 03:32:43,165][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.20598025619983673, 'train/grad_norm_pre_clip': 13.595603942871094, 'timing/data': 0.00668863405007869, 'timing/model': 1.5172807640046813, 'learning_rate/qwen_vl_interface': 3.128894996268837e-06, 'learning_rate/action_model': 3.128894996268837e-05, 'epoch': 1.41}) +[2026-06-16 03:32:49,244][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.17312511801719666, 'train/grad_norm_pre_clip': 8.28555679321289, 'timing/data': 0.006562627968378365, 'timing/model': 1.5131357019999996, 'learning_rate/qwen_vl_interface': 3.1261705127298934e-06, 'learning_rate/action_model': 3.1261705127298935e-05, 'epoch': 1.41}) +[2026-06-16 03:32:55,369][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.23277594149112701, 'train/grad_norm_pre_clip': 11.971498489379883, 'timing/data': 0.006684897991362959, 'timing/model': 1.5364948379574344, 'learning_rate/qwen_vl_interface': 3.123446902217028e-06, 'learning_rate/action_model': 3.1234469022170283e-05, 'epoch': 1.41}) +[2026-06-16 03:33:01,353][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.02577001042664051, 'train/grad_norm_pre_clip': 2.0897440910339355, 'timing/data': 0.006988824985455722, 'timing/model': 1.6416013049893081, 'learning_rate/qwen_vl_interface': 3.1207241658498145e-06, 'learning_rate/action_model': 3.1207241658498145e-05, 'epoch': 1.41}) +[2026-06-16 03:33:07,262][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.05768129974603653, 'train/grad_norm_pre_clip': 6.189231872558594, 'timing/data': 0.006487677979748696, 'timing/model': 1.4840322579839267, 'learning_rate/qwen_vl_interface': 3.1180023047474667e-06, 'learning_rate/action_model': 3.1180023047474664e-05, 'epoch': 1.41}) +[2026-06-16 03:33:13,263][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.12438558787107468, 'train/grad_norm_pre_clip': 17.028982162475586, 'timing/data': 0.00705987901892513, 'timing/model': 1.4595613430137746, 'learning_rate/qwen_vl_interface': 3.1152813200288386e-06, 'learning_rate/action_model': 3.115281320028838e-05, 'epoch': 1.41}) +[2026-06-16 03:33:19,264][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.12608972191810608, 'train/grad_norm_pre_clip': 10.992278099060059, 'timing/data': 0.00671493198024109, 'timing/model': 1.460459013003856, 'learning_rate/qwen_vl_interface': 3.1125612128124244e-06, 'learning_rate/action_model': 3.112561212812424e-05, 'epoch': 1.41}) +[2026-06-16 03:33:25,346][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.13252300024032593, 'train/grad_norm_pre_clip': 12.641525268554688, 'timing/data': 0.007518094033002853, 'timing/model': 1.5093994050403126, 'learning_rate/qwen_vl_interface': 3.1098419842163573e-06, 'learning_rate/action_model': 3.109841984216357e-05, 'epoch': 1.41}) +[2026-06-16 03:33:31,397][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.02263483963906765, 'train/grad_norm_pre_clip': 1.8949192762374878, 'timing/data': 0.006601655972190201, 'timing/model': 1.6928430880070664, 'learning_rate/qwen_vl_interface': 3.10712363535841e-06, 'learning_rate/action_model': 3.10712363535841e-05, 'epoch': 1.41}) +[2026-06-16 03:33:37,338][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.05663366988301277, 'train/grad_norm_pre_clip': 3.839421033859253, 'timing/data': 0.006953302014153451, 'timing/model': 1.5592135500046425, 'learning_rate/qwen_vl_interface': 3.1044061673559926e-06, 'learning_rate/action_model': 3.1044061673559925e-05, 'epoch': 1.41}) +[2026-06-16 03:33:43,365][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.039050810039043427, 'train/grad_norm_pre_clip': 4.959326267242432, 'timing/data': 0.007337922987062484, 'timing/model': 1.4612063270178623, 'learning_rate/qwen_vl_interface': 3.1016895813261556e-06, 'learning_rate/action_model': 3.101689581326155e-05, 'epoch': 1.41}) +[2026-06-16 03:33:49,252][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.15948104858398438, 'train/grad_norm_pre_clip': 9.830958366394043, 'timing/data': 0.0067494289833121, 'timing/model': 1.5165097240242176, 'learning_rate/qwen_vl_interface': 3.098973878385581e-06, 'learning_rate/action_model': 3.098973878385581e-05, 'epoch': 1.41}) +[2026-06-16 03:33:55,231][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.023613696917891502, 'train/grad_norm_pre_clip': 2.5599236488342285, 'timing/data': 0.006827814970165491, 'timing/model': 1.5199263609829359, 'learning_rate/qwen_vl_interface': 3.096259059650594e-06, 'learning_rate/action_model': 3.096259059650594e-05, 'epoch': 1.41}) +[2026-06-16 03:34:01,162][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.12682153284549713, 'train/grad_norm_pre_clip': 8.669971466064453, 'timing/data': 0.006810449995100498, 'timing/model': 1.4590273260255344, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 1.41}) +[2026-06-16 03:34:07,099][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.16930794715881348, 'train/grad_norm_pre_clip': 7.701037406921387, 'timing/data': 0.0066963229910470545, 'timing/model': 1.6942408570321277, 'learning_rate/qwen_vl_interface': 3.090832079260855e-06, 'learning_rate/action_model': 3.0908320792608544e-05, 'epoch': 1.41}) +[2026-06-16 03:34:13,005][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.053648632019758224, 'train/grad_norm_pre_clip': 15.155092239379883, 'timing/data': 0.007658625021576881, 'timing/model': 1.5230742869898677, 'learning_rate/qwen_vl_interface': 3.088119919836929e-06, 'learning_rate/action_model': 3.088119919836929e-05, 'epoch': 1.41}) +[2026-06-16 03:34:19,028][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.0112972566857934, 'train/grad_norm_pre_clip': 1.4008039236068726, 'timing/data': 0.006647171976510435, 'timing/model': 1.5212580359657295, 'learning_rate/qwen_vl_interface': 3.0854086490802433e-06, 'learning_rate/action_model': 3.085408649080243e-05, 'epoch': 1.41}) +[2026-06-16 03:34:25,064][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.058335039764642715, 'train/grad_norm_pre_clip': 9.361105918884277, 'timing/data': 0.00666327599901706, 'timing/model': 1.470973870018497, 'learning_rate/qwen_vl_interface': 3.082698268105299e-06, 'learning_rate/action_model': 3.082698268105299e-05, 'epoch': 1.42}) +[2026-06-16 03:34:30,968][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.04237612709403038, 'train/grad_norm_pre_clip': 9.796062469482422, 'timing/data': 0.0067788250162266195, 'timing/model': 1.4642447549849749, 'learning_rate/qwen_vl_interface': 3.0799887780262295e-06, 'learning_rate/action_model': 3.0799887780262296e-05, 'epoch': 1.42}) +[2026-06-16 03:34:36,873][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.18638718128204346, 'train/grad_norm_pre_clip': 9.060220718383789, 'timing/data': 0.006628474046010524, 'timing/model': 1.4763209380325861, 'learning_rate/qwen_vl_interface': 3.0772801799568054e-06, 'learning_rate/action_model': 3.0772801799568054e-05, 'epoch': 1.42}) +[2026-06-16 03:34:42,819][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.13439925014972687, 'train/grad_norm_pre_clip': 7.461788177490234, 'timing/data': 0.006532824016176164, 'timing/model': 1.7165570110082626, 'learning_rate/qwen_vl_interface': 3.07457247501043e-06, 'learning_rate/action_model': 3.0745724750104296e-05, 'epoch': 1.42}) +[2026-06-16 03:34:48,562][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.0456472747027874, 'train/grad_norm_pre_clip': 8.049378395080566, 'timing/data': 0.0066592899966053665, 'timing/model': 1.4575466940295883, 'learning_rate/qwen_vl_interface': 3.0718656643001353e-06, 'learning_rate/action_model': 3.071865664300135e-05, 'epoch': 1.42}) +[2026-06-16 03:34:54,563][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.17707069218158722, 'train/grad_norm_pre_clip': 9.854987144470215, 'timing/data': 0.006660952989477664, 'timing/model': 1.4461177989724092, 'learning_rate/qwen_vl_interface': 3.069159748938591e-06, 'learning_rate/action_model': 3.069159748938591e-05, 'epoch': 1.42}) +[2026-06-16 03:35:00,463][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.08319012075662613, 'train/grad_norm_pre_clip': 8.59915828704834, 'timing/data': 0.0068784430040977895, 'timing/model': 1.453056342958007, 'learning_rate/qwen_vl_interface': 3.066454730038096e-06, 'learning_rate/action_model': 3.0664547300380956e-05, 'epoch': 1.42}) +[2026-06-16 03:35:06,439][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.09672828018665314, 'train/grad_norm_pre_clip': 16.666019439697266, 'timing/data': 0.006857372005470097, 'timing/model': 1.5078784080105834, 'learning_rate/qwen_vl_interface': 3.0637506087105816e-06, 'learning_rate/action_model': 3.0637506087105815e-05, 'epoch': 1.42}) +[2026-06-16 03:35:12,359][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.1334851086139679, 'train/grad_norm_pre_clip': 6.2727508544921875, 'timing/data': 0.006731697008945048, 'timing/model': 1.4553441439638846, 'learning_rate/qwen_vl_interface': 3.0610473860676103e-06, 'learning_rate/action_model': 3.06104738606761e-05, 'epoch': 1.42}) +[2026-06-16 03:35:18,297][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.05984417721629143, 'train/grad_norm_pre_clip': 4.625506401062012, 'timing/data': 0.006719940982293338, 'timing/model': 1.6914822779945098, 'learning_rate/qwen_vl_interface': 3.0583450632203748e-06, 'learning_rate/action_model': 3.058345063220375e-05, 'epoch': 1.42}) +[2026-06-16 03:35:23,963][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.06773823499679565, 'train/grad_norm_pre_clip': 6.166794300079346, 'timing/data': 0.006368425965774804, 'timing/model': 1.4584007970406674, 'learning_rate/qwen_vl_interface': 3.055643641279699e-06, 'learning_rate/action_model': 3.055643641279699e-05, 'epoch': 1.42}) +[2026-06-16 03:44:25,152][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.04223504289984703, 'train/grad_norm_pre_clip': 6.84653377532959, 'eval/loss': 0.05040019005537033, 'eval/flap_precision': 0.9391480730223124, 'eval/flap_recall': 0.9278557114228457, 'eval/flap_f1': 0.933467741935484, 'eval/flap_rate_model': 0.13694444444444445, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.933467741935484, 'timing/data': 0.0067370059550739825, 'timing/model': 1.459227160026785, 'learning_rate/qwen_vl_interface': 3.0529431213560334e-06, 'learning_rate/action_model': 3.0529431213560334e-05, 'epoch': 1.42}) +[2026-06-16 03:44:25,152][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3300_state +[2026-06-16 03:44:25,156][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 03:44:30,807][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3300_state/model.safetensors +[2026-06-16 03:44:44,373][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3300_state/optimizer.bin +[2026-06-16 03:44:44,374][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3300_state/sampler.bin +[2026-06-16 03:44:44,374][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3300_state/sampler_1.bin +[2026-06-16 03:44:44,376][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3300_state/random_states_0.pkl +[2026-06-16 03:44:44,486][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3300 +[2026-06-16 03:44:44,487][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 03:44:44,497][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 03:44:50,141][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.20244118571281433, 'train/grad_norm_pre_clip': 17.07724380493164, 'timing/data': 0.006818632013164461, 'timing/model': 1.4581924789818004, 'learning_rate/qwen_vl_interface': 3.0502435045594596e-06, 'learning_rate/action_model': 3.0502435045594595e-05, 'epoch': 1.42}) +[2026-06-16 03:44:56,243][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.0526561439037323, 'train/grad_norm_pre_clip': 5.709694862365723, 'timing/data': 0.0073758510407060385, 'timing/model': 1.478399190993514, 'learning_rate/qwen_vl_interface': 3.047544791999689e-06, 'learning_rate/action_model': 3.0475447919996892e-05, 'epoch': 1.42}) +[2026-06-16 03:45:02,047][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.05520210787653923, 'train/grad_norm_pre_clip': 3.8361105918884277, 'timing/data': 0.006870539044030011, 'timing/model': 1.45909425400896, 'learning_rate/qwen_vl_interface': 3.044846984786061e-06, 'learning_rate/action_model': 3.044846984786061e-05, 'epoch': 1.42}) +[2026-06-16 03:45:08,039][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.12335222214460373, 'train/grad_norm_pre_clip': 7.755761623382568, 'timing/data': 0.006371295021381229, 'timing/model': 1.5484114299761131, 'learning_rate/qwen_vl_interface': 3.0421500840275413e-06, 'learning_rate/action_model': 3.0421500840275413e-05, 'epoch': 1.42}) +[2026-06-16 03:45:14,228][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.010648931376636028, 'train/grad_norm_pre_clip': 1.2500319480895996, 'timing/data': 0.007092684973031282, 'timing/model': 1.5723855599644594, 'learning_rate/qwen_vl_interface': 3.039454090832727e-06, 'learning_rate/action_model': 3.0394540908327268e-05, 'epoch': 1.42}) +[2026-06-16 03:45:20,185][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.18188297748565674, 'train/grad_norm_pre_clip': 12.845720291137695, 'timing/data': 0.006423971964977682, 'timing/model': 1.6910001739743166, 'learning_rate/qwen_vl_interface': 3.0367590063098333e-06, 'learning_rate/action_model': 3.0367590063098333e-05, 'epoch': 1.42}) +[2026-06-16 03:45:25,854][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.03116619773209095, 'train/grad_norm_pre_clip': 6.252157688140869, 'timing/data': 0.006883478024974465, 'timing/model': 1.4615289400098845, 'learning_rate/qwen_vl_interface': 3.0340648315667115e-06, 'learning_rate/action_model': 3.0340648315667113e-05, 'epoch': 1.42}) +[2026-06-16 03:45:31,854][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.01865493506193161, 'train/grad_norm_pre_clip': 4.6282806396484375, 'timing/data': 0.006445734994485974, 'timing/model': 1.4585793529986404, 'learning_rate/qwen_vl_interface': 3.0313715677108336e-06, 'learning_rate/action_model': 3.0313715677108335e-05, 'epoch': 1.42}) +[2026-06-16 03:45:37,839][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.03320055082440376, 'train/grad_norm_pre_clip': 6.478625774383545, 'timing/data': 0.006653289950918406, 'timing/model': 1.5426565099623986, 'learning_rate/qwen_vl_interface': 3.0286792158492996e-06, 'learning_rate/action_model': 3.0286792158492994e-05, 'epoch': 1.42}) +[2026-06-16 03:45:43,756][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.20222246646881104, 'train/grad_norm_pre_clip': 10.202688217163086, 'timing/data': 0.006744864978827536, 'timing/model': 1.4554050830192864, 'learning_rate/qwen_vl_interface': 3.0259877770888334e-06, 'learning_rate/action_model': 3.0259877770888333e-05, 'epoch': 1.42}) +[2026-06-16 03:45:49,739][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.0967182144522667, 'train/grad_norm_pre_clip': 20.50813865661621, 'timing/data': 0.006617637001909316, 'timing/model': 1.507987154996954, 'learning_rate/qwen_vl_interface': 3.0232972525357836e-06, 'learning_rate/action_model': 3.0232972525357834e-05, 'epoch': 1.42}) +[2026-06-16 03:45:55,716][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.012223332189023495, 'train/grad_norm_pre_clip': 5.381176471710205, 'timing/data': 0.006848106975667179, 'timing/model': 1.7120468490174972, 'learning_rate/qwen_vl_interface': 3.020607643296123e-06, 'learning_rate/action_model': 3.0206076432961226e-05, 'epoch': 1.43}) +[2026-06-16 03:46:01,746][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.073989637196064, 'train/grad_norm_pre_clip': 7.833427429199219, 'timing/data': 0.006858312000986189, 'timing/model': 1.7421124760294333, 'learning_rate/qwen_vl_interface': 3.017918950475449e-06, 'learning_rate/action_model': 3.017918950475449e-05, 'epoch': 1.43}) +[2026-06-16 03:46:07,542][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.19049601256847382, 'train/grad_norm_pre_clip': 17.356067657470703, 'timing/data': 0.006988602981436998, 'timing/model': 1.5118521410040557, 'learning_rate/qwen_vl_interface': 3.0152311751789815e-06, 'learning_rate/action_model': 3.0152311751789814e-05, 'epoch': 1.43}) +[2026-06-16 03:46:13,560][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.25789108872413635, 'train/grad_norm_pre_clip': 12.210332870483398, 'timing/data': 0.0067396280355751514, 'timing/model': 1.4844070640392601, 'learning_rate/qwen_vl_interface': 3.0125443185115655e-06, 'learning_rate/action_model': 3.0125443185115655e-05, 'epoch': 1.43}) +[2026-06-16 03:46:19,559][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.09299994260072708, 'train/grad_norm_pre_clip': 11.979181289672852, 'timing/data': 0.006636215024627745, 'timing/model': 1.4690001819981262, 'learning_rate/qwen_vl_interface': 3.0098583815776615e-06, 'learning_rate/action_model': 3.0098583815776615e-05, 'epoch': 1.43}) +[2026-06-16 03:46:25,560][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.019962234422564507, 'train/grad_norm_pre_clip': 2.6791746616363525, 'timing/data': 0.007359288982115686, 'timing/model': 1.46873564098496, 'learning_rate/qwen_vl_interface': 3.0071733654813595e-06, 'learning_rate/action_model': 3.0071733654813595e-05, 'epoch': 1.43}) +[2026-06-16 03:46:31,504][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.12404219061136246, 'train/grad_norm_pre_clip': 12.345193862915039, 'timing/data': 0.006674472009763122, 'timing/model': 1.684699236007873, 'learning_rate/qwen_vl_interface': 3.0044892713263685e-06, 'learning_rate/action_model': 3.0044892713263682e-05, 'epoch': 1.43}) +[2026-06-16 03:46:37,256][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.0030420690309256315, 'train/grad_norm_pre_clip': 0.667408287525177, 'timing/data': 0.006495207955595106, 'timing/model': 1.458311236987356, 'learning_rate/qwen_vl_interface': 3.001806100216017e-06, 'learning_rate/action_model': 3.001806100216017e-05, 'epoch': 1.43}) +[2026-06-16 03:46:43,260][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.029113225638866425, 'train/grad_norm_pre_clip': 7.279674053192139, 'timing/data': 0.006906263006385416, 'timing/model': 1.4582918450469151, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 1.43}) +[2026-06-16 03:46:49,260][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.06465569883584976, 'train/grad_norm_pre_clip': 10.359055519104004, 'timing/data': 0.006831371982116252, 'timing/model': 1.4583833960350603, 'learning_rate/qwen_vl_interface': 2.996442531540657e-06, 'learning_rate/action_model': 2.9964425315406568e-05, 'epoch': 1.43}) +[2026-06-16 03:46:55,245][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.19845519959926605, 'train/grad_norm_pre_clip': 11.827136993408203, 'timing/data': 0.006489741033874452, 'timing/model': 1.5101596679887734, 'learning_rate/qwen_vl_interface': 2.993762136180408e-06, 'learning_rate/action_model': 2.993762136180408e-05, 'epoch': 1.43}) +[2026-06-16 03:47:01,253][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.13717025518417358, 'train/grad_norm_pre_clip': 11.414643287658691, 'timing/data': 0.006553273997269571, 'timing/model': 1.5094141979934648, 'learning_rate/qwen_vl_interface': 2.9910826682743195e-06, 'learning_rate/action_model': 2.9910826682743192e-05, 'epoch': 1.43}) +[2026-06-16 03:47:07,340][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.08972115069627762, 'train/grad_norm_pre_clip': 8.140691757202148, 'timing/data': 0.006606042035855353, 'timing/model': 1.6856533070094883, 'learning_rate/qwen_vl_interface': 2.9884041289238192e-06, 'learning_rate/action_model': 2.988404128923819e-05, 'epoch': 1.43}) +[2026-06-16 03:47:13,062][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.04199032858014107, 'train/grad_norm_pre_clip': 7.085377216339111, 'timing/data': 0.0067080919980071485, 'timing/model': 1.4578620420070365, 'learning_rate/qwen_vl_interface': 2.9857265192299557e-06, 'learning_rate/action_model': 2.9857265192299556e-05, 'epoch': 1.43}) +[2026-06-16 03:47:19,077][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.09338092058897018, 'train/grad_norm_pre_clip': 5.875355243682861, 'timing/data': 0.00658973801182583, 'timing/model': 1.4760716970195062, 'learning_rate/qwen_vl_interface': 2.9830498402933894e-06, 'learning_rate/action_model': 2.9830498402933894e-05, 'epoch': 1.43}) +[2026-06-16 03:47:24,961][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.083549864590168, 'train/grad_norm_pre_clip': 11.072315216064453, 'timing/data': 0.006372287985868752, 'timing/model': 1.4565513139823452, 'learning_rate/qwen_vl_interface': 2.9803740932144025e-06, 'learning_rate/action_model': 2.9803740932144024e-05, 'epoch': 1.43}) +[2026-06-16 03:47:31,054][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.09566814452409744, 'train/grad_norm_pre_clip': 7.849071025848389, 'timing/data': 0.006433898990508169, 'timing/model': 1.5146754009765573, 'learning_rate/qwen_vl_interface': 2.9776992790928953e-06, 'learning_rate/action_model': 2.9776992790928952e-05, 'epoch': 1.43}) +[2026-06-16 03:47:37,162][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.11037007719278336, 'train/grad_norm_pre_clip': 8.521808624267578, 'timing/data': 0.0072624769527465105, 'timing/model': 1.474761895951815, 'learning_rate/qwen_vl_interface': 2.975025399028382e-06, 'learning_rate/action_model': 2.975025399028382e-05, 'epoch': 1.43}) +[2026-06-16 03:47:43,054][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.16149671375751495, 'train/grad_norm_pre_clip': 24.210922241210938, 'timing/data': 0.006495059002190828, 'timing/model': 1.6489052940160036, 'learning_rate/qwen_vl_interface': 2.9723524541199943e-06, 'learning_rate/action_model': 2.972352454119994e-05, 'epoch': 1.43}) +[2026-06-16 03:47:48,863][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.07742372900247574, 'train/grad_norm_pre_clip': 6.843290328979492, 'timing/data': 0.006808409001678228, 'timing/model': 1.4585029410081916, 'learning_rate/qwen_vl_interface': 2.969680445466478e-06, 'learning_rate/action_model': 2.969680445466478e-05, 'epoch': 1.43}) +[2026-06-16 03:47:54,864][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.04996754601597786, 'train/grad_norm_pre_clip': 10.932445526123047, 'timing/data': 0.006713189999572933, 'timing/model': 1.4600081779644825, 'learning_rate/qwen_vl_interface': 2.9670093741661955e-06, 'learning_rate/action_model': 2.9670093741661954e-05, 'epoch': 1.43}) +[2026-06-16 03:48:00,860][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.015454857610166073, 'train/grad_norm_pre_clip': 3.03242826461792, 'timing/data': 0.006477223010733724, 'timing/model': 1.4783936929889023, 'learning_rate/qwen_vl_interface': 2.964339241317123e-06, 'learning_rate/action_model': 2.964339241317123e-05, 'epoch': 1.43}) +[2026-06-16 03:48:06,965][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.1201944574713707, 'train/grad_norm_pre_clip': 12.802066802978516, 'timing/data': 0.006649035960435867, 'timing/model': 1.437037703988608, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 1.43}) +[2026-06-16 03:48:12,890][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.1446615606546402, 'train/grad_norm_pre_clip': 12.7990140914917, 'timing/data': 0.006890048971399665, 'timing/model': 1.6828415360068902, 'learning_rate/qwen_vl_interface': 2.9590017953625854e-06, 'learning_rate/action_model': 2.9590017953625853e-05, 'epoch': 1.43}) +[2026-06-16 03:48:18,560][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.09159193187952042, 'train/grad_norm_pre_clip': 11.485871315002441, 'timing/data': 0.0071088719996623695, 'timing/model': 1.4546726280241273, 'learning_rate/qwen_vl_interface': 2.9563344844511455e-06, 'learning_rate/action_model': 2.9563344844511455e-05, 'epoch': 1.44}) +[2026-06-16 03:48:24,562][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.14239417016506195, 'train/grad_norm_pre_clip': 23.53750991821289, 'timing/data': 0.006496721995063126, 'timing/model': 1.4596869440283626, 'learning_rate/qwen_vl_interface': 2.9536681163789568e-06, 'learning_rate/action_model': 2.9536681163789564e-05, 'epoch': 1.44}) +[2026-06-16 03:48:30,557][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.02832520753145218, 'train/grad_norm_pre_clip': 8.873008728027344, 'timing/data': 0.006790974002797157, 'timing/model': 1.5132510989787988, 'learning_rate/qwen_vl_interface': 2.9510026922420653e-06, 'learning_rate/action_model': 2.9510026922420653e-05, 'epoch': 1.44}) +[2026-06-16 03:48:36,464][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.010171720758080482, 'train/grad_norm_pre_clip': 1.4051406383514404, 'timing/data': 0.00664793502073735, 'timing/model': 1.4623438359703869, 'learning_rate/qwen_vl_interface': 2.9483382131361273e-06, 'learning_rate/action_model': 2.948338213136127e-05, 'epoch': 1.44}) +[2026-06-16 03:48:42,362][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.40131068229675293, 'train/grad_norm_pre_clip': 14.96164321899414, 'timing/data': 0.006827217992395163, 'timing/model': 1.4749283609562553, 'learning_rate/qwen_vl_interface': 2.9456746801564072e-06, 'learning_rate/action_model': 2.945674680156407e-05, 'epoch': 1.44}) +[2026-06-16 03:48:48,300][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.04754802584648132, 'train/grad_norm_pre_clip': 5.585269927978516, 'timing/data': 0.006514367996715009, 'timing/model': 1.6938311720150523, 'learning_rate/qwen_vl_interface': 2.9430120943977854e-06, 'learning_rate/action_model': 2.9430120943977853e-05, 'epoch': 1.44}) +[2026-06-16 03:48:53,964][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.03318101912736893, 'train/grad_norm_pre_clip': 3.8704721927642822, 'timing/data': 0.006941125961020589, 'timing/model': 1.457733383984305, 'learning_rate/qwen_vl_interface': 2.940350456954748e-06, 'learning_rate/action_model': 2.940350456954748e-05, 'epoch': 1.44}) +[2026-06-16 03:49:00,062][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.11063936352729797, 'train/grad_norm_pre_clip': 8.18016529083252, 'timing/data': 0.006776394031476229, 'timing/model': 1.4936474810237996, 'learning_rate/qwen_vl_interface': 2.9376897689213952e-06, 'learning_rate/action_model': 2.937689768921395e-05, 'epoch': 1.44}) +[2026-06-16 03:49:06,063][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.22511593997478485, 'train/grad_norm_pre_clip': 20.20901870727539, 'timing/data': 0.006439758988562971, 'timing/model': 1.4934103649575263, 'learning_rate/qwen_vl_interface': 2.9350300313914364e-06, 'learning_rate/action_model': 2.9350300313914363e-05, 'epoch': 1.44}) +[2026-06-16 03:49:11,963][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.014867338351905346, 'train/grad_norm_pre_clip': 1.8272289037704468, 'timing/data': 0.006718778982758522, 'timing/model': 1.4594484330154955, 'learning_rate/qwen_vl_interface': 2.9323712454581867e-06, 'learning_rate/action_model': 2.9323712454581866e-05, 'epoch': 1.44}) +[2026-06-16 03:49:18,046][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.03849714994430542, 'train/grad_norm_pre_clip': 7.062516212463379, 'timing/data': 0.006703696970362216, 'timing/model': 1.5148709139903076, 'learning_rate/qwen_vl_interface': 2.929713412214572e-06, 'learning_rate/action_model': 2.929713412214572e-05, 'epoch': 1.44}) +[2026-06-16 03:49:23,996][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.37603816390037537, 'train/grad_norm_pre_clip': 19.05350685119629, 'timing/data': 0.006793948996346444, 'timing/model': 1.6909444630146027, 'learning_rate/qwen_vl_interface': 2.9270565327531294e-06, 'learning_rate/action_model': 2.927056532753129e-05, 'epoch': 1.44}) +[2026-06-16 03:49:29,663][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.018836339935660362, 'train/grad_norm_pre_clip': 3.269070625305176, 'timing/data': 0.007261286955326796, 'timing/model': 1.4580126469954848, 'learning_rate/qwen_vl_interface': 2.924400608166003e-06, 'learning_rate/action_model': 2.924400608166003e-05, 'epoch': 1.44}) +[2026-06-16 03:49:35,741][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.19725154340267181, 'train/grad_norm_pre_clip': 9.076271057128906, 'timing/data': 0.006735773000400513, 'timing/model': 1.5140065180021338, 'learning_rate/qwen_vl_interface': 2.921745639544939e-06, 'learning_rate/action_model': 2.921745639544939e-05, 'epoch': 1.44}) +[2026-06-16 03:49:41,863][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.025262143462896347, 'train/grad_norm_pre_clip': 4.268805980682373, 'timing/data': 0.006916498998180032, 'timing/model': 1.47818911599461, 'learning_rate/qwen_vl_interface': 2.919091627981297e-06, 'learning_rate/action_model': 2.919091627981297e-05, 'epoch': 1.44}) +[2026-06-16 03:49:47,848][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.014957762323319912, 'train/grad_norm_pre_clip': 2.078552007675171, 'timing/data': 0.007417747052386403, 'timing/model': 1.5137236059526913, 'learning_rate/qwen_vl_interface': 2.9164385745660396e-06, 'learning_rate/action_model': 2.9164385745660394e-05, 'epoch': 1.44}) +[2026-06-16 03:49:53,763][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.18620523810386658, 'train/grad_norm_pre_clip': 18.708599090576172, 'timing/data': 0.006983378028962761, 'timing/model': 1.4575909519917332, 'learning_rate/qwen_vl_interface': 2.9137864803897385e-06, 'learning_rate/action_model': 2.9137864803897387e-05, 'epoch': 1.44}) +[2026-06-16 03:49:59,697][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.05523161217570305, 'train/grad_norm_pre_clip': 6.477005481719971, 'timing/data': 0.006663388048764318, 'timing/model': 1.6911758710048161, 'learning_rate/qwen_vl_interface': 2.9111353465425696e-06, 'learning_rate/action_model': 2.9111353465425694e-05, 'epoch': 1.44}) +[2026-06-16 03:50:05,363][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.0922437310218811, 'train/grad_norm_pre_clip': 6.782427787780762, 'timing/data': 0.0071483299834653735, 'timing/model': 1.4561792369931936, 'learning_rate/qwen_vl_interface': 2.908485174114312e-06, 'learning_rate/action_model': 2.908485174114312e-05, 'epoch': 1.44}) +[2026-06-16 03:50:11,464][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.14624299108982086, 'train/grad_norm_pre_clip': 9.697919845581055, 'timing/data': 0.006963649997487664, 'timing/model': 1.4596398569992743, 'learning_rate/qwen_vl_interface': 2.9058359641943523e-06, 'learning_rate/action_model': 2.9058359641943523e-05, 'epoch': 1.44}) +[2026-06-16 03:50:17,383][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.0007279214332811534, 'train/grad_norm_pre_clip': 0.1186286136507988, 'timing/data': 0.006651661009527743, 'timing/model': 1.4683415430481546, 'learning_rate/qwen_vl_interface': 2.9031877178716823e-06, 'learning_rate/action_model': 2.903187717871682e-05, 'epoch': 1.44}) +[2026-06-16 03:50:23,264][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.1994297057390213, 'train/grad_norm_pre_clip': 14.427488327026367, 'timing/data': 0.006001013971399516, 'timing/model': 1.4666187500115484, 'learning_rate/qwen_vl_interface': 2.900540436234894e-06, 'learning_rate/action_model': 2.9005404362348937e-05, 'epoch': 1.44}) +[2026-06-16 03:50:29,248][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.15909722447395325, 'train/grad_norm_pre_clip': 16.79151725769043, 'timing/data': 0.006692470982670784, 'timing/model': 1.5148472560103983, 'learning_rate/qwen_vl_interface': 2.897894120372185e-06, 'learning_rate/action_model': 2.897894120372185e-05, 'epoch': 1.44}) +[2026-06-16 03:50:35,220][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.08986221253871918, 'train/grad_norm_pre_clip': 5.031527996063232, 'timing/data': 0.006663097999989986, 'timing/model': 1.717329590988811, 'learning_rate/qwen_vl_interface': 2.8952487713713585e-06, 'learning_rate/action_model': 2.8952487713713582e-05, 'epoch': 1.45}) +[2026-06-16 03:50:40,980][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.10392044484615326, 'train/grad_norm_pre_clip': 13.9241304397583, 'timing/data': 0.0068458570167422295, 'timing/model': 1.4771977860364132, 'learning_rate/qwen_vl_interface': 2.8926043903198135e-06, 'learning_rate/action_model': 2.8926043903198135e-05, 'epoch': 1.45}) +[2026-06-16 03:50:46,863][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.02309795655310154, 'train/grad_norm_pre_clip': 2.9079220294952393, 'timing/data': 0.007223711989354342, 'timing/model': 1.4568441270384938, 'learning_rate/qwen_vl_interface': 2.8899609783045574e-06, 'learning_rate/action_model': 2.8899609783045573e-05, 'epoch': 1.45}) +[2026-06-16 03:50:52,834][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.0694977343082428, 'train/grad_norm_pre_clip': 10.580205917358398, 'timing/data': 0.00719797401688993, 'timing/model': 1.5313757020048797, 'learning_rate/qwen_vl_interface': 2.8873185364121965e-06, 'learning_rate/action_model': 2.8873185364121963e-05, 'epoch': 1.45}) +[2026-06-16 03:50:58,771][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.23353363573551178, 'train/grad_norm_pre_clip': 11.767973899841309, 'timing/data': 0.0066740140318870544, 'timing/model': 1.4674283299827948, 'learning_rate/qwen_vl_interface': 2.8846770657289386e-06, 'learning_rate/action_model': 2.8846770657289386e-05, 'epoch': 1.45}) +[2026-06-16 03:51:04,764][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.19501514732837677, 'train/grad_norm_pre_clip': 17.208251953125, 'timing/data': 0.006775262008886784, 'timing/model': 1.5258903760113753, 'learning_rate/qwen_vl_interface': 2.882036567340593e-06, 'learning_rate/action_model': 2.8820365673405932e-05, 'epoch': 1.45}) +[2026-06-16 03:51:10,744][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.09860166907310486, 'train/grad_norm_pre_clip': 8.039061546325684, 'timing/data': 0.006770945969037712, 'timing/model': 1.7403785820351914, 'learning_rate/qwen_vl_interface': 2.8793970423325684e-06, 'learning_rate/action_model': 2.8793970423325685e-05, 'epoch': 1.45}) +[2026-06-16 03:51:16,462][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.09570077061653137, 'train/grad_norm_pre_clip': 5.785646438598633, 'timing/data': 0.006723145954310894, 'timing/model': 1.446262304962147, 'learning_rate/qwen_vl_interface': 2.8767584917898756e-06, 'learning_rate/action_model': 2.8767584917898754e-05, 'epoch': 1.45}) +[2026-06-16 03:51:22,364][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.07629542052745819, 'train/grad_norm_pre_clip': 8.123119354248047, 'timing/data': 0.006582049012649804, 'timing/model': 1.4467606670223176, 'learning_rate/qwen_vl_interface': 2.874120916797121e-06, 'learning_rate/action_model': 2.8741209167971207e-05, 'epoch': 1.45}) +[2026-06-16 03:51:28,362][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.0639515072107315, 'train/grad_norm_pre_clip': 6.235346794128418, 'timing/data': 0.006791016960050911, 'timing/model': 1.4472596400300972, 'learning_rate/qwen_vl_interface': 2.8714843184385124e-06, 'learning_rate/action_model': 2.871484318438512e-05, 'epoch': 1.45}) +[2026-06-16 03:51:34,262][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.02213146537542343, 'train/grad_norm_pre_clip': 3.0054266452789307, 'timing/data': 0.006761012016795576, 'timing/model': 1.4455172659945674, 'learning_rate/qwen_vl_interface': 2.8688486977978575e-06, 'learning_rate/action_model': 2.8688486977978572e-05, 'epoch': 1.45}) +[2026-06-16 03:51:40,264][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.20459982752799988, 'train/grad_norm_pre_clip': 10.771275520324707, 'timing/data': 0.006714363000355661, 'timing/model': 1.452461597044021, 'learning_rate/qwen_vl_interface': 2.8662140559585584e-06, 'learning_rate/action_model': 2.866214055958558e-05, 'epoch': 1.45}) +[2026-06-16 03:51:46,193][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.009770341217517853, 'train/grad_norm_pre_clip': 1.2008881568908691, 'timing/data': 0.006616880011279136, 'timing/model': 1.6906824489706196, 'learning_rate/qwen_vl_interface': 2.8635803940036162e-06, 'learning_rate/action_model': 2.863580394003616e-05, 'epoch': 1.45}) +[2026-06-16 03:51:51,865][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.07140768319368362, 'train/grad_norm_pre_clip': 11.04702091217041, 'timing/data': 0.006697789009194821, 'timing/model': 1.461277608992532, 'learning_rate/qwen_vl_interface': 2.8609477130156317e-06, 'learning_rate/action_model': 2.8609477130156314e-05, 'epoch': 1.45}) +[2026-06-16 03:51:57,866][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.0901130735874176, 'train/grad_norm_pre_clip': 8.50485897064209, 'timing/data': 0.006519276998005807, 'timing/model': 1.460551790019963, 'learning_rate/qwen_vl_interface': 2.8583160140767997e-06, 'learning_rate/action_model': 2.8583160140767994e-05, 'epoch': 1.45}) +[2026-06-16 03:52:03,765][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.31919410824775696, 'train/grad_norm_pre_clip': 17.366823196411133, 'timing/data': 0.006922399974428117, 'timing/model': 1.4574350949842483, 'learning_rate/qwen_vl_interface': 2.8556852982689125e-06, 'learning_rate/action_model': 2.8556852982689126e-05, 'epoch': 1.45}) +[2026-06-16 03:52:09,862][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.09439792484045029, 'train/grad_norm_pre_clip': 8.22346305847168, 'timing/data': 0.006946654990315437, 'timing/model': 1.45981309103081, 'learning_rate/qwen_vl_interface': 2.8530555666733584e-06, 'learning_rate/action_model': 2.8530555666733583e-05, 'epoch': 1.45}) +[2026-06-16 03:52:15,863][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.11188612133264542, 'train/grad_norm_pre_clip': 9.2372465133667, 'timing/data': 0.006554519000928849, 'timing/model': 1.5093356929719448, 'learning_rate/qwen_vl_interface': 2.85042682037112e-06, 'learning_rate/action_model': 2.85042682037112e-05, 'epoch': 1.45}) +[2026-06-16 03:52:21,789][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.11544637382030487, 'train/grad_norm_pre_clip': 10.08269214630127, 'timing/data': 0.006967717024963349, 'timing/model': 1.6866527930251323, 'learning_rate/qwen_vl_interface': 2.8477990604427764e-06, 'learning_rate/action_model': 2.8477990604427763e-05, 'epoch': 1.45}) +[2026-06-16 03:52:27,716][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.14987938106060028, 'train/grad_norm_pre_clip': 9.014203071594238, 'timing/data': 0.0069899390218779445, 'timing/model': 1.6112163300276734, 'learning_rate/qwen_vl_interface': 2.8451722879684994e-06, 'learning_rate/action_model': 2.8451722879684994e-05, 'epoch': 1.45}) +[2026-06-16 03:52:33,565][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.05957822874188423, 'train/grad_norm_pre_clip': 6.33321475982666, 'timing/data': 0.006475072004832327, 'timing/model': 1.4589938989956863, 'learning_rate/qwen_vl_interface': 2.842546504028059e-06, 'learning_rate/action_model': 2.8425465040280586e-05, 'epoch': 1.45}) +[2026-06-16 03:52:39,467][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.19894173741340637, 'train/grad_norm_pre_clip': 14.93570613861084, 'timing/data': 0.006882180983666331, 'timing/model': 1.4621971669839695, 'learning_rate/qwen_vl_interface': 2.83992170970081e-06, 'learning_rate/action_model': 2.8399217097008102e-05, 'epoch': 1.45}) +[2026-06-16 03:52:45,488][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.0809529647231102, 'train/grad_norm_pre_clip': 14.654732704162598, 'timing/data': 0.006702215003315359, 'timing/model': 1.5034480870235711, 'learning_rate/qwen_vl_interface': 2.837297906065711e-06, 'learning_rate/action_model': 2.8372979060657107e-05, 'epoch': 1.45}) +[2026-06-16 03:52:51,450][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.11237850040197372, 'train/grad_norm_pre_clip': 9.010384559631348, 'timing/data': 0.006806178949773312, 'timing/model': 1.5142197560053319, 'learning_rate/qwen_vl_interface': 2.834675094201306e-06, 'learning_rate/action_model': 2.8346750942013063e-05, 'epoch': 1.46}) +[2026-06-16 03:52:57,391][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.17561683058738708, 'train/grad_norm_pre_clip': 5.968276500701904, 'timing/data': 0.006560957990586758, 'timing/model': 1.6839472790015861, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 1.46}) +[2026-06-16 03:53:03,505][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.14391261339187622, 'train/grad_norm_pre_clip': 12.19996452331543, 'timing/data': 0.007082594966050237, 'timing/model': 1.8480392369674519, 'learning_rate/qwen_vl_interface': 2.829432450096731e-06, 'learning_rate/action_model': 2.8294324500967306e-05, 'epoch': 1.46}) +[2026-06-16 03:53:09,166][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.03468756750226021, 'train/grad_norm_pre_clip': 3.8132808208465576, 'timing/data': 0.0071030460530892015, 'timing/model': 1.4603556679794565, 'learning_rate/qwen_vl_interface': 2.8268126200116118e-06, 'learning_rate/action_model': 2.8268126200116117e-05, 'epoch': 1.46}) +[2026-06-16 03:53:15,177][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.0962127223610878, 'train/grad_norm_pre_clip': 12.727036476135254, 'timing/data': 0.007218846003524959, 'timing/model': 1.4716285329777747, 'learning_rate/qwen_vl_interface': 2.824193786007294e-06, 'learning_rate/action_model': 2.8241937860072937e-05, 'epoch': 1.46}) +[2026-06-16 03:53:21,259][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.09068606793880463, 'train/grad_norm_pre_clip': 11.219501495361328, 'timing/data': 0.006836413987912238, 'timing/model': 1.490252157964278, 'learning_rate/qwen_vl_interface': 2.821575949160279e-06, 'learning_rate/action_model': 2.8215759491602787e-05, 'epoch': 1.46}) +[2026-06-16 03:53:27,277][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.3489280641078949, 'train/grad_norm_pre_clip': 17.50174903869629, 'timing/data': 0.006873378006275743, 'timing/model': 1.5373811730532907, 'learning_rate/qwen_vl_interface': 2.8189591105466625e-06, 'learning_rate/action_model': 2.8189591105466624e-05, 'epoch': 1.46}) +[2026-06-16 03:53:33,376][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.04667821526527405, 'train/grad_norm_pre_clip': 5.970891952514648, 'timing/data': 0.006667396053671837, 'timing/model': 1.8021444419864565, 'learning_rate/qwen_vl_interface': 2.816343271242126e-06, 'learning_rate/action_model': 2.816343271242126e-05, 'epoch': 1.46}) +[2026-06-16 03:53:39,422][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.13844434916973114, 'train/grad_norm_pre_clip': 7.080307960510254, 'timing/data': 0.006566656986251473, 'timing/model': 1.7179354730178602, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 1.46}) +[2026-06-16 03:53:45,261][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.01655297353863716, 'train/grad_norm_pre_clip': 2.702608108520508, 'timing/data': 0.006560397974681109, 'timing/model': 1.4827689540106803, 'learning_rate/qwen_vl_interface': 2.81111459486098e-06, 'learning_rate/action_model': 2.81111459486098e-05, 'epoch': 1.46}) +[2026-06-16 03:53:51,256][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.028570927679538727, 'train/grad_norm_pre_clip': 3.2026448249816895, 'timing/data': 0.006800150033086538, 'timing/model': 1.4495511590503156, 'learning_rate/qwen_vl_interface': 2.8085017599336806e-06, 'learning_rate/action_model': 2.8085017599336806e-05, 'epoch': 1.46}) +[2026-06-16 03:53:57,259][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.20620369911193848, 'train/grad_norm_pre_clip': 10.270951271057129, 'timing/data': 0.006660825980361551, 'timing/model': 1.4823035020381212, 'learning_rate/qwen_vl_interface': 2.8058899286140828e-06, 'learning_rate/action_model': 2.805889928614083e-05, 'epoch': 1.46}) +[2026-06-16 03:54:03,274][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.03358779847621918, 'train/grad_norm_pre_clip': 4.281094074249268, 'timing/data': 0.0076889279880560935, 'timing/model': 1.532076848030556, 'learning_rate/qwen_vl_interface': 2.8032791019758127e-06, 'learning_rate/action_model': 2.8032791019758127e-05, 'epoch': 1.46}) +[2026-06-16 03:54:09,345][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.039885926991701126, 'train/grad_norm_pre_clip': 3.702322244644165, 'timing/data': 0.00652439403347671, 'timing/model': 1.5105183200212196, 'learning_rate/qwen_vl_interface': 2.800669281092084e-06, 'learning_rate/action_model': 2.8006692810920838e-05, 'epoch': 1.46}) +[2026-06-16 03:54:15,351][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.11693384498357773, 'train/grad_norm_pre_clip': 15.121749877929688, 'timing/data': 0.006862821988761425, 'timing/model': 1.6437716740183532, 'learning_rate/qwen_vl_interface': 2.7980604670356938e-06, 'learning_rate/action_model': 2.7980604670356937e-05, 'epoch': 1.46}) +[2026-06-16 03:54:21,161][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.09396960586309433, 'train/grad_norm_pre_clip': 12.105788230895996, 'timing/data': 0.006912516022566706, 'timing/model': 1.4549005899461918, 'learning_rate/qwen_vl_interface': 2.795452660879027e-06, 'learning_rate/action_model': 2.7954526608790273e-05, 'epoch': 1.46}) +[2026-06-16 03:54:27,274][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.07138418406248093, 'train/grad_norm_pre_clip': 6.723633289337158, 'timing/data': 0.00713914999505505, 'timing/model': 1.494436590000987, 'learning_rate/qwen_vl_interface': 2.7928458636940565e-06, 'learning_rate/action_model': 2.7928458636940564e-05, 'epoch': 1.46}) +[2026-06-16 03:54:33,263][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.06568248569965363, 'train/grad_norm_pre_clip': 5.86909294128418, 'timing/data': 0.006566145981196314, 'timing/model': 1.4524796630139463, 'learning_rate/qwen_vl_interface': 2.790240076552336e-06, 'learning_rate/action_model': 2.7902400765523355e-05, 'epoch': 1.46}) +[2026-06-16 04:03:35,252][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.10968740284442902, 'train/grad_norm_pre_clip': 15.291075706481934, 'eval/loss': 0.049617230892181396, 'eval/flap_precision': 0.9256360078277887, 'eval/flap_recall': 0.9478957915831663, 'eval/flap_f1': 0.9366336633663366, 'eval/flap_rate_model': 0.14194444444444446, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9366336633663366, 'timing/data': 0.006643811007961631, 'timing/model': 1.4491880690329708, 'learning_rate/qwen_vl_interface': 2.787635300525006e-06, 'learning_rate/action_model': 2.787635300525006e-05, 'epoch': 1.46}) +[2026-06-16 04:03:35,253][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3400_state +[2026-06-16 04:03:35,257][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 04:03:40,694][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3400_state/model.safetensors +[2026-06-16 04:03:48,834][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3400_state/optimizer.bin +[2026-06-16 04:03:48,835][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3400_state/sampler.bin +[2026-06-16 04:03:48,835][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3400_state/sampler_1.bin +[2026-06-16 04:03:48,836][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3400_state/random_states_0.pkl +[2026-06-16 04:03:48,938][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3400 +[2026-06-16 04:03:48,938][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 04:03:48,948][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 04:03:54,837][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.030225012451410294, 'train/grad_norm_pre_clip': 3.326465129852295, 'timing/data': 0.006927920039743185, 'timing/model': 1.4217904760153033, 'learning_rate/qwen_vl_interface': 2.7850315366827933e-06, 'learning_rate/action_model': 2.7850315366827934e-05, 'epoch': 1.46}) +[2026-06-16 04:04:00,655][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.022346317768096924, 'train/grad_norm_pre_clip': 6.038670063018799, 'timing/data': 0.006893039040733129, 'timing/model': 1.4695573080098256, 'learning_rate/qwen_vl_interface': 2.7824287860960076e-06, 'learning_rate/action_model': 2.7824287860960075e-05, 'epoch': 1.46}) +[2026-06-16 04:04:06,565][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.0607820525765419, 'train/grad_norm_pre_clip': 6.171885967254639, 'timing/data': 0.006837557011749595, 'timing/model': 1.679341631999705, 'learning_rate/qwen_vl_interface': 2.7798270498345377e-06, 'learning_rate/action_model': 2.7798270498345376e-05, 'epoch': 1.46}) +[2026-06-16 04:04:12,339][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.0899580642580986, 'train/grad_norm_pre_clip': 5.916259288787842, 'timing/data': 0.007535593991633505, 'timing/model': 1.4955415820004418, 'learning_rate/qwen_vl_interface': 2.777226328967861e-06, 'learning_rate/action_model': 2.7772263289678607e-05, 'epoch': 1.46}) +[2026-06-16 04:04:18,250][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.03247995302081108, 'train/grad_norm_pre_clip': 7.305848598480225, 'timing/data': 0.006935731042176485, 'timing/model': 1.4601498660049401, 'learning_rate/qwen_vl_interface': 2.774626624565035e-06, 'learning_rate/action_model': 2.774626624565035e-05, 'epoch': 1.47}) +[2026-06-16 04:04:24,149][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.026745639741420746, 'train/grad_norm_pre_clip': 5.828290939331055, 'timing/data': 0.0067595019936561584, 'timing/model': 1.4592521979939193, 'learning_rate/qwen_vl_interface': 2.772027937694701e-06, 'learning_rate/action_model': 2.772027937694701e-05, 'epoch': 1.47}) +[2026-06-16 04:04:30,044][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.11281610280275345, 'train/grad_norm_pre_clip': 7.41803503036499, 'timing/data': 0.006785950972698629, 'timing/model': 1.4259246989968233, 'learning_rate/qwen_vl_interface': 2.7694302694250806e-06, 'learning_rate/action_model': 2.7694302694250808e-05, 'epoch': 1.47}) +[2026-06-16 04:04:35,962][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.0732252225279808, 'train/grad_norm_pre_clip': 4.872800350189209, 'timing/data': 0.006993627001065761, 'timing/model': 1.4819246220286004, 'learning_rate/qwen_vl_interface': 2.7668336208239775e-06, 'learning_rate/action_model': 2.7668336208239776e-05, 'epoch': 1.47}) +[2026-06-16 04:04:41,886][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.06448415666818619, 'train/grad_norm_pre_clip': 8.997852325439453, 'timing/data': 0.006667346984613687, 'timing/model': 1.6959638149710372, 'learning_rate/qwen_vl_interface': 2.7642379929587755e-06, 'learning_rate/action_model': 2.7642379929587754e-05, 'epoch': 1.47}) +[2026-06-16 04:04:47,651][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.1613931506872177, 'train/grad_norm_pre_clip': 14.523543357849121, 'timing/data': 0.006769466970581561, 'timing/model': 1.4757672779960558, 'learning_rate/qwen_vl_interface': 2.7616433868964387e-06, 'learning_rate/action_model': 2.7616433868964388e-05, 'epoch': 1.47}) +[2026-06-16 04:04:53,560][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.025054683908820152, 'train/grad_norm_pre_clip': 4.4677205085754395, 'timing/data': 0.006830586993601173, 'timing/model': 1.467013583984226, 'learning_rate/qwen_vl_interface': 2.7590498037035122e-06, 'learning_rate/action_model': 2.7590498037035124e-05, 'epoch': 1.47}) +[2026-06-16 04:04:59,457][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.005171495024114847, 'train/grad_norm_pre_clip': 1.1256905794143677, 'timing/data': 0.007195513986516744, 'timing/model': 1.4587202979600988, 'learning_rate/qwen_vl_interface': 2.7564572444461218e-06, 'learning_rate/action_model': 2.7564572444461217e-05, 'epoch': 1.47}) +[2026-06-16 04:05:05,436][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.15536782145500183, 'train/grad_norm_pre_clip': 16.391691207885742, 'timing/data': 0.0068675309885293245, 'timing/model': 1.5133624729933217, 'learning_rate/qwen_vl_interface': 2.7538657101899674e-06, 'learning_rate/action_model': 2.7538657101899673e-05, 'epoch': 1.47}) +[2026-06-16 04:05:11,354][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.25131726264953613, 'train/grad_norm_pre_clip': 13.028040885925293, 'timing/data': 0.007027306011877954, 'timing/model': 1.4552603220217861, 'learning_rate/qwen_vl_interface': 2.751275202000332e-06, 'learning_rate/action_model': 2.7512752020003317e-05, 'epoch': 1.47}) +[2026-06-16 04:05:17,292][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.04908505827188492, 'train/grad_norm_pre_clip': 7.679964542388916, 'timing/data': 0.00689938897266984, 'timing/model': 1.6963331520091742, 'learning_rate/qwen_vl_interface': 2.7486857209420763e-06, 'learning_rate/action_model': 2.7486857209420763e-05, 'epoch': 1.47}) +[2026-06-16 04:05:22,957][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.004057714715600014, 'train/grad_norm_pre_clip': 0.6762111186981201, 'timing/data': 0.006646415975410491, 'timing/model': 1.4612802909687161, 'learning_rate/qwen_vl_interface': 2.7460972680796377e-06, 'learning_rate/action_model': 2.7460972680796375e-05, 'epoch': 1.47}) +[2026-06-16 04:05:28,954][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.22332532703876495, 'train/grad_norm_pre_clip': 9.517516136169434, 'timing/data': 0.00650484801735729, 'timing/model': 1.4587549869902432, 'learning_rate/qwen_vl_interface': 2.7435098444770314e-06, 'learning_rate/action_model': 2.7435098444770315e-05, 'epoch': 1.47}) +[2026-06-16 04:05:34,861][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.0662769079208374, 'train/grad_norm_pre_clip': 4.866196632385254, 'timing/data': 0.007250915979966521, 'timing/model': 1.4653776409686543, 'learning_rate/qwen_vl_interface': 2.7409234511978495e-06, 'learning_rate/action_model': 2.7409234511978492e-05, 'epoch': 1.47}) +[2026-06-16 04:05:40,757][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.06807685643434525, 'train/grad_norm_pre_clip': 6.090024948120117, 'timing/data': 0.006669350026641041, 'timing/model': 1.461340754001867, 'learning_rate/qwen_vl_interface': 2.7383380893052606e-06, 'learning_rate/action_model': 2.7383380893052608e-05, 'epoch': 1.47}) +[2026-06-16 04:05:46,732][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.1555163413286209, 'train/grad_norm_pre_clip': 5.938554763793945, 'timing/data': 0.006791010033339262, 'timing/model': 1.4960389150073752, 'learning_rate/qwen_vl_interface': 2.73575375986201e-06, 'learning_rate/action_model': 2.7357537598620096e-05, 'epoch': 1.47}) +[2026-06-16 04:05:52,687][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.013408391736447811, 'train/grad_norm_pre_clip': 5.869661331176758, 'timing/data': 0.006823856965638697, 'timing/model': 1.6935982800205238, 'learning_rate/qwen_vl_interface': 2.733170463930417e-06, 'learning_rate/action_model': 2.7331704639304167e-05, 'epoch': 1.47}) +[2026-06-16 04:05:58,445][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.05190064385533333, 'train/grad_norm_pre_clip': 6.507439613342285, 'timing/data': 0.006817768036853522, 'timing/model': 1.4521375219919719, 'learning_rate/qwen_vl_interface': 2.7305882025723774e-06, 'learning_rate/action_model': 2.7305882025723774e-05, 'epoch': 1.47}) +[2026-06-16 04:06:04,424][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.07924309372901917, 'train/grad_norm_pre_clip': 8.974885940551758, 'timing/data': 0.006537675973959267, 'timing/model': 1.53288726095343, 'learning_rate/qwen_vl_interface': 2.7280069768493626e-06, 'learning_rate/action_model': 2.7280069768493628e-05, 'epoch': 1.47}) +[2026-06-16 04:06:10,369][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.0746648982167244, 'train/grad_norm_pre_clip': 10.749361991882324, 'timing/data': 0.007890846987720579, 'timing/model': 1.4836794710136019, 'learning_rate/qwen_vl_interface': 2.7254267878224144e-06, 'learning_rate/action_model': 2.7254267878224143e-05, 'epoch': 1.47}) +[2026-06-16 04:06:16,253][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.027934852987527847, 'train/grad_norm_pre_clip': 4.678913593292236, 'timing/data': 0.006771370011847466, 'timing/model': 1.4605281690019183, 'learning_rate/qwen_vl_interface': 2.7228476365521524e-06, 'learning_rate/action_model': 2.722847636552152e-05, 'epoch': 1.47}) +[2026-06-16 04:06:22,233][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.03984512388706207, 'train/grad_norm_pre_clip': 6.303332805633545, 'timing/data': 0.006779571995139122, 'timing/model': 1.4994655949994922, 'learning_rate/qwen_vl_interface': 2.7202695240987674e-06, 'learning_rate/action_model': 2.7202695240987676e-05, 'epoch': 1.47}) +[2026-06-16 04:06:28,189][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.0720609501004219, 'train/grad_norm_pre_clip': 6.844400405883789, 'timing/data': 0.006738030991982669, 'timing/model': 1.6958936390001327, 'learning_rate/qwen_vl_interface': 2.717692451522025e-06, 'learning_rate/action_model': 2.7176924515220248e-05, 'epoch': 1.47}) +[2026-06-16 04:06:33,859][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.19805046916007996, 'train/grad_norm_pre_clip': 11.030794143676758, 'timing/data': 0.006434424023609608, 'timing/model': 1.4653221949702129, 'learning_rate/qwen_vl_interface': 2.7151164198812624e-06, 'learning_rate/action_model': 2.7151164198812622e-05, 'epoch': 1.47}) +[2026-06-16 04:06:39,752][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.06544815748929977, 'train/grad_norm_pre_clip': 7.713496685028076, 'timing/data': 0.006734796974342316, 'timing/model': 1.459661700995639, 'learning_rate/qwen_vl_interface': 2.712541430235388e-06, 'learning_rate/action_model': 2.712541430235388e-05, 'epoch': 1.48}) +[2026-06-16 04:06:45,652][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.12285115569829941, 'train/grad_norm_pre_clip': 9.843770980834961, 'timing/data': 0.006410578964278102, 'timing/model': 1.4590670380275697, 'learning_rate/qwen_vl_interface': 2.7099674836428842e-06, 'learning_rate/action_model': 2.709967483642884e-05, 'epoch': 1.48}) +[2026-06-16 04:06:51,554][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.013818304985761642, 'train/grad_norm_pre_clip': 1.1174860000610352, 'timing/data': 0.00663084298139438, 'timing/model': 1.4594422010122798, 'learning_rate/qwen_vl_interface': 2.7073945811618018e-06, 'learning_rate/action_model': 2.7073945811618017e-05, 'epoch': 1.48}) +[2026-06-16 04:06:57,458][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.03746098279953003, 'train/grad_norm_pre_clip': 4.169968128204346, 'timing/data': 0.006479741015937179, 'timing/model': 1.4318349540117197, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 1.48}) +[2026-06-16 04:07:03,383][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.06966562569141388, 'train/grad_norm_pre_clip': 3.494971513748169, 'timing/data': 0.006557235028594732, 'timing/model': 1.6887239330098964, 'learning_rate/qwen_vl_interface': 2.7022519127639687e-06, 'learning_rate/action_model': 2.7022519127639682e-05, 'epoch': 1.48}) +[2026-06-16 04:07:09,054][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.08910363167524338, 'train/grad_norm_pre_clip': 12.073920249938965, 'timing/data': 0.006617543986067176, 'timing/model': 1.459121152991429, 'learning_rate/qwen_vl_interface': 2.6996821489611713e-06, 'learning_rate/action_model': 2.699682148961171e-05, 'epoch': 1.48}) +[2026-06-16 04:07:15,055][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.14575637876987457, 'train/grad_norm_pre_clip': 13.846027374267578, 'timing/data': 0.006549715006258339, 'timing/model': 1.4603045450057834, 'learning_rate/qwen_vl_interface': 2.697113433497712e-06, 'learning_rate/action_model': 2.6971134334977115e-05, 'epoch': 1.48}) +[2026-06-16 04:07:20,963][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.03805786743760109, 'train/grad_norm_pre_clip': 3.3631832599639893, 'timing/data': 0.007678344030864537, 'timing/model': 1.46546464698622, 'learning_rate/qwen_vl_interface': 2.6945457674294863e-06, 'learning_rate/action_model': 2.694545767429486e-05, 'epoch': 1.48}) +[2026-06-16 04:07:26,856][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.13401858508586884, 'train/grad_norm_pre_clip': 29.44683837890625, 'timing/data': 0.006579047010745853, 'timing/model': 1.4585577889811248, 'learning_rate/qwen_vl_interface': 2.6919791518119676e-06, 'learning_rate/action_model': 2.6919791518119674e-05, 'epoch': 1.48}) +[2026-06-16 04:07:32,757][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.06424625217914581, 'train/grad_norm_pre_clip': 10.105222702026367, 'timing/data': 0.00673891301266849, 'timing/model': 1.4314297569799237, 'learning_rate/qwen_vl_interface': 2.6894135877001946e-06, 'learning_rate/action_model': 2.689413587700195e-05, 'epoch': 1.48}) +[2026-06-16 04:07:38,869][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.1309366375207901, 'train/grad_norm_pre_clip': 11.570521354675293, 'timing/data': 0.008189356012735516, 'timing/model': 1.7589366509928368, 'learning_rate/qwen_vl_interface': 2.686849076148775e-06, 'learning_rate/action_model': 2.6868490761487747e-05, 'epoch': 1.48}) +[2026-06-16 04:07:44,702][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.03339621424674988, 'train/grad_norm_pre_clip': 12.163841247558594, 'timing/data': 0.0066303329658694565, 'timing/model': 1.6041434209910221, 'learning_rate/qwen_vl_interface': 2.68428561821188e-06, 'learning_rate/action_model': 2.6842856182118802e-05, 'epoch': 1.48}) +[2026-06-16 04:07:50,354][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.09627874940633774, 'train/grad_norm_pre_clip': 11.094060897827148, 'timing/data': 0.006711566005833447, 'timing/model': 1.4488520179875195, 'learning_rate/qwen_vl_interface': 2.681723214943253e-06, 'learning_rate/action_model': 2.681723214943253e-05, 'epoch': 1.48}) +[2026-06-16 04:07:56,255][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.03033270500600338, 'train/grad_norm_pre_clip': 5.957774639129639, 'timing/data': 0.006973511015530676, 'timing/model': 1.4586647510295734, 'learning_rate/qwen_vl_interface': 2.6791618673961996e-06, 'learning_rate/action_model': 2.6791618673961994e-05, 'epoch': 1.48}) +[2026-06-16 04:08:02,161][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.05266120657324791, 'train/grad_norm_pre_clip': 9.489335060119629, 'timing/data': 0.0066207319614477456, 'timing/model': 1.4647805740241893, 'learning_rate/qwen_vl_interface': 2.676601576623594e-06, 'learning_rate/action_model': 2.6766015766235937e-05, 'epoch': 1.48}) +[2026-06-16 04:08:08,133][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.08627171814441681, 'train/grad_norm_pre_clip': 15.62218952178955, 'timing/data': 0.006666720029897988, 'timing/model': 1.501662538968958, 'learning_rate/qwen_vl_interface': 2.674042343677874e-06, 'learning_rate/action_model': 2.6740423436778738e-05, 'epoch': 1.48}) +[2026-06-16 04:08:14,080][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.015289582312107086, 'train/grad_norm_pre_clip': 6.513119697570801, 'timing/data': 0.006836470041889697, 'timing/model': 1.6845137950149365, 'learning_rate/qwen_vl_interface': 2.6714841696110456e-06, 'learning_rate/action_model': 2.6714841696110456e-05, 'epoch': 1.48}) +[2026-06-16 04:08:19,986][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.1305856555700302, 'train/grad_norm_pre_clip': 7.2960205078125, 'timing/data': 0.00685886200517416, 'timing/model': 1.68848395900568, 'learning_rate/qwen_vl_interface': 2.6689270554746743e-06, 'learning_rate/action_model': 2.6689270554746743e-05, 'epoch': 1.48}) +[2026-06-16 04:08:25,655][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.1498890519142151, 'train/grad_norm_pre_clip': 13.438772201538086, 'timing/data': 0.006885140959639102, 'timing/model': 1.457107416994404, 'learning_rate/qwen_vl_interface': 2.6663710023198956e-06, 'learning_rate/action_model': 2.6663710023198956e-05, 'epoch': 1.48}) +[2026-06-16 04:08:31,560][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.1554248034954071, 'train/grad_norm_pre_clip': 14.332379341125488, 'timing/data': 0.007355151989031583, 'timing/model': 1.4639850810053758, 'learning_rate/qwen_vl_interface': 2.663816011197405e-06, 'learning_rate/action_model': 2.6638160111974048e-05, 'epoch': 1.48}) +[2026-06-16 04:08:37,553][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.18124325573444366, 'train/grad_norm_pre_clip': 10.187912940979004, 'timing/data': 0.00707882602000609, 'timing/model': 1.4866157039650716, 'learning_rate/qwen_vl_interface': 2.661262083157463e-06, 'learning_rate/action_model': 2.661262083157463e-05, 'epoch': 1.48}) +[2026-06-16 04:08:43,527][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.1710028499364853, 'train/grad_norm_pre_clip': 11.80637264251709, 'timing/data': 0.006781317992135882, 'timing/model': 1.49897991202306, 'learning_rate/qwen_vl_interface': 2.6587092192498943e-06, 'learning_rate/action_model': 2.658709219249894e-05, 'epoch': 1.48}) +[2026-06-16 04:08:49,485][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.10416491329669952, 'train/grad_norm_pre_clip': 16.46114158630371, 'timing/data': 0.006820567010436207, 'timing/model': 1.6897934299777262, 'learning_rate/qwen_vl_interface': 2.656157420524084e-06, 'learning_rate/action_model': 2.6561574205240842e-05, 'epoch': 1.48}) +[2026-06-16 04:08:55,154][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.15403030812740326, 'train/grad_norm_pre_clip': 12.529967308044434, 'timing/data': 0.0067526360508054495, 'timing/model': 1.4587380660232157, 'learning_rate/qwen_vl_interface': 2.6536066880289795e-06, 'learning_rate/action_model': 2.6536066880289794e-05, 'epoch': 1.49}) +[2026-06-16 04:09:01,055][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.16963914036750793, 'train/grad_norm_pre_clip': 18.061548233032227, 'timing/data': 0.006656393990851939, 'timing/model': 1.4600715120323002, 'learning_rate/qwen_vl_interface': 2.6510570228130916e-06, 'learning_rate/action_model': 2.6510570228130914e-05, 'epoch': 1.49}) +[2026-06-16 04:09:06,954][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.03490104153752327, 'train/grad_norm_pre_clip': 7.050565719604492, 'timing/data': 0.00644984096288681, 'timing/model': 1.4586565340287052, 'learning_rate/qwen_vl_interface': 2.648508425924491e-06, 'learning_rate/action_model': 2.6485084259244907e-05, 'epoch': 1.49}) +[2026-06-16 04:09:12,855][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.010955766774713993, 'train/grad_norm_pre_clip': 2.272067070007324, 'timing/data': 0.007426657015457749, 'timing/model': 1.4307664500083774, 'learning_rate/qwen_vl_interface': 2.6459608984108093e-06, 'learning_rate/action_model': 2.645960898410809e-05, 'epoch': 1.49}) +[2026-06-16 04:09:18,744][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.1800692081451416, 'train/grad_norm_pre_clip': 16.90060043334961, 'timing/data': 0.006549577985424548, 'timing/model': 1.504606154980138, 'learning_rate/qwen_vl_interface': 2.6434144413192414e-06, 'learning_rate/action_model': 2.6434144413192414e-05, 'epoch': 1.49}) +[2026-06-16 04:09:24,630][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.005799581296741962, 'train/grad_norm_pre_clip': 1.384257197380066, 'timing/data': 0.006519262969959527, 'timing/model': 1.637384583009407, 'learning_rate/qwen_vl_interface': 2.6408690556965347e-06, 'learning_rate/action_model': 2.640869055696535e-05, 'epoch': 1.49}) +[2026-06-16 04:09:30,354][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.04166345298290253, 'train/grad_norm_pre_clip': 4.09556770324707, 'timing/data': 0.006745856022462249, 'timing/model': 1.459338674962055, 'learning_rate/qwen_vl_interface': 2.6383247425890047e-06, 'learning_rate/action_model': 2.6383247425890047e-05, 'epoch': 1.49}) +[2026-06-16 04:09:36,352][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.005886752158403397, 'train/grad_norm_pre_clip': 0.6169397830963135, 'timing/data': 0.006540743983350694, 'timing/model': 1.4602846310008317, 'learning_rate/qwen_vl_interface': 2.6357815030425204e-06, 'learning_rate/action_model': 2.6357815030425205e-05, 'epoch': 1.49}) +[2026-06-16 04:09:42,251][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.22217707335948944, 'train/grad_norm_pre_clip': 8.624990463256836, 'timing/data': 0.006685767963062972, 'timing/model': 1.45667634595884, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 1.49}) +[2026-06-16 04:09:48,259][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.17661070823669434, 'train/grad_norm_pre_clip': 13.414426803588867, 'timing/data': 0.006435008952394128, 'timing/model': 1.5170629150234163, 'learning_rate/qwen_vl_interface': 2.6306982488139704e-06, 'learning_rate/action_model': 2.6306982488139702e-05, 'epoch': 1.49}) +[2026-06-16 04:09:54,229][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.2529688775539398, 'train/grad_norm_pre_clip': 16.79869270324707, 'timing/data': 0.006604427995625883, 'timing/model': 1.4965412230230868, 'learning_rate/qwen_vl_interface': 2.628158236221438e-06, 'learning_rate/action_model': 2.628158236221438e-05, 'epoch': 1.49}) +[2026-06-16 04:10:00,146][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.07407952100038528, 'train/grad_norm_pre_clip': 4.213472843170166, 'timing/data': 0.006541735026985407, 'timing/model': 1.6485452380147763, 'learning_rate/qwen_vl_interface': 2.6256193013690183e-06, 'learning_rate/action_model': 2.6256193013690182e-05, 'epoch': 1.49}) +[2026-06-16 04:10:05,856][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.006287400145083666, 'train/grad_norm_pre_clip': 0.914419949054718, 'timing/data': 0.006642383988946676, 'timing/model': 1.456723775016144, 'learning_rate/qwen_vl_interface': 2.623081445300373e-06, 'learning_rate/action_model': 2.6230814453003728e-05, 'epoch': 1.49}) +[2026-06-16 04:10:11,861][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.25097090005874634, 'train/grad_norm_pre_clip': 14.924020767211914, 'timing/data': 0.006908194976858795, 'timing/model': 1.4640952459885739, 'learning_rate/qwen_vl_interface': 2.6205446690587183e-06, 'learning_rate/action_model': 2.6205446690587182e-05, 'epoch': 1.49}) +[2026-06-16 04:10:17,852][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.10730060189962387, 'train/grad_norm_pre_clip': 5.308382987976074, 'timing/data': 0.006695592019241303, 'timing/model': 1.4591102360282093, 'learning_rate/qwen_vl_interface': 2.618008973686828e-06, 'learning_rate/action_model': 2.618008973686828e-05, 'epoch': 1.49}) +[2026-06-16 04:10:23,864][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.04098257049918175, 'train/grad_norm_pre_clip': 8.815810203552246, 'timing/data': 0.006715441006235778, 'timing/model': 1.5266601239563897, 'learning_rate/qwen_vl_interface': 2.615474360227028e-06, 'learning_rate/action_model': 2.615474360227028e-05, 'epoch': 1.49}) +[2026-06-16 04:10:29,780][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.04411784186959267, 'train/grad_norm_pre_clip': 9.554509162902832, 'timing/data': 0.00749906396958977, 'timing/model': 1.6857268980238587, 'learning_rate/qwen_vl_interface': 2.612940829721204e-06, 'learning_rate/action_model': 2.612940829721204e-05, 'epoch': 1.49}) +[2026-06-16 04:10:35,463][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.12825068831443787, 'train/grad_norm_pre_clip': 6.894985675811768, 'timing/data': 0.006972648960072547, 'timing/model': 1.4678833190118894, 'learning_rate/qwen_vl_interface': 2.6104083832107948e-06, 'learning_rate/action_model': 2.6104083832107945e-05, 'epoch': 1.49}) +[2026-06-16 04:10:41,356][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.2446455955505371, 'train/grad_norm_pre_clip': 19.609468460083008, 'timing/data': 0.006613361008930951, 'timing/model': 1.4582800939679146, 'learning_rate/qwen_vl_interface': 2.607877021736793e-06, 'learning_rate/action_model': 2.6078770217367926e-05, 'epoch': 1.49}) +[2026-06-16 04:10:47,165][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.12716597318649292, 'train/grad_norm_pre_clip': 14.994400978088379, 'timing/data': 0.006762409990187734, 'timing/model': 1.4652385579538532, 'learning_rate/qwen_vl_interface': 2.605346746339745e-06, 'learning_rate/action_model': 2.605346746339745e-05, 'epoch': 1.49}) +[2026-06-16 04:10:53,055][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.17111705243587494, 'train/grad_norm_pre_clip': 12.87265396118164, 'timing/data': 0.006547513999976218, 'timing/model': 1.4576653799740598, 'learning_rate/qwen_vl_interface': 2.602817558059753e-06, 'learning_rate/action_model': 2.6028175580597532e-05, 'epoch': 1.49}) +[2026-06-16 04:10:59,002][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.06432195007801056, 'train/grad_norm_pre_clip': 6.038970470428467, 'timing/data': 0.006956875964533538, 'timing/model': 1.5087133669876494, 'learning_rate/qwen_vl_interface': 2.60028945793647e-06, 'learning_rate/action_model': 2.60028945793647e-05, 'epoch': 1.49}) +[2026-06-16 04:11:04,890][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.013992603868246078, 'train/grad_norm_pre_clip': 2.0321972370147705, 'timing/data': 0.007989826030097902, 'timing/model': 1.6903626960120164, 'learning_rate/qwen_vl_interface': 2.5977624470091024e-06, 'learning_rate/action_model': 2.5977624470091023e-05, 'epoch': 1.49}) +[2026-06-16 04:11:10,555][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.14752061665058136, 'train/grad_norm_pre_clip': 10.669979095458984, 'timing/data': 0.006756371003575623, 'timing/model': 1.4454715749598108, 'learning_rate/qwen_vl_interface': 2.59523652631641e-06, 'learning_rate/action_model': 2.5952365263164097e-05, 'epoch': 1.5}) +[2026-06-16 04:11:16,456][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.09877705574035645, 'train/grad_norm_pre_clip': 8.648244857788086, 'timing/data': 0.006668852001894265, 'timing/model': 1.4604198649758473, 'learning_rate/qwen_vl_interface': 2.5927116968967024e-06, 'learning_rate/action_model': 2.5927116968967025e-05, 'epoch': 1.5}) +[2026-06-16 04:11:22,555][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.11577539145946503, 'train/grad_norm_pre_clip': 11.273858070373535, 'timing/data': 0.006785734964068979, 'timing/model': 1.4795425169868395, 'learning_rate/qwen_vl_interface': 2.5901879597878417e-06, 'learning_rate/action_model': 2.5901879597878415e-05, 'epoch': 1.5}) +[2026-06-16 04:11:28,454][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.0587073415517807, 'train/grad_norm_pre_clip': 8.744187355041504, 'timing/data': 0.006542056042235345, 'timing/model': 1.4579461149987765, 'learning_rate/qwen_vl_interface': 2.5876653160272395e-06, 'learning_rate/action_model': 2.5876653160272397e-05, 'epoch': 1.5}) +[2026-06-16 04:11:34,353][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.1067509725689888, 'train/grad_norm_pre_clip': 15.377036094665527, 'timing/data': 0.006902916997205466, 'timing/model': 1.4686005580006167, 'learning_rate/qwen_vl_interface': 2.5851437666518607e-06, 'learning_rate/action_model': 2.5851437666518607e-05, 'epoch': 1.5}) +[2026-06-16 04:11:40,290][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.14772853255271912, 'train/grad_norm_pre_clip': 14.60481071472168, 'timing/data': 0.006484220968559384, 'timing/model': 1.6939819499966688, 'learning_rate/qwen_vl_interface': 2.5826233126982198e-06, 'learning_rate/action_model': 2.5826233126982196e-05, 'epoch': 1.5}) +[2026-06-16 04:11:45,956][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.09038514643907547, 'train/grad_norm_pre_clip': 6.647436141967773, 'timing/data': 0.006610937009099871, 'timing/model': 1.462725845980458, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 1.5}) +[2026-06-16 04:11:51,909][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.2439088672399521, 'train/grad_norm_pre_clip': 18.85423469543457, 'timing/data': 0.009098316018935293, 'timing/model': 1.5122456419630907, 'learning_rate/qwen_vl_interface': 2.577585695199954e-06, 'learning_rate/action_model': 2.5775856951999538e-05, 'epoch': 1.5}) +[2026-06-16 04:11:57,858][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.07291693240404129, 'train/grad_norm_pre_clip': 6.616799354553223, 'timing/data': 0.00661144801415503, 'timing/model': 1.4656429990427569, 'learning_rate/qwen_vl_interface': 2.575068533726101e-06, 'learning_rate/action_model': 2.575068533726101e-05, 'epoch': 1.5}) +[2026-06-16 04:12:03,764][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.07515448331832886, 'train/grad_norm_pre_clip': 8.856019020080566, 'timing/data': 0.006698024983052164, 'timing/model': 1.439144556003157, 'learning_rate/qwen_vl_interface': 2.5725524718155333e-06, 'learning_rate/action_model': 2.572552471815533e-05, 'epoch': 1.5}) +[2026-06-16 04:12:09,654][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.1098603829741478, 'train/grad_norm_pre_clip': 5.415701866149902, 'timing/data': 0.006706357991788536, 'timing/model': 1.4581561150262132, 'learning_rate/qwen_vl_interface': 2.570037510502508e-06, 'learning_rate/action_model': 2.5700375105025083e-05, 'epoch': 1.5}) +[2026-06-16 04:12:15,688][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.08963680267333984, 'train/grad_norm_pre_clip': 6.446550369262695, 'timing/data': 0.007369502971414477, 'timing/model': 1.6861908769933507, 'learning_rate/qwen_vl_interface': 2.5675236508208313e-06, 'learning_rate/action_model': 2.5675236508208313e-05, 'epoch': 1.5}) +[2026-06-16 04:12:21,451][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.019605256617069244, 'train/grad_norm_pre_clip': 3.408033609390259, 'timing/data': 0.006778984039556235, 'timing/model': 1.4593107359833084, 'learning_rate/qwen_vl_interface': 2.565010893803856e-06, 'learning_rate/action_model': 2.565010893803856e-05, 'epoch': 1.5}) +[2026-06-16 04:12:27,352][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.02514856681227684, 'train/grad_norm_pre_clip': 2.400038480758667, 'timing/data': 0.006737193034496158, 'timing/model': 1.460093133035116, 'learning_rate/qwen_vl_interface': 2.562499240484482e-06, 'learning_rate/action_model': 2.5624992404844816e-05, 'epoch': 1.5}) +[2026-06-16 04:12:33,335][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.18041610717773438, 'train/grad_norm_pre_clip': 11.68876838684082, 'timing/data': 0.006605328992009163, 'timing/model': 1.5396784850163385, 'learning_rate/qwen_vl_interface': 2.559988691895155e-06, 'learning_rate/action_model': 2.559988691895155e-05, 'epoch': 1.5}) +[2026-06-16 04:12:39,256][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.11188827455043793, 'train/grad_norm_pre_clip': 8.829132080078125, 'timing/data': 0.006778033042792231, 'timing/model': 1.4591057020006701, 'learning_rate/qwen_vl_interface': 2.5574792490678648e-06, 'learning_rate/action_model': 2.5574792490678644e-05, 'epoch': 1.5}) +[2026-06-16 04:12:45,230][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.10565037280321121, 'train/grad_norm_pre_clip': 6.943935871124268, 'timing/data': 0.0067691399599425495, 'timing/model': 1.4984290059655905, 'learning_rate/qwen_vl_interface': 2.5549709130341493e-06, 'learning_rate/action_model': 2.5549709130341493e-05, 'epoch': 1.5}) +[2026-06-16 04:12:51,282][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.021122602745890617, 'train/grad_norm_pre_clip': 2.402864456176758, 'timing/data': 0.0067142590414732695, 'timing/model': 1.7093210619641468, 'learning_rate/qwen_vl_interface': 2.5524636848250924e-06, 'learning_rate/action_model': 2.552463684825092e-05, 'epoch': 1.5}) +[2026-06-16 04:12:56,953][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.032844047993421555, 'train/grad_norm_pre_clip': 2.6297028064727783, 'timing/data': 0.006995222996920347, 'timing/model': 1.4562563889776357, 'learning_rate/qwen_vl_interface': 2.5499575654713186e-06, 'learning_rate/action_model': 2.5499575654713185e-05, 'epoch': 1.5}) +[2026-06-16 04:13:02,852][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.0676356852054596, 'train/grad_norm_pre_clip': 17.255474090576172, 'timing/data': 0.006898500025272369, 'timing/model': 1.456686718040146, 'learning_rate/qwen_vl_interface': 2.547452556003001e-06, 'learning_rate/action_model': 2.547452556003001e-05, 'epoch': 1.5}) +[2026-06-16 04:13:08,853][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.2875709533691406, 'train/grad_norm_pre_clip': 20.724132537841797, 'timing/data': 0.007213674020022154, 'timing/model': 1.4868633329751901, 'learning_rate/qwen_vl_interface': 2.5449486574498535e-06, 'learning_rate/action_model': 2.5449486574498534e-05, 'epoch': 1.5}) +[2026-06-16 04:13:14,754][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.12132273614406586, 'train/grad_norm_pre_clip': 13.486472129821777, 'timing/data': 0.00674535398138687, 'timing/model': 1.4590863510384224, 'learning_rate/qwen_vl_interface': 2.5424458708411335e-06, 'learning_rate/action_model': 2.5424458708411335e-05, 'epoch': 1.5}) +[2026-06-16 04:13:20,691][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.1599530577659607, 'train/grad_norm_pre_clip': 10.618404388427734, 'timing/data': 0.0068252719938755035, 'timing/model': 1.463932066981215, 'learning_rate/qwen_vl_interface': 2.5399441972056437e-06, 'learning_rate/action_model': 2.5399441972056438e-05, 'epoch': 1.5}) +[2026-06-16 04:13:26,578][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.05237443000078201, 'train/grad_norm_pre_clip': 5.30016565322876, 'timing/data': 0.006392966024577618, 'timing/model': 1.6889952199999243, 'learning_rate/qwen_vl_interface': 2.537443637571726e-06, 'learning_rate/action_model': 2.537443637571726e-05, 'epoch': 1.51}) +[2026-06-16 04:13:32,254][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.03905867785215378, 'train/grad_norm_pre_clip': 5.547409534454346, 'timing/data': 0.006881063978653401, 'timing/model': 1.4480993559700437, 'learning_rate/qwen_vl_interface': 2.534944192967269e-06, 'learning_rate/action_model': 2.534944192967269e-05, 'epoch': 1.51}) +[2026-06-16 04:22:25,051][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.02712918259203434, 'train/grad_norm_pre_clip': 3.396909236907959, 'eval/loss': 0.04842057451605797, 'eval/flap_precision': 0.9373737373737374, 'eval/flap_recall': 0.9298597194388778, 'eval/flap_f1': 0.9336016096579477, 'eval/flap_rate_model': 0.1375, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9336016096579477, 'timing/data': 0.006764271995052695, 'timing/model': 1.458747438970022, 'learning_rate/qwen_vl_interface': 2.532445864419695e-06, 'learning_rate/action_model': 2.5324458644196952e-05, 'epoch': 1.51}) +[2026-06-16 04:22:25,051][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3500_state +[2026-06-16 04:22:25,055][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 04:22:30,644][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3500_state/model.safetensors +[2026-06-16 04:22:38,831][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3500_state/optimizer.bin +[2026-06-16 04:22:38,831][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3500_state/sampler.bin +[2026-06-16 04:22:38,831][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-16 04:22:38,832][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-16 04:22:38,963][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-16 04:22:38,964][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 04:22:38,973][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 04:22:44,940][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.06202926114201546, 'train/grad_norm_pre_clip': 6.902623653411865, 'timing/data': 0.006418893986847252, 'timing/model': 1.4748523050220683, 'learning_rate/qwen_vl_interface': 2.5299486529559758e-06, 'learning_rate/action_model': 2.5299486529559756e-05, 'epoch': 1.51}) +[2026-06-16 04:22:51,046][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.11386600881814957, 'train/grad_norm_pre_clip': 10.246618270874023, 'timing/data': 0.006286797986831516, 'timing/model': 1.4756086669513024, 'learning_rate/qwen_vl_interface': 2.5274525596026202e-06, 'learning_rate/action_model': 2.52745255960262e-05, 'epoch': 1.51}) +[2026-06-16 04:22:56,946][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.09593795239925385, 'train/grad_norm_pre_clip': 7.1064910888671875, 'timing/data': 0.006524197990074754, 'timing/model': 1.4570762009825557, 'learning_rate/qwen_vl_interface': 2.524957585385677e-06, 'learning_rate/action_model': 2.524957585385677e-05, 'epoch': 1.51}) +[2026-06-16 04:23:02,947][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.022013908252120018, 'train/grad_norm_pre_clip': 3.5428555011749268, 'timing/data': 0.006725494051352143, 'timing/model': 1.456516316975467, 'learning_rate/qwen_vl_interface': 2.522463731330738e-06, 'learning_rate/action_model': 2.5224637313307374e-05, 'epoch': 1.51}) +[2026-06-16 04:23:08,833][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.21227145195007324, 'train/grad_norm_pre_clip': 14.310493469238281, 'timing/data': 0.006620959960855544, 'timing/model': 1.4226258759736083, 'learning_rate/qwen_vl_interface': 2.5199709984629294e-06, 'learning_rate/action_model': 2.5199709984629292e-05, 'epoch': 1.51}) +[2026-06-16 04:23:15,082][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.1297108381986618, 'train/grad_norm_pre_clip': 14.025935173034668, 'timing/data': 0.006704212981276214, 'timing/model': 1.7121021730126813, 'learning_rate/qwen_vl_interface': 2.517479387806922e-06, 'learning_rate/action_model': 2.517479387806922e-05, 'epoch': 1.51}) +[2026-06-16 04:23:20,942][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.009645756334066391, 'train/grad_norm_pre_clip': 1.4264510869979858, 'timing/data': 0.006875624996609986, 'timing/model': 1.651817583013326, 'learning_rate/qwen_vl_interface': 2.5149889003869217e-06, 'learning_rate/action_model': 2.5149889003869214e-05, 'epoch': 1.51}) +[2026-06-16 04:23:26,647][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.36703115701675415, 'train/grad_norm_pre_clip': 11.91995620727539, 'timing/data': 0.006718983990140259, 'timing/model': 1.4611038199509494, 'learning_rate/qwen_vl_interface': 2.5124995372266736e-06, 'learning_rate/action_model': 2.5124995372266735e-05, 'epoch': 1.51}) +[2026-06-16 04:23:32,646][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.05047907680273056, 'train/grad_norm_pre_clip': 4.753206729888916, 'timing/data': 0.006746595026925206, 'timing/model': 1.4595043759909458, 'learning_rate/qwen_vl_interface': 2.5100112993494626e-06, 'learning_rate/action_model': 2.5100112993494624e-05, 'epoch': 1.51}) +[2026-06-16 04:23:38,546][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.009098710492253304, 'train/grad_norm_pre_clip': 1.1498970985412598, 'timing/data': 0.006628561008255929, 'timing/model': 1.4611869209911674, 'learning_rate/qwen_vl_interface': 2.5075241877781085e-06, 'learning_rate/action_model': 2.5075241877781086e-05, 'epoch': 1.51}) +[2026-06-16 04:23:44,534][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.13227395713329315, 'train/grad_norm_pre_clip': 6.938337326049805, 'timing/data': 0.0067266260157339275, 'timing/model': 1.5181894699926488, 'learning_rate/qwen_vl_interface': 2.5050382035349673e-06, 'learning_rate/action_model': 2.5050382035349674e-05, 'epoch': 1.51}) +[2026-06-16 04:23:50,452][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.11284578591585159, 'train/grad_norm_pre_clip': 7.920121192932129, 'timing/data': 0.007230898016132414, 'timing/model': 1.4285559229901992, 'learning_rate/qwen_vl_interface': 2.5025533476419343e-06, 'learning_rate/action_model': 2.5025533476419343e-05, 'epoch': 1.51}) +[2026-06-16 04:23:56,346][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.003708179807290435, 'train/grad_norm_pre_clip': 0.6176151037216187, 'timing/data': 0.0066511849872767925, 'timing/model': 1.6536423210054636, 'learning_rate/qwen_vl_interface': 2.5000696211204394e-06, 'learning_rate/action_model': 2.5000696211204394e-05, 'epoch': 1.51}) +[2026-06-16 04:24:02,061][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.1004427820444107, 'train/grad_norm_pre_clip': 12.101045608520508, 'timing/data': 0.00738246098626405, 'timing/model': 1.4631097759702243, 'learning_rate/qwen_vl_interface': 2.4975870249914495e-06, 'learning_rate/action_model': 2.4975870249914497e-05, 'epoch': 1.51}) +[2026-06-16 04:24:08,053][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.10037247091531754, 'train/grad_norm_pre_clip': 12.341836929321289, 'timing/data': 0.006680538994260132, 'timing/model': 1.4626284569967538, 'learning_rate/qwen_vl_interface': 2.4951055602754664e-06, 'learning_rate/action_model': 2.4951055602754665e-05, 'epoch': 1.51}) +[2026-06-16 04:24:14,003][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.004576108418405056, 'train/grad_norm_pre_clip': 1.0588940382003784, 'timing/data': 0.006622702989261597, 'timing/model': 1.5057715449947864, 'learning_rate/qwen_vl_interface': 2.4926252279925255e-06, 'learning_rate/action_model': 2.4926252279925254e-05, 'epoch': 1.51}) +[2026-06-16 04:24:19,945][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.12320569157600403, 'train/grad_norm_pre_clip': 11.155326843261719, 'timing/data': 0.006684234016574919, 'timing/model': 1.505611518980004, 'learning_rate/qwen_vl_interface': 2.490146029162199e-06, 'learning_rate/action_model': 2.4901460291621988e-05, 'epoch': 1.51}) +[2026-06-16 04:24:25,880][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.04057886824011803, 'train/grad_norm_pre_clip': 5.523334503173828, 'timing/data': 0.006628420960623771, 'timing/model': 1.6870585650322028, 'learning_rate/qwen_vl_interface': 2.4876679648035905e-06, 'learning_rate/action_model': 2.4876679648035903e-05, 'epoch': 1.51}) +[2026-06-16 04:24:31,550][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.16529710590839386, 'train/grad_norm_pre_clip': 8.829428672790527, 'timing/data': 0.006720677018165588, 'timing/model': 1.455614173959475, 'learning_rate/qwen_vl_interface': 2.4851910359353392e-06, 'learning_rate/action_model': 2.485191035935339e-05, 'epoch': 1.51}) +[2026-06-16 04:24:37,574][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.016963893547654152, 'train/grad_norm_pre_clip': 2.393132209777832, 'timing/data': 0.006659496983047575, 'timing/model': 1.4582629990181886, 'learning_rate/qwen_vl_interface': 2.4827152435756185e-06, 'learning_rate/action_model': 2.4827152435756185e-05, 'epoch': 1.51}) +[2026-06-16 04:24:43,453][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.10483717173337936, 'train/grad_norm_pre_clip': 10.258024215698242, 'timing/data': 0.00659668497974053, 'timing/model': 1.4521505659795366, 'learning_rate/qwen_vl_interface': 2.4802405887421307e-06, 'learning_rate/action_model': 2.4802405887421306e-05, 'epoch': 1.52}) +[2026-06-16 04:24:49,380][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.07883793115615845, 'train/grad_norm_pre_clip': 14.169153213500977, 'timing/data': 0.006828025041613728, 'timing/model': 1.486770396004431, 'learning_rate/qwen_vl_interface': 2.4777670724521134e-06, 'learning_rate/action_model': 2.4777670724521134e-05, 'epoch': 1.52}) +[2026-06-16 04:24:55,271][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.08921187371015549, 'train/grad_norm_pre_clip': 8.040609359741211, 'timing/data': 0.007038555049803108, 'timing/model': 1.4923231540014967, 'learning_rate/qwen_vl_interface': 2.475294695722337e-06, 'learning_rate/action_model': 2.475294695722337e-05, 'epoch': 1.52}) +[2026-06-16 04:25:01,367][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.1684538722038269, 'train/grad_norm_pre_clip': 16.811771392822266, 'timing/data': 0.006862606038339436, 'timing/model': 1.7386848379974253, 'learning_rate/qwen_vl_interface': 2.4728234595691006e-06, 'learning_rate/action_model': 2.4728234595691005e-05, 'epoch': 1.52}) +[2026-06-16 04:25:07,154][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.087560273706913, 'train/grad_norm_pre_clip': 12.232685089111328, 'timing/data': 0.007698706001974642, 'timing/model': 1.4758720399695449, 'learning_rate/qwen_vl_interface': 2.4703533650082368e-06, 'learning_rate/action_model': 2.4703533650082366e-05, 'epoch': 1.52}) +[2026-06-16 04:25:13,155][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.14791949093341827, 'train/grad_norm_pre_clip': 8.829835891723633, 'timing/data': 0.006577104970347136, 'timing/model': 1.4604785930132493, 'learning_rate/qwen_vl_interface': 2.4678844130551103e-06, 'learning_rate/action_model': 2.4678844130551103e-05, 'epoch': 1.52}) +[2026-06-16 04:25:19,061][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.03286856785416603, 'train/grad_norm_pre_clip': 6.538578987121582, 'timing/data': 0.006439042976126075, 'timing/model': 1.4657007249770686, 'learning_rate/qwen_vl_interface': 2.4654166047246102e-06, 'learning_rate/action_model': 2.46541660472461e-05, 'epoch': 1.52}) +[2026-06-16 04:25:25,069][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.0883861780166626, 'train/grad_norm_pre_clip': 8.881770133972168, 'timing/data': 0.00719822000246495, 'timing/model': 1.4940018590423279, 'learning_rate/qwen_vl_interface': 2.4629499410311606e-06, 'learning_rate/action_model': 2.462949941031161e-05, 'epoch': 1.52}) +[2026-06-16 04:25:30,972][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.012361641973257065, 'train/grad_norm_pre_clip': 3.045252561569214, 'timing/data': 0.007145421986933798, 'timing/model': 1.487101275997702, 'learning_rate/qwen_vl_interface': 2.460484422988715e-06, 'learning_rate/action_model': 2.4604844229887146e-05, 'epoch': 1.52}) +[2026-06-16 04:25:36,899][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.00597428297623992, 'train/grad_norm_pre_clip': 1.3132926225662231, 'timing/data': 0.006938798003830016, 'timing/model': 1.7049809390446171, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 1.52}) +[2026-06-16 04:25:42,562][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.009028165601193905, 'train/grad_norm_pre_clip': 2.9329652786254883, 'timing/data': 0.007225078996270895, 'timing/model': 1.464662693964783, 'learning_rate/qwen_vl_interface': 2.455556827910286e-06, 'learning_rate/action_model': 2.455556827910286e-05, 'epoch': 1.52}) +[2026-06-16 04:25:48,472][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.11975470185279846, 'train/grad_norm_pre_clip': 10.445469856262207, 'timing/data': 0.006664283981081098, 'timing/model': 1.4766445810091682, 'learning_rate/qwen_vl_interface': 2.453094752899853e-06, 'learning_rate/action_model': 2.453094752899853e-05, 'epoch': 1.52}) +[2026-06-16 04:25:54,367][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.06276893615722656, 'train/grad_norm_pre_clip': 5.511228561401367, 'timing/data': 0.006634810997638851, 'timing/model': 1.471463357040193, 'learning_rate/qwen_vl_interface': 2.4506338275915173e-06, 'learning_rate/action_model': 2.4506338275915172e-05, 'epoch': 1.52}) +[2026-06-16 04:26:00,273][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.056185074150562286, 'train/grad_norm_pre_clip': 6.030656814575195, 'timing/data': 0.006822727038525045, 'timing/model': 1.4500134499976411, 'learning_rate/qwen_vl_interface': 2.4481740529968735e-06, 'learning_rate/action_model': 2.4481740529968734e-05, 'epoch': 1.52}) +[2026-06-16 04:26:06,053][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.11981746554374695, 'train/grad_norm_pre_clip': 14.817272186279297, 'timing/data': 0.00691917899530381, 'timing/model': 1.461927838041447, 'learning_rate/qwen_vl_interface': 2.4457154301270427e-06, 'learning_rate/action_model': 2.4457154301270427e-05, 'epoch': 1.52}) +[2026-06-16 04:26:11,987][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.22868113219738007, 'train/grad_norm_pre_clip': 21.729835510253906, 'timing/data': 0.006758042029105127, 'timing/model': 1.6923757789772935, 'learning_rate/qwen_vl_interface': 2.443257959992671e-06, 'learning_rate/action_model': 2.443257959992671e-05, 'epoch': 1.52}) +[2026-06-16 04:26:17,658][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.003201462095603347, 'train/grad_norm_pre_clip': 0.6490975022315979, 'timing/data': 0.006468386040069163, 'timing/model': 1.4640674239490181, 'learning_rate/qwen_vl_interface': 2.440801643603933e-06, 'learning_rate/action_model': 2.440801643603933e-05, 'epoch': 1.52}) +[2026-06-16 04:26:23,654][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.1443960964679718, 'train/grad_norm_pre_clip': 13.296066284179688, 'timing/data': 0.006558728986419737, 'timing/model': 1.4587025599903427, 'learning_rate/qwen_vl_interface': 2.4383464819705252e-06, 'learning_rate/action_model': 2.438346481970525e-05, 'epoch': 1.52}) +[2026-06-16 04:26:29,554][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.13146936893463135, 'train/grad_norm_pre_clip': 11.61631965637207, 'timing/data': 0.006555843981914222, 'timing/model': 1.4579262069892138, 'learning_rate/qwen_vl_interface': 2.435892476101674e-06, 'learning_rate/action_model': 2.435892476101674e-05, 'epoch': 1.52}) +[2026-06-16 04:26:35,352][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.10612127929925919, 'train/grad_norm_pre_clip': 10.483083724975586, 'timing/data': 0.006160582997836173, 'timing/model': 1.4592392190243118, 'learning_rate/qwen_vl_interface': 2.433439627006128e-06, 'learning_rate/action_model': 2.4334396270061277e-05, 'epoch': 1.52}) +[2026-06-16 04:26:41,256][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.2983244061470032, 'train/grad_norm_pre_clip': 23.14071273803711, 'timing/data': 0.006585007009562105, 'timing/model': 1.44836167502217, 'learning_rate/qwen_vl_interface': 2.4309879356921602e-06, 'learning_rate/action_model': 2.43098793569216e-05, 'epoch': 1.52}) +[2026-06-16 04:26:47,198][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.24633681774139404, 'train/grad_norm_pre_clip': 16.934511184692383, 'timing/data': 0.006569765042513609, 'timing/model': 1.7011577439843677, 'learning_rate/qwen_vl_interface': 2.428537403167569e-06, 'learning_rate/action_model': 2.4285374031675687e-05, 'epoch': 1.52}) +[2026-06-16 04:26:52,862][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.11361150443553925, 'train/grad_norm_pre_clip': 10.227313041687012, 'timing/data': 0.006598296982701868, 'timing/model': 1.4652034700266086, 'learning_rate/qwen_vl_interface': 2.4260880304396765e-06, 'learning_rate/action_model': 2.4260880304396765e-05, 'epoch': 1.52}) +[2026-06-16 04:26:58,854][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.24908298254013062, 'train/grad_norm_pre_clip': 14.490154266357422, 'timing/data': 0.006631616037338972, 'timing/model': 1.4574221529765055, 'learning_rate/qwen_vl_interface': 2.4236398185153245e-06, 'learning_rate/action_model': 2.4236398185153245e-05, 'epoch': 1.52}) +[2026-06-16 04:27:04,756][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.11769693344831467, 'train/grad_norm_pre_clip': 9.718838691711426, 'timing/data': 0.006305056042037904, 'timing/model': 1.4561595539562404, 'learning_rate/qwen_vl_interface': 2.4211927684008823e-06, 'learning_rate/action_model': 2.4211927684008824e-05, 'epoch': 1.53}) +[2026-06-16 04:27:10,654][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.0013327362248674035, 'train/grad_norm_pre_clip': 0.21868939697742462, 'timing/data': 0.006400776037480682, 'timing/model': 1.4563193480134942, 'learning_rate/qwen_vl_interface': 2.4187468811022398e-06, 'learning_rate/action_model': 2.4187468811022396e-05, 'epoch': 1.53}) +[2026-06-16 04:27:16,627][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.04953666403889656, 'train/grad_norm_pre_clip': 4.05333137512207, 'timing/data': 0.0062620220123790205, 'timing/model': 1.4993971009971574, 'learning_rate/qwen_vl_interface': 2.416302157624809e-06, 'learning_rate/action_model': 2.4163021576248088e-05, 'epoch': 1.53}) +[2026-06-16 04:27:22,686][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.09641324728727341, 'train/grad_norm_pre_clip': 7.496089458465576, 'timing/data': 0.006473614019341767, 'timing/model': 1.6906979299965315, 'learning_rate/qwen_vl_interface': 2.4138585989735227e-06, 'learning_rate/action_model': 2.4138585989735226e-05, 'epoch': 1.53}) +[2026-06-16 04:27:28,352][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.01624372974038124, 'train/grad_norm_pre_clip': 2.1360294818878174, 'timing/data': 0.006415948970243335, 'timing/model': 1.4568123620119877, 'learning_rate/qwen_vl_interface': 2.4114162061528357e-06, 'learning_rate/action_model': 2.4114162061528354e-05, 'epoch': 1.53}) +[2026-06-16 04:27:34,255][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.24298784136772156, 'train/grad_norm_pre_clip': 14.081682205200195, 'timing/data': 0.00585338199744001, 'timing/model': 1.4458335799863562, 'learning_rate/qwen_vl_interface': 2.408974980166724e-06, 'learning_rate/action_model': 2.408974980166724e-05, 'epoch': 1.53}) +[2026-06-16 04:27:40,153][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.10839226096868515, 'train/grad_norm_pre_clip': 9.519180297851562, 'timing/data': 0.006060575949959457, 'timing/model': 1.4580440220306627, 'learning_rate/qwen_vl_interface': 2.406534922018683e-06, 'learning_rate/action_model': 2.4065349220186827e-05, 'epoch': 1.53}) +[2026-06-16 04:27:46,054][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.015408987179398537, 'train/grad_norm_pre_clip': 1.4291068315505981, 'timing/data': 0.00694467598805204, 'timing/model': 1.4590995850157924, 'learning_rate/qwen_vl_interface': 2.4040960327117287e-06, 'learning_rate/action_model': 2.4040960327117283e-05, 'epoch': 1.53}) +[2026-06-16 04:27:52,038][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.08557241410017014, 'train/grad_norm_pre_clip': 9.460671424865723, 'timing/data': 0.005855623981915414, 'timing/model': 1.5060449369484559, 'learning_rate/qwen_vl_interface': 2.4016583132483976e-06, 'learning_rate/action_model': 2.4016583132483976e-05, 'epoch': 1.53}) +[2026-06-16 04:27:57,987][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.14021062850952148, 'train/grad_norm_pre_clip': 8.084745407104492, 'timing/data': 0.006698113982565701, 'timing/model': 1.6873452229774557, 'learning_rate/qwen_vl_interface': 2.399221764630741e-06, 'learning_rate/action_model': 2.3992217646307406e-05, 'epoch': 1.53}) +[2026-06-16 04:28:03,657][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.01749745011329651, 'train/grad_norm_pre_clip': 2.692739486694336, 'timing/data': 0.006689390982501209, 'timing/model': 1.4595862080459483, 'learning_rate/qwen_vl_interface': 2.3967863878603334e-06, 'learning_rate/action_model': 2.3967863878603332e-05, 'epoch': 1.53}) +[2026-06-16 04:28:09,635][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.081162229180336, 'train/grad_norm_pre_clip': 7.665642738342285, 'timing/data': 0.007246960944030434, 'timing/model': 1.53865300596226, 'learning_rate/qwen_vl_interface': 2.394352183938266e-06, 'learning_rate/action_model': 2.3943521839382664e-05, 'epoch': 1.53}) +[2026-06-16 04:28:15,554][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.11650365591049194, 'train/grad_norm_pre_clip': 10.185751914978027, 'timing/data': 0.006725003011524677, 'timing/model': 1.4579450419987552, 'learning_rate/qwen_vl_interface': 2.39191915386515e-06, 'learning_rate/action_model': 2.39191915386515e-05, 'epoch': 1.53}) +[2026-06-16 04:28:21,553][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.07772820442914963, 'train/grad_norm_pre_clip': 11.687755584716797, 'timing/data': 0.006482436030637473, 'timing/model': 1.4807983220089227, 'learning_rate/qwen_vl_interface': 2.3894872986411095e-06, 'learning_rate/action_model': 2.3894872986411094e-05, 'epoch': 1.53}) +[2026-06-16 04:28:27,550][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.2043478935956955, 'train/grad_norm_pre_clip': 37.6746711730957, 'timing/data': 0.0065701749990694225, 'timing/model': 1.522005603008438, 'learning_rate/qwen_vl_interface': 2.3870566192657903e-06, 'learning_rate/action_model': 2.38705661926579e-05, 'epoch': 1.53}) +[2026-06-16 04:28:33,580][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.00520610436797142, 'train/grad_norm_pre_clip': 1.418832540512085, 'timing/data': 0.006803318974561989, 'timing/model': 1.782032106013503, 'learning_rate/qwen_vl_interface': 2.3846271167383515e-06, 'learning_rate/action_model': 2.3846271167383515e-05, 'epoch': 1.53}) +[2026-06-16 04:28:39,397][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.01947759836912155, 'train/grad_norm_pre_clip': 5.281049728393555, 'timing/data': 0.006230696046259254, 'timing/model': 1.6042990270070732, 'learning_rate/qwen_vl_interface': 2.382198792057471e-06, 'learning_rate/action_model': 2.3821987920574707e-05, 'epoch': 1.53}) +[2026-06-16 04:28:45,053][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.13517126441001892, 'train/grad_norm_pre_clip': 13.534915924072266, 'timing/data': 0.0070944170001894236, 'timing/model': 1.4557815970038064, 'learning_rate/qwen_vl_interface': 2.3797716462213402e-06, 'learning_rate/action_model': 2.3797716462213402e-05, 'epoch': 1.53}) +[2026-06-16 04:28:50,965][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.032755348831415176, 'train/grad_norm_pre_clip': 3.2204580307006836, 'timing/data': 0.006500262999907136, 'timing/model': 1.4728225159924477, 'learning_rate/qwen_vl_interface': 2.37734568022767e-06, 'learning_rate/action_model': 2.3773456802276697e-05, 'epoch': 1.53}) +[2026-06-16 04:28:56,857][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.060770563781261444, 'train/grad_norm_pre_clip': 9.09151840209961, 'timing/data': 0.006570175988599658, 'timing/model': 1.4557753569679335, 'learning_rate/qwen_vl_interface': 2.3749208950736783e-06, 'learning_rate/action_model': 2.374920895073678e-05, 'epoch': 1.53}) +[2026-06-16 04:29:02,792][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.12277445197105408, 'train/grad_norm_pre_clip': 10.32218074798584, 'timing/data': 0.006392784998752177, 'timing/model': 1.4635782759869471, 'learning_rate/qwen_vl_interface': 2.372497291756106e-06, 'learning_rate/action_model': 2.3724972917561058e-05, 'epoch': 1.53}) +[2026-06-16 04:29:08,686][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.055746130645275116, 'train/grad_norm_pre_clip': 9.011455535888672, 'timing/data': 0.006709831010084599, 'timing/model': 1.6890473490348086, 'learning_rate/qwen_vl_interface': 2.3700748712712043e-06, 'learning_rate/action_model': 2.370074871271204e-05, 'epoch': 1.53}) +[2026-06-16 04:29:14,586][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.0854092687368393, 'train/grad_norm_pre_clip': 7.828191757202148, 'timing/data': 0.006573260005097836, 'timing/model': 1.688960561004933, 'learning_rate/qwen_vl_interface': 2.3676536346147388e-06, 'learning_rate/action_model': 2.3676536346147388e-05, 'epoch': 1.53}) +[2026-06-16 04:29:20,256][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.04400196671485901, 'train/grad_norm_pre_clip': 9.091300964355469, 'timing/data': 0.006380948005244136, 'timing/model': 1.4587795740226284, 'learning_rate/qwen_vl_interface': 2.365233582781988e-06, 'learning_rate/action_model': 2.365233582781988e-05, 'epoch': 1.54}) +[2026-06-16 04:29:26,357][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.17499208450317383, 'train/grad_norm_pre_clip': 10.93155574798584, 'timing/data': 0.006810820021200925, 'timing/model': 1.4468368039815687, 'learning_rate/qwen_vl_interface': 2.3628147167677444e-06, 'learning_rate/action_model': 2.3628147167677443e-05, 'epoch': 1.54}) +[2026-06-16 04:29:32,353][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.06389990448951721, 'train/grad_norm_pre_clip': 4.740839958190918, 'timing/data': 0.006504017976112664, 'timing/model': 1.4676255670492537, 'learning_rate/qwen_vl_interface': 2.3603970375663123e-06, 'learning_rate/action_model': 2.3603970375663125e-05, 'epoch': 1.54}) +[2026-06-16 04:29:38,339][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.09309953451156616, 'train/grad_norm_pre_clip': 7.511288642883301, 'timing/data': 0.006377601996064186, 'timing/model': 1.5017705979989842, 'learning_rate/qwen_vl_interface': 2.35798054617151e-06, 'learning_rate/action_model': 2.3579805461715096e-05, 'epoch': 1.54}) +[2026-06-16 04:29:44,281][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.049122270196676254, 'train/grad_norm_pre_clip': 9.741758346557617, 'timing/data': 0.006504078977741301, 'timing/model': 1.6907797240419313, 'learning_rate/qwen_vl_interface': 2.355565243576662e-06, 'learning_rate/action_model': 2.3555652435766617e-05, 'epoch': 1.54}) +[2026-06-16 04:29:50,131][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.1208818331360817, 'train/grad_norm_pre_clip': 9.668730735778809, 'timing/data': 0.006466913037002087, 'timing/model': 1.639021501003299, 'learning_rate/qwen_vl_interface': 2.35315113077461e-06, 'learning_rate/action_model': 2.3531511307746098e-05, 'epoch': 1.54}) +[2026-06-16 04:29:55,862][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.041550714522600174, 'train/grad_norm_pre_clip': 6.818750381469727, 'timing/data': 0.006611396966036409, 'timing/model': 1.4708427769946866, 'learning_rate/qwen_vl_interface': 2.3507382087577047e-06, 'learning_rate/action_model': 2.350738208757705e-05, 'epoch': 1.54}) +[2026-06-16 04:30:01,763][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.0411289781332016, 'train/grad_norm_pre_clip': 4.767451286315918, 'timing/data': 0.007329403015319258, 'timing/model': 1.4683991629863158, 'learning_rate/qwen_vl_interface': 2.348326478517807e-06, 'learning_rate/action_model': 2.3483264785178072e-05, 'epoch': 1.54}) +[2026-06-16 04:30:07,854][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.012685923837125301, 'train/grad_norm_pre_clip': 2.577816963195801, 'timing/data': 0.0075201730360277, 'timing/model': 1.4461156699690036, 'learning_rate/qwen_vl_interface': 2.3459159410462907e-06, 'learning_rate/action_model': 2.3459159410462905e-05, 'epoch': 1.54}) +[2026-06-16 04:30:13,761][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.0940389558672905, 'train/grad_norm_pre_clip': 12.436895370483398, 'timing/data': 0.006669411959592253, 'timing/model': 1.4340173809905536, 'learning_rate/qwen_vl_interface': 2.343506597334031e-06, 'learning_rate/action_model': 2.343506597334031e-05, 'epoch': 1.54}) +[2026-06-16 04:30:19,737][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.06971113383769989, 'train/grad_norm_pre_clip': 6.433400630950928, 'timing/data': 0.006699124991428107, 'timing/model': 1.5061914270045236, 'learning_rate/qwen_vl_interface': 2.341098448371423e-06, 'learning_rate/action_model': 2.3410984483714227e-05, 'epoch': 1.54}) +[2026-06-16 04:30:25,649][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.07817390561103821, 'train/grad_norm_pre_clip': 5.645942211151123, 'timing/data': 0.006555614003445953, 'timing/model': 1.650826352997683, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 1.54}) +[2026-06-16 04:30:31,354][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.11484644562005997, 'train/grad_norm_pre_clip': 6.719059944152832, 'timing/data': 0.0063077189843170345, 'timing/model': 1.460660143988207, 'learning_rate/qwen_vl_interface': 2.33628573865426e-06, 'learning_rate/action_model': 2.3362857386542596e-05, 'epoch': 1.54}) +[2026-06-16 04:30:37,354][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.027875715866684914, 'train/grad_norm_pre_clip': 3.387450933456421, 'timing/data': 0.006531468010507524, 'timing/model': 1.4618805689970031, 'learning_rate/qwen_vl_interface': 2.3338811798780282e-06, 'learning_rate/action_model': 2.333881179878028e-05, 'epoch': 1.54}) +[2026-06-16 04:30:43,254][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.17195838689804077, 'train/grad_norm_pre_clip': 10.894015312194824, 'timing/data': 0.006588101969100535, 'timing/model': 1.4565262319520116, 'learning_rate/qwen_vl_interface': 2.3314778198080933e-06, 'learning_rate/action_model': 2.331477819808093e-05, 'epoch': 1.54}) +[2026-06-16 04:30:49,238][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.0871901735663414, 'train/grad_norm_pre_clip': 7.7030744552612305, 'timing/data': 0.006551417987793684, 'timing/model': 1.5017723040073179, 'learning_rate/qwen_vl_interface': 2.3290756594323834e-06, 'learning_rate/action_model': 2.3290756594323836e-05, 'epoch': 1.54}) +[2026-06-16 04:30:55,186][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.1157042533159256, 'train/grad_norm_pre_clip': 10.64333724975586, 'timing/data': 0.006383611005730927, 'timing/model': 1.6833813340053894, 'learning_rate/qwen_vl_interface': 2.3266746997383378e-06, 'learning_rate/action_model': 2.3266746997383377e-05, 'epoch': 1.54}) +[2026-06-16 04:31:00,852][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.1463213860988617, 'train/grad_norm_pre_clip': 7.954274654388428, 'timing/data': 0.006100705009885132, 'timing/model': 1.4591919219819829, 'learning_rate/qwen_vl_interface': 2.324274941712899e-06, 'learning_rate/action_model': 2.324274941712899e-05, 'epoch': 1.54}) +[2026-06-16 04:31:06,799][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.07157783210277557, 'train/grad_norm_pre_clip': 9.072822570800781, 'timing/data': 0.006620328989811242, 'timing/model': 1.50526990898652, 'learning_rate/qwen_vl_interface': 2.321876386342519e-06, 'learning_rate/action_model': 2.3218763863425187e-05, 'epoch': 1.54}) +[2026-06-16 04:31:12,656][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.24914675951004028, 'train/grad_norm_pre_clip': 23.4088077545166, 'timing/data': 0.006481765012722462, 'timing/model': 1.4645804180181585, 'learning_rate/qwen_vl_interface': 2.31947903461315e-06, 'learning_rate/action_model': 2.3194790346131502e-05, 'epoch': 1.54}) +[2026-06-16 04:31:18,558][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.007520870305597782, 'train/grad_norm_pre_clip': 1.9577624797821045, 'timing/data': 0.006471280998084694, 'timing/model': 1.46981565997703, 'learning_rate/qwen_vl_interface': 2.3170828875102544e-06, 'learning_rate/action_model': 2.3170828875102543e-05, 'epoch': 1.54}) +[2026-06-16 04:31:24,452][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.04088226705789566, 'train/grad_norm_pre_clip': 5.345946311950684, 'timing/data': 0.007406695978716016, 'timing/model': 1.4607860049582087, 'learning_rate/qwen_vl_interface': 2.3146879460187984e-06, 'learning_rate/action_model': 2.3146879460187982e-05, 'epoch': 1.54}) +[2026-06-16 04:31:30,428][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.14419317245483398, 'train/grad_norm_pre_clip': 12.656150817871094, 'timing/data': 0.006574030965566635, 'timing/model': 1.7221061699674465, 'learning_rate/qwen_vl_interface': 2.3122942111232506e-06, 'learning_rate/action_model': 2.3122942111232506e-05, 'epoch': 1.54}) +[2026-06-16 04:31:36,150][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.038783200085163116, 'train/grad_norm_pre_clip': 1.9342542886734009, 'timing/data': 0.006804689997807145, 'timing/model': 1.4565607180120423, 'learning_rate/qwen_vl_interface': 2.3099016838075857e-06, 'learning_rate/action_model': 2.3099016838075856e-05, 'epoch': 1.55}) +[2026-06-16 04:31:42,155][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.010255986824631691, 'train/grad_norm_pre_clip': 1.81168794631958, 'timing/data': 0.0068104389938525856, 'timing/model': 1.4590868249651976, 'learning_rate/qwen_vl_interface': 2.307510365055282e-06, 'learning_rate/action_model': 2.307510365055282e-05, 'epoch': 1.55}) +[2026-06-16 04:31:48,056][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.17320393025875092, 'train/grad_norm_pre_clip': 11.327132225036621, 'timing/data': 0.006644945009611547, 'timing/model': 1.4607941550202668, 'learning_rate/qwen_vl_interface': 2.3051202558493196e-06, 'learning_rate/action_model': 2.3051202558493192e-05, 'epoch': 1.55}) +[2026-06-16 04:31:53,963][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.13499221205711365, 'train/grad_norm_pre_clip': 7.76375675201416, 'timing/data': 0.007376210996881127, 'timing/model': 1.4687001860002056, 'learning_rate/qwen_vl_interface': 2.302731357172182e-06, 'learning_rate/action_model': 2.302731357172182e-05, 'epoch': 1.55}) +[2026-06-16 04:31:59,856][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.08212186396121979, 'train/grad_norm_pre_clip': 8.40322494506836, 'timing/data': 0.006683322018943727, 'timing/model': 1.4696136089623906, 'learning_rate/qwen_vl_interface': 2.3003436700058556e-06, 'learning_rate/action_model': 2.3003436700058554e-05, 'epoch': 1.55}) +[2026-06-16 04:32:05,791][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.08782265335321426, 'train/grad_norm_pre_clip': 6.9311065673828125, 'timing/data': 0.0067577119916677475, 'timing/model': 1.6946016639703885, 'learning_rate/qwen_vl_interface': 2.297957195331828e-06, 'learning_rate/action_model': 2.2979571953318276e-05, 'epoch': 1.55}) +[2026-06-16 04:32:11,455][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.08605314046144485, 'train/grad_norm_pre_clip': 5.400221824645996, 'timing/data': 0.006846512027550489, 'timing/model': 1.4597142760176212, 'learning_rate/qwen_vl_interface': 2.2955719341310903e-06, 'learning_rate/action_model': 2.2955719341310903e-05, 'epoch': 1.55}) +[2026-06-16 04:32:17,351][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.058102093636989594, 'train/grad_norm_pre_clip': 8.796655654907227, 'timing/data': 0.0067855530069209635, 'timing/model': 1.4594989589531906, 'learning_rate/qwen_vl_interface': 2.2931878873841304e-06, 'learning_rate/action_model': 2.2931878873841302e-05, 'epoch': 1.55}) +[2026-06-16 04:32:23,353][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.014055001549422741, 'train/grad_norm_pre_clip': 1.65101158618927, 'timing/data': 0.006552178994752467, 'timing/model': 1.4754443360143341, 'learning_rate/qwen_vl_interface': 2.290805056070942e-06, 'learning_rate/action_model': 2.290805056070942e-05, 'epoch': 1.55}) +[2026-06-16 04:41:31,756][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.06681232154369354, 'train/grad_norm_pre_clip': 7.969670295715332, 'eval/loss': 0.05014069750905037, 'eval/flap_precision': 0.9325396825396826, 'eval/flap_recall': 0.9418837675350702, 'eval/flap_f1': 0.9371884346959123, 'eval/flap_rate_model': 0.14, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9371884346959122, 'timing/data': 0.00662521697813645, 'timing/model': 1.5088086280156858, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 1.55}) +[2026-06-16 04:41:31,756][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3600_state +[2026-06-16 04:41:31,760][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 04:41:37,127][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3600_state/model.safetensors +[2026-06-16 04:41:50,264][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3600_state/optimizer.bin +[2026-06-16 04:41:50,265][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3600_state/sampler.bin +[2026-06-16 04:41:50,265][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3600_state/sampler_1.bin +[2026-06-16 04:41:50,266][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3600_state/random_states_0.pkl +[2026-06-16 04:41:50,371][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3600 +[2026-06-16 04:41:50,372][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 04:41:50,381][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 04:41:56,436][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.15232743322849274, 'train/grad_norm_pre_clip': 9.168088912963867, 'timing/data': 0.006928040995262563, 'timing/model': 1.4953698099707253, 'learning_rate/qwen_vl_interface': 2.2860430436633447e-06, 'learning_rate/action_model': 2.2860430436633447e-05, 'epoch': 1.55}) +[2026-06-16 04:42:02,476][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.018669795244932175, 'train/grad_norm_pre_clip': 3.264676094055176, 'timing/data': 0.006740611046552658, 'timing/model': 1.5296585370087996, 'learning_rate/qwen_vl_interface': 2.28366386452642e-06, 'learning_rate/action_model': 2.28366386452642e-05, 'epoch': 1.55}) +[2026-06-16 04:42:08,487][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.17620815336704254, 'train/grad_norm_pre_clip': 20.223865509033203, 'timing/data': 0.006628495000768453, 'timing/model': 1.6646032329881564, 'learning_rate/qwen_vl_interface': 2.2812859047382326e-06, 'learning_rate/action_model': 2.2812859047382323e-05, 'epoch': 1.55}) +[2026-06-16 04:42:14,329][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.007577486801892519, 'train/grad_norm_pre_clip': 1.3098090887069702, 'timing/data': 0.015403837023768574, 'timing/model': 1.5291753600467928, 'learning_rate/qwen_vl_interface': 2.2789091652762712e-06, 'learning_rate/action_model': 2.2789091652762712e-05, 'epoch': 1.55}) +[2026-06-16 04:42:20,623][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.20774123072624207, 'train/grad_norm_pre_clip': 16.079830169677734, 'timing/data': 0.006738157011568546, 'timing/model': 1.5698647529934533, 'learning_rate/qwen_vl_interface': 2.276533647117524e-06, 'learning_rate/action_model': 2.2765336471175242e-05, 'epoch': 1.55}) +[2026-06-16 04:42:26,755][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.013191103003919125, 'train/grad_norm_pre_clip': 1.9124670028686523, 'timing/data': 0.007150063989683986, 'timing/model': 1.4614900479791686, 'learning_rate/qwen_vl_interface': 2.2741593512384773e-06, 'learning_rate/action_model': 2.2741593512384772e-05, 'epoch': 1.55}) +[2026-06-16 04:42:32,734][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.3156696856021881, 'train/grad_norm_pre_clip': 21.883142471313477, 'timing/data': 0.019564970978535712, 'timing/model': 1.5193255899939686, 'learning_rate/qwen_vl_interface': 2.2717862786151153e-06, 'learning_rate/action_model': 2.271786278615115e-05, 'epoch': 1.55}) +[2026-06-16 04:42:38,751][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.12816785275936127, 'train/grad_norm_pre_clip': 11.80445671081543, 'timing/data': 0.006402249971870333, 'timing/model': 1.4988429310033098, 'learning_rate/qwen_vl_interface': 2.2694144302229152e-06, 'learning_rate/action_model': 2.2694144302229152e-05, 'epoch': 1.55}) +[2026-06-16 04:42:44,792][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.05631749704480171, 'train/grad_norm_pre_clip': 6.138664245605469, 'timing/data': 0.006707350956276059, 'timing/model': 1.7101115050027147, 'learning_rate/qwen_vl_interface': 2.2670438070368568e-06, 'learning_rate/action_model': 2.2670438070368566e-05, 'epoch': 1.55}) +[2026-06-16 04:42:50,456][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.012851422652602196, 'train/grad_norm_pre_clip': 1.8059288263320923, 'timing/data': 0.006755971990060061, 'timing/model': 1.4555937240365893, 'learning_rate/qwen_vl_interface': 2.264674410031413e-06, 'learning_rate/action_model': 2.2646744100314127e-05, 'epoch': 1.55}) +[2026-06-16 04:42:56,489][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.1601925790309906, 'train/grad_norm_pre_clip': 14.975642204284668, 'timing/data': 0.00683117302833125, 'timing/model': 1.4918488609837368, 'learning_rate/qwen_vl_interface': 2.2623062401805537e-06, 'learning_rate/action_model': 2.2623062401805535e-05, 'epoch': 1.55}) +[2026-06-16 04:43:02,433][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.056893058121204376, 'train/grad_norm_pre_clip': 12.513798713684082, 'timing/data': 0.006617277045734227, 'timing/model': 1.5320305859786458, 'learning_rate/qwen_vl_interface': 2.2599392984577436e-06, 'learning_rate/action_model': 2.2599392984577438e-05, 'epoch': 1.55}) +[2026-06-16 04:43:08,561][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.259249746799469, 'train/grad_norm_pre_clip': 19.723709106445312, 'timing/data': 0.0065725999884307384, 'timing/model': 1.4621689670020714, 'learning_rate/qwen_vl_interface': 2.2575735858359436e-06, 'learning_rate/action_model': 2.257573585835944e-05, 'epoch': 1.55}) +[2026-06-16 04:43:14,561][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.06229385361075401, 'train/grad_norm_pre_clip': 8.790684700012207, 'timing/data': 0.006478191004134715, 'timing/model': 1.5284641219768673, 'learning_rate/qwen_vl_interface': 2.255209103287609e-06, 'learning_rate/action_model': 2.255209103287609e-05, 'epoch': 1.56}) +[2026-06-16 04:43:20,610][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.08260594308376312, 'train/grad_norm_pre_clip': 12.931699752807617, 'timing/data': 0.007670229009818286, 'timing/model': 1.7039530740003102, 'learning_rate/qwen_vl_interface': 2.2528458517846886e-06, 'learning_rate/action_model': 2.252845851784689e-05, 'epoch': 1.56}) +[2026-06-16 04:43:26,445][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.1065930500626564, 'train/grad_norm_pre_clip': 8.982416152954102, 'timing/data': 0.0066409109858796, 'timing/model': 1.5081567110028118, 'learning_rate/qwen_vl_interface': 2.250483832298627e-06, 'learning_rate/action_model': 2.250483832298627e-05, 'epoch': 1.56}) +[2026-06-16 04:43:32,355][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.25189730525016785, 'train/grad_norm_pre_clip': 18.422771453857422, 'timing/data': 0.007394531974568963, 'timing/model': 1.4315759009914473, 'learning_rate/qwen_vl_interface': 2.2481230458003608e-06, 'learning_rate/action_model': 2.2481230458003606e-05, 'epoch': 1.56}) +[2026-06-16 04:43:38,348][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.3933747410774231, 'train/grad_norm_pre_clip': 22.44167709350586, 'timing/data': 0.008484517980832607, 'timing/model': 1.4472856349893846, 'learning_rate/qwen_vl_interface': 2.2457634932603184e-06, 'learning_rate/action_model': 2.2457634932603183e-05, 'epoch': 1.56}) +[2026-06-16 04:43:44,259][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.16763320565223694, 'train/grad_norm_pre_clip': 11.31347942352295, 'timing/data': 0.006645377026870847, 'timing/model': 1.4589618730242364, 'learning_rate/qwen_vl_interface': 2.2434051756484255e-06, 'learning_rate/action_model': 2.2434051756484253e-05, 'epoch': 1.56}) +[2026-06-16 04:43:50,270][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.035499949008226395, 'train/grad_norm_pre_clip': 4.702183246612549, 'timing/data': 0.006664654996711761, 'timing/model': 1.5090054709580727, 'learning_rate/qwen_vl_interface': 2.2410480939340945e-06, 'learning_rate/action_model': 2.2410480939340945e-05, 'epoch': 1.56}) +[2026-06-16 04:43:56,184][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.07731465995311737, 'train/grad_norm_pre_clip': 11.103549003601074, 'timing/data': 0.006144218030385673, 'timing/model': 1.680458213959355, 'learning_rate/qwen_vl_interface': 2.238692249086234e-06, 'learning_rate/action_model': 2.2386922490862336e-05, 'epoch': 1.56}) +[2026-06-16 04:44:01,849][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.10143276304006577, 'train/grad_norm_pre_clip': 8.886542320251465, 'timing/data': 0.006500792980659753, 'timing/model': 1.4566484700189903, 'learning_rate/qwen_vl_interface': 2.2363376420732426e-06, 'learning_rate/action_model': 2.2363376420732423e-05, 'epoch': 1.56}) +[2026-06-16 04:44:08,123][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.05067699775099754, 'train/grad_norm_pre_clip': 4.566769123077393, 'timing/data': 0.005942189018242061, 'timing/model': 1.6324994250317104, 'learning_rate/qwen_vl_interface': 2.233984273863011e-06, 'learning_rate/action_model': 2.233984273863011e-05, 'epoch': 1.56}) +[2026-06-16 04:44:14,224][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.0748201236128807, 'train/grad_norm_pre_clip': 6.665594577789307, 'timing/data': 0.0058409899938851595, 'timing/model': 1.5430316130514257, 'learning_rate/qwen_vl_interface': 2.231632145422919e-06, 'learning_rate/action_model': 2.231632145422919e-05, 'epoch': 1.56}) +[2026-06-16 04:44:20,059][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.08706508576869965, 'train/grad_norm_pre_clip': 9.91029167175293, 'timing/data': 0.006778963026590645, 'timing/model': 1.4580321160028689, 'learning_rate/qwen_vl_interface': 2.2292812577198383e-06, 'learning_rate/action_model': 2.229281257719838e-05, 'epoch': 1.56}) +[2026-06-16 04:44:26,011][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.21706008911132812, 'train/grad_norm_pre_clip': 20.19635772705078, 'timing/data': 0.006710081011988223, 'timing/model': 1.4836577160167508, 'learning_rate/qwen_vl_interface': 2.2269316117201307e-06, 'learning_rate/action_model': 2.2269316117201307e-05, 'epoch': 1.56}) +[2026-06-16 04:44:31,893][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.03605932742357254, 'train/grad_norm_pre_clip': 6.8457231521606445, 'timing/data': 0.006710471003316343, 'timing/model': 1.6936485470505431, 'learning_rate/qwen_vl_interface': 2.224583208389646e-06, 'learning_rate/action_model': 2.224583208389646e-05, 'epoch': 1.56}) +[2026-06-16 04:44:37,712][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.040522512048482895, 'train/grad_norm_pre_clip': 7.285869121551514, 'timing/data': 0.006706946005579084, 'timing/model': 1.6073832080001011, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 1.56}) +[2026-06-16 04:44:43,460][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.044913530349731445, 'train/grad_norm_pre_clip': 7.010931968688965, 'timing/data': 0.006487211969215423, 'timing/model': 1.457438685989473, 'learning_rate/qwen_vl_interface': 2.219890133597194e-06, 'learning_rate/action_model': 2.219890133597194e-05, 'epoch': 1.56}) +[2026-06-16 04:44:49,429][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.12701915204524994, 'train/grad_norm_pre_clip': 23.62538719177246, 'timing/data': 0.007252079027239233, 'timing/model': 1.5238586920313537, 'learning_rate/qwen_vl_interface': 2.217545464064374e-06, 'learning_rate/action_model': 2.217545464064374e-05, 'epoch': 1.56}) +[2026-06-16 04:44:55,462][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.10044494271278381, 'train/grad_norm_pre_clip': 7.584471225738525, 'timing/data': 0.00862486200639978, 'timing/model': 1.4723134380183183, 'learning_rate/qwen_vl_interface': 2.215202041059067e-06, 'learning_rate/action_model': 2.215202041059067e-05, 'epoch': 1.56}) +[2026-06-16 04:45:01,542][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.04062081128358841, 'train/grad_norm_pre_clip': 7.017831325531006, 'timing/data': 0.006459900992922485, 'timing/model': 1.583493729995098, 'learning_rate/qwen_vl_interface': 2.212859865544566e-06, 'learning_rate/action_model': 2.212859865544566e-05, 'epoch': 1.56}) +[2026-06-16 04:45:07,616][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.02379533089697361, 'train/grad_norm_pre_clip': 4.855329990386963, 'timing/data': 0.014873540960252285, 'timing/model': 1.7356654800241813, 'learning_rate/qwen_vl_interface': 2.210518938483652e-06, 'learning_rate/action_model': 2.210518938483652e-05, 'epoch': 1.56}) +[2026-06-16 04:45:13,702][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.013793883845210075, 'train/grad_norm_pre_clip': 3.5714688301086426, 'timing/data': 0.01844017702387646, 'timing/model': 1.7514663400361314, 'learning_rate/qwen_vl_interface': 2.20817926083859e-06, 'learning_rate/action_model': 2.20817926083859e-05, 'epoch': 1.56}) +[2026-06-16 04:45:19,363][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.14077633619308472, 'train/grad_norm_pre_clip': 11.0516939163208, 'timing/data': 0.007112252002116293, 'timing/model': 1.456553513999097, 'learning_rate/qwen_vl_interface': 2.205840833571135e-06, 'learning_rate/action_model': 2.205840833571135e-05, 'epoch': 1.56}) +[2026-06-16 04:45:25,460][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.13320770859718323, 'train/grad_norm_pre_clip': 18.668195724487305, 'timing/data': 0.00763660401571542, 'timing/model': 1.4580431790091097, 'learning_rate/qwen_vl_interface': 2.2035036576425264e-06, 'learning_rate/action_model': 2.203503657642526e-05, 'epoch': 1.56}) +[2026-06-16 04:45:31,462][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.04634242132306099, 'train/grad_norm_pre_clip': 4.73828649520874, 'timing/data': 0.007955222972668707, 'timing/model': 1.471279067976866, 'learning_rate/qwen_vl_interface': 2.2011677340134867e-06, 'learning_rate/action_model': 2.2011677340134867e-05, 'epoch': 1.56}) +[2026-06-16 04:45:37,560][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.048954546451568604, 'train/grad_norm_pre_clip': 4.554941177368164, 'timing/data': 0.018016181013081223, 'timing/model': 1.5113553100381978, 'learning_rate/qwen_vl_interface': 2.198833063644227e-06, 'learning_rate/action_model': 2.198833063644227e-05, 'epoch': 1.57}) +[2026-06-16 04:45:43,592][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.11718541383743286, 'train/grad_norm_pre_clip': 13.365484237670898, 'timing/data': 0.00644902401836589, 'timing/model': 1.7197548910044134, 'learning_rate/qwen_vl_interface': 2.196499647494443e-06, 'learning_rate/action_model': 2.196499647494443e-05, 'epoch': 1.57}) +[2026-06-16 04:45:49,442][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.034681536257267, 'train/grad_norm_pre_clip': 4.107337474822998, 'timing/data': 0.006696147960610688, 'timing/model': 1.6415350229945034, 'learning_rate/qwen_vl_interface': 2.1941674865233144e-06, 'learning_rate/action_model': 2.1941674865233143e-05, 'epoch': 1.57}) +[2026-06-16 04:45:55,159][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.03309239447116852, 'train/grad_norm_pre_clip': 11.833084106445312, 'timing/data': 0.00666896702023223, 'timing/model': 1.460448040976189, 'learning_rate/qwen_vl_interface': 2.1918365816895013e-06, 'learning_rate/action_model': 2.191836581689501e-05, 'epoch': 1.57}) +[2026-06-16 04:46:01,260][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.2009914070367813, 'train/grad_norm_pre_clip': 14.298479080200195, 'timing/data': 0.0069054560153745115, 'timing/model': 1.4455342380097136, 'learning_rate/qwen_vl_interface': 2.189506933951153e-06, 'learning_rate/action_model': 2.1895069339511525e-05, 'epoch': 1.57}) +[2026-06-16 04:46:07,160][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.08617321401834488, 'train/grad_norm_pre_clip': 8.101438522338867, 'timing/data': 0.006512266001664102, 'timing/model': 1.455098127014935, 'learning_rate/qwen_vl_interface': 2.187178544265898e-06, 'learning_rate/action_model': 2.1871785442658983e-05, 'epoch': 1.57}) +[2026-06-16 04:46:13,239][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.10925499349832535, 'train/grad_norm_pre_clip': 5.9018659591674805, 'timing/data': 0.00657576898811385, 'timing/model': 1.509788409981411, 'learning_rate/qwen_vl_interface': 2.184851413590852e-06, 'learning_rate/action_model': 2.1848514135908518e-05, 'epoch': 1.57}) +[2026-06-16 04:46:19,271][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.04055915027856827, 'train/grad_norm_pre_clip': 8.154779434204102, 'timing/data': 0.007424490002449602, 'timing/model': 1.5352886890177615, 'learning_rate/qwen_vl_interface': 2.1825255428826087e-06, 'learning_rate/action_model': 2.182525542882609e-05, 'epoch': 1.57}) +[2026-06-16 04:46:25,146][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.1663123220205307, 'train/grad_norm_pre_clip': 14.848770141601562, 'timing/data': 0.006785337987821549, 'timing/model': 1.645081154012587, 'learning_rate/qwen_vl_interface': 2.180200933097246e-06, 'learning_rate/action_model': 2.180200933097246e-05, 'epoch': 1.57}) +[2026-06-16 04:46:30,859][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.011821861378848553, 'train/grad_norm_pre_clip': 1.5306799411773682, 'timing/data': 0.006577662017662078, 'timing/model': 1.4592700600042008, 'learning_rate/qwen_vl_interface': 2.1778775851903235e-06, 'learning_rate/action_model': 2.177877585190323e-05, 'epoch': 1.57}) +[2026-06-16 04:46:37,062][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.1225859597325325, 'train/grad_norm_pre_clip': 15.674416542053223, 'timing/data': 0.022983664995990694, 'timing/model': 1.461020915012341, 'learning_rate/qwen_vl_interface': 2.175555500116881e-06, 'learning_rate/action_model': 2.175555500116881e-05, 'epoch': 1.57}) +[2026-06-16 04:46:43,062][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.05167451128363609, 'train/grad_norm_pre_clip': 8.355146408081055, 'timing/data': 0.0068008709931746125, 'timing/model': 1.4568615229800344, 'learning_rate/qwen_vl_interface': 2.1732346788314404e-06, 'learning_rate/action_model': 2.1732346788314402e-05, 'epoch': 1.57}) +[2026-06-16 04:46:49,061][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.37948092818260193, 'train/grad_norm_pre_clip': 26.19919204711914, 'timing/data': 0.006813188025262207, 'timing/model': 1.5020271679968573, 'learning_rate/qwen_vl_interface': 2.170915122288006e-06, 'learning_rate/action_model': 2.170915122288006e-05, 'epoch': 1.57}) +[2026-06-16 04:46:55,000][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.08633963763713837, 'train/grad_norm_pre_clip': 8.470074653625488, 'timing/data': 0.01089728600345552, 'timing/model': 1.6895055329659954, 'learning_rate/qwen_vl_interface': 2.168596831440055e-06, 'learning_rate/action_model': 2.1685968314400552e-05, 'epoch': 1.57}) +[2026-06-16 04:47:00,744][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.10409596562385559, 'train/grad_norm_pre_clip': 11.959803581237793, 'timing/data': 0.014548695995472372, 'timing/model': 1.5311453550239094, 'learning_rate/qwen_vl_interface': 2.166279807240553e-06, 'learning_rate/action_model': 2.1662798072405528e-05, 'epoch': 1.57}) +[2026-06-16 04:47:06,762][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.15106303989887238, 'train/grad_norm_pre_clip': 10.504322052001953, 'timing/data': 0.0067856270470656455, 'timing/model': 1.4612748399958946, 'learning_rate/qwen_vl_interface': 2.1639640506419394e-06, 'learning_rate/action_model': 2.163964050641939e-05, 'epoch': 1.57}) +[2026-06-16 04:47:12,558][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.11386309564113617, 'train/grad_norm_pre_clip': 11.369501113891602, 'timing/data': 0.006721012992784381, 'timing/model': 1.4596013440168463, 'learning_rate/qwen_vl_interface': 2.161649562596135e-06, 'learning_rate/action_model': 2.161649562596135e-05, 'epoch': 1.57}) +[2026-06-16 04:47:18,457][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.16728833317756653, 'train/grad_norm_pre_clip': 10.776662826538086, 'timing/data': 0.007261188002303243, 'timing/model': 1.4644870809861459, 'learning_rate/qwen_vl_interface': 2.1593363440545383e-06, 'learning_rate/action_model': 2.1593363440545384e-05, 'epoch': 1.57}) +[2026-06-16 04:47:24,485][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.07942201942205429, 'train/grad_norm_pre_clip': 4.582993984222412, 'timing/data': 0.006658168975263834, 'timing/model': 1.5008669790113345, 'learning_rate/qwen_vl_interface': 2.157024395968025e-06, 'learning_rate/action_model': 2.1570243959680252e-05, 'epoch': 1.57}) +[2026-06-16 04:47:30,394][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.30007898807525635, 'train/grad_norm_pre_clip': 11.864514350891113, 'timing/data': 0.006769662955775857, 'timing/model': 1.6932618059800006, 'learning_rate/qwen_vl_interface': 2.154713719286952e-06, 'learning_rate/action_model': 2.154713719286952e-05, 'epoch': 1.57}) +[2026-06-16 04:47:36,065][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.007780003827065229, 'train/grad_norm_pre_clip': 1.6449079513549805, 'timing/data': 0.006764785968698561, 'timing/model': 1.462595017044805, 'learning_rate/qwen_vl_interface': 2.152404314961149e-06, 'learning_rate/action_model': 2.152404314961149e-05, 'epoch': 1.57}) +[2026-06-16 04:47:42,063][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.008064771071076393, 'train/grad_norm_pre_clip': 1.7606589794158936, 'timing/data': 0.006558583001606166, 'timing/model': 1.463810048007872, 'learning_rate/qwen_vl_interface': 2.1500961839399264e-06, 'learning_rate/action_model': 2.1500961839399262e-05, 'epoch': 1.57}) +[2026-06-16 04:47:48,058][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.08653348684310913, 'train/grad_norm_pre_clip': 6.600050449371338, 'timing/data': 0.007155811996199191, 'timing/model': 1.4596445819479413, 'learning_rate/qwen_vl_interface': 2.1477893271720683e-06, 'learning_rate/action_model': 2.1477893271720682e-05, 'epoch': 1.57}) +[2026-06-16 04:47:53,959][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.13229435682296753, 'train/grad_norm_pre_clip': 7.963153839111328, 'timing/data': 0.006605911010410637, 'timing/model': 1.4524319280171767, 'learning_rate/qwen_vl_interface': 2.1454837456058385e-06, 'learning_rate/action_model': 2.1454837456058386e-05, 'epoch': 1.58}) +[2026-06-16 04:47:59,858][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.002493566134944558, 'train/grad_norm_pre_clip': 0.3670161962509155, 'timing/data': 0.006763754005078226, 'timing/model': 1.4717818360077217, 'learning_rate/qwen_vl_interface': 2.1431794401889722e-06, 'learning_rate/action_model': 2.143179440188972e-05, 'epoch': 1.58}) +[2026-06-16 04:48:05,793][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.05244407057762146, 'train/grad_norm_pre_clip': 5.861772060394287, 'timing/data': 0.00739097804762423, 'timing/model': 1.689158875029534, 'learning_rate/qwen_vl_interface': 2.1408764118686837e-06, 'learning_rate/action_model': 2.1408764118686838e-05, 'epoch': 1.58}) +[2026-06-16 04:48:11,470][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.009989460930228233, 'train/grad_norm_pre_clip': 1.4597676992416382, 'timing/data': 0.0064674270106479526, 'timing/model': 1.4647426820010878, 'learning_rate/qwen_vl_interface': 2.1385746615916598e-06, 'learning_rate/action_model': 2.1385746615916596e-05, 'epoch': 1.58}) +[2026-06-16 04:48:17,458][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.09349890798330307, 'train/grad_norm_pre_clip': 12.639076232910156, 'timing/data': 0.006842429982498288, 'timing/model': 1.457313780032564, 'learning_rate/qwen_vl_interface': 2.1362741903040636e-06, 'learning_rate/action_model': 2.1362741903040633e-05, 'epoch': 1.58}) +[2026-06-16 04:48:23,461][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.16811533272266388, 'train/grad_norm_pre_clip': 16.2879638671875, 'timing/data': 0.007040982018224895, 'timing/model': 1.4923118089791387, 'learning_rate/qwen_vl_interface': 2.1339749989515318e-06, 'learning_rate/action_model': 2.1339749989515316e-05, 'epoch': 1.58}) +[2026-06-16 04:48:29,445][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.17594976723194122, 'train/grad_norm_pre_clip': 14.13123607635498, 'timing/data': 0.0070996389840729535, 'timing/model': 1.513842387998011, 'learning_rate/qwen_vl_interface': 2.131677088479175e-06, 'learning_rate/action_model': 2.1316770884791752e-05, 'epoch': 1.58}) +[2026-06-16 04:48:35,361][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.028190631419420242, 'train/grad_norm_pre_clip': 2.722463846206665, 'timing/data': 0.006717615004163235, 'timing/model': 1.4558766940026544, 'learning_rate/qwen_vl_interface': 2.1293804598315787e-06, 'learning_rate/action_model': 2.1293804598315787e-05, 'epoch': 1.58}) +[2026-06-16 04:48:41,295][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.12753938138484955, 'train/grad_norm_pre_clip': 10.043922424316406, 'timing/data': 0.00678047799738124, 'timing/model': 1.6905836700461805, 'learning_rate/qwen_vl_interface': 2.1270851139527986e-06, 'learning_rate/action_model': 2.1270851139527983e-05, 'epoch': 1.58}) +[2026-06-16 04:48:47,022][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.08826061338186264, 'train/grad_norm_pre_clip': 7.673192977905273, 'timing/data': 0.006703754013869911, 'timing/model': 1.5111995859770104, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 1.58}) +[2026-06-16 04:48:52,961][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.15842565894126892, 'train/grad_norm_pre_clip': 12.993672370910645, 'timing/data': 0.006508847989607602, 'timing/model': 1.4625609560171142, 'learning_rate/qwen_vl_interface': 2.1224982742752814e-06, 'learning_rate/action_model': 2.1224982742752812e-05, 'epoch': 1.58}) +[2026-06-16 04:48:58,857][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.052293553948402405, 'train/grad_norm_pre_clip': 6.268005847930908, 'timing/data': 0.00635738397249952, 'timing/model': 1.4567857430083677, 'learning_rate/qwen_vl_interface': 2.1202067823620213e-06, 'learning_rate/action_model': 2.1202067823620212e-05, 'epoch': 1.58}) +[2026-06-16 04:49:04,758][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.06027995049953461, 'train/grad_norm_pre_clip': 4.645204067230225, 'timing/data': 0.006549537007231265, 'timing/model': 1.4601090240175836, 'learning_rate/qwen_vl_interface': 2.117916576988531e-06, 'learning_rate/action_model': 2.117916576988531e-05, 'epoch': 1.58}) +[2026-06-16 04:49:10,738][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.02223115973174572, 'train/grad_norm_pre_clip': 4.315081596374512, 'timing/data': 0.006999970006290823, 'timing/model': 1.5065063299844041, 'learning_rate/qwen_vl_interface': 2.1156276590962257e-06, 'learning_rate/action_model': 2.1156276590962257e-05, 'epoch': 1.58}) +[2026-06-16 04:49:16,685][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.03802505135536194, 'train/grad_norm_pre_clip': 5.966679573059082, 'timing/data': 0.006742611003573984, 'timing/model': 1.6865860499674454, 'learning_rate/qwen_vl_interface': 2.113340029625994e-06, 'learning_rate/action_model': 2.1133400296259938e-05, 'epoch': 1.58}) +[2026-06-16 04:49:22,362][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.19133003056049347, 'train/grad_norm_pre_clip': 24.685104370117188, 'timing/data': 0.0065729409689083695, 'timing/model': 1.4581765760085545, 'learning_rate/qwen_vl_interface': 2.111053689518194e-06, 'learning_rate/action_model': 2.111053689518194e-05, 'epoch': 1.58}) +[2026-06-16 04:49:28,359][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.09570316970348358, 'train/grad_norm_pre_clip': 11.362120628356934, 'timing/data': 0.006587622978258878, 'timing/model': 1.4590149369905703, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 1.58}) +[2026-06-16 04:49:34,261][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.14238004386425018, 'train/grad_norm_pre_clip': 18.315645217895508, 'timing/data': 0.007202779001090676, 'timing/model': 1.4562482390319929, 'learning_rate/qwen_vl_interface': 2.106484881148672e-06, 'learning_rate/action_model': 2.1064848811486718e-05, 'epoch': 1.58}) +[2026-06-16 04:49:40,158][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.03269117698073387, 'train/grad_norm_pre_clip': 6.406215667724609, 'timing/data': 0.006905480986461043, 'timing/model': 1.4581951440195553, 'learning_rate/qwen_vl_interface': 2.1042024147650137e-06, 'learning_rate/action_model': 2.1042024147650137e-05, 'epoch': 1.58}) +[2026-06-16 04:49:46,146][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.06767861545085907, 'train/grad_norm_pre_clip': 7.466673374176025, 'timing/data': 0.007107828976586461, 'timing/model': 1.5085835559875704, 'learning_rate/qwen_vl_interface': 2.1019212414999163e-06, 'learning_rate/action_model': 2.101921241499916e-05, 'epoch': 1.58}) +[2026-06-16 04:49:52,096][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.042590927332639694, 'train/grad_norm_pre_clip': 4.394250392913818, 'timing/data': 0.006772073975298554, 'timing/model': 1.6918713560444303, 'learning_rate/qwen_vl_interface': 2.099641362291082e-06, 'learning_rate/action_model': 2.099641362291082e-05, 'epoch': 1.58}) +[2026-06-16 04:49:57,765][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.20362362265586853, 'train/grad_norm_pre_clip': 13.521830558776855, 'timing/data': 0.006684235006105155, 'timing/model': 1.460246588976588, 'learning_rate/qwen_vl_interface': 2.097362778075685e-06, 'learning_rate/action_model': 2.097362778075685e-05, 'epoch': 1.58}) +[2026-06-16 04:50:03,763][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.1724872887134552, 'train/grad_norm_pre_clip': 15.089715957641602, 'timing/data': 0.006893653015140444, 'timing/model': 1.4586200619814917, 'learning_rate/qwen_vl_interface': 2.0950854897903663e-06, 'learning_rate/action_model': 2.095085489790366e-05, 'epoch': 1.58}) +[2026-06-16 04:50:09,769][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.009090675041079521, 'train/grad_norm_pre_clip': 0.9646338224411011, 'timing/data': 0.006691435002721846, 'timing/model': 1.4644842160050757, 'learning_rate/qwen_vl_interface': 2.092809498371229e-06, 'learning_rate/action_model': 2.0928094983712292e-05, 'epoch': 1.59}) +[2026-06-16 04:50:15,871][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.1910960078239441, 'train/grad_norm_pre_clip': 16.309951782226562, 'timing/data': 0.00792980898404494, 'timing/model': 1.6016339939669706, 'learning_rate/qwen_vl_interface': 2.0905348047538498e-06, 'learning_rate/action_model': 2.0905348047538497e-05, 'epoch': 1.59}) +[2026-06-16 04:50:21,947][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.26806968450546265, 'train/grad_norm_pre_clip': 18.486278533935547, 'timing/data': 0.006800656032282859, 'timing/model': 1.5096305619808845, 'learning_rate/qwen_vl_interface': 2.0882614098732686e-06, 'learning_rate/action_model': 2.0882614098732685e-05, 'epoch': 1.59}) +[2026-06-16 04:50:27,892][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.05850639194250107, 'train/grad_norm_pre_clip': 9.259930610656738, 'timing/data': 0.006421568046789616, 'timing/model': 1.6869185399846174, 'learning_rate/qwen_vl_interface': 2.085989314663993e-06, 'learning_rate/action_model': 2.0859893146639927e-05, 'epoch': 1.59}) +[2026-06-16 04:50:33,561][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.09539549797773361, 'train/grad_norm_pre_clip': 10.259310722351074, 'timing/data': 0.0065455600270070136, 'timing/model': 1.4582048330339603, 'learning_rate/qwen_vl_interface': 2.083718520059995e-06, 'learning_rate/action_model': 2.0837185200599946e-05, 'epoch': 1.59}) +[2026-06-16 04:50:39,570][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.004512235522270203, 'train/grad_norm_pre_clip': 1.1676909923553467, 'timing/data': 0.006741949007846415, 'timing/model': 1.4705952500225976, 'learning_rate/qwen_vl_interface': 2.0814490269947117e-06, 'learning_rate/action_model': 2.0814490269947118e-05, 'epoch': 1.59}) +[2026-06-16 04:50:45,566][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.04365168884396553, 'train/grad_norm_pre_clip': 7.809628009796143, 'timing/data': 0.007229958020616323, 'timing/model': 1.4654485500068404, 'learning_rate/qwen_vl_interface': 2.079180836401047e-06, 'learning_rate/action_model': 2.079180836401047e-05, 'epoch': 1.59}) +[2026-06-16 04:50:51,574][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.023476170375943184, 'train/grad_norm_pre_clip': 4.3522772789001465, 'timing/data': 0.006651294999755919, 'timing/model': 1.4942059899913147, 'learning_rate/qwen_vl_interface': 2.0769139492113687e-06, 'learning_rate/action_model': 2.076913949211369e-05, 'epoch': 1.59}) +[2026-06-16 04:50:57,557][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.172934427857399, 'train/grad_norm_pre_clip': 14.934605598449707, 'timing/data': 0.0064830470364540815, 'timing/model': 1.5166226419969462, 'learning_rate/qwen_vl_interface': 2.0746483663575075e-06, 'learning_rate/action_model': 2.0746483663575076e-05, 'epoch': 1.59}) +[2026-06-16 04:51:03,489][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.16309650242328644, 'train/grad_norm_pre_clip': 15.757457733154297, 'timing/data': 0.0065059110056608915, 'timing/model': 1.6837698830058798, 'learning_rate/qwen_vl_interface': 2.0723840887707595e-06, 'learning_rate/action_model': 2.0723840887707594e-05, 'epoch': 1.59}) +[2026-06-16 04:51:09,307][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.15071527659893036, 'train/grad_norm_pre_clip': 10.196052551269531, 'timing/data': 0.007250847993418574, 'timing/model': 1.6027135149925016, 'learning_rate/qwen_vl_interface': 2.0701211173818854e-06, 'learning_rate/action_model': 2.0701211173818853e-05, 'epoch': 1.59}) +[2026-06-16 04:51:14,963][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.04374904930591583, 'train/grad_norm_pre_clip': 3.353672981262207, 'timing/data': 0.006606840004678816, 'timing/model': 1.4584875919972546, 'learning_rate/qwen_vl_interface': 2.0678594531211036e-06, 'learning_rate/action_model': 2.0678594531211033e-05, 'epoch': 1.59}) +[2026-06-16 04:51:20,863][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.08745002001523972, 'train/grad_norm_pre_clip': 10.608230590820312, 'timing/data': 0.006576655025128275, 'timing/model': 1.458354917995166, 'learning_rate/qwen_vl_interface': 2.0655990969181e-06, 'learning_rate/action_model': 2.0655990969181003e-05, 'epoch': 1.59}) +[2026-06-16 04:51:26,864][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.03643748536705971, 'train/grad_norm_pre_clip': 5.334771633148193, 'timing/data': 0.0067868540063500404, 'timing/model': 1.4741528519662097, 'learning_rate/qwen_vl_interface': 2.0633400497020236e-06, 'learning_rate/action_model': 2.0633400497020232e-05, 'epoch': 1.59}) +[2026-06-16 04:51:32,869][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.019306661561131477, 'train/grad_norm_pre_clip': 1.8342026472091675, 'timing/data': 0.00737941701663658, 'timing/model': 1.537459768995177, 'learning_rate/qwen_vl_interface': 2.061082312401481e-06, 'learning_rate/action_model': 2.061082312401481e-05, 'epoch': 1.59}) +[2026-06-16 04:51:38,791][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.15318934619426727, 'train/grad_norm_pre_clip': 9.599205017089844, 'timing/data': 0.006611545977648348, 'timing/model': 1.6868143280153163, 'learning_rate/qwen_vl_interface': 2.058825885944544e-06, 'learning_rate/action_model': 2.0588258859445442e-05, 'epoch': 1.59}) +[2026-06-16 05:00:45,949][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.2438623309135437, 'train/grad_norm_pre_clip': 15.844361305236816, 'eval/loss': 0.04968015477061272, 'eval/flap_precision': 0.9456740442655935, 'eval/flap_recall': 0.9418837675350702, 'eval/flap_f1': 0.9437751004016065, 'eval/flap_rate_model': 0.13805555555555554, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9437751004016065, 'timing/data': 0.006747555977199227, 'timing/model': 1.6927573079592548, 'learning_rate/qwen_vl_interface': 2.0565707712587447e-06, 'learning_rate/action_model': 2.0565707712587444e-05, 'epoch': 1.59}) +[2026-06-16 05:00:45,950][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3700_state +[2026-06-16 05:00:45,954][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 05:00:51,377][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3700_state/model.safetensors +[2026-06-16 05:00:59,507][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3700_state/optimizer.bin +[2026-06-16 05:00:59,508][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3700_state/sampler.bin +[2026-06-16 05:00:59,508][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3700_state/sampler_1.bin +[2026-06-16 05:00:59,509][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3700_state/random_states_0.pkl +[2026-06-16 05:00:59,615][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3700 +[2026-06-16 05:00:59,616][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 05:00:59,625][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 05:01:05,573][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.04765130579471588, 'train/grad_norm_pre_clip': 10.393070220947266, 'timing/data': 0.006901477987412363, 'timing/model': 1.6949571029981598, 'learning_rate/qwen_vl_interface': 2.0543169692710747e-06, 'learning_rate/action_model': 2.054316969271075e-05, 'epoch': 1.59}) +[2026-06-16 05:01:11,242][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.07070856541395187, 'train/grad_norm_pre_clip': 5.6881422996521, 'timing/data': 0.0066633569658733904, 'timing/model': 1.458929164975416, 'learning_rate/qwen_vl_interface': 2.0520644809079873e-06, 'learning_rate/action_model': 2.0520644809079873e-05, 'epoch': 1.59}) +[2026-06-16 05:01:17,343][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.03299921005964279, 'train/grad_norm_pre_clip': 3.7568795680999756, 'timing/data': 0.011324949038680643, 'timing/model': 1.4587959530181251, 'learning_rate/qwen_vl_interface': 2.049813307095395e-06, 'learning_rate/action_model': 2.0498133070953947e-05, 'epoch': 1.59}) +[2026-06-16 05:01:23,292][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.17070366442203522, 'train/grad_norm_pre_clip': 11.086983680725098, 'timing/data': 0.015082870027981699, 'timing/model': 1.489157045027241, 'learning_rate/qwen_vl_interface': 2.047563448758672e-06, 'learning_rate/action_model': 2.0475634487586717e-05, 'epoch': 1.59}) +[2026-06-16 05:01:29,264][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.0027863802388310432, 'train/grad_norm_pre_clip': 0.5089419484138489, 'timing/data': 0.0070333919720724225, 'timing/model': 1.4895996900158934, 'learning_rate/qwen_vl_interface': 2.045314906822647e-06, 'learning_rate/action_model': 2.0453149068226467e-05, 'epoch': 1.59}) +[2026-06-16 05:01:35,227][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.011558925732970238, 'train/grad_norm_pre_clip': 1.6256201267242432, 'timing/data': 0.0067827829625457525, 'timing/model': 1.4982021459727548, 'learning_rate/qwen_vl_interface': 2.0430676822116124e-06, 'learning_rate/action_model': 2.0430676822116123e-05, 'epoch': 1.59}) +[2026-06-16 05:01:41,257][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.032497093081474304, 'train/grad_norm_pre_clip': 3.883131504058838, 'timing/data': 0.011483831971418113, 'timing/model': 1.7612775769666769, 'learning_rate/qwen_vl_interface': 2.040821775849319e-06, 'learning_rate/action_model': 2.0408217758493188e-05, 'epoch': 1.6}) +[2026-06-16 05:01:46,949][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.011476175859570503, 'train/grad_norm_pre_clip': 1.6400814056396484, 'timing/data': 0.0069172310177236795, 'timing/model': 1.458533898985479, 'learning_rate/qwen_vl_interface': 2.03857718865897e-06, 'learning_rate/action_model': 2.0385771886589702e-05, 'epoch': 1.6}) +[2026-06-16 05:01:52,850][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.03514881804585457, 'train/grad_norm_pre_clip': 3.8831169605255127, 'timing/data': 0.006711838999763131, 'timing/model': 1.457864157971926, 'learning_rate/qwen_vl_interface': 2.0363339215632337e-06, 'learning_rate/action_model': 2.0363339215632334e-05, 'epoch': 1.6}) +[2026-06-16 05:01:58,751][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.024137230589985847, 'train/grad_norm_pre_clip': 5.377825736999512, 'timing/data': 0.006717236014083028, 'timing/model': 1.4585486500291154, 'learning_rate/qwen_vl_interface': 2.034091975484231e-06, 'learning_rate/action_model': 2.0340919754842307e-05, 'epoch': 1.6}) +[2026-06-16 05:02:04,653][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.06038618087768555, 'train/grad_norm_pre_clip': 8.57533073425293, 'timing/data': 0.006796083005610853, 'timing/model': 1.4322480850387365, 'learning_rate/qwen_vl_interface': 2.0318513513435415e-06, 'learning_rate/action_model': 2.0318513513435415e-05, 'epoch': 1.6}) +[2026-06-16 05:02:10,557][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.0559425912797451, 'train/grad_norm_pre_clip': 7.468115329742432, 'timing/data': 0.006619232008233666, 'timing/model': 1.491316027007997, 'learning_rate/qwen_vl_interface': 2.0296120500622018e-06, 'learning_rate/action_model': 2.0296120500622017e-05, 'epoch': 1.6}) +[2026-06-16 05:02:16,463][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.15808898210525513, 'train/grad_norm_pre_clip': 11.694795608520508, 'timing/data': 0.006714121962431818, 'timing/model': 1.663179385010153, 'learning_rate/qwen_vl_interface': 2.0273740725607034e-06, 'learning_rate/action_model': 2.0273740725607032e-05, 'epoch': 1.6}) +[2026-06-16 05:02:22,352][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.10512463748455048, 'train/grad_norm_pre_clip': 17.24555206298828, 'timing/data': 0.006647043977864087, 'timing/model': 1.459035516018048, 'learning_rate/qwen_vl_interface': 2.0251374197589942e-06, 'learning_rate/action_model': 2.0251374197589943e-05, 'epoch': 1.6}) +[2026-06-16 05:02:28,436][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.05356074497103691, 'train/grad_norm_pre_clip': 10.319072723388672, 'timing/data': 0.006997800024691969, 'timing/model': 1.5429359470144846, 'learning_rate/qwen_vl_interface': 2.0229020925764785e-06, 'learning_rate/action_model': 2.0229020925764785e-05, 'epoch': 1.6}) +[2026-06-16 05:02:34,452][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.1238916739821434, 'train/grad_norm_pre_clip': 13.990303039550781, 'timing/data': 0.006966574001125991, 'timing/model': 1.4568221869994886, 'learning_rate/qwen_vl_interface': 2.0206680919320145e-06, 'learning_rate/action_model': 2.020668091932014e-05, 'epoch': 1.6}) +[2026-06-16 05:02:40,434][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.008151886984705925, 'train/grad_norm_pre_clip': 1.1633996963500977, 'timing/data': 0.006792786996811628, 'timing/model': 1.5044969010050409, 'learning_rate/qwen_vl_interface': 2.0184354187439164e-06, 'learning_rate/action_model': 2.0184354187439162e-05, 'epoch': 1.6}) +[2026-06-16 05:02:46,354][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.06132460758090019, 'train/grad_norm_pre_clip': 4.138250350952148, 'timing/data': 0.006754560978151858, 'timing/model': 1.4810105580254458, 'learning_rate/qwen_vl_interface': 2.016204073929949e-06, 'learning_rate/action_model': 2.016204073929949e-05, 'epoch': 1.6}) +[2026-06-16 05:02:52,287][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.09832635521888733, 'train/grad_norm_pre_clip': 6.505383014678955, 'timing/data': 0.0067342419642955065, 'timing/model': 1.6907248820061795, 'learning_rate/qwen_vl_interface': 2.0139740584073363e-06, 'learning_rate/action_model': 2.0139740584073364e-05, 'epoch': 1.6}) +[2026-06-16 05:02:57,952][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.29929473996162415, 'train/grad_norm_pre_clip': 11.809961318969727, 'timing/data': 0.006584661023225635, 'timing/model': 1.4564847020083107, 'learning_rate/qwen_vl_interface': 2.011745373092753e-06, 'learning_rate/action_model': 2.011745373092753e-05, 'epoch': 1.6}) +[2026-06-16 05:03:03,860][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.048045117408037186, 'train/grad_norm_pre_clip': 8.957249641418457, 'timing/data': 0.007099918962921947, 'timing/model': 1.463060285022948, 'learning_rate/qwen_vl_interface': 2.0095180189023267e-06, 'learning_rate/action_model': 2.0095180189023265e-05, 'epoch': 1.6}) +[2026-06-16 05:03:09,754][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.02098514884710312, 'train/grad_norm_pre_clip': 2.4366936683654785, 'timing/data': 0.007774251978844404, 'timing/model': 1.4562199430074543, 'learning_rate/qwen_vl_interface': 2.0072919967516393e-06, 'learning_rate/action_model': 2.0072919967516392e-05, 'epoch': 1.6}) +[2026-06-16 05:03:15,660][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.14178431034088135, 'train/grad_norm_pre_clip': 14.209754943847656, 'timing/data': 0.0066049209563061595, 'timing/model': 1.4617850079666823, 'learning_rate/qwen_vl_interface': 2.0050673075557245e-06, 'learning_rate/action_model': 2.0050673075557243e-05, 'epoch': 1.6}) +[2026-06-16 05:03:21,556][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.06933232396841049, 'train/grad_norm_pre_clip': 5.2846903800964355, 'timing/data': 0.006869931006804109, 'timing/model': 1.4719390980317257, 'learning_rate/qwen_vl_interface': 2.0028439522290677e-06, 'learning_rate/action_model': 2.0028439522290676e-05, 'epoch': 1.6}) +[2026-06-16 05:03:27,488][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.0029853342566639185, 'train/grad_norm_pre_clip': 0.5966445207595825, 'timing/data': 0.006675305019598454, 'timing/model': 1.690199535980355, 'learning_rate/qwen_vl_interface': 2.000621931685606e-06, 'learning_rate/action_model': 2.000621931685606e-05, 'epoch': 1.6}) +[2026-06-16 05:03:33,157][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.08580807596445084, 'train/grad_norm_pre_clip': 15.367294311523438, 'timing/data': 0.0070654190494678915, 'timing/model': 1.4600412679719739, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 1.6}) +[2026-06-16 05:03:39,057][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.008349849842488766, 'train/grad_norm_pre_clip': 2.644639730453491, 'timing/data': 0.007067561964504421, 'timing/model': 1.4607612140243873, 'learning_rate/qwen_vl_interface': 1.9961818986012767e-06, 'learning_rate/action_model': 1.9961818986012763e-05, 'epoch': 1.6}) +[2026-06-16 05:03:45,022][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.06959997117519379, 'train/grad_norm_pre_clip': 9.36202335357666, 'timing/data': 0.006777435017284006, 'timing/model': 1.522681292030029, 'learning_rate/qwen_vl_interface': 1.9939638878855372e-06, 'learning_rate/action_model': 1.993963887885537e-05, 'epoch': 1.6}) +[2026-06-16 05:03:50,958][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.04040377959609032, 'train/grad_norm_pre_clip': 5.206623077392578, 'timing/data': 0.006572072976268828, 'timing/model': 1.4338920699665323, 'learning_rate/qwen_vl_interface': 1.991747215603252e-06, 'learning_rate/action_model': 1.991747215603252e-05, 'epoch': 1.6}) +[2026-06-16 05:03:56,857][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.05585668236017227, 'train/grad_norm_pre_clip': 3.307886838912964, 'timing/data': 0.006512985972221941, 'timing/model': 1.460895499039907, 'learning_rate/qwen_vl_interface': 1.9895318826656116e-06, 'learning_rate/action_model': 1.9895318826656114e-05, 'epoch': 1.6}) +[2026-06-16 05:04:02,791][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.014676027931272984, 'train/grad_norm_pre_clip': 2.7206952571868896, 'timing/data': 0.006794379965867847, 'timing/model': 1.691464987990912, 'learning_rate/qwen_vl_interface': 1.9873178899832553e-06, 'learning_rate/action_model': 1.9873178899832554e-05, 'epoch': 1.61}) +[2026-06-16 05:04:08,454][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.005229236092418432, 'train/grad_norm_pre_clip': 1.033473014831543, 'timing/data': 0.006728342035785317, 'timing/model': 1.4552018900285475, 'learning_rate/qwen_vl_interface': 1.985105238466272e-06, 'learning_rate/action_model': 1.9851052384662722e-05, 'epoch': 1.61}) +[2026-06-16 05:04:14,356][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.11523672938346863, 'train/grad_norm_pre_clip': 20.980581283569336, 'timing/data': 0.006942367006558925, 'timing/model': 1.4590486980159767, 'learning_rate/qwen_vl_interface': 1.9828939290242e-06, 'learning_rate/action_model': 1.9828939290241998e-05, 'epoch': 1.61}) +[2026-06-16 05:04:20,256][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.009669242426753044, 'train/grad_norm_pre_clip': 5.0786967277526855, 'timing/data': 0.006747551029548049, 'timing/model': 1.4556668869918212, 'learning_rate/qwen_vl_interface': 1.980683962566024e-06, 'learning_rate/action_model': 1.980683962566024e-05, 'epoch': 1.61}) +[2026-06-16 05:04:26,256][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.08788947761058807, 'train/grad_norm_pre_clip': 13.230606079101562, 'timing/data': 0.007268938992638141, 'timing/model': 1.458566019020509, 'learning_rate/qwen_vl_interface': 1.9784753400001784e-06, 'learning_rate/action_model': 1.978475340000178e-05, 'epoch': 1.61}) +[2026-06-16 05:04:32,233][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.12336591631174088, 'train/grad_norm_pre_clip': 17.065135955810547, 'timing/data': 0.007029375003185123, 'timing/model': 1.5032662329613231, 'learning_rate/qwen_vl_interface': 1.9762680622345443e-06, 'learning_rate/action_model': 1.9762680622345442e-05, 'epoch': 1.61}) +[2026-06-16 05:04:38,184][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.0695272907614708, 'train/grad_norm_pre_clip': 8.952354431152344, 'timing/data': 0.006822520983405411, 'timing/model': 1.686329853022471, 'learning_rate/qwen_vl_interface': 1.9740621301764513e-06, 'learning_rate/action_model': 1.9740621301764513e-05, 'epoch': 1.61}) +[2026-06-16 05:04:43,852][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.3286634683609009, 'train/grad_norm_pre_clip': 22.210405349731445, 'timing/data': 0.006825245043728501, 'timing/model': 1.4576205959892832, 'learning_rate/qwen_vl_interface': 1.971857544732672e-06, 'learning_rate/action_model': 1.971857544732672e-05, 'epoch': 1.61}) +[2026-06-16 05:04:49,754][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.08209938555955887, 'train/grad_norm_pre_clip': 16.084251403808594, 'timing/data': 0.006576738960575312, 'timing/model': 1.4610916909878142, 'learning_rate/qwen_vl_interface': 1.96965430680943e-06, 'learning_rate/action_model': 1.96965430680943e-05, 'epoch': 1.61}) +[2026-06-16 05:04:55,653][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.009630434215068817, 'train/grad_norm_pre_clip': 2.325974702835083, 'timing/data': 0.006699600024148822, 'timing/model': 1.4607784560066648, 'learning_rate/qwen_vl_interface': 1.967452417312393e-06, 'learning_rate/action_model': 1.9674524173123927e-05, 'epoch': 1.61}) +[2026-06-16 05:05:01,552][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.007849257439374924, 'train/grad_norm_pre_clip': 1.1956241130828857, 'timing/data': 0.007218543963972479, 'timing/model': 1.4583124130149372, 'learning_rate/qwen_vl_interface': 1.9652518771466746e-06, 'learning_rate/action_model': 1.9652518771466747e-05, 'epoch': 1.61}) +[2026-06-16 05:05:07,749][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.14405028522014618, 'train/grad_norm_pre_clip': 13.839892387390137, 'timing/data': 0.006089832982979715, 'timing/model': 1.715509324974846, 'learning_rate/qwen_vl_interface': 1.963052687216834e-06, 'learning_rate/action_model': 1.963052687216834e-05, 'epoch': 1.61}) +[2026-06-16 05:05:13,693][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.03626077622175217, 'train/grad_norm_pre_clip': 10.270751953125, 'timing/data': 0.006892784032970667, 'timing/model': 1.7104717760230415, 'learning_rate/qwen_vl_interface': 1.960854848426874e-06, 'learning_rate/action_model': 1.960854848426874e-05, 'epoch': 1.61}) +[2026-06-16 05:05:19,344][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.027667537331581116, 'train/grad_norm_pre_clip': 8.013096809387207, 'timing/data': 0.006727840984240174, 'timing/model': 1.4540862520225346, 'learning_rate/qwen_vl_interface': 1.9586583616802446e-06, 'learning_rate/action_model': 1.9586583616802446e-05, 'epoch': 1.61}) +[2026-06-16 05:05:25,255][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.01066916435956955, 'train/grad_norm_pre_clip': 2.9593021869659424, 'timing/data': 0.006889970041811466, 'timing/model': 1.459856059984304, 'learning_rate/qwen_vl_interface': 1.956463227879838e-06, 'learning_rate/action_model': 1.9564632278798377e-05, 'epoch': 1.61}) +[2026-06-16 05:05:31,162][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.23595327138900757, 'train/grad_norm_pre_clip': 18.3437442779541, 'timing/data': 0.006616166967432946, 'timing/model': 1.467044175020419, 'learning_rate/qwen_vl_interface': 1.95426944792799e-06, 'learning_rate/action_model': 1.95426944792799e-05, 'epoch': 1.61}) +[2026-06-16 05:05:37,056][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.03824210166931152, 'train/grad_norm_pre_clip': 3.932447910308838, 'timing/data': 0.006635214958805591, 'timing/model': 1.460152137966361, 'learning_rate/qwen_vl_interface': 1.9520770227264817e-06, 'learning_rate/action_model': 1.9520770227264817e-05, 'epoch': 1.61}) +[2026-06-16 05:05:42,960][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.24737495183944702, 'train/grad_norm_pre_clip': 14.620881080627441, 'timing/data': 0.006667862995527685, 'timing/model': 1.4337904580170289, 'learning_rate/qwen_vl_interface': 1.9498859531765373e-06, 'learning_rate/action_model': 1.9498859531765372e-05, 'epoch': 1.61}) +[2026-06-16 05:05:48,885][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.016708634793758392, 'train/grad_norm_pre_clip': 2.7361910343170166, 'timing/data': 0.006734321010299027, 'timing/model': 1.686769257998094, 'learning_rate/qwen_vl_interface': 1.9476962401788204e-06, 'learning_rate/action_model': 1.9476962401788204e-05, 'epoch': 1.61}) +[2026-06-16 05:05:54,552][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.08712352067232132, 'train/grad_norm_pre_clip': 11.071237564086914, 'timing/data': 0.007058387971483171, 'timing/model': 1.458635515999049, 'learning_rate/qwen_vl_interface': 1.9455078846334397e-06, 'learning_rate/action_model': 1.9455078846334395e-05, 'epoch': 1.61}) +[2026-06-16 05:06:00,465][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.08768819272518158, 'train/grad_norm_pre_clip': 10.066530227661133, 'timing/data': 0.006955577002372593, 'timing/model': 1.4710696630063467, 'learning_rate/qwen_vl_interface': 1.9433208874399465e-06, 'learning_rate/action_model': 1.9433208874399464e-05, 'epoch': 1.61}) +[2026-06-16 05:06:06,456][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.26257070899009705, 'train/grad_norm_pre_clip': 18.396076202392578, 'timing/data': 0.007174078025855124, 'timing/model': 1.4628856150084175, 'learning_rate/qwen_vl_interface': 1.9411352494973314e-06, 'learning_rate/action_model': 1.9411352494973315e-05, 'epoch': 1.61}) +[2026-06-16 05:06:12,355][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.0475144237279892, 'train/grad_norm_pre_clip': 4.38851261138916, 'timing/data': 0.007098786998540163, 'timing/model': 1.4557721930323169, 'learning_rate/qwen_vl_interface': 1.938950971704031e-06, 'learning_rate/action_model': 1.9389509717040308e-05, 'epoch': 1.61}) +[2026-06-16 05:06:18,261][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.09244381636381149, 'train/grad_norm_pre_clip': 14.430272102355957, 'timing/data': 0.006495730020105839, 'timing/model': 1.4346723459893838, 'learning_rate/qwen_vl_interface': 1.936768054957915e-06, 'learning_rate/action_model': 1.9367680549579147e-05, 'epoch': 1.62}) +[2026-06-16 05:06:24,181][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.13094714283943176, 'train/grad_norm_pre_clip': 16.337932586669922, 'timing/data': 0.006546714983414859, 'timing/model': 1.6830890289857052, 'learning_rate/qwen_vl_interface': 1.9345865001562994e-06, 'learning_rate/action_model': 1.9345865001562994e-05, 'epoch': 1.62}) +[2026-06-16 05:06:30,000][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.08938967436552048, 'train/grad_norm_pre_clip': 14.239358901977539, 'timing/data': 0.006730124994646758, 'timing/model': 1.6084443110157736, 'learning_rate/qwen_vl_interface': 1.9324063081959394e-06, 'learning_rate/action_model': 1.9324063081959395e-05, 'epoch': 1.62}) +[2026-06-16 05:06:35,657][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.005567690823227167, 'train/grad_norm_pre_clip': 1.3649051189422607, 'timing/data': 0.006578291999176145, 'timing/model': 1.4617012160015292, 'learning_rate/qwen_vl_interface': 1.9302274799730304e-06, 'learning_rate/action_model': 1.9302274799730306e-05, 'epoch': 1.62}) +[2026-06-16 05:06:41,656][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.09106287360191345, 'train/grad_norm_pre_clip': 22.772300720214844, 'timing/data': 0.006657807971350849, 'timing/model': 1.4573210059897974, 'learning_rate/qwen_vl_interface': 1.9280500163832055e-06, 'learning_rate/action_model': 1.9280500163832055e-05, 'epoch': 1.62}) +[2026-06-16 05:06:47,652][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.008049292489886284, 'train/grad_norm_pre_clip': 1.7948068380355835, 'timing/data': 0.006583057984244078, 'timing/model': 1.4569370480021462, 'learning_rate/qwen_vl_interface': 1.925873918321537e-06, 'learning_rate/action_model': 1.925873918321537e-05, 'epoch': 1.62}) +[2026-06-16 05:06:53,554][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.05245574563741684, 'train/grad_norm_pre_clip': 5.9911980628967285, 'timing/data': 0.007159636006690562, 'timing/model': 1.4302698689862154, 'learning_rate/qwen_vl_interface': 1.92369918668254e-06, 'learning_rate/action_model': 1.92369918668254e-05, 'epoch': 1.62}) +[2026-06-16 05:06:59,483][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.049912579357624054, 'train/grad_norm_pre_clip': 7.562801361083984, 'timing/data': 0.006655394972767681, 'timing/model': 1.6892410020227544, 'learning_rate/qwen_vl_interface': 1.9215258223601605e-06, 'learning_rate/action_model': 1.9215258223601605e-05, 'epoch': 1.62}) +[2026-06-16 05:07:05,380][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.022944610565900803, 'train/grad_norm_pre_clip': 3.2229080200195312, 'timing/data': 0.006832195038441569, 'timing/model': 1.6882471760036424, 'learning_rate/qwen_vl_interface': 1.9193538262477874e-06, 'learning_rate/action_model': 1.9193538262477874e-05, 'epoch': 1.62}) +[2026-06-16 05:07:11,052][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.040292952209711075, 'train/grad_norm_pre_clip': 6.685702323913574, 'timing/data': 0.006918552040588111, 'timing/model': 1.4589165600482374, 'learning_rate/qwen_vl_interface': 1.917183199238247e-06, 'learning_rate/action_model': 1.917183199238247e-05, 'epoch': 1.62}) +[2026-06-16 05:07:16,957][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.05918225273489952, 'train/grad_norm_pre_clip': 13.814913749694824, 'timing/data': 0.0066473030019551516, 'timing/model': 1.4604954299866222, 'learning_rate/qwen_vl_interface': 1.9150139422238016e-06, 'learning_rate/action_model': 1.9150139422238018e-05, 'epoch': 1.62}) +[2026-06-16 05:07:22,954][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.049283888190984726, 'train/grad_norm_pre_clip': 9.270645141601562, 'timing/data': 0.006577550026122481, 'timing/model': 1.4582660970045254, 'learning_rate/qwen_vl_interface': 1.9128460560961513e-06, 'learning_rate/action_model': 1.9128460560961514e-05, 'epoch': 1.62}) +[2026-06-16 05:07:28,852][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.0014423185493797064, 'train/grad_norm_pre_clip': 0.3257597088813782, 'timing/data': 0.006655164994299412, 'timing/model': 1.4302286509773694, 'learning_rate/qwen_vl_interface': 1.910679541746432e-06, 'learning_rate/action_model': 1.9106795417464323e-05, 'epoch': 1.62}) +[2026-06-16 05:07:34,775][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.1602335423231125, 'train/grad_norm_pre_clip': 5.879319667816162, 'timing/data': 0.006917320017237216, 'timing/model': 1.6808494110009633, 'learning_rate/qwen_vl_interface': 1.9085144000652166e-06, 'learning_rate/action_model': 1.9085144000652166e-05, 'epoch': 1.62}) +[2026-06-16 05:07:40,629][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.03862900659441948, 'train/grad_norm_pre_clip': 4.799127578735352, 'timing/data': 0.006467807979788631, 'timing/model': 1.6339492570259608, 'learning_rate/qwen_vl_interface': 1.9063506319425116e-06, 'learning_rate/action_model': 1.9063506319425114e-05, 'epoch': 1.62}) +[2026-06-16 05:07:46,353][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.04331105202436447, 'train/grad_norm_pre_clip': 4.969873428344727, 'timing/data': 0.006681243015918881, 'timing/model': 1.45912254601717, 'learning_rate/qwen_vl_interface': 1.9041882382677613e-06, 'learning_rate/action_model': 1.9041882382677612e-05, 'epoch': 1.62}) +[2026-06-16 05:07:52,255][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.14873363077640533, 'train/grad_norm_pre_clip': 11.677321434020996, 'timing/data': 0.006546453980263323, 'timing/model': 1.4597500069648959, 'learning_rate/qwen_vl_interface': 1.9020272199298452e-06, 'learning_rate/action_model': 1.902027219929845e-05, 'epoch': 1.62}) +[2026-06-16 05:07:58,153][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.032151367515325546, 'train/grad_norm_pre_clip': 9.328507423400879, 'timing/data': 0.006605801987461746, 'timing/model': 1.457310784026049, 'learning_rate/qwen_vl_interface': 1.8998675778170736e-06, 'learning_rate/action_model': 1.8998675778170734e-05, 'epoch': 1.62}) +[2026-06-16 05:08:04,051][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.06599970906972885, 'train/grad_norm_pre_clip': 11.489509582519531, 'timing/data': 0.006434909999370575, 'timing/model': 1.4334811020526104, 'learning_rate/qwen_vl_interface': 1.8977093128171955e-06, 'learning_rate/action_model': 1.8977093128171954e-05, 'epoch': 1.62}) +[2026-06-16 05:08:09,949][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.014537259005010128, 'train/grad_norm_pre_clip': 2.4735629558563232, 'timing/data': 0.007309627020731568, 'timing/model': 1.4277863589813933, 'learning_rate/qwen_vl_interface': 1.8955524258173919e-06, 'learning_rate/action_model': 1.895552425817392e-05, 'epoch': 1.62}) +[2026-06-16 05:08:15,840][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.03799443319439888, 'train/grad_norm_pre_clip': 5.753812789916992, 'timing/data': 0.006372629024554044, 'timing/model': 1.6495678850333206, 'learning_rate/qwen_vl_interface': 1.8933969177042769e-06, 'learning_rate/action_model': 1.8933969177042768e-05, 'epoch': 1.62}) +[2026-06-16 05:08:21,683][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.03720283508300781, 'train/grad_norm_pre_clip': 5.319523334503174, 'timing/data': 0.00669010600540787, 'timing/model': 1.4585710790124722, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 1.62}) +[2026-06-16 05:08:27,657][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.2578662037849426, 'train/grad_norm_pre_clip': 19.770450592041016, 'timing/data': 0.006828069046605378, 'timing/model': 1.4602341029676609, 'learning_rate/qwen_vl_interface': 1.8890900416817398e-06, 'learning_rate/action_model': 1.8890900416817397e-05, 'epoch': 1.62}) +[2026-06-16 05:08:33,657][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.013108032755553722, 'train/grad_norm_pre_clip': 2.856458902359009, 'timing/data': 0.006704586965497583, 'timing/model': 1.4617509020026773, 'learning_rate/qwen_vl_interface': 1.8869386755427117e-06, 'learning_rate/action_model': 1.8869386755427114e-05, 'epoch': 1.63}) +[2026-06-16 05:08:39,755][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.18937167525291443, 'train/grad_norm_pre_clip': 11.643375396728516, 'timing/data': 0.00706124200951308, 'timing/model': 1.5055239200009964, 'learning_rate/qwen_vl_interface': 1.8847886918311597e-06, 'learning_rate/action_model': 1.8847886918311598e-05, 'epoch': 1.63}) +[2026-06-16 05:08:45,685][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.25008347630500793, 'train/grad_norm_pre_clip': 35.82965850830078, 'timing/data': 0.010458171018399298, 'timing/model': 1.6838937270222232, 'learning_rate/qwen_vl_interface': 1.8826400914308616e-06, 'learning_rate/action_model': 1.8826400914308613e-05, 'epoch': 1.63}) +[2026-06-16 05:08:51,506][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.0010761566227301955, 'train/grad_norm_pre_clip': 0.27911436557769775, 'timing/data': 0.013702424999792129, 'timing/model': 1.5614810439874418, 'learning_rate/qwen_vl_interface': 1.880492875225025e-06, 'learning_rate/action_model': 1.8804928752250252e-05, 'epoch': 1.63}) +[2026-06-16 05:08:57,455][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.05232209712266922, 'train/grad_norm_pre_clip': 6.823968410491943, 'timing/data': 0.007003906997852027, 'timing/model': 1.4611316919908859, 'learning_rate/qwen_vl_interface': 1.8783470440962922e-06, 'learning_rate/action_model': 1.878347044096292e-05, 'epoch': 1.63}) +[2026-06-16 05:09:03,256][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.10370957106351852, 'train/grad_norm_pre_clip': 12.730159759521484, 'timing/data': 0.007308386033400893, 'timing/model': 1.4566931779845618, 'learning_rate/qwen_vl_interface': 1.8762025989267293e-06, 'learning_rate/action_model': 1.876202598926729e-05, 'epoch': 1.63}) +[2026-06-16 05:09:09,155][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.19237405061721802, 'train/grad_norm_pre_clip': 11.17971134185791, 'timing/data': 0.006535146967507899, 'timing/model': 1.4589618039899506, 'learning_rate/qwen_vl_interface': 1.874059540597839e-06, 'learning_rate/action_model': 1.874059540597839e-05, 'epoch': 1.63}) +[2026-06-16 05:09:15,056][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.3953764736652374, 'train/grad_norm_pre_clip': 20.227041244506836, 'timing/data': 0.006812105013523251, 'timing/model': 1.4771197639638558, 'learning_rate/qwen_vl_interface': 1.8719178699905508e-06, 'learning_rate/action_model': 1.8719178699905508e-05, 'epoch': 1.63}) +[2026-06-16 05:09:20,993][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.3553926646709442, 'train/grad_norm_pre_clip': 9.043478012084961, 'timing/data': 0.006815129017923027, 'timing/model': 1.694329667952843, 'learning_rate/qwen_vl_interface': 1.8697775879852242e-06, 'learning_rate/action_model': 1.8697775879852242e-05, 'epoch': 1.63}) +[2026-06-16 05:09:26,656][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.04227036237716675, 'train/grad_norm_pre_clip': 6.726134300231934, 'timing/data': 0.006755751965101808, 'timing/model': 1.4581928959814832, 'learning_rate/qwen_vl_interface': 1.86763869546165e-06, 'learning_rate/action_model': 1.86763869546165e-05, 'epoch': 1.63}) +[2026-06-16 05:09:32,656][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.013579640537500381, 'train/grad_norm_pre_clip': 3.6693966388702393, 'timing/data': 0.006658038997557014, 'timing/model': 1.4590861279866658, 'learning_rate/qwen_vl_interface': 1.865501193299045e-06, 'learning_rate/action_model': 1.8655011932990447e-05, 'epoch': 1.63}) +[2026-06-16 05:09:38,558][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.011959978379309177, 'train/grad_norm_pre_clip': 2.3932366371154785, 'timing/data': 0.006631600030232221, 'timing/model': 1.455984300991986, 'learning_rate/qwen_vl_interface': 1.8633650823760555e-06, 'learning_rate/action_model': 1.8633650823760555e-05, 'epoch': 1.63}) +[2026-06-16 05:09:44,458][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.0040536606684327126, 'train/grad_norm_pre_clip': 1.574709177017212, 'timing/data': 0.0066932199988514185, 'timing/model': 1.4593812780221924, 'learning_rate/qwen_vl_interface': 1.8612303635707562e-06, 'learning_rate/action_model': 1.861230363570756e-05, 'epoch': 1.63}) +[2026-06-16 05:09:50,369][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.00449402118101716, 'train/grad_norm_pre_clip': 1.3726762533187866, 'timing/data': 0.00699273101054132, 'timing/model': 1.4458954999572597, 'learning_rate/qwen_vl_interface': 1.8590970377606494e-06, 'learning_rate/action_model': 1.859097037760649e-05, 'epoch': 1.63}) +[2026-06-16 05:09:56,286][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.027346232905983925, 'train/grad_norm_pre_clip': 6.756114959716797, 'timing/data': 0.00650109798880294, 'timing/model': 1.6930326739675365, 'learning_rate/qwen_vl_interface': 1.8569651058226667e-06, 'learning_rate/action_model': 1.8569651058226666e-05, 'epoch': 1.63}) +[2026-06-16 05:10:01,952][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.07541916519403458, 'train/grad_norm_pre_clip': 13.939111709594727, 'timing/data': 0.006356433965265751, 'timing/model': 1.4613442900008522, 'learning_rate/qwen_vl_interface': 1.854834568633161e-06, 'learning_rate/action_model': 1.854834568633161e-05, 'epoch': 1.63}) +[2026-06-16 05:10:07,851][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.08919767290353775, 'train/grad_norm_pre_clip': 15.063880920410156, 'timing/data': 0.006446586980018765, 'timing/model': 1.4587499600020237, 'learning_rate/qwen_vl_interface': 1.8527054270679174e-06, 'learning_rate/action_model': 1.8527054270679172e-05, 'epoch': 1.63}) +[2026-06-16 05:10:13,761][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.12631088495254517, 'train/grad_norm_pre_clip': 15.530535697937012, 'timing/data': 0.007117753964848816, 'timing/model': 1.4672227470437065, 'learning_rate/qwen_vl_interface': 1.850577682002146e-06, 'learning_rate/action_model': 1.850577682002146e-05, 'epoch': 1.63}) +[2026-06-16 05:10:19,654][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.08837141841650009, 'train/grad_norm_pre_clip': 14.71448040008545, 'timing/data': 0.006803322001360357, 'timing/model': 1.4329763969872147, 'learning_rate/qwen_vl_interface': 1.848451334310482e-06, 'learning_rate/action_model': 1.848451334310482e-05, 'epoch': 1.63}) +[2026-06-16 05:10:25,560][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.08568963408470154, 'train/grad_norm_pre_clip': 10.60250473022461, 'timing/data': 0.006642736028879881, 'timing/model': 1.4658298729918897, 'learning_rate/qwen_vl_interface': 1.8463263848669865e-06, 'learning_rate/action_model': 1.8463263848669864e-05, 'epoch': 1.63}) +[2026-06-16 05:10:31,496][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.05630872771143913, 'train/grad_norm_pre_clip': 5.9786763191223145, 'timing/data': 0.006744275975506753, 'timing/model': 1.704104834003374, 'learning_rate/qwen_vl_interface': 1.8442028345451468e-06, 'learning_rate/action_model': 1.8442028345451468e-05, 'epoch': 1.63}) +[2026-06-16 05:10:37,153][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.07104163616895676, 'train/grad_norm_pre_clip': 10.415914535522461, 'timing/data': 0.006467818049713969, 'timing/model': 1.4589209819678217, 'learning_rate/qwen_vl_interface': 1.8420806842178746e-06, 'learning_rate/action_model': 1.8420806842178745e-05, 'epoch': 1.63}) +[2026-06-16 05:10:43,053][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.22510813176631927, 'train/grad_norm_pre_clip': 27.76748275756836, 'timing/data': 0.006587233976460993, 'timing/model': 1.460274011013098, 'learning_rate/qwen_vl_interface': 1.8399599347575034e-06, 'learning_rate/action_model': 1.8399599347575033e-05, 'epoch': 1.63}) +[2026-06-16 05:20:01,652][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.17216141521930695, 'train/grad_norm_pre_clip': 34.24286651611328, 'eval/loss': 0.058521583676338196, 'eval/flap_precision': 0.9259259259259259, 'eval/flap_recall': 0.9519038076152304, 'eval/flap_f1': 0.9387351778656126, 'eval/flap_rate_model': 0.1425, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9387351778656126, 'timing/data': 0.0075427599949762225, 'timing/model': 1.458150945021771, 'learning_rate/qwen_vl_interface': 1.8378405870357943e-06, 'learning_rate/action_model': 1.837840587035794e-05, 'epoch': 1.64}) +[2026-06-16 05:20:01,652][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3800_state +[2026-06-16 05:20:01,657][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 05:20:07,086][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3800_state/model.safetensors +[2026-06-16 05:20:15,130][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3800_state/optimizer.bin +[2026-06-16 05:20:15,130][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3800_state/sampler.bin +[2026-06-16 05:20:15,130][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3800_state/sampler_1.bin +[2026-06-16 05:20:15,131][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3800_state/random_states_0.pkl +[2026-06-16 05:20:15,236][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3800 +[2026-06-16 05:20:15,236][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 05:20:15,261][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 05:20:21,233][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.003406266449019313, 'train/grad_norm_pre_clip': 1.1267350912094116, 'timing/data': 0.006843659037258476, 'timing/model': 1.4590226439759135, 'learning_rate/qwen_vl_interface': 1.8357226419239316e-06, 'learning_rate/action_model': 1.8357226419239315e-05, 'epoch': 1.64}) +[2026-06-16 05:20:27,138][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.05362006276845932, 'train/grad_norm_pre_clip': 7.732306003570557, 'timing/data': 0.007076000969391316, 'timing/model': 1.4584133010357618, 'learning_rate/qwen_vl_interface': 1.8336061002925224e-06, 'learning_rate/action_model': 1.8336061002925224e-05, 'epoch': 1.64}) +[2026-06-16 05:20:33,054][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.045599088072776794, 'train/grad_norm_pre_clip': 4.942208290100098, 'timing/data': 0.007141536974813789, 'timing/model': 1.4564845170243643, 'learning_rate/qwen_vl_interface': 1.8314909630115985e-06, 'learning_rate/action_model': 1.8314909630115983e-05, 'epoch': 1.64}) +[2026-06-16 05:20:38,950][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.11825088411569595, 'train/grad_norm_pre_clip': 18.723129272460938, 'timing/data': 0.007797021011356264, 'timing/model': 1.4324218889814802, 'learning_rate/qwen_vl_interface': 1.8293772309506092e-06, 'learning_rate/action_model': 1.829377230950609e-05, 'epoch': 1.64}) +[2026-06-16 05:20:44,975][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.14925481379032135, 'train/grad_norm_pre_clip': 15.41783332824707, 'timing/data': 0.009003286017104983, 'timing/model': 1.7819388899952173, 'learning_rate/qwen_vl_interface': 1.8272649049784326e-06, 'learning_rate/action_model': 1.8272649049784323e-05, 'epoch': 1.64}) +[2026-06-16 05:20:50,744][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.0031849557999521494, 'train/grad_norm_pre_clip': 0.667091429233551, 'timing/data': 0.007132383994758129, 'timing/model': 1.4577392539940774, 'learning_rate/qwen_vl_interface': 1.8251539859633653e-06, 'learning_rate/action_model': 1.825153985963365e-05, 'epoch': 1.64}) +[2026-06-16 05:20:56,848][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.05951108783483505, 'train/grad_norm_pre_clip': 7.990994930267334, 'timing/data': 0.00737493101041764, 'timing/model': 1.4579798320191912, 'learning_rate/qwen_vl_interface': 1.8230444747731265e-06, 'learning_rate/action_model': 1.8230444747731264e-05, 'epoch': 1.64}) +[2026-06-16 05:21:02,847][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.09158454090356827, 'train/grad_norm_pre_clip': 11.930492401123047, 'timing/data': 0.007404674019198865, 'timing/model': 1.456312071008142, 'learning_rate/qwen_vl_interface': 1.820936372274857e-06, 'learning_rate/action_model': 1.820936372274857e-05, 'epoch': 1.64}) +[2026-06-16 05:21:08,759][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.03411025553941727, 'train/grad_norm_pre_clip': 6.589091777801514, 'timing/data': 0.006921883032191545, 'timing/model': 1.4710280460421927, 'learning_rate/qwen_vl_interface': 1.8188296793351178e-06, 'learning_rate/action_model': 1.818829679335118e-05, 'epoch': 1.64}) +[2026-06-16 05:21:14,735][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.012595994397997856, 'train/grad_norm_pre_clip': 5.09816312789917, 'timing/data': 0.006668881047517061, 'timing/model': 1.5060522190178744, 'learning_rate/qwen_vl_interface': 1.8167243968198905e-06, 'learning_rate/action_model': 1.8167243968198905e-05, 'epoch': 1.64}) +[2026-06-16 05:21:20,682][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.3948190212249756, 'train/grad_norm_pre_clip': 15.31352710723877, 'timing/data': 0.006861164991278201, 'timing/model': 1.690099816012662, 'learning_rate/qwen_vl_interface': 1.8146205255945775e-06, 'learning_rate/action_model': 1.8146205255945777e-05, 'epoch': 1.64}) +[2026-06-16 05:21:26,360][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.06314735859632492, 'train/grad_norm_pre_clip': 21.793611526489258, 'timing/data': 0.00667421001708135, 'timing/model': 1.467427850002423, 'learning_rate/qwen_vl_interface': 1.8125180665240007e-06, 'learning_rate/action_model': 1.8125180665240008e-05, 'epoch': 1.64}) +[2026-06-16 05:21:32,251][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.07723035663366318, 'train/grad_norm_pre_clip': 6.8398942947387695, 'timing/data': 0.006739334960002452, 'timing/model': 1.460159199021291, 'learning_rate/qwen_vl_interface': 1.8104170204724025e-06, 'learning_rate/action_model': 1.8104170204724024e-05, 'epoch': 1.64}) +[2026-06-16 05:21:38,161][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.03979603946208954, 'train/grad_norm_pre_clip': 4.168299198150635, 'timing/data': 0.0068375300033949316, 'timing/model': 1.4664327229838818, 'learning_rate/qwen_vl_interface': 1.8083173883034435e-06, 'learning_rate/action_model': 1.8083173883034435e-05, 'epoch': 1.64}) +[2026-06-16 05:21:44,060][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.20694468915462494, 'train/grad_norm_pre_clip': 16.089012145996094, 'timing/data': 0.007517139951232821, 'timing/model': 1.4630737439729273, 'learning_rate/qwen_vl_interface': 1.8062191708802013e-06, 'learning_rate/action_model': 1.8062191708802012e-05, 'epoch': 1.64}) +[2026-06-16 05:21:49,958][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.04456665739417076, 'train/grad_norm_pre_clip': 8.679977416992188, 'timing/data': 0.006517820002045482, 'timing/model': 1.4325649100355804, 'learning_rate/qwen_vl_interface': 1.804122369065175e-06, 'learning_rate/action_model': 1.8041223690651748e-05, 'epoch': 1.64}) +[2026-06-16 05:21:55,890][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.0029762517660856247, 'train/grad_norm_pre_clip': 1.068954348564148, 'timing/data': 0.007353799999691546, 'timing/model': 1.6990081049734727, 'learning_rate/qwen_vl_interface': 1.8020269837202806e-06, 'learning_rate/action_model': 1.8020269837202804e-05, 'epoch': 1.64}) +[2026-06-16 05:22:01,698][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.028073623776435852, 'train/grad_norm_pre_clip': 3.4653897285461426, 'timing/data': 0.007199182000476867, 'timing/model': 1.6077720450120978, 'learning_rate/qwen_vl_interface': 1.7999330157068517e-06, 'learning_rate/action_model': 1.7999330157068515e-05, 'epoch': 1.64}) +[2026-06-16 05:22:07,353][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.17596435546875, 'train/grad_norm_pre_clip': 7.924301624298096, 'timing/data': 0.006860873952973634, 'timing/model': 1.4593160480144434, 'learning_rate/qwen_vl_interface': 1.7978404658856393e-06, 'learning_rate/action_model': 1.7978404658856394e-05, 'epoch': 1.64}) +[2026-06-16 05:22:13,261][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.023924702778458595, 'train/grad_norm_pre_clip': 5.352138042449951, 'timing/data': 0.007142789021600038, 'timing/model': 1.4683764410438016, 'learning_rate/qwen_vl_interface': 1.7957493351168126e-06, 'learning_rate/action_model': 1.7957493351168126e-05, 'epoch': 1.64}) +[2026-06-16 05:22:19,253][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.048321932554244995, 'train/grad_norm_pre_clip': 10.866049766540527, 'timing/data': 0.0067732149618677795, 'timing/model': 1.4593766929465346, 'learning_rate/qwen_vl_interface': 1.7936596242599552e-06, 'learning_rate/action_model': 1.7936596242599553e-05, 'epoch': 1.64}) +[2026-06-16 05:22:25,211][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.025330420583486557, 'train/grad_norm_pre_clip': 2.8615200519561768, 'timing/data': 0.006790690997149795, 'timing/model': 1.4834617609740235, 'learning_rate/qwen_vl_interface': 1.7915713341740692e-06, 'learning_rate/action_model': 1.791571334174069e-05, 'epoch': 1.64}) +[2026-06-16 05:22:31,083][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.015676824375987053, 'train/grad_norm_pre_clip': 4.849480152130127, 'timing/data': 0.006683072017040104, 'timing/model': 1.6853632209822536, 'learning_rate/qwen_vl_interface': 1.7894844657175716e-06, 'learning_rate/action_model': 1.7894844657175717e-05, 'epoch': 1.64}) +[2026-06-16 05:22:36,985][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.020677806809544563, 'train/grad_norm_pre_clip': 7.738611698150635, 'timing/data': 0.006690052978228778, 'timing/model': 1.6754340450279415, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 1.65}) +[2026-06-16 05:22:42,655][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.010291102342307568, 'train/grad_norm_pre_clip': 1.9301373958587646, 'timing/data': 0.006653268996160477, 'timing/model': 1.4565700159873813, 'learning_rate/qwen_vl_interface': 1.7853149971234883e-06, 'learning_rate/action_model': 1.785314997123488e-05, 'epoch': 1.65}) +[2026-06-16 05:22:48,554][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.02736630104482174, 'train/grad_norm_pre_clip': 5.2738566398620605, 'timing/data': 0.00663224799791351, 'timing/model': 1.4577976139844395, 'learning_rate/qwen_vl_interface': 1.7832323986998128e-06, 'learning_rate/action_model': 1.783232398699813e-05, 'epoch': 1.65}) +[2026-06-16 05:22:54,460][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.09134989976882935, 'train/grad_norm_pre_clip': 17.441259384155273, 'timing/data': 0.0072068130248226225, 'timing/model': 1.4648266779840924, 'learning_rate/qwen_vl_interface': 1.7811512253333468e-06, 'learning_rate/action_model': 1.7811512253333467e-05, 'epoch': 1.65}) +[2026-06-16 05:23:00,466][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.06294980645179749, 'train/grad_norm_pre_clip': 12.23330020904541, 'timing/data': 0.00661760603543371, 'timing/model': 1.4895532520022243, 'learning_rate/qwen_vl_interface': 1.7790714778795823e-06, 'learning_rate/action_model': 1.779071477879582e-05, 'epoch': 1.65}) +[2026-06-16 05:23:06,386][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.13155415654182434, 'train/grad_norm_pre_clip': 36.44356918334961, 'timing/data': 0.006660439015831798, 'timing/model': 1.694179742014967, 'learning_rate/qwen_vl_interface': 1.7769931571934246e-06, 'learning_rate/action_model': 1.7769931571934246e-05, 'epoch': 1.65}) +[2026-06-16 05:23:12,051][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.008791612461209297, 'train/grad_norm_pre_clip': 1.4061533212661743, 'timing/data': 0.006815827975515276, 'timing/model': 1.4555736529873684, 'learning_rate/qwen_vl_interface': 1.7749162641291933e-06, 'learning_rate/action_model': 1.7749162641291933e-05, 'epoch': 1.65}) +[2026-06-16 05:23:17,951][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.118738554418087, 'train/grad_norm_pre_clip': 14.453813552856445, 'timing/data': 0.0064180229674093425, 'timing/model': 1.4577935920096934, 'learning_rate/qwen_vl_interface': 1.7728407995406206e-06, 'learning_rate/action_model': 1.7728407995406205e-05, 'epoch': 1.65}) +[2026-06-16 05:23:23,851][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.10111456364393234, 'train/grad_norm_pre_clip': 6.650951862335205, 'timing/data': 0.0067244539968669415, 'timing/model': 1.4576519869733602, 'learning_rate/qwen_vl_interface': 1.7707667642808513e-06, 'learning_rate/action_model': 1.7707667642808512e-05, 'epoch': 1.65}) +[2026-06-16 05:23:29,750][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.03949207440018654, 'train/grad_norm_pre_clip': 8.087430953979492, 'timing/data': 0.006608422962017357, 'timing/model': 1.4301720979856327, 'learning_rate/qwen_vl_interface': 1.7686941592024423e-06, 'learning_rate/action_model': 1.7686941592024423e-05, 'epoch': 1.65}) +[2026-06-16 05:23:35,643][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.08200924098491669, 'train/grad_norm_pre_clip': 8.745444297790527, 'timing/data': 0.006693759001791477, 'timing/model': 1.5036554670077749, 'learning_rate/qwen_vl_interface': 1.7666229851573662e-06, 'learning_rate/action_model': 1.7666229851573664e-05, 'epoch': 1.65}) +[2026-06-16 05:23:41,531][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.2154846042394638, 'train/grad_norm_pre_clip': 21.770105361938477, 'timing/data': 0.006720938952639699, 'timing/model': 1.6375013700453565, 'learning_rate/qwen_vl_interface': 1.7645532429970006e-06, 'learning_rate/action_model': 1.7645532429970004e-05, 'epoch': 1.65}) +[2026-06-16 05:23:47,249][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.17511360347270966, 'train/grad_norm_pre_clip': 16.38192367553711, 'timing/data': 0.006419715995434672, 'timing/model': 1.455997102020774, 'learning_rate/qwen_vl_interface': 1.7624849335721402e-06, 'learning_rate/action_model': 1.7624849335721402e-05, 'epoch': 1.65}) +[2026-06-16 05:23:53,150][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.088108591735363, 'train/grad_norm_pre_clip': 12.801058769226074, 'timing/data': 0.0066839539795182645, 'timing/model': 1.4584416009602137, 'learning_rate/qwen_vl_interface': 1.7604180577329883e-06, 'learning_rate/action_model': 1.7604180577329883e-05, 'epoch': 1.65}) +[2026-06-16 05:23:59,052][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.16172900795936584, 'train/grad_norm_pre_clip': 7.526435852050781, 'timing/data': 0.006514155014883727, 'timing/model': 1.4580354770296253, 'learning_rate/qwen_vl_interface': 1.7583526163291604e-06, 'learning_rate/action_model': 1.7583526163291604e-05, 'epoch': 1.65}) +[2026-06-16 05:24:04,950][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.08306003361940384, 'train/grad_norm_pre_clip': 9.72514820098877, 'timing/data': 0.0066207609488628805, 'timing/model': 1.4329230599687435, 'learning_rate/qwen_vl_interface': 1.7562886102096807e-06, 'learning_rate/action_model': 1.7562886102096807e-05, 'epoch': 1.65}) +[2026-06-16 05:24:10,852][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.00016259655239991844, 'train/grad_norm_pre_clip': 0.02253430336713791, 'timing/data': 0.006588243006262928, 'timing/model': 1.4301017109537497, 'learning_rate/qwen_vl_interface': 1.7542260402229838e-06, 'learning_rate/action_model': 1.7542260402229838e-05, 'epoch': 1.65}) +[2026-06-16 05:24:16,738][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.03974252566695213, 'train/grad_norm_pre_clip': 4.424752712249756, 'timing/data': 0.006480685027781874, 'timing/model': 1.6443662870442495, 'learning_rate/qwen_vl_interface': 1.7521649072169152e-06, 'learning_rate/action_model': 1.752164907216915e-05, 'epoch': 1.65}) +[2026-06-16 05:24:22,468][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.07988261431455612, 'train/grad_norm_pre_clip': 9.239205360412598, 'timing/data': 0.007630627020262182, 'timing/model': 1.4712585409870371, 'learning_rate/qwen_vl_interface': 1.7501052120387285e-06, 'learning_rate/action_model': 1.7501052120387285e-05, 'epoch': 1.65}) +[2026-06-16 05:24:28,355][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.09763934463262558, 'train/grad_norm_pre_clip': 44.49424362182617, 'timing/data': 0.007636796042788774, 'timing/model': 1.4603725159540772, 'learning_rate/qwen_vl_interface': 1.7480469555350866e-06, 'learning_rate/action_model': 1.7480469555350865e-05, 'epoch': 1.65}) +[2026-06-16 05:24:34,255][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.2140730321407318, 'train/grad_norm_pre_clip': 19.922500610351562, 'timing/data': 0.007300901983398944, 'timing/model': 1.4561186399660073, 'learning_rate/qwen_vl_interface': 1.7459901385520619e-06, 'learning_rate/action_model': 1.745990138552062e-05, 'epoch': 1.65}) +[2026-06-16 05:24:40,234][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.031293608248233795, 'train/grad_norm_pre_clip': 3.5604076385498047, 'timing/data': 0.006831972976215184, 'timing/model': 1.501869207015261, 'learning_rate/qwen_vl_interface': 1.7439347619351324e-06, 'learning_rate/action_model': 1.7439347619351324e-05, 'epoch': 1.65}) +[2026-06-16 05:24:46,184][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.13743796944618225, 'train/grad_norm_pre_clip': 16.09069061279297, 'timing/data': 0.0067044239840470254, 'timing/model': 1.6873611700139008, 'learning_rate/qwen_vl_interface': 1.741880826529186e-06, 'learning_rate/action_model': 1.741880826529186e-05, 'epoch': 1.65}) +[2026-06-16 05:24:51,854][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.042438022792339325, 'train/grad_norm_pre_clip': 7.723976135253906, 'timing/data': 0.006644034991040826, 'timing/model': 1.457596481021028, 'learning_rate/qwen_vl_interface': 1.7398283331785205e-06, 'learning_rate/action_model': 1.7398283331785203e-05, 'epoch': 1.66}) +[2026-06-16 05:24:57,755][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.024684224277734756, 'train/grad_norm_pre_clip': 1.4136853218078613, 'timing/data': 0.006630917021539062, 'timing/model': 1.4594739660387859, 'learning_rate/qwen_vl_interface': 1.737777282726835e-06, 'learning_rate/action_model': 1.737777282726835e-05, 'epoch': 1.66}) +[2026-06-16 05:25:03,554][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.0134139284491539, 'train/grad_norm_pre_clip': 2.190150260925293, 'timing/data': 0.006710504007060081, 'timing/model': 1.4560128760058433, 'learning_rate/qwen_vl_interface': 1.7357276760172403e-06, 'learning_rate/action_model': 1.7357276760172404e-05, 'epoch': 1.66}) +[2026-06-16 05:25:09,454][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.17850209772586823, 'train/grad_norm_pre_clip': 63.69629669189453, 'timing/data': 0.00675094302278012, 'timing/model': 1.4648105219821446, 'learning_rate/qwen_vl_interface': 1.733679513892252e-06, 'learning_rate/action_model': 1.7336795138922517e-05, 'epoch': 1.66}) +[2026-06-16 05:25:15,456][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.24420404434204102, 'train/grad_norm_pre_clip': 14.547170639038086, 'timing/data': 0.006790541985537857, 'timing/model': 1.4779893250088207, 'learning_rate/qwen_vl_interface': 1.7316327971937926e-06, 'learning_rate/action_model': 1.7316327971937926e-05, 'epoch': 1.66}) +[2026-06-16 05:25:21,390][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.21259500086307526, 'train/grad_norm_pre_clip': 30.300670623779297, 'timing/data': 0.0067219199845567346, 'timing/model': 1.6936747439904138, 'learning_rate/qwen_vl_interface': 1.7295875267631904e-06, 'learning_rate/action_model': 1.7295875267631904e-05, 'epoch': 1.66}) +[2026-06-16 05:25:27,056][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.10594076663255692, 'train/grad_norm_pre_clip': 15.39638900756836, 'timing/data': 0.006592890014871955, 'timing/model': 1.4580142190097831, 'learning_rate/qwen_vl_interface': 1.7275437034411778e-06, 'learning_rate/action_model': 1.7275437034411776e-05, 'epoch': 1.66}) +[2026-06-16 05:25:33,057][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.14104144275188446, 'train/grad_norm_pre_clip': 10.58320140838623, 'timing/data': 0.0069721179897896945, 'timing/model': 1.4597275819978677, 'learning_rate/qwen_vl_interface': 1.7255013280678941e-06, 'learning_rate/action_model': 1.725501328067894e-05, 'epoch': 1.66}) +[2026-06-16 05:25:38,955][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.0511842705309391, 'train/grad_norm_pre_clip': 9.106109619140625, 'timing/data': 0.006447255960665643, 'timing/model': 1.4595603679772466, 'learning_rate/qwen_vl_interface': 1.7234604014828824e-06, 'learning_rate/action_model': 1.7234604014828823e-05, 'epoch': 1.66}) +[2026-06-16 05:25:44,934][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.15179693698883057, 'train/grad_norm_pre_clip': 16.8253173828125, 'timing/data': 0.006997865973971784, 'timing/model': 1.4497108270297758, 'learning_rate/qwen_vl_interface': 1.7214209245250906e-06, 'learning_rate/action_model': 1.7214209245250905e-05, 'epoch': 1.66}) +[2026-06-16 05:25:50,941][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.05124961584806442, 'train/grad_norm_pre_clip': 9.112456321716309, 'timing/data': 0.0065897959866560996, 'timing/model': 1.4577357270172797, 'learning_rate/qwen_vl_interface': 1.7193828980328723e-06, 'learning_rate/action_model': 1.719382898032872e-05, 'epoch': 1.66}) +[2026-06-16 05:25:57,001][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.059136129915714264, 'train/grad_norm_pre_clip': 9.655929565429688, 'timing/data': 0.006599951011594385, 'timing/model': 1.7167359800077975, 'learning_rate/qwen_vl_interface': 1.71734632284398e-06, 'learning_rate/action_model': 1.7173463228439798e-05, 'epoch': 1.66}) +[2026-06-16 05:26:02,654][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.010890808887779713, 'train/grad_norm_pre_clip': 5.055582523345947, 'timing/data': 0.006662953004706651, 'timing/model': 1.4594045670237392, 'learning_rate/qwen_vl_interface': 1.7153111997955744e-06, 'learning_rate/action_model': 1.7153111997955743e-05, 'epoch': 1.66}) +[2026-06-16 05:26:08,553][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.0058203707449138165, 'train/grad_norm_pre_clip': 1.2281932830810547, 'timing/data': 0.006402099970728159, 'timing/model': 1.4582176830153912, 'learning_rate/qwen_vl_interface': 1.7132775297242175e-06, 'learning_rate/action_model': 1.7132775297242172e-05, 'epoch': 1.66}) +[2026-06-16 05:26:14,453][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.14952987432479858, 'train/grad_norm_pre_clip': 16.50646209716797, 'timing/data': 0.006303225003648549, 'timing/model': 1.45806175499456, 'learning_rate/qwen_vl_interface': 1.7112453134658747e-06, 'learning_rate/action_model': 1.7112453134658747e-05, 'epoch': 1.66}) +[2026-06-16 05:26:20,357][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.06323853880167007, 'train/grad_norm_pre_clip': 6.301235675811768, 'timing/data': 0.006397683988325298, 'timing/model': 1.4328513839864172, 'learning_rate/qwen_vl_interface': 1.7092145518559129e-06, 'learning_rate/action_model': 1.709214551855913e-05, 'epoch': 1.66}) +[2026-06-16 05:26:26,153][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.05928557738661766, 'train/grad_norm_pre_clip': 10.650771141052246, 'timing/data': 0.006150890025310218, 'timing/model': 1.4565094240242615, 'learning_rate/qwen_vl_interface': 1.7071852457291013e-06, 'learning_rate/action_model': 1.707185245729101e-05, 'epoch': 1.66}) +[2026-06-16 05:26:32,087][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.10215559601783752, 'train/grad_norm_pre_clip': 13.610791206359863, 'timing/data': 0.00649293401511386, 'timing/model': 1.6922786879586056, 'learning_rate/qwen_vl_interface': 1.7051573959196122e-06, 'learning_rate/action_model': 1.7051573959196122e-05, 'epoch': 1.66}) +[2026-06-16 05:26:37,761][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.07272262126207352, 'train/grad_norm_pre_clip': 18.94123077392578, 'timing/data': 0.007484331959858537, 'timing/model': 1.4616831800085492, 'learning_rate/qwen_vl_interface': 1.7031310032610166e-06, 'learning_rate/action_model': 1.7031310032610163e-05, 'epoch': 1.66}) +[2026-06-16 05:26:43,655][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.00919784139841795, 'train/grad_norm_pre_clip': 2.033707857131958, 'timing/data': 0.006810980965383351, 'timing/model': 1.4562233589822426, 'learning_rate/qwen_vl_interface': 1.701106068586289e-06, 'learning_rate/action_model': 1.701106068586289e-05, 'epoch': 1.66}) +[2026-06-16 05:26:49,554][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.20355968177318573, 'train/grad_norm_pre_clip': 14.24213981628418, 'timing/data': 0.006643765023909509, 'timing/model': 1.4563107099966146, 'learning_rate/qwen_vl_interface': 1.6990825927278032e-06, 'learning_rate/action_model': 1.699082592727803e-05, 'epoch': 1.66}) +[2026-06-16 05:26:55,354][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.0023216893896460533, 'train/grad_norm_pre_clip': 0.5400858521461487, 'timing/data': 0.0068369999644346535, 'timing/model': 1.4570086379535496, 'learning_rate/qwen_vl_interface': 1.6970605765173344e-06, 'learning_rate/action_model': 1.6970605765173346e-05, 'epoch': 1.66}) +[2026-06-16 05:27:01,254][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.31070005893707275, 'train/grad_norm_pre_clip': 13.809250831604004, 'timing/data': 0.006763512035831809, 'timing/model': 1.4454987549688667, 'learning_rate/qwen_vl_interface': 1.6950400207860557e-06, 'learning_rate/action_model': 1.6950400207860557e-05, 'epoch': 1.66}) +[2026-06-16 05:27:07,185][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.07901597023010254, 'train/grad_norm_pre_clip': 20.22055435180664, 'timing/data': 0.006483499950263649, 'timing/model': 1.6890267049893737, 'learning_rate/qwen_vl_interface': 1.6930209263645412e-06, 'learning_rate/action_model': 1.6930209263645412e-05, 'epoch': 1.67}) +[2026-06-16 05:27:12,855][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.04110441729426384, 'train/grad_norm_pre_clip': 6.472129821777344, 'timing/data': 0.0065127029665745795, 'timing/model': 1.459027294011321, 'learning_rate/qwen_vl_interface': 1.6910032940827656e-06, 'learning_rate/action_model': 1.6910032940827657e-05, 'epoch': 1.67}) +[2026-06-16 05:27:18,755][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.23342587053775787, 'train/grad_norm_pre_clip': 16.00810432434082, 'timing/data': 0.006262924987822771, 'timing/model': 1.4614972390118055, 'learning_rate/qwen_vl_interface': 1.6889871247701003e-06, 'learning_rate/action_model': 1.6889871247701002e-05, 'epoch': 1.67}) +[2026-06-16 05:27:24,655][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.0935848131775856, 'train/grad_norm_pre_clip': 10.859929084777832, 'timing/data': 0.006892281002365053, 'timing/model': 1.460911028960254, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 1.67}) +[2026-06-16 05:27:30,553][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.16057796776294708, 'train/grad_norm_pre_clip': 15.514641761779785, 'timing/data': 0.006578049040399492, 'timing/model': 1.4579917850205675, 'learning_rate/qwen_vl_interface': 1.6849591783665847e-06, 'learning_rate/action_model': 1.6849591783665847e-05, 'epoch': 1.67}) +[2026-06-16 05:27:36,456][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.08371087163686752, 'train/grad_norm_pre_clip': 9.408563613891602, 'timing/data': 0.006487535953056067, 'timing/model': 1.428858199971728, 'learning_rate/qwen_vl_interface': 1.6829474029314713e-06, 'learning_rate/action_model': 1.6829474029314713e-05, 'epoch': 1.67}) +[2026-06-16 05:27:42,391][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.10879375040531158, 'train/grad_norm_pre_clip': 15.650707244873047, 'timing/data': 0.007161527988500893, 'timing/model': 1.7023027300019749, 'learning_rate/qwen_vl_interface': 1.6809370937769408e-06, 'learning_rate/action_model': 1.680937093776941e-05, 'epoch': 1.67}) +[2026-06-16 05:27:48,049][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.03407870978116989, 'train/grad_norm_pre_clip': 4.8358964920043945, 'timing/data': 0.006689553032629192, 'timing/model': 1.458354914037045, 'learning_rate/qwen_vl_interface': 1.678928251729356e-06, 'learning_rate/action_model': 1.6789282517293557e-05, 'epoch': 1.67}) +[2026-06-16 05:27:53,950][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.38026726245880127, 'train/grad_norm_pre_clip': 31.681413650512695, 'timing/data': 0.006663894979283214, 'timing/model': 1.4591959830140695, 'learning_rate/qwen_vl_interface': 1.6769208776144767e-06, 'learning_rate/action_model': 1.6769208776144764e-05, 'epoch': 1.67}) +[2026-06-16 05:27:59,848][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.12549875676631927, 'train/grad_norm_pre_clip': 15.287897109985352, 'timing/data': 0.006679737998638302, 'timing/model': 1.4568547090166248, 'learning_rate/qwen_vl_interface': 1.6749149722574564e-06, 'learning_rate/action_model': 1.6749149722574564e-05, 'epoch': 1.67}) +[2026-06-16 05:28:05,748][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.13459771871566772, 'train/grad_norm_pre_clip': 12.739883422851562, 'timing/data': 0.006642232998274267, 'timing/model': 1.4604665160295554, 'learning_rate/qwen_vl_interface': 1.6729105364828486e-06, 'learning_rate/action_model': 1.6729105364828484e-05, 'epoch': 1.67}) +[2026-06-16 05:28:11,650][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.12438412010669708, 'train/grad_norm_pre_clip': 14.522156715393066, 'timing/data': 0.006547825003508478, 'timing/model': 1.4271735710208304, 'learning_rate/qwen_vl_interface': 1.6709075711146e-06, 'learning_rate/action_model': 1.6709075711146e-05, 'epoch': 1.67}) +[2026-06-16 05:28:17,590][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.026217734441161156, 'train/grad_norm_pre_clip': 3.0311379432678223, 'timing/data': 0.007440968009177595, 'timing/model': 1.6959282319876365, 'learning_rate/qwen_vl_interface': 1.6689060769760556e-06, 'learning_rate/action_model': 1.6689060769760557e-05, 'epoch': 1.67}) +[2026-06-16 05:28:23,250][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.13972535729408264, 'train/grad_norm_pre_clip': 12.893938064575195, 'timing/data': 0.006800666975323111, 'timing/model': 1.4582757069729269, 'learning_rate/qwen_vl_interface': 1.6669060548899528e-06, 'learning_rate/action_model': 1.6669060548899525e-05, 'epoch': 1.67}) +[2026-06-16 05:28:29,152][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.034429874271154404, 'train/grad_norm_pre_clip': 5.714568138122559, 'timing/data': 0.006615142978262156, 'timing/model': 1.4596938340109773, 'learning_rate/qwen_vl_interface': 1.6649075056784253e-06, 'learning_rate/action_model': 1.6649075056784253e-05, 'epoch': 1.67}) +[2026-06-16 05:28:35,053][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.24690544605255127, 'train/grad_norm_pre_clip': 16.304981231689453, 'timing/data': 0.0067010700004175305, 'timing/model': 1.4608818950364366, 'learning_rate/qwen_vl_interface': 1.6629104301630019e-06, 'learning_rate/action_model': 1.662910430163002e-05, 'epoch': 1.67}) +[2026-06-16 05:28:40,950][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.14847658574581146, 'train/grad_norm_pre_clip': 19.1693172454834, 'timing/data': 0.006871199992019683, 'timing/model': 1.4560926689882763, 'learning_rate/qwen_vl_interface': 1.6609148291646043e-06, 'learning_rate/action_model': 1.6609148291646042e-05, 'epoch': 1.67}) +[2026-06-16 05:28:46,854][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.09009373188018799, 'train/grad_norm_pre_clip': 13.270087242126465, 'timing/data': 0.006633981014601886, 'timing/model': 1.4317332890350372, 'learning_rate/qwen_vl_interface': 1.658920703503549e-06, 'learning_rate/action_model': 1.6589207035035488e-05, 'epoch': 1.67}) +[2026-06-16 05:28:52,778][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.012471825815737247, 'train/grad_norm_pre_clip': 2.4197471141815186, 'timing/data': 0.006639158993493766, 'timing/model': 1.6838758590165526, 'learning_rate/qwen_vl_interface': 1.6569280539995467e-06, 'learning_rate/action_model': 1.6569280539995467e-05, 'epoch': 1.67}) +[2026-06-16 05:28:58,616][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.11332018673419952, 'train/grad_norm_pre_clip': 11.498642921447754, 'timing/data': 0.007597790041472763, 'timing/model': 1.6216564110363834, 'learning_rate/qwen_vl_interface': 1.6549368814716977e-06, 'learning_rate/action_model': 1.6549368814716977e-05, 'epoch': 1.67}) +[2026-06-16 05:29:04,354][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.007325512822717428, 'train/grad_norm_pre_clip': 1.4448702335357666, 'timing/data': 0.006977726996410638, 'timing/model': 1.457709211972542, 'learning_rate/qwen_vl_interface': 1.6529471867385009e-06, 'learning_rate/action_model': 1.6529471867385007e-05, 'epoch': 1.67}) +[2026-06-16 05:29:10,258][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.12479671835899353, 'train/grad_norm_pre_clip': 12.851156234741211, 'timing/data': 0.006599561020266265, 'timing/model': 1.4602614300092682, 'learning_rate/qwen_vl_interface': 1.6509589706178417e-06, 'learning_rate/action_model': 1.6509589706178417e-05, 'epoch': 1.67}) +[2026-06-16 05:29:16,254][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.16485051810741425, 'train/grad_norm_pre_clip': 15.1467924118042, 'timing/data': 0.006810721009969711, 'timing/model': 1.4574944900232367, 'learning_rate/qwen_vl_interface': 1.648972233927002e-06, 'learning_rate/action_model': 1.648972233927002e-05, 'epoch': 1.67}) +[2026-06-16 05:29:22,161][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.05957283452153206, 'train/grad_norm_pre_clip': 12.036885261535645, 'timing/data': 0.006670454982668161, 'timing/model': 1.43242841004394, 'learning_rate/qwen_vl_interface': 1.646986977482654e-06, 'learning_rate/action_model': 1.6469869774826538e-05, 'epoch': 1.68}) +[2026-06-16 05:29:28,078][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.09707866609096527, 'train/grad_norm_pre_clip': 18.291641235351562, 'timing/data': 0.007540375052485615, 'timing/model': 1.6686293939710595, 'learning_rate/qwen_vl_interface': 1.6450032021008618e-06, 'learning_rate/action_model': 1.6450032021008617e-05, 'epoch': 1.68}) +[2026-06-16 05:29:33,992][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.030912606045603752, 'train/grad_norm_pre_clip': 6.487478256225586, 'timing/data': 0.006943296000827104, 'timing/model': 1.6985498400172219, 'learning_rate/qwen_vl_interface': 1.6430209085970812e-06, 'learning_rate/action_model': 1.643020908597081e-05, 'epoch': 1.68}) +[2026-06-16 05:29:39,654][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.01152975857257843, 'train/grad_norm_pre_clip': 2.1045286655426025, 'timing/data': 0.006732436013408005, 'timing/model': 1.4602236609789543, 'learning_rate/qwen_vl_interface': 1.641040097786157e-06, 'learning_rate/action_model': 1.641040097786157e-05, 'epoch': 1.68}) +[2026-06-16 05:29:45,554][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.11673320084810257, 'train/grad_norm_pre_clip': 5.652029037475586, 'timing/data': 0.006554884021170437, 'timing/model': 1.4584524260135368, 'learning_rate/qwen_vl_interface': 1.6390607704823267e-06, 'learning_rate/action_model': 1.6390607704823265e-05, 'epoch': 1.68}) +[2026-06-16 05:29:51,554][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.23437504470348358, 'train/grad_norm_pre_clip': 11.856285095214844, 'timing/data': 0.006426054984331131, 'timing/model': 1.4455850339727476, 'learning_rate/qwen_vl_interface': 1.6370829274992162e-06, 'learning_rate/action_model': 1.6370829274992162e-05, 'epoch': 1.68}) +[2026-06-16 05:29:57,553][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.07281762361526489, 'train/grad_norm_pre_clip': 14.963479042053223, 'timing/data': 0.006636695004999638, 'timing/model': 1.4981553870020434, 'learning_rate/qwen_vl_interface': 1.6351065696498422e-06, 'learning_rate/action_model': 1.6351065696498422e-05, 'epoch': 1.68}) +[2026-06-16 05:39:14,550][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.027122678235173225, 'train/grad_norm_pre_clip': 8.70996379852295, 'eval/loss': 0.04857712611556053, 'eval/flap_precision': 0.9508196721311475, 'eval/flap_recall': 0.9298597194388778, 'eval/flap_f1': 0.9402228976697062, 'eval/flap_rate_model': 0.13555555555555557, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9402228976697061, 'timing/data': 0.006808337988331914, 'timing/model': 1.6879750089719892, 'learning_rate/qwen_vl_interface': 1.6331316977466117e-06, 'learning_rate/action_model': 1.6331316977466115e-05, 'epoch': 1.68}) +[2026-06-16 05:39:14,551][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3900_state +[2026-06-16 05:39:14,556][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 05:39:19,977][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3900_state/model.safetensors +[2026-06-16 05:39:28,029][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3900_state/optimizer.bin +[2026-06-16 05:39:28,030][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3900_state/sampler.bin +[2026-06-16 05:39:28,030][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3900_state/sampler_1.bin +[2026-06-16 05:39:28,031][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_3900_state/random_states_0.pkl +[2026-06-16 05:39:28,142][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3900 +[2026-06-16 05:39:28,143][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 05:39:28,153][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 05:39:34,101][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.17579875886440277, 'train/grad_norm_pre_clip': 12.086437225341797, 'timing/data': 0.0074808080098591745, 'timing/model': 1.7224613250000402, 'learning_rate/qwen_vl_interface': 1.6311583126013198e-06, 'learning_rate/action_model': 1.6311583126013198e-05, 'epoch': 1.68}) +[2026-06-16 05:39:39,844][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.0019272193312644958, 'train/grad_norm_pre_clip': 0.28339120745658875, 'timing/data': 0.006857790984213352, 'timing/model': 1.5629764719633386, 'learning_rate/qwen_vl_interface': 1.6291864150251488e-06, 'learning_rate/action_model': 1.629186415025149e-05, 'epoch': 1.68}) +[2026-06-16 05:39:45,556][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.09924115985631943, 'train/grad_norm_pre_clip': 8.959259033203125, 'timing/data': 0.006717735028360039, 'timing/model': 1.4735358160105534, 'learning_rate/qwen_vl_interface': 1.6272160058286727e-06, 'learning_rate/action_model': 1.6272160058286726e-05, 'epoch': 1.68}) +[2026-06-16 05:39:51,444][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.0021532706450670958, 'train/grad_norm_pre_clip': 0.28542089462280273, 'timing/data': 0.006994522002059966, 'timing/model': 1.459602995950263, 'learning_rate/qwen_vl_interface': 1.6252470858218519e-06, 'learning_rate/action_model': 1.6252470858218517e-05, 'epoch': 1.68}) +[2026-06-16 05:39:57,443][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.1145162284374237, 'train/grad_norm_pre_clip': 10.26074504852295, 'timing/data': 0.010538757022004575, 'timing/model': 1.465902414987795, 'learning_rate/qwen_vl_interface': 1.6232796558140357e-06, 'learning_rate/action_model': 1.6232796558140357e-05, 'epoch': 1.68}) +[2026-06-16 05:40:03,534][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.11336931586265564, 'train/grad_norm_pre_clip': 11.997995376586914, 'timing/data': 0.014548256993293762, 'timing/model': 1.499402829038445, 'learning_rate/qwen_vl_interface': 1.6213137166139592e-06, 'learning_rate/action_model': 1.621313716613959e-05, 'epoch': 1.68}) +[2026-06-16 05:40:09,480][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.024314245209097862, 'train/grad_norm_pre_clip': 3.586400270462036, 'timing/data': 0.006548175006173551, 'timing/model': 1.6987368480185978, 'learning_rate/qwen_vl_interface': 1.6193492690297463e-06, 'learning_rate/action_model': 1.6193492690297464e-05, 'epoch': 1.68}) +[2026-06-16 05:40:15,145][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.12464284896850586, 'train/grad_norm_pre_clip': 8.755362510681152, 'timing/data': 0.006823609990533441, 'timing/model': 1.4598636100417934, 'learning_rate/qwen_vl_interface': 1.6173863138689078e-06, 'learning_rate/action_model': 1.617386313868908e-05, 'epoch': 1.68}) +[2026-06-16 05:40:21,047][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.12632159888744354, 'train/grad_norm_pre_clip': 8.73141860961914, 'timing/data': 0.006788018043152988, 'timing/model': 1.459348973003216, 'learning_rate/qwen_vl_interface': 1.6154248519383396e-06, 'learning_rate/action_model': 1.6154248519383395e-05, 'epoch': 1.68}) +[2026-06-16 05:40:26,949][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.09023751318454742, 'train/grad_norm_pre_clip': 7.493931770324707, 'timing/data': 0.006487976002972573, 'timing/model': 1.4616752869915217, 'learning_rate/qwen_vl_interface': 1.6134648840443248e-06, 'learning_rate/action_model': 1.613464884044325e-05, 'epoch': 1.68}) +[2026-06-16 05:40:32,853][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.021764438599348068, 'train/grad_norm_pre_clip': 2.7941319942474365, 'timing/data': 0.006938249978702515, 'timing/model': 1.4326886750059202, 'learning_rate/qwen_vl_interface': 1.6115064109925334e-06, 'learning_rate/action_model': 1.6115064109925336e-05, 'epoch': 1.68}) +[2026-06-16 05:40:38,751][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.14210127294063568, 'train/grad_norm_pre_clip': 11.306599617004395, 'timing/data': 0.006942045001778752, 'timing/model': 1.4847755579976365, 'learning_rate/qwen_vl_interface': 1.6095494335880173e-06, 'learning_rate/action_model': 1.6095494335880174e-05, 'epoch': 1.68}) +[2026-06-16 05:40:44,655][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.1130823865532875, 'train/grad_norm_pre_clip': 10.10061264038086, 'timing/data': 0.006642635038588196, 'timing/model': 1.6619723339681514, 'learning_rate/qwen_vl_interface': 1.6075939526352175e-06, 'learning_rate/action_model': 1.6075939526352175e-05, 'epoch': 1.68}) +[2026-06-16 05:40:50,352][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.020490771159529686, 'train/grad_norm_pre_clip': 2.2522525787353516, 'timing/data': 0.0065004859934560955, 'timing/model': 1.4594910939922556, 'learning_rate/qwen_vl_interface': 1.6056399689379576e-06, 'learning_rate/action_model': 1.6056399689379575e-05, 'epoch': 1.68}) +[2026-06-16 05:40:56,353][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.025108754634857178, 'train/grad_norm_pre_clip': 2.8004002571105957, 'timing/data': 0.006215376022737473, 'timing/model': 1.460205245995894, 'learning_rate/qwen_vl_interface': 1.6036874832994473e-06, 'learning_rate/action_model': 1.603687483299447e-05, 'epoch': 1.68}) +[2026-06-16 05:41:02,252][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.16660144925117493, 'train/grad_norm_pre_clip': 10.847670555114746, 'timing/data': 0.006783731980249286, 'timing/model': 1.458967323997058, 'learning_rate/qwen_vl_interface': 1.6017364965222796e-06, 'learning_rate/action_model': 1.6017364965222793e-05, 'epoch': 1.68}) +[2026-06-16 05:41:08,156][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.034991554915905, 'train/grad_norm_pre_clip': 9.517987251281738, 'timing/data': 0.006417281983885914, 'timing/model': 1.4301278890343383, 'learning_rate/qwen_vl_interface': 1.5997870094084305e-06, 'learning_rate/action_model': 1.5997870094084306e-05, 'epoch': 1.69}) +[2026-06-16 05:41:14,067][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.0983930304646492, 'train/grad_norm_pre_clip': 7.7517876625061035, 'timing/data': 0.006572471000254154, 'timing/model': 1.4878034550347365, 'learning_rate/qwen_vl_interface': 1.5978390227592616e-06, 'learning_rate/action_model': 1.5978390227592615e-05, 'epoch': 1.69}) +[2026-06-16 05:41:19,999][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.03257337212562561, 'train/grad_norm_pre_clip': 10.120996475219727, 'timing/data': 0.006606029986869544, 'timing/model': 1.7035235569928773, 'learning_rate/qwen_vl_interface': 1.5958925373755164e-06, 'learning_rate/action_model': 1.5958925373755164e-05, 'epoch': 1.69}) +[2026-06-16 05:41:25,655][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.19316789507865906, 'train/grad_norm_pre_clip': 10.229194641113281, 'timing/data': 0.005818450998049229, 'timing/model': 1.459656308987178, 'learning_rate/qwen_vl_interface': 1.5939475540573213e-06, 'learning_rate/action_model': 1.5939475540573212e-05, 'epoch': 1.69}) +[2026-06-16 05:41:31,555][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.00465902267023921, 'train/grad_norm_pre_clip': 1.6061949729919434, 'timing/data': 0.006097853009123355, 'timing/model': 1.4604468530160375, 'learning_rate/qwen_vl_interface': 1.5920040736041867e-06, 'learning_rate/action_model': 1.5920040736041866e-05, 'epoch': 1.69}) +[2026-06-16 05:41:37,555][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.0671185776591301, 'train/grad_norm_pre_clip': 6.435245990753174, 'timing/data': 0.006590857985429466, 'timing/model': 1.487970453046728, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 1.69}) +[2026-06-16 05:41:43,454][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.08485554903745651, 'train/grad_norm_pre_clip': 8.528665542602539, 'timing/data': 0.006634532008320093, 'timing/model': 1.4582203880418092, 'learning_rate/qwen_vl_interface': 1.5881216244880407e-06, 'learning_rate/action_model': 1.5881216244880407e-05, 'epoch': 1.69}) +[2026-06-16 05:41:49,356][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.17220444977283478, 'train/grad_norm_pre_clip': 17.228504180908203, 'timing/data': 0.006713829003274441, 'timing/model': 1.4692592699429952, 'learning_rate/qwen_vl_interface': 1.5861826574209587e-06, 'learning_rate/action_model': 1.5861826574209587e-05, 'epoch': 1.69}) +[2026-06-16 05:41:55,289][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.025898467749357224, 'train/grad_norm_pre_clip': 8.102622032165527, 'timing/data': 0.006696673983242363, 'timing/model': 1.6929099120316096, 'learning_rate/qwen_vl_interface': 1.5842451964107916e-06, 'learning_rate/action_model': 1.5842451964107916e-05, 'epoch': 1.69}) +[2026-06-16 05:42:00,955][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.09346385300159454, 'train/grad_norm_pre_clip': 11.158071517944336, 'timing/data': 0.0064731750171631575, 'timing/model': 1.461286059988197, 'learning_rate/qwen_vl_interface': 1.5823092422539567e-06, 'learning_rate/action_model': 1.582309242253957e-05, 'epoch': 1.69}) +[2026-06-16 05:42:06,955][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.032062601298093796, 'train/grad_norm_pre_clip': 4.3909711837768555, 'timing/data': 0.00658704194938764, 'timing/model': 1.4596382739837281, 'learning_rate/qwen_vl_interface': 1.5803747957462514e-06, 'learning_rate/action_model': 1.5803747957462514e-05, 'epoch': 1.69}) +[2026-06-16 05:42:12,953][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.10595329105854034, 'train/grad_norm_pre_clip': 11.308479309082031, 'timing/data': 0.007042012002784759, 'timing/model': 1.462422222015448, 'learning_rate/qwen_vl_interface': 1.5784418576828541e-06, 'learning_rate/action_model': 1.5784418576828543e-05, 'epoch': 1.69}) +[2026-06-16 05:42:18,860][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.29907092452049255, 'train/grad_norm_pre_clip': 19.763093948364258, 'timing/data': 0.006774467998184264, 'timing/model': 1.4333189380122349, 'learning_rate/qwen_vl_interface': 1.576510428858322e-06, 'learning_rate/action_model': 1.5765104288583217e-05, 'epoch': 1.69}) +[2026-06-16 05:42:24,655][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.053868912160396576, 'train/grad_norm_pre_clip': 9.800332069396973, 'timing/data': 0.006659169041085988, 'timing/model': 1.4547581850201823, 'learning_rate/qwen_vl_interface': 1.5745805100665922e-06, 'learning_rate/action_model': 1.5745805100665923e-05, 'epoch': 1.69}) +[2026-06-16 05:42:30,598][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.03483569994568825, 'train/grad_norm_pre_clip': 9.760815620422363, 'timing/data': 0.00662848399952054, 'timing/model': 1.7026053100125864, 'learning_rate/qwen_vl_interface': 1.572652102100983e-06, 'learning_rate/action_model': 1.572652102100983e-05, 'epoch': 1.69}) +[2026-06-16 05:42:36,254][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.03408612683415413, 'train/grad_norm_pre_clip': 10.445136070251465, 'timing/data': 0.00679808302083984, 'timing/model': 1.457848522986751, 'learning_rate/qwen_vl_interface': 1.5707252057541898e-06, 'learning_rate/action_model': 1.5707252057541896e-05, 'epoch': 1.69}) +[2026-06-16 05:42:42,318][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.20330481231212616, 'train/grad_norm_pre_clip': 16.11690330505371, 'timing/data': 0.006775729998480529, 'timing/model': 1.5238540269783698, 'learning_rate/qwen_vl_interface': 1.5687998218182849e-06, 'learning_rate/action_model': 1.5687998218182847e-05, 'epoch': 1.69}) +[2026-06-16 05:42:48,355][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.030400313436985016, 'train/grad_norm_pre_clip': 5.739621162414551, 'timing/data': 0.006687148998025805, 'timing/model': 1.4760309180128388, 'learning_rate/qwen_vl_interface': 1.5668759510847235e-06, 'learning_rate/action_model': 1.5668759510847233e-05, 'epoch': 1.69}) +[2026-06-16 05:42:54,162][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.09575723856687546, 'train/grad_norm_pre_clip': 8.644232749938965, 'timing/data': 0.006642153020948172, 'timing/model': 1.4626517810393125, 'learning_rate/qwen_vl_interface': 1.5649535943443331e-06, 'learning_rate/action_model': 1.564953594344333e-05, 'epoch': 1.69}) +[2026-06-16 05:43:00,065][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.12730486690998077, 'train/grad_norm_pre_clip': 9.36632251739502, 'timing/data': 0.0067330870078876615, 'timing/model': 1.4580028009950183, 'learning_rate/qwen_vl_interface': 1.5630327523873244e-06, 'learning_rate/action_model': 1.563032752387324e-05, 'epoch': 1.69}) +[2026-06-16 05:43:06,034][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.10953577607870102, 'train/grad_norm_pre_clip': 11.256498336791992, 'timing/data': 0.006423812999855727, 'timing/model': 1.7418043519719504, 'learning_rate/qwen_vl_interface': 1.561113426003282e-06, 'learning_rate/action_model': 1.561113426003282e-05, 'epoch': 1.69}) +[2026-06-16 05:43:11,749][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.16532354056835175, 'train/grad_norm_pre_clip': 15.269499778747559, 'timing/data': 0.006891880999319255, 'timing/model': 1.4576006490387954, 'learning_rate/qwen_vl_interface': 1.5591956159811686e-06, 'learning_rate/action_model': 1.5591956159811687e-05, 'epoch': 1.69}) +[2026-06-16 05:43:17,769][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.2659173309803009, 'train/grad_norm_pre_clip': 18.153289794921875, 'timing/data': 0.0066707649966701865, 'timing/model': 1.4748799860244617, 'learning_rate/qwen_vl_interface': 1.5572793231093235e-06, 'learning_rate/action_model': 1.5572793231093236e-05, 'epoch': 1.69}) +[2026-06-16 05:43:23,662][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.07635480165481567, 'train/grad_norm_pre_clip': 10.02994441986084, 'timing/data': 0.006540555041283369, 'timing/model': 1.4651451819809154, 'learning_rate/qwen_vl_interface': 1.5553645481754626e-06, 'learning_rate/action_model': 1.5553645481754626e-05, 'epoch': 1.7}) +[2026-06-16 05:43:29,553][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.059879597276449203, 'train/grad_norm_pre_clip': 6.998565673828125, 'timing/data': 0.006731233967002481, 'timing/model': 1.4491285340045579, 'learning_rate/qwen_vl_interface': 1.5534512919666775e-06, 'learning_rate/action_model': 1.5534512919666776e-05, 'epoch': 1.7}) +[2026-06-16 05:43:35,453][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.004378276411443949, 'train/grad_norm_pre_clip': 1.7419421672821045, 'timing/data': 0.006554344028700143, 'timing/model': 1.4328688889509067, 'learning_rate/qwen_vl_interface': 1.551539555269435e-06, 'learning_rate/action_model': 1.5515395552694348e-05, 'epoch': 1.7}) +[2026-06-16 05:43:41,378][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.3015153408050537, 'train/grad_norm_pre_clip': 13.029951095581055, 'timing/data': 0.007546545006334782, 'timing/model': 1.6855574289802462, 'learning_rate/qwen_vl_interface': 1.5496293388695787e-06, 'learning_rate/action_model': 1.5496293388695786e-05, 'epoch': 1.7}) +[2026-06-16 05:43:47,052][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.2612307369709015, 'train/grad_norm_pre_clip': 10.18042278289795, 'timing/data': 0.0069546730373986065, 'timing/model': 1.4584573520114645, 'learning_rate/qwen_vl_interface': 1.547720643552327e-06, 'learning_rate/action_model': 1.547720643552327e-05, 'epoch': 1.7}) +[2026-06-16 05:43:53,060][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.14211857318878174, 'train/grad_norm_pre_clip': 16.238933563232422, 'timing/data': 0.006420748017262667, 'timing/model': 1.4683198439888656, 'learning_rate/qwen_vl_interface': 1.5458134701022706e-06, 'learning_rate/action_model': 1.5458134701022705e-05, 'epoch': 1.7}) +[2026-06-16 05:43:58,964][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.09148790687322617, 'train/grad_norm_pre_clip': 9.181001663208008, 'timing/data': 0.007013269001618028, 'timing/model': 1.4540081049781293, 'learning_rate/qwen_vl_interface': 1.5439078193033765e-06, 'learning_rate/action_model': 1.5439078193033766e-05, 'epoch': 1.7}) +[2026-06-16 05:44:04,850][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.07288393378257751, 'train/grad_norm_pre_clip': 11.831228256225586, 'timing/data': 0.006579421053174883, 'timing/model': 1.4512400520034134, 'learning_rate/qwen_vl_interface': 1.5420036919389872e-06, 'learning_rate/action_model': 1.5420036919389873e-05, 'epoch': 1.7}) +[2026-06-16 05:44:10,865][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.16503043472766876, 'train/grad_norm_pre_clip': 16.79328155517578, 'timing/data': 0.006704034982249141, 'timing/model': 1.529932285018731, 'learning_rate/qwen_vl_interface': 1.540101088791817e-06, 'learning_rate/action_model': 1.5401010887918168e-05, 'epoch': 1.7}) +[2026-06-16 05:44:16,784][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.16026973724365234, 'train/grad_norm_pre_clip': 13.014575004577637, 'timing/data': 0.006546463002450764, 'timing/model': 1.6916910060099326, 'learning_rate/qwen_vl_interface': 1.5382000106439532e-06, 'learning_rate/action_model': 1.5382000106439532e-05, 'epoch': 1.7}) +[2026-06-16 05:44:22,466][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.012683101929724216, 'train/grad_norm_pre_clip': 1.5644567012786865, 'timing/data': 0.006562125985510647, 'timing/model': 1.4731226419680752, 'learning_rate/qwen_vl_interface': 1.5363004582768584e-06, 'learning_rate/action_model': 1.5363004582768584e-05, 'epoch': 1.7}) +[2026-06-16 05:44:28,353][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.009214606136083603, 'train/grad_norm_pre_clip': 1.3524410724639893, 'timing/data': 0.006397343007847667, 'timing/model': 1.4593408229993656, 'learning_rate/qwen_vl_interface': 1.534402432471367e-06, 'learning_rate/action_model': 1.534402432471367e-05, 'epoch': 1.7}) +[2026-06-16 05:44:34,253][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.09111201763153076, 'train/grad_norm_pre_clip': 5.413901329040527, 'timing/data': 0.006546913995407522, 'timing/model': 1.458380661031697, 'learning_rate/qwen_vl_interface': 1.532505934007685e-06, 'learning_rate/action_model': 1.532505934007685e-05, 'epoch': 1.7}) +[2026-06-16 05:44:40,270][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.1696571558713913, 'train/grad_norm_pre_clip': 9.512494087219238, 'timing/data': 0.006491572014056146, 'timing/model': 1.496963823970873, 'learning_rate/qwen_vl_interface': 1.5306109636653915e-06, 'learning_rate/action_model': 1.5306109636653914e-05, 'epoch': 1.7}) +[2026-06-16 05:44:46,156][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.22798189520835876, 'train/grad_norm_pre_clip': 16.459178924560547, 'timing/data': 0.0065505689708516, 'timing/model': 1.4330391230178066, 'learning_rate/qwen_vl_interface': 1.5287175222234383e-06, 'learning_rate/action_model': 1.528717522223438e-05, 'epoch': 1.7}) +[2026-06-16 05:44:52,079][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.023462843149900436, 'train/grad_norm_pre_clip': 4.067590236663818, 'timing/data': 0.006467066006734967, 'timing/model': 1.6830864860094152, 'learning_rate/qwen_vl_interface': 1.5268256104601457e-06, 'learning_rate/action_model': 1.5268256104601454e-05, 'epoch': 1.7}) +[2026-06-16 05:44:57,903][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.009686751291155815, 'train/grad_norm_pre_clip': 2.1441025733947754, 'timing/data': 0.006389481946825981, 'timing/model': 1.6100761250127107, 'learning_rate/qwen_vl_interface': 1.5249352291532075e-06, 'learning_rate/action_model': 1.5249352291532073e-05, 'epoch': 1.7}) +[2026-06-16 05:45:03,555][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.052067361772060394, 'train/grad_norm_pre_clip': 6.172170639038086, 'timing/data': 0.006496189045719802, 'timing/model': 1.459065439994447, 'learning_rate/qwen_vl_interface': 1.523046379079689e-06, 'learning_rate/action_model': 1.523046379079689e-05, 'epoch': 1.7}) +[2026-06-16 05:45:09,463][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.12047573179006577, 'train/grad_norm_pre_clip': 6.60139274597168, 'timing/data': 0.006670995964668691, 'timing/model': 1.4694797339616343, 'learning_rate/qwen_vl_interface': 1.5211590610160244e-06, 'learning_rate/action_model': 1.5211590610160242e-05, 'epoch': 1.7}) +[2026-06-16 05:45:15,351][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.19995884597301483, 'train/grad_norm_pre_clip': 9.118708610534668, 'timing/data': 0.006660500017460436, 'timing/model': 1.456460346991662, 'learning_rate/qwen_vl_interface': 1.5192732757380193e-06, 'learning_rate/action_model': 1.5192732757380194e-05, 'epoch': 1.7}) +[2026-06-16 05:45:21,256][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.07179049402475357, 'train/grad_norm_pre_clip': 5.591326713562012, 'timing/data': 0.006673519033938646, 'timing/model': 1.4328767550177872, 'learning_rate/qwen_vl_interface': 1.517389024020849e-06, 'learning_rate/action_model': 1.5173890240208488e-05, 'epoch': 1.7}) +[2026-06-16 05:45:27,178][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.12825125455856323, 'train/grad_norm_pre_clip': 13.920595169067383, 'timing/data': 0.006504732009489089, 'timing/model': 1.6824601959669963, 'learning_rate/qwen_vl_interface': 1.515506306639057e-06, 'learning_rate/action_model': 1.515506306639057e-05, 'epoch': 1.7}) +[2026-06-16 05:45:33,082][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.055655576288700104, 'train/grad_norm_pre_clip': 8.653339385986328, 'timing/data': 0.006408409972209483, 'timing/model': 1.6873079890501685, 'learning_rate/qwen_vl_interface': 1.5136251243665589e-06, 'learning_rate/action_model': 1.5136251243665589e-05, 'epoch': 1.7}) +[2026-06-16 05:45:38,755][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.04573897644877434, 'train/grad_norm_pre_clip': 5.344249248504639, 'timing/data': 0.006370434013660997, 'timing/model': 1.4601728400448337, 'learning_rate/qwen_vl_interface': 1.5117454779766372e-06, 'learning_rate/action_model': 1.5117454779766371e-05, 'epoch': 1.71}) +[2026-06-16 05:45:44,654][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.01070620957762003, 'train/grad_norm_pre_clip': 1.0407650470733643, 'timing/data': 0.006226981990039349, 'timing/model': 1.4613005190039985, 'learning_rate/qwen_vl_interface': 1.5098673682419428e-06, 'learning_rate/action_model': 1.5098673682419429e-05, 'epoch': 1.71}) +[2026-06-16 05:45:50,554][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.09380010515451431, 'train/grad_norm_pre_clip': 7.080433368682861, 'timing/data': 0.005957727029453963, 'timing/model': 1.4602917060256004, 'learning_rate/qwen_vl_interface': 1.5079907959344977e-06, 'learning_rate/action_model': 1.5079907959344978e-05, 'epoch': 1.71}) +[2026-06-16 05:45:56,460][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.029921192675828934, 'train/grad_norm_pre_clip': 6.464155197143555, 'timing/data': 0.006259720947127789, 'timing/model': 1.4328983279992826, 'learning_rate/qwen_vl_interface': 1.5061157618256873e-06, 'learning_rate/action_model': 1.5061157618256871e-05, 'epoch': 1.71}) +[2026-06-16 05:46:02,386][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.17262238264083862, 'train/grad_norm_pre_clip': 8.734504699707031, 'timing/data': 0.006090701965149492, 'timing/model': 1.6900038570165634, 'learning_rate/qwen_vl_interface': 1.5042422666862675e-06, 'learning_rate/action_model': 1.5042422666862673e-05, 'epoch': 1.71}) +[2026-06-16 05:46:08,054][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.046264830976724625, 'train/grad_norm_pre_clip': 15.548672676086426, 'timing/data': 0.006368761009071022, 'timing/model': 1.4589260370121337, 'learning_rate/qwen_vl_interface': 1.5023703112863628e-06, 'learning_rate/action_model': 1.5023703112863626e-05, 'epoch': 1.71}) +[2026-06-16 05:46:13,953][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.06195902079343796, 'train/grad_norm_pre_clip': 7.634067058563232, 'timing/data': 0.00616151699796319, 'timing/model': 1.4592431929777376, 'learning_rate/qwen_vl_interface': 1.5004998963954629e-06, 'learning_rate/action_model': 1.5004998963954628e-05, 'epoch': 1.71}) +[2026-06-16 05:46:19,853][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.035984110087156296, 'train/grad_norm_pre_clip': 5.5377421379089355, 'timing/data': 0.006208945997059345, 'timing/model': 1.4575541180092841, 'learning_rate/qwen_vl_interface': 1.498631022782424e-06, 'learning_rate/action_model': 1.4986310227824241e-05, 'epoch': 1.71}) +[2026-06-16 05:46:25,758][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.0920911654829979, 'train/grad_norm_pre_clip': 8.072880744934082, 'timing/data': 0.0067721750237978995, 'timing/model': 1.4331086070160381, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 1.71}) +[2026-06-16 05:46:31,652][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.11264313012361526, 'train/grad_norm_pre_clip': 7.901533603668213, 'timing/data': 0.006559332017786801, 'timing/model': 1.5071849980158731, 'learning_rate/qwen_vl_interface': 1.4948979024621915e-06, 'learning_rate/action_model': 1.4948979024621914e-05, 'epoch': 1.71}) +[2026-06-16 05:46:37,545][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.03237764537334442, 'train/grad_norm_pre_clip': 3.2499358654022217, 'timing/data': 0.005808205984067172, 'timing/model': 1.655488770978991, 'learning_rate/qwen_vl_interface': 1.493033657289542e-06, 'learning_rate/action_model': 1.493033657289542e-05, 'epoch': 1.71}) +[2026-06-16 05:46:43,270][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.12761901319026947, 'train/grad_norm_pre_clip': 6.729938983917236, 'timing/data': 0.006239501002710313, 'timing/model': 1.478947227995377, 'learning_rate/qwen_vl_interface': 1.4911709564638428e-06, 'learning_rate/action_model': 1.4911709564638427e-05, 'epoch': 1.71}) +[2026-06-16 05:46:49,248][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.01846054196357727, 'train/grad_norm_pre_clip': 2.0673129558563232, 'timing/data': 0.006794407963752747, 'timing/model': 1.4571000410360284, 'learning_rate/qwen_vl_interface': 1.4893098007507817e-06, 'learning_rate/action_model': 1.4893098007507816e-05, 'epoch': 1.71}) +[2026-06-16 05:46:55,153][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.0052099875174462795, 'train/grad_norm_pre_clip': 1.1799324750900269, 'timing/data': 0.006658046972006559, 'timing/model': 1.4701427829568274, 'learning_rate/qwen_vl_interface': 1.487450190915406e-06, 'learning_rate/action_model': 1.487450190915406e-05, 'epoch': 1.71}) +[2026-06-16 05:47:01,069][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.07990024983882904, 'train/grad_norm_pre_clip': 7.890178203582764, 'timing/data': 0.0071241630357690156, 'timing/model': 1.4449443250196055, 'learning_rate/qwen_vl_interface': 1.4855921277221336e-06, 'learning_rate/action_model': 1.4855921277221336e-05, 'epoch': 1.71}) +[2026-06-16 05:47:06,972][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.04808219149708748, 'train/grad_norm_pre_clip': 8.881957054138184, 'timing/data': 0.006501206022221595, 'timing/model': 1.4445664990344085, 'learning_rate/qwen_vl_interface': 1.4837356119347423e-06, 'learning_rate/action_model': 1.4837356119347423e-05, 'epoch': 1.71}) +[2026-06-16 05:47:12,872][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.012293288484215736, 'train/grad_norm_pre_clip': 2.1675782203674316, 'timing/data': 0.006977307028137147, 'timing/model': 1.683738199993968, 'learning_rate/qwen_vl_interface': 1.4818806443163774e-06, 'learning_rate/action_model': 1.4818806443163775e-05, 'epoch': 1.71}) +[2026-06-16 05:47:18,557][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.0685194879770279, 'train/grad_norm_pre_clip': 7.1093573570251465, 'timing/data': 0.007273312017787248, 'timing/model': 1.464602270978503, 'learning_rate/qwen_vl_interface': 1.480027225629547e-06, 'learning_rate/action_model': 1.480027225629547e-05, 'epoch': 1.71}) +[2026-06-16 05:47:24,462][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.07829063385725021, 'train/grad_norm_pre_clip': 8.551519393920898, 'timing/data': 0.006714179995469749, 'timing/model': 1.4680404600221664, 'learning_rate/qwen_vl_interface': 1.4781753566361182e-06, 'learning_rate/action_model': 1.4781753566361182e-05, 'epoch': 1.71}) +[2026-06-16 05:47:30,352][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.22228914499282837, 'train/grad_norm_pre_clip': 12.460476875305176, 'timing/data': 0.006527565012220293, 'timing/model': 1.4567273639841005, 'learning_rate/qwen_vl_interface': 1.4763250380973255e-06, 'learning_rate/action_model': 1.4763250380973254e-05, 'epoch': 1.71}) +[2026-06-16 05:47:36,323][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.10432113707065582, 'train/grad_norm_pre_clip': 9.693389892578125, 'timing/data': 0.006830630009062588, 'timing/model': 1.4937638909905218, 'learning_rate/qwen_vl_interface': 1.4744762707737662e-06, 'learning_rate/action_model': 1.4744762707737662e-05, 'epoch': 1.71}) +[2026-06-16 05:47:42,333][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.04063277691602707, 'train/grad_norm_pre_clip': 5.018294811248779, 'timing/data': 0.006677736004348844, 'timing/model': 1.7415865930379368, 'learning_rate/qwen_vl_interface': 1.4726290554253977e-06, 'learning_rate/action_model': 1.4726290554253977e-05, 'epoch': 1.71}) +[2026-06-16 05:47:48,052][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.16644485294818878, 'train/grad_norm_pre_clip': 14.057558059692383, 'timing/data': 0.0061108809895813465, 'timing/model': 1.4482169409748167, 'learning_rate/qwen_vl_interface': 1.4707833928115406e-06, 'learning_rate/action_model': 1.4707833928115405e-05, 'epoch': 1.71}) +[2026-06-16 05:47:54,050][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.11997583508491516, 'train/grad_norm_pre_clip': 20.680538177490234, 'timing/data': 0.00660466798581183, 'timing/model': 1.459661872999277, 'learning_rate/qwen_vl_interface': 1.4689392836908775e-06, 'learning_rate/action_model': 1.4689392836908774e-05, 'epoch': 1.72}) +[2026-06-16 05:47:59,850][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.026312008500099182, 'train/grad_norm_pre_clip': 5.160192489624023, 'timing/data': 0.006583987968042493, 'timing/model': 1.4577452930388972, 'learning_rate/qwen_vl_interface': 1.4670967288214514e-06, 'learning_rate/action_model': 1.4670967288214513e-05, 'epoch': 1.72}) +[2026-06-16 05:48:05,750][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.03076533041894436, 'train/grad_norm_pre_clip': 2.690701961517334, 'timing/data': 0.0065505689708516, 'timing/model': 1.4609825360239483, 'learning_rate/qwen_vl_interface': 1.4652557289606687e-06, 'learning_rate/action_model': 1.4652557289606685e-05, 'epoch': 1.72}) +[2026-06-16 05:48:11,646][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.15321294963359833, 'train/grad_norm_pre_clip': 15.042816162109375, 'timing/data': 0.007185383001342416, 'timing/model': 1.47506704099942, 'learning_rate/qwen_vl_interface': 1.4634162848652918e-06, 'learning_rate/action_model': 1.4634162848652919e-05, 'epoch': 1.72}) +[2026-06-16 05:48:17,598][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.12378309667110443, 'train/grad_norm_pre_clip': 16.81252098083496, 'timing/data': 0.0066598900011740625, 'timing/model': 1.708486016956158, 'learning_rate/qwen_vl_interface': 1.4615783972914494e-06, 'learning_rate/action_model': 1.4615783972914493e-05, 'epoch': 1.72}) +[2026-06-16 05:48:23,263][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.035049330443143845, 'train/grad_norm_pre_clip': 4.177243232727051, 'timing/data': 0.006547634955495596, 'timing/model': 1.4697505880030803, 'learning_rate/qwen_vl_interface': 1.459742066994626e-06, 'learning_rate/action_model': 1.4597420669946261e-05, 'epoch': 1.72}) +[2026-06-16 05:48:29,253][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.19036109745502472, 'train/grad_norm_pre_clip': 41.12210464477539, 'timing/data': 0.006510740960948169, 'timing/model': 1.4584388139774092, 'learning_rate/qwen_vl_interface': 1.4579072947296687e-06, 'learning_rate/action_model': 1.4579072947296687e-05, 'epoch': 1.72}) +[2026-06-16 05:48:35,152][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.11483719944953918, 'train/grad_norm_pre_clip': 9.178624153137207, 'timing/data': 0.006482678989414126, 'timing/model': 1.4586032149964012, 'learning_rate/qwen_vl_interface': 1.4560740812507823e-06, 'learning_rate/action_model': 1.4560740812507822e-05, 'epoch': 1.72}) +[2026-06-16 05:48:41,053][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.01973317377269268, 'train/grad_norm_pre_clip': 3.083343029022217, 'timing/data': 0.00660430797142908, 'timing/model': 1.4623072859831154, 'learning_rate/qwen_vl_interface': 1.4542424273115322e-06, 'learning_rate/action_model': 1.454242427311532e-05, 'epoch': 1.72}) +[2026-06-16 05:48:46,955][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.10368171334266663, 'train/grad_norm_pre_clip': 8.647589683532715, 'timing/data': 0.006978338002227247, 'timing/model': 1.4651331540080719, 'learning_rate/qwen_vl_interface': 1.4524123336648419e-06, 'learning_rate/action_model': 1.4524123336648418e-05, 'epoch': 1.72}) +[2026-06-16 05:48:52,894][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.022654466331005096, 'train/grad_norm_pre_clip': 4.045026779174805, 'timing/data': 0.006696844007819891, 'timing/model': 1.686845856020227, 'learning_rate/qwen_vl_interface': 1.4505838010629942e-06, 'learning_rate/action_model': 1.4505838010629941e-05, 'epoch': 1.72}) +[2026-06-16 05:48:58,562][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.18781104683876038, 'train/grad_norm_pre_clip': 16.438249588012695, 'timing/data': 0.006644496985245496, 'timing/model': 1.4671391059528105, 'learning_rate/qwen_vl_interface': 1.4487568302576284e-06, 'learning_rate/action_model': 1.4487568302576282e-05, 'epoch': 1.72}) +[2026-06-16 05:49:04,454][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.26419299840927124, 'train/grad_norm_pre_clip': 15.93242073059082, 'timing/data': 0.0069978569517843425, 'timing/model': 1.459343775990419, 'learning_rate/qwen_vl_interface': 1.446931421999746e-06, 'learning_rate/action_model': 1.4469314219997459e-05, 'epoch': 1.72}) +[2026-06-16 05:49:10,450][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.25889840722084045, 'train/grad_norm_pre_clip': 17.43899917602539, 'timing/data': 0.006411925016436726, 'timing/model': 1.476786742976401, 'learning_rate/qwen_vl_interface': 1.4451075770397004e-06, 'learning_rate/action_model': 1.4451075770397004e-05, 'epoch': 1.72}) +[2026-06-16 05:58:03,652][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.18572735786437988, 'train/grad_norm_pre_clip': 10.261578559875488, 'eval/loss': 0.0491250678896904, 'eval/flap_precision': 0.9608247422680413, 'eval/flap_recall': 0.9338677354709419, 'eval/flap_f1': 0.9471544715447155, 'eval/flap_rate_model': 0.13472222222222222, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9471544715447157, 'timing/data': 0.006300630979239941, 'timing/model': 1.4301728520076722, 'learning_rate/qwen_vl_interface': 1.4432852961272068e-06, 'learning_rate/action_model': 1.4432852961272066e-05, 'epoch': 1.72}) +[2026-06-16 05:58:03,653][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4000_state +[2026-06-16 05:58:03,657][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 05:58:09,160][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4000_state/model.safetensors +[2026-06-16 05:58:17,370][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4000_state/optimizer.bin +[2026-06-16 05:58:17,371][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4000_state/sampler.bin +[2026-06-16 05:58:17,371][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-16 05:58:17,372][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-16 05:58:17,486][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-16 05:58:17,486][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 05:58:17,496][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 05:58:23,339][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.037631209939718246, 'train/grad_norm_pre_clip': 4.581520080566406, 'timing/data': 0.006946817971765995, 'timing/model': 1.4593456269940361, 'learning_rate/qwen_vl_interface': 1.4414645800113353e-06, 'learning_rate/action_model': 1.4414645800113352e-05, 'epoch': 1.72}) +[2026-06-16 05:58:29,237][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.21126249432563782, 'train/grad_norm_pre_clip': 10.200246810913086, 'timing/data': 0.006386553985066712, 'timing/model': 1.4255642429925501, 'learning_rate/qwen_vl_interface': 1.4396454294405138e-06, 'learning_rate/action_model': 1.4396454294405136e-05, 'epoch': 1.72}) +[2026-06-16 05:58:35,179][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 0.09129812568426132, 'train/grad_norm_pre_clip': 9.674236297607422, 'timing/data': 0.006685133033897728, 'timing/model': 1.698248266999144, 'learning_rate/qwen_vl_interface': 1.4378278451625271e-06, 'learning_rate/action_model': 1.4378278451625271e-05, 'epoch': 1.72}) +[2026-06-16 05:58:40,843][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.09868308156728745, 'train/grad_norm_pre_clip': 11.506811141967773, 'timing/data': 0.006703330029267818, 'timing/model': 1.4597950030001812, 'learning_rate/qwen_vl_interface': 1.4360118279245143e-06, 'learning_rate/action_model': 1.4360118279245144e-05, 'epoch': 1.72}) +[2026-06-16 05:58:46,846][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.0681622326374054, 'train/grad_norm_pre_clip': 3.0043046474456787, 'timing/data': 0.0064236290054395795, 'timing/model': 1.460821607033722, 'learning_rate/qwen_vl_interface': 1.434197378472973e-06, 'learning_rate/action_model': 1.4341973784729729e-05, 'epoch': 1.72}) +[2026-06-16 05:58:52,757][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.16922242939472198, 'train/grad_norm_pre_clip': 13.519757270812988, 'timing/data': 0.007162234978750348, 'timing/model': 1.4631086729932576, 'learning_rate/qwen_vl_interface': 1.4323844975537535e-06, 'learning_rate/action_model': 1.4323844975537534e-05, 'epoch': 1.72}) +[2026-06-16 05:58:58,644][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.02924213372170925, 'train/grad_norm_pre_clip': 3.504405975341797, 'timing/data': 0.006733966001775116, 'timing/model': 1.4575744229950942, 'learning_rate/qwen_vl_interface': 1.4305731859120628e-06, 'learning_rate/action_model': 1.4305731859120628e-05, 'epoch': 1.72}) +[2026-06-16 05:59:04,548][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.24161984026432037, 'train/grad_norm_pre_clip': 14.177995681762695, 'timing/data': 0.0066619099816307425, 'timing/model': 1.4270356000051834, 'learning_rate/qwen_vl_interface': 1.4287634442924647e-06, 'learning_rate/action_model': 1.4287634442924646e-05, 'epoch': 1.72}) +[2026-06-16 05:59:10,481][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.06762735545635223, 'train/grad_norm_pre_clip': 7.905840873718262, 'timing/data': 0.007511287985835224, 'timing/model': 1.6903042030171491, 'learning_rate/qwen_vl_interface': 1.4269552734388725e-06, 'learning_rate/action_model': 1.4269552734388724e-05, 'epoch': 1.72}) +[2026-06-16 05:59:16,162][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.06343068182468414, 'train/grad_norm_pre_clip': 11.373077392578125, 'timing/data': 0.006652235984802246, 'timing/model': 1.4695704629993998, 'learning_rate/qwen_vl_interface': 1.4251486740945584e-06, 'learning_rate/action_model': 1.4251486740945583e-05, 'epoch': 1.73}) +[2026-06-16 05:59:22,048][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.0476415790617466, 'train/grad_norm_pre_clip': 7.720709323883057, 'timing/data': 0.0073021610151045024, 'timing/model': 1.459859925962519, 'learning_rate/qwen_vl_interface': 1.423343647002147e-06, 'learning_rate/action_model': 1.423343647002147e-05, 'epoch': 1.73}) +[2026-06-16 05:59:27,950][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.017752505838871002, 'train/grad_norm_pre_clip': 3.165026903152466, 'timing/data': 0.006659546983428299, 'timing/model': 1.4608291229815222, 'learning_rate/qwen_vl_interface': 1.421540192903617e-06, 'learning_rate/action_model': 1.421540192903617e-05, 'epoch': 1.73}) +[2026-06-16 05:59:33,859][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 0.04633916914463043, 'train/grad_norm_pre_clip': 2.9670989513397217, 'timing/data': 0.006709931010846049, 'timing/model': 1.4686058340012096, 'learning_rate/qwen_vl_interface': 1.4197383125402997e-06, 'learning_rate/action_model': 1.4197383125402997e-05, 'epoch': 1.73}) +[2026-06-16 05:59:39,826][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.20882244408130646, 'train/grad_norm_pre_clip': 9.623037338256836, 'timing/data': 0.0064862919971346855, 'timing/model': 1.4927805460174568, 'learning_rate/qwen_vl_interface': 1.4179380066528807e-06, 'learning_rate/action_model': 1.4179380066528807e-05, 'epoch': 1.73}) +[2026-06-16 05:59:45,782][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.12571972608566284, 'train/grad_norm_pre_clip': 16.11452293395996, 'timing/data': 0.006247962010093033, 'timing/model': 1.688053031975869, 'learning_rate/qwen_vl_interface': 1.4161392759813984e-06, 'learning_rate/action_model': 1.4161392759813983e-05, 'epoch': 1.73}) +[2026-06-16 05:59:51,612][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.11662068963050842, 'train/grad_norm_pre_clip': 6.760902404785156, 'timing/data': 0.007263995998073369, 'timing/model': 1.5183736850158311, 'learning_rate/qwen_vl_interface': 1.4143421212652411e-06, 'learning_rate/action_model': 1.414342121265241e-05, 'epoch': 1.73}) +[2026-06-16 05:59:57,462][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.15772810578346252, 'train/grad_norm_pre_clip': 11.584773063659668, 'timing/data': 0.007051132968626916, 'timing/model': 1.4686759369797073, 'learning_rate/qwen_vl_interface': 1.4125465432431529e-06, 'learning_rate/action_model': 1.4125465432431529e-05, 'epoch': 1.73}) +[2026-06-16 06:00:03,362][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.002817663364112377, 'train/grad_norm_pre_clip': 0.3833102881908417, 'timing/data': 0.006550837017130107, 'timing/model': 1.4673314120154828, 'learning_rate/qwen_vl_interface': 1.4107525426532273e-06, 'learning_rate/action_model': 1.4107525426532272e-05, 'epoch': 1.73}) +[2026-06-16 06:00:09,260][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.09055537730455399, 'train/grad_norm_pre_clip': 9.04932975769043, 'timing/data': 0.006812251987867057, 'timing/model': 1.4650898689869791, 'learning_rate/qwen_vl_interface': 1.4089601202329116e-06, 'learning_rate/action_model': 1.4089601202329114e-05, 'epoch': 1.73}) +[2026-06-16 06:00:15,157][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.02708897367119789, 'train/grad_norm_pre_clip': 7.063029766082764, 'timing/data': 0.006682751991320401, 'timing/model': 1.4329743990092538, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 1.73}) +[2026-06-16 06:00:21,078][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.010584775358438492, 'train/grad_norm_pre_clip': 1.86423921585083, 'timing/data': 0.006672005984000862, 'timing/model': 1.6848285400192253, 'learning_rate/qwen_vl_interface': 1.4053800128476426e-06, 'learning_rate/action_model': 1.4053800128476427e-05, 'epoch': 1.73}) +[2026-06-16 06:00:26,908][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.1304994821548462, 'train/grad_norm_pre_clip': 14.683609008789062, 'timing/data': 0.006388790032360703, 'timing/model': 1.61341505398741, 'learning_rate/qwen_vl_interface': 1.403592329354338e-06, 'learning_rate/action_model': 1.403592329354338e-05, 'epoch': 1.73}) +[2026-06-16 06:00:32,698][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.2274855077266693, 'train/grad_norm_pre_clip': 15.30355167388916, 'timing/data': 0.00699568196432665, 'timing/model': 1.4489057540195063, 'learning_rate/qwen_vl_interface': 1.4018062269739354e-06, 'learning_rate/action_model': 1.4018062269739355e-05, 'epoch': 1.73}) +[2026-06-16 06:00:38,667][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.03409555181860924, 'train/grad_norm_pre_clip': 5.304696083068848, 'timing/data': 0.006749680964276195, 'timing/model': 1.4689684819895774, 'learning_rate/qwen_vl_interface': 1.4000217064406339e-06, 'learning_rate/action_model': 1.4000217064406337e-05, 'epoch': 1.73}) +[2026-06-16 06:00:44,558][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.08360453695058823, 'train/grad_norm_pre_clip': 8.664925575256348, 'timing/data': 0.006696141965221614, 'timing/model': 1.4573548219632357, 'learning_rate/qwen_vl_interface': 1.3982387684879836e-06, 'learning_rate/action_model': 1.3982387684879836e-05, 'epoch': 1.73}) +[2026-06-16 06:00:50,562][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.12635888159275055, 'train/grad_norm_pre_clip': 8.806159973144531, 'timing/data': 0.006699767021927983, 'timing/model': 1.4463796409545466, 'learning_rate/qwen_vl_interface': 1.3964574138488804e-06, 'learning_rate/action_model': 1.3964574138488803e-05, 'epoch': 1.73}) +[2026-06-16 06:00:56,535][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.126888245344162, 'train/grad_norm_pre_clip': 7.256666660308838, 'timing/data': 0.006967480992898345, 'timing/model': 1.7365185930393636, 'learning_rate/qwen_vl_interface': 1.3946776432555743e-06, 'learning_rate/action_model': 1.3946776432555741e-05, 'epoch': 1.73}) +[2026-06-16 06:01:02,602][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.05402250215411186, 'train/grad_norm_pre_clip': 5.427219390869141, 'timing/data': 0.006748248997610062, 'timing/model': 1.7108169290004298, 'learning_rate/qwen_vl_interface': 1.3928994574396595e-06, 'learning_rate/action_model': 1.3928994574396594e-05, 'epoch': 1.73}) +[2026-06-16 06:01:08,448][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.014664789661765099, 'train/grad_norm_pre_clip': 1.5115123987197876, 'timing/data': 0.007070042018312961, 'timing/model': 1.565718038007617, 'learning_rate/qwen_vl_interface': 1.391122857132083e-06, 'learning_rate/action_model': 1.3911228571320831e-05, 'epoch': 1.73}) +[2026-06-16 06:01:14,372][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.04947538673877716, 'train/grad_norm_pre_clip': 8.75576114654541, 'timing/data': 0.008205751015339047, 'timing/model': 1.474243321979884, 'learning_rate/qwen_vl_interface': 1.3893478430631376e-06, 'learning_rate/action_model': 1.3893478430631376e-05, 'epoch': 1.73}) +[2026-06-16 06:01:20,251][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.2159423679113388, 'train/grad_norm_pre_clip': 13.228999137878418, 'timing/data': 0.006569786986801773, 'timing/model': 1.4516934959683567, 'learning_rate/qwen_vl_interface': 1.3875744159624652e-06, 'learning_rate/action_model': 1.3875744159624652e-05, 'epoch': 1.73}) +[2026-06-16 06:01:26,157][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.0078340545296669, 'train/grad_norm_pre_clip': 0.7427345514297485, 'timing/data': 0.007066858001053333, 'timing/model': 1.4309205600293353, 'learning_rate/qwen_vl_interface': 1.3858025765590525e-06, 'learning_rate/action_model': 1.3858025765590525e-05, 'epoch': 1.73}) +[2026-06-16 06:01:32,085][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.19713065028190613, 'train/grad_norm_pre_clip': 15.184911727905273, 'timing/data': 0.006540253991261125, 'timing/model': 1.6922174880164675, 'learning_rate/qwen_vl_interface': 1.384032325581237e-06, 'learning_rate/action_model': 1.3840323255812371e-05, 'epoch': 1.74}) +[2026-06-16 06:01:37,946][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.04589654877781868, 'train/grad_norm_pre_clip': 5.175567150115967, 'timing/data': 0.006494396016933024, 'timing/model': 1.6561787250102498, 'learning_rate/qwen_vl_interface': 1.3822636637567025e-06, 'learning_rate/action_model': 1.3822636637567024e-05, 'epoch': 1.74}) +[2026-06-16 06:01:43,653][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.0011673381086438894, 'train/grad_norm_pre_clip': 0.2018725723028183, 'timing/data': 0.006885331007651985, 'timing/model': 1.45884966396261, 'learning_rate/qwen_vl_interface': 1.3804965918124788e-06, 'learning_rate/action_model': 1.3804965918124787e-05, 'epoch': 1.74}) +[2026-06-16 06:01:49,565][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.17622989416122437, 'train/grad_norm_pre_clip': 12.504279136657715, 'timing/data': 0.007487435999792069, 'timing/model': 1.4673820549505763, 'learning_rate/qwen_vl_interface': 1.3787311104749418e-06, 'learning_rate/action_model': 1.3787311104749418e-05, 'epoch': 1.74}) +[2026-06-16 06:01:55,449][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.07576118409633636, 'train/grad_norm_pre_clip': 6.368229866027832, 'timing/data': 0.006789339997339994, 'timing/model': 1.4566515979822725, 'learning_rate/qwen_vl_interface': 1.3769672204698149e-06, 'learning_rate/action_model': 1.3769672204698147e-05, 'epoch': 1.74}) +[2026-06-16 06:02:01,351][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.053749363869428635, 'train/grad_norm_pre_clip': 8.100198745727539, 'timing/data': 0.007358045957516879, 'timing/model': 1.4321015870082192, 'learning_rate/qwen_vl_interface': 1.3752049225221664e-06, 'learning_rate/action_model': 1.3752049225221663e-05, 'epoch': 1.74}) +[2026-06-16 06:02:07,257][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.08540251106023788, 'train/grad_norm_pre_clip': 11.99167251586914, 'timing/data': 0.006720448960550129, 'timing/model': 1.433686863980256, 'learning_rate/qwen_vl_interface': 1.3734442173564103e-06, 'learning_rate/action_model': 1.3734442173564102e-05, 'epoch': 1.74}) +[2026-06-16 06:02:13,142][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.0929197296500206, 'train/grad_norm_pre_clip': 9.366777420043945, 'timing/data': 0.006552030972670764, 'timing/model': 1.646429770044051, 'learning_rate/qwen_vl_interface': 1.3716851056963062e-06, 'learning_rate/action_model': 1.3716851056963062e-05, 'epoch': 1.74}) +[2026-06-16 06:02:18,854][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.15089203417301178, 'train/grad_norm_pre_clip': 15.01203727722168, 'timing/data': 0.006694400974083692, 'timing/model': 1.458171408972703, 'learning_rate/qwen_vl_interface': 1.3699275882649597e-06, 'learning_rate/action_model': 1.3699275882649596e-05, 'epoch': 1.74}) +[2026-06-16 06:02:24,755][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.16451212763786316, 'train/grad_norm_pre_clip': 12.423043251037598, 'timing/data': 0.007351906970143318, 'timing/model': 1.4581808680086397, 'learning_rate/qwen_vl_interface': 1.368171665784817e-06, 'learning_rate/action_model': 1.3681716657848171e-05, 'epoch': 1.74}) +[2026-06-16 06:02:30,654][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.07538165152072906, 'train/grad_norm_pre_clip': 7.777403354644775, 'timing/data': 0.00679063203278929, 'timing/model': 1.4576220230082981, 'learning_rate/qwen_vl_interface': 1.3664173389776735e-06, 'learning_rate/action_model': 1.3664173389776735e-05, 'epoch': 1.74}) +[2026-06-16 06:02:36,670][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.10340661555528641, 'train/grad_norm_pre_clip': 12.308338165283203, 'timing/data': 0.006781759962905198, 'timing/model': 1.5299768329714425, 'learning_rate/qwen_vl_interface': 1.3646646085646658e-06, 'learning_rate/action_model': 1.3646646085646659e-05, 'epoch': 1.74}) +[2026-06-16 06:02:42,683][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.043424494564533234, 'train/grad_norm_pre_clip': 6.07125997543335, 'timing/data': 0.006697725970298052, 'timing/model': 1.6907534579513595, 'learning_rate/qwen_vl_interface': 1.3629134752662756e-06, 'learning_rate/action_model': 1.3629134752662756e-05, 'epoch': 1.74}) +[2026-06-16 06:02:48,356][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 0.04230250045657158, 'train/grad_norm_pre_clip': 6.452188491821289, 'timing/data': 0.006900424021296203, 'timing/model': 1.459828082995955, 'learning_rate/qwen_vl_interface': 1.3611639398023282e-06, 'learning_rate/action_model': 1.3611639398023283e-05, 'epoch': 1.74}) +[2026-06-16 06:02:54,257][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.08365505188703537, 'train/grad_norm_pre_clip': 12.978015899658203, 'timing/data': 0.007654001994524151, 'timing/model': 1.4568619570345618, 'learning_rate/qwen_vl_interface': 1.35941600289199e-06, 'learning_rate/action_model': 1.35941600289199e-05, 'epoch': 1.74}) +[2026-06-16 06:03:00,056][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.006423827726393938, 'train/grad_norm_pre_clip': 1.1344220638275146, 'timing/data': 0.006686640030238777, 'timing/model': 1.4589593459968455, 'learning_rate/qwen_vl_interface': 1.3576696652537735e-06, 'learning_rate/action_model': 1.3576696652537736e-05, 'epoch': 1.74}) +[2026-06-16 06:03:05,957][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.08425689488649368, 'train/grad_norm_pre_clip': 9.035394668579102, 'timing/data': 0.006980602047406137, 'timing/model': 1.4749519409961067, 'learning_rate/qwen_vl_interface': 1.355924927605531e-06, 'learning_rate/action_model': 1.3559249276055311e-05, 'epoch': 1.74}) +[2026-06-16 06:03:11,856][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.010635470040142536, 'train/grad_norm_pre_clip': 2.14285945892334, 'timing/data': 0.007174947997555137, 'timing/model': 1.4683557790121995, 'learning_rate/qwen_vl_interface': 1.3541817906644591e-06, 'learning_rate/action_model': 1.3541817906644591e-05, 'epoch': 1.74}) +[2026-06-16 06:03:17,788][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.19214428961277008, 'train/grad_norm_pre_clip': 14.983333587646484, 'timing/data': 0.006735011003911495, 'timing/model': 1.6882693200022914, 'learning_rate/qwen_vl_interface': 1.3524402551470955e-06, 'learning_rate/action_model': 1.3524402551470956e-05, 'epoch': 1.74}) +[2026-06-16 06:03:23,457][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.020969891920685768, 'train/grad_norm_pre_clip': 2.385956048965454, 'timing/data': 0.006678287987597287, 'timing/model': 1.4569416079903021, 'learning_rate/qwen_vl_interface': 1.3507003217693196e-06, 'learning_rate/action_model': 1.3507003217693196e-05, 'epoch': 1.74}) +[2026-06-16 06:03:29,354][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.024302121251821518, 'train/grad_norm_pre_clip': 5.654204368591309, 'timing/data': 0.006741930963471532, 'timing/model': 1.4556585969985463, 'learning_rate/qwen_vl_interface': 1.3489619912463515e-06, 'learning_rate/action_model': 1.3489619912463514e-05, 'epoch': 1.74}) +[2026-06-16 06:03:35,255][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.14471285045146942, 'train/grad_norm_pre_clip': 11.761655807495117, 'timing/data': 0.006918630970176309, 'timing/model': 1.4583172589773312, 'learning_rate/qwen_vl_interface': 1.3472252642927534e-06, 'learning_rate/action_model': 1.3472252642927533e-05, 'epoch': 1.74}) +[2026-06-16 06:03:41,257][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 0.22439230978488922, 'train/grad_norm_pre_clip': 19.384441375732422, 'timing/data': 0.006659039994701743, 'timing/model': 1.4626920640002936, 'learning_rate/qwen_vl_interface': 1.3454901416224284e-06, 'learning_rate/action_model': 1.3454901416224284e-05, 'epoch': 1.74}) +[2026-06-16 06:03:47,157][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 0.018172545358538628, 'train/grad_norm_pre_clip': 3.1243627071380615, 'timing/data': 0.007260754995513707, 'timing/model': 1.4664804179919884, 'learning_rate/qwen_vl_interface': 1.34375662394862e-06, 'learning_rate/action_model': 1.34375662394862e-05, 'epoch': 1.75}) +[2026-06-16 06:03:53,087][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.053382739424705505, 'train/grad_norm_pre_clip': 8.251256942749023, 'timing/data': 0.0066855280310846865, 'timing/model': 1.6882049120031297, 'learning_rate/qwen_vl_interface': 1.3420247119839116e-06, 'learning_rate/action_model': 1.3420247119839117e-05, 'epoch': 1.75}) +[2026-06-16 06:03:58,856][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.1592840999364853, 'train/grad_norm_pre_clip': 15.887176513671875, 'timing/data': 0.006790822953917086, 'timing/model': 1.4443869440001436, 'learning_rate/qwen_vl_interface': 1.340294406440227e-06, 'learning_rate/action_model': 1.340294406440227e-05, 'epoch': 1.75}) +[2026-06-16 06:04:04,775][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.3687606453895569, 'train/grad_norm_pre_clip': 14.093107223510742, 'timing/data': 0.006563280010595918, 'timing/model': 1.4673640170367435, 'learning_rate/qwen_vl_interface': 1.3385657080288297e-06, 'learning_rate/action_model': 1.3385657080288297e-05, 'epoch': 1.75}) +[2026-06-16 06:04:10,748][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.01475119311362505, 'train/grad_norm_pre_clip': 2.260408401489258, 'timing/data': 0.006542538001667708, 'timing/model': 1.5168324489495717, 'learning_rate/qwen_vl_interface': 1.336838617460322e-06, 'learning_rate/action_model': 1.336838617460322e-05, 'epoch': 1.75}) +[2026-06-16 06:04:16,659][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.07496048510074615, 'train/grad_norm_pre_clip': 10.070168495178223, 'timing/data': 0.0066452500177547336, 'timing/model': 1.4353272640146315, 'learning_rate/qwen_vl_interface': 1.335113135444646e-06, 'learning_rate/action_model': 1.335113135444646e-05, 'epoch': 1.75}) +[2026-06-16 06:04:22,555][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 0.022495336830615997, 'train/grad_norm_pre_clip': 4.082225799560547, 'timing/data': 0.006580584973562509, 'timing/model': 1.459213805035688, 'learning_rate/qwen_vl_interface': 1.3333892626910823e-06, 'learning_rate/action_model': 1.3333892626910823e-05, 'epoch': 1.75}) +[2026-06-16 06:04:28,491][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.03639034554362297, 'train/grad_norm_pre_clip': 6.525270462036133, 'timing/data': 0.006784604978747666, 'timing/model': 1.6935341189964674, 'learning_rate/qwen_vl_interface': 1.3316669999082486e-06, 'learning_rate/action_model': 1.3316669999082484e-05, 'epoch': 1.75}) +[2026-06-16 06:04:34,238][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.2802526652812958, 'train/grad_norm_pre_clip': 14.185418128967285, 'timing/data': 0.006831574020907283, 'timing/model': 1.539634105982259, 'learning_rate/qwen_vl_interface': 1.3299463478041024e-06, 'learning_rate/action_model': 1.3299463478041024e-05, 'epoch': 1.75}) +[2026-06-16 06:04:40,153][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.024261560291051865, 'train/grad_norm_pre_clip': 2.1045351028442383, 'timing/data': 0.007238582998979837, 'timing/model': 1.45818146498641, 'learning_rate/qwen_vl_interface': 1.3282273070859396e-06, 'learning_rate/action_model': 1.3282273070859396e-05, 'epoch': 1.75}) +[2026-06-16 06:04:46,055][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 0.0058207823894917965, 'train/grad_norm_pre_clip': 1.8529515266418457, 'timing/data': 0.006526675948407501, 'timing/model': 1.4606276829726994, 'learning_rate/qwen_vl_interface': 1.3265098784603924e-06, 'learning_rate/action_model': 1.3265098784603924e-05, 'epoch': 1.75}) +[2026-06-16 06:04:51,957][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.13669168949127197, 'train/grad_norm_pre_clip': 14.263806343078613, 'timing/data': 0.006795872002840042, 'timing/model': 1.4618834120337851, 'learning_rate/qwen_vl_interface': 1.324794062633431e-06, 'learning_rate/action_model': 1.324794062633431e-05, 'epoch': 1.75}) +[2026-06-16 06:04:57,927][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.010942335240542889, 'train/grad_norm_pre_clip': 1.8947734832763672, 'timing/data': 0.006839255976956338, 'timing/model': 1.4940802209894173, 'learning_rate/qwen_vl_interface': 1.323079860310363e-06, 'learning_rate/action_model': 1.3230798603103628e-05, 'epoch': 1.75}) +[2026-06-16 06:05:03,890][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.09245356172323227, 'train/grad_norm_pre_clip': 6.124184608459473, 'timing/data': 0.006763985031284392, 'timing/model': 1.6950796759920195, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 1.75}) +[2026-06-16 06:05:09,664][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.023425014689564705, 'train/grad_norm_pre_clip': 3.9321820735931396, 'timing/data': 0.007393540989141911, 'timing/model': 1.4682780699804425, 'learning_rate/qwen_vl_interface': 1.3196562989938175e-06, 'learning_rate/action_model': 1.3196562989938174e-05, 'epoch': 1.75}) +[2026-06-16 06:05:15,653][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.021670345216989517, 'train/grad_norm_pre_clip': 5.955005645751953, 'timing/data': 0.006660492974333465, 'timing/model': 1.4583792370394804, 'learning_rate/qwen_vl_interface': 1.3179469414076357e-06, 'learning_rate/action_model': 1.3179469414076356e-05, 'epoch': 1.75}) +[2026-06-16 06:05:21,555][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.32199832797050476, 'train/grad_norm_pre_clip': 14.453856468200684, 'timing/data': 0.006617590028326958, 'timing/model': 1.4620121999760158, 'learning_rate/qwen_vl_interface': 1.3162392001399401e-06, 'learning_rate/action_model': 1.3162392001399401e-05, 'epoch': 1.75}) +[2026-06-16 06:05:27,452][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.016596294939517975, 'train/grad_norm_pre_clip': 2.073838710784912, 'timing/data': 0.006485486053861678, 'timing/model': 1.4598703120136634, 'learning_rate/qwen_vl_interface': 1.3145330758927182e-06, 'learning_rate/action_model': 1.3145330758927182e-05, 'epoch': 1.75}) +[2026-06-16 06:05:33,357][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.27669617533683777, 'train/grad_norm_pre_clip': 18.676475524902344, 'timing/data': 0.006345219037029892, 'timing/model': 1.4312835920136422, 'learning_rate/qwen_vl_interface': 1.3128285693672935e-06, 'learning_rate/action_model': 1.3128285693672934e-05, 'epoch': 1.75}) +[2026-06-16 06:05:39,282][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.09314803034067154, 'train/grad_norm_pre_clip': 11.618130683898926, 'timing/data': 0.006599642976652831, 'timing/model': 1.6875279020168819, 'learning_rate/qwen_vl_interface': 1.311125681264325e-06, 'learning_rate/action_model': 1.311125681264325e-05, 'epoch': 1.75}) +[2026-06-16 06:05:44,951][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.2966058552265167, 'train/grad_norm_pre_clip': 19.46782112121582, 'timing/data': 0.00661040999693796, 'timing/model': 1.4578486629761755, 'learning_rate/qwen_vl_interface': 1.309424412283803e-06, 'learning_rate/action_model': 1.3094244122838029e-05, 'epoch': 1.75}) +[2026-06-16 06:05:50,853][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 0.061355821788311005, 'train/grad_norm_pre_clip': 4.720505237579346, 'timing/data': 0.006470282969530672, 'timing/model': 1.458870375994593, 'learning_rate/qwen_vl_interface': 1.307724763125057e-06, 'learning_rate/action_model': 1.307724763125057e-05, 'epoch': 1.75}) +[2026-06-16 06:05:56,841][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.006249279715120792, 'train/grad_norm_pre_clip': 0.8867076635360718, 'timing/data': 0.006620004016440362, 'timing/model': 1.5438944110064767, 'learning_rate/qwen_vl_interface': 1.3060267344867478e-06, 'learning_rate/action_model': 1.3060267344867478e-05, 'epoch': 1.75}) +[2026-06-16 06:06:02,765][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.06364072114229202, 'train/grad_norm_pre_clip': 9.506144523620605, 'timing/data': 0.006653713004197925, 'timing/model': 1.4674792360165156, 'learning_rate/qwen_vl_interface': 1.3043303270668715e-06, 'learning_rate/action_model': 1.3043303270668713e-05, 'epoch': 1.76}) +[2026-06-16 06:06:08,739][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.009704873897135258, 'train/grad_norm_pre_clip': 1.6209861040115356, 'timing/data': 0.006444464961532503, 'timing/model': 1.508923573011998, 'learning_rate/qwen_vl_interface': 1.3026355415627565e-06, 'learning_rate/action_model': 1.3026355415627564e-05, 'epoch': 1.76}) +[2026-06-16 06:06:14,689][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.12818893790245056, 'train/grad_norm_pre_clip': 16.75456428527832, 'timing/data': 0.006614265032112598, 'timing/model': 1.6925408570095897, 'learning_rate/qwen_vl_interface': 1.3009423786710653e-06, 'learning_rate/action_model': 1.3009423786710653e-05, 'epoch': 1.76}) +[2026-06-16 06:06:20,359][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.11081670969724655, 'train/grad_norm_pre_clip': 12.42639446258545, 'timing/data': 0.006637399026658386, 'timing/model': 1.4615525120170787, 'learning_rate/qwen_vl_interface': 1.2992508390877933e-06, 'learning_rate/action_model': 1.2992508390877933e-05, 'epoch': 1.76}) +[2026-06-16 06:06:26,255][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.14124441146850586, 'train/grad_norm_pre_clip': 12.090716361999512, 'timing/data': 0.006347302987705916, 'timing/model': 1.460717645008117, 'learning_rate/qwen_vl_interface': 1.297560923508269e-06, 'learning_rate/action_model': 1.297560923508269e-05, 'epoch': 1.76}) +[2026-06-16 06:06:32,155][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.013356891460716724, 'train/grad_norm_pre_clip': 1.2703380584716797, 'timing/data': 0.006621646985877305, 'timing/model': 1.4577637710026465, 'learning_rate/qwen_vl_interface': 1.295872632627152e-06, 'learning_rate/action_model': 1.2958726326271519e-05, 'epoch': 1.76}) +[2026-06-16 06:06:38,057][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 0.20326967537403107, 'train/grad_norm_pre_clip': 10.934710502624512, 'timing/data': 0.006522890995256603, 'timing/model': 1.4576231440296397, 'learning_rate/qwen_vl_interface': 1.2941859671384364e-06, 'learning_rate/action_model': 1.2941859671384362e-05, 'epoch': 1.76}) +[2026-06-16 06:06:43,960][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.008534128777682781, 'train/grad_norm_pre_clip': 1.4883157014846802, 'timing/data': 0.006886687013320625, 'timing/model': 1.4337886780267581, 'learning_rate/qwen_vl_interface': 1.2925009277354444e-06, 'learning_rate/action_model': 1.2925009277354444e-05, 'epoch': 1.76}) +[2026-06-16 06:06:49,881][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.12015838176012039, 'train/grad_norm_pre_clip': 14.943170547485352, 'timing/data': 0.006687782995868474, 'timing/model': 1.6840107729658484, 'learning_rate/qwen_vl_interface': 1.2908175151108335e-06, 'learning_rate/action_model': 1.2908175151108334e-05, 'epoch': 1.76}) +[2026-06-16 06:06:55,710][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.013550142757594585, 'train/grad_norm_pre_clip': 2.01204252243042, 'timing/data': 0.009079750976525247, 'timing/model': 1.6099605310009792, 'learning_rate/qwen_vl_interface': 1.2891357299565899e-06, 'learning_rate/action_model': 1.2891357299565898e-05, 'epoch': 1.76}) +[2026-06-16 06:07:01,353][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.0029651615768671036, 'train/grad_norm_pre_clip': 1.020369529724121, 'timing/data': 0.006864002963993698, 'timing/model': 1.459959758038167, 'learning_rate/qwen_vl_interface': 1.2874555729640333e-06, 'learning_rate/action_model': 1.2874555729640334e-05, 'epoch': 1.76}) +[2026-06-16 06:07:07,252][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.00798711646348238, 'train/grad_norm_pre_clip': 1.417669653892517, 'timing/data': 0.006441201025154442, 'timing/model': 1.4592015619855374, 'learning_rate/qwen_vl_interface': 1.2857770448238124e-06, 'learning_rate/action_model': 1.2857770448238123e-05, 'epoch': 1.76}) +[2026-06-16 06:07:13,250][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.02896205522119999, 'train/grad_norm_pre_clip': 4.783805847167969, 'timing/data': 0.006486768019385636, 'timing/model': 1.4876050369930454, 'learning_rate/qwen_vl_interface': 1.2841001462259067e-06, 'learning_rate/action_model': 1.2841001462259067e-05, 'epoch': 1.76}) +[2026-06-16 06:07:19,155][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.013329138979315758, 'train/grad_norm_pre_clip': 1.875296711921692, 'timing/data': 0.007375975954346359, 'timing/model': 1.4297358329640701, 'learning_rate/qwen_vl_interface': 1.282424877859626e-06, 'learning_rate/action_model': 1.282424877859626e-05, 'epoch': 1.76}) +[2026-06-16 06:07:25,093][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.0348043218255043, 'train/grad_norm_pre_clip': 8.03760051727295, 'timing/data': 0.0063828660058788955, 'timing/model': 1.6871199490269646, 'learning_rate/qwen_vl_interface': 1.2807512404136088e-06, 'learning_rate/action_model': 1.2807512404136087e-05, 'epoch': 1.76}) +[2026-06-16 06:07:31,099][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.09339234977960587, 'train/grad_norm_pre_clip': 15.938802719116211, 'timing/data': 0.005968426004983485, 'timing/model': 1.7119094199733809, 'learning_rate/qwen_vl_interface': 1.2790792345758255e-06, 'learning_rate/action_model': 1.2790792345758254e-05, 'epoch': 1.76}) +[2026-06-16 06:07:36,750][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.3223167359828949, 'train/grad_norm_pre_clip': 13.7192964553833, 'timing/data': 0.007225705019664019, 'timing/model': 1.4622600140282884, 'learning_rate/qwen_vl_interface': 1.2774088610335752e-06, 'learning_rate/action_model': 1.2774088610335752e-05, 'epoch': 1.76}) +[2026-06-16 06:07:42,723][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 0.0068182311952114105, 'train/grad_norm_pre_clip': 1.773644208908081, 'timing/data': 0.00660785601940006, 'timing/model': 1.5277180969715118, 'learning_rate/qwen_vl_interface': 1.2757401204734827e-06, 'learning_rate/action_model': 1.2757401204734826e-05, 'epoch': 1.76}) +[2026-06-16 06:07:48,651][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.040125422179698944, 'train/grad_norm_pre_clip': 8.931947708129883, 'timing/data': 0.006475963047705591, 'timing/model': 1.45652175496798, 'learning_rate/qwen_vl_interface': 1.2740730135815064e-06, 'learning_rate/action_model': 1.2740730135815062e-05, 'epoch': 1.76}) +[2026-06-16 06:07:54,558][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 0.037386663258075714, 'train/grad_norm_pre_clip': 7.871553897857666, 'timing/data': 0.007330049003940076, 'timing/model': 1.4283764869906008, 'learning_rate/qwen_vl_interface': 1.2724075410429298e-06, 'learning_rate/action_model': 1.2724075410429299e-05, 'epoch': 1.76}) +[2026-06-16 06:08:00,584][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.24304048717021942, 'train/grad_norm_pre_clip': 16.845157623291016, 'timing/data': 0.006777615984901786, 'timing/model': 1.6934553749742918, 'learning_rate/qwen_vl_interface': 1.2707437035423664e-06, 'learning_rate/action_model': 1.2707437035423664e-05, 'epoch': 1.76}) +[2026-06-16 06:17:14,941][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.20743802189826965, 'train/grad_norm_pre_clip': 12.798150062561035, 'eval/loss': 0.05049687251448631, 'eval/flap_precision': 0.9415322580645161, 'eval/flap_recall': 0.935871743486974, 'eval/flap_f1': 0.9386934673366835, 'eval/flap_rate_model': 0.13777777777777778, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9386934673366834, 'timing/data': 0.006658231024630368, 'timing/model': 1.4569533140165731, 'learning_rate/qwen_vl_interface': 1.2690815017637573e-06, 'learning_rate/action_model': 1.2690815017637572e-05, 'epoch': 1.76}) +[2026-06-16 06:17:14,942][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4100_state +[2026-06-16 06:17:14,957][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 06:17:20,763][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4100_state/model.safetensors +[2026-06-16 06:17:34,711][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4100_state/optimizer.bin +[2026-06-16 06:17:34,712][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4100_state/sampler.bin +[2026-06-16 06:17:34,712][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4100_state/sampler_1.bin +[2026-06-16 06:17:34,713][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4100_state/random_states_0.pkl +[2026-06-16 06:17:34,827][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4100 +[2026-06-16 06:17:34,828][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 06:17:34,838][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 06:17:40,881][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.14273415505886078, 'train/grad_norm_pre_clip': 13.028252601623535, 'timing/data': 0.0070239410270005465, 'timing/model': 1.669536016008351, 'learning_rate/qwen_vl_interface': 1.2674209363903694e-06, 'learning_rate/action_model': 1.2674209363903693e-05, 'epoch': 1.76}) +[2026-06-16 06:17:46,545][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.24271753430366516, 'train/grad_norm_pre_clip': 10.444119453430176, 'timing/data': 0.006582261004950851, 'timing/model': 1.4588396240142174, 'learning_rate/qwen_vl_interface': 1.2657620081048005e-06, 'learning_rate/action_model': 1.2657620081048006e-05, 'epoch': 1.76}) +[2026-06-16 06:17:52,547][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.16130144894123077, 'train/grad_norm_pre_clip': 14.46827507019043, 'timing/data': 0.006784469005651772, 'timing/model': 1.4594645560136996, 'learning_rate/qwen_vl_interface': 1.2641047175889714e-06, 'learning_rate/action_model': 1.2641047175889714e-05, 'epoch': 1.77}) +[2026-06-16 06:17:58,554][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.038561247289180756, 'train/grad_norm_pre_clip': 17.462326049804688, 'timing/data': 0.007659886032342911, 'timing/model': 1.5592309159692377, 'learning_rate/qwen_vl_interface': 1.2624490655241332e-06, 'learning_rate/action_model': 1.262449065524133e-05, 'epoch': 1.77}) +[2026-06-16 06:18:04,452][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.2162054032087326, 'train/grad_norm_pre_clip': 14.02345085144043, 'timing/data': 0.00665465701604262, 'timing/model': 1.4320316729717888, 'learning_rate/qwen_vl_interface': 1.2607950525908602e-06, 'learning_rate/action_model': 1.2607950525908601e-05, 'epoch': 1.77}) +[2026-06-16 06:18:10,358][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.18949615955352783, 'train/grad_norm_pre_clip': 18.954715728759766, 'timing/data': 0.006609852018300444, 'timing/model': 1.4330102049862035, 'learning_rate/qwen_vl_interface': 1.2591426794690575e-06, 'learning_rate/action_model': 1.2591426794690576e-05, 'epoch': 1.77}) +[2026-06-16 06:18:16,242][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.09324383735656738, 'train/grad_norm_pre_clip': 15.543376922607422, 'timing/data': 0.006504176999442279, 'timing/model': 1.6501811160123907, 'learning_rate/qwen_vl_interface': 1.25749194683795e-06, 'learning_rate/action_model': 1.25749194683795e-05, 'epoch': 1.77}) +[2026-06-16 06:18:21,960][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.041918955743312836, 'train/grad_norm_pre_clip': 7.515026569366455, 'timing/data': 0.007406282995361835, 'timing/model': 1.4638708099955693, 'learning_rate/qwen_vl_interface': 1.2558428553760933e-06, 'learning_rate/action_model': 1.2558428553760932e-05, 'epoch': 1.77}) +[2026-06-16 06:18:27,955][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.08312208950519562, 'train/grad_norm_pre_clip': 8.777021408081055, 'timing/data': 0.0066646719933487475, 'timing/model': 1.4597691600210965, 'learning_rate/qwen_vl_interface': 1.2541954057613659e-06, 'learning_rate/action_model': 1.2541954057613658e-05, 'epoch': 1.77}) +[2026-06-16 06:18:33,862][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.06517720967531204, 'train/grad_norm_pre_clip': 10.192018508911133, 'timing/data': 0.006823646021075547, 'timing/model': 1.4341078079887666, 'learning_rate/qwen_vl_interface': 1.2525495986709727e-06, 'learning_rate/action_model': 1.2525495986709726e-05, 'epoch': 1.77}) +[2026-06-16 06:18:39,859][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.062301088124513626, 'train/grad_norm_pre_clip': 10.556551933288574, 'timing/data': 0.007081425050273538, 'timing/model': 1.4402714560274035, 'learning_rate/qwen_vl_interface': 1.2509054347814423e-06, 'learning_rate/action_model': 1.2509054347814423e-05, 'epoch': 1.77}) +[2026-06-16 06:18:45,811][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.07205954194068909, 'train/grad_norm_pre_clip': 7.485838890075684, 'timing/data': 0.006574189988896251, 'timing/model': 1.7119056060328148, 'learning_rate/qwen_vl_interface': 1.2492629147686288e-06, 'learning_rate/action_model': 1.2492629147686289e-05, 'epoch': 1.77}) +[2026-06-16 06:18:51,468][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.18575023114681244, 'train/grad_norm_pre_clip': 16.48851203918457, 'timing/data': 0.006504547025542706, 'timing/model': 1.454869192035403, 'learning_rate/qwen_vl_interface': 1.2476220393077098e-06, 'learning_rate/action_model': 1.2476220393077098e-05, 'epoch': 1.77}) +[2026-06-16 06:18:57,371][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.08354689925909042, 'train/grad_norm_pre_clip': 13.520845413208008, 'timing/data': 0.00669267401099205, 'timing/model': 1.474285261996556, 'learning_rate/qwen_vl_interface': 1.2459828090731877e-06, 'learning_rate/action_model': 1.2459828090731876e-05, 'epoch': 1.77}) +[2026-06-16 06:19:03,168][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.04086306318640709, 'train/grad_norm_pre_clip': 8.293401718139648, 'timing/data': 0.00685206800699234, 'timing/model': 1.46687830099836, 'learning_rate/qwen_vl_interface': 1.2443452247388876e-06, 'learning_rate/action_model': 1.2443452247388877e-05, 'epoch': 1.77}) +[2026-06-16 06:19:09,063][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.024966828525066376, 'train/grad_norm_pre_clip': 15.134629249572754, 'timing/data': 0.007029498985502869, 'timing/model': 1.4688057680032216, 'learning_rate/qwen_vl_interface': 1.2427092869779603e-06, 'learning_rate/action_model': 1.2427092869779602e-05, 'epoch': 1.77}) +[2026-06-16 06:19:14,973][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.1303647756576538, 'train/grad_norm_pre_clip': 15.378375053405762, 'timing/data': 0.0071169079747051, 'timing/model': 1.4898099030251615, 'learning_rate/qwen_vl_interface': 1.2410749964628751e-06, 'learning_rate/action_model': 1.2410749964628751e-05, 'epoch': 1.77}) +[2026-06-16 06:19:20,897][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.04889679327607155, 'train/grad_norm_pre_clip': 14.904460906982422, 'timing/data': 0.006399773003067821, 'timing/model': 1.6920984179596417, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 1.77}) +[2026-06-16 06:19:26,563][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.19713355600833893, 'train/grad_norm_pre_clip': 15.91799545288086, 'timing/data': 0.006707074993755668, 'timing/model': 1.458892212016508, 'learning_rate/qwen_vl_interface': 1.2378113598567394e-06, 'learning_rate/action_model': 1.2378113598567392e-05, 'epoch': 1.77}) +[2026-06-16 06:19:32,457][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.03492167219519615, 'train/grad_norm_pre_clip': 7.08491325378418, 'timing/data': 0.006304031994659454, 'timing/model': 1.4583076550043188, 'learning_rate/qwen_vl_interface': 1.2361820151072454e-06, 'learning_rate/action_model': 1.2361820151072453e-05, 'epoch': 1.77}) +[2026-06-16 06:19:38,366][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.01353328675031662, 'train/grad_norm_pre_clip': 2.106942653656006, 'timing/data': 0.006629259965848178, 'timing/model': 1.4673052510479465, 'learning_rate/qwen_vl_interface': 1.23455432028671e-06, 'learning_rate/action_model': 1.23455432028671e-05, 'epoch': 1.77}) +[2026-06-16 06:19:44,343][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.0630311369895935, 'train/grad_norm_pre_clip': 7.305873870849609, 'timing/data': 0.0065671089687384665, 'timing/model': 1.5097734869923443, 'learning_rate/qwen_vl_interface': 1.2329282760642167e-06, 'learning_rate/action_model': 1.2329282760642166e-05, 'epoch': 1.77}) +[2026-06-16 06:19:50,339][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.10561715811491013, 'train/grad_norm_pre_clip': 13.911942481994629, 'timing/data': 0.006486169993877411, 'timing/model': 1.5509511459968053, 'learning_rate/qwen_vl_interface': 1.2313038831081716e-06, 'learning_rate/action_model': 1.2313038831081717e-05, 'epoch': 1.77}) +[2026-06-16 06:19:56,297][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.3314940333366394, 'train/grad_norm_pre_clip': 29.86981201171875, 'timing/data': 0.006604222988244146, 'timing/model': 1.6892196889966726, 'learning_rate/qwen_vl_interface': 1.2296811420863002e-06, 'learning_rate/action_model': 1.2296811420863003e-05, 'epoch': 1.77}) +[2026-06-16 06:20:01,962][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.009854697622358799, 'train/grad_norm_pre_clip': 1.9798856973648071, 'timing/data': 0.006551616999786347, 'timing/model': 1.46002892695833, 'learning_rate/qwen_vl_interface': 1.2280600536656504e-06, 'learning_rate/action_model': 1.2280600536656503e-05, 'epoch': 1.77}) +[2026-06-16 06:20:07,962][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.03139767423272133, 'train/grad_norm_pre_clip': 11.2184476852417, 'timing/data': 0.0062035649898462, 'timing/model': 1.462431903986726, 'learning_rate/qwen_vl_interface': 1.2264406185125907e-06, 'learning_rate/action_model': 1.2264406185125907e-05, 'epoch': 1.78}) +[2026-06-16 06:20:13,863][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.19951725006103516, 'train/grad_norm_pre_clip': 15.854857444763184, 'timing/data': 0.006730780005455017, 'timing/model': 1.4540182310156524, 'learning_rate/qwen_vl_interface': 1.2248228372928098e-06, 'learning_rate/action_model': 1.2248228372928097e-05, 'epoch': 1.78}) +[2026-06-16 06:20:19,844][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.007926980033516884, 'train/grad_norm_pre_clip': 0.955657422542572, 'timing/data': 0.006519790040329099, 'timing/model': 1.5146263079950586, 'learning_rate/qwen_vl_interface': 1.2232067106713174e-06, 'learning_rate/action_model': 1.2232067106713174e-05, 'epoch': 1.78}) +[2026-06-16 06:20:25,860][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.05946497991681099, 'train/grad_norm_pre_clip': 10.521880149841309, 'timing/data': 0.006935540994163603, 'timing/model': 1.5118938410305418, 'learning_rate/qwen_vl_interface': 1.22159223931244e-06, 'learning_rate/action_model': 1.2215922393124399e-05, 'epoch': 1.78}) +[2026-06-16 06:20:31,792][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.08684302121400833, 'train/grad_norm_pre_clip': 7.929557800292969, 'timing/data': 0.007234390010125935, 'timing/model': 1.680217432032805, 'learning_rate/qwen_vl_interface': 1.2199794238798266e-06, 'learning_rate/action_model': 1.2199794238798264e-05, 'epoch': 1.78}) +[2026-06-16 06:20:37,562][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.14033451676368713, 'train/grad_norm_pre_clip': 10.507865905761719, 'timing/data': 0.006825999997090548, 'timing/model': 1.5106636459822766, 'learning_rate/qwen_vl_interface': 1.2183682650364447e-06, 'learning_rate/action_model': 1.2183682650364447e-05, 'epoch': 1.78}) +[2026-06-16 06:20:43,745][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.14813175797462463, 'train/grad_norm_pre_clip': 11.89022445678711, 'timing/data': 0.007314418035093695, 'timing/model': 1.564193973026704, 'learning_rate/qwen_vl_interface': 1.216758763444581e-06, 'learning_rate/action_model': 1.2167587634445809e-05, 'epoch': 1.78}) +[2026-06-16 06:20:49,759][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.10582179576158524, 'train/grad_norm_pre_clip': 8.000483512878418, 'timing/data': 0.006515643035527319, 'timing/model': 1.4772833320312202, 'learning_rate/qwen_vl_interface': 1.215150919765841e-06, 'learning_rate/action_model': 1.215150919765841e-05, 'epoch': 1.78}) +[2026-06-16 06:20:55,814][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.06761156767606735, 'train/grad_norm_pre_clip': 9.220077514648438, 'timing/data': 0.007491839991416782, 'timing/model': 1.531866026984062, 'learning_rate/qwen_vl_interface': 1.213544734661148e-06, 'learning_rate/action_model': 1.2135447346611478e-05, 'epoch': 1.78}) +[2026-06-16 06:21:01,961][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.019492553547024727, 'train/grad_norm_pre_clip': 5.949527263641357, 'timing/data': 0.006676018994767219, 'timing/model': 1.471604321966879, 'learning_rate/qwen_vl_interface': 1.2119402087907448e-06, 'learning_rate/action_model': 1.2119402087907448e-05, 'epoch': 1.78}) +[2026-06-16 06:21:07,916][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.21372929215431213, 'train/grad_norm_pre_clip': 18.076698303222656, 'timing/data': 0.0067148159723728895, 'timing/model': 1.7121257279650308, 'learning_rate/qwen_vl_interface': 1.2103373428141903e-06, 'learning_rate/action_model': 1.2103373428141903e-05, 'epoch': 1.78}) +[2026-06-16 06:21:13,784][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.05802588537335396, 'train/grad_norm_pre_clip': 5.858415126800537, 'timing/data': 0.0069488699664361775, 'timing/model': 1.4636002629995346, 'learning_rate/qwen_vl_interface': 1.2087361373903633e-06, 'learning_rate/action_model': 1.2087361373903633e-05, 'epoch': 1.78}) +[2026-06-16 06:21:19,968][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.051352955400943756, 'train/grad_norm_pre_clip': 6.6594672203063965, 'timing/data': 0.006606868002563715, 'timing/model': 1.4930933449650183, 'learning_rate/qwen_vl_interface': 1.2071365931774584e-06, 'learning_rate/action_model': 1.2071365931774584e-05, 'epoch': 1.78}) +[2026-06-16 06:21:26,060][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.11565494537353516, 'train/grad_norm_pre_clip': 14.73729133605957, 'timing/data': 0.00674154597800225, 'timing/model': 1.4773057859856635, 'learning_rate/qwen_vl_interface': 1.2055387108329885e-06, 'learning_rate/action_model': 1.2055387108329886e-05, 'epoch': 1.78}) +[2026-06-16 06:21:32,040][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.14471520483493805, 'train/grad_norm_pre_clip': 23.26026153564453, 'timing/data': 0.006899638043250889, 'timing/model': 1.5312905319733545, 'learning_rate/qwen_vl_interface': 1.2039424910137807e-06, 'learning_rate/action_model': 1.2039424910137806e-05, 'epoch': 1.78}) +[2026-06-16 06:21:38,283][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.18435636162757874, 'train/grad_norm_pre_clip': 11.656436920166016, 'timing/data': 0.006800633040256798, 'timing/model': 1.5920232529751956, 'learning_rate/qwen_vl_interface': 1.2023479343759822e-06, 'learning_rate/action_model': 1.2023479343759823e-05, 'epoch': 1.78}) +[2026-06-16 06:21:44,478][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.1057046502828598, 'train/grad_norm_pre_clip': 8.884187698364258, 'timing/data': 0.006608459982089698, 'timing/model': 1.7607906199991703, 'learning_rate/qwen_vl_interface': 1.2007550415750546e-06, 'learning_rate/action_model': 1.2007550415750545e-05, 'epoch': 1.78}) +[2026-06-16 06:21:50,163][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.13932721316814423, 'train/grad_norm_pre_clip': 12.916582107543945, 'timing/data': 0.0073201769846491516, 'timing/model': 1.4527492839843035, 'learning_rate/qwen_vl_interface': 1.1991638132657759e-06, 'learning_rate/action_model': 1.199163813265776e-05, 'epoch': 1.78}) +[2026-06-16 06:21:56,163][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.08100233972072601, 'train/grad_norm_pre_clip': 12.533839225769043, 'timing/data': 0.006591424986254424, 'timing/model': 1.4620721180108376, 'learning_rate/qwen_vl_interface': 1.1975742501022406e-06, 'learning_rate/action_model': 1.1975742501022405e-05, 'epoch': 1.78}) +[2026-06-16 06:22:02,261][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.10443565249443054, 'train/grad_norm_pre_clip': 16.36898422241211, 'timing/data': 0.006659416016191244, 'timing/model': 1.4874063730239868, 'learning_rate/qwen_vl_interface': 1.1959863527378569e-06, 'learning_rate/action_model': 1.1959863527378567e-05, 'epoch': 1.78}) +[2026-06-16 06:22:08,164][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.17871616780757904, 'train/grad_norm_pre_clip': 8.646587371826172, 'timing/data': 0.006816465989686549, 'timing/model': 1.4592466450412758, 'learning_rate/qwen_vl_interface': 1.1944001218253505e-06, 'learning_rate/action_model': 1.1944001218253503e-05, 'epoch': 1.78}) +[2026-06-16 06:22:14,158][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.11509209871292114, 'train/grad_norm_pre_clip': 16.75319480895996, 'timing/data': 0.006776858004741371, 'timing/model': 1.5139832479762845, 'learning_rate/qwen_vl_interface': 1.1928155580167608e-06, 'learning_rate/action_model': 1.1928155580167608e-05, 'epoch': 1.78}) +[2026-06-16 06:22:20,088][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.08577431738376617, 'train/grad_norm_pre_clip': 9.592388153076172, 'timing/data': 0.006624454050324857, 'timing/model': 1.6861463399836794, 'learning_rate/qwen_vl_interface': 1.1912326619634412e-06, 'learning_rate/action_model': 1.1912326619634412e-05, 'epoch': 1.78}) +[2026-06-16 06:22:25,761][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.026627354323863983, 'train/grad_norm_pre_clip': 3.5822596549987793, 'timing/data': 0.007010500994510949, 'timing/model': 1.4581599819939584, 'learning_rate/qwen_vl_interface': 1.1896514343160632e-06, 'learning_rate/action_model': 1.189651434316063e-05, 'epoch': 1.79}) +[2026-06-16 06:22:31,664][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.08366809040307999, 'train/grad_norm_pre_clip': 6.422110080718994, 'timing/data': 0.006741505989339203, 'timing/model': 1.455808344995603, 'learning_rate/qwen_vl_interface': 1.1880718757246062e-06, 'learning_rate/action_model': 1.1880718757246063e-05, 'epoch': 1.79}) +[2026-06-16 06:22:38,159][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.05116185173392296, 'train/grad_norm_pre_clip': 6.887231349945068, 'timing/data': 0.006526289973407984, 'timing/model': 1.5208534499979578, 'learning_rate/qwen_vl_interface': 1.186493986838369e-06, 'learning_rate/action_model': 1.186493986838369e-05, 'epoch': 1.79}) +[2026-06-16 06:22:44,471][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.269575297832489, 'train/grad_norm_pre_clip': 6.716997146606445, 'timing/data': 0.006670620990917087, 'timing/model': 1.486428797012195, 'learning_rate/qwen_vl_interface': 1.1849177683059612e-06, 'learning_rate/action_model': 1.1849177683059612e-05, 'epoch': 1.79}) +[2026-06-16 06:22:50,482][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.027342915534973145, 'train/grad_norm_pre_clip': 3.128674030303955, 'timing/data': 0.006997902994044125, 'timing/model': 1.5108723620069213, 'learning_rate/qwen_vl_interface': 1.1833432207753075e-06, 'learning_rate/action_model': 1.1833432207753076e-05, 'epoch': 1.79}) +[2026-06-16 06:22:56,653][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.06119788438081741, 'train/grad_norm_pre_clip': 9.380257606506348, 'timing/data': 0.011173053993843496, 'timing/model': 1.8429043020005338, 'learning_rate/qwen_vl_interface': 1.1817703448936448e-06, 'learning_rate/action_model': 1.181770344893645e-05, 'epoch': 1.79}) +[2026-06-16 06:23:02,634][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.10936207324266434, 'train/grad_norm_pre_clip': 9.979826927185059, 'timing/data': 0.007019395008683205, 'timing/model': 1.6269800760201178, 'learning_rate/qwen_vl_interface': 1.1801991413075228e-06, 'learning_rate/action_model': 1.1801991413075227e-05, 'epoch': 1.79}) +[2026-06-16 06:23:08,563][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.011558174155652523, 'train/grad_norm_pre_clip': 2.8086416721343994, 'timing/data': 0.006994577997829765, 'timing/model': 1.5507590250344947, 'learning_rate/qwen_vl_interface': 1.1786296106628038e-06, 'learning_rate/action_model': 1.1786296106628036e-05, 'epoch': 1.79}) +[2026-06-16 06:23:14,836][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.24146242439746857, 'train/grad_norm_pre_clip': 20.692745208740234, 'timing/data': 0.00643972196849063, 'timing/model': 1.5637587169767357, 'learning_rate/qwen_vl_interface': 1.1770617536046633e-06, 'learning_rate/action_model': 1.1770617536046632e-05, 'epoch': 1.79}) +[2026-06-16 06:23:21,160][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.031888969242572784, 'train/grad_norm_pre_clip': 3.4377219676971436, 'timing/data': 0.007032022986095399, 'timing/model': 1.578272536979057, 'learning_rate/qwen_vl_interface': 1.1754955707775873e-06, 'learning_rate/action_model': 1.1754955707775872e-05, 'epoch': 1.79}) +[2026-06-16 06:23:27,285][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.02096124179661274, 'train/grad_norm_pre_clip': 12.267289161682129, 'timing/data': 0.006875843973830342, 'timing/model': 1.5359961650101468, 'learning_rate/qwen_vl_interface': 1.1739310628253753e-06, 'learning_rate/action_model': 1.1739310628253753e-05, 'epoch': 1.79}) +[2026-06-16 06:23:33,405][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.1238107681274414, 'train/grad_norm_pre_clip': 9.76744556427002, 'timing/data': 0.006523424992337823, 'timing/model': 1.757105233031325, 'learning_rate/qwen_vl_interface': 1.172368230391136e-06, 'learning_rate/action_model': 1.1723682303911359e-05, 'epoch': 1.79}) +[2026-06-16 06:23:39,288][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.15767884254455566, 'train/grad_norm_pre_clip': 13.343618392944336, 'timing/data': 0.006604955007787794, 'timing/model': 1.6860246120486408, 'learning_rate/qwen_vl_interface': 1.1708070741172928e-06, 'learning_rate/action_model': 1.1708070741172928e-05, 'epoch': 1.79}) +[2026-06-16 06:23:44,962][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.20808039605617523, 'train/grad_norm_pre_clip': 10.08858871459961, 'timing/data': 0.0063947459566406906, 'timing/model': 1.4598465519957244, 'learning_rate/qwen_vl_interface': 1.1692475946455755e-06, 'learning_rate/action_model': 1.1692475946455755e-05, 'epoch': 1.79}) +[2026-06-16 06:23:50,861][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.008157718926668167, 'train/grad_norm_pre_clip': 1.3595271110534668, 'timing/data': 0.006646786991041154, 'timing/model': 1.4575827129883692, 'learning_rate/qwen_vl_interface': 1.1676897926170289e-06, 'learning_rate/action_model': 1.1676897926170289e-05, 'epoch': 1.79}) +[2026-06-16 06:23:56,861][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.07687442749738693, 'train/grad_norm_pre_clip': 10.571084022521973, 'timing/data': 0.0064143239869736135, 'timing/model': 1.4528752400074154, 'learning_rate/qwen_vl_interface': 1.166133668672006e-06, 'learning_rate/action_model': 1.166133668672006e-05, 'epoch': 1.79}) +[2026-06-16 06:24:02,840][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.07315007597208023, 'train/grad_norm_pre_clip': 11.50973129272461, 'timing/data': 0.006581240042578429, 'timing/model': 1.5083899779710919, 'learning_rate/qwen_vl_interface': 1.1645792234501706e-06, 'learning_rate/action_model': 1.1645792234501706e-05, 'epoch': 1.79}) +[2026-06-16 06:24:08,791][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.07131677865982056, 'train/grad_norm_pre_clip': 11.882055282592773, 'timing/data': 0.007396759989205748, 'timing/model': 1.6893899199785665, 'learning_rate/qwen_vl_interface': 1.1630264575904972e-06, 'learning_rate/action_model': 1.1630264575904973e-05, 'epoch': 1.79}) +[2026-06-16 06:24:14,646][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.0810648575425148, 'train/grad_norm_pre_clip': 4.8854594230651855, 'timing/data': 0.006660807004664093, 'timing/model': 1.6403145769727416, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 1.79}) +[2026-06-16 06:24:20,363][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.015476313419640064, 'train/grad_norm_pre_clip': 2.092158794403076, 'timing/data': 0.0066069780150428414, 'timing/model': 1.4582128280308098, 'learning_rate/qwen_vl_interface': 1.1599259665100767e-06, 'learning_rate/action_model': 1.1599259665100767e-05, 'epoch': 1.79}) +[2026-06-16 06:24:26,362][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.03400144726037979, 'train/grad_norm_pre_clip': 9.635257720947266, 'timing/data': 0.006722844031173736, 'timing/model': 1.4538225279538892, 'learning_rate/qwen_vl_interface': 1.1583782425638252e-06, 'learning_rate/action_model': 1.158378242563825e-05, 'epoch': 1.79}) +[2026-06-16 06:24:32,438][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.005937648005783558, 'train/grad_norm_pre_clip': 1.3264641761779785, 'timing/data': 0.006758055998943746, 'timing/model': 1.5536082569742575, 'learning_rate/qwen_vl_interface': 1.1568322005287237e-06, 'learning_rate/action_model': 1.1568322005287235e-05, 'epoch': 1.79}) +[2026-06-16 06:24:38,453][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.04072196036577225, 'train/grad_norm_pre_clip': 5.671295642852783, 'timing/data': 0.006543159019201994, 'timing/model': 1.5135379080311395, 'learning_rate/qwen_vl_interface': 1.1552878410402923e-06, 'learning_rate/action_model': 1.1552878410402924e-05, 'epoch': 1.79}) +[2026-06-16 06:24:44,474][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.12858811020851135, 'train/grad_norm_pre_clip': 15.741238594055176, 'timing/data': 0.007526606030296534, 'timing/model': 1.5396189599996433, 'learning_rate/qwen_vl_interface': 1.1537451647333594e-06, 'learning_rate/action_model': 1.1537451647333593e-05, 'epoch': 1.8}) +[2026-06-16 06:24:50,551][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.14579345285892487, 'train/grad_norm_pre_clip': 15.912660598754883, 'timing/data': 0.006489780033007264, 'timing/model': 1.6490340260206722, 'learning_rate/qwen_vl_interface': 1.1522041722420594e-06, 'learning_rate/action_model': 1.1522041722420594e-05, 'epoch': 1.8}) +[2026-06-16 06:24:56,461][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.11997099965810776, 'train/grad_norm_pre_clip': 13.048199653625488, 'timing/data': 0.006532092986162752, 'timing/model': 1.4590690819895826, 'learning_rate/qwen_vl_interface': 1.1506648641998365e-06, 'learning_rate/action_model': 1.1506648641998365e-05, 'epoch': 1.8}) +[2026-06-16 06:25:02,774][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 0.03688603639602661, 'train/grad_norm_pre_clip': 4.386887073516846, 'timing/data': 0.006703135033603758, 'timing/model': 1.5936053669429384, 'learning_rate/qwen_vl_interface': 1.1491272412394433e-06, 'learning_rate/action_model': 1.1491272412394433e-05, 'epoch': 1.8}) +[2026-06-16 06:25:08,858][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.25140368938446045, 'train/grad_norm_pre_clip': 17.4007568359375, 'timing/data': 0.006574183993507177, 'timing/model': 1.4535912919673137, 'learning_rate/qwen_vl_interface': 1.1475913039929373e-06, 'learning_rate/action_model': 1.1475913039929372e-05, 'epoch': 1.8}) +[2026-06-16 06:25:14,876][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.010237803682684898, 'train/grad_norm_pre_clip': 3.09173846244812, 'timing/data': 0.006822069000918418, 'timing/model': 1.5185969249578193, 'learning_rate/qwen_vl_interface': 1.1460570530916844e-06, 'learning_rate/action_model': 1.1460570530916844e-05, 'epoch': 1.8}) +[2026-06-16 06:25:20,973][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.265455961227417, 'train/grad_norm_pre_clip': 12.625885009765625, 'timing/data': 0.006968064000830054, 'timing/model': 1.790168025996536, 'learning_rate/qwen_vl_interface': 1.1445244891663581e-06, 'learning_rate/action_model': 1.1445244891663581e-05, 'epoch': 1.8}) +[2026-06-16 06:25:26,937][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 0.11847524344921112, 'train/grad_norm_pre_clip': 13.184647560119629, 'timing/data': 0.006657667981926352, 'timing/model': 1.5823818210046738, 'learning_rate/qwen_vl_interface': 1.1429936128469366e-06, 'learning_rate/action_model': 1.1429936128469366e-05, 'epoch': 1.8}) +[2026-06-16 06:25:32,964][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.08701751381158829, 'train/grad_norm_pre_clip': 10.261804580688477, 'timing/data': 0.006868868018500507, 'timing/model': 1.4584207339794375, 'learning_rate/qwen_vl_interface': 1.1414644247627056e-06, 'learning_rate/action_model': 1.1414644247627058e-05, 'epoch': 1.8}) +[2026-06-16 06:25:38,843][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.2867652177810669, 'train/grad_norm_pre_clip': 19.758962631225586, 'timing/data': 0.007227746013086289, 'timing/model': 1.5321259979973547, 'learning_rate/qwen_vl_interface': 1.1399369255422568e-06, 'learning_rate/action_model': 1.1399369255422568e-05, 'epoch': 1.8}) +[2026-06-16 06:25:45,040][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.01432422548532486, 'train/grad_norm_pre_clip': 3.520087480545044, 'timing/data': 0.006821918999776244, 'timing/model': 1.5370957020204514, 'learning_rate/qwen_vl_interface': 1.138411115813488e-06, 'learning_rate/action_model': 1.138411115813488e-05, 'epoch': 1.8}) +[2026-06-16 06:25:51,139][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.17769350111484528, 'train/grad_norm_pre_clip': 10.10058879852295, 'timing/data': 0.006462490011472255, 'timing/model': 1.5380034690024331, 'learning_rate/qwen_vl_interface': 1.1368869962036006e-06, 'learning_rate/action_model': 1.1368869962036005e-05, 'epoch': 1.8}) +[2026-06-16 06:25:57,402][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.06642217189073563, 'train/grad_norm_pre_clip': 4.820126533508301, 'timing/data': 0.007921127020381391, 'timing/model': 1.8170216800062917, 'learning_rate/qwen_vl_interface': 1.1353645673391025e-06, 'learning_rate/action_model': 1.1353645673391026e-05, 'epoch': 1.8}) +[2026-06-16 06:26:03,261][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.01876489445567131, 'train/grad_norm_pre_clip': 2.6204826831817627, 'timing/data': 0.006784432975109667, 'timing/model': 1.4609588169842027, 'learning_rate/qwen_vl_interface': 1.1338438298458078e-06, 'learning_rate/action_model': 1.1338438298458078e-05, 'epoch': 1.8}) +[2026-06-16 06:26:09,352][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.04556986689567566, 'train/grad_norm_pre_clip': 6.2864179611206055, 'timing/data': 0.006999409990385175, 'timing/model': 1.5450295920018107, 'learning_rate/qwen_vl_interface': 1.1323247843488345e-06, 'learning_rate/action_model': 1.1323247843488344e-05, 'epoch': 1.8}) +[2026-06-16 06:26:15,633][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.06898374855518341, 'train/grad_norm_pre_clip': 11.227173805236816, 'timing/data': 0.00888367200968787, 'timing/model': 1.551887484965846, 'learning_rate/qwen_vl_interface': 1.130807431472603e-06, 'learning_rate/action_model': 1.130807431472603e-05, 'epoch': 1.8}) +[2026-06-16 06:26:21,665][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.3595260977745056, 'train/grad_norm_pre_clip': 26.849563598632812, 'timing/data': 0.007021202007308602, 'timing/model': 1.4649717169813812, 'learning_rate/qwen_vl_interface': 1.1292917718408416e-06, 'learning_rate/action_model': 1.1292917718408416e-05, 'epoch': 1.8}) +[2026-06-16 06:26:27,566][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.057137515395879745, 'train/grad_norm_pre_clip': 8.48495101928711, 'timing/data': 0.010078321036417037, 'timing/model': 1.4433348110178486, 'learning_rate/qwen_vl_interface': 1.12777780607658e-06, 'learning_rate/action_model': 1.12777780607658e-05, 'epoch': 1.8}) +[2026-06-16 06:26:33,599][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.2877204418182373, 'train/grad_norm_pre_clip': 15.77616024017334, 'timing/data': 0.014760700985789299, 'timing/model': 1.7496751790167764, 'learning_rate/qwen_vl_interface': 1.1262655348021523e-06, 'learning_rate/action_model': 1.1262655348021523e-05, 'epoch': 1.8}) +[2026-06-16 06:26:39,264][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.17187103629112244, 'train/grad_norm_pre_clip': 11.831239700317383, 'timing/data': 0.007671369006857276, 'timing/model': 1.4567858600057662, 'learning_rate/qwen_vl_interface': 1.1247549586391966e-06, 'learning_rate/action_model': 1.1247549586391964e-05, 'epoch': 1.8}) +[2026-06-16 06:26:45,266][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.04007688909769058, 'train/grad_norm_pre_clip': 7.003425121307373, 'timing/data': 0.007223568973131478, 'timing/model': 1.4597199090057984, 'learning_rate/qwen_vl_interface': 1.1232460782086534e-06, 'learning_rate/action_model': 1.1232460782086534e-05, 'epoch': 1.8}) +[2026-06-16 06:26:51,362][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.15151193737983704, 'train/grad_norm_pre_clip': 10.775778770446777, 'timing/data': 0.006985999993048608, 'timing/model': 1.473156874999404, 'learning_rate/qwen_vl_interface': 1.1217388941307658e-06, 'learning_rate/action_model': 1.1217388941307658e-05, 'epoch': 1.8}) +[2026-06-16 06:26:57,341][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.07797746360301971, 'train/grad_norm_pre_clip': 12.859103202819824, 'timing/data': 0.006625780020840466, 'timing/model': 1.511416515975725, 'learning_rate/qwen_vl_interface': 1.1202334070250807e-06, 'learning_rate/action_model': 1.1202334070250806e-05, 'epoch': 1.8}) +[2026-06-16 06:27:03,261][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.11757422983646393, 'train/grad_norm_pre_clip': 9.383441925048828, 'timing/data': 0.006593161961063743, 'timing/model': 1.456008936977014, 'learning_rate/qwen_vl_interface': 1.1187296175104472e-06, 'learning_rate/action_model': 1.1187296175104471e-05, 'epoch': 1.81}) +[2026-06-16 06:27:09,192][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.052600834518671036, 'train/grad_norm_pre_clip': 11.352219581604004, 'timing/data': 0.006640250969212502, 'timing/model': 1.6930940090096556, 'learning_rate/qwen_vl_interface': 1.1172275262050158e-06, 'learning_rate/action_model': 1.1172275262050159e-05, 'epoch': 1.81}) +[2026-06-16 06:27:14,859][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.13129118084907532, 'train/grad_norm_pre_clip': 9.27746295928955, 'timing/data': 0.006776470982003957, 'timing/model': 1.4572280279826373, 'learning_rate/qwen_vl_interface': 1.1157271337262401e-06, 'learning_rate/action_model': 1.11572713372624e-05, 'epoch': 1.81}) +[2026-06-16 06:27:20,759][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 0.08353491872549057, 'train/grad_norm_pre_clip': 9.698564529418945, 'timing/data': 0.0066196099505759776, 'timing/model': 1.4573592740343884, 'learning_rate/qwen_vl_interface': 1.1142284406908751e-06, 'learning_rate/action_model': 1.114228440690875e-05, 'epoch': 1.81}) +[2026-06-16 06:27:26,663][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.07824301719665527, 'train/grad_norm_pre_clip': 8.951835632324219, 'timing/data': 0.006672158022411168, 'timing/model': 1.4608832120429724, 'learning_rate/qwen_vl_interface': 1.1127314477149758e-06, 'learning_rate/action_model': 1.1127314477149759e-05, 'epoch': 1.81}) +[2026-06-16 06:36:21,459][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.0010573193430900574, 'train/grad_norm_pre_clip': 0.15460246801376343, 'eval/loss': 0.04723859578371048, 'eval/flap_precision': 0.9434343434343434, 'eval/flap_recall': 0.935871743486974, 'eval/flap_f1': 0.9396378269617707, 'eval/flap_rate_model': 0.1375, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9396378269617706, 'timing/data': 0.007524542044848204, 'timing/model': 1.4580991179682314, 'learning_rate/qwen_vl_interface': 1.1112361554139003e-06, 'learning_rate/action_model': 1.1112361554139004e-05, 'epoch': 1.81}) +[2026-06-16 06:36:21,460][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4200_state +[2026-06-16 06:36:21,464][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 06:36:26,972][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4200_state/model.safetensors +[2026-06-16 06:36:40,990][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4200_state/optimizer.bin +[2026-06-16 06:36:40,991][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4200_state/sampler.bin +[2026-06-16 06:36:40,991][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4200_state/sampler_1.bin +[2026-06-16 06:36:40,992][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4200_state/random_states_0.pkl +[2026-06-16 06:36:41,173][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4200 +[2026-06-16 06:36:41,173][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 06:36:41,183][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 06:36:47,138][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.054323602467775345, 'train/grad_norm_pre_clip': 8.245607376098633, 'timing/data': 0.006212430016603321, 'timing/model': 1.4603051030426286, 'learning_rate/qwen_vl_interface': 1.1097425644023065e-06, 'learning_rate/action_model': 1.1097425644023065e-05, 'epoch': 1.81}) +[2026-06-16 06:36:53,038][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 0.15183371305465698, 'train/grad_norm_pre_clip': 12.73612117767334, 'timing/data': 0.0060993030201643705, 'timing/model': 1.4556192770251073, 'learning_rate/qwen_vl_interface': 1.1082506752941529e-06, 'learning_rate/action_model': 1.1082506752941528e-05, 'epoch': 1.81}) +[2026-06-16 06:36:59,041][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.16738586127758026, 'train/grad_norm_pre_clip': 15.123041152954102, 'timing/data': 0.006871438992675394, 'timing/model': 1.4281115640187636, 'learning_rate/qwen_vl_interface': 1.1067604887027004e-06, 'learning_rate/action_model': 1.1067604887027004e-05, 'epoch': 1.81}) +[2026-06-16 06:37:04,984][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.18264496326446533, 'train/grad_norm_pre_clip': 10.838791847229004, 'timing/data': 0.007398214016575366, 'timing/model': 1.6749592859996483, 'learning_rate/qwen_vl_interface': 1.1052720052405066e-06, 'learning_rate/action_model': 1.1052720052405064e-05, 'epoch': 1.81}) +[2026-06-16 06:37:10,762][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.044827308505773544, 'train/grad_norm_pre_clip': 7.247294902801514, 'timing/data': 0.006946720008272678, 'timing/model': 1.4645180120132864, 'learning_rate/qwen_vl_interface': 1.1037852255194309e-06, 'learning_rate/action_model': 1.1037852255194308e-05, 'epoch': 1.81}) +[2026-06-16 06:37:16,849][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 0.24732589721679688, 'train/grad_norm_pre_clip': 11.559743881225586, 'timing/data': 0.011626515013631433, 'timing/model': 1.4679553269525059, 'learning_rate/qwen_vl_interface': 1.1023001501506329e-06, 'learning_rate/action_model': 1.102300150150633e-05, 'epoch': 1.81}) +[2026-06-16 06:37:22,854][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.20797206461429596, 'train/grad_norm_pre_clip': 10.980551719665527, 'timing/data': 0.006642952968832105, 'timing/model': 1.4427349000470713, 'learning_rate/qwen_vl_interface': 1.1008167797445717e-06, 'learning_rate/action_model': 1.1008167797445718e-05, 'epoch': 1.81}) +[2026-06-16 06:37:28,763][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.08832218497991562, 'train/grad_norm_pre_clip': 10.880356788635254, 'timing/data': 0.006850778008811176, 'timing/model': 1.4489207590231672, 'learning_rate/qwen_vl_interface': 1.099335114911003e-06, 'learning_rate/action_model': 1.0993351149110031e-05, 'epoch': 1.81}) +[2026-06-16 06:37:34,760][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.05503208562731743, 'train/grad_norm_pre_clip': 5.5764689445495605, 'timing/data': 0.0063295820145867765, 'timing/model': 1.5187217780039646, 'learning_rate/qwen_vl_interface': 1.0978551562589846e-06, 'learning_rate/action_model': 1.0978551562589845e-05, 'epoch': 1.81}) +[2026-06-16 06:37:40,832][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.021749285981059074, 'train/grad_norm_pre_clip': 2.313197612762451, 'timing/data': 0.007368420017883182, 'timing/model': 1.7203937539597973, 'learning_rate/qwen_vl_interface': 1.0963769043968708e-06, 'learning_rate/action_model': 1.0963769043968708e-05, 'epoch': 1.81}) +[2026-06-16 06:37:46,558][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.035374149680137634, 'train/grad_norm_pre_clip': 11.968498229980469, 'timing/data': 0.006617686012759805, 'timing/model': 1.44548155303346, 'learning_rate/qwen_vl_interface': 1.0949003599323158e-06, 'learning_rate/action_model': 1.0949003599323157e-05, 'epoch': 1.81}) +[2026-06-16 06:37:52,558][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 0.19988971948623657, 'train/grad_norm_pre_clip': 13.222816467285156, 'timing/data': 0.0065369170042686164, 'timing/model': 1.4637848989805207, 'learning_rate/qwen_vl_interface': 1.0934255234722708e-06, 'learning_rate/action_model': 1.0934255234722706e-05, 'epoch': 1.81}) +[2026-06-16 06:37:58,363][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.062225136905908585, 'train/grad_norm_pre_clip': 5.274863243103027, 'timing/data': 0.006704944011289626, 'timing/model': 1.4479832500219345, 'learning_rate/qwen_vl_interface': 1.0919523956229858e-06, 'learning_rate/action_model': 1.0919523956229857e-05, 'epoch': 1.81}) +[2026-06-16 06:38:04,360][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.01496201567351818, 'train/grad_norm_pre_clip': 1.762926459312439, 'timing/data': 0.00665304803987965, 'timing/model': 1.4914873010129668, 'learning_rate/qwen_vl_interface': 1.0904809769900075e-06, 'learning_rate/action_model': 1.0904809769900073e-05, 'epoch': 1.81}) +[2026-06-16 06:38:10,539][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.042044151574373245, 'train/grad_norm_pre_clip': 7.331071376800537, 'timing/data': 0.006705375039018691, 'timing/model': 1.504656313045416, 'learning_rate/qwen_vl_interface': 1.0890112681781821e-06, 'learning_rate/action_model': 1.089011268178182e-05, 'epoch': 1.81}) +[2026-06-16 06:38:16,498][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.06732839345932007, 'train/grad_norm_pre_clip': 5.888692378997803, 'timing/data': 0.006636323989368975, 'timing/model': 1.6969081080169417, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 1.81}) +[2026-06-16 06:38:22,161][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.06350814551115036, 'train/grad_norm_pre_clip': 8.22159194946289, 'timing/data': 0.006528344994876534, 'timing/model': 1.45994020299986, 'learning_rate/qwen_vl_interface': 1.0860769824338477e-06, 'learning_rate/action_model': 1.0860769824338478e-05, 'epoch': 1.81}) +[2026-06-16 06:38:28,158][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.004371847026050091, 'train/grad_norm_pre_clip': 1.6723700761795044, 'timing/data': 0.00643793202470988, 'timing/model': 1.510678877006285, 'learning_rate/qwen_vl_interface': 1.084612406707514e-06, 'learning_rate/action_model': 1.084612406707514e-05, 'epoch': 1.81}) +[2026-06-16 06:38:34,059][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.10697641968727112, 'train/grad_norm_pre_clip': 8.395237922668457, 'timing/data': 0.006983955041505396, 'timing/model': 1.4585560329724103, 'learning_rate/qwen_vl_interface': 1.0831495432146795e-06, 'learning_rate/action_model': 1.0831495432146795e-05, 'epoch': 1.82}) +[2026-06-16 06:38:40,059][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.11615054309368134, 'train/grad_norm_pre_clip': 8.593486785888672, 'timing/data': 0.006784049968700856, 'timing/model': 1.5129706349689513, 'learning_rate/qwen_vl_interface': 1.0816883925566717e-06, 'learning_rate/action_model': 1.0816883925566716e-05, 'epoch': 1.82}) +[2026-06-16 06:38:46,061][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.052753299474716187, 'train/grad_norm_pre_clip': 7.451632499694824, 'timing/data': 0.006456559000071138, 'timing/model': 1.4533755669835955, 'learning_rate/qwen_vl_interface': 1.0802289553341153e-06, 'learning_rate/action_model': 1.0802289553341152e-05, 'epoch': 1.82}) +[2026-06-16 06:38:52,000][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.09844869375228882, 'train/grad_norm_pre_clip': 4.19781494140625, 'timing/data': 0.0064938029972836375, 'timing/model': 1.7004323599976487, 'learning_rate/qwen_vl_interface': 1.0787712321469289e-06, 'learning_rate/action_model': 1.0787712321469288e-05, 'epoch': 1.82}) +[2026-06-16 06:38:57,889][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.03585215285420418, 'train/grad_norm_pre_clip': 5.315279006958008, 'timing/data': 0.006535344989970326, 'timing/model': 1.6885442999773659, 'learning_rate/qwen_vl_interface': 1.077315223594328e-06, 'learning_rate/action_model': 1.077315223594328e-05, 'epoch': 1.82}) +[2026-06-16 06:39:03,654][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.2552945017814636, 'train/grad_norm_pre_clip': 19.97981834411621, 'timing/data': 0.006446202984079719, 'timing/model': 1.4787452960154042, 'learning_rate/qwen_vl_interface': 1.0758609302748227e-06, 'learning_rate/action_model': 1.0758609302748227e-05, 'epoch': 1.82}) +[2026-06-16 06:39:09,658][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.1991756707429886, 'train/grad_norm_pre_clip': 9.723621368408203, 'timing/data': 0.006652346986811608, 'timing/model': 1.4759993439656682, 'learning_rate/qwen_vl_interface': 1.0744083527862182e-06, 'learning_rate/action_model': 1.074408352786218e-05, 'epoch': 1.82}) +[2026-06-16 06:39:15,659][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.050283849239349365, 'train/grad_norm_pre_clip': 6.940041542053223, 'timing/data': 0.006729811022523791, 'timing/model': 1.4784181310096756, 'learning_rate/qwen_vl_interface': 1.0729574917256144e-06, 'learning_rate/action_model': 1.0729574917256145e-05, 'epoch': 1.82}) +[2026-06-16 06:39:21,647][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.13421806693077087, 'train/grad_norm_pre_clip': 13.668901443481445, 'timing/data': 0.006762388977222145, 'timing/model': 1.5019978990312666, 'learning_rate/qwen_vl_interface': 1.0715083476894047e-06, 'learning_rate/action_model': 1.0715083476894047e-05, 'epoch': 1.82}) +[2026-06-16 06:39:27,597][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 0.0897473618388176, 'train/grad_norm_pre_clip': 10.739702224731445, 'timing/data': 0.006784771976526827, 'timing/model': 1.6992353969835676, 'learning_rate/qwen_vl_interface': 1.0700609212732773e-06, 'learning_rate/action_model': 1.0700609212732774e-05, 'epoch': 1.82}) +[2026-06-16 06:39:33,256][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.12293443828821182, 'train/grad_norm_pre_clip': 5.677216053009033, 'timing/data': 0.006473213026765734, 'timing/model': 1.4563604450086132, 'learning_rate/qwen_vl_interface': 1.0686152130722155e-06, 'learning_rate/action_model': 1.0686152130722154e-05, 'epoch': 1.82}) +[2026-06-16 06:39:39,357][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.11643523722887039, 'train/grad_norm_pre_clip': 6.946942329406738, 'timing/data': 0.006681650003883988, 'timing/model': 1.4885296099819243, 'learning_rate/qwen_vl_interface': 1.0671712236804947e-06, 'learning_rate/action_model': 1.0671712236804946e-05, 'epoch': 1.82}) +[2026-06-16 06:39:45,359][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 0.0023412799928337336, 'train/grad_norm_pre_clip': 0.32707372307777405, 'timing/data': 0.0066041359677910805, 'timing/model': 1.481165494013112, 'learning_rate/qwen_vl_interface': 1.0657289536916845e-06, 'learning_rate/action_model': 1.0657289536916845e-05, 'epoch': 1.82}) +[2026-06-16 06:39:51,358][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.01198155339807272, 'train/grad_norm_pre_clip': 3.17193603515625, 'timing/data': 0.0069092249614186585, 'timing/model': 1.4913133319932967, 'learning_rate/qwen_vl_interface': 1.0642884036986481e-06, 'learning_rate/action_model': 1.064288403698648e-05, 'epoch': 1.82}) +[2026-06-16 06:39:57,355][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.009784412570297718, 'train/grad_norm_pre_clip': 1.1135503053665161, 'timing/data': 0.006723030004650354, 'timing/model': 1.434119498997461, 'learning_rate/qwen_vl_interface': 1.0628495742935404e-06, 'learning_rate/action_model': 1.0628495742935405e-05, 'epoch': 1.82}) +[2026-06-16 06:40:03,385][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 0.0586179755628109, 'train/grad_norm_pre_clip': 6.498199939727783, 'timing/data': 0.006610594980884343, 'timing/model': 1.7056912340340205, 'learning_rate/qwen_vl_interface': 1.0614124660678103e-06, 'learning_rate/action_model': 1.0614124660678102e-05, 'epoch': 1.82}) +[2026-06-16 06:40:09,058][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.2963891625404358, 'train/grad_norm_pre_clip': 23.16543197631836, 'timing/data': 0.006650583993177861, 'timing/model': 1.4587947470135987, 'learning_rate/qwen_vl_interface': 1.0599770796121988e-06, 'learning_rate/action_model': 1.0599770796121987e-05, 'epoch': 1.82}) +[2026-06-16 06:40:15,058][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.07661670446395874, 'train/grad_norm_pre_clip': 5.543765068054199, 'timing/data': 0.006603674963116646, 'timing/model': 1.460058155003935, 'learning_rate/qwen_vl_interface': 1.0585434155167394e-06, 'learning_rate/action_model': 1.0585434155167393e-05, 'epoch': 1.82}) +[2026-06-16 06:40:21,054][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.07943855226039886, 'train/grad_norm_pre_clip': 9.725615501403809, 'timing/data': 0.006612798024434596, 'timing/model': 1.4573332340223715, 'learning_rate/qwen_vl_interface': 1.0571114743707556e-06, 'learning_rate/action_model': 1.0571114743707555e-05, 'epoch': 1.82}) +[2026-06-16 06:40:27,030][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.132362499833107, 'train/grad_norm_pre_clip': 8.28418254852295, 'timing/data': 0.006832600978668779, 'timing/model': 1.5012920929584652, 'learning_rate/qwen_vl_interface': 1.0556812567628661e-06, 'learning_rate/action_model': 1.055681256762866e-05, 'epoch': 1.82}) +[2026-06-16 06:40:33,050][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.025614283978939056, 'train/grad_norm_pre_clip': 1.5894813537597656, 'timing/data': 0.006429579050745815, 'timing/model': 1.5039830799796619, 'learning_rate/qwen_vl_interface': 1.0542527632809783e-06, 'learning_rate/action_model': 1.0542527632809784e-05, 'epoch': 1.82}) +[2026-06-16 06:40:39,098][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.11078749597072601, 'train/grad_norm_pre_clip': 13.298519134521484, 'timing/data': 0.006369450013153255, 'timing/model': 1.7174612989765592, 'learning_rate/qwen_vl_interface': 1.0528259945122925e-06, 'learning_rate/action_model': 1.0528259945122925e-05, 'epoch': 1.82}) +[2026-06-16 06:40:44,874][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.1490941345691681, 'train/grad_norm_pre_clip': 16.636966705322266, 'timing/data': 0.006713777023833245, 'timing/model': 1.4954657279886305, 'learning_rate/qwen_vl_interface': 1.0514009510432988e-06, 'learning_rate/action_model': 1.0514009510432988e-05, 'epoch': 1.82}) +[2026-06-16 06:40:50,860][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.14321346580982208, 'train/grad_norm_pre_clip': 12.874940872192383, 'timing/data': 0.006916314945556223, 'timing/model': 1.4616171090165153, 'learning_rate/qwen_vl_interface': 1.0499776334597796e-06, 'learning_rate/action_model': 1.0499776334597797e-05, 'epoch': 1.83}) +[2026-06-16 06:40:56,940][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.006556390784680843, 'train/grad_norm_pre_clip': 0.6971730589866638, 'timing/data': 0.0071337950066663325, 'timing/model': 1.5441797910025343, 'learning_rate/qwen_vl_interface': 1.0485560423468064e-06, 'learning_rate/action_model': 1.0485560423468063e-05, 'epoch': 1.83}) +[2026-06-16 06:41:02,952][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.26398414373397827, 'train/grad_norm_pre_clip': 14.432429313659668, 'timing/data': 0.006643583998084068, 'timing/model': 1.5028370369691402, 'learning_rate/qwen_vl_interface': 1.0471361782887421e-06, 'learning_rate/action_model': 1.0471361782887421e-05, 'epoch': 1.83}) +[2026-06-16 06:41:08,897][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.23471882939338684, 'train/grad_norm_pre_clip': 10.264897346496582, 'timing/data': 0.0067526340135373175, 'timing/model': 1.6998888699454255, 'learning_rate/qwen_vl_interface': 1.0457180418692385e-06, 'learning_rate/action_model': 1.0457180418692383e-05, 'epoch': 1.83}) +[2026-06-16 06:41:14,561][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.018576104193925858, 'train/grad_norm_pre_clip': 2.774052858352661, 'timing/data': 0.007540683029219508, 'timing/model': 1.465250779001508, 'learning_rate/qwen_vl_interface': 1.0443016336712395e-06, 'learning_rate/action_model': 1.0443016336712393e-05, 'epoch': 1.83}) +[2026-06-16 06:41:20,555][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.10356874763965607, 'train/grad_norm_pre_clip': 8.147196769714355, 'timing/data': 0.006600530003197491, 'timing/model': 1.4594927169964649, 'learning_rate/qwen_vl_interface': 1.0428869542769744e-06, 'learning_rate/action_model': 1.0428869542769744e-05, 'epoch': 1.83}) +[2026-06-16 06:41:26,355][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.12841829657554626, 'train/grad_norm_pre_clip': 9.779288291931152, 'timing/data': 0.007232921023387462, 'timing/model': 1.4451741310185753, 'learning_rate/qwen_vl_interface': 1.0414740042679663e-06, 'learning_rate/action_model': 1.0414740042679662e-05, 'epoch': 1.83}) +[2026-06-16 06:41:32,257][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.0169115848839283, 'train/grad_norm_pre_clip': 2.994774580001831, 'timing/data': 0.006781227013561875, 'timing/model': 1.4547708579921164, 'learning_rate/qwen_vl_interface': 1.040062784225025e-06, 'learning_rate/action_model': 1.0400627842250249e-05, 'epoch': 1.83}) +[2026-06-16 06:41:38,155][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.04522016644477844, 'train/grad_norm_pre_clip': 6.891185283660889, 'timing/data': 0.007589554996229708, 'timing/model': 1.465947769000195, 'learning_rate/qwen_vl_interface': 1.0386532947282496e-06, 'learning_rate/action_model': 1.0386532947282495e-05, 'epoch': 1.83}) +[2026-06-16 06:41:44,289][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.06593295186758041, 'train/grad_norm_pre_clip': 9.430940628051758, 'timing/data': 0.0066786049865186214, 'timing/model': 1.7068811970530078, 'learning_rate/qwen_vl_interface': 1.0372455363570282e-06, 'learning_rate/action_model': 1.0372455363570283e-05, 'epoch': 1.83}) +[2026-06-16 06:41:49,954][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.02356504462659359, 'train/grad_norm_pre_clip': 3.7005300521850586, 'timing/data': 0.0072010940057225525, 'timing/model': 1.4526456779567525, 'learning_rate/qwen_vl_interface': 1.035839509690039e-06, 'learning_rate/action_model': 1.0358395096900388e-05, 'epoch': 1.83}) +[2026-06-16 06:41:55,858][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.06035511568188667, 'train/grad_norm_pre_clip': 9.768263816833496, 'timing/data': 0.006290825025644153, 'timing/model': 1.4585331979906186, 'learning_rate/qwen_vl_interface': 1.0344352153052429e-06, 'learning_rate/action_model': 1.0344352153052427e-05, 'epoch': 1.83}) +[2026-06-16 06:42:01,759][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.027565432712435722, 'train/grad_norm_pre_clip': 3.657731056213379, 'timing/data': 0.006318564992398024, 'timing/model': 1.4565838390262797, 'learning_rate/qwen_vl_interface': 1.033032653779894e-06, 'learning_rate/action_model': 1.033032653779894e-05, 'epoch': 1.83}) +[2026-06-16 06:42:07,657][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.051371872425079346, 'train/grad_norm_pre_clip': 6.982600212097168, 'timing/data': 0.006505360011942685, 'timing/model': 1.4536827289848588, 'learning_rate/qwen_vl_interface': 1.0316318256905325e-06, 'learning_rate/action_model': 1.0316318256905324e-05, 'epoch': 1.83}) +[2026-06-16 06:42:13,556][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.08582457154989243, 'train/grad_norm_pre_clip': 11.875455856323242, 'timing/data': 0.0067380929831415415, 'timing/model': 1.467882387980353, 'learning_rate/qwen_vl_interface': 1.0302327316129856e-06, 'learning_rate/action_model': 1.0302327316129855e-05, 'epoch': 1.83}) +[2026-06-16 06:42:19,579][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.1700877994298935, 'train/grad_norm_pre_clip': 10.651679039001465, 'timing/data': 0.006458832009229809, 'timing/model': 1.7780423570075072, 'learning_rate/qwen_vl_interface': 1.028835372122368e-06, 'learning_rate/action_model': 1.0288353721223682e-05, 'epoch': 1.83}) +[2026-06-16 06:42:25,256][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 0.04564819484949112, 'train/grad_norm_pre_clip': 4.1553120613098145, 'timing/data': 0.006577695952728391, 'timing/model': 1.4588582700234838, 'learning_rate/qwen_vl_interface': 1.0274397477930822e-06, 'learning_rate/action_model': 1.0274397477930822e-05, 'epoch': 1.83}) +[2026-06-16 06:42:31,158][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.0049039265140891075, 'train/grad_norm_pre_clip': 1.0142607688903809, 'timing/data': 0.006577886990271509, 'timing/model': 1.4607181249884889, 'learning_rate/qwen_vl_interface': 1.0260458591988164e-06, 'learning_rate/action_model': 1.0260458591988165e-05, 'epoch': 1.83}) +[2026-06-16 06:42:37,151][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.09985680133104324, 'train/grad_norm_pre_clip': 6.9910664558410645, 'timing/data': 0.006576234998647124, 'timing/model': 1.4590474650030956, 'learning_rate/qwen_vl_interface': 1.0246537069125445e-06, 'learning_rate/action_model': 1.0246537069125444e-05, 'epoch': 1.83}) +[2026-06-16 06:42:43,157][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.019008636474609375, 'train/grad_norm_pre_clip': 2.075976848602295, 'timing/data': 0.006780655006878078, 'timing/model': 1.4329527859808877, 'learning_rate/qwen_vl_interface': 1.0232632915065286e-06, 'learning_rate/action_model': 1.0232632915065285e-05, 'epoch': 1.83}) +[2026-06-16 06:42:49,054][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.007161244284361601, 'train/grad_norm_pre_clip': 1.0819849967956543, 'timing/data': 0.006508785008918494, 'timing/model': 1.448122360976413, 'learning_rate/qwen_vl_interface': 1.021874613552316e-06, 'learning_rate/action_model': 1.0218746135523157e-05, 'epoch': 1.83}) +[2026-06-16 06:42:55,097][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.06853263080120087, 'train/grad_norm_pre_clip': 3.963432550430298, 'timing/data': 0.006253639992792159, 'timing/model': 1.7047265210421756, 'learning_rate/qwen_vl_interface': 1.0204876736207384e-06, 'learning_rate/action_model': 1.0204876736207384e-05, 'epoch': 1.83}) +[2026-06-16 06:43:00,957][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.13535523414611816, 'train/grad_norm_pre_clip': 8.217646598815918, 'timing/data': 0.00647775997640565, 'timing/model': 1.4830263280309737, 'learning_rate/qwen_vl_interface': 1.0191024722819163e-06, 'learning_rate/action_model': 1.0191024722819162e-05, 'epoch': 1.83}) +[2026-06-16 06:43:06,967][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.03186187893152237, 'train/grad_norm_pre_clip': 4.795596599578857, 'timing/data': 0.006672304996754974, 'timing/model': 1.4655404320219532, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 1.84}) +[2026-06-16 06:43:12,934][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.061456333845853806, 'train/grad_norm_pre_clip': 12.11214542388916, 'timing/data': 0.006427877000533044, 'timing/model': 1.5303870800416917, 'learning_rate/qwen_vl_interface': 1.016337287659437e-06, 'learning_rate/action_model': 1.0163372876594368e-05, 'epoch': 1.84}) +[2026-06-16 06:43:19,134][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.1370663046836853, 'train/grad_norm_pre_clip': 9.704536437988281, 'timing/data': 0.006828584999311715, 'timing/model': 1.5904951160191558, 'learning_rate/qwen_vl_interface': 1.0149573055124427e-06, 'learning_rate/action_model': 1.0149573055124426e-05, 'epoch': 1.84}) +[2026-06-16 06:43:25,139][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.11056086421012878, 'train/grad_norm_pre_clip': 9.48202896118164, 'timing/data': 0.0066873879986815155, 'timing/model': 1.5050964819965884, 'learning_rate/qwen_vl_interface': 1.013579064231529e-06, 'learning_rate/action_model': 1.013579064231529e-05, 'epoch': 1.84}) +[2026-06-16 06:43:31,095][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.07472839206457138, 'train/grad_norm_pre_clip': 10.290929794311523, 'timing/data': 0.00670429301680997, 'timing/model': 1.6967398339766078, 'learning_rate/qwen_vl_interface': 1.012202564383239e-06, 'learning_rate/action_model': 1.0122025643832388e-05, 'epoch': 1.84}) +[2026-06-16 06:43:36,829][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.06035659834742546, 'train/grad_norm_pre_clip': 8.287141799926758, 'timing/data': 0.007323774974793196, 'timing/model': 1.5292677889810875, 'learning_rate/qwen_vl_interface': 1.0108278065333978e-06, 'learning_rate/action_model': 1.0108278065333978e-05, 'epoch': 1.84}) +[2026-06-16 06:43:42,861][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.129047229886055, 'train/grad_norm_pre_clip': 10.160197257995605, 'timing/data': 0.006663142004981637, 'timing/model': 1.4612995170173235, 'learning_rate/qwen_vl_interface': 1.0094547912471183e-06, 'learning_rate/action_model': 1.0094547912471184e-05, 'epoch': 1.84}) +[2026-06-16 06:43:48,856][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.07161951810121536, 'train/grad_norm_pre_clip': 11.302271842956543, 'timing/data': 0.0075097979861311615, 'timing/model': 1.4522789549664594, 'learning_rate/qwen_vl_interface': 1.0080835190887949e-06, 'learning_rate/action_model': 1.0080835190887947e-05, 'epoch': 1.84}) +[2026-06-16 06:43:54,770][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.3533123731613159, 'train/grad_norm_pre_clip': 16.72199249267578, 'timing/data': 0.006585208000615239, 'timing/model': 1.471468064992223, 'learning_rate/qwen_vl_interface': 1.0067139906221052e-06, 'learning_rate/action_model': 1.0067139906221052e-05, 'epoch': 1.84}) +[2026-06-16 06:44:00,825][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 0.17656210064888, 'train/grad_norm_pre_clip': 22.014686584472656, 'timing/data': 0.006574703031219542, 'timing/model': 1.4922844370012172, 'learning_rate/qwen_vl_interface': 1.0053462064100112e-06, 'learning_rate/action_model': 1.0053462064100112e-05, 'epoch': 1.84}) +[2026-06-16 06:44:06,791][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.07529334723949432, 'train/grad_norm_pre_clip': 7.238701343536377, 'timing/data': 0.006359175022225827, 'timing/model': 1.6870863149524666, 'learning_rate/qwen_vl_interface': 1.003980167014757e-06, 'learning_rate/action_model': 1.003980167014757e-05, 'epoch': 1.84}) +[2026-06-16 06:44:12,460][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.05652712285518646, 'train/grad_norm_pre_clip': 11.101943969726562, 'timing/data': 0.007016252027824521, 'timing/model': 1.459715071017854, 'learning_rate/qwen_vl_interface': 1.0026158729978702e-06, 'learning_rate/action_model': 1.0026158729978702e-05, 'epoch': 1.84}) +[2026-06-16 06:44:18,465][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.10134614259004593, 'train/grad_norm_pre_clip': 5.992616653442383, 'timing/data': 0.0075604330049827695, 'timing/model': 1.467100856010802, 'learning_rate/qwen_vl_interface': 1.0012533249201604e-06, 'learning_rate/action_model': 1.0012533249201602e-05, 'epoch': 1.84}) +[2026-06-16 06:44:24,777][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.2262469083070755, 'train/grad_norm_pre_clip': 18.852764129638672, 'timing/data': 0.010039750020951033, 'timing/model': 1.5498644200270064, 'learning_rate/qwen_vl_interface': 9.99892523341719e-07, 'learning_rate/action_model': 9.99892523341719e-06, 'epoch': 1.84}) +[2026-06-16 06:44:30,848][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.03255210444331169, 'train/grad_norm_pre_clip': 3.7776851654052734, 'timing/data': 0.006606249022297561, 'timing/model': 1.5437547919573262, 'learning_rate/qwen_vl_interface': 9.98533468821922e-07, 'learning_rate/action_model': 9.98533468821922e-06, 'epoch': 1.84}) +[2026-06-16 06:44:36,937][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.015958571806550026, 'train/grad_norm_pre_clip': 6.695594787597656, 'timing/data': 0.006597515020985156, 'timing/model': 1.504603313980624, 'learning_rate/qwen_vl_interface': 9.971761619194231e-07, 'learning_rate/action_model': 9.97176161919423e-06, 'epoch': 1.84}) +[2026-06-16 06:44:42,895][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 0.04857976362109184, 'train/grad_norm_pre_clip': 9.148526191711426, 'timing/data': 0.006170227017719299, 'timing/model': 1.6993465349660255, 'learning_rate/qwen_vl_interface': 9.958206031921606e-07, 'learning_rate/action_model': 9.958206031921604e-06, 'epoch': 1.84}) +[2026-06-16 06:44:48,742][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.028431467711925507, 'train/grad_norm_pre_clip': 3.5783004760742188, 'timing/data': 0.006284294999204576, 'timing/model': 1.5435033589601517, 'learning_rate/qwen_vl_interface': 9.944667931973537e-07, 'learning_rate/action_model': 9.944667931973536e-06, 'epoch': 1.84}) +[2026-06-16 06:44:54,659][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.12175048142671585, 'train/grad_norm_pre_clip': 9.687236785888672, 'timing/data': 0.00818349898327142, 'timing/model': 1.4475336939794943, 'learning_rate/qwen_vl_interface': 9.93114732491503e-07, 'learning_rate/action_model': 9.93114732491503e-06, 'epoch': 1.84}) +[2026-06-16 06:45:00,555][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.12235257029533386, 'train/grad_norm_pre_clip': 13.626891136169434, 'timing/data': 0.006596974970307201, 'timing/model': 1.4595059069688432, 'learning_rate/qwen_vl_interface': 9.91764421630389e-07, 'learning_rate/action_model': 9.91764421630389e-06, 'epoch': 1.84}) +[2026-06-16 06:45:06,456][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.18418802320957184, 'train/grad_norm_pre_clip': 11.151863098144531, 'timing/data': 0.006054105993825942, 'timing/model': 1.4599878140143119, 'learning_rate/qwen_vl_interface': 9.904158611690732e-07, 'learning_rate/action_model': 9.90415861169073e-06, 'epoch': 1.84}) +[2026-06-16 06:45:12,453][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.09888448566198349, 'train/grad_norm_pre_clip': 7.9494781494140625, 'timing/data': 0.006529666017740965, 'timing/model': 1.5059790919767693, 'learning_rate/qwen_vl_interface': 9.890690516618987e-07, 'learning_rate/action_model': 9.890690516618985e-06, 'epoch': 1.84}) +[2026-06-16 06:45:18,645][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.06963910907506943, 'train/grad_norm_pre_clip': 5.1232123374938965, 'timing/data': 0.007674330030567944, 'timing/model': 1.7585020040278323, 'learning_rate/qwen_vl_interface': 9.877239936624875e-07, 'learning_rate/action_model': 9.877239936624875e-06, 'epoch': 1.84}) +[2026-06-16 06:45:24,506][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.0533759780228138, 'train/grad_norm_pre_clip': 5.453879356384277, 'timing/data': 0.006878989981487393, 'timing/model': 1.6138348600361496, 'learning_rate/qwen_vl_interface': 9.863806877237421e-07, 'learning_rate/action_model': 9.863806877237422e-06, 'epoch': 1.85}) +[2026-06-16 06:45:30,157][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.06367131322622299, 'train/grad_norm_pre_clip': 7.608201503753662, 'timing/data': 0.006359997030813247, 'timing/model': 1.45713886001613, 'learning_rate/qwen_vl_interface': 9.850391343978454e-07, 'learning_rate/action_model': 9.850391343978454e-06, 'epoch': 1.85}) +[2026-06-16 06:45:36,056][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.003774581477046013, 'train/grad_norm_pre_clip': 0.5374689102172852, 'timing/data': 0.006427895976230502, 'timing/model': 1.4579396869521588, 'learning_rate/qwen_vl_interface': 9.836993342362596e-07, 'learning_rate/action_model': 9.836993342362596e-06, 'epoch': 1.85}) +[2026-06-16 06:45:42,064][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.015908971428871155, 'train/grad_norm_pre_clip': 6.601849555969238, 'timing/data': 0.006503658019937575, 'timing/model': 1.4655396570451558, 'learning_rate/qwen_vl_interface': 9.823612877897246e-07, 'learning_rate/action_model': 9.823612877897245e-06, 'epoch': 1.85}) +[2026-06-16 06:45:48,061][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.08880975097417831, 'train/grad_norm_pre_clip': 6.5800299644470215, 'timing/data': 0.007402870978694409, 'timing/model': 1.4878887150553055, 'learning_rate/qwen_vl_interface': 9.81024995608261e-07, 'learning_rate/action_model': 9.810249956082609e-06, 'epoch': 1.85}) +[2026-06-16 06:45:54,023][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.0017185123870149255, 'train/grad_norm_pre_clip': 0.20811371505260468, 'timing/data': 0.006306959025096148, 'timing/model': 1.7306241070036776, 'learning_rate/qwen_vl_interface': 9.796904582411685e-07, 'learning_rate/action_model': 9.796904582411684e-06, 'epoch': 1.85}) +[2026-06-16 06:46:00,106][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.003252696245908737, 'train/grad_norm_pre_clip': 0.36809682846069336, 'timing/data': 0.006691865040920675, 'timing/model': 1.8002682600053959, 'learning_rate/qwen_vl_interface': 9.78357676237025e-07, 'learning_rate/action_model': 9.78357676237025e-06, 'epoch': 1.85}) +[2026-06-16 06:46:05,763][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 0.2505936622619629, 'train/grad_norm_pre_clip': 14.56890869140625, 'timing/data': 0.006795366993173957, 'timing/model': 1.469804208027199, 'learning_rate/qwen_vl_interface': 9.77026650143687e-07, 'learning_rate/action_model': 9.77026650143687e-06, 'epoch': 1.85}) +[2026-06-16 06:46:11,856][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.025954730808734894, 'train/grad_norm_pre_clip': 3.4815542697906494, 'timing/data': 0.006583093956578523, 'timing/model': 1.454361287993379, 'learning_rate/qwen_vl_interface': 9.756973805082882e-07, 'learning_rate/action_model': 9.756973805082883e-06, 'epoch': 1.85}) +[2026-06-16 06:46:17,944][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.038649335503578186, 'train/grad_norm_pre_clip': 7.6099653244018555, 'timing/data': 0.007643483986612409, 'timing/model': 1.634953380038496, 'learning_rate/qwen_vl_interface': 9.743698678772434e-07, 'learning_rate/action_model': 9.743698678772433e-06, 'epoch': 1.85}) +[2026-06-16 06:46:23,976][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.02321256324648857, 'train/grad_norm_pre_clip': 1.854185700416565, 'timing/data': 0.00697655399562791, 'timing/model': 1.5142946560517885, 'learning_rate/qwen_vl_interface': 9.730441127962395e-07, 'learning_rate/action_model': 9.730441127962395e-06, 'epoch': 1.85}) +[2026-06-16 06:46:29,971][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.06962496787309647, 'train/grad_norm_pre_clip': 6.092844486236572, 'timing/data': 0.018833626993000507, 'timing/model': 1.7543534400174394, 'learning_rate/qwen_vl_interface': 9.717201158102465e-07, 'learning_rate/action_model': 9.717201158102466e-06, 'epoch': 1.85}) +[2026-06-16 06:55:25,153][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.04756157472729683, 'train/grad_norm_pre_clip': 8.619868278503418, 'eval/loss': 0.047702230513095856, 'eval/flap_precision': 0.9416498993963782, 'eval/flap_recall': 0.9378757515030061, 'eval/flap_f1': 0.9397590361445783, 'eval/flap_rate_model': 0.13805555555555554, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9397590361445783, 'timing/data': 0.018141397973522544, 'timing/model': 1.6936727309948765, 'learning_rate/qwen_vl_interface': 9.703978774635094e-07, 'learning_rate/action_model': 9.703978774635095e-06, 'epoch': 1.85}) +[2026-06-16 06:55:25,154][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4300_state +[2026-06-16 06:55:25,158][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 06:55:30,510][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4300_state/model.safetensors +[2026-06-16 06:55:43,695][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4300_state/optimizer.bin +[2026-06-16 06:55:43,696][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4300_state/sampler.bin +[2026-06-16 06:55:43,696][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4300_state/sampler_1.bin +[2026-06-16 06:55:43,697][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4300_state/random_states_0.pkl +[2026-06-16 06:55:44,031][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4300 +[2026-06-16 06:55:44,032][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 06:55:44,042][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 06:55:50,020][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.025638364255428314, 'train/grad_norm_pre_clip': 4.274115085601807, 'timing/data': 0.006388948007952422, 'timing/model': 1.7430266770534217, 'learning_rate/qwen_vl_interface': 9.6907739829955e-07, 'learning_rate/action_model': 9.6907739829955e-06, 'epoch': 1.85}) +[2026-06-16 06:55:55,859][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.1204630583524704, 'train/grad_norm_pre_clip': 11.134969711303711, 'timing/data': 0.0067799839889630675, 'timing/model': 1.580407579021994, 'learning_rate/qwen_vl_interface': 9.677586788611681e-07, 'learning_rate/action_model': 9.677586788611681e-06, 'epoch': 1.85}) +[2026-06-16 06:56:01,546][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.03490150347352028, 'train/grad_norm_pre_clip': 4.947842597961426, 'timing/data': 0.006720186036545783, 'timing/model': 1.460140596027486, 'learning_rate/qwen_vl_interface': 9.664417196904382e-07, 'learning_rate/action_model': 9.664417196904382e-06, 'epoch': 1.85}) +[2026-06-16 06:56:07,548][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 0.00403389660641551, 'train/grad_norm_pre_clip': 0.7292549014091492, 'timing/data': 0.006780985975638032, 'timing/model': 1.461363194976002, 'learning_rate/qwen_vl_interface': 9.651265213287134e-07, 'learning_rate/action_model': 9.651265213287134e-06, 'epoch': 1.85}) +[2026-06-16 06:56:13,831][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.08005699515342712, 'train/grad_norm_pre_clip': 8.456076622009277, 'timing/data': 0.010936879029031843, 'timing/model': 1.5635940850479528, 'learning_rate/qwen_vl_interface': 9.638130843166218e-07, 'learning_rate/action_model': 9.638130843166219e-06, 'epoch': 1.85}) +[2026-06-16 06:56:20,162][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.018886994570493698, 'train/grad_norm_pre_clip': 3.5872199535369873, 'timing/data': 0.007040917000267655, 'timing/model': 1.567414884979371, 'learning_rate/qwen_vl_interface': 9.62501409194068e-07, 'learning_rate/action_model': 9.625014091940678e-06, 'epoch': 1.85}) +[2026-06-16 06:56:26,225][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.053703512996435165, 'train/grad_norm_pre_clip': 5.8036699295043945, 'timing/data': 0.006862794980406761, 'timing/model': 1.536977038020268, 'learning_rate/qwen_vl_interface': 9.611914965002317e-07, 'learning_rate/action_model': 9.611914965002316e-06, 'epoch': 1.85}) +[2026-06-16 06:56:32,427][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.017892559990286827, 'train/grad_norm_pre_clip': 2.0949933528900146, 'timing/data': 0.006829827034380287, 'timing/model': 1.7593184919678606, 'learning_rate/qwen_vl_interface': 9.598833467735686e-07, 'learning_rate/action_model': 9.598833467735685e-06, 'epoch': 1.85}) +[2026-06-16 06:56:38,152][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.1727878898382187, 'train/grad_norm_pre_clip': 26.420751571655273, 'timing/data': 0.007095557986758649, 'timing/model': 1.4605120210326277, 'learning_rate/qwen_vl_interface': 9.585769605518097e-07, 'learning_rate/action_model': 9.585769605518097e-06, 'epoch': 1.85}) +[2026-06-16 06:56:44,048][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.015606685541570187, 'train/grad_norm_pre_clip': 2.6430294513702393, 'timing/data': 0.006716491014230996, 'timing/model': 1.4516223939717747, 'learning_rate/qwen_vl_interface': 9.572723383719611e-07, 'learning_rate/action_model': 9.572723383719612e-06, 'epoch': 1.85}) +[2026-06-16 06:56:50,031][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.04838549718260765, 'train/grad_norm_pre_clip': 5.695097923278809, 'timing/data': 0.006675810960587114, 'timing/model': 1.5319078200263903, 'learning_rate/qwen_vl_interface': 9.559694807703041e-07, 'learning_rate/action_model': 9.55969480770304e-06, 'epoch': 1.85}) +[2026-06-16 06:56:56,124][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.07626175880432129, 'train/grad_norm_pre_clip': 7.682483673095703, 'timing/data': 0.006624255038332194, 'timing/model': 1.5250120259588584, 'learning_rate/qwen_vl_interface': 9.546683882823945e-07, 'learning_rate/action_model': 9.546683882823945e-06, 'epoch': 1.86}) +[2026-06-16 06:57:02,252][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.16466206312179565, 'train/grad_norm_pre_clip': 15.368701934814453, 'timing/data': 0.007507413974963129, 'timing/model': 1.5079143139882945, 'learning_rate/qwen_vl_interface': 9.533690614430608e-07, 'learning_rate/action_model': 9.533690614430608e-06, 'epoch': 1.86}) +[2026-06-16 06:57:08,308][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.11139436811208725, 'train/grad_norm_pre_clip': 6.384092330932617, 'timing/data': 0.006724411970935762, 'timing/model': 1.7116387090063654, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 1.86}) +[2026-06-16 06:57:14,135][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.0064411163330078125, 'train/grad_norm_pre_clip': 1.0621750354766846, 'timing/data': 0.006800374016165733, 'timing/model': 1.5298684889567085, 'learning_rate/qwen_vl_interface': 9.50775706845814e-07, 'learning_rate/action_model': 9.50775706845814e-06, 'epoch': 1.86}) +[2026-06-16 06:57:20,213][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.01716896891593933, 'train/grad_norm_pre_clip': 2.7824814319610596, 'timing/data': 0.00743486697319895, 'timing/model': 1.5109607439953834, 'learning_rate/qwen_vl_interface': 9.49481680153931e-07, 'learning_rate/action_model': 9.494816801539309e-06, 'epoch': 1.86}) +[2026-06-16 06:57:26,212][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.17144379019737244, 'train/grad_norm_pre_clip': 18.428388595581055, 'timing/data': 0.006710141024086624, 'timing/model': 1.5121601389837451, 'learning_rate/qwen_vl_interface': 9.481894212426848e-07, 'learning_rate/action_model': 9.481894212426847e-06, 'epoch': 1.86}) +[2026-06-16 06:57:32,160][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.18948917090892792, 'train/grad_norm_pre_clip': 9.377558708190918, 'timing/data': 0.006532519997563213, 'timing/model': 1.4497347139986232, 'learning_rate/qwen_vl_interface': 9.468989306432738e-07, 'learning_rate/action_model': 9.468989306432738e-06, 'epoch': 1.86}) +[2026-06-16 06:57:38,161][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.02537783421576023, 'train/grad_norm_pre_clip': 5.830069065093994, 'timing/data': 0.006738672964274883, 'timing/model': 1.4683895769994706, 'learning_rate/qwen_vl_interface': 9.456102088861702e-07, 'learning_rate/action_model': 9.456102088861701e-06, 'epoch': 1.86}) +[2026-06-16 06:57:44,197][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.014705441892147064, 'train/grad_norm_pre_clip': 1.6444894075393677, 'timing/data': 0.0068379289587028325, 'timing/model': 1.7963626540149562, 'learning_rate/qwen_vl_interface': 9.443232565011188e-07, 'learning_rate/action_model': 9.443232565011187e-06, 'epoch': 1.86}) +[2026-06-16 06:57:49,861][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.056613314896821976, 'train/grad_norm_pre_clip': 14.166733741760254, 'timing/data': 0.007395719992928207, 'timing/model': 1.464174305961933, 'learning_rate/qwen_vl_interface': 9.430380740171369e-07, 'learning_rate/action_model': 9.430380740171368e-06, 'epoch': 1.86}) +[2026-06-16 06:57:55,856][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.11748700588941574, 'train/grad_norm_pre_clip': 9.608538627624512, 'timing/data': 0.006659957987722009, 'timing/model': 1.4582417490310036, 'learning_rate/qwen_vl_interface': 9.417546619625143e-07, 'learning_rate/action_model': 9.417546619625143e-06, 'epoch': 1.86}) +[2026-06-16 06:58:01,757][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.028946302831172943, 'train/grad_norm_pre_clip': 5.459405422210693, 'timing/data': 0.006829125981312245, 'timing/model': 1.459299874957651, 'learning_rate/qwen_vl_interface': 9.404730208648144e-07, 'learning_rate/action_model': 9.404730208648144e-06, 'epoch': 1.86}) +[2026-06-16 06:58:07,655][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.002997680101543665, 'train/grad_norm_pre_clip': 1.0402605533599854, 'timing/data': 0.006767264974769205, 'timing/model': 1.4601094259996898, 'learning_rate/qwen_vl_interface': 9.3919315125087e-07, 'learning_rate/action_model': 9.3919315125087e-06, 'epoch': 1.86}) +[2026-06-16 06:58:13,642][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.22825075685977936, 'train/grad_norm_pre_clip': 8.733309745788574, 'timing/data': 0.006612506986130029, 'timing/model': 1.5052216289914213, 'learning_rate/qwen_vl_interface': 9.379150536467879e-07, 'learning_rate/action_model': 9.379150536467878e-06, 'epoch': 1.86}) +[2026-06-16 06:58:19,598][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.20909391343593597, 'train/grad_norm_pre_clip': 11.015958786010742, 'timing/data': 0.006718532997183502, 'timing/model': 1.7001302560092881, 'learning_rate/qwen_vl_interface': 9.366387285779453e-07, 'learning_rate/action_model': 9.366387285779453e-06, 'epoch': 1.86}) +[2026-06-16 06:58:25,256][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.06431952118873596, 'train/grad_norm_pre_clip': 6.694969177246094, 'timing/data': 0.0066173350205644965, 'timing/model': 1.44754483201541, 'learning_rate/qwen_vl_interface': 9.353641765689915e-07, 'learning_rate/action_model': 9.353641765689915e-06, 'epoch': 1.86}) +[2026-06-16 06:58:31,158][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.05994017794728279, 'train/grad_norm_pre_clip': 3.9361391067504883, 'timing/data': 0.006629366020206362, 'timing/model': 1.4608541009947658, 'learning_rate/qwen_vl_interface': 9.340913981438462e-07, 'learning_rate/action_model': 9.340913981438463e-06, 'epoch': 1.86}) +[2026-06-16 06:58:37,064][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.06279713660478592, 'train/grad_norm_pre_clip': 10.387701988220215, 'timing/data': 0.006537790992297232, 'timing/model': 1.4677516429801472, 'learning_rate/qwen_vl_interface': 9.328203938257019e-07, 'learning_rate/action_model': 9.328203938257018e-06, 'epoch': 1.86}) +[2026-06-16 06:58:42,951][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.090119369328022, 'train/grad_norm_pre_clip': 4.316306114196777, 'timing/data': 0.006744205020368099, 'timing/model': 1.4579718429595232, 'learning_rate/qwen_vl_interface': 9.315511641370191e-07, 'learning_rate/action_model': 9.315511641370191e-06, 'epoch': 1.86}) +[2026-06-16 06:58:48,858][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.0370604433119297, 'train/grad_norm_pre_clip': 9.051858901977539, 'timing/data': 0.006742232013493776, 'timing/model': 1.4313661910127848, 'learning_rate/qwen_vl_interface': 9.302837095995306e-07, 'learning_rate/action_model': 9.302837095995306e-06, 'epoch': 1.86}) +[2026-06-16 06:58:54,788][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.09281167387962341, 'train/grad_norm_pre_clip': 6.587157726287842, 'timing/data': 0.006557781016454101, 'timing/model': 1.6925694629899226, 'learning_rate/qwen_vl_interface': 9.290180307342392e-07, 'learning_rate/action_model': 9.290180307342391e-06, 'epoch': 1.86}) +[2026-06-16 06:59:00,456][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 0.03091840073466301, 'train/grad_norm_pre_clip': 6.510698318481445, 'timing/data': 0.006798073998652399, 'timing/model': 1.4581874660216272, 'learning_rate/qwen_vl_interface': 9.277541280614183e-07, 'learning_rate/action_model': 9.277541280614183e-06, 'epoch': 1.86}) +[2026-06-16 06:59:06,355][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.11127250641584396, 'train/grad_norm_pre_clip': 13.269061088562012, 'timing/data': 0.006250638980418444, 'timing/model': 1.4444952189805917, 'learning_rate/qwen_vl_interface': 9.264920021006093e-07, 'learning_rate/action_model': 9.264920021006093e-06, 'epoch': 1.86}) +[2026-06-16 06:59:12,252][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.009210566058754921, 'train/grad_norm_pre_clip': 2.1122348308563232, 'timing/data': 0.0065137160127051175, 'timing/model': 1.4581637260271236, 'learning_rate/qwen_vl_interface': 9.252316533706246e-07, 'learning_rate/action_model': 9.252316533706246e-06, 'epoch': 1.87}) +[2026-06-16 06:59:18,153][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.19472919404506683, 'train/grad_norm_pre_clip': 10.904276847839355, 'timing/data': 0.00660976703511551, 'timing/model': 1.4549201849731617, 'learning_rate/qwen_vl_interface': 9.239730823895469e-07, 'learning_rate/action_model': 9.23973082389547e-06, 'epoch': 1.87}) +[2026-06-16 06:59:24,057][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.23095940053462982, 'train/grad_norm_pre_clip': 14.440889358520508, 'timing/data': 0.006388710986357182, 'timing/model': 1.432073144998867, 'learning_rate/qwen_vl_interface': 9.227162896747264e-07, 'learning_rate/action_model': 9.227162896747264e-06, 'epoch': 1.87}) +[2026-06-16 06:59:29,991][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.009196387603878975, 'train/grad_norm_pre_clip': 2.394465446472168, 'timing/data': 0.006779065995942801, 'timing/model': 1.6909751140046865, 'learning_rate/qwen_vl_interface': 9.214612757427833e-07, 'learning_rate/action_model': 9.214612757427833e-06, 'epoch': 1.87}) +[2026-06-16 06:59:35,658][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.12985031306743622, 'train/grad_norm_pre_clip': 16.43012237548828, 'timing/data': 0.0073174479766748846, 'timing/model': 1.4591372089926153, 'learning_rate/qwen_vl_interface': 9.202080411096071e-07, 'learning_rate/action_model': 9.20208041109607e-06, 'epoch': 1.87}) +[2026-06-16 06:59:41,662][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.12927591800689697, 'train/grad_norm_pre_clip': 10.459579467773438, 'timing/data': 0.006656965997535735, 'timing/model': 1.4636647540028207, 'learning_rate/qwen_vl_interface': 9.189565862903543e-07, 'learning_rate/action_model': 9.189565862903543e-06, 'epoch': 1.87}) +[2026-06-16 06:59:47,563][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.042096350342035294, 'train/grad_norm_pre_clip': 5.576199531555176, 'timing/data': 0.006452806002926081, 'timing/model': 1.4649114210042171, 'learning_rate/qwen_vl_interface': 9.17706911799451e-07, 'learning_rate/action_model': 9.17706911799451e-06, 'epoch': 1.87}) +[2026-06-16 06:59:53,457][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.14364488422870636, 'train/grad_norm_pre_clip': 12.892755508422852, 'timing/data': 0.006593221973162144, 'timing/model': 1.4548178399563767, 'learning_rate/qwen_vl_interface': 9.164590181505912e-07, 'learning_rate/action_model': 9.164590181505911e-06, 'epoch': 1.87}) +[2026-06-16 06:59:59,442][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.01816733181476593, 'train/grad_norm_pre_clip': 3.1182777881622314, 'timing/data': 0.00727463501971215, 'timing/model': 1.5054589349892922, 'learning_rate/qwen_vl_interface': 9.15212905856738e-07, 'learning_rate/action_model': 9.15212905856738e-06, 'epoch': 1.87}) +[2026-06-16 07:00:05,385][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.012752815149724483, 'train/grad_norm_pre_clip': 4.766163349151611, 'timing/data': 0.006511872983537614, 'timing/model': 1.6846263809711672, 'learning_rate/qwen_vl_interface': 9.139685754301189e-07, 'learning_rate/action_model': 9.139685754301189e-06, 'epoch': 1.87}) +[2026-06-16 07:00:11,206][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.18798844516277313, 'train/grad_norm_pre_clip': 11.39513874053955, 'timing/data': 0.0067489020293578506, 'timing/model': 1.6042916770093143, 'learning_rate/qwen_vl_interface': 9.127260273822324e-07, 'learning_rate/action_model': 9.127260273822325e-06, 'epoch': 1.87}) +[2026-06-16 07:00:16,853][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 0.06692563742399216, 'train/grad_norm_pre_clip': 5.704127311706543, 'timing/data': 0.006708020984660834, 'timing/model': 1.4459426029934548, 'learning_rate/qwen_vl_interface': 9.114852622238435e-07, 'learning_rate/action_model': 9.114852622238435e-06, 'epoch': 1.87}) +[2026-06-16 07:00:22,763][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.02828720025718212, 'train/grad_norm_pre_clip': 3.343841075897217, 'timing/data': 0.007562999031506479, 'timing/model': 1.4676576100173406, 'learning_rate/qwen_vl_interface': 9.102462804649827e-07, 'learning_rate/action_model': 9.102462804649827e-06, 'epoch': 1.87}) +[2026-06-16 07:00:28,661][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.25842300057411194, 'train/grad_norm_pre_clip': 13.002115249633789, 'timing/data': 0.007946163008455187, 'timing/model': 1.4621186329750344, 'learning_rate/qwen_vl_interface': 9.090090826149489e-07, 'learning_rate/action_model': 9.09009082614949e-06, 'epoch': 1.87}) +[2026-06-16 07:00:34,557][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.060679856687784195, 'train/grad_norm_pre_clip': 5.7197418212890625, 'timing/data': 0.006584398972336203, 'timing/model': 1.4346314980066381, 'learning_rate/qwen_vl_interface': 9.077736691823081e-07, 'learning_rate/action_model': 9.07773669182308e-06, 'epoch': 1.87}) +[2026-06-16 07:00:40,488][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.10318046808242798, 'train/grad_norm_pre_clip': 12.528339385986328, 'timing/data': 0.006807518017012626, 'timing/model': 1.6923492699861526, 'learning_rate/qwen_vl_interface': 9.065400406748912e-07, 'learning_rate/action_model': 9.06540040674891e-06, 'epoch': 1.87}) +[2026-06-16 07:00:46,389][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.10570446401834488, 'train/grad_norm_pre_clip': 10.652177810668945, 'timing/data': 0.007047449995297939, 'timing/model': 1.690623993985355, 'learning_rate/qwen_vl_interface': 9.053081975997969e-07, 'learning_rate/action_model': 9.053081975997968e-06, 'epoch': 1.87}) +[2026-06-16 07:00:52,059][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.1011190414428711, 'train/grad_norm_pre_clip': 11.933952331542969, 'timing/data': 0.006560464040376246, 'timing/model': 1.4613911269698292, 'learning_rate/qwen_vl_interface': 9.040781404633894e-07, 'learning_rate/action_model': 9.040781404633895e-06, 'epoch': 1.87}) +[2026-06-16 07:00:57,959][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 0.28992435336112976, 'train/grad_norm_pre_clip': 20.210487365722656, 'timing/data': 0.006671597016975284, 'timing/model': 1.4598975530243479, 'learning_rate/qwen_vl_interface': 9.028498697712982e-07, 'learning_rate/action_model': 9.028498697712981e-06, 'epoch': 1.87}) +[2026-06-16 07:01:03,853][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.25332996249198914, 'train/grad_norm_pre_clip': 18.962738037109375, 'timing/data': 0.006520455004647374, 'timing/model': 1.4557386470260099, 'learning_rate/qwen_vl_interface': 9.016233860284195e-07, 'learning_rate/action_model': 9.016233860284195e-06, 'epoch': 1.87}) +[2026-06-16 07:01:09,774][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.0211019329726696, 'train/grad_norm_pre_clip': 2.9669289588928223, 'timing/data': 0.006637686979956925, 'timing/model': 1.4461426690104418, 'learning_rate/qwen_vl_interface': 9.003986897389142e-07, 'learning_rate/action_model': 9.003986897389141e-06, 'epoch': 1.87}) +[2026-06-16 07:01:15,683][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.014762242324650288, 'train/grad_norm_pre_clip': 2.489903450012207, 'timing/data': 0.0065102800144813955, 'timing/model': 1.6835563680506311, 'learning_rate/qwen_vl_interface': 8.991757814062094e-07, 'learning_rate/action_model': 8.991757814062093e-06, 'epoch': 1.87}) +[2026-06-16 07:01:21,534][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.0025127672124654055, 'train/grad_norm_pre_clip': 0.7642930746078491, 'timing/data': 0.006776101014111191, 'timing/model': 1.6374122700071894, 'learning_rate/qwen_vl_interface': 8.979546615329954e-07, 'learning_rate/action_model': 8.979546615329953e-06, 'epoch': 1.87}) +[2026-06-16 07:01:27,254][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.005263470113277435, 'train/grad_norm_pre_clip': 0.9479976296424866, 'timing/data': 0.006686879962217063, 'timing/model': 1.458810493000783, 'learning_rate/qwen_vl_interface': 8.967353306212286e-07, 'learning_rate/action_model': 8.967353306212286e-06, 'epoch': 1.88}) +[2026-06-16 07:01:33,154][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.14620524644851685, 'train/grad_norm_pre_clip': 9.613425254821777, 'timing/data': 0.0064509829971939325, 'timing/model': 1.4604221089975908, 'learning_rate/qwen_vl_interface': 8.955177891721303e-07, 'learning_rate/action_model': 8.955177891721303e-06, 'epoch': 1.88}) +[2026-06-16 07:01:39,053][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.026601625606417656, 'train/grad_norm_pre_clip': 11.774293899536133, 'timing/data': 0.006760898977518082, 'timing/model': 1.4587028929963708, 'learning_rate/qwen_vl_interface': 8.943020376861858e-07, 'learning_rate/action_model': 8.943020376861857e-06, 'epoch': 1.88}) +[2026-06-16 07:01:44,961][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.17457646131515503, 'train/grad_norm_pre_clip': 14.821203231811523, 'timing/data': 0.006933952972758561, 'timing/model': 1.4340055850334466, 'learning_rate/qwen_vl_interface': 8.930880766631444e-07, 'learning_rate/action_model': 8.930880766631445e-06, 'epoch': 1.88}) +[2026-06-16 07:01:50,857][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.0774679034948349, 'train/grad_norm_pre_clip': 7.058228969573975, 'timing/data': 0.006430551991797984, 'timing/model': 1.432362501975149, 'learning_rate/qwen_vl_interface': 8.918759066020201e-07, 'learning_rate/action_model': 8.9187590660202e-06, 'epoch': 1.88}) +[2026-06-16 07:01:56,745][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 0.012574350461363792, 'train/grad_norm_pre_clip': 2.0290169715881348, 'timing/data': 0.00656505097867921, 'timing/model': 1.646794967004098, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 1.88}) +[2026-06-16 07:02:02,456][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.07474943995475769, 'train/grad_norm_pre_clip': 9.09377670288086, 'timing/data': 0.006720508972648531, 'timing/model': 1.4586770620080642, 'learning_rate/qwen_vl_interface': 8.894569413578938e-07, 'learning_rate/action_model': 8.894569413578937e-06, 'epoch': 1.88}) +[2026-06-16 07:02:08,357][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.11934106051921844, 'train/grad_norm_pre_clip': 9.851412773132324, 'timing/data': 0.006981373007874936, 'timing/model': 1.4583067210041918, 'learning_rate/qwen_vl_interface': 8.882501471692379e-07, 'learning_rate/action_model': 8.882501471692378e-06, 'epoch': 1.88}) +[2026-06-16 07:02:14,339][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.02195947989821434, 'train/grad_norm_pre_clip': 4.53139066696167, 'timing/data': 0.006543509021867067, 'timing/model': 1.5063430199516006, 'learning_rate/qwen_vl_interface': 8.870451459311894e-07, 'learning_rate/action_model': 8.870451459311893e-06, 'epoch': 1.88}) +[2026-06-16 07:02:20,340][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.13277578353881836, 'train/grad_norm_pre_clip': 13.140645027160645, 'timing/data': 0.007062361983116716, 'timing/model': 1.5076994289993308, 'learning_rate/qwen_vl_interface': 8.85841938139077e-07, 'learning_rate/action_model': 8.85841938139077e-06, 'epoch': 1.88}) +[2026-06-16 07:02:26,364][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.0305336806923151, 'train/grad_norm_pre_clip': 7.437119483947754, 'timing/data': 0.006670996022876352, 'timing/model': 1.7634288279805332, 'learning_rate/qwen_vl_interface': 8.846405242874955e-07, 'learning_rate/action_model': 8.846405242874954e-06, 'epoch': 1.88}) +[2026-06-16 07:02:32,058][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.0033436762169003487, 'train/grad_norm_pre_clip': 0.3462909162044525, 'timing/data': 0.0063129490008577704, 'timing/model': 1.459098024002742, 'learning_rate/qwen_vl_interface': 8.834409048703004e-07, 'learning_rate/action_model': 8.834409048703005e-06, 'epoch': 1.88}) +[2026-06-16 07:02:38,051][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.03784462437033653, 'train/grad_norm_pre_clip': 7.638352870941162, 'timing/data': 0.006446474988479167, 'timing/model': 1.5113060110015795, 'learning_rate/qwen_vl_interface': 8.822430803806098e-07, 'learning_rate/action_model': 8.822430803806097e-06, 'epoch': 1.88}) +[2026-06-16 07:02:43,857][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.04332651570439339, 'train/grad_norm_pre_clip': 9.709514617919922, 'timing/data': 0.006574144004844129, 'timing/model': 1.457706626970321, 'learning_rate/qwen_vl_interface': 8.810470513108041e-07, 'learning_rate/action_model': 8.81047051310804e-06, 'epoch': 1.88}) +[2026-06-16 07:02:49,762][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 0.03485354781150818, 'train/grad_norm_pre_clip': 8.296609878540039, 'timing/data': 0.006680649996269494, 'timing/model': 1.4635457340045832, 'learning_rate/qwen_vl_interface': 8.798528181525262e-07, 'learning_rate/action_model': 8.798528181525261e-06, 'epoch': 1.88}) +[2026-06-16 07:02:55,757][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.22179606556892395, 'train/grad_norm_pre_clip': 8.476800918579102, 'timing/data': 0.006589034979697317, 'timing/model': 1.4821150320349261, 'learning_rate/qwen_vl_interface': 8.786603813966798e-07, 'learning_rate/action_model': 8.786603813966798e-06, 'epoch': 1.88}) +[2026-06-16 07:03:01,799][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 0.20620593428611755, 'train/grad_norm_pre_clip': 17.818994522094727, 'timing/data': 0.007132735045161098, 'timing/model': 1.7004302630084567, 'learning_rate/qwen_vl_interface': 8.774697415334304e-07, 'learning_rate/action_model': 8.774697415334304e-06, 'epoch': 1.88}) +[2026-06-16 07:03:07,562][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.21407583355903625, 'train/grad_norm_pre_clip': 21.45600700378418, 'timing/data': 0.006758373987395316, 'timing/model': 1.4654990729759447, 'learning_rate/qwen_vl_interface': 8.762808990522056e-07, 'learning_rate/action_model': 8.762808990522057e-06, 'epoch': 1.88}) +[2026-06-16 07:03:13,456][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.024716932326555252, 'train/grad_norm_pre_clip': 4.142848968505859, 'timing/data': 0.006517770001664758, 'timing/model': 1.4588441549567506, 'learning_rate/qwen_vl_interface': 8.750938544416934e-07, 'learning_rate/action_model': 8.750938544416934e-06, 'epoch': 1.88}) +[2026-06-16 07:03:19,357][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.049732618033885956, 'train/grad_norm_pre_clip': 12.224413871765137, 'timing/data': 0.006734990049153566, 'timing/model': 1.4574979299795814, 'learning_rate/qwen_vl_interface': 8.739086081898437e-07, 'learning_rate/action_model': 8.739086081898437e-06, 'epoch': 1.88}) +[2026-06-16 07:03:25,260][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 0.20990830659866333, 'train/grad_norm_pre_clip': 18.682706832885742, 'timing/data': 0.007154788007028401, 'timing/model': 1.4581196199869737, 'learning_rate/qwen_vl_interface': 8.727251607838647e-07, 'learning_rate/action_model': 8.727251607838647e-06, 'epoch': 1.88}) +[2026-06-16 07:03:31,172][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.1833527535200119, 'train/grad_norm_pre_clip': 17.16653823852539, 'timing/data': 0.006401989026926458, 'timing/model': 1.4869874819996767, 'learning_rate/qwen_vl_interface': 8.715435127102281e-07, 'learning_rate/action_model': 8.715435127102281e-06, 'epoch': 1.88}) +[2026-06-16 07:03:37,098][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.030351290479302406, 'train/grad_norm_pre_clip': 5.343417167663574, 'timing/data': 0.0064725439879111946, 'timing/model': 1.700243275030516, 'learning_rate/qwen_vl_interface': 8.703636644546651e-07, 'learning_rate/action_model': 8.703636644546651e-06, 'epoch': 1.88}) +[2026-06-16 07:03:42,757][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.025060631334781647, 'train/grad_norm_pre_clip': 1.947608470916748, 'timing/data': 0.006481456977780908, 'timing/model': 1.4591022520326078, 'learning_rate/qwen_vl_interface': 8.69185616502166e-07, 'learning_rate/action_model': 8.69185616502166e-06, 'epoch': 1.89}) +[2026-06-16 07:03:48,758][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.042139191180467606, 'train/grad_norm_pre_clip': 7.340802192687988, 'timing/data': 0.006544420029968023, 'timing/model': 1.4809104059822857, 'learning_rate/qwen_vl_interface': 8.680093693369821e-07, 'learning_rate/action_model': 8.68009369336982e-06, 'epoch': 1.89}) +[2026-06-16 07:03:54,764][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.06575118750333786, 'train/grad_norm_pre_clip': 10.420502662658691, 'timing/data': 0.0070754309999756515, 'timing/model': 1.5649461330031045, 'learning_rate/qwen_vl_interface': 8.66834923442624e-07, 'learning_rate/action_model': 8.66834923442624e-06, 'epoch': 1.89}) +[2026-06-16 07:04:00,771][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.20195749402046204, 'train/grad_norm_pre_clip': 20.97333335876465, 'timing/data': 0.006734579044859856, 'timing/model': 1.449959785968531, 'learning_rate/qwen_vl_interface': 8.65662279301863e-07, 'learning_rate/action_model': 8.656622793018628e-06, 'epoch': 1.89}) +[2026-06-16 07:04:06,660][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.10295987874269485, 'train/grad_norm_pre_clip': 10.322333335876465, 'timing/data': 0.006599249958526343, 'timing/model': 1.4614031200180762, 'learning_rate/qwen_vl_interface': 8.644914373967274e-07, 'learning_rate/action_model': 8.644914373967273e-06, 'epoch': 1.89}) +[2026-06-16 07:04:12,588][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 0.05091835558414459, 'train/grad_norm_pre_clip': 10.284038543701172, 'timing/data': 0.0067987440270371735, 'timing/model': 1.6902912700315937, 'learning_rate/qwen_vl_interface': 8.633223982085068e-07, 'learning_rate/action_model': 8.633223982085068e-06, 'epoch': 1.89}) +[2026-06-16 07:04:18,258][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.13584782183170319, 'train/grad_norm_pre_clip': 13.194443702697754, 'timing/data': 0.006554583960678428, 'timing/model': 1.4571160540217534, 'learning_rate/qwen_vl_interface': 8.621551622177499e-07, 'learning_rate/action_model': 8.621551622177499e-06, 'epoch': 1.89}) +[2026-06-16 07:04:24,158][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 0.044000789523124695, 'train/grad_norm_pre_clip': 8.980648040771484, 'timing/data': 0.006489017978310585, 'timing/model': 1.4591014999896288, 'learning_rate/qwen_vl_interface': 8.609897299042615e-07, 'learning_rate/action_model': 8.609897299042614e-06, 'epoch': 1.89}) +[2026-06-16 07:04:30,056][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.029224777594208717, 'train/grad_norm_pre_clip': 5.215259552001953, 'timing/data': 0.00707029300974682, 'timing/model': 1.4569884469965473, 'learning_rate/qwen_vl_interface': 8.598261017471088e-07, 'learning_rate/action_model': 8.598261017471088e-06, 'epoch': 1.89}) +[2026-06-16 07:04:35,956][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.18432970345020294, 'train/grad_norm_pre_clip': 19.483644485473633, 'timing/data': 0.006649604008998722, 'timing/model': 1.4386282490449958, 'learning_rate/qwen_vl_interface': 8.586642782246133e-07, 'learning_rate/action_model': 8.586642782246131e-06, 'epoch': 1.89}) +[2026-06-16 07:04:41,938][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.009117468260228634, 'train/grad_norm_pre_clip': 1.1943559646606445, 'timing/data': 0.006959991005714983, 'timing/model': 1.5085460999980569, 'learning_rate/qwen_vl_interface': 8.57504259814358e-07, 'learning_rate/action_model': 8.57504259814358e-06, 'epoch': 1.89}) +[2026-06-16 07:04:47,889][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.009693280793726444, 'train/grad_norm_pre_clip': 1.349137544631958, 'timing/data': 0.0067049150238744915, 'timing/model': 1.691248640010599, 'learning_rate/qwen_vl_interface': 8.563460469931823e-07, 'learning_rate/action_model': 8.563460469931822e-06, 'epoch': 1.89}) +[2026-06-16 07:04:53,555][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.10041200369596481, 'train/grad_norm_pre_clip': 12.312640190124512, 'timing/data': 0.0064924529870040715, 'timing/model': 1.4586956740240566, 'learning_rate/qwen_vl_interface': 8.551896402371834e-07, 'learning_rate/action_model': 8.551896402371834e-06, 'epoch': 1.89}) +[2026-06-16 07:04:59,462][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.02998725324869156, 'train/grad_norm_pre_clip': 3.5795445442199707, 'timing/data': 0.006899642001371831, 'timing/model': 1.467666277952958, 'learning_rate/qwen_vl_interface': 8.540350400217177e-07, 'learning_rate/action_model': 8.540350400217176e-06, 'epoch': 1.89}) +[2026-06-16 07:05:05,354][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 0.051109202206134796, 'train/grad_norm_pre_clip': 23.582290649414062, 'timing/data': 0.006780696974601597, 'timing/model': 1.457081650965847, 'learning_rate/qwen_vl_interface': 8.528822468213964e-07, 'learning_rate/action_model': 8.528822468213964e-06, 'epoch': 1.89}) +[2026-06-16 07:05:11,256][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.09309975057840347, 'train/grad_norm_pre_clip': 5.509460926055908, 'timing/data': 0.006730583030730486, 'timing/model': 1.463543385034427, 'learning_rate/qwen_vl_interface': 8.517312611100898e-07, 'learning_rate/action_model': 8.517312611100897e-06, 'epoch': 1.89}) +[2026-06-16 07:05:17,238][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.19674146175384521, 'train/grad_norm_pre_clip': 17.707304000854492, 'timing/data': 0.006823009985964745, 'timing/model': 1.5082071569631808, 'learning_rate/qwen_vl_interface': 8.505820833609249e-07, 'learning_rate/action_model': 8.50582083360925e-06, 'epoch': 1.89}) +[2026-06-16 07:05:23,184][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.04410594701766968, 'train/grad_norm_pre_clip': 6.339455604553223, 'timing/data': 0.006543027004227042, 'timing/model': 1.6853132900432684, 'learning_rate/qwen_vl_interface': 8.494347140462847e-07, 'learning_rate/action_model': 8.494347140462846e-06, 'epoch': 1.89}) +[2026-06-16 07:05:28,864][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.15474866330623627, 'train/grad_norm_pre_clip': 14.685680389404297, 'timing/data': 0.006749360007233918, 'timing/model': 1.4647103099850938, 'learning_rate/qwen_vl_interface': 8.482891536378105e-07, 'learning_rate/action_model': 8.482891536378105e-06, 'epoch': 1.89}) +[2026-06-16 07:14:31,965][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.022164760157465935, 'train/grad_norm_pre_clip': 2.6603472232818604, 'eval/loss': 0.04802977293729782, 'eval/flap_precision': 0.9436619718309859, 'eval/flap_recall': 0.9398797595190381, 'eval/flap_f1': 0.9417670682730924, 'eval/flap_rate_model': 0.13805555555555554, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9417670682730924, 'timing/data': 0.00662194297183305, 'timing/model': 1.4561430210014805, 'learning_rate/qwen_vl_interface': 8.471454026063972e-07, 'learning_rate/action_model': 8.47145402606397e-06, 'epoch': 1.89}) +[2026-06-16 07:14:31,966][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4400_state +[2026-06-16 07:14:31,971][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 07:14:37,447][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4400_state/model.safetensors +[2026-06-16 07:14:50,877][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4400_state/optimizer.bin +[2026-06-16 07:14:50,878][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4400_state/sampler.bin +[2026-06-16 07:14:50,878][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4400_state/sampler_1.bin +[2026-06-16 07:14:50,879][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4400_state/random_states_0.pkl +[2026-06-16 07:14:50,982][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4400 +[2026-06-16 07:14:50,982][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 07:14:50,992][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 07:14:56,757][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.10363348573446274, 'train/grad_norm_pre_clip': 21.32111358642578, 'timing/data': 0.007755800033919513, 'timing/model': 1.4716753020184115, 'learning_rate/qwen_vl_interface': 8.460034614221988e-07, 'learning_rate/action_model': 8.460034614221988e-06, 'epoch': 1.89}) +[2026-06-16 07:15:02,750][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.07658977061510086, 'train/grad_norm_pre_clip': 9.987048149108887, 'timing/data': 0.0067882470320910215, 'timing/model': 1.4601525089819916, 'learning_rate/qwen_vl_interface': 8.448633305546236e-07, 'learning_rate/action_model': 8.448633305546235e-06, 'epoch': 1.89}) +[2026-06-16 07:15:08,753][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.01200201828032732, 'train/grad_norm_pre_clip': 2.3230156898498535, 'timing/data': 0.006535686028655618, 'timing/model': 1.48103230597917, 'learning_rate/qwen_vl_interface': 8.437250104723369e-07, 'learning_rate/action_model': 8.437250104723369e-06, 'epoch': 1.89}) +[2026-06-16 07:15:14,985][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.04233288764953613, 'train/grad_norm_pre_clip': 18.275602340698242, 'timing/data': 0.006405373045708984, 'timing/model': 1.4933203620021231, 'learning_rate/qwen_vl_interface': 8.42588501643259e-07, 'learning_rate/action_model': 8.42588501643259e-06, 'epoch': 1.89}) +[2026-06-16 07:15:20,967][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.08014161139726639, 'train/grad_norm_pre_clip': 8.094876289367676, 'timing/data': 0.007190378033556044, 'timing/model': 1.4899426120100543, 'learning_rate/qwen_vl_interface': 8.414538045345655e-07, 'learning_rate/action_model': 8.414538045345654e-06, 'epoch': 1.9}) +[2026-06-16 07:15:27,049][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.06842090934515, 'train/grad_norm_pre_clip': 4.748346328735352, 'timing/data': 0.00674399099079892, 'timing/model': 1.7817126139998436, 'learning_rate/qwen_vl_interface': 8.403209196126871e-07, 'learning_rate/action_model': 8.403209196126871e-06, 'epoch': 1.9}) +[2026-06-16 07:15:33,124][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.08026392757892609, 'train/grad_norm_pre_clip': 9.14654541015625, 'timing/data': 0.006831891019828618, 'timing/model': 1.5482289910432883, 'learning_rate/qwen_vl_interface': 8.391898473433111e-07, 'learning_rate/action_model': 8.39189847343311e-06, 'epoch': 1.9}) +[2026-06-16 07:15:39,364][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.1571948230266571, 'train/grad_norm_pre_clip': 7.369835376739502, 'timing/data': 0.006522156007122248, 'timing/model': 1.6001467409660108, 'learning_rate/qwen_vl_interface': 8.380605881913778e-07, 'learning_rate/action_model': 8.380605881913779e-06, 'epoch': 1.9}) +[2026-06-16 07:15:45,627][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 0.0523202158510685, 'train/grad_norm_pre_clip': 7.889088153839111, 'timing/data': 0.006455107009969652, 'timing/model': 1.580206423997879, 'learning_rate/qwen_vl_interface': 8.369331426210832e-07, 'learning_rate/action_model': 8.369331426210832e-06, 'epoch': 1.9}) +[2026-06-16 07:15:51,663][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.08606769889593124, 'train/grad_norm_pre_clip': 11.426199913024902, 'timing/data': 0.006488184968475252, 'timing/model': 1.4391381919849664, 'learning_rate/qwen_vl_interface': 8.358075110958785e-07, 'learning_rate/action_model': 8.358075110958784e-06, 'epoch': 1.9}) +[2026-06-16 07:15:57,684][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.07066407054662704, 'train/grad_norm_pre_clip': 4.907467842102051, 'timing/data': 0.006660269049461931, 'timing/model': 1.7219794070115313, 'learning_rate/qwen_vl_interface': 8.346836940784661e-07, 'learning_rate/action_model': 8.346836940784661e-06, 'epoch': 1.9}) +[2026-06-16 07:16:03,751][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.1237873062491417, 'train/grad_norm_pre_clip': 10.855303764343262, 'timing/data': 0.0066786459647119045, 'timing/model': 1.6375972829991952, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 1.9}) +[2026-06-16 07:16:09,546][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.16749125719070435, 'train/grad_norm_pre_clip': 26.844579696655273, 'timing/data': 0.006615802994929254, 'timing/model': 1.5099517289781943, 'learning_rate/qwen_vl_interface': 8.324415054141111e-07, 'learning_rate/action_model': 8.32441505414111e-06, 'epoch': 1.9}) +[2026-06-16 07:16:15,628][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.05866770073771477, 'train/grad_norm_pre_clip': 7.7131853103637695, 'timing/data': 0.006620770029257983, 'timing/model': 1.520614626002498, 'learning_rate/qwen_vl_interface': 8.313231346888476e-07, 'learning_rate/action_model': 8.313231346888476e-06, 'epoch': 1.9}) +[2026-06-16 07:16:21,683][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.018519137054681778, 'train/grad_norm_pre_clip': 2.9413275718688965, 'timing/data': 0.006438482960220426, 'timing/model': 1.4839502850081772, 'learning_rate/qwen_vl_interface': 8.302065803147353e-07, 'learning_rate/action_model': 8.302065803147353e-06, 'epoch': 1.9}) +[2026-06-16 07:16:27,922][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.052822045981884, 'train/grad_norm_pre_clip': 8.665371894836426, 'timing/data': 0.0068390219821594656, 'timing/model': 1.520550910034217, 'learning_rate/qwen_vl_interface': 8.29091842750748e-07, 'learning_rate/action_model': 8.29091842750748e-06, 'epoch': 1.9}) +[2026-06-16 07:16:34,078][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 0.12248518317937851, 'train/grad_norm_pre_clip': 6.342512130737305, 'timing/data': 0.006806934019550681, 'timing/model': 1.526608941028826, 'learning_rate/qwen_vl_interface': 8.279789224551116e-07, 'learning_rate/action_model': 8.279789224551116e-06, 'epoch': 1.9}) +[2026-06-16 07:16:40,050][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.019845856353640556, 'train/grad_norm_pre_clip': 2.1190176010131836, 'timing/data': 0.006646387977525592, 'timing/model': 1.6296131300041452, 'learning_rate/qwen_vl_interface': 8.268678198853058e-07, 'learning_rate/action_model': 8.268678198853059e-06, 'epoch': 1.9}) +[2026-06-16 07:16:46,015][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.13712266087532043, 'train/grad_norm_pre_clip': 12.261489868164062, 'timing/data': 0.00684313802048564, 'timing/model': 1.51112605095841, 'learning_rate/qwen_vl_interface': 8.257585354980628e-07, 'learning_rate/action_model': 8.257585354980628e-06, 'epoch': 1.9}) +[2026-06-16 07:16:52,063][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.05896669998764992, 'train/grad_norm_pre_clip': 10.272406578063965, 'timing/data': 0.0067453240044415, 'timing/model': 1.5217256140313111, 'learning_rate/qwen_vl_interface': 8.246510697493681e-07, 'learning_rate/action_model': 8.246510697493682e-06, 'epoch': 1.9}) +[2026-06-16 07:16:58,224][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.20145194232463837, 'train/grad_norm_pre_clip': 7.549299240112305, 'timing/data': 0.007348962011747062, 'timing/model': 1.5436997109791264, 'learning_rate/qwen_vl_interface': 8.235454230944584e-07, 'learning_rate/action_model': 8.235454230944584e-06, 'epoch': 1.9}) +[2026-06-16 07:17:04,360][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.027682239189743996, 'train/grad_norm_pre_clip': 4.753824234008789, 'timing/data': 0.006834715022705495, 'timing/model': 1.4877541130408645, 'learning_rate/qwen_vl_interface': 8.224415959878234e-07, 'learning_rate/action_model': 8.224415959878234e-06, 'epoch': 1.9}) +[2026-06-16 07:17:10,603][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.18032316863536835, 'train/grad_norm_pre_clip': 14.00189208984375, 'timing/data': 0.006556786014698446, 'timing/model': 1.7664555459632538, 'learning_rate/qwen_vl_interface': 8.213395888832052e-07, 'learning_rate/action_model': 8.213395888832051e-06, 'epoch': 1.9}) +[2026-06-16 07:17:16,432][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.009182271547615528, 'train/grad_norm_pre_clip': 2.726163148880005, 'timing/data': 0.006569374992977828, 'timing/model': 1.5297840749844909, 'learning_rate/qwen_vl_interface': 8.202394022335958e-07, 'learning_rate/action_model': 8.202394022335957e-06, 'epoch': 1.9}) +[2026-06-16 07:17:22,570][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 0.0043014357797801495, 'train/grad_norm_pre_clip': 0.7050471305847168, 'timing/data': 0.006630615040194243, 'timing/model': 1.5065611799946055, 'learning_rate/qwen_vl_interface': 8.191410364912424e-07, 'learning_rate/action_model': 8.191410364912424e-06, 'epoch': 1.9}) +[2026-06-16 07:17:28,371][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.16261495649814606, 'train/grad_norm_pre_clip': 15.509791374206543, 'timing/data': 0.007124772004317492, 'timing/model': 1.4707993949996307, 'learning_rate/qwen_vl_interface': 8.1804449210764e-07, 'learning_rate/action_model': 8.1804449210764e-06, 'epoch': 1.9}) +[2026-06-16 07:17:34,267][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.022861814126372337, 'train/grad_norm_pre_clip': 5.656394958496094, 'timing/data': 0.007315242022741586, 'timing/model': 1.4711333670420572, 'learning_rate/qwen_vl_interface': 8.169497695335377e-07, 'learning_rate/action_model': 8.169497695335376e-06, 'epoch': 1.9}) +[2026-06-16 07:17:40,167][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 0.06924788653850555, 'train/grad_norm_pre_clip': 10.156243324279785, 'timing/data': 0.006647039030212909, 'timing/model': 1.4847669830196537, 'learning_rate/qwen_vl_interface': 8.158568692189342e-07, 'learning_rate/action_model': 8.158568692189342e-06, 'epoch': 1.91}) +[2026-06-16 07:17:46,128][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.019521938636898994, 'train/grad_norm_pre_clip': 16.698610305786133, 'timing/data': 0.0064380120020359755, 'timing/model': 1.725752738013398, 'learning_rate/qwen_vl_interface': 8.147657916130797e-07, 'learning_rate/action_model': 8.147657916130796e-06, 'epoch': 1.91}) +[2026-06-16 07:17:51,876][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.028652802109718323, 'train/grad_norm_pre_clip': 6.2616705894470215, 'timing/data': 0.006398833997081965, 'timing/model': 1.4740858880104497, 'learning_rate/qwen_vl_interface': 8.136765371644753e-07, 'learning_rate/action_model': 8.136765371644753e-06, 'epoch': 1.91}) +[2026-06-16 07:17:57,978][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.10264332592487335, 'train/grad_norm_pre_clip': 15.526725769042969, 'timing/data': 0.006537206994835287, 'timing/model': 1.4749367680051364, 'learning_rate/qwen_vl_interface': 8.125891063208713e-07, 'learning_rate/action_model': 8.125891063208712e-06, 'epoch': 1.91}) +[2026-06-16 07:18:03,870][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.040717292577028275, 'train/grad_norm_pre_clip': 9.464461326599121, 'timing/data': 0.006911508040502667, 'timing/model': 1.4656390890013427, 'learning_rate/qwen_vl_interface': 8.115034995292698e-07, 'learning_rate/action_model': 8.115034995292698e-06, 'epoch': 1.91}) +[2026-06-16 07:18:09,781][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.033157434314489365, 'train/grad_norm_pre_clip': 4.006600856781006, 'timing/data': 0.006304544978775084, 'timing/model': 1.47646498499671, 'learning_rate/qwen_vl_interface': 8.104197172359235e-07, 'learning_rate/action_model': 8.104197172359234e-06, 'epoch': 1.91}) +[2026-06-16 07:18:15,712][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.16135255992412567, 'train/grad_norm_pre_clip': 19.189218521118164, 'timing/data': 0.006559200002811849, 'timing/model': 1.5199889759533107, 'learning_rate/qwen_vl_interface': 8.093377598863334e-07, 'learning_rate/action_model': 8.093377598863334e-06, 'epoch': 1.91}) +[2026-06-16 07:18:21,792][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.1532486081123352, 'train/grad_norm_pre_clip': 12.659635543823242, 'timing/data': 0.007256875978782773, 'timing/model': 1.78625283902511, 'learning_rate/qwen_vl_interface': 8.082576279252519e-07, 'learning_rate/action_model': 8.082576279252518e-06, 'epoch': 1.91}) +[2026-06-16 07:18:27,742][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.07304645329713821, 'train/grad_norm_pre_clip': 8.8339262008667, 'timing/data': 0.006937156023923308, 'timing/model': 1.5370759220095351, 'learning_rate/qwen_vl_interface': 8.071793217966803e-07, 'learning_rate/action_model': 8.071793217966802e-06, 'epoch': 1.91}) +[2026-06-16 07:18:33,763][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.1346874088048935, 'train/grad_norm_pre_clip': 10.549520492553711, 'timing/data': 0.0064819559920579195, 'timing/model': 1.4592554330010898, 'learning_rate/qwen_vl_interface': 8.061028419438691e-07, 'learning_rate/action_model': 8.06102841943869e-06, 'epoch': 1.91}) +[2026-06-16 07:18:39,864][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 0.24544450640678406, 'train/grad_norm_pre_clip': 10.761258125305176, 'timing/data': 0.007014419999904931, 'timing/model': 1.5152127059991471, 'learning_rate/qwen_vl_interface': 8.050281888093193e-07, 'learning_rate/action_model': 8.050281888093192e-06, 'epoch': 1.91}) +[2026-06-16 07:18:45,848][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.007267547305673361, 'train/grad_norm_pre_clip': 2.7429702281951904, 'timing/data': 0.006699627032503486, 'timing/model': 1.5116183270001784, 'learning_rate/qwen_vl_interface': 8.039553628347794e-07, 'learning_rate/action_model': 8.039553628347793e-06, 'epoch': 1.91}) +[2026-06-16 07:18:51,863][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.0021072966046631336, 'train/grad_norm_pre_clip': 0.28092777729034424, 'timing/data': 0.006818681955337524, 'timing/model': 1.4587934709852561, 'learning_rate/qwen_vl_interface': 8.028843644612483e-07, 'learning_rate/action_model': 8.028843644612482e-06, 'epoch': 1.91}) +[2026-06-16 07:18:58,058][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.09210983663797379, 'train/grad_norm_pre_clip': 14.226207733154297, 'timing/data': 0.006836357002612203, 'timing/model': 1.7850607479922473, 'learning_rate/qwen_vl_interface': 8.01815194128973e-07, 'learning_rate/action_model': 8.01815194128973e-06, 'epoch': 1.91}) +[2026-06-16 07:19:03,962][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.02219795435667038, 'train/grad_norm_pre_clip': 2.682439088821411, 'timing/data': 0.006773335044272244, 'timing/model': 1.4875798099674284, 'learning_rate/qwen_vl_interface': 8.007478522774481e-07, 'learning_rate/action_model': 8.00747852277448e-06, 'epoch': 1.91}) +[2026-06-16 07:19:10,177][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.005428317468613386, 'train/grad_norm_pre_clip': 0.9255146384239197, 'timing/data': 0.0070111650275066495, 'timing/model': 1.5044726110063493, 'learning_rate/qwen_vl_interface': 7.996823393454197e-07, 'learning_rate/action_model': 7.996823393454196e-06, 'epoch': 1.91}) +[2026-06-16 07:19:16,435][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.10582774877548218, 'train/grad_norm_pre_clip': 8.330140113830566, 'timing/data': 0.006542034039739519, 'timing/model': 1.5635640539694577, 'learning_rate/qwen_vl_interface': 7.986186557708773e-07, 'learning_rate/action_model': 7.986186557708771e-06, 'epoch': 1.91}) +[2026-06-16 07:19:22,759][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.06605341285467148, 'train/grad_norm_pre_clip': 8.743075370788574, 'timing/data': 0.006703501974698156, 'timing/model': 1.5972709120251238, 'learning_rate/qwen_vl_interface': 7.975568019910617e-07, 'learning_rate/action_model': 7.975568019910617e-06, 'epoch': 1.91}) +[2026-06-16 07:19:28,954][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.1703195720911026, 'train/grad_norm_pre_clip': 13.8506441116333, 'timing/data': 0.006580481014680117, 'timing/model': 1.5126967270043679, 'learning_rate/qwen_vl_interface': 7.964967784424618e-07, 'learning_rate/action_model': 7.964967784424618e-06, 'epoch': 1.91}) +[2026-06-16 07:19:35,108][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.0717923641204834, 'train/grad_norm_pre_clip': 12.87059497833252, 'timing/data': 0.007021069002803415, 'timing/model': 1.743005711992737, 'learning_rate/qwen_vl_interface': 7.954385855608129e-07, 'learning_rate/action_model': 7.95438585560813e-06, 'epoch': 1.91}) +[2026-06-16 07:19:40,947][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.007944482378661633, 'train/grad_norm_pre_clip': 1.1419719457626343, 'timing/data': 0.0066888510482385755, 'timing/model': 1.509972819010727, 'learning_rate/qwen_vl_interface': 7.943822237810974e-07, 'learning_rate/action_model': 7.943822237810975e-06, 'epoch': 1.91}) +[2026-06-16 07:19:47,094][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.03247041255235672, 'train/grad_norm_pre_clip': 5.306941032409668, 'timing/data': 0.0065245889709331095, 'timing/model': 1.4999895520159043, 'learning_rate/qwen_vl_interface': 7.93327693537547e-07, 'learning_rate/action_model': 7.93327693537547e-06, 'epoch': 1.91}) +[2026-06-16 07:19:53,239][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.05881103500723839, 'train/grad_norm_pre_clip': 7.898251056671143, 'timing/data': 0.006488886021543294, 'timing/model': 1.5642405049875379, 'learning_rate/qwen_vl_interface': 7.92274995263638e-07, 'learning_rate/action_model': 7.92274995263638e-06, 'epoch': 1.91}) +[2026-06-16 07:19:59,278][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.08506495505571365, 'train/grad_norm_pre_clip': 10.648719787597656, 'timing/data': 0.0066882899845950305, 'timing/model': 1.4755281929974444, 'learning_rate/qwen_vl_interface': 7.912241293920958e-07, 'learning_rate/action_model': 7.912241293920959e-06, 'epoch': 1.92}) +[2026-06-16 07:20:05,245][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.013398947194218636, 'train/grad_norm_pre_clip': 1.6439812183380127, 'timing/data': 0.006751322012860328, 'timing/model': 1.5104912490351126, 'learning_rate/qwen_vl_interface': 7.90175096354891e-07, 'learning_rate/action_model': 7.90175096354891e-06, 'epoch': 1.92}) +[2026-06-16 07:20:11,192][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.034835703670978546, 'train/grad_norm_pre_clip': 11.330357551574707, 'timing/data': 0.006491160020232201, 'timing/model': 1.6898176289978437, 'learning_rate/qwen_vl_interface': 7.891278965832431e-07, 'learning_rate/action_model': 7.891278965832432e-06, 'epoch': 1.92}) +[2026-06-16 07:20:16,862][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.12313122302293777, 'train/grad_norm_pre_clip': 15.393401145935059, 'timing/data': 0.006685905973426998, 'timing/model': 1.459465169988107, 'learning_rate/qwen_vl_interface': 7.880825305076138e-07, 'learning_rate/action_model': 7.880825305076139e-06, 'epoch': 1.92}) +[2026-06-16 07:20:23,120][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.22099880874156952, 'train/grad_norm_pre_clip': 15.186494827270508, 'timing/data': 0.006437691044993699, 'timing/model': 1.5645729209645651, 'learning_rate/qwen_vl_interface': 7.870389985577148e-07, 'learning_rate/action_model': 7.870389985577148e-06, 'epoch': 1.92}) +[2026-06-16 07:20:29,140][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.10523472726345062, 'train/grad_norm_pre_clip': 9.357181549072266, 'timing/data': 0.0069162549916654825, 'timing/model': 1.508659811981488, 'learning_rate/qwen_vl_interface': 7.859973011625028e-07, 'learning_rate/action_model': 7.859973011625026e-06, 'epoch': 1.92}) +[2026-06-16 07:20:35,123][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.16547609865665436, 'train/grad_norm_pre_clip': 10.989659309387207, 'timing/data': 0.0065204830025322735, 'timing/model': 1.5224678249796852, 'learning_rate/qwen_vl_interface': 7.849574387501801e-07, 'learning_rate/action_model': 7.8495743875018e-06, 'epoch': 1.92}) +[2026-06-16 07:20:41,242][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.09308009594678879, 'train/grad_norm_pre_clip': 6.303339958190918, 'timing/data': 0.0067011790233664215, 'timing/model': 1.5083625229890458, 'learning_rate/qwen_vl_interface': 7.839194117481952e-07, 'learning_rate/action_model': 7.839194117481952e-06, 'epoch': 1.92}) +[2026-06-16 07:20:47,400][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.21182021498680115, 'train/grad_norm_pre_clip': 17.18162727355957, 'timing/data': 0.006639928964432329, 'timing/model': 1.7941496249986812, 'learning_rate/qwen_vl_interface': 7.828832205832418e-07, 'learning_rate/action_model': 7.828832205832418e-06, 'epoch': 1.92}) +[2026-06-16 07:20:53,362][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.014875722117722034, 'train/grad_norm_pre_clip': 4.367336750030518, 'timing/data': 0.006992917973548174, 'timing/model': 1.4730119350133464, 'learning_rate/qwen_vl_interface': 7.818488656812581e-07, 'learning_rate/action_model': 7.818488656812582e-06, 'epoch': 1.92}) +[2026-06-16 07:20:59,671][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 0.11044181138277054, 'train/grad_norm_pre_clip': 17.65287971496582, 'timing/data': 0.006795327004510909, 'timing/model': 1.5930164309684187, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 1.92}) +[2026-06-16 07:21:05,933][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.10769245028495789, 'train/grad_norm_pre_clip': 4.725207805633545, 'timing/data': 0.006760255026165396, 'timing/model': 1.5543273629737087, 'learning_rate/qwen_vl_interface': 7.797856663661836e-07, 'learning_rate/action_model': 7.797856663661835e-06, 'epoch': 1.92}) +[2026-06-16 07:21:12,237][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.021696172654628754, 'train/grad_norm_pre_clip': 3.4588072299957275, 'timing/data': 0.006355870980769396, 'timing/model': 1.5536729490268044, 'learning_rate/qwen_vl_interface': 7.787568228011959e-07, 'learning_rate/action_model': 7.78756822801196e-06, 'epoch': 1.92}) +[2026-06-16 07:21:18,409][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.05559905618429184, 'train/grad_norm_pre_clip': 10.320259094238281, 'timing/data': 0.006564307026565075, 'timing/model': 1.5001726999762468, 'learning_rate/qwen_vl_interface': 7.777298171953858e-07, 'learning_rate/action_model': 7.777298171953857e-06, 'epoch': 1.92}) +[2026-06-16 07:21:24,559][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.16842657327651978, 'train/grad_norm_pre_clip': 9.33092212677002, 'timing/data': 0.006532799976412207, 'timing/model': 1.7804201210383326, 'learning_rate/qwen_vl_interface': 7.767046499709144e-07, 'learning_rate/action_model': 7.767046499709143e-06, 'epoch': 1.92}) +[2026-06-16 07:21:30,609][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.1270580291748047, 'train/grad_norm_pre_clip': 10.8861665725708, 'timing/data': 0.0073507039924152195, 'timing/model': 1.609255388029851, 'learning_rate/qwen_vl_interface': 7.756813215491904e-07, 'learning_rate/action_model': 7.756813215491904e-06, 'epoch': 1.92}) +[2026-06-16 07:21:36,415][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.04941745847463608, 'train/grad_norm_pre_clip': 6.837327480316162, 'timing/data': 0.006455217022448778, 'timing/model': 1.5427533219917677, 'learning_rate/qwen_vl_interface': 7.746598323508655e-07, 'learning_rate/action_model': 7.746598323508655e-06, 'epoch': 1.92}) +[2026-06-16 07:21:42,518][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.06918704509735107, 'train/grad_norm_pre_clip': 14.10498332977295, 'timing/data': 0.00662567705148831, 'timing/model': 1.5186201469623484, 'learning_rate/qwen_vl_interface': 7.736401827958351e-07, 'learning_rate/action_model': 7.73640182795835e-06, 'epoch': 1.92}) +[2026-06-16 07:21:48,569][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 0.0110967718064785, 'train/grad_norm_pre_clip': 3.085407257080078, 'timing/data': 0.007150510034989566, 'timing/model': 1.4644670690177009, 'learning_rate/qwen_vl_interface': 7.726223733032382e-07, 'learning_rate/action_model': 7.72622373303238e-06, 'epoch': 1.92}) +[2026-06-16 07:21:54,551][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.016783805564045906, 'train/grad_norm_pre_clip': 1.9997104406356812, 'timing/data': 0.0064658530172891915, 'timing/model': 1.5109648740035482, 'learning_rate/qwen_vl_interface': 7.716064042914584e-07, 'learning_rate/action_model': 7.716064042914584e-06, 'epoch': 1.92}) +[2026-06-16 07:22:00,486][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.15468084812164307, 'train/grad_norm_pre_clip': 10.676444053649902, 'timing/data': 0.006677373021375388, 'timing/model': 1.6852854180033319, 'learning_rate/qwen_vl_interface': 7.705922761781226e-07, 'learning_rate/action_model': 7.705922761781226e-06, 'epoch': 1.92}) +[2026-06-16 07:22:06,463][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.19202862679958344, 'train/grad_norm_pre_clip': 15.097868919372559, 'timing/data': 0.006818060006480664, 'timing/model': 1.7617018889868632, 'learning_rate/qwen_vl_interface': 7.695799893800996e-07, 'learning_rate/action_model': 7.695799893800996e-06, 'epoch': 1.92}) +[2026-06-16 07:22:12,229][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.07024725526571274, 'train/grad_norm_pre_clip': 15.432228088378906, 'timing/data': 0.006476277019828558, 'timing/model': 1.523114636016544, 'learning_rate/qwen_vl_interface': 7.685695443135028e-07, 'learning_rate/action_model': 7.685695443135027e-06, 'epoch': 1.92}) +[2026-06-16 07:22:18,435][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 0.0023036878556013107, 'train/grad_norm_pre_clip': 0.4093526005744934, 'timing/data': 0.0066155020031146705, 'timing/model': 1.5615858540404588, 'learning_rate/qwen_vl_interface': 7.675609413936888e-07, 'learning_rate/action_model': 7.675609413936888e-06, 'epoch': 1.93}) +[2026-06-16 07:22:24,726][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.08126971870660782, 'train/grad_norm_pre_clip': 9.630539894104004, 'timing/data': 0.007243937056045979, 'timing/model': 1.568223538051825, 'learning_rate/qwen_vl_interface': 7.665541810352541e-07, 'learning_rate/action_model': 7.66554181035254e-06, 'epoch': 1.93}) +[2026-06-16 07:22:30,844][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.06342960149049759, 'train/grad_norm_pre_clip': 9.310810089111328, 'timing/data': 0.0074177730130031705, 'timing/model': 1.510810523002874, 'learning_rate/qwen_vl_interface': 7.655492636520415e-07, 'learning_rate/action_model': 7.655492636520414e-06, 'epoch': 1.93}) +[2026-06-16 07:22:36,992][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.01917572133243084, 'train/grad_norm_pre_clip': 3.4895167350769043, 'timing/data': 0.006599389016628265, 'timing/model': 1.6780972790438682, 'learning_rate/qwen_vl_interface': 7.64546189657134e-07, 'learning_rate/action_model': 7.645461896571339e-06, 'epoch': 1.93}) +[2026-06-16 07:22:43,208][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.0524827316403389, 'train/grad_norm_pre_clip': 6.707449436187744, 'timing/data': 0.0069255990092642605, 'timing/model': 1.7350193510064855, 'learning_rate/qwen_vl_interface': 7.635449594628571e-07, 'learning_rate/action_model': 7.635449594628571e-06, 'epoch': 1.93}) +[2026-06-16 07:22:49,028][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.07322020828723907, 'train/grad_norm_pre_clip': 5.496025085449219, 'timing/data': 0.006475426023826003, 'timing/model': 1.5354416880290955, 'learning_rate/qwen_vl_interface': 7.625455734807803e-07, 'learning_rate/action_model': 7.625455734807803e-06, 'epoch': 1.93}) +[2026-06-16 07:22:55,177][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.05578059330582619, 'train/grad_norm_pre_clip': 9.601837158203125, 'timing/data': 0.0066710939863696694, 'timing/model': 1.5003331080079079, 'learning_rate/qwen_vl_interface': 7.615480321217116e-07, 'learning_rate/action_model': 7.615480321217117e-06, 'epoch': 1.93}) +[2026-06-16 07:23:01,431][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.07397248595952988, 'train/grad_norm_pre_clip': 7.442111492156982, 'timing/data': 0.006682531035039574, 'timing/model': 1.5497099689673632, 'learning_rate/qwen_vl_interface': 7.605523357957028e-07, 'learning_rate/action_model': 7.605523357957028e-06, 'epoch': 1.93}) +[2026-06-16 07:23:07,570][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.030537990853190422, 'train/grad_norm_pre_clip': 4.885286331176758, 'timing/data': 0.006638106016907841, 'timing/model': 1.4854275079560466, 'learning_rate/qwen_vl_interface': 7.59558484912048e-07, 'learning_rate/action_model': 7.59558484912048e-06, 'epoch': 1.93}) +[2026-06-16 07:23:13,559][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.007808836176991463, 'train/grad_norm_pre_clip': 0.6321704387664795, 'timing/data': 0.0067978500155732036, 'timing/model': 1.436647753987927, 'learning_rate/qwen_vl_interface': 7.585664798792807e-07, 'learning_rate/action_model': 7.585664798792807e-06, 'epoch': 1.93}) +[2026-06-16 07:23:19,530][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.030435778200626373, 'train/grad_norm_pre_clip': 10.583680152893066, 'timing/data': 0.006669921975117177, 'timing/model': 1.7269494149950333, 'learning_rate/qwen_vl_interface': 7.575763211051779e-07, 'learning_rate/action_model': 7.575763211051778e-06, 'epoch': 1.93}) +[2026-06-16 07:23:25,262][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.055457502603530884, 'train/grad_norm_pre_clip': 10.961201667785645, 'timing/data': 0.006686426990199834, 'timing/model': 1.4454459479893558, 'learning_rate/qwen_vl_interface': 7.565880089967553e-07, 'learning_rate/action_model': 7.565880089967553e-06, 'epoch': 1.93}) +[2026-06-16 07:23:31,384][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.19170795381069183, 'train/grad_norm_pre_clip': 12.981489181518555, 'timing/data': 0.006833211984485388, 'timing/model': 1.4764902879833244, 'learning_rate/qwen_vl_interface': 7.556015439602725e-07, 'learning_rate/action_model': 7.556015439602725e-06, 'epoch': 1.93}) +[2026-06-16 07:23:37,368][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.015766313299536705, 'train/grad_norm_pre_clip': 3.502321720123291, 'timing/data': 0.00725195900304243, 'timing/model': 1.4575956109911203, 'learning_rate/qwen_vl_interface': 7.546169264012252e-07, 'learning_rate/action_model': 7.5461692640122525e-06, 'epoch': 1.93}) +[2026-06-16 07:23:43,362][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.017752381041646004, 'train/grad_norm_pre_clip': 1.7281252145767212, 'timing/data': 0.006936305027920753, 'timing/model': 1.4618555569904856, 'learning_rate/qwen_vl_interface': 7.536341567243544e-07, 'learning_rate/action_model': 7.536341567243543e-06, 'epoch': 1.93}) +[2026-06-16 07:23:49,295][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.03967612609267235, 'train/grad_norm_pre_clip': 8.460917472839355, 'timing/data': 0.006940070015843958, 'timing/model': 1.6914166280184872, 'learning_rate/qwen_vl_interface': 7.526532353336395e-07, 'learning_rate/action_model': 7.526532353336395e-06, 'epoch': 1.93}) +[2026-06-16 07:23:54,960][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.02673913724720478, 'train/grad_norm_pre_clip': 6.094754695892334, 'timing/data': 0.006492051004897803, 'timing/model': 1.4592167540104128, 'learning_rate/qwen_vl_interface': 7.516741626322994e-07, 'learning_rate/action_model': 7.516741626322994e-06, 'epoch': 1.93}) +[2026-06-16 07:24:00,962][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.11097944527864456, 'train/grad_norm_pre_clip': 13.493935585021973, 'timing/data': 0.00640375999500975, 'timing/model': 1.4608854789985344, 'learning_rate/qwen_vl_interface': 7.506969390227945e-07, 'learning_rate/action_model': 7.506969390227945e-06, 'epoch': 1.93}) +[2026-06-16 07:24:06,761][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.18792815506458282, 'train/grad_norm_pre_clip': 12.90554141998291, 'timing/data': 0.006669401016552001, 'timing/model': 1.459325022995472, 'learning_rate/qwen_vl_interface': 7.497215649068244e-07, 'learning_rate/action_model': 7.497215649068244e-06, 'epoch': 1.93}) +[2026-06-16 07:24:12,659][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.062169283628463745, 'train/grad_norm_pre_clip': 5.182910442352295, 'timing/data': 0.00653216999489814, 'timing/model': 1.4603393740253523, 'learning_rate/qwen_vl_interface': 7.48748040685329e-07, 'learning_rate/action_model': 7.487480406853289e-06, 'epoch': 1.93}) +[2026-06-16 07:24:18,561][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.06935003399848938, 'train/grad_norm_pre_clip': 8.531407356262207, 'timing/data': 0.006756258953828365, 'timing/model': 1.474056264036335, 'learning_rate/qwen_vl_interface': 7.477763667584867e-07, 'learning_rate/action_model': 7.477763667584866e-06, 'epoch': 1.93}) +[2026-06-16 07:24:24,523][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.031196583062410355, 'train/grad_norm_pre_clip': 7.221706390380859, 'timing/data': 0.006491860025562346, 'timing/model': 1.7202514359960333, 'learning_rate/qwen_vl_interface': 7.468065435257159e-07, 'learning_rate/action_model': 7.468065435257159e-06, 'epoch': 1.93}) +[2026-06-16 07:24:30,360][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.2348162680864334, 'train/grad_norm_pre_clip': 18.58811378479004, 'timing/data': 0.006702269951347262, 'timing/model': 1.459698060003575, 'learning_rate/qwen_vl_interface': 7.458385713856745e-07, 'learning_rate/action_model': 7.458385713856744e-06, 'epoch': 1.93}) +[2026-06-16 07:24:36,364][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.08691469579935074, 'train/grad_norm_pre_clip': 9.922994613647461, 'timing/data': 0.006672576011624187, 'timing/model': 1.4590063720243052, 'learning_rate/qwen_vl_interface': 7.448724507362605e-07, 'learning_rate/action_model': 7.448724507362605e-06, 'epoch': 1.93}) +[2026-06-16 07:24:42,361][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.027329793199896812, 'train/grad_norm_pre_clip': 13.80643367767334, 'timing/data': 0.0072173470398411155, 'timing/model': 1.4886306429980323, 'learning_rate/qwen_vl_interface': 7.439081819746078e-07, 'learning_rate/action_model': 7.4390818197460775e-06, 'epoch': 1.94}) +[2026-06-16 07:24:48,628][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.02633027918636799, 'train/grad_norm_pre_clip': 12.607423782348633, 'timing/data': 0.008117611985653639, 'timing/model': 1.5440285920049064, 'learning_rate/qwen_vl_interface': 7.42945765497091e-07, 'learning_rate/action_model': 7.42945765497091e-06, 'epoch': 1.94}) +[2026-06-16 07:33:54,648][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.14432914555072784, 'train/grad_norm_pre_clip': 14.823736190795898, 'eval/loss': 0.04814671725034714, 'eval/flap_precision': 0.9491869918699187, 'eval/flap_recall': 0.935871743486974, 'eval/flap_f1': 0.9424823410696267, 'eval/flap_rate_model': 0.13666666666666666, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9424823410696266, 'timing/data': 0.007224848959594965, 'timing/model': 1.6051399260177277, 'learning_rate/qwen_vl_interface': 7.41985201699324e-07, 'learning_rate/action_model': 7.41985201699324e-06, 'epoch': 1.94}) +[2026-06-16 07:33:54,649][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4500_state +[2026-06-16 07:33:54,654][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 07:34:00,021][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4500_state/model.safetensors +[2026-06-16 07:34:13,000][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4500_state/optimizer.bin +[2026-06-16 07:34:13,000][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4500_state/sampler.bin +[2026-06-16 07:34:13,000][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4500_state/sampler_1.bin +[2026-06-16 07:34:13,001][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4500_state/random_states_0.pkl +[2026-06-16 07:34:13,104][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-06-16 07:34:13,104][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 07:34:13,114][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 07:34:19,133][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.08630141615867615, 'train/grad_norm_pre_clip': 11.021689414978027, 'timing/data': 0.006761399039532989, 'timing/model': 1.4929599959868938, 'learning_rate/qwen_vl_interface': 7.410264909761581e-07, 'learning_rate/action_model': 7.41026490976158e-06, 'epoch': 1.94}) +[2026-06-16 07:34:24,939][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.1402311772108078, 'train/grad_norm_pre_clip': 16.887855529785156, 'timing/data': 0.006412745977286249, 'timing/model': 1.4443431010004133, 'learning_rate/qwen_vl_interface': 7.40069633721683e-07, 'learning_rate/action_model': 7.400696337216829e-06, 'epoch': 1.94}) +[2026-06-16 07:34:30,894][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.09151086956262589, 'train/grad_norm_pre_clip': 6.319180488586426, 'timing/data': 0.006712818052619696, 'timing/model': 1.7098844649735838, 'learning_rate/qwen_vl_interface': 7.391146303292262e-07, 'learning_rate/action_model': 7.391146303292262e-06, 'epoch': 1.94}) +[2026-06-16 07:34:36,557][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.17471013963222504, 'train/grad_norm_pre_clip': 12.680828094482422, 'timing/data': 0.0070394090143963695, 'timing/model': 1.4700927220401354, 'learning_rate/qwen_vl_interface': 7.381614811913538e-07, 'learning_rate/action_model': 7.381614811913537e-06, 'epoch': 1.94}) +[2026-06-16 07:34:42,448][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.21533693373203278, 'train/grad_norm_pre_clip': 14.980106353759766, 'timing/data': 0.00659379200078547, 'timing/model': 1.4607550370274112, 'learning_rate/qwen_vl_interface': 7.372101866998693e-07, 'learning_rate/action_model': 7.372101866998693e-06, 'epoch': 1.94}) +[2026-06-16 07:34:48,446][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.08125579357147217, 'train/grad_norm_pre_clip': 5.357120513916016, 'timing/data': 0.00681999500375241, 'timing/model': 1.4555480320123024, 'learning_rate/qwen_vl_interface': 7.362607472458147e-07, 'learning_rate/action_model': 7.362607472458147e-06, 'epoch': 1.94}) +[2026-06-16 07:34:54,248][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.15809641778469086, 'train/grad_norm_pre_clip': 13.291670799255371, 'timing/data': 0.006368117989040911, 'timing/model': 1.4541232960182242, 'learning_rate/qwen_vl_interface': 7.353131632194682e-07, 'learning_rate/action_model': 7.3531316321946816e-06, 'epoch': 1.94}) +[2026-06-16 07:35:00,149][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.062291085720062256, 'train/grad_norm_pre_clip': 15.454033851623535, 'timing/data': 0.006339225976262242, 'timing/model': 1.4507272129994817, 'learning_rate/qwen_vl_interface': 7.343674350103455e-07, 'learning_rate/action_model': 7.343674350103455e-06, 'epoch': 1.94}) +[2026-06-16 07:35:06,083][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.07771430164575577, 'train/grad_norm_pre_clip': 11.40524673461914, 'timing/data': 0.006489045976195484, 'timing/model': 1.6900688229943626, 'learning_rate/qwen_vl_interface': 7.334235630071995e-07, 'learning_rate/action_model': 7.334235630071995e-06, 'epoch': 1.94}) +[2026-06-16 07:35:11,753][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.1892823427915573, 'train/grad_norm_pre_clip': 10.097146034240723, 'timing/data': 0.00638481299392879, 'timing/model': 1.4601258509792387, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 1.94}) +[2026-06-16 07:35:17,753][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 0.012992775067687035, 'train/grad_norm_pre_clip': 1.598636507987976, 'timing/data': 0.006842045986559242, 'timing/model': 1.5024113230174407, 'learning_rate/qwen_vl_interface': 7.315413891700363e-07, 'learning_rate/action_model': 7.315413891700362e-06, 'epoch': 1.94}) +[2026-06-16 07:35:23,763][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.14379675686359406, 'train/grad_norm_pre_clip': 8.173059463500977, 'timing/data': 0.0064904679893516, 'timing/model': 1.4732745840447024, 'learning_rate/qwen_vl_interface': 7.306030881097092e-07, 'learning_rate/action_model': 7.3060308810970925e-06, 'epoch': 1.94}) +[2026-06-16 07:35:29,653][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.055063240230083466, 'train/grad_norm_pre_clip': 6.362916946411133, 'timing/data': 0.006569744960870594, 'timing/model': 1.459584238997195, 'learning_rate/qwen_vl_interface': 7.296666448027406e-07, 'learning_rate/action_model': 7.296666448027406e-06, 'epoch': 1.94}) +[2026-06-16 07:35:35,556][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.014673981815576553, 'train/grad_norm_pre_clip': 1.957567811012268, 'timing/data': 0.006748979038093239, 'timing/model': 1.432099629950244, 'learning_rate/qwen_vl_interface': 7.287320596340664e-07, 'learning_rate/action_model': 7.287320596340663e-06, 'epoch': 1.94}) +[2026-06-16 07:35:41,484][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.010333762504160404, 'train/grad_norm_pre_clip': 1.1797488927841187, 'timing/data': 0.006550797028467059, 'timing/model': 1.6764404000132345, 'learning_rate/qwen_vl_interface': 7.277993329878593e-07, 'learning_rate/action_model': 7.277993329878594e-06, 'epoch': 1.94}) +[2026-06-16 07:35:47,158][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.01970219612121582, 'train/grad_norm_pre_clip': 3.099591016769409, 'timing/data': 0.006828885991126299, 'timing/model': 1.4616581400041468, 'learning_rate/qwen_vl_interface': 7.268684652475277e-07, 'learning_rate/action_model': 7.268684652475277e-06, 'epoch': 1.94}) +[2026-06-16 07:35:53,052][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.008926398120820522, 'train/grad_norm_pre_clip': 3.3282034397125244, 'timing/data': 0.00658119103172794, 'timing/model': 1.4595308830030262, 'learning_rate/qwen_vl_interface': 7.259394567957169e-07, 'learning_rate/action_model': 7.259394567957169e-06, 'epoch': 1.94}) +[2026-06-16 07:35:58,954][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.014092128723859787, 'train/grad_norm_pre_clip': 2.1992461681365967, 'timing/data': 0.006414004019461572, 'timing/model': 1.4594885489786975, 'learning_rate/qwen_vl_interface': 7.250123080143066e-07, 'learning_rate/action_model': 7.2501230801430665e-06, 'epoch': 1.94}) +[2026-06-16 07:36:04,860][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.04688860848546028, 'train/grad_norm_pre_clip': 8.871984481811523, 'timing/data': 0.0064891650108620524, 'timing/model': 1.4563222819706425, 'learning_rate/qwen_vl_interface': 7.240870192844127e-07, 'learning_rate/action_model': 7.240870192844127e-06, 'epoch': 1.94}) +[2026-06-16 07:36:10,840][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.17097386717796326, 'train/grad_norm_pre_clip': 13.734899520874023, 'timing/data': 0.006498288013972342, 'timing/model': 1.5048615650157444, 'learning_rate/qwen_vl_interface': 7.231635909863858e-07, 'learning_rate/action_model': 7.2316359098638585e-06, 'epoch': 1.94}) +[2026-06-16 07:36:16,786][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 0.06331861019134521, 'train/grad_norm_pre_clip': 8.764238357543945, 'timing/data': 0.0066671669483184814, 'timing/model': 1.687459281005431, 'learning_rate/qwen_vl_interface': 7.222420234998138e-07, 'learning_rate/action_model': 7.2224202349981385e-06, 'epoch': 1.95}) +[2026-06-16 07:36:22,461][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.0022922155912965536, 'train/grad_norm_pre_clip': 0.3861304521560669, 'timing/data': 0.0065124090178869665, 'timing/model': 1.4649335659923963, 'learning_rate/qwen_vl_interface': 7.213223172035175e-07, 'learning_rate/action_model': 7.213223172035174e-06, 'epoch': 1.95}) +[2026-06-16 07:36:28,362][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.012744003906846046, 'train/grad_norm_pre_clip': 1.7713971138000488, 'timing/data': 0.006337361002806574, 'timing/model': 1.464898255013395, 'learning_rate/qwen_vl_interface': 7.204044724755528e-07, 'learning_rate/action_model': 7.204044724755528e-06, 'epoch': 1.95}) +[2026-06-16 07:36:34,255][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.004683579318225384, 'train/grad_norm_pre_clip': 1.3001205921173096, 'timing/data': 0.006705764040816575, 'timing/model': 1.4598398110247217, 'learning_rate/qwen_vl_interface': 7.194884896932132e-07, 'learning_rate/action_model': 7.194884896932132e-06, 'epoch': 1.95}) +[2026-06-16 07:36:40,156][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.08957865089178085, 'train/grad_norm_pre_clip': 10.333123207092285, 'timing/data': 0.006109032954555005, 'timing/model': 1.459192389040254, 'learning_rate/qwen_vl_interface': 7.185743692330221e-07, 'learning_rate/action_model': 7.185743692330221e-06, 'epoch': 1.95}) +[2026-06-16 07:36:46,271][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 0.06783095747232437, 'train/grad_norm_pre_clip': 7.053196907043457, 'timing/data': 0.006820171023719013, 'timing/model': 1.49837284500245, 'learning_rate/qwen_vl_interface': 7.176621114707408e-07, 'learning_rate/action_model': 7.176621114707407e-06, 'epoch': 1.95}) +[2026-06-16 07:36:52,288][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.03016788698732853, 'train/grad_norm_pre_clip': 4.383768081665039, 'timing/data': 0.006602099980227649, 'timing/model': 1.6791153689846396, 'learning_rate/qwen_vl_interface': 7.167517167813642e-07, 'learning_rate/action_model': 7.167517167813642e-06, 'epoch': 1.95}) +[2026-06-16 07:36:58,110][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.34373459219932556, 'train/grad_norm_pre_clip': 11.250314712524414, 'timing/data': 0.007026455015875399, 'timing/model': 1.6062348799896426, 'learning_rate/qwen_vl_interface': 7.15843185539121e-07, 'learning_rate/action_model': 7.15843185539121e-06, 'epoch': 1.95}) +[2026-06-16 07:37:03,755][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.10762055218219757, 'train/grad_norm_pre_clip': 11.180224418640137, 'timing/data': 0.006772049993742257, 'timing/model': 1.451049270981457, 'learning_rate/qwen_vl_interface': 7.149365181174739e-07, 'learning_rate/action_model': 7.149365181174739e-06, 'epoch': 1.95}) +[2026-06-16 07:37:09,656][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.027770206332206726, 'train/grad_norm_pre_clip': 3.079071044921875, 'timing/data': 0.006567918986547738, 'timing/model': 1.45324907702161, 'learning_rate/qwen_vl_interface': 7.14031714889121e-07, 'learning_rate/action_model': 7.14031714889121e-06, 'epoch': 1.95}) +[2026-06-16 07:37:15,556][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.14782482385635376, 'train/grad_norm_pre_clip': 11.62362003326416, 'timing/data': 0.006449754990171641, 'timing/model': 1.4481685649952851, 'learning_rate/qwen_vl_interface': 7.131287762259907e-07, 'learning_rate/action_model': 7.131287762259907e-06, 'epoch': 1.95}) +[2026-06-16 07:37:21,560][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.15205880999565125, 'train/grad_norm_pre_clip': 16.449716567993164, 'timing/data': 0.007278675970155746, 'timing/model': 1.48592942598043, 'learning_rate/qwen_vl_interface': 7.12227702499248e-07, 'learning_rate/action_model': 7.12227702499248e-06, 'epoch': 1.95}) +[2026-06-16 07:37:27,487][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.01035546325147152, 'train/grad_norm_pre_clip': 0.9187253713607788, 'timing/data': 0.006562301015947014, 'timing/model': 1.6841644159867428, 'learning_rate/qwen_vl_interface': 7.113284940792903e-07, 'learning_rate/action_model': 7.113284940792903e-06, 'epoch': 1.95}) +[2026-06-16 07:37:33,388][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.0597214438021183, 'train/grad_norm_pre_clip': 9.622977256774902, 'timing/data': 0.006287625001277775, 'timing/model': 1.6859005239675753, 'learning_rate/qwen_vl_interface': 7.104311513357476e-07, 'learning_rate/action_model': 7.1043115133574755e-06, 'epoch': 1.95}) +[2026-06-16 07:37:39,057][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.26262253522872925, 'train/grad_norm_pre_clip': 10.497876167297363, 'timing/data': 0.006369195994921029, 'timing/model': 1.4587585550034419, 'learning_rate/qwen_vl_interface': 7.095356746374845e-07, 'learning_rate/action_model': 7.095356746374844e-06, 'epoch': 1.95}) +[2026-06-16 07:37:44,957][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.003658162197098136, 'train/grad_norm_pre_clip': 0.5119538903236389, 'timing/data': 0.006900258013047278, 'timing/model': 1.4581008729874156, 'learning_rate/qwen_vl_interface': 7.086420643525973e-07, 'learning_rate/action_model': 7.086420643525972e-06, 'epoch': 1.95}) +[2026-06-16 07:37:50,856][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.1680019199848175, 'train/grad_norm_pre_clip': 16.985963821411133, 'timing/data': 0.006309707998298109, 'timing/model': 1.4578940350329503, 'learning_rate/qwen_vl_interface': 7.077503208484151e-07, 'learning_rate/action_model': 7.0775032084841515e-06, 'epoch': 1.95}) +[2026-06-16 07:37:56,838][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.01279912143945694, 'train/grad_norm_pre_clip': 1.4819871187210083, 'timing/data': 0.006161710014566779, 'timing/model': 1.5063652039971203, 'learning_rate/qwen_vl_interface': 7.068604444915003e-07, 'learning_rate/action_model': 7.068604444915003e-06, 'epoch': 1.95}) +[2026-06-16 07:38:02,788][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.06840828061103821, 'train/grad_norm_pre_clip': 18.919782638549805, 'timing/data': 0.006255788030102849, 'timing/model': 1.690670553012751, 'learning_rate/qwen_vl_interface': 7.059724356476475e-07, 'learning_rate/action_model': 7.059724356476474e-06, 'epoch': 1.95}) +[2026-06-16 07:38:08,462][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.11425264179706573, 'train/grad_norm_pre_clip': 10.70739459991455, 'timing/data': 0.006568849028553814, 'timing/model': 1.464195131964516, 'learning_rate/qwen_vl_interface': 7.050862946818836e-07, 'learning_rate/action_model': 7.050862946818836e-06, 'epoch': 1.95}) +[2026-06-16 07:38:14,455][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.035013601183891296, 'train/grad_norm_pre_clip': 11.740951538085938, 'timing/data': 0.006159786018542945, 'timing/model': 1.4577360120019875, 'learning_rate/qwen_vl_interface': 7.042020219584673e-07, 'learning_rate/action_model': 7.042020219584673e-06, 'epoch': 1.95}) +[2026-06-16 07:38:20,456][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.21060055494308472, 'train/grad_norm_pre_clip': 16.587430953979492, 'timing/data': 0.006874950020574033, 'timing/model': 1.4755358489928767, 'learning_rate/qwen_vl_interface': 7.0331961784089e-07, 'learning_rate/action_model': 7.033196178408899e-06, 'epoch': 1.95}) +[2026-06-16 07:38:26,442][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.012361913919448853, 'train/grad_norm_pre_clip': 2.7078304290771484, 'timing/data': 0.006744727026671171, 'timing/model': 1.511483162001241, 'learning_rate/qwen_vl_interface': 7.024390826918748e-07, 'learning_rate/action_model': 7.024390826918747e-06, 'epoch': 1.95}) +[2026-06-16 07:38:32,422][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.08252842724323273, 'train/grad_norm_pre_clip': 21.1724853515625, 'timing/data': 0.0068646849831566215, 'timing/model': 1.5663450720021501, 'learning_rate/qwen_vl_interface': 7.01560416873376e-07, 'learning_rate/action_model': 7.015604168733759e-06, 'epoch': 1.96}) +[2026-06-16 07:38:38,355][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.009156221523880959, 'train/grad_norm_pre_clip': 1.5532163381576538, 'timing/data': 0.006992573034949601, 'timing/model': 1.6565522590535693, 'learning_rate/qwen_vl_interface': 7.006836207465803e-07, 'learning_rate/action_model': 7.006836207465803e-06, 'epoch': 1.96}) +[2026-06-16 07:38:44,064][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 0.10121787339448929, 'train/grad_norm_pre_clip': 10.518536567687988, 'timing/data': 0.00873694202164188, 'timing/model': 1.4561268810066395, 'learning_rate/qwen_vl_interface': 6.998086946719058e-07, 'learning_rate/action_model': 6.998086946719057e-06, 'epoch': 1.96}) +[2026-06-16 07:38:50,061][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.12689334154129028, 'train/grad_norm_pre_clip': 19.670530319213867, 'timing/data': 0.00656963896472007, 'timing/model': 1.4657285370049067, 'learning_rate/qwen_vl_interface': 6.989356390090008e-07, 'learning_rate/action_model': 6.989356390090007e-06, 'epoch': 1.96}) +[2026-06-16 07:38:55,959][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.22554939985275269, 'train/grad_norm_pre_clip': 17.672771453857422, 'timing/data': 0.007261669030413032, 'timing/model': 1.457112264004536, 'learning_rate/qwen_vl_interface': 6.980644541167459e-07, 'learning_rate/action_model': 6.980644541167458e-06, 'epoch': 1.96}) +[2026-06-16 07:39:01,855][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.014078283682465553, 'train/grad_norm_pre_clip': 4.317407131195068, 'timing/data': 0.005938799004070461, 'timing/model': 1.4329066529753618, 'learning_rate/qwen_vl_interface': 6.971951403532524e-07, 'learning_rate/action_model': 6.971951403532523e-06, 'epoch': 1.96}) +[2026-06-16 07:39:07,858][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.003961893264204264, 'train/grad_norm_pre_clip': 1.027966022491455, 'timing/data': 0.006242806965019554, 'timing/model': 1.5206713550142013, 'learning_rate/qwen_vl_interface': 6.963276980758622e-07, 'learning_rate/action_model': 6.963276980758622e-06, 'epoch': 1.96}) +[2026-06-16 07:39:13,746][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.11516933888196945, 'train/grad_norm_pre_clip': 5.5669331550598145, 'timing/data': 0.0065306409960612655, 'timing/model': 1.6527219979907386, 'learning_rate/qwen_vl_interface': 6.954621276411485e-07, 'learning_rate/action_model': 6.954621276411484e-06, 'epoch': 1.96}) +[2026-06-16 07:39:19,464][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 0.051091477274894714, 'train/grad_norm_pre_clip': 8.175758361816406, 'timing/data': 0.006503009994048625, 'timing/model': 1.4657749829930253, 'learning_rate/qwen_vl_interface': 6.945984294049136e-07, 'learning_rate/action_model': 6.9459842940491364e-06, 'epoch': 1.96}) +[2026-06-16 07:39:25,357][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.027398426085710526, 'train/grad_norm_pre_clip': 6.846786975860596, 'timing/data': 0.006756143004167825, 'timing/model': 1.4584493339643814, 'learning_rate/qwen_vl_interface': 6.937366037221919e-07, 'learning_rate/action_model': 6.937366037221918e-06, 'epoch': 1.96}) +[2026-06-16 07:39:31,256][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.005120815709233284, 'train/grad_norm_pre_clip': 0.4840836226940155, 'timing/data': 0.006686570995952934, 'timing/model': 1.4545204580063, 'learning_rate/qwen_vl_interface': 6.928766509472471e-07, 'learning_rate/action_model': 6.92876650947247e-06, 'epoch': 1.96}) +[2026-06-16 07:39:37,264][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 0.002040717750787735, 'train/grad_norm_pre_clip': 0.34913691878318787, 'timing/data': 0.006502920005004853, 'timing/model': 1.5249743009917438, 'learning_rate/qwen_vl_interface': 6.920185714335737e-07, 'learning_rate/action_model': 6.9201857143357375e-06, 'epoch': 1.96}) +[2026-06-16 07:39:43,188][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.016854729503393173, 'train/grad_norm_pre_clip': 3.7202415466308594, 'timing/data': 0.006499653973150998, 'timing/model': 1.6904509260202758, 'learning_rate/qwen_vl_interface': 6.911623655338959e-07, 'learning_rate/action_model': 6.911623655338959e-06, 'epoch': 1.96}) +[2026-06-16 07:39:48,857][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.1493598073720932, 'train/grad_norm_pre_clip': 9.801896095275879, 'timing/data': 0.006730595021508634, 'timing/model': 1.4590186459827237, 'learning_rate/qwen_vl_interface': 6.903080336001666e-07, 'learning_rate/action_model': 6.903080336001666e-06, 'epoch': 1.96}) +[2026-06-16 07:39:54,856][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.08738637715578079, 'train/grad_norm_pre_clip': 12.667737007141113, 'timing/data': 0.006359348015394062, 'timing/model': 1.4570294489967637, 'learning_rate/qwen_vl_interface': 6.894555759835707e-07, 'learning_rate/action_model': 6.894555759835707e-06, 'epoch': 1.96}) +[2026-06-16 07:40:00,760][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.03128175437450409, 'train/grad_norm_pre_clip': 6.376038551330566, 'timing/data': 0.006799306022003293, 'timing/model': 1.458895843999926, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 1.96}) +[2026-06-16 07:40:06,664][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.19284111261367798, 'train/grad_norm_pre_clip': 25.974824905395508, 'timing/data': 0.006592991994693875, 'timing/model': 1.468312720011454, 'learning_rate/qwen_vl_interface': 6.877562851026584e-07, 'learning_rate/action_model': 6.877562851026584e-06, 'epoch': 1.96}) +[2026-06-16 07:40:12,556][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.09653375297784805, 'train/grad_norm_pre_clip': 8.671648979187012, 'timing/data': 0.006570058001670986, 'timing/model': 1.4795028080116026, 'learning_rate/qwen_vl_interface': 6.869094525368572e-07, 'learning_rate/action_model': 6.8690945253685715e-06, 'epoch': 1.96}) +[2026-06-16 07:40:18,490][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.008752034045755863, 'train/grad_norm_pre_clip': 2.4169442653656006, 'timing/data': 0.006394019001163542, 'timing/model': 1.6841151060070843, 'learning_rate/qwen_vl_interface': 6.860644956852168e-07, 'learning_rate/action_model': 6.860644956852168e-06, 'epoch': 1.96}) +[2026-06-16 07:40:24,159][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.0499214269220829, 'train/grad_norm_pre_clip': 8.302522659301758, 'timing/data': 0.006578630011063069, 'timing/model': 1.458571225986816, 'learning_rate/qwen_vl_interface': 6.852214148950667e-07, 'learning_rate/action_model': 6.852214148950667e-06, 'epoch': 1.96}) +[2026-06-16 07:40:30,160][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.07331280410289764, 'train/grad_norm_pre_clip': 16.992483139038086, 'timing/data': 0.007316606992390007, 'timing/model': 1.4592866459861398, 'learning_rate/qwen_vl_interface': 6.843802105129667e-07, 'learning_rate/action_model': 6.843802105129666e-06, 'epoch': 1.96}) +[2026-06-16 07:40:36,056][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.04466614872217178, 'train/grad_norm_pre_clip': 9.821702003479004, 'timing/data': 0.0065421870094724, 'timing/model': 1.4583091929671355, 'learning_rate/qwen_vl_interface': 6.835408828847035e-07, 'learning_rate/action_model': 6.8354088288470356e-06, 'epoch': 1.96}) +[2026-06-16 07:40:41,955][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.15340584516525269, 'train/grad_norm_pre_clip': 9.544120788574219, 'timing/data': 0.0062729299534112215, 'timing/model': 1.4597636220278218, 'learning_rate/qwen_vl_interface': 6.827034323552929e-07, 'learning_rate/action_model': 6.827034323552929e-06, 'epoch': 1.96}) +[2026-06-16 07:40:47,855][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.02702455408871174, 'train/grad_norm_pre_clip': 4.206943511962891, 'timing/data': 0.00640887999907136, 'timing/model': 1.4701622950378805, 'learning_rate/qwen_vl_interface': 6.818678592689794e-07, 'learning_rate/action_model': 6.818678592689794e-06, 'epoch': 1.97}) +[2026-06-16 07:40:53,786][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.05294111371040344, 'train/grad_norm_pre_clip': 7.040132522583008, 'timing/data': 0.00647739099804312, 'timing/model': 1.6868835809873417, 'learning_rate/qwen_vl_interface': 6.810341639692357e-07, 'learning_rate/action_model': 6.810341639692357e-06, 'epoch': 1.97}) +[2026-06-16 07:40:59,525][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 0.13666465878486633, 'train/grad_norm_pre_clip': 7.226036071777344, 'timing/data': 0.007730907003860921, 'timing/model': 1.5266140879830346, 'learning_rate/qwen_vl_interface': 6.802023467987631e-07, 'learning_rate/action_model': 6.802023467987631e-06, 'epoch': 1.97}) +[2026-06-16 07:41:05,464][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.17307217419147491, 'train/grad_norm_pre_clip': 16.785676956176758, 'timing/data': 0.007310207991395146, 'timing/model': 1.4639776659896597, 'learning_rate/qwen_vl_interface': 6.793724080994884e-07, 'learning_rate/action_model': 6.793724080994884e-06, 'epoch': 1.97}) +[2026-06-16 07:41:11,361][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.1109253466129303, 'train/grad_norm_pre_clip': 13.155527114868164, 'timing/data': 0.006751093023922294, 'timing/model': 1.4583196709863842, 'learning_rate/qwen_vl_interface': 6.785443482125693e-07, 'learning_rate/action_model': 6.7854434821256915e-06, 'epoch': 1.97}) +[2026-06-16 07:41:17,256][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.3559225797653198, 'train/grad_norm_pre_clip': 12.357218742370605, 'timing/data': 0.006551810016389936, 'timing/model': 1.4338304689736106, 'learning_rate/qwen_vl_interface': 6.777181674783895e-07, 'learning_rate/action_model': 6.777181674783895e-06, 'epoch': 1.97}) +[2026-06-16 07:41:23,159][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.040199313312768936, 'train/grad_norm_pre_clip': 8.052694320678711, 'timing/data': 0.0065486159874126315, 'timing/model': 1.4629453200032003, 'learning_rate/qwen_vl_interface': 6.76893866236561e-07, 'learning_rate/action_model': 6.7689386623656094e-06, 'epoch': 1.97}) +[2026-06-16 07:41:29,086][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.013067767024040222, 'train/grad_norm_pre_clip': 1.9791754484176636, 'timing/data': 0.006546642049215734, 'timing/model': 1.6885509630083106, 'learning_rate/qwen_vl_interface': 6.76071444825923e-07, 'learning_rate/action_model': 6.76071444825923e-06, 'epoch': 1.97}) +[2026-06-16 07:41:34,758][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 0.27052173018455505, 'train/grad_norm_pre_clip': 27.244186401367188, 'timing/data': 0.006310864991974086, 'timing/model': 1.4580899559659883, 'learning_rate/qwen_vl_interface': 6.752509035845413e-07, 'learning_rate/action_model': 6.7525090358454126e-06, 'epoch': 1.97}) +[2026-06-16 07:41:40,854][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.0755426213145256, 'train/grad_norm_pre_clip': 12.181194305419922, 'timing/data': 0.0056945569813251495, 'timing/model': 1.456453374994453, 'learning_rate/qwen_vl_interface': 6.744322428497095e-07, 'learning_rate/action_model': 6.744322428497095e-06, 'epoch': 1.97}) +[2026-06-16 07:41:46,753][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.18453654646873474, 'train/grad_norm_pre_clip': 10.981851577758789, 'timing/data': 0.006198399001732469, 'timing/model': 1.457652549026534, 'learning_rate/qwen_vl_interface': 6.736154629579487e-07, 'learning_rate/action_model': 6.736154629579487e-06, 'epoch': 1.97}) +[2026-06-16 07:41:52,664][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.0038519087247550488, 'train/grad_norm_pre_clip': 0.49087369441986084, 'timing/data': 0.006501915981061757, 'timing/model': 1.4686126710148528, 'learning_rate/qwen_vl_interface': 6.728005642450055e-07, 'learning_rate/action_model': 6.7280056424500554e-06, 'epoch': 1.97}) +[2026-06-16 07:41:58,639][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 0.10587262362241745, 'train/grad_norm_pre_clip': 7.5135979652404785, 'timing/data': 0.006635372992604971, 'timing/model': 1.5073631420382299, 'learning_rate/qwen_vl_interface': 6.719875470458553e-07, 'learning_rate/action_model': 6.719875470458553e-06, 'epoch': 1.97}) +[2026-06-16 07:42:04,588][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.10146063566207886, 'train/grad_norm_pre_clip': 12.160223960876465, 'timing/data': 0.0067049950012005866, 'timing/model': 1.687493651988916, 'learning_rate/qwen_vl_interface': 6.711764116946972e-07, 'learning_rate/action_model': 6.7117641169469715e-06, 'epoch': 1.97}) +[2026-06-16 07:42:10,255][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.04949042946100235, 'train/grad_norm_pre_clip': 17.4935359954834, 'timing/data': 0.006204226985573769, 'timing/model': 1.4550209140288644, 'learning_rate/qwen_vl_interface': 6.703671585249587e-07, 'learning_rate/action_model': 6.703671585249586e-06, 'epoch': 1.97}) +[2026-06-16 07:42:16,157][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.05653730034828186, 'train/grad_norm_pre_clip': 11.315244674682617, 'timing/data': 0.00637367699528113, 'timing/model': 1.4521365859545767, 'learning_rate/qwen_vl_interface': 6.69559787869294e-07, 'learning_rate/action_model': 6.69559787869294e-06, 'epoch': 1.97}) +[2026-06-16 07:42:22,053][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.2969124913215637, 'train/grad_norm_pre_clip': 23.80813217163086, 'timing/data': 0.006149426044430584, 'timing/model': 1.4573177169659175, 'learning_rate/qwen_vl_interface': 6.68754300059582e-07, 'learning_rate/action_model': 6.687543000595819e-06, 'epoch': 1.97}) +[2026-06-16 07:42:27,954][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.16949054598808289, 'train/grad_norm_pre_clip': 12.582342147827148, 'timing/data': 0.006392313982360065, 'timing/model': 1.4581790870288387, 'learning_rate/qwen_vl_interface': 6.67950695426929e-07, 'learning_rate/action_model': 6.679506954269291e-06, 'epoch': 1.97}) +[2026-06-16 07:42:33,935][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.3506678640842438, 'train/grad_norm_pre_clip': 16.05295181274414, 'timing/data': 0.006773766013793647, 'timing/model': 1.5062265580054373, 'learning_rate/qwen_vl_interface': 6.671489743016651e-07, 'learning_rate/action_model': 6.671489743016651e-06, 'epoch': 1.97}) +[2026-06-16 07:42:39,886][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.2139892876148224, 'train/grad_norm_pre_clip': 17.29410171508789, 'timing/data': 0.006578036991413683, 'timing/model': 1.6864982030238025, 'learning_rate/qwen_vl_interface': 6.663491370133487e-07, 'learning_rate/action_model': 6.663491370133487e-06, 'epoch': 1.97}) +[2026-06-16 07:42:45,560][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.07507606595754623, 'train/grad_norm_pre_clip': 8.315055847167969, 'timing/data': 0.006779433984775096, 'timing/model': 1.4577343959826976, 'learning_rate/qwen_vl_interface': 6.655511838907629e-07, 'learning_rate/action_model': 6.655511838907629e-06, 'epoch': 1.97}) +[2026-06-16 07:42:51,558][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.12714026868343353, 'train/grad_norm_pre_clip': 16.955747604370117, 'timing/data': 0.0062450970290228724, 'timing/model': 1.4616564229945652, 'learning_rate/qwen_vl_interface': 6.647551152619149e-07, 'learning_rate/action_model': 6.647551152619148e-06, 'epoch': 1.97}) +[2026-06-16 07:42:57,563][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.03990209847688675, 'train/grad_norm_pre_clip': 6.818812370300293, 'timing/data': 0.006452611996792257, 'timing/model': 1.464176376000978, 'learning_rate/qwen_vl_interface': 6.639609314540396e-07, 'learning_rate/action_model': 6.639609314540396e-06, 'epoch': 1.97}) +[2026-06-16 07:43:03,555][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.13589127361774445, 'train/grad_norm_pre_clip': 11.457047462463379, 'timing/data': 0.006432391994167119, 'timing/model': 1.4549867340247147, 'learning_rate/qwen_vl_interface': 6.631686327935954e-07, 'learning_rate/action_model': 6.631686327935954e-06, 'epoch': 1.97}) +[2026-06-16 07:43:09,464][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.011407729238271713, 'train/grad_norm_pre_clip': 1.264691948890686, 'timing/data': 0.006578108004759997, 'timing/model': 1.4370922340312973, 'learning_rate/qwen_vl_interface': 6.623782196062665e-07, 'learning_rate/action_model': 6.623782196062664e-06, 'epoch': 1.98}) +[2026-06-16 07:43:15,390][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.27741721272468567, 'train/grad_norm_pre_clip': 22.794065475463867, 'timing/data': 0.0064804640132933855, 'timing/model': 1.6887495130067691, 'learning_rate/qwen_vl_interface': 6.615896922169622e-07, 'learning_rate/action_model': 6.615896922169622e-06, 'epoch': 1.98}) +[2026-06-16 07:43:21,136][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.16435618698596954, 'train/grad_norm_pre_clip': 17.59264373779297, 'timing/data': 0.0064994310378097, 'timing/model': 1.5050876620225608, 'learning_rate/qwen_vl_interface': 6.608030509498156e-07, 'learning_rate/action_model': 6.608030509498156e-06, 'epoch': 1.98}) +[2026-06-16 07:43:27,058][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.03757200390100479, 'train/grad_norm_pre_clip': 1.970209002494812, 'timing/data': 0.00647242198465392, 'timing/model': 1.458644672005903, 'learning_rate/qwen_vl_interface': 6.600182961281865e-07, 'learning_rate/action_model': 6.600182961281864e-06, 'epoch': 1.98}) +[2026-06-16 07:43:32,956][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.025326106697320938, 'train/grad_norm_pre_clip': 7.365777969360352, 'timing/data': 0.006728969048708677, 'timing/model': 1.4573188229696825, 'learning_rate/qwen_vl_interface': 6.592354280746561e-07, 'learning_rate/action_model': 6.5923542807465595e-06, 'epoch': 1.98}) +[2026-06-16 07:43:38,857][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 0.03395829349756241, 'train/grad_norm_pre_clip': 5.199209213256836, 'timing/data': 0.006479651026893407, 'timing/model': 1.4533279489842243, 'learning_rate/qwen_vl_interface': 6.584544471110335e-07, 'learning_rate/action_model': 6.584544471110335e-06, 'epoch': 1.98}) +[2026-06-16 07:43:44,863][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.0020947072189301252, 'train/grad_norm_pre_clip': 0.31469887495040894, 'timing/data': 0.006569213001057506, 'timing/model': 1.529975380981341, 'learning_rate/qwen_vl_interface': 6.576753535583495e-07, 'learning_rate/action_model': 6.576753535583495e-06, 'epoch': 1.98}) +[2026-06-16 07:43:50,887][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.004860495217144489, 'train/grad_norm_pre_clip': 0.959041953086853, 'timing/data': 0.007092374027706683, 'timing/model': 1.6828966479515657, 'learning_rate/qwen_vl_interface': 6.5689814773686e-07, 'learning_rate/action_model': 6.568981477368599e-06, 'epoch': 1.98}) +[2026-06-16 07:43:56,887][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.20927074551582336, 'train/grad_norm_pre_clip': 11.2108154296875, 'timing/data': 0.006550226011313498, 'timing/model': 1.6868402760010213, 'learning_rate/qwen_vl_interface': 6.56122829966046e-07, 'learning_rate/action_model': 6.56122829966046e-06, 'epoch': 1.98}) +[2026-06-16 07:53:05,848][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.028466304764151573, 'train/grad_norm_pre_clip': 4.3716936111450195, 'eval/loss': 0.04794754087924957, 'eval/flap_precision': 0.9473684210526315, 'eval/flap_recall': 0.9378757515030061, 'eval/flap_f1': 0.9425981873111783, 'eval/flap_rate_model': 0.13722222222222222, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9425981873111783, 'timing/data': 0.007452473975718021, 'timing/model': 1.4728957479819655, 'learning_rate/qwen_vl_interface': 6.553494005646101e-07, 'learning_rate/action_model': 6.553494005646101e-06, 'epoch': 1.98}) +[2026-06-16 07:53:05,848][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4600_state +[2026-06-16 07:53:05,866][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 07:53:11,172][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4600_state/model.safetensors +[2026-06-16 07:53:18,237][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4600_state/optimizer.bin +[2026-06-16 07:53:18,237][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4600_state/sampler.bin +[2026-06-16 07:53:18,238][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4600_state/sampler_1.bin +[2026-06-16 07:53:18,239][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4600_state/random_states_0.pkl +[2026-06-16 07:53:18,481][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4600 +[2026-06-16 07:53:18,481][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 07:53:18,491][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 07:53:24,232][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.10271896421909332, 'train/grad_norm_pre_clip': 10.629127502441406, 'timing/data': 0.0063879420049488544, 'timing/model': 1.4505806989618577, 'learning_rate/qwen_vl_interface': 6.545778598504807e-07, 'learning_rate/action_model': 6.545778598504807e-06, 'epoch': 1.98}) +[2026-06-16 07:53:30,332][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.05618537589907646, 'train/grad_norm_pre_clip': 5.695318222045898, 'timing/data': 0.005995735002215952, 'timing/model': 1.665114564995747, 'learning_rate/qwen_vl_interface': 6.538082081408093e-07, 'learning_rate/action_model': 6.538082081408093e-06, 'epoch': 1.98}) +[2026-06-16 07:53:36,628][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 0.13879987597465515, 'train/grad_norm_pre_clip': 10.189208030700684, 'timing/data': 0.0069682959583587945, 'timing/model': 1.4517193129868247, 'learning_rate/qwen_vl_interface': 6.530404457519697e-07, 'learning_rate/action_model': 6.530404457519697e-06, 'epoch': 1.98}) +[2026-06-16 07:53:42,645][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.12713027000427246, 'train/grad_norm_pre_clip': 7.731017589569092, 'timing/data': 0.006894198013469577, 'timing/model': 1.4585903119877912, 'learning_rate/qwen_vl_interface': 6.522745729995614e-07, 'learning_rate/action_model': 6.522745729995613e-06, 'epoch': 1.98}) +[2026-06-16 07:53:48,633][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.1755126416683197, 'train/grad_norm_pre_clip': 19.757125854492188, 'timing/data': 0.006692890950944275, 'timing/model': 1.5002553919912316, 'learning_rate/qwen_vl_interface': 6.515105901984037e-07, 'learning_rate/action_model': 6.515105901984037e-06, 'epoch': 1.98}) +[2026-06-16 07:53:54,976][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.07430734485387802, 'train/grad_norm_pre_clip': 9.933497428894043, 'timing/data': 0.012633975013159215, 'timing/model': 1.8000065349624492, 'learning_rate/qwen_vl_interface': 6.50748497662543e-07, 'learning_rate/action_model': 6.507484976625429e-06, 'epoch': 1.98}) +[2026-06-16 07:54:00,759][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.03131446614861488, 'train/grad_norm_pre_clip': 10.053671836853027, 'timing/data': 0.006281275011133403, 'timing/model': 1.475355401984416, 'learning_rate/qwen_vl_interface': 6.499882957052459e-07, 'learning_rate/action_model': 6.4998829570524586e-06, 'epoch': 1.98}) +[2026-06-16 07:54:06,845][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.2708912193775177, 'train/grad_norm_pre_clip': 14.48868179321289, 'timing/data': 0.007188150018919259, 'timing/model': 1.576536362990737, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 1.98}) +[2026-06-16 07:54:13,140][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.009506560862064362, 'train/grad_norm_pre_clip': 1.3164476156234741, 'timing/data': 0.0073330330196768045, 'timing/model': 1.5205165239749476, 'learning_rate/qwen_vl_interface': 6.484735647755259e-07, 'learning_rate/action_model': 6.48473564775526e-06, 'epoch': 1.98}) +[2026-06-16 07:54:19,236][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.031720731407403946, 'train/grad_norm_pre_clip': 7.089595317840576, 'timing/data': 0.006662475992925465, 'timing/model': 1.535787586006336, 'learning_rate/qwen_vl_interface': 6.477190364257524e-07, 'learning_rate/action_model': 6.477190364257524e-06, 'epoch': 1.98}) +[2026-06-16 07:54:25,371][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.08696828037500381, 'train/grad_norm_pre_clip': 11.474276542663574, 'timing/data': 0.006848119024652988, 'timing/model': 1.6192368870251812, 'learning_rate/qwen_vl_interface': 6.469663998998393e-07, 'learning_rate/action_model': 6.469663998998393e-06, 'epoch': 1.98}) +[2026-06-16 07:54:31,560][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.03111112490296364, 'train/grad_norm_pre_clip': 3.5933287143707275, 'timing/data': 0.0064989959937520325, 'timing/model': 1.6679421979933977, 'learning_rate/qwen_vl_interface': 6.462156555071681e-07, 'learning_rate/action_model': 6.46215655507168e-06, 'epoch': 1.98}) +[2026-06-16 07:54:37,640][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 0.16707205772399902, 'train/grad_norm_pre_clip': 11.70670223236084, 'timing/data': 0.006817994988523424, 'timing/model': 1.551464103977196, 'learning_rate/qwen_vl_interface': 6.454668035563401e-07, 'learning_rate/action_model': 6.454668035563401e-06, 'epoch': 1.98}) +[2026-06-16 07:54:43,655][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.19003617763519287, 'train/grad_norm_pre_clip': 16.341899871826172, 'timing/data': 0.0063802910153754056, 'timing/model': 1.4593323840526864, 'learning_rate/qwen_vl_interface': 6.447198443551809e-07, 'learning_rate/action_model': 6.4471984435518075e-06, 'epoch': 1.99}) +[2026-06-16 07:54:49,716][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.16648420691490173, 'train/grad_norm_pre_clip': 13.442893981933594, 'timing/data': 0.0066861609811894596, 'timing/model': 1.5535529080079868, 'learning_rate/qwen_vl_interface': 6.439747782107373e-07, 'learning_rate/action_model': 6.4397477821073735e-06, 'epoch': 1.99}) +[2026-06-16 07:54:55,757][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 0.0018910939106717706, 'train/grad_norm_pre_clip': 0.28395023941993713, 'timing/data': 0.006438145996071398, 'timing/model': 1.436502262018621, 'learning_rate/qwen_vl_interface': 6.432316054292779e-07, 'learning_rate/action_model': 6.43231605429278e-06, 'epoch': 1.99}) +[2026-06-16 07:55:01,756][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.006235474720597267, 'train/grad_norm_pre_clip': 0.9414721727371216, 'timing/data': 0.007385890989098698, 'timing/model': 1.4970126299886033, 'learning_rate/qwen_vl_interface': 6.424903263162924e-07, 'learning_rate/action_model': 6.424903263162923e-06, 'epoch': 1.99}) +[2026-06-16 07:55:07,697][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.1140790805220604, 'train/grad_norm_pre_clip': 17.169151306152344, 'timing/data': 0.007059841009322554, 'timing/model': 1.698372115031816, 'learning_rate/qwen_vl_interface': 6.417509411764924e-07, 'learning_rate/action_model': 6.4175094117649244e-06, 'epoch': 1.99}) +[2026-06-16 07:55:13,358][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.009818557649850845, 'train/grad_norm_pre_clip': 1.759063959121704, 'timing/data': 0.006854177976492792, 'timing/model': 1.4576774659799412, 'learning_rate/qwen_vl_interface': 6.410134503138122e-07, 'learning_rate/action_model': 6.4101345031381224e-06, 'epoch': 1.99}) +[2026-06-16 07:55:19,357][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.017294617369771004, 'train/grad_norm_pre_clip': 3.9174816608428955, 'timing/data': 0.006785167031921446, 'timing/model': 1.4624600620009005, 'learning_rate/qwen_vl_interface': 6.402778540314054e-07, 'learning_rate/action_model': 6.402778540314054e-06, 'epoch': 1.99}) +[2026-06-16 07:55:25,355][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.15178842842578888, 'train/grad_norm_pre_clip': 14.203484535217285, 'timing/data': 0.006590579985640943, 'timing/model': 1.4561372809694149, 'learning_rate/qwen_vl_interface': 6.39544152631649e-07, 'learning_rate/action_model': 6.3954415263164895e-06, 'epoch': 1.99}) +[2026-06-16 07:55:31,374][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 0.18196265399456024, 'train/grad_norm_pre_clip': 13.73236083984375, 'timing/data': 0.0065896789892576635, 'timing/model': 1.4753135809442028, 'learning_rate/qwen_vl_interface': 6.388123464161388e-07, 'learning_rate/action_model': 6.388123464161389e-06, 'epoch': 1.99}) +[2026-06-16 07:55:37,267][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.0918988287448883, 'train/grad_norm_pre_clip': 12.545391082763672, 'timing/data': 0.007191083976067603, 'timing/model': 1.4790854080347344, 'learning_rate/qwen_vl_interface': 6.380824356856936e-07, 'learning_rate/action_model': 6.380824356856936e-06, 'epoch': 1.99}) +[2026-06-16 07:55:43,192][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.014530626125633717, 'train/grad_norm_pre_clip': 2.438967227935791, 'timing/data': 0.007141030975617468, 'timing/model': 1.6907689759973437, 'learning_rate/qwen_vl_interface': 6.373544207403513e-07, 'learning_rate/action_model': 6.373544207403513e-06, 'epoch': 1.99}) +[2026-06-16 07:55:48,859][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 0.08162045478820801, 'train/grad_norm_pre_clip': 30.143049240112305, 'timing/data': 0.006584540999028832, 'timing/model': 1.459512942994479, 'learning_rate/qwen_vl_interface': 6.366283018793723e-07, 'learning_rate/action_model': 6.366283018793723e-06, 'epoch': 1.99}) +[2026-06-16 07:55:54,860][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.206154927611351, 'train/grad_norm_pre_clip': 15.859334945678711, 'timing/data': 0.006655896955635399, 'timing/model': 1.4615606830338947, 'learning_rate/qwen_vl_interface': 6.359040794012358e-07, 'learning_rate/action_model': 6.359040794012358e-06, 'epoch': 1.99}) +[2026-06-16 07:56:00,862][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.014041121117770672, 'train/grad_norm_pre_clip': 1.6901307106018066, 'timing/data': 0.006757926021236926, 'timing/model': 1.459036883956287, 'learning_rate/qwen_vl_interface': 6.351817536036431e-07, 'learning_rate/action_model': 6.351817536036431e-06, 'epoch': 1.99}) +[2026-06-16 07:56:06,862][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.19140325486660004, 'train/grad_norm_pre_clip': 10.019715309143066, 'timing/data': 0.0066755450097844005, 'timing/model': 1.4370761890313588, 'learning_rate/qwen_vl_interface': 6.344613247835135e-07, 'learning_rate/action_model': 6.344613247835135e-06, 'epoch': 1.99}) +[2026-06-16 07:56:12,759][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.019040200859308243, 'train/grad_norm_pre_clip': 4.145923137664795, 'timing/data': 0.006595516984816641, 'timing/model': 1.4593187909922563, 'learning_rate/qwen_vl_interface': 6.337427932369898e-07, 'learning_rate/action_model': 6.337427932369898e-06, 'epoch': 1.99}) +[2026-06-16 07:56:18,894][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.0029937371145933867, 'train/grad_norm_pre_clip': 0.32961657643318176, 'timing/data': 0.00878739298786968, 'timing/model': 1.8131247729761526, 'learning_rate/qwen_vl_interface': 6.330261592594328e-07, 'learning_rate/action_model': 6.330261592594327e-06, 'epoch': 1.99}) +[2026-06-16 07:56:24,664][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.041664645075798035, 'train/grad_norm_pre_clip': 7.205855846405029, 'timing/data': 0.007298270997125655, 'timing/model': 1.4639736649696715, 'learning_rate/qwen_vl_interface': 6.323114231454229e-07, 'learning_rate/action_model': 6.323114231454229e-06, 'epoch': 1.99}) +[2026-06-16 07:56:30,660][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.11092080920934677, 'train/grad_norm_pre_clip': 11.951997756958008, 'timing/data': 0.006735734001267701, 'timing/model': 1.4606955869821832, 'learning_rate/qwen_vl_interface': 6.315985851887618e-07, 'learning_rate/action_model': 6.315985851887618e-06, 'epoch': 1.99}) +[2026-06-16 07:56:36,661][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.1384100317955017, 'train/grad_norm_pre_clip': 11.669763565063477, 'timing/data': 0.0066721499897539616, 'timing/model': 1.4806626010104083, 'learning_rate/qwen_vl_interface': 6.308876456824697e-07, 'learning_rate/action_model': 6.3088764568246975e-06, 'epoch': 1.99}) +[2026-06-16 07:56:42,662][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.24911269545555115, 'train/grad_norm_pre_clip': 13.573057174682617, 'timing/data': 0.006704858038574457, 'timing/model': 1.4580375169753097, 'learning_rate/qwen_vl_interface': 6.301786049187875e-07, 'learning_rate/action_model': 6.3017860491878745e-06, 'epoch': 1.99}) +[2026-06-16 07:56:48,654][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.004608636721968651, 'train/grad_norm_pre_clip': 0.6707064509391785, 'timing/data': 0.006675985001493245, 'timing/model': 1.5098807960166596, 'learning_rate/qwen_vl_interface': 6.294714631891755e-07, 'learning_rate/action_model': 6.294714631891754e-06, 'epoch': 1.99}) +[2026-06-16 07:56:54,609][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.07820092886686325, 'train/grad_norm_pre_clip': 9.975537300109863, 'timing/data': 0.006564472045283765, 'timing/model': 1.6934444610378705, 'learning_rate/qwen_vl_interface': 6.287662207843115e-07, 'learning_rate/action_model': 6.287662207843114e-06, 'epoch': 1.99}) +[2026-06-16 07:57:00,263][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.07532429695129395, 'train/grad_norm_pre_clip': 6.551929950714111, 'timing/data': 0.006589739001356065, 'timing/model': 1.460773543047253, 'learning_rate/qwen_vl_interface': 6.280628779940956e-07, 'learning_rate/action_model': 6.280628779940956e-06, 'epoch': 2.0}) +[2026-06-16 07:57:06,302][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.008315261453390121, 'train/grad_norm_pre_clip': 1.4357458353042603, 'timing/data': 0.0068273189826868474, 'timing/model': 1.4958867690293118, 'learning_rate/qwen_vl_interface': 6.273614351076445e-07, 'learning_rate/action_model': 6.273614351076445e-06, 'epoch': 2.0}) +[2026-06-16 07:57:12,214][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.08572319149971008, 'train/grad_norm_pre_clip': 7.976571559906006, 'timing/data': 0.007027252984698862, 'timing/model': 1.5118741460028104, 'learning_rate/qwen_vl_interface': 6.266618924132951e-07, 'learning_rate/action_model': 6.266618924132951e-06, 'epoch': 2.0}) +[2026-06-16 07:57:18,063][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.06202348321676254, 'train/grad_norm_pre_clip': 10.340063095092773, 'timing/data': 0.006541978043969721, 'timing/model': 1.4575425679795444, 'learning_rate/qwen_vl_interface': 6.25964250198603e-07, 'learning_rate/action_model': 6.2596425019860294e-06, 'epoch': 2.0}) +[2026-06-16 07:57:24,059][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.02938961051404476, 'train/grad_norm_pre_clip': 6.933207035064697, 'timing/data': 0.0065520930220372975, 'timing/model': 1.4911156969610602, 'learning_rate/qwen_vl_interface': 6.25268508750342e-07, 'learning_rate/action_model': 6.25268508750342e-06, 'epoch': 2.0}) +[2026-06-16 07:57:29,993][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.034454718232154846, 'train/grad_norm_pre_clip': 18.898380279541016, 'timing/data': 0.0071604689583182335, 'timing/model': 1.6931508849957027, 'learning_rate/qwen_vl_interface': 6.245746683545062e-07, 'learning_rate/action_model': 6.245746683545062e-06, 'epoch': 2.0}) +[2026-06-16 07:57:35,661][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.0011808769777417183, 'train/grad_norm_pre_clip': 0.18028952181339264, 'timing/data': 0.007337699003983289, 'timing/model': 1.4413205729797482, 'learning_rate/qwen_vl_interface': 6.238827292963063e-07, 'learning_rate/action_model': 6.238827292963063e-06, 'epoch': 2.0}) +[2026-06-16 07:57:41,744][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 0.04176970571279526, 'train/grad_norm_pre_clip': 8.912483215332031, 'timing/data': 0.00697381398640573, 'timing/model': 1.5120788699714467, 'learning_rate/qwen_vl_interface': 6.231926918601731e-07, 'learning_rate/action_model': 6.2319269186017305e-06, 'epoch': 2.0}) +[2026-06-16 07:57:47,863][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 0.008529010228812695, 'train/grad_norm_pre_clip': 1.4364700317382812, 'timing/data': 0.006386640015989542, 'timing/model': 1.4944363399990834, 'learning_rate/qwen_vl_interface': 6.22504556329754e-07, 'learning_rate/action_model': 6.22504556329754e-06, 'epoch': 2.0}) +[2026-06-16 07:57:53,862][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.05476997792720795, 'train/grad_norm_pre_clip': 14.666624069213867, 'timing/data': 0.006672330026049167, 'timing/model': 1.5011831179726869, 'learning_rate/qwen_vl_interface': 6.218183229879164e-07, 'learning_rate/action_model': 6.218183229879164e-06, 'epoch': 2.0}) +[2026-06-16 07:57:59,859][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.012976947240531445, 'train/grad_norm_pre_clip': 3.2586636543273926, 'timing/data': 0.006833437015302479, 'timing/model': 1.4995124959968962, 'learning_rate/qwen_vl_interface': 6.211339921167447e-07, 'learning_rate/action_model': 6.211339921167446e-06, 'epoch': 2.0}) +[2026-06-16 07:58:05,783][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.019746053963899612, 'train/grad_norm_pre_clip': 2.891676187515259, 'timing/data': 0.006454179005231708, 'timing/model': 1.6798054509563372, 'learning_rate/qwen_vl_interface': 6.204515639975415e-07, 'learning_rate/action_model': 6.204515639975414e-06, 'epoch': 2.0}) +[2026-06-16 07:58:11,960][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.02069767937064171, 'train/grad_norm_pre_clip': 2.3858540058135986, 'timing/data': 0.007396857021376491, 'timing/model': 1.4349467830033973, 'learning_rate/qwen_vl_interface': 6.197710389108262e-07, 'learning_rate/action_model': 6.197710389108262e-06, 'epoch': 2.0}) +[2026-06-16 07:58:17,959][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.0071477158926427364, 'train/grad_norm_pre_clip': 1.6851472854614258, 'timing/data': 0.006597060011699796, 'timing/model': 1.4584099329658784, 'learning_rate/qwen_vl_interface': 6.190924171363378e-07, 'learning_rate/action_model': 6.190924171363378e-06, 'epoch': 2.0}) +[2026-06-16 07:58:23,960][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.2086753398180008, 'train/grad_norm_pre_clip': 16.81669044494629, 'timing/data': 0.006414429983124137, 'timing/model': 1.459403040993493, 'learning_rate/qwen_vl_interface': 6.184156989530326e-07, 'learning_rate/action_model': 6.184156989530325e-06, 'epoch': 2.0}) +[2026-06-16 07:58:29,968][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.05166742205619812, 'train/grad_norm_pre_clip': 6.120194911956787, 'timing/data': 0.006876721046864986, 'timing/model': 1.5324440270196646, 'learning_rate/qwen_vl_interface': 6.17740884639083e-07, 'learning_rate/action_model': 6.177408846390829e-06, 'epoch': 2.0}) +[2026-06-16 07:58:35,894][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.001309191226027906, 'train/grad_norm_pre_clip': 0.17952312529087067, 'timing/data': 0.006345328991301358, 'timing/model': 1.6937023969949223, 'learning_rate/qwen_vl_interface': 6.170679744718798e-07, 'learning_rate/action_model': 6.1706797447187975e-06, 'epoch': 2.0}) +[2026-06-16 07:58:41,980][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.0040618679486215115, 'train/grad_norm_pre_clip': 0.9490207433700562, 'timing/data': 0.006718316988553852, 'timing/model': 1.6796069480478764, 'learning_rate/qwen_vl_interface': 6.163969687280311e-07, 'learning_rate/action_model': 6.163969687280311e-06, 'epoch': 2.0}) +[2026-06-16 07:58:47,660][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.12881189584732056, 'train/grad_norm_pre_clip': 18.857271194458008, 'timing/data': 0.00662963796639815, 'timing/model': 1.4564673589775339, 'learning_rate/qwen_vl_interface': 6.157278676833627e-07, 'learning_rate/action_model': 6.157278676833627e-06, 'epoch': 2.0}) +[2026-06-16 07:58:53,661][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.187856063246727, 'train/grad_norm_pre_clip': 8.799436569213867, 'timing/data': 0.006612983008380979, 'timing/model': 1.4604938319534995, 'learning_rate/qwen_vl_interface': 6.150606716129155e-07, 'learning_rate/action_model': 6.150606716129154e-06, 'epoch': 2.0}) +[2026-06-16 07:58:59,660][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.22011864185333252, 'train/grad_norm_pre_clip': 13.874979972839355, 'timing/data': 0.007531724986620247, 'timing/model': 1.4765513120219111, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 2.0}) +[2026-06-16 07:59:05,642][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 0.07997167110443115, 'train/grad_norm_pre_clip': 9.81985092163086, 'timing/data': 0.007096064044162631, 'timing/model': 1.5089989330153912, 'learning_rate/qwen_vl_interface': 6.137319954909407e-07, 'learning_rate/action_model': 6.137319954909406e-06, 'epoch': 2.0}) +[2026-06-16 07:59:11,592][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.084210105240345, 'train/grad_norm_pre_clip': 9.319297790527344, 'timing/data': 0.00646181998308748, 'timing/model': 1.6887063839822076, 'learning_rate/qwen_vl_interface': 6.130705159855815e-07, 'learning_rate/action_model': 6.130705159855814e-06, 'epoch': 2.0}) +[2026-06-16 07:59:17,379][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.13386785984039307, 'train/grad_norm_pre_clip': 13.778008460998535, 'timing/data': 0.0068259050021879375, 'timing/model': 1.5748178669600748, 'learning_rate/qwen_vl_interface': 6.124109425467805e-07, 'learning_rate/action_model': 6.124109425467805e-06, 'epoch': 2.01}) +[2026-06-16 07:59:23,060][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.12661047279834747, 'train/grad_norm_pre_clip': 26.59431266784668, 'timing/data': 0.007136322965379804, 'timing/model': 1.4592794320196845, 'learning_rate/qwen_vl_interface': 6.117532754456642e-07, 'learning_rate/action_model': 6.117532754456642e-06, 'epoch': 2.01}) +[2026-06-16 07:59:28,962][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.054663728922605515, 'train/grad_norm_pre_clip': 14.34372615814209, 'timing/data': 0.006547926983330399, 'timing/model': 1.4524447089643218, 'learning_rate/qwen_vl_interface': 6.110975149525746e-07, 'learning_rate/action_model': 6.110975149525746e-06, 'epoch': 2.01}) +[2026-06-16 07:59:34,863][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.010536029934883118, 'train/grad_norm_pre_clip': 1.8247331380844116, 'timing/data': 0.006935426965355873, 'timing/model': 1.4594125819858164, 'learning_rate/qwen_vl_interface': 6.104436613370692e-07, 'learning_rate/action_model': 6.104436613370692e-06, 'epoch': 2.01}) +[2026-06-16 07:59:40,925][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.01270047202706337, 'train/grad_norm_pre_clip': 2.101992607116699, 'timing/data': 0.006656557030510157, 'timing/model': 1.5284852070035413, 'learning_rate/qwen_vl_interface': 6.09791714867924e-07, 'learning_rate/action_model': 6.0979171486792396e-06, 'epoch': 2.01}) +[2026-06-16 07:59:46,881][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.0644836574792862, 'train/grad_norm_pre_clip': 14.464278221130371, 'timing/data': 0.006644919980317354, 'timing/model': 1.7375942970393226, 'learning_rate/qwen_vl_interface': 6.091416758131287e-07, 'learning_rate/action_model': 6.091416758131286e-06, 'epoch': 2.01}) +[2026-06-16 07:59:52,559][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.22665101289749146, 'train/grad_norm_pre_clip': 16.859994888305664, 'timing/data': 0.006639922969043255, 'timing/model': 1.4549974270048551, 'learning_rate/qwen_vl_interface': 6.084935444398895e-07, 'learning_rate/action_model': 6.084935444398894e-06, 'epoch': 2.01}) +[2026-06-16 07:59:58,562][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.0014071848709136248, 'train/grad_norm_pre_clip': 0.2954312562942505, 'timing/data': 0.00823800399666652, 'timing/model': 1.4484487469890155, 'learning_rate/qwen_vl_interface': 6.078473210146294e-07, 'learning_rate/action_model': 6.078473210146294e-06, 'epoch': 2.01}) +[2026-06-16 08:00:04,461][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 0.0534069798886776, 'train/grad_norm_pre_clip': 12.098353385925293, 'timing/data': 0.007254697033204138, 'timing/model': 1.4474588860175572, 'learning_rate/qwen_vl_interface': 6.072030058029862e-07, 'learning_rate/action_model': 6.072030058029861e-06, 'epoch': 2.01}) +[2026-06-16 08:00:10,362][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 0.06070876494050026, 'train/grad_norm_pre_clip': 11.244380950927734, 'timing/data': 0.006519646034575999, 'timing/model': 1.4414365109987557, 'learning_rate/qwen_vl_interface': 6.065605990698136e-07, 'learning_rate/action_model': 6.065605990698136e-06, 'epoch': 2.01}) +[2026-06-16 08:00:16,339][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.1178058460354805, 'train/grad_norm_pre_clip': 13.51259994506836, 'timing/data': 0.006438606011215597, 'timing/model': 1.5053974220063537, 'learning_rate/qwen_vl_interface': 6.059201010791811e-07, 'learning_rate/action_model': 6.059201010791811e-06, 'epoch': 2.01}) +[2026-06-16 08:00:22,294][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.08010939508676529, 'train/grad_norm_pre_clip': 11.43840217590332, 'timing/data': 0.0063954530050978065, 'timing/model': 1.6879452600260265, 'learning_rate/qwen_vl_interface': 6.052815120943721e-07, 'learning_rate/action_model': 6.052815120943721e-06, 'epoch': 2.01}) +[2026-06-16 08:00:27,971][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.23738159239292145, 'train/grad_norm_pre_clip': 16.56688690185547, 'timing/data': 0.0071181760285981, 'timing/model': 1.4709295500069857, 'learning_rate/qwen_vl_interface': 6.046448323778874e-07, 'learning_rate/action_model': 6.046448323778874e-06, 'epoch': 2.01}) +[2026-06-16 08:00:33,963][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.007471583783626556, 'train/grad_norm_pre_clip': 1.0660669803619385, 'timing/data': 0.006480737996753305, 'timing/model': 1.4599329539923929, 'learning_rate/qwen_vl_interface': 6.040100621914419e-07, 'learning_rate/action_model': 6.040100621914418e-06, 'epoch': 2.01}) +[2026-06-16 08:00:39,863][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.04117791727185249, 'train/grad_norm_pre_clip': 11.425783157348633, 'timing/data': 0.00657765002688393, 'timing/model': 1.4537873269873671, 'learning_rate/qwen_vl_interface': 6.033772017959652e-07, 'learning_rate/action_model': 6.033772017959652e-06, 'epoch': 2.01}) +[2026-06-16 08:00:45,788][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.010559194721281528, 'train/grad_norm_pre_clip': 2.8390777111053467, 'timing/data': 0.006444704020395875, 'timing/model': 1.4594767080270685, 'learning_rate/qwen_vl_interface': 6.027462514516022e-07, 'learning_rate/action_model': 6.0274625145160214e-06, 'epoch': 2.01}) +[2026-06-16 08:00:51,737][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.009326042607426643, 'train/grad_norm_pre_clip': 2.943455219268799, 'timing/data': 0.006475570029579103, 'timing/model': 1.5070640609483235, 'learning_rate/qwen_vl_interface': 6.021172114177139e-07, 'learning_rate/action_model': 6.021172114177139e-06, 'epoch': 2.01}) +[2026-06-16 08:00:57,647][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.13862551748752594, 'train/grad_norm_pre_clip': 8.868532180786133, 'timing/data': 0.006513736967463046, 'timing/model': 1.6494450829923153, 'learning_rate/qwen_vl_interface': 6.014900819528734e-07, 'learning_rate/action_model': 6.0149008195287335e-06, 'epoch': 2.01}) +[2026-06-16 08:01:03,358][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.09682722389698029, 'train/grad_norm_pre_clip': 13.313253402709961, 'timing/data': 0.006414471019525081, 'timing/model': 1.4588191930088215, 'learning_rate/qwen_vl_interface': 6.008648633148707e-07, 'learning_rate/action_model': 6.008648633148707e-06, 'epoch': 2.01}) +[2026-06-16 08:01:09,359][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.006398052908480167, 'train/grad_norm_pre_clip': 1.5426372289657593, 'timing/data': 0.0066493559861555696, 'timing/model': 1.4586645939853042, 'learning_rate/qwen_vl_interface': 6.002415557607097e-07, 'learning_rate/action_model': 6.002415557607096e-06, 'epoch': 2.01}) +[2026-06-16 08:01:15,263][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.03202883154153824, 'train/grad_norm_pre_clip': 8.383522033691406, 'timing/data': 0.008046472037676722, 'timing/model': 1.457704919972457, 'learning_rate/qwen_vl_interface': 5.996201595466088e-07, 'learning_rate/action_model': 5.996201595466088e-06, 'epoch': 2.01}) +[2026-06-16 08:01:21,342][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.18476170301437378, 'train/grad_norm_pre_clip': 16.99578285217285, 'timing/data': 0.006690216017886996, 'timing/model': 1.5089677610085346, 'learning_rate/qwen_vl_interface': 5.990006749279998e-07, 'learning_rate/action_model': 5.990006749279998e-06, 'epoch': 2.01}) +[2026-06-16 08:01:27,341][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.002886588219553232, 'train/grad_norm_pre_clip': 0.46370697021484375, 'timing/data': 0.0069641590234823525, 'timing/model': 1.50872945797164, 'learning_rate/qwen_vl_interface': 5.983831021595296e-07, 'learning_rate/action_model': 5.9838310215952956e-06, 'epoch': 2.01}) +[2026-06-16 08:01:33,273][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.022430788725614548, 'train/grad_norm_pre_clip': 9.564347267150879, 'timing/data': 0.006755302019882947, 'timing/model': 1.6702580279670656, 'learning_rate/qwen_vl_interface': 5.977674414950595e-07, 'learning_rate/action_model': 5.9776744149505955e-06, 'epoch': 2.01}) +[2026-06-16 08:01:38,964][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.0016556621994823217, 'train/grad_norm_pre_clip': 0.3989088237285614, 'timing/data': 0.006851463986095041, 'timing/model': 1.448907813988626, 'learning_rate/qwen_vl_interface': 5.971536931876635e-07, 'learning_rate/action_model': 5.971536931876634e-06, 'epoch': 2.02}) +[2026-06-16 08:01:44,963][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.041262656450271606, 'train/grad_norm_pre_clip': 7.153811931610107, 'timing/data': 0.0067009920021519065, 'timing/model': 1.4548365579685196, 'learning_rate/qwen_vl_interface': 5.965418574896315e-07, 'learning_rate/action_model': 5.9654185748963145e-06, 'epoch': 2.02}) +[2026-06-16 08:01:50,963][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.14260724186897278, 'train/grad_norm_pre_clip': 11.005696296691895, 'timing/data': 0.007239455007947981, 'timing/model': 1.4756960850208998, 'learning_rate/qwen_vl_interface': 5.959319346524646e-07, 'learning_rate/action_model': 5.959319346524646e-06, 'epoch': 2.02}) +[2026-06-16 08:01:56,946][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.19185586273670197, 'train/grad_norm_pre_clip': 16.224063873291016, 'timing/data': 0.006708482978865504, 'timing/model': 1.511216017010156, 'learning_rate/qwen_vl_interface': 5.953239249268804e-07, 'learning_rate/action_model': 5.953239249268804e-06, 'epoch': 2.02}) +[2026-06-16 08:02:02,862][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.22388167679309845, 'train/grad_norm_pre_clip': 12.738602638244629, 'timing/data': 0.0068454049760475755, 'timing/model': 1.4558287019608542, 'learning_rate/qwen_vl_interface': 5.947178285628081e-07, 'learning_rate/action_model': 5.94717828562808e-06, 'epoch': 2.02}) +[2026-06-16 08:02:08,794][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.011123827658593655, 'train/grad_norm_pre_clip': 1.5499637126922607, 'timing/data': 0.006860827037598938, 'timing/model': 1.690268013975583, 'learning_rate/qwen_vl_interface': 5.941136458093907e-07, 'learning_rate/action_model': 5.941136458093907e-06, 'epoch': 2.02}) +[2026-06-16 08:02:14,663][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.03985407203435898, 'train/grad_norm_pre_clip': 8.521181106567383, 'timing/data': 0.00648398301564157, 'timing/model': 1.461101368011441, 'learning_rate/qwen_vl_interface': 5.935113769149855e-07, 'learning_rate/action_model': 5.935113769149856e-06, 'epoch': 2.02}) +[2026-06-16 08:02:20,663][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.008805743418633938, 'train/grad_norm_pre_clip': 0.9031305313110352, 'timing/data': 0.007346181955654174, 'timing/model': 1.4583985980134457, 'learning_rate/qwen_vl_interface': 5.92911022127163e-07, 'learning_rate/action_model': 5.929110221271629e-06, 'epoch': 2.02}) +[2026-06-16 08:02:26,662][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.010888413526117802, 'train/grad_norm_pre_clip': 2.0317625999450684, 'timing/data': 0.00675581197720021, 'timing/model': 1.4872332819504663, 'learning_rate/qwen_vl_interface': 5.92312581692705e-07, 'learning_rate/action_model': 5.923125816927049e-06, 'epoch': 2.02}) +[2026-06-16 08:02:32,854][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.09900100529193878, 'train/grad_norm_pre_clip': 19.15667724609375, 'timing/data': 0.006757325027137995, 'timing/model': 1.514591182989534, 'learning_rate/qwen_vl_interface': 5.917160558576088e-07, 'learning_rate/action_model': 5.917160558576088e-06, 'epoch': 2.02}) +[2026-06-16 08:02:38,840][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.18159310519695282, 'train/grad_norm_pre_clip': 15.776084899902344, 'timing/data': 0.006588866992387921, 'timing/model': 1.5144726569997147, 'learning_rate/qwen_vl_interface': 5.911214448670829e-07, 'learning_rate/action_model': 5.911214448670828e-06, 'epoch': 2.02}) +[2026-06-16 08:02:44,898][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.027941172942519188, 'train/grad_norm_pre_clip': 4.813224792480469, 'timing/data': 0.006690967013128102, 'timing/model': 1.6944544819998555, 'learning_rate/qwen_vl_interface': 5.905287489655499e-07, 'learning_rate/action_model': 5.905287489655499e-06, 'epoch': 2.02}) +[2026-06-16 08:02:50,661][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 0.21284617483615875, 'train/grad_norm_pre_clip': 16.174304962158203, 'timing/data': 0.006499123992398381, 'timing/model': 1.5145689439959824, 'learning_rate/qwen_vl_interface': 5.899379683966445e-07, 'learning_rate/action_model': 5.899379683966445e-06, 'epoch': 2.02}) +[2026-06-16 08:02:56,664][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.042653750628232956, 'train/grad_norm_pre_clip': 9.093663215637207, 'timing/data': 0.006759067007806152, 'timing/model': 1.4602359519922175, 'learning_rate/qwen_vl_interface': 5.893491034032142e-07, 'learning_rate/action_model': 5.893491034032142e-06, 'epoch': 2.02}) +[2026-06-16 08:03:02,707][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.04830223321914673, 'train/grad_norm_pre_clip': 6.371560096740723, 'timing/data': 0.006752878020051867, 'timing/model': 1.5214209249825217, 'learning_rate/qwen_vl_interface': 5.887621542273196e-07, 'learning_rate/action_model': 5.887621542273196e-06, 'epoch': 2.02}) +[2026-06-16 08:03:08,747][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.09692828357219696, 'train/grad_norm_pre_clip': 17.346290588378906, 'timing/data': 0.006871281948406249, 'timing/model': 1.5130071599851362, 'learning_rate/qwen_vl_interface': 5.881771211102327e-07, 'learning_rate/action_model': 5.881771211102327e-06, 'epoch': 2.02}) +[2026-06-16 08:12:16,652][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.11192776262760162, 'train/grad_norm_pre_clip': 12.312230110168457, 'eval/loss': 0.04941743612289429, 'eval/flap_precision': 0.9415322580645161, 'eval/flap_recall': 0.935871743486974, 'eval/flap_f1': 0.9386934673366835, 'eval/flap_rate_model': 0.13777777777777778, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9386934673366834, 'timing/data': 0.0067195589654147625, 'timing/model': 1.7230566359939985, 'learning_rate/qwen_vl_interface': 5.875940042924391e-07, 'learning_rate/action_model': 5.875940042924391e-06, 'epoch': 2.02}) +[2026-06-16 08:12:16,652][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4700_state +[2026-06-16 08:12:16,657][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 08:12:22,244][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4700_state/model.safetensors +[2026-06-16 08:12:35,305][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4700_state/optimizer.bin +[2026-06-16 08:12:35,307][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4700_state/sampler.bin +[2026-06-16 08:12:35,307][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4700_state/sampler_1.bin +[2026-06-16 08:12:35,308][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4700_state/random_states_0.pkl +[2026-06-16 08:12:35,415][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4700 +[2026-06-16 08:12:35,415][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 08:12:35,425][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 08:12:41,433][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.019266413524746895, 'train/grad_norm_pre_clip': 5.007233142852783, 'timing/data': 0.006815851025748998, 'timing/model': 1.4582303030183539, 'learning_rate/qwen_vl_interface': 5.870128040136368e-07, 'learning_rate/action_model': 5.870128040136367e-06, 'epoch': 2.02}) +[2026-06-16 08:12:47,378][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.09655976295471191, 'train/grad_norm_pre_clip': 10.18177604675293, 'timing/data': 0.006449171982239932, 'timing/model': 1.683696089021396, 'learning_rate/qwen_vl_interface': 5.864335205127332e-07, 'learning_rate/action_model': 5.864335205127332e-06, 'epoch': 2.02}) +[2026-06-16 08:12:53,040][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.1284957081079483, 'train/grad_norm_pre_clip': 16.101655960083008, 'timing/data': 0.0068899900070391595, 'timing/model': 1.4610732620349154, 'learning_rate/qwen_vl_interface': 5.858561540278513e-07, 'learning_rate/action_model': 5.858561540278513e-06, 'epoch': 2.02}) +[2026-06-16 08:12:59,042][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 0.07465121150016785, 'train/grad_norm_pre_clip': 16.547134399414062, 'timing/data': 0.006588877993635833, 'timing/model': 1.4641914799576625, 'learning_rate/qwen_vl_interface': 5.852807047963244e-07, 'learning_rate/action_model': 5.852807047963243e-06, 'epoch': 2.02}) +[2026-06-16 08:13:04,958][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.027679137885570526, 'train/grad_norm_pre_clip': 8.717752456665039, 'timing/data': 0.0063981880084611475, 'timing/model': 1.4764844930032268, 'learning_rate/qwen_vl_interface': 5.847071730546974e-07, 'learning_rate/action_model': 5.847071730546974e-06, 'epoch': 2.02}) +[2026-06-16 08:13:10,849][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.002859534928575158, 'train/grad_norm_pre_clip': 0.6524534821510315, 'timing/data': 0.006411198002751917, 'timing/model': 1.431692746002227, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 2.02}) +[2026-06-16 08:13:16,743][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 0.018681179732084274, 'train/grad_norm_pre_clip': 4.6647257804870605, 'timing/data': 0.007011809968389571, 'timing/model': 1.478864903037902, 'learning_rate/qwen_vl_interface': 5.835658629833866e-07, 'learning_rate/action_model': 5.8356586298338654e-06, 'epoch': 2.03}) +[2026-06-16 08:13:22,625][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.05837453156709671, 'train/grad_norm_pre_clip': 8.746650695800781, 'timing/data': 0.006455922964960337, 'timing/model': 1.6401992869796231, 'learning_rate/qwen_vl_interface': 5.829980851228512e-07, 'learning_rate/action_model': 5.829980851228512e-06, 'epoch': 2.03}) +[2026-06-16 08:13:28,244][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.002628397662192583, 'train/grad_norm_pre_clip': 0.874215304851532, 'timing/data': 0.0071497519966214895, 'timing/model': 1.457774632028304, 'learning_rate/qwen_vl_interface': 5.824322256905155e-07, 'learning_rate/action_model': 5.824322256905154e-06, 'epoch': 2.03}) +[2026-06-16 08:13:34,245][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.07705891132354736, 'train/grad_norm_pre_clip': 8.94469928741455, 'timing/data': 0.0062479679472744465, 'timing/model': 1.4578433380229399, 'learning_rate/qwen_vl_interface': 5.818682849189818e-07, 'learning_rate/action_model': 5.818682849189817e-06, 'epoch': 2.03}) +[2026-06-16 08:13:40,145][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.11548362672328949, 'train/grad_norm_pre_clip': 8.828757286071777, 'timing/data': 0.006400833022780716, 'timing/model': 1.4565167230321094, 'learning_rate/qwen_vl_interface': 5.813062630400653e-07, 'learning_rate/action_model': 5.813062630400653e-06, 'epoch': 2.03}) +[2026-06-16 08:13:46,139][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 0.067796491086483, 'train/grad_norm_pre_clip': 6.485507965087891, 'timing/data': 0.00665063998894766, 'timing/model': 1.4970146599807777, 'learning_rate/qwen_vl_interface': 5.807461602847922e-07, 'learning_rate/action_model': 5.807461602847922e-06, 'epoch': 2.03}) +[2026-06-16 08:13:52,076][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.02297121472656727, 'train/grad_norm_pre_clip': 4.3784565925598145, 'timing/data': 0.006600216031074524, 'timing/model': 1.6909154820023105, 'learning_rate/qwen_vl_interface': 5.801879768833998e-07, 'learning_rate/action_model': 5.801879768833998e-06, 'epoch': 2.03}) +[2026-06-16 08:13:57,849][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.18012984097003937, 'train/grad_norm_pre_clip': 16.487380981445312, 'timing/data': 0.006511305982712656, 'timing/model': 1.5636760970228352, 'learning_rate/qwen_vl_interface': 5.796317130653363e-07, 'learning_rate/action_model': 5.796317130653363e-06, 'epoch': 2.03}) +[2026-06-16 08:14:03,548][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.0792558565735817, 'train/grad_norm_pre_clip': 14.512553215026855, 'timing/data': 0.006868900964036584, 'timing/model': 1.4591808289987966, 'learning_rate/qwen_vl_interface': 5.790773690592592e-07, 'learning_rate/action_model': 5.7907736905925925e-06, 'epoch': 2.03}) +[2026-06-16 08:14:09,547][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.06429941952228546, 'train/grad_norm_pre_clip': 5.458482265472412, 'timing/data': 0.006365412031300366, 'timing/model': 1.4582160040154122, 'learning_rate/qwen_vl_interface': 5.7852494509304e-07, 'learning_rate/action_model': 5.7852494509304e-06, 'epoch': 2.03}) +[2026-06-16 08:14:15,464][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.07695724815130234, 'train/grad_norm_pre_clip': 10.145143508911133, 'timing/data': 0.006774222012609243, 'timing/model': 1.4731075339950621, 'learning_rate/qwen_vl_interface': 5.779744413937592e-07, 'learning_rate/action_model': 5.779744413937592e-06, 'epoch': 2.03}) +[2026-06-16 08:14:21,434][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.24207274615764618, 'train/grad_norm_pre_clip': 24.150545120239258, 'timing/data': 0.00675842899363488, 'timing/model': 1.4950953589868732, 'learning_rate/qwen_vl_interface': 5.774258581877073e-07, 'learning_rate/action_model': 5.774258581877073e-06, 'epoch': 2.03}) +[2026-06-16 08:14:27,388][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.06282973289489746, 'train/grad_norm_pre_clip': 13.079414367675781, 'timing/data': 0.013461907976306975, 'timing/model': 1.6910122279659845, 'learning_rate/qwen_vl_interface': 5.768791957003866e-07, 'learning_rate/action_model': 5.768791957003865e-06, 'epoch': 2.03}) +[2026-06-16 08:14:33,054][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.019793875515460968, 'train/grad_norm_pre_clip': 13.556879043579102, 'timing/data': 0.012620611989405006, 'timing/model': 1.4556078860186972, 'learning_rate/qwen_vl_interface': 5.763344541565091e-07, 'learning_rate/action_model': 5.7633445415650905e-06, 'epoch': 2.03}) +[2026-06-16 08:14:38,954][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.005966012366116047, 'train/grad_norm_pre_clip': 1.0724513530731201, 'timing/data': 0.012449310976080596, 'timing/model': 1.4428906929679215, 'learning_rate/qwen_vl_interface': 5.75791633779998e-07, 'learning_rate/action_model': 5.75791633779998e-06, 'epoch': 2.03}) +[2026-06-16 08:14:44,854][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 0.023450126871466637, 'train/grad_norm_pre_clip': 6.157995223999023, 'timing/data': 0.013795269012916833, 'timing/model': 1.4521347639965825, 'learning_rate/qwen_vl_interface': 5.752507347939857e-07, 'learning_rate/action_model': 5.752507347939858e-06, 'epoch': 2.03}) +[2026-06-16 08:14:50,755][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.08707043528556824, 'train/grad_norm_pre_clip': 16.170024871826172, 'timing/data': 0.012616417021490633, 'timing/model': 1.4509310059947893, 'learning_rate/qwen_vl_interface': 5.74711757420816e-07, 'learning_rate/action_model': 5.747117574208159e-06, 'epoch': 2.03}) +[2026-06-16 08:14:56,747][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.08829908072948456, 'train/grad_norm_pre_clip': 14.665521621704102, 'timing/data': 0.013130060979165137, 'timing/model': 1.5138797639519908, 'learning_rate/qwen_vl_interface': 5.74174701882042e-07, 'learning_rate/action_model': 5.74174701882042e-06, 'epoch': 2.03}) +[2026-06-16 08:15:02,681][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.12176081538200378, 'train/grad_norm_pre_clip': 14.592958450317383, 'timing/data': 0.01325412403093651, 'timing/model': 1.678678063035477, 'learning_rate/qwen_vl_interface': 5.736395683984262e-07, 'learning_rate/action_model': 5.736395683984261e-06, 'epoch': 2.03}) +[2026-06-16 08:15:08,457][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.0034044908825308084, 'train/grad_norm_pre_clip': 1.6133222579956055, 'timing/data': 0.01308627799153328, 'timing/model': 1.5549427589867264, 'learning_rate/qwen_vl_interface': 5.731063571899421e-07, 'learning_rate/action_model': 5.731063571899421e-06, 'epoch': 2.03}) +[2026-06-16 08:15:14,152][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.04639134928584099, 'train/grad_norm_pre_clip': 10.411042213439941, 'timing/data': 0.01229916198644787, 'timing/model': 1.4535986589617096, 'learning_rate/qwen_vl_interface': 5.72575068475773e-07, 'learning_rate/action_model': 5.72575068475773e-06, 'epoch': 2.03}) +[2026-06-16 08:15:20,148][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.25961849093437195, 'train/grad_norm_pre_clip': 10.943602561950684, 'timing/data': 0.012187076965346932, 'timing/model': 1.4518445399589837, 'learning_rate/qwen_vl_interface': 5.720457024743115e-07, 'learning_rate/action_model': 5.720457024743115e-06, 'epoch': 2.03}) +[2026-06-16 08:15:26,048][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.007640424184501171, 'train/grad_norm_pre_clip': 2.0073914527893066, 'timing/data': 0.012248017010279, 'timing/model': 1.4530219180160202, 'learning_rate/qwen_vl_interface': 5.715182594031601e-07, 'learning_rate/action_model': 5.7151825940316e-06, 'epoch': 2.03}) +[2026-06-16 08:15:31,948][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.023519935086369514, 'train/grad_norm_pre_clip': 8.962294578552246, 'timing/data': 0.011714994034264237, 'timing/model': 1.4259324099984951, 'learning_rate/qwen_vl_interface': 5.709927394791305e-07, 'learning_rate/action_model': 5.709927394791305e-06, 'epoch': 2.04}) +[2026-06-16 08:15:37,975][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.05563017353415489, 'train/grad_norm_pre_clip': 7.705474376678467, 'timing/data': 0.015907665016129613, 'timing/model': 1.6659861350199208, 'learning_rate/qwen_vl_interface': 5.70469142918244e-07, 'learning_rate/action_model': 5.704691429182441e-06, 'epoch': 2.04}) +[2026-06-16 08:15:43,763][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.10567791014909744, 'train/grad_norm_pre_clip': 9.963669776916504, 'timing/data': 0.012337478983681649, 'timing/model': 1.5537948079872876, 'learning_rate/qwen_vl_interface': 5.699474699357317e-07, 'learning_rate/action_model': 5.699474699357317e-06, 'epoch': 2.04}) +[2026-06-16 08:15:49,449][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.0025886031799018383, 'train/grad_norm_pre_clip': 0.5257833003997803, 'timing/data': 0.012502481986302882, 'timing/model': 1.4446161740343086, 'learning_rate/qwen_vl_interface': 5.694277207460331e-07, 'learning_rate/action_model': 5.694277207460332e-06, 'epoch': 2.04}) +[2026-06-16 08:15:55,350][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.006417753174901009, 'train/grad_norm_pre_clip': 2.023853302001953, 'timing/data': 0.012375935970339924, 'timing/model': 1.4413334949640557, 'learning_rate/qwen_vl_interface': 5.689098955627982e-07, 'learning_rate/action_model': 5.689098955627982e-06, 'epoch': 2.04}) +[2026-06-16 08:16:01,256][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.040201179683208466, 'train/grad_norm_pre_clip': 6.666463851928711, 'timing/data': 0.01384047803003341, 'timing/model': 1.4594334089779295, 'learning_rate/qwen_vl_interface': 5.683939945988853e-07, 'learning_rate/action_model': 5.683939945988853e-06, 'epoch': 2.04}) +[2026-06-16 08:16:07,248][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.08879119157791138, 'train/grad_norm_pre_clip': 5.207498073577881, 'timing/data': 0.01257809397066012, 'timing/model': 1.5013887070235796, 'learning_rate/qwen_vl_interface': 5.678800180663608e-07, 'learning_rate/action_model': 5.678800180663609e-06, 'epoch': 2.04}) +[2026-06-16 08:16:13,190][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.11727806180715561, 'train/grad_norm_pre_clip': 11.7238130569458, 'timing/data': 0.012174942006822675, 'timing/model': 1.6866033199476078, 'learning_rate/qwen_vl_interface': 5.673679661765018e-07, 'learning_rate/action_model': 5.6736796617650175e-06, 'epoch': 2.04}) +[2026-06-16 08:16:18,950][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.03950655087828636, 'train/grad_norm_pre_clip': 9.255088806152344, 'timing/data': 0.012001477007288486, 'timing/model': 1.443744999007322, 'learning_rate/qwen_vl_interface': 5.668578391397938e-07, 'learning_rate/action_model': 5.668578391397937e-06, 'epoch': 2.04}) +[2026-06-16 08:16:25,029][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 0.07134880125522614, 'train/grad_norm_pre_clip': 17.16761016845703, 'timing/data': 0.012698031030595303, 'timing/model': 1.5345397720229812, 'learning_rate/qwen_vl_interface': 5.663496371659303e-07, 'learning_rate/action_model': 5.663496371659303e-06, 'epoch': 2.04}) +[2026-06-16 08:16:30,952][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.09922735393047333, 'train/grad_norm_pre_clip': 8.044018745422363, 'timing/data': 0.012725150969345123, 'timing/model': 1.4531229000422172, 'learning_rate/qwen_vl_interface': 5.658433604638141e-07, 'learning_rate/action_model': 5.65843360463814e-06, 'epoch': 2.04}) +[2026-06-16 08:16:36,852][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.09132850915193558, 'train/grad_norm_pre_clip': 14.427184104919434, 'timing/data': 0.012545116012915969, 'timing/model': 1.4508773440029472, 'learning_rate/qwen_vl_interface': 5.653390092415559e-07, 'learning_rate/action_model': 5.653390092415559e-06, 'epoch': 2.04}) +[2026-06-16 08:16:42,838][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 0.1615191102027893, 'train/grad_norm_pre_clip': 9.792069435119629, 'timing/data': 0.011733523046132177, 'timing/model': 1.5037829980137758, 'learning_rate/qwen_vl_interface': 5.648365837064759e-07, 'learning_rate/action_model': 5.648365837064759e-06, 'epoch': 2.04}) +[2026-06-16 08:16:48,789][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.039305876940488815, 'train/grad_norm_pre_clip': 9.73905086517334, 'timing/data': 0.012753902992699295, 'timing/model': 1.6898250540252775, 'learning_rate/qwen_vl_interface': 5.643360840651025e-07, 'learning_rate/action_model': 5.643360840651025e-06, 'epoch': 2.04}) +[2026-06-16 08:16:54,456][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.14501267671585083, 'train/grad_norm_pre_clip': 12.750983238220215, 'timing/data': 0.013543273962568492, 'timing/model': 1.4541021640179679, 'learning_rate/qwen_vl_interface': 5.638375105231719e-07, 'learning_rate/action_model': 5.638375105231719e-06, 'epoch': 2.04}) +[2026-06-16 08:17:00,452][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.22340770065784454, 'train/grad_norm_pre_clip': 7.597999095916748, 'timing/data': 0.012335258012171835, 'timing/model': 1.4522204420063645, 'learning_rate/qwen_vl_interface': 5.633408632856294e-07, 'learning_rate/action_model': 5.6334086328562945e-06, 'epoch': 2.04}) +[2026-06-16 08:17:06,352][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.20688894391059875, 'train/grad_norm_pre_clip': 22.001781463623047, 'timing/data': 0.012777319003362209, 'timing/model': 1.4522152669960633, 'learning_rate/qwen_vl_interface': 5.628461425566268e-07, 'learning_rate/action_model': 5.6284614255662675e-06, 'epoch': 2.04}) +[2026-06-16 08:17:12,352][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.038731981068849564, 'train/grad_norm_pre_clip': 13.124832153320312, 'timing/data': 0.012773973983712494, 'timing/model': 1.4714792220038362, 'learning_rate/qwen_vl_interface': 5.62353348539526e-07, 'learning_rate/action_model': 5.623533485395259e-06, 'epoch': 2.04}) +[2026-06-16 08:17:18,342][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 0.10413499176502228, 'train/grad_norm_pre_clip': 11.337944030761719, 'timing/data': 0.012969781993888319, 'timing/model': 1.5078835979802534, 'learning_rate/qwen_vl_interface': 5.618624814368956e-07, 'learning_rate/action_model': 5.618624814368956e-06, 'epoch': 2.04}) +[2026-06-16 08:17:24,281][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.06534626334905624, 'train/grad_norm_pre_clip': 5.724053382873535, 'timing/data': 0.012542583979666233, 'timing/model': 1.681159096013289, 'learning_rate/qwen_vl_interface': 5.613735414505126e-07, 'learning_rate/action_model': 5.613735414505125e-06, 'epoch': 2.04}) +[2026-06-16 08:17:30,180][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.03564973920583725, 'train/grad_norm_pre_clip': 4.63710355758667, 'timing/data': 0.01225430000340566, 'timing/model': 1.6800853930180892, 'learning_rate/qwen_vl_interface': 5.608865287813617e-07, 'learning_rate/action_model': 5.608865287813616e-06, 'epoch': 2.04}) +[2026-06-16 08:17:35,858][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.21149349212646484, 'train/grad_norm_pre_clip': 21.243242263793945, 'timing/data': 0.012404831999447197, 'timing/model': 1.4586958109866828, 'learning_rate/qwen_vl_interface': 5.604014436296362e-07, 'learning_rate/action_model': 5.604014436296362e-06, 'epoch': 2.04}) +[2026-06-16 08:17:41,853][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.011699861846864223, 'train/grad_norm_pre_clip': 2.534618616104126, 'timing/data': 0.012305315991397947, 'timing/model': 1.4556030870298855, 'learning_rate/qwen_vl_interface': 5.599182861947348e-07, 'learning_rate/action_model': 5.599182861947347e-06, 'epoch': 2.04}) +[2026-06-16 08:17:47,848][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.011444734409451485, 'train/grad_norm_pre_clip': 1.5342795848846436, 'timing/data': 0.012885618954896927, 'timing/model': 1.4475681289914064, 'learning_rate/qwen_vl_interface': 5.594370566752665e-07, 'learning_rate/action_model': 5.5943705667526645e-06, 'epoch': 2.05}) +[2026-06-16 08:17:53,872][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.0901804268360138, 'train/grad_norm_pre_clip': 13.030426025390625, 'timing/data': 0.013093474961351603, 'timing/model': 1.5114894110010937, 'learning_rate/qwen_vl_interface': 5.589577552690456e-07, 'learning_rate/action_model': 5.589577552690456e-06, 'epoch': 2.05}) +[2026-06-16 08:17:59,787][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.16834181547164917, 'train/grad_norm_pre_clip': 7.310615062713623, 'timing/data': 0.012715548975393176, 'timing/model': 1.689965507015586, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 2.05}) +[2026-06-16 08:18:05,659][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.010676465928554535, 'train/grad_norm_pre_clip': 1.223151683807373, 'timing/data': 0.014779413002543151, 'timing/model': 1.4610400719684549, 'learning_rate/qwen_vl_interface': 5.580049375836466e-07, 'learning_rate/action_model': 5.580049375836466e-06, 'epoch': 2.05}) +[2026-06-16 08:18:11,650][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.12104251980781555, 'train/grad_norm_pre_clip': 12.440056800842285, 'timing/data': 0.011676992988213897, 'timing/model': 1.4553591620060615, 'learning_rate/qwen_vl_interface': 5.575314216961356e-07, 'learning_rate/action_model': 5.575314216961356e-06, 'epoch': 2.05}) +[2026-06-16 08:18:17,654][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.06590323895215988, 'train/grad_norm_pre_clip': 9.199894905090332, 'timing/data': 0.012182575999759138, 'timing/model': 1.4576722120400518, 'learning_rate/qwen_vl_interface': 5.570598347052073e-07, 'learning_rate/action_model': 5.570598347052072e-06, 'epoch': 2.05}) +[2026-06-16 08:18:23,558][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.145285964012146, 'train/grad_norm_pre_clip': 7.745916843414307, 'timing/data': 0.011949605017434806, 'timing/model': 1.4359739890205674, 'learning_rate/qwen_vl_interface': 5.565901768047126e-07, 'learning_rate/action_model': 5.5659017680471255e-06, 'epoch': 2.05}) +[2026-06-16 08:18:29,450][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.1499386727809906, 'train/grad_norm_pre_clip': 14.189400672912598, 'timing/data': 0.012351805984508246, 'timing/model': 1.508093143987935, 'learning_rate/qwen_vl_interface': 5.561224481877105e-07, 'learning_rate/action_model': 5.561224481877105e-06, 'epoch': 2.05}) +[2026-06-16 08:18:35,341][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.15493343770503998, 'train/grad_norm_pre_clip': 13.640167236328125, 'timing/data': 0.01216013397788629, 'timing/model': 1.6392148269806057, 'learning_rate/qwen_vl_interface': 5.55656649046467e-07, 'learning_rate/action_model': 5.556566490464669e-06, 'epoch': 2.05}) +[2026-06-16 08:18:41,058][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 0.092807836830616, 'train/grad_norm_pre_clip': 15.681787490844727, 'timing/data': 0.012359827000182122, 'timing/model': 1.461319078982342, 'learning_rate/qwen_vl_interface': 5.551927795724544e-07, 'learning_rate/action_model': 5.551927795724543e-06, 'epoch': 2.05}) +[2026-06-16 08:18:47,052][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.2167782038450241, 'train/grad_norm_pre_clip': 24.747873306274414, 'timing/data': 0.012374980025924742, 'timing/model': 1.4548204359598458, 'learning_rate/qwen_vl_interface': 5.547308399563515e-07, 'learning_rate/action_model': 5.547308399563514e-06, 'epoch': 2.05}) +[2026-06-16 08:18:53,055][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.048545800149440765, 'train/grad_norm_pre_clip': 9.107132911682129, 'timing/data': 0.01188836497021839, 'timing/model': 1.4608088579843752, 'learning_rate/qwen_vl_interface': 5.542708303880442e-07, 'learning_rate/action_model': 5.542708303880442e-06, 'epoch': 2.05}) +[2026-06-16 08:18:58,959][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.009878256358206272, 'train/grad_norm_pre_clip': 1.2670143842697144, 'timing/data': 0.012394919991493225, 'timing/model': 1.4356214010040276, 'learning_rate/qwen_vl_interface': 5.538127510566258e-07, 'learning_rate/action_model': 5.538127510566258e-06, 'epoch': 2.05}) +[2026-06-16 08:19:05,030][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.08353759348392487, 'train/grad_norm_pre_clip': 11.71345043182373, 'timing/data': 0.012128919013775885, 'timing/model': 1.4977137119858526, 'learning_rate/qwen_vl_interface': 5.533566021503944e-07, 'learning_rate/action_model': 5.533566021503943e-06, 'epoch': 2.05}) +[2026-06-16 08:19:10,939][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 0.12014149129390717, 'train/grad_norm_pre_clip': 9.92566204071045, 'timing/data': 0.012454106996301562, 'timing/model': 1.6406328559969552, 'learning_rate/qwen_vl_interface': 5.529023838568566e-07, 'learning_rate/action_model': 5.529023838568566e-06, 'epoch': 2.05}) +[2026-06-16 08:19:16,651][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.049562979489564896, 'train/grad_norm_pre_clip': 6.793166637420654, 'timing/data': 0.012095499958377331, 'timing/model': 1.4533695239806548, 'learning_rate/qwen_vl_interface': 5.524500963627239e-07, 'learning_rate/action_model': 5.524500963627239e-06, 'epoch': 2.05}) +[2026-06-16 08:19:22,551][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 0.21936747431755066, 'train/grad_norm_pre_clip': 15.074600219726562, 'timing/data': 0.012176088988780975, 'timing/model': 1.4519937100121751, 'learning_rate/qwen_vl_interface': 5.519997398539145e-07, 'learning_rate/action_model': 5.519997398539145e-06, 'epoch': 2.05}) +[2026-06-16 08:19:28,551][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.3844728171825409, 'train/grad_norm_pre_clip': 14.725403785705566, 'timing/data': 0.011862898012623191, 'timing/model': 1.4892214909777977, 'learning_rate/qwen_vl_interface': 5.515513145155531e-07, 'learning_rate/action_model': 5.515513145155531e-06, 'epoch': 2.05}) +[2026-06-16 08:19:34,550][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.10106009244918823, 'train/grad_norm_pre_clip': 15.793049812316895, 'timing/data': 0.012328131997492164, 'timing/model': 1.4944185889908113, 'learning_rate/qwen_vl_interface': 5.511048205319702e-07, 'learning_rate/action_model': 5.511048205319701e-06, 'epoch': 2.05}) +[2026-06-16 08:19:40,482][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 0.11683490127325058, 'train/grad_norm_pre_clip': 31.38072967529297, 'timing/data': 0.012494056019932032, 'timing/model': 1.6830464120139368, 'learning_rate/qwen_vl_interface': 5.506602580867034e-07, 'learning_rate/action_model': 5.506602580867034e-06, 'epoch': 2.05}) +[2026-06-16 08:19:46,191][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.10468535125255585, 'train/grad_norm_pre_clip': 10.886566162109375, 'timing/data': 0.012369964038953185, 'timing/model': 1.4910539169795811, 'learning_rate/qwen_vl_interface': 5.502176273624942e-07, 'learning_rate/action_model': 5.502176273624942e-06, 'epoch': 2.05}) +[2026-06-16 08:19:52,155][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.01798834279179573, 'train/grad_norm_pre_clip': 3.2894253730773926, 'timing/data': 0.012322094000410289, 'timing/model': 1.4527710360125639, 'learning_rate/qwen_vl_interface': 5.497769285412921e-07, 'learning_rate/action_model': 5.497769285412921e-06, 'epoch': 2.05}) +[2026-06-16 08:19:57,952][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.11203769594430923, 'train/grad_norm_pre_clip': 11.476519584655762, 'timing/data': 0.012462069978937507, 'timing/model': 1.4543777240323834, 'learning_rate/qwen_vl_interface': 5.493381618042519e-07, 'learning_rate/action_model': 5.493381618042518e-06, 'epoch': 2.05}) +[2026-06-16 08:20:04,058][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.13248209655284882, 'train/grad_norm_pre_clip': 14.01962661743164, 'timing/data': 0.014802149962633848, 'timing/model': 1.4574238290078938, 'learning_rate/qwen_vl_interface': 5.489013273317343e-07, 'learning_rate/action_model': 5.489013273317343e-06, 'epoch': 2.06}) +[2026-06-16 08:20:10,035][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.08343645185232162, 'train/grad_norm_pre_clip': 16.23044204711914, 'timing/data': 0.01168408599914983, 'timing/model': 1.5511536539997905, 'learning_rate/qwen_vl_interface': 5.484664253033042e-07, 'learning_rate/action_model': 5.484664253033042e-06, 'epoch': 2.06}) +[2026-06-16 08:20:15,999][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.060760922729969025, 'train/grad_norm_pre_clip': 6.701086521148682, 'timing/data': 0.012407079979311675, 'timing/model': 1.7030650849919766, 'learning_rate/qwen_vl_interface': 5.480334558977351e-07, 'learning_rate/action_model': 5.480334558977351e-06, 'epoch': 2.06}) +[2026-06-16 08:20:21,649][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.013941783457994461, 'train/grad_norm_pre_clip': 2.8830387592315674, 'timing/data': 0.012235487985890359, 'timing/model': 1.4524187879869714, 'learning_rate/qwen_vl_interface': 5.476024192930029e-07, 'learning_rate/action_model': 5.47602419293003e-06, 'epoch': 2.06}) +[2026-06-16 08:20:27,646][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.0724945068359375, 'train/grad_norm_pre_clip': 13.539742469787598, 'timing/data': 0.012395533034577966, 'timing/model': 1.4335395690286532, 'learning_rate/qwen_vl_interface': 5.471733156662914e-07, 'learning_rate/action_model': 5.4717331566629135e-06, 'epoch': 2.06}) +[2026-06-16 08:20:33,837][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 0.041328828781843185, 'train/grad_norm_pre_clip': 8.617409706115723, 'timing/data': 0.035454683005809784, 'timing/model': 1.5777009280282073, 'learning_rate/qwen_vl_interface': 5.467461451939882e-07, 'learning_rate/action_model': 5.467461451939882e-06, 'epoch': 2.06}) +[2026-06-16 08:20:39,756][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.05983554944396019, 'train/grad_norm_pre_clip': 15.329316139221191, 'timing/data': 0.012625631992705166, 'timing/model': 1.4512006840086542, 'learning_rate/qwen_vl_interface': 5.463209080516873e-07, 'learning_rate/action_model': 5.463209080516873e-06, 'epoch': 2.06}) +[2026-06-16 08:20:45,899][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.16491983830928802, 'train/grad_norm_pre_clip': 16.948095321655273, 'timing/data': 0.01250816899118945, 'timing/model': 1.532632154005114, 'learning_rate/qwen_vl_interface': 5.458976044141875e-07, 'learning_rate/action_model': 5.458976044141875e-06, 'epoch': 2.06}) +[2026-06-16 08:20:51,891][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.2062598466873169, 'train/grad_norm_pre_clip': 15.296131134033203, 'timing/data': 0.012562698975671083, 'timing/model': 1.6873520810040645, 'learning_rate/qwen_vl_interface': 5.454762344554931e-07, 'learning_rate/action_model': 5.4547623445549316e-06, 'epoch': 2.06}) +[2026-06-16 08:20:57,656][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 0.16205409169197083, 'train/grad_norm_pre_clip': 21.245349884033203, 'timing/data': 0.013114841014612466, 'timing/model': 1.4525355400401168, 'learning_rate/qwen_vl_interface': 5.450567983488137e-07, 'learning_rate/action_model': 5.450567983488137e-06, 'epoch': 2.06}) +[2026-06-16 08:21:03,657][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.20329928398132324, 'train/grad_norm_pre_clip': 11.972332954406738, 'timing/data': 0.011485374998301268, 'timing/model': 1.4471521499799564, 'learning_rate/qwen_vl_interface': 5.446392962665629e-07, 'learning_rate/action_model': 5.4463929626656296e-06, 'epoch': 2.06}) +[2026-06-16 08:21:09,621][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.0022470124531537294, 'train/grad_norm_pre_clip': 0.6309153437614441, 'timing/data': 0.012214648013468832, 'timing/model': 1.5163126129773445, 'learning_rate/qwen_vl_interface': 5.442237283803605e-07, 'learning_rate/action_model': 5.442237283803604e-06, 'epoch': 2.06}) +[2026-06-16 08:21:15,645][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.08722957968711853, 'train/grad_norm_pre_clip': 11.579010963439941, 'timing/data': 0.012322646973188967, 'timing/model': 1.5019004190107808, 'learning_rate/qwen_vl_interface': 5.43810094861031e-07, 'learning_rate/action_model': 5.43810094861031e-06, 'epoch': 2.06}) +[2026-06-16 08:21:21,646][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.0510057732462883, 'train/grad_norm_pre_clip': 6.717188835144043, 'timing/data': 0.018567698949482292, 'timing/model': 1.5263529240037315, 'learning_rate/qwen_vl_interface': 5.433983958786027e-07, 'learning_rate/action_model': 5.433983958786027e-06, 'epoch': 2.06}) +[2026-06-16 08:21:27,590][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.10141140967607498, 'train/grad_norm_pre_clip': 9.806424140930176, 'timing/data': 0.012408594018779695, 'timing/model': 1.6895965799922124, 'learning_rate/qwen_vl_interface': 5.429886316023102e-07, 'learning_rate/action_model': 5.429886316023101e-06, 'epoch': 2.06}) +[2026-06-16 08:21:33,353][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.041566044092178345, 'train/grad_norm_pre_clip': 6.721459865570068, 'timing/data': 0.01283225603401661, 'timing/model': 1.4543974620173685, 'learning_rate/qwen_vl_interface': 5.425808022005916e-07, 'learning_rate/action_model': 5.425808022005916e-06, 'epoch': 2.06}) +[2026-06-16 08:21:39,351][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.16692301630973816, 'train/grad_norm_pre_clip': 15.072338104248047, 'timing/data': 0.012343928043264896, 'timing/model': 1.4536036779754795, 'learning_rate/qwen_vl_interface': 5.421749078410903e-07, 'learning_rate/action_model': 5.421749078410903e-06, 'epoch': 2.06}) +[2026-06-16 08:21:45,251][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.035092584788799286, 'train/grad_norm_pre_clip': 7.960064888000488, 'timing/data': 0.012984320055693388, 'timing/model': 1.4503780090017244, 'learning_rate/qwen_vl_interface': 5.417709486906547e-07, 'learning_rate/action_model': 5.417709486906547e-06, 'epoch': 2.06}) +[2026-06-16 08:21:51,153][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.04465358331799507, 'train/grad_norm_pre_clip': 4.679446220397949, 'timing/data': 0.011860407947096974, 'timing/model': 1.479750327009242, 'learning_rate/qwen_vl_interface': 5.413689249153367e-07, 'learning_rate/action_model': 5.413689249153366e-06, 'epoch': 2.06}) +[2026-06-16 08:21:57,053][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.1256341189146042, 'train/grad_norm_pre_clip': 19.3681697845459, 'timing/data': 0.011692971049342304, 'timing/model': 1.4441573479562066, 'learning_rate/qwen_vl_interface': 5.409688366803931e-07, 'learning_rate/action_model': 5.409688366803931e-06, 'epoch': 2.06}) +[2026-06-16 08:22:03,085][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.11230984330177307, 'train/grad_norm_pre_clip': 8.813270568847656, 'timing/data': 0.012140478997025639, 'timing/model': 1.6834979450213723, 'learning_rate/qwen_vl_interface': 5.405706841502853e-07, 'learning_rate/action_model': 5.4057068415028525e-06, 'epoch': 2.06}) +[2026-06-16 08:22:08,752][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 0.01088665146380663, 'train/grad_norm_pre_clip': 1.2497895956039429, 'timing/data': 0.011920996999833733, 'timing/model': 1.4529506730032153, 'learning_rate/qwen_vl_interface': 5.401744674886781e-07, 'learning_rate/action_model': 5.40174467488678e-06, 'epoch': 2.06}) +[2026-06-16 08:22:14,754][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.09705889970064163, 'train/grad_norm_pre_clip': 13.715941429138184, 'timing/data': 0.011945503007154912, 'timing/model': 1.455130981048569, 'learning_rate/qwen_vl_interface': 5.397801868584418e-07, 'learning_rate/action_model': 5.397801868584417e-06, 'epoch': 2.06}) +[2026-06-16 08:22:20,662][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.04565960913896561, 'train/grad_norm_pre_clip': 7.592650890350342, 'timing/data': 0.01238833402749151, 'timing/model': 1.4572955960175022, 'learning_rate/qwen_vl_interface': 5.393878424216502e-07, 'learning_rate/action_model': 5.393878424216501e-06, 'epoch': 2.06}) +[2026-06-16 08:31:12,846][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.1107020154595375, 'train/grad_norm_pre_clip': 10.511608123779297, 'eval/loss': 0.047290075570344925, 'eval/flap_precision': 0.947261663286004, 'eval/flap_recall': 0.935871743486974, 'eval/flap_f1': 0.9415322580645161, 'eval/flap_rate_model': 0.13694444444444445, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9415322580645161, 'timing/data': 0.012541139032691717, 'timing/model': 1.4557193900109269, 'learning_rate/qwen_vl_interface': 5.389974343395813e-07, 'learning_rate/action_model': 5.389974343395812e-06, 'epoch': 2.07}) +[2026-06-16 08:31:12,847][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4800_state +[2026-06-16 08:31:12,853][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 08:31:18,898][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4800_state/model.safetensors +[2026-06-16 08:31:27,199][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4800_state/optimizer.bin +[2026-06-16 08:31:27,200][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4800_state/sampler.bin +[2026-06-16 08:31:27,200][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4800_state/sampler_1.bin +[2026-06-16 08:31:27,201][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4800_state/random_states_0.pkl +[2026-06-16 08:31:27,307][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4800 +[2026-06-16 08:31:27,307][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 08:31:27,317][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 08:31:33,139][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 0.24146290123462677, 'train/grad_norm_pre_clip': 14.916505813598633, 'timing/data': 0.013370364962611347, 'timing/model': 1.4523148829466663, 'learning_rate/qwen_vl_interface': 5.38608962772717e-07, 'learning_rate/action_model': 5.38608962772717e-06, 'epoch': 2.07}) +[2026-06-16 08:31:39,068][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.011709065176546574, 'train/grad_norm_pre_clip': 2.7063167095184326, 'timing/data': 0.012174005038104951, 'timing/model': 1.4747401289641857, 'learning_rate/qwen_vl_interface': 5.382224278807429e-07, 'learning_rate/action_model': 5.382224278807429e-06, 'epoch': 2.07}) +[2026-06-16 08:31:44,943][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.009270760230720043, 'train/grad_norm_pre_clip': 2.0725531578063965, 'timing/data': 0.012245821009855717, 'timing/model': 1.433217612968292, 'learning_rate/qwen_vl_interface': 5.378378298225496e-07, 'learning_rate/action_model': 5.378378298225496e-06, 'epoch': 2.07}) +[2026-06-16 08:31:50,978][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.12032324820756912, 'train/grad_norm_pre_clip': 9.502249717712402, 'timing/data': 0.0188002809882164, 'timing/model': 1.682859444001224, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 2.07}) +[2026-06-16 08:31:56,844][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.012611549347639084, 'train/grad_norm_pre_clip': 2.5684163570404053, 'timing/data': 0.018806821026373655, 'timing/model': 1.457657956983894, 'learning_rate/qwen_vl_interface': 5.37074444839083e-07, 'learning_rate/action_model': 5.370744448390829e-06, 'epoch': 2.07}) +[2026-06-16 08:32:02,842][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 0.1046145036816597, 'train/grad_norm_pre_clip': 9.768799781799316, 'timing/data': 0.01292962598381564, 'timing/model': 1.4522453199606389, 'learning_rate/qwen_vl_interface': 5.366956582276085e-07, 'learning_rate/action_model': 5.366956582276084e-06, 'epoch': 2.07}) +[2026-06-16 08:32:08,844][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.029704995453357697, 'train/grad_norm_pre_clip': 7.228445529937744, 'timing/data': 0.012941775028593838, 'timing/model': 1.4542609690106474, 'learning_rate/qwen_vl_interface': 5.363188090775122e-07, 'learning_rate/action_model': 5.363188090775122e-06, 'epoch': 2.07}) +[2026-06-16 08:32:14,844][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.17151440680027008, 'train/grad_norm_pre_clip': 14.962305068969727, 'timing/data': 0.012098575010895729, 'timing/model': 1.436906007991638, 'learning_rate/qwen_vl_interface': 5.359438975437018e-07, 'learning_rate/action_model': 5.359438975437017e-06, 'epoch': 2.07}) +[2026-06-16 08:32:20,822][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.030090143904089928, 'train/grad_norm_pre_clip': 7.625658988952637, 'timing/data': 0.01586066698655486, 'timing/model': 1.4917560940375552, 'learning_rate/qwen_vl_interface': 5.355709237802895e-07, 'learning_rate/action_model': 5.3557092378028955e-06, 'epoch': 2.07}) +[2026-06-16 08:32:26,783][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.023682082071900368, 'train/grad_norm_pre_clip': 7.467801094055176, 'timing/data': 0.012440196995157748, 'timing/model': 1.7883432069793344, 'learning_rate/qwen_vl_interface': 5.351998879405911e-07, 'learning_rate/action_model': 5.351998879405911e-06, 'epoch': 2.07}) +[2026-06-16 08:32:32,557][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.00832676887512207, 'train/grad_norm_pre_clip': 3.131218194961548, 'timing/data': 0.012760618003085256, 'timing/model': 1.4626693410100415, 'learning_rate/qwen_vl_interface': 5.348307901771256e-07, 'learning_rate/action_model': 5.348307901771256e-06, 'epoch': 2.07}) +[2026-06-16 08:32:38,451][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.07071708142757416, 'train/grad_norm_pre_clip': 14.872714042663574, 'timing/data': 0.012364485010039061, 'timing/model': 1.4540195720037445, 'learning_rate/qwen_vl_interface': 5.344636306416147e-07, 'learning_rate/action_model': 5.344636306416146e-06, 'epoch': 2.07}) +[2026-06-16 08:32:44,252][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.017340384423732758, 'train/grad_norm_pre_clip': 2.0399701595306396, 'timing/data': 0.011898740951437503, 'timing/model': 1.4550607980345376, 'learning_rate/qwen_vl_interface': 5.340984094849836e-07, 'learning_rate/action_model': 5.340984094849836e-06, 'epoch': 2.07}) +[2026-06-16 08:32:50,352][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.10407667607069016, 'train/grad_norm_pre_clip': 14.379660606384277, 'timing/data': 0.012308113044127822, 'timing/model': 1.4545325639774092, 'learning_rate/qwen_vl_interface': 5.337351268573606e-07, 'learning_rate/action_model': 5.337351268573606e-06, 'epoch': 2.07}) +[2026-06-16 08:32:56,354][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.0037691774778068066, 'train/grad_norm_pre_clip': 0.6478581428527832, 'timing/data': 0.01263712701620534, 'timing/model': 1.4531148730311543, 'learning_rate/qwen_vl_interface': 5.33373782908079e-07, 'learning_rate/action_model': 5.333737829080789e-06, 'epoch': 2.07}) +[2026-06-16 08:33:02,283][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.10808407515287399, 'train/grad_norm_pre_clip': 10.598228454589844, 'timing/data': 0.012359439046122134, 'timing/model': 1.686320708016865, 'learning_rate/qwen_vl_interface': 5.330143777856718e-07, 'learning_rate/action_model': 5.330143777856718e-06, 'epoch': 2.07}) +[2026-06-16 08:33:08,159][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 0.018459318205714226, 'train/grad_norm_pre_clip': 2.981006622314453, 'timing/data': 0.011944768019020557, 'timing/model': 1.5613935370347463, 'learning_rate/qwen_vl_interface': 5.326569116378786e-07, 'learning_rate/action_model': 5.326569116378786e-06, 'epoch': 2.07}) +[2026-06-16 08:33:14,151][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.17340679466724396, 'train/grad_norm_pre_clip': 6.553343772888184, 'timing/data': 0.013055201037786901, 'timing/model': 1.4546016700332984, 'learning_rate/qwen_vl_interface': 5.323013846116387e-07, 'learning_rate/action_model': 5.323013846116386e-06, 'epoch': 2.07}) +[2026-06-16 08:33:20,056][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.07349788397550583, 'train/grad_norm_pre_clip': 8.736662864685059, 'timing/data': 0.012253082008101046, 'timing/model': 1.4609338470036164, 'learning_rate/qwen_vl_interface': 5.31947796853097e-07, 'learning_rate/action_model': 5.3194779685309696e-06, 'epoch': 2.07}) +[2026-06-16 08:33:25,951][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.20989517867565155, 'train/grad_norm_pre_clip': 23.667545318603516, 'timing/data': 0.012682783999480307, 'timing/model': 1.4499645659816451, 'learning_rate/qwen_vl_interface': 5.315961485075992e-07, 'learning_rate/action_model': 5.315961485075993e-06, 'epoch': 2.07}) +[2026-06-16 08:33:31,854][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.011089367792010307, 'train/grad_norm_pre_clip': 1.3725178241729736, 'timing/data': 0.012652278994210064, 'timing/model': 1.4277788879699074, 'learning_rate/qwen_vl_interface': 5.312464397196958e-07, 'learning_rate/action_model': 5.312464397196958e-06, 'epoch': 2.07}) +[2026-06-16 08:33:37,778][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.07684400677680969, 'train/grad_norm_pre_clip': 10.818942070007324, 'timing/data': 0.013904571009334177, 'timing/model': 1.679330940998625, 'learning_rate/qwen_vl_interface': 5.308986706331377e-07, 'learning_rate/action_model': 5.308986706331376e-06, 'epoch': 2.07}) +[2026-06-16 08:33:43,696][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.03811664879322052, 'train/grad_norm_pre_clip': 6.524897575378418, 'timing/data': 0.01239994796924293, 'timing/model': 1.699542462010868, 'learning_rate/qwen_vl_interface': 5.305528413908797e-07, 'learning_rate/action_model': 5.305528413908798e-06, 'epoch': 2.08}) +[2026-06-16 08:33:49,351][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.2755267322063446, 'train/grad_norm_pre_clip': 12.459953308105469, 'timing/data': 0.01311027305200696, 'timing/model': 1.4510067329974845, 'learning_rate/qwen_vl_interface': 5.302089521350799e-07, 'learning_rate/action_model': 5.3020895213507996e-06, 'epoch': 2.08}) +[2026-06-16 08:33:55,255][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.09420997649431229, 'train/grad_norm_pre_clip': 11.272804260253906, 'timing/data': 0.01269158796640113, 'timing/model': 1.452669752005022, 'learning_rate/qwen_vl_interface': 5.298670030070977e-07, 'learning_rate/action_model': 5.298670030070976e-06, 'epoch': 2.08}) +[2026-06-16 08:34:01,152][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.024360911920666695, 'train/grad_norm_pre_clip': 6.219041347503662, 'timing/data': 0.01259356300579384, 'timing/model': 1.4372821220313199, 'learning_rate/qwen_vl_interface': 5.295269941474957e-07, 'learning_rate/action_model': 5.295269941474957e-06, 'epoch': 2.08}) +[2026-06-16 08:34:07,262][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.015089647844433784, 'train/grad_norm_pre_clip': 3.4580466747283936, 'timing/data': 0.01233414199668914, 'timing/model': 1.47442114102887, 'learning_rate/qwen_vl_interface': 5.291889256960384e-07, 'learning_rate/action_model': 5.291889256960384e-06, 'epoch': 2.08}) +[2026-06-16 08:34:13,191][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.015508790500462055, 'train/grad_norm_pre_clip': 1.445923089981079, 'timing/data': 0.012112506025005132, 'timing/model': 1.6919950799783692, 'learning_rate/qwen_vl_interface': 5.288527977916929e-07, 'learning_rate/action_model': 5.288527977916929e-06, 'epoch': 2.08}) +[2026-06-16 08:34:18,854][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.2361159473657608, 'train/grad_norm_pre_clip': 7.770964622497559, 'timing/data': 0.012916467967443168, 'timing/model': 1.4391053830040619, 'learning_rate/qwen_vl_interface': 5.28518610572629e-07, 'learning_rate/action_model': 5.28518610572629e-06, 'epoch': 2.08}) +[2026-06-16 08:34:24,850][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.05031043291091919, 'train/grad_norm_pre_clip': 4.139890670776367, 'timing/data': 0.012755942007061094, 'timing/model': 1.4547893439885229, 'learning_rate/qwen_vl_interface': 5.281863641762184e-07, 'learning_rate/action_model': 5.281863641762184e-06, 'epoch': 2.08}) +[2026-06-16 08:34:30,749][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 0.08907782286405563, 'train/grad_norm_pre_clip': 9.016828536987305, 'timing/data': 0.012865172000601888, 'timing/model': 1.4525890810182318, 'learning_rate/qwen_vl_interface': 5.278560587390355e-07, 'learning_rate/action_model': 5.278560587390355e-06, 'epoch': 2.08}) +[2026-06-16 08:34:36,657][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.05457548797130585, 'train/grad_norm_pre_clip': 6.7003397941589355, 'timing/data': 0.013137433037627488, 'timing/model': 1.4287347680074163, 'learning_rate/qwen_vl_interface': 5.275276943968557e-07, 'learning_rate/action_model': 5.275276943968557e-06, 'epoch': 2.08}) +[2026-06-16 08:34:42,550][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.08497391641139984, 'train/grad_norm_pre_clip': 12.272212982177734, 'timing/data': 0.01304377504857257, 'timing/model': 1.4989399219630286, 'learning_rate/qwen_vl_interface': 5.272012712846566e-07, 'learning_rate/action_model': 5.2720127128465654e-06, 'epoch': 2.08}) +[2026-06-16 08:34:48,429][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.030603326857089996, 'train/grad_norm_pre_clip': 8.264473915100098, 'timing/data': 0.012689563969615847, 'timing/model': 1.631921851017978, 'learning_rate/qwen_vl_interface': 5.268767895366198e-07, 'learning_rate/action_model': 5.268767895366198e-06, 'epoch': 2.08}) +[2026-06-16 08:34:54,150][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.008303429931402206, 'train/grad_norm_pre_clip': 1.414454698562622, 'timing/data': 0.012185523053631186, 'timing/model': 1.453775077010505, 'learning_rate/qwen_vl_interface': 5.265542492861268e-07, 'learning_rate/action_model': 5.2655424928612676e-06, 'epoch': 2.08}) +[2026-06-16 08:35:00,056][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 0.0473179966211319, 'train/grad_norm_pre_clip': 7.111335277557373, 'timing/data': 0.013349245011340827, 'timing/model': 1.4600360990152694, 'learning_rate/qwen_vl_interface': 5.262336506657617e-07, 'learning_rate/action_model': 5.2623365066576165e-06, 'epoch': 2.08}) +[2026-06-16 08:35:05,948][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.03226673603057861, 'train/grad_norm_pre_clip': 4.40323543548584, 'timing/data': 0.012539233022835106, 'timing/model': 1.4538750289939344, 'learning_rate/qwen_vl_interface': 5.259149938073103e-07, 'learning_rate/action_model': 5.259149938073102e-06, 'epoch': 2.08}) +[2026-06-16 08:35:11,849][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.022474627941846848, 'train/grad_norm_pre_clip': 3.6806013584136963, 'timing/data': 0.012472385016735643, 'timing/model': 1.4294138540280983, 'learning_rate/qwen_vl_interface': 5.255982788417609e-07, 'learning_rate/action_model': 5.255982788417608e-06, 'epoch': 2.08}) +[2026-06-16 08:35:17,866][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.009303731843829155, 'train/grad_norm_pre_clip': 1.5947102308273315, 'timing/data': 0.012512123037595302, 'timing/model': 1.5246985960402526, 'learning_rate/qwen_vl_interface': 5.252835058993029e-07, 'learning_rate/action_model': 5.252835058993029e-06, 'epoch': 2.08}) +[2026-06-16 08:35:23,736][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.06953100860118866, 'train/grad_norm_pre_clip': 7.242705821990967, 'timing/data': 0.011465955001767725, 'timing/model': 1.6413156969938427, 'learning_rate/qwen_vl_interface': 5.249706751093274e-07, 'learning_rate/action_model': 5.249706751093273e-06, 'epoch': 2.08}) +[2026-06-16 08:35:29,556][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 0.10023342818021774, 'train/grad_norm_pre_clip': 10.517090797424316, 'timing/data': 0.01304437598446384, 'timing/model': 1.4726305269869044, 'learning_rate/qwen_vl_interface': 5.246597866004273e-07, 'learning_rate/action_model': 5.246597866004273e-06, 'epoch': 2.08}) +[2026-06-16 08:35:35,556][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.19828195869922638, 'train/grad_norm_pre_clip': 17.80797576904297, 'timing/data': 0.01308112102560699, 'timing/model': 1.4592093490064144, 'learning_rate/qwen_vl_interface': 5.243508405003975e-07, 'learning_rate/action_model': 5.243508405003975e-06, 'epoch': 2.08}) +[2026-06-16 08:35:41,544][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.0832807719707489, 'train/grad_norm_pre_clip': 9.347616195678711, 'timing/data': 0.013442381983622909, 'timing/model': 1.504375004034955, 'learning_rate/qwen_vl_interface': 5.240438369362336e-07, 'learning_rate/action_model': 5.2404383693623366e-06, 'epoch': 2.08}) +[2026-06-16 08:35:47,549][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.12196655571460724, 'train/grad_norm_pre_clip': 6.588201999664307, 'timing/data': 0.013268075999803841, 'timing/model': 1.5068349660141394, 'learning_rate/qwen_vl_interface': 5.237387760341335e-07, 'learning_rate/action_model': 5.237387760341335e-06, 'epoch': 2.08}) +[2026-06-16 08:35:53,489][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.2514299750328064, 'train/grad_norm_pre_clip': 8.633318901062012, 'timing/data': 0.013633993978146464, 'timing/model': 1.6868209330132231, 'learning_rate/qwen_vl_interface': 5.234356579194963e-07, 'learning_rate/action_model': 5.234356579194963e-06, 'epoch': 2.08}) +[2026-06-16 08:35:59,155][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.04501062631607056, 'train/grad_norm_pre_clip': 10.13783073425293, 'timing/data': 0.012580114009324461, 'timing/model': 1.4542429439607076, 'learning_rate/qwen_vl_interface': 5.231344827169218e-07, 'learning_rate/action_model': 5.231344827169217e-06, 'epoch': 2.09}) +[2026-06-16 08:36:05,155][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.08867572247982025, 'train/grad_norm_pre_clip': 11.021611213684082, 'timing/data': 0.012533345026895404, 'timing/model': 1.4550112769939005, 'learning_rate/qwen_vl_interface': 5.228352505502119e-07, 'learning_rate/action_model': 5.228352505502119e-06, 'epoch': 2.09}) +[2026-06-16 08:36:10,954][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.16009502112865448, 'train/grad_norm_pre_clip': 24.980072021484375, 'timing/data': 0.012467347958590835, 'timing/model': 1.456988897989504, 'learning_rate/qwen_vl_interface': 5.225379615423701e-07, 'learning_rate/action_model': 5.2253796154237e-06, 'epoch': 2.09}) +[2026-06-16 08:36:16,855][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.041542213410139084, 'train/grad_norm_pre_clip': 16.336183547973633, 'timing/data': 0.014454891963396221, 'timing/model': 1.4724908720236272, 'learning_rate/qwen_vl_interface': 5.222426158156008e-07, 'learning_rate/action_model': 5.222426158156008e-06, 'epoch': 2.09}) +[2026-06-16 08:36:22,751][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.08938565105199814, 'train/grad_norm_pre_clip': 9.500970840454102, 'timing/data': 0.012622426031157374, 'timing/model': 1.4733786100405268, 'learning_rate/qwen_vl_interface': 5.219492134913084e-07, 'learning_rate/action_model': 5.219492134913084e-06, 'epoch': 2.09}) +[2026-06-16 08:36:28,684][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.1209581270813942, 'train/grad_norm_pre_clip': 5.95564603805542, 'timing/data': 0.012932882993482053, 'timing/model': 1.6846549339825287, 'learning_rate/qwen_vl_interface': 5.216577546901008e-07, 'learning_rate/action_model': 5.216577546901007e-06, 'epoch': 2.09}) +[2026-06-16 08:36:34,350][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 0.1072586253285408, 'train/grad_norm_pre_clip': 12.894047737121582, 'timing/data': 0.01233414199668914, 'timing/model': 1.451363802945707, 'learning_rate/qwen_vl_interface': 5.213682395317852e-07, 'learning_rate/action_model': 5.213682395317851e-06, 'epoch': 2.09}) +[2026-06-16 08:36:40,259][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 0.1572158932685852, 'train/grad_norm_pre_clip': 9.433699607849121, 'timing/data': 0.01232096302555874, 'timing/model': 1.4612233779625967, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 2.09}) +[2026-06-16 08:36:46,150][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.013142488896846771, 'train/grad_norm_pre_clip': 1.6658196449279785, 'timing/data': 0.012302596005611122, 'timing/model': 1.4532169629819691, 'learning_rate/qwen_vl_interface': 5.207950406190654e-07, 'learning_rate/action_model': 5.207950406190653e-06, 'epoch': 2.09}) +[2026-06-16 08:36:52,051][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.060935765504837036, 'train/grad_norm_pre_clip': 6.563363075256348, 'timing/data': 0.012174026982393116, 'timing/model': 1.455854405008722, 'learning_rate/qwen_vl_interface': 5.205113571002818e-07, 'learning_rate/action_model': 5.205113571002818e-06, 'epoch': 2.09}) +[2026-06-16 08:36:57,961][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.07119183242321014, 'train/grad_norm_pre_clip': 12.499432563781738, 'timing/data': 0.012577309971675277, 'timing/model': 1.4729693519766442, 'learning_rate/qwen_vl_interface': 5.202296176956308e-07, 'learning_rate/action_model': 5.202296176956308e-06, 'epoch': 2.09}) +[2026-06-16 08:37:03,884][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 0.009778312407433987, 'train/grad_norm_pre_clip': 1.4054474830627441, 'timing/data': 0.012520045042037964, 'timing/model': 1.6836838009767234, 'learning_rate/qwen_vl_interface': 5.19949822520925e-07, 'learning_rate/action_model': 5.19949822520925e-06, 'epoch': 2.09}) +[2026-06-16 08:37:09,553][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.05356935039162636, 'train/grad_norm_pre_clip': 9.419928550720215, 'timing/data': 0.012570090009830892, 'timing/model': 1.4544056609738618, 'learning_rate/qwen_vl_interface': 5.196719716911777e-07, 'learning_rate/action_model': 5.1967197169117764e-06, 'epoch': 2.09}) +[2026-06-16 08:37:15,552][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.11087799817323685, 'train/grad_norm_pre_clip': 17.482257843017578, 'timing/data': 0.012350086006335914, 'timing/model': 1.4344821480335668, 'learning_rate/qwen_vl_interface': 5.193960653206029e-07, 'learning_rate/action_model': 5.193960653206029e-06, 'epoch': 2.09}) +[2026-06-16 08:37:21,451][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.021520575508475304, 'train/grad_norm_pre_clip': 3.6290435791015625, 'timing/data': 0.012386779999360442, 'timing/model': 1.451345824985765, 'learning_rate/qwen_vl_interface': 5.191221035226144e-07, 'learning_rate/action_model': 5.191221035226144e-06, 'epoch': 2.09}) +[2026-06-16 08:37:27,445][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.059566374868154526, 'train/grad_norm_pre_clip': 12.960860252380371, 'timing/data': 0.012551532010547817, 'timing/model': 1.5027976569836028, 'learning_rate/qwen_vl_interface': 5.188500864098292e-07, 'learning_rate/action_model': 5.1885008640982915e-06, 'epoch': 2.09}) +[2026-06-16 08:37:33,251][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.024012304842472076, 'train/grad_norm_pre_clip': 3.4796133041381836, 'timing/data': 0.012739558005705476, 'timing/model': 1.4521750080166385, 'learning_rate/qwen_vl_interface': 5.185800140940613e-07, 'learning_rate/action_model': 5.185800140940613e-06, 'epoch': 2.09}) +[2026-06-16 08:37:39,183][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 0.11900940537452698, 'train/grad_norm_pre_clip': 9.065629959106445, 'timing/data': 0.012463232036679983, 'timing/model': 1.6854592469753698, 'learning_rate/qwen_vl_interface': 5.183118866863288e-07, 'learning_rate/action_model': 5.1831188668632875e-06, 'epoch': 2.09}) +[2026-06-16 08:37:44,857][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.05902821943163872, 'train/grad_norm_pre_clip': 126.876220703125, 'timing/data': 0.012940893997438252, 'timing/model': 1.454440574045293, 'learning_rate/qwen_vl_interface': 5.180457042968484e-07, 'learning_rate/action_model': 5.1804570429684835e-06, 'epoch': 2.09}) +[2026-06-16 08:37:50,854][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.003381589660421014, 'train/grad_norm_pre_clip': 0.33690133690834045, 'timing/data': 0.012842189986258745, 'timing/model': 1.4515407910221256, 'learning_rate/qwen_vl_interface': 5.17781467035037e-07, 'learning_rate/action_model': 5.177814670350371e-06, 'epoch': 2.09}) +[2026-06-16 08:37:56,753][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.00950862467288971, 'train/grad_norm_pre_clip': 0.8922788500785828, 'timing/data': 0.012478974997065961, 'timing/model': 1.451924218039494, 'learning_rate/qwen_vl_interface': 5.175191750095133e-07, 'learning_rate/action_model': 5.175191750095132e-06, 'epoch': 2.09}) +[2026-06-16 08:38:02,550][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.19148047268390656, 'train/grad_norm_pre_clip': 21.592674255371094, 'timing/data': 0.013276318029966205, 'timing/model': 1.4518337870249525, 'learning_rate/qwen_vl_interface': 5.172588283280953e-07, 'learning_rate/action_model': 5.172588283280953e-06, 'epoch': 2.09}) +[2026-06-16 08:38:08,448][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.1705741137266159, 'train/grad_norm_pre_clip': 17.63473892211914, 'timing/data': 0.012944661022629589, 'timing/model': 1.4402096300036646, 'learning_rate/qwen_vl_interface': 5.170004270978019e-07, 'learning_rate/action_model': 5.170004270978018e-06, 'epoch': 2.09}) +[2026-06-16 08:38:14,382][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 0.023357324302196503, 'train/grad_norm_pre_clip': 2.802313804626465, 'timing/data': 0.012454069044906646, 'timing/model': 1.6842470390256494, 'learning_rate/qwen_vl_interface': 5.167439714248512e-07, 'learning_rate/action_model': 5.1674397142485116e-06, 'epoch': 2.1}) +[2026-06-16 08:38:20,152][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.16275367140769958, 'train/grad_norm_pre_clip': 12.007465362548828, 'timing/data': 0.013073018984869123, 'timing/model': 1.454723392962478, 'learning_rate/qwen_vl_interface': 5.164894614146633e-07, 'learning_rate/action_model': 5.164894614146634e-06, 'epoch': 2.1}) +[2026-06-16 08:38:26,054][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.005844296887516975, 'train/grad_norm_pre_clip': 0.6881507635116577, 'timing/data': 0.01314710796577856, 'timing/model': 1.4559464629855938, 'learning_rate/qwen_vl_interface': 5.16236897171858e-07, 'learning_rate/action_model': 5.16236897171858e-06, 'epoch': 2.1}) +[2026-06-16 08:38:31,951][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.06596998125314713, 'train/grad_norm_pre_clip': 16.162466049194336, 'timing/data': 0.013002264953684062, 'timing/model': 1.4537733480101451, 'learning_rate/qwen_vl_interface': 5.15986278800254e-07, 'learning_rate/action_model': 5.15986278800254e-06, 'epoch': 2.1}) +[2026-06-16 08:38:37,855][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.07333436608314514, 'train/grad_norm_pre_clip': 9.152284622192383, 'timing/data': 0.012704517983365804, 'timing/model': 1.452618421986699, 'learning_rate/qwen_vl_interface': 5.157376064028719e-07, 'learning_rate/action_model': 5.157376064028718e-06, 'epoch': 2.1}) +[2026-06-16 08:38:43,856][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.05849550664424896, 'train/grad_norm_pre_clip': 6.901041030883789, 'timing/data': 0.013019710953813046, 'timing/model': 1.5238960669958033, 'learning_rate/qwen_vl_interface': 5.154908800819307e-07, 'learning_rate/action_model': 5.154908800819306e-06, 'epoch': 2.1}) +[2026-06-16 08:38:49,785][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.1627255231142044, 'train/grad_norm_pre_clip': 18.814117431640625, 'timing/data': 0.013144585012923926, 'timing/model': 1.6860876580467448, 'learning_rate/qwen_vl_interface': 5.152460999388504e-07, 'learning_rate/action_model': 5.152460999388503e-06, 'epoch': 2.1}) +[2026-06-16 08:38:55,455][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.10675173252820969, 'train/grad_norm_pre_clip': 16.396167755126953, 'timing/data': 0.01322089700261131, 'timing/model': 1.45153752702754, 'learning_rate/qwen_vl_interface': 5.150032660742517e-07, 'learning_rate/action_model': 5.150032660742517e-06, 'epoch': 2.1}) +[2026-06-16 08:39:01,357][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.03639870136976242, 'train/grad_norm_pre_clip': 12.907951354980469, 'timing/data': 0.013290580012835562, 'timing/model': 1.453916587983258, 'learning_rate/qwen_vl_interface': 5.147623785879535e-07, 'learning_rate/action_model': 5.147623785879534e-06, 'epoch': 2.1}) +[2026-06-16 08:39:07,356][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.16775304079055786, 'train/grad_norm_pre_clip': 9.507529258728027, 'timing/data': 0.013006902008783072, 'timing/model': 1.45394633396063, 'learning_rate/qwen_vl_interface': 5.145234375789758e-07, 'learning_rate/action_model': 5.145234375789758e-06, 'epoch': 2.1}) +[2026-06-16 08:39:13,257][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.1364097148180008, 'train/grad_norm_pre_clip': 8.91690444946289, 'timing/data': 0.012741170998197049, 'timing/model': 1.4552533280220814, 'learning_rate/qwen_vl_interface': 5.142864431455387e-07, 'learning_rate/action_model': 5.1428644314553875e-06, 'epoch': 2.1}) +[2026-06-16 08:39:19,232][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.04746131971478462, 'train/grad_norm_pre_clip': 8.421372413635254, 'timing/data': 0.01229859102750197, 'timing/model': 1.4995268419734202, 'learning_rate/qwen_vl_interface': 5.140513953850612e-07, 'learning_rate/action_model': 5.140513953850612e-06, 'epoch': 2.1}) +[2026-06-16 08:39:25,182][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.09874429553747177, 'train/grad_norm_pre_clip': 11.678056716918945, 'timing/data': 0.013654154958203435, 'timing/model': 1.679622030002065, 'learning_rate/qwen_vl_interface': 5.138182943941623e-07, 'learning_rate/action_model': 5.138182943941623e-06, 'epoch': 2.1}) +[2026-06-16 08:39:30,854][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.1064685583114624, 'train/grad_norm_pre_clip': 9.568041801452637, 'timing/data': 0.013341103971470147, 'timing/model': 1.4506756089976989, 'learning_rate/qwen_vl_interface': 5.135871402686615e-07, 'learning_rate/action_model': 5.135871402686615e-06, 'epoch': 2.1}) +[2026-06-16 08:39:36,755][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.04788149893283844, 'train/grad_norm_pre_clip': 5.863098621368408, 'timing/data': 0.012851844017859548, 'timing/model': 1.4528048429638147, 'learning_rate/qwen_vl_interface': 5.133579331035777e-07, 'learning_rate/action_model': 5.133579331035777e-06, 'epoch': 2.1}) +[2026-06-16 08:39:42,654][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.016753224655985832, 'train/grad_norm_pre_clip': 2.406049966812134, 'timing/data': 0.011694852961227298, 'timing/model': 1.4580460640136153, 'learning_rate/qwen_vl_interface': 5.131306729931284e-07, 'learning_rate/action_model': 5.1313067299312836e-06, 'epoch': 2.1}) +[2026-06-16 08:39:48,552][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.02188105322420597, 'train/grad_norm_pre_clip': 4.350532054901123, 'timing/data': 0.013174789026379585, 'timing/model': 1.448903676995542, 'learning_rate/qwen_vl_interface': 5.129053600307326e-07, 'learning_rate/action_model': 5.129053600307326e-06, 'epoch': 2.1}) +[2026-06-16 08:39:54,508][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.015432192012667656, 'train/grad_norm_pre_clip': 2.581387758255005, 'timing/data': 0.012712048017419875, 'timing/model': 1.4794257449684665, 'learning_rate/qwen_vl_interface': 5.126819943090078e-07, 'learning_rate/action_model': 5.1268199430900775e-06, 'epoch': 2.1}) +[2026-06-16 08:40:00,378][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.13827283680438995, 'train/grad_norm_pre_clip': 15.028719902038574, 'timing/data': 0.012326241005212069, 'timing/model': 1.6801716989721172, 'learning_rate/qwen_vl_interface': 5.124605759197706e-07, 'learning_rate/action_model': 5.124605759197706e-06, 'epoch': 2.1}) +[2026-06-16 08:40:06,280][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.13591785728931427, 'train/grad_norm_pre_clip': 12.204483032226562, 'timing/data': 0.012426498986314982, 'timing/model': 1.6809091859613545, 'learning_rate/qwen_vl_interface': 5.122411049540383e-07, 'learning_rate/action_model': 5.1224110495403825e-06, 'epoch': 2.1}) +[2026-06-16 08:40:11,950][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.005352243315428495, 'train/grad_norm_pre_clip': 0.8560517430305481, 'timing/data': 0.012486908002756536, 'timing/model': 1.4505845469539054, 'learning_rate/qwen_vl_interface': 5.120235815020269e-07, 'learning_rate/action_model': 5.120235815020269e-06, 'epoch': 2.1}) +[2026-06-16 08:40:17,852][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.1243959367275238, 'train/grad_norm_pre_clip': 9.780485153198242, 'timing/data': 0.01320348196895793, 'timing/model': 1.4529642079724, 'learning_rate/qwen_vl_interface': 5.118080056531524e-07, 'learning_rate/action_model': 5.118080056531524e-06, 'epoch': 2.1}) +[2026-06-16 08:40:23,754][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.037660229951143265, 'train/grad_norm_pre_clip': 14.008776664733887, 'timing/data': 0.013786600029561669, 'timing/model': 1.441675204958301, 'learning_rate/qwen_vl_interface': 5.115943774960295e-07, 'learning_rate/action_model': 5.115943774960295e-06, 'epoch': 2.1}) +[2026-06-16 08:40:29,738][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.07506705820560455, 'train/grad_norm_pre_clip': 20.51785659790039, 'timing/data': 0.01262166496599093, 'timing/model': 1.5003397170221433, 'learning_rate/qwen_vl_interface': 5.113826971184727e-07, 'learning_rate/action_model': 5.113826971184727e-06, 'epoch': 2.1}) +[2026-06-16 08:40:35,701][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.089974544942379, 'train/grad_norm_pre_clip': 16.57923698425293, 'timing/data': 0.012686742004007101, 'timing/model': 1.7007521120249294, 'learning_rate/qwen_vl_interface': 5.111729646074954e-07, 'learning_rate/action_model': 5.111729646074954e-06, 'epoch': 2.11}) +[2026-06-16 08:40:41,351][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.11929488182067871, 'train/grad_norm_pre_clip': 16.61185646057129, 'timing/data': 0.01300732302479446, 'timing/model': 1.454030372959096, 'learning_rate/qwen_vl_interface': 5.10965180049312e-07, 'learning_rate/action_model': 5.10965180049312e-06, 'epoch': 2.11}) +[2026-06-16 08:40:47,250][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.2181078940629959, 'train/grad_norm_pre_clip': 14.127418518066406, 'timing/data': 0.012751336034853011, 'timing/model': 1.4513233620091341, 'learning_rate/qwen_vl_interface': 5.107593435293336e-07, 'learning_rate/action_model': 5.107593435293336e-06, 'epoch': 2.11}) +[2026-06-16 08:40:53,256][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.049691300839185715, 'train/grad_norm_pre_clip': 6.378232479095459, 'timing/data': 0.012976778030861169, 'timing/model': 1.4525808669859543, 'learning_rate/qwen_vl_interface': 5.105554551321724e-07, 'learning_rate/action_model': 5.105554551321723e-06, 'epoch': 2.11}) +[2026-06-16 08:40:59,230][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.05227665230631828, 'train/grad_norm_pre_clip': 9.485963821411133, 'timing/data': 0.012463829945772886, 'timing/model': 1.4987408259767108, 'learning_rate/qwen_vl_interface': 5.103535149416389e-07, 'learning_rate/action_model': 5.1035351494163885e-06, 'epoch': 2.11}) +[2026-06-16 08:41:05,150][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.025798795744776726, 'train/grad_norm_pre_clip': 5.743332862854004, 'timing/data': 0.012715349963400513, 'timing/model': 1.5056567360297777, 'learning_rate/qwen_vl_interface': 5.101535230407436e-07, 'learning_rate/action_model': 5.101535230407436e-06, 'epoch': 2.11}) +[2026-06-16 08:41:11,034][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.0019359844736754894, 'train/grad_norm_pre_clip': 0.5943434834480286, 'timing/data': 0.012698424980044365, 'timing/model': 1.6358213539933786, 'learning_rate/qwen_vl_interface': 5.099554795116956e-07, 'learning_rate/action_model': 5.0995547951169565e-06, 'epoch': 2.11}) +[2026-06-16 08:50:14,349][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.0656171441078186, 'train/grad_norm_pre_clip': 13.614215850830078, 'eval/loss': 0.04743180796504021, 'eval/flap_precision': 0.9473684210526315, 'eval/flap_recall': 0.9378757515030061, 'eval/flap_f1': 0.9425981873111783, 'eval/flap_rate_model': 0.13722222222222222, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9425981873111783, 'timing/data': 0.012890206999145448, 'timing/model': 1.4513924829661846, 'learning_rate/qwen_vl_interface': 5.097593844359024e-07, 'learning_rate/action_model': 5.0975938443590245e-06, 'epoch': 2.11}) +[2026-06-16 08:50:14,349][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4900_state +[2026-06-16 08:50:14,353][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 08:50:20,496][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4900_state/model.safetensors +[2026-06-16 08:50:28,519][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4900_state/optimizer.bin +[2026-06-16 08:50:28,520][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4900_state/sampler.bin +[2026-06-16 08:50:28,520][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4900_state/sampler_1.bin +[2026-06-16 08:50:28,521][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_4900_state/random_states_0.pkl +[2026-06-16 08:50:28,633][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4900 +[2026-06-16 08:50:28,634][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 08:50:28,643][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 08:50:34,741][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.00598464161157608, 'train/grad_norm_pre_clip': 0.8109943270683289, 'timing/data': 0.01259788399329409, 'timing/model': 1.5564254589844495, 'learning_rate/qwen_vl_interface': 5.095652378939718e-07, 'learning_rate/action_model': 5.095652378939717e-06, 'epoch': 2.11}) +[2026-06-16 08:50:40,854][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.007461670786142349, 'train/grad_norm_pre_clip': 0.9916667342185974, 'timing/data': 0.013292996969539672, 'timing/model': 1.5161217610002495, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 2.11}) +[2026-06-16 08:50:46,865][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.011360852047801018, 'train/grad_norm_pre_clip': 1.997484564781189, 'timing/data': 0.012583984003867954, 'timing/model': 1.5141264130361378, 'learning_rate/qwen_vl_interface': 5.091827907301227e-07, 'learning_rate/action_model': 5.091827907301227e-06, 'epoch': 2.11}) +[2026-06-16 08:50:52,753][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.003750766860321164, 'train/grad_norm_pre_clip': 0.7575343251228333, 'timing/data': 0.013114614994265139, 'timing/model': 1.4571590599953197, 'learning_rate/qwen_vl_interface': 5.089944902654134e-07, 'learning_rate/action_model': 5.089944902654134e-06, 'epoch': 2.11}) +[2026-06-16 08:50:58,760][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.07011616975069046, 'train/grad_norm_pre_clip': 7.676285266876221, 'timing/data': 0.012091820011846721, 'timing/model': 1.5280025139800273, 'learning_rate/qwen_vl_interface': 5.088081386489856e-07, 'learning_rate/action_model': 5.088081386489857e-06, 'epoch': 2.11}) +[2026-06-16 08:51:04,546][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.026757173240184784, 'train/grad_norm_pre_clip': 6.98668909072876, 'timing/data': 0.012691071024164557, 'timing/model': 1.4391868640086614, 'learning_rate/qwen_vl_interface': 5.086237359574419e-07, 'learning_rate/action_model': 5.086237359574419e-06, 'epoch': 2.11}) +[2026-06-16 08:51:10,482][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.05494827777147293, 'train/grad_norm_pre_clip': 10.2459716796875, 'timing/data': 0.013834893994498998, 'timing/model': 1.6849014919716865, 'learning_rate/qwen_vl_interface': 5.084412822665825e-07, 'learning_rate/action_model': 5.084412822665825e-06, 'epoch': 2.11}) +[2026-06-16 08:51:16,150][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.06335864961147308, 'train/grad_norm_pre_clip': 6.696113586425781, 'timing/data': 0.01247377198887989, 'timing/model': 1.4543823970016092, 'learning_rate/qwen_vl_interface': 5.082607776514081e-07, 'learning_rate/action_model': 5.082607776514081e-06, 'epoch': 2.11}) +[2026-06-16 08:51:22,151][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.22400829195976257, 'train/grad_norm_pre_clip': 16.012422561645508, 'timing/data': 0.012567128986120224, 'timing/model': 1.45231781702023, 'learning_rate/qwen_vl_interface': 5.080822221861167e-07, 'learning_rate/action_model': 5.080822221861166e-06, 'epoch': 2.11}) +[2026-06-16 08:51:28,152][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.057465117424726486, 'train/grad_norm_pre_clip': 14.760384559631348, 'timing/data': 0.01281583501258865, 'timing/model': 1.4472065459704027, 'learning_rate/qwen_vl_interface': 5.07905615944106e-07, 'learning_rate/action_model': 5.0790561594410595e-06, 'epoch': 2.11}) +[2026-06-16 08:51:34,159][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.2193431258201599, 'train/grad_norm_pre_clip': 15.047471046447754, 'timing/data': 0.015414458001032472, 'timing/model': 1.5327687940443866, 'learning_rate/qwen_vl_interface': 5.077309589979716e-07, 'learning_rate/action_model': 5.077309589979715e-06, 'epoch': 2.11}) +[2026-06-16 08:51:40,154][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.038458701223134995, 'train/grad_norm_pre_clip': 7.063968658447266, 'timing/data': 0.014470309019088745, 'timing/model': 1.4523072859738022, 'learning_rate/qwen_vl_interface': 5.075582514195089e-07, 'learning_rate/action_model': 5.0755825141950895e-06, 'epoch': 2.11}) +[2026-06-16 08:51:46,185][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 0.019784530624747276, 'train/grad_norm_pre_clip': 2.986525535583496, 'timing/data': 0.01376953802537173, 'timing/model': 1.71276595199015, 'learning_rate/qwen_vl_interface': 5.073874932797115e-07, 'learning_rate/action_model': 5.0738749327971155e-06, 'epoch': 2.11}) +[2026-06-16 08:51:51,856][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.06372018158435822, 'train/grad_norm_pre_clip': 5.557075023651123, 'timing/data': 0.012503686011768878, 'timing/model': 1.4559712019981816, 'learning_rate/qwen_vl_interface': 5.072186846487713e-07, 'learning_rate/action_model': 5.072186846487713e-06, 'epoch': 2.11}) +[2026-06-16 08:51:57,951][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 0.046921953558921814, 'train/grad_norm_pre_clip': 12.246204376220703, 'timing/data': 0.01187406899407506, 'timing/model': 1.4290518970228732, 'learning_rate/qwen_vl_interface': 5.070518255960796e-07, 'learning_rate/action_model': 5.070518255960795e-06, 'epoch': 2.11}) +[2026-06-16 08:52:04,169][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.37130069732666016, 'train/grad_norm_pre_clip': 14.829676628112793, 'timing/data': 0.013059033022727817, 'timing/model': 1.6739448019652627, 'learning_rate/qwen_vl_interface': 5.068869161902256e-07, 'learning_rate/action_model': 5.068869161902256e-06, 'epoch': 2.12}) +[2026-06-16 08:52:10,060][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.14664748311042786, 'train/grad_norm_pre_clip': 8.403399467468262, 'timing/data': 0.012220459000673145, 'timing/model': 1.434592705045361, 'learning_rate/qwen_vl_interface': 5.067239564989968e-07, 'learning_rate/action_model': 5.067239564989968e-06, 'epoch': 2.12}) +[2026-06-16 08:52:15,857][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 0.0076791271567344666, 'train/grad_norm_pre_clip': 1.456929326057434, 'timing/data': 0.012415245990268886, 'timing/model': 1.4577725949930027, 'learning_rate/qwen_vl_interface': 5.065629465893808e-07, 'learning_rate/action_model': 5.065629465893807e-06, 'epoch': 2.12}) +[2026-06-16 08:52:21,787][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.005155341234058142, 'train/grad_norm_pre_clip': 0.9609241485595703, 'timing/data': 0.012502263998612761, 'timing/model': 1.6842813029652461, 'learning_rate/qwen_vl_interface': 5.064038865275625e-07, 'learning_rate/action_model': 5.0640388652756245e-06, 'epoch': 2.12}) +[2026-06-16 08:52:27,765][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.03755934536457062, 'train/grad_norm_pre_clip': 9.366074562072754, 'timing/data': 0.013646116014569998, 'timing/model': 1.595163180027157, 'learning_rate/qwen_vl_interface': 5.062467763789242e-07, 'learning_rate/action_model': 5.062467763789242e-06, 'epoch': 2.12}) +[2026-06-16 08:52:33,756][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.11939579248428345, 'train/grad_norm_pre_clip': 7.177204608917236, 'timing/data': 0.012191246030852199, 'timing/model': 1.4469931009807624, 'learning_rate/qwen_vl_interface': 5.060916162080496e-07, 'learning_rate/action_model': 5.060916162080496e-06, 'epoch': 2.12}) +[2026-06-16 08:52:39,758][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.1915031224489212, 'train/grad_norm_pre_clip': 10.36959457397461, 'timing/data': 0.01239931199233979, 'timing/model': 1.4616827009595, 'learning_rate/qwen_vl_interface': 5.059384060787182e-07, 'learning_rate/action_model': 5.059384060787183e-06, 'epoch': 2.12}) +[2026-06-16 08:52:45,552][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 0.1692691147327423, 'train/grad_norm_pre_clip': 14.53580379486084, 'timing/data': 0.01199420599732548, 'timing/model': 1.4546695519820787, 'learning_rate/qwen_vl_interface': 5.05787146053909e-07, 'learning_rate/action_model': 5.0578714605390904e-06, 'epoch': 2.12}) +[2026-06-16 08:52:51,454][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.005040834192186594, 'train/grad_norm_pre_clip': 2.0446016788482666, 'timing/data': 0.012539387971628457, 'timing/model': 1.4446856009890325, 'learning_rate/qwen_vl_interface': 5.056378361957998e-07, 'learning_rate/action_model': 5.056378361957998e-06, 'epoch': 2.12}) +[2026-06-16 08:52:57,589][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.05711059272289276, 'train/grad_norm_pre_clip': 14.82824993133545, 'timing/data': 0.012662538967560977, 'timing/model': 1.6873378069722094, 'learning_rate/qwen_vl_interface': 5.05490476565766e-07, 'learning_rate/action_model': 5.054904765657659e-06, 'epoch': 2.12}) +[2026-06-16 08:53:03,255][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.019650183618068695, 'train/grad_norm_pre_clip': 3.68293833732605, 'timing/data': 0.01215276902075857, 'timing/model': 1.4536378280026838, 'learning_rate/qwen_vl_interface': 5.053450672243811e-07, 'learning_rate/action_model': 5.0534506722438114e-06, 'epoch': 2.12}) +[2026-06-16 08:53:09,354][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 0.07787690311670303, 'train/grad_norm_pre_clip': 8.67623519897461, 'timing/data': 0.013495733961462975, 'timing/model': 1.442526779952459, 'learning_rate/qwen_vl_interface': 5.052016082314184e-07, 'learning_rate/action_model': 5.052016082314184e-06, 'epoch': 2.12}) +[2026-06-16 08:53:15,254][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.009476544335484505, 'train/grad_norm_pre_clip': 2.529078722000122, 'timing/data': 0.013612176000606269, 'timing/model': 1.4518920159898698, 'learning_rate/qwen_vl_interface': 5.050600996458473e-07, 'learning_rate/action_model': 5.050600996458473e-06, 'epoch': 2.12}) +[2026-06-16 08:53:21,158][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 0.17672748863697052, 'train/grad_norm_pre_clip': 14.744185447692871, 'timing/data': 0.012397438986226916, 'timing/model': 1.4581408690428361, 'learning_rate/qwen_vl_interface': 5.049205415258373e-07, 'learning_rate/action_model': 5.049205415258373e-06, 'epoch': 2.12}) +[2026-06-16 08:53:27,139][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.18709774315357208, 'train/grad_norm_pre_clip': 18.1378116607666, 'timing/data': 0.012516985007096082, 'timing/model': 1.5008492580382153, 'learning_rate/qwen_vl_interface': 5.047829339287557e-07, 'learning_rate/action_model': 5.047829339287557e-06, 'epoch': 2.12}) +[2026-06-16 08:53:33,088][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.0039304932579398155, 'train/grad_norm_pre_clip': 0.6427643895149231, 'timing/data': 0.012443015992175788, 'timing/model': 1.6866715059732087, 'learning_rate/qwen_vl_interface': 5.046472769111668e-07, 'learning_rate/action_model': 5.046472769111667e-06, 'epoch': 2.12}) +[2026-06-16 08:53:38,758][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.16698254644870758, 'train/grad_norm_pre_clip': 18.025480270385742, 'timing/data': 0.012867400015238672, 'timing/model': 1.4567708889953792, 'learning_rate/qwen_vl_interface': 5.045135705288353e-07, 'learning_rate/action_model': 5.0451357052883525e-06, 'epoch': 2.12}) +[2026-06-16 08:53:44,756][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.25006377696990967, 'train/grad_norm_pre_clip': 23.06236457824707, 'timing/data': 0.012455545016564429, 'timing/model': 1.4526154829654843, 'learning_rate/qwen_vl_interface': 5.043818148367221e-07, 'learning_rate/action_model': 5.043818148367221e-06, 'epoch': 2.12}) +[2026-06-16 08:53:50,754][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.1217704713344574, 'train/grad_norm_pre_clip': 11.520034790039062, 'timing/data': 0.012309999961871654, 'timing/model': 1.451370761031285, 'learning_rate/qwen_vl_interface': 5.042520098889869e-07, 'learning_rate/action_model': 5.0425200988898685e-06, 'epoch': 2.12}) +[2026-06-16 08:53:56,846][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.014204532839357853, 'train/grad_norm_pre_clip': 2.1648168563842773, 'timing/data': 0.012796184979379177, 'timing/model': 1.5551354059716687, 'learning_rate/qwen_vl_interface': 5.041241557389879e-07, 'learning_rate/action_model': 5.041241557389879e-06, 'epoch': 2.12}) +[2026-06-16 08:54:02,871][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.2658771276473999, 'train/grad_norm_pre_clip': 12.38163948059082, 'timing/data': 0.012346153031103313, 'timing/model': 1.530507872987073, 'learning_rate/qwen_vl_interface': 5.039982524392814e-07, 'learning_rate/action_model': 5.039982524392814e-06, 'epoch': 2.12}) +[2026-06-16 08:54:08,882][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.1936948597431183, 'train/grad_norm_pre_clip': 15.110624313354492, 'timing/data': 0.013321767968591303, 'timing/model': 1.6775006239768118, 'learning_rate/qwen_vl_interface': 5.038743000416207e-07, 'learning_rate/action_model': 5.0387430004162064e-06, 'epoch': 2.12}) +[2026-06-16 08:54:14,558][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.09632154554128647, 'train/grad_norm_pre_clip': 11.298327445983887, 'timing/data': 0.013048216002061963, 'timing/model': 1.4527819310314953, 'learning_rate/qwen_vl_interface': 5.037522985969588e-07, 'learning_rate/action_model': 5.037522985969587e-06, 'epoch': 2.12}) +[2026-06-16 08:54:20,553][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.11267103999853134, 'train/grad_norm_pre_clip': 10.117106437683105, 'timing/data': 0.012464928033296019, 'timing/model': 1.449302970955614, 'learning_rate/qwen_vl_interface': 5.036322481554451e-07, 'learning_rate/action_model': 5.036322481554451e-06, 'epoch': 2.13}) +[2026-06-16 08:54:26,838][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.08003439754247665, 'train/grad_norm_pre_clip': 9.698549270629883, 'timing/data': 0.012944834015797824, 'timing/model': 1.5526307030231692, 'learning_rate/qwen_vl_interface': 5.035141487664285e-07, 'learning_rate/action_model': 5.035141487664285e-06, 'epoch': 2.13}) +[2026-06-16 08:54:33,022][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.08717488497495651, 'train/grad_norm_pre_clip': 8.008018493652344, 'timing/data': 0.012752290989737958, 'timing/model': 1.5390542980167083, 'learning_rate/qwen_vl_interface': 5.03398000478455e-07, 'learning_rate/action_model': 5.03398000478455e-06, 'epoch': 2.13}) +[2026-06-16 08:54:39,234][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.004594269674271345, 'train/grad_norm_pre_clip': 0.667877197265625, 'timing/data': 0.013493500999175012, 'timing/model': 1.5253664589836262, 'learning_rate/qwen_vl_interface': 5.032838033392686e-07, 'learning_rate/action_model': 5.032838033392685e-06, 'epoch': 2.13}) +[2026-06-16 08:54:45,455][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.04100511223077774, 'train/grad_norm_pre_clip': 3.6916370391845703, 'timing/data': 0.012579896952956915, 'timing/model': 1.745997645019088, 'learning_rate/qwen_vl_interface': 5.031715573958111e-07, 'learning_rate/action_model': 5.031715573958111e-06, 'epoch': 2.13}) +[2026-06-16 08:54:51,491][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.06208376958966255, 'train/grad_norm_pre_clip': 7.193927764892578, 'timing/data': 0.01241297199157998, 'timing/model': 1.7096039419993758, 'learning_rate/qwen_vl_interface': 5.030612626942235e-07, 'learning_rate/action_model': 5.030612626942235e-06, 'epoch': 2.13}) +[2026-06-16 08:54:57,232][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.04382915794849396, 'train/grad_norm_pre_clip': 8.805194854736328, 'timing/data': 0.012374034966342151, 'timing/model': 1.5278967480408028, 'learning_rate/qwen_vl_interface': 5.029529192798433e-07, 'learning_rate/action_model': 5.029529192798433e-06, 'epoch': 2.13}) +[2026-06-16 08:55:03,155][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.11431493610143661, 'train/grad_norm_pre_clip': 14.42368221282959, 'timing/data': 0.012581920018419623, 'timing/model': 1.454890882014297, 'learning_rate/qwen_vl_interface': 5.028465271972058e-07, 'learning_rate/action_model': 5.028465271972058e-06, 'epoch': 2.13}) +[2026-06-16 08:55:09,052][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.006368239410221577, 'train/grad_norm_pre_clip': 1.0969970226287842, 'timing/data': 0.012786781007889658, 'timing/model': 1.451195283967536, 'learning_rate/qwen_vl_interface': 5.027420864900457e-07, 'learning_rate/action_model': 5.027420864900456e-06, 'epoch': 2.13}) +[2026-06-16 08:55:15,116][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.050290707498788834, 'train/grad_norm_pre_clip': 9.475263595581055, 'timing/data': 0.01516018301481381, 'timing/model': 1.5772744560381398, 'learning_rate/qwen_vl_interface': 5.026395972012943e-07, 'learning_rate/action_model': 5.0263959720129436e-06, 'epoch': 2.13}) +[2026-06-16 08:55:21,010][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.05070773512125015, 'train/grad_norm_pre_clip': 7.909759998321533, 'timing/data': 0.012917024025227875, 'timing/model': 1.704739959037397, 'learning_rate/qwen_vl_interface': 5.025390593730809e-07, 'learning_rate/action_model': 5.025390593730809e-06, 'epoch': 2.13}) +[2026-06-16 08:55:26,787][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.01607704721391201, 'train/grad_norm_pre_clip': 2.325434923171997, 'timing/data': 0.012504646962042898, 'timing/model': 1.4848493110039271, 'learning_rate/qwen_vl_interface': 5.024404730467325e-07, 'learning_rate/action_model': 5.024404730467325e-06, 'epoch': 2.13}) +[2026-06-16 08:55:32,776][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.1787952333688736, 'train/grad_norm_pre_clip': 8.275745391845703, 'timing/data': 0.012349027034360915, 'timing/model': 1.4731406850041822, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 2.13}) +[2026-06-16 08:55:38,662][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.05412490665912628, 'train/grad_norm_pre_clip': 15.144161224365234, 'timing/data': 0.012821471958886832, 'timing/model': 1.4550854709814303, 'learning_rate/qwen_vl_interface': 5.022491550609311e-07, 'learning_rate/action_model': 5.022491550609311e-06, 'epoch': 2.13}) +[2026-06-16 08:55:44,674][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.15611445903778076, 'train/grad_norm_pre_clip': 6.217624664306641, 'timing/data': 0.012685892987065017, 'timing/model': 1.5444996429723687, 'learning_rate/qwen_vl_interface': 5.021564234801213e-07, 'learning_rate/action_model': 5.021564234801213e-06, 'epoch': 2.13}) +[2026-06-16 08:55:50,585][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.02324887551367283, 'train/grad_norm_pre_clip': 7.184616565704346, 'timing/data': 0.013150907005183399, 'timing/model': 1.5318541980232112, 'learning_rate/qwen_vl_interface': 5.020656435584636e-07, 'learning_rate/action_model': 5.020656435584635e-06, 'epoch': 2.13}) +[2026-06-16 08:55:56,474][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.008638128638267517, 'train/grad_norm_pre_clip': 1.3963731527328491, 'timing/data': 0.012643862050026655, 'timing/model': 1.6734208039706573, 'learning_rate/qwen_vl_interface': 5.019768153332752e-07, 'learning_rate/action_model': 5.019768153332752e-06, 'epoch': 2.13}) +[2026-06-16 08:56:02,157][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.008185111917555332, 'train/grad_norm_pre_clip': 1.3309170007705688, 'timing/data': 0.012394813995342702, 'timing/model': 1.456596328993328, 'learning_rate/qwen_vl_interface': 5.018899388410693e-07, 'learning_rate/action_model': 5.0188993884106935e-06, 'epoch': 2.13}) +[2026-06-16 08:56:08,156][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 0.07381212711334229, 'train/grad_norm_pre_clip': 10.52541732788086, 'timing/data': 0.012458026991225779, 'timing/model': 1.4544407269568183, 'learning_rate/qwen_vl_interface': 5.018050141175576e-07, 'learning_rate/action_model': 5.018050141175575e-06, 'epoch': 2.13}) +[2026-06-16 08:56:14,056][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.009105231612920761, 'train/grad_norm_pre_clip': 1.8332631587982178, 'timing/data': 0.01232232799520716, 'timing/model': 1.4567215309944004, 'learning_rate/qwen_vl_interface': 5.017220411976496e-07, 'learning_rate/action_model': 5.0172204119764965e-06, 'epoch': 2.13}) +[2026-06-16 08:56:20,035][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.29062920808792114, 'train/grad_norm_pre_clip': 19.416976928710938, 'timing/data': 0.012585415039211512, 'timing/model': 1.5022294389782473, 'learning_rate/qwen_vl_interface': 5.016410201154523e-07, 'learning_rate/action_model': 5.016410201154523e-06, 'epoch': 2.13}) +[2026-06-16 08:56:26,069][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.07346264272928238, 'train/grad_norm_pre_clip': 8.70451831817627, 'timing/data': 0.01256612699944526, 'timing/model': 1.4444541219854727, 'learning_rate/qwen_vl_interface': 5.015619509042709e-07, 'learning_rate/action_model': 5.015619509042709e-06, 'epoch': 2.13}) +[2026-06-16 08:56:31,962][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.20917266607284546, 'train/grad_norm_pre_clip': 32.97877883911133, 'timing/data': 0.012621307978406549, 'timing/model': 1.6542772839893587, 'learning_rate/qwen_vl_interface': 5.014848335966067e-07, 'learning_rate/action_model': 5.014848335966067e-06, 'epoch': 2.13}) +[2026-06-16 08:56:37,658][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 0.024070212617516518, 'train/grad_norm_pre_clip': 4.0517578125, 'timing/data': 0.012757729040458798, 'timing/model': 1.4538704120204784, 'learning_rate/qwen_vl_interface': 5.014096682241608e-07, 'learning_rate/action_model': 5.014096682241608e-06, 'epoch': 2.14}) +[2026-06-16 08:56:43,656][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.15357612073421478, 'train/grad_norm_pre_clip': 11.897218704223633, 'timing/data': 0.013009419024456292, 'timing/model': 1.4517773919505998, 'learning_rate/qwen_vl_interface': 5.013364548178299e-07, 'learning_rate/action_model': 5.013364548178299e-06, 'epoch': 2.14}) +[2026-06-16 08:56:49,665][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.04280763119459152, 'train/grad_norm_pre_clip': 12.175469398498535, 'timing/data': 0.012753241986501962, 'timing/model': 1.4673232509521767, 'learning_rate/qwen_vl_interface': 5.012651934077101e-07, 'learning_rate/action_model': 5.0126519340771e-06, 'epoch': 2.14}) +[2026-06-16 08:56:55,765][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 0.13274210691452026, 'train/grad_norm_pre_clip': 16.19558334350586, 'timing/data': 0.012306014017667621, 'timing/model': 1.5192828610306606, 'learning_rate/qwen_vl_interface': 5.01195884023094e-07, 'learning_rate/action_model': 5.0119588402309395e-06, 'epoch': 2.14}) +[2026-06-16 08:57:01,780][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.07559555023908615, 'train/grad_norm_pre_clip': 10.351578712463379, 'timing/data': 0.012233096989803016, 'timing/model': 1.777572953957133, 'learning_rate/qwen_vl_interface': 5.011285266924714e-07, 'learning_rate/action_model': 5.011285266924714e-06, 'epoch': 2.14}) +[2026-06-16 08:57:07,458][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.1435670554637909, 'train/grad_norm_pre_clip': 25.044485092163086, 'timing/data': 0.013881242019124329, 'timing/model': 1.4551664140308276, 'learning_rate/qwen_vl_interface': 5.010631214435312e-07, 'learning_rate/action_model': 5.010631214435312e-06, 'epoch': 2.14}) +[2026-06-16 08:57:13,454][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 0.05923088267445564, 'train/grad_norm_pre_clip': 12.14868450164795, 'timing/data': 0.012667515024077147, 'timing/model': 1.4524515279917978, 'learning_rate/qwen_vl_interface': 5.009996683031586e-07, 'learning_rate/action_model': 5.009996683031586e-06, 'epoch': 2.14}) +[2026-06-16 08:57:19,255][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.14314588904380798, 'train/grad_norm_pre_clip': 10.79556655883789, 'timing/data': 0.012877053988631815, 'timing/model': 1.4538697139942087, 'learning_rate/qwen_vl_interface': 5.009381672974375e-07, 'learning_rate/action_model': 5.009381672974375e-06, 'epoch': 2.14}) +[2026-06-16 08:57:25,153][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.21170812845230103, 'train/grad_norm_pre_clip': 19.57928466796875, 'timing/data': 0.012443096027709544, 'timing/model': 1.4532737669651397, 'learning_rate/qwen_vl_interface': 5.008786184516475e-07, 'learning_rate/action_model': 5.008786184516474e-06, 'epoch': 2.14}) +[2026-06-16 08:57:31,056][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 0.03482460603117943, 'train/grad_norm_pre_clip': 11.649054527282715, 'timing/data': 0.012510845030192286, 'timing/model': 1.4735810100100935, 'learning_rate/qwen_vl_interface': 5.008210217902676e-07, 'learning_rate/action_model': 5.008210217902676e-06, 'epoch': 2.14}) +[2026-06-16 08:57:36,986][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.017485927790403366, 'train/grad_norm_pre_clip': 6.447540760040283, 'timing/data': 0.012270601990167052, 'timing/model': 1.6844278569915332, 'learning_rate/qwen_vl_interface': 5.007653773369737e-07, 'learning_rate/action_model': 5.007653773369736e-06, 'epoch': 2.14}) +[2026-06-16 08:57:42,656][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.026781801134347916, 'train/grad_norm_pre_clip': 10.108412742614746, 'timing/data': 0.012471858004573733, 'timing/model': 1.4522008749772795, 'learning_rate/qwen_vl_interface': 5.007116851146386e-07, 'learning_rate/action_model': 5.007116851146386e-06, 'epoch': 2.14}) +[2026-06-16 08:57:48,653][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.00842798687517643, 'train/grad_norm_pre_clip': 2.08353853225708, 'timing/data': 0.012269038998056203, 'timing/model': 1.4522166380193084, 'learning_rate/qwen_vl_interface': 5.006599451453334e-07, 'learning_rate/action_model': 5.006599451453334e-06, 'epoch': 2.14}) +[2026-06-16 08:57:54,556][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.18935994803905487, 'train/grad_norm_pre_clip': 19.10979652404785, 'timing/data': 0.01235595700563863, 'timing/model': 1.44782765902346, 'learning_rate/qwen_vl_interface': 5.006101574503268e-07, 'learning_rate/action_model': 5.006101574503267e-06, 'epoch': 2.14}) +[2026-06-16 08:58:00,457][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.0708092525601387, 'train/grad_norm_pre_clip': 10.628561019897461, 'timing/data': 0.012470435991417617, 'timing/model': 1.4548654719837941, 'learning_rate/qwen_vl_interface': 5.005623220500839e-07, 'learning_rate/action_model': 5.005623220500839e-06, 'epoch': 2.14}) +[2026-06-16 08:58:06,354][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.16729208827018738, 'train/grad_norm_pre_clip': 12.948747634887695, 'timing/data': 0.012575579981785268, 'timing/model': 1.4646992880152538, 'learning_rate/qwen_vl_interface': 5.005164389642687e-07, 'learning_rate/action_model': 5.005164389642687e-06, 'epoch': 2.14}) +[2026-06-16 08:58:12,292][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.07506930828094482, 'train/grad_norm_pre_clip': 5.9822587966918945, 'timing/data': 0.012381846026983112, 'timing/model': 1.6892640449805185, 'learning_rate/qwen_vl_interface': 5.004725082117423e-07, 'learning_rate/action_model': 5.004725082117422e-06, 'epoch': 2.14}) +[2026-06-16 08:58:17,957][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.1568799912929535, 'train/grad_norm_pre_clip': 9.311779022216797, 'timing/data': 0.013395496003795415, 'timing/model': 1.4520133640035056, 'learning_rate/qwen_vl_interface': 5.004305298105616e-07, 'learning_rate/action_model': 5.004305298105616e-06, 'epoch': 2.14}) +[2026-06-16 08:58:23,956][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.16817045211791992, 'train/grad_norm_pre_clip': 8.961698532104492, 'timing/data': 0.012287446996197104, 'timing/model': 1.4537966569769196, 'learning_rate/qwen_vl_interface': 5.003905037779834e-07, 'learning_rate/action_model': 5.0039050377798345e-06, 'epoch': 2.14}) +[2026-06-16 08:58:30,084][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.199232280254364, 'train/grad_norm_pre_clip': 18.318187713623047, 'timing/data': 0.013565466972067952, 'timing/model': 1.511941634991672, 'learning_rate/qwen_vl_interface': 5.003524301304608e-07, 'learning_rate/action_model': 5.003524301304608e-06, 'epoch': 2.14}) +[2026-06-16 08:58:36,040][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.039758551865816116, 'train/grad_norm_pre_clip': 5.466022968292236, 'timing/data': 0.011484695016406476, 'timing/model': 1.5036556639824994, 'learning_rate/qwen_vl_interface': 5.003163088836445e-07, 'learning_rate/action_model': 5.003163088836446e-06, 'epoch': 2.14}) +[2026-06-16 08:58:41,968][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.1971478909254074, 'train/grad_norm_pre_clip': 24.655317306518555, 'timing/data': 0.012518025992903858, 'timing/model': 1.4615177129744552, 'learning_rate/qwen_vl_interface': 5.002821400523818e-07, 'learning_rate/action_model': 5.002821400523818e-06, 'epoch': 2.14}) +[2026-06-16 08:58:47,888][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 0.014586446806788445, 'train/grad_norm_pre_clip': 3.9703776836395264, 'timing/data': 0.012405561050400138, 'timing/model': 1.6801465460448526, 'learning_rate/qwen_vl_interface': 5.002499236507192e-07, 'learning_rate/action_model': 5.0024992365071925e-06, 'epoch': 2.14}) +[2026-06-16 08:58:53,636][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.02766125276684761, 'train/grad_norm_pre_clip': 5.860057830810547, 'timing/data': 0.013226197974290699, 'timing/model': 1.5053081219666637, 'learning_rate/qwen_vl_interface': 5.002196596918993e-07, 'learning_rate/action_model': 5.002196596918993e-06, 'epoch': 2.14}) +[2026-06-16 08:58:59,659][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.24158963561058044, 'train/grad_norm_pre_clip': 15.146860122680664, 'timing/data': 0.014820072043221444, 'timing/model': 1.4514283870230429, 'learning_rate/qwen_vl_interface': 5.001913481883623e-07, 'learning_rate/action_model': 5.001913481883623e-06, 'epoch': 2.15}) +[2026-06-16 08:59:05,655][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.09386716037988663, 'train/grad_norm_pre_clip': 12.851141929626465, 'timing/data': 0.012274136999621987, 'timing/model': 1.4531435400131159, 'learning_rate/qwen_vl_interface': 5.001649891517459e-07, 'learning_rate/action_model': 5.001649891517459e-06, 'epoch': 2.15}) +[2026-06-16 08:59:11,557][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.07210026681423187, 'train/grad_norm_pre_clip': 7.354985237121582, 'timing/data': 0.014055186999030411, 'timing/model': 1.4540615999721922, 'learning_rate/qwen_vl_interface': 5.001405825928858e-07, 'learning_rate/action_model': 5.0014058259288575e-06, 'epoch': 2.15}) +[2026-06-16 08:59:17,538][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.2520427703857422, 'train/grad_norm_pre_clip': 13.95708179473877, 'timing/data': 0.012257281981874257, 'timing/model': 1.5041315489797853, 'learning_rate/qwen_vl_interface': 5.001181285218143e-07, 'learning_rate/action_model': 5.0011812852181434e-06, 'epoch': 2.15}) +[2026-06-16 08:59:23,486][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.021863840520381927, 'train/grad_norm_pre_clip': 3.4404284954071045, 'timing/data': 0.012055676023010164, 'timing/model': 1.6789439509739168, 'learning_rate/qwen_vl_interface': 5.000976269477613e-07, 'learning_rate/action_model': 5.0009762694776125e-06, 'epoch': 2.15}) +[2026-06-16 08:59:29,158][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.010657845064997673, 'train/grad_norm_pre_clip': 3.110590934753418, 'timing/data': 0.012419531005434692, 'timing/model': 1.4554291589884087, 'learning_rate/qwen_vl_interface': 5.000790778791542e-07, 'learning_rate/action_model': 5.000790778791543e-06, 'epoch': 2.15}) +[2026-06-16 08:59:35,157][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.007453525438904762, 'train/grad_norm_pre_clip': 1.624962568283081, 'timing/data': 0.012499056989327073, 'timing/model': 1.4543795639765449, 'learning_rate/qwen_vl_interface': 5.000624813236185e-07, 'learning_rate/action_model': 5.000624813236185e-06, 'epoch': 2.15}) +[2026-06-16 08:59:41,059][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.031027816236019135, 'train/grad_norm_pre_clip': 3.9860928058624268, 'timing/data': 0.01251577294897288, 'timing/model': 1.4524527189787477, 'learning_rate/qwen_vl_interface': 5.000478372879752e-07, 'learning_rate/action_model': 5.000478372879752e-06, 'epoch': 2.15}) +[2026-06-16 08:59:46,958][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.008672542870044708, 'train/grad_norm_pre_clip': 2.6653475761413574, 'timing/data': 0.012398059014230967, 'timing/model': 1.452705750009045, 'learning_rate/qwen_vl_interface': 5.000351457782452e-07, 'learning_rate/action_model': 5.000351457782451e-06, 'epoch': 2.15}) +[2026-06-16 08:59:52,966][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 0.04281890392303467, 'train/grad_norm_pre_clip': 8.173797607421875, 'timing/data': 0.01271343301050365, 'timing/model': 1.5286204399890266, 'learning_rate/qwen_vl_interface': 5.000244067996445e-07, 'learning_rate/action_model': 5.000244067996445e-06, 'epoch': 2.15}) +[2026-06-16 08:59:58,887][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.21193952858448029, 'train/grad_norm_pre_clip': 16.404848098754883, 'timing/data': 0.012296119006350636, 'timing/model': 1.685740095039364, 'learning_rate/qwen_vl_interface': 5.000156203565884e-07, 'learning_rate/action_model': 5.000156203565883e-06, 'epoch': 2.15}) +[2026-06-16 09:00:04,555][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.07894277572631836, 'train/grad_norm_pre_clip': 12.01097583770752, 'timing/data': 0.012704260007012635, 'timing/model': 1.453231121005956, 'learning_rate/qwen_vl_interface': 5.000087864526878e-07, 'learning_rate/action_model': 5.000087864526878e-06, 'epoch': 2.15}) +[2026-06-16 09:00:10,500][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.003192906267940998, 'train/grad_norm_pre_clip': 0.3960980176925659, 'timing/data': 0.013187970966100693, 'timing/model': 1.4963987459777854, 'learning_rate/qwen_vl_interface': 5.000039050907523e-07, 'learning_rate/action_model': 5.000039050907523e-06, 'epoch': 2.15}) +[2026-06-16 09:00:16,458][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.2534717619419098, 'train/grad_norm_pre_clip': 19.535655975341797, 'timing/data': 0.012518486008048058, 'timing/model': 1.4505255190306343, 'learning_rate/qwen_vl_interface': 5.000009762727885e-07, 'learning_rate/action_model': 5.000009762727884e-06, 'epoch': 2.15}) +[2026-06-16 09:09:20,447][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.047330863773822784, 'train/grad_norm_pre_clip': 10.813126564025879, 'eval/loss': 0.04780582711100578, 'eval/flap_precision': 0.9435483870967742, 'eval/flap_recall': 0.9378757515030061, 'eval/flap_f1': 0.9407035175879397, 'eval/flap_rate_model': 0.13777777777777778, 'eval/flap_rate_teacher': 0.1386111111111111, 'eval/flap_cc_f1': 0.9407035175879397, 'timing/data': 0.012459589983336627, 'timing/model': 1.453433986986056, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 2.15}) +[2026-06-16 09:09:20,447][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_5000_state +[2026-06-16 09:09:20,452][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 09:09:25,922][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_5000_state/model.safetensors +[2026-06-16 09:09:39,881][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_5000_state/optimizer.bin +[2026-06-16 09:09:39,882][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_5000_state/sampler.bin +[2026-06-16 09:09:39,882][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_5000_state/sampler_1.bin +[2026-06-16 09:09:39,883][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints/steps_5000_state/random_states_0.pkl +[2026-06-16 09:09:40,023][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-06-16 09:09:40,023][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 09:09:40,033][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 09:09:40,033][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /workspace/starVLA/results/Checkpoints/flappy_fix_latency_1_200ep/checkpoints +[2026-06-16 09:09:41,998][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..ffa1ed16fa7cd08f0c83739d659e9d2a8bd4e827 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_1_200ep/summary.jsonl @@ -0,0 +1,50 @@ +{"steps": 100} +{"steps": 200} +{"steps": 300} +{"steps": 400} +{"steps": 500} +{"steps": 600} +{"steps": 700} +{"steps": 800} +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500} +{"steps": 1600} +{"steps": 1700} +{"steps": 1800} +{"steps": 1900} +{"steps": 2000} +{"steps": 2100} +{"steps": 2200} +{"steps": 2300} +{"steps": 2400} +{"steps": 2500} +{"steps": 2600} +{"steps": 2700} +{"steps": 2800} +{"steps": 2900} +{"steps": 3000} +{"steps": 3100} +{"steps": 3200} +{"steps": 3300} +{"steps": 3400} +{"steps": 3500} +{"steps": 3600} +{"steps": 3700} +{"steps": 3800} +{"steps": 3900} +{"steps": 4000} +{"steps": 4100} +{"steps": 4200} +{"steps": 4300} +{"steps": 4400} +{"steps": 4500} +{"steps": 4600} +{"steps": 4700} +{"steps": 4800} +{"steps": 4900} +{"steps": 5000} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/checkpoints/steps_5000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/checkpoints/steps_5000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..565c43d6aa94ca639ff799db226d20175005285e --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/checkpoints/steps_5000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:aabcc8e3a22947e22d161d324de0e613a03300f39fa1dd28bffa2df927e994cf +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..c9ead13be69443dbe94e6dc526e5f30bed9c8b94 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/config.full.yaml @@ -0,0 +1,199 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_fix_latency_2_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: null + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_2_200ep +output_dir: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..afc1f3c511268717a97d95ad6d363c20fd2df66d --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/config.yaml @@ -0,0 +1,71 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep +rl_games: + env_eval: + enabled: true + mid_train: + enabled: false + interval_steps: 100 + task: flappy +run_id: flappy_fix_latency_2_200ep +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_action_classification: true + eval_interval: 100 + eval_num_batches: 100 + freeze_modules: '' + gradient_accumulation_steps: 4 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + pretrained_checkpoint: null + reload_modules: null + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..aa03e6cbf8e375d660528e17fafddb4804b596b7 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.6028500199317932, + 0.3971499800682068, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4890792667865753, + 0.4890792667865753, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 330734, + "num_trajectories": 180 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..0caa254376d5805f972b0b0411e0b5188e3634e3 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.7959861159324646, + 0.2040138840675354, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4030573070049286, + 0.4030573070049286, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 72000, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/hydra/.hydra/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..2357031161be8f029217a695f78ed8ed6593785c --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/hydra/.hydra/config.yaml @@ -0,0 +1,197 @@ +framework: + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_fix_latency_2_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + config_name: null + source_subdir: null + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: null + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_2_200ep +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/hydra/.hydra/hydra.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..4c639eaba6f15c5c408940c250790500bc0eff49 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/hydra/.hydra/hydra.yaml @@ -0,0 +1,252 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=single + - run_id=flappy_fix_latency_2_200ep + - run_root_dir=/workspace/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=false + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=5000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=100 + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=4 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9, 0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=2 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.shuffle=true + - datasets.vla_data.action_balance.enabled=false + - datasets.vla_data.action_balance.strategy=balanced_epoch + - datasets.vla_data.action_balance.action_key=action_id + - datasets.vla_data.action_balance.target_flap_fraction=0.3 + - datasets.vla_data.action_balance.noop_id=0 + - datasets.vla_data.action_balance.flap_id=1 + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_2_200ep + - datasets.vla_data.data_mix=flappy_train__bridge + - datasets.vla_data.eval_data_mix=flappy_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=flappy + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=1 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.mid_train.enabled=false + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[0] + - rl_games.env_eval.post_train.enabled=false + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=false,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_balance.action_key=action_id,datasets.vla_data.action_balance.enabled=false,datasets.vla_data.action_balance.flap_id=1,datasets.vla_data.action_balance.noop_id=0,datasets.vla_data.action_balance.strategy=balanced_epoch,datasets.vla_data.action_balance.target_flap_fraction=0.3,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=flappy_train__bridge,datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_2_200ep,datasets.vla_data.eval_data_mix=flappy_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.shuffle=true,datasets.vla_data.video_backend=torchvision_av,env=flappy,framework.action_model.action_dim=7,framework.action_model.action_env_dim=2,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=1,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=false,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[0],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=false,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=flappy,run_id=flappy_fix_latency_2_200ep,run_root_dir=/workspace/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=100,trainer.gradient_accumulation_steps=4,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=5000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9, + 0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.3 + version_base: '1.1' + cwd: /workspace/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/hydra/.hydra/overrides.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..fecde000e7fd19e53e0459001ab577a301a11fd5 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/hydra/.hydra/overrides.yaml @@ -0,0 +1,91 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=single +- run_id=flappy_fix_latency_2_200ep +- run_root_dir=/workspace/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=false +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=5000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=100 +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=4 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9, 0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=2 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.shuffle=true +- datasets.vla_data.action_balance.enabled=false +- datasets.vla_data.action_balance.strategy=balanced_epoch +- datasets.vla_data.action_balance.action_key=action_id +- datasets.vla_data.action_balance.target_flap_fraction=0.3 +- datasets.vla_data.action_balance.noop_id=0 +- datasets.vla_data.action_balance.flap_id=1 +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_2_200ep +- datasets.vla_data.data_mix=flappy_train__bridge +- datasets.vla_data.eval_data_mix=flappy_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=flappy +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=1 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.mid_train.enabled=false +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[0] +- rl_games.env_eval.post_train.enabled=false +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/hydra/train_starvla_hydra.log b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..7ddf8d0b55283951b0f2cc0eae2a4ace0a2bfa7b --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/hydra/train_starvla_hydra.log @@ -0,0 +1,5552 @@ +[2026-06-15 18:45:48,141][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-15 18:45:48,160][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-15 18:46:00,577][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-15 18:46:00,621][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-15 18:46:01,460][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-15 18:46:01,460][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-15 18:46:01,460][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-15 18:46:01,460][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-15 18:46:01,461][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-15 18:46:01,461][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-15 18:46:01,461][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-15 18:46:01,461][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-15 18:46:01,461][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-15 18:46:01,461][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-15 18:46:01,461][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-15 18:46:01,461][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-15 18:46:01,461][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-15 18:46:01,462][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-15 18:46:01,462][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-15 18:46:01,462][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-15 18:46:01,462][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-15 18:46:01,462][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-15 18:46:01,462][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-15 18:46:01,462][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-15 18:46:01,462][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-15 18:46:01,463][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-15 18:46:01,463][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-15 18:46:01,463][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-15 18:46:01,463][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-15 18:46:01,463][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-15 18:46:01,463][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-15 18:46:01,463][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-15 18:46:01,463][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-15 18:46:01,463][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-15 18:46:01,464][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-15 18:46:01,528][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-15 18:46:01,572][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/config.full.yaml +[2026-06-15 18:46:01,587][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/config.yaml +[2026-06-15 18:46:01,587][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-15 18:46:38,719][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-15 18:46:38,719][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-06-15 18:46:38,720][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-15 18:46:38,720][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 4 +[2026-06-15 18:46:38,720][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-15 18:47:10,650][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 0.6953649520874023, 'train/grad_norm_pre_clip': 5.6823954582214355, 'timing/data': 0.011559074744582176, 'timing/model': 1.2822359800338745, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-15 18:47:15,517][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 0.6938610076904297, 'train/grad_norm_pre_clip': 0.8743993043899536, 'timing/data': 0.013546345755457878, 'timing/model': 1.2506431695073843, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-15 18:47:20,520][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 0.6918092966079712, 'train/grad_norm_pre_clip': 2.3196492195129395, 'timing/data': 0.014509351924061775, 'timing/model': 1.4448218941688538, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-06-15 18:47:25,736][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 0.6913321018218994, 'train/grad_norm_pre_clip': 0.6348549723625183, 'timing/data': 0.012306047603487968, 'timing/model': 1.4579042065888643, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-06-15 18:47:30,518][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 0.6878237128257751, 'train/grad_norm_pre_clip': 0.8284586668014526, 'timing/data': 0.015466950833797455, 'timing/model': 1.2456006594002247, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-06-15 18:47:35,502][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 0.6514891982078552, 'train/grad_norm_pre_clip': 4.231412887573242, 'timing/data': 0.019389593973755836, 'timing/model': 1.30930351652205, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-06-15 18:47:40,381][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 0.6654630303382874, 'train/grad_norm_pre_clip': 1.069857120513916, 'timing/data': 0.012927768751978874, 'timing/model': 1.2172320205718279, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-06-15 18:47:45,560][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.7075253129005432, 'train/grad_norm_pre_clip': 2.2960171699523926, 'timing/data': 0.01418025977909565, 'timing/model': 1.4847473204135895, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-06-15 18:47:51,054][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 0.6252374053001404, 'train/grad_norm_pre_clip': 1.131171464920044, 'timing/data': 0.013530390337109566, 'timing/model': 1.3819865565747023, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-06-15 18:47:55,925][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.733879029750824, 'train/grad_norm_pre_clip': 3.6766347885131836, 'timing/data': 0.014141622930765152, 'timing/model': 1.2570607792586088, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-15 18:48:00,955][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.6710817813873291, 'train/grad_norm_pre_clip': 1.7728900909423828, 'timing/data': 0.015304932370781898, 'timing/model': 1.4781273622065783, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-06-15 18:48:05,745][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.7524063587188721, 'train/grad_norm_pre_clip': 4.326694965362549, 'timing/data': 0.012467917054891586, 'timing/model': 1.2608650904148817, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0}) +[2026-06-15 18:48:10,559][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.6723101139068604, 'train/grad_norm_pre_clip': 1.8448882102966309, 'timing/data': 0.012626003473997116, 'timing/model': 1.2524420637637377, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.01}) +[2026-06-15 18:48:15,537][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.5964439511299133, 'train/grad_norm_pre_clip': 0.939233124256134, 'timing/data': 0.012875735759735107, 'timing/model': 1.2600756119936705, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.01}) +[2026-06-15 18:48:20,533][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.6197341680526733, 'train/grad_norm_pre_clip': 0.4757549464702606, 'timing/data': 0.01642744615674019, 'timing/model': 1.320472227409482, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.01}) +[2026-06-15 18:48:25,642][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.6202621459960938, 'train/grad_norm_pre_clip': 0.4004499316215515, 'timing/data': 0.012658750638365746, 'timing/model': 1.3602932691574097, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 18:48:30,452][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.7857952117919922, 'train/grad_norm_pre_clip': 5.570484638214111, 'timing/data': 0.012658411636948586, 'timing/model': 1.2733061630278826, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 18:48:35,445][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.7518805861473083, 'train/grad_norm_pre_clip': 4.5512800216674805, 'timing/data': 0.011914003640413284, 'timing/model': 1.3343912325799465, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.01}) +[2026-06-15 18:48:40,237][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.6969744563102722, 'train/grad_norm_pre_clip': 2.1569743156433105, 'timing/data': 0.01271681860089302, 'timing/model': 1.2517464086413383, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.01}) +[2026-06-15 18:48:45,132][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.6718900203704834, 'train/grad_norm_pre_clip': 0.3864869773387909, 'timing/data': 0.01331750676035881, 'timing/model': 1.3263100050389767, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-15 18:48:49,943][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.6676563620567322, 'train/grad_norm_pre_clip': 1.7279311418533325, 'timing/data': 0.012701323255896568, 'timing/model': 1.2149568181484938, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.01}) +[2026-06-15 18:48:54,744][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.6903775334358215, 'train/grad_norm_pre_clip': 0.47071388363838196, 'timing/data': 0.013116434216499329, 'timing/model': 1.2560089826583862, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 18:48:59,760][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.6776753067970276, 'train/grad_norm_pre_clip': 5.428708553314209, 'timing/data': 0.012577719986438751, 'timing/model': 1.2803018912672997, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 18:49:04,708][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.6775094270706177, 'train/grad_norm_pre_clip': 5.533832550048828, 'timing/data': 0.015069495886564255, 'timing/model': 1.300712838768959, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.01}) +[2026-06-15 18:49:09,660][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.6746218800544739, 'train/grad_norm_pre_clip': 2.683002233505249, 'timing/data': 0.012991378083825111, 'timing/model': 1.248488748446107, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.01}) +[2026-06-15 18:49:14,546][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.6967321038246155, 'train/grad_norm_pre_clip': 2.334792375564575, 'timing/data': 0.012585312128067017, 'timing/model': 1.2570262011140585, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.01}) +[2026-06-15 18:49:19,345][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.6857238411903381, 'train/grad_norm_pre_clip': 1.2251204252243042, 'timing/data': 0.013017157092690468, 'timing/model': 1.252176258713007, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.01}) +[2026-06-15 18:49:24,243][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.6557568311691284, 'train/grad_norm_pre_clip': 1.8318887948989868, 'timing/data': 0.01304721087217331, 'timing/model': 1.3443336710333824, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 18:49:29,045][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.6433625221252441, 'train/grad_norm_pre_clip': 2.135551929473877, 'timing/data': 0.01278768852353096, 'timing/model': 1.247783076018095, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.01}) +[2026-06-15 18:49:33,971][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.6446296572685242, 'train/grad_norm_pre_clip': 0.7154413461685181, 'timing/data': 0.013050360605120659, 'timing/model': 1.2639287300407887, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-15 18:49:39,244][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.6209415793418884, 'train/grad_norm_pre_clip': 1.5216808319091797, 'timing/data': 0.01450185477733612, 'timing/model': 1.552803022786975, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.01}) +[2026-06-15 18:49:44,254][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.5823749303817749, 'train/grad_norm_pre_clip': 1.0392240285873413, 'timing/data': 0.012643452733755112, 'timing/model': 1.3399669267237186, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.01}) +[2026-06-15 18:49:50,134][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.8198704719543457, 'train/grad_norm_pre_clip': 7.029882907867432, 'timing/data': 0.01211131177842617, 'timing/model': 1.8410166390240192, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.01}) +[2026-06-15 18:49:55,142][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.6511622071266174, 'train/grad_norm_pre_clip': 3.7866361141204834, 'timing/data': 0.013520075008273125, 'timing/model': 1.3458202257752419, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.01}) +[2026-06-15 18:49:59,944][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.7445846199989319, 'train/grad_norm_pre_clip': 6.066600322723389, 'timing/data': 0.012042736634612083, 'timing/model': 1.2365891318768263, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.01}) +[2026-06-15 18:50:04,745][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.6815060377120972, 'train/grad_norm_pre_clip': 5.200607776641846, 'timing/data': 0.012295080348849297, 'timing/model': 1.2518440764397383, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.01}) +[2026-06-15 18:50:09,576][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.7208424210548401, 'train/grad_norm_pre_clip': 5.967237949371338, 'timing/data': 0.01326294057071209, 'timing/model': 1.2649025861173868, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.01}) +[2026-06-15 18:50:14,757][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.6095817685127258, 'train/grad_norm_pre_clip': 1.8836207389831543, 'timing/data': 0.012534037232398987, 'timing/model': 1.262587457895279, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.01}) +[2026-06-15 18:50:19,644][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.5857824683189392, 'train/grad_norm_pre_clip': 6.637852191925049, 'timing/data': 0.012330971658229828, 'timing/model': 1.2341781072318554, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 18:50:24,459][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.6153339743614197, 'train/grad_norm_pre_clip': 7.105679512023926, 'timing/data': 0.012677060440182686, 'timing/model': 1.2353077754378319, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.02}) +[2026-06-15 18:50:29,348][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.6213220953941345, 'train/grad_norm_pre_clip': 17.6669864654541, 'timing/data': 0.013309309259057045, 'timing/model': 1.2506171222776175, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.02}) +[2026-06-15 18:50:34,156][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.7150598764419556, 'train/grad_norm_pre_clip': 6.761688709259033, 'timing/data': 0.012407416477799416, 'timing/model': 1.2551690880209208, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.02}) +[2026-06-15 18:50:39,086][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.6983003616333008, 'train/grad_norm_pre_clip': 10.691108703613281, 'timing/data': 0.012635864317417145, 'timing/model': 1.2739309947937727, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.02}) +[2026-06-15 18:50:44,064][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.5337405800819397, 'train/grad_norm_pre_clip': 5.592596530914307, 'timing/data': 0.015295485034584999, 'timing/model': 1.2641363348811865, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 18:50:48,952][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.5802640914916992, 'train/grad_norm_pre_clip': 20.177513122558594, 'timing/data': 0.01263311319053173, 'timing/model': 1.23374461568892, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.02}) +[2026-06-15 18:50:53,948][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.6368723511695862, 'train/grad_norm_pre_clip': 13.737288475036621, 'timing/data': 0.012870822101831436, 'timing/model': 1.4477182440459728, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.02}) +[2026-06-15 18:50:58,925][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.7953289747238159, 'train/grad_norm_pre_clip': 19.63357162475586, 'timing/data': 0.013045961037278175, 'timing/model': 1.3241430427879095, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.02}) +[2026-06-15 18:51:03,757][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.8406080603599548, 'train/grad_norm_pre_clip': 18.088998794555664, 'timing/data': 0.012968016788363457, 'timing/model': 1.2476453762501478, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.02}) +[2026-06-15 18:51:08,671][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.6770475506782532, 'train/grad_norm_pre_clip': 7.9095072746276855, 'timing/data': 0.012761043384671211, 'timing/model': 1.2716793809086084, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.02}) +[2026-06-15 18:51:13,680][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.5277400016784668, 'train/grad_norm_pre_clip': 4.46175479888916, 'timing/data': 0.012451035901904106, 'timing/model': 1.27915252558887, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.02}) +[2026-06-15 18:51:18,546][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.5740368366241455, 'train/grad_norm_pre_clip': 5.117498397827148, 'timing/data': 0.013315025717020035, 'timing/model': 1.2484532352536917, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 18:51:23,937][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.7155932188034058, 'train/grad_norm_pre_clip': 14.676297187805176, 'timing/data': 0.013411292806267738, 'timing/model': 1.8390712924301624, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.02}) +[2026-06-15 18:51:28,749][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.5284402370452881, 'train/grad_norm_pre_clip': 6.3115010261535645, 'timing/data': 0.012648055329918861, 'timing/model': 1.2517734672874212, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.02}) +[2026-06-15 18:51:33,550][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.685431957244873, 'train/grad_norm_pre_clip': 8.88254165649414, 'timing/data': 0.011892914772033691, 'timing/model': 1.253806820139289, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.02}) +[2026-06-15 18:51:38,504][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.6257445812225342, 'train/grad_norm_pre_clip': 13.560232162475586, 'timing/data': 0.01206188090145588, 'timing/model': 1.3094406481832266, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.02}) +[2026-06-15 18:51:43,457][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.564083993434906, 'train/grad_norm_pre_clip': 7.56689977645874, 'timing/data': 0.011852696537971497, 'timing/model': 1.2469460312277079, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 18:51:48,354][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.4936544597148895, 'train/grad_norm_pre_clip': 11.772976875305176, 'timing/data': 0.014764178544282913, 'timing/model': 1.2534342892467976, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.02}) +[2026-06-15 18:51:53,254][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.6251440644264221, 'train/grad_norm_pre_clip': 17.94013786315918, 'timing/data': 0.012802651152014732, 'timing/model': 1.2557514291256666, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.02}) +[2026-06-15 18:51:58,159][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.6125572919845581, 'train/grad_norm_pre_clip': 20.852895736694336, 'timing/data': 0.013058414682745934, 'timing/model': 1.2547869086265564, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.02}) +[2026-06-15 18:52:03,262][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.6102402210235596, 'train/grad_norm_pre_clip': 14.011717796325684, 'timing/data': 0.012184612452983856, 'timing/model': 1.371389266103506, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.02}) +[2026-06-15 18:52:08,260][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.5400649905204773, 'train/grad_norm_pre_clip': 12.50018310546875, 'timing/data': 0.012863269075751305, 'timing/model': 1.3400248978286982, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.02}) +[2026-06-15 18:52:13,155][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.5083160400390625, 'train/grad_norm_pre_clip': 5.263453483581543, 'timing/data': 0.012309933081269264, 'timing/model': 1.2560982760041952, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.02}) +[2026-06-15 18:52:18,069][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.5785739421844482, 'train/grad_norm_pre_clip': 4.966517925262451, 'timing/data': 0.012764660641551018, 'timing/model': 1.2489299569278955, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.02}) +[2026-06-15 18:52:22,973][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.6107025146484375, 'train/grad_norm_pre_clip': 4.781861305236816, 'timing/data': 0.012597421184182167, 'timing/model': 1.2548293452709913, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.02}) +[2026-06-15 18:52:28,062][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.5858097672462463, 'train/grad_norm_pre_clip': 6.610975742340088, 'timing/data': 0.013615002855658531, 'timing/model': 1.2713252510875463, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.03}) +[2026-06-15 18:52:32,953][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.5466405153274536, 'train/grad_norm_pre_clip': 5.610494136810303, 'timing/data': 0.01335659809410572, 'timing/model': 1.252746356651187, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.03}) +[2026-06-15 18:52:37,851][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.5915202498435974, 'train/grad_norm_pre_clip': 2.5863683223724365, 'timing/data': 0.018626920878887177, 'timing/model': 1.237107578665018, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.03}) +[2026-06-15 18:52:42,772][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.6012626886367798, 'train/grad_norm_pre_clip': 3.0627729892730713, 'timing/data': 0.012991068884730339, 'timing/model': 1.283803142607212, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.03}) +[2026-06-15 18:52:48,177][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.46188050508499146, 'train/grad_norm_pre_clip': 3.5541751384735107, 'timing/data': 0.018944215029478073, 'timing/model': 1.3351883832365274, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.03}) +[2026-06-15 18:52:53,250][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.4979674816131592, 'train/grad_norm_pre_clip': 5.752820014953613, 'timing/data': 0.01245160587131977, 'timing/model': 1.2591267954558134, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.03}) +[2026-06-15 18:52:58,269][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.640041172504425, 'train/grad_norm_pre_clip': 12.068479537963867, 'timing/data': 0.012968650087714195, 'timing/model': 1.270481165498495, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.03}) +[2026-06-15 18:53:03,273][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.5784414410591125, 'train/grad_norm_pre_clip': 13.219377517700195, 'timing/data': 0.011834532022476196, 'timing/model': 1.2748832888901234, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.03}) +[2026-06-15 18:53:08,148][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.3695666491985321, 'train/grad_norm_pre_clip': 2.6825366020202637, 'timing/data': 0.013065217062830925, 'timing/model': 1.2515122052282095, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.03}) +[2026-06-15 18:53:12,960][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.46759316325187683, 'train/grad_norm_pre_clip': 9.87741756439209, 'timing/data': 0.012763772159814835, 'timing/model': 1.2592764589935541, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.03}) +[2026-06-15 18:53:17,875][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.2937176823616028, 'train/grad_norm_pre_clip': 6.049681186676025, 'timing/data': 0.0172087661921978, 'timing/model': 1.2522506844252348, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.03}) +[2026-06-15 18:53:22,829][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 0.4301876723766327, 'train/grad_norm_pre_clip': 114.37755584716797, 'timing/data': 0.01277589239180088, 'timing/model': 1.3050175923854113, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.03}) +[2026-06-15 18:53:28,451][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.4230019748210907, 'train/grad_norm_pre_clip': 7.601662635803223, 'timing/data': 0.012588119134306908, 'timing/model': 1.3511417489498854, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.03}) +[2026-06-15 18:53:33,466][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.8821499347686768, 'train/grad_norm_pre_clip': 23.095108032226562, 'timing/data': 0.011555872857570648, 'timing/model': 1.275725832208991, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.03}) +[2026-06-15 18:53:38,760][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.5819803476333618, 'train/grad_norm_pre_clip': 21.969066619873047, 'timing/data': 0.01196935772895813, 'timing/model': 1.560203792527318, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.03}) +[2026-06-15 18:53:44,010][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.683251678943634, 'train/grad_norm_pre_clip': 14.038250923156738, 'timing/data': 0.017785247415304184, 'timing/model': 1.379765847697854, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.03}) +[2026-06-15 18:53:49,358][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.6412600874900818, 'train/grad_norm_pre_clip': 10.649118423461914, 'timing/data': 0.020465435460209846, 'timing/model': 1.3386239763349295, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.03}) +[2026-06-15 18:53:54,674][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.3608018755912781, 'train/grad_norm_pre_clip': 5.234973430633545, 'timing/data': 0.015199892222881317, 'timing/model': 1.3673572465777397, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.03}) +[2026-06-15 18:54:00,268][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.4548056125640869, 'train/grad_norm_pre_clip': 4.9897918701171875, 'timing/data': 0.018239852041006088, 'timing/model': 1.3881291691213846, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.03}) +[2026-06-15 18:54:05,921][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.40885382890701294, 'train/grad_norm_pre_clip': 3.7938029766082764, 'timing/data': 0.019767891615629196, 'timing/model': 1.461334077641368, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.03}) +[2026-06-15 18:54:11,475][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.507046103477478, 'train/grad_norm_pre_clip': 4.500901222229004, 'timing/data': 0.022991731762886047, 'timing/model': 1.4027739632874727, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.03}) +[2026-06-15 18:54:16,560][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.3292650282382965, 'train/grad_norm_pre_clip': 5.446850776672363, 'timing/data': 0.013210462406277657, 'timing/model': 1.2586955465376377, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.03}) +[2026-06-15 18:54:21,551][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.5481966137886047, 'train/grad_norm_pre_clip': 15.122845649719238, 'timing/data': 0.01371118426322937, 'timing/model': 1.243345843628049, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.03}) +[2026-06-15 18:54:26,756][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.3999304175376892, 'train/grad_norm_pre_clip': 4.311548709869385, 'timing/data': 0.017497407272458076, 'timing/model': 1.3477651719003916, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.03}) +[2026-06-15 18:54:31,836][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.4602227807044983, 'train/grad_norm_pre_clip': 9.318939208984375, 'timing/data': 0.014079181477427483, 'timing/model': 1.322174770757556, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.03}) +[2026-06-15 18:54:36,765][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.4348880350589752, 'train/grad_norm_pre_clip': 6.327298641204834, 'timing/data': 0.012402232736349106, 'timing/model': 1.2551042400300503, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.03}) +[2026-06-15 18:54:41,864][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.45022323727607727, 'train/grad_norm_pre_clip': 3.0629069805145264, 'timing/data': 0.01401522010564804, 'timing/model': 1.2692569941282272, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.04}) +[2026-06-15 18:54:47,042][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.4487628936767578, 'train/grad_norm_pre_clip': 3.5137741565704346, 'timing/data': 0.011821506544947624, 'timing/model': 1.3467567171901464, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.04}) +[2026-06-15 18:54:51,845][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.3653678894042969, 'train/grad_norm_pre_clip': 5.540731430053711, 'timing/data': 0.012551840394735336, 'timing/model': 1.249110234901309, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.04}) +[2026-06-15 18:54:56,646][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.3200666010379791, 'train/grad_norm_pre_clip': 2.9317069053649902, 'timing/data': 0.012781992554664612, 'timing/model': 1.257616700604558, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.04}) +[2026-06-15 18:55:01,447][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.4081973731517792, 'train/grad_norm_pre_clip': 4.118079662322998, 'timing/data': 0.011909333989024162, 'timing/model': 1.2581983357667923, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.04}) +[2026-06-15 18:55:06,252][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.4307279884815216, 'train/grad_norm_pre_clip': 14.819358825683594, 'timing/data': 0.01309264823794365, 'timing/model': 1.2538116462528706, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.04}) +[2026-06-15 18:55:11,150][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.41345667839050293, 'train/grad_norm_pre_clip': 21.516408920288086, 'timing/data': 0.013896225020289421, 'timing/model': 1.2361380737274885, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.04}) +[2026-06-15 18:55:16,347][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.5250341892242432, 'train/grad_norm_pre_clip': 18.73723030090332, 'timing/data': 0.013194642961025238, 'timing/model': 1.24240586720407, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.04}) +[2026-06-15 18:55:21,367][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 1.1137739419937134, 'train/grad_norm_pre_clip': 18.969980239868164, 'timing/data': 0.017257215455174446, 'timing/model': 1.3470959346741438, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.04}) +[2026-06-15 18:57:27,308][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpe2s0xqib/test.c -o /tmp/tmpe2s0xqib/test.o +[2026-06-15 18:57:27,337][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpe2s0xqib/test.o -laio -o /tmp/tmpe2s0xqib/a.out +[2026-06-15 18:57:27,861][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpj4r4strz/test.c -o /tmp/tmpj4r4strz/test.o +[2026-06-15 18:57:27,887][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpj4r4strz/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpj4r4strz/a.out +[2026-06-15 18:57:27,942][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpc5dwxyuk/test.c -o /tmp/tmpc5dwxyuk/test.o +[2026-06-15 18:57:27,968][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpc5dwxyuk/test.o -laio -o /tmp/tmpc5dwxyuk/a.out +[2026-06-15 19:03:25,951][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.6895526051521301, 'train/grad_norm_pre_clip': 54.99896240234375, 'eval/loss': 0.37620729207992554, 'eval/flap_precision': 0.6140684410646388, 'eval/flap_recall': 0.8765264586160109, 'eval/flap_f1': 0.7221911682504192, 'eval/flap_rate_model': 0.2922222222222222, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.7221911682504192, 'timing/data': 0.011938365176320076, 'timing/model': 1.2565926481038332, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.04}) +[2026-06-15 19:03:25,952][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_100_state +[2026-06-15 19:03:25,959][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 19:03:34,334][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_100_state/model.safetensors +[2026-06-15 19:03:48,998][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_100_state/optimizer.bin +[2026-06-15 19:03:48,999][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_100_state/sampler.bin +[2026-06-15 19:03:48,999][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_100_state/sampler_1.bin +[2026-06-15 19:03:49,000][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-15 19:03:49,003][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-15 19:03:49,004][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 19:03:49,012][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 19:03:54,795][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.6257591843605042, 'train/grad_norm_pre_clip': 22.94671630859375, 'timing/data': 0.01031578704714775, 'timing/model': 1.4384132474660873, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 0.04}) +[2026-06-15 19:04:00,749][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.5150156617164612, 'train/grad_norm_pre_clip': 17.4416446685791, 'timing/data': 0.015244515612721443, 'timing/model': 1.5213141646236181, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 0.04}) +[2026-06-15 19:04:06,725][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.5122390985488892, 'train/grad_norm_pre_clip': 7.560738563537598, 'timing/data': 0.011819949373602867, 'timing/model': 1.6498572323471308, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 0.04}) +[2026-06-15 19:04:12,471][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.524395227432251, 'train/grad_norm_pre_clip': 9.261170387268066, 'timing/data': 0.016024835407733917, 'timing/model': 1.4313479959964752, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 0.04}) +[2026-06-15 19:04:18,373][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.5620490908622742, 'train/grad_norm_pre_clip': 10.744336128234863, 'timing/data': 0.0153788011521101, 'timing/model': 1.439180115237832, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 0.04}) +[2026-06-15 19:04:24,175][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.5180409550666809, 'train/grad_norm_pre_clip': 5.710319519042969, 'timing/data': 0.010966768488287926, 'timing/model': 1.4296815358102322, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 0.04}) +[2026-06-15 19:04:30,079][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.5354279279708862, 'train/grad_norm_pre_clip': 4.670665740966797, 'timing/data': 0.01048796996474266, 'timing/model': 1.4441013149917126, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 0.04}) +[2026-06-15 19:04:36,133][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.4139496386051178, 'train/grad_norm_pre_clip': 7.462329387664795, 'timing/data': 0.012204395607113838, 'timing/model': 1.5442617032676935, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 0.04}) +[2026-06-15 19:04:42,253][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.5257267951965332, 'train/grad_norm_pre_clip': 4.373129367828369, 'timing/data': 0.013196747750043869, 'timing/model': 1.7571870405226946, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 0.04}) +[2026-06-15 19:04:48,139][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.4918507933616638, 'train/grad_norm_pre_clip': 5.303348064422607, 'timing/data': 0.012728378176689148, 'timing/model': 1.5427652355283499, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.04}) +[2026-06-15 19:04:54,271][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.3991169035434723, 'train/grad_norm_pre_clip': 7.947640419006348, 'timing/data': 0.012124566361308098, 'timing/model': 1.5757893361151218, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 0.04}) +[2026-06-15 19:05:00,774][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.4239080250263214, 'train/grad_norm_pre_clip': 7.328357219696045, 'timing/data': 0.012878576293587685, 'timing/model': 1.5893911365419626, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 0.04}) +[2026-06-15 19:05:07,180][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.45355647802352905, 'train/grad_norm_pre_clip': 4.455695629119873, 'timing/data': 0.01296224258840084, 'timing/model': 1.550900388509035, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 0.04}) +[2026-06-15 19:05:13,479][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.2881810963153839, 'train/grad_norm_pre_clip': 11.478075981140137, 'timing/data': 0.013097729533910751, 'timing/model': 1.5600861515849829, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 0.04}) +[2026-06-15 19:05:19,824][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.6167723536491394, 'train/grad_norm_pre_clip': 9.376981735229492, 'timing/data': 0.012688769027590752, 'timing/model': 1.8010620214045048, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 0.04}) +[2026-06-15 19:05:25,882][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.6053109765052795, 'train/grad_norm_pre_clip': 8.688023567199707, 'timing/data': 0.012410849332809448, 'timing/model': 1.560438361018896, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 0.04}) +[2026-06-15 19:05:32,479][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.3867817521095276, 'train/grad_norm_pre_clip': 7.233063697814941, 'timing/data': 0.01211150735616684, 'timing/model': 1.5725183691829443, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 0.05}) +[2026-06-15 19:05:38,844][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.31333526968955994, 'train/grad_norm_pre_clip': 4.745509147644043, 'timing/data': 0.012319672852754593, 'timing/model': 1.5044815000146627, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 0.05}) +[2026-06-15 19:05:45,047][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.5463556051254272, 'train/grad_norm_pre_clip': 12.722293853759766, 'timing/data': 0.012620627880096436, 'timing/model': 1.5303716771304607, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 0.05}) +[2026-06-15 19:05:51,243][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.4312005043029785, 'train/grad_norm_pre_clip': 7.68502950668335, 'timing/data': 0.013122066855430603, 'timing/model': 1.5141245257109404, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.05}) +[2026-06-15 19:05:57,203][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.488573282957077, 'train/grad_norm_pre_clip': 8.243599891662598, 'timing/data': 0.010542139410972595, 'timing/model': 1.6740442253649235, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 0.05}) +[2026-06-15 19:06:02,887][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.3124884366989136, 'train/grad_norm_pre_clip': 10.855141639709473, 'timing/data': 0.0112540852278471, 'timing/model': 1.448468567803502, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 0.05}) +[2026-06-15 19:06:08,890][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.4482584297657013, 'train/grad_norm_pre_clip': 6.932565212249756, 'timing/data': 0.010541200637817383, 'timing/model': 1.5616857409477234, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 0.05}) +[2026-06-15 19:06:14,781][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.5499709844589233, 'train/grad_norm_pre_clip': 6.0039801597595215, 'timing/data': 0.011293536052107811, 'timing/model': 1.4548131357878447, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 0.05}) +[2026-06-15 19:06:21,378][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.6226928234100342, 'train/grad_norm_pre_clip': 7.361837863922119, 'timing/data': 0.01084950938820839, 'timing/model': 1.5586615595966578, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 0.05}) +[2026-06-15 19:06:27,676][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.29954707622528076, 'train/grad_norm_pre_clip': 16.307029724121094, 'timing/data': 0.012260789051651955, 'timing/model': 1.518846521154046, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 0.05}) +[2026-06-15 19:06:34,016][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.4339091181755066, 'train/grad_norm_pre_clip': 18.66168212890625, 'timing/data': 0.012737032026052475, 'timing/model': 1.786458745598793, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 0.05}) +[2026-06-15 19:06:40,075][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.385786235332489, 'train/grad_norm_pre_clip': 8.292863845825195, 'timing/data': 0.012452609837055206, 'timing/model': 1.5547876749187708, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 0.05}) +[2026-06-15 19:06:46,553][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.4437907934188843, 'train/grad_norm_pre_clip': 16.075254440307617, 'timing/data': 0.012658866122364998, 'timing/model': 1.6230227276682854, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 0.05}) +[2026-06-15 19:06:52,766][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.45356690883636475, 'train/grad_norm_pre_clip': 38.54984664916992, 'timing/data': 0.012475579977035522, 'timing/model': 1.5503928791731596, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.05}) +[2026-06-15 19:06:59,146][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.3767856955528259, 'train/grad_norm_pre_clip': 6.309640884399414, 'timing/data': 0.03176407143473625, 'timing/model': 1.5052185766398907, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 0.05}) +[2026-06-15 19:07:05,344][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.2624562978744507, 'train/grad_norm_pre_clip': 5.139845848083496, 'timing/data': 0.011646127328276634, 'timing/model': 1.5384818855673075, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 0.05}) +[2026-06-15 19:07:11,486][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.2746107578277588, 'train/grad_norm_pre_clip': 6.623098850250244, 'timing/data': 0.012142283841967583, 'timing/model': 1.751384811475873, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 0.05}) +[2026-06-15 19:07:17,442][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.30222654342651367, 'train/grad_norm_pre_clip': 7.768138408660889, 'timing/data': 0.01209191232919693, 'timing/model': 1.5086094252765179, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 0.05}) +[2026-06-15 19:07:23,630][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.1668901890516281, 'train/grad_norm_pre_clip': 7.021686553955078, 'timing/data': 0.012303773313760757, 'timing/model': 1.5539509896188974, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.05}) +[2026-06-15 19:07:29,763][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.5719524025917053, 'train/grad_norm_pre_clip': 64.38138580322266, 'timing/data': 0.012452015653252602, 'timing/model': 1.54569511115551, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 0.05}) +[2026-06-15 19:07:36,149][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 1.4513616561889648, 'train/grad_norm_pre_clip': 104.72686004638672, 'timing/data': 0.012938084080815315, 'timing/model': 1.5491131749004126, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 0.05}) +[2026-06-15 19:07:42,441][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.8812086582183838, 'train/grad_norm_pre_clip': 60.30991744995117, 'timing/data': 0.012491371482610703, 'timing/model': 1.5080559309571981, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 0.05}) +[2026-06-15 19:07:48,718][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.3689625859260559, 'train/grad_norm_pre_clip': 8.345800399780273, 'timing/data': 0.013893036171793938, 'timing/model': 1.8266069516539574, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 0.05}) +[2026-06-15 19:07:54,767][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.4166554808616638, 'train/grad_norm_pre_clip': 11.476691246032715, 'timing/data': 0.012441795319318771, 'timing/model': 1.5731733925640583, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.05}) +[2026-06-15 19:08:01,270][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.5059816241264343, 'train/grad_norm_pre_clip': 13.235543251037598, 'timing/data': 0.013299582526087761, 'timing/model': 1.5269735362380743, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 0.05}) +[2026-06-15 19:08:07,565][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.31568610668182373, 'train/grad_norm_pre_clip': 12.276432037353516, 'timing/data': 0.012215035036206245, 'timing/model': 1.565188067033887, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 0.05}) +[2026-06-15 19:08:13,869][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.33955374360084534, 'train/grad_norm_pre_clip': 11.329293251037598, 'timing/data': 0.019985174760222435, 'timing/model': 1.55262778326869, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 0.06}) +[2026-06-15 19:08:20,259][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.4919147789478302, 'train/grad_norm_pre_clip': 24.133386611938477, 'timing/data': 0.018365327268838882, 'timing/model': 1.5679077692329884, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 0.06}) +[2026-06-15 19:08:26,470][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.3373374938964844, 'train/grad_norm_pre_clip': 14.468968391418457, 'timing/data': 0.012139931321144104, 'timing/model': 1.7559393048286438, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 0.06}) +[2026-06-15 19:08:32,774][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.36724987626075745, 'train/grad_norm_pre_clip': 8.924036979675293, 'timing/data': 0.01127026230096817, 'timing/model': 1.7792114485055208, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 0.06}) +[2026-06-15 19:08:38,604][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.38014188408851624, 'train/grad_norm_pre_clip': 13.777669906616211, 'timing/data': 0.011128200218081474, 'timing/model': 1.4479682706296444, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 0.06}) +[2026-06-15 19:08:44,491][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.42578861117362976, 'train/grad_norm_pre_clip': 11.102985382080078, 'timing/data': 0.011113254353404045, 'timing/model': 1.4404871109873056, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 0.06}) +[2026-06-15 19:08:50,382][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.8929845690727234, 'train/grad_norm_pre_clip': 25.74066925048828, 'timing/data': 0.010183000937104225, 'timing/model': 1.437118973582983, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.06}) +[2026-06-15 19:08:56,289][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.5678400993347168, 'train/grad_norm_pre_clip': 10.726646423339844, 'timing/data': 0.010198036208748817, 'timing/model': 1.4156862627714872, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.06}) +[2026-06-15 19:09:02,225][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.4364469647407532, 'train/grad_norm_pre_clip': 9.616485595703125, 'timing/data': 0.011420989409089088, 'timing/model': 1.6806303076446056, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 0.06}) +[2026-06-15 19:09:07,881][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.40498441457748413, 'train/grad_norm_pre_clip': 4.182466983795166, 'timing/data': 0.011004138737916946, 'timing/model': 1.431819075718522, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 0.06}) +[2026-06-15 19:09:13,797][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.587107241153717, 'train/grad_norm_pre_clip': 14.174885749816895, 'timing/data': 0.010992005467414856, 'timing/model': 1.4539938289672136, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 0.06}) +[2026-06-15 19:09:19,839][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.396762877702713, 'train/grad_norm_pre_clip': 8.487641334533691, 'timing/data': 0.013033231720328331, 'timing/model': 1.5190021842718124, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 0.06}) +[2026-06-15 19:09:26,044][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.509322464466095, 'train/grad_norm_pre_clip': 13.712817192077637, 'timing/data': 0.01273295283317566, 'timing/model': 1.5410092324018478, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 0.06}) +[2026-06-15 19:09:32,172][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.36164194345474243, 'train/grad_norm_pre_clip': 6.4074931144714355, 'timing/data': 0.012099087238311768, 'timing/model': 1.4978403467684984, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 0.06}) +[2026-06-15 19:09:38,456][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.45231524109840393, 'train/grad_norm_pre_clip': 6.357168197631836, 'timing/data': 0.011872710660099983, 'timing/model': 1.7591530289500952, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 0.06}) +[2026-06-15 19:09:44,449][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.2413424849510193, 'train/grad_norm_pre_clip': 6.625472068786621, 'timing/data': 0.012448001652956009, 'timing/model': 1.5672278385609388, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 0.06}) +[2026-06-15 19:09:50,742][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.5006401538848877, 'train/grad_norm_pre_clip': 10.349830627441406, 'timing/data': 0.0118671003729105, 'timing/model': 1.5518400054425001, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 0.06}) +[2026-06-15 19:09:56,949][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.23335105180740356, 'train/grad_norm_pre_clip': 4.671407699584961, 'timing/data': 0.012680351734161377, 'timing/model': 1.5605333596467972, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.06}) +[2026-06-15 19:10:03,273][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.5992698669433594, 'train/grad_norm_pre_clip': 19.418685913085938, 'timing/data': 0.013024820014834404, 'timing/model': 1.5685992389917374, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 0.06}) +[2026-06-15 19:10:09,549][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.3323291540145874, 'train/grad_norm_pre_clip': 17.122617721557617, 'timing/data': 0.012373803183436394, 'timing/model': 1.5648418068885803, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 0.06}) +[2026-06-15 19:10:16,114][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.35419777035713196, 'train/grad_norm_pre_clip': 16.475889205932617, 'timing/data': 0.018007583916187286, 'timing/model': 1.8595386855304241, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 0.06}) +[2026-06-15 19:10:22,945][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.4879995882511139, 'train/grad_norm_pre_clip': 24.70724868774414, 'timing/data': 0.020091557875275612, 'timing/model': 1.6801433861255646, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 0.06}) +[2026-06-15 19:10:29,305][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.32728394865989685, 'train/grad_norm_pre_clip': 8.661782264709473, 'timing/data': 0.011604225262999535, 'timing/model': 1.5477725956588984, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 0.06}) +[2026-06-15 19:10:35,593][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.27905455231666565, 'train/grad_norm_pre_clip': 10.412144660949707, 'timing/data': 0.012041721493005753, 'timing/model': 1.5454778503626585, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 0.06}) +[2026-06-15 19:10:41,920][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.48097217082977295, 'train/grad_norm_pre_clip': 15.182086944580078, 'timing/data': 0.01350279338657856, 'timing/model': 1.5466604754328728, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 0.06}) +[2026-06-15 19:10:48,522][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.3312852084636688, 'train/grad_norm_pre_clip': 7.3851423263549805, 'timing/data': 0.011532656848430634, 'timing/model': 1.784141743555665, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 0.07}) +[2026-06-15 19:10:54,577][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.4344395697116852, 'train/grad_norm_pre_clip': 13.33521556854248, 'timing/data': 0.012448552995920181, 'timing/model': 1.5401515159755945, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 0.07}) +[2026-06-15 19:11:01,302][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.16733857989311218, 'train/grad_norm_pre_clip': 5.637725353240967, 'timing/data': 0.012317582964897156, 'timing/model': 1.6226889044046402, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.07}) +[2026-06-15 19:11:07,989][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.23940202593803406, 'train/grad_norm_pre_clip': 10.335057258605957, 'timing/data': 0.013045962899923325, 'timing/model': 1.665369100868702, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 0.07}) +[2026-06-15 19:11:14,610][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.5145256519317627, 'train/grad_norm_pre_clip': 19.90628433227539, 'timing/data': 0.012481411918997765, 'timing/model': 1.5781119894236326, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 0.07}) +[2026-06-15 19:11:20,999][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.5269585251808167, 'train/grad_norm_pre_clip': 17.724477767944336, 'timing/data': 0.012622440233826637, 'timing/model': 1.6181483566761017, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 0.07}) +[2026-06-15 19:11:27,867][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.33223748207092285, 'train/grad_norm_pre_clip': 9.64694881439209, 'timing/data': 0.012029394507408142, 'timing/model': 1.8406690694391727, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 0.07}) +[2026-06-15 19:11:33,976][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.47400984168052673, 'train/grad_norm_pre_clip': 14.091753005981445, 'timing/data': 0.012708496302366257, 'timing/model': 1.5275124981999397, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 0.07}) +[2026-06-15 19:11:40,377][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.42800015211105347, 'train/grad_norm_pre_clip': 9.846426963806152, 'timing/data': 0.013288373127579689, 'timing/model': 1.5417966358363628, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 0.07}) +[2026-06-15 19:11:46,738][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.5088513493537903, 'train/grad_norm_pre_clip': 10.745437622070312, 'timing/data': 0.012209462001919746, 'timing/model': 1.5592157654464245, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 0.07}) +[2026-06-15 19:11:52,957][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.44115954637527466, 'train/grad_norm_pre_clip': 6.519452095031738, 'timing/data': 0.00994214415550232, 'timing/model': 1.6602680142968893, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 0.07}) +[2026-06-15 19:11:59,143][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.4554843604564667, 'train/grad_norm_pre_clip': 13.365137100219727, 'timing/data': 0.01276814378798008, 'timing/model': 1.4856914710253477, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 0.07}) +[2026-06-15 19:12:04,987][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.5443095564842224, 'train/grad_norm_pre_clip': 9.416001319885254, 'timing/data': 0.011661842465400696, 'timing/model': 1.6576046831905842, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.07}) +[2026-06-15 19:12:10,680][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.3196408152580261, 'train/grad_norm_pre_clip': 6.018702507019043, 'timing/data': 0.01016772910952568, 'timing/model': 1.4550598077476025, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 0.07}) +[2026-06-15 19:12:16,583][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.2968251705169678, 'train/grad_norm_pre_clip': 4.361067295074463, 'timing/data': 0.010563338175415993, 'timing/model': 1.4426210708916187, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 0.07}) +[2026-06-15 19:12:22,480][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.33258911967277527, 'train/grad_norm_pre_clip': 12.309675216674805, 'timing/data': 0.009756715968251228, 'timing/model': 1.4545492269098759, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 0.07}) +[2026-06-15 19:12:28,381][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.5906557440757751, 'train/grad_norm_pre_clip': 10.486757278442383, 'timing/data': 0.009653998538851738, 'timing/model': 1.4388669654726982, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 0.07}) +[2026-06-15 19:12:34,280][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.4384256899356842, 'train/grad_norm_pre_clip': 12.69680404663086, 'timing/data': 0.010159973055124283, 'timing/model': 1.4783503413200378, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 0.07}) +[2026-06-15 19:12:40,186][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.24295984208583832, 'train/grad_norm_pre_clip': 7.588339328765869, 'timing/data': 0.009802108630537987, 'timing/model': 1.6402707230299711, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 0.07}) +[2026-06-15 19:12:45,897][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.3865174353122711, 'train/grad_norm_pre_clip': 14.46019172668457, 'timing/data': 0.010482843965291977, 'timing/model': 1.4555657282471657, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 0.07}) +[2026-06-15 19:12:51,794][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.24305836856365204, 'train/grad_norm_pre_clip': 15.53459358215332, 'timing/data': 0.010836875066161156, 'timing/model': 1.444593032822013, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 0.07}) +[2026-06-15 19:12:57,689][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.48632991313934326, 'train/grad_norm_pre_clip': 15.602582931518555, 'timing/data': 0.011798221617937088, 'timing/model': 1.4391919020563364, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 0.07}) +[2026-06-15 19:13:03,481][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.2444409281015396, 'train/grad_norm_pre_clip': 20.273929595947266, 'timing/data': 0.010748542845249176, 'timing/model': 1.453690206632018, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.07}) +[2026-06-15 19:13:09,542][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.4014548361301422, 'train/grad_norm_pre_clip': 14.605171203613281, 'timing/data': 0.01017536036670208, 'timing/model': 1.4886010605841875, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 0.07}) +[2026-06-15 19:13:15,787][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.2982407510280609, 'train/grad_norm_pre_clip': 10.081781387329102, 'timing/data': 0.011963970959186554, 'timing/model': 1.8085772152990103, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 0.07}) +[2026-06-15 19:13:21,833][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.16458426415920258, 'train/grad_norm_pre_clip': 5.740553855895996, 'timing/data': 0.01234625093638897, 'timing/model': 1.5544349234551191, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 0.07}) +[2026-06-15 19:13:28,235][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.2417597621679306, 'train/grad_norm_pre_clip': 5.98413610458374, 'timing/data': 0.014318816363811493, 'timing/model': 1.5995615720748901, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 0.08}) +[2026-06-15 19:13:34,529][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.3692851960659027, 'train/grad_norm_pre_clip': 7.293747901916504, 'timing/data': 0.014880284667015076, 'timing/model': 1.5686938799917698, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 0.08}) +[2026-06-15 19:13:40,787][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.3932490944862366, 'train/grad_norm_pre_clip': 11.24608325958252, 'timing/data': 0.012257320806384087, 'timing/model': 1.5607779044657946, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 0.08}) +[2026-06-15 19:13:47,184][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.29605385661125183, 'train/grad_norm_pre_clip': 8.739174842834473, 'timing/data': 0.01272585615515709, 'timing/model': 1.612293278798461, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 0.08}) +[2026-06-15 19:13:53,487][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.28836941719055176, 'train/grad_norm_pre_clip': 10.685478210449219, 'timing/data': 0.012786049395799637, 'timing/model': 1.761927342042327, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.08}) +[2026-06-15 19:13:59,597][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.26527342200279236, 'train/grad_norm_pre_clip': 10.959639549255371, 'timing/data': 0.01212167739868164, 'timing/model': 1.577228868380189, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 0.08}) +[2026-06-15 19:21:50,393][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.40756145119667053, 'train/grad_norm_pre_clip': 14.493865013122559, 'eval/loss': 0.20939967036247253, 'eval/flap_precision': 0.6827651515151515, 'eval/flap_recall': 0.9782903663500678, 'eval/flap_f1': 0.8042387060791969, 'eval/flap_rate_model': 0.29333333333333333, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8042387060791969, 'timing/data': 0.012683659791946411, 'timing/model': 1.5319031812250614, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.08}) +[2026-06-15 19:21:50,396][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_200_state +[2026-06-15 19:21:50,402][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 19:21:58,680][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_200_state/model.safetensors +[2026-06-15 19:22:12,739][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_200_state/optimizer.bin +[2026-06-15 19:22:12,740][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_200_state/sampler.bin +[2026-06-15 19:22:12,740][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_200_state/sampler_1.bin +[2026-06-15 19:22:12,741][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-15 19:22:12,841][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-15 19:22:12,842][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 19:22:12,850][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 19:22:18,529][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.36058831214904785, 'train/grad_norm_pre_clip': 10.466704368591309, 'timing/data': 0.01090221293270588, 'timing/model': 1.475527424365282, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 0.08}) +[2026-06-15 19:22:25,189][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.4515170454978943, 'train/grad_norm_pre_clip': 12.476136207580566, 'timing/data': 0.014719067141413689, 'timing/model': 1.5589357037097216, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 0.08}) +[2026-06-15 19:22:31,590][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.30474409461021423, 'train/grad_norm_pre_clip': 8.508352279663086, 'timing/data': 0.013863755390048027, 'timing/model': 1.5613809116184711, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 0.08}) +[2026-06-15 19:22:38,249][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.5186912417411804, 'train/grad_norm_pre_clip': 9.05171012878418, 'timing/data': 0.011150192469358444, 'timing/model': 1.6991036031395197, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 0.08}) +[2026-06-15 19:22:44,440][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.23315845429897308, 'train/grad_norm_pre_clip': 8.927083969116211, 'timing/data': 0.01189814880490303, 'timing/model': 1.5735908299684525, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.08}) +[2026-06-15 19:22:50,659][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.28042107820510864, 'train/grad_norm_pre_clip': 6.221262454986572, 'timing/data': 0.011398212984204292, 'timing/model': 1.7474829889833927, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 0.08}) +[2026-06-15 19:22:56,770][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.6060979962348938, 'train/grad_norm_pre_clip': 25.23383140563965, 'timing/data': 0.012518491595983505, 'timing/model': 1.557101109996438, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 0.08}) +[2026-06-15 19:23:03,070][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.32269182801246643, 'train/grad_norm_pre_clip': 6.243686199188232, 'timing/data': 0.013425527140498161, 'timing/model': 1.5533157717436552, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 0.08}) +[2026-06-15 19:23:09,368][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.3204279839992523, 'train/grad_norm_pre_clip': 6.170331001281738, 'timing/data': 0.013177765533328056, 'timing/model': 1.5513718388974667, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 0.08}) +[2026-06-15 19:23:15,864][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.3012814223766327, 'train/grad_norm_pre_clip': 9.954532623291016, 'timing/data': 0.013138623908162117, 'timing/model': 1.509421143680811, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.08}) +[2026-06-15 19:23:22,200][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.4583684802055359, 'train/grad_norm_pre_clip': 12.65805435180664, 'timing/data': 0.011949170380830765, 'timing/model': 1.7870706617832184, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 0.08}) +[2026-06-15 19:23:28,459][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.2273314744234085, 'train/grad_norm_pre_clip': 4.915033340454102, 'timing/data': 0.01286543719470501, 'timing/model': 1.734317997470498, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 0.08}) +[2026-06-15 19:23:34,571][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.3309882879257202, 'train/grad_norm_pre_clip': 14.015440940856934, 'timing/data': 0.012379676103591919, 'timing/model': 1.551163237541914, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 0.08}) +[2026-06-15 19:23:40,875][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.2891162037849426, 'train/grad_norm_pre_clip': 17.51253318786621, 'timing/data': 0.012811986729502678, 'timing/model': 1.5561457443982363, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 0.08}) +[2026-06-15 19:23:47,584][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.528221070766449, 'train/grad_norm_pre_clip': 12.750551223754883, 'timing/data': 0.0124345812946558, 'timing/model': 1.568660955876112, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 0.08}) +[2026-06-15 19:23:53,831][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.3215579688549042, 'train/grad_norm_pre_clip': 10.794120788574219, 'timing/data': 0.01298539713025093, 'timing/model': 1.475784843787551, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 0.08}) +[2026-06-15 19:23:59,779][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.2489417940378189, 'train/grad_norm_pre_clip': 8.799481391906738, 'timing/data': 0.011081095784902573, 'timing/model': 1.4262497909367085, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 0.08}) +[2026-06-15 19:24:05,687][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.24534055590629578, 'train/grad_norm_pre_clip': 8.53939151763916, 'timing/data': 0.01076153852045536, 'timing/model': 1.6223542355000973, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 0.08}) +[2026-06-15 19:24:11,743][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.28356045484542847, 'train/grad_norm_pre_clip': 5.234655857086182, 'timing/data': 0.010988783091306686, 'timing/model': 1.4847796112298965, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 0.08}) +[2026-06-15 19:24:17,728][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.5144352912902832, 'train/grad_norm_pre_clip': 10.828970909118652, 'timing/data': 0.01070033572614193, 'timing/model': 1.4604485370218754, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.09}) +[2026-06-15 19:24:23,714][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.2933851182460785, 'train/grad_norm_pre_clip': 9.179598808288574, 'timing/data': 0.010449284687638283, 'timing/model': 1.453539939597249, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 0.09}) +[2026-06-15 19:24:29,629][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.32007622718811035, 'train/grad_norm_pre_clip': 8.217461585998535, 'timing/data': 0.010047078132629395, 'timing/model': 1.4567092936486006, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 0.09}) +[2026-06-15 19:24:35,813][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.2554017901420593, 'train/grad_norm_pre_clip': 10.617673873901367, 'timing/data': 0.01200869120657444, 'timing/model': 1.737739922478795, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 0.09}) +[2026-06-15 19:24:41,516][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.4028950333595276, 'train/grad_norm_pre_clip': 19.121936798095703, 'timing/data': 0.010342137888073921, 'timing/model': 1.4557859599590302, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 0.09}) +[2026-06-15 19:24:47,418][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.6832517981529236, 'train/grad_norm_pre_clip': 20.21133041381836, 'timing/data': 0.010830281302332878, 'timing/model': 1.4577671717852354, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 0.09}) +[2026-06-15 19:24:53,224][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.28795939683914185, 'train/grad_norm_pre_clip': 5.489530563354492, 'timing/data': 0.014237560331821442, 'timing/model': 1.454746350646019, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 0.09}) +[2026-06-15 19:24:59,136][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.36187881231307983, 'train/grad_norm_pre_clip': 5.6976189613342285, 'timing/data': 0.011505907401442528, 'timing/model': 1.4583615846931934, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 0.09}) +[2026-06-15 19:25:05,133][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.3730451762676239, 'train/grad_norm_pre_clip': 8.676812171936035, 'timing/data': 0.01134442538022995, 'timing/model': 1.519610472023487, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 0.09}) +[2026-06-15 19:25:10,981][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.5128065347671509, 'train/grad_norm_pre_clip': 13.316902160644531, 'timing/data': 0.010866358876228333, 'timing/model': 1.616021966561675, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 0.09}) +[2026-06-15 19:25:17,331][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.36172571778297424, 'train/grad_norm_pre_clip': 9.423879623413086, 'timing/data': 0.013811929151415825, 'timing/model': 1.6938011720776558, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.09}) +[2026-06-15 19:25:23,323][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.3837582767009735, 'train/grad_norm_pre_clip': 3.8930959701538086, 'timing/data': 0.009482499212026596, 'timing/model': 1.5356862600892782, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 0.09}) +[2026-06-15 19:25:29,337][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.22023387253284454, 'train/grad_norm_pre_clip': 3.6647229194641113, 'timing/data': 0.009927680715918541, 'timing/model': 1.4653491154313087, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 0.09}) +[2026-06-15 19:25:35,418][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.4433111250400543, 'train/grad_norm_pre_clip': 14.273321151733398, 'timing/data': 0.01029910333454609, 'timing/model': 1.5316089559346437, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 0.09}) +[2026-06-15 19:25:41,309][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.2317907214164734, 'train/grad_norm_pre_clip': 11.094340324401855, 'timing/data': 0.010463463142514229, 'timing/model': 1.5095563791692257, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 0.09}) +[2026-06-15 19:25:47,405][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.3990175426006317, 'train/grad_norm_pre_clip': 18.295272827148438, 'timing/data': 0.011259187012910843, 'timing/model': 1.7006439454853535, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 0.09}) +[2026-06-15 19:25:53,085][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.2995617091655731, 'train/grad_norm_pre_clip': 18.175994873046875, 'timing/data': 0.011148719117045403, 'timing/model': 1.4407842084765434, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 0.09}) +[2026-06-15 19:25:59,103][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.29541799426078796, 'train/grad_norm_pre_clip': 9.636698722839355, 'timing/data': 0.011224955320358276, 'timing/model': 1.4765339568257332, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 0.09}) +[2026-06-15 19:26:05,290][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.2809019088745117, 'train/grad_norm_pre_clip': 9.696067810058594, 'timing/data': 0.011063303798437119, 'timing/model': 1.5060051772743464, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 0.09}) +[2026-06-15 19:26:11,447][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.326132595539093, 'train/grad_norm_pre_clip': 16.37302589416504, 'timing/data': 0.01190405897796154, 'timing/model': 1.5172602906823158, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 0.09}) +[2026-06-15 19:26:17,749][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.5097858309745789, 'train/grad_norm_pre_clip': 20.241453170776367, 'timing/data': 0.012224618345499039, 'timing/model': 1.5439567398279905, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.09}) +[2026-06-15 19:26:23,986][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.3873198926448822, 'train/grad_norm_pre_clip': 11.197426795959473, 'timing/data': 0.013220669701695442, 'timing/model': 1.7999181393533945, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 0.09}) +[2026-06-15 19:26:30,051][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.374838650226593, 'train/grad_norm_pre_clip': 15.172563552856445, 'timing/data': 0.012781212106347084, 'timing/model': 1.5693184304982424, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 0.09}) +[2026-06-15 19:26:36,254][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.18018846213817596, 'train/grad_norm_pre_clip': 12.479422569274902, 'timing/data': 0.013191316276788712, 'timing/model': 1.5396537110209465, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 0.09}) +[2026-06-15 19:26:42,574][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.723952054977417, 'train/grad_norm_pre_clip': 32.55342102050781, 'timing/data': 0.01217091828584671, 'timing/model': 1.5867301151156425, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 0.09}) +[2026-06-15 19:26:48,864][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.5136566758155823, 'train/grad_norm_pre_clip': 19.544843673706055, 'timing/data': 0.013188095763325691, 'timing/model': 1.5475946310907602, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 0.09}) +[2026-06-15 19:26:55,872][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.6688106656074524, 'train/grad_norm_pre_clip': 22.52878189086914, 'timing/data': 0.012345539405941963, 'timing/model': 1.5358556564897299, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 0.1}) +[2026-06-15 19:27:02,100][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.349651038646698, 'train/grad_norm_pre_clip': 13.123702049255371, 'timing/data': 0.011736208572983742, 'timing/model': 1.7274931259453297, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.1}) +[2026-06-15 19:27:08,143][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.2917095422744751, 'train/grad_norm_pre_clip': 5.572112560272217, 'timing/data': 0.013756485655903816, 'timing/model': 1.5611220244318247, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 0.1}) +[2026-06-15 19:27:14,578][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.6146559715270996, 'train/grad_norm_pre_clip': 11.173638343811035, 'timing/data': 0.012230733409523964, 'timing/model': 1.5987447835505009, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 0.1}) +[2026-06-15 19:27:20,948][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.33540022373199463, 'train/grad_norm_pre_clip': 8.713528633117676, 'timing/data': 0.013081321492791176, 'timing/model': 1.563413567841053, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.1}) +[2026-06-15 19:27:27,047][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.4339880645275116, 'train/grad_norm_pre_clip': 11.008612632751465, 'timing/data': 0.013026263564825058, 'timing/model': 1.553235100582242, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 0.1}) +[2026-06-15 19:27:33,247][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.25563445687294006, 'train/grad_norm_pre_clip': 6.969004154205322, 'timing/data': 0.013055408373475075, 'timing/model': 1.503871861845255, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 0.1}) +[2026-06-15 19:27:39,499][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.2215137779712677, 'train/grad_norm_pre_clip': 3.622771739959717, 'timing/data': 0.012896470725536346, 'timing/model': 1.8193140551447868, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 0.1}) +[2026-06-15 19:27:45,542][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.4597776532173157, 'train/grad_norm_pre_clip': 14.706647872924805, 'timing/data': 0.01252790167927742, 'timing/model': 1.5618835929781199, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 0.1}) +[2026-06-15 19:27:51,747][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.3688139319419861, 'train/grad_norm_pre_clip': 12.99852180480957, 'timing/data': 0.012023909017443657, 'timing/model': 1.5436575561761856, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 0.1}) +[2026-06-15 19:27:57,979][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.21262764930725098, 'train/grad_norm_pre_clip': 5.646860122680664, 'timing/data': 0.012044500559568405, 'timing/model': 1.5722611136734486, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 0.1}) +[2026-06-15 19:28:04,248][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.6349222660064697, 'train/grad_norm_pre_clip': 18.34854507446289, 'timing/data': 0.012287996709346771, 'timing/model': 1.553490824997425, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 0.1}) +[2026-06-15 19:28:10,451][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.44758203625679016, 'train/grad_norm_pre_clip': 6.54263162612915, 'timing/data': 0.012451333925127983, 'timing/model': 1.4900652207434177, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 0.1}) +[2026-06-15 19:28:16,646][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.32309770584106445, 'train/grad_norm_pre_clip': 6.075307846069336, 'timing/data': 0.01487351581454277, 'timing/model': 1.815276611596346, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 0.1}) +[2026-06-15 19:28:22,639][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.13177259266376495, 'train/grad_norm_pre_clip': 5.232865810394287, 'timing/data': 0.012285128235816956, 'timing/model': 1.5997050683945417, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.1}) +[2026-06-15 19:28:28,845][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.2507903277873993, 'train/grad_norm_pre_clip': 6.895757675170898, 'timing/data': 0.01265903189778328, 'timing/model': 1.5516926497220993, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 0.1}) +[2026-06-15 19:28:35,045][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.4630506634712219, 'train/grad_norm_pre_clip': 12.840288162231445, 'timing/data': 0.012459678575396538, 'timing/model': 1.5573204960674047, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 0.1}) +[2026-06-15 19:28:41,267][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.18690477311611176, 'train/grad_norm_pre_clip': 5.72249174118042, 'timing/data': 0.013048851862549782, 'timing/model': 1.5318466909229755, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 0.1}) +[2026-06-15 19:28:47,691][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.24951547384262085, 'train/grad_norm_pre_clip': 9.06702995300293, 'timing/data': 0.013070879504084587, 'timing/model': 1.5646155439317226, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 0.1}) +[2026-06-15 19:28:54,587][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.3188457787036896, 'train/grad_norm_pre_clip': 8.674660682678223, 'timing/data': 0.012414533644914627, 'timing/model': 1.809205986559391, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 0.1}) +[2026-06-15 19:29:00,824][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.22196745872497559, 'train/grad_norm_pre_clip': 5.393778324127197, 'timing/data': 0.01311882771551609, 'timing/model': 1.7428170405328274, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 0.1}) +[2026-06-15 19:29:06,843][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.11333660781383514, 'train/grad_norm_pre_clip': 6.012779235839844, 'timing/data': 0.013338001444935799, 'timing/model': 1.5602008774876595, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 0.1}) +[2026-06-15 19:29:13,342][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.29837465286254883, 'train/grad_norm_pre_clip': 11.293578147888184, 'timing/data': 0.03374497778713703, 'timing/model': 1.6301699317991734, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 0.1}) +[2026-06-15 19:29:19,635][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.35673296451568604, 'train/grad_norm_pre_clip': 15.71542739868164, 'timing/data': 0.012013599276542664, 'timing/model': 1.6438955068588257, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 0.1}) +[2026-06-15 19:29:25,859][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.4803010821342468, 'train/grad_norm_pre_clip': 12.81292724609375, 'timing/data': 0.013114843517541885, 'timing/model': 1.517375886440277, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.1}) +[2026-06-15 19:29:32,303][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.4255915880203247, 'train/grad_norm_pre_clip': 19.175395965576172, 'timing/data': 0.013279026374220848, 'timing/model': 1.807431399822235, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 0.1}) +[2026-06-15 19:29:38,447][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.2925833761692047, 'train/grad_norm_pre_clip': 18.28829002380371, 'timing/data': 0.012323327362537384, 'timing/model': 1.537518011406064, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 0.11}) +[2026-06-15 19:29:44,673][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.2792108952999115, 'train/grad_norm_pre_clip': 9.133354187011719, 'timing/data': 0.012792598456144333, 'timing/model': 1.5536092910915613, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 0.11}) +[2026-06-15 19:29:50,943][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.26808470487594604, 'train/grad_norm_pre_clip': 19.43116569519043, 'timing/data': 0.012765895575284958, 'timing/model': 1.4815671034157276, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 0.11}) +[2026-06-15 19:29:57,238][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.45117658376693726, 'train/grad_norm_pre_clip': 15.628181457519531, 'timing/data': 0.01256239041686058, 'timing/model': 1.5915150735527277, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.11}) +[2026-06-15 19:30:03,475][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.44192081689834595, 'train/grad_norm_pre_clip': 20.385944366455078, 'timing/data': 0.012210270389914513, 'timing/model': 1.5297592598944902, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 0.11}) +[2026-06-15 19:30:09,735][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.3051316738128662, 'train/grad_norm_pre_clip': 8.037019729614258, 'timing/data': 0.01273326575756073, 'timing/model': 1.7392266802489758, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 0.11}) +[2026-06-15 19:30:15,758][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.2604759931564331, 'train/grad_norm_pre_clip': 26.179250717163086, 'timing/data': 0.012947678565979004, 'timing/model': 1.5313730109483004, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 0.11}) +[2026-06-15 19:30:22,178][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.20614171028137207, 'train/grad_norm_pre_clip': 5.358424186706543, 'timing/data': 0.012327488511800766, 'timing/model': 1.5942809116095304, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 0.11}) +[2026-06-15 19:30:28,574][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.3131379783153534, 'train/grad_norm_pre_clip': 7.142933368682861, 'timing/data': 0.012791534885764122, 'timing/model': 1.5710868816822767, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.11}) +[2026-06-15 19:30:34,973][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.29841819405555725, 'train/grad_norm_pre_clip': 8.944432258605957, 'timing/data': 0.0125461146235466, 'timing/model': 1.6104240603744984, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 0.11}) +[2026-06-15 19:30:41,390][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.37618371844291687, 'train/grad_norm_pre_clip': 6.99162483215332, 'timing/data': 0.012699944898486137, 'timing/model': 1.6069001946598291, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 0.11}) +[2026-06-15 19:30:47,952][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.38991430401802063, 'train/grad_norm_pre_clip': 10.632672309875488, 'timing/data': 0.012593179941177368, 'timing/model': 1.7827270943671465, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 0.11}) +[2026-06-15 19:30:54,112][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.23000481724739075, 'train/grad_norm_pre_clip': 6.9974751472473145, 'timing/data': 0.014250867068767548, 'timing/model': 1.5445422288030386, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 0.11}) +[2026-06-15 19:31:00,481][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.19109317660331726, 'train/grad_norm_pre_clip': 6.2748613357543945, 'timing/data': 0.012778382748365402, 'timing/model': 1.549547677859664, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 0.11}) +[2026-06-15 19:31:06,795][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.18838557600975037, 'train/grad_norm_pre_clip': 8.323176383972168, 'timing/data': 0.012055279687047005, 'timing/model': 1.5716438628733158, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 0.11}) +[2026-06-15 19:31:13,191][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.16837573051452637, 'train/grad_norm_pre_clip': 7.231723785400391, 'timing/data': 0.012210585176944733, 'timing/model': 1.5737262461334467, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 0.11}) +[2026-06-15 19:31:19,525][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.6443915367126465, 'train/grad_norm_pre_clip': 33.570438385009766, 'timing/data': 0.01301567256450653, 'timing/model': 1.7737748883664608, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 0.11}) +[2026-06-15 19:31:25,590][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.3993995189666748, 'train/grad_norm_pre_clip': 15.853438377380371, 'timing/data': 0.012865094467997551, 'timing/model': 1.5322453416883945, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 0.11}) +[2026-06-15 19:31:31,891][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.21995893120765686, 'train/grad_norm_pre_clip': 10.75341796875, 'timing/data': 0.012303467839956284, 'timing/model': 1.542165920138359, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.11}) +[2026-06-15 19:31:38,050][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.3762367069721222, 'train/grad_norm_pre_clip': 30.4831600189209, 'timing/data': 0.01281476579606533, 'timing/model': 1.5373658053576946, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 0.11}) +[2026-06-15 19:31:44,148][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.7035928964614868, 'train/grad_norm_pre_clip': 34.21062088012695, 'timing/data': 0.012359831482172012, 'timing/model': 1.5565922148525715, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 0.11}) +[2026-06-15 19:31:50,250][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.6556038856506348, 'train/grad_norm_pre_clip': 47.3809700012207, 'timing/data': 0.012440219521522522, 'timing/model': 1.485220555216074, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 0.11}) +[2026-06-15 19:31:56,560][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.2849160432815552, 'train/grad_norm_pre_clip': 23.17378807067871, 'timing/data': 0.01240580528974533, 'timing/model': 1.8618138637393713, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 0.11}) +[2026-06-15 19:32:02,655][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.18608678877353668, 'train/grad_norm_pre_clip': 15.045454025268555, 'timing/data': 0.012691739946603775, 'timing/model': 1.5317142568528652, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 0.11}) +[2026-06-15 19:32:09,051][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.3859599828720093, 'train/grad_norm_pre_clip': 20.951875686645508, 'timing/data': 0.011092755943536758, 'timing/model': 1.6419896241277456, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.11}) +[2026-06-15 19:32:15,437][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.4910694658756256, 'train/grad_norm_pre_clip': 28.764354705810547, 'timing/data': 0.011302275583148003, 'timing/model': 1.6411075666546822, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 0.11}) +[2026-06-15 19:32:21,546][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.4469835162162781, 'train/grad_norm_pre_clip': 41.103858947753906, 'timing/data': 0.01250704936683178, 'timing/model': 1.5560488756746054, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 0.12}) +[2026-06-15 19:32:27,724][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.23240689933300018, 'train/grad_norm_pre_clip': 10.742268562316895, 'timing/data': 0.012225985527038574, 'timing/model': 1.5516332052648067, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 0.12}) +[2026-06-15 19:39:52,302][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.5433692336082458, 'train/grad_norm_pre_clip': 11.249452590942383, 'eval/loss': 0.1698370724916458, 'eval/flap_precision': 0.9312977099236641, 'eval/flap_recall': 0.6621438263229308, 'eval/flap_f1': 0.7739888977002378, 'eval/flap_rate_model': 0.14555555555555555, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.7739888977002379, 'timing/data': 0.011385906487703323, 'timing/model': 1.8081430960446596, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.12}) +[2026-06-15 19:39:52,303][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_300_state +[2026-06-15 19:39:52,308][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 19:39:59,450][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_300_state/model.safetensors +[2026-06-15 19:40:13,847][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_300_state/optimizer.bin +[2026-06-15 19:40:13,847][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_300_state/sampler.bin +[2026-06-15 19:40:13,847][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_300_state/sampler_1.bin +[2026-06-15 19:40:13,848][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-15 19:40:13,959][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-15 19:40:13,959][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 19:40:13,967][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 19:40:19,890][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.48759135603904724, 'train/grad_norm_pre_clip': 33.17839050292969, 'timing/data': 0.009724659845232964, 'timing/model': 1.6795969419181347, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 0.12}) +[2026-06-15 19:40:25,753][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.2367551475763321, 'train/grad_norm_pre_clip': 16.97817039489746, 'timing/data': 0.010531321167945862, 'timing/model': 1.6428775507956743, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 0.12}) +[2026-06-15 19:40:31,579][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.24160507321357727, 'train/grad_norm_pre_clip': 15.707741737365723, 'timing/data': 0.010239783674478531, 'timing/model': 1.4674751535058022, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 0.12}) +[2026-06-15 19:40:37,733][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.3815111517906189, 'train/grad_norm_pre_clip': 10.14627742767334, 'timing/data': 0.012638211250305176, 'timing/model': 1.5512237511575222, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 0.12}) +[2026-06-15 19:40:43,637][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.18667779862880707, 'train/grad_norm_pre_clip': 4.611798286437988, 'timing/data': 0.010986251756548882, 'timing/model': 1.5236456468701363, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 0.12}) +[2026-06-15 19:40:49,788][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.2525726854801178, 'train/grad_norm_pre_clip': 10.146714210510254, 'timing/data': 0.012633521109819412, 'timing/model': 1.5587362069636583, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 0.12}) +[2026-06-15 19:40:56,926][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.18968625366687775, 'train/grad_norm_pre_clip': 12.73231315612793, 'timing/data': 0.010728316381573677, 'timing/model': 1.8115285113453865, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 0.12}) +[2026-06-15 19:41:03,095][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.2582858204841614, 'train/grad_norm_pre_clip': 10.254754066467285, 'timing/data': 0.0148308165371418, 'timing/model': 1.678595745936036, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 0.12}) +[2026-06-15 19:41:09,383][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.29111677408218384, 'train/grad_norm_pre_clip': 10.448928833007812, 'timing/data': 0.013176916167140007, 'timing/model': 1.57421232201159, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 0.12}) +[2026-06-15 19:41:15,774][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.18723885715007782, 'train/grad_norm_pre_clip': 12.451249122619629, 'timing/data': 0.01381935179233551, 'timing/model': 1.5571004282683134, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.12}) +[2026-06-15 19:41:22,186][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.14248454570770264, 'train/grad_norm_pre_clip': 5.832310676574707, 'timing/data': 0.01410749927163124, 'timing/model': 1.5274427309632301, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 0.12}) +[2026-06-15 19:41:28,758][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.4117511212825775, 'train/grad_norm_pre_clip': 9.814964294433594, 'timing/data': 0.014277944341301918, 'timing/model': 1.5919148158282042, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 0.12}) +[2026-06-15 19:41:35,111][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.4738222360610962, 'train/grad_norm_pre_clip': 12.179606437683105, 'timing/data': 0.014307305216789246, 'timing/model': 1.7904240880161524, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 0.12}) +[2026-06-15 19:41:41,307][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.2740764617919922, 'train/grad_norm_pre_clip': 13.098174095153809, 'timing/data': 0.011998435482382774, 'timing/model': 1.6913206744939089, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 0.12}) +[2026-06-15 19:41:47,481][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.2367001622915268, 'train/grad_norm_pre_clip': 7.073999404907227, 'timing/data': 0.017299378290772438, 'timing/model': 1.547809237614274, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 0.12}) +[2026-06-15 19:41:53,788][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.21630524098873138, 'train/grad_norm_pre_clip': 6.677618980407715, 'timing/data': 0.01314421184360981, 'timing/model': 1.555892614647746, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 0.12}) +[2026-06-15 19:42:00,080][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.3996621072292328, 'train/grad_norm_pre_clip': 21.243167877197266, 'timing/data': 0.013421742245554924, 'timing/model': 1.5563638526946306, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 0.12}) +[2026-06-15 19:42:06,383][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.35124433040618896, 'train/grad_norm_pre_clip': 16.770475387573242, 'timing/data': 0.013822175562381744, 'timing/model': 1.5170209035277367, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 0.12}) +[2026-06-15 19:42:13,030][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.25018608570098877, 'train/grad_norm_pre_clip': 9.563288688659668, 'timing/data': 0.013011449947953224, 'timing/model': 1.8066350799053907, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 0.12}) +[2026-06-15 19:42:19,070][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.10630286484956741, 'train/grad_norm_pre_clip': 2.962397813796997, 'timing/data': 0.011941229924559593, 'timing/model': 1.5537966340780258, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.12}) +[2026-06-15 19:42:25,470][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.1926145702600479, 'train/grad_norm_pre_clip': 11.293282508850098, 'timing/data': 0.01279069110751152, 'timing/model': 1.6499611753970385, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 0.12}) +[2026-06-15 19:42:31,756][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.5705693960189819, 'train/grad_norm_pre_clip': 11.528432846069336, 'timing/data': 0.012118939310312271, 'timing/model': 1.5541020799428225, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 0.12}) +[2026-06-15 19:42:37,958][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.46309754252433777, 'train/grad_norm_pre_clip': 16.52890396118164, 'timing/data': 0.011366436257958412, 'timing/model': 1.544099822640419, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 0.13}) +[2026-06-15 19:42:44,331][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.16132129728794098, 'train/grad_norm_pre_clip': 10.425389289855957, 'timing/data': 0.012377023696899414, 'timing/model': 1.5549158193171024, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 0.13}) +[2026-06-15 19:42:50,569][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.07485634088516235, 'train/grad_norm_pre_clip': 4.495593547821045, 'timing/data': 0.01149263046681881, 'timing/model': 1.7835328876972198, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 0.13}) +[2026-06-15 19:42:56,657][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.6378294229507446, 'train/grad_norm_pre_clip': 18.824649810791016, 'timing/data': 0.011898044496774673, 'timing/model': 1.540623327717185, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 0.13}) +[2026-06-15 19:43:02,950][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.28522101044654846, 'train/grad_norm_pre_clip': 34.69849395751953, 'timing/data': 0.012006843462586403, 'timing/model': 1.545280136168003, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 0.13}) +[2026-06-15 19:43:09,260][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.24359416961669922, 'train/grad_norm_pre_clip': 16.16136932373047, 'timing/data': 0.01193808764219284, 'timing/model': 1.5478271152824163, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 0.13}) +[2026-06-15 19:43:15,349][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.34603267908096313, 'train/grad_norm_pre_clip': 25.670116424560547, 'timing/data': 0.01160566508769989, 'timing/model': 1.5035013761371374, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 0.13}) +[2026-06-15 19:43:21,560][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.43211299180984497, 'train/grad_norm_pre_clip': 22.4054012298584, 'timing/data': 0.012537788599729538, 'timing/model': 1.5669855307787657, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.13}) +[2026-06-15 19:43:27,798][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.3358212113380432, 'train/grad_norm_pre_clip': 13.465337753295898, 'timing/data': 0.011599132791161537, 'timing/model': 1.801629999652505, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 0.13}) +[2026-06-15 19:43:34,105][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.40408796072006226, 'train/grad_norm_pre_clip': 8.849000930786133, 'timing/data': 0.012411097064614296, 'timing/model': 1.8697462193667889, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 0.13}) +[2026-06-15 19:43:40,246][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.34672120213508606, 'train/grad_norm_pre_clip': 18.448820114135742, 'timing/data': 0.011909721419215202, 'timing/model': 1.555258670821786, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 0.13}) +[2026-06-15 19:43:46,361][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.29395273327827454, 'train/grad_norm_pre_clip': 9.090324401855469, 'timing/data': 0.01255984976887703, 'timing/model': 1.4872644878923893, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 0.13}) +[2026-06-15 19:43:52,654][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.33100900053977966, 'train/grad_norm_pre_clip': 11.345486640930176, 'timing/data': 0.01303388737142086, 'timing/model': 1.5585429277271032, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 0.13}) +[2026-06-15 19:43:58,845][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.20758739113807678, 'train/grad_norm_pre_clip': 5.751999855041504, 'timing/data': 0.011869523674249649, 'timing/model': 1.5628561489284039, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 0.13}) +[2026-06-15 19:44:05,488][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.24386654794216156, 'train/grad_norm_pre_clip': 5.598018646240234, 'timing/data': 0.012684237211942673, 'timing/model': 1.7997103240340948, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 0.13}) +[2026-06-15 19:44:11,563][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.15459315478801727, 'train/grad_norm_pre_clip': 6.313194274902344, 'timing/data': 0.01301933079957962, 'timing/model': 1.5536321252584457, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 0.13}) +[2026-06-15 19:44:17,863][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.5140328407287598, 'train/grad_norm_pre_clip': 14.07620906829834, 'timing/data': 0.012710744515061378, 'timing/model': 1.5645061619579792, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 0.13}) +[2026-06-15 19:44:24,161][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.4184127151966095, 'train/grad_norm_pre_clip': 13.002762794494629, 'timing/data': 0.012104922905564308, 'timing/model': 1.5665277279913425, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.13}) +[2026-06-15 19:44:30,374][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.45602813363075256, 'train/grad_norm_pre_clip': 12.462076187133789, 'timing/data': 0.012024272233247757, 'timing/model': 1.5039195511490107, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 0.13}) +[2026-06-15 19:44:36,642][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.2624712288379669, 'train/grad_norm_pre_clip': 12.728659629821777, 'timing/data': 0.013168461620807648, 'timing/model': 1.5635970681905746, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 0.13}) +[2026-06-15 19:44:42,907][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.2603558897972107, 'train/grad_norm_pre_clip': 10.855321884155273, 'timing/data': 0.011797547340393066, 'timing/model': 1.8112704064697027, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 0.13}) +[2026-06-15 19:44:48,948][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.2984618842601776, 'train/grad_norm_pre_clip': 10.844058990478516, 'timing/data': 0.01278752088546753, 'timing/model': 1.5380470994859934, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 0.13}) +[2026-06-15 19:44:55,155][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.16212177276611328, 'train/grad_norm_pre_clip': 4.802700996398926, 'timing/data': 0.012122338637709618, 'timing/model': 1.542961023747921, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.13}) +[2026-06-15 19:45:01,460][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.2123742252588272, 'train/grad_norm_pre_clip': 6.350131511688232, 'timing/data': 0.012622741982340813, 'timing/model': 1.5601024702191353, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 0.13}) +[2026-06-15 19:45:07,643][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.13317923247814178, 'train/grad_norm_pre_clip': 6.02307653427124, 'timing/data': 0.012697681784629822, 'timing/model': 1.506738344207406, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 0.13}) +[2026-06-15 19:45:13,744][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.2560597062110901, 'train/grad_norm_pre_clip': 9.281922340393066, 'timing/data': 0.012756725773215294, 'timing/model': 1.4847592450678349, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 0.13}) +[2026-06-15 19:45:20,079][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.13965070247650146, 'train/grad_norm_pre_clip': 4.731906414031982, 'timing/data': 0.013128060847520828, 'timing/model': 1.7619829270988703, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 0.14}) +[2026-06-15 19:45:26,157][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.5416407585144043, 'train/grad_norm_pre_clip': 24.42733383178711, 'timing/data': 0.012444216758012772, 'timing/model': 1.550613658502698, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.14}) +[2026-06-15 19:45:32,460][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.4352581202983856, 'train/grad_norm_pre_clip': 26.396568298339844, 'timing/data': 0.012813106179237366, 'timing/model': 1.552186779677868, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 0.14}) +[2026-06-15 19:45:38,662][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.6008324027061462, 'train/grad_norm_pre_clip': 24.678102493286133, 'timing/data': 0.012442577630281448, 'timing/model': 1.5462957601994276, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 0.14}) +[2026-06-15 19:45:44,943][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.5725428462028503, 'train/grad_norm_pre_clip': 17.675334930419922, 'timing/data': 0.012003172188997269, 'timing/model': 1.5004608165472746, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 0.14}) +[2026-06-15 19:45:51,306][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.1498475819826126, 'train/grad_norm_pre_clip': 10.934450149536133, 'timing/data': 0.01239333488047123, 'timing/model': 1.9284263364970684, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 0.14}) +[2026-06-15 19:45:57,458][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.0756944790482521, 'train/grad_norm_pre_clip': 3.8911914825439453, 'timing/data': 0.01237129233777523, 'timing/model': 1.5460034105926752, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 0.14}) +[2026-06-15 19:46:03,688][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.27948668599128723, 'train/grad_norm_pre_clip': 21.794532775878906, 'timing/data': 0.012868201360106468, 'timing/model': 1.5739524848759174, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 0.14}) +[2026-06-15 19:46:09,875][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.1274922788143158, 'train/grad_norm_pre_clip': 8.266420364379883, 'timing/data': 0.012760691344738007, 'timing/model': 1.5555497780442238, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 0.14}) +[2026-06-15 19:46:16,047][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.09683628380298615, 'train/grad_norm_pre_clip': 4.486093521118164, 'timing/data': 0.012338865548372269, 'timing/model': 1.558943884447217, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 0.14}) +[2026-06-15 19:46:22,238][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.15290498733520508, 'train/grad_norm_pre_clip': 13.516362190246582, 'timing/data': 0.012794453650712967, 'timing/model': 1.5606382638216019, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 0.14}) +[2026-06-15 19:46:28,519][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.2164144366979599, 'train/grad_norm_pre_clip': 9.762839317321777, 'timing/data': 0.012140845879912376, 'timing/model': 1.747707137838006, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.14}) +[2026-06-15 19:46:34,571][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.1319393664598465, 'train/grad_norm_pre_clip': 6.038562297821045, 'timing/data': 0.012514252215623856, 'timing/model': 1.5813097897917032, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 0.14}) +[2026-06-15 19:46:40,867][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.13064682483673096, 'train/grad_norm_pre_clip': 14.485231399536133, 'timing/data': 0.01158120296895504, 'timing/model': 1.562184313312173, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 0.14}) +[2026-06-15 19:46:47,374][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.4784877300262451, 'train/grad_norm_pre_clip': 28.52141571044922, 'timing/data': 0.0138760507106781, 'timing/model': 1.7332298941910267, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 0.14}) +[2026-06-15 19:46:53,854][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.11069027334451675, 'train/grad_norm_pre_clip': 10.664776802062988, 'timing/data': 0.012850373983383179, 'timing/model': 1.7114586681127548, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 0.14}) +[2026-06-15 19:47:00,093][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.06105076149106026, 'train/grad_norm_pre_clip': 11.269607543945312, 'timing/data': 0.01270853728055954, 'timing/model': 1.5199231635779142, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 0.14}) +[2026-06-15 19:47:06,513][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.21391606330871582, 'train/grad_norm_pre_clip': 14.225005149841309, 'timing/data': 0.012997699901461601, 'timing/model': 1.7837497405707836, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 0.14}) +[2026-06-15 19:47:12,698][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.36554041504859924, 'train/grad_norm_pre_clip': 17.186199188232422, 'timing/data': 0.014439297839999199, 'timing/model': 1.5368299763649702, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 0.14}) +[2026-06-15 19:47:19,169][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.24161484837532043, 'train/grad_norm_pre_clip': 13.163772583007812, 'timing/data': 0.01421450823545456, 'timing/model': 1.6401579603552818, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 0.14}) +[2026-06-15 19:47:25,583][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.3168179988861084, 'train/grad_norm_pre_clip': 15.819849967956543, 'timing/data': 0.012190597131848335, 'timing/model': 1.5607722271233797, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 0.14}) +[2026-06-15 19:47:31,988][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.7202390432357788, 'train/grad_norm_pre_clip': 32.40492248535156, 'timing/data': 0.012817373499274254, 'timing/model': 1.5863738097250462, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.14}) +[2026-06-15 19:47:38,273][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.5479941964149475, 'train/grad_norm_pre_clip': 13.153142929077148, 'timing/data': 0.012740552425384521, 'timing/model': 1.5467695891857147, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 0.14}) +[2026-06-15 19:47:44,612][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.01969938352704048, 'train/grad_norm_pre_clip': 2.7789669036865234, 'timing/data': 0.011794036254286766, 'timing/model': 1.781582998111844, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 0.14}) +[2026-06-15 19:47:50,643][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.20647311210632324, 'train/grad_norm_pre_clip': 8.273591995239258, 'timing/data': 0.012540128082036972, 'timing/model': 1.535620991140604, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 0.14}) +[2026-06-15 19:47:56,842][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.7577691674232483, 'train/grad_norm_pre_clip': 35.03153991699219, 'timing/data': 0.011371348053216934, 'timing/model': 1.559346167370677, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 0.14}) +[2026-06-15 19:48:03,042][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.4069881737232208, 'train/grad_norm_pre_clip': 28.425907135009766, 'timing/data': 0.013302024453878403, 'timing/model': 1.5519932322204113, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 0.15}) +[2026-06-15 19:48:09,148][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.17853888869285583, 'train/grad_norm_pre_clip': 9.519248008728027, 'timing/data': 0.012659966945648193, 'timing/model': 1.4860415030270815, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 0.15}) +[2026-06-15 19:48:16,088][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.5783771872520447, 'train/grad_norm_pre_clip': 18.983739852905273, 'timing/data': 0.016323285177350044, 'timing/model': 2.19082061201334, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 0.15}) +[2026-06-15 19:48:22,431][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.2422284632921219, 'train/grad_norm_pre_clip': 7.5418782234191895, 'timing/data': 0.012125788256525993, 'timing/model': 1.8052163887768984, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 0.15}) +[2026-06-15 19:48:28,441][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.304121732711792, 'train/grad_norm_pre_clip': 8.064424514770508, 'timing/data': 0.012151828035712242, 'timing/model': 1.5173382423818111, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 0.15}) +[2026-06-15 19:48:34,843][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.20287536084651947, 'train/grad_norm_pre_clip': 8.106307029724121, 'timing/data': 0.013743691146373749, 'timing/model': 1.5633895304054022, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.15}) +[2026-06-15 19:48:41,072][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.3116844892501831, 'train/grad_norm_pre_clip': 9.48231029510498, 'timing/data': 0.012575389817357063, 'timing/model': 1.5850094426423311, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 0.15}) +[2026-06-15 19:48:47,361][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.37094080448150635, 'train/grad_norm_pre_clip': 7.199596405029297, 'timing/data': 0.01154782623052597, 'timing/model': 1.572832241654396, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 0.15}) +[2026-06-15 19:48:53,543][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.3025321960449219, 'train/grad_norm_pre_clip': 7.8946051597595215, 'timing/data': 0.012953057885169983, 'timing/model': 1.5116069745272398, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 0.15}) +[2026-06-15 19:48:59,871][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.29476386308670044, 'train/grad_norm_pre_clip': 3.8379909992218018, 'timing/data': 0.011884655803442001, 'timing/model': 1.7895677965134382, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 0.15}) +[2026-06-15 19:49:05,944][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.31478047370910645, 'train/grad_norm_pre_clip': 9.7468843460083, 'timing/data': 0.011897673830389977, 'timing/model': 1.5581829026341438, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 0.15}) +[2026-06-15 19:49:12,144][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.2704472243785858, 'train/grad_norm_pre_clip': 8.093697547912598, 'timing/data': 0.01249019242823124, 'timing/model': 1.5527633670717478, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 0.15}) +[2026-06-15 19:49:18,256][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.3599996268749237, 'train/grad_norm_pre_clip': 10.536920547485352, 'timing/data': 0.012255718931555748, 'timing/model': 1.5651336591690779, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 0.15}) +[2026-06-15 19:49:24,459][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.3271688222885132, 'train/grad_norm_pre_clip': 8.193071365356445, 'timing/data': 0.012875564396381378, 'timing/model': 1.5510059297084808, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 0.15}) +[2026-06-15 19:49:30,745][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.30953913927078247, 'train/grad_norm_pre_clip': 9.236828804016113, 'timing/data': 0.0132354237139225, 'timing/model': 1.5545324329286814, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 0.15}) +[2026-06-15 19:49:37,402][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.3218047618865967, 'train/grad_norm_pre_clip': 5.225984573364258, 'timing/data': 0.014678193256258965, 'timing/model': 1.9046342745423317, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.15}) +[2026-06-15 19:49:43,654][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.31903794407844543, 'train/grad_norm_pre_clip': 11.596860885620117, 'timing/data': 0.013316966593265533, 'timing/model': 1.5609531048685312, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 0.15}) +[2026-06-15 19:49:49,972][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.16984185576438904, 'train/grad_norm_pre_clip': 17.11548614501953, 'timing/data': 0.013191916048526764, 'timing/model': 1.5583339221775532, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 0.15}) +[2026-06-15 19:49:56,194][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.21979722380638123, 'train/grad_norm_pre_clip': 7.854018688201904, 'timing/data': 0.011662142351269722, 'timing/model': 1.5214289035648108, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 0.15}) +[2026-06-15 19:50:02,445][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.4817362129688263, 'train/grad_norm_pre_clip': 12.602404594421387, 'timing/data': 0.012416651472449303, 'timing/model': 1.554591516032815, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.15}) +[2026-06-15 19:50:08,645][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.270271360874176, 'train/grad_norm_pre_clip': 20.706003189086914, 'timing/data': 0.013027140870690346, 'timing/model': 1.5127000603824854, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 0.15}) +[2026-06-15 19:50:14,882][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.10698101669549942, 'train/grad_norm_pre_clip': 4.985708236694336, 'timing/data': 0.012494351714849472, 'timing/model': 1.7975441720336676, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 0.15}) +[2026-06-15 19:50:21,116][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.13909964263439178, 'train/grad_norm_pre_clip': 10.088353157043457, 'timing/data': 0.012622382491827011, 'timing/model': 1.7311680987477303, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 0.15}) +[2026-06-15 19:50:27,138][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.16952447593212128, 'train/grad_norm_pre_clip': 28.17595672607422, 'timing/data': 0.01126069389283657, 'timing/model': 1.5597115885466337, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 0.15}) +[2026-06-15 19:50:33,339][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.12822626531124115, 'train/grad_norm_pre_clip': 9.1530122756958, 'timing/data': 0.013128982856869698, 'timing/model': 1.5625592451542616, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 0.15}) +[2026-06-15 19:58:00,935][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.29072314500808716, 'train/grad_norm_pre_clip': 11.109748840332031, 'eval/loss': 0.1663447469472885, 'eval/flap_precision': 0.9015873015873016, 'eval/flap_recall': 0.7706919945725916, 'eval/flap_f1': 0.8310168251645942, 'eval/flap_rate_model': 0.175, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8310168251645942, 'timing/data': 0.011573955416679382, 'timing/model': 1.5661475509405136, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.15}) +[2026-06-15 19:58:00,936][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_400_state +[2026-06-15 19:58:00,939][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 19:58:07,157][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_400_state/model.safetensors +[2026-06-15 19:58:21,563][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_400_state/optimizer.bin +[2026-06-15 19:58:21,564][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_400_state/sampler.bin +[2026-06-15 19:58:21,564][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_400_state/sampler_1.bin +[2026-06-15 19:58:21,565][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-15 19:58:21,663][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-15 19:58:21,664][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 19:58:21,672][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 19:58:27,598][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.44861048460006714, 'train/grad_norm_pre_clip': 13.892101287841797, 'timing/data': 0.010485174134373665, 'timing/model': 1.4532749466598034, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 0.16}) +[2026-06-15 19:58:33,519][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.16617506742477417, 'train/grad_norm_pre_clip': 12.383772850036621, 'timing/data': 0.012055087834596634, 'timing/model': 1.5055371932685375, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 0.16}) +[2026-06-15 19:58:39,320][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.20346999168395996, 'train/grad_norm_pre_clip': 18.15629005432129, 'timing/data': 0.011891067028045654, 'timing/model': 1.471044186502695, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 0.16}) +[2026-06-15 19:58:45,234][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.21334964036941528, 'train/grad_norm_pre_clip': 12.644291877746582, 'timing/data': 0.010922158136963844, 'timing/model': 1.6854918953031301, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 0.16}) +[2026-06-15 19:58:51,174][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.22995589673519135, 'train/grad_norm_pre_clip': 18.383411407470703, 'timing/data': 0.012981276959180832, 'timing/model': 1.5625892747193575, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 0.16}) +[2026-06-15 19:58:57,473][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.3528706729412079, 'train/grad_norm_pre_clip': 14.940934181213379, 'timing/data': 0.013061443343758583, 'timing/model': 1.5675762370228767, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 0.16}) +[2026-06-15 19:59:03,898][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.22300906479358673, 'train/grad_norm_pre_clip': 10.987215042114258, 'timing/data': 0.0118086077272892, 'timing/model': 1.5874966941773891, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 0.16}) +[2026-06-15 19:59:10,171][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.295981764793396, 'train/grad_norm_pre_clip': 11.743810653686523, 'timing/data': 0.012617850676178932, 'timing/model': 1.5623179245740175, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 0.16}) +[2026-06-15 19:59:16,502][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.2866356074810028, 'train/grad_norm_pre_clip': 9.150825500488281, 'timing/data': 0.013043493032455444, 'timing/model': 1.79628137499094, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 0.16}) +[2026-06-15 19:59:22,709][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.2508009076118469, 'train/grad_norm_pre_clip': 8.832139015197754, 'timing/data': 0.01246483251452446, 'timing/model': 1.7321357354521751, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.16}) +[2026-06-15 19:59:28,647][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.11577355861663818, 'train/grad_norm_pre_clip': 6.3589558601379395, 'timing/data': 0.01269318163394928, 'timing/model': 1.4830246455967426, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 0.16}) +[2026-06-15 19:59:34,838][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.313808411359787, 'train/grad_norm_pre_clip': 8.194497108459473, 'timing/data': 0.01187724806368351, 'timing/model': 1.5636641513556242, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 0.16}) +[2026-06-15 19:59:41,036][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.3633345365524292, 'train/grad_norm_pre_clip': 7.593324661254883, 'timing/data': 0.011303139850497246, 'timing/model': 1.5514271222054958, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 0.16}) +[2026-06-15 19:59:47,262][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.27980268001556396, 'train/grad_norm_pre_clip': 6.17257022857666, 'timing/data': 0.012583917006850243, 'timing/model': 1.5688149370253086, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 0.16}) +[2026-06-15 19:59:53,361][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.23986120522022247, 'train/grad_norm_pre_clip': 6.8084845542907715, 'timing/data': 0.012435587123036385, 'timing/model': 1.4883679244667292, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.16}) +[2026-06-15 19:59:59,538][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.3653305470943451, 'train/grad_norm_pre_clip': 9.477520942687988, 'timing/data': 0.012496890500187874, 'timing/model': 1.755267584696412, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 0.16}) +[2026-06-15 20:00:05,944][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.1751733422279358, 'train/grad_norm_pre_clip': 8.051980018615723, 'timing/data': 0.012571888044476509, 'timing/model': 1.7481031958013773, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 0.16}) +[2026-06-15 20:00:12,943][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.2619052231311798, 'train/grad_norm_pre_clip': 5.174084186553955, 'timing/data': 0.01381184346973896, 'timing/model': 1.6561620775610209, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 0.16}) +[2026-06-15 20:00:19,743][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.37869083881378174, 'train/grad_norm_pre_clip': 11.549373626708984, 'timing/data': 0.012702826410531998, 'timing/model': 1.9554063398391008, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 0.16}) +[2026-06-15 20:00:25,943][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.20829929411411285, 'train/grad_norm_pre_clip': 8.532466888427734, 'timing/data': 0.012525852769613266, 'timing/model': 1.5152771640568972, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.16}) +[2026-06-15 20:00:32,189][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.11906161904335022, 'train/grad_norm_pre_clip': 5.5849432945251465, 'timing/data': 0.0130649134516716, 'timing/model': 1.8077990058809519, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 0.16}) +[2026-06-15 20:00:38,364][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.14634308218955994, 'train/grad_norm_pre_clip': 9.919591903686523, 'timing/data': 0.01053394004702568, 'timing/model': 1.6532668638974428, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 0.16}) +[2026-06-15 20:00:44,692][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.24725903570652008, 'train/grad_norm_pre_clip': 36.36274719238281, 'timing/data': 0.01259683258831501, 'timing/model': 1.5679319091141224, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 0.16}) +[2026-06-15 20:00:50,873][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.2035982459783554, 'train/grad_norm_pre_clip': 12.796507835388184, 'timing/data': 0.0131157748401165, 'timing/model': 1.5400295723229647, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 0.16}) +[2026-06-15 20:00:57,175][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.1717672348022461, 'train/grad_norm_pre_clip': 14.20058822631836, 'timing/data': 0.013280361890792847, 'timing/model': 1.5518556851893663, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 0.16}) +[2026-06-15 20:01:03,475][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.3680187165737152, 'train/grad_norm_pre_clip': 10.68893814086914, 'timing/data': 0.012751540169119835, 'timing/model': 1.561725489795208, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 0.16}) +[2026-06-15 20:01:09,816][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.2717021405696869, 'train/grad_norm_pre_clip': 9.847562789916992, 'timing/data': 0.012608053162693977, 'timing/model': 1.7949217818677425, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 0.17}) +[2026-06-15 20:01:15,978][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.29467761516571045, 'train/grad_norm_pre_clip': 11.105871200561523, 'timing/data': 0.013113943859934807, 'timing/model': 1.653962904587388, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 0.17}) +[2026-06-15 20:01:22,272][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.2772727608680725, 'train/grad_norm_pre_clip': 16.418975830078125, 'timing/data': 0.012924203649163246, 'timing/model': 1.5520895514637232, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 0.17}) +[2026-06-15 20:01:28,620][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.25873687863349915, 'train/grad_norm_pre_clip': 10.221277236938477, 'timing/data': 0.011665217578411102, 'timing/model': 1.5457244347780943, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.17}) +[2026-06-15 20:01:34,848][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.16421747207641602, 'train/grad_norm_pre_clip': 14.801393508911133, 'timing/data': 0.01319892331957817, 'timing/model': 1.5629139207303524, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 0.17}) +[2026-06-15 20:01:41,456][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.5187108516693115, 'train/grad_norm_pre_clip': 33.84763717651367, 'timing/data': 0.020713940262794495, 'timing/model': 1.6963095776736736, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 0.17}) +[2026-06-15 20:01:48,116][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.2700052857398987, 'train/grad_norm_pre_clip': 16.794363021850586, 'timing/data': 0.014656824991106987, 'timing/model': 1.792333658784628, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 0.17}) +[2026-06-15 20:01:54,147][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.3997565805912018, 'train/grad_norm_pre_clip': 16.244234085083008, 'timing/data': 0.012735921889543533, 'timing/model': 1.5520551074296236, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 0.17}) +[2026-06-15 20:02:00,344][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.24456143379211426, 'train/grad_norm_pre_clip': 8.052167892456055, 'timing/data': 0.01383676752448082, 'timing/model': 1.553936282172799, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 0.17}) +[2026-06-15 20:02:06,548][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.3015664517879486, 'train/grad_norm_pre_clip': 9.473652839660645, 'timing/data': 0.012677066028118134, 'timing/model': 1.5585353840142488, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 0.17}) +[2026-06-15 20:02:12,859][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.19882778823375702, 'train/grad_norm_pre_clip': 13.766258239746094, 'timing/data': 0.012291785329580307, 'timing/model': 1.587147792801261, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 0.17}) +[2026-06-15 20:02:19,189][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.3576948940753937, 'train/grad_norm_pre_clip': 8.839275360107422, 'timing/data': 0.013105569407343864, 'timing/model': 1.5913875307887793, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 0.17}) +[2026-06-15 20:02:25,510][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.20133629441261292, 'train/grad_norm_pre_clip': 4.638193130493164, 'timing/data': 0.01295478455722332, 'timing/model': 1.792747851461172, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 0.17}) +[2026-06-15 20:02:31,584][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.20882777869701385, 'train/grad_norm_pre_clip': 5.911658763885498, 'timing/data': 0.012464853003621101, 'timing/model': 1.567555958405137, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.17}) +[2026-06-15 20:02:38,081][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.1734243631362915, 'train/grad_norm_pre_clip': 6.169336795806885, 'timing/data': 0.011374322697520256, 'timing/model': 1.5638311170041561, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 0.17}) +[2026-06-15 20:02:44,384][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.3181552290916443, 'train/grad_norm_pre_clip': 10.182571411132812, 'timing/data': 0.012394534423947334, 'timing/model': 1.553219050168991, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 0.17}) +[2026-06-15 20:02:50,530][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.14899450540542603, 'train/grad_norm_pre_clip': 4.330554485321045, 'timing/data': 0.010724442079663277, 'timing/model': 1.467962559312582, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.17}) +[2026-06-15 20:02:56,519][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.20861127972602844, 'train/grad_norm_pre_clip': 7.17411994934082, 'timing/data': 0.011186383664608002, 'timing/model': 1.5245118252933025, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 0.17}) +[2026-06-15 20:03:02,408][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.41284364461898804, 'train/grad_norm_pre_clip': 16.429014205932617, 'timing/data': 0.00965350866317749, 'timing/model': 1.675975302234292, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 0.17}) +[2026-06-15 20:03:08,089][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.36874616146087646, 'train/grad_norm_pre_clip': 13.212273597717285, 'timing/data': 0.012054860591888428, 'timing/model': 1.4601807985454798, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 0.17}) +[2026-06-15 20:03:13,988][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.24709679186344147, 'train/grad_norm_pre_clip': 10.270960807800293, 'timing/data': 0.010997915640473366, 'timing/model': 1.4570751059800386, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 0.17}) +[2026-06-15 20:03:19,799][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.22406935691833496, 'train/grad_norm_pre_clip': 11.384379386901855, 'timing/data': 0.011198442429304123, 'timing/model': 1.4689223766326904, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 0.17}) +[2026-06-15 20:03:25,696][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.6328737735748291, 'train/grad_norm_pre_clip': 16.942642211914062, 'timing/data': 0.01036088541150093, 'timing/model': 1.462719963863492, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 0.17}) +[2026-06-15 20:03:31,685][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.3732127249240875, 'train/grad_norm_pre_clip': 18.412153244018555, 'timing/data': 0.011243315413594246, 'timing/model': 1.5051356423646212, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.17}) +[2026-06-15 20:03:37,594][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.5587481260299683, 'train/grad_norm_pre_clip': 21.597929000854492, 'timing/data': 0.012217944487929344, 'timing/model': 1.6589929200708866, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 0.17}) +[2026-06-15 20:03:43,288][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.23005184531211853, 'train/grad_norm_pre_clip': 8.626538276672363, 'timing/data': 0.009949000552296638, 'timing/model': 1.4540353398770094, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 0.17}) +[2026-06-15 20:03:49,189][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.20388923585414886, 'train/grad_norm_pre_clip': 9.617876052856445, 'timing/data': 0.011255316436290741, 'timing/model': 1.4518350418657064, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 0.18}) +[2026-06-15 20:03:55,089][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.22028809785842896, 'train/grad_norm_pre_clip': 13.180021286010742, 'timing/data': 0.011158330366015434, 'timing/model': 1.4561485964804888, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 0.18}) +[2026-06-15 20:04:00,986][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.06643452495336533, 'train/grad_norm_pre_clip': 3.7187862396240234, 'timing/data': 0.011385016143321991, 'timing/model': 1.4520292431116104, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 0.18}) +[2026-06-15 20:04:06,895][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.16784843802452087, 'train/grad_norm_pre_clip': 9.31714153289795, 'timing/data': 0.010737884789705276, 'timing/model': 1.4586274679750204, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 0.18}) +[2026-06-15 20:04:12,793][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.10591807961463928, 'train/grad_norm_pre_clip': 8.80597972869873, 'timing/data': 0.010342782363295555, 'timing/model': 1.6597126200795174, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 0.18}) +[2026-06-15 20:04:18,492][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.2319258749485016, 'train/grad_norm_pre_clip': 11.393730163574219, 'timing/data': 0.011178461834788322, 'timing/model': 1.4554403591901064, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 0.18}) +[2026-06-15 20:04:24,391][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.2477262318134308, 'train/grad_norm_pre_clip': 7.32093620300293, 'timing/data': 0.010968264192342758, 'timing/model': 1.4572047255933285, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 0.18}) +[2026-06-15 20:04:30,187][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.05854547768831253, 'train/grad_norm_pre_clip': 5.143606662750244, 'timing/data': 0.009949062019586563, 'timing/model': 1.4508179239928722, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.18}) +[2026-06-15 20:04:36,445][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.12585441768169403, 'train/grad_norm_pre_clip': 9.988216400146484, 'timing/data': 0.012036098167300224, 'timing/model': 1.553027719259262, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 0.18}) +[2026-06-15 20:04:42,645][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.07345907390117645, 'train/grad_norm_pre_clip': 17.43534278869629, 'timing/data': 0.01218581572175026, 'timing/model': 1.509904108941555, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 0.18}) +[2026-06-15 20:04:49,115][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.16770398616790771, 'train/grad_norm_pre_clip': 17.877670288085938, 'timing/data': 0.011501133441925049, 'timing/model': 2.041554855182767, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 0.18}) +[2026-06-15 20:04:55,345][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.43387570977211, 'train/grad_norm_pre_clip': 25.888216018676758, 'timing/data': 0.013667073100805283, 'timing/model': 1.750216729938984, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 0.18}) +[2026-06-15 20:05:01,843][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.41336533427238464, 'train/grad_norm_pre_clip': 16.76459503173828, 'timing/data': 0.012861944735050201, 'timing/model': 1.55678547360003, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 0.18}) +[2026-06-15 20:05:07,944][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.2149062156677246, 'train/grad_norm_pre_clip': 10.211172103881836, 'timing/data': 0.012150906026363373, 'timing/model': 1.5370168667286634, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 0.18}) +[2026-06-15 20:05:14,328][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.43646323680877686, 'train/grad_norm_pre_clip': 16.12760353088379, 'timing/data': 0.015098238363862038, 'timing/model': 1.687662836164236, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 0.18}) +[2026-06-15 20:05:21,239][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.13831453025341034, 'train/grad_norm_pre_clip': 13.27038288116455, 'timing/data': 0.017930418252944946, 'timing/model': 1.5917937997728586, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 0.18}) +[2026-06-15 20:05:28,264][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.27184510231018066, 'train/grad_norm_pre_clip': 13.853178024291992, 'timing/data': 0.018398525193333626, 'timing/model': 2.068765301257372, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 0.18}) +[2026-06-15 20:05:34,930][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.1912303864955902, 'train/grad_norm_pre_clip': 7.200125217437744, 'timing/data': 0.019151508808135986, 'timing/model': 1.7382976729422808, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.18}) +[2026-06-15 20:05:41,877][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.4554362893104553, 'train/grad_norm_pre_clip': 16.59208106994629, 'timing/data': 0.012530729174613953, 'timing/model': 1.5720463320612907, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 0.18}) +[2026-06-15 20:05:48,716][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.3591777980327606, 'train/grad_norm_pre_clip': 14.083049774169922, 'timing/data': 0.0193195603787899, 'timing/model': 1.7253712974488735, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 0.18}) +[2026-06-15 20:05:55,729][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.0673450380563736, 'train/grad_norm_pre_clip': 3.536320209503174, 'timing/data': 0.01771821826696396, 'timing/model': 1.6811560317873955, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 0.18}) +[2026-06-15 20:06:02,217][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.2939012050628662, 'train/grad_norm_pre_clip': 9.52133560180664, 'timing/data': 0.01312229409813881, 'timing/model': 1.5420215595513582, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 0.18}) +[2026-06-15 20:06:08,477][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.23759402334690094, 'train/grad_norm_pre_clip': 14.843788146972656, 'timing/data': 0.012990331277251244, 'timing/model': 1.7931058127433062, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 0.18}) +[2026-06-15 20:06:14,993][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.14641878008842468, 'train/grad_norm_pre_clip': 8.720770835876465, 'timing/data': 0.01750730350613594, 'timing/model': 1.6320098787546158, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 0.18}) +[2026-06-15 20:06:21,272][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.19831903278827667, 'train/grad_norm_pre_clip': 5.430628776550293, 'timing/data': 0.013060266152024269, 'timing/model': 1.5486053172498941, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 0.18}) +[2026-06-15 20:06:27,775][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.2929600179195404, 'train/grad_norm_pre_clip': 15.678874015808105, 'timing/data': 0.012595536187291145, 'timing/model': 1.5554614067077637, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 0.18}) +[2026-06-15 20:06:34,300][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.4304783046245575, 'train/grad_norm_pre_clip': 12.584604263305664, 'timing/data': 0.013077760115265846, 'timing/model': 1.5885811150074005, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 0.19}) +[2026-06-15 20:06:40,778][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.20226654410362244, 'train/grad_norm_pre_clip': 6.619656562805176, 'timing/data': 0.012632757425308228, 'timing/model': 1.5246378779411316, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.19}) +[2026-06-15 20:06:47,268][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.16868053376674652, 'train/grad_norm_pre_clip': 9.041001319885254, 'timing/data': 0.01271551288664341, 'timing/model': 1.830533878877759, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 0.19}) +[2026-06-15 20:06:53,442][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.21923670172691345, 'train/grad_norm_pre_clip': 5.862260818481445, 'timing/data': 0.012516090646386147, 'timing/model': 1.6308494936674833, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 0.19}) +[2026-06-15 20:06:59,636][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.28410661220550537, 'train/grad_norm_pre_clip': 8.035374641418457, 'timing/data': 0.013056429103016853, 'timing/model': 1.5558203924447298, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 0.19}) +[2026-06-15 20:07:05,841][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.3149661719799042, 'train/grad_norm_pre_clip': 6.518610000610352, 'timing/data': 0.012500882148742676, 'timing/model': 1.552302323281765, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 0.19}) +[2026-06-15 20:07:12,179][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.4250418543815613, 'train/grad_norm_pre_clip': 15.034090042114258, 'timing/data': 0.013380393385887146, 'timing/model': 1.5433514136821032, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.19}) +[2026-06-15 20:07:18,536][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.16487443447113037, 'train/grad_norm_pre_clip': 4.66918420791626, 'timing/data': 0.012908482924103737, 'timing/model': 1.8179464563727379, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 0.19}) +[2026-06-15 20:07:24,882][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.19029177725315094, 'train/grad_norm_pre_clip': 5.154379367828369, 'timing/data': 0.012108063325285912, 'timing/model': 1.5460287686437368, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 0.19}) +[2026-06-15 20:07:31,173][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.24780264496803284, 'train/grad_norm_pre_clip': 6.471077919006348, 'timing/data': 0.013076422736048698, 'timing/model': 1.540793463587761, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 0.19}) +[2026-06-15 20:07:37,375][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.20483261346817017, 'train/grad_norm_pre_clip': 4.195549964904785, 'timing/data': 0.011981332674622536, 'timing/model': 1.552844000980258, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 0.19}) +[2026-06-15 20:07:43,875][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.4933026134967804, 'train/grad_norm_pre_clip': 11.428943634033203, 'timing/data': 0.012941556051373482, 'timing/model': 1.7501970902085304, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 0.19}) +[2026-06-15 20:07:50,290][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.1612084060907364, 'train/grad_norm_pre_clip': 5.737983226776123, 'timing/data': 0.013813143596053123, 'timing/model': 1.5796653423458338, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 0.19}) +[2026-06-15 20:07:56,904][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.15243858098983765, 'train/grad_norm_pre_clip': 8.234745979309082, 'timing/data': 0.013167249038815498, 'timing/model': 1.782493844628334, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.19}) +[2026-06-15 20:08:03,073][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.08703313767910004, 'train/grad_norm_pre_clip': 3.8909928798675537, 'timing/data': 0.013641998171806335, 'timing/model': 1.5422986261546612, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 0.19}) +[2026-06-15 20:08:09,493][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.42942070960998535, 'train/grad_norm_pre_clip': 30.40403175354004, 'timing/data': 0.014715198427438736, 'timing/model': 1.5551966074854136, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 0.19}) +[2026-06-15 20:08:15,987][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.18877437710762024, 'train/grad_norm_pre_clip': 15.499567031860352, 'timing/data': 0.013596244156360626, 'timing/model': 1.548301111906767, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 0.19}) +[2026-06-15 20:08:22,367][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.12247513234615326, 'train/grad_norm_pre_clip': 10.255915641784668, 'timing/data': 0.01533549651503563, 'timing/model': 1.5516234040260315, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 0.19}) +[2026-06-15 20:08:28,691][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.23913021385669708, 'train/grad_norm_pre_clip': 17.043020248413086, 'timing/data': 0.014142943546175957, 'timing/model': 1.5421556700021029, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 0.19}) +[2026-06-15 20:08:35,063][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.05051795020699501, 'train/grad_norm_pre_clip': 5.488379955291748, 'timing/data': 0.01284881867468357, 'timing/model': 1.8333456106483936, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 0.19}) +[2026-06-15 20:08:41,399][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.5388379096984863, 'train/grad_norm_pre_clip': 19.602767944335938, 'timing/data': 0.012748092412948608, 'timing/model': 1.5698030441999435, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 0.19}) +[2026-06-15 20:16:31,389][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.8494249582290649, 'train/grad_norm_pre_clip': 21.20408821105957, 'eval/loss': 0.27182701230049133, 'eval/flap_precision': 0.7110894941634242, 'eval/flap_recall': 0.9918588873812755, 'eval/flap_f1': 0.8283286118980171, 'eval/flap_rate_model': 0.28555555555555556, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8283286118980172, 'timing/data': 0.012435643002390862, 'timing/model': 1.5461838580667973, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 0.19}) +[2026-06-15 20:16:31,390][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_500_state +[2026-06-15 20:16:31,394][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:16:38,505][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_500_state/model.safetensors +[2026-06-15 20:16:56,460][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_500_state/optimizer.bin +[2026-06-15 20:16:56,461][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_500_state/sampler.bin +[2026-06-15 20:16:56,461][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_500_state/sampler_1.bin +[2026-06-15 20:16:56,462][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-15 20:16:56,563][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-15 20:16:56,563][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:16:56,572][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:17:02,302][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.963638961315155, 'train/grad_norm_pre_clip': 40.794490814208984, 'timing/data': 0.011070050299167633, 'timing/model': 1.4645006824284792, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 0.19}) +[2026-06-15 20:17:08,204][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.457842081785202, 'train/grad_norm_pre_clip': 20.888072967529297, 'timing/data': 0.010850800201296806, 'timing/model': 1.458059599623084, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 0.19}) +[2026-06-15 20:17:14,207][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.39227065443992615, 'train/grad_norm_pre_clip': 13.425825119018555, 'timing/data': 0.010245239362120628, 'timing/model': 1.4657001681625843, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 0.19}) +[2026-06-15 20:17:20,219][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.21259531378746033, 'train/grad_norm_pre_clip': 9.611626625061035, 'timing/data': 0.010321687906980515, 'timing/model': 1.474715756252408, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 0.2}) +[2026-06-15 20:17:26,128][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.4886851906776428, 'train/grad_norm_pre_clip': 16.681804656982422, 'timing/data': 0.010218864306807518, 'timing/model': 1.679947316646576, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 0.2}) +[2026-06-15 20:17:32,049][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.15590012073516846, 'train/grad_norm_pre_clip': 11.06656551361084, 'timing/data': 0.011267812922596931, 'timing/model': 1.7235837392508984, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 0.2}) +[2026-06-15 20:17:38,029][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.43473607301712036, 'train/grad_norm_pre_clip': 18.903926849365234, 'timing/data': 0.013300709426403046, 'timing/model': 1.5480012316256762, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 0.2}) +[2026-06-15 20:17:44,127][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.09486722946166992, 'train/grad_norm_pre_clip': 4.3354172706604, 'timing/data': 0.013141617178916931, 'timing/model': 1.547806367278099, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 0.2}) +[2026-06-15 20:17:50,345][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.15609213709831238, 'train/grad_norm_pre_clip': 9.384635925292969, 'timing/data': 0.012476984411478043, 'timing/model': 1.5385737195611, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 0.2}) +[2026-06-15 20:17:56,544][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.1484411209821701, 'train/grad_norm_pre_clip': 5.065042972564697, 'timing/data': 0.013254843652248383, 'timing/model': 1.523429524153471, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 0.2}) +[2026-06-15 20:18:03,175][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.19823046028614044, 'train/grad_norm_pre_clip': 6.949150085449219, 'timing/data': 0.01397986151278019, 'timing/model': 1.8843977600336075, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 0.2}) +[2026-06-15 20:18:09,233][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.245345801115036, 'train/grad_norm_pre_clip': 4.0547027587890625, 'timing/data': 0.012684887275099754, 'timing/model': 1.5554739907383919, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 0.2}) +[2026-06-15 20:18:15,583][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.24507424235343933, 'train/grad_norm_pre_clip': 6.26650333404541, 'timing/data': 0.013461995869874954, 'timing/model': 1.6155102290213108, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 0.2}) +[2026-06-15 20:18:21,988][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.22411997616291046, 'train/grad_norm_pre_clip': 5.294307231903076, 'timing/data': 0.013109274208545685, 'timing/model': 1.5451660249382257, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 0.2}) +[2026-06-15 20:18:28,375][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.32516932487487793, 'train/grad_norm_pre_clip': 10.676804542541504, 'timing/data': 0.012647651135921478, 'timing/model': 1.520151423290372, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 0.2}) +[2026-06-15 20:18:34,791][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.1748138964176178, 'train/grad_norm_pre_clip': 3.8605918884277344, 'timing/data': 0.013238081708550453, 'timing/model': 1.5879761464893818, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 0.2}) +[2026-06-15 20:18:41,137][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.21892595291137695, 'train/grad_norm_pre_clip': 4.886127948760986, 'timing/data': 0.012236429378390312, 'timing/model': 1.7993323374539614, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 0.2}) +[2026-06-15 20:18:47,176][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.22932910919189453, 'train/grad_norm_pre_clip': 4.1387248039245605, 'timing/data': 0.01157490722835064, 'timing/model': 1.5534809418022633, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 0.2}) +[2026-06-15 20:18:53,476][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.23514527082443237, 'train/grad_norm_pre_clip': 7.016620635986328, 'timing/data': 0.01316930167376995, 'timing/model': 1.5532641410827637, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 0.2}) +[2026-06-15 20:18:59,778][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.36891549825668335, 'train/grad_norm_pre_clip': 14.367609977722168, 'timing/data': 0.012037184089422226, 'timing/model': 1.5568155366927385, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.2}) +[2026-06-15 20:19:06,264][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.3545210063457489, 'train/grad_norm_pre_clip': 9.592818260192871, 'timing/data': 0.01156482845544815, 'timing/model': 1.6189603470265865, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 0.2}) +[2026-06-15 20:19:12,671][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.12441522628068924, 'train/grad_norm_pre_clip': 3.137190818786621, 'timing/data': 0.012505004182457924, 'timing/model': 1.576663812622428, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 0.2}) +[2026-06-15 20:19:19,007][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.18372192978858948, 'train/grad_norm_pre_clip': 6.579905033111572, 'timing/data': 0.012867746874690056, 'timing/model': 1.763588022440672, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 0.2}) +[2026-06-15 20:19:25,301][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.17537185549736023, 'train/grad_norm_pre_clip': 7.486320495605469, 'timing/data': 0.011717790737748146, 'timing/model': 1.7593827210366726, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 0.2}) +[2026-06-15 20:19:31,382][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.2744666635990143, 'train/grad_norm_pre_clip': 10.531204223632812, 'timing/data': 0.012310093268752098, 'timing/model': 1.5453509837388992, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 0.2}) +[2026-06-15 20:19:37,919][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.25074970722198486, 'train/grad_norm_pre_clip': 11.297200202941895, 'timing/data': 0.013764815405011177, 'timing/model': 1.6014017574489117, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 0.2}) +[2026-06-15 20:19:44,787][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.27365362644195557, 'train/grad_norm_pre_clip': 11.543874740600586, 'timing/data': 0.012577574700117111, 'timing/model': 1.553792605176568, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 0.2}) +[2026-06-15 20:19:51,284][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.0829799547791481, 'train/grad_norm_pre_clip': 5.107192516326904, 'timing/data': 0.014733841642737389, 'timing/model': 1.5580403581261635, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 0.2}) +[2026-06-15 20:19:57,631][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.26830148696899414, 'train/grad_norm_pre_clip': 12.390193939208984, 'timing/data': 0.013132985681295395, 'timing/model': 1.7797423731535673, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 0.2}) +[2026-06-15 20:20:03,774][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.12733867764472961, 'train/grad_norm_pre_clip': 6.135601997375488, 'timing/data': 0.013964841142296791, 'timing/model': 1.595602435991168, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 0.21}) +[2026-06-15 20:20:10,082][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.35613301396369934, 'train/grad_norm_pre_clip': 15.203215599060059, 'timing/data': 0.012753227725625038, 'timing/model': 1.5618246532976627, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 0.21}) +[2026-06-15 20:20:16,396][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.44973069429397583, 'train/grad_norm_pre_clip': 15.759709358215332, 'timing/data': 0.01499524712562561, 'timing/model': 1.551560863852501, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 0.21}) +[2026-06-15 20:20:22,797][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.2092779278755188, 'train/grad_norm_pre_clip': 9.569751739501953, 'timing/data': 0.017405686900019646, 'timing/model': 1.5738239623606205, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 0.21}) +[2026-06-15 20:20:29,075][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.20218683779239655, 'train/grad_norm_pre_clip': 7.863595008850098, 'timing/data': 0.013355787843465805, 'timing/model': 1.5742962677031755, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 0.21}) +[2026-06-15 20:20:35,560][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.31501540541648865, 'train/grad_norm_pre_clip': 18.19964027404785, 'timing/data': 0.0122881680727005, 'timing/model': 1.8237298633903265, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 0.21}) +[2026-06-15 20:20:41,670][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.23768387734889984, 'train/grad_norm_pre_clip': 15.644794464111328, 'timing/data': 0.011604076251387596, 'timing/model': 1.5484716892242432, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 0.21}) +[2026-06-15 20:20:48,067][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.37732601165771484, 'train/grad_norm_pre_clip': 12.020499229431152, 'timing/data': 0.01394629292190075, 'timing/model': 1.550872404128313, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 0.21}) +[2026-06-15 20:20:54,384][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.12506423890590668, 'train/grad_norm_pre_clip': 10.770156860351562, 'timing/data': 0.011777425184845924, 'timing/model': 1.5376715864986181, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 0.21}) +[2026-06-15 20:21:00,788][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.25755441188812256, 'train/grad_norm_pre_clip': 7.271237373352051, 'timing/data': 0.012073665857315063, 'timing/model': 1.5969862956553698, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 0.21}) +[2026-06-15 20:21:07,372][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.25377801060676575, 'train/grad_norm_pre_clip': 6.174048900604248, 'timing/data': 0.012277938425540924, 'timing/model': 1.552789170295, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 0.21}) +[2026-06-15 20:21:13,686][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.1507553607225418, 'train/grad_norm_pre_clip': 7.234947681427002, 'timing/data': 0.01353432983160019, 'timing/model': 1.7613606359809637, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.21}) +[2026-06-15 20:21:19,747][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.20211991667747498, 'train/grad_norm_pre_clip': 12.407620429992676, 'timing/data': 0.013725720345973969, 'timing/model': 1.5539948921650648, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 0.21}) +[2026-06-15 20:21:26,242][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.3967885971069336, 'train/grad_norm_pre_clip': 18.14117431640625, 'timing/data': 0.013512244448065758, 'timing/model': 1.552681552246213, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 0.21}) +[2026-06-15 20:21:32,495][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.3729532063007355, 'train/grad_norm_pre_clip': 25.22492790222168, 'timing/data': 0.012873632833361626, 'timing/model': 1.5948169864714146, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 0.21}) +[2026-06-15 20:21:39,398][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.20146359503269196, 'train/grad_norm_pre_clip': 7.132994651794434, 'timing/data': 0.012610780075192451, 'timing/model': 1.5825591199100018, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 0.21}) +[2026-06-15 20:21:45,829][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.3405066430568695, 'train/grad_norm_pre_clip': 9.760577201843262, 'timing/data': 0.012164764106273651, 'timing/model': 1.8156605288386345, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 0.21}) +[2026-06-15 20:21:51,892][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.3017004132270813, 'train/grad_norm_pre_clip': 5.4044060707092285, 'timing/data': 0.011696139350533485, 'timing/model': 1.5794712994247675, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 0.21}) +[2026-06-15 20:21:58,280][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.2792406678199768, 'train/grad_norm_pre_clip': 10.240063667297363, 'timing/data': 0.01357927918434143, 'timing/model': 1.5590745117515326, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 0.21}) +[2026-06-15 20:22:04,586][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.21445274353027344, 'train/grad_norm_pre_clip': 7.071736812591553, 'timing/data': 0.014045912772417068, 'timing/model': 1.563402807340026, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 0.21}) +[2026-06-15 20:22:10,952][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.037888117134571075, 'train/grad_norm_pre_clip': 1.6505931615829468, 'timing/data': 0.014301275834441185, 'timing/model': 1.6095543317496777, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 0.21}) +[2026-06-15 20:22:17,261][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.11534377932548523, 'train/grad_norm_pre_clip': 5.769264221191406, 'timing/data': 0.012720208615064621, 'timing/model': 1.5149496085941792, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 0.21}) +[2026-06-15 20:22:23,607][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.11128538846969604, 'train/grad_norm_pre_clip': 7.39889669418335, 'timing/data': 0.012870574370026588, 'timing/model': 1.8053712528198957, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 0.21}) +[2026-06-15 20:22:29,736][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.20593956112861633, 'train/grad_norm_pre_clip': 6.967522144317627, 'timing/data': 0.011449983343482018, 'timing/model': 1.5673361420631409, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 0.21}) +[2026-06-15 20:22:35,869][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.3482228219509125, 'train/grad_norm_pre_clip': 22.308759689331055, 'timing/data': 0.011299898847937584, 'timing/model': 1.4968153052031994, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 0.21}) +[2026-06-15 20:22:42,146][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.3042166829109192, 'train/grad_norm_pre_clip': 25.407978057861328, 'timing/data': 0.01408480480313301, 'timing/model': 1.5609958972781897, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.21}) +[2026-06-15 20:22:48,249][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.04726213961839676, 'train/grad_norm_pre_clip': 3.2624752521514893, 'timing/data': 0.012391695752739906, 'timing/model': 1.4910659231245518, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 0.22}) +[2026-06-15 20:22:54,351][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.1591414511203766, 'train/grad_norm_pre_clip': 9.591102600097656, 'timing/data': 0.011547747999429703, 'timing/model': 1.524807807058096, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 0.22}) +[2026-06-15 20:23:00,630][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.16496269404888153, 'train/grad_norm_pre_clip': 9.238970756530762, 'timing/data': 0.012288602069020271, 'timing/model': 1.7550089862197638, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 0.22}) +[2026-06-15 20:23:06,837][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.5691257119178772, 'train/grad_norm_pre_clip': 68.70027160644531, 'timing/data': 0.013329384848475456, 'timing/model': 1.7554894592612982, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 0.22}) +[2026-06-15 20:23:12,963][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.11488719284534454, 'train/grad_norm_pre_clip': 9.852398872375488, 'timing/data': 0.01178492046892643, 'timing/model': 1.4877770487219095, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 0.22}) +[2026-06-15 20:23:19,252][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.21873289346694946, 'train/grad_norm_pre_clip': 13.887833595275879, 'timing/data': 0.012370672076940536, 'timing/model': 1.5754474308341742, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 0.22}) +[2026-06-15 20:23:25,535][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.1667482703924179, 'train/grad_norm_pre_clip': 10.846612930297852, 'timing/data': 0.012288656085729599, 'timing/model': 1.5591472368687391, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 0.22}) +[2026-06-15 20:23:31,710][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.1912870705127716, 'train/grad_norm_pre_clip': 12.302119255065918, 'timing/data': 0.012431761249899864, 'timing/model': 1.5362350270152092, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 0.22}) +[2026-06-15 20:23:38,011][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.262735515832901, 'train/grad_norm_pre_clip': 12.475528717041016, 'timing/data': 0.012070320546627045, 'timing/model': 1.805167742073536, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 0.22}) +[2026-06-15 20:23:44,080][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.11684904992580414, 'train/grad_norm_pre_clip': 8.690176010131836, 'timing/data': 0.012315493077039719, 'timing/model': 1.5656678918749094, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 0.22}) +[2026-06-15 20:23:50,441][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.09007924795150757, 'train/grad_norm_pre_clip': 8.420599937438965, 'timing/data': 0.0121281947940588, 'timing/model': 1.5580830425024033, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 0.22}) +[2026-06-15 20:23:56,672][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.04818178713321686, 'train/grad_norm_pre_clip': 4.194741725921631, 'timing/data': 0.012267053127288818, 'timing/model': 1.5896644908934832, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 0.22}) +[2026-06-15 20:24:02,979][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.3554263710975647, 'train/grad_norm_pre_clip': 42.43460464477539, 'timing/data': 0.012855585664510727, 'timing/model': 1.5639121290296316, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 0.22}) +[2026-06-15 20:24:09,476][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.13711372017860413, 'train/grad_norm_pre_clip': 9.981755256652832, 'timing/data': 0.01202489621937275, 'timing/model': 1.5650916639715433, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 0.22}) +[2026-06-15 20:24:15,813][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.6934071779251099, 'train/grad_norm_pre_clip': 38.714263916015625, 'timing/data': 0.011828005313873291, 'timing/model': 1.797053400427103, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 0.22}) +[2026-06-15 20:24:21,872][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.17652367055416107, 'train/grad_norm_pre_clip': 5.340568542480469, 'timing/data': 0.012304538860917091, 'timing/model': 1.5592854209244251, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 0.22}) +[2026-06-15 20:24:28,172][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.5635913014411926, 'train/grad_norm_pre_clip': 27.939504623413086, 'timing/data': 0.01267269067466259, 'timing/model': 1.5563327837735415, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 0.22}) +[2026-06-15 20:24:34,470][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.09180697798728943, 'train/grad_norm_pre_clip': 4.7524213790893555, 'timing/data': 0.012475378811359406, 'timing/model': 1.5555924121290445, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 0.22}) +[2026-06-15 20:24:40,769][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.28193527460098267, 'train/grad_norm_pre_clip': 14.192728042602539, 'timing/data': 0.012576274573802948, 'timing/model': 1.55685412324965, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 0.22}) +[2026-06-15 20:24:47,070][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.16869540512561798, 'train/grad_norm_pre_clip': 10.852692604064941, 'timing/data': 0.01227322407066822, 'timing/model': 1.514120688661933, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 0.22}) +[2026-06-15 20:24:53,454][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.13545183837413788, 'train/grad_norm_pre_clip': 6.270475387573242, 'timing/data': 0.011773521080613136, 'timing/model': 1.8416377790272236, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 0.22}) +[2026-06-15 20:24:59,568][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.14253541827201843, 'train/grad_norm_pre_clip': 12.066015243530273, 'timing/data': 0.012648520991206169, 'timing/model': 1.5508503913879395, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 0.22}) +[2026-06-15 20:25:06,372][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.27888914942741394, 'train/grad_norm_pre_clip': 8.625266075134277, 'timing/data': 0.012309381738305092, 'timing/model': 1.548773132264614, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 0.22}) +[2026-06-15 20:25:12,670][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.17381376028060913, 'train/grad_norm_pre_clip': 9.05599308013916, 'timing/data': 0.012532608583569527, 'timing/model': 1.5485060643404722, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 0.22}) +[2026-06-15 20:25:18,968][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.37326961755752563, 'train/grad_norm_pre_clip': 16.495954513549805, 'timing/data': 0.01182144321501255, 'timing/model': 1.5556168369948864, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 0.22}) +[2026-06-15 20:25:25,272][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.19798767566680908, 'train/grad_norm_pre_clip': 5.06011962890625, 'timing/data': 0.012074682861566544, 'timing/model': 1.5150108486413956, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 0.22}) +[2026-06-15 20:25:31,608][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.13997013866901398, 'train/grad_norm_pre_clip': 3.9450314044952393, 'timing/data': 0.012450916692614555, 'timing/model': 1.7908519674092531, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 0.23}) +[2026-06-15 20:25:37,669][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.39338424801826477, 'train/grad_norm_pre_clip': 14.164045333862305, 'timing/data': 0.01174953207373619, 'timing/model': 1.5538689326494932, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 0.23}) +[2026-06-15 20:25:43,838][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.1976461410522461, 'train/grad_norm_pre_clip': 5.293525695800781, 'timing/data': 0.011964647099375725, 'timing/model': 1.4823055230081081, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 0.23}) +[2026-06-15 20:25:49,995][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.09906276315450668, 'train/grad_norm_pre_clip': 3.810114622116089, 'timing/data': 0.011160558089613914, 'timing/model': 1.529436782002449, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 0.23}) +[2026-06-15 20:25:56,257][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.11285868287086487, 'train/grad_norm_pre_clip': 5.168182373046875, 'timing/data': 0.013088038191199303, 'timing/model': 1.5670798886567354, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 0.23}) +[2026-06-15 20:26:02,440][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.1560133844614029, 'train/grad_norm_pre_clip': 4.732734203338623, 'timing/data': 0.012821432203054428, 'timing/model': 1.4783460311591625, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 0.23}) +[2026-06-15 20:26:08,693][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.14213311672210693, 'train/grad_norm_pre_clip': 5.117446422576904, 'timing/data': 0.012414524331688881, 'timing/model': 1.7928410414606333, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 0.23}) +[2026-06-15 20:26:14,835][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.5262433886528015, 'train/grad_norm_pre_clip': 11.509164810180664, 'timing/data': 0.01284470409154892, 'timing/model': 1.7058878391981125, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 0.23}) +[2026-06-15 20:26:20,839][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.20410649478435516, 'train/grad_norm_pre_clip': 5.2400288581848145, 'timing/data': 0.01256791315972805, 'timing/model': 1.5559576116502285, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.23}) +[2026-06-15 20:26:26,936][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.4320541322231293, 'train/grad_norm_pre_clip': 18.09408950805664, 'timing/data': 0.011526022106409073, 'timing/model': 1.5447972510010004, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 0.23}) +[2026-06-15 20:26:33,141][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.16765861213207245, 'train/grad_norm_pre_clip': 11.79020881652832, 'timing/data': 0.012732824310660362, 'timing/model': 1.553745275363326, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 0.23}) +[2026-06-15 20:26:39,338][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.01815148815512657, 'train/grad_norm_pre_clip': 1.3037395477294922, 'timing/data': 0.012365087866783142, 'timing/model': 1.521132841706276, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 0.23}) +[2026-06-15 20:26:45,639][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.40188145637512207, 'train/grad_norm_pre_clip': 18.7437686920166, 'timing/data': 0.01070399209856987, 'timing/model': 1.8200048804283142, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 0.23}) +[2026-06-15 20:26:51,670][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.23169264197349548, 'train/grad_norm_pre_clip': 6.792322158813477, 'timing/data': 0.011639116331934929, 'timing/model': 1.5535717345774174, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 0.23}) +[2026-06-15 20:26:58,074][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.20956408977508545, 'train/grad_norm_pre_clip': 13.559480667114258, 'timing/data': 0.012993639335036278, 'timing/model': 1.5609547272324562, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 0.23}) +[2026-06-15 20:27:04,370][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.19618955254554749, 'train/grad_norm_pre_clip': 9.77447509765625, 'timing/data': 0.012614032253623009, 'timing/model': 1.5558517631143332, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 0.23}) +[2026-06-15 20:27:10,679][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.08870856463909149, 'train/grad_norm_pre_clip': 6.713503360748291, 'timing/data': 0.01191776618361473, 'timing/model': 1.5147483199834824, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 0.23}) +[2026-06-15 20:27:16,970][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.08478005975484848, 'train/grad_norm_pre_clip': 7.953597545623779, 'timing/data': 0.012644520029425621, 'timing/model': 1.536027079448104, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 0.23}) +[2026-06-15 20:34:56,575][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.26536399126052856, 'train/grad_norm_pre_clip': 5.401891231536865, 'eval/loss': 0.11320653557777405, 'eval/flap_precision': 0.8963068181818182, 'eval/flap_recall': 0.8561736770691994, 'eval/flap_f1': 0.8757807078417765, 'eval/flap_rate_model': 0.19555555555555557, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8757807078417765, 'timing/data': 0.01303640753030777, 'timing/model': 1.7346723824739456, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 0.23}) +[2026-06-15 20:34:56,575][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_600_state +[2026-06-15 20:34:56,594][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:35:02,537][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_600_state/model.safetensors +[2026-06-15 20:35:16,882][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_600_state/optimizer.bin +[2026-06-15 20:35:16,883][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_600_state/sampler.bin +[2026-06-15 20:35:16,883][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_600_state/sampler_1.bin +[2026-06-15 20:35:16,884][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_600_state/random_states_0.pkl +[2026-06-15 20:35:16,970][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600 +[2026-06-15 20:35:16,970][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:35:16,979][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:35:23,204][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.1316639482975006, 'train/grad_norm_pre_clip': 5.363007545471191, 'timing/data': 0.013957051560282707, 'timing/model': 1.8365270961076021, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 0.23}) +[2026-06-15 20:35:29,264][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.19720256328582764, 'train/grad_norm_pre_clip': 10.653711318969727, 'timing/data': 0.01597268506884575, 'timing/model': 1.582831623032689, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 0.23}) +[2026-06-15 20:35:35,533][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.1977311372756958, 'train/grad_norm_pre_clip': 10.709331512451172, 'timing/data': 0.015116993337869644, 'timing/model': 1.554689571261406, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 0.23}) +[2026-06-15 20:35:42,180][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.2536763846874237, 'train/grad_norm_pre_clip': 7.679709434509277, 'timing/data': 0.019114535301923752, 'timing/model': 1.7169922851026058, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 0.23}) +[2026-06-15 20:35:48,520][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.1462440937757492, 'train/grad_norm_pre_clip': 9.399834632873535, 'timing/data': 0.0163101889193058, 'timing/model': 1.6285802721977234, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 0.23}) +[2026-06-15 20:35:54,665][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.14141210913658142, 'train/grad_norm_pre_clip': 11.264564514160156, 'timing/data': 0.01605873927474022, 'timing/model': 1.5279118176549673, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 0.23}) +[2026-06-15 20:36:00,820][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.08580493181943893, 'train/grad_norm_pre_clip': 8.511969566345215, 'timing/data': 0.01343272253870964, 'timing/model': 1.7429003659635782, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 0.23}) +[2026-06-15 20:36:06,759][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.3530368506908417, 'train/grad_norm_pre_clip': 12.085783004760742, 'timing/data': 0.01702500879764557, 'timing/model': 1.6685800962150097, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 0.24}) +[2026-06-15 20:36:12,776][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.19558276236057281, 'train/grad_norm_pre_clip': 13.25314998626709, 'timing/data': 0.009568672627210617, 'timing/model': 1.4576845616102219, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 0.24}) +[2026-06-15 20:36:18,680][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.16728843748569489, 'train/grad_norm_pre_clip': 10.448138236999512, 'timing/data': 0.010948112234473228, 'timing/model': 1.444932982325554, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 0.24}) +[2026-06-15 20:36:24,575][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.3776665925979614, 'train/grad_norm_pre_clip': 17.204057693481445, 'timing/data': 0.01054903119802475, 'timing/model': 1.4504483435302973, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 0.24}) +[2026-06-15 20:36:30,485][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.03789334371685982, 'train/grad_norm_pre_clip': 1.7012932300567627, 'timing/data': 0.010920576751232147, 'timing/model': 1.498593831434846, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 0.24}) +[2026-06-15 20:36:36,400][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.26542577147483826, 'train/grad_norm_pre_clip': 9.090563774108887, 'timing/data': 0.011172221973538399, 'timing/model': 1.681908218190074, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 0.24}) +[2026-06-15 20:36:42,082][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.2779281735420227, 'train/grad_norm_pre_clip': 13.768891334533691, 'timing/data': 0.010987590998411179, 'timing/model': 1.4506068099290133, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 0.24}) +[2026-06-15 20:36:47,986][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.24363015592098236, 'train/grad_norm_pre_clip': 17.465587615966797, 'timing/data': 0.011218074709177017, 'timing/model': 1.467891002073884, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 0.24}) +[2026-06-15 20:36:53,887][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.19570764899253845, 'train/grad_norm_pre_clip': 4.8649067878723145, 'timing/data': 0.010006438940763474, 'timing/model': 1.4680627193301916, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 0.24}) +[2026-06-15 20:36:59,785][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.07705020904541016, 'train/grad_norm_pre_clip': 7.789323329925537, 'timing/data': 0.009815337136387825, 'timing/model': 1.4655900672078133, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 0.24}) +[2026-06-15 20:37:05,668][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.20210185647010803, 'train/grad_norm_pre_clip': 9.022378921508789, 'timing/data': 0.010047899559140205, 'timing/model': 1.496814277023077, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 0.24}) +[2026-06-15 20:37:11,616][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.25868815183639526, 'train/grad_norm_pre_clip': 12.23416519165039, 'timing/data': 0.01093083806335926, 'timing/model': 1.6965954657644033, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 0.24}) +[2026-06-15 20:37:17,669][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.15499182045459747, 'train/grad_norm_pre_clip': 4.6111979484558105, 'timing/data': 0.016209464520215988, 'timing/model': 1.5907707009464502, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 0.24}) +[2026-06-15 20:37:23,835][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.16588541865348816, 'train/grad_norm_pre_clip': 7.658841133117676, 'timing/data': 0.011175161227583885, 'timing/model': 1.529123367741704, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 0.24}) +[2026-06-15 20:37:30,068][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.16678181290626526, 'train/grad_norm_pre_clip': 9.493738174438477, 'timing/data': 0.014847584068775177, 'timing/model': 1.5628505609929562, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 0.24}) +[2026-06-15 20:37:36,339][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.07190603762865067, 'train/grad_norm_pre_clip': 4.760917663574219, 'timing/data': 0.014427674934267998, 'timing/model': 1.4996349811553955, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 0.24}) +[2026-06-15 20:37:42,480][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.19154119491577148, 'train/grad_norm_pre_clip': 9.374805450439453, 'timing/data': 0.00988335907459259, 'timing/model': 1.4325657691806555, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 0.24}) +[2026-06-15 20:37:48,678][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.23682694137096405, 'train/grad_norm_pre_clip': 15.298025131225586, 'timing/data': 0.012840557843446732, 'timing/model': 1.749802179634571, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.24}) +[2026-06-15 20:37:54,635][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.2280479222536087, 'train/grad_norm_pre_clip': 8.03620719909668, 'timing/data': 0.011247511953115463, 'timing/model': 1.484085911884904, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 0.24}) +[2026-06-15 20:38:00,733][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.05144215747714043, 'train/grad_norm_pre_clip': 4.370303153991699, 'timing/data': 0.011545265093445778, 'timing/model': 1.5392110627144575, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 0.24}) +[2026-06-15 20:38:06,926][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.10238127410411835, 'train/grad_norm_pre_clip': 6.914186000823975, 'timing/data': 0.012662878260016441, 'timing/model': 1.5484641883522272, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 0.24}) +[2026-06-15 20:38:13,046][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.10787076503038406, 'train/grad_norm_pre_clip': 6.982759952545166, 'timing/data': 0.012655066326260567, 'timing/model': 1.4812130834907293, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 0.24}) +[2026-06-15 20:38:19,240][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.06540423631668091, 'train/grad_norm_pre_clip': 5.420105457305908, 'timing/data': 0.011980555951595306, 'timing/model': 1.532811339944601, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 0.24}) +[2026-06-15 20:38:25,369][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.22140559554100037, 'train/grad_norm_pre_clip': 9.103581428527832, 'timing/data': 0.013049956411123276, 'timing/model': 1.7145854756236076, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 0.24}) +[2026-06-15 20:38:31,334][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.21298106014728546, 'train/grad_norm_pre_clip': 16.037099838256836, 'timing/data': 0.012693982571363449, 'timing/model': 1.5229815859347582, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 0.24}) +[2026-06-15 20:38:37,434][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.09151982516050339, 'train/grad_norm_pre_clip': 6.805119037628174, 'timing/data': 0.012447454035282135, 'timing/model': 1.489413945004344, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 0.24}) +[2026-06-15 20:38:43,520][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.35236239433288574, 'train/grad_norm_pre_clip': 16.879791259765625, 'timing/data': 0.011728329584002495, 'timing/model': 1.5517934877425432, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 0.25}) +[2026-06-15 20:38:49,767][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.12542682886123657, 'train/grad_norm_pre_clip': 19.95637321472168, 'timing/data': 0.011599240824580193, 'timing/model': 1.553648678585887, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 0.25}) +[2026-06-15 20:38:56,419][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.11831199377775192, 'train/grad_norm_pre_clip': 8.039978981018066, 'timing/data': 0.01887800358235836, 'timing/model': 1.6628116015344858, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 0.25}) +[2026-06-15 20:39:03,363][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.10640934854745865, 'train/grad_norm_pre_clip': 6.139134407043457, 'timing/data': 0.02495967224240303, 'timing/model': 1.9131003208458424, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 0.25}) +[2026-06-15 20:39:09,486][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.028913935646414757, 'train/grad_norm_pre_clip': 5.317960262298584, 'timing/data': 0.011672429740428925, 'timing/model': 1.5446560196578503, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 0.25}) +[2026-06-15 20:39:15,882][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.12600430846214294, 'train/grad_norm_pre_clip': 9.301959037780762, 'timing/data': 0.011698814108967781, 'timing/model': 1.5446543339639902, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.25}) +[2026-06-15 20:39:22,190][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.43944892287254333, 'train/grad_norm_pre_clip': 14.196770668029785, 'timing/data': 0.013581722974777222, 'timing/model': 1.5595230851322412, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 0.25}) +[2026-06-15 20:39:28,489][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.28188610076904297, 'train/grad_norm_pre_clip': 10.537959098815918, 'timing/data': 0.012722738087177277, 'timing/model': 1.5483566112816334, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 0.25}) +[2026-06-15 20:39:34,863][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.10199485719203949, 'train/grad_norm_pre_clip': 14.459989547729492, 'timing/data': 0.012506650760769844, 'timing/model': 1.5627944357693195, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 0.25}) +[2026-06-15 20:39:41,293][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.4461810290813446, 'train/grad_norm_pre_clip': 19.126123428344727, 'timing/data': 0.013549940660595894, 'timing/model': 1.811054952442646, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 0.25}) +[2026-06-15 20:39:47,232][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.05461696907877922, 'train/grad_norm_pre_clip': 5.37482213973999, 'timing/data': 0.012575052678585052, 'timing/model': 1.5266954153776169, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 0.25}) +[2026-06-15 20:39:53,334][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.18774962425231934, 'train/grad_norm_pre_clip': 15.131743431091309, 'timing/data': 0.012323733419179916, 'timing/model': 1.5437023770064116, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 0.25}) +[2026-06-15 20:39:59,179][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.13989433646202087, 'train/grad_norm_pre_clip': 8.492932319641113, 'timing/data': 0.010485516861081123, 'timing/model': 1.452683947980404, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 0.25}) +[2026-06-15 20:40:05,082][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.10089796036481857, 'train/grad_norm_pre_clip': 4.053798675537109, 'timing/data': 0.01062685251235962, 'timing/model': 1.4577749334275723, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 0.25}) +[2026-06-15 20:40:10,980][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.15750989317893982, 'train/grad_norm_pre_clip': 10.033835411071777, 'timing/data': 0.010383270680904388, 'timing/model': 1.4570260811597109, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 0.25}) +[2026-06-15 20:40:16,879][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.49811071157455444, 'train/grad_norm_pre_clip': 14.549586296081543, 'timing/data': 0.010606436058878899, 'timing/model': 1.6560702845454216, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 0.25}) +[2026-06-15 20:40:22,842][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.305124431848526, 'train/grad_norm_pre_clip': 7.4209113121032715, 'timing/data': 0.01027807965874672, 'timing/model': 1.683753751218319, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 0.25}) +[2026-06-15 20:40:28,626][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.3282715976238251, 'train/grad_norm_pre_clip': 19.263423919677734, 'timing/data': 0.01121133379638195, 'timing/model': 1.470608089119196, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 0.25}) +[2026-06-15 20:40:34,618][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.1639941930770874, 'train/grad_norm_pre_clip': 9.418333053588867, 'timing/data': 0.010422986000776291, 'timing/model': 1.509747987613082, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 0.25}) +[2026-06-15 20:40:40,822][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.21464122831821442, 'train/grad_norm_pre_clip': 5.7821784019470215, 'timing/data': 0.011647883802652359, 'timing/model': 1.5419745184481144, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 0.25}) +[2026-06-15 20:40:46,953][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.15419597923755646, 'train/grad_norm_pre_clip': 4.372828006744385, 'timing/data': 0.01678554154932499, 'timing/model': 1.5397747252136469, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 0.25}) +[2026-06-15 20:40:53,161][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.1028064489364624, 'train/grad_norm_pre_clip': 4.080408096313477, 'timing/data': 0.013064542785286903, 'timing/model': 1.7712824307382107, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 0.25}) +[2026-06-15 20:40:59,219][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.311648428440094, 'train/grad_norm_pre_clip': 7.677615165710449, 'timing/data': 0.012544028460979462, 'timing/model': 1.5375491827726364, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 0.25}) +[2026-06-15 20:41:05,729][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.09967684000730515, 'train/grad_norm_pre_clip': 6.244476318359375, 'timing/data': 0.018916679546236992, 'timing/model': 1.6590154487639666, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 0.25}) +[2026-06-15 20:41:12,626][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.21522143483161926, 'train/grad_norm_pre_clip': 11.79732894897461, 'timing/data': 0.018602708354592323, 'timing/model': 1.7328452616930008, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 0.25}) +[2026-06-15 20:41:19,570][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.19897516071796417, 'train/grad_norm_pre_clip': 7.923460483551025, 'timing/data': 0.018061617389321327, 'timing/model': 1.6534493118524551, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 0.26}) +[2026-06-15 20:41:26,072][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.3235357105731964, 'train/grad_norm_pre_clip': 13.112397193908691, 'timing/data': 0.012881672009825706, 'timing/model': 1.548380471765995, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.26}) +[2026-06-15 20:41:32,775][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.3550988733768463, 'train/grad_norm_pre_clip': 8.581221580505371, 'timing/data': 0.012839587405323982, 'timing/model': 1.8579555675387383, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 0.26}) +[2026-06-15 20:41:38,865][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.19115599989891052, 'train/grad_norm_pre_clip': 4.849575042724609, 'timing/data': 0.011939549818634987, 'timing/model': 1.5531300492584705, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 0.26}) +[2026-06-15 20:41:45,165][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.2697214186191559, 'train/grad_norm_pre_clip': 4.702699184417725, 'timing/data': 0.013026043772697449, 'timing/model': 1.5520749539136887, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 0.26}) +[2026-06-15 20:41:51,467][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.23600584268569946, 'train/grad_norm_pre_clip': 8.608820915222168, 'timing/data': 0.01281639002263546, 'timing/model': 1.5565707739442587, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 0.26}) +[2026-06-15 20:41:57,866][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.286081999540329, 'train/grad_norm_pre_clip': 10.318492889404297, 'timing/data': 0.012069262564182281, 'timing/model': 1.5645318627357483, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 0.26}) +[2026-06-15 20:42:04,137][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.11196765303611755, 'train/grad_norm_pre_clip': 3.572679281234741, 'timing/data': 0.011109694838523865, 'timing/model': 1.4802744202315807, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 0.26}) +[2026-06-15 20:42:10,238][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.09057635813951492, 'train/grad_norm_pre_clip': 3.2341885566711426, 'timing/data': 0.012619657441973686, 'timing/model': 1.694901991635561, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 0.26}) +[2026-06-15 20:42:16,417][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.05112629383802414, 'train/grad_norm_pre_clip': 2.066816806793213, 'timing/data': 0.012816928327083588, 'timing/model': 1.7355454955250025, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 0.26}) +[2026-06-15 20:42:22,625][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.1574493646621704, 'train/grad_norm_pre_clip': 9.68714714050293, 'timing/data': 0.011712247505784035, 'timing/model': 1.544758616015315, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 0.26}) +[2026-06-15 20:42:28,968][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.3840521574020386, 'train/grad_norm_pre_clip': 16.04928207397461, 'timing/data': 0.012415310367941856, 'timing/model': 1.498820437118411, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 0.26}) +[2026-06-15 20:42:35,228][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.16916194558143616, 'train/grad_norm_pre_clip': 5.804520130157471, 'timing/data': 0.011721273884177208, 'timing/model': 1.4746617432683706, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 0.26}) +[2026-06-15 20:42:41,550][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.40517663955688477, 'train/grad_norm_pre_clip': 16.33436393737793, 'timing/data': 0.013411100953817368, 'timing/model': 1.7471615634858608, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 0.26}) +[2026-06-15 20:42:47,537][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.332588255405426, 'train/grad_norm_pre_clip': 18.4428768157959, 'timing/data': 0.012378158047795296, 'timing/model': 1.5452370941638947, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 0.26}) +[2026-06-15 20:42:54,176][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.3851124942302704, 'train/grad_norm_pre_clip': 20.719036102294922, 'timing/data': 0.013023627921938896, 'timing/model': 1.5313177034258842, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 0.26}) +[2026-06-15 20:43:00,457][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.502913236618042, 'train/grad_norm_pre_clip': 20.13347625732422, 'timing/data': 0.018977733328938484, 'timing/model': 1.6016300488263369, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 0.26}) +[2026-06-15 20:43:06,880][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.20810694992542267, 'train/grad_norm_pre_clip': 11.726116180419922, 'timing/data': 0.013856528326869011, 'timing/model': 1.5401105266064405, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 0.26}) +[2026-06-15 20:43:13,709][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.25122278928756714, 'train/grad_norm_pre_clip': 7.847662925720215, 'timing/data': 0.015545964241027832, 'timing/model': 1.6498041804879904, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 0.26}) +[2026-06-15 20:43:20,401][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.21077372133731842, 'train/grad_norm_pre_clip': 9.21895694732666, 'timing/data': 0.012712981551885605, 'timing/model': 1.893549744039774, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 0.26}) +[2026-06-15 20:43:26,468][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.16471992433071136, 'train/grad_norm_pre_clip': 15.105690002441406, 'timing/data': 0.012708723545074463, 'timing/model': 1.5578092467039824, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 0.26}) +[2026-06-15 20:43:32,876][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.07731419056653976, 'train/grad_norm_pre_clip': 7.820648670196533, 'timing/data': 0.012068213894963264, 'timing/model': 1.5579488761723042, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 0.26}) +[2026-06-15 20:43:39,173][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.2210191786289215, 'train/grad_norm_pre_clip': 20.42858123779297, 'timing/data': 0.012646012008190155, 'timing/model': 1.5435748044401407, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 0.26}) +[2026-06-15 20:43:45,646][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.28197771310806274, 'train/grad_norm_pre_clip': 14.065320014953613, 'timing/data': 0.01080239750444889, 'timing/model': 1.6214708406478167, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 0.26}) +[2026-06-15 20:43:51,739][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.705986738204956, 'train/grad_norm_pre_clip': 25.54799461364746, 'timing/data': 0.01286889798939228, 'timing/model': 1.4829303920269012, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 0.26}) +[2026-06-15 20:43:57,946][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.231200709939003, 'train/grad_norm_pre_clip': 5.7397541999816895, 'timing/data': 0.013665355741977692, 'timing/model': 1.7395889069885015, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 0.26}) +[2026-06-15 20:44:03,953][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.30673229694366455, 'train/grad_norm_pre_clip': 8.977259635925293, 'timing/data': 0.015603562816977501, 'timing/model': 1.5320899486541748, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 0.27}) +[2026-06-15 20:44:10,373][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.09803803265094757, 'train/grad_norm_pre_clip': 2.9773221015930176, 'timing/data': 0.01937517710030079, 'timing/model': 1.5433840937912464, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 0.27}) +[2026-06-15 20:44:16,760][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.29279977083206177, 'train/grad_norm_pre_clip': 12.0844144821167, 'timing/data': 0.013196233659982681, 'timing/model': 1.5694837179034948, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 0.27}) +[2026-06-15 20:44:23,042][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.14098015427589417, 'train/grad_norm_pre_clip': 6.170985698699951, 'timing/data': 0.015580687671899796, 'timing/model': 1.504038903862238, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.27}) +[2026-06-15 20:44:29,247][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.0812637060880661, 'train/grad_norm_pre_clip': 4.21087121963501, 'timing/data': 0.012788884341716766, 'timing/model': 1.5582715068012476, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 0.27}) +[2026-06-15 20:44:35,530][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.11957048624753952, 'train/grad_norm_pre_clip': 5.824039459228516, 'timing/data': 0.011734174564480782, 'timing/model': 1.834606185555458, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 0.27}) +[2026-06-15 20:44:41,438][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.3754633367061615, 'train/grad_norm_pre_clip': 11.134121894836426, 'timing/data': 0.012184374034404755, 'timing/model': 1.501448756083846, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 0.27}) +[2026-06-15 20:44:47,537][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.25121405720710754, 'train/grad_norm_pre_clip': 12.29266357421875, 'timing/data': 0.012874335050582886, 'timing/model': 1.534553924575448, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 0.27}) +[2026-06-15 20:44:53,740][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.2198551744222641, 'train/grad_norm_pre_clip': 6.38364839553833, 'timing/data': 0.012146228924393654, 'timing/model': 1.5541734360158443, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 0.27}) +[2026-06-15 20:44:59,844][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.07093574851751328, 'train/grad_norm_pre_clip': 3.914935350418091, 'timing/data': 0.013298260048031807, 'timing/model': 1.5432945974171162, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 0.27}) +[2026-06-15 20:45:06,395][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.1529851108789444, 'train/grad_norm_pre_clip': 7.307392597198486, 'timing/data': 0.01830919273197651, 'timing/model': 1.7125727031379938, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.27}) +[2026-06-15 20:45:13,364][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.19440428912639618, 'train/grad_norm_pre_clip': 9.796442985534668, 'timing/data': 0.017235880717635155, 'timing/model': 2.078259291127324, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 0.27}) +[2026-06-15 20:45:19,929][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.2984994351863861, 'train/grad_norm_pre_clip': 13.635614395141602, 'timing/data': 0.019720526412129402, 'timing/model': 1.6602030638605356, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 0.27}) +[2026-06-15 20:45:26,826][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.10606103390455246, 'train/grad_norm_pre_clip': 10.469500541687012, 'timing/data': 0.018225008621811867, 'timing/model': 1.662076411768794, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 0.27}) +[2026-06-15 20:45:33,336][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.25336357951164246, 'train/grad_norm_pre_clip': 11.561995506286621, 'timing/data': 0.011816175654530525, 'timing/model': 1.5454285517334938, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 0.27}) +[2026-06-15 20:53:30,792][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.2093064785003662, 'train/grad_norm_pre_clip': 16.417190551757812, 'eval/loss': 0.1016061082482338, 'eval/flap_precision': 0.899581589958159, 'eval/flap_recall': 0.8751696065128901, 'eval/flap_f1': 0.8872077028885833, 'eval/flap_rate_model': 0.19916666666666666, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8872077028885833, 'timing/data': 0.01256379671394825, 'timing/model': 1.5477141104638577, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 0.27}) +[2026-06-15 20:53:30,793][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_700_state +[2026-06-15 20:53:30,797][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:53:36,726][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_700_state/model.safetensors +[2026-06-15 20:53:50,416][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_700_state/optimizer.bin +[2026-06-15 20:53:50,416][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_700_state/sampler.bin +[2026-06-15 20:53:50,416][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_700_state/sampler_1.bin +[2026-06-15 20:53:50,417][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_700_state/random_states_0.pkl +[2026-06-15 20:53:50,514][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700 +[2026-06-15 20:53:50,514][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:53:50,523][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:53:56,398][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.19532346725463867, 'train/grad_norm_pre_clip': 12.540177345275879, 'timing/data': 0.00997145101428032, 'timing/model': 1.4462345093488693, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 0.27}) +[2026-06-15 20:54:02,304][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.1005842387676239, 'train/grad_norm_pre_clip': 4.031957626342773, 'timing/data': 0.011579984799027443, 'timing/model': 1.4354410097002983, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 0.27}) +[2026-06-15 20:54:08,102][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.1906859129667282, 'train/grad_norm_pre_clip': 12.023046493530273, 'timing/data': 0.011297084391117096, 'timing/model': 1.4435898959636688, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 0.27}) +[2026-06-15 20:54:13,865][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.32730838656425476, 'train/grad_norm_pre_clip': 13.388528823852539, 'timing/data': 0.012057624757289886, 'timing/model': 1.4961480777710676, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 0.27}) +[2026-06-15 20:54:19,475][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.1528596431016922, 'train/grad_norm_pre_clip': 6.354780197143555, 'timing/data': 0.010044269263744354, 'timing/model': 1.4457567557692528, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 0.27}) +[2026-06-15 20:54:25,278][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.26172125339508057, 'train/grad_norm_pre_clip': 18.509294509887695, 'timing/data': 0.010917959734797478, 'timing/model': 1.4505782313644886, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 0.27}) +[2026-06-15 20:54:31,176][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.10020112246274948, 'train/grad_norm_pre_clip': 4.376757621765137, 'timing/data': 0.010643808171153069, 'timing/model': 1.455624995753169, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 0.27}) +[2026-06-15 20:54:37,074][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.22627480328083038, 'train/grad_norm_pre_clip': 8.70257568359375, 'timing/data': 0.009820591658353806, 'timing/model': 1.4202968683093786, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 0.27}) +[2026-06-15 20:54:42,980][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.233086496591568, 'train/grad_norm_pre_clip': 6.266672611236572, 'timing/data': 0.010234873741865158, 'timing/model': 1.649805972352624, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 0.27}) +[2026-06-15 20:54:48,676][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.2505742907524109, 'train/grad_norm_pre_clip': 12.94452953338623, 'timing/data': 0.010389579460024834, 'timing/model': 1.4318170435726643, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 0.27}) +[2026-06-15 20:54:54,579][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.4669833183288574, 'train/grad_norm_pre_clip': 16.9835262298584, 'timing/data': 0.01100754551589489, 'timing/model': 1.4485028237104416, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 0.28}) +[2026-06-15 20:55:00,479][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.22263851761817932, 'train/grad_norm_pre_clip': 7.781375408172607, 'timing/data': 0.016629870980978012, 'timing/model': 1.4506539311259985, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 0.28}) +[2026-06-15 20:55:06,381][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.26421040296554565, 'train/grad_norm_pre_clip': 8.941143035888672, 'timing/data': 0.010968737304210663, 'timing/model': 1.4551355186849833, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 0.28}) +[2026-06-15 20:55:12,286][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.08161020278930664, 'train/grad_norm_pre_clip': 3.5509698390960693, 'timing/data': 0.011075612157583237, 'timing/model': 1.497631473466754, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 0.28}) +[2026-06-15 20:55:18,185][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.24776192009449005, 'train/grad_norm_pre_clip': 8.417337417602539, 'timing/data': 0.010856619104743004, 'timing/model': 1.6614206973463297, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 0.28}) +[2026-06-15 20:55:23,883][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.18500670790672302, 'train/grad_norm_pre_clip': 5.600353240966797, 'timing/data': 0.011238528415560722, 'timing/model': 1.4595465920865536, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 0.28}) +[2026-06-15 20:55:29,806][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.2468334287405014, 'train/grad_norm_pre_clip': 8.363468170166016, 'timing/data': 0.0108388252556324, 'timing/model': 1.4765269868075848, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 0.28}) +[2026-06-15 20:55:35,980][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.14359602332115173, 'train/grad_norm_pre_clip': 7.442277908325195, 'timing/data': 0.010894663631916046, 'timing/model': 1.4483035951852798, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 0.28}) +[2026-06-15 20:55:41,884][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.35530388355255127, 'train/grad_norm_pre_clip': 14.14304256439209, 'timing/data': 0.011132966727018356, 'timing/model': 1.4616371728479862, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 0.28}) +[2026-06-15 20:55:48,026][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.28894180059432983, 'train/grad_norm_pre_clip': 11.879555702209473, 'timing/data': 0.01250390149652958, 'timing/model': 1.559935700148344, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 0.28}) +[2026-06-15 20:55:54,177][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.12639851868152618, 'train/grad_norm_pre_clip': 4.920924186706543, 'timing/data': 0.012883055955171585, 'timing/model': 1.7813548520207405, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 0.28}) +[2026-06-15 20:56:00,352][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.18291924893856049, 'train/grad_norm_pre_clip': 6.31239652633667, 'timing/data': 0.012375470250844955, 'timing/model': 1.7614789083600044, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 0.28}) +[2026-06-15 20:56:06,442][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.4038645923137665, 'train/grad_norm_pre_clip': 19.315872192382812, 'timing/data': 0.012809816747903824, 'timing/model': 1.4817701876163483, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 0.28}) +[2026-06-15 20:56:12,638][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.32766038179397583, 'train/grad_norm_pre_clip': 22.11037254333496, 'timing/data': 0.01220872811973095, 'timing/model': 1.5495949927717447, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 0.28}) +[2026-06-15 20:56:18,673][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.10777364671230316, 'train/grad_norm_pre_clip': 6.146875858306885, 'timing/data': 0.012500623241066933, 'timing/model': 1.4991560354828835, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 0.28}) +[2026-06-15 20:56:24,842][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.4287073612213135, 'train/grad_norm_pre_clip': 20.511018753051758, 'timing/data': 0.012076042592525482, 'timing/model': 1.5234613176435232, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 0.28}) +[2026-06-15 20:56:30,986][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.15450957417488098, 'train/grad_norm_pre_clip': 19.483779907226562, 'timing/data': 0.014029575511813164, 'timing/model': 1.7815448101609945, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 0.28}) +[2026-06-15 20:56:37,140][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.3283468186855316, 'train/grad_norm_pre_clip': 21.602895736694336, 'timing/data': 0.011862339451909065, 'timing/model': 1.5657257605344057, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 0.28}) +[2026-06-15 20:56:43,242][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.1605842113494873, 'train/grad_norm_pre_clip': 8.908458709716797, 'timing/data': 0.013161120936274529, 'timing/model': 1.5097450707107782, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 0.28}) +[2026-06-15 20:56:49,338][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.12457747757434845, 'train/grad_norm_pre_clip': 7.759832859039307, 'timing/data': 0.01350797712802887, 'timing/model': 1.5338564347475767, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.28}) +[2026-06-15 20:56:55,184][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.3721123933792114, 'train/grad_norm_pre_clip': 14.120145797729492, 'timing/data': 0.010418388992547989, 'timing/model': 1.4598748851567507, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 0.28}) +[2026-06-15 20:57:00,983][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.3914559781551361, 'train/grad_norm_pre_clip': 13.310355186462402, 'timing/data': 0.01035955548286438, 'timing/model': 1.4451427217572927, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 0.28}) +[2026-06-15 20:57:06,831][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.25689712166786194, 'train/grad_norm_pre_clip': 8.746299743652344, 'timing/data': 0.01113218069076538, 'timing/model': 1.6024574525654316, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 0.28}) +[2026-06-15 20:57:12,484][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.2935803532600403, 'train/grad_norm_pre_clip': 13.316324234008789, 'timing/data': 0.010436993092298508, 'timing/model': 1.4557419624179602, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 0.28}) +[2026-06-15 20:57:18,385][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.4729040861129761, 'train/grad_norm_pre_clip': 11.289423942565918, 'timing/data': 0.010848376899957657, 'timing/model': 1.4565918799489737, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 0.28}) +[2026-06-15 20:57:24,283][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.21916796267032623, 'train/grad_norm_pre_clip': 8.346287727355957, 'timing/data': 0.010586487129330635, 'timing/model': 1.4613203909248114, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 0.28}) +[2026-06-15 20:57:30,182][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.17929036915302277, 'train/grad_norm_pre_clip': 6.326100826263428, 'timing/data': 0.011305548250675201, 'timing/model': 1.4501168448477983, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.29}) +[2026-06-15 20:57:36,084][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.22155819833278656, 'train/grad_norm_pre_clip': 6.087711811065674, 'timing/data': 0.010401707142591476, 'timing/model': 1.464626682922244, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 0.29}) +[2026-06-15 20:57:41,943][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.16069379448890686, 'train/grad_norm_pre_clip': 8.338677406311035, 'timing/data': 0.011202486231923103, 'timing/model': 1.613316560164094, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 0.29}) +[2026-06-15 20:57:47,584][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.23795747756958008, 'train/grad_norm_pre_clip': 5.819955825805664, 'timing/data': 0.010939301922917366, 'timing/model': 1.4509232267737389, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 0.29}) +[2026-06-15 20:57:53,484][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.3063933849334717, 'train/grad_norm_pre_clip': 6.707716464996338, 'timing/data': 0.010344177484512329, 'timing/model': 1.4543867446482182, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 0.29}) +[2026-06-15 20:57:59,384][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.23274902999401093, 'train/grad_norm_pre_clip': 4.92919921875, 'timing/data': 0.010834712535142899, 'timing/model': 1.4537803642451763, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 0.29}) +[2026-06-15 20:58:05,284][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.24404795467853546, 'train/grad_norm_pre_clip': 4.846412181854248, 'timing/data': 0.01153586432337761, 'timing/model': 1.475032301619649, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 0.29}) +[2026-06-15 20:58:11,188][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.2954883873462677, 'train/grad_norm_pre_clip': 7.171014308929443, 'timing/data': 0.010293785482645035, 'timing/model': 1.649902867153287, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 0.29}) +[2026-06-15 20:58:16,884][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.12537223100662231, 'train/grad_norm_pre_clip': 3.3946855068206787, 'timing/data': 0.010755334049463272, 'timing/model': 1.454545360058546, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 0.29}) +[2026-06-15 20:58:22,786][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.17811056971549988, 'train/grad_norm_pre_clip': 4.014373779296875, 'timing/data': 0.009887311607599258, 'timing/model': 1.4530080389231443, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 0.29}) +[2026-06-15 20:58:28,489][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.16564004123210907, 'train/grad_norm_pre_clip': 5.9947662353515625, 'timing/data': 0.01040695607662201, 'timing/model': 1.459118116647005, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 0.29}) +[2026-06-15 20:58:34,291][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.16870087385177612, 'train/grad_norm_pre_clip': 8.503504753112793, 'timing/data': 0.01142864115536213, 'timing/model': 1.4641099963337183, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 0.29}) +[2026-06-15 20:58:40,099][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.4077741801738739, 'train/grad_norm_pre_clip': 8.995226860046387, 'timing/data': 0.010546976700425148, 'timing/model': 1.426282074302435, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 0.29}) +[2026-06-15 20:58:46,381][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.07989392429590225, 'train/grad_norm_pre_clip': 11.196250915527344, 'timing/data': 0.012417364865541458, 'timing/model': 1.79928545281291, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 0.29}) +[2026-06-15 20:58:52,534][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.37798357009887695, 'train/grad_norm_pre_clip': 18.400161743164062, 'timing/data': 0.012502651661634445, 'timing/model': 1.644759001210332, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 0.29}) +[2026-06-15 20:58:58,654][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.4253077208995819, 'train/grad_norm_pre_clip': 28.6904296875, 'timing/data': 0.012561926618218422, 'timing/model': 1.5182993318885565, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 0.29}) +[2026-06-15 20:59:04,838][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.17215192317962646, 'train/grad_norm_pre_clip': 15.01389217376709, 'timing/data': 0.013669941574335098, 'timing/model': 1.5442776810377836, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 0.29}) +[2026-06-15 20:59:10,942][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.16495168209075928, 'train/grad_norm_pre_clip': 13.790386199951172, 'timing/data': 0.012558821588754654, 'timing/model': 1.5524253565818071, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 0.29}) +[2026-06-15 20:59:17,126][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.42403146624565125, 'train/grad_norm_pre_clip': 16.059268951416016, 'timing/data': 0.012221317738294601, 'timing/model': 1.5533098708838224, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 0.29}) +[2026-06-15 20:59:23,287][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.2828083038330078, 'train/grad_norm_pre_clip': 12.731053352355957, 'timing/data': 0.011719759553670883, 'timing/model': 1.7934699151664972, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 0.29}) +[2026-06-15 20:59:29,305][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.2755398750305176, 'train/grad_norm_pre_clip': 14.999591827392578, 'timing/data': 0.01286027580499649, 'timing/model': 1.5363212171941996, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 0.29}) +[2026-06-15 20:59:35,536][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.42552900314331055, 'train/grad_norm_pre_clip': 23.02332878112793, 'timing/data': 0.011988097801804543, 'timing/model': 1.55302307382226, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 0.29}) +[2026-06-15 20:59:41,740][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.1440127044916153, 'train/grad_norm_pre_clip': 5.820819854736328, 'timing/data': 0.012945428490638733, 'timing/model': 1.5489551480859518, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 0.29}) +[2026-06-15 20:59:47,941][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.2355993539094925, 'train/grad_norm_pre_clip': 12.431605339050293, 'timing/data': 0.01263146661221981, 'timing/model': 1.5401740595698357, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 0.29}) +[2026-06-15 20:59:54,141][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.07901807874441147, 'train/grad_norm_pre_clip': 5.2609944343566895, 'timing/data': 0.012235680595040321, 'timing/model': 1.5549465101212263, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 0.29}) +[2026-06-15 21:00:00,273][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.18238377571105957, 'train/grad_norm_pre_clip': 8.344470977783203, 'timing/data': 0.012841576710343361, 'timing/model': 1.7285643126815557, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 0.29}) +[2026-06-15 21:00:06,328][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.19204461574554443, 'train/grad_norm_pre_clip': 7.544445037841797, 'timing/data': 0.012750552967190742, 'timing/model': 1.5622529070824385, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 0.3}) +[2026-06-15 21:00:12,464][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.14576029777526855, 'train/grad_norm_pre_clip': 6.634995937347412, 'timing/data': 0.012458467856049538, 'timing/model': 1.5718688890337944, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 0.3}) +[2026-06-15 21:00:18,664][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.280056893825531, 'train/grad_norm_pre_clip': 8.921013832092285, 'timing/data': 0.013073556125164032, 'timing/model': 1.5218224618583918, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.3}) +[2026-06-15 21:00:24,840][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.08604395389556885, 'train/grad_norm_pre_clip': 7.159809112548828, 'timing/data': 0.01211848296225071, 'timing/model': 1.555511076003313, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 0.3}) +[2026-06-15 21:00:30,936][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.08127380907535553, 'train/grad_norm_pre_clip': 5.704520225524902, 'timing/data': 0.013728432357311249, 'timing/model': 1.47931619733572, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 0.3}) +[2026-06-15 21:00:37,171][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.22756806015968323, 'train/grad_norm_pre_clip': 12.272747993469238, 'timing/data': 0.012828122824430466, 'timing/model': 1.802142210304737, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 0.3}) +[2026-06-15 21:00:43,142][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.23606713116168976, 'train/grad_norm_pre_clip': 8.93822193145752, 'timing/data': 0.013465926051139832, 'timing/model': 1.5107003785669804, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 0.3}) +[2026-06-15 21:00:49,451][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.25045663118362427, 'train/grad_norm_pre_clip': 8.583163261413574, 'timing/data': 0.013324957340955734, 'timing/model': 1.6750646531581879, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 0.3}) +[2026-06-15 21:00:55,869][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.30265170335769653, 'train/grad_norm_pre_clip': 8.911643981933594, 'timing/data': 0.011356502771377563, 'timing/model': 1.5967319440096617, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 0.3}) +[2026-06-15 21:01:02,083][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.3576904237270355, 'train/grad_norm_pre_clip': 11.472879409790039, 'timing/data': 0.011861748993396759, 'timing/model': 1.5652390457689762, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 0.3}) +[2026-06-15 21:01:08,695][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.07668016105890274, 'train/grad_norm_pre_clip': 4.304683685302734, 'timing/data': 0.014190152287483215, 'timing/model': 1.579762103036046, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 0.3}) +[2026-06-15 21:01:15,076][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.14409245550632477, 'train/grad_norm_pre_clip': 9.592854499816895, 'timing/data': 0.014297347515821457, 'timing/model': 1.8346865698695183, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 0.3}) +[2026-06-15 21:01:21,393][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.1761588454246521, 'train/grad_norm_pre_clip': 8.182302474975586, 'timing/data': 0.01603787951171398, 'timing/model': 1.5877697952091694, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 0.3}) +[2026-06-15 21:01:27,740][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.16681116819381714, 'train/grad_norm_pre_clip': 7.451217174530029, 'timing/data': 0.012977028265595436, 'timing/model': 1.5398692525923252, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 0.3}) +[2026-06-15 21:01:33,585][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.16406582295894623, 'train/grad_norm_pre_clip': 7.813756942749023, 'timing/data': 0.01116703636944294, 'timing/model': 1.4545146524906158, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 0.3}) +[2026-06-15 21:01:39,385][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.14397646486759186, 'train/grad_norm_pre_clip': 6.9983367919921875, 'timing/data': 0.010251568630337715, 'timing/model': 1.453803637996316, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 0.3}) +[2026-06-15 21:01:45,286][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.0774514228105545, 'train/grad_norm_pre_clip': 3.7911126613616943, 'timing/data': 0.01063087210059166, 'timing/model': 1.4492605794221163, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 0.3}) +[2026-06-15 21:01:51,638][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.1769554615020752, 'train/grad_norm_pre_clip': 5.9038920402526855, 'timing/data': 0.013207780197262764, 'timing/model': 1.8339863680303097, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 0.3}) +[2026-06-15 21:01:57,552][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.060311585664749146, 'train/grad_norm_pre_clip': 2.526093006134033, 'timing/data': 0.013142261654138565, 'timing/model': 1.4835672117769718, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 0.3}) +[2026-06-15 21:02:03,741][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.1466687023639679, 'train/grad_norm_pre_clip': 7.154332160949707, 'timing/data': 0.013229146599769592, 'timing/model': 1.499497963115573, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 0.3}) +[2026-06-15 21:02:10,061][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.01728183403611183, 'train/grad_norm_pre_clip': 1.5905652046203613, 'timing/data': 0.012178352102637291, 'timing/model': 1.5441065300256014, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 0.3}) +[2026-06-15 21:02:16,261][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.1396063268184662, 'train/grad_norm_pre_clip': 7.92185115814209, 'timing/data': 0.013415290042757988, 'timing/model': 1.536458559334278, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 0.3}) +[2026-06-15 21:02:22,551][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.05055120959877968, 'train/grad_norm_pre_clip': 5.1425275802612305, 'timing/data': 0.014004377648234367, 'timing/model': 1.5140074882656336, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 0.3}) +[2026-06-15 21:02:28,786][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.19013822078704834, 'train/grad_norm_pre_clip': 16.349971771240234, 'timing/data': 0.012078113853931427, 'timing/model': 1.7930426951497793, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.3}) +[2026-06-15 21:02:35,025][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.045905157923698425, 'train/grad_norm_pre_clip': 6.247809410095215, 'timing/data': 0.013303620740771294, 'timing/model': 1.72645771689713, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 0.3}) +[2026-06-15 21:02:41,049][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.26327890157699585, 'train/grad_norm_pre_clip': 10.197498321533203, 'timing/data': 0.012881863862276077, 'timing/model': 1.5619981605559587, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 0.3}) +[2026-06-15 21:02:47,347][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.4905107319355011, 'train/grad_norm_pre_clip': 23.007383346557617, 'timing/data': 0.012699557468295097, 'timing/model': 1.558647532016039, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 0.31}) +[2026-06-15 21:02:53,547][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.7013146877288818, 'train/grad_norm_pre_clip': 32.53125, 'timing/data': 0.012802353128790855, 'timing/model': 1.559389865025878, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 0.31}) +[2026-06-15 21:02:59,757][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.40941596031188965, 'train/grad_norm_pre_clip': 13.640701293945312, 'timing/data': 0.01224815659224987, 'timing/model': 1.5285151097923517, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 0.31}) +[2026-06-15 21:03:06,088][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.10414022207260132, 'train/grad_norm_pre_clip': 11.479907035827637, 'timing/data': 0.013464370742440224, 'timing/model': 1.7979072369635105, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 0.31}) +[2026-06-15 21:03:12,147][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.26246511936187744, 'train/grad_norm_pre_clip': 18.152183532714844, 'timing/data': 0.01164235733449459, 'timing/model': 1.556673651561141, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 0.31}) +[2026-06-15 21:03:18,350][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.0350971519947052, 'train/grad_norm_pre_clip': 5.956549644470215, 'timing/data': 0.011582789942622185, 'timing/model': 1.5515658408403397, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 0.31}) +[2026-06-15 21:03:24,656][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.2999689280986786, 'train/grad_norm_pre_clip': 14.386012077331543, 'timing/data': 0.011670220643281937, 'timing/model': 1.661632090806961, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 0.31}) +[2026-06-15 21:03:30,843][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.25006628036499023, 'train/grad_norm_pre_clip': 12.027094841003418, 'timing/data': 0.012651143595576286, 'timing/model': 1.48840238340199, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 0.31}) +[2026-06-15 21:03:37,042][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.15034960210323334, 'train/grad_norm_pre_clip': 20.492685317993164, 'timing/data': 0.012574620544910431, 'timing/model': 1.549707179889083, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 0.31}) +[2026-06-15 21:03:43,232][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.10335056483745575, 'train/grad_norm_pre_clip': 7.397607326507568, 'timing/data': 0.012637503445148468, 'timing/model': 1.7209273464977741, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 0.31}) +[2026-06-15 21:03:49,237][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.051662568002939224, 'train/grad_norm_pre_clip': 5.396174430847168, 'timing/data': 0.012829218059778214, 'timing/model': 1.5172897446900606, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 0.31}) +[2026-06-15 21:11:10,867][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.24209900200366974, 'train/grad_norm_pre_clip': 6.337982177734375, 'eval/loss': 0.10994815826416016, 'eval/flap_precision': 0.8626716604244694, 'eval/flap_recall': 0.937584803256445, 'eval/flap_f1': 0.8985695708712614, 'eval/flap_rate_model': 0.2225, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8985695708712614, 'timing/data': 0.014008497819304466, 'timing/model': 1.5377746261656284, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.31}) +[2026-06-15 21:11:10,868][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_800_state +[2026-06-15 21:11:10,875][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 21:11:21,964][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_800_state/model.safetensors +[2026-06-15 21:11:33,257][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_800_state/optimizer.bin +[2026-06-15 21:11:33,257][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_800_state/sampler.bin +[2026-06-15 21:11:33,258][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_800_state/sampler_1.bin +[2026-06-15 21:11:33,259][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_800_state/random_states_0.pkl +[2026-06-15 21:11:33,359][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800 +[2026-06-15 21:11:33,359][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 21:11:33,367][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 21:11:39,421][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.4713733494281769, 'train/grad_norm_pre_clip': 27.792800903320312, 'timing/data': 0.012885719537734985, 'timing/model': 1.452140124514699, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 0.31}) +[2026-06-15 21:11:45,424][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.17104049026966095, 'train/grad_norm_pre_clip': 14.004708290100098, 'timing/data': 0.015477115288376808, 'timing/model': 1.5566555634140968, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 0.31}) +[2026-06-15 21:11:51,506][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.15135878324508667, 'train/grad_norm_pre_clip': 10.643597602844238, 'timing/data': 0.02768649533390999, 'timing/model': 1.5321469232439995, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 0.31}) +[2026-06-15 21:11:57,620][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.20961782336235046, 'train/grad_norm_pre_clip': 4.8246169090271, 'timing/data': 0.013789379969239235, 'timing/model': 1.541225392371416, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 0.31}) +[2026-06-15 21:12:03,901][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.09765429049730301, 'train/grad_norm_pre_clip': 7.0736870765686035, 'timing/data': 0.017205674201250076, 'timing/model': 1.7930133901536465, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 0.31}) +[2026-06-15 21:12:09,841][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.3210364878177643, 'train/grad_norm_pre_clip': 16.202268600463867, 'timing/data': 0.009897971525788307, 'timing/model': 1.6209203004837036, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 0.31}) +[2026-06-15 21:12:15,471][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.07509345561265945, 'train/grad_norm_pre_clip': 3.5972282886505127, 'timing/data': 0.011102071031928062, 'timing/model': 1.4377211034297943, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 0.31}) +[2026-06-15 21:12:21,377][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.12968751788139343, 'train/grad_norm_pre_clip': 10.021777153015137, 'timing/data': 0.010727083310484886, 'timing/model': 1.4533127322793007, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 0.31}) +[2026-06-15 21:12:27,277][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.14327757060527802, 'train/grad_norm_pre_clip': 9.402259826660156, 'timing/data': 0.010762516409158707, 'timing/model': 1.4612325876951218, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 0.31}) +[2026-06-15 21:12:33,182][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.32007840275764465, 'train/grad_norm_pre_clip': 14.411864280700684, 'timing/data': 0.01124933548271656, 'timing/model': 1.4945373944938183, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 0.31}) +[2026-06-15 21:12:39,080][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.16736149787902832, 'train/grad_norm_pre_clip': 10.316354751586914, 'timing/data': 0.009584162384271622, 'timing/model': 1.4740263409912586, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 0.31}) +[2026-06-15 21:12:45,217][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.15721789002418518, 'train/grad_norm_pre_clip': 10.518630027770996, 'timing/data': 0.011327922344207764, 'timing/model': 1.7387171052396297, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 0.31}) +[2026-06-15 21:12:51,039][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.19049306213855743, 'train/grad_norm_pre_clip': 13.148808479309082, 'timing/data': 0.011847976595163345, 'timing/model': 1.4691387061029673, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 0.31}) +[2026-06-15 21:12:57,091][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.20334279537200928, 'train/grad_norm_pre_clip': 9.082793235778809, 'timing/data': 0.011857600882649422, 'timing/model': 1.513187501579523, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 0.32}) +[2026-06-15 21:13:03,247][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.16961884498596191, 'train/grad_norm_pre_clip': 7.979831218719482, 'timing/data': 0.012436715885996819, 'timing/model': 1.4763979148119688, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 0.32}) +[2026-06-15 21:13:09,333][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.18206430971622467, 'train/grad_norm_pre_clip': 8.527828216552734, 'timing/data': 0.01249694637954235, 'timing/model': 1.5355869513005018, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 0.32}) +[2026-06-15 21:13:15,417][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.1249060332775116, 'train/grad_norm_pre_clip': 9.16395378112793, 'timing/data': 0.012808967381715775, 'timing/model': 1.738315410912037, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 0.32}) +[2026-06-15 21:13:21,251][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.16506977379322052, 'train/grad_norm_pre_clip': 6.460197448730469, 'timing/data': 0.013274457305669785, 'timing/model': 1.4755013976246119, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 0.32}) +[2026-06-15 21:13:27,400][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.1584256887435913, 'train/grad_norm_pre_clip': 11.05677318572998, 'timing/data': 0.01214616373181343, 'timing/model': 1.5241648256778717, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 0.32}) +[2026-06-15 21:13:33,351][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.23889082670211792, 'train/grad_norm_pre_clip': 13.285223007202148, 'timing/data': 0.011784430593252182, 'timing/model': 1.4791167043149471, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 0.32}) +[2026-06-15 21:13:39,350][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.08507572114467621, 'train/grad_norm_pre_clip': 4.4609785079956055, 'timing/data': 0.012736508622765541, 'timing/model': 1.480686942115426, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 0.32}) +[2026-06-15 21:13:45,337][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.06866366416215897, 'train/grad_norm_pre_clip': 10.712331771850586, 'timing/data': 0.01207059621810913, 'timing/model': 1.4888910856097937, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 0.32}) +[2026-06-15 21:13:51,412][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.11067716777324677, 'train/grad_norm_pre_clip': 6.230166435241699, 'timing/data': 0.011576032266020775, 'timing/model': 1.7379846666008234, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 0.32}) +[2026-06-15 21:13:57,263][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.1689995527267456, 'train/grad_norm_pre_clip': 11.265186309814453, 'timing/data': 0.013417916372418404, 'timing/model': 1.4884925913065672, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 0.32}) +[2026-06-15 21:14:03,312][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.23133856058120728, 'train/grad_norm_pre_clip': 7.4160966873168945, 'timing/data': 0.011927859857678413, 'timing/model': 1.5337544716894627, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 0.32}) +[2026-06-15 21:14:09,278][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.10964033007621765, 'train/grad_norm_pre_clip': 5.652531147003174, 'timing/data': 0.010828794911503792, 'timing/model': 1.4638941511511803, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 0.32}) +[2026-06-15 21:14:15,079][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.23423364758491516, 'train/grad_norm_pre_clip': 11.102234840393066, 'timing/data': 0.009994691237807274, 'timing/model': 1.46002596616745, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 0.32}) +[2026-06-15 21:14:21,281][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.2529439330101013, 'train/grad_norm_pre_clip': 11.760683059692383, 'timing/data': 0.0107541773468256, 'timing/model': 1.6620893143117428, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 0.32}) +[2026-06-15 21:14:27,818][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.14585615694522858, 'train/grad_norm_pre_clip': 4.616509437561035, 'timing/data': 0.015066120773553848, 'timing/model': 1.7598962616175413, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 0.32}) +[2026-06-15 21:14:34,108][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.405121386051178, 'train/grad_norm_pre_clip': 6.476192474365234, 'timing/data': 0.019124696031212807, 'timing/model': 1.6356120128184557, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 0.32}) +[2026-06-15 21:14:40,808][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.15155470371246338, 'train/grad_norm_pre_clip': 5.983487606048584, 'timing/data': 0.017592042684555054, 'timing/model': 1.6426241397857666, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 0.32}) +[2026-06-15 21:14:47,498][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.13046647608280182, 'train/grad_norm_pre_clip': 7.662569522857666, 'timing/data': 0.014612527564167976, 'timing/model': 1.5936559233814478, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 0.32}) +[2026-06-15 21:14:54,097][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.27687638998031616, 'train/grad_norm_pre_clip': 13.052346229553223, 'timing/data': 0.018948927521705627, 'timing/model': 1.5830297823995352, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 0.32}) +[2026-06-15 21:15:00,456][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.09154780209064484, 'train/grad_norm_pre_clip': 5.666027069091797, 'timing/data': 0.013512205332517624, 'timing/model': 1.6417896263301373, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 0.32}) +[2026-06-15 21:15:06,927][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.14927570521831512, 'train/grad_norm_pre_clip': 6.339055061340332, 'timing/data': 0.013493508100509644, 'timing/model': 1.9853278305381536, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.32}) +[2026-06-15 21:15:12,907][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.1404278576374054, 'train/grad_norm_pre_clip': 4.949909687042236, 'timing/data': 0.012412330135703087, 'timing/model': 1.5034342370927334, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 0.32}) +[2026-06-15 21:15:18,800][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.14513373374938965, 'train/grad_norm_pre_clip': 9.012604713439941, 'timing/data': 0.010897673666477203, 'timing/model': 1.449895391240716, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 0.32}) +[2026-06-15 21:15:24,677][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.3409740924835205, 'train/grad_norm_pre_clip': 13.990921974182129, 'timing/data': 0.011472566053271294, 'timing/model': 1.418742449954152, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 0.32}) +[2026-06-15 21:15:30,481][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.08127537369728088, 'train/grad_norm_pre_clip': 5.668440341949463, 'timing/data': 0.010388456284999847, 'timing/model': 1.4619591571390629, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 0.32}) +[2026-06-15 21:15:36,373][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.13945752382278442, 'train/grad_norm_pre_clip': 4.870945453643799, 'timing/data': 0.010454092174768448, 'timing/model': 1.4225252214819193, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.33}) +[2026-06-15 21:15:42,303][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.27691832184791565, 'train/grad_norm_pre_clip': 13.968795776367188, 'timing/data': 0.010498318821191788, 'timing/model': 1.6713544055819511, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 0.33}) +[2026-06-15 21:15:47,986][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.0761188343167305, 'train/grad_norm_pre_clip': 4.869462490081787, 'timing/data': 0.010372322052717209, 'timing/model': 1.448176559060812, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 0.33}) +[2026-06-15 21:15:53,881][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.06949926167726517, 'train/grad_norm_pre_clip': 5.161500930786133, 'timing/data': 0.010249055922031403, 'timing/model': 1.451884776353836, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 0.33}) +[2026-06-15 21:15:59,677][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.38675981760025024, 'train/grad_norm_pre_clip': 12.664570808410645, 'timing/data': 0.01122107170522213, 'timing/model': 1.4537325333803892, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 0.33}) +[2026-06-15 21:16:05,482][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.09576564282178879, 'train/grad_norm_pre_clip': 9.08498764038086, 'timing/data': 0.010395599529147148, 'timing/model': 1.4589104764163494, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 0.33}) +[2026-06-15 21:16:11,375][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.18606798350811005, 'train/grad_norm_pre_clip': 11.96535873413086, 'timing/data': 0.010455138981342316, 'timing/model': 1.431624261662364, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 0.33}) +[2026-06-15 21:16:17,275][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.24471595883369446, 'train/grad_norm_pre_clip': 13.881566047668457, 'timing/data': 0.010183580219745636, 'timing/model': 1.6556302029639482, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 0.33}) +[2026-06-15 21:16:23,016][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.1454164683818817, 'train/grad_norm_pre_clip': 8.14010238647461, 'timing/data': 0.01088634878396988, 'timing/model': 1.4959361273795366, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 0.33}) +[2026-06-15 21:16:28,875][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.09121942520141602, 'train/grad_norm_pre_clip': 4.569103717803955, 'timing/data': 0.011141035705804825, 'timing/model': 1.456791816279292, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 0.33}) +[2026-06-15 21:16:34,772][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.12631303071975708, 'train/grad_norm_pre_clip': 18.816944122314453, 'timing/data': 0.011423183605074883, 'timing/model': 1.4518918637186289, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.33}) +[2026-06-15 21:16:40,728][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.10384385287761688, 'train/grad_norm_pre_clip': 8.565752983093262, 'timing/data': 0.013676736503839493, 'timing/model': 1.5281682927161455, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 0.33}) +[2026-06-15 21:16:46,834][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.10843633115291595, 'train/grad_norm_pre_clip': 9.95818042755127, 'timing/data': 0.012132694944739342, 'timing/model': 1.4703707043081522, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 0.33}) +[2026-06-15 21:16:52,976][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.09623777866363525, 'train/grad_norm_pre_clip': 6.03989315032959, 'timing/data': 0.012526044622063637, 'timing/model': 1.7779957596212626, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 0.33}) +[2026-06-15 21:16:58,920][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.16724690794944763, 'train/grad_norm_pre_clip': 12.896269798278809, 'timing/data': 0.012500409036874771, 'timing/model': 1.516912043094635, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 0.33}) +[2026-06-15 21:17:05,033][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.24209657311439514, 'train/grad_norm_pre_clip': 11.990412712097168, 'timing/data': 0.011779909953474998, 'timing/model': 1.5413774903863668, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 0.33}) +[2026-06-15 21:17:11,038][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.47181016206741333, 'train/grad_norm_pre_clip': 23.64549446105957, 'timing/data': 0.012254191562533379, 'timing/model': 1.5320850759744644, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 0.33}) +[2026-06-15 21:17:17,138][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.1832592785358429, 'train/grad_norm_pre_clip': 13.429825782775879, 'timing/data': 0.011384239420294762, 'timing/model': 1.5365403797477484, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 0.33}) +[2026-06-15 21:17:23,240][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.08529748022556305, 'train/grad_norm_pre_clip': 9.677621841430664, 'timing/data': 0.011604299768805504, 'timing/model': 1.54793306812644, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 0.33}) +[2026-06-15 21:17:29,278][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.28443634510040283, 'train/grad_norm_pre_clip': 13.40034008026123, 'timing/data': 0.010053599253296852, 'timing/model': 1.6879574432969093, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 0.33}) +[2026-06-15 21:17:35,402][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.49872297048568726, 'train/grad_norm_pre_clip': 22.409990310668945, 'timing/data': 0.010885057970881462, 'timing/model': 1.8771228175610304, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.33}) +[2026-06-15 21:17:41,329][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.20571745932102203, 'train/grad_norm_pre_clip': 9.664970397949219, 'timing/data': 0.012457147240638733, 'timing/model': 1.5386025328189135, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 0.33}) +[2026-06-15 21:17:47,448][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.17155097424983978, 'train/grad_norm_pre_clip': 10.898725509643555, 'timing/data': 0.012904446572065353, 'timing/model': 1.545803751796484, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 0.33}) +[2026-06-15 21:17:53,567][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.13816283643245697, 'train/grad_norm_pre_clip': 7.906748294830322, 'timing/data': 0.011873485520482063, 'timing/model': 1.4868387673050165, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 0.33}) +[2026-06-15 21:17:59,944][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.18332292139530182, 'train/grad_norm_pre_clip': 9.403039932250977, 'timing/data': 0.013146670535206795, 'timing/model': 1.5397503357380629, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 0.33}) +[2026-06-15 21:18:06,099][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.11755787581205368, 'train/grad_norm_pre_clip': 5.490954399108887, 'timing/data': 0.012073332443833351, 'timing/model': 1.7967832889407873, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 0.33}) +[2026-06-15 21:18:12,129][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.28600046038627625, 'train/grad_norm_pre_clip': 15.760616302490234, 'timing/data': 0.013049617409706116, 'timing/model': 1.515802575275302, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 0.34}) +[2026-06-15 21:18:18,344][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.1649245172739029, 'train/grad_norm_pre_clip': 8.311843872070312, 'timing/data': 0.01346401497721672, 'timing/model': 1.5478602163493633, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 0.34}) +[2026-06-15 21:18:24,541][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.24546894431114197, 'train/grad_norm_pre_clip': 7.305130958557129, 'timing/data': 0.012584583833813667, 'timing/model': 1.541590815410018, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 0.34}) +[2026-06-15 21:18:30,647][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.1741919219493866, 'train/grad_norm_pre_clip': 9.340925216674805, 'timing/data': 0.012052102014422417, 'timing/model': 1.476039981469512, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 0.34}) +[2026-06-15 21:18:36,726][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.17000648379325867, 'train/grad_norm_pre_clip': 4.063102722167969, 'timing/data': 0.012717343866825104, 'timing/model': 1.5268380381166935, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.34}) +[2026-06-15 21:18:42,854][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.2220514565706253, 'train/grad_norm_pre_clip': 10.570755004882812, 'timing/data': 0.012526484206318855, 'timing/model': 1.7451337650418282, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 0.34}) +[2026-06-15 21:18:48,858][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.13538654148578644, 'train/grad_norm_pre_clip': 7.327899932861328, 'timing/data': 0.012407386675477028, 'timing/model': 1.5528527926653624, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 0.34}) +[2026-06-15 21:18:55,059][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.25907087326049805, 'train/grad_norm_pre_clip': 7.184938907623291, 'timing/data': 0.013152128085494041, 'timing/model': 1.550570908933878, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 0.34}) +[2026-06-15 21:19:01,258][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.142319917678833, 'train/grad_norm_pre_clip': 6.530084609985352, 'timing/data': 0.013782422989606857, 'timing/model': 1.5246688593178988, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 0.34}) +[2026-06-15 21:19:07,338][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.2707599401473999, 'train/grad_norm_pre_clip': 7.304408073425293, 'timing/data': 0.011056609451770782, 'timing/model': 1.4598369114100933, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 0.34}) +[2026-06-15 21:19:13,221][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.1608610600233078, 'train/grad_norm_pre_clip': 14.319985389709473, 'timing/data': 0.011284511536359787, 'timing/model': 1.45291218906641, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 0.34}) +[2026-06-15 21:19:19,250][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.20492222905158997, 'train/grad_norm_pre_clip': 8.369176864624023, 'timing/data': 0.01068590022623539, 'timing/model': 1.8273916486650705, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 0.34}) +[2026-06-15 21:19:24,885][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.07265538722276688, 'train/grad_norm_pre_clip': 2.677262544631958, 'timing/data': 0.010752325877547264, 'timing/model': 1.4535590820014477, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 0.34}) +[2026-06-15 21:19:30,770][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.06058407202363014, 'train/grad_norm_pre_clip': 2.21103835105896, 'timing/data': 0.011450054123997688, 'timing/model': 1.4503304157406092, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 0.34}) +[2026-06-15 21:19:36,681][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.28603681921958923, 'train/grad_norm_pre_clip': 9.494089126586914, 'timing/data': 0.010426001623272896, 'timing/model': 1.4561464916914701, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 0.34}) +[2026-06-15 21:19:42,582][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.0409274585545063, 'train/grad_norm_pre_clip': 2.7923855781555176, 'timing/data': 0.011162880808115005, 'timing/model': 1.4439632520079613, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 0.34}) +[2026-06-15 21:19:48,484][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.05920930951833725, 'train/grad_norm_pre_clip': 4.286746025085449, 'timing/data': 0.010763904079794884, 'timing/model': 1.6548040453344584, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 0.34}) +[2026-06-15 21:19:54,220][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.05758120119571686, 'train/grad_norm_pre_clip': 2.8035776615142822, 'timing/data': 0.009711261838674545, 'timing/model': 1.4571006372570992, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 0.34}) +[2026-06-15 21:20:00,225][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.12751789391040802, 'train/grad_norm_pre_clip': 9.929491996765137, 'timing/data': 0.01060870848596096, 'timing/model': 1.459760520607233, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.34}) +[2026-06-15 21:20:06,009][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.2484470009803772, 'train/grad_norm_pre_clip': 15.480535507202148, 'timing/data': 0.010461151599884033, 'timing/model': 1.4521738328039646, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 0.34}) +[2026-06-15 21:20:12,265][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.14213809370994568, 'train/grad_norm_pre_clip': 13.19992733001709, 'timing/data': 0.012576334178447723, 'timing/model': 1.545592064037919, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 0.34}) +[2026-06-15 21:20:18,567][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.1310749650001526, 'train/grad_norm_pre_clip': 9.04447078704834, 'timing/data': 0.011397412046790123, 'timing/model': 1.5513432566076517, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 0.34}) +[2026-06-15 21:20:25,333][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.3502354621887207, 'train/grad_norm_pre_clip': 17.90115737915039, 'timing/data': 0.01720176450908184, 'timing/model': 1.9655966013669968, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 0.34}) +[2026-06-15 21:20:31,380][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.23205146193504333, 'train/grad_norm_pre_clip': 17.92426109313965, 'timing/data': 0.01654326356947422, 'timing/model': 1.5402824524790049, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 0.34}) +[2026-06-15 21:20:38,062][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.09832137078046799, 'train/grad_norm_pre_clip': 7.102894306182861, 'timing/data': 0.01538437232375145, 'timing/model': 1.6143594197928905, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 0.34}) +[2026-06-15 21:20:44,767][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.033396296203136444, 'train/grad_norm_pre_clip': 4.725578784942627, 'timing/data': 0.013635808601975441, 'timing/model': 1.5464813392609358, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 0.34}) +[2026-06-15 21:20:51,243][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.38558048009872437, 'train/grad_norm_pre_clip': 25.4382381439209, 'timing/data': 0.01728907972574234, 'timing/model': 1.6134685035794973, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 0.35}) +[2026-06-15 21:20:57,856][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.21667735278606415, 'train/grad_norm_pre_clip': 20.844276428222656, 'timing/data': 0.014357216656208038, 'timing/model': 1.5480528865009546, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 0.35}) +[2026-06-15 21:21:04,194][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.29992371797561646, 'train/grad_norm_pre_clip': 23.77528953552246, 'timing/data': 0.015431845560669899, 'timing/model': 1.874372873455286, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 0.35}) +[2026-06-15 21:21:10,521][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.4017334282398224, 'train/grad_norm_pre_clip': 13.818918228149414, 'timing/data': 0.01671488769352436, 'timing/model': 1.641083288937807, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 0.35}) +[2026-06-15 21:21:17,417][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.06757953017950058, 'train/grad_norm_pre_clip': 5.0774712562561035, 'timing/data': 0.024204764515161514, 'timing/model': 1.7334897257387638, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 0.35}) +[2026-06-15 21:21:23,925][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.31373512744903564, 'train/grad_norm_pre_clip': 14.037302017211914, 'timing/data': 0.01224382221698761, 'timing/model': 1.5316555425524712, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 0.35}) +[2026-06-15 21:21:30,248][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.08069737255573273, 'train/grad_norm_pre_clip': 5.110514163970947, 'timing/data': 0.019541019573807716, 'timing/model': 1.6380000989884138, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 0.35}) +[2026-06-15 21:21:36,138][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.09588652104139328, 'train/grad_norm_pre_clip': 4.543095588684082, 'timing/data': 0.01545015349984169, 'timing/model': 1.479451086372137, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 0.35}) +[2026-06-15 21:29:12,810][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.39973706007003784, 'train/grad_norm_pre_clip': 30.816421508789062, 'eval/loss': 0.1492151916027069, 'eval/flap_precision': 0.9519071310116086, 'eval/flap_recall': 0.7788331071913162, 'eval/flap_f1': 0.8567164179104477, 'eval/flap_rate_model': 0.1675, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8567164179104478, 'timing/data': 0.012820228934288025, 'timing/model': 1.7692960388958454, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 0.35}) +[2026-06-15 21:29:12,811][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_900_state +[2026-06-15 21:29:12,815][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 21:29:18,734][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_900_state/model.safetensors +[2026-06-15 21:29:32,506][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_900_state/optimizer.bin +[2026-06-15 21:29:32,507][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_900_state/sampler.bin +[2026-06-15 21:29:32,507][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_900_state/sampler_1.bin +[2026-06-15 21:29:32,508][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_900_state/random_states_0.pkl +[2026-06-15 21:29:32,613][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-06-15 21:29:32,614][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 21:29:32,623][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 21:29:38,567][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.09559861570596695, 'train/grad_norm_pre_clip': 6.812185764312744, 'timing/data': 0.012802887707948685, 'timing/model': 1.7012785356491804, 'learning_rate/qwen_vl_interface': 9.387268552285024e-06, 'learning_rate/action_model': 9.387268552285025e-05, 'epoch': 0.35}) +[2026-06-15 21:29:44,862][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.16518621146678925, 'train/grad_norm_pre_clip': 15.407369613647461, 'timing/data': 0.013105280697345734, 'timing/model': 1.8769856411963701, 'learning_rate/qwen_vl_interface': 9.385771559309125e-06, 'learning_rate/action_model': 9.385771559309125e-05, 'epoch': 0.35}) +[2026-06-15 21:29:51,229][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.3559071719646454, 'train/grad_norm_pre_clip': 26.841848373413086, 'timing/data': 0.012499768286943436, 'timing/model': 1.627228058874607, 'learning_rate/qwen_vl_interface': 9.38427286627376e-06, 'learning_rate/action_model': 9.38427286627376e-05, 'epoch': 0.35}) +[2026-06-15 21:29:57,565][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.323954313993454, 'train/grad_norm_pre_clip': 15.410510063171387, 'timing/data': 0.012779032811522484, 'timing/model': 1.5630995035171509, 'learning_rate/qwen_vl_interface': 9.382772473794985e-06, 'learning_rate/action_model': 9.382772473794985e-05, 'epoch': 0.35}) +[2026-06-15 21:30:03,858][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.18053193390369415, 'train/grad_norm_pre_clip': 10.221189498901367, 'timing/data': 0.013624686747789383, 'timing/model': 1.5476174261420965, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.35}) +[2026-06-15 21:30:10,158][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.20396187901496887, 'train/grad_norm_pre_clip': 13.461851119995117, 'timing/data': 0.029514912515878677, 'timing/model': 1.531727571040392, 'learning_rate/qwen_vl_interface': 9.37976659297492e-06, 'learning_rate/action_model': 9.379766592974921e-05, 'epoch': 0.35}) +[2026-06-15 21:30:16,463][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.35323265194892883, 'train/grad_norm_pre_clip': 12.601702690124512, 'timing/data': 0.014749711379408836, 'timing/model': 1.5345344003289938, 'learning_rate/qwen_vl_interface': 9.378261105869237e-06, 'learning_rate/action_model': 9.378261105869237e-05, 'epoch': 0.35}) +[2026-06-15 21:30:22,800][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.11142374575138092, 'train/grad_norm_pre_clip': 4.685494899749756, 'timing/data': 0.013832034543156624, 'timing/model': 1.7945117801427841, 'learning_rate/qwen_vl_interface': 9.376753921791348e-06, 'learning_rate/action_model': 9.376753921791348e-05, 'epoch': 0.35}) +[2026-06-15 21:30:28,859][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.38810431957244873, 'train/grad_norm_pre_clip': 11.036473274230957, 'timing/data': 0.014099761843681335, 'timing/model': 1.556227095425129, 'learning_rate/qwen_vl_interface': 9.375245041360805e-06, 'learning_rate/action_model': 9.375245041360804e-05, 'epoch': 0.35}) +[2026-06-15 21:30:35,159][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.26537448167800903, 'train/grad_norm_pre_clip': 12.94902229309082, 'timing/data': 0.013886913657188416, 'timing/model': 1.543523008003831, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 0.35}) +[2026-06-15 21:30:41,472][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.11195942014455795, 'train/grad_norm_pre_clip': 3.964942455291748, 'timing/data': 0.012812471017241478, 'timing/model': 1.5674235504120588, 'learning_rate/qwen_vl_interface': 9.372222193923421e-06, 'learning_rate/action_model': 9.372222193923421e-05, 'epoch': 0.35}) +[2026-06-15 21:30:47,761][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.1897694617509842, 'train/grad_norm_pre_clip': 6.8050971031188965, 'timing/data': 0.011792052537202835, 'timing/model': 1.5561053715646267, 'learning_rate/qwen_vl_interface': 9.370708228159158e-06, 'learning_rate/action_model': 9.370708228159159e-05, 'epoch': 0.35}) +[2026-06-15 21:30:53,965][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.20921632647514343, 'train/grad_norm_pre_clip': 7.972151279449463, 'timing/data': 0.013111904263496399, 'timing/model': 1.5179538298398256, 'learning_rate/qwen_vl_interface': 9.369192568527398e-06, 'learning_rate/action_model': 9.369192568527398e-05, 'epoch': 0.35}) +[2026-06-15 21:31:00,321][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.1492530107498169, 'train/grad_norm_pre_clip': 6.1676249504089355, 'timing/data': 0.012743990868330002, 'timing/model': 1.8143811039626598, 'learning_rate/qwen_vl_interface': 9.367675215651168e-06, 'learning_rate/action_model': 9.367675215651168e-05, 'epoch': 0.35}) +[2026-06-15 21:31:06,365][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.1545821726322174, 'train/grad_norm_pre_clip': 4.191322326660156, 'timing/data': 0.012749329209327698, 'timing/model': 1.5560720935463905, 'learning_rate/qwen_vl_interface': 9.366156170154194e-06, 'learning_rate/action_model': 9.366156170154194e-05, 'epoch': 0.35}) +[2026-06-15 21:31:12,665][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.10657542198896408, 'train/grad_norm_pre_clip': 6.786237716674805, 'timing/data': 0.01313391886651516, 'timing/model': 1.5552878770977259, 'learning_rate/qwen_vl_interface': 9.3646354326609e-06, 'learning_rate/action_model': 9.364635432660899e-05, 'epoch': 0.35}) +[2026-06-15 21:31:18,973][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.17770667374134064, 'train/grad_norm_pre_clip': 5.886345386505127, 'timing/data': 0.013037638738751411, 'timing/model': 1.561924945563078, 'learning_rate/qwen_vl_interface': 9.3631130037964e-06, 'learning_rate/action_model': 9.363113003796401e-05, 'epoch': 0.35}) +[2026-06-15 21:31:25,266][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.06046301871538162, 'train/grad_norm_pre_clip': 2.7412216663360596, 'timing/data': 0.013181403279304504, 'timing/model': 1.5586095731705427, 'learning_rate/qwen_vl_interface': 9.361588884186513e-06, 'learning_rate/action_model': 9.361588884186513e-05, 'epoch': 0.36}) +[2026-06-15 21:31:31,561][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.25515905022621155, 'train/grad_norm_pre_clip': 11.18206787109375, 'timing/data': 0.012579765170812607, 'timing/model': 1.511954439803958, 'learning_rate/qwen_vl_interface': 9.360063074457744e-06, 'learning_rate/action_model': 9.360063074457743e-05, 'epoch': 0.36}) +[2026-06-15 21:31:37,899][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.4231579303741455, 'train/grad_norm_pre_clip': 13.424073219299316, 'timing/data': 0.011624550446867943, 'timing/model': 1.7903686314821243, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 0.36}) +[2026-06-15 21:31:43,969][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.1759784072637558, 'train/grad_norm_pre_clip': 10.94747543334961, 'timing/data': 0.012525195255875587, 'timing/model': 1.5571663323789835, 'learning_rate/qwen_vl_interface': 9.357006387153065e-06, 'learning_rate/action_model': 9.357006387153065e-05, 'epoch': 0.36}) +[2026-06-15 21:31:50,268][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.07383634150028229, 'train/grad_norm_pre_clip': 5.7444257736206055, 'timing/data': 0.012940552085638046, 'timing/model': 1.5552179142832756, 'learning_rate/qwen_vl_interface': 9.355475510833643e-06, 'learning_rate/action_model': 9.355475510833643e-05, 'epoch': 0.36}) +[2026-06-15 21:31:56,669][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.17142905294895172, 'train/grad_norm_pre_clip': 8.047947883605957, 'timing/data': 0.013116158545017242, 'timing/model': 1.657484708353877, 'learning_rate/qwen_vl_interface': 9.353942946908318e-06, 'learning_rate/action_model': 9.353942946908317e-05, 'epoch': 0.36}) +[2026-06-15 21:32:02,974][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.020413538441061974, 'train/grad_norm_pre_clip': 1.946776032447815, 'timing/data': 0.013860329985618591, 'timing/model': 1.5518095716834068, 'learning_rate/qwen_vl_interface': 9.352408696007065e-06, 'learning_rate/action_model': 9.352408696007065e-05, 'epoch': 0.36}) +[2026-06-15 21:32:09,485][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.10888925939798355, 'train/grad_norm_pre_clip': 6.802077293395996, 'timing/data': 0.01594018004834652, 'timing/model': 1.7218161914497614, 'learning_rate/qwen_vl_interface': 9.350872758760558e-06, 'learning_rate/action_model': 9.350872758760559e-05, 'epoch': 0.36}) +[2026-06-15 21:32:15,829][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.10870310664176941, 'train/grad_norm_pre_clip': 7.719545364379883, 'timing/data': 0.013042392209172249, 'timing/model': 1.8162423092871904, 'learning_rate/qwen_vl_interface': 9.349335135800164e-06, 'learning_rate/action_model': 9.349335135800163e-05, 'epoch': 0.36}) +[2026-06-15 21:32:22,478][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.0816834568977356, 'train/grad_norm_pre_clip': 9.072654724121094, 'timing/data': 0.013215996325016022, 'timing/model': 1.889913471415639, 'learning_rate/qwen_vl_interface': 9.347795827757942e-06, 'learning_rate/action_model': 9.347795827757941e-05, 'epoch': 0.36}) +[2026-06-15 21:32:28,811][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.38110843300819397, 'train/grad_norm_pre_clip': 13.57249641418457, 'timing/data': 0.0183455441147089, 'timing/model': 1.6326872631907463, 'learning_rate/qwen_vl_interface': 9.346254835266643e-06, 'learning_rate/action_model': 9.346254835266642e-05, 'epoch': 0.36}) +[2026-06-15 21:32:35,066][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.2351088523864746, 'train/grad_norm_pre_clip': 14.967276573181152, 'timing/data': 0.012724978849291801, 'timing/model': 1.554112497717142, 'learning_rate/qwen_vl_interface': 9.344712158959709e-06, 'learning_rate/action_model': 9.344712158959709e-05, 'epoch': 0.36}) +[2026-06-15 21:32:41,364][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.324769526720047, 'train/grad_norm_pre_clip': 18.35906982421875, 'timing/data': 0.012695390731096268, 'timing/model': 1.5532460771501064, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 0.36}) +[2026-06-15 21:32:47,665][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.14897575974464417, 'train/grad_norm_pre_clip': 11.373285293579102, 'timing/data': 0.012336574494838715, 'timing/model': 1.5185199081897736, 'learning_rate/qwen_vl_interface': 9.341621757436177e-06, 'learning_rate/action_model': 9.341621757436177e-05, 'epoch': 0.36}) +[2026-06-15 21:32:54,054][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.09304627776145935, 'train/grad_norm_pre_clip': 6.084325313568115, 'timing/data': 0.012054473161697388, 'timing/model': 1.8186258729547262, 'learning_rate/qwen_vl_interface': 9.340074033489924e-06, 'learning_rate/action_model': 9.340074033489923e-05, 'epoch': 0.36}) +[2026-06-15 21:33:00,073][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.05959697440266609, 'train/grad_norm_pre_clip': 6.5912556648254395, 'timing/data': 0.01235789991915226, 'timing/model': 1.5515130832791328, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.36}) +[2026-06-15 21:33:06,373][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.12530116736888885, 'train/grad_norm_pre_clip': 8.898909568786621, 'timing/data': 0.011752240359783173, 'timing/model': 1.5353754051029682, 'learning_rate/qwen_vl_interface': 9.336973542409505e-06, 'learning_rate/action_model': 9.336973542409504e-05, 'epoch': 0.36}) +[2026-06-15 21:33:12,713][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.20674680173397064, 'train/grad_norm_pre_clip': 23.701799392700195, 'timing/data': 0.01691613160073757, 'timing/model': 1.54951729811728, 'learning_rate/qwen_vl_interface': 9.33542077654983e-06, 'learning_rate/action_model': 9.33542077654983e-05, 'epoch': 0.36}) +[2026-06-15 21:33:19,049][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.2945866286754608, 'train/grad_norm_pre_clip': 18.88377571105957, 'timing/data': 0.011693399399518967, 'timing/model': 1.5718092396855354, 'learning_rate/qwen_vl_interface': 9.333866331327995e-06, 'learning_rate/action_model': 9.333866331327995e-05, 'epoch': 0.36}) +[2026-06-15 21:33:25,047][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.2409074902534485, 'train/grad_norm_pre_clip': 15.765600204467773, 'timing/data': 0.013018999248743057, 'timing/model': 1.4957445282489061, 'learning_rate/qwen_vl_interface': 9.332310207382972e-06, 'learning_rate/action_model': 9.332310207382972e-05, 'epoch': 0.36}) +[2026-06-15 21:33:31,222][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.41874784231185913, 'train/grad_norm_pre_clip': 38.065895080566406, 'timing/data': 0.011855065822601318, 'timing/model': 1.7394290920346975, 'learning_rate/qwen_vl_interface': 9.330752405354426e-06, 'learning_rate/action_model': 9.330752405354426e-05, 'epoch': 0.36}) +[2026-06-15 21:33:37,171][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.08741272240877151, 'train/grad_norm_pre_clip': 7.365462303161621, 'timing/data': 0.01203136332333088, 'timing/model': 1.579092500731349, 'learning_rate/qwen_vl_interface': 9.32919292588271e-06, 'learning_rate/action_model': 9.32919292588271e-05, 'epoch': 0.36}) +[2026-06-15 21:33:43,582][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.597327470779419, 'train/grad_norm_pre_clip': 30.620731353759766, 'timing/data': 0.013606121763586998, 'timing/model': 1.5581735912710428, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.36}) +[2026-06-15 21:33:49,907][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.20746317505836487, 'train/grad_norm_pre_clip': 8.513435363769531, 'timing/data': 0.014502309262752533, 'timing/model': 1.55678048171103, 'learning_rate/qwen_vl_interface': 9.326068937174627e-06, 'learning_rate/action_model': 9.326068937174626e-05, 'epoch': 0.36}) +[2026-06-15 21:33:56,284][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.2926700711250305, 'train/grad_norm_pre_clip': 6.099906921386719, 'timing/data': 0.012093840166926384, 'timing/model': 1.5830099117010832, 'learning_rate/qwen_vl_interface': 9.324504429222412e-06, 'learning_rate/action_model': 9.324504429222412e-05, 'epoch': 0.36}) +[2026-06-15 21:34:02,571][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.14945288002490997, 'train/grad_norm_pre_clip': 7.515312194824219, 'timing/data': 0.012196477502584457, 'timing/model': 1.52968336828053, 'learning_rate/qwen_vl_interface': 9.322938246395336e-06, 'learning_rate/action_model': 9.322938246395337e-05, 'epoch': 0.36}) +[2026-06-15 21:34:08,890][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.2249940186738968, 'train/grad_norm_pre_clip': 11.10307788848877, 'timing/data': 0.013145811855793, 'timing/model': 1.7517046891152859, 'learning_rate/qwen_vl_interface': 9.321370389337196e-06, 'learning_rate/action_model': 9.321370389337196e-05, 'epoch': 0.37}) +[2026-06-15 21:34:14,933][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.3707005977630615, 'train/grad_norm_pre_clip': 11.327436447143555, 'timing/data': 0.012274278327822685, 'timing/model': 1.5343189183622599, 'learning_rate/qwen_vl_interface': 9.319800858692478e-06, 'learning_rate/action_model': 9.319800858692476e-05, 'epoch': 0.37}) +[2026-06-15 21:34:21,067][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.28279194235801697, 'train/grad_norm_pre_clip': 8.955784797668457, 'timing/data': 0.01259358786046505, 'timing/model': 1.5400552041828632, 'learning_rate/qwen_vl_interface': 9.318229655106357e-06, 'learning_rate/action_model': 9.318229655106356e-05, 'epoch': 0.37}) +[2026-06-15 21:34:27,365][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.3635857105255127, 'train/grad_norm_pre_clip': 12.038575172424316, 'timing/data': 0.013174263760447502, 'timing/model': 1.565610771998763, 'learning_rate/qwen_vl_interface': 9.316656779224693e-06, 'learning_rate/action_model': 9.316656779224693e-05, 'epoch': 0.37}) +[2026-06-15 21:34:33,573][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.20121972262859344, 'train/grad_norm_pre_clip': 5.235189914703369, 'timing/data': 0.013016702607274055, 'timing/model': 1.50701997987926, 'learning_rate/qwen_vl_interface': 9.31508223169404e-06, 'learning_rate/action_model': 9.31508223169404e-05, 'epoch': 0.37}) +[2026-06-15 21:34:39,807][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.2702943980693817, 'train/grad_norm_pre_clip': 7.091606140136719, 'timing/data': 0.013045098632574081, 'timing/model': 1.7822811100631952, 'learning_rate/qwen_vl_interface': 9.313506013161633e-06, 'learning_rate/action_model': 9.313506013161633e-05, 'epoch': 0.37}) +[2026-06-15 21:34:45,765][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.3685271441936493, 'train/grad_norm_pre_clip': 14.776504516601562, 'timing/data': 0.013101551681756973, 'timing/model': 1.522296566516161, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 0.37}) +[2026-06-15 21:34:52,033][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.21071429550647736, 'train/grad_norm_pre_clip': 6.265281677246094, 'timing/data': 0.01258731260895729, 'timing/model': 1.6108475886285305, 'learning_rate/qwen_vl_interface': 9.310348565683938e-06, 'learning_rate/action_model': 9.310348565683939e-05, 'epoch': 0.37}) +[2026-06-15 21:34:58,252][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.21847598254680634, 'train/grad_norm_pre_clip': 5.868122100830078, 'timing/data': 0.013749675825238228, 'timing/model': 1.5554967932403088, 'learning_rate/qwen_vl_interface': 9.30876733803656e-06, 'learning_rate/action_model': 9.308767338036559e-05, 'epoch': 0.37}) +[2026-06-15 21:35:04,319][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.11303739994764328, 'train/grad_norm_pre_clip': 3.6217870712280273, 'timing/data': 0.012161599472165108, 'timing/model': 1.5400865990668535, 'learning_rate/qwen_vl_interface': 9.30718444198324e-06, 'learning_rate/action_model': 9.307184441983239e-05, 'epoch': 0.37}) +[2026-06-15 21:35:10,448][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.3749995529651642, 'train/grad_norm_pre_clip': 6.906598091125488, 'timing/data': 0.01211673766374588, 'timing/model': 1.5003641173243523, 'learning_rate/qwen_vl_interface': 9.30559987817465e-06, 'learning_rate/action_model': 9.30559987817465e-05, 'epoch': 0.37}) +[2026-06-15 21:35:16,990][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.16959470510482788, 'train/grad_norm_pre_clip': 4.472731113433838, 'timing/data': 0.01273849606513977, 'timing/model': 1.7935263682156801, 'learning_rate/qwen_vl_interface': 9.304013647262143e-06, 'learning_rate/action_model': 9.304013647262143e-05, 'epoch': 0.37}) +[2026-06-15 21:35:22,936][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.1500360369682312, 'train/grad_norm_pre_clip': 5.048038482666016, 'timing/data': 0.012305429205298424, 'timing/model': 1.5381443202495575, 'learning_rate/qwen_vl_interface': 9.302425749897761e-06, 'learning_rate/action_model': 9.302425749897761e-05, 'epoch': 0.37}) +[2026-06-15 21:35:28,898][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.10140814632177353, 'train/grad_norm_pre_clip': 5.958983898162842, 'timing/data': 0.011694692075252533, 'timing/model': 1.467264398932457, 'learning_rate/qwen_vl_interface': 9.300836186734225e-06, 'learning_rate/action_model': 9.300836186734226e-05, 'epoch': 0.37}) +[2026-06-15 21:35:34,782][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.3045533001422882, 'train/grad_norm_pre_clip': 7.9529924392700195, 'timing/data': 0.011015687137842178, 'timing/model': 1.4582986067980528, 'learning_rate/qwen_vl_interface': 9.299244958424947e-06, 'learning_rate/action_model': 9.299244958424946e-05, 'epoch': 0.37}) +[2026-06-15 21:35:40,679][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.2163175344467163, 'train/grad_norm_pre_clip': 11.521099090576172, 'timing/data': 0.010350795462727547, 'timing/model': 1.4549897741526365, 'learning_rate/qwen_vl_interface': 9.29765206562402e-06, 'learning_rate/action_model': 9.297652065624018e-05, 'epoch': 0.37}) +[2026-06-15 21:35:46,613][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.1338212490081787, 'train/grad_norm_pre_clip': 6.261387348175049, 'timing/data': 0.011022701859474182, 'timing/model': 1.4969598315656185, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 0.37}) +[2026-06-15 21:35:52,525][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.10950159281492233, 'train/grad_norm_pre_clip': 8.60978889465332, 'timing/data': 0.011414431035518646, 'timing/model': 1.6547387670725584, 'learning_rate/qwen_vl_interface': 9.294461289167013e-06, 'learning_rate/action_model': 9.294461289167014e-05, 'epoch': 0.37}) +[2026-06-15 21:35:58,220][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.2730858325958252, 'train/grad_norm_pre_clip': 15.168874740600586, 'timing/data': 0.01089552417397499, 'timing/model': 1.4574303328990936, 'learning_rate/qwen_vl_interface': 9.292863406822543e-06, 'learning_rate/action_model': 9.292863406822542e-05, 'epoch': 0.37}) +[2026-06-15 21:36:04,206][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.12632432579994202, 'train/grad_norm_pre_clip': 4.86808443069458, 'timing/data': 0.010864835232496262, 'timing/model': 1.5228573586791754, 'learning_rate/qwen_vl_interface': 9.291263862609638e-06, 'learning_rate/action_model': 9.291263862609638e-05, 'epoch': 0.37}) +[2026-06-15 21:36:10,113][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.07278350740671158, 'train/grad_norm_pre_clip': 4.60497522354126, 'timing/data': 0.011415723711252213, 'timing/model': 1.4533490166068077, 'learning_rate/qwen_vl_interface': 9.28966265718581e-06, 'learning_rate/action_model': 9.28966265718581e-05, 'epoch': 0.37}) +[2026-06-15 21:36:16,019][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.2465616762638092, 'train/grad_norm_pre_clip': 8.956875801086426, 'timing/data': 0.010674210265278816, 'timing/model': 1.4558570086956024, 'learning_rate/qwen_vl_interface': 9.288059791209256e-06, 'learning_rate/action_model': 9.288059791209256e-05, 'epoch': 0.37}) +[2026-06-15 21:36:21,933][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.08474506437778473, 'train/grad_norm_pre_clip': 4.291271686553955, 'timing/data': 0.010887634009122849, 'timing/model': 1.4786085523664951, 'learning_rate/qwen_vl_interface': 9.286455265338853e-06, 'learning_rate/action_model': 9.286455265338853e-05, 'epoch': 0.37}) +[2026-06-15 21:36:27,838][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.2687990963459015, 'train/grad_norm_pre_clip': 12.16909122467041, 'timing/data': 0.010954737663269043, 'timing/model': 1.6822356525808573, 'learning_rate/qwen_vl_interface': 9.284849080234161e-06, 'learning_rate/action_model': 9.28484908023416e-05, 'epoch': 0.37}) +[2026-06-15 21:36:34,335][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.10369332879781723, 'train/grad_norm_pre_clip': 7.946263790130615, 'timing/data': 0.011712079867720604, 'timing/model': 1.46355558373034, 'learning_rate/qwen_vl_interface': 9.283241236555421e-06, 'learning_rate/action_model': 9.283241236555421e-05, 'epoch': 0.37}) +[2026-06-15 21:36:40,540][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.3883911371231079, 'train/grad_norm_pre_clip': 12.435551643371582, 'timing/data': 0.013669570907950401, 'timing/model': 1.5403976123780012, 'learning_rate/qwen_vl_interface': 9.281631734963557e-06, 'learning_rate/action_model': 9.281631734963557e-05, 'epoch': 0.38}) +[2026-06-15 21:36:46,690][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.43146997690200806, 'train/grad_norm_pre_clip': 29.7310733795166, 'timing/data': 0.01105666533112526, 'timing/model': 1.5104870274662971, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 0.38}) +[2026-06-15 21:36:52,481][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.04707252234220505, 'train/grad_norm_pre_clip': 3.8747167587280273, 'timing/data': 0.010634360834956169, 'timing/model': 1.4505126103758812, 'learning_rate/qwen_vl_interface': 9.278407760687562e-06, 'learning_rate/action_model': 9.278407760687562e-05, 'epoch': 0.38}) +[2026-06-15 21:36:58,379][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.375052273273468, 'train/grad_norm_pre_clip': 14.1573486328125, 'timing/data': 0.010231735184788704, 'timing/model': 1.4409903436899185, 'learning_rate/qwen_vl_interface': 9.276793289328686e-06, 'learning_rate/action_model': 9.276793289328685e-05, 'epoch': 0.38}) +[2026-06-15 21:37:04,510][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.16559769213199615, 'train/grad_norm_pre_clip': 12.2810640335083, 'timing/data': 0.01123516634106636, 'timing/model': 1.6498974598944187, 'learning_rate/qwen_vl_interface': 9.27517716270719e-06, 'learning_rate/action_model': 9.275177162707191e-05, 'epoch': 0.38}) +[2026-06-15 21:37:10,210][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.03052874654531479, 'train/grad_norm_pre_clip': 2.4199042320251465, 'timing/data': 0.009922118857502937, 'timing/model': 1.4535524640232325, 'learning_rate/qwen_vl_interface': 9.273559381487409e-06, 'learning_rate/action_model': 9.273559381487409e-05, 'epoch': 0.38}) +[2026-06-15 21:37:16,112][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.09874287247657776, 'train/grad_norm_pre_clip': 16.502172470092773, 'timing/data': 0.01028561219573021, 'timing/model': 1.4550313372164965, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.38}) +[2026-06-15 21:37:22,010][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.35833287239074707, 'train/grad_norm_pre_clip': 10.613051414489746, 'timing/data': 0.011130871251225471, 'timing/model': 1.449417231604457, 'learning_rate/qwen_vl_interface': 9.2703188579137e-06, 'learning_rate/action_model': 9.2703188579137e-05, 'epoch': 0.38}) +[2026-06-15 21:37:27,909][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.271318644285202, 'train/grad_norm_pre_clip': 9.331229209899902, 'timing/data': 0.010451585054397583, 'timing/model': 1.4529521334916353, 'learning_rate/qwen_vl_interface': 9.26869611689183e-06, 'learning_rate/action_model': 9.26869611689183e-05, 'epoch': 0.38}) +[2026-06-15 21:37:33,815][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.04268130660057068, 'train/grad_norm_pre_clip': 9.338216781616211, 'timing/data': 0.01050734706223011, 'timing/model': 1.4460313580930233, 'learning_rate/qwen_vl_interface': 9.267071723935784e-06, 'learning_rate/action_model': 9.267071723935784e-05, 'epoch': 0.38}) +[2026-06-15 21:37:39,714][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.22886419296264648, 'train/grad_norm_pre_clip': 15.806002616882324, 'timing/data': 0.010785235092043877, 'timing/model': 1.6608779057860374, 'learning_rate/qwen_vl_interface': 9.26544567971329e-06, 'learning_rate/action_model': 9.26544567971329e-05, 'epoch': 0.38}) +[2026-06-15 21:37:45,421][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.2705998122692108, 'train/grad_norm_pre_clip': 10.75424575805664, 'timing/data': 0.011521900072693825, 'timing/model': 1.4611310008913279, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 0.38}) +[2026-06-15 21:37:51,305][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.12091647833585739, 'train/grad_norm_pre_clip': 7.972076892852783, 'timing/data': 0.011039849370718002, 'timing/model': 1.4532386995851994, 'learning_rate/qwen_vl_interface': 9.262188640143262e-06, 'learning_rate/action_model': 9.262188640143262e-05, 'epoch': 0.38}) +[2026-06-15 21:37:57,190][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.22868910431861877, 'train/grad_norm_pre_clip': 16.099027633666992, 'timing/data': 0.011744698509573936, 'timing/model': 1.4606589730829, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.38}) +[2026-06-15 21:38:03,082][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.31648558378219604, 'train/grad_norm_pre_clip': 14.32918643951416, 'timing/data': 0.011728884652256966, 'timing/model': 1.4541995637118816, 'learning_rate/qwen_vl_interface': 9.258925003537127e-06, 'learning_rate/action_model': 9.258925003537126e-05, 'epoch': 0.38}) +[2026-06-15 21:38:09,021][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.3119460940361023, 'train/grad_norm_pre_clip': 8.840660095214844, 'timing/data': 0.0112775769084692, 'timing/model': 1.4801276829093695, 'learning_rate/qwen_vl_interface': 9.257290713022041e-06, 'learning_rate/action_model': 9.257290713022041e-05, 'epoch': 0.38}) +[2026-06-15 21:38:15,135][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.45272716879844666, 'train/grad_norm_pre_clip': 13.873525619506836, 'timing/data': 0.010870255529880524, 'timing/model': 1.673125982284546, 'learning_rate/qwen_vl_interface': 9.255654775261112e-06, 'learning_rate/action_model': 9.255654775261112e-05, 'epoch': 0.38}) +[2026-06-15 21:38:21,805][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.06326843798160553, 'train/grad_norm_pre_clip': 6.463525295257568, 'timing/data': 0.013557907193899155, 'timing/model': 2.082050021737814, 'learning_rate/qwen_vl_interface': 9.254017190926813e-06, 'learning_rate/action_model': 9.254017190926813e-05, 'epoch': 0.38}) +[2026-06-15 21:38:28,405][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.08993835747241974, 'train/grad_norm_pre_clip': 7.08416223526001, 'timing/data': 0.01263371855020523, 'timing/model': 1.5887704566121101, 'learning_rate/qwen_vl_interface': 9.25237796069229e-06, 'learning_rate/action_model': 9.25237796069229e-05, 'epoch': 0.38}) +[2026-06-15 21:38:35,063][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.1241445541381836, 'train/grad_norm_pre_clip': 6.4694037437438965, 'timing/data': 0.016843317076563835, 'timing/model': 1.5176612045615911, 'learning_rate/qwen_vl_interface': 9.250737085231371e-06, 'learning_rate/action_model': 9.250737085231372e-05, 'epoch': 0.38}) +[2026-06-15 21:38:41,342][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.1677069216966629, 'train/grad_norm_pre_clip': 6.2029314041137695, 'timing/data': 0.012479979544878006, 'timing/model': 1.556363569572568, 'learning_rate/qwen_vl_interface': 9.24909456521856e-06, 'learning_rate/action_model': 9.249094565218559e-05, 'epoch': 0.38}) +[2026-06-15 21:38:47,436][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.15191347897052765, 'train/grad_norm_pre_clip': 7.232162952423096, 'timing/data': 0.012176509946584702, 'timing/model': 1.4931528884917498, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 0.38}) +[2026-06-15 21:38:53,871][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.2558709979057312, 'train/grad_norm_pre_clip': 12.67628288269043, 'timing/data': 0.012946443632245064, 'timing/model': 1.7685985434800386, 'learning_rate/qwen_vl_interface': 9.245804594238637e-06, 'learning_rate/action_model': 9.245804594238636e-05, 'epoch': 0.38}) +[2026-06-15 21:39:00,190][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.1489313542842865, 'train/grad_norm_pre_clip': 4.863381862640381, 'timing/data': 0.012731235474348068, 'timing/model': 1.7469949685037136, 'learning_rate/qwen_vl_interface': 9.244157144623909e-06, 'learning_rate/action_model': 9.244157144623909e-05, 'epoch': 0.38}) +[2026-06-15 21:39:06,269][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.12148130685091019, 'train/grad_norm_pre_clip': 3.8251640796661377, 'timing/data': 0.011851591989398003, 'timing/model': 1.529654024168849, 'learning_rate/qwen_vl_interface': 9.242508053162051e-06, 'learning_rate/action_model': 9.242508053162051e-05, 'epoch': 0.38}) +[2026-06-15 21:39:12,563][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.29745569825172424, 'train/grad_norm_pre_clip': 12.873458862304688, 'timing/data': 0.013103719800710678, 'timing/model': 1.5515609849244356, 'learning_rate/qwen_vl_interface': 9.240857320530944e-06, 'learning_rate/action_model': 9.240857320530943e-05, 'epoch': 0.38}) +[2026-06-15 21:39:18,967][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.14721035957336426, 'train/grad_norm_pre_clip': 11.23918628692627, 'timing/data': 0.01239699125289917, 'timing/model': 1.5516350828111172, 'learning_rate/qwen_vl_interface': 9.23920494740914e-06, 'learning_rate/action_model': 9.23920494740914e-05, 'epoch': 0.39}) +[2026-06-15 21:39:25,389][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.29841262102127075, 'train/grad_norm_pre_clip': 11.579176902770996, 'timing/data': 0.012746535241603851, 'timing/model': 1.5793986655771732, 'learning_rate/qwen_vl_interface': 9.237550934475868e-06, 'learning_rate/action_model': 9.237550934475867e-05, 'epoch': 0.39}) +[2026-06-15 21:39:31,717][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.2400347739458084, 'train/grad_norm_pre_clip': 9.754467010498047, 'timing/data': 0.011667069047689438, 'timing/model': 1.7988953776657581, 'learning_rate/qwen_vl_interface': 9.235895282411029e-06, 'learning_rate/action_model': 9.235895282411029e-05, 'epoch': 0.39}) +[2026-06-15 21:39:37,766][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.17630186676979065, 'train/grad_norm_pre_clip': 8.993285179138184, 'timing/data': 0.012589965015649796, 'timing/model': 1.5475458204746246, 'learning_rate/qwen_vl_interface': 9.2342379918952e-06, 'learning_rate/action_model': 9.2342379918952e-05, 'epoch': 0.39}) +[2026-06-15 21:39:44,620][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.22776629030704498, 'train/grad_norm_pre_clip': 5.570355415344238, 'timing/data': 0.021156731992959976, 'timing/model': 1.7425310611724854, 'learning_rate/qwen_vl_interface': 9.232579063609631e-06, 'learning_rate/action_model': 9.232579063609631e-05, 'epoch': 0.39}) +[2026-06-15 21:47:14,145][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.046648651361465454, 'train/grad_norm_pre_clip': 2.898453950881958, 'eval/loss': 0.10180556774139404, 'eval/flap_precision': 0.8628640776699029, 'eval/flap_recall': 0.9647218453188603, 'eval/flap_f1': 0.9109545163356823, 'eval/flap_rate_model': 0.2288888888888889, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9109545163356823, 'timing/data': 0.013344775885343552, 'timing/model': 1.569240365177393, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 0.39}) +[2026-06-15 21:47:14,146][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1000_state +[2026-06-15 21:47:14,153][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 21:47:25,077][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1000_state/model.safetensors +[2026-06-15 21:47:41,793][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1000_state/optimizer.bin +[2026-06-15 21:47:41,793][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1000_state/sampler.bin +[2026-06-15 21:47:41,794][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-15 21:47:41,795][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-15 21:47:41,885][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-15 21:47:41,885][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 21:47:41,894][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 21:47:47,816][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.27707168459892273, 'train/grad_norm_pre_clip': 6.904143333435059, 'timing/data': 0.021547844633460045, 'timing/model': 1.4581664707511663, 'learning_rate/qwen_vl_interface': 9.229256296457635e-06, 'learning_rate/action_model': 9.229256296457634e-05, 'epoch': 0.39}) +[2026-06-15 21:47:53,709][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.15482176840305328, 'train/grad_norm_pre_clip': 8.479676246643066, 'timing/data': 0.011467602103948593, 'timing/model': 1.4673150479793549, 'learning_rate/qwen_vl_interface': 9.22759245895707e-06, 'learning_rate/action_model': 9.227592458957071e-05, 'epoch': 0.39}) +[2026-06-15 21:47:59,610][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.13718809187412262, 'train/grad_norm_pre_clip': 6.934897422790527, 'timing/data': 0.011211564764380455, 'timing/model': 1.449069894850254, 'learning_rate/qwen_vl_interface': 9.225926986418494e-06, 'learning_rate/action_model': 9.225926986418494e-05, 'epoch': 0.39}) +[2026-06-15 21:48:05,512][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.08066097646951675, 'train/grad_norm_pre_clip': 5.839199542999268, 'timing/data': 0.010258344933390617, 'timing/model': 1.666830848902464, 'learning_rate/qwen_vl_interface': 9.224259879526519e-06, 'learning_rate/action_model': 9.224259879526519e-05, 'epoch': 0.39}) +[2026-06-15 21:48:11,207][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.1644056737422943, 'train/grad_norm_pre_clip': 13.251150131225586, 'timing/data': 0.011089669540524483, 'timing/model': 1.4619317632168531, 'learning_rate/qwen_vl_interface': 9.222591138966427e-06, 'learning_rate/action_model': 9.222591138966426e-05, 'epoch': 0.39}) +[2026-06-15 21:48:17,072][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.15148241817951202, 'train/grad_norm_pre_clip': 7.6851277351379395, 'timing/data': 0.010856779292225838, 'timing/model': 1.4588325340300798, 'learning_rate/qwen_vl_interface': 9.220920765424175e-06, 'learning_rate/action_model': 9.220920765424174e-05, 'epoch': 0.39}) +[2026-06-15 21:48:22,970][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.1432236135005951, 'train/grad_norm_pre_clip': 5.594466686248779, 'timing/data': 0.01071387529373169, 'timing/model': 1.4532493259757757, 'learning_rate/qwen_vl_interface': 9.219248759586392e-06, 'learning_rate/action_model': 9.219248759586391e-05, 'epoch': 0.39}) +[2026-06-15 21:48:28,919][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.20982463657855988, 'train/grad_norm_pre_clip': 9.627006530761719, 'timing/data': 0.011412255465984344, 'timing/model': 1.4538291040807962, 'learning_rate/qwen_vl_interface': 9.217575122140375e-06, 'learning_rate/action_model': 9.217575122140375e-05, 'epoch': 0.39}) +[2026-06-15 21:48:34,816][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.05571255832910538, 'train/grad_norm_pre_clip': 3.6628308296203613, 'timing/data': 0.009957749396562576, 'timing/model': 1.5389472171664238, 'learning_rate/qwen_vl_interface': 9.215899853774094e-06, 'learning_rate/action_model': 9.215899853774093e-05, 'epoch': 0.39}) +[2026-06-15 21:48:40,881][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.16803108155727386, 'train/grad_norm_pre_clip': 12.82093620300293, 'timing/data': 0.011126544326543808, 'timing/model': 1.7021545264869928, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.39}) +[2026-06-15 21:48:46,961][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.10582462698221207, 'train/grad_norm_pre_clip': 8.098608016967773, 'timing/data': 0.012881709262728691, 'timing/model': 1.5181398019194603, 'learning_rate/qwen_vl_interface': 9.212544427035967e-06, 'learning_rate/action_model': 9.212544427035967e-05, 'epoch': 0.39}) +[2026-06-15 21:48:53,259][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.09825637191534042, 'train/grad_norm_pre_clip': 5.530412197113037, 'timing/data': 0.012671306729316711, 'timing/model': 1.5152200926095247, 'learning_rate/qwen_vl_interface': 9.21086427004341e-06, 'learning_rate/action_model': 9.210864270043411e-05, 'epoch': 0.39}) +[2026-06-15 21:48:59,776][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.09007647633552551, 'train/grad_norm_pre_clip': 6.440918922424316, 'timing/data': 0.01372908242046833, 'timing/model': 1.5646999180316925, 'learning_rate/qwen_vl_interface': 9.209182484889168e-06, 'learning_rate/action_model': 9.209182484889167e-05, 'epoch': 0.39}) +[2026-06-15 21:49:06,194][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.14625072479248047, 'train/grad_norm_pre_clip': 5.291211128234863, 'timing/data': 0.012935571372509003, 'timing/model': 1.6232928987592459, 'learning_rate/qwen_vl_interface': 9.207499072264557e-06, 'learning_rate/action_model': 9.207499072264558e-05, 'epoch': 0.39}) +[2026-06-15 21:49:12,630][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.3057187497615814, 'train/grad_norm_pre_clip': 10.514937400817871, 'timing/data': 0.01866593025624752, 'timing/model': 1.5738143287599087, 'learning_rate/qwen_vl_interface': 9.205814032861566e-06, 'learning_rate/action_model': 9.205814032861566e-05, 'epoch': 0.39}) +[2026-06-15 21:49:18,863][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.17617912590503693, 'train/grad_norm_pre_clip': 10.742250442504883, 'timing/data': 0.012470068410038948, 'timing/model': 1.7552637476474047, 'learning_rate/qwen_vl_interface': 9.204127367372849e-06, 'learning_rate/action_model': 9.204127367372849e-05, 'epoch': 0.39}) +[2026-06-15 21:49:24,976][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.23323842883110046, 'train/grad_norm_pre_clip': 11.013616561889648, 'timing/data': 0.012212194502353668, 'timing/model': 1.5641053840517998, 'learning_rate/qwen_vl_interface': 9.202439076491732e-06, 'learning_rate/action_model': 9.202439076491732e-05, 'epoch': 0.39}) +[2026-06-15 21:49:31,564][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.12141477316617966, 'train/grad_norm_pre_clip': 4.17929744720459, 'timing/data': 0.01218332163989544, 'timing/model': 1.755155060440302, 'learning_rate/qwen_vl_interface': 9.200749160912208e-06, 'learning_rate/action_model': 9.200749160912208e-05, 'epoch': 0.39}) +[2026-06-15 21:49:37,876][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.10897287726402283, 'train/grad_norm_pre_clip': 7.008778095245361, 'timing/data': 0.012478096410632133, 'timing/model': 1.5582463797181845, 'learning_rate/qwen_vl_interface': 9.199057621328937e-06, 'learning_rate/action_model': 9.199057621328935e-05, 'epoch': 0.39}) +[2026-06-15 21:49:44,158][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.07722728699445724, 'train/grad_norm_pre_clip': 3.4698846340179443, 'timing/data': 0.012500887736678123, 'timing/model': 1.5061052441596985, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 0.39}) +[2026-06-15 21:49:50,360][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.10131266713142395, 'train/grad_norm_pre_clip': 8.998044967651367, 'timing/data': 0.012112218886613846, 'timing/model': 1.7501898352056742, 'learning_rate/qwen_vl_interface': 9.19566967293313e-06, 'learning_rate/action_model': 9.19566967293313e-05, 'epoch': 0.4}) +[2026-06-15 21:49:56,365][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.2272028923034668, 'train/grad_norm_pre_clip': 11.253982543945312, 'timing/data': 0.013274047523736954, 'timing/model': 1.5517729464918375, 'learning_rate/qwen_vl_interface': 9.193973265513253e-06, 'learning_rate/action_model': 9.193973265513253e-05, 'epoch': 0.4}) +[2026-06-15 21:50:02,969][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.20599967241287231, 'train/grad_norm_pre_clip': 15.238222122192383, 'timing/data': 0.013295689597725868, 'timing/model': 1.5615519173443317, 'learning_rate/qwen_vl_interface': 9.192275236874946e-06, 'learning_rate/action_model': 9.192275236874946e-05, 'epoch': 0.4}) +[2026-06-15 21:50:09,168][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.10888668149709702, 'train/grad_norm_pre_clip': 18.04337501525879, 'timing/data': 0.014809951186180115, 'timing/model': 1.5480769854038954, 'learning_rate/qwen_vl_interface': 9.190575587716199e-06, 'learning_rate/action_model': 9.190575587716199e-05, 'epoch': 0.4}) +[2026-06-15 21:50:15,477][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.252594918012619, 'train/grad_norm_pre_clip': 27.75362777709961, 'timing/data': 0.014406716451048851, 'timing/model': 1.5529739204794168, 'learning_rate/qwen_vl_interface': 9.188874318735677e-06, 'learning_rate/action_model': 9.188874318735676e-05, 'epoch': 0.4}) +[2026-06-15 21:50:22,085][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.13179829716682434, 'train/grad_norm_pre_clip': 5.486799716949463, 'timing/data': 0.013874735683202744, 'timing/model': 1.6626599393785, 'learning_rate/qwen_vl_interface': 9.187171430632707e-06, 'learning_rate/action_model': 9.187171430632707e-05, 'epoch': 0.4}) +[2026-06-15 21:50:28,623][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.4074060022830963, 'train/grad_norm_pre_clip': 26.440034866333008, 'timing/data': 0.01216346025466919, 'timing/model': 1.9138204231858253, 'learning_rate/qwen_vl_interface': 9.185466924107283e-06, 'learning_rate/action_model': 9.185466924107282e-05, 'epoch': 0.4}) +[2026-06-15 21:50:34,679][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.3494150638580322, 'train/grad_norm_pre_clip': 12.387070655822754, 'timing/data': 0.012204261496663094, 'timing/model': 1.5206228867173195, 'learning_rate/qwen_vl_interface': 9.18376079986006e-06, 'learning_rate/action_model': 9.183760799860061e-05, 'epoch': 0.4}) +[2026-06-15 21:50:41,063][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.17878861725330353, 'train/grad_norm_pre_clip': 38.173526763916016, 'timing/data': 0.013645071536302567, 'timing/model': 1.544619895517826, 'learning_rate/qwen_vl_interface': 9.182053058592365e-06, 'learning_rate/action_model': 9.182053058592365e-05, 'epoch': 0.4}) +[2026-06-15 21:50:47,360][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.07263261079788208, 'train/grad_norm_pre_clip': 6.402219295501709, 'timing/data': 0.012123921886086464, 'timing/model': 1.5329743772745132, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 0.4}) +[2026-06-15 21:50:53,666][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.13971403241157532, 'train/grad_norm_pre_clip': 3.422419309616089, 'timing/data': 0.014368072152137756, 'timing/model': 1.5475403033196926, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.4}) +[2026-06-15 21:50:59,963][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.10585521161556244, 'train/grad_norm_pre_clip': 4.532573699951172, 'timing/data': 0.012071080505847931, 'timing/model': 1.551667159423232, 'learning_rate/qwen_vl_interface': 9.176920139689638e-06, 'learning_rate/action_model': 9.176920139689637e-05, 'epoch': 0.4}) +[2026-06-15 21:51:06,262][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.07074737548828125, 'train/grad_norm_pre_clip': 2.275235652923584, 'timing/data': 0.01341409981250763, 'timing/model': 1.7452887762337923, 'learning_rate/qwen_vl_interface': 9.175205937366568e-06, 'learning_rate/action_model': 9.175205937366569e-05, 'epoch': 0.4}) +[2026-06-15 21:51:12,364][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.22064274549484253, 'train/grad_norm_pre_clip': 9.260001182556152, 'timing/data': 0.012653559446334839, 'timing/model': 1.5554595105350018, 'learning_rate/qwen_vl_interface': 9.173490121539608e-06, 'learning_rate/action_model': 9.173490121539608e-05, 'epoch': 0.4}) +[2026-06-15 21:51:18,530][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.09488189220428467, 'train/grad_norm_pre_clip': 8.500126838684082, 'timing/data': 0.013908250257372856, 'timing/model': 1.4706657472997904, 'learning_rate/qwen_vl_interface': 9.171772692914062e-06, 'learning_rate/action_model': 9.171772692914061e-05, 'epoch': 0.4}) +[2026-06-15 21:51:24,726][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.16618385910987854, 'train/grad_norm_pre_clip': 6.498237133026123, 'timing/data': 0.012652771547436714, 'timing/model': 1.5490753147751093, 'learning_rate/qwen_vl_interface': 9.1700536521959e-06, 'learning_rate/action_model': 9.170053652195899e-05, 'epoch': 0.4}) +[2026-06-15 21:51:30,929][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.11398587375879288, 'train/grad_norm_pre_clip': 3.625577211380005, 'timing/data': 0.013855937868356705, 'timing/model': 1.5103246252983809, 'learning_rate/qwen_vl_interface': 9.168333000091754e-06, 'learning_rate/action_model': 9.168333000091754e-05, 'epoch': 0.4}) +[2026-06-15 21:51:36,878][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.055747125297784805, 'train/grad_norm_pre_clip': 2.5576798915863037, 'timing/data': 0.0108365248888731, 'timing/model': 1.4510667379945517, 'learning_rate/qwen_vl_interface': 9.16661073730892e-06, 'learning_rate/action_model': 9.166610737308919e-05, 'epoch': 0.4}) +[2026-06-15 21:51:42,738][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.05715456232428551, 'train/grad_norm_pre_clip': 3.752326488494873, 'timing/data': 0.010934252291917801, 'timing/model': 1.597614450380206, 'learning_rate/qwen_vl_interface': 9.164886864555355e-06, 'learning_rate/action_model': 9.164886864555355e-05, 'epoch': 0.4}) +[2026-06-15 21:51:48,375][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.053399309515953064, 'train/grad_norm_pre_clip': 2.7039952278137207, 'timing/data': 0.01129738800227642, 'timing/model': 1.4357099160552025, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 0.4}) +[2026-06-15 21:51:54,279][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.09239586442708969, 'train/grad_norm_pre_clip': 7.9715800285339355, 'timing/data': 0.010549819096922874, 'timing/model': 1.4406225811690092, 'learning_rate/qwen_vl_interface': 9.16143429197117e-06, 'learning_rate/action_model': 9.16143429197117e-05, 'epoch': 0.4}) +[2026-06-15 21:52:00,178][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.12178091704845428, 'train/grad_norm_pre_clip': 7.901038646697998, 'timing/data': 0.01222647912800312, 'timing/model': 1.444549884647131, 'learning_rate/qwen_vl_interface': 9.159705593559773e-06, 'learning_rate/action_model': 9.159705593559773e-05, 'epoch': 0.4}) +[2026-06-15 21:52:06,073][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.10386435687541962, 'train/grad_norm_pre_clip': 10.370361328125, 'timing/data': 0.010812126100063324, 'timing/model': 1.4293766152113676, 'learning_rate/qwen_vl_interface': 9.157975288016089e-06, 'learning_rate/action_model': 9.157975288016088e-05, 'epoch': 0.4}) +[2026-06-15 21:52:11,969][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.1699836403131485, 'train/grad_norm_pre_clip': 13.15367603302002, 'timing/data': 0.011513687670230865, 'timing/model': 1.4133453331887722, 'learning_rate/qwen_vl_interface': 9.156243376051381e-06, 'learning_rate/action_model': 9.156243376051381e-05, 'epoch': 0.4}) +[2026-06-15 21:52:17,842][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.1450732797384262, 'train/grad_norm_pre_clip': 14.37158489227295, 'timing/data': 0.010952992364764214, 'timing/model': 1.6018251236528158, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.4}) +[2026-06-15 21:52:23,474][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.13048353791236877, 'train/grad_norm_pre_clip': 25.105995178222656, 'timing/data': 0.010628240182995796, 'timing/model': 1.4375508036464453, 'learning_rate/qwen_vl_interface': 9.152774735707248e-06, 'learning_rate/action_model': 9.152774735707247e-05, 'epoch': 0.4}) +[2026-06-15 21:52:29,373][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.05855342745780945, 'train/grad_norm_pre_clip': 15.73808765411377, 'timing/data': 0.010599847882986069, 'timing/model': 1.4427329655736685, 'learning_rate/qwen_vl_interface': 9.15103800875365e-06, 'learning_rate/action_model': 9.15103800875365e-05, 'epoch': 0.41}) +[2026-06-15 21:52:35,176][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.3428521752357483, 'train/grad_norm_pre_clip': 12.473934173583984, 'timing/data': 0.010683722794055939, 'timing/model': 1.4455051086843014, 'learning_rate/qwen_vl_interface': 9.149299678230681e-06, 'learning_rate/action_model': 9.149299678230681e-05, 'epoch': 0.41}) +[2026-06-15 21:52:41,077][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.16758696734905243, 'train/grad_norm_pre_clip': 11.7835693359375, 'timing/data': 0.011554934084415436, 'timing/model': 1.4501360580325127, 'learning_rate/qwen_vl_interface': 9.147559744852905e-06, 'learning_rate/action_model': 9.147559744852905e-05, 'epoch': 0.41}) +[2026-06-15 21:52:47,335][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.01893371343612671, 'train/grad_norm_pre_clip': 3.3918607234954834, 'timing/data': 0.012238284572958946, 'timing/model': 1.5923118628561497, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 0.41}) +[2026-06-15 21:52:53,546][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.33774086833000183, 'train/grad_norm_pre_clip': 23.918930053710938, 'timing/data': 0.013172060251235962, 'timing/model': 1.7514700461179018, 'learning_rate/qwen_vl_interface': 9.14407507239447e-06, 'learning_rate/action_model': 9.14407507239447e-05, 'epoch': 0.41}) +[2026-06-15 21:52:59,737][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.2506634294986725, 'train/grad_norm_pre_clip': 20.546415328979492, 'timing/data': 0.011358445510268211, 'timing/model': 1.5539394840598106, 'learning_rate/qwen_vl_interface': 9.142330334746227e-06, 'learning_rate/action_model': 9.142330334746227e-05, 'epoch': 0.41}) +[2026-06-15 21:53:05,932][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.356819748878479, 'train/grad_norm_pre_clip': 20.903884887695312, 'timing/data': 0.012559011578559875, 'timing/model': 1.5491892211139202, 'learning_rate/qwen_vl_interface': 9.14058399710801e-06, 'learning_rate/action_model': 9.14058399710801e-05, 'epoch': 0.41}) +[2026-06-15 21:53:12,023][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.18540802597999573, 'train/grad_norm_pre_clip': 12.173863410949707, 'timing/data': 0.012545526027679443, 'timing/model': 1.5450823046267033, 'learning_rate/qwen_vl_interface': 9.138836060197673e-06, 'learning_rate/action_model': 9.138836060197672e-05, 'epoch': 0.41}) +[2026-06-15 21:53:18,218][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.158685103058815, 'train/grad_norm_pre_clip': 6.990239143371582, 'timing/data': 0.0130236204713583, 'timing/model': 1.5392052624374628, 'learning_rate/qwen_vl_interface': 9.137086524733725e-06, 'learning_rate/action_model': 9.137086524733724e-05, 'epoch': 0.41}) +[2026-06-15 21:53:24,427][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.31112241744995117, 'train/grad_norm_pre_clip': 14.81502628326416, 'timing/data': 0.011859351769089699, 'timing/model': 1.4948013611137867, 'learning_rate/qwen_vl_interface': 9.135335391435336e-06, 'learning_rate/action_model': 9.135335391435335e-05, 'epoch': 0.41}) +[2026-06-15 21:53:30,671][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.11127030849456787, 'train/grad_norm_pre_clip': 11.196609497070312, 'timing/data': 0.028236148878932, 'timing/model': 1.7527583707123995, 'learning_rate/qwen_vl_interface': 9.13358266102233e-06, 'learning_rate/action_model': 9.133582661022328e-05, 'epoch': 0.41}) +[2026-06-15 21:53:36,741][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.05108524113893509, 'train/grad_norm_pre_clip': 5.2244367599487305, 'timing/data': 0.013043869286775589, 'timing/model': 1.5398694202303886, 'learning_rate/qwen_vl_interface': 9.131828334215185e-06, 'learning_rate/action_model': 9.131828334215185e-05, 'epoch': 0.41}) +[2026-06-15 21:53:42,942][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.061322811990976334, 'train/grad_norm_pre_clip': 8.95024299621582, 'timing/data': 0.015188829973340034, 'timing/model': 1.5485367141664028, 'learning_rate/qwen_vl_interface': 9.130072411735043e-06, 'learning_rate/action_model': 9.130072411735043e-05, 'epoch': 0.41}) +[2026-06-15 21:53:49,151][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.09137044847011566, 'train/grad_norm_pre_clip': 6.770573616027832, 'timing/data': 0.016261272132396698, 'timing/model': 1.5613439548760653, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 0.41}) +[2026-06-15 21:53:55,443][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.12479574233293533, 'train/grad_norm_pre_clip': 7.372476577758789, 'timing/data': 0.014309220016002655, 'timing/model': 1.5418157447129488, 'learning_rate/qwen_vl_interface': 9.126555782643592e-06, 'learning_rate/action_model': 9.126555782643591e-05, 'epoch': 0.41}) +[2026-06-15 21:54:01,644][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.2556005120277405, 'train/grad_norm_pre_clip': 10.126444816589355, 'timing/data': 0.01258075051009655, 'timing/model': 1.5211703833192587, 'learning_rate/qwen_vl_interface': 9.124795077477835e-06, 'learning_rate/action_model': 9.124795077477834e-05, 'epoch': 0.41}) +[2026-06-15 21:54:08,161][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.38960036635398865, 'train/grad_norm_pre_clip': 26.525976181030273, 'timing/data': 0.013042712584137917, 'timing/model': 1.7792427446693182, 'learning_rate/qwen_vl_interface': 9.123032779530186e-06, 'learning_rate/action_model': 9.123032779530187e-05, 'epoch': 0.41}) +[2026-06-15 21:54:14,528][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.32481077313423157, 'train/grad_norm_pre_clip': 12.483561515808105, 'timing/data': 0.015217415988445282, 'timing/model': 1.8171913605183363, 'learning_rate/qwen_vl_interface': 9.12126888952506e-06, 'learning_rate/action_model': 9.12126888952506e-05, 'epoch': 0.41}) +[2026-06-15 21:54:20,625][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.08478409051895142, 'train/grad_norm_pre_clip': 7.047303676605225, 'timing/data': 0.015189135447144508, 'timing/model': 1.5854099001735449, 'learning_rate/qwen_vl_interface': 9.119503408187521e-06, 'learning_rate/action_model': 9.119503408187522e-05, 'epoch': 0.41}) +[2026-06-15 21:54:27,510][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.14608411490917206, 'train/grad_norm_pre_clip': 6.749758243560791, 'timing/data': 0.017516063526272774, 'timing/model': 1.6435557138174772, 'learning_rate/qwen_vl_interface': 9.117736336243299e-06, 'learning_rate/action_model': 9.117736336243298e-05, 'epoch': 0.41}) +[2026-06-15 21:54:34,561][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.08896578103303909, 'train/grad_norm_pre_clip': 5.386985778808594, 'timing/data': 0.012420887127518654, 'timing/model': 1.560088463127613, 'learning_rate/qwen_vl_interface': 9.115967674418765e-06, 'learning_rate/action_model': 9.115967674418765e-05, 'epoch': 0.41}) +[2026-06-15 21:54:40,863][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.07223371416330338, 'train/grad_norm_pre_clip': 4.559093475341797, 'timing/data': 0.011991433799266815, 'timing/model': 1.5508639179170132, 'learning_rate/qwen_vl_interface': 9.114197423440949e-06, 'learning_rate/action_model': 9.11419742344095e-05, 'epoch': 0.41}) +[2026-06-15 21:54:47,182][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.33041617274284363, 'train/grad_norm_pre_clip': 12.716039657592773, 'timing/data': 0.012322401627898216, 'timing/model': 1.7725251484662294, 'learning_rate/qwen_vl_interface': 9.112425584037536e-06, 'learning_rate/action_model': 9.112425584037536e-05, 'epoch': 0.41}) +[2026-06-15 21:54:53,264][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.20916032791137695, 'train/grad_norm_pre_clip': 12.312445640563965, 'timing/data': 0.012090608477592468, 'timing/model': 1.5425544884055853, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 0.41}) +[2026-06-15 21:54:59,560][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.22313210368156433, 'train/grad_norm_pre_clip': 10.320597648620605, 'timing/data': 0.012378238141536713, 'timing/model': 1.547958429902792, 'learning_rate/qwen_vl_interface': 9.108877142867918e-06, 'learning_rate/action_model': 9.108877142867918e-05, 'epoch': 0.41}) +[2026-06-15 21:55:05,867][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.1300964504480362, 'train/grad_norm_pre_clip': 8.126948356628418, 'timing/data': 0.012925151735544205, 'timing/model': 1.5457532815635204, 'learning_rate/qwen_vl_interface': 9.107100542560342e-06, 'learning_rate/action_model': 9.107100542560342e-05, 'epoch': 0.41}) +[2026-06-15 21:55:12,148][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.14640896022319794, 'train/grad_norm_pre_clip': 6.927704334259033, 'timing/data': 0.011611595749855042, 'timing/model': 1.5323528107255697, 'learning_rate/qwen_vl_interface': 9.105322356744429e-06, 'learning_rate/action_model': 9.105322356744429e-05, 'epoch': 0.42}) +[2026-06-15 21:55:18,133][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.10074571520090103, 'train/grad_norm_pre_clip': 9.389364242553711, 'timing/data': 0.011769905686378479, 'timing/model': 1.4832363557070494, 'learning_rate/qwen_vl_interface': 9.103542586151121e-06, 'learning_rate/action_model': 9.10354258615112e-05, 'epoch': 0.42}) +[2026-06-15 21:55:24,178][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.4542391002178192, 'train/grad_norm_pre_clip': 15.560866355895996, 'timing/data': 0.01158975251019001, 'timing/model': 1.6892072521150112, 'learning_rate/qwen_vl_interface': 9.101761231512017e-06, 'learning_rate/action_model': 9.101761231512017e-05, 'epoch': 0.42}) +[2026-06-15 21:55:30,135][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.23859740793704987, 'train/grad_norm_pre_clip': 12.844186782836914, 'timing/data': 0.012477666139602661, 'timing/model': 1.5361412838101387, 'learning_rate/qwen_vl_interface': 9.099978293559365e-06, 'learning_rate/action_model': 9.099978293559366e-05, 'epoch': 0.42}) +[2026-06-15 21:55:36,571][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.09695952385663986, 'train/grad_norm_pre_clip': 10.504056930541992, 'timing/data': 0.013264814391732216, 'timing/model': 1.5611196756362915, 'learning_rate/qwen_vl_interface': 9.098193773026065e-06, 'learning_rate/action_model': 9.098193773026064e-05, 'epoch': 0.42}) +[2026-06-15 21:55:43,004][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.061174217611551285, 'train/grad_norm_pre_clip': 6.141226291656494, 'timing/data': 0.013148780912160873, 'timing/model': 1.588754864409566, 'learning_rate/qwen_vl_interface': 9.096407670645663e-06, 'learning_rate/action_model': 9.096407670645662e-05, 'epoch': 0.42}) +[2026-06-15 21:55:49,260][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.17109684646129608, 'train/grad_norm_pre_clip': 7.533446788787842, 'timing/data': 0.012317648157477379, 'timing/model': 1.5085052717477083, 'learning_rate/qwen_vl_interface': 9.094619987152357e-06, 'learning_rate/action_model': 9.094619987152358e-05, 'epoch': 0.42}) +[2026-06-15 21:55:55,462][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.12908780574798584, 'train/grad_norm_pre_clip': 10.31092357635498, 'timing/data': 0.011898582801222801, 'timing/model': 1.5474024564027786, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.42}) +[2026-06-15 21:56:01,942][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.020974868908524513, 'train/grad_norm_pre_clip': 1.0801705121994019, 'timing/data': 0.011927016079425812, 'timing/model': 1.7298660688102245, 'learning_rate/qwen_vl_interface': 9.09103987976709e-06, 'learning_rate/action_model': 9.09103987976709e-05, 'epoch': 0.42}) +[2026-06-15 21:56:07,962][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.24222227931022644, 'train/grad_norm_pre_clip': 8.36782455444336, 'timing/data': 0.012392180040478706, 'timing/model': 1.5425418820232153, 'learning_rate/qwen_vl_interface': 9.089247457346774e-06, 'learning_rate/action_model': 9.089247457346773e-05, 'epoch': 0.42}) +[2026-06-15 21:56:14,256][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.371188223361969, 'train/grad_norm_pre_clip': 15.258654594421387, 'timing/data': 0.011111568659543991, 'timing/model': 1.547781690955162, 'learning_rate/qwen_vl_interface': 9.087453456756849e-06, 'learning_rate/action_model': 9.087453456756849e-05, 'epoch': 0.42}) +[2026-06-15 21:56:20,573][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.14062809944152832, 'train/grad_norm_pre_clip': 7.052487850189209, 'timing/data': 0.01002315990626812, 'timing/model': 1.5662194527685642, 'learning_rate/qwen_vl_interface': 9.085657878734759e-06, 'learning_rate/action_model': 9.085657878734758e-05, 'epoch': 0.42}) +[2026-06-15 21:56:27,260][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.14307692646980286, 'train/grad_norm_pre_clip': 7.599001407623291, 'timing/data': 0.013737568631768227, 'timing/model': 1.5589588731527328, 'learning_rate/qwen_vl_interface': 9.083860724018603e-06, 'learning_rate/action_model': 9.083860724018602e-05, 'epoch': 0.42}) +[2026-06-15 21:56:33,466][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.11788463592529297, 'train/grad_norm_pre_clip': 9.46163558959961, 'timing/data': 0.012699127197265625, 'timing/model': 1.736148327589035, 'learning_rate/qwen_vl_interface': 9.08206199334712e-06, 'learning_rate/action_model': 9.082061993347119e-05, 'epoch': 0.42}) +[2026-06-15 21:56:39,705][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.3238323926925659, 'train/grad_norm_pre_clip': 28.843313217163086, 'timing/data': 0.011339997872710228, 'timing/model': 1.6796547397971153, 'learning_rate/qwen_vl_interface': 9.0802616874597e-06, 'learning_rate/action_model': 9.0802616874597e-05, 'epoch': 0.42}) +[2026-06-15 21:56:46,493][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.010850233025848866, 'train/grad_norm_pre_clip': 1.3529821634292603, 'timing/data': 0.01412939839065075, 'timing/model': 1.6452048867940903, 'learning_rate/qwen_vl_interface': 9.078459807096385e-06, 'learning_rate/action_model': 9.078459807096385e-05, 'epoch': 0.42}) +[2026-06-15 21:56:52,882][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.09581194818019867, 'train/grad_norm_pre_clip': 6.475162506103516, 'timing/data': 0.012566521763801575, 'timing/model': 1.540947437286377, 'learning_rate/qwen_vl_interface': 9.076656352997854e-06, 'learning_rate/action_model': 9.076656352997853e-05, 'epoch': 0.42}) +[2026-06-15 21:56:59,324][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.017628271132707596, 'train/grad_norm_pre_clip': 4.102265357971191, 'timing/data': 0.01810101419687271, 'timing/model': 1.6381235402077436, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 0.42}) +[2026-06-15 21:57:05,559][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.06792589277029037, 'train/grad_norm_pre_clip': 11.826790809631348, 'timing/data': 0.01262698508799076, 'timing/model': 1.5117288995534182, 'learning_rate/qwen_vl_interface': 9.073044726561128e-06, 'learning_rate/action_model': 9.073044726561129e-05, 'epoch': 0.42}) +[2026-06-15 21:57:11,810][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.30378443002700806, 'train/grad_norm_pre_clip': 12.682318687438965, 'timing/data': 0.012253619730472565, 'timing/model': 1.80025496147573, 'learning_rate/qwen_vl_interface': 9.071236555707536e-06, 'learning_rate/action_model': 9.071236555707537e-05, 'epoch': 0.42}) +[2026-06-15 21:57:17,879][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.41999825835227966, 'train/grad_norm_pre_clip': 22.524641036987305, 'timing/data': 0.012849589809775352, 'timing/model': 1.5666768364608288, 'learning_rate/qwen_vl_interface': 9.069426814087937e-06, 'learning_rate/action_model': 9.069426814087937e-05, 'epoch': 0.42}) +[2026-06-15 21:57:24,167][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.06906765699386597, 'train/grad_norm_pre_clip': 13.51388168334961, 'timing/data': 0.012702185660600662, 'timing/model': 1.5556051041930914, 'learning_rate/qwen_vl_interface': 9.067615502446247e-06, 'learning_rate/action_model': 9.067615502446246e-05, 'epoch': 0.42}) +[2026-06-15 21:57:30,465][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.28135961294174194, 'train/grad_norm_pre_clip': 25.55377197265625, 'timing/data': 0.011736977845430374, 'timing/model': 1.5560156907886267, 'learning_rate/qwen_vl_interface': 9.06580262152703e-06, 'learning_rate/action_model': 9.065802621527028e-05, 'epoch': 0.42}) +[2026-06-15 21:57:36,766][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.1921498328447342, 'train/grad_norm_pre_clip': 15.495628356933594, 'timing/data': 0.012660885229706764, 'timing/model': 1.5558287482708693, 'learning_rate/qwen_vl_interface': 9.063988172075486e-06, 'learning_rate/action_model': 9.063988172075486e-05, 'epoch': 0.42}) +[2026-06-15 21:57:43,065][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.008402352221310139, 'train/grad_norm_pre_clip': 3.542050838470459, 'timing/data': 0.01287454180419445, 'timing/model': 1.5528489332646132, 'learning_rate/qwen_vl_interface': 9.062172154837475e-06, 'learning_rate/action_model': 9.062172154837475e-05, 'epoch': 0.42}) +[2026-06-15 21:57:49,404][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.23960080742835999, 'train/grad_norm_pre_clip': 22.03043556213379, 'timing/data': 0.012759307399392128, 'timing/model': 1.7837332841008902, 'learning_rate/qwen_vl_interface': 9.060354570559487e-06, 'learning_rate/action_model': 9.060354570559487e-05, 'epoch': 0.42}) +[2026-06-15 21:57:55,335][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.21190397441387177, 'train/grad_norm_pre_clip': 31.854867935180664, 'timing/data': 0.011781768873333931, 'timing/model': 1.535697365179658, 'learning_rate/qwen_vl_interface': 9.058535419988667e-06, 'learning_rate/action_model': 9.058535419988667e-05, 'epoch': 0.43}) +[2026-06-15 22:05:16,551][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.28425613045692444, 'train/grad_norm_pre_clip': 37.77574920654297, 'eval/loss': 0.10695761442184448, 'eval/flap_precision': 0.948051948051948, 'eval/flap_recall': 0.8914518317503393, 'eval/flap_f1': 0.9188811188811189, 'eval/flap_rate_model': 0.1925, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.918881118881119, 'timing/data': 0.011795870959758759, 'timing/model': 1.4424377158284187, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 0.43}) +[2026-06-15 22:05:16,552][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1100_state +[2026-06-15 22:05:16,557][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 22:05:22,839][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1100_state/model.safetensors +[2026-06-15 22:05:38,110][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1100_state/optimizer.bin +[2026-06-15 22:05:38,111][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1100_state/sampler.bin +[2026-06-15 22:05:38,111][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1100_state/sampler_1.bin +[2026-06-15 22:05:38,112][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1100_state/random_states_0.pkl +[2026-06-15 22:05:38,208][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-06-15 22:05:38,209][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 22:05:38,217][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 22:05:43,910][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.05718613415956497, 'train/grad_norm_pre_clip': 6.3817243576049805, 'timing/data': 0.010883018374443054, 'timing/model': 1.4768401961773634, 'learning_rate/qwen_vl_interface': 9.0548924229603e-06, 'learning_rate/action_model': 9.054892422960301e-05, 'epoch': 0.43}) +[2026-06-15 22:05:49,934][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.4301464855670929, 'train/grad_norm_pre_clip': 25.34037971496582, 'timing/data': 0.010734155774116516, 'timing/model': 1.4909090586006641, 'learning_rate/qwen_vl_interface': 9.053068578000256e-06, 'learning_rate/action_model': 9.053068578000256e-05, 'epoch': 0.43}) +[2026-06-15 22:05:55,900][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.28013473749160767, 'train/grad_norm_pre_clip': 21.162092208862305, 'timing/data': 0.011012125760316849, 'timing/model': 1.4519343636929989, 'learning_rate/qwen_vl_interface': 9.051243169742372e-06, 'learning_rate/action_model': 9.051243169742372e-05, 'epoch': 0.43}) +[2026-06-15 22:06:01,818][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.06945033371448517, 'train/grad_norm_pre_clip': 9.773555755615234, 'timing/data': 0.01021917536854744, 'timing/model': 1.4508567154407501, 'learning_rate/qwen_vl_interface': 9.049416198937008e-06, 'learning_rate/action_model': 9.049416198937007e-05, 'epoch': 0.43}) +[2026-06-15 22:06:07,721][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.08967416733503342, 'train/grad_norm_pre_clip': 20.24090576171875, 'timing/data': 0.009876556694507599, 'timing/model': 1.5219025146216154, 'learning_rate/qwen_vl_interface': 9.047587666335159e-06, 'learning_rate/action_model': 9.047587666335159e-05, 'epoch': 0.43}) +[2026-06-15 22:06:13,506][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.06893263757228851, 'train/grad_norm_pre_clip': 9.478460311889648, 'timing/data': 0.010901320725679398, 'timing/model': 1.5434448756277561, 'learning_rate/qwen_vl_interface': 9.045757572688469e-06, 'learning_rate/action_model': 9.04575757268847e-05, 'epoch': 0.43}) +[2026-06-15 22:06:19,213][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.2028995305299759, 'train/grad_norm_pre_clip': 16.60440444946289, 'timing/data': 0.011088632047176361, 'timing/model': 1.4495559819042683, 'learning_rate/qwen_vl_interface': 9.043925918749218e-06, 'learning_rate/action_model': 9.043925918749218e-05, 'epoch': 0.43}) +[2026-06-15 22:06:25,106][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.06129928678274155, 'train/grad_norm_pre_clip': 17.358030319213867, 'timing/data': 0.010761739686131477, 'timing/model': 1.4515196662396193, 'learning_rate/qwen_vl_interface': 9.042092705270333e-06, 'learning_rate/action_model': 9.042092705270333e-05, 'epoch': 0.43}) +[2026-06-15 22:06:31,005][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.06372448801994324, 'train/grad_norm_pre_clip': 6.322160720825195, 'timing/data': 0.010195976123213768, 'timing/model': 1.451899979263544, 'learning_rate/qwen_vl_interface': 9.040257933005375e-06, 'learning_rate/action_model': 9.040257933005374e-05, 'epoch': 0.43}) +[2026-06-15 22:06:36,941][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.28499582409858704, 'train/grad_norm_pre_clip': 22.677827835083008, 'timing/data': 0.01106988824903965, 'timing/model': 1.463014144450426, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 0.43}) +[2026-06-15 22:06:43,113][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.3427576720714569, 'train/grad_norm_pre_clip': 6.118840217590332, 'timing/data': 0.01282135583460331, 'timing/model': 1.809026988223195, 'learning_rate/qwen_vl_interface': 9.036583715134709e-06, 'learning_rate/action_model': 9.036583715134708e-05, 'epoch': 0.43}) +[2026-06-15 22:06:49,265][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.63890141248703, 'train/grad_norm_pre_clip': 30.1822509765625, 'timing/data': 0.01187647134065628, 'timing/model': 1.5491046160459518, 'learning_rate/qwen_vl_interface': 9.034744271039332e-06, 'learning_rate/action_model': 9.034744271039332e-05, 'epoch': 0.43}) +[2026-06-15 22:06:55,432][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.14535722136497498, 'train/grad_norm_pre_clip': 11.271977424621582, 'timing/data': 0.011624548584222794, 'timing/model': 1.549355959519744, 'learning_rate/qwen_vl_interface': 9.03290327117855e-06, 'learning_rate/action_model': 9.032903271178549e-05, 'epoch': 0.43}) +[2026-06-15 22:07:01,279][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.22647985816001892, 'train/grad_norm_pre_clip': 15.42543888092041, 'timing/data': 0.010933103039860725, 'timing/model': 1.4425221886485815, 'learning_rate/qwen_vl_interface': 9.031060716309122e-06, 'learning_rate/action_model': 9.031060716309121e-05, 'epoch': 0.43}) +[2026-06-15 22:07:07,194][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.14999063313007355, 'train/grad_norm_pre_clip': 12.8882474899292, 'timing/data': 0.010860631242394447, 'timing/model': 1.4484758451581001, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.43}) +[2026-06-15 22:07:13,084][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.14658555388450623, 'train/grad_norm_pre_clip': 5.46821928024292, 'timing/data': 0.010445632040500641, 'timing/model': 1.414351837709546, 'learning_rate/qwen_vl_interface': 9.027370944574604e-06, 'learning_rate/action_model': 9.027370944574604e-05, 'epoch': 0.43}) +[2026-06-15 22:07:19,132][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.24278299510478973, 'train/grad_norm_pre_clip': 13.525166511535645, 'timing/data': 0.010315297171473503, 'timing/model': 1.7641560956835747, 'learning_rate/qwen_vl_interface': 9.025523729226236e-06, 'learning_rate/action_model': 9.025523729226236e-05, 'epoch': 0.43}) +[2026-06-15 22:07:24,816][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.07070586830377579, 'train/grad_norm_pre_clip': 5.600308895111084, 'timing/data': 0.010293537750840187, 'timing/model': 1.4595329873263836, 'learning_rate/qwen_vl_interface': 9.023674961902677e-06, 'learning_rate/action_model': 9.023674961902676e-05, 'epoch': 0.43}) +[2026-06-15 22:07:30,723][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.4272564649581909, 'train/grad_norm_pre_clip': 11.233741760253906, 'timing/data': 0.010347677394747734, 'timing/model': 1.4657036606222391, 'learning_rate/qwen_vl_interface': 9.021824643363883e-06, 'learning_rate/action_model': 9.021824643363883e-05, 'epoch': 0.43}) +[2026-06-15 22:07:36,631][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.19290989637374878, 'train/grad_norm_pre_clip': 11.935040473937988, 'timing/data': 0.011366387829184532, 'timing/model': 1.4668901655822992, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 0.43}) +[2026-06-15 22:07:42,828][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.2909345030784607, 'train/grad_norm_pre_clip': 8.790668487548828, 'timing/data': 0.009892703965306282, 'timing/model': 1.6318082120269537, 'learning_rate/qwen_vl_interface': 9.018119355683623e-06, 'learning_rate/action_model': 9.018119355683623e-05, 'epoch': 0.43}) +[2026-06-15 22:07:48,712][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.2078525871038437, 'train/grad_norm_pre_clip': 11.34313678741455, 'timing/data': 0.010332910344004631, 'timing/model': 1.4540309645235538, 'learning_rate/qwen_vl_interface': 9.016264388065258e-06, 'learning_rate/action_model': 9.016264388065258e-05, 'epoch': 0.43}) +[2026-06-15 22:07:54,710][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.11067084223031998, 'train/grad_norm_pre_clip': 4.6982340812683105, 'timing/data': 0.01164381392300129, 'timing/model': 1.6430310048162937, 'learning_rate/qwen_vl_interface': 9.014407872277869e-06, 'learning_rate/action_model': 9.014407872277869e-05, 'epoch': 0.43}) +[2026-06-15 22:08:00,954][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.08558981120586395, 'train/grad_norm_pre_clip': 3.9854156970977783, 'timing/data': 0.0152658112347126, 'timing/model': 1.7717043198645115, 'learning_rate/qwen_vl_interface': 9.012549809084595e-06, 'learning_rate/action_model': 9.012549809084595e-05, 'epoch': 0.44}) +[2026-06-15 22:08:06,587][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.11140687763690948, 'train/grad_norm_pre_clip': 4.675559997558594, 'timing/data': 0.010435270145535469, 'timing/model': 1.4511222951114178, 'learning_rate/qwen_vl_interface': 9.01069019924922e-06, 'learning_rate/action_model': 9.010690199249219e-05, 'epoch': 0.44}) +[2026-06-15 22:08:12,485][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.16434244811534882, 'train/grad_norm_pre_clip': 4.9265289306640625, 'timing/data': 0.011140992864966393, 'timing/model': 1.4455901216715574, 'learning_rate/qwen_vl_interface': 9.008829043536157e-06, 'learning_rate/action_model': 9.008829043536156e-05, 'epoch': 0.44}) +[2026-06-15 22:08:18,391][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.2629956305027008, 'train/grad_norm_pre_clip': 15.120308876037598, 'timing/data': 0.010441865772008896, 'timing/model': 1.4551341887563467, 'learning_rate/qwen_vl_interface': 9.006966342710458e-06, 'learning_rate/action_model': 9.006966342710457e-05, 'epoch': 0.44}) +[2026-06-15 22:08:24,301][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.15582697093486786, 'train/grad_norm_pre_clip': 12.173855781555176, 'timing/data': 0.011790815740823746, 'timing/model': 1.4485392142087221, 'learning_rate/qwen_vl_interface': 9.005102097537809e-06, 'learning_rate/action_model': 9.005102097537809e-05, 'epoch': 0.44}) +[2026-06-15 22:08:30,248][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.12319188565015793, 'train/grad_norm_pre_clip': 6.103833198547363, 'timing/data': 0.016332469880580902, 'timing/model': 1.6987904012203217, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.44}) +[2026-06-15 22:08:35,901][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.13315904140472412, 'train/grad_norm_pre_clip': 4.770022869110107, 'timing/data': 0.012559618800878525, 'timing/model': 1.4701234884560108, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 0.44}) +[2026-06-15 22:08:41,905][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.09683580696582794, 'train/grad_norm_pre_clip': 6.489483833312988, 'timing/data': 0.012356555089354515, 'timing/model': 1.4722521807998419, 'learning_rate/qwen_vl_interface': 8.999500103604538e-06, 'learning_rate/action_model': 8.999500103604536e-05, 'epoch': 0.44}) +[2026-06-15 22:08:47,807][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.1440984606742859, 'train/grad_norm_pre_clip': 4.214000701904297, 'timing/data': 0.011818528175354004, 'timing/model': 1.476207448169589, 'learning_rate/qwen_vl_interface': 8.997629688713639e-06, 'learning_rate/action_model': 8.997629688713638e-05, 'epoch': 0.44}) +[2026-06-15 22:08:53,739][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.09910885989665985, 'train/grad_norm_pre_clip': 4.76268196105957, 'timing/data': 0.019100354984402657, 'timing/model': 1.4645908437669277, 'learning_rate/qwen_vl_interface': 8.995757733313733e-06, 'learning_rate/action_model': 8.995757733313732e-05, 'epoch': 0.44}) +[2026-06-15 22:08:59,632][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.09092482179403305, 'train/grad_norm_pre_clip': 6.605806350708008, 'timing/data': 0.010597212240099907, 'timing/model': 1.5405562203377485, 'learning_rate/qwen_vl_interface': 8.993884238174314e-06, 'learning_rate/action_model': 8.993884238174314e-05, 'epoch': 0.44}) +[2026-06-15 22:09:05,966][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.02209235541522503, 'train/grad_norm_pre_clip': 2.365140438079834, 'timing/data': 0.011695656925439835, 'timing/model': 1.8486189432442188, 'learning_rate/qwen_vl_interface': 8.992009204065505e-06, 'learning_rate/action_model': 8.992009204065505e-05, 'epoch': 0.44}) +[2026-06-15 22:09:12,084][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.13957586884498596, 'train/grad_norm_pre_clip': 7.7950215339660645, 'timing/data': 0.013593675568699837, 'timing/model': 1.55880656093359, 'learning_rate/qwen_vl_interface': 8.990132631758058e-06, 'learning_rate/action_model': 8.990132631758059e-05, 'epoch': 0.44}) +[2026-06-15 22:09:18,497][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.05995460972189903, 'train/grad_norm_pre_clip': 11.964502334594727, 'timing/data': 0.012134233489632607, 'timing/model': 1.5850753914564848, 'learning_rate/qwen_vl_interface': 8.988254522023364e-06, 'learning_rate/action_model': 8.988254522023363e-05, 'epoch': 0.44}) +[2026-06-15 22:09:24,788][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.25763148069381714, 'train/grad_norm_pre_clip': 32.472503662109375, 'timing/data': 0.012876365333795547, 'timing/model': 1.572418412193656, 'learning_rate/qwen_vl_interface': 8.986374875633441e-06, 'learning_rate/action_model': 8.98637487563344e-05, 'epoch': 0.44}) +[2026-06-15 22:09:31,089][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.21302853524684906, 'train/grad_norm_pre_clip': 27.188335418701172, 'timing/data': 0.01367867924273014, 'timing/model': 1.5416361633688211, 'learning_rate/qwen_vl_interface': 8.984493693360944e-06, 'learning_rate/action_model': 8.984493693360943e-05, 'epoch': 0.44}) +[2026-06-15 22:09:37,387][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.13267959654331207, 'train/grad_norm_pre_clip': 8.402892112731934, 'timing/data': 0.012572210282087326, 'timing/model': 1.5618282239884138, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 0.44}) +[2026-06-15 22:09:43,667][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.06360959261655807, 'train/grad_norm_pre_clip': 7.421630382537842, 'timing/data': 0.011802004650235176, 'timing/model': 1.7367471400648355, 'learning_rate/qwen_vl_interface': 8.98072672426198e-06, 'learning_rate/action_model': 8.980726724261981e-05, 'epoch': 0.44}) +[2026-06-15 22:09:49,772][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.1151532307267189, 'train/grad_norm_pre_clip': 15.870810508728027, 'timing/data': 0.01300443708896637, 'timing/model': 1.5418565440922976, 'learning_rate/qwen_vl_interface': 8.978840938983977e-06, 'learning_rate/action_model': 8.978840938983977e-05, 'epoch': 0.44}) +[2026-06-15 22:09:56,136][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.11189419031143188, 'train/grad_norm_pre_clip': 9.915371894836426, 'timing/data': 0.013168176636099815, 'timing/model': 1.5495668463408947, 'learning_rate/qwen_vl_interface': 8.976953620920313e-06, 'learning_rate/action_model': 8.976953620920312e-05, 'epoch': 0.44}) +[2026-06-15 22:10:02,435][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.31016936898231506, 'train/grad_norm_pre_clip': 13.746338844299316, 'timing/data': 0.011867363005876541, 'timing/model': 1.5498386062681675, 'learning_rate/qwen_vl_interface': 8.975064770846794e-06, 'learning_rate/action_model': 8.975064770846794e-05, 'epoch': 0.44}) +[2026-06-15 22:10:08,568][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.09571152925491333, 'train/grad_norm_pre_clip': 8.325394630432129, 'timing/data': 0.013660965487360954, 'timing/model': 1.5143931303173304, 'learning_rate/qwen_vl_interface': 8.973174389539856e-06, 'learning_rate/action_model': 8.973174389539855e-05, 'epoch': 0.44}) +[2026-06-15 22:10:15,174][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.07388085871934891, 'train/grad_norm_pre_clip': 8.047188758850098, 'timing/data': 0.012696808204054832, 'timing/model': 1.8178543150424957, 'learning_rate/qwen_vl_interface': 8.971282477776563e-06, 'learning_rate/action_model': 8.971282477776563e-05, 'epoch': 0.44}) +[2026-06-15 22:10:21,492][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.444036066532135, 'train/grad_norm_pre_clip': 13.166572570800781, 'timing/data': 0.014637179672718048, 'timing/model': 1.5757150333374739, 'learning_rate/qwen_vl_interface': 8.969389036334609e-06, 'learning_rate/action_model': 8.969389036334609e-05, 'epoch': 0.44}) +[2026-06-15 22:10:27,980][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.03687230125069618, 'train/grad_norm_pre_clip': 4.739160537719727, 'timing/data': 0.012633165344595909, 'timing/model': 1.5497931949794292, 'learning_rate/qwen_vl_interface': 8.967494065992315e-06, 'learning_rate/action_model': 8.967494065992315e-05, 'epoch': 0.44}) +[2026-06-15 22:10:34,578][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.023134319111704826, 'train/grad_norm_pre_clip': 4.144922733306885, 'timing/data': 0.013229530304670334, 'timing/model': 1.7532405741512775, 'learning_rate/qwen_vl_interface': 8.965597567528634e-06, 'learning_rate/action_model': 8.965597567528633e-05, 'epoch': 0.44}) +[2026-06-15 22:10:40,889][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.3269268572330475, 'train/grad_norm_pre_clip': 8.448051452636719, 'timing/data': 0.014086026698350906, 'timing/model': 1.559562610462308, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.45}) +[2026-06-15 22:10:47,182][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.09619039297103882, 'train/grad_norm_pre_clip': 5.412232398986816, 'timing/data': 0.01309066079556942, 'timing/model': 1.5754058565944433, 'learning_rate/qwen_vl_interface': 8.961799989356048e-06, 'learning_rate/action_model': 8.961799989356048e-05, 'epoch': 0.45}) +[2026-06-15 22:10:53,518][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.17074312269687653, 'train/grad_norm_pre_clip': 12.8922700881958, 'timing/data': 0.013503268361091614, 'timing/model': 1.7997717391699553, 'learning_rate/qwen_vl_interface': 8.959898911208185e-06, 'learning_rate/action_model': 8.959898911208185e-05, 'epoch': 0.45}) +[2026-06-15 22:11:00,372][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.11086215078830719, 'train/grad_norm_pre_clip': 7.4970197677612305, 'timing/data': 0.011697955429553986, 'timing/model': 1.5583481043577194, 'learning_rate/qwen_vl_interface': 8.957996308061015e-06, 'learning_rate/action_model': 8.957996308061014e-05, 'epoch': 0.45}) +[2026-06-15 22:11:06,784][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.36577099561691284, 'train/grad_norm_pre_clip': 8.530294418334961, 'timing/data': 0.012244323268532753, 'timing/model': 1.5676312856376171, 'learning_rate/qwen_vl_interface': 8.956092180696626e-06, 'learning_rate/action_model': 8.956092180696625e-05, 'epoch': 0.45}) +[2026-06-15 22:11:13,093][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.1542292982339859, 'train/grad_norm_pre_clip': 7.735238075256348, 'timing/data': 0.011965524405241013, 'timing/model': 1.574408520013094, 'learning_rate/qwen_vl_interface': 8.95418652989773e-06, 'learning_rate/action_model': 8.95418652989773e-05, 'epoch': 0.45}) +[2026-06-15 22:11:19,372][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.36635449528694153, 'train/grad_norm_pre_clip': 13.84746265411377, 'timing/data': 0.012481145560741425, 'timing/model': 1.5358442552387714, 'learning_rate/qwen_vl_interface': 8.952279356447675e-06, 'learning_rate/action_model': 8.952279356447675e-05, 'epoch': 0.45}) +[2026-06-15 22:11:25,672][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.14371487498283386, 'train/grad_norm_pre_clip': 5.358113765716553, 'timing/data': 0.013510536402463913, 'timing/model': 1.558502659201622, 'learning_rate/qwen_vl_interface': 8.950370661130421e-06, 'learning_rate/action_model': 8.950370661130422e-05, 'epoch': 0.45}) +[2026-06-15 22:11:32,014][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.16842952370643616, 'train/grad_norm_pre_clip': 6.335630416870117, 'timing/data': 0.012656915932893753, 'timing/model': 1.7968607228249311, 'learning_rate/qwen_vl_interface': 8.948460444730566e-06, 'learning_rate/action_model': 8.948460444730566e-05, 'epoch': 0.45}) +[2026-06-15 22:11:38,172][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.06329592317342758, 'train/grad_norm_pre_clip': 4.410772800445557, 'timing/data': 0.012181824073195457, 'timing/model': 1.5563598293811083, 'learning_rate/qwen_vl_interface': 8.946548708033324e-06, 'learning_rate/action_model': 8.946548708033323e-05, 'epoch': 0.45}) +[2026-06-15 22:11:44,541][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.09473175555467606, 'train/grad_norm_pre_clip': 7.549411296844482, 'timing/data': 0.012308968231081963, 'timing/model': 1.532156553119421, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 0.45}) +[2026-06-15 22:11:50,740][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.2231036275625229, 'train/grad_norm_pre_clip': 10.321554183959961, 'timing/data': 0.012436022982001305, 'timing/model': 1.5567026361823082, 'learning_rate/qwen_vl_interface': 8.942720676890677e-06, 'learning_rate/action_model': 8.942720676890677e-05, 'epoch': 0.45}) +[2026-06-15 22:11:56,844][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.19702482223510742, 'train/grad_norm_pre_clip': 9.577274322509766, 'timing/data': 0.012364471331238747, 'timing/model': 1.4830755908042192, 'learning_rate/qwen_vl_interface': 8.940804384018833e-06, 'learning_rate/action_model': 8.940804384018832e-05, 'epoch': 0.45}) +[2026-06-15 22:12:03,082][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.15932315587997437, 'train/grad_norm_pre_clip': 9.413077354431152, 'timing/data': 0.013024188578128815, 'timing/model': 1.6056472714990377, 'learning_rate/qwen_vl_interface': 8.93888657399672e-06, 'learning_rate/action_model': 8.93888657399672e-05, 'epoch': 0.45}) +[2026-06-15 22:12:09,890][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.11787554621696472, 'train/grad_norm_pre_clip': 4.610143661499023, 'timing/data': 0.013870032504200935, 'timing/model': 1.8021099083125591, 'learning_rate/qwen_vl_interface': 8.936967247612678e-06, 'learning_rate/action_model': 8.936967247612678e-05, 'epoch': 0.45}) +[2026-06-15 22:12:15,852][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.08754333853721619, 'train/grad_norm_pre_clip': 4.53568172454834, 'timing/data': 0.012908605858683586, 'timing/model': 1.4844599515199661, 'learning_rate/qwen_vl_interface': 8.93504640565567e-06, 'learning_rate/action_model': 8.93504640565567e-05, 'epoch': 0.45}) +[2026-06-15 22:12:22,055][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.09092646837234497, 'train/grad_norm_pre_clip': 6.120616912841797, 'timing/data': 0.01310189999639988, 'timing/model': 1.483984725549817, 'learning_rate/qwen_vl_interface': 8.933124048915278e-06, 'learning_rate/action_model': 8.933124048915278e-05, 'epoch': 0.45}) +[2026-06-15 22:12:28,458][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.12402569502592087, 'train/grad_norm_pre_clip': 10.378137588500977, 'timing/data': 0.013006919994950294, 'timing/model': 1.7732838597148657, 'learning_rate/qwen_vl_interface': 8.931200178181715e-06, 'learning_rate/action_model': 8.931200178181715e-05, 'epoch': 0.45}) +[2026-06-15 22:12:34,535][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.09135235100984573, 'train/grad_norm_pre_clip': 6.469967842102051, 'timing/data': 0.012503955513238907, 'timing/model': 1.5469528120011091, 'learning_rate/qwen_vl_interface': 8.929274794245812e-06, 'learning_rate/action_model': 8.929274794245812e-05, 'epoch': 0.45}) +[2026-06-15 22:12:40,768][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.31011247634887695, 'train/grad_norm_pre_clip': 22.39993667602539, 'timing/data': 0.013421451672911644, 'timing/model': 1.4937646761536598, 'learning_rate/qwen_vl_interface': 8.927347897899018e-06, 'learning_rate/action_model': 8.927347897899016e-05, 'epoch': 0.45}) +[2026-06-15 22:12:46,985][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.05824124813079834, 'train/grad_norm_pre_clip': 3.974776029586792, 'timing/data': 0.011881601065397263, 'timing/model': 1.788597920909524, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 0.45}) +[2026-06-15 22:12:52,938][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.21792921423912048, 'train/grad_norm_pre_clip': 11.48687744140625, 'timing/data': 0.012592136859893799, 'timing/model': 1.5385446399450302, 'learning_rate/qwen_vl_interface': 8.923489571141679e-06, 'learning_rate/action_model': 8.923489571141679e-05, 'epoch': 0.45}) +[2026-06-15 22:12:59,035][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.13209223747253418, 'train/grad_norm_pre_clip': 4.892486572265625, 'timing/data': 0.012242298573255539, 'timing/model': 1.5002704001963139, 'learning_rate/qwen_vl_interface': 8.921558142317147e-06, 'learning_rate/action_model': 8.921558142317147e-05, 'epoch': 0.45}) +[2026-06-15 22:13:05,238][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.6465821862220764, 'train/grad_norm_pre_clip': 14.289371490478516, 'timing/data': 0.012775074690580368, 'timing/model': 1.5359527226537466, 'learning_rate/qwen_vl_interface': 8.91962520425375e-06, 'learning_rate/action_model': 8.91962520425375e-05, 'epoch': 0.45}) +[2026-06-15 22:13:11,460][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.3693735599517822, 'train/grad_norm_pre_clip': 24.080781936645508, 'timing/data': 0.012239797040820122, 'timing/model': 1.561453839763999, 'learning_rate/qwen_vl_interface': 8.917690757746044e-06, 'learning_rate/action_model': 8.917690757746044e-05, 'epoch': 0.45}) +[2026-06-15 22:13:17,768][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.18134787678718567, 'train/grad_norm_pre_clip': 28.80377960205078, 'timing/data': 0.01321486383676529, 'timing/model': 1.5085068885236979, 'learning_rate/qwen_vl_interface': 8.915754803589208e-06, 'learning_rate/action_model': 8.915754803589209e-05, 'epoch': 0.45}) +[2026-06-15 22:13:24,984][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.07071232050657272, 'train/grad_norm_pre_clip': 3.6572046279907227, 'timing/data': 0.013160830363631248, 'timing/model': 1.8015740737318993, 'learning_rate/qwen_vl_interface': 8.913817342579042e-06, 'learning_rate/action_model': 8.913817342579042e-05, 'epoch': 0.46}) +[2026-06-15 22:13:31,558][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.17264848947525024, 'train/grad_norm_pre_clip': 10.224995613098145, 'timing/data': 0.011045122519135475, 'timing/model': 1.9452939070761204, 'learning_rate/qwen_vl_interface': 8.91187837551196e-06, 'learning_rate/action_model': 8.911878375511961e-05, 'epoch': 0.46}) +[2026-06-15 22:13:37,940][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.15732629597187042, 'train/grad_norm_pre_clip': 8.312850952148438, 'timing/data': 0.012793481349945068, 'timing/model': 1.557364098727703, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.46}) +[2026-06-15 22:13:44,133][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.08355022966861725, 'train/grad_norm_pre_clip': 7.603479862213135, 'timing/data': 0.012171579524874687, 'timing/model': 1.608223058283329, 'learning_rate/qwen_vl_interface': 8.907995926395814e-06, 'learning_rate/action_model': 8.907995926395815e-05, 'epoch': 0.46}) +[2026-06-15 22:13:50,447][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.11569983512163162, 'train/grad_norm_pre_clip': 6.967716217041016, 'timing/data': 0.012794148176908493, 'timing/model': 1.5582438558340073, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 0.46}) +[2026-06-15 22:13:56,657][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.16249501705169678, 'train/grad_norm_pre_clip': 5.365876197814941, 'timing/data': 0.01251455582678318, 'timing/model': 1.4897543583065271, 'learning_rate/qwen_vl_interface': 8.904107462624484e-06, 'learning_rate/action_model': 8.904107462624483e-05, 'epoch': 0.46}) +[2026-06-15 22:14:02,897][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.2833530306816101, 'train/grad_norm_pre_clip': 10.949203491210938, 'timing/data': 0.012149125337600708, 'timing/model': 1.7905821334570646, 'learning_rate/qwen_vl_interface': 8.902160977240739e-06, 'learning_rate/action_model': 8.902160977240739e-05, 'epoch': 0.46}) +[2026-06-15 22:14:08,948][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.08660890907049179, 'train/grad_norm_pre_clip': 10.775303840637207, 'timing/data': 0.012020736932754517, 'timing/model': 1.5639825407415628, 'learning_rate/qwen_vl_interface': 8.90021299059157e-06, 'learning_rate/action_model': 8.90021299059157e-05, 'epoch': 0.46}) +[2026-06-15 22:14:15,153][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.2592836022377014, 'train/grad_norm_pre_clip': 19.05078887939453, 'timing/data': 0.012723695486783981, 'timing/model': 1.568770268931985, 'learning_rate/qwen_vl_interface': 8.898263503477722e-06, 'learning_rate/action_model': 8.898263503477721e-05, 'epoch': 0.46}) +[2026-06-15 22:14:21,383][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.12722663581371307, 'train/grad_norm_pre_clip': 7.639601230621338, 'timing/data': 0.023124035447835922, 'timing/model': 1.5911690518260002, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.46}) +[2026-06-15 22:14:27,755][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.24334275722503662, 'train/grad_norm_pre_clip': 12.087228775024414, 'timing/data': 0.013228638097643852, 'timing/model': 1.5753607004880905, 'learning_rate/qwen_vl_interface': 8.894360031062043e-06, 'learning_rate/action_model': 8.894360031062043e-05, 'epoch': 0.46}) +[2026-06-15 22:14:33,940][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.20143437385559082, 'train/grad_norm_pre_clip': 21.495136260986328, 'timing/data': 0.02560289017856121, 'timing/model': 1.4881192985922098, 'learning_rate/qwen_vl_interface': 8.892406047364783e-06, 'learning_rate/action_model': 8.892406047364784e-05, 'epoch': 0.46}) +[2026-06-15 22:14:40,362][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.31660011410713196, 'train/grad_norm_pre_clip': 12.28442668914795, 'timing/data': 0.012824960052967072, 'timing/model': 1.976612040773034, 'learning_rate/qwen_vl_interface': 8.890450566411984e-06, 'learning_rate/action_model': 8.890450566411984e-05, 'epoch': 0.46}) +[2026-06-15 22:14:46,665][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.17610692977905273, 'train/grad_norm_pre_clip': 10.574344635009766, 'timing/data': 0.015168683603405952, 'timing/model': 1.7730159610509872, 'learning_rate/qwen_vl_interface': 8.888493589007467e-06, 'learning_rate/action_model': 8.888493589007468e-05, 'epoch': 0.46}) +[2026-06-15 22:14:53,052][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.10689948499202728, 'train/grad_norm_pre_clip': 11.77894401550293, 'timing/data': 0.016067756339907646, 'timing/model': 1.7691689357161522, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 0.46}) +[2026-06-15 22:14:59,632][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.09577330946922302, 'train/grad_norm_pre_clip': 8.002193450927734, 'timing/data': 0.018469292670488358, 'timing/model': 1.6685915403068066, 'learning_rate/qwen_vl_interface': 8.884575148061662e-06, 'learning_rate/action_model': 8.88457514806166e-05, 'epoch': 0.46}) +[2026-06-15 22:15:06,681][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.1486748307943344, 'train/grad_norm_pre_clip': 11.89115047454834, 'timing/data': 0.021332846954464912, 'timing/model': 1.7616830728948116, 'learning_rate/qwen_vl_interface': 8.882613686131095e-06, 'learning_rate/action_model': 8.882613686131093e-05, 'epoch': 0.46}) +[2026-06-15 22:15:13,610][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.042438358068466187, 'train/grad_norm_pre_clip': 5.4318952560424805, 'timing/data': 0.022934002801775932, 'timing/model': 1.7143771033734083, 'learning_rate/qwen_vl_interface': 8.880650730970254e-06, 'learning_rate/action_model': 8.880650730970254e-05, 'epoch': 0.46}) +[2026-06-15 22:15:20,018][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.254883348941803, 'train/grad_norm_pre_clip': 12.672637939453125, 'timing/data': 0.013927139341831207, 'timing/model': 1.79884366504848, 'learning_rate/qwen_vl_interface': 8.878686283386042e-06, 'learning_rate/action_model': 8.878686283386042e-05, 'epoch': 0.46}) +[2026-06-15 22:15:26,074][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.020125756040215492, 'train/grad_norm_pre_clip': 2.6115410327911377, 'timing/data': 0.013777736574411392, 'timing/model': 1.5552258267998695, 'learning_rate/qwen_vl_interface': 8.876720344185967e-06, 'learning_rate/action_model': 8.876720344185967e-05, 'epoch': 0.46}) +[2026-06-15 22:15:32,374][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.21753227710723877, 'train/grad_norm_pre_clip': 8.551061630249023, 'timing/data': 0.013441404327750206, 'timing/model': 1.5584332887083292, 'learning_rate/qwen_vl_interface': 8.874752914178149e-06, 'learning_rate/action_model': 8.87475291417815e-05, 'epoch': 0.46}) +[2026-06-15 22:15:38,676][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.21642915904521942, 'train/grad_norm_pre_clip': 10.352214813232422, 'timing/data': 0.012301713228225708, 'timing/model': 1.5567046664655209, 'learning_rate/qwen_vl_interface': 8.87278399417133e-06, 'learning_rate/action_model': 8.872783994171329e-05, 'epoch': 0.46}) +[2026-06-15 22:15:45,075][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.31049802899360657, 'train/grad_norm_pre_clip': 11.342724800109863, 'timing/data': 0.012230079621076584, 'timing/model': 1.5865212697535753, 'learning_rate/qwen_vl_interface': 8.870813584974852e-06, 'learning_rate/action_model': 8.870813584974852e-05, 'epoch': 0.46}) +[2026-06-15 22:15:51,751][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.05406704545021057, 'train/grad_norm_pre_clip': 4.43029260635376, 'timing/data': 0.01753590814769268, 'timing/model': 1.6322914734482765, 'learning_rate/qwen_vl_interface': 8.868841687398683e-06, 'learning_rate/action_model': 8.868841687398683e-05, 'epoch': 0.46}) +[2026-06-15 22:23:32,751][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.3236440420150757, 'train/grad_norm_pre_clip': 9.355981826782227, 'eval/loss': 0.08176028728485107, 'eval/flap_precision': 0.8937259923175416, 'eval/flap_recall': 0.9470827679782904, 'eval/flap_f1': 0.919631093544137, 'eval/flap_rate_model': 0.21694444444444444, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9196310935441369, 'timing/data': 0.018079649657011032, 'timing/model': 1.9573477618396282, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 0.46}) +[2026-06-15 22:23:32,751][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1200_state +[2026-06-15 22:23:32,767][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 22:23:38,820][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1200_state/model.safetensors +[2026-06-15 22:23:52,505][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1200_state/optimizer.bin +[2026-06-15 22:23:52,506][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1200_state/sampler.bin +[2026-06-15 22:23:52,506][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1200_state/sampler_1.bin +[2026-06-15 22:23:52,507][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1200_state/random_states_0.pkl +[2026-06-15 22:23:52,610][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-06-15 22:23:52,611][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 22:23:52,619][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 22:23:58,550][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.043152548372745514, 'train/grad_norm_pre_clip': 7.6008524894714355, 'timing/data': 0.010893257334828377, 'timing/model': 1.7000591978430748, 'learning_rate/qwen_vl_interface': 8.864893430350159e-06, 'learning_rate/action_model': 8.864893430350159e-05, 'epoch': 0.46}) +[2026-06-15 22:24:04,797][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.248020201921463, 'train/grad_norm_pre_clip': 7.329273223876953, 'timing/data': 0.011635653674602509, 'timing/model': 2.0108118318021297, 'learning_rate/qwen_vl_interface': 8.862917072500785e-06, 'learning_rate/action_model': 8.862917072500784e-05, 'epoch': 0.47}) +[2026-06-15 22:24:11,528][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.11482028663158417, 'train/grad_norm_pre_clip': 8.372306823730469, 'timing/data': 0.015169858932495117, 'timing/model': 1.5933280009776354, 'learning_rate/qwen_vl_interface': 8.860939229517674e-06, 'learning_rate/action_model': 8.860939229517674e-05, 'epoch': 0.47}) +[2026-06-15 22:24:17,964][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.08075329661369324, 'train/grad_norm_pre_clip': 6.492944240570068, 'timing/data': 0.013056579977273941, 'timing/model': 1.5537959691137075, 'learning_rate/qwen_vl_interface': 8.858959902213844e-06, 'learning_rate/action_model': 8.858959902213845e-05, 'epoch': 0.47}) +[2026-06-15 22:24:24,298][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.07105744630098343, 'train/grad_norm_pre_clip': 4.42965841293335, 'timing/data': 0.01301594264805317, 'timing/model': 1.5648283511400223, 'learning_rate/qwen_vl_interface': 8.856979091402921e-06, 'learning_rate/action_model': 8.85697909140292e-05, 'epoch': 0.47}) +[2026-06-15 22:24:30,573][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.17770153284072876, 'train/grad_norm_pre_clip': 8.083613395690918, 'timing/data': 0.012975608929991722, 'timing/model': 1.5858581829816103, 'learning_rate/qwen_vl_interface': 8.854996797899139e-06, 'learning_rate/action_model': 8.854996797899139e-05, 'epoch': 0.47}) +[2026-06-15 22:24:36,891][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.1471114605665207, 'train/grad_norm_pre_clip': 3.527350425720215, 'timing/data': 0.012093598023056984, 'timing/model': 1.77193814329803, 'learning_rate/qwen_vl_interface': 8.853013022517348e-06, 'learning_rate/action_model': 8.853013022517348e-05, 'epoch': 0.47}) +[2026-06-15 22:24:43,070][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.14930979907512665, 'train/grad_norm_pre_clip': 8.805161476135254, 'timing/data': 0.01144016906619072, 'timing/model': 1.567890552803874, 'learning_rate/qwen_vl_interface': 8.851027766073002e-06, 'learning_rate/action_model': 8.851027766073001e-05, 'epoch': 0.47}) +[2026-06-15 22:24:49,480][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.26640763878822327, 'train/grad_norm_pre_clip': 9.858766555786133, 'timing/data': 0.013077616691589355, 'timing/model': 1.5972102731466293, 'learning_rate/qwen_vl_interface': 8.84904102938216e-06, 'learning_rate/action_model': 8.84904102938216e-05, 'epoch': 0.47}) +[2026-06-15 22:24:55,738][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.12861867249011993, 'train/grad_norm_pre_clip': 15.660653114318848, 'timing/data': 0.01268845982849598, 'timing/model': 1.5610289089381695, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 0.47}) +[2026-06-15 22:25:01,846][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.09151424467563629, 'train/grad_norm_pre_clip': 7.5341997146606445, 'timing/data': 0.011502450332045555, 'timing/model': 1.5344814974814653, 'learning_rate/qwen_vl_interface': 8.845063118528303e-06, 'learning_rate/action_model': 8.845063118528304e-05, 'epoch': 0.47}) +[2026-06-15 22:25:07,681][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.2229962944984436, 'train/grad_norm_pre_clip': 6.712004661560059, 'timing/data': 0.010974464938044548, 'timing/model': 1.4184252675622702, 'learning_rate/qwen_vl_interface': 8.843071946000455e-06, 'learning_rate/action_model': 8.843071946000456e-05, 'epoch': 0.47}) +[2026-06-15 22:25:13,591][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.12703940272331238, 'train/grad_norm_pre_clip': 10.057852745056152, 'timing/data': 0.010378405451774597, 'timing/model': 1.6695004738867283, 'learning_rate/qwen_vl_interface': 8.841079296496451e-06, 'learning_rate/action_model': 8.84107929649645e-05, 'epoch': 0.47}) +[2026-06-15 22:25:19,478][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.49015259742736816, 'train/grad_norm_pre_clip': 9.445720672607422, 'timing/data': 0.011351434513926506, 'timing/model': 1.4547981582581997, 'learning_rate/qwen_vl_interface': 8.839085170835396e-06, 'learning_rate/action_model': 8.839085170835395e-05, 'epoch': 0.47}) +[2026-06-15 22:25:25,393][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.11853227019309998, 'train/grad_norm_pre_clip': 8.274369239807129, 'timing/data': 0.01013149693608284, 'timing/model': 1.4628061205148697, 'learning_rate/qwen_vl_interface': 8.837089569837e-06, 'learning_rate/action_model': 8.837089569836998e-05, 'epoch': 0.47}) +[2026-06-15 22:25:31,382][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.11852246522903442, 'train/grad_norm_pre_clip': 5.170013427734375, 'timing/data': 0.009582454338669777, 'timing/model': 1.4489820282906294, 'learning_rate/qwen_vl_interface': 8.835092494321575e-06, 'learning_rate/action_model': 8.835092494321575e-05, 'epoch': 0.47}) +[2026-06-15 22:25:37,633][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.08927418291568756, 'train/grad_norm_pre_clip': 6.334549427032471, 'timing/data': 0.0117568951100111, 'timing/model': 1.5529438443481922, 'learning_rate/qwen_vl_interface': 8.833093945110048e-06, 'learning_rate/action_model': 8.833093945110047e-05, 'epoch': 0.47}) +[2026-06-15 22:25:43,741][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.11711522191762924, 'train/grad_norm_pre_clip': 6.903232574462891, 'timing/data': 0.012242008000612259, 'timing/model': 1.5438137110322714, 'learning_rate/qwen_vl_interface': 8.831093923023946e-06, 'learning_rate/action_model': 8.831093923023945e-05, 'epoch': 0.47}) +[2026-06-15 22:25:49,987][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.042206816375255585, 'train/grad_norm_pre_clip': 4.087066173553467, 'timing/data': 0.014416413381695747, 'timing/model': 1.8019531443715096, 'learning_rate/qwen_vl_interface': 8.8290924288854e-06, 'learning_rate/action_model': 8.8290924288854e-05, 'epoch': 0.47}) +[2026-06-15 22:25:56,136][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.17619313299655914, 'train/grad_norm_pre_clip': 9.872551918029785, 'timing/data': 0.012663954868912697, 'timing/model': 1.554731260985136, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.47}) +[2026-06-15 22:26:02,253][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.34235790371894836, 'train/grad_norm_pre_clip': 14.060982704162598, 'timing/data': 0.012556703761219978, 'timing/model': 1.485767886042595, 'learning_rate/qwen_vl_interface': 8.825085027742545e-06, 'learning_rate/action_model': 8.825085027742544e-05, 'epoch': 0.47}) +[2026-06-15 22:26:09,147][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.3046596050262451, 'train/grad_norm_pre_clip': 7.762209892272949, 'timing/data': 0.011670006439089775, 'timing/model': 2.2654113322496414, 'learning_rate/qwen_vl_interface': 8.823079122385523e-06, 'learning_rate/action_model': 8.823079122385523e-05, 'epoch': 0.47}) +[2026-06-15 22:26:15,340][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.017662033438682556, 'train/grad_norm_pre_clip': 2.7276382446289062, 'timing/data': 0.011504625901579857, 'timing/model': 1.5545708201825619, 'learning_rate/qwen_vl_interface': 8.821071748270644e-06, 'learning_rate/action_model': 8.821071748270644e-05, 'epoch': 0.47}) +[2026-06-15 22:26:21,447][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.03990565612912178, 'train/grad_norm_pre_clip': 5.129390239715576, 'timing/data': 0.01227717287838459, 'timing/model': 1.5018505025655031, 'learning_rate/qwen_vl_interface': 8.81906290622306e-06, 'learning_rate/action_model': 8.81906290622306e-05, 'epoch': 0.47}) +[2026-06-15 22:26:27,700][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.307589054107666, 'train/grad_norm_pre_clip': 11.310659408569336, 'timing/data': 0.012774132192134857, 'timing/model': 1.7947584483772516, 'learning_rate/qwen_vl_interface': 8.81705259706853e-06, 'learning_rate/action_model': 8.817052597068529e-05, 'epoch': 0.47}) +[2026-06-15 22:26:33,740][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.08606042712926865, 'train/grad_norm_pre_clip': 5.679851531982422, 'timing/data': 0.01265755109488964, 'timing/model': 1.5573637448251247, 'learning_rate/qwen_vl_interface': 8.815040821633416e-06, 'learning_rate/action_model': 8.815040821633416e-05, 'epoch': 0.47}) +[2026-06-15 22:26:40,146][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.07565314322710037, 'train/grad_norm_pre_clip': 5.246359825134277, 'timing/data': 0.012295674532651901, 'timing/model': 1.5703238304704428, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.47}) +[2026-06-15 22:26:46,335][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.17600509524345398, 'train/grad_norm_pre_clip': 8.11907958984375, 'timing/data': 0.01368601806461811, 'timing/model': 1.5545058064162731, 'learning_rate/qwen_vl_interface': 8.8110128752299e-06, 'learning_rate/action_model': 8.8110128752299e-05, 'epoch': 0.48}) +[2026-06-15 22:26:52,466][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.045797199010849, 'train/grad_norm_pre_clip': 4.508969306945801, 'timing/data': 0.012439819052815437, 'timing/model': 1.50086067058146, 'learning_rate/qwen_vl_interface': 8.808996705917235e-06, 'learning_rate/action_model': 8.808996705917235e-05, 'epoch': 0.48}) +[2026-06-15 22:26:58,658][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.22089911997318268, 'train/grad_norm_pre_clip': 6.490782260894775, 'timing/data': 0.01129171997308731, 'timing/model': 1.5563355647027493, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 0.48}) +[2026-06-15 22:27:04,804][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.3765200972557068, 'train/grad_norm_pre_clip': 18.09420394897461, 'timing/data': 0.01234966330230236, 'timing/model': 1.746870445087552, 'learning_rate/qwen_vl_interface': 8.804959979213945e-06, 'learning_rate/action_model': 8.804959979213945e-05, 'epoch': 0.48}) +[2026-06-15 22:27:10,848][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.2921564280986786, 'train/grad_norm_pre_clip': 16.417842864990234, 'timing/data': 0.012371031567454338, 'timing/model': 1.5713081620633602, 'learning_rate/qwen_vl_interface': 8.802939423482667e-06, 'learning_rate/action_model': 8.802939423482666e-05, 'epoch': 0.48}) +[2026-06-15 22:27:17,244][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.412072092294693, 'train/grad_norm_pre_clip': 14.790180206298828, 'timing/data': 0.01299702376127243, 'timing/model': 1.6564561277627945, 'learning_rate/qwen_vl_interface': 8.800917407272198e-06, 'learning_rate/action_model': 8.800917407272198e-05, 'epoch': 0.48}) +[2026-06-15 22:27:23,193][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.07957375049591064, 'train/grad_norm_pre_clip': 4.441450119018555, 'timing/data': 0.010994156822562218, 'timing/model': 1.4553934149444103, 'learning_rate/qwen_vl_interface': 8.798893931413711e-06, 'learning_rate/action_model': 8.79889393141371e-05, 'epoch': 0.48}) +[2026-06-15 22:27:29,175][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.2598678767681122, 'train/grad_norm_pre_clip': 7.772704124450684, 'timing/data': 0.010485192760825157, 'timing/model': 1.545551961287856, 'learning_rate/qwen_vl_interface': 8.796868996738985e-06, 'learning_rate/action_model': 8.796868996738984e-05, 'epoch': 0.48}) +[2026-06-15 22:27:35,097][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.05018078163266182, 'train/grad_norm_pre_clip': 3.7029755115509033, 'timing/data': 0.010587736964225769, 'timing/model': 1.5038898009806871, 'learning_rate/qwen_vl_interface': 8.794842604080389e-06, 'learning_rate/action_model': 8.794842604080389e-05, 'epoch': 0.48}) +[2026-06-15 22:27:40,995][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.0625862181186676, 'train/grad_norm_pre_clip': 2.617886543273926, 'timing/data': 0.01046256348490715, 'timing/model': 1.675595449283719, 'learning_rate/qwen_vl_interface': 8.7928147542709e-06, 'learning_rate/action_model': 8.7928147542709e-05, 'epoch': 0.48}) +[2026-06-15 22:27:46,688][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.17562849819660187, 'train/grad_norm_pre_clip': 5.38173770904541, 'timing/data': 0.011384505778551102, 'timing/model': 1.4387630801647902, 'learning_rate/qwen_vl_interface': 8.790785448144088e-06, 'learning_rate/action_model': 8.790785448144088e-05, 'epoch': 0.48}) +[2026-06-15 22:27:52,589][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.10711871087551117, 'train/grad_norm_pre_clip': 6.3868727684021, 'timing/data': 0.010830124840140343, 'timing/model': 1.455915568396449, 'learning_rate/qwen_vl_interface': 8.788754686534126e-06, 'learning_rate/action_model': 8.788754686534127e-05, 'epoch': 0.48}) +[2026-06-15 22:27:58,487][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.03826384246349335, 'train/grad_norm_pre_clip': 3.3230819702148438, 'timing/data': 0.010941000655293465, 'timing/model': 1.4582813642919064, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 0.48}) +[2026-06-15 22:28:04,390][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.14838825166225433, 'train/grad_norm_pre_clip': 9.096662521362305, 'timing/data': 0.011122342199087143, 'timing/model': 1.458984375, 'learning_rate/qwen_vl_interface': 8.784688800204427e-06, 'learning_rate/action_model': 8.784688800204426e-05, 'epoch': 0.48}) +[2026-06-15 22:28:10,294][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.24667324125766754, 'train/grad_norm_pre_clip': 8.763570785522461, 'timing/data': 0.011221861466765404, 'timing/model': 1.4287501014769077, 'learning_rate/qwen_vl_interface': 8.782653677156022e-06, 'learning_rate/action_model': 8.782653677156022e-05, 'epoch': 0.48}) +[2026-06-15 22:28:16,200][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.11897562444210052, 'train/grad_norm_pre_clip': 5.966414451599121, 'timing/data': 0.011002115905284882, 'timing/model': 1.6528879906982183, 'learning_rate/qwen_vl_interface': 8.780617101967129e-06, 'learning_rate/action_model': 8.780617101967129e-05, 'epoch': 0.48}) +[2026-06-15 22:28:21,895][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.2943439483642578, 'train/grad_norm_pre_clip': 9.227704048156738, 'timing/data': 0.010971296578645706, 'timing/model': 1.4693233389407396, 'learning_rate/qwen_vl_interface': 8.77857907547491e-06, 'learning_rate/action_model': 8.77857907547491e-05, 'epoch': 0.48}) +[2026-06-15 22:28:27,795][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.022245796397328377, 'train/grad_norm_pre_clip': 1.655242919921875, 'timing/data': 0.012198982760310173, 'timing/model': 1.468762455508113, 'learning_rate/qwen_vl_interface': 8.776539598517119e-06, 'learning_rate/action_model': 8.776539598517118e-05, 'epoch': 0.48}) +[2026-06-15 22:28:33,695][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.08285222947597504, 'train/grad_norm_pre_clip': 6.474428653717041, 'timing/data': 0.010257216170430183, 'timing/model': 1.466475959867239, 'learning_rate/qwen_vl_interface': 8.774498671932108e-06, 'learning_rate/action_model': 8.774498671932107e-05, 'epoch': 0.48}) +[2026-06-15 22:28:39,695][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.13847433030605316, 'train/grad_norm_pre_clip': 5.479798316955566, 'timing/data': 0.010819068178534508, 'timing/model': 1.454404305666685, 'learning_rate/qwen_vl_interface': 8.772456296558824e-06, 'learning_rate/action_model': 8.772456296558824e-05, 'epoch': 0.48}) +[2026-06-15 22:28:45,606][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.1449953317642212, 'train/grad_norm_pre_clip': 7.9841461181640625, 'timing/data': 0.010767215862870216, 'timing/model': 1.4363651555031538, 'learning_rate/qwen_vl_interface': 8.77041247323681e-06, 'learning_rate/action_model': 8.770412473236811e-05, 'epoch': 0.48}) +[2026-06-15 22:28:52,302][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.1476166695356369, 'train/grad_norm_pre_clip': 8.05355167388916, 'timing/data': 0.01191037893295288, 'timing/model': 2.1040855329483747, 'learning_rate/qwen_vl_interface': 8.76836720280621e-06, 'learning_rate/action_model': 8.768367202806209e-05, 'epoch': 0.48}) +[2026-06-15 22:28:59,128][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.06853241473436356, 'train/grad_norm_pre_clip': 3.7984321117401123, 'timing/data': 0.01469324342906475, 'timing/model': 1.7039370518177748, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 0.48}) +[2026-06-15 22:29:05,147][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.0382288321852684, 'train/grad_norm_pre_clip': 3.303600788116455, 'timing/data': 0.01267351396381855, 'timing/model': 1.5617683697491884, 'learning_rate/qwen_vl_interface': 8.76427232398276e-06, 'learning_rate/action_model': 8.76427232398276e-05, 'epoch': 0.48}) +[2026-06-15 22:29:11,346][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.038074344396591187, 'train/grad_norm_pre_clip': 3.9007389545440674, 'timing/data': 0.016575945541262627, 'timing/model': 1.5574673172086477, 'learning_rate/qwen_vl_interface': 8.762222717273167e-06, 'learning_rate/action_model': 8.762222717273167e-05, 'epoch': 0.48}) +[2026-06-15 22:29:17,544][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.10299619287252426, 'train/grad_norm_pre_clip': 9.506675720214844, 'timing/data': 0.013092609122395515, 'timing/model': 1.551622949540615, 'learning_rate/qwen_vl_interface': 8.76017166682148e-06, 'learning_rate/action_model': 8.760171666821481e-05, 'epoch': 0.48}) +[2026-06-15 22:29:24,140][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.13260217010974884, 'train/grad_norm_pre_clip': 8.882380485534668, 'timing/data': 0.013051904737949371, 'timing/model': 1.6030129734426737, 'learning_rate/qwen_vl_interface': 8.758119173470815e-06, 'learning_rate/action_model': 8.758119173470815e-05, 'epoch': 0.49}) +[2026-06-15 22:29:30,591][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.13781240582466125, 'train/grad_norm_pre_clip': 7.108937740325928, 'timing/data': 0.011587493121623993, 'timing/model': 1.808355636894703, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 0.49}) +[2026-06-15 22:29:36,739][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.20837129652500153, 'train/grad_norm_pre_clip': 4.325923442840576, 'timing/data': 0.017926625907421112, 'timing/model': 1.5946349427103996, 'learning_rate/qwen_vl_interface': 8.75400986144794e-06, 'learning_rate/action_model': 8.75400986144794e-05, 'epoch': 0.49}) +[2026-06-15 22:29:43,206][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.01282090600579977, 'train/grad_norm_pre_clip': 2.3981149196624756, 'timing/data': 0.015001365914940834, 'timing/model': 1.5593612305819988, 'learning_rate/qwen_vl_interface': 8.751953044464913e-06, 'learning_rate/action_model': 8.751953044464913e-05, 'epoch': 0.49}) +[2026-06-15 22:29:49,976][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.3113170564174652, 'train/grad_norm_pre_clip': 15.149018287658691, 'timing/data': 0.01182224415242672, 'timing/model': 1.5600492525845766, 'learning_rate/qwen_vl_interface': 8.749894787961272e-06, 'learning_rate/action_model': 8.749894787961271e-05, 'epoch': 0.49}) +[2026-06-15 22:29:56,287][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.0742950588464737, 'train/grad_norm_pre_clip': 12.944169998168945, 'timing/data': 0.013561643660068512, 'timing/model': 1.5255790408700705, 'learning_rate/qwen_vl_interface': 8.747835092783086e-06, 'learning_rate/action_model': 8.747835092783085e-05, 'epoch': 0.49}) +[2026-06-15 22:30:02,487][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.025172999128699303, 'train/grad_norm_pre_clip': 4.6912126541137695, 'timing/data': 0.012732703238725662, 'timing/model': 1.5286486651748419, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 0.49}) +[2026-06-15 22:30:08,819][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.06876841932535172, 'train/grad_norm_pre_clip': 11.721890449523926, 'timing/data': 0.012638412415981293, 'timing/model': 1.801322216168046, 'learning_rate/qwen_vl_interface': 8.743711389790321e-06, 'learning_rate/action_model': 8.74371138979032e-05, 'epoch': 0.49}) +[2026-06-15 22:30:14,897][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.4551744759082794, 'train/grad_norm_pre_clip': 16.767702102661133, 'timing/data': 0.012569870799779892, 'timing/model': 1.5840121060609818, 'learning_rate/qwen_vl_interface': 8.741647383670839e-06, 'learning_rate/action_model': 8.741647383670839e-05, 'epoch': 0.49}) +[2026-06-15 22:30:21,297][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.36694440245628357, 'train/grad_norm_pre_clip': 21.76354217529297, 'timing/data': 0.01231350190937519, 'timing/model': 1.5368597563356161, 'learning_rate/qwen_vl_interface': 8.739581942267011e-06, 'learning_rate/action_model': 8.739581942267011e-05, 'epoch': 0.49}) +[2026-06-15 22:30:27,685][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.21566317975521088, 'train/grad_norm_pre_clip': 16.69776725769043, 'timing/data': 0.011603562161326408, 'timing/model': 1.5400082636624575, 'learning_rate/qwen_vl_interface': 8.737515066427861e-06, 'learning_rate/action_model': 8.73751506642786e-05, 'epoch': 0.49}) +[2026-06-15 22:30:34,091][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.07441018521785736, 'train/grad_norm_pre_clip': 3.8714466094970703, 'timing/data': 0.011996414512395859, 'timing/model': 1.6071395855396986, 'learning_rate/qwen_vl_interface': 8.735446757003e-06, 'learning_rate/action_model': 8.735446757003e-05, 'epoch': 0.49}) +[2026-06-15 22:30:40,340][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.20319657027721405, 'train/grad_norm_pre_clip': 12.521629333496094, 'timing/data': 0.01260978914797306, 'timing/model': 1.5613477621227503, 'learning_rate/qwen_vl_interface': 8.733377014842635e-06, 'learning_rate/action_model': 8.733377014842635e-05, 'epoch': 0.49}) +[2026-06-15 22:30:46,552][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.03002486191689968, 'train/grad_norm_pre_clip': 3.8309760093688965, 'timing/data': 0.012919776141643524, 'timing/model': 1.7410355862230062, 'learning_rate/qwen_vl_interface': 8.731305840797557e-06, 'learning_rate/action_model': 8.731305840797558e-05, 'epoch': 0.49}) +[2026-06-15 22:30:52,769][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.13942940533161163, 'train/grad_norm_pre_clip': 8.256369590759277, 'timing/data': 0.012019338086247444, 'timing/model': 1.6786536816507578, 'learning_rate/qwen_vl_interface': 8.72923323571915e-06, 'learning_rate/action_model': 8.72923323571915e-05, 'epoch': 0.49}) +[2026-06-15 22:30:59,286][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.412945955991745, 'train/grad_norm_pre_clip': 13.225098609924316, 'timing/data': 0.01430915854871273, 'timing/model': 1.5729802809655666, 'learning_rate/qwen_vl_interface': 8.727159200459381e-06, 'learning_rate/action_model': 8.727159200459381e-05, 'epoch': 0.49}) +[2026-06-15 22:31:05,685][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.2141832709312439, 'train/grad_norm_pre_clip': 7.402444362640381, 'timing/data': 0.01177867129445076, 'timing/model': 1.5499015804380178, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 0.49}) +[2026-06-15 22:31:12,074][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.047749683260917664, 'train/grad_norm_pre_clip': 3.356980562210083, 'timing/data': 0.011704003438353539, 'timing/model': 1.5727336090058088, 'learning_rate/qwen_vl_interface': 8.723006842806575e-06, 'learning_rate/action_model': 8.723006842806576e-05, 'epoch': 0.49}) +[2026-06-15 22:31:18,410][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.17431066930294037, 'train/grad_norm_pre_clip': 12.806414604187012, 'timing/data': 0.011402307078242302, 'timing/model': 1.7925974018871784, 'learning_rate/qwen_vl_interface': 8.720928522120419e-06, 'learning_rate/action_model': 8.720928522120418e-05, 'epoch': 0.49}) +[2026-06-15 22:31:24,472][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.16795876622200012, 'train/grad_norm_pre_clip': 11.039758682250977, 'timing/data': 0.011688213795423508, 'timing/model': 1.542356699705124, 'learning_rate/qwen_vl_interface': 8.718848774666655e-06, 'learning_rate/action_model': 8.718848774666655e-05, 'epoch': 0.49}) +[2026-06-15 22:31:30,971][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.30023542046546936, 'train/grad_norm_pre_clip': 19.673500061035156, 'timing/data': 0.01322358101606369, 'timing/model': 1.5529594887048006, 'learning_rate/qwen_vl_interface': 8.716767601300189e-06, 'learning_rate/action_model': 8.716767601300188e-05, 'epoch': 0.49}) +[2026-06-15 22:31:37,171][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.27902933955192566, 'train/grad_norm_pre_clip': 8.08282470703125, 'timing/data': 0.01241990551352501, 'timing/model': 1.5520701091736555, 'learning_rate/qwen_vl_interface': 8.714685002876514e-06, 'learning_rate/action_model': 8.714685002876513e-05, 'epoch': 0.49}) +[2026-06-15 22:31:43,371][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.12255099415779114, 'train/grad_norm_pre_clip': 7.617125511169434, 'timing/data': 0.012782899662852287, 'timing/model': 1.5465116389095783, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.49}) +[2026-06-15 22:31:49,614][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.12461046874523163, 'train/grad_norm_pre_clip': 6.773924827575684, 'timing/data': 0.010693488642573357, 'timing/model': 1.5153119675815105, 'learning_rate/qwen_vl_interface': 8.710515534282428e-06, 'learning_rate/action_model': 8.710515534282428e-05, 'epoch': 0.49}) +[2026-06-15 22:31:55,522][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.15068411827087402, 'train/grad_norm_pre_clip': 6.70623254776001, 'timing/data': 0.01078350655734539, 'timing/model': 1.6644397526979446, 'learning_rate/qwen_vl_interface': 8.708428665825931e-06, 'learning_rate/action_model': 8.708428665825931e-05, 'epoch': 0.49}) +[2026-06-15 22:32:01,182][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.2405046969652176, 'train/grad_norm_pre_clip': 11.763635635375977, 'timing/data': 0.010473240166902542, 'timing/model': 1.4579712487757206, 'learning_rate/qwen_vl_interface': 8.706340375740045e-06, 'learning_rate/action_model': 8.706340375740045e-05, 'epoch': 0.49}) +[2026-06-15 22:32:07,086][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.08111243695020676, 'train/grad_norm_pre_clip': 3.594348669052124, 'timing/data': 0.010334327816963196, 'timing/model': 1.454611824825406, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 0.5}) +[2026-06-15 22:32:13,277][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.08593310415744781, 'train/grad_norm_pre_clip': 5.932294845581055, 'timing/data': 0.010517949238419533, 'timing/model': 1.749726064503193, 'learning_rate/qwen_vl_interface': 8.70215953411436e-06, 'learning_rate/action_model': 8.70215953411436e-05, 'epoch': 0.5}) +[2026-06-15 22:32:19,191][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.3196403384208679, 'train/grad_norm_pre_clip': 11.985642433166504, 'timing/data': 0.010172789916396141, 'timing/model': 1.4644648227840662, 'learning_rate/qwen_vl_interface': 8.70006698429315e-06, 'learning_rate/action_model': 8.700066984293149e-05, 'epoch': 0.5}) +[2026-06-15 22:32:25,090][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.20411741733551025, 'train/grad_norm_pre_clip': 10.976678848266602, 'timing/data': 0.010443441569805145, 'timing/model': 1.4861762505024672, 'learning_rate/qwen_vl_interface': 8.697973016279722e-06, 'learning_rate/action_model': 8.697973016279722e-05, 'epoch': 0.5}) +[2026-06-15 22:32:30,995][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.16327853500843048, 'train/grad_norm_pre_clip': 6.687685012817383, 'timing/data': 0.009772483259439468, 'timing/model': 1.6664730627089739, 'learning_rate/qwen_vl_interface': 8.695877630934827e-06, 'learning_rate/action_model': 8.695877630934826e-05, 'epoch': 0.5}) +[2026-06-15 22:32:36,945][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.08124590665102005, 'train/grad_norm_pre_clip': 13.2590913772583, 'timing/data': 0.01201414130628109, 'timing/model': 1.5333735328167677, 'learning_rate/qwen_vl_interface': 8.693780829119801e-06, 'learning_rate/action_model': 8.693780829119801e-05, 'epoch': 0.5}) +[2026-06-15 22:32:43,142][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.17585912346839905, 'train/grad_norm_pre_clip': 7.076813220977783, 'timing/data': 0.012180408462882042, 'timing/model': 1.5268779788166285, 'learning_rate/qwen_vl_interface': 8.69168261169656e-06, 'learning_rate/action_model': 8.691682611696559e-05, 'epoch': 0.5}) +[2026-06-15 22:32:49,606][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.30304446816444397, 'train/grad_norm_pre_clip': 9.183507919311523, 'timing/data': 0.014880532398819923, 'timing/model': 1.552585519850254, 'learning_rate/qwen_vl_interface': 8.689582979527599e-06, 'learning_rate/action_model': 8.689582979527599e-05, 'epoch': 0.5}) +[2026-06-15 22:32:56,332][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.16215060651302338, 'train/grad_norm_pre_clip': 7.125270843505859, 'timing/data': 0.019006945192813873, 'timing/model': 1.664216049015522, 'learning_rate/qwen_vl_interface': 8.687481933476e-06, 'learning_rate/action_model': 8.687481933476e-05, 'epoch': 0.5}) +[2026-06-15 22:33:02,570][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.38656142354011536, 'train/grad_norm_pre_clip': 23.435026168823242, 'timing/data': 0.012071896344423294, 'timing/model': 1.5235916301608086, 'learning_rate/qwen_vl_interface': 8.685379474405423e-06, 'learning_rate/action_model': 8.685379474405423e-05, 'epoch': 0.5}) +[2026-06-15 22:33:08,912][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.44026944041252136, 'train/grad_norm_pre_clip': 14.556169509887695, 'timing/data': 0.012106534093618393, 'timing/model': 1.777801088988781, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.5}) +[2026-06-15 22:33:14,993][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.09168922156095505, 'train/grad_norm_pre_clip': 4.88240909576416, 'timing/data': 0.013820579275488853, 'timing/model': 1.5501572620123625, 'learning_rate/qwen_vl_interface': 8.681170320664883e-06, 'learning_rate/action_model': 8.681170320664883e-05, 'epoch': 0.5}) +[2026-06-15 22:33:21,410][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.20692867040634155, 'train/grad_norm_pre_clip': 7.078794479370117, 'timing/data': 0.011760160326957703, 'timing/model': 1.555561751127243, 'learning_rate/qwen_vl_interface': 8.679063627725143e-06, 'learning_rate/action_model': 8.679063627725144e-05, 'epoch': 0.5}) +[2026-06-15 22:33:27,805][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.08231467008590698, 'train/grad_norm_pre_clip': 6.168626308441162, 'timing/data': 0.014070501551032066, 'timing/model': 1.5647691395133734, 'learning_rate/qwen_vl_interface': 8.676955525226876e-06, 'learning_rate/action_model': 8.676955525226875e-05, 'epoch': 0.5}) +[2026-06-15 22:33:33,897][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.24014395475387573, 'train/grad_norm_pre_clip': 9.194761276245117, 'timing/data': 0.010025175288319588, 'timing/model': 1.5071019139140844, 'learning_rate/qwen_vl_interface': 8.674846014036636e-06, 'learning_rate/action_model': 8.674846014036636e-05, 'epoch': 0.5}) +[2026-06-15 22:33:39,936][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.09583403170108795, 'train/grad_norm_pre_clip': 6.90585470199585, 'timing/data': 0.010985143482685089, 'timing/model': 1.4649527948349714, 'learning_rate/qwen_vl_interface': 8.672735095021569e-06, 'learning_rate/action_model': 8.672735095021569e-05, 'epoch': 0.5}) +[2026-06-15 22:33:46,442][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.3974423110485077, 'train/grad_norm_pre_clip': 10.292888641357422, 'timing/data': 0.014027716591954231, 'timing/model': 1.8959477003663778, 'learning_rate/qwen_vl_interface': 8.670622769049393e-06, 'learning_rate/action_model': 8.670622769049392e-05, 'epoch': 0.5}) +[2026-06-15 22:33:52,479][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.07066062837839127, 'train/grad_norm_pre_clip': 3.3021085262298584, 'timing/data': 0.013844551518559456, 'timing/model': 1.5367900524288416, 'learning_rate/qwen_vl_interface': 8.668509036988406e-06, 'learning_rate/action_model': 8.668509036988405e-05, 'epoch': 0.5}) +[2026-06-15 22:33:58,746][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.15438176691532135, 'train/grad_norm_pre_clip': 7.070822238922119, 'timing/data': 0.013316478580236435, 'timing/model': 1.5582708083093166, 'learning_rate/qwen_vl_interface': 8.66639389970748e-06, 'learning_rate/action_model': 8.666393899707479e-05, 'epoch': 0.5}) +[2026-06-15 22:34:04,847][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.12983663380146027, 'train/grad_norm_pre_clip': 7.1698760986328125, 'timing/data': 0.01342041976749897, 'timing/model': 1.554568987339735, 'learning_rate/qwen_vl_interface': 8.66427735807607e-06, 'learning_rate/action_model': 8.664277358076069e-05, 'epoch': 0.5}) +[2026-06-15 22:41:43,350][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.10563893616199493, 'train/grad_norm_pre_clip': 5.830327033996582, 'eval/loss': 0.07988234609365463, 'eval/flap_precision': 0.8919949174078781, 'eval/flap_recall': 0.9525101763907734, 'eval/flap_f1': 0.9212598425196852, 'eval/flap_rate_model': 0.21861111111111112, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9212598425196851, 'timing/data': 0.012227186933159828, 'timing/model': 1.5519299302250147, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 0.5}) +[2026-06-15 22:41:43,351][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1300_state +[2026-06-15 22:41:43,358][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 22:41:55,153][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1300_state/model.safetensors +[2026-06-15 22:42:12,183][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1300_state/optimizer.bin +[2026-06-15 22:42:12,184][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1300_state/sampler.bin +[2026-06-15 22:42:12,184][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1300_state/sampler_1.bin +[2026-06-15 22:42:12,185][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1300_state/random_states_0.pkl +[2026-06-15 22:42:12,274][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-06-15 22:42:12,274][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 22:42:12,282][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 22:42:18,524][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.2270655333995819, 'train/grad_norm_pre_clip': 8.684493064880371, 'timing/data': 0.012911058962345123, 'timing/model': 1.5268871765583754, 'learning_rate/qwen_vl_interface': 8.660040065242499e-06, 'learning_rate/action_model': 8.660040065242498e-05, 'epoch': 0.5}) +[2026-06-15 22:42:24,503][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.09885209798812866, 'train/grad_norm_pre_clip': 3.342791795730591, 'timing/data': 0.01052347756922245, 'timing/model': 1.4375427048653364, 'learning_rate/qwen_vl_interface': 8.657919315782127e-06, 'learning_rate/action_model': 8.657919315782127e-05, 'epoch': 0.5}) +[2026-06-15 22:42:30,400][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.18112173676490784, 'train/grad_norm_pre_clip': 10.54075813293457, 'timing/data': 0.011597616598010063, 'timing/model': 1.4622723385691643, 'learning_rate/qwen_vl_interface': 8.655797165454853e-06, 'learning_rate/action_model': 8.655797165454853e-05, 'epoch': 0.5}) +[2026-06-15 22:42:36,311][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.31907087564468384, 'train/grad_norm_pre_clip': 28.868118286132812, 'timing/data': 0.009849347174167633, 'timing/model': 1.6617076825350523, 'learning_rate/qwen_vl_interface': 8.653673615133013e-06, 'learning_rate/action_model': 8.653673615133013e-05, 'epoch': 0.5}) +[2026-06-15 22:42:42,002][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.1513098180294037, 'train/grad_norm_pre_clip': 8.268166542053223, 'timing/data': 0.011939145624637604, 'timing/model': 1.450796877965331, 'learning_rate/qwen_vl_interface': 8.651548665689519e-06, 'learning_rate/action_model': 8.651548665689518e-05, 'epoch': 0.51}) +[2026-06-15 22:42:47,913][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.33123311400413513, 'train/grad_norm_pre_clip': 8.40202522277832, 'timing/data': 0.012196080759167671, 'timing/model': 1.4593223333358765, 'learning_rate/qwen_vl_interface': 8.649422317997855e-06, 'learning_rate/action_model': 8.649422317997856e-05, 'epoch': 0.51}) +[2026-06-15 22:42:54,076][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.3913556635379791, 'train/grad_norm_pre_clip': 14.179412841796875, 'timing/data': 0.013275595381855965, 'timing/model': 1.5599605534225702, 'learning_rate/qwen_vl_interface': 8.647294572932082e-06, 'learning_rate/action_model': 8.647294572932082e-05, 'epoch': 0.51}) +[2026-06-15 22:43:00,372][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.2571222484111786, 'train/grad_norm_pre_clip': 15.404706001281738, 'timing/data': 0.015431711450219154, 'timing/model': 1.5235987417399883, 'learning_rate/qwen_vl_interface': 8.64516543136684e-06, 'learning_rate/action_model': 8.64516543136684e-05, 'epoch': 0.51}) +[2026-06-15 22:43:06,768][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.17350253462791443, 'train/grad_norm_pre_clip': 10.961291313171387, 'timing/data': 0.014322338625788689, 'timing/model': 1.5151793472468853, 'learning_rate/qwen_vl_interface': 8.643034894177335e-06, 'learning_rate/action_model': 8.643034894177335e-05, 'epoch': 0.51}) +[2026-06-15 22:43:13,132][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.12547528743743896, 'train/grad_norm_pre_clip': 6.907463073730469, 'timing/data': 0.013000769540667534, 'timing/model': 1.8125641159713268, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 0.51}) +[2026-06-15 22:43:19,183][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.21049083769321442, 'train/grad_norm_pre_clip': 6.11299467086792, 'timing/data': 0.011874141171574593, 'timing/model': 1.5462571568787098, 'learning_rate/qwen_vl_interface': 8.638769636429245e-06, 'learning_rate/action_model': 8.638769636429245e-05, 'epoch': 0.51}) +[2026-06-15 22:43:25,678][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.12319371849298477, 'train/grad_norm_pre_clip': 5.62902307510376, 'timing/data': 0.012692360207438469, 'timing/model': 1.6589240562170744, 'learning_rate/qwen_vl_interface': 8.636634917623946e-06, 'learning_rate/action_model': 8.636634917623944e-05, 'epoch': 0.51}) +[2026-06-15 22:43:32,229][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.17933407425880432, 'train/grad_norm_pre_clip': 12.703655242919922, 'timing/data': 0.01613939367234707, 'timing/model': 1.60809894092381, 'learning_rate/qwen_vl_interface': 8.634498806700955e-06, 'learning_rate/action_model': 8.634498806700954e-05, 'epoch': 0.51}) +[2026-06-15 22:43:38,935][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.06812755763530731, 'train/grad_norm_pre_clip': 5.333306312561035, 'timing/data': 0.018817482516169548, 'timing/model': 1.6099751014262438, 'learning_rate/qwen_vl_interface': 8.63236130453835e-06, 'learning_rate/action_model': 8.63236130453835e-05, 'epoch': 0.51}) +[2026-06-15 22:43:45,558][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.2485598623752594, 'train/grad_norm_pre_clip': 7.486690998077393, 'timing/data': 0.01620147004723549, 'timing/model': 1.5557904224842787, 'learning_rate/qwen_vl_interface': 8.630222412014777e-06, 'learning_rate/action_model': 8.630222412014776e-05, 'epoch': 0.51}) +[2026-06-15 22:43:51,809][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.05504690110683441, 'train/grad_norm_pre_clip': 3.0314230918884277, 'timing/data': 0.013817807659506798, 'timing/model': 1.8192583601921797, 'learning_rate/qwen_vl_interface': 8.628082130009451e-06, 'learning_rate/action_model': 8.628082130009451e-05, 'epoch': 0.51}) +[2026-06-15 22:43:57,834][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.12319104373455048, 'train/grad_norm_pre_clip': 6.653126239776611, 'timing/data': 0.012151896953582764, 'timing/model': 1.5205074306577444, 'learning_rate/qwen_vl_interface': 8.62594045940216e-06, 'learning_rate/action_model': 8.62594045940216e-05, 'epoch': 0.51}) +[2026-06-15 22:44:03,932][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.16226860880851746, 'train/grad_norm_pre_clip': 5.652074813842773, 'timing/data': 0.012404564768075943, 'timing/model': 1.5451492052525282, 'learning_rate/qwen_vl_interface': 8.623797401073272e-06, 'learning_rate/action_model': 8.623797401073271e-05, 'epoch': 0.51}) +[2026-06-15 22:44:10,034][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.07990890741348267, 'train/grad_norm_pre_clip': 6.891188144683838, 'timing/data': 0.011771483346819878, 'timing/model': 1.5507710166275501, 'learning_rate/qwen_vl_interface': 8.62165295590371e-06, 'learning_rate/action_model': 8.621652955903709e-05, 'epoch': 0.51}) +[2026-06-15 22:44:16,134][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.11318383365869522, 'train/grad_norm_pre_clip': 4.27441930770874, 'timing/data': 0.012787392362952232, 'timing/model': 1.5385286901146173, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 0.51}) +[2026-06-15 22:44:22,237][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.038484543561935425, 'train/grad_norm_pre_clip': 3.4635276794433594, 'timing/data': 0.01277921162545681, 'timing/model': 1.4996233489364386, 'learning_rate/qwen_vl_interface': 8.61735990856914e-06, 'learning_rate/action_model': 8.617359908569139e-05, 'epoch': 0.51}) +[2026-06-15 22:44:28,371][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.0721035972237587, 'train/grad_norm_pre_clip': 9.099488258361816, 'timing/data': 0.012568870559334755, 'timing/model': 1.782953929156065, 'learning_rate/qwen_vl_interface': 8.615211308168842e-06, 'learning_rate/action_model': 8.61521130816884e-05, 'epoch': 0.51}) +[2026-06-15 22:44:34,510][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.39388105273246765, 'train/grad_norm_pre_clip': 9.959912300109863, 'timing/data': 0.013219974935054779, 'timing/model': 1.7203589640557766, 'learning_rate/qwen_vl_interface': 8.613061324457289e-06, 'learning_rate/action_model': 8.613061324457288e-05, 'epoch': 0.51}) +[2026-06-15 22:44:40,440][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.1621856987476349, 'train/grad_norm_pre_clip': 8.561177253723145, 'timing/data': 0.011199507862329483, 'timing/model': 1.5520087778568268, 'learning_rate/qwen_vl_interface': 8.610909958318262e-06, 'learning_rate/action_model': 8.610909958318261e-05, 'epoch': 0.51}) +[2026-06-15 22:44:46,537][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.11196716129779816, 'train/grad_norm_pre_clip': 6.594451904296875, 'timing/data': 0.012542685493826866, 'timing/model': 1.5357196051627398, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.51}) +[2026-06-15 22:44:53,174][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.18963736295700073, 'train/grad_norm_pre_clip': 7.0856804847717285, 'timing/data': 0.012506311759352684, 'timing/model': 1.9822987206280231, 'learning_rate/qwen_vl_interface': 8.606603082295724e-06, 'learning_rate/action_model': 8.606603082295724e-05, 'epoch': 0.51}) +[2026-06-15 22:44:59,679][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.11874042451381683, 'train/grad_norm_pre_clip': 8.760396003723145, 'timing/data': 0.012840047478675842, 'timing/model': 1.5218337941914797, 'learning_rate/qwen_vl_interface': 8.60444757418261e-06, 'learning_rate/action_model': 8.604447574182609e-05, 'epoch': 0.51}) +[2026-06-15 22:45:06,005][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.03519432619214058, 'train/grad_norm_pre_clip': 3.344249963760376, 'timing/data': 0.014016218483448029, 'timing/model': 1.758527623489499, 'learning_rate/qwen_vl_interface': 8.602290687182805e-06, 'learning_rate/action_model': 8.602290687182806e-05, 'epoch': 0.51}) +[2026-06-15 22:45:12,064][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.016192618757486343, 'train/grad_norm_pre_clip': 1.9916229248046875, 'timing/data': 0.013641363009810448, 'timing/model': 1.529912892729044, 'learning_rate/qwen_vl_interface': 8.600132422182929e-06, 'learning_rate/action_model': 8.600132422182928e-05, 'epoch': 0.51}) +[2026-06-15 22:45:18,235][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.15322908759117126, 'train/grad_norm_pre_clip': 14.447823524475098, 'timing/data': 0.012136507779359818, 'timing/model': 1.513022817671299, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 0.51}) +[2026-06-15 22:45:24,430][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.13984669744968414, 'train/grad_norm_pre_clip': 9.146367073059082, 'timing/data': 0.011533791199326515, 'timing/model': 1.5141295157372952, 'learning_rate/qwen_vl_interface': 8.595811761732239e-06, 'learning_rate/action_model': 8.595811761732238e-05, 'epoch': 0.52}) +[2026-06-15 22:45:30,636][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.3355492949485779, 'train/grad_norm_pre_clip': 10.516618728637695, 'timing/data': 0.012473540380597115, 'timing/model': 1.5006409622728825, 'learning_rate/qwen_vl_interface': 8.593649368057489e-06, 'learning_rate/action_model': 8.59364936805749e-05, 'epoch': 0.52}) +[2026-06-15 22:45:36,634][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.31296008825302124, 'train/grad_norm_pre_clip': 28.119632720947266, 'timing/data': 0.011679336428642273, 'timing/model': 1.487632606178522, 'learning_rate/qwen_vl_interface': 8.591485599934786e-06, 'learning_rate/action_model': 8.591485599934785e-05, 'epoch': 0.52}) +[2026-06-15 22:45:42,755][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.2562943994998932, 'train/grad_norm_pre_clip': 10.85092830657959, 'timing/data': 0.011886820197105408, 'timing/model': 1.7607014570385218, 'learning_rate/qwen_vl_interface': 8.58932045825357e-06, 'learning_rate/action_model': 8.589320458253569e-05, 'epoch': 0.52}) +[2026-06-15 22:45:48,656][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.5289958715438843, 'train/grad_norm_pre_clip': 18.192846298217773, 'timing/data': 0.012511052191257477, 'timing/model': 1.5615551117807627, 'learning_rate/qwen_vl_interface': 8.58715394390385e-06, 'learning_rate/action_model': 8.58715394390385e-05, 'epoch': 0.52}) +[2026-06-15 22:45:54,838][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.15313492715358734, 'train/grad_norm_pre_clip': 9.979842185974121, 'timing/data': 0.012539742514491081, 'timing/model': 1.5555212628096342, 'learning_rate/qwen_vl_interface': 8.5849860577762e-06, 'learning_rate/action_model': 8.5849860577762e-05, 'epoch': 0.52}) +[2026-06-15 22:46:01,059][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.1472979635000229, 'train/grad_norm_pre_clip': 7.743558883666992, 'timing/data': 0.012959310784935951, 'timing/model': 1.5753870606422424, 'learning_rate/qwen_vl_interface': 8.582816800761756e-06, 'learning_rate/action_model': 8.582816800761755e-05, 'epoch': 0.52}) +[2026-06-15 22:46:07,148][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.06727080792188644, 'train/grad_norm_pre_clip': 3.6979589462280273, 'timing/data': 0.011802280321717262, 'timing/model': 1.5140387192368507, 'learning_rate/qwen_vl_interface': 8.580646173752214e-06, 'learning_rate/action_model': 8.580646173752213e-05, 'epoch': 0.52}) +[2026-06-15 22:46:13,016][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.025184130296111107, 'train/grad_norm_pre_clip': 1.5999488830566406, 'timing/data': 0.010571898892521858, 'timing/model': 1.4499855302274227, 'learning_rate/qwen_vl_interface': 8.578474177639841e-06, 'learning_rate/action_model': 8.578474177639842e-05, 'epoch': 0.52}) +[2026-06-15 22:46:18,875][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.07899155467748642, 'train/grad_norm_pre_clip': 8.766584396362305, 'timing/data': 0.010532088577747345, 'timing/model': 1.613435585051775, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 0.52}) +[2026-06-15 22:46:24,726][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.042082998901605606, 'train/grad_norm_pre_clip': 6.864889144897461, 'timing/data': 0.011193931102752686, 'timing/model': 1.457751002162695, 'learning_rate/qwen_vl_interface': 8.574126081678463e-06, 'learning_rate/action_model': 8.574126081678462e-05, 'epoch': 0.52}) +[2026-06-15 22:46:30,823][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.09681697189807892, 'train/grad_norm_pre_clip': 5.481386661529541, 'timing/data': 0.010052572935819626, 'timing/model': 1.4481746833771467, 'learning_rate/qwen_vl_interface': 8.571949983616796e-06, 'learning_rate/action_model': 8.571949983616795e-05, 'epoch': 0.52}) +[2026-06-15 22:46:36,717][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.1340799480676651, 'train/grad_norm_pre_clip': 9.19624137878418, 'timing/data': 0.012347228825092316, 'timing/model': 1.4524255376309156, 'learning_rate/qwen_vl_interface': 8.569772520026971e-06, 'learning_rate/action_model': 8.569772520026971e-05, 'epoch': 0.52}) +[2026-06-15 22:46:42,940][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.09072153270244598, 'train/grad_norm_pre_clip': 4.755789756774902, 'timing/data': 0.012504231184720993, 'timing/model': 1.6163420900702477, 'learning_rate/qwen_vl_interface': 8.567593691804061e-06, 'learning_rate/action_model': 8.567593691804061e-05, 'epoch': 0.52}) +[2026-06-15 22:46:49,409][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.251987487077713, 'train/grad_norm_pre_clip': 12.697669982910156, 'timing/data': 0.012659719213843346, 'timing/model': 1.8031021505594254, 'learning_rate/qwen_vl_interface': 8.565413499843703e-06, 'learning_rate/action_model': 8.565413499843703e-05, 'epoch': 0.52}) +[2026-06-15 22:46:55,396][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.009253118187189102, 'train/grad_norm_pre_clip': 1.66431725025177, 'timing/data': 0.012207096442580223, 'timing/model': 1.5260040704160929, 'learning_rate/qwen_vl_interface': 8.563231945042087e-06, 'learning_rate/action_model': 8.563231945042088e-05, 'epoch': 0.52}) +[2026-06-15 22:47:01,482][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.2523370087146759, 'train/grad_norm_pre_clip': 10.434694290161133, 'timing/data': 0.011725671589374542, 'timing/model': 1.4234082736074924, 'learning_rate/qwen_vl_interface': 8.56104902829597e-06, 'learning_rate/action_model': 8.56104902829597e-05, 'epoch': 0.52}) +[2026-06-15 22:47:07,284][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.03836551681160927, 'train/grad_norm_pre_clip': 3.9665439128875732, 'timing/data': 0.009960098192095757, 'timing/model': 1.4562562853097916, 'learning_rate/qwen_vl_interface': 8.558864750502669e-06, 'learning_rate/action_model': 8.558864750502668e-05, 'epoch': 0.52}) +[2026-06-15 22:47:13,090][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.1416010856628418, 'train/grad_norm_pre_clip': 6.88977575302124, 'timing/data': 0.009663471952080727, 'timing/model': 1.4622025843709707, 'learning_rate/qwen_vl_interface': 8.556679112560055e-06, 'learning_rate/action_model': 8.556679112560054e-05, 'epoch': 0.52}) +[2026-06-15 22:47:18,986][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.06312283128499985, 'train/grad_norm_pre_clip': 9.857502937316895, 'timing/data': 0.010454179719090462, 'timing/model': 1.491555469110608, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 0.52}) +[2026-06-15 22:47:24,891][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.23739774525165558, 'train/grad_norm_pre_clip': 12.652668952941895, 'timing/data': 0.010639140382409096, 'timing/model': 1.658828891813755, 'learning_rate/qwen_vl_interface': 8.55230375982118e-06, 'learning_rate/action_model': 8.55230375982118e-05, 'epoch': 0.52}) +[2026-06-15 22:47:30,586][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.14463244378566742, 'train/grad_norm_pre_clip': 10.586531639099121, 'timing/data': 0.01397392712533474, 'timing/model': 1.4579695519059896, 'learning_rate/qwen_vl_interface': 8.550114046823463e-06, 'learning_rate/action_model': 8.550114046823463e-05, 'epoch': 0.52}) +[2026-06-15 22:47:36,483][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.04926520213484764, 'train/grad_norm_pre_clip': 4.0056352615356445, 'timing/data': 0.011197876185178757, 'timing/model': 1.4550933986902237, 'learning_rate/qwen_vl_interface': 8.547922977273518e-06, 'learning_rate/action_model': 8.547922977273518e-05, 'epoch': 0.52}) +[2026-06-15 22:47:42,381][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.03666914254426956, 'train/grad_norm_pre_clip': 2.528635263442993, 'timing/data': 0.01026618666946888, 'timing/model': 1.4522696025669575, 'learning_rate/qwen_vl_interface': 8.54573055207201e-06, 'learning_rate/action_model': 8.545730552072008e-05, 'epoch': 0.52}) +[2026-06-15 22:47:48,185][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.11959439516067505, 'train/grad_norm_pre_clip': 9.37136173248291, 'timing/data': 0.010748371481895447, 'timing/model': 1.4576970431953669, 'learning_rate/qwen_vl_interface': 8.543536772120161e-06, 'learning_rate/action_model': 8.54353677212016e-05, 'epoch': 0.52}) +[2026-06-15 22:47:54,084][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.21732079982757568, 'train/grad_norm_pre_clip': 18.531665802001953, 'timing/data': 0.010125787928700447, 'timing/model': 1.491290159523487, 'learning_rate/qwen_vl_interface': 8.541341638319754e-06, 'learning_rate/action_model': 8.541341638319754e-05, 'epoch': 0.52}) +[2026-06-15 22:48:00,004][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.01592153310775757, 'train/grad_norm_pre_clip': 1.1313222646713257, 'timing/data': 0.011459605768322945, 'timing/model': 1.668062200769782, 'learning_rate/qwen_vl_interface': 8.539145151573126e-06, 'learning_rate/action_model': 8.539145151573125e-05, 'epoch': 0.53}) +[2026-06-15 22:48:05,693][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.09129540622234344, 'train/grad_norm_pre_clip': 11.058143615722656, 'timing/data': 0.01117529347538948, 'timing/model': 1.4540135022252798, 'learning_rate/qwen_vl_interface': 8.536947312783167e-06, 'learning_rate/action_model': 8.536947312783166e-05, 'epoch': 0.53}) +[2026-06-15 22:48:11,594][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.515160083770752, 'train/grad_norm_pre_clip': 23.87833023071289, 'timing/data': 0.010744316503405571, 'timing/model': 1.4541083220392466, 'learning_rate/qwen_vl_interface': 8.534748122853326e-06, 'learning_rate/action_model': 8.534748122853325e-05, 'epoch': 0.53}) +[2026-06-15 22:48:17,493][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.1706218123435974, 'train/grad_norm_pre_clip': 7.1857147216796875, 'timing/data': 0.0105845145881176, 'timing/model': 1.4524646196514368, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.53}) +[2026-06-15 22:48:23,395][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.3883366882801056, 'train/grad_norm_pre_clip': 17.026302337646484, 'timing/data': 0.011370167136192322, 'timing/model': 1.4349840208888054, 'learning_rate/qwen_vl_interface': 8.530345693190573e-06, 'learning_rate/action_model': 8.530345693190572e-05, 'epoch': 0.53}) +[2026-06-15 22:48:29,477][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.05870755389332771, 'train/grad_norm_pre_clip': 7.621776103973389, 'timing/data': 0.017380520701408386, 'timing/model': 1.5735940132290125, 'learning_rate/qwen_vl_interface': 8.528142455267328e-06, 'learning_rate/action_model': 8.528142455267329e-05, 'epoch': 0.53}) +[2026-06-15 22:48:35,783][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.13367480039596558, 'train/grad_norm_pre_clip': 12.192729949951172, 'timing/data': 0.012861751019954681, 'timing/model': 1.7667355220764875, 'learning_rate/qwen_vl_interface': 8.52593786982355e-06, 'learning_rate/action_model': 8.525937869823551e-05, 'epoch': 0.53}) +[2026-06-15 22:48:41,732][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.3597806692123413, 'train/grad_norm_pre_clip': 11.625717163085938, 'timing/data': 0.011848557740449905, 'timing/model': 1.533658642321825, 'learning_rate/qwen_vl_interface': 8.523731937765456e-06, 'learning_rate/action_model': 8.523731937765456e-05, 'epoch': 0.53}) +[2026-06-15 22:48:47,967][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.0594821460545063, 'train/grad_norm_pre_clip': 5.100400924682617, 'timing/data': 0.01577075570821762, 'timing/model': 1.5716213546693325, 'learning_rate/qwen_vl_interface': 8.521524659999823e-06, 'learning_rate/action_model': 8.521524659999822e-05, 'epoch': 0.53}) +[2026-06-15 22:48:54,215][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.15597811341285706, 'train/grad_norm_pre_clip': 11.564091682434082, 'timing/data': 0.010638952255249023, 'timing/model': 1.5169390924274921, 'learning_rate/qwen_vl_interface': 8.519316037433976e-06, 'learning_rate/action_model': 8.519316037433977e-05, 'epoch': 0.53}) +[2026-06-15 22:49:00,082][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.10418911278247833, 'train/grad_norm_pre_clip': 6.516470432281494, 'timing/data': 0.011347604915499687, 'timing/model': 1.4197724480181932, 'learning_rate/qwen_vl_interface': 8.517106070975801e-06, 'learning_rate/action_model': 8.5171060709758e-05, 'epoch': 0.53}) +[2026-06-15 22:49:05,982][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.2009926289319992, 'train/grad_norm_pre_clip': 9.181132316589355, 'timing/data': 0.01055426150560379, 'timing/model': 1.443480845540762, 'learning_rate/qwen_vl_interface': 8.514894761533729e-06, 'learning_rate/action_model': 8.514894761533728e-05, 'epoch': 0.53}) +[2026-06-15 22:49:11,897][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.12072121351957321, 'train/grad_norm_pre_clip': 9.609450340270996, 'timing/data': 0.010519901290535927, 'timing/model': 1.663051038980484, 'learning_rate/qwen_vl_interface': 8.512682110016747e-06, 'learning_rate/action_model': 8.512682110016747e-05, 'epoch': 0.53}) +[2026-06-15 22:49:17,581][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.32094112038612366, 'train/grad_norm_pre_clip': 7.026394844055176, 'timing/data': 0.010455427691340446, 'timing/model': 1.4543383792042732, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 0.53}) +[2026-06-15 22:49:23,483][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.3266216516494751, 'train/grad_norm_pre_clip': 11.724141120910645, 'timing/data': 0.010064685717225075, 'timing/model': 1.4547331631183624, 'learning_rate/qwen_vl_interface': 8.50825278439675e-06, 'learning_rate/action_model': 8.508252784396749e-05, 'epoch': 0.53}) +[2026-06-15 22:49:29,279][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.2108066976070404, 'train/grad_norm_pre_clip': 13.25953197479248, 'timing/data': 0.010406119748950005, 'timing/model': 1.4500875137746334, 'learning_rate/qwen_vl_interface': 8.506036112114465e-06, 'learning_rate/action_model': 8.506036112114464e-05, 'epoch': 0.53}) +[2026-06-15 22:49:35,095][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.2632093131542206, 'train/grad_norm_pre_clip': 7.367688179016113, 'timing/data': 0.010402360931038857, 'timing/model': 1.46806132607162, 'learning_rate/qwen_vl_interface': 8.503818101398725e-06, 'learning_rate/action_model': 8.503818101398725e-05, 'epoch': 0.53}) +[2026-06-15 22:49:41,115][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.029848741367459297, 'train/grad_norm_pre_clip': 2.212028741836548, 'timing/data': 0.010431783273816109, 'timing/model': 1.5205726604908705, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.53}) +[2026-06-15 22:49:47,369][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.13116632401943207, 'train/grad_norm_pre_clip': 7.386049747467041, 'timing/data': 0.011197144165635109, 'timing/model': 1.9852246306836605, 'learning_rate/qwen_vl_interface': 8.499378068314395e-06, 'learning_rate/action_model': 8.499378068314395e-05, 'epoch': 0.53}) +[2026-06-15 22:49:53,439][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.10656553506851196, 'train/grad_norm_pre_clip': 3.934109687805176, 'timing/data': 0.011618182063102722, 'timing/model': 1.6488416735082865, 'learning_rate/qwen_vl_interface': 8.497156047770932e-06, 'learning_rate/action_model': 8.497156047770933e-05, 'epoch': 0.53}) +[2026-06-15 22:49:59,534][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.09893438220024109, 'train/grad_norm_pre_clip': 5.507835388183594, 'timing/data': 0.013464048504829407, 'timing/model': 1.461877716705203, 'learning_rate/qwen_vl_interface': 8.494932692444276e-06, 'learning_rate/action_model': 8.494932692444276e-05, 'epoch': 0.53}) +[2026-06-15 22:50:05,719][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.06563935428857803, 'train/grad_norm_pre_clip': 3.1120669841766357, 'timing/data': 0.013705529272556305, 'timing/model': 1.4527586195617914, 'learning_rate/qwen_vl_interface': 8.492708003248362e-06, 'learning_rate/action_model': 8.492708003248362e-05, 'epoch': 0.53}) +[2026-06-15 22:50:11,827][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.019983017817139626, 'train/grad_norm_pre_clip': 1.2161448001861572, 'timing/data': 0.01594029739499092, 'timing/model': 1.523411214351654, 'learning_rate/qwen_vl_interface': 8.490481981097674e-06, 'learning_rate/action_model': 8.490481981097674e-05, 'epoch': 0.53}) +[2026-06-15 22:50:17,842][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.2946438193321228, 'train/grad_norm_pre_clip': 7.016838550567627, 'timing/data': 0.01147177442908287, 'timing/model': 1.4649490173906088, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 0.53}) +[2026-06-15 22:50:23,793][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.17376120388507843, 'train/grad_norm_pre_clip': 6.0745768547058105, 'timing/data': 0.010427450761198997, 'timing/model': 1.6619253400713205, 'learning_rate/qwen_vl_interface': 8.486025941592664e-06, 'learning_rate/action_model': 8.486025941592664e-05, 'epoch': 0.53}) +[2026-06-15 22:50:29,495][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.3260616958141327, 'train/grad_norm_pre_clip': 18.87388038635254, 'timing/data': 0.01006365567445755, 'timing/model': 1.4509374257177114, 'learning_rate/qwen_vl_interface': 8.483795926070054e-06, 'learning_rate/action_model': 8.483795926070052e-05, 'epoch': 0.53}) +[2026-06-15 22:50:35,394][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.11133212596178055, 'train/grad_norm_pre_clip': 8.280810356140137, 'timing/data': 0.010419521480798721, 'timing/model': 1.445703538134694, 'learning_rate/qwen_vl_interface': 8.481564581256087e-06, 'learning_rate/action_model': 8.481564581256087e-05, 'epoch': 0.54}) +[2026-06-15 22:50:41,287][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.13547269999980927, 'train/grad_norm_pre_clip': 11.183500289916992, 'timing/data': 0.010450752452015877, 'timing/model': 1.4547751694917679, 'learning_rate/qwen_vl_interface': 8.479331908067987e-06, 'learning_rate/action_model': 8.479331908067987e-05, 'epoch': 0.54}) +[2026-06-15 22:50:47,186][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.035187140107154846, 'train/grad_norm_pre_clip': 3.116612195968628, 'timing/data': 0.01108112744987011, 'timing/model': 1.4584956504404545, 'learning_rate/qwen_vl_interface': 8.477097907423524e-06, 'learning_rate/action_model': 8.477097907423524e-05, 'epoch': 0.54}) +[2026-06-15 22:50:53,081][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.20128273963928223, 'train/grad_norm_pre_clip': 12.76973819732666, 'timing/data': 0.010822366923093796, 'timing/model': 1.4423884190618992, 'learning_rate/qwen_vl_interface': 8.474862580241006e-06, 'learning_rate/action_model': 8.474862580241006e-05, 'epoch': 0.54}) +[2026-06-15 22:50:58,990][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.09944567084312439, 'train/grad_norm_pre_clip': 12.852916717529297, 'timing/data': 0.012632448226213455, 'timing/model': 1.6562310066074133, 'learning_rate/qwen_vl_interface': 8.472625927439299e-06, 'learning_rate/action_model': 8.472625927439299e-05, 'epoch': 0.54}) +[2026-06-15 22:51:04,891][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.07928764075040817, 'train/grad_norm_pre_clip': 6.225752353668213, 'timing/data': 0.011276565492153168, 'timing/model': 1.6603171397000551, 'learning_rate/qwen_vl_interface': 8.4703879499378e-06, 'learning_rate/action_model': 8.4703879499378e-05, 'epoch': 0.54}) +[2026-06-15 22:51:10,584][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.14120817184448242, 'train/grad_norm_pre_clip': 18.613447189331055, 'timing/data': 0.011303804814815521, 'timing/model': 1.4376532789319754, 'learning_rate/qwen_vl_interface': 8.46814864865646e-06, 'learning_rate/action_model': 8.46814864865646e-05, 'epoch': 0.54}) +[2026-06-15 22:51:16,485][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.1493212878704071, 'train/grad_norm_pre_clip': 7.170654296875, 'timing/data': 0.011011267080903053, 'timing/model': 1.4531890098005533, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 0.54}) +[2026-06-15 22:51:22,384][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.12342580407857895, 'train/grad_norm_pre_clip': 16.9200496673584, 'timing/data': 0.010892124846577644, 'timing/model': 1.454049689695239, 'learning_rate/qwen_vl_interface': 8.463666078436769e-06, 'learning_rate/action_model': 8.463666078436769e-05, 'epoch': 0.54}) +[2026-06-15 22:51:28,297][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.10806913673877716, 'train/grad_norm_pre_clip': 16.074108123779297, 'timing/data': 0.010980788618326187, 'timing/model': 1.42466614022851, 'learning_rate/qwen_vl_interface': 8.46142281134103e-06, 'learning_rate/action_model': 8.461422811341031e-05, 'epoch': 0.54}) +[2026-06-15 22:51:34,195][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.14764173328876495, 'train/grad_norm_pre_clip': 15.087101936340332, 'timing/data': 0.011470777913928032, 'timing/model': 1.6633966397494078, 'learning_rate/qwen_vl_interface': 8.459178224150683e-06, 'learning_rate/action_model': 8.459178224150683e-05, 'epoch': 0.54}) +[2026-06-15 22:51:40,280][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.22953863441944122, 'train/grad_norm_pre_clip': 9.897847175598145, 'timing/data': 0.014770196750760078, 'timing/model': 1.5474461000412703, 'learning_rate/qwen_vl_interface': 8.456932317788389e-06, 'learning_rate/action_model': 8.456932317788389e-05, 'epoch': 0.54}) +[2026-06-15 22:51:46,778][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.0874771997332573, 'train/grad_norm_pre_clip': 6.672316551208496, 'timing/data': 0.010402785614132881, 'timing/model': 1.429240059107542, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 0.54}) +[2026-06-15 22:51:52,679][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.024324189871549606, 'train/grad_norm_pre_clip': 2.5891385078430176, 'timing/data': 0.010861996561288834, 'timing/model': 1.4551700595766306, 'learning_rate/qwen_vl_interface': 8.452436551241328e-06, 'learning_rate/action_model': 8.452436551241328e-05, 'epoch': 0.54}) +[2026-06-15 22:51:58,941][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.13034461438655853, 'train/grad_norm_pre_clip': 7.895651340484619, 'timing/data': 0.017124241217970848, 'timing/model': 1.4872006196528673, 'learning_rate/qwen_vl_interface': 8.450186692904605e-06, 'learning_rate/action_model': 8.450186692904606e-05, 'epoch': 0.54}) +[2026-06-15 22:52:05,049][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.15853849053382874, 'train/grad_norm_pre_clip': 14.952655792236328, 'timing/data': 0.01556583121418953, 'timing/model': 1.5353791452944279, 'learning_rate/qwen_vl_interface': 8.447935519092014e-06, 'learning_rate/action_model': 8.447935519092014e-05, 'epoch': 0.54}) +[2026-06-15 22:52:11,202][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.12195923179388046, 'train/grad_norm_pre_clip': 7.686537742614746, 'timing/data': 0.013647451996803284, 'timing/model': 1.7018714249134064, 'learning_rate/qwen_vl_interface': 8.445683030728925e-06, 'learning_rate/action_model': 8.445683030728925e-05, 'epoch': 0.54}) +[2026-06-15 23:00:12,537][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.16250218451023102, 'train/grad_norm_pre_clip': 17.306657791137695, 'eval/loss': 0.09556012600660324, 'eval/flap_precision': 0.9602941176470589, 'eval/flap_recall': 0.8860244233378561, 'eval/flap_f1': 0.9216654904728299, 'eval/flap_rate_model': 0.18888888888888888, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9216654904728299, 'timing/data': 0.012921785935759544, 'timing/model': 1.5430039297789335, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 0.54}) +[2026-06-15 23:00:12,538][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1400_state +[2026-06-15 23:00:12,544][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:00:18,529][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1400_state/model.safetensors +[2026-06-15 23:00:31,988][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1400_state/optimizer.bin +[2026-06-15 23:00:31,989][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1400_state/sampler.bin +[2026-06-15 23:00:31,989][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1400_state/sampler_1.bin +[2026-06-15 23:00:31,990][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1400_state/random_states_0.pkl +[2026-06-15 23:00:32,092][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-06-15 23:00:32,092][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:00:32,100][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:00:38,035][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.056950643658638, 'train/grad_norm_pre_clip': 7.322772026062012, 'timing/data': 0.010397931560873985, 'timing/model': 1.6972643285989761, 'learning_rate/qwen_vl_interface': 8.441174114055457e-06, 'learning_rate/action_model': 8.441174114055456e-05, 'epoch': 0.54}) +[2026-06-15 23:00:43,701][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.0672815591096878, 'train/grad_norm_pre_clip': 8.695283889770508, 'timing/data': 0.011448154225945473, 'timing/model': 1.45769901573658, 'learning_rate/qwen_vl_interface': 8.43891768759852e-06, 'learning_rate/action_model': 8.43891768759852e-05, 'epoch': 0.54}) +[2026-06-15 23:00:49,603][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.4279981851577759, 'train/grad_norm_pre_clip': 17.52819061279297, 'timing/data': 0.010879036039113998, 'timing/model': 1.4598944187164307, 'learning_rate/qwen_vl_interface': 8.436659950297978e-06, 'learning_rate/action_model': 8.436659950297977e-05, 'epoch': 0.54}) +[2026-06-15 23:00:55,509][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.28301483392715454, 'train/grad_norm_pre_clip': 10.718393325805664, 'timing/data': 0.011261677369475365, 'timing/model': 1.4628493338823318, 'learning_rate/qwen_vl_interface': 8.4344009030819e-06, 'learning_rate/action_model': 8.4344009030819e-05, 'epoch': 0.54}) +[2026-06-15 23:01:01,386][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.219296395778656, 'train/grad_norm_pre_clip': 15.244400024414062, 'timing/data': 0.010428844019770622, 'timing/model': 1.4363671503961086, 'learning_rate/qwen_vl_interface': 8.432140546878899e-06, 'learning_rate/action_model': 8.432140546878898e-05, 'epoch': 0.54}) +[2026-06-15 23:01:07,671][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.28964000940322876, 'train/grad_norm_pre_clip': 15.066142082214355, 'timing/data': 0.013247435912489891, 'timing/model': 1.7904992289841175, 'learning_rate/qwen_vl_interface': 8.429878882618116e-06, 'learning_rate/action_model': 8.429878882618116e-05, 'epoch': 0.54}) +[2026-06-15 23:01:14,167][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.5202773213386536, 'train/grad_norm_pre_clip': 26.344654083251953, 'timing/data': 0.024788618087768555, 'timing/model': 1.5235844980925322, 'learning_rate/qwen_vl_interface': 8.42761591122924e-06, 'learning_rate/action_model': 8.427615911229241e-05, 'epoch': 0.54}) +[2026-06-15 23:01:20,868][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.1085720807313919, 'train/grad_norm_pre_clip': 11.217009544372559, 'timing/data': 0.013113806024193764, 'timing/model': 1.653580678626895, 'learning_rate/qwen_vl_interface': 8.425351633642491e-06, 'learning_rate/action_model': 8.425351633642492e-05, 'epoch': 0.54}) +[2026-06-15 23:01:27,721][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.10233870148658752, 'train/grad_norm_pre_clip': 7.437058448791504, 'timing/data': 0.015113640576601028, 'timing/model': 1.564348716288805, 'learning_rate/qwen_vl_interface': 8.423086050788632e-06, 'learning_rate/action_model': 8.423086050788631e-05, 'epoch': 0.55}) +[2026-06-15 23:01:34,230][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.10412436723709106, 'train/grad_norm_pre_clip': 12.696830749511719, 'timing/data': 0.01521691307425499, 'timing/model': 1.580522071570158, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 0.55}) +[2026-06-15 23:01:40,881][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.10907492786645889, 'train/grad_norm_pre_clip': 7.203219890594482, 'timing/data': 0.014026075601577759, 'timing/model': 1.6153866201639175, 'learning_rate/qwen_vl_interface': 8.418550973005289e-06, 'learning_rate/action_model': 8.418550973005289e-05, 'epoch': 0.55}) +[2026-06-15 23:01:47,436][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.16741366684436798, 'train/grad_norm_pre_clip': 9.056901931762695, 'timing/data': 0.015367504209280014, 'timing/model': 1.798522600904107, 'learning_rate/qwen_vl_interface': 8.416281479940005e-06, 'learning_rate/action_model': 8.416281479940005e-05, 'epoch': 0.55}) +[2026-06-15 23:01:53,821][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.2693396806716919, 'train/grad_norm_pre_clip': 13.9036226272583, 'timing/data': 0.019080093130469322, 'timing/model': 1.6660181675106287, 'learning_rate/qwen_vl_interface': 8.414010685336008e-06, 'learning_rate/action_model': 8.414010685336008e-05, 'epoch': 0.55}) +[2026-06-15 23:02:00,818][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.02911607176065445, 'train/grad_norm_pre_clip': 2.3892831802368164, 'timing/data': 0.02008107490837574, 'timing/model': 1.7458112351596355, 'learning_rate/qwen_vl_interface': 8.411738590126733e-06, 'learning_rate/action_model': 8.411738590126732e-05, 'epoch': 0.55}) +[2026-06-15 23:02:08,269][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.14491340517997742, 'train/grad_norm_pre_clip': 12.499536514282227, 'timing/data': 0.014766452834010124, 'timing/model': 1.5410611722618341, 'learning_rate/qwen_vl_interface': 8.409465195246152e-06, 'learning_rate/action_model': 8.409465195246152e-05, 'epoch': 0.55}) +[2026-06-15 23:02:14,569][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.31270548701286316, 'train/grad_norm_pre_clip': 8.562483787536621, 'timing/data': 0.012921120971441269, 'timing/model': 1.5518704634159803, 'learning_rate/qwen_vl_interface': 8.407190501628772e-06, 'learning_rate/action_model': 8.407190501628772e-05, 'epoch': 0.55}) +[2026-06-15 23:02:20,868][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.260016530752182, 'train/grad_norm_pre_clip': 14.31004810333252, 'timing/data': 0.01255139522254467, 'timing/model': 1.5469814073294401, 'learning_rate/qwen_vl_interface': 8.404914510209635e-06, 'learning_rate/action_model': 8.404914510209635e-05, 'epoch': 0.55}) +[2026-06-15 23:02:27,679][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.2638286054134369, 'train/grad_norm_pre_clip': 13.923072814941406, 'timing/data': 0.014857884496450424, 'timing/model': 1.8460998497903347, 'learning_rate/qwen_vl_interface': 8.402637221924315e-06, 'learning_rate/action_model': 8.402637221924314e-05, 'epoch': 0.55}) +[2026-06-15 23:02:33,913][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.10564853996038437, 'train/grad_norm_pre_clip': 5.753415107727051, 'timing/data': 0.01498529501259327, 'timing/model': 1.5736563559621572, 'learning_rate/qwen_vl_interface': 8.400358637708919e-06, 'learning_rate/action_model': 8.400358637708918e-05, 'epoch': 0.55}) +[2026-06-15 23:02:40,389][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.14902932941913605, 'train/grad_norm_pre_clip': 9.961849212646484, 'timing/data': 0.014119785279035568, 'timing/model': 1.542409436777234, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 0.55}) +[2026-06-15 23:02:46,823][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.1807650774717331, 'train/grad_norm_pre_clip': 5.946606636047363, 'timing/data': 0.01324119046330452, 'timing/model': 1.6497169323265553, 'learning_rate/qwen_vl_interface': 8.395797585234987e-06, 'learning_rate/action_model': 8.395797585234988e-05, 'epoch': 0.55}) +[2026-06-15 23:02:52,934][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.11025510728359222, 'train/grad_norm_pre_clip': 4.322889804840088, 'timing/data': 0.012382403016090393, 'timing/model': 1.5028795171529055, 'learning_rate/qwen_vl_interface': 8.39351511885133e-06, 'learning_rate/action_model': 8.393515118851329e-05, 'epoch': 0.55}) +[2026-06-15 23:02:59,044][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.1858828067779541, 'train/grad_norm_pre_clip': 12.53477668762207, 'timing/data': 0.01196310855448246, 'timing/model': 1.4932666532695293, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.55}) +[2026-06-15 23:03:05,271][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.2247820943593979, 'train/grad_norm_pre_clip': 15.140442848205566, 'timing/data': 0.013265572488307953, 'timing/model': 1.750570259988308, 'learning_rate/qwen_vl_interface': 8.388946310481807e-06, 'learning_rate/action_model': 8.388946310481807e-05, 'epoch': 0.55}) +[2026-06-15 23:03:11,383][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.15435872972011566, 'train/grad_norm_pre_clip': 8.315503120422363, 'timing/data': 0.012569643557071686, 'timing/model': 1.5579905472695827, 'learning_rate/qwen_vl_interface': 8.386659970374007e-06, 'learning_rate/action_model': 8.386659970374006e-05, 'epoch': 0.55}) +[2026-06-15 23:03:17,762][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.09256207942962646, 'train/grad_norm_pre_clip': 4.548832893371582, 'timing/data': 0.0127827487885952, 'timing/model': 1.5289678554981947, 'learning_rate/qwen_vl_interface': 8.384372340903775e-06, 'learning_rate/action_model': 8.384372340903775e-05, 'epoch': 0.55}) +[2026-06-15 23:03:24,158][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.0754103809595108, 'train/grad_norm_pre_clip': 8.50998592376709, 'timing/data': 0.01391090638935566, 'timing/model': 1.5066909622401, 'learning_rate/qwen_vl_interface': 8.382083423011472e-06, 'learning_rate/action_model': 8.382083423011471e-05, 'epoch': 0.55}) +[2026-06-15 23:03:30,269][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.09529444575309753, 'train/grad_norm_pre_clip': 4.306748867034912, 'timing/data': 0.013481056317687035, 'timing/model': 1.5519734527915716, 'learning_rate/qwen_vl_interface': 8.379793217637979e-06, 'learning_rate/action_model': 8.379793217637979e-05, 'epoch': 0.55}) +[2026-06-15 23:03:36,679][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.24200451374053955, 'train/grad_norm_pre_clip': 5.3822479248046875, 'timing/data': 0.015537198632955551, 'timing/model': 1.5972358919680119, 'learning_rate/qwen_vl_interface': 8.377501725724719e-06, 'learning_rate/action_model': 8.377501725724719e-05, 'epoch': 0.55}) +[2026-06-15 23:03:43,068][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.06554768234491348, 'train/grad_norm_pre_clip': 3.2420237064361572, 'timing/data': 0.012791460379958153, 'timing/model': 1.751875290647149, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.55}) +[2026-06-15 23:03:49,179][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.10569890588521957, 'train/grad_norm_pre_clip': 8.843856811523438, 'timing/data': 0.012557340785861015, 'timing/model': 1.5589826330542564, 'learning_rate/qwen_vl_interface': 8.372914886047203e-06, 'learning_rate/action_model': 8.372914886047203e-05, 'epoch': 0.55}) +[2026-06-15 23:03:55,565][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.11382604390382767, 'train/grad_norm_pre_clip': 6.5401387214660645, 'timing/data': 0.013922479003667831, 'timing/model': 1.547292174771428, 'learning_rate/qwen_vl_interface': 8.370619540168422e-06, 'learning_rate/action_model': 8.370619540168423e-05, 'epoch': 0.55}) +[2026-06-15 23:04:01,776][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.2161841243505478, 'train/grad_norm_pre_clip': 9.598212242126465, 'timing/data': 0.015672871842980385, 'timing/model': 1.5401099305599928, 'learning_rate/qwen_vl_interface': 8.368322911520826e-06, 'learning_rate/action_model': 8.368322911520826e-05, 'epoch': 0.55}) +[2026-06-15 23:04:08,076][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.18494698405265808, 'train/grad_norm_pre_clip': 9.993947982788086, 'timing/data': 0.014269990846514702, 'timing/model': 1.5301241781562567, 'learning_rate/qwen_vl_interface': 8.36602500104847e-06, 'learning_rate/action_model': 8.36602500104847e-05, 'epoch': 0.55}) +[2026-06-15 23:04:14,386][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.027393534779548645, 'train/grad_norm_pre_clip': 2.3559601306915283, 'timing/data': 0.01310400664806366, 'timing/model': 1.519900020211935, 'learning_rate/qwen_vl_interface': 8.36372580969594e-06, 'learning_rate/action_model': 8.36372580969594e-05, 'epoch': 0.56}) +[2026-06-15 23:04:20,777][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.13718333840370178, 'train/grad_norm_pre_clip': 8.579151153564453, 'timing/data': 0.015820885077118874, 'timing/model': 1.8067800756543875, 'learning_rate/qwen_vl_interface': 8.361425338408344e-06, 'learning_rate/action_model': 8.361425338408343e-05, 'epoch': 0.56}) +[2026-06-15 23:04:26,871][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.17502953112125397, 'train/grad_norm_pre_clip': 10.272159576416016, 'timing/data': 0.012401124462485313, 'timing/model': 1.5237562842667103, 'learning_rate/qwen_vl_interface': 8.359123588131319e-06, 'learning_rate/action_model': 8.359123588131319e-05, 'epoch': 0.56}) +[2026-06-15 23:04:33,169][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.2807900905609131, 'train/grad_norm_pre_clip': 14.934815406799316, 'timing/data': 0.012570548802614212, 'timing/model': 1.5285765565931797, 'learning_rate/qwen_vl_interface': 8.35682055981103e-06, 'learning_rate/action_model': 8.35682055981103e-05, 'epoch': 0.56}) +[2026-06-15 23:04:39,508][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.20749710500240326, 'train/grad_norm_pre_clip': 11.114559173583984, 'timing/data': 0.011451039463281631, 'timing/model': 1.5920918714255095, 'learning_rate/qwen_vl_interface': 8.354516254394163e-06, 'learning_rate/action_model': 8.354516254394162e-05, 'epoch': 0.56}) +[2026-06-15 23:04:45,793][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.21889767050743103, 'train/grad_norm_pre_clip': 14.934083938598633, 'timing/data': 0.012032534927129745, 'timing/model': 1.5733670722693205, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 0.56}) +[2026-06-15 23:04:52,536][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.19904597103595734, 'train/grad_norm_pre_clip': 8.918803215026855, 'timing/data': 0.017721910029649734, 'timing/model': 1.5847954843193293, 'learning_rate/qwen_vl_interface': 8.349903816060074e-06, 'learning_rate/action_model': 8.349903816060075e-05, 'epoch': 0.56}) +[2026-06-15 23:04:59,266][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.1295279860496521, 'train/grad_norm_pre_clip': 10.30491828918457, 'timing/data': 0.01781037077307701, 'timing/model': 1.9383762907236814, 'learning_rate/qwen_vl_interface': 8.34759568503885e-06, 'learning_rate/action_model': 8.347595685038851e-05, 'epoch': 0.56}) +[2026-06-15 23:05:05,739][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.10098890215158463, 'train/grad_norm_pre_clip': 11.058022499084473, 'timing/data': 0.017557447776198387, 'timing/model': 1.6624558810144663, 'learning_rate/qwen_vl_interface': 8.345286280713049e-06, 'learning_rate/action_model': 8.345286280713048e-05, 'epoch': 0.56}) +[2026-06-15 23:05:12,567][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.013246907852590084, 'train/grad_norm_pre_clip': 0.9950096011161804, 'timing/data': 0.01843789592385292, 'timing/model': 1.6783337239176035, 'learning_rate/qwen_vl_interface': 8.342975604031974e-06, 'learning_rate/action_model': 8.342975604031974e-05, 'epoch': 0.56}) +[2026-06-15 23:05:19,345][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.37000197172164917, 'train/grad_norm_pre_clip': 30.046045303344727, 'timing/data': 0.01830541342496872, 'timing/model': 1.6670667063444853, 'learning_rate/qwen_vl_interface': 8.340663655945462e-06, 'learning_rate/action_model': 8.340663655945462e-05, 'epoch': 0.56}) +[2026-06-15 23:05:26,198][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.27090370655059814, 'train/grad_norm_pre_clip': 16.321882247924805, 'timing/data': 0.019360100850462914, 'timing/model': 1.6818780358880758, 'learning_rate/qwen_vl_interface': 8.338350437403866e-06, 'learning_rate/action_model': 8.338350437403865e-05, 'epoch': 0.56}) +[2026-06-15 23:05:32,959][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.26514932513237, 'train/grad_norm_pre_clip': 23.50596046447754, 'timing/data': 0.018828073516488075, 'timing/model': 1.6492155268788338, 'learning_rate/qwen_vl_interface': 8.336035949358063e-06, 'learning_rate/action_model': 8.336035949358062e-05, 'epoch': 0.56}) +[2026-06-15 23:05:39,462][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.19104479253292084, 'train/grad_norm_pre_clip': 11.52939510345459, 'timing/data': 0.012964203953742981, 'timing/model': 1.746081206947565, 'learning_rate/qwen_vl_interface': 8.333720192759448e-06, 'learning_rate/action_model': 8.333720192759448e-05, 'epoch': 0.56}) +[2026-06-15 23:05:45,724][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.2552888095378876, 'train/grad_norm_pre_clip': 11.553394317626953, 'timing/data': 0.012823408469557762, 'timing/model': 1.737048676237464, 'learning_rate/qwen_vl_interface': 8.331403168559946e-06, 'learning_rate/action_model': 8.331403168559945e-05, 'epoch': 0.56}) +[2026-06-15 23:05:51,744][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.0429874062538147, 'train/grad_norm_pre_clip': 4.758054256439209, 'timing/data': 0.012964334338903427, 'timing/model': 1.5530741587281227, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 0.56}) +[2026-06-15 23:05:58,040][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.19190971553325653, 'train/grad_norm_pre_clip': 11.907402038574219, 'timing/data': 0.016507908701896667, 'timing/model': 1.6380106043070555, 'learning_rate/qwen_vl_interface': 8.32676532116856e-06, 'learning_rate/action_model': 8.326765321168559e-05, 'epoch': 0.56}) +[2026-06-15 23:06:04,466][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.2772206664085388, 'train/grad_norm_pre_clip': 12.630825996398926, 'timing/data': 0.016193876042962074, 'timing/model': 1.5494091175496578, 'learning_rate/qwen_vl_interface': 8.324444499883119e-06, 'learning_rate/action_model': 8.324444499883119e-05, 'epoch': 0.56}) +[2026-06-15 23:06:10,756][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.22836866974830627, 'train/grad_norm_pre_clip': 10.53380298614502, 'timing/data': 0.016232872381806374, 'timing/model': 1.5356775131076574, 'learning_rate/qwen_vl_interface': 8.322122414809678e-06, 'learning_rate/action_model': 8.322122414809677e-05, 'epoch': 0.56}) +[2026-06-15 23:06:16,946][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.13207709789276123, 'train/grad_norm_pre_clip': 9.064696311950684, 'timing/data': 0.013185666874051094, 'timing/model': 1.7606778033077717, 'learning_rate/qwen_vl_interface': 8.319799066902755e-06, 'learning_rate/action_model': 8.319799066902755e-05, 'epoch': 0.56}) +[2026-06-15 23:06:23,138][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.0692346915602684, 'train/grad_norm_pre_clip': 7.33001184463501, 'timing/data': 0.014285480603575706, 'timing/model': 1.8585720714181662, 'learning_rate/qwen_vl_interface': 8.317474457117393e-06, 'learning_rate/action_model': 8.317474457117392e-05, 'epoch': 0.56}) +[2026-06-15 23:06:29,026][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.246253103017807, 'train/grad_norm_pre_clip': 8.6133394241333, 'timing/data': 0.010963791981339455, 'timing/model': 1.5403484795242548, 'learning_rate/qwen_vl_interface': 8.315148586409148e-06, 'learning_rate/action_model': 8.315148586409148e-05, 'epoch': 0.56}) +[2026-06-15 23:06:34,925][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.15325097739696503, 'train/grad_norm_pre_clip': 13.650822639465332, 'timing/data': 0.01016322337090969, 'timing/model': 1.4580250978469849, 'learning_rate/qwen_vl_interface': 8.312821455734104e-06, 'learning_rate/action_model': 8.312821455734103e-05, 'epoch': 0.56}) +[2026-06-15 23:06:40,821][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.13355274498462677, 'train/grad_norm_pre_clip': 10.745209693908691, 'timing/data': 0.010652007535099983, 'timing/model': 1.4526668563485146, 'learning_rate/qwen_vl_interface': 8.31049306604885e-06, 'learning_rate/action_model': 8.310493066048849e-05, 'epoch': 0.56}) +[2026-06-15 23:06:46,815][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.07736916840076447, 'train/grad_norm_pre_clip': 4.065188884735107, 'timing/data': 0.011031262576580048, 'timing/model': 1.5248303208500147, 'learning_rate/qwen_vl_interface': 8.3081634183105e-06, 'learning_rate/action_model': 8.308163418310501e-05, 'epoch': 0.56}) +[2026-06-15 23:06:52,812][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.0915897786617279, 'train/grad_norm_pre_clip': 4.781978607177734, 'timing/data': 0.010994981974363327, 'timing/model': 1.7284773364663124, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 0.57}) +[2026-06-15 23:06:58,708][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.1339104324579239, 'train/grad_norm_pre_clip': 7.062338829040527, 'timing/data': 0.011544084176421165, 'timing/model': 1.5239383038133383, 'learning_rate/qwen_vl_interface': 8.303500352505557e-06, 'learning_rate/action_model': 8.303500352505557e-05, 'epoch': 0.57}) +[2026-06-15 23:07:04,685][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.1545802801847458, 'train/grad_norm_pre_clip': 9.606241226196289, 'timing/data': 0.011328702792525291, 'timing/model': 1.4426498804241419, 'learning_rate/qwen_vl_interface': 8.301166936355773e-06, 'learning_rate/action_model': 8.301166936355773e-05, 'epoch': 0.57}) +[2026-06-15 23:07:10,588][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.2186346799135208, 'train/grad_norm_pre_clip': 7.6320624351501465, 'timing/data': 0.010144177824258804, 'timing/model': 1.445114802569151, 'learning_rate/qwen_vl_interface': 8.298832265986514e-06, 'learning_rate/action_model': 8.298832265986513e-05, 'epoch': 0.57}) +[2026-06-15 23:07:16,488][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.24004162847995758, 'train/grad_norm_pre_clip': 12.864788055419922, 'timing/data': 0.01109934039413929, 'timing/model': 1.4479852207005024, 'learning_rate/qwen_vl_interface': 8.296496342357475e-06, 'learning_rate/action_model': 8.296496342357474e-05, 'epoch': 0.57}) +[2026-06-15 23:07:22,429][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.2392221838235855, 'train/grad_norm_pre_clip': 8.054300308227539, 'timing/data': 0.010373080149292946, 'timing/model': 1.539379209280014, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 0.57}) +[2026-06-15 23:07:28,275][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.15150824189186096, 'train/grad_norm_pre_clip': 6.245942115783691, 'timing/data': 0.010018577799201012, 'timing/model': 1.6009112857282162, 'learning_rate/qwen_vl_interface': 8.29182073916141e-06, 'learning_rate/action_model': 8.29182073916141e-05, 'epoch': 0.57}) +[2026-06-15 23:07:34,126][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.1890103965997696, 'train/grad_norm_pre_clip': 7.717260837554932, 'timing/data': 0.011829350143671036, 'timing/model': 1.456529188901186, 'learning_rate/qwen_vl_interface': 8.28948106151635e-06, 'learning_rate/action_model': 8.28948106151635e-05, 'epoch': 0.57}) +[2026-06-15 23:07:40,031][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.15886761248111725, 'train/grad_norm_pre_clip': 7.968163967132568, 'timing/data': 0.010918274521827698, 'timing/model': 1.4573494270443916, 'learning_rate/qwen_vl_interface': 8.287140134455435e-06, 'learning_rate/action_model': 8.287140134455435e-05, 'epoch': 0.57}) +[2026-06-15 23:07:45,923][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.07206974178552628, 'train/grad_norm_pre_clip': 4.377898693084717, 'timing/data': 0.010833853855729103, 'timing/model': 1.458597494289279, 'learning_rate/qwen_vl_interface': 8.284797958940934e-06, 'learning_rate/action_model': 8.284797958940935e-05, 'epoch': 0.57}) +[2026-06-15 23:07:51,835][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.1374516636133194, 'train/grad_norm_pre_clip': 4.556503772735596, 'timing/data': 0.011185459792613983, 'timing/model': 1.4608846250921488, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 0.57}) +[2026-06-15 23:07:57,922][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.26734259724617004, 'train/grad_norm_pre_clip': 9.935818672180176, 'timing/data': 0.012721186503767967, 'timing/model': 1.5159864965826273, 'learning_rate/qwen_vl_interface': 8.280109866402805e-06, 'learning_rate/action_model': 8.280109866402806e-05, 'epoch': 0.57}) +[2026-06-15 23:08:03,884][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.10026675462722778, 'train/grad_norm_pre_clip': 5.280952453613281, 'timing/data': 0.012065785005688667, 'timing/model': 1.6361858770251274, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.57}) +[2026-06-15 23:08:09,948][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.1249527633190155, 'train/grad_norm_pre_clip': 14.421964645385742, 'timing/data': 0.01025058887898922, 'timing/model': 1.548506960272789, 'learning_rate/qwen_vl_interface': 8.275416791610356e-06, 'learning_rate/action_model': 8.275416791610355e-05, 'epoch': 0.57}) +[2026-06-15 23:08:16,241][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.17229555547237396, 'train/grad_norm_pre_clip': 5.899147033691406, 'timing/data': 0.013347536325454712, 'timing/model': 1.4785747863352299, 'learning_rate/qwen_vl_interface': 8.27306838827987e-06, 'learning_rate/action_model': 8.273068388279871e-05, 'epoch': 0.57}) +[2026-06-15 23:08:22,231][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.07456310093402863, 'train/grad_norm_pre_clip': 4.04578161239624, 'timing/data': 0.012660309672355652, 'timing/model': 1.5367239639163017, 'learning_rate/qwen_vl_interface': 8.270718742280162e-06, 'learning_rate/action_model': 8.270718742280162e-05, 'epoch': 0.57}) +[2026-06-15 23:08:28,209][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.11342447996139526, 'train/grad_norm_pre_clip': 4.428936004638672, 'timing/data': 0.01232239417731762, 'timing/model': 1.5268995594233274, 'learning_rate/qwen_vl_interface': 8.268367854577084e-06, 'learning_rate/action_model': 8.268367854577082e-05, 'epoch': 0.57}) +[2026-06-15 23:08:34,221][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.13260284066200256, 'train/grad_norm_pre_clip': 7.242731094360352, 'timing/data': 0.01140686683356762, 'timing/model': 1.7342810481786728, 'learning_rate/qwen_vl_interface': 8.266015726136991e-06, 'learning_rate/action_model': 8.266015726136991e-05, 'epoch': 0.57}) +[2026-06-15 23:08:40,024][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.1485392153263092, 'train/grad_norm_pre_clip': 7.177488327026367, 'timing/data': 0.011381767690181732, 'timing/model': 1.5378697253763676, 'learning_rate/qwen_vl_interface': 8.26366235792676e-06, 'learning_rate/action_model': 8.26366235792676e-05, 'epoch': 0.57}) +[2026-06-15 23:08:46,016][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.05400237813591957, 'train/grad_norm_pre_clip': 4.703453063964844, 'timing/data': 0.011016888543963432, 'timing/model': 1.5255952514708042, 'learning_rate/qwen_vl_interface': 8.261307750913769e-06, 'learning_rate/action_model': 8.261307750913769e-05, 'epoch': 0.57}) +[2026-06-15 23:08:51,896][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.041052501648664474, 'train/grad_norm_pre_clip': 2.6521430015563965, 'timing/data': 0.01033823937177658, 'timing/model': 1.5130787435919046, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 0.57}) +[2026-06-15 23:08:57,691][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.06334337592124939, 'train/grad_norm_pre_clip': 4.195317268371582, 'timing/data': 0.010060442611575127, 'timing/model': 1.443868512287736, 'learning_rate/qwen_vl_interface': 8.256594824351577e-06, 'learning_rate/action_model': 8.256594824351576e-05, 'epoch': 0.57}) +[2026-06-15 23:09:03,670][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.29188272356987, 'train/grad_norm_pre_clip': 8.491083145141602, 'timing/data': 0.011101428419351578, 'timing/model': 1.4918313790112734, 'learning_rate/qwen_vl_interface': 8.254236506739683e-06, 'learning_rate/action_model': 8.254236506739681e-05, 'epoch': 0.57}) +[2026-06-15 23:09:09,589][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.025302929803729057, 'train/grad_norm_pre_clip': 4.025911331176758, 'timing/data': 0.010765440762043, 'timing/model': 1.6416122056543827, 'learning_rate/qwen_vl_interface': 8.25187695419964e-06, 'learning_rate/action_model': 8.25187695419964e-05, 'epoch': 0.57}) +[2026-06-15 23:09:15,440][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.12045643478631973, 'train/grad_norm_pre_clip': 12.61342716217041, 'timing/data': 0.01275039091706276, 'timing/model': 1.5531695950776339, 'learning_rate/qwen_vl_interface': 8.249516167701373e-06, 'learning_rate/action_model': 8.249516167701374e-05, 'epoch': 0.57}) +[2026-06-15 23:09:21,626][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.45359450578689575, 'train/grad_norm_pre_clip': 14.59793472290039, 'timing/data': 0.013018287718296051, 'timing/model': 1.5453220549970865, 'learning_rate/qwen_vl_interface': 8.247154148215311e-06, 'learning_rate/action_model': 8.24715414821531e-05, 'epoch': 0.57}) +[2026-06-15 23:09:27,850][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.36788618564605713, 'train/grad_norm_pre_clip': 15.708869934082031, 'timing/data': 0.013521751388907433, 'timing/model': 1.5475735422223806, 'learning_rate/qwen_vl_interface': 8.24479089671239e-06, 'learning_rate/action_model': 8.24479089671239e-05, 'epoch': 0.58}) +[2026-06-15 23:09:33,961][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.38103047013282776, 'train/grad_norm_pre_clip': 20.03048324584961, 'timing/data': 0.01262018270790577, 'timing/model': 1.5544872134923935, 'learning_rate/qwen_vl_interface': 8.242426414164056e-06, 'learning_rate/action_model': 8.242426414164056e-05, 'epoch': 0.58}) +[2026-06-15 23:09:40,157][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.19343504309654236, 'train/grad_norm_pre_clip': 13.50640869140625, 'timing/data': 0.011834317818284035, 'timing/model': 1.539254356175661, 'learning_rate/qwen_vl_interface': 8.240060701542257e-06, 'learning_rate/action_model': 8.240060701542257e-05, 'epoch': 0.58}) +[2026-06-15 23:09:46,198][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.1914847493171692, 'train/grad_norm_pre_clip': 12.710148811340332, 'timing/data': 0.010144839063286781, 'timing/model': 1.6851061340421438, 'learning_rate/qwen_vl_interface': 8.237693759819447e-06, 'learning_rate/action_model': 8.237693759819448e-05, 'epoch': 0.58}) +[2026-06-15 23:09:51,893][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.005567991640418768, 'train/grad_norm_pre_clip': 1.1250399351119995, 'timing/data': 0.009860353544354439, 'timing/model': 1.4444548077881336, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 0.58}) +[2026-06-15 23:09:57,792][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.3207939863204956, 'train/grad_norm_pre_clip': 15.55095100402832, 'timing/data': 0.010077035054564476, 'timing/model': 1.4403506200760603, 'learning_rate/qwen_vl_interface': 8.232956192963143e-06, 'learning_rate/action_model': 8.232956192963144e-05, 'epoch': 0.58}) +[2026-06-15 23:10:03,697][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.24301722645759583, 'train/grad_norm_pre_clip': 15.393527030944824, 'timing/data': 0.01139463484287262, 'timing/model': 1.4461136721074581, 'learning_rate/qwen_vl_interface': 8.230585569777086e-06, 'learning_rate/action_model': 8.230585569777085e-05, 'epoch': 0.58}) +[2026-06-15 23:10:09,590][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.5892621278762817, 'train/grad_norm_pre_clip': 19.048187255859375, 'timing/data': 0.010881125926971436, 'timing/model': 1.4282545428723097, 'learning_rate/qwen_vl_interface': 8.228213721384885e-06, 'learning_rate/action_model': 8.228213721384885e-05, 'epoch': 0.58}) +[2026-06-15 23:10:15,688][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.444231778383255, 'train/grad_norm_pre_clip': 15.991133689880371, 'timing/data': 0.010840574279427528, 'timing/model': 1.461111718788743, 'learning_rate/qwen_vl_interface': 8.225840648761523e-06, 'learning_rate/action_model': 8.225840648761523e-05, 'epoch': 0.58}) +[2026-06-15 23:10:21,583][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.03614543750882149, 'train/grad_norm_pre_clip': 17.140926361083984, 'timing/data': 0.010150054469704628, 'timing/model': 1.649079617112875, 'learning_rate/qwen_vl_interface': 8.223466352882476e-06, 'learning_rate/action_model': 8.223466352882475e-05, 'epoch': 0.58}) +[2026-06-15 23:10:27,287][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.45199257135391235, 'train/grad_norm_pre_clip': 29.229312896728516, 'timing/data': 0.010955007746815681, 'timing/model': 1.4477300029248, 'learning_rate/qwen_vl_interface': 8.22109083472373e-06, 'learning_rate/action_model': 8.22109083472373e-05, 'epoch': 0.58}) +[2026-06-15 23:10:33,208][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.10734128206968307, 'train/grad_norm_pre_clip': 6.678073883056641, 'timing/data': 0.010832469910383224, 'timing/model': 1.474425945430994, 'learning_rate/qwen_vl_interface': 8.218714095261769e-06, 'learning_rate/action_model': 8.218714095261768e-05, 'epoch': 0.58}) +[2026-06-15 23:10:39,092][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.3316463530063629, 'train/grad_norm_pre_clip': 22.66439437866211, 'timing/data': 0.01059836894273758, 'timing/model': 1.4537656735628843, 'learning_rate/qwen_vl_interface': 8.21633613547358e-06, 'learning_rate/action_model': 8.21633613547358e-05, 'epoch': 0.58}) +[2026-06-15 23:10:44,878][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.0815158411860466, 'train/grad_norm_pre_clip': 5.545979022979736, 'timing/data': 0.010384231805801392, 'timing/model': 1.4424366876482964, 'learning_rate/qwen_vl_interface': 8.213956956336656e-06, 'learning_rate/action_model': 8.213956956336656e-05, 'epoch': 0.58}) +[2026-06-15 23:18:25,816][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.11420834809541702, 'train/grad_norm_pre_clip': 6.189693927764893, 'eval/loss': 0.06713058054447174, 'eval/flap_precision': 0.9328859060402684, 'eval/flap_recall': 0.9430122116689281, 'eval/flap_f1': 0.9379217273954117, 'eval/flap_rate_model': 0.20694444444444443, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9379217273954117, 'timing/data': 0.010701728984713554, 'timing/model': 1.4153763744980097, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.58}) +[2026-06-15 23:18:25,817][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1500_state +[2026-06-15 23:18:25,823][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:18:34,655][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1500_state/model.safetensors +[2026-06-15 23:18:50,929][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1500_state/optimizer.bin +[2026-06-15 23:18:50,929][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1500_state/sampler.bin +[2026-06-15 23:18:50,930][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-15 23:18:50,931][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-15 23:18:51,019][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-15 23:18:51,019][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:18:51,027][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:18:56,902][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.428238183259964, 'train/grad_norm_pre_clip': 17.894046783447266, 'timing/data': 0.01028454676270485, 'timing/model': 1.427489073947072, 'learning_rate/qwen_vl_interface': 8.209194943929059e-06, 'learning_rate/action_model': 8.20919494392906e-05, 'epoch': 0.58}) +[2026-06-15 23:19:02,800][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.18622314929962158, 'train/grad_norm_pre_clip': 9.446183204650879, 'timing/data': 0.010911963880062103, 'timing/model': 1.4634916745126247, 'learning_rate/qwen_vl_interface': 8.20681211261587e-06, 'learning_rate/action_model': 8.20681211261587e-05, 'epoch': 0.58}) +[2026-06-15 23:19:08,704][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.1412009745836258, 'train/grad_norm_pre_clip': 9.796512603759766, 'timing/data': 0.010900421068072319, 'timing/model': 1.6404954884201288, 'learning_rate/qwen_vl_interface': 8.204428065868911e-06, 'learning_rate/action_model': 8.204428065868911e-05, 'epoch': 0.58}) +[2026-06-15 23:19:14,403][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.12143730372190475, 'train/grad_norm_pre_clip': 6.685035228729248, 'timing/data': 0.011188767850399017, 'timing/model': 1.4511165227741003, 'learning_rate/qwen_vl_interface': 8.202042804668173e-06, 'learning_rate/action_model': 8.202042804668172e-05, 'epoch': 0.58}) +[2026-06-15 23:19:20,404][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.1285545378923416, 'train/grad_norm_pre_clip': 6.7089972496032715, 'timing/data': 0.012182369828224182, 'timing/model': 1.4509343765676022, 'learning_rate/qwen_vl_interface': 8.199656329994145e-06, 'learning_rate/action_model': 8.199656329994145e-05, 'epoch': 0.58}) +[2026-06-15 23:19:26,305][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.21830089390277863, 'train/grad_norm_pre_clip': 13.52292537689209, 'timing/data': 0.010651109740138054, 'timing/model': 1.454224780201912, 'learning_rate/qwen_vl_interface': 8.19726864282782e-06, 'learning_rate/action_model': 8.197268642827818e-05, 'epoch': 0.58}) +[2026-06-15 23:19:32,131][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.17176178097724915, 'train/grad_norm_pre_clip': 5.330009937286377, 'timing/data': 0.010486844927072525, 'timing/model': 1.5070178229361773, 'learning_rate/qwen_vl_interface': 8.194879744150682e-06, 'learning_rate/action_model': 8.194879744150681e-05, 'epoch': 0.58}) +[2026-06-15 23:19:37,980][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.24540290236473083, 'train/grad_norm_pre_clip': 6.387383937835693, 'timing/data': 0.013377930968999863, 'timing/model': 1.4353149197995663, 'learning_rate/qwen_vl_interface': 8.192489634944718e-06, 'learning_rate/action_model': 8.192489634944718e-05, 'epoch': 0.58}) +[2026-06-15 23:19:43,878][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.062452178448438644, 'train/grad_norm_pre_clip': 4.777509689331055, 'timing/data': 0.010814541950821877, 'timing/model': 1.653707580640912, 'learning_rate/qwen_vl_interface': 8.190098316192416e-06, 'learning_rate/action_model': 8.190098316192416e-05, 'epoch': 0.58}) +[2026-06-15 23:19:49,583][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.25278952717781067, 'train/grad_norm_pre_clip': 5.4917802810668945, 'timing/data': 0.01037675328552723, 'timing/model': 1.4572648089379072, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 0.58}) +[2026-06-15 23:19:55,481][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.15806648135185242, 'train/grad_norm_pre_clip': 5.962818622589111, 'timing/data': 0.011238807812333107, 'timing/model': 1.4543907530605793, 'learning_rate/qwen_vl_interface': 8.185312053981204e-06, 'learning_rate/action_model': 8.185312053981204e-05, 'epoch': 0.58}) +[2026-06-15 23:20:01,594][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.14496764540672302, 'train/grad_norm_pre_clip': 6.9747443199157715, 'timing/data': 0.012761963531374931, 'timing/model': 1.6500105764716864, 'learning_rate/qwen_vl_interface': 8.182917112489747e-06, 'learning_rate/action_model': 8.182917112489746e-05, 'epoch': 0.59}) +[2026-06-15 23:20:07,883][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.26172858476638794, 'train/grad_norm_pre_clip': 8.407585144042969, 'timing/data': 0.013110028579831123, 'timing/model': 1.5545158255845308, 'learning_rate/qwen_vl_interface': 8.18052096538685e-06, 'learning_rate/action_model': 8.18052096538685e-05, 'epoch': 0.59}) +[2026-06-15 23:20:14,318][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.1667521595954895, 'train/grad_norm_pre_clip': 3.7480757236480713, 'timing/data': 0.0143735371530056, 'timing/model': 1.6031214762479067, 'learning_rate/qwen_vl_interface': 8.178123613657483e-06, 'learning_rate/action_model': 8.178123613657482e-05, 'epoch': 0.59}) +[2026-06-15 23:20:20,680][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.11815953999757767, 'train/grad_norm_pre_clip': 3.658691883087158, 'timing/data': 0.028087079524993896, 'timing/model': 1.8304189350456, 'learning_rate/qwen_vl_interface': 8.175725058287102e-06, 'learning_rate/action_model': 8.175725058287102e-05, 'epoch': 0.59}) +[2026-06-15 23:20:26,990][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.08670224994421005, 'train/grad_norm_pre_clip': 3.9938137531280518, 'timing/data': 0.01637907698750496, 'timing/model': 1.551151305437088, 'learning_rate/qwen_vl_interface': 8.173325300261664e-06, 'learning_rate/action_model': 8.173325300261663e-05, 'epoch': 0.59}) +[2026-06-15 23:20:33,371][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.08792238682508469, 'train/grad_norm_pre_clip': 5.361588478088379, 'timing/data': 0.01543419249355793, 'timing/model': 1.6397684011608362, 'learning_rate/qwen_vl_interface': 8.170924340567617e-06, 'learning_rate/action_model': 8.170924340567616e-05, 'epoch': 0.59}) +[2026-06-15 23:20:39,641][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.08388454467058182, 'train/grad_norm_pre_clip': 4.965169906616211, 'timing/data': 0.011686228215694427, 'timing/model': 1.5468906629830599, 'learning_rate/qwen_vl_interface': 8.168522180191909e-06, 'learning_rate/action_model': 8.168522180191909e-05, 'epoch': 0.59}) +[2026-06-15 23:20:45,839][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.15460583567619324, 'train/grad_norm_pre_clip': 5.683926582336426, 'timing/data': 0.012813810259103775, 'timing/model': 1.5510518942028284, 'learning_rate/qwen_vl_interface': 8.166118820121972e-06, 'learning_rate/action_model': 8.166118820121973e-05, 'epoch': 0.59}) +[2026-06-15 23:20:52,139][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.146454319357872, 'train/grad_norm_pre_clip': 6.326501846313477, 'timing/data': 0.013037560507655144, 'timing/model': 1.5540364664047956, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 0.59}) +[2026-06-15 23:20:58,272][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.22317226231098175, 'train/grad_norm_pre_clip': 10.309977531433105, 'timing/data': 0.012266865000128746, 'timing/model': 1.735839568078518, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.59}) +[2026-06-15 23:21:04,341][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.07252126187086105, 'train/grad_norm_pre_clip': 5.19499397277832, 'timing/data': 0.011992543935775757, 'timing/model': 1.5579339638352394, 'learning_rate/qwen_vl_interface': 8.158901551628579e-06, 'learning_rate/action_model': 8.158901551628579e-05, 'epoch': 0.59}) +[2026-06-15 23:21:10,437][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.060354020446538925, 'train/grad_norm_pre_clip': 10.50456714630127, 'timing/data': 0.012393742799758911, 'timing/model': 1.5153258312493563, 'learning_rate/qwen_vl_interface': 8.156493402665971e-06, 'learning_rate/action_model': 8.15649340266597e-05, 'epoch': 0.59}) +[2026-06-15 23:21:16,539][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.07734816521406174, 'train/grad_norm_pre_clip': 7.165558815002441, 'timing/data': 0.012544140219688416, 'timing/model': 1.5375597532838583, 'learning_rate/qwen_vl_interface': 8.154084058953713e-06, 'learning_rate/action_model': 8.154084058953713e-05, 'epoch': 0.59}) +[2026-06-15 23:21:22,590][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.24511991441249847, 'train/grad_norm_pre_clip': 21.612457275390625, 'timing/data': 0.01017846167087555, 'timing/model': 1.4486146438866854, 'learning_rate/qwen_vl_interface': 8.151673521482193e-06, 'learning_rate/action_model': 8.151673521482194e-05, 'epoch': 0.59}) +[2026-06-15 23:21:28,496][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.018336869776248932, 'train/grad_norm_pre_clip': 1.757142186164856, 'timing/data': 0.010287594050168991, 'timing/model': 1.4276878200471401, 'learning_rate/qwen_vl_interface': 8.149261791242297e-06, 'learning_rate/action_model': 8.149261791242297e-05, 'epoch': 0.59}) +[2026-06-15 23:21:34,391][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.12962482869625092, 'train/grad_norm_pre_clip': 4.982252597808838, 'timing/data': 0.01009838841855526, 'timing/model': 1.6476309709250927, 'learning_rate/qwen_vl_interface': 8.146848869225391e-06, 'learning_rate/action_model': 8.146848869225391e-05, 'epoch': 0.59}) +[2026-06-15 23:21:40,500][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.023472554981708527, 'train/grad_norm_pre_clip': 3.831505298614502, 'timing/data': 0.010458843782544136, 'timing/model': 1.6582945212721825, 'learning_rate/qwen_vl_interface': 8.14443475642334e-06, 'learning_rate/action_model': 8.144434756423339e-05, 'epoch': 0.59}) +[2026-06-15 23:21:46,192][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.3454315662384033, 'train/grad_norm_pre_clip': 13.0096435546875, 'timing/data': 0.010779345408082008, 'timing/model': 1.4478277824819088, 'learning_rate/qwen_vl_interface': 8.142019453828493e-06, 'learning_rate/action_model': 8.142019453828493e-05, 'epoch': 0.59}) +[2026-06-15 23:21:52,095][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.1967458724975586, 'train/grad_norm_pre_clip': 13.548792839050293, 'timing/data': 0.010548504069447517, 'timing/model': 1.446508727967739, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 0.59}) +[2026-06-15 23:21:57,991][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.040218185633420944, 'train/grad_norm_pre_clip': 7.259700298309326, 'timing/data': 0.009666793048381805, 'timing/model': 1.4496263228356838, 'learning_rate/qwen_vl_interface': 8.137185283232255e-06, 'learning_rate/action_model': 8.137185283232256e-05, 'epoch': 0.59}) +[2026-06-15 23:22:03,905][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.050711728632450104, 'train/grad_norm_pre_clip': 7.575284481048584, 'timing/data': 0.009961068630218506, 'timing/model': 1.4322296306490898, 'learning_rate/qwen_vl_interface': 8.134766417218014e-06, 'learning_rate/action_model': 8.134766417218012e-05, 'epoch': 0.59}) +[2026-06-15 23:22:09,813][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.3819281756877899, 'train/grad_norm_pre_clip': 19.30760955810547, 'timing/data': 0.01042640395462513, 'timing/model': 1.6681476589292288, 'learning_rate/qwen_vl_interface': 8.132346365385262e-06, 'learning_rate/action_model': 8.132346365385261e-05, 'epoch': 0.59}) +[2026-06-15 23:22:15,497][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.08081476390361786, 'train/grad_norm_pre_clip': 10.958041191101074, 'timing/data': 0.01042848825454712, 'timing/model': 1.4518011678010225, 'learning_rate/qwen_vl_interface': 8.129925128728796e-06, 'learning_rate/action_model': 8.129925128728796e-05, 'epoch': 0.59}) +[2026-06-15 23:22:21,412][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.3428287208080292, 'train/grad_norm_pre_clip': 25.571142196655273, 'timing/data': 0.01150842197239399, 'timing/model': 1.464907368645072, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 0.59}) +[2026-06-15 23:22:27,297][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.10455098003149033, 'train/grad_norm_pre_clip': 7.4708404541015625, 'timing/data': 0.010554514825344086, 'timing/model': 1.4398252312093973, 'learning_rate/qwen_vl_interface': 8.125079104926324e-06, 'learning_rate/action_model': 8.125079104926323e-05, 'epoch': 0.59}) +[2026-06-15 23:22:33,196][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.3976784944534302, 'train/grad_norm_pre_clip': 28.769079208374023, 'timing/data': 0.010926607996225357, 'timing/model': 1.441109037026763, 'learning_rate/qwen_vl_interface': 8.122654319772332e-06, 'learning_rate/action_model': 8.122654319772332e-05, 'epoch': 0.59}) +[2026-06-15 23:22:39,073][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.038154374808073044, 'train/grad_norm_pre_clip': 2.4798238277435303, 'timing/data': 0.010679511353373528, 'timing/model': 1.4946599677205086, 'learning_rate/qwen_vl_interface': 8.120228353778661e-06, 'learning_rate/action_model': 8.12022835377866e-05, 'epoch': 0.6}) +[2026-06-15 23:22:44,944][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.08988351374864578, 'train/grad_norm_pre_clip': 15.860891342163086, 'timing/data': 0.010208377614617348, 'timing/model': 1.600244291126728, 'learning_rate/qwen_vl_interface': 8.117801207942528e-06, 'learning_rate/action_model': 8.117801207942528e-05, 'epoch': 0.6}) +[2026-06-15 23:22:50,599][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.2501552402973175, 'train/grad_norm_pre_clip': 14.590563774108887, 'timing/data': 0.010806923732161522, 'timing/model': 1.4590680785477161, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 0.6}) +[2026-06-15 23:22:56,613][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.016488807275891304, 'train/grad_norm_pre_clip': 2.126943588256836, 'timing/data': 0.010345296934247017, 'timing/model': 1.5467153415083885, 'learning_rate/qwen_vl_interface': 8.11294338073421e-06, 'learning_rate/action_model': 8.11294338073421e-05, 'epoch': 0.6}) +[2026-06-15 23:23:02,477][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.16626542806625366, 'train/grad_norm_pre_clip': 14.014033317565918, 'timing/data': 0.010786909610033035, 'timing/model': 1.4482217207551003, 'learning_rate/qwen_vl_interface': 8.110512701358892e-06, 'learning_rate/action_model': 8.110512701358891e-05, 'epoch': 0.6}) +[2026-06-15 23:23:08,778][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.10963760316371918, 'train/grad_norm_pre_clip': 8.201210021972656, 'timing/data': 0.01257101260125637, 'timing/model': 1.554980343207717, 'learning_rate/qwen_vl_interface': 8.108080846134852e-06, 'learning_rate/action_model': 8.108080846134852e-05, 'epoch': 0.6}) +[2026-06-15 23:23:14,941][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.01746298000216484, 'train/grad_norm_pre_clip': 1.213138222694397, 'timing/data': 0.012116845697164536, 'timing/model': 1.4869954846799374, 'learning_rate/qwen_vl_interface': 8.105647816061734e-06, 'learning_rate/action_model': 8.105647816061735e-05, 'epoch': 0.6}) +[2026-06-15 23:23:21,030][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.3043215572834015, 'train/grad_norm_pre_clip': 13.910492897033691, 'timing/data': 0.012252753600478172, 'timing/model': 1.6656245049089193, 'learning_rate/qwen_vl_interface': 8.103213612139667e-06, 'learning_rate/action_model': 8.103213612139667e-05, 'epoch': 0.6}) +[2026-06-15 23:23:26,993][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.10465463995933533, 'train/grad_norm_pre_clip': 15.5311279296875, 'timing/data': 0.012959584593772888, 'timing/model': 1.5196669269353151, 'learning_rate/qwen_vl_interface': 8.100778235369262e-06, 'learning_rate/action_model': 8.10077823536926e-05, 'epoch': 0.6}) +[2026-06-15 23:23:33,260][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.25151246786117554, 'train/grad_norm_pre_clip': 19.549428939819336, 'timing/data': 0.011954942718148232, 'timing/model': 1.5700959134846926, 'learning_rate/qwen_vl_interface': 8.098341686751605e-06, 'learning_rate/action_model': 8.098341686751605e-05, 'epoch': 0.6}) +[2026-06-15 23:23:39,451][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.06752408295869827, 'train/grad_norm_pre_clip': 4.035189628601074, 'timing/data': 0.013059983029961586, 'timing/model': 1.5416992008686066, 'learning_rate/qwen_vl_interface': 8.095903967288273e-06, 'learning_rate/action_model': 8.095903967288272e-05, 'epoch': 0.6}) +[2026-06-15 23:23:45,764][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.008447093889117241, 'train/grad_norm_pre_clip': 1.0649641752243042, 'timing/data': 0.01179317943751812, 'timing/model': 1.5786847416311502, 'learning_rate/qwen_vl_interface': 8.093465077981318e-06, 'learning_rate/action_model': 8.093465077981318e-05, 'epoch': 0.6}) +[2026-06-15 23:23:52,074][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.14090991020202637, 'train/grad_norm_pre_clip': 10.611392974853516, 'timing/data': 0.012512534856796265, 'timing/model': 1.7594554089009762, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 0.6}) +[2026-06-15 23:23:58,144][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.05498480424284935, 'train/grad_norm_pre_clip': 6.772032260894775, 'timing/data': 0.01212223805487156, 'timing/model': 1.5545298494398594, 'learning_rate/qwen_vl_interface': 8.088583793847165e-06, 'learning_rate/action_model': 8.088583793847164e-05, 'epoch': 0.6}) +[2026-06-15 23:24:04,344][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.18861901760101318, 'train/grad_norm_pre_clip': 5.135129451751709, 'timing/data': 0.011776451021432877, 'timing/model': 1.5315914675593376, 'learning_rate/qwen_vl_interface': 8.086141401026478e-06, 'learning_rate/action_model': 8.086141401026478e-05, 'epoch': 0.6}) +[2026-06-15 23:24:10,441][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.17054258286952972, 'train/grad_norm_pre_clip': 9.960932731628418, 'timing/data': 0.011686448007822037, 'timing/model': 1.5581063255667686, 'learning_rate/qwen_vl_interface': 8.083697842375193e-06, 'learning_rate/action_model': 8.083697842375193e-05, 'epoch': 0.6}) +[2026-06-15 23:24:16,545][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.1359335482120514, 'train/grad_norm_pre_clip': 11.644204139709473, 'timing/data': 0.012109063565731049, 'timing/model': 1.5536827333271503, 'learning_rate/qwen_vl_interface': 8.081253118897762e-06, 'learning_rate/action_model': 8.081253118897761e-05, 'epoch': 0.6}) +[2026-06-15 23:24:22,660][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.2051563411951065, 'train/grad_norm_pre_clip': 20.1403865814209, 'timing/data': 0.012475479394197464, 'timing/model': 1.4863079003989697, 'learning_rate/qwen_vl_interface': 8.07880723159912e-06, 'learning_rate/action_model': 8.07880723159912e-05, 'epoch': 0.6}) +[2026-06-15 23:24:28,861][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.1440441757440567, 'train/grad_norm_pre_clip': 10.732966423034668, 'timing/data': 0.012097738683223724, 'timing/model': 1.7846818286925554, 'learning_rate/qwen_vl_interface': 8.076360181484676e-06, 'learning_rate/action_model': 8.076360181484676e-05, 'epoch': 0.6}) +[2026-06-15 23:24:34,842][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.02426513470709324, 'train/grad_norm_pre_clip': 2.4995388984680176, 'timing/data': 0.012887248769402504, 'timing/model': 1.5440771523863077, 'learning_rate/qwen_vl_interface': 8.073911969560326e-06, 'learning_rate/action_model': 8.073911969560326e-05, 'epoch': 0.6}) +[2026-06-15 23:24:41,043][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.2118569016456604, 'train/grad_norm_pre_clip': 9.02227783203125, 'timing/data': 0.012351740151643753, 'timing/model': 1.5577952098101377, 'learning_rate/qwen_vl_interface': 8.071462596832432e-06, 'learning_rate/action_model': 8.071462596832432e-05, 'epoch': 0.6}) +[2026-06-15 23:24:47,244][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.3120558559894562, 'train/grad_norm_pre_clip': 13.19360637664795, 'timing/data': 0.01158100739121437, 'timing/model': 1.5531618930399418, 'learning_rate/qwen_vl_interface': 8.06901206430784e-06, 'learning_rate/action_model': 8.06901206430784e-05, 'epoch': 0.6}) +[2026-06-15 23:24:53,343][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.13809269666671753, 'train/grad_norm_pre_clip': 8.98160457611084, 'timing/data': 0.012068541720509529, 'timing/model': 1.547355830669403, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 0.6}) +[2026-06-15 23:24:59,646][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.4165569841861725, 'train/grad_norm_pre_clip': 18.391550064086914, 'timing/data': 0.012712199240922928, 'timing/model': 1.5281522907316685, 'learning_rate/qwen_vl_interface': 8.064107523898327e-06, 'learning_rate/action_model': 8.064107523898326e-05, 'epoch': 0.6}) +[2026-06-15 23:25:05,789][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.4376392960548401, 'train/grad_norm_pre_clip': 19.57771110534668, 'timing/data': 0.011951714754104614, 'timing/model': 1.716226726770401, 'learning_rate/qwen_vl_interface': 8.061653518029476e-06, 'learning_rate/action_model': 8.061653518029476e-05, 'epoch': 0.6}) +[2026-06-15 23:25:11,827][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.16828779876232147, 'train/grad_norm_pre_clip': 10.35572624206543, 'timing/data': 0.013343824073672295, 'timing/model': 1.550538880750537, 'learning_rate/qwen_vl_interface': 8.059198356396069e-06, 'learning_rate/action_model': 8.059198356396068e-05, 'epoch': 0.6}) +[2026-06-15 23:25:18,033][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.1369876265525818, 'train/grad_norm_pre_clip': 7.676822185516357, 'timing/data': 0.013120833784341812, 'timing/model': 1.5552556291222572, 'learning_rate/qwen_vl_interface': 8.05674204000733e-06, 'learning_rate/action_model': 8.05674204000733e-05, 'epoch': 0.61}) +[2026-06-15 23:25:24,259][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.33550024032592773, 'train/grad_norm_pre_clip': 7.4377312660217285, 'timing/data': 0.013177711516618729, 'timing/model': 1.5414915103465319, 'learning_rate/qwen_vl_interface': 8.05428456987296e-06, 'learning_rate/action_model': 8.05428456987296e-05, 'epoch': 0.61}) +[2026-06-15 23:25:30,458][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.1998850554227829, 'train/grad_norm_pre_clip': 6.996608257293701, 'timing/data': 0.012458786368370056, 'timing/model': 1.54762651771307, 'learning_rate/qwen_vl_interface': 8.05182594700313e-06, 'learning_rate/action_model': 8.05182594700313e-05, 'epoch': 0.61}) +[2026-06-15 23:25:36,844][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.06986238807439804, 'train/grad_norm_pre_clip': 4.518219470977783, 'timing/data': 0.01230422779917717, 'timing/model': 1.5485425256192684, 'learning_rate/qwen_vl_interface': 8.049366172408484e-06, 'learning_rate/action_model': 8.049366172408485e-05, 'epoch': 0.61}) +[2026-06-15 23:25:43,194][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.13229748606681824, 'train/grad_norm_pre_clip': 8.74763011932373, 'timing/data': 0.012866564095020294, 'timing/model': 1.7425639387220144, 'learning_rate/qwen_vl_interface': 8.04690524710015e-06, 'learning_rate/action_model': 8.046905247100151e-05, 'epoch': 0.61}) +[2026-06-15 23:25:49,270][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.30754920840263367, 'train/grad_norm_pre_clip': 15.504704475402832, 'timing/data': 0.012772105634212494, 'timing/model': 1.560605976730585, 'learning_rate/qwen_vl_interface': 8.044443172089715e-06, 'learning_rate/action_model': 8.044443172089714e-05, 'epoch': 0.61}) +[2026-06-15 23:25:55,678][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.1448114961385727, 'train/grad_norm_pre_clip': 5.754133701324463, 'timing/data': 0.011982902884483337, 'timing/model': 1.5569261778146029, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.61}) +[2026-06-15 23:26:02,088][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.5506370067596436, 'train/grad_norm_pre_clip': 20.053987503051758, 'timing/data': 0.010470660403370857, 'timing/model': 1.543475715443492, 'learning_rate/qwen_vl_interface': 8.039515577011287e-06, 'learning_rate/action_model': 8.039515577011288e-05, 'epoch': 0.61}) +[2026-06-15 23:26:08,378][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.33752748370170593, 'train/grad_norm_pre_clip': 13.50062084197998, 'timing/data': 0.012401653453707695, 'timing/model': 1.540800217539072, 'learning_rate/qwen_vl_interface': 8.037050058968841e-06, 'learning_rate/action_model': 8.03705005896884e-05, 'epoch': 0.61}) +[2026-06-15 23:26:14,781][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.07159175723791122, 'train/grad_norm_pre_clip': 4.065396308898926, 'timing/data': 0.0164348091930151, 'timing/model': 1.5922643672674894, 'learning_rate/qwen_vl_interface': 8.034583395275392e-06, 'learning_rate/action_model': 8.034583395275392e-05, 'epoch': 0.61}) +[2026-06-15 23:26:21,120][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.07875421643257141, 'train/grad_norm_pre_clip': 3.897524118423462, 'timing/data': 0.014192137867212296, 'timing/model': 1.7694942317903042, 'learning_rate/qwen_vl_interface': 8.032115586944893e-06, 'learning_rate/action_model': 8.032115586944892e-05, 'epoch': 0.61}) +[2026-06-15 23:26:27,189][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.16371646523475647, 'train/grad_norm_pre_clip': 4.183848857879639, 'timing/data': 0.01397259533405304, 'timing/model': 1.5405425410717726, 'learning_rate/qwen_vl_interface': 8.029646634991762e-06, 'learning_rate/action_model': 8.029646634991762e-05, 'epoch': 0.61}) +[2026-06-15 23:26:33,581][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.12897734344005585, 'train/grad_norm_pre_clip': 4.3852643966674805, 'timing/data': 0.010923197492957115, 'timing/model': 1.5587700586766005, 'learning_rate/qwen_vl_interface': 8.0271765404309e-06, 'learning_rate/action_model': 8.0271765404309e-05, 'epoch': 0.61}) +[2026-06-15 23:26:39,893][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.1372525840997696, 'train/grad_norm_pre_clip': 7.626558780670166, 'timing/data': 0.012930268421769142, 'timing/model': 1.5800755992531776, 'learning_rate/qwen_vl_interface': 8.024705304277664e-06, 'learning_rate/action_model': 8.024705304277664e-05, 'epoch': 0.61}) +[2026-06-15 23:26:46,172][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.06565156579017639, 'train/grad_norm_pre_clip': 3.8351478576660156, 'timing/data': 0.012955274432897568, 'timing/model': 1.556120129302144, 'learning_rate/qwen_vl_interface': 8.022232927547888e-06, 'learning_rate/action_model': 8.022232927547886e-05, 'epoch': 0.61}) +[2026-06-15 23:26:52,507][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.16762682795524597, 'train/grad_norm_pre_clip': 10.662510871887207, 'timing/data': 0.012177148833870888, 'timing/model': 1.543439982458949, 'learning_rate/qwen_vl_interface': 8.01975941125787e-06, 'learning_rate/action_model': 8.01975941125787e-05, 'epoch': 0.61}) +[2026-06-15 23:26:58,818][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.16156810522079468, 'train/grad_norm_pre_clip': 8.946747779846191, 'timing/data': 0.013074124231934547, 'timing/model': 1.8031601812690496, 'learning_rate/qwen_vl_interface': 8.017284756424381e-06, 'learning_rate/action_model': 8.017284756424381e-05, 'epoch': 0.61}) +[2026-06-15 23:27:04,883][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.2113550901412964, 'train/grad_norm_pre_clip': 4.634838581085205, 'timing/data': 0.011483851820230484, 'timing/model': 1.5263531226664782, 'learning_rate/qwen_vl_interface': 8.01480896406466e-06, 'learning_rate/action_model': 8.01480896406466e-05, 'epoch': 0.61}) +[2026-06-15 23:27:11,396][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.07036097347736359, 'train/grad_norm_pre_clip': 4.188699722290039, 'timing/data': 0.013127600774168968, 'timing/model': 1.546048590913415, 'learning_rate/qwen_vl_interface': 8.01233203519641e-06, 'learning_rate/action_model': 8.012332035196409e-05, 'epoch': 0.61}) +[2026-06-15 23:27:17,668][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.25982996821403503, 'train/grad_norm_pre_clip': 5.310091972351074, 'timing/data': 0.01223115622997284, 'timing/model': 1.5192416217178106, 'learning_rate/qwen_vl_interface': 8.009853970837801e-06, 'learning_rate/action_model': 8.009853970837801e-05, 'epoch': 0.61}) +[2026-06-15 23:27:24,085][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.05450674891471863, 'train/grad_norm_pre_clip': 3.9275646209716797, 'timing/data': 0.013283869251608849, 'timing/model': 1.5871651470661163, 'learning_rate/qwen_vl_interface': 8.007374772007474e-06, 'learning_rate/action_model': 8.007374772007474e-05, 'epoch': 0.61}) +[2026-06-15 23:27:30,497][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.24370737373828888, 'train/grad_norm_pre_clip': 7.085195541381836, 'timing/data': 0.018784604966640472, 'timing/model': 1.6050709579139948, 'learning_rate/qwen_vl_interface': 8.004894439724533e-06, 'learning_rate/action_model': 8.004894439724533e-05, 'epoch': 0.61}) +[2026-06-15 23:27:37,057][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.04481683298945427, 'train/grad_norm_pre_clip': 7.66007661819458, 'timing/data': 0.01546376384794712, 'timing/model': 1.9186953008174896, 'learning_rate/qwen_vl_interface': 8.002412975008552e-06, 'learning_rate/action_model': 8.002412975008552e-05, 'epoch': 0.61}) +[2026-06-15 23:27:43,136][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.23640665411949158, 'train/grad_norm_pre_clip': 19.839981079101562, 'timing/data': 0.012926086783409119, 'timing/model': 1.517650030553341, 'learning_rate/qwen_vl_interface': 7.999930378879562e-06, 'learning_rate/action_model': 7.999930378879561e-05, 'epoch': 0.61}) +[2026-06-15 23:27:49,354][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.07424739003181458, 'train/grad_norm_pre_clip': 6.919114112854004, 'timing/data': 0.013211017474532127, 'timing/model': 1.5531916692852974, 'learning_rate/qwen_vl_interface': 7.997446652358067e-06, 'learning_rate/action_model': 7.997446652358068e-05, 'epoch': 0.61}) +[2026-06-15 23:27:55,744][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.16311924159526825, 'train/grad_norm_pre_clip': 5.721022129058838, 'timing/data': 0.027409998700022697, 'timing/model': 1.5225429590791464, 'learning_rate/qwen_vl_interface': 7.994961796465035e-06, 'learning_rate/action_model': 7.994961796465035e-05, 'epoch': 0.61}) +[2026-06-15 23:28:01,935][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.109584279358387, 'train/grad_norm_pre_clip': 13.395867347717285, 'timing/data': 0.012689732015132904, 'timing/model': 1.5093656089156866, 'learning_rate/qwen_vl_interface': 7.992475812221894e-06, 'learning_rate/action_model': 7.992475812221894e-05, 'epoch': 0.62}) +[2026-06-15 23:28:08,138][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.09180187433958054, 'train/grad_norm_pre_clip': 14.265453338623047, 'timing/data': 0.013445889577269554, 'timing/model': 1.5459974985569715, 'learning_rate/qwen_vl_interface': 7.989988700650538e-06, 'learning_rate/action_model': 7.989988700650538e-05, 'epoch': 0.62}) +[2026-06-15 23:28:14,077][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.45497506856918335, 'train/grad_norm_pre_clip': 14.164198875427246, 'timing/data': 0.011457027867436409, 'timing/model': 1.6501109208911657, 'learning_rate/qwen_vl_interface': 7.987500462773326e-06, 'learning_rate/action_model': 7.987500462773326e-05, 'epoch': 0.62}) +[2026-06-15 23:28:19,993][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.1464601308107376, 'train/grad_norm_pre_clip': 9.82203483581543, 'timing/data': 0.011717453598976135, 'timing/model': 1.6668735947459936, 'learning_rate/qwen_vl_interface': 7.985011099613078e-06, 'learning_rate/action_model': 7.985011099613077e-05, 'epoch': 0.62}) +[2026-06-15 23:28:25,837][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.05789843201637268, 'train/grad_norm_pre_clip': 2.398549795150757, 'timing/data': 0.013364480808377266, 'timing/model': 1.5340250488370657, 'learning_rate/qwen_vl_interface': 7.98252061219308e-06, 'learning_rate/action_model': 7.982520612193078e-05, 'epoch': 0.62}) +[2026-06-15 23:28:32,144][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.29799240827560425, 'train/grad_norm_pre_clip': 14.392112731933594, 'timing/data': 0.02271435782313347, 'timing/model': 1.5095267463475466, 'learning_rate/qwen_vl_interface': 7.980029001537073e-06, 'learning_rate/action_model': 7.980029001537072e-05, 'epoch': 0.62}) +[2026-06-15 23:28:38,364][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.3279711604118347, 'train/grad_norm_pre_clip': 14.694576263427734, 'timing/data': 0.015914594754576683, 'timing/model': 1.4975878540426493, 'learning_rate/qwen_vl_interface': 7.977536268669264e-06, 'learning_rate/action_model': 7.977536268669264e-05, 'epoch': 0.62}) +[2026-06-15 23:28:44,378][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.2236516773700714, 'train/grad_norm_pre_clip': 14.353983879089355, 'timing/data': 0.010175017639994621, 'timing/model': 1.441616902127862, 'learning_rate/qwen_vl_interface': 7.975042414614324e-06, 'learning_rate/action_model': 7.975042414614324e-05, 'epoch': 0.62}) +[2026-06-15 23:28:50,290][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.20866189897060394, 'train/grad_norm_pre_clip': 17.48918914794922, 'timing/data': 0.011872043833136559, 'timing/model': 1.6626878641545773, 'learning_rate/qwen_vl_interface': 7.972547440397382e-06, 'learning_rate/action_model': 7.972547440397383e-05, 'epoch': 0.62}) +[2026-06-15 23:28:55,983][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.07906167954206467, 'train/grad_norm_pre_clip': 5.4318671226501465, 'timing/data': 0.02069324254989624, 'timing/model': 1.443136153742671, 'learning_rate/qwen_vl_interface': 7.970051347044026e-06, 'learning_rate/action_model': 7.970051347044025e-05, 'epoch': 0.62}) +[2026-06-15 23:36:28,840][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.17295899987220764, 'train/grad_norm_pre_clip': 10.600360870361328, 'eval/loss': 0.06741061806678772, 'eval/flap_precision': 0.9402985074626866, 'eval/flap_recall': 0.9402985074626866, 'eval/flap_f1': 0.9402985074626865, 'eval/flap_rate_model': 0.20472222222222222, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9402985074626866, 'timing/data': 0.013409018516540527, 'timing/model': 1.425777593627572, 'learning_rate/qwen_vl_interface': 7.967554135580307e-06, 'learning_rate/action_model': 7.967554135580307e-05, 'epoch': 0.62}) +[2026-06-15 23:36:28,841][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1600_state +[2026-06-15 23:36:28,845][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:36:35,012][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1600_state/model.safetensors +[2026-06-15 23:36:49,424][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1600_state/optimizer.bin +[2026-06-15 23:36:49,425][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1600_state/sampler.bin +[2026-06-15 23:36:49,425][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1600_state/sampler_1.bin +[2026-06-15 23:36:49,426][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1600_state/random_states_0.pkl +[2026-06-15 23:36:49,470][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1600 +[2026-06-15 23:36:49,470][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:36:49,675][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:36:55,507][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.052333660423755646, 'train/grad_norm_pre_clip': 6.687155246734619, 'timing/data': 0.011075925081968307, 'timing/model': 1.454727428033948, 'learning_rate/qwen_vl_interface': 7.965055807032734e-06, 'learning_rate/action_model': 7.965055807032735e-05, 'epoch': 0.62}) +[2026-06-15 23:37:01,513][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.04209711775183678, 'train/grad_norm_pre_clip': 3.448502540588379, 'timing/data': 0.010947054252028465, 'timing/model': 1.454391360282898, 'learning_rate/qwen_vl_interface': 7.962556362428275e-06, 'learning_rate/action_model': 7.962556362428274e-05, 'epoch': 0.62}) +[2026-06-15 23:37:07,418][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.18567931652069092, 'train/grad_norm_pre_clip': 15.714367866516113, 'timing/data': 0.011636560782790184, 'timing/model': 1.4579398725181818, 'learning_rate/qwen_vl_interface': 7.960055802794358e-06, 'learning_rate/action_model': 7.960055802794357e-05, 'epoch': 0.62}) +[2026-06-15 23:37:13,287][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.03429201617836952, 'train/grad_norm_pre_clip': 5.780483722686768, 'timing/data': 0.009888667613267899, 'timing/model': 1.4493802227079868, 'learning_rate/qwen_vl_interface': 7.957554129158866e-06, 'learning_rate/action_model': 7.957554129158866e-05, 'epoch': 0.62}) +[2026-06-15 23:37:19,224][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.18367701768875122, 'train/grad_norm_pre_clip': 7.212746620178223, 'timing/data': 0.010194620117545128, 'timing/model': 1.6908697932958603, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 0.62}) +[2026-06-15 23:37:24,891][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.20389339327812195, 'train/grad_norm_pre_clip': 11.054634094238281, 'timing/data': 0.010812301188707352, 'timing/model': 1.4407956674695015, 'learning_rate/qwen_vl_interface': 7.952547443997e-06, 'learning_rate/action_model': 7.952547443997e-05, 'epoch': 0.62}) +[2026-06-15 23:37:30,789][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.22052544355392456, 'train/grad_norm_pre_clip': 18.525535583496094, 'timing/data': 0.010412080213427544, 'timing/model': 1.4365512002259493, 'learning_rate/qwen_vl_interface': 7.950042434528681e-06, 'learning_rate/action_model': 7.950042434528682e-05, 'epoch': 0.62}) +[2026-06-15 23:37:36,595][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.19385235011577606, 'train/grad_norm_pre_clip': 7.176607131958008, 'timing/data': 0.011475495994091034, 'timing/model': 1.4468383081257343, 'learning_rate/qwen_vl_interface': 7.94753631517491e-06, 'learning_rate/action_model': 7.947536315174909e-05, 'epoch': 0.62}) +[2026-06-15 23:37:42,495][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.21593602001667023, 'train/grad_norm_pre_clip': 10.378026008605957, 'timing/data': 0.010182151570916176, 'timing/model': 1.4497562907636166, 'learning_rate/qwen_vl_interface': 7.945029086965854e-06, 'learning_rate/action_model': 7.945029086965853e-05, 'epoch': 0.62}) +[2026-06-15 23:37:48,411][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.2044019103050232, 'train/grad_norm_pre_clip': 3.688636064529419, 'timing/data': 0.009701190516352654, 'timing/model': 1.4347241688519716, 'learning_rate/qwen_vl_interface': 7.942520750932138e-06, 'learning_rate/action_model': 7.942520750932138e-05, 'epoch': 0.62}) +[2026-06-15 23:37:54,298][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.24115082621574402, 'train/grad_norm_pre_clip': 7.23398494720459, 'timing/data': 0.010791026055812836, 'timing/model': 1.6428249832242727, 'learning_rate/qwen_vl_interface': 7.940011308104848e-06, 'learning_rate/action_model': 7.940011308104848e-05, 'epoch': 0.62}) +[2026-06-15 23:38:00,216][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.02014051377773285, 'train/grad_norm_pre_clip': 1.5525354146957397, 'timing/data': 0.010392269119620323, 'timing/model': 1.6662290152162313, 'learning_rate/qwen_vl_interface': 7.93750075951552e-06, 'learning_rate/action_model': 7.93750075951552e-05, 'epoch': 0.62}) +[2026-06-15 23:38:06,467][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.09460911154747009, 'train/grad_norm_pre_clip': 9.104864120483398, 'timing/data': 0.037186138331890106, 'timing/model': 1.5309301763772964, 'learning_rate/qwen_vl_interface': 7.934989106196145e-06, 'learning_rate/action_model': 7.934989106196146e-05, 'epoch': 0.62}) +[2026-06-15 23:38:12,847][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.1133878156542778, 'train/grad_norm_pre_clip': 15.213637351989746, 'timing/data': 0.012298496440052986, 'timing/model': 1.559046681970358, 'learning_rate/qwen_vl_interface': 7.93247634917917e-06, 'learning_rate/action_model': 7.93247634917917e-05, 'epoch': 0.62}) +[2026-06-15 23:38:19,055][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.1296275109052658, 'train/grad_norm_pre_clip': 7.688848972320557, 'timing/data': 0.012216892093420029, 'timing/model': 1.5596789103001356, 'learning_rate/qwen_vl_interface': 7.929962489497494e-06, 'learning_rate/action_model': 7.929962489497493e-05, 'epoch': 0.63}) +[2026-06-15 23:38:25,251][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.19121702015399933, 'train/grad_norm_pre_clip': 10.354455947875977, 'timing/data': 0.012324212118983269, 'timing/model': 1.4995374269783497, 'learning_rate/qwen_vl_interface': 7.927447528184469e-06, 'learning_rate/action_model': 7.927447528184468e-05, 'epoch': 0.63}) +[2026-06-15 23:38:31,572][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.4655734598636627, 'train/grad_norm_pre_clip': 23.066972732543945, 'timing/data': 0.01226186752319336, 'timing/model': 1.7841174975037575, 'learning_rate/qwen_vl_interface': 7.924931466273901e-06, 'learning_rate/action_model': 7.924931466273902e-05, 'epoch': 0.63}) +[2026-06-15 23:38:37,652][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.1524251401424408, 'train/grad_norm_pre_clip': 10.64207649230957, 'timing/data': 0.011664127930998802, 'timing/model': 1.5615816041827202, 'learning_rate/qwen_vl_interface': 7.92241430480005e-06, 'learning_rate/action_model': 7.922414304800049e-05, 'epoch': 0.63}) +[2026-06-15 23:38:43,951][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.1639045923948288, 'train/grad_norm_pre_clip': 10.35934829711914, 'timing/data': 0.013030484318733215, 'timing/model': 1.5526382271200418, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.63}) +[2026-06-15 23:38:50,256][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.03924601897597313, 'train/grad_norm_pre_clip': 2.5918760299682617, 'timing/data': 0.01265474408864975, 'timing/model': 1.5575529877096415, 'learning_rate/qwen_vl_interface': 7.91737668730178e-06, 'learning_rate/action_model': 7.91737668730178e-05, 'epoch': 0.63}) +[2026-06-15 23:38:56,565][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.15813696384429932, 'train/grad_norm_pre_clip': 9.393872261047363, 'timing/data': 0.011581694707274437, 'timing/model': 1.5808279439806938, 'learning_rate/qwen_vl_interface': 7.91485623334814e-06, 'learning_rate/action_model': 7.91485623334814e-05, 'epoch': 0.63}) +[2026-06-15 23:39:02,852][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.06442814320325851, 'train/grad_norm_pre_clip': 4.201730251312256, 'timing/data': 0.01371869258582592, 'timing/model': 1.5094414111226797, 'learning_rate/qwen_vl_interface': 7.91233468397276e-06, 'learning_rate/action_model': 7.912334683972761e-05, 'epoch': 0.63}) +[2026-06-15 23:39:09,157][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.20659705996513367, 'train/grad_norm_pre_clip': 11.690510749816895, 'timing/data': 0.011917490512132645, 'timing/model': 1.7294689826667309, 'learning_rate/qwen_vl_interface': 7.90981204021216e-06, 'learning_rate/action_model': 7.90981204021216e-05, 'epoch': 0.63}) +[2026-06-15 23:39:15,151][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.22839316725730896, 'train/grad_norm_pre_clip': 14.416144371032715, 'timing/data': 0.012781569734215736, 'timing/model': 1.541031552478671, 'learning_rate/qwen_vl_interface': 7.907288303103298e-06, 'learning_rate/action_model': 7.907288303103298e-05, 'epoch': 0.63}) +[2026-06-15 23:39:21,452][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.21758930385112762, 'train/grad_norm_pre_clip': 9.771539688110352, 'timing/data': 0.01231776736676693, 'timing/model': 1.5334979593753815, 'learning_rate/qwen_vl_interface': 7.90476347368359e-06, 'learning_rate/action_model': 7.90476347368359e-05, 'epoch': 0.63}) +[2026-06-15 23:39:27,650][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.1059139147400856, 'train/grad_norm_pre_clip': 3.7979605197906494, 'timing/data': 0.012716459110379219, 'timing/model': 1.5409078001976013, 'learning_rate/qwen_vl_interface': 7.902237552990897e-06, 'learning_rate/action_model': 7.902237552990897e-05, 'epoch': 0.63}) +[2026-06-15 23:39:33,894][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.12800180912017822, 'train/grad_norm_pre_clip': 6.42877721786499, 'timing/data': 0.010889101773500443, 'timing/model': 1.49439219199121, 'learning_rate/qwen_vl_interface': 7.89971054206353e-06, 'learning_rate/action_model': 7.89971054206353e-05, 'epoch': 0.63}) +[2026-06-15 23:39:39,788][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.21908780932426453, 'train/grad_norm_pre_clip': 9.350004196166992, 'timing/data': 0.010673874989151955, 'timing/model': 1.5049665663391352, 'learning_rate/qwen_vl_interface': 7.897182441940247e-06, 'learning_rate/action_model': 7.897182441940247e-05, 'epoch': 0.63}) +[2026-06-15 23:39:45,688][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.25895893573760986, 'train/grad_norm_pre_clip': 8.585227966308594, 'timing/data': 0.010842520743608475, 'timing/model': 1.6469336450099945, 'learning_rate/qwen_vl_interface': 7.894653253660256e-06, 'learning_rate/action_model': 7.894653253660256e-05, 'epoch': 0.63}) +[2026-06-15 23:39:51,479][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.09495479613542557, 'train/grad_norm_pre_clip': 11.788640975952148, 'timing/data': 0.010778328403830528, 'timing/model': 1.5412156581878662, 'learning_rate/qwen_vl_interface': 7.89212297826321e-06, 'learning_rate/action_model': 7.892122978263209e-05, 'epoch': 0.63}) +[2026-06-15 23:39:57,381][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.14438849687576294, 'train/grad_norm_pre_clip': 6.940544128417969, 'timing/data': 0.011534415185451508, 'timing/model': 1.4527976904064417, 'learning_rate/qwen_vl_interface': 7.889591616789206e-06, 'learning_rate/action_model': 7.889591616789206e-05, 'epoch': 0.63}) +[2026-06-15 23:40:03,280][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.07403691112995148, 'train/grad_norm_pre_clip': 7.942310810089111, 'timing/data': 0.011008374392986298, 'timing/model': 1.4184080604463816, 'learning_rate/qwen_vl_interface': 7.887059170278799e-06, 'learning_rate/action_model': 7.887059170278797e-05, 'epoch': 0.63}) +[2026-06-15 23:40:09,180][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.05292578041553497, 'train/grad_norm_pre_clip': 3.237708330154419, 'timing/data': 0.010414443910121918, 'timing/model': 1.4172189235687256, 'learning_rate/qwen_vl_interface': 7.884525639772974e-06, 'learning_rate/action_model': 7.884525639772974e-05, 'epoch': 0.63}) +[2026-06-15 23:40:15,081][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.1409398913383484, 'train/grad_norm_pre_clip': 10.28232192993164, 'timing/data': 0.01096564345061779, 'timing/model': 1.4761846959590912, 'learning_rate/qwen_vl_interface': 7.881991026313176e-06, 'learning_rate/action_model': 7.881991026313175e-05, 'epoch': 0.63}) +[2026-06-15 23:40:21,368][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.22560691833496094, 'train/grad_norm_pre_clip': 15.580904006958008, 'timing/data': 0.012543361634016037, 'timing/model': 1.790996514260769, 'learning_rate/qwen_vl_interface': 7.879455330941283e-06, 'learning_rate/action_model': 7.879455330941282e-05, 'epoch': 0.63}) +[2026-06-15 23:40:27,544][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.2587784230709076, 'train/grad_norm_pre_clip': 19.164905548095703, 'timing/data': 0.012108659371733665, 'timing/model': 1.5516352858394384, 'learning_rate/qwen_vl_interface': 7.876918554699627e-06, 'learning_rate/action_model': 7.876918554699627e-05, 'epoch': 0.63}) +[2026-06-15 23:40:33,718][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.1640930026769638, 'train/grad_norm_pre_clip': 9.108044624328613, 'timing/data': 0.012645300477743149, 'timing/model': 1.5482582654803991, 'learning_rate/qwen_vl_interface': 7.874380698630982e-06, 'learning_rate/action_model': 7.874380698630982e-05, 'epoch': 0.63}) +[2026-06-15 23:40:39,939][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.10715701431035995, 'train/grad_norm_pre_clip': 6.213925361633301, 'timing/data': 0.011664077639579773, 'timing/model': 1.5523989778012037, 'learning_rate/qwen_vl_interface': 7.871841763778562e-06, 'learning_rate/action_model': 7.871841763778563e-05, 'epoch': 0.63}) +[2026-06-15 23:40:46,144][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.018292345106601715, 'train/grad_norm_pre_clip': 1.9095590114593506, 'timing/data': 0.011908797547221184, 'timing/model': 1.5189014989882708, 'learning_rate/qwen_vl_interface': 7.86930175118603e-06, 'learning_rate/action_model': 7.86930175118603e-05, 'epoch': 0.63}) +[2026-06-15 23:40:52,343][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.2562711536884308, 'train/grad_norm_pre_clip': 18.686613082885742, 'timing/data': 0.012111062183976173, 'timing/model': 1.536228857934475, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.63}) +[2026-06-15 23:40:58,540][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.27303317189216614, 'train/grad_norm_pre_clip': 15.557513236999512, 'timing/data': 0.011596975848078728, 'timing/model': 1.7530722301453352, 'learning_rate/qwen_vl_interface': 7.86421849695748e-06, 'learning_rate/action_model': 7.864218496957481e-05, 'epoch': 0.64}) +[2026-06-15 23:41:04,543][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.09805913269519806, 'train/grad_norm_pre_clip': 6.401612758636475, 'timing/data': 0.012987315654754639, 'timing/model': 1.554546369239688, 'learning_rate/qwen_vl_interface': 7.861675257410998e-06, 'learning_rate/action_model': 7.861675257410997e-05, 'epoch': 0.64}) +[2026-06-15 23:41:10,744][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.13743583858013153, 'train/grad_norm_pre_clip': 10.738895416259766, 'timing/data': 0.012184364721179008, 'timing/model': 1.5600633341819048, 'learning_rate/qwen_vl_interface': 7.859130944303468e-06, 'learning_rate/action_model': 7.859130944303467e-05, 'epoch': 0.64}) +[2026-06-15 23:41:16,942][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.1289873868227005, 'train/grad_norm_pre_clip': 11.765069961547852, 'timing/data': 0.011850612238049507, 'timing/model': 1.5568582098931074, 'learning_rate/qwen_vl_interface': 7.856585558680761e-06, 'learning_rate/action_model': 7.856585558680761e-05, 'epoch': 0.64}) +[2026-06-15 23:41:22,789][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.19051873683929443, 'train/grad_norm_pre_clip': 11.071504592895508, 'timing/data': 0.010240606963634491, 'timing/model': 1.460563438013196, 'learning_rate/qwen_vl_interface': 7.854039101589191e-06, 'learning_rate/action_model': 7.854039101589191e-05, 'epoch': 0.64}) +[2026-06-15 23:41:28,688][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.051474086940288544, 'train/grad_norm_pre_clip': 4.045446395874023, 'timing/data': 0.010662015527486801, 'timing/model': 1.4468912240117788, 'learning_rate/qwen_vl_interface': 7.851491574075509e-06, 'learning_rate/action_model': 7.851491574075509e-05, 'epoch': 0.64}) +[2026-06-15 23:41:34,556][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.09460586309432983, 'train/grad_norm_pre_clip': 5.610445976257324, 'timing/data': 0.010718969628214836, 'timing/model': 1.6218265611678362, 'learning_rate/qwen_vl_interface': 7.848942977186909e-06, 'learning_rate/action_model': 7.848942977186908e-05, 'epoch': 0.64}) +[2026-06-15 23:41:40,182][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.05580463632941246, 'train/grad_norm_pre_clip': 4.484016418457031, 'timing/data': 0.010570492595434189, 'timing/model': 1.4508107155561447, 'learning_rate/qwen_vl_interface': 7.84639331197102e-06, 'learning_rate/action_model': 7.84639331197102e-05, 'epoch': 0.64}) +[2026-06-15 23:41:46,081][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.15980012714862823, 'train/grad_norm_pre_clip': 12.462797164916992, 'timing/data': 0.009621907025575638, 'timing/model': 1.4476106278598309, 'learning_rate/qwen_vl_interface': 7.843842579475918e-06, 'learning_rate/action_model': 7.843842579475918e-05, 'epoch': 0.64}) +[2026-06-15 23:41:51,882][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.2453964501619339, 'train/grad_norm_pre_clip': 9.296730995178223, 'timing/data': 0.010166114196181297, 'timing/model': 1.4529066681861877, 'learning_rate/qwen_vl_interface': 7.841290780750107e-06, 'learning_rate/action_model': 7.841290780750107e-05, 'epoch': 0.64}) +[2026-06-15 23:41:57,684][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.133988618850708, 'train/grad_norm_pre_clip': 13.455358505249023, 'timing/data': 0.010052034631371498, 'timing/model': 1.446910798549652, 'learning_rate/qwen_vl_interface': 7.838737916842538e-06, 'learning_rate/action_model': 7.838737916842538e-05, 'epoch': 0.64}) +[2026-06-15 23:42:03,582][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.14698919653892517, 'train/grad_norm_pre_clip': 6.329489231109619, 'timing/data': 0.0116419717669487, 'timing/model': 1.4352078922092915, 'learning_rate/qwen_vl_interface': 7.836183988802596e-06, 'learning_rate/action_model': 7.836183988802595e-05, 'epoch': 0.64}) +[2026-06-15 23:42:09,478][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.11412879824638367, 'train/grad_norm_pre_clip': 8.400785446166992, 'timing/data': 0.010274725034832954, 'timing/model': 1.6359011139720678, 'learning_rate/qwen_vl_interface': 7.833628997680107e-06, 'learning_rate/action_model': 7.833628997680107e-05, 'epoch': 0.64}) +[2026-06-15 23:42:15,181][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.317577987909317, 'train/grad_norm_pre_clip': 4.440154075622559, 'timing/data': 0.011103024706244469, 'timing/model': 1.440939787775278, 'learning_rate/qwen_vl_interface': 7.831072944525329e-06, 'learning_rate/action_model': 7.831072944525328e-05, 'epoch': 0.64}) +[2026-06-15 23:42:21,080][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.08309578895568848, 'train/grad_norm_pre_clip': 5.956480979919434, 'timing/data': 0.009929532185196877, 'timing/model': 1.4228634592145681, 'learning_rate/qwen_vl_interface': 7.828515830388956e-06, 'learning_rate/action_model': 7.828515830388956e-05, 'epoch': 0.64}) +[2026-06-15 23:42:26,980][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.025949832051992416, 'train/grad_norm_pre_clip': 2.81512188911438, 'timing/data': 0.0107806995511055, 'timing/model': 1.4413197003304958, 'learning_rate/qwen_vl_interface': 7.825957656322129e-06, 'learning_rate/action_model': 7.825957656322128e-05, 'epoch': 0.64}) +[2026-06-15 23:42:32,881][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.24247775971889496, 'train/grad_norm_pre_clip': 13.523239135742188, 'timing/data': 0.01036033220589161, 'timing/model': 1.4388998467475176, 'learning_rate/qwen_vl_interface': 7.823398423376406e-06, 'learning_rate/action_model': 7.823398423376406e-05, 'epoch': 0.64}) +[2026-06-15 23:42:38,803][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.1465427130460739, 'train/grad_norm_pre_clip': 9.07890796661377, 'timing/data': 0.010545141994953156, 'timing/model': 1.4261401295661926, 'learning_rate/qwen_vl_interface': 7.820838132603801e-06, 'learning_rate/action_model': 7.8208381326038e-05, 'epoch': 0.64}) +[2026-06-15 23:42:44,697][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.026154395192861557, 'train/grad_norm_pre_clip': 2.973306179046631, 'timing/data': 0.010889900848269463, 'timing/model': 1.6602977141737938, 'learning_rate/qwen_vl_interface': 7.818276785056748e-06, 'learning_rate/action_model': 7.818276785056749e-05, 'epoch': 0.64}) +[2026-06-15 23:42:50,754][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.0625641867518425, 'train/grad_norm_pre_clip': 6.249004364013672, 'timing/data': 0.013076836243271828, 'timing/model': 1.5455130394548178, 'learning_rate/qwen_vl_interface': 7.815714381788123e-06, 'learning_rate/action_model': 7.815714381788122e-05, 'epoch': 0.64}) +[2026-06-15 23:42:57,066][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.08736944943666458, 'train/grad_norm_pre_clip': 5.980703353881836, 'timing/data': 0.014741197228431702, 'timing/model': 1.5632748510688543, 'learning_rate/qwen_vl_interface': 7.813150923851227e-06, 'learning_rate/action_model': 7.813150923851227e-05, 'epoch': 0.64}) +[2026-06-15 23:43:03,368][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.16487017273902893, 'train/grad_norm_pre_clip': 11.696735382080078, 'timing/data': 0.014328591525554657, 'timing/model': 1.563181972131133, 'learning_rate/qwen_vl_interface': 7.810586412299807e-06, 'learning_rate/action_model': 7.810586412299806e-05, 'epoch': 0.64}) +[2026-06-15 23:43:09,652][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.033750537782907486, 'train/grad_norm_pre_clip': 5.351086139678955, 'timing/data': 0.012978361919522285, 'timing/model': 1.5546514727175236, 'learning_rate/qwen_vl_interface': 7.808020848188034e-06, 'learning_rate/action_model': 7.808020848188034e-05, 'epoch': 0.64}) +[2026-06-15 23:43:15,962][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.30101293325424194, 'train/grad_norm_pre_clip': 23.676713943481445, 'timing/data': 0.012043939903378487, 'timing/model': 1.582528006285429, 'learning_rate/qwen_vl_interface': 7.805454232570517e-06, 'learning_rate/action_model': 7.805454232570517e-05, 'epoch': 0.64}) +[2026-06-15 23:43:22,305][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.21286985278129578, 'train/grad_norm_pre_clip': 15.955605506896973, 'timing/data': 0.012327587231993675, 'timing/model': 1.7716176006942987, 'learning_rate/qwen_vl_interface': 7.80288656650229e-06, 'learning_rate/action_model': 7.80288656650229e-05, 'epoch': 0.64}) +[2026-06-15 23:43:28,555][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.15211136639118195, 'train/grad_norm_pre_clip': 18.328298568725586, 'timing/data': 0.013417385518550873, 'timing/model': 1.7397655863314867, 'learning_rate/qwen_vl_interface': 7.800317851038829e-06, 'learning_rate/action_model': 7.800317851038828e-05, 'epoch': 0.64}) +[2026-06-15 23:43:34,674][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.0873379185795784, 'train/grad_norm_pre_clip': 11.947227478027344, 'timing/data': 0.012876495718955994, 'timing/model': 1.549792455509305, 'learning_rate/qwen_vl_interface': 7.797748087236033e-06, 'learning_rate/action_model': 7.797748087236033e-05, 'epoch': 0.65}) +[2026-06-15 23:43:41,050][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.2650130093097687, 'train/grad_norm_pre_clip': 8.191876411437988, 'timing/data': 0.013000179082155228, 'timing/model': 1.5319695323705673, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.65}) +[2026-06-15 23:43:47,262][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.3839961886405945, 'train/grad_norm_pre_clip': 34.25952911376953, 'timing/data': 0.013057820498943329, 'timing/model': 1.5324786845594645, 'learning_rate/qwen_vl_interface': 7.792605418838196e-06, 'learning_rate/action_model': 7.792605418838197e-05, 'epoch': 0.65}) +[2026-06-15 23:43:53,651][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.12530362606048584, 'train/grad_norm_pre_clip': 8.882855415344238, 'timing/data': 0.012745065614581108, 'timing/model': 1.5664214212447405, 'learning_rate/qwen_vl_interface': 7.790032516357116e-06, 'learning_rate/action_model': 7.790032516357117e-05, 'epoch': 0.65}) +[2026-06-15 23:43:59,890][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.3385674059391022, 'train/grad_norm_pre_clip': 8.631729125976562, 'timing/data': 0.013114474713802338, 'timing/model': 1.789978090673685, 'learning_rate/qwen_vl_interface': 7.787458569764613e-06, 'learning_rate/action_model': 7.787458569764612e-05, 'epoch': 0.65}) +[2026-06-15 23:44:05,952][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.06274476647377014, 'train/grad_norm_pre_clip': 7.928426742553711, 'timing/data': 0.013046035543084145, 'timing/model': 1.5564306657761335, 'learning_rate/qwen_vl_interface': 7.784883580118738e-06, 'learning_rate/action_model': 7.784883580118738e-05, 'epoch': 0.65}) +[2026-06-15 23:44:12,356][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.07691133767366409, 'train/grad_norm_pre_clip': 9.526179313659668, 'timing/data': 0.013421053066849709, 'timing/model': 1.5525910947471857, 'learning_rate/qwen_vl_interface': 7.782307548477975e-06, 'learning_rate/action_model': 7.782307548477975e-05, 'epoch': 0.65}) +[2026-06-15 23:44:18,651][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.06743936240673065, 'train/grad_norm_pre_clip': 9.287169456481934, 'timing/data': 0.012295596301555634, 'timing/model': 1.5545902084559202, 'learning_rate/qwen_vl_interface': 7.779730475901233e-06, 'learning_rate/action_model': 7.779730475901232e-05, 'epoch': 0.65}) +[2026-06-15 23:44:24,854][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.23124656081199646, 'train/grad_norm_pre_clip': 11.57943058013916, 'timing/data': 0.013802602887153625, 'timing/model': 1.516205009073019, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 0.65}) +[2026-06-15 23:44:31,157][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.1285119354724884, 'train/grad_norm_pre_clip': 5.920028209686279, 'timing/data': 0.014558495953679085, 'timing/model': 1.573596928268671, 'learning_rate/qwen_vl_interface': 7.774573212177587e-06, 'learning_rate/action_model': 7.774573212177588e-05, 'epoch': 0.65}) +[2026-06-15 23:44:37,350][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.07587900012731552, 'train/grad_norm_pre_clip': 3.181809186935425, 'timing/data': 0.012270171195268631, 'timing/model': 1.7497607357800007, 'learning_rate/qwen_vl_interface': 7.77199302315064e-06, 'learning_rate/action_model': 7.77199302315064e-05, 'epoch': 0.65}) +[2026-06-15 23:44:43,353][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.09560427069664001, 'train/grad_norm_pre_clip': 4.102834224700928, 'timing/data': 0.012817693874239922, 'timing/model': 1.5540291853249073, 'learning_rate/qwen_vl_interface': 7.769411797427622e-06, 'learning_rate/action_model': 7.769411797427623e-05, 'epoch': 0.65}) +[2026-06-15 23:44:49,651][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.11834923177957535, 'train/grad_norm_pre_clip': 5.61898136138916, 'timing/data': 0.013936016708612442, 'timing/model': 1.5523139908909798, 'learning_rate/qwen_vl_interface': 7.766829536069584e-06, 'learning_rate/action_model': 7.766829536069584e-05, 'epoch': 0.65}) +[2026-06-15 23:44:55,594][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.09281578660011292, 'train/grad_norm_pre_clip': 12.63339614868164, 'timing/data': 0.010417722165584564, 'timing/model': 1.4530124012380838, 'learning_rate/qwen_vl_interface': 7.764246240137991e-06, 'learning_rate/action_model': 7.764246240137991e-05, 'epoch': 0.65}) +[2026-06-15 23:45:01,494][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.30894920229911804, 'train/grad_norm_pre_clip': 17.987340927124023, 'timing/data': 0.009865252301096916, 'timing/model': 1.447484351694584, 'learning_rate/qwen_vl_interface': 7.761661910694741e-06, 'learning_rate/action_model': 7.761661910694741e-05, 'epoch': 0.65}) +[2026-06-15 23:45:07,393][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.052771929651498795, 'train/grad_norm_pre_clip': 3.3591487407684326, 'timing/data': 0.009881705045700073, 'timing/model': 1.4583394508808851, 'learning_rate/qwen_vl_interface': 7.759076548802153e-06, 'learning_rate/action_model': 7.759076548802152e-05, 'epoch': 0.65}) +[2026-06-15 23:45:13,252][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.118026502430439, 'train/grad_norm_pre_clip': 4.979308605194092, 'timing/data': 0.010400466620922089, 'timing/model': 1.6071748286485672, 'learning_rate/qwen_vl_interface': 7.75649015552297e-06, 'learning_rate/action_model': 7.75649015552297e-05, 'epoch': 0.65}) +[2026-06-15 23:45:18,885][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.12327876687049866, 'train/grad_norm_pre_clip': 14.376423835754395, 'timing/data': 0.011441607028245926, 'timing/model': 1.4526662696152925, 'learning_rate/qwen_vl_interface': 7.753902731920364e-06, 'learning_rate/action_model': 7.753902731920364e-05, 'epoch': 0.65}) +[2026-06-15 23:45:24,782][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.09150971472263336, 'train/grad_norm_pre_clip': 19.189083099365234, 'timing/data': 0.010958492755889893, 'timing/model': 1.443066406995058, 'learning_rate/qwen_vl_interface': 7.751314279057925e-06, 'learning_rate/action_model': 7.751314279057923e-05, 'epoch': 0.65}) +[2026-06-15 23:45:30,695][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.20406904816627502, 'train/grad_norm_pre_clip': 6.999400615692139, 'timing/data': 0.010433048009872437, 'timing/model': 1.46278715133667, 'learning_rate/qwen_vl_interface': 7.74872479799967e-06, 'learning_rate/action_model': 7.74872479799967e-05, 'epoch': 0.65}) +[2026-06-15 23:45:36,586][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.03039933741092682, 'train/grad_norm_pre_clip': 1.6524817943572998, 'timing/data': 0.010343315079808235, 'timing/model': 1.4320220313966274, 'learning_rate/qwen_vl_interface': 7.746134289810034e-06, 'learning_rate/action_model': 7.746134289810033e-05, 'epoch': 0.65}) +[2026-06-15 23:45:42,499][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.08721759915351868, 'train/grad_norm_pre_clip': 5.626883506774902, 'timing/data': 0.011442525312304497, 'timing/model': 1.6607144959270954, 'learning_rate/qwen_vl_interface': 7.74354275555388e-06, 'learning_rate/action_model': 7.74354275555388e-05, 'epoch': 0.65}) +[2026-06-15 23:45:48,190][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.07055100053548813, 'train/grad_norm_pre_clip': 3.9340968132019043, 'timing/data': 0.010713836178183556, 'timing/model': 1.4551893770694733, 'learning_rate/qwen_vl_interface': 7.740950196296488e-06, 'learning_rate/action_model': 7.740950196296488e-05, 'epoch': 0.65}) +[2026-06-15 23:45:54,584][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.04449721425771713, 'train/grad_norm_pre_clip': 3.876657247543335, 'timing/data': 0.012183114886283875, 'timing/model': 1.5864416658878326, 'learning_rate/qwen_vl_interface': 7.738356613103562e-06, 'learning_rate/action_model': 7.73835661310356e-05, 'epoch': 0.65}) +[2026-06-15 23:46:01,046][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.1647833287715912, 'train/grad_norm_pre_clip': 10.460776329040527, 'timing/data': 0.014979606494307518, 'timing/model': 1.5997677072882652, 'learning_rate/qwen_vl_interface': 7.735762007041226e-06, 'learning_rate/action_model': 7.735762007041226e-05, 'epoch': 0.65}) +[2026-06-15 23:46:07,377][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.06579963862895966, 'train/grad_norm_pre_clip': 5.230018138885498, 'timing/data': 0.012517815455794334, 'timing/model': 1.5338050797581673, 'learning_rate/qwen_vl_interface': 7.733166379176023e-06, 'learning_rate/action_model': 7.733166379176023e-05, 'epoch': 0.65}) +[2026-06-15 23:46:13,687][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.3146698474884033, 'train/grad_norm_pre_clip': 20.372722625732422, 'timing/data': 0.01238175481557846, 'timing/model': 1.5361729338765144, 'learning_rate/qwen_vl_interface': 7.730569730574921e-06, 'learning_rate/action_model': 7.73056973057492e-05, 'epoch': 0.66}) +[2026-06-15 23:46:20,032][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.25400352478027344, 'train/grad_norm_pre_clip': 19.388015747070312, 'timing/data': 0.01301456056535244, 'timing/model': 1.7675351984798908, 'learning_rate/qwen_vl_interface': 7.727972062305299e-06, 'learning_rate/action_model': 7.727972062305299e-05, 'epoch': 0.66}) +[2026-06-15 23:46:26,077][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.534135103225708, 'train/grad_norm_pre_clip': 22.206287384033203, 'timing/data': 0.011272571980953217, 'timing/model': 1.5259179789572954, 'learning_rate/qwen_vl_interface': 7.725373375434965e-06, 'learning_rate/action_model': 7.725373375434965e-05, 'epoch': 0.66}) +[2026-06-15 23:46:32,357][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.04464636743068695, 'train/grad_norm_pre_clip': 4.662339687347412, 'timing/data': 0.012160604819655418, 'timing/model': 1.5410782173275948, 'learning_rate/qwen_vl_interface': 7.722773671032141e-06, 'learning_rate/action_model': 7.722773671032142e-05, 'epoch': 0.66}) +[2026-06-15 23:46:38,544][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.11484014987945557, 'train/grad_norm_pre_clip': 13.326234817504883, 'timing/data': 0.013037523254752159, 'timing/model': 1.532006548717618, 'learning_rate/qwen_vl_interface': 7.720172950165466e-06, 'learning_rate/action_model': 7.720172950165465e-05, 'epoch': 0.66}) +[2026-06-15 23:46:44,644][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.08050414174795151, 'train/grad_norm_pre_clip': 11.533473014831543, 'timing/data': 0.013398651033639908, 'timing/model': 1.5258134957402945, 'learning_rate/qwen_vl_interface': 7.717571213903996e-06, 'learning_rate/action_model': 7.717571213903996e-05, 'epoch': 0.66}) +[2026-06-15 23:46:50,850][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.0519043430685997, 'train/grad_norm_pre_clip': 9.759145736694336, 'timing/data': 0.013765700161457062, 'timing/model': 1.5684127639979124, 'learning_rate/qwen_vl_interface': 7.714968463317207e-06, 'learning_rate/action_model': 7.714968463317207e-05, 'epoch': 0.66}) +[2026-06-15 23:54:02,746][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.5483821630477905, 'train/grad_norm_pre_clip': 19.511676788330078, 'eval/loss': 0.09499765932559967, 'eval/flap_precision': 0.8929032258064517, 'eval/flap_recall': 0.9389416553595658, 'eval/flap_f1': 0.9153439153439153, 'eval/flap_rate_model': 0.2152777777777778, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9153439153439153, 'timing/data': 0.012661263346672058, 'timing/model': 1.8010014444589615, 'learning_rate/qwen_vl_interface': 7.712364699474996e-06, 'learning_rate/action_model': 7.712364699474996e-05, 'epoch': 0.66}) +[2026-06-15 23:54:02,747][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1700_state +[2026-06-15 23:54:02,750][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:54:08,912][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1700_state/model.safetensors +[2026-06-15 23:54:23,414][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1700_state/optimizer.bin +[2026-06-15 23:54:23,414][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1700_state/sampler.bin +[2026-06-15 23:54:23,415][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1700_state/sampler_1.bin +[2026-06-15 23:54:23,416][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1700_state/random_states_0.pkl +[2026-06-15 23:54:23,448][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1700 +[2026-06-15 23:54:23,448][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:54:23,456][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:54:29,347][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.007223934400826693, 'train/grad_norm_pre_clip': 0.7875608801841736, 'timing/data': 0.01033153384923935, 'timing/model': 1.6790467910468578, 'learning_rate/qwen_vl_interface': 7.709759923447667e-06, 'learning_rate/action_model': 7.709759923447665e-05, 'epoch': 0.66}) +[2026-06-15 23:54:35,222][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.3881702721118927, 'train/grad_norm_pre_clip': 10.499147415161133, 'timing/data': 0.011303747072815895, 'timing/model': 1.6571192126721144, 'learning_rate/qwen_vl_interface': 7.707154136305945e-06, 'learning_rate/action_model': 7.707154136305945e-05, 'epoch': 0.66}) +[2026-06-15 23:54:40,906][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.1619393229484558, 'train/grad_norm_pre_clip': 10.944377899169922, 'timing/data': 0.010374784469604492, 'timing/model': 1.448220606893301, 'learning_rate/qwen_vl_interface': 7.704547339120974e-06, 'learning_rate/action_model': 7.704547339120974e-05, 'epoch': 0.66}) +[2026-06-15 23:54:46,881][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.12731444835662842, 'train/grad_norm_pre_clip': 12.909282684326172, 'timing/data': 0.010082779452204704, 'timing/model': 1.4578715115785599, 'learning_rate/qwen_vl_interface': 7.70193953296431e-06, 'learning_rate/action_model': 7.70193953296431e-05, 'epoch': 0.66}) +[2026-06-15 23:54:52,784][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.0234069786965847, 'train/grad_norm_pre_clip': 1.8528831005096436, 'timing/data': 0.010187847539782524, 'timing/model': 1.4498815406113863, 'learning_rate/qwen_vl_interface': 7.69933071890792e-06, 'learning_rate/action_model': 7.699330718907919e-05, 'epoch': 0.66}) +[2026-06-15 23:54:58,587][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.21627624332904816, 'train/grad_norm_pre_clip': 13.097332000732422, 'timing/data': 0.016199497506022453, 'timing/model': 1.4483335968106985, 'learning_rate/qwen_vl_interface': 7.696720898024188e-06, 'learning_rate/action_model': 7.69672089802419e-05, 'epoch': 0.66}) +[2026-06-15 23:55:04,488][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.003602050244808197, 'train/grad_norm_pre_clip': 0.7748599648475647, 'timing/data': 0.01013302244246006, 'timing/model': 1.4588614832609892, 'learning_rate/qwen_vl_interface': 7.69411007138592e-06, 'learning_rate/action_model': 7.694110071385919e-05, 'epoch': 0.66}) +[2026-06-15 23:55:10,390][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.07829379290342331, 'train/grad_norm_pre_clip': 5.6627631187438965, 'timing/data': 0.011097924783825874, 'timing/model': 1.6578576918691397, 'learning_rate/qwen_vl_interface': 7.691498240066322e-06, 'learning_rate/action_model': 7.691498240066322e-05, 'epoch': 0.66}) +[2026-06-15 23:55:16,089][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.20373426377773285, 'train/grad_norm_pre_clip': 10.909443855285645, 'timing/data': 0.01513710804283619, 'timing/model': 1.4488250352442265, 'learning_rate/qwen_vl_interface': 7.688885405139022e-06, 'learning_rate/action_model': 7.688885405139021e-05, 'epoch': 0.66}) +[2026-06-15 23:55:21,990][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.1216001808643341, 'train/grad_norm_pre_clip': 9.425844192504883, 'timing/data': 0.0137651227414608, 'timing/model': 1.4538891911506653, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.66}) +[2026-06-15 23:55:27,889][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.2903911769390106, 'train/grad_norm_pre_clip': 14.765549659729004, 'timing/data': 0.010050093755126, 'timing/model': 1.457525871694088, 'learning_rate/qwen_vl_interface': 7.683656728757873e-06, 'learning_rate/action_model': 7.683656728757872e-05, 'epoch': 0.66}) +[2026-06-15 23:55:33,788][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.26528000831604004, 'train/grad_norm_pre_clip': 24.971969604492188, 'timing/data': 0.010292138904333115, 'timing/model': 1.472641421481967, 'learning_rate/qwen_vl_interface': 7.681040889453337e-06, 'learning_rate/action_model': 7.681040889453336e-05, 'epoch': 0.66}) +[2026-06-15 23:55:39,680][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.3013205826282501, 'train/grad_norm_pre_clip': 15.533636093139648, 'timing/data': 0.010012546554207802, 'timing/model': 1.4407393280416727, 'learning_rate/qwen_vl_interface': 7.678424050839722e-06, 'learning_rate/action_model': 7.67842405083972e-05, 'epoch': 0.66}) +[2026-06-15 23:55:45,588][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.1555607169866562, 'train/grad_norm_pre_clip': 5.574622631072998, 'timing/data': 0.010620653629302979, 'timing/model': 1.6618840470910072, 'learning_rate/qwen_vl_interface': 7.675806213992706e-06, 'learning_rate/action_model': 7.675806213992707e-05, 'epoch': 0.66}) +[2026-06-15 23:55:51,436][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.09337050467729568, 'train/grad_norm_pre_clip': 3.649144172668457, 'timing/data': 0.012965351343154907, 'timing/model': 1.5533129386603832, 'learning_rate/qwen_vl_interface': 7.673187379988388e-06, 'learning_rate/action_model': 7.673187379988388e-05, 'epoch': 0.66}) +[2026-06-15 23:55:57,645][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.015683986246585846, 'train/grad_norm_pre_clip': 1.1216604709625244, 'timing/data': 0.01255340501666069, 'timing/model': 1.5600747652351856, 'learning_rate/qwen_vl_interface': 7.67056754990327e-06, 'learning_rate/action_model': 7.670567549903269e-05, 'epoch': 0.66}) +[2026-06-15 23:56:03,847][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.11065123230218887, 'train/grad_norm_pre_clip': 7.30963659286499, 'timing/data': 0.013649405911564827, 'timing/model': 1.5576509851962328, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.66}) +[2026-06-15 23:56:10,036][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.08974459767341614, 'train/grad_norm_pre_clip': 7.5025787353515625, 'timing/data': 0.01258002407848835, 'timing/model': 1.551440829411149, 'learning_rate/qwen_vl_interface': 7.665324905798694e-06, 'learning_rate/action_model': 7.665324905798695e-05, 'epoch': 0.66}) +[2026-06-15 23:56:16,139][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.16726070642471313, 'train/grad_norm_pre_clip': 15.029831886291504, 'timing/data': 0.012665417045354843, 'timing/model': 1.4762097131460905, 'learning_rate/qwen_vl_interface': 7.66270209393429e-06, 'learning_rate/action_model': 7.66270209393429e-05, 'epoch': 0.67}) +[2026-06-15 23:56:22,273][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.5518674850463867, 'train/grad_norm_pre_clip': 10.721603393554688, 'timing/data': 0.01176903024315834, 'timing/model': 1.7746680062264204, 'learning_rate/qwen_vl_interface': 7.66007829029919e-06, 'learning_rate/action_model': 7.660078290299191e-05, 'epoch': 0.67}) +[2026-06-15 23:56:28,238][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.24096670746803284, 'train/grad_norm_pre_clip': 22.282058715820312, 'timing/data': 0.011331625282764435, 'timing/model': 1.5392746329307556, 'learning_rate/qwen_vl_interface': 7.657453495971942e-06, 'learning_rate/action_model': 7.657453495971943e-05, 'epoch': 0.67}) +[2026-06-15 23:56:34,334][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.20494773983955383, 'train/grad_norm_pre_clip': 12.58581256866455, 'timing/data': 0.013076098635792732, 'timing/model': 1.5365133117884398, 'learning_rate/qwen_vl_interface': 7.6548277120315e-06, 'learning_rate/action_model': 7.6548277120315e-05, 'epoch': 0.67}) +[2026-06-15 23:56:40,434][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.49988439679145813, 'train/grad_norm_pre_clip': 14.8282470703125, 'timing/data': 0.012051524594426155, 'timing/model': 1.5410397313535213, 'learning_rate/qwen_vl_interface': 7.652200939557225e-06, 'learning_rate/action_model': 7.652200939557225e-05, 'epoch': 0.67}) +[2026-06-15 23:56:46,638][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.030393874272704124, 'train/grad_norm_pre_clip': 2.079624891281128, 'timing/data': 0.012982569634914398, 'timing/model': 1.5465675219893456, 'learning_rate/qwen_vl_interface': 7.649573179628881e-06, 'learning_rate/action_model': 7.649573179628881e-05, 'epoch': 0.67}) +[2026-06-15 23:56:52,738][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.2702900171279907, 'train/grad_norm_pre_clip': 8.14217472076416, 'timing/data': 0.012494122609496117, 'timing/model': 1.4974720161408186, 'learning_rate/qwen_vl_interface': 7.646944433326645e-06, 'learning_rate/action_model': 7.646944433326643e-05, 'epoch': 0.67}) +[2026-06-15 23:56:58,880][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.23816299438476562, 'train/grad_norm_pre_clip': 17.75009536743164, 'timing/data': 0.011757692322134972, 'timing/model': 1.7893011532723904, 'learning_rate/qwen_vl_interface': 7.64431470173109e-06, 'learning_rate/action_model': 7.64431470173109e-05, 'epoch': 0.67}) +[2026-06-15 23:57:05,023][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.10149334371089935, 'train/grad_norm_pre_clip': 7.645995616912842, 'timing/data': 0.013056067749857903, 'timing/model': 1.6912959553301334, 'learning_rate/qwen_vl_interface': 7.641683985923203e-06, 'learning_rate/action_model': 7.641683985923203e-05, 'epoch': 0.67}) +[2026-06-15 23:57:11,048][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.03605597838759422, 'train/grad_norm_pre_clip': 2.628678560256958, 'timing/data': 0.013029143214225769, 'timing/model': 1.5588021073490381, 'learning_rate/qwen_vl_interface': 7.63905228698437e-06, 'learning_rate/action_model': 7.63905228698437e-05, 'epoch': 0.67}) +[2026-06-15 23:57:17,158][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.16144613921642303, 'train/grad_norm_pre_clip': 4.381807327270508, 'timing/data': 0.012522688135504723, 'timing/model': 1.4872960038483143, 'learning_rate/qwen_vl_interface': 7.636419605996385e-06, 'learning_rate/action_model': 7.636419605996384e-05, 'epoch': 0.67}) +[2026-06-15 23:57:23,646][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.11515405029058456, 'train/grad_norm_pre_clip': 7.504403591156006, 'timing/data': 0.012583836913108826, 'timing/model': 1.5516707114875317, 'learning_rate/qwen_vl_interface': 7.633785944041444e-06, 'learning_rate/action_model': 7.633785944041444e-05, 'epoch': 0.67}) +[2026-06-15 23:57:29,845][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.2642900049686432, 'train/grad_norm_pre_clip': 12.097944259643555, 'timing/data': 0.012116730213165283, 'timing/model': 1.5167165845632553, 'learning_rate/qwen_vl_interface': 7.631151302202145e-06, 'learning_rate/action_model': 7.631151302202144e-05, 'epoch': 0.67}) +[2026-06-15 23:57:35,973][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.46274685859680176, 'train/grad_norm_pre_clip': 19.386756896972656, 'timing/data': 0.01062445342540741, 'timing/model': 1.692601127550006, 'learning_rate/qwen_vl_interface': 7.628515681561488e-06, 'learning_rate/action_model': 7.628515681561487e-05, 'epoch': 0.67}) +[2026-06-15 23:57:41,687][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.1971234828233719, 'train/grad_norm_pre_clip': 4.993325710296631, 'timing/data': 0.01119263470172882, 'timing/model': 1.45959354005754, 'learning_rate/qwen_vl_interface': 7.62587908320288e-06, 'learning_rate/action_model': 7.62587908320288e-05, 'epoch': 0.67}) +[2026-06-15 23:57:47,594][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.030964743345975876, 'train/grad_norm_pre_clip': 2.317688226699829, 'timing/data': 0.010506082326173782, 'timing/model': 1.4680167101323605, 'learning_rate/qwen_vl_interface': 7.6232415082101255e-06, 'learning_rate/action_model': 7.623241508210124e-05, 'epoch': 0.67}) +[2026-06-15 23:57:53,494][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.06854697316884995, 'train/grad_norm_pre_clip': 2.829540491104126, 'timing/data': 0.010630467906594276, 'timing/model': 1.4644078388810158, 'learning_rate/qwen_vl_interface': 7.620602957667432e-06, 'learning_rate/action_model': 7.620602957667432e-05, 'epoch': 0.67}) +[2026-06-15 23:57:59,392][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.11686984449625015, 'train/grad_norm_pre_clip': 6.045090675354004, 'timing/data': 0.011080993339419365, 'timing/model': 1.4438876044005156, 'learning_rate/qwen_vl_interface': 7.617963432659408e-06, 'learning_rate/action_model': 7.617963432659408e-05, 'epoch': 0.67}) +[2026-06-15 23:58:05,186][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.08852429687976837, 'train/grad_norm_pre_clip': 4.717060089111328, 'timing/data': 0.01071849837899208, 'timing/model': 1.4394607171416283, 'learning_rate/qwen_vl_interface': 7.615322934271063e-06, 'learning_rate/action_model': 7.615322934271064e-05, 'epoch': 0.67}) +[2026-06-15 23:58:11,026][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.06849483400583267, 'train/grad_norm_pre_clip': 6.9528326988220215, 'timing/data': 0.011039981618523598, 'timing/model': 1.592220213264227, 'learning_rate/qwen_vl_interface': 7.6126814635878055e-06, 'learning_rate/action_model': 7.612681463587805e-05, 'epoch': 0.67}) +[2026-06-15 23:58:16,681][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.12919804453849792, 'train/grad_norm_pre_clip': 8.953486442565918, 'timing/data': 0.011322835460305214, 'timing/model': 1.4512211121618748, 'learning_rate/qwen_vl_interface': 7.6100390216954445e-06, 'learning_rate/action_model': 7.610039021695445e-05, 'epoch': 0.67}) +[2026-06-15 23:58:22,585][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.10231060534715652, 'train/grad_norm_pre_clip': 3.5559308528900146, 'timing/data': 0.01103035919368267, 'timing/model': 1.4444118440151215, 'learning_rate/qwen_vl_interface': 7.607395609680189e-06, 'learning_rate/action_model': 7.60739560968019e-05, 'epoch': 0.67}) +[2026-06-15 23:58:28,494][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.08861680328845978, 'train/grad_norm_pre_clip': 5.920454025268555, 'timing/data': 0.011853504925966263, 'timing/model': 1.45887260325253, 'learning_rate/qwen_vl_interface': 7.604751228628644e-06, 'learning_rate/action_model': 7.604751228628644e-05, 'epoch': 0.67}) +[2026-06-15 23:58:34,382][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.044501908123493195, 'train/grad_norm_pre_clip': 2.8337130546569824, 'timing/data': 0.010900238528847694, 'timing/model': 1.427062839269638, 'learning_rate/qwen_vl_interface': 7.602105879627814e-06, 'learning_rate/action_model': 7.602105879627814e-05, 'epoch': 0.67}) +[2026-06-15 23:58:40,290][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.12252920866012573, 'train/grad_norm_pre_clip': 6.142753601074219, 'timing/data': 0.011000923812389374, 'timing/model': 1.46497599221766, 'learning_rate/qwen_vl_interface': 7.599459563765107e-06, 'learning_rate/action_model': 7.599459563765106e-05, 'epoch': 0.67}) +[2026-06-15 23:58:46,147][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.017273208126425743, 'train/grad_norm_pre_clip': 1.3853093385696411, 'timing/data': 0.01112469844520092, 'timing/model': 1.6101963855326176, 'learning_rate/qwen_vl_interface': 7.596812282128319e-06, 'learning_rate/action_model': 7.596812282128319e-05, 'epoch': 0.67}) +[2026-06-15 23:58:51,795][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.14607220888137817, 'train/grad_norm_pre_clip': 12.050354957580566, 'timing/data': 0.010229572653770447, 'timing/model': 1.4611835442483425, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 0.68}) +[2026-06-15 23:58:57,907][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.06929033249616623, 'train/grad_norm_pre_clip': 9.087998390197754, 'timing/data': 0.010188041254878044, 'timing/model': 1.6243443712592125, 'learning_rate/qwen_vl_interface': 7.59151482588569e-06, 'learning_rate/action_model': 7.59151482588569e-05, 'epoch': 0.68}) +[2026-06-15 23:59:03,812][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.15919296443462372, 'train/grad_norm_pre_clip': 6.858726978302002, 'timing/data': 0.01082373596727848, 'timing/model': 1.4492060989141464, 'learning_rate/qwen_vl_interface': 7.588864653457433e-06, 'learning_rate/action_model': 7.588864653457432e-05, 'epoch': 0.68}) +[2026-06-15 23:59:10,172][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.0669761672616005, 'train/grad_norm_pre_clip': 4.501565456390381, 'timing/data': 0.010508878156542778, 'timing/model': 1.7761941812932491, 'learning_rate/qwen_vl_interface': 7.5862135196102634e-06, 'learning_rate/action_model': 7.586213519610263e-05, 'epoch': 0.68}) +[2026-06-15 23:59:16,596][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.00476804468780756, 'train/grad_norm_pre_clip': 0.4754060208797455, 'timing/data': 0.013047371059656143, 'timing/model': 1.7555756643414497, 'learning_rate/qwen_vl_interface': 7.583561425433962e-06, 'learning_rate/action_model': 7.583561425433962e-05, 'epoch': 0.68}) +[2026-06-15 23:59:22,763][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.2615892291069031, 'train/grad_norm_pre_clip': 19.903705596923828, 'timing/data': 0.012938261032104492, 'timing/model': 1.540526956319809, 'learning_rate/qwen_vl_interface': 7.580908372018706e-06, 'learning_rate/action_model': 7.580908372018705e-05, 'epoch': 0.68}) +[2026-06-15 23:59:29,079][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.13791991770267487, 'train/grad_norm_pre_clip': 13.737671852111816, 'timing/data': 0.012858184054493904, 'timing/model': 1.5442749448120594, 'learning_rate/qwen_vl_interface': 7.5782543604550655e-06, 'learning_rate/action_model': 7.578254360455064e-05, 'epoch': 0.68}) +[2026-06-15 23:59:35,281][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.018673591315746307, 'train/grad_norm_pre_clip': 3.1760852336883545, 'timing/data': 0.012566717341542244, 'timing/model': 1.555478347465396, 'learning_rate/qwen_vl_interface': 7.575599391834e-06, 'learning_rate/action_model': 7.575599391834e-05, 'epoch': 0.68}) +[2026-06-15 23:59:41,573][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.16363607347011566, 'train/grad_norm_pre_clip': 8.758646965026855, 'timing/data': 0.013018250465393066, 'timing/model': 1.536513488739729, 'learning_rate/qwen_vl_interface': 7.572943467246871e-06, 'learning_rate/action_model': 7.57294346724687e-05, 'epoch': 0.68}) +[2026-06-15 23:59:47,842][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.11516697704792023, 'train/grad_norm_pre_clip': 6.366059303283691, 'timing/data': 0.013051886111497879, 'timing/model': 1.489916367456317, 'learning_rate/qwen_vl_interface': 7.570286587785428e-06, 'learning_rate/action_model': 7.570286587785428e-05, 'epoch': 0.68}) +[2026-06-15 23:59:53,977][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.13451647758483887, 'train/grad_norm_pre_clip': 12.44436264038086, 'timing/data': 0.013112079352140427, 'timing/model': 1.759662002325058, 'learning_rate/qwen_vl_interface': 7.567628754541815e-06, 'learning_rate/action_model': 7.567628754541814e-05, 'epoch': 0.68}) +[2026-06-16 00:00:00,048][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.40475159883499146, 'train/grad_norm_pre_clip': 19.0684871673584, 'timing/data': 0.01286388374865055, 'timing/model': 1.5602538082748652, 'learning_rate/qwen_vl_interface': 7.564969968608565e-06, 'learning_rate/action_model': 7.564969968608565e-05, 'epoch': 0.68}) +[2026-06-16 00:00:06,247][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.23172584176063538, 'train/grad_norm_pre_clip': 14.31870174407959, 'timing/data': 0.012710502371191978, 'timing/model': 1.5237754881381989, 'learning_rate/qwen_vl_interface': 7.562310231078605e-06, 'learning_rate/action_model': 7.562310231078604e-05, 'epoch': 0.68}) +[2026-06-16 00:00:12,473][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.0757012665271759, 'train/grad_norm_pre_clip': 4.612424373626709, 'timing/data': 0.011700911447405815, 'timing/model': 1.5290489960461855, 'learning_rate/qwen_vl_interface': 7.559649543045252e-06, 'learning_rate/action_model': 7.559649543045252e-05, 'epoch': 0.68}) +[2026-06-16 00:00:18,769][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.0626847892999649, 'train/grad_norm_pre_clip': 12.724285125732422, 'timing/data': 0.012597883120179176, 'timing/model': 1.5153280049562454, 'learning_rate/qwen_vl_interface': 7.556987905602215e-06, 'learning_rate/action_model': 7.556987905602215e-05, 'epoch': 0.68}) +[2026-06-16 00:00:25,079][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.04160644859075546, 'train/grad_norm_pre_clip': 6.430386543273926, 'timing/data': 0.012403223663568497, 'timing/model': 1.5153730232268572, 'learning_rate/qwen_vl_interface': 7.554325319843593e-06, 'learning_rate/action_model': 7.554325319843592e-05, 'epoch': 0.68}) +[2026-06-16 00:00:31,408][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.3174264132976532, 'train/grad_norm_pre_clip': 13.780116081237793, 'timing/data': 0.013438297435641289, 'timing/model': 1.7555325087159872, 'learning_rate/qwen_vl_interface': 7.551661786863874e-06, 'learning_rate/action_model': 7.551661786863874e-05, 'epoch': 0.68}) +[2026-06-16 00:00:37,472][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.15684551000595093, 'train/grad_norm_pre_clip': 14.349018096923828, 'timing/data': 0.013125313445925713, 'timing/model': 1.538145575672388, 'learning_rate/qwen_vl_interface': 7.548997307757936e-06, 'learning_rate/action_model': 7.548997307757935e-05, 'epoch': 0.68}) +[2026-06-16 00:00:43,886][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.11447322368621826, 'train/grad_norm_pre_clip': 6.3997721672058105, 'timing/data': 0.012146556749939919, 'timing/model': 1.5687279049307108, 'learning_rate/qwen_vl_interface': 7.546331883621044e-06, 'learning_rate/action_model': 7.546331883621044e-05, 'epoch': 0.68}) +[2026-06-16 00:00:50,187][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.27510377764701843, 'train/grad_norm_pre_clip': 6.939737796783447, 'timing/data': 0.01257309503853321, 'timing/model': 1.5586571749299765, 'learning_rate/qwen_vl_interface': 7.543665515548857e-06, 'learning_rate/action_model': 7.543665515548856e-05, 'epoch': 0.68}) +[2026-06-16 00:00:56,569][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.27600985765457153, 'train/grad_norm_pre_clip': 19.516530990600586, 'timing/data': 0.01221439242362976, 'timing/model': 1.574653159826994, 'learning_rate/qwen_vl_interface': 7.540998204637413e-06, 'learning_rate/action_model': 7.540998204637413e-05, 'epoch': 0.68}) +[2026-06-16 00:01:02,873][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.042461272329092026, 'train/grad_norm_pre_clip': 7.301154613494873, 'timing/data': 0.012978743761777878, 'timing/model': 1.5164835955947638, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.68}) +[2026-06-16 00:01:09,182][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.15315435826778412, 'train/grad_norm_pre_clip': 7.163990020751953, 'timing/data': 0.01353803463280201, 'timing/model': 1.788981581106782, 'learning_rate/qwen_vl_interface': 7.535660758682877e-06, 'learning_rate/action_model': 7.535660758682876e-05, 'epoch': 0.68}) +[2026-06-16 00:01:15,238][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.16552114486694336, 'train/grad_norm_pre_clip': 16.110797882080078, 'timing/data': 0.01289842464029789, 'timing/model': 1.553049298003316, 'learning_rate/qwen_vl_interface': 7.532990625833806e-06, 'learning_rate/action_model': 7.532990625833805e-05, 'epoch': 0.68}) +[2026-06-16 00:01:21,442][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.21733775734901428, 'train/grad_norm_pre_clip': 15.672624588012695, 'timing/data': 0.012619497254490852, 'timing/model': 1.5558213591575623, 'learning_rate/qwen_vl_interface': 7.5303195545335225e-06, 'learning_rate/action_model': 7.530319554533522e-05, 'epoch': 0.68}) +[2026-06-16 00:01:27,644][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.08406752347946167, 'train/grad_norm_pre_clip': 6.723327159881592, 'timing/data': 0.013049833476543427, 'timing/model': 1.5549499280750751, 'learning_rate/qwen_vl_interface': 7.527647545880007e-06, 'learning_rate/action_model': 7.527647545880007e-05, 'epoch': 0.69}) +[2026-06-16 00:01:33,746][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.23180606961250305, 'train/grad_norm_pre_clip': 8.691396713256836, 'timing/data': 0.013089852407574654, 'timing/model': 1.5527799855917692, 'learning_rate/qwen_vl_interface': 7.52497460097162e-06, 'learning_rate/action_model': 7.52497460097162e-05, 'epoch': 0.69}) +[2026-06-16 00:01:39,944][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.27150219678878784, 'train/grad_norm_pre_clip': 11.54996109008789, 'timing/data': 0.01244240254163742, 'timing/model': 1.5109767131507397, 'learning_rate/qwen_vl_interface': 7.522300720907105e-06, 'learning_rate/action_model': 7.522300720907105e-05, 'epoch': 0.69}) +[2026-06-16 00:01:46,079][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.0847063884139061, 'train/grad_norm_pre_clip': 5.9117302894592285, 'timing/data': 0.0111034344881773, 'timing/model': 1.7666169926524162, 'learning_rate/qwen_vl_interface': 7.519625906785601e-06, 'learning_rate/action_model': 7.5196259067856e-05, 'epoch': 0.69}) +[2026-06-16 00:01:52,147][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.05282694101333618, 'train/grad_norm_pre_clip': 5.423880100250244, 'timing/data': 0.012855520471930504, 'timing/model': 1.564466830343008, 'learning_rate/qwen_vl_interface': 7.516950159706614e-06, 'learning_rate/action_model': 7.516950159706614e-05, 'epoch': 0.69}) +[2026-06-16 00:01:58,344][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.08933300524950027, 'train/grad_norm_pre_clip': 7.032886028289795, 'timing/data': 0.013882016763091087, 'timing/model': 1.5561859160661697, 'learning_rate/qwen_vl_interface': 7.514273480770047e-06, 'learning_rate/action_model': 7.514273480770047e-05, 'epoch': 0.69}) +[2026-06-16 00:02:04,296][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.2128525823354721, 'train/grad_norm_pre_clip': 11.23379898071289, 'timing/data': 0.01148655079305172, 'timing/model': 1.459016928449273, 'learning_rate/qwen_vl_interface': 7.511595871076181e-06, 'learning_rate/action_model': 7.511595871076181e-05, 'epoch': 0.69}) +[2026-06-16 00:02:10,182][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.06709915399551392, 'train/grad_norm_pre_clip': 8.829597473144531, 'timing/data': 0.010448239743709564, 'timing/model': 1.4295049980282784, 'learning_rate/qwen_vl_interface': 7.50891733172568e-06, 'learning_rate/action_model': 7.50891733172568e-05, 'epoch': 0.69}) +[2026-06-16 00:02:16,159][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.1262555569410324, 'train/grad_norm_pre_clip': 7.278568744659424, 'timing/data': 0.012082751840353012, 'timing/model': 1.5033970642834902, 'learning_rate/qwen_vl_interface': 7.506237863819593e-06, 'learning_rate/action_model': 7.506237863819593e-05, 'epoch': 0.69}) +[2026-06-16 00:02:22,404][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.09990133345127106, 'train/grad_norm_pre_clip': 9.001126289367676, 'timing/data': 0.013342196121811867, 'timing/model': 1.797899192199111, 'learning_rate/qwen_vl_interface': 7.503557468459345e-06, 'learning_rate/action_model': 7.503557468459344e-05, 'epoch': 0.69}) +[2026-06-16 00:02:28,441][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.1594119817018509, 'train/grad_norm_pre_clip': 7.620095252990723, 'timing/data': 0.013075053691864014, 'timing/model': 1.5515215769410133, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.69}) +[2026-06-16 00:02:34,644][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.04737791791558266, 'train/grad_norm_pre_clip': 4.001708507537842, 'timing/data': 0.011479279026389122, 'timing/model': 1.5608903486281633, 'learning_rate/qwen_vl_interface': 7.498193899783983e-06, 'learning_rate/action_model': 7.498193899783983e-05, 'epoch': 0.69}) +[2026-06-16 00:02:40,844][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.03845769539475441, 'train/grad_norm_pre_clip': 2.4734394550323486, 'timing/data': 0.012310655787587166, 'timing/model': 1.5361492969095707, 'learning_rate/qwen_vl_interface': 7.495510728673631e-06, 'learning_rate/action_model': 7.495510728673632e-05, 'epoch': 0.69}) +[2026-06-16 00:02:47,042][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.20518726110458374, 'train/grad_norm_pre_clip': 12.875574111938477, 'timing/data': 0.012907426804304123, 'timing/model': 1.5281973276287317, 'learning_rate/qwen_vl_interface': 7.492826634518642e-06, 'learning_rate/action_model': 7.492826634518641e-05, 'epoch': 0.69}) +[2026-06-16 00:02:53,341][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.2822421193122864, 'train/grad_norm_pre_clip': 11.719695091247559, 'timing/data': 0.012476813048124313, 'timing/model': 1.5366421174257994, 'learning_rate/qwen_vl_interface': 7.4901416184223404e-06, 'learning_rate/action_model': 7.49014161842234e-05, 'epoch': 0.69}) +[2026-06-16 00:02:59,202][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.1769678145647049, 'train/grad_norm_pre_clip': 9.95262336730957, 'timing/data': 0.010875299572944641, 'timing/model': 1.6638453602790833, 'learning_rate/qwen_vl_interface': 7.487455681488437e-06, 'learning_rate/action_model': 7.487455681488436e-05, 'epoch': 0.69}) +[2026-06-16 00:03:04,883][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.16015134751796722, 'train/grad_norm_pre_clip': 5.822872638702393, 'timing/data': 0.010779540985822678, 'timing/model': 1.4402118045836687, 'learning_rate/qwen_vl_interface': 7.484768824821018e-06, 'learning_rate/action_model': 7.484768824821018e-05, 'epoch': 0.69}) +[2026-06-16 00:03:10,784][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.0804472491145134, 'train/grad_norm_pre_clip': 5.38889741897583, 'timing/data': 0.011015830561518669, 'timing/model': 1.449677912518382, 'learning_rate/qwen_vl_interface': 7.482081049524551e-06, 'learning_rate/action_model': 7.482081049524551e-05, 'epoch': 0.69}) +[2026-06-16 00:03:16,681][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.2625627815723419, 'train/grad_norm_pre_clip': 15.67250919342041, 'timing/data': 0.01008305698633194, 'timing/model': 1.4545567873865366, 'learning_rate/qwen_vl_interface': 7.479392356703878e-06, 'learning_rate/action_model': 7.479392356703878e-05, 'epoch': 0.69}) +[2026-06-16 00:03:22,583][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.05819482356309891, 'train/grad_norm_pre_clip': 3.4471230506896973, 'timing/data': 0.010958489030599594, 'timing/model': 1.4551566187292337, 'learning_rate/qwen_vl_interface': 7.476702747464219e-06, 'learning_rate/action_model': 7.476702747464219e-05, 'epoch': 0.69}) +[2026-06-16 00:03:28,482][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.026161503046751022, 'train/grad_norm_pre_clip': 2.1703720092773438, 'timing/data': 0.009658996015787125, 'timing/model': 1.4316675122827291, 'learning_rate/qwen_vl_interface': 7.474012222911169e-06, 'learning_rate/action_model': 7.474012222911169e-05, 'epoch': 0.69}) +[2026-06-16 00:03:34,380][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.2551485002040863, 'train/grad_norm_pre_clip': 15.293282508850098, 'timing/data': 0.01105082780122757, 'timing/model': 1.6509432569146156, 'learning_rate/qwen_vl_interface': 7.471320784150703e-06, 'learning_rate/action_model': 7.471320784150703e-05, 'epoch': 0.69}) +[2026-06-16 00:03:40,283][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.22492454946041107, 'train/grad_norm_pre_clip': 3.1488537788391113, 'timing/data': 0.00980977900326252, 'timing/model': 1.6506478376686573, 'learning_rate/qwen_vl_interface': 7.468628432289169e-06, 'learning_rate/action_model': 7.468628432289168e-05, 'epoch': 0.69}) +[2026-06-16 00:03:45,981][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.010377069935202599, 'train/grad_norm_pre_clip': 0.9885770082473755, 'timing/data': 0.011293318122625351, 'timing/model': 1.4532315507531166, 'learning_rate/qwen_vl_interface': 7.465935168433292e-06, 'learning_rate/action_model': 7.465935168433292e-05, 'epoch': 0.69}) +[2026-06-16 00:03:51,881][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.02471686340868473, 'train/grad_norm_pre_clip': 1.6250724792480469, 'timing/data': 0.009998124092817307, 'timing/model': 1.454146958887577, 'learning_rate/qwen_vl_interface': 7.463240993690168e-06, 'learning_rate/action_model': 7.463240993690168e-05, 'epoch': 0.69}) +[2026-06-16 00:03:57,782][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.1890258491039276, 'train/grad_norm_pre_clip': 8.36376953125, 'timing/data': 0.010384956374764442, 'timing/model': 1.4550700392574072, 'learning_rate/qwen_vl_interface': 7.460545909167276e-06, 'learning_rate/action_model': 7.460545909167275e-05, 'epoch': 0.69}) +[2026-06-16 00:04:03,686][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.06918509304523468, 'train/grad_norm_pre_clip': 5.459967613220215, 'timing/data': 0.010757872834801674, 'timing/model': 1.421320155262947, 'learning_rate/qwen_vl_interface': 7.457849915972459e-06, 'learning_rate/action_model': 7.457849915972458e-05, 'epoch': 0.7}) +[2026-06-16 00:04:09,584][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.14806528389453888, 'train/grad_norm_pre_clip': 13.140108108520508, 'timing/data': 0.010983418673276901, 'timing/model': 1.6530887186527252, 'learning_rate/qwen_vl_interface': 7.455153015213941e-06, 'learning_rate/action_model': 7.45515301521394e-05, 'epoch': 0.7}) +[2026-06-16 00:04:15,293][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.1397707611322403, 'train/grad_norm_pre_clip': 5.320542812347412, 'timing/data': 0.011154089123010635, 'timing/model': 1.464551841840148, 'learning_rate/qwen_vl_interface': 7.452455208000311e-06, 'learning_rate/action_model': 7.452455208000311e-05, 'epoch': 0.7}) +[2026-06-16 00:04:21,182][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.18498539924621582, 'train/grad_norm_pre_clip': 10.955448150634766, 'timing/data': 0.010227451100945473, 'timing/model': 1.453910119831562, 'learning_rate/qwen_vl_interface': 7.449756495440541e-06, 'learning_rate/action_model': 7.44975649544054e-05, 'epoch': 0.7}) +[2026-06-16 00:11:27,444][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.04554227739572525, 'train/grad_norm_pre_clip': 7.478892803192139, 'eval/loss': 0.07345417141914368, 'eval/flap_precision': 0.9299191374663073, 'eval/flap_recall': 0.9362279511533242, 'eval/flap_f1': 0.9330628803245435, 'eval/flap_rate_model': 0.2061111111111111, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9330628803245435, 'timing/data': 0.010388443246483803, 'timing/model': 1.458339735865593, 'learning_rate/qwen_vl_interface': 7.447056878643969e-06, 'learning_rate/action_model': 7.447056878643968e-05, 'epoch': 0.7}) +[2026-06-16 00:11:27,445][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1800_state +[2026-06-16 00:11:27,449][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 00:11:33,577][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1800_state/model.safetensors +[2026-06-16 00:11:48,669][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1800_state/optimizer.bin +[2026-06-16 00:11:48,670][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1800_state/sampler.bin +[2026-06-16 00:11:48,670][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1800_state/sampler_1.bin +[2026-06-16 00:11:48,672][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1800_state/random_states_0.pkl +[2026-06-16 00:11:48,706][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1800 +[2026-06-16 00:11:48,706][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 00:11:48,715][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 00:11:54,601][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.2707749307155609, 'train/grad_norm_pre_clip': 9.860091209411621, 'timing/data': 0.011604061350226402, 'timing/model': 1.4520674291998148, 'learning_rate/qwen_vl_interface': 7.444356358720303e-06, 'learning_rate/action_model': 7.444356358720302e-05, 'epoch': 0.7}) +[2026-06-16 00:12:00,502][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.09556650370359421, 'train/grad_norm_pre_clip': 9.295950889587402, 'timing/data': 0.011443240568041801, 'timing/model': 1.4517550524324179, 'learning_rate/qwen_vl_interface': 7.4416549367796255e-06, 'learning_rate/action_model': 7.441654936779625e-05, 'epoch': 0.7}) +[2026-06-16 00:12:06,403][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.19921541213989258, 'train/grad_norm_pre_clip': 18.672237396240234, 'timing/data': 0.010900799185037613, 'timing/model': 1.4390067160129547, 'learning_rate/qwen_vl_interface': 7.4389526139323895e-06, 'learning_rate/action_model': 7.438952613932389e-05, 'epoch': 0.7}) +[2026-06-16 00:12:12,274][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.26137834787368774, 'train/grad_norm_pre_clip': 10.771505355834961, 'timing/data': 0.010913338512182236, 'timing/model': 1.6542003490030766, 'learning_rate/qwen_vl_interface': 7.436249391289419e-06, 'learning_rate/action_model': 7.436249391289418e-05, 'epoch': 0.7}) +[2026-06-16 00:12:18,126][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.18298207223415375, 'train/grad_norm_pre_clip': 12.762600898742676, 'timing/data': 0.010496323928236961, 'timing/model': 1.6053930558264256, 'learning_rate/qwen_vl_interface': 7.433545269961906e-06, 'learning_rate/action_model': 7.433545269961906e-05, 'epoch': 0.7}) +[2026-06-16 00:12:24,174][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.045560307800769806, 'train/grad_norm_pre_clip': 5.349353313446045, 'timing/data': 0.012409001588821411, 'timing/model': 1.546942038461566, 'learning_rate/qwen_vl_interface': 7.430840251061411e-06, 'learning_rate/action_model': 7.430840251061411e-05, 'epoch': 0.7}) +[2026-06-16 00:12:30,474][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.20795540511608124, 'train/grad_norm_pre_clip': 8.726606369018555, 'timing/data': 0.012332260608673096, 'timing/model': 1.5457225497812033, 'learning_rate/qwen_vl_interface': 7.428134335699866e-06, 'learning_rate/action_model': 7.428134335699865e-05, 'epoch': 0.7}) +[2026-06-16 00:12:36,767][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.19893985986709595, 'train/grad_norm_pre_clip': 22.090999603271484, 'timing/data': 0.011874334886670113, 'timing/model': 1.5524667911231518, 'learning_rate/qwen_vl_interface': 7.425427524989572e-06, 'learning_rate/action_model': 7.425427524989572e-05, 'epoch': 0.7}) +[2026-06-16 00:12:43,080][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.16273289918899536, 'train/grad_norm_pre_clip': 7.845548152923584, 'timing/data': 0.013325296342372894, 'timing/model': 1.5139409005641937, 'learning_rate/qwen_vl_interface': 7.422719820043194e-06, 'learning_rate/action_model': 7.422719820043193e-05, 'epoch': 0.7}) +[2026-06-16 00:12:49,374][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.043134406208992004, 'train/grad_norm_pre_clip': 3.140132188796997, 'timing/data': 0.012591924518346786, 'timing/model': 1.542153699323535, 'learning_rate/qwen_vl_interface': 7.420011221973771e-06, 'learning_rate/action_model': 7.420011221973771e-05, 'epoch': 0.7}) +[2026-06-16 00:12:55,667][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.2107725590467453, 'train/grad_norm_pre_clip': 15.790112495422363, 'timing/data': 0.012557629495859146, 'timing/model': 1.7485547214746475, 'learning_rate/qwen_vl_interface': 7.417301731894702e-06, 'learning_rate/action_model': 7.417301731894701e-05, 'epoch': 0.7}) +[2026-06-16 00:13:01,775][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.07014964520931244, 'train/grad_norm_pre_clip': 4.71121883392334, 'timing/data': 0.012511665001511574, 'timing/model': 1.5541225038468838, 'learning_rate/qwen_vl_interface': 7.414591350919756e-06, 'learning_rate/action_model': 7.414591350919756e-05, 'epoch': 0.7}) +[2026-06-16 00:13:08,182][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.2094205915927887, 'train/grad_norm_pre_clip': 7.658817768096924, 'timing/data': 0.012999895960092545, 'timing/model': 1.5365613028407097, 'learning_rate/qwen_vl_interface': 7.4118800801630715e-06, 'learning_rate/action_model': 7.411880080163072e-05, 'epoch': 0.7}) +[2026-06-16 00:13:14,586][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.0623897910118103, 'train/grad_norm_pre_clip': 3.901575803756714, 'timing/data': 0.012041684240102768, 'timing/model': 1.5486273244023323, 'learning_rate/qwen_vl_interface': 7.409167920739146e-06, 'learning_rate/action_model': 7.409167920739147e-05, 'epoch': 0.7}) +[2026-06-16 00:13:20,990][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.08763112872838974, 'train/grad_norm_pre_clip': 4.684324264526367, 'timing/data': 0.011917201802134514, 'timing/model': 1.580417612567544, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.7}) +[2026-06-16 00:13:27,329][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.11348506063222885, 'train/grad_norm_pre_clip': 3.6977639198303223, 'timing/data': 0.012859150767326355, 'timing/model': 1.807426257058978, 'learning_rate/qwen_vl_interface': 7.403740940349408e-06, 'learning_rate/action_model': 7.403740940349408e-05, 'epoch': 0.7}) +[2026-06-16 00:13:33,326][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.06186613440513611, 'train/grad_norm_pre_clip': 11.29239273071289, 'timing/data': 0.009810274466872215, 'timing/model': 1.5032030045986176, 'learning_rate/qwen_vl_interface': 7.401026121614421e-06, 'learning_rate/action_model': 7.40102612161442e-05, 'epoch': 0.7}) +[2026-06-16 00:13:39,430][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.09178618341684341, 'train/grad_norm_pre_clip': 8.179499626159668, 'timing/data': 0.010697375982999802, 'timing/model': 1.5324113294482231, 'learning_rate/qwen_vl_interface': 7.398310418673846e-06, 'learning_rate/action_model': 7.398310418673846e-05, 'epoch': 0.7}) +[2026-06-16 00:13:45,199][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.18359969556331635, 'train/grad_norm_pre_clip': 8.533554077148438, 'timing/data': 0.010620808228850365, 'timing/model': 1.4496649000793695, 'learning_rate/qwen_vl_interface': 7.395593832644007e-06, 'learning_rate/action_model': 7.395593832644007e-05, 'epoch': 0.7}) +[2026-06-16 00:13:51,085][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.307401180267334, 'train/grad_norm_pre_clip': 12.195798873901367, 'timing/data': 0.01075655221939087, 'timing/model': 1.4419681746512651, 'learning_rate/qwen_vl_interface': 7.39287636464159e-06, 'learning_rate/action_model': 7.39287636464159e-05, 'epoch': 0.7}) +[2026-06-16 00:13:56,987][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.047497618943452835, 'train/grad_norm_pre_clip': 5.815485954284668, 'timing/data': 0.010843675583600998, 'timing/model': 1.4901806693524122, 'learning_rate/qwen_vl_interface': 7.3901580157836425e-06, 'learning_rate/action_model': 7.390158015783643e-05, 'epoch': 0.7}) +[2026-06-16 00:14:02,887][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.015205012634396553, 'train/grad_norm_pre_clip': 1.2492480278015137, 'timing/data': 0.009641656652092934, 'timing/model': 1.6325447838753462, 'learning_rate/qwen_vl_interface': 7.387438787187576e-06, 'learning_rate/action_model': 7.387438787187576e-05, 'epoch': 0.71}) +[2026-06-16 00:14:08,583][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.18058358132839203, 'train/grad_norm_pre_clip': 8.930401802062988, 'timing/data': 0.010513611137866974, 'timing/model': 1.4284206461161375, 'learning_rate/qwen_vl_interface': 7.3847186799711625e-06, 'learning_rate/action_model': 7.384718679971162e-05, 'epoch': 0.71}) +[2026-06-16 00:14:14,490][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.17056307196617126, 'train/grad_norm_pre_clip': 10.068853378295898, 'timing/data': 0.011374406516551971, 'timing/model': 1.4371724147349596, 'learning_rate/qwen_vl_interface': 7.381997695252535e-06, 'learning_rate/action_model': 7.381997695252535e-05, 'epoch': 0.71}) +[2026-06-16 00:14:20,385][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.06918255239725113, 'train/grad_norm_pre_clip': 5.036041736602783, 'timing/data': 0.010445356369018555, 'timing/model': 1.443452263250947, 'learning_rate/qwen_vl_interface': 7.379275834150186e-06, 'learning_rate/action_model': 7.379275834150186e-05, 'epoch': 0.71}) +[2026-06-16 00:14:26,181][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.020674491301178932, 'train/grad_norm_pre_clip': 1.129992961883545, 'timing/data': 0.011484045535326004, 'timing/model': 1.4552678111940622, 'learning_rate/qwen_vl_interface': 7.376553097782972e-06, 'learning_rate/action_model': 7.376553097782972e-05, 'epoch': 0.71}) +[2026-06-16 00:14:32,080][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.17846056818962097, 'train/grad_norm_pre_clip': 11.018585205078125, 'timing/data': 0.012491084635257721, 'timing/model': 1.4912150986492634, 'learning_rate/qwen_vl_interface': 7.373829487270109e-06, 'learning_rate/action_model': 7.373829487270109e-05, 'epoch': 0.71}) +[2026-06-16 00:14:37,974][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.05566025525331497, 'train/grad_norm_pre_clip': 5.104090213775635, 'timing/data': 0.011240869760513306, 'timing/model': 1.6321697253733873, 'learning_rate/qwen_vl_interface': 7.371105003731165e-06, 'learning_rate/action_model': 7.371105003731164e-05, 'epoch': 0.71}) +[2026-06-16 00:14:43,680][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.253136545419693, 'train/grad_norm_pre_clip': 20.97317886352539, 'timing/data': 0.011058364063501358, 'timing/model': 1.434665383771062, 'learning_rate/qwen_vl_interface': 7.3683796482860755e-06, 'learning_rate/action_model': 7.368379648286075e-05, 'epoch': 0.71}) +[2026-06-16 00:14:49,588][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.0728851929306984, 'train/grad_norm_pre_clip': 8.18471908569336, 'timing/data': 0.01070486381649971, 'timing/model': 1.450587797909975, 'learning_rate/qwen_vl_interface': 7.365653422055131e-06, 'learning_rate/action_model': 7.365653422055131e-05, 'epoch': 0.71}) +[2026-06-16 00:14:55,489][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.01943574659526348, 'train/grad_norm_pre_clip': 1.2794060707092285, 'timing/data': 0.01134464144706726, 'timing/model': 1.4470234978944063, 'learning_rate/qwen_vl_interface': 7.3629263261589816e-06, 'learning_rate/action_model': 7.362926326158981e-05, 'epoch': 0.71}) +[2026-06-16 00:15:01,390][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.014547400176525116, 'train/grad_norm_pre_clip': 1.692055106163025, 'timing/data': 0.010715538635849953, 'timing/model': 1.4320603869855404, 'learning_rate/qwen_vl_interface': 7.36019836171863e-06, 'learning_rate/action_model': 7.360198361718629e-05, 'epoch': 0.71}) +[2026-06-16 00:15:07,289][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.26667478680610657, 'train/grad_norm_pre_clip': 13.561387062072754, 'timing/data': 0.01002984493970871, 'timing/model': 1.492487657815218, 'learning_rate/qwen_vl_interface': 7.357469529855441e-06, 'learning_rate/action_model': 7.35746952985544e-05, 'epoch': 0.71}) +[2026-06-16 00:15:13,181][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.17263609170913696, 'train/grad_norm_pre_clip': 7.138169765472412, 'timing/data': 0.010599391534924507, 'timing/model': 1.6549095157533884, 'learning_rate/qwen_vl_interface': 7.354739831691134e-06, 'learning_rate/action_model': 7.354739831691135e-05, 'epoch': 0.71}) +[2026-06-16 00:15:18,881][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.11624694615602493, 'train/grad_norm_pre_clip': 5.471112251281738, 'timing/data': 0.010079314932227135, 'timing/model': 1.4532982874661684, 'learning_rate/qwen_vl_interface': 7.352009268347786e-06, 'learning_rate/action_model': 7.352009268347786e-05, 'epoch': 0.71}) +[2026-06-16 00:15:24,790][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.22924816608428955, 'train/grad_norm_pre_clip': 24.407480239868164, 'timing/data': 0.01100078970193863, 'timing/model': 1.457776602357626, 'learning_rate/qwen_vl_interface': 7.349277840947828e-06, 'learning_rate/action_model': 7.349277840947828e-05, 'epoch': 0.71}) +[2026-06-16 00:15:30,719][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.09262844175100327, 'train/grad_norm_pre_clip': 6.045297145843506, 'timing/data': 0.010849615558981895, 'timing/model': 1.4524759259074926, 'learning_rate/qwen_vl_interface': 7.346545550614046e-06, 'learning_rate/action_model': 7.346545550614046e-05, 'epoch': 0.71}) +[2026-06-16 00:15:36,579][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.016615191474556923, 'train/grad_norm_pre_clip': 2.7882931232452393, 'timing/data': 0.01012636348605156, 'timing/model': 1.424265667796135, 'learning_rate/qwen_vl_interface': 7.343812398469582e-06, 'learning_rate/action_model': 7.343812398469581e-05, 'epoch': 0.71}) +[2026-06-16 00:15:42,480][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.23335203528404236, 'train/grad_norm_pre_clip': 15.972896575927734, 'timing/data': 0.010711297392845154, 'timing/model': 1.4942560959607363, 'learning_rate/qwen_vl_interface': 7.34107838563793e-06, 'learning_rate/action_model': 7.341078385637929e-05, 'epoch': 0.71}) +[2026-06-16 00:15:48,384][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.01764105260372162, 'train/grad_norm_pre_clip': 2.455853223800659, 'timing/data': 0.009416429325938225, 'timing/model': 1.6578415166586637, 'learning_rate/qwen_vl_interface': 7.338343513242941e-06, 'learning_rate/action_model': 7.338343513242941e-05, 'epoch': 0.71}) +[2026-06-16 00:15:54,086][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.311353862285614, 'train/grad_norm_pre_clip': 15.77267837524414, 'timing/data': 0.01072554849088192, 'timing/model': 1.460213290527463, 'learning_rate/qwen_vl_interface': 7.335607782408818e-06, 'learning_rate/action_model': 7.335607782408818e-05, 'epoch': 0.71}) +[2026-06-16 00:15:59,984][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.5086550116539001, 'train/grad_norm_pre_clip': 19.455047607421875, 'timing/data': 0.010009566321969032, 'timing/model': 1.4578453600406647, 'learning_rate/qwen_vl_interface': 7.3328711942601145e-06, 'learning_rate/action_model': 7.332871194260115e-05, 'epoch': 0.71}) +[2026-06-16 00:16:05,783][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.13734844326972961, 'train/grad_norm_pre_clip': 10.508707046508789, 'timing/data': 0.0108683742582798, 'timing/model': 1.4521746877580881, 'learning_rate/qwen_vl_interface': 7.33013374992174e-06, 'learning_rate/action_model': 7.33013374992174e-05, 'epoch': 0.71}) +[2026-06-16 00:16:11,581][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.294792503118515, 'train/grad_norm_pre_clip': 23.925678253173828, 'timing/data': 0.010924477130174637, 'timing/model': 1.4456119500100613, 'learning_rate/qwen_vl_interface': 7.327395450518956e-06, 'learning_rate/action_model': 7.327395450518956e-05, 'epoch': 0.71}) +[2026-06-16 00:16:17,487][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.2106875330209732, 'train/grad_norm_pre_clip': 13.263955116271973, 'timing/data': 0.011322816833853722, 'timing/model': 1.4949737507849932, 'learning_rate/qwen_vl_interface': 7.324656297177372e-06, 'learning_rate/action_model': 7.324656297177371e-05, 'epoch': 0.71}) +[2026-06-16 00:16:23,467][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.22002850472927094, 'train/grad_norm_pre_clip': 5.17549991607666, 'timing/data': 0.010777510702610016, 'timing/model': 1.7437791302800179, 'learning_rate/qwen_vl_interface': 7.32191629102295e-06, 'learning_rate/action_model': 7.32191629102295e-05, 'epoch': 0.71}) +[2026-06-16 00:16:29,459][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.15020200610160828, 'train/grad_norm_pre_clip': 15.270973205566406, 'timing/data': 0.01238170638680458, 'timing/model': 1.4853954315185547, 'learning_rate/qwen_vl_interface': 7.319175433182004e-06, 'learning_rate/action_model': 7.319175433182004e-05, 'epoch': 0.71}) +[2026-06-16 00:16:35,566][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.13831163942813873, 'train/grad_norm_pre_clip': 11.789657592773438, 'timing/data': 0.01245284266769886, 'timing/model': 1.5046826228499413, 'learning_rate/qwen_vl_interface': 7.316433724781198e-06, 'learning_rate/action_model': 7.316433724781197e-05, 'epoch': 0.72}) +[2026-06-16 00:16:41,662][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.1143166795372963, 'train/grad_norm_pre_clip': 5.484840393066406, 'timing/data': 0.012706328183412552, 'timing/model': 1.577305432409048, 'learning_rate/qwen_vl_interface': 7.313691166947543e-06, 'learning_rate/action_model': 7.313691166947543e-05, 'epoch': 0.72}) +[2026-06-16 00:16:47,962][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.07947540283203125, 'train/grad_norm_pre_clip': 5.290705680847168, 'timing/data': 0.011946143582463264, 'timing/model': 1.575144700706005, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.72}) +[2026-06-16 00:16:54,247][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.08008423447608948, 'train/grad_norm_pre_clip': 3.877948760986328, 'timing/data': 0.01273874007165432, 'timing/model': 1.5706252120435238, 'learning_rate/qwen_vl_interface': 7.308203507491485e-06, 'learning_rate/action_model': 7.308203507491484e-05, 'epoch': 0.72}) +[2026-06-16 00:17:00,510][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.06019434705376625, 'train/grad_norm_pre_clip': 4.171360015869141, 'timing/data': 0.01254560798406601, 'timing/model': 1.8239284083247185, 'learning_rate/qwen_vl_interface': 7.3054584081248515e-06, 'learning_rate/action_model': 7.305458408124852e-05, 'epoch': 0.72}) +[2026-06-16 00:17:06,549][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.04726165533065796, 'train/grad_norm_pre_clip': 4.876519680023193, 'timing/data': 0.012231273576617241, 'timing/model': 1.5280606914311647, 'learning_rate/qwen_vl_interface': 7.302712463836907e-06, 'learning_rate/action_model': 7.302712463836907e-05, 'epoch': 0.72}) +[2026-06-16 00:17:12,742][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.18455706536769867, 'train/grad_norm_pre_clip': 9.77033805847168, 'timing/data': 0.013026081025600433, 'timing/model': 1.546453708782792, 'learning_rate/qwen_vl_interface': 7.2999656757564086e-06, 'learning_rate/action_model': 7.299965675756409e-05, 'epoch': 0.72}) +[2026-06-16 00:17:18,946][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.1512213498353958, 'train/grad_norm_pre_clip': 9.56443977355957, 'timing/data': 0.012529395520687103, 'timing/model': 1.55798758007586, 'learning_rate/qwen_vl_interface': 7.297218045012453e-06, 'learning_rate/action_model': 7.297218045012453e-05, 'epoch': 0.72}) +[2026-06-16 00:17:25,161][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.09189651161432266, 'train/grad_norm_pre_clip': 11.360357284545898, 'timing/data': 0.011830147355794907, 'timing/model': 1.5732564367353916, 'learning_rate/qwen_vl_interface': 7.29446957273449e-06, 'learning_rate/action_model': 7.294469572734489e-05, 'epoch': 0.72}) +[2026-06-16 00:17:31,563][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.14090879261493683, 'train/grad_norm_pre_clip': 9.513607025146484, 'timing/data': 0.012183081358671188, 'timing/model': 1.5814189743250608, 'learning_rate/qwen_vl_interface': 7.2917202600523104e-06, 'learning_rate/action_model': 7.29172026005231e-05, 'epoch': 0.72}) +[2026-06-16 00:17:37,763][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.07469256222248077, 'train/grad_norm_pre_clip': 5.28878927230835, 'timing/data': 0.01246984489262104, 'timing/model': 1.6950863748788834, 'learning_rate/qwen_vl_interface': 7.2889701080960565e-06, 'learning_rate/action_model': 7.288970108096056e-05, 'epoch': 0.72}) +[2026-06-16 00:17:43,564][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.020775271579623222, 'train/grad_norm_pre_clip': 1.4697837829589844, 'timing/data': 0.009448472410440445, 'timing/model': 1.6336324382573366, 'learning_rate/qwen_vl_interface': 7.286219117996209e-06, 'learning_rate/action_model': 7.286219117996207e-05, 'epoch': 0.72}) +[2026-06-16 00:17:49,182][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.21913160383701324, 'train/grad_norm_pre_clip': 12.92842960357666, 'timing/data': 0.010639248415827751, 'timing/model': 1.4477411266416311, 'learning_rate/qwen_vl_interface': 7.283467290883598e-06, 'learning_rate/action_model': 7.283467290883598e-05, 'epoch': 0.72}) +[2026-06-16 00:17:55,120][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.08044271171092987, 'train/grad_norm_pre_clip': 7.039404392242432, 'timing/data': 0.009942257776856422, 'timing/model': 1.4536628685891628, 'learning_rate/qwen_vl_interface': 7.2807146278893925e-06, 'learning_rate/action_model': 7.280714627889393e-05, 'epoch': 0.72}) +[2026-06-16 00:18:01,119][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.15328890085220337, 'train/grad_norm_pre_clip': 7.66383695602417, 'timing/data': 0.009327724575996399, 'timing/model': 1.4708571210503578, 'learning_rate/qwen_vl_interface': 7.277961130145111e-06, 'learning_rate/action_model': 7.277961130145111e-05, 'epoch': 0.72}) +[2026-06-16 00:18:07,007][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.11000647395849228, 'train/grad_norm_pre_clip': 6.384427070617676, 'timing/data': 0.009798374027013779, 'timing/model': 1.4519404899328947, 'learning_rate/qwen_vl_interface': 7.2752067987826116e-06, 'learning_rate/action_model': 7.275206798782611e-05, 'epoch': 0.72}) +[2026-06-16 00:18:12,917][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.06419215351343155, 'train/grad_norm_pre_clip': 9.204005241394043, 'timing/data': 0.009830635040998459, 'timing/model': 1.6651015374809504, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.72}) +[2026-06-16 00:18:18,619][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.239481121301651, 'train/grad_norm_pre_clip': 11.554183959960938, 'timing/data': 0.011104844510555267, 'timing/model': 1.4572315588593483, 'learning_rate/qwen_vl_interface': 7.269695639732108e-06, 'learning_rate/action_model': 7.269695639732108e-05, 'epoch': 0.72}) +[2026-06-16 00:18:24,630][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.09588849544525146, 'train/grad_norm_pre_clip': 8.359301567077637, 'timing/data': 0.01054343581199646, 'timing/model': 1.4712493158876896, 'learning_rate/qwen_vl_interface': 7.266938814309533e-06, 'learning_rate/action_model': 7.266938814309533e-05, 'epoch': 0.72}) +[2026-06-16 00:18:30,775][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.11316058039665222, 'train/grad_norm_pre_clip': 5.334068775177002, 'timing/data': 0.011298295110464096, 'timing/model': 1.556864507496357, 'learning_rate/qwen_vl_interface': 7.264181159799598e-06, 'learning_rate/action_model': 7.264181159799597e-05, 'epoch': 0.72}) +[2026-06-16 00:18:37,091][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.058335039764642715, 'train/grad_norm_pre_clip': 4.7473835945129395, 'timing/data': 0.012317940592765808, 'timing/model': 1.5595537759363651, 'learning_rate/qwen_vl_interface': 7.261422677335871e-06, 'learning_rate/action_model': 7.261422677335871e-05, 'epoch': 0.72}) +[2026-06-16 00:18:43,288][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.2735937535762787, 'train/grad_norm_pre_clip': 16.074338912963867, 'timing/data': 0.013190671801567078, 'timing/model': 1.535734448581934, 'learning_rate/qwen_vl_interface': 7.258663368052259e-06, 'learning_rate/action_model': 7.258663368052259e-05, 'epoch': 0.72}) +[2026-06-16 00:18:49,579][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.09400254487991333, 'train/grad_norm_pre_clip': 5.238927841186523, 'timing/data': 0.010970313102006912, 'timing/model': 1.763223646208644, 'learning_rate/qwen_vl_interface': 7.2559032330830105e-06, 'learning_rate/action_model': 7.25590323308301e-05, 'epoch': 0.72}) +[2026-06-16 00:18:55,646][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.3006915748119354, 'train/grad_norm_pre_clip': 15.021156311035156, 'timing/data': 0.012635359540581703, 'timing/model': 1.5592993628233671, 'learning_rate/qwen_vl_interface': 7.253142273562712e-06, 'learning_rate/action_model': 7.253142273562712e-05, 'epoch': 0.72}) +[2026-06-16 00:19:01,845][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.19845949113368988, 'train/grad_norm_pre_clip': 8.894790649414062, 'timing/data': 0.012212945148348808, 'timing/model': 1.558056116104126, 'learning_rate/qwen_vl_interface': 7.25038049062629e-06, 'learning_rate/action_model': 7.250380490626289e-05, 'epoch': 0.72}) +[2026-06-16 00:19:08,152][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.2553316652774811, 'train/grad_norm_pre_clip': 21.533262252807617, 'timing/data': 0.012490812689065933, 'timing/model': 1.5767598915845156, 'learning_rate/qwen_vl_interface': 7.24761788540901e-06, 'learning_rate/action_model': 7.24761788540901e-05, 'epoch': 0.72}) +[2026-06-16 00:19:14,371][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.06932879984378815, 'train/grad_norm_pre_clip': 4.349401950836182, 'timing/data': 0.01270754262804985, 'timing/model': 1.5104922782629728, 'learning_rate/qwen_vl_interface': 7.244854459046473e-06, 'learning_rate/action_model': 7.244854459046473e-05, 'epoch': 0.73}) +[2026-06-16 00:19:20,566][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.06653404980897903, 'train/grad_norm_pre_clip': 4.188286781311035, 'timing/data': 0.011855363845825195, 'timing/model': 1.5262687280774117, 'learning_rate/qwen_vl_interface': 7.24209021267462e-06, 'learning_rate/action_model': 7.24209021267462e-05, 'epoch': 0.73}) +[2026-06-16 00:19:26,846][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.05149875953793526, 'train/grad_norm_pre_clip': 2.518526554107666, 'timing/data': 0.013513609766960144, 'timing/model': 1.7391719296574593, 'learning_rate/qwen_vl_interface': 7.23932514742973e-06, 'learning_rate/action_model': 7.23932514742973e-05, 'epoch': 0.73}) +[2026-06-16 00:19:32,862][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.08935961872339249, 'train/grad_norm_pre_clip': 3.751009225845337, 'timing/data': 0.012972254306077957, 'timing/model': 1.5676100868731737, 'learning_rate/qwen_vl_interface': 7.236559264448413e-06, 'learning_rate/action_model': 7.236559264448412e-05, 'epoch': 0.73}) +[2026-06-16 00:19:39,157][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.07167627662420273, 'train/grad_norm_pre_clip': 7.370177745819092, 'timing/data': 0.011454453691840172, 'timing/model': 1.5431013945490122, 'learning_rate/qwen_vl_interface': 7.233792564867623e-06, 'learning_rate/action_model': 7.233792564867623e-05, 'epoch': 0.73}) +[2026-06-16 00:19:45,357][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.16532863676548004, 'train/grad_norm_pre_clip': 5.1078996658325195, 'timing/data': 0.012750662863254547, 'timing/model': 1.5063647255301476, 'learning_rate/qwen_vl_interface': 7.231025049824644e-06, 'learning_rate/action_model': 7.231025049824643e-05, 'epoch': 0.73}) +[2026-06-16 00:19:51,546][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.2632398307323456, 'train/grad_norm_pre_clip': 8.82878589630127, 'timing/data': 0.013377785682678223, 'timing/model': 1.5169525444507599, 'learning_rate/qwen_vl_interface': 7.228256720457097e-06, 'learning_rate/action_model': 7.228256720457096e-05, 'epoch': 0.73}) +[2026-06-16 00:19:57,774][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.07674965262413025, 'train/grad_norm_pre_clip': 5.417971611022949, 'timing/data': 0.011595828458666801, 'timing/model': 1.787058426067233, 'learning_rate/qwen_vl_interface': 7.225487577902937e-06, 'learning_rate/action_model': 7.225487577902936e-05, 'epoch': 0.73}) +[2026-06-16 00:20:03,844][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.046622101217508316, 'train/grad_norm_pre_clip': 2.542750835418701, 'timing/data': 0.013242727145552635, 'timing/model': 1.5603866055607796, 'learning_rate/qwen_vl_interface': 7.222717623300455e-06, 'learning_rate/action_model': 7.222717623300455e-05, 'epoch': 0.73}) +[2026-06-16 00:20:10,044][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.19547328352928162, 'train/grad_norm_pre_clip': 7.031624794006348, 'timing/data': 0.012585420161485672, 'timing/model': 1.5629777647554874, 'learning_rate/qwen_vl_interface': 7.219946857788273e-06, 'learning_rate/action_model': 7.219946857788272e-05, 'epoch': 0.73}) +[2026-06-16 00:20:16,163][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.19441968202590942, 'train/grad_norm_pre_clip': 7.802000522613525, 'timing/data': 0.012112867087125778, 'timing/model': 1.5857068356126547, 'learning_rate/qwen_vl_interface': 7.21717528250535e-06, 'learning_rate/action_model': 7.21717528250535e-05, 'epoch': 0.73}) +[2026-06-16 00:20:22,452][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.16582876443862915, 'train/grad_norm_pre_clip': 8.139755249023438, 'timing/data': 0.012362141162157059, 'timing/model': 1.5632785521447659, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 0.73}) +[2026-06-16 00:20:28,557][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.106959268450737, 'train/grad_norm_pre_clip': 7.8990631103515625, 'timing/data': 0.012619460001587868, 'timing/model': 1.5144058130681515, 'learning_rate/qwen_vl_interface': 7.211629707184768e-06, 'learning_rate/action_model': 7.211629707184768e-05, 'epoch': 0.73}) +[2026-06-16 00:20:34,888][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.12948882579803467, 'train/grad_norm_pre_clip': 8.238977432250977, 'timing/data': 0.012626193463802338, 'timing/model': 1.7853041123598814, 'learning_rate/qwen_vl_interface': 7.2088557094266885e-06, 'learning_rate/action_model': 7.208855709426688e-05, 'epoch': 0.73}) +[2026-06-16 00:20:41,186][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.12176894396543503, 'train/grad_norm_pre_clip': 6.156548023223877, 'timing/data': 0.012697594240307808, 'timing/model': 1.566046228632331, 'learning_rate/qwen_vl_interface': 7.2060809064570195e-06, 'learning_rate/action_model': 7.20608090645702e-05, 'epoch': 0.73}) +[2026-06-16 00:20:47,587][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.01931314915418625, 'train/grad_norm_pre_clip': 1.2850062847137451, 'timing/data': 0.012791940942406654, 'timing/model': 1.566575065255165, 'learning_rate/qwen_vl_interface': 7.203305299416375e-06, 'learning_rate/action_model': 7.203305299416375e-05, 'epoch': 0.73}) +[2026-06-16 00:20:53,889][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.183371901512146, 'train/grad_norm_pre_clip': 9.463034629821777, 'timing/data': 0.012875568121671677, 'timing/model': 1.5669186804443598, 'learning_rate/qwen_vl_interface': 7.200528889445707e-06, 'learning_rate/action_model': 7.200528889445706e-05, 'epoch': 0.73}) +[2026-06-16 00:21:00,183][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.361924946308136, 'train/grad_norm_pre_clip': 9.841242790222168, 'timing/data': 0.011451881378889084, 'timing/model': 1.564417002722621, 'learning_rate/qwen_vl_interface': 7.19775167768629e-06, 'learning_rate/action_model': 7.19775167768629e-05, 'epoch': 0.73}) +[2026-06-16 00:21:06,603][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.15461701154708862, 'train/grad_norm_pre_clip': 10.552261352539062, 'timing/data': 0.01239042729139328, 'timing/model': 1.6181621216237545, 'learning_rate/qwen_vl_interface': 7.194973665279731e-06, 'learning_rate/action_model': 7.19497366527973e-05, 'epoch': 0.73}) +[2026-06-16 00:21:12,931][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.12089326977729797, 'train/grad_norm_pre_clip': 3.716764211654663, 'timing/data': 0.013760266825556755, 'timing/model': 1.78442020714283, 'learning_rate/qwen_vl_interface': 7.1921948533679666e-06, 'learning_rate/action_model': 7.192194853367967e-05, 'epoch': 0.73}) +[2026-06-16 00:21:18,976][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.044491834938526154, 'train/grad_norm_pre_clip': 3.4547436237335205, 'timing/data': 0.012873701751232147, 'timing/model': 1.554104533046484, 'learning_rate/qwen_vl_interface': 7.1894152430932605e-06, 'learning_rate/action_model': 7.189415243093261e-05, 'epoch': 0.73}) +[2026-06-16 00:21:25,276][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.02886572852730751, 'train/grad_norm_pre_clip': 2.8546528816223145, 'timing/data': 0.011674955487251282, 'timing/model': 1.5642056409269571, 'learning_rate/qwen_vl_interface': 7.186634835598205e-06, 'learning_rate/action_model': 7.186634835598205e-05, 'epoch': 0.73}) +[2026-06-16 00:21:31,676][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.031939782202243805, 'train/grad_norm_pre_clip': 2.752288818359375, 'timing/data': 0.013568613678216934, 'timing/model': 1.521022766828537, 'learning_rate/qwen_vl_interface': 7.183853632025724e-06, 'learning_rate/action_model': 7.183853632025723e-05, 'epoch': 0.73}) +[2026-06-16 00:21:37,980][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.15745101869106293, 'train/grad_norm_pre_clip': 9.722770690917969, 'timing/data': 0.012511635199189186, 'timing/model': 1.5190922506153584, 'learning_rate/qwen_vl_interface': 7.181071633519062e-06, 'learning_rate/action_model': 7.181071633519062e-05, 'epoch': 0.73}) +[2026-06-16 00:21:44,275][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.21882657706737518, 'train/grad_norm_pre_clip': 6.55069637298584, 'timing/data': 0.011705458164215088, 'timing/model': 1.5458871237933636, 'learning_rate/qwen_vl_interface': 7.178288841221795e-06, 'learning_rate/action_model': 7.178288841221794e-05, 'epoch': 0.73}) +[2026-06-16 00:21:50,574][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.14213380217552185, 'train/grad_norm_pre_clip': 8.979968070983887, 'timing/data': 0.01249481551349163, 'timing/model': 1.7902248408645391, 'learning_rate/qwen_vl_interface': 7.1755052562778215e-06, 'learning_rate/action_model': 7.175505256277822e-05, 'epoch': 0.73}) +[2026-06-16 00:29:14,465][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.2563594877719879, 'train/grad_norm_pre_clip': 5.413326263427734, 'eval/loss': 0.05513652041554451, 'eval/flap_precision': 0.9165596919127086, 'eval/flap_recall': 0.9687924016282226, 'eval/flap_f1': 0.941952506596306, 'eval/flap_rate_model': 0.21638888888888888, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.941952506596306, 'timing/data': 0.012860730290412903, 'timing/model': 1.5533787924796343, 'learning_rate/qwen_vl_interface': 7.1727208798313704e-06, 'learning_rate/action_model': 7.172720879831371e-05, 'epoch': 0.74}) +[2026-06-16 00:29:14,466][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1900_state +[2026-06-16 00:29:14,473][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 00:29:25,131][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1900_state/model.safetensors +[2026-06-16 00:29:35,732][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1900_state/optimizer.bin +[2026-06-16 00:29:35,732][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1900_state/sampler.bin +[2026-06-16 00:29:35,733][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1900_state/sampler_1.bin +[2026-06-16 00:29:35,734][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_1900_state/random_states_0.pkl +[2026-06-16 00:29:35,803][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1900 +[2026-06-16 00:29:35,803][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 00:29:35,811][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 00:29:41,964][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.2916560173034668, 'train/grad_norm_pre_clip': 9.471929550170898, 'timing/data': 0.012349752709269524, 'timing/model': 1.7841380573809147, 'learning_rate/qwen_vl_interface': 7.169935713026994e-06, 'learning_rate/action_model': 7.169935713026994e-05, 'epoch': 0.74}) +[2026-06-16 00:29:48,131][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.077844999730587, 'train/grad_norm_pre_clip': 6.40632963180542, 'timing/data': 0.01371670886874199, 'timing/model': 1.4881788864731789, 'learning_rate/qwen_vl_interface': 7.1671497570095684e-06, 'learning_rate/action_model': 7.167149757009568e-05, 'epoch': 0.74}) +[2026-06-16 00:29:54,250][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.15818804502487183, 'train/grad_norm_pre_clip': 7.8343892097473145, 'timing/data': 0.012945396825671196, 'timing/model': 1.479343205690384, 'learning_rate/qwen_vl_interface': 7.164363012924295e-06, 'learning_rate/action_model': 7.164363012924294e-05, 'epoch': 0.74}) +[2026-06-16 00:30:00,433][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.07740525901317596, 'train/grad_norm_pre_clip': 3.50518798828125, 'timing/data': 0.012979023158550262, 'timing/model': 1.4690001253038645, 'learning_rate/qwen_vl_interface': 7.161575481916699e-06, 'learning_rate/action_model': 7.1615754819167e-05, 'epoch': 0.74}) +[2026-06-16 00:30:06,741][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.043547872453927994, 'train/grad_norm_pre_clip': 3.1746201515197754, 'timing/data': 0.011983077973127365, 'timing/model': 1.6432445999234915, 'learning_rate/qwen_vl_interface': 7.158787165132627e-06, 'learning_rate/action_model': 7.158787165132626e-05, 'epoch': 0.74}) +[2026-06-16 00:30:12,881][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.13338302075862885, 'train/grad_norm_pre_clip': 6.834331512451172, 'timing/data': 0.010813694447278976, 'timing/model': 1.4425264913588762, 'learning_rate/qwen_vl_interface': 7.1559980637182535e-06, 'learning_rate/action_model': 7.155998063718254e-05, 'epoch': 0.74}) +[2026-06-16 00:30:18,782][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.16236940026283264, 'train/grad_norm_pre_clip': 7.8292694091796875, 'timing/data': 0.011553781107068062, 'timing/model': 1.639296893030405, 'learning_rate/qwen_vl_interface': 7.153208178820071e-06, 'learning_rate/action_model': 7.153208178820072e-05, 'epoch': 0.74}) +[2026-06-16 00:30:24,486][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.15154050290584564, 'train/grad_norm_pre_clip': 7.544152736663818, 'timing/data': 0.010207461193203926, 'timing/model': 1.4559185430407524, 'learning_rate/qwen_vl_interface': 7.150417511584898e-06, 'learning_rate/action_model': 7.150417511584898e-05, 'epoch': 0.74}) +[2026-06-16 00:30:30,387][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.16219638288021088, 'train/grad_norm_pre_clip': 6.833129405975342, 'timing/data': 0.010310059413313866, 'timing/model': 1.4537550881505013, 'learning_rate/qwen_vl_interface': 7.147626063159872e-06, 'learning_rate/action_model': 7.147626063159872e-05, 'epoch': 0.74}) +[2026-06-16 00:30:36,288][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.14059261977672577, 'train/grad_norm_pre_clip': 9.583962440490723, 'timing/data': 0.01022709533572197, 'timing/model': 1.4539979677647352, 'learning_rate/qwen_vl_interface': 7.14483383469245e-06, 'learning_rate/action_model': 7.144833834692449e-05, 'epoch': 0.74}) +[2026-06-16 00:30:42,187][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.06162377819418907, 'train/grad_norm_pre_clip': 5.053524494171143, 'timing/data': 0.01098012924194336, 'timing/model': 1.4735073409974575, 'learning_rate/qwen_vl_interface': 7.142040827330411e-06, 'learning_rate/action_model': 7.142040827330411e-05, 'epoch': 0.74}) +[2026-06-16 00:30:48,079][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.1579175889492035, 'train/grad_norm_pre_clip': 6.981147766113281, 'timing/data': 0.010674048215150833, 'timing/model': 1.438301496207714, 'learning_rate/qwen_vl_interface': 7.139247042221859e-06, 'learning_rate/action_model': 7.139247042221858e-05, 'epoch': 0.74}) +[2026-06-16 00:30:54,041][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.13384704291820526, 'train/grad_norm_pre_clip': 5.889822483062744, 'timing/data': 0.010559691116213799, 'timing/model': 1.7057656683027744, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.74}) +[2026-06-16 00:31:00,016][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.021913059055805206, 'train/grad_norm_pre_clip': 1.4149388074874878, 'timing/data': 0.011361628770828247, 'timing/model': 1.5307866018265486, 'learning_rate/qwen_vl_interface': 7.133657143359203e-06, 'learning_rate/action_model': 7.133657143359202e-05, 'epoch': 0.74}) +[2026-06-16 00:31:06,236][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.12116359919309616, 'train/grad_norm_pre_clip': 5.872037887573242, 'timing/data': 0.012228403240442276, 'timing/model': 1.5506444051861763, 'learning_rate/qwen_vl_interface': 7.130861031902897e-06, 'learning_rate/action_model': 7.130861031902897e-05, 'epoch': 0.74}) +[2026-06-16 00:31:12,341][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.1819174736738205, 'train/grad_norm_pre_clip': 4.992373943328857, 'timing/data': 0.015178600326180458, 'timing/model': 1.4898494593799114, 'learning_rate/qwen_vl_interface': 7.128064147295667e-06, 'learning_rate/action_model': 7.128064147295667e-05, 'epoch': 0.74}) +[2026-06-16 00:31:18,414][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.058855459094047546, 'train/grad_norm_pre_clip': 3.6421239376068115, 'timing/data': 0.011338727548718452, 'timing/model': 1.5310510825365782, 'learning_rate/qwen_vl_interface': 7.125266490687206e-06, 'learning_rate/action_model': 7.125266490687206e-05, 'epoch': 0.74}) +[2026-06-16 00:31:24,450][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.1727236807346344, 'train/grad_norm_pre_clip': 8.580401420593262, 'timing/data': 0.012589136138558388, 'timing/model': 1.5374341122806072, 'learning_rate/qwen_vl_interface': 7.1224680632275265e-06, 'learning_rate/action_model': 7.122468063227527e-05, 'epoch': 0.74}) +[2026-06-16 00:31:30,415][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.09851240366697311, 'train/grad_norm_pre_clip': 7.3742570877075195, 'timing/data': 0.01076783612370491, 'timing/model': 1.6531477216631174, 'learning_rate/qwen_vl_interface': 7.1196688660669555e-06, 'learning_rate/action_model': 7.119668866066955e-05, 'epoch': 0.74}) +[2026-06-16 00:31:36,114][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.051202256232500076, 'train/grad_norm_pre_clip': 4.662549018859863, 'timing/data': 0.010150454938411713, 'timing/model': 1.4527097195386887, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.74}) +[2026-06-16 00:31:42,484][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.10576662421226501, 'train/grad_norm_pre_clip': 5.836271286010742, 'timing/data': 0.013652252033352852, 'timing/model': 1.5541512873023748, 'learning_rate/qwen_vl_interface': 7.114068167246033e-06, 'learning_rate/action_model': 7.114068167246033e-05, 'epoch': 0.74}) +[2026-06-16 00:31:48,783][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.1769750714302063, 'train/grad_norm_pre_clip': 11.024460792541504, 'timing/data': 0.01277739554643631, 'timing/model': 1.543799813836813, 'learning_rate/qwen_vl_interface': 7.111266667887916e-06, 'learning_rate/action_model': 7.111266667887915e-05, 'epoch': 0.74}) +[2026-06-16 00:31:55,052][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.27161550521850586, 'train/grad_norm_pre_clip': 9.07369613647461, 'timing/data': 0.013924209401011467, 'timing/model': 1.5628369767218828, 'learning_rate/qwen_vl_interface': 7.108464403433378e-06, 'learning_rate/action_model': 7.108464403433378e-05, 'epoch': 0.74}) +[2026-06-16 00:32:01,253][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.027964532375335693, 'train/grad_norm_pre_clip': 1.493192195892334, 'timing/data': 0.012015551328659058, 'timing/model': 1.5188612658530474, 'learning_rate/qwen_vl_interface': 7.105661375034324e-06, 'learning_rate/action_model': 7.105661375034324e-05, 'epoch': 0.74}) +[2026-06-16 00:32:07,591][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.08032083511352539, 'train/grad_norm_pre_clip': 5.0777435302734375, 'timing/data': 0.013567660003900528, 'timing/model': 1.8078135643154383, 'learning_rate/qwen_vl_interface': 7.102857583842973e-06, 'learning_rate/action_model': 7.102857583842974e-05, 'epoch': 0.75}) +[2026-06-16 00:32:13,826][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.21830682456493378, 'train/grad_norm_pre_clip': 7.618496894836426, 'timing/data': 0.013043424114584923, 'timing/model': 1.7281257212162018, 'learning_rate/qwen_vl_interface': 7.100053031011859e-06, 'learning_rate/action_model': 7.100053031011859e-05, 'epoch': 0.75}) +[2026-06-16 00:32:19,852][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.19663573801517487, 'train/grad_norm_pre_clip': 5.371036052703857, 'timing/data': 0.012042015790939331, 'timing/model': 1.5702030900865793, 'learning_rate/qwen_vl_interface': 7.097247717693827e-06, 'learning_rate/action_model': 7.097247717693827e-05, 'epoch': 0.75}) +[2026-06-16 00:32:26,153][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.1970837116241455, 'train/grad_norm_pre_clip': 11.810393333435059, 'timing/data': 0.01290547288954258, 'timing/model': 1.5640496034175158, 'learning_rate/qwen_vl_interface': 7.094441645042033e-06, 'learning_rate/action_model': 7.094441645042032e-05, 'epoch': 0.75}) +[2026-06-16 00:32:32,447][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.3177552819252014, 'train/grad_norm_pre_clip': 18.573322296142578, 'timing/data': 0.013078967109322548, 'timing/model': 1.5512980557978153, 'learning_rate/qwen_vl_interface': 7.09163481420995e-06, 'learning_rate/action_model': 7.09163481420995e-05, 'epoch': 0.75}) +[2026-06-16 00:32:38,642][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.024317748844623566, 'train/grad_norm_pre_clip': 2.828187942504883, 'timing/data': 0.012556377798318863, 'timing/model': 1.497177453711629, 'learning_rate/qwen_vl_interface': 7.0888272263513595e-06, 'learning_rate/action_model': 7.088827226351359e-05, 'epoch': 0.75}) +[2026-06-16 00:32:45,178][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.17330677807331085, 'train/grad_norm_pre_clip': 11.541889190673828, 'timing/data': 0.013002928346395493, 'timing/model': 1.7672522850334644, 'learning_rate/qwen_vl_interface': 7.086018882620354e-06, 'learning_rate/action_model': 7.086018882620354e-05, 'epoch': 0.75}) +[2026-06-16 00:32:51,252][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.0423361100256443, 'train/grad_norm_pre_clip': 1.9360469579696655, 'timing/data': 0.012785997241735458, 'timing/model': 1.5529099330306053, 'learning_rate/qwen_vl_interface': 7.083209784171341e-06, 'learning_rate/action_model': 7.083209784171341e-05, 'epoch': 0.75}) +[2026-06-16 00:32:57,549][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.06636077165603638, 'train/grad_norm_pre_clip': 10.602581024169922, 'timing/data': 0.013118680566549301, 'timing/model': 1.5505506973713636, 'learning_rate/qwen_vl_interface': 7.080399932159029e-06, 'learning_rate/action_model': 7.080399932159029e-05, 'epoch': 0.75}) +[2026-06-16 00:33:03,750][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.09278208017349243, 'train/grad_norm_pre_clip': 7.767446041107178, 'timing/data': 0.011849112808704376, 'timing/model': 1.543165098875761, 'learning_rate/qwen_vl_interface': 7.077589327738448e-06, 'learning_rate/action_model': 7.077589327738447e-05, 'epoch': 0.75}) +[2026-06-16 00:33:09,949][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.24733209609985352, 'train/grad_norm_pre_clip': 15.026117324829102, 'timing/data': 0.012892654165625572, 'timing/model': 1.5050513092428446, 'learning_rate/qwen_vl_interface': 7.0747779720649265e-06, 'learning_rate/action_model': 7.074777972064927e-05, 'epoch': 0.75}) +[2026-06-16 00:33:16,157][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.022012516856193542, 'train/grad_norm_pre_clip': 1.7432736158370972, 'timing/data': 0.012761266902089119, 'timing/model': 1.516492087393999, 'learning_rate/qwen_vl_interface': 7.071965866294108e-06, 'learning_rate/action_model': 7.071965866294108e-05, 'epoch': 0.75}) +[2026-06-16 00:33:22,474][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.009480009786784649, 'train/grad_norm_pre_clip': 0.7433788180351257, 'timing/data': 0.012026943266391754, 'timing/model': 1.761136194691062, 'learning_rate/qwen_vl_interface': 7.0691530115819454e-06, 'learning_rate/action_model': 7.069153011581945e-05, 'epoch': 0.75}) +[2026-06-16 00:33:28,568][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.22570808231830597, 'train/grad_norm_pre_clip': 10.398702621459961, 'timing/data': 0.011852560564875603, 'timing/model': 1.547138487920165, 'learning_rate/qwen_vl_interface': 7.066339409084694e-06, 'learning_rate/action_model': 7.066339409084694e-05, 'epoch': 0.75}) +[2026-06-16 00:33:34,982][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.2225784957408905, 'train/grad_norm_pre_clip': 37.887847900390625, 'timing/data': 0.012484384700655937, 'timing/model': 1.56563881598413, 'learning_rate/qwen_vl_interface': 7.0635250599589215e-06, 'learning_rate/action_model': 7.06352505995892e-05, 'epoch': 0.75}) +[2026-06-16 00:33:41,266][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.25341537594795227, 'train/grad_norm_pre_clip': 23.984708786010742, 'timing/data': 0.012540856376290321, 'timing/model': 1.5498335231095552, 'learning_rate/qwen_vl_interface': 7.060709965361499e-06, 'learning_rate/action_model': 7.060709965361499e-05, 'epoch': 0.75}) +[2026-06-16 00:33:47,557][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.024133343249559402, 'train/grad_norm_pre_clip': 1.5947239398956299, 'timing/data': 0.012071778997778893, 'timing/model': 1.5122142434120178, 'learning_rate/qwen_vl_interface': 7.057894126449606e-06, 'learning_rate/action_model': 7.057894126449606e-05, 'epoch': 0.75}) +[2026-06-16 00:33:53,664][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.1137004941701889, 'train/grad_norm_pre_clip': 6.464354991912842, 'timing/data': 0.01290198601782322, 'timing/model': 1.504913056269288, 'learning_rate/qwen_vl_interface': 7.055077544380728e-06, 'learning_rate/action_model': 7.055077544380728e-05, 'epoch': 0.75}) +[2026-06-16 00:33:59,971][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.12471409142017365, 'train/grad_norm_pre_clip': 6.742273330688477, 'timing/data': 0.012938316911458969, 'timing/model': 1.7696377858519554, 'learning_rate/qwen_vl_interface': 7.0522602203126554e-06, 'learning_rate/action_model': 7.052260220312656e-05, 'epoch': 0.75}) +[2026-06-16 00:34:06,056][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.22738733887672424, 'train/grad_norm_pre_clip': 12.708108901977539, 'timing/data': 0.013108018785715103, 'timing/model': 1.5453493651002645, 'learning_rate/qwen_vl_interface': 7.0494421554034845e-06, 'learning_rate/action_model': 7.049442155403484e-05, 'epoch': 0.75}) +[2026-06-16 00:34:12,349][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.05225466564297676, 'train/grad_norm_pre_clip': 5.899774074554443, 'timing/data': 0.012851005420088768, 'timing/model': 1.494065336883068, 'learning_rate/qwen_vl_interface': 7.046623350811612e-06, 'learning_rate/action_model': 7.046623350811612e-05, 'epoch': 0.75}) +[2026-06-16 00:34:18,556][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.09796663373708725, 'train/grad_norm_pre_clip': 5.950987815856934, 'timing/data': 0.011889295652508736, 'timing/model': 1.558394331485033, 'learning_rate/qwen_vl_interface': 7.043803807695747e-06, 'learning_rate/action_model': 7.043803807695747e-05, 'epoch': 0.75}) +[2026-06-16 00:34:24,856][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.15528148412704468, 'train/grad_norm_pre_clip': 9.119303703308105, 'timing/data': 0.012693444266915321, 'timing/model': 1.4935087766498327, 'learning_rate/qwen_vl_interface': 7.040983527214893e-06, 'learning_rate/action_model': 7.040983527214893e-05, 'epoch': 0.75}) +[2026-06-16 00:34:31,194][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.16990946233272552, 'train/grad_norm_pre_clip': 7.5264506340026855, 'timing/data': 0.012410994619131088, 'timing/model': 1.7997931335121393, 'learning_rate/qwen_vl_interface': 7.038162510528363e-06, 'learning_rate/action_model': 7.038162510528362e-05, 'epoch': 0.75}) +[2026-06-16 00:34:37,250][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.04526180773973465, 'train/grad_norm_pre_clip': 6.899406433105469, 'timing/data': 0.013003826141357422, 'timing/model': 1.54749945551157, 'learning_rate/qwen_vl_interface': 7.03534075879577e-06, 'learning_rate/action_model': 7.03534075879577e-05, 'epoch': 0.75}) +[2026-06-16 00:34:43,465][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.1746395379304886, 'train/grad_norm_pre_clip': 13.767539024353027, 'timing/data': 0.012727787718176842, 'timing/model': 1.5667826011776924, 'learning_rate/qwen_vl_interface': 7.032518273177028e-06, 'learning_rate/action_model': 7.032518273177028e-05, 'epoch': 0.75}) +[2026-06-16 00:34:49,655][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.09118474274873734, 'train/grad_norm_pre_clip': 19.839630126953125, 'timing/data': 0.013814510777592659, 'timing/model': 1.5571864042431116, 'learning_rate/qwen_vl_interface': 7.029695054832357e-06, 'learning_rate/action_model': 7.029695054832357e-05, 'epoch': 0.76}) +[2026-06-16 00:34:55,944][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.2728244364261627, 'train/grad_norm_pre_clip': 24.42095947265625, 'timing/data': 0.013020835816860199, 'timing/model': 1.5684548504650593, 'learning_rate/qwen_vl_interface': 7.026871104922273e-06, 'learning_rate/action_model': 7.026871104922273e-05, 'epoch': 0.76}) +[2026-06-16 00:35:02,151][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.15228164196014404, 'train/grad_norm_pre_clip': 14.312512397766113, 'timing/data': 0.013212587684392929, 'timing/model': 1.5437818206846714, 'learning_rate/qwen_vl_interface': 7.024046424607596e-06, 'learning_rate/action_model': 7.024046424607595e-05, 'epoch': 0.76}) +[2026-06-16 00:35:08,407][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.1616460531949997, 'train/grad_norm_pre_clip': 14.360322952270508, 'timing/data': 0.012036014348268509, 'timing/model': 1.8113323617726564, 'learning_rate/qwen_vl_interface': 7.021221015049445e-06, 'learning_rate/action_model': 7.021221015049446e-05, 'epoch': 0.76}) +[2026-06-16 00:35:14,453][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.20481736958026886, 'train/grad_norm_pre_clip': 16.79305648803711, 'timing/data': 0.012719912454485893, 'timing/model': 1.5564323402941227, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 0.76}) +[2026-06-16 00:35:20,847][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.10564139485359192, 'train/grad_norm_pre_clip': 7.216736316680908, 'timing/data': 0.012678952887654305, 'timing/model': 1.5448537096381187, 'learning_rate/qwen_vl_interface': 7.015568012848699e-06, 'learning_rate/action_model': 7.015568012848699e-05, 'epoch': 0.76}) +[2026-06-16 00:35:26,793][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.12067274749279022, 'train/grad_norm_pre_clip': 16.139724731445312, 'timing/data': 0.010122966021299362, 'timing/model': 1.458244251087308, 'learning_rate/qwen_vl_interface': 7.01274042252984e-06, 'learning_rate/action_model': 7.01274042252984e-05, 'epoch': 0.76}) +[2026-06-16 00:35:32,694][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.15046831965446472, 'train/grad_norm_pre_clip': 11.70344066619873, 'timing/data': 0.010873276740312576, 'timing/model': 1.4449847526848316, 'learning_rate/qwen_vl_interface': 7.009912107614978e-06, 'learning_rate/action_model': 7.009912107614978e-05, 'epoch': 0.76}) +[2026-06-16 00:35:38,594][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.025184391066432, 'train/grad_norm_pre_clip': 3.3745276927948, 'timing/data': 0.01023382693529129, 'timing/model': 1.4719068966805935, 'learning_rate/qwen_vl_interface': 7.0070830692667244e-06, 'learning_rate/action_model': 7.007083069266724e-05, 'epoch': 0.76}) +[2026-06-16 00:35:44,490][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.0740201324224472, 'train/grad_norm_pre_clip': 8.407905578613281, 'timing/data': 0.01042463444173336, 'timing/model': 1.644180878996849, 'learning_rate/qwen_vl_interface': 7.004253308647993e-06, 'learning_rate/action_model': 7.004253308647994e-05, 'epoch': 0.76}) +[2026-06-16 00:35:50,186][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.06475945562124252, 'train/grad_norm_pre_clip': 9.308229446411133, 'timing/data': 0.011225683614611626, 'timing/model': 1.4436625391244888, 'learning_rate/qwen_vl_interface': 7.001422826921991e-06, 'learning_rate/action_model': 7.00142282692199e-05, 'epoch': 0.76}) +[2026-06-16 00:35:56,086][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.33298587799072266, 'train/grad_norm_pre_clip': 8.79959487915039, 'timing/data': 0.010718340054154396, 'timing/model': 1.4495400674641132, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 0.76}) +[2026-06-16 00:36:01,994][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.012092554941773415, 'train/grad_norm_pre_clip': 1.250963807106018, 'timing/data': 0.010247047990560532, 'timing/model': 1.4499339498579502, 'learning_rate/qwen_vl_interface': 6.995759704802485e-06, 'learning_rate/action_model': 6.995759704802485e-05, 'epoch': 0.76}) +[2026-06-16 00:36:07,882][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.05225130915641785, 'train/grad_norm_pre_clip': 11.069673538208008, 'timing/data': 0.010929951444268227, 'timing/model': 1.438226044178009, 'learning_rate/qwen_vl_interface': 6.992927066736876e-06, 'learning_rate/action_model': 6.992927066736875e-05, 'epoch': 0.76}) +[2026-06-16 00:36:13,793][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.10136330127716064, 'train/grad_norm_pre_clip': 9.03447151184082, 'timing/data': 0.010637450963258743, 'timing/model': 1.492739100009203, 'learning_rate/qwen_vl_interface': 6.990093712219787e-06, 'learning_rate/action_model': 6.990093712219787e-05, 'epoch': 0.76}) +[2026-06-16 00:36:19,689][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.004247844684869051, 'train/grad_norm_pre_clip': 0.34965604543685913, 'timing/data': 0.010528337210416794, 'timing/model': 1.6604881137609482, 'learning_rate/qwen_vl_interface': 6.9872596424159e-06, 'learning_rate/action_model': 6.9872596424159e-05, 'epoch': 0.76}) +[2026-06-16 00:36:25,396][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.09138760715723038, 'train/grad_norm_pre_clip': 14.171976089477539, 'timing/data': 0.01157686859369278, 'timing/model': 1.448943680152297, 'learning_rate/qwen_vl_interface': 6.9844248584902e-06, 'learning_rate/action_model': 6.984424858490199e-05, 'epoch': 0.76}) +[2026-06-16 00:36:31,285][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.21465657651424408, 'train/grad_norm_pre_clip': 13.79564380645752, 'timing/data': 0.010733714327216148, 'timing/model': 1.457655232399702, 'learning_rate/qwen_vl_interface': 6.981589361607952e-06, 'learning_rate/action_model': 6.981589361607953e-05, 'epoch': 0.76}) +[2026-06-16 00:36:37,179][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.04881707951426506, 'train/grad_norm_pre_clip': 10.87646770477295, 'timing/data': 0.010427035391330719, 'timing/model': 1.4529510401189327, 'learning_rate/qwen_vl_interface': 6.9787531529347274e-06, 'learning_rate/action_model': 6.978753152934727e-05, 'epoch': 0.76}) +[2026-06-16 00:36:43,083][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.21414531767368317, 'train/grad_norm_pre_clip': 18.469982147216797, 'timing/data': 0.0103409793227911, 'timing/model': 1.4793044235557318, 'learning_rate/qwen_vl_interface': 6.97591623363638e-06, 'learning_rate/action_model': 6.97591623363638e-05, 'epoch': 0.76}) +[2026-06-16 00:36:49,234][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.3307705223560333, 'train/grad_norm_pre_clip': 17.791948318481445, 'timing/data': 0.013085341081023216, 'timing/model': 1.4834465719759464, 'learning_rate/qwen_vl_interface': 6.9730786048790644e-06, 'learning_rate/action_model': 6.973078604879064e-05, 'epoch': 0.76}) +[2026-06-16 00:36:55,083][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.027858208864927292, 'train/grad_norm_pre_clip': 3.995039939880371, 'timing/data': 0.011214066296815872, 'timing/model': 1.655345058068633, 'learning_rate/qwen_vl_interface': 6.970240267829221e-06, 'learning_rate/action_model': 6.970240267829222e-05, 'epoch': 0.76}) +[2026-06-16 00:37:00,783][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.1495678424835205, 'train/grad_norm_pre_clip': 8.868284225463867, 'timing/data': 0.009802203625440598, 'timing/model': 1.4449902903288603, 'learning_rate/qwen_vl_interface': 6.967401223653582e-06, 'learning_rate/action_model': 6.967401223653581e-05, 'epoch': 0.76}) +[2026-06-16 00:37:06,683][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.043022520840168, 'train/grad_norm_pre_clip': 8.643407821655273, 'timing/data': 0.010690050199627876, 'timing/model': 1.4506051875650883, 'learning_rate/qwen_vl_interface': 6.9645614735191724e-06, 'learning_rate/action_model': 6.964561473519172e-05, 'epoch': 0.76}) +[2026-06-16 00:37:12,483][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.07992482930421829, 'train/grad_norm_pre_clip': 6.204268932342529, 'timing/data': 0.010641567409038544, 'timing/model': 1.4548136703670025, 'learning_rate/qwen_vl_interface': 6.961721018593305e-06, 'learning_rate/action_model': 6.961721018593305e-05, 'epoch': 0.76}) +[2026-06-16 00:37:18,383][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.19309599697589874, 'train/grad_norm_pre_clip': 7.3017096519470215, 'timing/data': 0.011575736105442047, 'timing/model': 1.4522137437015772, 'learning_rate/qwen_vl_interface': 6.958879860043585e-06, 'learning_rate/action_model': 6.958879860043585e-05, 'epoch': 0.76}) +[2026-06-16 00:37:24,282][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.0391000472009182, 'train/grad_norm_pre_clip': 4.846049785614014, 'timing/data': 0.009850569069385529, 'timing/model': 1.43558319658041, 'learning_rate/qwen_vl_interface': 6.9560379990379025e-06, 'learning_rate/action_model': 6.956037999037902e-05, 'epoch': 0.77}) +[2026-06-16 00:37:30,191][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.08239950984716415, 'train/grad_norm_pre_clip': 12.855324745178223, 'timing/data': 0.009860558435320854, 'timing/model': 1.6622586213052273, 'learning_rate/qwen_vl_interface': 6.953195436744442e-06, 'learning_rate/action_model': 6.953195436744442e-05, 'epoch': 0.77}) +[2026-06-16 00:37:35,886][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.21880726516246796, 'train/grad_norm_pre_clip': 11.051030158996582, 'timing/data': 0.010145207867026329, 'timing/model': 1.4568185545504093, 'learning_rate/qwen_vl_interface': 6.950352174331676e-06, 'learning_rate/action_model': 6.950352174331675e-05, 'epoch': 0.77}) +[2026-06-16 00:37:41,784][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.2270595282316208, 'train/grad_norm_pre_clip': 14.813127517700195, 'timing/data': 0.011024592444300652, 'timing/model': 1.4530811309814453, 'learning_rate/qwen_vl_interface': 6.947508212968356e-06, 'learning_rate/action_model': 6.947508212968356e-05, 'epoch': 0.77}) +[2026-06-16 00:37:47,685][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.009683539159595966, 'train/grad_norm_pre_clip': 0.9961959719657898, 'timing/data': 0.010302398353815079, 'timing/model': 1.4478379227221012, 'learning_rate/qwen_vl_interface': 6.9446635538235295e-06, 'learning_rate/action_model': 6.944663553823529e-05, 'epoch': 0.77}) +[2026-06-16 00:37:53,583][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.08734779804944992, 'train/grad_norm_pre_clip': 3.893827199935913, 'timing/data': 0.01069994829595089, 'timing/model': 1.4430961981415749, 'learning_rate/qwen_vl_interface': 6.94181819806653e-06, 'learning_rate/action_model': 6.94181819806653e-05, 'epoch': 0.77}) +[2026-06-16 00:37:59,482][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.07108908146619797, 'train/grad_norm_pre_clip': 4.819019317626953, 'timing/data': 0.009991155937314034, 'timing/model': 1.4498778488487005, 'learning_rate/qwen_vl_interface': 6.938972146866975e-06, 'learning_rate/action_model': 6.938972146866974e-05, 'epoch': 0.77}) +[2026-06-16 00:38:05,390][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.053522463887929916, 'train/grad_norm_pre_clip': 6.377292633056641, 'timing/data': 0.01064014621078968, 'timing/model': 1.6582892891019583, 'learning_rate/qwen_vl_interface': 6.936125401394769e-06, 'learning_rate/action_model': 6.936125401394768e-05, 'epoch': 0.77}) +[2026-06-16 00:38:11,080][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.030835598707199097, 'train/grad_norm_pre_clip': 2.801424741744995, 'timing/data': 0.010639240965247154, 'timing/model': 1.4526996929198503, 'learning_rate/qwen_vl_interface': 6.933277962820101e-06, 'learning_rate/action_model': 6.9332779628201e-05, 'epoch': 0.77}) +[2026-06-16 00:38:17,651][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.061743464320898056, 'train/grad_norm_pre_clip': 5.552354335784912, 'timing/data': 0.012059222906827927, 'timing/model': 1.556308699771762, 'learning_rate/qwen_vl_interface': 6.930429832313444e-06, 'learning_rate/action_model': 6.930429832313444e-05, 'epoch': 0.77}) +[2026-06-16 00:38:23,941][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.19231447577476501, 'train/grad_norm_pre_clip': 15.517385482788086, 'timing/data': 0.01362907886505127, 'timing/model': 1.5426813513040543, 'learning_rate/qwen_vl_interface': 6.92758101104556e-06, 'learning_rate/action_model': 6.92758101104556e-05, 'epoch': 0.77}) +[2026-06-16 00:38:30,153][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.04485601931810379, 'train/grad_norm_pre_clip': 6.447232246398926, 'timing/data': 0.012218078598380089, 'timing/model': 1.5590766798704863, 'learning_rate/qwen_vl_interface': 6.9247315001874885e-06, 'learning_rate/action_model': 6.924731500187488e-05, 'epoch': 0.77}) +[2026-06-16 00:38:36,296][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.12140632420778275, 'train/grad_norm_pre_clip': 8.18242073059082, 'timing/data': 0.010903222486376762, 'timing/model': 1.4532712176442146, 'learning_rate/qwen_vl_interface': 6.9218813009105605e-06, 'learning_rate/action_model': 6.92188130091056e-05, 'epoch': 0.77}) +[2026-06-16 00:38:42,190][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.01001864206045866, 'train/grad_norm_pre_clip': 1.3611416816711426, 'timing/data': 0.011180393397808075, 'timing/model': 1.6528850700706244, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.77}) +[2026-06-16 00:38:48,069][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.03989928215742111, 'train/grad_norm_pre_clip': 4.162353038787842, 'timing/data': 0.010906735435128212, 'timing/model': 1.6301753651350737, 'learning_rate/qwen_vl_interface': 6.916178841786839e-06, 'learning_rate/action_model': 6.916178841786838e-05, 'epoch': 0.77}) +[2026-06-16 00:38:53,787][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.09666340798139572, 'train/grad_norm_pre_clip': 5.461154460906982, 'timing/data': 0.010666275396943092, 'timing/model': 1.456453062593937, 'learning_rate/qwen_vl_interface': 6.913326584284115e-06, 'learning_rate/action_model': 6.913326584284115e-05, 'epoch': 0.77}) +[2026-06-16 00:38:59,689][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.2059854418039322, 'train/grad_norm_pre_clip': 25.969926834106445, 'timing/data': 0.010875940322875977, 'timing/model': 1.4246109183877707, 'learning_rate/qwen_vl_interface': 6.910473643050662e-06, 'learning_rate/action_model': 6.910473643050662e-05, 'epoch': 0.77}) +[2026-06-16 00:39:05,594][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.0790531262755394, 'train/grad_norm_pre_clip': 25.287067413330078, 'timing/data': 0.01008530892431736, 'timing/model': 1.4581900276243687, 'learning_rate/qwen_vl_interface': 6.907620019259215e-06, 'learning_rate/action_model': 6.907620019259216e-05, 'epoch': 0.77}) +[2026-06-16 00:39:11,492][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.013024548999965191, 'train/grad_norm_pre_clip': 2.073657512664795, 'timing/data': 0.010048748925328255, 'timing/model': 1.4424248524010181, 'learning_rate/qwen_vl_interface': 6.9047657140827935e-06, 'learning_rate/action_model': 6.904765714082794e-05, 'epoch': 0.77}) +[2026-06-16 00:39:17,386][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.3858347237110138, 'train/grad_norm_pre_clip': 19.031225204467773, 'timing/data': 0.01001683995127678, 'timing/model': 1.653208438307047, 'learning_rate/qwen_vl_interface': 6.901910728694694e-06, 'learning_rate/action_model': 6.901910728694694e-05, 'epoch': 0.77}) +[2026-06-16 00:39:23,360][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.2623654007911682, 'train/grad_norm_pre_clip': 20.235164642333984, 'timing/data': 0.012605009600520134, 'timing/model': 1.5387349221855402, 'learning_rate/qwen_vl_interface': 6.899055064268489e-06, 'learning_rate/action_model': 6.899055064268489e-05, 'epoch': 0.77}) +[2026-06-16 00:39:29,563][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.29853469133377075, 'train/grad_norm_pre_clip': 15.927433013916016, 'timing/data': 0.011480286717414856, 'timing/model': 1.542478395625949, 'learning_rate/qwen_vl_interface': 6.896198721978038e-06, 'learning_rate/action_model': 6.896198721978038e-05, 'epoch': 0.77}) +[2026-06-16 00:39:35,855][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.07401608675718307, 'train/grad_norm_pre_clip': 5.870882987976074, 'timing/data': 0.012261917814612389, 'timing/model': 1.5527571234852076, 'learning_rate/qwen_vl_interface': 6.893341702997475e-06, 'learning_rate/action_model': 6.893341702997475e-05, 'epoch': 0.77}) +[2026-06-16 00:47:46,952][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.08366727083921432, 'train/grad_norm_pre_clip': 19.313108444213867, 'eval/loss': 0.06600970029830933, 'eval/flap_precision': 0.9358288770053476, 'eval/flap_recall': 0.9497964721845319, 'eval/flap_f1': 0.9427609427609428, 'eval/flap_rate_model': 0.20777777777777778, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9427609427609428, 'timing/data': 0.012517232447862625, 'timing/model': 1.5643445178866386, 'learning_rate/qwen_vl_interface': 6.890484008501211e-06, 'learning_rate/action_model': 6.890484008501211e-05, 'epoch': 0.77}) +[2026-06-16 00:47:46,953][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2000_state +[2026-06-16 00:47:46,960][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 00:47:54,465][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2000_state/model.safetensors +[2026-06-16 00:48:03,951][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2000_state/optimizer.bin +[2026-06-16 00:48:03,951][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2000_state/sampler.bin +[2026-06-16 00:48:03,952][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-16 00:48:03,953][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-16 00:48:04,061][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-16 00:48:04,061][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 00:48:04,070][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 00:48:10,015][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.24393330514431, 'train/grad_norm_pre_clip': 11.894477844238281, 'timing/data': 0.010624485090374947, 'timing/model': 1.4508874006569386, 'learning_rate/qwen_vl_interface': 6.887625639663941e-06, 'learning_rate/action_model': 6.88762563966394e-05, 'epoch': 0.77}) +[2026-06-16 00:48:16,564][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.18618085980415344, 'train/grad_norm_pre_clip': 11.880622863769531, 'timing/data': 0.015400739386677742, 'timing/model': 1.5979256518185139, 'learning_rate/qwen_vl_interface': 6.884766597660626e-06, 'learning_rate/action_model': 6.884766597660626e-05, 'epoch': 0.77}) +[2026-06-16 00:48:22,882][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.045603930950164795, 'train/grad_norm_pre_clip': 8.736760139465332, 'timing/data': 0.01221042312681675, 'timing/model': 1.7900390587747097, 'learning_rate/qwen_vl_interface': 6.881906883666515e-06, 'learning_rate/action_model': 6.881906883666515e-05, 'epoch': 0.78}) +[2026-06-16 00:48:29,028][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.06023956835269928, 'train/grad_norm_pre_clip': 14.9968900680542, 'timing/data': 0.013490308076143265, 'timing/model': 1.6790724210441113, 'learning_rate/qwen_vl_interface': 6.879046498857125e-06, 'learning_rate/action_model': 6.879046498857124e-05, 'epoch': 0.78}) +[2026-06-16 00:48:34,916][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.4267682433128357, 'train/grad_norm_pre_clip': 28.86698341369629, 'timing/data': 0.012022653594613075, 'timing/model': 1.4539162907749414, 'learning_rate/qwen_vl_interface': 6.876185444408255e-06, 'learning_rate/action_model': 6.876185444408255e-05, 'epoch': 0.78}) +[2026-06-16 00:48:40,913][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.28230494260787964, 'train/grad_norm_pre_clip': 20.21778678894043, 'timing/data': 0.013493850827217102, 'timing/model': 1.4580112155526876, 'learning_rate/qwen_vl_interface': 6.873323721495976e-06, 'learning_rate/action_model': 6.873323721495976e-05, 'epoch': 0.78}) +[2026-06-16 00:48:46,921][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.19492198526859283, 'train/grad_norm_pre_clip': 18.56702995300293, 'timing/data': 0.01202765665948391, 'timing/model': 1.4562533386051655, 'learning_rate/qwen_vl_interface': 6.870461331296632e-06, 'learning_rate/action_model': 6.870461331296632e-05, 'epoch': 0.78}) +[2026-06-16 00:48:52,912][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.24660059809684753, 'train/grad_norm_pre_clip': 16.206134796142578, 'timing/data': 0.012036003172397614, 'timing/model': 1.5019940081983805, 'learning_rate/qwen_vl_interface': 6.8675982749868455e-06, 'learning_rate/action_model': 6.867598274986845e-05, 'epoch': 0.78}) +[2026-06-16 00:48:58,929][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.1839534193277359, 'train/grad_norm_pre_clip': 14.026527404785156, 'timing/data': 0.011156029999256134, 'timing/model': 1.4584139958024025, 'learning_rate/qwen_vl_interface': 6.864734553743509e-06, 'learning_rate/action_model': 6.86473455374351e-05, 'epoch': 0.78}) +[2026-06-16 00:49:04,828][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.0941002145409584, 'train/grad_norm_pre_clip': 14.690469741821289, 'timing/data': 0.009844489395618439, 'timing/model': 1.622769283130765, 'learning_rate/qwen_vl_interface': 6.861870168743793e-06, 'learning_rate/action_model': 6.861870168743793e-05, 'epoch': 0.78}) +[2026-06-16 00:49:10,461][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.021523844450712204, 'train/grad_norm_pre_clip': 2.1915929317474365, 'timing/data': 0.010885270312428474, 'timing/model': 1.4322381019592285, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 0.78}) +[2026-06-16 00:49:16,270][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.05745796114206314, 'train/grad_norm_pre_clip': 8.045530319213867, 'timing/data': 0.010600175708532333, 'timing/model': 1.441205658018589, 'learning_rate/qwen_vl_interface': 6.85613941218525e-06, 'learning_rate/action_model': 6.85613941218525e-05, 'epoch': 0.78}) +[2026-06-16 00:49:22,270][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.015143859200179577, 'train/grad_norm_pre_clip': 2.219299554824829, 'timing/data': 0.010851522907614708, 'timing/model': 1.4338358007371426, 'learning_rate/qwen_vl_interface': 6.8532730429821215e-06, 'learning_rate/action_model': 6.853273042982122e-05, 'epoch': 0.78}) +[2026-06-16 00:49:28,077][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.10055582225322723, 'train/grad_norm_pre_clip': 6.419556140899658, 'timing/data': 0.011173328384757042, 'timing/model': 1.4111513122916222, 'learning_rate/qwen_vl_interface': 6.850406014734005e-06, 'learning_rate/action_model': 6.850406014734005e-05, 'epoch': 0.78}) +[2026-06-16 00:49:34,075][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.03605005890130997, 'train/grad_norm_pre_clip': 2.4576303958892822, 'timing/data': 0.01259375549852848, 'timing/model': 1.700671585276723, 'learning_rate/qwen_vl_interface': 6.8475383286194295e-06, 'learning_rate/action_model': 6.84753832861943e-05, 'epoch': 0.78}) +[2026-06-16 00:49:39,942][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.017357952892780304, 'train/grad_norm_pre_clip': 2.785733222961426, 'timing/data': 0.012814553454518318, 'timing/model': 1.4692713767290115, 'learning_rate/qwen_vl_interface': 6.844669985817192e-06, 'learning_rate/action_model': 6.844669985817192e-05, 'epoch': 0.78}) +[2026-06-16 00:49:46,009][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.09544412046670914, 'train/grad_norm_pre_clip': 6.3048601150512695, 'timing/data': 0.012461764737963676, 'timing/model': 1.5304462611675262, 'learning_rate/qwen_vl_interface': 6.8418009875063594e-06, 'learning_rate/action_model': 6.84180098750636e-05, 'epoch': 0.78}) +[2026-06-16 00:49:51,999][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.19221912324428558, 'train/grad_norm_pre_clip': 13.69438648223877, 'timing/data': 0.0129226204007864, 'timing/model': 1.5208864286541939, 'learning_rate/qwen_vl_interface': 6.838931334866267e-06, 'learning_rate/action_model': 6.838931334866266e-05, 'epoch': 0.78}) +[2026-06-16 00:49:58,127][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.11910787224769592, 'train/grad_norm_pre_clip': 7.364483833312988, 'timing/data': 0.013276981189846992, 'timing/model': 1.470988156273961, 'learning_rate/qwen_vl_interface': 6.836061029076523e-06, 'learning_rate/action_model': 6.836061029076522e-05, 'epoch': 0.78}) +[2026-06-16 00:50:04,141][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.07788509875535965, 'train/grad_norm_pre_clip': 8.631217956542969, 'timing/data': 0.011851916089653969, 'timing/model': 1.4707716014236212, 'learning_rate/qwen_vl_interface': 6.833190071317e-06, 'learning_rate/action_model': 6.833190071316999e-05, 'epoch': 0.78}) +[2026-06-16 00:50:10,392][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.10033179819583893, 'train/grad_norm_pre_clip': 12.806647300720215, 'timing/data': 0.011027220636606216, 'timing/model': 1.7684281952679157, 'learning_rate/qwen_vl_interface': 6.830318462767841e-06, 'learning_rate/action_model': 6.830318462767841e-05, 'epoch': 0.78}) +[2026-06-16 00:50:16,458][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.10460110753774643, 'train/grad_norm_pre_clip': 5.775740146636963, 'timing/data': 0.01295069046318531, 'timing/model': 1.519418090581894, 'learning_rate/qwen_vl_interface': 6.827446204609456e-06, 'learning_rate/action_model': 6.827446204609456e-05, 'epoch': 0.78}) +[2026-06-16 00:50:22,878][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.1390085369348526, 'train/grad_norm_pre_clip': 12.391263961791992, 'timing/data': 0.014853132888674736, 'timing/model': 1.5528760720044374, 'learning_rate/qwen_vl_interface': 6.824573298022522e-06, 'learning_rate/action_model': 6.824573298022522e-05, 'epoch': 0.78}) +[2026-06-16 00:50:29,196][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.09058137238025665, 'train/grad_norm_pre_clip': 8.216850280761719, 'timing/data': 0.01591828465461731, 'timing/model': 1.5586817488074303, 'learning_rate/qwen_vl_interface': 6.8216997441879805e-06, 'learning_rate/action_model': 6.821699744187981e-05, 'epoch': 0.78}) +[2026-06-16 00:50:35,699][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.06867274641990662, 'train/grad_norm_pre_clip': 8.86113166809082, 'timing/data': 0.013930927962064743, 'timing/model': 1.529748035594821, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 0.78}) +[2026-06-16 00:50:42,018][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.015840502455830574, 'train/grad_norm_pre_clip': 2.462432861328125, 'timing/data': 0.012058146297931671, 'timing/model': 1.4690297339111567, 'learning_rate/qwen_vl_interface': 6.815950699501186e-06, 'learning_rate/action_model': 6.815950699501186e-05, 'epoch': 0.78}) +[2026-06-16 00:50:48,168][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.03022933006286621, 'train/grad_norm_pre_clip': 2.9779324531555176, 'timing/data': 0.012524593621492386, 'timing/model': 1.7756479531526566, 'learning_rate/qwen_vl_interface': 6.813075211012148e-06, 'learning_rate/action_model': 6.813075211012147e-05, 'epoch': 0.78}) +[2026-06-16 00:50:54,129][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.3214043378829956, 'train/grad_norm_pre_clip': 12.568553924560547, 'timing/data': 0.01165623590350151, 'timing/model': 1.5282582379877567, 'learning_rate/qwen_vl_interface': 6.81019908000193e-06, 'learning_rate/action_model': 6.81019908000193e-05, 'epoch': 0.78}) +[2026-06-16 00:51:00,228][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.01145844254642725, 'train/grad_norm_pre_clip': 1.0575907230377197, 'timing/data': 0.011621875688433647, 'timing/model': 1.5172489993274212, 'learning_rate/qwen_vl_interface': 6.807322307652806e-06, 'learning_rate/action_model': 6.807322307652806e-05, 'epoch': 0.79}) +[2026-06-16 00:51:06,330][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.01896250806748867, 'train/grad_norm_pre_clip': 6.029750823974609, 'timing/data': 0.011952711269259453, 'timing/model': 1.4847912788391113, 'learning_rate/qwen_vl_interface': 6.804444895147306e-06, 'learning_rate/action_model': 6.804444895147306e-05, 'epoch': 0.79}) +[2026-06-16 00:51:12,447][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.04926840215921402, 'train/grad_norm_pre_clip': 5.442132472991943, 'timing/data': 0.012393780052661896, 'timing/model': 1.472902376204729, 'learning_rate/qwen_vl_interface': 6.8015668436682255e-06, 'learning_rate/action_model': 6.801566843668226e-05, 'epoch': 0.79}) +[2026-06-16 00:51:18,481][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.23802076280117035, 'train/grad_norm_pre_clip': 7.000408172607422, 'timing/data': 0.01033770851790905, 'timing/model': 1.4110390935093164, 'learning_rate/qwen_vl_interface': 6.7986881543986245e-06, 'learning_rate/action_model': 6.798688154398624e-05, 'epoch': 0.79}) +[2026-06-16 00:51:24,451][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.10807360708713531, 'train/grad_norm_pre_clip': 13.19875431060791, 'timing/data': 0.012065257877111435, 'timing/model': 1.7107947710901499, 'learning_rate/qwen_vl_interface': 6.795808828521822e-06, 'learning_rate/action_model': 6.795808828521821e-05, 'epoch': 0.79}) +[2026-06-16 00:51:30,330][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.09421917796134949, 'train/grad_norm_pre_clip': 10.071063041687012, 'timing/data': 0.012235786765813828, 'timing/model': 1.5444461796432734, 'learning_rate/qwen_vl_interface': 6.792928867221401e-06, 'learning_rate/action_model': 6.792928867221401e-05, 'epoch': 0.79}) +[2026-06-16 00:51:36,436][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.08325996994972229, 'train/grad_norm_pre_clip': 12.119054794311523, 'timing/data': 0.012537308037281036, 'timing/model': 1.5510819479823112, 'learning_rate/qwen_vl_interface': 6.790048271681202e-06, 'learning_rate/action_model': 6.790048271681202e-05, 'epoch': 0.79}) +[2026-06-16 00:51:42,445][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.07253046333789825, 'train/grad_norm_pre_clip': 12.947076797485352, 'timing/data': 0.011484364047646523, 'timing/model': 1.4696197640150785, 'learning_rate/qwen_vl_interface': 6.7871670430853335e-06, 'learning_rate/action_model': 6.787167043085333e-05, 'epoch': 0.79}) +[2026-06-16 00:51:48,537][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.032165899872779846, 'train/grad_norm_pre_clip': 10.983880043029785, 'timing/data': 0.012484760954976082, 'timing/model': 1.4680459797382355, 'learning_rate/qwen_vl_interface': 6.784285182618156e-06, 'learning_rate/action_model': 6.784285182618156e-05, 'epoch': 0.79}) +[2026-06-16 00:51:54,631][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.3111351430416107, 'train/grad_norm_pre_clip': 19.0911922454834, 'timing/data': 0.011574482545256615, 'timing/model': 1.50237450376153, 'learning_rate/qwen_vl_interface': 6.781402691464297e-06, 'learning_rate/action_model': 6.781402691464297e-05, 'epoch': 0.79}) +[2026-06-16 00:52:00,790][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.010717264376580715, 'train/grad_norm_pre_clip': 1.454663872718811, 'timing/data': 0.012281225994229317, 'timing/model': 1.80219123698771, 'learning_rate/qwen_vl_interface': 6.778519570808636e-06, 'learning_rate/action_model': 6.778519570808636e-05, 'epoch': 0.79}) +[2026-06-16 00:52:06,733][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.1032203659415245, 'train/grad_norm_pre_clip': 7.368175983428955, 'timing/data': 0.01141277328133583, 'timing/model': 1.5519065987318754, 'learning_rate/qwen_vl_interface': 6.775635821836317e-06, 'learning_rate/action_model': 6.775635821836317e-05, 'epoch': 0.79}) +[2026-06-16 00:52:12,905][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.03586629033088684, 'train/grad_norm_pre_clip': 17.73607063293457, 'timing/data': 0.01532403938472271, 'timing/model': 1.593376036733389, 'learning_rate/qwen_vl_interface': 6.77275144573274e-06, 'learning_rate/action_model': 6.77275144573274e-05, 'epoch': 0.79}) +[2026-06-16 00:52:19,496][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.0237586572766304, 'train/grad_norm_pre_clip': 5.638242721557617, 'timing/data': 0.016204385086894035, 'timing/model': 1.6210713181644678, 'learning_rate/qwen_vl_interface': 6.769866443683564e-06, 'learning_rate/action_model': 6.769866443683563e-05, 'epoch': 0.79}) +[2026-06-16 00:52:25,767][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.22061218321323395, 'train/grad_norm_pre_clip': 10.423718452453613, 'timing/data': 0.012609859928488731, 'timing/model': 1.5728011559695005, 'learning_rate/qwen_vl_interface': 6.766980816874701e-06, 'learning_rate/action_model': 6.7669808168747e-05, 'epoch': 0.79}) +[2026-06-16 00:52:31,732][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.21290460228919983, 'train/grad_norm_pre_clip': 9.211047172546387, 'timing/data': 0.01007305458188057, 'timing/model': 1.4511158224195242, 'learning_rate/qwen_vl_interface': 6.764094566492327e-06, 'learning_rate/action_model': 6.764094566492328e-05, 'epoch': 0.79}) +[2026-06-16 00:52:37,599][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.034230709075927734, 'train/grad_norm_pre_clip': 10.915435791015625, 'timing/data': 0.009956678375601768, 'timing/model': 1.6563250999897718, 'learning_rate/qwen_vl_interface': 6.7612076937228685e-06, 'learning_rate/action_model': 6.761207693722869e-05, 'epoch': 0.79}) +[2026-06-16 00:52:43,302][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.08889275789260864, 'train/grad_norm_pre_clip': 17.381996154785156, 'timing/data': 0.011102808639407158, 'timing/model': 1.444794375449419, 'learning_rate/qwen_vl_interface': 6.7583201997530126e-06, 'learning_rate/action_model': 6.758320199753012e-05, 'epoch': 0.79}) +[2026-06-16 00:52:49,212][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.29554274678230286, 'train/grad_norm_pre_clip': 15.419843673706055, 'timing/data': 0.010873297229409218, 'timing/model': 1.449391731992364, 'learning_rate/qwen_vl_interface': 6.755432085769693e-06, 'learning_rate/action_model': 6.755432085769694e-05, 'epoch': 0.79}) +[2026-06-16 00:52:55,214][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.43207240104675293, 'train/grad_norm_pre_clip': 31.534379959106445, 'timing/data': 0.010179964825510979, 'timing/model': 1.457585772499442, 'learning_rate/qwen_vl_interface': 6.75254335296011e-06, 'learning_rate/action_model': 6.75254335296011e-05, 'epoch': 0.79}) +[2026-06-16 00:53:01,115][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.10983531922101974, 'train/grad_norm_pre_clip': 18.156936645507812, 'timing/data': 0.01099274680018425, 'timing/model': 1.4511509034782648, 'learning_rate/qwen_vl_interface': 6.749654002511712e-06, 'learning_rate/action_model': 6.749654002511712e-05, 'epoch': 0.79}) +[2026-06-16 00:53:07,286][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.13515998423099518, 'train/grad_norm_pre_clip': 2.8589744567871094, 'timing/data': 0.01256895437836647, 'timing/model': 1.5327200051397085, 'learning_rate/qwen_vl_interface': 6.7467640356122e-06, 'learning_rate/action_model': 6.746764035612199e-05, 'epoch': 0.79}) +[2026-06-16 00:53:13,628][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.3108338713645935, 'train/grad_norm_pre_clip': 17.2603816986084, 'timing/data': 0.013563243672251701, 'timing/model': 1.790558248758316, 'learning_rate/qwen_vl_interface': 6.74387345344953e-06, 'learning_rate/action_model': 6.74387345344953e-05, 'epoch': 0.79}) +[2026-06-16 00:53:19,662][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.04897978901863098, 'train/grad_norm_pre_clip': 5.825046539306641, 'timing/data': 0.012746594846248627, 'timing/model': 1.5552953705191612, 'learning_rate/qwen_vl_interface': 6.740982257211911e-06, 'learning_rate/action_model': 6.740982257211911e-05, 'epoch': 0.79}) +[2026-06-16 00:53:25,965][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.23836226761341095, 'train/grad_norm_pre_clip': 11.42842960357666, 'timing/data': 0.012137945741415024, 'timing/model': 1.554005166515708, 'learning_rate/qwen_vl_interface': 6.738090448087806e-06, 'learning_rate/action_model': 6.738090448087806e-05, 'epoch': 0.79}) +[2026-06-16 00:53:32,166][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.023557886481285095, 'train/grad_norm_pre_clip': 9.991181373596191, 'timing/data': 0.0122743621468544, 'timing/model': 1.5602113604545593, 'learning_rate/qwen_vl_interface': 6.735198027265928e-06, 'learning_rate/action_model': 6.735198027265928e-05, 'epoch': 0.79}) +[2026-06-16 00:53:38,452][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.12731467187404633, 'train/grad_norm_pre_clip': 10.98238754272461, 'timing/data': 0.01270870491862297, 'timing/model': 1.5481182392686605, 'learning_rate/qwen_vl_interface': 6.7323049959352424e-06, 'learning_rate/action_model': 6.732304995935243e-05, 'epoch': 0.8}) +[2026-06-16 00:53:44,671][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.07912202179431915, 'train/grad_norm_pre_clip': 15.098998069763184, 'timing/data': 0.011945391073822975, 'timing/model': 1.5053844787180424, 'learning_rate/qwen_vl_interface': 6.729411355284962e-06, 'learning_rate/action_model': 6.729411355284963e-05, 'epoch': 0.8}) +[2026-06-16 00:53:50,880][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.1413470208644867, 'train/grad_norm_pre_clip': 9.52760124206543, 'timing/data': 0.013377778232097626, 'timing/model': 1.7679143697023392, 'learning_rate/qwen_vl_interface': 6.726517106504559e-06, 'learning_rate/action_model': 6.726517106504558e-05, 'epoch': 0.8}) +[2026-06-16 00:53:57,111][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.3061155378818512, 'train/grad_norm_pre_clip': 29.3387508392334, 'timing/data': 0.011959483847022057, 'timing/model': 1.7225413173437119, 'learning_rate/qwen_vl_interface': 6.723622250783745e-06, 'learning_rate/action_model': 6.723622250783744e-05, 'epoch': 0.8}) +[2026-06-16 00:54:03,149][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.11245281249284744, 'train/grad_norm_pre_clip': 8.816569328308105, 'timing/data': 0.02817343920469284, 'timing/model': 1.5512789078056812, 'learning_rate/qwen_vl_interface': 6.720726789312487e-06, 'learning_rate/action_model': 6.720726789312487e-05, 'epoch': 0.8}) +[2026-06-16 00:54:09,428][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.09176451712846756, 'train/grad_norm_pre_clip': 9.969963073730469, 'timing/data': 0.017608076333999634, 'timing/model': 1.581450069323182, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.8}) +[2026-06-16 00:54:15,850][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.05195022001862526, 'train/grad_norm_pre_clip': 4.165393352508545, 'timing/data': 0.012369703501462936, 'timing/model': 1.5297148786485195, 'learning_rate/qwen_vl_interface': 6.714934053879749e-06, 'learning_rate/action_model': 6.71493405387975e-05, 'epoch': 0.8}) +[2026-06-16 00:54:22,125][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.05328356474637985, 'train/grad_norm_pre_clip': 4.496798038482666, 'timing/data': 0.013009889051318169, 'timing/model': 1.460347844287753, 'learning_rate/qwen_vl_interface': 6.712036782299442e-06, 'learning_rate/action_model': 6.712036782299443e-05, 'epoch': 0.8}) +[2026-06-16 00:54:28,466][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.1113172397017479, 'train/grad_norm_pre_clip': 8.68098258972168, 'timing/data': 0.01718888059258461, 'timing/model': 1.8123184405267239, 'learning_rate/qwen_vl_interface': 6.709138909731041e-06, 'learning_rate/action_model': 6.709138909731041e-05, 'epoch': 0.8}) +[2026-06-16 00:54:34,688][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.19929111003875732, 'train/grad_norm_pre_clip': 11.774189949035645, 'timing/data': 0.014282476156949997, 'timing/model': 1.54574828222394, 'learning_rate/qwen_vl_interface': 6.706240437365751e-06, 'learning_rate/action_model': 6.70624043736575e-05, 'epoch': 0.8}) +[2026-06-16 00:54:41,063][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.07905904203653336, 'train/grad_norm_pre_clip': 3.7260489463806152, 'timing/data': 0.011490393429994583, 'timing/model': 1.54600715264678, 'learning_rate/qwen_vl_interface': 6.703341366395023e-06, 'learning_rate/action_model': 6.703341366395023e-05, 'epoch': 0.8}) +[2026-06-16 00:54:47,146][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.11637280881404877, 'train/grad_norm_pre_clip': 10.314042091369629, 'timing/data': 0.01287633553147316, 'timing/model': 1.4704143106937408, 'learning_rate/qwen_vl_interface': 6.700441698010558e-06, 'learning_rate/action_model': 6.700441698010558e-05, 'epoch': 0.8}) +[2026-06-16 00:54:53,225][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.08835551887750626, 'train/grad_norm_pre_clip': 6.553226470947266, 'timing/data': 0.011794252321124077, 'timing/model': 1.4725642688572407, 'learning_rate/qwen_vl_interface': 6.697541433404299e-06, 'learning_rate/action_model': 6.697541433404299e-05, 'epoch': 0.8}) +[2026-06-16 00:54:59,234][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.021278994157910347, 'train/grad_norm_pre_clip': 1.4953616857528687, 'timing/data': 0.011453879997134209, 'timing/model': 1.5489535126835108, 'learning_rate/qwen_vl_interface': 6.6946405737684325e-06, 'learning_rate/action_model': 6.694640573768433e-05, 'epoch': 0.8}) +[2026-06-16 00:55:05,251][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.13362707197666168, 'train/grad_norm_pre_clip': 4.183146953582764, 'timing/data': 0.012839125469326973, 'timing/model': 1.6729288790374994, 'learning_rate/qwen_vl_interface': 6.691739120295397e-06, 'learning_rate/action_model': 6.691739120295397e-05, 'epoch': 0.8}) +[2026-06-16 00:55:11,122][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.052541978657245636, 'train/grad_norm_pre_clip': 9.854625701904297, 'timing/data': 0.013290677219629288, 'timing/model': 1.5402392745018005, 'learning_rate/qwen_vl_interface': 6.688837074177868e-06, 'learning_rate/action_model': 6.688837074177868e-05, 'epoch': 0.8}) +[2026-06-16 00:55:17,100][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.23107695579528809, 'train/grad_norm_pre_clip': 9.080381393432617, 'timing/data': 0.009964095428586006, 'timing/model': 1.416236149147153, 'learning_rate/qwen_vl_interface': 6.68593443660877e-06, 'learning_rate/action_model': 6.685934436608769e-05, 'epoch': 0.8}) +[2026-06-16 00:55:22,977][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.0964738056063652, 'train/grad_norm_pre_clip': 6.095695495605469, 'timing/data': 0.010351600125432014, 'timing/model': 1.4551192801445723, 'learning_rate/qwen_vl_interface': 6.6830312087812625e-06, 'learning_rate/action_model': 6.683031208781262e-05, 'epoch': 0.8}) +[2026-06-16 00:55:28,862][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.24886517226696014, 'train/grad_norm_pre_clip': 12.521061897277832, 'timing/data': 0.010277122259140015, 'timing/model': 1.435012487694621, 'learning_rate/qwen_vl_interface': 6.680127391888754e-06, 'learning_rate/action_model': 6.680127391888754e-05, 'epoch': 0.8}) +[2026-06-16 00:55:34,662][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.026581725105643272, 'train/grad_norm_pre_clip': 1.7849689722061157, 'timing/data': 0.01042107678949833, 'timing/model': 1.452004723250866, 'learning_rate/qwen_vl_interface': 6.677222987124897e-06, 'learning_rate/action_model': 6.677222987124897e-05, 'epoch': 0.8}) +[2026-06-16 00:55:40,419][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.015928540378808975, 'train/grad_norm_pre_clip': 0.9075528383255005, 'timing/data': 0.01073257066309452, 'timing/model': 1.5933722406625748, 'learning_rate/qwen_vl_interface': 6.67431799568358e-06, 'learning_rate/action_model': 6.67431799568358e-05, 'epoch': 0.8}) +[2026-06-16 00:55:46,076][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.03356831148266792, 'train/grad_norm_pre_clip': 3.1487321853637695, 'timing/data': 0.00999012216925621, 'timing/model': 1.4596764240413904, 'learning_rate/qwen_vl_interface': 6.671412418758939e-06, 'learning_rate/action_model': 6.671412418758938e-05, 'epoch': 0.8}) +[2026-06-16 00:55:51,965][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.12477539479732513, 'train/grad_norm_pre_clip': 14.829482078552246, 'timing/data': 0.010762728750705719, 'timing/model': 1.4495060928165913, 'learning_rate/qwen_vl_interface': 6.668506257545341e-06, 'learning_rate/action_model': 6.668506257545341e-05, 'epoch': 0.8}) +[2026-06-16 00:55:57,668][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.06622271239757538, 'train/grad_norm_pre_clip': 8.551271438598633, 'timing/data': 0.010711735114455223, 'timing/model': 1.4247246608138084, 'learning_rate/qwen_vl_interface': 6.665599513237403e-06, 'learning_rate/action_model': 6.665599513237402e-05, 'epoch': 0.8}) +[2026-06-16 00:56:03,471][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.015482299961149693, 'train/grad_norm_pre_clip': 1.2833155393600464, 'timing/data': 0.010721204802393913, 'timing/model': 1.4415284395217896, 'learning_rate/qwen_vl_interface': 6.662692187029978e-06, 'learning_rate/action_model': 6.662692187029978e-05, 'epoch': 0.8}) +[2026-06-16 00:56:09,269][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.104183129966259, 'train/grad_norm_pre_clip': 5.479187488555908, 'timing/data': 0.010852200910449028, 'timing/model': 1.6518559250980616, 'learning_rate/qwen_vl_interface': 6.659784280118156e-06, 'learning_rate/action_model': 6.659784280118156e-05, 'epoch': 0.8}) +[2026-06-16 00:56:14,869][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.27431705594062805, 'train/grad_norm_pre_clip': 19.666601181030273, 'timing/data': 0.009816082194447517, 'timing/model': 1.4528869185596704, 'learning_rate/qwen_vl_interface': 6.65687579369727e-06, 'learning_rate/action_model': 6.656875793697271e-05, 'epoch': 0.81}) +[2026-06-16 00:56:20,671][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.056489959359169006, 'train/grad_norm_pre_clip': 9.68222427368164, 'timing/data': 0.010472847148776054, 'timing/model': 1.4405237697064877, 'learning_rate/qwen_vl_interface': 6.653966728962885e-06, 'learning_rate/action_model': 6.653966728962885e-05, 'epoch': 0.81}) +[2026-06-16 00:56:26,367][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.03639739751815796, 'train/grad_norm_pre_clip': 8.213476181030273, 'timing/data': 0.010178910568356514, 'timing/model': 1.4425531756132841, 'learning_rate/qwen_vl_interface': 6.651057087110813e-06, 'learning_rate/action_model': 6.651057087110813e-05, 'epoch': 0.81}) +[2026-06-16 00:56:32,073][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.054905690252780914, 'train/grad_norm_pre_clip': 9.217033386230469, 'timing/data': 0.009962543845176697, 'timing/model': 1.455057317391038, 'learning_rate/qwen_vl_interface': 6.6481468693370924e-06, 'learning_rate/action_model': 6.648146869337091e-05, 'epoch': 0.81}) +[2026-06-16 00:56:37,886][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.04374014958739281, 'train/grad_norm_pre_clip': 7.027448654174805, 'timing/data': 0.010768415406346321, 'timing/model': 1.4118153601884842, 'learning_rate/qwen_vl_interface': 6.645236076838009e-06, 'learning_rate/action_model': 6.645236076838009e-05, 'epoch': 0.81}) +[2026-06-16 00:56:43,765][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.3031267523765564, 'train/grad_norm_pre_clip': 22.646833419799805, 'timing/data': 0.010059809312224388, 'timing/model': 1.6368402652442455, 'learning_rate/qwen_vl_interface': 6.642324710810075e-06, 'learning_rate/action_model': 6.642324710810076e-05, 'epoch': 0.81}) +[2026-06-16 00:56:49,375][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.0456324964761734, 'train/grad_norm_pre_clip': 10.825801849365234, 'timing/data': 0.009843580424785614, 'timing/model': 1.4482901748269796, 'learning_rate/qwen_vl_interface': 6.639412772450047e-06, 'learning_rate/action_model': 6.639412772450047e-05, 'epoch': 0.81}) +[2026-06-16 00:56:55,281][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.1155262440443039, 'train/grad_norm_pre_clip': 12.606959342956543, 'timing/data': 0.009768497198820114, 'timing/model': 1.4674972165375948, 'learning_rate/qwen_vl_interface': 6.6365002629549104e-06, 'learning_rate/action_model': 6.63650026295491e-05, 'epoch': 0.81}) +[2026-06-16 00:57:01,072][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.15347060561180115, 'train/grad_norm_pre_clip': 21.357728958129883, 'timing/data': 0.009696800261735916, 'timing/model': 1.4321675952523947, 'learning_rate/qwen_vl_interface': 6.633587183521887e-06, 'learning_rate/action_model': 6.633587183521887e-05, 'epoch': 0.81}) +[2026-06-16 00:57:06,870][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.28243136405944824, 'train/grad_norm_pre_clip': 20.315610885620117, 'timing/data': 0.00972278043627739, 'timing/model': 1.4589866064488888, 'learning_rate/qwen_vl_interface': 6.6306735353484365e-06, 'learning_rate/action_model': 6.630673535348436e-05, 'epoch': 0.81}) +[2026-06-16 00:57:12,582][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.05555688962340355, 'train/grad_norm_pre_clip': 26.60942268371582, 'timing/data': 0.010105280205607414, 'timing/model': 1.4114404153078794, 'learning_rate/qwen_vl_interface': 6.627759319632246e-06, 'learning_rate/action_model': 6.627759319632247e-05, 'epoch': 0.81}) +[2026-06-16 00:57:18,480][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.13368608057498932, 'train/grad_norm_pre_clip': 18.391565322875977, 'timing/data': 0.01020619086921215, 'timing/model': 1.6430654730647802, 'learning_rate/qwen_vl_interface': 6.624844537571245e-06, 'learning_rate/action_model': 6.624844537571245e-05, 'epoch': 0.81}) +[2026-06-16 00:57:24,183][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.048551157116889954, 'train/grad_norm_pre_clip': 11.143020629882812, 'timing/data': 0.00950586423277855, 'timing/model': 1.4463018141686916, 'learning_rate/qwen_vl_interface': 6.621929190363585e-06, 'learning_rate/action_model': 6.621929190363585e-05, 'epoch': 0.81}) +[2026-06-16 00:57:30,530][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.15120479464530945, 'train/grad_norm_pre_clip': 15.88591194152832, 'timing/data': 0.01190781220793724, 'timing/model': 1.4995120149105787, 'learning_rate/qwen_vl_interface': 6.619013279207657e-06, 'learning_rate/action_model': 6.619013279207656e-05, 'epoch': 0.81}) +[2026-06-16 00:57:36,409][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.020595107227563858, 'train/grad_norm_pre_clip': 3.7874066829681396, 'timing/data': 0.009958511218428612, 'timing/model': 1.4492833446711302, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 0.81}) +[2026-06-16 00:57:42,270][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.0616464838385582, 'train/grad_norm_pre_clip': 13.155248641967773, 'timing/data': 0.00994274951517582, 'timing/model': 1.4441994596272707, 'learning_rate/qwen_vl_interface': 6.6131797698457135e-06, 'learning_rate/action_model': 6.613179769845713e-05, 'epoch': 0.81}) +[2026-06-16 00:57:48,065][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.011803898960351944, 'train/grad_norm_pre_clip': 3.1905148029327393, 'timing/data': 0.010385250672698021, 'timing/model': 1.4736797288060188, 'learning_rate/qwen_vl_interface': 6.610262174037633e-06, 'learning_rate/action_model': 6.610262174037633e-05, 'epoch': 0.81}) +[2026-06-16 00:57:53,873][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.2158745527267456, 'train/grad_norm_pre_clip': 22.209260940551758, 'timing/data': 0.010313445702195168, 'timing/model': 1.6444542445242405, 'learning_rate/qwen_vl_interface': 6.607344019077155e-06, 'learning_rate/action_model': 6.607344019077156e-05, 'epoch': 0.81}) +[2026-06-16 00:57:59,481][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.14916543662548065, 'train/grad_norm_pre_clip': 17.13884735107422, 'timing/data': 0.010909931734204292, 'timing/model': 1.45773646235466, 'learning_rate/qwen_vl_interface': 6.604425306163822e-06, 'learning_rate/action_model': 6.604425306163822e-05, 'epoch': 0.81}) +[2026-06-16 01:05:49,143][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.3276979327201843, 'train/grad_norm_pre_clip': 29.297468185424805, 'eval/loss': 0.06287051737308502, 'eval/flap_precision': 0.9247730220492867, 'eval/flap_recall': 0.9674355495251018, 'eval/flap_f1': 0.9456233421750664, 'eval/flap_rate_model': 0.21416666666666667, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9456233421750663, 'timing/data': 0.011301932856440544, 'timing/model': 1.4461898263543844, 'learning_rate/qwen_vl_interface': 6.6015060364974045e-06, 'learning_rate/action_model': 6.601506036497404e-05, 'epoch': 0.81}) +[2026-06-16 01:05:49,144][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2100_state +[2026-06-16 01:05:49,147][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 01:05:55,132][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2100_state/model.safetensors +[2026-06-16 01:06:09,262][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2100_state/optimizer.bin +[2026-06-16 01:06:09,262][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2100_state/sampler.bin +[2026-06-16 01:06:09,262][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2100_state/sampler_1.bin +[2026-06-16 01:06:09,263][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2100_state/random_states_0.pkl +[2026-06-16 01:06:09,351][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2100 +[2026-06-16 01:06:09,351][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 01:06:09,359][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 01:06:15,005][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.17096848785877228, 'train/grad_norm_pre_clip': 12.204948425292969, 'timing/data': 0.010952625423669815, 'timing/model': 1.4540666844695807, 'learning_rate/qwen_vl_interface': 6.598586211277907e-06, 'learning_rate/action_model': 6.598586211277907e-05, 'epoch': 0.81}) +[2026-06-16 01:06:20,912][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.03307672590017319, 'train/grad_norm_pre_clip': 6.6069159507751465, 'timing/data': 0.009957637637853622, 'timing/model': 1.4611301515251398, 'learning_rate/qwen_vl_interface': 6.59566583170556e-06, 'learning_rate/action_model': 6.59566583170556e-05, 'epoch': 0.81}) +[2026-06-16 01:06:27,163][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.09987081587314606, 'train/grad_norm_pre_clip': 11.921760559082031, 'timing/data': 0.012180544435977936, 'timing/model': 1.5504965744912624, 'learning_rate/qwen_vl_interface': 6.59274489898082e-06, 'learning_rate/action_model': 6.592744898980819e-05, 'epoch': 0.81}) +[2026-06-16 01:06:33,463][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.08214715868234634, 'train/grad_norm_pre_clip': 24.57614517211914, 'timing/data': 0.012730970978736877, 'timing/model': 1.5108630526810884, 'learning_rate/qwen_vl_interface': 6.5898234143043716e-06, 'learning_rate/action_model': 6.589823414304371e-05, 'epoch': 0.81}) +[2026-06-16 01:06:39,766][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.08241420984268188, 'train/grad_norm_pre_clip': 6.585299968719482, 'timing/data': 0.012679798528552055, 'timing/model': 1.5474495720118284, 'learning_rate/qwen_vl_interface': 6.5869013788771264e-06, 'learning_rate/action_model': 6.586901378877126e-05, 'epoch': 0.81}) +[2026-06-16 01:06:46,108][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.20854690670967102, 'train/grad_norm_pre_clip': 8.887931823730469, 'timing/data': 0.012830914929509163, 'timing/model': 1.788615208119154, 'learning_rate/qwen_vl_interface': 6.583978793900223e-06, 'learning_rate/action_model': 6.583978793900223e-05, 'epoch': 0.82}) +[2026-06-16 01:06:52,170][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.06978175044059753, 'train/grad_norm_pre_clip': 7.110709190368652, 'timing/data': 0.011475779116153717, 'timing/model': 1.5517710484564304, 'learning_rate/qwen_vl_interface': 6.581055660575028e-06, 'learning_rate/action_model': 6.581055660575027e-05, 'epoch': 0.82}) +[2026-06-16 01:06:58,475][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.2248043566942215, 'train/grad_norm_pre_clip': 24.742210388183594, 'timing/data': 0.011736087501049042, 'timing/model': 1.5560628529638052, 'learning_rate/qwen_vl_interface': 6.578131980103128e-06, 'learning_rate/action_model': 6.578131980103128e-05, 'epoch': 0.82}) +[2026-06-16 01:07:04,720][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.19400402903556824, 'train/grad_norm_pre_clip': 46.13678741455078, 'timing/data': 0.011153658851981163, 'timing/model': 1.4995430149137974, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 0.82}) +[2026-06-16 01:07:10,635][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.011743078008294106, 'train/grad_norm_pre_clip': 2.9935007095336914, 'timing/data': 0.009783700108528137, 'timing/model': 1.4587868247181177, 'learning_rate/qwen_vl_interface': 6.572282982526702e-06, 'learning_rate/action_model': 6.572282982526702e-05, 'epoch': 0.82}) +[2026-06-16 01:07:16,608][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.21633458137512207, 'train/grad_norm_pre_clip': 25.668832778930664, 'timing/data': 0.011350128799676895, 'timing/model': 1.5288788229227066, 'learning_rate/qwen_vl_interface': 6.5693576678264755e-06, 'learning_rate/action_model': 6.569357667826476e-05, 'epoch': 0.82}) +[2026-06-16 01:07:22,644][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.08935562521219254, 'train/grad_norm_pre_clip': 6.907886505126953, 'timing/data': 0.010600607842206955, 'timing/model': 1.754161387681961, 'learning_rate/qwen_vl_interface': 6.566431810788148e-06, 'learning_rate/action_model': 6.566431810788148e-05, 'epoch': 0.82}) +[2026-06-16 01:07:28,290][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.3117089867591858, 'train/grad_norm_pre_clip': 18.413883209228516, 'timing/data': 0.010665450245141983, 'timing/model': 1.4500135146081448, 'learning_rate/qwen_vl_interface': 6.5635054126144306e-06, 'learning_rate/action_model': 6.56350541261443e-05, 'epoch': 0.82}) +[2026-06-16 01:07:34,183][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.4812062978744507, 'train/grad_norm_pre_clip': 19.391250610351562, 'timing/data': 0.011150689795613289, 'timing/model': 1.4506063405424356, 'learning_rate/qwen_vl_interface': 6.560578474508254e-06, 'learning_rate/action_model': 6.560578474508253e-05, 'epoch': 0.82}) +[2026-06-16 01:07:39,989][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.25666457414627075, 'train/grad_norm_pre_clip': 9.3317232131958, 'timing/data': 0.010936805978417397, 'timing/model': 1.4559757970273495, 'learning_rate/qwen_vl_interface': 6.5576509976727685e-06, 'learning_rate/action_model': 6.557650997672769e-05, 'epoch': 0.82}) +[2026-06-16 01:07:46,249][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.015725016593933105, 'train/grad_norm_pre_clip': 2.9362878799438477, 'timing/data': 0.013244371861219406, 'timing/model': 1.5504490043967962, 'learning_rate/qwen_vl_interface': 6.554722983311354e-06, 'learning_rate/action_model': 6.554722983311355e-05, 'epoch': 0.82}) +[2026-06-16 01:07:52,441][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.11464429646730423, 'train/grad_norm_pre_clip': 10.555265426635742, 'timing/data': 0.014448940753936768, 'timing/model': 1.486392429098487, 'learning_rate/qwen_vl_interface': 6.551794432627605e-06, 'learning_rate/action_model': 6.551794432627605e-05, 'epoch': 0.82}) +[2026-06-16 01:07:58,363][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.16547738015651703, 'train/grad_norm_pre_clip': 24.594575881958008, 'timing/data': 0.01154927909374237, 'timing/model': 1.6893299147486687, 'learning_rate/qwen_vl_interface': 6.548865346825338e-06, 'learning_rate/action_model': 6.548865346825338e-05, 'epoch': 0.82}) +[2026-06-16 01:08:04,115][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.1416221708059311, 'train/grad_norm_pre_clip': 6.075078010559082, 'timing/data': 0.011780787259340286, 'timing/model': 1.4512709360569715, 'learning_rate/qwen_vl_interface': 6.545935727108589e-06, 'learning_rate/action_model': 6.545935727108588e-05, 'epoch': 0.82}) +[2026-06-16 01:08:10,016][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.20926550030708313, 'train/grad_norm_pre_clip': 22.115930557250977, 'timing/data': 0.011760219931602478, 'timing/model': 1.451593041419983, 'learning_rate/qwen_vl_interface': 6.543005574681613e-06, 'learning_rate/action_model': 6.543005574681613e-05, 'epoch': 0.82}) +[2026-06-16 01:08:15,919][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.20058228075504303, 'train/grad_norm_pre_clip': 9.408963203430176, 'timing/data': 0.010595068335533142, 'timing/model': 1.4525443520396948, 'learning_rate/qwen_vl_interface': 6.5400748907488886e-06, 'learning_rate/action_model': 6.540074890748888e-05, 'epoch': 0.82}) +[2026-06-16 01:08:21,882][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.06076844036579132, 'train/grad_norm_pre_clip': 4.806589603424072, 'timing/data': 0.010001523420214653, 'timing/model': 1.5146333165466785, 'learning_rate/qwen_vl_interface': 6.537143676515105e-06, 'learning_rate/action_model': 6.537143676515105e-05, 'epoch': 0.82}) +[2026-06-16 01:08:28,271][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.22474227845668793, 'train/grad_norm_pre_clip': 9.7927885055542, 'timing/data': 0.013523964211344719, 'timing/model': 1.5653552915900946, 'learning_rate/qwen_vl_interface': 6.534211933185176e-06, 'learning_rate/action_model': 6.534211933185176e-05, 'epoch': 0.82}) +[2026-06-16 01:08:34,573][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.09346476942300797, 'train/grad_norm_pre_clip': 5.876769542694092, 'timing/data': 0.01230895146727562, 'timing/model': 1.7804942559450865, 'learning_rate/qwen_vl_interface': 6.531279661964231e-06, 'learning_rate/action_model': 6.531279661964231e-05, 'epoch': 0.82}) +[2026-06-16 01:08:40,819][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.10465152561664581, 'train/grad_norm_pre_clip': 4.839321136474609, 'timing/data': 0.012426791712641716, 'timing/model': 1.7159260045737028, 'learning_rate/qwen_vl_interface': 6.528346864057616e-06, 'learning_rate/action_model': 6.528346864057615e-05, 'epoch': 0.82}) +[2026-06-16 01:08:46,866][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.1150549128651619, 'train/grad_norm_pre_clip': 12.567497253417969, 'timing/data': 0.012049449607729912, 'timing/model': 1.5456298124045134, 'learning_rate/qwen_vl_interface': 6.5254135406708906e-06, 'learning_rate/action_model': 6.52541354067089e-05, 'epoch': 0.82}) +[2026-06-16 01:08:53,151][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.10195005685091019, 'train/grad_norm_pre_clip': 5.327526569366455, 'timing/data': 0.011632159352302551, 'timing/model': 1.5307960864156485, 'learning_rate/qwen_vl_interface': 6.522479693009836e-06, 'learning_rate/action_model': 6.522479693009837e-05, 'epoch': 0.82}) +[2026-06-16 01:08:59,350][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.14924447238445282, 'train/grad_norm_pre_clip': 11.24545955657959, 'timing/data': 0.013403598219156265, 'timing/model': 1.5275069884955883, 'learning_rate/qwen_vl_interface': 6.519545322280446e-06, 'learning_rate/action_model': 6.519545322280446e-05, 'epoch': 0.82}) +[2026-06-16 01:09:05,294][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.0824086144566536, 'train/grad_norm_pre_clip': 6.360136985778809, 'timing/data': 0.010816741734743118, 'timing/model': 1.427587015554309, 'learning_rate/qwen_vl_interface': 6.5166104296889275e-06, 'learning_rate/action_model': 6.516610429688927e-05, 'epoch': 0.82}) +[2026-06-16 01:09:11,199][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.1603744626045227, 'train/grad_norm_pre_clip': 6.267747402191162, 'timing/data': 0.01161164790391922, 'timing/model': 1.6446928419172764, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.82}) +[2026-06-16 01:09:16,885][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.08869930356740952, 'train/grad_norm_pre_clip': 11.068435668945312, 'timing/data': 0.009931905195116997, 'timing/model': 1.4370849840342999, 'learning_rate/qwen_vl_interface': 6.510739083745422e-06, 'learning_rate/action_model': 6.510739083745422e-05, 'epoch': 0.82}) +[2026-06-16 01:09:22,784][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.14281471073627472, 'train/grad_norm_pre_clip': 7.074254512786865, 'timing/data': 0.0113496333360672, 'timing/model': 1.4474314395338297, 'learning_rate/qwen_vl_interface': 6.507802632806919e-06, 'learning_rate/action_model': 6.507802632806919e-05, 'epoch': 0.83}) +[2026-06-16 01:09:28,682][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.13917838037014008, 'train/grad_norm_pre_clip': 6.596592426300049, 'timing/data': 0.010332487523555756, 'timing/model': 1.4418278969824314, 'learning_rate/qwen_vl_interface': 6.5048656648332646e-06, 'learning_rate/action_model': 6.504865664833264e-05, 'epoch': 0.83}) +[2026-06-16 01:09:34,585][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.0729047954082489, 'train/grad_norm_pre_clip': 3.3463761806488037, 'timing/data': 0.011372042819857597, 'timing/model': 1.4184003248810768, 'learning_rate/qwen_vl_interface': 6.501928181031737e-06, 'learning_rate/action_model': 6.501928181031737e-05, 'epoch': 0.83}) +[2026-06-16 01:09:40,421][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.12939821183681488, 'train/grad_norm_pre_clip': 6.948318958282471, 'timing/data': 0.01078198291361332, 'timing/model': 1.4460860006511211, 'learning_rate/qwen_vl_interface': 6.498990182609823e-06, 'learning_rate/action_model': 6.498990182609823e-05, 'epoch': 0.83}) +[2026-06-16 01:09:46,627][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.07282641530036926, 'train/grad_norm_pre_clip': 3.028053045272827, 'timing/data': 0.011688176542520523, 'timing/model': 1.7253712322562933, 'learning_rate/qwen_vl_interface': 6.4960516707752255e-06, 'learning_rate/action_model': 6.496051670775226e-05, 'epoch': 0.83}) +[2026-06-16 01:09:52,643][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.1385008990764618, 'train/grad_norm_pre_clip': 5.687015533447266, 'timing/data': 0.013071384280920029, 'timing/model': 1.5564467385411263, 'learning_rate/qwen_vl_interface': 6.493112646735851e-06, 'learning_rate/action_model': 6.493112646735851e-05, 'epoch': 0.83}) +[2026-06-16 01:09:58,842][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.02180521748960018, 'train/grad_norm_pre_clip': 2.349935531616211, 'timing/data': 0.012614406645298004, 'timing/model': 1.560116345062852, 'learning_rate/qwen_vl_interface': 6.490173111699824e-06, 'learning_rate/action_model': 6.490173111699825e-05, 'epoch': 0.83}) +[2026-06-16 01:10:05,042][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.27575093507766724, 'train/grad_norm_pre_clip': 11.308110237121582, 'timing/data': 0.012149738147854805, 'timing/model': 1.5487469919025898, 'learning_rate/qwen_vl_interface': 6.487233066875479e-06, 'learning_rate/action_model': 6.487233066875479e-05, 'epoch': 0.83}) +[2026-06-16 01:10:11,247][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.12713375687599182, 'train/grad_norm_pre_clip': 6.679997444152832, 'timing/data': 0.012564817443490028, 'timing/model': 1.5098196268081665, 'learning_rate/qwen_vl_interface': 6.484292513471352e-06, 'learning_rate/action_model': 6.484292513471352e-05, 'epoch': 0.83}) +[2026-06-16 01:10:17,349][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.05626004934310913, 'train/grad_norm_pre_clip': 2.3625948429107666, 'timing/data': 0.01213851198554039, 'timing/model': 1.5013336036354303, 'learning_rate/qwen_vl_interface': 6.4813514526961975e-06, 'learning_rate/action_model': 6.481351452696198e-05, 'epoch': 0.83}) +[2026-06-16 01:10:23,434][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.057779502123594284, 'train/grad_norm_pre_clip': 4.482041835784912, 'timing/data': 0.011237449944019318, 'timing/model': 1.6500591523945332, 'learning_rate/qwen_vl_interface': 6.478409885758972e-06, 'learning_rate/action_model': 6.478409885758971e-05, 'epoch': 0.83}) +[2026-06-16 01:10:29,085][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.14405988156795502, 'train/grad_norm_pre_clip': 12.881567001342773, 'timing/data': 0.010672803968191147, 'timing/model': 1.4535998459905386, 'learning_rate/qwen_vl_interface': 6.475467813868844e-06, 'learning_rate/action_model': 6.475467813868844e-05, 'epoch': 0.83}) +[2026-06-16 01:10:34,986][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.12973764538764954, 'train/grad_norm_pre_clip': 7.586562633514404, 'timing/data': 0.010772915557026863, 'timing/model': 1.4533348958939314, 'learning_rate/qwen_vl_interface': 6.472525238235187e-06, 'learning_rate/action_model': 6.472525238235186e-05, 'epoch': 0.83}) +[2026-06-16 01:10:40,887][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.009228498674929142, 'train/grad_norm_pre_clip': 1.6114518642425537, 'timing/data': 0.009985420852899551, 'timing/model': 1.4528064765036106, 'learning_rate/qwen_vl_interface': 6.4695821600675834e-06, 'learning_rate/action_model': 6.469582160067583e-05, 'epoch': 0.83}) +[2026-06-16 01:10:46,784][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.06713906675577164, 'train/grad_norm_pre_clip': 4.6968231201171875, 'timing/data': 0.010174877941608429, 'timing/model': 1.4428764283657074, 'learning_rate/qwen_vl_interface': 6.466638580575824e-06, 'learning_rate/action_model': 6.466638580575824e-05, 'epoch': 0.83}) +[2026-06-16 01:10:52,690][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.39211905002593994, 'train/grad_norm_pre_clip': 14.014819145202637, 'timing/data': 0.010918457061052322, 'timing/model': 1.657621392980218, 'learning_rate/qwen_vl_interface': 6.463694500969898e-06, 'learning_rate/action_model': 6.463694500969898e-05, 'epoch': 0.83}) +[2026-06-16 01:10:58,395][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.04843902215361595, 'train/grad_norm_pre_clip': 7.053808689117432, 'timing/data': 0.010878609493374825, 'timing/model': 1.4401343911886215, 'learning_rate/qwen_vl_interface': 6.460749922460008e-06, 'learning_rate/action_model': 6.460749922460008e-05, 'epoch': 0.83}) +[2026-06-16 01:11:04,293][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.26812657713890076, 'train/grad_norm_pre_clip': 29.58918571472168, 'timing/data': 0.011330883949995041, 'timing/model': 1.4451180286705494, 'learning_rate/qwen_vl_interface': 6.457804846256559e-06, 'learning_rate/action_model': 6.457804846256559e-05, 'epoch': 0.83}) +[2026-06-16 01:11:10,295][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.030021309852600098, 'train/grad_norm_pre_clip': 4.217162132263184, 'timing/data': 0.014262823387980461, 'timing/model': 1.5493141431361437, 'learning_rate/qwen_vl_interface': 6.45485927357016e-06, 'learning_rate/action_model': 6.454859273570159e-05, 'epoch': 0.83}) +[2026-06-16 01:11:16,572][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.061030227690935135, 'train/grad_norm_pre_clip': 6.604347229003906, 'timing/data': 0.012141184881329536, 'timing/model': 1.5600127447396517, 'learning_rate/qwen_vl_interface': 6.451913205611626e-06, 'learning_rate/action_model': 6.451913205611626e-05, 'epoch': 0.83}) +[2026-06-16 01:11:22,871][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.11965484172105789, 'train/grad_norm_pre_clip': 11.375418663024902, 'timing/data': 0.012934554368257523, 'timing/model': 1.5521555356681347, 'learning_rate/qwen_vl_interface': 6.448966643591974e-06, 'learning_rate/action_model': 6.448966643591974e-05, 'epoch': 0.83}) +[2026-06-16 01:11:29,318][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.12282325327396393, 'train/grad_norm_pre_clip': 6.942805767059326, 'timing/data': 0.03241925872862339, 'timing/model': 1.767884312197566, 'learning_rate/qwen_vl_interface': 6.446019588722423e-06, 'learning_rate/action_model': 6.446019588722422e-05, 'epoch': 0.83}) +[2026-06-16 01:11:35,387][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.07740962505340576, 'train/grad_norm_pre_clip': 11.691001892089844, 'timing/data': 0.012405851855874062, 'timing/model': 1.5583387464284897, 'learning_rate/qwen_vl_interface': 6.443072042214398e-06, 'learning_rate/action_model': 6.443072042214397e-05, 'epoch': 0.83}) +[2026-06-16 01:11:41,681][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.01236063614487648, 'train/grad_norm_pre_clip': 1.8289525508880615, 'timing/data': 0.012393586337566376, 'timing/model': 1.5482295975089073, 'learning_rate/qwen_vl_interface': 6.440124005279521e-06, 'learning_rate/action_model': 6.440124005279521e-05, 'epoch': 0.83}) +[2026-06-16 01:11:48,288][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.031201744452118874, 'train/grad_norm_pre_clip': 2.735919952392578, 'timing/data': 0.013098880648612976, 'timing/model': 1.5569230541586876, 'learning_rate/qwen_vl_interface': 6.43717547912962e-06, 'learning_rate/action_model': 6.437175479129621e-05, 'epoch': 0.83}) +[2026-06-16 01:11:54,806][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.10717620700597763, 'train/grad_norm_pre_clip': 7.586886882781982, 'timing/data': 0.012886608019471169, 'timing/model': 1.677712732926011, 'learning_rate/qwen_vl_interface': 6.434226464976725e-06, 'learning_rate/action_model': 6.434226464976724e-05, 'epoch': 0.83}) +[2026-06-16 01:12:01,939][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.0655975341796875, 'train/grad_norm_pre_clip': 11.445988655090332, 'timing/data': 0.019138243049383163, 'timing/model': 1.7438839245587587, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 0.84}) +[2026-06-16 01:12:08,962][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.010860728099942207, 'train/grad_norm_pre_clip': 1.5172975063323975, 'timing/data': 0.01900869980454445, 'timing/model': 1.9940750189125538, 'learning_rate/qwen_vl_interface': 6.428326977511059e-06, 'learning_rate/action_model': 6.428326977511058e-05, 'epoch': 0.84}) +[2026-06-16 01:12:15,732][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.006322762928903103, 'train/grad_norm_pre_clip': 1.6317318677902222, 'timing/data': 0.018230093643069267, 'timing/model': 1.7207091059535742, 'learning_rate/qwen_vl_interface': 6.425376506623344e-06, 'learning_rate/action_model': 6.425376506623344e-05, 'epoch': 0.84}) +[2026-06-16 01:12:22,636][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.00390987703576684, 'train/grad_norm_pre_clip': 1.0400886535644531, 'timing/data': 0.018257856369018555, 'timing/model': 1.7498320266604424, 'learning_rate/qwen_vl_interface': 6.422425552582747e-06, 'learning_rate/action_model': 6.422425552582747e-05, 'epoch': 0.84}) +[2026-06-16 01:12:29,640][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.30969324707984924, 'train/grad_norm_pre_clip': 11.896856307983398, 'timing/data': 0.018972374498844147, 'timing/model': 1.7421980053186417, 'learning_rate/qwen_vl_interface': 6.41947411660229e-06, 'learning_rate/action_model': 6.41947411660229e-05, 'epoch': 0.84}) +[2026-06-16 01:12:35,948][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.005202450789511204, 'train/grad_norm_pre_clip': 1.6486343145370483, 'timing/data': 0.012902189046144485, 'timing/model': 1.4986140839755535, 'learning_rate/qwen_vl_interface': 6.4165221998951995e-06, 'learning_rate/action_model': 6.4165221998952e-05, 'epoch': 0.84}) +[2026-06-16 01:12:42,250][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.057190436869859695, 'train/grad_norm_pre_clip': 10.090608596801758, 'timing/data': 0.012000946328043938, 'timing/model': 1.5634528510272503, 'learning_rate/qwen_vl_interface': 6.413569803674896e-06, 'learning_rate/action_model': 6.413569803674896e-05, 'epoch': 0.84}) +[2026-06-16 01:12:48,513][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.07075172662734985, 'train/grad_norm_pre_clip': 16.04995346069336, 'timing/data': 0.012069590389728546, 'timing/model': 1.779687674716115, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 0.84}) +[2026-06-16 01:12:54,549][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.03075292520225048, 'train/grad_norm_pre_clip': 5.631943702697754, 'timing/data': 0.013120466843247414, 'timing/model': 1.530747577548027, 'learning_rate/qwen_vl_interface': 6.40766357754932e-06, 'learning_rate/action_model': 6.407663577549319e-05, 'epoch': 0.84}) +[2026-06-16 01:13:00,751][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.2759611904621124, 'train/grad_norm_pre_clip': 11.60709285736084, 'timing/data': 0.012597564607858658, 'timing/model': 1.5546280313283205, 'learning_rate/qwen_vl_interface': 6.404709750071875e-06, 'learning_rate/action_model': 6.404709750071874e-05, 'epoch': 0.84}) +[2026-06-16 01:13:06,947][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.07229924947023392, 'train/grad_norm_pre_clip': 11.34853744506836, 'timing/data': 0.011938374489545822, 'timing/model': 1.5267009884119034, 'learning_rate/qwen_vl_interface': 6.401755447936869e-06, 'learning_rate/action_model': 6.401755447936869e-05, 'epoch': 0.84}) +[2026-06-16 01:13:13,149][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.218875914812088, 'train/grad_norm_pre_clip': 16.439756393432617, 'timing/data': 0.012337187305092812, 'timing/model': 1.5167166087776423, 'learning_rate/qwen_vl_interface': 6.3988006723587036e-06, 'learning_rate/action_model': 6.398800672358703e-05, 'epoch': 0.84}) +[2026-06-16 01:13:19,450][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.15526621043682098, 'train/grad_norm_pre_clip': 13.743785858154297, 'timing/data': 0.011701157316565514, 'timing/model': 1.528514053672552, 'learning_rate/qwen_vl_interface': 6.395845424551975e-06, 'learning_rate/action_model': 6.395845424551975e-05, 'epoch': 0.84}) +[2026-06-16 01:13:25,685][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.09666673839092255, 'train/grad_norm_pre_clip': 11.9932279586792, 'timing/data': 0.012705961242318153, 'timing/model': 1.7503938674926758, 'learning_rate/qwen_vl_interface': 6.392889705731475e-06, 'learning_rate/action_model': 6.392889705731476e-05, 'epoch': 0.84}) +[2026-06-16 01:13:31,748][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.012611555866897106, 'train/grad_norm_pre_clip': 7.491371154785156, 'timing/data': 0.014032011851668358, 'timing/model': 1.5314095374196768, 'learning_rate/qwen_vl_interface': 6.389933517112188e-06, 'learning_rate/action_model': 6.389933517112187e-05, 'epoch': 0.84}) +[2026-06-16 01:13:38,047][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.052335698157548904, 'train/grad_norm_pre_clip': 15.16662883758545, 'timing/data': 0.013118354603648186, 'timing/model': 1.5544507838785648, 'learning_rate/qwen_vl_interface': 6.386976859909293e-06, 'learning_rate/action_model': 6.386976859909292e-05, 'epoch': 0.84}) +[2026-06-16 01:13:44,250][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.2776065766811371, 'train/grad_norm_pre_clip': 20.534032821655273, 'timing/data': 0.011942684650421143, 'timing/model': 1.5403979383409023, 'learning_rate/qwen_vl_interface': 6.384019735338156e-06, 'learning_rate/action_model': 6.384019735338156e-05, 'epoch': 0.84}) +[2026-06-16 01:13:50,446][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.5984463691711426, 'train/grad_norm_pre_clip': 38.62454605102539, 'timing/data': 0.012861216440796852, 'timing/model': 1.520049074664712, 'learning_rate/qwen_vl_interface': 6.381062144614342e-06, 'learning_rate/action_model': 6.381062144614342e-05, 'epoch': 0.84}) +[2026-06-16 01:13:56,749][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.08057812601327896, 'train/grad_norm_pre_clip': 12.563116073608398, 'timing/data': 0.012819565832614899, 'timing/model': 1.5620131995528936, 'learning_rate/qwen_vl_interface': 6.378104088953605e-06, 'learning_rate/action_model': 6.378104088953605e-05, 'epoch': 0.84}) +[2026-06-16 01:14:03,004][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.3221225440502167, 'train/grad_norm_pre_clip': 15.862269401550293, 'timing/data': 0.012636935338377953, 'timing/model': 1.8059760108590126, 'learning_rate/qwen_vl_interface': 6.375145569571889e-06, 'learning_rate/action_model': 6.375145569571889e-05, 'epoch': 0.84}) +[2026-06-16 01:14:09,055][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.20877528190612793, 'train/grad_norm_pre_clip': 19.7763671875, 'timing/data': 0.012643245980143547, 'timing/model': 1.5520148687064648, 'learning_rate/qwen_vl_interface': 6.372186587685329e-06, 'learning_rate/action_model': 6.372186587685329e-05, 'epoch': 0.84}) +[2026-06-16 01:14:15,350][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.2065490484237671, 'train/grad_norm_pre_clip': 10.905277252197266, 'timing/data': 0.012714631855487823, 'timing/model': 1.5508989449590445, 'learning_rate/qwen_vl_interface': 6.369227144510252e-06, 'learning_rate/action_model': 6.369227144510252e-05, 'epoch': 0.84}) +[2026-06-16 01:14:21,550][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.05528685823082924, 'train/grad_norm_pre_clip': 5.712629795074463, 'timing/data': 0.012635497376322746, 'timing/model': 1.5518873501569033, 'learning_rate/qwen_vl_interface': 6.3662672412631736e-06, 'learning_rate/action_model': 6.366267241263173e-05, 'epoch': 0.84}) +[2026-06-16 01:14:27,747][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.06595202535390854, 'train/grad_norm_pre_clip': 10.981732368469238, 'timing/data': 0.012402519583702087, 'timing/model': 1.5517089944332838, 'learning_rate/qwen_vl_interface': 6.3633068791607954e-06, 'learning_rate/action_model': 6.363306879160795e-05, 'epoch': 0.84}) +[2026-06-16 01:14:33,943][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.046551384031772614, 'train/grad_norm_pre_clip': 4.512518882751465, 'timing/data': 0.01202751137316227, 'timing/model': 1.4886299092322588, 'learning_rate/qwen_vl_interface': 6.360346059420014e-06, 'learning_rate/action_model': 6.360346059420014e-05, 'epoch': 0.84}) +[2026-06-16 01:14:39,792][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.07973767071962357, 'train/grad_norm_pre_clip': 7.120151519775391, 'timing/data': 0.009971888735890388, 'timing/model': 1.6530278734862804, 'learning_rate/qwen_vl_interface': 6.357384783257909e-06, 'learning_rate/action_model': 6.357384783257908e-05, 'epoch': 0.84}) +[2026-06-16 01:14:45,493][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.21500451862812042, 'train/grad_norm_pre_clip': 14.93298625946045, 'timing/data': 0.01095418818295002, 'timing/model': 1.452913798391819, 'learning_rate/qwen_vl_interface': 6.354423051891748e-06, 'learning_rate/action_model': 6.354423051891749e-05, 'epoch': 0.85}) +[2026-06-16 01:14:51,384][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.0188506618142128, 'train/grad_norm_pre_clip': 2.324361801147461, 'timing/data': 0.010715346783399582, 'timing/model': 1.44098568148911, 'learning_rate/qwen_vl_interface': 6.35146086653899e-06, 'learning_rate/action_model': 6.35146086653899e-05, 'epoch': 0.85}) +[2026-06-16 01:14:57,294][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.08482765406370163, 'train/grad_norm_pre_clip': 5.539858341217041, 'timing/data': 0.009929956868290901, 'timing/model': 1.4467803128063679, 'learning_rate/qwen_vl_interface': 6.348498228417273e-06, 'learning_rate/action_model': 6.348498228417273e-05, 'epoch': 0.85}) +[2026-06-16 01:15:03,198][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.08440171927213669, 'train/grad_norm_pre_clip': 16.178354263305664, 'timing/data': 0.010144896805286407, 'timing/model': 1.4572133403271437, 'learning_rate/qwen_vl_interface': 6.34553513874443e-06, 'learning_rate/action_model': 6.34553513874443e-05, 'epoch': 0.85}) +[2026-06-16 01:15:09,091][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.2503759562969208, 'train/grad_norm_pre_clip': 10.268607139587402, 'timing/data': 0.010278381407260895, 'timing/model': 1.4447229243814945, 'learning_rate/qwen_vl_interface': 6.342571598738473e-06, 'learning_rate/action_model': 6.342571598738472e-05, 'epoch': 0.85}) +[2026-06-16 01:15:14,985][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.01417857687920332, 'train/grad_norm_pre_clip': 2.289233684539795, 'timing/data': 0.009930325672030449, 'timing/model': 1.6462776903063059, 'learning_rate/qwen_vl_interface': 6.339607609617603e-06, 'learning_rate/action_model': 6.339607609617603e-05, 'epoch': 0.85}) +[2026-06-16 01:15:20,879][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.16050302982330322, 'train/grad_norm_pre_clip': 15.928568840026855, 'timing/data': 0.010479660704731941, 'timing/model': 1.6404042188078165, 'learning_rate/qwen_vl_interface': 6.336643172600202e-06, 'learning_rate/action_model': 6.336643172600202e-05, 'epoch': 0.85}) +[2026-06-16 01:15:26,593][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.028671547770500183, 'train/grad_norm_pre_clip': 5.343113422393799, 'timing/data': 0.009764105081558228, 'timing/model': 1.450822401791811, 'learning_rate/qwen_vl_interface': 6.333678288904838e-06, 'learning_rate/action_model': 6.333678288904837e-05, 'epoch': 0.85}) +[2026-06-16 01:15:32,650][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.025266533717513084, 'train/grad_norm_pre_clip': 5.649644374847412, 'timing/data': 0.013444008305668831, 'timing/model': 1.548322543501854, 'learning_rate/qwen_vl_interface': 6.330712959750264e-06, 'learning_rate/action_model': 6.330712959750263e-05, 'epoch': 0.85}) +[2026-06-16 01:15:38,842][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.08568461984395981, 'train/grad_norm_pre_clip': 5.417969226837158, 'timing/data': 0.011826897040009499, 'timing/model': 1.498759051784873, 'learning_rate/qwen_vl_interface': 6.3277471863554145e-06, 'learning_rate/action_model': 6.327747186355414e-05, 'epoch': 0.85}) +[2026-06-16 01:15:44,908][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.1405484676361084, 'train/grad_norm_pre_clip': 9.617170333862305, 'timing/data': 0.010737759992480278, 'timing/model': 1.437215767800808, 'learning_rate/qwen_vl_interface': 6.324780969939406e-06, 'learning_rate/action_model': 6.324780969939405e-05, 'epoch': 0.85}) +[2026-06-16 01:15:50,797][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.06397745013237, 'train/grad_norm_pre_clip': 18.9645938873291, 'timing/data': 0.009783374145627022, 'timing/model': 1.650373362004757, 'learning_rate/qwen_vl_interface': 6.32181431172154e-06, 'learning_rate/action_model': 6.32181431172154e-05, 'epoch': 0.85}) +[2026-06-16 01:15:56,501][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.09978656470775604, 'train/grad_norm_pre_clip': 9.058389663696289, 'timing/data': 0.011531040072441101, 'timing/model': 1.4398670215159655, 'learning_rate/qwen_vl_interface': 6.318847212921295e-06, 'learning_rate/action_model': 6.318847212921295e-05, 'epoch': 0.85}) +[2026-06-16 01:16:02,404][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.07444940507411957, 'train/grad_norm_pre_clip': 9.586095809936523, 'timing/data': 0.010070830583572388, 'timing/model': 1.4458117429167032, 'learning_rate/qwen_vl_interface': 6.3158796747583364e-06, 'learning_rate/action_model': 6.315879674758336e-05, 'epoch': 0.85}) +[2026-06-16 01:16:08,502][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.1179099753499031, 'train/grad_norm_pre_clip': 10.94459056854248, 'timing/data': 0.010201269760727882, 'timing/model': 1.4451936390250921, 'learning_rate/qwen_vl_interface': 6.312911698452503e-06, 'learning_rate/action_model': 6.312911698452502e-05, 'epoch': 0.85}) +[2026-06-16 01:16:14,486][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.17854899168014526, 'train/grad_norm_pre_clip': 15.106152534484863, 'timing/data': 0.010152401402592659, 'timing/model': 1.505676256492734, 'learning_rate/qwen_vl_interface': 6.309943285223824e-06, 'learning_rate/action_model': 6.309943285223824e-05, 'epoch': 0.85}) +[2026-06-16 01:24:15,637][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.05128104239702225, 'train/grad_norm_pre_clip': 5.902050495147705, 'eval/loss': 0.08880278468132019, 'eval/flap_precision': 0.9856230031948882, 'eval/flap_recall': 0.8371777476255088, 'eval/flap_f1': 0.9053558327219369, 'eval/flap_rate_model': 0.1738888888888889, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.905355832721937, 'timing/data': 0.010638684034347534, 'timing/model': 1.5087002199143171, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.85}) +[2026-06-16 01:24:15,638][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2200_state +[2026-06-16 01:24:15,642][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 01:24:21,440][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2200_state/model.safetensors +[2026-06-16 01:24:34,881][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2200_state/optimizer.bin +[2026-06-16 01:24:34,881][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2200_state/sampler.bin +[2026-06-16 01:24:34,881][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2200_state/sampler_1.bin +[2026-06-16 01:24:34,882][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2200_state/random_states_0.pkl +[2026-06-16 01:24:34,988][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2200 +[2026-06-16 01:24:34,988][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 01:24:34,996][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 01:24:41,569][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.04307563975453377, 'train/grad_norm_pre_clip': 11.271952629089355, 'timing/data': 0.013753199949860573, 'timing/model': 1.573867341503501, 'learning_rate/qwen_vl_interface': 6.304005152878902e-06, 'learning_rate/action_model': 6.304005152878901e-05, 'epoch': 0.85}) +[2026-06-16 01:24:48,322][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.11306662857532501, 'train/grad_norm_pre_clip': 9.550832748413086, 'timing/data': 0.014920154586434364, 'timing/model': 1.951169217005372, 'learning_rate/qwen_vl_interface': 6.301035436203602e-06, 'learning_rate/action_model': 6.301035436203602e-05, 'epoch': 0.85}) +[2026-06-16 01:24:54,993][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.07668942213058472, 'train/grad_norm_pre_clip': 8.434296607971191, 'timing/data': 0.016915317624807358, 'timing/model': 1.9519133362919092, 'learning_rate/qwen_vl_interface': 6.298065287487337e-06, 'learning_rate/action_model': 6.298065287487337e-05, 'epoch': 0.85}) +[2026-06-16 01:25:01,420][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.11907000094652176, 'train/grad_norm_pre_clip': 7.694005966186523, 'timing/data': 0.024483196437358856, 'timing/model': 1.7306461445987225, 'learning_rate/qwen_vl_interface': 6.295094707951019e-06, 'learning_rate/action_model': 6.295094707951019e-05, 'epoch': 0.85}) +[2026-06-16 01:25:07,802][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.0676216334104538, 'train/grad_norm_pre_clip': 5.686626434326172, 'timing/data': 0.014890370890498161, 'timing/model': 1.5762653518468142, 'learning_rate/qwen_vl_interface': 6.292123698815745e-06, 'learning_rate/action_model': 6.292123698815745e-05, 'epoch': 0.85}) +[2026-06-16 01:25:14,377][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.40247273445129395, 'train/grad_norm_pre_clip': 14.303735733032227, 'timing/data': 0.01535925455391407, 'timing/model': 1.612055193632841, 'learning_rate/qwen_vl_interface': 6.289152261302783e-06, 'learning_rate/action_model': 6.289152261302784e-05, 'epoch': 0.85}) +[2026-06-16 01:25:20,889][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.14447763562202454, 'train/grad_norm_pre_clip': 22.587081909179688, 'timing/data': 0.018321869894862175, 'timing/model': 1.5341636221855879, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 0.85}) +[2026-06-16 01:25:27,321][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.0332004651427269, 'train/grad_norm_pre_clip': 4.9138970375061035, 'timing/data': 0.01705031841993332, 'timing/model': 1.5807129871100187, 'learning_rate/qwen_vl_interface': 6.283208106029753e-06, 'learning_rate/action_model': 6.283208106029752e-05, 'epoch': 0.85}) +[2026-06-16 01:25:33,635][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.04317278414964676, 'train/grad_norm_pre_clip': 4.701192378997803, 'timing/data': 0.018379300832748413, 'timing/model': 1.771302754059434, 'learning_rate/qwen_vl_interface': 6.280235390713102e-06, 'learning_rate/action_model': 6.280235390713102e-05, 'epoch': 0.85}) +[2026-06-16 01:25:39,897][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.09148970246315002, 'train/grad_norm_pre_clip': 9.578542709350586, 'timing/data': 0.01688801869750023, 'timing/model': 1.6238242536783218, 'learning_rate/qwen_vl_interface': 6.277262251905597e-06, 'learning_rate/action_model': 6.277262251905597e-05, 'epoch': 0.86}) +[2026-06-16 01:25:46,606][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.0562407523393631, 'train/grad_norm_pre_clip': 5.270678520202637, 'timing/data': 0.01678679697215557, 'timing/model': 1.64140291698277, 'learning_rate/qwen_vl_interface': 6.274288690829385e-06, 'learning_rate/action_model': 6.274288690829384e-05, 'epoch': 0.86}) +[2026-06-16 01:25:53,298][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.09614399075508118, 'train/grad_norm_pre_clip': 7.5895538330078125, 'timing/data': 0.019587412476539612, 'timing/model': 1.6094300597906113, 'learning_rate/qwen_vl_interface': 6.271314708706781e-06, 'learning_rate/action_model': 6.271314708706782e-05, 'epoch': 0.86}) +[2026-06-16 01:25:59,811][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.0902993381023407, 'train/grad_norm_pre_clip': 9.659158706665039, 'timing/data': 0.017494356259703636, 'timing/model': 1.5617108680307865, 'learning_rate/qwen_vl_interface': 6.26834030676028e-06, 'learning_rate/action_model': 6.26834030676028e-05, 'epoch': 0.86}) +[2026-06-16 01:26:06,451][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.11404373496770859, 'train/grad_norm_pre_clip': 6.345771789550781, 'timing/data': 0.012269152328372002, 'timing/model': 1.8552102521061897, 'learning_rate/qwen_vl_interface': 6.265365486212545e-06, 'learning_rate/action_model': 6.265365486212545e-05, 'epoch': 0.86}) +[2026-06-16 01:26:12,658][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.037215352058410645, 'train/grad_norm_pre_clip': 6.6067399978637695, 'timing/data': 0.0185744296759367, 'timing/model': 1.6511671226471663, 'learning_rate/qwen_vl_interface': 6.26239024828641e-06, 'learning_rate/action_model': 6.26239024828641e-05, 'epoch': 0.86}) +[2026-06-16 01:26:19,228][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.0779988095164299, 'train/grad_norm_pre_clip': 4.096742153167725, 'timing/data': 0.01893569529056549, 'timing/model': 1.6306541170924902, 'learning_rate/qwen_vl_interface': 6.259414594204888e-06, 'learning_rate/action_model': 6.259414594204888e-05, 'epoch': 0.86}) +[2026-06-16 01:26:25,618][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.03317543864250183, 'train/grad_norm_pre_clip': 2.54923939704895, 'timing/data': 0.012650277465581894, 'timing/model': 1.6131833102554083, 'learning_rate/qwen_vl_interface': 6.256438525191155e-06, 'learning_rate/action_model': 6.256438525191156e-05, 'epoch': 0.86}) +[2026-06-16 01:26:32,144][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.29991963505744934, 'train/grad_norm_pre_clip': 16.422361373901367, 'timing/data': 0.017943955957889557, 'timing/model': 1.6549121029675007, 'learning_rate/qwen_vl_interface': 6.253462042468561e-06, 'learning_rate/action_model': 6.25346204246856e-05, 'epoch': 0.86}) +[2026-06-16 01:26:38,722][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.07397541403770447, 'train/grad_norm_pre_clip': 4.056596755981445, 'timing/data': 0.01869010552763939, 'timing/model': 1.6355165801942348, 'learning_rate/qwen_vl_interface': 6.2504851472606246e-06, 'learning_rate/action_model': 6.250485147260625e-05, 'epoch': 0.86}) +[2026-06-16 01:26:44,992][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.18950743973255157, 'train/grad_norm_pre_clip': 15.48570728302002, 'timing/data': 0.01198875717818737, 'timing/model': 1.7991756554692984, 'learning_rate/qwen_vl_interface': 6.247507840791035e-06, 'learning_rate/action_model': 6.247507840791035e-05, 'epoch': 0.86}) +[2026-06-16 01:26:51,054][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.12041991204023361, 'train/grad_norm_pre_clip': 11.004368782043457, 'timing/data': 0.012952456250786781, 'timing/model': 1.5312457084655762, 'learning_rate/qwen_vl_interface': 6.244530124283651e-06, 'learning_rate/action_model': 6.24453012428365e-05, 'epoch': 0.86}) +[2026-06-16 01:26:57,358][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.08997582644224167, 'train/grad_norm_pre_clip': 8.33364200592041, 'timing/data': 0.011841531842947006, 'timing/model': 1.561933808028698, 'learning_rate/qwen_vl_interface': 6.2415519989625016e-06, 'learning_rate/action_model': 6.241551998962501e-05, 'epoch': 0.86}) +[2026-06-16 01:27:03,501][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.048495423048734665, 'train/grad_norm_pre_clip': 5.0449018478393555, 'timing/data': 0.011429719626903534, 'timing/model': 1.4983693026006222, 'learning_rate/qwen_vl_interface': 6.238573466051781e-06, 'learning_rate/action_model': 6.23857346605178e-05, 'epoch': 0.86}) +[2026-06-16 01:27:10,045][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.0953882560133934, 'train/grad_norm_pre_clip': 5.881357669830322, 'timing/data': 0.012303028255701065, 'timing/model': 1.548263667151332, 'learning_rate/qwen_vl_interface': 6.2355945267758465e-06, 'learning_rate/action_model': 6.235594526775846e-05, 'epoch': 0.86}) +[2026-06-16 01:27:16,345][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.009611892513930798, 'train/grad_norm_pre_clip': 1.3114991188049316, 'timing/data': 0.013462314382195473, 'timing/model': 1.5526533871889114, 'learning_rate/qwen_vl_interface': 6.232615182359231e-06, 'learning_rate/action_model': 6.232615182359231e-05, 'epoch': 0.86}) +[2026-06-16 01:27:22,570][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.10901633650064468, 'train/grad_norm_pre_clip': 11.721089363098145, 'timing/data': 0.012635312974452972, 'timing/model': 1.7800349406898022, 'learning_rate/qwen_vl_interface': 6.229635434026634e-06, 'learning_rate/action_model': 6.229635434026634e-05, 'epoch': 0.86}) +[2026-06-16 01:27:28,639][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.09320758283138275, 'train/grad_norm_pre_clip': 9.736494064331055, 'timing/data': 0.011524150148034096, 'timing/model': 1.5319761168211699, 'learning_rate/qwen_vl_interface': 6.226655283002912e-06, 'learning_rate/action_model': 6.226655283002911e-05, 'epoch': 0.86}) +[2026-06-16 01:27:34,744][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.023043369874358177, 'train/grad_norm_pre_clip': 2.41096830368042, 'timing/data': 0.012281280010938644, 'timing/model': 1.4801684692502022, 'learning_rate/qwen_vl_interface': 6.223674730513096e-06, 'learning_rate/action_model': 6.223674730513096e-05, 'epoch': 0.86}) +[2026-06-16 01:27:40,939][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.25502094626426697, 'train/grad_norm_pre_clip': 23.513376235961914, 'timing/data': 0.01249915175139904, 'timing/model': 1.5123654063791037, 'learning_rate/qwen_vl_interface': 6.220693777782375e-06, 'learning_rate/action_model': 6.220693777782375e-05, 'epoch': 0.86}) +[2026-06-16 01:27:47,150][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.0589916817843914, 'train/grad_norm_pre_clip': 13.265701293945312, 'timing/data': 0.013484127819538116, 'timing/model': 1.4837845489382744, 'learning_rate/qwen_vl_interface': 6.21771242603611e-06, 'learning_rate/action_model': 6.21771242603611e-05, 'epoch': 0.86}) +[2026-06-16 01:27:53,338][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.04526279494166374, 'train/grad_norm_pre_clip': 8.844928741455078, 'timing/data': 0.013149812817573547, 'timing/model': 1.510996650904417, 'learning_rate/qwen_vl_interface': 6.21473067649982e-06, 'learning_rate/action_model': 6.21473067649982e-05, 'epoch': 0.86}) +[2026-06-16 01:27:59,572][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.039043884724378586, 'train/grad_norm_pre_clip': 9.31554126739502, 'timing/data': 0.013016343116760254, 'timing/model': 1.747997859492898, 'learning_rate/qwen_vl_interface': 6.211748530399189e-06, 'learning_rate/action_model': 6.211748530399189e-05, 'epoch': 0.86}) +[2026-06-16 01:28:05,630][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.22852501273155212, 'train/grad_norm_pre_clip': 15.788986206054688, 'timing/data': 0.011601964011788368, 'timing/model': 1.5212965272367, 'learning_rate/qwen_vl_interface': 6.208765988960067e-06, 'learning_rate/action_model': 6.208765988960066e-05, 'epoch': 0.86}) +[2026-06-16 01:28:11,698][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.003638227004557848, 'train/grad_norm_pre_clip': 0.6227094531059265, 'timing/data': 0.010247116908431053, 'timing/model': 1.45473668910563, 'learning_rate/qwen_vl_interface': 6.2057830534084615e-06, 'learning_rate/action_model': 6.205783053408461e-05, 'epoch': 0.86}) +[2026-06-16 01:28:17,584][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.15428388118743896, 'train/grad_norm_pre_clip': 13.734089851379395, 'timing/data': 0.010205458849668503, 'timing/model': 1.426252108067274, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 0.86}) +[2026-06-16 01:28:23,387][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.34266337752342224, 'train/grad_norm_pre_clip': 14.536227226257324, 'timing/data': 0.010803446173667908, 'timing/model': 1.4389530755579472, 'learning_rate/qwen_vl_interface': 6.199816004872655e-06, 'learning_rate/action_model': 6.199816004872655e-05, 'epoch': 0.87}) +[2026-06-16 01:28:29,279][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.17090117931365967, 'train/grad_norm_pre_clip': 21.329402923583984, 'timing/data': 0.01033330149948597, 'timing/model': 1.4164738953113556, 'learning_rate/qwen_vl_interface': 6.196831894341281e-06, 'learning_rate/action_model': 6.19683189434128e-05, 'epoch': 0.87}) +[2026-06-16 01:28:35,165][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.022198911756277084, 'train/grad_norm_pre_clip': 10.51009464263916, 'timing/data': 0.011431694030761719, 'timing/model': 1.6291558817029, 'learning_rate/qwen_vl_interface': 6.1938473946030795e-06, 'learning_rate/action_model': 6.193847394603079e-05, 'epoch': 0.87}) +[2026-06-16 01:28:40,779][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.0654325783252716, 'train/grad_norm_pre_clip': 8.771780967712402, 'timing/data': 0.010768657550215721, 'timing/model': 1.4454173762351274, 'learning_rate/qwen_vl_interface': 6.1908625068848695e-06, 'learning_rate/action_model': 6.19086250688487e-05, 'epoch': 0.87}) +[2026-06-16 01:28:46,676][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.003545597195625305, 'train/grad_norm_pre_clip': 0.3984053432941437, 'timing/data': 0.010402029380202293, 'timing/model': 1.4375581219792366, 'learning_rate/qwen_vl_interface': 6.187877232413619e-06, 'learning_rate/action_model': 6.18787723241362e-05, 'epoch': 0.87}) +[2026-06-16 01:28:52,480][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.04907551780343056, 'train/grad_norm_pre_clip': 6.810282230377197, 'timing/data': 0.010511068627238274, 'timing/model': 1.4482219871133566, 'learning_rate/qwen_vl_interface': 6.184891572416467e-06, 'learning_rate/action_model': 6.184891572416466e-05, 'epoch': 0.87}) +[2026-06-16 01:28:58,379][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.11556770652532578, 'train/grad_norm_pre_clip': 6.26511287689209, 'timing/data': 0.010853942483663559, 'timing/model': 1.4398279022425413, 'learning_rate/qwen_vl_interface': 6.181905528120703e-06, 'learning_rate/action_model': 6.181905528120703e-05, 'epoch': 0.87}) +[2026-06-16 01:29:04,314][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.08679452538490295, 'train/grad_norm_pre_clip': 11.141938209533691, 'timing/data': 0.009977607056498528, 'timing/model': 1.4610223975032568, 'learning_rate/qwen_vl_interface': 6.178919100753778e-06, 'learning_rate/action_model': 6.178919100753778e-05, 'epoch': 0.87}) +[2026-06-16 01:29:10,307][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.027955317869782448, 'train/grad_norm_pre_clip': 7.774988651275635, 'timing/data': 0.01099320687353611, 'timing/model': 1.6403957717120647, 'learning_rate/qwen_vl_interface': 6.175932291543299e-06, 'learning_rate/action_model': 6.175932291543298e-05, 'epoch': 0.87}) +[2026-06-16 01:29:16,364][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.004335425328463316, 'train/grad_norm_pre_clip': 0.7841615080833435, 'timing/data': 0.012428967282176018, 'timing/model': 1.5310712475329638, 'learning_rate/qwen_vl_interface': 6.172945101717029e-06, 'learning_rate/action_model': 6.17294510171703e-05, 'epoch': 0.87}) +[2026-06-16 01:29:22,616][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.004243983421474695, 'train/grad_norm_pre_clip': 0.6609145998954773, 'timing/data': 0.011035468429327011, 'timing/model': 1.4772700034081936, 'learning_rate/qwen_vl_interface': 6.169957532502892e-06, 'learning_rate/action_model': 6.169957532502892e-05, 'epoch': 0.87}) +[2026-06-16 01:29:28,707][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.01936950534582138, 'train/grad_norm_pre_clip': 4.4411211013793945, 'timing/data': 0.010734030976891518, 'timing/model': 1.5227083191275597, 'learning_rate/qwen_vl_interface': 6.166969585128962e-06, 'learning_rate/action_model': 6.166969585128962e-05, 'epoch': 0.87}) +[2026-06-16 01:29:34,808][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.1292085498571396, 'train/grad_norm_pre_clip': 23.649307250976562, 'timing/data': 0.010018985718488693, 'timing/model': 1.5275779571384192, 'learning_rate/qwen_vl_interface': 6.163981260823473e-06, 'learning_rate/action_model': 6.163981260823473e-05, 'epoch': 0.87}) +[2026-06-16 01:29:40,920][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.2855510115623474, 'train/grad_norm_pre_clip': 21.785051345825195, 'timing/data': 0.01013236865401268, 'timing/model': 1.520665792748332, 'learning_rate/qwen_vl_interface': 6.16099256081481e-06, 'learning_rate/action_model': 6.16099256081481e-05, 'epoch': 0.87}) +[2026-06-16 01:29:47,209][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.05280807986855507, 'train/grad_norm_pre_clip': 36.732730865478516, 'timing/data': 0.01243683136999607, 'timing/model': 1.7789214793592691, 'learning_rate/qwen_vl_interface': 6.15800348633152e-06, 'learning_rate/action_model': 6.15800348633152e-05, 'epoch': 0.87}) +[2026-06-16 01:29:53,722][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.026092665269970894, 'train/grad_norm_pre_clip': 3.6196444034576416, 'timing/data': 0.017181117087602615, 'timing/model': 1.7357619311660528, 'learning_rate/qwen_vl_interface': 6.155014038602293e-06, 'learning_rate/action_model': 6.155014038602292e-05, 'epoch': 0.87}) +[2026-06-16 01:30:00,562][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.0739186480641365, 'train/grad_norm_pre_clip': 9.73387622833252, 'timing/data': 0.015755079686641693, 'timing/model': 1.6093469113111496, 'learning_rate/qwen_vl_interface': 6.152024218855979e-06, 'learning_rate/action_model': 6.15202421885598e-05, 'epoch': 0.87}) +[2026-06-16 01:30:07,391][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.004701234865933657, 'train/grad_norm_pre_clip': 0.5434229969978333, 'timing/data': 0.01354915276169777, 'timing/model': 1.5282560717314482, 'learning_rate/qwen_vl_interface': 6.149034028321583e-06, 'learning_rate/action_model': 6.149034028321583e-05, 'epoch': 0.87}) +[2026-06-16 01:30:13,928][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.11930504441261292, 'train/grad_norm_pre_clip': 7.005909442901611, 'timing/data': 0.016941597685217857, 'timing/model': 1.6365461759269238, 'learning_rate/qwen_vl_interface': 6.1460434682282566e-06, 'learning_rate/action_model': 6.146043468228256e-05, 'epoch': 0.87}) +[2026-06-16 01:30:20,269][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.16017462313175201, 'train/grad_norm_pre_clip': 11.185795783996582, 'timing/data': 0.011346522718667984, 'timing/model': 1.5365753769874573, 'learning_rate/qwen_vl_interface': 6.143052539805308e-06, 'learning_rate/action_model': 6.143052539805308e-05, 'epoch': 0.87}) +[2026-06-16 01:30:26,613][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.10700056701898575, 'train/grad_norm_pre_clip': 21.106054306030273, 'timing/data': 0.011699911206960678, 'timing/model': 1.7661319728940725, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 0.87}) +[2026-06-16 01:30:32,882][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.0029092119075357914, 'train/grad_norm_pre_clip': 0.6302772164344788, 'timing/data': 0.013132985681295395, 'timing/model': 1.7205308731645346, 'learning_rate/qwen_vl_interface': 6.137069582888521e-06, 'learning_rate/action_model': 6.137069582888521e-05, 'epoch': 0.87}) +[2026-06-16 01:30:38,944][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.21933354437351227, 'train/grad_norm_pre_clip': 34.819271087646484, 'timing/data': 0.012372571974992752, 'timing/model': 1.5137319397181273, 'learning_rate/qwen_vl_interface': 6.13407755685405e-06, 'learning_rate/action_model': 6.13407755685405e-05, 'epoch': 0.87}) +[2026-06-16 01:30:45,162][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.11205180734395981, 'train/grad_norm_pre_clip': 25.073707580566406, 'timing/data': 0.01235293596982956, 'timing/model': 1.572828434407711, 'learning_rate/qwen_vl_interface': 6.131085167408692e-06, 'learning_rate/action_model': 6.131085167408692e-05, 'epoch': 0.87}) +[2026-06-16 01:30:51,087][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.0774451196193695, 'train/grad_norm_pre_clip': 17.111263275146484, 'timing/data': 0.010267049074172974, 'timing/model': 1.4566396810114384, 'learning_rate/qwen_vl_interface': 6.128092415782502e-06, 'learning_rate/action_model': 6.128092415782501e-05, 'epoch': 0.87}) +[2026-06-16 01:30:56,985][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.006570721510797739, 'train/grad_norm_pre_clip': 1.8204227685928345, 'timing/data': 0.011423183605074883, 'timing/model': 1.4223638288676739, 'learning_rate/qwen_vl_interface': 6.125099303205688e-06, 'learning_rate/action_model': 6.125099303205688e-05, 'epoch': 0.88}) +[2026-06-16 01:31:02,893][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.20669417083263397, 'train/grad_norm_pre_clip': 16.787057876586914, 'timing/data': 0.010257596150040627, 'timing/model': 1.644076868891716, 'learning_rate/qwen_vl_interface': 6.122105830908607e-06, 'learning_rate/action_model': 6.122105830908606e-05, 'epoch': 0.88}) +[2026-06-16 01:31:08,585][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.1863275021314621, 'train/grad_norm_pre_clip': 10.78583812713623, 'timing/data': 0.010862685739994049, 'timing/model': 1.4357615131884813, 'learning_rate/qwen_vl_interface': 6.11911200012176e-06, 'learning_rate/action_model': 6.11911200012176e-05, 'epoch': 0.88}) +[2026-06-16 01:31:14,483][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.0013074165908619761, 'train/grad_norm_pre_clip': 0.272621750831604, 'timing/data': 0.009142352268099785, 'timing/model': 1.4558867514133453, 'learning_rate/qwen_vl_interface': 6.116117812075799e-06, 'learning_rate/action_model': 6.116117812075799e-05, 'epoch': 0.88}) +[2026-06-16 01:31:20,386][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.02252119965851307, 'train/grad_norm_pre_clip': 6.469943046569824, 'timing/data': 0.010319916531443596, 'timing/model': 1.4593638349324465, 'learning_rate/qwen_vl_interface': 6.113123268001524e-06, 'learning_rate/action_model': 6.113123268001524e-05, 'epoch': 0.88}) +[2026-06-16 01:31:26,279][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.11644640564918518, 'train/grad_norm_pre_clip': 7.665858745574951, 'timing/data': 0.009700026363134384, 'timing/model': 1.450583415105939, 'learning_rate/qwen_vl_interface': 6.110128369129873e-06, 'learning_rate/action_model': 6.110128369129873e-05, 'epoch': 0.88}) +[2026-06-16 01:31:32,081][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.006219516042619944, 'train/grad_norm_pre_clip': 2.015909194946289, 'timing/data': 0.010100647807121277, 'timing/model': 1.4398020338267088, 'learning_rate/qwen_vl_interface': 6.107133116691943e-06, 'learning_rate/action_model': 6.107133116691942e-05, 'epoch': 0.88}) +[2026-06-16 01:31:37,919][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.25466370582580566, 'train/grad_norm_pre_clip': 12.453814506530762, 'timing/data': 0.009777655825018883, 'timing/model': 1.5914477799087763, 'learning_rate/qwen_vl_interface': 6.104137511918963e-06, 'learning_rate/action_model': 6.104137511918963e-05, 'epoch': 0.88}) +[2026-06-16 01:31:43,580][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.0694257915019989, 'train/grad_norm_pre_clip': 7.943427562713623, 'timing/data': 0.009976178407669067, 'timing/model': 1.4529608525335789, 'learning_rate/qwen_vl_interface': 6.1011415560423155e-06, 'learning_rate/action_model': 6.101141556042316e-05, 'epoch': 0.88}) +[2026-06-16 01:31:49,478][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.18196071684360504, 'train/grad_norm_pre_clip': 15.221118927001953, 'timing/data': 0.010481618344783783, 'timing/model': 1.4494313336908817, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.88}) +[2026-06-16 01:31:55,379][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.21009647846221924, 'train/grad_norm_pre_clip': 15.028457641601562, 'timing/data': 0.010129166767001152, 'timing/model': 1.4467472322285175, 'learning_rate/qwen_vl_interface': 6.09514859590426e-06, 'learning_rate/action_model': 6.09514859590426e-05, 'epoch': 0.88}) +[2026-06-16 01:32:01,360][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.030312100425362587, 'train/grad_norm_pre_clip': 16.071931838989258, 'timing/data': 0.009916212409734726, 'timing/model': 1.5293567404150963, 'learning_rate/qwen_vl_interface': 6.09215159410633e-06, 'learning_rate/action_model': 6.09215159410633e-05, 'epoch': 0.88}) +[2026-06-16 01:32:07,650][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.013867327012121677, 'train/grad_norm_pre_clip': 1.799735426902771, 'timing/data': 0.011979363858699799, 'timing/model': 1.5433551035821438, 'learning_rate/qwen_vl_interface': 6.089154246131692e-06, 'learning_rate/action_model': 6.089154246131692e-05, 'epoch': 0.88}) +[2026-06-16 01:32:13,755][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.008616281673312187, 'train/grad_norm_pre_clip': 0.8335216641426086, 'timing/data': 0.012306759133934975, 'timing/model': 1.7109629847109318, 'learning_rate/qwen_vl_interface': 6.08615655321244e-06, 'learning_rate/action_model': 6.08615655321244e-05, 'epoch': 0.88}) +[2026-06-16 01:32:20,018][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.1366773098707199, 'train/grad_norm_pre_clip': 12.26806640625, 'timing/data': 0.009921936318278313, 'timing/model': 1.5802912879735231, 'learning_rate/qwen_vl_interface': 6.083158516580814e-06, 'learning_rate/action_model': 6.083158516580814e-05, 'epoch': 0.88}) +[2026-06-16 01:32:26,017][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.08590691536664963, 'train/grad_norm_pre_clip': 15.121406555175781, 'timing/data': 0.010906774550676346, 'timing/model': 1.4560196865350008, 'learning_rate/qwen_vl_interface': 6.080160137469192e-06, 'learning_rate/action_model': 6.0801601374691916e-05, 'epoch': 0.88}) +[2026-06-16 01:32:31,918][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.006323171779513359, 'train/grad_norm_pre_clip': 1.3031851053237915, 'timing/data': 0.010197848081588745, 'timing/model': 1.4627622440457344, 'learning_rate/qwen_vl_interface': 6.077161417110097e-06, 'learning_rate/action_model': 6.0771614171100965e-05, 'epoch': 0.88}) +[2026-06-16 01:32:38,115][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.3580845594406128, 'train/grad_norm_pre_clip': 20.72947120666504, 'timing/data': 0.009661344811320305, 'timing/model': 1.7166331466287374, 'learning_rate/qwen_vl_interface': 6.0741623567361864e-06, 'learning_rate/action_model': 6.074162356736186e-05, 'epoch': 0.88}) +[2026-06-16 01:32:44,013][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.012930137105286121, 'train/grad_norm_pre_clip': 2.4986820220947266, 'timing/data': 0.01066475361585617, 'timing/model': 1.6586011350154877, 'learning_rate/qwen_vl_interface': 6.071162957580262e-06, 'learning_rate/action_model': 6.071162957580262e-05, 'epoch': 0.88}) +[2026-06-16 01:32:49,710][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.004393140785396099, 'train/grad_norm_pre_clip': 0.9219728708267212, 'timing/data': 0.011040104553103447, 'timing/model': 1.4501192905008793, 'learning_rate/qwen_vl_interface': 6.068163220875266e-06, 'learning_rate/action_model': 6.068163220875266e-05, 'epoch': 0.88}) +[2026-06-16 01:32:55,961][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.002978360978886485, 'train/grad_norm_pre_clip': 0.5470899343490601, 'timing/data': 0.012304192408919334, 'timing/model': 1.5927528887987137, 'learning_rate/qwen_vl_interface': 6.065163147854273e-06, 'learning_rate/action_model': 6.065163147854273e-05, 'epoch': 0.88}) +[2026-06-16 01:33:02,045][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.14530010521411896, 'train/grad_norm_pre_clip': 14.454874038696289, 'timing/data': 0.012252012267708778, 'timing/model': 1.552014783024788, 'learning_rate/qwen_vl_interface': 6.062162739750502e-06, 'learning_rate/action_model': 6.062162739750502e-05, 'epoch': 0.88}) +[2026-06-16 01:33:08,140][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.11871412396430969, 'train/grad_norm_pre_clip': 21.68402862548828, 'timing/data': 0.013507021591067314, 'timing/model': 1.5249222554266453, 'learning_rate/qwen_vl_interface': 6.059161997797306e-06, 'learning_rate/action_model': 6.0591619977973056e-05, 'epoch': 0.88}) +[2026-06-16 01:33:14,237][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.40086305141448975, 'train/grad_norm_pre_clip': 16.789104461669922, 'timing/data': 0.012210620567202568, 'timing/model': 1.541555443778634, 'learning_rate/qwen_vl_interface': 6.056160923228178e-06, 'learning_rate/action_model': 6.056160923228178e-05, 'epoch': 0.88}) +[2026-06-16 01:33:20,380][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.09567417949438095, 'train/grad_norm_pre_clip': 22.337072372436523, 'timing/data': 0.012387510389089584, 'timing/model': 1.7845696676522493, 'learning_rate/qwen_vl_interface': 6.0531595172767455e-06, 'learning_rate/action_model': 6.0531595172767455e-05, 'epoch': 0.88}) +[2026-06-16 01:33:26,434][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.025037283077836037, 'train/grad_norm_pre_clip': 9.909651756286621, 'timing/data': 0.012510977685451508, 'timing/model': 1.5518183670938015, 'learning_rate/qwen_vl_interface': 6.050157781176774e-06, 'learning_rate/action_model': 6.050157781176774e-05, 'epoch': 0.88}) +[2026-06-16 01:33:32,537][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.1463867872953415, 'train/grad_norm_pre_clip': 13.263872146606445, 'timing/data': 0.011993089690804482, 'timing/model': 1.535922832787037, 'learning_rate/qwen_vl_interface': 6.047155716162164e-06, 'learning_rate/action_model': 6.047155716162164e-05, 'epoch': 0.89}) +[2026-06-16 01:33:38,728][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.011354073882102966, 'train/grad_norm_pre_clip': 1.9715173244476318, 'timing/data': 0.012437613680958748, 'timing/model': 1.5478351917117834, 'learning_rate/qwen_vl_interface': 6.044153323466948e-06, 'learning_rate/action_model': 6.044153323466948e-05, 'epoch': 0.89}) +[2026-06-16 01:33:44,837][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.0996466651558876, 'train/grad_norm_pre_clip': 12.151212692260742, 'timing/data': 0.012428266927599907, 'timing/model': 1.5437377467751503, 'learning_rate/qwen_vl_interface': 6.0411506043253e-06, 'learning_rate/action_model': 6.0411506043253e-05, 'epoch': 0.89}) +[2026-06-16 01:33:50,932][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.002138565294444561, 'train/grad_norm_pre_clip': 0.3131430447101593, 'timing/data': 0.012667868286371231, 'timing/model': 1.4926056638360023, 'learning_rate/qwen_vl_interface': 6.038147559971523e-06, 'learning_rate/action_model': 6.038147559971523e-05, 'epoch': 0.89}) +[2026-06-16 01:33:57,167][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.18491338193416595, 'train/grad_norm_pre_clip': 13.8471040725708, 'timing/data': 0.01279175654053688, 'timing/model': 1.789317511022091, 'learning_rate/qwen_vl_interface': 6.035144191640055e-06, 'learning_rate/action_model': 6.035144191640055e-05, 'epoch': 0.89}) +[2026-06-16 01:34:03,133][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.03353974223136902, 'train/grad_norm_pre_clip': 17.138381958007812, 'timing/data': 0.01170506700873375, 'timing/model': 1.5378935243934393, 'learning_rate/qwen_vl_interface': 6.032140500565467e-06, 'learning_rate/action_model': 6.0321405005654665e-05, 'epoch': 0.89}) +[2026-06-16 01:34:09,339][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.10698521882295609, 'train/grad_norm_pre_clip': 18.31177520751953, 'timing/data': 0.012881727889180183, 'timing/model': 1.5440093614161015, 'learning_rate/qwen_vl_interface': 6.029136487982463e-06, 'learning_rate/action_model': 6.029136487982463e-05, 'epoch': 0.89}) +[2026-06-16 01:34:15,540][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.4133426249027252, 'train/grad_norm_pre_clip': 24.108917236328125, 'timing/data': 0.012555468827486038, 'timing/model': 1.5087028481066227, 'learning_rate/qwen_vl_interface': 6.02613215512588e-06, 'learning_rate/action_model': 6.0261321551258796e-05, 'epoch': 0.89}) +[2026-06-16 01:34:21,737][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.025354832410812378, 'train/grad_norm_pre_clip': 3.835177183151245, 'timing/data': 0.01258070394396782, 'timing/model': 1.493046909570694, 'learning_rate/qwen_vl_interface': 6.023127503230684e-06, 'learning_rate/action_model': 6.023127503230683e-05, 'epoch': 0.89}) +[2026-06-16 01:34:27,888][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.11102119088172913, 'train/grad_norm_pre_clip': 9.844756126403809, 'timing/data': 0.011654900386929512, 'timing/model': 1.4714988637715578, 'learning_rate/qwen_vl_interface': 6.020122533531976e-06, 'learning_rate/action_model': 6.0201225335319755e-05, 'epoch': 0.89}) +[2026-06-16 01:34:33,772][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.2445148378610611, 'train/grad_norm_pre_clip': 15.834456443786621, 'timing/data': 0.010863680392503738, 'timing/model': 1.6465866938233376, 'learning_rate/qwen_vl_interface': 6.017117247264983e-06, 'learning_rate/action_model': 6.017117247264983e-05, 'epoch': 0.89}) +[2026-06-16 01:34:39,480][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.07346304506063461, 'train/grad_norm_pre_clip': 9.52535343170166, 'timing/data': 0.011766452342271805, 'timing/model': 1.4514832738786936, 'learning_rate/qwen_vl_interface': 6.014111645665069e-06, 'learning_rate/action_model': 6.0141116456650684e-05, 'epoch': 0.89}) +[2026-06-16 01:34:45,381][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.06210726127028465, 'train/grad_norm_pre_clip': 17.132991790771484, 'timing/data': 0.011852052062749863, 'timing/model': 1.4299833066761494, 'learning_rate/qwen_vl_interface': 6.011105729967719e-06, 'learning_rate/action_model': 6.0111057299677184e-05, 'epoch': 0.89}) +[2026-06-16 01:42:47,678][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.09080660343170166, 'train/grad_norm_pre_clip': 13.997525215148926, 'eval/loss': 0.06123311072587967, 'eval/flap_precision': 0.9323607427055703, 'eval/flap_recall': 0.9538670284938942, 'eval/flap_f1': 0.9429912810194501, 'eval/flap_rate_model': 0.20944444444444443, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9429912810194501, 'timing/data': 0.010507635772228241, 'timing/model': 1.4439127743244171, 'learning_rate/qwen_vl_interface': 6.008099501408552e-06, 'learning_rate/action_model': 6.008099501408552e-05, 'epoch': 0.89}) +[2026-06-16 01:42:47,679][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2300_state +[2026-06-16 01:42:47,683][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 01:42:53,755][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2300_state/model.safetensors +[2026-06-16 01:43:08,237][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2300_state/optimizer.bin +[2026-06-16 01:43:08,237][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2300_state/sampler.bin +[2026-06-16 01:43:08,237][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2300_state/sampler_1.bin +[2026-06-16 01:43:08,238][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2300_state/random_states_0.pkl +[2026-06-16 01:43:08,330][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2300 +[2026-06-16 01:43:08,331][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 01:43:08,339][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 01:43:14,530][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.25919434428215027, 'train/grad_norm_pre_clip': 19.937570571899414, 'timing/data': 0.012139780446887016, 'timing/model': 1.5136214066296816, 'learning_rate/qwen_vl_interface': 6.0050929612233165e-06, 'learning_rate/action_model': 6.005092961223316e-05, 'epoch': 0.89}) +[2026-06-16 01:43:20,433][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.20113706588745117, 'train/grad_norm_pre_clip': 19.742813110351562, 'timing/data': 0.010392680764198303, 'timing/model': 1.4640358984470367, 'learning_rate/qwen_vl_interface': 6.002086110647888e-06, 'learning_rate/action_model': 6.002086110647888e-05, 'epoch': 0.89}) +[2026-06-16 01:43:26,312][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.2055419385433197, 'train/grad_norm_pre_clip': 11.7644681930542, 'timing/data': 0.011721622198820114, 'timing/model': 1.4711196850985289, 'learning_rate/qwen_vl_interface': 5.999078950918267e-06, 'learning_rate/action_model': 5.999078950918267e-05, 'epoch': 0.89}) +[2026-06-16 01:43:32,411][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.2247103452682495, 'train/grad_norm_pre_clip': 39.13772201538086, 'timing/data': 0.01629389449954033, 'timing/model': 1.4914484396576881, 'learning_rate/qwen_vl_interface': 5.996071483270585e-06, 'learning_rate/action_model': 5.996071483270585e-05, 'epoch': 0.89}) +[2026-06-16 01:43:38,336][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.04016721993684769, 'train/grad_norm_pre_clip': 6.869581699371338, 'timing/data': 0.010295916348695755, 'timing/model': 1.6742748450487852, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.89}) +[2026-06-16 01:43:44,023][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.04951256513595581, 'train/grad_norm_pre_clip': 2.913646936416626, 'timing/data': 0.019632697105407715, 'timing/model': 1.4499195646494627, 'learning_rate/qwen_vl_interface': 5.990055629166187e-06, 'learning_rate/action_model': 5.990055629166187e-05, 'epoch': 0.89}) +[2026-06-16 01:43:49,885][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.10581228882074356, 'train/grad_norm_pre_clip': 8.228814125061035, 'timing/data': 0.009982319548726082, 'timing/model': 1.4520602114498615, 'learning_rate/qwen_vl_interface': 5.987047245182363e-06, 'learning_rate/action_model': 5.987047245182362e-05, 'epoch': 0.89}) +[2026-06-16 01:43:55,784][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.037519339472055435, 'train/grad_norm_pre_clip': 14.33845329284668, 'timing/data': 0.011953111737966537, 'timing/model': 1.4436961691826582, 'learning_rate/qwen_vl_interface': 5.984038558226255e-06, 'learning_rate/action_model': 5.9840385582262546e-05, 'epoch': 0.89}) +[2026-06-16 01:44:01,700][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.04761577770113945, 'train/grad_norm_pre_clip': 10.577105522155762, 'timing/data': 0.01015125960111618, 'timing/model': 1.464250210672617, 'learning_rate/qwen_vl_interface': 5.981029569534625e-06, 'learning_rate/action_model': 5.981029569534625e-05, 'epoch': 0.89}) +[2026-06-16 01:44:08,003][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.07474447786808014, 'train/grad_norm_pre_clip': 15.644299507141113, 'timing/data': 0.012798594310879707, 'timing/model': 1.491721423342824, 'learning_rate/qwen_vl_interface': 5.978020280344351e-06, 'learning_rate/action_model': 5.9780202803443506e-05, 'epoch': 0.89}) +[2026-06-16 01:44:13,901][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.10197582840919495, 'train/grad_norm_pre_clip': 17.20896339416504, 'timing/data': 0.012207947671413422, 'timing/model': 1.6453491616994143, 'learning_rate/qwen_vl_interface': 5.97501069189244e-06, 'learning_rate/action_model': 5.97501069189244e-05, 'epoch': 0.89}) +[2026-06-16 01:44:20,376][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.2119046449661255, 'train/grad_norm_pre_clip': 17.504459381103516, 'timing/data': 0.012781204655766487, 'timing/model': 1.8843848202377558, 'learning_rate/qwen_vl_interface': 5.97200080541602e-06, 'learning_rate/action_model': 5.97200080541602e-05, 'epoch': 0.89}) +[2026-06-16 01:44:26,884][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.08067432790994644, 'train/grad_norm_pre_clip': 9.958699226379395, 'timing/data': 0.011752752587199211, 'timing/model': 1.5326500330120325, 'learning_rate/qwen_vl_interface': 5.968990622152341e-06, 'learning_rate/action_model': 5.9689906221523404e-05, 'epoch': 0.9}) +[2026-06-16 01:44:33,380][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.03073250874876976, 'train/grad_norm_pre_clip': 3.554439067840576, 'timing/data': 0.011182988062500954, 'timing/model': 1.5410140547901392, 'learning_rate/qwen_vl_interface': 5.965980143338774e-06, 'learning_rate/action_model': 5.965980143338774e-05, 'epoch': 0.9}) +[2026-06-16 01:44:39,894][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.09613261371850967, 'train/grad_norm_pre_clip': 8.067395210266113, 'timing/data': 0.01187194138765335, 'timing/model': 1.5683156438171864, 'learning_rate/qwen_vl_interface': 5.962969370212819e-06, 'learning_rate/action_model': 5.9629693702128185e-05, 'epoch': 0.9}) +[2026-06-16 01:44:46,268][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.02691853791475296, 'train/grad_norm_pre_clip': 2.174156904220581, 'timing/data': 0.01473335362970829, 'timing/model': 1.58903718739748, 'learning_rate/qwen_vl_interface': 5.959958304012085e-06, 'learning_rate/action_model': 5.9599583040120845e-05, 'epoch': 0.9}) +[2026-06-16 01:44:52,602][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.11500317603349686, 'train/grad_norm_pre_clip': 7.0530104637146, 'timing/data': 0.012649796903133392, 'timing/model': 1.7871195077896118, 'learning_rate/qwen_vl_interface': 5.956946945974309e-06, 'learning_rate/action_model': 5.9569469459743085e-05, 'epoch': 0.9}) +[2026-06-16 01:44:58,892][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.1477542519569397, 'train/grad_norm_pre_clip': 11.374141693115234, 'timing/data': 0.012467360123991966, 'timing/model': 1.7778488919138908, 'learning_rate/qwen_vl_interface': 5.953935297337347e-06, 'learning_rate/action_model': 5.953935297337347e-05, 'epoch': 0.9}) +[2026-06-16 01:45:04,969][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.07562939822673798, 'train/grad_norm_pre_clip': 6.548341274261475, 'timing/data': 0.013824427500367165, 'timing/model': 1.5446116346865892, 'learning_rate/qwen_vl_interface': 5.950923359339175e-06, 'learning_rate/action_model': 5.950923359339174e-05, 'epoch': 0.9}) +[2026-06-16 01:45:11,266][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.015359653159976006, 'train/grad_norm_pre_clip': 1.4937108755111694, 'timing/data': 0.013076193630695343, 'timing/model': 1.5429581888020039, 'learning_rate/qwen_vl_interface': 5.947911133217886e-06, 'learning_rate/action_model': 5.947911133217886e-05, 'epoch': 0.9}) +[2026-06-16 01:45:17,574][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.06987567991018295, 'train/grad_norm_pre_clip': 8.322022438049316, 'timing/data': 0.013039814308285713, 'timing/model': 1.5400517247617245, 'learning_rate/qwen_vl_interface': 5.944898620211693e-06, 'learning_rate/action_model': 5.944898620211693e-05, 'epoch': 0.9}) +[2026-06-16 01:45:23,875][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.4511443078517914, 'train/grad_norm_pre_clip': 16.362306594848633, 'timing/data': 0.013126371428370476, 'timing/model': 1.5137968361377716, 'learning_rate/qwen_vl_interface': 5.941885821558925e-06, 'learning_rate/action_model': 5.9418858215589245e-05, 'epoch': 0.9}) +[2026-06-16 01:45:30,196][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.18827423453330994, 'train/grad_norm_pre_clip': 12.214587211608887, 'timing/data': 0.011976229026913643, 'timing/model': 1.7663528732955456, 'learning_rate/qwen_vl_interface': 5.938872738498031e-06, 'learning_rate/action_model': 5.93887273849803e-05, 'epoch': 0.9}) +[2026-06-16 01:45:36,373][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.1054278165102005, 'train/grad_norm_pre_clip': 5.793278217315674, 'timing/data': 0.012432072311639786, 'timing/model': 1.5557311996817589, 'learning_rate/qwen_vl_interface': 5.935859372267575e-06, 'learning_rate/action_model': 5.935859372267575e-05, 'epoch': 0.9}) +[2026-06-16 01:45:42,670][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.014793839305639267, 'train/grad_norm_pre_clip': 1.9309608936309814, 'timing/data': 0.011776058003306389, 'timing/model': 1.5552069153636694, 'learning_rate/qwen_vl_interface': 5.9328457241062396e-06, 'learning_rate/action_model': 5.9328457241062394e-05, 'epoch': 0.9}) +[2026-06-16 01:45:48,978][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.24843165278434753, 'train/grad_norm_pre_clip': 5.417750835418701, 'timing/data': 0.012281255796551704, 'timing/model': 1.5457641519606113, 'learning_rate/qwen_vl_interface': 5.929831795252819e-06, 'learning_rate/action_model': 5.9298317952528184e-05, 'epoch': 0.9}) +[2026-06-16 01:45:55,270][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.11097753793001175, 'train/grad_norm_pre_clip': 8.074721336364746, 'timing/data': 0.014939900487661362, 'timing/model': 1.5114057008177042, 'learning_rate/qwen_vl_interface': 5.926817586946227e-06, 'learning_rate/action_model': 5.926817586946227e-05, 'epoch': 0.9}) +[2026-06-16 01:46:01,567][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.051956597715616226, 'train/grad_norm_pre_clip': 5.720032691955566, 'timing/data': 0.012646082788705826, 'timing/model': 1.5852937046438456, 'learning_rate/qwen_vl_interface': 5.923803100425491e-06, 'learning_rate/action_model': 5.92380310042549e-05, 'epoch': 0.9}) +[2026-06-16 01:46:07,863][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.09240737557411194, 'train/grad_norm_pre_clip': 13.635303497314453, 'timing/data': 0.014214379712939262, 'timing/model': 1.7403413951396942, 'learning_rate/qwen_vl_interface': 5.92078833692975e-06, 'learning_rate/action_model': 5.92078833692975e-05, 'epoch': 0.9}) +[2026-06-16 01:46:13,979][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.23919320106506348, 'train/grad_norm_pre_clip': 28.60736083984375, 'timing/data': 0.023265600204467773, 'timing/model': 1.542379517108202, 'learning_rate/qwen_vl_interface': 5.917773297698264e-06, 'learning_rate/action_model': 5.9177732976982636e-05, 'epoch': 0.9}) +[2026-06-16 01:46:20,138][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.24774986505508423, 'train/grad_norm_pre_clip': 11.097158432006836, 'timing/data': 0.011194439604878426, 'timing/model': 1.4822630602866411, 'learning_rate/qwen_vl_interface': 5.9147579839704e-06, 'learning_rate/action_model': 5.9147579839704e-05, 'epoch': 0.9}) +[2026-06-16 01:46:26,233][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.19875502586364746, 'train/grad_norm_pre_clip': 8.542794227600098, 'timing/data': 0.017410114407539368, 'timing/model': 1.4579945001751184, 'learning_rate/qwen_vl_interface': 5.911742396985637e-06, 'learning_rate/action_model': 5.9117423969856364e-05, 'epoch': 0.9}) +[2026-06-16 01:46:32,090][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.11266964673995972, 'train/grad_norm_pre_clip': 21.598787307739258, 'timing/data': 0.010690178722143173, 'timing/model': 1.4495038241147995, 'learning_rate/qwen_vl_interface': 5.908726537983571e-06, 'learning_rate/action_model': 5.908726537983571e-05, 'epoch': 0.9}) +[2026-06-16 01:46:37,988][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.14169971644878387, 'train/grad_norm_pre_clip': 14.831648826599121, 'timing/data': 0.010889913886785507, 'timing/model': 1.46527886018157, 'learning_rate/qwen_vl_interface': 5.9057104082039085e-06, 'learning_rate/action_model': 5.905710408203908e-05, 'epoch': 0.9}) +[2026-06-16 01:46:43,849][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.012672959826886654, 'train/grad_norm_pre_clip': 3.112602472305298, 'timing/data': 0.017200402915477753, 'timing/model': 1.6144463997334242, 'learning_rate/qwen_vl_interface': 5.902694008886466e-06, 'learning_rate/action_model': 5.902694008886466e-05, 'epoch': 0.9}) +[2026-06-16 01:46:49,482][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.00567982392385602, 'train/grad_norm_pre_clip': 0.5970571637153625, 'timing/data': 0.010760761797428131, 'timing/model': 1.4516281355172396, 'learning_rate/qwen_vl_interface': 5.8996773412711694e-06, 'learning_rate/action_model': 5.899677341271169e-05, 'epoch': 0.9}) +[2026-06-16 01:46:55,382][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.061182938516139984, 'train/grad_norm_pre_clip': 5.727722644805908, 'timing/data': 0.010240452364087105, 'timing/model': 1.4555686302483082, 'learning_rate/qwen_vl_interface': 5.896660406598059e-06, 'learning_rate/action_model': 5.896660406598059e-05, 'epoch': 0.9}) +[2026-06-16 01:47:01,277][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.108038529753685, 'train/grad_norm_pre_clip': 9.00339126586914, 'timing/data': 0.010528013110160828, 'timing/model': 1.4486603662371635, 'learning_rate/qwen_vl_interface': 5.893643206107283e-06, 'learning_rate/action_model': 5.893643206107282e-05, 'epoch': 0.9}) +[2026-06-16 01:47:07,177][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.015341189689934254, 'train/grad_norm_pre_clip': 1.8622722625732422, 'timing/data': 0.011966882273554802, 'timing/model': 1.4142659697681665, 'learning_rate/qwen_vl_interface': 5.890625741039096e-06, 'learning_rate/action_model': 5.890625741039095e-05, 'epoch': 0.91}) +[2026-06-16 01:47:13,074][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.22035415470600128, 'train/grad_norm_pre_clip': 18.519397735595703, 'timing/data': 0.011221775785088539, 'timing/model': 1.6399139892309904, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 0.91}) +[2026-06-16 01:47:18,779][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.11885339766740799, 'train/grad_norm_pre_clip': 9.122867584228516, 'timing/data': 0.011477096006274223, 'timing/model': 1.4410405103117228, 'learning_rate/qwen_vl_interface': 5.884590022132064e-06, 'learning_rate/action_model': 5.884590022132064e-05, 'epoch': 0.91}) +[2026-06-16 01:47:24,688][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.07678326219320297, 'train/grad_norm_pre_clip': 7.4751877784729, 'timing/data': 0.010173194110393524, 'timing/model': 1.4387761242687702, 'learning_rate/qwen_vl_interface': 5.881571770774278e-06, 'learning_rate/action_model': 5.881571770774278e-05, 'epoch': 0.91}) +[2026-06-16 01:47:30,485][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.0029894323088228703, 'train/grad_norm_pre_clip': 0.21824060380458832, 'timing/data': 0.011180516332387924, 'timing/model': 1.440735625103116, 'learning_rate/qwen_vl_interface': 5.878553259801192e-06, 'learning_rate/action_model': 5.878553259801192e-05, 'epoch': 0.91}) +[2026-06-16 01:47:36,534][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.21534819900989532, 'train/grad_norm_pre_clip': 16.683929443359375, 'timing/data': 0.012922912836074829, 'timing/model': 1.5347170885652304, 'learning_rate/qwen_vl_interface': 5.8755344904536025e-06, 'learning_rate/action_model': 5.8755344904536025e-05, 'epoch': 0.91}) +[2026-06-16 01:47:42,647][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.04006803408265114, 'train/grad_norm_pre_clip': 3.436666488647461, 'timing/data': 0.012057680636644363, 'timing/model': 1.5013701245188713, 'learning_rate/qwen_vl_interface': 5.872515463972415e-06, 'learning_rate/action_model': 5.8725154639724146e-05, 'epoch': 0.91}) +[2026-06-16 01:47:48,790][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.08172563463449478, 'train/grad_norm_pre_clip': 7.041662693023682, 'timing/data': 0.012430243194103241, 'timing/model': 1.7716780193150043, 'learning_rate/qwen_vl_interface': 5.869496181598633e-06, 'learning_rate/action_model': 5.869496181598633e-05, 'epoch': 0.91}) +[2026-06-16 01:47:54,734][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.08324287086725235, 'train/grad_norm_pre_clip': 6.5988593101501465, 'timing/data': 0.01255466602742672, 'timing/model': 1.5361258313059807, 'learning_rate/qwen_vl_interface': 5.866476644573371e-06, 'learning_rate/action_model': 5.866476644573371e-05, 'epoch': 0.91}) +[2026-06-16 01:48:01,181][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.01082705520093441, 'train/grad_norm_pre_clip': 1.3307477235794067, 'timing/data': 0.010852320119738579, 'timing/model': 1.7591943629086018, 'learning_rate/qwen_vl_interface': 5.863456854137848e-06, 'learning_rate/action_model': 5.8634568541378474e-05, 'epoch': 0.91}) +[2026-06-16 01:48:07,472][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.041158467531204224, 'train/grad_norm_pre_clip': 7.0381550788879395, 'timing/data': 0.012014804407954216, 'timing/model': 1.53871650993824, 'learning_rate/qwen_vl_interface': 5.860436811533383e-06, 'learning_rate/action_model': 5.860436811533383e-05, 'epoch': 0.91}) +[2026-06-16 01:48:13,775][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.14015570282936096, 'train/grad_norm_pre_clip': 9.710686683654785, 'timing/data': 0.012798141688108444, 'timing/model': 1.5547579787671566, 'learning_rate/qwen_vl_interface': 5.857416518001404e-06, 'learning_rate/action_model': 5.857416518001404e-05, 'epoch': 0.91}) +[2026-06-16 01:48:20,083][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.21275857090950012, 'train/grad_norm_pre_clip': 9.742502212524414, 'timing/data': 0.012495312839746475, 'timing/model': 1.5588975884020329, 'learning_rate/qwen_vl_interface': 5.854395974783439e-06, 'learning_rate/action_model': 5.8543959747834386e-05, 'epoch': 0.91}) +[2026-06-16 01:48:26,455][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.03933897614479065, 'train/grad_norm_pre_clip': 10.986573219299316, 'timing/data': 0.013230158016085625, 'timing/model': 1.817218055948615, 'learning_rate/qwen_vl_interface': 5.851375183121119e-06, 'learning_rate/action_model': 5.8513751831211186e-05, 'epoch': 0.91}) +[2026-06-16 01:48:32,778][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.07276646047830582, 'train/grad_norm_pre_clip': 10.957635879516602, 'timing/data': 0.014670073986053467, 'timing/model': 1.5468167215585709, 'learning_rate/qwen_vl_interface': 5.848354144256175e-06, 'learning_rate/action_model': 5.8483541442561754e-05, 'epoch': 0.91}) +[2026-06-16 01:48:39,063][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.008616236038506031, 'train/grad_norm_pre_clip': 1.228007435798645, 'timing/data': 0.013942983001470566, 'timing/model': 1.5113175753504038, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 0.91}) +[2026-06-16 01:48:45,211][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.08589013665914536, 'train/grad_norm_pre_clip': 12.09541130065918, 'timing/data': 0.010499946773052216, 'timing/model': 1.528708377853036, 'learning_rate/qwen_vl_interface': 5.842311329885866e-06, 'learning_rate/action_model': 5.842311329885866e-05, 'epoch': 0.91}) +[2026-06-16 01:48:51,278][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.05021508410573006, 'train/grad_norm_pre_clip': 4.242032051086426, 'timing/data': 0.011253779754042625, 'timing/model': 1.4213974326848984, 'learning_rate/qwen_vl_interface': 5.839289556864472e-06, 'learning_rate/action_model': 5.8392895568644714e-05, 'epoch': 0.91}) +[2026-06-16 01:48:57,205][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.012397189624607563, 'train/grad_norm_pre_clip': 2.9601826667785645, 'timing/data': 0.010431282222270966, 'timing/model': 1.4864765275269747, 'learning_rate/qwen_vl_interface': 5.836267541608398e-06, 'learning_rate/action_model': 5.836267541608398e-05, 'epoch': 0.91}) +[2026-06-16 01:49:03,504][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.08008419722318649, 'train/grad_norm_pre_clip': 5.276918411254883, 'timing/data': 0.013389967381954193, 'timing/model': 1.7460331730544567, 'learning_rate/qwen_vl_interface': 5.833245285359887e-06, 'learning_rate/action_model': 5.833245285359886e-05, 'epoch': 0.91}) +[2026-06-16 01:49:10,170][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.03116677515208721, 'train/grad_norm_pre_clip': 14.463607788085938, 'timing/data': 0.01264890842139721, 'timing/model': 1.5522000566124916, 'learning_rate/qwen_vl_interface': 5.830222789361268e-06, 'learning_rate/action_model': 5.8302227893612684e-05, 'epoch': 0.91}) +[2026-06-16 01:49:16,475][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.0018651258433237672, 'train/grad_norm_pre_clip': 0.2153063863515854, 'timing/data': 0.014183558523654938, 'timing/model': 1.5533283967524767, 'learning_rate/qwen_vl_interface': 5.8272000548549776e-06, 'learning_rate/action_model': 5.827200054854978e-05, 'epoch': 0.91}) +[2026-06-16 01:49:22,773][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.07932060956954956, 'train/grad_norm_pre_clip': 13.597685813903809, 'timing/data': 0.011644378304481506, 'timing/model': 1.5568940546363592, 'learning_rate/qwen_vl_interface': 5.824177083083547e-06, 'learning_rate/action_model': 5.8241770830835475e-05, 'epoch': 0.91}) +[2026-06-16 01:49:29,074][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.16071870923042297, 'train/grad_norm_pre_clip': 24.349361419677734, 'timing/data': 0.014078782871365547, 'timing/model': 1.5513028036803007, 'learning_rate/qwen_vl_interface': 5.8211538752896056e-06, 'learning_rate/action_model': 5.821153875289606e-05, 'epoch': 0.91}) +[2026-06-16 01:49:35,402][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.057269733399152756, 'train/grad_norm_pre_clip': 26.87766456604004, 'timing/data': 0.013693949207663536, 'timing/model': 1.5263248831033707, 'learning_rate/qwen_vl_interface': 5.8181304327158814e-06, 'learning_rate/action_model': 5.8181304327158814e-05, 'epoch': 0.91}) +[2026-06-16 01:49:41,712][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.01016507763415575, 'train/grad_norm_pre_clip': 2.5537726879119873, 'timing/data': 0.014941612258553505, 'timing/model': 1.7839960400015116, 'learning_rate/qwen_vl_interface': 5.8151067566051965e-06, 'learning_rate/action_model': 5.815106756605196e-05, 'epoch': 0.91}) +[2026-06-16 01:49:47,781][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.07409454137086868, 'train/grad_norm_pre_clip': 7.399503707885742, 'timing/data': 0.01321222260594368, 'timing/model': 1.550741944462061, 'learning_rate/qwen_vl_interface': 5.81208284820047e-06, 'learning_rate/action_model': 5.81208284820047e-05, 'epoch': 0.92}) +[2026-06-16 01:49:54,068][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.42949432134628296, 'train/grad_norm_pre_clip': 25.48980140686035, 'timing/data': 0.012609077617526054, 'timing/model': 1.5233839824795723, 'learning_rate/qwen_vl_interface': 5.809058708744716e-06, 'learning_rate/action_model': 5.809058708744716e-05, 'epoch': 0.92}) +[2026-06-16 01:50:00,365][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.009165508672595024, 'train/grad_norm_pre_clip': 1.0252761840820312, 'timing/data': 0.01156851090490818, 'timing/model': 1.525232158601284, 'learning_rate/qwen_vl_interface': 5.8060343394810445e-06, 'learning_rate/action_model': 5.806034339481045e-05, 'epoch': 0.92}) +[2026-06-16 01:50:06,802][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.1560850888490677, 'train/grad_norm_pre_clip': 6.764654636383057, 'timing/data': 0.012680493295192719, 'timing/model': 1.5987748764455318, 'learning_rate/qwen_vl_interface': 5.803009741652659e-06, 'learning_rate/action_model': 5.8030097416526585e-05, 'epoch': 0.92}) +[2026-06-16 01:50:13,171][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.01911172829568386, 'train/grad_norm_pre_clip': 4.150325298309326, 'timing/data': 0.013462832197546959, 'timing/model': 1.589933842420578, 'learning_rate/qwen_vl_interface': 5.799984916502859e-06, 'learning_rate/action_model': 5.799984916502859e-05, 'epoch': 0.92}) +[2026-06-16 01:50:19,498][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.2653178870677948, 'train/grad_norm_pre_clip': 16.337797164916992, 'timing/data': 0.012014301493763924, 'timing/model': 1.7543040663003922, 'learning_rate/qwen_vl_interface': 5.7969598652750334e-06, 'learning_rate/action_model': 5.796959865275033e-05, 'epoch': 0.92}) +[2026-06-16 01:50:25,571][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.10043374449014664, 'train/grad_norm_pre_clip': 15.765563011169434, 'timing/data': 0.012522667646408081, 'timing/model': 1.5137765184044838, 'learning_rate/qwen_vl_interface': 5.793934589212666e-06, 'learning_rate/action_model': 5.793934589212666e-05, 'epoch': 0.92}) +[2026-06-16 01:50:31,872][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.1362510323524475, 'train/grad_norm_pre_clip': 20.61403465270996, 'timing/data': 0.01288016326725483, 'timing/model': 1.5267181098461151, 'learning_rate/qwen_vl_interface': 5.7909090895593375e-06, 'learning_rate/action_model': 5.790909089559338e-05, 'epoch': 0.92}) +[2026-06-16 01:50:37,896][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.016574157401919365, 'train/grad_norm_pre_clip': 9.360991477966309, 'timing/data': 0.01655212603509426, 'timing/model': 1.4477159958332777, 'learning_rate/qwen_vl_interface': 5.7878833675587135e-06, 'learning_rate/action_model': 5.787883367558713e-05, 'epoch': 0.92}) +[2026-06-16 01:50:43,854][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.356943815946579, 'train/grad_norm_pre_clip': 10.886334419250488, 'timing/data': 0.010490763932466507, 'timing/model': 1.5188139211386442, 'learning_rate/qwen_vl_interface': 5.784857424454556e-06, 'learning_rate/action_model': 5.7848574244545563e-05, 'epoch': 0.92}) +[2026-06-16 01:50:50,036][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.3401874601840973, 'train/grad_norm_pre_clip': 20.775434494018555, 'timing/data': 0.012334881350398064, 'timing/model': 1.4883222877979279, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 0.92}) +[2026-06-16 01:50:56,155][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.19650331139564514, 'train/grad_norm_pre_clip': 14.162393569946289, 'timing/data': 0.013188280165195465, 'timing/model': 1.7484980132430792, 'learning_rate/qwen_vl_interface': 5.778804879911129e-06, 'learning_rate/action_model': 5.778804879911128e-05, 'epoch': 0.92}) +[2026-06-16 01:51:02,343][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.07825888693332672, 'train/grad_norm_pre_clip': 9.521212577819824, 'timing/data': 0.012231698259711266, 'timing/model': 1.528086656704545, 'learning_rate/qwen_vl_interface': 5.775778280959832e-06, 'learning_rate/action_model': 5.7757782809598315e-05, 'epoch': 0.92}) +[2026-06-16 01:51:08,590][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.016800304874777794, 'train/grad_norm_pre_clip': 3.0904786586761475, 'timing/data': 0.013674605637788773, 'timing/model': 1.683258593082428, 'learning_rate/qwen_vl_interface': 5.7727514658809435e-06, 'learning_rate/action_model': 5.772751465880944e-05, 'epoch': 0.92}) +[2026-06-16 01:51:14,966][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.013250085525214672, 'train/grad_norm_pre_clip': 2.6840946674346924, 'timing/data': 0.012281915172934532, 'timing/model': 1.5344107188284397, 'learning_rate/qwen_vl_interface': 5.769724435918676e-06, 'learning_rate/action_model': 5.7697244359186755e-05, 'epoch': 0.92}) +[2026-06-16 01:51:21,267][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.08461848646402359, 'train/grad_norm_pre_clip': 11.228042602539062, 'timing/data': 0.01251242496073246, 'timing/model': 1.537877669557929, 'learning_rate/qwen_vl_interface': 5.766697192317323e-06, 'learning_rate/action_model': 5.766697192317323e-05, 'epoch': 0.92}) +[2026-06-16 01:51:27,547][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.07471879571676254, 'train/grad_norm_pre_clip': 9.76347541809082, 'timing/data': 0.01158624142408371, 'timing/model': 1.4901021402329206, 'learning_rate/qwen_vl_interface': 5.763669736321271e-06, 'learning_rate/action_model': 5.763669736321271e-05, 'epoch': 0.92}) +[2026-06-16 01:51:33,692][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.15168802440166473, 'train/grad_norm_pre_clip': 6.637903690338135, 'timing/data': 0.011454958468675613, 'timing/model': 1.7871597949415445, 'learning_rate/qwen_vl_interface': 5.760642069174994e-06, 'learning_rate/action_model': 5.760642069174994e-05, 'epoch': 0.92}) +[2026-06-16 01:51:39,826][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.0515865758061409, 'train/grad_norm_pre_clip': 14.487845420837402, 'timing/data': 0.011496327817440033, 'timing/model': 1.730356002226472, 'learning_rate/qwen_vl_interface': 5.757614192123051e-06, 'learning_rate/action_model': 5.757614192123051e-05, 'epoch': 0.92}) +[2026-06-16 01:51:45,756][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.020635394379496574, 'train/grad_norm_pre_clip': 2.540707588195801, 'timing/data': 0.01238919235765934, 'timing/model': 1.5556324739009142, 'learning_rate/qwen_vl_interface': 5.754586106410089e-06, 'learning_rate/action_model': 5.754586106410088e-05, 'epoch': 0.92}) +[2026-06-16 01:51:51,931][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.12548835575580597, 'train/grad_norm_pre_clip': 12.744081497192383, 'timing/data': 0.012854868546128273, 'timing/model': 1.546296687796712, 'learning_rate/qwen_vl_interface': 5.751557813280838e-06, 'learning_rate/action_model': 5.751557813280838e-05, 'epoch': 0.92}) +[2026-06-16 01:51:57,774][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.023604828864336014, 'train/grad_norm_pre_clip': 3.3468070030212402, 'timing/data': 0.011238371953368187, 'timing/model': 1.4522929973900318, 'learning_rate/qwen_vl_interface': 5.7485293139801146e-06, 'learning_rate/action_model': 5.7485293139801146e-05, 'epoch': 0.92}) +[2026-06-16 01:52:03,818][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.010193301364779472, 'train/grad_norm_pre_clip': 1.6568386554718018, 'timing/data': 0.011275941506028175, 'timing/model': 1.531878475099802, 'learning_rate/qwen_vl_interface': 5.745500609752821e-06, 'learning_rate/action_model': 5.745500609752821e-05, 'epoch': 0.92}) +[2026-06-16 01:52:10,104][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.11447799950838089, 'train/grad_norm_pre_clip': 18.64565658569336, 'timing/data': 0.013046720996499062, 'timing/model': 1.7577260583639145, 'learning_rate/qwen_vl_interface': 5.742471701843943e-06, 'learning_rate/action_model': 5.742471701843943e-05, 'epoch': 0.92}) +[2026-06-16 01:52:16,169][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.0030056843534111977, 'train/grad_norm_pre_clip': 0.42451605200767517, 'timing/data': 0.01313001848757267, 'timing/model': 1.5197239480912685, 'learning_rate/qwen_vl_interface': 5.739442591498553e-06, 'learning_rate/action_model': 5.739442591498553e-05, 'epoch': 0.92}) +[2026-06-16 01:52:22,583][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.03064907155930996, 'train/grad_norm_pre_clip': 5.974395275115967, 'timing/data': 0.012768812477588654, 'timing/model': 1.531499058008194, 'learning_rate/qwen_vl_interface': 5.7364132799617985e-06, 'learning_rate/action_model': 5.7364132799617984e-05, 'epoch': 0.92}) +[2026-06-16 01:52:28,879][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.22322259843349457, 'train/grad_norm_pre_clip': 6.093856334686279, 'timing/data': 0.012223862111568451, 'timing/model': 1.5570718944072723, 'learning_rate/qwen_vl_interface': 5.733383768478923e-06, 'learning_rate/action_model': 5.7333837684789225e-05, 'epoch': 0.93}) +[2026-06-16 01:52:35,195][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.08181895315647125, 'train/grad_norm_pre_clip': 5.312394618988037, 'timing/data': 0.01285560429096222, 'timing/model': 1.5357668064534664, 'learning_rate/qwen_vl_interface': 5.730354058295234e-06, 'learning_rate/action_model': 5.730354058295234e-05, 'epoch': 0.93}) +[2026-06-16 01:52:41,481][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.008037340827286243, 'train/grad_norm_pre_clip': 2.7193801403045654, 'timing/data': 0.012475261464715004, 'timing/model': 1.6045401152223349, 'learning_rate/qwen_vl_interface': 5.72732415065614e-06, 'learning_rate/action_model': 5.727324150656139e-05, 'epoch': 0.93}) +[2026-06-16 01:52:47,790][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.22409726679325104, 'train/grad_norm_pre_clip': 15.139004707336426, 'timing/data': 0.012532886117696762, 'timing/model': 1.7499150522053242, 'learning_rate/qwen_vl_interface': 5.724294046807115e-06, 'learning_rate/action_model': 5.724294046807115e-05, 'epoch': 0.93}) +[2026-06-16 01:52:53,887][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.07529886066913605, 'train/grad_norm_pre_clip': 16.535667419433594, 'timing/data': 0.013039765879511833, 'timing/model': 1.5594991222023964, 'learning_rate/qwen_vl_interface': 5.7212637479937234e-06, 'learning_rate/action_model': 5.721263747993724e-05, 'epoch': 0.93}) +[2026-06-16 01:53:00,568][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.08180831372737885, 'train/grad_norm_pre_clip': 7.899681091308594, 'timing/data': 0.01378406397998333, 'timing/model': 1.6495088692754507, 'learning_rate/qwen_vl_interface': 5.718233255461607e-06, 'learning_rate/action_model': 5.718233255461607e-05, 'epoch': 0.93}) +[2026-06-16 01:53:06,870][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.40035903453826904, 'train/grad_norm_pre_clip': 12.499462127685547, 'timing/data': 0.01276877149939537, 'timing/model': 1.5545210018754005, 'learning_rate/qwen_vl_interface': 5.715202570456486e-06, 'learning_rate/action_model': 5.715202570456486e-05, 'epoch': 0.93}) +[2026-06-16 01:53:13,364][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.08316448330879211, 'train/grad_norm_pre_clip': 11.804555892944336, 'timing/data': 0.013461336493492126, 'timing/model': 1.701382253319025, 'learning_rate/qwen_vl_interface': 5.7121716942241615e-06, 'learning_rate/action_model': 5.712171694224161e-05, 'epoch': 0.93}) +[2026-06-16 01:53:19,674][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.09292903542518616, 'train/grad_norm_pre_clip': 21.73636817932129, 'timing/data': 0.012957144528627396, 'timing/model': 1.5440408885478973, 'learning_rate/qwen_vl_interface': 5.709140628010509e-06, 'learning_rate/action_model': 5.7091406280105085e-05, 'epoch': 0.93}) +[2026-06-16 02:00:37,969][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.13989871740341187, 'train/grad_norm_pre_clip': 19.94243812561035, 'eval/loss': 0.055620573461055756, 'eval/flap_precision': 0.9559834938101788, 'eval/flap_recall': 0.9430122116689281, 'eval/flap_f1': 0.9494535519125683, 'eval/flap_rate_model': 0.20194444444444445, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9494535519125684, 'timing/data': 0.01362922415137291, 'timing/model': 1.713436234742403, 'learning_rate/qwen_vl_interface': 5.706109373061489e-06, 'learning_rate/action_model': 5.706109373061489e-05, 'epoch': 0.93}) +[2026-06-16 02:00:37,970][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2400_state +[2026-06-16 02:00:37,974][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:00:43,790][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2400_state/model.safetensors +[2026-06-16 02:00:56,989][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2400_state/optimizer.bin +[2026-06-16 02:00:56,990][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2400_state/sampler.bin +[2026-06-16 02:00:56,990][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2400_state/sampler_1.bin +[2026-06-16 02:00:56,991][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2400_state/random_states_0.pkl +[2026-06-16 02:00:57,094][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2400 +[2026-06-16 02:00:57,094][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:00:57,104][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:01:03,425][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.10920947790145874, 'train/grad_norm_pre_clip': 18.202533721923828, 'timing/data': 0.01078776828944683, 'timing/model': 1.6581780556589365, 'learning_rate/qwen_vl_interface': 5.703077930623134e-06, 'learning_rate/action_model': 5.7030779306231344e-05, 'epoch': 0.93}) +[2026-06-16 02:01:09,164][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.08719678223133087, 'train/grad_norm_pre_clip': 18.92455291748047, 'timing/data': 0.011202948167920113, 'timing/model': 1.4508983362466097, 'learning_rate/qwen_vl_interface': 5.7000463019415565e-06, 'learning_rate/action_model': 5.7000463019415563e-05, 'epoch': 0.93}) +[2026-06-16 02:01:15,099][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.05676058679819107, 'train/grad_norm_pre_clip': 8.583874702453613, 'timing/data': 0.011034024879336357, 'timing/model': 1.4526299647986889, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 0.93}) +[2026-06-16 02:01:21,002][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.028431322425603867, 'train/grad_norm_pre_clip': 9.380537986755371, 'timing/data': 0.011477667838335037, 'timing/model': 1.4496757127344608, 'learning_rate/qwen_vl_interface': 5.693982490833561e-06, 'learning_rate/action_model': 5.693982490833561e-05, 'epoch': 0.93}) +[2026-06-16 02:01:26,773][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.02637570910155773, 'train/grad_norm_pre_clip': 2.1816213130950928, 'timing/data': 0.011000849306583405, 'timing/model': 1.4533290322870016, 'learning_rate/qwen_vl_interface': 5.690950310899745e-06, 'learning_rate/action_model': 5.6909503108997456e-05, 'epoch': 0.93}) +[2026-06-16 02:01:32,566][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.10733494162559509, 'train/grad_norm_pre_clip': 6.69575309753418, 'timing/data': 0.01064666174352169, 'timing/model': 1.410533081740141, 'learning_rate/qwen_vl_interface': 5.687917949707913e-06, 'learning_rate/action_model': 5.687917949707913e-05, 'epoch': 0.93}) +[2026-06-16 02:01:38,432][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.25316351652145386, 'train/grad_norm_pre_clip': 15.206171035766602, 'timing/data': 0.017771661281585693, 'timing/model': 1.6056854762136936, 'learning_rate/qwen_vl_interface': 5.684885408504555e-06, 'learning_rate/action_model': 5.684885408504554e-05, 'epoch': 0.93}) +[2026-06-16 02:01:44,072][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.21399228274822235, 'train/grad_norm_pre_clip': 14.0872802734375, 'timing/data': 0.011026065796613693, 'timing/model': 1.4520322643220425, 'learning_rate/qwen_vl_interface': 5.681852688536231e-06, 'learning_rate/action_model': 5.681852688536231e-05, 'epoch': 0.93}) +[2026-06-16 02:01:49,974][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.013044769875705242, 'train/grad_norm_pre_clip': 1.1055991649627686, 'timing/data': 0.01101464033126831, 'timing/model': 1.4539756756275892, 'learning_rate/qwen_vl_interface': 5.67881979104958e-06, 'learning_rate/action_model': 5.67881979104958e-05, 'epoch': 0.93}) +[2026-06-16 02:01:55,882][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.03513287752866745, 'train/grad_norm_pre_clip': 4.995228290557861, 'timing/data': 0.010760152712464333, 'timing/model': 1.4600971061736345, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 0.93}) +[2026-06-16 02:02:01,779][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.056244708597660065, 'train/grad_norm_pre_clip': 6.346026420593262, 'timing/data': 0.011091914027929306, 'timing/model': 1.4523363560438156, 'learning_rate/qwen_vl_interface': 5.672753468508204e-06, 'learning_rate/action_model': 5.672753468508204e-05, 'epoch': 0.93}) +[2026-06-16 02:02:07,587][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.010927719995379448, 'train/grad_norm_pre_clip': 1.5281028747558594, 'timing/data': 0.010658018290996552, 'timing/model': 1.4117690958082676, 'learning_rate/qwen_vl_interface': 5.6697200459471165e-06, 'learning_rate/action_model': 5.669720045947117e-05, 'epoch': 0.93}) +[2026-06-16 02:02:13,477][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.06868034601211548, 'train/grad_norm_pre_clip': 8.393495559692383, 'timing/data': 0.010371293872594833, 'timing/model': 1.6474376861006021, 'learning_rate/qwen_vl_interface': 5.666686450854973e-06, 'learning_rate/action_model': 5.666686450854973e-05, 'epoch': 0.93}) +[2026-06-16 02:02:19,181][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.010012433864176273, 'train/grad_norm_pre_clip': 3.0736732482910156, 'timing/data': 0.011699369177222252, 'timing/model': 1.450931129977107, 'learning_rate/qwen_vl_interface': 5.663652684478768e-06, 'learning_rate/action_model': 5.6636526844787676e-05, 'epoch': 0.93}) +[2026-06-16 02:02:25,279][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.17470449209213257, 'train/grad_norm_pre_clip': 10.363280296325684, 'timing/data': 0.00963592529296875, 'timing/model': 1.451501740142703, 'learning_rate/qwen_vl_interface': 5.660618748065569e-06, 'learning_rate/action_model': 5.6606187480655695e-05, 'epoch': 0.93}) +[2026-06-16 02:02:31,172][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.26057639718055725, 'train/grad_norm_pre_clip': 9.975396156311035, 'timing/data': 0.01130468025803566, 'timing/model': 1.4602008685469627, 'learning_rate/qwen_vl_interface': 5.657584642862514e-06, 'learning_rate/action_model': 5.657584642862513e-05, 'epoch': 0.94}) +[2026-06-16 02:02:36,976][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.1538405418395996, 'train/grad_norm_pre_clip': 13.466291427612305, 'timing/data': 0.01012241281569004, 'timing/model': 1.4539405908435583, 'learning_rate/qwen_vl_interface': 5.654550370116808e-06, 'learning_rate/action_model': 5.654550370116808e-05, 'epoch': 0.94}) +[2026-06-16 02:02:42,876][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.020811185240745544, 'train/grad_norm_pre_clip': 1.8251615762710571, 'timing/data': 0.010060256347060204, 'timing/model': 1.475001873448491, 'learning_rate/qwen_vl_interface': 5.651515931075726e-06, 'learning_rate/action_model': 5.651515931075726e-05, 'epoch': 0.94}) +[2026-06-16 02:02:48,769][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.17855329811573029, 'train/grad_norm_pre_clip': 11.173192024230957, 'timing/data': 0.010422153398394585, 'timing/model': 1.6421384196728468, 'learning_rate/qwen_vl_interface': 5.6484813269866126e-06, 'learning_rate/action_model': 5.648481326986612e-05, 'epoch': 0.94}) +[2026-06-16 02:02:54,476][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.08917804807424545, 'train/grad_norm_pre_clip': 13.437854766845703, 'timing/data': 0.011213099583983421, 'timing/model': 1.4498018100857735, 'learning_rate/qwen_vl_interface': 5.645446559096875e-06, 'learning_rate/action_model': 5.645446559096875e-05, 'epoch': 0.94}) +[2026-06-16 02:03:00,382][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.12808647751808167, 'train/grad_norm_pre_clip': 7.333944797515869, 'timing/data': 0.01149066537618637, 'timing/model': 1.453784890472889, 'learning_rate/qwen_vl_interface': 5.6424116286539965e-06, 'learning_rate/action_model': 5.642411628653996e-05, 'epoch': 0.94}) +[2026-06-16 02:03:06,277][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.20187297463417053, 'train/grad_norm_pre_clip': 22.566213607788086, 'timing/data': 0.010753439739346504, 'timing/model': 1.451666221022606, 'learning_rate/qwen_vl_interface': 5.639376536905517e-06, 'learning_rate/action_model': 5.6393765369055175e-05, 'epoch': 0.94}) +[2026-06-16 02:03:12,177][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.20814001560211182, 'train/grad_norm_pre_clip': 12.496309280395508, 'timing/data': 0.010649232193827629, 'timing/model': 1.4425276014953852, 'learning_rate/qwen_vl_interface': 5.636341285099055e-06, 'learning_rate/action_model': 5.636341285099055e-05, 'epoch': 0.94}) +[2026-06-16 02:03:18,143][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.10457400232553482, 'train/grad_norm_pre_clip': 15.226975440979004, 'timing/data': 0.012822551652789116, 'timing/model': 1.4971472844481468, 'learning_rate/qwen_vl_interface': 5.6333058744822835e-06, 'learning_rate/action_model': 5.633305874482283e-05, 'epoch': 0.94}) +[2026-06-16 02:03:24,290][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.0053981002420187, 'train/grad_norm_pre_clip': 0.3217175006866455, 'timing/data': 0.013024820014834404, 'timing/model': 1.7720050271600485, 'learning_rate/qwen_vl_interface': 5.630270306302944e-06, 'learning_rate/action_model': 5.6302703063029436e-05, 'epoch': 0.94}) +[2026-06-16 02:03:30,238][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.4131183922290802, 'train/grad_norm_pre_clip': 11.117477416992188, 'timing/data': 0.013745913282036781, 'timing/model': 1.5134775079786777, 'learning_rate/qwen_vl_interface': 5.627234581808848e-06, 'learning_rate/action_model': 5.627234581808848e-05, 'epoch': 0.94}) +[2026-06-16 02:03:36,669][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.022737210616469383, 'train/grad_norm_pre_clip': 3.243173599243164, 'timing/data': 0.012646500021219254, 'timing/model': 1.5379741191864014, 'learning_rate/qwen_vl_interface': 5.624198702247863e-06, 'learning_rate/action_model': 5.624198702247863e-05, 'epoch': 0.94}) +[2026-06-16 02:03:42,969][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.10486399382352829, 'train/grad_norm_pre_clip': 11.447515487670898, 'timing/data': 0.013021910563111305, 'timing/model': 1.517918311059475, 'learning_rate/qwen_vl_interface': 5.621162668867925e-06, 'learning_rate/action_model': 5.621162668867925e-05, 'epoch': 0.94}) +[2026-06-16 02:03:49,352][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.07478750497102737, 'train/grad_norm_pre_clip': 3.838554859161377, 'timing/data': 0.012518145143985748, 'timing/model': 1.5680381916463375, 'learning_rate/qwen_vl_interface': 5.618126482917037e-06, 'learning_rate/action_model': 5.618126482917036e-05, 'epoch': 0.94}) +[2026-06-16 02:03:55,634][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.24194854497909546, 'train/grad_norm_pre_clip': 14.866304397583008, 'timing/data': 0.011456452310085297, 'timing/model': 1.5045265350490808, 'learning_rate/qwen_vl_interface': 5.6150901456432546e-06, 'learning_rate/action_model': 5.615090145643255e-05, 'epoch': 0.94}) +[2026-06-16 02:04:01,758][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.05774853006005287, 'train/grad_norm_pre_clip': 13.028217315673828, 'timing/data': 0.012874383479356766, 'timing/model': 1.7618699446320534, 'learning_rate/qwen_vl_interface': 5.612053658294705e-06, 'learning_rate/action_model': 5.6120536582947045e-05, 'epoch': 0.94}) +[2026-06-16 02:04:07,637][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.12885227799415588, 'train/grad_norm_pre_clip': 6.802921772003174, 'timing/data': 0.012446394190192223, 'timing/model': 1.4976436253637075, 'learning_rate/qwen_vl_interface': 5.609017022119572e-06, 'learning_rate/action_model': 5.609017022119573e-05, 'epoch': 0.94}) +[2026-06-16 02:04:13,771][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.004598428029567003, 'train/grad_norm_pre_clip': 0.37835848331451416, 'timing/data': 0.011907437816262245, 'timing/model': 1.5657075606286526, 'learning_rate/qwen_vl_interface': 5.605980238366104e-06, 'learning_rate/action_model': 5.605980238366103e-05, 'epoch': 0.94}) +[2026-06-16 02:04:20,171][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.023499753326177597, 'train/grad_norm_pre_clip': 3.1378872394561768, 'timing/data': 0.01338709145784378, 'timing/model': 1.556102318689227, 'learning_rate/qwen_vl_interface': 5.602943308282606e-06, 'learning_rate/action_model': 5.602943308282605e-05, 'epoch': 0.94}) +[2026-06-16 02:04:26,568][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.01971627026796341, 'train/grad_norm_pre_clip': 1.8189716339111328, 'timing/data': 0.0139449592679739, 'timing/model': 1.5170166678726673, 'learning_rate/qwen_vl_interface': 5.599906233117445e-06, 'learning_rate/action_model': 5.5999062331174446e-05, 'epoch': 0.94}) +[2026-06-16 02:04:32,982][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.08400363475084305, 'train/grad_norm_pre_clip': 8.34037971496582, 'timing/data': 0.012341707944869995, 'timing/model': 1.5758764035999775, 'learning_rate/qwen_vl_interface': 5.596869014119048e-06, 'learning_rate/action_model': 5.596869014119048e-05, 'epoch': 0.94}) +[2026-06-16 02:04:39,300][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.057522110641002655, 'train/grad_norm_pre_clip': 4.33018684387207, 'timing/data': 0.012750398367643356, 'timing/model': 1.7498744688928127, 'learning_rate/qwen_vl_interface': 5.593831652535902e-06, 'learning_rate/action_model': 5.593831652535902e-05, 'epoch': 0.94}) +[2026-06-16 02:04:45,372][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.08533783257007599, 'train/grad_norm_pre_clip': 9.539327621459961, 'timing/data': 0.012812856584787369, 'timing/model': 1.5244386997073889, 'learning_rate/qwen_vl_interface': 5.590794149616552e-06, 'learning_rate/action_model': 5.590794149616552e-05, 'epoch': 0.94}) +[2026-06-16 02:04:51,638][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.21344758570194244, 'train/grad_norm_pre_clip': 5.9544219970703125, 'timing/data': 0.012475375086069107, 'timing/model': 1.5495891086757183, 'learning_rate/qwen_vl_interface': 5.587756506609601e-06, 'learning_rate/action_model': 5.5877565066096006e-05, 'epoch': 0.94}) +[2026-06-16 02:04:57,755][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.2288188934326172, 'train/grad_norm_pre_clip': 7.530545234680176, 'timing/data': 0.012873934581875801, 'timing/model': 1.525152424350381, 'learning_rate/qwen_vl_interface': 5.584718724763707e-06, 'learning_rate/action_model': 5.584718724763707e-05, 'epoch': 0.94}) +[2026-06-16 02:05:03,951][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.10520687699317932, 'train/grad_norm_pre_clip': 10.249605178833008, 'timing/data': 0.012699050828814507, 'timing/model': 1.5403776075690985, 'learning_rate/qwen_vl_interface': 5.581680805327588e-06, 'learning_rate/action_model': 5.5816808053275885e-05, 'epoch': 0.94}) +[2026-06-16 02:05:10,271][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.023486303165555, 'train/grad_norm_pre_clip': 4.011368751525879, 'timing/data': 0.012128418311476707, 'timing/model': 1.690044928342104, 'learning_rate/qwen_vl_interface': 5.578642749550019e-06, 'learning_rate/action_model': 5.578642749550019e-05, 'epoch': 0.95}) +[2026-06-16 02:05:16,614][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.01689271815121174, 'train/grad_norm_pre_clip': 2.865048408508301, 'timing/data': 0.013230700045824051, 'timing/model': 1.7602185383439064, 'learning_rate/qwen_vl_interface': 5.575604558679831e-06, 'learning_rate/action_model': 5.57560455867983e-05, 'epoch': 0.95}) +[2026-06-16 02:05:22,792][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.09016021341085434, 'train/grad_norm_pre_clip': 3.761014461517334, 'timing/data': 0.013165896758437157, 'timing/model': 1.5659771002829075, 'learning_rate/qwen_vl_interface': 5.5725662339659085e-06, 'learning_rate/action_model': 5.572566233965908e-05, 'epoch': 0.95}) +[2026-06-16 02:05:29,682][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.08287076652050018, 'train/grad_norm_pre_clip': 9.061240196228027, 'timing/data': 0.010719714686274529, 'timing/model': 1.5421828646212816, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 0.95}) +[2026-06-16 02:05:36,178][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.17006471753120422, 'train/grad_norm_pre_clip': 12.599437713623047, 'timing/data': 0.01416841708123684, 'timing/model': 1.5599705800414085, 'learning_rate/qwen_vl_interface': 5.566489188002671e-06, 'learning_rate/action_model': 5.56648918800267e-05, 'epoch': 0.95}) +[2026-06-16 02:05:42,474][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.2245522141456604, 'train/grad_norm_pre_clip': 19.585561752319336, 'timing/data': 0.011732900515198708, 'timing/model': 1.5529334545135498, 'learning_rate/qwen_vl_interface': 5.563450469251403e-06, 'learning_rate/action_model': 5.563450469251402e-05, 'epoch': 0.95}) +[2026-06-16 02:05:48,859][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.2027561366558075, 'train/grad_norm_pre_clip': 12.851744651794434, 'timing/data': 0.012594686821103096, 'timing/model': 1.5892993248999119, 'learning_rate/qwen_vl_interface': 5.560411621652488e-06, 'learning_rate/action_model': 5.560411621652488e-05, 'epoch': 0.95}) +[2026-06-16 02:05:55,108][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.14213351905345917, 'train/grad_norm_pre_clip': 38.583866119384766, 'timing/data': 0.015250956639647484, 'timing/model': 1.77634122595191, 'learning_rate/qwen_vl_interface': 5.557372646455081e-06, 'learning_rate/action_model': 5.55737264645508e-05, 'epoch': 0.95}) +[2026-06-16 02:06:01,404][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.17334945499897003, 'train/grad_norm_pre_clip': 9.87766170501709, 'timing/data': 0.012878874316811562, 'timing/model': 1.7791532836854458, 'learning_rate/qwen_vl_interface': 5.554333544908388e-06, 'learning_rate/action_model': 5.5543335449083875e-05, 'epoch': 0.95}) +[2026-06-16 02:06:07,606][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.26021313667297363, 'train/grad_norm_pre_clip': 19.558656692504883, 'timing/data': 0.013706283643841743, 'timing/model': 1.5518059432506561, 'learning_rate/qwen_vl_interface': 5.551294318261671e-06, 'learning_rate/action_model': 5.5512943182616705e-05, 'epoch': 0.95}) +[2026-06-16 02:06:14,012][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.05186733230948448, 'train/grad_norm_pre_clip': 4.983249187469482, 'timing/data': 0.016380079090595245, 'timing/model': 1.5534937679767609, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 0.95}) +[2026-06-16 02:06:20,514][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.021029334515333176, 'train/grad_norm_pre_clip': 2.4525747299194336, 'timing/data': 0.015346847474575043, 'timing/model': 1.6325714811682701, 'learning_rate/qwen_vl_interface': 5.545215494665458e-06, 'learning_rate/action_model': 5.545215494665458e-05, 'epoch': 0.95}) +[2026-06-16 02:06:26,997][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.045988310128450394, 'train/grad_norm_pre_clip': 9.041716575622559, 'timing/data': 0.01876373589038849, 'timing/model': 1.5485118981450796, 'learning_rate/qwen_vl_interface': 5.5421759002147376e-06, 'learning_rate/action_model': 5.542175900214738e-05, 'epoch': 0.95}) +[2026-06-16 02:06:33,319][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.031117763370275497, 'train/grad_norm_pre_clip': 8.141568183898926, 'timing/data': 0.012871267274022102, 'timing/model': 1.8098848331719637, 'learning_rate/qwen_vl_interface': 5.53913618566154e-06, 'learning_rate/action_model': 5.53913618566154e-05, 'epoch': 0.95}) +[2026-06-16 02:06:39,337][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.26301947236061096, 'train/grad_norm_pre_clip': 12.095132827758789, 'timing/data': 0.012253133580088615, 'timing/model': 1.5595178846269846, 'learning_rate/qwen_vl_interface': 5.536096352255381e-06, 'learning_rate/action_model': 5.536096352255381e-05, 'epoch': 0.95}) +[2026-06-16 02:06:45,441][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.04656195268034935, 'train/grad_norm_pre_clip': 7.887118339538574, 'timing/data': 0.012053845450282097, 'timing/model': 1.5363485980778933, 'learning_rate/qwen_vl_interface': 5.533056401245817e-06, 'learning_rate/action_model': 5.5330564012458164e-05, 'epoch': 0.95}) +[2026-06-16 02:06:51,553][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.0909792110323906, 'train/grad_norm_pre_clip': 4.24714469909668, 'timing/data': 0.012830065563321114, 'timing/model': 1.482601199299097, 'learning_rate/qwen_vl_interface': 5.530016333882456e-06, 'learning_rate/action_model': 5.5300163338824565e-05, 'epoch': 0.95}) +[2026-06-16 02:06:57,826][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.17557349801063538, 'train/grad_norm_pre_clip': 4.787358283996582, 'timing/data': 0.012447604909539223, 'timing/model': 1.5492112636566162, 'learning_rate/qwen_vl_interface': 5.52697615141496e-06, 'learning_rate/action_model': 5.52697615141496e-05, 'epoch': 0.95}) +[2026-06-16 02:07:04,041][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.1179940402507782, 'train/grad_norm_pre_clip': 5.115443229675293, 'timing/data': 0.012582104653120041, 'timing/model': 1.5474982354789972, 'learning_rate/qwen_vl_interface': 5.52393585509303e-06, 'learning_rate/action_model': 5.52393585509303e-05, 'epoch': 0.95}) +[2026-06-16 02:07:10,120][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.04812634736299515, 'train/grad_norm_pre_clip': 7.2849345207214355, 'timing/data': 0.012072889134287834, 'timing/model': 1.699892483651638, 'learning_rate/qwen_vl_interface': 5.520895446166418e-06, 'learning_rate/action_model': 5.5208954461664175e-05, 'epoch': 0.95}) +[2026-06-16 02:07:16,041][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.17532116174697876, 'train/grad_norm_pre_clip': 9.155553817749023, 'timing/data': 0.012796906754374504, 'timing/model': 1.5135121829807758, 'learning_rate/qwen_vl_interface': 5.517854925884919e-06, 'learning_rate/action_model': 5.517854925884919e-05, 'epoch': 0.95}) +[2026-06-16 02:07:22,341][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.08733116090297699, 'train/grad_norm_pre_clip': 7.731072425842285, 'timing/data': 0.011973470449447632, 'timing/model': 1.5301548596471548, 'learning_rate/qwen_vl_interface': 5.514814295498379e-06, 'learning_rate/action_model': 5.514814295498378e-05, 'epoch': 0.95}) +[2026-06-16 02:07:28,551][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.07717646658420563, 'train/grad_norm_pre_clip': 4.9626874923706055, 'timing/data': 0.012326743453741074, 'timing/model': 1.5443289279937744, 'learning_rate/qwen_vl_interface': 5.5117735562566846e-06, 'learning_rate/action_model': 5.5117735562566844e-05, 'epoch': 0.95}) +[2026-06-16 02:07:35,113][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.10849674046039581, 'train/grad_norm_pre_clip': 7.393424987792969, 'timing/data': 0.01753612793982029, 'timing/model': 1.600109675899148, 'learning_rate/qwen_vl_interface': 5.5087327094097685e-06, 'learning_rate/action_model': 5.508732709409769e-05, 'epoch': 0.95}) +[2026-06-16 02:07:41,573][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.12110768258571625, 'train/grad_norm_pre_clip': 5.532446384429932, 'timing/data': 0.013344883918762207, 'timing/model': 1.5394900869578123, 'learning_rate/qwen_vl_interface': 5.505691756207609e-06, 'learning_rate/action_model': 5.505691756207609e-05, 'epoch': 0.95}) +[2026-06-16 02:07:47,855][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.0174628347158432, 'train/grad_norm_pre_clip': 2.646564245223999, 'timing/data': 0.012302275747060776, 'timing/model': 1.703119020909071, 'learning_rate/qwen_vl_interface': 5.502650697900228e-06, 'learning_rate/action_model': 5.502650697900228e-05, 'epoch': 0.95}) +[2026-06-16 02:07:53,867][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.02165355533361435, 'train/grad_norm_pre_clip': 1.9558888673782349, 'timing/data': 0.012580668553709984, 'timing/model': 1.5262115262448788, 'learning_rate/qwen_vl_interface': 5.499609535737688e-06, 'learning_rate/action_model': 5.4996095357376875e-05, 'epoch': 0.96}) +[2026-06-16 02:08:00,174][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.0571797676384449, 'train/grad_norm_pre_clip': 4.530552387237549, 'timing/data': 0.012792147696018219, 'timing/model': 1.5296405367553234, 'learning_rate/qwen_vl_interface': 5.496568270970096e-06, 'learning_rate/action_model': 5.496568270970096e-05, 'epoch': 0.96}) +[2026-06-16 02:08:06,443][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.21671274304389954, 'train/grad_norm_pre_clip': 12.093864440917969, 'timing/data': 0.01162375696003437, 'timing/model': 1.5554753113538027, 'learning_rate/qwen_vl_interface': 5.4935269048476015e-06, 'learning_rate/action_model': 5.493526904847601e-05, 'epoch': 0.96}) +[2026-06-16 02:08:12,540][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.2699478268623352, 'train/grad_norm_pre_clip': 13.557921409606934, 'timing/data': 0.01194746419787407, 'timing/model': 1.4970749039202929, 'learning_rate/qwen_vl_interface': 5.490485438620397e-06, 'learning_rate/action_model': 5.4904854386203965e-05, 'epoch': 0.96}) +[2026-06-16 02:08:18,678][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.08284521847963333, 'train/grad_norm_pre_clip': 5.761648654937744, 'timing/data': 0.012088214978575706, 'timing/model': 1.772959453985095, 'learning_rate/qwen_vl_interface': 5.487443873538712e-06, 'learning_rate/action_model': 5.4874438735387115e-05, 'epoch': 0.96}) +[2026-06-16 02:08:24,641][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.03715631738305092, 'train/grad_norm_pre_clip': 3.949687957763672, 'timing/data': 0.012588180601596832, 'timing/model': 1.5428387895226479, 'learning_rate/qwen_vl_interface': 5.484402210852819e-06, 'learning_rate/action_model': 5.484402210852818e-05, 'epoch': 0.96}) +[2026-06-16 02:08:30,840][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.034452442079782486, 'train/grad_norm_pre_clip': 1.9692370891571045, 'timing/data': 0.013157419860363007, 'timing/model': 1.5571902077645063, 'learning_rate/qwen_vl_interface': 5.48136045181303e-06, 'learning_rate/action_model': 5.48136045181303e-05, 'epoch': 0.96}) +[2026-06-16 02:08:36,836][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.1239943876862526, 'train/grad_norm_pre_clip': 12.41335391998291, 'timing/data': 0.012754363939166069, 'timing/model': 1.5254604388028383, 'learning_rate/qwen_vl_interface': 5.478318597669699e-06, 'learning_rate/action_model': 5.4783185976696995e-05, 'epoch': 0.96}) +[2026-06-16 02:08:42,942][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.12278881669044495, 'train/grad_norm_pre_clip': 9.622188568115234, 'timing/data': 0.011823184788227081, 'timing/model': 1.5504138115793467, 'learning_rate/qwen_vl_interface': 5.475276649673218e-06, 'learning_rate/action_model': 5.4752766496732174e-05, 'epoch': 0.96}) +[2026-06-16 02:08:49,276][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.1013890728354454, 'train/grad_norm_pre_clip': 12.051445007324219, 'timing/data': 0.012429488822817802, 'timing/model': 1.6984398141503334, 'learning_rate/qwen_vl_interface': 5.472234609074014e-06, 'learning_rate/action_model': 5.472234609074014e-05, 'epoch': 0.96}) +[2026-06-16 02:08:55,629][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.08329007029533386, 'train/grad_norm_pre_clip': 7.592592716217041, 'timing/data': 0.013097820803523064, 'timing/model': 1.805891951546073, 'learning_rate/qwen_vl_interface': 5.46919247712256e-06, 'learning_rate/action_model': 5.46919247712256e-05, 'epoch': 0.96}) +[2026-06-16 02:09:01,743][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.04180655628442764, 'train/grad_norm_pre_clip': 4.841804027557373, 'timing/data': 0.0125910434871912, 'timing/model': 1.507770661264658, 'learning_rate/qwen_vl_interface': 5.466150255069356e-06, 'learning_rate/action_model': 5.4661502550693555e-05, 'epoch': 0.96}) +[2026-06-16 02:09:07,944][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.33594846725463867, 'train/grad_norm_pre_clip': 8.231362342834473, 'timing/data': 0.01323457807302475, 'timing/model': 1.5619741845875978, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 0.96}) +[2026-06-16 02:09:14,194][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.13899242877960205, 'train/grad_norm_pre_clip': 8.005581855773926, 'timing/data': 0.01455005630850792, 'timing/model': 1.5554999094456434, 'learning_rate/qwen_vl_interface': 5.46006554565991e-06, 'learning_rate/action_model': 5.4600655456599105e-05, 'epoch': 0.96}) +[2026-06-16 02:09:20,646][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.0312509760260582, 'train/grad_norm_pre_clip': 5.247705936431885, 'timing/data': 0.015432815998792648, 'timing/model': 1.5815030075609684, 'learning_rate/qwen_vl_interface': 5.457023060804865e-06, 'learning_rate/action_model': 5.457023060804864e-05, 'epoch': 0.96}) +[2026-06-16 02:09:27,038][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.26199421286582947, 'train/grad_norm_pre_clip': 13.712793350219727, 'timing/data': 0.01900065317749977, 'timing/model': 1.5748667120933533, 'learning_rate/qwen_vl_interface': 5.453980490850459e-06, 'learning_rate/action_model': 5.453980490850459e-05, 'epoch': 0.96}) +[2026-06-16 02:09:33,502][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.024547308683395386, 'train/grad_norm_pre_clip': 1.9069994688034058, 'timing/data': 0.013596823439002037, 'timing/model': 1.7549356743693352, 'learning_rate/qwen_vl_interface': 5.4509378370473785e-06, 'learning_rate/action_model': 5.450937837047378e-05, 'epoch': 0.96}) +[2026-06-16 02:09:39,574][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.20148903131484985, 'train/grad_norm_pre_clip': 11.749991416931152, 'timing/data': 0.01240803487598896, 'timing/model': 1.5537933483719826, 'learning_rate/qwen_vl_interface': 5.4478951006463424e-06, 'learning_rate/action_model': 5.447895100646342e-05, 'epoch': 0.96}) +[2026-06-16 02:09:45,873][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.036428045481443405, 'train/grad_norm_pre_clip': 2.5352115631103516, 'timing/data': 0.0121691282838583, 'timing/model': 1.5427043233066797, 'learning_rate/qwen_vl_interface': 5.444852282898106e-06, 'learning_rate/action_model': 5.444852282898106e-05, 'epoch': 0.96}) +[2026-06-16 02:09:52,422][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.11641474813222885, 'train/grad_norm_pre_clip': 8.518537521362305, 'timing/data': 0.019035104662179947, 'timing/model': 1.623765967786312, 'learning_rate/qwen_vl_interface': 5.441809385053456e-06, 'learning_rate/action_model': 5.4418093850534555e-05, 'epoch': 0.96}) +[2026-06-16 02:09:58,670][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.1204986646771431, 'train/grad_norm_pre_clip': 10.949007987976074, 'timing/data': 0.011893901973962784, 'timing/model': 1.520465089008212, 'learning_rate/qwen_vl_interface': 5.438766408363214e-06, 'learning_rate/action_model': 5.438766408363214e-05, 'epoch': 0.96}) +[2026-06-16 02:10:04,974][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.015567594207823277, 'train/grad_norm_pre_clip': 2.149815320968628, 'timing/data': 0.012067044153809547, 'timing/model': 1.5377828516066074, 'learning_rate/qwen_vl_interface': 5.43572335407823e-06, 'learning_rate/action_model': 5.43572335407823e-05, 'epoch': 0.96}) +[2026-06-16 02:10:11,512][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.03302085027098656, 'train/grad_norm_pre_clip': 6.872976779937744, 'timing/data': 0.013261871412396431, 'timing/model': 1.7933431193232536, 'learning_rate/qwen_vl_interface': 5.432680223449392e-06, 'learning_rate/action_model': 5.4326802234493924e-05, 'epoch': 0.96}) +[2026-06-16 02:10:17,575][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.10915986448526382, 'train/grad_norm_pre_clip': 9.53903865814209, 'timing/data': 0.012059647589921951, 'timing/model': 1.550909023731947, 'learning_rate/qwen_vl_interface': 5.429637017727614e-06, 'learning_rate/action_model': 5.4296370177276144e-05, 'epoch': 0.96}) +[2026-06-16 02:10:23,902][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.023972727358341217, 'train/grad_norm_pre_clip': 5.677492618560791, 'timing/data': 0.011912178248167038, 'timing/model': 1.5437142625451088, 'learning_rate/qwen_vl_interface': 5.426593738163845e-06, 'learning_rate/action_model': 5.4265937381638446e-05, 'epoch': 0.96}) +[2026-06-16 02:10:30,472][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.1259928047657013, 'train/grad_norm_pre_clip': 8.781405448913574, 'timing/data': 0.012831896543502808, 'timing/model': 1.7570781391113997, 'learning_rate/qwen_vl_interface': 5.42355038600906e-06, 'learning_rate/action_model': 5.42355038600906e-05, 'epoch': 0.96}) +[2026-06-16 02:10:36,878][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.07836228609085083, 'train/grad_norm_pre_clip': 5.536607265472412, 'timing/data': 0.013296118006110191, 'timing/model': 1.558385668322444, 'learning_rate/qwen_vl_interface': 5.420506962514267e-06, 'learning_rate/action_model': 5.420506962514267e-05, 'epoch': 0.97}) +[2026-06-16 02:10:43,274][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.12803925573825836, 'train/grad_norm_pre_clip': 13.818011283874512, 'timing/data': 0.014883168041706085, 'timing/model': 1.5984113570302725, 'learning_rate/qwen_vl_interface': 5.4174634689305015e-06, 'learning_rate/action_model': 5.417463468930501e-05, 'epoch': 0.97}) +[2026-06-16 02:10:49,575][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.04417016729712486, 'train/grad_norm_pre_clip': 3.3262698650360107, 'timing/data': 0.013471769168972969, 'timing/model': 1.743404321372509, 'learning_rate/qwen_vl_interface': 5.41441990650883e-06, 'learning_rate/action_model': 5.4144199065088295e-05, 'epoch': 0.97}) +[2026-06-16 02:10:55,646][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.045126549899578094, 'train/grad_norm_pre_clip': 3.9575603008270264, 'timing/data': 0.015257786959409714, 'timing/model': 1.5273291878402233, 'learning_rate/qwen_vl_interface': 5.411376276500345e-06, 'learning_rate/action_model': 5.411376276500345e-05, 'epoch': 0.97}) +[2026-06-16 02:11:01,591][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.01569349691271782, 'train/grad_norm_pre_clip': 2.6378939151763916, 'timing/data': 0.012726152315735817, 'timing/model': 1.4538706336170435, 'learning_rate/qwen_vl_interface': 5.408332580156168e-06, 'learning_rate/action_model': 5.408332580156168e-05, 'epoch': 0.97}) +[2026-06-16 02:11:07,390][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.00785504188388586, 'train/grad_norm_pre_clip': 1.5718008279800415, 'timing/data': 0.010902762413024902, 'timing/model': 1.4653068035840988, 'learning_rate/qwen_vl_interface': 5.40528881872745e-06, 'learning_rate/action_model': 5.40528881872745e-05, 'epoch': 0.97}) +[2026-06-16 02:18:19,141][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.17600686848163605, 'train/grad_norm_pre_clip': 14.004118919372559, 'eval/loss': 0.05921969935297966, 'eval/flap_precision': 0.9541666666666667, 'eval/flap_recall': 0.932157394843962, 'eval/flap_f1': 0.9430336307481125, 'eval/flap_rate_model': 0.2, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9430336307481125, 'timing/data': 0.011158064007759094, 'timing/model': 1.4554000347852707, 'learning_rate/qwen_vl_interface': 5.402244993465364e-06, 'learning_rate/action_model': 5.402244993465364e-05, 'epoch': 0.97}) +[2026-06-16 02:18:19,141][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2500_state +[2026-06-16 02:18:19,145][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:18:25,095][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2500_state/model.safetensors +[2026-06-16 02:18:39,479][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2500_state/optimizer.bin +[2026-06-16 02:18:39,480][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2500_state/sampler.bin +[2026-06-16 02:18:39,480][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-16 02:18:39,481][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-16 02:18:39,567][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-16 02:18:39,567][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:18:39,575][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:18:45,502][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.05840175971388817, 'train/grad_norm_pre_clip': 12.593381881713867, 'timing/data': 0.011178335174918175, 'timing/model': 1.4573351852595806, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 0.97}) +[2026-06-16 02:18:51,558][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.1157289370894432, 'train/grad_norm_pre_clip': 5.11034631729126, 'timing/data': 0.013094857335090637, 'timing/model': 1.532703498378396, 'learning_rate/qwen_vl_interface': 5.396157156445918e-06, 'learning_rate/action_model': 5.396157156445918e-05, 'epoch': 0.97}) +[2026-06-16 02:18:58,059][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.006458217743784189, 'train/grad_norm_pre_clip': 1.0571664571762085, 'timing/data': 0.012058332562446594, 'timing/model': 1.5179747361689806, 'learning_rate/qwen_vl_interface': 5.39311314719104e-06, 'learning_rate/action_model': 5.39311314719104e-05, 'epoch': 0.97}) +[2026-06-16 02:19:04,530][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.05923440679907799, 'train/grad_norm_pre_clip': 2.962573289871216, 'timing/data': 0.014165326952934265, 'timing/model': 1.9231574311852455, 'learning_rate/qwen_vl_interface': 5.390069079107755e-06, 'learning_rate/action_model': 5.390069079107755e-05, 'epoch': 0.97}) +[2026-06-16 02:19:10,557][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.01812814548611641, 'train/grad_norm_pre_clip': 2.689659595489502, 'timing/data': 0.012268545106053352, 'timing/model': 1.5400660801678896, 'learning_rate/qwen_vl_interface': 5.387024953447361e-06, 'learning_rate/action_model': 5.387024953447361e-05, 'epoch': 0.97}) +[2026-06-16 02:19:16,788][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.013920520432293415, 'train/grad_norm_pre_clip': 3.8371963500976562, 'timing/data': 0.011921059340238571, 'timing/model': 1.5659086778759956, 'learning_rate/qwen_vl_interface': 5.383980771461184e-06, 'learning_rate/action_model': 5.3839807714611846e-05, 'epoch': 0.97}) +[2026-06-16 02:19:22,973][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.007001180201768875, 'train/grad_norm_pre_clip': 1.5612238645553589, 'timing/data': 0.012274429202079773, 'timing/model': 1.5508544240146875, 'learning_rate/qwen_vl_interface': 5.380936534400572e-06, 'learning_rate/action_model': 5.3809365344005726e-05, 'epoch': 0.97}) +[2026-06-16 02:19:29,285][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.0025338451378047466, 'train/grad_norm_pre_clip': 0.24704109132289886, 'timing/data': 0.012779047712683678, 'timing/model': 1.5457393117249012, 'learning_rate/qwen_vl_interface': 5.377892243516899e-06, 'learning_rate/action_model': 5.3778922435168985e-05, 'epoch': 0.97}) +[2026-06-16 02:19:35,689][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.3833787143230438, 'train/grad_norm_pre_clip': 30.53288459777832, 'timing/data': 0.013014717027544975, 'timing/model': 1.613438280299306, 'learning_rate/qwen_vl_interface': 5.3748479000615525e-06, 'learning_rate/action_model': 5.374847900061553e-05, 'epoch': 0.97}) +[2026-06-16 02:19:42,060][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.1753571331501007, 'train/grad_norm_pre_clip': 17.475387573242188, 'timing/data': 0.012601133435964584, 'timing/model': 1.7972224559634924, 'learning_rate/qwen_vl_interface': 5.371803505285954e-06, 'learning_rate/action_model': 5.371803505285954e-05, 'epoch': 0.97}) +[2026-06-16 02:19:48,049][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.07366455346345901, 'train/grad_norm_pre_clip': 11.517775535583496, 'timing/data': 0.013898167759180069, 'timing/model': 1.552095839753747, 'learning_rate/qwen_vl_interface': 5.368759060441534e-06, 'learning_rate/action_model': 5.368759060441534e-05, 'epoch': 0.97}) +[2026-06-16 02:19:54,248][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.19416192173957825, 'train/grad_norm_pre_clip': 8.806302070617676, 'timing/data': 0.012254636734724045, 'timing/model': 1.5409275144338608, 'learning_rate/qwen_vl_interface': 5.365714566779749e-06, 'learning_rate/action_model': 5.3657145667797484e-05, 'epoch': 0.97}) +[2026-06-16 02:20:00,446][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.1420898735523224, 'train/grad_norm_pre_clip': 6.448753833770752, 'timing/data': 0.012532398104667664, 'timing/model': 1.5312787294387817, 'learning_rate/qwen_vl_interface': 5.362670025552077e-06, 'learning_rate/action_model': 5.362670025552077e-05, 'epoch': 0.97}) +[2026-06-16 02:20:06,652][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.02983730286359787, 'train/grad_norm_pre_clip': 7.677248001098633, 'timing/data': 0.011662330478429794, 'timing/model': 1.5581087619066238, 'learning_rate/qwen_vl_interface': 5.359625438010013e-06, 'learning_rate/action_model': 5.359625438010013e-05, 'epoch': 0.97}) +[2026-06-16 02:20:12,848][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.007225935813039541, 'train/grad_norm_pre_clip': 1.4566668272018433, 'timing/data': 0.011333661153912544, 'timing/model': 1.510179404169321, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 0.97}) +[2026-06-16 02:20:19,072][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.006659998558461666, 'train/grad_norm_pre_clip': 2.03458309173584, 'timing/data': 0.013473089784383774, 'timing/model': 1.7762942854315042, 'learning_rate/qwen_vl_interface': 5.353536128988791e-06, 'learning_rate/action_model': 5.353536128988791e-05, 'epoch': 0.97}) +[2026-06-16 02:20:25,374][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.3762570321559906, 'train/grad_norm_pre_clip': 18.107328414916992, 'timing/data': 0.01230895146727562, 'timing/model': 1.7749564200639725, 'learning_rate/qwen_vl_interface': 5.350491410012717e-06, 'learning_rate/action_model': 5.350491410012717e-05, 'epoch': 0.97}) +[2026-06-16 02:20:31,452][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.28498831391334534, 'train/grad_norm_pre_clip': 14.81008529663086, 'timing/data': 0.01333252526819706, 'timing/model': 1.5511314645409584, 'learning_rate/qwen_vl_interface': 5.347446649728419e-06, 'learning_rate/action_model': 5.347446649728418e-05, 'epoch': 0.97}) +[2026-06-16 02:20:37,656][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.09997967630624771, 'train/grad_norm_pre_clip': 8.486495018005371, 'timing/data': 0.013289203867316246, 'timing/model': 1.5551671627908945, 'learning_rate/qwen_vl_interface': 5.344401849387484e-06, 'learning_rate/action_model': 5.344401849387484e-05, 'epoch': 0.97}) +[2026-06-16 02:20:43,951][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.3279991149902344, 'train/grad_norm_pre_clip': 23.10478401184082, 'timing/data': 0.012906990945339203, 'timing/model': 1.5500709433108568, 'learning_rate/qwen_vl_interface': 5.3413570102415144e-06, 'learning_rate/action_model': 5.3413570102415144e-05, 'epoch': 0.98}) +[2026-06-16 02:20:50,100][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.2521803081035614, 'train/grad_norm_pre_clip': 15.888933181762695, 'timing/data': 0.010898295789957047, 'timing/model': 1.4339132234454155, 'learning_rate/qwen_vl_interface': 5.338312133542131e-06, 'learning_rate/action_model': 5.3383121335421305e-05, 'epoch': 0.98}) +[2026-06-16 02:20:55,997][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.1715654879808426, 'train/grad_norm_pre_clip': 17.562583923339844, 'timing/data': 0.011570582166314125, 'timing/model': 1.6539038214832544, 'learning_rate/qwen_vl_interface': 5.335267220540963e-06, 'learning_rate/action_model': 5.335267220540963e-05, 'epoch': 0.98}) +[2026-06-16 02:21:01,696][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.07540884613990784, 'train/grad_norm_pre_clip': 12.446996688842773, 'timing/data': 0.010361267253756523, 'timing/model': 1.4518975857645273, 'learning_rate/qwen_vl_interface': 5.332222272489662e-06, 'learning_rate/action_model': 5.332222272489662e-05, 'epoch': 0.98}) +[2026-06-16 02:21:07,595][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.009481528773903847, 'train/grad_norm_pre_clip': 2.140727996826172, 'timing/data': 0.011223981156945229, 'timing/model': 1.4505762122571468, 'learning_rate/qwen_vl_interface': 5.3291772906398895e-06, 'learning_rate/action_model': 5.32917729063989e-05, 'epoch': 0.98}) +[2026-06-16 02:21:13,527][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.08145569264888763, 'train/grad_norm_pre_clip': 13.673362731933594, 'timing/data': 0.01134515181183815, 'timing/model': 1.4793343897908926, 'learning_rate/qwen_vl_interface': 5.326132276243327e-06, 'learning_rate/action_model': 5.3261322762433266e-05, 'epoch': 0.98}) +[2026-06-16 02:21:19,394][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.04923307150602341, 'train/grad_norm_pre_clip': 5.182424068450928, 'timing/data': 0.011154977604746819, 'timing/model': 1.4375707674771547, 'learning_rate/qwen_vl_interface': 5.323087230551659e-06, 'learning_rate/action_model': 5.323087230551659e-05, 'epoch': 0.98}) +[2026-06-16 02:21:25,730][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.12578940391540527, 'train/grad_norm_pre_clip': 13.996256828308105, 'timing/data': 0.01182355172932148, 'timing/model': 1.502966644242406, 'learning_rate/qwen_vl_interface': 5.3200421548165916e-06, 'learning_rate/action_model': 5.3200421548165914e-05, 'epoch': 0.98}) +[2026-06-16 02:21:31,763][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.04336674138903618, 'train/grad_norm_pre_clip': 7.557978630065918, 'timing/data': 0.010132437571883202, 'timing/model': 1.596365112811327, 'learning_rate/qwen_vl_interface': 5.3169970502898405e-06, 'learning_rate/action_model': 5.31699705028984e-05, 'epoch': 0.98}) +[2026-06-16 02:21:37,743][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.12783029675483704, 'train/grad_norm_pre_clip': 4.858959197998047, 'timing/data': 0.014490047469735146, 'timing/model': 1.6638972256332636, 'learning_rate/qwen_vl_interface': 5.31395191822313e-06, 'learning_rate/action_model': 5.31395191822313e-05, 'epoch': 0.98}) +[2026-06-16 02:21:43,920][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.034799724817276, 'train/grad_norm_pre_clip': 8.704668998718262, 'timing/data': 0.010258357971906662, 'timing/model': 1.536410054191947, 'learning_rate/qwen_vl_interface': 5.310906759868205e-06, 'learning_rate/action_model': 5.310906759868205e-05, 'epoch': 0.98}) +[2026-06-16 02:21:50,328][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.16199584305286407, 'train/grad_norm_pre_clip': 10.565515518188477, 'timing/data': 0.01239071972668171, 'timing/model': 1.53573034144938, 'learning_rate/qwen_vl_interface': 5.30786157647681e-06, 'learning_rate/action_model': 5.30786157647681e-05, 'epoch': 0.98}) +[2026-06-16 02:21:56,336][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.042790353298187256, 'train/grad_norm_pre_clip': 8.742594718933105, 'timing/data': 0.010054519400000572, 'timing/model': 1.4575831554830074, 'learning_rate/qwen_vl_interface': 5.304816369300709e-06, 'learning_rate/action_model': 5.304816369300709e-05, 'epoch': 0.98}) +[2026-06-16 02:22:02,527][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.13024824857711792, 'train/grad_norm_pre_clip': 10.223390579223633, 'timing/data': 0.009564543142914772, 'timing/model': 1.5280371028929949, 'learning_rate/qwen_vl_interface': 5.301771139591666e-06, 'learning_rate/action_model': 5.301771139591666e-05, 'epoch': 0.98}) +[2026-06-16 02:22:08,592][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.11363691091537476, 'train/grad_norm_pre_clip': 5.734871864318848, 'timing/data': 0.010378975421190262, 'timing/model': 1.7014275696128607, 'learning_rate/qwen_vl_interface': 5.298725888601465e-06, 'learning_rate/action_model': 5.298725888601465e-05, 'epoch': 0.98}) +[2026-06-16 02:22:14,529][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.02256188355386257, 'train/grad_norm_pre_clip': 2.7218210697174072, 'timing/data': 0.011448100209236145, 'timing/model': 1.5437771696597338, 'learning_rate/qwen_vl_interface': 5.295680617581891e-06, 'learning_rate/action_model': 5.2956806175818904e-05, 'epoch': 0.98}) +[2026-06-16 02:22:20,505][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.038799822330474854, 'train/grad_norm_pre_clip': 18.412691116333008, 'timing/data': 0.01067960262298584, 'timing/model': 1.5319710280746222, 'learning_rate/qwen_vl_interface': 5.2926353277847395e-06, 'learning_rate/action_model': 5.292635327784739e-05, 'epoch': 0.98}) +[2026-06-16 02:22:26,636][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.18828782439231873, 'train/grad_norm_pre_clip': 7.327692985534668, 'timing/data': 0.010959092527627945, 'timing/model': 1.5487479101866484, 'learning_rate/qwen_vl_interface': 5.2895900204618175e-06, 'learning_rate/action_model': 5.289590020461818e-05, 'epoch': 0.98}) +[2026-06-16 02:22:32,611][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.12474417686462402, 'train/grad_norm_pre_clip': 8.598443984985352, 'timing/data': 0.010827179998159409, 'timing/model': 1.4846560284495354, 'learning_rate/qwen_vl_interface': 5.286544696864933e-06, 'learning_rate/action_model': 5.2865446968649324e-05, 'epoch': 0.98}) +[2026-06-16 02:22:38,529][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.12826886773109436, 'train/grad_norm_pre_clip': 15.897222518920898, 'timing/data': 0.010553833097219467, 'timing/model': 1.6579992361366749, 'learning_rate/qwen_vl_interface': 5.283499358245904e-06, 'learning_rate/action_model': 5.2834993582459034e-05, 'epoch': 0.98}) +[2026-06-16 02:22:44,217][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.1477733701467514, 'train/grad_norm_pre_clip': 6.373338222503662, 'timing/data': 0.009678823873400688, 'timing/model': 1.451992891728878, 'learning_rate/qwen_vl_interface': 5.280454005856553e-06, 'learning_rate/action_model': 5.2804540058565534e-05, 'epoch': 0.98}) +[2026-06-16 02:22:50,277][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.0023534512147307396, 'train/grad_norm_pre_clip': 0.18630144000053406, 'timing/data': 0.01103072427213192, 'timing/model': 1.497815353795886, 'learning_rate/qwen_vl_interface': 5.277408640948711e-06, 'learning_rate/action_model': 5.277408640948711e-05, 'epoch': 0.98}) +[2026-06-16 02:22:56,112][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.1132953092455864, 'train/grad_norm_pre_clip': 17.56503677368164, 'timing/data': 0.011164018884301186, 'timing/model': 1.4539848323911428, 'learning_rate/qwen_vl_interface': 5.274363264774213e-06, 'learning_rate/action_model': 5.274363264774212e-05, 'epoch': 0.98}) +[2026-06-16 02:23:01,925][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.05409117788076401, 'train/grad_norm_pre_clip': 13.358357429504395, 'timing/data': 0.01057988777756691, 'timing/model': 1.450963955372572, 'learning_rate/qwen_vl_interface': 5.271317878584897e-06, 'learning_rate/action_model': 5.2713178785848974e-05, 'epoch': 0.98}) +[2026-06-16 02:23:07,787][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.22471952438354492, 'train/grad_norm_pre_clip': 8.832891464233398, 'timing/data': 0.010235704481601715, 'timing/model': 1.4993227440863848, 'learning_rate/qwen_vl_interface': 5.268272483632606e-06, 'learning_rate/action_model': 5.268272483632606e-05, 'epoch': 0.98}) +[2026-06-16 02:23:13,889][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.09133455157279968, 'train/grad_norm_pre_clip': 6.031317710876465, 'timing/data': 0.01011621206998825, 'timing/model': 1.6345057617872953, 'learning_rate/qwen_vl_interface': 5.265227081169187e-06, 'learning_rate/action_model': 5.265227081169186e-05, 'epoch': 0.98}) +[2026-06-16 02:23:19,588][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.05341147631406784, 'train/grad_norm_pre_clip': 9.28722858428955, 'timing/data': 0.011135924607515335, 'timing/model': 1.4605164676904678, 'learning_rate/qwen_vl_interface': 5.2621816724464895e-06, 'learning_rate/action_model': 5.262181672446489e-05, 'epoch': 0.99}) +[2026-06-16 02:23:25,482][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.017707744613289833, 'train/grad_norm_pre_clip': 1.1937980651855469, 'timing/data': 0.010120632126927376, 'timing/model': 1.4498471673578024, 'learning_rate/qwen_vl_interface': 5.259136258716367e-06, 'learning_rate/action_model': 5.2591362587163675e-05, 'epoch': 0.99}) +[2026-06-16 02:23:31,393][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.17994549870491028, 'train/grad_norm_pre_clip': 11.544049263000488, 'timing/data': 0.010409509763121605, 'timing/model': 1.4679620489478111, 'learning_rate/qwen_vl_interface': 5.256090841230675e-06, 'learning_rate/action_model': 5.256090841230674e-05, 'epoch': 0.99}) +[2026-06-16 02:23:37,279][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.09587985277175903, 'train/grad_norm_pre_clip': 11.650474548339844, 'timing/data': 0.010721907019615173, 'timing/model': 1.4522865749895573, 'learning_rate/qwen_vl_interface': 5.253045421241266e-06, 'learning_rate/action_model': 5.253045421241266e-05, 'epoch': 0.99}) +[2026-06-16 02:23:43,181][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.09708940982818604, 'train/grad_norm_pre_clip': 5.6176981925964355, 'timing/data': 0.010930558666586876, 'timing/model': 1.4893923606723547, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.99}) +[2026-06-16 02:23:49,085][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.05479605868458748, 'train/grad_norm_pre_clip': 6.197235584259033, 'timing/data': 0.010690664872527122, 'timing/model': 1.6545487269759178, 'learning_rate/qwen_vl_interface': 5.246954578758735e-06, 'learning_rate/action_model': 5.246954578758735e-05, 'epoch': 0.99}) +[2026-06-16 02:23:54,781][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.01843613013625145, 'train/grad_norm_pre_clip': 1.444374442100525, 'timing/data': 0.010454388335347176, 'timing/model': 1.4463768247514963, 'learning_rate/qwen_vl_interface': 5.243909158769327e-06, 'learning_rate/action_model': 5.2439091587693264e-05, 'epoch': 0.99}) +[2026-06-16 02:24:00,685][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.12988612055778503, 'train/grad_norm_pre_clip': 9.149599075317383, 'timing/data': 0.011207405477762222, 'timing/model': 1.4537306930869818, 'learning_rate/qwen_vl_interface': 5.240863741283634e-06, 'learning_rate/action_model': 5.2408637412836336e-05, 'epoch': 0.99}) +[2026-06-16 02:24:06,584][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.11226656287908554, 'train/grad_norm_pre_clip': 7.684329032897949, 'timing/data': 0.010740470141172409, 'timing/model': 1.450056254863739, 'learning_rate/qwen_vl_interface': 5.2378183275535125e-06, 'learning_rate/action_model': 5.237818327553512e-05, 'epoch': 0.99}) +[2026-06-16 02:24:12,479][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.029481705278158188, 'train/grad_norm_pre_clip': 3.3529481887817383, 'timing/data': 0.010605907067656517, 'timing/model': 1.4170627277344465, 'learning_rate/qwen_vl_interface': 5.234772918830815e-06, 'learning_rate/action_model': 5.234772918830815e-05, 'epoch': 0.99}) +[2026-06-16 02:24:18,379][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.16691626608371735, 'train/grad_norm_pre_clip': 7.174160480499268, 'timing/data': 0.01210096850991249, 'timing/model': 1.4725999310612679, 'learning_rate/qwen_vl_interface': 5.231727516367395e-06, 'learning_rate/action_model': 5.231727516367395e-05, 'epoch': 0.99}) +[2026-06-16 02:24:24,276][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.11490824818611145, 'train/grad_norm_pre_clip': 8.308960914611816, 'timing/data': 0.009755011647939682, 'timing/model': 1.6418252531439066, 'learning_rate/qwen_vl_interface': 5.228682121415104e-06, 'learning_rate/action_model': 5.228682121415104e-05, 'epoch': 0.99}) +[2026-06-16 02:24:29,981][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.024202605709433556, 'train/grad_norm_pre_clip': 2.588041067123413, 'timing/data': 0.011311640962958336, 'timing/model': 1.4420406054705381, 'learning_rate/qwen_vl_interface': 5.225636735225789e-06, 'learning_rate/action_model': 5.225636735225788e-05, 'epoch': 0.99}) +[2026-06-16 02:24:35,880][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.029506921768188477, 'train/grad_norm_pre_clip': 7.876106262207031, 'timing/data': 0.010629216209053993, 'timing/model': 1.4475746639072895, 'learning_rate/qwen_vl_interface': 5.22259135905129e-06, 'learning_rate/action_model': 5.2225913590512895e-05, 'epoch': 0.99}) +[2026-06-16 02:24:42,386][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.06750106066465378, 'train/grad_norm_pre_clip': 9.810687065124512, 'timing/data': 0.013542547821998596, 'timing/model': 1.4518653638660908, 'learning_rate/qwen_vl_interface': 5.219545994143448e-06, 'learning_rate/action_model': 5.219545994143448e-05, 'epoch': 0.99}) +[2026-06-16 02:24:48,182][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.07631316781044006, 'train/grad_norm_pre_clip': 11.716723442077637, 'timing/data': 0.009802399203181267, 'timing/model': 1.4507221095263958, 'learning_rate/qwen_vl_interface': 5.216500641754097e-06, 'learning_rate/action_model': 5.216500641754097e-05, 'epoch': 0.99}) +[2026-06-16 02:24:54,114][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.06613949686288834, 'train/grad_norm_pre_clip': 39.63457489013672, 'timing/data': 0.010744061321020126, 'timing/model': 1.4742560759186745, 'learning_rate/qwen_vl_interface': 5.213455303135067e-06, 'learning_rate/action_model': 5.2134553031350673e-05, 'epoch': 0.99}) +[2026-06-16 02:25:00,017][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.07218968123197556, 'train/grad_norm_pre_clip': 12.756122589111328, 'timing/data': 0.010659093037247658, 'timing/model': 1.6563998684287071, 'learning_rate/qwen_vl_interface': 5.2104099795381836e-06, 'learning_rate/action_model': 5.210409979538183e-05, 'epoch': 0.99}) +[2026-06-16 02:25:05,729][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.12857657670974731, 'train/grad_norm_pre_clip': 10.014362335205078, 'timing/data': 0.01123814657330513, 'timing/model': 1.465757628902793, 'learning_rate/qwen_vl_interface': 5.207364672215261e-06, 'learning_rate/action_model': 5.2073646722152604e-05, 'epoch': 0.99}) +[2026-06-16 02:25:11,715][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.030658384785056114, 'train/grad_norm_pre_clip': 3.2922379970550537, 'timing/data': 0.010675286874175072, 'timing/model': 1.4512065406888723, 'learning_rate/qwen_vl_interface': 5.2043193824181095e-06, 'learning_rate/action_model': 5.2043193824181094e-05, 'epoch': 0.99}) +[2026-06-16 02:25:17,614][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.03187613561749458, 'train/grad_norm_pre_clip': 5.425629615783691, 'timing/data': 0.011161979287862778, 'timing/model': 1.4513491988182068, 'learning_rate/qwen_vl_interface': 5.201274111398535e-06, 'learning_rate/action_model': 5.2012741113985356e-05, 'epoch': 0.99}) +[2026-06-16 02:25:23,621][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.025400230661034584, 'train/grad_norm_pre_clip': 2.6845998764038086, 'timing/data': 0.010186778381466866, 'timing/model': 1.4615971222519875, 'learning_rate/qwen_vl_interface': 5.198228860408334e-06, 'learning_rate/action_model': 5.198228860408334e-05, 'epoch': 0.99}) +[2026-06-16 02:25:29,613][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.04457030072808266, 'train/grad_norm_pre_clip': 6.278814792633057, 'timing/data': 0.010221470147371292, 'timing/model': 1.5068659633398056, 'learning_rate/qwen_vl_interface': 5.195183630699292e-06, 'learning_rate/action_model': 5.195183630699292e-05, 'epoch': 0.99}) +[2026-06-16 02:25:35,631][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.04502849653363228, 'train/grad_norm_pre_clip': 7.831565856933594, 'timing/data': 0.014559004455804825, 'timing/model': 1.7580985222011805, 'learning_rate/qwen_vl_interface': 5.192138423523189e-06, 'learning_rate/action_model': 5.1921384235231895e-05, 'epoch': 0.99}) +[2026-06-16 02:25:41,589][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.3587060272693634, 'train/grad_norm_pre_clip': 17.096187591552734, 'timing/data': 0.009660784155130386, 'timing/model': 1.5691721271723509, 'learning_rate/qwen_vl_interface': 5.189093240131795e-06, 'learning_rate/action_model': 5.189093240131795e-05, 'epoch': 0.99}) +[2026-06-16 02:25:47,579][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.24536949396133423, 'train/grad_norm_pre_clip': 14.64449405670166, 'timing/data': 0.010033216327428818, 'timing/model': 1.4421484805643559, 'learning_rate/qwen_vl_interface': 5.186048081776869e-06, 'learning_rate/action_model': 5.1860480817768685e-05, 'epoch': 1.0}) +[2026-06-16 02:25:53,478][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.12603889405727386, 'train/grad_norm_pre_clip': 18.011980056762695, 'timing/data': 0.010374926030635834, 'timing/model': 1.4517981093376875, 'learning_rate/qwen_vl_interface': 5.183002949710161e-06, 'learning_rate/action_model': 5.18300294971016e-05, 'epoch': 1.0}) +[2026-06-16 02:25:59,381][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.004324715584516525, 'train/grad_norm_pre_clip': 0.41904136538505554, 'timing/data': 0.009857814759016037, 'timing/model': 1.4303399976342916, 'learning_rate/qwen_vl_interface': 5.179957845183411e-06, 'learning_rate/action_model': 5.179957845183411e-05, 'epoch': 1.0}) +[2026-06-16 02:26:05,384][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.1652943342924118, 'train/grad_norm_pre_clip': 11.079792022705078, 'timing/data': 0.010114723816514015, 'timing/model': 1.4953547436743975, 'learning_rate/qwen_vl_interface': 5.176912769448343e-06, 'learning_rate/action_model': 5.176912769448343e-05, 'epoch': 1.0}) +[2026-06-16 02:26:11,302][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.015437038615345955, 'train/grad_norm_pre_clip': 3.7014617919921875, 'timing/data': 0.009987140074372292, 'timing/model': 1.647142143920064, 'learning_rate/qwen_vl_interface': 5.173867723756676e-06, 'learning_rate/action_model': 5.173867723756676e-05, 'epoch': 1.0}) +[2026-06-16 02:26:16,991][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.13744975626468658, 'train/grad_norm_pre_clip': 8.87923812866211, 'timing/data': 0.010216299444437027, 'timing/model': 1.4318216387182474, 'learning_rate/qwen_vl_interface': 5.170822709360112e-06, 'learning_rate/action_model': 5.170822709360111e-05, 'epoch': 1.0}) +[2026-06-16 02:26:22,887][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.08818399906158447, 'train/grad_norm_pre_clip': 12.320659637451172, 'timing/data': 0.011003747582435608, 'timing/model': 1.4494620487093925, 'learning_rate/qwen_vl_interface': 5.16777772751034e-06, 'learning_rate/action_model': 5.16777772751034e-05, 'epoch': 1.0}) +[2026-06-16 02:26:28,795][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.04299916326999664, 'train/grad_norm_pre_clip': 10.969375610351562, 'timing/data': 0.01064985990524292, 'timing/model': 1.4542104937136173, 'learning_rate/qwen_vl_interface': 5.16473277945904e-06, 'learning_rate/action_model': 5.1647327794590396e-05, 'epoch': 1.0}) +[2026-06-16 02:26:34,684][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.03295086696743965, 'train/grad_norm_pre_clip': 5.344318389892578, 'timing/data': 0.01125900074839592, 'timing/model': 1.438036011531949, 'learning_rate/qwen_vl_interface': 5.161687866457872e-06, 'learning_rate/action_model': 5.161687866457871e-05, 'epoch': 1.0}) +[2026-06-16 02:26:40,587][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.026774659752845764, 'train/grad_norm_pre_clip': 5.78224515914917, 'timing/data': 0.010551266372203827, 'timing/model': 1.4730189461261034, 'learning_rate/qwen_vl_interface': 5.158642989758487e-06, 'learning_rate/action_model': 5.158642989758487e-05, 'epoch': 1.0}) +[2026-06-16 02:26:46,692][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.31095775961875916, 'train/grad_norm_pre_clip': 16.018766403198242, 'timing/data': 0.010422008112072945, 'timing/model': 1.8546007312834263, 'learning_rate/qwen_vl_interface': 5.155598150612519e-06, 'learning_rate/action_model': 5.1555981506125185e-05, 'epoch': 1.0}) +[2026-06-16 02:26:52,705][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.012765135616064072, 'train/grad_norm_pre_clip': 1.3190803527832031, 'timing/data': 0.010779254138469696, 'timing/model': 1.6398778352886438, 'learning_rate/qwen_vl_interface': 5.152553350271584e-06, 'learning_rate/action_model': 5.152553350271584e-05, 'epoch': 1.0}) +[2026-06-16 02:26:58,402][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.06526127457618713, 'train/grad_norm_pre_clip': 4.404139041900635, 'timing/data': 0.010252928361296654, 'timing/model': 1.4593424517661333, 'learning_rate/qwen_vl_interface': 5.149508589987286e-06, 'learning_rate/action_model': 5.1495085899872855e-05, 'epoch': 1.0}) +[2026-06-16 02:27:03,584][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.05187926068902016, 'train/grad_norm_pre_clip': 5.910617828369141, 'timing/data': 0.00024638883769512177, 'timing/model': 0.7370494809001684, 'learning_rate/qwen_vl_interface': 5.14646387101121e-06, 'learning_rate/action_model': 5.14646387101121e-05, 'epoch': 1.0}) +[2026-06-16 02:27:11,476][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.0141116539016366, 'train/grad_norm_pre_clip': 6.286667346954346, 'timing/data': 0.043360088020563126, 'timing/model': 1.6688790321350098, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 1.0}) +[2026-06-16 02:27:17,878][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.10361742973327637, 'train/grad_norm_pre_clip': 7.8058013916015625, 'timing/data': 0.0405802633613348, 'timing/model': 1.5589013565331697, 'learning_rate/qwen_vl_interface': 5.140374561989988e-06, 'learning_rate/action_model': 5.140374561989988e-05, 'epoch': 1.0}) +[2026-06-16 02:27:24,216][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.03125752508640289, 'train/grad_norm_pre_clip': 5.880345821380615, 'timing/data': 0.04106205701828003, 'timing/model': 1.7611168436706066, 'learning_rate/qwen_vl_interface': 5.137329974447924e-06, 'learning_rate/action_model': 5.137329974447924e-05, 'epoch': 1.0}) +[2026-06-16 02:27:30,281][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.022155337035655975, 'train/grad_norm_pre_clip': 10.627053260803223, 'timing/data': 0.04195313900709152, 'timing/model': 1.5234099924564362, 'learning_rate/qwen_vl_interface': 5.1342854332202516e-06, 'learning_rate/action_model': 5.1342854332202514e-05, 'epoch': 1.0}) +[2026-06-16 02:27:36,576][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.06363215297460556, 'train/grad_norm_pre_clip': 5.649112701416016, 'timing/data': 0.04207848571240902, 'timing/model': 1.522580873221159, 'learning_rate/qwen_vl_interface': 5.131240939558467e-06, 'learning_rate/action_model': 5.131240939558467e-05, 'epoch': 1.0}) +[2026-06-16 02:27:42,875][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.02883232943713665, 'train/grad_norm_pre_clip': 8.766976356506348, 'timing/data': 0.040706995874643326, 'timing/model': 1.5221751257777214, 'learning_rate/qwen_vl_interface': 5.128196494714048e-06, 'learning_rate/action_model': 5.128196494714047e-05, 'epoch': 1.0}) +[2026-06-16 02:27:49,158][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.03849126771092415, 'train/grad_norm_pre_clip': 16.001434326171875, 'timing/data': 0.04009022377431393, 'timing/model': 1.4944459609687328, 'learning_rate/qwen_vl_interface': 5.125152099938448e-06, 'learning_rate/action_model': 5.125152099938447e-05, 'epoch': 1.0}) +[2026-06-16 02:27:55,449][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.02867591753602028, 'train/grad_norm_pre_clip': 7.470026969909668, 'timing/data': 0.04190541245043278, 'timing/model': 1.5423957742750645, 'learning_rate/qwen_vl_interface': 5.122107756483104e-06, 'learning_rate/action_model': 5.122107756483103e-05, 'epoch': 1.0}) +[2026-06-16 02:28:01,573][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.21377603709697723, 'train/grad_norm_pre_clip': 9.937789916992188, 'timing/data': 0.04231667332351208, 'timing/model': 1.7151319924741983, 'learning_rate/qwen_vl_interface': 5.119063465599429e-06, 'learning_rate/action_model': 5.1190634655994285e-05, 'epoch': 1.0}) +[2026-06-16 02:28:07,554][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.003914397209882736, 'train/grad_norm_pre_clip': 0.6317076683044434, 'timing/data': 0.04075324535369873, 'timing/model': 1.5349624585360289, 'learning_rate/qwen_vl_interface': 5.116019228538818e-06, 'learning_rate/action_model': 5.116019228538818e-05, 'epoch': 1.0}) +[2026-06-16 02:28:13,749][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.004320832900702953, 'train/grad_norm_pre_clip': 0.6986322999000549, 'timing/data': 0.041256485506892204, 'timing/model': 1.5463125240057707, 'learning_rate/qwen_vl_interface': 5.112975046552641e-06, 'learning_rate/action_model': 5.1129750465526404e-05, 'epoch': 1.0}) +[2026-06-16 02:28:19,847][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.10397972166538239, 'train/grad_norm_pre_clip': 11.155569076538086, 'timing/data': 0.04119277559220791, 'timing/model': 1.5278647728264332, 'learning_rate/qwen_vl_interface': 5.109930920892246e-06, 'learning_rate/action_model': 5.1099309208922464e-05, 'epoch': 1.0}) +[2026-06-16 02:28:26,054][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.026975184679031372, 'train/grad_norm_pre_clip': 26.433664321899414, 'timing/data': 0.039800362661480904, 'timing/model': 1.5357047989964485, 'learning_rate/qwen_vl_interface': 5.10688685280896e-06, 'learning_rate/action_model': 5.10688685280896e-05, 'epoch': 1.01}) +[2026-06-16 02:28:32,152][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.0018194590229541063, 'train/grad_norm_pre_clip': 0.32352161407470703, 'timing/data': 0.04158352129161358, 'timing/model': 1.5577459875494242, 'learning_rate/qwen_vl_interface': 5.103842843554082e-06, 'learning_rate/action_model': 5.103842843554082e-05, 'epoch': 1.01}) +[2026-06-16 02:28:38,311][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.06230223923921585, 'train/grad_norm_pre_clip': 7.353724479675293, 'timing/data': 0.03966492973268032, 'timing/model': 1.7611430920660496, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 1.01}) +[2026-06-16 02:35:59,127][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.04295088350772858, 'train/grad_norm_pre_clip': 7.567426681518555, 'eval/loss': 0.06219562515616417, 'eval/flap_precision': 0.9467213114754098, 'eval/flap_recall': 0.9402985074626866, 'eval/flap_f1': 0.943498978897209, 'eval/flap_rate_model': 0.20333333333333334, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.943498978897209, 'timing/data': 0.04226808436214924, 'timing/model': 1.5250205975025892, 'learning_rate/qwen_vl_interface': 5.097755006534638e-06, 'learning_rate/action_model': 5.097755006534638e-05, 'epoch': 1.01}) +[2026-06-16 02:35:59,128][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2600_state +[2026-06-16 02:35:59,134][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:36:06,974][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2600_state/model.safetensors +[2026-06-16 02:36:20,872][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2600_state/optimizer.bin +[2026-06-16 02:36:20,873][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2600_state/sampler.bin +[2026-06-16 02:36:20,873][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2600_state/sampler_1.bin +[2026-06-16 02:36:20,875][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2600_state/random_states_0.pkl +[2026-06-16 02:36:20,997][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2600 +[2026-06-16 02:36:20,998][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:36:21,013][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:36:27,336][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.18429134786128998, 'train/grad_norm_pre_clip': 17.631053924560547, 'timing/data': 0.05698288790881634, 'timing/model': 1.72165803052485, 'learning_rate/qwen_vl_interface': 5.094711181272551e-06, 'learning_rate/action_model': 5.094711181272551e-05, 'epoch': 1.01}) +[2026-06-16 02:36:33,133][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.047474734485149384, 'train/grad_norm_pre_clip': 7.2569661140441895, 'timing/data': 0.041171854361891747, 'timing/model': 1.536660809069872, 'learning_rate/qwen_vl_interface': 5.091667419843833e-06, 'learning_rate/action_model': 5.091667419843833e-05, 'epoch': 1.01}) +[2026-06-16 02:36:39,229][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.015956418588757515, 'train/grad_norm_pre_clip': 4.715312480926514, 'timing/data': 0.039744770154356956, 'timing/model': 1.5264936424791813, 'learning_rate/qwen_vl_interface': 5.088623723499657e-06, 'learning_rate/action_model': 5.088623723499657e-05, 'epoch': 1.01}) +[2026-06-16 02:36:45,334][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.12884262204170227, 'train/grad_norm_pre_clip': 10.562652587890625, 'timing/data': 0.04058384522795677, 'timing/model': 1.5356167908757925, 'learning_rate/qwen_vl_interface': 5.085580093491171e-06, 'learning_rate/action_model': 5.085580093491171e-05, 'epoch': 1.01}) +[2026-06-16 02:36:51,439][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.03499500826001167, 'train/grad_norm_pre_clip': 10.262173652648926, 'timing/data': 0.03981565497815609, 'timing/model': 1.4857558012008667, 'learning_rate/qwen_vl_interface': 5.0825365310695e-06, 'learning_rate/action_model': 5.082536531069499e-05, 'epoch': 1.01}) +[2026-06-16 02:36:57,289][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.2330349236726761, 'train/grad_norm_pre_clip': 24.34765625, 'timing/data': 0.033158738166093826, 'timing/model': 1.6303015407174826, 'learning_rate/qwen_vl_interface': 5.079493037485735e-06, 'learning_rate/action_model': 5.0794930374857344e-05, 'epoch': 1.01}) +[2026-06-16 02:37:02,981][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.1602310985326767, 'train/grad_norm_pre_clip': 5.695999622344971, 'timing/data': 0.03437438793480396, 'timing/model': 1.431538024917245, 'learning_rate/qwen_vl_interface': 5.076449613990941e-06, 'learning_rate/action_model': 5.076449613990941e-05, 'epoch': 1.01}) +[2026-06-16 02:37:08,885][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.060896821320056915, 'train/grad_norm_pre_clip': 31.064735412597656, 'timing/data': 0.03375277854502201, 'timing/model': 1.4309478867799044, 'learning_rate/qwen_vl_interface': 5.073406261836157e-06, 'learning_rate/action_model': 5.0734062618361565e-05, 'epoch': 1.01}) +[2026-06-16 02:37:14,783][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.15533624589443207, 'train/grad_norm_pre_clip': 24.157501220703125, 'timing/data': 0.03310792334377766, 'timing/model': 1.429160201922059, 'learning_rate/qwen_vl_interface': 5.070362982272386e-06, 'learning_rate/action_model': 5.070362982272385e-05, 'epoch': 1.01}) +[2026-06-16 02:37:20,682][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.15172043442726135, 'train/grad_norm_pre_clip': 6.568496227264404, 'timing/data': 0.033121995627880096, 'timing/model': 1.4305826872587204, 'learning_rate/qwen_vl_interface': 5.067319776550607e-06, 'learning_rate/action_model': 5.067319776550607e-05, 'epoch': 1.01}) +[2026-06-16 02:37:26,590][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.08313315361738205, 'train/grad_norm_pre_clip': 13.175565719604492, 'timing/data': 0.033641744405031204, 'timing/model': 1.4222675804048777, 'learning_rate/qwen_vl_interface': 5.064276645921768e-06, 'learning_rate/action_model': 5.064276645921768e-05, 'epoch': 1.01}) +[2026-06-16 02:37:32,493][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.0335378423333168, 'train/grad_norm_pre_clip': 4.710267543792725, 'timing/data': 0.0329043623059988, 'timing/model': 1.6397032123059034, 'learning_rate/qwen_vl_interface': 5.061233591636787e-06, 'learning_rate/action_model': 5.061233591636787e-05, 'epoch': 1.01}) +[2026-06-16 02:37:38,579][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.007679967675358057, 'train/grad_norm_pre_clip': 3.117368698120117, 'timing/data': 0.04087166301906109, 'timing/model': 1.652529140934348, 'learning_rate/qwen_vl_interface': 5.058190614946544e-06, 'learning_rate/action_model': 5.058190614946544e-05, 'epoch': 1.01}) +[2026-06-16 02:37:44,645][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.11272906512022018, 'train/grad_norm_pre_clip': 9.459307670593262, 'timing/data': 0.042792389169335365, 'timing/model': 1.526167443022132, 'learning_rate/qwen_vl_interface': 5.055147717101895e-06, 'learning_rate/action_model': 5.055147717101894e-05, 'epoch': 1.01}) +[2026-06-16 02:37:50,945][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.10278518497943878, 'train/grad_norm_pre_clip': 10.483854293823242, 'timing/data': 0.04234310984611511, 'timing/model': 1.5176447238773108, 'learning_rate/qwen_vl_interface': 5.052104899353658e-06, 'learning_rate/action_model': 5.0521048993536577e-05, 'epoch': 1.01}) +[2026-06-16 02:37:57,147][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.23733538389205933, 'train/grad_norm_pre_clip': 31.15034294128418, 'timing/data': 0.04147132858633995, 'timing/model': 1.4860724080353975, 'learning_rate/qwen_vl_interface': 5.049062162952622e-06, 'learning_rate/action_model': 5.0490621629526214e-05, 'epoch': 1.01}) +[2026-06-16 02:38:03,343][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.05243061110377312, 'train/grad_norm_pre_clip': 10.59315013885498, 'timing/data': 0.04142150841653347, 'timing/model': 1.5144253838807344, 'learning_rate/qwen_vl_interface': 5.046019509149541e-06, 'learning_rate/action_model': 5.046019509149541e-05, 'epoch': 1.01}) +[2026-06-16 02:38:09,529][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.06538322567939758, 'train/grad_norm_pre_clip': 5.906461715698242, 'timing/data': 0.040312690660357475, 'timing/model': 1.7751835118979216, 'learning_rate/qwen_vl_interface': 5.042976939195135e-06, 'learning_rate/action_model': 5.0429769391951355e-05, 'epoch': 1.01}) +[2026-06-16 02:38:15,746][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.062489256262779236, 'train/grad_norm_pre_clip': 11.850234031677246, 'timing/data': 0.04119168780744076, 'timing/model': 1.721808172762394, 'learning_rate/qwen_vl_interface': 5.039934454340091e-06, 'learning_rate/action_model': 5.0399344543400907e-05, 'epoch': 1.01}) +[2026-06-16 02:38:21,748][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.21769878268241882, 'train/grad_norm_pre_clip': 7.408224105834961, 'timing/data': 0.04092315398156643, 'timing/model': 1.533685328438878, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 1.01}) +[2026-06-16 02:38:27,890][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.10947369039058685, 'train/grad_norm_pre_clip': 6.4344563484191895, 'timing/data': 0.04142061062157154, 'timing/model': 1.5240942239761353, 'learning_rate/qwen_vl_interface': 5.033849744930647e-06, 'learning_rate/action_model': 5.033849744930646e-05, 'epoch': 1.01}) +[2026-06-16 02:38:34,150][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.1360999047756195, 'train/grad_norm_pre_clip': 8.831068992614746, 'timing/data': 0.040829725563526154, 'timing/model': 1.5007808171212673, 'learning_rate/qwen_vl_interface': 5.030807522877443e-06, 'learning_rate/action_model': 5.030807522877443e-05, 'epoch': 1.01}) +[2026-06-16 02:38:40,458][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.09434226155281067, 'train/grad_norm_pre_clip': 4.527894496917725, 'timing/data': 0.04118581302464008, 'timing/model': 1.5089113898575306, 'learning_rate/qwen_vl_interface': 5.027765390925987e-06, 'learning_rate/action_model': 5.027765390925987e-05, 'epoch': 1.02}) +[2026-06-16 02:38:46,697][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.15892305970191956, 'train/grad_norm_pre_clip': 10.752608299255371, 'timing/data': 0.04005436599254608, 'timing/model': 1.7588558197021484, 'learning_rate/qwen_vl_interface': 5.024723350326783e-06, 'learning_rate/action_model': 5.024723350326783e-05, 'epoch': 1.02}) +[2026-06-16 02:38:53,008][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.004587728530168533, 'train/grad_norm_pre_clip': 0.42351165413856506, 'timing/data': 0.04118550196290016, 'timing/model': 1.783155543729663, 'learning_rate/qwen_vl_interface': 5.021681402330303e-06, 'learning_rate/action_model': 5.021681402330303e-05, 'epoch': 1.02}) +[2026-06-16 02:38:59,071][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.15727488696575165, 'train/grad_norm_pre_clip': 10.810026168823242, 'timing/data': 0.04123768024146557, 'timing/model': 1.5366226248443127, 'learning_rate/qwen_vl_interface': 5.018639548186972e-06, 'learning_rate/action_model': 5.0186395481869716e-05, 'epoch': 1.02}) +[2026-06-16 02:39:05,257][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.009643019177019596, 'train/grad_norm_pre_clip': 0.634360134601593, 'timing/data': 0.04138060100376606, 'timing/model': 1.5019894167780876, 'learning_rate/qwen_vl_interface': 5.015597789147184e-06, 'learning_rate/action_model': 5.015597789147184e-05, 'epoch': 1.02}) +[2026-06-16 02:39:11,672][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.09356763958930969, 'train/grad_norm_pre_clip': 7.3922882080078125, 'timing/data': 0.0418210681527853, 'timing/model': 1.6366804335266352, 'learning_rate/qwen_vl_interface': 5.012556126461291e-06, 'learning_rate/action_model': 5.012556126461291e-05, 'epoch': 1.02}) +[2026-06-16 02:39:18,340][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.03056320734322071, 'train/grad_norm_pre_clip': 3.6516923904418945, 'timing/data': 0.05774254538118839, 'timing/model': 1.7425938788801432, 'learning_rate/qwen_vl_interface': 5.0095145613796044e-06, 'learning_rate/action_model': 5.0095145613796046e-05, 'epoch': 1.02}) +[2026-06-16 02:39:24,949][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.04168999195098877, 'train/grad_norm_pre_clip': 7.225548267364502, 'timing/data': 0.041188446804881096, 'timing/model': 1.5427829753607512, 'learning_rate/qwen_vl_interface': 5.0064730951524e-06, 'learning_rate/action_model': 5.0064730951524e-05, 'epoch': 1.02}) +[2026-06-16 02:39:31,218][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.05472399666905403, 'train/grad_norm_pre_clip': 9.149144172668457, 'timing/data': 0.0421154759824276, 'timing/model': 1.7606368325650692, 'learning_rate/qwen_vl_interface': 5.003431729029905e-06, 'learning_rate/action_model': 5.003431729029905e-05, 'epoch': 1.02}) +[2026-06-16 02:39:37,280][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.08761068433523178, 'train/grad_norm_pre_clip': 13.551356315612793, 'timing/data': 0.04204885847866535, 'timing/model': 1.5247886683791876, 'learning_rate/qwen_vl_interface': 5.000390464262314e-06, 'learning_rate/action_model': 5.0003904642623136e-05, 'epoch': 1.02}) +[2026-06-16 02:39:43,581][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.017178045585751534, 'train/grad_norm_pre_clip': 1.8916305303573608, 'timing/data': 0.041058095172047615, 'timing/model': 1.5233814921230078, 'learning_rate/qwen_vl_interface': 4.997349302099774e-06, 'learning_rate/action_model': 4.997349302099773e-05, 'epoch': 1.02}) +[2026-06-16 02:39:49,901][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.045331019908189774, 'train/grad_norm_pre_clip': 3.9117910861968994, 'timing/data': 0.04153989814221859, 'timing/model': 1.6406055558472872, 'learning_rate/qwen_vl_interface': 4.994308243792391e-06, 'learning_rate/action_model': 4.994308243792391e-05, 'epoch': 1.02}) +[2026-06-16 02:39:56,104][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.04215038940310478, 'train/grad_norm_pre_clip': 5.072832107543945, 'timing/data': 0.04350366070866585, 'timing/model': 1.5441426243633032, 'learning_rate/qwen_vl_interface': 4.991267290590233e-06, 'learning_rate/action_model': 4.991267290590233e-05, 'epoch': 1.02}) +[2026-06-16 02:40:02,385][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.1607484370470047, 'train/grad_norm_pre_clip': 10.893362045288086, 'timing/data': 0.042503390461206436, 'timing/model': 1.5605474393814802, 'learning_rate/qwen_vl_interface': 4.988226443743317e-06, 'learning_rate/action_model': 4.9882264437433174e-05, 'epoch': 1.02}) +[2026-06-16 02:40:08,773][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.10871690511703491, 'train/grad_norm_pre_clip': 12.727972030639648, 'timing/data': 0.046214018017053604, 'timing/model': 1.8066956009715796, 'learning_rate/qwen_vl_interface': 4.985185704501623e-06, 'learning_rate/action_model': 4.985185704501623e-05, 'epoch': 1.02}) +[2026-06-16 02:40:14,861][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.07761263102293015, 'train/grad_norm_pre_clip': 12.469282150268555, 'timing/data': 0.04385833069682121, 'timing/model': 1.4972999226301908, 'learning_rate/qwen_vl_interface': 4.982145074115082e-06, 'learning_rate/action_model': 4.982145074115082e-05, 'epoch': 1.02}) +[2026-06-16 02:40:21,046][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.1283458024263382, 'train/grad_norm_pre_clip': 16.68389892578125, 'timing/data': 0.04212195612490177, 'timing/model': 1.5327301863580942, 'learning_rate/qwen_vl_interface': 4.979104553833584e-06, 'learning_rate/action_model': 4.979104553833584e-05, 'epoch': 1.02}) +[2026-06-16 02:40:27,293][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.059321921318769455, 'train/grad_norm_pre_clip': 4.822775840759277, 'timing/data': 0.04284844920039177, 'timing/model': 1.5284733325242996, 'learning_rate/qwen_vl_interface': 4.9760641449069724e-06, 'learning_rate/action_model': 4.976064144906972e-05, 'epoch': 1.02}) +[2026-06-16 02:40:33,590][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.01170318853110075, 'train/grad_norm_pre_clip': 4.478997230529785, 'timing/data': 0.040628932416439056, 'timing/model': 1.522760633379221, 'learning_rate/qwen_vl_interface': 4.973023848585042e-06, 'learning_rate/action_model': 4.973023848585042e-05, 'epoch': 1.02}) +[2026-06-16 02:40:40,007][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.039591312408447266, 'train/grad_norm_pre_clip': 6.177189826965332, 'timing/data': 0.04313924349844456, 'timing/model': 1.7794362306594849, 'learning_rate/qwen_vl_interface': 4.969983666117545e-06, 'learning_rate/action_model': 4.9699836661175446e-05, 'epoch': 1.02}) +[2026-06-16 02:40:46,059][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.021583573892712593, 'train/grad_norm_pre_clip': 4.232626914978027, 'timing/data': 0.042752498760819435, 'timing/model': 1.532477056607604, 'learning_rate/qwen_vl_interface': 4.966943598754185e-06, 'learning_rate/action_model': 4.9669435987541854e-05, 'epoch': 1.02}) +[2026-06-16 02:40:52,250][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.005502898246049881, 'train/grad_norm_pre_clip': 1.0823187828063965, 'timing/data': 0.04196457378566265, 'timing/model': 1.4906009938567877, 'learning_rate/qwen_vl_interface': 4.963903647744621e-06, 'learning_rate/action_model': 4.96390364774462e-05, 'epoch': 1.02}) +[2026-06-16 02:40:58,853][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.19517728686332703, 'train/grad_norm_pre_clip': 7.367279529571533, 'timing/data': 0.055982403457164764, 'timing/model': 1.6289511639624834, 'learning_rate/qwen_vl_interface': 4.9608638143384604e-06, 'learning_rate/action_model': 4.960863814338461e-05, 'epoch': 1.02}) +[2026-06-16 02:41:05,679][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.1406848281621933, 'train/grad_norm_pre_clip': 19.900978088378906, 'timing/data': 0.04155925288796425, 'timing/model': 1.6651718225330114, 'learning_rate/qwen_vl_interface': 4.9578240997852636e-06, 'learning_rate/action_model': 4.957824099785264e-05, 'epoch': 1.02}) +[2026-06-16 02:41:12,062][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.20185519754886627, 'train/grad_norm_pre_clip': 15.379271507263184, 'timing/data': 0.033473193645477295, 'timing/model': 1.5462523344904184, 'learning_rate/qwen_vl_interface': 4.954784505334544e-06, 'learning_rate/action_model': 4.9547845053345434e-05, 'epoch': 1.02}) +[2026-06-16 02:41:18,294][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.10605449974536896, 'train/grad_norm_pre_clip': 6.785065174102783, 'timing/data': 0.03259738348424435, 'timing/model': 1.7755255177617073, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 1.02}) +[2026-06-16 02:41:24,547][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.14637932181358337, 'train/grad_norm_pre_clip': 11.38059139251709, 'timing/data': 0.0328510757535696, 'timing/model': 1.7296018544584513, 'learning_rate/qwen_vl_interface': 4.94870568173833e-06, 'learning_rate/action_model': 4.9487056817383307e-05, 'epoch': 1.03}) +[2026-06-16 02:41:30,561][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.036074165254831314, 'train/grad_norm_pre_clip': 6.727998733520508, 'timing/data': 0.03216783329844475, 'timing/model': 1.5273862332105637, 'learning_rate/qwen_vl_interface': 4.945666455091614e-06, 'learning_rate/action_model': 4.9456664550916136e-05, 'epoch': 1.03}) +[2026-06-16 02:41:37,056][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.05476367473602295, 'train/grad_norm_pre_clip': 8.827466011047363, 'timing/data': 0.03207097202539444, 'timing/model': 1.5271114837378263, 'learning_rate/qwen_vl_interface': 4.942627353544921e-06, 'learning_rate/action_model': 4.942627353544921e-05, 'epoch': 1.03}) +[2026-06-16 02:41:43,257][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.11228091269731522, 'train/grad_norm_pre_clip': 8.60122299194336, 'timing/data': 0.031599072739481926, 'timing/model': 1.4879203662276268, 'learning_rate/qwen_vl_interface': 4.939588378347513e-06, 'learning_rate/action_model': 4.9395883783475125e-05, 'epoch': 1.03}) +[2026-06-16 02:41:49,550][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.11080992966890335, 'train/grad_norm_pre_clip': 11.876238822937012, 'timing/data': 0.03237926587462425, 'timing/model': 1.5536967758089304, 'learning_rate/qwen_vl_interface': 4.936549530748597e-06, 'learning_rate/action_model': 4.9365495307485963e-05, 'epoch': 1.03}) +[2026-06-16 02:41:55,791][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.0049730632454156876, 'train/grad_norm_pre_clip': 1.0088080167770386, 'timing/data': 0.03170582838356495, 'timing/model': 1.742375260218978, 'learning_rate/qwen_vl_interface': 4.9335108119973286e-06, 'learning_rate/action_model': 4.933510811997328e-05, 'epoch': 1.03}) +[2026-06-16 02:42:02,021][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.5144990682601929, 'train/grad_norm_pre_clip': 27.428300857543945, 'timing/data': 0.033663200214505196, 'timing/model': 1.7010733149945736, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 1.03}) +[2026-06-16 02:42:08,446][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.04900902882218361, 'train/grad_norm_pre_clip': 4.0448503494262695, 'timing/data': 0.030345452949404716, 'timing/model': 1.6334679797291756, 'learning_rate/qwen_vl_interface': 4.927433766034093e-06, 'learning_rate/action_model': 4.9274337660340934e-05, 'epoch': 1.03}) +[2026-06-16 02:42:14,656][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.1439715027809143, 'train/grad_norm_pre_clip': 10.699039459228516, 'timing/data': 0.03238176181912422, 'timing/model': 1.5426553804427385, 'learning_rate/qwen_vl_interface': 4.92439544132017e-06, 'learning_rate/action_model': 4.92439544132017e-05, 'epoch': 1.03}) +[2026-06-16 02:42:20,898][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.21050141751766205, 'train/grad_norm_pre_clip': 10.54964542388916, 'timing/data': 0.034124914556741714, 'timing/model': 1.546329140663147, 'learning_rate/qwen_vl_interface': 4.92135725044998e-06, 'learning_rate/action_model': 4.92135725044998e-05, 'epoch': 1.03}) +[2026-06-16 02:42:27,947][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.021270183846354485, 'train/grad_norm_pre_clip': 4.474385738372803, 'timing/data': 0.08074376545846462, 'timing/model': 1.7218271624296904, 'learning_rate/qwen_vl_interface': 4.918319194672412e-06, 'learning_rate/action_model': 4.918319194672412e-05, 'epoch': 1.03}) +[2026-06-16 02:42:34,881][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.1979065090417862, 'train/grad_norm_pre_clip': 11.642731666564941, 'timing/data': 0.042578199878335, 'timing/model': 1.9990357290953398, 'learning_rate/qwen_vl_interface': 4.9152812752362935e-06, 'learning_rate/action_model': 4.915281275236293e-05, 'epoch': 1.03}) +[2026-06-16 02:42:41,819][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.0657135546207428, 'train/grad_norm_pre_clip': 3.119763135910034, 'timing/data': 0.0409254040569067, 'timing/model': 2.010834489017725, 'learning_rate/qwen_vl_interface': 4.9122434933904e-06, 'learning_rate/action_model': 4.9122434933904e-05, 'epoch': 1.03}) +[2026-06-16 02:42:48,403][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.14911523461341858, 'train/grad_norm_pre_clip': 14.040431022644043, 'timing/data': 0.05206935293972492, 'timing/model': 1.6725661233067513, 'learning_rate/qwen_vl_interface': 4.909205850383449e-06, 'learning_rate/action_model': 4.9092058503834485e-05, 'epoch': 1.03}) +[2026-06-16 02:42:55,062][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.13508225977420807, 'train/grad_norm_pre_clip': 13.596293449401855, 'timing/data': 0.03488658554852009, 'timing/model': 1.552973872050643, 'learning_rate/qwen_vl_interface': 4.906168347464098e-06, 'learning_rate/action_model': 4.906168347464098e-05, 'epoch': 1.03}) +[2026-06-16 02:43:01,252][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.18210230767726898, 'train/grad_norm_pre_clip': 7.039907455444336, 'timing/data': 0.03385167568922043, 'timing/model': 1.5350312292575836, 'learning_rate/qwen_vl_interface': 4.903130985880955e-06, 'learning_rate/action_model': 4.903130985880954e-05, 'epoch': 1.03}) +[2026-06-16 02:43:07,197][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.2533574402332306, 'train/grad_norm_pre_clip': 19.08183479309082, 'timing/data': 0.02684212476015091, 'timing/model': 1.4313599709421396, 'learning_rate/qwen_vl_interface': 4.900093766882558e-06, 'learning_rate/action_model': 4.900093766882558e-05, 'epoch': 1.03}) +[2026-06-16 02:43:13,102][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.04808751493692398, 'train/grad_norm_pre_clip': 4.528818607330322, 'timing/data': 0.026616288349032402, 'timing/model': 1.6469897609204054, 'learning_rate/qwen_vl_interface': 4.897056691717397e-06, 'learning_rate/action_model': 4.8970566917173966e-05, 'epoch': 1.03}) +[2026-06-16 02:43:18,791][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.06036766618490219, 'train/grad_norm_pre_clip': 6.031740188598633, 'timing/data': 0.026113228872418404, 'timing/model': 1.4376393388956785, 'learning_rate/qwen_vl_interface': 4.894019761633898e-06, 'learning_rate/action_model': 4.894019761633898e-05, 'epoch': 1.03}) +[2026-06-16 02:43:24,691][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.0452798455953598, 'train/grad_norm_pre_clip': 3.25307297706604, 'timing/data': 0.025523778051137924, 'timing/model': 1.4373516887426376, 'learning_rate/qwen_vl_interface': 4.8909829778804304e-06, 'learning_rate/action_model': 4.89098297788043e-05, 'epoch': 1.03}) +[2026-06-16 02:43:30,591][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.0878685861825943, 'train/grad_norm_pre_clip': 6.655977249145508, 'timing/data': 0.02612253464758396, 'timing/model': 1.4369289688766003, 'learning_rate/qwen_vl_interface': 4.887946341705298e-06, 'learning_rate/action_model': 4.887946341705297e-05, 'epoch': 1.03}) +[2026-06-16 02:43:36,493][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.028263187035918236, 'train/grad_norm_pre_clip': 8.40356731414795, 'timing/data': 0.02618316002190113, 'timing/model': 1.4277638662606478, 'learning_rate/qwen_vl_interface': 4.8849098543567465e-06, 'learning_rate/action_model': 4.884909854356747e-05, 'epoch': 1.03}) +[2026-06-16 02:43:42,315][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.2199358195066452, 'train/grad_norm_pre_clip': 14.440475463867188, 'timing/data': 0.0275042112916708, 'timing/model': 1.4389133267104626, 'learning_rate/qwen_vl_interface': 4.881873517082965e-06, 'learning_rate/action_model': 4.8818735170829655e-05, 'epoch': 1.03}) +[2026-06-16 02:43:48,148][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.1333211362361908, 'train/grad_norm_pre_clip': 7.860350131988525, 'timing/data': 0.027071155607700348, 'timing/model': 1.5939934439957142, 'learning_rate/qwen_vl_interface': 4.878837331132076e-06, 'learning_rate/action_model': 4.8788373311320754e-05, 'epoch': 1.03}) +[2026-06-16 02:43:53,792][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.18359357118606567, 'train/grad_norm_pre_clip': 8.801304817199707, 'timing/data': 0.027030320838093758, 'timing/model': 1.4287993479520082, 'learning_rate/qwen_vl_interface': 4.87580129775214e-06, 'learning_rate/action_model': 4.87580129775214e-05, 'epoch': 1.03}) +[2026-06-16 02:43:59,683][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.07933413237333298, 'train/grad_norm_pre_clip': 9.8085355758667, 'timing/data': 0.02617420256137848, 'timing/model': 1.429125927388668, 'learning_rate/qwen_vl_interface': 4.872765418191156e-06, 'learning_rate/action_model': 4.872765418191155e-05, 'epoch': 1.03}) +[2026-06-16 02:44:05,583][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.15463528037071228, 'train/grad_norm_pre_clip': 11.055339813232422, 'timing/data': 0.026405716314911842, 'timing/model': 1.4220683388411999, 'learning_rate/qwen_vl_interface': 4.8697296936970575e-06, 'learning_rate/action_model': 4.869729693697057e-05, 'epoch': 1.04}) +[2026-06-16 02:44:11,592][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.20740193128585815, 'train/grad_norm_pre_clip': 9.616686820983887, 'timing/data': 0.026336178183555603, 'timing/model': 1.4964831322431564, 'learning_rate/qwen_vl_interface': 4.8666941255177185e-06, 'learning_rate/action_model': 4.8666941255177185e-05, 'epoch': 1.04}) +[2026-06-16 02:44:17,500][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.040134821087121964, 'train/grad_norm_pre_clip': 3.4122776985168457, 'timing/data': 0.026784010231494904, 'timing/model': 1.4400448650121689, 'learning_rate/qwen_vl_interface': 4.863658714900945e-06, 'learning_rate/action_model': 4.863658714900945e-05, 'epoch': 1.04}) +[2026-06-16 02:44:23,558][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.0332307443022728, 'train/grad_norm_pre_clip': 1.5181169509887695, 'timing/data': 0.03180915117263794, 'timing/model': 1.7622034065425396, 'learning_rate/qwen_vl_interface': 4.860623463094483e-06, 'learning_rate/action_model': 4.860623463094483e-05, 'epoch': 1.04}) +[2026-06-16 02:44:29,549][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.0077701592817902565, 'train/grad_norm_pre_clip': 0.8367686867713928, 'timing/data': 0.031487325206398964, 'timing/model': 1.5328111425042152, 'learning_rate/qwen_vl_interface': 4.857588371346006e-06, 'learning_rate/action_model': 4.8575883713460057e-05, 'epoch': 1.04}) +[2026-06-16 02:44:35,756][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.15188054740428925, 'train/grad_norm_pre_clip': 9.365428924560547, 'timing/data': 0.033088380470871925, 'timing/model': 1.5221004225313663, 'learning_rate/qwen_vl_interface': 4.854553440903127e-06, 'learning_rate/action_model': 4.854553440903126e-05, 'epoch': 1.04}) +[2026-06-16 02:44:41,950][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.1556219607591629, 'train/grad_norm_pre_clip': 14.908615112304688, 'timing/data': 0.03230326063930988, 'timing/model': 1.5080835651606321, 'learning_rate/qwen_vl_interface': 4.85151867301339e-06, 'learning_rate/action_model': 4.85151867301339e-05, 'epoch': 1.04}) +[2026-06-16 02:44:48,247][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.03397836163640022, 'train/grad_norm_pre_clip': 2.1936933994293213, 'timing/data': 0.032693833112716675, 'timing/model': 1.5330001078546047, 'learning_rate/qwen_vl_interface': 4.848484068924274e-06, 'learning_rate/action_model': 4.848484068924275e-05, 'epoch': 1.04}) +[2026-06-16 02:44:54,501][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.12855495512485504, 'train/grad_norm_pre_clip': 8.553606986999512, 'timing/data': 0.03248624503612518, 'timing/model': 1.7809428684413433, 'learning_rate/qwen_vl_interface': 4.845449629883193e-06, 'learning_rate/action_model': 4.8454496298831934e-05, 'epoch': 1.04}) +[2026-06-16 02:45:01,827][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.12435200810432434, 'train/grad_norm_pre_clip': 8.865519523620605, 'timing/data': 0.04731919430196285, 'timing/model': 1.7286341488361359, 'learning_rate/qwen_vl_interface': 4.842415357137487e-06, 'learning_rate/action_model': 4.842415357137487e-05, 'epoch': 1.04}) +[2026-06-16 02:45:08,936][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.07331989705562592, 'train/grad_norm_pre_clip': 11.20358943939209, 'timing/data': 0.04839631915092468, 'timing/model': 1.7212358172982931, 'learning_rate/qwen_vl_interface': 4.839381251934432e-06, 'learning_rate/action_model': 4.8393812519344316e-05, 'epoch': 1.04}) +[2026-06-16 02:45:15,737][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.010690716095268726, 'train/grad_norm_pre_clip': 1.3903648853302002, 'timing/data': 0.04651248641312122, 'timing/model': 1.5814680475741625, 'learning_rate/qwen_vl_interface': 4.836347315521233e-06, 'learning_rate/action_model': 4.836347315521233e-05, 'epoch': 1.04}) +[2026-06-16 02:45:22,596][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.22954733669757843, 'train/grad_norm_pre_clip': 10.309439659118652, 'timing/data': 0.05067247897386551, 'timing/model': 1.6666358429938555, 'learning_rate/qwen_vl_interface': 4.833313549145029e-06, 'learning_rate/action_model': 4.833313549145029e-05, 'epoch': 1.04}) +[2026-06-16 02:45:29,439][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.06098369136452675, 'train/grad_norm_pre_clip': 4.559554576873779, 'timing/data': 0.04841272719204426, 'timing/model': 1.624307220801711, 'learning_rate/qwen_vl_interface': 4.830279954052884e-06, 'learning_rate/action_model': 4.8302799540528836e-05, 'epoch': 1.04}) +[2026-06-16 02:45:36,430][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.1502894014120102, 'train/grad_norm_pre_clip': 13.155197143554688, 'timing/data': 0.04563058912754059, 'timing/model': 2.0103302020579576, 'learning_rate/qwen_vl_interface': 4.827246531491797e-06, 'learning_rate/action_model': 4.827246531491797e-05, 'epoch': 1.04}) +[2026-06-16 02:45:43,099][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.0963420495390892, 'train/grad_norm_pre_clip': 16.35677146911621, 'timing/data': 0.04541347175836563, 'timing/model': 1.716872725635767, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 1.04}) +[2026-06-16 02:45:50,160][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.05684215575456619, 'train/grad_norm_pre_clip': 5.377486228942871, 'timing/data': 0.05652174726128578, 'timing/model': 1.7723499033600092, 'learning_rate/qwen_vl_interface': 4.821180208950422e-06, 'learning_rate/action_model': 4.821180208950421e-05, 'epoch': 1.04}) +[2026-06-16 02:45:57,132][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.05632155388593674, 'train/grad_norm_pre_clip': 7.342297077178955, 'timing/data': 0.044457728043198586, 'timing/model': 1.7189778555184603, 'learning_rate/qwen_vl_interface': 4.81814731146377e-06, 'learning_rate/action_model': 4.8181473114637695e-05, 'epoch': 1.04}) +[2026-06-16 02:46:03,993][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.28829464316368103, 'train/grad_norm_pre_clip': 9.583488464355469, 'timing/data': 0.04594207741320133, 'timing/model': 1.6501555908471346, 'learning_rate/qwen_vl_interface': 4.8151145914954465e-06, 'learning_rate/action_model': 4.815114591495446e-05, 'epoch': 1.04}) +[2026-06-16 02:46:10,246][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.1066228449344635, 'train/grad_norm_pre_clip': 8.764049530029297, 'timing/data': 0.032705727964639664, 'timing/model': 1.5166310351341963, 'learning_rate/qwen_vl_interface': 4.812082050292088e-06, 'learning_rate/action_model': 4.8120820502920876e-05, 'epoch': 1.04}) +[2026-06-16 02:46:16,501][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.010710247792303562, 'train/grad_norm_pre_clip': 1.2902685403823853, 'timing/data': 0.03205196559429169, 'timing/model': 1.7934621404856443, 'learning_rate/qwen_vl_interface': 4.809049689100256e-06, 'learning_rate/action_model': 4.8090496891002555e-05, 'epoch': 1.04}) +[2026-06-16 02:46:22,548][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.010319100692868233, 'train/grad_norm_pre_clip': 1.0912641286849976, 'timing/data': 0.031387271359562874, 'timing/model': 1.5318749193102121, 'learning_rate/qwen_vl_interface': 4.80601750916644e-06, 'learning_rate/action_model': 4.8060175091664404e-05, 'epoch': 1.04}) +[2026-06-16 02:46:28,489][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.008125606924295425, 'train/grad_norm_pre_clip': 0.7477065920829773, 'timing/data': 0.025193439796566963, 'timing/model': 1.4330345802009106, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 1.04}) +[2026-06-16 02:46:34,389][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.10122092068195343, 'train/grad_norm_pre_clip': 14.376566886901855, 'timing/data': 0.025480497628450394, 'timing/model': 1.4386670235544443, 'learning_rate/qwen_vl_interface': 4.799953698058444e-06, 'learning_rate/action_model': 4.7999536980584434e-05, 'epoch': 1.04}) +[2026-06-16 02:46:40,287][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.10613703727722168, 'train/grad_norm_pre_clip': 5.9196977615356445, 'timing/data': 0.026497891172766685, 'timing/model': 1.4311750885099173, 'learning_rate/qwen_vl_interface': 4.796922069376867e-06, 'learning_rate/action_model': 4.796922069376867e-05, 'epoch': 1.04}) +[2026-06-16 02:54:17,760][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.03675809130072594, 'train/grad_norm_pre_clip': 6.596818447113037, 'eval/loss': 0.049443911761045456, 'eval/flap_precision': 0.9436997319034852, 'eval/flap_recall': 0.9552238805970149, 'eval/flap_f1': 0.9494268374915712, 'eval/flap_rate_model': 0.20722222222222222, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9494268374915712, 'timing/data': 0.026713887229561806, 'timing/model': 1.4633416961878538, 'learning_rate/qwen_vl_interface': 4.793890626938511e-06, 'learning_rate/action_model': 4.793890626938511e-05, 'epoch': 1.04}) +[2026-06-16 02:54:17,760][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2700_state +[2026-06-16 02:54:17,764][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:54:23,691][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2700_state/model.safetensors +[2026-06-16 02:54:33,989][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2700_state/optimizer.bin +[2026-06-16 02:54:33,989][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2700_state/sampler.bin +[2026-06-16 02:54:33,989][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2700_state/sampler_1.bin +[2026-06-16 02:54:33,990][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2700_state/random_states_0.pkl +[2026-06-16 02:54:34,088][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2700 +[2026-06-16 02:54:34,088][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:54:34,097][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:54:40,120][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.034900035709142685, 'train/grad_norm_pre_clip': 4.665510177612305, 'timing/data': 0.038031645119190216, 'timing/model': 1.5007664803415537, 'learning_rate/qwen_vl_interface': 4.7908593719894914e-06, 'learning_rate/action_model': 4.790859371989491e-05, 'epoch': 1.05}) +[2026-06-16 02:54:46,026][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.05174944922327995, 'train/grad_norm_pre_clip': 6.042603492736816, 'timing/data': 0.02629830688238144, 'timing/model': 1.6563734095543623, 'learning_rate/qwen_vl_interface': 4.78782830577584e-06, 'learning_rate/action_model': 4.7878283057758396e-05, 'epoch': 1.05}) +[2026-06-16 02:54:51,801][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.12585707008838654, 'train/grad_norm_pre_clip': 9.680818557739258, 'timing/data': 0.027749627828598022, 'timing/model': 1.5203746054321527, 'learning_rate/qwen_vl_interface': 4.784797429543513e-06, 'learning_rate/action_model': 4.784797429543513e-05, 'epoch': 1.05}) +[2026-06-16 02:54:57,803][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.03421027958393097, 'train/grad_norm_pre_clip': 7.926631450653076, 'timing/data': 0.025643784552812576, 'timing/model': 1.5209222082048655, 'learning_rate/qwen_vl_interface': 4.781766744538393e-06, 'learning_rate/action_model': 4.781766744538393e-05, 'epoch': 1.05}) +[2026-06-16 02:55:03,696][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.0435141883790493, 'train/grad_norm_pre_clip': 3.8384132385253906, 'timing/data': 0.02683999203145504, 'timing/model': 1.513251792639494, 'learning_rate/qwen_vl_interface': 4.778736252006277e-06, 'learning_rate/action_model': 4.7787362520062765e-05, 'epoch': 1.05}) +[2026-06-16 02:55:09,520][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.2422291785478592, 'train/grad_norm_pre_clip': 13.198114395141602, 'timing/data': 0.026153385639190674, 'timing/model': 1.4482431896030903, 'learning_rate/qwen_vl_interface': 4.775705953192885e-06, 'learning_rate/action_model': 4.7757059531928856e-05, 'epoch': 1.05}) +[2026-06-16 02:55:15,379][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.1609804630279541, 'train/grad_norm_pre_clip': 11.874455451965332, 'timing/data': 0.025930095463991165, 'timing/model': 1.4279041159898043, 'learning_rate/qwen_vl_interface': 4.7726758493438615e-06, 'learning_rate/action_model': 4.7726758493438613e-05, 'epoch': 1.05}) +[2026-06-16 02:55:21,288][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.27139198780059814, 'train/grad_norm_pre_clip': 32.84832763671875, 'timing/data': 0.026648296043276787, 'timing/model': 1.6458440832793713, 'learning_rate/qwen_vl_interface': 4.769645941704766e-06, 'learning_rate/action_model': 4.769645941704766e-05, 'epoch': 1.05}) +[2026-06-16 02:55:26,979][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.3715190887451172, 'train/grad_norm_pre_clip': 18.4888973236084, 'timing/data': 0.026777803897857666, 'timing/model': 1.435989709571004, 'learning_rate/qwen_vl_interface': 4.766616231521081e-06, 'learning_rate/action_model': 4.7666162315210806e-05, 'epoch': 1.05}) +[2026-06-16 02:55:32,879][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.1612684279680252, 'train/grad_norm_pre_clip': 9.619853019714355, 'timing/data': 0.025809958577156067, 'timing/model': 1.437422700226307, 'learning_rate/qwen_vl_interface': 4.763586720038202e-06, 'learning_rate/action_model': 4.7635867200382014e-05, 'epoch': 1.05}) +[2026-06-16 02:55:38,682][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.009529038332402706, 'train/grad_norm_pre_clip': 1.6981697082519531, 'timing/data': 0.025593889877200127, 'timing/model': 1.4271922204643488, 'learning_rate/qwen_vl_interface': 4.76055740850145e-06, 'learning_rate/action_model': 4.76055740850145e-05, 'epoch': 1.05}) +[2026-06-16 02:55:44,876][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.04724428802728653, 'train/grad_norm_pre_clip': 3.978745460510254, 'timing/data': 0.031753841787576675, 'timing/model': 1.5241288859397173, 'learning_rate/qwen_vl_interface': 4.757528298156058e-06, 'learning_rate/action_model': 4.757528298156058e-05, 'epoch': 1.05}) +[2026-06-16 02:55:51,373][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.14384473860263824, 'train/grad_norm_pre_clip': 11.176957130432129, 'timing/data': 0.03308121673762798, 'timing/model': 1.6824494879692793, 'learning_rate/qwen_vl_interface': 4.754499390247181e-06, 'learning_rate/action_model': 4.754499390247181e-05, 'epoch': 1.05}) +[2026-06-16 02:55:57,938][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.10313235968351364, 'train/grad_norm_pre_clip': 13.76681137084961, 'timing/data': 0.027950728312134743, 'timing/model': 1.709764352068305, 'learning_rate/qwen_vl_interface': 4.751470686019887e-06, 'learning_rate/action_model': 4.751470686019888e-05, 'epoch': 1.05}) +[2026-06-16 02:56:03,974][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.32145974040031433, 'train/grad_norm_pre_clip': 20.12221336364746, 'timing/data': 0.030995739623904228, 'timing/model': 1.6709565035998821, 'learning_rate/qwen_vl_interface': 4.7484421867191645e-06, 'learning_rate/action_model': 4.748442186719165e-05, 'epoch': 1.05}) +[2026-06-16 02:56:10,380][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.0042032781057059765, 'train/grad_norm_pre_clip': 1.1266517639160156, 'timing/data': 0.03758993744850159, 'timing/model': 1.5310595389455557, 'learning_rate/qwen_vl_interface': 4.745413893589913e-06, 'learning_rate/action_model': 4.7454138935899125e-05, 'epoch': 1.05}) +[2026-06-16 02:56:16,682][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.03514017537236214, 'train/grad_norm_pre_clip': 14.882826805114746, 'timing/data': 0.042863087728619576, 'timing/model': 1.5216147992759943, 'learning_rate/qwen_vl_interface': 4.7423858078769506e-06, 'learning_rate/action_model': 4.742385807876951e-05, 'epoch': 1.05}) +[2026-06-16 02:56:23,170][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.12849274277687073, 'train/grad_norm_pre_clip': 7.5199456214904785, 'timing/data': 0.03590086288750172, 'timing/model': 1.5972453858703375, 'learning_rate/qwen_vl_interface': 4.739357930825008e-06, 'learning_rate/action_model': 4.739357930825008e-05, 'epoch': 1.05}) +[2026-06-16 02:56:29,580][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.048300959169864655, 'train/grad_norm_pre_clip': 7.318377494812012, 'timing/data': 0.035322921350598335, 'timing/model': 1.584511524066329, 'learning_rate/qwen_vl_interface': 4.736330263678731e-06, 'learning_rate/action_model': 4.736330263678731e-05, 'epoch': 1.05}) +[2026-06-16 02:56:35,919][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.0009452244266867638, 'train/grad_norm_pre_clip': 0.1020292267203331, 'timing/data': 0.03422229923307896, 'timing/model': 1.7716342192143202, 'learning_rate/qwen_vl_interface': 4.733302807682679e-06, 'learning_rate/action_model': 4.733302807682679e-05, 'epoch': 1.05}) +[2026-06-16 02:56:41,981][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.08626189082860947, 'train/grad_norm_pre_clip': 9.77907943725586, 'timing/data': 0.035193296149373055, 'timing/model': 1.5312269926071167, 'learning_rate/qwen_vl_interface': 4.7302755640813255e-06, 'learning_rate/action_model': 4.7302755640813256e-05, 'epoch': 1.05}) +[2026-06-16 02:56:48,281][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.08387894183397293, 'train/grad_norm_pre_clip': 20.08421516418457, 'timing/data': 0.031246626749634743, 'timing/model': 1.540890047326684, 'learning_rate/qwen_vl_interface': 4.727248534119058e-06, 'learning_rate/action_model': 4.7272485341190574e-05, 'epoch': 1.05}) +[2026-06-16 02:56:54,667][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.04269606992602348, 'train/grad_norm_pre_clip': 4.7762579917907715, 'timing/data': 0.03282465972006321, 'timing/model': 1.595046928152442, 'learning_rate/qwen_vl_interface': 4.72422171904017e-06, 'learning_rate/action_model': 4.7242217190401697e-05, 'epoch': 1.05}) +[2026-06-16 02:57:01,063][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.10217097401618958, 'train/grad_norm_pre_clip': 8.843487739562988, 'timing/data': 0.03176136501133442, 'timing/model': 1.5882233399897814, 'learning_rate/qwen_vl_interface': 4.721195120088873e-06, 'learning_rate/action_model': 4.721195120088873e-05, 'epoch': 1.05}) +[2026-06-16 02:57:08,183][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.12057708948850632, 'train/grad_norm_pre_clip': 5.026678562164307, 'timing/data': 0.034790387377142906, 'timing/model': 1.5902568642050028, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 1.05}) +[2026-06-16 02:57:14,590][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.020977558568120003, 'train/grad_norm_pre_clip': 2.341440439224243, 'timing/data': 0.031529296189546585, 'timing/model': 1.819767750799656, 'learning_rate/qwen_vl_interface': 4.715142575545445e-06, 'learning_rate/action_model': 4.7151425755454454e-05, 'epoch': 1.06}) +[2026-06-16 02:57:20,847][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.2548648715019226, 'train/grad_norm_pre_clip': 10.265955924987793, 'timing/data': 0.03223409689962864, 'timing/model': 1.7018236964941025, 'learning_rate/qwen_vl_interface': 4.712116632441287e-06, 'learning_rate/action_model': 4.7121166324412866e-05, 'epoch': 1.06}) +[2026-06-16 02:57:27,001][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.05954313650727272, 'train/grad_norm_pre_clip': 9.287910461425781, 'timing/data': 0.036944374442100525, 'timing/model': 1.6098650377243757, 'learning_rate/qwen_vl_interface': 4.709090910440664e-06, 'learning_rate/action_model': 4.7090909104406634e-05, 'epoch': 1.06}) +[2026-06-16 02:57:33,249][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.09476497769355774, 'train/grad_norm_pre_clip': 14.717400550842285, 'timing/data': 0.032130250707268715, 'timing/model': 1.5149699971079826, 'learning_rate/qwen_vl_interface': 4.7060654107873335e-06, 'learning_rate/action_model': 4.7060654107873335e-05, 'epoch': 1.06}) +[2026-06-16 02:57:39,747][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.02581864967942238, 'train/grad_norm_pre_clip': 2.286102771759033, 'timing/data': 0.0332459881901741, 'timing/model': 1.5106369964778423, 'learning_rate/qwen_vl_interface': 4.7030401347249685e-06, 'learning_rate/action_model': 4.703040134724968e-05, 'epoch': 1.06}) +[2026-06-16 02:57:46,047][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.01212065014988184, 'train/grad_norm_pre_clip': 2.4790196418762207, 'timing/data': 0.03251199051737785, 'timing/model': 1.5369794834405184, 'learning_rate/qwen_vl_interface': 4.700015083497143e-06, 'learning_rate/action_model': 4.700015083497142e-05, 'epoch': 1.06}) +[2026-06-16 02:57:52,290][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.1438102275133133, 'train/grad_norm_pre_clip': 15.352432250976562, 'timing/data': 0.03244071267545223, 'timing/model': 1.7419119216501713, 'learning_rate/qwen_vl_interface': 4.696990258347342e-06, 'learning_rate/action_model': 4.696990258347342e-05, 'epoch': 1.06}) +[2026-06-16 02:57:58,546][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.16195423901081085, 'train/grad_norm_pre_clip': 18.079500198364258, 'timing/data': 0.03475543111562729, 'timing/model': 1.7047688327729702, 'learning_rate/qwen_vl_interface': 4.693965660518957e-06, 'learning_rate/action_model': 4.693965660518957e-05, 'epoch': 1.06}) +[2026-06-16 02:58:04,526][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.10958418995141983, 'train/grad_norm_pre_clip': 9.354738235473633, 'timing/data': 0.02616397850215435, 'timing/model': 1.5313474405556917, 'learning_rate/qwen_vl_interface': 4.690941291255285e-06, 'learning_rate/action_model': 4.6909412912552844e-05, 'epoch': 1.06}) +[2026-06-16 02:58:10,533][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.1803773194551468, 'train/grad_norm_pre_clip': 7.346717834472656, 'timing/data': 0.02610168047249317, 'timing/model': 1.531892066821456, 'learning_rate/qwen_vl_interface': 4.68791715179953e-06, 'learning_rate/action_model': 4.68791715179953e-05, 'epoch': 1.06}) +[2026-06-16 02:58:16,520][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.04989541321992874, 'train/grad_norm_pre_clip': 7.1787309646606445, 'timing/data': 0.027447199448943138, 'timing/model': 1.5156804863363504, 'learning_rate/qwen_vl_interface': 4.684893243394804e-06, 'learning_rate/action_model': 4.6848932433948036e-05, 'epoch': 1.06}) +[2026-06-16 02:58:22,521][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.2597883939743042, 'train/grad_norm_pre_clip': 15.776163101196289, 'timing/data': 0.02635880745947361, 'timing/model': 1.506819287315011, 'learning_rate/qwen_vl_interface': 4.68186956728412e-06, 'learning_rate/action_model': 4.68186956728412e-05, 'epoch': 1.06}) +[2026-06-16 02:58:28,528][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.20392952859401703, 'train/grad_norm_pre_clip': 11.129842758178711, 'timing/data': 0.03120375983417034, 'timing/model': 1.733184952288866, 'learning_rate/qwen_vl_interface': 4.678846124710395e-06, 'learning_rate/action_model': 4.678846124710394e-05, 'epoch': 1.06}) +[2026-06-16 02:58:34,204][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.11158905923366547, 'train/grad_norm_pre_clip': 9.746818542480469, 'timing/data': 0.026152240112423897, 'timing/model': 1.451358649879694, 'learning_rate/qwen_vl_interface': 4.675822916916454e-06, 'learning_rate/action_model': 4.6758229169164536e-05, 'epoch': 1.06}) +[2026-06-16 02:58:40,097][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.16548781096935272, 'train/grad_norm_pre_clip': 10.364249229431152, 'timing/data': 0.0269144456833601, 'timing/model': 1.4521600008010864, 'learning_rate/qwen_vl_interface': 4.6727999451450236e-06, 'learning_rate/action_model': 4.672799945145023e-05, 'epoch': 1.06}) +[2026-06-16 02:58:45,997][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.09980102628469467, 'train/grad_norm_pre_clip': 6.148072242736816, 'timing/data': 0.025659872218966484, 'timing/model': 1.4416952673345804, 'learning_rate/qwen_vl_interface': 4.669777210638733e-06, 'learning_rate/action_model': 4.669777210638733e-05, 'epoch': 1.06}) +[2026-06-16 02:58:52,000][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.0646066814661026, 'train/grad_norm_pre_clip': 2.5312483310699463, 'timing/data': 0.02751391939818859, 'timing/model': 1.5130458679050207, 'learning_rate/qwen_vl_interface': 4.666754714640115e-06, 'learning_rate/action_model': 4.666754714640115e-05, 'epoch': 1.06}) +[2026-06-16 02:58:57,903][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.11309295892715454, 'train/grad_norm_pre_clip': 7.907665729522705, 'timing/data': 0.025925640016794205, 'timing/model': 1.5053077768534422, 'learning_rate/qwen_vl_interface': 4.663732458391602e-06, 'learning_rate/action_model': 4.663732458391601e-05, 'epoch': 1.06}) +[2026-06-16 02:59:04,143][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.030837133526802063, 'train/grad_norm_pre_clip': 2.8966002464294434, 'timing/data': 0.0314355343580246, 'timing/model': 1.7347111348062754, 'learning_rate/qwen_vl_interface': 4.6607104431355295e-06, 'learning_rate/action_model': 4.66071044313553e-05, 'epoch': 1.06}) +[2026-06-16 02:59:10,149][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.057520750910043716, 'train/grad_norm_pre_clip': 6.149730205535889, 'timing/data': 0.033195434138178825, 'timing/model': 1.5370969101786613, 'learning_rate/qwen_vl_interface': 4.657688670114135e-06, 'learning_rate/action_model': 4.657688670114135e-05, 'epoch': 1.06}) +[2026-06-16 02:59:16,348][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.049982499331235886, 'train/grad_norm_pre_clip': 3.7821693420410156, 'timing/data': 0.030794886872172356, 'timing/model': 1.5324431955814362, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 1.06}) +[2026-06-16 02:59:22,550][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.12293446063995361, 'train/grad_norm_pre_clip': 5.1130452156066895, 'timing/data': 0.03138483315706253, 'timing/model': 1.5356547106057405, 'learning_rate/qwen_vl_interface': 4.651645855743825e-06, 'learning_rate/action_model': 4.651645855743825e-05, 'epoch': 1.06}) +[2026-06-16 02:59:28,742][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.03908867388963699, 'train/grad_norm_pre_clip': 2.0545036792755127, 'timing/data': 0.030876964330673218, 'timing/model': 1.4897614736109972, 'learning_rate/qwen_vl_interface': 4.648624816878883e-06, 'learning_rate/action_model': 4.6486248168788825e-05, 'epoch': 1.06}) +[2026-06-16 02:59:34,950][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.05748667195439339, 'train/grad_norm_pre_clip': 3.4520809650421143, 'timing/data': 0.031175976619124413, 'timing/model': 1.5465776715427637, 'learning_rate/qwen_vl_interface': 4.6456040252165615e-06, 'learning_rate/action_model': 4.645604025216562e-05, 'epoch': 1.06}) +[2026-06-16 02:59:41,152][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.005916799418628216, 'train/grad_norm_pre_clip': 0.8801791071891785, 'timing/data': 0.032565442845225334, 'timing/model': 1.7369412537664175, 'learning_rate/qwen_vl_interface': 4.642583481998596e-06, 'learning_rate/action_model': 4.642583481998596e-05, 'epoch': 1.06}) +[2026-06-16 02:59:47,149][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.27029091119766235, 'train/grad_norm_pre_clip': 13.662585258483887, 'timing/data': 0.03182920813560486, 'timing/model': 1.5357591677457094, 'learning_rate/qwen_vl_interface': 4.639563188466616e-06, 'learning_rate/action_model': 4.6395631884666165e-05, 'epoch': 1.06}) +[2026-06-16 02:59:53,302][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.13084521889686584, 'train/grad_norm_pre_clip': 6.003938674926758, 'timing/data': 0.027303943410515785, 'timing/model': 1.4961989149451256, 'learning_rate/qwen_vl_interface': 4.636543145862152e-06, 'learning_rate/action_model': 4.6365431458621523e-05, 'epoch': 1.07}) +[2026-06-16 02:59:59,290][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.0045041912235319614, 'train/grad_norm_pre_clip': 0.6339015364646912, 'timing/data': 0.02553476393222809, 'timing/model': 1.4399493802338839, 'learning_rate/qwen_vl_interface': 4.633523355426629e-06, 'learning_rate/action_model': 4.6335233554266295e-05, 'epoch': 1.07}) +[2026-06-16 03:00:05,187][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.02818112075328827, 'train/grad_norm_pre_clip': 8.647375106811523, 'timing/data': 0.025761742144823074, 'timing/model': 1.4211675692349672, 'learning_rate/qwen_vl_interface': 4.6305038184013675e-06, 'learning_rate/action_model': 4.6305038184013675e-05, 'epoch': 1.07}) +[2026-06-16 03:00:11,196][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.059628985822200775, 'train/grad_norm_pre_clip': 3.7404375076293945, 'timing/data': 0.0266768466681242, 'timing/model': 1.7501796875149012, 'learning_rate/qwen_vl_interface': 4.627484536027589e-06, 'learning_rate/action_model': 4.627484536027588e-05, 'epoch': 1.07}) +[2026-06-16 03:00:17,255][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.014862162992358208, 'train/grad_norm_pre_clip': 2.4513964653015137, 'timing/data': 0.031554631888866425, 'timing/model': 1.5396985821425915, 'learning_rate/qwen_vl_interface': 4.624465509546399e-06, 'learning_rate/action_model': 4.6244655095463986e-05, 'epoch': 1.07}) +[2026-06-16 03:00:23,463][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.08393019437789917, 'train/grad_norm_pre_clip': 9.224533081054688, 'timing/data': 0.033441150560975075, 'timing/model': 1.5371818281710148, 'learning_rate/qwen_vl_interface': 4.6214467401988106e-06, 'learning_rate/action_model': 4.621446740198811e-05, 'epoch': 1.07}) +[2026-06-16 03:00:29,662][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.06735485047101974, 'train/grad_norm_pre_clip': 4.779604911804199, 'timing/data': 0.03398727625608444, 'timing/model': 1.5341885443776846, 'learning_rate/qwen_vl_interface': 4.6184282292257245e-06, 'learning_rate/action_model': 4.618428229225725e-05, 'epoch': 1.07}) +[2026-06-16 03:00:35,847][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.04736510291695595, 'train/grad_norm_pre_clip': 2.599553346633911, 'timing/data': 0.03404770419001579, 'timing/model': 1.5271680783480406, 'learning_rate/qwen_vl_interface': 4.615409977867938e-06, 'learning_rate/action_model': 4.615409977867937e-05, 'epoch': 1.07}) +[2026-06-16 03:00:42,449][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.009767984971404076, 'train/grad_norm_pre_clip': 1.6551882028579712, 'timing/data': 0.03186305612325668, 'timing/model': 1.531511101871729, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 1.07}) +[2026-06-16 03:00:48,692][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.14465175569057465, 'train/grad_norm_pre_clip': 9.935277938842773, 'timing/data': 0.03249874338507652, 'timing/model': 1.7703413497656584, 'learning_rate/qwen_vl_interface': 4.609374258960908e-06, 'learning_rate/action_model': 4.6093742589609074e-05, 'epoch': 1.07}) +[2026-06-16 03:00:54,749][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.1292811632156372, 'train/grad_norm_pre_clip': 16.095462799072266, 'timing/data': 0.03110363706946373, 'timing/model': 1.533850908279419, 'learning_rate/qwen_vl_interface': 4.60635679389272e-06, 'learning_rate/action_model': 4.60635679389272e-05, 'epoch': 1.07}) +[2026-06-16 03:01:01,049][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.3662472665309906, 'train/grad_norm_pre_clip': 15.385375022888184, 'timing/data': 0.035318523645401, 'timing/model': 1.5369542930275202, 'learning_rate/qwen_vl_interface': 4.603339593401943e-06, 'learning_rate/action_model': 4.603339593401942e-05, 'epoch': 1.07}) +[2026-06-16 03:01:07,248][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.16865119338035583, 'train/grad_norm_pre_clip': 7.650055408477783, 'timing/data': 0.03363896906375885, 'timing/model': 1.5335265845060349, 'learning_rate/qwen_vl_interface': 4.600322658728832e-06, 'learning_rate/action_model': 4.600322658728832e-05, 'epoch': 1.07}) +[2026-06-16 03:01:13,344][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.01853439025580883, 'train/grad_norm_pre_clip': 3.496500015258789, 'timing/data': 0.03279459848999977, 'timing/model': 1.5297542195767164, 'learning_rate/qwen_vl_interface': 4.597305991113537e-06, 'learning_rate/action_model': 4.597305991113537e-05, 'epoch': 1.07}) +[2026-06-16 03:01:19,447][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.05781460180878639, 'train/grad_norm_pre_clip': 7.500576019287109, 'timing/data': 0.03115294687449932, 'timing/model': 1.525295214727521, 'learning_rate/qwen_vl_interface': 4.5942895917960935e-06, 'learning_rate/action_model': 4.5942895917960935e-05, 'epoch': 1.07}) +[2026-06-16 03:01:25,692][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.07855407148599625, 'train/grad_norm_pre_clip': 12.947182655334473, 'timing/data': 0.032688532024621964, 'timing/model': 1.7929909005761147, 'learning_rate/qwen_vl_interface': 4.59127346201643e-06, 'learning_rate/action_model': 4.59127346201643e-05, 'epoch': 1.07}) +[2026-06-16 03:01:31,747][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.07095400243997574, 'train/grad_norm_pre_clip': 7.401782989501953, 'timing/data': 0.03201320394873619, 'timing/model': 1.5360846295952797, 'learning_rate/qwen_vl_interface': 4.588257603014364e-06, 'learning_rate/action_model': 4.588257603014364e-05, 'epoch': 1.07}) +[2026-06-16 03:01:37,944][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.1645633578300476, 'train/grad_norm_pre_clip': 9.427257537841797, 'timing/data': 0.03243919089436531, 'timing/model': 1.534168677404523, 'learning_rate/qwen_vl_interface': 4.585242016029602e-06, 'learning_rate/action_model': 4.5852420160296015e-05, 'epoch': 1.07}) +[2026-06-16 03:01:44,180][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.01776680164039135, 'train/grad_norm_pre_clip': 1.9922856092453003, 'timing/data': 0.03220335580408573, 'timing/model': 1.5304216500371695, 'learning_rate/qwen_vl_interface': 4.582226702301737e-06, 'learning_rate/action_model': 4.5822267023017375e-05, 'epoch': 1.07}) +[2026-06-16 03:01:50,584][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.04534842446446419, 'train/grad_norm_pre_clip': 2.592080593109131, 'timing/data': 0.03148754686117172, 'timing/model': 1.5963269975036383, 'learning_rate/qwen_vl_interface': 4.57921166307025e-06, 'learning_rate/action_model': 4.57921166307025e-05, 'epoch': 1.07}) +[2026-06-16 03:01:56,875][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.011082999408245087, 'train/grad_norm_pre_clip': 1.8733694553375244, 'timing/data': 0.03243291936814785, 'timing/model': 1.5582656431943178, 'learning_rate/qwen_vl_interface': 4.576196899574511e-06, 'learning_rate/action_model': 4.576196899574511e-05, 'epoch': 1.07}) +[2026-06-16 03:02:03,225][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.05802680552005768, 'train/grad_norm_pre_clip': 20.487232208251953, 'timing/data': 0.032822027802467346, 'timing/model': 1.7622743248939514, 'learning_rate/qwen_vl_interface': 4.573182413053775e-06, 'learning_rate/action_model': 4.5731824130537746e-05, 'epoch': 1.07}) +[2026-06-16 03:02:09,297][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.036200862377882004, 'train/grad_norm_pre_clip': 2.9858670234680176, 'timing/data': 0.03312770463526249, 'timing/model': 1.5277267787605524, 'learning_rate/qwen_vl_interface': 4.570168204747183e-06, 'learning_rate/action_model': 4.570168204747183e-05, 'epoch': 1.07}) +[2026-06-16 03:02:15,815][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.06541195511817932, 'train/grad_norm_pre_clip': 5.740769863128662, 'timing/data': 0.03722249157726765, 'timing/model': 1.630058664828539, 'learning_rate/qwen_vl_interface': 4.567154275893762e-06, 'learning_rate/action_model': 4.5671542758937624e-05, 'epoch': 1.07}) +[2026-06-16 03:02:22,182][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.135535329580307, 'train/grad_norm_pre_clip': 11.992767333984375, 'timing/data': 0.034928007051348686, 'timing/model': 1.5969125293195248, 'learning_rate/qwen_vl_interface': 4.564140627732426e-06, 'learning_rate/action_model': 4.564140627732426e-05, 'epoch': 1.07}) +[2026-06-16 03:02:28,497][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.014924176037311554, 'train/grad_norm_pre_clip': 2.8590571880340576, 'timing/data': 0.0367073155939579, 'timing/model': 1.5983823854476213, 'learning_rate/qwen_vl_interface': 4.56112726150197e-06, 'learning_rate/action_model': 4.56112726150197e-05, 'epoch': 1.07}) +[2026-06-16 03:02:34,889][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.05264577642083168, 'train/grad_norm_pre_clip': 8.45251750946045, 'timing/data': 0.034043870866298676, 'timing/model': 1.5664270631968975, 'learning_rate/qwen_vl_interface': 4.558114178441076e-06, 'learning_rate/action_model': 4.558114178441076e-05, 'epoch': 1.08}) +[2026-06-16 03:02:41,139][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.07762981206178665, 'train/grad_norm_pre_clip': 6.565985202789307, 'timing/data': 0.026747193187475204, 'timing/model': 1.7428477834910154, 'learning_rate/qwen_vl_interface': 4.555101379788308e-06, 'learning_rate/action_model': 4.555101379788308e-05, 'epoch': 1.08}) +[2026-06-16 03:02:46,798][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.05504297465085983, 'train/grad_norm_pre_clip': 8.108421325683594, 'timing/data': 0.025927945971488953, 'timing/model': 1.4435734283179045, 'learning_rate/qwen_vl_interface': 4.552088866782115e-06, 'learning_rate/action_model': 4.552088866782115e-05, 'epoch': 1.08}) +[2026-06-16 03:02:52,695][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.021433072164654732, 'train/grad_norm_pre_clip': 1.963509440422058, 'timing/data': 0.026805831119418144, 'timing/model': 1.4301852453500032, 'learning_rate/qwen_vl_interface': 4.5490766406608265e-06, 'learning_rate/action_model': 4.549076640660827e-05, 'epoch': 1.08}) +[2026-06-16 03:02:58,599][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.008454778231680393, 'train/grad_norm_pre_clip': 2.379287004470825, 'timing/data': 0.025638915598392487, 'timing/model': 1.4361687675118446, 'learning_rate/qwen_vl_interface': 4.546064702662655e-06, 'learning_rate/action_model': 4.546064702662655e-05, 'epoch': 1.08}) +[2026-06-16 03:03:04,405][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.18749666213989258, 'train/grad_norm_pre_clip': 13.306998252868652, 'timing/data': 0.02563750557601452, 'timing/model': 1.4346965085715055, 'learning_rate/qwen_vl_interface': 4.543053054025693e-06, 'learning_rate/action_model': 4.5430530540256926e-05, 'epoch': 1.08}) +[2026-06-16 03:03:10,398][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.1675952672958374, 'train/grad_norm_pre_clip': 16.935020446777344, 'timing/data': 0.02496720850467682, 'timing/model': 1.4887615013867617, 'learning_rate/qwen_vl_interface': 4.540041695987917e-06, 'learning_rate/action_model': 4.540041695987917e-05, 'epoch': 1.08}) +[2026-06-16 03:03:16,313][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.06489703804254532, 'train/grad_norm_pre_clip': 11.00046157836914, 'timing/data': 0.02516517974436283, 'timing/model': 1.6593124512583017, 'learning_rate/qwen_vl_interface': 4.537030629787182e-06, 'learning_rate/action_model': 4.5370306297871826e-05, 'epoch': 1.08}) +[2026-06-16 03:03:22,006][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.3633245527744293, 'train/grad_norm_pre_clip': 30.717666625976562, 'timing/data': 0.026215754449367523, 'timing/model': 1.45715300925076, 'learning_rate/qwen_vl_interface': 4.534019856661226e-06, 'learning_rate/action_model': 4.534019856661226e-05, 'epoch': 1.08}) +[2026-06-16 03:03:27,992][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.4189222455024719, 'train/grad_norm_pre_clip': 22.85727310180664, 'timing/data': 0.042739398777484894, 'timing/model': 1.4470599368214607, 'learning_rate/qwen_vl_interface': 4.531009377847661e-06, 'learning_rate/action_model': 4.531009377847661e-05, 'epoch': 1.08}) +[2026-06-16 03:03:33,888][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.2290063053369522, 'train/grad_norm_pre_clip': 18.238874435424805, 'timing/data': 0.02614414319396019, 'timing/model': 1.4446395207196474, 'learning_rate/qwen_vl_interface': 4.527999194583982e-06, 'learning_rate/action_model': 4.527999194583981e-05, 'epoch': 1.08}) +[2026-06-16 03:03:39,685][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.08598018437623978, 'train/grad_norm_pre_clip': 8.99320125579834, 'timing/data': 0.025937093421816826, 'timing/model': 1.4379125777631998, 'learning_rate/qwen_vl_interface': 4.52498930810756e-06, 'learning_rate/action_model': 4.52498930810756e-05, 'epoch': 1.08}) +[2026-06-16 03:03:45,584][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.27811920642852783, 'train/grad_norm_pre_clip': 19.275556564331055, 'timing/data': 0.02609306573867798, 'timing/model': 1.4172892794013023, 'learning_rate/qwen_vl_interface': 4.521979719655648e-06, 'learning_rate/action_model': 4.5219797196556484e-05, 'epoch': 1.08}) +[2026-06-16 03:03:51,493][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.03733738139271736, 'train/grad_norm_pre_clip': 12.108538627624512, 'timing/data': 0.0248720720410347, 'timing/model': 1.6312061436474323, 'learning_rate/qwen_vl_interface': 4.518970430465377e-06, 'learning_rate/action_model': 4.518970430465376e-05, 'epoch': 1.08}) +[2026-06-16 03:03:57,177][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.148919939994812, 'train/grad_norm_pre_clip': 24.543855667114258, 'timing/data': 0.025218866765499115, 'timing/model': 1.432473435997963, 'learning_rate/qwen_vl_interface': 4.515961441773745e-06, 'learning_rate/action_model': 4.515961441773745e-05, 'epoch': 1.08}) +[2026-06-16 03:04:03,078][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.0184854157269001, 'train/grad_norm_pre_clip': 3.691986322402954, 'timing/data': 0.026484929025173187, 'timing/model': 1.4184761736541986, 'learning_rate/qwen_vl_interface': 4.5129527548176385e-06, 'learning_rate/action_model': 4.512952754817638e-05, 'epoch': 1.08}) +[2026-06-16 03:04:08,882][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.020428728312253952, 'train/grad_norm_pre_clip': 4.860313892364502, 'timing/data': 0.02602001652121544, 'timing/model': 1.4357531014829874, 'learning_rate/qwen_vl_interface': 4.509944370833813e-06, 'learning_rate/action_model': 4.509944370833813e-05, 'epoch': 1.08}) +[2026-06-16 03:04:14,782][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.015038884244859219, 'train/grad_norm_pre_clip': 2.7707738876342773, 'timing/data': 0.02514573000371456, 'timing/model': 1.4465904273092747, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 1.08}) +[2026-06-16 03:04:20,689][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.02839028835296631, 'train/grad_norm_pre_clip': 5.250579833984375, 'timing/data': 0.026133526116609573, 'timing/model': 1.4886525683104992, 'learning_rate/qwen_vl_interface': 4.503928516729416e-06, 'learning_rate/action_model': 4.503928516729416e-05, 'epoch': 1.08}) +[2026-06-16 03:04:26,589][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.28331297636032104, 'train/grad_norm_pre_clip': 12.946578025817871, 'timing/data': 0.02584994025528431, 'timing/model': 1.6364794038236141, 'learning_rate/qwen_vl_interface': 4.500921049081734e-06, 'learning_rate/action_model': 4.500921049081734e-05, 'epoch': 1.08}) +[2026-06-16 03:04:32,403][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.26140445470809937, 'train/grad_norm_pre_clip': 12.374204635620117, 'timing/data': 0.026347273960709572, 'timing/model': 1.5512757953256369, 'learning_rate/qwen_vl_interface': 4.497913889352113e-06, 'learning_rate/action_model': 4.4979138893521126e-05, 'epoch': 1.08}) +[2026-06-16 03:04:38,179][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.08915143460035324, 'train/grad_norm_pre_clip': 14.01138973236084, 'timing/data': 0.025696758180856705, 'timing/model': 1.4207044690847397, 'learning_rate/qwen_vl_interface': 4.494907038776684e-06, 'learning_rate/action_model': 4.4949070387766836e-05, 'epoch': 1.08}) +[2026-06-16 03:11:58,338][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.1316954642534256, 'train/grad_norm_pre_clip': 7.696872234344482, 'eval/loss': 0.05244449898600578, 'eval/flap_precision': 0.9584487534626038, 'eval/flap_recall': 0.9389416553595658, 'eval/flap_f1': 0.9485949280328992, 'eval/flap_rate_model': 0.20055555555555554, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9485949280328991, 'timing/data': 0.025390611961483955, 'timing/model': 1.432735750451684, 'learning_rate/qwen_vl_interface': 4.49190049859145e-06, 'learning_rate/action_model': 4.49190049859145e-05, 'epoch': 1.08}) +[2026-06-16 03:11:58,339][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2800_state +[2026-06-16 03:11:58,343][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 03:12:04,210][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2800_state/model.safetensors +[2026-06-16 03:12:13,924][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2800_state/optimizer.bin +[2026-06-16 03:12:13,925][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2800_state/sampler.bin +[2026-06-16 03:12:13,925][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2800_state/sampler_1.bin +[2026-06-16 03:12:13,927][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2800_state/random_states_0.pkl +[2026-06-16 03:12:14,032][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2800 +[2026-06-16 03:12:14,032][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 03:12:14,042][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 03:12:19,874][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.05706111341714859, 'train/grad_norm_pre_clip': 7.853425979614258, 'timing/data': 0.02928958460688591, 'timing/model': 1.4484741799533367, 'learning_rate/qwen_vl_interface': 4.488894270032284e-06, 'learning_rate/action_model': 4.4888942700322834e-05, 'epoch': 1.08}) +[2026-06-16 03:12:25,673][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.07113734632730484, 'train/grad_norm_pre_clip': 10.300508499145508, 'timing/data': 0.02605435997247696, 'timing/model': 1.4482079613953829, 'learning_rate/qwen_vl_interface': 4.485888354334933e-06, 'learning_rate/action_model': 4.4858883543349334e-05, 'epoch': 1.08}) +[2026-06-16 03:12:31,371][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.2141822874546051, 'train/grad_norm_pre_clip': 8.997146606445312, 'timing/data': 0.024981748312711716, 'timing/model': 1.4267845395952463, 'learning_rate/qwen_vl_interface': 4.482882752735017e-06, 'learning_rate/action_model': 4.482882752735017e-05, 'epoch': 1.08}) +[2026-06-16 03:12:37,175][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.0744093805551529, 'train/grad_norm_pre_clip': 5.298084735870361, 'timing/data': 0.026759784668684006, 'timing/model': 1.4303039833903313, 'learning_rate/qwen_vl_interface': 4.479877466468027e-06, 'learning_rate/action_model': 4.479877466468026e-05, 'epoch': 1.09}) +[2026-06-16 03:12:42,990][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.04614182934165001, 'train/grad_norm_pre_clip': 6.4875617027282715, 'timing/data': 0.025639094412326813, 'timing/model': 1.650867948308587, 'learning_rate/qwen_vl_interface': 4.476872496769318e-06, 'learning_rate/action_model': 4.476872496769318e-05, 'epoch': 1.09}) +[2026-06-16 03:12:48,677][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.14827369153499603, 'train/grad_norm_pre_clip': 9.207484245300293, 'timing/data': 0.026443256065249443, 'timing/model': 1.4380719345062971, 'learning_rate/qwen_vl_interface': 4.473867844874122e-06, 'learning_rate/action_model': 4.473867844874122e-05, 'epoch': 1.09}) +[2026-06-16 03:12:54,580][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.014891581609845161, 'train/grad_norm_pre_clip': 1.3240357637405396, 'timing/data': 0.026064282283186913, 'timing/model': 1.4284168686717749, 'learning_rate/qwen_vl_interface': 4.470863512017539e-06, 'learning_rate/action_model': 4.470863512017539e-05, 'epoch': 1.09}) +[2026-06-16 03:13:00,482][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.28396075963974, 'train/grad_norm_pre_clip': 9.665388107299805, 'timing/data': 0.025056499987840652, 'timing/model': 1.4393805880099535, 'learning_rate/qwen_vl_interface': 4.4678594994345344e-06, 'learning_rate/action_model': 4.467859499434534e-05, 'epoch': 1.09}) +[2026-06-16 03:13:06,382][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.18938696384429932, 'train/grad_norm_pre_clip': 10.38714599609375, 'timing/data': 0.025058485567569733, 'timing/model': 1.4380963500589132, 'learning_rate/qwen_vl_interface': 4.4648558083599475e-06, 'learning_rate/action_model': 4.464855808359947e-05, 'epoch': 1.09}) +[2026-06-16 03:13:12,336][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.06705842167139053, 'train/grad_norm_pre_clip': 7.389303684234619, 'timing/data': 0.0260626133531332, 'timing/model': 1.661275178194046, 'learning_rate/qwen_vl_interface': 4.461852440028479e-06, 'learning_rate/action_model': 4.461852440028479e-05, 'epoch': 1.09}) +[2026-06-16 03:13:18,115][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.08376243710517883, 'train/grad_norm_pre_clip': 10.679887771606445, 'timing/data': 0.025318685919046402, 'timing/model': 1.534751608967781, 'learning_rate/qwen_vl_interface': 4.458849395674702e-06, 'learning_rate/action_model': 4.458849395674701e-05, 'epoch': 1.09}) +[2026-06-16 03:13:24,077][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.054044488817453384, 'train/grad_norm_pre_clip': 5.951695442199707, 'timing/data': 0.025696992874145508, 'timing/model': 1.5031114090234041, 'learning_rate/qwen_vl_interface': 4.455846676533052e-06, 'learning_rate/action_model': 4.4558466765330525e-05, 'epoch': 1.09}) +[2026-06-16 03:13:29,927][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.12583395838737488, 'train/grad_norm_pre_clip': 12.75548267364502, 'timing/data': 0.02518647164106369, 'timing/model': 1.4539120476692915, 'learning_rate/qwen_vl_interface': 4.452844283837838e-06, 'learning_rate/action_model': 4.452844283837838e-05, 'epoch': 1.09}) +[2026-06-16 03:13:35,832][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.10916510969400406, 'train/grad_norm_pre_clip': 6.883100986480713, 'timing/data': 0.025830287486314774, 'timing/model': 1.455361358821392, 'learning_rate/qwen_vl_interface': 4.449842218823228e-06, 'learning_rate/action_model': 4.449842218823228e-05, 'epoch': 1.09}) +[2026-06-16 03:13:41,687][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.047997333109378815, 'train/grad_norm_pre_clip': 4.356739521026611, 'timing/data': 0.026008395478129387, 'timing/model': 1.4655910711735487, 'learning_rate/qwen_vl_interface': 4.446840482723256e-06, 'learning_rate/action_model': 4.446840482723255e-05, 'epoch': 1.09}) +[2026-06-16 03:13:47,591][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.013083728030323982, 'train/grad_norm_pre_clip': 1.4619563817977905, 'timing/data': 0.02594023197889328, 'timing/model': 1.6439554169774055, 'learning_rate/qwen_vl_interface': 4.443839076771823e-06, 'learning_rate/action_model': 4.443839076771823e-05, 'epoch': 1.09}) +[2026-06-16 03:13:53,487][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.07153439521789551, 'train/grad_norm_pre_clip': 3.808699131011963, 'timing/data': 0.025235697627067566, 'timing/model': 1.6395361237227917, 'learning_rate/qwen_vl_interface': 4.4408380022026945e-06, 'learning_rate/action_model': 4.440838002202694e-05, 'epoch': 1.09}) +[2026-06-16 03:13:59,190][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.012515898793935776, 'train/grad_norm_pre_clip': 1.2319691181182861, 'timing/data': 0.026162676513195038, 'timing/model': 1.4402557890862226, 'learning_rate/qwen_vl_interface': 4.437837260249499e-06, 'learning_rate/action_model': 4.437837260249499e-05, 'epoch': 1.09}) +[2026-06-16 03:14:05,517][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.022954825311899185, 'train/grad_norm_pre_clip': 4.73308801651001, 'timing/data': 0.02612970396876335, 'timing/model': 1.6270744819194078, 'learning_rate/qwen_vl_interface': 4.434836852145728e-06, 'learning_rate/action_model': 4.434836852145728e-05, 'epoch': 1.09}) +[2026-06-16 03:14:11,518][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.08673355728387833, 'train/grad_norm_pre_clip': 7.844159126281738, 'timing/data': 0.030183741822838783, 'timing/model': 1.530698360875249, 'learning_rate/qwen_vl_interface': 4.431836779124736e-06, 'learning_rate/action_model': 4.431836779124735e-05, 'epoch': 1.09}) +[2026-06-16 03:14:17,529][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.03273184597492218, 'train/grad_norm_pre_clip': 2.241774797439575, 'timing/data': 0.03158400394022465, 'timing/model': 1.4943419508635998, 'learning_rate/qwen_vl_interface': 4.428837042419738e-06, 'learning_rate/action_model': 4.428837042419738e-05, 'epoch': 1.09}) +[2026-06-16 03:14:23,534][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.05908938869833946, 'train/grad_norm_pre_clip': 4.146002292633057, 'timing/data': 0.029267728328704834, 'timing/model': 1.6775979679077864, 'learning_rate/qwen_vl_interface': 4.425837643263815e-06, 'learning_rate/action_model': 4.425837643263814e-05, 'epoch': 1.09}) +[2026-06-16 03:14:29,550][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.18959438800811768, 'train/grad_norm_pre_clip': 14.543404579162598, 'timing/data': 0.0333238560706377, 'timing/model': 1.5132781974971294, 'learning_rate/qwen_vl_interface': 4.4228385828899045e-06, 'learning_rate/action_model': 4.4228385828899046e-05, 'epoch': 1.09}) +[2026-06-16 03:14:35,757][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.1381216198205948, 'train/grad_norm_pre_clip': 7.009180068969727, 'timing/data': 0.031952135264873505, 'timing/model': 1.5303936582058668, 'learning_rate/qwen_vl_interface': 4.419839862530809e-06, 'learning_rate/action_model': 4.419839862530809e-05, 'epoch': 1.09}) +[2026-06-16 03:14:41,851][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.0935615673661232, 'train/grad_norm_pre_clip': 8.232526779174805, 'timing/data': 0.035163240507245064, 'timing/model': 1.515182988718152, 'learning_rate/qwen_vl_interface': 4.416841483419188e-06, 'learning_rate/action_model': 4.416841483419188e-05, 'epoch': 1.09}) +[2026-06-16 03:14:48,040][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.019054457545280457, 'train/grad_norm_pre_clip': 4.005983829498291, 'timing/data': 0.04128115065395832, 'timing/model': 1.5610474050045013, 'learning_rate/qwen_vl_interface': 4.413843446787561e-06, 'learning_rate/action_model': 4.413843446787561e-05, 'epoch': 1.09}) +[2026-06-16 03:14:54,250][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.012515771202743053, 'train/grad_norm_pre_clip': 2.4148964881896973, 'timing/data': 0.04091424308717251, 'timing/model': 1.5430913846939802, 'learning_rate/qwen_vl_interface': 4.41084575386831e-06, 'learning_rate/action_model': 4.4108457538683093e-05, 'epoch': 1.09}) +[2026-06-16 03:15:00,435][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.005528951529413462, 'train/grad_norm_pre_clip': 0.6204884648323059, 'timing/data': 0.03957943618297577, 'timing/model': 1.6784167122095823, 'learning_rate/qwen_vl_interface': 4.40784840589367e-06, 'learning_rate/action_model': 4.4078484058936706e-05, 'epoch': 1.09}) +[2026-06-16 03:15:06,183][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.24412265419960022, 'train/grad_norm_pre_clip': 13.247600555419922, 'timing/data': 0.025875616818666458, 'timing/model': 1.424522491171956, 'learning_rate/qwen_vl_interface': 4.404851404095742e-06, 'learning_rate/action_model': 4.404851404095742e-05, 'epoch': 1.09}) +[2026-06-16 03:15:12,156][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.16724561154842377, 'train/grad_norm_pre_clip': 7.294576168060303, 'timing/data': 0.03177485056221485, 'timing/model': 1.4836323913186789, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 1.1}) +[2026-06-16 03:15:18,337][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.16655556857585907, 'train/grad_norm_pre_clip': 8.124983787536621, 'timing/data': 0.03202771954238415, 'timing/model': 1.5254677049815655, 'learning_rate/qwen_vl_interface': 4.398858443957686e-06, 'learning_rate/action_model': 4.398858443957686e-05, 'epoch': 1.1}) +[2026-06-16 03:15:24,673][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.04922253638505936, 'train/grad_norm_pre_clip': 13.053426742553711, 'timing/data': 0.04089026153087616, 'timing/model': 1.5699708256870508, 'learning_rate/qwen_vl_interface': 4.395862488081038e-06, 'learning_rate/action_model': 4.395862488081038e-05, 'epoch': 1.1}) +[2026-06-16 03:15:30,854][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.15234419703483582, 'train/grad_norm_pre_clip': 10.536005973815918, 'timing/data': 0.03602417930960655, 'timing/model': 1.5758621226996183, 'learning_rate/qwen_vl_interface': 4.392866883308059e-06, 'learning_rate/action_model': 4.392866883308059e-05, 'epoch': 1.1}) +[2026-06-16 03:15:36,994][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.02112293802201748, 'train/grad_norm_pre_clip': 1.7667728662490845, 'timing/data': 0.034292347729206085, 'timing/model': 1.7603682074695826, 'learning_rate/qwen_vl_interface': 4.389871630870127e-06, 'learning_rate/action_model': 4.389871630870127e-05, 'epoch': 1.1}) +[2026-06-16 03:15:43,060][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.1763831079006195, 'train/grad_norm_pre_clip': 20.557382583618164, 'timing/data': 0.035625746473670006, 'timing/model': 1.5411702655255795, 'learning_rate/qwen_vl_interface': 4.386876731998478e-06, 'learning_rate/action_model': 4.386876731998478e-05, 'epoch': 1.1}) +[2026-06-16 03:15:49,325][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.0031770244240760803, 'train/grad_norm_pre_clip': 0.5531242489814758, 'timing/data': 0.027510035783052444, 'timing/model': 1.533430952578783, 'learning_rate/qwen_vl_interface': 4.3838821879242006e-06, 'learning_rate/action_model': 4.3838821879242e-05, 'epoch': 1.1}) +[2026-06-16 03:15:55,377][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.01621433161199093, 'train/grad_norm_pre_clip': 1.5932443141937256, 'timing/data': 0.0315288919955492, 'timing/model': 1.522384999319911, 'learning_rate/qwen_vl_interface': 4.38088799987824e-06, 'learning_rate/action_model': 4.38088799987824e-05, 'epoch': 1.1}) +[2026-06-16 03:16:01,547][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.05871584638953209, 'train/grad_norm_pre_clip': 9.088273048400879, 'timing/data': 0.033630892634391785, 'timing/model': 1.5269320514053106, 'learning_rate/qwen_vl_interface': 4.3778941690913935e-06, 'learning_rate/action_model': 4.3778941690913934e-05, 'epoch': 1.1}) +[2026-06-16 03:16:07,746][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.15867500007152557, 'train/grad_norm_pre_clip': 10.820399284362793, 'timing/data': 0.03344152122735977, 'timing/model': 1.5305881053209305, 'learning_rate/qwen_vl_interface': 4.374900696794311e-06, 'learning_rate/action_model': 4.374900696794311e-05, 'epoch': 1.1}) +[2026-06-16 03:16:13,969][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.08478198200464249, 'train/grad_norm_pre_clip': 6.133737564086914, 'timing/data': 0.0321168452501297, 'timing/model': 1.7585061248391867, 'learning_rate/qwen_vl_interface': 4.371907584217499e-06, 'learning_rate/action_model': 4.371907584217498e-05, 'epoch': 1.1}) +[2026-06-16 03:16:19,945][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.28497496247291565, 'train/grad_norm_pre_clip': 12.965681076049805, 'timing/data': 0.03081376478075981, 'timing/model': 1.5355991143733263, 'learning_rate/qwen_vl_interface': 4.368914832591309e-06, 'learning_rate/action_model': 4.368914832591309e-05, 'epoch': 1.1}) +[2026-06-16 03:16:26,141][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.04741264507174492, 'train/grad_norm_pre_clip': 7.111693859100342, 'timing/data': 0.03275864012539387, 'timing/model': 1.5317102335393429, 'learning_rate/qwen_vl_interface': 4.365922443145949e-06, 'learning_rate/action_model': 4.365922443145949e-05, 'epoch': 1.1}) +[2026-06-16 03:16:32,347][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.10107635706663132, 'train/grad_norm_pre_clip': 6.342442989349365, 'timing/data': 0.03205917775630951, 'timing/model': 1.5361663736402988, 'learning_rate/qwen_vl_interface': 4.362930417111479e-06, 'learning_rate/action_model': 4.362930417111479e-05, 'epoch': 1.1}) +[2026-06-16 03:16:38,440][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.1371961534023285, 'train/grad_norm_pre_clip': 21.615310668945312, 'timing/data': 0.03198188357055187, 'timing/model': 1.485839620232582, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 1.1}) +[2026-06-16 03:16:44,286][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.011048845015466213, 'train/grad_norm_pre_clip': 4.10957145690918, 'timing/data': 0.026388019323349, 'timing/model': 1.4590662810951471, 'learning_rate/qwen_vl_interface': 4.3569474601946934e-06, 'learning_rate/action_model': 4.356947460194694e-05, 'epoch': 1.1}) +[2026-06-16 03:16:50,185][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.09796641021966934, 'train/grad_norm_pre_clip': 9.463212013244629, 'timing/data': 0.026576479896903038, 'timing/model': 1.6365288849920034, 'learning_rate/qwen_vl_interface': 4.3539565317717446e-06, 'learning_rate/action_model': 4.353956531771745e-05, 'epoch': 1.1}) +[2026-06-16 03:16:55,889][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.12332911044359207, 'train/grad_norm_pre_clip': 12.579850196838379, 'timing/data': 0.026322124525904655, 'timing/model': 1.4371178690344095, 'learning_rate/qwen_vl_interface': 4.350965971678419e-06, 'learning_rate/action_model': 4.3509659716784185e-05, 'epoch': 1.1}) +[2026-06-16 03:17:01,794][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.003560778219252825, 'train/grad_norm_pre_clip': 0.6632110476493835, 'timing/data': 0.02661096304655075, 'timing/model': 1.427582211792469, 'learning_rate/qwen_vl_interface': 4.3479757811440225e-06, 'learning_rate/action_model': 4.347975781144022e-05, 'epoch': 1.1}) +[2026-06-16 03:17:07,838][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.0052933283150196075, 'train/grad_norm_pre_clip': 0.8275511860847473, 'timing/data': 0.03167624585330486, 'timing/model': 1.5484074857085943, 'learning_rate/qwen_vl_interface': 4.34498596139771e-06, 'learning_rate/action_model': 4.34498596139771e-05, 'epoch': 1.1}) +[2026-06-16 03:17:13,937][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.005913798697292805, 'train/grad_norm_pre_clip': 1.0014317035675049, 'timing/data': 0.032778024673461914, 'timing/model': 1.5008818618953228, 'learning_rate/qwen_vl_interface': 4.341996513668483e-06, 'learning_rate/action_model': 4.3419965136684824e-05, 'epoch': 1.1}) +[2026-06-16 03:17:20,146][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.034627147018909454, 'train/grad_norm_pre_clip': 10.262040138244629, 'timing/data': 0.031187277287244797, 'timing/model': 1.4791019782423973, 'learning_rate/qwen_vl_interface': 4.33900743918519e-06, 'learning_rate/action_model': 4.33900743918519e-05, 'epoch': 1.1}) +[2026-06-16 03:17:26,459][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.021928012371063232, 'train/grad_norm_pre_clip': 3.5395689010620117, 'timing/data': 0.03122207336127758, 'timing/model': 1.7450374756008387, 'learning_rate/qwen_vl_interface': 4.336018739176529e-06, 'learning_rate/action_model': 4.33601873917653e-05, 'epoch': 1.1}) +[2026-06-16 03:17:32,446][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.10820510238409042, 'train/grad_norm_pre_clip': 11.335915565490723, 'timing/data': 0.03083995170891285, 'timing/model': 1.531111216172576, 'learning_rate/qwen_vl_interface': 4.333030414871041e-06, 'learning_rate/action_model': 4.33303041487104e-05, 'epoch': 1.1}) +[2026-06-16 03:17:38,660][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.12153080105781555, 'train/grad_norm_pre_clip': 18.088836669921875, 'timing/data': 0.03309977613389492, 'timing/model': 1.5424163602292538, 'learning_rate/qwen_vl_interface': 4.330042467497111e-06, 'learning_rate/action_model': 4.330042467497111e-05, 'epoch': 1.1}) +[2026-06-16 03:17:44,853][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.10760831832885742, 'train/grad_norm_pre_clip': 23.470558166503906, 'timing/data': 0.03143136948347092, 'timing/model': 1.5394940096884966, 'learning_rate/qwen_vl_interface': 4.327054898282973e-06, 'learning_rate/action_model': 4.3270548982829727e-05, 'epoch': 1.1}) +[2026-06-16 03:17:50,948][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.05155973881483078, 'train/grad_norm_pre_clip': 8.367462158203125, 'timing/data': 0.032646410167217255, 'timing/model': 1.523057509213686, 'learning_rate/qwen_vl_interface': 4.324067708456703e-06, 'learning_rate/action_model': 4.3240677084567036e-05, 'epoch': 1.11}) +[2026-06-16 03:17:57,240][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.03240213170647621, 'train/grad_norm_pre_clip': 2.7711613178253174, 'timing/data': 0.0326873455196619, 'timing/model': 1.5647488087415695, 'learning_rate/qwen_vl_interface': 4.321080899246225e-06, 'learning_rate/action_model': 4.321080899246225e-05, 'epoch': 1.11}) +[2026-06-16 03:18:03,393][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.24374271929264069, 'train/grad_norm_pre_clip': 9.97800064086914, 'timing/data': 0.03158790245652199, 'timing/model': 1.7671931944787502, 'learning_rate/qwen_vl_interface': 4.318094471879299e-06, 'learning_rate/action_model': 4.3180944718792985e-05, 'epoch': 1.11}) +[2026-06-16 03:18:09,748][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.15709495544433594, 'train/grad_norm_pre_clip': 10.2525634765625, 'timing/data': 0.03184380941092968, 'timing/model': 1.5997702647000551, 'learning_rate/qwen_vl_interface': 4.315108427583535e-06, 'learning_rate/action_model': 4.315108427583535e-05, 'epoch': 1.11}) +[2026-06-16 03:18:15,959][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.15161916613578796, 'train/grad_norm_pre_clip': 13.366658210754395, 'timing/data': 0.03260795772075653, 'timing/model': 1.5502357687801123, 'learning_rate/qwen_vl_interface': 4.312122767586381e-06, 'learning_rate/action_model': 4.3121227675863815e-05, 'epoch': 1.11}) +[2026-06-16 03:18:22,051][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.12883096933364868, 'train/grad_norm_pre_clip': 23.767574310302734, 'timing/data': 0.030996622517704964, 'timing/model': 1.5382207911461592, 'learning_rate/qwen_vl_interface': 4.3091374931151324e-06, 'learning_rate/action_model': 4.309137493115132e-05, 'epoch': 1.11}) +[2026-06-16 03:18:28,163][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.15599122643470764, 'train/grad_norm_pre_clip': 28.799579620361328, 'timing/data': 0.03230576030910015, 'timing/model': 1.5538168046623468, 'learning_rate/qwen_vl_interface': 4.306152605396921e-06, 'learning_rate/action_model': 4.306152605396921e-05, 'epoch': 1.11}) +[2026-06-16 03:18:34,359][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.044373948127031326, 'train/grad_norm_pre_clip': 17.997478485107422, 'timing/data': 0.03127626143395901, 'timing/model': 1.5110557600855827, 'learning_rate/qwen_vl_interface': 4.30316810565872e-06, 'learning_rate/action_model': 4.3031681056587206e-05, 'epoch': 1.11}) +[2026-06-16 03:18:40,611][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.024429436773061752, 'train/grad_norm_pre_clip': 3.2313642501831055, 'timing/data': 0.03626318834722042, 'timing/model': 1.794037640094757, 'learning_rate/qwen_vl_interface': 4.300183995127347e-06, 'learning_rate/action_model': 4.300183995127346e-05, 'epoch': 1.11}) +[2026-06-16 03:18:46,662][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.16473229229450226, 'train/grad_norm_pre_clip': 11.794958114624023, 'timing/data': 0.031418826431035995, 'timing/model': 1.5586955212056637, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 1.11}) +[2026-06-16 03:18:52,845][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.017160305753350258, 'train/grad_norm_pre_clip': 3.4288792610168457, 'timing/data': 0.03034113347530365, 'timing/model': 1.5481196101754904, 'learning_rate/qwen_vl_interface': 4.29421694659154e-06, 'learning_rate/action_model': 4.294216946591539e-05, 'epoch': 1.11}) +[2026-06-16 03:18:58,944][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.2988879382610321, 'train/grad_norm_pre_clip': 19.84984588623047, 'timing/data': 0.031998688355088234, 'timing/model': 1.5416603293269873, 'learning_rate/qwen_vl_interface': 4.2912340110399345e-06, 'learning_rate/action_model': 4.291234011039934e-05, 'epoch': 1.11}) +[2026-06-16 03:19:05,143][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.007922505028545856, 'train/grad_norm_pre_clip': 1.6336113214492798, 'timing/data': 0.03343724086880684, 'timing/model': 1.5349037609994411, 'learning_rate/qwen_vl_interface': 4.2882514696008125e-06, 'learning_rate/action_model': 4.2882514696008125e-05, 'epoch': 1.11}) +[2026-06-16 03:19:11,242][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.031080212444067, 'train/grad_norm_pre_clip': 4.577581405639648, 'timing/data': 0.031991953030228615, 'timing/model': 1.490143919363618, 'learning_rate/qwen_vl_interface': 4.2852693235001815e-06, 'learning_rate/action_model': 4.2852693235001815e-05, 'epoch': 1.11}) +[2026-06-16 03:19:17,487][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.05648105591535568, 'train/grad_norm_pre_clip': 6.759799957275391, 'timing/data': 0.03298506326973438, 'timing/model': 1.7786218039691448, 'learning_rate/qwen_vl_interface': 4.2822875739638915e-06, 'learning_rate/action_model': 4.282287573963891e-05, 'epoch': 1.11}) +[2026-06-16 03:19:23,610][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.10688169300556183, 'train/grad_norm_pre_clip': 6.895208358764648, 'timing/data': 0.03227893449366093, 'timing/model': 1.674636559560895, 'learning_rate/qwen_vl_interface': 4.279306222217625e-06, 'learning_rate/action_model': 4.279306222217626e-05, 'epoch': 1.11}) +[2026-06-16 03:19:29,618][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.3890334367752075, 'train/grad_norm_pre_clip': 28.78926658630371, 'timing/data': 0.03144095651805401, 'timing/model': 1.5470975749194622, 'learning_rate/qwen_vl_interface': 4.276325269486906e-06, 'learning_rate/action_model': 4.276325269486906e-05, 'epoch': 1.11}) +[2026-06-16 03:19:35,752][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.3228847086429596, 'train/grad_norm_pre_clip': 12.130865097045898, 'timing/data': 0.03284268267452717, 'timing/model': 1.4868159554898739, 'learning_rate/qwen_vl_interface': 4.273344716997089e-06, 'learning_rate/action_model': 4.273344716997089e-05, 'epoch': 1.11}) +[2026-06-16 03:19:41,951][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.06480921804904938, 'train/grad_norm_pre_clip': 9.196429252624512, 'timing/data': 0.03165494278073311, 'timing/model': 1.5359653886407614, 'learning_rate/qwen_vl_interface': 4.270364565973368e-06, 'learning_rate/action_model': 4.270364565973367e-05, 'epoch': 1.11}) +[2026-06-16 03:19:47,789][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.0009856510441750288, 'train/grad_norm_pre_clip': 0.10960933566093445, 'timing/data': 0.025869805365800858, 'timing/model': 1.4474432319402695, 'learning_rate/qwen_vl_interface': 4.267384817640768e-06, 'learning_rate/action_model': 4.2673848176407685e-05, 'epoch': 1.11}) +[2026-06-16 03:19:53,712][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.1568097323179245, 'train/grad_norm_pre_clip': 12.461875915527344, 'timing/data': 0.025439497083425522, 'timing/model': 1.6674148235470057, 'learning_rate/qwen_vl_interface': 4.264405473224155e-06, 'learning_rate/action_model': 4.2644054732241546e-05, 'epoch': 1.11}) +[2026-06-16 03:19:59,614][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.11518062651157379, 'train/grad_norm_pre_clip': 46.6943359375, 'timing/data': 0.027056032791733742, 'timing/model': 1.6658266801387072, 'learning_rate/qwen_vl_interface': 4.261426533948221e-06, 'learning_rate/action_model': 4.2614265339482213e-05, 'epoch': 1.11}) +[2026-06-16 03:20:05,786][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.08421799540519714, 'train/grad_norm_pre_clip': 6.970989227294922, 'timing/data': 0.03479268029332161, 'timing/model': 1.5491337459534407, 'learning_rate/qwen_vl_interface': 4.258448001037499e-06, 'learning_rate/action_model': 4.2584480010374985e-05, 'epoch': 1.11}) +[2026-06-16 03:20:12,331][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.023396579548716545, 'train/grad_norm_pre_clip': 4.106247901916504, 'timing/data': 0.0486082024872303, 'timing/model': 1.7106733694672585, 'learning_rate/qwen_vl_interface': 4.255469875716348e-06, 'learning_rate/action_model': 4.255469875716348e-05, 'epoch': 1.11}) +[2026-06-16 03:20:19,209][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.13872569799423218, 'train/grad_norm_pre_clip': 11.05309772491455, 'timing/data': 0.04649733379483223, 'timing/model': 1.7020039521157742, 'learning_rate/qwen_vl_interface': 4.252492159208966e-06, 'learning_rate/action_model': 4.252492159208966e-05, 'epoch': 1.11}) +[2026-06-16 03:20:26,235][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.04717395827174187, 'train/grad_norm_pre_clip': 11.802230834960938, 'timing/data': 0.04812180623412132, 'timing/model': 1.7156593818217516, 'learning_rate/qwen_vl_interface': 4.2495148527393766e-06, 'learning_rate/action_model': 4.2495148527393766e-05, 'epoch': 1.11}) +[2026-06-16 03:20:32,605][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.2497003674507141, 'train/grad_norm_pre_clip': 9.951112747192383, 'timing/data': 0.03265640139579773, 'timing/model': 1.797788094729185, 'learning_rate/qwen_vl_interface': 4.246537957531441e-06, 'learning_rate/action_model': 4.2465379575314404e-05, 'epoch': 1.12}) +[2026-06-16 03:20:38,649][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.04315902665257454, 'train/grad_norm_pre_clip': 6.69486141204834, 'timing/data': 0.032480400055646896, 'timing/model': 1.5360967479646206, 'learning_rate/qwen_vl_interface': 4.243561474808846e-06, 'learning_rate/action_model': 4.2435614748088456e-05, 'epoch': 1.12}) +[2026-06-16 03:20:44,972][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.30297204852104187, 'train/grad_norm_pre_clip': 14.595404624938965, 'timing/data': 0.038974080234766006, 'timing/model': 1.6525003872811794, 'learning_rate/qwen_vl_interface': 4.240585405795112e-06, 'learning_rate/action_model': 4.240585405795112e-05, 'epoch': 1.12}) +[2026-06-16 03:20:51,148][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.01838233508169651, 'train/grad_norm_pre_clip': 1.9704744815826416, 'timing/data': 0.032105810940265656, 'timing/model': 1.538807863369584, 'learning_rate/qwen_vl_interface': 4.237609751713589e-06, 'learning_rate/action_model': 4.2376097517135896e-05, 'epoch': 1.12}) +[2026-06-16 03:20:57,342][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.05648060142993927, 'train/grad_norm_pre_clip': 9.248130798339844, 'timing/data': 0.03235063888132572, 'timing/model': 1.4965061135590076, 'learning_rate/qwen_vl_interface': 4.234634513787455e-06, 'learning_rate/action_model': 4.2346345137874556e-05, 'epoch': 1.12}) +[2026-06-16 03:21:03,445][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.11743053793907166, 'train/grad_norm_pre_clip': 4.556338787078857, 'timing/data': 0.03126608766615391, 'timing/model': 1.5043573528528214, 'learning_rate/qwen_vl_interface': 4.2316596932397195e-06, 'learning_rate/action_model': 4.2316596932397195e-05, 'epoch': 1.12}) +[2026-06-16 03:21:09,628][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.04057147353887558, 'train/grad_norm_pre_clip': 4.369653701782227, 'timing/data': 0.030803987756371498, 'timing/model': 1.747858701273799, 'learning_rate/qwen_vl_interface': 4.22868529129322e-06, 'learning_rate/action_model': 4.22868529129322e-05, 'epoch': 1.12}) +[2026-06-16 03:21:15,646][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.09290274977684021, 'train/grad_norm_pre_clip': 8.901859283447266, 'timing/data': 0.03179865702986717, 'timing/model': 1.558005953207612, 'learning_rate/qwen_vl_interface': 4.2257113091706165e-06, 'learning_rate/action_model': 4.2257113091706165e-05, 'epoch': 1.12}) +[2026-06-16 03:21:21,864][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.08618917316198349, 'train/grad_norm_pre_clip': 15.852339744567871, 'timing/data': 0.03147738054394722, 'timing/model': 1.5388528686016798, 'learning_rate/qwen_vl_interface': 4.222737748094404e-06, 'learning_rate/action_model': 4.222737748094404e-05, 'epoch': 1.12}) +[2026-06-16 03:21:28,065][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.1390487253665924, 'train/grad_norm_pre_clip': 11.076297760009766, 'timing/data': 0.031944284215569496, 'timing/model': 1.5353788379579782, 'learning_rate/qwen_vl_interface': 4.219764609286901e-06, 'learning_rate/action_model': 4.219764609286901e-05, 'epoch': 1.12}) +[2026-06-16 03:21:34,245][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.03370547667145729, 'train/grad_norm_pre_clip': 5.732207775115967, 'timing/data': 0.03319905139505863, 'timing/model': 1.5365778654813766, 'learning_rate/qwen_vl_interface': 4.21679189397025e-06, 'learning_rate/action_model': 4.21679189397025e-05, 'epoch': 1.12}) +[2026-06-16 03:21:40,089][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.2133038491010666, 'train/grad_norm_pre_clip': 12.128528594970703, 'timing/data': 0.024770870804786682, 'timing/model': 1.4388497788459063, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 1.12}) +[2026-06-16 03:21:45,950][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.22375719249248505, 'train/grad_norm_pre_clip': 9.882952690124512, 'timing/data': 0.026046765968203545, 'timing/model': 1.5967369694262743, 'learning_rate/qwen_vl_interface': 4.210847738697219e-06, 'learning_rate/action_model': 4.210847738697218e-05, 'epoch': 1.12}) +[2026-06-16 03:21:51,591][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.04048510640859604, 'train/grad_norm_pre_clip': 7.299717426300049, 'timing/data': 0.02517133206129074, 'timing/model': 1.4408099446445704, 'learning_rate/qwen_vl_interface': 4.207876301184256e-06, 'learning_rate/action_model': 4.207876301184256e-05, 'epoch': 1.12}) +[2026-06-16 03:21:57,494][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.04609990864992142, 'train/grad_norm_pre_clip': 5.448387145996094, 'timing/data': 0.025724759325385094, 'timing/model': 1.4330152422189713, 'learning_rate/qwen_vl_interface': 4.2049052920489825e-06, 'learning_rate/action_model': 4.2049052920489825e-05, 'epoch': 1.12}) +[2026-06-16 03:22:03,389][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.2704053819179535, 'train/grad_norm_pre_clip': 10.439047813415527, 'timing/data': 0.025885626673698425, 'timing/model': 1.436255481094122, 'learning_rate/qwen_vl_interface': 4.201934712512666e-06, 'learning_rate/action_model': 4.2019347125126656e-05, 'epoch': 1.12}) +[2026-06-16 03:22:09,290][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.19413277506828308, 'train/grad_norm_pre_clip': 12.713436126708984, 'timing/data': 0.025122342631220818, 'timing/model': 1.4294482320547104, 'learning_rate/qwen_vl_interface': 4.1989645637964e-06, 'learning_rate/action_model': 4.1989645637964e-05, 'epoch': 1.12}) +[2026-06-16 03:22:15,196][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.028190013021230698, 'train/grad_norm_pre_clip': 2.199336051940918, 'timing/data': 0.026101035997271538, 'timing/model': 1.6447939220815897, 'learning_rate/qwen_vl_interface': 4.195994847121101e-06, 'learning_rate/action_model': 4.195994847121101e-05, 'epoch': 1.12}) +[2026-06-16 03:29:48,437][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.06043345853686333, 'train/grad_norm_pre_clip': 5.036750316619873, 'eval/loss': 0.045467205345630646, 'eval/flap_precision': 0.9354413702239789, 'eval/flap_recall': 0.9633649932157394, 'eval/flap_f1': 0.9491978609625668, 'eval/flap_rate_model': 0.21083333333333334, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9491978609625668, 'timing/data': 0.02607564814388752, 'timing/model': 1.4351787716150284, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 1.12}) +[2026-06-16 03:29:48,438][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2900_state +[2026-06-16 03:29:48,442][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 03:29:54,291][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2900_state/model.safetensors +[2026-06-16 03:30:04,489][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2900_state/optimizer.bin +[2026-06-16 03:30:04,489][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2900_state/sampler.bin +[2026-06-16 03:30:04,489][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2900_state/sampler_1.bin +[2026-06-16 03:30:04,490][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_2900_state/random_states_0.pkl +[2026-06-16 03:30:04,579][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2900 +[2026-06-16 03:30:04,579][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 03:30:04,588][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 03:30:10,659][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.17115405201911926, 'train/grad_norm_pre_clip': 8.839723587036133, 'timing/data': 0.02918008528649807, 'timing/model': 1.7712297551333904, 'learning_rate/qwen_vl_interface': 4.19005671477618e-06, 'learning_rate/action_model': 4.1900567147761794e-05, 'epoch': 1.12}) +[2026-06-16 03:30:16,336][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.028647569939494133, 'train/grad_norm_pre_clip': 3.4797258377075195, 'timing/data': 0.025501832365989685, 'timing/model': 1.4667949862778187, 'learning_rate/qwen_vl_interface': 4.1870883015474975e-06, 'learning_rate/action_model': 4.187088301547497e-05, 'epoch': 1.12}) +[2026-06-16 03:30:22,780][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.03627561032772064, 'train/grad_norm_pre_clip': 3.7796990871429443, 'timing/data': 0.03652677498757839, 'timing/model': 1.6095244903117418, 'learning_rate/qwen_vl_interface': 4.1841203252416655e-06, 'learning_rate/action_model': 4.1841203252416655e-05, 'epoch': 1.12}) +[2026-06-16 03:30:29,562][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.012098913080990314, 'train/grad_norm_pre_clip': 1.9273796081542969, 'timing/data': 0.04356682486832142, 'timing/model': 1.6824620384722948, 'learning_rate/qwen_vl_interface': 4.181152787078705e-06, 'learning_rate/action_model': 4.1811527870787056e-05, 'epoch': 1.12}) +[2026-06-16 03:30:36,178][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.005342929624021053, 'train/grad_norm_pre_clip': 0.6438032984733582, 'timing/data': 0.03263104893267155, 'timing/model': 1.5425166860222816, 'learning_rate/qwen_vl_interface': 4.178185688278462e-06, 'learning_rate/action_model': 4.1781856882784616e-05, 'epoch': 1.12}) +[2026-06-16 03:30:42,486][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.1896612048149109, 'train/grad_norm_pre_clip': 9.285316467285156, 'timing/data': 0.03037126548588276, 'timing/model': 1.5760314799845219, 'learning_rate/qwen_vl_interface': 4.175219030060595e-06, 'learning_rate/action_model': 4.175219030060595e-05, 'epoch': 1.12}) +[2026-06-16 03:30:48,802][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.27609527111053467, 'train/grad_norm_pre_clip': 10.231575965881348, 'timing/data': 0.031542832031846046, 'timing/model': 1.8174541052430868, 'learning_rate/qwen_vl_interface': 4.1722528136445875e-06, 'learning_rate/action_model': 4.172252813644588e-05, 'epoch': 1.13}) +[2026-06-16 03:30:54,740][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.08320801705121994, 'train/grad_norm_pre_clip': 3.614114284515381, 'timing/data': 0.03239890746772289, 'timing/model': 1.5417944341897964, 'learning_rate/qwen_vl_interface': 4.169287040249738e-06, 'learning_rate/action_model': 4.169287040249738e-05, 'epoch': 1.13}) +[2026-06-16 03:31:01,041][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.039216361939907074, 'train/grad_norm_pre_clip': 2.7628884315490723, 'timing/data': 0.03129596821963787, 'timing/model': 1.5572826657444239, 'learning_rate/qwen_vl_interface': 4.166321711095164e-06, 'learning_rate/action_model': 4.166321711095164e-05, 'epoch': 1.13}) +[2026-06-16 03:31:07,142][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.0808994248509407, 'train/grad_norm_pre_clip': 5.266337871551514, 'timing/data': 0.032881928607821465, 'timing/model': 1.5430835150182247, 'learning_rate/qwen_vl_interface': 4.1633568273998e-06, 'learning_rate/action_model': 4.1633568273998e-05, 'epoch': 1.13}) +[2026-06-16 03:31:13,234][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.05267300829291344, 'train/grad_norm_pre_clip': 3.5324552059173584, 'timing/data': 0.030905747786164284, 'timing/model': 1.5309825129806995, 'learning_rate/qwen_vl_interface': 4.1603923903823995e-06, 'learning_rate/action_model': 4.160392390382399e-05, 'epoch': 1.13}) +[2026-06-16 03:31:19,131][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.03745787590742111, 'train/grad_norm_pre_clip': 4.577226638793945, 'timing/data': 0.030510224401950836, 'timing/model': 1.475320352241397, 'learning_rate/qwen_vl_interface': 4.1574284012615276e-06, 'learning_rate/action_model': 4.157428401261528e-05, 'epoch': 1.13}) +[2026-06-16 03:31:25,501][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.04312265291810036, 'train/grad_norm_pre_clip': 6.727691650390625, 'timing/data': 0.0362964179366827, 'timing/model': 1.8976755235344172, 'learning_rate/qwen_vl_interface': 4.154464861255571e-06, 'learning_rate/action_model': 4.154464861255571e-05, 'epoch': 1.13}) +[2026-06-16 03:31:31,437][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.1920233517885208, 'train/grad_norm_pre_clip': 9.217976570129395, 'timing/data': 0.031275125220417976, 'timing/model': 1.5306404028087854, 'learning_rate/qwen_vl_interface': 4.151501771582728e-06, 'learning_rate/action_model': 4.151501771582728e-05, 'epoch': 1.13}) +[2026-06-16 03:31:37,542][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.044763654470443726, 'train/grad_norm_pre_clip': 10.434535026550293, 'timing/data': 0.032210174947977066, 'timing/model': 1.5168342590332031, 'learning_rate/qwen_vl_interface': 4.148539133461012e-06, 'learning_rate/action_model': 4.1485391334610116e-05, 'epoch': 1.13}) +[2026-06-16 03:31:43,660][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.14977504312992096, 'train/grad_norm_pre_clip': 11.390466690063477, 'timing/data': 0.03166952729225159, 'timing/model': 1.526248099282384, 'learning_rate/qwen_vl_interface': 4.145576948108254e-06, 'learning_rate/action_model': 4.145576948108253e-05, 'epoch': 1.13}) +[2026-06-16 03:31:49,650][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.09396897256374359, 'train/grad_norm_pre_clip': 11.812657356262207, 'timing/data': 0.03143595531582832, 'timing/model': 1.512597132474184, 'learning_rate/qwen_vl_interface': 4.142615216742093e-06, 'learning_rate/action_model': 4.142615216742093e-05, 'epoch': 1.13}) +[2026-06-16 03:31:55,768][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.09678459912538528, 'train/grad_norm_pre_clip': 9.136113166809082, 'timing/data': 0.033172059804201126, 'timing/model': 1.4991839919239283, 'learning_rate/qwen_vl_interface': 4.139653940579988e-06, 'learning_rate/action_model': 4.1396539405799874e-05, 'epoch': 1.13}) +[2026-06-16 03:32:01,990][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.06504809111356735, 'train/grad_norm_pre_clip': 7.427217483520508, 'timing/data': 0.031640417873859406, 'timing/model': 1.7810041792690754, 'learning_rate/qwen_vl_interface': 4.136693120839206e-06, 'learning_rate/action_model': 4.136693120839205e-05, 'epoch': 1.13}) +[2026-06-16 03:32:07,946][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.15571409463882446, 'train/grad_norm_pre_clip': 9.417527198791504, 'timing/data': 0.03194956108927727, 'timing/model': 1.5180269330739975, 'learning_rate/qwen_vl_interface': 4.133732758736828e-06, 'learning_rate/action_model': 4.133732758736828e-05, 'epoch': 1.13}) +[2026-06-16 03:32:14,144][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.14841392636299133, 'train/grad_norm_pre_clip': 9.394484519958496, 'timing/data': 0.033298516646027565, 'timing/model': 1.5347483269870281, 'learning_rate/qwen_vl_interface': 4.13077285548975e-06, 'learning_rate/action_model': 4.13077285548975e-05, 'epoch': 1.13}) +[2026-06-16 03:32:20,342][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.03941578418016434, 'train/grad_norm_pre_clip': 3.3227474689483643, 'timing/data': 0.03221728838980198, 'timing/model': 1.5299724955111742, 'learning_rate/qwen_vl_interface': 4.127813412314672e-06, 'learning_rate/action_model': 4.1278134123146725e-05, 'epoch': 1.13}) +[2026-06-16 03:32:26,446][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.029729414731264114, 'train/grad_norm_pre_clip': 2.2535464763641357, 'timing/data': 0.031138040125370026, 'timing/model': 1.527050320059061, 'learning_rate/qwen_vl_interface': 4.124854430428113e-06, 'learning_rate/action_model': 4.1248544304281126e-05, 'epoch': 1.13}) +[2026-06-16 03:32:32,648][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.04781132563948631, 'train/grad_norm_pre_clip': 3.6291403770446777, 'timing/data': 0.033635104075074196, 'timing/model': 1.4871727880090475, 'learning_rate/qwen_vl_interface': 4.121895911046397e-06, 'learning_rate/action_model': 4.121895911046396e-05, 'epoch': 1.13}) +[2026-06-16 03:32:38,899][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.07310202717781067, 'train/grad_norm_pre_clip': 6.703815937042236, 'timing/data': 0.0326794795691967, 'timing/model': 1.7903411649167538, 'learning_rate/qwen_vl_interface': 4.118937855385659e-06, 'learning_rate/action_model': 4.1189378553856584e-05, 'epoch': 1.13}) +[2026-06-16 03:32:44,950][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.02513543888926506, 'train/grad_norm_pre_clip': 1.405693769454956, 'timing/data': 0.030700551345944405, 'timing/model': 1.5404399763792753, 'learning_rate/qwen_vl_interface': 4.115980264661844e-06, 'learning_rate/action_model': 4.115980264661844e-05, 'epoch': 1.13}) +[2026-06-16 03:32:51,167][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.04064966365695, 'train/grad_norm_pre_clip': 4.1497297286987305, 'timing/data': 0.03127871826291084, 'timing/model': 1.5588487330824137, 'learning_rate/qwen_vl_interface': 4.113023140090709e-06, 'learning_rate/action_model': 4.113023140090709e-05, 'epoch': 1.13}) +[2026-06-16 03:32:57,364][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.0422147698700428, 'train/grad_norm_pre_clip': 5.949779987335205, 'timing/data': 0.0318036712706089, 'timing/model': 1.5620767027139664, 'learning_rate/qwen_vl_interface': 4.110066482887812e-06, 'learning_rate/action_model': 4.110066482887812e-05, 'epoch': 1.13}) +[2026-06-16 03:33:03,459][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.06162644922733307, 'train/grad_norm_pre_clip': 6.883623123168945, 'timing/data': 0.03114817850291729, 'timing/model': 1.5154981892555952, 'learning_rate/qwen_vl_interface': 4.107110294268525e-06, 'learning_rate/action_model': 4.107110294268525e-05, 'epoch': 1.13}) +[2026-06-16 03:33:09,558][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.059538863599300385, 'train/grad_norm_pre_clip': 5.622194290161133, 'timing/data': 0.031837619841098785, 'timing/model': 1.5290197376161814, 'learning_rate/qwen_vl_interface': 4.104154575448026e-06, 'learning_rate/action_model': 4.104154575448025e-05, 'epoch': 1.13}) +[2026-06-16 03:33:15,724][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.025928761810064316, 'train/grad_norm_pre_clip': 2.6778340339660645, 'timing/data': 0.03230350278317928, 'timing/model': 1.8068262953311205, 'learning_rate/qwen_vl_interface': 4.101199327641297e-06, 'learning_rate/action_model': 4.101199327641297e-05, 'epoch': 1.13}) +[2026-06-16 03:33:21,746][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.066141776740551, 'train/grad_norm_pre_clip': 9.165782928466797, 'timing/data': 0.03174618259072304, 'timing/model': 1.5510617960244417, 'learning_rate/qwen_vl_interface': 4.098244552063133e-06, 'learning_rate/action_model': 4.098244552063132e-05, 'epoch': 1.13}) +[2026-06-16 03:33:27,945][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.07287817448377609, 'train/grad_norm_pre_clip': 7.024491310119629, 'timing/data': 0.03195665776729584, 'timing/model': 1.5341815743595362, 'learning_rate/qwen_vl_interface': 4.0952902499281254e-06, 'learning_rate/action_model': 4.095290249928125e-05, 'epoch': 1.14}) +[2026-06-16 03:33:34,143][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.18908549845218658, 'train/grad_norm_pre_clip': 11.701617240905762, 'timing/data': 0.03238355740904808, 'timing/model': 1.5116806644946337, 'learning_rate/qwen_vl_interface': 4.092336422450681e-06, 'learning_rate/action_model': 4.0923364224506804e-05, 'epoch': 1.14}) +[2026-06-16 03:33:40,246][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.005431319121271372, 'train/grad_norm_pre_clip': 0.6950052976608276, 'timing/data': 0.03178131394088268, 'timing/model': 1.5114725679159164, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 1.14}) +[2026-06-16 03:33:46,560][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.061778947710990906, 'train/grad_norm_pre_clip': 9.817875862121582, 'timing/data': 0.03187473677098751, 'timing/model': 1.5550088807940483, 'learning_rate/qwen_vl_interface': 4.086430196325106e-06, 'learning_rate/action_model': 4.086430196325106e-05, 'epoch': 1.14}) +[2026-06-16 03:33:52,791][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.40955108404159546, 'train/grad_norm_pre_clip': 12.363495826721191, 'timing/data': 0.040617069229483604, 'timing/model': 1.7724370639771223, 'learning_rate/qwen_vl_interface': 4.083477800104802e-06, 'learning_rate/action_model': 4.083477800104802e-05, 'epoch': 1.14}) +[2026-06-16 03:33:59,003][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.13955582678318024, 'train/grad_norm_pre_clip': 15.365816116333008, 'timing/data': 0.03177617862820625, 'timing/model': 1.6662805192172527, 'learning_rate/qwen_vl_interface': 4.080525883397711e-06, 'learning_rate/action_model': 4.080525883397711e-05, 'epoch': 1.14}) +[2026-06-16 03:34:04,961][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.0025751362554728985, 'train/grad_norm_pre_clip': 0.7706683874130249, 'timing/data': 0.030485162511467934, 'timing/model': 1.4907470978796482, 'learning_rate/qwen_vl_interface': 4.077574447417256e-06, 'learning_rate/action_model': 4.077574447417256e-05, 'epoch': 1.14}) +[2026-06-16 03:34:10,897][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.02864585816860199, 'train/grad_norm_pre_clip': 4.530617713928223, 'timing/data': 0.024638574570417404, 'timing/model': 1.4259036295115948, 'learning_rate/qwen_vl_interface': 4.074623493376657e-06, 'learning_rate/action_model': 4.074623493376657e-05, 'epoch': 1.14}) +[2026-06-16 03:34:16,790][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.06736179441213608, 'train/grad_norm_pre_clip': 14.49300479888916, 'timing/data': 0.027747193351387978, 'timing/model': 1.4254764001816511, 'learning_rate/qwen_vl_interface': 4.071673022488944e-06, 'learning_rate/action_model': 4.071673022488944e-05, 'epoch': 1.14}) +[2026-06-16 03:34:22,690][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.0038272743113338947, 'train/grad_norm_pre_clip': 0.5711231827735901, 'timing/data': 0.026042424142360687, 'timing/model': 1.42447180300951, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 1.14}) +[2026-06-16 03:34:28,737][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.16918425261974335, 'train/grad_norm_pre_clip': 11.149892807006836, 'timing/data': 0.026625964790582657, 'timing/model': 1.744119668379426, 'learning_rate/qwen_vl_interface': 4.065773535023277e-06, 'learning_rate/action_model': 4.0657735350232766e-05, 'epoch': 1.14}) +[2026-06-16 03:34:34,759][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.08381732553243637, 'train/grad_norm_pre_clip': 4.1013641357421875, 'timing/data': 0.024629464372992516, 'timing/model': 1.7662145532667637, 'learning_rate/qwen_vl_interface': 4.062824520870382e-06, 'learning_rate/action_model': 4.062824520870381e-05, 'epoch': 1.14}) +[2026-06-16 03:34:40,530][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.0820240005850792, 'train/grad_norm_pre_clip': 12.963432312011719, 'timing/data': 0.02626345120370388, 'timing/model': 1.5371508672833443, 'learning_rate/qwen_vl_interface': 4.059875994720481e-06, 'learning_rate/action_model': 4.059875994720481e-05, 'epoch': 1.14}) +[2026-06-16 03:34:46,525][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.002434404566884041, 'train/grad_norm_pre_clip': 0.7726824283599854, 'timing/data': 0.024739684537053108, 'timing/model': 1.5358023829758167, 'learning_rate/qwen_vl_interface': 4.056927957785605e-06, 'learning_rate/action_model': 4.056927957785605e-05, 'epoch': 1.14}) +[2026-06-16 03:34:52,396][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.005967068485915661, 'train/grad_norm_pre_clip': 0.7397370934486389, 'timing/data': 0.024105887860059738, 'timing/model': 1.4251207038760185, 'learning_rate/qwen_vl_interface': 4.0539804112775785e-06, 'learning_rate/action_model': 4.053980411277578e-05, 'epoch': 1.14}) +[2026-06-16 03:34:58,599][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.048555098474025726, 'train/grad_norm_pre_clip': 5.87846040725708, 'timing/data': 0.025528639554977417, 'timing/model': 1.704769292846322, 'learning_rate/qwen_vl_interface': 4.051033356408027e-06, 'learning_rate/action_model': 4.0510333564080264e-05, 'epoch': 1.14}) +[2026-06-16 03:35:04,505][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.008992738090455532, 'train/grad_norm_pre_clip': 1.1180702447891235, 'timing/data': 0.027082528918981552, 'timing/model': 1.6530301962047815, 'learning_rate/qwen_vl_interface': 4.048086794388375e-06, 'learning_rate/action_model': 4.048086794388375e-05, 'epoch': 1.14}) +[2026-06-16 03:35:10,204][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.06159578263759613, 'train/grad_norm_pre_clip': 12.521385192871094, 'timing/data': 0.026188723742961884, 'timing/model': 1.4387876689434052, 'learning_rate/qwen_vl_interface': 4.045140726429841e-06, 'learning_rate/action_model': 4.0451407264298406e-05, 'epoch': 1.14}) +[2026-06-16 03:35:16,353][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.06253618001937866, 'train/grad_norm_pre_clip': 8.661051750183105, 'timing/data': 0.02837548404932022, 'timing/model': 1.636304346844554, 'learning_rate/qwen_vl_interface': 4.042195153743443e-06, 'learning_rate/action_model': 4.042195153743443e-05, 'epoch': 1.14}) +[2026-06-16 03:35:22,290][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.09832006692886353, 'train/grad_norm_pre_clip': 11.492258071899414, 'timing/data': 0.025903409346938133, 'timing/model': 1.4367071241140366, 'learning_rate/qwen_vl_interface': 4.039250077539993e-06, 'learning_rate/action_model': 4.039250077539993e-05, 'epoch': 1.14}) +[2026-06-16 03:35:28,293][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.10782194882631302, 'train/grad_norm_pre_clip': 7.54045295715332, 'timing/data': 0.02637716755270958, 'timing/model': 1.5076480377465487, 'learning_rate/qwen_vl_interface': 4.036305499030104e-06, 'learning_rate/action_model': 4.0363054990301036e-05, 'epoch': 1.14}) +[2026-06-16 03:35:34,108][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.1878204643726349, 'train/grad_norm_pre_clip': 11.126495361328125, 'timing/data': 0.02668832801282406, 'timing/model': 1.4349350985139608, 'learning_rate/qwen_vl_interface': 4.033361419424178e-06, 'learning_rate/action_model': 4.033361419424178e-05, 'epoch': 1.14}) +[2026-06-16 03:35:39,947][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.03371242433786392, 'train/grad_norm_pre_clip': 2.619946002960205, 'timing/data': 0.025918880477547646, 'timing/model': 1.5919355023652315, 'learning_rate/qwen_vl_interface': 4.030417839932418e-06, 'learning_rate/action_model': 4.030417839932417e-05, 'epoch': 1.14}) +[2026-06-16 03:35:45,592][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.005693129263818264, 'train/grad_norm_pre_clip': 0.8667473793029785, 'timing/data': 0.02514858916401863, 'timing/model': 1.4416188225150108, 'learning_rate/qwen_vl_interface': 4.027474761764815e-06, 'learning_rate/action_model': 4.027474761764814e-05, 'epoch': 1.14}) +[2026-06-16 03:35:51,499][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.013270869851112366, 'train/grad_norm_pre_clip': 2.054617404937744, 'timing/data': 0.025210296735167503, 'timing/model': 1.4469002801924944, 'learning_rate/qwen_vl_interface': 4.024532186131158e-06, 'learning_rate/action_model': 4.0245321861311575e-05, 'epoch': 1.14}) +[2026-06-16 03:35:57,402][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.13224676251411438, 'train/grad_norm_pre_clip': 26.31156349182129, 'timing/data': 0.025595465674996376, 'timing/model': 1.4477160964161158, 'learning_rate/qwen_vl_interface': 4.021590114241029e-06, 'learning_rate/action_model': 4.021590114241029e-05, 'epoch': 1.14}) +[2026-06-16 03:36:03,314][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.07318010181188583, 'train/grad_norm_pre_clip': 9.984857559204102, 'timing/data': 0.02624106965959072, 'timing/model': 1.4388119503855705, 'learning_rate/qwen_vl_interface': 4.018648547303804e-06, 'learning_rate/action_model': 4.0186485473038036e-05, 'epoch': 1.15}) +[2026-06-16 03:36:09,311][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.018635934218764305, 'train/grad_norm_pre_clip': 6.748518943786621, 'timing/data': 0.024613622575998306, 'timing/model': 1.5472497344017029, 'learning_rate/qwen_vl_interface': 4.015707486528649e-06, 'learning_rate/action_model': 4.015707486528649e-05, 'epoch': 1.15}) +[2026-06-16 03:36:15,166][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.08147159963846207, 'train/grad_norm_pre_clip': 19.301008224487305, 'timing/data': 0.025769315659999847, 'timing/model': 1.6035046614706516, 'learning_rate/qwen_vl_interface': 4.012766933124523e-06, 'learning_rate/action_model': 4.012766933124523e-05, 'epoch': 1.15}) +[2026-06-16 03:36:21,069][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.11647486686706543, 'train/grad_norm_pre_clip': 12.818076133728027, 'timing/data': 0.031817490234971046, 'timing/model': 1.5348126646131277, 'learning_rate/qwen_vl_interface': 4.009826888300176e-06, 'learning_rate/action_model': 4.0098268883001755e-05, 'epoch': 1.15}) +[2026-06-16 03:36:26,888][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.20283165574073792, 'train/grad_norm_pre_clip': 15.00035572052002, 'timing/data': 0.025966748595237732, 'timing/model': 1.4518700800836086, 'learning_rate/qwen_vl_interface': 4.00688735326415e-06, 'learning_rate/action_model': 4.0068873532641495e-05, 'epoch': 1.15}) +[2026-06-16 03:36:32,788][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.05236285552382469, 'train/grad_norm_pre_clip': 15.478801727294922, 'timing/data': 0.02583114430308342, 'timing/model': 1.4505914449691772, 'learning_rate/qwen_vl_interface': 4.0039483292247765e-06, 'learning_rate/action_model': 4.003948329224776e-05, 'epoch': 1.15}) +[2026-06-16 03:36:38,688][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.16291135549545288, 'train/grad_norm_pre_clip': 17.96994400024414, 'timing/data': 0.026353253051638603, 'timing/model': 1.4727012384682894, 'learning_rate/qwen_vl_interface': 4.001009817390178e-06, 'learning_rate/action_model': 4.001009817390178e-05, 'epoch': 1.15}) +[2026-06-16 03:36:44,590][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.16493889689445496, 'train/grad_norm_pre_clip': 14.722515106201172, 'timing/data': 0.026349078863859177, 'timing/model': 1.65178856626153, 'learning_rate/qwen_vl_interface': 3.998071818968264e-06, 'learning_rate/action_model': 3.998071818968264e-05, 'epoch': 1.15}) +[2026-06-16 03:36:50,290][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.1986912190914154, 'train/grad_norm_pre_clip': 13.327213287353516, 'timing/data': 0.025745974853634834, 'timing/model': 1.4487220849841833, 'learning_rate/qwen_vl_interface': 3.995134335166736e-06, 'learning_rate/action_model': 3.995134335166736e-05, 'epoch': 1.15}) +[2026-06-16 03:36:56,224][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.13142864406108856, 'train/grad_norm_pre_clip': 13.046891212463379, 'timing/data': 0.02627861127257347, 'timing/model': 1.4767936076968908, 'learning_rate/qwen_vl_interface': 3.992197367193083e-06, 'learning_rate/action_model': 3.992197367193082e-05, 'epoch': 1.15}) +[2026-06-16 03:37:02,044][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.006536542437970638, 'train/grad_norm_pre_clip': 1.8899768590927124, 'timing/data': 0.0268991868942976, 'timing/model': 1.4711739588528872, 'learning_rate/qwen_vl_interface': 3.98926091625458e-06, 'learning_rate/action_model': 3.9892609162545795e-05, 'epoch': 1.15}) +[2026-06-16 03:37:07,927][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.310383141040802, 'train/grad_norm_pre_clip': 8.830460548400879, 'timing/data': 0.025126684457063675, 'timing/model': 1.532187007367611, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 1.15}) +[2026-06-16 03:37:13,814][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.15891598165035248, 'train/grad_norm_pre_clip': 12.950469017028809, 'timing/data': 0.02552592195570469, 'timing/model': 1.4695521797984838, 'learning_rate/qwen_vl_interface': 3.983389570311073e-06, 'learning_rate/action_model': 3.9833895703110724e-05, 'epoch': 1.15}) +[2026-06-16 03:37:19,701][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.08056604862213135, 'train/grad_norm_pre_clip': 4.876960277557373, 'timing/data': 0.025370119139552116, 'timing/model': 1.6553502473980188, 'learning_rate/qwen_vl_interface': 3.9804546777195546e-06, 'learning_rate/action_model': 3.980454677719554e-05, 'epoch': 1.15}) +[2026-06-16 03:37:25,434][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.02022172138094902, 'train/grad_norm_pre_clip': 9.478269577026367, 'timing/data': 0.026195067912340164, 'timing/model': 1.4648354146629572, 'learning_rate/qwen_vl_interface': 3.977520306990164e-06, 'learning_rate/action_model': 3.977520306990163e-05, 'epoch': 1.15}) +[2026-06-16 03:37:31,334][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.0767107903957367, 'train/grad_norm_pre_clip': 8.307859420776367, 'timing/data': 0.02555187977850437, 'timing/model': 1.459138199687004, 'learning_rate/qwen_vl_interface': 3.974586459329109e-06, 'learning_rate/action_model': 3.9745864593291085e-05, 'epoch': 1.15}) +[2026-06-16 03:37:37,326][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.00904582068324089, 'train/grad_norm_pre_clip': 2.2225608825683594, 'timing/data': 0.024862254038453102, 'timing/model': 1.5399772766977549, 'learning_rate/qwen_vl_interface': 3.9716531359423855e-06, 'learning_rate/action_model': 3.971653135942385e-05, 'epoch': 1.15}) +[2026-06-16 03:37:43,201][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.039871904999017715, 'train/grad_norm_pre_clip': 5.754057884216309, 'timing/data': 0.02443123608827591, 'timing/model': 1.45777521841228, 'learning_rate/qwen_vl_interface': 3.968720338035769e-06, 'learning_rate/action_model': 3.968720338035769e-05, 'epoch': 1.15}) +[2026-06-16 03:37:49,252][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.023806514218449593, 'train/grad_norm_pre_clip': 5.291821479797363, 'timing/data': 0.03215251490473747, 'timing/model': 1.5995261501520872, 'learning_rate/qwen_vl_interface': 3.965788066814824e-06, 'learning_rate/action_model': 3.965788066814823e-05, 'epoch': 1.15}) +[2026-06-16 03:37:55,496][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.1448335349559784, 'train/grad_norm_pre_clip': 7.321427345275879, 'timing/data': 0.032178161665797234, 'timing/model': 1.7848477773368359, 'learning_rate/qwen_vl_interface': 3.962856323484895e-06, 'learning_rate/action_model': 3.9628563234848945e-05, 'epoch': 1.15}) +[2026-06-16 03:38:01,549][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.19241200387477875, 'train/grad_norm_pre_clip': 16.656198501586914, 'timing/data': 0.03277242369949818, 'timing/model': 1.5352421011775732, 'learning_rate/qwen_vl_interface': 3.959925109251112e-06, 'learning_rate/action_model': 3.959925109251112e-05, 'epoch': 1.15}) +[2026-06-16 03:38:07,748][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.005081376060843468, 'train/grad_norm_pre_clip': 0.755683183670044, 'timing/data': 0.031706953421235085, 'timing/model': 1.5414110478013754, 'learning_rate/qwen_vl_interface': 3.956994425318387e-06, 'learning_rate/action_model': 3.956994425318387e-05, 'epoch': 1.15}) +[2026-06-16 03:38:13,945][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.10613011568784714, 'train/grad_norm_pre_clip': 11.329460144042969, 'timing/data': 0.03222567029297352, 'timing/model': 1.5253757983446121, 'learning_rate/qwen_vl_interface': 3.9540642728914115e-06, 'learning_rate/action_model': 3.9540642728914115e-05, 'epoch': 1.15}) +[2026-06-16 03:38:20,072][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.12031016498804092, 'train/grad_norm_pre_clip': 5.731673240661621, 'timing/data': 0.03302946127951145, 'timing/model': 1.5004477929323912, 'learning_rate/qwen_vl_interface': 3.951134653174664e-06, 'learning_rate/action_model': 3.951134653174663e-05, 'epoch': 1.15}) +[2026-06-16 03:38:26,253][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.11602084338665009, 'train/grad_norm_pre_clip': 13.229303359985352, 'timing/data': 0.03342572785913944, 'timing/model': 1.5263031385838985, 'learning_rate/qwen_vl_interface': 3.948205567372397e-06, 'learning_rate/action_model': 3.948205567372397e-05, 'epoch': 1.15}) +[2026-06-16 03:38:32,508][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.04484681040048599, 'train/grad_norm_pre_clip': 4.125152587890625, 'timing/data': 0.03293560445308685, 'timing/model': 1.803286211565137, 'learning_rate/qwen_vl_interface': 3.945277016688647e-06, 'learning_rate/action_model': 3.945277016688647e-05, 'epoch': 1.15}) +[2026-06-16 03:38:38,658][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.44824278354644775, 'train/grad_norm_pre_clip': 23.679296493530273, 'timing/data': 0.032763632014393806, 'timing/model': 1.5143878925591707, 'learning_rate/qwen_vl_interface': 3.9423490023272326e-06, 'learning_rate/action_model': 3.9423490023272324e-05, 'epoch': 1.16}) +[2026-06-16 03:38:44,853][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.00552408816292882, 'train/grad_norm_pre_clip': 0.4939368963241577, 'timing/data': 0.031321996822953224, 'timing/model': 1.5228637102991343, 'learning_rate/qwen_vl_interface': 3.939421525491748e-06, 'learning_rate/action_model': 3.9394215254917483e-05, 'epoch': 1.16}) +[2026-06-16 03:38:51,195][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.19996435940265656, 'train/grad_norm_pre_clip': 17.88081169128418, 'timing/data': 0.032116301357746124, 'timing/model': 1.6224903035908937, 'learning_rate/qwen_vl_interface': 3.9364945873855705e-06, 'learning_rate/action_model': 3.93649458738557e-05, 'epoch': 1.16}) +[2026-06-16 03:38:57,804][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.16451199352741241, 'train/grad_norm_pre_clip': 10.297226905822754, 'timing/data': 0.03279932960867882, 'timing/model': 1.6356988362967968, 'learning_rate/qwen_vl_interface': 3.9335681892118536e-06, 'learning_rate/action_model': 3.933568189211854e-05, 'epoch': 1.16}) +[2026-06-16 03:39:04,280][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.01697014458477497, 'train/grad_norm_pre_clip': 3.5525643825531006, 'timing/data': 0.032927922904491425, 'timing/model': 1.5809776652604342, 'learning_rate/qwen_vl_interface': 3.930642332173527e-06, 'learning_rate/action_model': 3.930642332173527e-05, 'epoch': 1.16}) +[2026-06-16 03:39:10,634][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.16679854691028595, 'train/grad_norm_pre_clip': 25.695743560791016, 'timing/data': 0.03268388845026493, 'timing/model': 1.780890490859747, 'learning_rate/qwen_vl_interface': 3.927717017473301e-06, 'learning_rate/action_model': 3.9277170174733006e-05, 'epoch': 1.16}) +[2026-06-16 03:39:16,689][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.08222360908985138, 'train/grad_norm_pre_clip': 8.233412742614746, 'timing/data': 0.03246612474322319, 'timing/model': 1.5437619648873806, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 1.16}) +[2026-06-16 03:39:23,221][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.004907016176730394, 'train/grad_norm_pre_clip': 0.4825482666492462, 'timing/data': 0.038102056831121445, 'timing/model': 1.6134571731090546, 'learning_rate/qwen_vl_interface': 3.921868019896873e-06, 'learning_rate/action_model': 3.921868019896873e-05, 'epoch': 1.16}) +[2026-06-16 03:39:29,726][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.16925722360610962, 'train/grad_norm_pre_clip': 11.74224853515625, 'timing/data': 0.04657946899533272, 'timing/model': 1.569787185639143, 'learning_rate/qwen_vl_interface': 3.918944339424974e-06, 'learning_rate/action_model': 3.9189443394249736e-05, 'epoch': 1.16}) +[2026-06-16 03:39:36,219][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.07773829996585846, 'train/grad_norm_pre_clip': 15.446538925170898, 'timing/data': 0.04520257189869881, 'timing/model': 1.5985793378204107, 'learning_rate/qwen_vl_interface': 3.9160212060997774e-06, 'learning_rate/action_model': 3.9160212060997774e-05, 'epoch': 1.16}) +[2026-06-16 03:39:42,733][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.0032849153503775597, 'train/grad_norm_pre_clip': 0.2677605450153351, 'timing/data': 0.04648414067924023, 'timing/model': 1.5842179786413908, 'learning_rate/qwen_vl_interface': 3.913098621122875e-06, 'learning_rate/action_model': 3.913098621122875e-05, 'epoch': 1.16}) +[2026-06-16 03:39:48,688][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.04841959848999977, 'train/grad_norm_pre_clip': 8.645476341247559, 'timing/data': 0.0253309179097414, 'timing/model': 1.641103284433484, 'learning_rate/qwen_vl_interface': 3.91017658569563e-06, 'learning_rate/action_model': 3.91017658569563e-05, 'epoch': 1.16}) +[2026-06-16 03:39:54,393][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.06124180182814598, 'train/grad_norm_pre_clip': 11.08900260925293, 'timing/data': 0.026809943839907646, 'timing/model': 1.439903000369668, 'learning_rate/qwen_vl_interface': 3.907255101019181e-06, 'learning_rate/action_model': 3.907255101019181e-05, 'epoch': 1.16}) +[2026-06-16 03:40:00,282][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.05280295014381409, 'train/grad_norm_pre_clip': 5.643263816833496, 'timing/data': 0.025672119110822678, 'timing/model': 1.4273255355656147, 'learning_rate/qwen_vl_interface': 3.904334168294442e-06, 'learning_rate/action_model': 3.9043341682944415e-05, 'epoch': 1.16}) +[2026-06-16 03:40:06,446][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.07658363878726959, 'train/grad_norm_pre_clip': 8.55333423614502, 'timing/data': 0.03278584964573383, 'timing/model': 1.5336809046566486, 'learning_rate/qwen_vl_interface': 3.901413788722093e-06, 'learning_rate/action_model': 3.901413788722093e-05, 'epoch': 1.16}) +[2026-06-16 03:47:22,342][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.16760334372520447, 'train/grad_norm_pre_clip': 12.573572158813477, 'eval/loss': 0.054805412888526917, 'eval/flap_precision': 0.9175126903553299, 'eval/flap_recall': 0.9810040705563093, 'eval/flap_f1': 0.9481967213114753, 'eval/flap_rate_model': 0.21888888888888888, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9481967213114751, 'timing/data': 0.03056272864341736, 'timing/model': 1.5143509898334742, 'learning_rate/qwen_vl_interface': 3.8984939635025966e-06, 'learning_rate/action_model': 3.898493963502596e-05, 'epoch': 1.16}) +[2026-06-16 03:47:22,343][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3000_state +[2026-06-16 03:47:22,347][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 03:47:28,122][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3000_state/model.safetensors +[2026-06-16 03:47:37,558][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3000_state/optimizer.bin +[2026-06-16 03:47:37,558][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3000_state/sampler.bin +[2026-06-16 03:47:37,558][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-16 03:47:37,559][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-16 03:47:37,661][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-16 03:47:37,661][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 03:47:37,670][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 03:47:43,964][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.06215260177850723, 'train/grad_norm_pre_clip': 7.867335796356201, 'timing/data': 0.03241666033864021, 'timing/model': 1.5160524398088455, 'learning_rate/qwen_vl_interface': 3.895574693836181e-06, 'learning_rate/action_model': 3.8955746938361804e-05, 'epoch': 1.16}) +[2026-06-16 03:47:50,054][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.1041724756360054, 'train/grad_norm_pre_clip': 16.80843734741211, 'timing/data': 0.03235273249447346, 'timing/model': 1.5602577589452267, 'learning_rate/qwen_vl_interface': 3.892655980922847e-06, 'learning_rate/action_model': 3.8926559809228466e-05, 'epoch': 1.16}) +[2026-06-16 03:47:56,559][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.29326388239860535, 'train/grad_norm_pre_clip': 22.799381256103516, 'timing/data': 0.0319965947419405, 'timing/model': 1.576400462538004, 'learning_rate/qwen_vl_interface': 3.889737825962368e-06, 'learning_rate/action_model': 3.889737825962368e-05, 'epoch': 1.16}) +[2026-06-16 03:48:02,794][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.06698952615261078, 'train/grad_norm_pre_clip': 8.555118560791016, 'timing/data': 0.031198855489492416, 'timing/model': 1.7674181442707777, 'learning_rate/qwen_vl_interface': 3.886820230154288e-06, 'learning_rate/action_model': 3.886820230154287e-05, 'epoch': 1.16}) +[2026-06-16 03:48:08,860][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.09863648563623428, 'train/grad_norm_pre_clip': 15.256725311279297, 'timing/data': 0.032271839678287506, 'timing/model': 1.5276369228959084, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 1.16}) +[2026-06-16 03:48:15,061][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.04213084653019905, 'train/grad_norm_pre_clip': 9.149386405944824, 'timing/data': 0.030312784016132355, 'timing/model': 1.5395004972815514, 'learning_rate/qwen_vl_interface': 3.880986720792344e-06, 'learning_rate/action_model': 3.880986720792344e-05, 'epoch': 1.16}) +[2026-06-16 03:48:21,164][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.1197173222899437, 'train/grad_norm_pre_clip': 12.58115291595459, 'timing/data': 0.032359927892684937, 'timing/model': 1.5387696325778961, 'learning_rate/qwen_vl_interface': 3.878070809636416e-06, 'learning_rate/action_model': 3.878070809636416e-05, 'epoch': 1.16}) +[2026-06-16 03:48:27,372][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.013221422210335732, 'train/grad_norm_pre_clip': 1.2477571964263916, 'timing/data': 0.03348875790834427, 'timing/model': 1.5413394439965487, 'learning_rate/qwen_vl_interface': 3.875155462428758e-06, 'learning_rate/action_model': 3.875155462428758e-05, 'epoch': 1.16}) +[2026-06-16 03:48:33,601][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.020262431353330612, 'train/grad_norm_pre_clip': 3.8458938598632812, 'timing/data': 0.031662099063396454, 'timing/model': 1.526295192539692, 'learning_rate/qwen_vl_interface': 3.872240680367754e-06, 'learning_rate/action_model': 3.872240680367754e-05, 'epoch': 1.16}) +[2026-06-16 03:48:39,914][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.041412580758333206, 'train/grad_norm_pre_clip': 2.841771125793457, 'timing/data': 0.032436372712254524, 'timing/model': 1.800546320155263, 'learning_rate/qwen_vl_interface': 3.869326464651566e-06, 'learning_rate/action_model': 3.869326464651566e-05, 'epoch': 1.16}) +[2026-06-16 03:48:45,872][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.04905565828084946, 'train/grad_norm_pre_clip': 7.468934535980225, 'timing/data': 0.03196088783442974, 'timing/model': 1.5921361912041903, 'learning_rate/qwen_vl_interface': 3.866412816478114e-06, 'learning_rate/action_model': 3.8664128164781144e-05, 'epoch': 1.17}) +[2026-06-16 03:48:52,146][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.04228067398071289, 'train/grad_norm_pre_clip': 5.169809818267822, 'timing/data': 0.03385987505316734, 'timing/model': 1.5403537806123495, 'learning_rate/qwen_vl_interface': 3.8634997370450915e-06, 'learning_rate/action_model': 3.8634997370450914e-05, 'epoch': 1.17}) +[2026-06-16 03:48:58,339][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.044783808290958405, 'train/grad_norm_pre_clip': 5.00722074508667, 'timing/data': 0.03248327784240246, 'timing/model': 1.5410576295107603, 'learning_rate/qwen_vl_interface': 3.860587227549956e-06, 'learning_rate/action_model': 3.860587227549955e-05, 'epoch': 1.17}) +[2026-06-16 03:49:04,538][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.012529335916042328, 'train/grad_norm_pre_clip': 1.7501590251922607, 'timing/data': 0.03440430201590061, 'timing/model': 1.53288234770298, 'learning_rate/qwen_vl_interface': 3.857675289189926e-06, 'learning_rate/action_model': 3.857675289189926e-05, 'epoch': 1.17}) +[2026-06-16 03:49:10,641][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.0021795802749693394, 'train/grad_norm_pre_clip': 0.3595297932624817, 'timing/data': 0.03083457425236702, 'timing/model': 1.4981097839772701, 'learning_rate/qwen_vl_interface': 3.854763923161993e-06, 'learning_rate/action_model': 3.854763923161992e-05, 'epoch': 1.17}) +[2026-06-16 03:49:16,775][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.04614594206213951, 'train/grad_norm_pre_clip': 4.438094139099121, 'timing/data': 0.03326638415455818, 'timing/model': 1.7674998305737972, 'learning_rate/qwen_vl_interface': 3.851853130662908e-06, 'learning_rate/action_model': 3.8518531306629077e-05, 'epoch': 1.17}) +[2026-06-16 03:49:23,014][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.0391903854906559, 'train/grad_norm_pre_clip': 7.35899019241333, 'timing/data': 0.03241730108857155, 'timing/model': 1.7968672402203083, 'learning_rate/qwen_vl_interface': 3.8489429128891884e-06, 'learning_rate/action_model': 3.848942912889188e-05, 'epoch': 1.17}) +[2026-06-16 03:49:28,952][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.07176099717617035, 'train/grad_norm_pre_clip': 11.460695266723633, 'timing/data': 0.03219764307141304, 'timing/model': 1.5225035715848207, 'learning_rate/qwen_vl_interface': 3.846033271037114e-06, 'learning_rate/action_model': 3.8460332710371136e-05, 'epoch': 1.17}) +[2026-06-16 03:49:35,149][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.09006275981664658, 'train/grad_norm_pre_clip': 6.014003753662109, 'timing/data': 0.03180772066116333, 'timing/model': 1.5386620685458183, 'learning_rate/qwen_vl_interface': 3.843124206302732e-06, 'learning_rate/action_model': 3.8431242063027316e-05, 'epoch': 1.17}) +[2026-06-16 03:49:41,241][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.14186044037342072, 'train/grad_norm_pre_clip': 5.521825313568115, 'timing/data': 0.032727329060435295, 'timing/model': 1.5349725224077702, 'learning_rate/qwen_vl_interface': 3.840215719881844e-06, 'learning_rate/action_model': 3.8402157198818445e-05, 'epoch': 1.17}) +[2026-06-16 03:49:47,357][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.062085580080747604, 'train/grad_norm_pre_clip': 4.428066253662109, 'timing/data': 0.03227187134325504, 'timing/model': 1.502771144732833, 'learning_rate/qwen_vl_interface': 3.837307812970022e-06, 'learning_rate/action_model': 3.837307812970022e-05, 'epoch': 1.17}) +[2026-06-16 03:49:53,518][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.16049696505069733, 'train/grad_norm_pre_clip': 22.664596557617188, 'timing/data': 0.0326352845877409, 'timing/model': 1.8096827398985624, 'learning_rate/qwen_vl_interface': 3.834400486762596e-06, 'learning_rate/action_model': 3.834400486762596e-05, 'epoch': 1.17}) +[2026-06-16 03:49:59,562][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.10871002823114395, 'train/grad_norm_pre_clip': 6.847458362579346, 'timing/data': 0.03203098103404045, 'timing/model': 1.5687583107501268, 'learning_rate/qwen_vl_interface': 3.8314937424546585e-06, 'learning_rate/action_model': 3.8314937424546584e-05, 'epoch': 1.17}) +[2026-06-16 03:50:05,747][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.012413883581757545, 'train/grad_norm_pre_clip': 2.5449795722961426, 'timing/data': 0.0331101194024086, 'timing/model': 1.5283448342233896, 'learning_rate/qwen_vl_interface': 3.828587581241064e-06, 'learning_rate/action_model': 3.8285875812410635e-05, 'epoch': 1.17}) +[2026-06-16 03:50:11,972][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.025796804577112198, 'train/grad_norm_pre_clip': 6.772104740142822, 'timing/data': 0.03156220726668835, 'timing/model': 1.562138944864273, 'learning_rate/qwen_vl_interface': 3.825682004316419e-06, 'learning_rate/action_model': 3.825682004316419e-05, 'epoch': 1.17}) +[2026-06-16 03:50:18,163][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.15976709127426147, 'train/grad_norm_pre_clip': 8.00932502746582, 'timing/data': 0.033290283754467964, 'timing/model': 1.4995580557733774, 'learning_rate/qwen_vl_interface': 3.822777012875104e-06, 'learning_rate/action_model': 3.822777012875103e-05, 'epoch': 1.17}) +[2026-06-16 03:50:24,448][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.02633049339056015, 'train/grad_norm_pre_clip': 3.5992531776428223, 'timing/data': 0.03223692066967487, 'timing/model': 1.5650905407965183, 'learning_rate/qwen_vl_interface': 3.819872608111248e-06, 'learning_rate/action_model': 3.819872608111248e-05, 'epoch': 1.17}) +[2026-06-16 03:50:30,535][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.09506980329751968, 'train/grad_norm_pre_clip': 5.393341064453125, 'timing/data': 0.03233577497303486, 'timing/model': 1.7244451083242893, 'learning_rate/qwen_vl_interface': 3.816968791218741e-06, 'learning_rate/action_model': 3.816968791218741e-05, 'epoch': 1.17}) +[2026-06-16 03:50:36,450][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.019146712496876717, 'train/grad_norm_pre_clip': 2.407288074493408, 'timing/data': 0.03131677955389023, 'timing/model': 1.5264641605317593, 'learning_rate/qwen_vl_interface': 3.8140655633912337e-06, 'learning_rate/action_model': 3.814065563391234e-05, 'epoch': 1.17}) +[2026-06-16 03:50:42,648][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.2955322861671448, 'train/grad_norm_pre_clip': 15.949808120727539, 'timing/data': 0.032700760290026665, 'timing/model': 1.5397506766021252, 'learning_rate/qwen_vl_interface': 3.811162925822133e-06, 'learning_rate/action_model': 3.811162925822132e-05, 'epoch': 1.17}) +[2026-06-16 03:50:48,759][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.008674106560647488, 'train/grad_norm_pre_clip': 1.1877259016036987, 'timing/data': 0.03592259623110294, 'timing/model': 1.5402149800211191, 'learning_rate/qwen_vl_interface': 3.8082608797046045e-06, 'learning_rate/action_model': 3.8082608797046045e-05, 'epoch': 1.17}) +[2026-06-16 03:50:55,144][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.013876141048967838, 'train/grad_norm_pre_clip': 4.204444885253906, 'timing/data': 0.036533182486891747, 'timing/model': 1.4917729999870062, 'learning_rate/qwen_vl_interface': 3.8053594262315687e-06, 'learning_rate/action_model': 3.805359426231568e-05, 'epoch': 1.17}) +[2026-06-16 03:51:01,343][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.13157495856285095, 'train/grad_norm_pre_clip': 13.535337448120117, 'timing/data': 0.03155771270394325, 'timing/model': 1.5264468044042587, 'learning_rate/qwen_vl_interface': 3.802458566595704e-06, 'learning_rate/action_model': 3.802458566595704e-05, 'epoch': 1.17}) +[2026-06-16 03:51:07,444][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.205709770321846, 'train/grad_norm_pre_clip': 12.749688148498535, 'timing/data': 0.0336285550147295, 'timing/model': 1.6868057195097208, 'learning_rate/qwen_vl_interface': 3.7995583019894444e-06, 'learning_rate/action_model': 3.7995583019894445e-05, 'epoch': 1.17}) +[2026-06-16 03:51:13,344][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.1962321549654007, 'train/grad_norm_pre_clip': 9.269429206848145, 'timing/data': 0.031447356566786766, 'timing/model': 1.4990731254220009, 'learning_rate/qwen_vl_interface': 3.7966586336049782e-06, 'learning_rate/action_model': 3.796658633604978e-05, 'epoch': 1.17}) +[2026-06-16 03:51:19,293][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.21171584725379944, 'train/grad_norm_pre_clip': 9.572242736816406, 'timing/data': 0.027639908716082573, 'timing/model': 1.4237977024167776, 'learning_rate/qwen_vl_interface': 3.7937595626342516e-06, 'learning_rate/action_model': 3.793759562634252e-05, 'epoch': 1.17}) +[2026-06-16 03:51:25,239][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.08656923472881317, 'train/grad_norm_pre_clip': 15.22647762298584, 'timing/data': 0.026716740801930428, 'timing/model': 1.4692489709705114, 'learning_rate/qwen_vl_interface': 3.79086109026896e-06, 'learning_rate/action_model': 3.7908610902689604e-05, 'epoch': 1.18}) +[2026-06-16 03:51:31,527][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.024649549275636673, 'train/grad_norm_pre_clip': 6.075740337371826, 'timing/data': 0.030222468078136444, 'timing/model': 1.5020599532872438, 'learning_rate/qwen_vl_interface': 3.7879632177005584e-06, 'learning_rate/action_model': 3.787963217700558e-05, 'epoch': 1.18}) +[2026-06-16 03:51:37,766][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.0011840590741485357, 'train/grad_norm_pre_clip': 0.11845169216394424, 'timing/data': 0.03239375352859497, 'timing/model': 1.7341260891407728, 'learning_rate/qwen_vl_interface': 3.7850659461202523e-06, 'learning_rate/action_model': 3.785065946120252e-05, 'epoch': 1.18}) +[2026-06-16 03:51:43,522][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.040447529405355453, 'train/grad_norm_pre_clip': 5.324073791503906, 'timing/data': 0.02948332019150257, 'timing/model': 1.521960174664855, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 1.18}) +[2026-06-16 03:51:49,522][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.002951784757897258, 'train/grad_norm_pre_clip': 0.5700505375862122, 'timing/data': 0.02712116576731205, 'timing/model': 1.5346595980226994, 'learning_rate/qwen_vl_interface': 3.779273210687514e-06, 'learning_rate/action_model': 3.779273210687514e-05, 'epoch': 1.18}) +[2026-06-16 03:51:55,424][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.05381181836128235, 'train/grad_norm_pre_clip': 4.810877323150635, 'timing/data': 0.026475798338651657, 'timing/model': 1.535099245607853, 'learning_rate/qwen_vl_interface': 3.776377749216257e-06, 'learning_rate/action_model': 3.776377749216257e-05, 'epoch': 1.18}) +[2026-06-16 03:52:01,318][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.05587755888700485, 'train/grad_norm_pre_clip': 6.322068214416504, 'timing/data': 0.03275041654706001, 'timing/model': 1.5239907521754503, 'learning_rate/qwen_vl_interface': 3.773482893495442e-06, 'learning_rate/action_model': 3.773482893495442e-05, 'epoch': 1.18}) +[2026-06-16 03:52:07,315][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.0056015183217823505, 'train/grad_norm_pre_clip': 0.8340734839439392, 'timing/data': 0.03481752425432205, 'timing/model': 1.5425112042576075, 'learning_rate/qwen_vl_interface': 3.7705886447150375e-06, 'learning_rate/action_model': 3.7705886447150376e-05, 'epoch': 1.18}) +[2026-06-16 03:52:13,430][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.005757725797593594, 'train/grad_norm_pre_clip': 0.8098742961883545, 'timing/data': 0.038943178951740265, 'timing/model': 1.7729164958000183, 'learning_rate/qwen_vl_interface': 3.767695004064759e-06, 'learning_rate/action_model': 3.767695004064759e-05, 'epoch': 1.18}) +[2026-06-16 03:52:19,481][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.10052594542503357, 'train/grad_norm_pre_clip': 13.659022331237793, 'timing/data': 0.03204681910574436, 'timing/model': 1.5202652495354414, 'learning_rate/qwen_vl_interface': 3.7648019727340725e-06, 'learning_rate/action_model': 3.764801972734072e-05, 'epoch': 1.18}) +[2026-06-16 03:52:25,808][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.06506125628948212, 'train/grad_norm_pre_clip': 8.139141082763672, 'timing/data': 0.03886908479034901, 'timing/model': 1.5504476334899664, 'learning_rate/qwen_vl_interface': 3.761909551912196e-06, 'learning_rate/action_model': 3.761909551912196e-05, 'epoch': 1.18}) +[2026-06-16 03:52:32,781][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.029740063473582268, 'train/grad_norm_pre_clip': 12.087883949279785, 'timing/data': 0.031526895239949226, 'timing/model': 1.5150798503309488, 'learning_rate/qwen_vl_interface': 3.75901774278809e-06, 'learning_rate/action_model': 3.7590177427880904e-05, 'epoch': 1.18}) +[2026-06-16 03:52:39,083][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.17387911677360535, 'train/grad_norm_pre_clip': 16.411846160888672, 'timing/data': 0.03183751180768013, 'timing/model': 1.51491329818964, 'learning_rate/qwen_vl_interface': 3.756126546550472e-06, 'learning_rate/action_model': 3.756126546550472e-05, 'epoch': 1.18}) +[2026-06-16 03:52:45,383][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.009491726756095886, 'train/grad_norm_pre_clip': 1.2073966264724731, 'timing/data': 0.03083810769021511, 'timing/model': 1.513718619942665, 'learning_rate/qwen_vl_interface': 3.753235964387802e-06, 'learning_rate/action_model': 3.753235964387802e-05, 'epoch': 1.18}) +[2026-06-16 03:52:51,840][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.22561344504356384, 'train/grad_norm_pre_clip': 19.294692993164062, 'timing/data': 0.03348162770271301, 'timing/model': 1.8397315945476294, 'learning_rate/qwen_vl_interface': 3.750345997488289e-06, 'learning_rate/action_model': 3.7503459974882886e-05, 'epoch': 1.18}) +[2026-06-16 03:52:57,887][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.25252437591552734, 'train/grad_norm_pre_clip': 11.198570251464844, 'timing/data': 0.03157494589686394, 'timing/model': 1.517622584477067, 'learning_rate/qwen_vl_interface': 3.7474566470398903e-06, 'learning_rate/action_model': 3.74745664703989e-05, 'epoch': 1.18}) +[2026-06-16 03:53:04,190][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.05712290108203888, 'train/grad_norm_pre_clip': 5.140963554382324, 'timing/data': 0.03226248733699322, 'timing/model': 1.5373298097401857, 'learning_rate/qwen_vl_interface': 3.744567914230307e-06, 'learning_rate/action_model': 3.744567914230307e-05, 'epoch': 1.18}) +[2026-06-16 03:53:10,906][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.013831665739417076, 'train/grad_norm_pre_clip': 3.9999730587005615, 'timing/data': 0.03872708044946194, 'timing/model': 1.6075061969459057, 'learning_rate/qwen_vl_interface': 3.7416798002469902e-06, 'learning_rate/action_model': 3.74167980024699e-05, 'epoch': 1.18}) +[2026-06-16 03:53:17,306][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.02310633659362793, 'train/grad_norm_pre_clip': 8.782747268676758, 'timing/data': 0.03549689054489136, 'timing/model': 1.5530126243829727, 'learning_rate/qwen_vl_interface': 3.7387923062771313e-06, 'learning_rate/action_model': 3.738792306277131e-05, 'epoch': 1.18}) +[2026-06-16 03:53:23,645][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.06406814604997635, 'train/grad_norm_pre_clip': 15.773710250854492, 'timing/data': 0.03282236494123936, 'timing/model': 1.5396024640649557, 'learning_rate/qwen_vl_interface': 3.7359054335076732e-06, 'learning_rate/action_model': 3.7359054335076734e-05, 'epoch': 1.18}) +[2026-06-16 03:53:29,912][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.08439194411039352, 'train/grad_norm_pre_clip': 6.290318489074707, 'timing/data': 0.031046560034155846, 'timing/model': 1.7973761688917875, 'learning_rate/qwen_vl_interface': 3.7330191831253006e-06, 'learning_rate/action_model': 3.7330191831253e-05, 'epoch': 1.18}) +[2026-06-16 03:53:35,945][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.1489674597978592, 'train/grad_norm_pre_clip': 16.24439239501953, 'timing/data': 0.031140148639678955, 'timing/model': 1.5048888474702835, 'learning_rate/qwen_vl_interface': 3.7301335563164382e-06, 'learning_rate/action_model': 3.730133556316438e-05, 'epoch': 1.18}) +[2026-06-16 03:53:42,143][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.13239923119544983, 'train/grad_norm_pre_clip': 7.926433086395264, 'timing/data': 0.0322866253554821, 'timing/model': 1.534780703485012, 'learning_rate/qwen_vl_interface': 3.727248554267261e-06, 'learning_rate/action_model': 3.727248554267261e-05, 'epoch': 1.18}) +[2026-06-16 03:53:48,359][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.051839396357536316, 'train/grad_norm_pre_clip': 6.219491958618164, 'timing/data': 0.03275291807949543, 'timing/model': 1.5154617056250572, 'learning_rate/qwen_vl_interface': 3.7243641781636838e-06, 'learning_rate/action_model': 3.7243641781636836e-05, 'epoch': 1.18}) +[2026-06-16 03:53:54,295][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.1514534205198288, 'train/grad_norm_pre_clip': 18.11985969543457, 'timing/data': 0.027444349601864815, 'timing/model': 1.432226400822401, 'learning_rate/qwen_vl_interface': 3.721480429191364e-06, 'learning_rate/action_model': 3.721480429191364e-05, 'epoch': 1.18}) +[2026-06-16 03:54:00,282][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.06754656881093979, 'train/grad_norm_pre_clip': 14.612454414367676, 'timing/data': 0.02714737132191658, 'timing/model': 1.4973601717501879, 'learning_rate/qwen_vl_interface': 3.7185973085357045e-06, 'learning_rate/action_model': 3.718597308535705e-05, 'epoch': 1.19}) +[2026-06-16 03:54:06,224][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.26069387793540955, 'train/grad_norm_pre_clip': 21.620941162109375, 'timing/data': 0.026820475235581398, 'timing/model': 1.6556343846023083, 'learning_rate/qwen_vl_interface': 3.7157148173818443e-06, 'learning_rate/action_model': 3.715714817381844e-05, 'epoch': 1.19}) +[2026-06-16 03:54:11,994][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.07072984427213669, 'train/grad_norm_pre_clip': 15.4092378616333, 'timing/data': 0.02684524841606617, 'timing/model': 1.4955666940659285, 'learning_rate/qwen_vl_interface': 3.7128329569146672e-06, 'learning_rate/action_model': 3.7128329569146666e-05, 'epoch': 1.19}) +[2026-06-16 03:54:17,887][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.14261852204799652, 'train/grad_norm_pre_clip': 7.707554340362549, 'timing/data': 0.025694018229842186, 'timing/model': 1.4278090968728065, 'learning_rate/qwen_vl_interface': 3.7099517283187977e-06, 'learning_rate/action_model': 3.7099517283187976e-05, 'epoch': 1.19}) +[2026-06-16 03:54:23,697][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.014625116251409054, 'train/grad_norm_pre_clip': 1.9213165044784546, 'timing/data': 0.026186438277363777, 'timing/model': 1.423425953835249, 'learning_rate/qwen_vl_interface': 3.7070711327786e-06, 'learning_rate/action_model': 3.7070711327785997e-05, 'epoch': 1.19}) +[2026-06-16 03:54:29,489][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.22739234566688538, 'train/grad_norm_pre_clip': 10.99299430847168, 'timing/data': 0.02650434896349907, 'timing/model': 1.4275127090513706, 'learning_rate/qwen_vl_interface': 3.704191171478179e-06, 'learning_rate/action_model': 3.704191171478179e-05, 'epoch': 1.19}) +[2026-06-16 03:54:35,653][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.10097381472587585, 'train/grad_norm_pre_clip': 14.303536415100098, 'timing/data': 0.033178091049194336, 'timing/model': 1.4916234854608774, 'learning_rate/qwen_vl_interface': 3.701311845601376e-06, 'learning_rate/action_model': 3.701311845601376e-05, 'epoch': 1.19}) +[2026-06-16 03:54:41,923][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.03672497346997261, 'train/grad_norm_pre_clip': 8.932029724121094, 'timing/data': 0.033541832119226456, 'timing/model': 1.8210950326174498, 'learning_rate/qwen_vl_interface': 3.6984331563317744e-06, 'learning_rate/action_model': 3.698433156331774e-05, 'epoch': 1.19}) +[2026-06-16 03:54:47,848][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.031864915043115616, 'train/grad_norm_pre_clip': 6.0383124351501465, 'timing/data': 0.03124169260263443, 'timing/model': 1.4988062102347612, 'learning_rate/qwen_vl_interface': 3.695555104852695e-06, 'learning_rate/action_model': 3.6955551048526946e-05, 'epoch': 1.19}) +[2026-06-16 03:54:54,066][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.03856249898672104, 'train/grad_norm_pre_clip': 7.33026123046875, 'timing/data': 0.03269371576607227, 'timing/model': 1.559790637344122, 'learning_rate/qwen_vl_interface': 3.6926776923471948e-06, 'learning_rate/action_model': 3.6926776923471946e-05, 'epoch': 1.19}) +[2026-06-16 03:55:00,140][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.2111383080482483, 'train/grad_norm_pre_clip': 8.951531410217285, 'timing/data': 0.033199720084667206, 'timing/model': 1.5255443323403597, 'learning_rate/qwen_vl_interface': 3.6898009199980715e-06, 'learning_rate/action_model': 3.6898009199980714e-05, 'epoch': 1.19}) +[2026-06-16 03:55:05,880][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.07463668286800385, 'train/grad_norm_pre_clip': 6.413533687591553, 'timing/data': 0.026555504649877548, 'timing/model': 1.4324065204709768, 'learning_rate/qwen_vl_interface': 3.686924788987855e-06, 'learning_rate/action_model': 3.686924788987855e-05, 'epoch': 1.19}) +[2026-06-16 03:55:11,780][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.27819597721099854, 'train/grad_norm_pre_clip': 13.028166770935059, 'timing/data': 0.025570852681994438, 'timing/model': 1.417900973930955, 'learning_rate/qwen_vl_interface': 3.6840493004988155e-06, 'learning_rate/action_model': 3.684049300498815e-05, 'epoch': 1.19}) +[2026-06-16 03:55:17,844][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.006310633383691311, 'train/grad_norm_pre_clip': 2.15181040763855, 'timing/data': 0.03187932074069977, 'timing/model': 1.7565055079758167, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 1.19}) +[2026-06-16 03:55:23,575][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.006158880889415741, 'train/grad_norm_pre_clip': 0.5997951626777649, 'timing/data': 0.027861682698130608, 'timing/model': 1.4921833518892527, 'learning_rate/qwen_vl_interface': 3.6783002558120206e-06, 'learning_rate/action_model': 3.67830025581202e-05, 'epoch': 1.19}) +[2026-06-16 03:55:29,937][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.01800158992409706, 'train/grad_norm_pre_clip': 2.1484580039978027, 'timing/data': 0.03161306492984295, 'timing/model': 1.5261655934154987, 'learning_rate/qwen_vl_interface': 3.6754267019774804e-06, 'learning_rate/action_model': 3.67542670197748e-05, 'epoch': 1.19}) +[2026-06-16 03:55:36,038][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.016247892752289772, 'train/grad_norm_pre_clip': 2.2581825256347656, 'timing/data': 0.03161722049117088, 'timing/model': 1.4999500196427107, 'learning_rate/qwen_vl_interface': 3.6725537953905453e-06, 'learning_rate/action_model': 3.672553795390545e-05, 'epoch': 1.19}) +[2026-06-16 03:55:42,145][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.004667338449507952, 'train/grad_norm_pre_clip': 0.6653912663459778, 'timing/data': 0.032102957367897034, 'timing/model': 1.5090086925774813, 'learning_rate/qwen_vl_interface': 3.6696815372321603e-06, 'learning_rate/action_model': 3.6696815372321604e-05, 'epoch': 1.19}) +[2026-06-16 03:55:48,244][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.13998740911483765, 'train/grad_norm_pre_clip': 8.811607360839844, 'timing/data': 0.03184656426310539, 'timing/model': 1.4868001751601696, 'learning_rate/qwen_vl_interface': 3.6668099286830023e-06, 'learning_rate/action_model': 3.666809928683002e-05, 'epoch': 1.19}) +[2026-06-16 03:55:54,486][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.02368786558508873, 'train/grad_norm_pre_clip': 5.849206447601318, 'timing/data': 0.03245631232857704, 'timing/model': 1.7562974598258734, 'learning_rate/qwen_vl_interface': 3.66393897092348e-06, 'learning_rate/action_model': 3.66393897092348e-05, 'epoch': 1.19}) +[2026-06-16 03:56:00,624][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.005413229111582041, 'train/grad_norm_pre_clip': 0.44736236333847046, 'timing/data': 0.03310029208660126, 'timing/model': 1.684260094538331, 'learning_rate/qwen_vl_interface': 3.661068665133735e-06, 'learning_rate/action_model': 3.661068665133735e-05, 'epoch': 1.19}) +[2026-06-16 03:56:06,649][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.00543673662468791, 'train/grad_norm_pre_clip': 1.0340673923492432, 'timing/data': 0.03592943958938122, 'timing/model': 1.5078066755086184, 'learning_rate/qwen_vl_interface': 3.6581990124936434e-06, 'learning_rate/action_model': 3.6581990124936434e-05, 'epoch': 1.19}) +[2026-06-16 03:56:12,755][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.13313093781471252, 'train/grad_norm_pre_clip': 8.914984703063965, 'timing/data': 0.03271756134927273, 'timing/model': 1.544109532609582, 'learning_rate/qwen_vl_interface': 3.65533001418281e-06, 'learning_rate/action_model': 3.655330014182809e-05, 'epoch': 1.19}) +[2026-06-16 03:56:19,061][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.11073853820562363, 'train/grad_norm_pre_clip': 16.46811866760254, 'timing/data': 0.033416250720620155, 'timing/model': 1.531106200069189, 'learning_rate/qwen_vl_interface': 3.652461671380572e-06, 'learning_rate/action_model': 3.652461671380572e-05, 'epoch': 1.19}) +[2026-06-16 03:56:25,251][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.17009815573692322, 'train/grad_norm_pre_clip': 16.77463150024414, 'timing/data': 0.03591969609260559, 'timing/model': 1.5148340836167336, 'learning_rate/qwen_vl_interface': 3.6495939852659966e-06, 'learning_rate/action_model': 3.649593985265997e-05, 'epoch': 1.19}) +[2026-06-16 03:56:31,680][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.198850616812706, 'train/grad_norm_pre_clip': 6.4446258544921875, 'timing/data': 0.030809802934527397, 'timing/model': 1.8621513862162828, 'learning_rate/qwen_vl_interface': 3.646726957017881e-06, 'learning_rate/action_model': 3.646726957017881e-05, 'epoch': 1.19}) +[2026-06-16 03:56:38,089][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.0327911451458931, 'train/grad_norm_pre_clip': 9.566492080688477, 'timing/data': 0.03206147067248821, 'timing/model': 1.7366819325834513, 'learning_rate/qwen_vl_interface': 3.643860587814752e-06, 'learning_rate/action_model': 3.643860587814752e-05, 'epoch': 1.2}) +[2026-06-16 03:56:44,173][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.20769160985946655, 'train/grad_norm_pre_clip': 13.923439025878906, 'timing/data': 0.03620913252234459, 'timing/model': 1.5226255152374506, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 1.2}) +[2026-06-16 03:56:50,488][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.07014327496290207, 'train/grad_norm_pre_clip': 7.314449787139893, 'timing/data': 0.03134072944521904, 'timing/model': 1.5218245722353458, 'learning_rate/qwen_vl_interface': 3.6381298312562095e-06, 'learning_rate/action_model': 3.6381298312562094e-05, 'epoch': 1.2}) +[2026-06-16 03:56:56,806][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.17338749766349792, 'train/grad_norm_pre_clip': 20.6447696685791, 'timing/data': 0.03326501324772835, 'timing/model': 1.5582634750753641, 'learning_rate/qwen_vl_interface': 3.635265446256493e-06, 'learning_rate/action_model': 3.6352654462564926e-05, 'epoch': 1.2}) +[2026-06-16 03:57:03,298][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.032744672149419785, 'train/grad_norm_pre_clip': 5.3050994873046875, 'timing/data': 0.04376590438187122, 'timing/model': 1.6052439641207457, 'learning_rate/qwen_vl_interface': 3.6324017250131565e-06, 'learning_rate/action_model': 3.6324017250131567e-05, 'epoch': 1.2}) +[2026-06-16 03:57:09,913][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.01030389778316021, 'train/grad_norm_pre_clip': 1.4827786684036255, 'timing/data': 0.03970376215875149, 'timing/model': 1.9052504524588585, 'learning_rate/qwen_vl_interface': 3.6295386687033696e-06, 'learning_rate/action_model': 3.6295386687033695e-05, 'epoch': 1.2}) +[2026-06-16 03:57:16,198][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.04852680861949921, 'train/grad_norm_pre_clip': 2.6311073303222656, 'timing/data': 0.03923540562391281, 'timing/model': 1.5940311383455992, 'learning_rate/qwen_vl_interface': 3.6266762785040253e-06, 'learning_rate/action_model': 3.626676278504025e-05, 'epoch': 1.2}) +[2026-06-16 03:57:22,596][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.028832240030169487, 'train/grad_norm_pre_clip': 3.0925240516662598, 'timing/data': 0.032093584537506104, 'timing/model': 1.5299316383898258, 'learning_rate/qwen_vl_interface': 3.6238145555917464e-06, 'learning_rate/action_model': 3.6238145555917465e-05, 'epoch': 1.2}) +[2026-06-16 03:57:28,878][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.18185803294181824, 'train/grad_norm_pre_clip': 11.227531433105469, 'timing/data': 0.031243178993463516, 'timing/model': 1.5184351317584515, 'learning_rate/qwen_vl_interface': 3.6209535011428757e-06, 'learning_rate/action_model': 3.620953501142876e-05, 'epoch': 1.2}) +[2026-06-16 03:57:35,227][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.033659446984529495, 'train/grad_norm_pre_clip': 3.9412741661071777, 'timing/data': 0.02559344470500946, 'timing/model': 1.5210619941353798, 'learning_rate/qwen_vl_interface': 3.6180931163334863e-06, 'learning_rate/action_model': 3.6180931163334863e-05, 'epoch': 1.2}) +[2026-06-16 03:57:41,682][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.07120151817798615, 'train/grad_norm_pre_clip': 6.742323398590088, 'timing/data': 0.03452572040259838, 'timing/model': 1.5666070021688938, 'learning_rate/qwen_vl_interface': 3.6152334023393747e-06, 'learning_rate/action_model': 3.6152334023393745e-05, 'epoch': 1.2}) +[2026-06-16 03:57:48,017][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.009582326747477055, 'train/grad_norm_pre_clip': 2.114417314529419, 'timing/data': 0.03707159869372845, 'timing/model': 1.7629795800894499, 'learning_rate/qwen_vl_interface': 3.6123743603360613e-06, 'learning_rate/action_model': 3.612374360336061e-05, 'epoch': 1.2}) +[2026-06-16 04:05:10,140][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.014492454007267952, 'train/grad_norm_pre_clip': 4.4261651039123535, 'eval/loss': 0.05541456490755081, 'eval/flap_precision': 0.9561042524005487, 'eval/flap_recall': 0.9457259158751696, 'eval/flap_f1': 0.9508867667121419, 'eval/flap_rate_model': 0.2025, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.950886766712142, 'timing/data': 0.038078151643276215, 'timing/model': 1.5708406381309032, 'learning_rate/qwen_vl_interface': 3.6095159914987893e-06, 'learning_rate/action_model': 3.6095159914987894e-05, 'epoch': 1.2}) +[2026-06-16 04:05:10,141][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3100_state +[2026-06-16 04:05:10,156][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 04:05:16,120][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3100_state/model.safetensors +[2026-06-16 04:05:26,315][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3100_state/optimizer.bin +[2026-06-16 04:05:26,315][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3100_state/sampler.bin +[2026-06-16 04:05:26,316][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3100_state/sampler_1.bin +[2026-06-16 04:05:26,317][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3100_state/random_states_0.pkl +[2026-06-16 04:05:26,467][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3100 +[2026-06-16 04:05:26,467][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 04:05:26,475][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 04:05:32,997][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.02149716578423977, 'train/grad_norm_pre_clip': 3.7925446033477783, 'timing/data': 0.02856414020061493, 'timing/model': 1.8975882828235626, 'learning_rate/qwen_vl_interface': 3.6066582970025266e-06, 'learning_rate/action_model': 3.6066582970025264e-05, 'epoch': 1.2}) +[2026-06-16 04:05:38,700][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.075717031955719, 'train/grad_norm_pre_clip': 5.288224220275879, 'timing/data': 0.027151891961693764, 'timing/model': 1.4361137431114912, 'learning_rate/qwen_vl_interface': 3.603801278021964e-06, 'learning_rate/action_model': 3.6038012780219635e-05, 'epoch': 1.2}) +[2026-06-16 04:05:44,739][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.005503915250301361, 'train/grad_norm_pre_clip': 0.8102797865867615, 'timing/data': 0.03199492581188679, 'timing/model': 1.5353320110589266, 'learning_rate/qwen_vl_interface': 3.6009449357315122e-06, 'learning_rate/action_model': 3.6009449357315125e-05, 'epoch': 1.2}) +[2026-06-16 04:05:51,289][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.21102814376354218, 'train/grad_norm_pre_clip': 15.634954452514648, 'timing/data': 0.046379564329981804, 'timing/model': 1.6796512734144926, 'learning_rate/qwen_vl_interface': 3.5980892713053084e-06, 'learning_rate/action_model': 3.598089271305308e-05, 'epoch': 1.2}) +[2026-06-16 04:05:58,169][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.29143986105918884, 'train/grad_norm_pre_clip': 25.616703033447266, 'timing/data': 0.044813117012381554, 'timing/model': 1.647510414943099, 'learning_rate/qwen_vl_interface': 3.5952342859172063e-06, 'learning_rate/action_model': 3.595234285917206e-05, 'epoch': 1.2}) +[2026-06-16 04:06:05,095][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.009193590842187405, 'train/grad_norm_pre_clip': 2.3915627002716064, 'timing/data': 0.04331100545823574, 'timing/model': 2.11360433883965, 'learning_rate/qwen_vl_interface': 3.592379980740785e-06, 'learning_rate/action_model': 3.5923799807407845e-05, 'epoch': 1.2}) +[2026-06-16 04:06:11,570][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.015194795094430447, 'train/grad_norm_pre_clip': 6.445866584777832, 'timing/data': 0.04206457361578941, 'timing/model': 1.6826728824526072, 'learning_rate/qwen_vl_interface': 3.5895263569493384e-06, 'learning_rate/action_model': 3.589526356949338e-05, 'epoch': 1.2}) +[2026-06-16 04:06:18,346][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.025926697999238968, 'train/grad_norm_pre_clip': 5.26963996887207, 'timing/data': 0.04382457956671715, 'timing/model': 1.6623538993299007, 'learning_rate/qwen_vl_interface': 3.586673415715885e-06, 'learning_rate/action_model': 3.586673415715885e-05, 'epoch': 1.2}) +[2026-06-16 04:06:25,019][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.06562395393848419, 'train/grad_norm_pre_clip': 7.240921497344971, 'timing/data': 0.045383889228105545, 'timing/model': 1.71189708635211, 'learning_rate/qwen_vl_interface': 3.5838211582131603e-06, 'learning_rate/action_model': 3.58382115821316e-05, 'epoch': 1.2}) +[2026-06-16 04:06:31,339][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.07203791290521622, 'train/grad_norm_pre_clip': 8.246932983398438, 'timing/data': 0.030661538243293762, 'timing/model': 1.5180256515741348, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 1.2}) +[2026-06-16 04:06:37,436][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.16696424782276154, 'train/grad_norm_pre_clip': 12.86583423614502, 'timing/data': 0.03254633955657482, 'timing/model': 1.4866667203605175, 'learning_rate/qwen_vl_interface': 3.5781186990894423e-06, 'learning_rate/action_model': 3.578118699089442e-05, 'epoch': 1.2}) +[2026-06-16 04:06:43,585][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.037138499319553375, 'train/grad_norm_pre_clip': 3.526623010635376, 'timing/data': 0.03225676529109478, 'timing/model': 1.7701227627694607, 'learning_rate/qwen_vl_interface': 3.5752684998125113e-06, 'learning_rate/action_model': 3.5752684998125116e-05, 'epoch': 1.2}) +[2026-06-16 04:06:49,534][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.08931291848421097, 'train/grad_norm_pre_clip': 16.97119140625, 'timing/data': 0.03230137564241886, 'timing/model': 1.4905490763485432, 'learning_rate/qwen_vl_interface': 3.5724189889544407e-06, 'learning_rate/action_model': 3.572418988954441e-05, 'epoch': 1.2}) +[2026-06-16 04:06:55,633][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.09960697591304779, 'train/grad_norm_pre_clip': 5.642821788787842, 'timing/data': 0.032015759497880936, 'timing/model': 1.4787502828985453, 'learning_rate/qwen_vl_interface': 3.569570167686556e-06, 'learning_rate/action_model': 3.569570167686556e-05, 'epoch': 1.21}) +[2026-06-16 04:07:01,835][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.2580947279930115, 'train/grad_norm_pre_clip': 13.60142993927002, 'timing/data': 0.0321191493421793, 'timing/model': 1.5199936293065548, 'learning_rate/qwen_vl_interface': 3.5667220371798995e-06, 'learning_rate/action_model': 3.566722037179899e-05, 'epoch': 1.21}) +[2026-06-16 04:07:07,834][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.03587862476706505, 'train/grad_norm_pre_clip': 7.3920087814331055, 'timing/data': 0.03127649985253811, 'timing/model': 1.536452405154705, 'learning_rate/qwen_vl_interface': 3.563874598605232e-06, 'learning_rate/action_model': 3.5638745986052316e-05, 'epoch': 1.21}) +[2026-06-16 04:07:13,984][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.004662771709263325, 'train/grad_norm_pre_clip': 0.5770587921142578, 'timing/data': 0.03251747041940689, 'timing/model': 1.5799475349485874, 'learning_rate/qwen_vl_interface': 3.561027853133025e-06, 'learning_rate/action_model': 3.561027853133025e-05, 'epoch': 1.21}) +[2026-06-16 04:07:20,176][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.024053655564785004, 'train/grad_norm_pre_clip': 5.1052680015563965, 'timing/data': 0.030999794602394104, 'timing/model': 1.7737532872706652, 'learning_rate/qwen_vl_interface': 3.558181801933472e-06, 'learning_rate/action_model': 3.5581818019334714e-05, 'epoch': 1.21}) +[2026-06-16 04:07:26,437][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.008117740973830223, 'train/grad_norm_pre_clip': 0.8306543827056885, 'timing/data': 0.031224017962813377, 'timing/model': 1.5358991511166096, 'learning_rate/qwen_vl_interface': 3.555336446176473e-06, 'learning_rate/action_model': 3.5553364461764726e-05, 'epoch': 1.21}) +[2026-06-16 04:07:32,640][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.015092332847416401, 'train/grad_norm_pre_clip': 1.2868925333023071, 'timing/data': 0.032312335446476936, 'timing/model': 1.55128070153296, 'learning_rate/qwen_vl_interface': 3.552491787031647e-06, 'learning_rate/action_model': 3.552491787031647e-05, 'epoch': 1.21}) +[2026-06-16 04:07:38,738][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.13564804196357727, 'train/grad_norm_pre_clip': 13.299341201782227, 'timing/data': 0.03164568170905113, 'timing/model': 1.5284581873565912, 'learning_rate/qwen_vl_interface': 3.5496478256683277e-06, 'learning_rate/action_model': 3.549647825668327e-05, 'epoch': 1.21}) +[2026-06-16 04:07:44,836][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.03358836472034454, 'train/grad_norm_pre_clip': 9.928215980529785, 'timing/data': 0.03131779097020626, 'timing/model': 1.4910675846040249, 'learning_rate/qwen_vl_interface': 3.5468045632555585e-06, 'learning_rate/action_model': 3.546804563255559e-05, 'epoch': 1.21}) +[2026-06-16 04:07:50,840][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.09172443300485611, 'train/grad_norm_pre_clip': 24.490867614746094, 'timing/data': 0.032493866980075836, 'timing/model': 1.5076328162103891, 'learning_rate/qwen_vl_interface': 3.5439620009620995e-06, 'learning_rate/action_model': 3.543962000962099e-05, 'epoch': 1.21}) +[2026-06-16 04:07:56,978][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.09324873238801956, 'train/grad_norm_pre_clip': 4.923399448394775, 'timing/data': 0.032702984288334846, 'timing/model': 1.7655920833349228, 'learning_rate/qwen_vl_interface': 3.541120139956419e-06, 'learning_rate/action_model': 3.5411201399564184e-05, 'epoch': 1.21}) +[2026-06-16 04:08:02,939][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.18124692142009735, 'train/grad_norm_pre_clip': 19.559282302856445, 'timing/data': 0.030503220856189728, 'timing/model': 1.5333767589181662, 'learning_rate/qwen_vl_interface': 3.538278981406697e-06, 'learning_rate/action_model': 3.5382789814066966e-05, 'epoch': 1.21}) +[2026-06-16 04:08:09,141][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.005816295742988586, 'train/grad_norm_pre_clip': 0.6962618827819824, 'timing/data': 0.031207909807562828, 'timing/model': 1.5395658761262894, 'learning_rate/qwen_vl_interface': 3.5354385264808295e-06, 'learning_rate/action_model': 3.535438526480829e-05, 'epoch': 1.21}) +[2026-06-16 04:08:15,240][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.044895101338624954, 'train/grad_norm_pre_clip': 9.984186172485352, 'timing/data': 0.0328308567404747, 'timing/model': 1.5288032293319702, 'learning_rate/qwen_vl_interface': 3.5325987763464196e-06, 'learning_rate/action_model': 3.5325987763464194e-05, 'epoch': 1.21}) +[2026-06-16 04:08:21,140][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.11442192643880844, 'train/grad_norm_pre_clip': 13.5549955368042, 'timing/data': 0.0320295337587595, 'timing/model': 1.522653417661786, 'learning_rate/qwen_vl_interface': 3.529759732170781e-06, 'learning_rate/action_model': 3.529759732170781e-05, 'epoch': 1.21}) +[2026-06-16 04:08:27,353][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.0914502739906311, 'train/grad_norm_pre_clip': 24.032474517822266, 'timing/data': 0.03172946348786354, 'timing/model': 1.5233565978705883, 'learning_rate/qwen_vl_interface': 3.5269213951209375e-06, 'learning_rate/action_model': 3.5269213951209374e-05, 'epoch': 1.21}) +[2026-06-16 04:08:33,483][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.038937538862228394, 'train/grad_norm_pre_clip': 9.050675392150879, 'timing/data': 0.032608602195978165, 'timing/model': 1.7768560573458672, 'learning_rate/qwen_vl_interface': 3.5240837663636207e-06, 'learning_rate/action_model': 3.5240837663636204e-05, 'epoch': 1.21}) +[2026-06-16 04:08:39,445][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.05248488485813141, 'train/grad_norm_pre_clip': 13.186357498168945, 'timing/data': 0.031871497631073, 'timing/model': 1.5394436959177256, 'learning_rate/qwen_vl_interface': 3.5212468470652745e-06, 'learning_rate/action_model': 3.521246847065274e-05, 'epoch': 1.21}) +[2026-06-16 04:08:45,742][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.03776414319872856, 'train/grad_norm_pre_clip': 7.025203227996826, 'timing/data': 0.03094720095396042, 'timing/model': 1.5341881774365902, 'learning_rate/qwen_vl_interface': 3.5184106383920487e-06, 'learning_rate/action_model': 3.5184106383920486e-05, 'epoch': 1.21}) +[2026-06-16 04:08:51,961][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.18148468434810638, 'train/grad_norm_pre_clip': 20.400894165039062, 'timing/data': 0.031233297660946846, 'timing/model': 1.5488484427332878, 'learning_rate/qwen_vl_interface': 3.5155751415098022e-06, 'learning_rate/action_model': 3.515575141509802e-05, 'epoch': 1.21}) +[2026-06-16 04:08:58,053][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.0009489959920756519, 'train/grad_norm_pre_clip': 0.18393324315547943, 'timing/data': 0.03086860477924347, 'timing/model': 1.5496536530554295, 'learning_rate/qwen_vl_interface': 3.5127403575841e-06, 'learning_rate/action_model': 3.5127403575841e-05, 'epoch': 1.21}) +[2026-06-16 04:09:03,988][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.07865197956562042, 'train/grad_norm_pre_clip': 7.794933319091797, 'timing/data': 0.025702236220240593, 'timing/model': 1.4312750957906246, 'learning_rate/qwen_vl_interface': 3.509906287780215e-06, 'learning_rate/action_model': 3.5099062877802144e-05, 'epoch': 1.21}) +[2026-06-16 04:09:09,903][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.11555591970682144, 'train/grad_norm_pre_clip': 11.979804992675781, 'timing/data': 0.026529228314757347, 'timing/model': 1.6385561246424913, 'learning_rate/qwen_vl_interface': 3.5070729332631247e-06, 'learning_rate/action_model': 3.507072933263125e-05, 'epoch': 1.21}) +[2026-06-16 04:09:15,589][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.0578213706612587, 'train/grad_norm_pre_clip': 11.51716136932373, 'timing/data': 0.026213843375444412, 'timing/model': 1.4403453804552555, 'learning_rate/qwen_vl_interface': 3.5042402951975167e-06, 'learning_rate/action_model': 3.5042402951975167e-05, 'epoch': 1.21}) +[2026-06-16 04:09:21,485][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.21946127712726593, 'train/grad_norm_pre_clip': 18.60100746154785, 'timing/data': 0.026263337582349777, 'timing/model': 1.4356726314872503, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 1.21}) +[2026-06-16 04:09:27,385][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.21170523762702942, 'train/grad_norm_pre_clip': 29.750080108642578, 'timing/data': 0.02570376731455326, 'timing/model': 1.438332935795188, 'learning_rate/qwen_vl_interface': 3.498577173078011e-06, 'learning_rate/action_model': 3.498577173078011e-05, 'epoch': 1.21}) +[2026-06-16 04:09:33,187][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.047811359167099, 'train/grad_norm_pre_clip': 17.933156967163086, 'timing/data': 0.026038728654384613, 'timing/model': 1.4406354110687971, 'learning_rate/qwen_vl_interface': 3.495746691352008e-06, 'learning_rate/action_model': 3.495746691352008e-05, 'epoch': 1.22}) +[2026-06-16 04:09:39,085][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.03733096271753311, 'train/grad_norm_pre_clip': 6.47805643081665, 'timing/data': 0.026113195344805717, 'timing/model': 1.4286190252751112, 'learning_rate/qwen_vl_interface': 3.4929169307332763e-06, 'learning_rate/action_model': 3.492916930733276e-05, 'epoch': 1.22}) +[2026-06-16 04:09:44,966][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.030957747250795364, 'train/grad_norm_pre_clip': 3.8326961994171143, 'timing/data': 0.025721171870827675, 'timing/model': 1.6188951283693314, 'learning_rate/qwen_vl_interface': 3.4900878923850234e-06, 'learning_rate/action_model': 3.490087892385023e-05, 'epoch': 1.22}) +[2026-06-16 04:09:50,590][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.21192476153373718, 'train/grad_norm_pre_clip': 30.318456649780273, 'timing/data': 0.02591353841125965, 'timing/model': 1.439974769949913, 'learning_rate/qwen_vl_interface': 3.4872595774701607e-06, 'learning_rate/action_model': 3.487259577470161e-05, 'epoch': 1.22}) +[2026-06-16 04:09:56,481][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.027112938463687897, 'train/grad_norm_pre_clip': 4.999779224395752, 'timing/data': 0.02674916945397854, 'timing/model': 1.4353925827890635, 'learning_rate/qwen_vl_interface': 3.4844319871513014e-06, 'learning_rate/action_model': 3.484431987151301e-05, 'epoch': 1.22}) +[2026-06-16 04:10:02,287][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.37902238965034485, 'train/grad_norm_pre_clip': 25.05327033996582, 'timing/data': 0.02551225572824478, 'timing/model': 1.4378450587391853, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 1.22}) +[2026-06-16 04:10:08,189][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.06187814101576805, 'train/grad_norm_pre_clip': 3.525498151779175, 'timing/data': 0.025927938520908356, 'timing/model': 1.4406400248408318, 'learning_rate/qwen_vl_interface': 3.478778984950556e-06, 'learning_rate/action_model': 3.478778984950556e-05, 'epoch': 1.22}) +[2026-06-16 04:10:14,087][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.009202997200191021, 'train/grad_norm_pre_clip': 4.459671497344971, 'timing/data': 0.025874702259898186, 'timing/model': 1.4323674272745848, 'learning_rate/qwen_vl_interface': 3.475953575392405e-06, 'learning_rate/action_model': 3.475953575392405e-05, 'epoch': 1.22}) +[2026-06-16 04:10:19,970][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.03263125568628311, 'train/grad_norm_pre_clip': 4.220536231994629, 'timing/data': 0.02711046300828457, 'timing/model': 1.6205207295715809, 'learning_rate/qwen_vl_interface': 3.473128895077728e-06, 'learning_rate/action_model': 3.473128895077728e-05, 'epoch': 1.22}) +[2026-06-16 04:10:25,681][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.006127126049250364, 'train/grad_norm_pre_clip': 1.1315133571624756, 'timing/data': 0.02642703615128994, 'timing/model': 1.534879956394434, 'learning_rate/qwen_vl_interface': 3.4703049451676453e-06, 'learning_rate/action_model': 3.4703049451676454e-05, 'epoch': 1.22}) +[2026-06-16 04:10:31,392][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.0019157846691086888, 'train/grad_norm_pre_clip': 0.41555890440940857, 'timing/data': 0.02538045309484005, 'timing/model': 1.4435119722038507, 'learning_rate/qwen_vl_interface': 3.4674817268229732e-06, 'learning_rate/action_model': 3.467481726822973e-05, 'epoch': 1.22}) +[2026-06-16 04:10:37,183][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.014833827503025532, 'train/grad_norm_pre_clip': 1.9890666007995605, 'timing/data': 0.027830341830849648, 'timing/model': 1.4305844381451607, 'learning_rate/qwen_vl_interface': 3.464659241204232e-06, 'learning_rate/action_model': 3.4646592412042314e-05, 'epoch': 1.22}) +[2026-06-16 04:10:43,242][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.006712114904075861, 'train/grad_norm_pre_clip': 1.0633410215377808, 'timing/data': 0.0321621298789978, 'timing/model': 1.556219631806016, 'learning_rate/qwen_vl_interface': 3.4618374894716377e-06, 'learning_rate/action_model': 3.461837489471638e-05, 'epoch': 1.22}) +[2026-06-16 04:10:49,446][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.0531245656311512, 'train/grad_norm_pre_clip': 8.158666610717773, 'timing/data': 0.032657694071531296, 'timing/model': 1.5337661281228065, 'learning_rate/qwen_vl_interface': 3.459016472785107e-06, 'learning_rate/action_model': 3.459016472785107e-05, 'epoch': 1.22}) +[2026-06-16 04:10:55,573][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.14246372878551483, 'train/grad_norm_pre_clip': 13.549519538879395, 'timing/data': 0.031245747581124306, 'timing/model': 1.7531017437577248, 'learning_rate/qwen_vl_interface': 3.456196192304254e-06, 'learning_rate/action_model': 3.456196192304254e-05, 'epoch': 1.22}) +[2026-06-16 04:11:01,712][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.014510013163089752, 'train/grad_norm_pre_clip': 4.182548999786377, 'timing/data': 0.03132680803537369, 'timing/model': 1.674694212153554, 'learning_rate/qwen_vl_interface': 3.4533766491883887e-06, 'learning_rate/action_model': 3.453376649188388e-05, 'epoch': 1.22}) +[2026-06-16 04:11:07,750][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.026280716061592102, 'train/grad_norm_pre_clip': 18.53652000427246, 'timing/data': 0.03158494830131531, 'timing/model': 1.539700549095869, 'learning_rate/qwen_vl_interface': 3.450557844596517e-06, 'learning_rate/action_model': 3.450557844596517e-05, 'epoch': 1.22}) +[2026-06-16 04:11:13,948][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.1316801756620407, 'train/grad_norm_pre_clip': 13.501941680908203, 'timing/data': 0.03144962713122368, 'timing/model': 1.537759143859148, 'learning_rate/qwen_vl_interface': 3.447739779687345e-06, 'learning_rate/action_model': 3.4477397796873446e-05, 'epoch': 1.22}) +[2026-06-16 04:11:20,145][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.044975269585847855, 'train/grad_norm_pre_clip': 11.757503509521484, 'timing/data': 0.031158920377492905, 'timing/model': 1.5322847217321396, 'learning_rate/qwen_vl_interface': 3.4449224556192722e-06, 'learning_rate/action_model': 3.444922455619272e-05, 'epoch': 1.22}) +[2026-06-16 04:11:25,993][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.03923949599266052, 'train/grad_norm_pre_clip': 8.230636596679688, 'timing/data': 0.02594245970249176, 'timing/model': 1.4377015680074692, 'learning_rate/qwen_vl_interface': 3.442105873550395e-06, 'learning_rate/action_model': 3.442105873550395e-05, 'epoch': 1.22}) +[2026-06-16 04:11:31,874][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.2449827790260315, 'train/grad_norm_pre_clip': 10.02928638458252, 'timing/data': 0.026541680097579956, 'timing/model': 1.6205241680145264, 'learning_rate/qwen_vl_interface': 3.4392900346385022e-06, 'learning_rate/action_model': 3.4392900346385024e-05, 'epoch': 1.22}) +[2026-06-16 04:11:37,488][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.00014328054385259748, 'train/grad_norm_pre_clip': 0.014809505082666874, 'timing/data': 0.026165097951889038, 'timing/model': 1.4359289333224297, 'learning_rate/qwen_vl_interface': 3.4364749400410796e-06, 'learning_rate/action_model': 3.43647494004108e-05, 'epoch': 1.22}) +[2026-06-16 04:11:43,385][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.34097352623939514, 'train/grad_norm_pre_clip': 11.90272045135498, 'timing/data': 0.02632226049900055, 'timing/model': 1.436370188370347, 'learning_rate/qwen_vl_interface': 3.4336605909153066e-06, 'learning_rate/action_model': 3.433660590915306e-05, 'epoch': 1.22}) +[2026-06-16 04:11:49,285][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.0022766259498894215, 'train/grad_norm_pre_clip': 0.9471263885498047, 'timing/data': 0.025578264147043228, 'timing/model': 1.4407119657844305, 'learning_rate/qwen_vl_interface': 3.430846988418057e-06, 'learning_rate/action_model': 3.430846988418057e-05, 'epoch': 1.22}) +[2026-06-16 04:11:55,187][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.01796751283109188, 'train/grad_norm_pre_clip': 3.9635050296783447, 'timing/data': 0.025951609015464783, 'timing/model': 1.453106241300702, 'learning_rate/qwen_vl_interface': 3.428034133705893e-06, 'learning_rate/action_model': 3.428034133705893e-05, 'epoch': 1.22}) +[2026-06-16 04:12:00,986][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.05037674680352211, 'train/grad_norm_pre_clip': 15.328874588012695, 'timing/data': 0.026335805654525757, 'timing/model': 1.4657256435602903, 'learning_rate/qwen_vl_interface': 3.4252220279350763e-06, 'learning_rate/action_model': 3.425222027935076e-05, 'epoch': 1.22}) +[2026-06-16 04:12:06,912][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.05458410084247589, 'train/grad_norm_pre_clip': 6.241765975952148, 'timing/data': 0.028312932699918747, 'timing/model': 1.5625099334865808, 'learning_rate/qwen_vl_interface': 3.4224106722615547e-06, 'learning_rate/action_model': 3.422410672261554e-05, 'epoch': 1.23}) +[2026-06-16 04:12:12,583][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.1746886968612671, 'train/grad_norm_pre_clip': 16.202852249145508, 'timing/data': 0.025262203067541122, 'timing/model': 1.4329697005450726, 'learning_rate/qwen_vl_interface': 3.419600067840973e-06, 'learning_rate/action_model': 3.419600067840973e-05, 'epoch': 1.23}) +[2026-06-16 04:12:18,488][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.022553782910108566, 'train/grad_norm_pre_clip': 6.899263381958008, 'timing/data': 0.026004984974861145, 'timing/model': 1.444557623937726, 'learning_rate/qwen_vl_interface': 3.4167902158286627e-06, 'learning_rate/action_model': 3.4167902158286626e-05, 'epoch': 1.23}) +[2026-06-16 04:12:24,381][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.03195882588624954, 'train/grad_norm_pre_clip': 7.5596442222595215, 'timing/data': 0.02673378214240074, 'timing/model': 1.4377954117953777, 'learning_rate/qwen_vl_interface': 3.4139811173796468e-06, 'learning_rate/action_model': 3.413981117379647e-05, 'epoch': 1.23}) +[2026-06-16 04:12:30,284][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.027712959796190262, 'train/grad_norm_pre_clip': 13.522468566894531, 'timing/data': 0.025506827980279922, 'timing/model': 1.4555196817964315, 'learning_rate/qwen_vl_interface': 3.411172773648642e-06, 'learning_rate/action_model': 3.411172773648642e-05, 'epoch': 1.23}) +[2026-06-16 04:12:36,187][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.02334793657064438, 'train/grad_norm_pre_clip': 6.651714324951172, 'timing/data': 0.026494856923818588, 'timing/model': 1.4584368411451578, 'learning_rate/qwen_vl_interface': 3.408365185790052e-06, 'learning_rate/action_model': 3.4083651857900515e-05, 'epoch': 1.23}) +[2026-06-16 04:12:42,438][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.0642508864402771, 'train/grad_norm_pre_clip': 7.53277587890625, 'timing/data': 0.0327193196862936, 'timing/model': 1.7235177494585514, 'learning_rate/qwen_vl_interface': 3.40555835495797e-06, 'learning_rate/action_model': 3.40555835495797e-05, 'epoch': 1.23}) +[2026-06-16 04:12:48,458][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.4074348211288452, 'train/grad_norm_pre_clip': 10.542556762695312, 'timing/data': 0.03142750635743141, 'timing/model': 1.5648760236799717, 'learning_rate/qwen_vl_interface': 3.4027522823061763e-06, 'learning_rate/action_model': 3.4027522823061765e-05, 'epoch': 1.23}) +[2026-06-16 04:12:54,644][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.07003434747457504, 'train/grad_norm_pre_clip': 7.300660610198975, 'timing/data': 0.031245680525898933, 'timing/model': 1.5341102164238691, 'learning_rate/qwen_vl_interface': 3.399946968988143e-06, 'learning_rate/action_model': 3.399946968988143e-05, 'epoch': 1.23}) +[2026-06-16 04:13:00,841][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.33688119053840637, 'train/grad_norm_pre_clip': 17.403812408447266, 'timing/data': 0.032039858400821686, 'timing/model': 1.53662683442235, 'learning_rate/qwen_vl_interface': 3.397142416157028e-06, 'learning_rate/action_model': 3.397142416157028e-05, 'epoch': 1.23}) +[2026-06-16 04:13:06,971][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.08061768859624863, 'train/grad_norm_pre_clip': 22.568511962890625, 'timing/data': 0.031162144616246223, 'timing/model': 1.5039785038679838, 'learning_rate/qwen_vl_interface': 3.3943386249656766e-06, 'learning_rate/action_model': 3.394338624965677e-05, 'epoch': 1.23}) +[2026-06-16 04:13:13,172][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.00020812664297409356, 'train/grad_norm_pre_clip': 0.06908179819583893, 'timing/data': 0.032026562839746475, 'timing/model': 1.7636972721666098, 'learning_rate/qwen_vl_interface': 3.391535596566624e-06, 'learning_rate/action_model': 3.391535596566624e-05, 'epoch': 1.23}) +[2026-06-16 04:13:18,890][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.11372663080692291, 'train/grad_norm_pre_clip': 27.19202995300293, 'timing/data': 0.024954454973340034, 'timing/model': 1.443302622064948, 'learning_rate/qwen_vl_interface': 3.388733332112086e-06, 'learning_rate/action_model': 3.3887333321120856e-05, 'epoch': 1.23}) +[2026-06-16 04:13:24,787][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.017057349905371666, 'train/grad_norm_pre_clip': 8.0634126663208, 'timing/data': 0.026807574555277824, 'timing/model': 1.4399052038788795, 'learning_rate/qwen_vl_interface': 3.3859318327539693e-06, 'learning_rate/action_model': 3.385931832753969e-05, 'epoch': 1.23}) +[2026-06-16 04:13:30,585][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.004321633838117123, 'train/grad_norm_pre_clip': 1.0502967834472656, 'timing/data': 0.025449661538004875, 'timing/model': 1.4386318139731884, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 1.23}) +[2026-06-16 04:13:36,385][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.020037123933434486, 'train/grad_norm_pre_clip': 7.47617244720459, 'timing/data': 0.0279280673712492, 'timing/model': 1.440445076674223, 'learning_rate/qwen_vl_interface': 3.3803311339330447e-06, 'learning_rate/action_model': 3.3803311339330444e-05, 'epoch': 1.23}) +[2026-06-16 04:13:42,287][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.008730277419090271, 'train/grad_norm_pre_clip': 2.616443157196045, 'timing/data': 0.027170859277248383, 'timing/model': 1.4750445410609245, 'learning_rate/qwen_vl_interface': 3.3775319367724746e-06, 'learning_rate/action_model': 3.3775319367724745e-05, 'epoch': 1.23}) +[2026-06-16 04:13:48,172][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.006816357374191284, 'train/grad_norm_pre_clip': 2.75765323638916, 'timing/data': 0.027877846732735634, 'timing/model': 1.612112134695053, 'learning_rate/qwen_vl_interface': 3.374733509312795e-06, 'learning_rate/action_model': 3.374733509312795e-05, 'epoch': 1.23}) +[2026-06-16 04:13:53,786][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.4138629734516144, 'train/grad_norm_pre_clip': 16.496850967407227, 'timing/data': 0.026061762124300003, 'timing/model': 1.434453522786498, 'learning_rate/qwen_vl_interface': 3.3719358527043345e-06, 'learning_rate/action_model': 3.371935852704334e-05, 'epoch': 1.23}) +[2026-06-16 04:13:59,682][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.0138363903388381, 'train/grad_norm_pre_clip': 6.848586559295654, 'timing/data': 0.02505127154290676, 'timing/model': 1.4239947590976954, 'learning_rate/qwen_vl_interface': 3.3691389680971047e-06, 'learning_rate/action_model': 3.3691389680971044e-05, 'epoch': 1.23}) +[2026-06-16 04:14:05,815][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.13663631677627563, 'train/grad_norm_pre_clip': 10.621376037597656, 'timing/data': 0.03753977455198765, 'timing/model': 1.6074829027056694, 'learning_rate/qwen_vl_interface': 3.3663428566407985e-06, 'learning_rate/action_model': 3.366342856640798e-05, 'epoch': 1.23}) +[2026-06-16 04:14:12,353][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.08328168094158173, 'train/grad_norm_pre_clip': 23.966045379638672, 'timing/data': 0.04409376345574856, 'timing/model': 1.6367253959178925, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 1.23}) +[2026-06-16 04:14:19,012][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.22778943181037903, 'train/grad_norm_pre_clip': 14.913077354431152, 'timing/data': 0.044302478432655334, 'timing/model': 1.6522126160562038, 'learning_rate/qwen_vl_interface': 3.3607529577781433e-06, 'learning_rate/action_model': 3.360752957778143e-05, 'epoch': 1.23}) +[2026-06-16 04:14:25,781][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.41847923398017883, 'train/grad_norm_pre_clip': 22.578922271728516, 'timing/data': 0.046032972633838654, 'timing/model': 1.9646254759281874, 'learning_rate/qwen_vl_interface': 3.3579591726695906e-06, 'learning_rate/action_model': 3.35795917266959e-05, 'epoch': 1.23}) +[2026-06-16 04:14:32,308][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.004791079554706812, 'train/grad_norm_pre_clip': 0.9332703948020935, 'timing/data': 0.04545826464891434, 'timing/model': 1.7011534366756678, 'learning_rate/qwen_vl_interface': 3.3551661653075522e-06, 'learning_rate/action_model': 3.355166165307552e-05, 'epoch': 1.23}) +[2026-06-16 04:14:39,032][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.004462590906769037, 'train/grad_norm_pre_clip': 1.134727120399475, 'timing/data': 0.04089261591434479, 'timing/model': 1.645096193999052, 'learning_rate/qwen_vl_interface': 3.35237393684013e-06, 'learning_rate/action_model': 3.35237393684013e-05, 'epoch': 1.23}) +[2026-06-16 04:14:45,238][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.39077329635620117, 'train/grad_norm_pre_clip': 25.485055923461914, 'timing/data': 0.031188417226076126, 'timing/model': 1.5367197133600712, 'learning_rate/qwen_vl_interface': 3.349582488415103e-06, 'learning_rate/action_model': 3.3495824884151026e-05, 'epoch': 1.24}) +[2026-06-16 04:14:51,183][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.008831370621919632, 'train/grad_norm_pre_clip': 2.1819331645965576, 'timing/data': 0.024720927700400352, 'timing/model': 1.4549808278679848, 'learning_rate/qwen_vl_interface': 3.3467918211799294e-06, 'learning_rate/action_model': 3.3467918211799295e-05, 'epoch': 1.24}) +[2026-06-16 04:14:57,087][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.03611501306295395, 'train/grad_norm_pre_clip': 10.156937599182129, 'timing/data': 0.02653033472597599, 'timing/model': 1.4663700107485056, 'learning_rate/qwen_vl_interface': 3.3440019362817476e-06, 'learning_rate/action_model': 3.344001936281747e-05, 'epoch': 1.24}) +[2026-06-16 04:15:02,979][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.16971631348133087, 'train/grad_norm_pre_clip': 18.68195915222168, 'timing/data': 0.026493912562727928, 'timing/model': 1.6313617136329412, 'learning_rate/qwen_vl_interface': 3.3412128348673745e-06, 'learning_rate/action_model': 3.341212834867374e-05, 'epoch': 1.24}) +[2026-06-16 04:15:08,686][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.008792636916041374, 'train/grad_norm_pre_clip': 5.520846843719482, 'timing/data': 0.027944842353463173, 'timing/model': 1.4370978623628616, 'learning_rate/qwen_vl_interface': 3.338424518083303e-06, 'learning_rate/action_model': 3.338424518083303e-05, 'epoch': 1.24}) +[2026-06-16 04:15:14,837][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.04215870797634125, 'train/grad_norm_pre_clip': 9.431024551391602, 'timing/data': 0.03330588527023792, 'timing/model': 1.474735924974084, 'learning_rate/qwen_vl_interface': 3.3356369870757056e-06, 'learning_rate/action_model': 3.335636987075705e-05, 'epoch': 1.24}) +[2026-06-16 04:15:20,796][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.08368393778800964, 'train/grad_norm_pre_clip': 6.080990791320801, 'timing/data': 0.025886686518788338, 'timing/model': 1.4437635261565447, 'learning_rate/qwen_vl_interface': 3.3328502429904327e-06, 'learning_rate/action_model': 3.3328502429904325e-05, 'epoch': 1.24}) +[2026-06-16 04:15:26,589][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.007449513301253319, 'train/grad_norm_pre_clip': 0.7355284690856934, 'timing/data': 0.025138815864920616, 'timing/model': 1.4234769884496927, 'learning_rate/qwen_vl_interface': 3.3300642869730066e-06, 'learning_rate/action_model': 3.330064286973006e-05, 'epoch': 1.24}) +[2026-06-16 04:24:04,642][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.1194840595126152, 'train/grad_norm_pre_clip': 19.287643432617188, 'eval/loss': 0.05262741446495056, 'eval/flap_precision': 0.9314359637774903, 'eval/flap_recall': 0.9769335142469471, 'eval/flap_f1': 0.9536423841059603, 'eval/flap_rate_model': 0.21472222222222223, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9536423841059603, 'timing/data': 0.02554345317184925, 'timing/model': 1.4267631527036428, 'learning_rate/qwen_vl_interface': 3.3272791201686294e-06, 'learning_rate/action_model': 3.327279120168629e-05, 'epoch': 1.24}) +[2026-06-16 04:24:04,644][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3200_state +[2026-06-16 04:24:04,651][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 04:24:10,683][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3200_state/model.safetensors +[2026-06-16 04:24:20,091][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3200_state/optimizer.bin +[2026-06-16 04:24:20,091][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3200_state/sampler.bin +[2026-06-16 04:24:20,091][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3200_state/sampler_1.bin +[2026-06-16 04:24:20,092][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3200_state/random_states_0.pkl +[2026-06-16 04:24:20,196][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3200 +[2026-06-16 04:24:20,196][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 04:24:20,225][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 04:24:26,059][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.006894370075315237, 'train/grad_norm_pre_clip': 0.6081625819206238, 'timing/data': 0.027023877948522568, 'timing/model': 1.424297409132123, 'learning_rate/qwen_vl_interface': 3.324494743722179e-06, 'learning_rate/action_model': 3.3244947437221785e-05, 'epoch': 1.24}) +[2026-06-16 04:24:31,862][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.16823738813400269, 'train/grad_norm_pre_clip': 7.546449184417725, 'timing/data': 0.02644941583275795, 'timing/model': 1.4752859883010387, 'learning_rate/qwen_vl_interface': 3.321711158778205e-06, 'learning_rate/action_model': 3.321711158778205e-05, 'epoch': 1.24}) +[2026-06-16 04:24:37,656][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.0318586491048336, 'train/grad_norm_pre_clip': 6.377735614776611, 'timing/data': 0.02608882635831833, 'timing/model': 1.6273069847375154, 'learning_rate/qwen_vl_interface': 3.3189283664809393e-06, 'learning_rate/action_model': 3.318928366480939e-05, 'epoch': 1.24}) +[2026-06-16 04:24:43,279][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.0009152798447757959, 'train/grad_norm_pre_clip': 0.08782663941383362, 'timing/data': 0.02717379480600357, 'timing/model': 1.445780634880066, 'learning_rate/qwen_vl_interface': 3.316146367974276e-06, 'learning_rate/action_model': 3.316146367974276e-05, 'epoch': 1.24}) +[2026-06-16 04:24:49,445][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.0405791811645031, 'train/grad_norm_pre_clip': 5.263889312744141, 'timing/data': 0.034777697175741196, 'timing/model': 1.5337007846683264, 'learning_rate/qwen_vl_interface': 3.3133651644017945e-06, 'learning_rate/action_model': 3.3133651644017944e-05, 'epoch': 1.24}) +[2026-06-16 04:24:55,533][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.02112828567624092, 'train/grad_norm_pre_clip': 3.992980718612671, 'timing/data': 0.033805735409259796, 'timing/model': 1.5340115632861853, 'learning_rate/qwen_vl_interface': 3.31058475690674e-06, 'learning_rate/action_model': 3.31058475690674e-05, 'epoch': 1.24}) +[2026-06-16 04:25:01,531][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.21295477449893951, 'train/grad_norm_pre_clip': 10.948009490966797, 'timing/data': 0.03313069976866245, 'timing/model': 1.5360044781118631, 'learning_rate/qwen_vl_interface': 3.3078051466320337e-06, 'learning_rate/action_model': 3.307805146632034e-05, 'epoch': 1.24}) +[2026-06-16 04:25:07,529][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.004237829707562923, 'train/grad_norm_pre_clip': 0.8000000715255737, 'timing/data': 0.026525426656007767, 'timing/model': 1.4838855676352978, 'learning_rate/qwen_vl_interface': 3.3050263347202712e-06, 'learning_rate/action_model': 3.305026334720271e-05, 'epoch': 1.24}) +[2026-06-16 04:25:13,675][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.007208469323813915, 'train/grad_norm_pre_clip': 0.7104856967926025, 'timing/data': 0.03505057655274868, 'timing/model': 1.768019177019596, 'learning_rate/qwen_vl_interface': 3.302248322313712e-06, 'learning_rate/action_model': 3.302248322313712e-05, 'epoch': 1.24}) +[2026-06-16 04:25:19,636][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.16678231954574585, 'train/grad_norm_pre_clip': 23.223894119262695, 'timing/data': 0.03350904583930969, 'timing/model': 1.5503030773252249, 'learning_rate/qwen_vl_interface': 3.299471110554295e-06, 'learning_rate/action_model': 3.299471110554295e-05, 'epoch': 1.24}) +[2026-06-16 04:25:26,222][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.06555987149477005, 'train/grad_norm_pre_clip': 22.56136131286621, 'timing/data': 0.04364383965730667, 'timing/model': 1.6005517970770597, 'learning_rate/qwen_vl_interface': 3.296694700583626e-06, 'learning_rate/action_model': 3.296694700583626e-05, 'epoch': 1.24}) +[2026-06-16 04:25:33,008][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.012974579818546772, 'train/grad_norm_pre_clip': 2.058342695236206, 'timing/data': 0.04714813642203808, 'timing/model': 1.713780913501978, 'learning_rate/qwen_vl_interface': 3.2939190935429833e-06, 'learning_rate/action_model': 3.293919093542983e-05, 'epoch': 1.24}) +[2026-06-16 04:25:39,819][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.11377144604921341, 'train/grad_norm_pre_clip': 13.527292251586914, 'timing/data': 0.05379839055240154, 'timing/model': 1.728326415643096, 'learning_rate/qwen_vl_interface': 3.2911442905733126e-06, 'learning_rate/action_model': 3.291144290573312e-05, 'epoch': 1.24}) +[2026-06-16 04:25:46,403][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.06345921009778976, 'train/grad_norm_pre_clip': 13.268258094787598, 'timing/data': 0.039168111979961395, 'timing/model': 1.5588103365153074, 'learning_rate/qwen_vl_interface': 3.2883702928152323e-06, 'learning_rate/action_model': 3.288370292815232e-05, 'epoch': 1.24}) +[2026-06-16 04:25:53,270][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.4104633033275604, 'train/grad_norm_pre_clip': 26.920621871948242, 'timing/data': 0.047151656821370125, 'timing/model': 1.9657868221402168, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 1.24}) +[2026-06-16 04:25:59,146][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.07457933574914932, 'train/grad_norm_pre_clip': 17.812530517578125, 'timing/data': 0.038373859599232674, 'timing/model': 1.536014599725604, 'learning_rate/qwen_vl_interface': 3.2828247174946528e-06, 'learning_rate/action_model': 3.2828247174946524e-05, 'epoch': 1.24}) +[2026-06-16 04:26:05,359][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.09927273541688919, 'train/grad_norm_pre_clip': 8.735408782958984, 'timing/data': 0.04333032667636871, 'timing/model': 1.5035931952297688, 'learning_rate/qwen_vl_interface': 3.280053142211729e-06, 'learning_rate/action_model': 3.280053142211729e-05, 'epoch': 1.25}) +[2026-06-16 04:26:11,563][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.12196236848831177, 'train/grad_norm_pre_clip': 10.525191307067871, 'timing/data': 0.0602306853979826, 'timing/model': 1.5209700800478458, 'learning_rate/qwen_vl_interface': 3.2772823766995475e-06, 'learning_rate/action_model': 3.277282376699548e-05, 'epoch': 1.25}) +[2026-06-16 04:26:17,488][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.02095964550971985, 'train/grad_norm_pre_clip': 3.1829140186309814, 'timing/data': 0.029673244804143906, 'timing/model': 1.4411966409534216, 'learning_rate/qwen_vl_interface': 3.2745124220970646e-06, 'learning_rate/action_model': 3.274512422097065e-05, 'epoch': 1.25}) +[2026-06-16 04:26:23,400][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.003378773806616664, 'train/grad_norm_pre_clip': 0.42954498529434204, 'timing/data': 0.02632300741970539, 'timing/model': 1.4462785758078098, 'learning_rate/qwen_vl_interface': 3.2717432795429057e-06, 'learning_rate/action_model': 3.271743279542905e-05, 'epoch': 1.25}) +[2026-06-16 04:26:29,284][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.0015045423060655594, 'train/grad_norm_pre_clip': 0.18246237933635712, 'timing/data': 0.028444185853004456, 'timing/model': 1.628473848104477, 'learning_rate/qwen_vl_interface': 3.2689749501753578e-06, 'learning_rate/action_model': 3.2689749501753574e-05, 'epoch': 1.25}) +[2026-06-16 04:26:34,991][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.002282230881974101, 'train/grad_norm_pre_clip': 0.5150952339172363, 'timing/data': 0.029864121228456497, 'timing/model': 1.4443507436662912, 'learning_rate/qwen_vl_interface': 3.2662074351323774e-06, 'learning_rate/action_model': 3.266207435132378e-05, 'epoch': 1.25}) +[2026-06-16 04:26:41,244][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.03982076793909073, 'train/grad_norm_pre_clip': 9.2863130569458, 'timing/data': 0.032324159517884254, 'timing/model': 1.544974660500884, 'learning_rate/qwen_vl_interface': 3.2634407355515875e-06, 'learning_rate/action_model': 3.2634407355515875e-05, 'epoch': 1.25}) +[2026-06-16 04:26:47,340][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.45229217410087585, 'train/grad_norm_pre_clip': 16.482162475585938, 'timing/data': 0.03157172165811062, 'timing/model': 1.5408023111522198, 'learning_rate/qwen_vl_interface': 3.2606748525702715e-06, 'learning_rate/action_model': 3.260674852570271e-05, 'epoch': 1.25}) +[2026-06-16 04:26:53,545][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.006583206355571747, 'train/grad_norm_pre_clip': 1.411716103553772, 'timing/data': 0.03235985338687897, 'timing/model': 1.5397416036576033, 'learning_rate/qwen_vl_interface': 3.2579097873253803e-06, 'learning_rate/action_model': 3.25790978732538e-05, 'epoch': 1.25}) +[2026-06-16 04:26:59,739][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.010603017173707485, 'train/grad_norm_pre_clip': 4.370006084442139, 'timing/data': 0.032408757135272026, 'timing/model': 1.5239354632794857, 'learning_rate/qwen_vl_interface': 3.2551455409535274e-06, 'learning_rate/action_model': 3.255145540953528e-05, 'epoch': 1.25}) +[2026-06-16 04:27:05,954][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.11649370938539505, 'train/grad_norm_pre_clip': 18.73641014099121, 'timing/data': 0.031853215768933296, 'timing/model': 1.7521291002631187, 'learning_rate/qwen_vl_interface': 3.2523821145909907e-06, 'learning_rate/action_model': 3.252382114590991e-05, 'epoch': 1.25}) +[2026-06-16 04:27:11,781][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.0016309164930135012, 'train/grad_norm_pre_clip': 0.2781025171279907, 'timing/data': 0.026028284803032875, 'timing/model': 1.5327189397066832, 'learning_rate/qwen_vl_interface': 3.249619509373712e-06, 'learning_rate/action_model': 3.249619509373712e-05, 'epoch': 1.25}) +[2026-06-16 04:27:17,486][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.2677212953567505, 'train/grad_norm_pre_clip': 16.679290771484375, 'timing/data': 0.026014599949121475, 'timing/model': 1.4417671337723732, 'learning_rate/qwen_vl_interface': 3.2468577264372905e-06, 'learning_rate/action_model': 3.24685772643729e-05, 'epoch': 1.25}) +[2026-06-16 04:27:23,391][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.0006420560530386865, 'train/grad_norm_pre_clip': 0.06764058768749237, 'timing/data': 0.028100699186325073, 'timing/model': 1.4395869839936495, 'learning_rate/qwen_vl_interface': 3.244096766916992e-06, 'learning_rate/action_model': 3.244096766916992e-05, 'epoch': 1.25}) +[2026-06-16 04:27:29,293][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.014203554019331932, 'train/grad_norm_pre_clip': 1.9248814582824707, 'timing/data': 0.025932297110557556, 'timing/model': 1.4413466211408377, 'learning_rate/qwen_vl_interface': 3.241336631947743e-06, 'learning_rate/action_model': 3.2413366319477427e-05, 'epoch': 1.25}) +[2026-06-16 04:27:35,195][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.1892624944448471, 'train/grad_norm_pre_clip': 35.75955581665039, 'timing/data': 0.025494806468486786, 'timing/model': 1.430648835375905, 'learning_rate/qwen_vl_interface': 3.238577322664131e-06, 'learning_rate/action_model': 3.238577322664131e-05, 'epoch': 1.25}) +[2026-06-16 04:27:41,094][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.3894450068473816, 'train/grad_norm_pre_clip': 15.360218048095703, 'timing/data': 0.02629026211798191, 'timing/model': 1.6409196984022856, 'learning_rate/qwen_vl_interface': 3.2358188402004026e-06, 'learning_rate/action_model': 3.2358188402004024e-05, 'epoch': 1.25}) +[2026-06-16 04:27:46,977][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.2583979666233063, 'train/grad_norm_pre_clip': 6.305166244506836, 'timing/data': 0.02664046920835972, 'timing/model': 1.6218281388282776, 'learning_rate/qwen_vl_interface': 3.2330611856904675e-06, 'learning_rate/action_model': 3.233061185690468e-05, 'epoch': 1.25}) +[2026-06-16 04:27:52,688][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.08667292445898056, 'train/grad_norm_pre_clip': 18.439712524414062, 'timing/data': 0.0342682208865881, 'timing/model': 1.4261777009814978, 'learning_rate/qwen_vl_interface': 3.2303043602678928e-06, 'learning_rate/action_model': 3.230304360267893e-05, 'epoch': 1.25}) +[2026-06-16 04:27:58,589][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.0688781589269638, 'train/grad_norm_pre_clip': 6.687500953674316, 'timing/data': 0.026538804173469543, 'timing/model': 1.4354866575449705, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 1.25}) +[2026-06-16 04:28:04,494][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.09805738180875778, 'train/grad_norm_pre_clip': 6.631331920623779, 'timing/data': 0.025686850771307945, 'timing/model': 1.4396657682955265, 'learning_rate/qwen_vl_interface': 3.2247932012173887e-06, 'learning_rate/action_model': 3.2247932012173884e-05, 'epoch': 1.25}) +[2026-06-16 04:28:10,772][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.005081657785922289, 'train/grad_norm_pre_clip': 0.6583746671676636, 'timing/data': 0.026627561077475548, 'timing/model': 1.4921956527978182, 'learning_rate/qwen_vl_interface': 3.2220388698548897e-06, 'learning_rate/action_model': 3.22203886985489e-05, 'epoch': 1.25}) +[2026-06-16 04:28:16,575][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.08640759438276291, 'train/grad_norm_pre_clip': 8.082104682922363, 'timing/data': 0.026480287313461304, 'timing/model': 1.6217343844473362, 'learning_rate/qwen_vl_interface': 3.21928537211061e-06, 'learning_rate/action_model': 3.21928537211061e-05, 'epoch': 1.25}) +[2026-06-16 04:28:22,190][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.2566676437854767, 'train/grad_norm_pre_clip': 17.096298217773438, 'timing/data': 0.025082604959607124, 'timing/model': 1.4374574162065983, 'learning_rate/qwen_vl_interface': 3.216532709116405e-06, 'learning_rate/action_model': 3.216532709116405e-05, 'epoch': 1.25}) +[2026-06-16 04:28:28,091][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.0006359557737596333, 'train/grad_norm_pre_clip': 0.07569225132465363, 'timing/data': 0.026785636320710182, 'timing/model': 1.4371645711362362, 'learning_rate/qwen_vl_interface': 3.2137808820037925e-06, 'learning_rate/action_model': 3.2137808820037924e-05, 'epoch': 1.25}) +[2026-06-16 04:28:33,993][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.151278555393219, 'train/grad_norm_pre_clip': 17.241483688354492, 'timing/data': 0.0261037889868021, 'timing/model': 1.4387645293027163, 'learning_rate/qwen_vl_interface': 3.211029891903945e-06, 'learning_rate/action_model': 3.211029891903945e-05, 'epoch': 1.25}) +[2026-06-16 04:28:39,987][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.12858889997005463, 'train/grad_norm_pre_clip': 7.571009635925293, 'timing/data': 0.02720191329717636, 'timing/model': 1.513551577925682, 'learning_rate/qwen_vl_interface': 3.2082797399476894e-06, 'learning_rate/action_model': 3.2082797399476895e-05, 'epoch': 1.26}) +[2026-06-16 04:28:45,807][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.1283523440361023, 'train/grad_norm_pre_clip': 15.458842277526855, 'timing/data': 0.025805147364735603, 'timing/model': 1.4341916795819998, 'learning_rate/qwen_vl_interface': 3.2055304272655114e-06, 'learning_rate/action_model': 3.205530427265511e-05, 'epoch': 1.26}) +[2026-06-16 04:28:51,627][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.1291818767786026, 'train/grad_norm_pre_clip': 14.667558670043945, 'timing/data': 0.02667292393743992, 'timing/model': 1.5740439649671316, 'learning_rate/qwen_vl_interface': 3.2027819549875477e-06, 'learning_rate/action_model': 3.202781954987548e-05, 'epoch': 1.26}) +[2026-06-16 04:28:57,289][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.1556626409292221, 'train/grad_norm_pre_clip': 18.322023391723633, 'timing/data': 0.026117343455553055, 'timing/model': 1.4355380199849606, 'learning_rate/qwen_vl_interface': 3.2000343242435926e-06, 'learning_rate/action_model': 3.2000343242435924e-05, 'epoch': 1.26}) +[2026-06-16 04:29:03,191][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.05205081030726433, 'train/grad_norm_pre_clip': 11.856945991516113, 'timing/data': 0.02777865342795849, 'timing/model': 1.4369825180619955, 'learning_rate/qwen_vl_interface': 3.1972875361630923e-06, 'learning_rate/action_model': 3.1972875361630923e-05, 'epoch': 1.26}) +[2026-06-16 04:29:09,188][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.15819379687309265, 'train/grad_norm_pre_clip': 16.466209411621094, 'timing/data': 0.026660075411200523, 'timing/model': 1.4356665834784508, 'learning_rate/qwen_vl_interface': 3.1945415918751484e-06, 'learning_rate/action_model': 3.194541591875148e-05, 'epoch': 1.26}) +[2026-06-16 04:29:15,086][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.04828065261244774, 'train/grad_norm_pre_clip': 12.543009757995605, 'timing/data': 0.025487588718533516, 'timing/model': 1.4278556481003761, 'learning_rate/qwen_vl_interface': 3.191796492508516e-06, 'learning_rate/action_model': 3.191796492508516e-05, 'epoch': 1.26}) +[2026-06-16 04:29:21,188][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.11411832273006439, 'train/grad_norm_pre_clip': 7.1010942459106445, 'timing/data': 0.026354031637310982, 'timing/model': 1.4270595647394657, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 1.26}) +[2026-06-16 04:29:27,029][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.009215980768203735, 'train/grad_norm_pre_clip': 1.0653374195098877, 'timing/data': 0.02550164796411991, 'timing/model': 1.575842224061489, 'learning_rate/qwen_vl_interface': 3.1863088330524577e-06, 'learning_rate/action_model': 3.186308833052458e-05, 'epoch': 1.26}) +[2026-06-16 04:29:33,040][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.01140030287206173, 'train/grad_norm_pre_clip': 6.638880729675293, 'timing/data': 0.03152942284941673, 'timing/model': 1.514990173280239, 'learning_rate/qwen_vl_interface': 3.183566275218803e-06, 'learning_rate/action_model': 3.1835662752188026e-05, 'epoch': 1.26}) +[2026-06-16 04:29:39,244][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.05837235972285271, 'train/grad_norm_pre_clip': 11.123981475830078, 'timing/data': 0.03231218084692955, 'timing/model': 1.5196553654968739, 'learning_rate/qwen_vl_interface': 3.180824566817996e-06, 'learning_rate/action_model': 3.1808245668179956e-05, 'epoch': 1.26}) +[2026-06-16 04:29:45,449][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.14621640741825104, 'train/grad_norm_pre_clip': 18.212127685546875, 'timing/data': 0.03203502483665943, 'timing/model': 1.498429724946618, 'learning_rate/qwen_vl_interface': 3.1780837089770524e-06, 'learning_rate/action_model': 3.1780837089770524e-05, 'epoch': 1.26}) +[2026-06-16 04:29:51,746][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.05463610589504242, 'train/grad_norm_pre_clip': 9.548087120056152, 'timing/data': 0.03202403336763382, 'timing/model': 1.5217120964080095, 'learning_rate/qwen_vl_interface': 3.17534370282263e-06, 'learning_rate/action_model': 3.1753437028226303e-05, 'epoch': 1.26}) +[2026-06-16 04:29:57,989][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.014044489711523056, 'train/grad_norm_pre_clip': 3.5761139392852783, 'timing/data': 0.03138880990445614, 'timing/model': 1.7585806977003813, 'learning_rate/qwen_vl_interface': 3.172604549481046e-06, 'learning_rate/action_model': 3.1726045494810455e-05, 'epoch': 1.26}) +[2026-06-16 04:30:04,045][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.19526511430740356, 'train/grad_norm_pre_clip': 14.888505935668945, 'timing/data': 0.0353511106222868, 'timing/model': 1.5151647944003344, 'learning_rate/qwen_vl_interface': 3.169866250078261e-06, 'learning_rate/action_model': 3.169866250078261e-05, 'epoch': 1.26}) +[2026-06-16 04:30:10,263][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.13968412578105927, 'train/grad_norm_pre_clip': 19.95104217529297, 'timing/data': 0.03216833434998989, 'timing/model': 1.5346939731389284, 'learning_rate/qwen_vl_interface': 3.167128805739887e-06, 'learning_rate/action_model': 3.1671288057398866e-05, 'epoch': 1.26}) +[2026-06-16 04:30:16,364][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.15230564773082733, 'train/grad_norm_pre_clip': 9.44659423828125, 'timing/data': 0.03409898839890957, 'timing/model': 1.5363976527005434, 'learning_rate/qwen_vl_interface': 3.1643922175911847e-06, 'learning_rate/action_model': 3.164392217591185e-05, 'epoch': 1.26}) +[2026-06-16 04:30:22,462][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.0653156265616417, 'train/grad_norm_pre_clip': 11.938726425170898, 'timing/data': 0.03292955085635185, 'timing/model': 1.5357943940907717, 'learning_rate/qwen_vl_interface': 3.1616564867570603e-06, 'learning_rate/action_model': 3.16165648675706e-05, 'epoch': 1.26}) +[2026-06-16 04:30:28,664][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.1747034192085266, 'train/grad_norm_pre_clip': 6.244226455688477, 'timing/data': 0.03233398124575615, 'timing/model': 1.5426887478679419, 'learning_rate/qwen_vl_interface': 3.1589216143620733e-06, 'learning_rate/action_model': 3.1589216143620734e-05, 'epoch': 1.26}) +[2026-06-16 04:30:34,904][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.11184253543615341, 'train/grad_norm_pre_clip': 11.417924880981445, 'timing/data': 0.03151634521782398, 'timing/model': 1.7782176863402128, 'learning_rate/qwen_vl_interface': 3.1561876015304215e-06, 'learning_rate/action_model': 3.156187601530421e-05, 'epoch': 1.26}) +[2026-06-16 04:30:40,972][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.09289420396089554, 'train/grad_norm_pre_clip': 23.61186981201172, 'timing/data': 0.03215157613158226, 'timing/model': 1.535767188295722, 'learning_rate/qwen_vl_interface': 3.1534544493859563e-06, 'learning_rate/action_model': 3.1534544493859566e-05, 'epoch': 1.26}) +[2026-06-16 04:30:47,255][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.0031118218321353197, 'train/grad_norm_pre_clip': 0.237756609916687, 'timing/data': 0.03029315173625946, 'timing/model': 1.51233633980155, 'learning_rate/qwen_vl_interface': 3.150722159052174e-06, 'learning_rate/action_model': 3.150722159052174e-05, 'epoch': 1.26}) +[2026-06-16 04:30:53,464][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.01806846261024475, 'train/grad_norm_pre_clip': 3.860819101333618, 'timing/data': 0.03127189166843891, 'timing/model': 1.5424560084939003, 'learning_rate/qwen_vl_interface': 3.147990731652215e-06, 'learning_rate/action_model': 3.147990731652215e-05, 'epoch': 1.26}) +[2026-06-16 04:30:59,488][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.13419020175933838, 'train/grad_norm_pre_clip': 9.898688316345215, 'timing/data': 0.026417596265673637, 'timing/model': 1.4738667998462915, 'learning_rate/qwen_vl_interface': 3.1452601683088673e-06, 'learning_rate/action_model': 3.145260168308867e-05, 'epoch': 1.26}) +[2026-06-16 04:31:05,386][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.26748648285865784, 'train/grad_norm_pre_clip': 8.53402042388916, 'timing/data': 0.02659783884882927, 'timing/model': 1.4507628586143255, 'learning_rate/qwen_vl_interface': 3.1425304701445608e-06, 'learning_rate/action_model': 3.1425304701445605e-05, 'epoch': 1.26}) +[2026-06-16 04:31:11,285][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.1963060349225998, 'train/grad_norm_pre_clip': 15.807577133178711, 'timing/data': 0.0270315483212471, 'timing/model': 1.6204643249511719, 'learning_rate/qwen_vl_interface': 3.139801638281372e-06, 'learning_rate/action_model': 3.139801638281372e-05, 'epoch': 1.26}) +[2026-06-16 04:31:16,994][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.12165333330631256, 'train/grad_norm_pre_clip': 9.261552810668945, 'timing/data': 0.02658792957663536, 'timing/model': 1.4253849107772112, 'learning_rate/qwen_vl_interface': 3.13707367384102e-06, 'learning_rate/action_model': 3.13707367384102e-05, 'epoch': 1.27}) +[2026-06-16 04:31:22,886][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.052695658057928085, 'train/grad_norm_pre_clip': 11.383045196533203, 'timing/data': 0.02580830827355385, 'timing/model': 1.4329086430370808, 'learning_rate/qwen_vl_interface': 3.134346577944869e-06, 'learning_rate/action_model': 3.134346577944869e-05, 'epoch': 1.27}) +[2026-06-16 04:31:28,789][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.09147217869758606, 'train/grad_norm_pre_clip': 11.636042594909668, 'timing/data': 0.025943448767066002, 'timing/model': 1.439482094720006, 'learning_rate/qwen_vl_interface': 3.1316203517139243e-06, 'learning_rate/action_model': 3.131620351713924e-05, 'epoch': 1.27}) +[2026-06-16 04:31:34,689][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.028959698975086212, 'train/grad_norm_pre_clip': 5.605719089508057, 'timing/data': 0.025404196232557297, 'timing/model': 1.457436565309763, 'learning_rate/qwen_vl_interface': 3.128894996268837e-06, 'learning_rate/action_model': 3.128894996268837e-05, 'epoch': 1.27}) +[2026-06-16 04:31:40,573][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.008335157297551632, 'train/grad_norm_pre_clip': 0.8828812837600708, 'timing/data': 0.026494355872273445, 'timing/model': 1.4970979671925306, 'learning_rate/qwen_vl_interface': 3.1261705127298934e-06, 'learning_rate/action_model': 3.1261705127298935e-05, 'epoch': 1.27}) +[2026-06-16 04:31:46,369][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.01881198026239872, 'train/grad_norm_pre_clip': 3.610700845718384, 'timing/data': 0.02570340596139431, 'timing/model': 1.6233841869980097, 'learning_rate/qwen_vl_interface': 3.123446902217028e-06, 'learning_rate/action_model': 3.1234469022170283e-05, 'epoch': 1.27}) +[2026-06-16 04:31:51,988][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.10489938408136368, 'train/grad_norm_pre_clip': 15.89310359954834, 'timing/data': 0.02573252096772194, 'timing/model': 1.4454591423273087, 'learning_rate/qwen_vl_interface': 3.1207241658498145e-06, 'learning_rate/action_model': 3.1207241658498145e-05, 'epoch': 1.27}) +[2026-06-16 04:31:57,891][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.007468762341886759, 'train/grad_norm_pre_clip': 1.304781436920166, 'timing/data': 0.026158956810832024, 'timing/model': 1.444883519783616, 'learning_rate/qwen_vl_interface': 3.1180023047474667e-06, 'learning_rate/action_model': 3.1180023047474664e-05, 'epoch': 1.27}) +[2026-06-16 04:32:03,791][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.1003042608499527, 'train/grad_norm_pre_clip': 9.608329772949219, 'timing/data': 0.025459326803684235, 'timing/model': 1.4451313726603985, 'learning_rate/qwen_vl_interface': 3.1152813200288386e-06, 'learning_rate/action_model': 3.115281320028838e-05, 'epoch': 1.27}) +[2026-06-16 04:32:09,592][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.0037049506790935993, 'train/grad_norm_pre_clip': 0.4180174767971039, 'timing/data': 0.025903282687067986, 'timing/model': 1.4442488607019186, 'learning_rate/qwen_vl_interface': 3.1125612128124244e-06, 'learning_rate/action_model': 3.112561212812424e-05, 'epoch': 1.27}) +[2026-06-16 04:32:15,487][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.005574690643697977, 'train/grad_norm_pre_clip': 0.5589695572853088, 'timing/data': 0.02616775408387184, 'timing/model': 1.4694944098591805, 'learning_rate/qwen_vl_interface': 3.1098419842163573e-06, 'learning_rate/action_model': 3.109841984216357e-05, 'epoch': 1.27}) +[2026-06-16 04:32:21,375][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.09049957990646362, 'train/grad_norm_pre_clip': 12.142232894897461, 'timing/data': 0.025740159675478935, 'timing/model': 1.6276124902069569, 'learning_rate/qwen_vl_interface': 3.10712363535841e-06, 'learning_rate/action_model': 3.10712363535841e-05, 'epoch': 1.27}) +[2026-06-16 04:32:26,981][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.08926820755004883, 'train/grad_norm_pre_clip': 13.34206771850586, 'timing/data': 0.025469236075878143, 'timing/model': 1.4369428176432848, 'learning_rate/qwen_vl_interface': 3.1044061673559926e-06, 'learning_rate/action_model': 3.1044061673559925e-05, 'epoch': 1.27}) +[2026-06-16 04:32:32,885][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.01542664784938097, 'train/grad_norm_pre_clip': 4.351236343383789, 'timing/data': 0.023548798635601997, 'timing/model': 1.4541513752192259, 'learning_rate/qwen_vl_interface': 3.1016895813261556e-06, 'learning_rate/action_model': 3.101689581326155e-05, 'epoch': 1.27}) +[2026-06-16 04:32:38,689][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.15102961659431458, 'train/grad_norm_pre_clip': 14.643892288208008, 'timing/data': 0.026028141379356384, 'timing/model': 1.4460057839751244, 'learning_rate/qwen_vl_interface': 3.098973878385581e-06, 'learning_rate/action_model': 3.098973878385581e-05, 'epoch': 1.27}) +[2026-06-16 04:32:44,481][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.006974507588893175, 'train/grad_norm_pre_clip': 0.7228183746337891, 'timing/data': 0.02534244768321514, 'timing/model': 1.4407502356916666, 'learning_rate/qwen_vl_interface': 3.096259059650594e-06, 'learning_rate/action_model': 3.096259059650594e-05, 'epoch': 1.27}) +[2026-06-16 04:32:50,388][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.006759051699191332, 'train/grad_norm_pre_clip': 0.7594503164291382, 'timing/data': 0.02500917576253414, 'timing/model': 1.470150288194418, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 1.27}) +[2026-06-16 04:32:56,275][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.04968596622347832, 'train/grad_norm_pre_clip': 8.876477241516113, 'timing/data': 0.025561310350894928, 'timing/model': 1.6348573956638575, 'learning_rate/qwen_vl_interface': 3.090832079260855e-06, 'learning_rate/action_model': 3.0908320792608544e-05, 'epoch': 1.27}) +[2026-06-16 04:33:01,879][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.1534828096628189, 'train/grad_norm_pre_clip': 8.795173645019531, 'timing/data': 0.026234721764922142, 'timing/model': 1.4409699272364378, 'learning_rate/qwen_vl_interface': 3.088119919836929e-06, 'learning_rate/action_model': 3.088119919836929e-05, 'epoch': 1.27}) +[2026-06-16 04:33:07,791][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.09489256143569946, 'train/grad_norm_pre_clip': 4.6772990226745605, 'timing/data': 0.025110526010394096, 'timing/model': 1.4466666597872972, 'learning_rate/qwen_vl_interface': 3.0854086490802433e-06, 'learning_rate/action_model': 3.085408649080243e-05, 'epoch': 1.27}) +[2026-06-16 04:33:13,693][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.00338844140060246, 'train/grad_norm_pre_clip': 1.0778406858444214, 'timing/data': 0.027770716696977615, 'timing/model': 1.447993140667677, 'learning_rate/qwen_vl_interface': 3.082698268105299e-06, 'learning_rate/action_model': 3.082698268105299e-05, 'epoch': 1.27}) +[2026-06-16 04:33:19,484][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.003402640111744404, 'train/grad_norm_pre_clip': 0.3481239676475525, 'timing/data': 0.02603055164217949, 'timing/model': 1.437341459095478, 'learning_rate/qwen_vl_interface': 3.0799887780262295e-06, 'learning_rate/action_model': 3.0799887780262296e-05, 'epoch': 1.27}) +[2026-06-16 04:33:25,379][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.05609412491321564, 'train/grad_norm_pre_clip': 5.964397430419922, 'timing/data': 0.023563511669635773, 'timing/model': 1.4633709639310837, 'learning_rate/qwen_vl_interface': 3.0772801799568054e-06, 'learning_rate/action_model': 3.0772801799568054e-05, 'epoch': 1.27}) +[2026-06-16 04:33:31,274][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.13050813972949982, 'train/grad_norm_pre_clip': 7.124476909637451, 'timing/data': 0.024802211672067642, 'timing/model': 1.6299996376037598, 'learning_rate/qwen_vl_interface': 3.07457247501043e-06, 'learning_rate/action_model': 3.0745724750104296e-05, 'epoch': 1.27}) +[2026-06-16 04:33:36,961][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.16801896691322327, 'train/grad_norm_pre_clip': 16.838598251342773, 'timing/data': 0.02437572367489338, 'timing/model': 1.4571648556739092, 'learning_rate/qwen_vl_interface': 3.0718656643001353e-06, 'learning_rate/action_model': 3.071865664300135e-05, 'epoch': 1.27}) +[2026-06-16 04:33:42,795][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.13115505874156952, 'train/grad_norm_pre_clip': 14.027100563049316, 'timing/data': 0.026857977733016014, 'timing/model': 1.4481258057057858, 'learning_rate/qwen_vl_interface': 3.069159748938591e-06, 'learning_rate/action_model': 3.069159748938591e-05, 'epoch': 1.27}) +[2026-06-16 04:33:48,579][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.10156188160181046, 'train/grad_norm_pre_clip': 9.863658905029297, 'timing/data': 0.02606736309826374, 'timing/model': 1.429159987717867, 'learning_rate/qwen_vl_interface': 3.066454730038096e-06, 'learning_rate/action_model': 3.0664547300380956e-05, 'epoch': 1.28}) +[2026-06-16 04:33:54,492][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.10147984325885773, 'train/grad_norm_pre_clip': 24.52766227722168, 'timing/data': 0.02648330293595791, 'timing/model': 1.4486038256436586, 'learning_rate/qwen_vl_interface': 3.0637506087105816e-06, 'learning_rate/action_model': 3.0637506087105815e-05, 'epoch': 1.28}) +[2026-06-16 04:34:00,397][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.07602010667324066, 'train/grad_norm_pre_clip': 6.761867523193359, 'timing/data': 0.027220964431762695, 'timing/model': 1.4408439081162214, 'learning_rate/qwen_vl_interface': 3.0610473860676103e-06, 'learning_rate/action_model': 3.06104738606761e-05, 'epoch': 1.28}) +[2026-06-16 04:34:06,272][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.012273543514311314, 'train/grad_norm_pre_clip': 3.648181200027466, 'timing/data': 0.027695607393980026, 'timing/model': 1.6108006611466408, 'learning_rate/qwen_vl_interface': 3.0583450632203748e-06, 'learning_rate/action_model': 3.058345063220375e-05, 'epoch': 1.28}) +[2026-06-16 04:34:11,980][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.035034872591495514, 'train/grad_norm_pre_clip': 3.828535318374634, 'timing/data': 0.025515004992485046, 'timing/model': 1.5298439841717482, 'learning_rate/qwen_vl_interface': 3.055643641279699e-06, 'learning_rate/action_model': 3.055643641279699e-05, 'epoch': 1.28}) +[2026-06-16 04:42:11,846][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.10501033812761307, 'train/grad_norm_pre_clip': 12.298232078552246, 'eval/loss': 0.04997913911938667, 'eval/flap_precision': 0.9564625850340136, 'eval/flap_recall': 0.9538670284938942, 'eval/flap_f1': 0.9551630434782609, 'eval/flap_rate_model': 0.20416666666666666, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.955163043478261, 'timing/data': 0.025973770767450333, 'timing/model': 1.4375358745455742, 'learning_rate/qwen_vl_interface': 3.0529431213560334e-06, 'learning_rate/action_model': 3.0529431213560334e-05, 'epoch': 1.28}) +[2026-06-16 04:42:11,847][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3300_state +[2026-06-16 04:42:11,851][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 04:42:17,768][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3300_state/model.safetensors +[2026-06-16 04:42:28,044][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3300_state/optimizer.bin +[2026-06-16 04:42:28,045][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3300_state/sampler.bin +[2026-06-16 04:42:28,045][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3300_state/sampler_1.bin +[2026-06-16 04:42:28,046][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3300_state/random_states_0.pkl +[2026-06-16 04:42:28,142][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3300 +[2026-06-16 04:42:28,142][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 04:42:28,150][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 04:42:33,892][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.23408514261245728, 'train/grad_norm_pre_clip': 10.927230834960938, 'timing/data': 0.0271791759878397, 'timing/model': 1.4260200634598732, 'learning_rate/qwen_vl_interface': 3.0502435045594596e-06, 'learning_rate/action_model': 3.0502435045594595e-05, 'epoch': 1.28}) +[2026-06-16 04:42:39,795][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.21236062049865723, 'train/grad_norm_pre_clip': 13.678009986877441, 'timing/data': 0.026582250371575356, 'timing/model': 1.4329572152346373, 'learning_rate/qwen_vl_interface': 3.047544791999689e-06, 'learning_rate/action_model': 3.0475447919996892e-05, 'epoch': 1.28}) +[2026-06-16 04:42:45,698][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.1254299432039261, 'train/grad_norm_pre_clip': 9.961316108703613, 'timing/data': 0.027013294398784637, 'timing/model': 1.4358272161334753, 'learning_rate/qwen_vl_interface': 3.044846984786061e-06, 'learning_rate/action_model': 3.044846984786061e-05, 'epoch': 1.28}) +[2026-06-16 04:42:51,499][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.08141840994358063, 'train/grad_norm_pre_clip': 9.237837791442871, 'timing/data': 0.026551663875579834, 'timing/model': 1.4316086992621422, 'learning_rate/qwen_vl_interface': 3.0421500840275413e-06, 'learning_rate/action_model': 3.0421500840275413e-05, 'epoch': 1.28}) +[2026-06-16 04:42:57,311][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.05191168189048767, 'train/grad_norm_pre_clip': 5.607336044311523, 'timing/data': 0.027032548561692238, 'timing/model': 1.4358990620821714, 'learning_rate/qwen_vl_interface': 3.039454090832727e-06, 'learning_rate/action_model': 3.0394540908327268e-05, 'epoch': 1.28}) +[2026-06-16 04:43:03,159][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.09000159800052643, 'train/grad_norm_pre_clip': 14.41650676727295, 'timing/data': 0.025962430983781815, 'timing/model': 1.6286660302430391, 'learning_rate/qwen_vl_interface': 3.0367590063098333e-06, 'learning_rate/action_model': 3.0367590063098333e-05, 'epoch': 1.28}) +[2026-06-16 04:43:08,775][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.14993466436862946, 'train/grad_norm_pre_clip': 10.657130241394043, 'timing/data': 0.03216094151139259, 'timing/model': 1.4256805330514908, 'learning_rate/qwen_vl_interface': 3.0340648315667115e-06, 'learning_rate/action_model': 3.0340648315667113e-05, 'epoch': 1.28}) +[2026-06-16 04:43:14,616][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.03204004466533661, 'train/grad_norm_pre_clip': 5.807342052459717, 'timing/data': 0.026640694588422775, 'timing/model': 1.442843072116375, 'learning_rate/qwen_vl_interface': 3.0313715677108336e-06, 'learning_rate/action_model': 3.0313715677108335e-05, 'epoch': 1.28}) +[2026-06-16 04:43:20,594][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.06825895607471466, 'train/grad_norm_pre_clip': 8.682429313659668, 'timing/data': 0.025850564241409302, 'timing/model': 1.5144039280712605, 'learning_rate/qwen_vl_interface': 3.0286792158492996e-06, 'learning_rate/action_model': 3.0286792158492994e-05, 'epoch': 1.28}) +[2026-06-16 04:43:26,606][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.22071243822574615, 'train/grad_norm_pre_clip': 9.159586906433105, 'timing/data': 0.026280175894498825, 'timing/model': 1.5078353211283684, 'learning_rate/qwen_vl_interface': 3.0259877770888334e-06, 'learning_rate/action_model': 3.0259877770888333e-05, 'epoch': 1.28}) +[2026-06-16 04:43:32,519][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.138151615858078, 'train/grad_norm_pre_clip': 11.177964210510254, 'timing/data': 0.026022369042038918, 'timing/model': 1.6438590791076422, 'learning_rate/qwen_vl_interface': 3.0232972525357836e-06, 'learning_rate/action_model': 3.0232972525357834e-05, 'epoch': 1.28}) +[2026-06-16 04:43:38,367][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.12419672310352325, 'train/grad_norm_pre_clip': 12.330490112304688, 'timing/data': 0.02739071659743786, 'timing/model': 1.5905427932739258, 'learning_rate/qwen_vl_interface': 3.020607643296123e-06, 'learning_rate/action_model': 3.0206076432961226e-05, 'epoch': 1.28}) +[2026-06-16 04:43:43,990][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.014280631206929684, 'train/grad_norm_pre_clip': 1.6665287017822266, 'timing/data': 0.025526337325572968, 'timing/model': 1.4405010901391506, 'learning_rate/qwen_vl_interface': 3.017918950475449e-06, 'learning_rate/action_model': 3.017918950475449e-05, 'epoch': 1.28}) +[2026-06-16 04:43:49,780][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.004835172090679407, 'train/grad_norm_pre_clip': 0.700150728225708, 'timing/data': 0.025680141523480415, 'timing/model': 1.436527531594038, 'learning_rate/qwen_vl_interface': 3.0152311751789815e-06, 'learning_rate/action_model': 3.0152311751789814e-05, 'epoch': 1.28}) +[2026-06-16 04:43:55,685][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.04981584846973419, 'train/grad_norm_pre_clip': 8.509559631347656, 'timing/data': 0.026301851496100426, 'timing/model': 1.4388430640101433, 'learning_rate/qwen_vl_interface': 3.0125443185115655e-06, 'learning_rate/action_model': 3.0125443185115655e-05, 'epoch': 1.28}) +[2026-06-16 04:44:01,583][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.20363806188106537, 'train/grad_norm_pre_clip': 10.882091522216797, 'timing/data': 0.02650594152510166, 'timing/model': 1.4127179440110922, 'learning_rate/qwen_vl_interface': 3.0098583815776615e-06, 'learning_rate/action_model': 3.0098583815776615e-05, 'epoch': 1.28}) +[2026-06-16 04:44:07,475][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.10097028315067291, 'train/grad_norm_pre_clip': 15.920787811279297, 'timing/data': 0.027092311531305313, 'timing/model': 1.4230166506022215, 'learning_rate/qwen_vl_interface': 3.0071733654813595e-06, 'learning_rate/action_model': 3.0071733654813595e-05, 'epoch': 1.28}) +[2026-06-16 04:44:13,215][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.08323442190885544, 'train/grad_norm_pre_clip': 12.51335334777832, 'timing/data': 0.026273779571056366, 'timing/model': 1.572922982275486, 'learning_rate/qwen_vl_interface': 3.0044892713263685e-06, 'learning_rate/action_model': 3.0044892713263682e-05, 'epoch': 1.28}) +[2026-06-16 04:44:18,877][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.1149778664112091, 'train/grad_norm_pre_clip': 14.438551902770996, 'timing/data': 0.02652382105588913, 'timing/model': 1.4349867310374975, 'learning_rate/qwen_vl_interface': 3.001806100216017e-06, 'learning_rate/action_model': 3.001806100216017e-05, 'epoch': 1.28}) +[2026-06-16 04:44:24,779][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.11743862926959991, 'train/grad_norm_pre_clip': 11.536136627197266, 'timing/data': 0.02581067569553852, 'timing/model': 1.4370654933154583, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 1.28}) +[2026-06-16 04:44:30,804][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.019122280180454254, 'train/grad_norm_pre_clip': 3.190929889678955, 'timing/data': 0.0264033991843462, 'timing/model': 1.5247089006006718, 'learning_rate/qwen_vl_interface': 2.996442531540657e-06, 'learning_rate/action_model': 2.9964425315406568e-05, 'epoch': 1.29}) +[2026-06-16 04:44:36,815][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.02681146375834942, 'train/grad_norm_pre_clip': 4.247776508331299, 'timing/data': 0.026074539870023727, 'timing/model': 1.517802806571126, 'learning_rate/qwen_vl_interface': 2.993762136180408e-06, 'learning_rate/action_model': 2.993762136180408e-05, 'epoch': 1.29}) +[2026-06-16 04:44:42,802][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.056645117700099945, 'train/grad_norm_pre_clip': 6.452848434448242, 'timing/data': 0.026324309408664703, 'timing/model': 1.7218404579907656, 'learning_rate/qwen_vl_interface': 2.9910826682743195e-06, 'learning_rate/action_model': 2.9910826682743192e-05, 'epoch': 1.29}) +[2026-06-16 04:44:48,600][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.08824901282787323, 'train/grad_norm_pre_clip': 2.5343410968780518, 'timing/data': 0.02675052545964718, 'timing/model': 1.5212584715336561, 'learning_rate/qwen_vl_interface': 2.9884041289238192e-06, 'learning_rate/action_model': 2.988404128923819e-05, 'epoch': 1.29}) +[2026-06-16 04:44:54,835][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.12337427586317062, 'train/grad_norm_pre_clip': 10.921124458312988, 'timing/data': 0.026176976040005684, 'timing/model': 1.64802055247128, 'learning_rate/qwen_vl_interface': 2.9857265192299557e-06, 'learning_rate/action_model': 2.9857265192299556e-05, 'epoch': 1.29}) +[2026-06-16 04:45:00,710][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.027012012898921967, 'train/grad_norm_pre_clip': 2.4891865253448486, 'timing/data': 0.026879431679844856, 'timing/model': 1.5305418204516172, 'learning_rate/qwen_vl_interface': 2.9830498402933894e-06, 'learning_rate/action_model': 2.9830498402933894e-05, 'epoch': 1.29}) +[2026-06-16 04:45:06,746][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.1022471934556961, 'train/grad_norm_pre_clip': 9.67529296875, 'timing/data': 0.0338395182043314, 'timing/model': 1.5202557854354382, 'learning_rate/qwen_vl_interface': 2.9803740932144025e-06, 'learning_rate/action_model': 2.9803740932144024e-05, 'epoch': 1.29}) +[2026-06-16 04:45:12,851][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.24081990122795105, 'train/grad_norm_pre_clip': 10.876221656799316, 'timing/data': 0.037865232676267624, 'timing/model': 1.5572012681514025, 'learning_rate/qwen_vl_interface': 2.9776992790928953e-06, 'learning_rate/action_model': 2.9776992790928952e-05, 'epoch': 1.29}) +[2026-06-16 04:45:18,969][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.1993131786584854, 'train/grad_norm_pre_clip': 11.098505020141602, 'timing/data': 0.0368953887373209, 'timing/model': 1.7344474010169506, 'learning_rate/qwen_vl_interface': 2.975025399028382e-06, 'learning_rate/action_model': 2.975025399028382e-05, 'epoch': 1.29}) +[2026-06-16 04:45:25,050][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.05013565346598625, 'train/grad_norm_pre_clip': 7.356333255767822, 'timing/data': 0.03608779422938824, 'timing/model': 1.5186510533094406, 'learning_rate/qwen_vl_interface': 2.9723524541199943e-06, 'learning_rate/action_model': 2.972352454119994e-05, 'epoch': 1.29}) +[2026-06-16 04:45:30,885][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.022170452401041985, 'train/grad_norm_pre_clip': 1.4745261669158936, 'timing/data': 0.02657756581902504, 'timing/model': 1.4308546278625727, 'learning_rate/qwen_vl_interface': 2.969680445466478e-06, 'learning_rate/action_model': 2.969680445466478e-05, 'epoch': 1.29}) +[2026-06-16 04:45:36,785][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.03215813264250755, 'train/grad_norm_pre_clip': 11.520845413208008, 'timing/data': 0.025825798511505127, 'timing/model': 1.433688459917903, 'learning_rate/qwen_vl_interface': 2.9670093741661955e-06, 'learning_rate/action_model': 2.9670093741661954e-05, 'epoch': 1.29}) +[2026-06-16 04:45:42,684][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.17249798774719238, 'train/grad_norm_pre_clip': 9.97321605682373, 'timing/data': 0.026068072766065598, 'timing/model': 1.4363610967993736, 'learning_rate/qwen_vl_interface': 2.964339241317123e-06, 'learning_rate/action_model': 2.964339241317123e-05, 'epoch': 1.29}) +[2026-06-16 04:45:48,841][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.02209961786866188, 'train/grad_norm_pre_clip': 1.9364935159683228, 'timing/data': 0.03196534886956215, 'timing/model': 1.5008774641901255, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 1.29}) +[2026-06-16 04:45:55,082][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.10990209132432938, 'train/grad_norm_pre_clip': 7.538702011108398, 'timing/data': 0.032924139872193336, 'timing/model': 1.7847090754657984, 'learning_rate/qwen_vl_interface': 2.9590017953625854e-06, 'learning_rate/action_model': 2.9590017953625853e-05, 'epoch': 1.29}) +[2026-06-16 04:46:01,143][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.3392466902732849, 'train/grad_norm_pre_clip': 13.398819923400879, 'timing/data': 0.03223680704832077, 'timing/model': 1.533549014478922, 'learning_rate/qwen_vl_interface': 2.9563344844511455e-06, 'learning_rate/action_model': 2.9563344844511455e-05, 'epoch': 1.29}) +[2026-06-16 04:46:07,339][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.04936368763446808, 'train/grad_norm_pre_clip': 6.004038333892822, 'timing/data': 0.031683435663580894, 'timing/model': 1.53362400457263, 'learning_rate/qwen_vl_interface': 2.9536681163789568e-06, 'learning_rate/action_model': 2.9536681163789564e-05, 'epoch': 1.29}) +[2026-06-16 04:46:13,547][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.009462249465286732, 'train/grad_norm_pre_clip': 1.0189263820648193, 'timing/data': 0.031407080590724945, 'timing/model': 1.5342101827263832, 'learning_rate/qwen_vl_interface': 2.9510026922420653e-06, 'learning_rate/action_model': 2.9510026922420653e-05, 'epoch': 1.29}) +[2026-06-16 04:46:19,741][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.18274779617786407, 'train/grad_norm_pre_clip': 11.703598022460938, 'timing/data': 0.03200400434434414, 'timing/model': 1.4805876649916172, 'learning_rate/qwen_vl_interface': 2.9483382131361273e-06, 'learning_rate/action_model': 2.948338213136127e-05, 'epoch': 1.29}) +[2026-06-16 04:46:25,951][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.10420557856559753, 'train/grad_norm_pre_clip': 22.1129093170166, 'timing/data': 0.03202328272163868, 'timing/model': 1.4833070617169142, 'learning_rate/qwen_vl_interface': 2.9456746801564072e-06, 'learning_rate/action_model': 2.945674680156407e-05, 'epoch': 1.29}) +[2026-06-16 04:46:32,187][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.01696833223104477, 'train/grad_norm_pre_clip': 2.2889604568481445, 'timing/data': 0.03268032521009445, 'timing/model': 1.769528714939952, 'learning_rate/qwen_vl_interface': 2.9430120943977854e-06, 'learning_rate/action_model': 2.9430120943977853e-05, 'epoch': 1.29}) +[2026-06-16 04:46:38,240][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.07924385368824005, 'train/grad_norm_pre_clip': 5.06648063659668, 'timing/data': 0.032676370814442635, 'timing/model': 1.5093845669180155, 'learning_rate/qwen_vl_interface': 2.940350456954748e-06, 'learning_rate/action_model': 2.940350456954748e-05, 'epoch': 1.29}) +[2026-06-16 04:46:44,448][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.14720773696899414, 'train/grad_norm_pre_clip': 6.58890438079834, 'timing/data': 0.032060982659459114, 'timing/model': 1.5472501870244741, 'learning_rate/qwen_vl_interface': 2.9376897689213952e-06, 'learning_rate/action_model': 2.937689768921395e-05, 'epoch': 1.29}) +[2026-06-16 04:46:50,642][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.020563963800668716, 'train/grad_norm_pre_clip': 2.2707407474517822, 'timing/data': 0.031233083456754684, 'timing/model': 1.533085212111473, 'learning_rate/qwen_vl_interface': 2.9350300313914364e-06, 'learning_rate/action_model': 2.9350300313914363e-05, 'epoch': 1.29}) +[2026-06-16 04:46:56,748][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.03600465878844261, 'train/grad_norm_pre_clip': 4.336299896240234, 'timing/data': 0.030246702954173088, 'timing/model': 1.5325665958225727, 'learning_rate/qwen_vl_interface': 2.9323712454581867e-06, 'learning_rate/action_model': 2.9323712454581866e-05, 'epoch': 1.29}) +[2026-06-16 04:47:02,942][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.1693439930677414, 'train/grad_norm_pre_clip': 8.195470809936523, 'timing/data': 0.03180382400751114, 'timing/model': 1.5037516467273235, 'learning_rate/qwen_vl_interface': 2.929713412214572e-06, 'learning_rate/action_model': 2.929713412214572e-05, 'epoch': 1.29}) +[2026-06-16 04:47:09,067][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.07709795981645584, 'train/grad_norm_pre_clip': 8.878897666931152, 'timing/data': 0.037360550835728645, 'timing/model': 1.7327543273568153, 'learning_rate/qwen_vl_interface': 2.9270565327531294e-06, 'learning_rate/action_model': 2.927056532753129e-05, 'epoch': 1.3}) +[2026-06-16 04:47:15,045][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.032725654542446136, 'train/grad_norm_pre_clip': 5.5753350257873535, 'timing/data': 0.030513277277350426, 'timing/model': 1.534008601680398, 'learning_rate/qwen_vl_interface': 2.924400608166003e-06, 'learning_rate/action_model': 2.924400608166003e-05, 'epoch': 1.3}) +[2026-06-16 04:47:21,244][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.17284923791885376, 'train/grad_norm_pre_clip': 14.223362922668457, 'timing/data': 0.03374730423092842, 'timing/model': 1.5398506298661232, 'learning_rate/qwen_vl_interface': 2.921745639544939e-06, 'learning_rate/action_model': 2.921745639544939e-05, 'epoch': 1.3}) +[2026-06-16 04:47:27,343][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.0396481417119503, 'train/grad_norm_pre_clip': 7.032351493835449, 'timing/data': 0.03223214112222195, 'timing/model': 1.545854663476348, 'learning_rate/qwen_vl_interface': 2.919091627981297e-06, 'learning_rate/action_model': 2.919091627981297e-05, 'epoch': 1.3}) +[2026-06-16 04:47:33,439][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.010560127906501293, 'train/grad_norm_pre_clip': 2.81036114692688, 'timing/data': 0.030612612143158913, 'timing/model': 1.5282034315168858, 'learning_rate/qwen_vl_interface': 2.9164385745660396e-06, 'learning_rate/action_model': 2.9164385745660394e-05, 'epoch': 1.3}) +[2026-06-16 04:47:39,542][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.09792645275592804, 'train/grad_norm_pre_clip': 10.10217571258545, 'timing/data': 0.031178345903754234, 'timing/model': 1.5001434963196516, 'learning_rate/qwen_vl_interface': 2.9137864803897385e-06, 'learning_rate/action_model': 2.9137864803897387e-05, 'epoch': 1.3}) +[2026-06-16 04:47:45,660][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.10013441741466522, 'train/grad_norm_pre_clip': 9.40494441986084, 'timing/data': 0.03251976892352104, 'timing/model': 1.7323337998241186, 'learning_rate/qwen_vl_interface': 2.9111353465425696e-06, 'learning_rate/action_model': 2.9111353465425694e-05, 'epoch': 1.3}) +[2026-06-16 04:47:51,640][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.1446562558412552, 'train/grad_norm_pre_clip': 8.270840644836426, 'timing/data': 0.03148555941879749, 'timing/model': 1.5349193084985018, 'learning_rate/qwen_vl_interface': 2.908485174114312e-06, 'learning_rate/action_model': 2.908485174114312e-05, 'epoch': 1.3}) +[2026-06-16 04:47:57,843][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.22243990004062653, 'train/grad_norm_pre_clip': 12.112659454345703, 'timing/data': 0.03242374211549759, 'timing/model': 1.5380928330123425, 'learning_rate/qwen_vl_interface': 2.9058359641943523e-06, 'learning_rate/action_model': 2.9058359641943523e-05, 'epoch': 1.3}) +[2026-06-16 04:48:03,941][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.15371154248714447, 'train/grad_norm_pre_clip': 12.616109848022461, 'timing/data': 0.03203661926090717, 'timing/model': 1.53308286331594, 'learning_rate/qwen_vl_interface': 2.9031877178716823e-06, 'learning_rate/action_model': 2.903187717871682e-05, 'epoch': 1.3}) +[2026-06-16 04:48:10,038][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.04741798713803291, 'train/grad_norm_pre_clip': 7.0754876136779785, 'timing/data': 0.03186691552400589, 'timing/model': 1.515533396974206, 'learning_rate/qwen_vl_interface': 2.900540436234894e-06, 'learning_rate/action_model': 2.9005404362348937e-05, 'epoch': 1.3}) +[2026-06-16 04:48:16,239][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.04430603235960007, 'train/grad_norm_pre_clip': 11.178502082824707, 'timing/data': 0.03133142739534378, 'timing/model': 1.487832959741354, 'learning_rate/qwen_vl_interface': 2.897894120372185e-06, 'learning_rate/action_model': 2.897894120372185e-05, 'epoch': 1.3}) +[2026-06-16 04:48:22,363][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.01606098562479019, 'train/grad_norm_pre_clip': 4.121921062469482, 'timing/data': 0.032263560220599174, 'timing/model': 1.7533419262617826, 'learning_rate/qwen_vl_interface': 2.8952487713713585e-06, 'learning_rate/action_model': 2.8952487713713582e-05, 'epoch': 1.3}) +[2026-06-16 04:48:28,343][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.01446579396724701, 'train/grad_norm_pre_clip': 1.317438006401062, 'timing/data': 0.031412770971655846, 'timing/model': 1.535657536238432, 'learning_rate/qwen_vl_interface': 2.8926043903198135e-06, 'learning_rate/action_model': 2.8926043903198135e-05, 'epoch': 1.3}) +[2026-06-16 04:48:34,540][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.25244805216789246, 'train/grad_norm_pre_clip': 13.475191116333008, 'timing/data': 0.03127329610288143, 'timing/model': 1.551784336566925, 'learning_rate/qwen_vl_interface': 2.8899609783045574e-06, 'learning_rate/action_model': 2.8899609783045573e-05, 'epoch': 1.3}) +[2026-06-16 04:48:40,643][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.07365385442972183, 'train/grad_norm_pre_clip': 10.740408897399902, 'timing/data': 0.029544943943619728, 'timing/model': 1.5280098374933004, 'learning_rate/qwen_vl_interface': 2.8873185364121965e-06, 'learning_rate/action_model': 2.8873185364121963e-05, 'epoch': 1.3}) +[2026-06-16 04:48:46,842][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.05362636223435402, 'train/grad_norm_pre_clip': 7.8937506675720215, 'timing/data': 0.03128267079591751, 'timing/model': 1.5318394843488932, 'learning_rate/qwen_vl_interface': 2.8846770657289386e-06, 'learning_rate/action_model': 2.8846770657289386e-05, 'epoch': 1.3}) +[2026-06-16 04:48:52,960][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.14266948401927948, 'train/grad_norm_pre_clip': 13.956304550170898, 'timing/data': 0.03161296434700489, 'timing/model': 1.4869337975978851, 'learning_rate/qwen_vl_interface': 2.882036567340593e-06, 'learning_rate/action_model': 2.8820365673405932e-05, 'epoch': 1.3}) +[2026-06-16 04:48:59,128][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.020432330667972565, 'train/grad_norm_pre_clip': 5.142208099365234, 'timing/data': 0.031097039580345154, 'timing/model': 1.71951699629426, 'learning_rate/qwen_vl_interface': 2.8793970423325684e-06, 'learning_rate/action_model': 2.8793970423325685e-05, 'epoch': 1.3}) +[2026-06-16 04:49:05,189][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.003167729824781418, 'train/grad_norm_pre_clip': 0.25225305557250977, 'timing/data': 0.03189156390726566, 'timing/model': 1.6688245218247175, 'learning_rate/qwen_vl_interface': 2.8767584917898756e-06, 'learning_rate/action_model': 2.8767584917898754e-05, 'epoch': 1.3}) +[2026-06-16 04:49:11,145][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.07542724162340164, 'train/grad_norm_pre_clip': 9.107316017150879, 'timing/data': 0.030824769288301468, 'timing/model': 1.5217543356120586, 'learning_rate/qwen_vl_interface': 2.874120916797121e-06, 'learning_rate/action_model': 2.8741209167971207e-05, 'epoch': 1.3}) +[2026-06-16 04:49:17,343][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.029390741139650345, 'train/grad_norm_pre_clip': 5.687203407287598, 'timing/data': 0.031356168910861015, 'timing/model': 1.508582690730691, 'learning_rate/qwen_vl_interface': 2.8714843184385124e-06, 'learning_rate/action_model': 2.871484318438512e-05, 'epoch': 1.3}) +[2026-06-16 04:49:23,543][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.1619030088186264, 'train/grad_norm_pre_clip': 29.196819305419922, 'timing/data': 0.031032567843794823, 'timing/model': 1.5296754408627748, 'learning_rate/qwen_vl_interface': 2.8688486977978575e-06, 'learning_rate/action_model': 2.8688486977978572e-05, 'epoch': 1.3}) +[2026-06-16 04:49:29,752][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.025912277400493622, 'train/grad_norm_pre_clip': 8.297898292541504, 'timing/data': 0.03223271295428276, 'timing/model': 1.5378290303051472, 'learning_rate/qwen_vl_interface': 2.8662140559585584e-06, 'learning_rate/action_model': 2.866214055958558e-05, 'epoch': 1.3}) +[2026-06-16 04:49:35,969][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.005619630217552185, 'train/grad_norm_pre_clip': 0.7380086779594421, 'timing/data': 0.030436599627137184, 'timing/model': 1.7587911002337933, 'learning_rate/qwen_vl_interface': 2.8635803940036162e-06, 'learning_rate/action_model': 2.863580394003616e-05, 'epoch': 1.3}) +[2026-06-16 04:49:42,117][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.06387216597795486, 'train/grad_norm_pre_clip': 9.961746215820312, 'timing/data': 0.030722763389348984, 'timing/model': 1.7003062069416046, 'learning_rate/qwen_vl_interface': 2.8609477130156317e-06, 'learning_rate/action_model': 2.8609477130156314e-05, 'epoch': 1.31}) +[2026-06-16 04:49:48,144][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.16018988192081451, 'train/grad_norm_pre_clip': 11.532416343688965, 'timing/data': 0.03289756551384926, 'timing/model': 1.5293981488794088, 'learning_rate/qwen_vl_interface': 2.8583160140767997e-06, 'learning_rate/action_model': 2.8583160140767994e-05, 'epoch': 1.31}) +[2026-06-16 04:49:54,353][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.1548822969198227, 'train/grad_norm_pre_clip': 7.269509792327881, 'timing/data': 0.03221954591572285, 'timing/model': 1.5391006376594305, 'learning_rate/qwen_vl_interface': 2.8556852982689125e-06, 'learning_rate/action_model': 2.8556852982689126e-05, 'epoch': 1.31}) +[2026-06-16 04:50:00,545][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.12638360261917114, 'train/grad_norm_pre_clip': 16.812332153320312, 'timing/data': 0.03261683136224747, 'timing/model': 1.533736925572157, 'learning_rate/qwen_vl_interface': 2.8530555666733584e-06, 'learning_rate/action_model': 2.8530555666733583e-05, 'epoch': 1.31}) +[2026-06-16 04:50:06,764][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.12178350239992142, 'train/grad_norm_pre_clip': 7.582998275756836, 'timing/data': 0.0320309866219759, 'timing/model': 1.4924930669367313, 'learning_rate/qwen_vl_interface': 2.85042682037112e-06, 'learning_rate/action_model': 2.85042682037112e-05, 'epoch': 1.31}) +[2026-06-16 04:50:12,975][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.050812728703022, 'train/grad_norm_pre_clip': 5.114083766937256, 'timing/data': 0.03155190870165825, 'timing/model': 1.7537104785442352, 'learning_rate/qwen_vl_interface': 2.8477990604427764e-06, 'learning_rate/action_model': 2.8477990604427763e-05, 'epoch': 1.31}) +[2026-06-16 04:50:19,044][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.004448191728442907, 'train/grad_norm_pre_clip': 0.42224183678627014, 'timing/data': 0.032736239954829216, 'timing/model': 1.5493583846837282, 'learning_rate/qwen_vl_interface': 2.8451722879684994e-06, 'learning_rate/action_model': 2.8451722879684994e-05, 'epoch': 1.31}) +[2026-06-16 04:50:25,247][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.2288948893547058, 'train/grad_norm_pre_clip': 12.207380294799805, 'timing/data': 0.03201861120760441, 'timing/model': 1.544550934806466, 'learning_rate/qwen_vl_interface': 2.842546504028059e-06, 'learning_rate/action_model': 2.8425465040280586e-05, 'epoch': 1.31}) +[2026-06-16 04:50:31,348][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.031412120908498764, 'train/grad_norm_pre_clip': 4.358173370361328, 'timing/data': 0.033757343888282776, 'timing/model': 1.5093013774603605, 'learning_rate/qwen_vl_interface': 2.83992170970081e-06, 'learning_rate/action_model': 2.8399217097008102e-05, 'epoch': 1.31}) +[2026-06-16 04:50:37,553][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.04730977863073349, 'train/grad_norm_pre_clip': 12.086986541748047, 'timing/data': 0.031665096059441566, 'timing/model': 1.487465975806117, 'learning_rate/qwen_vl_interface': 2.837297906065711e-06, 'learning_rate/action_model': 2.8372979060657107e-05, 'epoch': 1.31}) +[2026-06-16 04:50:43,644][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.053536221385002136, 'train/grad_norm_pre_clip': 10.28528118133545, 'timing/data': 0.032344743609428406, 'timing/model': 1.5433628782629967, 'learning_rate/qwen_vl_interface': 2.834675094201306e-06, 'learning_rate/action_model': 2.8346750942013063e-05, 'epoch': 1.31}) +[2026-06-16 04:50:49,709][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.10422387719154358, 'train/grad_norm_pre_clip': 18.288564682006836, 'timing/data': 0.033352602273225784, 'timing/model': 1.6960108000785112, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 1.31}) +[2026-06-16 04:50:55,659][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.12696708738803864, 'train/grad_norm_pre_clip': 9.813192367553711, 'timing/data': 0.03238464705646038, 'timing/model': 1.545392731204629, 'learning_rate/qwen_vl_interface': 2.829432450096731e-06, 'learning_rate/action_model': 2.8294324500967306e-05, 'epoch': 1.31}) +[2026-06-16 04:51:01,874][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.06997182965278625, 'train/grad_norm_pre_clip': 10.607815742492676, 'timing/data': 0.03249797411262989, 'timing/model': 1.5263924896717072, 'learning_rate/qwen_vl_interface': 2.8268126200116118e-06, 'learning_rate/action_model': 2.8268126200116117e-05, 'epoch': 1.31}) +[2026-06-16 04:51:08,275][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.13003085553646088, 'train/grad_norm_pre_clip': 11.176351547241211, 'timing/data': 0.03453228808939457, 'timing/model': 1.5656377580016851, 'learning_rate/qwen_vl_interface': 2.824193786007294e-06, 'learning_rate/action_model': 2.8241937860072937e-05, 'epoch': 1.31}) +[2026-06-16 04:51:14,779][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.005256896838545799, 'train/grad_norm_pre_clip': 0.940383791923523, 'timing/data': 0.03286450542509556, 'timing/model': 1.588142916560173, 'learning_rate/qwen_vl_interface': 2.821575949160279e-06, 'learning_rate/action_model': 2.8215759491602787e-05, 'epoch': 1.31}) +[2026-06-16 04:51:21,217][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.1977926790714264, 'train/grad_norm_pre_clip': 15.125068664550781, 'timing/data': 0.02912888675928116, 'timing/model': 1.5107240211218596, 'learning_rate/qwen_vl_interface': 2.8189591105466625e-06, 'learning_rate/action_model': 2.8189591105466624e-05, 'epoch': 1.31}) +[2026-06-16 04:51:27,256][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.07390753924846649, 'train/grad_norm_pre_clip': 10.260811805725098, 'timing/data': 0.03088032267987728, 'timing/model': 1.7696827296167612, 'learning_rate/qwen_vl_interface': 2.816343271242126e-06, 'learning_rate/action_model': 2.816343271242126e-05, 'epoch': 1.31}) +[2026-06-16 04:51:33,277][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.17431025207042694, 'train/grad_norm_pre_clip': 13.947690963745117, 'timing/data': 0.031855132430791855, 'timing/model': 1.534317770972848, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 1.31}) +[2026-06-16 04:51:39,608][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.009627788327634335, 'train/grad_norm_pre_clip': 2.0542848110198975, 'timing/data': 0.03662371635437012, 'timing/model': 1.5468682367354631, 'learning_rate/qwen_vl_interface': 2.81111459486098e-06, 'learning_rate/action_model': 2.81111459486098e-05, 'epoch': 1.31}) +[2026-06-16 04:51:46,115][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.03601907938718796, 'train/grad_norm_pre_clip': 5.063807010650635, 'timing/data': 0.041439078748226166, 'timing/model': 1.6384079214185476, 'learning_rate/qwen_vl_interface': 2.8085017599336806e-06, 'learning_rate/action_model': 2.8085017599336806e-05, 'epoch': 1.31}) +[2026-06-16 04:51:52,626][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.032796118408441544, 'train/grad_norm_pre_clip': 11.11352825164795, 'timing/data': 0.04763488844037056, 'timing/model': 1.6319376770406961, 'learning_rate/qwen_vl_interface': 2.8058899286140828e-06, 'learning_rate/action_model': 2.805889928614083e-05, 'epoch': 1.31}) +[2026-06-16 04:51:59,109][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.08258849382400513, 'train/grad_norm_pre_clip': 13.428055763244629, 'timing/data': 0.045491116121411324, 'timing/model': 1.9002995640039444, 'learning_rate/qwen_vl_interface': 2.8032791019758127e-06, 'learning_rate/action_model': 2.8032791019758127e-05, 'epoch': 1.31}) +[2026-06-16 04:52:05,412][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.042490404099226, 'train/grad_norm_pre_clip': 4.704468727111816, 'timing/data': 0.03505774028599262, 'timing/model': 1.6200214009732008, 'learning_rate/qwen_vl_interface': 2.800669281092084e-06, 'learning_rate/action_model': 2.8006692810920838e-05, 'epoch': 1.31}) +[2026-06-16 04:52:11,840][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.061647962778806686, 'train/grad_norm_pre_clip': 10.925604820251465, 'timing/data': 0.031258558854460716, 'timing/model': 1.5242228507995605, 'learning_rate/qwen_vl_interface': 2.7980604670356938e-06, 'learning_rate/action_model': 2.7980604670356937e-05, 'epoch': 1.31}) +[2026-06-16 04:52:17,955][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.12408613413572311, 'train/grad_norm_pre_clip': 17.688783645629883, 'timing/data': 0.031717097386717796, 'timing/model': 1.5572670325636864, 'learning_rate/qwen_vl_interface': 2.795452660879027e-06, 'learning_rate/action_model': 2.7954526608790273e-05, 'epoch': 1.31}) +[2026-06-16 04:52:24,053][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.042884428054094315, 'train/grad_norm_pre_clip': 3.8585448265075684, 'timing/data': 0.0318459440022707, 'timing/model': 1.534345407038927, 'learning_rate/qwen_vl_interface': 2.7928458636940565e-06, 'learning_rate/action_model': 2.7928458636940564e-05, 'epoch': 1.32}) +[2026-06-16 04:52:30,148][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.05941575765609741, 'train/grad_norm_pre_clip': 4.749176025390625, 'timing/data': 0.03233468160033226, 'timing/model': 1.4938248079270124, 'learning_rate/qwen_vl_interface': 2.790240076552336e-06, 'learning_rate/action_model': 2.7902400765523355e-05, 'epoch': 1.32}) +[2026-06-16 05:01:26,046][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.09762661904096603, 'train/grad_norm_pre_clip': 11.843008995056152, 'eval/loss': 0.04603276029229164, 'eval/flap_precision': 0.9629120879120879, 'eval/flap_recall': 0.9511533242876526, 'eval/flap_f1': 0.9569965870307168, 'eval/flap_rate_model': 0.20222222222222222, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9569965870307169, 'timing/data': 0.030852101743221283, 'timing/model': 1.7524403370916843, 'learning_rate/qwen_vl_interface': 2.787635300525006e-06, 'learning_rate/action_model': 2.787635300525006e-05, 'epoch': 1.32}) +[2026-06-16 05:01:26,047][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3400_state +[2026-06-16 05:01:26,051][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 05:01:32,037][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3400_state/model.safetensors +[2026-06-16 05:01:41,433][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3400_state/optimizer.bin +[2026-06-16 05:01:41,434][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3400_state/sampler.bin +[2026-06-16 05:01:41,434][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3400_state/sampler_1.bin +[2026-06-16 05:01:41,435][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3400_state/random_states_0.pkl +[2026-06-16 05:01:41,536][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3400 +[2026-06-16 05:01:41,537][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 05:01:41,545][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 05:01:47,504][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.021667877212166786, 'train/grad_norm_pre_clip': 4.089461803436279, 'timing/data': 0.027016369625926018, 'timing/model': 1.64440350048244, 'learning_rate/qwen_vl_interface': 2.7850315366827933e-06, 'learning_rate/action_model': 2.7850315366827934e-05, 'epoch': 1.32}) +[2026-06-16 05:01:53,571][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.03869963809847832, 'train/grad_norm_pre_clip': 4.79851770401001, 'timing/data': 0.032087091356515884, 'timing/model': 1.7252616938203573, 'learning_rate/qwen_vl_interface': 2.7824287860960076e-06, 'learning_rate/action_model': 2.7824287860960075e-05, 'epoch': 1.32}) +[2026-06-16 05:01:59,654][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.034368451684713364, 'train/grad_norm_pre_clip': 6.378804683685303, 'timing/data': 0.03232160583138466, 'timing/model': 1.5276154670864344, 'learning_rate/qwen_vl_interface': 2.7798270498345377e-06, 'learning_rate/action_model': 2.7798270498345376e-05, 'epoch': 1.32}) +[2026-06-16 05:02:06,160][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.012005959637463093, 'train/grad_norm_pre_clip': 2.643770694732666, 'timing/data': 0.03299279324710369, 'timing/model': 1.8111396674066782, 'learning_rate/qwen_vl_interface': 2.777226328967861e-06, 'learning_rate/action_model': 2.7772263289678607e-05, 'epoch': 1.32}) +[2026-06-16 05:02:12,520][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.04154352471232414, 'train/grad_norm_pre_clip': 11.680274963378906, 'timing/data': 0.026716554537415504, 'timing/model': 1.6230043601244688, 'learning_rate/qwen_vl_interface': 2.774626624565035e-06, 'learning_rate/action_model': 2.774626624565035e-05, 'epoch': 1.32}) +[2026-06-16 05:02:18,401][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.013007743284106255, 'train/grad_norm_pre_clip': 2.5159754753112793, 'timing/data': 0.02505507506430149, 'timing/model': 1.4416876453906298, 'learning_rate/qwen_vl_interface': 2.772027937694701e-06, 'learning_rate/action_model': 2.772027937694701e-05, 'epoch': 1.32}) +[2026-06-16 05:02:24,604][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.0024533583782613277, 'train/grad_norm_pre_clip': 0.383470743894577, 'timing/data': 0.029607806354761124, 'timing/model': 1.4489007201045752, 'learning_rate/qwen_vl_interface': 2.7694302694250806e-06, 'learning_rate/action_model': 2.7694302694250808e-05, 'epoch': 1.32}) +[2026-06-16 05:02:30,581][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.09467282891273499, 'train/grad_norm_pre_clip': 6.873831748962402, 'timing/data': 0.03007569909095764, 'timing/model': 1.7059320528060198, 'learning_rate/qwen_vl_interface': 2.7668336208239775e-06, 'learning_rate/action_model': 2.7668336208239776e-05, 'epoch': 1.32}) +[2026-06-16 05:02:36,400][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.007804617285728455, 'train/grad_norm_pre_clip': 1.1160672903060913, 'timing/data': 0.03232588618993759, 'timing/model': 1.520234253257513, 'learning_rate/qwen_vl_interface': 2.7642379929587755e-06, 'learning_rate/action_model': 2.7642379929587754e-05, 'epoch': 1.32}) +[2026-06-16 05:02:42,280][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.007190991193056107, 'train/grad_norm_pre_clip': 2.8003013134002686, 'timing/data': 0.02855011075735092, 'timing/model': 1.4366568513214588, 'learning_rate/qwen_vl_interface': 2.7616433868964387e-06, 'learning_rate/action_model': 2.7616433868964388e-05, 'epoch': 1.32}) +[2026-06-16 05:02:48,278][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.08472483605146408, 'train/grad_norm_pre_clip': 9.033738136291504, 'timing/data': 0.029395584017038345, 'timing/model': 1.4277084488421679, 'learning_rate/qwen_vl_interface': 2.7590498037035122e-06, 'learning_rate/action_model': 2.7590498037035124e-05, 'epoch': 1.32}) +[2026-06-16 05:02:54,076][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.017202157527208328, 'train/grad_norm_pre_clip': 3.8198273181915283, 'timing/data': 0.02582673355937004, 'timing/model': 1.4253535699099302, 'learning_rate/qwen_vl_interface': 2.7564572444461218e-06, 'learning_rate/action_model': 2.7564572444461217e-05, 'epoch': 1.32}) +[2026-06-16 05:02:59,977][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.04251781851053238, 'train/grad_norm_pre_clip': 12.47391414642334, 'timing/data': 0.027296477928757668, 'timing/model': 1.442662438377738, 'learning_rate/qwen_vl_interface': 2.7538657101899674e-06, 'learning_rate/action_model': 2.7538657101899673e-05, 'epoch': 1.32}) +[2026-06-16 05:03:06,200][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.191658616065979, 'train/grad_norm_pre_clip': 15.32114028930664, 'timing/data': 0.03162275440990925, 'timing/model': 1.751983217895031, 'learning_rate/qwen_vl_interface': 2.751275202000332e-06, 'learning_rate/action_model': 2.7512752020003317e-05, 'epoch': 1.32}) +[2026-06-16 05:03:12,285][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.1596653163433075, 'train/grad_norm_pre_clip': 19.014694213867188, 'timing/data': 0.032511573284864426, 'timing/model': 1.5551581885665655, 'learning_rate/qwen_vl_interface': 2.7486857209420763e-06, 'learning_rate/action_model': 2.7486857209420763e-05, 'epoch': 1.32}) +[2026-06-16 05:03:18,567][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.10461997985839844, 'train/grad_norm_pre_clip': 16.51788902282715, 'timing/data': 0.03215792216360569, 'timing/model': 1.5347756836563349, 'learning_rate/qwen_vl_interface': 2.7460972680796377e-06, 'learning_rate/action_model': 2.7460972680796375e-05, 'epoch': 1.32}) +[2026-06-16 05:03:24,649][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.22398848831653595, 'train/grad_norm_pre_clip': 20.152645111083984, 'timing/data': 0.032444750890135765, 'timing/model': 1.5520594995468855, 'learning_rate/qwen_vl_interface': 2.7435098444770314e-06, 'learning_rate/action_model': 2.7435098444770315e-05, 'epoch': 1.32}) +[2026-06-16 05:03:30,737][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.056101590394973755, 'train/grad_norm_pre_clip': 11.511207580566406, 'timing/data': 0.03114662691950798, 'timing/model': 1.5418574921786785, 'learning_rate/qwen_vl_interface': 2.7409234511978495e-06, 'learning_rate/action_model': 2.7409234511978492e-05, 'epoch': 1.32}) +[2026-06-16 05:03:36,840][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.06568986922502518, 'train/grad_norm_pre_clip': 9.055078506469727, 'timing/data': 0.03152225911617279, 'timing/model': 1.5259532108902931, 'learning_rate/qwen_vl_interface': 2.7383380893052606e-06, 'learning_rate/action_model': 2.7383380893052608e-05, 'epoch': 1.32}) +[2026-06-16 05:03:42,955][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.029864732176065445, 'train/grad_norm_pre_clip': 7.697659015655518, 'timing/data': 0.0313020721077919, 'timing/model': 1.7591639962047338, 'learning_rate/qwen_vl_interface': 2.73575375986201e-06, 'learning_rate/action_model': 2.7357537598620096e-05, 'epoch': 1.32}) +[2026-06-16 05:03:48,957][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.062441904097795486, 'train/grad_norm_pre_clip': 9.835616111755371, 'timing/data': 0.029789326712489128, 'timing/model': 1.4994021486490965, 'learning_rate/qwen_vl_interface': 2.733170463930417e-06, 'learning_rate/action_model': 2.7331704639304167e-05, 'epoch': 1.32}) +[2026-06-16 05:03:55,083][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.053185462951660156, 'train/grad_norm_pre_clip': 39.958229064941406, 'timing/data': 0.032868875190615654, 'timing/model': 1.5713521931320429, 'learning_rate/qwen_vl_interface': 2.7305882025723774e-06, 'learning_rate/action_model': 2.7305882025723774e-05, 'epoch': 1.32}) +[2026-06-16 05:04:01,375][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.03204302117228508, 'train/grad_norm_pre_clip': 4.568165302276611, 'timing/data': 0.033139484003186226, 'timing/model': 1.5401208084076643, 'learning_rate/qwen_vl_interface': 2.7280069768493626e-06, 'learning_rate/action_model': 2.7280069768493628e-05, 'epoch': 1.32}) +[2026-06-16 05:04:07,770][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.1347576379776001, 'train/grad_norm_pre_clip': 11.635139465332031, 'timing/data': 0.032975537702441216, 'timing/model': 1.539212130010128, 'learning_rate/qwen_vl_interface': 2.7254267878224144e-06, 'learning_rate/action_model': 2.7254267878224143e-05, 'epoch': 1.33}) +[2026-06-16 05:04:14,210][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.004543286748230457, 'train/grad_norm_pre_clip': 1.7854360342025757, 'timing/data': 0.036794388666749, 'timing/model': 1.5753629729151726, 'learning_rate/qwen_vl_interface': 2.7228476365521524e-06, 'learning_rate/action_model': 2.722847636552152e-05, 'epoch': 1.33}) +[2026-06-16 05:04:20,650][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.006009765900671482, 'train/grad_norm_pre_clip': 1.716072678565979, 'timing/data': 0.03214294649660587, 'timing/model': 1.8711073566228151, 'learning_rate/qwen_vl_interface': 2.7202695240987674e-06, 'learning_rate/action_model': 2.7202695240987676e-05, 'epoch': 1.33}) +[2026-06-16 05:04:26,793][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.11535868048667908, 'train/grad_norm_pre_clip': 11.807537078857422, 'timing/data': 0.03184143267571926, 'timing/model': 1.6782861277461052, 'learning_rate/qwen_vl_interface': 2.717692451522025e-06, 'learning_rate/action_model': 2.7176924515220248e-05, 'epoch': 1.33}) +[2026-06-16 05:04:32,779][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.12447821348905563, 'train/grad_norm_pre_clip': 12.016464233398438, 'timing/data': 0.03211292810738087, 'timing/model': 1.5319417640566826, 'learning_rate/qwen_vl_interface': 2.7151164198812624e-06, 'learning_rate/action_model': 2.7151164198812622e-05, 'epoch': 1.33}) +[2026-06-16 05:04:39,072][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.23311497271060944, 'train/grad_norm_pre_clip': 26.424346923828125, 'timing/data': 0.032975105568766594, 'timing/model': 1.5155427306890488, 'learning_rate/qwen_vl_interface': 2.712541430235388e-06, 'learning_rate/action_model': 2.712541430235388e-05, 'epoch': 1.33}) +[2026-06-16 05:04:45,496][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.04032811522483826, 'train/grad_norm_pre_clip': 6.847994327545166, 'timing/data': 0.03600935451686382, 'timing/model': 1.6138176415115595, 'learning_rate/qwen_vl_interface': 2.7099674836428842e-06, 'learning_rate/action_model': 2.709967483642884e-05, 'epoch': 1.33}) +[2026-06-16 05:04:51,987][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.07584501057863235, 'train/grad_norm_pre_clip': 8.801444053649902, 'timing/data': 0.044144587591290474, 'timing/model': 1.5889318995177746, 'learning_rate/qwen_vl_interface': 2.7073945811618018e-06, 'learning_rate/action_model': 2.7073945811618017e-05, 'epoch': 1.33}) +[2026-06-16 05:04:58,286][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.0055230893194675446, 'train/grad_norm_pre_clip': 1.27440345287323, 'timing/data': 0.03190052509307861, 'timing/model': 1.7420137021690607, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 1.33}) +[2026-06-16 05:05:04,601][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.004240651614964008, 'train/grad_norm_pre_clip': 1.2879172563552856, 'timing/data': 0.03543262742459774, 'timing/model': 1.7476627863943577, 'learning_rate/qwen_vl_interface': 2.7022519127639687e-06, 'learning_rate/action_model': 2.7022519127639682e-05, 'epoch': 1.33}) +[2026-06-16 05:05:10,780][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.004984336905181408, 'train/grad_norm_pre_clip': 0.7004427313804626, 'timing/data': 0.031537262722849846, 'timing/model': 1.5978738106787205, 'learning_rate/qwen_vl_interface': 2.6996821489611713e-06, 'learning_rate/action_model': 2.699682148961171e-05, 'epoch': 1.33}) +[2026-06-16 05:05:17,305][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.0016753548989072442, 'train/grad_norm_pre_clip': 0.38141384720802307, 'timing/data': 0.0316806361079216, 'timing/model': 1.5389730744063854, 'learning_rate/qwen_vl_interface': 2.697113433497712e-06, 'learning_rate/action_model': 2.6971134334977115e-05, 'epoch': 1.33}) +[2026-06-16 05:05:23,700][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.042470164597034454, 'train/grad_norm_pre_clip': 8.485884666442871, 'timing/data': 0.024517331272363663, 'timing/model': 1.5215196963399649, 'learning_rate/qwen_vl_interface': 2.6945457674294863e-06, 'learning_rate/action_model': 2.694545767429486e-05, 'epoch': 1.33}) +[2026-06-16 05:05:29,717][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.09439972788095474, 'train/grad_norm_pre_clip': 14.691225051879883, 'timing/data': 0.026868751272559166, 'timing/model': 1.5246369391679764, 'learning_rate/qwen_vl_interface': 2.6919791518119676e-06, 'learning_rate/action_model': 2.6919791518119674e-05, 'epoch': 1.33}) +[2026-06-16 05:05:35,808][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.06694254279136658, 'train/grad_norm_pre_clip': 11.630400657653809, 'timing/data': 0.02553669922053814, 'timing/model': 1.7249557692557573, 'learning_rate/qwen_vl_interface': 2.6894135877001946e-06, 'learning_rate/action_model': 2.689413587700195e-05, 'epoch': 1.33}) +[2026-06-16 05:05:41,609][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.08982113748788834, 'train/grad_norm_pre_clip': 10.265841484069824, 'timing/data': 0.02572387084364891, 'timing/model': 1.5249513760209084, 'learning_rate/qwen_vl_interface': 2.686849076148775e-06, 'learning_rate/action_model': 2.6868490761487747e-05, 'epoch': 1.33}) +[2026-06-16 05:05:47,536][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.4720780849456787, 'train/grad_norm_pre_clip': 26.24314308166504, 'timing/data': 0.026517026126384735, 'timing/model': 1.4596113190054893, 'learning_rate/qwen_vl_interface': 2.68428561821188e-06, 'learning_rate/action_model': 2.6842856182118802e-05, 'epoch': 1.33}) +[2026-06-16 05:05:53,506][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.017361905425786972, 'train/grad_norm_pre_clip': 1.5283405780792236, 'timing/data': 0.025355882942676544, 'timing/model': 1.522997533902526, 'learning_rate/qwen_vl_interface': 2.681723214943253e-06, 'learning_rate/action_model': 2.681723214943253e-05, 'epoch': 1.33}) +[2026-06-16 05:05:59,731][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.003522315062582493, 'train/grad_norm_pre_clip': 1.8674458265304565, 'timing/data': 0.028873320668935776, 'timing/model': 1.6226481199264526, 'learning_rate/qwen_vl_interface': 2.6791618673961996e-06, 'learning_rate/action_model': 2.6791618673961994e-05, 'epoch': 1.33}) +[2026-06-16 05:06:05,624][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.13336560130119324, 'train/grad_norm_pre_clip': 27.65363311767578, 'timing/data': 0.026383915916085243, 'timing/model': 1.522555572912097, 'learning_rate/qwen_vl_interface': 2.676601576623594e-06, 'learning_rate/action_model': 2.6766015766235937e-05, 'epoch': 1.33}) +[2026-06-16 05:06:12,182][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.2524363100528717, 'train/grad_norm_pre_clip': 22.34577751159668, 'timing/data': 0.03416425921022892, 'timing/model': 1.9037330690771341, 'learning_rate/qwen_vl_interface': 2.674042343677874e-06, 'learning_rate/action_model': 2.6740423436778738e-05, 'epoch': 1.33}) +[2026-06-16 05:06:18,610][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.05605538189411163, 'train/grad_norm_pre_clip': 16.870058059692383, 'timing/data': 0.037839945405721664, 'timing/model': 1.6472891960293055, 'learning_rate/qwen_vl_interface': 2.6714841696110456e-06, 'learning_rate/action_model': 2.6714841696110456e-05, 'epoch': 1.33}) +[2026-06-16 05:06:25,421][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.07482212036848068, 'train/grad_norm_pre_clip': 12.00326156616211, 'timing/data': 0.043514665216207504, 'timing/model': 1.6048436276614666, 'learning_rate/qwen_vl_interface': 2.6689270554746743e-06, 'learning_rate/action_model': 2.6689270554746743e-05, 'epoch': 1.33}) +[2026-06-16 05:06:31,784][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.12535974383354187, 'train/grad_norm_pre_clip': 9.705402374267578, 'timing/data': 0.03178497590124607, 'timing/model': 1.5879693012684584, 'learning_rate/qwen_vl_interface': 2.6663710023198956e-06, 'learning_rate/action_model': 2.6663710023198956e-05, 'epoch': 1.33}) +[2026-06-16 05:06:38,277][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.10828377306461334, 'train/grad_norm_pre_clip': 13.347433090209961, 'timing/data': 0.03207653388381004, 'timing/model': 1.571698671206832, 'learning_rate/qwen_vl_interface': 2.663816011197405e-06, 'learning_rate/action_model': 2.6638160111974048e-05, 'epoch': 1.33}) +[2026-06-16 05:06:44,726][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.12979769706726074, 'train/grad_norm_pre_clip': 15.458871841430664, 'timing/data': 0.038716476410627365, 'timing/model': 1.6175491325557232, 'learning_rate/qwen_vl_interface': 2.661262083157463e-06, 'learning_rate/action_model': 2.661262083157463e-05, 'epoch': 1.33}) +[2026-06-16 05:06:51,302][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.03635923191905022, 'train/grad_norm_pre_clip': 7.562878131866455, 'timing/data': 0.03474353440105915, 'timing/model': 1.8326252475380898, 'learning_rate/qwen_vl_interface': 2.6587092192498943e-06, 'learning_rate/action_model': 2.658709219249894e-05, 'epoch': 1.34}) +[2026-06-16 05:06:57,501][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.017148585990071297, 'train/grad_norm_pre_clip': 5.856657028198242, 'timing/data': 0.04262189567089081, 'timing/model': 1.586225738748908, 'learning_rate/qwen_vl_interface': 2.656157420524084e-06, 'learning_rate/action_model': 2.6561574205240842e-05, 'epoch': 1.34}) +[2026-06-16 05:07:04,091][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.29282641410827637, 'train/grad_norm_pre_clip': 16.787431716918945, 'timing/data': 0.03251222334802151, 'timing/model': 1.649316892027855, 'learning_rate/qwen_vl_interface': 2.6536066880289795e-06, 'learning_rate/action_model': 2.6536066880289794e-05, 'epoch': 1.34}) +[2026-06-16 05:07:10,905][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.007834980264306068, 'train/grad_norm_pre_clip': 3.000375986099243, 'timing/data': 0.045257365331053734, 'timing/model': 1.5977629572153091, 'learning_rate/qwen_vl_interface': 2.6510570228130916e-06, 'learning_rate/action_model': 2.6510570228130914e-05, 'epoch': 1.34}) +[2026-06-16 05:07:17,241][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.011704623699188232, 'train/grad_norm_pre_clip': 1.7532601356506348, 'timing/data': 0.03516756184399128, 'timing/model': 1.4938551746308804, 'learning_rate/qwen_vl_interface': 2.648508425924491e-06, 'learning_rate/action_model': 2.6485084259244907e-05, 'epoch': 1.34}) +[2026-06-16 05:07:23,459][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.13519105315208435, 'train/grad_norm_pre_clip': 14.011655807495117, 'timing/data': 0.031687404960393906, 'timing/model': 1.7615163810551167, 'learning_rate/qwen_vl_interface': 2.6459608984108093e-06, 'learning_rate/action_model': 2.645960898410809e-05, 'epoch': 1.34}) +[2026-06-16 05:07:29,185][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.007026989944279194, 'train/grad_norm_pre_clip': 1.2214857339859009, 'timing/data': 0.024132831022143364, 'timing/model': 1.442931242287159, 'learning_rate/qwen_vl_interface': 2.6434144413192414e-06, 'learning_rate/action_model': 2.6434144413192414e-05, 'epoch': 1.34}) +[2026-06-16 05:07:35,080][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.2799466550350189, 'train/grad_norm_pre_clip': 15.843082427978516, 'timing/data': 0.024836987257003784, 'timing/model': 1.4378885123878717, 'learning_rate/qwen_vl_interface': 2.6408690556965347e-06, 'learning_rate/action_model': 2.640869055696535e-05, 'epoch': 1.34}) +[2026-06-16 05:07:40,882][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.0675821602344513, 'train/grad_norm_pre_clip': 15.201006889343262, 'timing/data': 0.025882862508296967, 'timing/model': 1.437621358782053, 'learning_rate/qwen_vl_interface': 2.6383247425890047e-06, 'learning_rate/action_model': 2.6383247425890047e-05, 'epoch': 1.34}) +[2026-06-16 05:07:46,685][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.06071613356471062, 'train/grad_norm_pre_clip': 11.654953002929688, 'timing/data': 0.026691820472478867, 'timing/model': 1.439568990841508, 'learning_rate/qwen_vl_interface': 2.6357815030425204e-06, 'learning_rate/action_model': 2.6357815030425205e-05, 'epoch': 1.34}) +[2026-06-16 05:07:52,492][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.08912376314401627, 'train/grad_norm_pre_clip': 12.75326919555664, 'timing/data': 0.023877425119280815, 'timing/model': 1.4442722648382187, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 1.34}) +[2026-06-16 05:07:58,373][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.0349297821521759, 'train/grad_norm_pre_clip': 5.384091377258301, 'timing/data': 0.02564559504389763, 'timing/model': 1.6282734293490648, 'learning_rate/qwen_vl_interface': 2.6306982488139704e-06, 'learning_rate/action_model': 2.6306982488139702e-05, 'epoch': 1.34}) +[2026-06-16 05:08:04,050][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.0011008200235664845, 'train/grad_norm_pre_clip': 0.18069112300872803, 'timing/data': 0.027244852855801582, 'timing/model': 1.5125727616250515, 'learning_rate/qwen_vl_interface': 2.628158236221438e-06, 'learning_rate/action_model': 2.628158236221438e-05, 'epoch': 1.34}) +[2026-06-16 05:08:10,255][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.14071765542030334, 'train/grad_norm_pre_clip': 17.352319717407227, 'timing/data': 0.033014703541994095, 'timing/model': 1.5420382618904114, 'learning_rate/qwen_vl_interface': 2.6256193013690183e-06, 'learning_rate/action_model': 2.6256193013690182e-05, 'epoch': 1.34}) +[2026-06-16 05:08:16,345][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.20987392961978912, 'train/grad_norm_pre_clip': 30.57581329345703, 'timing/data': 0.032077616080641747, 'timing/model': 1.535780157893896, 'learning_rate/qwen_vl_interface': 2.623081445300373e-06, 'learning_rate/action_model': 2.6230814453003728e-05, 'epoch': 1.34}) +[2026-06-16 05:08:22,569][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.012725764885544777, 'train/grad_norm_pre_clip': 3.884654998779297, 'timing/data': 0.03167632967233658, 'timing/model': 1.5593494977802038, 'learning_rate/qwen_vl_interface': 2.6205446690587183e-06, 'learning_rate/action_model': 2.6205446690587182e-05, 'epoch': 1.34}) +[2026-06-16 05:08:28,764][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.0016684102592989802, 'train/grad_norm_pre_clip': 0.3459664583206177, 'timing/data': 0.03174957074224949, 'timing/model': 1.5331563521176577, 'learning_rate/qwen_vl_interface': 2.618008973686828e-06, 'learning_rate/action_model': 2.618008973686828e-05, 'epoch': 1.34}) +[2026-06-16 05:08:34,990][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.02294243685901165, 'train/grad_norm_pre_clip': 6.115611553192139, 'timing/data': 0.03209041431546211, 'timing/model': 1.7725567575544119, 'learning_rate/qwen_vl_interface': 2.615474360227028e-06, 'learning_rate/action_model': 2.615474360227028e-05, 'epoch': 1.34}) +[2026-06-16 05:08:40,945][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.001204920350573957, 'train/grad_norm_pre_clip': 0.20176051557064056, 'timing/data': 0.031000878661870956, 'timing/model': 1.5210252776741982, 'learning_rate/qwen_vl_interface': 2.612940829721204e-06, 'learning_rate/action_model': 2.612940829721204e-05, 'epoch': 1.34}) +[2026-06-16 05:08:47,146][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.06485652923583984, 'train/grad_norm_pre_clip': 5.735449314117432, 'timing/data': 0.03181697241961956, 'timing/model': 1.5498230569064617, 'learning_rate/qwen_vl_interface': 2.6104083832107948e-06, 'learning_rate/action_model': 2.6104083832107945e-05, 'epoch': 1.34}) +[2026-06-16 05:08:53,341][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.05351575091481209, 'train/grad_norm_pre_clip': 7.272963523864746, 'timing/data': 0.03084949217736721, 'timing/model': 1.5327203329652548, 'learning_rate/qwen_vl_interface': 2.607877021736793e-06, 'learning_rate/action_model': 2.6078770217367926e-05, 'epoch': 1.34}) +[2026-06-16 05:08:59,440][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.02740057185292244, 'train/grad_norm_pre_clip': 8.872489929199219, 'timing/data': 0.030734948813915253, 'timing/model': 1.4874334707856178, 'learning_rate/qwen_vl_interface': 2.605346746339745e-06, 'learning_rate/action_model': 2.605346746339745e-05, 'epoch': 1.34}) +[2026-06-16 05:09:05,542][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.11201480031013489, 'train/grad_norm_pre_clip': 16.532337188720703, 'timing/data': 0.031279781833291054, 'timing/model': 1.4958238117396832, 'learning_rate/qwen_vl_interface': 2.602817558059753e-06, 'learning_rate/action_model': 2.6028175580597532e-05, 'epoch': 1.34}) +[2026-06-16 05:09:11,655][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.1597575545310974, 'train/grad_norm_pre_clip': 15.02953052520752, 'timing/data': 0.03118312358856201, 'timing/model': 1.724480563774705, 'learning_rate/qwen_vl_interface': 2.60028945793647e-06, 'learning_rate/action_model': 2.60028945793647e-05, 'epoch': 1.34}) +[2026-06-16 05:09:17,283][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.10634321719408035, 'train/grad_norm_pre_clip': 6.607079982757568, 'timing/data': 0.027050526812672615, 'timing/model': 1.4360877480357885, 'learning_rate/qwen_vl_interface': 2.5977624470091024e-06, 'learning_rate/action_model': 2.5977624470091023e-05, 'epoch': 1.34}) +[2026-06-16 05:09:23,387][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.004304702393710613, 'train/grad_norm_pre_clip': 1.2068607807159424, 'timing/data': 0.027273127809166908, 'timing/model': 1.4308655932545662, 'learning_rate/qwen_vl_interface': 2.59523652631641e-06, 'learning_rate/action_model': 2.5952365263164097e-05, 'epoch': 1.34}) +[2026-06-16 05:09:29,517][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.09293540567159653, 'train/grad_norm_pre_clip': 4.899598121643066, 'timing/data': 0.026943389326334, 'timing/model': 1.5354073103517294, 'learning_rate/qwen_vl_interface': 2.5927116968967024e-06, 'learning_rate/action_model': 2.5927116968967025e-05, 'epoch': 1.35}) +[2026-06-16 05:09:35,496][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.1569817215204239, 'train/grad_norm_pre_clip': 35.86294174194336, 'timing/data': 0.027306178584694862, 'timing/model': 1.5113105550408363, 'learning_rate/qwen_vl_interface': 2.5901879597878417e-06, 'learning_rate/action_model': 2.5901879597878415e-05, 'epoch': 1.35}) +[2026-06-16 05:09:41,518][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.24273398518562317, 'train/grad_norm_pre_clip': 6.487544059753418, 'timing/data': 0.02556683123111725, 'timing/model': 1.507154431194067, 'learning_rate/qwen_vl_interface': 2.5876653160272395e-06, 'learning_rate/action_model': 2.5876653160272397e-05, 'epoch': 1.35}) +[2026-06-16 05:09:47,493][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.33607035875320435, 'train/grad_norm_pre_clip': 17.37164306640625, 'timing/data': 0.02581092342734337, 'timing/model': 1.7067264970391989, 'learning_rate/qwen_vl_interface': 2.5851437666518607e-06, 'learning_rate/action_model': 2.5851437666518607e-05, 'epoch': 1.35}) +[2026-06-16 05:09:53,307][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.12207558751106262, 'train/grad_norm_pre_clip': 18.837169647216797, 'timing/data': 0.02549690380692482, 'timing/model': 1.5121518727391958, 'learning_rate/qwen_vl_interface': 2.5826233126982198e-06, 'learning_rate/action_model': 2.5826233126982196e-05, 'epoch': 1.35}) +[2026-06-16 05:09:59,290][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.08724477887153625, 'train/grad_norm_pre_clip': 4.569446086883545, 'timing/data': 0.02497703582048416, 'timing/model': 1.5063764546066523, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 1.35}) +[2026-06-16 05:10:05,220][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.025144554674625397, 'train/grad_norm_pre_clip': 6.928483486175537, 'timing/data': 0.026455383747816086, 'timing/model': 1.4464520364999771, 'learning_rate/qwen_vl_interface': 2.577585695199954e-06, 'learning_rate/action_model': 2.5775856951999538e-05, 'epoch': 1.35}) +[2026-06-16 05:10:11,218][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.04625317081809044, 'train/grad_norm_pre_clip': 17.53619384765625, 'timing/data': 0.026047291234135628, 'timing/model': 1.5319279991090298, 'learning_rate/qwen_vl_interface': 2.575068533726101e-06, 'learning_rate/action_model': 2.575068533726101e-05, 'epoch': 1.35}) +[2026-06-16 05:10:17,210][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.040611717849969864, 'train/grad_norm_pre_clip': 10.307327270507812, 'timing/data': 0.02666143886744976, 'timing/model': 1.514688590541482, 'learning_rate/qwen_vl_interface': 2.5725524718155333e-06, 'learning_rate/action_model': 2.572552471815533e-05, 'epoch': 1.35}) +[2026-06-16 05:10:23,374][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.19088074564933777, 'train/grad_norm_pre_clip': 23.462745666503906, 'timing/data': 0.026102587580680847, 'timing/model': 1.8961722441017628, 'learning_rate/qwen_vl_interface': 2.570037510502508e-06, 'learning_rate/action_model': 2.5700375105025083e-05, 'epoch': 1.35}) +[2026-06-16 05:10:29,332][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.0033611610997468233, 'train/grad_norm_pre_clip': 0.6952448487281799, 'timing/data': 0.03226293995976448, 'timing/model': 1.5324396435171366, 'learning_rate/qwen_vl_interface': 2.5675236508208313e-06, 'learning_rate/action_model': 2.5675236508208313e-05, 'epoch': 1.35}) +[2026-06-16 05:10:35,537][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.023845477029681206, 'train/grad_norm_pre_clip': 4.298908710479736, 'timing/data': 0.030914535745978355, 'timing/model': 1.5247327238321304, 'learning_rate/qwen_vl_interface': 2.565010893803856e-06, 'learning_rate/action_model': 2.565010893803856e-05, 'epoch': 1.35}) +[2026-06-16 05:10:41,631][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.16866572201251984, 'train/grad_norm_pre_clip': 10.44671630859375, 'timing/data': 0.03148560971021652, 'timing/model': 1.5175987742841244, 'learning_rate/qwen_vl_interface': 2.562499240484482e-06, 'learning_rate/action_model': 2.5624992404844816e-05, 'epoch': 1.35}) +[2026-06-16 05:10:47,742][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.03558903932571411, 'train/grad_norm_pre_clip': 13.697723388671875, 'timing/data': 0.03173116035759449, 'timing/model': 1.537540651857853, 'learning_rate/qwen_vl_interface': 2.559988691895155e-06, 'learning_rate/action_model': 2.559988691895155e-05, 'epoch': 1.35}) +[2026-06-16 05:10:53,851][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.15402168035507202, 'train/grad_norm_pre_clip': 28.509199142456055, 'timing/data': 0.03279984928667545, 'timing/model': 1.4928167890757322, 'learning_rate/qwen_vl_interface': 2.5574792490678648e-06, 'learning_rate/action_model': 2.5574792490678644e-05, 'epoch': 1.35}) +[2026-06-16 05:10:59,966][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.1343747079372406, 'train/grad_norm_pre_clip': 16.479198455810547, 'timing/data': 0.03237288258969784, 'timing/model': 1.7474145386368036, 'learning_rate/qwen_vl_interface': 2.5549709130341493e-06, 'learning_rate/action_model': 2.5549709130341493e-05, 'epoch': 1.35}) +[2026-06-16 05:11:05,941][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.02308540605008602, 'train/grad_norm_pre_clip': 3.9434916973114014, 'timing/data': 0.03133873641490936, 'timing/model': 1.5354130156338215, 'learning_rate/qwen_vl_interface': 2.5524636848250924e-06, 'learning_rate/action_model': 2.552463684825092e-05, 'epoch': 1.35}) +[2026-06-16 05:11:11,886][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.1864088624715805, 'train/grad_norm_pre_clip': 20.213367462158203, 'timing/data': 0.025577174499630928, 'timing/model': 1.4347590636461973, 'learning_rate/qwen_vl_interface': 2.5499575654713186e-06, 'learning_rate/action_model': 2.5499575654713185e-05, 'epoch': 1.35}) +[2026-06-16 05:11:18,140][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.08166114240884781, 'train/grad_norm_pre_clip': 5.289852142333984, 'timing/data': 0.031348977237939835, 'timing/model': 1.556083431467414, 'learning_rate/qwen_vl_interface': 2.547452556003001e-06, 'learning_rate/action_model': 2.547452556003001e-05, 'epoch': 1.35}) +[2026-06-16 05:11:24,243][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.06575171649456024, 'train/grad_norm_pre_clip': 14.993938446044922, 'timing/data': 0.03180031664669514, 'timing/model': 1.5308376383036375, 'learning_rate/qwen_vl_interface': 2.5449486574498535e-06, 'learning_rate/action_model': 2.5449486574498534e-05, 'epoch': 1.35}) +[2026-06-16 05:11:30,340][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.19965122640132904, 'train/grad_norm_pre_clip': 9.529091835021973, 'timing/data': 0.030782511457800865, 'timing/model': 1.4882103353738785, 'learning_rate/qwen_vl_interface': 2.5424458708411335e-06, 'learning_rate/action_model': 2.5424458708411335e-05, 'epoch': 1.35}) +[2026-06-16 05:11:36,455][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.004906924441456795, 'train/grad_norm_pre_clip': 1.1464548110961914, 'timing/data': 0.03205845318734646, 'timing/model': 1.7408404629677534, 'learning_rate/qwen_vl_interface': 2.5399441972056437e-06, 'learning_rate/action_model': 2.5399441972056438e-05, 'epoch': 1.35}) +[2026-06-16 05:11:42,603][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.01262312475591898, 'train/grad_norm_pre_clip': 1.8900595903396606, 'timing/data': 0.03312317095696926, 'timing/model': 1.708670249208808, 'learning_rate/qwen_vl_interface': 2.537443637571726e-06, 'learning_rate/action_model': 2.537443637571726e-05, 'epoch': 1.35}) +[2026-06-16 05:11:48,536][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.13137154281139374, 'train/grad_norm_pre_clip': 9.99681282043457, 'timing/data': 0.0327520240098238, 'timing/model': 1.520390061661601, 'learning_rate/qwen_vl_interface': 2.534944192967269e-06, 'learning_rate/action_model': 2.534944192967269e-05, 'epoch': 1.35}) +[2026-06-16 05:19:20,198][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.013764502480626106, 'train/grad_norm_pre_clip': 1.7785067558288574, 'eval/loss': 0.05133847892284393, 'eval/flap_precision': 0.9692737430167597, 'eval/flap_recall': 0.9416553595658074, 'eval/flap_f1': 0.9552649690295939, 'eval/flap_rate_model': 0.1988888888888889, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9552649690295939, 'timing/data': 0.030589627102017403, 'timing/model': 1.526369011029601, 'learning_rate/qwen_vl_interface': 2.532445864419695e-06, 'learning_rate/action_model': 2.5324458644196952e-05, 'epoch': 1.35}) +[2026-06-16 05:19:20,198][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3500_state +[2026-06-16 05:19:20,202][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 05:19:26,170][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3500_state/model.safetensors +[2026-06-16 05:19:36,456][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3500_state/optimizer.bin +[2026-06-16 05:19:36,457][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3500_state/sampler.bin +[2026-06-16 05:19:36,457][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-16 05:19:36,458][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-16 05:19:36,545][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-16 05:19:36,545][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 05:19:36,553][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 05:19:42,398][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.29795798659324646, 'train/grad_norm_pre_clip': 15.09940242767334, 'timing/data': 0.025703400373458862, 'timing/model': 1.4373018760234118, 'learning_rate/qwen_vl_interface': 2.5299486529559758e-06, 'learning_rate/action_model': 2.5299486529559756e-05, 'epoch': 1.35}) +[2026-06-16 05:19:48,199][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.006364407949149609, 'train/grad_norm_pre_clip': 1.164517879486084, 'timing/data': 0.02718278020620346, 'timing/model': 1.4330848045647144, 'learning_rate/qwen_vl_interface': 2.5274525596026202e-06, 'learning_rate/action_model': 2.52745255960262e-05, 'epoch': 1.36}) +[2026-06-16 05:19:54,108][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.03224094584584236, 'train/grad_norm_pre_clip': 4.265556335449219, 'timing/data': 0.02535596862435341, 'timing/model': 1.4411397147923708, 'learning_rate/qwen_vl_interface': 2.524957585385677e-06, 'learning_rate/action_model': 2.524957585385677e-05, 'epoch': 1.36}) +[2026-06-16 05:20:00,104][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.35306304693222046, 'train/grad_norm_pre_clip': 22.00201416015625, 'timing/data': 0.02675258368253708, 'timing/model': 1.519162019714713, 'learning_rate/qwen_vl_interface': 2.522463731330738e-06, 'learning_rate/action_model': 2.5224637313307374e-05, 'epoch': 1.36}) +[2026-06-16 05:20:06,098][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.04904618486762047, 'train/grad_norm_pre_clip': 7.553275108337402, 'timing/data': 0.026025256142020226, 'timing/model': 1.7079042978584766, 'learning_rate/qwen_vl_interface': 2.5199709984629294e-06, 'learning_rate/action_model': 2.5199709984629292e-05, 'epoch': 1.36}) +[2026-06-16 05:20:11,889][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.10093863308429718, 'train/grad_norm_pre_clip': 14.475329399108887, 'timing/data': 0.025892581790685654, 'timing/model': 1.5071121603250504, 'learning_rate/qwen_vl_interface': 2.517479387806922e-06, 'learning_rate/action_model': 2.517479387806922e-05, 'epoch': 1.36}) +[2026-06-16 05:20:17,819][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.04112325608730316, 'train/grad_norm_pre_clip': 5.583261966705322, 'timing/data': 0.02679450623691082, 'timing/model': 1.443235481157899, 'learning_rate/qwen_vl_interface': 2.5149889003869217e-06, 'learning_rate/action_model': 2.5149889003869214e-05, 'epoch': 1.36}) +[2026-06-16 05:20:23,803][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.00976861547678709, 'train/grad_norm_pre_clip': 2.006406784057617, 'timing/data': 0.02971551939845085, 'timing/model': 1.5182093922048807, 'learning_rate/qwen_vl_interface': 2.5124995372266736e-06, 'learning_rate/action_model': 2.5124995372266735e-05, 'epoch': 1.36}) +[2026-06-16 05:20:29,910][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.016189096495509148, 'train/grad_norm_pre_clip': 3.781559705734253, 'timing/data': 0.03333084657788277, 'timing/model': 1.5167958196252584, 'learning_rate/qwen_vl_interface': 2.5100112993494626e-06, 'learning_rate/action_model': 2.5100112993494624e-05, 'epoch': 1.36}) +[2026-06-16 05:20:35,909][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.012908780947327614, 'train/grad_norm_pre_clip': 4.916750431060791, 'timing/data': 0.02694038115441799, 'timing/model': 1.5205039028078318, 'learning_rate/qwen_vl_interface': 2.5075241877781085e-06, 'learning_rate/action_model': 2.5075241877781086e-05, 'epoch': 1.36}) +[2026-06-16 05:20:41,771][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.006965996697545052, 'train/grad_norm_pre_clip': 1.6971276998519897, 'timing/data': 0.025643091648817062, 'timing/model': 1.5985874198377132, 'learning_rate/qwen_vl_interface': 2.5050382035349673e-06, 'learning_rate/action_model': 2.5050382035349674e-05, 'epoch': 1.36}) +[2026-06-16 05:20:47,565][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.10501769930124283, 'train/grad_norm_pre_clip': 8.642290115356445, 'timing/data': 0.026774588972330093, 'timing/model': 1.616783456876874, 'learning_rate/qwen_vl_interface': 2.5025533476419343e-06, 'learning_rate/action_model': 2.5025533476419343e-05, 'epoch': 1.36}) +[2026-06-16 05:20:53,179][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.03982183709740639, 'train/grad_norm_pre_clip': 10.272547721862793, 'timing/data': 0.026729736477136612, 'timing/model': 1.4325553718954325, 'learning_rate/qwen_vl_interface': 2.5000696211204394e-06, 'learning_rate/action_model': 2.5000696211204394e-05, 'epoch': 1.36}) +[2026-06-16 05:20:58,978][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.39680105447769165, 'train/grad_norm_pre_clip': 25.913124084472656, 'timing/data': 0.027112144976854324, 'timing/model': 1.420696308836341, 'learning_rate/qwen_vl_interface': 2.4975870249914495e-06, 'learning_rate/action_model': 2.4975870249914497e-05, 'epoch': 1.36}) +[2026-06-16 05:21:04,879][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.01692153699696064, 'train/grad_norm_pre_clip': 2.3098666667938232, 'timing/data': 0.026160191744565964, 'timing/model': 1.4291201755404472, 'learning_rate/qwen_vl_interface': 2.4951055602754664e-06, 'learning_rate/action_model': 2.4951055602754665e-05, 'epoch': 1.36}) +[2026-06-16 05:21:10,790][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.01099257729947567, 'train/grad_norm_pre_clip': 1.5404325723648071, 'timing/data': 0.02497319132089615, 'timing/model': 1.4191853683441877, 'learning_rate/qwen_vl_interface': 2.4926252279925255e-06, 'learning_rate/action_model': 2.4926252279925254e-05, 'epoch': 1.36}) +[2026-06-16 05:21:16,879][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.01275462843477726, 'train/grad_norm_pre_clip': 2.0372703075408936, 'timing/data': 0.02602117881178856, 'timing/model': 1.623821560293436, 'learning_rate/qwen_vl_interface': 2.490146029162199e-06, 'learning_rate/action_model': 2.4901460291621988e-05, 'epoch': 1.36}) +[2026-06-16 05:21:22,585][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.20024853944778442, 'train/grad_norm_pre_clip': 14.649543762207031, 'timing/data': 0.026491081342101097, 'timing/model': 1.4360698238015175, 'learning_rate/qwen_vl_interface': 2.4876679648035905e-06, 'learning_rate/action_model': 2.4876679648035903e-05, 'epoch': 1.36}) +[2026-06-16 05:21:28,499][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.032179806381464005, 'train/grad_norm_pre_clip': 8.903793334960938, 'timing/data': 0.027480782940983772, 'timing/model': 1.439980486407876, 'learning_rate/qwen_vl_interface': 2.4851910359353392e-06, 'learning_rate/action_model': 2.485191035935339e-05, 'epoch': 1.36}) +[2026-06-16 05:21:34,787][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.14154602587223053, 'train/grad_norm_pre_clip': 28.16595458984375, 'timing/data': 0.02594747580587864, 'timing/model': 1.4348934926092625, 'learning_rate/qwen_vl_interface': 2.4827152435756185e-06, 'learning_rate/action_model': 2.4827152435756185e-05, 'epoch': 1.36}) +[2026-06-16 05:21:40,681][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.06462815403938293, 'train/grad_norm_pre_clip': 10.358720779418945, 'timing/data': 0.025393182411789894, 'timing/model': 1.4231560919433832, 'learning_rate/qwen_vl_interface': 2.4802405887421307e-06, 'learning_rate/action_model': 2.4802405887421306e-05, 'epoch': 1.36}) +[2026-06-16 05:21:46,487][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.00969261396676302, 'train/grad_norm_pre_clip': 2.744536876678467, 'timing/data': 0.02554665319621563, 'timing/model': 1.4209367483854294, 'learning_rate/qwen_vl_interface': 2.4777670724521134e-06, 'learning_rate/action_model': 2.4777670724521134e-05, 'epoch': 1.36}) +[2026-06-16 05:21:52,311][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.000274973688647151, 'train/grad_norm_pre_clip': 0.036247074604034424, 'timing/data': 0.026643695309758186, 'timing/model': 1.5522739607840776, 'learning_rate/qwen_vl_interface': 2.475294695722337e-06, 'learning_rate/action_model': 2.475294695722337e-05, 'epoch': 1.36}) +[2026-06-16 05:21:57,987][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.10830648988485336, 'train/grad_norm_pre_clip': 5.930959224700928, 'timing/data': 0.02739257924258709, 'timing/model': 1.4349729474633932, 'learning_rate/qwen_vl_interface': 2.4728234595691006e-06, 'learning_rate/action_model': 2.4728234595691005e-05, 'epoch': 1.36}) +[2026-06-16 05:22:03,884][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.13143958151340485, 'train/grad_norm_pre_clip': 21.626535415649414, 'timing/data': 0.02601027674973011, 'timing/model': 1.430670564994216, 'learning_rate/qwen_vl_interface': 2.4703533650082368e-06, 'learning_rate/action_model': 2.4703533650082366e-05, 'epoch': 1.36}) +[2026-06-16 05:22:09,782][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.12518641352653503, 'train/grad_norm_pre_clip': 14.73061466217041, 'timing/data': 0.02592560276389122, 'timing/model': 1.4267854634672403, 'learning_rate/qwen_vl_interface': 2.4678844130551103e-06, 'learning_rate/action_model': 2.4678844130551103e-05, 'epoch': 1.36}) +[2026-06-16 05:22:15,679][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.16916489601135254, 'train/grad_norm_pre_clip': 6.79030179977417, 'timing/data': 0.02591768093407154, 'timing/model': 1.4210404753684998, 'learning_rate/qwen_vl_interface': 2.4654166047246102e-06, 'learning_rate/action_model': 2.46541660472461e-05, 'epoch': 1.37}) +[2026-06-16 05:22:21,579][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.09476927667856216, 'train/grad_norm_pre_clip': 19.593229293823242, 'timing/data': 0.026671184226870537, 'timing/model': 1.4572637993842363, 'learning_rate/qwen_vl_interface': 2.4629499410311606e-06, 'learning_rate/action_model': 2.462949941031161e-05, 'epoch': 1.37}) +[2026-06-16 05:22:27,413][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.2826462984085083, 'train/grad_norm_pre_clip': 12.64541244506836, 'timing/data': 0.025660302489995956, 'timing/model': 1.5723031368106604, 'learning_rate/qwen_vl_interface': 2.460484422988715e-06, 'learning_rate/action_model': 2.4604844229887146e-05, 'epoch': 1.37}) +[2026-06-16 05:22:33,088][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.10592342168092728, 'train/grad_norm_pre_clip': 13.59504222869873, 'timing/data': 0.02751368097960949, 'timing/model': 1.4399624355137348, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 1.37}) +[2026-06-16 05:22:38,974][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.19124993681907654, 'train/grad_norm_pre_clip': 14.692037582397461, 'timing/data': 0.02605385147035122, 'timing/model': 1.4204930402338505, 'learning_rate/qwen_vl_interface': 2.455556827910286e-06, 'learning_rate/action_model': 2.455556827910286e-05, 'epoch': 1.37}) +[2026-06-16 05:22:44,778][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.025182852521538734, 'train/grad_norm_pre_clip': 8.164319038391113, 'timing/data': 0.026418883353471756, 'timing/model': 1.437576787546277, 'learning_rate/qwen_vl_interface': 2.453094752899853e-06, 'learning_rate/action_model': 2.453094752899853e-05, 'epoch': 1.37}) +[2026-06-16 05:22:50,827][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.10541925579309464, 'train/grad_norm_pre_clip': 11.843664169311523, 'timing/data': 0.026083873584866524, 'timing/model': 1.5672916918992996, 'learning_rate/qwen_vl_interface': 2.4506338275915173e-06, 'learning_rate/action_model': 2.4506338275915172e-05, 'epoch': 1.37}) +[2026-06-16 05:22:56,960][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.047303542494773865, 'train/grad_norm_pre_clip': 12.546401023864746, 'timing/data': 0.03225068189203739, 'timing/model': 1.7575945500284433, 'learning_rate/qwen_vl_interface': 2.4481740529968735e-06, 'learning_rate/action_model': 2.4481740529968734e-05, 'epoch': 1.37}) +[2026-06-16 05:23:02,944][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.023512277752161026, 'train/grad_norm_pre_clip': 4.9674577713012695, 'timing/data': 0.031876757740974426, 'timing/model': 1.5459784343838692, 'learning_rate/qwen_vl_interface': 2.4457154301270427e-06, 'learning_rate/action_model': 2.4457154301270427e-05, 'epoch': 1.37}) +[2026-06-16 05:23:08,878][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.15247704088687897, 'train/grad_norm_pre_clip': 21.3404483795166, 'timing/data': 0.026412876322865486, 'timing/model': 1.4274117927998304, 'learning_rate/qwen_vl_interface': 2.443257959992671e-06, 'learning_rate/action_model': 2.443257959992671e-05, 'epoch': 1.37}) +[2026-06-16 05:23:14,676][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.059259530156850815, 'train/grad_norm_pre_clip': 7.718303680419922, 'timing/data': 0.02680007554590702, 'timing/model': 1.4335985966026783, 'learning_rate/qwen_vl_interface': 2.440801643603933e-06, 'learning_rate/action_model': 2.440801643603933e-05, 'epoch': 1.37}) +[2026-06-16 05:23:20,379][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.1251395344734192, 'train/grad_norm_pre_clip': 15.022721290588379, 'timing/data': 0.02624545618891716, 'timing/model': 1.4383345395326614, 'learning_rate/qwen_vl_interface': 2.4383464819705252e-06, 'learning_rate/action_model': 2.438346481970525e-05, 'epoch': 1.37}) +[2026-06-16 05:23:26,280][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.19757165014743805, 'train/grad_norm_pre_clip': 7.334802627563477, 'timing/data': 0.026073608547449112, 'timing/model': 1.488608120009303, 'learning_rate/qwen_vl_interface': 2.435892476101674e-06, 'learning_rate/action_model': 2.435892476101674e-05, 'epoch': 1.37}) +[2026-06-16 05:23:32,158][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.01341660600155592, 'train/grad_norm_pre_clip': 2.351363182067871, 'timing/data': 0.02674318104982376, 'timing/model': 1.6200241409242153, 'learning_rate/qwen_vl_interface': 2.433439627006128e-06, 'learning_rate/action_model': 2.4334396270061277e-05, 'epoch': 1.37}) +[2026-06-16 05:23:37,779][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.01019398495554924, 'train/grad_norm_pre_clip': 1.6889139413833618, 'timing/data': 0.02740451507270336, 'timing/model': 1.4322633314877748, 'learning_rate/qwen_vl_interface': 2.4309879356921602e-06, 'learning_rate/action_model': 2.43098793569216e-05, 'epoch': 1.37}) +[2026-06-16 05:23:43,679][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.09698709845542908, 'train/grad_norm_pre_clip': 7.496113300323486, 'timing/data': 0.026641571894288063, 'timing/model': 1.4315002486109734, 'learning_rate/qwen_vl_interface': 2.428537403167569e-06, 'learning_rate/action_model': 2.4285374031675687e-05, 'epoch': 1.37}) +[2026-06-16 05:23:49,679][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.09052912145853043, 'train/grad_norm_pre_clip': 6.991764545440674, 'timing/data': 0.027323145419359207, 'timing/model': 1.431925343349576, 'learning_rate/qwen_vl_interface': 2.4260880304396765e-06, 'learning_rate/action_model': 2.4260880304396765e-05, 'epoch': 1.37}) +[2026-06-16 05:23:55,580][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.053148508071899414, 'train/grad_norm_pre_clip': 6.0250020027160645, 'timing/data': 0.027121970430016518, 'timing/model': 1.4384006448090076, 'learning_rate/qwen_vl_interface': 2.4236398185153245e-06, 'learning_rate/action_model': 2.4236398185153245e-05, 'epoch': 1.37}) +[2026-06-16 05:24:01,477][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.12561601400375366, 'train/grad_norm_pre_clip': 25.555988311767578, 'timing/data': 0.026465611532330513, 'timing/model': 1.4810564629733562, 'learning_rate/qwen_vl_interface': 2.4211927684008823e-06, 'learning_rate/action_model': 2.4211927684008824e-05, 'epoch': 1.37}) +[2026-06-16 05:24:07,370][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.05974159389734268, 'train/grad_norm_pre_clip': 8.122025489807129, 'timing/data': 0.027925563976168633, 'timing/model': 1.6273850090801716, 'learning_rate/qwen_vl_interface': 2.4187468811022398e-06, 'learning_rate/action_model': 2.4187468811022396e-05, 'epoch': 1.37}) +[2026-06-16 05:24:12,980][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.07487461715936661, 'train/grad_norm_pre_clip': 5.885087490081787, 'timing/data': 0.02725416049361229, 'timing/model': 1.4378284625709057, 'learning_rate/qwen_vl_interface': 2.416302157624809e-06, 'learning_rate/action_model': 2.4163021576248088e-05, 'epoch': 1.37}) +[2026-06-16 05:24:18,879][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.018792718648910522, 'train/grad_norm_pre_clip': 3.170579433441162, 'timing/data': 0.026592137292027473, 'timing/model': 1.4356501884758472, 'learning_rate/qwen_vl_interface': 2.4138585989735227e-06, 'learning_rate/action_model': 2.4138585989735226e-05, 'epoch': 1.37}) +[2026-06-16 05:24:24,778][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.23544901609420776, 'train/grad_norm_pre_clip': 16.753183364868164, 'timing/data': 0.026190225034952164, 'timing/model': 1.4381894133985043, 'learning_rate/qwen_vl_interface': 2.4114162061528357e-06, 'learning_rate/action_model': 2.4114162061528354e-05, 'epoch': 1.37}) +[2026-06-16 05:24:30,678][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.0035905728582292795, 'train/grad_norm_pre_clip': 0.5295313000679016, 'timing/data': 0.02619049698114395, 'timing/model': 1.4508920777589083, 'learning_rate/qwen_vl_interface': 2.408974980166724e-06, 'learning_rate/action_model': 2.408974980166724e-05, 'epoch': 1.37}) +[2026-06-16 05:24:36,555][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.12751545011997223, 'train/grad_norm_pre_clip': 22.88788414001465, 'timing/data': 0.026135748252272606, 'timing/model': 1.4741760529577732, 'learning_rate/qwen_vl_interface': 2.406534922018683e-06, 'learning_rate/action_model': 2.4065349220186827e-05, 'epoch': 1.37}) +[2026-06-16 05:24:42,371][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.018252870067954063, 'train/grad_norm_pre_clip': 6.849729061126709, 'timing/data': 0.02668835036456585, 'timing/model': 1.629515128210187, 'learning_rate/qwen_vl_interface': 2.4040960327117287e-06, 'learning_rate/action_model': 2.4040960327117283e-05, 'epoch': 1.37}) +[2026-06-16 05:24:48,335][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.3022172749042511, 'train/grad_norm_pre_clip': 12.151419639587402, 'timing/data': 0.03189893625676632, 'timing/model': 1.5402694772928953, 'learning_rate/qwen_vl_interface': 2.4016583132483976e-06, 'learning_rate/action_model': 2.4016583132483976e-05, 'epoch': 1.38}) +[2026-06-16 05:24:54,538][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.16800978779792786, 'train/grad_norm_pre_clip': 8.342101097106934, 'timing/data': 0.03205127455294132, 'timing/model': 1.5332979038357735, 'learning_rate/qwen_vl_interface': 2.399221764630741e-06, 'learning_rate/action_model': 2.3992217646307406e-05, 'epoch': 1.38}) +[2026-06-16 05:25:00,640][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.04474696144461632, 'train/grad_norm_pre_clip': 6.047703742980957, 'timing/data': 0.03199809603393078, 'timing/model': 1.5343251805752516, 'learning_rate/qwen_vl_interface': 2.3967863878603334e-06, 'learning_rate/action_model': 2.3967863878603332e-05, 'epoch': 1.38}) +[2026-06-16 05:25:06,734][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.10923261940479279, 'train/grad_norm_pre_clip': 13.00610065460205, 'timing/data': 0.032118966802954674, 'timing/model': 1.5268241111189127, 'learning_rate/qwen_vl_interface': 2.394352183938266e-06, 'learning_rate/action_model': 2.3943521839382664e-05, 'epoch': 1.38}) +[2026-06-16 05:25:12,944][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.07938873767852783, 'train/grad_norm_pre_clip': 14.11292839050293, 'timing/data': 0.03336907923221588, 'timing/model': 1.4808597415685654, 'learning_rate/qwen_vl_interface': 2.39191915386515e-06, 'learning_rate/action_model': 2.39191915386515e-05, 'epoch': 1.38}) +[2026-06-16 05:25:19,230][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.00371304783038795, 'train/grad_norm_pre_clip': 0.5816822052001953, 'timing/data': 0.045782098546624184, 'timing/model': 1.8085936941206455, 'learning_rate/qwen_vl_interface': 2.3894872986411095e-06, 'learning_rate/action_model': 2.3894872986411094e-05, 'epoch': 1.38}) +[2026-06-16 05:25:25,137][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.03985122591257095, 'train/grad_norm_pre_clip': 5.747642517089844, 'timing/data': 0.03192451223731041, 'timing/model': 1.5323073156177998, 'learning_rate/qwen_vl_interface': 2.3870566192657903e-06, 'learning_rate/action_model': 2.38705661926579e-05, 'epoch': 1.38}) +[2026-06-16 05:25:31,337][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.26364949345588684, 'train/grad_norm_pre_clip': 10.8414888381958, 'timing/data': 0.03138367086648941, 'timing/model': 1.537650253623724, 'learning_rate/qwen_vl_interface': 2.3846271167383515e-06, 'learning_rate/action_model': 2.3846271167383515e-05, 'epoch': 1.38}) +[2026-06-16 05:25:37,439][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.1534590870141983, 'train/grad_norm_pre_clip': 16.113122940063477, 'timing/data': 0.047016747295856476, 'timing/model': 1.511965051293373, 'learning_rate/qwen_vl_interface': 2.382198792057471e-06, 'learning_rate/action_model': 2.3821987920574707e-05, 'epoch': 1.38}) +[2026-06-16 05:25:43,751][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.06647053360939026, 'train/grad_norm_pre_clip': 20.751150131225586, 'timing/data': 0.033567894250154495, 'timing/model': 1.5282339006662369, 'learning_rate/qwen_vl_interface': 2.3797716462213402e-06, 'learning_rate/action_model': 2.3797716462213402e-05, 'epoch': 1.38}) +[2026-06-16 05:25:49,832][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.011739620007574558, 'train/grad_norm_pre_clip': 1.4125819206237793, 'timing/data': 0.031594378873705864, 'timing/model': 1.4726014472544193, 'learning_rate/qwen_vl_interface': 2.37734568022767e-06, 'learning_rate/action_model': 2.3773456802276697e-05, 'epoch': 1.38}) +[2026-06-16 05:25:55,967][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.11273346841335297, 'train/grad_norm_pre_clip': 12.09225082397461, 'timing/data': 0.031072916463017464, 'timing/model': 1.7499777302145958, 'learning_rate/qwen_vl_interface': 2.3749208950736783e-06, 'learning_rate/action_model': 2.374920895073678e-05, 'epoch': 1.38}) +[2026-06-16 05:26:01,934][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.03125564008951187, 'train/grad_norm_pre_clip': 5.054133892059326, 'timing/data': 0.03055139258503914, 'timing/model': 1.5553542524576187, 'learning_rate/qwen_vl_interface': 2.372497291756106e-06, 'learning_rate/action_model': 2.3724972917561058e-05, 'epoch': 1.38}) +[2026-06-16 05:26:08,142][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.03907572105526924, 'train/grad_norm_pre_clip': 12.629289627075195, 'timing/data': 0.032128484919667244, 'timing/model': 1.5407818835228682, 'learning_rate/qwen_vl_interface': 2.3700748712712043e-06, 'learning_rate/action_model': 2.370074871271204e-05, 'epoch': 1.38}) +[2026-06-16 05:26:14,334][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.17084398865699768, 'train/grad_norm_pre_clip': 7.7343597412109375, 'timing/data': 0.03407934866845608, 'timing/model': 1.525069735944271, 'learning_rate/qwen_vl_interface': 2.3676536346147388e-06, 'learning_rate/action_model': 2.3676536346147388e-05, 'epoch': 1.38}) +[2026-06-16 05:26:20,434][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.2844998240470886, 'train/grad_norm_pre_clip': 24.266624450683594, 'timing/data': 0.032422399148344994, 'timing/model': 1.5032680742442608, 'learning_rate/qwen_vl_interface': 2.365233582781988e-06, 'learning_rate/action_model': 2.365233582781988e-05, 'epoch': 1.38}) +[2026-06-16 05:26:26,636][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.14850279688835144, 'train/grad_norm_pre_clip': 9.358409881591797, 'timing/data': 0.03168838284909725, 'timing/model': 1.4882302805781364, 'learning_rate/qwen_vl_interface': 2.3628147167677444e-06, 'learning_rate/action_model': 2.3628147167677443e-05, 'epoch': 1.38}) +[2026-06-16 05:26:32,873][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.07631982117891312, 'train/grad_norm_pre_clip': 6.986816883087158, 'timing/data': 0.03187950700521469, 'timing/model': 1.7727659977972507, 'learning_rate/qwen_vl_interface': 2.3603970375663123e-06, 'learning_rate/action_model': 2.3603970375663125e-05, 'epoch': 1.38}) +[2026-06-16 05:26:38,837][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.027233539149165154, 'train/grad_norm_pre_clip': 6.677207946777344, 'timing/data': 0.03250479698181152, 'timing/model': 1.5190456230193377, 'learning_rate/qwen_vl_interface': 2.35798054617151e-06, 'learning_rate/action_model': 2.3579805461715096e-05, 'epoch': 1.38}) +[2026-06-16 05:26:44,876][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.014196970500051975, 'train/grad_norm_pre_clip': 5.446424961090088, 'timing/data': 0.025265857577323914, 'timing/model': 1.473475620150566, 'learning_rate/qwen_vl_interface': 2.355565243576662e-06, 'learning_rate/action_model': 2.3555652435766617e-05, 'epoch': 1.38}) +[2026-06-16 05:26:50,673][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.024638261646032333, 'train/grad_norm_pre_clip': 8.234968185424805, 'timing/data': 0.0271226204931736, 'timing/model': 1.4351963363587856, 'learning_rate/qwen_vl_interface': 2.35315113077461e-06, 'learning_rate/action_model': 2.3531511307746098e-05, 'epoch': 1.38}) +[2026-06-16 05:26:56,485][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.10380382090806961, 'train/grad_norm_pre_clip': 12.726673126220703, 'timing/data': 0.026935981586575508, 'timing/model': 1.4455801583826542, 'learning_rate/qwen_vl_interface': 2.3507382087577047e-06, 'learning_rate/action_model': 2.350738208757705e-05, 'epoch': 1.38}) +[2026-06-16 05:27:02,379][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.018813664093613625, 'train/grad_norm_pre_clip': 6.17973518371582, 'timing/data': 0.02887614443898201, 'timing/model': 1.427690191194415, 'learning_rate/qwen_vl_interface': 2.348326478517807e-06, 'learning_rate/action_model': 2.3483264785178072e-05, 'epoch': 1.38}) +[2026-06-16 05:27:08,256][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.1388501226902008, 'train/grad_norm_pre_clip': 14.225415229797363, 'timing/data': 0.026932939887046814, 'timing/model': 1.6105654519051313, 'learning_rate/qwen_vl_interface': 2.3459159410462907e-06, 'learning_rate/action_model': 2.3459159410462905e-05, 'epoch': 1.38}) +[2026-06-16 05:27:13,973][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.08103685081005096, 'train/grad_norm_pre_clip': 5.087209701538086, 'timing/data': 0.02533050999045372, 'timing/model': 1.5331547241657972, 'learning_rate/qwen_vl_interface': 2.343506597334031e-06, 'learning_rate/action_model': 2.343506597334031e-05, 'epoch': 1.38}) +[2026-06-16 05:27:19,583][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.06814318895339966, 'train/grad_norm_pre_clip': 4.425628185272217, 'timing/data': 0.025905361399054527, 'timing/model': 1.4400806222110987, 'learning_rate/qwen_vl_interface': 2.341098448371423e-06, 'learning_rate/action_model': 2.3410984483714227e-05, 'epoch': 1.38}) +[2026-06-16 05:27:25,479][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.005029910709708929, 'train/grad_norm_pre_clip': 1.0257872343063354, 'timing/data': 0.026893991976976395, 'timing/model': 1.4421862848103046, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 1.39}) +[2026-06-16 05:27:31,378][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.1270793229341507, 'train/grad_norm_pre_clip': 8.147467613220215, 'timing/data': 0.02702338807284832, 'timing/model': 1.4382309950888157, 'learning_rate/qwen_vl_interface': 2.33628573865426e-06, 'learning_rate/action_model': 2.3362857386542596e-05, 'epoch': 1.39}) +[2026-06-16 05:27:37,278][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.13940423727035522, 'train/grad_norm_pre_clip': 15.671652793884277, 'timing/data': 0.025746358558535576, 'timing/model': 1.4574833400547504, 'learning_rate/qwen_vl_interface': 2.3338811798780282e-06, 'learning_rate/action_model': 2.333881179878028e-05, 'epoch': 1.39}) +[2026-06-16 05:27:43,167][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.024052508175373077, 'train/grad_norm_pre_clip': 5.776633262634277, 'timing/data': 0.027315188199281693, 'timing/model': 1.6255762297660112, 'learning_rate/qwen_vl_interface': 2.3314778198080933e-06, 'learning_rate/action_model': 2.331477819808093e-05, 'epoch': 1.39}) +[2026-06-16 05:27:48,976][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.01454089768230915, 'train/grad_norm_pre_clip': 1.279337763786316, 'timing/data': 0.027339279651641846, 'timing/model': 1.6331580877304077, 'learning_rate/qwen_vl_interface': 2.3290756594323834e-06, 'learning_rate/action_model': 2.3290756594323836e-05, 'epoch': 1.39}) +[2026-06-16 05:27:54,685][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.005514111369848251, 'train/grad_norm_pre_clip': 1.2988083362579346, 'timing/data': 0.025538092479109764, 'timing/model': 1.4413058403879404, 'learning_rate/qwen_vl_interface': 2.3266746997383378e-06, 'learning_rate/action_model': 2.3266746997383377e-05, 'epoch': 1.39}) +[2026-06-16 05:28:00,580][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.25200608372688293, 'train/grad_norm_pre_clip': 15.096620559692383, 'timing/data': 0.02562546730041504, 'timing/model': 1.4400607123970985, 'learning_rate/qwen_vl_interface': 2.324274941712899e-06, 'learning_rate/action_model': 2.324274941712899e-05, 'epoch': 1.39}) +[2026-06-16 05:28:06,536][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.11435486376285553, 'train/grad_norm_pre_clip': 9.885047912597656, 'timing/data': 0.03217826969921589, 'timing/model': 1.4890249762684107, 'learning_rate/qwen_vl_interface': 2.321876386342519e-06, 'learning_rate/action_model': 2.3218763863425187e-05, 'epoch': 1.39}) +[2026-06-16 05:28:12,641][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.007735455874353647, 'train/grad_norm_pre_clip': 1.9951473474502563, 'timing/data': 0.031618690118193626, 'timing/model': 1.4763502050191164, 'learning_rate/qwen_vl_interface': 2.31947903461315e-06, 'learning_rate/action_model': 2.3194790346131502e-05, 'epoch': 1.39}) +[2026-06-16 05:28:18,767][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.02948138862848282, 'train/grad_norm_pre_clip': 12.328386306762695, 'timing/data': 0.03175622783601284, 'timing/model': 1.765450021252036, 'learning_rate/qwen_vl_interface': 2.3170828875102544e-06, 'learning_rate/action_model': 2.3170828875102543e-05, 'epoch': 1.39}) +[2026-06-16 05:28:24,738][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.007389899808913469, 'train/grad_norm_pre_clip': 1.5571106672286987, 'timing/data': 0.032424554228782654, 'timing/model': 1.5076307151466608, 'learning_rate/qwen_vl_interface': 2.3146879460187984e-06, 'learning_rate/action_model': 2.3146879460187982e-05, 'epoch': 1.39}) +[2026-06-16 05:28:30,836][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.02070944383740425, 'train/grad_norm_pre_clip': 2.1958675384521484, 'timing/data': 0.031731344759464264, 'timing/model': 1.520998302847147, 'learning_rate/qwen_vl_interface': 2.3122942111232506e-06, 'learning_rate/action_model': 2.3122942111232506e-05, 'epoch': 1.39}) +[2026-06-16 05:28:36,945][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.11890648305416107, 'train/grad_norm_pre_clip': 7.538701057434082, 'timing/data': 0.032382749021053314, 'timing/model': 1.527349580079317, 'learning_rate/qwen_vl_interface': 2.3099016838075857e-06, 'learning_rate/action_model': 2.3099016838075856e-05, 'epoch': 1.39}) +[2026-06-16 05:28:43,135][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.3455018997192383, 'train/grad_norm_pre_clip': 14.805753707885742, 'timing/data': 0.03307870402932167, 'timing/model': 1.4993018843233585, 'learning_rate/qwen_vl_interface': 2.307510365055282e-06, 'learning_rate/action_model': 2.307510365055282e-05, 'epoch': 1.39}) +[2026-06-16 05:28:49,377][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.0016001074109226465, 'train/grad_norm_pre_clip': 0.2658858597278595, 'timing/data': 0.030967770144343376, 'timing/model': 1.5982582122087479, 'learning_rate/qwen_vl_interface': 2.3051202558493196e-06, 'learning_rate/action_model': 2.3051202558493192e-05, 'epoch': 1.39}) +[2026-06-16 05:28:55,940][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.008886621333658695, 'train/grad_norm_pre_clip': 1.5040161609649658, 'timing/data': 0.03980960324406624, 'timing/model': 1.875300507992506, 'learning_rate/qwen_vl_interface': 2.302731357172182e-06, 'learning_rate/action_model': 2.302731357172182e-05, 'epoch': 1.39}) +[2026-06-16 05:29:02,606][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.07867205142974854, 'train/grad_norm_pre_clip': 14.023046493530273, 'timing/data': 0.03642282634973526, 'timing/model': 1.6310448739677668, 'learning_rate/qwen_vl_interface': 2.3003436700058556e-06, 'learning_rate/action_model': 2.3003436700058554e-05, 'epoch': 1.39}) +[2026-06-16 05:29:09,727][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.015284196473658085, 'train/grad_norm_pre_clip': 2.2871222496032715, 'timing/data': 0.05126658268272877, 'timing/model': 1.7083014212548733, 'learning_rate/qwen_vl_interface': 2.297957195331828e-06, 'learning_rate/action_model': 2.2979571953318276e-05, 'epoch': 1.39}) +[2026-06-16 05:29:16,365][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.01750466600060463, 'train/grad_norm_pre_clip': 2.7764151096343994, 'timing/data': 0.03317066468298435, 'timing/model': 1.5206464100629091, 'learning_rate/qwen_vl_interface': 2.2955719341310903e-06, 'learning_rate/action_model': 2.2955719341310903e-05, 'epoch': 1.39}) +[2026-06-16 05:29:22,895][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.0766819640994072, 'train/grad_norm_pre_clip': 6.12615966796875, 'timing/data': 0.04484411142766476, 'timing/model': 1.6940420009195805, 'learning_rate/qwen_vl_interface': 2.2931878873841304e-06, 'learning_rate/action_model': 2.2931878873841302e-05, 'epoch': 1.39}) +[2026-06-16 05:29:29,315][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.10754819959402084, 'train/grad_norm_pre_clip': 6.748364448547363, 'timing/data': 0.03779046796262264, 'timing/model': 1.6153647806495428, 'learning_rate/qwen_vl_interface': 2.290805056070942e-06, 'learning_rate/action_model': 2.290805056070942e-05, 'epoch': 1.39}) +[2026-06-16 05:37:01,526][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.1235659271478653, 'train/grad_norm_pre_clip': 25.084476470947266, 'eval/loss': 0.040428709238767624, 'eval/flap_precision': 0.9492656875834445, 'eval/flap_recall': 0.9647218453188603, 'eval/flap_f1': 0.9569313593539703, 'eval/flap_rate_model': 0.20805555555555555, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9569313593539703, 'timing/data': 0.04189600981771946, 'timing/model': 1.9048816747963428, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 1.39}) +[2026-06-16 05:37:01,527][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3600_state +[2026-06-16 05:37:01,541][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 05:37:07,328][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3600_state/model.safetensors +[2026-06-16 05:37:21,417][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3600_state/optimizer.bin +[2026-06-16 05:37:21,418][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3600_state/sampler.bin +[2026-06-16 05:37:21,418][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3600_state/sampler_1.bin +[2026-06-16 05:37:21,419][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3600_state/random_states_0.pkl +[2026-06-16 05:37:21,518][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3600 +[2026-06-16 05:37:21,519][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 05:37:21,527][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 05:37:27,384][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.10316622257232666, 'train/grad_norm_pre_clip': 7.905230522155762, 'timing/data': 0.026442991569638252, 'timing/model': 1.6284299828112125, 'learning_rate/qwen_vl_interface': 2.2860430436633447e-06, 'learning_rate/action_model': 2.2860430436633447e-05, 'epoch': 1.39}) +[2026-06-16 05:37:33,216][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.08982394635677338, 'train/grad_norm_pre_clip': 4.690190315246582, 'timing/data': 0.03729900345206261, 'timing/model': 1.517263300716877, 'learning_rate/qwen_vl_interface': 2.28366386452642e-06, 'learning_rate/action_model': 2.28366386452642e-05, 'epoch': 1.39}) +[2026-06-16 05:37:39,657][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.10108921676874161, 'train/grad_norm_pre_clip': 15.291975975036621, 'timing/data': 0.03380821458995342, 'timing/model': 1.561900295317173, 'learning_rate/qwen_vl_interface': 2.2812859047382326e-06, 'learning_rate/action_model': 2.2812859047382323e-05, 'epoch': 1.39}) +[2026-06-16 05:37:45,862][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.017128977924585342, 'train/grad_norm_pre_clip': 4.468823432922363, 'timing/data': 0.031702032312750816, 'timing/model': 1.519978079944849, 'learning_rate/qwen_vl_interface': 2.2789091652762712e-06, 'learning_rate/action_model': 2.2789091652762712e-05, 'epoch': 1.39}) +[2026-06-16 05:37:51,964][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.012264229357242584, 'train/grad_norm_pre_clip': 1.9475048780441284, 'timing/data': 0.03328445181250572, 'timing/model': 1.5342847909778357, 'learning_rate/qwen_vl_interface': 2.276533647117524e-06, 'learning_rate/action_model': 2.2765336471175242e-05, 'epoch': 1.4}) +[2026-06-16 05:37:58,129][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.062169261276721954, 'train/grad_norm_pre_clip': 14.171182632446289, 'timing/data': 0.03204216808080673, 'timing/model': 1.517801409587264, 'learning_rate/qwen_vl_interface': 2.2741593512384773e-06, 'learning_rate/action_model': 2.2741593512384772e-05, 'epoch': 1.4}) +[2026-06-16 05:38:04,312][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.0864206850528717, 'train/grad_norm_pre_clip': 9.924896240234375, 'timing/data': 0.03203074261546135, 'timing/model': 1.6982591301202774, 'learning_rate/qwen_vl_interface': 2.2717862786151153e-06, 'learning_rate/action_model': 2.271786278615115e-05, 'epoch': 1.4}) +[2026-06-16 05:38:10,230][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.10055109858512878, 'train/grad_norm_pre_clip': 9.54305648803711, 'timing/data': 0.03156882897019386, 'timing/model': 1.5299680568277836, 'learning_rate/qwen_vl_interface': 2.2694144302229152e-06, 'learning_rate/action_model': 2.2694144302229152e-05, 'epoch': 1.4}) +[2026-06-16 05:38:16,338][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.09336669743061066, 'train/grad_norm_pre_clip': 10.139541625976562, 'timing/data': 0.03264516405761242, 'timing/model': 1.5091412626206875, 'learning_rate/qwen_vl_interface': 2.2670438070368568e-06, 'learning_rate/action_model': 2.2670438070368566e-05, 'epoch': 1.4}) +[2026-06-16 05:38:22,439][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.07441118359565735, 'train/grad_norm_pre_clip': 3.861067533493042, 'timing/data': 0.03094898909330368, 'timing/model': 1.5383409317582846, 'learning_rate/qwen_vl_interface': 2.264674410031413e-06, 'learning_rate/action_model': 2.2646744100314127e-05, 'epoch': 1.4}) +[2026-06-16 05:38:28,532][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.10413393378257751, 'train/grad_norm_pre_clip': 8.969725608825684, 'timing/data': 0.032483210787177086, 'timing/model': 1.4967219736427069, 'learning_rate/qwen_vl_interface': 2.2623062401805537e-06, 'learning_rate/action_model': 2.2623062401805535e-05, 'epoch': 1.4}) +[2026-06-16 05:38:34,634][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.18587073683738708, 'train/grad_norm_pre_clip': 13.494617462158203, 'timing/data': 0.03216574527323246, 'timing/model': 1.5419276524335146, 'learning_rate/qwen_vl_interface': 2.2599392984577436e-06, 'learning_rate/action_model': 2.2599392984577438e-05, 'epoch': 1.4}) +[2026-06-16 05:38:40,749][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.0739520788192749, 'train/grad_norm_pre_clip': 8.423821449279785, 'timing/data': 0.03166571259498596, 'timing/model': 1.7410540338605642, 'learning_rate/qwen_vl_interface': 2.2575735858359436e-06, 'learning_rate/action_model': 2.257573585835944e-05, 'epoch': 1.4}) +[2026-06-16 05:38:46,634][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.004830889869481325, 'train/grad_norm_pre_clip': 0.8206647038459778, 'timing/data': 0.03134996071457863, 'timing/model': 1.5301939323544502, 'learning_rate/qwen_vl_interface': 2.255209103287609e-06, 'learning_rate/action_model': 2.255209103287609e-05, 'epoch': 1.4}) +[2026-06-16 05:38:52,837][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.11317407339811325, 'train/grad_norm_pre_clip': 3.4009244441986084, 'timing/data': 0.03209078125655651, 'timing/model': 1.5350044276565313, 'learning_rate/qwen_vl_interface': 2.2528458517846886e-06, 'learning_rate/action_model': 2.252845851784689e-05, 'epoch': 1.4}) +[2026-06-16 05:38:58,936][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.06268757581710815, 'train/grad_norm_pre_clip': 3.2103664875030518, 'timing/data': 0.03107961267232895, 'timing/model': 1.517928533256054, 'learning_rate/qwen_vl_interface': 2.250483832298627e-06, 'learning_rate/action_model': 2.250483832298627e-05, 'epoch': 1.4}) +[2026-06-16 05:39:05,038][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.02051149494946003, 'train/grad_norm_pre_clip': 1.7453430891036987, 'timing/data': 0.032520247623324394, 'timing/model': 1.523967307060957, 'learning_rate/qwen_vl_interface': 2.2481230458003608e-06, 'learning_rate/action_model': 2.2481230458003606e-05, 'epoch': 1.4}) +[2026-06-16 05:39:11,236][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.22039470076560974, 'train/grad_norm_pre_clip': 15.924649238586426, 'timing/data': 0.030426908284425735, 'timing/model': 1.530326932668686, 'learning_rate/qwen_vl_interface': 2.2457634932603184e-06, 'learning_rate/action_model': 2.2457634932603183e-05, 'epoch': 1.4}) +[2026-06-16 05:39:17,453][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.0023543343413621187, 'train/grad_norm_pre_clip': 0.41618266701698303, 'timing/data': 0.03185330890119076, 'timing/model': 1.7377731576561928, 'learning_rate/qwen_vl_interface': 2.2434051756484255e-06, 'learning_rate/action_model': 2.2434051756484253e-05, 'epoch': 1.4}) +[2026-06-16 05:39:23,344][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.07687323540449142, 'train/grad_norm_pre_clip': 10.45620059967041, 'timing/data': 0.03183206357061863, 'timing/model': 1.4809787329286337, 'learning_rate/qwen_vl_interface': 2.2410480939340945e-06, 'learning_rate/action_model': 2.2410480939340945e-05, 'epoch': 1.4}) +[2026-06-16 05:39:29,339][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.05880816653370857, 'train/grad_norm_pre_clip': 16.685134887695312, 'timing/data': 0.029802069067955017, 'timing/model': 1.4856468848884106, 'learning_rate/qwen_vl_interface': 2.238692249086234e-06, 'learning_rate/action_model': 2.2386922490862336e-05, 'epoch': 1.4}) +[2026-06-16 05:39:35,439][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.009413225576281548, 'train/grad_norm_pre_clip': 0.9084727168083191, 'timing/data': 0.031725842505693436, 'timing/model': 1.4909641165286303, 'learning_rate/qwen_vl_interface': 2.2363376420732426e-06, 'learning_rate/action_model': 2.2363376420732423e-05, 'epoch': 1.4}) +[2026-06-16 05:39:41,704][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.009472953155636787, 'train/grad_norm_pre_clip': 1.5371811389923096, 'timing/data': 0.032508667558431625, 'timing/model': 1.5354356467723846, 'learning_rate/qwen_vl_interface': 2.233984273863011e-06, 'learning_rate/action_model': 2.233984273863011e-05, 'epoch': 1.4}) +[2026-06-16 05:39:47,940][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.0009240249637514353, 'train/grad_norm_pre_clip': 0.10580140352249146, 'timing/data': 0.03174752742052078, 'timing/model': 1.5430121440440416, 'learning_rate/qwen_vl_interface': 2.231632145422919e-06, 'learning_rate/action_model': 2.231632145422919e-05, 'epoch': 1.4}) +[2026-06-16 05:39:54,124][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.0122801149263978, 'train/grad_norm_pre_clip': 5.723255634307861, 'timing/data': 0.03129948116838932, 'timing/model': 1.756724614650011, 'learning_rate/qwen_vl_interface': 2.2292812577198383e-06, 'learning_rate/action_model': 2.229281257719838e-05, 'epoch': 1.4}) +[2026-06-16 05:40:00,139][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.013596754521131516, 'train/grad_norm_pre_clip': 1.5838907957077026, 'timing/data': 0.031235460191965103, 'timing/model': 1.5408591516315937, 'learning_rate/qwen_vl_interface': 2.2269316117201307e-06, 'learning_rate/action_model': 2.2269316117201307e-05, 'epoch': 1.4}) +[2026-06-16 05:40:06,347][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.004424617160111666, 'train/grad_norm_pre_clip': 0.5433185696601868, 'timing/data': 0.03264933079481125, 'timing/model': 1.5553377214819193, 'learning_rate/qwen_vl_interface': 2.224583208389646e-06, 'learning_rate/action_model': 2.224583208389646e-05, 'epoch': 1.4}) +[2026-06-16 05:40:12,540][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.22641561925411224, 'train/grad_norm_pre_clip': 33.89748764038086, 'timing/data': 0.03225459158420563, 'timing/model': 1.5189278032630682, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 1.4}) +[2026-06-16 05:40:18,638][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.008996933698654175, 'train/grad_norm_pre_clip': 1.0680917501449585, 'timing/data': 0.031652430072426796, 'timing/model': 1.5073820650577545, 'learning_rate/qwen_vl_interface': 2.219890133597194e-06, 'learning_rate/action_model': 2.219890133597194e-05, 'epoch': 1.4}) +[2026-06-16 05:40:24,580][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.03915538266301155, 'train/grad_norm_pre_clip': 4.825211048126221, 'timing/data': 0.024879692122340202, 'timing/model': 1.4314061719924212, 'learning_rate/qwen_vl_interface': 2.217545464064374e-06, 'learning_rate/action_model': 2.217545464064374e-05, 'epoch': 1.4}) +[2026-06-16 05:40:30,457][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.011400388553738594, 'train/grad_norm_pre_clip': 7.08645486831665, 'timing/data': 0.0253714919090271, 'timing/model': 1.6102588269859552, 'learning_rate/qwen_vl_interface': 2.215202041059067e-06, 'learning_rate/action_model': 2.215202041059067e-05, 'epoch': 1.41}) +[2026-06-16 05:40:36,092][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.1311725527048111, 'train/grad_norm_pre_clip': 26.23352813720703, 'timing/data': 0.027437260374426842, 'timing/model': 1.4418851975351572, 'learning_rate/qwen_vl_interface': 2.212859865544566e-06, 'learning_rate/action_model': 2.212859865544566e-05, 'epoch': 1.41}) +[2026-06-16 05:40:41,994][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.02805607207119465, 'train/grad_norm_pre_clip': 10.887086868286133, 'timing/data': 0.025284819304943085, 'timing/model': 1.4506933949887753, 'learning_rate/qwen_vl_interface': 2.210518938483652e-06, 'learning_rate/action_model': 2.210518938483652e-05, 'epoch': 1.41}) +[2026-06-16 05:40:48,371][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.014161466620862484, 'train/grad_norm_pre_clip': 8.397713661193848, 'timing/data': 0.032672781497240067, 'timing/model': 1.531343987211585, 'learning_rate/qwen_vl_interface': 2.20817926083859e-06, 'learning_rate/action_model': 2.20817926083859e-05, 'epoch': 1.41}) +[2026-06-16 05:40:54,756][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.08901804685592651, 'train/grad_norm_pre_clip': 10.457986831665039, 'timing/data': 0.03981229104101658, 'timing/model': 1.522861523553729, 'learning_rate/qwen_vl_interface': 2.205840833571135e-06, 'learning_rate/action_model': 2.205840833571135e-05, 'epoch': 1.41}) +[2026-06-16 05:41:01,048][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.11677994579076767, 'train/grad_norm_pre_clip': 7.966493129730225, 'timing/data': 0.03604126535356045, 'timing/model': 1.5720543172210455, 'learning_rate/qwen_vl_interface': 2.2035036576425264e-06, 'learning_rate/action_model': 2.203503657642526e-05, 'epoch': 1.41}) +[2026-06-16 05:41:07,247][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.33198654651641846, 'train/grad_norm_pre_clip': 22.98520278930664, 'timing/data': 0.03622567653656006, 'timing/model': 1.7375689465552568, 'learning_rate/qwen_vl_interface': 2.2011677340134867e-06, 'learning_rate/action_model': 2.2011677340134867e-05, 'epoch': 1.41}) +[2026-06-16 05:41:13,241][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.0032757248263806105, 'train/grad_norm_pre_clip': 0.5844184756278992, 'timing/data': 0.034493330866098404, 'timing/model': 1.4884021747857332, 'learning_rate/qwen_vl_interface': 2.198833063644227e-06, 'learning_rate/action_model': 2.198833063644227e-05, 'epoch': 1.41}) +[2026-06-16 05:41:19,443][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.05145224183797836, 'train/grad_norm_pre_clip': 10.556107521057129, 'timing/data': 0.03604687191545963, 'timing/model': 1.5244618318974972, 'learning_rate/qwen_vl_interface': 2.196499647494443e-06, 'learning_rate/action_model': 2.196499647494443e-05, 'epoch': 1.41}) +[2026-06-16 05:41:25,540][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.23565782606601715, 'train/grad_norm_pre_clip': 25.618135452270508, 'timing/data': 0.03494521789252758, 'timing/model': 1.5296281892806292, 'learning_rate/qwen_vl_interface': 2.1941674865233144e-06, 'learning_rate/action_model': 2.1941674865233143e-05, 'epoch': 1.41}) +[2026-06-16 05:41:31,641][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.010937659069895744, 'train/grad_norm_pre_clip': 3.448479175567627, 'timing/data': 0.03221964091062546, 'timing/model': 1.5300128366798162, 'learning_rate/qwen_vl_interface': 2.1918365816895013e-06, 'learning_rate/action_model': 2.191836581689501e-05, 'epoch': 1.41}) +[2026-06-16 05:41:37,939][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.014551771804690361, 'train/grad_norm_pre_clip': 1.7870770692825317, 'timing/data': 0.03535312041640282, 'timing/model': 1.5162188578397036, 'learning_rate/qwen_vl_interface': 2.189506933951153e-06, 'learning_rate/action_model': 2.1895069339511525e-05, 'epoch': 1.41}) +[2026-06-16 05:41:44,144][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.006357007194310427, 'train/grad_norm_pre_clip': 0.9150403141975403, 'timing/data': 0.03161297179758549, 'timing/model': 1.7309325579553843, 'learning_rate/qwen_vl_interface': 2.187178544265898e-06, 'learning_rate/action_model': 2.1871785442658983e-05, 'epoch': 1.41}) +[2026-06-16 05:41:50,037][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.06067571043968201, 'train/grad_norm_pre_clip': 7.421716690063477, 'timing/data': 0.033710647374391556, 'timing/model': 1.5256387777626514, 'learning_rate/qwen_vl_interface': 2.184851413590852e-06, 'learning_rate/action_model': 2.1848514135908518e-05, 'epoch': 1.41}) +[2026-06-16 05:41:56,237][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.017862241715192795, 'train/grad_norm_pre_clip': 7.044828414916992, 'timing/data': 0.030127378180623055, 'timing/model': 1.5421355422586203, 'learning_rate/qwen_vl_interface': 2.1825255428826087e-06, 'learning_rate/action_model': 2.182525542882609e-05, 'epoch': 1.41}) +[2026-06-16 05:42:02,337][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.03780433163046837, 'train/grad_norm_pre_clip': 4.495031833648682, 'timing/data': 0.03125188499689102, 'timing/model': 1.5369700286537409, 'learning_rate/qwen_vl_interface': 2.180200933097246e-06, 'learning_rate/action_model': 2.180200933097246e-05, 'epoch': 1.41}) +[2026-06-16 05:42:08,637][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.030657662078738213, 'train/grad_norm_pre_clip': 6.850855350494385, 'timing/data': 0.03311034478247166, 'timing/model': 1.6287744119763374, 'learning_rate/qwen_vl_interface': 2.1778775851903235e-06, 'learning_rate/action_model': 2.177877585190323e-05, 'epoch': 1.41}) +[2026-06-16 05:42:14,736][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.11512670665979385, 'train/grad_norm_pre_clip': 12.182114601135254, 'timing/data': 0.03182613663375378, 'timing/model': 1.5019597206264734, 'learning_rate/qwen_vl_interface': 2.175555500116881e-06, 'learning_rate/action_model': 2.175555500116881e-05, 'epoch': 1.41}) +[2026-06-16 05:42:20,946][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.04473910480737686, 'train/grad_norm_pre_clip': 9.40683650970459, 'timing/data': 0.031633732840418816, 'timing/model': 1.7413203194737434, 'learning_rate/qwen_vl_interface': 2.1732346788314404e-06, 'learning_rate/action_model': 2.1732346788314402e-05, 'epoch': 1.41}) +[2026-06-16 05:42:26,963][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.01773189567029476, 'train/grad_norm_pre_clip': 5.787208557128906, 'timing/data': 0.031898193061351776, 'timing/model': 1.6623365841805935, 'learning_rate/qwen_vl_interface': 2.170915122288006e-06, 'learning_rate/action_model': 2.170915122288006e-05, 'epoch': 1.41}) +[2026-06-16 05:42:32,936][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.03786846250295639, 'train/grad_norm_pre_clip': 5.3776164054870605, 'timing/data': 0.0315858107060194, 'timing/model': 1.5353864822536707, 'learning_rate/qwen_vl_interface': 2.168596831440055e-06, 'learning_rate/action_model': 2.1685968314400552e-05, 'epoch': 1.41}) +[2026-06-16 05:42:39,035][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.00027719969511963427, 'train/grad_norm_pre_clip': 0.04532178118824959, 'timing/data': 0.03256812319159508, 'timing/model': 1.5268192775547504, 'learning_rate/qwen_vl_interface': 2.166279807240553e-06, 'learning_rate/action_model': 2.1662798072405528e-05, 'epoch': 1.41}) +[2026-06-16 05:42:45,133][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.24417516589164734, 'train/grad_norm_pre_clip': 44.28108596801758, 'timing/data': 0.030906902626156807, 'timing/model': 1.5233762152493, 'learning_rate/qwen_vl_interface': 2.1639640506419394e-06, 'learning_rate/action_model': 2.163964050641939e-05, 'epoch': 1.41}) +[2026-06-16 05:42:51,230][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.02262398786842823, 'train/grad_norm_pre_clip': 4.295614242553711, 'timing/data': 0.03127117641270161, 'timing/model': 1.477127904072404, 'learning_rate/qwen_vl_interface': 2.161649562596135e-06, 'learning_rate/action_model': 2.161649562596135e-05, 'epoch': 1.41}) +[2026-06-16 05:42:57,347][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.12534794211387634, 'train/grad_norm_pre_clip': 10.666203498840332, 'timing/data': 0.035887714475393295, 'timing/model': 1.717084176838398, 'learning_rate/qwen_vl_interface': 2.1593363440545383e-06, 'learning_rate/action_model': 2.1593363440545384e-05, 'epoch': 1.41}) +[2026-06-16 05:43:03,625][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.08969032019376755, 'train/grad_norm_pre_clip': 10.94721794128418, 'timing/data': 0.03131052665412426, 'timing/model': 1.7101426757872105, 'learning_rate/qwen_vl_interface': 2.157024395968025e-06, 'learning_rate/action_model': 2.1570243959680252e-05, 'epoch': 1.41}) +[2026-06-16 05:43:09,738][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.04311518371105194, 'train/grad_norm_pre_clip': 6.622003078460693, 'timing/data': 0.03046991117298603, 'timing/model': 1.6305737793445587, 'learning_rate/qwen_vl_interface': 2.154713719286952e-06, 'learning_rate/action_model': 2.154713719286952e-05, 'epoch': 1.42}) +[2026-06-16 05:43:16,041][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.04554666951298714, 'train/grad_norm_pre_clip': 26.937089920043945, 'timing/data': 0.03281703032553196, 'timing/model': 1.5164458025246859, 'learning_rate/qwen_vl_interface': 2.152404314961149e-06, 'learning_rate/action_model': 2.152404314961149e-05, 'epoch': 1.42}) +[2026-06-16 05:43:22,135][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.3138730525970459, 'train/grad_norm_pre_clip': 20.589033126831055, 'timing/data': 0.03189827688038349, 'timing/model': 1.531379645690322, 'learning_rate/qwen_vl_interface': 2.1500961839399264e-06, 'learning_rate/action_model': 2.1500961839399262e-05, 'epoch': 1.42}) +[2026-06-16 05:43:28,234][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.0012252797605469823, 'train/grad_norm_pre_clip': 0.20491772890090942, 'timing/data': 0.03180324845016003, 'timing/model': 1.4917941223829985, 'learning_rate/qwen_vl_interface': 2.1477893271720683e-06, 'learning_rate/action_model': 2.1477893271720682e-05, 'epoch': 1.42}) +[2026-06-16 05:43:34,354][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.01002067606896162, 'train/grad_norm_pre_clip': 3.841899871826172, 'timing/data': 0.03185673989355564, 'timing/model': 1.7477492969483137, 'learning_rate/qwen_vl_interface': 2.1454837456058385e-06, 'learning_rate/action_model': 2.1454837456058386e-05, 'epoch': 1.42}) +[2026-06-16 05:43:40,233][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.0003226857224944979, 'train/grad_norm_pre_clip': 0.0666378065943718, 'timing/data': 0.03162689320743084, 'timing/model': 1.5321768578141928, 'learning_rate/qwen_vl_interface': 2.1431794401889722e-06, 'learning_rate/action_model': 2.143179440188972e-05, 'epoch': 1.42}) +[2026-06-16 05:43:46,335][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.204034224152565, 'train/grad_norm_pre_clip': 27.934463500976562, 'timing/data': 0.031198007985949516, 'timing/model': 1.518742635846138, 'learning_rate/qwen_vl_interface': 2.1408764118686837e-06, 'learning_rate/action_model': 2.1408764118686838e-05, 'epoch': 1.42}) +[2026-06-16 05:43:52,432][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.13728481531143188, 'train/grad_norm_pre_clip': 23.117918014526367, 'timing/data': 0.03324352204799652, 'timing/model': 1.5194759741425514, 'learning_rate/qwen_vl_interface': 2.1385746615916598e-06, 'learning_rate/action_model': 2.1385746615916596e-05, 'epoch': 1.42}) +[2026-06-16 05:43:58,532][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.0029026055708527565, 'train/grad_norm_pre_clip': 0.6991587281227112, 'timing/data': 0.03323686495423317, 'timing/model': 1.49040481261909, 'learning_rate/qwen_vl_interface': 2.1362741903040636e-06, 'learning_rate/action_model': 2.1362741903040633e-05, 'epoch': 1.42}) +[2026-06-16 05:44:04,542][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.20810391008853912, 'train/grad_norm_pre_clip': 12.818207740783691, 'timing/data': 0.0305436160415411, 'timing/model': 1.4762274250388145, 'learning_rate/qwen_vl_interface': 2.1339749989515318e-06, 'learning_rate/action_model': 2.1339749989515316e-05, 'epoch': 1.42}) +[2026-06-16 05:44:10,593][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.09531434625387192, 'train/grad_norm_pre_clip': 8.346220016479492, 'timing/data': 0.030891472473740578, 'timing/model': 1.6615301072597504, 'learning_rate/qwen_vl_interface': 2.131677088479175e-06, 'learning_rate/action_model': 2.1316770884791752e-05, 'epoch': 1.42}) +[2026-06-16 05:44:16,539][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.04584015905857086, 'train/grad_norm_pre_clip': 10.626178741455078, 'timing/data': 0.031075315549969673, 'timing/model': 1.5324390195310116, 'learning_rate/qwen_vl_interface': 2.1293804598315787e-06, 'learning_rate/action_model': 2.1293804598315787e-05, 'epoch': 1.42}) +[2026-06-16 05:44:22,712][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.025710932910442352, 'train/grad_norm_pre_clip': 8.728585243225098, 'timing/data': 0.024926455691456795, 'timing/model': 1.527162792161107, 'learning_rate/qwen_vl_interface': 2.1270851139527986e-06, 'learning_rate/action_model': 2.1270851139527983e-05, 'epoch': 1.42}) +[2026-06-16 05:44:28,622][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.1314050555229187, 'train/grad_norm_pre_clip': 12.795087814331055, 'timing/data': 0.025393785908818245, 'timing/model': 1.4459353871643543, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 1.42}) +[2026-06-16 05:44:34,817][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.18881455063819885, 'train/grad_norm_pre_clip': 7.793420791625977, 'timing/data': 0.027753343805670738, 'timing/model': 1.5352095980197191, 'learning_rate/qwen_vl_interface': 2.1224982742752814e-06, 'learning_rate/action_model': 2.1224982742752812e-05, 'epoch': 1.42}) +[2026-06-16 05:44:40,822][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.001959406305104494, 'train/grad_norm_pre_clip': 0.3917292058467865, 'timing/data': 0.02558920532464981, 'timing/model': 1.507679397240281, 'learning_rate/qwen_vl_interface': 2.1202067823620213e-06, 'learning_rate/action_model': 2.1202067823620212e-05, 'epoch': 1.42}) +[2026-06-16 05:44:46,666][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.060675062239170074, 'train/grad_norm_pre_clip': 10.81235122680664, 'timing/data': 0.0248734001070261, 'timing/model': 1.5872910842299461, 'learning_rate/qwen_vl_interface': 2.117916576988531e-06, 'learning_rate/action_model': 2.117916576988531e-05, 'epoch': 1.42}) +[2026-06-16 05:44:52,507][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.0047339703887701035, 'train/grad_norm_pre_clip': 1.973665714263916, 'timing/data': 0.027822915464639664, 'timing/model': 1.5288611631840467, 'learning_rate/qwen_vl_interface': 2.1156276590962257e-06, 'learning_rate/action_model': 2.1156276590962257e-05, 'epoch': 1.42}) +[2026-06-16 05:44:58,620][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.0929969772696495, 'train/grad_norm_pre_clip': 7.120568752288818, 'timing/data': 0.02566029317677021, 'timing/model': 1.4461457133293152, 'learning_rate/qwen_vl_interface': 2.113340029625994e-06, 'learning_rate/action_model': 2.1133400296259938e-05, 'epoch': 1.42}) +[2026-06-16 05:45:04,491][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.11924974620342255, 'train/grad_norm_pre_clip': 17.154916763305664, 'timing/data': 0.025707215070724487, 'timing/model': 1.4153672773391008, 'learning_rate/qwen_vl_interface': 2.111053689518194e-06, 'learning_rate/action_model': 2.111053689518194e-05, 'epoch': 1.42}) +[2026-06-16 05:45:10,381][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.043470948934555054, 'train/grad_norm_pre_clip': 12.14187240600586, 'timing/data': 0.027072882279753685, 'timing/model': 1.4469464402645826, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 1.42}) +[2026-06-16 05:45:16,253][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.043275799602270126, 'train/grad_norm_pre_clip': 7.716400623321533, 'timing/data': 0.026392243802547455, 'timing/model': 1.598788220435381, 'learning_rate/qwen_vl_interface': 2.106484881148672e-06, 'learning_rate/action_model': 2.1064848811486718e-05, 'epoch': 1.42}) +[2026-06-16 05:45:21,880][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.024880969896912575, 'train/grad_norm_pre_clip': 8.410667419433594, 'timing/data': 0.025507615879178047, 'timing/model': 1.428165988996625, 'learning_rate/qwen_vl_interface': 2.1042024147650137e-06, 'learning_rate/action_model': 2.1042024147650137e-05, 'epoch': 1.42}) +[2026-06-16 05:45:27,681][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.052353814244270325, 'train/grad_norm_pre_clip': 15.034987449645996, 'timing/data': 0.025408605113625526, 'timing/model': 1.417631732299924, 'learning_rate/qwen_vl_interface': 2.1019212414999163e-06, 'learning_rate/action_model': 2.101921241499916e-05, 'epoch': 1.42}) +[2026-06-16 05:45:33,395][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.0004504077660385519, 'train/grad_norm_pre_clip': 0.10998120158910751, 'timing/data': 0.026205504313111305, 'timing/model': 1.4523008596152067, 'learning_rate/qwen_vl_interface': 2.099641362291082e-06, 'learning_rate/action_model': 2.099641362291082e-05, 'epoch': 1.42}) +[2026-06-16 05:45:39,309][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.08980461955070496, 'train/grad_norm_pre_clip': 13.689859390258789, 'timing/data': 0.02685864455997944, 'timing/model': 1.4549769330769777, 'learning_rate/qwen_vl_interface': 2.097362778075685e-06, 'learning_rate/action_model': 2.097362778075685e-05, 'epoch': 1.43}) +[2026-06-16 05:45:45,274][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.01679876260459423, 'train/grad_norm_pre_clip': 1.0075891017913818, 'timing/data': 0.026333732530474663, 'timing/model': 1.4936650395393372, 'learning_rate/qwen_vl_interface': 2.0950854897903663e-06, 'learning_rate/action_model': 2.095085489790366e-05, 'epoch': 1.43}) +[2026-06-16 05:45:51,361][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.0164946299046278, 'train/grad_norm_pre_clip': 3.649249315261841, 'timing/data': 0.027086541056632996, 'timing/model': 1.6149389427155256, 'learning_rate/qwen_vl_interface': 2.092809498371229e-06, 'learning_rate/action_model': 2.0928094983712292e-05, 'epoch': 1.43}) +[2026-06-16 05:45:56,981][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.004708784632384777, 'train/grad_norm_pre_clip': 3.5773885250091553, 'timing/data': 0.02578330785036087, 'timing/model': 1.4356064405292273, 'learning_rate/qwen_vl_interface': 2.0905348047538498e-06, 'learning_rate/action_model': 2.0905348047538497e-05, 'epoch': 1.43}) +[2026-06-16 05:46:02,880][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.27233392000198364, 'train/grad_norm_pre_clip': 16.16301727294922, 'timing/data': 0.02631249837577343, 'timing/model': 1.4299544151872396, 'learning_rate/qwen_vl_interface': 2.0882614098732686e-06, 'learning_rate/action_model': 2.0882614098732685e-05, 'epoch': 1.43}) +[2026-06-16 05:46:08,782][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.074358269572258, 'train/grad_norm_pre_clip': 12.611884117126465, 'timing/data': 0.02616002783179283, 'timing/model': 1.4309644717723131, 'learning_rate/qwen_vl_interface': 2.085989314663993e-06, 'learning_rate/action_model': 2.0859893146639927e-05, 'epoch': 1.43}) +[2026-06-16 05:46:14,681][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.08291873335838318, 'train/grad_norm_pre_clip': 11.037833213806152, 'timing/data': 0.025689570233225822, 'timing/model': 1.4326439406722784, 'learning_rate/qwen_vl_interface': 2.083718520059995e-06, 'learning_rate/action_model': 2.0837185200599946e-05, 'epoch': 1.43}) +[2026-06-16 05:46:20,582][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.007825431413948536, 'train/grad_norm_pre_clip': 2.0987374782562256, 'timing/data': 0.026707537472248077, 'timing/model': 1.48707802221179, 'learning_rate/qwen_vl_interface': 2.0814490269947117e-06, 'learning_rate/action_model': 2.0814490269947118e-05, 'epoch': 1.43}) +[2026-06-16 05:46:26,457][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.026717519387602806, 'train/grad_norm_pre_clip': 6.557201385498047, 'timing/data': 0.026272524148225784, 'timing/model': 1.6095059644430876, 'learning_rate/qwen_vl_interface': 2.079180836401047e-06, 'learning_rate/action_model': 2.079180836401047e-05, 'epoch': 1.43}) +[2026-06-16 05:46:32,091][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.12024002522230148, 'train/grad_norm_pre_clip': 10.45042896270752, 'timing/data': 0.027990035712718964, 'timing/model': 1.4408413711935282, 'learning_rate/qwen_vl_interface': 2.0769139492113687e-06, 'learning_rate/action_model': 2.076913949211369e-05, 'epoch': 1.43}) +[2026-06-16 05:46:38,579][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.15221934020519257, 'train/grad_norm_pre_clip': 8.332472801208496, 'timing/data': 0.026177216321229935, 'timing/model': 1.433716306462884, 'learning_rate/qwen_vl_interface': 2.0746483663575075e-06, 'learning_rate/action_model': 2.0746483663575076e-05, 'epoch': 1.43}) +[2026-06-16 05:46:44,523][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.013255171477794647, 'train/grad_norm_pre_clip': 2.8290696144104004, 'timing/data': 0.026877274736762047, 'timing/model': 1.4460810627788305, 'learning_rate/qwen_vl_interface': 2.0723840887707595e-06, 'learning_rate/action_model': 2.0723840887707594e-05, 'epoch': 1.43}) +[2026-06-16 05:46:50,507][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.015942366793751717, 'train/grad_norm_pre_clip': 2.6335649490356445, 'timing/data': 0.02565379999577999, 'timing/model': 1.5292532164603472, 'learning_rate/qwen_vl_interface': 2.0701211173818854e-06, 'learning_rate/action_model': 2.0701211173818853e-05, 'epoch': 1.43}) +[2026-06-16 05:46:56,608][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.008890638127923012, 'train/grad_norm_pre_clip': 4.014597415924072, 'timing/data': 0.025940846651792526, 'timing/model': 1.4957886431366205, 'learning_rate/qwen_vl_interface': 2.0678594531211036e-06, 'learning_rate/action_model': 2.0678594531211033e-05, 'epoch': 1.43}) +[2026-06-16 05:47:02,495][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.3352997899055481, 'train/grad_norm_pre_clip': 4.753530502319336, 'timing/data': 0.027822095900774002, 'timing/model': 1.6408195663243532, 'learning_rate/qwen_vl_interface': 2.0655990969181e-06, 'learning_rate/action_model': 2.0655990969181003e-05, 'epoch': 1.43}) +[2026-06-16 05:47:08,296][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.04296628758311272, 'train/grad_norm_pre_clip': 20.130849838256836, 'timing/data': 0.025272123515605927, 'timing/model': 1.5168097727000713, 'learning_rate/qwen_vl_interface': 2.0633400497020236e-06, 'learning_rate/action_model': 2.0633400497020232e-05, 'epoch': 1.43}) +[2026-06-16 05:47:14,436][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.05766090005636215, 'train/grad_norm_pre_clip': 7.203924655914307, 'timing/data': 0.031967801973223686, 'timing/model': 1.5546411219984293, 'learning_rate/qwen_vl_interface': 2.061082312401481e-06, 'learning_rate/action_model': 2.061082312401481e-05, 'epoch': 1.43}) +[2026-06-16 05:47:20,425][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.16317269206047058, 'train/grad_norm_pre_clip': 19.19805908203125, 'timing/data': 0.02828861027956009, 'timing/model': 1.538400100544095, 'learning_rate/qwen_vl_interface': 2.058825885944544e-06, 'learning_rate/action_model': 2.0588258859445442e-05, 'epoch': 1.43}) +[2026-06-16 05:54:27,236][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.10993481427431107, 'train/grad_norm_pre_clip': 13.173314094543457, 'eval/loss': 0.0438065230846405, 'eval/flap_precision': 0.9591280653950953, 'eval/flap_recall': 0.9552238805970149, 'eval/flap_f1': 0.9571719918422841, 'eval/flap_rate_model': 0.2038888888888889, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.957171991842284, 'timing/data': 0.02784259244799614, 'timing/model': 1.4463939182460308, 'learning_rate/qwen_vl_interface': 2.0565707712587447e-06, 'learning_rate/action_model': 2.0565707712587444e-05, 'epoch': 1.43}) +[2026-06-16 05:54:27,236][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3700_state +[2026-06-16 05:54:27,240][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 05:54:33,174][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3700_state/model.safetensors +[2026-06-16 05:54:42,769][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3700_state/optimizer.bin +[2026-06-16 05:54:42,770][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3700_state/sampler.bin +[2026-06-16 05:54:42,770][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3700_state/sampler_1.bin +[2026-06-16 05:54:42,771][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3700_state/random_states_0.pkl +[2026-06-16 05:54:42,876][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3700 +[2026-06-16 05:54:42,876][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 05:54:42,885][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 05:54:48,919][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.054282888770103455, 'train/grad_norm_pre_clip': 14.282947540283203, 'timing/data': 0.032905323430895805, 'timing/model': 1.489641746506095, 'learning_rate/qwen_vl_interface': 2.0543169692710747e-06, 'learning_rate/action_model': 2.054316969271075e-05, 'epoch': 1.43}) +[2026-06-16 05:54:54,889][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.0009433472296223044, 'train/grad_norm_pre_clip': 0.18303436040878296, 'timing/data': 0.03017566166818142, 'timing/model': 1.4908081833273172, 'learning_rate/qwen_vl_interface': 2.0520644809079873e-06, 'learning_rate/action_model': 2.0520644809079873e-05, 'epoch': 1.43}) +[2026-06-16 05:55:00,732][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.06847106665372849, 'train/grad_norm_pre_clip': 5.227058410644531, 'timing/data': 0.031146839261054993, 'timing/model': 1.470563007518649, 'learning_rate/qwen_vl_interface': 2.049813307095395e-06, 'learning_rate/action_model': 2.0498133070953947e-05, 'epoch': 1.43}) +[2026-06-16 05:55:06,724][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.07888923585414886, 'train/grad_norm_pre_clip': 11.498769760131836, 'timing/data': 0.031391095370054245, 'timing/model': 1.641948290169239, 'learning_rate/qwen_vl_interface': 2.047563448758672e-06, 'learning_rate/action_model': 2.0475634487586717e-05, 'epoch': 1.43}) +[2026-06-16 05:55:12,617][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.16341865062713623, 'train/grad_norm_pre_clip': 18.014917373657227, 'timing/data': 0.03424966521561146, 'timing/model': 1.5125296209007502, 'learning_rate/qwen_vl_interface': 2.045314906822647e-06, 'learning_rate/action_model': 2.0453149068226467e-05, 'epoch': 1.43}) +[2026-06-16 05:55:18,513][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.02927701361477375, 'train/grad_norm_pre_clip': 11.306112289428711, 'timing/data': 0.031061658635735512, 'timing/model': 1.446767084300518, 'learning_rate/qwen_vl_interface': 2.0430676822116124e-06, 'learning_rate/action_model': 2.0430676822116123e-05, 'epoch': 1.43}) +[2026-06-16 05:55:24,408][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.07476387917995453, 'train/grad_norm_pre_clip': 9.685138702392578, 'timing/data': 0.04576826095581055, 'timing/model': 1.5166172161698341, 'learning_rate/qwen_vl_interface': 2.040821775849319e-06, 'learning_rate/action_model': 2.0408217758493188e-05, 'epoch': 1.43}) +[2026-06-16 05:55:30,413][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.02618754468858242, 'train/grad_norm_pre_clip': 4.429313659667969, 'timing/data': 0.03291526436805725, 'timing/model': 1.445319626480341, 'learning_rate/qwen_vl_interface': 2.03857718865897e-06, 'learning_rate/action_model': 2.0385771886589702e-05, 'epoch': 1.44}) +[2026-06-16 05:55:36,266][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.0032719240989536047, 'train/grad_norm_pre_clip': 0.735235869884491, 'timing/data': 0.0317262876778841, 'timing/model': 1.6707364972680807, 'learning_rate/qwen_vl_interface': 2.0363339215632337e-06, 'learning_rate/action_model': 2.0363339215632334e-05, 'epoch': 1.44}) +[2026-06-16 05:55:41,889][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.009454240091145039, 'train/grad_norm_pre_clip': 3.4492363929748535, 'timing/data': 0.03122730925679207, 'timing/model': 1.4459253568202257, 'learning_rate/qwen_vl_interface': 2.034091975484231e-06, 'learning_rate/action_model': 2.0340919754842307e-05, 'epoch': 1.44}) +[2026-06-16 05:55:47,747][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.17237067222595215, 'train/grad_norm_pre_clip': 20.97679328918457, 'timing/data': 0.031271906569600105, 'timing/model': 1.444526944309473, 'learning_rate/qwen_vl_interface': 2.0318513513435415e-06, 'learning_rate/action_model': 2.0318513513435415e-05, 'epoch': 1.44}) +[2026-06-16 05:55:53,556][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.0064097740687429905, 'train/grad_norm_pre_clip': 0.671567440032959, 'timing/data': 0.031637733802199364, 'timing/model': 1.4463498461991549, 'learning_rate/qwen_vl_interface': 2.0296120500622018e-06, 'learning_rate/action_model': 2.0296120500622017e-05, 'epoch': 1.44}) +[2026-06-16 05:55:59,289][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.03450985625386238, 'train/grad_norm_pre_clip': 8.467497825622559, 'timing/data': 0.024378133937716484, 'timing/model': 1.388638213276863, 'learning_rate/qwen_vl_interface': 2.0273740725607034e-06, 'learning_rate/action_model': 2.0273740725607032e-05, 'epoch': 1.44}) +[2026-06-16 05:56:04,965][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.011147037148475647, 'train/grad_norm_pre_clip': 1.742593765258789, 'timing/data': 0.030378814786672592, 'timing/model': 1.4779688101261854, 'learning_rate/qwen_vl_interface': 2.0251374197589942e-06, 'learning_rate/action_model': 2.0251374197589943e-05, 'epoch': 1.44}) +[2026-06-16 05:56:10,536][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.003432430559769273, 'train/grad_norm_pre_clip': 0.7026329636573792, 'timing/data': 0.024983657523989677, 'timing/model': 1.5602229926735163, 'learning_rate/qwen_vl_interface': 2.0229020925764785e-06, 'learning_rate/action_model': 2.0229020925764785e-05, 'epoch': 1.44}) +[2026-06-16 05:56:15,945][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.06839506328105927, 'train/grad_norm_pre_clip': 9.982422828674316, 'timing/data': 0.02677946537733078, 'timing/model': 1.3929709400981665, 'learning_rate/qwen_vl_interface': 2.0206680919320145e-06, 'learning_rate/action_model': 2.020668091932014e-05, 'epoch': 1.44}) +[2026-06-16 05:56:21,554][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.04413284733891487, 'train/grad_norm_pre_clip': 10.865557670593262, 'timing/data': 0.0246493648737669, 'timing/model': 1.3906202223151922, 'learning_rate/qwen_vl_interface': 2.0184354187439164e-06, 'learning_rate/action_model': 2.0184354187439162e-05, 'epoch': 1.44}) +[2026-06-16 05:56:27,101][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.03923410922288895, 'train/grad_norm_pre_clip': 5.655479431152344, 'timing/data': 0.02496904693543911, 'timing/model': 1.3920896854251623, 'learning_rate/qwen_vl_interface': 2.016204073929949e-06, 'learning_rate/action_model': 2.016204073929949e-05, 'epoch': 1.44}) +[2026-06-16 05:56:32,683][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.00026755392900668085, 'train/grad_norm_pre_clip': 0.0626940131187439, 'timing/data': 0.027660904452204704, 'timing/model': 1.3936306089162827, 'learning_rate/qwen_vl_interface': 2.0139740584073363e-06, 'learning_rate/action_model': 2.0139740584073364e-05, 'epoch': 1.44}) +[2026-06-16 05:56:38,272][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.11175954341888428, 'train/grad_norm_pre_clip': 10.391274452209473, 'timing/data': 0.02492811530828476, 'timing/model': 1.3928307630121708, 'learning_rate/qwen_vl_interface': 2.011745373092753e-06, 'learning_rate/action_model': 2.011745373092753e-05, 'epoch': 1.44}) +[2026-06-16 05:56:43,850][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.0024903323501348495, 'train/grad_norm_pre_clip': 0.7026467323303223, 'timing/data': 0.02536655217409134, 'timing/model': 1.561507822945714, 'learning_rate/qwen_vl_interface': 2.0095180189023267e-06, 'learning_rate/action_model': 2.0095180189023265e-05, 'epoch': 1.44}) +[2026-06-16 05:56:49,350][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.1370888203382492, 'train/grad_norm_pre_clip': 9.246017456054688, 'timing/data': 0.025939730927348137, 'timing/model': 1.4728396199643612, 'learning_rate/qwen_vl_interface': 2.0072919967516393e-06, 'learning_rate/action_model': 2.0072919967516392e-05, 'epoch': 1.44}) +[2026-06-16 05:56:54,770][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.1821732521057129, 'train/grad_norm_pre_clip': 26.73770523071289, 'timing/data': 0.025111770257353783, 'timing/model': 1.3925813771784306, 'learning_rate/qwen_vl_interface': 2.0050673075557245e-06, 'learning_rate/action_model': 2.0050673075557243e-05, 'epoch': 1.44}) +[2026-06-16 05:57:00,433][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.028261393308639526, 'train/grad_norm_pre_clip': 10.742615699768066, 'timing/data': 0.02682766132056713, 'timing/model': 1.3971882592886686, 'learning_rate/qwen_vl_interface': 2.0028439522290677e-06, 'learning_rate/action_model': 2.0028439522290676e-05, 'epoch': 1.44}) +[2026-06-16 05:57:06,027][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.00765629205852747, 'train/grad_norm_pre_clip': 1.8785635232925415, 'timing/data': 0.0261673703789711, 'timing/model': 1.3949047718197107, 'learning_rate/qwen_vl_interface': 2.000621931685606e-06, 'learning_rate/action_model': 2.000621931685606e-05, 'epoch': 1.44}) +[2026-06-16 05:57:11,655][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.0054568033665418625, 'train/grad_norm_pre_clip': 2.3480138778686523, 'timing/data': 0.025528056547045708, 'timing/model': 1.3963510431349277, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 1.44}) +[2026-06-16 05:57:17,268][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.10894676297903061, 'train/grad_norm_pre_clip': 13.465227127075195, 'timing/data': 0.02833244577050209, 'timing/model': 1.57001275382936, 'learning_rate/qwen_vl_interface': 1.9961818986012767e-06, 'learning_rate/action_model': 1.9961818986012763e-05, 'epoch': 1.44}) +[2026-06-16 05:57:22,877][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.2001509666442871, 'train/grad_norm_pre_clip': 17.951269149780273, 'timing/data': 0.026243412867188454, 'timing/model': 1.568112574517727, 'learning_rate/qwen_vl_interface': 1.9939638878855372e-06, 'learning_rate/action_model': 1.993963887885537e-05, 'epoch': 1.44}) +[2026-06-16 05:57:28,315][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.04011757671833038, 'train/grad_norm_pre_clip': 6.320266246795654, 'timing/data': 0.02611098811030388, 'timing/model': 1.3970402255654335, 'learning_rate/qwen_vl_interface': 1.991747215603252e-06, 'learning_rate/action_model': 1.991747215603252e-05, 'epoch': 1.44}) +[2026-06-16 05:57:33,912][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.15059088170528412, 'train/grad_norm_pre_clip': 14.855413436889648, 'timing/data': 0.026047509163618088, 'timing/model': 1.3966529723256826, 'learning_rate/qwen_vl_interface': 1.9895318826656116e-06, 'learning_rate/action_model': 1.9895318826656114e-05, 'epoch': 1.44}) +[2026-06-16 05:57:39,509][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.12470733374357224, 'train/grad_norm_pre_clip': 13.040382385253906, 'timing/data': 0.026233896613121033, 'timing/model': 1.4003027435392141, 'learning_rate/qwen_vl_interface': 1.9873178899832553e-06, 'learning_rate/action_model': 1.9873178899832554e-05, 'epoch': 1.44}) +[2026-06-16 05:57:45,141][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.0024059792049229145, 'train/grad_norm_pre_clip': 0.5580805540084839, 'timing/data': 0.02695273980498314, 'timing/model': 1.4013932440429926, 'learning_rate/qwen_vl_interface': 1.985105238466272e-06, 'learning_rate/action_model': 1.9851052384662722e-05, 'epoch': 1.44}) +[2026-06-16 05:57:50,755][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.003144612768664956, 'train/grad_norm_pre_clip': 0.46439483761787415, 'timing/data': 0.025357156991958618, 'timing/model': 1.5738060139119625, 'learning_rate/qwen_vl_interface': 1.9828939290242e-06, 'learning_rate/action_model': 1.9828939290241998e-05, 'epoch': 1.44}) +[2026-06-16 05:57:56,191][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.12138239294290543, 'train/grad_norm_pre_clip': 8.616868019104004, 'timing/data': 0.027445603162050247, 'timing/model': 1.398422982543707, 'learning_rate/qwen_vl_interface': 1.980683962566024e-06, 'learning_rate/action_model': 1.980683962566024e-05, 'epoch': 1.45}) +[2026-06-16 05:58:01,820][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.2443179041147232, 'train/grad_norm_pre_clip': 31.339570999145508, 'timing/data': 0.026135466992855072, 'timing/model': 1.400578860193491, 'learning_rate/qwen_vl_interface': 1.9784753400001784e-06, 'learning_rate/action_model': 1.978475340000178e-05, 'epoch': 1.45}) +[2026-06-16 05:58:07,414][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.008326752111315727, 'train/grad_norm_pre_clip': 1.5329030752182007, 'timing/data': 0.025547459721565247, 'timing/model': 1.397529212757945, 'learning_rate/qwen_vl_interface': 1.9762680622345443e-06, 'learning_rate/action_model': 1.9762680622345442e-05, 'epoch': 1.45}) +[2026-06-16 05:58:13,048][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.08202304691076279, 'train/grad_norm_pre_clip': 23.832874298095703, 'timing/data': 0.026118027046322823, 'timing/model': 1.3929674476385117, 'learning_rate/qwen_vl_interface': 1.9740621301764513e-06, 'learning_rate/action_model': 1.9740621301764513e-05, 'epoch': 1.45}) +[2026-06-16 05:58:18,602][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.2527441680431366, 'train/grad_norm_pre_clip': 9.912349700927734, 'timing/data': 0.025355087593197823, 'timing/model': 1.3949782215058804, 'learning_rate/qwen_vl_interface': 1.971857544732672e-06, 'learning_rate/action_model': 1.971857544732672e-05, 'epoch': 1.45}) +[2026-06-16 05:58:24,155][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.03261854127049446, 'train/grad_norm_pre_clip': 11.132781982421875, 'timing/data': 0.027083154767751694, 'timing/model': 1.5135839991271496, 'learning_rate/qwen_vl_interface': 1.96965430680943e-06, 'learning_rate/action_model': 1.96965430680943e-05, 'epoch': 1.45}) +[2026-06-16 05:58:29,583][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.20821550488471985, 'train/grad_norm_pre_clip': 35.40183639526367, 'timing/data': 0.02696857415139675, 'timing/model': 1.3998390156775713, 'learning_rate/qwen_vl_interface': 1.967452417312393e-06, 'learning_rate/action_model': 1.9674524173123927e-05, 'epoch': 1.45}) +[2026-06-16 05:58:35,225][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.06034388393163681, 'train/grad_norm_pre_clip': 4.881478786468506, 'timing/data': 0.026468859985470772, 'timing/model': 1.400600302964449, 'learning_rate/qwen_vl_interface': 1.9652518771466746e-06, 'learning_rate/action_model': 1.9652518771466747e-05, 'epoch': 1.45}) +[2026-06-16 05:58:40,821][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.12294666469097137, 'train/grad_norm_pre_clip': 5.8657073974609375, 'timing/data': 0.025354618206620216, 'timing/model': 1.3969394583255053, 'learning_rate/qwen_vl_interface': 1.963052687216834e-06, 'learning_rate/action_model': 1.963052687216834e-05, 'epoch': 1.45}) +[2026-06-16 05:58:46,422][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.14012329280376434, 'train/grad_norm_pre_clip': 10.302680015563965, 'timing/data': 0.0263979472219944, 'timing/model': 1.3967881053686142, 'learning_rate/qwen_vl_interface': 1.960854848426874e-06, 'learning_rate/action_model': 1.960854848426874e-05, 'epoch': 1.45}) +[2026-06-16 05:58:52,027][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.008410017937421799, 'train/grad_norm_pre_clip': 2.068296194076538, 'timing/data': 0.02641315944492817, 'timing/model': 1.3966337386518717, 'learning_rate/qwen_vl_interface': 1.9586583616802446e-06, 'learning_rate/action_model': 1.9586583616802446e-05, 'epoch': 1.45}) +[2026-06-16 05:58:57,578][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.03751238062977791, 'train/grad_norm_pre_clip': 12.095239639282227, 'timing/data': 0.025935404002666473, 'timing/model': 1.5148509666323662, 'learning_rate/qwen_vl_interface': 1.956463227879838e-06, 'learning_rate/action_model': 1.9564632278798377e-05, 'epoch': 1.45}) +[2026-06-16 05:59:03,007][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.11632346361875534, 'train/grad_norm_pre_clip': 13.96801471710205, 'timing/data': 0.026049647480249405, 'timing/model': 1.3960442300885916, 'learning_rate/qwen_vl_interface': 1.95426944792799e-06, 'learning_rate/action_model': 1.95426944792799e-05, 'epoch': 1.45}) +[2026-06-16 05:59:08,632][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.04844203218817711, 'train/grad_norm_pre_clip': 10.330869674682617, 'timing/data': 0.02534356713294983, 'timing/model': 1.3964684326201677, 'learning_rate/qwen_vl_interface': 1.9520770227264817e-06, 'learning_rate/action_model': 1.9520770227264817e-05, 'epoch': 1.45}) +[2026-06-16 05:59:14,243][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.06104956194758415, 'train/grad_norm_pre_clip': 15.391193389892578, 'timing/data': 0.026086430996656418, 'timing/model': 1.3982287775725126, 'learning_rate/qwen_vl_interface': 1.9498859531765373e-06, 'learning_rate/action_model': 1.9498859531765372e-05, 'epoch': 1.45}) +[2026-06-16 05:59:19,875][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.0019292397191748023, 'train/grad_norm_pre_clip': 0.41400691866874695, 'timing/data': 0.02633548714220524, 'timing/model': 1.3951476532965899, 'learning_rate/qwen_vl_interface': 1.9476962401788204e-06, 'learning_rate/action_model': 1.9476962401788204e-05, 'epoch': 1.45}) +[2026-06-16 05:59:25,492][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.006316420156508684, 'train/grad_norm_pre_clip': 0.9509788751602173, 'timing/data': 0.025917932391166687, 'timing/model': 1.5688973553478718, 'learning_rate/qwen_vl_interface': 1.9455078846334397e-06, 'learning_rate/action_model': 1.9455078846334395e-05, 'epoch': 1.45}) +[2026-06-16 05:59:30,935][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.031167365610599518, 'train/grad_norm_pre_clip': 17.64507484436035, 'timing/data': 0.02633766084909439, 'timing/model': 1.3967211693525314, 'learning_rate/qwen_vl_interface': 1.9433208874399465e-06, 'learning_rate/action_model': 1.9433208874399464e-05, 'epoch': 1.45}) +[2026-06-16 05:59:36,582][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.0012199769262224436, 'train/grad_norm_pre_clip': 0.18826892971992493, 'timing/data': 0.027249330654740334, 'timing/model': 1.397824365645647, 'learning_rate/qwen_vl_interface': 1.9411352494973314e-06, 'learning_rate/action_model': 1.9411352494973315e-05, 'epoch': 1.45}) +[2026-06-16 05:59:42,104][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.4415878653526306, 'train/grad_norm_pre_clip': 24.948318481445312, 'timing/data': 0.02642563171684742, 'timing/model': 1.401450116187334, 'learning_rate/qwen_vl_interface': 1.938950971704031e-06, 'learning_rate/action_model': 1.9389509717040308e-05, 'epoch': 1.45}) +[2026-06-16 05:59:47,641][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.00017992232460528612, 'train/grad_norm_pre_clip': 0.03523539379239082, 'timing/data': 0.026788165792822838, 'timing/model': 1.3976628836244345, 'learning_rate/qwen_vl_interface': 1.936768054957915e-06, 'learning_rate/action_model': 1.9367680549579147e-05, 'epoch': 1.45}) +[2026-06-16 05:59:53,284][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.17264290153980255, 'train/grad_norm_pre_clip': 13.911044120788574, 'timing/data': 0.02629874274134636, 'timing/model': 1.398479724302888, 'learning_rate/qwen_vl_interface': 1.9345865001562994e-06, 'learning_rate/action_model': 1.9345865001562994e-05, 'epoch': 1.45}) +[2026-06-16 05:59:58,896][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.026354674249887466, 'train/grad_norm_pre_clip': 6.073163032531738, 'timing/data': 0.026160918176174164, 'timing/model': 1.5659414324909449, 'learning_rate/qwen_vl_interface': 1.9324063081959394e-06, 'learning_rate/action_model': 1.9324063081959395e-05, 'epoch': 1.45}) +[2026-06-16 06:00:04,347][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.26277366280555725, 'train/grad_norm_pre_clip': 15.274876594543457, 'timing/data': 0.02676633559167385, 'timing/model': 1.4037628211081028, 'learning_rate/qwen_vl_interface': 1.9302274799730304e-06, 'learning_rate/action_model': 1.9302274799730306e-05, 'epoch': 1.45}) +[2026-06-16 06:00:10,035][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.1854623556137085, 'train/grad_norm_pre_clip': 11.521636962890625, 'timing/data': 0.026229195296764374, 'timing/model': 1.399486467242241, 'learning_rate/qwen_vl_interface': 1.9280500163832055e-06, 'learning_rate/action_model': 1.9280500163832055e-05, 'epoch': 1.45}) +[2026-06-16 06:00:15,668][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.1430807113647461, 'train/grad_norm_pre_clip': 18.471784591674805, 'timing/data': 0.026406900957226753, 'timing/model': 1.408936643972993, 'learning_rate/qwen_vl_interface': 1.925873918321537e-06, 'learning_rate/action_model': 1.925873918321537e-05, 'epoch': 1.45}) +[2026-06-16 06:00:22,045][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.0011643101461231709, 'train/grad_norm_pre_clip': 0.15781676769256592, 'timing/data': 0.03080303594470024, 'timing/model': 1.4993910901248455, 'learning_rate/qwen_vl_interface': 1.92369918668254e-06, 'learning_rate/action_model': 1.92369918668254e-05, 'epoch': 1.46}) +[2026-06-16 06:00:28,076][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.09644461423158646, 'train/grad_norm_pre_clip': 10.609660148620605, 'timing/data': 0.03259549103677273, 'timing/model': 1.5060135964304209, 'learning_rate/qwen_vl_interface': 1.9215258223601605e-06, 'learning_rate/action_model': 1.9215258223601605e-05, 'epoch': 1.46}) +[2026-06-16 06:00:34,457][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.09085346758365631, 'train/grad_norm_pre_clip': 14.204328536987305, 'timing/data': 0.039919404312968254, 'timing/model': 1.7993874661624432, 'learning_rate/qwen_vl_interface': 1.9193538262477874e-06, 'learning_rate/action_model': 1.9193538262477874e-05, 'epoch': 1.46}) +[2026-06-16 06:00:40,518][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.07009681314229965, 'train/grad_norm_pre_clip': 11.342291831970215, 'timing/data': 0.041082676500082016, 'timing/model': 1.5436126049607992, 'learning_rate/qwen_vl_interface': 1.917183199238247e-06, 'learning_rate/action_model': 1.917183199238247e-05, 'epoch': 1.46}) +[2026-06-16 06:00:46,734][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.018547216430306435, 'train/grad_norm_pre_clip': 2.586021661758423, 'timing/data': 0.031155936419963837, 'timing/model': 1.4838256128132343, 'learning_rate/qwen_vl_interface': 1.9150139422238016e-06, 'learning_rate/action_model': 1.9150139422238018e-05, 'epoch': 1.46}) +[2026-06-16 06:00:52,718][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.16541649401187897, 'train/grad_norm_pre_clip': 16.453678131103516, 'timing/data': 0.031936608254909515, 'timing/model': 1.4830863270908594, 'learning_rate/qwen_vl_interface': 1.9128460560961513e-06, 'learning_rate/action_model': 1.9128460560961514e-05, 'epoch': 1.46}) +[2026-06-16 06:00:58,732][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.13615912199020386, 'train/grad_norm_pre_clip': 14.017931938171387, 'timing/data': 0.031112531200051308, 'timing/model': 1.4835320077836514, 'learning_rate/qwen_vl_interface': 1.910679541746432e-06, 'learning_rate/action_model': 1.9106795417464323e-05, 'epoch': 1.46}) +[2026-06-16 06:01:04,716][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.02287297695875168, 'train/grad_norm_pre_clip': 2.7653579711914062, 'timing/data': 0.03237318433821201, 'timing/model': 1.4851452894508839, 'learning_rate/qwen_vl_interface': 1.9085144000652166e-06, 'learning_rate/action_model': 1.9085144000652166e-05, 'epoch': 1.46}) +[2026-06-16 06:01:10,705][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.07382712513208389, 'train/grad_norm_pre_clip': 12.24573802947998, 'timing/data': 0.03165235370397568, 'timing/model': 1.6833523605018854, 'learning_rate/qwen_vl_interface': 1.9063506319425116e-06, 'learning_rate/action_model': 1.9063506319425114e-05, 'epoch': 1.46}) +[2026-06-16 06:01:16,420][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.0019343075109645724, 'train/grad_norm_pre_clip': 0.3120228946208954, 'timing/data': 0.03228738531470299, 'timing/model': 1.4476713482290506, 'learning_rate/qwen_vl_interface': 1.9041882382677613e-06, 'learning_rate/action_model': 1.9041882382677612e-05, 'epoch': 1.46}) +[2026-06-16 06:01:22,053][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.03620036318898201, 'train/grad_norm_pre_clip': 12.657506942749023, 'timing/data': 0.02637166902422905, 'timing/model': 1.3970711510628462, 'learning_rate/qwen_vl_interface': 1.9020272199298452e-06, 'learning_rate/action_model': 1.902027219929845e-05, 'epoch': 1.46}) +[2026-06-16 06:01:27,662][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.004421536345034838, 'train/grad_norm_pre_clip': 0.7996723055839539, 'timing/data': 0.025507979094982147, 'timing/model': 1.4050485733896494, 'learning_rate/qwen_vl_interface': 1.8998675778170736e-06, 'learning_rate/action_model': 1.8998675778170734e-05, 'epoch': 1.46}) +[2026-06-16 06:01:33,238][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.14039131999015808, 'train/grad_norm_pre_clip': 10.172895431518555, 'timing/data': 0.026011275127530098, 'timing/model': 1.396714212372899, 'learning_rate/qwen_vl_interface': 1.8977093128171955e-06, 'learning_rate/action_model': 1.8977093128171954e-05, 'epoch': 1.46}) +[2026-06-16 06:01:38,881][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.024158790707588196, 'train/grad_norm_pre_clip': 5.896144866943359, 'timing/data': 0.025510458275675774, 'timing/model': 1.3913720920681953, 'learning_rate/qwen_vl_interface': 1.8955524258173919e-06, 'learning_rate/action_model': 1.895552425817392e-05, 'epoch': 1.46}) +[2026-06-16 06:01:44,511][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.028770921751856804, 'train/grad_norm_pre_clip': 20.270902633666992, 'timing/data': 0.02614780142903328, 'timing/model': 1.5747416447848082, 'learning_rate/qwen_vl_interface': 1.8933969177042769e-06, 'learning_rate/action_model': 1.8933969177042768e-05, 'epoch': 1.46}) +[2026-06-16 06:01:49,988][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.006249390076845884, 'train/grad_norm_pre_clip': 1.2644786834716797, 'timing/data': 0.025092437863349915, 'timing/model': 1.3947067242115736, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 1.46}) +[2026-06-16 06:01:55,659][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.03914094343781471, 'train/grad_norm_pre_clip': 28.661619186401367, 'timing/data': 0.02585330232977867, 'timing/model': 1.4011217094957829, 'learning_rate/qwen_vl_interface': 1.8890900416817398e-06, 'learning_rate/action_model': 1.8890900416817397e-05, 'epoch': 1.46}) +[2026-06-16 06:02:01,243][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.0003886528720613569, 'train/grad_norm_pre_clip': 0.07672416418790817, 'timing/data': 0.02615472488105297, 'timing/model': 1.3988038338720798, 'learning_rate/qwen_vl_interface': 1.8869386755427117e-06, 'learning_rate/action_model': 1.8869386755427114e-05, 'epoch': 1.46}) +[2026-06-16 06:02:07,010][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.0901830792427063, 'train/grad_norm_pre_clip': 8.846266746520996, 'timing/data': 0.03477686457335949, 'timing/model': 1.596083004027605, 'learning_rate/qwen_vl_interface': 1.8847886918311597e-06, 'learning_rate/action_model': 1.8847886918311598e-05, 'epoch': 1.46}) +[2026-06-16 06:02:12,912][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.00046796698006801307, 'train/grad_norm_pre_clip': 0.0768827348947525, 'timing/data': 0.03386988304555416, 'timing/model': 1.464641721919179, 'learning_rate/qwen_vl_interface': 1.8826400914308616e-06, 'learning_rate/action_model': 1.8826400914308613e-05, 'epoch': 1.46}) +[2026-06-16 06:02:18,835][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.030182011425495148, 'train/grad_norm_pre_clip': 6.534542560577393, 'timing/data': 0.03440755978226662, 'timing/model': 1.664335235953331, 'learning_rate/qwen_vl_interface': 1.880492875225025e-06, 'learning_rate/action_model': 1.8804928752250252e-05, 'epoch': 1.46}) +[2026-06-16 06:02:24,519][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.011684195138514042, 'train/grad_norm_pre_clip': 4.072813510894775, 'timing/data': 0.03434109687805176, 'timing/model': 1.4603870064020157, 'learning_rate/qwen_vl_interface': 1.8783470440962922e-06, 'learning_rate/action_model': 1.878347044096292e-05, 'epoch': 1.46}) +[2026-06-16 06:02:30,520][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.20307710766792297, 'train/grad_norm_pre_clip': 7.953860282897949, 'timing/data': 0.03096933476626873, 'timing/model': 1.528106302022934, 'learning_rate/qwen_vl_interface': 1.8762025989267293e-06, 'learning_rate/action_model': 1.876202598926729e-05, 'epoch': 1.46}) +[2026-06-16 06:02:36,743][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.0016721008578315377, 'train/grad_norm_pre_clip': 0.35555607080459595, 'timing/data': 0.03381245769560337, 'timing/model': 1.7502705790102482, 'learning_rate/qwen_vl_interface': 1.874059540597839e-06, 'learning_rate/action_model': 1.874059540597839e-05, 'epoch': 1.46}) +[2026-06-16 06:02:42,623][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.10315766930580139, 'train/grad_norm_pre_clip': 13.444005012512207, 'timing/data': 0.041178856045007706, 'timing/model': 1.4573111273348331, 'learning_rate/qwen_vl_interface': 1.8719178699905508e-06, 'learning_rate/action_model': 1.8719178699905508e-05, 'epoch': 1.46}) +[2026-06-16 06:02:48,549][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.01601334847509861, 'train/grad_norm_pre_clip': 2.2701075077056885, 'timing/data': 0.032432448118925095, 'timing/model': 1.4571976773440838, 'learning_rate/qwen_vl_interface': 1.8697775879852242e-06, 'learning_rate/action_model': 1.8697775879852242e-05, 'epoch': 1.46}) +[2026-06-16 06:02:54,456][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.09958526492118835, 'train/grad_norm_pre_clip': 14.834897994995117, 'timing/data': 0.03444887883961201, 'timing/model': 1.6763820108026266, 'learning_rate/qwen_vl_interface': 1.86763869546165e-06, 'learning_rate/action_model': 1.86763869546165e-05, 'epoch': 1.47}) +[2026-06-16 06:03:00,337][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.05494683235883713, 'train/grad_norm_pre_clip': 8.55231761932373, 'timing/data': 0.03289595618844032, 'timing/model': 1.4606884215027094, 'learning_rate/qwen_vl_interface': 1.865501193299045e-06, 'learning_rate/action_model': 1.8655011932990447e-05, 'epoch': 1.47}) +[2026-06-16 06:03:06,247][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.012589728459715843, 'train/grad_norm_pre_clip': 2.164497137069702, 'timing/data': 0.033219145610928535, 'timing/model': 1.4570334032177925, 'learning_rate/qwen_vl_interface': 1.8633650823760555e-06, 'learning_rate/action_model': 1.8633650823760555e-05, 'epoch': 1.47}) +[2026-06-16 06:03:12,124][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.009716169908642769, 'train/grad_norm_pre_clip': 1.0678577423095703, 'timing/data': 0.030327413231134415, 'timing/model': 1.4552756249904633, 'learning_rate/qwen_vl_interface': 1.8612303635707562e-06, 'learning_rate/action_model': 1.861230363570756e-05, 'epoch': 1.47}) +[2026-06-16 06:03:18,000][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.013390561565756798, 'train/grad_norm_pre_clip': 1.5019748210906982, 'timing/data': 0.03178765997290611, 'timing/model': 1.4589392058551311, 'learning_rate/qwen_vl_interface': 1.8590970377606494e-06, 'learning_rate/action_model': 1.859097037760649e-05, 'epoch': 1.47}) +[2026-06-16 06:03:23,904][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.05184655636548996, 'train/grad_norm_pre_clip': 8.018607139587402, 'timing/data': 0.03162672370672226, 'timing/model': 1.4569496158510447, 'learning_rate/qwen_vl_interface': 1.8569651058226667e-06, 'learning_rate/action_model': 1.8569651058226666e-05, 'epoch': 1.47}) +[2026-06-16 06:03:29,773][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.030344324186444283, 'train/grad_norm_pre_clip': 5.998843193054199, 'timing/data': 0.0316240843385458, 'timing/model': 1.656374106183648, 'learning_rate/qwen_vl_interface': 1.854834568633161e-06, 'learning_rate/action_model': 1.854834568633161e-05, 'epoch': 1.47}) +[2026-06-16 06:03:35,600][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.08438780903816223, 'train/grad_norm_pre_clip': 11.616172790527344, 'timing/data': 0.03122769482433796, 'timing/model': 1.6111863125115633, 'learning_rate/qwen_vl_interface': 1.8527054270679174e-06, 'learning_rate/action_model': 1.8527054270679172e-05, 'epoch': 1.47}) +[2026-06-16 06:03:41,289][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.0021539952140301466, 'train/grad_norm_pre_clip': 0.7406597137451172, 'timing/data': 0.03122432716190815, 'timing/model': 1.4554597083479166, 'learning_rate/qwen_vl_interface': 1.850577682002146e-06, 'learning_rate/action_model': 1.850577682002146e-05, 'epoch': 1.47}) +[2026-06-16 06:03:47,212][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.0019276137463748455, 'train/grad_norm_pre_clip': 0.4366266429424286, 'timing/data': 0.03108988329768181, 'timing/model': 1.4523608274757862, 'learning_rate/qwen_vl_interface': 1.848451334310482e-06, 'learning_rate/action_model': 1.848451334310482e-05, 'epoch': 1.47}) +[2026-06-16 06:03:52,819][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.11896776407957077, 'train/grad_norm_pre_clip': 17.835128784179688, 'timing/data': 0.025849275290966034, 'timing/model': 1.3982511777430773, 'learning_rate/qwen_vl_interface': 1.8463263848669865e-06, 'learning_rate/action_model': 1.8463263848669864e-05, 'epoch': 1.47}) +[2026-06-16 06:03:58,514][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.07164380699396133, 'train/grad_norm_pre_clip': 22.154035568237305, 'timing/data': 0.026784492656588554, 'timing/model': 1.3990004993975163, 'learning_rate/qwen_vl_interface': 1.8442028345451468e-06, 'learning_rate/action_model': 1.8442028345451468e-05, 'epoch': 1.47}) +[2026-06-16 06:04:04,130][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.4286271631717682, 'train/grad_norm_pre_clip': 33.61179733276367, 'timing/data': 0.02541142702102661, 'timing/model': 1.5683813001960516, 'learning_rate/qwen_vl_interface': 1.8420806842178746e-06, 'learning_rate/action_model': 1.8420806842178745e-05, 'epoch': 1.47}) +[2026-06-16 06:04:09,769][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.008197267539799213, 'train/grad_norm_pre_clip': 2.1274020671844482, 'timing/data': 0.027399558573961258, 'timing/model': 1.5746962390840054, 'learning_rate/qwen_vl_interface': 1.8399599347575034e-06, 'learning_rate/action_model': 1.8399599347575033e-05, 'epoch': 1.47}) +[2026-06-16 06:11:17,337][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.017055204138159752, 'train/grad_norm_pre_clip': 4.008745193481445, 'eval/loss': 0.04648386687040329, 'eval/flap_precision': 0.9617486338797814, 'eval/flap_recall': 0.9552238805970149, 'eval/flap_f1': 0.9584751531654188, 'eval/flap_rate_model': 0.20333333333333334, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9584751531654188, 'timing/data': 0.02755908854305744, 'timing/model': 1.3973112050443888, 'learning_rate/qwen_vl_interface': 1.8378405870357943e-06, 'learning_rate/action_model': 1.837840587035794e-05, 'epoch': 1.47}) +[2026-06-16 06:11:17,338][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3800_state +[2026-06-16 06:11:17,342][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 06:11:23,207][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3800_state/model.safetensors +[2026-06-16 06:11:33,286][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3800_state/optimizer.bin +[2026-06-16 06:11:33,287][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3800_state/sampler.bin +[2026-06-16 06:11:33,287][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3800_state/sampler_1.bin +[2026-06-16 06:11:33,288][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3800_state/random_states_0.pkl +[2026-06-16 06:11:33,386][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3800 +[2026-06-16 06:11:33,386][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 06:11:33,394][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 06:11:39,090][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.1493464708328247, 'train/grad_norm_pre_clip': 16.504817962646484, 'timing/data': 0.02682545967400074, 'timing/model': 1.4360624514520168, 'learning_rate/qwen_vl_interface': 1.8357226419239316e-06, 'learning_rate/action_model': 1.8357226419239315e-05, 'epoch': 1.47}) +[2026-06-16 06:11:44,997][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.11995163559913635, 'train/grad_norm_pre_clip': 12.086732864379883, 'timing/data': 0.02750602550804615, 'timing/model': 1.4435331337153912, 'learning_rate/qwen_vl_interface': 1.8336061002925224e-06, 'learning_rate/action_model': 1.8336061002925224e-05, 'epoch': 1.47}) +[2026-06-16 06:11:50,902][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.1310989409685135, 'train/grad_norm_pre_clip': 14.520004272460938, 'timing/data': 0.026406999677419662, 'timing/model': 1.4308785796165466, 'learning_rate/qwen_vl_interface': 1.8314909630115985e-06, 'learning_rate/action_model': 1.8314909630115983e-05, 'epoch': 1.47}) +[2026-06-16 06:11:56,893][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.07663599401712418, 'train/grad_norm_pre_clip': 4.556574821472168, 'timing/data': 0.03196814842522144, 'timing/model': 1.5150846429169178, 'learning_rate/qwen_vl_interface': 1.8293772309506092e-06, 'learning_rate/action_model': 1.829377230950609e-05, 'epoch': 1.47}) +[2026-06-16 06:12:02,797][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.24475368857383728, 'train/grad_norm_pre_clip': 13.769098281860352, 'timing/data': 0.026029445230960846, 'timing/model': 1.434354668483138, 'learning_rate/qwen_vl_interface': 1.8272649049784326e-06, 'learning_rate/action_model': 1.8272649049784323e-05, 'epoch': 1.47}) +[2026-06-16 06:12:08,683][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.2755088806152344, 'train/grad_norm_pre_clip': 11.486653327941895, 'timing/data': 0.025382770225405693, 'timing/model': 1.6221828740090132, 'learning_rate/qwen_vl_interface': 1.8251539859633653e-06, 'learning_rate/action_model': 1.825153985963365e-05, 'epoch': 1.47}) +[2026-06-16 06:12:14,402][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.005955197848379612, 'train/grad_norm_pre_clip': 1.3915302753448486, 'timing/data': 0.02553188055753708, 'timing/model': 1.4407967757433653, 'learning_rate/qwen_vl_interface': 1.8230444747731265e-06, 'learning_rate/action_model': 1.8230444747731264e-05, 'epoch': 1.47}) +[2026-06-16 06:12:20,271][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.0028082409407943487, 'train/grad_norm_pre_clip': 0.5834982395172119, 'timing/data': 0.03048049286007881, 'timing/model': 1.4344468750059605, 'learning_rate/qwen_vl_interface': 1.820936372274857e-06, 'learning_rate/action_model': 1.820936372274857e-05, 'epoch': 1.47}) +[2026-06-16 06:12:26,690][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.30927908420562744, 'train/grad_norm_pre_clip': 16.220664978027344, 'timing/data': 0.032132040709257126, 'timing/model': 1.629748746752739, 'learning_rate/qwen_vl_interface': 1.8188296793351178e-06, 'learning_rate/action_model': 1.818829679335118e-05, 'epoch': 1.47}) +[2026-06-16 06:12:33,173][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.007078371476382017, 'train/grad_norm_pre_clip': 1.4379302263259888, 'timing/data': 0.03113817796111107, 'timing/model': 1.5234568044543266, 'learning_rate/qwen_vl_interface': 1.8167243968198905e-06, 'learning_rate/action_model': 1.8167243968198905e-05, 'epoch': 1.47}) +[2026-06-16 06:12:39,462][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.0015619074692949653, 'train/grad_norm_pre_clip': 0.29028335213661194, 'timing/data': 0.031970176845788956, 'timing/model': 1.5566944535821676, 'learning_rate/qwen_vl_interface': 1.8146205255945775e-06, 'learning_rate/action_model': 1.8146205255945777e-05, 'epoch': 1.47}) +[2026-06-16 06:12:45,735][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.12051621824502945, 'train/grad_norm_pre_clip': 22.39763069152832, 'timing/data': 0.03496871702373028, 'timing/model': 1.7001716680824757, 'learning_rate/qwen_vl_interface': 1.8125180665240007e-06, 'learning_rate/action_model': 1.8125180665240008e-05, 'epoch': 1.48}) +[2026-06-16 06:12:51,630][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.007310357876121998, 'train/grad_norm_pre_clip': 1.5611398220062256, 'timing/data': 0.033141011372208595, 'timing/model': 1.528095444664359, 'learning_rate/qwen_vl_interface': 1.8104170204724025e-06, 'learning_rate/action_model': 1.8104170204724024e-05, 'epoch': 1.48}) +[2026-06-16 06:12:57,730][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.26677632331848145, 'train/grad_norm_pre_clip': 16.327903747558594, 'timing/data': 0.032423002645373344, 'timing/model': 1.5203084331005812, 'learning_rate/qwen_vl_interface': 1.8083173883034435e-06, 'learning_rate/action_model': 1.8083173883034435e-05, 'epoch': 1.48}) +[2026-06-16 06:13:03,717][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.13191170990467072, 'train/grad_norm_pre_clip': 32.63932418823242, 'timing/data': 0.028861427679657936, 'timing/model': 1.4428680166602135, 'learning_rate/qwen_vl_interface': 1.8062191708802013e-06, 'learning_rate/action_model': 1.8062191708802012e-05, 'epoch': 1.48}) +[2026-06-16 06:13:09,616][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.024173706769943237, 'train/grad_norm_pre_clip': 4.976064682006836, 'timing/data': 0.02812405489385128, 'timing/model': 1.4441275801509619, 'learning_rate/qwen_vl_interface': 1.804122369065175e-06, 'learning_rate/action_model': 1.8041223690651748e-05, 'epoch': 1.48}) +[2026-06-16 06:13:15,504][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.1350259929895401, 'train/grad_norm_pre_clip': 23.85866928100586, 'timing/data': 0.02703634463250637, 'timing/model': 1.493681101128459, 'learning_rate/qwen_vl_interface': 1.8020269837202806e-06, 'learning_rate/action_model': 1.8020269837202804e-05, 'epoch': 1.48}) +[2026-06-16 06:13:21,499][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.0625639408826828, 'train/grad_norm_pre_clip': 12.421287536621094, 'timing/data': 0.02886143885552883, 'timing/model': 1.7227880228310823, 'learning_rate/qwen_vl_interface': 1.7999330157068517e-06, 'learning_rate/action_model': 1.7999330157068515e-05, 'epoch': 1.48}) +[2026-06-16 06:13:27,222][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.3294970691204071, 'train/grad_norm_pre_clip': 20.353395462036133, 'timing/data': 0.0254780575633049, 'timing/model': 1.4473765715956688, 'learning_rate/qwen_vl_interface': 1.7978404658856393e-06, 'learning_rate/action_model': 1.7978404658856394e-05, 'epoch': 1.48}) +[2026-06-16 06:13:33,422][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.1424906998872757, 'train/grad_norm_pre_clip': 26.260854721069336, 'timing/data': 0.027483735233545303, 'timing/model': 1.5157474651932716, 'learning_rate/qwen_vl_interface': 1.7957493351168126e-06, 'learning_rate/action_model': 1.7957493351168126e-05, 'epoch': 1.48}) +[2026-06-16 06:13:39,426][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.05209144577383995, 'train/grad_norm_pre_clip': 6.339400768280029, 'timing/data': 0.027365071699023247, 'timing/model': 1.539719857275486, 'learning_rate/qwen_vl_interface': 1.7936596242599552e-06, 'learning_rate/action_model': 1.7936596242599553e-05, 'epoch': 1.48}) +[2026-06-16 06:13:45,358][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.007259248290210962, 'train/grad_norm_pre_clip': 1.512622356414795, 'timing/data': 0.026667404919862747, 'timing/model': 1.480335272848606, 'learning_rate/qwen_vl_interface': 1.7915713341740692e-06, 'learning_rate/action_model': 1.791571334174069e-05, 'epoch': 1.48}) +[2026-06-16 06:13:51,283][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.06102217361330986, 'train/grad_norm_pre_clip': 4.549969673156738, 'timing/data': 0.025027886033058167, 'timing/model': 1.551891639828682, 'learning_rate/qwen_vl_interface': 1.7894844657175716e-06, 'learning_rate/action_model': 1.7894844657175717e-05, 'epoch': 1.48}) +[2026-06-16 06:13:57,559][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.0011699701426550746, 'train/grad_norm_pre_clip': 0.18520358204841614, 'timing/data': 0.03269588015973568, 'timing/model': 1.7031336762011051, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 1.48}) +[2026-06-16 06:14:03,569][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.002383038867264986, 'train/grad_norm_pre_clip': 0.2889554500579834, 'timing/data': 0.03178367577493191, 'timing/model': 1.519041020423174, 'learning_rate/qwen_vl_interface': 1.7853149971234883e-06, 'learning_rate/action_model': 1.785314997123488e-05, 'epoch': 1.48}) +[2026-06-16 06:14:09,782][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.0334334634244442, 'train/grad_norm_pre_clip': 13.582934379577637, 'timing/data': 0.03249269537627697, 'timing/model': 1.543398428708315, 'learning_rate/qwen_vl_interface': 1.7832323986998128e-06, 'learning_rate/action_model': 1.783232398699813e-05, 'epoch': 1.48}) +[2026-06-16 06:14:16,047][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.008749689906835556, 'train/grad_norm_pre_clip': 1.6134154796600342, 'timing/data': 0.03111620992422104, 'timing/model': 1.51799507252872, 'learning_rate/qwen_vl_interface': 1.7811512253333468e-06, 'learning_rate/action_model': 1.7811512253333467e-05, 'epoch': 1.48}) +[2026-06-16 06:14:22,135][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.08946726471185684, 'train/grad_norm_pre_clip': 18.918699264526367, 'timing/data': 0.032676417380571365, 'timing/model': 1.5060757379978895, 'learning_rate/qwen_vl_interface': 1.7790714778795823e-06, 'learning_rate/action_model': 1.779071477879582e-05, 'epoch': 1.48}) +[2026-06-16 06:14:28,534][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.0014321956550702453, 'train/grad_norm_pre_clip': 0.3869892954826355, 'timing/data': 0.03202291578054428, 'timing/model': 1.5059629455208778, 'learning_rate/qwen_vl_interface': 1.7769931571934246e-06, 'learning_rate/action_model': 1.7769931571934246e-05, 'epoch': 1.48}) +[2026-06-16 06:14:34,659][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.004722573794424534, 'train/grad_norm_pre_clip': 1.4044004678726196, 'timing/data': 0.03253774158656597, 'timing/model': 1.7554301135241985, 'learning_rate/qwen_vl_interface': 1.7749162641291933e-06, 'learning_rate/action_model': 1.7749162641291933e-05, 'epoch': 1.48}) +[2026-06-16 06:14:40,775][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.0016221536789089441, 'train/grad_norm_pre_clip': 0.29197582602500916, 'timing/data': 0.031305400654673576, 'timing/model': 1.6740405801683664, 'learning_rate/qwen_vl_interface': 1.7728407995406206e-06, 'learning_rate/action_model': 1.7728407995406205e-05, 'epoch': 1.48}) +[2026-06-16 06:14:47,132][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.015562187880277634, 'train/grad_norm_pre_clip': 2.309199333190918, 'timing/data': 0.033191412687301636, 'timing/model': 1.5346298720687628, 'learning_rate/qwen_vl_interface': 1.7707667642808513e-06, 'learning_rate/action_model': 1.7707667642808512e-05, 'epoch': 1.48}) +[2026-06-16 06:14:53,233][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.012314804829657078, 'train/grad_norm_pre_clip': 2.2307655811309814, 'timing/data': 0.033158985897898674, 'timing/model': 1.5418323781341314, 'learning_rate/qwen_vl_interface': 1.7686941592024423e-06, 'learning_rate/action_model': 1.7686941592024423e-05, 'epoch': 1.48}) +[2026-06-16 06:14:59,247][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.0057717738673090935, 'train/grad_norm_pre_clip': 1.931678056716919, 'timing/data': 0.03288051299750805, 'timing/model': 1.5471016466617584, 'learning_rate/qwen_vl_interface': 1.7666229851573662e-06, 'learning_rate/action_model': 1.7666229851573664e-05, 'epoch': 1.48}) +[2026-06-16 06:15:05,337][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.00016485110973007977, 'train/grad_norm_pre_clip': 0.03109024465084076, 'timing/data': 0.03277211636304855, 'timing/model': 1.5038232747465372, 'learning_rate/qwen_vl_interface': 1.7645532429970006e-06, 'learning_rate/action_model': 1.7645532429970004e-05, 'epoch': 1.48}) +[2026-06-16 06:15:11,449][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.0039703152142465115, 'train/grad_norm_pre_clip': 0.5819461941719055, 'timing/data': 0.03109753504395485, 'timing/model': 1.7492483705282211, 'learning_rate/qwen_vl_interface': 1.7624849335721402e-06, 'learning_rate/action_model': 1.7624849335721402e-05, 'epoch': 1.48}) +[2026-06-16 06:15:17,331][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.02671351283788681, 'train/grad_norm_pre_clip': 7.27180814743042, 'timing/data': 0.03169046342372894, 'timing/model': 1.5208391845226288, 'learning_rate/qwen_vl_interface': 1.7604180577329883e-06, 'learning_rate/action_model': 1.7604180577329883e-05, 'epoch': 1.48}) +[2026-06-16 06:15:23,428][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.015066980384290218, 'train/grad_norm_pre_clip': 2.425670862197876, 'timing/data': 0.03159008361399174, 'timing/model': 1.5184486974030733, 'learning_rate/qwen_vl_interface': 1.7583526163291604e-06, 'learning_rate/action_model': 1.7583526163291604e-05, 'epoch': 1.49}) +[2026-06-16 06:15:29,429][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.13126011192798615, 'train/grad_norm_pre_clip': 14.954702377319336, 'timing/data': 0.03144572861492634, 'timing/model': 1.5123629290610552, 'learning_rate/qwen_vl_interface': 1.7562886102096807e-06, 'learning_rate/action_model': 1.7562886102096807e-05, 'epoch': 1.49}) +[2026-06-16 06:15:35,434][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.22853444516658783, 'train/grad_norm_pre_clip': 9.69711685180664, 'timing/data': 0.03136158734560013, 'timing/model': 1.5344836432486773, 'learning_rate/qwen_vl_interface': 1.7542260402229838e-06, 'learning_rate/action_model': 1.7542260402229838e-05, 'epoch': 1.49}) +[2026-06-16 06:15:41,533][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.000802172755356878, 'train/grad_norm_pre_clip': 0.22058400511741638, 'timing/data': 0.030727749690413475, 'timing/model': 1.4982355404645205, 'learning_rate/qwen_vl_interface': 1.7521649072169152e-06, 'learning_rate/action_model': 1.752164907216915e-05, 'epoch': 1.49}) +[2026-06-16 06:15:47,636][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.02476944588124752, 'train/grad_norm_pre_clip': 5.645750999450684, 'timing/data': 0.03171119466423988, 'timing/model': 1.7229722738265991, 'learning_rate/qwen_vl_interface': 1.7501052120387285e-06, 'learning_rate/action_model': 1.7501052120387285e-05, 'epoch': 1.49}) +[2026-06-16 06:15:53,476][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.000703528115991503, 'train/grad_norm_pre_clip': 0.27188366651535034, 'timing/data': 0.028101108968257904, 'timing/model': 1.4740803837776184, 'learning_rate/qwen_vl_interface': 1.7480469555350866e-06, 'learning_rate/action_model': 1.7480469555350865e-05, 'epoch': 1.49}) +[2026-06-16 06:15:59,375][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.020194074138998985, 'train/grad_norm_pre_clip': 3.8735597133636475, 'timing/data': 0.02600734867155552, 'timing/model': 1.420295462012291, 'learning_rate/qwen_vl_interface': 1.7459901385520619e-06, 'learning_rate/action_model': 1.745990138552062e-05, 'epoch': 1.49}) +[2026-06-16 06:16:05,184][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.011331753805279732, 'train/grad_norm_pre_clip': 2.2889742851257324, 'timing/data': 0.027492200955748558, 'timing/model': 1.427256504073739, 'learning_rate/qwen_vl_interface': 1.7439347619351324e-06, 'learning_rate/action_model': 1.7439347619351324e-05, 'epoch': 1.49}) +[2026-06-16 06:16:10,991][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.07539455592632294, 'train/grad_norm_pre_clip': 9.330047607421875, 'timing/data': 0.02675282396376133, 'timing/model': 1.4501050040125847, 'learning_rate/qwen_vl_interface': 1.741880826529186e-06, 'learning_rate/action_model': 1.741880826529186e-05, 'epoch': 1.49}) +[2026-06-16 06:16:17,560][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.3442382216453552, 'train/grad_norm_pre_clip': 13.88884162902832, 'timing/data': 0.036290375515818596, 'timing/model': 1.5517393462359905, 'learning_rate/qwen_vl_interface': 1.7398283331785205e-06, 'learning_rate/action_model': 1.7398283331785203e-05, 'epoch': 1.49}) +[2026-06-16 06:16:23,394][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.008854959160089493, 'train/grad_norm_pre_clip': 1.4522367715835571, 'timing/data': 0.024934159591794014, 'timing/model': 1.6238994915038347, 'learning_rate/qwen_vl_interface': 1.737777282726835e-06, 'learning_rate/action_model': 1.737777282726835e-05, 'epoch': 1.49}) +[2026-06-16 06:16:29,110][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.0016822668258100748, 'train/grad_norm_pre_clip': 0.4006289839744568, 'timing/data': 0.024484459310770035, 'timing/model': 1.4416122920811176, 'learning_rate/qwen_vl_interface': 1.7357276760172403e-06, 'learning_rate/action_model': 1.7357276760172404e-05, 'epoch': 1.49}) +[2026-06-16 06:16:35,224][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.15790030360221863, 'train/grad_norm_pre_clip': 6.357778549194336, 'timing/data': 0.027540311217308044, 'timing/model': 1.519068231806159, 'learning_rate/qwen_vl_interface': 1.733679513892252e-06, 'learning_rate/action_model': 1.7336795138922517e-05, 'epoch': 1.49}) +[2026-06-16 06:16:41,202][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.019891614094376564, 'train/grad_norm_pre_clip': 5.299429416656494, 'timing/data': 0.026030126959085464, 'timing/model': 1.5181315522640944, 'learning_rate/qwen_vl_interface': 1.7316327971937926e-06, 'learning_rate/action_model': 1.7316327971937926e-05, 'epoch': 1.49}) +[2026-06-16 06:16:47,480][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.029784943908452988, 'train/grad_norm_pre_clip': 9.538405418395996, 'timing/data': 0.03293232247233391, 'timing/model': 1.5277212839573622, 'learning_rate/qwen_vl_interface': 1.7295875267631904e-06, 'learning_rate/action_model': 1.7295875267631904e-05, 'epoch': 1.49}) +[2026-06-16 06:16:54,672][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.26742926239967346, 'train/grad_norm_pre_clip': 23.52866554260254, 'timing/data': 0.031327059492468834, 'timing/model': 2.0074488315731287, 'learning_rate/qwen_vl_interface': 1.7275437034411778e-06, 'learning_rate/action_model': 1.7275437034411776e-05, 'epoch': 1.49}) +[2026-06-16 06:17:01,168][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.05351162329316139, 'train/grad_norm_pre_clip': 14.194795608520508, 'timing/data': 0.031414253637194633, 'timing/model': 1.7351489067077637, 'learning_rate/qwen_vl_interface': 1.7255013280678941e-06, 'learning_rate/action_model': 1.725501328067894e-05, 'epoch': 1.49}) +[2026-06-16 06:17:07,308][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.004582632798701525, 'train/grad_norm_pre_clip': 1.728738784790039, 'timing/data': 0.032751334831118584, 'timing/model': 1.6456497479230165, 'learning_rate/qwen_vl_interface': 1.7234604014828824e-06, 'learning_rate/action_model': 1.7234604014828823e-05, 'epoch': 1.49}) +[2026-06-16 06:17:13,359][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.25778162479400635, 'train/grad_norm_pre_clip': 14.75933837890625, 'timing/data': 0.032893043011426926, 'timing/model': 1.5068664476275444, 'learning_rate/qwen_vl_interface': 1.7214209245250906e-06, 'learning_rate/action_model': 1.7214209245250905e-05, 'epoch': 1.49}) +[2026-06-16 06:17:19,564][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.015689760446548462, 'train/grad_norm_pre_clip': 21.665546417236328, 'timing/data': 0.03297995217144489, 'timing/model': 1.51225271448493, 'learning_rate/qwen_vl_interface': 1.7193828980328723e-06, 'learning_rate/action_model': 1.719382898032872e-05, 'epoch': 1.49}) +[2026-06-16 06:17:25,762][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.16003365814685822, 'train/grad_norm_pre_clip': 17.34130859375, 'timing/data': 0.033927636221051216, 'timing/model': 1.5128710363060236, 'learning_rate/qwen_vl_interface': 1.71734632284398e-06, 'learning_rate/action_model': 1.7173463228439798e-05, 'epoch': 1.49}) +[2026-06-16 06:17:31,974][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.20947973430156708, 'train/grad_norm_pre_clip': 17.90546226501465, 'timing/data': 0.03203338012099266, 'timing/model': 1.5062377732247114, 'learning_rate/qwen_vl_interface': 1.7153111997955744e-06, 'learning_rate/action_model': 1.7153111997955743e-05, 'epoch': 1.49}) +[2026-06-16 06:17:38,294][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.022415366023778915, 'train/grad_norm_pre_clip': 13.301735877990723, 'timing/data': 0.030446110293269157, 'timing/model': 1.7401901502162218, 'learning_rate/qwen_vl_interface': 1.7132775297242175e-06, 'learning_rate/action_model': 1.7132775297242172e-05, 'epoch': 1.49}) +[2026-06-16 06:17:44,800][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.002923231339082122, 'train/grad_norm_pre_clip': 0.378740519285202, 'timing/data': 0.03231929428875446, 'timing/model': 1.7680874466896057, 'learning_rate/qwen_vl_interface': 1.7112453134658747e-06, 'learning_rate/action_model': 1.7112453134658747e-05, 'epoch': 1.49}) +[2026-06-16 06:17:50,865][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.04432706907391548, 'train/grad_norm_pre_clip': 19.422996520996094, 'timing/data': 0.03260400891304016, 'timing/model': 1.5339210126549006, 'learning_rate/qwen_vl_interface': 1.7092145518559129e-06, 'learning_rate/action_model': 1.709214551855913e-05, 'epoch': 1.49}) +[2026-06-16 06:17:57,064][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.020655475556850433, 'train/grad_norm_pre_clip': 9.844468116760254, 'timing/data': 0.0325778815895319, 'timing/model': 1.532310288399458, 'learning_rate/qwen_vl_interface': 1.7071852457291013e-06, 'learning_rate/action_model': 1.707185245729101e-05, 'epoch': 1.5}) +[2026-06-16 06:18:03,263][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.015605133958160877, 'train/grad_norm_pre_clip': 11.801630020141602, 'timing/data': 0.03180520795285702, 'timing/model': 1.5286102760583162, 'learning_rate/qwen_vl_interface': 1.7051573959196122e-06, 'learning_rate/action_model': 1.7051573959196122e-05, 'epoch': 1.5}) +[2026-06-16 06:18:09,445][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.04955694451928139, 'train/grad_norm_pre_clip': 8.055977821350098, 'timing/data': 0.03237984515726566, 'timing/model': 1.5019826665520668, 'learning_rate/qwen_vl_interface': 1.7031310032610166e-06, 'learning_rate/action_model': 1.7031310032610163e-05, 'epoch': 1.5}) +[2026-06-16 06:18:15,545][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.39905744791030884, 'train/grad_norm_pre_clip': 24.47775650024414, 'timing/data': 0.031591931357979774, 'timing/model': 1.7513155229389668, 'learning_rate/qwen_vl_interface': 1.701106068586289e-06, 'learning_rate/action_model': 1.701106068586289e-05, 'epoch': 1.5}) +[2026-06-16 06:18:21,191][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.12476629763841629, 'train/grad_norm_pre_clip': 18.66115951538086, 'timing/data': 0.027157962322235107, 'timing/model': 1.448536328971386, 'learning_rate/qwen_vl_interface': 1.6990825927278032e-06, 'learning_rate/action_model': 1.699082592727803e-05, 'epoch': 1.5}) +[2026-06-16 06:18:27,216][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.11140985786914825, 'train/grad_norm_pre_clip': 7.160735130310059, 'timing/data': 0.029595140367746353, 'timing/model': 1.5313006397336721, 'learning_rate/qwen_vl_interface': 1.6970605765173344e-06, 'learning_rate/action_model': 1.6970605765173346e-05, 'epoch': 1.5}) +[2026-06-16 06:18:33,326][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.08416643738746643, 'train/grad_norm_pre_clip': 32.7028923034668, 'timing/data': 0.028275251388549805, 'timing/model': 1.536311537027359, 'learning_rate/qwen_vl_interface': 1.6950400207860557e-06, 'learning_rate/action_model': 1.6950400207860557e-05, 'epoch': 1.5}) +[2026-06-16 06:18:39,813][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.02715155854821205, 'train/grad_norm_pre_clip': 13.911505699157715, 'timing/data': 0.04608922824263573, 'timing/model': 1.7050333935767412, 'learning_rate/qwen_vl_interface': 1.6930209263645412e-06, 'learning_rate/action_model': 1.6930209263645412e-05, 'epoch': 1.5}) +[2026-06-16 06:18:46,296][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.1134597584605217, 'train/grad_norm_pre_clip': 15.455389976501465, 'timing/data': 0.04682129994034767, 'timing/model': 1.56373823992908, 'learning_rate/qwen_vl_interface': 1.6910032940827656e-06, 'learning_rate/action_model': 1.6910032940827657e-05, 'epoch': 1.5}) +[2026-06-16 06:18:52,510][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.002574521815404296, 'train/grad_norm_pre_clip': 1.6636650562286377, 'timing/data': 0.0320366695523262, 'timing/model': 1.6681206319481134, 'learning_rate/qwen_vl_interface': 1.6889871247701003e-06, 'learning_rate/action_model': 1.6889871247701002e-05, 'epoch': 1.5}) +[2026-06-16 06:18:59,158][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.15772102773189545, 'train/grad_norm_pre_clip': 32.14613723754883, 'timing/data': 0.032873496413230896, 'timing/model': 2.125005157664418, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 1.5}) +[2026-06-16 06:19:05,367][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.027271773666143417, 'train/grad_norm_pre_clip': 15.212116241455078, 'timing/data': 0.03557979501783848, 'timing/model': 1.521626679226756, 'learning_rate/qwen_vl_interface': 1.6849591783665847e-06, 'learning_rate/action_model': 1.6849591783665847e-05, 'epoch': 1.5}) +[2026-06-16 06:19:11,561][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.08242225646972656, 'train/grad_norm_pre_clip': 10.789417266845703, 'timing/data': 0.03251384757459164, 'timing/model': 1.5328171700239182, 'learning_rate/qwen_vl_interface': 1.6829474029314713e-06, 'learning_rate/action_model': 1.6829474029314713e-05, 'epoch': 1.5}) +[2026-06-16 06:19:17,800][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.1926274299621582, 'train/grad_norm_pre_clip': 21.825702667236328, 'timing/data': 0.025346610695123672, 'timing/model': 1.5500376019626856, 'learning_rate/qwen_vl_interface': 1.6809370937769408e-06, 'learning_rate/action_model': 1.680937093776941e-05, 'epoch': 1.5}) +[2026-06-16 06:19:23,913][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.007093227002769709, 'train/grad_norm_pre_clip': 1.7646868228912354, 'timing/data': 0.02572786994278431, 'timing/model': 1.522156249731779, 'learning_rate/qwen_vl_interface': 1.678928251729356e-06, 'learning_rate/action_model': 1.6789282517293557e-05, 'epoch': 1.5}) +[2026-06-16 06:19:29,843][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.14061076939105988, 'train/grad_norm_pre_clip': 11.097784996032715, 'timing/data': 0.026895835995674133, 'timing/model': 1.6632189527153969, 'learning_rate/qwen_vl_interface': 1.6769208776144767e-06, 'learning_rate/action_model': 1.6769208776144764e-05, 'epoch': 1.5}) +[2026-06-16 06:19:35,607][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.03363840654492378, 'train/grad_norm_pre_clip': 7.103517532348633, 'timing/data': 0.026856273412704468, 'timing/model': 1.5240336507558823, 'learning_rate/qwen_vl_interface': 1.6749149722574564e-06, 'learning_rate/action_model': 1.6749149722574564e-05, 'epoch': 1.5}) +[2026-06-16 06:19:41,618][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.23937560617923737, 'train/grad_norm_pre_clip': 14.391181945800781, 'timing/data': 0.025972066447138786, 'timing/model': 1.528712997213006, 'learning_rate/qwen_vl_interface': 1.6729105364828486e-06, 'learning_rate/action_model': 1.6729105364828484e-05, 'epoch': 1.5}) +[2026-06-16 06:19:47,601][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.011370842345058918, 'train/grad_norm_pre_clip': 4.371057987213135, 'timing/data': 0.026004686951637268, 'timing/model': 1.5198195353150368, 'learning_rate/qwen_vl_interface': 1.6709075711146e-06, 'learning_rate/action_model': 1.6709075711146e-05, 'epoch': 1.5}) +[2026-06-16 06:19:53,571][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.05591389909386635, 'train/grad_norm_pre_clip': 22.70157814025879, 'timing/data': 0.027361173182725906, 'timing/model': 1.4884341601282358, 'learning_rate/qwen_vl_interface': 1.6689060769760556e-06, 'learning_rate/action_model': 1.6689060769760557e-05, 'epoch': 1.5}) +[2026-06-16 06:19:59,477][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.04364149272441864, 'train/grad_norm_pre_clip': 7.721349716186523, 'timing/data': 0.02820395678281784, 'timing/model': 1.623571990057826, 'learning_rate/qwen_vl_interface': 1.6669060548899528e-06, 'learning_rate/action_model': 1.6669060548899525e-05, 'epoch': 1.5}) +[2026-06-16 06:20:05,191][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.09217432886362076, 'train/grad_norm_pre_clip': 25.61652946472168, 'timing/data': 0.029415978118777275, 'timing/model': 1.4261870365589857, 'learning_rate/qwen_vl_interface': 1.6649075056784253e-06, 'learning_rate/action_model': 1.6649075056784253e-05, 'epoch': 1.5}) +[2026-06-16 06:20:11,314][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.07635978609323502, 'train/grad_norm_pre_clip': 24.910926818847656, 'timing/data': 0.026174863800406456, 'timing/model': 1.6231652740389109, 'learning_rate/qwen_vl_interface': 1.6629104301630019e-06, 'learning_rate/action_model': 1.662910430163002e-05, 'epoch': 1.5}) +[2026-06-16 06:20:17,197][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.08106549829244614, 'train/grad_norm_pre_clip': 3.499309778213501, 'timing/data': 0.02539360150694847, 'timing/model': 1.5141402017325163, 'learning_rate/qwen_vl_interface': 1.6609148291646043e-06, 'learning_rate/action_model': 1.6609148291646042e-05, 'epoch': 1.5}) +[2026-06-16 06:20:23,105][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.06438082456588745, 'train/grad_norm_pre_clip': 8.310391426086426, 'timing/data': 0.026389699429273605, 'timing/model': 1.5229975804686546, 'learning_rate/qwen_vl_interface': 1.658920703503549e-06, 'learning_rate/action_model': 1.6589207035035488e-05, 'epoch': 1.5}) +[2026-06-16 06:20:29,006][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.0625896155834198, 'train/grad_norm_pre_clip': 7.724667549133301, 'timing/data': 0.026565443724393845, 'timing/model': 1.481056498363614, 'learning_rate/qwen_vl_interface': 1.6569280539995467e-06, 'learning_rate/action_model': 1.6569280539995467e-05, 'epoch': 1.5}) +[2026-06-16 06:20:35,093][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.02027302049100399, 'train/grad_norm_pre_clip': 5.1054840087890625, 'timing/data': 0.02534211240708828, 'timing/model': 1.8162690289318562, 'learning_rate/qwen_vl_interface': 1.6549368814716977e-06, 'learning_rate/action_model': 1.6549368814716977e-05, 'epoch': 1.51}) +[2026-06-16 06:20:41,181][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.14066524803638458, 'train/grad_norm_pre_clip': 21.777751922607422, 'timing/data': 0.0332427266985178, 'timing/model': 1.535094365477562, 'learning_rate/qwen_vl_interface': 1.6529471867385009e-06, 'learning_rate/action_model': 1.6529471867385007e-05, 'epoch': 1.51}) +[2026-06-16 06:20:47,462][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.334843248128891, 'train/grad_norm_pre_clip': 35.843143463134766, 'timing/data': 0.03166693262755871, 'timing/model': 1.5289612002670765, 'learning_rate/qwen_vl_interface': 1.6509589706178417e-06, 'learning_rate/action_model': 1.6509589706178417e-05, 'epoch': 1.51}) +[2026-06-16 06:20:53,660][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.15849962830543518, 'train/grad_norm_pre_clip': 28.792654037475586, 'timing/data': 0.03237790986895561, 'timing/model': 1.5283212400972843, 'learning_rate/qwen_vl_interface': 1.648972233927002e-06, 'learning_rate/action_model': 1.648972233927002e-05, 'epoch': 1.51}) +[2026-06-16 06:20:59,860][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.05249436944723129, 'train/grad_norm_pre_clip': 4.492186069488525, 'timing/data': 0.03217180073261261, 'timing/model': 1.5288354922086, 'learning_rate/qwen_vl_interface': 1.646986977482654e-06, 'learning_rate/action_model': 1.6469869774826538e-05, 'epoch': 1.51}) +[2026-06-16 06:21:06,062][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.10884202271699905, 'train/grad_norm_pre_clip': 18.992053985595703, 'timing/data': 0.032398831099271774, 'timing/model': 1.5406545028090477, 'learning_rate/qwen_vl_interface': 1.6450032021008618e-06, 'learning_rate/action_model': 1.6450032021008617e-05, 'epoch': 1.51}) +[2026-06-16 06:21:12,269][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.07986294478178024, 'train/grad_norm_pre_clip': 16.057172775268555, 'timing/data': 0.033030539751052856, 'timing/model': 1.7348755728453398, 'learning_rate/qwen_vl_interface': 1.6430209085970812e-06, 'learning_rate/action_model': 1.643020908597081e-05, 'epoch': 1.51}) +[2026-06-16 06:21:18,264][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.06375575810670853, 'train/grad_norm_pre_clip': 9.219438552856445, 'timing/data': 0.03247321769595146, 'timing/model': 1.5309562366455793, 'learning_rate/qwen_vl_interface': 1.641040097786157e-06, 'learning_rate/action_model': 1.641040097786157e-05, 'epoch': 1.51}) +[2026-06-16 06:21:24,332][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.08441077172756195, 'train/grad_norm_pre_clip': 8.802584648132324, 'timing/data': 0.032447103410959244, 'timing/model': 1.5174234192818403, 'learning_rate/qwen_vl_interface': 1.6390607704823267e-06, 'learning_rate/action_model': 1.6390607704823265e-05, 'epoch': 1.51}) +[2026-06-16 06:21:30,432][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.014250042848289013, 'train/grad_norm_pre_clip': 7.002510070800781, 'timing/data': 0.03225498087704182, 'timing/model': 1.5337116047739983, 'learning_rate/qwen_vl_interface': 1.6370829274992162e-06, 'learning_rate/action_model': 1.6370829274992162e-05, 'epoch': 1.51}) +[2026-06-16 06:21:36,736][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.37169593572616577, 'train/grad_norm_pre_clip': 18.003934860229492, 'timing/data': 0.03288185968995094, 'timing/model': 1.5034442488104105, 'learning_rate/qwen_vl_interface': 1.6351065696498422e-06, 'learning_rate/action_model': 1.6351065696498422e-05, 'epoch': 1.51}) +[2026-06-16 06:28:55,097][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.0019273062935099006, 'train/grad_norm_pre_clip': 0.2808607220649719, 'eval/loss': 0.04687529429793358, 'eval/flap_precision': 0.9461235216819974, 'eval/flap_recall': 0.9769335142469471, 'eval/flap_f1': 0.9612817089452604, 'eval/flap_rate_model': 0.21138888888888888, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9612817089452604, 'timing/data': 0.03191487491130829, 'timing/model': 1.5443839635699987, 'learning_rate/qwen_vl_interface': 1.6331316977466117e-06, 'learning_rate/action_model': 1.6331316977466115e-05, 'epoch': 1.51}) +[2026-06-16 06:28:55,098][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3900_state +[2026-06-16 06:28:55,110][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 06:29:03,345][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3900_state/model.safetensors +[2026-06-16 06:29:17,221][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3900_state/optimizer.bin +[2026-06-16 06:29:17,221][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3900_state/sampler.bin +[2026-06-16 06:29:17,221][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3900_state/sampler_1.bin +[2026-06-16 06:29:17,223][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_3900_state/random_states_0.pkl +[2026-06-16 06:29:17,334][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3900 +[2026-06-16 06:29:17,334][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 06:29:17,400][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 06:29:23,754][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.07116404175758362, 'train/grad_norm_pre_clip': 15.947209358215332, 'timing/data': 0.0329772774130106, 'timing/model': 1.5467913560569286, 'learning_rate/qwen_vl_interface': 1.6311583126013198e-06, 'learning_rate/action_model': 1.6311583126013198e-05, 'epoch': 1.51}) +[2026-06-16 06:29:29,991][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.04608249291777611, 'train/grad_norm_pre_clip': 4.302645683288574, 'timing/data': 0.03230697475373745, 'timing/model': 1.7754715774208307, 'learning_rate/qwen_vl_interface': 1.6291864150251488e-06, 'learning_rate/action_model': 1.629186415025149e-05, 'epoch': 1.51}) +[2026-06-16 06:29:36,052][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.002675105584785342, 'train/grad_norm_pre_clip': 1.1683363914489746, 'timing/data': 0.03562760166823864, 'timing/model': 1.520547028630972, 'learning_rate/qwen_vl_interface': 1.6272160058286727e-06, 'learning_rate/action_model': 1.6272160058286726e-05, 'epoch': 1.51}) +[2026-06-16 06:29:42,242][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.17569422721862793, 'train/grad_norm_pre_clip': 14.171897888183594, 'timing/data': 0.03203825652599335, 'timing/model': 1.5207290779799223, 'learning_rate/qwen_vl_interface': 1.6252470858218519e-06, 'learning_rate/action_model': 1.6252470858218517e-05, 'epoch': 1.51}) +[2026-06-16 06:29:48,357][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.0013906736858189106, 'train/grad_norm_pre_clip': 0.28443843126296997, 'timing/data': 0.03413783945143223, 'timing/model': 1.5432923696935177, 'learning_rate/qwen_vl_interface': 1.6232796558140357e-06, 'learning_rate/action_model': 1.6232796558140357e-05, 'epoch': 1.51}) +[2026-06-16 06:29:54,445][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.03278299421072006, 'train/grad_norm_pre_clip': 10.932470321655273, 'timing/data': 0.0318369846791029, 'timing/model': 1.492857314646244, 'learning_rate/qwen_vl_interface': 1.6213137166139592e-06, 'learning_rate/action_model': 1.621313716613959e-05, 'epoch': 1.51}) +[2026-06-16 06:30:00,759][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.0767548680305481, 'train/grad_norm_pre_clip': 7.094457149505615, 'timing/data': 0.044608091935515404, 'timing/model': 1.5641843620687723, 'learning_rate/qwen_vl_interface': 1.6193492690297463e-06, 'learning_rate/action_model': 1.6193492690297464e-05, 'epoch': 1.51}) +[2026-06-16 06:30:06,944][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.09302699565887451, 'train/grad_norm_pre_clip': 11.359281539916992, 'timing/data': 0.03297866880893707, 'timing/model': 1.7051528096199036, 'learning_rate/qwen_vl_interface': 1.6173863138689078e-06, 'learning_rate/action_model': 1.617386313868908e-05, 'epoch': 1.51}) +[2026-06-16 06:30:12,718][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.015256590209901333, 'train/grad_norm_pre_clip': 1.1926530599594116, 'timing/data': 0.032276805490255356, 'timing/model': 1.525069311261177, 'learning_rate/qwen_vl_interface': 1.6154248519383396e-06, 'learning_rate/action_model': 1.6154248519383395e-05, 'epoch': 1.51}) +[2026-06-16 06:30:18,820][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.10704227536916733, 'train/grad_norm_pre_clip': 11.459661483764648, 'timing/data': 0.03239774703979492, 'timing/model': 1.5341519322246313, 'learning_rate/qwen_vl_interface': 1.6134648840443248e-06, 'learning_rate/action_model': 1.613464884044325e-05, 'epoch': 1.51}) +[2026-06-16 06:30:24,853][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.006510540843009949, 'train/grad_norm_pre_clip': 2.005521774291992, 'timing/data': 0.03481004387140274, 'timing/model': 1.5620392207056284, 'learning_rate/qwen_vl_interface': 1.6115064109925334e-06, 'learning_rate/action_model': 1.6115064109925336e-05, 'epoch': 1.51}) +[2026-06-16 06:30:31,070][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.046088747680187225, 'train/grad_norm_pre_clip': 17.692459106445312, 'timing/data': 0.033657338470220566, 'timing/model': 1.540711523965001, 'learning_rate/qwen_vl_interface': 1.6095494335880173e-06, 'learning_rate/action_model': 1.6095494335880174e-05, 'epoch': 1.51}) +[2026-06-16 06:30:37,284][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.0030075327958911657, 'train/grad_norm_pre_clip': 1.0362889766693115, 'timing/data': 0.031995126977562904, 'timing/model': 1.5098693557083607, 'learning_rate/qwen_vl_interface': 1.6075939526352175e-06, 'learning_rate/action_model': 1.6075939526352175e-05, 'epoch': 1.51}) +[2026-06-16 06:30:43,633][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.025643665343523026, 'train/grad_norm_pre_clip': 9.043973922729492, 'timing/data': 0.034864770248532295, 'timing/model': 1.7803941369056702, 'learning_rate/qwen_vl_interface': 1.6056399689379576e-06, 'learning_rate/action_model': 1.6056399689379575e-05, 'epoch': 1.51}) +[2026-06-16 06:30:49,661][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.07911474257707596, 'train/grad_norm_pre_clip': 7.993592262268066, 'timing/data': 0.03287573903799057, 'timing/model': 1.5327674429863691, 'learning_rate/qwen_vl_interface': 1.6036874832994473e-06, 'learning_rate/action_model': 1.603687483299447e-05, 'epoch': 1.52}) +[2026-06-16 06:30:55,856][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.2383808195590973, 'train/grad_norm_pre_clip': 16.219757080078125, 'timing/data': 0.03206028416752815, 'timing/model': 1.5313474256545305, 'learning_rate/qwen_vl_interface': 1.6017364965222796e-06, 'learning_rate/action_model': 1.6017364965222793e-05, 'epoch': 1.52}) +[2026-06-16 06:31:02,068][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.034271273761987686, 'train/grad_norm_pre_clip': 3.6819419860839844, 'timing/data': 0.03143472038209438, 'timing/model': 1.5205510165542364, 'learning_rate/qwen_vl_interface': 1.5997870094084305e-06, 'learning_rate/action_model': 1.5997870094084306e-05, 'epoch': 1.52}) +[2026-06-16 06:31:08,262][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.11362723261117935, 'train/grad_norm_pre_clip': 7.501465320587158, 'timing/data': 0.03197849169373512, 'timing/model': 1.5332863759249449, 'learning_rate/qwen_vl_interface': 1.5978390227592616e-06, 'learning_rate/action_model': 1.5978390227592615e-05, 'epoch': 1.52}) +[2026-06-16 06:31:14,538][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.28834739327430725, 'train/grad_norm_pre_clip': 22.737327575683594, 'timing/data': 0.03263818100094795, 'timing/model': 1.5622551031410694, 'learning_rate/qwen_vl_interface': 1.5958925373755164e-06, 'learning_rate/action_model': 1.5958925373755164e-05, 'epoch': 1.52}) +[2026-06-16 06:31:20,761][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.05401262268424034, 'train/grad_norm_pre_clip': 17.841459274291992, 'timing/data': 0.03121969662606716, 'timing/model': 1.7241562344133854, 'learning_rate/qwen_vl_interface': 1.5939475540573213e-06, 'learning_rate/action_model': 1.5939475540573212e-05, 'epoch': 1.52}) +[2026-06-16 06:31:26,902][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.06629859656095505, 'train/grad_norm_pre_clip': 6.834187984466553, 'timing/data': 0.03209994360804558, 'timing/model': 1.6606627572327852, 'learning_rate/qwen_vl_interface': 1.5920040736041867e-06, 'learning_rate/action_model': 1.5920040736041866e-05, 'epoch': 1.52}) +[2026-06-16 06:31:33,077][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.003363273572176695, 'train/grad_norm_pre_clip': 0.3067390024662018, 'timing/data': 0.031161559745669365, 'timing/model': 1.5410423912107944, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 1.52}) +[2026-06-16 06:31:39,306][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.005922017619013786, 'train/grad_norm_pre_clip': 1.0941624641418457, 'timing/data': 0.027415383607149124, 'timing/model': 1.4793555922806263, 'learning_rate/qwen_vl_interface': 1.5881216244880407e-06, 'learning_rate/action_model': 1.5881216244880407e-05, 'epoch': 1.52}) +[2026-06-16 06:31:45,221][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.00877995416522026, 'train/grad_norm_pre_clip': 4.529979228973389, 'timing/data': 0.025903793051838875, 'timing/model': 1.4506932403892279, 'learning_rate/qwen_vl_interface': 1.5861826574209587e-06, 'learning_rate/action_model': 1.5861826574209587e-05, 'epoch': 1.52}) +[2026-06-16 06:31:51,310][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.03909548372030258, 'train/grad_norm_pre_clip': 27.571609497070312, 'timing/data': 0.026994699612259865, 'timing/model': 1.5281046330928802, 'learning_rate/qwen_vl_interface': 1.5842451964107916e-06, 'learning_rate/action_model': 1.5842451964107916e-05, 'epoch': 1.52}) +[2026-06-16 06:31:57,304][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.10206709057092667, 'train/grad_norm_pre_clip': 4.1269402503967285, 'timing/data': 0.027528338134288788, 'timing/model': 1.7303040605038404, 'learning_rate/qwen_vl_interface': 1.5823092422539567e-06, 'learning_rate/action_model': 1.582309242253957e-05, 'epoch': 1.52}) +[2026-06-16 06:32:03,677][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.0947338342666626, 'train/grad_norm_pre_clip': 7.823190689086914, 'timing/data': 0.035301199182868004, 'timing/model': 1.7356750015169382, 'learning_rate/qwen_vl_interface': 1.5803747957462514e-06, 'learning_rate/action_model': 1.5803747957462514e-05, 'epoch': 1.52}) +[2026-06-16 06:32:10,069][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.09701615571975708, 'train/grad_norm_pre_clip': 8.144570350646973, 'timing/data': 0.03363075852394104, 'timing/model': 1.5396293513476849, 'learning_rate/qwen_vl_interface': 1.5784418576828541e-06, 'learning_rate/action_model': 1.5784418576828543e-05, 'epoch': 1.52}) +[2026-06-16 06:32:16,231][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.2843300700187683, 'train/grad_norm_pre_clip': 20.095260620117188, 'timing/data': 0.03043188713490963, 'timing/model': 1.535019937902689, 'learning_rate/qwen_vl_interface': 1.576510428858322e-06, 'learning_rate/action_model': 1.5765104288583217e-05, 'epoch': 1.52}) +[2026-06-16 06:32:22,333][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.0001098523716791533, 'train/grad_norm_pre_clip': 0.013702652417123318, 'timing/data': 0.030857114121317863, 'timing/model': 1.5348125323653221, 'learning_rate/qwen_vl_interface': 1.5745805100665922e-06, 'learning_rate/action_model': 1.5745805100665923e-05, 'epoch': 1.52}) +[2026-06-16 06:32:28,437][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.13923601806163788, 'train/grad_norm_pre_clip': 10.125028610229492, 'timing/data': 0.03052687831223011, 'timing/model': 1.5190491005778313, 'learning_rate/qwen_vl_interface': 1.572652102100983e-06, 'learning_rate/action_model': 1.572652102100983e-05, 'epoch': 1.52}) +[2026-06-16 06:32:34,540][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.11718077957630157, 'train/grad_norm_pre_clip': 13.426165580749512, 'timing/data': 0.03153548389673233, 'timing/model': 1.738891964778304, 'learning_rate/qwen_vl_interface': 1.5707252057541898e-06, 'learning_rate/action_model': 1.5707252057541896e-05, 'epoch': 1.52}) +[2026-06-16 06:32:40,433][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.17490209639072418, 'train/grad_norm_pre_clip': 10.572718620300293, 'timing/data': 0.030675988644361496, 'timing/model': 1.5319981928914785, 'learning_rate/qwen_vl_interface': 1.5687998218182849e-06, 'learning_rate/action_model': 1.5687998218182847e-05, 'epoch': 1.52}) +[2026-06-16 06:32:46,531][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.03209099546074867, 'train/grad_norm_pre_clip': 25.06379508972168, 'timing/data': 0.03041553869843483, 'timing/model': 1.5132109317928553, 'learning_rate/qwen_vl_interface': 1.5668759510847235e-06, 'learning_rate/action_model': 1.5668759510847233e-05, 'epoch': 1.52}) +[2026-06-16 06:32:52,819][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.04009709879755974, 'train/grad_norm_pre_clip': 5.493594646453857, 'timing/data': 0.037546198815107346, 'timing/model': 1.573031997308135, 'learning_rate/qwen_vl_interface': 1.5649535943443331e-06, 'learning_rate/action_model': 1.564953594344333e-05, 'epoch': 1.52}) +[2026-06-16 06:32:59,264][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.03801889717578888, 'train/grad_norm_pre_clip': 4.7343621253967285, 'timing/data': 0.03177928924560547, 'timing/model': 1.5850719921290874, 'learning_rate/qwen_vl_interface': 1.5630327523873244e-06, 'learning_rate/action_model': 1.563032752387324e-05, 'epoch': 1.52}) +[2026-06-16 06:33:05,461][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.023306598886847496, 'train/grad_norm_pre_clip': 3.7262520790100098, 'timing/data': 0.03196340426802635, 'timing/model': 1.5706504415720701, 'learning_rate/qwen_vl_interface': 1.561113426003282e-06, 'learning_rate/action_model': 1.561113426003282e-05, 'epoch': 1.52}) +[2026-06-16 06:33:11,616][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.00952738244086504, 'train/grad_norm_pre_clip': 0.7673202157020569, 'timing/data': 0.02986746095120907, 'timing/model': 1.687100188806653, 'learning_rate/qwen_vl_interface': 1.5591956159811686e-06, 'learning_rate/action_model': 1.5591956159811687e-05, 'epoch': 1.52}) +[2026-06-16 06:33:17,671][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.009791147895157337, 'train/grad_norm_pre_clip': 5.267819404602051, 'timing/data': 0.03155668079853058, 'timing/model': 1.54208168014884, 'learning_rate/qwen_vl_interface': 1.5572793231093235e-06, 'learning_rate/action_model': 1.5572793231093236e-05, 'epoch': 1.52}) +[2026-06-16 06:33:23,977][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.08319512009620667, 'train/grad_norm_pre_clip': 16.201608657836914, 'timing/data': 0.03254295885562897, 'timing/model': 1.5283538736402988, 'learning_rate/qwen_vl_interface': 1.5553645481754626e-06, 'learning_rate/action_model': 1.5553645481754626e-05, 'epoch': 1.52}) +[2026-06-16 06:33:30,273][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.012766499072313309, 'train/grad_norm_pre_clip': 3.6432130336761475, 'timing/data': 0.0319732241332531, 'timing/model': 1.5325396377593279, 'learning_rate/qwen_vl_interface': 1.5534512919666775e-06, 'learning_rate/action_model': 1.5534512919666776e-05, 'epoch': 1.53}) +[2026-06-16 06:33:36,739][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.015353242866694927, 'train/grad_norm_pre_clip': 1.669645071029663, 'timing/data': 0.031000986695289612, 'timing/model': 1.4744160436093807, 'learning_rate/qwen_vl_interface': 1.551539555269435e-06, 'learning_rate/action_model': 1.5515395552694348e-05, 'epoch': 1.53}) +[2026-06-16 06:33:43,025][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.06642356514930725, 'train/grad_norm_pre_clip': 8.084925651550293, 'timing/data': 0.03167545981705189, 'timing/model': 1.7188686467707157, 'learning_rate/qwen_vl_interface': 1.5496293388695787e-06, 'learning_rate/action_model': 1.5496293388695786e-05, 'epoch': 1.53}) +[2026-06-16 06:33:49,103][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.13212379813194275, 'train/grad_norm_pre_clip': 20.67142677307129, 'timing/data': 0.03114272467792034, 'timing/model': 1.6824906151741743, 'learning_rate/qwen_vl_interface': 1.547720643552327e-06, 'learning_rate/action_model': 1.547720643552327e-05, 'epoch': 1.53}) +[2026-06-16 06:33:55,036][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.11566212773323059, 'train/grad_norm_pre_clip': 18.5855712890625, 'timing/data': 0.03176356665790081, 'timing/model': 1.485232688486576, 'learning_rate/qwen_vl_interface': 1.5458134701022706e-06, 'learning_rate/action_model': 1.5458134701022705e-05, 'epoch': 1.53}) +[2026-06-16 06:34:01,235][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.013580012135207653, 'train/grad_norm_pre_clip': 4.150051116943359, 'timing/data': 0.03321433998644352, 'timing/model': 1.5174667686223984, 'learning_rate/qwen_vl_interface': 1.5439078193033765e-06, 'learning_rate/action_model': 1.5439078193033766e-05, 'epoch': 1.53}) +[2026-06-16 06:34:07,337][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.028686268255114555, 'train/grad_norm_pre_clip': 4.400692939758301, 'timing/data': 0.030749760568141937, 'timing/model': 1.5332911144942045, 'learning_rate/qwen_vl_interface': 1.5420036919389872e-06, 'learning_rate/action_model': 1.5420036919389873e-05, 'epoch': 1.53}) +[2026-06-16 06:34:13,436][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.044756755232810974, 'train/grad_norm_pre_clip': 4.5729804039001465, 'timing/data': 0.03151938132941723, 'timing/model': 1.4920580815523863, 'learning_rate/qwen_vl_interface': 1.540101088791817e-06, 'learning_rate/action_model': 1.5401010887918168e-05, 'epoch': 1.53}) +[2026-06-16 06:34:20,041][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.07949788868427277, 'train/grad_norm_pre_clip': 4.940633773803711, 'timing/data': 0.03163010627031326, 'timing/model': 1.835652519017458, 'learning_rate/qwen_vl_interface': 1.5382000106439532e-06, 'learning_rate/action_model': 1.5382000106439532e-05, 'epoch': 1.53}) +[2026-06-16 06:34:25,931][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.02722393348813057, 'train/grad_norm_pre_clip': 3.1086325645446777, 'timing/data': 0.0319259837269783, 'timing/model': 1.524375680834055, 'learning_rate/qwen_vl_interface': 1.5363004582768584e-06, 'learning_rate/action_model': 1.5363004582768584e-05, 'epoch': 1.53}) +[2026-06-16 06:34:32,037][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.031714826822280884, 'train/grad_norm_pre_clip': 3.548382043838501, 'timing/data': 0.03199849650263786, 'timing/model': 1.5176597572863102, 'learning_rate/qwen_vl_interface': 1.534402432471367e-06, 'learning_rate/action_model': 1.534402432471367e-05, 'epoch': 1.53}) +[2026-06-16 06:34:38,038][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.23450259864330292, 'train/grad_norm_pre_clip': 14.953978538513184, 'timing/data': 0.031008195132017136, 'timing/model': 1.5191658921539783, 'learning_rate/qwen_vl_interface': 1.532505934007685e-06, 'learning_rate/action_model': 1.532505934007685e-05, 'epoch': 1.53}) +[2026-06-16 06:34:44,033][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.01975858584046364, 'train/grad_norm_pre_clip': 6.447850704193115, 'timing/data': 0.034726813435554504, 'timing/model': 1.5246789678931236, 'learning_rate/qwen_vl_interface': 1.5306109636653915e-06, 'learning_rate/action_model': 1.5306109636653914e-05, 'epoch': 1.53}) +[2026-06-16 06:34:50,133][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.04852214828133583, 'train/grad_norm_pre_clip': 31.144634246826172, 'timing/data': 0.0324892308562994, 'timing/model': 1.5509406495839357, 'learning_rate/qwen_vl_interface': 1.5287175222234383e-06, 'learning_rate/action_model': 1.528717522223438e-05, 'epoch': 1.53}) +[2026-06-16 06:34:56,243][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.12194932997226715, 'train/grad_norm_pre_clip': 14.6647367477417, 'timing/data': 0.031650012359023094, 'timing/model': 1.740733776241541, 'learning_rate/qwen_vl_interface': 1.5268256104601457e-06, 'learning_rate/action_model': 1.5268256104601454e-05, 'epoch': 1.53}) +[2026-06-16 06:35:02,137][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.007099370937794447, 'train/grad_norm_pre_clip': 2.37134051322937, 'timing/data': 0.032478610053658485, 'timing/model': 1.53431286662817, 'learning_rate/qwen_vl_interface': 1.5249352291532075e-06, 'learning_rate/action_model': 1.5249352291532073e-05, 'epoch': 1.53}) +[2026-06-16 06:35:08,235][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.07009333372116089, 'train/grad_norm_pre_clip': 5.702725887298584, 'timing/data': 0.03175182081758976, 'timing/model': 1.4888433944433928, 'learning_rate/qwen_vl_interface': 1.523046379079689e-06, 'learning_rate/action_model': 1.523046379079689e-05, 'epoch': 1.53}) +[2026-06-16 06:35:14,461][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.005849670618772507, 'train/grad_norm_pre_clip': 2.7324655055999756, 'timing/data': 0.031554073095321655, 'timing/model': 1.5274464581161737, 'learning_rate/qwen_vl_interface': 1.5211590610160244e-06, 'learning_rate/action_model': 1.5211590610160242e-05, 'epoch': 1.53}) +[2026-06-16 06:35:20,666][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.07320854812860489, 'train/grad_norm_pre_clip': 6.942813396453857, 'timing/data': 0.03230748511850834, 'timing/model': 1.5309020075947046, 'learning_rate/qwen_vl_interface': 1.5192732757380193e-06, 'learning_rate/action_model': 1.5192732757380194e-05, 'epoch': 1.53}) +[2026-06-16 06:35:26,966][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.10866527259349823, 'train/grad_norm_pre_clip': 11.99217700958252, 'timing/data': 0.03272315673530102, 'timing/model': 1.5567489862442017, 'learning_rate/qwen_vl_interface': 1.517389024020849e-06, 'learning_rate/action_model': 1.5173890240208488e-05, 'epoch': 1.53}) +[2026-06-16 06:35:33,191][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.05182545259594917, 'train/grad_norm_pre_clip': 19.316104888916016, 'timing/data': 0.03153618983924389, 'timing/model': 1.7567426357418299, 'learning_rate/qwen_vl_interface': 1.515506306639057e-06, 'learning_rate/action_model': 1.515506306639057e-05, 'epoch': 1.53}) +[2026-06-16 06:35:39,271][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.06048155575990677, 'train/grad_norm_pre_clip': 11.31633472442627, 'timing/data': 0.03239882551133633, 'timing/model': 1.540783142670989, 'learning_rate/qwen_vl_interface': 1.5136251243665589e-06, 'learning_rate/action_model': 1.5136251243665589e-05, 'epoch': 1.53}) +[2026-06-16 06:35:45,578][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.003876077476888895, 'train/grad_norm_pre_clip': 1.0617001056671143, 'timing/data': 0.031108984723687172, 'timing/model': 1.5492256712168455, 'learning_rate/qwen_vl_interface': 1.5117454779766372e-06, 'learning_rate/action_model': 1.5117454779766371e-05, 'epoch': 1.53}) +[2026-06-16 06:35:51,859][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.13685660064220428, 'train/grad_norm_pre_clip': 14.916244506835938, 'timing/data': 0.03202989138662815, 'timing/model': 1.5048127248883247, 'learning_rate/qwen_vl_interface': 1.5098673682419428e-06, 'learning_rate/action_model': 1.5098673682419429e-05, 'epoch': 1.53}) +[2026-06-16 06:35:57,936][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.13088633120059967, 'train/grad_norm_pre_clip': 13.14627456665039, 'timing/data': 0.032593827694654465, 'timing/model': 1.4732643961906433, 'learning_rate/qwen_vl_interface': 1.5079907959344977e-06, 'learning_rate/action_model': 1.5079907959344978e-05, 'epoch': 1.53}) +[2026-06-16 06:36:04,031][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.018310291692614555, 'train/grad_norm_pre_clip': 2.6966638565063477, 'timing/data': 0.03202947787940502, 'timing/model': 1.525165667757392, 'learning_rate/qwen_vl_interface': 1.5061157618256873e-06, 'learning_rate/action_model': 1.5061157618256871e-05, 'epoch': 1.53}) +[2026-06-16 06:36:10,289][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.016935067251324654, 'train/grad_norm_pre_clip': 3.3609566688537598, 'timing/data': 0.030927177518606186, 'timing/model': 1.757386703044176, 'learning_rate/qwen_vl_interface': 1.5042422666862675e-06, 'learning_rate/action_model': 1.5042422666862673e-05, 'epoch': 1.54}) +[2026-06-16 06:36:16,468][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.17079517245292664, 'train/grad_norm_pre_clip': 20.846891403198242, 'timing/data': 0.030703239142894745, 'timing/model': 1.6218529269099236, 'learning_rate/qwen_vl_interface': 1.5023703112863628e-06, 'learning_rate/action_model': 1.5023703112863626e-05, 'epoch': 1.54}) +[2026-06-16 06:36:22,760][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.007250107824802399, 'train/grad_norm_pre_clip': 2.7676517963409424, 'timing/data': 0.031859464943408966, 'timing/model': 1.5087937228381634, 'learning_rate/qwen_vl_interface': 1.5004998963954629e-06, 'learning_rate/action_model': 1.5004998963954628e-05, 'epoch': 1.54}) +[2026-06-16 06:36:29,026][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.024750839918851852, 'train/grad_norm_pre_clip': 7.960302829742432, 'timing/data': 0.033490993082523346, 'timing/model': 1.5960807800292969, 'learning_rate/qwen_vl_interface': 1.498631022782424e-06, 'learning_rate/action_model': 1.4986310227824241e-05, 'epoch': 1.54}) +[2026-06-16 06:36:35,259][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.008180057629942894, 'train/grad_norm_pre_clip': 1.448598861694336, 'timing/data': 0.03187456168234348, 'timing/model': 1.5263494960963726, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 1.54}) +[2026-06-16 06:36:41,371][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.018744807690382004, 'train/grad_norm_pre_clip': 3.8995769023895264, 'timing/data': 0.02528887242078781, 'timing/model': 1.4472345598042011, 'learning_rate/qwen_vl_interface': 1.4948979024621915e-06, 'learning_rate/action_model': 1.4948979024621914e-05, 'epoch': 1.54}) +[2026-06-16 06:36:47,385][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.09619235247373581, 'train/grad_norm_pre_clip': 14.677549362182617, 'timing/data': 0.028101623058319092, 'timing/model': 1.6965146847069263, 'learning_rate/qwen_vl_interface': 1.493033657289542e-06, 'learning_rate/action_model': 1.493033657289542e-05, 'epoch': 1.54}) +[2026-06-16 06:36:53,127][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.1760483831167221, 'train/grad_norm_pre_clip': 13.929221153259277, 'timing/data': 0.026721419766545296, 'timing/model': 1.4491396844387054, 'learning_rate/qwen_vl_interface': 1.4911709564638428e-06, 'learning_rate/action_model': 1.4911709564638427e-05, 'epoch': 1.54}) +[2026-06-16 06:36:59,014][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.010376201011240482, 'train/grad_norm_pre_clip': 3.624593496322632, 'timing/data': 0.026689402759075165, 'timing/model': 1.4389089904725552, 'learning_rate/qwen_vl_interface': 1.4893098007507817e-06, 'learning_rate/action_model': 1.4893098007507816e-05, 'epoch': 1.54}) +[2026-06-16 06:37:04,908][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.021274972707033157, 'train/grad_norm_pre_clip': 13.546292304992676, 'timing/data': 0.02562260441482067, 'timing/model': 1.4400397147983313, 'learning_rate/qwen_vl_interface': 1.487450190915406e-06, 'learning_rate/action_model': 1.487450190915406e-05, 'epoch': 1.54}) +[2026-06-16 06:37:10,716][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.020922640338540077, 'train/grad_norm_pre_clip': 3.2811362743377686, 'timing/data': 0.02682291716337204, 'timing/model': 1.4417376276105642, 'learning_rate/qwen_vl_interface': 1.4855921277221336e-06, 'learning_rate/action_model': 1.4855921277221336e-05, 'epoch': 1.54}) +[2026-06-16 06:37:16,605][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.1386014074087143, 'train/grad_norm_pre_clip': 10.567502975463867, 'timing/data': 0.02523070015013218, 'timing/model': 1.4424586798995733, 'learning_rate/qwen_vl_interface': 1.4837356119347423e-06, 'learning_rate/action_model': 1.4837356119347423e-05, 'epoch': 1.54}) +[2026-06-16 06:37:22,461][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.06042155250906944, 'train/grad_norm_pre_clip': 26.992109298706055, 'timing/data': 0.025747276842594147, 'timing/model': 1.5866412241011858, 'learning_rate/qwen_vl_interface': 1.4818806443163774e-06, 'learning_rate/action_model': 1.4818806443163775e-05, 'epoch': 1.54}) +[2026-06-16 06:37:28,177][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.002304143738001585, 'train/grad_norm_pre_clip': 0.9482210874557495, 'timing/data': 0.02529500424861908, 'timing/model': 1.4397692643105984, 'learning_rate/qwen_vl_interface': 1.480027225629547e-06, 'learning_rate/action_model': 1.480027225629547e-05, 'epoch': 1.54}) +[2026-06-16 06:37:34,075][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.07297325879335403, 'train/grad_norm_pre_clip': 7.145954132080078, 'timing/data': 0.02647285908460617, 'timing/model': 1.433765534311533, 'learning_rate/qwen_vl_interface': 1.4781753566361182e-06, 'learning_rate/action_model': 1.4781753566361182e-05, 'epoch': 1.54}) +[2026-06-16 06:37:39,879][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.018572969362139702, 'train/grad_norm_pre_clip': 4.585050106048584, 'timing/data': 0.025039752945303917, 'timing/model': 1.4405673556029797, 'learning_rate/qwen_vl_interface': 1.4763250380973255e-06, 'learning_rate/action_model': 1.4763250380973254e-05, 'epoch': 1.54}) +[2026-06-16 06:37:45,677][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.01434209942817688, 'train/grad_norm_pre_clip': 5.8108110427856445, 'timing/data': 0.02712850831449032, 'timing/model': 1.4344819579273462, 'learning_rate/qwen_vl_interface': 1.4744762707737662e-06, 'learning_rate/action_model': 1.4744762707737662e-05, 'epoch': 1.54}) +[2026-06-16 06:37:51,684][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.15550220012664795, 'train/grad_norm_pre_clip': 10.131355285644531, 'timing/data': 0.026023289188742638, 'timing/model': 1.4180038776248693, 'learning_rate/qwen_vl_interface': 1.4726290554253977e-06, 'learning_rate/action_model': 1.4726290554253977e-05, 'epoch': 1.54}) +[2026-06-16 06:37:57,648][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.03607943281531334, 'train/grad_norm_pre_clip': 4.810746669769287, 'timing/data': 0.025663452222943306, 'timing/model': 1.6066414564847946, 'learning_rate/qwen_vl_interface': 1.4707833928115406e-06, 'learning_rate/action_model': 1.4707833928115405e-05, 'epoch': 1.54}) +[2026-06-16 06:38:03,275][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.002371616894379258, 'train/grad_norm_pre_clip': 0.5008905529975891, 'timing/data': 0.026197079569101334, 'timing/model': 1.4365643803030252, 'learning_rate/qwen_vl_interface': 1.4689392836908775e-06, 'learning_rate/action_model': 1.4689392836908774e-05, 'epoch': 1.54}) +[2026-06-16 06:38:09,077][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.06535030901432037, 'train/grad_norm_pre_clip': 10.294593811035156, 'timing/data': 0.02723412960767746, 'timing/model': 1.4353029802441597, 'learning_rate/qwen_vl_interface': 1.4670967288214514e-06, 'learning_rate/action_model': 1.4670967288214513e-05, 'epoch': 1.54}) +[2026-06-16 06:38:14,875][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.16778762638568878, 'train/grad_norm_pre_clip': 19.974740982055664, 'timing/data': 0.026497842743992805, 'timing/model': 1.4328933339565992, 'learning_rate/qwen_vl_interface': 1.4652557289606687e-06, 'learning_rate/action_model': 1.4652557289606685e-05, 'epoch': 1.54}) +[2026-06-16 06:38:20,838][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.025077270343899727, 'train/grad_norm_pre_clip': 6.031102180480957, 'timing/data': 0.03244563192129135, 'timing/model': 1.478055253624916, 'learning_rate/qwen_vl_interface': 1.4634162848652918e-06, 'learning_rate/action_model': 1.4634162848652919e-05, 'epoch': 1.54}) +[2026-06-16 06:38:26,933][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.07924488186836243, 'train/grad_norm_pre_clip': 34.98427963256836, 'timing/data': 0.03114723041653633, 'timing/model': 1.4943020287901163, 'learning_rate/qwen_vl_interface': 1.4615783972914494e-06, 'learning_rate/action_model': 1.4615783972914493e-05, 'epoch': 1.54}) +[2026-06-16 06:38:33,093][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.08344011008739471, 'train/grad_norm_pre_clip': 8.573108673095703, 'timing/data': 0.03226601891219616, 'timing/model': 1.7911019828170538, 'learning_rate/qwen_vl_interface': 1.459742066994626e-06, 'learning_rate/action_model': 1.4597420669946261e-05, 'epoch': 1.54}) +[2026-06-16 06:38:39,302][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.08880399912595749, 'train/grad_norm_pre_clip': 5.354371547698975, 'timing/data': 0.032159579917788506, 'timing/model': 1.6692524775862694, 'learning_rate/qwen_vl_interface': 1.4579072947296687e-06, 'learning_rate/action_model': 1.4579072947296687e-05, 'epoch': 1.54}) +[2026-06-16 06:38:45,359][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.05590822175145149, 'train/grad_norm_pre_clip': 11.37126350402832, 'timing/data': 0.03095318004488945, 'timing/model': 1.5318356696516275, 'learning_rate/qwen_vl_interface': 1.4560740812507823e-06, 'learning_rate/action_model': 1.4560740812507822e-05, 'epoch': 1.55}) +[2026-06-16 06:38:51,561][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.004948381334543228, 'train/grad_norm_pre_clip': 1.104252576828003, 'timing/data': 0.03250870853662491, 'timing/model': 1.517531393095851, 'learning_rate/qwen_vl_interface': 1.4542424273115322e-06, 'learning_rate/action_model': 1.454242427311532e-05, 'epoch': 1.55}) +[2026-06-16 06:38:57,775][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.021301910281181335, 'train/grad_norm_pre_clip': 12.841059684753418, 'timing/data': 0.030763640999794006, 'timing/model': 1.5174945089966059, 'learning_rate/qwen_vl_interface': 1.4524123336648419e-06, 'learning_rate/action_model': 1.4524123336648418e-05, 'epoch': 1.55}) +[2026-06-16 06:39:04,158][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.18379732966423035, 'train/grad_norm_pre_clip': 32.52254867553711, 'timing/data': 0.03221288323402405, 'timing/model': 1.574073864147067, 'learning_rate/qwen_vl_interface': 1.4505838010629942e-06, 'learning_rate/action_model': 1.4505838010629941e-05, 'epoch': 1.55}) +[2026-06-16 06:39:10,364][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.06395532190799713, 'train/grad_norm_pre_clip': 18.09713363647461, 'timing/data': 0.032970983535051346, 'timing/model': 1.7109518721699715, 'learning_rate/qwen_vl_interface': 1.4487568302576284e-06, 'learning_rate/action_model': 1.4487568302576282e-05, 'epoch': 1.55}) +[2026-06-16 06:39:16,495][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.002771573606878519, 'train/grad_norm_pre_clip': 0.46203169226646423, 'timing/data': 0.0318458117544651, 'timing/model': 1.6780674792826176, 'learning_rate/qwen_vl_interface': 1.446931421999746e-06, 'learning_rate/action_model': 1.4469314219997459e-05, 'epoch': 1.55}) +[2026-06-16 06:39:22,441][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.07145702093839645, 'train/grad_norm_pre_clip': 5.0310540199279785, 'timing/data': 0.032874543219804764, 'timing/model': 1.5052817407995462, 'learning_rate/qwen_vl_interface': 1.4451075770397004e-06, 'learning_rate/action_model': 1.4451075770397004e-05, 'epoch': 1.55}) +[2026-06-16 06:46:35,675][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.19232696294784546, 'train/grad_norm_pre_clip': 17.045879364013672, 'eval/loss': 0.046897973865270615, 'eval/flap_precision': 0.9627586206896551, 'eval/flap_recall': 0.9470827679782904, 'eval/flap_f1': 0.9548563611491109, 'eval/flap_rate_model': 0.2013888888888889, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9548563611491108, 'timing/data': 0.031131820753216743, 'timing/model': 1.509055895730853, 'learning_rate/qwen_vl_interface': 1.4432852961272068e-06, 'learning_rate/action_model': 1.4432852961272066e-05, 'epoch': 1.55}) +[2026-06-16 06:46:35,676][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4000_state +[2026-06-16 06:46:35,680][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 06:46:41,464][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4000_state/model.safetensors +[2026-06-16 06:46:50,771][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4000_state/optimizer.bin +[2026-06-16 06:46:50,772][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4000_state/sampler.bin +[2026-06-16 06:46:50,772][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-16 06:46:50,773][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-16 06:46:50,874][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-16 06:46:50,874][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 06:46:50,883][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 06:46:56,789][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.30314937233924866, 'train/grad_norm_pre_clip': 25.560108184814453, 'timing/data': 0.025300180539488792, 'timing/model': 1.4362199809402227, 'learning_rate/qwen_vl_interface': 1.4414645800113353e-06, 'learning_rate/action_model': 1.4414645800113352e-05, 'epoch': 1.55}) +[2026-06-16 06:47:02,697][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.009314735420048237, 'train/grad_norm_pre_clip': 0.799500584602356, 'timing/data': 0.02760658785700798, 'timing/model': 1.4393446296453476, 'learning_rate/qwen_vl_interface': 1.4396454294405138e-06, 'learning_rate/action_model': 1.4396454294405136e-05, 'epoch': 1.55}) +[2026-06-16 06:47:08,603][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 0.019409533590078354, 'train/grad_norm_pre_clip': 6.134542942047119, 'timing/data': 0.02591523714363575, 'timing/model': 1.4306934494525194, 'learning_rate/qwen_vl_interface': 1.4378278451625271e-06, 'learning_rate/action_model': 1.4378278451625271e-05, 'epoch': 1.55}) +[2026-06-16 06:47:14,659][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.0806470438838005, 'train/grad_norm_pre_clip': 23.221939086914062, 'timing/data': 0.03297489508986473, 'timing/model': 1.4946301728487015, 'learning_rate/qwen_vl_interface': 1.4360118279245143e-06, 'learning_rate/action_model': 1.4360118279245144e-05, 'epoch': 1.55}) +[2026-06-16 06:47:20,605][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.09972096234560013, 'train/grad_norm_pre_clip': 14.496617317199707, 'timing/data': 0.03206232003867626, 'timing/model': 1.6314846985042095, 'learning_rate/qwen_vl_interface': 1.434197378472973e-06, 'learning_rate/action_model': 1.4341973784729729e-05, 'epoch': 1.55}) +[2026-06-16 06:47:26,423][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.02756015583872795, 'train/grad_norm_pre_clip': 12.064332008361816, 'timing/data': 0.03279891610145569, 'timing/model': 1.5217973552644253, 'learning_rate/qwen_vl_interface': 1.4323844975537535e-06, 'learning_rate/action_model': 1.4323844975537534e-05, 'epoch': 1.55}) +[2026-06-16 06:47:32,553][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.17949599027633667, 'train/grad_norm_pre_clip': 54.63922882080078, 'timing/data': 0.03222692012786865, 'timing/model': 1.5579888112843037, 'learning_rate/qwen_vl_interface': 1.4305731859120628e-06, 'learning_rate/action_model': 1.4305731859120628e-05, 'epoch': 1.55}) +[2026-06-16 06:47:38,373][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.09874788671731949, 'train/grad_norm_pre_clip': 7.179281234741211, 'timing/data': 0.02655252069234848, 'timing/model': 1.4461042489856482, 'learning_rate/qwen_vl_interface': 1.4287634442924647e-06, 'learning_rate/action_model': 1.4287634442924646e-05, 'epoch': 1.55}) +[2026-06-16 06:47:44,172][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.08033356070518494, 'train/grad_norm_pre_clip': 11.671467781066895, 'timing/data': 0.026596250012516975, 'timing/model': 1.4538721647113562, 'learning_rate/qwen_vl_interface': 1.4269552734388725e-06, 'learning_rate/action_model': 1.4269552734388724e-05, 'epoch': 1.55}) +[2026-06-16 06:47:50,052][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.021543331444263458, 'train/grad_norm_pre_clip': 9.984899520874023, 'timing/data': 0.02582496404647827, 'timing/model': 1.6187751200050116, 'learning_rate/qwen_vl_interface': 1.4251486740945584e-06, 'learning_rate/action_model': 1.4251486740945583e-05, 'epoch': 1.55}) +[2026-06-16 06:47:55,671][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.1397077590227127, 'train/grad_norm_pre_clip': 9.197826385498047, 'timing/data': 0.029306042939424515, 'timing/model': 1.4321332331746817, 'learning_rate/qwen_vl_interface': 1.423343647002147e-06, 'learning_rate/action_model': 1.423343647002147e-05, 'epoch': 1.55}) +[2026-06-16 06:48:01,577][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.11519087851047516, 'train/grad_norm_pre_clip': 15.651007652282715, 'timing/data': 0.026912664994597435, 'timing/model': 1.4379575438797474, 'learning_rate/qwen_vl_interface': 1.421540192903617e-06, 'learning_rate/action_model': 1.421540192903617e-05, 'epoch': 1.55}) +[2026-06-16 06:48:07,378][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 0.028859244659543037, 'train/grad_norm_pre_clip': 14.089475631713867, 'timing/data': 0.026000576093792915, 'timing/model': 1.44310980476439, 'learning_rate/qwen_vl_interface': 1.4197383125402997e-06, 'learning_rate/action_model': 1.4197383125402997e-05, 'epoch': 1.55}) +[2026-06-16 06:48:13,176][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.0015232032164931297, 'train/grad_norm_pre_clip': 0.27658140659332275, 'timing/data': 0.028453409671783447, 'timing/model': 1.4355279672890902, 'learning_rate/qwen_vl_interface': 1.4179380066528807e-06, 'learning_rate/action_model': 1.4179380066528807e-05, 'epoch': 1.55}) +[2026-06-16 06:48:18,977][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.0034368708729743958, 'train/grad_norm_pre_clip': 0.5607478618621826, 'timing/data': 0.027038592845201492, 'timing/model': 1.4075929149985313, 'learning_rate/qwen_vl_interface': 1.4161392759813984e-06, 'learning_rate/action_model': 1.4161392759813983e-05, 'epoch': 1.55}) +[2026-06-16 06:48:24,857][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.006277075968682766, 'train/grad_norm_pre_clip': 0.8087501525878906, 'timing/data': 0.02627173438668251, 'timing/model': 1.6178543120622635, 'learning_rate/qwen_vl_interface': 1.4143421212652411e-06, 'learning_rate/action_model': 1.414342121265241e-05, 'epoch': 1.55}) +[2026-06-16 06:48:30,730][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.02434031292796135, 'train/grad_norm_pre_clip': 7.70379114151001, 'timing/data': 0.03168310597538948, 'timing/model': 1.515310188755393, 'learning_rate/qwen_vl_interface': 1.4125465432431529e-06, 'learning_rate/action_model': 1.4125465432431529e-05, 'epoch': 1.55}) +[2026-06-16 06:48:36,674][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.1553221493959427, 'train/grad_norm_pre_clip': 12.95566463470459, 'timing/data': 0.026426855474710464, 'timing/model': 1.436408057808876, 'learning_rate/qwen_vl_interface': 1.4107525426532273e-06, 'learning_rate/action_model': 1.4107525426532272e-05, 'epoch': 1.56}) +[2026-06-16 06:48:42,476][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.008171227760612965, 'train/grad_norm_pre_clip': 2.651942729949951, 'timing/data': 0.026885665953159332, 'timing/model': 1.4379800837486982, 'learning_rate/qwen_vl_interface': 1.4089601202329116e-06, 'learning_rate/action_model': 1.4089601202329114e-05, 'epoch': 1.56}) +[2026-06-16 06:48:48,375][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.09174580872058868, 'train/grad_norm_pre_clip': 11.776223182678223, 'timing/data': 0.026002593338489532, 'timing/model': 1.4345742166042328, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 1.56}) +[2026-06-16 06:48:54,190][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.23129859566688538, 'train/grad_norm_pre_clip': 13.002643585205078, 'timing/data': 0.027244841679930687, 'timing/model': 1.4115054868161678, 'learning_rate/qwen_vl_interface': 1.4053800128476426e-06, 'learning_rate/action_model': 1.4053800128476427e-05, 'epoch': 1.56}) +[2026-06-16 06:49:00,058][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.18622660636901855, 'train/grad_norm_pre_clip': 21.347639083862305, 'timing/data': 0.026218058541417122, 'timing/model': 1.6169179920107126, 'learning_rate/qwen_vl_interface': 1.403592329354338e-06, 'learning_rate/action_model': 1.403592329354338e-05, 'epoch': 1.56}) +[2026-06-16 06:49:05,675][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.0660654604434967, 'train/grad_norm_pre_clip': 10.622855186462402, 'timing/data': 0.027179209515452385, 'timing/model': 1.4323476497083902, 'learning_rate/qwen_vl_interface': 1.4018062269739354e-06, 'learning_rate/action_model': 1.4018062269739355e-05, 'epoch': 1.56}) +[2026-06-16 06:49:11,578][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.2068021297454834, 'train/grad_norm_pre_clip': 10.669271469116211, 'timing/data': 0.026165304705500603, 'timing/model': 1.435654267668724, 'learning_rate/qwen_vl_interface': 1.4000217064406339e-06, 'learning_rate/action_model': 1.4000217064406337e-05, 'epoch': 1.56}) +[2026-06-16 06:49:17,481][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.1216278001666069, 'train/grad_norm_pre_clip': 22.631711959838867, 'timing/data': 0.026742272078990936, 'timing/model': 1.4400871843099594, 'learning_rate/qwen_vl_interface': 1.3982387684879836e-06, 'learning_rate/action_model': 1.3982387684879836e-05, 'epoch': 1.56}) +[2026-06-16 06:49:23,377][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.09715298563241959, 'train/grad_norm_pre_clip': 12.434998512268066, 'timing/data': 0.027287624776363373, 'timing/model': 1.441144784912467, 'learning_rate/qwen_vl_interface': 1.3964574138488804e-06, 'learning_rate/action_model': 1.3964574138488803e-05, 'epoch': 1.56}) +[2026-06-16 06:49:29,279][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.0925784632563591, 'train/grad_norm_pre_clip': 10.755060195922852, 'timing/data': 0.02697908692061901, 'timing/model': 1.4711180441081524, 'learning_rate/qwen_vl_interface': 1.3946776432555743e-06, 'learning_rate/action_model': 1.3946776432555741e-05, 'epoch': 1.56}) +[2026-06-16 06:49:35,171][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.03972203657031059, 'train/grad_norm_pre_clip': 22.523895263671875, 'timing/data': 0.027276504784822464, 'timing/model': 1.626332936808467, 'learning_rate/qwen_vl_interface': 1.3928994574396595e-06, 'learning_rate/action_model': 1.3928994574396594e-05, 'epoch': 1.56}) +[2026-06-16 06:49:40,824][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.3265596628189087, 'train/grad_norm_pre_clip': 18.352890014648438, 'timing/data': 0.02582080289721489, 'timing/model': 1.4499115999788046, 'learning_rate/qwen_vl_interface': 1.391122857132083e-06, 'learning_rate/action_model': 1.3911228571320831e-05, 'epoch': 1.56}) +[2026-06-16 06:49:46,724][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.008180512115359306, 'train/grad_norm_pre_clip': 1.6314914226531982, 'timing/data': 0.02617059089243412, 'timing/model': 1.4492733050137758, 'learning_rate/qwen_vl_interface': 1.3893478430631376e-06, 'learning_rate/action_model': 1.3893478430631376e-05, 'epoch': 1.56}) +[2026-06-16 06:49:52,617][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.03548682481050491, 'train/grad_norm_pre_clip': 11.554205894470215, 'timing/data': 0.011854501441121101, 'timing/model': 1.458441387861967, 'learning_rate/qwen_vl_interface': 1.3875744159624652e-06, 'learning_rate/action_model': 1.3875744159624652e-05, 'epoch': 1.56}) +[2026-06-16 06:49:58,507][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.06501247733831406, 'train/grad_norm_pre_clip': 3.4158084392547607, 'timing/data': 0.010221520438790321, 'timing/model': 1.4481239672750235, 'learning_rate/qwen_vl_interface': 1.3858025765590525e-06, 'learning_rate/action_model': 1.3858025765590525e-05, 'epoch': 1.56}) +[2026-06-16 06:50:04,511][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.1325872242450714, 'train/grad_norm_pre_clip': 23.96068572998047, 'timing/data': 0.009573962539434433, 'timing/model': 1.4582345727831125, 'learning_rate/qwen_vl_interface': 1.384032325581237e-06, 'learning_rate/action_model': 1.3840323255812371e-05, 'epoch': 1.56}) +[2026-06-16 06:50:10,602][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.009066413156688213, 'train/grad_norm_pre_clip': 2.0363636016845703, 'timing/data': 0.02184215374290943, 'timing/model': 1.7184697482734919, 'learning_rate/qwen_vl_interface': 1.3822636637567025e-06, 'learning_rate/action_model': 1.3822636637567024e-05, 'epoch': 1.56}) +[2026-06-16 06:50:16,523][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.03053484484553337, 'train/grad_norm_pre_clip': 12.748820304870605, 'timing/data': 0.02377990446984768, 'timing/model': 1.5302539821714163, 'learning_rate/qwen_vl_interface': 1.3804965918124788e-06, 'learning_rate/action_model': 1.3804965918124787e-05, 'epoch': 1.56}) +[2026-06-16 06:50:22,609][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.18101991713047028, 'train/grad_norm_pre_clip': 9.30808162689209, 'timing/data': 0.01926606148481369, 'timing/model': 1.483558475971222, 'learning_rate/qwen_vl_interface': 1.3787311104749418e-06, 'learning_rate/action_model': 1.3787311104749418e-05, 'epoch': 1.56}) +[2026-06-16 06:50:28,475][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.01512875221669674, 'train/grad_norm_pre_clip': 1.739418864250183, 'timing/data': 0.018196173012256622, 'timing/model': 1.4300149828195572, 'learning_rate/qwen_vl_interface': 1.3769672204698149e-06, 'learning_rate/action_model': 1.3769672204698147e-05, 'epoch': 1.56}) +[2026-06-16 06:50:34,278][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.01445956900715828, 'train/grad_norm_pre_clip': 4.487961769104004, 'timing/data': 0.01983119174838066, 'timing/model': 1.4411898627877235, 'learning_rate/qwen_vl_interface': 1.3752049225221664e-06, 'learning_rate/action_model': 1.3752049225221663e-05, 'epoch': 1.56}) +[2026-06-16 06:50:40,177][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.09139035642147064, 'train/grad_norm_pre_clip': 12.883596420288086, 'timing/data': 0.018820855766534805, 'timing/model': 1.4559294134378433, 'learning_rate/qwen_vl_interface': 1.3734442173564103e-06, 'learning_rate/action_model': 1.3734442173564102e-05, 'epoch': 1.56}) +[2026-06-16 06:50:46,051][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.16687919199466705, 'train/grad_norm_pre_clip': 19.031707763671875, 'timing/data': 0.019335288554430008, 'timing/model': 1.6142197828739882, 'learning_rate/qwen_vl_interface': 1.3716851056963062e-06, 'learning_rate/action_model': 1.3716851056963062e-05, 'epoch': 1.56}) +[2026-06-16 06:50:51,763][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.09242761880159378, 'train/grad_norm_pre_clip': 35.56162643432617, 'timing/data': 0.018642988055944443, 'timing/model': 1.5286455228924751, 'learning_rate/qwen_vl_interface': 1.3699275882649597e-06, 'learning_rate/action_model': 1.3699275882649596e-05, 'epoch': 1.56}) +[2026-06-16 06:50:57,936][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.08179443329572678, 'train/grad_norm_pre_clip': 5.989335536956787, 'timing/data': 0.02332867868244648, 'timing/model': 1.510509829968214, 'learning_rate/qwen_vl_interface': 1.368171665784817e-06, 'learning_rate/action_model': 1.3681716657848171e-05, 'epoch': 1.56}) +[2026-06-16 06:51:04,030][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.01846907287836075, 'train/grad_norm_pre_clip': 5.869511604309082, 'timing/data': 0.022407924756407738, 'timing/model': 1.5201187878847122, 'learning_rate/qwen_vl_interface': 1.3664173389776735e-06, 'learning_rate/action_model': 1.3664173389776735e-05, 'epoch': 1.56}) +[2026-06-16 06:51:10,134][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.26903751492500305, 'train/grad_norm_pre_clip': 12.915419578552246, 'timing/data': 0.022356107831001282, 'timing/model': 1.5388468261808157, 'learning_rate/qwen_vl_interface': 1.3646646085646658e-06, 'learning_rate/action_model': 1.3646646085646659e-05, 'epoch': 1.57}) +[2026-06-16 06:51:16,234][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.09067761898040771, 'train/grad_norm_pre_clip': 6.157706260681152, 'timing/data': 0.022786905989050865, 'timing/model': 1.5169098991900682, 'learning_rate/qwen_vl_interface': 1.3629134752662756e-06, 'learning_rate/action_model': 1.3629134752662756e-05, 'epoch': 1.57}) +[2026-06-16 06:51:22,337][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 0.07611234486103058, 'train/grad_norm_pre_clip': 29.04706382751465, 'timing/data': 0.022328611463308334, 'timing/model': 1.7458176966756582, 'learning_rate/qwen_vl_interface': 1.3611639398023282e-06, 'learning_rate/action_model': 1.3611639398023283e-05, 'epoch': 1.57}) +[2026-06-16 06:51:28,402][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.08582604676485062, 'train/grad_norm_pre_clip': 14.600696563720703, 'timing/data': 0.021577028557658195, 'timing/model': 1.709334710612893, 'learning_rate/qwen_vl_interface': 1.35941600289199e-06, 'learning_rate/action_model': 1.35941600289199e-05, 'epoch': 1.57}) +[2026-06-16 06:51:34,332][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.023512890562415123, 'train/grad_norm_pre_clip': 3.457291603088379, 'timing/data': 0.021126773208379745, 'timing/model': 1.540387123823166, 'learning_rate/qwen_vl_interface': 1.3576696652537735e-06, 'learning_rate/action_model': 1.3576696652537736e-05, 'epoch': 1.57}) +[2026-06-16 06:51:40,532][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.10667543113231659, 'train/grad_norm_pre_clip': 25.315092086791992, 'timing/data': 0.0211422648280859, 'timing/model': 1.5420928448438644, 'learning_rate/qwen_vl_interface': 1.355924927605531e-06, 'learning_rate/action_model': 1.3559249276055311e-05, 'epoch': 1.57}) +[2026-06-16 06:51:46,723][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.17100787162780762, 'train/grad_norm_pre_clip': 15.110657691955566, 'timing/data': 0.02196844108402729, 'timing/model': 1.5420735329389572, 'learning_rate/qwen_vl_interface': 1.3541817906644591e-06, 'learning_rate/action_model': 1.3541817906644591e-05, 'epoch': 1.57}) +[2026-06-16 06:51:52,835][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.04611924663186073, 'train/grad_norm_pre_clip': 10.077361106872559, 'timing/data': 0.024309491738677025, 'timing/model': 1.5128412880003452, 'learning_rate/qwen_vl_interface': 1.3524402551470955e-06, 'learning_rate/action_model': 1.3524402551470956e-05, 'epoch': 1.57}) +[2026-06-16 06:51:59,055][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.16917411983013153, 'train/grad_norm_pre_clip': 12.919361114501953, 'timing/data': 0.021036673337221146, 'timing/model': 1.8609752729535103, 'learning_rate/qwen_vl_interface': 1.3507003217693196e-06, 'learning_rate/action_model': 1.3507003217693196e-05, 'epoch': 1.57}) +[2026-06-16 06:52:05,135][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.007206672802567482, 'train/grad_norm_pre_clip': 1.4232597351074219, 'timing/data': 0.021631503477692604, 'timing/model': 1.5323904398828745, 'learning_rate/qwen_vl_interface': 1.3489619912463515e-06, 'learning_rate/action_model': 1.3489619912463514e-05, 'epoch': 1.57}) +[2026-06-16 06:52:11,341][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.08024894446134567, 'train/grad_norm_pre_clip': 6.522856712341309, 'timing/data': 0.021621162071824074, 'timing/model': 1.5452254824340343, 'learning_rate/qwen_vl_interface': 1.3472252642927534e-06, 'learning_rate/action_model': 1.3472252642927533e-05, 'epoch': 1.57}) +[2026-06-16 06:52:17,515][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 0.1964515745639801, 'train/grad_norm_pre_clip': 15.084945678710938, 'timing/data': 0.021294258534908295, 'timing/model': 1.5445985458791256, 'learning_rate/qwen_vl_interface': 1.3454901416224284e-06, 'learning_rate/action_model': 1.3454901416224284e-05, 'epoch': 1.57}) +[2026-06-16 06:52:23,633][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 0.010685061104595661, 'train/grad_norm_pre_clip': 1.0857391357421875, 'timing/data': 0.02209142968058586, 'timing/model': 1.5156462844461203, 'learning_rate/qwen_vl_interface': 1.34375662394862e-06, 'learning_rate/action_model': 1.34375662394862e-05, 'epoch': 1.57}) +[2026-06-16 06:52:29,636][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.0910269096493721, 'train/grad_norm_pre_clip': 6.242886066436768, 'timing/data': 0.02185368351638317, 'timing/model': 1.5069936644285917, 'learning_rate/qwen_vl_interface': 1.3420247119839116e-06, 'learning_rate/action_model': 1.3420247119839117e-05, 'epoch': 1.57}) +[2026-06-16 06:52:35,784][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.0743643268942833, 'train/grad_norm_pre_clip': 10.255644798278809, 'timing/data': 0.022767500951886177, 'timing/model': 1.6988945920020342, 'learning_rate/qwen_vl_interface': 1.340294406440227e-06, 'learning_rate/action_model': 1.340294406440227e-05, 'epoch': 1.57}) +[2026-06-16 06:52:41,834][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.00785452127456665, 'train/grad_norm_pre_clip': 2.65207839012146, 'timing/data': 0.021176915615797043, 'timing/model': 1.5454683899879456, 'learning_rate/qwen_vl_interface': 1.3385657080288297e-06, 'learning_rate/action_model': 1.3385657080288297e-05, 'epoch': 1.57}) +[2026-06-16 06:52:48,042][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.11271990835666656, 'train/grad_norm_pre_clip': 11.390020370483398, 'timing/data': 0.022206295281648636, 'timing/model': 1.5445575378835201, 'learning_rate/qwen_vl_interface': 1.336838617460322e-06, 'learning_rate/action_model': 1.336838617460322e-05, 'epoch': 1.57}) +[2026-06-16 06:52:54,236][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.01717865653336048, 'train/grad_norm_pre_clip': 3.6719729900360107, 'timing/data': 0.024189753457903862, 'timing/model': 1.5398190300911665, 'learning_rate/qwen_vl_interface': 1.335113135444646e-06, 'learning_rate/action_model': 1.335113135444646e-05, 'epoch': 1.57}) +[2026-06-16 06:53:00,433][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 0.02287030778825283, 'train/grad_norm_pre_clip': 9.334863662719727, 'timing/data': 0.022232644259929657, 'timing/model': 1.5087867751717567, 'learning_rate/qwen_vl_interface': 1.3333892626910823e-06, 'learning_rate/action_model': 1.3333892626910823e-05, 'epoch': 1.57}) +[2026-06-16 06:53:06,627][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.007844814099371433, 'train/grad_norm_pre_clip': 0.71791672706604, 'timing/data': 0.022528808563947678, 'timing/model': 1.5492563247680664, 'learning_rate/qwen_vl_interface': 1.3316669999082486e-06, 'learning_rate/action_model': 1.3316669999082484e-05, 'epoch': 1.57}) +[2026-06-16 06:53:12,907][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.018241912126541138, 'train/grad_norm_pre_clip': 2.9731028079986572, 'timing/data': 0.022413387894630432, 'timing/model': 1.704239645972848, 'learning_rate/qwen_vl_interface': 1.3299463478041024e-06, 'learning_rate/action_model': 1.3299463478041024e-05, 'epoch': 1.57}) +[2026-06-16 06:53:19,136][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.03637577220797539, 'train/grad_norm_pre_clip': 7.078047752380371, 'timing/data': 0.02376825176179409, 'timing/model': 1.520880438387394, 'learning_rate/qwen_vl_interface': 1.3282273070859396e-06, 'learning_rate/action_model': 1.3282273070859396e-05, 'epoch': 1.57}) +[2026-06-16 06:53:25,377][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 0.18872633576393127, 'train/grad_norm_pre_clip': 11.653959274291992, 'timing/data': 0.025508932769298553, 'timing/model': 1.5548203978687525, 'learning_rate/qwen_vl_interface': 1.3265098784603924e-06, 'learning_rate/action_model': 1.3265098784603924e-05, 'epoch': 1.57}) +[2026-06-16 06:53:31,851][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.034020621329545975, 'train/grad_norm_pre_clip': 17.55926513671875, 'timing/data': 0.025340156629681587, 'timing/model': 1.6285120099782944, 'learning_rate/qwen_vl_interface': 1.324794062633431e-06, 'learning_rate/action_model': 1.324794062633431e-05, 'epoch': 1.57}) +[2026-06-16 06:53:38,140][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.009984144940972328, 'train/grad_norm_pre_clip': 4.256527423858643, 'timing/data': 0.027437608689069748, 'timing/model': 1.5285439547151327, 'learning_rate/qwen_vl_interface': 1.323079860310363e-06, 'learning_rate/action_model': 1.3230798603103628e-05, 'epoch': 1.57}) +[2026-06-16 06:53:44,343][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.042679790407419205, 'train/grad_norm_pre_clip': 7.956838607788086, 'timing/data': 0.02322564646601677, 'timing/model': 1.7455552257597446, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 1.57}) +[2026-06-16 06:53:50,441][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.11192778497934341, 'train/grad_norm_pre_clip': 6.914502143859863, 'timing/data': 0.024613680317997932, 'timing/model': 1.5556329879909754, 'learning_rate/qwen_vl_interface': 1.3196562989938175e-06, 'learning_rate/action_model': 1.3196562989938174e-05, 'epoch': 1.58}) +[2026-06-16 06:53:56,632][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.25698715448379517, 'train/grad_norm_pre_clip': 21.026601791381836, 'timing/data': 0.022609157487750053, 'timing/model': 1.5335411727428436, 'learning_rate/qwen_vl_interface': 1.3179469414076357e-06, 'learning_rate/action_model': 1.3179469414076356e-05, 'epoch': 1.58}) +[2026-06-16 06:54:02,746][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.04344351589679718, 'train/grad_norm_pre_clip': 13.586257934570312, 'timing/data': 0.02369343675673008, 'timing/model': 1.5471345353871584, 'learning_rate/qwen_vl_interface': 1.3162392001399401e-06, 'learning_rate/action_model': 1.3162392001399401e-05, 'epoch': 1.58}) +[2026-06-16 06:54:08,916][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.005994302686303854, 'train/grad_norm_pre_clip': 1.0205072164535522, 'timing/data': 0.021914348006248474, 'timing/model': 1.5501676853746176, 'learning_rate/qwen_vl_interface': 1.3145330758927182e-06, 'learning_rate/action_model': 1.3145330758927182e-05, 'epoch': 1.58}) +[2026-06-16 06:54:14,935][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.006324253045022488, 'train/grad_norm_pre_clip': 0.9401861429214478, 'timing/data': 0.02117174118757248, 'timing/model': 1.475809721276164, 'learning_rate/qwen_vl_interface': 1.3128285693672935e-06, 'learning_rate/action_model': 1.3128285693672934e-05, 'epoch': 1.58}) +[2026-06-16 06:54:21,076][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.05469527840614319, 'train/grad_norm_pre_clip': 20.99491310119629, 'timing/data': 0.021814187988638878, 'timing/model': 1.7872438058257103, 'learning_rate/qwen_vl_interface': 1.311125681264325e-06, 'learning_rate/action_model': 1.311125681264325e-05, 'epoch': 1.58}) +[2026-06-16 06:54:27,058][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.00519840931519866, 'train/grad_norm_pre_clip': 1.792457938194275, 'timing/data': 0.022805601358413696, 'timing/model': 1.5605698134750128, 'learning_rate/qwen_vl_interface': 1.309424412283803e-06, 'learning_rate/action_model': 1.3094244122838029e-05, 'epoch': 1.58}) +[2026-06-16 06:54:33,338][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 0.08403105288743973, 'train/grad_norm_pre_clip': 14.096186637878418, 'timing/data': 0.022801490500569344, 'timing/model': 1.5561726205050945, 'learning_rate/qwen_vl_interface': 1.307724763125057e-06, 'learning_rate/action_model': 1.307724763125057e-05, 'epoch': 1.58}) +[2026-06-16 06:54:39,436][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.024959580972790718, 'train/grad_norm_pre_clip': 3.283538579940796, 'timing/data': 0.021318016573786736, 'timing/model': 1.5371546391397715, 'learning_rate/qwen_vl_interface': 1.3060267344867478e-06, 'learning_rate/action_model': 1.3060267344867478e-05, 'epoch': 1.58}) +[2026-06-16 06:54:45,647][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.002392273861914873, 'train/grad_norm_pre_clip': 0.32050904631614685, 'timing/data': 0.024616628885269165, 'timing/model': 1.532351028174162, 'learning_rate/qwen_vl_interface': 1.3043303270668715e-06, 'learning_rate/action_model': 1.3043303270668713e-05, 'epoch': 1.58}) +[2026-06-16 06:54:51,861][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.019265858456492424, 'train/grad_norm_pre_clip': 6.5758538246154785, 'timing/data': 0.025417320430278778, 'timing/model': 1.5270355362445116, 'learning_rate/qwen_vl_interface': 1.3026355415627565e-06, 'learning_rate/action_model': 1.3026355415627564e-05, 'epoch': 1.58}) +[2026-06-16 06:54:58,144][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.0695313960313797, 'train/grad_norm_pre_clip': 8.862645149230957, 'timing/data': 0.024333056062459946, 'timing/model': 1.8363461066037416, 'learning_rate/qwen_vl_interface': 1.3009423786710653e-06, 'learning_rate/action_model': 1.3009423786710653e-05, 'epoch': 1.58}) +[2026-06-16 06:55:04,037][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.11722329258918762, 'train/grad_norm_pre_clip': 20.549320220947266, 'timing/data': 0.023751206696033478, 'timing/model': 1.5426441077142954, 'learning_rate/qwen_vl_interface': 1.2992508390877933e-06, 'learning_rate/action_model': 1.2992508390877933e-05, 'epoch': 1.58}) +[2026-06-16 06:55:10,236][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.17075635492801666, 'train/grad_norm_pre_clip': 8.638833999633789, 'timing/data': 0.02315746806561947, 'timing/model': 1.5446473229676485, 'learning_rate/qwen_vl_interface': 1.297560923508269e-06, 'learning_rate/action_model': 1.297560923508269e-05, 'epoch': 1.58}) +[2026-06-16 06:55:16,432][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.05464952066540718, 'train/grad_norm_pre_clip': 9.174800872802734, 'timing/data': 0.022013675421476364, 'timing/model': 1.537874672561884, 'learning_rate/qwen_vl_interface': 1.295872632627152e-06, 'learning_rate/action_model': 1.2958726326271519e-05, 'epoch': 1.58}) +[2026-06-16 06:55:22,532][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 0.26421183347702026, 'train/grad_norm_pre_clip': 37.06964874267578, 'timing/data': 0.022698944434523582, 'timing/model': 1.5384606923907995, 'learning_rate/qwen_vl_interface': 1.2941859671384364e-06, 'learning_rate/action_model': 1.2941859671384362e-05, 'epoch': 1.58}) +[2026-06-16 06:55:28,732][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.0189689751714468, 'train/grad_norm_pre_clip': 7.7120513916015625, 'timing/data': 0.022384434938430786, 'timing/model': 1.5326306000351906, 'learning_rate/qwen_vl_interface': 1.2925009277354444e-06, 'learning_rate/action_model': 1.2925009277354444e-05, 'epoch': 1.58}) +[2026-06-16 06:55:34,944][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.15856146812438965, 'train/grad_norm_pre_clip': 13.827054977416992, 'timing/data': 0.02304108440876007, 'timing/model': 1.7469792552292347, 'learning_rate/qwen_vl_interface': 1.2908175151108335e-06, 'learning_rate/action_model': 1.2908175151108334e-05, 'epoch': 1.58}) +[2026-06-16 06:55:40,837][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.21051324903964996, 'train/grad_norm_pre_clip': 13.785000801086426, 'timing/data': 0.022317029535770416, 'timing/model': 1.5429731849581003, 'learning_rate/qwen_vl_interface': 1.2891357299565899e-06, 'learning_rate/action_model': 1.2891357299565898e-05, 'epoch': 1.58}) +[2026-06-16 06:55:47,044][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.1682320386171341, 'train/grad_norm_pre_clip': 14.02140998840332, 'timing/data': 0.023927852511405945, 'timing/model': 1.5339573603123426, 'learning_rate/qwen_vl_interface': 1.2874555729640333e-06, 'learning_rate/action_model': 1.2874555729640334e-05, 'epoch': 1.58}) +[2026-06-16 06:55:53,121][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.11305461823940277, 'train/grad_norm_pre_clip': 17.93488883972168, 'timing/data': 0.02059541456401348, 'timing/model': 1.554925600066781, 'learning_rate/qwen_vl_interface': 1.2857770448238124e-06, 'learning_rate/action_model': 1.2857770448238123e-05, 'epoch': 1.58}) +[2026-06-16 06:55:59,016][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.20945581793785095, 'train/grad_norm_pre_clip': 15.216880798339844, 'timing/data': 0.018738901242613792, 'timing/model': 1.4512381814420223, 'learning_rate/qwen_vl_interface': 1.2841001462259067e-06, 'learning_rate/action_model': 1.2841001462259067e-05, 'epoch': 1.58}) +[2026-06-16 06:56:05,009][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.24350711703300476, 'train/grad_norm_pre_clip': 27.732688903808594, 'timing/data': 0.018935609608888626, 'timing/model': 1.4576271064579487, 'learning_rate/qwen_vl_interface': 1.282424877859626e-06, 'learning_rate/action_model': 1.282424877859626e-05, 'epoch': 1.58}) +[2026-06-16 06:56:10,900][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.0002101573918480426, 'train/grad_norm_pre_clip': 0.049718957394361496, 'timing/data': 0.018109366297721863, 'timing/model': 1.6225659158080816, 'learning_rate/qwen_vl_interface': 1.2807512404136088e-06, 'learning_rate/action_model': 1.2807512404136087e-05, 'epoch': 1.58}) +[2026-06-16 06:56:16,611][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.10556885600090027, 'train/grad_norm_pre_clip': 9.888439178466797, 'timing/data': 0.019019804894924164, 'timing/model': 1.4440543204545975, 'learning_rate/qwen_vl_interface': 1.2790792345758255e-06, 'learning_rate/action_model': 1.2790792345758254e-05, 'epoch': 1.58}) +[2026-06-16 06:56:22,515][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.09922928363084793, 'train/grad_norm_pre_clip': 10.505207061767578, 'timing/data': 0.01781458780169487, 'timing/model': 1.4446386359632015, 'learning_rate/qwen_vl_interface': 1.2774088610335752e-06, 'learning_rate/action_model': 1.2774088610335752e-05, 'epoch': 1.58}) +[2026-06-16 06:56:28,510][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 0.022716570645570755, 'train/grad_norm_pre_clip': 5.547913551330566, 'timing/data': 0.01865444891154766, 'timing/model': 1.4458384290337563, 'learning_rate/qwen_vl_interface': 1.2757401204734827e-06, 'learning_rate/action_model': 1.2757401204734826e-05, 'epoch': 1.59}) +[2026-06-16 06:56:34,409][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.18375657498836517, 'train/grad_norm_pre_clip': 90.29678344726562, 'timing/data': 0.018802929669618607, 'timing/model': 1.4420709926635027, 'learning_rate/qwen_vl_interface': 1.2740730135815064e-06, 'learning_rate/action_model': 1.2740730135815062e-05, 'epoch': 1.59}) +[2026-06-16 06:56:40,307][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 0.10065820068120956, 'train/grad_norm_pre_clip': 9.372617721557617, 'timing/data': 0.017621347680687904, 'timing/model': 1.459528984501958, 'learning_rate/qwen_vl_interface': 1.2724075410429298e-06, 'learning_rate/action_model': 1.2724075410429299e-05, 'epoch': 1.59}) +[2026-06-16 06:56:46,189][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.014624162577092648, 'train/grad_norm_pre_clip': 4.8164777755737305, 'timing/data': 0.017895136028528214, 'timing/model': 1.6195014081895351, 'learning_rate/qwen_vl_interface': 1.2707437035423664e-06, 'learning_rate/action_model': 1.2707437035423664e-05, 'epoch': 1.59}) +[2026-06-16 07:04:07,625][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.014792745001614094, 'train/grad_norm_pre_clip': 2.271841287612915, 'eval/loss': 0.04235098883509636, 'eval/flap_precision': 0.9620081411126187, 'eval/flap_recall': 0.9620081411126187, 'eval/flap_f1': 0.9620081411126187, 'eval/flap_rate_model': 0.20472222222222222, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9620081411126186, 'timing/data': 0.018031328916549683, 'timing/model': 1.44540592469275, 'learning_rate/qwen_vl_interface': 1.2690815017637573e-06, 'learning_rate/action_model': 1.2690815017637572e-05, 'epoch': 1.59}) +[2026-06-16 07:04:07,626][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4100_state +[2026-06-16 07:04:07,640][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 07:04:13,560][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4100_state/model.safetensors +[2026-06-16 07:04:28,029][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4100_state/optimizer.bin +[2026-06-16 07:04:28,029][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4100_state/sampler.bin +[2026-06-16 07:04:28,030][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4100_state/sampler_1.bin +[2026-06-16 07:04:28,030][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4100_state/random_states_0.pkl +[2026-06-16 07:04:28,136][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4100 +[2026-06-16 07:04:28,136][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 07:04:28,144][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 07:04:33,986][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.007344446610659361, 'train/grad_norm_pre_clip': 0.53858882188797, 'timing/data': 0.018709812313318253, 'timing/model': 1.6314545776695013, 'learning_rate/qwen_vl_interface': 1.2674209363903694e-06, 'learning_rate/action_model': 1.2674209363903693e-05, 'epoch': 1.59}) +[2026-06-16 07:04:39,679][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.0024808647576719522, 'train/grad_norm_pre_clip': 0.5180067420005798, 'timing/data': 0.01857518032193184, 'timing/model': 1.4388168659061193, 'learning_rate/qwen_vl_interface': 1.2657620081048005e-06, 'learning_rate/action_model': 1.2657620081048006e-05, 'epoch': 1.59}) +[2026-06-16 07:04:45,636][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.0939357802271843, 'train/grad_norm_pre_clip': 15.418636322021484, 'timing/data': 0.017858322709798813, 'timing/model': 1.4757346417754889, 'learning_rate/qwen_vl_interface': 1.2641047175889714e-06, 'learning_rate/action_model': 1.2641047175889714e-05, 'epoch': 1.59}) +[2026-06-16 07:04:51,489][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.16753311455249786, 'train/grad_norm_pre_clip': 16.788230895996094, 'timing/data': 0.019624531269073486, 'timing/model': 1.434744605794549, 'learning_rate/qwen_vl_interface': 1.2624490655241332e-06, 'learning_rate/action_model': 1.262449065524133e-05, 'epoch': 1.59}) +[2026-06-16 07:04:57,360][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.1588594764471054, 'train/grad_norm_pre_clip': 20.82722282409668, 'timing/data': 0.02085128054022789, 'timing/model': 1.438819209113717, 'learning_rate/qwen_vl_interface': 1.2607950525908602e-06, 'learning_rate/action_model': 1.2607950525908601e-05, 'epoch': 1.59}) +[2026-06-16 07:05:03,167][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.11316913366317749, 'train/grad_norm_pre_clip': 15.924195289611816, 'timing/data': 0.027889473363757133, 'timing/model': 1.4449964351952076, 'learning_rate/qwen_vl_interface': 1.2591426794690575e-06, 'learning_rate/action_model': 1.2591426794690576e-05, 'epoch': 1.59}) +[2026-06-16 07:05:09,192][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.02835768833756447, 'train/grad_norm_pre_clip': 13.331351280212402, 'timing/data': 0.022667624056339264, 'timing/model': 1.6351393423974514, 'learning_rate/qwen_vl_interface': 1.25749194683795e-06, 'learning_rate/action_model': 1.25749194683795e-05, 'epoch': 1.59}) +[2026-06-16 07:05:14,864][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.09752130508422852, 'train/grad_norm_pre_clip': 17.10163116455078, 'timing/data': 0.01847163215279579, 'timing/model': 1.4308689329773188, 'learning_rate/qwen_vl_interface': 1.2558428553760933e-06, 'learning_rate/action_model': 1.2558428553760932e-05, 'epoch': 1.59}) +[2026-06-16 07:05:20,692][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.0984799861907959, 'train/grad_norm_pre_clip': 7.614214897155762, 'timing/data': 0.01924275979399681, 'timing/model': 1.4343696404248476, 'learning_rate/qwen_vl_interface': 1.2541954057613659e-06, 'learning_rate/action_model': 1.2541954057613658e-05, 'epoch': 1.59}) +[2026-06-16 07:05:26,591][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.03947825729846954, 'train/grad_norm_pre_clip': 8.692025184631348, 'timing/data': 0.017730172723531723, 'timing/model': 1.428978530690074, 'learning_rate/qwen_vl_interface': 1.2525495986709727e-06, 'learning_rate/action_model': 1.2525495986709726e-05, 'epoch': 1.59}) +[2026-06-16 07:05:32,495][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.00765134347602725, 'train/grad_norm_pre_clip': 1.1219695806503296, 'timing/data': 0.01847856305539608, 'timing/model': 1.4437120910733938, 'learning_rate/qwen_vl_interface': 1.2509054347814423e-06, 'learning_rate/action_model': 1.2509054347814423e-05, 'epoch': 1.59}) +[2026-06-16 07:05:38,300][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.08368902653455734, 'train/grad_norm_pre_clip': 20.964223861694336, 'timing/data': 0.018052777275443077, 'timing/model': 1.4319730307906866, 'learning_rate/qwen_vl_interface': 1.2492629147686288e-06, 'learning_rate/action_model': 1.2492629147686289e-05, 'epoch': 1.59}) +[2026-06-16 07:05:44,159][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.11977046728134155, 'train/grad_norm_pre_clip': 25.930601119995117, 'timing/data': 0.019126923754811287, 'timing/model': 1.6096553169190884, 'learning_rate/qwen_vl_interface': 1.2476220393077098e-06, 'learning_rate/action_model': 1.2476220393077098e-05, 'epoch': 1.59}) +[2026-06-16 07:05:49,833][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.003150931093841791, 'train/grad_norm_pre_clip': 0.4091375470161438, 'timing/data': 0.022301796823740005, 'timing/model': 1.454984748736024, 'learning_rate/qwen_vl_interface': 1.2459828090731877e-06, 'learning_rate/action_model': 1.2459828090731876e-05, 'epoch': 1.59}) +[2026-06-16 07:05:55,899][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.002770835068076849, 'train/grad_norm_pre_clip': 0.5909168124198914, 'timing/data': 0.017719972878694534, 'timing/model': 1.4345272034406662, 'learning_rate/qwen_vl_interface': 1.2443452247388876e-06, 'learning_rate/action_model': 1.2443452247388877e-05, 'epoch': 1.59}) +[2026-06-16 07:06:01,791][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.06858360022306442, 'train/grad_norm_pre_clip': 10.441091537475586, 'timing/data': 0.016973936930298805, 'timing/model': 1.4381404854357243, 'learning_rate/qwen_vl_interface': 1.2427092869779603e-06, 'learning_rate/action_model': 1.2427092869779602e-05, 'epoch': 1.59}) +[2026-06-16 07:06:07,665][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.22518537938594818, 'train/grad_norm_pre_clip': 28.25465965270996, 'timing/data': 0.017739394679665565, 'timing/model': 1.4514556396752596, 'learning_rate/qwen_vl_interface': 1.2410749964628751e-06, 'learning_rate/action_model': 1.2410749964628751e-05, 'epoch': 1.59}) +[2026-06-16 07:06:13,375][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.06990446150302887, 'train/grad_norm_pre_clip': 4.118963718414307, 'timing/data': 0.018872829154133797, 'timing/model': 1.4053804036229849, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 1.59}) +[2026-06-16 07:06:19,158][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.22566765546798706, 'train/grad_norm_pre_clip': 18.328655242919922, 'timing/data': 0.0171649307012558, 'timing/model': 1.6356559805572033, 'learning_rate/qwen_vl_interface': 1.2378113598567394e-06, 'learning_rate/action_model': 1.2378113598567392e-05, 'epoch': 1.59}) +[2026-06-16 07:06:25,028][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.06551940739154816, 'train/grad_norm_pre_clip': 6.860774993896484, 'timing/data': 0.022549506276845932, 'timing/model': 1.5392857752740383, 'learning_rate/qwen_vl_interface': 1.2361820151072454e-06, 'learning_rate/action_model': 1.2361820151072453e-05, 'epoch': 1.59}) +[2026-06-16 07:06:31,226][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.015779005363583565, 'train/grad_norm_pre_clip': 5.970219135284424, 'timing/data': 0.022536879405379295, 'timing/model': 1.5404751915484667, 'learning_rate/qwen_vl_interface': 1.23455432028671e-06, 'learning_rate/action_model': 1.23455432028671e-05, 'epoch': 1.59}) +[2026-06-16 07:06:37,322][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.0014593747910112143, 'train/grad_norm_pre_clip': 0.16893887519836426, 'timing/data': 0.022845765575766563, 'timing/model': 1.5364607572555542, 'learning_rate/qwen_vl_interface': 1.2329282760642167e-06, 'learning_rate/action_model': 1.2329282760642166e-05, 'epoch': 1.6}) +[2026-06-16 07:06:43,725][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.06327342987060547, 'train/grad_norm_pre_clip': 23.096033096313477, 'timing/data': 0.023053016513586044, 'timing/model': 1.5485513918101788, 'learning_rate/qwen_vl_interface': 1.2313038831081716e-06, 'learning_rate/action_model': 1.2313038831081717e-05, 'epoch': 1.6}) +[2026-06-16 07:06:49,572][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.15178464353084564, 'train/grad_norm_pre_clip': 13.906559944152832, 'timing/data': 0.01889755204319954, 'timing/model': 1.4834751356393099, 'learning_rate/qwen_vl_interface': 1.2296811420863002e-06, 'learning_rate/action_model': 1.2296811420863003e-05, 'epoch': 1.6}) +[2026-06-16 07:06:55,498][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.007601647637784481, 'train/grad_norm_pre_clip': 6.0816569328308105, 'timing/data': 0.018316160887479782, 'timing/model': 1.6316798254847527, 'learning_rate/qwen_vl_interface': 1.2280600536656504e-06, 'learning_rate/action_model': 1.2280600536656503e-05, 'epoch': 1.6}) +[2026-06-16 07:07:01,197][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.056738290935754776, 'train/grad_norm_pre_clip': 9.95048999786377, 'timing/data': 0.018509944900870323, 'timing/model': 1.4385067280381918, 'learning_rate/qwen_vl_interface': 1.2264406185125907e-06, 'learning_rate/action_model': 1.2264406185125907e-05, 'epoch': 1.6}) +[2026-06-16 07:07:07,099][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.17703141272068024, 'train/grad_norm_pre_clip': 10.996018409729004, 'timing/data': 0.018670525401830673, 'timing/model': 1.4421146679669619, 'learning_rate/qwen_vl_interface': 1.2248228372928098e-06, 'learning_rate/action_model': 1.2248228372928097e-05, 'epoch': 1.6}) +[2026-06-16 07:07:13,112][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.00806551706045866, 'train/grad_norm_pre_clip': 0.6625372171401978, 'timing/data': 0.018054787069559097, 'timing/model': 1.4374430887401104, 'learning_rate/qwen_vl_interface': 1.2232067106713174e-06, 'learning_rate/action_model': 1.2232067106713174e-05, 'epoch': 1.6}) +[2026-06-16 07:07:18,872][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.1077345535159111, 'train/grad_norm_pre_clip': 13.26406192779541, 'timing/data': 0.018276046961545944, 'timing/model': 1.427780320867896, 'learning_rate/qwen_vl_interface': 1.22159223931244e-06, 'learning_rate/action_model': 1.2215922393124399e-05, 'epoch': 1.6}) +[2026-06-16 07:07:24,674][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.015552363358438015, 'train/grad_norm_pre_clip': 4.038922309875488, 'timing/data': 0.017704825848340988, 'timing/model': 1.4591991528868675, 'learning_rate/qwen_vl_interface': 1.2199794238798266e-06, 'learning_rate/action_model': 1.2199794238798264e-05, 'epoch': 1.6}) +[2026-06-16 07:07:30,443][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.11436262726783752, 'train/grad_norm_pre_clip': 14.505261421203613, 'timing/data': 0.019405798986554146, 'timing/model': 1.6097622625529766, 'learning_rate/qwen_vl_interface': 1.2183682650364447e-06, 'learning_rate/action_model': 1.2183682650364447e-05, 'epoch': 1.6}) +[2026-06-16 07:07:36,069][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.019966337829828262, 'train/grad_norm_pre_clip': 4.4865217208862305, 'timing/data': 0.018654564395546913, 'timing/model': 1.445596855133772, 'learning_rate/qwen_vl_interface': 1.216758763444581e-06, 'learning_rate/action_model': 1.2167587634445809e-05, 'epoch': 1.6}) +[2026-06-16 07:07:41,898][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.03845963627099991, 'train/grad_norm_pre_clip': 7.168929100036621, 'timing/data': 0.017702609300613403, 'timing/model': 1.4446030091494322, 'learning_rate/qwen_vl_interface': 1.215150919765841e-06, 'learning_rate/action_model': 1.215150919765841e-05, 'epoch': 1.6}) +[2026-06-16 07:07:47,960][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.0031549655832350254, 'train/grad_norm_pre_clip': 1.6444129943847656, 'timing/data': 0.022027183324098587, 'timing/model': 1.557913612574339, 'learning_rate/qwen_vl_interface': 1.213544734661148e-06, 'learning_rate/action_model': 1.2135447346611478e-05, 'epoch': 1.6}) +[2026-06-16 07:07:54,148][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.03348546102643013, 'train/grad_norm_pre_clip': 11.999892234802246, 'timing/data': 0.029626166447997093, 'timing/model': 1.507067235186696, 'learning_rate/qwen_vl_interface': 1.2119402087907448e-06, 'learning_rate/action_model': 1.2119402087907448e-05, 'epoch': 1.6}) +[2026-06-16 07:08:00,455][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.15980912744998932, 'train/grad_norm_pre_clip': 7.659643650054932, 'timing/data': 0.023281993344426155, 'timing/model': 1.5728164333850145, 'learning_rate/qwen_vl_interface': 1.2103373428141903e-06, 'learning_rate/action_model': 1.2103373428141903e-05, 'epoch': 1.6}) +[2026-06-16 07:08:06,680][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.027226818725466728, 'train/grad_norm_pre_clip': 8.695965766906738, 'timing/data': 0.02280091866850853, 'timing/model': 1.7558966055512428, 'learning_rate/qwen_vl_interface': 1.2087361373903633e-06, 'learning_rate/action_model': 1.2087361373903633e-05, 'epoch': 1.6}) +[2026-06-16 07:08:12,720][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.013070084154605865, 'train/grad_norm_pre_clip': 1.6420543193817139, 'timing/data': 0.022305920720100403, 'timing/model': 1.5274722278118134, 'learning_rate/qwen_vl_interface': 1.2071365931774584e-06, 'learning_rate/action_model': 1.2071365931774584e-05, 'epoch': 1.6}) +[2026-06-16 07:08:18,958][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.060176171362400055, 'train/grad_norm_pre_clip': 9.85158634185791, 'timing/data': 0.02082650549709797, 'timing/model': 1.5445889998227358, 'learning_rate/qwen_vl_interface': 1.2055387108329885e-06, 'learning_rate/action_model': 1.2055387108329886e-05, 'epoch': 1.6}) +[2026-06-16 07:08:25,160][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.06336276978254318, 'train/grad_norm_pre_clip': 6.216650485992432, 'timing/data': 0.022116128355264664, 'timing/model': 1.5278781466186047, 'learning_rate/qwen_vl_interface': 1.2039424910137807e-06, 'learning_rate/action_model': 1.2039424910137806e-05, 'epoch': 1.6}) +[2026-06-16 07:08:31,354][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.012521013617515564, 'train/grad_norm_pre_clip': 1.4193987846374512, 'timing/data': 0.022467153146862984, 'timing/model': 1.5378284491598606, 'learning_rate/qwen_vl_interface': 1.2023479343759822e-06, 'learning_rate/action_model': 1.2023479343759823e-05, 'epoch': 1.6}) +[2026-06-16 07:08:37,559][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.011998788453638554, 'train/grad_norm_pre_clip': 5.3812456130981445, 'timing/data': 0.021292535588145256, 'timing/model': 1.5229878909885883, 'learning_rate/qwen_vl_interface': 1.2007550415750546e-06, 'learning_rate/action_model': 1.2007550415750545e-05, 'epoch': 1.6}) +[2026-06-16 07:08:43,767][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.381612092256546, 'train/grad_norm_pre_clip': 16.496997833251953, 'timing/data': 0.02175760827958584, 'timing/model': 1.7512783166021109, 'learning_rate/qwen_vl_interface': 1.1991638132657759e-06, 'learning_rate/action_model': 1.199163813265776e-05, 'epoch': 1.6}) +[2026-06-16 07:08:49,759][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.010653693228960037, 'train/grad_norm_pre_clip': 2.6105058193206787, 'timing/data': 0.02208208665251732, 'timing/model': 1.5285837184637785, 'learning_rate/qwen_vl_interface': 1.1975742501022406e-06, 'learning_rate/action_model': 1.1975742501022405e-05, 'epoch': 1.6}) +[2026-06-16 07:08:55,958][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.13547201454639435, 'train/grad_norm_pre_clip': 13.390493392944336, 'timing/data': 0.02216765470802784, 'timing/model': 1.5397213287651539, 'learning_rate/qwen_vl_interface': 1.1959863527378569e-06, 'learning_rate/action_model': 1.1959863527378567e-05, 'epoch': 1.6}) +[2026-06-16 07:09:02,159][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.10071226209402084, 'train/grad_norm_pre_clip': 20.43842315673828, 'timing/data': 0.02140350267291069, 'timing/model': 1.5454815477132797, 'learning_rate/qwen_vl_interface': 1.1944001218253505e-06, 'learning_rate/action_model': 1.1944001218253503e-05, 'epoch': 1.6}) +[2026-06-16 07:09:08,763][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.04674006626009941, 'train/grad_norm_pre_clip': 4.989546298980713, 'timing/data': 0.022902894765138626, 'timing/model': 1.5471870061010122, 'learning_rate/qwen_vl_interface': 1.1928155580167608e-06, 'learning_rate/action_model': 1.1928155580167608e-05, 'epoch': 1.6}) +[2026-06-16 07:09:15,055][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.047364264726638794, 'train/grad_norm_pre_clip': 22.984310150146484, 'timing/data': 0.023328861221671104, 'timing/model': 1.5180083457380533, 'learning_rate/qwen_vl_interface': 1.1912326619634412e-06, 'learning_rate/action_model': 1.1912326619634412e-05, 'epoch': 1.61}) +[2026-06-16 07:09:21,266][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.3012244701385498, 'train/grad_norm_pre_clip': 13.608339309692383, 'timing/data': 0.022238463163375854, 'timing/model': 1.748991210013628, 'learning_rate/qwen_vl_interface': 1.1896514343160632e-06, 'learning_rate/action_model': 1.189651434316063e-05, 'epoch': 1.61}) +[2026-06-16 07:09:27,370][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.0855882540345192, 'train/grad_norm_pre_clip': 8.55164623260498, 'timing/data': 0.023204483091831207, 'timing/model': 1.6763308122754097, 'learning_rate/qwen_vl_interface': 1.1880718757246062e-06, 'learning_rate/action_model': 1.1880718757246063e-05, 'epoch': 1.61}) +[2026-06-16 07:09:33,328][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.03553350269794464, 'train/grad_norm_pre_clip': 8.270998001098633, 'timing/data': 0.022149283438920975, 'timing/model': 1.541221434250474, 'learning_rate/qwen_vl_interface': 1.186493986838369e-06, 'learning_rate/action_model': 1.186493986838369e-05, 'epoch': 1.61}) +[2026-06-16 07:09:39,549][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.013476659543812275, 'train/grad_norm_pre_clip': 6.24539041519165, 'timing/data': 0.02263735421001911, 'timing/model': 1.474556153640151, 'learning_rate/qwen_vl_interface': 1.1849177683059612e-06, 'learning_rate/action_model': 1.1849177683059612e-05, 'epoch': 1.61}) +[2026-06-16 07:09:45,631][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.21721386909484863, 'train/grad_norm_pre_clip': 33.76442337036133, 'timing/data': 0.022321103140711784, 'timing/model': 1.5458198431879282, 'learning_rate/qwen_vl_interface': 1.1833432207753075e-06, 'learning_rate/action_model': 1.1833432207753076e-05, 'epoch': 1.61}) +[2026-06-16 07:09:51,726][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.12694121897220612, 'train/grad_norm_pre_clip': 14.244426727294922, 'timing/data': 0.021564532071352005, 'timing/model': 1.5218707136809826, 'learning_rate/qwen_vl_interface': 1.1817703448936448e-06, 'learning_rate/action_model': 1.181770344893645e-05, 'epoch': 1.61}) +[2026-06-16 07:09:57,661][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.27356284856796265, 'train/grad_norm_pre_clip': 16.103153228759766, 'timing/data': 0.01741613820195198, 'timing/model': 1.6299085896462202, 'learning_rate/qwen_vl_interface': 1.1801991413075228e-06, 'learning_rate/action_model': 1.1801991413075227e-05, 'epoch': 1.61}) +[2026-06-16 07:10:03,466][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.1015404686331749, 'train/grad_norm_pre_clip': 17.487018585205078, 'timing/data': 0.018283547833561897, 'timing/model': 1.6336129121482372, 'learning_rate/qwen_vl_interface': 1.1786296106628038e-06, 'learning_rate/action_model': 1.1786296106628036e-05, 'epoch': 1.61}) +[2026-06-16 07:10:09,069][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.09844652563333511, 'train/grad_norm_pre_clip': 9.545854568481445, 'timing/data': 0.017713351175189018, 'timing/model': 1.4385517686605453, 'learning_rate/qwen_vl_interface': 1.1770617536046633e-06, 'learning_rate/action_model': 1.1770617536046632e-05, 'epoch': 1.61}) +[2026-06-16 07:10:15,334][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.02159886062145233, 'train/grad_norm_pre_clip': 3.411104917526245, 'timing/data': 0.022992976009845734, 'timing/model': 1.549274992197752, 'learning_rate/qwen_vl_interface': 1.1754955707775873e-06, 'learning_rate/action_model': 1.1754955707775872e-05, 'epoch': 1.61}) +[2026-06-16 07:10:21,537][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.07659011334180832, 'train/grad_norm_pre_clip': 9.604987144470215, 'timing/data': 0.021219652146100998, 'timing/model': 1.644566334784031, 'learning_rate/qwen_vl_interface': 1.1739310628253753e-06, 'learning_rate/action_model': 1.1739310628253753e-05, 'epoch': 1.61}) +[2026-06-16 07:10:27,731][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.17713917791843414, 'train/grad_norm_pre_clip': 8.68545913696289, 'timing/data': 0.021642759442329407, 'timing/model': 1.4866344518959522, 'learning_rate/qwen_vl_interface': 1.172368230391136e-06, 'learning_rate/action_model': 1.1723682303911359e-05, 'epoch': 1.61}) +[2026-06-16 07:10:33,849][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.004594719037413597, 'train/grad_norm_pre_clip': 1.0676600933074951, 'timing/data': 0.021660765632987022, 'timing/model': 1.750990068539977, 'learning_rate/qwen_vl_interface': 1.1708070741172928e-06, 'learning_rate/action_model': 1.1708070741172928e-05, 'epoch': 1.61}) +[2026-06-16 07:10:39,729][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.05943906307220459, 'train/grad_norm_pre_clip': 9.249123573303223, 'timing/data': 0.02118256315588951, 'timing/model': 1.5338490959256887, 'learning_rate/qwen_vl_interface': 1.1692475946455755e-06, 'learning_rate/action_model': 1.1692475946455755e-05, 'epoch': 1.61}) +[2026-06-16 07:10:45,931][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.022212952375411987, 'train/grad_norm_pre_clip': 3.465977191925049, 'timing/data': 0.022014031186699867, 'timing/model': 1.53629563562572, 'learning_rate/qwen_vl_interface': 1.1676897926170289e-06, 'learning_rate/action_model': 1.1676897926170289e-05, 'epoch': 1.61}) +[2026-06-16 07:10:52,035][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.0362689271569252, 'train/grad_norm_pre_clip': 10.855777740478516, 'timing/data': 0.023442059755325317, 'timing/model': 1.512693390250206, 'learning_rate/qwen_vl_interface': 1.166133668672006e-06, 'learning_rate/action_model': 1.166133668672006e-05, 'epoch': 1.61}) +[2026-06-16 07:10:58,140][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.14224480092525482, 'train/grad_norm_pre_clip': 6.1535468101501465, 'timing/data': 0.022040631622076035, 'timing/model': 1.5197989139705896, 'learning_rate/qwen_vl_interface': 1.1645792234501706e-06, 'learning_rate/action_model': 1.1645792234501706e-05, 'epoch': 1.61}) +[2026-06-16 07:11:04,127][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.0044789728708565235, 'train/grad_norm_pre_clip': 0.8298912644386292, 'timing/data': 0.02247416414320469, 'timing/model': 1.5254725944250822, 'learning_rate/qwen_vl_interface': 1.1630264575904972e-06, 'learning_rate/action_model': 1.1630264575904973e-05, 'epoch': 1.61}) +[2026-06-16 07:11:10,206][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.10218897461891174, 'train/grad_norm_pre_clip': 14.559602737426758, 'timing/data': 0.022663023322820663, 'timing/model': 1.7173831034451723, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 1.61}) +[2026-06-16 07:11:16,133][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.008286934345960617, 'train/grad_norm_pre_clip': 1.0613678693771362, 'timing/data': 0.023167748004198074, 'timing/model': 1.5443147979676723, 'learning_rate/qwen_vl_interface': 1.1599259665100767e-06, 'learning_rate/action_model': 1.1599259665100767e-05, 'epoch': 1.61}) +[2026-06-16 07:11:22,330][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.14479388296604156, 'train/grad_norm_pre_clip': 11.050432205200195, 'timing/data': 0.022564128041267395, 'timing/model': 1.5429955124855042, 'learning_rate/qwen_vl_interface': 1.1583782425638252e-06, 'learning_rate/action_model': 1.158378242563825e-05, 'epoch': 1.61}) +[2026-06-16 07:11:28,433][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.022922281175851822, 'train/grad_norm_pre_clip': 2.1385793685913086, 'timing/data': 0.022650500759482384, 'timing/model': 1.5482101496309042, 'learning_rate/qwen_vl_interface': 1.1568322005287237e-06, 'learning_rate/action_model': 1.1568322005287235e-05, 'epoch': 1.61}) +[2026-06-16 07:11:34,532][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.07405415177345276, 'train/grad_norm_pre_clip': 16.98999786376953, 'timing/data': 0.02273494191467762, 'timing/model': 1.5253248065710068, 'learning_rate/qwen_vl_interface': 1.1552878410402923e-06, 'learning_rate/action_model': 1.1552878410402924e-05, 'epoch': 1.61}) +[2026-06-16 07:11:40,545][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.00884825550019741, 'train/grad_norm_pre_clip': 1.9849904775619507, 'timing/data': 0.022596631199121475, 'timing/model': 1.47828958183527, 'learning_rate/qwen_vl_interface': 1.1537451647333594e-06, 'learning_rate/action_model': 1.1537451647333593e-05, 'epoch': 1.61}) +[2026-06-16 07:11:46,612][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.07995234429836273, 'train/grad_norm_pre_clip': 11.471341133117676, 'timing/data': 0.022938992828130722, 'timing/model': 1.7251482717692852, 'learning_rate/qwen_vl_interface': 1.1522041722420594e-06, 'learning_rate/action_model': 1.1522041722420594e-05, 'epoch': 1.62}) +[2026-06-16 07:11:52,530][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.006770499981939793, 'train/grad_norm_pre_clip': 1.2378047704696655, 'timing/data': 0.022648731246590614, 'timing/model': 1.541624791920185, 'learning_rate/qwen_vl_interface': 1.1506648641998365e-06, 'learning_rate/action_model': 1.1506648641998365e-05, 'epoch': 1.62}) +[2026-06-16 07:11:58,633][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 0.040674224495887756, 'train/grad_norm_pre_clip': 16.54413414001465, 'timing/data': 0.021304864436388016, 'timing/model': 1.5470598991960287, 'learning_rate/qwen_vl_interface': 1.1491272412394433e-06, 'learning_rate/action_model': 1.1491272412394433e-05, 'epoch': 1.62}) +[2026-06-16 07:12:04,734][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.21271708607673645, 'train/grad_norm_pre_clip': 7.002330303192139, 'timing/data': 0.021715741604566574, 'timing/model': 1.545582439750433, 'learning_rate/qwen_vl_interface': 1.1475913039929373e-06, 'learning_rate/action_model': 1.1475913039929372e-05, 'epoch': 1.62}) +[2026-06-16 07:12:10,829][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.03614982217550278, 'train/grad_norm_pre_clip': 5.696824550628662, 'timing/data': 0.023126304149627686, 'timing/model': 1.5146043077111244, 'learning_rate/qwen_vl_interface': 1.1460570530916844e-06, 'learning_rate/action_model': 1.1460570530916844e-05, 'epoch': 1.62}) +[2026-06-16 07:12:16,944][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.019360052421689034, 'train/grad_norm_pre_clip': 2.0492377281188965, 'timing/data': 0.022610142827033997, 'timing/model': 1.7634602580219507, 'learning_rate/qwen_vl_interface': 1.1445244891663581e-06, 'learning_rate/action_model': 1.1445244891663581e-05, 'epoch': 1.62}) +[2026-06-16 07:12:22,931][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 0.07662448287010193, 'train/grad_norm_pre_clip': 6.2818217277526855, 'timing/data': 0.024730436503887177, 'timing/model': 1.5359348505735397, 'learning_rate/qwen_vl_interface': 1.1429936128469366e-06, 'learning_rate/action_model': 1.1429936128469366e-05, 'epoch': 1.62}) +[2026-06-16 07:12:29,016][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.3803694248199463, 'train/grad_norm_pre_clip': 23.76600456237793, 'timing/data': 0.022404305636882782, 'timing/model': 1.5386083405464888, 'learning_rate/qwen_vl_interface': 1.1414644247627056e-06, 'learning_rate/action_model': 1.1414644247627058e-05, 'epoch': 1.62}) +[2026-06-16 07:12:35,032][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.017429035156965256, 'train/grad_norm_pre_clip': 3.3983027935028076, 'timing/data': 0.02201109565794468, 'timing/model': 1.5450017116963863, 'learning_rate/qwen_vl_interface': 1.1399369255422568e-06, 'learning_rate/action_model': 1.1399369255422568e-05, 'epoch': 1.62}) +[2026-06-16 07:12:41,106][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.15401500463485718, 'train/grad_norm_pre_clip': 23.38329315185547, 'timing/data': 0.022242946550250053, 'timing/model': 1.5258871018886566, 'learning_rate/qwen_vl_interface': 1.138411115813488e-06, 'learning_rate/action_model': 1.138411115813488e-05, 'epoch': 1.62}) +[2026-06-16 07:12:47,133][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.025175023823976517, 'train/grad_norm_pre_clip': 3.3892621994018555, 'timing/data': 0.021432077512145042, 'timing/model': 1.4712314177304506, 'learning_rate/qwen_vl_interface': 1.1368869962036006e-06, 'learning_rate/action_model': 1.1368869962036005e-05, 'epoch': 1.62}) +[2026-06-16 07:12:53,274][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.02904634363949299, 'train/grad_norm_pre_clip': 4.695971488952637, 'timing/data': 0.02494068257510662, 'timing/model': 1.7852863445878029, 'learning_rate/qwen_vl_interface': 1.1353645673391025e-06, 'learning_rate/action_model': 1.1353645673391026e-05, 'epoch': 1.62}) +[2026-06-16 07:12:59,327][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.03815017640590668, 'train/grad_norm_pre_clip': 9.949007987976074, 'timing/data': 0.026408422738313675, 'timing/model': 1.5420851409435272, 'learning_rate/qwen_vl_interface': 1.1338438298458078e-06, 'learning_rate/action_model': 1.1338438298458078e-05, 'epoch': 1.62}) +[2026-06-16 07:13:05,430][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.006229727063328028, 'train/grad_norm_pre_clip': 0.5416719317436218, 'timing/data': 0.022493693977594376, 'timing/model': 1.5436333268880844, 'learning_rate/qwen_vl_interface': 1.1323247843488345e-06, 'learning_rate/action_model': 1.1323247843488344e-05, 'epoch': 1.62}) +[2026-06-16 07:13:11,530][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.0629127025604248, 'train/grad_norm_pre_clip': 19.16142463684082, 'timing/data': 0.023143310099840164, 'timing/model': 1.532963888719678, 'learning_rate/qwen_vl_interface': 1.130807431472603e-06, 'learning_rate/action_model': 1.130807431472603e-05, 'epoch': 1.62}) +[2026-06-16 07:13:17,527][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.0009545334032736719, 'train/grad_norm_pre_clip': 0.12335741519927979, 'timing/data': 0.02155836671590805, 'timing/model': 1.540137005969882, 'learning_rate/qwen_vl_interface': 1.1292917718408416e-06, 'learning_rate/action_model': 1.1292917718408416e-05, 'epoch': 1.62}) +[2026-06-16 07:13:23,625][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.04502802714705467, 'train/grad_norm_pre_clip': 5.043008327484131, 'timing/data': 0.021790891885757446, 'timing/model': 1.5433551501482725, 'learning_rate/qwen_vl_interface': 1.12777780607658e-06, 'learning_rate/action_model': 1.12777780607658e-05, 'epoch': 1.62}) +[2026-06-16 07:13:29,465][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.025708932429552078, 'train/grad_norm_pre_clip': 7.053754806518555, 'timing/data': 0.018054913729429245, 'timing/model': 1.634898278862238, 'learning_rate/qwen_vl_interface': 1.1262655348021523e-06, 'learning_rate/action_model': 1.1262655348021523e-05, 'epoch': 1.62}) +[2026-06-16 07:13:35,070][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.05045964941382408, 'train/grad_norm_pre_clip': 20.936426162719727, 'timing/data': 0.01844615489244461, 'timing/model': 1.4384830128401518, 'learning_rate/qwen_vl_interface': 1.1247549586391966e-06, 'learning_rate/action_model': 1.1247549586391964e-05, 'epoch': 1.62}) +[2026-06-16 07:13:40,776][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.12093155831098557, 'train/grad_norm_pre_clip': 10.597569465637207, 'timing/data': 0.019440360367298126, 'timing/model': 1.4423906151205301, 'learning_rate/qwen_vl_interface': 1.1232460782086534e-06, 'learning_rate/action_model': 1.1232460782086534e-05, 'epoch': 1.62}) +[2026-06-16 07:13:46,581][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.021006178110837936, 'train/grad_norm_pre_clip': 5.872375011444092, 'timing/data': 0.018483445048332214, 'timing/model': 1.4326815363019705, 'learning_rate/qwen_vl_interface': 1.1217388941307658e-06, 'learning_rate/action_model': 1.1217388941307658e-05, 'epoch': 1.62}) +[2026-06-16 07:13:52,478][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.15893149375915527, 'train/grad_norm_pre_clip': 11.849874496459961, 'timing/data': 0.018510127440094948, 'timing/model': 1.439569415524602, 'learning_rate/qwen_vl_interface': 1.1202334070250807e-06, 'learning_rate/action_model': 1.1202334070250806e-05, 'epoch': 1.62}) +[2026-06-16 07:13:58,389][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.2101915180683136, 'train/grad_norm_pre_clip': 15.896350860595703, 'timing/data': 0.017490850761532784, 'timing/model': 1.563001610338688, 'learning_rate/qwen_vl_interface': 1.1187296175104472e-06, 'learning_rate/action_model': 1.1187296175104471e-05, 'epoch': 1.62}) +[2026-06-16 07:14:04,251][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.0011451509781181812, 'train/grad_norm_pre_clip': 0.20134013891220093, 'timing/data': 0.019300105050206184, 'timing/model': 1.6140574049204588, 'learning_rate/qwen_vl_interface': 1.1172275262050158e-06, 'learning_rate/action_model': 1.1172275262050159e-05, 'epoch': 1.62}) +[2026-06-16 07:14:09,927][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.1696184128522873, 'train/grad_norm_pre_clip': 30.698833465576172, 'timing/data': 0.018613362684845924, 'timing/model': 1.4966513831168413, 'learning_rate/qwen_vl_interface': 1.1157271337262401e-06, 'learning_rate/action_model': 1.11572713372624e-05, 'epoch': 1.62}) +[2026-06-16 07:14:16,132][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 0.03073885850608349, 'train/grad_norm_pre_clip': 5.903652191162109, 'timing/data': 0.02232147753238678, 'timing/model': 1.5382046196609735, 'learning_rate/qwen_vl_interface': 1.1142284406908751e-06, 'learning_rate/action_model': 1.114228440690875e-05, 'epoch': 1.62}) +[2026-06-16 07:14:22,233][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.0018288358114659786, 'train/grad_norm_pre_clip': 0.5225504636764526, 'timing/data': 0.022338755428791046, 'timing/model': 1.5464750956743956, 'learning_rate/qwen_vl_interface': 1.1127314477149758e-06, 'learning_rate/action_model': 1.1127314477149759e-05, 'epoch': 1.63}) +[2026-06-16 07:21:35,806][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.4677260220050812, 'train/grad_norm_pre_clip': 32.37561798095703, 'eval/loss': 0.04270390421152115, 'eval/flap_precision': 0.9631147540983607, 'eval/flap_recall': 0.9565807327001357, 'eval/flap_f1': 0.9598366235534377, 'eval/flap_rate_model': 0.20333333333333334, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9598366235534376, 'timing/data': 0.022139351814985275, 'timing/model': 1.5457576792687178, 'learning_rate/qwen_vl_interface': 1.1112361554139003e-06, 'learning_rate/action_model': 1.1112361554139004e-05, 'epoch': 1.63}) +[2026-06-16 07:21:35,807][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4200_state +[2026-06-16 07:21:35,813][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 07:21:43,603][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4200_state/model.safetensors +[2026-06-16 07:21:53,798][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4200_state/optimizer.bin +[2026-06-16 07:21:53,799][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4200_state/sampler.bin +[2026-06-16 07:21:53,799][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4200_state/sampler_1.bin +[2026-06-16 07:21:53,800][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4200_state/random_states_0.pkl +[2026-06-16 07:21:53,898][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4200 +[2026-06-16 07:21:53,898][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 07:21:53,906][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 07:21:59,793][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.20198661088943481, 'train/grad_norm_pre_clip': 15.204080581665039, 'timing/data': 0.01895871013402939, 'timing/model': 1.4306953642517328, 'learning_rate/qwen_vl_interface': 1.1097425644023065e-06, 'learning_rate/action_model': 1.1097425644023065e-05, 'epoch': 1.63}) +[2026-06-16 07:22:05,595][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 0.009191361255943775, 'train/grad_norm_pre_clip': 1.1921921968460083, 'timing/data': 0.018148932605981827, 'timing/model': 1.4457511771470308, 'learning_rate/qwen_vl_interface': 1.1082506752941529e-06, 'learning_rate/action_model': 1.1082506752941528e-05, 'epoch': 1.63}) +[2026-06-16 07:22:11,498][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.1766682267189026, 'train/grad_norm_pre_clip': 18.117477416992188, 'timing/data': 0.018773570656776428, 'timing/model': 1.4681290164589882, 'learning_rate/qwen_vl_interface': 1.1067604887027004e-06, 'learning_rate/action_model': 1.1067604887027004e-05, 'epoch': 1.63}) +[2026-06-16 07:22:17,401][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.04734513908624649, 'train/grad_norm_pre_clip': 14.77194595336914, 'timing/data': 0.019251663237810135, 'timing/model': 1.6505199521780014, 'learning_rate/qwen_vl_interface': 1.1052720052405066e-06, 'learning_rate/action_model': 1.1052720052405064e-05, 'epoch': 1.63}) +[2026-06-16 07:22:23,071][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.17799898982048035, 'train/grad_norm_pre_clip': 17.34963607788086, 'timing/data': 0.01793849468231201, 'timing/model': 1.4050777275115252, 'learning_rate/qwen_vl_interface': 1.1037852255194309e-06, 'learning_rate/action_model': 1.1037852255194308e-05, 'epoch': 1.63}) +[2026-06-16 07:22:28,884][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 0.001841872581280768, 'train/grad_norm_pre_clip': 0.43481874465942383, 'timing/data': 0.019002223387360573, 'timing/model': 1.447246428579092, 'learning_rate/qwen_vl_interface': 1.1023001501506329e-06, 'learning_rate/action_model': 1.102300150150633e-05, 'epoch': 1.63}) +[2026-06-16 07:22:34,775][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.026560666039586067, 'train/grad_norm_pre_clip': 12.162588119506836, 'timing/data': 0.01749161258339882, 'timing/model': 1.4482737015932798, 'learning_rate/qwen_vl_interface': 1.1008167797445717e-06, 'learning_rate/action_model': 1.1008167797445718e-05, 'epoch': 1.63}) +[2026-06-16 07:22:40,576][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.08257998526096344, 'train/grad_norm_pre_clip': 14.716064453125, 'timing/data': 0.019705310463905334, 'timing/model': 1.440810453146696, 'learning_rate/qwen_vl_interface': 1.099335114911003e-06, 'learning_rate/action_model': 1.0993351149110031e-05, 'epoch': 1.63}) +[2026-06-16 07:22:46,495][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.060620736330747604, 'train/grad_norm_pre_clip': 19.825679779052734, 'timing/data': 0.019190000370144844, 'timing/model': 1.4532689210027456, 'learning_rate/qwen_vl_interface': 1.0978551562589846e-06, 'learning_rate/action_model': 1.0978551562589845e-05, 'epoch': 1.63}) +[2026-06-16 07:22:52,290][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.017342248931527138, 'train/grad_norm_pre_clip': 3.9593732357025146, 'timing/data': 0.018270593136548996, 'timing/model': 1.5621471777558327, 'learning_rate/qwen_vl_interface': 1.0963769043968708e-06, 'learning_rate/action_model': 1.0963769043968708e-05, 'epoch': 1.63}) +[2026-06-16 07:22:58,018][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.06890393048524857, 'train/grad_norm_pre_clip': 6.7300238609313965, 'timing/data': 0.01933273673057556, 'timing/model': 1.4651301987469196, 'learning_rate/qwen_vl_interface': 1.0949003599323158e-06, 'learning_rate/action_model': 1.0949003599323157e-05, 'epoch': 1.63}) +[2026-06-16 07:23:04,305][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 0.11220797896385193, 'train/grad_norm_pre_clip': 20.228731155395508, 'timing/data': 0.022975299507379532, 'timing/model': 1.4580822195857763, 'learning_rate/qwen_vl_interface': 1.0934255234722708e-06, 'learning_rate/action_model': 1.0934255234722706e-05, 'epoch': 1.63}) +[2026-06-16 07:23:10,734][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.06676198542118073, 'train/grad_norm_pre_clip': 19.550113677978516, 'timing/data': 0.022813452407717705, 'timing/model': 1.5307941120117903, 'learning_rate/qwen_vl_interface': 1.0919523956229858e-06, 'learning_rate/action_model': 1.0919523956229857e-05, 'epoch': 1.63}) +[2026-06-16 07:23:16,836][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.01199278049170971, 'train/grad_norm_pre_clip': 5.560705661773682, 'timing/data': 0.023673899471759796, 'timing/model': 1.5204117130488157, 'learning_rate/qwen_vl_interface': 1.0904809769900075e-06, 'learning_rate/action_model': 1.0904809769900073e-05, 'epoch': 1.63}) +[2026-06-16 07:23:23,133][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.03992901369929314, 'train/grad_norm_pre_clip': 4.692152976989746, 'timing/data': 0.021700384095311165, 'timing/model': 1.841414362192154, 'learning_rate/qwen_vl_interface': 1.0890112681781821e-06, 'learning_rate/action_model': 1.089011268178182e-05, 'epoch': 1.63}) +[2026-06-16 07:23:29,033][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.007555563934147358, 'train/grad_norm_pre_clip': 1.8033521175384521, 'timing/data': 0.02315591834485531, 'timing/model': 1.540515162050724, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 1.63}) +[2026-06-16 07:23:35,133][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.10212302207946777, 'train/grad_norm_pre_clip': 12.157194137573242, 'timing/data': 0.022145073860883713, 'timing/model': 1.541789211332798, 'learning_rate/qwen_vl_interface': 1.0860769824338477e-06, 'learning_rate/action_model': 1.0860769824338478e-05, 'epoch': 1.63}) +[2026-06-16 07:23:41,236][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.0841975286602974, 'train/grad_norm_pre_clip': 10.457537651062012, 'timing/data': 0.022144468501210213, 'timing/model': 1.5431029833853245, 'learning_rate/qwen_vl_interface': 1.084612406707514e-06, 'learning_rate/action_model': 1.084612406707514e-05, 'epoch': 1.63}) +[2026-06-16 07:23:47,332][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.05263196676969528, 'train/grad_norm_pre_clip': 11.383627891540527, 'timing/data': 0.021873855963349342, 'timing/model': 1.5213740542531013, 'learning_rate/qwen_vl_interface': 1.0831495432146795e-06, 'learning_rate/action_model': 1.0831495432146795e-05, 'epoch': 1.63}) +[2026-06-16 07:23:53,333][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.12400709837675095, 'train/grad_norm_pre_clip': 24.488502502441406, 'timing/data': 0.022860124707221985, 'timing/model': 1.5025832522660494, 'learning_rate/qwen_vl_interface': 1.0816883925566717e-06, 'learning_rate/action_model': 1.0816883925566716e-05, 'epoch': 1.63}) +[2026-06-16 07:23:59,401][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.025003768503665924, 'train/grad_norm_pre_clip': 3.3647501468658447, 'timing/data': 0.023685386404395103, 'timing/model': 1.70701165497303, 'learning_rate/qwen_vl_interface': 1.0802289553341153e-06, 'learning_rate/action_model': 1.0802289553341152e-05, 'epoch': 1.63}) +[2026-06-16 07:24:05,340][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.10364029556512833, 'train/grad_norm_pre_clip': 7.203773021697998, 'timing/data': 0.023440413177013397, 'timing/model': 1.5440057013183832, 'learning_rate/qwen_vl_interface': 1.0787712321469289e-06, 'learning_rate/action_model': 1.0787712321469288e-05, 'epoch': 1.63}) +[2026-06-16 07:24:11,446][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.005870644003152847, 'train/grad_norm_pre_clip': 2.8149123191833496, 'timing/data': 0.022094959393143654, 'timing/model': 1.5520586762577295, 'learning_rate/qwen_vl_interface': 1.077315223594328e-06, 'learning_rate/action_model': 1.077315223594328e-05, 'epoch': 1.63}) +[2026-06-16 07:24:17,536][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.07615100592374802, 'train/grad_norm_pre_clip': 19.122482299804688, 'timing/data': 0.02203330025076866, 'timing/model': 1.5307829193770885, 'learning_rate/qwen_vl_interface': 1.0758609302748227e-06, 'learning_rate/action_model': 1.0758609302748227e-05, 'epoch': 1.63}) +[2026-06-16 07:24:23,639][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.035824500024318695, 'train/grad_norm_pre_clip': 7.354759216308594, 'timing/data': 0.02261955477297306, 'timing/model': 1.5235729217529297, 'learning_rate/qwen_vl_interface': 1.0744083527862182e-06, 'learning_rate/action_model': 1.074408352786218e-05, 'epoch': 1.64}) +[2026-06-16 07:24:29,658][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.007481606211513281, 'train/grad_norm_pre_clip': 1.7767623662948608, 'timing/data': 0.022200996056199074, 'timing/model': 1.486308166757226, 'learning_rate/qwen_vl_interface': 1.0729574917256144e-06, 'learning_rate/action_model': 1.0729574917256145e-05, 'epoch': 1.64}) +[2026-06-16 07:24:35,721][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.007518246304243803, 'train/grad_norm_pre_clip': 0.9893423914909363, 'timing/data': 0.02191995084285736, 'timing/model': 1.724997028708458, 'learning_rate/qwen_vl_interface': 1.0715083476894047e-06, 'learning_rate/action_model': 1.0715083476894047e-05, 'epoch': 1.64}) +[2026-06-16 07:24:41,636][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 0.0111060980707407, 'train/grad_norm_pre_clip': 1.3598893880844116, 'timing/data': 0.021990671753883362, 'timing/model': 1.544091571122408, 'learning_rate/qwen_vl_interface': 1.0700609212732773e-06, 'learning_rate/action_model': 1.0700609212732774e-05, 'epoch': 1.64}) +[2026-06-16 07:24:47,483][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.019160807132720947, 'train/grad_norm_pre_clip': 12.883784294128418, 'timing/data': 0.0187942236661911, 'timing/model': 1.4450468756258488, 'learning_rate/qwen_vl_interface': 1.0686152130722155e-06, 'learning_rate/action_model': 1.0686152130722154e-05, 'epoch': 1.64}) +[2026-06-16 07:24:53,385][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.09615793079137802, 'train/grad_norm_pre_clip': 17.130874633789062, 'timing/data': 0.018734194338321686, 'timing/model': 1.445806672796607, 'learning_rate/qwen_vl_interface': 1.0671712236804947e-06, 'learning_rate/action_model': 1.0671712236804946e-05, 'epoch': 1.64}) +[2026-06-16 07:24:59,280][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 0.06695287674665451, 'train/grad_norm_pre_clip': 5.113563537597656, 'timing/data': 0.018585216253995895, 'timing/model': 1.4545152727514505, 'learning_rate/qwen_vl_interface': 1.0657289536916845e-06, 'learning_rate/action_model': 1.0657289536916845e-05, 'epoch': 1.64}) +[2026-06-16 07:25:05,159][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.020293591544032097, 'train/grad_norm_pre_clip': 3.382960796356201, 'timing/data': 0.018586935475468636, 'timing/model': 1.6222465354949236, 'learning_rate/qwen_vl_interface': 1.0642884036986481e-06, 'learning_rate/action_model': 1.064288403698648e-05, 'epoch': 1.64}) +[2026-06-16 07:25:10,782][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.2866726219654083, 'train/grad_norm_pre_clip': 18.41820526123047, 'timing/data': 0.01878243125975132, 'timing/model': 1.4422545358538628, 'learning_rate/qwen_vl_interface': 1.0628495742935404e-06, 'learning_rate/action_model': 1.0628495742935405e-05, 'epoch': 1.64}) +[2026-06-16 07:25:16,781][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 0.2772412896156311, 'train/grad_norm_pre_clip': 14.045853614807129, 'timing/data': 0.01837102882564068, 'timing/model': 1.4451966751366854, 'learning_rate/qwen_vl_interface': 1.0614124660678103e-06, 'learning_rate/action_model': 1.0614124660678102e-05, 'epoch': 1.64}) +[2026-06-16 07:25:22,482][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.07950448244810104, 'train/grad_norm_pre_clip': 7.969701766967773, 'timing/data': 0.018796173855662346, 'timing/model': 1.4455132614821196, 'learning_rate/qwen_vl_interface': 1.0599770796121988e-06, 'learning_rate/action_model': 1.0599770796121987e-05, 'epoch': 1.64}) +[2026-06-16 07:25:28,280][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.0969175398349762, 'train/grad_norm_pre_clip': 21.296457290649414, 'timing/data': 0.018771016970276833, 'timing/model': 1.442340375855565, 'learning_rate/qwen_vl_interface': 1.0585434155167394e-06, 'learning_rate/action_model': 1.0585434155167393e-05, 'epoch': 1.64}) +[2026-06-16 07:25:34,093][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.02943681739270687, 'train/grad_norm_pre_clip': 4.020589828491211, 'timing/data': 0.01966186799108982, 'timing/model': 1.4159465413540602, 'learning_rate/qwen_vl_interface': 1.0571114743707556e-06, 'learning_rate/action_model': 1.0571114743707555e-05, 'epoch': 1.64}) +[2026-06-16 07:25:39,971][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.22571620345115662, 'train/grad_norm_pre_clip': 4.451505184173584, 'timing/data': 0.018908685073256493, 'timing/model': 1.6354734729975462, 'learning_rate/qwen_vl_interface': 1.0556812567628661e-06, 'learning_rate/action_model': 1.055681256762866e-05, 'epoch': 1.64}) +[2026-06-16 07:25:45,583][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.07196709513664246, 'train/grad_norm_pre_clip': 12.427133560180664, 'timing/data': 0.018667930737137794, 'timing/model': 1.4472323898226023, 'learning_rate/qwen_vl_interface': 1.0542527632809783e-06, 'learning_rate/action_model': 1.0542527632809784e-05, 'epoch': 1.64}) +[2026-06-16 07:25:51,482][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.00630346592515707, 'train/grad_norm_pre_clip': 1.234176754951477, 'timing/data': 0.01735253818333149, 'timing/model': 1.4460009187459946, 'learning_rate/qwen_vl_interface': 1.0528259945122925e-06, 'learning_rate/action_model': 1.0528259945122925e-05, 'epoch': 1.64}) +[2026-06-16 07:25:57,380][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.06027192622423172, 'train/grad_norm_pre_clip': 5.245941162109375, 'timing/data': 0.018887387588620186, 'timing/model': 1.443574147298932, 'learning_rate/qwen_vl_interface': 1.0514009510432988e-06, 'learning_rate/action_model': 1.0514009510432988e-05, 'epoch': 1.64}) +[2026-06-16 07:26:03,283][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.07281076163053513, 'train/grad_norm_pre_clip': 6.460565090179443, 'timing/data': 0.01974497362971306, 'timing/model': 1.4442487638443708, 'learning_rate/qwen_vl_interface': 1.0499776334597796e-06, 'learning_rate/action_model': 1.0499776334597797e-05, 'epoch': 1.64}) +[2026-06-16 07:26:09,511][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.0022095004096627235, 'train/grad_norm_pre_clip': 0.481833279132843, 'timing/data': 0.019599638879299164, 'timing/model': 1.4531811736524105, 'learning_rate/qwen_vl_interface': 1.0485560423468064e-06, 'learning_rate/action_model': 1.0485560423468063e-05, 'epoch': 1.64}) +[2026-06-16 07:26:15,499][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.025683965533971786, 'train/grad_norm_pre_clip': 14.064648628234863, 'timing/data': 0.019312575459480286, 'timing/model': 1.6453688349574804, 'learning_rate/qwen_vl_interface': 1.0471361782887421e-06, 'learning_rate/action_model': 1.0471361782887421e-05, 'epoch': 1.64}) +[2026-06-16 07:26:21,216][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.009472946636378765, 'train/grad_norm_pre_clip': 1.2056849002838135, 'timing/data': 0.01876143552362919, 'timing/model': 1.448805058375001, 'learning_rate/qwen_vl_interface': 1.0457180418692385e-06, 'learning_rate/action_model': 1.0457180418692383e-05, 'epoch': 1.64}) +[2026-06-16 07:26:27,118][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.09374626725912094, 'train/grad_norm_pre_clip': 8.520545959472656, 'timing/data': 0.018018202856183052, 'timing/model': 1.4485147092491388, 'learning_rate/qwen_vl_interface': 1.0443016336712395e-06, 'learning_rate/action_model': 1.0443016336712393e-05, 'epoch': 1.64}) +[2026-06-16 07:26:33,022][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.006020728964358568, 'train/grad_norm_pre_clip': 1.8852912187576294, 'timing/data': 0.01940814219415188, 'timing/model': 1.4484015386551619, 'learning_rate/qwen_vl_interface': 1.0428869542769744e-06, 'learning_rate/action_model': 1.0428869542769744e-05, 'epoch': 1.64}) +[2026-06-16 07:26:38,913][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.18951761722564697, 'train/grad_norm_pre_clip': 13.17732048034668, 'timing/data': 0.01913132146000862, 'timing/model': 1.446403743699193, 'learning_rate/qwen_vl_interface': 1.0414740042679663e-06, 'learning_rate/action_model': 1.0414740042679662e-05, 'epoch': 1.64}) +[2026-06-16 07:26:44,812][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.038667961955070496, 'train/grad_norm_pre_clip': 11.648799896240234, 'timing/data': 0.01918191649019718, 'timing/model': 1.4493668898940086, 'learning_rate/qwen_vl_interface': 1.040062784225025e-06, 'learning_rate/action_model': 1.0400627842250249e-05, 'epoch': 1.64}) +[2026-06-16 07:26:50,662][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.09761054068803787, 'train/grad_norm_pre_clip': 9.292806625366211, 'timing/data': 0.01883842423558235, 'timing/model': 1.6264333818107843, 'learning_rate/qwen_vl_interface': 1.0386532947282496e-06, 'learning_rate/action_model': 1.0386532947282495e-05, 'epoch': 1.64}) +[2026-06-16 07:26:56,287][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.005587995983660221, 'train/grad_norm_pre_clip': 1.3199384212493896, 'timing/data': 0.018085384741425514, 'timing/model': 1.4350078999996185, 'learning_rate/qwen_vl_interface': 1.0372455363570282e-06, 'learning_rate/action_model': 1.0372455363570283e-05, 'epoch': 1.65}) +[2026-06-16 07:27:02,737][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.13211601972579956, 'train/grad_norm_pre_clip': 11.015003204345703, 'timing/data': 0.023224487900733948, 'timing/model': 1.5459008775651455, 'learning_rate/qwen_vl_interface': 1.035839509690039e-06, 'learning_rate/action_model': 1.0358395096900388e-05, 'epoch': 1.65}) +[2026-06-16 07:27:08,938][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.20267760753631592, 'train/grad_norm_pre_clip': 22.85542869567871, 'timing/data': 0.022864622995257378, 'timing/model': 1.6437566839158535, 'learning_rate/qwen_vl_interface': 1.0344352153052429e-06, 'learning_rate/action_model': 1.0344352153052427e-05, 'epoch': 1.65}) +[2026-06-16 07:27:15,266][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.208296000957489, 'train/grad_norm_pre_clip': 12.108211517333984, 'timing/data': 0.02231026627123356, 'timing/model': 1.5256149619817734, 'learning_rate/qwen_vl_interface': 1.033032653779894e-06, 'learning_rate/action_model': 1.033032653779894e-05, 'epoch': 1.65}) +[2026-06-16 07:27:21,464][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.17561763525009155, 'train/grad_norm_pre_clip': 15.029501914978027, 'timing/data': 0.024317393079400063, 'timing/model': 1.5177514404058456, 'learning_rate/qwen_vl_interface': 1.0316318256905325e-06, 'learning_rate/action_model': 1.0316318256905324e-05, 'epoch': 1.65}) +[2026-06-16 07:27:27,683][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.08396324515342712, 'train/grad_norm_pre_clip': 8.951541900634766, 'timing/data': 0.023297252133488655, 'timing/model': 1.7611130010336637, 'learning_rate/qwen_vl_interface': 1.0302327316129856e-06, 'learning_rate/action_model': 1.0302327316129855e-05, 'epoch': 1.65}) +[2026-06-16 07:27:33,762][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.0012665825197473168, 'train/grad_norm_pre_clip': 0.28021714091300964, 'timing/data': 0.02250536158680916, 'timing/model': 1.532036691904068, 'learning_rate/qwen_vl_interface': 1.028835372122368e-06, 'learning_rate/action_model': 1.0288353721223682e-05, 'epoch': 1.65}) +[2026-06-16 07:27:40,105][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 0.0016652269987389445, 'train/grad_norm_pre_clip': 0.589453399181366, 'timing/data': 0.029885010793805122, 'timing/model': 1.630224460735917, 'learning_rate/qwen_vl_interface': 1.0274397477930822e-06, 'learning_rate/action_model': 1.0274397477930822e-05, 'epoch': 1.65}) +[2026-06-16 07:27:46,494][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.030353572219610214, 'train/grad_norm_pre_clip': 9.54896068572998, 'timing/data': 0.03245255723595619, 'timing/model': 1.5333237685263157, 'learning_rate/qwen_vl_interface': 1.0260458591988164e-06, 'learning_rate/action_model': 1.0260458591988165e-05, 'epoch': 1.65}) +[2026-06-16 07:27:52,974][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.03359565511345863, 'train/grad_norm_pre_clip': 9.792486190795898, 'timing/data': 0.022964054718613625, 'timing/model': 1.541615666821599, 'learning_rate/qwen_vl_interface': 1.0246537069125445e-06, 'learning_rate/action_model': 1.0246537069125444e-05, 'epoch': 1.65}) +[2026-06-16 07:27:59,173][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.07377516478300095, 'train/grad_norm_pre_clip': 11.988182067871094, 'timing/data': 0.022729411721229553, 'timing/model': 1.5102971214801073, 'learning_rate/qwen_vl_interface': 1.0232632915065286e-06, 'learning_rate/action_model': 1.0232632915065285e-05, 'epoch': 1.65}) +[2026-06-16 07:28:05,498][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.02311197854578495, 'train/grad_norm_pre_clip': 4.348830699920654, 'timing/data': 0.022591223940253258, 'timing/model': 1.7735866028815508, 'learning_rate/qwen_vl_interface': 1.021874613552316e-06, 'learning_rate/action_model': 1.0218746135523157e-05, 'epoch': 1.65}) +[2026-06-16 07:28:11,566][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.13146580755710602, 'train/grad_norm_pre_clip': 10.855037689208984, 'timing/data': 0.02222646400332451, 'timing/model': 1.5426228046417236, 'learning_rate/qwen_vl_interface': 1.0204876736207384e-06, 'learning_rate/action_model': 1.0204876736207384e-05, 'epoch': 1.65}) +[2026-06-16 07:28:17,764][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.06492221355438232, 'train/grad_norm_pre_clip': 16.41805648803711, 'timing/data': 0.02302895113825798, 'timing/model': 1.5362661834806204, 'learning_rate/qwen_vl_interface': 1.0191024722819163e-06, 'learning_rate/action_model': 1.0191024722819162e-05, 'epoch': 1.65}) +[2026-06-16 07:28:23,832][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.10854897648096085, 'train/grad_norm_pre_clip': 11.015477180480957, 'timing/data': 0.02240597829222679, 'timing/model': 1.5404786560684443, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 1.65}) +[2026-06-16 07:28:29,931][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.0011670972453430295, 'train/grad_norm_pre_clip': 0.2593643069267273, 'timing/data': 0.022345801815390587, 'timing/model': 1.5188251975923777, 'learning_rate/qwen_vl_interface': 1.016337287659437e-06, 'learning_rate/action_model': 1.0163372876594368e-05, 'epoch': 1.65}) +[2026-06-16 07:28:36,034][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.027824219316244125, 'train/grad_norm_pre_clip': 6.267847537994385, 'timing/data': 0.022167326882481575, 'timing/model': 1.507002528756857, 'learning_rate/qwen_vl_interface': 1.0149573055124427e-06, 'learning_rate/action_model': 1.0149573055124426e-05, 'epoch': 1.65}) +[2026-06-16 07:28:42,146][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.035975586622953415, 'train/grad_norm_pre_clip': 7.069951057434082, 'timing/data': 0.022781815379858017, 'timing/model': 1.758484072983265, 'learning_rate/qwen_vl_interface': 1.013579064231529e-06, 'learning_rate/action_model': 1.013579064231529e-05, 'epoch': 1.65}) +[2026-06-16 07:28:48,170][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.0012561205076053739, 'train/grad_norm_pre_clip': 0.41935300827026367, 'timing/data': 0.02306446246802807, 'timing/model': 1.6644343342632055, 'learning_rate/qwen_vl_interface': 1.012202564383239e-06, 'learning_rate/action_model': 1.0122025643832388e-05, 'epoch': 1.65}) +[2026-06-16 07:28:54,137][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.10130327194929123, 'train/grad_norm_pre_clip': 7.971098899841309, 'timing/data': 0.02227426879107952, 'timing/model': 1.5474611278623343, 'learning_rate/qwen_vl_interface': 1.0108278065333978e-06, 'learning_rate/action_model': 1.0108278065333978e-05, 'epoch': 1.65}) +[2026-06-16 07:29:00,231][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.009863593615591526, 'train/grad_norm_pre_clip': 2.5430614948272705, 'timing/data': 0.023299168795347214, 'timing/model': 1.537605358287692, 'learning_rate/qwen_vl_interface': 1.0094547912471183e-06, 'learning_rate/action_model': 1.0094547912471184e-05, 'epoch': 1.65}) +[2026-06-16 07:29:06,334][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.014606744982302189, 'train/grad_norm_pre_clip': 6.749065399169922, 'timing/data': 0.023366620764136314, 'timing/model': 1.5454836245626211, 'learning_rate/qwen_vl_interface': 1.0080835190887949e-06, 'learning_rate/action_model': 1.0080835190887947e-05, 'epoch': 1.65}) +[2026-06-16 07:29:12,439][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.028818616643548012, 'train/grad_norm_pre_clip': 4.0633649826049805, 'timing/data': 0.021838532760739326, 'timing/model': 1.5339817199856043, 'learning_rate/qwen_vl_interface': 1.0067139906221052e-06, 'learning_rate/action_model': 1.0067139906221052e-05, 'epoch': 1.65}) +[2026-06-16 07:29:18,549][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 0.01459740661084652, 'train/grad_norm_pre_clip': 5.505152702331543, 'timing/data': 0.023335546255111694, 'timing/model': 1.756454385817051, 'learning_rate/qwen_vl_interface': 1.0053462064100112e-06, 'learning_rate/action_model': 1.0053462064100112e-05, 'epoch': 1.65}) +[2026-06-16 07:29:24,473][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.39456889033317566, 'train/grad_norm_pre_clip': 24.077560424804688, 'timing/data': 0.018529308959841728, 'timing/model': 1.654609490185976, 'learning_rate/qwen_vl_interface': 1.003980167014757e-06, 'learning_rate/action_model': 1.003980167014757e-05, 'epoch': 1.65}) +[2026-06-16 07:29:30,082][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.13614340126514435, 'train/grad_norm_pre_clip': 11.17005443572998, 'timing/data': 0.019186293706297874, 'timing/model': 1.444066571071744, 'learning_rate/qwen_vl_interface': 1.0026158729978702e-06, 'learning_rate/action_model': 1.0026158729978702e-05, 'epoch': 1.65}) +[2026-06-16 07:29:35,880][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.19132281839847565, 'train/grad_norm_pre_clip': 11.637517929077148, 'timing/data': 0.018528085201978683, 'timing/model': 1.4479430001229048, 'learning_rate/qwen_vl_interface': 1.0012533249201604e-06, 'learning_rate/action_model': 1.0012533249201602e-05, 'epoch': 1.66}) +[2026-06-16 07:29:41,778][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.01849006488919258, 'train/grad_norm_pre_clip': 4.559744358062744, 'timing/data': 0.018665386363863945, 'timing/model': 1.4453651271760464, 'learning_rate/qwen_vl_interface': 9.99892523341719e-07, 'learning_rate/action_model': 9.99892523341719e-06, 'epoch': 1.66}) +[2026-06-16 07:29:47,680][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.2531590163707733, 'train/grad_norm_pre_clip': 9.070767402648926, 'timing/data': 0.019324487075209618, 'timing/model': 1.4530739970505238, 'learning_rate/qwen_vl_interface': 9.98533468821922e-07, 'learning_rate/action_model': 9.98533468821922e-06, 'epoch': 1.66}) +[2026-06-16 07:29:53,573][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.019818048924207687, 'train/grad_norm_pre_clip': 6.235299587249756, 'timing/data': 0.01940848119556904, 'timing/model': 1.6349301096051931, 'learning_rate/qwen_vl_interface': 9.971761619194231e-07, 'learning_rate/action_model': 9.97176161919423e-06, 'epoch': 1.66}) +[2026-06-16 07:29:59,178][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 0.13339464366436005, 'train/grad_norm_pre_clip': 14.46907901763916, 'timing/data': 0.019382664933800697, 'timing/model': 1.4435253124684095, 'learning_rate/qwen_vl_interface': 9.958206031921606e-07, 'learning_rate/action_model': 9.958206031921604e-06, 'epoch': 1.66}) +[2026-06-16 07:30:04,984][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.031051423400640488, 'train/grad_norm_pre_clip': 11.206202507019043, 'timing/data': 0.018749255686998367, 'timing/model': 1.4442859068512917, 'learning_rate/qwen_vl_interface': 9.944667931973537e-07, 'learning_rate/action_model': 9.944667931973536e-06, 'epoch': 1.66}) +[2026-06-16 07:30:11,079][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.0006556063890457153, 'train/grad_norm_pre_clip': 0.08868682384490967, 'timing/data': 0.018401650711894035, 'timing/model': 1.4440263826400042, 'learning_rate/qwen_vl_interface': 9.93114732491503e-07, 'learning_rate/action_model': 9.93114732491503e-06, 'epoch': 1.66}) +[2026-06-16 07:30:17,116][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.16182361543178558, 'train/grad_norm_pre_clip': 12.665122985839844, 'timing/data': 0.01996571384370327, 'timing/model': 1.489184083417058, 'learning_rate/qwen_vl_interface': 9.91764421630389e-07, 'learning_rate/action_model': 9.91764421630389e-06, 'epoch': 1.66}) +[2026-06-16 07:30:23,112][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.011636455543339252, 'train/grad_norm_pre_clip': 1.150245189666748, 'timing/data': 0.017741138115525246, 'timing/model': 1.5248835682868958, 'learning_rate/qwen_vl_interface': 9.904158611690732e-07, 'learning_rate/action_model': 9.90415861169073e-06, 'epoch': 1.66}) +[2026-06-16 07:30:29,129][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.2753007709980011, 'train/grad_norm_pre_clip': 8.107266426086426, 'timing/data': 0.01899116113781929, 'timing/model': 1.6467457450926304, 'learning_rate/qwen_vl_interface': 9.890690516618987e-07, 'learning_rate/action_model': 9.890690516618985e-06, 'epoch': 1.66}) +[2026-06-16 07:30:34,827][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.17871226370334625, 'train/grad_norm_pre_clip': 18.286279678344727, 'timing/data': 0.01814889721572399, 'timing/model': 1.4545131772756577, 'learning_rate/qwen_vl_interface': 9.877239936624875e-07, 'learning_rate/action_model': 9.877239936624875e-06, 'epoch': 1.66}) +[2026-06-16 07:30:40,913][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.05524561181664467, 'train/grad_norm_pre_clip': 7.71474552154541, 'timing/data': 0.01929536461830139, 'timing/model': 1.4482395760715008, 'learning_rate/qwen_vl_interface': 9.863806877237421e-07, 'learning_rate/action_model': 9.863806877237422e-06, 'epoch': 1.66}) +[2026-06-16 07:30:46,815][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.03275994956493378, 'train/grad_norm_pre_clip': 22.81520652770996, 'timing/data': 0.017142077907919884, 'timing/model': 1.4437337517738342, 'learning_rate/qwen_vl_interface': 9.850391343978454e-07, 'learning_rate/action_model': 9.850391343978454e-06, 'epoch': 1.66}) +[2026-06-16 07:30:52,703][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.0912293940782547, 'train/grad_norm_pre_clip': 6.593712329864502, 'timing/data': 0.018013732507824898, 'timing/model': 1.4469065554440022, 'learning_rate/qwen_vl_interface': 9.836993342362596e-07, 'learning_rate/action_model': 9.836993342362596e-06, 'epoch': 1.66}) +[2026-06-16 07:30:58,743][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.00018672304577194154, 'train/grad_norm_pre_clip': 0.018753493204712868, 'timing/data': 0.02196843922138214, 'timing/model': 1.478844290599227, 'learning_rate/qwen_vl_interface': 9.823612877897246e-07, 'learning_rate/action_model': 9.823612877897245e-06, 'epoch': 1.66}) +[2026-06-16 07:31:04,800][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.019544128328561783, 'train/grad_norm_pre_clip': 9.21341609954834, 'timing/data': 0.02276061661541462, 'timing/model': 1.7065768856555223, 'learning_rate/qwen_vl_interface': 9.81024995608261e-07, 'learning_rate/action_model': 9.810249956082609e-06, 'epoch': 1.66}) +[2026-06-16 07:31:10,579][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.07307954132556915, 'train/grad_norm_pre_clip': 6.200174808502197, 'timing/data': 0.017811618745326996, 'timing/model': 1.4425117373466492, 'learning_rate/qwen_vl_interface': 9.796904582411685e-07, 'learning_rate/action_model': 9.796904582411684e-06, 'epoch': 1.66}) +[2026-06-16 07:31:16,474][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.008580051362514496, 'train/grad_norm_pre_clip': 2.6069211959838867, 'timing/data': 0.01842387020587921, 'timing/model': 1.438269516453147, 'learning_rate/qwen_vl_interface': 9.78357676237025e-07, 'learning_rate/action_model': 9.78357676237025e-06, 'epoch': 1.66}) +[2026-06-16 07:31:22,282][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 0.002283899812027812, 'train/grad_norm_pre_clip': 0.6151410937309265, 'timing/data': 0.018113626167178154, 'timing/model': 1.4466451443731785, 'learning_rate/qwen_vl_interface': 9.77026650143687e-07, 'learning_rate/action_model': 9.77026650143687e-06, 'epoch': 1.66}) +[2026-06-16 07:31:28,178][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.14265021681785583, 'train/grad_norm_pre_clip': 22.62050437927246, 'timing/data': 0.018581591546535492, 'timing/model': 1.443517992272973, 'learning_rate/qwen_vl_interface': 9.756973805082882e-07, 'learning_rate/action_model': 9.756973805082883e-06, 'epoch': 1.66}) +[2026-06-16 07:31:34,074][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.21497748792171478, 'train/grad_norm_pre_clip': 19.746843338012695, 'timing/data': 0.019353345036506653, 'timing/model': 1.6356301307678223, 'learning_rate/qwen_vl_interface': 9.743698678772434e-07, 'learning_rate/action_model': 9.743698678772433e-06, 'epoch': 1.66}) +[2026-06-16 07:31:39,680][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.12577500939369202, 'train/grad_norm_pre_clip': 14.166815757751465, 'timing/data': 0.018767548725008965, 'timing/model': 1.4442207477986813, 'learning_rate/qwen_vl_interface': 9.730441127962395e-07, 'learning_rate/action_model': 9.730441127962395e-06, 'epoch': 1.66}) +[2026-06-16 07:31:45,580][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.08393631875514984, 'train/grad_norm_pre_clip': 13.508502960205078, 'timing/data': 0.017566034570336342, 'timing/model': 1.4459162708371878, 'learning_rate/qwen_vl_interface': 9.717201158102465e-07, 'learning_rate/action_model': 9.717201158102466e-06, 'epoch': 1.66}) +[2026-06-16 07:38:59,316][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.11380649358034134, 'train/grad_norm_pre_clip': 28.16388511657715, 'eval/loss': 0.04076360538601875, 'eval/flap_precision': 0.952191235059761, 'eval/flap_recall': 0.9728629579375848, 'eval/flap_f1': 0.9624161073825503, 'eval/flap_rate_model': 0.20916666666666667, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9624161073825505, 'timing/data': 0.018772291019558907, 'timing/model': 1.4389139600098133, 'learning_rate/qwen_vl_interface': 9.703978774635094e-07, 'learning_rate/action_model': 9.703978774635095e-06, 'epoch': 1.66}) +[2026-06-16 07:38:59,317][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4300_state +[2026-06-16 07:38:59,323][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 07:39:07,400][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4300_state/model.safetensors +[2026-06-16 07:39:18,671][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4300_state/optimizer.bin +[2026-06-16 07:39:18,672][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4300_state/sampler.bin +[2026-06-16 07:39:18,672][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4300_state/sampler_1.bin +[2026-06-16 07:39:18,673][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4300_state/random_states_0.pkl +[2026-06-16 07:39:18,772][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4300 +[2026-06-16 07:39:18,772][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 07:39:18,781][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 07:39:25,142][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.02512478083372116, 'train/grad_norm_pre_clip': 3.239708185195923, 'timing/data': 0.023148534819483757, 'timing/model': 1.541824160143733, 'learning_rate/qwen_vl_interface': 9.6907739829955e-07, 'learning_rate/action_model': 9.6907739829955e-06, 'epoch': 1.66}) +[2026-06-16 07:39:31,346][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.038634512573480606, 'train/grad_norm_pre_clip': 8.88756275177002, 'timing/data': 0.023452043533325195, 'timing/model': 1.5433844588696957, 'learning_rate/qwen_vl_interface': 9.677586788611681e-07, 'learning_rate/action_model': 9.677586788611681e-06, 'epoch': 1.66}) +[2026-06-16 07:39:37,546][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.007843530736863613, 'train/grad_norm_pre_clip': 1.158692717552185, 'timing/data': 0.02176133543252945, 'timing/model': 1.5224510934203863, 'learning_rate/qwen_vl_interface': 9.664417196904382e-07, 'learning_rate/action_model': 9.664417196904382e-06, 'epoch': 1.67}) +[2026-06-16 07:39:43,750][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 0.16808928549289703, 'train/grad_norm_pre_clip': 6.338055610656738, 'timing/data': 0.02238583005964756, 'timing/model': 1.5566717758774757, 'learning_rate/qwen_vl_interface': 9.651265213287134e-07, 'learning_rate/action_model': 9.651265213287134e-06, 'epoch': 1.67}) +[2026-06-16 07:39:49,571][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.0021558695007115602, 'train/grad_norm_pre_clip': 0.6062459349632263, 'timing/data': 0.017622673884034157, 'timing/model': 1.620286464691162, 'learning_rate/qwen_vl_interface': 9.638130843166218e-07, 'learning_rate/action_model': 9.638130843166219e-06, 'epoch': 1.67}) +[2026-06-16 07:39:55,202][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.1768781989812851, 'train/grad_norm_pre_clip': 13.54720687866211, 'timing/data': 0.018666256219148636, 'timing/model': 1.4405859671533108, 'learning_rate/qwen_vl_interface': 9.62501409194068e-07, 'learning_rate/action_model': 9.625014091940678e-06, 'epoch': 1.67}) +[2026-06-16 07:40:01,095][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.25741249322891235, 'train/grad_norm_pre_clip': 9.346278190612793, 'timing/data': 0.01888870820403099, 'timing/model': 1.439471511170268, 'learning_rate/qwen_vl_interface': 9.611914965002317e-07, 'learning_rate/action_model': 9.611914965002316e-06, 'epoch': 1.67}) +[2026-06-16 07:40:06,997][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.0029211437795311213, 'train/grad_norm_pre_clip': 0.3043847382068634, 'timing/data': 0.018413837999105453, 'timing/model': 1.446299271658063, 'learning_rate/qwen_vl_interface': 9.598833467735686e-07, 'learning_rate/action_model': 9.598833467735685e-06, 'epoch': 1.67}) +[2026-06-16 07:40:13,224][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.18155401945114136, 'train/grad_norm_pre_clip': 11.261758804321289, 'timing/data': 0.018167637288570404, 'timing/model': 1.454827370122075, 'learning_rate/qwen_vl_interface': 9.585769605518097e-07, 'learning_rate/action_model': 9.585769605518097e-06, 'epoch': 1.67}) +[2026-06-16 07:40:19,158][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.17395809292793274, 'train/grad_norm_pre_clip': 14.673237800598145, 'timing/data': 0.022169681265950203, 'timing/model': 1.5277576986700296, 'learning_rate/qwen_vl_interface': 9.572723383719611e-07, 'learning_rate/action_model': 9.572723383719612e-06, 'epoch': 1.67}) +[2026-06-16 07:40:25,387][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.01518993079662323, 'train/grad_norm_pre_clip': 4.497972011566162, 'timing/data': 0.021497882902622223, 'timing/model': 1.7694922275841236, 'learning_rate/qwen_vl_interface': 9.559694807703041e-07, 'learning_rate/action_model': 9.55969480770304e-06, 'epoch': 1.67}) +[2026-06-16 07:40:31,459][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.015395021997392178, 'train/grad_norm_pre_clip': 4.148104667663574, 'timing/data': 0.021968623623251915, 'timing/model': 1.542121209204197, 'learning_rate/qwen_vl_interface': 9.546683882823945e-07, 'learning_rate/action_model': 9.546683882823945e-06, 'epoch': 1.67}) +[2026-06-16 07:40:37,660][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.2287379801273346, 'train/grad_norm_pre_clip': 10.917720794677734, 'timing/data': 0.02206810563802719, 'timing/model': 1.5456796828657389, 'learning_rate/qwen_vl_interface': 9.533690614430608e-07, 'learning_rate/action_model': 9.533690614430608e-06, 'epoch': 1.67}) +[2026-06-16 07:40:43,862][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.021698584780097008, 'train/grad_norm_pre_clip': 9.566727638244629, 'timing/data': 0.0222458653151989, 'timing/model': 1.5429456401616335, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 1.67}) +[2026-06-16 07:40:50,058][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.023052332922816277, 'train/grad_norm_pre_clip': 4.986093521118164, 'timing/data': 0.022224461659789085, 'timing/model': 1.5385274030268192, 'learning_rate/qwen_vl_interface': 9.50775706845814e-07, 'learning_rate/action_model': 9.50775706845814e-06, 'epoch': 1.67}) +[2026-06-16 07:40:56,262][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.20909187197685242, 'train/grad_norm_pre_clip': 21.176658630371094, 'timing/data': 0.02192586660385132, 'timing/model': 1.518642956390977, 'learning_rate/qwen_vl_interface': 9.49481680153931e-07, 'learning_rate/action_model': 9.494816801539309e-06, 'epoch': 1.67}) +[2026-06-16 07:41:02,481][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.10614284127950668, 'train/grad_norm_pre_clip': 11.986688613891602, 'timing/data': 0.02215389534831047, 'timing/model': 1.7642682921141386, 'learning_rate/qwen_vl_interface': 9.481894212426848e-07, 'learning_rate/action_model': 9.481894212426847e-06, 'epoch': 1.67}) +[2026-06-16 07:41:08,665][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.003482808591797948, 'train/grad_norm_pre_clip': 0.7196987271308899, 'timing/data': 0.023021413013339043, 'timing/model': 1.645551798865199, 'learning_rate/qwen_vl_interface': 9.468989306432738e-07, 'learning_rate/action_model': 9.468989306432738e-06, 'epoch': 1.67}) +[2026-06-16 07:41:14,862][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.007569984998553991, 'train/grad_norm_pre_clip': 0.9982863068580627, 'timing/data': 0.022402053698897362, 'timing/model': 1.5234887916594744, 'learning_rate/qwen_vl_interface': 9.456102088861702e-07, 'learning_rate/action_model': 9.456102088861701e-06, 'epoch': 1.67}) +[2026-06-16 07:41:20,976][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.01709696836769581, 'train/grad_norm_pre_clip': 1.6289494037628174, 'timing/data': 0.020493699237704277, 'timing/model': 1.5316862259060144, 'learning_rate/qwen_vl_interface': 9.443232565011188e-07, 'learning_rate/action_model': 9.443232565011187e-06, 'epoch': 1.67}) +[2026-06-16 07:41:27,277][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.020454823970794678, 'train/grad_norm_pre_clip': 1.9607964754104614, 'timing/data': 0.024371908977627754, 'timing/model': 1.5479406919330359, 'learning_rate/qwen_vl_interface': 9.430380740171369e-07, 'learning_rate/action_model': 9.430380740171368e-06, 'epoch': 1.67}) +[2026-06-16 07:41:33,876][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.0025035885628312826, 'train/grad_norm_pre_clip': 0.3049183785915375, 'timing/data': 0.022026427090168, 'timing/model': 1.544830972328782, 'learning_rate/qwen_vl_interface': 9.417546619625143e-07, 'learning_rate/action_model': 9.417546619625143e-06, 'epoch': 1.67}) +[2026-06-16 07:41:40,209][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.00890292413532734, 'train/grad_norm_pre_clip': 1.6003175973892212, 'timing/data': 0.02336651086807251, 'timing/model': 1.771068288013339, 'learning_rate/qwen_vl_interface': 9.404730208648144e-07, 'learning_rate/action_model': 9.404730208648144e-06, 'epoch': 1.67}) +[2026-06-16 07:41:46,270][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.053943682461977005, 'train/grad_norm_pre_clip': 3.2368178367614746, 'timing/data': 0.02220129407942295, 'timing/model': 1.5277728978544474, 'learning_rate/qwen_vl_interface': 9.3919315125087e-07, 'learning_rate/action_model': 9.3919315125087e-06, 'epoch': 1.67}) +[2026-06-16 07:41:52,462][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.1409188210964203, 'train/grad_norm_pre_clip': 6.837255001068115, 'timing/data': 0.022145388647913933, 'timing/model': 1.550012907013297, 'learning_rate/qwen_vl_interface': 9.379150536467879e-07, 'learning_rate/action_model': 9.379150536467878e-06, 'epoch': 1.67}) +[2026-06-16 07:41:58,408][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.004612289369106293, 'train/grad_norm_pre_clip': 0.6172312498092651, 'timing/data': 0.017678240314126015, 'timing/model': 1.447238627821207, 'learning_rate/qwen_vl_interface': 9.366387285779453e-07, 'learning_rate/action_model': 9.366387285779453e-06, 'epoch': 1.67}) +[2026-06-16 07:42:04,701][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.054834168404340744, 'train/grad_norm_pre_clip': 8.385246276855469, 'timing/data': 0.02536538802087307, 'timing/model': 1.5463640950620174, 'learning_rate/qwen_vl_interface': 9.353641765689915e-07, 'learning_rate/action_model': 9.353641765689915e-06, 'epoch': 1.67}) +[2026-06-16 07:42:10,965][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.005255572497844696, 'train/grad_norm_pre_clip': 0.9781049489974976, 'timing/data': 0.02247472107410431, 'timing/model': 1.520181730389595, 'learning_rate/qwen_vl_interface': 9.340913981438462e-07, 'learning_rate/action_model': 9.340913981438463e-06, 'epoch': 1.68}) +[2026-06-16 07:42:17,303][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.04203120619058609, 'train/grad_norm_pre_clip': 13.179033279418945, 'timing/data': 0.04030889272689819, 'timing/model': 1.8398156315088272, 'learning_rate/qwen_vl_interface': 9.328203938257019e-07, 'learning_rate/action_model': 9.328203938257018e-06, 'epoch': 1.68}) +[2026-06-16 07:42:23,361][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.0038612126372754574, 'train/grad_norm_pre_clip': 0.9660305976867676, 'timing/data': 0.024303872138261795, 'timing/model': 1.5154185108840466, 'learning_rate/qwen_vl_interface': 9.315511641370191e-07, 'learning_rate/action_model': 9.315511641370191e-06, 'epoch': 1.68}) +[2026-06-16 07:42:29,563][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.17046169936656952, 'train/grad_norm_pre_clip': 13.397237777709961, 'timing/data': 0.02244597114622593, 'timing/model': 1.5098494160920382, 'learning_rate/qwen_vl_interface': 9.302837095995306e-07, 'learning_rate/action_model': 9.302837095995306e-06, 'epoch': 1.68}) +[2026-06-16 07:42:35,665][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.0025543593801558018, 'train/grad_norm_pre_clip': 0.4275188446044922, 'timing/data': 0.02119750902056694, 'timing/model': 1.546959163621068, 'learning_rate/qwen_vl_interface': 9.290180307342392e-07, 'learning_rate/action_model': 9.290180307342391e-06, 'epoch': 1.68}) +[2026-06-16 07:42:41,962][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 0.00030097734997980297, 'train/grad_norm_pre_clip': 0.08269957453012466, 'timing/data': 0.023047324270009995, 'timing/model': 1.5438440032303333, 'learning_rate/qwen_vl_interface': 9.277541280614183e-07, 'learning_rate/action_model': 9.277541280614183e-06, 'epoch': 1.68}) +[2026-06-16 07:42:48,130][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.1350603848695755, 'train/grad_norm_pre_clip': 17.459327697753906, 'timing/data': 0.021911047399044037, 'timing/model': 1.526088023558259, 'learning_rate/qwen_vl_interface': 9.264920021006093e-07, 'learning_rate/action_model': 9.264920021006093e-06, 'epoch': 1.68}) +[2026-06-16 07:42:54,234][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.09404855966567993, 'train/grad_norm_pre_clip': 12.414050102233887, 'timing/data': 0.023628896102309227, 'timing/model': 1.7428190596401691, 'learning_rate/qwen_vl_interface': 9.252316533706246e-07, 'learning_rate/action_model': 9.252316533706246e-06, 'epoch': 1.68}) +[2026-06-16 07:43:00,274][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.002402821322903037, 'train/grad_norm_pre_clip': 0.45121878385543823, 'timing/data': 0.022227490320801735, 'timing/model': 1.6860443092882633, 'learning_rate/qwen_vl_interface': 9.239730823895469e-07, 'learning_rate/action_model': 9.23973082389547e-06, 'epoch': 1.68}) +[2026-06-16 07:43:06,237][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.0673336312174797, 'train/grad_norm_pre_clip': 15.613382339477539, 'timing/data': 0.02184816263616085, 'timing/model': 1.5466525983065367, 'learning_rate/qwen_vl_interface': 9.227162896747264e-07, 'learning_rate/action_model': 9.227162896747264e-06, 'epoch': 1.68}) +[2026-06-16 07:43:12,333][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.2879156172275543, 'train/grad_norm_pre_clip': 14.672852516174316, 'timing/data': 0.02471485547721386, 'timing/model': 1.5371901895850897, 'learning_rate/qwen_vl_interface': 9.214612757427833e-07, 'learning_rate/action_model': 9.214612757427833e-06, 'epoch': 1.68}) +[2026-06-16 07:43:18,434][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.03503990173339844, 'train/grad_norm_pre_clip': 5.298305034637451, 'timing/data': 0.02121848054230213, 'timing/model': 1.5440966859459877, 'learning_rate/qwen_vl_interface': 9.202080411096071e-07, 'learning_rate/action_model': 9.20208041109607e-06, 'epoch': 1.68}) +[2026-06-16 07:43:24,531][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.04597495123744011, 'train/grad_norm_pre_clip': 16.235597610473633, 'timing/data': 0.02060089074075222, 'timing/model': 1.528688883408904, 'learning_rate/qwen_vl_interface': 9.189565862903543e-07, 'learning_rate/action_model': 9.189565862903543e-06, 'epoch': 1.68}) +[2026-06-16 07:43:30,452][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.06342560052871704, 'train/grad_norm_pre_clip': 16.684913635253906, 'timing/data': 0.017975540831685066, 'timing/model': 1.6301027536392212, 'learning_rate/qwen_vl_interface': 9.17706911799451e-07, 'learning_rate/action_model': 9.17706911799451e-06, 'epoch': 1.68}) +[2026-06-16 07:43:36,258][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.03663654625415802, 'train/grad_norm_pre_clip': 4.326044082641602, 'timing/data': 0.01859981194138527, 'timing/model': 1.6241593770682812, 'learning_rate/qwen_vl_interface': 9.164590181505912e-07, 'learning_rate/action_model': 9.164590181505911e-06, 'epoch': 1.68}) +[2026-06-16 07:43:41,881][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.004498652648180723, 'train/grad_norm_pre_clip': 0.5065910220146179, 'timing/data': 0.02005958929657936, 'timing/model': 1.4432700891047716, 'learning_rate/qwen_vl_interface': 9.15212905856738e-07, 'learning_rate/action_model': 9.15212905856738e-06, 'epoch': 1.68}) +[2026-06-16 07:43:47,680][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.02107292227447033, 'train/grad_norm_pre_clip': 12.90889835357666, 'timing/data': 0.018421173095703125, 'timing/model': 1.4440012462437153, 'learning_rate/qwen_vl_interface': 9.139685754301189e-07, 'learning_rate/action_model': 9.139685754301189e-06, 'epoch': 1.68}) +[2026-06-16 07:43:53,632][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.09011183679103851, 'train/grad_norm_pre_clip': 12.072168350219727, 'timing/data': 0.018514858558773994, 'timing/model': 1.497058155015111, 'learning_rate/qwen_vl_interface': 9.127260273822324e-07, 'learning_rate/action_model': 9.127260273822325e-06, 'epoch': 1.68}) +[2026-06-16 07:43:59,480][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 0.15402264893054962, 'train/grad_norm_pre_clip': 24.245649337768555, 'timing/data': 0.01850116439163685, 'timing/model': 1.4378055892884731, 'learning_rate/qwen_vl_interface': 9.114852622238435e-07, 'learning_rate/action_model': 9.114852622238435e-06, 'epoch': 1.68}) +[2026-06-16 07:44:05,369][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.2157016545534134, 'train/grad_norm_pre_clip': 22.50933265686035, 'timing/data': 0.018239855766296387, 'timing/model': 1.6344220116734505, 'learning_rate/qwen_vl_interface': 9.102462804649827e-07, 'learning_rate/action_model': 9.102462804649827e-06, 'epoch': 1.68}) +[2026-06-16 07:44:10,982][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.22639872133731842, 'train/grad_norm_pre_clip': 9.676301002502441, 'timing/data': 0.019355207681655884, 'timing/model': 1.4455386437475681, 'learning_rate/qwen_vl_interface': 9.090090826149489e-07, 'learning_rate/action_model': 9.09009082614949e-06, 'epoch': 1.68}) +[2026-06-16 07:44:16,942][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.05224020034074783, 'train/grad_norm_pre_clip': 3.206897497177124, 'timing/data': 0.022309722378849983, 'timing/model': 1.4994299244135618, 'learning_rate/qwen_vl_interface': 9.077736691823081e-07, 'learning_rate/action_model': 9.07773669182308e-06, 'epoch': 1.68}) +[2026-06-16 07:44:23,037][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.004636554978787899, 'train/grad_norm_pre_clip': 1.1244369745254517, 'timing/data': 0.022982660681009293, 'timing/model': 1.5436941962689161, 'learning_rate/qwen_vl_interface': 9.065400406748912e-07, 'learning_rate/action_model': 9.06540040674891e-06, 'epoch': 1.68}) +[2026-06-16 07:44:29,133][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.014465691521763802, 'train/grad_norm_pre_clip': 1.6097170114517212, 'timing/data': 0.02262922003865242, 'timing/model': 1.5295742191374302, 'learning_rate/qwen_vl_interface': 9.053081975997969e-07, 'learning_rate/action_model': 9.053081975997968e-06, 'epoch': 1.68}) +[2026-06-16 07:44:35,132][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.11262262612581253, 'train/grad_norm_pre_clip': 12.534394264221191, 'timing/data': 0.022440608590841293, 'timing/model': 1.5046072993427515, 'learning_rate/qwen_vl_interface': 9.040781404633894e-07, 'learning_rate/action_model': 9.040781404633895e-06, 'epoch': 1.68}) +[2026-06-16 07:44:41,187][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 0.06887339800596237, 'train/grad_norm_pre_clip': 22.921567916870117, 'timing/data': 0.0211406908929348, 'timing/model': 1.6964613869786263, 'learning_rate/qwen_vl_interface': 9.028498697712982e-07, 'learning_rate/action_model': 9.028498697712981e-06, 'epoch': 1.68}) +[2026-06-16 07:44:47,131][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.008112376555800438, 'train/grad_norm_pre_clip': 1.2349356412887573, 'timing/data': 0.023580776527523994, 'timing/model': 1.5390857197344303, 'learning_rate/qwen_vl_interface': 9.016233860284195e-07, 'learning_rate/action_model': 9.016233860284195e-06, 'epoch': 1.69}) +[2026-06-16 07:44:53,230][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.18951813876628876, 'train/grad_norm_pre_clip': 10.89180850982666, 'timing/data': 0.021457504481077194, 'timing/model': 1.5420316085219383, 'learning_rate/qwen_vl_interface': 9.003986897389142e-07, 'learning_rate/action_model': 9.003986897389141e-06, 'epoch': 1.69}) +[2026-06-16 07:44:59,335][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.1264152228832245, 'train/grad_norm_pre_clip': 6.032370090484619, 'timing/data': 0.02138504944741726, 'timing/model': 1.5474544614553452, 'learning_rate/qwen_vl_interface': 8.991757814062094e-07, 'learning_rate/action_model': 8.991757814062093e-06, 'epoch': 1.69}) +[2026-06-16 07:45:05,436][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.05651814490556717, 'train/grad_norm_pre_clip': 12.257529258728027, 'timing/data': 0.021697435528039932, 'timing/model': 1.5179157424718142, 'learning_rate/qwen_vl_interface': 8.979546615329954e-07, 'learning_rate/action_model': 8.979546615329953e-06, 'epoch': 1.69}) +[2026-06-16 07:45:11,444][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.008937501348555088, 'train/grad_norm_pre_clip': 1.4357702732086182, 'timing/data': 0.021554600447416306, 'timing/model': 1.477177880704403, 'learning_rate/qwen_vl_interface': 8.967353306212286e-07, 'learning_rate/action_model': 8.967353306212286e-06, 'epoch': 1.69}) +[2026-06-16 07:45:17,552][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.004860307089984417, 'train/grad_norm_pre_clip': 0.8627557754516602, 'timing/data': 0.022793838754296303, 'timing/model': 1.7315752655267715, 'learning_rate/qwen_vl_interface': 8.955177891721303e-07, 'learning_rate/action_model': 8.955177891721303e-06, 'epoch': 1.69}) +[2026-06-16 07:45:23,563][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.01613732799887657, 'train/grad_norm_pre_clip': 2.452298879623413, 'timing/data': 0.022014549002051353, 'timing/model': 1.5410652048885822, 'learning_rate/qwen_vl_interface': 8.943020376861858e-07, 'learning_rate/action_model': 8.943020376861857e-06, 'epoch': 1.69}) +[2026-06-16 07:45:29,766][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.06848472356796265, 'train/grad_norm_pre_clip': 10.05532169342041, 'timing/data': 0.021467555314302444, 'timing/model': 1.5454709865152836, 'learning_rate/qwen_vl_interface': 8.930880766631444e-07, 'learning_rate/action_model': 8.930880766631445e-06, 'epoch': 1.69}) +[2026-06-16 07:45:35,975][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.09037280082702637, 'train/grad_norm_pre_clip': 10.743738174438477, 'timing/data': 0.022878743708133698, 'timing/model': 1.5553926639258862, 'learning_rate/qwen_vl_interface': 8.918759066020201e-07, 'learning_rate/action_model': 8.9187590660202e-06, 'epoch': 1.69}) +[2026-06-16 07:45:42,265][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 0.01858706772327423, 'train/grad_norm_pre_clip': 3.933769941329956, 'timing/data': 0.021730925887823105, 'timing/model': 1.5216044690459967, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 1.69}) +[2026-06-16 07:45:48,480][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.12481964379549026, 'train/grad_norm_pre_clip': 7.435490608215332, 'timing/data': 0.02122756838798523, 'timing/model': 1.7579653467983007, 'learning_rate/qwen_vl_interface': 8.894569413578938e-07, 'learning_rate/action_model': 8.894569413578937e-06, 'epoch': 1.69}) +[2026-06-16 07:45:54,563][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.005240988451987505, 'train/grad_norm_pre_clip': 1.6928119659423828, 'timing/data': 0.02192658931016922, 'timing/model': 1.5390913132578135, 'learning_rate/qwen_vl_interface': 8.882501471692379e-07, 'learning_rate/action_model': 8.882501471692378e-06, 'epoch': 1.69}) +[2026-06-16 07:46:00,764][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.002681440208107233, 'train/grad_norm_pre_clip': 0.7509584426879883, 'timing/data': 0.021739674732089043, 'timing/model': 1.5399266257882118, 'learning_rate/qwen_vl_interface': 8.870451459311894e-07, 'learning_rate/action_model': 8.870451459311893e-06, 'epoch': 1.69}) +[2026-06-16 07:46:06,877][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.06761156022548676, 'train/grad_norm_pre_clip': 16.648000717163086, 'timing/data': 0.022572115063667297, 'timing/model': 1.5543785188347101, 'learning_rate/qwen_vl_interface': 8.85841938139077e-07, 'learning_rate/action_model': 8.85841938139077e-06, 'epoch': 1.69}) +[2026-06-16 07:46:13,175][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.05214284360408783, 'train/grad_norm_pre_clip': 11.619258880615234, 'timing/data': 0.020627832040190697, 'timing/model': 1.5471030063927174, 'learning_rate/qwen_vl_interface': 8.846405242874955e-07, 'learning_rate/action_model': 8.846405242874954e-06, 'epoch': 1.69}) +[2026-06-16 07:46:19,464][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.04285547137260437, 'train/grad_norm_pre_clip': 5.555448055267334, 'timing/data': 0.02124864049255848, 'timing/model': 1.5770034492015839, 'learning_rate/qwen_vl_interface': 8.834409048703004e-07, 'learning_rate/action_model': 8.834409048703005e-06, 'epoch': 1.69}) +[2026-06-16 07:46:25,709][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.015367941930890083, 'train/grad_norm_pre_clip': 3.5713415145874023, 'timing/data': 0.021857112646102905, 'timing/model': 1.7874081209301949, 'learning_rate/qwen_vl_interface': 8.822430803806098e-07, 'learning_rate/action_model': 8.822430803806097e-06, 'epoch': 1.69}) +[2026-06-16 07:46:31,980][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.01061482448130846, 'train/grad_norm_pre_clip': 3.7543575763702393, 'timing/data': 0.021968988701701164, 'timing/model': 1.5600974280387163, 'learning_rate/qwen_vl_interface': 8.810470513108041e-07, 'learning_rate/action_model': 8.81047051310804e-06, 'epoch': 1.69}) +[2026-06-16 07:46:38,264][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 0.29138752818107605, 'train/grad_norm_pre_clip': 25.60112190246582, 'timing/data': 0.022064747288823128, 'timing/model': 1.5442168191075325, 'learning_rate/qwen_vl_interface': 8.798528181525262e-07, 'learning_rate/action_model': 8.798528181525261e-06, 'epoch': 1.69}) +[2026-06-16 07:46:44,476][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.009596512652933598, 'train/grad_norm_pre_clip': 1.443663477897644, 'timing/data': 0.022295096889138222, 'timing/model': 1.5091058909893036, 'learning_rate/qwen_vl_interface': 8.786603813966798e-07, 'learning_rate/action_model': 8.786603813966798e-06, 'epoch': 1.69}) +[2026-06-16 07:46:50,775][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 0.0016629272140562534, 'train/grad_norm_pre_clip': 0.21938671171665192, 'timing/data': 0.025275539606809616, 'timing/model': 1.5488311666995287, 'learning_rate/qwen_vl_interface': 8.774697415334304e-07, 'learning_rate/action_model': 8.774697415334304e-06, 'epoch': 1.69}) +[2026-06-16 07:46:57,065][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.10213887691497803, 'train/grad_norm_pre_clip': 12.125391006469727, 'timing/data': 0.022655978798866272, 'timing/model': 1.5508214104920626, 'learning_rate/qwen_vl_interface': 8.762808990522056e-07, 'learning_rate/action_model': 8.762808990522057e-06, 'epoch': 1.69}) +[2026-06-16 07:47:03,263][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.033267486840486526, 'train/grad_norm_pre_clip': 11.78050422668457, 'timing/data': 0.02160368114709854, 'timing/model': 1.772884238511324, 'learning_rate/qwen_vl_interface': 8.750938544416934e-07, 'learning_rate/action_model': 8.750938544416934e-06, 'epoch': 1.69}) +[2026-06-16 07:47:09,135][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.0013020142214372754, 'train/grad_norm_pre_clip': 0.5580072999000549, 'timing/data': 0.02320292219519615, 'timing/model': 1.5420637894421816, 'learning_rate/qwen_vl_interface': 8.739086081898437e-07, 'learning_rate/action_model': 8.739086081898437e-06, 'epoch': 1.69}) +[2026-06-16 07:47:15,234][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 0.05764801427721977, 'train/grad_norm_pre_clip': 7.261584758758545, 'timing/data': 0.022489221766591072, 'timing/model': 1.5402228124439716, 'learning_rate/qwen_vl_interface': 8.727251607838647e-07, 'learning_rate/action_model': 8.727251607838647e-06, 'epoch': 1.69}) +[2026-06-16 07:47:21,336][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.12846609950065613, 'train/grad_norm_pre_clip': 16.089855194091797, 'timing/data': 0.022660836577415466, 'timing/model': 1.5462474673986435, 'learning_rate/qwen_vl_interface': 8.715435127102281e-07, 'learning_rate/action_model': 8.715435127102281e-06, 'epoch': 1.69}) +[2026-06-16 07:47:27,436][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.09755059331655502, 'train/grad_norm_pre_clip': 8.411651611328125, 'timing/data': 0.023522626608610153, 'timing/model': 1.5428918953984976, 'learning_rate/qwen_vl_interface': 8.703636644546651e-07, 'learning_rate/action_model': 8.703636644546651e-06, 'epoch': 1.7}) +[2026-06-16 07:47:33,936][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.2090938538312912, 'train/grad_norm_pre_clip': 33.94489288330078, 'timing/data': 0.020189249888062477, 'timing/model': 1.6040350962430239, 'learning_rate/qwen_vl_interface': 8.69185616502166e-07, 'learning_rate/action_model': 8.69185616502166e-06, 'epoch': 1.7}) +[2026-06-16 07:47:40,163][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.060917455703020096, 'train/grad_norm_pre_clip': 13.060492515563965, 'timing/data': 0.022253748029470444, 'timing/model': 1.7662845887243748, 'learning_rate/qwen_vl_interface': 8.680093693369821e-07, 'learning_rate/action_model': 8.68009369336982e-06, 'epoch': 1.7}) +[2026-06-16 07:47:46,577][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.10749551653862, 'train/grad_norm_pre_clip': 10.502181053161621, 'timing/data': 0.02308543026447296, 'timing/model': 1.5838358849287033, 'learning_rate/qwen_vl_interface': 8.66834923442624e-07, 'learning_rate/action_model': 8.66834923442624e-06, 'epoch': 1.7}) +[2026-06-16 07:47:52,867][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.004670761991292238, 'train/grad_norm_pre_clip': 0.9775660634040833, 'timing/data': 0.02253098599612713, 'timing/model': 1.5461653806269169, 'learning_rate/qwen_vl_interface': 8.65662279301863e-07, 'learning_rate/action_model': 8.656622793018628e-06, 'epoch': 1.7}) +[2026-06-16 07:47:58,968][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.07688621431589127, 'train/grad_norm_pre_clip': 8.698265075683594, 'timing/data': 0.02195010334253311, 'timing/model': 1.5439463593065739, 'learning_rate/qwen_vl_interface': 8.644914373967274e-07, 'learning_rate/action_model': 8.644914373967273e-06, 'epoch': 1.7}) +[2026-06-16 07:48:05,171][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 0.0368591770529747, 'train/grad_norm_pre_clip': 12.945642471313477, 'timing/data': 0.023274777457118034, 'timing/model': 1.5486285239458084, 'learning_rate/qwen_vl_interface': 8.633223982085068e-07, 'learning_rate/action_model': 8.633223982085068e-06, 'epoch': 1.7}) +[2026-06-16 07:48:11,372][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.04983361437916756, 'train/grad_norm_pre_clip': 17.146400451660156, 'timing/data': 0.022451192140579224, 'timing/model': 1.533704524859786, 'learning_rate/qwen_vl_interface': 8.621551622177499e-07, 'learning_rate/action_model': 8.621551622177499e-06, 'epoch': 1.7}) +[2026-06-16 07:48:17,655][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 0.182262122631073, 'train/grad_norm_pre_clip': 19.567874908447266, 'timing/data': 0.02195197157561779, 'timing/model': 1.75520059466362, 'learning_rate/qwen_vl_interface': 8.609897299042615e-07, 'learning_rate/action_model': 8.609897299042614e-06, 'epoch': 1.7}) +[2026-06-16 07:48:23,544][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.1592564731836319, 'train/grad_norm_pre_clip': 20.663496017456055, 'timing/data': 0.02225811965763569, 'timing/model': 1.5508878398686647, 'learning_rate/qwen_vl_interface': 8.598261017471088e-07, 'learning_rate/action_model': 8.598261017471088e-06, 'epoch': 1.7}) +[2026-06-16 07:48:29,640][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.003266168525442481, 'train/grad_norm_pre_clip': 0.9222447872161865, 'timing/data': 0.021844593808054924, 'timing/model': 1.553119633346796, 'learning_rate/qwen_vl_interface': 8.586642782246133e-07, 'learning_rate/action_model': 8.586642782246131e-06, 'epoch': 1.7}) +[2026-06-16 07:48:35,754][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.27113428711891174, 'train/grad_norm_pre_clip': 12.536836624145508, 'timing/data': 0.021222248673439026, 'timing/model': 1.567654225975275, 'learning_rate/qwen_vl_interface': 8.57504259814358e-07, 'learning_rate/action_model': 8.57504259814358e-06, 'epoch': 1.7}) +[2026-06-16 07:48:41,846][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.03843195363879204, 'train/grad_norm_pre_clip': 26.680747985839844, 'timing/data': 0.021977337077260017, 'timing/model': 1.5576819200068712, 'learning_rate/qwen_vl_interface': 8.563460469931823e-07, 'learning_rate/action_model': 8.563460469931822e-06, 'epoch': 1.7}) +[2026-06-16 07:48:47,964][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.08470011502504349, 'train/grad_norm_pre_clip': 14.163269996643066, 'timing/data': 0.022145992144942284, 'timing/model': 1.5582461226731539, 'learning_rate/qwen_vl_interface': 8.551896402371834e-07, 'learning_rate/action_model': 8.551896402371834e-06, 'epoch': 1.7}) +[2026-06-16 07:48:54,214][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.0010633390629664063, 'train/grad_norm_pre_clip': 0.1407414972782135, 'timing/data': 0.022072700783610344, 'timing/model': 1.787105904892087, 'learning_rate/qwen_vl_interface': 8.540350400217177e-07, 'learning_rate/action_model': 8.540350400217176e-06, 'epoch': 1.7}) +[2026-06-16 07:49:00,281][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 0.031267549842596054, 'train/grad_norm_pre_clip': 10.80803108215332, 'timing/data': 0.021585294976830482, 'timing/model': 1.5625931080430746, 'learning_rate/qwen_vl_interface': 8.528822468213964e-07, 'learning_rate/action_model': 8.528822468213964e-06, 'epoch': 1.7}) +[2026-06-16 07:49:06,604][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.060942355543375015, 'train/grad_norm_pre_clip': 9.05080795288086, 'timing/data': 0.02432352676987648, 'timing/model': 1.5635277349501848, 'learning_rate/qwen_vl_interface': 8.517312611100898e-07, 'learning_rate/action_model': 8.517312611100897e-06, 'epoch': 1.7}) +[2026-06-16 07:49:12,988][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.26022109389305115, 'train/grad_norm_pre_clip': 11.661535263061523, 'timing/data': 0.021759619936347008, 'timing/model': 1.5546111818403006, 'learning_rate/qwen_vl_interface': 8.505820833609249e-07, 'learning_rate/action_model': 8.50582083360925e-06, 'epoch': 1.7}) +[2026-06-16 07:49:19,369][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.020151900127530098, 'train/grad_norm_pre_clip': 3.356160879135132, 'timing/data': 0.021227780729532242, 'timing/model': 1.543528413400054, 'learning_rate/qwen_vl_interface': 8.494347140462847e-07, 'learning_rate/action_model': 8.494347140462846e-06, 'epoch': 1.7}) +[2026-06-16 07:49:25,783][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.03391982242465019, 'train/grad_norm_pre_clip': 9.37876033782959, 'timing/data': 0.020337320864200592, 'timing/model': 1.5902530271559954, 'learning_rate/qwen_vl_interface': 8.482891536378105e-07, 'learning_rate/action_model': 8.482891536378105e-06, 'epoch': 1.7}) +[2026-06-16 07:56:34,103][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.038176119327545166, 'train/grad_norm_pre_clip': 7.043071269989014, 'eval/loss': 0.041674818843603134, 'eval/flap_precision': 0.9571619812583668, 'eval/flap_recall': 0.9701492537313433, 'eval/flap_f1': 0.963611859838275, 'eval/flap_rate_model': 0.2075, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9636118598382749, 'timing/data': 0.0225162822753191, 'timing/model': 1.739309286698699, 'learning_rate/qwen_vl_interface': 8.471454026063972e-07, 'learning_rate/action_model': 8.47145402606397e-06, 'epoch': 1.7}) +[2026-06-16 07:56:34,104][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4400_state +[2026-06-16 07:56:34,109][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 07:56:42,067][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4400_state/model.safetensors +[2026-06-16 07:56:52,914][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4400_state/optimizer.bin +[2026-06-16 07:56:52,914][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4400_state/sampler.bin +[2026-06-16 07:56:52,914][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4400_state/sampler_1.bin +[2026-06-16 07:56:52,915][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4400_state/random_states_0.pkl +[2026-06-16 07:56:53,019][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4400 +[2026-06-16 07:56:53,019][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 07:56:53,028][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 07:56:58,889][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.00108453503344208, 'train/grad_norm_pre_clip': 0.1907634288072586, 'timing/data': 0.018866240978240967, 'timing/model': 1.4520136509090662, 'learning_rate/qwen_vl_interface': 8.460034614221988e-07, 'learning_rate/action_model': 8.460034614221988e-06, 'epoch': 1.7}) +[2026-06-16 07:57:04,781][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.1432318091392517, 'train/grad_norm_pre_clip': 7.766788482666016, 'timing/data': 0.018217042088508606, 'timing/model': 1.6354652401059866, 'learning_rate/qwen_vl_interface': 8.448633305546236e-07, 'learning_rate/action_model': 8.448633305546235e-06, 'epoch': 1.7}) +[2026-06-16 07:57:10,501][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.11767245829105377, 'train/grad_norm_pre_clip': 19.500030517578125, 'timing/data': 0.018571268767118454, 'timing/model': 1.45037885569036, 'learning_rate/qwen_vl_interface': 8.437250104723369e-07, 'learning_rate/action_model': 8.437250104723369e-06, 'epoch': 1.7}) +[2026-06-16 07:57:16,406][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.3038947582244873, 'train/grad_norm_pre_clip': 10.192376136779785, 'timing/data': 0.01994895748794079, 'timing/model': 1.4513709619641304, 'learning_rate/qwen_vl_interface': 8.42588501643259e-07, 'learning_rate/action_model': 8.42588501643259e-06, 'epoch': 1.7}) +[2026-06-16 07:57:22,298][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.019808778539299965, 'train/grad_norm_pre_clip': 4.122622966766357, 'timing/data': 0.019563056528568268, 'timing/model': 1.442467451095581, 'learning_rate/qwen_vl_interface': 8.414538045345655e-07, 'learning_rate/action_model': 8.414538045345654e-06, 'epoch': 1.7}) +[2026-06-16 07:57:28,202][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.0009276880882680416, 'train/grad_norm_pre_clip': 0.2073107361793518, 'timing/data': 0.019331077113747597, 'timing/model': 1.4576081465929747, 'learning_rate/qwen_vl_interface': 8.403209196126871e-07, 'learning_rate/action_model': 8.403209196126871e-06, 'epoch': 1.71}) +[2026-06-16 07:57:34,278][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.04610623046755791, 'train/grad_norm_pre_clip': 5.3818583488464355, 'timing/data': 0.019191622734069824, 'timing/model': 1.8227678090333939, 'learning_rate/qwen_vl_interface': 8.391898473433111e-07, 'learning_rate/action_model': 8.39189847343311e-06, 'epoch': 1.71}) +[2026-06-16 07:57:40,382][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.07228446751832962, 'train/grad_norm_pre_clip': 10.645493507385254, 'timing/data': 0.021224601194262505, 'timing/model': 1.5556297358125448, 'learning_rate/qwen_vl_interface': 8.380605881913778e-07, 'learning_rate/action_model': 8.380605881913779e-06, 'epoch': 1.71}) +[2026-06-16 07:57:46,679][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 0.1057349294424057, 'train/grad_norm_pre_clip': 12.457788467407227, 'timing/data': 0.022225873544812202, 'timing/model': 1.5523349232971668, 'learning_rate/qwen_vl_interface': 8.369331426210832e-07, 'learning_rate/action_model': 8.369331426210832e-06, 'epoch': 1.71}) +[2026-06-16 07:57:52,868][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.000684925529640168, 'train/grad_norm_pre_clip': 0.10463307052850723, 'timing/data': 0.02384783700108528, 'timing/model': 1.5545519683510065, 'learning_rate/qwen_vl_interface': 8.358075110958785e-07, 'learning_rate/action_model': 8.358075110958784e-06, 'epoch': 1.71}) +[2026-06-16 07:57:59,030][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.008775243535637856, 'train/grad_norm_pre_clip': 1.1944704055786133, 'timing/data': 0.021621275693178177, 'timing/model': 1.5480945147573948, 'learning_rate/qwen_vl_interface': 8.346836940784661e-07, 'learning_rate/action_model': 8.346836940784661e-06, 'epoch': 1.71}) +[2026-06-16 07:58:05,129][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.007823231630027294, 'train/grad_norm_pre_clip': 1.3838560581207275, 'timing/data': 0.0226011760532856, 'timing/model': 1.5382164679467678, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 1.71}) +[2026-06-16 07:58:11,245][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.07586299628019333, 'train/grad_norm_pre_clip': 8.378790855407715, 'timing/data': 0.022361116483807564, 'timing/model': 1.7637324947863817, 'learning_rate/qwen_vl_interface': 8.324415054141111e-07, 'learning_rate/action_model': 8.32441505414111e-06, 'epoch': 1.71}) +[2026-06-16 07:58:17,395][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.00810977816581726, 'train/grad_norm_pre_clip': 0.9820513129234314, 'timing/data': 0.021986568346619606, 'timing/model': 1.811064500361681, 'learning_rate/qwen_vl_interface': 8.313231346888476e-07, 'learning_rate/action_model': 8.313231346888476e-06, 'epoch': 1.71}) +[2026-06-16 07:58:23,332][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.0023087216541171074, 'train/grad_norm_pre_clip': 0.45761898159980774, 'timing/data': 0.0223858542740345, 'timing/model': 1.5448949970304966, 'learning_rate/qwen_vl_interface': 8.302065803147353e-07, 'learning_rate/action_model': 8.302065803147353e-06, 'epoch': 1.71}) +[2026-06-16 07:58:29,436][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.009313303045928478, 'train/grad_norm_pre_clip': 1.7207896709442139, 'timing/data': 0.02249941974878311, 'timing/model': 1.5479361284524202, 'learning_rate/qwen_vl_interface': 8.29091842750748e-07, 'learning_rate/action_model': 8.29091842750748e-06, 'epoch': 1.71}) +[2026-06-16 07:58:35,536][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 0.25818008184432983, 'train/grad_norm_pre_clip': 19.14838218688965, 'timing/data': 0.02350255288183689, 'timing/model': 1.5432734079658985, 'learning_rate/qwen_vl_interface': 8.279789224551116e-07, 'learning_rate/action_model': 8.279789224551116e-06, 'epoch': 1.71}) +[2026-06-16 07:58:41,633][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.03735978528857231, 'train/grad_norm_pre_clip': 6.055016040802002, 'timing/data': 0.022085554897785187, 'timing/model': 1.5091822743415833, 'learning_rate/qwen_vl_interface': 8.268678198853058e-07, 'learning_rate/action_model': 8.268678198853059e-06, 'epoch': 1.71}) +[2026-06-16 07:58:47,675][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.04538702964782715, 'train/grad_norm_pre_clip': 8.80931568145752, 'timing/data': 0.021846817806363106, 'timing/model': 1.702508857473731, 'learning_rate/qwen_vl_interface': 8.257585354980628e-07, 'learning_rate/action_model': 8.257585354980628e-06, 'epoch': 1.71}) +[2026-06-16 07:58:53,621][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.15340715646743774, 'train/grad_norm_pre_clip': 10.345588684082031, 'timing/data': 0.022153489291667938, 'timing/model': 1.541240081191063, 'learning_rate/qwen_vl_interface': 8.246510697493681e-07, 'learning_rate/action_model': 8.246510697493682e-06, 'epoch': 1.71}) +[2026-06-16 07:59:00,043][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.006174983456730843, 'train/grad_norm_pre_clip': 0.7378426790237427, 'timing/data': 0.02219209633767605, 'timing/model': 1.5463023390620947, 'learning_rate/qwen_vl_interface': 8.235454230944584e-07, 'learning_rate/action_model': 8.235454230944584e-06, 'epoch': 1.71}) +[2026-06-16 07:59:06,134][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.008774517104029655, 'train/grad_norm_pre_clip': 4.3314528465271, 'timing/data': 0.022701110690832138, 'timing/model': 1.5309521071612835, 'learning_rate/qwen_vl_interface': 8.224415959878234e-07, 'learning_rate/action_model': 8.224415959878234e-06, 'epoch': 1.71}) +[2026-06-16 07:59:12,229][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.034154411405324936, 'train/grad_norm_pre_clip': 5.20531702041626, 'timing/data': 0.022419748827815056, 'timing/model': 1.521714536473155, 'learning_rate/qwen_vl_interface': 8.213395888832052e-07, 'learning_rate/action_model': 8.213395888832051e-06, 'epoch': 1.71}) +[2026-06-16 07:59:18,349][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.006344669498503208, 'train/grad_norm_pre_clip': 0.9633008241653442, 'timing/data': 0.022726086899638176, 'timing/model': 1.516447702422738, 'learning_rate/qwen_vl_interface': 8.202394022335958e-07, 'learning_rate/action_model': 8.202394022335957e-06, 'epoch': 1.71}) +[2026-06-16 07:59:24,574][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 0.10297911614179611, 'train/grad_norm_pre_clip': 10.223687171936035, 'timing/data': 0.023380529135465622, 'timing/model': 1.7302877120673656, 'learning_rate/qwen_vl_interface': 8.191410364912424e-07, 'learning_rate/action_model': 8.191410364912424e-06, 'epoch': 1.71}) +[2026-06-16 07:59:30,675][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.49545446038246155, 'train/grad_norm_pre_clip': 18.577241897583008, 'timing/data': 0.02245483174920082, 'timing/model': 1.529949253425002, 'learning_rate/qwen_vl_interface': 8.1804449210764e-07, 'learning_rate/action_model': 8.1804449210764e-06, 'epoch': 1.71}) +[2026-06-16 07:59:36,975][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.019988469779491425, 'train/grad_norm_pre_clip': 6.945185661315918, 'timing/data': 0.022929081693291664, 'timing/model': 1.52969503775239, 'learning_rate/qwen_vl_interface': 8.169497695335377e-07, 'learning_rate/action_model': 8.169497695335376e-06, 'epoch': 1.71}) +[2026-06-16 07:59:43,281][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 0.03798326849937439, 'train/grad_norm_pre_clip': 5.875532150268555, 'timing/data': 0.022322794422507286, 'timing/model': 1.5446472205221653, 'learning_rate/qwen_vl_interface': 8.158568692189342e-07, 'learning_rate/action_model': 8.158568692189342e-06, 'epoch': 1.71}) +[2026-06-16 07:59:49,603][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.07371305674314499, 'train/grad_norm_pre_clip': 13.60245418548584, 'timing/data': 0.0227511003613472, 'timing/model': 1.5657644513994455, 'learning_rate/qwen_vl_interface': 8.147657916130797e-07, 'learning_rate/action_model': 8.147657916130796e-06, 'epoch': 1.71}) +[2026-06-16 07:59:55,976][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.11404187232255936, 'train/grad_norm_pre_clip': 8.437889099121094, 'timing/data': 0.02275046333670616, 'timing/model': 1.548990236595273, 'learning_rate/qwen_vl_interface': 8.136765371644753e-07, 'learning_rate/action_model': 8.136765371644753e-06, 'epoch': 1.71}) +[2026-06-16 08:00:02,386][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.02462727017700672, 'train/grad_norm_pre_clip': 14.518168449401855, 'timing/data': 0.021841084584593773, 'timing/model': 1.7715272083878517, 'learning_rate/qwen_vl_interface': 8.125891063208713e-07, 'learning_rate/action_model': 8.125891063208712e-06, 'epoch': 1.71}) +[2026-06-16 08:00:08,435][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.11668175458908081, 'train/grad_norm_pre_clip': 8.014472007751465, 'timing/data': 0.02165061980485916, 'timing/model': 1.5254434775561094, 'learning_rate/qwen_vl_interface': 8.115034995292698e-07, 'learning_rate/action_model': 8.115034995292698e-06, 'epoch': 1.72}) +[2026-06-16 08:00:14,534][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.017837420105934143, 'train/grad_norm_pre_clip': 3.992220163345337, 'timing/data': 0.02287915162742138, 'timing/model': 1.5322551243007183, 'learning_rate/qwen_vl_interface': 8.104197172359235e-07, 'learning_rate/action_model': 8.104197172359234e-06, 'epoch': 1.72}) +[2026-06-16 08:00:20,632][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.010711890645325184, 'train/grad_norm_pre_clip': 1.2369729280471802, 'timing/data': 0.02261212095618248, 'timing/model': 1.5292993988841772, 'learning_rate/qwen_vl_interface': 8.093377598863334e-07, 'learning_rate/action_model': 8.093377598863334e-06, 'epoch': 1.72}) +[2026-06-16 08:00:26,632][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.12266264855861664, 'train/grad_norm_pre_clip': 9.737027168273926, 'timing/data': 0.021018516272306442, 'timing/model': 1.500913642346859, 'learning_rate/qwen_vl_interface': 8.082576279252519e-07, 'learning_rate/action_model': 8.082576279252518e-06, 'epoch': 1.72}) +[2026-06-16 08:00:32,759][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.007392521481961012, 'train/grad_norm_pre_clip': 0.9576124548912048, 'timing/data': 0.02302452363073826, 'timing/model': 1.5798224285244942, 'learning_rate/qwen_vl_interface': 8.071793217966803e-07, 'learning_rate/action_model': 8.071793217966802e-06, 'epoch': 1.72}) +[2026-06-16 08:00:38,985][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.0046095154248178005, 'train/grad_norm_pre_clip': 1.1242849826812744, 'timing/data': 0.022180281579494476, 'timing/model': 1.7580357994884253, 'learning_rate/qwen_vl_interface': 8.061028419438691e-07, 'learning_rate/action_model': 8.06102841943869e-06, 'epoch': 1.72}) +[2026-06-16 08:00:45,065][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 0.005127571057528257, 'train/grad_norm_pre_clip': 0.9647186994552612, 'timing/data': 0.021849090233445168, 'timing/model': 1.5448976643383503, 'learning_rate/qwen_vl_interface': 8.050281888093193e-07, 'learning_rate/action_model': 8.050281888093192e-06, 'epoch': 1.72}) +[2026-06-16 08:00:51,009][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.206040158867836, 'train/grad_norm_pre_clip': 11.270092964172363, 'timing/data': 0.020702168345451355, 'timing/model': 1.446439342573285, 'learning_rate/qwen_vl_interface': 8.039553628347794e-07, 'learning_rate/action_model': 8.039553628347793e-06, 'epoch': 1.72}) +[2026-06-16 08:00:56,914][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.005753210745751858, 'train/grad_norm_pre_clip': 1.2276109457015991, 'timing/data': 0.018947672098875046, 'timing/model': 1.4488051012158394, 'learning_rate/qwen_vl_interface': 8.028843644612483e-07, 'learning_rate/action_model': 8.028843644612482e-06, 'epoch': 1.72}) +[2026-06-16 08:01:02,721][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.03201081231236458, 'train/grad_norm_pre_clip': 8.92293930053711, 'timing/data': 0.017658906057476997, 'timing/model': 1.4471541233360767, 'learning_rate/qwen_vl_interface': 8.01815194128973e-07, 'learning_rate/action_model': 8.01815194128973e-06, 'epoch': 1.72}) +[2026-06-16 08:01:08,579][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.03366391733288765, 'train/grad_norm_pre_clip': 23.71088409423828, 'timing/data': 0.01825493387877941, 'timing/model': 1.4569569304585457, 'learning_rate/qwen_vl_interface': 8.007478522774481e-07, 'learning_rate/action_model': 8.00747852277448e-06, 'epoch': 1.72}) +[2026-06-16 08:01:14,467][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.004096712917089462, 'train/grad_norm_pre_clip': 1.3889920711517334, 'timing/data': 0.01873805560171604, 'timing/model': 1.6330166272819042, 'learning_rate/qwen_vl_interface': 7.996823393454197e-07, 'learning_rate/action_model': 7.996823393454196e-06, 'epoch': 1.72}) +[2026-06-16 08:01:20,086][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.052057020366191864, 'train/grad_norm_pre_clip': 19.81862449645996, 'timing/data': 0.01846739836037159, 'timing/model': 1.4490594789385796, 'learning_rate/qwen_vl_interface': 7.986186557708773e-07, 'learning_rate/action_model': 7.986186557708771e-06, 'epoch': 1.72}) +[2026-06-16 08:01:25,982][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.038471661508083344, 'train/grad_norm_pre_clip': 4.72323751449585, 'timing/data': 0.018205657601356506, 'timing/model': 1.4486989732831717, 'learning_rate/qwen_vl_interface': 7.975568019910617e-07, 'learning_rate/action_model': 7.975568019910617e-06, 'epoch': 1.72}) +[2026-06-16 08:01:31,887][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.10431946814060211, 'train/grad_norm_pre_clip': 13.979777336120605, 'timing/data': 0.02300878055393696, 'timing/model': 1.446499440819025, 'learning_rate/qwen_vl_interface': 7.964967784424618e-07, 'learning_rate/action_model': 7.964967784424618e-06, 'epoch': 1.72}) +[2026-06-16 08:01:37,698][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.11222407221794128, 'train/grad_norm_pre_clip': 12.946368217468262, 'timing/data': 0.01755497045814991, 'timing/model': 1.4578801430761814, 'learning_rate/qwen_vl_interface': 7.954385855608129e-07, 'learning_rate/action_model': 7.95438585560813e-06, 'epoch': 1.72}) +[2026-06-16 08:01:43,581][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.00710709486156702, 'train/grad_norm_pre_clip': 1.1636788845062256, 'timing/data': 0.019026288762688637, 'timing/model': 1.4585894271731377, 'learning_rate/qwen_vl_interface': 7.943822237810974e-07, 'learning_rate/action_model': 7.943822237810975e-06, 'epoch': 1.72}) +[2026-06-16 08:01:49,468][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.16977669298648834, 'train/grad_norm_pre_clip': 18.465370178222656, 'timing/data': 0.019855501130223274, 'timing/model': 1.632560571655631, 'learning_rate/qwen_vl_interface': 7.93327693537547e-07, 'learning_rate/action_model': 7.93327693537547e-06, 'epoch': 1.72}) +[2026-06-16 08:01:55,081][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.005258200224488974, 'train/grad_norm_pre_clip': 0.6577907800674438, 'timing/data': 0.018988531082868576, 'timing/model': 1.4452437367290258, 'learning_rate/qwen_vl_interface': 7.92274995263638e-07, 'learning_rate/action_model': 7.92274995263638e-06, 'epoch': 1.72}) +[2026-06-16 08:02:00,886][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.020705755800008774, 'train/grad_norm_pre_clip': 3.3618931770324707, 'timing/data': 0.01841689832508564, 'timing/model': 1.4447008408606052, 'learning_rate/qwen_vl_interface': 7.912241293920958e-07, 'learning_rate/action_model': 7.912241293920959e-06, 'epoch': 1.72}) +[2026-06-16 08:02:06,881][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.011140074580907822, 'train/grad_norm_pre_clip': 1.7661751508712769, 'timing/data': 0.01786419190466404, 'timing/model': 1.4461972061544657, 'learning_rate/qwen_vl_interface': 7.90175096354891e-07, 'learning_rate/action_model': 7.90175096354891e-06, 'epoch': 1.72}) +[2026-06-16 08:02:12,957][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.06566344201564789, 'train/grad_norm_pre_clip': 20.49420928955078, 'timing/data': 0.022212771698832512, 'timing/model': 1.4826066363602877, 'learning_rate/qwen_vl_interface': 7.891278965832431e-07, 'learning_rate/action_model': 7.891278965832432e-06, 'epoch': 1.72}) +[2026-06-16 08:02:18,981][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.021771302446722984, 'train/grad_norm_pre_clip': 3.9472570419311523, 'timing/data': 0.02450575679540634, 'timing/model': 1.4666214063763618, 'learning_rate/qwen_vl_interface': 7.880825305076138e-07, 'learning_rate/action_model': 7.880825305076139e-06, 'epoch': 1.72}) +[2026-06-16 08:02:24,893][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.04521390423178673, 'train/grad_norm_pre_clip': 8.437034606933594, 'timing/data': 0.01871182955801487, 'timing/model': 1.6503200680017471, 'learning_rate/qwen_vl_interface': 7.870389985577148e-07, 'learning_rate/action_model': 7.870389985577148e-06, 'epoch': 1.72}) +[2026-06-16 08:02:30,581][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.09769414365291595, 'train/grad_norm_pre_clip': 15.933066368103027, 'timing/data': 0.018856121227145195, 'timing/model': 1.4448216743767262, 'learning_rate/qwen_vl_interface': 7.859973011625028e-07, 'learning_rate/action_model': 7.859973011625026e-06, 'epoch': 1.72}) +[2026-06-16 08:02:36,381][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.09666396677494049, 'train/grad_norm_pre_clip': 15.225131034851074, 'timing/data': 0.018656199797987938, 'timing/model': 1.444427479058504, 'learning_rate/qwen_vl_interface': 7.849574387501801e-07, 'learning_rate/action_model': 7.8495743875018e-06, 'epoch': 1.72}) +[2026-06-16 08:02:42,181][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.03558824583888054, 'train/grad_norm_pre_clip': 6.334922790527344, 'timing/data': 0.018238116055727005, 'timing/model': 1.4461499527096748, 'learning_rate/qwen_vl_interface': 7.839194117481952e-07, 'learning_rate/action_model': 7.839194117481952e-06, 'epoch': 1.73}) +[2026-06-16 08:02:48,102][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.008544610813260078, 'train/grad_norm_pre_clip': 3.1173453330993652, 'timing/data': 0.01902368664741516, 'timing/model': 1.4523510318249464, 'learning_rate/qwen_vl_interface': 7.828832205832418e-07, 'learning_rate/action_model': 7.828832205832418e-06, 'epoch': 1.73}) +[2026-06-16 08:02:53,986][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.0703524574637413, 'train/grad_norm_pre_clip': 22.783098220825195, 'timing/data': 0.018346795812249184, 'timing/model': 1.4653460271656513, 'learning_rate/qwen_vl_interface': 7.818488656812581e-07, 'learning_rate/action_model': 7.818488656812582e-06, 'epoch': 1.73}) +[2026-06-16 08:02:59,867][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 0.007314331363886595, 'train/grad_norm_pre_clip': 1.500887155532837, 'timing/data': 0.0186535082757473, 'timing/model': 1.631156399846077, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 1.73}) +[2026-06-16 08:03:05,576][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.008897362276911736, 'train/grad_norm_pre_clip': 2.909249782562256, 'timing/data': 0.018854917958378792, 'timing/model': 1.5306079909205437, 'learning_rate/qwen_vl_interface': 7.797856663661836e-07, 'learning_rate/action_model': 7.797856663661835e-06, 'epoch': 1.73}) +[2026-06-16 08:03:11,283][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.016881853342056274, 'train/grad_norm_pre_clip': 4.182898044586182, 'timing/data': 0.018995782360434532, 'timing/model': 1.4433456920087337, 'learning_rate/qwen_vl_interface': 7.787568228011959e-07, 'learning_rate/action_model': 7.78756822801196e-06, 'epoch': 1.73}) +[2026-06-16 08:03:17,085][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.10256671160459518, 'train/grad_norm_pre_clip': 11.700758934020996, 'timing/data': 0.018792714923620224, 'timing/model': 1.446598555892706, 'learning_rate/qwen_vl_interface': 7.777298171953858e-07, 'learning_rate/action_model': 7.777298171953857e-06, 'epoch': 1.73}) +[2026-06-16 08:03:22,983][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.06775218993425369, 'train/grad_norm_pre_clip': 25.306194305419922, 'timing/data': 0.01917550526559353, 'timing/model': 1.4465980604290962, 'learning_rate/qwen_vl_interface': 7.767046499709144e-07, 'learning_rate/action_model': 7.767046499709143e-06, 'epoch': 1.73}) +[2026-06-16 08:03:28,883][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.0031978574115782976, 'train/grad_norm_pre_clip': 0.8426958918571472, 'timing/data': 0.0192946195602417, 'timing/model': 1.4617987740784883, 'learning_rate/qwen_vl_interface': 7.756813215491904e-07, 'learning_rate/action_model': 7.756813215491904e-06, 'epoch': 1.73}) +[2026-06-16 08:03:34,855][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.12463495135307312, 'train/grad_norm_pre_clip': 5.530434608459473, 'timing/data': 0.017866304144263268, 'timing/model': 1.6186848133802414, 'learning_rate/qwen_vl_interface': 7.746598323508655e-07, 'learning_rate/action_model': 7.746598323508655e-06, 'epoch': 1.73}) +[2026-06-16 08:03:40,690][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.0790596604347229, 'train/grad_norm_pre_clip': 17.624042510986328, 'timing/data': 0.018018923699855804, 'timing/model': 1.65162025205791, 'learning_rate/qwen_vl_interface': 7.736401827958351e-07, 'learning_rate/action_model': 7.73640182795835e-06, 'epoch': 1.73}) +[2026-06-16 08:03:46,386][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 0.22214274108409882, 'train/grad_norm_pre_clip': 11.94648265838623, 'timing/data': 0.019287968054413795, 'timing/model': 1.4424521699547768, 'learning_rate/qwen_vl_interface': 7.726223733032382e-07, 'learning_rate/action_model': 7.72622373303238e-06, 'epoch': 1.73}) +[2026-06-16 08:03:52,289][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.028498297557234764, 'train/grad_norm_pre_clip': 5.836219787597656, 'timing/data': 0.018391883000731468, 'timing/model': 1.4421953037381172, 'learning_rate/qwen_vl_interface': 7.716064042914584e-07, 'learning_rate/action_model': 7.716064042914584e-06, 'epoch': 1.73}) +[2026-06-16 08:03:58,441][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.03658600151538849, 'train/grad_norm_pre_clip': 8.818406105041504, 'timing/data': 0.022856025025248528, 'timing/model': 1.5470983311533928, 'learning_rate/qwen_vl_interface': 7.705922761781226e-07, 'learning_rate/action_model': 7.705922761781226e-06, 'epoch': 1.73}) +[2026-06-16 08:04:04,536][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.0327603779733181, 'train/grad_norm_pre_clip': 6.113483905792236, 'timing/data': 0.022967003285884857, 'timing/model': 1.5206255242228508, 'learning_rate/qwen_vl_interface': 7.695799893800996e-07, 'learning_rate/action_model': 7.695799893800996e-06, 'epoch': 1.73}) +[2026-06-16 08:04:10,659][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.005553443916141987, 'train/grad_norm_pre_clip': 1.6955592632293701, 'timing/data': 0.02524786815047264, 'timing/model': 1.7640007194131613, 'learning_rate/qwen_vl_interface': 7.685695443135028e-07, 'learning_rate/action_model': 7.685695443135027e-06, 'epoch': 1.73}) +[2026-06-16 08:04:16,740][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 0.19549039006233215, 'train/grad_norm_pre_clip': 16.756160736083984, 'timing/data': 0.023366672918200493, 'timing/model': 1.5461685489863157, 'learning_rate/qwen_vl_interface': 7.675609413936888e-07, 'learning_rate/action_model': 7.675609413936888e-06, 'epoch': 1.73}) +[2026-06-16 08:04:22,838][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.19319139420986176, 'train/grad_norm_pre_clip': 28.56626319885254, 'timing/data': 0.022876473143696785, 'timing/model': 1.5422090776264668, 'learning_rate/qwen_vl_interface': 7.665541810352541e-07, 'learning_rate/action_model': 7.66554181035254e-06, 'epoch': 1.73}) +[2026-06-16 08:04:28,938][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.08046698570251465, 'train/grad_norm_pre_clip': 5.7980146408081055, 'timing/data': 0.023012252524495125, 'timing/model': 1.5409273598343134, 'learning_rate/qwen_vl_interface': 7.655492636520415e-07, 'learning_rate/action_model': 7.655492636520414e-06, 'epoch': 1.73}) +[2026-06-16 08:04:35,038][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.04748211055994034, 'train/grad_norm_pre_clip': 10.994370460510254, 'timing/data': 0.02258477546274662, 'timing/model': 1.5013875365257263, 'learning_rate/qwen_vl_interface': 7.64546189657134e-07, 'learning_rate/action_model': 7.645461896571339e-06, 'epoch': 1.73}) +[2026-06-16 08:04:41,038][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.00709637301042676, 'train/grad_norm_pre_clip': 2.098513603210449, 'timing/data': 0.021073544397950172, 'timing/model': 1.5015471875667572, 'learning_rate/qwen_vl_interface': 7.635449594628571e-07, 'learning_rate/action_model': 7.635449594628571e-06, 'epoch': 1.73}) +[2026-06-16 08:04:47,101][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.020356561988592148, 'train/grad_norm_pre_clip': 5.1080145835876465, 'timing/data': 0.02197205275297165, 'timing/model': 1.7071597129106522, 'learning_rate/qwen_vl_interface': 7.625455734807803e-07, 'learning_rate/action_model': 7.625455734807803e-06, 'epoch': 1.73}) +[2026-06-16 08:04:53,037][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.025910640135407448, 'train/grad_norm_pre_clip': 5.663922309875488, 'timing/data': 0.021644089370965958, 'timing/model': 1.5427297670394182, 'learning_rate/qwen_vl_interface': 7.615480321217116e-07, 'learning_rate/action_model': 7.615480321217117e-06, 'epoch': 1.73}) +[2026-06-16 08:04:59,145][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.15648822486400604, 'train/grad_norm_pre_clip': 10.580340385437012, 'timing/data': 0.023978158831596375, 'timing/model': 1.54881089925766, 'learning_rate/qwen_vl_interface': 7.605523357957028e-07, 'learning_rate/action_model': 7.605523357957028e-06, 'epoch': 1.73}) +[2026-06-16 08:05:05,235][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.17852847278118134, 'train/grad_norm_pre_clip': 10.603564262390137, 'timing/data': 0.02063879929482937, 'timing/model': 1.5399810448288918, 'learning_rate/qwen_vl_interface': 7.59558484912048e-07, 'learning_rate/action_model': 7.59558484912048e-06, 'epoch': 1.73}) +[2026-06-16 08:05:11,640][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.15419605374336243, 'train/grad_norm_pre_clip': 13.919139862060547, 'timing/data': 0.025796489790081978, 'timing/model': 1.5248595084995031, 'learning_rate/qwen_vl_interface': 7.585664798792807e-07, 'learning_rate/action_model': 7.585664798792807e-06, 'epoch': 1.74}) +[2026-06-16 08:05:17,842][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.1505592316389084, 'train/grad_norm_pre_clip': 5.793468952178955, 'timing/data': 0.024139702320098877, 'timing/model': 1.6454213727265596, 'learning_rate/qwen_vl_interface': 7.575763211051779e-07, 'learning_rate/action_model': 7.575763211051778e-06, 'epoch': 1.74}) +[2026-06-16 08:05:24,361][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.18129409849643707, 'train/grad_norm_pre_clip': 8.298473358154297, 'timing/data': 0.02116822451353073, 'timing/model': 1.7275393959134817, 'learning_rate/qwen_vl_interface': 7.565880089967553e-07, 'learning_rate/action_model': 7.565880089967553e-06, 'epoch': 1.74}) +[2026-06-16 08:05:30,519][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.05165676772594452, 'train/grad_norm_pre_clip': 10.183835983276367, 'timing/data': 0.027726346626877785, 'timing/model': 1.6406985353678465, 'learning_rate/qwen_vl_interface': 7.556015439602725e-07, 'learning_rate/action_model': 7.556015439602725e-06, 'epoch': 1.74}) +[2026-06-16 08:05:36,965][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.10082804411649704, 'train/grad_norm_pre_clip': 16.740190505981445, 'timing/data': 0.02143944799900055, 'timing/model': 1.5505082327872515, 'learning_rate/qwen_vl_interface': 7.546169264012252e-07, 'learning_rate/action_model': 7.5461692640122525e-06, 'epoch': 1.74}) +[2026-06-16 08:05:43,264][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.03342636302113533, 'train/grad_norm_pre_clip': 6.169662952423096, 'timing/data': 0.03172084502875805, 'timing/model': 1.5279129184782505, 'learning_rate/qwen_vl_interface': 7.536341567243544e-07, 'learning_rate/action_model': 7.536341567243543e-06, 'epoch': 1.74}) +[2026-06-16 08:05:49,690][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.0019172544125467539, 'train/grad_norm_pre_clip': 0.44381460547447205, 'timing/data': 0.04511777125298977, 'timing/model': 1.5811620652675629, 'learning_rate/qwen_vl_interface': 7.526532353336395e-07, 'learning_rate/action_model': 7.526532353336395e-06, 'epoch': 1.74}) +[2026-06-16 08:05:55,944][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.10150294005870819, 'train/grad_norm_pre_clip': 13.539313316345215, 'timing/data': 0.03158131428062916, 'timing/model': 1.7456254810094833, 'learning_rate/qwen_vl_interface': 7.516741626322994e-07, 'learning_rate/action_model': 7.516741626322994e-06, 'epoch': 1.74}) +[2026-06-16 08:06:01,743][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.023103341460227966, 'train/grad_norm_pre_clip': 5.134233474731445, 'timing/data': 0.03150578774511814, 'timing/model': 1.5184086710214615, 'learning_rate/qwen_vl_interface': 7.506969390227945e-07, 'learning_rate/action_model': 7.506969390227945e-06, 'epoch': 1.74}) +[2026-06-16 08:06:07,989][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.05177346616983414, 'train/grad_norm_pre_clip': 7.611845970153809, 'timing/data': 0.0343317985534668, 'timing/model': 1.5311629474163055, 'learning_rate/qwen_vl_interface': 7.497215649068244e-07, 'learning_rate/action_model': 7.497215649068244e-06, 'epoch': 1.74}) +[2026-06-16 08:06:14,180][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.10451675206422806, 'train/grad_norm_pre_clip': 7.467231273651123, 'timing/data': 0.03234330751001835, 'timing/model': 1.5430125780403614, 'learning_rate/qwen_vl_interface': 7.48748040685329e-07, 'learning_rate/action_model': 7.487480406853289e-06, 'epoch': 1.74}) +[2026-06-16 08:06:20,363][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.02143247053027153, 'train/grad_norm_pre_clip': 7.495148658752441, 'timing/data': 0.03203357569873333, 'timing/model': 1.5300018582493067, 'learning_rate/qwen_vl_interface': 7.477763667584867e-07, 'learning_rate/action_model': 7.477763667584866e-06, 'epoch': 1.74}) +[2026-06-16 08:06:26,568][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.00282203103415668, 'train/grad_norm_pre_clip': 0.44572019577026367, 'timing/data': 0.030365094542503357, 'timing/model': 1.5611604005098343, 'learning_rate/qwen_vl_interface': 7.468065435257159e-07, 'learning_rate/action_model': 7.468065435257159e-06, 'epoch': 1.74}) +[2026-06-16 08:06:32,793][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.12200465798377991, 'train/grad_norm_pre_clip': 17.635499954223633, 'timing/data': 0.031765835359692574, 'timing/model': 1.7575295008718967, 'learning_rate/qwen_vl_interface': 7.458385713856745e-07, 'learning_rate/action_model': 7.458385713856744e-06, 'epoch': 1.74}) +[2026-06-16 08:06:38,871][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.03877795487642288, 'train/grad_norm_pre_clip': 6.202877998352051, 'timing/data': 0.03180459886789322, 'timing/model': 1.5417960565537214, 'learning_rate/qwen_vl_interface': 7.448724507362605e-07, 'learning_rate/action_model': 7.448724507362605e-06, 'epoch': 1.74}) +[2026-06-16 08:06:45,134][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.0037143398076295853, 'train/grad_norm_pre_clip': 1.2820602655410767, 'timing/data': 0.03267952427268028, 'timing/model': 1.5359057802706957, 'learning_rate/qwen_vl_interface': 7.439081819746078e-07, 'learning_rate/action_model': 7.4390818197460775e-06, 'epoch': 1.74}) +[2026-06-16 08:06:51,235][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.024859026074409485, 'train/grad_norm_pre_clip': 9.806403160095215, 'timing/data': 0.03191451542079449, 'timing/model': 1.5344175435602665, 'learning_rate/qwen_vl_interface': 7.42945765497091e-07, 'learning_rate/action_model': 7.42945765497091e-06, 'epoch': 1.74}) +[2026-06-16 08:14:03,338][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.11511050164699554, 'train/grad_norm_pre_clip': 7.774658679962158, 'eval/loss': 0.042317140847444534, 'eval/flap_precision': 0.9557046979865772, 'eval/flap_recall': 0.966078697421981, 'eval/flap_f1': 0.9608636977058029, 'eval/flap_rate_model': 0.20694444444444443, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9608636977058028, 'timing/data': 0.03109085187315941, 'timing/model': 1.5373590551316738, 'learning_rate/qwen_vl_interface': 7.41985201699324e-07, 'learning_rate/action_model': 7.41985201699324e-06, 'epoch': 1.74}) +[2026-06-16 08:14:03,339][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4500_state +[2026-06-16 08:14:03,343][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 08:14:09,275][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4500_state/model.safetensors +[2026-06-16 08:14:19,604][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4500_state/optimizer.bin +[2026-06-16 08:14:19,605][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4500_state/sampler.bin +[2026-06-16 08:14:19,605][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4500_state/sampler_1.bin +[2026-06-16 08:14:19,606][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4500_state/random_states_0.pkl +[2026-06-16 08:14:19,705][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-06-16 08:14:19,705][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 08:14:19,714][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 08:14:25,845][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.004445851780474186, 'train/grad_norm_pre_clip': 0.9630623459815979, 'timing/data': 0.03254491649568081, 'timing/model': 1.534250408411026, 'learning_rate/qwen_vl_interface': 7.410264909761581e-07, 'learning_rate/action_model': 7.41026490976158e-06, 'epoch': 1.74}) +[2026-06-16 08:14:32,047][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.024355819448828697, 'train/grad_norm_pre_clip': 4.40286922454834, 'timing/data': 0.030490489676594734, 'timing/model': 1.5222092382609844, 'learning_rate/qwen_vl_interface': 7.40069633721683e-07, 'learning_rate/action_model': 7.400696337216829e-06, 'epoch': 1.74}) +[2026-06-16 08:14:38,147][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.0371914878487587, 'train/grad_norm_pre_clip': 9.441892623901367, 'timing/data': 0.031870605424046516, 'timing/model': 1.5504320599138737, 'learning_rate/qwen_vl_interface': 7.391146303292262e-07, 'learning_rate/action_model': 7.391146303292262e-06, 'epoch': 1.74}) +[2026-06-16 08:14:44,333][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.006950678303837776, 'train/grad_norm_pre_clip': 0.8229415416717529, 'timing/data': 0.031142234802246094, 'timing/model': 1.7130396086722612, 'learning_rate/qwen_vl_interface': 7.381614811913538e-07, 'learning_rate/action_model': 7.381614811913537e-06, 'epoch': 1.74}) +[2026-06-16 08:14:50,352][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.03400309383869171, 'train/grad_norm_pre_clip': 18.385068893432617, 'timing/data': 0.03121856600046158, 'timing/model': 1.531266102567315, 'learning_rate/qwen_vl_interface': 7.372101866998693e-07, 'learning_rate/action_model': 7.372101866998693e-06, 'epoch': 1.74}) +[2026-06-16 08:14:56,554][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.009933373890817165, 'train/grad_norm_pre_clip': 2.973933458328247, 'timing/data': 0.03139008767902851, 'timing/model': 1.5350266192108393, 'learning_rate/qwen_vl_interface': 7.362607472458147e-07, 'learning_rate/action_model': 7.362607472458147e-06, 'epoch': 1.74}) +[2026-06-16 08:15:02,757][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.2184477150440216, 'train/grad_norm_pre_clip': 20.469358444213867, 'timing/data': 0.030848108232021332, 'timing/model': 1.533907551318407, 'learning_rate/qwen_vl_interface': 7.353131632194682e-07, 'learning_rate/action_model': 7.3531316321946816e-06, 'epoch': 1.74}) +[2026-06-16 08:15:08,954][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.0230389591306448, 'train/grad_norm_pre_clip': 16.07562828063965, 'timing/data': 0.03162823244929314, 'timing/model': 1.5238210447132587, 'learning_rate/qwen_vl_interface': 7.343674350103455e-07, 'learning_rate/action_model': 7.343674350103455e-06, 'epoch': 1.74}) +[2026-06-16 08:15:15,249][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.005450710188597441, 'train/grad_norm_pre_clip': 1.2323931455612183, 'timing/data': 0.03098638914525509, 'timing/model': 1.6529024504125118, 'learning_rate/qwen_vl_interface': 7.334235630071995e-07, 'learning_rate/action_model': 7.334235630071995e-06, 'epoch': 1.75}) +[2026-06-16 08:15:21,483][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.1607203483581543, 'train/grad_norm_pre_clip': 11.134358406066895, 'timing/data': 0.030514515936374664, 'timing/model': 1.7589051239192486, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 1.75}) +[2026-06-16 08:15:27,558][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 0.2630214989185333, 'train/grad_norm_pre_clip': 7.552671432495117, 'timing/data': 0.03116758167743683, 'timing/model': 1.5311553236097097, 'learning_rate/qwen_vl_interface': 7.315413891700363e-07, 'learning_rate/action_model': 7.315413891700362e-06, 'epoch': 1.75}) +[2026-06-16 08:15:33,873][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.24785318970680237, 'train/grad_norm_pre_clip': 53.24689483642578, 'timing/data': 0.033697811886668205, 'timing/model': 1.5300633516162634, 'learning_rate/qwen_vl_interface': 7.306030881097092e-07, 'learning_rate/action_model': 7.3060308810970925e-06, 'epoch': 1.75}) +[2026-06-16 08:15:40,260][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.35976386070251465, 'train/grad_norm_pre_clip': 32.4979133605957, 'timing/data': 0.03328937664628029, 'timing/model': 1.5390031542629004, 'learning_rate/qwen_vl_interface': 7.296666448027406e-07, 'learning_rate/action_model': 7.296666448027406e-06, 'epoch': 1.75}) +[2026-06-16 08:15:46,459][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.0007643250864930451, 'train/grad_norm_pre_clip': 0.1209903284907341, 'timing/data': 0.03233429230749607, 'timing/model': 1.5276770126074553, 'learning_rate/qwen_vl_interface': 7.287320596340664e-07, 'learning_rate/action_model': 7.287320596340663e-06, 'epoch': 1.75}) +[2026-06-16 08:15:52,661][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.02756681479513645, 'train/grad_norm_pre_clip': 12.63235855102539, 'timing/data': 0.03212357684969902, 'timing/model': 1.5396910160779953, 'learning_rate/qwen_vl_interface': 7.277993329878593e-07, 'learning_rate/action_model': 7.277993329878594e-06, 'epoch': 1.75}) +[2026-06-16 08:15:58,882][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.03208201378583908, 'train/grad_norm_pre_clip': 13.669533729553223, 'timing/data': 0.03147582709789276, 'timing/model': 1.758796188980341, 'learning_rate/qwen_vl_interface': 7.268684652475277e-07, 'learning_rate/action_model': 7.268684652475277e-06, 'epoch': 1.75}) +[2026-06-16 08:16:04,935][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.023655934259295464, 'train/grad_norm_pre_clip': 12.688719749450684, 'timing/data': 0.03151947259902954, 'timing/model': 1.5071283169090748, 'learning_rate/qwen_vl_interface': 7.259394567957169e-07, 'learning_rate/action_model': 7.259394567957169e-06, 'epoch': 1.75}) +[2026-06-16 08:16:10,929][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.005638163071125746, 'train/grad_norm_pre_clip': 1.0210808515548706, 'timing/data': 0.031203830614686012, 'timing/model': 1.5210488233715296, 'learning_rate/qwen_vl_interface': 7.250123080143066e-07, 'learning_rate/action_model': 7.2501230801430665e-06, 'epoch': 1.75}) +[2026-06-16 08:16:17,031][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.16240178048610687, 'train/grad_norm_pre_clip': 11.958695411682129, 'timing/data': 0.030454708263278008, 'timing/model': 1.537384184077382, 'learning_rate/qwen_vl_interface': 7.240870192844127e-07, 'learning_rate/action_model': 7.240870192844127e-06, 'epoch': 1.75}) +[2026-06-16 08:16:22,775][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.02219315618276596, 'train/grad_norm_pre_clip': 8.15693187713623, 'timing/data': 0.026287203654646873, 'timing/model': 1.4389213752001524, 'learning_rate/qwen_vl_interface': 7.231635909863858e-07, 'learning_rate/action_model': 7.2316359098638585e-06, 'epoch': 1.75}) +[2026-06-16 08:16:28,578][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 0.000384190323529765, 'train/grad_norm_pre_clip': 0.05776849389076233, 'timing/data': 0.02574121206998825, 'timing/model': 1.4650457240641117, 'learning_rate/qwen_vl_interface': 7.222420234998138e-07, 'learning_rate/action_model': 7.2224202349981385e-06, 'epoch': 1.75}) +[2026-06-16 08:16:34,461][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.016082001850008965, 'train/grad_norm_pre_clip': 3.716491937637329, 'timing/data': 0.026439402252435684, 'timing/model': 1.6230682265013456, 'learning_rate/qwen_vl_interface': 7.213223172035175e-07, 'learning_rate/action_model': 7.213223172035174e-06, 'epoch': 1.75}) +[2026-06-16 08:16:40,079][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.0802074521780014, 'train/grad_norm_pre_clip': 6.806365013122559, 'timing/data': 0.02664176933467388, 'timing/model': 1.4385925475507975, 'learning_rate/qwen_vl_interface': 7.204044724755528e-07, 'learning_rate/action_model': 7.204044724755528e-06, 'epoch': 1.75}) +[2026-06-16 08:16:46,335][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.11328572034835815, 'train/grad_norm_pre_clip': 11.982280731201172, 'timing/data': 0.029902826994657516, 'timing/model': 1.5340061206370592, 'learning_rate/qwen_vl_interface': 7.194884896932132e-07, 'learning_rate/action_model': 7.194884896932132e-06, 'epoch': 1.75}) +[2026-06-16 08:16:52,331][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.012208622880280018, 'train/grad_norm_pre_clip': 4.038143634796143, 'timing/data': 0.03244406543672085, 'timing/model': 1.522102426737547, 'learning_rate/qwen_vl_interface': 7.185743692330221e-07, 'learning_rate/action_model': 7.185743692330221e-06, 'epoch': 1.75}) +[2026-06-16 08:16:58,335][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 0.060712460428476334, 'train/grad_norm_pre_clip': 14.31347942352295, 'timing/data': 0.03116506151854992, 'timing/model': 1.5367195457220078, 'learning_rate/qwen_vl_interface': 7.176621114707408e-07, 'learning_rate/action_model': 7.176621114707407e-06, 'epoch': 1.75}) +[2026-06-16 08:17:04,180][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.11438024789094925, 'train/grad_norm_pre_clip': 12.561811447143555, 'timing/data': 0.025079883635044098, 'timing/model': 1.4518267344683409, 'learning_rate/qwen_vl_interface': 7.167517167813642e-07, 'learning_rate/action_model': 7.167517167813642e-06, 'epoch': 1.75}) +[2026-06-16 08:17:10,074][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.32053372263908386, 'train/grad_norm_pre_clip': 19.65535545349121, 'timing/data': 0.026449060067534447, 'timing/model': 1.6291837133467197, 'learning_rate/qwen_vl_interface': 7.15843185539121e-07, 'learning_rate/action_model': 7.15843185539121e-06, 'epoch': 1.75}) +[2026-06-16 08:17:15,835][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.006156535353511572, 'train/grad_norm_pre_clip': 0.6695771217346191, 'timing/data': 0.02690701372921467, 'timing/model': 1.5629188492894173, 'learning_rate/qwen_vl_interface': 7.149365181174739e-07, 'learning_rate/action_model': 7.149365181174739e-06, 'epoch': 1.75}) +[2026-06-16 08:17:22,058][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.07706888020038605, 'train/grad_norm_pre_clip': 21.560653686523438, 'timing/data': 0.039633505046367645, 'timing/model': 1.7440270520746708, 'learning_rate/qwen_vl_interface': 7.14031714889121e-07, 'learning_rate/action_model': 7.14031714889121e-06, 'epoch': 1.75}) +[2026-06-16 08:17:28,362][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.01264764554798603, 'train/grad_norm_pre_clip': 1.8784230947494507, 'timing/data': 0.0330544114112854, 'timing/model': 1.5233406368643045, 'learning_rate/qwen_vl_interface': 7.131287762259907e-07, 'learning_rate/action_model': 7.131287762259907e-06, 'epoch': 1.75}) +[2026-06-16 08:17:34,562][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.16466686129570007, 'train/grad_norm_pre_clip': 10.174210548400879, 'timing/data': 0.031634699553251266, 'timing/model': 1.5316565334796906, 'learning_rate/qwen_vl_interface': 7.12227702499248e-07, 'learning_rate/action_model': 7.12227702499248e-06, 'epoch': 1.75}) +[2026-06-16 08:17:41,063][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.03562743589282036, 'train/grad_norm_pre_clip': 7.088682174682617, 'timing/data': 0.031637076288461685, 'timing/model': 1.581994378939271, 'learning_rate/qwen_vl_interface': 7.113284940792903e-07, 'learning_rate/action_model': 7.113284940792903e-06, 'epoch': 1.75}) +[2026-06-16 08:17:47,281][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.1583474576473236, 'train/grad_norm_pre_clip': 13.186044692993164, 'timing/data': 0.0316519308835268, 'timing/model': 1.750509625300765, 'learning_rate/qwen_vl_interface': 7.104311513357476e-07, 'learning_rate/action_model': 7.1043115133574755e-06, 'epoch': 1.75}) +[2026-06-16 08:17:53,364][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.05205304175615311, 'train/grad_norm_pre_clip': 9.750934600830078, 'timing/data': 0.031158247962594032, 'timing/model': 1.5353999324142933, 'learning_rate/qwen_vl_interface': 7.095356746374845e-07, 'learning_rate/action_model': 7.095356746374844e-06, 'epoch': 1.76}) +[2026-06-16 08:17:59,563][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.10729032754898071, 'train/grad_norm_pre_clip': 44.3579216003418, 'timing/data': 0.030732885003089905, 'timing/model': 1.5371236559003592, 'learning_rate/qwen_vl_interface': 7.086420643525973e-07, 'learning_rate/action_model': 7.086420643525972e-06, 'epoch': 1.76}) +[2026-06-16 08:18:05,663][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.017209060490131378, 'train/grad_norm_pre_clip': 4.56267786026001, 'timing/data': 0.03153925761580467, 'timing/model': 1.5338590648025274, 'learning_rate/qwen_vl_interface': 7.077503208484151e-07, 'learning_rate/action_model': 7.0775032084841515e-06, 'epoch': 1.76}) +[2026-06-16 08:18:11,867][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.026639070361852646, 'train/grad_norm_pre_clip': 9.9590425491333, 'timing/data': 0.03248344361782074, 'timing/model': 1.5389905516058207, 'learning_rate/qwen_vl_interface': 7.068604444915003e-07, 'learning_rate/action_model': 7.068604444915003e-06, 'epoch': 1.76}) +[2026-06-16 08:18:18,266][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.05602037161588669, 'train/grad_norm_pre_clip': 9.07106876373291, 'timing/data': 0.03185122087597847, 'timing/model': 1.5198984034359455, 'learning_rate/qwen_vl_interface': 7.059724356476475e-07, 'learning_rate/action_model': 7.059724356476474e-06, 'epoch': 1.76}) +[2026-06-16 08:18:24,485][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.007751115132123232, 'train/grad_norm_pre_clip': 0.7134749293327332, 'timing/data': 0.032304875552654266, 'timing/model': 1.7572989575564861, 'learning_rate/qwen_vl_interface': 7.050862946818836e-07, 'learning_rate/action_model': 7.050862946818836e-06, 'epoch': 1.76}) +[2026-06-16 08:18:30,702][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.005435039754956961, 'train/grad_norm_pre_clip': 0.951626718044281, 'timing/data': 0.031851304695010185, 'timing/model': 1.6349524278193712, 'learning_rate/qwen_vl_interface': 7.042020219584673e-07, 'learning_rate/action_model': 7.042020219584673e-06, 'epoch': 1.76}) +[2026-06-16 08:18:36,874][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.021720217540860176, 'train/grad_norm_pre_clip': 4.56591796875, 'timing/data': 0.02979021705687046, 'timing/model': 1.5284488257020712, 'learning_rate/qwen_vl_interface': 7.0331961784089e-07, 'learning_rate/action_model': 7.033196178408899e-06, 'epoch': 1.76}) +[2026-06-16 08:18:43,068][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.0061553106643259525, 'train/grad_norm_pre_clip': 1.9225205183029175, 'timing/data': 0.03204579837620258, 'timing/model': 1.533625353127718, 'learning_rate/qwen_vl_interface': 7.024390826918748e-07, 'learning_rate/action_model': 7.024390826918747e-06, 'epoch': 1.76}) +[2026-06-16 08:18:49,262][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.04673586040735245, 'train/grad_norm_pre_clip': 11.785916328430176, 'timing/data': 0.031336596235632896, 'timing/model': 1.5356515794992447, 'learning_rate/qwen_vl_interface': 7.01560416873376e-07, 'learning_rate/action_model': 7.015604168733759e-06, 'epoch': 1.76}) +[2026-06-16 08:18:55,462][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.06508701294660568, 'train/grad_norm_pre_clip': 20.829044342041016, 'timing/data': 0.03162714093923569, 'timing/model': 1.5195319429039955, 'learning_rate/qwen_vl_interface': 7.006836207465803e-07, 'learning_rate/action_model': 7.006836207465803e-06, 'epoch': 1.76}) +[2026-06-16 08:19:01,677][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 0.2569703459739685, 'train/grad_norm_pre_clip': 3.8001277446746826, 'timing/data': 0.047128526493906975, 'timing/model': 1.7526436112821102, 'learning_rate/qwen_vl_interface': 6.998086946719058e-07, 'learning_rate/action_model': 6.998086946719057e-06, 'epoch': 1.76}) +[2026-06-16 08:19:07,800][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.007133403327316046, 'train/grad_norm_pre_clip': 2.834641695022583, 'timing/data': 0.032959455624222755, 'timing/model': 1.7007044330239296, 'learning_rate/qwen_vl_interface': 6.989356390090008e-07, 'learning_rate/action_model': 6.989356390090007e-06, 'epoch': 1.76}) +[2026-06-16 08:19:13,764][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.0034961162600666285, 'train/grad_norm_pre_clip': 1.5256747007369995, 'timing/data': 0.03161695785820484, 'timing/model': 1.5488606486469507, 'learning_rate/qwen_vl_interface': 6.980644541167459e-07, 'learning_rate/action_model': 6.980644541167458e-06, 'epoch': 1.76}) +[2026-06-16 08:19:19,962][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.020282965153455734, 'train/grad_norm_pre_clip': 4.645607948303223, 'timing/data': 0.03188890963792801, 'timing/model': 1.5331220477819443, 'learning_rate/qwen_vl_interface': 6.971951403532524e-07, 'learning_rate/action_model': 6.971951403532523e-06, 'epoch': 1.76}) +[2026-06-16 08:19:26,161][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.0005911768530495465, 'train/grad_norm_pre_clip': 0.2041197270154953, 'timing/data': 0.045129936188459396, 'timing/model': 1.5165409091860056, 'learning_rate/qwen_vl_interface': 6.963276980758622e-07, 'learning_rate/action_model': 6.963276980758622e-06, 'epoch': 1.76}) +[2026-06-16 08:19:32,461][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.0022384668700397015, 'train/grad_norm_pre_clip': 0.2909011244773865, 'timing/data': 0.04064717888832092, 'timing/model': 1.5707348082214594, 'learning_rate/qwen_vl_interface': 6.954621276411485e-07, 'learning_rate/action_model': 6.954621276411484e-06, 'epoch': 1.76}) +[2026-06-16 08:19:38,691][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 0.14853589236736298, 'train/grad_norm_pre_clip': 7.329405307769775, 'timing/data': 0.03641894832253456, 'timing/model': 1.747199360281229, 'learning_rate/qwen_vl_interface': 6.945984294049136e-07, 'learning_rate/action_model': 6.9459842940491364e-06, 'epoch': 1.76}) +[2026-06-16 08:19:44,764][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.0722932443022728, 'train/grad_norm_pre_clip': 27.183515548706055, 'timing/data': 0.032033566385507584, 'timing/model': 1.534629039466381, 'learning_rate/qwen_vl_interface': 6.937366037221919e-07, 'learning_rate/action_model': 6.937366037221918e-06, 'epoch': 1.76}) +[2026-06-16 08:19:50,969][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.14189088344573975, 'train/grad_norm_pre_clip': 20.138858795166016, 'timing/data': 0.03284516744315624, 'timing/model': 1.5390901025384665, 'learning_rate/qwen_vl_interface': 6.928766509472471e-07, 'learning_rate/action_model': 6.92876650947247e-06, 'epoch': 1.76}) +[2026-06-16 08:19:57,260][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 0.33665332198143005, 'train/grad_norm_pre_clip': 18.577238082885742, 'timing/data': 0.03073262795805931, 'timing/model': 1.533187149092555, 'learning_rate/qwen_vl_interface': 6.920185714335737e-07, 'learning_rate/action_model': 6.9201857143357375e-06, 'epoch': 1.76}) +[2026-06-16 08:20:03,462][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.1345861256122589, 'train/grad_norm_pre_clip': 6.611974239349365, 'timing/data': 0.030833611264824867, 'timing/model': 1.5211708545684814, 'learning_rate/qwen_vl_interface': 6.911623655338959e-07, 'learning_rate/action_model': 6.911623655338959e-06, 'epoch': 1.76}) +[2026-06-16 08:20:09,977][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.009791792370378971, 'train/grad_norm_pre_clip': 4.463946342468262, 'timing/data': 0.03158383443951607, 'timing/model': 1.5933632887899876, 'learning_rate/qwen_vl_interface': 6.903080336001666e-07, 'learning_rate/action_model': 6.903080336001666e-06, 'epoch': 1.76}) +[2026-06-16 08:20:16,527][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.07550203800201416, 'train/grad_norm_pre_clip': 15.272758483886719, 'timing/data': 0.16984783299267292, 'timing/model': 1.8141107112169266, 'learning_rate/qwen_vl_interface': 6.894555759835707e-07, 'learning_rate/action_model': 6.894555759835707e-06, 'epoch': 1.76}) +[2026-06-16 08:20:22,773][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.006513826549053192, 'train/grad_norm_pre_clip': 0.9495323300361633, 'timing/data': 0.034807536751031876, 'timing/model': 1.539945349097252, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 1.76}) +[2026-06-16 08:20:29,172][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.09778152406215668, 'train/grad_norm_pre_clip': 12.58834171295166, 'timing/data': 0.03438577428460121, 'timing/model': 1.5362516157329082, 'learning_rate/qwen_vl_interface': 6.877562851026584e-07, 'learning_rate/action_model': 6.877562851026584e-06, 'epoch': 1.76}) +[2026-06-16 08:20:35,474][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.01235877349972725, 'train/grad_norm_pre_clip': 3.7143235206604004, 'timing/data': 0.03205156698822975, 'timing/model': 1.5328000150620937, 'learning_rate/qwen_vl_interface': 6.869094525368572e-07, 'learning_rate/action_model': 6.8690945253685715e-06, 'epoch': 1.77}) +[2026-06-16 08:20:42,169][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.011722282506525517, 'train/grad_norm_pre_clip': 1.9857293367385864, 'timing/data': 0.03160353749990463, 'timing/model': 1.507119532674551, 'learning_rate/qwen_vl_interface': 6.860644956852168e-07, 'learning_rate/action_model': 6.860644956852168e-06, 'epoch': 1.77}) +[2026-06-16 08:20:48,358][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.1918209195137024, 'train/grad_norm_pre_clip': 20.208518981933594, 'timing/data': 0.03138147480785847, 'timing/model': 1.509599981829524, 'learning_rate/qwen_vl_interface': 6.852214148950667e-07, 'learning_rate/action_model': 6.852214148950667e-06, 'epoch': 1.77}) +[2026-06-16 08:20:54,537][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.0756959617137909, 'train/grad_norm_pre_clip': 11.856739044189453, 'timing/data': 0.032993972301483154, 'timing/model': 1.6968582887202501, 'learning_rate/qwen_vl_interface': 6.843802105129667e-07, 'learning_rate/action_model': 6.843802105129666e-06, 'epoch': 1.77}) +[2026-06-16 08:21:00,764][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.001057891990058124, 'train/grad_norm_pre_clip': 0.2533011734485626, 'timing/data': 0.032190386205911636, 'timing/model': 1.5419168509542942, 'learning_rate/qwen_vl_interface': 6.835408828847035e-07, 'learning_rate/action_model': 6.8354088288470356e-06, 'epoch': 1.77}) +[2026-06-16 08:21:06,962][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.07727047801017761, 'train/grad_norm_pre_clip': 7.139344215393066, 'timing/data': 0.031959397718310356, 'timing/model': 1.535612653940916, 'learning_rate/qwen_vl_interface': 6.827034323552929e-07, 'learning_rate/action_model': 6.827034323552929e-06, 'epoch': 1.77}) +[2026-06-16 08:21:13,163][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.010609767399728298, 'train/grad_norm_pre_clip': 1.8670786619186401, 'timing/data': 0.03204051963984966, 'timing/model': 1.5364293418824673, 'learning_rate/qwen_vl_interface': 6.818678592689794e-07, 'learning_rate/action_model': 6.818678592689794e-06, 'epoch': 1.77}) +[2026-06-16 08:21:19,356][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.005646688863635063, 'train/grad_norm_pre_clip': 2.069606304168701, 'timing/data': 0.03532896563410759, 'timing/model': 1.5022748317569494, 'learning_rate/qwen_vl_interface': 6.810341639692357e-07, 'learning_rate/action_model': 6.810341639692357e-06, 'epoch': 1.77}) +[2026-06-16 08:21:25,563][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 0.0043158759362995625, 'train/grad_norm_pre_clip': 0.6444789171218872, 'timing/data': 0.03238155320286751, 'timing/model': 1.7332660369575024, 'learning_rate/qwen_vl_interface': 6.802023467987631e-07, 'learning_rate/action_model': 6.802023467987631e-06, 'epoch': 1.77}) +[2026-06-16 08:21:31,560][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.01740829460322857, 'train/grad_norm_pre_clip': 8.43302059173584, 'timing/data': 0.03257192671298981, 'timing/model': 1.5333511400967836, 'learning_rate/qwen_vl_interface': 6.793724080994884e-07, 'learning_rate/action_model': 6.793724080994884e-06, 'epoch': 1.77}) +[2026-06-16 08:21:37,731][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.08810823410749435, 'train/grad_norm_pre_clip': 10.206706047058105, 'timing/data': 0.03218343108892441, 'timing/model': 1.5357172340154648, 'learning_rate/qwen_vl_interface': 6.785443482125693e-07, 'learning_rate/action_model': 6.7854434821256915e-06, 'epoch': 1.77}) +[2026-06-16 08:21:43,475][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.010737386532127857, 'train/grad_norm_pre_clip': 1.288490653038025, 'timing/data': 0.02694150060415268, 'timing/model': 1.4377092495560646, 'learning_rate/qwen_vl_interface': 6.777181674783895e-07, 'learning_rate/action_model': 6.777181674783895e-06, 'epoch': 1.77}) +[2026-06-16 08:21:49,178][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.05153997614979744, 'train/grad_norm_pre_clip': 16.61506462097168, 'timing/data': 0.026266086846590042, 'timing/model': 1.440009968355298, 'learning_rate/qwen_vl_interface': 6.76893866236561e-07, 'learning_rate/action_model': 6.7689386623656094e-06, 'epoch': 1.77}) +[2026-06-16 08:21:54,980][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.002606512513011694, 'train/grad_norm_pre_clip': 0.573553204536438, 'timing/data': 0.026303356513381004, 'timing/model': 1.4140343256294727, 'learning_rate/qwen_vl_interface': 6.76071444825923e-07, 'learning_rate/action_model': 6.76071444825923e-06, 'epoch': 1.77}) +[2026-06-16 08:22:01,131][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 0.045070748776197433, 'train/grad_norm_pre_clip': 12.757713317871094, 'timing/data': 0.025965619832277298, 'timing/model': 1.830967815592885, 'learning_rate/qwen_vl_interface': 6.752509035845413e-07, 'learning_rate/action_model': 6.7525090358454126e-06, 'epoch': 1.77}) +[2026-06-16 08:22:06,894][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.0006524922209791839, 'train/grad_norm_pre_clip': 0.1525614857673645, 'timing/data': 0.028568485751748085, 'timing/model': 1.5107941906899214, 'learning_rate/qwen_vl_interface': 6.744322428497095e-07, 'learning_rate/action_model': 6.744322428497095e-06, 'epoch': 1.77}) +[2026-06-16 08:22:12,901][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.007904554717242718, 'train/grad_norm_pre_clip': 1.5066773891448975, 'timing/data': 0.02601782977581024, 'timing/model': 1.5210434067994356, 'learning_rate/qwen_vl_interface': 6.736154629579487e-07, 'learning_rate/action_model': 6.736154629579487e-06, 'epoch': 1.77}) +[2026-06-16 08:22:18,820][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.017959043383598328, 'train/grad_norm_pre_clip': 3.5950381755828857, 'timing/data': 0.02543884888291359, 'timing/model': 1.4418468605726957, 'learning_rate/qwen_vl_interface': 6.728005642450055e-07, 'learning_rate/action_model': 6.7280056424500554e-06, 'epoch': 1.77}) +[2026-06-16 08:22:24,960][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 0.0004758866271004081, 'train/grad_norm_pre_clip': 0.06590979546308517, 'timing/data': 0.029102357104420662, 'timing/model': 1.5850374232977629, 'learning_rate/qwen_vl_interface': 6.719875470458553e-07, 'learning_rate/action_model': 6.719875470458553e-06, 'epoch': 1.77}) +[2026-06-16 08:22:31,179][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.10575851798057556, 'train/grad_norm_pre_clip': 10.347488403320312, 'timing/data': 0.029598545283079147, 'timing/model': 1.5744378082454205, 'learning_rate/qwen_vl_interface': 6.711764116946972e-07, 'learning_rate/action_model': 6.7117641169469715e-06, 'epoch': 1.77}) +[2026-06-16 08:22:37,571][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.10326044261455536, 'train/grad_norm_pre_clip': 10.013916969299316, 'timing/data': 0.031289564445614815, 'timing/model': 1.746344255283475, 'learning_rate/qwen_vl_interface': 6.703671585249587e-07, 'learning_rate/action_model': 6.703671585249586e-06, 'epoch': 1.77}) +[2026-06-16 08:22:43,563][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.024876903742551804, 'train/grad_norm_pre_clip': 14.525317192077637, 'timing/data': 0.030940834432840347, 'timing/model': 1.530107956379652, 'learning_rate/qwen_vl_interface': 6.69559787869294e-07, 'learning_rate/action_model': 6.69559787869294e-06, 'epoch': 1.77}) +[2026-06-16 08:22:49,763][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.0028127371333539486, 'train/grad_norm_pre_clip': 0.494863361120224, 'timing/data': 0.03212844394147396, 'timing/model': 1.5246163979172707, 'learning_rate/qwen_vl_interface': 6.68754300059582e-07, 'learning_rate/action_model': 6.687543000595819e-06, 'epoch': 1.77}) +[2026-06-16 08:22:55,931][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.018043547868728638, 'train/grad_norm_pre_clip': 16.999832153320312, 'timing/data': 0.03209063783288002, 'timing/model': 1.547442253679037, 'learning_rate/qwen_vl_interface': 6.67950695426929e-07, 'learning_rate/action_model': 6.679506954269291e-06, 'epoch': 1.77}) +[2026-06-16 08:23:02,029][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.25639182329177856, 'train/grad_norm_pre_clip': 15.173460960388184, 'timing/data': 0.03188764117658138, 'timing/model': 1.5194667130708694, 'learning_rate/qwen_vl_interface': 6.671489743016651e-07, 'learning_rate/action_model': 6.671489743016651e-06, 'epoch': 1.77}) +[2026-06-16 08:23:08,133][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.06629447638988495, 'train/grad_norm_pre_clip': 6.524111270904541, 'timing/data': 0.03153346851468086, 'timing/model': 1.5486272368580103, 'learning_rate/qwen_vl_interface': 6.663491370133487e-07, 'learning_rate/action_model': 6.663491370133487e-06, 'epoch': 1.77}) +[2026-06-16 08:23:14,459][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.020916184410452843, 'train/grad_norm_pre_clip': 4.3696794509887695, 'timing/data': 0.029726848006248474, 'timing/model': 1.9600749127566814, 'learning_rate/qwen_vl_interface': 6.655511838907629e-07, 'learning_rate/action_model': 6.655511838907629e-06, 'epoch': 1.78}) +[2026-06-16 08:23:20,329][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.003797621699050069, 'train/grad_norm_pre_clip': 1.2837899923324585, 'timing/data': 0.03246252425014973, 'timing/model': 1.5166744012385607, 'learning_rate/qwen_vl_interface': 6.647551152619149e-07, 'learning_rate/action_model': 6.647551152619148e-06, 'epoch': 1.78}) +[2026-06-16 08:23:26,675][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.06293344497680664, 'train/grad_norm_pre_clip': 20.4963436126709, 'timing/data': 0.03494033217430115, 'timing/model': 1.5341482255607843, 'learning_rate/qwen_vl_interface': 6.639609314540396e-07, 'learning_rate/action_model': 6.639609314540396e-06, 'epoch': 1.78}) +[2026-06-16 08:23:33,016][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.07158473879098892, 'train/grad_norm_pre_clip': 15.085222244262695, 'timing/data': 0.033523401245474815, 'timing/model': 1.5705320294946432, 'learning_rate/qwen_vl_interface': 6.631686327935954e-07, 'learning_rate/action_model': 6.631686327935954e-06, 'epoch': 1.78}) +[2026-06-16 08:23:39,163][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.0047769430093467236, 'train/grad_norm_pre_clip': 4.173465251922607, 'timing/data': 0.032222842797636986, 'timing/model': 1.5393326170742512, 'learning_rate/qwen_vl_interface': 6.623782196062665e-07, 'learning_rate/action_model': 6.623782196062664e-06, 'epoch': 1.78}) +[2026-06-16 08:23:45,359][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.11081009358167648, 'train/grad_norm_pre_clip': 11.007452964782715, 'timing/data': 0.031750598922371864, 'timing/model': 1.5200500674545765, 'learning_rate/qwen_vl_interface': 6.615896922169622e-07, 'learning_rate/action_model': 6.615896922169622e-06, 'epoch': 1.78}) +[2026-06-16 08:23:51,580][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.0789971873164177, 'train/grad_norm_pre_clip': 11.361373901367188, 'timing/data': 0.03242109529674053, 'timing/model': 1.7451645638793707, 'learning_rate/qwen_vl_interface': 6.608030509498156e-07, 'learning_rate/action_model': 6.608030509498156e-06, 'epoch': 1.78}) +[2026-06-16 08:23:57,668][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.05998310074210167, 'train/grad_norm_pre_clip': 5.48374605178833, 'timing/data': 0.031866271048784256, 'timing/model': 1.5391315966844559, 'learning_rate/qwen_vl_interface': 6.600182961281865e-07, 'learning_rate/action_model': 6.600182961281864e-06, 'epoch': 1.78}) +[2026-06-16 08:24:03,858][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.03921873867511749, 'train/grad_norm_pre_clip': 13.28465461730957, 'timing/data': 0.030703343451023102, 'timing/model': 1.5281698573380709, 'learning_rate/qwen_vl_interface': 6.592354280746561e-07, 'learning_rate/action_model': 6.5923542807465595e-06, 'epoch': 1.78}) +[2026-06-16 08:24:09,967][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 0.008465316146612167, 'train/grad_norm_pre_clip': 2.0096755027770996, 'timing/data': 0.03130628727376461, 'timing/model': 1.5395579524338245, 'learning_rate/qwen_vl_interface': 6.584544471110335e-07, 'learning_rate/action_model': 6.584544471110335e-06, 'epoch': 1.78}) +[2026-06-16 08:24:16,163][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.01460692472755909, 'train/grad_norm_pre_clip': 8.696215629577637, 'timing/data': 0.03129695542156696, 'timing/model': 1.5038300566375256, 'learning_rate/qwen_vl_interface': 6.576753535583495e-07, 'learning_rate/action_model': 6.576753535583495e-06, 'epoch': 1.78}) +[2026-06-16 08:24:22,463][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.008696298114955425, 'train/grad_norm_pre_clip': 0.6955997943878174, 'timing/data': 0.03181313909590244, 'timing/model': 1.5194506831467152, 'learning_rate/qwen_vl_interface': 6.5689814773686e-07, 'learning_rate/action_model': 6.568981477368599e-06, 'epoch': 1.78}) +[2026-06-16 08:24:28,534][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.08816909044981003, 'train/grad_norm_pre_clip': 10.1821870803833, 'timing/data': 0.0311625674366951, 'timing/model': 1.7293955441564322, 'learning_rate/qwen_vl_interface': 6.56122829966046e-07, 'learning_rate/action_model': 6.56122829966046e-06, 'epoch': 1.78}) +[2026-06-16 08:31:38,339][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.005581401288509369, 'train/grad_norm_pre_clip': 3.0229265689849854, 'eval/loss': 0.043226517736911774, 'eval/flap_precision': 0.9578804347826086, 'eval/flap_recall': 0.9565807327001357, 'eval/flap_f1': 0.9572301425661914, 'eval/flap_rate_model': 0.20444444444444446, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9572301425661913, 'timing/data': 0.03114854358136654, 'timing/model': 1.5232782438397408, 'learning_rate/qwen_vl_interface': 6.553494005646101e-07, 'learning_rate/action_model': 6.553494005646101e-06, 'epoch': 1.78}) +[2026-06-16 08:31:38,339][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4600_state +[2026-06-16 08:31:38,343][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 08:31:44,334][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4600_state/model.safetensors +[2026-06-16 08:31:53,508][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4600_state/optimizer.bin +[2026-06-16 08:31:53,508][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4600_state/sampler.bin +[2026-06-16 08:31:53,509][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4600_state/sampler_1.bin +[2026-06-16 08:31:53,509][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4600_state/random_states_0.pkl +[2026-06-16 08:31:53,615][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4600 +[2026-06-16 08:31:53,615][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 08:31:53,623][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 08:31:59,490][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.09951842576265335, 'train/grad_norm_pre_clip': 15.397851943969727, 'timing/data': 0.026638595387339592, 'timing/model': 1.6348921041935682, 'learning_rate/qwen_vl_interface': 6.545778598504807e-07, 'learning_rate/action_model': 6.545778598504807e-06, 'epoch': 1.78}) +[2026-06-16 08:32:05,207][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.07129483669996262, 'train/grad_norm_pre_clip': 17.844146728515625, 'timing/data': 0.024987252429127693, 'timing/model': 1.4505210518836975, 'learning_rate/qwen_vl_interface': 6.538082081408093e-07, 'learning_rate/action_model': 6.538082081408093e-06, 'epoch': 1.78}) +[2026-06-16 08:32:11,093][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 0.027465397492051125, 'train/grad_norm_pre_clip': 10.5276517868042, 'timing/data': 0.024893920868635178, 'timing/model': 1.43741038069129, 'learning_rate/qwen_vl_interface': 6.530404457519697e-07, 'learning_rate/action_model': 6.530404457519697e-06, 'epoch': 1.78}) +[2026-06-16 08:32:17,312][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.00886473711580038, 'train/grad_norm_pre_clip': 1.0325801372528076, 'timing/data': 0.026527300477027893, 'timing/model': 1.6140207089483738, 'learning_rate/qwen_vl_interface': 6.522745729995614e-07, 'learning_rate/action_model': 6.522745729995613e-06, 'epoch': 1.78}) +[2026-06-16 08:32:23,477][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.009404012002050877, 'train/grad_norm_pre_clip': 3.134629964828491, 'timing/data': 0.03492443822324276, 'timing/model': 1.5356520898640156, 'learning_rate/qwen_vl_interface': 6.515105901984037e-07, 'learning_rate/action_model': 6.515105901984037e-06, 'epoch': 1.78}) +[2026-06-16 08:32:29,750][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.0028091727290302515, 'train/grad_norm_pre_clip': 1.1475166082382202, 'timing/data': 0.03414931707084179, 'timing/model': 1.5082662869244814, 'learning_rate/qwen_vl_interface': 6.50748497662543e-07, 'learning_rate/action_model': 6.507484976625429e-06, 'epoch': 1.78}) +[2026-06-16 08:32:35,837][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.033834025263786316, 'train/grad_norm_pre_clip': 25.080665588378906, 'timing/data': 0.032179445028305054, 'timing/model': 1.737271811813116, 'learning_rate/qwen_vl_interface': 6.499882957052459e-07, 'learning_rate/action_model': 6.4998829570524586e-06, 'epoch': 1.78}) +[2026-06-16 08:32:41,728][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.007112117484211922, 'train/grad_norm_pre_clip': 1.373923897743225, 'timing/data': 0.029993252828717232, 'timing/model': 1.5289149638265371, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 1.78}) +[2026-06-16 08:32:47,828][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.0989394560456276, 'train/grad_norm_pre_clip': 7.170047283172607, 'timing/data': 0.030198244377970695, 'timing/model': 1.5355463791638613, 'learning_rate/qwen_vl_interface': 6.484735647755259e-07, 'learning_rate/action_model': 6.48473564775526e-06, 'epoch': 1.78}) +[2026-06-16 08:32:53,930][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.002710111439228058, 'train/grad_norm_pre_clip': 0.7191554307937622, 'timing/data': 0.03116508759558201, 'timing/model': 1.5351084154099226, 'learning_rate/qwen_vl_interface': 6.477190364257524e-07, 'learning_rate/action_model': 6.477190364257524e-06, 'epoch': 1.78}) +[2026-06-16 08:33:00,155][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.05616173520684242, 'train/grad_norm_pre_clip': 29.34347152709961, 'timing/data': 0.029545633122324944, 'timing/model': 1.6441943366080523, 'learning_rate/qwen_vl_interface': 6.469663998998393e-07, 'learning_rate/action_model': 6.469663998998393e-06, 'epoch': 1.78}) +[2026-06-16 08:33:06,261][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.1626186966896057, 'train/grad_norm_pre_clip': 19.557214736938477, 'timing/data': 0.030253373086452484, 'timing/model': 1.739198373630643, 'learning_rate/qwen_vl_interface': 6.462156555071681e-07, 'learning_rate/action_model': 6.46215655507168e-06, 'epoch': 1.78}) +[2026-06-16 08:33:12,142][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 0.004998267628252506, 'train/grad_norm_pre_clip': 1.9088597297668457, 'timing/data': 0.030859461054205894, 'timing/model': 1.5293856374919415, 'learning_rate/qwen_vl_interface': 6.454668035563401e-07, 'learning_rate/action_model': 6.454668035563401e-06, 'epoch': 1.79}) +[2026-06-16 08:33:18,245][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.00398523174226284, 'train/grad_norm_pre_clip': 0.706997811794281, 'timing/data': 0.03054799884557724, 'timing/model': 1.5476512350142002, 'learning_rate/qwen_vl_interface': 6.447198443551809e-07, 'learning_rate/action_model': 6.4471984435518075e-06, 'epoch': 1.79}) +[2026-06-16 08:33:24,395][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.010367592796683311, 'train/grad_norm_pre_clip': 3.4375505447387695, 'timing/data': 0.031783679500222206, 'timing/model': 1.5442567635327578, 'learning_rate/qwen_vl_interface': 6.439747782107373e-07, 'learning_rate/action_model': 6.4397477821073735e-06, 'epoch': 1.79}) +[2026-06-16 08:33:30,673][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 0.1610061377286911, 'train/grad_norm_pre_clip': 12.783346176147461, 'timing/data': 0.031403591856360435, 'timing/model': 1.539105324074626, 'learning_rate/qwen_vl_interface': 6.432316054292779e-07, 'learning_rate/action_model': 6.43231605429278e-06, 'epoch': 1.79}) +[2026-06-16 08:33:37,078][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.03139563649892807, 'train/grad_norm_pre_clip': 32.61660385131836, 'timing/data': 0.030552877113223076, 'timing/model': 1.5928076077252626, 'learning_rate/qwen_vl_interface': 6.424903263162924e-07, 'learning_rate/action_model': 6.424903263162923e-06, 'epoch': 1.79}) +[2026-06-16 08:33:43,374][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.07451887428760529, 'train/grad_norm_pre_clip': 10.590195655822754, 'timing/data': 0.03169330768287182, 'timing/model': 1.7407292984426022, 'learning_rate/qwen_vl_interface': 6.417509411764924e-07, 'learning_rate/action_model': 6.4175094117649244e-06, 'epoch': 1.79}) +[2026-06-16 08:33:49,683][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.0014401547377929091, 'train/grad_norm_pre_clip': 0.3738773465156555, 'timing/data': 0.03230556473135948, 'timing/model': 1.7504382878541946, 'learning_rate/qwen_vl_interface': 6.410134503138122e-07, 'learning_rate/action_model': 6.4101345031381224e-06, 'epoch': 1.79}) +[2026-06-16 08:33:56,069][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.03685392066836357, 'train/grad_norm_pre_clip': 12.053524017333984, 'timing/data': 0.03340156748890877, 'timing/model': 1.5342974718660116, 'learning_rate/qwen_vl_interface': 6.402778540314054e-07, 'learning_rate/action_model': 6.402778540314054e-06, 'epoch': 1.79}) +[2026-06-16 08:34:02,485][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.0023498849477618933, 'train/grad_norm_pre_clip': 0.9002461433410645, 'timing/data': 0.03789816237986088, 'timing/model': 1.7233039997518063, 'learning_rate/qwen_vl_interface': 6.39544152631649e-07, 'learning_rate/action_model': 6.3954415263164895e-06, 'epoch': 1.79}) +[2026-06-16 08:34:08,878][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 0.16096073389053345, 'train/grad_norm_pre_clip': 17.815797805786133, 'timing/data': 0.033322008326649666, 'timing/model': 1.531208660453558, 'learning_rate/qwen_vl_interface': 6.388123464161388e-07, 'learning_rate/action_model': 6.388123464161389e-06, 'epoch': 1.79}) +[2026-06-16 08:34:15,278][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.0714503675699234, 'train/grad_norm_pre_clip': 7.993490219116211, 'timing/data': 0.03175973892211914, 'timing/model': 1.5938210766762495, 'learning_rate/qwen_vl_interface': 6.380824356856936e-07, 'learning_rate/action_model': 6.380824356856936e-06, 'epoch': 1.79}) +[2026-06-16 08:34:21,574][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.023992929607629776, 'train/grad_norm_pre_clip': 4.5877366065979, 'timing/data': 0.029617490246891975, 'timing/model': 1.745319364592433, 'learning_rate/qwen_vl_interface': 6.373544207403513e-07, 'learning_rate/action_model': 6.373544207403513e-06, 'epoch': 1.79}) +[2026-06-16 08:34:27,665][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 0.0004709357162937522, 'train/grad_norm_pre_clip': 0.07172410935163498, 'timing/data': 0.031065404415130615, 'timing/model': 1.5374819431453943, 'learning_rate/qwen_vl_interface': 6.366283018793723e-07, 'learning_rate/action_model': 6.366283018793723e-06, 'epoch': 1.79}) +[2026-06-16 08:34:33,963][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.052836913615465164, 'train/grad_norm_pre_clip': 20.980445861816406, 'timing/data': 0.03177388571202755, 'timing/model': 1.5337275806814432, 'learning_rate/qwen_vl_interface': 6.359040794012358e-07, 'learning_rate/action_model': 6.359040794012358e-06, 'epoch': 1.79}) +[2026-06-16 08:34:40,168][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.1165347769856453, 'train/grad_norm_pre_clip': 16.918359756469727, 'timing/data': 0.030835818499326706, 'timing/model': 1.5349502488970757, 'learning_rate/qwen_vl_interface': 6.351817536036431e-07, 'learning_rate/action_model': 6.351817536036431e-06, 'epoch': 1.79}) +[2026-06-16 08:34:46,551][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.06285402178764343, 'train/grad_norm_pre_clip': 19.896625518798828, 'timing/data': 0.031285250559449196, 'timing/model': 1.601024441421032, 'learning_rate/qwen_vl_interface': 6.344613247835135e-07, 'learning_rate/action_model': 6.344613247835135e-06, 'epoch': 1.79}) +[2026-06-16 08:34:52,866][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.0013240885455161333, 'train/grad_norm_pre_clip': 0.43870776891708374, 'timing/data': 0.03313959576189518, 'timing/model': 1.5713597685098648, 'learning_rate/qwen_vl_interface': 6.337427932369898e-07, 'learning_rate/action_model': 6.337427932369898e-06, 'epoch': 1.79}) +[2026-06-16 08:34:59,009][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.02279515564441681, 'train/grad_norm_pre_clip': 17.898090362548828, 'timing/data': 0.03157687932252884, 'timing/model': 1.705653078854084, 'learning_rate/qwen_vl_interface': 6.330261592594328e-07, 'learning_rate/action_model': 6.330261592594327e-06, 'epoch': 1.79}) +[2026-06-16 08:35:04,951][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.0424831360578537, 'train/grad_norm_pre_clip': 7.521399974822998, 'timing/data': 0.032510505989193916, 'timing/model': 1.5466894544661045, 'learning_rate/qwen_vl_interface': 6.323114231454229e-07, 'learning_rate/action_model': 6.323114231454229e-06, 'epoch': 1.79}) +[2026-06-16 08:35:11,036][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.23293344676494598, 'train/grad_norm_pre_clip': 32.20986557006836, 'timing/data': 0.03247922286391258, 'timing/model': 1.5336703471839428, 'learning_rate/qwen_vl_interface': 6.315985851887618e-07, 'learning_rate/action_model': 6.315985851887618e-06, 'epoch': 1.79}) +[2026-06-16 08:35:17,137][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.03686376288533211, 'train/grad_norm_pre_clip': 12.723555564880371, 'timing/data': 0.03311576135456562, 'timing/model': 1.534840689972043, 'learning_rate/qwen_vl_interface': 6.308876456824697e-07, 'learning_rate/action_model': 6.3088764568246975e-06, 'epoch': 1.79}) +[2026-06-16 08:35:23,239][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.12127853184938431, 'train/grad_norm_pre_clip': 9.328742027282715, 'timing/data': 0.032302189618349075, 'timing/model': 1.5250699799507856, 'learning_rate/qwen_vl_interface': 6.301786049187875e-07, 'learning_rate/action_model': 6.3017860491878745e-06, 'epoch': 1.79}) +[2026-06-16 08:35:29,384][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.1277303844690323, 'train/grad_norm_pre_clip': 12.682836532592773, 'timing/data': 0.031658196821808815, 'timing/model': 1.5894096307456493, 'learning_rate/qwen_vl_interface': 6.294714631891755e-07, 'learning_rate/action_model': 6.294714631891754e-06, 'epoch': 1.79}) +[2026-06-16 08:35:35,792][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.06628544628620148, 'train/grad_norm_pre_clip': 9.472372055053711, 'timing/data': 0.037270184606313705, 'timing/model': 1.7483989764004946, 'learning_rate/qwen_vl_interface': 6.287662207843115e-07, 'learning_rate/action_model': 6.287662207843114e-06, 'epoch': 1.79}) +[2026-06-16 08:35:41,955][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.003870171494781971, 'train/grad_norm_pre_clip': 0.6700531244277954, 'timing/data': 0.03568051941692829, 'timing/model': 1.5945868771523237, 'learning_rate/qwen_vl_interface': 6.280628779940956e-07, 'learning_rate/action_model': 6.280628779940956e-06, 'epoch': 1.79}) +[2026-06-16 08:35:48,210][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.12793131172657013, 'train/grad_norm_pre_clip': 16.45044708251953, 'timing/data': 0.03317602537572384, 'timing/model': 1.559933889657259, 'learning_rate/qwen_vl_interface': 6.273614351076445e-07, 'learning_rate/action_model': 6.273614351076445e-06, 'epoch': 1.79}) +[2026-06-16 08:35:54,467][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.1688481569290161, 'train/grad_norm_pre_clip': 17.4390811920166, 'timing/data': 0.032700126990675926, 'timing/model': 1.5325678028166294, 'learning_rate/qwen_vl_interface': 6.266618924132951e-07, 'learning_rate/action_model': 6.266618924132951e-06, 'epoch': 1.8}) +[2026-06-16 08:36:00,768][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.09381389617919922, 'train/grad_norm_pre_clip': 8.816136360168457, 'timing/data': 0.032415829598903656, 'timing/model': 1.5269819479435682, 'learning_rate/qwen_vl_interface': 6.25964250198603e-07, 'learning_rate/action_model': 6.2596425019860294e-06, 'epoch': 1.8}) +[2026-06-16 08:36:07,076][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.012459577061235905, 'train/grad_norm_pre_clip': 3.0591864585876465, 'timing/data': 0.03541480749845505, 'timing/model': 1.5521992202848196, 'learning_rate/qwen_vl_interface': 6.25268508750342e-07, 'learning_rate/action_model': 6.25268508750342e-06, 'epoch': 1.8}) +[2026-06-16 08:36:13,378][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.09252527356147766, 'train/grad_norm_pre_clip': 15.382845878601074, 'timing/data': 0.031873080879449844, 'timing/model': 1.7390419840812683, 'learning_rate/qwen_vl_interface': 6.245746683545062e-07, 'learning_rate/action_model': 6.245746683545062e-06, 'epoch': 1.8}) +[2026-06-16 08:36:19,465][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.15813229978084564, 'train/grad_norm_pre_clip': 48.47894287109375, 'timing/data': 0.03128010593354702, 'timing/model': 1.531192883849144, 'learning_rate/qwen_vl_interface': 6.238827292963063e-07, 'learning_rate/action_model': 6.238827292963063e-06, 'epoch': 1.8}) +[2026-06-16 08:36:25,771][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 0.0013238992542028427, 'train/grad_norm_pre_clip': 0.18358880281448364, 'timing/data': 0.03191355988383293, 'timing/model': 1.5392909925431013, 'learning_rate/qwen_vl_interface': 6.231926918601731e-07, 'learning_rate/action_model': 6.2319269186017305e-06, 'epoch': 1.8}) +[2026-06-16 08:36:32,063][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 0.037077393382787704, 'train/grad_norm_pre_clip': 7.571651458740234, 'timing/data': 0.032398950308561325, 'timing/model': 1.505919510498643, 'learning_rate/qwen_vl_interface': 6.22504556329754e-07, 'learning_rate/action_model': 6.22504556329754e-06, 'epoch': 1.8}) +[2026-06-16 08:36:38,341][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.027430284768342972, 'train/grad_norm_pre_clip': 14.456192970275879, 'timing/data': 0.030820637941360474, 'timing/model': 1.4765399191528559, 'learning_rate/qwen_vl_interface': 6.218183229879164e-07, 'learning_rate/action_model': 6.218183229879164e-06, 'epoch': 1.8}) +[2026-06-16 08:36:44,636][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.09217442572116852, 'train/grad_norm_pre_clip': 7.947964668273926, 'timing/data': 0.03149569407105446, 'timing/model': 1.5045992825180292, 'learning_rate/qwen_vl_interface': 6.211339921167447e-07, 'learning_rate/action_model': 6.211339921167446e-06, 'epoch': 1.8}) +[2026-06-16 08:36:50,767][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.12822945415973663, 'train/grad_norm_pre_clip': 12.2775297164917, 'timing/data': 0.03123227134346962, 'timing/model': 1.76558343693614, 'learning_rate/qwen_vl_interface': 6.204515639975415e-07, 'learning_rate/action_model': 6.204515639975414e-06, 'epoch': 1.8}) +[2026-06-16 08:36:57,042][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.026531683281064034, 'train/grad_norm_pre_clip': 7.999688625335693, 'timing/data': 0.035098982974886894, 'timing/model': 1.53258584626019, 'learning_rate/qwen_vl_interface': 6.197710389108262e-07, 'learning_rate/action_model': 6.197710389108262e-06, 'epoch': 1.8}) +[2026-06-16 08:37:03,237][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.058409299701452255, 'train/grad_norm_pre_clip': 8.987339973449707, 'timing/data': 0.03292039409279823, 'timing/model': 1.5314593762159348, 'learning_rate/qwen_vl_interface': 6.190924171363378e-07, 'learning_rate/action_model': 6.190924171363378e-06, 'epoch': 1.8}) +[2026-06-16 08:37:09,440][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.018215658143162727, 'train/grad_norm_pre_clip': 6.728272914886475, 'timing/data': 0.032554011791944504, 'timing/model': 1.5332952104508877, 'learning_rate/qwen_vl_interface': 6.184156989530326e-07, 'learning_rate/action_model': 6.184156989530325e-06, 'epoch': 1.8}) +[2026-06-16 08:37:15,638][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.0053135696798563, 'train/grad_norm_pre_clip': 1.8698776960372925, 'timing/data': 0.032217005267739296, 'timing/model': 1.5267370212823153, 'learning_rate/qwen_vl_interface': 6.17740884639083e-07, 'learning_rate/action_model': 6.177408846390829e-06, 'epoch': 1.8}) +[2026-06-16 08:37:21,733][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.06630515307188034, 'train/grad_norm_pre_clip': 13.116217613220215, 'timing/data': 0.03055865503847599, 'timing/model': 1.4788303319364786, 'learning_rate/qwen_vl_interface': 6.170679744718798e-07, 'learning_rate/action_model': 6.1706797447187975e-06, 'epoch': 1.8}) +[2026-06-16 08:37:27,840][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.004465400241315365, 'train/grad_norm_pre_clip': 2.7105519771575928, 'timing/data': 0.03156728111207485, 'timing/model': 1.7242264710366726, 'learning_rate/qwen_vl_interface': 6.163969687280311e-07, 'learning_rate/action_model': 6.163969687280311e-06, 'epoch': 1.8}) +[2026-06-16 08:37:33,845][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.22712092101573944, 'train/grad_norm_pre_clip': 20.401704788208008, 'timing/data': 0.03498885966837406, 'timing/model': 1.5209932122379541, 'learning_rate/qwen_vl_interface': 6.157278676833627e-07, 'learning_rate/action_model': 6.157278676833627e-06, 'epoch': 1.8}) +[2026-06-16 08:37:40,137][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.006501997821033001, 'train/grad_norm_pre_clip': 1.8118468523025513, 'timing/data': 0.03225432522594929, 'timing/model': 1.5379718746989965, 'learning_rate/qwen_vl_interface': 6.150606716129155e-07, 'learning_rate/action_model': 6.150606716129154e-06, 'epoch': 1.8}) +[2026-06-16 08:37:46,343][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.37855279445648193, 'train/grad_norm_pre_clip': 30.972694396972656, 'timing/data': 0.0313278716057539, 'timing/model': 1.6545290891081095, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 1.8}) +[2026-06-16 08:37:52,456][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 0.001101038884371519, 'train/grad_norm_pre_clip': 0.15563778579235077, 'timing/data': 0.03180485963821411, 'timing/model': 1.5458710007369518, 'learning_rate/qwen_vl_interface': 6.137319954909407e-07, 'learning_rate/action_model': 6.137319954909406e-06, 'epoch': 1.8}) +[2026-06-16 08:37:58,784][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.020373912528157234, 'train/grad_norm_pre_clip': 9.77236270904541, 'timing/data': 0.03457590937614441, 'timing/model': 1.5651648249477148, 'learning_rate/qwen_vl_interface': 6.130705159855815e-07, 'learning_rate/action_model': 6.130705159855814e-06, 'epoch': 1.8}) +[2026-06-16 08:38:05,069][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.007193376775830984, 'train/grad_norm_pre_clip': 3.910501003265381, 'timing/data': 0.03374161384999752, 'timing/model': 1.7285583727061749, 'learning_rate/qwen_vl_interface': 6.124109425467805e-07, 'learning_rate/action_model': 6.124109425467805e-06, 'epoch': 1.8}) +[2026-06-16 08:38:11,068][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.09426077455282211, 'train/grad_norm_pre_clip': 8.82639217376709, 'timing/data': 0.0325031578540802, 'timing/model': 1.5293886233121157, 'learning_rate/qwen_vl_interface': 6.117532754456642e-07, 'learning_rate/action_model': 6.117532754456642e-06, 'epoch': 1.8}) +[2026-06-16 08:38:17,268][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.006994988303631544, 'train/grad_norm_pre_clip': 7.47807502746582, 'timing/data': 0.031937940046191216, 'timing/model': 1.5304281637072563, 'learning_rate/qwen_vl_interface': 6.110975149525746e-07, 'learning_rate/action_model': 6.110975149525746e-06, 'epoch': 1.8}) +[2026-06-16 08:38:23,565][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.001721537089906633, 'train/grad_norm_pre_clip': 0.9736199975013733, 'timing/data': 0.03160756826400757, 'timing/model': 1.510050741955638, 'learning_rate/qwen_vl_interface': 6.104436613370692e-07, 'learning_rate/action_model': 6.104436613370692e-06, 'epoch': 1.8}) +[2026-06-16 08:38:29,764][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.12744061648845673, 'train/grad_norm_pre_clip': 10.653998374938965, 'timing/data': 0.032636985182762146, 'timing/model': 1.5212259367108345, 'learning_rate/qwen_vl_interface': 6.09791714867924e-07, 'learning_rate/action_model': 6.0979171486792396e-06, 'epoch': 1.81}) +[2026-06-16 08:38:36,075][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.16778914630413055, 'train/grad_norm_pre_clip': 10.396571159362793, 'timing/data': 0.032697662711143494, 'timing/model': 1.5779210552573204, 'learning_rate/qwen_vl_interface': 6.091416758131287e-07, 'learning_rate/action_model': 6.091416758131286e-06, 'epoch': 1.81}) +[2026-06-16 08:38:42,475][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.08676204085350037, 'train/grad_norm_pre_clip': 13.980493545532227, 'timing/data': 0.03282998129725456, 'timing/model': 1.732464337721467, 'learning_rate/qwen_vl_interface': 6.084935444398895e-07, 'learning_rate/action_model': 6.084935444398894e-06, 'epoch': 1.81}) +[2026-06-16 08:38:48,678][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.0037243568804115057, 'train/grad_norm_pre_clip': 0.7321529388427734, 'timing/data': 0.04241716489195824, 'timing/model': 1.6103029251098633, 'learning_rate/qwen_vl_interface': 6.078473210146294e-07, 'learning_rate/action_model': 6.078473210146294e-06, 'epoch': 1.81}) +[2026-06-16 08:38:55,077][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 0.01245291717350483, 'train/grad_norm_pre_clip': 2.926893711090088, 'timing/data': 0.032684119418263435, 'timing/model': 1.5172856636345387, 'learning_rate/qwen_vl_interface': 6.072030058029862e-07, 'learning_rate/action_model': 6.072030058029861e-06, 'epoch': 1.81}) +[2026-06-16 08:39:01,254][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 0.007693649735301733, 'train/grad_norm_pre_clip': 3.365884304046631, 'timing/data': 0.032109497115015984, 'timing/model': 1.5504329781979322, 'learning_rate/qwen_vl_interface': 6.065605990698136e-07, 'learning_rate/action_model': 6.065605990698136e-06, 'epoch': 1.81}) +[2026-06-16 08:39:07,343][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.19411584734916687, 'train/grad_norm_pre_clip': 27.295019149780273, 'timing/data': 0.032225148752331734, 'timing/model': 1.519336998462677, 'learning_rate/qwen_vl_interface': 6.059201010791811e-07, 'learning_rate/action_model': 6.059201010791811e-06, 'epoch': 1.81}) +[2026-06-16 08:39:13,440][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.010619045235216618, 'train/grad_norm_pre_clip': 4.222362995147705, 'timing/data': 0.03111989051103592, 'timing/model': 1.5043497942388058, 'learning_rate/qwen_vl_interface': 6.052815120943721e-07, 'learning_rate/action_model': 6.052815120943721e-06, 'epoch': 1.81}) +[2026-06-16 08:39:19,542][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.07833417505025864, 'train/grad_norm_pre_clip': 10.830862998962402, 'timing/data': 0.03133048489689827, 'timing/model': 1.7261593509465456, 'learning_rate/qwen_vl_interface': 6.046448323778874e-07, 'learning_rate/action_model': 6.046448323778874e-06, 'epoch': 1.81}) +[2026-06-16 08:39:25,574][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.020412081852555275, 'train/grad_norm_pre_clip': 5.549227714538574, 'timing/data': 0.031048236414790154, 'timing/model': 1.6685078851878643, 'learning_rate/qwen_vl_interface': 6.040100621914419e-07, 'learning_rate/action_model': 6.040100621914418e-06, 'epoch': 1.81}) +[2026-06-16 08:39:31,558][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.14007316529750824, 'train/grad_norm_pre_clip': 16.903690338134766, 'timing/data': 0.03302730433642864, 'timing/model': 1.5514534674584866, 'learning_rate/qwen_vl_interface': 6.033772017959652e-07, 'learning_rate/action_model': 6.033772017959652e-06, 'epoch': 1.81}) +[2026-06-16 08:39:37,740][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.05905504897236824, 'train/grad_norm_pre_clip': 20.755823135375977, 'timing/data': 0.03139118477702141, 'timing/model': 1.5406890287995338, 'learning_rate/qwen_vl_interface': 6.027462514516022e-07, 'learning_rate/action_model': 6.0274625145160214e-06, 'epoch': 1.81}) +[2026-06-16 08:39:43,843][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.13417597115039825, 'train/grad_norm_pre_clip': 53.00794219970703, 'timing/data': 0.039273349568247795, 'timing/model': 1.5059728249907494, 'learning_rate/qwen_vl_interface': 6.021172114177139e-07, 'learning_rate/action_model': 6.021172114177139e-06, 'epoch': 1.81}) +[2026-06-16 08:39:50,372][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.006467948667705059, 'train/grad_norm_pre_clip': 2.440673351287842, 'timing/data': 0.03044489212334156, 'timing/model': 1.6584743037819862, 'learning_rate/qwen_vl_interface': 6.014900819528734e-07, 'learning_rate/action_model': 6.0149008195287335e-06, 'epoch': 1.81}) +[2026-06-16 08:39:56,677][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.033365361392498016, 'train/grad_norm_pre_clip': 7.0521440505981445, 'timing/data': 0.03135468065738678, 'timing/model': 1.7435471136122942, 'learning_rate/qwen_vl_interface': 6.008648633148707e-07, 'learning_rate/action_model': 6.008648633148707e-06, 'epoch': 1.81}) +[2026-06-16 08:40:02,973][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.01185555662959814, 'train/grad_norm_pre_clip': 4.864119529724121, 'timing/data': 0.03180895186960697, 'timing/model': 1.7340776603668928, 'learning_rate/qwen_vl_interface': 6.002415557607097e-07, 'learning_rate/action_model': 6.002415557607096e-06, 'epoch': 1.81}) +[2026-06-16 08:40:09,064][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.010626115836203098, 'train/grad_norm_pre_clip': 1.9469245672225952, 'timing/data': 0.03117220103740692, 'timing/model': 1.5318374559283257, 'learning_rate/qwen_vl_interface': 5.996201595466088e-07, 'learning_rate/action_model': 5.996201595466088e-06, 'epoch': 1.81}) +[2026-06-16 08:40:15,277][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.012071644887328148, 'train/grad_norm_pre_clip': 2.7864670753479004, 'timing/data': 0.032056303694844246, 'timing/model': 1.5385647621005774, 'learning_rate/qwen_vl_interface': 5.990006749279998e-07, 'learning_rate/action_model': 5.990006749279998e-06, 'epoch': 1.81}) +[2026-06-16 08:40:21,567][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.28484389185905457, 'train/grad_norm_pre_clip': 42.90269088745117, 'timing/data': 0.032615307718515396, 'timing/model': 1.5304435715079308, 'learning_rate/qwen_vl_interface': 5.983831021595296e-07, 'learning_rate/action_model': 5.9838310215952956e-06, 'epoch': 1.81}) +[2026-06-16 08:40:27,869][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.0033286847174167633, 'train/grad_norm_pre_clip': 1.303345799446106, 'timing/data': 0.03299958072602749, 'timing/model': 1.5157594084739685, 'learning_rate/qwen_vl_interface': 5.977674414950595e-07, 'learning_rate/action_model': 5.9776744149505955e-06, 'epoch': 1.81}) +[2026-06-16 08:40:34,082][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.06241939589381218, 'train/grad_norm_pre_clip': 26.607633590698242, 'timing/data': 0.032334594056010246, 'timing/model': 1.747039994224906, 'learning_rate/qwen_vl_interface': 5.971536931876635e-07, 'learning_rate/action_model': 5.971536931876634e-06, 'epoch': 1.81}) +[2026-06-16 08:40:40,172][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.04541179910302162, 'train/grad_norm_pre_clip': 21.67201805114746, 'timing/data': 0.03174062632024288, 'timing/model': 1.540123114362359, 'learning_rate/qwen_vl_interface': 5.965418574896315e-07, 'learning_rate/action_model': 5.9654185748963145e-06, 'epoch': 1.81}) +[2026-06-16 08:40:46,563][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.014142109081149101, 'train/grad_norm_pre_clip': 2.3715662956237793, 'timing/data': 0.032584423199296, 'timing/model': 1.530658258125186, 'learning_rate/qwen_vl_interface': 5.959319346524646e-07, 'learning_rate/action_model': 5.959319346524646e-06, 'epoch': 1.81}) +[2026-06-16 08:40:52,948][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.06933438777923584, 'train/grad_norm_pre_clip': 11.813729286193848, 'timing/data': 0.03295108303427696, 'timing/model': 1.5626779329031706, 'learning_rate/qwen_vl_interface': 5.953239249268804e-07, 'learning_rate/action_model': 5.953239249268804e-06, 'epoch': 1.81}) +[2026-06-16 08:40:59,034][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.16436435282230377, 'train/grad_norm_pre_clip': 22.01112937927246, 'timing/data': 0.03308304026722908, 'timing/model': 1.5159029122442007, 'learning_rate/qwen_vl_interface': 5.947178285628081e-07, 'learning_rate/action_model': 5.94717828562808e-06, 'epoch': 1.81}) +[2026-06-16 08:41:05,136][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.0017377983313053846, 'train/grad_norm_pre_clip': 0.5374326109886169, 'timing/data': 0.03167029097676277, 'timing/model': 1.5002395957708359, 'learning_rate/qwen_vl_interface': 5.941136458093907e-07, 'learning_rate/action_model': 5.941136458093907e-06, 'epoch': 1.81}) +[2026-06-16 08:41:11,204][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.20806513726711273, 'train/grad_norm_pre_clip': 21.75702667236328, 'timing/data': 0.033552370965480804, 'timing/model': 1.6983049884438515, 'learning_rate/qwen_vl_interface': 5.935113769149855e-07, 'learning_rate/action_model': 5.935113769149856e-06, 'epoch': 1.82}) +[2026-06-16 08:41:17,157][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.1436883509159088, 'train/grad_norm_pre_clip': 11.762429237365723, 'timing/data': 0.03184117190539837, 'timing/model': 1.553897749632597, 'learning_rate/qwen_vl_interface': 5.92911022127163e-07, 'learning_rate/action_model': 5.929110221271629e-06, 'epoch': 1.82}) +[2026-06-16 08:41:23,634][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.017977800220251083, 'train/grad_norm_pre_clip': 3.4510860443115234, 'timing/data': 0.03197525255382061, 'timing/model': 1.5328613165766, 'learning_rate/qwen_vl_interface': 5.92312581692705e-07, 'learning_rate/action_model': 5.923125816927049e-06, 'epoch': 1.82}) +[2026-06-16 08:41:29,739][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.23317192494869232, 'train/grad_norm_pre_clip': 31.617828369140625, 'timing/data': 0.03432866372168064, 'timing/model': 1.498605327680707, 'learning_rate/qwen_vl_interface': 5.917160558576088e-07, 'learning_rate/action_model': 5.917160558576088e-06, 'epoch': 1.82}) +[2026-06-16 08:41:35,818][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.09177250415086746, 'train/grad_norm_pre_clip': 25.3942813873291, 'timing/data': 0.03302750922739506, 'timing/model': 1.4529160857200623, 'learning_rate/qwen_vl_interface': 5.911214448670829e-07, 'learning_rate/action_model': 5.911214448670828e-06, 'epoch': 1.82}) +[2026-06-16 08:41:41,680][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.027049997821450233, 'train/grad_norm_pre_clip': 10.927675247192383, 'timing/data': 0.026336999610066414, 'timing/model': 1.46109863743186, 'learning_rate/qwen_vl_interface': 5.905287489655499e-07, 'learning_rate/action_model': 5.905287489655499e-06, 'epoch': 1.82}) +[2026-06-16 08:41:47,511][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 0.004825187381356955, 'train/grad_norm_pre_clip': 0.5395038723945618, 'timing/data': 0.026887455955147743, 'timing/model': 1.5655391700565815, 'learning_rate/qwen_vl_interface': 5.899379683966445e-07, 'learning_rate/action_model': 5.899379683966445e-06, 'epoch': 1.82}) +[2026-06-16 08:41:53,214][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.2712998390197754, 'train/grad_norm_pre_clip': 24.08814239501953, 'timing/data': 0.02660929039120674, 'timing/model': 1.4641172178089619, 'learning_rate/qwen_vl_interface': 5.893491034032142e-07, 'learning_rate/action_model': 5.893491034032142e-06, 'epoch': 1.82}) +[2026-06-16 08:41:59,125][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.26438361406326294, 'train/grad_norm_pre_clip': 4.472399711608887, 'timing/data': 0.026595095172524452, 'timing/model': 1.4527644477784634, 'learning_rate/qwen_vl_interface': 5.887621542273196e-07, 'learning_rate/action_model': 5.887621542273196e-06, 'epoch': 1.82}) +[2026-06-16 08:42:05,110][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.2608124315738678, 'train/grad_norm_pre_clip': 19.36025619506836, 'timing/data': 0.026998788118362427, 'timing/model': 1.5308998133987188, 'learning_rate/qwen_vl_interface': 5.881771211102327e-07, 'learning_rate/action_model': 5.881771211102327e-06, 'epoch': 1.82}) +[2026-06-16 08:49:21,049][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.001515475451014936, 'train/grad_norm_pre_clip': 0.22431209683418274, 'eval/loss': 0.043809812515974045, 'eval/flap_precision': 0.9546666666666667, 'eval/flap_recall': 0.9715061058344641, 'eval/flap_f1': 0.9630127774041695, 'eval/flap_rate_model': 0.20833333333333334, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9630127774041696, 'timing/data': 0.027397960424423218, 'timing/model': 1.4474841188639402, 'learning_rate/qwen_vl_interface': 5.875940042924391e-07, 'learning_rate/action_model': 5.875940042924391e-06, 'epoch': 1.82}) +[2026-06-16 08:49:21,050][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4700_state +[2026-06-16 08:49:21,054][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 08:49:27,225][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4700_state/model.safetensors +[2026-06-16 08:49:37,532][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4700_state/optimizer.bin +[2026-06-16 08:49:37,532][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4700_state/sampler.bin +[2026-06-16 08:49:37,532][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4700_state/sampler_1.bin +[2026-06-16 08:49:37,533][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4700_state/random_states_0.pkl +[2026-06-16 08:49:37,625][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4700 +[2026-06-16 08:49:37,625][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 08:49:37,677][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 08:49:44,023][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.11271774768829346, 'train/grad_norm_pre_clip': 12.43276596069336, 'timing/data': 0.03301446698606014, 'timing/model': 1.5564020797610283, 'learning_rate/qwen_vl_interface': 5.870128040136368e-07, 'learning_rate/action_model': 5.870128040136367e-06, 'epoch': 1.82}) +[2026-06-16 08:49:50,280][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.19551612436771393, 'train/grad_norm_pre_clip': 26.048933029174805, 'timing/data': 0.03372318297624588, 'timing/model': 1.754706010222435, 'learning_rate/qwen_vl_interface': 5.864335205127332e-07, 'learning_rate/action_model': 5.864335205127332e-06, 'epoch': 1.82}) +[2026-06-16 08:49:56,544][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.19592273235321045, 'train/grad_norm_pre_clip': 14.666812896728516, 'timing/data': 0.03399920277297497, 'timing/model': 1.729772599413991, 'learning_rate/qwen_vl_interface': 5.858561540278513e-07, 'learning_rate/action_model': 5.858561540278513e-06, 'epoch': 1.82}) +[2026-06-16 08:50:02,521][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 0.0986148864030838, 'train/grad_norm_pre_clip': 25.793834686279297, 'timing/data': 0.032081183046102524, 'timing/model': 1.4876857120543718, 'learning_rate/qwen_vl_interface': 5.852807047963244e-07, 'learning_rate/action_model': 5.852807047963243e-06, 'epoch': 1.82}) +[2026-06-16 08:50:08,730][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.14434055984020233, 'train/grad_norm_pre_clip': 13.350519180297852, 'timing/data': 0.03310009092092514, 'timing/model': 1.5273559596389532, 'learning_rate/qwen_vl_interface': 5.847071730546974e-07, 'learning_rate/action_model': 5.847071730546974e-06, 'epoch': 1.82}) +[2026-06-16 08:50:14,925][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.0033709390554577112, 'train/grad_norm_pre_clip': 1.2883743047714233, 'timing/data': 0.03310655243694782, 'timing/model': 1.5323869176208973, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 1.82}) +[2026-06-16 08:50:21,030][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 0.18291659653186798, 'train/grad_norm_pre_clip': 15.755439758300781, 'timing/data': 0.03643452003598213, 'timing/model': 1.4891182780265808, 'learning_rate/qwen_vl_interface': 5.835658629833866e-07, 'learning_rate/action_model': 5.8356586298338654e-06, 'epoch': 1.82}) +[2026-06-16 08:50:27,144][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.205780491232872, 'train/grad_norm_pre_clip': 10.206417083740234, 'timing/data': 0.02651839330792427, 'timing/model': 1.4740981254726648, 'learning_rate/qwen_vl_interface': 5.829980851228512e-07, 'learning_rate/action_model': 5.829980851228512e-06, 'epoch': 1.82}) +[2026-06-16 08:50:33,268][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.11161814630031586, 'train/grad_norm_pre_clip': 15.458226203918457, 'timing/data': 0.027640970423817635, 'timing/model': 1.7925638109445572, 'learning_rate/qwen_vl_interface': 5.824322256905155e-07, 'learning_rate/action_model': 5.824322256905154e-06, 'epoch': 1.82}) +[2026-06-16 08:50:38,906][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.0027821632102131844, 'train/grad_norm_pre_clip': 0.7452993392944336, 'timing/data': 0.0277422908693552, 'timing/model': 1.440269835293293, 'learning_rate/qwen_vl_interface': 5.818682849189818e-07, 'learning_rate/action_model': 5.818682849189817e-06, 'epoch': 1.82}) +[2026-06-16 08:50:45,019][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.08593673259019852, 'train/grad_norm_pre_clip': 13.952112197875977, 'timing/data': 0.026439320296049118, 'timing/model': 1.5397771839052439, 'learning_rate/qwen_vl_interface': 5.813062630400653e-07, 'learning_rate/action_model': 5.813062630400653e-06, 'epoch': 1.82}) +[2026-06-16 08:50:50,916][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 0.16998080909252167, 'train/grad_norm_pre_clip': 10.83234977722168, 'timing/data': 0.03486582078039646, 'timing/model': 1.438344655558467, 'learning_rate/qwen_vl_interface': 5.807461602847922e-07, 'learning_rate/action_model': 5.807461602847922e-06, 'epoch': 1.82}) +[2026-06-16 08:50:56,923][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.0029352803248912096, 'train/grad_norm_pre_clip': 0.659549355506897, 'timing/data': 0.02679680474102497, 'timing/model': 1.4520845487713814, 'learning_rate/qwen_vl_interface': 5.801879768833998e-07, 'learning_rate/action_model': 5.801879768833998e-06, 'epoch': 1.82}) +[2026-06-16 08:51:03,012][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.1303170770406723, 'train/grad_norm_pre_clip': 9.998117446899414, 'timing/data': 0.044905541464686394, 'timing/model': 1.641526073217392, 'learning_rate/qwen_vl_interface': 5.796317130653363e-07, 'learning_rate/action_model': 5.796317130653363e-06, 'epoch': 1.82}) +[2026-06-16 08:51:09,574][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.005245519336313009, 'train/grad_norm_pre_clip': 2.6810715198516846, 'timing/data': 0.04275699704885483, 'timing/model': 1.8982573971152306, 'learning_rate/qwen_vl_interface': 5.790773690592592e-07, 'learning_rate/action_model': 5.7907736905925925e-06, 'epoch': 1.82}) +[2026-06-16 08:51:15,761][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.2058182805776596, 'train/grad_norm_pre_clip': 10.940018653869629, 'timing/data': 0.031923653557896614, 'timing/model': 1.5368558522313833, 'learning_rate/qwen_vl_interface': 5.7852494509304e-07, 'learning_rate/action_model': 5.7852494509304e-06, 'epoch': 1.83}) +[2026-06-16 08:51:22,071][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.2710362672805786, 'train/grad_norm_pre_clip': 17.90293312072754, 'timing/data': 0.03149905800819397, 'timing/model': 1.537853704765439, 'learning_rate/qwen_vl_interface': 5.779744413937592e-07, 'learning_rate/action_model': 5.779744413937592e-06, 'epoch': 1.83}) +[2026-06-16 08:51:28,279][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.027720730751752853, 'train/grad_norm_pre_clip': 4.0328474044799805, 'timing/data': 0.03258830867707729, 'timing/model': 1.557558499276638, 'learning_rate/qwen_vl_interface': 5.774258581877073e-07, 'learning_rate/action_model': 5.774258581877073e-06, 'epoch': 1.83}) +[2026-06-16 08:51:34,505][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.026621568948030472, 'train/grad_norm_pre_clip': 6.820637226104736, 'timing/data': 0.026097454130649567, 'timing/model': 1.470095569267869, 'learning_rate/qwen_vl_interface': 5.768791957003866e-07, 'learning_rate/action_model': 5.768791957003865e-06, 'epoch': 1.83}) +[2026-06-16 08:51:40,406][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.02267739549279213, 'train/grad_norm_pre_clip': 4.924310207366943, 'timing/data': 0.026304690167307854, 'timing/model': 1.4880128055810928, 'learning_rate/qwen_vl_interface': 5.763344541565091e-07, 'learning_rate/action_model': 5.7633445415650905e-06, 'epoch': 1.83}) +[2026-06-16 08:51:46,302][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.1185946837067604, 'train/grad_norm_pre_clip': 7.6697564125061035, 'timing/data': 0.026660114526748657, 'timing/model': 1.6336972787976265, 'learning_rate/qwen_vl_interface': 5.75791633779998e-07, 'learning_rate/action_model': 5.75791633779998e-06, 'epoch': 1.83}) +[2026-06-16 08:51:52,013][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 0.00022636450012214482, 'train/grad_norm_pre_clip': 0.0367618054151535, 'timing/data': 0.02626398764550686, 'timing/model': 1.445284228771925, 'learning_rate/qwen_vl_interface': 5.752507347939857e-07, 'learning_rate/action_model': 5.752507347939858e-06, 'epoch': 1.83}) +[2026-06-16 08:51:57,911][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.19930337369441986, 'train/grad_norm_pre_clip': 10.491790771484375, 'timing/data': 0.025535564869642258, 'timing/model': 1.4431227631866932, 'learning_rate/qwen_vl_interface': 5.74711757420816e-07, 'learning_rate/action_model': 5.747117574208159e-06, 'epoch': 1.83}) +[2026-06-16 08:52:03,813][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.012888411059975624, 'train/grad_norm_pre_clip': 2.253300666809082, 'timing/data': 0.027031585574150085, 'timing/model': 1.442934324964881, 'learning_rate/qwen_vl_interface': 5.74174701882042e-07, 'learning_rate/action_model': 5.74174701882042e-06, 'epoch': 1.83}) +[2026-06-16 08:52:09,617][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.005808120593428612, 'train/grad_norm_pre_clip': 0.5517119765281677, 'timing/data': 0.02579493448138237, 'timing/model': 1.4484215304255486, 'learning_rate/qwen_vl_interface': 5.736395683984262e-07, 'learning_rate/action_model': 5.736395683984261e-06, 'epoch': 1.83}) +[2026-06-16 08:52:15,510][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.03175564110279083, 'train/grad_norm_pre_clip': 8.10550308227539, 'timing/data': 0.026577049866318703, 'timing/model': 1.4432907663285732, 'learning_rate/qwen_vl_interface': 5.731063571899421e-07, 'learning_rate/action_model': 5.731063571899421e-06, 'epoch': 1.83}) +[2026-06-16 08:52:21,410][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.009672741405665874, 'train/grad_norm_pre_clip': 5.132038116455078, 'timing/data': 0.026677744463086128, 'timing/model': 1.64017073251307, 'learning_rate/qwen_vl_interface': 5.72575068475773e-07, 'learning_rate/action_model': 5.72575068475773e-06, 'epoch': 1.83}) +[2026-06-16 08:52:27,540][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.11500915139913559, 'train/grad_norm_pre_clip': 8.246954917907715, 'timing/data': 0.025876903906464577, 'timing/model': 1.6356524880975485, 'learning_rate/qwen_vl_interface': 5.720457024743115e-07, 'learning_rate/action_model': 5.720457024743115e-06, 'epoch': 1.83}) +[2026-06-16 08:52:34,026][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.027959192171692848, 'train/grad_norm_pre_clip': 16.794015884399414, 'timing/data': 0.031624192371964455, 'timing/model': 1.6505944896489382, 'learning_rate/qwen_vl_interface': 5.715182594031601e-07, 'learning_rate/action_model': 5.7151825940316e-06, 'epoch': 1.83}) +[2026-06-16 08:52:40,292][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.005967398639768362, 'train/grad_norm_pre_clip': 0.5875575542449951, 'timing/data': 0.03250798024237156, 'timing/model': 1.5535659324377775, 'learning_rate/qwen_vl_interface': 5.709927394791305e-07, 'learning_rate/action_model': 5.709927394791305e-06, 'epoch': 1.83}) +[2026-06-16 08:52:46,571][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.11831369996070862, 'train/grad_norm_pre_clip': 7.121823310852051, 'timing/data': 0.027691584080457687, 'timing/model': 1.546787928789854, 'learning_rate/qwen_vl_interface': 5.70469142918244e-07, 'learning_rate/action_model': 5.704691429182441e-06, 'epoch': 1.83}) +[2026-06-16 08:52:52,983][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.003408899297937751, 'train/grad_norm_pre_clip': 1.1625584363937378, 'timing/data': 0.030980762094259262, 'timing/model': 1.5245925802737474, 'learning_rate/qwen_vl_interface': 5.699474699357317e-07, 'learning_rate/action_model': 5.699474699357317e-06, 'epoch': 1.83}) +[2026-06-16 08:52:59,366][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.09565572440624237, 'train/grad_norm_pre_clip': 31.04171371459961, 'timing/data': 0.03216802887618542, 'timing/model': 1.7622585967183113, 'learning_rate/qwen_vl_interface': 5.694277207460331e-07, 'learning_rate/action_model': 5.694277207460332e-06, 'epoch': 1.83}) +[2026-06-16 08:53:05,337][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.005169543903321028, 'train/grad_norm_pre_clip': 1.7780650854110718, 'timing/data': 0.03210742212831974, 'timing/model': 1.5447525661438704, 'learning_rate/qwen_vl_interface': 5.689098955627982e-07, 'learning_rate/action_model': 5.689098955627982e-06, 'epoch': 1.83}) +[2026-06-16 08:53:11,543][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.015356084331870079, 'train/grad_norm_pre_clip': 5.569282531738281, 'timing/data': 0.031313683837652206, 'timing/model': 1.5604750886559486, 'learning_rate/qwen_vl_interface': 5.683939945988853e-07, 'learning_rate/action_model': 5.683939945988853e-06, 'epoch': 1.83}) +[2026-06-16 08:53:17,839][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.034404970705509186, 'train/grad_norm_pre_clip': 13.816848754882812, 'timing/data': 0.03304632380604744, 'timing/model': 1.538589283823967, 'learning_rate/qwen_vl_interface': 5.678800180663608e-07, 'learning_rate/action_model': 5.678800180663609e-06, 'epoch': 1.83}) +[2026-06-16 08:53:23,939][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.01129189133644104, 'train/grad_norm_pre_clip': 2.573946714401245, 'timing/data': 0.030341697856783867, 'timing/model': 1.5354737155139446, 'learning_rate/qwen_vl_interface': 5.673679661765018e-07, 'learning_rate/action_model': 5.6736796617650175e-06, 'epoch': 1.83}) +[2026-06-16 08:53:30,064][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.0008273679413832724, 'train/grad_norm_pre_clip': 0.3457570970058441, 'timing/data': 0.03181176073849201, 'timing/model': 1.5029618795961142, 'learning_rate/qwen_vl_interface': 5.668578391397938e-07, 'learning_rate/action_model': 5.668578391397937e-06, 'epoch': 1.83}) +[2026-06-16 08:53:36,470][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 0.0035419620107859373, 'train/grad_norm_pre_clip': 0.8707354664802551, 'timing/data': 0.03222627192735672, 'timing/model': 1.7534806802868843, 'learning_rate/qwen_vl_interface': 5.663496371659303e-07, 'learning_rate/action_model': 5.663496371659303e-06, 'epoch': 1.83}) +[2026-06-16 08:53:42,448][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.005913161672651768, 'train/grad_norm_pre_clip': 0.5680162310600281, 'timing/data': 0.03237982839345932, 'timing/model': 1.5392827838659286, 'learning_rate/qwen_vl_interface': 5.658433604638141e-07, 'learning_rate/action_model': 5.65843360463814e-06, 'epoch': 1.83}) +[2026-06-16 08:53:48,645][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.12343809753656387, 'train/grad_norm_pre_clip': 8.514809608459473, 'timing/data': 0.03158550523221493, 'timing/model': 1.5337994527071714, 'learning_rate/qwen_vl_interface': 5.653390092415559e-07, 'learning_rate/action_model': 5.653390092415559e-06, 'epoch': 1.83}) +[2026-06-16 08:53:54,756][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 0.004134567454457283, 'train/grad_norm_pre_clip': 0.8650951981544495, 'timing/data': 0.03258916921913624, 'timing/model': 1.552014047279954, 'learning_rate/qwen_vl_interface': 5.648365837064759e-07, 'learning_rate/action_model': 5.648365837064759e-06, 'epoch': 1.84}) +[2026-06-16 08:54:00,941][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.003879150375723839, 'train/grad_norm_pre_clip': 0.7783670425415039, 'timing/data': 0.03212489187717438, 'timing/model': 1.5326303448528051, 'learning_rate/qwen_vl_interface': 5.643360840651025e-07, 'learning_rate/action_model': 5.643360840651025e-06, 'epoch': 1.84}) +[2026-06-16 08:54:07,040][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.0017024438129737973, 'train/grad_norm_pre_clip': 0.33445072174072266, 'timing/data': 0.030154960229992867, 'timing/model': 1.4945098087191582, 'learning_rate/qwen_vl_interface': 5.638375105231719e-07, 'learning_rate/action_model': 5.638375105231719e-06, 'epoch': 1.84}) +[2026-06-16 08:54:13,325][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.011529488489031792, 'train/grad_norm_pre_clip': 3.721543073654175, 'timing/data': 0.03408535569906235, 'timing/model': 1.825236115604639, 'learning_rate/qwen_vl_interface': 5.633408632856294e-07, 'learning_rate/action_model': 5.6334086328562945e-06, 'epoch': 1.84}) +[2026-06-16 08:54:19,889][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.19090336561203003, 'train/grad_norm_pre_clip': 18.04644775390625, 'timing/data': 0.0332300178706646, 'timing/model': 1.872471883893013, 'learning_rate/qwen_vl_interface': 5.628461425566268e-07, 'learning_rate/action_model': 5.6284614255662675e-06, 'epoch': 1.84}) +[2026-06-16 08:54:26,158][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.00398373743519187, 'train/grad_norm_pre_clip': 1.7526557445526123, 'timing/data': 0.033258477225899696, 'timing/model': 1.648438721895218, 'learning_rate/qwen_vl_interface': 5.62353348539526e-07, 'learning_rate/action_model': 5.623533485395259e-06, 'epoch': 1.84}) +[2026-06-16 08:54:32,473][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 0.0006662914529442787, 'train/grad_norm_pre_clip': 0.1248326301574707, 'timing/data': 0.03372540883719921, 'timing/model': 1.542653564363718, 'learning_rate/qwen_vl_interface': 5.618624814368956e-07, 'learning_rate/action_model': 5.618624814368956e-06, 'epoch': 1.84}) +[2026-06-16 08:54:38,866][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.028702368959784508, 'train/grad_norm_pre_clip': 18.800743103027344, 'timing/data': 0.03280181623995304, 'timing/model': 1.6239292658865452, 'learning_rate/qwen_vl_interface': 5.613735414505126e-07, 'learning_rate/action_model': 5.613735414505125e-06, 'epoch': 1.84}) +[2026-06-16 08:54:45,174][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.03350931033492088, 'train/grad_norm_pre_clip': 8.252245903015137, 'timing/data': 0.03489556163549423, 'timing/model': 1.5123921521008015, 'learning_rate/qwen_vl_interface': 5.608865287813617e-07, 'learning_rate/action_model': 5.608865287813616e-06, 'epoch': 1.84}) +[2026-06-16 08:54:51,479][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.23451125621795654, 'train/grad_norm_pre_clip': 8.294333457946777, 'timing/data': 0.03495748154819012, 'timing/model': 1.7408415861427784, 'learning_rate/qwen_vl_interface': 5.604014436296362e-07, 'learning_rate/action_model': 5.604014436296362e-06, 'epoch': 1.84}) +[2026-06-16 08:54:57,874][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.018016837537288666, 'train/grad_norm_pre_clip': 8.06615924835205, 'timing/data': 0.03151031769812107, 'timing/model': 1.7430491279810667, 'learning_rate/qwen_vl_interface': 5.599182861947348e-07, 'learning_rate/action_model': 5.599182861947347e-06, 'epoch': 1.84}) +[2026-06-16 08:55:04,063][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.056640930473804474, 'train/grad_norm_pre_clip': 6.395127773284912, 'timing/data': 0.03262539766728878, 'timing/model': 1.5284775830805302, 'learning_rate/qwen_vl_interface': 5.594370566752665e-07, 'learning_rate/action_model': 5.5943705667526645e-06, 'epoch': 1.84}) +[2026-06-16 08:55:10,267][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.09703625738620758, 'train/grad_norm_pre_clip': 4.7516865730285645, 'timing/data': 0.031859174370765686, 'timing/model': 1.534969113767147, 'learning_rate/qwen_vl_interface': 5.589577552690456e-07, 'learning_rate/action_model': 5.589577552690456e-06, 'epoch': 1.84}) +[2026-06-16 08:55:16,465][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.010030007921159267, 'train/grad_norm_pre_clip': 3.9930667877197266, 'timing/data': 0.030874410644173622, 'timing/model': 1.5360074378550053, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 1.84}) +[2026-06-16 08:55:22,679][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.0031529527623206377, 'train/grad_norm_pre_clip': 0.5813062787055969, 'timing/data': 0.032314589247107506, 'timing/model': 1.5349265467375517, 'learning_rate/qwen_vl_interface': 5.580049375836466e-07, 'learning_rate/action_model': 5.580049375836466e-06, 'epoch': 1.84}) +[2026-06-16 08:55:28,869][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.0019477094756439328, 'train/grad_norm_pre_clip': 0.42105886340141296, 'timing/data': 0.027313681319355965, 'timing/model': 1.6294788531959057, 'learning_rate/qwen_vl_interface': 5.575314216961356e-07, 'learning_rate/action_model': 5.575314216961356e-06, 'epoch': 1.84}) +[2026-06-16 08:55:34,477][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.13507527112960815, 'train/grad_norm_pre_clip': 14.241155624389648, 'timing/data': 0.02787865325808525, 'timing/model': 1.429273271933198, 'learning_rate/qwen_vl_interface': 5.570598347052073e-07, 'learning_rate/action_model': 5.570598347052072e-06, 'epoch': 1.84}) +[2026-06-16 08:55:40,378][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.3242928385734558, 'train/grad_norm_pre_clip': 13.428095817565918, 'timing/data': 0.02614663913846016, 'timing/model': 1.4368770327419043, 'learning_rate/qwen_vl_interface': 5.565901768047126e-07, 'learning_rate/action_model': 5.5659017680471255e-06, 'epoch': 1.84}) +[2026-06-16 08:55:46,533][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.011664064601063728, 'train/grad_norm_pre_clip': 1.8119553327560425, 'timing/data': 0.032111791893839836, 'timing/model': 1.533601215109229, 'learning_rate/qwen_vl_interface': 5.561224481877105e-07, 'learning_rate/action_model': 5.561224481877105e-06, 'epoch': 1.84}) +[2026-06-16 08:55:52,636][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.11723294854164124, 'train/grad_norm_pre_clip': 18.935626983642578, 'timing/data': 0.03222385607659817, 'timing/model': 1.5164325442165136, 'learning_rate/qwen_vl_interface': 5.55656649046467e-07, 'learning_rate/action_model': 5.556566490464669e-06, 'epoch': 1.84}) +[2026-06-16 08:55:58,843][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 0.015350925736129284, 'train/grad_norm_pre_clip': 4.394026279449463, 'timing/data': 0.04640408232808113, 'timing/model': 1.5276341680437326, 'learning_rate/qwen_vl_interface': 5.551927795724544e-07, 'learning_rate/action_model': 5.551927795724543e-06, 'epoch': 1.84}) +[2026-06-16 08:56:05,049][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.011887757107615471, 'train/grad_norm_pre_clip': 5.889849662780762, 'timing/data': 0.033475860953330994, 'timing/model': 1.744972487911582, 'learning_rate/qwen_vl_interface': 5.547308399563515e-07, 'learning_rate/action_model': 5.547308399563514e-06, 'epoch': 1.84}) +[2026-06-16 08:56:10,938][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.07772523164749146, 'train/grad_norm_pre_clip': 4.862720966339111, 'timing/data': 0.03281691484153271, 'timing/model': 1.5332247484475374, 'learning_rate/qwen_vl_interface': 5.542708303880442e-07, 'learning_rate/action_model': 5.542708303880442e-06, 'epoch': 1.84}) +[2026-06-16 08:56:16,980][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.0024382376577705145, 'train/grad_norm_pre_clip': 0.5393928289413452, 'timing/data': 0.02806680090725422, 'timing/model': 1.456931872293353, 'learning_rate/qwen_vl_interface': 5.538127510566258e-07, 'learning_rate/action_model': 5.538127510566258e-06, 'epoch': 1.84}) +[2026-06-16 08:56:22,920][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.0034083116333931684, 'train/grad_norm_pre_clip': 0.5831413269042969, 'timing/data': 0.02714623138308525, 'timing/model': 1.4459788836538792, 'learning_rate/qwen_vl_interface': 5.533566021503944e-07, 'learning_rate/action_model': 5.533566021503943e-06, 'epoch': 1.84}) +[2026-06-16 08:56:28,781][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 0.0745009183883667, 'train/grad_norm_pre_clip': 5.012874603271484, 'timing/data': 0.036123860627412796, 'timing/model': 1.4236446879804134, 'learning_rate/qwen_vl_interface': 5.529023838568566e-07, 'learning_rate/action_model': 5.529023838568566e-06, 'epoch': 1.84}) +[2026-06-16 08:56:34,677][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.09110219031572342, 'train/grad_norm_pre_clip': 20.780136108398438, 'timing/data': 0.02731768786907196, 'timing/model': 1.4365278482437134, 'learning_rate/qwen_vl_interface': 5.524500963627239e-07, 'learning_rate/action_model': 5.524500963627239e-06, 'epoch': 1.85}) +[2026-06-16 08:56:40,521][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 0.0008901113760657609, 'train/grad_norm_pre_clip': 0.15406155586242676, 'timing/data': 0.02709081955254078, 'timing/model': 1.5648707412183285, 'learning_rate/qwen_vl_interface': 5.519997398539145e-07, 'learning_rate/action_model': 5.519997398539145e-06, 'epoch': 1.85}) +[2026-06-16 08:56:46,179][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.023768942803144455, 'train/grad_norm_pre_clip': 4.1571831703186035, 'timing/data': 0.027235083281993866, 'timing/model': 1.4358254913240671, 'learning_rate/qwen_vl_interface': 5.515513145155531e-07, 'learning_rate/action_model': 5.515513145155531e-06, 'epoch': 1.85}) +[2026-06-16 08:56:52,081][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.0057776207104325294, 'train/grad_norm_pre_clip': 0.9857482314109802, 'timing/data': 0.032051824033260345, 'timing/model': 1.4311182741075754, 'learning_rate/qwen_vl_interface': 5.511048205319702e-07, 'learning_rate/action_model': 5.511048205319701e-06, 'epoch': 1.85}) +[2026-06-16 08:56:58,169][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 0.14781717956066132, 'train/grad_norm_pre_clip': 11.943864822387695, 'timing/data': 0.033146535977721214, 'timing/model': 1.5653345733880997, 'learning_rate/qwen_vl_interface': 5.506602580867034e-07, 'learning_rate/action_model': 5.506602580867034e-06, 'epoch': 1.85}) +[2026-06-16 08:57:04,338][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.004341277293860912, 'train/grad_norm_pre_clip': 1.5379300117492676, 'timing/data': 0.03219393081963062, 'timing/model': 1.496700905263424, 'learning_rate/qwen_vl_interface': 5.502176273624942e-07, 'learning_rate/action_model': 5.502176273624942e-06, 'epoch': 1.85}) +[2026-06-16 08:57:10,759][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.048321112990379333, 'train/grad_norm_pre_clip': 14.435473442077637, 'timing/data': 0.026622628793120384, 'timing/model': 1.6801692917943, 'learning_rate/qwen_vl_interface': 5.497769285412921e-07, 'learning_rate/action_model': 5.497769285412921e-06, 'epoch': 1.85}) +[2026-06-16 08:57:16,520][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.03132481873035431, 'train/grad_norm_pre_clip': 11.681421279907227, 'timing/data': 0.036170559003949165, 'timing/model': 1.5191252417862415, 'learning_rate/qwen_vl_interface': 5.493381618042519e-07, 'learning_rate/action_model': 5.493381618042518e-06, 'epoch': 1.85}) +[2026-06-16 08:57:22,746][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.0059545207768678665, 'train/grad_norm_pre_clip': 5.874473571777344, 'timing/data': 0.03165535070002079, 'timing/model': 1.5067517068237066, 'learning_rate/qwen_vl_interface': 5.489013273317343e-07, 'learning_rate/action_model': 5.489013273317343e-06, 'epoch': 1.85}) +[2026-06-16 08:57:28,842][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.002813967177644372, 'train/grad_norm_pre_clip': 0.370363712310791, 'timing/data': 0.03200782090425491, 'timing/model': 1.5366112478077412, 'learning_rate/qwen_vl_interface': 5.484664253033042e-07, 'learning_rate/action_model': 5.484664253033042e-06, 'epoch': 1.85}) +[2026-06-16 08:57:35,073][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.04728580638766289, 'train/grad_norm_pre_clip': 15.111295700073242, 'timing/data': 0.031263455748558044, 'timing/model': 1.5412644352763891, 'learning_rate/qwen_vl_interface': 5.480334558977351e-07, 'learning_rate/action_model': 5.480334558977351e-06, 'epoch': 1.85}) +[2026-06-16 08:57:41,363][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.0269601009786129, 'train/grad_norm_pre_clip': 11.830449104309082, 'timing/data': 0.032309675589203835, 'timing/model': 1.5659992936998606, 'learning_rate/qwen_vl_interface': 5.476024192930029e-07, 'learning_rate/action_model': 5.47602419293003e-06, 'epoch': 1.85}) +[2026-06-16 08:57:47,578][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.045390818268060684, 'train/grad_norm_pre_clip': 8.327205657958984, 'timing/data': 0.029958851635456085, 'timing/model': 1.7367195077240467, 'learning_rate/qwen_vl_interface': 5.471733156662914e-07, 'learning_rate/action_model': 5.4717331566629135e-06, 'epoch': 1.85}) +[2026-06-16 08:57:53,675][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 0.0038204011507332325, 'train/grad_norm_pre_clip': 1.2999885082244873, 'timing/data': 0.032387517392635345, 'timing/model': 1.5259573142975569, 'learning_rate/qwen_vl_interface': 5.467461451939882e-07, 'learning_rate/action_model': 5.467461451939882e-06, 'epoch': 1.85}) +[2026-06-16 08:58:00,075][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.1498122364282608, 'train/grad_norm_pre_clip': 27.00992202758789, 'timing/data': 0.03524800390005112, 'timing/model': 1.5272260718047619, 'learning_rate/qwen_vl_interface': 5.463209080516873e-07, 'learning_rate/action_model': 5.463209080516873e-06, 'epoch': 1.85}) +[2026-06-16 08:58:06,365][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.10784284770488739, 'train/grad_norm_pre_clip': 23.68097686767578, 'timing/data': 0.030879219993948936, 'timing/model': 1.5129726696759462, 'learning_rate/qwen_vl_interface': 5.458976044141875e-07, 'learning_rate/action_model': 5.458976044141875e-06, 'epoch': 1.85}) +[2026-06-16 08:58:12,579][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.004171076230704784, 'train/grad_norm_pre_clip': 2.417278528213501, 'timing/data': 0.0312917809933424, 'timing/model': 1.537738274782896, 'learning_rate/qwen_vl_interface': 5.454762344554931e-07, 'learning_rate/action_model': 5.4547623445549316e-06, 'epoch': 1.85}) +[2026-06-16 08:58:18,862][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 4.518303830991499e-05, 'train/grad_norm_pre_clip': 0.00862742867320776, 'timing/data': 0.03224384784698486, 'timing/model': 1.5234206076711416, 'learning_rate/qwen_vl_interface': 5.450567983488137e-07, 'learning_rate/action_model': 5.450567983488137e-06, 'epoch': 1.85}) +[2026-06-16 08:58:25,074][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.12742339074611664, 'train/grad_norm_pre_clip': 9.209356307983398, 'timing/data': 0.031631264835596085, 'timing/model': 1.744164114817977, 'learning_rate/qwen_vl_interface': 5.446392962665629e-07, 'learning_rate/action_model': 5.4463929626656296e-06, 'epoch': 1.85}) +[2026-06-16 08:58:31,159][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.005380481947213411, 'train/grad_norm_pre_clip': 0.589569628238678, 'timing/data': 0.03142003156244755, 'timing/model': 1.5096274241805077, 'learning_rate/qwen_vl_interface': 5.442237283803605e-07, 'learning_rate/action_model': 5.442237283803604e-06, 'epoch': 1.85}) +[2026-06-16 08:58:37,363][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.0701618418097496, 'train/grad_norm_pre_clip': 8.637049674987793, 'timing/data': 0.031935347244143486, 'timing/model': 1.5224575065076351, 'learning_rate/qwen_vl_interface': 5.43810094861031e-07, 'learning_rate/action_model': 5.43810094861031e-06, 'epoch': 1.85}) +[2026-06-16 08:58:43,564][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.09195384383201599, 'train/grad_norm_pre_clip': 24.199010848999023, 'timing/data': 0.03143983148038387, 'timing/model': 1.5371689051389694, 'learning_rate/qwen_vl_interface': 5.433983958786027e-07, 'learning_rate/action_model': 5.433983958786027e-06, 'epoch': 1.85}) +[2026-06-16 08:58:49,768][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.2560468912124634, 'train/grad_norm_pre_clip': 13.578433990478516, 'timing/data': 0.0317107904702425, 'timing/model': 1.5004356130957603, 'learning_rate/qwen_vl_interface': 5.429886316023102e-07, 'learning_rate/action_model': 5.429886316023101e-06, 'epoch': 1.85}) +[2026-06-16 08:58:55,968][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.19948308169841766, 'train/grad_norm_pre_clip': 14.430486679077148, 'timing/data': 0.034170132130384445, 'timing/model': 1.5383948683738708, 'learning_rate/qwen_vl_interface': 5.425808022005916e-07, 'learning_rate/action_model': 5.425808022005916e-06, 'epoch': 1.85}) +[2026-06-16 08:59:02,176][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.0020271025132387877, 'train/grad_norm_pre_clip': 0.5392800569534302, 'timing/data': 0.03120596520602703, 'timing/model': 1.7387233655899763, 'learning_rate/qwen_vl_interface': 5.421749078410903e-07, 'learning_rate/action_model': 5.421749078410903e-06, 'epoch': 1.85}) +[2026-06-16 08:59:08,272][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.004808774683624506, 'train/grad_norm_pre_clip': 1.6051329374313354, 'timing/data': 0.031458357349038124, 'timing/model': 1.5276098269969225, 'learning_rate/qwen_vl_interface': 5.417709486906547e-07, 'learning_rate/action_model': 5.417709486906547e-06, 'epoch': 1.85}) +[2026-06-16 08:59:14,565][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.09953872859477997, 'train/grad_norm_pre_clip': 4.364861011505127, 'timing/data': 0.032584188506007195, 'timing/model': 1.5226338636130095, 'learning_rate/qwen_vl_interface': 5.413689249153367e-07, 'learning_rate/action_model': 5.413689249153366e-06, 'epoch': 1.86}) +[2026-06-16 08:59:21,564][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.012546371668577194, 'train/grad_norm_pre_clip': 5.583584785461426, 'timing/data': 0.03196503967046738, 'timing/model': 1.5407290309667587, 'learning_rate/qwen_vl_interface': 5.409688366803931e-07, 'learning_rate/action_model': 5.409688366803931e-06, 'epoch': 1.86}) +[2026-06-16 08:59:27,661][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.03913503140211105, 'train/grad_norm_pre_clip': 7.503859519958496, 'timing/data': 0.0310567244887352, 'timing/model': 1.527691164985299, 'learning_rate/qwen_vl_interface': 5.405706841502853e-07, 'learning_rate/action_model': 5.4057068415028525e-06, 'epoch': 1.86}) +[2026-06-16 08:59:34,069][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 0.10120023041963577, 'train/grad_norm_pre_clip': 7.70676326751709, 'timing/data': 0.030821800231933594, 'timing/model': 1.5182224549353123, 'learning_rate/qwen_vl_interface': 5.401744674886781e-07, 'learning_rate/action_model': 5.40174467488678e-06, 'epoch': 1.86}) +[2026-06-16 08:59:40,270][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.059415049850940704, 'train/grad_norm_pre_clip': 11.679508209228516, 'timing/data': 0.031000815331935883, 'timing/model': 1.7398461811244488, 'learning_rate/qwen_vl_interface': 5.397801868584418e-07, 'learning_rate/action_model': 5.397801868584417e-06, 'epoch': 1.86}) +[2026-06-16 08:59:46,267][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.0010160076199099422, 'train/grad_norm_pre_clip': 0.31929922103881836, 'timing/data': 0.03168438374996185, 'timing/model': 1.535579226911068, 'learning_rate/qwen_vl_interface': 5.393878424216502e-07, 'learning_rate/action_model': 5.393878424216501e-06, 'epoch': 1.86}) +[2026-06-16 09:07:06,441][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.04587344080209732, 'train/grad_norm_pre_clip': 10.843644142150879, 'eval/loss': 0.044175341725349426, 'eval/flap_precision': 0.9582772543741588, 'eval/flap_recall': 0.966078697421981, 'eval/flap_f1': 0.9621621621621621, 'eval/flap_rate_model': 0.2063888888888889, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9621621621621622, 'timing/data': 0.0313727967441082, 'timing/model': 1.5349065866321325, 'learning_rate/qwen_vl_interface': 5.389974343395813e-07, 'learning_rate/action_model': 5.389974343395812e-06, 'epoch': 1.86}) +[2026-06-16 09:07:06,442][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4800_state +[2026-06-16 09:07:06,446][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 09:07:12,527][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4800_state/model.safetensors +[2026-06-16 09:07:27,050][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4800_state/optimizer.bin +[2026-06-16 09:07:27,050][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4800_state/sampler.bin +[2026-06-16 09:07:27,050][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4800_state/sampler_1.bin +[2026-06-16 09:07:27,051][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4800_state/random_states_0.pkl +[2026-06-16 09:07:27,151][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4800 +[2026-06-16 09:07:27,151][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 09:07:27,159][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 09:07:32,789][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 0.004694295581430197, 'train/grad_norm_pre_clip': 0.3786761462688446, 'timing/data': 0.02739734761416912, 'timing/model': 1.4301952123641968, 'learning_rate/qwen_vl_interface': 5.38608962772717e-07, 'learning_rate/action_model': 5.38608962772717e-06, 'epoch': 1.86}) +[2026-06-16 09:07:39,047][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.26854878664016724, 'train/grad_norm_pre_clip': 33.11454391479492, 'timing/data': 0.0313704963773489, 'timing/model': 1.53477924130857, 'learning_rate/qwen_vl_interface': 5.382224278807429e-07, 'learning_rate/action_model': 5.382224278807429e-06, 'epoch': 1.86}) +[2026-06-16 09:07:45,095][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.08361062407493591, 'train/grad_norm_pre_clip': 12.956669807434082, 'timing/data': 0.02854778803884983, 'timing/model': 1.433520583435893, 'learning_rate/qwen_vl_interface': 5.378378298225496e-07, 'learning_rate/action_model': 5.378378298225496e-06, 'epoch': 1.86}) +[2026-06-16 09:07:50,998][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.038067951798439026, 'train/grad_norm_pre_clip': 5.533152103424072, 'timing/data': 0.028205137699842453, 'timing/model': 1.447894912213087, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 1.86}) +[2026-06-16 09:07:56,800][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.05019213631749153, 'train/grad_norm_pre_clip': 31.4198055267334, 'timing/data': 0.026611771434545517, 'timing/model': 1.4467602036893368, 'learning_rate/qwen_vl_interface': 5.37074444839083e-07, 'learning_rate/action_model': 5.370744448390829e-06, 'epoch': 1.86}) +[2026-06-16 09:08:02,698][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 0.015369528904557228, 'train/grad_norm_pre_clip': 5.626628875732422, 'timing/data': 0.0322505421936512, 'timing/model': 1.6530073955655098, 'learning_rate/qwen_vl_interface': 5.366956582276085e-07, 'learning_rate/action_model': 5.366956582276084e-06, 'epoch': 1.86}) +[2026-06-16 09:08:08,560][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.009125304408371449, 'train/grad_norm_pre_clip': 1.3065754175186157, 'timing/data': 0.03277331590652466, 'timing/model': 1.5499357394874096, 'learning_rate/qwen_vl_interface': 5.363188090775122e-07, 'learning_rate/action_model': 5.363188090775122e-06, 'epoch': 1.86}) +[2026-06-16 09:08:14,858][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.03865819796919823, 'train/grad_norm_pre_clip': 9.090444564819336, 'timing/data': 0.03295401856303215, 'timing/model': 1.5330457985401154, 'learning_rate/qwen_vl_interface': 5.359438975437018e-07, 'learning_rate/action_model': 5.359438975437017e-06, 'epoch': 1.86}) +[2026-06-16 09:08:21,165][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.0011247512884438038, 'train/grad_norm_pre_clip': 0.1749633103609085, 'timing/data': 0.0358512457460165, 'timing/model': 1.5317969676107168, 'learning_rate/qwen_vl_interface': 5.355709237802895e-07, 'learning_rate/action_model': 5.3557092378028955e-06, 'epoch': 1.86}) +[2026-06-16 09:08:27,679][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.185830220580101, 'train/grad_norm_pre_clip': 13.6353178024292, 'timing/data': 0.03232368640601635, 'timing/model': 1.5417387522757053, 'learning_rate/qwen_vl_interface': 5.351998879405911e-07, 'learning_rate/action_model': 5.351998879405911e-06, 'epoch': 1.86}) +[2026-06-16 09:08:34,186][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.007013794034719467, 'train/grad_norm_pre_clip': 2.2196874618530273, 'timing/data': 0.031435100361704826, 'timing/model': 1.7565231211483479, 'learning_rate/qwen_vl_interface': 5.348307901771256e-07, 'learning_rate/action_model': 5.348307901771256e-06, 'epoch': 1.86}) +[2026-06-16 09:08:40,437][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.003628023434430361, 'train/grad_norm_pre_clip': 0.4187505841255188, 'timing/data': 0.03531778044998646, 'timing/model': 1.7066348996013403, 'learning_rate/qwen_vl_interface': 5.344636306416147e-07, 'learning_rate/action_model': 5.344636306416146e-06, 'epoch': 1.86}) +[2026-06-16 09:08:46,464][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.0911661833524704, 'train/grad_norm_pre_clip': 27.396533966064453, 'timing/data': 0.036464691162109375, 'timing/model': 1.5263223759829998, 'learning_rate/qwen_vl_interface': 5.340984094849836e-07, 'learning_rate/action_model': 5.340984094849836e-06, 'epoch': 1.86}) +[2026-06-16 09:08:52,683][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.02336544543504715, 'train/grad_norm_pre_clip': 16.120820999145508, 'timing/data': 0.03158881142735481, 'timing/model': 1.5531931761652231, 'learning_rate/qwen_vl_interface': 5.337351268573606e-07, 'learning_rate/action_model': 5.337351268573606e-06, 'epoch': 1.86}) +[2026-06-16 09:08:59,008][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.07483035326004028, 'train/grad_norm_pre_clip': 15.554804801940918, 'timing/data': 0.037753963842988014, 'timing/model': 1.5604855753481388, 'learning_rate/qwen_vl_interface': 5.33373782908079e-07, 'learning_rate/action_model': 5.333737829080789e-06, 'epoch': 1.86}) +[2026-06-16 09:09:05,384][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.0029115616343915462, 'train/grad_norm_pre_clip': 0.4014102816581726, 'timing/data': 0.03360961750149727, 'timing/model': 1.5396875534206629, 'learning_rate/qwen_vl_interface': 5.330143777856718e-07, 'learning_rate/action_model': 5.330143777856718e-06, 'epoch': 1.86}) +[2026-06-16 09:09:11,672][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 0.005251491442322731, 'train/grad_norm_pre_clip': 2.563852310180664, 'timing/data': 0.05643260106444359, 'timing/model': 1.5082133319228888, 'learning_rate/qwen_vl_interface': 5.326569116378786e-07, 'learning_rate/action_model': 5.326569116378786e-06, 'epoch': 1.86}) +[2026-06-16 09:09:18,049][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.0004969116416759789, 'train/grad_norm_pre_clip': 0.08227380365133286, 'timing/data': 0.031607264652848244, 'timing/model': 1.7237639874219894, 'learning_rate/qwen_vl_interface': 5.323013846116387e-07, 'learning_rate/action_model': 5.323013846116386e-06, 'epoch': 1.86}) +[2026-06-16 09:09:24,062][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.09142563492059708, 'train/grad_norm_pre_clip': 15.782856941223145, 'timing/data': 0.04195578396320343, 'timing/model': 1.525585725903511, 'learning_rate/qwen_vl_interface': 5.31947796853097e-07, 'learning_rate/action_model': 5.3194779685309696e-06, 'epoch': 1.87}) +[2026-06-16 09:09:30,236][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.21089217066764832, 'train/grad_norm_pre_clip': 27.397802352905273, 'timing/data': 0.041689034551382065, 'timing/model': 1.4938393123447895, 'learning_rate/qwen_vl_interface': 5.315961485075992e-07, 'learning_rate/action_model': 5.315961485075993e-06, 'epoch': 1.87}) +[2026-06-16 09:09:36,334][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.1099424809217453, 'train/grad_norm_pre_clip': 20.96129035949707, 'timing/data': 0.04286020062863827, 'timing/model': 1.5264316722750664, 'learning_rate/qwen_vl_interface': 5.312464397196958e-07, 'learning_rate/action_model': 5.312464397196958e-06, 'epoch': 1.87}) +[2026-06-16 09:09:42,275][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.12907981872558594, 'train/grad_norm_pre_clip': 11.999201774597168, 'timing/data': 0.03877136670053005, 'timing/model': 1.5084694251418114, 'learning_rate/qwen_vl_interface': 5.308986706331377e-07, 'learning_rate/action_model': 5.308986706331376e-06, 'epoch': 1.87}) +[2026-06-16 09:09:48,664][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.01794213429093361, 'train/grad_norm_pre_clip': 2.204927444458008, 'timing/data': 0.04187941551208496, 'timing/model': 1.53721272200346, 'learning_rate/qwen_vl_interface': 5.305528413908797e-07, 'learning_rate/action_model': 5.305528413908798e-06, 'epoch': 1.87}) +[2026-06-16 09:09:54,879][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.002783677540719509, 'train/grad_norm_pre_clip': 0.8017622828483582, 'timing/data': 0.04155168682336807, 'timing/model': 1.7407169006764889, 'learning_rate/qwen_vl_interface': 5.302089521350799e-07, 'learning_rate/action_model': 5.3020895213507996e-06, 'epoch': 1.87}) +[2026-06-16 09:10:00,968][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.029080189764499664, 'train/grad_norm_pre_clip': 7.8947954177856445, 'timing/data': 0.04308925196528435, 'timing/model': 1.528274169191718, 'learning_rate/qwen_vl_interface': 5.298670030070977e-07, 'learning_rate/action_model': 5.298670030070976e-06, 'epoch': 1.87}) +[2026-06-16 09:10:07,271][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.029039572924375534, 'train/grad_norm_pre_clip': 7.9932026863098145, 'timing/data': 0.04187106341123581, 'timing/model': 1.5423343535512686, 'learning_rate/qwen_vl_interface': 5.295269941474957e-07, 'learning_rate/action_model': 5.295269941474957e-06, 'epoch': 1.87}) +[2026-06-16 09:10:13,467][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.05636338144540787, 'train/grad_norm_pre_clip': 11.90006160736084, 'timing/data': 0.042986929416656494, 'timing/model': 1.5259296242147684, 'learning_rate/qwen_vl_interface': 5.291889256960384e-07, 'learning_rate/action_model': 5.291889256960384e-06, 'epoch': 1.87}) +[2026-06-16 09:10:19,968][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.00298582436516881, 'train/grad_norm_pre_clip': 0.43654006719589233, 'timing/data': 0.0420349296182394, 'timing/model': 1.5092963874340057, 'learning_rate/qwen_vl_interface': 5.288527977916929e-07, 'learning_rate/action_model': 5.288527977916929e-06, 'epoch': 1.87}) +[2026-06-16 09:10:26,319][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.007764194160699844, 'train/grad_norm_pre_clip': 3.8649046421051025, 'timing/data': 0.036753756925463676, 'timing/model': 1.7054691947996616, 'learning_rate/qwen_vl_interface': 5.28518610572629e-07, 'learning_rate/action_model': 5.28518610572629e-06, 'epoch': 1.87}) +[2026-06-16 09:10:32,031][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.0007456640596501529, 'train/grad_norm_pre_clip': 0.18282872438430786, 'timing/data': 0.032841211184859276, 'timing/model': 1.4571137242019176, 'learning_rate/qwen_vl_interface': 5.281863641762184e-07, 'learning_rate/action_model': 5.281863641762184e-06, 'epoch': 1.87}) +[2026-06-16 09:10:38,018][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 0.0008015265921130776, 'train/grad_norm_pre_clip': 0.14141596853733063, 'timing/data': 0.033413130789995193, 'timing/model': 1.538295242935419, 'learning_rate/qwen_vl_interface': 5.278560587390355e-07, 'learning_rate/action_model': 5.278560587390355e-06, 'epoch': 1.87}) +[2026-06-16 09:10:44,030][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.06621103733778, 'train/grad_norm_pre_clip': 7.626757621765137, 'timing/data': 0.03763406723737717, 'timing/model': 1.5429289527237415, 'learning_rate/qwen_vl_interface': 5.275276943968557e-07, 'learning_rate/action_model': 5.275276943968557e-06, 'epoch': 1.87}) +[2026-06-16 09:10:49,934][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.06922594457864761, 'train/grad_norm_pre_clip': 4.132960319519043, 'timing/data': 0.03455718792974949, 'timing/model': 1.4567866586148739, 'learning_rate/qwen_vl_interface': 5.272012712846566e-07, 'learning_rate/action_model': 5.2720127128465654e-06, 'epoch': 1.87}) +[2026-06-16 09:10:55,940][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.0157223641872406, 'train/grad_norm_pre_clip': 7.48493766784668, 'timing/data': 0.03986169211566448, 'timing/model': 1.5615099016577005, 'learning_rate/qwen_vl_interface': 5.268767895366198e-07, 'learning_rate/action_model': 5.268767895366198e-06, 'epoch': 1.87}) +[2026-06-16 09:11:01,982][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.0028873267583549023, 'train/grad_norm_pre_clip': 0.4309980571269989, 'timing/data': 0.033074649050831795, 'timing/model': 1.7034237794578075, 'learning_rate/qwen_vl_interface': 5.265542492861268e-07, 'learning_rate/action_model': 5.2655424928612676e-06, 'epoch': 1.87}) +[2026-06-16 09:11:07,689][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 0.012162226252257824, 'train/grad_norm_pre_clip': 2.957008123397827, 'timing/data': 0.03301401063799858, 'timing/model': 1.4132920484989882, 'learning_rate/qwen_vl_interface': 5.262336506657617e-07, 'learning_rate/action_model': 5.2623365066576165e-06, 'epoch': 1.87}) +[2026-06-16 09:11:13,577][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.012338375672698021, 'train/grad_norm_pre_clip': 3.442146062850952, 'timing/data': 0.03424454666674137, 'timing/model': 1.4271727316081524, 'learning_rate/qwen_vl_interface': 5.259149938073103e-07, 'learning_rate/action_model': 5.259149938073102e-06, 'epoch': 1.87}) +[2026-06-16 09:11:19,377][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.020201101899147034, 'train/grad_norm_pre_clip': 5.183985233306885, 'timing/data': 0.0343187041580677, 'timing/model': 1.4180654734373093, 'learning_rate/qwen_vl_interface': 5.255982788417609e-07, 'learning_rate/action_model': 5.255982788417608e-06, 'epoch': 1.87}) +[2026-06-16 09:11:25,183][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.07741212099790573, 'train/grad_norm_pre_clip': 6.130545616149902, 'timing/data': 0.033698612824082375, 'timing/model': 1.4350920878350735, 'learning_rate/qwen_vl_interface': 5.252835058993029e-07, 'learning_rate/action_model': 5.252835058993029e-06, 'epoch': 1.87}) +[2026-06-16 09:11:31,079][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.05620958283543587, 'train/grad_norm_pre_clip': 14.0083646774292, 'timing/data': 0.03354961983859539, 'timing/model': 1.4519505370408297, 'learning_rate/qwen_vl_interface': 5.249706751093274e-07, 'learning_rate/action_model': 5.249706751093273e-06, 'epoch': 1.87}) +[2026-06-16 09:11:36,949][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 0.0003434007230680436, 'train/grad_norm_pre_clip': 0.07925064861774445, 'timing/data': 0.03315779007971287, 'timing/model': 1.6019912995398045, 'learning_rate/qwen_vl_interface': 5.246597866004273e-07, 'learning_rate/action_model': 5.246597866004273e-06, 'epoch': 1.87}) +[2026-06-16 09:11:42,681][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.11193247139453888, 'train/grad_norm_pre_clip': 10.445540428161621, 'timing/data': 0.0333685576915741, 'timing/model': 1.4365291856229305, 'learning_rate/qwen_vl_interface': 5.243508405003975e-07, 'learning_rate/action_model': 5.243508405003975e-06, 'epoch': 1.87}) +[2026-06-16 09:11:48,483][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.005916232708841562, 'train/grad_norm_pre_clip': 1.598495602607727, 'timing/data': 0.03413878381252289, 'timing/model': 1.4329008050262928, 'learning_rate/qwen_vl_interface': 5.240438369362336e-07, 'learning_rate/action_model': 5.2404383693623366e-06, 'epoch': 1.87}) +[2026-06-16 09:11:54,183][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.004829484038054943, 'train/grad_norm_pre_clip': 1.0582391023635864, 'timing/data': 0.02682037092745304, 'timing/model': 1.4425257239490747, 'learning_rate/qwen_vl_interface': 5.237387760341335e-07, 'learning_rate/action_model': 5.237387760341335e-06, 'epoch': 1.87}) +[2026-06-16 09:11:59,886][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.004603174049407244, 'train/grad_norm_pre_clip': 1.2721059322357178, 'timing/data': 0.029162723571062088, 'timing/model': 1.4304924309253693, 'learning_rate/qwen_vl_interface': 5.234356579194963e-07, 'learning_rate/action_model': 5.234356579194963e-06, 'epoch': 1.88}) +[2026-06-16 09:12:05,682][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.01413918100297451, 'train/grad_norm_pre_clip': 5.876669883728027, 'timing/data': 0.0254436656832695, 'timing/model': 1.4592619314789772, 'learning_rate/qwen_vl_interface': 5.231344827169218e-07, 'learning_rate/action_model': 5.231344827169217e-06, 'epoch': 1.88}) +[2026-06-16 09:12:11,556][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.11223595589399338, 'train/grad_norm_pre_clip': 12.627495765686035, 'timing/data': 0.026850076392292976, 'timing/model': 1.6102679129689932, 'learning_rate/qwen_vl_interface': 5.228352505502119e-07, 'learning_rate/action_model': 5.228352505502119e-06, 'epoch': 1.88}) +[2026-06-16 09:12:17,181][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.002770602237433195, 'train/grad_norm_pre_clip': 1.3153215646743774, 'timing/data': 0.02675023302435875, 'timing/model': 1.4379108920693398, 'learning_rate/qwen_vl_interface': 5.225379615423701e-07, 'learning_rate/action_model': 5.2253796154237e-06, 'epoch': 1.88}) +[2026-06-16 09:12:22,984][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.07797576487064362, 'train/grad_norm_pre_clip': 43.829002380371094, 'timing/data': 0.02758861519396305, 'timing/model': 1.4398565478622913, 'learning_rate/qwen_vl_interface': 5.222426158156008e-07, 'learning_rate/action_model': 5.222426158156008e-06, 'epoch': 1.88}) +[2026-06-16 09:12:28,880][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.07529537379741669, 'train/grad_norm_pre_clip': 6.90803861618042, 'timing/data': 0.027256987988948822, 'timing/model': 1.438036810606718, 'learning_rate/qwen_vl_interface': 5.219492134913084e-07, 'learning_rate/action_model': 5.219492134913084e-06, 'epoch': 1.88}) +[2026-06-16 09:12:34,783][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.0022510758135467768, 'train/grad_norm_pre_clip': 0.5853313207626343, 'timing/data': 0.025562329217791557, 'timing/model': 1.458064828068018, 'learning_rate/qwen_vl_interface': 5.216577546901008e-07, 'learning_rate/action_model': 5.216577546901007e-06, 'epoch': 1.88}) +[2026-06-16 09:12:40,699][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 0.08404359221458435, 'train/grad_norm_pre_clip': 13.078969955444336, 'timing/data': 0.028233719989657402, 'timing/model': 1.6209493391215801, 'learning_rate/qwen_vl_interface': 5.213682395317852e-07, 'learning_rate/action_model': 5.213682395317851e-06, 'epoch': 1.88}) +[2026-06-16 09:12:47,145][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 0.10115871578454971, 'train/grad_norm_pre_clip': 14.875395774841309, 'timing/data': 0.026435619220137596, 'timing/model': 2.169108025729656, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 1.88}) +[2026-06-16 09:12:53,067][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.11692550778388977, 'train/grad_norm_pre_clip': 12.154293060302734, 'timing/data': 0.032087136059999466, 'timing/model': 1.5244709700345993, 'learning_rate/qwen_vl_interface': 5.207950406190654e-07, 'learning_rate/action_model': 5.207950406190653e-06, 'epoch': 1.88}) +[2026-06-16 09:12:59,268][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.11592038720846176, 'train/grad_norm_pre_clip': 13.371919631958008, 'timing/data': 0.036274779587984085, 'timing/model': 1.5177195128053427, 'learning_rate/qwen_vl_interface': 5.205113571002818e-07, 'learning_rate/action_model': 5.205113571002818e-06, 'epoch': 1.88}) +[2026-06-16 09:13:05,829][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.000412014895118773, 'train/grad_norm_pre_clip': 0.09545711427927017, 'timing/data': 0.030624007806181908, 'timing/model': 1.6083093360066414, 'learning_rate/qwen_vl_interface': 5.202296176956308e-07, 'learning_rate/action_model': 5.202296176956308e-06, 'epoch': 1.88}) +[2026-06-16 09:13:12,258][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 0.007048949599266052, 'train/grad_norm_pre_clip': 1.2508670091629028, 'timing/data': 0.031967466697096825, 'timing/model': 1.5924215037375689, 'learning_rate/qwen_vl_interface': 5.19949822520925e-07, 'learning_rate/action_model': 5.19949822520925e-06, 'epoch': 1.88}) +[2026-06-16 09:13:18,478][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.0025400202721357346, 'train/grad_norm_pre_clip': 0.4347573518753052, 'timing/data': 0.03210656903684139, 'timing/model': 1.5124297197908163, 'learning_rate/qwen_vl_interface': 5.196719716911777e-07, 'learning_rate/action_model': 5.1967197169117764e-06, 'epoch': 1.88}) +[2026-06-16 09:13:24,736][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.029716262593865395, 'train/grad_norm_pre_clip': 7.6857218742370605, 'timing/data': 0.03283506818115711, 'timing/model': 1.7014349233359098, 'learning_rate/qwen_vl_interface': 5.193960653206029e-07, 'learning_rate/action_model': 5.193960653206029e-06, 'epoch': 1.88}) +[2026-06-16 09:13:30,778][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.001148219220340252, 'train/grad_norm_pre_clip': 0.17375342547893524, 'timing/data': 0.03484959527850151, 'timing/model': 1.538744430989027, 'learning_rate/qwen_vl_interface': 5.191221035226144e-07, 'learning_rate/action_model': 5.191221035226144e-06, 'epoch': 1.88}) +[2026-06-16 09:13:37,076][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.08135098963975906, 'train/grad_norm_pre_clip': 26.675975799560547, 'timing/data': 0.03260476514697075, 'timing/model': 1.5436969362199306, 'learning_rate/qwen_vl_interface': 5.188500864098292e-07, 'learning_rate/action_model': 5.1885008640982915e-06, 'epoch': 1.88}) +[2026-06-16 09:13:43,237][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.17835009098052979, 'train/grad_norm_pre_clip': 23.152864456176758, 'timing/data': 0.03145116567611694, 'timing/model': 1.5323790330439806, 'learning_rate/qwen_vl_interface': 5.185800140940613e-07, 'learning_rate/action_model': 5.185800140940613e-06, 'epoch': 1.88}) +[2026-06-16 09:13:49,339][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 0.01985805854201317, 'train/grad_norm_pre_clip': 16.519245147705078, 'timing/data': 0.03188827447593212, 'timing/model': 1.5246095396578312, 'learning_rate/qwen_vl_interface': 5.183118866863288e-07, 'learning_rate/action_model': 5.1831188668632875e-06, 'epoch': 1.88}) +[2026-06-16 09:13:55,446][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.0011093519860878587, 'train/grad_norm_pre_clip': 0.2810172438621521, 'timing/data': 0.031558720394968987, 'timing/model': 1.7453537825495005, 'learning_rate/qwen_vl_interface': 5.180457042968484e-07, 'learning_rate/action_model': 5.1804570429684835e-06, 'epoch': 1.88}) +[2026-06-16 09:14:01,337][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.00015796646766830236, 'train/grad_norm_pre_clip': 0.02634623646736145, 'timing/data': 0.031407177448272705, 'timing/model': 1.5451127532869577, 'learning_rate/qwen_vl_interface': 5.17781467035037e-07, 'learning_rate/action_model': 5.177814670350371e-06, 'epoch': 1.88}) +[2026-06-16 09:14:07,438][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.02851124480366707, 'train/grad_norm_pre_clip': 12.025980949401855, 'timing/data': 0.032038187608122826, 'timing/model': 1.5407516304403543, 'learning_rate/qwen_vl_interface': 5.175191750095133e-07, 'learning_rate/action_model': 5.175191750095132e-06, 'epoch': 1.88}) +[2026-06-16 09:14:13,333][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.009215536527335644, 'train/grad_norm_pre_clip': 2.1569342613220215, 'timing/data': 0.031153762713074684, 'timing/model': 1.5216797795146704, 'learning_rate/qwen_vl_interface': 5.172588283280953e-07, 'learning_rate/action_model': 5.172588283280953e-06, 'epoch': 1.88}) +[2026-06-16 09:14:19,334][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.17259663343429565, 'train/grad_norm_pre_clip': 29.07216453552246, 'timing/data': 0.02964102104306221, 'timing/model': 1.5346679203212261, 'learning_rate/qwen_vl_interface': 5.170004270978019e-07, 'learning_rate/action_model': 5.170004270978018e-06, 'epoch': 1.88}) +[2026-06-16 09:14:25,333][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 0.20162473618984222, 'train/grad_norm_pre_clip': 20.88785171508789, 'timing/data': 0.030548343434929848, 'timing/model': 1.4857260007411242, 'learning_rate/qwen_vl_interface': 5.167439714248512e-07, 'learning_rate/action_model': 5.1674397142485116e-06, 'epoch': 1.88}) +[2026-06-16 09:14:31,453][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.0037858369760215282, 'train/grad_norm_pre_clip': 0.9976491928100586, 'timing/data': 0.032629434019327164, 'timing/model': 1.774669075384736, 'learning_rate/qwen_vl_interface': 5.164894614146633e-07, 'learning_rate/action_model': 5.164894614146634e-06, 'epoch': 1.88}) +[2026-06-16 09:14:37,339][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.1327275037765503, 'train/grad_norm_pre_clip': 6.924399375915527, 'timing/data': 0.032609619200229645, 'timing/model': 1.5406929589807987, 'learning_rate/qwen_vl_interface': 5.16236897171858e-07, 'learning_rate/action_model': 5.16236897171858e-06, 'epoch': 1.89}) +[2026-06-16 09:14:43,338][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.0005119962734170258, 'train/grad_norm_pre_clip': 0.08998320996761322, 'timing/data': 0.03228803351521492, 'timing/model': 1.5282078683376312, 'learning_rate/qwen_vl_interface': 5.15986278800254e-07, 'learning_rate/action_model': 5.15986278800254e-06, 'epoch': 1.89}) +[2026-06-16 09:14:49,538][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.007199717685580254, 'train/grad_norm_pre_clip': 1.994841456413269, 'timing/data': 0.033402178436517715, 'timing/model': 1.5357838198542595, 'learning_rate/qwen_vl_interface': 5.157376064028719e-07, 'learning_rate/action_model': 5.157376064028718e-06, 'epoch': 1.89}) +[2026-06-16 09:14:55,677][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.0070336610078811646, 'train/grad_norm_pre_clip': 1.7966850996017456, 'timing/data': 0.03370071388781071, 'timing/model': 1.5647371597588062, 'learning_rate/qwen_vl_interface': 5.154908800819307e-07, 'learning_rate/action_model': 5.154908800819306e-06, 'epoch': 1.89}) +[2026-06-16 09:15:02,114][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.05790019780397415, 'train/grad_norm_pre_clip': 6.740362167358398, 'timing/data': 0.03211312182247639, 'timing/model': 1.5385107230395079, 'learning_rate/qwen_vl_interface': 5.152460999388504e-07, 'learning_rate/action_model': 5.152460999388503e-06, 'epoch': 1.89}) +[2026-06-16 09:15:08,544][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.00454030092805624, 'train/grad_norm_pre_clip': 0.8515712022781372, 'timing/data': 0.03204047493636608, 'timing/model': 1.8090564664453268, 'learning_rate/qwen_vl_interface': 5.150032660742517e-07, 'learning_rate/action_model': 5.150032660742517e-06, 'epoch': 1.89}) +[2026-06-16 09:15:14,580][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.013508381322026253, 'train/grad_norm_pre_clip': 4.53155517578125, 'timing/data': 0.032134346663951874, 'timing/model': 1.5518400445580482, 'learning_rate/qwen_vl_interface': 5.147623785879535e-07, 'learning_rate/action_model': 5.147623785879534e-06, 'epoch': 1.89}) +[2026-06-16 09:15:20,869][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.2099892646074295, 'train/grad_norm_pre_clip': 11.52662467956543, 'timing/data': 0.03201392851769924, 'timing/model': 1.5399438608437777, 'learning_rate/qwen_vl_interface': 5.145234375789758e-07, 'learning_rate/action_model': 5.145234375789758e-06, 'epoch': 1.89}) +[2026-06-16 09:15:27,162][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.015161058865487576, 'train/grad_norm_pre_clip': 4.123385429382324, 'timing/data': 0.03429365158081055, 'timing/model': 1.5432493314146996, 'learning_rate/qwen_vl_interface': 5.142864431455387e-07, 'learning_rate/action_model': 5.1428644314553875e-06, 'epoch': 1.89}) +[2026-06-16 09:15:33,524][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.11244791001081467, 'train/grad_norm_pre_clip': 10.882259368896484, 'timing/data': 0.03903377987444401, 'timing/model': 1.6383410040289164, 'learning_rate/qwen_vl_interface': 5.140513953850612e-07, 'learning_rate/action_model': 5.140513953850612e-06, 'epoch': 1.89}) +[2026-06-16 09:15:39,668][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.07442285120487213, 'train/grad_norm_pre_clip': 15.668371200561523, 'timing/data': 0.03335470147430897, 'timing/model': 1.5065296348184347, 'learning_rate/qwen_vl_interface': 5.138182943941623e-07, 'learning_rate/action_model': 5.138182943941623e-06, 'epoch': 1.89}) +[2026-06-16 09:15:45,981][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.004168751649558544, 'train/grad_norm_pre_clip': 0.9168540239334106, 'timing/data': 0.03202190808951855, 'timing/model': 1.7459311373531818, 'learning_rate/qwen_vl_interface': 5.135871402686615e-07, 'learning_rate/action_model': 5.135871402686615e-06, 'epoch': 1.89}) +[2026-06-16 09:15:52,072][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.03820161148905754, 'train/grad_norm_pre_clip': 8.072219848632812, 'timing/data': 0.05028534308075905, 'timing/model': 1.5229757782071829, 'learning_rate/qwen_vl_interface': 5.133579331035777e-07, 'learning_rate/action_model': 5.133579331035777e-06, 'epoch': 1.89}) +[2026-06-16 09:15:58,466][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.1450280398130417, 'train/grad_norm_pre_clip': 12.130781173706055, 'timing/data': 0.03446688503026962, 'timing/model': 1.5277985762804747, 'learning_rate/qwen_vl_interface': 5.131306729931284e-07, 'learning_rate/action_model': 5.1313067299312836e-06, 'epoch': 1.89}) +[2026-06-16 09:16:04,668][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.08041081577539444, 'train/grad_norm_pre_clip': 15.115256309509277, 'timing/data': 0.03228103555738926, 'timing/model': 1.5480520464479923, 'learning_rate/qwen_vl_interface': 5.129053600307326e-07, 'learning_rate/action_model': 5.129053600307326e-06, 'epoch': 1.89}) +[2026-06-16 09:16:10,664][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.14531615376472473, 'train/grad_norm_pre_clip': 15.43973159790039, 'timing/data': 0.031219428405165672, 'timing/model': 1.5315915923565626, 'learning_rate/qwen_vl_interface': 5.126819943090078e-07, 'learning_rate/action_model': 5.1268199430900775e-06, 'epoch': 1.89}) +[2026-06-16 09:16:16,862][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.17561954259872437, 'train/grad_norm_pre_clip': 25.17110824584961, 'timing/data': 0.03161824494600296, 'timing/model': 1.55124169588089, 'learning_rate/qwen_vl_interface': 5.124605759197706e-07, 'learning_rate/action_model': 5.124605759197706e-06, 'epoch': 1.89}) +[2026-06-16 09:16:23,077][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.014771886169910431, 'train/grad_norm_pre_clip': 6.893841743469238, 'timing/data': 0.0312038566917181, 'timing/model': 1.752670545130968, 'learning_rate/qwen_vl_interface': 5.122411049540383e-07, 'learning_rate/action_model': 5.1224110495403825e-06, 'epoch': 1.89}) +[2026-06-16 09:16:29,172][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.004664212930947542, 'train/grad_norm_pre_clip': 0.7255708575248718, 'timing/data': 0.03217090293765068, 'timing/model': 1.5423364900052547, 'learning_rate/qwen_vl_interface': 5.120235815020269e-07, 'learning_rate/action_model': 5.120235815020269e-06, 'epoch': 1.89}) +[2026-06-16 09:16:35,467][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.048947226256132126, 'train/grad_norm_pre_clip': 6.942065715789795, 'timing/data': 0.03213569335639477, 'timing/model': 1.5411899369210005, 'learning_rate/qwen_vl_interface': 5.118080056531524e-07, 'learning_rate/action_model': 5.118080056531524e-06, 'epoch': 1.89}) +[2026-06-16 09:16:41,675][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.043479327112436295, 'train/grad_norm_pre_clip': 16.94379234313965, 'timing/data': 0.031221015378832817, 'timing/model': 1.5529136378318071, 'learning_rate/qwen_vl_interface': 5.115943774960295e-07, 'learning_rate/action_model': 5.115943774960295e-06, 'epoch': 1.89}) +[2026-06-16 09:16:47,905][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.003578332718461752, 'train/grad_norm_pre_clip': 0.8612317442893982, 'timing/data': 0.03364185988903046, 'timing/model': 1.554566776379943, 'learning_rate/qwen_vl_interface': 5.113826971184727e-07, 'learning_rate/action_model': 5.113826971184727e-06, 'epoch': 1.89}) +[2026-06-16 09:16:54,402][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.003354925662279129, 'train/grad_norm_pre_clip': 1.8098297119140625, 'timing/data': 0.03166260942816734, 'timing/model': 1.7039492074400187, 'learning_rate/qwen_vl_interface': 5.111729646074954e-07, 'learning_rate/action_model': 5.111729646074954e-06, 'epoch': 1.89}) +[2026-06-16 09:17:00,771][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.025623368099331856, 'train/grad_norm_pre_clip': 7.157865047454834, 'timing/data': 0.03184136562049389, 'timing/model': 1.7395510002970695, 'learning_rate/qwen_vl_interface': 5.10965180049312e-07, 'learning_rate/action_model': 5.10965180049312e-06, 'epoch': 1.89}) +[2026-06-16 09:17:06,764][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.06548881530761719, 'train/grad_norm_pre_clip': 20.96876335144043, 'timing/data': 0.03076602704823017, 'timing/model': 1.5379612259566784, 'learning_rate/qwen_vl_interface': 5.107593435293336e-07, 'learning_rate/action_model': 5.107593435293336e-06, 'epoch': 1.89}) +[2026-06-16 09:17:12,969][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.28113365173339844, 'train/grad_norm_pre_clip': 31.758241653442383, 'timing/data': 0.031506191939115524, 'timing/model': 1.5416840314865112, 'learning_rate/qwen_vl_interface': 5.105554551321724e-07, 'learning_rate/action_model': 5.105554551321723e-06, 'epoch': 1.89}) +[2026-06-16 09:17:19,171][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.003537176176905632, 'train/grad_norm_pre_clip': 0.8603262901306152, 'timing/data': 0.031023211777210236, 'timing/model': 1.5444880370050669, 'learning_rate/qwen_vl_interface': 5.103535149416389e-07, 'learning_rate/action_model': 5.1035351494163885e-06, 'epoch': 1.9}) +[2026-06-16 09:17:25,370][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.0033880257979035378, 'train/grad_norm_pre_clip': 0.47945499420166016, 'timing/data': 0.030815210193395615, 'timing/model': 1.5408340152353048, 'learning_rate/qwen_vl_interface': 5.101535230407436e-07, 'learning_rate/action_model': 5.101535230407436e-06, 'epoch': 1.9}) +[2026-06-16 09:17:31,564][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.07196620106697083, 'train/grad_norm_pre_clip': 25.749666213989258, 'timing/data': 0.029947366565465927, 'timing/model': 1.5249743796885014, 'learning_rate/qwen_vl_interface': 5.099554795116956e-07, 'learning_rate/action_model': 5.0995547951169565e-06, 'epoch': 1.9}) +[2026-06-16 09:24:49,718][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.14350323379039764, 'train/grad_norm_pre_clip': 35.71126937866211, 'eval/loss': 0.04624339938163757, 'eval/flap_precision': 0.9558823529411765, 'eval/flap_recall': 0.9701492537313433, 'eval/flap_f1': 0.962962962962963, 'eval/flap_rate_model': 0.20777777777777778, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9629629629629631, 'timing/data': 0.03158704936504364, 'timing/model': 1.7596970275044441, 'learning_rate/qwen_vl_interface': 5.097593844359024e-07, 'learning_rate/action_model': 5.0975938443590245e-06, 'epoch': 1.9}) +[2026-06-16 09:24:49,719][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4900_state +[2026-06-16 09:24:49,724][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 09:24:57,989][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4900_state/model.safetensors +[2026-06-16 09:25:12,830][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4900_state/optimizer.bin +[2026-06-16 09:25:12,830][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4900_state/sampler.bin +[2026-06-16 09:25:12,831][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4900_state/sampler_1.bin +[2026-06-16 09:25:12,832][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_4900_state/random_states_0.pkl +[2026-06-16 09:25:12,960][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4900 +[2026-06-16 09:25:12,960][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 09:25:13,018][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 09:25:19,012][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.020321452990174294, 'train/grad_norm_pre_clip': 6.76624059677124, 'timing/data': 0.030658820644021034, 'timing/model': 1.500422164797783, 'learning_rate/qwen_vl_interface': 5.095652378939718e-07, 'learning_rate/action_model': 5.095652378939717e-06, 'epoch': 1.9}) +[2026-06-16 09:25:25,456][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.13082659244537354, 'train/grad_norm_pre_clip': 24.50665855407715, 'timing/data': 0.03345964103937149, 'timing/model': 1.7582951243966818, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 1.9}) +[2026-06-16 09:25:31,223][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.0010488121770322323, 'train/grad_norm_pre_clip': 0.16790297627449036, 'timing/data': 0.03226301074028015, 'timing/model': 1.5360845010727644, 'learning_rate/qwen_vl_interface': 5.091827907301227e-07, 'learning_rate/action_model': 5.091827907301227e-06, 'epoch': 1.9}) +[2026-06-16 09:25:37,359][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.0007294563692994416, 'train/grad_norm_pre_clip': 0.1942286491394043, 'timing/data': 0.02758532389998436, 'timing/model': 1.5816771034151316, 'learning_rate/qwen_vl_interface': 5.089944902654134e-07, 'learning_rate/action_model': 5.089944902654134e-06, 'epoch': 1.9}) +[2026-06-16 09:25:43,102][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.018228229135274887, 'train/grad_norm_pre_clip': 3.459596872329712, 'timing/data': 0.026274746283888817, 'timing/model': 1.440266052260995, 'learning_rate/qwen_vl_interface': 5.088081386489856e-07, 'learning_rate/action_model': 5.088081386489857e-06, 'epoch': 1.9}) +[2026-06-16 09:25:49,121][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.3180537819862366, 'train/grad_norm_pre_clip': 10.801629066467285, 'timing/data': 0.025590237230062485, 'timing/model': 1.535114299505949, 'learning_rate/qwen_vl_interface': 5.086237359574419e-07, 'learning_rate/action_model': 5.086237359574419e-06, 'epoch': 1.9}) +[2026-06-16 09:25:55,114][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.04169628769159317, 'train/grad_norm_pre_clip': 13.212882041931152, 'timing/data': 0.027246292680501938, 'timing/model': 1.531339069828391, 'learning_rate/qwen_vl_interface': 5.084412822665825e-07, 'learning_rate/action_model': 5.084412822665825e-06, 'epoch': 1.9}) +[2026-06-16 09:26:01,390][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.11459711194038391, 'train/grad_norm_pre_clip': 22.540281295776367, 'timing/data': 0.02572254091501236, 'timing/model': 1.7659500669687986, 'learning_rate/qwen_vl_interface': 5.082607776514081e-07, 'learning_rate/action_model': 5.082607776514081e-06, 'epoch': 1.9}) +[2026-06-16 09:26:07,315][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.002028561430051923, 'train/grad_norm_pre_clip': 0.3814249634742737, 'timing/data': 0.025177033618092537, 'timing/model': 1.5311350859701633, 'learning_rate/qwen_vl_interface': 5.080822221861167e-07, 'learning_rate/action_model': 5.080822221861166e-06, 'epoch': 1.9}) +[2026-06-16 09:26:13,212][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.20822782814502716, 'train/grad_norm_pre_clip': 8.084807395935059, 'timing/data': 0.02559327706694603, 'timing/model': 1.4489501044154167, 'learning_rate/qwen_vl_interface': 5.07905615944106e-07, 'learning_rate/action_model': 5.0790561594410595e-06, 'epoch': 1.9}) +[2026-06-16 09:26:19,080][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.08643301576375961, 'train/grad_norm_pre_clip': 15.698493003845215, 'timing/data': 0.027446234598755836, 'timing/model': 1.4196039326488972, 'learning_rate/qwen_vl_interface': 5.077309589979716e-07, 'learning_rate/action_model': 5.077309589979715e-06, 'epoch': 1.9}) +[2026-06-16 09:26:24,983][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.012032550759613514, 'train/grad_norm_pre_clip': 2.6156423091888428, 'timing/data': 0.028530290350317955, 'timing/model': 1.4265366680920124, 'learning_rate/qwen_vl_interface': 5.075582514195089e-07, 'learning_rate/action_model': 5.0755825141950895e-06, 'epoch': 1.9}) +[2026-06-16 09:26:30,959][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 0.007076175883412361, 'train/grad_norm_pre_clip': 1.92924165725708, 'timing/data': 0.028872961178421974, 'timing/model': 1.4849036168307066, 'learning_rate/qwen_vl_interface': 5.073874932797115e-07, 'learning_rate/action_model': 5.0738749327971155e-06, 'epoch': 1.9}) +[2026-06-16 09:26:36,777][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.16210485994815826, 'train/grad_norm_pre_clip': 14.550323486328125, 'timing/data': 0.02499948814511299, 'timing/model': 1.6294268257915974, 'learning_rate/qwen_vl_interface': 5.072186846487713e-07, 'learning_rate/action_model': 5.072186846487713e-06, 'epoch': 1.9}) +[2026-06-16 09:26:42,651][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 0.18727660179138184, 'train/grad_norm_pre_clip': 9.123250961303711, 'timing/data': 0.024862879887223244, 'timing/model': 1.6032762192189693, 'learning_rate/qwen_vl_interface': 5.070518255960796e-07, 'learning_rate/action_model': 5.070518255960795e-06, 'epoch': 1.9}) +[2026-06-16 09:26:48,438][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.12737931311130524, 'train/grad_norm_pre_clip': 9.280040740966797, 'timing/data': 0.031746745109558105, 'timing/model': 1.5410732477903366, 'learning_rate/qwen_vl_interface': 5.068869161902256e-07, 'learning_rate/action_model': 5.068869161902256e-06, 'epoch': 1.9}) +[2026-06-16 09:26:54,535][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.028760969638824463, 'train/grad_norm_pre_clip': 9.637725830078125, 'timing/data': 0.031267885118722916, 'timing/model': 1.535977728664875, 'learning_rate/qwen_vl_interface': 5.067239564989968e-07, 'learning_rate/action_model': 5.067239564989968e-06, 'epoch': 1.9}) +[2026-06-16 09:27:00,635][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 0.27146852016448975, 'train/grad_norm_pre_clip': 30.05897331237793, 'timing/data': 0.030766434967517853, 'timing/model': 1.5362143144011497, 'learning_rate/qwen_vl_interface': 5.065629465893808e-07, 'learning_rate/action_model': 5.065629465893807e-06, 'epoch': 1.9}) +[2026-06-16 09:27:06,733][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.0006008714553900063, 'train/grad_norm_pre_clip': 0.09419197589159012, 'timing/data': 0.031857239082455635, 'timing/model': 1.5217731911689043, 'learning_rate/qwen_vl_interface': 5.064038865275625e-07, 'learning_rate/action_model': 5.0640388652756245e-06, 'epoch': 1.9}) +[2026-06-16 09:27:13,151][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.0003315388457849622, 'train/grad_norm_pre_clip': 0.04318246990442276, 'timing/data': 0.030180571600794792, 'timing/model': 1.7417149804532528, 'learning_rate/qwen_vl_interface': 5.062467763789242e-07, 'learning_rate/action_model': 5.062467763789242e-06, 'epoch': 1.9}) +[2026-06-16 09:27:19,040][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.013993175700306892, 'train/grad_norm_pre_clip': 6.291609764099121, 'timing/data': 0.033003538846969604, 'timing/model': 1.5415131356567144, 'learning_rate/qwen_vl_interface': 5.060916162080496e-07, 'learning_rate/action_model': 5.060916162080496e-06, 'epoch': 1.9}) +[2026-06-16 09:27:25,238][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.0024951843079179525, 'train/grad_norm_pre_clip': 0.5302427411079407, 'timing/data': 0.03126552887260914, 'timing/model': 1.554413303732872, 'learning_rate/qwen_vl_interface': 5.059384060787182e-07, 'learning_rate/action_model': 5.059384060787183e-06, 'epoch': 1.9}) +[2026-06-16 09:27:31,334][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 0.001910957507789135, 'train/grad_norm_pre_clip': 0.5118570923805237, 'timing/data': 0.03029731474816799, 'timing/model': 1.5353384856134653, 'learning_rate/qwen_vl_interface': 5.05787146053909e-07, 'learning_rate/action_model': 5.0578714605390904e-06, 'epoch': 1.91}) +[2026-06-16 09:27:37,433][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.15506844222545624, 'train/grad_norm_pre_clip': 10.58411693572998, 'timing/data': 0.03077208437025547, 'timing/model': 1.5128011386841536, 'learning_rate/qwen_vl_interface': 5.056378361957998e-07, 'learning_rate/action_model': 5.056378361957998e-06, 'epoch': 1.91}) +[2026-06-16 09:27:43,433][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.186852365732193, 'train/grad_norm_pre_clip': 14.989877700805664, 'timing/data': 0.031654203310608864, 'timing/model': 1.482618598267436, 'learning_rate/qwen_vl_interface': 5.05490476565766e-07, 'learning_rate/action_model': 5.054904765657659e-06, 'epoch': 1.91}) +[2026-06-16 09:27:49,500][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.009031422436237335, 'train/grad_norm_pre_clip': 1.8936562538146973, 'timing/data': 0.03245856612920761, 'timing/model': 1.6837381925433874, 'learning_rate/qwen_vl_interface': 5.053450672243811e-07, 'learning_rate/action_model': 5.0534506722438114e-06, 'epoch': 1.91}) +[2026-06-16 09:27:55,437][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 0.00044970534509047866, 'train/grad_norm_pre_clip': 0.1408560872077942, 'timing/data': 0.03191426582634449, 'timing/model': 1.5394307170063257, 'learning_rate/qwen_vl_interface': 5.052016082314184e-07, 'learning_rate/action_model': 5.052016082314184e-06, 'epoch': 1.91}) +[2026-06-16 09:28:01,756][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.020046822726726532, 'train/grad_norm_pre_clip': 5.660635948181152, 'timing/data': 0.033361440524458885, 'timing/model': 1.74864761531353, 'learning_rate/qwen_vl_interface': 5.050600996458473e-07, 'learning_rate/action_model': 5.050600996458473e-06, 'epoch': 1.91}) +[2026-06-16 09:28:07,839][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 0.06408808380365372, 'train/grad_norm_pre_clip': 19.696386337280273, 'timing/data': 0.03337709978222847, 'timing/model': 1.534699097275734, 'learning_rate/qwen_vl_interface': 5.049205415258373e-07, 'learning_rate/action_model': 5.049205415258373e-06, 'epoch': 1.91}) +[2026-06-16 09:28:13,956][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.08618384599685669, 'train/grad_norm_pre_clip': 7.19127082824707, 'timing/data': 0.03275061212480068, 'timing/model': 1.5169767327606678, 'learning_rate/qwen_vl_interface': 5.047829339287557e-07, 'learning_rate/action_model': 5.047829339287557e-06, 'epoch': 1.91}) +[2026-06-16 09:28:20,561][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.002630142495036125, 'train/grad_norm_pre_clip': 0.519252598285675, 'timing/data': 0.030768556520342827, 'timing/model': 1.7634681966155767, 'learning_rate/qwen_vl_interface': 5.046472769111668e-07, 'learning_rate/action_model': 5.046472769111667e-06, 'epoch': 1.91}) +[2026-06-16 09:28:27,026][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.018460573628544807, 'train/grad_norm_pre_clip': 13.744940757751465, 'timing/data': 0.03303765878081322, 'timing/model': 1.7104904055595398, 'learning_rate/qwen_vl_interface': 5.045135705288353e-07, 'learning_rate/action_model': 5.0451357052883525e-06, 'epoch': 1.91}) +[2026-06-16 09:28:32,924][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.007774718105792999, 'train/grad_norm_pre_clip': 1.5638428926467896, 'timing/data': 0.031196359544992447, 'timing/model': 1.5448060482740402, 'learning_rate/qwen_vl_interface': 5.043818148367221e-07, 'learning_rate/action_model': 5.043818148367221e-06, 'epoch': 1.91}) +[2026-06-16 09:28:39,143][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.1983603537082672, 'train/grad_norm_pre_clip': 16.813207626342773, 'timing/data': 0.03245260939002037, 'timing/model': 1.5381691940128803, 'learning_rate/qwen_vl_interface': 5.042520098889869e-07, 'learning_rate/action_model': 5.0425200988898685e-06, 'epoch': 1.91}) +[2026-06-16 09:28:45,335][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.19712908565998077, 'train/grad_norm_pre_clip': 10.798043251037598, 'timing/data': 0.031217504292726517, 'timing/model': 1.5351136848330498, 'learning_rate/qwen_vl_interface': 5.041241557389879e-07, 'learning_rate/action_model': 5.041241557389879e-06, 'epoch': 1.91}) +[2026-06-16 09:28:51,536][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.002866297960281372, 'train/grad_norm_pre_clip': 1.2631394863128662, 'timing/data': 0.030800053849816322, 'timing/model': 1.5330837275832891, 'learning_rate/qwen_vl_interface': 5.039982524392814e-07, 'learning_rate/action_model': 5.039982524392814e-06, 'epoch': 1.91}) +[2026-06-16 09:28:57,844][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.008549542166292667, 'train/grad_norm_pre_clip': 3.8905622959136963, 'timing/data': 0.06166066415607929, 'timing/model': 1.7301416713744402, 'learning_rate/qwen_vl_interface': 5.038743000416207e-07, 'learning_rate/action_model': 5.0387430004162064e-06, 'epoch': 1.91}) +[2026-06-16 09:29:03,839][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.0005619045696221292, 'train/grad_norm_pre_clip': 0.16400355100631714, 'timing/data': 0.031210249289870262, 'timing/model': 1.5419198051095009, 'learning_rate/qwen_vl_interface': 5.037522985969588e-07, 'learning_rate/action_model': 5.037522985969587e-06, 'epoch': 1.91}) +[2026-06-16 09:29:10,039][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.00987872201949358, 'train/grad_norm_pre_clip': 4.014404296875, 'timing/data': 0.03318956308066845, 'timing/model': 1.5208071637898684, 'learning_rate/qwen_vl_interface': 5.036322481554451e-07, 'learning_rate/action_model': 5.036322481554451e-06, 'epoch': 1.91}) +[2026-06-16 09:29:16,135][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.05115313082933426, 'train/grad_norm_pre_clip': 8.255294799804688, 'timing/data': 0.031159592792391777, 'timing/model': 1.5373204927891493, 'learning_rate/qwen_vl_interface': 5.035141487664285e-07, 'learning_rate/action_model': 5.035141487664285e-06, 'epoch': 1.91}) +[2026-06-16 09:29:22,139][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.0014411464799195528, 'train/grad_norm_pre_clip': 0.41283631324768066, 'timing/data': 0.03160805068910122, 'timing/model': 1.5422268118709326, 'learning_rate/qwen_vl_interface': 5.03398000478455e-07, 'learning_rate/action_model': 5.03398000478455e-06, 'epoch': 1.91}) +[2026-06-16 09:29:28,137][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.08015678822994232, 'train/grad_norm_pre_clip': 10.749860763549805, 'timing/data': 0.0319429449737072, 'timing/model': 1.4771033469587564, 'learning_rate/qwen_vl_interface': 5.032838033392686e-07, 'learning_rate/action_model': 5.032838033392685e-06, 'epoch': 1.91}) +[2026-06-16 09:29:34,243][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.09456964582204819, 'train/grad_norm_pre_clip': 30.353256225585938, 'timing/data': 0.030765358358621597, 'timing/model': 1.7460646089166403, 'learning_rate/qwen_vl_interface': 5.031715573958111e-07, 'learning_rate/action_model': 5.031715573958111e-06, 'epoch': 1.91}) +[2026-06-16 09:29:40,138][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.0009376804810017347, 'train/grad_norm_pre_clip': 0.20970416069030762, 'timing/data': 0.029955798760056496, 'timing/model': 1.558732982724905, 'learning_rate/qwen_vl_interface': 5.030612626942235e-07, 'learning_rate/action_model': 5.030612626942235e-06, 'epoch': 1.91}) +[2026-06-16 09:29:46,241][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.07408788055181503, 'train/grad_norm_pre_clip': 22.153289794921875, 'timing/data': 0.03027547150850296, 'timing/model': 1.4788088947534561, 'learning_rate/qwen_vl_interface': 5.029529192798433e-07, 'learning_rate/action_model': 5.029529192798433e-06, 'epoch': 1.91}) +[2026-06-16 09:29:52,344][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.013881213031709194, 'train/grad_norm_pre_clip': 7.111000061035156, 'timing/data': 0.036694472655653954, 'timing/model': 1.568884253501892, 'learning_rate/qwen_vl_interface': 5.028465271972058e-07, 'learning_rate/action_model': 5.028465271972058e-06, 'epoch': 1.91}) +[2026-06-16 09:29:58,851][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.002564786234870553, 'train/grad_norm_pre_clip': 0.8871942162513733, 'timing/data': 0.033468229696154594, 'timing/model': 1.5245286840945482, 'learning_rate/qwen_vl_interface': 5.027420864900457e-07, 'learning_rate/action_model': 5.027420864900456e-06, 'epoch': 1.91}) +[2026-06-16 09:30:04,929][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.0651230737566948, 'train/grad_norm_pre_clip': 39.97038269042969, 'timing/data': 0.029701896011829376, 'timing/model': 1.4791562259197235, 'learning_rate/qwen_vl_interface': 5.026395972012943e-07, 'learning_rate/action_model': 5.0263959720129436e-06, 'epoch': 1.91}) +[2026-06-16 09:30:11,314][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.10752223432064056, 'train/grad_norm_pre_clip': 12.383979797363281, 'timing/data': 0.03548017144203186, 'timing/model': 1.7755661495029926, 'learning_rate/qwen_vl_interface': 5.025390593730809e-07, 'learning_rate/action_model': 5.025390593730809e-06, 'epoch': 1.92}) +[2026-06-16 09:30:17,622][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.21870584785938263, 'train/grad_norm_pre_clip': 21.589054107666016, 'timing/data': 0.030732590705156326, 'timing/model': 1.6403160355985165, 'learning_rate/qwen_vl_interface': 5.024404730467325e-07, 'learning_rate/action_model': 5.024404730467325e-06, 'epoch': 1.92}) +[2026-06-16 09:30:24,053][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.003487766720354557, 'train/grad_norm_pre_clip': 1.4447108507156372, 'timing/data': 0.036760732531547546, 'timing/model': 1.587137158960104, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 1.92}) +[2026-06-16 09:30:30,275][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.07105712592601776, 'train/grad_norm_pre_clip': 5.281132698059082, 'timing/data': 0.03167653642594814, 'timing/model': 1.5315890721976757, 'learning_rate/qwen_vl_interface': 5.022491550609311e-07, 'learning_rate/action_model': 5.022491550609311e-06, 'epoch': 1.92}) +[2026-06-16 09:30:36,566][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.028692876920104027, 'train/grad_norm_pre_clip': 37.68321228027344, 'timing/data': 0.028538279235363007, 'timing/model': 1.5235043484717607, 'learning_rate/qwen_vl_interface': 5.021564234801213e-07, 'learning_rate/action_model': 5.021564234801213e-06, 'epoch': 1.92}) +[2026-06-16 09:30:42,769][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.16757233440876007, 'train/grad_norm_pre_clip': 17.14649772644043, 'timing/data': 0.030679354444146156, 'timing/model': 1.5516414400190115, 'learning_rate/qwen_vl_interface': 5.020656435584636e-07, 'learning_rate/action_model': 5.020656435584635e-06, 'epoch': 1.92}) +[2026-06-16 09:30:49,107][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.010587316937744617, 'train/grad_norm_pre_clip': 3.385812997817993, 'timing/data': 0.034095581620931625, 'timing/model': 1.7631535083055496, 'learning_rate/qwen_vl_interface': 5.019768153332752e-07, 'learning_rate/action_model': 5.019768153332752e-06, 'epoch': 1.92}) +[2026-06-16 09:30:55,182][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.03295978903770447, 'train/grad_norm_pre_clip': 17.522071838378906, 'timing/data': 0.0341024175286293, 'timing/model': 1.5383241288363934, 'learning_rate/qwen_vl_interface': 5.018899388410693e-07, 'learning_rate/action_model': 5.0188993884106935e-06, 'epoch': 1.92}) +[2026-06-16 09:31:01,579][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 0.0003247639106120914, 'train/grad_norm_pre_clip': 0.07503855228424072, 'timing/data': 0.03574960120022297, 'timing/model': 1.523193832486868, 'learning_rate/qwen_vl_interface': 5.018050141175576e-07, 'learning_rate/action_model': 5.018050141175575e-06, 'epoch': 1.92}) +[2026-06-16 09:31:07,961][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.27039408683776855, 'train/grad_norm_pre_clip': 12.603998184204102, 'timing/data': 0.030972890555858612, 'timing/model': 1.5314545724540949, 'learning_rate/qwen_vl_interface': 5.017220411976496e-07, 'learning_rate/action_model': 5.0172204119764965e-06, 'epoch': 1.92}) +[2026-06-16 09:31:14,039][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.1668025702238083, 'train/grad_norm_pre_clip': 21.43507194519043, 'timing/data': 0.03238489292562008, 'timing/model': 1.5384333599358797, 'learning_rate/qwen_vl_interface': 5.016410201154523e-07, 'learning_rate/action_model': 5.016410201154523e-06, 'epoch': 1.92}) +[2026-06-16 09:31:20,134][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.0363592728972435, 'train/grad_norm_pre_clip': 11.76132869720459, 'timing/data': 0.031729307025671005, 'timing/model': 1.5046998281031847, 'learning_rate/qwen_vl_interface': 5.015619509042709e-07, 'learning_rate/action_model': 5.015619509042709e-06, 'epoch': 1.92}) +[2026-06-16 09:31:26,248][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.0007728746859356761, 'train/grad_norm_pre_clip': 0.20415742695331573, 'timing/data': 0.0317066814750433, 'timing/model': 1.7462056819349527, 'learning_rate/qwen_vl_interface': 5.014848335966067e-07, 'learning_rate/action_model': 5.014848335966067e-06, 'epoch': 1.92}) +[2026-06-16 09:31:32,143][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 0.018463393673300743, 'train/grad_norm_pre_clip': 25.752870559692383, 'timing/data': 0.032700302079319954, 'timing/model': 1.5360996760427952, 'learning_rate/qwen_vl_interface': 5.014096682241608e-07, 'learning_rate/action_model': 5.014096682241608e-06, 'epoch': 1.92}) +[2026-06-16 09:31:38,342][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.00983443483710289, 'train/grad_norm_pre_clip': 8.357745170593262, 'timing/data': 0.031081339344382286, 'timing/model': 1.556450868025422, 'learning_rate/qwen_vl_interface': 5.013364548178299e-07, 'learning_rate/action_model': 5.013364548178299e-06, 'epoch': 1.92}) +[2026-06-16 09:31:44,439][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.008913409896194935, 'train/grad_norm_pre_clip': 6.8758745193481445, 'timing/data': 0.0333983413875103, 'timing/model': 1.542818894609809, 'learning_rate/qwen_vl_interface': 5.012651934077101e-07, 'learning_rate/action_model': 5.0126519340771e-06, 'epoch': 1.92}) +[2026-06-16 09:31:50,537][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 0.004537524655461311, 'train/grad_norm_pre_clip': 1.0720940828323364, 'timing/data': 0.03189351595938206, 'timing/model': 1.5321160275489092, 'learning_rate/qwen_vl_interface': 5.01195884023094e-07, 'learning_rate/action_model': 5.0119588402309395e-06, 'epoch': 1.92}) +[2026-06-16 09:31:56,385][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.07223686575889587, 'train/grad_norm_pre_clip': 10.83350944519043, 'timing/data': 0.026744728907942772, 'timing/model': 1.4615071676671505, 'learning_rate/qwen_vl_interface': 5.011285266924714e-07, 'learning_rate/action_model': 5.011285266924714e-06, 'epoch': 1.92}) +[2026-06-16 09:32:02,437][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.009753536432981491, 'train/grad_norm_pre_clip': 2.3486225605010986, 'timing/data': 0.030855456367135048, 'timing/model': 1.788739724084735, 'learning_rate/qwen_vl_interface': 5.010631214435312e-07, 'learning_rate/action_model': 5.010631214435312e-06, 'epoch': 1.92}) +[2026-06-16 09:32:08,368][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 0.0018729348666965961, 'train/grad_norm_pre_clip': 0.2945708930492401, 'timing/data': 0.04328066483139992, 'timing/model': 1.5486753229051828, 'learning_rate/qwen_vl_interface': 5.009996683031586e-07, 'learning_rate/action_model': 5.009996683031586e-06, 'epoch': 1.92}) +[2026-06-16 09:32:14,636][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.04758041724562645, 'train/grad_norm_pre_clip': 15.19874095916748, 'timing/data': 0.032153159379959106, 'timing/model': 1.5467599295079708, 'learning_rate/qwen_vl_interface': 5.009381672974375e-07, 'learning_rate/action_model': 5.009381672974375e-06, 'epoch': 1.92}) +[2026-06-16 09:32:20,717][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.001762928324751556, 'train/grad_norm_pre_clip': 0.3782469630241394, 'timing/data': 0.027354568243026733, 'timing/model': 1.5309557802975178, 'learning_rate/qwen_vl_interface': 5.008786184516475e-07, 'learning_rate/action_model': 5.008786184516474e-06, 'epoch': 1.92}) +[2026-06-16 09:32:26,793][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 0.014292017556726933, 'train/grad_norm_pre_clip': 10.147833824157715, 'timing/data': 0.025265108793973923, 'timing/model': 1.6067784558981657, 'learning_rate/qwen_vl_interface': 5.008210217902676e-07, 'learning_rate/action_model': 5.008210217902676e-06, 'epoch': 1.92}) +[2026-06-16 09:32:32,880][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.05893910303711891, 'train/grad_norm_pre_clip': 16.33155632019043, 'timing/data': 0.024760544300079346, 'timing/model': 1.4245722834020853, 'learning_rate/qwen_vl_interface': 5.007653773369737e-07, 'learning_rate/action_model': 5.007653773369736e-06, 'epoch': 1.92}) +[2026-06-16 09:32:38,936][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.004706060979515314, 'train/grad_norm_pre_clip': 5.161438941955566, 'timing/data': 0.02590353414416313, 'timing/model': 1.7509801387786865, 'learning_rate/qwen_vl_interface': 5.007116851146386e-07, 'learning_rate/action_model': 5.007116851146386e-06, 'epoch': 1.92}) +[2026-06-16 09:32:44,726][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.005903645418584347, 'train/grad_norm_pre_clip': 1.5751351118087769, 'timing/data': 0.029711386188864708, 'timing/model': 1.4586325585842133, 'learning_rate/qwen_vl_interface': 5.006599451453334e-07, 'learning_rate/action_model': 5.006599451453334e-06, 'epoch': 1.93}) +[2026-06-16 09:32:50,746][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.20935803651809692, 'train/grad_norm_pre_clip': 9.444384574890137, 'timing/data': 0.03142860345542431, 'timing/model': 1.4739993084222078, 'learning_rate/qwen_vl_interface': 5.006101574503268e-07, 'learning_rate/action_model': 5.006101574503267e-06, 'epoch': 1.93}) +[2026-06-16 09:32:56,853][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.09188748151063919, 'train/grad_norm_pre_clip': 7.435336589813232, 'timing/data': 0.026195868849754333, 'timing/model': 1.5421204790472984, 'learning_rate/qwen_vl_interface': 5.005623220500839e-07, 'learning_rate/action_model': 5.005623220500839e-06, 'epoch': 1.93}) +[2026-06-16 09:33:03,084][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.000417268747696653, 'train/grad_norm_pre_clip': 0.09743134677410126, 'timing/data': 0.03249309957027435, 'timing/model': 1.5387912020087242, 'learning_rate/qwen_vl_interface': 5.005164389642687e-07, 'learning_rate/action_model': 5.005164389642687e-06, 'epoch': 1.93}) +[2026-06-16 09:33:09,382][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.1768999993801117, 'train/grad_norm_pre_clip': 31.582399368286133, 'timing/data': 0.029935939237475395, 'timing/model': 1.5282186456024647, 'learning_rate/qwen_vl_interface': 5.004725082117423e-07, 'learning_rate/action_model': 5.004725082117422e-06, 'epoch': 1.93}) +[2026-06-16 09:33:15,741][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.08226119726896286, 'train/grad_norm_pre_clip': 33.082855224609375, 'timing/data': 0.03200776129961014, 'timing/model': 1.7756063044071198, 'learning_rate/qwen_vl_interface': 5.004305298105616e-07, 'learning_rate/action_model': 5.004305298105616e-06, 'epoch': 1.93}) +[2026-06-16 09:33:21,910][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.018810153007507324, 'train/grad_norm_pre_clip': 5.012979507446289, 'timing/data': 0.031878456473350525, 'timing/model': 1.6550748571753502, 'learning_rate/qwen_vl_interface': 5.003905037779834e-07, 'learning_rate/action_model': 5.0039050377798345e-06, 'epoch': 1.93}) +[2026-06-16 09:33:27,984][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.34498122334480286, 'train/grad_norm_pre_clip': 15.014751434326172, 'timing/data': 0.03076831065118313, 'timing/model': 1.523916631937027, 'learning_rate/qwen_vl_interface': 5.003524301304608e-07, 'learning_rate/action_model': 5.003524301304608e-06, 'epoch': 1.93}) +[2026-06-16 09:33:34,274][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.008634761907160282, 'train/grad_norm_pre_clip': 11.513140678405762, 'timing/data': 0.03096948191523552, 'timing/model': 1.5383575484156609, 'learning_rate/qwen_vl_interface': 5.003163088836445e-07, 'learning_rate/action_model': 5.003163088836446e-06, 'epoch': 1.93}) +[2026-06-16 09:33:40,578][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.0070840525440871716, 'train/grad_norm_pre_clip': 1.9562780857086182, 'timing/data': 0.030931027606129646, 'timing/model': 1.5434229411184788, 'learning_rate/qwen_vl_interface': 5.002821400523818e-07, 'learning_rate/action_model': 5.002821400523818e-06, 'epoch': 1.93}) +[2026-06-16 09:33:46,942][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 0.07116842269897461, 'train/grad_norm_pre_clip': 27.170379638671875, 'timing/data': 0.03203248046338558, 'timing/model': 1.48958545550704, 'learning_rate/qwen_vl_interface': 5.002499236507192e-07, 'learning_rate/action_model': 5.0024992365071925e-06, 'epoch': 1.93}) +[2026-06-16 09:33:53,171][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.015082626603543758, 'train/grad_norm_pre_clip': 4.402805805206299, 'timing/data': 0.0320298969745636, 'timing/model': 1.7575012724846601, 'learning_rate/qwen_vl_interface': 5.002196596918993e-07, 'learning_rate/action_model': 5.002196596918993e-06, 'epoch': 1.93}) +[2026-06-16 09:33:59,361][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.03557915240526199, 'train/grad_norm_pre_clip': 25.066133499145508, 'timing/data': 0.03211500495672226, 'timing/model': 1.7520908005535603, 'learning_rate/qwen_vl_interface': 5.001913481883623e-07, 'learning_rate/action_model': 5.001913481883623e-06, 'epoch': 1.93}) +[2026-06-16 09:34:05,340][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.1526186317205429, 'train/grad_norm_pre_clip': 12.102739334106445, 'timing/data': 0.03182006999850273, 'timing/model': 1.5020293798297644, 'learning_rate/qwen_vl_interface': 5.001649891517459e-07, 'learning_rate/action_model': 5.001649891517459e-06, 'epoch': 1.93}) +[2026-06-16 09:34:11,647][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.025724003091454506, 'train/grad_norm_pre_clip': 7.50661039352417, 'timing/data': 0.03290271386504173, 'timing/model': 1.5308069568127394, 'learning_rate/qwen_vl_interface': 5.001405825928858e-07, 'learning_rate/action_model': 5.0014058259288575e-06, 'epoch': 1.93}) +[2026-06-16 09:34:18,053][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.004121190868318081, 'train/grad_norm_pre_clip': 0.6235864758491516, 'timing/data': 0.034177444875240326, 'timing/model': 1.6427629571408033, 'learning_rate/qwen_vl_interface': 5.001181285218143e-07, 'learning_rate/action_model': 5.0011812852181434e-06, 'epoch': 1.93}) +[2026-06-16 09:34:24,248][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.1327296644449234, 'train/grad_norm_pre_clip': 6.901005744934082, 'timing/data': 0.03242347203195095, 'timing/model': 1.5252791717648506, 'learning_rate/qwen_vl_interface': 5.000976269477613e-07, 'learning_rate/action_model': 5.0009762694776125e-06, 'epoch': 1.93}) +[2026-06-16 09:34:30,478][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.05045134946703911, 'train/grad_norm_pre_clip': 10.177327156066895, 'timing/data': 0.03259347938001156, 'timing/model': 1.765146804973483, 'learning_rate/qwen_vl_interface': 5.000790778791542e-07, 'learning_rate/action_model': 5.000790778791543e-06, 'epoch': 1.93}) +[2026-06-16 09:34:36,472][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.13387024402618408, 'train/grad_norm_pre_clip': 9.342684745788574, 'timing/data': 0.031228551641106606, 'timing/model': 1.531601544469595, 'learning_rate/qwen_vl_interface': 5.000624813236185e-07, 'learning_rate/action_model': 5.000624813236185e-06, 'epoch': 1.93}) +[2026-06-16 09:34:42,767][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.0022285799495875835, 'train/grad_norm_pre_clip': 0.3648325800895691, 'timing/data': 0.03199109062552452, 'timing/model': 1.5107559002935886, 'learning_rate/qwen_vl_interface': 5.000478372879752e-07, 'learning_rate/action_model': 5.000478372879752e-06, 'epoch': 1.93}) +[2026-06-16 09:34:48,971][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.11239791661500931, 'train/grad_norm_pre_clip': 13.621885299682617, 'timing/data': 0.03223777562379837, 'timing/model': 1.5177211854606867, 'learning_rate/qwen_vl_interface': 5.000351457782452e-07, 'learning_rate/action_model': 5.000351457782451e-06, 'epoch': 1.93}) +[2026-06-16 09:34:55,020][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 0.017895130440592766, 'train/grad_norm_pre_clip': 7.729355335235596, 'timing/data': 0.025859177112579346, 'timing/model': 1.4614011012017727, 'learning_rate/qwen_vl_interface': 5.000244067996445e-07, 'learning_rate/action_model': 5.000244067996445e-06, 'epoch': 1.93}) +[2026-06-16 09:35:01,084][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.0013676730450242758, 'train/grad_norm_pre_clip': 0.4415513277053833, 'timing/data': 0.030942603945732117, 'timing/model': 1.5090389270335436, 'learning_rate/qwen_vl_interface': 5.000156203565884e-07, 'learning_rate/action_model': 5.000156203565883e-06, 'epoch': 1.93}) +[2026-06-16 09:35:07,327][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.00697926664724946, 'train/grad_norm_pre_clip': 1.8967065811157227, 'timing/data': 0.03083040751516819, 'timing/model': 1.6814876236021519, 'learning_rate/qwen_vl_interface': 5.000087864526878e-07, 'learning_rate/action_model': 5.000087864526878e-06, 'epoch': 1.93}) +[2026-06-16 09:35:13,365][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.0017646297346800566, 'train/grad_norm_pre_clip': 0.6394241452217102, 'timing/data': 0.0312027707695961, 'timing/model': 1.5097277499735355, 'learning_rate/qwen_vl_interface': 5.000039050907523e-07, 'learning_rate/action_model': 5.000039050907523e-06, 'epoch': 1.93}) +[2026-06-16 09:35:19,673][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.015442447736859322, 'train/grad_norm_pre_clip': 4.643595218658447, 'timing/data': 0.03134850412607193, 'timing/model': 1.5279419757425785, 'learning_rate/qwen_vl_interface': 5.000009762727885e-07, 'learning_rate/action_model': 5.000009762727884e-06, 'epoch': 1.93}) +[2026-06-16 09:42:33,810][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.20379847288131714, 'train/grad_norm_pre_clip': 21.844890594482422, 'eval/loss': 0.04701593145728111, 'eval/flap_precision': 0.9598393574297188, 'eval/flap_recall': 0.9728629579375848, 'eval/flap_f1': 0.9663072776280323, 'eval/flap_rate_model': 0.2075, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9663072776280323, 'timing/data': 0.030853038653731346, 'timing/model': 1.511435840278864, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 1.94}) +[2026-06-16 09:42:33,811][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_5000_state +[2026-06-16 09:42:33,817][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 09:42:40,689][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_5000_state/model.safetensors +[2026-06-16 09:42:50,120][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_5000_state/optimizer.bin +[2026-06-16 09:42:50,121][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_5000_state/sampler.bin +[2026-06-16 09:42:50,121][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_5000_state/sampler_1.bin +[2026-06-16 09:42:50,122][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints/steps_5000_state/random_states_0.pkl +[2026-06-16 09:42:50,227][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-06-16 09:42:50,227][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 09:42:50,281][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 09:42:50,281][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /workspace/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep/checkpoints +[2026-06-16 09:42:52,282][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..ffa1ed16fa7cd08f0c83739d659e9d2a8bd4e827 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep/summary.jsonl @@ -0,0 +1,50 @@ +{"steps": 100} +{"steps": 200} +{"steps": 300} +{"steps": 400} +{"steps": 500} +{"steps": 600} +{"steps": 700} +{"steps": 800} +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500} +{"steps": 1600} +{"steps": 1700} +{"steps": 1800} +{"steps": 1900} +{"steps": 2000} +{"steps": 2100} +{"steps": 2200} +{"steps": 2300} +{"steps": 2400} +{"steps": 2500} +{"steps": 2600} +{"steps": 2700} +{"steps": 2800} +{"steps": 2900} +{"steps": 3000} +{"steps": 3100} +{"steps": 3200} +{"steps": 3300} +{"steps": 3400} +{"steps": 3500} +{"steps": 3600} +{"steps": 3700} +{"steps": 3800} +{"steps": 3900} +{"steps": 4000} +{"steps": 4100} +{"steps": 4200} +{"steps": 4300} +{"steps": 4400} +{"steps": 4500} +{"steps": 4600} +{"steps": 4700} +{"steps": 4800} +{"steps": 4900} +{"steps": 5000} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/checkpoints/steps_5000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/checkpoints/steps_5000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9b098e38f746e082b50323df404218adff240e55 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/checkpoints/steps_5000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5da01d6eeb7a46861da90b0a8dbc364a33f92d542d25079bcc86280b10617418 +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..fbdf51172ce46ee8c6c91aeae0eedfe46fe47503 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/config.full.yaml @@ -0,0 +1,199 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_fix_latency_3_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: null + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_3_200ep +output_dir: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..a9e1fa82de9e3d798e97ae46453b709c50c8104c --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/config.yaml @@ -0,0 +1,71 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep +rl_games: + env_eval: + enabled: true + mid_train: + enabled: false + interval_steps: 100 + task: flappy +run_id: flappy_fix_latency_3_200ep +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_action_classification: true + eval_interval: 100 + eval_num_batches: 100 + freeze_modules: '' + gradient_accumulation_steps: 4 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + pretrained_checkpoint: null + reload_modules: null + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..f00fda49b35be32792bf4764fffbfe54fc4f0761 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.5971351861953735, + 0.40286484360694885, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.49107804894447327, + 0.49107804894447327, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 316528, + "num_trajectories": 180 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..c7de099f26b1fc338da09fe9b570a7ef1fd35fc4 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8040833473205566, + 0.19591666758060455, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.39686188101768494, + 0.39686188101768494, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 72000, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/hydra/.hydra/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..14eeeb2778efee3be2e13da1370f5182ffc0279b --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/hydra/.hydra/config.yaml @@ -0,0 +1,197 @@ +framework: + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_fix_latency_3_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + config_name: null + source_subdir: null + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: null + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_3_200ep +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/hydra/.hydra/hydra.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..bff9a928aafe484e8ce810b1ee13911f3bbc31b0 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/hydra/.hydra/hydra.yaml @@ -0,0 +1,252 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=single + - run_id=flappy_fix_latency_3_200ep + - run_root_dir=/workspace/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=false + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=5000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=100 + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=4 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9, 0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=2 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.shuffle=true + - datasets.vla_data.action_balance.enabled=false + - datasets.vla_data.action_balance.strategy=balanced_epoch + - datasets.vla_data.action_balance.action_key=action_id + - datasets.vla_data.action_balance.target_flap_fraction=0.3 + - datasets.vla_data.action_balance.noop_id=0 + - datasets.vla_data.action_balance.flap_id=1 + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_3_200ep + - datasets.vla_data.data_mix=flappy_train__bridge + - datasets.vla_data.eval_data_mix=flappy_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=flappy + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=1 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.mid_train.enabled=false + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[0] + - rl_games.env_eval.post_train.enabled=false + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=false,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_balance.action_key=action_id,datasets.vla_data.action_balance.enabled=false,datasets.vla_data.action_balance.flap_id=1,datasets.vla_data.action_balance.noop_id=0,datasets.vla_data.action_balance.strategy=balanced_epoch,datasets.vla_data.action_balance.target_flap_fraction=0.3,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=flappy_train__bridge,datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_3_200ep,datasets.vla_data.eval_data_mix=flappy_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.shuffle=true,datasets.vla_data.video_backend=torchvision_av,env=flappy,framework.action_model.action_dim=7,framework.action_model.action_env_dim=2,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=1,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=false,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[0],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=false,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=flappy,run_id=flappy_fix_latency_3_200ep,run_root_dir=/workspace/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=100,trainer.gradient_accumulation_steps=4,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=5000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9, + 0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.3 + version_base: '1.1' + cwd: /workspace/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/hydra/.hydra/overrides.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..a5ed8974cf6ea2f24671306315a29d98728e81db --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/hydra/.hydra/overrides.yaml @@ -0,0 +1,91 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=single +- run_id=flappy_fix_latency_3_200ep +- run_root_dir=/workspace/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=false +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=5000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=100 +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=4 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9, 0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=2 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.shuffle=true +- datasets.vla_data.action_balance.enabled=false +- datasets.vla_data.action_balance.strategy=balanced_epoch +- datasets.vla_data.action_balance.action_key=action_id +- datasets.vla_data.action_balance.target_flap_fraction=0.3 +- datasets.vla_data.action_balance.noop_id=0 +- datasets.vla_data.action_balance.flap_id=1 +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_3_200ep +- datasets.vla_data.data_mix=flappy_train__bridge +- datasets.vla_data.eval_data_mix=flappy_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=flappy +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=1 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.mid_train.enabled=false +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[0] +- rl_games.env_eval.post_train.enabled=false +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/hydra/train_starvla_hydra.log b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..3cc754b15ce86fa6bd6b3314408dea956862018a --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/hydra/train_starvla_hydra.log @@ -0,0 +1,5552 @@ +[2026-06-15 19:04:59,578][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-15 19:04:59,601][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-15 19:05:11,062][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-15 19:05:11,118][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-15 19:05:11,935][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-15 19:05:11,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-15 19:05:11,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-15 19:05:11,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-15 19:05:11,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-15 19:05:11,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-15 19:05:11,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-15 19:05:11,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-15 19:05:11,937][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-15 19:05:11,937][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-15 19:05:11,937][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-15 19:05:11,937][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-15 19:05:11,937][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-15 19:05:11,937][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-15 19:05:11,937][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-15 19:05:11,937][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-15 19:05:11,938][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-15 19:05:11,938][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-15 19:05:11,938][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-15 19:05:11,938][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-15 19:05:11,938][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-15 19:05:11,938][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-15 19:05:11,938][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-15 19:05:11,939][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-15 19:05:11,939][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-15 19:05:11,939][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-15 19:05:11,939][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-15 19:05:11,939][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-15 19:05:11,939][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-15 19:05:11,939][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-15 19:05:11,939][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-15 19:05:11,995][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-15 19:05:12,040][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/config.full.yaml +[2026-06-15 19:05:12,049][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/config.yaml +[2026-06-15 19:05:12,050][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-15 19:05:34,201][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-15 19:05:34,201][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-06-15 19:05:34,201][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-15 19:05:34,201][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 4 +[2026-06-15 19:05:34,201][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-15 19:06:08,127][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 0.6950794458389282, 'train/grad_norm_pre_clip': 3.6724846363067627, 'timing/data': 0.043561811093240976, 'timing/model': 1.260313581675291, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-15 19:06:12,799][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 0.6942895650863647, 'train/grad_norm_pre_clip': 1.6813071966171265, 'timing/data': 0.03551572421565652, 'timing/model': 1.2286668540909886, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-15 19:06:17,290][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 0.6915071606636047, 'train/grad_norm_pre_clip': 2.299996852874756, 'timing/data': 0.028540743980556726, 'timing/model': 1.1589012839831412, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-06-15 19:06:21,875][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 0.6848978996276855, 'train/grad_norm_pre_clip': 2.7885327339172363, 'timing/data': 0.03131660399958491, 'timing/model': 1.135405187960714, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-06-15 19:06:26,690][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 0.7062727808952332, 'train/grad_norm_pre_clip': 2.62735652923584, 'timing/data': 0.03570600692182779, 'timing/model': 1.1548062763176858, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-06-15 19:06:31,274][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 0.6661126017570496, 'train/grad_norm_pre_clip': 2.6778438091278076, 'timing/data': 0.03140298929065466, 'timing/model': 1.2218046472407877, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-06-15 19:06:35,889][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 0.6579855680465698, 'train/grad_norm_pre_clip': 2.1500065326690674, 'timing/data': 0.02790414495393634, 'timing/model': 1.148560875095427, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-06-15 19:06:40,487][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.636990487575531, 'train/grad_norm_pre_clip': 2.2407455444335938, 'timing/data': 0.03177957981824875, 'timing/model': 1.1746844127774239, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-06-15 19:06:45,076][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 0.7823224663734436, 'train/grad_norm_pre_clip': 5.882976055145264, 'timing/data': 0.024584801867604256, 'timing/model': 1.2302614920772612, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-06-15 19:06:49,592][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.6627289652824402, 'train/grad_norm_pre_clip': 0.8302857875823975, 'timing/data': 0.0284759858623147, 'timing/model': 1.1485068569891155, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-15 19:06:54,409][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.6204701662063599, 'train/grad_norm_pre_clip': 0.4858384430408478, 'timing/data': 0.019467113073915243, 'timing/model': 1.4731300729326904, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-06-15 19:06:58,905][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.6692717671394348, 'train/grad_norm_pre_clip': 1.6360288858413696, 'timing/data': 0.016764372121542692, 'timing/model': 1.160908174701035, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0}) +[2026-06-15 19:07:03,455][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.6709311008453369, 'train/grad_norm_pre_clip': 1.815356969833374, 'timing/data': 0.018005219288170338, 'timing/model': 1.2103828210383654, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.01}) +[2026-06-15 19:07:07,996][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.7455207109451294, 'train/grad_norm_pre_clip': 4.213377475738525, 'timing/data': 0.019720991607755423, 'timing/model': 1.1571488659828901, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.01}) +[2026-06-15 19:07:12,577][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.7350157499313354, 'train/grad_norm_pre_clip': 3.776216506958008, 'timing/data': 0.019224913325160742, 'timing/model': 1.2326418063603342, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.01}) +[2026-06-15 19:07:17,200][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.602380633354187, 'train/grad_norm_pre_clip': 1.7184579372406006, 'timing/data': 0.01927768299356103, 'timing/model': 1.2500822637230158, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 19:07:21,907][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.712407648563385, 'train/grad_norm_pre_clip': 2.873831033706665, 'timing/data': 0.018331628292798996, 'timing/model': 1.2612291327677667, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 19:07:26,576][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.6295365691184998, 'train/grad_norm_pre_clip': 1.8746718168258667, 'timing/data': 0.018060382921248674, 'timing/model': 1.223925950936973, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.01}) +[2026-06-15 19:07:31,185][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.658058762550354, 'train/grad_norm_pre_clip': 0.5383052229881287, 'timing/data': 0.01869768975302577, 'timing/model': 1.231009264010936, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.01}) +[2026-06-15 19:07:35,833][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.6971765160560608, 'train/grad_norm_pre_clip': 1.949150562286377, 'timing/data': 0.018039578571915627, 'timing/model': 1.2092756209895015, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-15 19:07:40,519][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.6483108997344971, 'train/grad_norm_pre_clip': 1.8129568099975586, 'timing/data': 0.019330274779349566, 'timing/model': 1.2701539839617908, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.01}) +[2026-06-15 19:07:45,084][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.704899787902832, 'train/grad_norm_pre_clip': 2.328106641769409, 'timing/data': 0.018541252706199884, 'timing/model': 1.233512079808861, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 19:07:49,808][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.6821631193161011, 'train/grad_norm_pre_clip': 0.7225078344345093, 'timing/data': 0.018968707881867886, 'timing/model': 1.2411992801353335, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 19:07:54,470][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.6897631883621216, 'train/grad_norm_pre_clip': 1.516922950744629, 'timing/data': 0.017531637102365494, 'timing/model': 1.2427381169982255, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.01}) +[2026-06-15 19:07:59,205][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.6660571694374084, 'train/grad_norm_pre_clip': 4.5716166496276855, 'timing/data': 0.019452711567282677, 'timing/model': 1.2406620858237147, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.01}) +[2026-06-15 19:08:03,906][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.6714301705360413, 'train/grad_norm_pre_clip': 4.0347514152526855, 'timing/data': 0.020240101031959057, 'timing/model': 1.2364794709719718, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.01}) +[2026-06-15 19:08:08,510][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.6780439615249634, 'train/grad_norm_pre_clip': 1.3085757493972778, 'timing/data': 0.01850323611870408, 'timing/model': 1.1887948769144714, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.01}) +[2026-06-15 19:08:13,287][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.6778265833854675, 'train/grad_norm_pre_clip': 1.5307925939559937, 'timing/data': 0.01966296974569559, 'timing/model': 1.250239981804043, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 19:08:17,908][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.6890416145324707, 'train/grad_norm_pre_clip': 4.69523286819458, 'timing/data': 0.018697755876928568, 'timing/model': 1.1719910162501037, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.01}) +[2026-06-15 19:08:22,643][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.6600711941719055, 'train/grad_norm_pre_clip': 7.686005592346191, 'timing/data': 0.019164140801876783, 'timing/model': 1.24998240172863, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-15 19:08:27,415][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.6555231809616089, 'train/grad_norm_pre_clip': 3.157658100128174, 'timing/data': 0.02253026282414794, 'timing/model': 1.2519468772225082, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.01}) +[2026-06-15 19:08:32,116][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.6232261657714844, 'train/grad_norm_pre_clip': 1.6551791429519653, 'timing/data': 0.018900964874774218, 'timing/model': 1.1703017139807343, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.01}) +[2026-06-15 19:08:36,886][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.6641365885734558, 'train/grad_norm_pre_clip': 3.568624258041382, 'timing/data': 0.0186231997795403, 'timing/model': 1.2368392599746585, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.01}) +[2026-06-15 19:08:41,614][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.6950137615203857, 'train/grad_norm_pre_clip': 5.474483013153076, 'timing/data': 0.01692565716803074, 'timing/model': 1.25551084196195, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.01}) +[2026-06-15 19:08:46,300][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.6463610529899597, 'train/grad_norm_pre_clip': 4.94097900390625, 'timing/data': 0.018985881004482508, 'timing/model': 1.2416644911281765, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.01}) +[2026-06-15 19:08:51,141][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.607109785079956, 'train/grad_norm_pre_clip': 4.796996593475342, 'timing/data': 0.019215164240449667, 'timing/model': 1.287073730956763, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.01}) +[2026-06-15 19:08:55,812][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.5663508772850037, 'train/grad_norm_pre_clip': 8.21845531463623, 'timing/data': 0.018669414799660444, 'timing/model': 1.2479022229090333, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.01}) +[2026-06-15 19:09:00,516][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.6353304982185364, 'train/grad_norm_pre_clip': 11.481087684631348, 'timing/data': 0.021443949081003666, 'timing/model': 1.2487521199509501, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.02}) +[2026-06-15 19:09:05,412][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.5764480829238892, 'train/grad_norm_pre_clip': 7.808916091918945, 'timing/data': 0.01921512931585312, 'timing/model': 1.1946344608440995, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 19:09:10,229][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.6565491557121277, 'train/grad_norm_pre_clip': 3.172970771789551, 'timing/data': 0.019080976024270058, 'timing/model': 1.2583034909330308, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.02}) +[2026-06-15 19:09:15,002][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.5601224303245544, 'train/grad_norm_pre_clip': 6.375847339630127, 'timing/data': 0.01952523086220026, 'timing/model': 1.2456680848263204, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.02}) +[2026-06-15 19:09:19,723][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.549798309803009, 'train/grad_norm_pre_clip': 6.225528717041016, 'timing/data': 0.01916267490014434, 'timing/model': 1.1940678819082677, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.02}) +[2026-06-15 19:09:24,513][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.4697257876396179, 'train/grad_norm_pre_clip': 6.183407783508301, 'timing/data': 0.019286504946649075, 'timing/model': 1.2527574910782278, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.02}) +[2026-06-15 19:09:29,214][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.504869818687439, 'train/grad_norm_pre_clip': 3.6395468711853027, 'timing/data': 0.018691105768084526, 'timing/model': 1.1826927410438657, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 19:09:33,895][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.383099228143692, 'train/grad_norm_pre_clip': 5.132894039154053, 'timing/data': 0.019715437665581703, 'timing/model': 1.2346943491138518, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.02}) +[2026-06-15 19:09:38,692][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.5359618663787842, 'train/grad_norm_pre_clip': 5.6183695793151855, 'timing/data': 0.019026518799364567, 'timing/model': 1.2444627080112696, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.02}) +[2026-06-15 19:09:43,415][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.5117762684822083, 'train/grad_norm_pre_clip': 14.747579574584961, 'timing/data': 0.017916984856128693, 'timing/model': 1.2144928630441427, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.02}) +[2026-06-15 19:09:48,210][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.8343796133995056, 'train/grad_norm_pre_clip': 40.2247314453125, 'timing/data': 0.01841584872454405, 'timing/model': 1.2445744662545621, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.02}) +[2026-06-15 19:09:53,015][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.4146016538143158, 'train/grad_norm_pre_clip': 25.242969512939453, 'timing/data': 0.021246942225843668, 'timing/model': 1.2109476439654827, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.02}) +[2026-06-15 19:09:57,723][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.3072759211063385, 'train/grad_norm_pre_clip': 6.532322406768799, 'timing/data': 0.016844897996634245, 'timing/model': 1.1815475160256028, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.02}) +[2026-06-15 19:10:02,491][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.5557382702827454, 'train/grad_norm_pre_clip': 23.01449203491211, 'timing/data': 0.018779780250042677, 'timing/model': 1.237819594796747, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 19:10:07,308][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.6811102032661438, 'train/grad_norm_pre_clip': 7.535217761993408, 'timing/data': 0.01905854232609272, 'timing/model': 1.2538228370249271, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.02}) +[2026-06-15 19:10:12,014][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.667302668094635, 'train/grad_norm_pre_clip': 13.372149467468262, 'timing/data': 0.018778687808662653, 'timing/model': 1.1834748112596571, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.02}) +[2026-06-15 19:10:16,721][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.6679757833480835, 'train/grad_norm_pre_clip': 9.758539199829102, 'timing/data': 0.019542935770004988, 'timing/model': 1.1864549783058465, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.02}) +[2026-06-15 19:10:21,513][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.5825536251068115, 'train/grad_norm_pre_clip': 20.197309494018555, 'timing/data': 0.019506934098899364, 'timing/model': 1.2504800837486982, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.02}) +[2026-06-15 19:10:26,319][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.5665510296821594, 'train/grad_norm_pre_clip': 17.255022048950195, 'timing/data': 0.02039824379608035, 'timing/model': 1.3469486450776458, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 19:10:31,123][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.4789898991584778, 'train/grad_norm_pre_clip': 6.984518527984619, 'timing/data': 0.01999776717275381, 'timing/model': 1.24415089096874, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.02}) +[2026-06-15 19:10:35,828][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.4484705328941345, 'train/grad_norm_pre_clip': 9.334809303283691, 'timing/data': 0.019146795850247145, 'timing/model': 1.2562315999530256, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.02}) +[2026-06-15 19:10:40,611][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.4155632257461548, 'train/grad_norm_pre_clip': 17.703420639038086, 'timing/data': 0.01937988819554448, 'timing/model': 1.2469529951922596, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.02}) +[2026-06-15 19:10:45,389][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.6068101525306702, 'train/grad_norm_pre_clip': 12.990472793579102, 'timing/data': 0.02013842621818185, 'timing/model': 1.2413963740691543, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.02}) +[2026-06-15 19:10:50,222][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.5181331634521484, 'train/grad_norm_pre_clip': 25.511016845703125, 'timing/data': 0.019569538068026304, 'timing/model': 1.2358342451043427, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.02}) +[2026-06-15 19:10:55,028][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.46414241194725037, 'train/grad_norm_pre_clip': 7.273467063903809, 'timing/data': 0.02111582364886999, 'timing/model': 1.2548502208665013, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.03}) +[2026-06-15 19:10:59,718][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.46091794967651367, 'train/grad_norm_pre_clip': 10.817440032958984, 'timing/data': 0.021165073849260807, 'timing/model': 1.2433892670087516, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.03}) +[2026-06-15 19:11:04,526][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.4233311712741852, 'train/grad_norm_pre_clip': 6.539572238922119, 'timing/data': 0.019952479284256697, 'timing/model': 1.2551384712569416, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.03}) +[2026-06-15 19:11:09,529][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.6433635354042053, 'train/grad_norm_pre_clip': 16.327207565307617, 'timing/data': 0.01992850797250867, 'timing/model': 1.2537553668953478, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.03}) +[2026-06-15 19:11:14,315][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.5048666596412659, 'train/grad_norm_pre_clip': 6.84199857711792, 'timing/data': 0.01873132260516286, 'timing/model': 1.2616100977174938, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.03}) +[2026-06-15 19:11:19,116][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.5152232646942139, 'train/grad_norm_pre_clip': 5.868174076080322, 'timing/data': 0.021992967929691076, 'timing/model': 1.2595783919095993, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.03}) +[2026-06-15 19:11:23,845][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.43098822236061096, 'train/grad_norm_pre_clip': 6.0883941650390625, 'timing/data': 0.018777833320200443, 'timing/model': 1.2018948569893837, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.03}) +[2026-06-15 19:11:28,621][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.36911043524742126, 'train/grad_norm_pre_clip': 5.09360408782959, 'timing/data': 0.019650531001389027, 'timing/model': 1.2445046929642558, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.03}) +[2026-06-15 19:11:33,394][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.5416786074638367, 'train/grad_norm_pre_clip': 24.517595291137695, 'timing/data': 0.018400336150079966, 'timing/model': 1.2437096131034195, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.03}) +[2026-06-15 19:11:38,118][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.5707727074623108, 'train/grad_norm_pre_clip': 29.40166473388672, 'timing/data': 0.01999201299622655, 'timing/model': 1.188067660201341, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.03}) +[2026-06-15 19:11:42,923][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.7404471635818481, 'train/grad_norm_pre_clip': 22.07059097290039, 'timing/data': 0.02305836509913206, 'timing/model': 1.2546411878429353, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.03}) +[2026-06-15 19:11:47,722][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.40818843245506287, 'train/grad_norm_pre_clip': 7.960685729980469, 'timing/data': 0.020167872309684753, 'timing/model': 1.250714076217264, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.03}) +[2026-06-15 19:11:52,488][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.5115309953689575, 'train/grad_norm_pre_clip': 8.347548484802246, 'timing/data': 0.019211021717637777, 'timing/model': 1.2400481230579317, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.03}) +[2026-06-15 19:11:57,218][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.8484440445899963, 'train/grad_norm_pre_clip': 37.07270812988281, 'timing/data': 0.018635463900864124, 'timing/model': 1.1814738931134343, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.03}) +[2026-06-15 19:12:02,009][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 0.5253087878227234, 'train/grad_norm_pre_clip': 17.948604583740234, 'timing/data': 0.019648313988000154, 'timing/model': 1.2490794518962502, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.03}) +[2026-06-15 19:12:06,806][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.4823644757270813, 'train/grad_norm_pre_clip': 11.976807594299316, 'timing/data': 0.019738804548978806, 'timing/model': 1.2579909549094737, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.03}) +[2026-06-15 19:12:11,603][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.48041924834251404, 'train/grad_norm_pre_clip': 7.411273956298828, 'timing/data': 0.019235958345234394, 'timing/model': 1.3105578967370093, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.03}) +[2026-06-15 19:12:16,315][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.4580838978290558, 'train/grad_norm_pre_clip': 11.462789535522461, 'timing/data': 0.019961314741522074, 'timing/model': 1.2447430128231645, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.03}) +[2026-06-15 19:12:21,007][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.5530476570129395, 'train/grad_norm_pre_clip': 11.452595710754395, 'timing/data': 0.01881350204348564, 'timing/model': 1.2449892121367157, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.03}) +[2026-06-15 19:12:25,808][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.5274460315704346, 'train/grad_norm_pre_clip': 7.484004974365234, 'timing/data': 0.019331696908921003, 'timing/model': 1.2564168712124228, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.03}) +[2026-06-15 19:12:30,625][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.5029457807540894, 'train/grad_norm_pre_clip': 7.3757476806640625, 'timing/data': 0.01881472021341324, 'timing/model': 1.232248009648174, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.03}) +[2026-06-15 19:12:35,409][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.510266125202179, 'train/grad_norm_pre_clip': 12.44406509399414, 'timing/data': 0.020510416943579912, 'timing/model': 1.2434983402490616, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.03}) +[2026-06-15 19:12:40,227][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.4063638150691986, 'train/grad_norm_pre_clip': 5.979213237762451, 'timing/data': 0.01843975530937314, 'timing/model': 1.2370743700303137, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.03}) +[2026-06-15 19:12:44,997][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.5525768995285034, 'train/grad_norm_pre_clip': 19.685325622558594, 'timing/data': 0.01950217504054308, 'timing/model': 1.2491270368918777, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.03}) +[2026-06-15 19:12:49,800][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.5996968746185303, 'train/grad_norm_pre_clip': 10.12064266204834, 'timing/data': 0.01867152564227581, 'timing/model': 1.2528534512966871, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.03}) +[2026-06-15 19:12:54,613][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.5747047662734985, 'train/grad_norm_pre_clip': 13.790938377380371, 'timing/data': 0.019655865151435137, 'timing/model': 1.2568178763613105, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.04}) +[2026-06-15 19:12:59,411][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.4384811520576477, 'train/grad_norm_pre_clip': 18.237085342407227, 'timing/data': 0.019510613288730383, 'timing/model': 1.2419026349671185, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.04}) +[2026-06-15 19:13:04,131][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.6744697093963623, 'train/grad_norm_pre_clip': 22.041624069213867, 'timing/data': 0.01950886193662882, 'timing/model': 1.2515915501862764, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.04}) +[2026-06-15 19:13:08,919][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.5948660969734192, 'train/grad_norm_pre_clip': 21.170757293701172, 'timing/data': 0.018537551164627075, 'timing/model': 1.2451475160196424, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.04}) +[2026-06-15 19:13:13,622][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.5605472326278687, 'train/grad_norm_pre_clip': 14.47829818725586, 'timing/data': 0.023514288943260908, 'timing/model': 1.2373284990899265, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.04}) +[2026-06-15 19:13:18,515][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.4299880564212799, 'train/grad_norm_pre_clip': 6.447131156921387, 'timing/data': 0.023535178042948246, 'timing/model': 1.2501162821426988, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.04}) +[2026-06-15 19:13:23,223][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.4344916045665741, 'train/grad_norm_pre_clip': 3.877480983734131, 'timing/data': 0.018780591897666454, 'timing/model': 1.2514035739004612, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.04}) +[2026-06-15 19:13:28,112][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.5518664717674255, 'train/grad_norm_pre_clip': 14.605691909790039, 'timing/data': 0.020374116022139788, 'timing/model': 1.255410262849182, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.04}) +[2026-06-15 19:13:32,887][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 1.0336359739303589, 'train/grad_norm_pre_clip': 18.750947952270508, 'timing/data': 0.01705117290839553, 'timing/model': 1.2423493890091777, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.04}) +[2026-06-15 19:13:37,606][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.4400854706764221, 'train/grad_norm_pre_clip': 3.765420913696289, 'timing/data': 0.019140363205224276, 'timing/model': 1.2404783121310174, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.04}) +[2026-06-15 19:13:42,432][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.315704345703125, 'train/grad_norm_pre_clip': 5.1773362159729, 'timing/data': 0.019315192941576242, 'timing/model': 1.2488466640934348, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.04}) +[2026-06-15 19:13:47,235][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.5648410320281982, 'train/grad_norm_pre_clip': 9.369698524475098, 'timing/data': 0.019204532261937857, 'timing/model': 1.2635596338659525, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.04}) +[2026-06-15 19:13:52,033][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.6810698509216309, 'train/grad_norm_pre_clip': 12.85167407989502, 'timing/data': 0.022416410967707634, 'timing/model': 1.2601135252043605, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.04}) +[2026-06-15 19:16:43,622][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpofg6st89/test.c -o /tmp/tmpofg6st89/test.o +[2026-06-15 19:16:43,657][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpofg6st89/test.o -laio -o /tmp/tmpofg6st89/a.out +[2026-06-15 19:16:44,183][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmppw3qjzzn/test.c -o /tmp/tmppw3qjzzn/test.o +[2026-06-15 19:16:44,216][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmppw3qjzzn/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmppw3qjzzn/a.out +[2026-06-15 19:16:44,272][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpekobrrsb/test.c -o /tmp/tmpekobrrsb/test.o +[2026-06-15 19:16:44,301][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpekobrrsb/test.o -laio -o /tmp/tmpekobrrsb/a.out +[2026-06-15 19:25:46,745][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.43821996450424194, 'train/grad_norm_pre_clip': 16.03993034362793, 'eval/loss': 0.32072678208351135, 'eval/flap_precision': 0.6392896781354052, 'eval/flap_recall': 0.7858117326057299, 'eval/flap_f1': 0.7050183598531212, 'eval/flap_rate_model': 0.25027777777777777, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.7050183598531212, 'timing/data': 0.018293501809239388, 'timing/model': 1.2474145479500294, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.04}) +[2026-06-15 19:25:46,746][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_100_state +[2026-06-15 19:25:46,752][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 19:25:55,368][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_100_state/model.safetensors +[2026-06-15 19:26:14,866][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_100_state/optimizer.bin +[2026-06-15 19:26:14,868][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_100_state/sampler.bin +[2026-06-15 19:26:14,868][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_100_state/sampler_1.bin +[2026-06-15 19:26:14,869][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-15 19:26:14,873][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-15 19:26:14,873][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 19:26:14,887][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 19:26:21,227][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.5063599944114685, 'train/grad_norm_pre_clip': 16.834922790527344, 'timing/data': 0.020277777686715126, 'timing/model': 1.557566481642425, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 0.04}) +[2026-06-15 19:26:27,437][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.40227019786834717, 'train/grad_norm_pre_clip': 3.954550266265869, 'timing/data': 0.020538039971143007, 'timing/model': 1.5489941998384893, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 0.04}) +[2026-06-15 19:26:33,683][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.5644968152046204, 'train/grad_norm_pre_clip': 14.227123260498047, 'timing/data': 0.021167803090065718, 'timing/model': 1.790140840690583, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 0.04}) +[2026-06-15 19:26:39,799][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.3647221028804779, 'train/grad_norm_pre_clip': 6.356745719909668, 'timing/data': 0.021220934111624956, 'timing/model': 1.5863161701709032, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 0.04}) +[2026-06-15 19:26:46,050][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.5076920390129089, 'train/grad_norm_pre_clip': 7.360154151916504, 'timing/data': 0.022450950928032398, 'timing/model': 1.557408683001995, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 0.04}) +[2026-06-15 19:26:52,344][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.3393617868423462, 'train/grad_norm_pre_clip': 5.369395732879639, 'timing/data': 0.020458167884498835, 'timing/model': 1.5540972840972245, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 0.04}) +[2026-06-15 19:26:58,625][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.37651029229164124, 'train/grad_norm_pre_clip': 4.983434677124023, 'timing/data': 0.02509116707369685, 'timing/model': 1.6063139238394797, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 0.04}) +[2026-06-15 19:27:04,845][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.27228590846061707, 'train/grad_norm_pre_clip': 4.385591506958008, 'timing/data': 0.02049299795180559, 'timing/model': 1.5490130698308349, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 0.04}) +[2026-06-15 19:27:11,083][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.40199601650238037, 'train/grad_norm_pre_clip': 9.49184513092041, 'timing/data': 0.020036954898387194, 'timing/model': 1.7860544510185719, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 0.04}) +[2026-06-15 19:27:17,150][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.5174397230148315, 'train/grad_norm_pre_clip': 14.411637306213379, 'timing/data': 0.022304928861558437, 'timing/model': 1.5360642271116376, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.04}) +[2026-06-15 19:27:23,453][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.4082837998867035, 'train/grad_norm_pre_clip': 8.169565200805664, 'timing/data': 0.022108986973762512, 'timing/model': 1.550637022126466, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 0.04}) +[2026-06-15 19:27:29,757][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.30087822675704956, 'train/grad_norm_pre_clip': 6.43845272064209, 'timing/data': 0.019659381825476885, 'timing/model': 1.5537725458852947, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 0.05}) +[2026-06-15 19:27:36,058][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.5823466181755066, 'train/grad_norm_pre_clip': 23.734989166259766, 'timing/data': 0.021032500080764294, 'timing/model': 1.5498748258687556, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 0.05}) +[2026-06-15 19:27:42,360][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.31828704476356506, 'train/grad_norm_pre_clip': 18.85915184020996, 'timing/data': 0.020493866875767708, 'timing/model': 1.5104955467395484, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 0.05}) +[2026-06-15 19:27:48,688][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.22457729279994965, 'train/grad_norm_pre_clip': 8.759284973144531, 'timing/data': 0.019773761741816998, 'timing/model': 1.7790760612115264, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 0.05}) +[2026-06-15 19:27:54,759][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.4315142333507538, 'train/grad_norm_pre_clip': 14.5382661819458, 'timing/data': 0.020991908852010965, 'timing/model': 1.549551879055798, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 0.05}) +[2026-06-15 19:28:01,058][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.7575641870498657, 'train/grad_norm_pre_clip': 16.18132209777832, 'timing/data': 0.02090412937104702, 'timing/model': 1.5590035528875887, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 0.05}) +[2026-06-15 19:28:07,353][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.4110638201236725, 'train/grad_norm_pre_clip': 42.0384635925293, 'timing/data': 0.020474466029554605, 'timing/model': 1.54565220605582, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 0.05}) +[2026-06-15 19:28:13,752][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.5669968128204346, 'train/grad_norm_pre_clip': 23.660476684570312, 'timing/data': 0.020564746111631393, 'timing/model': 1.579226867761463, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 0.05}) +[2026-06-15 19:28:20,155][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.3472740054130554, 'train/grad_norm_pre_clip': 17.045011520385742, 'timing/data': 0.02037638518959284, 'timing/model': 1.5957651338540018, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.05}) +[2026-06-15 19:28:26,505][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.4398956298828125, 'train/grad_norm_pre_clip': 29.90725326538086, 'timing/data': 0.02145057125017047, 'timing/model': 1.7880948656238616, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 0.05}) +[2026-06-15 19:28:32,562][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.25394347310066223, 'train/grad_norm_pre_clip': 15.522726058959961, 'timing/data': 0.020826198160648346, 'timing/model': 1.5544974128715694, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 0.05}) +[2026-06-15 19:28:38,887][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.29222095012664795, 'train/grad_norm_pre_clip': 10.5040864944458, 'timing/data': 0.020626366138458252, 'timing/model': 1.5772126377560198, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 0.05}) +[2026-06-15 19:28:45,254][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.33386537432670593, 'train/grad_norm_pre_clip': 15.409778594970703, 'timing/data': 0.020721056032925844, 'timing/model': 1.5710099167190492, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 0.05}) +[2026-06-15 19:28:51,654][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.3552438020706177, 'train/grad_norm_pre_clip': 10.451329231262207, 'timing/data': 0.020923397038131952, 'timing/model': 1.5542501010932028, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 0.05}) +[2026-06-15 19:28:57,999][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.2772792875766754, 'train/grad_norm_pre_clip': 20.601665496826172, 'timing/data': 0.021292250137776136, 'timing/model': 1.549358957912773, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 0.05}) +[2026-06-15 19:29:04,329][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.37208229303359985, 'train/grad_norm_pre_clip': 7.182112693786621, 'timing/data': 0.020567444618791342, 'timing/model': 1.804732614196837, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 0.05}) +[2026-06-15 19:29:10,737][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.3142251968383789, 'train/grad_norm_pre_clip': 8.216596603393555, 'timing/data': 0.020214643329381943, 'timing/model': 1.6486747730523348, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 0.05}) +[2026-06-15 19:29:17,065][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.30862462520599365, 'train/grad_norm_pre_clip': 9.020200729370117, 'timing/data': 0.02106040669605136, 'timing/model': 1.566078919917345, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 0.05}) +[2026-06-15 19:29:23,366][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.37661680579185486, 'train/grad_norm_pre_clip': 19.10415267944336, 'timing/data': 0.023387731984257698, 'timing/model': 1.553076087962836, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.05}) +[2026-06-15 19:29:29,745][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.30190175771713257, 'train/grad_norm_pre_clip': 9.613755226135254, 'timing/data': 0.01951038697734475, 'timing/model': 1.5805956427939236, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 0.05}) +[2026-06-15 19:29:36,073][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.3930111527442932, 'train/grad_norm_pre_clip': 13.850037574768066, 'timing/data': 0.019904169719666243, 'timing/model': 1.5421314481645823, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 0.05}) +[2026-06-15 19:29:42,511][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.23018139600753784, 'train/grad_norm_pre_clip': 15.771553993225098, 'timing/data': 0.022187544032931328, 'timing/model': 1.7937664603814483, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 0.05}) +[2026-06-15 19:29:48,735][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.3578856885433197, 'train/grad_norm_pre_clip': 20.184301376342773, 'timing/data': 0.02210932271555066, 'timing/model': 1.7186776651069522, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 0.05}) +[2026-06-15 19:29:54,810][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.2857016921043396, 'train/grad_norm_pre_clip': 8.605198860168457, 'timing/data': 0.020808294881135225, 'timing/model': 1.5956713240593672, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.05}) +[2026-06-15 19:30:01,089][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.6151934266090393, 'train/grad_norm_pre_clip': 35.89986038208008, 'timing/data': 0.021566390059888363, 'timing/model': 1.5759164839982986, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 0.05}) +[2026-06-15 19:30:07,512][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.4724236726760864, 'train/grad_norm_pre_clip': 19.349706649780273, 'timing/data': 0.021024946123361588, 'timing/model': 1.6075427308678627, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 0.06}) +[2026-06-15 19:30:13,773][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.2765207290649414, 'train/grad_norm_pre_clip': 14.609533309936523, 'timing/data': 0.020585543010383844, 'timing/model': 1.5269658779725432, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 0.06}) +[2026-06-15 19:30:20,095][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.28462278842926025, 'train/grad_norm_pre_clip': 11.763858795166016, 'timing/data': 0.02336103981360793, 'timing/model': 1.7688144510611892, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 0.06}) +[2026-06-15 19:30:26,501][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.5113676190376282, 'train/grad_norm_pre_clip': 18.87097930908203, 'timing/data': 0.021015744656324387, 'timing/model': 1.806415742263198, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.06}) +[2026-06-15 19:30:32,568][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.40718042850494385, 'train/grad_norm_pre_clip': 19.69118881225586, 'timing/data': 0.02017494011670351, 'timing/model': 1.5488911257125437, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 0.06}) +[2026-06-15 19:30:38,869][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.2893257141113281, 'train/grad_norm_pre_clip': 7.273366451263428, 'timing/data': 0.022042850963771343, 'timing/model': 1.5516846817918122, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 0.06}) +[2026-06-15 19:30:45,357][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.7750146389007568, 'train/grad_norm_pre_clip': 52.414283752441406, 'timing/data': 0.021343247033655643, 'timing/model': 1.6657553277909756, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 0.06}) +[2026-06-15 19:30:51,747][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.563769519329071, 'train/grad_norm_pre_clip': 35.66195297241211, 'timing/data': 0.021510177291929722, 'timing/model': 1.5908007621765137, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 0.06}) +[2026-06-15 19:30:58,107][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.42035067081451416, 'train/grad_norm_pre_clip': 16.83432388305664, 'timing/data': 0.020942464005202055, 'timing/model': 1.7901102229952812, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 0.06}) +[2026-06-15 19:31:04,175][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.3584233820438385, 'train/grad_norm_pre_clip': 16.3569393157959, 'timing/data': 0.02061012340709567, 'timing/model': 1.5503705409355462, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 0.06}) +[2026-06-15 19:31:10,566][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.5292676687240601, 'train/grad_norm_pre_clip': 26.913101196289062, 'timing/data': 0.021121964789927006, 'timing/model': 1.551745519042015, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 0.06}) +[2026-06-15 19:31:16,865][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.9604073762893677, 'train/grad_norm_pre_clip': 41.82682800292969, 'timing/data': 0.02126440592110157, 'timing/model': 1.552242322359234, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 0.06}) +[2026-06-15 19:31:23,195][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.4979386329650879, 'train/grad_norm_pre_clip': 40.225502014160156, 'timing/data': 0.02076867315918207, 'timing/model': 1.5478857760317624, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.06}) +[2026-06-15 19:31:29,566][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.2994846999645233, 'train/grad_norm_pre_clip': 12.189861297607422, 'timing/data': 0.02193122962489724, 'timing/model': 1.575830680783838, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.06}) +[2026-06-15 19:31:35,862][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.3058517575263977, 'train/grad_norm_pre_clip': 11.292652130126953, 'timing/data': 0.021053203847259283, 'timing/model': 1.7423345921561122, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 0.06}) +[2026-06-15 19:31:42,045][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.6625829935073853, 'train/grad_norm_pre_clip': 53.94601058959961, 'timing/data': 0.021041593980044127, 'timing/model': 1.630999956279993, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 0.06}) +[2026-06-15 19:31:48,268][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.4126063585281372, 'train/grad_norm_pre_clip': 45.777347564697266, 'timing/data': 0.021613926626741886, 'timing/model': 1.5437369830906391, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 0.06}) +[2026-06-15 19:31:54,568][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.5113959908485413, 'train/grad_norm_pre_clip': 9.55112075805664, 'timing/data': 0.020675621926784515, 'timing/model': 1.5568114998750389, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 0.06}) +[2026-06-15 19:32:01,059][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.5409514904022217, 'train/grad_norm_pre_clip': 25.299379348754883, 'timing/data': 0.02011703886091709, 'timing/model': 1.5917230057530105, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 0.06}) +[2026-06-15 19:32:07,460][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.35504350066185, 'train/grad_norm_pre_clip': 6.271706581115723, 'timing/data': 0.02051133057102561, 'timing/model': 1.5938347592018545, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 0.06}) +[2026-06-15 19:32:13,768][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.5689612627029419, 'train/grad_norm_pre_clip': 19.268959045410156, 'timing/data': 0.02009001700207591, 'timing/model': 1.7538139871321619, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 0.06}) +[2026-06-15 19:32:19,862][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.535481333732605, 'train/grad_norm_pre_clip': 11.353595733642578, 'timing/data': 0.020634061191231012, 'timing/model': 1.5449070241302252, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 0.06}) +[2026-06-15 19:32:26,168][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.48448610305786133, 'train/grad_norm_pre_clip': 11.290903091430664, 'timing/data': 0.021013572812080383, 'timing/model': 1.5541132800281048, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 0.06}) +[2026-06-15 19:32:32,465][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.21478164196014404, 'train/grad_norm_pre_clip': 7.80718469619751, 'timing/data': 0.020964832976460457, 'timing/model': 1.5477498536929488, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.06}) +[2026-06-15 19:32:38,871][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.4185712933540344, 'train/grad_norm_pre_clip': 5.933242321014404, 'timing/data': 0.021361923776566982, 'timing/model': 1.6182130672968924, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 0.07}) +[2026-06-15 19:32:45,221][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.43382593989372253, 'train/grad_norm_pre_clip': 11.719344139099121, 'timing/data': 0.02217931905761361, 'timing/model': 1.806512823794037, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 0.07}) +[2026-06-15 19:32:51,267][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.4135209918022156, 'train/grad_norm_pre_clip': 9.967572212219238, 'timing/data': 0.02125689433887601, 'timing/model': 1.5541307660751045, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 0.07}) +[2026-06-15 19:32:57,564][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.44461143016815186, 'train/grad_norm_pre_clip': 9.563761711120605, 'timing/data': 0.0212115328758955, 'timing/model': 1.5486083459109068, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 0.07}) +[2026-06-15 19:33:03,768][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.2811925709247589, 'train/grad_norm_pre_clip': 13.795652389526367, 'timing/data': 0.02035491866990924, 'timing/model': 1.5441471310332417, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 0.07}) +[2026-06-15 19:33:10,071][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.3996405303478241, 'train/grad_norm_pre_clip': 7.297816276550293, 'timing/data': 0.02045052871108055, 'timing/model': 1.5548607558012009, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 0.07}) +[2026-06-15 19:33:16,365][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.3165474534034729, 'train/grad_norm_pre_clip': 6.686971664428711, 'timing/data': 0.020545358769595623, 'timing/model': 1.548239469062537, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 0.07}) +[2026-06-15 19:33:22,734][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.39397385716438293, 'train/grad_norm_pre_clip': 7.5167036056518555, 'timing/data': 0.02156650461256504, 'timing/model': 1.8120995396748185, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 0.07}) +[2026-06-15 19:33:28,786][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.3724750280380249, 'train/grad_norm_pre_clip': 17.081573486328125, 'timing/data': 0.020763221196830273, 'timing/model': 1.54180229594931, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 0.07}) +[2026-06-15 19:33:35,171][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.34175315499305725, 'train/grad_norm_pre_clip': 7.229446887969971, 'timing/data': 0.021364992950111628, 'timing/model': 1.5577055611647666, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.07}) +[2026-06-15 19:33:41,463][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.42644479870796204, 'train/grad_norm_pre_clip': 12.299019813537598, 'timing/data': 0.021860715933144093, 'timing/model': 1.548704442102462, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 0.07}) +[2026-06-15 19:33:47,767][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.23987819254398346, 'train/grad_norm_pre_clip': 4.973797798156738, 'timing/data': 0.021284783259034157, 'timing/model': 1.5514177298173308, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 0.07}) +[2026-06-15 19:33:54,142][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.31192097067832947, 'train/grad_norm_pre_clip': 7.53944730758667, 'timing/data': 0.020653780084103346, 'timing/model': 1.6321394378319383, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 0.07}) +[2026-06-15 19:34:00,494][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.32652437686920166, 'train/grad_norm_pre_clip': 7.601837158203125, 'timing/data': 0.018064294010400772, 'timing/model': 1.844638234935701, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 0.07}) +[2026-06-15 19:34:06,661][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.3540346622467041, 'train/grad_norm_pre_clip': 25.271303176879883, 'timing/data': 0.02056765789166093, 'timing/model': 1.5589047609828413, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 0.07}) +[2026-06-15 19:34:12,960][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.45813485980033875, 'train/grad_norm_pre_clip': 16.020036697387695, 'timing/data': 0.021714515052735806, 'timing/model': 1.5456427866593003, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 0.07}) +[2026-06-15 19:34:19,270][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.3344852328300476, 'train/grad_norm_pre_clip': 20.085769653320312, 'timing/data': 0.021444147918373346, 'timing/model': 1.5436547501012683, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 0.07}) +[2026-06-15 19:34:25,560][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.4314591586589813, 'train/grad_norm_pre_clip': 49.719051361083984, 'timing/data': 0.021465007681399584, 'timing/model': 1.5177094051614404, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 0.07}) +[2026-06-15 19:34:31,859][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.7640119194984436, 'train/grad_norm_pre_clip': 40.160430908203125, 'timing/data': 0.020791137125343084, 'timing/model': 1.5497703743167222, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 0.07}) +[2026-06-15 19:34:38,194][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.8150393962860107, 'train/grad_norm_pre_clip': 31.900667190551758, 'timing/data': 0.021980410907417536, 'timing/model': 1.7548400950618088, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.07}) +[2026-06-15 19:34:44,265][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.5893721580505371, 'train/grad_norm_pre_clip': 33.03632736206055, 'timing/data': 0.02181413071230054, 'timing/model': 1.5486851721070707, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 0.07}) +[2026-06-15 19:34:50,566][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.3318467140197754, 'train/grad_norm_pre_clip': 25.440982818603516, 'timing/data': 0.02174880215898156, 'timing/model': 1.5550308781675994, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 0.07}) +[2026-06-15 19:34:56,964][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.39468592405319214, 'train/grad_norm_pre_clip': 9.467765808105469, 'timing/data': 0.02108905976638198, 'timing/model': 1.5592917269095778, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 0.07}) +[2026-06-15 19:35:03,337][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.22161714732646942, 'train/grad_norm_pre_clip': 11.40071964263916, 'timing/data': 0.020495297852903605, 'timing/model': 1.6258533149957657, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 0.07}) +[2026-06-15 19:35:09,632][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.3247668147087097, 'train/grad_norm_pre_clip': 10.747333526611328, 'timing/data': 0.02049452904611826, 'timing/model': 1.5815671822056174, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 0.07}) +[2026-06-15 19:35:15,892][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.3813264071941376, 'train/grad_norm_pre_clip': 15.632757186889648, 'timing/data': 0.019383435137569904, 'timing/model': 1.7820878829807043, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 0.08}) +[2026-06-15 19:35:21,976][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.25419652462005615, 'train/grad_norm_pre_clip': 6.269298076629639, 'timing/data': 0.021113282069563866, 'timing/model': 1.5492862379178405, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 0.08}) +[2026-06-15 19:35:28,270][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.3539491891860962, 'train/grad_norm_pre_clip': 18.39293670654297, 'timing/data': 0.0208980911411345, 'timing/model': 1.5567490672692657, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 0.08}) +[2026-06-15 19:35:34,565][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.21961955726146698, 'train/grad_norm_pre_clip': 10.998851776123047, 'timing/data': 0.021173511631786823, 'timing/model': 1.5521152652800083, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 0.08}) +[2026-06-15 19:35:40,872][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.24346330761909485, 'train/grad_norm_pre_clip': 7.1225714683532715, 'timing/data': 0.01988503895699978, 'timing/model': 1.5561080500483513, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.08}) +[2026-06-15 19:35:47,213][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.18061038851737976, 'train/grad_norm_pre_clip': 13.623771667480469, 'timing/data': 0.02210337668657303, 'timing/model': 1.5625091348774731, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 0.08}) +[2026-06-15 19:35:53,604][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.37331047654151917, 'train/grad_norm_pre_clip': 14.717140197753906, 'timing/data': 0.019845580216497183, 'timing/model': 1.8088266388513148, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 0.08}) +[2026-06-15 19:35:59,689][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.43484967947006226, 'train/grad_norm_pre_clip': 20.663799285888672, 'timing/data': 0.027073755394667387, 'timing/model': 1.5637406120076776, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 0.08}) +[2026-06-15 19:36:05,969][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.715147852897644, 'train/grad_norm_pre_clip': 18.58837890625, 'timing/data': 0.02071770327165723, 'timing/model': 1.5592226558364928, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 0.08}) +[2026-06-15 19:36:12,390][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.14955370128154755, 'train/grad_norm_pre_clip': 18.65339469909668, 'timing/data': 0.022893741726875305, 'timing/model': 1.6659582369029522, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 0.08}) +[2026-06-15 19:36:18,673][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.28435954451560974, 'train/grad_norm_pre_clip': 9.299210548400879, 'timing/data': 0.02141304686665535, 'timing/model': 1.5676319072954357, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 0.08}) +[2026-06-15 19:36:24,987][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.25354039669036865, 'train/grad_norm_pre_clip': 9.689739227294922, 'timing/data': 0.021622806787490845, 'timing/model': 1.5399818569421768, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 0.08}) +[2026-06-15 19:36:31,323][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.3795663118362427, 'train/grad_norm_pre_clip': 24.24373435974121, 'timing/data': 0.02610930521041155, 'timing/model': 1.7669381583109498, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.08}) +[2026-06-15 19:36:37,365][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.6125909090042114, 'train/grad_norm_pre_clip': 43.839576721191406, 'timing/data': 0.020407263655215502, 'timing/model': 1.5386189189739525, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 0.08}) +[2026-06-15 19:47:58,523][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.6563220024108887, 'train/grad_norm_pre_clip': 31.438732147216797, 'eval/loss': 0.1973312646150589, 'eval/flap_precision': 0.8274760383386581, 'eval/flap_recall': 0.7066848567530696, 'eval/flap_f1': 0.7623252391464311, 'eval/flap_rate_model': 0.1738888888888889, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.7623252391464311, 'timing/data': 0.021262238267809153, 'timing/model': 1.5583262797445059, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.08}) +[2026-06-15 19:47:58,524][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_200_state +[2026-06-15 19:47:58,530][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 19:48:07,147][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_200_state/model.safetensors +[2026-06-15 19:48:27,414][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_200_state/optimizer.bin +[2026-06-15 19:48:27,415][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_200_state/sampler.bin +[2026-06-15 19:48:27,416][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_200_state/sampler_1.bin +[2026-06-15 19:48:27,418][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-15 19:48:27,620][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-15 19:48:27,621][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 19:48:27,637][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 19:48:34,048][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.457278311252594, 'train/grad_norm_pre_clip': 12.028284072875977, 'timing/data': 0.027647562325000763, 'timing/model': 1.8631865549832582, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 0.08}) +[2026-06-15 19:48:40,132][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.4284825325012207, 'train/grad_norm_pre_clip': 7.236093044281006, 'timing/data': 0.023430590983480215, 'timing/model': 1.5609167711809278, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 0.08}) +[2026-06-15 19:48:46,363][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.3963071405887604, 'train/grad_norm_pre_clip': 9.053413391113281, 'timing/data': 0.020396368112415075, 'timing/model': 1.5760073587298393, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 0.08}) +[2026-06-15 19:48:52,644][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.2744212746620178, 'train/grad_norm_pre_clip': 6.679386138916016, 'timing/data': 0.022180526051670313, 'timing/model': 1.547845862340182, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 0.08}) +[2026-06-15 19:48:58,942][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.4465635120868683, 'train/grad_norm_pre_clip': 9.143243789672852, 'timing/data': 0.021309142000973225, 'timing/model': 1.565690094139427, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.08}) +[2026-06-15 19:49:05,126][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.34999293088912964, 'train/grad_norm_pre_clip': 5.790792465209961, 'timing/data': 0.022124625742435455, 'timing/model': 1.5696164187975228, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 0.08}) +[2026-06-15 19:49:11,341][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.41370251774787903, 'train/grad_norm_pre_clip': 11.978887557983398, 'timing/data': 0.021768405102193356, 'timing/model': 1.7466569542884827, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 0.08}) +[2026-06-15 19:49:17,354][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.38013169169425964, 'train/grad_norm_pre_clip': 10.903827667236328, 'timing/data': 0.021249532233923674, 'timing/model': 1.5591020612046123, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 0.08}) +[2026-06-15 19:49:23,645][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.38682201504707336, 'train/grad_norm_pre_clip': 7.661871910095215, 'timing/data': 0.021904106251895428, 'timing/model': 1.548756378237158, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 0.08}) +[2026-06-15 19:49:29,842][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.20077656209468842, 'train/grad_norm_pre_clip': 8.456083297729492, 'timing/data': 0.020469298120588064, 'timing/model': 1.5413340390659869, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.08}) +[2026-06-15 19:49:36,043][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.28706857562065125, 'train/grad_norm_pre_clip': 7.177474021911621, 'timing/data': 0.021004100795835257, 'timing/model': 1.5148697383701801, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 0.09}) +[2026-06-15 19:49:42,324][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.24289089441299438, 'train/grad_norm_pre_clip': 8.674004554748535, 'timing/data': 0.020462557673454285, 'timing/model': 1.5702730808407068, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 0.09}) +[2026-06-15 19:49:48,546][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.49645179510116577, 'train/grad_norm_pre_clip': 37.091407775878906, 'timing/data': 0.011039022821933031, 'timing/model': 1.7482042638584971, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 0.09}) +[2026-06-15 19:49:54,654][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.26886802911758423, 'train/grad_norm_pre_clip': 16.718002319335938, 'timing/data': 0.011415654793381691, 'timing/model': 1.5652516479603946, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 0.09}) +[2026-06-15 19:50:00,955][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.12360422313213348, 'train/grad_norm_pre_clip': 10.563070297241211, 'timing/data': 0.011020913254469633, 'timing/model': 1.5641006492078304, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 0.09}) +[2026-06-15 19:50:07,260][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.23006027936935425, 'train/grad_norm_pre_clip': 17.51976203918457, 'timing/data': 0.012547191232442856, 'timing/model': 1.5692796688526869, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 0.09}) +[2026-06-15 19:50:13,554][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.6452605724334717, 'train/grad_norm_pre_clip': 31.064756393432617, 'timing/data': 0.010974023025482893, 'timing/model': 1.5146295819431543, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 0.09}) +[2026-06-15 19:50:19,897][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.1625184565782547, 'train/grad_norm_pre_clip': 9.339861869812012, 'timing/data': 0.011091383174061775, 'timing/model': 1.7982024890370667, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 0.09}) +[2026-06-15 19:50:25,961][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.42293524742126465, 'train/grad_norm_pre_clip': 27.972482681274414, 'timing/data': 0.011528436094522476, 'timing/model': 1.5310219717212021, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 0.09}) +[2026-06-15 19:50:32,259][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.0829358771443367, 'train/grad_norm_pre_clip': 5.83635950088501, 'timing/data': 0.011412895284593105, 'timing/model': 1.5573669448494911, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.09}) +[2026-06-15 19:50:38,461][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.2498374879360199, 'train/grad_norm_pre_clip': 13.65011978149414, 'timing/data': 0.01119485404342413, 'timing/model': 1.560310305096209, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 0.09}) +[2026-06-15 19:50:44,761][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.1303916573524475, 'train/grad_norm_pre_clip': 9.960960388183594, 'timing/data': 0.01096578361466527, 'timing/model': 1.5676667359657586, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 0.09}) +[2026-06-15 19:50:51,052][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.5156023502349854, 'train/grad_norm_pre_clip': 12.381437301635742, 'timing/data': 0.010702542029321194, 'timing/model': 1.5693250191397965, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 0.09}) +[2026-06-15 19:50:57,515][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.3052564561367035, 'train/grad_norm_pre_clip': 20.714672088623047, 'timing/data': 0.03022255701944232, 'timing/model': 1.8977144109085202, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 0.09}) +[2026-06-15 19:51:03,561][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.2336834967136383, 'train/grad_norm_pre_clip': 11.340802192687988, 'timing/data': 0.011160133872181177, 'timing/model': 1.5431684432551265, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 0.09}) +[2026-06-15 19:51:09,865][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.25579580664634705, 'train/grad_norm_pre_clip': 10.120780944824219, 'timing/data': 0.011094664223492146, 'timing/model': 1.5587274441495538, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 0.09}) +[2026-06-15 19:51:16,293][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.26477810740470886, 'train/grad_norm_pre_clip': 10.416733741760254, 'timing/data': 0.01126095512881875, 'timing/model': 1.5972342370077968, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 0.09}) +[2026-06-15 19:51:22,789][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.39943888783454895, 'train/grad_norm_pre_clip': 11.182241439819336, 'timing/data': 0.011559995822608471, 'timing/model': 1.6074576321989298, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 0.09}) +[2026-06-15 19:51:29,068][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.44212841987609863, 'train/grad_norm_pre_clip': 14.446889877319336, 'timing/data': 0.011705337092280388, 'timing/model': 1.5608770498074591, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 0.09}) +[2026-06-15 19:51:35,396][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.27655282616615295, 'train/grad_norm_pre_clip': 7.997403621673584, 'timing/data': 0.011371336411684752, 'timing/model': 1.8009820063598454, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.09}) +[2026-06-15 19:51:41,463][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.2156861126422882, 'train/grad_norm_pre_clip': 7.862110614776611, 'timing/data': 0.010781112127006054, 'timing/model': 1.5675684232264757, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 0.09}) +[2026-06-15 19:51:47,859][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.237616628408432, 'train/grad_norm_pre_clip': 8.870927810668945, 'timing/data': 0.010824273340404034, 'timing/model': 1.557336307130754, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 0.09}) +[2026-06-15 19:51:54,171][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.4562191367149353, 'train/grad_norm_pre_clip': 23.535972595214844, 'timing/data': 0.010630012024194002, 'timing/model': 1.567521742079407, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 0.09}) +[2026-06-15 19:52:00,477][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.25793126225471497, 'train/grad_norm_pre_clip': 11.693217277526855, 'timing/data': 0.012034059036523104, 'timing/model': 1.5316112306900322, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 0.09}) +[2026-06-15 19:52:06,761][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.42976900935173035, 'train/grad_norm_pre_clip': 9.326051712036133, 'timing/data': 0.011327285319566727, 'timing/model': 1.5412726528011262, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 0.1}) +[2026-06-15 19:52:13,227][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.23534832894802094, 'train/grad_norm_pre_clip': 5.103386878967285, 'timing/data': 0.021922937128692865, 'timing/model': 1.8492032359354198, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 0.1}) +[2026-06-15 19:52:19,271][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.2677527666091919, 'train/grad_norm_pre_clip': 10.864368438720703, 'timing/data': 0.010383261367678642, 'timing/model': 1.5674131200648844, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 0.1}) +[2026-06-15 19:52:25,565][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.4635869264602661, 'train/grad_norm_pre_clip': 12.36182689666748, 'timing/data': 0.01099715381860733, 'timing/model': 1.5631060739979148, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 0.1}) +[2026-06-15 19:52:31,867][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.32141777873039246, 'train/grad_norm_pre_clip': 13.926366806030273, 'timing/data': 0.01085771294310689, 'timing/model': 1.572085693012923, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 0.1}) +[2026-06-15 19:52:38,164][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.22469012439250946, 'train/grad_norm_pre_clip': 13.763534545898438, 'timing/data': 0.011853109113872051, 'timing/model': 1.5570578225888312, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.1}) +[2026-06-15 19:52:44,472][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.326541543006897, 'train/grad_norm_pre_clip': 8.702664375305176, 'timing/data': 0.010965873952955008, 'timing/model': 1.5279573341831565, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 0.1}) +[2026-06-15 19:52:50,803][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.3395931124687195, 'train/grad_norm_pre_clip': 9.154637336730957, 'timing/data': 0.011254374869167805, 'timing/model': 1.8027264289557934, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 0.1}) +[2026-06-15 19:52:56,997][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.17505595088005066, 'train/grad_norm_pre_clip': 9.709901809692383, 'timing/data': 0.010203159879893064, 'timing/model': 1.5650383750908077, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 0.1}) +[2026-06-15 19:53:03,480][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.44331270456314087, 'train/grad_norm_pre_clip': 9.502130508422852, 'timing/data': 0.011445705778896809, 'timing/model': 1.557327541988343, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 0.1}) +[2026-06-15 19:53:09,785][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.37771186232566833, 'train/grad_norm_pre_clip': 10.71811580657959, 'timing/data': 0.011184655129909515, 'timing/model': 1.5815228349529207, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 0.1}) +[2026-06-15 19:53:16,075][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.22325532138347626, 'train/grad_norm_pre_clip': 8.768767356872559, 'timing/data': 0.01086351415142417, 'timing/model': 1.5678070010617375, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 0.1}) +[2026-06-15 19:53:22,483][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.2516746520996094, 'train/grad_norm_pre_clip': 6.658356189727783, 'timing/data': 0.010999334044754505, 'timing/model': 1.6112346895970404, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.1}) +[2026-06-15 19:53:29,109][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.3044557273387909, 'train/grad_norm_pre_clip': 22.55011558532715, 'timing/data': 0.009815528057515621, 'timing/model': 1.9085529898293316, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 0.1}) +[2026-06-15 19:53:35,289][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.5714868307113647, 'train/grad_norm_pre_clip': 9.491632461547852, 'timing/data': 0.010791482869535685, 'timing/model': 1.601716980803758, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 0.1}) +[2026-06-15 19:53:41,595][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.1812845915555954, 'train/grad_norm_pre_clip': 4.796756744384766, 'timing/data': 0.010535132139921188, 'timing/model': 1.5909069920890033, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.1}) +[2026-06-15 19:53:47,890][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.27143460512161255, 'train/grad_norm_pre_clip': 12.636021614074707, 'timing/data': 0.011443255934864283, 'timing/model': 1.5897722709923983, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 0.1}) +[2026-06-15 19:53:54,165][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.24832631647586823, 'train/grad_norm_pre_clip': 10.447476387023926, 'timing/data': 0.010699742939323187, 'timing/model': 1.5518680666573346, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 0.1}) +[2026-06-15 19:54:00,548][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.26472005248069763, 'train/grad_norm_pre_clip': 11.771641731262207, 'timing/data': 0.011192325036972761, 'timing/model': 1.5857642679475248, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 0.1}) +[2026-06-15 19:54:07,018][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.20450641214847565, 'train/grad_norm_pre_clip': 9.639881134033203, 'timing/data': 0.011511616874486208, 'timing/model': 1.8211439861916006, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 0.1}) +[2026-06-15 19:54:13,156][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.17075146734714508, 'train/grad_norm_pre_clip': 12.54137897491455, 'timing/data': 0.012062740046530962, 'timing/model': 1.5990036008879542, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 0.1}) +[2026-06-15 19:54:19,464][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.4786509573459625, 'train/grad_norm_pre_clip': 24.998001098632812, 'timing/data': 0.011718017980456352, 'timing/model': 1.5377515610307455, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 0.1}) +[2026-06-15 19:54:25,881][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.16233986616134644, 'train/grad_norm_pre_clip': 9.456938743591309, 'timing/data': 0.012244090903550386, 'timing/model': 1.634758535772562, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 0.1}) +[2026-06-15 19:54:32,188][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.34568095207214355, 'train/grad_norm_pre_clip': 25.62000274658203, 'timing/data': 0.01062884321436286, 'timing/model': 1.574470920022577, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 0.1}) +[2026-06-15 19:54:38,572][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.19456490874290466, 'train/grad_norm_pre_clip': 10.405491828918457, 'timing/data': 0.01103669498115778, 'timing/model': 1.524014674127102, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 0.1}) +[2026-06-15 19:54:44,983][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.3523212969303131, 'train/grad_norm_pre_clip': 16.81218147277832, 'timing/data': 0.011184276081621647, 'timing/model': 1.8504914562217891, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.11}) +[2026-06-15 19:54:51,228][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.24134108424186707, 'train/grad_norm_pre_clip': 12.13969898223877, 'timing/data': 0.010078979656100273, 'timing/model': 1.7198090120218694, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 0.11}) +[2026-06-15 19:54:57,267][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.6438300609588623, 'train/grad_norm_pre_clip': 25.009958267211914, 'timing/data': 0.011011585127562284, 'timing/model': 1.557196858804673, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 0.11}) +[2026-06-15 19:55:03,856][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.21587656438350677, 'train/grad_norm_pre_clip': 12.227629661560059, 'timing/data': 0.011115055065602064, 'timing/model': 1.5558233647607267, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 0.11}) +[2026-06-15 19:55:10,272][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.0843101516366005, 'train/grad_norm_pre_clip': 4.275847434997559, 'timing/data': 0.010763504076749086, 'timing/model': 1.586164629086852, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 0.11}) +[2026-06-15 19:55:16,671][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.3017260432243347, 'train/grad_norm_pre_clip': 17.20279312133789, 'timing/data': 0.010866484139114618, 'timing/model': 1.5517532820813358, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 0.11}) +[2026-06-15 19:55:23,038][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.19739001989364624, 'train/grad_norm_pre_clip': 11.948845863342285, 'timing/data': 0.010614593047648668, 'timing/model': 1.824468060862273, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 0.11}) +[2026-06-15 19:55:29,517][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.1975497454404831, 'train/grad_norm_pre_clip': 12.91758918762207, 'timing/data': 0.012033849954605103, 'timing/model': 1.7975363400764763, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 0.11}) +[2026-06-15 19:55:35,575][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.3208054304122925, 'train/grad_norm_pre_clip': 9.328775405883789, 'timing/data': 0.011113086249679327, 'timing/model': 1.5662203426472843, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 0.11}) +[2026-06-15 19:55:41,884][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.5580253601074219, 'train/grad_norm_pre_clip': 16.096494674682617, 'timing/data': 0.010620723012834787, 'timing/model': 1.5758716738782823, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 0.11}) +[2026-06-15 19:55:48,270][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.4444536566734314, 'train/grad_norm_pre_clip': 21.80935287475586, 'timing/data': 0.01146309683099389, 'timing/model': 1.554718571715057, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.11}) +[2026-06-15 19:55:54,573][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.2728692591190338, 'train/grad_norm_pre_clip': 10.404498100280762, 'timing/data': 0.01083910372108221, 'timing/model': 1.5286195650696754, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 0.11}) +[2026-06-15 19:56:00,920][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.5399773716926575, 'train/grad_norm_pre_clip': 14.372503280639648, 'timing/data': 0.011027075350284576, 'timing/model': 1.8061914793215692, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 0.11}) +[2026-06-15 19:56:06,974][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.34374797344207764, 'train/grad_norm_pre_clip': 10.446090698242188, 'timing/data': 0.01130656711757183, 'timing/model': 1.5601864382624626, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 0.11}) +[2026-06-15 19:56:13,281][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.44027402997016907, 'train/grad_norm_pre_clip': 12.647072792053223, 'timing/data': 0.01263479283079505, 'timing/model': 1.5736338179558516, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 0.11}) +[2026-06-15 19:56:19,670][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.1374257653951645, 'train/grad_norm_pre_clip': 6.671651840209961, 'timing/data': 0.01102561503648758, 'timing/model': 1.571466090157628, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.11}) +[2026-06-15 19:56:26,062][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.23448053002357483, 'train/grad_norm_pre_clip': 9.353166580200195, 'timing/data': 0.011002464685589075, 'timing/model': 1.6004666881635785, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 0.11}) +[2026-06-15 19:56:32,440][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.34247323870658875, 'train/grad_norm_pre_clip': 13.845319747924805, 'timing/data': 0.01115590613335371, 'timing/model': 1.654632190708071, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 0.11}) +[2026-06-15 19:56:38,764][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.3493770658969879, 'train/grad_norm_pre_clip': 14.695384979248047, 'timing/data': 0.011228295974433422, 'timing/model': 1.7519351299852133, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 0.11}) +[2026-06-15 19:56:44,922][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.3537168502807617, 'train/grad_norm_pre_clip': 13.79413890838623, 'timing/data': 0.0109850550070405, 'timing/model': 1.6051376801915467, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 0.11}) +[2026-06-15 19:56:51,370][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.25366461277008057, 'train/grad_norm_pre_clip': 5.681841850280762, 'timing/data': 0.01048390194773674, 'timing/model': 1.6532037728466094, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.11}) +[2026-06-15 19:56:57,722][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.4319112002849579, 'train/grad_norm_pre_clip': 9.938785552978516, 'timing/data': 0.011543336790055037, 'timing/model': 1.5919769098982215, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 0.11}) +[2026-06-15 19:57:03,976][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.15586057305335999, 'train/grad_norm_pre_clip': 6.151057720184326, 'timing/data': 0.01091586472466588, 'timing/model': 1.5456248801201582, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 0.11}) +[2026-06-15 19:57:10,344][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.4025033116340637, 'train/grad_norm_pre_clip': 17.63404655456543, 'timing/data': 0.011422597337514162, 'timing/model': 1.5855635930784047, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 0.11}) +[2026-06-15 19:57:16,772][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.46461114287376404, 'train/grad_norm_pre_clip': 33.439781188964844, 'timing/data': 0.011157456319779158, 'timing/model': 1.7517807469703257, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 0.11}) +[2026-06-15 19:57:22,965][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.47249913215637207, 'train/grad_norm_pre_clip': 32.970245361328125, 'timing/data': 0.011259666178375483, 'timing/model': 1.5612139161676168, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 0.12}) +[2026-06-15 19:57:29,267][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.17735667526721954, 'train/grad_norm_pre_clip': 7.271036148071289, 'timing/data': 0.011076655704528093, 'timing/model': 1.556886007077992, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 0.12}) +[2026-06-15 19:57:35,675][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.2979869246482849, 'train/grad_norm_pre_clip': 8.713915824890137, 'timing/data': 0.010606973432004452, 'timing/model': 1.568865149281919, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 0.12}) +[2026-06-15 19:57:41,963][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.1182437315583229, 'train/grad_norm_pre_clip': 4.740376949310303, 'timing/data': 0.01074045430868864, 'timing/model': 1.5227510607801378, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 0.12}) +[2026-06-15 19:57:48,302][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.38784289360046387, 'train/grad_norm_pre_clip': 15.247865676879883, 'timing/data': 0.01039518229663372, 'timing/model': 1.7959815030917525, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 0.12}) +[2026-06-15 19:57:54,371][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.21844162046909332, 'train/grad_norm_pre_clip': 6.594229221343994, 'timing/data': 0.011017085053026676, 'timing/model': 1.5647171856835485, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.12}) +[2026-06-15 19:58:00,673][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.2931751608848572, 'train/grad_norm_pre_clip': 8.252572059631348, 'timing/data': 0.010607433039695024, 'timing/model': 1.5677840039134026, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 0.12}) +[2026-06-15 19:58:06,867][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.2229926735162735, 'train/grad_norm_pre_clip': 11.926592826843262, 'timing/data': 0.010685533285140991, 'timing/model': 1.5609022718854249, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 0.12}) +[2026-06-15 19:58:13,272][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.3811075687408447, 'train/grad_norm_pre_clip': 13.095649719238281, 'timing/data': 0.011002765037119389, 'timing/model': 1.6508235950022936, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 0.12}) +[2026-06-15 19:58:19,571][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.3815315365791321, 'train/grad_norm_pre_clip': 17.543033599853516, 'timing/data': 0.01090701436623931, 'timing/model': 1.5597936711274087, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 0.12}) +[2026-06-15 19:58:25,909][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.22578439116477966, 'train/grad_norm_pre_clip': 9.814435005187988, 'timing/data': 0.011286837048828602, 'timing/model': 1.8000334091484547, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 0.12}) +[2026-06-15 19:58:31,971][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.22957244515419006, 'train/grad_norm_pre_clip': 8.406964302062988, 'timing/data': 0.010835614055395126, 'timing/model': 1.5570610407739878, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.12}) +[2026-06-15 19:58:38,265][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.5483612418174744, 'train/grad_norm_pre_clip': 23.894393920898438, 'timing/data': 0.011031534988433123, 'timing/model': 1.5586295207031071, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 0.12}) +[2026-06-15 19:58:44,666][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.3534707725048065, 'train/grad_norm_pre_clip': 12.967988967895508, 'timing/data': 0.010975615121424198, 'timing/model': 1.5636728387326002, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 0.12}) +[2026-06-15 19:58:50,971][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.21697868406772614, 'train/grad_norm_pre_clip': 9.4119291305542, 'timing/data': 0.011676257941871881, 'timing/model': 1.5682510640472174, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 0.12}) +[2026-06-15 20:10:12,422][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.23960058391094208, 'train/grad_norm_pre_clip': 9.294364929199219, 'eval/loss': 0.16394226253032684, 'eval/flap_precision': 0.8473389355742297, 'eval/flap_recall': 0.82537517053206, 'eval/flap_f1': 0.8362128541810644, 'eval/flap_rate_model': 0.19833333333333333, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.8362128541810644, 'timing/data': 0.011065836064517498, 'timing/model': 1.5530413589440286, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.12}) +[2026-06-15 20:10:12,423][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_300_state +[2026-06-15 20:10:12,430][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:10:20,988][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_300_state/model.safetensors +[2026-06-15 20:10:41,071][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_300_state/optimizer.bin +[2026-06-15 20:10:41,072][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_300_state/sampler.bin +[2026-06-15 20:10:41,072][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_300_state/sampler_1.bin +[2026-06-15 20:10:41,073][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-15 20:10:41,261][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-15 20:10:41,262][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:10:41,277][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:10:47,635][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.24055486917495728, 'train/grad_norm_pre_clip': 8.22108268737793, 'timing/data': 0.011405594646930695, 'timing/model': 1.5488384277559817, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 0.12}) +[2026-06-15 20:10:53,845][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.44351819157600403, 'train/grad_norm_pre_clip': 24.565780639648438, 'timing/data': 0.009991846978664398, 'timing/model': 1.538808252196759, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 0.12}) +[2026-06-15 20:11:00,172][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.3293302357196808, 'train/grad_norm_pre_clip': 9.932758331298828, 'timing/data': 0.01413657097145915, 'timing/model': 1.7993726572021842, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 0.12}) +[2026-06-15 20:11:06,241][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.2823663353919983, 'train/grad_norm_pre_clip': 10.037088394165039, 'timing/data': 0.011322743725031614, 'timing/model': 1.5609930511564016, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 0.12}) +[2026-06-15 20:11:12,542][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.3536687195301056, 'train/grad_norm_pre_clip': 13.489105224609375, 'timing/data': 0.01143236504867673, 'timing/model': 1.5614267736673355, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 0.12}) +[2026-06-15 20:11:18,848][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.229986771941185, 'train/grad_norm_pre_clip': 13.889236450195312, 'timing/data': 0.011856928002089262, 'timing/model': 1.560148196760565, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 0.12}) +[2026-06-15 20:11:25,150][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.32764074206352234, 'train/grad_norm_pre_clip': 10.205927848815918, 'timing/data': 0.010485460050404072, 'timing/model': 1.5391425103880465, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 0.12}) +[2026-06-15 20:11:31,488][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.20100055634975433, 'train/grad_norm_pre_clip': 8.077561378479004, 'timing/data': 0.010734890587627888, 'timing/model': 1.5626874840818346, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 0.12}) +[2026-06-15 20:11:37,783][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.16506530344486237, 'train/grad_norm_pre_clip': 7.690500736236572, 'timing/data': 0.011626805178821087, 'timing/model': 1.7912542899139225, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 0.12}) +[2026-06-15 20:11:43,874][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.20005322992801666, 'train/grad_norm_pre_clip': 11.645831108093262, 'timing/data': 0.010160807054489851, 'timing/model': 1.5771911502815783, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.13}) +[2026-06-15 20:11:50,152][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.3602929413318634, 'train/grad_norm_pre_clip': 13.484512329101562, 'timing/data': 0.010768420062959194, 'timing/model': 1.557910312898457, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 0.13}) +[2026-06-15 20:11:56,356][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.1835024505853653, 'train/grad_norm_pre_clip': 7.718051910400391, 'timing/data': 0.01020487817004323, 'timing/model': 1.5605305191129446, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 0.13}) +[2026-06-15 20:12:02,650][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.16760221123695374, 'train/grad_norm_pre_clip': 6.248000621795654, 'timing/data': 0.010063846129924059, 'timing/model': 1.5592697937972844, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 0.13}) +[2026-06-15 20:12:08,950][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.352468341588974, 'train/grad_norm_pre_clip': 20.414283752441406, 'timing/data': 0.010275078006088734, 'timing/model': 1.5303660230711102, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 0.13}) +[2026-06-15 20:12:15,278][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.2586755156517029, 'train/grad_norm_pre_clip': 10.731477737426758, 'timing/data': 0.011482934933155775, 'timing/model': 1.7907320307567716, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 0.13}) +[2026-06-15 20:12:21,349][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.12779460847377777, 'train/grad_norm_pre_clip': 4.985286235809326, 'timing/data': 0.011170032899826765, 'timing/model': 1.5581313441507518, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 0.13}) +[2026-06-15 20:12:27,667][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.14073845744132996, 'train/grad_norm_pre_clip': 16.408687591552734, 'timing/data': 0.011647204868495464, 'timing/model': 1.5638998989015818, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 0.13}) +[2026-06-15 20:12:34,060][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.13364629447460175, 'train/grad_norm_pre_clip': 5.730810642242432, 'timing/data': 0.011304113082587719, 'timing/model': 1.565467260312289, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 0.13}) +[2026-06-15 20:12:40,356][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.22951123118400574, 'train/grad_norm_pre_clip': 7.849694728851318, 'timing/data': 0.011151982005685568, 'timing/model': 1.5571213350631297, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 0.13}) +[2026-06-15 20:12:46,659][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.31443044543266296, 'train/grad_norm_pre_clip': 11.864238739013672, 'timing/data': 0.010968330781906843, 'timing/model': 1.527424902189523, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.13}) +[2026-06-15 20:12:52,990][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.11044076085090637, 'train/grad_norm_pre_clip': 9.462085723876953, 'timing/data': 0.012312729377299547, 'timing/model': 1.793999016750604, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 0.13}) +[2026-06-15 20:12:59,397][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.19160965085029602, 'train/grad_norm_pre_clip': 6.464025020599365, 'timing/data': 0.01144039398059249, 'timing/model': 1.8010026281699538, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 0.13}) +[2026-06-15 20:13:05,465][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.24535909295082092, 'train/grad_norm_pre_clip': 8.671761512756348, 'timing/data': 0.012838530819863081, 'timing/model': 1.5534917959012091, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 0.13}) +[2026-06-15 20:13:11,763][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.19681531190872192, 'train/grad_norm_pre_clip': 17.449934005737305, 'timing/data': 0.011316393036395311, 'timing/model': 1.5566066061146557, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 0.13}) +[2026-06-15 20:13:18,067][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.974096417427063, 'train/grad_norm_pre_clip': 75.97057342529297, 'timing/data': 0.011977367103099823, 'timing/model': 1.558116838335991, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 0.13}) +[2026-06-15 20:13:24,372][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.2107078582048416, 'train/grad_norm_pre_clip': 15.83574104309082, 'timing/data': 0.011375054251402617, 'timing/model': 1.524127050768584, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 0.13}) +[2026-06-15 20:13:30,799][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.265540212392807, 'train/grad_norm_pre_clip': 16.461925506591797, 'timing/data': 0.011850626207888126, 'timing/model': 1.7881000358611345, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 0.13}) +[2026-06-15 20:13:36,862][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.2014852911233902, 'train/grad_norm_pre_clip': 19.40563201904297, 'timing/data': 0.011539463885128498, 'timing/model': 1.5526944166049361, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 0.13}) +[2026-06-15 20:13:43,173][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.42906370759010315, 'train/grad_norm_pre_clip': 17.654232025146484, 'timing/data': 0.011508853640407324, 'timing/model': 1.5706502897664905, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 0.13}) +[2026-06-15 20:13:49,477][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.29156649112701416, 'train/grad_norm_pre_clip': 9.715052604675293, 'timing/data': 0.01142823277041316, 'timing/model': 1.570243161637336, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.13}) +[2026-06-15 20:13:55,768][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.16646653413772583, 'train/grad_norm_pre_clip': 9.838143348693848, 'timing/data': 0.010769989807158709, 'timing/model': 1.5326195959933102, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 0.13}) +[2026-06-15 20:14:02,050][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.33336248993873596, 'train/grad_norm_pre_clip': 10.638550758361816, 'timing/data': 0.0096784639172256, 'timing/model': 1.5855561727657914, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 0.13}) +[2026-06-15 20:14:08,413][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.3676704168319702, 'train/grad_norm_pre_clip': 12.545938491821289, 'timing/data': 0.010827039834111929, 'timing/model': 1.8002623580396175, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 0.13}) +[2026-06-15 20:14:14,461][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.32098954916000366, 'train/grad_norm_pre_clip': 19.255733489990234, 'timing/data': 0.011551355011761189, 'timing/model': 1.5543087772093713, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 0.14}) +[2026-06-15 20:14:20,770][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.1922679990530014, 'train/grad_norm_pre_clip': 6.447939872741699, 'timing/data': 0.010394377168267965, 'timing/model': 1.5551698058843613, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 0.14}) +[2026-06-15 20:14:27,274][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.21450568735599518, 'train/grad_norm_pre_clip': 5.603684425354004, 'timing/data': 0.010966470930725336, 'timing/model': 1.5585746681317687, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 0.14}) +[2026-06-15 20:14:33,659][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.2738841474056244, 'train/grad_norm_pre_clip': 8.208283424377441, 'timing/data': 0.010863150004297495, 'timing/model': 1.5351254921406507, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 0.14}) +[2026-06-15 20:14:40,062][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.46417003870010376, 'train/grad_norm_pre_clip': 11.68095588684082, 'timing/data': 0.011401193216443062, 'timing/model': 1.6033735638484359, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 0.14}) +[2026-06-15 20:14:46,368][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.30449196696281433, 'train/grad_norm_pre_clip': 47.735137939453125, 'timing/data': 0.010619389358907938, 'timing/model': 1.7387017449364066, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 0.14}) +[2026-06-15 20:14:52,614][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.28262653946876526, 'train/grad_norm_pre_clip': 8.675911903381348, 'timing/data': 0.009734553750604391, 'timing/model': 1.673870470840484, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.14}) +[2026-06-15 20:14:58,995][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.42445388436317444, 'train/grad_norm_pre_clip': 12.27020263671875, 'timing/data': 0.01155050378292799, 'timing/model': 1.5757372328080237, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 0.14}) +[2026-06-15 20:15:05,380][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.30922552943229675, 'train/grad_norm_pre_clip': 15.272953987121582, 'timing/data': 0.011173401959240437, 'timing/model': 1.5677866330370307, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 0.14}) +[2026-06-15 20:15:11,780][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.22802682220935822, 'train/grad_norm_pre_clip': 9.244485855102539, 'timing/data': 0.011907455045729876, 'timing/model': 1.618449626956135, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 0.14}) +[2026-06-15 20:15:18,235][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.2561003863811493, 'train/grad_norm_pre_clip': 31.284879684448242, 'timing/data': 0.010907469782978296, 'timing/model': 1.8204216728918254, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 0.14}) +[2026-06-15 20:15:24,273][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.3165092468261719, 'train/grad_norm_pre_clip': 19.065204620361328, 'timing/data': 0.01153983362019062, 'timing/model': 1.552645152900368, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.14}) +[2026-06-15 20:15:30,764][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.3582264482975006, 'train/grad_norm_pre_clip': 13.815170288085938, 'timing/data': 0.011139231733977795, 'timing/model': 1.6407703720033169, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 0.14}) +[2026-06-15 20:15:37,172][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.2937539219856262, 'train/grad_norm_pre_clip': 14.453980445861816, 'timing/data': 0.010589148383587599, 'timing/model': 1.5968051846139133, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 0.14}) +[2026-06-15 20:15:43,577][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.15295183658599854, 'train/grad_norm_pre_clip': 2.8511805534362793, 'timing/data': 0.010880299843847752, 'timing/model': 1.5781748867593706, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 0.14}) +[2026-06-15 20:15:49,886][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.18110719323158264, 'train/grad_norm_pre_clip': 6.486083030700684, 'timing/data': 0.011437132954597473, 'timing/model': 1.5867636892944574, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 0.14}) +[2026-06-15 20:15:56,235][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.409517377614975, 'train/grad_norm_pre_clip': 15.71204662322998, 'timing/data': 0.011399061884731054, 'timing/model': 1.8234339728951454, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.14}) +[2026-06-15 20:16:02,271][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.24217194318771362, 'train/grad_norm_pre_clip': 12.915421485900879, 'timing/data': 0.01230049692094326, 'timing/model': 1.565872984007001, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 0.14}) +[2026-06-15 20:16:08,576][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.06916216760873795, 'train/grad_norm_pre_clip': 3.3457283973693848, 'timing/data': 0.010857569985091686, 'timing/model': 1.5671772360801697, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 0.14}) +[2026-06-15 20:16:14,906][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.1268959790468216, 'train/grad_norm_pre_clip': 5.662763595581055, 'timing/data': 0.010841238778084517, 'timing/model': 1.5993013666011393, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 0.14}) +[2026-06-15 20:16:21,262][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.28369420766830444, 'train/grad_norm_pre_clip': 11.486666679382324, 'timing/data': 0.010577667970210314, 'timing/model': 1.5469545759260654, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 0.14}) +[2026-06-15 20:16:27,661][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.26881539821624756, 'train/grad_norm_pre_clip': 10.83857536315918, 'timing/data': 0.011140970978885889, 'timing/model': 1.5744563122279942, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 0.14}) +[2026-06-15 20:16:34,019][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.4106850028038025, 'train/grad_norm_pre_clip': 13.298616409301758, 'timing/data': 0.010649728123098612, 'timing/model': 1.8159341169521213, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 0.14}) +[2026-06-15 20:16:40,082][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.13865599036216736, 'train/grad_norm_pre_clip': 11.66440486907959, 'timing/data': 0.011044529732316732, 'timing/model': 1.5745696718804538, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 0.14}) +[2026-06-15 20:16:46,371][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.12025368213653564, 'train/grad_norm_pre_clip': 13.924543380737305, 'timing/data': 0.010968639049679041, 'timing/model': 1.567509570159018, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 0.14}) +[2026-06-15 20:16:52,790][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.35634559392929077, 'train/grad_norm_pre_clip': 28.471412658691406, 'timing/data': 0.010347327217459679, 'timing/model': 1.606155514717102, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 0.15}) +[2026-06-15 20:16:59,074][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.1163640096783638, 'train/grad_norm_pre_clip': 10.43863582611084, 'timing/data': 0.011176310945302248, 'timing/model': 1.5281473849900067, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.15}) +[2026-06-15 20:17:05,374][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.027470657601952553, 'train/grad_norm_pre_clip': 2.8321547508239746, 'timing/data': 0.010756619274616241, 'timing/model': 1.5525401732884347, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 0.15}) +[2026-06-15 20:17:11,698][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.18019255995750427, 'train/grad_norm_pre_clip': 16.65194320678711, 'timing/data': 0.010094485245645046, 'timing/model': 1.7948780599981546, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 0.15}) +[2026-06-15 20:17:17,805][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.31799134612083435, 'train/grad_norm_pre_clip': 13.440937995910645, 'timing/data': 0.010908809956163168, 'timing/model': 1.5960820717737079, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 0.15}) +[2026-06-15 20:17:24,171][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.4498758018016815, 'train/grad_norm_pre_clip': 9.638249397277832, 'timing/data': 0.011171811260282993, 'timing/model': 1.5548120932653546, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 0.15}) +[2026-06-15 20:17:30,553][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.640153706073761, 'train/grad_norm_pre_clip': 14.370264053344727, 'timing/data': 0.011075889691710472, 'timing/model': 1.641689172014594, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 0.15}) +[2026-06-15 20:17:36,889][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.15874715149402618, 'train/grad_norm_pre_clip': 8.110580444335938, 'timing/data': 0.01088894996792078, 'timing/model': 1.579567528795451, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 0.15}) +[2026-06-15 20:17:43,253][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.2295834720134735, 'train/grad_norm_pre_clip': 9.217787742614746, 'timing/data': 0.010569177102297544, 'timing/model': 1.5902435230091214, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 0.15}) +[2026-06-15 20:17:49,724][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.4604927897453308, 'train/grad_norm_pre_clip': 13.107295036315918, 'timing/data': 0.010890118777751923, 'timing/model': 1.8114466797560453, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 0.15}) +[2026-06-15 20:17:55,771][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.4367033541202545, 'train/grad_norm_pre_clip': 8.08796501159668, 'timing/data': 0.010718097910284996, 'timing/model': 1.5627046548761427, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 0.15}) +[2026-06-15 20:18:02,078][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.30852463841438293, 'train/grad_norm_pre_clip': 14.457181930541992, 'timing/data': 0.010412706062197685, 'timing/model': 1.574332954827696, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.15}) +[2026-06-15 20:18:08,393][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.38708701729774475, 'train/grad_norm_pre_clip': 13.04499626159668, 'timing/data': 0.010570806916803122, 'timing/model': 1.5878318240866065, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 0.15}) +[2026-06-15 20:18:14,797][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.3091791570186615, 'train/grad_norm_pre_clip': 8.834648132324219, 'timing/data': 0.010900979395955801, 'timing/model': 1.5694445902481675, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 0.15}) +[2026-06-15 20:18:21,172][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.2053840309381485, 'train/grad_norm_pre_clip': 4.245620250701904, 'timing/data': 0.011888014152646065, 'timing/model': 1.6098470147699118, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 0.15}) +[2026-06-15 20:18:27,508][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.23622457683086395, 'train/grad_norm_pre_clip': 5.22385835647583, 'timing/data': 0.010994620155543089, 'timing/model': 1.7870095097459853, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 0.15}) +[2026-06-15 20:18:33,587][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.3512524962425232, 'train/grad_norm_pre_clip': 7.829273700714111, 'timing/data': 0.011531992815434933, 'timing/model': 1.5421980801038444, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 0.15}) +[2026-06-15 20:18:39,996][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.24482904374599457, 'train/grad_norm_pre_clip': 6.501448154449463, 'timing/data': 0.011460721958428621, 'timing/model': 1.617178772110492, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 0.15}) +[2026-06-15 20:18:46,383][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.2746471166610718, 'train/grad_norm_pre_clip': 7.9297404289245605, 'timing/data': 0.010628737043589354, 'timing/model': 1.562562343198806, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 0.15}) +[2026-06-15 20:18:52,682][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.20359091460704803, 'train/grad_norm_pre_clip': 8.75676441192627, 'timing/data': 0.009762733243405819, 'timing/model': 1.5719375908374786, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 0.15}) +[2026-06-15 20:18:59,057][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.2490762174129486, 'train/grad_norm_pre_clip': 8.589571952819824, 'timing/data': 0.011259851045906544, 'timing/model': 1.5987131912261248, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 0.15}) +[2026-06-15 20:19:05,400][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.3323753774166107, 'train/grad_norm_pre_clip': 7.9335103034973145, 'timing/data': 0.010876159183681011, 'timing/model': 1.7942790249362588, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.15}) +[2026-06-15 20:19:11,575][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.29795941710472107, 'train/grad_norm_pre_clip': 11.159764289855957, 'timing/data': 0.011649422813206911, 'timing/model': 1.5632190918549895, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 0.15}) +[2026-06-15 20:19:18,091][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.3734143376350403, 'train/grad_norm_pre_clip': 17.64027214050293, 'timing/data': 0.011362091172486544, 'timing/model': 1.5909318989142776, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 0.15}) +[2026-06-15 20:19:24,366][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.14988355338573456, 'train/grad_norm_pre_clip': 7.055160045623779, 'timing/data': 0.011451731901615858, 'timing/model': 1.5471047381870449, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 0.15}) +[2026-06-15 20:19:30,677][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.2586755156517029, 'train/grad_norm_pre_clip': 9.92253589630127, 'timing/data': 0.011193579994142056, 'timing/model': 1.5579812750220299, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 0.16}) +[2026-06-15 20:19:36,973][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.3720894157886505, 'train/grad_norm_pre_clip': 20.6626033782959, 'timing/data': 0.012294176034629345, 'timing/model': 1.530465183313936, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 0.16}) +[2026-06-15 20:19:43,400][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.1875862330198288, 'train/grad_norm_pre_clip': 13.530927658081055, 'timing/data': 0.01168367313221097, 'timing/model': 1.814094346947968, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 0.16}) +[2026-06-15 20:19:49,626][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.13957194983959198, 'train/grad_norm_pre_clip': 20.66412925720215, 'timing/data': 0.01093186903744936, 'timing/model': 1.6927859401330352, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 0.16}) +[2026-06-15 20:19:55,662][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.22915413975715637, 'train/grad_norm_pre_clip': 24.859294891357422, 'timing/data': 0.010955719277262688, 'timing/model': 1.5591709669679403, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 0.16}) +[2026-06-15 20:20:01,963][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.1362057477235794, 'train/grad_norm_pre_clip': 9.396805763244629, 'timing/data': 0.011676173191517591, 'timing/model': 1.5562111479230225, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 0.16}) +[2026-06-15 20:20:08,277][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.403764009475708, 'train/grad_norm_pre_clip': 14.433721542358398, 'timing/data': 0.011938754934817553, 'timing/model': 1.559672082774341, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.16}) +[2026-06-15 20:20:14,682][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.38146114349365234, 'train/grad_norm_pre_clip': 21.11574935913086, 'timing/data': 0.011946363840252161, 'timing/model': 1.563119508791715, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 0.16}) +[2026-06-15 20:20:21,004][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.15564817190170288, 'train/grad_norm_pre_clip': 11.260730743408203, 'timing/data': 0.010890199337154627, 'timing/model': 1.7897818349301815, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 0.16}) +[2026-06-15 20:20:27,318][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.20383517444133759, 'train/grad_norm_pre_clip': 14.600784301757812, 'timing/data': 0.010855668224394321, 'timing/model': 1.8134169289842248, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 0.16}) +[2026-06-15 20:20:33,472][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.29959842562675476, 'train/grad_norm_pre_clip': 13.201201438903809, 'timing/data': 0.010726436972618103, 'timing/model': 1.5487182629294693, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.16}) +[2026-06-15 20:20:39,770][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.2958206236362457, 'train/grad_norm_pre_clip': 14.170507431030273, 'timing/data': 0.01039596600458026, 'timing/model': 1.5608369642868638, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 0.16}) +[2026-06-15 20:20:46,243][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.33828431367874146, 'train/grad_norm_pre_clip': 12.363601684570312, 'timing/data': 0.011264260858297348, 'timing/model': 1.6658285045996308, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 0.16}) +[2026-06-15 20:20:52,572][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.2015060931444168, 'train/grad_norm_pre_clip': 12.587512016296387, 'timing/data': 0.010791318025439978, 'timing/model': 1.5316890222020447, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 0.16}) +[2026-06-15 20:20:59,110][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.20577381551265717, 'train/grad_norm_pre_clip': 11.278953552246094, 'timing/data': 0.01130241109058261, 'timing/model': 1.7900038952939212, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 0.16}) +[2026-06-15 20:21:05,167][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.32276055216789246, 'train/grad_norm_pre_clip': 14.104653358459473, 'timing/data': 0.011075369082391262, 'timing/model': 1.562012119218707, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 0.16}) +[2026-06-15 20:32:17,750][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.33431917428970337, 'train/grad_norm_pre_clip': 10.07381534576416, 'eval/loss': 0.15543502569198608, 'eval/flap_precision': 0.7950727883538634, 'eval/flap_recall': 0.9686221009549796, 'eval/flap_f1': 0.8733087330873308, 'eval/flap_rate_model': 0.24805555555555556, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.8733087330873308, 'timing/data': 0.010875748936086893, 'timing/model': 1.553712261840701, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.16}) +[2026-06-15 20:32:17,751][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_400_state +[2026-06-15 20:32:17,758][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:32:27,221][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_400_state/model.safetensors +[2026-06-15 20:32:47,889][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_400_state/optimizer.bin +[2026-06-15 20:32:47,890][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_400_state/sampler.bin +[2026-06-15 20:32:47,890][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_400_state/sampler_1.bin +[2026-06-15 20:32:47,892][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-15 20:32:48,083][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-15 20:32:48,083][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:32:48,100][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:32:54,126][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.4651520848274231, 'train/grad_norm_pre_clip': 9.597777366638184, 'timing/data': 0.011084458325058222, 'timing/model': 1.5629095640033484, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 0.16}) +[2026-06-15 20:33:00,441][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.24332937598228455, 'train/grad_norm_pre_clip': 5.728870868682861, 'timing/data': 0.01109665771946311, 'timing/model': 1.5653515658341348, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 0.16}) +[2026-06-15 20:33:06,642][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.22724111378192902, 'train/grad_norm_pre_clip': 5.723072052001953, 'timing/data': 0.010823387186974287, 'timing/model': 1.5589644121937454, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 0.16}) +[2026-06-15 20:33:12,842][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.21571457386016846, 'train/grad_norm_pre_clip': 10.577322006225586, 'timing/data': 0.010741916950792074, 'timing/model': 1.5241290568374097, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 0.16}) +[2026-06-15 20:33:19,051][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.3032039403915405, 'train/grad_norm_pre_clip': 10.650983810424805, 'timing/data': 0.011148269288241863, 'timing/model': 1.5030022449791431, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 0.16}) +[2026-06-15 20:33:25,352][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.29781389236450195, 'train/grad_norm_pre_clip': 13.53488826751709, 'timing/data': 0.010914096608757973, 'timing/model': 1.760094276163727, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 0.16}) +[2026-06-15 20:33:31,448][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.2003142237663269, 'train/grad_norm_pre_clip': 7.851293087005615, 'timing/data': 0.011510870885103941, 'timing/model': 1.5577931418083608, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 0.16}) +[2026-06-15 20:33:37,755][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.1918044239282608, 'train/grad_norm_pre_clip': 15.148148536682129, 'timing/data': 0.011543930042535067, 'timing/model': 1.563258521258831, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 0.16}) +[2026-06-15 20:33:44,056][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.17599409818649292, 'train/grad_norm_pre_clip': 8.842445373535156, 'timing/data': 0.011048257816582918, 'timing/model': 1.5246345871128142, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 0.17}) +[2026-06-15 20:33:50,460][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.17435401678085327, 'train/grad_norm_pre_clip': 6.483120441436768, 'timing/data': 0.010727276094257832, 'timing/model': 1.5925387530587614, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.17}) +[2026-06-15 20:33:56,757][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.09894412010908127, 'train/grad_norm_pre_clip': 5.19795560836792, 'timing/data': 0.011011388152837753, 'timing/model': 1.519356207922101, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 0.17}) +[2026-06-15 20:34:03,063][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.0213321540504694, 'train/grad_norm_pre_clip': 3.20157790184021, 'timing/data': 0.011113607790321112, 'timing/model': 1.7656092806719244, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 0.17}) +[2026-06-15 20:34:09,156][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.21251380443572998, 'train/grad_norm_pre_clip': 38.62518310546875, 'timing/data': 0.011153647676110268, 'timing/model': 1.5564279309473932, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 0.17}) +[2026-06-15 20:34:15,459][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.45355740189552307, 'train/grad_norm_pre_clip': 44.56239700317383, 'timing/data': 0.011001246981322765, 'timing/model': 1.5574351768009365, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 0.17}) +[2026-06-15 20:34:21,867][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.39516016840934753, 'train/grad_norm_pre_clip': 9.701945304870605, 'timing/data': 0.011345779057592154, 'timing/model': 1.576767096761614, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.17}) +[2026-06-15 20:34:28,169][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.3598414361476898, 'train/grad_norm_pre_clip': 40.37567901611328, 'timing/data': 0.011461290065199137, 'timing/model': 1.521009094081819, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 0.17}) +[2026-06-15 20:34:34,499][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.6165491938591003, 'train/grad_norm_pre_clip': 42.779136657714844, 'timing/data': 0.012465516105294228, 'timing/model': 1.798304668162018, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 0.17}) +[2026-06-15 20:34:40,673][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.3444470465183258, 'train/grad_norm_pre_clip': 32.15315246582031, 'timing/data': 0.013594882097095251, 'timing/model': 1.5550567209720612, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 0.17}) +[2026-06-15 20:34:46,965][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.17606444656848907, 'train/grad_norm_pre_clip': 7.972348690032959, 'timing/data': 0.011255679186433554, 'timing/model': 1.5559906749986112, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 0.17}) +[2026-06-15 20:34:53,171][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.5370590686798096, 'train/grad_norm_pre_clip': 35.05251693725586, 'timing/data': 0.01077812584117055, 'timing/model': 1.5694467658177018, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.17}) +[2026-06-15 20:34:59,465][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.3750077188014984, 'train/grad_norm_pre_clip': 23.420001983642578, 'timing/data': 0.010839737020432949, 'timing/model': 1.5562531962059438, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 0.17}) +[2026-06-15 20:35:05,760][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.5913316607475281, 'train/grad_norm_pre_clip': 43.46760940551758, 'timing/data': 0.0107354368083179, 'timing/model': 1.5664091380313039, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 0.17}) +[2026-06-15 20:35:12,129][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.17895616590976715, 'train/grad_norm_pre_clip': 11.686385154724121, 'timing/data': 0.011581820901483297, 'timing/model': 1.818348489701748, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 0.17}) +[2026-06-15 20:35:18,168][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.24677874147891998, 'train/grad_norm_pre_clip': 16.535776138305664, 'timing/data': 0.010845727287232876, 'timing/model': 1.5327883013524115, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 0.17}) +[2026-06-15 20:35:24,466][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.2871168553829193, 'train/grad_norm_pre_clip': 10.816056251525879, 'timing/data': 0.011729651130735874, 'timing/model': 1.5631641899235547, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 0.17}) +[2026-06-15 20:35:30,771][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.2636539936065674, 'train/grad_norm_pre_clip': 13.438063621520996, 'timing/data': 0.010841676965355873, 'timing/model': 1.563032909296453, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 0.17}) +[2026-06-15 20:35:37,065][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.2714183032512665, 'train/grad_norm_pre_clip': 16.222705841064453, 'timing/data': 0.011297260411083698, 'timing/model': 1.5629347069188952, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 0.17}) +[2026-06-15 20:35:43,365][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.5785394310951233, 'train/grad_norm_pre_clip': 16.482830047607422, 'timing/data': 0.010980158112943172, 'timing/model': 1.5607695560902357, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 0.17}) +[2026-06-15 20:35:49,803][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.16008661687374115, 'train/grad_norm_pre_clip': 5.286272048950195, 'timing/data': 0.011216088198125362, 'timing/model': 1.8060581320896745, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 0.17}) +[2026-06-15 20:35:55,972][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.16187052428722382, 'train/grad_norm_pre_clip': 2.7251129150390625, 'timing/data': 0.011061558034271002, 'timing/model': 1.586769601330161, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.17}) +[2026-06-15 20:36:02,263][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.4188269376754761, 'train/grad_norm_pre_clip': 10.754144668579102, 'timing/data': 0.011764102149754763, 'timing/model': 1.552895613014698, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 0.17}) +[2026-06-15 20:36:08,573][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.31416094303131104, 'train/grad_norm_pre_clip': 15.461104393005371, 'timing/data': 0.010344504844397306, 'timing/model': 1.5630206246860325, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 0.17}) +[2026-06-15 20:36:14,869][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.2555849254131317, 'train/grad_norm_pre_clip': 16.61342430114746, 'timing/data': 0.011545910965651274, 'timing/model': 1.5337189519777894, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 0.18}) +[2026-06-15 20:36:21,275][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.20120559632778168, 'train/grad_norm_pre_clip': 5.581292152404785, 'timing/data': 0.011318949051201344, 'timing/model': 1.5670960061252117, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 0.18}) +[2026-06-15 20:36:27,697][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.19410528242588043, 'train/grad_norm_pre_clip': 10.41557502746582, 'timing/data': 0.01061976607888937, 'timing/model': 1.8170445263385773, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 0.18}) +[2026-06-15 20:36:33,775][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.2112956941127777, 'train/grad_norm_pre_clip': 7.319127082824707, 'timing/data': 0.0120555330067873, 'timing/model': 1.56308189406991, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 0.18}) +[2026-06-15 20:36:40,069][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.1831662505865097, 'train/grad_norm_pre_clip': 6.5950775146484375, 'timing/data': 0.011084227822721004, 'timing/model': 1.551676413975656, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 0.18}) +[2026-06-15 20:36:46,373][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.34660962224006653, 'train/grad_norm_pre_clip': 6.209660053253174, 'timing/data': 0.0104286540299654, 'timing/model': 1.5577528160065413, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 0.18}) +[2026-06-15 20:36:52,678][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.5054789185523987, 'train/grad_norm_pre_clip': 11.831990242004395, 'timing/data': 0.010769666638225317, 'timing/model': 1.5631990525871515, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 0.18}) +[2026-06-15 20:36:58,979][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.33594533801078796, 'train/grad_norm_pre_clip': 12.478612899780273, 'timing/data': 0.010394965298473835, 'timing/model': 1.540584844071418, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.18}) +[2026-06-15 20:37:05,302][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.192531555891037, 'train/grad_norm_pre_clip': 10.210753440856934, 'timing/data': 0.010590964928269386, 'timing/model': 1.7905028141103685, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 0.18}) +[2026-06-15 20:37:11,366][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.262557715177536, 'train/grad_norm_pre_clip': 9.515753746032715, 'timing/data': 0.00993784237653017, 'timing/model': 1.555795962922275, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 0.18}) +[2026-06-15 20:37:17,672][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.2301822453737259, 'train/grad_norm_pre_clip': 7.797399997711182, 'timing/data': 0.01097150705754757, 'timing/model': 1.5695861051790416, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.18}) +[2026-06-15 20:37:23,974][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.3496032655239105, 'train/grad_norm_pre_clip': 12.953707695007324, 'timing/data': 0.011937092058360577, 'timing/model': 1.5658628758974373, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 0.18}) +[2026-06-15 20:37:30,277][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.07002130150794983, 'train/grad_norm_pre_clip': 2.824650526046753, 'timing/data': 0.013166688848286867, 'timing/model': 1.5588552979752421, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 0.18}) +[2026-06-15 20:37:36,683][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.32929739356040955, 'train/grad_norm_pre_clip': 16.540571212768555, 'timing/data': 0.010963617358356714, 'timing/model': 1.547877190168947, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 0.18}) +[2026-06-15 20:37:43,004][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.1845068484544754, 'train/grad_norm_pre_clip': 5.927417278289795, 'timing/data': 0.0108907469548285, 'timing/model': 1.793314044829458, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 0.18}) +[2026-06-15 20:37:49,153][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.15369902551174164, 'train/grad_norm_pre_clip': 6.526403427124023, 'timing/data': 0.012573066167533398, 'timing/model': 1.6435474413447082, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 0.18}) +[2026-06-15 20:37:55,472][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.04926278442144394, 'train/grad_norm_pre_clip': 11.55648422241211, 'timing/data': 0.01059540593996644, 'timing/model': 1.5429422832094133, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 0.18}) +[2026-06-15 20:38:01,856][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.1982232928276062, 'train/grad_norm_pre_clip': 6.815553188323975, 'timing/data': 0.010770667344331741, 'timing/model': 1.6019211909733713, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.18}) +[2026-06-15 20:38:08,170][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.03153374046087265, 'train/grad_norm_pre_clip': 1.5412662029266357, 'timing/data': 0.01089711720123887, 'timing/model': 1.559253877028823, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 0.18}) +[2026-06-15 20:38:14,470][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.39408743381500244, 'train/grad_norm_pre_clip': 15.763384819030762, 'timing/data': 0.010882426984608173, 'timing/model': 1.5282957353629172, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 0.18}) +[2026-06-15 20:38:20,802][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.37497878074645996, 'train/grad_norm_pre_clip': 13.627607345581055, 'timing/data': 0.010739366058260202, 'timing/model': 1.7826965670101345, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 0.18}) +[2026-06-15 20:38:26,882][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.035689905285835266, 'train/grad_norm_pre_clip': 5.452155113220215, 'timing/data': 0.010750147048383951, 'timing/model': 1.5719206230714917, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 0.18}) +[2026-06-15 20:38:33,171][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.34038904309272766, 'train/grad_norm_pre_clip': 18.457599639892578, 'timing/data': 0.011680600699037313, 'timing/model': 1.5620215213857591, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 0.18}) +[2026-06-15 20:38:39,469][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.5285115838050842, 'train/grad_norm_pre_clip': 26.06019401550293, 'timing/data': 0.010456024669110775, 'timing/model': 1.5643853931687772, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 0.18}) +[2026-06-15 20:38:45,773][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.2337356060743332, 'train/grad_norm_pre_clip': 9.619638442993164, 'timing/data': 0.01127520902082324, 'timing/model': 1.5682998928241432, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 0.18}) +[2026-06-15 20:38:52,164][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.2967106103897095, 'train/grad_norm_pre_clip': 25.840736389160156, 'timing/data': 0.011446440126746893, 'timing/model': 1.597468786407262, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 0.19}) +[2026-06-15 20:38:58,595][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.46216416358947754, 'train/grad_norm_pre_clip': 46.24840545654297, 'timing/data': 0.010646676179021597, 'timing/model': 1.8068952122703195, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 0.19}) +[2026-06-15 20:39:04,831][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.4890359044075012, 'train/grad_norm_pre_clip': 30.972763061523438, 'timing/data': 0.010859436821192503, 'timing/model': 1.6946827340871096, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.19}) +[2026-06-15 20:39:10,872][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.24135716259479523, 'train/grad_norm_pre_clip': 20.70304298400879, 'timing/data': 0.011824232060462236, 'timing/model': 1.5622879303991795, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 0.19}) +[2026-06-15 20:39:17,253][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.12945449352264404, 'train/grad_norm_pre_clip': 5.488702774047852, 'timing/data': 0.01083405688405037, 'timing/model': 1.6310089798644185, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 0.19}) +[2026-06-15 20:39:23,660][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.35577192902565, 'train/grad_norm_pre_clip': 29.921092987060547, 'timing/data': 0.010706366039812565, 'timing/model': 1.6028364817611873, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 0.19}) +[2026-06-15 20:39:29,973][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.09084401279687881, 'train/grad_norm_pre_clip': 10.92935848236084, 'timing/data': 0.010075573343783617, 'timing/model': 1.530754464212805, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 0.19}) +[2026-06-15 20:39:36,374][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.1916206032037735, 'train/grad_norm_pre_clip': 9.478029251098633, 'timing/data': 0.012093523051589727, 'timing/model': 1.8698059190064669, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 0.19}) +[2026-06-15 20:39:42,714][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.2052173912525177, 'train/grad_norm_pre_clip': 9.8665771484375, 'timing/data': 0.010892766993492842, 'timing/model': 1.8087593889795244, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 0.19}) +[2026-06-15 20:39:48,774][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.24522726237773895, 'train/grad_norm_pre_clip': 4.306815147399902, 'timing/data': 0.011031128000468016, 'timing/model': 1.5636112047359347, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 0.19}) +[2026-06-15 20:39:55,075][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.1377158910036087, 'train/grad_norm_pre_clip': 7.625694274902344, 'timing/data': 0.01155824027955532, 'timing/model': 1.5650715129449964, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 0.19}) +[2026-06-15 20:40:01,373][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.1281951665878296, 'train/grad_norm_pre_clip': 2.144287109375, 'timing/data': 0.010915666818618774, 'timing/model': 1.56805583788082, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 0.19}) +[2026-06-15 20:40:07,664][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.2979915142059326, 'train/grad_norm_pre_clip': 7.8410563468933105, 'timing/data': 0.01129250880330801, 'timing/model': 1.5265605701133609, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.19}) +[2026-06-15 20:40:14,032][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.27716317772865295, 'train/grad_norm_pre_clip': 8.718892097473145, 'timing/data': 0.010578874964267015, 'timing/model': 1.813080569729209, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 0.19}) +[2026-06-15 20:40:20,090][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.24570807814598083, 'train/grad_norm_pre_clip': 10.944127082824707, 'timing/data': 0.012529695872217417, 'timing/model': 1.5463843629695475, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 0.19}) +[2026-06-15 20:40:26,487][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.15478181838989258, 'train/grad_norm_pre_clip': 6.021087646484375, 'timing/data': 0.012818696908652782, 'timing/model': 1.5692816632799804, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 0.19}) +[2026-06-15 20:40:32,781][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.2858983278274536, 'train/grad_norm_pre_clip': 10.145599365234375, 'timing/data': 0.011172888800501823, 'timing/model': 1.5712901530787349, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 0.19}) +[2026-06-15 20:40:39,069][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.19042176008224487, 'train/grad_norm_pre_clip': 5.473996162414551, 'timing/data': 0.011659310664981604, 'timing/model': 1.5307947807013988, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 0.19}) +[2026-06-15 20:40:45,370][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.12987661361694336, 'train/grad_norm_pre_clip': 5.377434730529785, 'timing/data': 0.010579266119748354, 'timing/model': 1.5934989382512867, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 0.19}) +[2026-06-15 20:40:51,694][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.33704516291618347, 'train/grad_norm_pre_clip': 10.606221199035645, 'timing/data': 0.010989317670464516, 'timing/model': 1.7779544238001108, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 0.19}) +[2026-06-15 20:40:57,772][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.18906110525131226, 'train/grad_norm_pre_clip': 7.825955390930176, 'timing/data': 0.011740230955183506, 'timing/model': 1.558220773935318, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 0.19}) +[2026-06-15 20:41:04,071][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.22416777908802032, 'train/grad_norm_pre_clip': 8.946008682250977, 'timing/data': 0.010700985789299011, 'timing/model': 1.560559316072613, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 0.19}) +[2026-06-15 20:41:10,473][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.20803707838058472, 'train/grad_norm_pre_clip': 13.943532943725586, 'timing/data': 0.009868842083960772, 'timing/model': 1.5564894238486886, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.19}) +[2026-06-15 20:41:16,840][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.2794630825519562, 'train/grad_norm_pre_clip': 7.797804355621338, 'timing/data': 0.010352014098316431, 'timing/model': 1.5901243300177157, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 0.19}) +[2026-06-15 20:41:23,259][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.351931631565094, 'train/grad_norm_pre_clip': 18.33451271057129, 'timing/data': 0.010621936060488224, 'timing/model': 1.6007886449806392, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 0.19}) +[2026-06-15 20:41:29,567][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.3033839464187622, 'train/grad_norm_pre_clip': 12.780827522277832, 'timing/data': 0.010399153921753168, 'timing/model': 1.7537031257525086, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 0.2}) +[2026-06-15 20:41:35,671][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.2979665696620941, 'train/grad_norm_pre_clip': 6.462737560272217, 'timing/data': 0.010443213861435652, 'timing/model': 1.5625498848967254, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 0.2}) +[2026-06-15 20:41:41,991][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.2995590567588806, 'train/grad_norm_pre_clip': 12.81518268585205, 'timing/data': 0.012041773181408644, 'timing/model': 1.5796346641145647, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.2}) +[2026-06-15 20:41:48,273][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.17397697269916534, 'train/grad_norm_pre_clip': 7.352206230163574, 'timing/data': 0.012256973888725042, 'timing/model': 1.5608304161578417, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 0.2}) +[2026-06-15 20:41:54,655][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.3476264476776123, 'train/grad_norm_pre_clip': 9.710663795471191, 'timing/data': 0.010797556024044752, 'timing/model': 1.601301767397672, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 0.2}) +[2026-06-15 20:42:01,016][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.18522202968597412, 'train/grad_norm_pre_clip': 5.366089820861816, 'timing/data': 0.01201258273795247, 'timing/model': 1.8027210202999413, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 0.2}) +[2026-06-15 20:42:07,071][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.23638200759887695, 'train/grad_norm_pre_clip': 6.996262073516846, 'timing/data': 0.010697825811803341, 'timing/model': 1.5513165052980185, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 0.2}) +[2026-06-15 20:42:13,367][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.22914443910121918, 'train/grad_norm_pre_clip': 7.278073787689209, 'timing/data': 0.010260173119604588, 'timing/model': 1.5554816769436002, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 0.2}) +[2026-06-15 20:42:19,671][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.25016555190086365, 'train/grad_norm_pre_clip': 6.483002662658691, 'timing/data': 0.010653926059603691, 'timing/model': 1.5645552230998874, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 0.2}) +[2026-06-15 20:42:25,981][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.10802067071199417, 'train/grad_norm_pre_clip': 5.296145915985107, 'timing/data': 0.011480559129267931, 'timing/model': 1.5770930899307132, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.2}) +[2026-06-15 20:42:32,272][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.15908558666706085, 'train/grad_norm_pre_clip': 5.450555324554443, 'timing/data': 0.011446710675954819, 'timing/model': 1.5773061979562044, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 0.2}) +[2026-06-15 20:42:38,717][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.1276818811893463, 'train/grad_norm_pre_clip': 4.291927337646484, 'timing/data': 0.01084098406136036, 'timing/model': 1.7969869058579206, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 0.2}) +[2026-06-15 20:42:44,773][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.1808958649635315, 'train/grad_norm_pre_clip': 7.2274885177612305, 'timing/data': 0.0107286237180233, 'timing/model': 1.5608153170906007, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 0.2}) +[2026-06-15 20:42:51,076][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.2974582612514496, 'train/grad_norm_pre_clip': 12.91762924194336, 'timing/data': 0.011414505075663328, 'timing/model': 1.5653499946929514, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 0.2}) +[2026-06-15 20:42:57,369][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.3136069178581238, 'train/grad_norm_pre_clip': 17.44501304626465, 'timing/data': 0.012809616979211569, 'timing/model': 1.5610906691290438, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 0.2}) +[2026-06-15 20:43:03,767][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.291848361492157, 'train/grad_norm_pre_clip': 22.076683044433594, 'timing/data': 0.011013096198439598, 'timing/model': 1.5684308800846338, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 0.2}) +[2026-06-15 20:43:10,073][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.1035689264535904, 'train/grad_norm_pre_clip': 5.473184108734131, 'timing/data': 0.010988947004079819, 'timing/model': 1.5782918692566454, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 0.2}) +[2026-06-15 20:54:42,824][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.20731811225414276, 'train/grad_norm_pre_clip': 7.581323623657227, 'eval/loss': 0.12368486821651459, 'eval/flap_precision': 0.8431372549019608, 'eval/flap_recall': 0.9386084583901774, 'eval/flap_f1': 0.8883150419625565, 'eval/flap_rate_model': 0.22666666666666666, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.8883150419625564, 'timing/data': 0.010756426956504583, 'timing/model': 1.800300344824791, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 0.2}) +[2026-06-15 20:54:42,826][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_500_state +[2026-06-15 20:54:42,834][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:54:52,114][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_500_state/model.safetensors +[2026-06-15 20:55:12,395][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_500_state/optimizer.bin +[2026-06-15 20:55:12,396][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_500_state/sampler.bin +[2026-06-15 20:55:12,396][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_500_state/sampler_1.bin +[2026-06-15 20:55:12,397][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-15 20:55:12,588][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-15 20:55:12,588][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:55:12,604][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:55:19,064][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.038410816341638565, 'train/grad_norm_pre_clip': 5.101771354675293, 'timing/data': 0.010995618999004364, 'timing/model': 1.9822846837341785, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 0.2}) +[2026-06-15 20:55:25,438][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.19516152143478394, 'train/grad_norm_pre_clip': 8.325628280639648, 'timing/data': 0.011663220822811127, 'timing/model': 1.8637758763507009, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 0.2}) +[2026-06-15 20:55:31,441][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.38492998480796814, 'train/grad_norm_pre_clip': 14.866209030151367, 'timing/data': 0.010823318269103765, 'timing/model': 1.5627640569582582, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 0.2}) +[2026-06-15 20:55:37,816][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.15737780928611755, 'train/grad_norm_pre_clip': 6.343805313110352, 'timing/data': 0.011394450906664133, 'timing/model': 1.637494680006057, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 0.2}) +[2026-06-15 20:55:44,230][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.1870148926973343, 'train/grad_norm_pre_clip': 9.549076080322266, 'timing/data': 0.011281710118055344, 'timing/model': 1.6518933791667223, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 0.2}) +[2026-06-15 20:55:50,557][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.13717317581176758, 'train/grad_norm_pre_clip': 8.72807502746582, 'timing/data': 0.011819601990282536, 'timing/model': 1.5739028081297874, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 0.2}) +[2026-06-15 20:55:56,950][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.15918105840682983, 'train/grad_norm_pre_clip': 9.941123962402344, 'timing/data': 0.010797978844493628, 'timing/model': 1.861120027024299, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 0.21}) +[2026-06-15 20:56:03,347][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.15154586732387543, 'train/grad_norm_pre_clip': 19.263877868652344, 'timing/data': 0.011734792031347752, 'timing/model': 1.8538587810471654, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 0.21}) +[2026-06-15 20:56:09,454][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.35458576679229736, 'train/grad_norm_pre_clip': 10.062583923339844, 'timing/data': 0.01538635604083538, 'timing/model': 1.560728969052434, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 0.21}) +[2026-06-15 20:56:15,851][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.1706032007932663, 'train/grad_norm_pre_clip': 6.6599907875061035, 'timing/data': 0.010301677044481039, 'timing/model': 1.5537324300967157, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 0.21}) +[2026-06-15 20:56:22,253][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.27217525243759155, 'train/grad_norm_pre_clip': 9.602834701538086, 'timing/data': 0.010668998584151268, 'timing/model': 1.5538834361359477, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 0.21}) +[2026-06-15 20:56:28,655][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.286648690700531, 'train/grad_norm_pre_clip': 12.926090240478516, 'timing/data': 0.012274555396288633, 'timing/model': 1.5674378909170628, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 0.21}) +[2026-06-15 20:56:35,061][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.09261409938335419, 'train/grad_norm_pre_clip': 4.456685543060303, 'timing/data': 0.010796449612826109, 'timing/model': 1.5701869162730873, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 0.21}) +[2026-06-15 20:56:41,477][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.18828599154949188, 'train/grad_norm_pre_clip': 6.949080944061279, 'timing/data': 0.010936540085822344, 'timing/model': 1.8716430980712175, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 0.21}) +[2026-06-15 20:56:47,574][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.3441370129585266, 'train/grad_norm_pre_clip': 15.300395965576172, 'timing/data': 0.010844559874385595, 'timing/model': 1.557826641947031, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 0.21}) +[2026-06-15 20:56:53,971][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.07251689583063126, 'train/grad_norm_pre_clip': 3.9583206176757812, 'timing/data': 0.01166562270373106, 'timing/model': 1.5664521176367998, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 0.21}) +[2026-06-15 20:57:00,263][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.18089638650417328, 'train/grad_norm_pre_clip': 7.269599914550781, 'timing/data': 0.011212061624974012, 'timing/model': 1.55503955995664, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 0.21}) +[2026-06-15 20:57:06,703][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.14560295641422272, 'train/grad_norm_pre_clip': 5.533788681030273, 'timing/data': 0.011299772188067436, 'timing/model': 1.590392435900867, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 0.21}) +[2026-06-15 20:57:13,065][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.2635933756828308, 'train/grad_norm_pre_clip': 7.43313455581665, 'timing/data': 0.012283424846827984, 'timing/model': 1.5723766339942813, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 0.21}) +[2026-06-15 20:57:19,454][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.37432312965393066, 'train/grad_norm_pre_clip': 12.579192161560059, 'timing/data': 0.011983634904026985, 'timing/model': 1.8563052853569388, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.21}) +[2026-06-15 20:57:25,560][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.1581675261259079, 'train/grad_norm_pre_clip': 5.813189506530762, 'timing/data': 0.01115451194345951, 'timing/model': 1.5598941249772906, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 0.21}) +[2026-06-15 20:57:31,965][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.26130959391593933, 'train/grad_norm_pre_clip': 10.892813682556152, 'timing/data': 0.011516952887177467, 'timing/model': 1.5657687010243535, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 0.21}) +[2026-06-15 20:57:38,475][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.2122219353914261, 'train/grad_norm_pre_clip': 9.143832206726074, 'timing/data': 0.012228705920279026, 'timing/model': 1.5655258130282164, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 0.21}) +[2026-06-15 20:57:44,868][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.08736740052700043, 'train/grad_norm_pre_clip': 6.787041187286377, 'timing/data': 0.010760301258414984, 'timing/model': 1.5644514029845595, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 0.21}) +[2026-06-15 20:57:51,366][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.20044317841529846, 'train/grad_norm_pre_clip': 8.722886085510254, 'timing/data': 0.011077342089265585, 'timing/model': 1.5619703982956707, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 0.21}) +[2026-06-15 20:57:57,757][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.11392942070960999, 'train/grad_norm_pre_clip': 4.349327564239502, 'timing/data': 0.01139102317392826, 'timing/model': 1.8504425501450896, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 0.21}) +[2026-06-15 20:58:04,179][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.12107907980680466, 'train/grad_norm_pre_clip': 6.152967929840088, 'timing/data': 0.01123391231521964, 'timing/model': 1.8754159710370004, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 0.21}) +[2026-06-15 20:58:10,265][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.13729886710643768, 'train/grad_norm_pre_clip': 5.822900772094727, 'timing/data': 0.01201195502653718, 'timing/model': 1.5366915711201727, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 0.21}) +[2026-06-15 20:58:16,675][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.20613302290439606, 'train/grad_norm_pre_clip': 11.121109962463379, 'timing/data': 0.011159891728311777, 'timing/model': 1.562496670987457, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 0.21}) +[2026-06-15 20:58:23,072][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.14886689186096191, 'train/grad_norm_pre_clip': 11.512112617492676, 'timing/data': 0.011601625010371208, 'timing/model': 1.5668303100392222, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 0.21}) +[2026-06-15 20:58:29,469][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.13594944775104523, 'train/grad_norm_pre_clip': 9.241219520568848, 'timing/data': 0.010543250944465399, 'timing/model': 1.5616439622826874, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 0.21}) +[2026-06-15 20:58:35,868][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.3720113933086395, 'train/grad_norm_pre_clip': 18.993366241455078, 'timing/data': 0.011553524062037468, 'timing/model': 1.8633159291930497, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 0.22}) +[2026-06-15 20:58:41,981][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.24241743981838226, 'train/grad_norm_pre_clip': 5.597600936889648, 'timing/data': 0.01099381223320961, 'timing/model': 1.573224621359259, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 0.22}) +[2026-06-15 20:58:48,366][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.07225155830383301, 'train/grad_norm_pre_clip': 11.55500602722168, 'timing/data': 0.011372713837772608, 'timing/model': 1.5586586771532893, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 0.22}) +[2026-06-15 20:58:54,771][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.08480042219161987, 'train/grad_norm_pre_clip': 5.476909160614014, 'timing/data': 0.01202292600646615, 'timing/model': 1.5566548220813274, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 0.22}) +[2026-06-15 20:59:01,173][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.33402225375175476, 'train/grad_norm_pre_clip': 24.882654190063477, 'timing/data': 0.01109612314030528, 'timing/model': 1.5681221866980195, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 0.22}) +[2026-06-15 20:59:07,468][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.40774276852607727, 'train/grad_norm_pre_clip': 21.607337951660156, 'timing/data': 0.01098216325044632, 'timing/model': 1.536024343688041, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 0.22}) +[2026-06-15 20:59:13,991][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.12629827857017517, 'train/grad_norm_pre_clip': 15.373969078063965, 'timing/data': 0.010186928790062666, 'timing/model': 1.8806713595986366, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 0.22}) +[2026-06-15 20:59:20,068][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.38779667019844055, 'train/grad_norm_pre_clip': 20.719074249267578, 'timing/data': 0.01101642195135355, 'timing/model': 1.5591193661093712, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 0.22}) +[2026-06-15 20:59:26,521][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.15281875431537628, 'train/grad_norm_pre_clip': 8.733449935913086, 'timing/data': 0.011038603261113167, 'timing/model': 1.6139943799935281, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 0.22}) +[2026-06-15 20:59:32,879][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.1469825655221939, 'train/grad_norm_pre_clip': 13.195781707763672, 'timing/data': 0.011154164094477892, 'timing/model': 1.5684565636329353, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.22}) +[2026-06-15 20:59:39,280][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.1019861027598381, 'train/grad_norm_pre_clip': 6.955463409423828, 'timing/data': 0.010821502190083265, 'timing/model': 1.5808394933119416, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 0.22}) +[2026-06-15 20:59:45,775][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.17501625418663025, 'train/grad_norm_pre_clip': 8.060094833374023, 'timing/data': 0.01043333113193512, 'timing/model': 1.566638727672398, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 0.22}) +[2026-06-15 20:59:52,229][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.09100323915481567, 'train/grad_norm_pre_clip': 5.321778297424316, 'timing/data': 0.01476972782984376, 'timing/model': 1.815203120931983, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 0.22}) +[2026-06-15 20:59:58,376][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.37366411089897156, 'train/grad_norm_pre_clip': 35.16638946533203, 'timing/data': 0.011094002984464169, 'timing/model': 1.5685954466462135, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 0.22}) +[2026-06-15 21:00:04,774][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.1654355674982071, 'train/grad_norm_pre_clip': 8.213338851928711, 'timing/data': 0.011605495121330023, 'timing/model': 1.5542925619520247, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 0.22}) +[2026-06-15 21:00:11,270][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.1273668110370636, 'train/grad_norm_pre_clip': 7.856249809265137, 'timing/data': 0.011004653293639421, 'timing/model': 1.553984060883522, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 0.22}) +[2026-06-15 21:00:17,672][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.2964816093444824, 'train/grad_norm_pre_clip': 15.46847152709961, 'timing/data': 0.011542504653334618, 'timing/model': 1.5617959420196712, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 0.22}) +[2026-06-15 21:00:24,171][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.5421096682548523, 'train/grad_norm_pre_clip': 13.993934631347656, 'timing/data': 0.011523565743118525, 'timing/model': 1.8628833089023829, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 0.22}) +[2026-06-15 21:00:30,265][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.13615646958351135, 'train/grad_norm_pre_clip': 10.578755378723145, 'timing/data': 0.011421034578233957, 'timing/model': 1.5394516931846738, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 0.22}) +[2026-06-15 21:00:36,666][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.1566963642835617, 'train/grad_norm_pre_clip': 8.627612113952637, 'timing/data': 0.011228064075112343, 'timing/model': 1.5607721162959933, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 0.22}) +[2026-06-15 21:00:42,966][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.2557394802570343, 'train/grad_norm_pre_clip': 6.802248001098633, 'timing/data': 0.01112714409828186, 'timing/model': 1.5386080080643296, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 0.22}) +[2026-06-15 21:00:49,267][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.21534501016139984, 'train/grad_norm_pre_clip': 9.899909019470215, 'timing/data': 0.010583322029560804, 'timing/model': 1.5643643997609615, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 0.22}) +[2026-06-15 21:00:55,575][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.2586132884025574, 'train/grad_norm_pre_clip': 19.734786987304688, 'timing/data': 0.011354364920407534, 'timing/model': 1.5255297338590026, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 0.22}) +[2026-06-15 21:01:02,076][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.21402305364608765, 'train/grad_norm_pre_clip': 5.265172481536865, 'timing/data': 0.01185322692617774, 'timing/model': 1.860498696565628, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.22}) +[2026-06-15 21:01:08,183][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.1826004832983017, 'train/grad_norm_pre_clip': 6.562077522277832, 'timing/data': 0.011389095336198807, 'timing/model': 1.5701853004284203, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 0.22}) +[2026-06-15 21:01:14,703][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.22707322239875793, 'train/grad_norm_pre_clip': 6.318752288818359, 'timing/data': 0.011534615885466337, 'timing/model': 1.6183716719970107, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 0.23}) +[2026-06-15 21:01:21,078][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.25632166862487793, 'train/grad_norm_pre_clip': 7.1234846115112305, 'timing/data': 0.011346925050020218, 'timing/model': 1.5624050670303404, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 0.23}) +[2026-06-15 21:01:27,471][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.2286628633737564, 'train/grad_norm_pre_clip': 6.325508117675781, 'timing/data': 0.01039343187585473, 'timing/model': 1.5636519799008965, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 0.23}) +[2026-06-15 21:01:33,853][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.17511038482189178, 'train/grad_norm_pre_clip': 14.302558898925781, 'timing/data': 0.010602052323520184, 'timing/model': 1.5924831051379442, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 0.23}) +[2026-06-15 21:01:40,256][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.10997170209884644, 'train/grad_norm_pre_clip': 3.966662645339966, 'timing/data': 0.011038303840905428, 'timing/model': 1.8368247635662556, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 0.23}) +[2026-06-15 21:01:46,371][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.11086341738700867, 'train/grad_norm_pre_clip': 5.013555526733398, 'timing/data': 0.010971253737807274, 'timing/model': 1.5617734589613974, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 0.23}) +[2026-06-15 21:01:52,771][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.17628797888755798, 'train/grad_norm_pre_clip': 7.145223140716553, 'timing/data': 0.0112265651114285, 'timing/model': 1.5600893599912524, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 0.23}) +[2026-06-15 21:01:59,198][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.14504152536392212, 'train/grad_norm_pre_clip': 9.193717002868652, 'timing/data': 0.010286471340805292, 'timing/model': 1.577627929393202, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 0.23}) +[2026-06-15 21:02:05,687][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.1873752325773239, 'train/grad_norm_pre_clip': 21.878734588623047, 'timing/data': 0.011672397144138813, 'timing/model': 1.5876218769699335, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 0.23}) +[2026-06-15 21:02:12,176][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.14176073670387268, 'train/grad_norm_pre_clip': 4.4208664894104, 'timing/data': 0.011281386017799377, 'timing/model': 1.5965906218625605, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 0.23}) +[2026-06-15 21:02:18,565][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.1726398468017578, 'train/grad_norm_pre_clip': 19.924266815185547, 'timing/data': 0.011444216128438711, 'timing/model': 1.8557525086216629, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 0.23}) +[2026-06-15 21:02:24,674][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.3071323335170746, 'train/grad_norm_pre_clip': 29.943212509155273, 'timing/data': 0.011152575258165598, 'timing/model': 1.5570724080316722, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 0.23}) +[2026-06-15 21:02:31,069][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.2001182585954666, 'train/grad_norm_pre_clip': 33.55896759033203, 'timing/data': 0.010862113907933235, 'timing/model': 1.5585504621267319, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 0.23}) +[2026-06-15 21:02:37,471][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.1587788611650467, 'train/grad_norm_pre_clip': 22.1256160736084, 'timing/data': 0.010782863944768906, 'timing/model': 1.5554482559673488, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 0.23}) +[2026-06-15 21:02:43,869][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.27049168944358826, 'train/grad_norm_pre_clip': 33.67122268676758, 'timing/data': 0.011133925057947636, 'timing/model': 1.5671452707611024, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 0.23}) +[2026-06-15 21:02:50,273][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.3310789465904236, 'train/grad_norm_pre_clip': 18.83008575439453, 'timing/data': 0.011503627058118582, 'timing/model': 1.568766145966947, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 0.23}) +[2026-06-15 21:02:56,662][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.6495214104652405, 'train/grad_norm_pre_clip': 71.41985321044922, 'timing/data': 0.01110491482540965, 'timing/model': 1.8526021502912045, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 0.23}) +[2026-06-15 21:03:02,872][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.22779235243797302, 'train/grad_norm_pre_clip': 8.81419563293457, 'timing/data': 0.010912994854152203, 'timing/model': 1.5586936976760626, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 0.23}) +[2026-06-15 21:03:09,273][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.35992589592933655, 'train/grad_norm_pre_clip': 17.534006118774414, 'timing/data': 0.010896035004407167, 'timing/model': 1.5598986209370196, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 0.23}) +[2026-06-15 21:03:15,573][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.33467569947242737, 'train/grad_norm_pre_clip': 21.08406639099121, 'timing/data': 0.012404709588736296, 'timing/model': 1.5599218076094985, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 0.23}) +[2026-06-15 21:03:21,977][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.19245053827762604, 'train/grad_norm_pre_clip': 12.241167068481445, 'timing/data': 0.011537767015397549, 'timing/model': 1.5742052039131522, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 0.23}) +[2026-06-15 21:03:28,374][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.26621371507644653, 'train/grad_norm_pre_clip': 10.74232292175293, 'timing/data': 0.010960345156490803, 'timing/model': 1.5729560959152877, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 0.23}) +[2026-06-15 21:03:34,822][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.1937868744134903, 'train/grad_norm_pre_clip': 4.487586498260498, 'timing/data': 0.01073068380355835, 'timing/model': 1.9097532536834478, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 0.23}) +[2026-06-15 21:03:40,885][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.2577979862689972, 'train/grad_norm_pre_clip': 15.840397834777832, 'timing/data': 0.011095334775745869, 'timing/model': 1.5706364098004997, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 0.23}) +[2026-06-15 21:03:47,384][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.45861318707466125, 'train/grad_norm_pre_clip': 11.801836967468262, 'timing/data': 0.013366254977881908, 'timing/model': 1.5746321720071137, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 0.23}) +[2026-06-15 21:03:53,770][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.17929066717624664, 'train/grad_norm_pre_clip': 9.524457931518555, 'timing/data': 0.010705963708460331, 'timing/model': 1.5554429301992059, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 0.24}) +[2026-06-15 21:04:00,173][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.1863829791545868, 'train/grad_norm_pre_clip': 7.5958452224731445, 'timing/data': 0.010887774173170328, 'timing/model': 1.55764497583732, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 0.24}) +[2026-06-15 21:04:06,576][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.14695407450199127, 'train/grad_norm_pre_clip': 4.339397430419922, 'timing/data': 0.01127581624314189, 'timing/model': 1.5693332110531628, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 0.24}) +[2026-06-15 21:04:13,151][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.1465824991464615, 'train/grad_norm_pre_clip': 12.043683052062988, 'timing/data': 0.010931595228612423, 'timing/model': 1.8625058056786656, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 0.24}) +[2026-06-15 21:04:19,273][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.631223201751709, 'train/grad_norm_pre_clip': 18.47689437866211, 'timing/data': 0.010603504255414009, 'timing/model': 1.5677871871739626, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 0.24}) +[2026-06-15 21:04:25,688][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.20800350606441498, 'train/grad_norm_pre_clip': 12.109273910522461, 'timing/data': 0.014601689763367176, 'timing/model': 1.5540967863053083, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 0.24}) +[2026-06-15 21:04:32,077][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.10444632917642593, 'train/grad_norm_pre_clip': 7.658205986022949, 'timing/data': 0.011668208986520767, 'timing/model': 1.563348620198667, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 0.24}) +[2026-06-15 21:04:38,696][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.09284330904483795, 'train/grad_norm_pre_clip': 6.584941387176514, 'timing/data': 0.011128205806016922, 'timing/model': 1.5752745559439063, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 0.24}) +[2026-06-15 21:04:45,189][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.16966612637043, 'train/grad_norm_pre_clip': 10.929134368896484, 'timing/data': 0.012074040248990059, 'timing/model': 1.579262258950621, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.24}) +[2026-06-15 21:04:51,624][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.18309061229228973, 'train/grad_norm_pre_clip': 16.916595458984375, 'timing/data': 0.011132576037198305, 'timing/model': 1.8921978529542685, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 0.24}) +[2026-06-15 21:04:57,916][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.6357198357582092, 'train/grad_norm_pre_clip': 22.356306076049805, 'timing/data': 0.0116076678968966, 'timing/model': 1.8013506531715393, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 0.24}) +[2026-06-15 21:05:03,970][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.5286079049110413, 'train/grad_norm_pre_clip': 27.29199981689453, 'timing/data': 0.011770328972488642, 'timing/model': 1.5539145502261817, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 0.24}) +[2026-06-15 21:05:10,371][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.9509003162384033, 'train/grad_norm_pre_clip': 23.296419143676758, 'timing/data': 0.01135083707049489, 'timing/model': 1.5586489760316908, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 0.24}) +[2026-06-15 21:05:16,776][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.31350502371788025, 'train/grad_norm_pre_clip': 19.50403594970703, 'timing/data': 0.011154336854815483, 'timing/model': 1.564505116082728, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 0.24}) +[2026-06-15 21:05:23,170][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.3172459900379181, 'train/grad_norm_pre_clip': 15.72802448272705, 'timing/data': 0.011374437250196934, 'timing/model': 1.5367009257897735, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 0.24}) +[2026-06-15 21:05:29,555][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.23287881910800934, 'train/grad_norm_pre_clip': 15.333162307739258, 'timing/data': 0.011281637009233236, 'timing/model': 1.835568297188729, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 0.24}) +[2026-06-15 21:05:35,966][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.27955153584480286, 'train/grad_norm_pre_clip': 8.79758358001709, 'timing/data': 0.010058882180601358, 'timing/model': 1.8599631949327886, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 0.24}) +[2026-06-15 21:05:42,074][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.41687414050102234, 'train/grad_norm_pre_clip': 21.45736312866211, 'timing/data': 0.0112798772752285, 'timing/model': 1.5644487882964313, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 0.24}) +[2026-06-15 21:16:56,232][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.2805851995944977, 'train/grad_norm_pre_clip': 11.123994827270508, 'eval/loss': 0.18048244714736938, 'eval/flap_precision': 0.7933701657458564, 'eval/flap_recall': 0.9795361527967258, 'eval/flap_f1': 0.8766788766788767, 'eval/flap_rate_model': 0.2513888888888889, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.8766788766788766, 'timing/data': 0.01178437890484929, 'timing/model': 1.5746121550910175, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 0.24}) +[2026-06-15 21:16:56,233][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_600_state +[2026-06-15 21:16:56,241][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 21:17:05,457][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_600_state/model.safetensors +[2026-06-15 21:17:25,470][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_600_state/optimizer.bin +[2026-06-15 21:17:25,471][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_600_state/sampler.bin +[2026-06-15 21:17:25,471][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_600_state/sampler_1.bin +[2026-06-15 21:17:25,473][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_600_state/random_states_0.pkl +[2026-06-15 21:17:25,662][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600 +[2026-06-15 21:17:25,663][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 21:17:25,679][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 21:17:32,139][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.15833422541618347, 'train/grad_norm_pre_clip': 6.081760883331299, 'timing/data': 0.011421565897762775, 'timing/model': 1.5304288710467517, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 0.24}) +[2026-06-15 21:17:38,443][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.09308165311813354, 'train/grad_norm_pre_clip': 2.150315046310425, 'timing/data': 0.011234506033360958, 'timing/model': 1.5713819428347051, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 0.24}) +[2026-06-15 21:17:44,809][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.2499140501022339, 'train/grad_norm_pre_clip': 15.963448524475098, 'timing/data': 0.011036986019462347, 'timing/model': 1.612962496932596, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 0.24}) +[2026-06-15 21:17:51,050][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.1645495593547821, 'train/grad_norm_pre_clip': 7.214656829833984, 'timing/data': 0.011038894765079021, 'timing/model': 1.5447346395812929, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 0.24}) +[2026-06-15 21:17:57,376][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.18781210482120514, 'train/grad_norm_pre_clip': 10.964564323425293, 'timing/data': 0.012802581302821636, 'timing/model': 1.7842314722947776, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 0.24}) +[2026-06-15 21:18:03,445][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.1904333084821701, 'train/grad_norm_pre_clip': 7.6086745262146, 'timing/data': 0.011407706886529922, 'timing/model': 1.5368704539723694, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 0.25}) +[2026-06-15 21:18:09,863][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.16721521317958832, 'train/grad_norm_pre_clip': 9.829426765441895, 'timing/data': 0.011463567148894072, 'timing/model': 1.5754297599196434, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 0.25}) +[2026-06-15 21:18:16,254][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.4321562349796295, 'train/grad_norm_pre_clip': 24.39643669128418, 'timing/data': 0.012498191092163324, 'timing/model': 1.5727309510111809, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 0.25}) +[2026-06-15 21:18:22,643][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.3326758146286011, 'train/grad_norm_pre_clip': 17.489233016967773, 'timing/data': 0.01117466576397419, 'timing/model': 1.5846805311739445, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 0.25}) +[2026-06-15 21:18:29,020][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.29880088567733765, 'train/grad_norm_pre_clip': 13.97242546081543, 'timing/data': 0.010805235244333744, 'timing/model': 1.8008992108516395, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 0.25}) +[2026-06-15 21:18:35,155][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.2655293643474579, 'train/grad_norm_pre_clip': 10.247212409973145, 'timing/data': 0.012899612076580524, 'timing/model': 1.5606232299469411, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 0.25}) +[2026-06-15 21:18:41,452][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.33582064509391785, 'train/grad_norm_pre_clip': 16.012365341186523, 'timing/data': 0.010733004659414291, 'timing/model': 1.5622772951610386, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 0.25}) +[2026-06-15 21:18:47,821][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.049266014248132706, 'train/grad_norm_pre_clip': 3.4244327545166016, 'timing/data': 0.011356287635862827, 'timing/model': 1.6255493643693626, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 0.25}) +[2026-06-15 21:18:54,060][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.2268543839454651, 'train/grad_norm_pre_clip': 16.24216651916504, 'timing/data': 0.010196623858064413, 'timing/model': 1.564502899069339, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 0.25}) +[2026-06-15 21:19:00,358][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.3378520607948303, 'train/grad_norm_pre_clip': 10.596640586853027, 'timing/data': 0.012015790212899446, 'timing/model': 1.534372937399894, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 0.25}) +[2026-06-15 21:19:06,744][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.13172155618667603, 'train/grad_norm_pre_clip': 4.489809513092041, 'timing/data': 0.010839506052434444, 'timing/model': 1.8400412718765438, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 0.25}) +[2026-06-15 21:19:12,875][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.19769088923931122, 'train/grad_norm_pre_clip': 7.992684841156006, 'timing/data': 0.01082569733262062, 'timing/model': 1.578039698768407, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 0.25}) +[2026-06-15 21:19:19,265][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.19111518561840057, 'train/grad_norm_pre_clip': 11.094285011291504, 'timing/data': 0.011317268945276737, 'timing/model': 1.5577783579938114, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 0.25}) +[2026-06-15 21:19:25,669][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.21939991414546967, 'train/grad_norm_pre_clip': 7.016668796539307, 'timing/data': 0.011231138370931149, 'timing/model': 1.5702397488057613, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 0.25}) +[2026-06-15 21:19:32,068][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.2697335481643677, 'train/grad_norm_pre_clip': 8.447959899902344, 'timing/data': 0.010837657377123833, 'timing/model': 1.5599338011816144, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 0.25}) +[2026-06-15 21:19:38,372][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.07768944650888443, 'train/grad_norm_pre_clip': 3.8703439235687256, 'timing/data': 0.010856376960873604, 'timing/model': 1.5356913753785193, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 0.25}) +[2026-06-15 21:19:44,741][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.1857730597257614, 'train/grad_norm_pre_clip': 6.552753448486328, 'timing/data': 0.01097568729892373, 'timing/model': 1.832377398852259, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 0.25}) +[2026-06-15 21:19:50,869][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.10176157206296921, 'train/grad_norm_pre_clip': 6.644115924835205, 'timing/data': 0.01069457596167922, 'timing/model': 1.5406066249124706, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 0.25}) +[2026-06-15 21:19:57,267][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.06895191222429276, 'train/grad_norm_pre_clip': 3.477322816848755, 'timing/data': 0.010003774892538786, 'timing/model': 1.5475917751900852, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 0.25}) +[2026-06-15 21:20:03,665][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.24120047688484192, 'train/grad_norm_pre_clip': 18.545997619628906, 'timing/data': 0.011032207868993282, 'timing/model': 1.5398536450229585, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.25}) +[2026-06-15 21:20:10,066][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.30527788400650024, 'train/grad_norm_pre_clip': 7.974703311920166, 'timing/data': 0.011070917826145887, 'timing/model': 1.5870042340829968, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 0.25}) +[2026-06-15 21:20:16,457][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.26928001642227173, 'train/grad_norm_pre_clip': 8.014948844909668, 'timing/data': 0.010904837865382433, 'timing/model': 1.603997000027448, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 0.25}) +[2026-06-15 21:20:22,841][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.40891218185424805, 'train/grad_norm_pre_clip': 19.927692413330078, 'timing/data': 0.011306548956781626, 'timing/model': 1.8349857116118073, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 0.25}) +[2026-06-15 21:20:28,868][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.093219093978405, 'train/grad_norm_pre_clip': 5.232244968414307, 'timing/data': 0.011041558347642422, 'timing/model': 1.562238465063274, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 0.25}) +[2026-06-15 21:20:35,262][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.31793418526649475, 'train/grad_norm_pre_clip': 12.80164909362793, 'timing/data': 0.01086110807955265, 'timing/model': 1.5529784578830004, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 0.25}) +[2026-06-15 21:20:41,773][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.2736523151397705, 'train/grad_norm_pre_clip': 8.573055267333984, 'timing/data': 0.012828215025365353, 'timing/model': 1.5826924480497837, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 0.26}) +[2026-06-15 21:20:48,084][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.24582558870315552, 'train/grad_norm_pre_clip': 5.279865264892578, 'timing/data': 0.011360160075128078, 'timing/model': 1.5775435836985707, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 0.26}) +[2026-06-15 21:20:54,491][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.30088651180267334, 'train/grad_norm_pre_clip': 16.526784896850586, 'timing/data': 0.011055558919906616, 'timing/model': 1.60696913395077, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 0.26}) +[2026-06-15 21:21:00,835][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.09110170602798462, 'train/grad_norm_pre_clip': 8.905773162841797, 'timing/data': 0.01132691977545619, 'timing/model': 1.8291191067546606, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 0.26}) +[2026-06-15 21:21:06,971][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.21088552474975586, 'train/grad_norm_pre_clip': 7.060052394866943, 'timing/data': 0.011591891292482615, 'timing/model': 1.5994855370372534, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 0.26}) +[2026-06-15 21:21:13,374][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.40964701771736145, 'train/grad_norm_pre_clip': 10.789334297180176, 'timing/data': 0.011297180317342281, 'timing/model': 1.5654860599897802, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 0.26}) +[2026-06-15 21:21:19,667][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.2345391809940338, 'train/grad_norm_pre_clip': 11.559496879577637, 'timing/data': 0.011367970146238804, 'timing/model': 1.556280312128365, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 0.26}) +[2026-06-15 21:21:26,073][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.2070530503988266, 'train/grad_norm_pre_clip': 8.75725269317627, 'timing/data': 0.011563180945813656, 'timing/model': 1.5686611998826265, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 0.26}) +[2026-06-15 21:21:32,464][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.2906447947025299, 'train/grad_norm_pre_clip': 7.654318809509277, 'timing/data': 0.010707518085837364, 'timing/model': 1.5768276629969478, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.26}) +[2026-06-15 21:21:38,844][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.24259069561958313, 'train/grad_norm_pre_clip': 12.045388221740723, 'timing/data': 0.011288529727607965, 'timing/model': 1.8385620089247823, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 0.26}) +[2026-06-15 21:21:44,976][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.18819290399551392, 'train/grad_norm_pre_clip': 10.672791481018066, 'timing/data': 0.01178722269833088, 'timing/model': 1.568914509844035, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 0.26}) +[2026-06-15 21:21:51,390][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.19030596315860748, 'train/grad_norm_pre_clip': 8.9805269241333, 'timing/data': 0.01085444912314415, 'timing/model': 1.573222259990871, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 0.26}) +[2026-06-15 21:21:57,868][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.5010385513305664, 'train/grad_norm_pre_clip': 21.69155502319336, 'timing/data': 0.011128069832921028, 'timing/model': 1.5508739622309804, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 0.26}) +[2026-06-15 21:22:04,286][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.352542519569397, 'train/grad_norm_pre_clip': 14.918663024902344, 'timing/data': 0.013746339827775955, 'timing/model': 1.5937373507767916, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 0.26}) +[2026-06-15 21:22:10,675][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.28673604130744934, 'train/grad_norm_pre_clip': 10.521138191223145, 'timing/data': 0.011329541448503733, 'timing/model': 1.5757105587981641, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 0.26}) +[2026-06-15 21:22:17,210][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.25191134214401245, 'train/grad_norm_pre_clip': 7.516295909881592, 'timing/data': 0.013384508900344372, 'timing/model': 1.9007452270016074, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 0.26}) +[2026-06-15 21:22:23,372][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.17131397128105164, 'train/grad_norm_pre_clip': 8.630047798156738, 'timing/data': 0.01104131992906332, 'timing/model': 1.5564953340217471, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 0.26}) +[2026-06-15 21:22:29,788][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.12293346226215363, 'train/grad_norm_pre_clip': 2.9292118549346924, 'timing/data': 0.011150610633194447, 'timing/model': 1.5779850538820028, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 0.26}) +[2026-06-15 21:22:36,166][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.16778896749019623, 'train/grad_norm_pre_clip': 6.045338153839111, 'timing/data': 0.012962737120687962, 'timing/model': 1.5540939788334072, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 0.26}) +[2026-06-15 21:22:42,573][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.1760004460811615, 'train/grad_norm_pre_clip': 8.137094497680664, 'timing/data': 0.010390767827630043, 'timing/model': 1.573348032310605, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 0.26}) +[2026-06-15 21:22:48,973][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.08188189566135406, 'train/grad_norm_pre_clip': 2.7973763942718506, 'timing/data': 0.011417021974921227, 'timing/model': 1.5934757078066468, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 0.26}) +[2026-06-15 21:22:55,428][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.15979492664337158, 'train/grad_norm_pre_clip': 11.120807647705078, 'timing/data': 0.01026415778324008, 'timing/model': 1.8270887737162411, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 0.26}) +[2026-06-15 21:23:01,657][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.1968793272972107, 'train/grad_norm_pre_clip': 9.74416732788086, 'timing/data': 0.010188946966081858, 'timing/model': 1.7534317839890718, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 0.26}) +[2026-06-15 21:23:07,881][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.14747202396392822, 'train/grad_norm_pre_clip': 10.399432182312012, 'timing/data': 0.011499293148517609, 'timing/model': 1.5732550481334329, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 0.26}) +[2026-06-15 21:23:14,273][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.26557299494743347, 'train/grad_norm_pre_clip': 15.338370323181152, 'timing/data': 0.010871950071305037, 'timing/model': 1.5652582198381424, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 0.26}) +[2026-06-15 21:23:20,740][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.3076860010623932, 'train/grad_norm_pre_clip': 18.703699111938477, 'timing/data': 0.01079915976151824, 'timing/model': 1.6294974512420595, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 0.27}) +[2026-06-15 21:23:27,172][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.07204309850931168, 'train/grad_norm_pre_clip': 5.9761643409729, 'timing/data': 0.010796660091727972, 'timing/model': 1.57984000723809, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 0.27}) +[2026-06-15 21:23:33,736][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.5124277472496033, 'train/grad_norm_pre_clip': 21.118505477905273, 'timing/data': 0.010347228962928057, 'timing/model': 1.8321788311004639, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 0.27}) +[2026-06-15 21:23:40,039][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.20150986313819885, 'train/grad_norm_pre_clip': 23.138940811157227, 'timing/data': 0.010218458250164986, 'timing/model': 1.7963425698690116, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 0.27}) +[2026-06-15 21:23:46,366][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.13916057348251343, 'train/grad_norm_pre_clip': 10.779300689697266, 'timing/data': 0.011584233958274126, 'timing/model': 1.6769760530442, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.27}) +[2026-06-15 21:23:52,824][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.16523894667625427, 'train/grad_norm_pre_clip': 16.133657455444336, 'timing/data': 0.011368672829121351, 'timing/model': 1.611801279708743, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 0.27}) +[2026-06-15 21:23:59,171][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.12069986760616302, 'train/grad_norm_pre_clip': 7.76392126083374, 'timing/data': 0.011719834059476852, 'timing/model': 1.562259895261377, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 0.27}) +[2026-06-15 21:24:05,570][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.1540839672088623, 'train/grad_norm_pre_clip': 12.359746932983398, 'timing/data': 0.010588659904897213, 'timing/model': 1.5685147577896714, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 0.27}) +[2026-06-15 21:24:12,046][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.2034296691417694, 'train/grad_norm_pre_clip': 13.874366760253906, 'timing/data': 0.012411307077854872, 'timing/model': 1.828675331082195, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 0.27}) +[2026-06-15 21:24:18,172][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.13273705542087555, 'train/grad_norm_pre_clip': 9.65212345123291, 'timing/data': 0.010858420748263597, 'timing/model': 1.5629168469458818, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 0.27}) +[2026-06-15 21:24:24,640][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.148678258061409, 'train/grad_norm_pre_clip': 8.513097763061523, 'timing/data': 0.010668640956282616, 'timing/model': 1.639349825680256, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 0.27}) +[2026-06-15 21:24:30,971][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.14271169900894165, 'train/grad_norm_pre_clip': 10.418298721313477, 'timing/data': 0.01079613110050559, 'timing/model': 1.567660114262253, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 0.27}) +[2026-06-15 21:24:37,468][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.16326533257961273, 'train/grad_norm_pre_clip': 12.031451225280762, 'timing/data': 0.01071111112833023, 'timing/model': 1.5938793239183724, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 0.27}) +[2026-06-15 21:24:43,836][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.13285985589027405, 'train/grad_norm_pre_clip': 9.617403030395508, 'timing/data': 0.010338339023292065, 'timing/model': 1.6286710957065225, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 0.27}) +[2026-06-15 21:24:50,184][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.11895202100276947, 'train/grad_norm_pre_clip': 9.337915420532227, 'timing/data': 0.010680751409381628, 'timing/model': 1.765528118237853, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 0.27}) +[2026-06-15 21:24:56,287][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.16485469043254852, 'train/grad_norm_pre_clip': 16.3162841796875, 'timing/data': 0.010710700880736113, 'timing/model': 1.5655325837433338, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 0.27}) +[2026-06-15 21:25:02,764][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.09665530920028687, 'train/grad_norm_pre_clip': 6.13588285446167, 'timing/data': 0.01238575717434287, 'timing/model': 1.66068111686036, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 0.27}) +[2026-06-15 21:25:09,174][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.20436006784439087, 'train/grad_norm_pre_clip': 16.746971130371094, 'timing/data': 0.0108985910192132, 'timing/model': 1.5667702169157565, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 0.27}) +[2026-06-15 21:25:15,568][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.29054465889930725, 'train/grad_norm_pre_clip': 13.608927726745605, 'timing/data': 0.01178171532228589, 'timing/model': 1.5971556319855154, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 0.27}) +[2026-06-15 21:25:21,965][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.34476590156555176, 'train/grad_norm_pre_clip': 8.996365547180176, 'timing/data': 0.010946952272206545, 'timing/model': 1.583340437617153, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 0.27}) +[2026-06-15 21:25:28,383][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.28603702783584595, 'train/grad_norm_pre_clip': 12.238319396972656, 'timing/data': 0.010890972334891558, 'timing/model': 1.8713425770401955, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 0.27}) +[2026-06-15 21:25:34,477][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.3405405580997467, 'train/grad_norm_pre_clip': 13.384289741516113, 'timing/data': 0.010801531840115786, 'timing/model': 1.5699780439026654, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 0.27}) +[2026-06-15 21:25:40,878][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.3268108665943146, 'train/grad_norm_pre_clip': 8.234711647033691, 'timing/data': 0.011478935368359089, 'timing/model': 1.565209184307605, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 0.27}) +[2026-06-15 21:25:47,366][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.26603078842163086, 'train/grad_norm_pre_clip': 11.294861793518066, 'timing/data': 0.011551055125892162, 'timing/model': 1.5741878980770707, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 0.27}) +[2026-06-15 21:25:53,774][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.21825329959392548, 'train/grad_norm_pre_clip': 6.902653217315674, 'timing/data': 0.012109746690839529, 'timing/model': 1.5707422723062336, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 0.27}) +[2026-06-15 21:26:00,172][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.2818544805049896, 'train/grad_norm_pre_clip': 5.835442543029785, 'timing/data': 0.011141032911837101, 'timing/model': 1.8603997039608657, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 0.28}) +[2026-06-15 21:26:06,279][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.2251165211200714, 'train/grad_norm_pre_clip': 11.173006057739258, 'timing/data': 0.010941722895950079, 'timing/model': 1.5450153076089919, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 0.28}) +[2026-06-15 21:26:12,677][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.19937832653522491, 'train/grad_norm_pre_clip': 6.569746494293213, 'timing/data': 0.011235434096306562, 'timing/model': 1.5478611271828413, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 0.28}) +[2026-06-15 21:26:19,083][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.2111932635307312, 'train/grad_norm_pre_clip': 11.076421737670898, 'timing/data': 0.011547484900802374, 'timing/model': 1.5485273380763829, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 0.28}) +[2026-06-15 21:26:25,483][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.2581622302532196, 'train/grad_norm_pre_clip': 6.431325435638428, 'timing/data': 0.010905992239713669, 'timing/model': 1.568005362059921, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 0.28}) +[2026-06-15 21:26:31,902][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.3063104450702667, 'train/grad_norm_pre_clip': 7.030819416046143, 'timing/data': 0.010773011948913336, 'timing/model': 1.6106962361373007, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 0.28}) +[2026-06-15 21:26:38,317][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.14402759075164795, 'train/grad_norm_pre_clip': 5.932766914367676, 'timing/data': 0.010888021904975176, 'timing/model': 1.895958109293133, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 0.28}) +[2026-06-15 21:26:44,501][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.16985002160072327, 'train/grad_norm_pre_clip': 7.406490325927734, 'timing/data': 0.01098271319642663, 'timing/model': 1.5607409360818565, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.28}) +[2026-06-15 21:26:51,002][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.2706044316291809, 'train/grad_norm_pre_clip': 5.590999126434326, 'timing/data': 0.011230743955820799, 'timing/model': 1.5869535561650991, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 0.28}) +[2026-06-15 21:26:57,579][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.19557565450668335, 'train/grad_norm_pre_clip': 8.062631607055664, 'timing/data': 0.011385663878172636, 'timing/model': 1.5664229542016983, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 0.28}) +[2026-06-15 21:27:03,972][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.2065798044204712, 'train/grad_norm_pre_clip': 11.693791389465332, 'timing/data': 0.011257113888859749, 'timing/model': 1.5664886613376439, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 0.28}) +[2026-06-15 21:27:10,276][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.4075612723827362, 'train/grad_norm_pre_clip': 11.832063674926758, 'timing/data': 0.011065373197197914, 'timing/model': 1.5342082921415567, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 0.28}) +[2026-06-15 21:27:16,663][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.375912606716156, 'train/grad_norm_pre_clip': 20.616212844848633, 'timing/data': 0.01106034405529499, 'timing/model': 1.8195642200298607, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 0.28}) +[2026-06-15 21:27:22,773][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.14967785775661469, 'train/grad_norm_pre_clip': 5.5998148918151855, 'timing/data': 0.011375184170901775, 'timing/model': 1.548938212916255, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 0.28}) +[2026-06-15 21:27:29,281][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.6220327615737915, 'train/grad_norm_pre_clip': 14.243627548217773, 'timing/data': 0.010969383176416159, 'timing/model': 1.5713532771915197, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.28}) +[2026-06-15 21:27:35,768][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.7724582552909851, 'train/grad_norm_pre_clip': 33.38308334350586, 'timing/data': 0.011281623970717192, 'timing/model': 1.5619494682177901, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 0.28}) +[2026-06-15 21:27:42,169][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.16502951085567474, 'train/grad_norm_pre_clip': 9.586735725402832, 'timing/data': 0.010925212875008583, 'timing/model': 1.5836728997528553, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 0.28}) +[2026-06-15 21:27:48,571][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.17193330824375153, 'train/grad_norm_pre_clip': 12.144633293151855, 'timing/data': 0.011968067847192287, 'timing/model': 1.5826127622276545, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 0.28}) +[2026-06-15 21:27:54,956][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.21916985511779785, 'train/grad_norm_pre_clip': 8.609063148498535, 'timing/data': 0.011868697125464678, 'timing/model': 1.8414903241209686, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 0.28}) +[2026-06-15 21:39:14,726][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.3697880804538727, 'train/grad_norm_pre_clip': 9.011520385742188, 'eval/loss': 0.10412628203630447, 'eval/flap_precision': 0.8584183673469388, 'eval/flap_recall': 0.9181446111869032, 'eval/flap_f1': 0.8872775214238628, 'eval/flap_rate_model': 0.21777777777777776, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.8872775214238628, 'timing/data': 0.01088384399190545, 'timing/model': 1.5613672607578337, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 0.28}) +[2026-06-15 21:39:14,727][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_700_state +[2026-06-15 21:39:14,734][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 21:39:23,883][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_700_state/model.safetensors +[2026-06-15 21:39:44,029][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_700_state/optimizer.bin +[2026-06-15 21:39:44,030][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_700_state/sampler.bin +[2026-06-15 21:39:44,030][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_700_state/sampler_1.bin +[2026-06-15 21:39:44,032][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_700_state/random_states_0.pkl +[2026-06-15 21:39:44,228][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700 +[2026-06-15 21:39:44,229][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 21:39:44,244][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 21:39:50,486][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.23240818083286285, 'train/grad_norm_pre_clip': 11.833636283874512, 'timing/data': 0.010077702812850475, 'timing/model': 1.8334592822939157, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 0.28}) +[2026-06-15 21:39:56,515][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.38727065920829773, 'train/grad_norm_pre_clip': 16.459455490112305, 'timing/data': 0.010567655321210623, 'timing/model': 1.554593469016254, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 0.28}) +[2026-06-15 21:40:02,721][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.12323852628469467, 'train/grad_norm_pre_clip': 8.896486282348633, 'timing/data': 0.01058690482750535, 'timing/model': 1.5571921621449292, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 0.28}) +[2026-06-15 21:40:08,943][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.21734581887722015, 'train/grad_norm_pre_clip': 6.872783660888672, 'timing/data': 0.012375883292406797, 'timing/model': 1.5718507850542665, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 0.28}) +[2026-06-15 21:40:15,037][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.291780561208725, 'train/grad_norm_pre_clip': 15.177531242370605, 'timing/data': 0.010639594867825508, 'timing/model': 1.5628234781324863, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 0.29}) +[2026-06-15 21:40:21,243][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.24946226179599762, 'train/grad_norm_pre_clip': 6.288971900939941, 'timing/data': 0.011013107374310493, 'timing/model': 1.527559211011976, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 0.29}) +[2026-06-15 21:40:27,589][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.2571392357349396, 'train/grad_norm_pre_clip': 11.340024948120117, 'timing/data': 0.01233630208298564, 'timing/model': 1.794822407886386, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 0.29}) +[2026-06-15 21:40:33,659][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.08253127336502075, 'train/grad_norm_pre_clip': 2.978001356124878, 'timing/data': 0.010776936076581478, 'timing/model': 1.5776330553926528, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 0.29}) +[2026-06-15 21:40:39,951][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.23878398537635803, 'train/grad_norm_pre_clip': 30.899749755859375, 'timing/data': 0.010442183818668127, 'timing/model': 1.5396662871353328, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 0.29}) +[2026-06-15 21:40:46,238][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.19255200028419495, 'train/grad_norm_pre_clip': 6.032233238220215, 'timing/data': 0.011091357097029686, 'timing/model': 1.6482450156472623, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 0.29}) +[2026-06-15 21:40:52,451][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.18973273038864136, 'train/grad_norm_pre_clip': 9.430808067321777, 'timing/data': 0.011107375845313072, 'timing/model': 1.5562927299179137, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 0.29}) +[2026-06-15 21:40:58,755][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.6253916025161743, 'train/grad_norm_pre_clip': 37.879764556884766, 'timing/data': 0.011550738010555506, 'timing/model': 1.5168538051657379, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 0.29}) +[2026-06-15 21:41:05,094][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.1465662568807602, 'train/grad_norm_pre_clip': 20.197317123413086, 'timing/data': 0.011687078978866339, 'timing/model': 1.8005237141624093, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 0.29}) +[2026-06-15 21:41:11,220][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.15639840066432953, 'train/grad_norm_pre_clip': 9.252217292785645, 'timing/data': 0.010879025794565678, 'timing/model': 1.6222682921215892, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 0.29}) +[2026-06-15 21:41:17,459][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.3647846579551697, 'train/grad_norm_pre_clip': 17.41866683959961, 'timing/data': 0.010583013761788607, 'timing/model': 1.5536079639568925, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 0.29}) +[2026-06-15 21:41:23,765][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.18178462982177734, 'train/grad_norm_pre_clip': 8.998796463012695, 'timing/data': 0.01187152974307537, 'timing/model': 1.570024753920734, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 0.29}) +[2026-06-15 21:41:30,193][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.21571588516235352, 'train/grad_norm_pre_clip': 11.554496765136719, 'timing/data': 0.011354058049619198, 'timing/model': 1.582801969256252, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 0.29}) +[2026-06-15 21:41:36,455][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.1551441252231598, 'train/grad_norm_pre_clip': 9.654964447021484, 'timing/data': 0.011522678192704916, 'timing/model': 1.5123229129239917, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 0.29}) +[2026-06-15 21:41:42,809][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.21442648768424988, 'train/grad_norm_pre_clip': 7.206221580505371, 'timing/data': 0.010424593929201365, 'timing/model': 1.8160085277631879, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 0.29}) +[2026-06-15 21:41:49,119][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.10330234467983246, 'train/grad_norm_pre_clip': 5.48013162612915, 'timing/data': 0.010604534763842821, 'timing/model': 1.8146122237667441, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 0.29}) +[2026-06-15 21:41:55,356][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.10948891192674637, 'train/grad_norm_pre_clip': 5.988481044769287, 'timing/data': 0.010815116111189127, 'timing/model': 1.5550780179910362, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 0.29}) +[2026-06-15 21:42:01,666][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.21894823014736176, 'train/grad_norm_pre_clip': 12.204498291015625, 'timing/data': 0.01059441501274705, 'timing/model': 1.5632123830728233, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 0.29}) +[2026-06-15 21:42:07,954][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.5012716054916382, 'train/grad_norm_pre_clip': 26.04953384399414, 'timing/data': 0.010350573807954788, 'timing/model': 1.562198560219258, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 0.29}) +[2026-06-15 21:42:14,257][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.11778690665960312, 'train/grad_norm_pre_clip': 8.374324798583984, 'timing/data': 0.010196063201874495, 'timing/model': 1.5183060360141098, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 0.29}) +[2026-06-15 21:42:20,693][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.06907951086759567, 'train/grad_norm_pre_clip': 4.242565631866455, 'timing/data': 0.010609405115246773, 'timing/model': 1.791198041755706, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 0.29}) +[2026-06-15 21:42:26,756][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.2591644525527954, 'train/grad_norm_pre_clip': 8.824640274047852, 'timing/data': 0.010569405276328325, 'timing/model': 1.563794040121138, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 0.29}) +[2026-06-15 21:42:33,058][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.320470929145813, 'train/grad_norm_pre_clip': 13.617132186889648, 'timing/data': 0.010196913033723831, 'timing/model': 1.5544154420495033, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 0.29}) +[2026-06-15 21:42:39,365][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.25931516289711, 'train/grad_norm_pre_clip': 12.73659610748291, 'timing/data': 0.010683675296604633, 'timing/model': 1.5685831629671156, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 0.29}) +[2026-06-15 21:42:45,658][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.20200741291046143, 'train/grad_norm_pre_clip': 11.286787033081055, 'timing/data': 0.010820135939866304, 'timing/model': 1.516655073966831, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 0.29}) +[2026-06-15 21:42:51,966][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.35673969984054565, 'train/grad_norm_pre_clip': 14.268837928771973, 'timing/data': 0.010454434435814619, 'timing/model': 1.5937372222542763, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.3}) +[2026-06-15 21:42:58,248][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.3039570450782776, 'train/grad_norm_pre_clip': 9.194708824157715, 'timing/data': 0.010344724170863628, 'timing/model': 1.745786237064749, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 0.3}) +[2026-06-15 21:43:04,464][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.501456618309021, 'train/grad_norm_pre_clip': 11.284187316894531, 'timing/data': 0.010437414050102234, 'timing/model': 1.561712509021163, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 0.3}) +[2026-06-15 21:43:10,766][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.1697542816400528, 'train/grad_norm_pre_clip': 7.326183319091797, 'timing/data': 0.010323324240744114, 'timing/model': 1.5592675092630088, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 0.3}) +[2026-06-15 21:43:17,058][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.1319204419851303, 'train/grad_norm_pre_clip': 7.171712398529053, 'timing/data': 0.010064451955258846, 'timing/model': 1.562421043869108, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 0.3}) +[2026-06-15 21:43:23,360][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.15525487065315247, 'train/grad_norm_pre_clip': 5.927177906036377, 'timing/data': 0.010190403088927269, 'timing/model': 1.529337463900447, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 0.3}) +[2026-06-15 21:43:29,729][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.2808977961540222, 'train/grad_norm_pre_clip': 13.129117965698242, 'timing/data': 0.010141723323613405, 'timing/model': 1.5775575898587704, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 0.3}) +[2026-06-15 21:43:35,966][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.19186431169509888, 'train/grad_norm_pre_clip': 6.5318284034729, 'timing/data': 0.010370643809437752, 'timing/model': 1.7579711847938597, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.3}) +[2026-06-15 21:43:42,060][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.15485352277755737, 'train/grad_norm_pre_clip': 4.095268726348877, 'timing/data': 0.010394063778221607, 'timing/model': 1.5592139340005815, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 0.3}) +[2026-06-15 21:43:48,360][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.1504426896572113, 'train/grad_norm_pre_clip': 6.2169671058654785, 'timing/data': 0.010255583096295595, 'timing/model': 1.5523160961456597, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 0.3}) +[2026-06-15 21:43:54,773][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.22504925727844238, 'train/grad_norm_pre_clip': 7.728254318237305, 'timing/data': 0.010073032695800066, 'timing/model': 1.5539324940182269, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 0.3}) +[2026-06-15 21:44:01,167][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.2637927234172821, 'train/grad_norm_pre_clip': 10.355512619018555, 'timing/data': 0.011097606737166643, 'timing/model': 1.6272865058854222, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 0.3}) +[2026-06-15 21:44:07,515][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.36477649211883545, 'train/grad_norm_pre_clip': 9.492899894714355, 'timing/data': 0.01102342689409852, 'timing/model': 1.8018415281549096, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 0.3}) +[2026-06-15 21:44:13,574][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.1923423558473587, 'train/grad_norm_pre_clip': 7.208767890930176, 'timing/data': 0.010862336028367281, 'timing/model': 1.565627175848931, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 0.3}) +[2026-06-15 21:44:19,871][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.14487096667289734, 'train/grad_norm_pre_clip': 9.021248817443848, 'timing/data': 0.011203307192772627, 'timing/model': 1.5627119750715792, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 0.3}) +[2026-06-15 21:44:26,242][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.0711037889122963, 'train/grad_norm_pre_clip': 2.9499149322509766, 'timing/data': 0.01011633314192295, 'timing/model': 1.6429725759662688, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 0.3}) +[2026-06-15 21:44:32,489][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.3030818700790405, 'train/grad_norm_pre_clip': 13.56414794921875, 'timing/data': 0.010608035139739513, 'timing/model': 1.57389828376472, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 0.3}) +[2026-06-15 21:44:38,776][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.18150100111961365, 'train/grad_norm_pre_clip': 9.113585472106934, 'timing/data': 0.010997646022588015, 'timing/model': 1.5834531760774553, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 0.3}) +[2026-06-15 21:44:45,116][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.08629675209522247, 'train/grad_norm_pre_clip': 9.362204551696777, 'timing/data': 0.011204357724636793, 'timing/model': 1.777578120585531, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 0.3}) +[2026-06-15 21:44:51,289][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.07326607406139374, 'train/grad_norm_pre_clip': 4.884758949279785, 'timing/data': 0.012087331153452396, 'timing/model': 1.5836838781833649, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 0.3}) +[2026-06-15 21:44:57,580][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.2226630300283432, 'train/grad_norm_pre_clip': 7.398292541503906, 'timing/data': 0.010870975907891989, 'timing/model': 1.5712338760495186, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 0.3}) +[2026-06-15 21:45:03,874][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.2076534926891327, 'train/grad_norm_pre_clip': 6.687291622161865, 'timing/data': 0.01085779583081603, 'timing/model': 1.5689581572078168, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 0.3}) +[2026-06-15 21:45:10,269][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.10715743154287338, 'train/grad_norm_pre_clip': 12.081270217895508, 'timing/data': 0.011400138027966022, 'timing/model': 1.5757459071464837, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 0.3}) +[2026-06-15 21:45:16,563][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.23004436492919922, 'train/grad_norm_pre_clip': 11.271757125854492, 'timing/data': 0.009877642150968313, 'timing/model': 1.552426238078624, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 0.3}) +[2026-06-15 21:45:22,906][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.09312497079372406, 'train/grad_norm_pre_clip': 7.670360088348389, 'timing/data': 0.010764146223664284, 'timing/model': 1.8021822306327522, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 0.3}) +[2026-06-15 21:45:28,976][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.3941950500011444, 'train/grad_norm_pre_clip': 15.863517761230469, 'timing/data': 0.011358247604221106, 'timing/model': 1.5710841799154878, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 0.31}) +[2026-06-15 21:45:35,277][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.09814220666885376, 'train/grad_norm_pre_clip': 4.591113090515137, 'timing/data': 0.011088226921856403, 'timing/model': 1.564423281699419, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 0.31}) +[2026-06-15 21:45:41,567][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.37567636370658875, 'train/grad_norm_pre_clip': 31.62746238708496, 'timing/data': 0.010395335033535957, 'timing/model': 1.55537795368582, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 0.31}) +[2026-06-15 21:45:47,967][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.14203844964504242, 'train/grad_norm_pre_clip': 11.822840690612793, 'timing/data': 0.011370047926902771, 'timing/model': 1.5271482048556209, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 0.31}) +[2026-06-15 21:45:54,268][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.13518604636192322, 'train/grad_norm_pre_clip': 9.387463569641113, 'timing/data': 0.010737895965576172, 'timing/model': 1.5528858648613095, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 0.31}) +[2026-06-15 21:46:00,611][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.2546026408672333, 'train/grad_norm_pre_clip': 10.035565376281738, 'timing/data': 0.010623555164784193, 'timing/model': 1.800373659003526, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 0.31}) +[2026-06-15 21:46:06,672][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.2052215188741684, 'train/grad_norm_pre_clip': 14.919638633728027, 'timing/data': 0.01123840780928731, 'timing/model': 1.5635068118572235, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 0.31}) +[2026-06-15 21:46:13,056][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.286523699760437, 'train/grad_norm_pre_clip': 20.04043197631836, 'timing/data': 0.011087587103247643, 'timing/model': 1.6501631517894566, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 0.31}) +[2026-06-15 21:46:19,428][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.13233430683612823, 'train/grad_norm_pre_clip': 10.452743530273438, 'timing/data': 0.011659649666398764, 'timing/model': 1.6202528849244118, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 0.31}) +[2026-06-15 21:46:25,674][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.10877320915460587, 'train/grad_norm_pre_clip': 7.203466892242432, 'timing/data': 0.010519314091652632, 'timing/model': 1.572148172184825, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 0.31}) +[2026-06-15 21:46:31,969][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.03668544813990593, 'train/grad_norm_pre_clip': 2.3561854362487793, 'timing/data': 0.011155208107084036, 'timing/model': 1.523468644823879, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.31}) +[2026-06-15 21:46:38,304][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.37057268619537354, 'train/grad_norm_pre_clip': 13.749855041503906, 'timing/data': 0.0132356658577919, 'timing/model': 1.7929627932608128, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 0.31}) +[2026-06-15 21:46:44,369][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.2786814868450165, 'train/grad_norm_pre_clip': 12.54049015045166, 'timing/data': 0.011281938292086124, 'timing/model': 1.5585644859820604, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 0.31}) +[2026-06-15 21:46:50,670][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.10457828640937805, 'train/grad_norm_pre_clip': 5.6603779792785645, 'timing/data': 0.010788326151669025, 'timing/model': 1.5614447090774775, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 0.31}) +[2026-06-15 21:46:56,984][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.15047046542167664, 'train/grad_norm_pre_clip': 8.056485176086426, 'timing/data': 0.011697540059685707, 'timing/model': 1.5658141570165753, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 0.31}) +[2026-06-15 21:47:03,266][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.11543220281600952, 'train/grad_norm_pre_clip': 8.516885757446289, 'timing/data': 0.010872756130993366, 'timing/model': 1.5551083427853882, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 0.31}) +[2026-06-15 21:47:09,611][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.16474269330501556, 'train/grad_norm_pre_clip': 7.043406009674072, 'timing/data': 0.010774155613034964, 'timing/model': 1.5685336412861943, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 0.31}) +[2026-06-15 21:47:15,900][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.1812247931957245, 'train/grad_norm_pre_clip': 11.316750526428223, 'timing/data': 0.011351339053362608, 'timing/model': 1.7728470307774842, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 0.31}) +[2026-06-15 21:47:21,970][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.3542028069496155, 'train/grad_norm_pre_clip': 10.924883842468262, 'timing/data': 0.011343228165060282, 'timing/model': 1.5710665229707956, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 0.31}) +[2026-06-15 21:47:28,268][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.1262415051460266, 'train/grad_norm_pre_clip': 8.071198463439941, 'timing/data': 0.010668846312910318, 'timing/model': 1.5652812886983156, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 0.31}) +[2026-06-15 21:47:34,571][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.161092609167099, 'train/grad_norm_pre_clip': 7.813799858093262, 'timing/data': 0.010556395165622234, 'timing/model': 1.5566652133129537, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 0.31}) +[2026-06-15 21:47:40,870][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.1513073444366455, 'train/grad_norm_pre_clip': 8.558869361877441, 'timing/data': 0.010608625132590532, 'timing/model': 1.5532457493245602, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 0.31}) +[2026-06-15 21:47:47,168][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.18664011359214783, 'train/grad_norm_pre_clip': 13.083944320678711, 'timing/data': 0.010930617339909077, 'timing/model': 1.5221757870167494, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 0.31}) +[2026-06-15 21:47:53,501][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.2915356159210205, 'train/grad_norm_pre_clip': 23.409744262695312, 'timing/data': 0.010905197355896235, 'timing/model': 1.7915814658626914, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 0.31}) +[2026-06-15 21:47:59,568][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.10310452431440353, 'train/grad_norm_pre_clip': 4.380577087402344, 'timing/data': 0.01165155041962862, 'timing/model': 1.5595972673036158, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 0.32}) +[2026-06-15 21:48:05,868][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.16238924860954285, 'train/grad_norm_pre_clip': 8.530400276184082, 'timing/data': 0.010931706987321377, 'timing/model': 1.5595976882614195, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 0.32}) +[2026-06-15 21:48:12,169][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.19534924626350403, 'train/grad_norm_pre_clip': 8.862419128417969, 'timing/data': 0.010766486171633005, 'timing/model': 1.561633056961, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 0.32}) +[2026-06-15 21:48:18,579][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.35335779190063477, 'train/grad_norm_pre_clip': 6.980277061462402, 'timing/data': 0.011486709117889404, 'timing/model': 1.584792897105217, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 0.32}) +[2026-06-15 21:48:24,905][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.368511438369751, 'train/grad_norm_pre_clip': 25.033531188964844, 'timing/data': 0.011145498137921095, 'timing/model': 1.555876384023577, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 0.32}) +[2026-06-15 21:48:31,206][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.09985508769750595, 'train/grad_norm_pre_clip': 5.546326637268066, 'timing/data': 0.011265277862548828, 'timing/model': 1.797486244700849, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 0.32}) +[2026-06-15 21:48:37,431][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.18554934859275818, 'train/grad_norm_pre_clip': 10.963261604309082, 'timing/data': 0.011744610033929348, 'timing/model': 1.7202830761671066, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 0.32}) +[2026-06-15 21:48:43,470][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.26419544219970703, 'train/grad_norm_pre_clip': 15.187273025512695, 'timing/data': 0.011804310139268637, 'timing/model': 1.5623327237553895, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.32}) +[2026-06-15 21:48:49,779][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.24814864993095398, 'train/grad_norm_pre_clip': 9.572510719299316, 'timing/data': 0.01093931682407856, 'timing/model': 1.564195541664958, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 0.32}) +[2026-06-15 21:48:56,177][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.16088412702083588, 'train/grad_norm_pre_clip': 8.340859413146973, 'timing/data': 0.011160137131810188, 'timing/model': 1.5633449689485133, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 0.32}) +[2026-06-15 21:49:02,471][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.18395306169986725, 'train/grad_norm_pre_clip': 12.962331771850586, 'timing/data': 0.010901765897870064, 'timing/model': 1.5245970343239605, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 0.32}) +[2026-06-15 21:49:08,804][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.2587748169898987, 'train/grad_norm_pre_clip': 7.0758442878723145, 'timing/data': 0.011438877787441015, 'timing/model': 1.7923342171125114, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 0.32}) +[2026-06-15 21:49:15,122][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.10417797416448593, 'train/grad_norm_pre_clip': 3.5604937076568604, 'timing/data': 0.011253977660089731, 'timing/model': 1.812711725011468, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 0.32}) +[2026-06-15 21:49:21,360][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.2364961802959442, 'train/grad_norm_pre_clip': 7.415164947509766, 'timing/data': 0.011620739940553904, 'timing/model': 1.5553403571248055, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 0.32}) +[2026-06-15 21:49:27,673][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.1734209954738617, 'train/grad_norm_pre_clip': 7.215564250946045, 'timing/data': 0.01076393574476242, 'timing/model': 1.5666260062716901, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 0.32}) +[2026-06-15 21:49:33,988][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.1914713978767395, 'train/grad_norm_pre_clip': 9.087687492370605, 'timing/data': 0.01189807103946805, 'timing/model': 1.5677773510105908, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 0.32}) +[2026-06-15 21:49:40,324][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.2565367817878723, 'train/grad_norm_pre_clip': 10.686193466186523, 'timing/data': 0.011354419402778149, 'timing/model': 1.5718688890337944, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 0.32}) +[2026-06-15 21:49:46,614][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.1624482274055481, 'train/grad_norm_pre_clip': 9.075521469116211, 'timing/data': 0.011784070171415806, 'timing/model': 1.809799744747579, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 0.32}) +[2026-06-15 21:49:52,671][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.19924402236938477, 'train/grad_norm_pre_clip': 7.208019733428955, 'timing/data': 0.011121456976979971, 'timing/model': 1.562158638611436, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 0.32}) +[2026-06-15 21:49:58,973][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.24082130193710327, 'train/grad_norm_pre_clip': 7.328659534454346, 'timing/data': 0.011175797786563635, 'timing/model': 1.5725050228647888, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 0.32}) +[2026-06-15 21:50:05,364][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.10865116864442825, 'train/grad_norm_pre_clip': 6.382639408111572, 'timing/data': 0.01084889704361558, 'timing/model': 1.5625961106270552, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 0.32}) +[2026-06-15 22:01:23,919][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.04836394265294075, 'train/grad_norm_pre_clip': 3.6523778438568115, 'eval/loss': 0.09448894113302231, 'eval/flap_precision': 0.9347181008902077, 'eval/flap_recall': 0.859481582537517, 'eval/flap_f1': 0.8955223880597014, 'eval/flap_rate_model': 0.18722222222222223, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.8955223880597015, 'timing/data': 0.011391518637537956, 'timing/model': 1.5852697677910328, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.32}) +[2026-06-15 22:01:23,920][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_800_state +[2026-06-15 22:01:23,926][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 22:01:33,057][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_800_state/model.safetensors +[2026-06-15 22:01:53,325][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_800_state/optimizer.bin +[2026-06-15 22:01:53,327][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_800_state/sampler.bin +[2026-06-15 22:01:53,327][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_800_state/sampler_1.bin +[2026-06-15 22:01:53,329][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_800_state/random_states_0.pkl +[2026-06-15 22:01:53,521][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800 +[2026-06-15 22:01:53,522][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 22:01:53,538][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 22:01:59,924][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.18018291890621185, 'train/grad_norm_pre_clip': 10.039727210998535, 'timing/data': 0.010739711113274097, 'timing/model': 1.5312619330361485, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 0.32}) +[2026-06-15 22:02:06,235][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.2868870198726654, 'train/grad_norm_pre_clip': 9.590590476989746, 'timing/data': 0.010530310217291117, 'timing/model': 1.5395932369865477, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 0.32}) +[2026-06-15 22:02:12,487][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.22102679312229156, 'train/grad_norm_pre_clip': 8.753506660461426, 'timing/data': 0.012333458289504051, 'timing/model': 1.8151024277321994, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 0.32}) +[2026-06-15 22:02:18,642][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.09570883959531784, 'train/grad_norm_pre_clip': 8.723796844482422, 'timing/data': 0.01064684009179473, 'timing/model': 1.6606829762458801, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 0.33}) +[2026-06-15 22:02:24,771][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.15007932484149933, 'train/grad_norm_pre_clip': 10.079590797424316, 'timing/data': 0.01095671160146594, 'timing/model': 1.6002989471890032, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 0.33}) +[2026-06-15 22:02:31,043][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.26893362402915955, 'train/grad_norm_pre_clip': 7.257351875305176, 'timing/data': 0.010343268048018217, 'timing/model': 1.4899304839782417, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 0.33}) +[2026-06-15 22:02:37,342][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.17682872712612152, 'train/grad_norm_pre_clip': 6.135939121246338, 'timing/data': 0.0109373708255589, 'timing/model': 1.5636944849975407, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 0.33}) +[2026-06-15 22:02:43,644][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.09430824220180511, 'train/grad_norm_pre_clip': 6.274994850158691, 'timing/data': 0.011392863932996988, 'timing/model': 1.528108132071793, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 0.33}) +[2026-06-15 22:02:49,828][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.022785039618611336, 'train/grad_norm_pre_clip': 2.446289300918579, 'timing/data': 0.010499699041247368, 'timing/model': 1.5669795898720622, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 0.33}) +[2026-06-15 22:02:56,059][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.3002781867980957, 'train/grad_norm_pre_clip': 10.55668830871582, 'timing/data': 0.010685570072382689, 'timing/model': 1.7785230451263487, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 0.33}) +[2026-06-15 22:03:02,349][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.5126115083694458, 'train/grad_norm_pre_clip': 23.8476619720459, 'timing/data': 0.010902551002800465, 'timing/model': 1.5713616809807718, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 0.33}) +[2026-06-15 22:03:08,651][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.051297347992658615, 'train/grad_norm_pre_clip': 6.652584552764893, 'timing/data': 0.010541579686105251, 'timing/model': 1.5617604511789978, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 0.33}) +[2026-06-15 22:03:14,953][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.08631980419158936, 'train/grad_norm_pre_clip': 6.877583026885986, 'timing/data': 0.011076512280851603, 'timing/model': 1.5592818749137223, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 0.33}) +[2026-06-15 22:03:21,258][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.1716405302286148, 'train/grad_norm_pre_clip': 14.940408706665039, 'timing/data': 0.010585949756205082, 'timing/model': 1.5404117917641997, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 0.33}) +[2026-06-15 22:03:27,558][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.4957112669944763, 'train/grad_norm_pre_clip': 12.045289039611816, 'timing/data': 0.01118632173165679, 'timing/model': 1.567348343320191, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 0.33}) +[2026-06-15 22:03:33,882][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.08545763045549393, 'train/grad_norm_pre_clip': 7.760544300079346, 'timing/data': 0.011955725960433483, 'timing/model': 1.7933114622719586, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 0.33}) +[2026-06-15 22:03:39,951][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.0816296935081482, 'train/grad_norm_pre_clip': 5.278547763824463, 'timing/data': 0.010447228793054819, 'timing/model': 1.563640776090324, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 0.33}) +[2026-06-15 22:03:46,256][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.49663710594177246, 'train/grad_norm_pre_clip': 11.147936820983887, 'timing/data': 0.011941264849156141, 'timing/model': 1.564842148218304, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 0.33}) +[2026-06-15 22:03:52,567][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.652949333190918, 'train/grad_norm_pre_clip': 21.81689453125, 'timing/data': 0.010870130266994238, 'timing/model': 1.5732856830582023, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 0.33}) +[2026-06-15 22:03:58,860][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.05327722057700157, 'train/grad_norm_pre_clip': 3.6766114234924316, 'timing/data': 0.011175952386111021, 'timing/model': 1.5369863002561033, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 0.33}) +[2026-06-15 22:04:05,159][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.04010983556509018, 'train/grad_norm_pre_clip': 2.6972529888153076, 'timing/data': 0.011949284933507442, 'timing/model': 1.5516421529464424, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 0.33}) +[2026-06-15 22:04:11,622][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.3294249475002289, 'train/grad_norm_pre_clip': 13.208422660827637, 'timing/data': 0.010999880731105804, 'timing/model': 1.8293287600390613, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 0.33}) +[2026-06-15 22:04:17,761][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.14313572645187378, 'train/grad_norm_pre_clip': 7.7187371253967285, 'timing/data': 0.011736564338207245, 'timing/model': 1.6396265840157866, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 0.33}) +[2026-06-15 22:04:24,269][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.3316618800163269, 'train/grad_norm_pre_clip': 12.255208015441895, 'timing/data': 0.010861839633435011, 'timing/model': 1.5906899715773761, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 0.33}) +[2026-06-15 22:04:30,666][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.06587668508291245, 'train/grad_norm_pre_clip': 5.998643398284912, 'timing/data': 0.01128186285495758, 'timing/model': 1.5654191500507295, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 0.33}) +[2026-06-15 22:04:36,963][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.4395645260810852, 'train/grad_norm_pre_clip': 13.461573600769043, 'timing/data': 0.01078855013474822, 'timing/model': 1.5177113763056695, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 0.33}) +[2026-06-15 22:04:43,333][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.20045071840286255, 'train/grad_norm_pre_clip': 8.989954948425293, 'timing/data': 0.010975970886647701, 'timing/model': 1.5786202042363584, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 0.33}) +[2026-06-15 22:04:49,597][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.27355486154556274, 'train/grad_norm_pre_clip': 7.104091167449951, 'timing/data': 0.011434533167630434, 'timing/model': 1.7787321261130273, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 0.33}) +[2026-06-15 22:04:55,661][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.22138231992721558, 'train/grad_norm_pre_clip': 8.575791358947754, 'timing/data': 0.01048674900084734, 'timing/model': 1.5623879912309349, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 0.34}) +[2026-06-15 22:05:01,962][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.07925944030284882, 'train/grad_norm_pre_clip': 4.401895046234131, 'timing/data': 0.011055480688810349, 'timing/model': 1.565646851900965, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 0.34}) +[2026-06-15 22:05:08,256][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.25756189227104187, 'train/grad_norm_pre_clip': 7.9776506423950195, 'timing/data': 0.011176900938153267, 'timing/model': 1.5542215462774038, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 0.34}) +[2026-06-15 22:05:14,566][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.1928044706583023, 'train/grad_norm_pre_clip': 5.225096225738525, 'timing/data': 0.010914530139416456, 'timing/model': 1.5624088398180902, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 0.34}) +[2026-06-15 22:05:20,945][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.2879830598831177, 'train/grad_norm_pre_clip': 18.518970489501953, 'timing/data': 0.01186928479000926, 'timing/model': 1.59605768090114, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 0.34}) +[2026-06-15 22:05:27,304][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.26129135489463806, 'train/grad_norm_pre_clip': 11.209833145141602, 'timing/data': 0.010890229605138302, 'timing/model': 1.79168371995911, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 0.34}) +[2026-06-15 22:05:33,467][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.14900678396224976, 'train/grad_norm_pre_clip': 9.615735054016113, 'timing/data': 0.011551972944289446, 'timing/model': 1.5609688819386065, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.34}) +[2026-06-15 22:05:39,773][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.2820718288421631, 'train/grad_norm_pre_clip': 11.641927719116211, 'timing/data': 0.011561273131519556, 'timing/model': 1.5681323716416955, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 0.34}) +[2026-06-15 22:05:46,070][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.1656196415424347, 'train/grad_norm_pre_clip': 5.077099323272705, 'timing/data': 0.010608118958771229, 'timing/model': 1.5592437167651951, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 0.34}) +[2026-06-15 22:05:52,367][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.08914284408092499, 'train/grad_norm_pre_clip': 3.6663620471954346, 'timing/data': 0.010588608682155609, 'timing/model': 1.562055726069957, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 0.34}) +[2026-06-15 22:05:58,665][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.062413040548563004, 'train/grad_norm_pre_clip': 4.553431510925293, 'timing/data': 0.010809368919581175, 'timing/model': 1.526431238744408, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 0.34}) +[2026-06-15 22:06:04,996][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.24175916612148285, 'train/grad_norm_pre_clip': 11.200114250183105, 'timing/data': 0.010790419299155474, 'timing/model': 1.7932082838378847, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.34}) +[2026-06-15 22:06:11,072][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.3693457841873169, 'train/grad_norm_pre_clip': 15.053193092346191, 'timing/data': 0.011985334567725658, 'timing/model': 1.563503383193165, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 0.34}) +[2026-06-15 22:06:17,359][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.1401217132806778, 'train/grad_norm_pre_clip': 7.487832546234131, 'timing/data': 0.010277477093040943, 'timing/model': 1.5623794640414417, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 0.34}) +[2026-06-15 22:06:23,661][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.025651022791862488, 'train/grad_norm_pre_clip': 2.4936699867248535, 'timing/data': 0.010177846066653728, 'timing/model': 1.5564750339835882, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 0.34}) +[2026-06-15 22:06:29,975][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.0527479387819767, 'train/grad_norm_pre_clip': 10.189956665039062, 'timing/data': 0.010290676727890968, 'timing/model': 1.5768655138090253, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 0.34}) +[2026-06-15 22:06:36,273][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.4657348394393921, 'train/grad_norm_pre_clip': 9.199831008911133, 'timing/data': 0.01118486188352108, 'timing/model': 1.526918611023575, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 0.34}) +[2026-06-15 22:06:42,613][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.11228176206350327, 'train/grad_norm_pre_clip': 11.611213684082031, 'timing/data': 0.011322191916406155, 'timing/model': 1.8047063741832972, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 0.34}) +[2026-06-15 22:06:48,828][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.08212391287088394, 'train/grad_norm_pre_clip': 13.814751625061035, 'timing/data': 0.011275711935013533, 'timing/model': 1.7191041870974004, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 0.34}) +[2026-06-15 22:06:54,960][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.2871882915496826, 'train/grad_norm_pre_clip': 29.90566062927246, 'timing/data': 0.010781798977404833, 'timing/model': 1.5934943379834294, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 0.34}) +[2026-06-15 22:07:01,271][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.20442375540733337, 'train/grad_norm_pre_clip': 14.28112506866455, 'timing/data': 0.010900620371103287, 'timing/model': 1.5532750100828707, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 0.34}) +[2026-06-15 22:07:07,667][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.06657732278108597, 'train/grad_norm_pre_clip': 8.022063255310059, 'timing/data': 0.010570188984274864, 'timing/model': 1.5847930810414255, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.34}) +[2026-06-15 22:07:14,173][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.3927838206291199, 'train/grad_norm_pre_clip': 25.225610733032227, 'timing/data': 0.01115704094991088, 'timing/model': 1.538152084685862, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 0.34}) +[2026-06-15 22:07:20,523][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.20595017075538635, 'train/grad_norm_pre_clip': 13.80567455291748, 'timing/data': 0.011209660675376654, 'timing/model': 1.8154323720373213, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 0.34}) +[2026-06-15 22:07:26,797][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.2542288899421692, 'train/grad_norm_pre_clip': 12.80795955657959, 'timing/data': 0.010898409876972437, 'timing/model': 1.7881253026425838, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 0.34}) +[2026-06-15 22:07:32,861][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.286335825920105, 'train/grad_norm_pre_clip': 27.526479721069336, 'timing/data': 0.010923470370471478, 'timing/model': 1.5568344821222126, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 0.35}) +[2026-06-15 22:07:39,170][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.09716479480266571, 'train/grad_norm_pre_clip': 10.979455947875977, 'timing/data': 0.011120641138404608, 'timing/model': 1.5688381642103195, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 0.35}) +[2026-06-15 22:07:45,462][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.0066173505038022995, 'train/grad_norm_pre_clip': 0.7194240689277649, 'timing/data': 0.011569222901016474, 'timing/model': 1.5490677608177066, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 0.35}) +[2026-06-15 22:07:51,860][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.45916998386383057, 'train/grad_norm_pre_clip': 18.56111717224121, 'timing/data': 0.010431458242237568, 'timing/model': 1.60161053808406, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 0.35}) +[2026-06-15 22:07:58,235][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.11149599403142929, 'train/grad_norm_pre_clip': 16.746356964111328, 'timing/data': 0.011530333198606968, 'timing/model': 1.8277634982950985, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 0.35}) +[2026-06-15 22:08:04,382][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.3342238664627075, 'train/grad_norm_pre_clip': 18.37367820739746, 'timing/data': 0.01135252183303237, 'timing/model': 1.571276473812759, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 0.35}) +[2026-06-15 22:08:10,672][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.12149787694215775, 'train/grad_norm_pre_clip': 15.3185396194458, 'timing/data': 0.011194751132279634, 'timing/model': 1.5569765558466315, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.35}) +[2026-06-15 22:08:17,072][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.26591068506240845, 'train/grad_norm_pre_clip': 9.220779418945312, 'timing/data': 0.011439032386988401, 'timing/model': 1.5872950218617916, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 0.35}) +[2026-06-15 22:08:23,381][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.31939083337783813, 'train/grad_norm_pre_clip': 23.487491607666016, 'timing/data': 0.011109380051493645, 'timing/model': 1.5354916830547154, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 0.35}) +[2026-06-15 22:08:29,677][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.1337876319885254, 'train/grad_norm_pre_clip': 11.269634246826172, 'timing/data': 0.01390533335506916, 'timing/model': 1.5974366227164865, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 0.35}) +[2026-06-15 22:08:36,064][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.13607797026634216, 'train/grad_norm_pre_clip': 13.04308032989502, 'timing/data': 0.010540538001805544, 'timing/model': 1.7452167142182589, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 0.35}) +[2026-06-15 22:08:42,165][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.23882600665092468, 'train/grad_norm_pre_clip': 10.177840232849121, 'timing/data': 0.010646477807313204, 'timing/model': 1.5630716113373637, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 0.35}) +[2026-06-15 22:08:48,462][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.15660257637500763, 'train/grad_norm_pre_clip': 9.077208518981934, 'timing/data': 0.010462417732924223, 'timing/model': 1.5600842447020113, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 0.35}) +[2026-06-15 22:08:54,768][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.24990150332450867, 'train/grad_norm_pre_clip': 10.511602401733398, 'timing/data': 0.010342186782509089, 'timing/model': 1.5635726279579103, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 0.35}) +[2026-06-15 22:09:01,066][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.17015062272548676, 'train/grad_norm_pre_clip': 7.3323540687561035, 'timing/data': 0.010650608222931623, 'timing/model': 1.5272284508682787, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 0.35}) +[2026-06-15 22:09:07,461][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.3576505780220032, 'train/grad_norm_pre_clip': 9.29178524017334, 'timing/data': 0.01268956670537591, 'timing/model': 1.605845196172595, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 0.35}) +[2026-06-15 22:09:13,873][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.07193048298358917, 'train/grad_norm_pre_clip': 6.965959548950195, 'timing/data': 0.0106005878187716, 'timing/model': 1.799219983164221, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.35}) +[2026-06-15 22:09:19,982][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.08728502690792084, 'train/grad_norm_pre_clip': 6.369183540344238, 'timing/data': 0.011828782968223095, 'timing/model': 1.5735685350373387, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 0.35}) +[2026-06-15 22:09:26,270][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.13978467881679535, 'train/grad_norm_pre_clip': 5.634537696838379, 'timing/data': 0.010940319392830133, 'timing/model': 1.5632598949596286, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 0.35}) +[2026-06-15 22:09:32,560][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.24677574634552002, 'train/grad_norm_pre_clip': 9.004015922546387, 'timing/data': 0.01015573600307107, 'timing/model': 1.5580844790674746, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 0.35}) +[2026-06-15 22:09:38,861][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.12138935178518295, 'train/grad_norm_pre_clip': 4.648613929748535, 'timing/data': 0.009648272767663002, 'timing/model': 1.525660970248282, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 0.35}) +[2026-06-15 22:09:45,206][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.2224389612674713, 'train/grad_norm_pre_clip': 5.684114933013916, 'timing/data': 0.009990124963223934, 'timing/model': 1.8082691710442305, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 0.35}) +[2026-06-15 22:09:51,381][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.24629396200180054, 'train/grad_norm_pre_clip': 14.494176864624023, 'timing/data': 0.00975740421563387, 'timing/model': 1.5783059648238122, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 0.35}) +[2026-06-15 22:09:57,775][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.23962315917015076, 'train/grad_norm_pre_clip': 5.8828444480896, 'timing/data': 0.010568497702479362, 'timing/model': 1.5766505650244653, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 0.35}) +[2026-06-15 22:10:04,117][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.34650036692619324, 'train/grad_norm_pre_clip': 18.122236251831055, 'timing/data': 0.010838689748197794, 'timing/model': 1.603179823141545, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 0.36}) +[2026-06-15 22:10:10,455][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.0994403287768364, 'train/grad_norm_pre_clip': 6.21401309967041, 'timing/data': 0.011085759848356247, 'timing/model': 1.6360048498027027, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 0.36}) +[2026-06-15 22:10:16,889][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.22465167939662933, 'train/grad_norm_pre_clip': 9.06672477722168, 'timing/data': 0.011494971811771393, 'timing/model': 1.5751600521616638, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 0.36}) +[2026-06-15 22:10:23,218][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.15287642180919647, 'train/grad_norm_pre_clip': 7.314493179321289, 'timing/data': 0.010951400268822908, 'timing/model': 1.8077113539911807, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 0.36}) +[2026-06-15 22:10:29,438][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.20885848999023438, 'train/grad_norm_pre_clip': 15.232973098754883, 'timing/data': 0.0127068180590868, 'timing/model': 1.590031925123185, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 0.36}) +[2026-06-15 22:10:35,771][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.10106701403856277, 'train/grad_norm_pre_clip': 13.31123161315918, 'timing/data': 0.011040839832276106, 'timing/model': 1.5721885520033538, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 0.36}) +[2026-06-15 22:10:42,061][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.09343907982110977, 'train/grad_norm_pre_clip': 6.833083629608154, 'timing/data': 0.01043567806482315, 'timing/model': 1.557792915031314, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.36}) +[2026-06-15 22:10:48,463][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.1639372855424881, 'train/grad_norm_pre_clip': 8.615323066711426, 'timing/data': 0.011081479955464602, 'timing/model': 1.560210573952645, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 0.36}) +[2026-06-15 22:10:54,760][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.11863640695810318, 'train/grad_norm_pre_clip': 5.432714462280273, 'timing/data': 0.010002365335822105, 'timing/model': 1.5592006570659578, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 0.36}) +[2026-06-15 22:11:01,092][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.16924956440925598, 'train/grad_norm_pre_clip': 12.219792366027832, 'timing/data': 0.010800718795508146, 'timing/model': 1.7944925902411342, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 0.36}) +[2026-06-15 22:11:07,169][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.028311626985669136, 'train/grad_norm_pre_clip': 3.795974016189575, 'timing/data': 0.011879234109073877, 'timing/model': 1.5653937519527972, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 0.36}) +[2026-06-15 22:11:13,474][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.18573260307312012, 'train/grad_norm_pre_clip': 6.963869094848633, 'timing/data': 0.010891689918935299, 'timing/model': 1.5663418639451265, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 0.36}) +[2026-06-15 22:11:19,773][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.14342904090881348, 'train/grad_norm_pre_clip': 11.076351165771484, 'timing/data': 0.010275986976921558, 'timing/model': 1.5610928181558847, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 0.36}) +[2026-06-15 22:11:26,070][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.2365676313638687, 'train/grad_norm_pre_clip': 9.807106971740723, 'timing/data': 0.01088540954515338, 'timing/model': 1.5368926571682096, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 0.36}) +[2026-06-15 22:11:32,376][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.11455005407333374, 'train/grad_norm_pre_clip': 5.90595817565918, 'timing/data': 0.01135057071223855, 'timing/model': 1.562809722032398, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 0.36}) +[2026-06-15 22:11:38,697][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.284853458404541, 'train/grad_norm_pre_clip': 16.153423309326172, 'timing/data': 0.01096116891130805, 'timing/model': 1.7936423239298165, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 0.36}) +[2026-06-15 22:11:44,765][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.08185175806283951, 'train/grad_norm_pre_clip': 5.236832618713379, 'timing/data': 0.010794788133352995, 'timing/model': 1.5584996198303998, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 0.36}) +[2026-06-15 22:11:51,059][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.06561515480279922, 'train/grad_norm_pre_clip': 7.2997846603393555, 'timing/data': 0.010274656116962433, 'timing/model': 1.5623356262221932, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 0.36}) +[2026-06-15 22:11:57,363][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.09540966898202896, 'train/grad_norm_pre_clip': 8.89179515838623, 'timing/data': 0.009979315102100372, 'timing/model': 1.530440220143646, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 0.36}) +[2026-06-15 22:12:03,773][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.14344406127929688, 'train/grad_norm_pre_clip': 10.34631633758545, 'timing/data': 0.010664257686585188, 'timing/model': 1.6742379870265722, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 0.36}) +[2026-06-15 22:12:10,161][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.19378122687339783, 'train/grad_norm_pre_clip': 13.059205055236816, 'timing/data': 0.011238350067287683, 'timing/model': 1.6030976241454482, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 0.36}) +[2026-06-15 22:12:16,506][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.48645898699760437, 'train/grad_norm_pre_clip': 20.08055877685547, 'timing/data': 0.011433860752731562, 'timing/model': 1.801412875764072, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 0.36}) +[2026-06-15 22:23:37,431][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.30706775188446045, 'train/grad_norm_pre_clip': 15.611613273620605, 'eval/loss': 0.15706634521484375, 'eval/flap_precision': 0.7794432548179872, 'eval/flap_recall': 0.9931787175989086, 'eval/flap_f1': 0.8734253149370126, 'eval/flap_rate_model': 0.2594444444444444, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.8734253149370127, 'timing/data': 0.01180450338870287, 'timing/model': 1.5848038182593882, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 0.36}) +[2026-06-15 22:23:37,432][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_900_state +[2026-06-15 22:23:37,439][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 22:23:46,287][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_900_state/model.safetensors +[2026-06-15 22:24:06,253][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_900_state/optimizer.bin +[2026-06-15 22:24:06,254][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_900_state/sampler.bin +[2026-06-15 22:24:06,254][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_900_state/sampler_1.bin +[2026-06-15 22:24:06,255][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_900_state/random_states_0.pkl +[2026-06-15 22:24:06,453][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-06-15 22:24:06,454][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 22:24:06,469][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 22:24:12,810][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.38029876351356506, 'train/grad_norm_pre_clip': 20.155691146850586, 'timing/data': 0.011669002007693052, 'timing/model': 1.8085198230110109, 'learning_rate/qwen_vl_interface': 9.387268552285024e-06, 'learning_rate/action_model': 9.387268552285025e-05, 'epoch': 0.36}) +[2026-06-15 22:24:18,945][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.14901724457740784, 'train/grad_norm_pre_clip': 7.6650238037109375, 'timing/data': 0.012225015088915825, 'timing/model': 1.576701431069523, 'learning_rate/qwen_vl_interface': 9.385771559309125e-06, 'learning_rate/action_model': 9.385771559309125e-05, 'epoch': 0.36}) +[2026-06-15 22:24:25,250][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.12494053691625595, 'train/grad_norm_pre_clip': 8.851288795471191, 'timing/data': 0.01289867702871561, 'timing/model': 1.5631521912291646, 'learning_rate/qwen_vl_interface': 9.38427286627376e-06, 'learning_rate/action_model': 9.38427286627376e-05, 'epoch': 0.37}) +[2026-06-15 22:24:31,540][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.015644341707229614, 'train/grad_norm_pre_clip': 2.261254072189331, 'timing/data': 0.010973609052598476, 'timing/model': 1.57544870628044, 'learning_rate/qwen_vl_interface': 9.382772473794985e-06, 'learning_rate/action_model': 9.382772473794985e-05, 'epoch': 0.37}) +[2026-06-15 22:24:37,843][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.24135029315948486, 'train/grad_norm_pre_clip': 12.319007873535156, 'timing/data': 0.010415077209472656, 'timing/model': 1.5647575268521905, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.37}) +[2026-06-15 22:24:44,038][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.20439347624778748, 'train/grad_norm_pre_clip': 13.780899047851562, 'timing/data': 0.010560006834566593, 'timing/model': 1.5124431350268424, 'learning_rate/qwen_vl_interface': 9.37976659297492e-06, 'learning_rate/action_model': 9.379766592974921e-05, 'epoch': 0.37}) +[2026-06-15 22:24:50,289][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.36825791001319885, 'train/grad_norm_pre_clip': 22.763370513916016, 'timing/data': 0.011733731720596552, 'timing/model': 1.8072839053347707, 'learning_rate/qwen_vl_interface': 9.378261105869237e-06, 'learning_rate/action_model': 9.378261105869237e-05, 'epoch': 0.37}) +[2026-06-15 22:24:56,581][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.04487781971693039, 'train/grad_norm_pre_clip': 2.747631311416626, 'timing/data': 0.013462180271744728, 'timing/model': 1.7642872482538223, 'learning_rate/qwen_vl_interface': 9.376753921791348e-06, 'learning_rate/action_model': 9.376753921791348e-05, 'epoch': 0.37}) +[2026-06-15 22:25:02,650][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.2837172746658325, 'train/grad_norm_pre_clip': 17.006671905517578, 'timing/data': 0.011766970157623291, 'timing/model': 1.554575421847403, 'learning_rate/qwen_vl_interface': 9.375245041360805e-06, 'learning_rate/action_model': 9.375245041360804e-05, 'epoch': 0.37}) +[2026-06-15 22:25:08,954][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.05239269137382507, 'train/grad_norm_pre_clip': 2.434774875640869, 'timing/data': 0.011534557212144136, 'timing/model': 1.5613098042085767, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 0.37}) +[2026-06-15 22:25:15,366][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.1499006450176239, 'train/grad_norm_pre_clip': 19.44881248474121, 'timing/data': 0.010097741149365902, 'timing/model': 1.5520815118215978, 'learning_rate/qwen_vl_interface': 9.372222193923421e-06, 'learning_rate/action_model': 9.372222193923421e-05, 'epoch': 0.37}) +[2026-06-15 22:25:21,665][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.0789574682712555, 'train/grad_norm_pre_clip': 5.288760185241699, 'timing/data': 0.01115092122927308, 'timing/model': 1.5187772950157523, 'learning_rate/qwen_vl_interface': 9.370708228159158e-06, 'learning_rate/action_model': 9.370708228159159e-05, 'epoch': 0.37}) +[2026-06-15 22:25:28,002][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.11250481009483337, 'train/grad_norm_pre_clip': 6.801211357116699, 'timing/data': 0.011270402930676937, 'timing/model': 1.8103344696573913, 'learning_rate/qwen_vl_interface': 9.369192568527398e-06, 'learning_rate/action_model': 9.369192568527398e-05, 'epoch': 0.37}) +[2026-06-15 22:25:34,057][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.11784503608942032, 'train/grad_norm_pre_clip': 14.021345138549805, 'timing/data': 0.01111252000555396, 'timing/model': 1.5425906819291413, 'learning_rate/qwen_vl_interface': 9.367675215651168e-06, 'learning_rate/action_model': 9.367675215651168e-05, 'epoch': 0.37}) +[2026-06-15 22:25:40,370][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.13385114073753357, 'train/grad_norm_pre_clip': 10.567357063293457, 'timing/data': 0.011823636945337057, 'timing/model': 1.5697846771217883, 'learning_rate/qwen_vl_interface': 9.366156170154194e-06, 'learning_rate/action_model': 9.366156170154194e-05, 'epoch': 0.37}) +[2026-06-15 22:25:46,663][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.053716130554676056, 'train/grad_norm_pre_clip': 10.329896926879883, 'timing/data': 0.011322000063955784, 'timing/model': 1.5609729010611773, 'learning_rate/qwen_vl_interface': 9.3646354326609e-06, 'learning_rate/action_model': 9.364635432660899e-05, 'epoch': 0.37}) +[2026-06-15 22:25:53,048][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.1417899876832962, 'train/grad_norm_pre_clip': 4.467811107635498, 'timing/data': 0.011432101018726826, 'timing/model': 1.5837103119120002, 'learning_rate/qwen_vl_interface': 9.3631130037964e-06, 'learning_rate/action_model': 9.363113003796401e-05, 'epoch': 0.37}) +[2026-06-15 22:25:59,761][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.128279447555542, 'train/grad_norm_pre_clip': 11.029311180114746, 'timing/data': 0.011588252149522305, 'timing/model': 1.5847669448703527, 'learning_rate/qwen_vl_interface': 9.361588884186513e-06, 'learning_rate/action_model': 9.361588884186513e-05, 'epoch': 0.37}) +[2026-06-15 22:26:06,050][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.35407158732414246, 'train/grad_norm_pre_clip': 30.363550186157227, 'timing/data': 0.01064949156716466, 'timing/model': 1.71078400593251, 'learning_rate/qwen_vl_interface': 9.360063074457744e-06, 'learning_rate/action_model': 9.360063074457743e-05, 'epoch': 0.37}) +[2026-06-15 22:26:12,158][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.10694210231304169, 'train/grad_norm_pre_clip': 9.735780715942383, 'timing/data': 0.011256278026849031, 'timing/model': 1.549473196733743, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 0.37}) +[2026-06-15 22:26:18,463][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.1591043472290039, 'train/grad_norm_pre_clip': 12.453603744506836, 'timing/data': 0.010925634298473597, 'timing/model': 1.5604022941552103, 'learning_rate/qwen_vl_interface': 9.357006387153065e-06, 'learning_rate/action_model': 9.357006387153065e-05, 'epoch': 0.37}) +[2026-06-15 22:26:24,758][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.1749218851327896, 'train/grad_norm_pre_clip': 10.93161678314209, 'timing/data': 0.011094855144619942, 'timing/model': 1.5618771747685969, 'learning_rate/qwen_vl_interface': 9.355475510833643e-06, 'learning_rate/action_model': 9.355475510833643e-05, 'epoch': 0.37}) +[2026-06-15 22:26:31,060][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.1295749992132187, 'train/grad_norm_pre_clip': 9.962907791137695, 'timing/data': 0.010544158983975649, 'timing/model': 1.5222194008529186, 'learning_rate/qwen_vl_interface': 9.353942946908318e-06, 'learning_rate/action_model': 9.353942946908317e-05, 'epoch': 0.37}) +[2026-06-15 22:26:37,355][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.055039212107658386, 'train/grad_norm_pre_clip': 5.186563014984131, 'timing/data': 0.01082704123109579, 'timing/model': 1.5246989927254617, 'learning_rate/qwen_vl_interface': 9.352408696007065e-06, 'learning_rate/action_model': 9.352408696007065e-05, 'epoch': 0.37}) +[2026-06-15 22:26:43,658][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.03236955404281616, 'train/grad_norm_pre_clip': 3.186532497406006, 'timing/data': 0.010964242275804281, 'timing/model': 1.760764071252197, 'learning_rate/qwen_vl_interface': 9.350872758760558e-06, 'learning_rate/action_model': 9.350872758760559e-05, 'epoch': 0.37}) +[2026-06-15 22:26:49,766][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.3777711093425751, 'train/grad_norm_pre_clip': 36.088050842285156, 'timing/data': 0.011342605110257864, 'timing/model': 1.5651778518222272, 'learning_rate/qwen_vl_interface': 9.349335135800164e-06, 'learning_rate/action_model': 9.349335135800163e-05, 'epoch': 0.37}) +[2026-06-15 22:26:56,063][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.10458306968212128, 'train/grad_norm_pre_clip': 8.022825241088867, 'timing/data': 0.012008531019091606, 'timing/model': 1.566062331199646, 'learning_rate/qwen_vl_interface': 9.347795827757942e-06, 'learning_rate/action_model': 9.347795827757941e-05, 'epoch': 0.37}) +[2026-06-15 22:27:02,362][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.23860414326190948, 'train/grad_norm_pre_clip': 16.830900192260742, 'timing/data': 0.011464334093034267, 'timing/model': 1.5592025900259614, 'learning_rate/qwen_vl_interface': 9.346254835266643e-06, 'learning_rate/action_model': 9.346254835266642e-05, 'epoch': 0.38}) +[2026-06-15 22:27:08,662][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.27527138590812683, 'train/grad_norm_pre_clip': 9.655049324035645, 'timing/data': 0.011083601042628288, 'timing/model': 1.5230651050806046, 'learning_rate/qwen_vl_interface': 9.344712158959709e-06, 'learning_rate/action_model': 9.344712158959709e-05, 'epoch': 0.38}) +[2026-06-15 22:27:15,002][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.036097340285778046, 'train/grad_norm_pre_clip': 6.703773021697998, 'timing/data': 0.011370322667062283, 'timing/model': 1.7948376312851906, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 0.38}) +[2026-06-15 22:27:21,058][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.12864024937152863, 'train/grad_norm_pre_clip': 16.572349548339844, 'timing/data': 0.01135682174935937, 'timing/model': 1.5504656569100916, 'learning_rate/qwen_vl_interface': 9.341621757436177e-06, 'learning_rate/action_model': 9.341621757436177e-05, 'epoch': 0.38}) +[2026-06-15 22:27:27,359][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.19490878283977509, 'train/grad_norm_pre_clip': 12.999567985534668, 'timing/data': 0.010968438349664211, 'timing/model': 1.5434781690128148, 'learning_rate/qwen_vl_interface': 9.340074033489924e-06, 'learning_rate/action_model': 9.340074033489923e-05, 'epoch': 0.38}) +[2026-06-15 22:27:33,560][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.06254443526268005, 'train/grad_norm_pre_clip': 11.405020713806152, 'timing/data': 0.010945287067443132, 'timing/model': 1.5617957599461079, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.38}) +[2026-06-15 22:27:39,853][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.08215754479169846, 'train/grad_norm_pre_clip': 10.173678398132324, 'timing/data': 0.01097953598946333, 'timing/model': 1.5501079759560525, 'learning_rate/qwen_vl_interface': 9.336973542409505e-06, 'learning_rate/action_model': 9.336973542409504e-05, 'epoch': 0.38}) +[2026-06-15 22:27:46,176][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.06474801152944565, 'train/grad_norm_pre_clip': 8.198067665100098, 'timing/data': 0.010863115079700947, 'timing/model': 1.5632479591295123, 'learning_rate/qwen_vl_interface': 9.33542077654983e-06, 'learning_rate/action_model': 9.33542077654983e-05, 'epoch': 0.38}) +[2026-06-15 22:27:52,505][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.467079758644104, 'train/grad_norm_pre_clip': 32.858734130859375, 'timing/data': 0.010263179894536734, 'timing/model': 1.8002924090251327, 'learning_rate/qwen_vl_interface': 9.333866331327995e-06, 'learning_rate/action_model': 9.333866331327995e-05, 'epoch': 0.38}) +[2026-06-15 22:27:58,563][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.16243144869804382, 'train/grad_norm_pre_clip': 23.426572799682617, 'timing/data': 0.011574251111596823, 'timing/model': 1.5641447068192065, 'learning_rate/qwen_vl_interface': 9.332310207382972e-06, 'learning_rate/action_model': 9.332310207382972e-05, 'epoch': 0.38}) +[2026-06-15 22:28:04,866][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.3529864549636841, 'train/grad_norm_pre_clip': 21.265905380249023, 'timing/data': 0.01286623114719987, 'timing/model': 1.5658784438855946, 'learning_rate/qwen_vl_interface': 9.330752405354426e-06, 'learning_rate/action_model': 9.330752405354426e-05, 'epoch': 0.38}) +[2026-06-15 22:28:11,162][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.07941380888223648, 'train/grad_norm_pre_clip': 26.41325569152832, 'timing/data': 0.011310367844998837, 'timing/model': 1.5536184697411954, 'learning_rate/qwen_vl_interface': 9.32919292588271e-06, 'learning_rate/action_model': 9.32919292588271e-05, 'epoch': 0.38}) +[2026-06-15 22:28:17,618][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.08678838610649109, 'train/grad_norm_pre_clip': 12.453804016113281, 'timing/data': 0.010630280710756779, 'timing/model': 1.626286803279072, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.38}) +[2026-06-15 22:28:23,945][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.22382476925849915, 'train/grad_norm_pre_clip': 10.298993110656738, 'timing/data': 0.011067943647503853, 'timing/model': 1.640885419677943, 'learning_rate/qwen_vl_interface': 9.326068937174627e-06, 'learning_rate/action_model': 9.326068937174626e-05, 'epoch': 0.38}) +[2026-06-15 22:28:30,209][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.08999920636415482, 'train/grad_norm_pre_clip': 10.746513366699219, 'timing/data': 0.011574368923902512, 'timing/model': 1.8007065439596772, 'learning_rate/qwen_vl_interface': 9.324504429222412e-06, 'learning_rate/action_model': 9.324504429222412e-05, 'epoch': 0.38}) +[2026-06-15 22:28:36,268][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.2264566421508789, 'train/grad_norm_pre_clip': 3.2890589237213135, 'timing/data': 0.010786450933665037, 'timing/model': 1.5605140356346965, 'learning_rate/qwen_vl_interface': 9.322938246395336e-06, 'learning_rate/action_model': 9.322938246395337e-05, 'epoch': 0.38}) +[2026-06-15 22:28:42,565][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.2642890214920044, 'train/grad_norm_pre_clip': 12.022652626037598, 'timing/data': 0.0109626823104918, 'timing/model': 1.5604839911684394, 'learning_rate/qwen_vl_interface': 9.321370389337196e-06, 'learning_rate/action_model': 9.321370389337196e-05, 'epoch': 0.38}) +[2026-06-15 22:28:48,971][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.12613868713378906, 'train/grad_norm_pre_clip': 4.839158535003662, 'timing/data': 0.011176783125847578, 'timing/model': 1.573634167201817, 'learning_rate/qwen_vl_interface': 9.319800858692478e-06, 'learning_rate/action_model': 9.319800858692476e-05, 'epoch': 0.38}) +[2026-06-15 22:28:55,264][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.11103612929582596, 'train/grad_norm_pre_clip': 5.900212287902832, 'timing/data': 0.010691928677260876, 'timing/model': 1.527418009005487, 'learning_rate/qwen_vl_interface': 9.318229655106357e-06, 'learning_rate/action_model': 9.318229655106356e-05, 'epoch': 0.38}) +[2026-06-15 22:29:01,565][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.14583148062229156, 'train/grad_norm_pre_clip': 7.589542865753174, 'timing/data': 0.011502345092594624, 'timing/model': 1.546859520021826, 'learning_rate/qwen_vl_interface': 9.316656779224693e-06, 'learning_rate/action_model': 9.316656779224693e-05, 'epoch': 0.38}) +[2026-06-15 22:29:07,908][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.1770707666873932, 'train/grad_norm_pre_clip': 6.505818843841553, 'timing/data': 0.010610627010464668, 'timing/model': 1.8040286470204592, 'learning_rate/qwen_vl_interface': 9.31508223169404e-06, 'learning_rate/action_model': 9.31508223169404e-05, 'epoch': 0.38}) +[2026-06-15 22:29:14,069][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.18995313346385956, 'train/grad_norm_pre_clip': 3.565237522125244, 'timing/data': 0.012187739834189415, 'timing/model': 1.5587675389833748, 'learning_rate/qwen_vl_interface': 9.313506013161633e-06, 'learning_rate/action_model': 9.313506013161633e-05, 'epoch': 0.38}) +[2026-06-15 22:29:20,361][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.4022134840488434, 'train/grad_norm_pre_clip': 13.851143836975098, 'timing/data': 0.011265912093222141, 'timing/model': 1.555445081088692, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 0.38}) +[2026-06-15 22:29:26,675][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.2723831832408905, 'train/grad_norm_pre_clip': 6.89453125, 'timing/data': 0.010566025972366333, 'timing/model': 1.5672123883850873, 'learning_rate/qwen_vl_interface': 9.310348565683938e-06, 'learning_rate/action_model': 9.310348565683939e-05, 'epoch': 0.38}) +[2026-06-15 22:29:32,967][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.2939741611480713, 'train/grad_norm_pre_clip': 9.479246139526367, 'timing/data': 0.011337252333760262, 'timing/model': 1.5605464968830347, 'learning_rate/qwen_vl_interface': 9.30876733803656e-06, 'learning_rate/action_model': 9.308767338036559e-05, 'epoch': 0.38}) +[2026-06-15 22:29:39,269][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.1325540542602539, 'train/grad_norm_pre_clip': 6.238526821136475, 'timing/data': 0.01211319724097848, 'timing/model': 1.5261638620868325, 'learning_rate/qwen_vl_interface': 9.30718444198324e-06, 'learning_rate/action_model': 9.307184441983239e-05, 'epoch': 0.39}) +[2026-06-15 22:29:45,603][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.23966814577579498, 'train/grad_norm_pre_clip': 5.219212532043457, 'timing/data': 0.010750825982540846, 'timing/model': 1.7973725288175046, 'learning_rate/qwen_vl_interface': 9.30559987817465e-06, 'learning_rate/action_model': 9.30559987817465e-05, 'epoch': 0.39}) +[2026-06-15 22:29:51,667][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.28338319063186646, 'train/grad_norm_pre_clip': 12.107531547546387, 'timing/data': 0.011134529020637274, 'timing/model': 1.562212270218879, 'learning_rate/qwen_vl_interface': 9.304013647262143e-06, 'learning_rate/action_model': 9.304013647262143e-05, 'epoch': 0.39}) +[2026-06-15 22:29:57,963][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.10348215699195862, 'train/grad_norm_pre_clip': 4.396270751953125, 'timing/data': 0.011091438122093678, 'timing/model': 1.554614454973489, 'learning_rate/qwen_vl_interface': 9.302425749897761e-06, 'learning_rate/action_model': 9.302425749897761e-05, 'epoch': 0.39}) +[2026-06-15 22:30:04,276][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.26085150241851807, 'train/grad_norm_pre_clip': 12.648158073425293, 'timing/data': 0.011357390321791172, 'timing/model': 1.5714158327318728, 'learning_rate/qwen_vl_interface': 9.300836186734225e-06, 'learning_rate/action_model': 9.300836186734226e-05, 'epoch': 0.39}) +[2026-06-15 22:30:10,577][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.1251317262649536, 'train/grad_norm_pre_clip': 11.313692092895508, 'timing/data': 0.011026166845113039, 'timing/model': 1.5712679871357977, 'learning_rate/qwen_vl_interface': 9.299244958424947e-06, 'learning_rate/action_model': 9.299244958424946e-05, 'epoch': 0.39}) +[2026-06-15 22:30:16,952][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.19392818212509155, 'train/grad_norm_pre_clip': 7.868234157562256, 'timing/data': 0.010744113940745592, 'timing/model': 1.5969755458645523, 'learning_rate/qwen_vl_interface': 9.29765206562402e-06, 'learning_rate/action_model': 9.297652065624018e-05, 'epoch': 0.39}) +[2026-06-15 22:30:23,319][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.19668298959732056, 'train/grad_norm_pre_clip': 4.748506546020508, 'timing/data': 0.012140775099396706, 'timing/model': 1.8153236317448318, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 0.39}) +[2026-06-15 22:30:29,372][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.2121240496635437, 'train/grad_norm_pre_clip': 14.036420822143555, 'timing/data': 0.011223276611417532, 'timing/model': 1.5635794098488986, 'learning_rate/qwen_vl_interface': 9.294461289167013e-06, 'learning_rate/action_model': 9.294461289167014e-05, 'epoch': 0.39}) +[2026-06-15 22:30:35,770][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.1885952204465866, 'train/grad_norm_pre_clip': 10.759696006774902, 'timing/data': 0.011538479011505842, 'timing/model': 1.5496794199571013, 'learning_rate/qwen_vl_interface': 9.292863406822543e-06, 'learning_rate/action_model': 9.292863406822542e-05, 'epoch': 0.39}) +[2026-06-15 22:30:42,173][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.24414972960948944, 'train/grad_norm_pre_clip': 8.872573852539062, 'timing/data': 0.011251047253608704, 'timing/model': 1.5807392578572035, 'learning_rate/qwen_vl_interface': 9.291263862609638e-06, 'learning_rate/action_model': 9.291263862609638e-05, 'epoch': 0.39}) +[2026-06-15 22:30:48,467][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.29627686738967896, 'train/grad_norm_pre_clip': 8.872968673706055, 'timing/data': 0.011171936057507992, 'timing/model': 1.5639909580349922, 'learning_rate/qwen_vl_interface': 9.28966265718581e-06, 'learning_rate/action_model': 9.28966265718581e-05, 'epoch': 0.39}) +[2026-06-15 22:30:54,875][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.24308903515338898, 'train/grad_norm_pre_clip': 26.928834915161133, 'timing/data': 0.011345146223902702, 'timing/model': 1.5677628479897976, 'learning_rate/qwen_vl_interface': 9.288059791209256e-06, 'learning_rate/action_model': 9.288059791209256e-05, 'epoch': 0.39}) +[2026-06-15 22:31:01,221][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.2043357789516449, 'train/grad_norm_pre_clip': 9.509013175964355, 'timing/data': 0.01111333491280675, 'timing/model': 1.820536460261792, 'learning_rate/qwen_vl_interface': 9.286455265338853e-06, 'learning_rate/action_model': 9.286455265338853e-05, 'epoch': 0.39}) +[2026-06-15 22:31:07,327][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.21093018352985382, 'train/grad_norm_pre_clip': 13.087541580200195, 'timing/data': 0.01194175984710455, 'timing/model': 1.6239162157289684, 'learning_rate/qwen_vl_interface': 9.284849080234161e-06, 'learning_rate/action_model': 9.28484908023416e-05, 'epoch': 0.39}) +[2026-06-15 22:31:13,565][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.1534079909324646, 'train/grad_norm_pre_clip': 8.766218185424805, 'timing/data': 0.010917033068835735, 'timing/model': 1.5637893718667328, 'learning_rate/qwen_vl_interface': 9.283241236555421e-06, 'learning_rate/action_model': 9.283241236555421e-05, 'epoch': 0.39}) +[2026-06-15 22:31:19,867][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.1273210048675537, 'train/grad_norm_pre_clip': 5.833622932434082, 'timing/data': 0.012421303894370794, 'timing/model': 1.5656634592451155, 'learning_rate/qwen_vl_interface': 9.281631734963557e-06, 'learning_rate/action_model': 9.281631734963557e-05, 'epoch': 0.39}) +[2026-06-15 22:31:26,178][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.19456999003887177, 'train/grad_norm_pre_clip': 11.285039901733398, 'timing/data': 0.01097498182207346, 'timing/model': 1.5329114380292594, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 0.39}) +[2026-06-15 22:31:32,468][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.16638317704200745, 'train/grad_norm_pre_clip': 10.56834888458252, 'timing/data': 0.010642730165272951, 'timing/model': 1.5304234731011093, 'learning_rate/qwen_vl_interface': 9.278407760687562e-06, 'learning_rate/action_model': 9.278407760687562e-05, 'epoch': 0.39}) +[2026-06-15 22:31:38,800][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.039742518216371536, 'train/grad_norm_pre_clip': 2.486401081085205, 'timing/data': 0.011297483928501606, 'timing/model': 1.796325343195349, 'learning_rate/qwen_vl_interface': 9.276793289328686e-06, 'learning_rate/action_model': 9.276793289328685e-05, 'epoch': 0.39}) +[2026-06-15 22:31:45,024][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.2870400547981262, 'train/grad_norm_pre_clip': 11.942987442016602, 'timing/data': 0.011626726016402245, 'timing/model': 1.7186530088074505, 'learning_rate/qwen_vl_interface': 9.27517716270719e-06, 'learning_rate/action_model': 9.275177162707191e-05, 'epoch': 0.39}) +[2026-06-15 22:31:51,070][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.2222975492477417, 'train/grad_norm_pre_clip': 8.764410972595215, 'timing/data': 0.01076373877003789, 'timing/model': 1.563706908840686, 'learning_rate/qwen_vl_interface': 9.273559381487409e-06, 'learning_rate/action_model': 9.273559381487409e-05, 'epoch': 0.39}) +[2026-06-15 22:31:57,355][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.1802307516336441, 'train/grad_norm_pre_clip': 11.656403541564941, 'timing/data': 0.010417186189442873, 'timing/model': 1.5570800071582198, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.39}) +[2026-06-15 22:32:03,658][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.2434651255607605, 'train/grad_norm_pre_clip': 14.085651397705078, 'timing/data': 0.011059680953621864, 'timing/model': 1.5598810031078756, 'learning_rate/qwen_vl_interface': 9.2703188579137e-06, 'learning_rate/action_model': 9.2703188579137e-05, 'epoch': 0.39}) +[2026-06-15 22:32:09,960][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.18995419144630432, 'train/grad_norm_pre_clip': 7.663307189941406, 'timing/data': 0.01058982778340578, 'timing/model': 1.5192493749782443, 'learning_rate/qwen_vl_interface': 9.26869611689183e-06, 'learning_rate/action_model': 9.26869611689183e-05, 'epoch': 0.4}) +[2026-06-15 22:32:16,326][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.052910055965185165, 'train/grad_norm_pre_clip': 7.441127300262451, 'timing/data': 0.012689971830695868, 'timing/model': 1.8190527106635273, 'learning_rate/qwen_vl_interface': 9.267071723935784e-06, 'learning_rate/action_model': 9.267071723935784e-05, 'epoch': 0.4}) +[2026-06-15 22:32:22,589][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.18865524232387543, 'train/grad_norm_pre_clip': 3.8241753578186035, 'timing/data': 0.010840227827429771, 'timing/model': 1.7465858981013298, 'learning_rate/qwen_vl_interface': 9.26544567971329e-06, 'learning_rate/action_model': 9.26544567971329e-05, 'epoch': 0.4}) +[2026-06-15 22:32:28,655][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.26659029722213745, 'train/grad_norm_pre_clip': 28.982290267944336, 'timing/data': 0.010853968095034361, 'timing/model': 1.556065820157528, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 0.4}) +[2026-06-15 22:32:34,985][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.11261346936225891, 'train/grad_norm_pre_clip': 14.1965913772583, 'timing/data': 0.012024005874991417, 'timing/model': 1.5868929708376527, 'learning_rate/qwen_vl_interface': 9.262188640143262e-06, 'learning_rate/action_model': 9.262188640143262e-05, 'epoch': 0.4}) +[2026-06-15 22:32:41,264][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.25097569823265076, 'train/grad_norm_pre_clip': 26.526622772216797, 'timing/data': 0.012822921853512526, 'timing/model': 1.5647251610644162, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.4}) +[2026-06-15 22:32:47,694][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.1228061318397522, 'train/grad_norm_pre_clip': 9.076057434082031, 'timing/data': 0.009644558187574148, 'timing/model': 1.6453581559471786, 'learning_rate/qwen_vl_interface': 9.258925003537127e-06, 'learning_rate/action_model': 9.258925003537126e-05, 'epoch': 0.4}) +[2026-06-15 22:32:54,017][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.3713158369064331, 'train/grad_norm_pre_clip': 27.356630325317383, 'timing/data': 0.010791746899485588, 'timing/model': 1.8246382484212518, 'learning_rate/qwen_vl_interface': 9.257290713022041e-06, 'learning_rate/action_model': 9.257290713022041e-05, 'epoch': 0.4}) +[2026-06-15 22:33:00,067][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.21639111638069153, 'train/grad_norm_pre_clip': 9.520840644836426, 'timing/data': 0.010816247202455997, 'timing/model': 1.557902846019715, 'learning_rate/qwen_vl_interface': 9.255654775261112e-06, 'learning_rate/action_model': 9.255654775261112e-05, 'epoch': 0.4}) +[2026-06-15 22:33:06,359][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.07010284811258316, 'train/grad_norm_pre_clip': 9.023735046386719, 'timing/data': 0.010165642015635967, 'timing/model': 1.5511943991295993, 'learning_rate/qwen_vl_interface': 9.254017190926813e-06, 'learning_rate/action_model': 9.254017190926813e-05, 'epoch': 0.4}) +[2026-06-15 22:33:12,799][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.3710225820541382, 'train/grad_norm_pre_clip': 20.83837890625, 'timing/data': 0.011036277748644352, 'timing/model': 1.6107093999162316, 'learning_rate/qwen_vl_interface': 9.25237796069229e-06, 'learning_rate/action_model': 9.25237796069229e-05, 'epoch': 0.4}) +[2026-06-15 22:33:19,159][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.18943196535110474, 'train/grad_norm_pre_clip': 12.617827415466309, 'timing/data': 0.010290882084518671, 'timing/model': 1.5354399839416146, 'learning_rate/qwen_vl_interface': 9.250737085231371e-06, 'learning_rate/action_model': 9.250737085231372e-05, 'epoch': 0.4}) +[2026-06-15 22:33:25,458][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.3014836013317108, 'train/grad_norm_pre_clip': 16.268041610717773, 'timing/data': 0.01027448195964098, 'timing/model': 1.5843917378224432, 'learning_rate/qwen_vl_interface': 9.24909456521856e-06, 'learning_rate/action_model': 9.249094565218559e-05, 'epoch': 0.4}) +[2026-06-15 22:33:31,750][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.16449612379074097, 'train/grad_norm_pre_clip': 7.089670658111572, 'timing/data': 0.010390933137387037, 'timing/model': 1.7531157559715211, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 0.4}) +[2026-06-15 22:33:37,963][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.14156661927700043, 'train/grad_norm_pre_clip': 8.738541603088379, 'timing/data': 0.010837454814463854, 'timing/model': 1.6576148751191795, 'learning_rate/qwen_vl_interface': 9.245804594238637e-06, 'learning_rate/action_model': 9.245804594238636e-05, 'epoch': 0.4}) +[2026-06-15 22:33:44,262][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.1219874769449234, 'train/grad_norm_pre_clip': 7.756184101104736, 'timing/data': 0.01064331317320466, 'timing/model': 1.5620564459823072, 'learning_rate/qwen_vl_interface': 9.244157144623909e-06, 'learning_rate/action_model': 9.244157144623909e-05, 'epoch': 0.4}) +[2026-06-15 22:33:50,560][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.12219041585922241, 'train/grad_norm_pre_clip': 5.461611747741699, 'timing/data': 0.010401311330497265, 'timing/model': 1.564620838034898, 'learning_rate/qwen_vl_interface': 9.242508053162051e-06, 'learning_rate/action_model': 9.242508053162051e-05, 'epoch': 0.4}) +[2026-06-15 22:33:56,958][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.15980164706707, 'train/grad_norm_pre_clip': 7.762094974517822, 'timing/data': 0.009476566221565008, 'timing/model': 1.5380528001114726, 'learning_rate/qwen_vl_interface': 9.240857320530944e-06, 'learning_rate/action_model': 9.240857320530943e-05, 'epoch': 0.4}) +[2026-06-15 22:34:03,263][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.16426102817058563, 'train/grad_norm_pre_clip': 5.967213153839111, 'timing/data': 0.010590153280645609, 'timing/model': 1.5315167107619345, 'learning_rate/qwen_vl_interface': 9.23920494740914e-06, 'learning_rate/action_model': 9.23920494740914e-05, 'epoch': 0.4}) +[2026-06-15 22:34:09,560][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.1723669022321701, 'train/grad_norm_pre_clip': 27.843687057495117, 'timing/data': 0.011628418695181608, 'timing/model': 1.75626813294366, 'learning_rate/qwen_vl_interface': 9.237550934475868e-06, 'learning_rate/action_model': 9.237550934475867e-05, 'epoch': 0.4}) +[2026-06-15 22:34:15,770][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.21399077773094177, 'train/grad_norm_pre_clip': 10.27531909942627, 'timing/data': 0.01155377784743905, 'timing/model': 1.5721207428723574, 'learning_rate/qwen_vl_interface': 9.235895282411029e-06, 'learning_rate/action_model': 9.235895282411029e-05, 'epoch': 0.4}) +[2026-06-15 22:34:22,064][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.23313401639461517, 'train/grad_norm_pre_clip': 10.668131828308105, 'timing/data': 0.010539841838181019, 'timing/model': 1.565838806796819, 'learning_rate/qwen_vl_interface': 9.2342379918952e-06, 'learning_rate/action_model': 9.2342379918952e-05, 'epoch': 0.4}) +[2026-06-15 22:34:28,533][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.09592338651418686, 'train/grad_norm_pre_clip': 4.671389102935791, 'timing/data': 0.01125401584431529, 'timing/model': 1.6465605678968132, 'learning_rate/qwen_vl_interface': 9.232579063609631e-06, 'learning_rate/action_model': 9.232579063609631e-05, 'epoch': 0.4}) +[2026-06-15 22:45:30,505][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.13138198852539062, 'train/grad_norm_pre_clip': 8.790584564208984, 'eval/loss': 0.07532093673944473, 'eval/flap_precision': 0.8693528693528694, 'eval/flap_recall': 0.9713506139154161, 'eval/flap_f1': 0.9175257731958764, 'eval/flap_rate_model': 0.2275, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9175257731958762, 'timing/data': 0.010711143258959055, 'timing/model': 1.5228396407328546, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 0.4}) +[2026-06-15 22:45:30,506][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1000_state +[2026-06-15 22:45:30,512][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 22:45:39,074][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1000_state/model.safetensors +[2026-06-15 22:45:59,001][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1000_state/optimizer.bin +[2026-06-15 22:45:59,001][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1000_state/sampler.bin +[2026-06-15 22:45:59,002][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-15 22:45:59,003][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-15 22:45:59,199][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-15 22:45:59,199][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 22:45:59,214][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 22:46:05,431][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.18861694633960724, 'train/grad_norm_pre_clip': 9.776796340942383, 'timing/data': 0.010689377784729004, 'timing/model': 1.5097715300507843, 'learning_rate/qwen_vl_interface': 9.229256296457635e-06, 'learning_rate/action_model': 9.229256296457634e-05, 'epoch': 0.4}) +[2026-06-15 22:46:11,644][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.25226202607154846, 'train/grad_norm_pre_clip': 18.07532501220703, 'timing/data': 0.010430408176034689, 'timing/model': 1.783575537148863, 'learning_rate/qwen_vl_interface': 9.22759245895707e-06, 'learning_rate/action_model': 9.227592458957071e-05, 'epoch': 0.41}) +[2026-06-15 22:46:18,016][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.35649409890174866, 'train/grad_norm_pre_clip': 18.9072208404541, 'timing/data': 0.009649632964283228, 'timing/model': 1.8104302249848843, 'learning_rate/qwen_vl_interface': 9.225926986418494e-06, 'learning_rate/action_model': 9.225926986418494e-05, 'epoch': 0.41}) +[2026-06-15 22:46:24,044][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.0789361447095871, 'train/grad_norm_pre_clip': 5.806077480316162, 'timing/data': 0.010836529079824686, 'timing/model': 1.5701924380846322, 'learning_rate/qwen_vl_interface': 9.224259879526519e-06, 'learning_rate/action_model': 9.224259879526519e-05, 'epoch': 0.41}) +[2026-06-15 22:46:30,338][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.13346664607524872, 'train/grad_norm_pre_clip': 10.487420082092285, 'timing/data': 0.010642118752002716, 'timing/model': 1.562842498999089, 'learning_rate/qwen_vl_interface': 9.222591138966427e-06, 'learning_rate/action_model': 9.222591138966426e-05, 'epoch': 0.41}) +[2026-06-15 22:46:36,653][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.10010392963886261, 'train/grad_norm_pre_clip': 10.616889953613281, 'timing/data': 0.0107115488499403, 'timing/model': 1.5577734727412462, 'learning_rate/qwen_vl_interface': 9.220920765424175e-06, 'learning_rate/action_model': 9.220920765424174e-05, 'epoch': 0.41}) +[2026-06-15 22:46:42,950][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.07428687810897827, 'train/grad_norm_pre_clip': 7.001244068145752, 'timing/data': 0.010517497546970844, 'timing/model': 1.5467724949121475, 'learning_rate/qwen_vl_interface': 9.219248759586392e-06, 'learning_rate/action_model': 9.219248759586391e-05, 'epoch': 0.41}) +[2026-06-15 22:46:49,250][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.16673430800437927, 'train/grad_norm_pre_clip': 9.214221000671387, 'timing/data': 0.011251571122556925, 'timing/model': 1.5828185249119997, 'learning_rate/qwen_vl_interface': 9.217575122140375e-06, 'learning_rate/action_model': 9.217575122140375e-05, 'epoch': 0.41}) +[2026-06-15 22:46:55,587][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.20922543108463287, 'train/grad_norm_pre_clip': 5.396793365478516, 'timing/data': 0.01131526194512844, 'timing/model': 1.797328329179436, 'learning_rate/qwen_vl_interface': 9.215899853774094e-06, 'learning_rate/action_model': 9.215899853774093e-05, 'epoch': 0.41}) +[2026-06-15 22:47:01,648][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.13549545407295227, 'train/grad_norm_pre_clip': 12.20091438293457, 'timing/data': 0.011199451982975006, 'timing/model': 1.5651181526482105, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.41}) +[2026-06-15 22:47:08,010][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.13207879662513733, 'train/grad_norm_pre_clip': 9.703607559204102, 'timing/data': 0.01121295103803277, 'timing/model': 1.621088061016053, 'learning_rate/qwen_vl_interface': 9.212544427035967e-06, 'learning_rate/action_model': 9.212544427035967e-05, 'epoch': 0.41}) +[2026-06-15 22:47:14,252][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.18771502375602722, 'train/grad_norm_pre_clip': 5.868302345275879, 'timing/data': 0.010806479025632143, 'timing/model': 1.5636805007234216, 'learning_rate/qwen_vl_interface': 9.21086427004341e-06, 'learning_rate/action_model': 9.210864270043411e-05, 'epoch': 0.41}) +[2026-06-15 22:47:20,655][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.015301929786801338, 'train/grad_norm_pre_clip': 1.0844799280166626, 'timing/data': 0.010847178753465414, 'timing/model': 1.5644048182293773, 'learning_rate/qwen_vl_interface': 9.209182484889168e-06, 'learning_rate/action_model': 9.209182484889167e-05, 'epoch': 0.41}) +[2026-06-15 22:47:26,952][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.17917832732200623, 'train/grad_norm_pre_clip': 15.146815299987793, 'timing/data': 0.010824969038367271, 'timing/model': 1.5623379577882588, 'learning_rate/qwen_vl_interface': 9.207499072264557e-06, 'learning_rate/action_model': 9.207499072264558e-05, 'epoch': 0.41}) +[2026-06-15 22:47:33,291][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.10720764845609665, 'train/grad_norm_pre_clip': 6.192152500152588, 'timing/data': 0.01056651771068573, 'timing/model': 1.786901003215462, 'learning_rate/qwen_vl_interface': 9.205814032861566e-06, 'learning_rate/action_model': 9.205814032861566e-05, 'epoch': 0.41}) +[2026-06-15 22:47:39,472][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.26355600357055664, 'train/grad_norm_pre_clip': 12.455751419067383, 'timing/data': 0.011175440158694983, 'timing/model': 1.6045145546086133, 'learning_rate/qwen_vl_interface': 9.204127367372849e-06, 'learning_rate/action_model': 9.204127367372849e-05, 'epoch': 0.41}) +[2026-06-15 22:47:45,757][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.1733713448047638, 'train/grad_norm_pre_clip': 10.156046867370605, 'timing/data': 0.010883858893066645, 'timing/model': 1.5640173279680312, 'learning_rate/qwen_vl_interface': 9.202439076491732e-06, 'learning_rate/action_model': 9.202439076491732e-05, 'epoch': 0.41}) +[2026-06-15 22:47:52,055][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.16609971225261688, 'train/grad_norm_pre_clip': 12.111109733581543, 'timing/data': 0.01122809061780572, 'timing/model': 1.5526182088069618, 'learning_rate/qwen_vl_interface': 9.200749160912208e-06, 'learning_rate/action_model': 9.200749160912208e-05, 'epoch': 0.41}) +[2026-06-15 22:47:58,432][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.04133031889796257, 'train/grad_norm_pre_clip': 3.9720895290374756, 'timing/data': 0.0111238700337708, 'timing/model': 1.5742826089262962, 'learning_rate/qwen_vl_interface': 9.199057621328937e-06, 'learning_rate/action_model': 9.199057621328935e-05, 'epoch': 0.41}) +[2026-06-15 22:48:04,654][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.23734214901924133, 'train/grad_norm_pre_clip': 15.617430686950684, 'timing/data': 0.012149255722761154, 'timing/model': 1.55827373592183, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 0.41}) +[2026-06-15 22:48:10,997][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.05208640173077583, 'train/grad_norm_pre_clip': 6.767524242401123, 'timing/data': 0.011007680092006922, 'timing/model': 1.8055561073124409, 'learning_rate/qwen_vl_interface': 9.19566967293313e-06, 'learning_rate/action_model': 9.19566967293313e-05, 'epoch': 0.41}) +[2026-06-15 22:48:17,057][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.23407621681690216, 'train/grad_norm_pre_clip': 16.979909896850586, 'timing/data': 0.011077359784394503, 'timing/model': 1.5614339909516275, 'learning_rate/qwen_vl_interface': 9.193973265513253e-06, 'learning_rate/action_model': 9.193973265513253e-05, 'epoch': 0.41}) +[2026-06-15 22:48:23,361][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.1368291825056076, 'train/grad_norm_pre_clip': 15.873318672180176, 'timing/data': 0.010806967969983816, 'timing/model': 1.565528026316315, 'learning_rate/qwen_vl_interface': 9.192275236874946e-06, 'learning_rate/action_model': 9.192275236874946e-05, 'epoch': 0.41}) +[2026-06-15 22:48:29,666][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.06116801127791405, 'train/grad_norm_pre_clip': 10.990894317626953, 'timing/data': 0.010785438120365143, 'timing/model': 1.5707021178677678, 'learning_rate/qwen_vl_interface': 9.190575587716199e-06, 'learning_rate/action_model': 9.190575587716199e-05, 'epoch': 0.41}) +[2026-06-15 22:48:35,956][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.41685646772384644, 'train/grad_norm_pre_clip': 14.615147590637207, 'timing/data': 0.010573347099125385, 'timing/model': 1.5593920410610735, 'learning_rate/qwen_vl_interface': 9.188874318735677e-06, 'learning_rate/action_model': 9.188874318735676e-05, 'epoch': 0.41}) +[2026-06-15 22:48:42,266][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.1603473722934723, 'train/grad_norm_pre_clip': 9.248005867004395, 'timing/data': 0.011286410968750715, 'timing/model': 1.5316133541055024, 'learning_rate/qwen_vl_interface': 9.187171430632707e-06, 'learning_rate/action_model': 9.187171430632707e-05, 'epoch': 0.41}) +[2026-06-15 22:48:48,593][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.2507646381855011, 'train/grad_norm_pre_clip': 14.742307662963867, 'timing/data': 0.010678918100893497, 'timing/model': 1.7879060092382133, 'learning_rate/qwen_vl_interface': 9.185466924107283e-06, 'learning_rate/action_model': 9.185466924107282e-05, 'epoch': 0.42}) +[2026-06-15 22:48:54,668][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.4294925034046173, 'train/grad_norm_pre_clip': 15.206056594848633, 'timing/data': 0.011444240808486938, 'timing/model': 1.5731122768484056, 'learning_rate/qwen_vl_interface': 9.18376079986006e-06, 'learning_rate/action_model': 9.183760799860061e-05, 'epoch': 0.42}) +[2026-06-15 22:49:00,966][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.10668080300092697, 'train/grad_norm_pre_clip': 12.964463233947754, 'timing/data': 0.011147779878228903, 'timing/model': 1.5646763946861029, 'learning_rate/qwen_vl_interface': 9.182053058592365e-06, 'learning_rate/action_model': 9.182053058592365e-05, 'epoch': 0.42}) +[2026-06-15 22:49:07,292][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.10435926169157028, 'train/grad_norm_pre_clip': 7.083316326141357, 'timing/data': 0.011298469733446836, 'timing/model': 1.549735279288143, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 0.42}) +[2026-06-15 22:49:13,561][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.1478043645620346, 'train/grad_norm_pre_clip': 6.9431257247924805, 'timing/data': 0.011034858878701925, 'timing/model': 1.5570061020553112, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.42}) +[2026-06-15 22:49:19,894][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.10786513239145279, 'train/grad_norm_pre_clip': 5.9349141120910645, 'timing/data': 0.011227699927985668, 'timing/model': 1.552281980868429, 'learning_rate/qwen_vl_interface': 9.176920139689638e-06, 'learning_rate/action_model': 9.176920139689637e-05, 'epoch': 0.42}) +[2026-06-15 22:49:26,201][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.23839472234249115, 'train/grad_norm_pre_clip': 9.302191734313965, 'timing/data': 0.010772848036140203, 'timing/model': 1.791670988779515, 'learning_rate/qwen_vl_interface': 9.175205937366568e-06, 'learning_rate/action_model': 9.175205937366569e-05, 'epoch': 0.42}) +[2026-06-15 22:49:32,272][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.10733073949813843, 'train/grad_norm_pre_clip': 3.9884254932403564, 'timing/data': 0.01189797418192029, 'timing/model': 1.5641801739111543, 'learning_rate/qwen_vl_interface': 9.173490121539608e-06, 'learning_rate/action_model': 9.173490121539608e-05, 'epoch': 0.42}) +[2026-06-15 22:49:38,570][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.06829067319631577, 'train/grad_norm_pre_clip': 4.243485450744629, 'timing/data': 0.010533566121011972, 'timing/model': 1.5669992938637733, 'learning_rate/qwen_vl_interface': 9.171772692914062e-06, 'learning_rate/action_model': 9.171772692914061e-05, 'epoch': 0.42}) +[2026-06-15 22:49:45,039][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.12974336743354797, 'train/grad_norm_pre_clip': 8.662046432495117, 'timing/data': 0.010000133886933327, 'timing/model': 1.5832055239006877, 'learning_rate/qwen_vl_interface': 9.1700536521959e-06, 'learning_rate/action_model': 9.170053652195899e-05, 'epoch': 0.42}) +[2026-06-15 22:49:51,265][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.1419793665409088, 'train/grad_norm_pre_clip': 6.468079566955566, 'timing/data': 0.011179229710251093, 'timing/model': 1.5681114690378308, 'learning_rate/qwen_vl_interface': 9.168333000091754e-06, 'learning_rate/action_model': 9.168333000091754e-05, 'epoch': 0.42}) +[2026-06-15 22:49:57,660][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.20465482771396637, 'train/grad_norm_pre_clip': 12.874711990356445, 'timing/data': 0.011035289149731398, 'timing/model': 1.5277732480317354, 'learning_rate/qwen_vl_interface': 9.16661073730892e-06, 'learning_rate/action_model': 9.166610737308919e-05, 'epoch': 0.42}) +[2026-06-15 22:50:04,004][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.074073426425457, 'train/grad_norm_pre_clip': 6.138857841491699, 'timing/data': 0.01193501427769661, 'timing/model': 1.8070002649910748, 'learning_rate/qwen_vl_interface': 9.164886864555355e-06, 'learning_rate/action_model': 9.164886864555355e-05, 'epoch': 0.42}) +[2026-06-15 22:50:10,163][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.2734919786453247, 'train/grad_norm_pre_clip': 13.240938186645508, 'timing/data': 0.011322641279548407, 'timing/model': 1.5620498899370432, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 0.42}) +[2026-06-15 22:50:16,453][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.08206843584775925, 'train/grad_norm_pre_clip': 6.101369857788086, 'timing/data': 0.010747416876256466, 'timing/model': 1.5582345151342452, 'learning_rate/qwen_vl_interface': 9.16143429197117e-06, 'learning_rate/action_model': 9.16143429197117e-05, 'epoch': 0.42}) +[2026-06-15 22:50:22,766][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.04122142121195793, 'train/grad_norm_pre_clip': 2.782283306121826, 'timing/data': 0.011171779129654169, 'timing/model': 1.570242963731289, 'learning_rate/qwen_vl_interface': 9.159705593559773e-06, 'learning_rate/action_model': 9.159705593559773e-05, 'epoch': 0.42}) +[2026-06-15 22:50:29,055][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.2585587203502655, 'train/grad_norm_pre_clip': 13.768172264099121, 'timing/data': 0.01090718898922205, 'timing/model': 1.5602179146371782, 'learning_rate/qwen_vl_interface': 9.157975288016089e-06, 'learning_rate/action_model': 9.157975288016088e-05, 'epoch': 0.42}) +[2026-06-15 22:50:35,596][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.06008019298315048, 'train/grad_norm_pre_clip': 7.411903381347656, 'timing/data': 0.013884124346077442, 'timing/model': 1.6297100097872317, 'learning_rate/qwen_vl_interface': 9.156243376051381e-06, 'learning_rate/action_model': 9.156243376051381e-05, 'epoch': 0.42}) +[2026-06-15 22:50:42,006][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.32473471760749817, 'train/grad_norm_pre_clip': 12.831634521484375, 'timing/data': 0.010693957097828388, 'timing/model': 1.7967016538605094, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.42}) +[2026-06-15 22:50:48,226][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.04891078919172287, 'train/grad_norm_pre_clip': 5.433910846710205, 'timing/data': 0.009787231683731079, 'timing/model': 1.6982052638195455, 'learning_rate/qwen_vl_interface': 9.152774735707248e-06, 'learning_rate/action_model': 9.152774735707247e-05, 'epoch': 0.42}) +[2026-06-15 22:50:54,271][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.3827303349971771, 'train/grad_norm_pre_clip': 19.136985778808594, 'timing/data': 0.010500164702534676, 'timing/model': 1.5431000641547143, 'learning_rate/qwen_vl_interface': 9.15103800875365e-06, 'learning_rate/action_model': 9.15103800875365e-05, 'epoch': 0.42}) +[2026-06-15 22:51:00,567][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.1263464391231537, 'train/grad_norm_pre_clip': 11.363264083862305, 'timing/data': 0.010779947973787785, 'timing/model': 1.56978306081146, 'learning_rate/qwen_vl_interface': 9.149299678230681e-06, 'learning_rate/action_model': 9.149299678230681e-05, 'epoch': 0.42}) +[2026-06-15 22:51:06,864][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.05052096024155617, 'train/grad_norm_pre_clip': 8.38298225402832, 'timing/data': 0.010622285772114992, 'timing/model': 1.5649825716391206, 'learning_rate/qwen_vl_interface': 9.147559744852905e-06, 'learning_rate/action_model': 9.147559744852905e-05, 'epoch': 0.42}) +[2026-06-15 22:51:13,159][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.0870557576417923, 'train/grad_norm_pre_clip': 9.706350326538086, 'timing/data': 0.01051308587193489, 'timing/model': 1.527929940726608, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 0.42}) +[2026-06-15 22:51:19,523][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.12137344479560852, 'train/grad_norm_pre_clip': 13.12855339050293, 'timing/data': 0.011756862048059702, 'timing/model': 1.8218512861058116, 'learning_rate/qwen_vl_interface': 9.14407507239447e-06, 'learning_rate/action_model': 9.14407507239447e-05, 'epoch': 0.43}) +[2026-06-15 22:51:25,831][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.1866321712732315, 'train/grad_norm_pre_clip': 11.354753494262695, 'timing/data': 0.010727507062256336, 'timing/model': 1.829665421973914, 'learning_rate/qwen_vl_interface': 9.142330334746227e-06, 'learning_rate/action_model': 9.142330334746227e-05, 'epoch': 0.43}) +[2026-06-15 22:51:32,170][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.6159324049949646, 'train/grad_norm_pre_clip': 33.42759323120117, 'timing/data': 0.010644005611538887, 'timing/model': 1.5657544978894293, 'learning_rate/qwen_vl_interface': 9.14058399710801e-06, 'learning_rate/action_model': 9.14058399710801e-05, 'epoch': 0.43}) +[2026-06-15 22:51:38,478][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.26759931445121765, 'train/grad_norm_pre_clip': 17.977508544921875, 'timing/data': 0.010885327123105526, 'timing/model': 1.5735827749595046, 'learning_rate/qwen_vl_interface': 9.138836060197673e-06, 'learning_rate/action_model': 9.138836060197672e-05, 'epoch': 0.43}) +[2026-06-15 22:51:44,769][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.10461927950382233, 'train/grad_norm_pre_clip': 8.926295280456543, 'timing/data': 0.012907329015433788, 'timing/model': 1.5618446380831301, 'learning_rate/qwen_vl_interface': 9.137086524733725e-06, 'learning_rate/action_model': 9.137086524733724e-05, 'epoch': 0.43}) +[2026-06-15 22:51:51,075][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.10392586886882782, 'train/grad_norm_pre_clip': 8.693906784057617, 'timing/data': 0.011330789886415005, 'timing/model': 1.5312084532342851, 'learning_rate/qwen_vl_interface': 9.135335391435336e-06, 'learning_rate/action_model': 9.135335391435335e-05, 'epoch': 0.43}) +[2026-06-15 22:51:57,399][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.07482089102268219, 'train/grad_norm_pre_clip': 4.13290548324585, 'timing/data': 0.01093403808772564, 'timing/model': 1.7944879606366158, 'learning_rate/qwen_vl_interface': 9.13358266102233e-06, 'learning_rate/action_model': 9.133582661022328e-05, 'epoch': 0.43}) +[2026-06-15 22:52:03,464][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.1373952478170395, 'train/grad_norm_pre_clip': 15.609967231750488, 'timing/data': 0.011783701833337545, 'timing/model': 1.5560440458357334, 'learning_rate/qwen_vl_interface': 9.131828334215185e-06, 'learning_rate/action_model': 9.131828334215185e-05, 'epoch': 0.43}) +[2026-06-15 22:52:09,763][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.024609293788671494, 'train/grad_norm_pre_clip': 1.8985857963562012, 'timing/data': 0.010771007277071476, 'timing/model': 1.5580395022407174, 'learning_rate/qwen_vl_interface': 9.130072411735043e-06, 'learning_rate/action_model': 9.130072411735043e-05, 'epoch': 0.43}) +[2026-06-15 22:52:16,066][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.3461643159389496, 'train/grad_norm_pre_clip': 8.148727416992188, 'timing/data': 0.010979767888784409, 'timing/model': 1.56540233688429, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 0.43}) +[2026-06-15 22:52:22,365][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.14091935753822327, 'train/grad_norm_pre_clip': 10.796758651733398, 'timing/data': 0.011274340096861124, 'timing/model': 1.5345551003701985, 'learning_rate/qwen_vl_interface': 9.126555782643592e-06, 'learning_rate/action_model': 9.126555782643591e-05, 'epoch': 0.43}) +[2026-06-15 22:52:28,665][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.18203429877758026, 'train/grad_norm_pre_clip': 14.569173812866211, 'timing/data': 0.010810646694153547, 'timing/model': 1.5909651727415621, 'learning_rate/qwen_vl_interface': 9.124795077477835e-06, 'learning_rate/action_model': 9.124795077477834e-05, 'epoch': 0.43}) +[2026-06-15 22:52:34,977][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.12262533605098724, 'train/grad_norm_pre_clip': 7.221132278442383, 'timing/data': 0.011285109911113977, 'timing/model': 1.7697488120757043, 'learning_rate/qwen_vl_interface': 9.123032779530186e-06, 'learning_rate/action_model': 9.123032779530187e-05, 'epoch': 0.43}) +[2026-06-15 22:52:41,071][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.11503732949495316, 'train/grad_norm_pre_clip': 7.270949840545654, 'timing/data': 0.011323720216751099, 'timing/model': 1.5667873290367424, 'learning_rate/qwen_vl_interface': 9.12126888952506e-06, 'learning_rate/action_model': 9.12126888952506e-05, 'epoch': 0.43}) +[2026-06-15 22:52:47,366][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.13060033321380615, 'train/grad_norm_pre_clip': 5.310980319976807, 'timing/data': 0.011394381057471037, 'timing/model': 1.5612994749099016, 'learning_rate/qwen_vl_interface': 9.119503408187521e-06, 'learning_rate/action_model': 9.119503408187522e-05, 'epoch': 0.43}) +[2026-06-15 22:52:53,671][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.1319291740655899, 'train/grad_norm_pre_clip': 5.558801651000977, 'timing/data': 0.01110686780884862, 'timing/model': 1.5702013089321554, 'learning_rate/qwen_vl_interface': 9.117736336243299e-06, 'learning_rate/action_model': 9.117736336243298e-05, 'epoch': 0.43}) +[2026-06-15 22:52:59,973][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.2352609634399414, 'train/grad_norm_pre_clip': 10.924601554870605, 'timing/data': 0.010876947082579136, 'timing/model': 1.5227470048703253, 'learning_rate/qwen_vl_interface': 9.115967674418765e-06, 'learning_rate/action_model': 9.115967674418765e-05, 'epoch': 0.43}) +[2026-06-15 22:53:06,372][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.050656355917453766, 'train/grad_norm_pre_clip': 5.319064617156982, 'timing/data': 0.016377725638449192, 'timing/model': 1.6000818689353764, 'learning_rate/qwen_vl_interface': 9.114197423440949e-06, 'learning_rate/action_model': 9.11419742344095e-05, 'epoch': 0.43}) +[2026-06-15 22:53:12,793][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.25301593542099, 'train/grad_norm_pre_clip': 15.783360481262207, 'timing/data': 0.011787491850554943, 'timing/model': 1.7978716059587896, 'learning_rate/qwen_vl_interface': 9.112425584037536e-06, 'learning_rate/action_model': 9.112425584037536e-05, 'epoch': 0.43}) +[2026-06-15 22:53:18,867][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.12258461117744446, 'train/grad_norm_pre_clip': 10.834647178649902, 'timing/data': 0.01171447103843093, 'timing/model': 1.5607675747014582, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 0.43}) +[2026-06-15 22:53:25,173][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.17164522409439087, 'train/grad_norm_pre_clip': 7.7158026695251465, 'timing/data': 0.010442154947668314, 'timing/model': 1.5655594058334827, 'learning_rate/qwen_vl_interface': 9.108877142867918e-06, 'learning_rate/action_model': 9.108877142867918e-05, 'epoch': 0.43}) +[2026-06-15 22:53:31,573][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.15002313256263733, 'train/grad_norm_pre_clip': 8.022618293762207, 'timing/data': 0.012023613322526217, 'timing/model': 1.571043351199478, 'learning_rate/qwen_vl_interface': 9.107100542560342e-06, 'learning_rate/action_model': 9.107100542560342e-05, 'epoch': 0.43}) +[2026-06-15 22:53:37,959][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.2359577715396881, 'train/grad_norm_pre_clip': 4.977614879608154, 'timing/data': 0.01202092319726944, 'timing/model': 1.5226925243623555, 'learning_rate/qwen_vl_interface': 9.105322356744429e-06, 'learning_rate/action_model': 9.105322356744429e-05, 'epoch': 0.43}) +[2026-06-15 22:53:44,487][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.26956096291542053, 'train/grad_norm_pre_clip': 7.043185234069824, 'timing/data': 0.011489150114357471, 'timing/model': 1.7885128357447684, 'learning_rate/qwen_vl_interface': 9.103542586151121e-06, 'learning_rate/action_model': 9.10354258615112e-05, 'epoch': 0.43}) +[2026-06-15 22:53:50,634][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.030191928148269653, 'train/grad_norm_pre_clip': 3.5395262241363525, 'timing/data': 0.01070834556594491, 'timing/model': 1.6310143661685288, 'learning_rate/qwen_vl_interface': 9.101761231512017e-06, 'learning_rate/action_model': 9.101761231512017e-05, 'epoch': 0.43}) +[2026-06-15 22:53:56,866][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.5854445695877075, 'train/grad_norm_pre_clip': 24.461057662963867, 'timing/data': 0.010335644241422415, 'timing/model': 1.557861769106239, 'learning_rate/qwen_vl_interface': 9.099978293559365e-06, 'learning_rate/action_model': 9.099978293559366e-05, 'epoch': 0.44}) +[2026-06-15 22:54:03,066][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.20040711760520935, 'train/grad_norm_pre_clip': 10.214534759521484, 'timing/data': 0.01039640512317419, 'timing/model': 1.5723298210650682, 'learning_rate/qwen_vl_interface': 9.098193773026065e-06, 'learning_rate/action_model': 9.098193773026064e-05, 'epoch': 0.44}) +[2026-06-15 22:54:09,362][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.3331161141395569, 'train/grad_norm_pre_clip': 20.254650115966797, 'timing/data': 0.010409425012767315, 'timing/model': 1.5630688788369298, 'learning_rate/qwen_vl_interface': 9.096407670645663e-06, 'learning_rate/action_model': 9.096407670645662e-05, 'epoch': 0.44}) +[2026-06-15 22:54:15,663][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.20507656037807465, 'train/grad_norm_pre_clip': 13.944112777709961, 'timing/data': 0.011025988031178713, 'timing/model': 1.5758934039622545, 'learning_rate/qwen_vl_interface': 9.094619987152357e-06, 'learning_rate/action_model': 9.094619987152358e-05, 'epoch': 0.44}) +[2026-06-15 22:54:22,007][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.03436814621090889, 'train/grad_norm_pre_clip': 2.6002326011657715, 'timing/data': 0.011507629882544279, 'timing/model': 1.801906144246459, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.44}) +[2026-06-15 22:54:28,120][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.05391770973801613, 'train/grad_norm_pre_clip': 2.4885783195495605, 'timing/data': 0.010657066013664007, 'timing/model': 1.619668086990714, 'learning_rate/qwen_vl_interface': 9.09103987976709e-06, 'learning_rate/action_model': 9.09103987976709e-05, 'epoch': 0.44}) +[2026-06-15 22:54:34,440][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.03806241601705551, 'train/grad_norm_pre_clip': 2.0341861248016357, 'timing/data': 0.011425069067627192, 'timing/model': 1.634692621883005, 'learning_rate/qwen_vl_interface': 9.089247457346774e-06, 'learning_rate/action_model': 9.089247457346773e-05, 'epoch': 0.44}) +[2026-06-15 22:54:40,663][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.26286396384239197, 'train/grad_norm_pre_clip': 13.023273468017578, 'timing/data': 0.010633016005158424, 'timing/model': 1.564802653156221, 'learning_rate/qwen_vl_interface': 9.087453456756849e-06, 'learning_rate/action_model': 9.087453456756849e-05, 'epoch': 0.44}) +[2026-06-15 22:54:47,066][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.08902014046907425, 'train/grad_norm_pre_clip': 5.941500663757324, 'timing/data': 0.011095398105680943, 'timing/model': 1.5618258439935744, 'learning_rate/qwen_vl_interface': 9.085657878734759e-06, 'learning_rate/action_model': 9.085657878734758e-05, 'epoch': 0.44}) +[2026-06-15 22:54:53,366][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.4338339865207672, 'train/grad_norm_pre_clip': 13.303251266479492, 'timing/data': 0.01081658573821187, 'timing/model': 1.5219131130725145, 'learning_rate/qwen_vl_interface': 9.083860724018603e-06, 'learning_rate/action_model': 9.083860724018602e-05, 'epoch': 0.44}) +[2026-06-15 22:54:59,712][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.10758179426193237, 'train/grad_norm_pre_clip': 7.602352619171143, 'timing/data': 0.01093139685690403, 'timing/model': 1.8157546450383961, 'learning_rate/qwen_vl_interface': 9.08206199334712e-06, 'learning_rate/action_model': 9.082061993347119e-05, 'epoch': 0.44}) +[2026-06-15 22:55:05,867][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.03167327493429184, 'train/grad_norm_pre_clip': 3.035303831100464, 'timing/data': 0.010928916744887829, 'timing/model': 1.5789266750216484, 'learning_rate/qwen_vl_interface': 9.0802616874597e-06, 'learning_rate/action_model': 9.0802616874597e-05, 'epoch': 0.44}) +[2026-06-15 22:55:12,160][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.2284521609544754, 'train/grad_norm_pre_clip': 4.822224140167236, 'timing/data': 0.010640976019203663, 'timing/model': 1.5580070433206856, 'learning_rate/qwen_vl_interface': 9.078459807096385e-06, 'learning_rate/action_model': 9.078459807096385e-05, 'epoch': 0.44}) +[2026-06-15 22:55:18,460][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.11684086173772812, 'train/grad_norm_pre_clip': 10.303736686706543, 'timing/data': 0.009625869803130627, 'timing/model': 1.5636173589155078, 'learning_rate/qwen_vl_interface': 9.076656352997854e-06, 'learning_rate/action_model': 9.076656352997853e-05, 'epoch': 0.44}) +[2026-06-15 22:55:24,758][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.040255725383758545, 'train/grad_norm_pre_clip': 9.558597564697266, 'timing/data': 0.009540488943457603, 'timing/model': 1.5282981130294502, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 0.44}) +[2026-06-15 22:55:31,065][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.20181606709957123, 'train/grad_norm_pre_clip': 7.343484401702881, 'timing/data': 0.0099246921017766, 'timing/model': 1.5617030044086277, 'learning_rate/qwen_vl_interface': 9.073044726561128e-06, 'learning_rate/action_model': 9.073044726561129e-05, 'epoch': 0.44}) +[2026-06-15 22:55:37,399][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.08630462735891342, 'train/grad_norm_pre_clip': 8.780117988586426, 'timing/data': 0.011371158994734287, 'timing/model': 1.7985368561930954, 'learning_rate/qwen_vl_interface': 9.071236555707536e-06, 'learning_rate/action_model': 9.071236555707537e-05, 'epoch': 0.44}) +[2026-06-15 22:55:43,560][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.2301524579524994, 'train/grad_norm_pre_clip': 16.197296142578125, 'timing/data': 0.011767780873924494, 'timing/model': 1.5529233627021313, 'learning_rate/qwen_vl_interface': 9.069426814087937e-06, 'learning_rate/action_model': 9.069426814087937e-05, 'epoch': 0.44}) +[2026-06-15 22:55:49,869][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.057671841233968735, 'train/grad_norm_pre_clip': 7.880616188049316, 'timing/data': 0.010680435225367546, 'timing/model': 1.572948934044689, 'learning_rate/qwen_vl_interface': 9.067615502446247e-06, 'learning_rate/action_model': 9.067615502446246e-05, 'epoch': 0.44}) +[2026-06-15 22:55:56,162][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.009451432153582573, 'train/grad_norm_pre_clip': 1.2249701023101807, 'timing/data': 0.011247779242694378, 'timing/model': 1.5650747898034751, 'learning_rate/qwen_vl_interface': 9.06580262152703e-06, 'learning_rate/action_model': 9.065802621527028e-05, 'epoch': 0.44}) +[2026-06-15 22:56:02,458][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.310269832611084, 'train/grad_norm_pre_clip': 26.40635108947754, 'timing/data': 0.010004532057791948, 'timing/model': 1.5576001489534974, 'learning_rate/qwen_vl_interface': 9.063988172075486e-06, 'learning_rate/action_model': 9.063988172075486e-05, 'epoch': 0.44}) +[2026-06-15 22:56:08,885][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.17129936814308167, 'train/grad_norm_pre_clip': 3.5034217834472656, 'timing/data': 0.011490109842270613, 'timing/model': 1.5884421369992197, 'learning_rate/qwen_vl_interface': 9.062172154837475e-06, 'learning_rate/action_model': 9.062172154837475e-05, 'epoch': 0.44}) +[2026-06-15 22:56:15,209][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.010443700477480888, 'train/grad_norm_pre_clip': 0.43338900804519653, 'timing/data': 0.011395418550819159, 'timing/model': 1.8084966489113867, 'learning_rate/qwen_vl_interface': 9.060354570559487e-06, 'learning_rate/action_model': 9.060354570559487e-05, 'epoch': 0.44}) +[2026-06-15 22:56:21,273][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.13933499157428741, 'train/grad_norm_pre_clip': 10.852575302124023, 'timing/data': 0.013555270154029131, 'timing/model': 1.5649938001297414, 'learning_rate/qwen_vl_interface': 9.058535419988667e-06, 'learning_rate/action_model': 9.058535419988667e-05, 'epoch': 0.44}) +[2026-06-15 23:07:19,099][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.17824654281139374, 'train/grad_norm_pre_clip': 14.758843421936035, 'eval/loss': 0.10692880302667618, 'eval/flap_precision': 0.9047619047619048, 'eval/flap_recall': 0.9072305593451568, 'eval/flap_f1': 0.9059945504087193, 'eval/flap_rate_model': 0.20416666666666666, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9059945504087193, 'timing/data': 0.010675115045160055, 'timing/model': 1.5615104702301323, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 0.44}) +[2026-06-15 23:07:19,100][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1100_state +[2026-06-15 23:07:19,106][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:07:27,781][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1100_state/model.safetensors +[2026-06-15 23:07:47,697][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1100_state/optimizer.bin +[2026-06-15 23:07:47,698][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1100_state/sampler.bin +[2026-06-15 23:07:47,698][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1100_state/sampler_1.bin +[2026-06-15 23:07:47,699][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1100_state/random_states_0.pkl +[2026-06-15 23:07:47,890][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-06-15 23:07:47,891][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:07:47,906][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:07:53,987][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.24341760575771332, 'train/grad_norm_pre_clip': 14.0646333694458, 'timing/data': 0.010955565143376589, 'timing/model': 1.6129260011948645, 'learning_rate/qwen_vl_interface': 9.0548924229603e-06, 'learning_rate/action_model': 9.054892422960301e-05, 'epoch': 0.45}) +[2026-06-15 23:08:00,399][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.39617836475372314, 'train/grad_norm_pre_clip': 12.817954063415527, 'timing/data': 0.011425747070461512, 'timing/model': 1.5919944983907044, 'learning_rate/qwen_vl_interface': 9.053068578000256e-06, 'learning_rate/action_model': 9.053068578000256e-05, 'epoch': 0.45}) +[2026-06-15 23:08:06,722][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.06262754648923874, 'train/grad_norm_pre_clip': 16.705921173095703, 'timing/data': 0.012112840078771114, 'timing/model': 1.6453314190730453, 'learning_rate/qwen_vl_interface': 9.051243169742372e-06, 'learning_rate/action_model': 9.051243169742372e-05, 'epoch': 0.45}) +[2026-06-15 23:08:12,842][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.4192531704902649, 'train/grad_norm_pre_clip': 48.060977935791016, 'timing/data': 0.0117041883058846, 'timing/model': 1.5648661609739065, 'learning_rate/qwen_vl_interface': 9.049416198937008e-06, 'learning_rate/action_model': 9.049416198937007e-05, 'epoch': 0.45}) +[2026-06-15 23:08:18,936][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.1260058581829071, 'train/grad_norm_pre_clip': 13.88043212890625, 'timing/data': 0.010748585220426321, 'timing/model': 1.4904614752158523, 'learning_rate/qwen_vl_interface': 9.047587666335159e-06, 'learning_rate/action_model': 9.047587666335159e-05, 'epoch': 0.45}) +[2026-06-15 23:08:25,195][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.041217703372240067, 'train/grad_norm_pre_clip': 11.919844627380371, 'timing/data': 0.010758403688669205, 'timing/model': 1.81454589497298, 'learning_rate/qwen_vl_interface': 9.045757572688469e-06, 'learning_rate/action_model': 9.04575757268847e-05, 'epoch': 0.45}) +[2026-06-15 23:08:31,254][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.09906808286905289, 'train/grad_norm_pre_clip': 13.825807571411133, 'timing/data': 0.011410348117351532, 'timing/model': 1.5499840429984033, 'learning_rate/qwen_vl_interface': 9.043925918749218e-06, 'learning_rate/action_model': 9.043925918749218e-05, 'epoch': 0.45}) +[2026-06-15 23:08:37,554][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.18321634829044342, 'train/grad_norm_pre_clip': 28.19695281982422, 'timing/data': 0.011711628641933203, 'timing/model': 1.5314627676270902, 'learning_rate/qwen_vl_interface': 9.042092705270333e-06, 'learning_rate/action_model': 9.042092705270333e-05, 'epoch': 0.45}) +[2026-06-15 23:08:43,978][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.15834538638591766, 'train/grad_norm_pre_clip': 8.991103172302246, 'timing/data': 0.011953730136156082, 'timing/model': 1.5867046378552914, 'learning_rate/qwen_vl_interface': 9.040257933005375e-06, 'learning_rate/action_model': 9.040257933005374e-05, 'epoch': 0.45}) +[2026-06-15 23:08:50,348][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.13919568061828613, 'train/grad_norm_pre_clip': 9.365737915039062, 'timing/data': 0.01163756800815463, 'timing/model': 1.5796803808771074, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 0.45}) +[2026-06-15 23:08:56,718][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.14807327091693878, 'train/grad_norm_pre_clip': 40.923973083496094, 'timing/data': 0.011716938111931086, 'timing/model': 1.7951478599570692, 'learning_rate/qwen_vl_interface': 9.036583715134709e-06, 'learning_rate/action_model': 9.036583715134708e-05, 'epoch': 0.45}) +[2026-06-15 23:09:02,993][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.2519645690917969, 'train/grad_norm_pre_clip': 26.253843307495117, 'timing/data': 0.010951715987175703, 'timing/model': 1.793854006100446, 'learning_rate/qwen_vl_interface': 9.034744271039332e-06, 'learning_rate/action_model': 9.034744271039332e-05, 'epoch': 0.45}) +[2026-06-15 23:09:09,067][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.1207713857293129, 'train/grad_norm_pre_clip': 9.123546600341797, 'timing/data': 0.011720808688551188, 'timing/model': 1.5532292728312314, 'learning_rate/qwen_vl_interface': 9.03290327117855e-06, 'learning_rate/action_model': 9.032903271178549e-05, 'epoch': 0.45}) +[2026-06-15 23:09:15,359][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.10892999172210693, 'train/grad_norm_pre_clip': 3.7850968837738037, 'timing/data': 0.010345393791794777, 'timing/model': 1.5771798831410706, 'learning_rate/qwen_vl_interface': 9.031060716309122e-06, 'learning_rate/action_model': 9.031060716309121e-05, 'epoch': 0.45}) +[2026-06-15 23:09:21,755][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.3297662138938904, 'train/grad_norm_pre_clip': 25.118446350097656, 'timing/data': 0.010875626001507044, 'timing/model': 1.5315146916545928, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.45}) +[2026-06-15 23:09:28,052][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.06896580755710602, 'train/grad_norm_pre_clip': 10.255820274353027, 'timing/data': 0.011442468035966158, 'timing/model': 1.530486358795315, 'learning_rate/qwen_vl_interface': 9.027370944574604e-06, 'learning_rate/action_model': 9.027370944574604e-05, 'epoch': 0.45}) +[2026-06-15 23:09:34,395][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.3292474150657654, 'train/grad_norm_pre_clip': 10.202305793762207, 'timing/data': 0.01144150784239173, 'timing/model': 1.5616826303303242, 'learning_rate/qwen_vl_interface': 9.025523729226236e-06, 'learning_rate/action_model': 9.025523729226236e-05, 'epoch': 0.45}) +[2026-06-15 23:09:40,657][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.4540805518627167, 'train/grad_norm_pre_clip': 22.1690731048584, 'timing/data': 0.011285077780485153, 'timing/model': 1.7619875632226467, 'learning_rate/qwen_vl_interface': 9.023674961902677e-06, 'learning_rate/action_model': 9.023674961902676e-05, 'epoch': 0.45}) +[2026-06-15 23:09:46,756][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.31015661358833313, 'train/grad_norm_pre_clip': 23.936174392700195, 'timing/data': 0.011468757875263691, 'timing/model': 1.55590614490211, 'learning_rate/qwen_vl_interface': 9.021824643363883e-06, 'learning_rate/action_model': 9.021824643363883e-05, 'epoch': 0.45}) +[2026-06-15 23:09:53,063][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.1088172197341919, 'train/grad_norm_pre_clip': 7.247957706451416, 'timing/data': 0.010820876341313124, 'timing/model': 1.5665982188656926, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 0.45}) +[2026-06-15 23:09:59,464][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.1760234385728836, 'train/grad_norm_pre_clip': 12.096226692199707, 'timing/data': 0.011671789921820164, 'timing/model': 1.5977138811722398, 'learning_rate/qwen_vl_interface': 9.018119355683623e-06, 'learning_rate/action_model': 9.018119355683623e-05, 'epoch': 0.45}) +[2026-06-15 23:10:05,761][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.03908747434616089, 'train/grad_norm_pre_clip': 3.0408287048339844, 'timing/data': 0.010988986119627953, 'timing/model': 1.5319790029898286, 'learning_rate/qwen_vl_interface': 9.016264388065258e-06, 'learning_rate/action_model': 9.016264388065258e-05, 'epoch': 0.45}) +[2026-06-15 23:10:12,091][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.06142263859510422, 'train/grad_norm_pre_clip': 5.303992748260498, 'timing/data': 0.010716574732214212, 'timing/model': 1.7597887488082051, 'learning_rate/qwen_vl_interface': 9.014407872277869e-06, 'learning_rate/action_model': 9.014407872277869e-05, 'epoch': 0.45}) +[2026-06-15 23:10:18,154][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.2087559551000595, 'train/grad_norm_pre_clip': 7.654935836791992, 'timing/data': 0.010782796423882246, 'timing/model': 1.5564259849488735, 'learning_rate/qwen_vl_interface': 9.012549809084595e-06, 'learning_rate/action_model': 9.012549809084595e-05, 'epoch': 0.45}) +[2026-06-15 23:10:24,456][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.18731920421123505, 'train/grad_norm_pre_clip': 3.8075332641601562, 'timing/data': 0.010849895887076855, 'timing/model': 1.5618073767982423, 'learning_rate/qwen_vl_interface': 9.01069019924922e-06, 'learning_rate/action_model': 9.010690199249219e-05, 'epoch': 0.45}) +[2026-06-15 23:10:30,759][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.3036334812641144, 'train/grad_norm_pre_clip': 12.274744033813477, 'timing/data': 0.010316373780369759, 'timing/model': 1.563349261879921, 'learning_rate/qwen_vl_interface': 9.008829043536157e-06, 'learning_rate/action_model': 9.008829043536156e-05, 'epoch': 0.46}) +[2026-06-15 23:10:37,086][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.5621840357780457, 'train/grad_norm_pre_clip': 20.008586883544922, 'timing/data': 0.011160696856677532, 'timing/model': 1.5692668668925762, 'learning_rate/qwen_vl_interface': 9.006966342710458e-06, 'learning_rate/action_model': 9.006966342710457e-05, 'epoch': 0.46}) +[2026-06-15 23:10:43,361][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.462210088968277, 'train/grad_norm_pre_clip': 15.139694213867188, 'timing/data': 0.011085767298936844, 'timing/model': 1.562766418326646, 'learning_rate/qwen_vl_interface': 9.005102097537809e-06, 'learning_rate/action_model': 9.005102097537809e-05, 'epoch': 0.46}) +[2026-06-15 23:10:49,797][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.20058207213878632, 'train/grad_norm_pre_clip': 6.635060787200928, 'timing/data': 0.011007026769220829, 'timing/model': 1.8121107541956007, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.46}) +[2026-06-15 23:10:55,862][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.3255499601364136, 'train/grad_norm_pre_clip': 7.8576788902282715, 'timing/data': 0.011550078634172678, 'timing/model': 1.5619762828573585, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 0.46}) +[2026-06-15 23:11:02,167][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.054231882095336914, 'train/grad_norm_pre_clip': 2.521245002746582, 'timing/data': 0.011158637702465057, 'timing/model': 1.5729119582101703, 'learning_rate/qwen_vl_interface': 8.999500103604538e-06, 'learning_rate/action_model': 8.999500103604536e-05, 'epoch': 0.46}) +[2026-06-15 23:11:08,456][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.20404373109340668, 'train/grad_norm_pre_clip': 9.15181827545166, 'timing/data': 0.010295764077454805, 'timing/model': 1.5602978821843863, 'learning_rate/qwen_vl_interface': 8.997629688713639e-06, 'learning_rate/action_model': 8.997629688713638e-05, 'epoch': 0.46}) +[2026-06-15 23:11:14,952][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.22239653766155243, 'train/grad_norm_pre_clip': 8.856595993041992, 'timing/data': 0.010958326980471611, 'timing/model': 1.5916285659186542, 'learning_rate/qwen_vl_interface': 8.995757733313733e-06, 'learning_rate/action_model': 8.995757733313732e-05, 'epoch': 0.46}) +[2026-06-15 23:11:21,300][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.06589508801698685, 'train/grad_norm_pre_clip': 4.971689224243164, 'timing/data': 0.010288164019584656, 'timing/model': 1.5581779698841274, 'learning_rate/qwen_vl_interface': 8.993884238174314e-06, 'learning_rate/action_model': 8.993884238174314e-05, 'epoch': 0.46}) +[2026-06-15 23:11:27,615][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.1231437623500824, 'train/grad_norm_pre_clip': 6.5958638191223145, 'timing/data': 0.011717029847204685, 'timing/model': 1.810290405061096, 'learning_rate/qwen_vl_interface': 8.992009204065505e-06, 'learning_rate/action_model': 8.992009204065505e-05, 'epoch': 0.46}) +[2026-06-15 23:11:33,666][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.08702167868614197, 'train/grad_norm_pre_clip': 4.706696510314941, 'timing/data': 0.010786005761474371, 'timing/model': 1.560970098245889, 'learning_rate/qwen_vl_interface': 8.990132631758058e-06, 'learning_rate/action_model': 8.990132631758059e-05, 'epoch': 0.46}) +[2026-06-15 23:11:40,039][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.3526695966720581, 'train/grad_norm_pre_clip': 10.182069778442383, 'timing/data': 0.010898767039179802, 'timing/model': 1.6220680503174663, 'learning_rate/qwen_vl_interface': 8.988254522023364e-06, 'learning_rate/action_model': 8.988254522023363e-05, 'epoch': 0.46}) +[2026-06-15 23:11:46,361][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.27304205298423767, 'train/grad_norm_pre_clip': 10.006436347961426, 'timing/data': 0.010696906130760908, 'timing/model': 1.5654898243956268, 'learning_rate/qwen_vl_interface': 8.986374875633441e-06, 'learning_rate/action_model': 8.98637487563344e-05, 'epoch': 0.46}) +[2026-06-15 23:11:52,659][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.09205809980630875, 'train/grad_norm_pre_clip': 5.137044429779053, 'timing/data': 0.011471550446003675, 'timing/model': 1.528217000886798, 'learning_rate/qwen_vl_interface': 8.984493693360944e-06, 'learning_rate/action_model': 8.984493693360943e-05, 'epoch': 0.46}) +[2026-06-15 23:11:58,962][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.15206581354141235, 'train/grad_norm_pre_clip': 6.572925090789795, 'timing/data': 0.011053048074245453, 'timing/model': 1.5457854131236672, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 0.46}) +[2026-06-15 23:12:05,303][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.028961673378944397, 'train/grad_norm_pre_clip': 1.187117576599121, 'timing/data': 0.010590245947241783, 'timing/model': 1.7964638890698552, 'learning_rate/qwen_vl_interface': 8.98072672426198e-06, 'learning_rate/action_model': 8.980726724261981e-05, 'epoch': 0.46}) +[2026-06-15 23:12:11,364][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.28580838441848755, 'train/grad_norm_pre_clip': 8.635567665100098, 'timing/data': 0.011181998997926712, 'timing/model': 1.5578027595765889, 'learning_rate/qwen_vl_interface': 8.978840938983977e-06, 'learning_rate/action_model': 8.978840938983977e-05, 'epoch': 0.46}) +[2026-06-15 23:12:17,669][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.0880672037601471, 'train/grad_norm_pre_clip': 8.164085388183594, 'timing/data': 0.011453850194811821, 'timing/model': 1.560618579853326, 'learning_rate/qwen_vl_interface': 8.976953620920313e-06, 'learning_rate/action_model': 8.976953620920312e-05, 'epoch': 0.46}) +[2026-06-15 23:12:24,084][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.11970149725675583, 'train/grad_norm_pre_clip': 5.948790550231934, 'timing/data': 0.011372540146112442, 'timing/model': 1.5945521052926779, 'learning_rate/qwen_vl_interface': 8.975064770846794e-06, 'learning_rate/action_model': 8.975064770846794e-05, 'epoch': 0.46}) +[2026-06-15 23:12:30,371][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.37621942162513733, 'train/grad_norm_pre_clip': 26.110124588012695, 'timing/data': 0.011799421161413193, 'timing/model': 1.5677605462260544, 'learning_rate/qwen_vl_interface': 8.973174389539856e-06, 'learning_rate/action_model': 8.973174389539855e-05, 'epoch': 0.46}) +[2026-06-15 23:12:36,741][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.06326214969158173, 'train/grad_norm_pre_clip': 4.38734245300293, 'timing/data': 0.01128458883613348, 'timing/model': 1.5889969961717725, 'learning_rate/qwen_vl_interface': 8.971282477776563e-06, 'learning_rate/action_model': 8.971282477776563e-05, 'epoch': 0.46}) +[2026-06-15 23:12:43,107][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.04909249022603035, 'train/grad_norm_pre_clip': 3.6615684032440186, 'timing/data': 0.010983448009938002, 'timing/model': 1.8001540559343994, 'learning_rate/qwen_vl_interface': 8.969389036334609e-06, 'learning_rate/action_model': 8.969389036334609e-05, 'epoch': 0.46}) +[2026-06-15 23:12:49,172][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.09869710355997086, 'train/grad_norm_pre_clip': 10.77419662475586, 'timing/data': 0.011621191166341305, 'timing/model': 1.5669367429800332, 'learning_rate/qwen_vl_interface': 8.967494065992315e-06, 'learning_rate/action_model': 8.967494065992315e-05, 'epoch': 0.46}) +[2026-06-15 23:12:55,470][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.10585443675518036, 'train/grad_norm_pre_clip': 7.243593692779541, 'timing/data': 0.011111578904092312, 'timing/model': 1.5618744282983243, 'learning_rate/qwen_vl_interface': 8.965597567528634e-06, 'learning_rate/action_model': 8.965597567528633e-05, 'epoch': 0.46}) +[2026-06-15 23:13:01,763][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.16034604609012604, 'train/grad_norm_pre_clip': 16.13884162902832, 'timing/data': 0.011161879170686007, 'timing/model': 1.5624758466146886, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.47}) +[2026-06-15 23:13:08,087][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.12189840525388718, 'train/grad_norm_pre_clip': 7.526144981384277, 'timing/data': 0.011354919988662004, 'timing/model': 1.5690839029848576, 'learning_rate/qwen_vl_interface': 8.961799989356048e-06, 'learning_rate/action_model': 8.961799989356048e-05, 'epoch': 0.47}) +[2026-06-15 23:13:14,463][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.014707366935908794, 'train/grad_norm_pre_clip': 1.6959068775177002, 'timing/data': 0.012717606034129858, 'timing/model': 1.6067817960865796, 'learning_rate/qwen_vl_interface': 8.959898911208185e-06, 'learning_rate/action_model': 8.959898911208185e-05, 'epoch': 0.47}) +[2026-06-15 23:13:20,802][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.18124458193778992, 'train/grad_norm_pre_clip': 6.549860000610352, 'timing/data': 0.011852192226797342, 'timing/model': 1.7931379908695817, 'learning_rate/qwen_vl_interface': 8.957996308061015e-06, 'learning_rate/action_model': 8.957996308061014e-05, 'epoch': 0.47}) +[2026-06-15 23:13:26,865][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.2465425282716751, 'train/grad_norm_pre_clip': 15.312884330749512, 'timing/data': 0.011639100965112448, 'timing/model': 1.5612775781191885, 'learning_rate/qwen_vl_interface': 8.956092180696626e-06, 'learning_rate/action_model': 8.956092180696625e-05, 'epoch': 0.47}) +[2026-06-15 23:13:33,265][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.0958760604262352, 'train/grad_norm_pre_clip': 7.080576419830322, 'timing/data': 0.01139354007318616, 'timing/model': 1.5614552558399737, 'learning_rate/qwen_vl_interface': 8.95418652989773e-06, 'learning_rate/action_model': 8.95418652989773e-05, 'epoch': 0.47}) +[2026-06-15 23:13:39,632][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.04962775856256485, 'train/grad_norm_pre_clip': 14.890264511108398, 'timing/data': 0.010978087782859802, 'timing/model': 1.6322564738802612, 'learning_rate/qwen_vl_interface': 8.952279356447675e-06, 'learning_rate/action_model': 8.952279356447675e-05, 'epoch': 0.47}) +[2026-06-15 23:13:45,866][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.07414697855710983, 'train/grad_norm_pre_clip': 10.696761131286621, 'timing/data': 0.011840101797133684, 'timing/model': 1.557227490004152, 'learning_rate/qwen_vl_interface': 8.950370661130421e-06, 'learning_rate/action_model': 8.950370661130422e-05, 'epoch': 0.47}) +[2026-06-15 23:13:52,239][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.21897120773792267, 'train/grad_norm_pre_clip': 17.621530532836914, 'timing/data': 0.01117386994883418, 'timing/model': 1.5832282500341535, 'learning_rate/qwen_vl_interface': 8.948460444730566e-06, 'learning_rate/action_model': 8.948460444730566e-05, 'epoch': 0.47}) +[2026-06-15 23:13:58,596][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.08431988954544067, 'train/grad_norm_pre_clip': 7.603888511657715, 'timing/data': 0.011245239991694689, 'timing/model': 1.7782897190190852, 'learning_rate/qwen_vl_interface': 8.946548708033324e-06, 'learning_rate/action_model': 8.946548708033323e-05, 'epoch': 0.47}) +[2026-06-15 23:14:04,672][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.24125835299491882, 'train/grad_norm_pre_clip': 16.32251739501953, 'timing/data': 0.011808551847934723, 'timing/model': 1.559818428941071, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 0.47}) +[2026-06-15 23:14:10,967][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.20058998465538025, 'train/grad_norm_pre_clip': 11.496504783630371, 'timing/data': 0.011676801834255457, 'timing/model': 1.5576181248761714, 'learning_rate/qwen_vl_interface': 8.942720676890677e-06, 'learning_rate/action_model': 8.942720676890677e-05, 'epoch': 0.47}) +[2026-06-15 23:14:17,272][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.5821611881256104, 'train/grad_norm_pre_clip': 21.100183486938477, 'timing/data': 0.011303049977868795, 'timing/model': 1.5650211842730641, 'learning_rate/qwen_vl_interface': 8.940804384018833e-06, 'learning_rate/action_model': 8.940804384018832e-05, 'epoch': 0.47}) +[2026-06-15 23:14:23,664][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.17886197566986084, 'train/grad_norm_pre_clip': 19.827228546142578, 'timing/data': 0.010916037950664759, 'timing/model': 1.5531727569177747, 'learning_rate/qwen_vl_interface': 8.93888657399672e-06, 'learning_rate/action_model': 8.93888657399672e-05, 'epoch': 0.47}) +[2026-06-15 23:14:30,013][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.3220233917236328, 'train/grad_norm_pre_clip': 20.227336883544922, 'timing/data': 0.011600971221923828, 'timing/model': 1.5641018617898226, 'learning_rate/qwen_vl_interface': 8.936967247612678e-06, 'learning_rate/action_model': 8.936967247612678e-05, 'epoch': 0.47}) +[2026-06-15 23:14:36,311][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.07255267351865768, 'train/grad_norm_pre_clip': 9.104450225830078, 'timing/data': 0.0108025879599154, 'timing/model': 1.8070186860859394, 'learning_rate/qwen_vl_interface': 8.93504640565567e-06, 'learning_rate/action_model': 8.93504640565567e-05, 'epoch': 0.47}) +[2026-06-15 23:14:42,524][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.18749640882015228, 'train/grad_norm_pre_clip': 7.24107027053833, 'timing/data': 0.01163695240393281, 'timing/model': 1.7136000371538103, 'learning_rate/qwen_vl_interface': 8.933124048915278e-06, 'learning_rate/action_model': 8.933124048915278e-05, 'epoch': 0.47}) +[2026-06-15 23:14:48,563][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.19436420500278473, 'train/grad_norm_pre_clip': 9.719663619995117, 'timing/data': 0.011405340861529112, 'timing/model': 1.5553514189086854, 'learning_rate/qwen_vl_interface': 8.931200178181715e-06, 'learning_rate/action_model': 8.931200178181715e-05, 'epoch': 0.47}) +[2026-06-15 23:14:54,858][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.1264703869819641, 'train/grad_norm_pre_clip': 14.417813301086426, 'timing/data': 0.01034193579107523, 'timing/model': 1.5603811270557344, 'learning_rate/qwen_vl_interface': 8.929274794245812e-06, 'learning_rate/action_model': 8.929274794245812e-05, 'epoch': 0.47}) +[2026-06-15 23:15:01,166][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.2896701395511627, 'train/grad_norm_pre_clip': 16.513071060180664, 'timing/data': 0.011702412273734808, 'timing/model': 1.5622873110696673, 'learning_rate/qwen_vl_interface': 8.927347897899018e-06, 'learning_rate/action_model': 8.927347897899016e-05, 'epoch': 0.47}) +[2026-06-15 23:15:07,456][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.0643124207854271, 'train/grad_norm_pre_clip': 4.785247325897217, 'timing/data': 0.011098168790340424, 'timing/model': 1.5157972588203847, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 0.47}) +[2026-06-15 23:15:13,790][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.3640206456184387, 'train/grad_norm_pre_clip': 12.160221099853516, 'timing/data': 0.011260609608143568, 'timing/model': 1.7929750112816691, 'learning_rate/qwen_vl_interface': 8.923489571141679e-06, 'learning_rate/action_model': 8.923489571141679e-05, 'epoch': 0.47}) +[2026-06-15 23:15:20,096][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.2101975381374359, 'train/grad_norm_pre_clip': 6.8427019119262695, 'timing/data': 0.01087418757379055, 'timing/model': 1.7965478831902146, 'learning_rate/qwen_vl_interface': 8.921558142317147e-06, 'learning_rate/action_model': 8.921558142317147e-05, 'epoch': 0.47}) +[2026-06-15 23:15:26,268][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.2034655511379242, 'train/grad_norm_pre_clip': 8.688155174255371, 'timing/data': 0.01105543877929449, 'timing/model': 1.5599219207651913, 'learning_rate/qwen_vl_interface': 8.91962520425375e-06, 'learning_rate/action_model': 8.91962520425375e-05, 'epoch': 0.47}) +[2026-06-15 23:15:32,572][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.20907892286777496, 'train/grad_norm_pre_clip': 8.913575172424316, 'timing/data': 0.01172558218240738, 'timing/model': 1.5583011587150395, 'learning_rate/qwen_vl_interface': 8.917690757746044e-06, 'learning_rate/action_model': 8.917690757746044e-05, 'epoch': 0.47}) +[2026-06-15 23:15:38,868][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.1578671783208847, 'train/grad_norm_pre_clip': 4.186645030975342, 'timing/data': 0.011410271283239126, 'timing/model': 1.5221939529292285, 'learning_rate/qwen_vl_interface': 8.915754803589208e-06, 'learning_rate/action_model': 8.915754803589209e-05, 'epoch': 0.48}) +[2026-06-15 23:15:45,272][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.1871476024389267, 'train/grad_norm_pre_clip': 4.427980422973633, 'timing/data': 0.011268380098044872, 'timing/model': 1.6075892699882388, 'learning_rate/qwen_vl_interface': 8.913817342579042e-06, 'learning_rate/action_model': 8.913817342579042e-05, 'epoch': 0.48}) +[2026-06-15 23:15:51,605][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.13941490650177002, 'train/grad_norm_pre_clip': 3.6554360389709473, 'timing/data': 0.01116515975445509, 'timing/model': 1.7950609852559865, 'learning_rate/qwen_vl_interface': 8.91187837551196e-06, 'learning_rate/action_model': 8.911878375511961e-05, 'epoch': 0.48}) +[2026-06-15 23:15:57,672][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.19411371648311615, 'train/grad_norm_pre_clip': 7.245616912841797, 'timing/data': 0.011309770867228508, 'timing/model': 1.5601489269174635, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.48}) +[2026-06-15 23:16:03,969][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.1853981614112854, 'train/grad_norm_pre_clip': 5.220479965209961, 'timing/data': 0.011399050708860159, 'timing/model': 1.5513607319444418, 'learning_rate/qwen_vl_interface': 8.907995926395814e-06, 'learning_rate/action_model': 8.907995926395815e-05, 'epoch': 0.48}) +[2026-06-15 23:16:10,302][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.14379079639911652, 'train/grad_norm_pre_clip': 5.86427116394043, 'timing/data': 0.01108550000935793, 'timing/model': 1.5992371020838618, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 0.48}) +[2026-06-15 23:16:16,685][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.11275891214609146, 'train/grad_norm_pre_clip': 5.686821937561035, 'timing/data': 0.014397585298866034, 'timing/model': 1.5996777480468154, 'learning_rate/qwen_vl_interface': 8.904107462624484e-06, 'learning_rate/action_model': 8.904107462624483e-05, 'epoch': 0.48}) +[2026-06-15 23:16:22,966][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.04526429623365402, 'train/grad_norm_pre_clip': 3.781137704849243, 'timing/data': 0.010479107033461332, 'timing/model': 1.580242226831615, 'learning_rate/qwen_vl_interface': 8.902160977240739e-06, 'learning_rate/action_model': 8.902160977240739e-05, 'epoch': 0.48}) +[2026-06-15 23:16:29,247][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.08606796711683273, 'train/grad_norm_pre_clip': 7.411816120147705, 'timing/data': 0.01101903012022376, 'timing/model': 1.748277676757425, 'learning_rate/qwen_vl_interface': 8.90021299059157e-06, 'learning_rate/action_model': 8.90021299059157e-05, 'epoch': 0.48}) +[2026-06-15 23:16:35,384][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.07978718727827072, 'train/grad_norm_pre_clip': 6.611082553863525, 'timing/data': 0.010907859075814486, 'timing/model': 1.5716241272166371, 'learning_rate/qwen_vl_interface': 8.898263503477722e-06, 'learning_rate/action_model': 8.898263503477721e-05, 'epoch': 0.48}) +[2026-06-15 23:16:41,766][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.33939412236213684, 'train/grad_norm_pre_clip': 21.956279754638672, 'timing/data': 0.013085679151117802, 'timing/model': 1.5604430800303817, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.48}) +[2026-06-15 23:16:48,073][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.056950464844703674, 'train/grad_norm_pre_clip': 8.5757474899292, 'timing/data': 0.01186456298455596, 'timing/model': 1.5672432049177587, 'learning_rate/qwen_vl_interface': 8.894360031062043e-06, 'learning_rate/action_model': 8.894360031062043e-05, 'epoch': 0.48}) +[2026-06-15 23:16:54,663][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.32997897267341614, 'train/grad_norm_pre_clip': 19.69253921508789, 'timing/data': 0.010995198972523212, 'timing/model': 1.5323223429732025, 'learning_rate/qwen_vl_interface': 8.892406047364783e-06, 'learning_rate/action_model': 8.892406047364784e-05, 'epoch': 0.48}) +[2026-06-15 23:17:01,053][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.03173546493053436, 'train/grad_norm_pre_clip': 5.277740955352783, 'timing/data': 0.012390105985105038, 'timing/model': 1.603116326034069, 'learning_rate/qwen_vl_interface': 8.890450566411984e-06, 'learning_rate/action_model': 8.890450566411984e-05, 'epoch': 0.48}) +[2026-06-15 23:17:07,372][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.32473641633987427, 'train/grad_norm_pre_clip': 13.94378662109375, 'timing/data': 0.010838169138878584, 'timing/model': 1.7671775012277067, 'learning_rate/qwen_vl_interface': 8.888493589007467e-06, 'learning_rate/action_model': 8.888493589007468e-05, 'epoch': 0.48}) +[2026-06-15 23:17:13,459][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.09130155295133591, 'train/grad_norm_pre_clip': 19.30150032043457, 'timing/data': 0.010126056149601936, 'timing/model': 1.5562542132101953, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 0.48}) +[2026-06-15 23:17:19,766][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.21391797065734863, 'train/grad_norm_pre_clip': 14.767003059387207, 'timing/data': 0.011379571165889502, 'timing/model': 1.5528087518177927, 'learning_rate/qwen_vl_interface': 8.884575148061662e-06, 'learning_rate/action_model': 8.88457514806166e-05, 'epoch': 0.48}) +[2026-06-15 23:17:26,068][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.2545213997364044, 'train/grad_norm_pre_clip': 9.4788179397583, 'timing/data': 0.01128537068143487, 'timing/model': 1.5522890430875123, 'learning_rate/qwen_vl_interface': 8.882613686131095e-06, 'learning_rate/action_model': 8.882613686131093e-05, 'epoch': 0.48}) +[2026-06-15 23:17:32,451][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.32594478130340576, 'train/grad_norm_pre_clip': 9.737054824829102, 'timing/data': 0.010889698751270771, 'timing/model': 1.5990792592056096, 'learning_rate/qwen_vl_interface': 8.880650730970254e-06, 'learning_rate/action_model': 8.880650730970254e-05, 'epoch': 0.48}) +[2026-06-15 23:17:38,805][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.21350663900375366, 'train/grad_norm_pre_clip': 10.302633285522461, 'timing/data': 0.011294040828943253, 'timing/model': 1.7929741609841585, 'learning_rate/qwen_vl_interface': 8.878686283386042e-06, 'learning_rate/action_model': 8.878686283386042e-05, 'epoch': 0.48}) +[2026-06-15 23:17:44,864][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.0649205818772316, 'train/grad_norm_pre_clip': 13.59693717956543, 'timing/data': 0.011697373352944851, 'timing/model': 1.5558324307203293, 'learning_rate/qwen_vl_interface': 8.876720344185967e-06, 'learning_rate/action_model': 8.876720344185967e-05, 'epoch': 0.48}) +[2026-06-15 23:17:51,174][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.07830200344324112, 'train/grad_norm_pre_clip': 18.926633834838867, 'timing/data': 0.011205891147255898, 'timing/model': 1.5716755958274007, 'learning_rate/qwen_vl_interface': 8.874752914178149e-06, 'learning_rate/action_model': 8.87475291417815e-05, 'epoch': 0.48}) +[2026-06-15 23:17:57,368][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.15565846860408783, 'train/grad_norm_pre_clip': 18.319595336914062, 'timing/data': 0.011524043045938015, 'timing/model': 1.5571917342022061, 'learning_rate/qwen_vl_interface': 8.87278399417133e-06, 'learning_rate/action_model': 8.872783994171329e-05, 'epoch': 0.48}) +[2026-06-15 23:18:03,666][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.01504136249423027, 'train/grad_norm_pre_clip': 7.488769054412842, 'timing/data': 0.011026919819414616, 'timing/model': 1.575624180957675, 'learning_rate/qwen_vl_interface': 8.870813584974852e-06, 'learning_rate/action_model': 8.870813584974852e-05, 'epoch': 0.48}) +[2026-06-15 23:18:09,964][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.031047821044921875, 'train/grad_norm_pre_clip': 8.259075164794922, 'timing/data': 0.01126445084810257, 'timing/model': 1.564757355954498, 'learning_rate/qwen_vl_interface': 8.868841687398683e-06, 'learning_rate/action_model': 8.868841687398683e-05, 'epoch': 0.48}) +[2026-06-15 23:29:08,125][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.1425909698009491, 'train/grad_norm_pre_clip': 27.044641494750977, 'eval/loss': 0.1090618297457695, 'eval/flap_precision': 0.9164345403899722, 'eval/flap_recall': 0.897680763983629, 'eval/flap_f1': 0.906960716747071, 'eval/flap_rate_model': 0.19944444444444445, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.906960716747071, 'timing/data': 0.01109826099127531, 'timing/model': 1.804075196851045, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 0.49}) +[2026-06-15 23:29:08,126][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1200_state +[2026-06-15 23:29:08,132][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:29:16,863][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1200_state/model.safetensors +[2026-06-15 23:29:36,532][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1200_state/optimizer.bin +[2026-06-15 23:29:36,532][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1200_state/sampler.bin +[2026-06-15 23:29:36,533][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1200_state/sampler_1.bin +[2026-06-15 23:29:36,534][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1200_state/random_states_0.pkl +[2026-06-15 23:29:36,724][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-06-15 23:29:36,724][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:29:36,739][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:29:43,106][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.15030856430530548, 'train/grad_norm_pre_clip': 5.95205020904541, 'timing/data': 0.011636105366051197, 'timing/model': 1.8471881817094982, 'learning_rate/qwen_vl_interface': 8.864893430350159e-06, 'learning_rate/action_model': 8.864893430350159e-05, 'epoch': 0.49}) +[2026-06-15 23:29:49,320][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.3417670726776123, 'train/grad_norm_pre_clip': 9.464919090270996, 'timing/data': 0.010266797617077827, 'timing/model': 1.7560083870775998, 'learning_rate/qwen_vl_interface': 8.862917072500785e-06, 'learning_rate/action_model': 8.862917072500784e-05, 'epoch': 0.49}) +[2026-06-15 23:29:55,311][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.5861345529556274, 'train/grad_norm_pre_clip': 77.27851104736328, 'timing/data': 0.011556994169950485, 'timing/model': 1.5577899194322526, 'learning_rate/qwen_vl_interface': 8.860939229517674e-06, 'learning_rate/action_model': 8.860939229517674e-05, 'epoch': 0.49}) +[2026-06-15 23:30:01,535][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.22280657291412354, 'train/grad_norm_pre_clip': 14.073911666870117, 'timing/data': 0.01147908391430974, 'timing/model': 1.5680509279482067, 'learning_rate/qwen_vl_interface': 8.858959902213844e-06, 'learning_rate/action_model': 8.858959902213845e-05, 'epoch': 0.49}) +[2026-06-15 23:30:07,724][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.7137607932090759, 'train/grad_norm_pre_clip': 25.769901275634766, 'timing/data': 0.01105562224984169, 'timing/model': 1.5599828301928937, 'learning_rate/qwen_vl_interface': 8.856979091402921e-06, 'learning_rate/action_model': 8.85697909140292e-05, 'epoch': 0.49}) +[2026-06-15 23:30:13,934][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.2121763527393341, 'train/grad_norm_pre_clip': 8.984519958496094, 'timing/data': 0.011402593925595284, 'timing/model': 1.5569783076643944, 'learning_rate/qwen_vl_interface': 8.854996797899139e-06, 'learning_rate/action_model': 8.854996797899139e-05, 'epoch': 0.49}) +[2026-06-15 23:30:20,146][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.20179373025894165, 'train/grad_norm_pre_clip': 9.705131530761719, 'timing/data': 0.011193013284355402, 'timing/model': 1.5439180368557572, 'learning_rate/qwen_vl_interface': 8.853013022517348e-06, 'learning_rate/action_model': 8.853013022517348e-05, 'epoch': 0.49}) +[2026-06-15 23:30:26,492][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.07775093615055084, 'train/grad_norm_pre_clip': 13.083499908447266, 'timing/data': 0.011921366211026907, 'timing/model': 1.8059053858742118, 'learning_rate/qwen_vl_interface': 8.851027766073002e-06, 'learning_rate/action_model': 8.851027766073001e-05, 'epoch': 0.49}) +[2026-06-15 23:30:32,546][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.12323877215385437, 'train/grad_norm_pre_clip': 10.440106391906738, 'timing/data': 0.01137520419433713, 'timing/model': 1.5624483949504793, 'learning_rate/qwen_vl_interface': 8.84904102938216e-06, 'learning_rate/action_model': 8.84904102938216e-05, 'epoch': 0.49}) +[2026-06-15 23:30:38,843][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.0499291867017746, 'train/grad_norm_pre_clip': 8.052902221679688, 'timing/data': 0.011015802156180143, 'timing/model': 1.5504862610250711, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 0.49}) +[2026-06-15 23:30:45,163][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.13398189842700958, 'train/grad_norm_pre_clip': 4.222113132476807, 'timing/data': 0.011103282682597637, 'timing/model': 1.5618023937568069, 'learning_rate/qwen_vl_interface': 8.845063118528303e-06, 'learning_rate/action_model': 8.845063118528304e-05, 'epoch': 0.49}) +[2026-06-15 23:30:51,560][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.21247072517871857, 'train/grad_norm_pre_clip': 7.340857028961182, 'timing/data': 0.012994850985705853, 'timing/model': 1.589638626202941, 'learning_rate/qwen_vl_interface': 8.843071946000455e-06, 'learning_rate/action_model': 8.843071946000456e-05, 'epoch': 0.49}) +[2026-06-15 23:30:57,846][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.12924863398075104, 'train/grad_norm_pre_clip': 7.583634853363037, 'timing/data': 0.011297822929918766, 'timing/model': 1.5374302216805518, 'learning_rate/qwen_vl_interface': 8.841079296496451e-06, 'learning_rate/action_model': 8.84107929649645e-05, 'epoch': 0.49}) +[2026-06-15 23:31:04,343][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.16049715876579285, 'train/grad_norm_pre_clip': 10.507970809936523, 'timing/data': 0.011386383790522814, 'timing/model': 1.8867896613664925, 'learning_rate/qwen_vl_interface': 8.839085170835396e-06, 'learning_rate/action_model': 8.839085170835395e-05, 'epoch': 0.49}) +[2026-06-15 23:31:10,459][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.23163765668869019, 'train/grad_norm_pre_clip': 8.430317878723145, 'timing/data': 0.011471183970570564, 'timing/model': 1.5648273820988834, 'learning_rate/qwen_vl_interface': 8.837089569837e-06, 'learning_rate/action_model': 8.837089569836998e-05, 'epoch': 0.49}) +[2026-06-15 23:31:16,754][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.20997609198093414, 'train/grad_norm_pre_clip': 8.09915542602539, 'timing/data': 0.011290962807834148, 'timing/model': 1.554059812799096, 'learning_rate/qwen_vl_interface': 8.835092494321575e-06, 'learning_rate/action_model': 8.835092494321575e-05, 'epoch': 0.49}) +[2026-06-15 23:31:23,146][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.05359144136309624, 'train/grad_norm_pre_clip': 2.5381312370300293, 'timing/data': 0.011085731908679008, 'timing/model': 1.592864464968443, 'learning_rate/qwen_vl_interface': 8.833093945110048e-06, 'learning_rate/action_model': 8.833093945110047e-05, 'epoch': 0.49}) +[2026-06-15 23:31:29,467][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.2377672791481018, 'train/grad_norm_pre_clip': 12.742929458618164, 'timing/data': 0.011405442841351032, 'timing/model': 1.5670403647236526, 'learning_rate/qwen_vl_interface': 8.831093923023946e-06, 'learning_rate/action_model': 8.831093923023945e-05, 'epoch': 0.49}) +[2026-06-15 23:31:35,764][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.12344127893447876, 'train/grad_norm_pre_clip': 5.9653730392456055, 'timing/data': 0.011220631655305624, 'timing/model': 1.5208347588777542, 'learning_rate/qwen_vl_interface': 8.8290924288854e-06, 'learning_rate/action_model': 8.8290924288854e-05, 'epoch': 0.49}) +[2026-06-15 23:31:42,102][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.16535839438438416, 'train/grad_norm_pre_clip': 6.962599754333496, 'timing/data': 0.011916795745491982, 'timing/model': 1.798335573170334, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.49}) +[2026-06-15 23:31:48,167][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.2684977650642395, 'train/grad_norm_pre_clip': 21.22981834411621, 'timing/data': 0.013015110976994038, 'timing/model': 1.5644882749766111, 'learning_rate/qwen_vl_interface': 8.825085027742545e-06, 'learning_rate/action_model': 8.825085027742544e-05, 'epoch': 0.49}) +[2026-06-15 23:31:54,458][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.04061231017112732, 'train/grad_norm_pre_clip': 2.6070761680603027, 'timing/data': 0.011513934005051851, 'timing/model': 1.5629821689799428, 'learning_rate/qwen_vl_interface': 8.823079122385523e-06, 'learning_rate/action_model': 8.823079122385523e-05, 'epoch': 0.49}) +[2026-06-15 23:32:00,762][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.18454892933368683, 'train/grad_norm_pre_clip': 9.278212547302246, 'timing/data': 0.0112869031727314, 'timing/model': 1.5622387360781431, 'learning_rate/qwen_vl_interface': 8.821071748270644e-06, 'learning_rate/action_model': 8.821071748270644e-05, 'epoch': 0.49}) +[2026-06-15 23:32:07,062][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.14831191301345825, 'train/grad_norm_pre_clip': 8.173529624938965, 'timing/data': 0.011907216161489487, 'timing/model': 1.5612545320764184, 'learning_rate/qwen_vl_interface': 8.81906290622306e-06, 'learning_rate/action_model': 8.81906290622306e-05, 'epoch': 0.49}) +[2026-06-15 23:32:13,366][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.026020262390375137, 'train/grad_norm_pre_clip': 1.569703221321106, 'timing/data': 0.011105162091553211, 'timing/model': 1.5216464470140636, 'learning_rate/qwen_vl_interface': 8.81705259706853e-06, 'learning_rate/action_model': 8.817052597068529e-05, 'epoch': 0.5}) +[2026-06-15 23:32:19,696][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.20675961673259735, 'train/grad_norm_pre_clip': 6.951544284820557, 'timing/data': 0.011044382117688656, 'timing/model': 1.7940111360512674, 'learning_rate/qwen_vl_interface': 8.815040821633416e-06, 'learning_rate/action_model': 8.815040821633416e-05, 'epoch': 0.5}) +[2026-06-15 23:32:25,759][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.14443379640579224, 'train/grad_norm_pre_clip': 8.175886154174805, 'timing/data': 0.011325253173708916, 'timing/model': 1.5482564736157656, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.5}) +[2026-06-15 23:32:32,056][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.15695622563362122, 'train/grad_norm_pre_clip': 10.540926933288574, 'timing/data': 0.010215488262474537, 'timing/model': 1.5656040660105646, 'learning_rate/qwen_vl_interface': 8.8110128752299e-06, 'learning_rate/action_model': 8.8110128752299e-05, 'epoch': 0.5}) +[2026-06-15 23:32:38,355][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.12779997289180756, 'train/grad_norm_pre_clip': 11.566689491271973, 'timing/data': 0.010087887290865183, 'timing/model': 1.5631162347272038, 'learning_rate/qwen_vl_interface': 8.808996705917235e-06, 'learning_rate/action_model': 8.808996705917235e-05, 'epoch': 0.5}) +[2026-06-15 23:32:44,655][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.34264835715293884, 'train/grad_norm_pre_clip': 20.008174896240234, 'timing/data': 0.010552580002695322, 'timing/model': 1.5639993092045188, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 0.5}) +[2026-06-15 23:32:50,958][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.044292956590652466, 'train/grad_norm_pre_clip': 6.356901168823242, 'timing/data': 0.010047786869108677, 'timing/model': 1.513286110945046, 'learning_rate/qwen_vl_interface': 8.804959979213945e-06, 'learning_rate/action_model': 8.804959979213945e-05, 'epoch': 0.5}) +[2026-06-15 23:32:57,304][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.44238755106925964, 'train/grad_norm_pre_clip': 17.596891403198242, 'timing/data': 0.010518020018935204, 'timing/model': 1.8134482828900218, 'learning_rate/qwen_vl_interface': 8.802939423482667e-06, 'learning_rate/action_model': 8.802939423482666e-05, 'epoch': 0.5}) +[2026-06-15 23:33:03,362][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.18785083293914795, 'train/grad_norm_pre_clip': 17.37632179260254, 'timing/data': 0.011647045146673918, 'timing/model': 1.5615503787994385, 'learning_rate/qwen_vl_interface': 8.800917407272198e-06, 'learning_rate/action_model': 8.800917407272198e-05, 'epoch': 0.5}) +[2026-06-15 23:33:09,660][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.15041491389274597, 'train/grad_norm_pre_clip': 7.345471382141113, 'timing/data': 0.01117445295676589, 'timing/model': 1.5590903791598976, 'learning_rate/qwen_vl_interface': 8.798893931413711e-06, 'learning_rate/action_model': 8.79889393141371e-05, 'epoch': 0.5}) +[2026-06-15 23:33:16,073][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.35314419865608215, 'train/grad_norm_pre_clip': 16.449260711669922, 'timing/data': 0.011656027287244797, 'timing/model': 1.5720536853186786, 'learning_rate/qwen_vl_interface': 8.796868996738985e-06, 'learning_rate/action_model': 8.796868996738984e-05, 'epoch': 0.5}) +[2026-06-15 23:33:22,359][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.2773810923099518, 'train/grad_norm_pre_clip': 11.136815071105957, 'timing/data': 0.01159023679792881, 'timing/model': 1.5607058801688254, 'learning_rate/qwen_vl_interface': 8.794842604080389e-06, 'learning_rate/action_model': 8.794842604080389e-05, 'epoch': 0.5}) +[2026-06-15 23:33:28,662][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.12586073577404022, 'train/grad_norm_pre_clip': 9.988302230834961, 'timing/data': 0.011189037002623081, 'timing/model': 1.532027519773692, 'learning_rate/qwen_vl_interface': 8.7928147542709e-06, 'learning_rate/action_model': 8.7928147542709e-05, 'epoch': 0.5}) +[2026-06-15 23:33:35,000][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.2149541825056076, 'train/grad_norm_pre_clip': 11.306601524353027, 'timing/data': 0.011376298032701015, 'timing/model': 1.8040784536860883, 'learning_rate/qwen_vl_interface': 8.790785448144088e-06, 'learning_rate/action_model': 8.790785448144088e-05, 'epoch': 0.5}) +[2026-06-15 23:33:41,229][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.13224312663078308, 'train/grad_norm_pre_clip': 6.301710605621338, 'timing/data': 0.01119870925322175, 'timing/model': 1.7188142761588097, 'learning_rate/qwen_vl_interface': 8.788754686534126e-06, 'learning_rate/action_model': 8.788754686534127e-05, 'epoch': 0.5}) +[2026-06-15 23:33:47,265][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.08420959115028381, 'train/grad_norm_pre_clip': 5.857232093811035, 'timing/data': 0.011440149042755365, 'timing/model': 1.5627125198952854, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 0.5}) +[2026-06-15 23:33:53,566][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.2840661406517029, 'train/grad_norm_pre_clip': 11.759068489074707, 'timing/data': 0.011743870098143816, 'timing/model': 1.55806370312348, 'learning_rate/qwen_vl_interface': 8.784688800204427e-06, 'learning_rate/action_model': 8.784688800204426e-05, 'epoch': 0.5}) +[2026-06-15 23:33:59,859][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.12031596153974533, 'train/grad_norm_pre_clip': 8.610696792602539, 'timing/data': 0.011005228850990534, 'timing/model': 1.5583738880231977, 'learning_rate/qwen_vl_interface': 8.782653677156022e-06, 'learning_rate/action_model': 8.782653677156022e-05, 'epoch': 0.5}) +[2026-06-15 23:34:06,264][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.13438546657562256, 'train/grad_norm_pre_clip': 8.968697547912598, 'timing/data': 0.011500430759042501, 'timing/model': 1.5696242339909077, 'learning_rate/qwen_vl_interface': 8.780617101967129e-06, 'learning_rate/action_model': 8.780617101967129e-05, 'epoch': 0.5}) +[2026-06-15 23:34:12,700][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.13390111923217773, 'train/grad_norm_pre_clip': 6.470911502838135, 'timing/data': 0.011120170820504427, 'timing/model': 1.826250268612057, 'learning_rate/qwen_vl_interface': 8.77857907547491e-06, 'learning_rate/action_model': 8.77857907547491e-05, 'epoch': 0.5}) +[2026-06-15 23:34:19,017][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.11127260327339172, 'train/grad_norm_pre_clip': 6.258545398712158, 'timing/data': 0.010732361115515232, 'timing/model': 1.8148415619507432, 'learning_rate/qwen_vl_interface': 8.776539598517119e-06, 'learning_rate/action_model': 8.776539598517118e-05, 'epoch': 0.5}) +[2026-06-15 23:34:25,059][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.14237698912620544, 'train/grad_norm_pre_clip': 15.930400848388672, 'timing/data': 0.011380531825125217, 'timing/model': 1.5513016232289374, 'learning_rate/qwen_vl_interface': 8.774498671932108e-06, 'learning_rate/action_model': 8.774498671932107e-05, 'epoch': 0.5}) +[2026-06-15 23:34:31,354][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.33371955156326294, 'train/grad_norm_pre_clip': 17.066585540771484, 'timing/data': 0.011019481811672449, 'timing/model': 1.552835885901004, 'learning_rate/qwen_vl_interface': 8.772456296558824e-06, 'learning_rate/action_model': 8.772456296558824e-05, 'epoch': 0.5}) +[2026-06-15 23:34:37,674][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.16342072188854218, 'train/grad_norm_pre_clip': 9.67087459564209, 'timing/data': 0.011118311900645494, 'timing/model': 1.5659139822237194, 'learning_rate/qwen_vl_interface': 8.77041247323681e-06, 'learning_rate/action_model': 8.770412473236811e-05, 'epoch': 0.5}) +[2026-06-15 23:34:43,973][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.2994650602340698, 'train/grad_norm_pre_clip': 8.270249366760254, 'timing/data': 0.01145148416981101, 'timing/model': 1.5259108440950513, 'learning_rate/qwen_vl_interface': 8.76836720280621e-06, 'learning_rate/action_model': 8.768367202806209e-05, 'epoch': 0.51}) +[2026-06-15 23:34:50,308][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.07756797224283218, 'train/grad_norm_pre_clip': 6.357613563537598, 'timing/data': 0.010971983894705772, 'timing/model': 1.8092956352047622, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 0.51}) +[2026-06-15 23:34:56,358][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.07790203392505646, 'train/grad_norm_pre_clip': 3.8263614177703857, 'timing/data': 0.011319533921778202, 'timing/model': 1.5548913138918579, 'learning_rate/qwen_vl_interface': 8.76427232398276e-06, 'learning_rate/action_model': 8.76427232398276e-05, 'epoch': 0.51}) +[2026-06-15 23:35:02,656][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.07275646924972534, 'train/grad_norm_pre_clip': 4.204984664916992, 'timing/data': 0.011784525122493505, 'timing/model': 1.548962493892759, 'learning_rate/qwen_vl_interface': 8.762222717273167e-06, 'learning_rate/action_model': 8.762222717273167e-05, 'epoch': 0.51}) +[2026-06-15 23:35:08,963][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.07428085803985596, 'train/grad_norm_pre_clip': 3.984274387359619, 'timing/data': 0.011316124815493822, 'timing/model': 1.5649093696847558, 'learning_rate/qwen_vl_interface': 8.76017166682148e-06, 'learning_rate/action_model': 8.760171666821481e-05, 'epoch': 0.51}) +[2026-06-15 23:35:15,324][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.07894797623157501, 'train/grad_norm_pre_clip': 5.296658039093018, 'timing/data': 0.012018656823784113, 'timing/model': 1.5768328541889787, 'learning_rate/qwen_vl_interface': 8.758119173470815e-06, 'learning_rate/action_model': 8.758119173470815e-05, 'epoch': 0.51}) +[2026-06-15 23:35:21,561][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.07830750942230225, 'train/grad_norm_pre_clip': 9.054033279418945, 'timing/data': 0.011797686107456684, 'timing/model': 1.5784655730240047, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 0.51}) +[2026-06-15 23:35:27,853][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.0406983345746994, 'train/grad_norm_pre_clip': 2.962287425994873, 'timing/data': 0.011686326004564762, 'timing/model': 1.747440509032458, 'learning_rate/qwen_vl_interface': 8.75400986144794e-06, 'learning_rate/action_model': 8.75400986144794e-05, 'epoch': 0.51}) +[2026-06-15 23:35:33,969][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.1739807277917862, 'train/grad_norm_pre_clip': 11.608800888061523, 'timing/data': 0.01143755717203021, 'timing/model': 1.5644964901730418, 'learning_rate/qwen_vl_interface': 8.751953044464913e-06, 'learning_rate/action_model': 8.751953044464913e-05, 'epoch': 0.51}) +[2026-06-15 23:35:40,259][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.1508156955242157, 'train/grad_norm_pre_clip': 7.931847095489502, 'timing/data': 0.011312686838209629, 'timing/model': 1.5520542948506773, 'learning_rate/qwen_vl_interface': 8.749894787961272e-06, 'learning_rate/action_model': 8.749894787961271e-05, 'epoch': 0.51}) +[2026-06-15 23:35:46,559][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.06865387409925461, 'train/grad_norm_pre_clip': 5.462488174438477, 'timing/data': 0.010827807243913412, 'timing/model': 1.5655593993142247, 'learning_rate/qwen_vl_interface': 8.747835092783086e-06, 'learning_rate/action_model': 8.747835092783085e-05, 'epoch': 0.51}) +[2026-06-15 23:35:52,855][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.037127263844013214, 'train/grad_norm_pre_clip': 1.9304596185684204, 'timing/data': 0.01047362620010972, 'timing/model': 1.5358111853711307, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 0.51}) +[2026-06-15 23:35:59,159][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.06763963401317596, 'train/grad_norm_pre_clip': 10.719654083251953, 'timing/data': 0.01042966591194272, 'timing/model': 1.5141357318498194, 'learning_rate/qwen_vl_interface': 8.743711389790321e-06, 'learning_rate/action_model': 8.74371138979032e-05, 'epoch': 0.51}) +[2026-06-15 23:36:05,452][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.05095000937581062, 'train/grad_norm_pre_clip': 2.333435297012329, 'timing/data': 0.010573426727205515, 'timing/model': 1.7576463078148663, 'learning_rate/qwen_vl_interface': 8.741647383670839e-06, 'learning_rate/action_model': 8.741647383670839e-05, 'epoch': 0.51}) +[2026-06-15 23:36:11,572][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.027066625654697418, 'train/grad_norm_pre_clip': 6.430027484893799, 'timing/data': 0.010177917312830687, 'timing/model': 1.5773997539654374, 'learning_rate/qwen_vl_interface': 8.739581942267011e-06, 'learning_rate/action_model': 8.739581942267011e-05, 'epoch': 0.51}) +[2026-06-15 23:36:17,861][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.014524171128869057, 'train/grad_norm_pre_clip': 2.213505744934082, 'timing/data': 0.010257226880639791, 'timing/model': 1.5543641848489642, 'learning_rate/qwen_vl_interface': 8.737515066427861e-06, 'learning_rate/action_model': 8.73751506642786e-05, 'epoch': 0.51}) +[2026-06-15 23:36:24,225][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.005651747342199087, 'train/grad_norm_pre_clip': 7.456860065460205, 'timing/data': 0.011186179239302874, 'timing/model': 1.6193794156424701, 'learning_rate/qwen_vl_interface': 8.735446757003e-06, 'learning_rate/action_model': 8.735446757003e-05, 'epoch': 0.51}) +[2026-06-15 23:36:30,455][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.013854801654815674, 'train/grad_norm_pre_clip': 4.247961521148682, 'timing/data': 0.011214219965040684, 'timing/model': 1.52687244489789, 'learning_rate/qwen_vl_interface': 8.733377014842635e-06, 'learning_rate/action_model': 8.733377014842635e-05, 'epoch': 0.51}) +[2026-06-15 23:36:36,785][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.019002806395292282, 'train/grad_norm_pre_clip': 13.261152267456055, 'timing/data': 0.01006105775013566, 'timing/model': 1.7882395777851343, 'learning_rate/qwen_vl_interface': 8.731305840797557e-06, 'learning_rate/action_model': 8.731305840797558e-05, 'epoch': 0.51}) +[2026-06-15 23:36:42,855][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.2480086237192154, 'train/grad_norm_pre_clip': 18.893844604492188, 'timing/data': 0.010751939378678799, 'timing/model': 1.5638073231093585, 'learning_rate/qwen_vl_interface': 8.72923323571915e-06, 'learning_rate/action_model': 8.72923323571915e-05, 'epoch': 0.51}) +[2026-06-15 23:36:49,258][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.15556401014328003, 'train/grad_norm_pre_clip': 9.627302169799805, 'timing/data': 0.010278498753905296, 'timing/model': 1.6601470792666078, 'learning_rate/qwen_vl_interface': 8.727159200459381e-06, 'learning_rate/action_model': 8.727159200459381e-05, 'epoch': 0.51}) +[2026-06-15 23:36:55,579][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.01385109405964613, 'train/grad_norm_pre_clip': 12.816043853759766, 'timing/data': 0.011697552166879177, 'timing/model': 1.5790444258600473, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 0.51}) +[2026-06-15 23:37:01,860][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.15865181386470795, 'train/grad_norm_pre_clip': 10.030319213867188, 'timing/data': 0.010920099914073944, 'timing/model': 1.542904012836516, 'learning_rate/qwen_vl_interface': 8.723006842806575e-06, 'learning_rate/action_model': 8.723006842806576e-05, 'epoch': 0.51}) +[2026-06-15 23:37:08,178][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 1.01833176612854, 'train/grad_norm_pre_clip': 54.606781005859375, 'timing/data': 0.011323411948978901, 'timing/model': 1.5630903369747102, 'learning_rate/qwen_vl_interface': 8.720928522120419e-06, 'learning_rate/action_model': 8.720928522120418e-05, 'epoch': 0.51}) +[2026-06-15 23:37:14,502][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.01829271949827671, 'train/grad_norm_pre_clip': 6.673907279968262, 'timing/data': 0.010666021145880222, 'timing/model': 1.7985414532013237, 'learning_rate/qwen_vl_interface': 8.718848774666655e-06, 'learning_rate/action_model': 8.718848774666655e-05, 'epoch': 0.51}) +[2026-06-15 23:37:20,678][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.05068519338965416, 'train/grad_norm_pre_clip': 18.145614624023438, 'timing/data': 0.011092063039541245, 'timing/model': 1.5790836568921804, 'learning_rate/qwen_vl_interface': 8.716767601300189e-06, 'learning_rate/action_model': 8.716767601300188e-05, 'epoch': 0.52}) +[2026-06-15 23:37:26,972][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.39650893211364746, 'train/grad_norm_pre_clip': 30.447378158569336, 'timing/data': 0.010742081794887781, 'timing/model': 1.5740817696787417, 'learning_rate/qwen_vl_interface': 8.714685002876514e-06, 'learning_rate/action_model': 8.714685002876513e-05, 'epoch': 0.52}) +[2026-06-15 23:37:33,273][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.30150532722473145, 'train/grad_norm_pre_clip': 18.0452880859375, 'timing/data': 0.01182247418910265, 'timing/model': 1.5691804639063776, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.52}) +[2026-06-15 23:37:39,559][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.2899326682090759, 'train/grad_norm_pre_clip': 35.66837692260742, 'timing/data': 0.010611792095005512, 'timing/model': 1.5652319891378284, 'learning_rate/qwen_vl_interface': 8.710515534282428e-06, 'learning_rate/action_model': 8.710515534282428e-05, 'epoch': 0.52}) +[2026-06-15 23:37:45,968][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.15116535127162933, 'train/grad_norm_pre_clip': 37.07880783081055, 'timing/data': 0.011335274204611778, 'timing/model': 1.54406627593562, 'learning_rate/qwen_vl_interface': 8.708428665825931e-06, 'learning_rate/action_model': 8.708428665825931e-05, 'epoch': 0.52}) +[2026-06-15 23:37:52,295][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.01199153158813715, 'train/grad_norm_pre_clip': 5.976186752319336, 'timing/data': 0.012130275834351778, 'timing/model': 1.7946176500990987, 'learning_rate/qwen_vl_interface': 8.706340375740045e-06, 'learning_rate/action_model': 8.706340375740045e-05, 'epoch': 0.52}) +[2026-06-15 23:37:58,367][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.005132016260176897, 'train/grad_norm_pre_clip': 1.2271884679794312, 'timing/data': 0.010636813007295132, 'timing/model': 1.556791024748236, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 0.52}) +[2026-06-15 23:38:04,763][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.26301226019859314, 'train/grad_norm_pre_clip': 12.691980361938477, 'timing/data': 0.01126639498397708, 'timing/model': 1.566338336095214, 'learning_rate/qwen_vl_interface': 8.70215953411436e-06, 'learning_rate/action_model': 8.70215953411436e-05, 'epoch': 0.52}) +[2026-06-15 23:38:11,067][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.07448583841323853, 'train/grad_norm_pre_clip': 22.934783935546875, 'timing/data': 0.0113443648442626, 'timing/model': 1.5626936983317137, 'learning_rate/qwen_vl_interface': 8.70006698429315e-06, 'learning_rate/action_model': 8.700066984293149e-05, 'epoch': 0.52}) +[2026-06-15 23:38:17,356][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.13138896226882935, 'train/grad_norm_pre_clip': 27.44280242919922, 'timing/data': 0.010658503975719213, 'timing/model': 1.518177189398557, 'learning_rate/qwen_vl_interface': 8.697973016279722e-06, 'learning_rate/action_model': 8.697973016279722e-05, 'epoch': 0.52}) +[2026-06-15 23:38:23,657][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.21688315272331238, 'train/grad_norm_pre_clip': 14.855939865112305, 'timing/data': 0.011204265058040619, 'timing/model': 1.5451483973301947, 'learning_rate/qwen_vl_interface': 8.695877630934827e-06, 'learning_rate/action_model': 8.695877630934826e-05, 'epoch': 0.52}) +[2026-06-15 23:38:29,985][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.3640982210636139, 'train/grad_norm_pre_clip': 29.785324096679688, 'timing/data': 0.010593024082481861, 'timing/model': 1.792257288005203, 'learning_rate/qwen_vl_interface': 8.693780829119801e-06, 'learning_rate/action_model': 8.693780829119801e-05, 'epoch': 0.52}) +[2026-06-15 23:38:36,162][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.40582093596458435, 'train/grad_norm_pre_clip': 9.342741012573242, 'timing/data': 0.010266464203596115, 'timing/model': 1.5787973999977112, 'learning_rate/qwen_vl_interface': 8.69168261169656e-06, 'learning_rate/action_model': 8.691682611696559e-05, 'epoch': 0.52}) +[2026-06-15 23:38:42,675][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.011617222800850868, 'train/grad_norm_pre_clip': 2.141186237335205, 'timing/data': 0.011091426014900208, 'timing/model': 1.6670629647560418, 'learning_rate/qwen_vl_interface': 8.689582979527599e-06, 'learning_rate/action_model': 8.689582979527599e-05, 'epoch': 0.52}) +[2026-06-15 23:38:49,065][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.16344018280506134, 'train/grad_norm_pre_clip': 24.062137603759766, 'timing/data': 0.010926025919616222, 'timing/model': 1.5673734000883996, 'learning_rate/qwen_vl_interface': 8.687481933476e-06, 'learning_rate/action_model': 8.687481933476e-05, 'epoch': 0.52}) +[2026-06-15 23:38:55,365][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.04274650290608406, 'train/grad_norm_pre_clip': 10.136112213134766, 'timing/data': 0.01100966613739729, 'timing/model': 1.5719466377049685, 'learning_rate/qwen_vl_interface': 8.685379474405423e-06, 'learning_rate/action_model': 8.685379474405423e-05, 'epoch': 0.52}) +[2026-06-15 23:39:01,673][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.010019891895353794, 'train/grad_norm_pre_clip': 2.6965348720550537, 'timing/data': 0.010753476060926914, 'timing/model': 1.5313376667909324, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.52}) +[2026-06-15 23:39:08,004][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.19831928610801697, 'train/grad_norm_pre_clip': 20.422809600830078, 'timing/data': 0.011653808411210775, 'timing/model': 1.8057715273462236, 'learning_rate/qwen_vl_interface': 8.681170320664883e-06, 'learning_rate/action_model': 8.681170320664883e-05, 'epoch': 0.52}) +[2026-06-15 23:39:14,064][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.1218433529138565, 'train/grad_norm_pre_clip': 18.304935455322266, 'timing/data': 0.011047806590795517, 'timing/model': 1.5723496158607304, 'learning_rate/qwen_vl_interface': 8.679063627725143e-06, 'learning_rate/action_model': 8.679063627725144e-05, 'epoch': 0.52}) +[2026-06-15 23:39:20,458][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.20935139060020447, 'train/grad_norm_pre_clip': 11.400398254394531, 'timing/data': 0.010689006187021732, 'timing/model': 1.592833542264998, 'learning_rate/qwen_vl_interface': 8.676955525226876e-06, 'learning_rate/action_model': 8.676955525226875e-05, 'epoch': 0.52}) +[2026-06-15 23:39:26,767][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.03745703026652336, 'train/grad_norm_pre_clip': 10.990532875061035, 'timing/data': 0.011870409827679396, 'timing/model': 1.5686983270570636, 'learning_rate/qwen_vl_interface': 8.674846014036636e-06, 'learning_rate/action_model': 8.674846014036636e-05, 'epoch': 0.52}) +[2026-06-15 23:39:33,078][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.07754873484373093, 'train/grad_norm_pre_clip': 9.725875854492188, 'timing/data': 0.011496559251099825, 'timing/model': 1.5761826098896563, 'learning_rate/qwen_vl_interface': 8.672735095021569e-06, 'learning_rate/action_model': 8.672735095021569e-05, 'epoch': 0.52}) +[2026-06-15 23:39:39,360][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.15383906662464142, 'train/grad_norm_pre_clip': 15.176647186279297, 'timing/data': 0.011308218818157911, 'timing/model': 1.5256810961291194, 'learning_rate/qwen_vl_interface': 8.670622769049393e-06, 'learning_rate/action_model': 8.670622769049392e-05, 'epoch': 0.52}) +[2026-06-15 23:39:45,797][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.04433057829737663, 'train/grad_norm_pre_clip': 6.325459003448486, 'timing/data': 0.011647130362689495, 'timing/model': 1.78156015323475, 'learning_rate/qwen_vl_interface': 8.668509036988406e-06, 'learning_rate/action_model': 8.668509036988405e-05, 'epoch': 0.52}) +[2026-06-15 23:39:51,863][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.09408880770206451, 'train/grad_norm_pre_clip': 8.600343704223633, 'timing/data': 0.01163657009601593, 'timing/model': 1.5587346348911524, 'learning_rate/qwen_vl_interface': 8.66639389970748e-06, 'learning_rate/action_model': 8.666393899707479e-05, 'epoch': 0.52}) +[2026-06-15 23:39:58,172][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.1766590029001236, 'train/grad_norm_pre_clip': 11.45142650604248, 'timing/data': 0.011711359955370426, 'timing/model': 1.568342812359333, 'learning_rate/qwen_vl_interface': 8.66427735807607e-06, 'learning_rate/action_model': 8.664277358076069e-05, 'epoch': 0.53}) +[2026-06-15 23:50:54,500][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.04363705590367317, 'train/grad_norm_pre_clip': 5.689990520477295, 'eval/loss': 0.06567003577947617, 'eval/flap_precision': 0.9217731421121251, 'eval/flap_recall': 0.9645293315143247, 'eval/flap_f1': 0.9426666666666668, 'eval/flap_rate_model': 0.21305555555555555, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9426666666666667, 'timing/data': 0.010925458278506994, 'timing/model': 1.5599602642469108, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 0.53}) +[2026-06-15 23:50:54,501][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1300_state +[2026-06-15 23:50:54,508][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:51:03,211][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1300_state/model.safetensors +[2026-06-15 23:51:23,130][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1300_state/optimizer.bin +[2026-06-15 23:51:23,131][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1300_state/sampler.bin +[2026-06-15 23:51:23,131][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1300_state/sampler_1.bin +[2026-06-15 23:51:23,132][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1300_state/random_states_0.pkl +[2026-06-15 23:51:23,329][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-06-15 23:51:23,330][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:51:23,345][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:51:29,635][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.05947952717542648, 'train/grad_norm_pre_clip': 6.260368824005127, 'timing/data': 0.011673684697598219, 'timing/model': 1.4922746187075973, 'learning_rate/qwen_vl_interface': 8.660040065242499e-06, 'learning_rate/action_model': 8.660040065242498e-05, 'epoch': 0.53}) +[2026-06-15 23:51:35,835][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.32669374346733093, 'train/grad_norm_pre_clip': 22.88638687133789, 'timing/data': 0.011771365068852901, 'timing/model': 1.5573153248988092, 'learning_rate/qwen_vl_interface': 8.657919315782127e-06, 'learning_rate/action_model': 8.657919315782127e-05, 'epoch': 0.53}) +[2026-06-15 23:51:42,067][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.16808438301086426, 'train/grad_norm_pre_clip': 10.211176872253418, 'timing/data': 0.01139072421938181, 'timing/model': 1.5935049033723772, 'learning_rate/qwen_vl_interface': 8.655797165454853e-06, 'learning_rate/action_model': 8.655797165454853e-05, 'epoch': 0.53}) +[2026-06-15 23:51:48,337][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.023518899455666542, 'train/grad_norm_pre_clip': 1.5642035007476807, 'timing/data': 0.01307313097640872, 'timing/model': 1.497203349135816, 'learning_rate/qwen_vl_interface': 8.653673615133013e-06, 'learning_rate/action_model': 8.653673615133013e-05, 'epoch': 0.53}) +[2026-06-15 23:51:54,577][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.1985262632369995, 'train/grad_norm_pre_clip': 7.229814529418945, 'timing/data': 0.011668744962662458, 'timing/model': 1.8060037489049137, 'learning_rate/qwen_vl_interface': 8.651548665689519e-06, 'learning_rate/action_model': 8.651548665689518e-05, 'epoch': 0.53}) +[2026-06-15 23:52:00,648][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.17836497724056244, 'train/grad_norm_pre_clip': 8.141708374023438, 'timing/data': 0.011136002838611603, 'timing/model': 1.5699062258936465, 'learning_rate/qwen_vl_interface': 8.649422317997855e-06, 'learning_rate/action_model': 8.649422317997856e-05, 'epoch': 0.53}) +[2026-06-15 23:52:06,947][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.18603180348873138, 'train/grad_norm_pre_clip': 7.68595027923584, 'timing/data': 0.013150871265679598, 'timing/model': 1.5620553833432496, 'learning_rate/qwen_vl_interface': 8.647294572932082e-06, 'learning_rate/action_model': 8.647294572932082e-05, 'epoch': 0.53}) +[2026-06-15 23:52:13,247][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.3873201012611389, 'train/grad_norm_pre_clip': 18.125951766967773, 'timing/data': 0.011402854230254889, 'timing/model': 1.5702364728786051, 'learning_rate/qwen_vl_interface': 8.64516543136684e-06, 'learning_rate/action_model': 8.64516543136684e-05, 'epoch': 0.53}) +[2026-06-15 23:52:19,542][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.1860022246837616, 'train/grad_norm_pre_clip': 3.1577420234680176, 'timing/data': 0.011348113883286715, 'timing/model': 1.5600946806371212, 'learning_rate/qwen_vl_interface': 8.643034894177335e-06, 'learning_rate/action_model': 8.643034894177335e-05, 'epoch': 0.53}) +[2026-06-15 23:52:25,762][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.18644239008426666, 'train/grad_norm_pre_clip': 20.323640823364258, 'timing/data': 0.011401053983718157, 'timing/model': 1.5311092361807823, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 0.53}) +[2026-06-15 23:52:32,078][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.22991658747196198, 'train/grad_norm_pre_clip': 6.958970069885254, 'timing/data': 0.010758871678262949, 'timing/model': 1.7932727918960154, 'learning_rate/qwen_vl_interface': 8.638769636429245e-06, 'learning_rate/action_model': 8.638769636429245e-05, 'epoch': 0.53}) +[2026-06-15 23:52:38,321][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.05766276642680168, 'train/grad_norm_pre_clip': 6.54280424118042, 'timing/data': 0.011855985969305038, 'timing/model': 1.7178666973486543, 'learning_rate/qwen_vl_interface': 8.636634917623946e-06, 'learning_rate/action_model': 8.636634917623944e-05, 'epoch': 0.53}) +[2026-06-15 23:52:44,545][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.05362311378121376, 'train/grad_norm_pre_clip': 3.0819718837738037, 'timing/data': 0.010997753124684095, 'timing/model': 1.5562473046593368, 'learning_rate/qwen_vl_interface': 8.634498806700955e-06, 'learning_rate/action_model': 8.634498806700954e-05, 'epoch': 0.53}) +[2026-06-15 23:52:50,852][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.1175495907664299, 'train/grad_norm_pre_clip': 6.219857692718506, 'timing/data': 0.011476135347038507, 'timing/model': 1.5613919929601252, 'learning_rate/qwen_vl_interface': 8.63236130453835e-06, 'learning_rate/action_model': 8.63236130453835e-05, 'epoch': 0.53}) +[2026-06-15 23:52:57,162][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.06916364282369614, 'train/grad_norm_pre_clip': 8.817728042602539, 'timing/data': 0.011490825098007917, 'timing/model': 1.5701766437850893, 'learning_rate/qwen_vl_interface': 8.630222412014777e-06, 'learning_rate/action_model': 8.630222412014776e-05, 'epoch': 0.53}) +[2026-06-15 23:53:03,453][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.2060842663049698, 'train/grad_norm_pre_clip': 7.817103385925293, 'timing/data': 0.011338504031300545, 'timing/model': 1.5220830631442368, 'learning_rate/qwen_vl_interface': 8.628082130009451e-06, 'learning_rate/action_model': 8.628082130009451e-05, 'epoch': 0.53}) +[2026-06-15 23:53:09,791][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.022331366315484047, 'train/grad_norm_pre_clip': 1.2245579957962036, 'timing/data': 0.011419444810599089, 'timing/model': 1.7993116951547563, 'learning_rate/qwen_vl_interface': 8.62594045940216e-06, 'learning_rate/action_model': 8.62594045940216e-05, 'epoch': 0.53}) +[2026-06-15 23:53:16,082][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.08557257801294327, 'train/grad_norm_pre_clip': 4.985179901123047, 'timing/data': 0.010829662904143333, 'timing/model': 1.7906123977154493, 'learning_rate/qwen_vl_interface': 8.623797401073272e-06, 'learning_rate/action_model': 8.623797401073271e-05, 'epoch': 0.53}) +[2026-06-15 23:53:22,161][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.14514674246311188, 'train/grad_norm_pre_clip': 4.4263081550598145, 'timing/data': 0.012123707216233015, 'timing/model': 1.5615828079171479, 'learning_rate/qwen_vl_interface': 8.62165295590371e-06, 'learning_rate/action_model': 8.621652955903709e-05, 'epoch': 0.53}) +[2026-06-15 23:53:28,453][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.03845054283738136, 'train/grad_norm_pre_clip': 8.927433013916016, 'timing/data': 0.011817547027021646, 'timing/model': 1.550880012102425, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 0.53}) +[2026-06-15 23:53:34,757][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.16885045170783997, 'train/grad_norm_pre_clip': 25.56673240661621, 'timing/data': 0.011047613807022572, 'timing/model': 1.56110272789374, 'learning_rate/qwen_vl_interface': 8.61735990856914e-06, 'learning_rate/action_model': 8.617359908569139e-05, 'epoch': 0.53}) +[2026-06-15 23:53:41,134][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.019741877913475037, 'train/grad_norm_pre_clip': 1.5592862367630005, 'timing/data': 0.01160340616479516, 'timing/model': 1.5831099622882903, 'learning_rate/qwen_vl_interface': 8.615211308168842e-06, 'learning_rate/action_model': 8.61521130816884e-05, 'epoch': 0.53}) +[2026-06-15 23:53:47,407][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.02130173332989216, 'train/grad_norm_pre_clip': 1.8327312469482422, 'timing/data': 0.01189809711650014, 'timing/model': 1.8082837858237326, 'learning_rate/qwen_vl_interface': 8.613061324457289e-06, 'learning_rate/action_model': 8.613061324457288e-05, 'epoch': 0.54}) +[2026-06-15 23:53:53,483][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.04727652668952942, 'train/grad_norm_pre_clip': 8.52452564239502, 'timing/data': 0.011294864118099213, 'timing/model': 1.577873864211142, 'learning_rate/qwen_vl_interface': 8.610909958318262e-06, 'learning_rate/action_model': 8.610909958318261e-05, 'epoch': 0.54}) +[2026-06-15 23:53:59,766][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.2794627547264099, 'train/grad_norm_pre_clip': 9.968459129333496, 'timing/data': 0.0114390947856009, 'timing/model': 1.5633961237035692, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.54}) +[2026-06-15 23:54:06,062][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.10670114308595657, 'train/grad_norm_pre_clip': 4.290114402770996, 'timing/data': 0.011483006179332733, 'timing/model': 1.5576204359531403, 'learning_rate/qwen_vl_interface': 8.606603082295724e-06, 'learning_rate/action_model': 8.606603082295724e-05, 'epoch': 0.54}) +[2026-06-15 23:54:12,448][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.11004650592803955, 'train/grad_norm_pre_clip': 19.390413284301758, 'timing/data': 0.011232024058699608, 'timing/model': 1.5364062190055847, 'learning_rate/qwen_vl_interface': 8.60444757418261e-06, 'learning_rate/action_model': 8.604447574182609e-05, 'epoch': 0.54}) +[2026-06-15 23:54:18,745][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.10363270342350006, 'train/grad_norm_pre_clip': 10.850628852844238, 'timing/data': 0.011330953799188137, 'timing/model': 1.5841882941313088, 'learning_rate/qwen_vl_interface': 8.602290687182805e-06, 'learning_rate/action_model': 8.602290687182806e-05, 'epoch': 0.54}) +[2026-06-15 23:54:25,044][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.006523448042571545, 'train/grad_norm_pre_clip': 2.060559034347534, 'timing/data': 0.011599486228078604, 'timing/model': 1.738092590123415, 'learning_rate/qwen_vl_interface': 8.600132422182929e-06, 'learning_rate/action_model': 8.600132422182928e-05, 'epoch': 0.54}) +[2026-06-15 23:54:31,227][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.09935164451599121, 'train/grad_norm_pre_clip': 10.901960372924805, 'timing/data': 0.010712612885981798, 'timing/model': 1.6316382042132318, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 0.54}) +[2026-06-15 23:54:37,456][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.263521671295166, 'train/grad_norm_pre_clip': 12.971467018127441, 'timing/data': 0.0113046751357615, 'timing/model': 1.5600721859373152, 'learning_rate/qwen_vl_interface': 8.595811761732239e-06, 'learning_rate/action_model': 8.595811761732238e-05, 'epoch': 0.54}) +[2026-06-15 23:54:43,757][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.09343316406011581, 'train/grad_norm_pre_clip': 12.762618064880371, 'timing/data': 0.010793792549520731, 'timing/model': 1.5630138427950442, 'learning_rate/qwen_vl_interface': 8.593649368057489e-06, 'learning_rate/action_model': 8.59364936805749e-05, 'epoch': 0.54}) +[2026-06-15 23:54:50,057][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.10993915796279907, 'train/grad_norm_pre_clip': 12.020350456237793, 'timing/data': 0.010961512569338083, 'timing/model': 1.5296051963232458, 'learning_rate/qwen_vl_interface': 8.591485599934786e-06, 'learning_rate/action_model': 8.591485599934785e-05, 'epoch': 0.54}) +[2026-06-15 23:54:56,362][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.6773592233657837, 'train/grad_norm_pre_clip': 22.15401840209961, 'timing/data': 0.01156329596415162, 'timing/model': 1.5181182459928095, 'learning_rate/qwen_vl_interface': 8.58932045825357e-06, 'learning_rate/action_model': 8.589320458253569e-05, 'epoch': 0.54}) +[2026-06-15 23:55:02,666][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.09207441657781601, 'train/grad_norm_pre_clip': 15.326395034790039, 'timing/data': 0.011499494779855013, 'timing/model': 1.7697664238512516, 'learning_rate/qwen_vl_interface': 8.58715394390385e-06, 'learning_rate/action_model': 8.58715394390385e-05, 'epoch': 0.54}) +[2026-06-15 23:55:08,760][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.2766024172306061, 'train/grad_norm_pre_clip': 8.070170402526855, 'timing/data': 0.011835306882858276, 'timing/model': 1.5577756948769093, 'learning_rate/qwen_vl_interface': 8.5849860577762e-06, 'learning_rate/action_model': 8.5849860577762e-05, 'epoch': 0.54}) +[2026-06-15 23:55:15,065][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.22058658301830292, 'train/grad_norm_pre_clip': 15.790064811706543, 'timing/data': 0.011658207047730684, 'timing/model': 1.5629402799531817, 'learning_rate/qwen_vl_interface': 8.582816800761756e-06, 'learning_rate/action_model': 8.582816800761755e-05, 'epoch': 0.54}) +[2026-06-15 23:55:21,458][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.37725478410720825, 'train/grad_norm_pre_clip': 12.699135780334473, 'timing/data': 0.011012133676558733, 'timing/model': 1.5659139677882195, 'learning_rate/qwen_vl_interface': 8.580646173752214e-06, 'learning_rate/action_model': 8.580646173752213e-05, 'epoch': 0.54}) +[2026-06-15 23:55:27,853][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.1535661369562149, 'train/grad_norm_pre_clip': 4.363565921783447, 'timing/data': 0.011595176067203283, 'timing/model': 1.5929003106430173, 'learning_rate/qwen_vl_interface': 8.578474177639841e-06, 'learning_rate/action_model': 8.578474177639842e-05, 'epoch': 0.54}) +[2026-06-15 23:55:34,207][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.07047425955533981, 'train/grad_norm_pre_clip': 6.0240325927734375, 'timing/data': 0.01095104357227683, 'timing/model': 1.8079246771521866, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 0.54}) +[2026-06-15 23:55:40,259][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.1299225091934204, 'train/grad_norm_pre_clip': 14.449280738830566, 'timing/data': 0.010966824367642403, 'timing/model': 1.5564637160860002, 'learning_rate/qwen_vl_interface': 8.574126081678463e-06, 'learning_rate/action_model': 8.574126081678462e-05, 'epoch': 0.54}) +[2026-06-15 23:55:46,557][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.0717172771692276, 'train/grad_norm_pre_clip': 5.032807350158691, 'timing/data': 0.011136244051158428, 'timing/model': 1.5681916777975857, 'learning_rate/qwen_vl_interface': 8.571949983616796e-06, 'learning_rate/action_model': 8.571949983616795e-05, 'epoch': 0.54}) +[2026-06-15 23:55:52,755][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.11628533899784088, 'train/grad_norm_pre_clip': 3.5858101844787598, 'timing/data': 0.010258270893245935, 'timing/model': 1.554685639217496, 'learning_rate/qwen_vl_interface': 8.569772520026971e-06, 'learning_rate/action_model': 8.569772520026971e-05, 'epoch': 0.54}) +[2026-06-15 23:55:59,054][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.20978662371635437, 'train/grad_norm_pre_clip': 16.751970291137695, 'timing/data': 0.010697592981159687, 'timing/model': 1.5520755830220878, 'learning_rate/qwen_vl_interface': 8.567593691804061e-06, 'learning_rate/action_model': 8.567593691804061e-05, 'epoch': 0.54}) +[2026-06-15 23:56:05,355][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.10844247043132782, 'train/grad_norm_pre_clip': 6.965167999267578, 'timing/data': 0.01074793329462409, 'timing/model': 1.5694148670881987, 'learning_rate/qwen_vl_interface': 8.565413499843703e-06, 'learning_rate/action_model': 8.565413499843703e-05, 'epoch': 0.54}) +[2026-06-15 23:56:11,694][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.08591452986001968, 'train/grad_norm_pre_clip': 7.424296855926514, 'timing/data': 0.011003794148564339, 'timing/model': 1.8006662530824542, 'learning_rate/qwen_vl_interface': 8.563231945042087e-06, 'learning_rate/action_model': 8.563231945042088e-05, 'epoch': 0.54}) +[2026-06-15 23:56:17,787][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.07114750146865845, 'train/grad_norm_pre_clip': 4.438152313232422, 'timing/data': 0.010573843028396368, 'timing/model': 1.5678732600063086, 'learning_rate/qwen_vl_interface': 8.56104902829597e-06, 'learning_rate/action_model': 8.56104902829597e-05, 'epoch': 0.54}) +[2026-06-15 23:56:24,065][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.04064570739865303, 'train/grad_norm_pre_clip': 4.950562953948975, 'timing/data': 0.011038355063647032, 'timing/model': 1.562921334989369, 'learning_rate/qwen_vl_interface': 8.558864750502669e-06, 'learning_rate/action_model': 8.558864750502668e-05, 'epoch': 0.55}) +[2026-06-15 23:56:30,359][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.048605624586343765, 'train/grad_norm_pre_clip': 3.7667250633239746, 'timing/data': 0.010870472993701696, 'timing/model': 1.5592451160773635, 'learning_rate/qwen_vl_interface': 8.556679112560055e-06, 'learning_rate/action_model': 8.556679112560054e-05, 'epoch': 0.55}) +[2026-06-15 23:56:36,664][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.09470418095588684, 'train/grad_norm_pre_clip': 9.336048126220703, 'timing/data': 0.011209855321794748, 'timing/model': 1.552313611842692, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 0.55}) +[2026-06-15 23:56:42,969][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.22116388380527496, 'train/grad_norm_pre_clip': 5.449916839599609, 'timing/data': 0.011143545154482126, 'timing/model': 1.5411072610877454, 'learning_rate/qwen_vl_interface': 8.55230375982118e-06, 'learning_rate/action_model': 8.55230375982118e-05, 'epoch': 0.55}) +[2026-06-15 23:56:49,304][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.11129025369882584, 'train/grad_norm_pre_clip': 6.9514265060424805, 'timing/data': 0.01374875521287322, 'timing/model': 1.7944068992510438, 'learning_rate/qwen_vl_interface': 8.550114046823463e-06, 'learning_rate/action_model': 8.550114046823463e-05, 'epoch': 0.55}) +[2026-06-15 23:56:55,569][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.1531001776456833, 'train/grad_norm_pre_clip': 8.365144729614258, 'timing/data': 0.011141314636915922, 'timing/model': 1.570306328125298, 'learning_rate/qwen_vl_interface': 8.547922977273518e-06, 'learning_rate/action_model': 8.547922977273518e-05, 'epoch': 0.55}) +[2026-06-15 23:57:02,016][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.2537083327770233, 'train/grad_norm_pre_clip': 10.399982452392578, 'timing/data': 0.010812052991241217, 'timing/model': 1.5809213859029114, 'learning_rate/qwen_vl_interface': 8.54573055207201e-06, 'learning_rate/action_model': 8.545730552072008e-05, 'epoch': 0.55}) +[2026-06-15 23:57:08,569][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.09218811988830566, 'train/grad_norm_pre_clip': 10.777125358581543, 'timing/data': 0.011299735866487026, 'timing/model': 1.5747525347396731, 'learning_rate/qwen_vl_interface': 8.543536772120161e-06, 'learning_rate/action_model': 8.54353677212016e-05, 'epoch': 0.55}) +[2026-06-15 23:57:14,873][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.1280621439218521, 'train/grad_norm_pre_clip': 6.966823101043701, 'timing/data': 0.011227484792470932, 'timing/model': 1.5421405569650233, 'learning_rate/qwen_vl_interface': 8.541341638319754e-06, 'learning_rate/action_model': 8.541341638319754e-05, 'epoch': 0.55}) +[2026-06-15 23:57:21,160][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.032770540565252304, 'train/grad_norm_pre_clip': 5.469846725463867, 'timing/data': 0.011228635907173157, 'timing/model': 1.546426938381046, 'learning_rate/qwen_vl_interface': 8.539145151573126e-06, 'learning_rate/action_model': 8.539145151573125e-05, 'epoch': 0.55}) +[2026-06-15 23:57:27,494][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.11366808414459229, 'train/grad_norm_pre_clip': 8.959708213806152, 'timing/data': 0.010990413837134838, 'timing/model': 1.793519985396415, 'learning_rate/qwen_vl_interface': 8.536947312783167e-06, 'learning_rate/action_model': 8.536947312783166e-05, 'epoch': 0.55}) +[2026-06-15 23:57:33,558][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.16400712728500366, 'train/grad_norm_pre_clip': 5.647512435913086, 'timing/data': 0.010673763230443, 'timing/model': 1.5470742699690163, 'learning_rate/qwen_vl_interface': 8.534748122853326e-06, 'learning_rate/action_model': 8.534748122853325e-05, 'epoch': 0.55}) +[2026-06-15 23:57:39,862][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.12556833028793335, 'train/grad_norm_pre_clip': 5.378664970397949, 'timing/data': 0.013685164973139763, 'timing/model': 1.5509225600399077, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.55}) +[2026-06-15 23:57:46,247][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.04337305948138237, 'train/grad_norm_pre_clip': 2.1577093601226807, 'timing/data': 0.011211855802685022, 'timing/model': 1.6469160127453506, 'learning_rate/qwen_vl_interface': 8.530345693190573e-06, 'learning_rate/action_model': 8.530345693190572e-05, 'epoch': 0.55}) +[2026-06-15 23:57:52,569][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.022012146189808846, 'train/grad_norm_pre_clip': 2.1881680488586426, 'timing/data': 0.010885274037718773, 'timing/model': 1.5593981118872762, 'learning_rate/qwen_vl_interface': 8.528142455267328e-06, 'learning_rate/action_model': 8.528142455267329e-05, 'epoch': 0.55}) +[2026-06-15 23:57:58,868][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.11176463216543198, 'train/grad_norm_pre_clip': 7.574152946472168, 'timing/data': 0.011107145342975855, 'timing/model': 1.5256136893294752, 'learning_rate/qwen_vl_interface': 8.52593786982355e-06, 'learning_rate/action_model': 8.525937869823551e-05, 'epoch': 0.55}) +[2026-06-15 23:58:05,272][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.1475159227848053, 'train/grad_norm_pre_clip': 12.139266014099121, 'timing/data': 0.011109744664281607, 'timing/model': 1.8562970091588795, 'learning_rate/qwen_vl_interface': 8.523731937765456e-06, 'learning_rate/action_model': 8.523731937765456e-05, 'epoch': 0.55}) +[2026-06-15 23:58:11,374][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.025227073580026627, 'train/grad_norm_pre_clip': 1.7407562732696533, 'timing/data': 0.012428631074726582, 'timing/model': 1.5473029338754714, 'learning_rate/qwen_vl_interface': 8.521524659999823e-06, 'learning_rate/action_model': 8.521524659999822e-05, 'epoch': 0.55}) +[2026-06-15 23:58:17,674][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.10698158293962479, 'train/grad_norm_pre_clip': 4.834968566894531, 'timing/data': 0.010903083719313145, 'timing/model': 1.5685228239744902, 'learning_rate/qwen_vl_interface': 8.519316037433976e-06, 'learning_rate/action_model': 8.519316037433977e-05, 'epoch': 0.55}) +[2026-06-15 23:58:23,960][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.5243242979049683, 'train/grad_norm_pre_clip': 11.32884407043457, 'timing/data': 0.010879415087401867, 'timing/model': 1.562818774022162, 'learning_rate/qwen_vl_interface': 8.517106070975801e-06, 'learning_rate/action_model': 8.5171060709758e-05, 'epoch': 0.55}) +[2026-06-15 23:58:30,264][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.2697274386882782, 'train/grad_norm_pre_clip': 14.201859474182129, 'timing/data': 0.010937975253909826, 'timing/model': 1.5654462990351021, 'learning_rate/qwen_vl_interface': 8.514894761533729e-06, 'learning_rate/action_model': 8.514894761533728e-05, 'epoch': 0.55}) +[2026-06-15 23:58:36,552][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.15837667882442474, 'train/grad_norm_pre_clip': 4.745169162750244, 'timing/data': 0.010868954937905073, 'timing/model': 1.525285858195275, 'learning_rate/qwen_vl_interface': 8.512682110016747e-06, 'learning_rate/action_model': 8.512682110016747e-05, 'epoch': 0.55}) +[2026-06-15 23:58:42,897][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.017779529094696045, 'train/grad_norm_pre_clip': 3.3529529571533203, 'timing/data': 0.01017472194507718, 'timing/model': 1.7800253932364285, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 0.55}) +[2026-06-15 23:58:48,969][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.1827755570411682, 'train/grad_norm_pre_clip': 11.696789741516113, 'timing/data': 0.011595367919653654, 'timing/model': 1.5665069138631225, 'learning_rate/qwen_vl_interface': 8.50825278439675e-06, 'learning_rate/action_model': 8.508252784396749e-05, 'epoch': 0.55}) +[2026-06-15 23:58:55,262][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.06173231825232506, 'train/grad_norm_pre_clip': 13.197482109069824, 'timing/data': 0.010807914193719625, 'timing/model': 1.559136806987226, 'learning_rate/qwen_vl_interface': 8.506036112114465e-06, 'learning_rate/action_model': 8.506036112114464e-05, 'epoch': 0.55}) +[2026-06-15 23:59:01,567][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.05006743222475052, 'train/grad_norm_pre_clip': 21.84521484375, 'timing/data': 0.01169360801577568, 'timing/model': 1.5574910836294293, 'learning_rate/qwen_vl_interface': 8.503818101398725e-06, 'learning_rate/action_model': 8.503818101398725e-05, 'epoch': 0.56}) +[2026-06-15 23:59:07,863][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.24122756719589233, 'train/grad_norm_pre_clip': 16.80916976928711, 'timing/data': 0.011083455290645361, 'timing/model': 1.5640568137168884, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.56}) +[2026-06-15 23:59:14,245][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.09792264550924301, 'train/grad_norm_pre_clip': 2.9280409812927246, 'timing/data': 0.01127779670059681, 'timing/model': 1.5920607219450176, 'learning_rate/qwen_vl_interface': 8.499378068314395e-06, 'learning_rate/action_model': 8.499378068314395e-05, 'epoch': 0.56}) +[2026-06-15 23:59:20,595][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.11278436332941055, 'train/grad_norm_pre_clip': 5.764167785644531, 'timing/data': 0.010316031984984875, 'timing/model': 1.7887736828997731, 'learning_rate/qwen_vl_interface': 8.497156047770932e-06, 'learning_rate/action_model': 8.497156047770933e-05, 'epoch': 0.56}) +[2026-06-15 23:59:26,672][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.15239986777305603, 'train/grad_norm_pre_clip': 8.351670265197754, 'timing/data': 0.01109555596485734, 'timing/model': 1.5655703502707183, 'learning_rate/qwen_vl_interface': 8.494932692444276e-06, 'learning_rate/action_model': 8.494932692444276e-05, 'epoch': 0.56}) +[2026-06-15 23:59:32,959][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.42350009083747864, 'train/grad_norm_pre_clip': 27.63384246826172, 'timing/data': 0.011161885689944029, 'timing/model': 1.5606226529926062, 'learning_rate/qwen_vl_interface': 8.492708003248362e-06, 'learning_rate/action_model': 8.492708003248362e-05, 'epoch': 0.56}) +[2026-06-15 23:59:39,257][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.12674283981323242, 'train/grad_norm_pre_clip': 16.891212463378906, 'timing/data': 0.010941915214061737, 'timing/model': 1.5627264161594212, 'learning_rate/qwen_vl_interface': 8.490481981097674e-06, 'learning_rate/action_model': 8.490481981097674e-05, 'epoch': 0.56}) +[2026-06-15 23:59:45,553][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.047857217490673065, 'train/grad_norm_pre_clip': 13.195157051086426, 'timing/data': 0.009735379833728075, 'timing/model': 1.5606631799601018, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 0.56}) +[2026-06-15 23:59:51,857][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.37530624866485596, 'train/grad_norm_pre_clip': 18.663333892822266, 'timing/data': 0.010007201228290796, 'timing/model': 1.5304624089039862, 'learning_rate/qwen_vl_interface': 8.486025941592664e-06, 'learning_rate/action_model': 8.486025941592664e-05, 'epoch': 0.56}) +[2026-06-15 23:59:58,321][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.021932603791356087, 'train/grad_norm_pre_clip': 2.063636541366577, 'timing/data': 0.010058551095426083, 'timing/model': 1.9113891078159213, 'learning_rate/qwen_vl_interface': 8.483795926070054e-06, 'learning_rate/action_model': 8.483795926070052e-05, 'epoch': 0.56}) +[2026-06-16 00:00:04,541][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.08569949865341187, 'train/grad_norm_pre_clip': 10.75151252746582, 'timing/data': 0.011012695264071226, 'timing/model': 1.7469202051870525, 'learning_rate/qwen_vl_interface': 8.481564581256087e-06, 'learning_rate/action_model': 8.481564581256087e-05, 'epoch': 0.56}) +[2026-06-16 00:00:10,580][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.06472145020961761, 'train/grad_norm_pre_clip': 9.219715118408203, 'timing/data': 0.01132931699976325, 'timing/model': 1.5739653571508825, 'learning_rate/qwen_vl_interface': 8.479331908067987e-06, 'learning_rate/action_model': 8.479331908067987e-05, 'epoch': 0.56}) +[2026-06-16 00:00:16,956][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.12220697849988937, 'train/grad_norm_pre_clip': 2.024996519088745, 'timing/data': 0.01103723468258977, 'timing/model': 1.6521530118770897, 'learning_rate/qwen_vl_interface': 8.477097907423524e-06, 'learning_rate/action_model': 8.477097907423524e-05, 'epoch': 0.56}) +[2026-06-16 00:00:23,374][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.033973243087530136, 'train/grad_norm_pre_clip': 6.981905937194824, 'timing/data': 0.012243860401213169, 'timing/model': 1.5568725932389498, 'learning_rate/qwen_vl_interface': 8.474862580241006e-06, 'learning_rate/action_model': 8.474862580241006e-05, 'epoch': 0.56}) +[2026-06-16 00:00:29,675][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.1874827742576599, 'train/grad_norm_pre_clip': 4.8916473388671875, 'timing/data': 0.011314217001199722, 'timing/model': 1.5264449403621256, 'learning_rate/qwen_vl_interface': 8.472625927439299e-06, 'learning_rate/action_model': 8.472625927439299e-05, 'epoch': 0.56}) +[2026-06-16 00:00:36,006][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.3897823691368103, 'train/grad_norm_pre_clip': 11.329726219177246, 'timing/data': 0.010756994131952524, 'timing/model': 1.8078130641952157, 'learning_rate/qwen_vl_interface': 8.4703879499378e-06, 'learning_rate/action_model': 8.4703879499378e-05, 'epoch': 0.56}) +[2026-06-16 00:00:42,309][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.16788522899150848, 'train/grad_norm_pre_clip': 7.433790683746338, 'timing/data': 0.009903959929943085, 'timing/model': 1.8107944591902196, 'learning_rate/qwen_vl_interface': 8.46814864865646e-06, 'learning_rate/action_model': 8.46814864865646e-05, 'epoch': 0.56}) +[2026-06-16 00:00:48,359][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.17285728454589844, 'train/grad_norm_pre_clip': 8.815862655639648, 'timing/data': 0.011510627809911966, 'timing/model': 1.5582487699575722, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 0.56}) +[2026-06-16 00:00:54,665][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.23488956689834595, 'train/grad_norm_pre_clip': 7.168096542358398, 'timing/data': 0.011093014851212502, 'timing/model': 1.5625702310353518, 'learning_rate/qwen_vl_interface': 8.463666078436769e-06, 'learning_rate/action_model': 8.463666078436769e-05, 'epoch': 0.56}) +[2026-06-16 00:01:00,965][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.0843229666352272, 'train/grad_norm_pre_clip': 4.4098005294799805, 'timing/data': 0.01147273788228631, 'timing/model': 1.5634270356968045, 'learning_rate/qwen_vl_interface': 8.46142281134103e-06, 'learning_rate/action_model': 8.461422811341031e-05, 'epoch': 0.56}) +[2026-06-16 00:01:07,277][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.08880426734685898, 'train/grad_norm_pre_clip': 11.047708511352539, 'timing/data': 0.011281576007604599, 'timing/model': 1.5432671871967614, 'learning_rate/qwen_vl_interface': 8.459178224150683e-06, 'learning_rate/action_model': 8.459178224150683e-05, 'epoch': 0.56}) +[2026-06-16 00:01:13,592][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.15476535260677338, 'train/grad_norm_pre_clip': 10.913642883300781, 'timing/data': 0.011102404911071062, 'timing/model': 1.7930455822497606, 'learning_rate/qwen_vl_interface': 8.456932317788389e-06, 'learning_rate/action_model': 8.456932317788389e-05, 'epoch': 0.56}) +[2026-06-16 00:01:19,667][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.0790855810046196, 'train/grad_norm_pre_clip': 5.302093982696533, 'timing/data': 0.010471623856574297, 'timing/model': 1.5698737120255828, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 0.56}) +[2026-06-16 00:01:25,966][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.06251068413257599, 'train/grad_norm_pre_clip': 2.5539889335632324, 'timing/data': 0.01087562507018447, 'timing/model': 1.5721674738451838, 'learning_rate/qwen_vl_interface': 8.452436551241328e-06, 'learning_rate/action_model': 8.452436551241328e-05, 'epoch': 0.56}) +[2026-06-16 00:01:32,276][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.14034481346607208, 'train/grad_norm_pre_clip': 11.860706329345703, 'timing/data': 0.011074015870690346, 'timing/model': 1.5736515522003174, 'learning_rate/qwen_vl_interface': 8.450186692904605e-06, 'learning_rate/action_model': 8.450186692904606e-05, 'epoch': 0.56}) +[2026-06-16 00:01:38,560][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.2922224998474121, 'train/grad_norm_pre_clip': 12.484957695007324, 'timing/data': 0.010863465256989002, 'timing/model': 1.521817249711603, 'learning_rate/qwen_vl_interface': 8.447935519092014e-06, 'learning_rate/action_model': 8.447935519092014e-05, 'epoch': 0.57}) +[2026-06-16 00:01:44,870][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.08731964975595474, 'train/grad_norm_pre_clip': 5.570126533508301, 'timing/data': 0.0121497199870646, 'timing/model': 1.5942831728607416, 'learning_rate/qwen_vl_interface': 8.445683030728925e-06, 'learning_rate/action_model': 8.445683030728925e-05, 'epoch': 0.57}) +[2026-06-16 00:12:39,426][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.1419016569852829, 'train/grad_norm_pre_clip': 11.587531089782715, 'eval/loss': 0.07577342540025711, 'eval/flap_precision': 0.9238095238095239, 'eval/flap_recall': 0.9263301500682128, 'eval/flap_f1': 0.9250681198910082, 'eval/flap_rate_model': 0.20416666666666666, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9250681198910082, 'timing/data': 0.011528157629072666, 'timing/model': 1.736383724026382, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 0.57}) +[2026-06-16 00:12:39,427][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1400_state +[2026-06-16 00:12:39,434][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 00:12:48,005][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1400_state/model.safetensors +[2026-06-16 00:13:07,844][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1400_state/optimizer.bin +[2026-06-16 00:13:07,845][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1400_state/sampler.bin +[2026-06-16 00:13:07,845][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1400_state/sampler_1.bin +[2026-06-16 00:13:07,847][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1400_state/random_states_0.pkl +[2026-06-16 00:13:08,048][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-06-16 00:13:08,049][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 00:13:08,064][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 00:13:14,417][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.13958740234375, 'train/grad_norm_pre_clip': 3.9939818382263184, 'timing/data': 0.01139098685234785, 'timing/model': 1.8413960742764175, 'learning_rate/qwen_vl_interface': 8.441174114055457e-06, 'learning_rate/action_model': 8.441174114055456e-05, 'epoch': 0.57}) +[2026-06-16 00:13:20,469][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.2871306836605072, 'train/grad_norm_pre_clip': 10.664815902709961, 'timing/data': 0.013853129930794239, 'timing/model': 1.5873417579568923, 'learning_rate/qwen_vl_interface': 8.43891768759852e-06, 'learning_rate/action_model': 8.43891768759852e-05, 'epoch': 0.57}) +[2026-06-16 00:13:26,735][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.07922087609767914, 'train/grad_norm_pre_clip': 13.051301002502441, 'timing/data': 0.011100455652922392, 'timing/model': 1.565273916348815, 'learning_rate/qwen_vl_interface': 8.436659950297978e-06, 'learning_rate/action_model': 8.436659950297977e-05, 'epoch': 0.57}) +[2026-06-16 00:13:32,943][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.1612684428691864, 'train/grad_norm_pre_clip': 6.033851623535156, 'timing/data': 0.011489816941320896, 'timing/model': 1.5681703193113208, 'learning_rate/qwen_vl_interface': 8.4344009030819e-06, 'learning_rate/action_model': 8.4344009030819e-05, 'epoch': 0.57}) +[2026-06-16 00:13:39,048][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.15729883313179016, 'train/grad_norm_pre_clip': 10.908426284790039, 'timing/data': 0.01221498055383563, 'timing/model': 1.57093719067052, 'learning_rate/qwen_vl_interface': 8.432140546878899e-06, 'learning_rate/action_model': 8.432140546878898e-05, 'epoch': 0.57}) +[2026-06-16 00:13:45,345][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.26030048727989197, 'train/grad_norm_pre_clip': 5.458954811096191, 'timing/data': 0.011768778786063194, 'timing/model': 1.5786353978328407, 'learning_rate/qwen_vl_interface': 8.429878882618116e-06, 'learning_rate/action_model': 8.429878882618116e-05, 'epoch': 0.57}) +[2026-06-16 00:13:51,662][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.0750446766614914, 'train/grad_norm_pre_clip': 8.158249855041504, 'timing/data': 0.010125970002263784, 'timing/model': 1.7804840877652168, 'learning_rate/qwen_vl_interface': 8.42761591122924e-06, 'learning_rate/action_model': 8.427615911229241e-05, 'epoch': 0.57}) +[2026-06-16 00:13:57,745][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.03417392820119858, 'train/grad_norm_pre_clip': 7.704128265380859, 'timing/data': 0.010826784186065197, 'timing/model': 1.567807471845299, 'learning_rate/qwen_vl_interface': 8.425351633642491e-06, 'learning_rate/action_model': 8.425351633642492e-05, 'epoch': 0.57}) +[2026-06-16 00:14:04,141][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.08260346949100494, 'train/grad_norm_pre_clip': 2.3227438926696777, 'timing/data': 0.01152484817430377, 'timing/model': 1.5683564627543092, 'learning_rate/qwen_vl_interface': 8.423086050788632e-06, 'learning_rate/action_model': 8.423086050788631e-05, 'epoch': 0.57}) +[2026-06-16 00:14:10,353][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.16827967762947083, 'train/grad_norm_pre_clip': 10.77343463897705, 'timing/data': 0.011389296967536211, 'timing/model': 1.5617685592733324, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 0.57}) +[2026-06-16 00:14:16,666][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.042987722903490067, 'train/grad_norm_pre_clip': 2.3884921073913574, 'timing/data': 0.011429877020418644, 'timing/model': 1.5508966450579464, 'learning_rate/qwen_vl_interface': 8.418550973005289e-06, 'learning_rate/action_model': 8.418550973005289e-05, 'epoch': 0.57}) +[2026-06-16 00:14:23,059][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.07131123542785645, 'train/grad_norm_pre_clip': 3.398909568786621, 'timing/data': 0.012586362194269896, 'timing/model': 1.596535971853882, 'learning_rate/qwen_vl_interface': 8.416281479940005e-06, 'learning_rate/action_model': 8.416281479940005e-05, 'epoch': 0.57}) +[2026-06-16 00:14:29,485][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.13943538069725037, 'train/grad_norm_pre_clip': 6.394433975219727, 'timing/data': 0.011720257811248302, 'timing/model': 1.7965840250253677, 'learning_rate/qwen_vl_interface': 8.414010685336008e-06, 'learning_rate/action_model': 8.414010685336008e-05, 'epoch': 0.57}) +[2026-06-16 00:14:35,732][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.008403210900723934, 'train/grad_norm_pre_clip': 1.509720802307129, 'timing/data': 0.011363646015524864, 'timing/model': 1.6350910309702158, 'learning_rate/qwen_vl_interface': 8.411738590126733e-06, 'learning_rate/action_model': 8.411738590126732e-05, 'epoch': 0.57}) +[2026-06-16 00:14:41,951][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.12225311994552612, 'train/grad_norm_pre_clip': 6.259674072265625, 'timing/data': 0.011199034750461578, 'timing/model': 1.5496424180455506, 'learning_rate/qwen_vl_interface': 8.409465195246152e-06, 'learning_rate/action_model': 8.409465195246152e-05, 'epoch': 0.57}) +[2026-06-16 00:14:48,281][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.23137232661247253, 'train/grad_norm_pre_clip': 13.139558792114258, 'timing/data': 0.011167516000568867, 'timing/model': 1.5738734970800579, 'learning_rate/qwen_vl_interface': 8.407190501628772e-06, 'learning_rate/action_model': 8.407190501628772e-05, 'epoch': 0.57}) +[2026-06-16 00:14:54,604][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.2598678171634674, 'train/grad_norm_pre_clip': 15.251067161560059, 'timing/data': 0.011393957305699587, 'timing/model': 1.6129408599808812, 'learning_rate/qwen_vl_interface': 8.404914510209635e-06, 'learning_rate/action_model': 8.404914510209635e-05, 'epoch': 0.57}) +[2026-06-16 00:15:00,854][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.19486922025680542, 'train/grad_norm_pre_clip': 10.050410270690918, 'timing/data': 0.010776703245937824, 'timing/model': 1.5473084771074355, 'learning_rate/qwen_vl_interface': 8.402637221924315e-06, 'learning_rate/action_model': 8.402637221924314e-05, 'epoch': 0.57}) +[2026-06-16 00:15:07,203][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.20343178510665894, 'train/grad_norm_pre_clip': 11.36691665649414, 'timing/data': 0.011414825916290283, 'timing/model': 1.811878011096269, 'learning_rate/qwen_vl_interface': 8.400358637708919e-06, 'learning_rate/action_model': 8.400358637708918e-05, 'epoch': 0.57}) +[2026-06-16 00:15:13,254][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.2660905122756958, 'train/grad_norm_pre_clip': 11.56479549407959, 'timing/data': 0.010902363806962967, 'timing/model': 1.5701762596145272, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 0.57}) +[2026-06-16 00:15:19,549][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.11432152986526489, 'train/grad_norm_pre_clip': 14.46035099029541, 'timing/data': 0.01066236337646842, 'timing/model': 1.5516873290762305, 'learning_rate/qwen_vl_interface': 8.395797585234987e-06, 'learning_rate/action_model': 8.395797585234988e-05, 'epoch': 0.57}) +[2026-06-16 00:15:25,857][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.09141674637794495, 'train/grad_norm_pre_clip': 4.78115701675415, 'timing/data': 0.011279236059635878, 'timing/model': 1.5667428337037563, 'learning_rate/qwen_vl_interface': 8.39351511885133e-06, 'learning_rate/action_model': 8.393515118851329e-05, 'epoch': 0.58}) +[2026-06-16 00:15:32,160][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.25685426592826843, 'train/grad_norm_pre_clip': 2.5344135761260986, 'timing/data': 0.010532571002840996, 'timing/model': 1.520837530028075, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.58}) +[2026-06-16 00:15:38,450][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.37044790387153625, 'train/grad_norm_pre_clip': 14.639720916748047, 'timing/data': 0.010115070268511772, 'timing/model': 1.5472034588456154, 'learning_rate/qwen_vl_interface': 8.388946310481807e-06, 'learning_rate/action_model': 8.388946310481807e-05, 'epoch': 0.58}) +[2026-06-16 00:15:44,785][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.4572685956954956, 'train/grad_norm_pre_clip': 7.763149738311768, 'timing/data': 0.011374906171113253, 'timing/model': 1.7869890141300857, 'learning_rate/qwen_vl_interface': 8.386659970374007e-06, 'learning_rate/action_model': 8.386659970374006e-05, 'epoch': 0.58}) +[2026-06-16 00:15:50,945][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.21658176183700562, 'train/grad_norm_pre_clip': 6.9472336769104, 'timing/data': 0.011058483738452196, 'timing/model': 1.648147906176746, 'learning_rate/qwen_vl_interface': 8.384372340903775e-06, 'learning_rate/action_model': 8.384372340903775e-05, 'epoch': 0.58}) +[2026-06-16 00:15:57,151][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.08207108825445175, 'train/grad_norm_pre_clip': 6.897251129150391, 'timing/data': 0.011803318280726671, 'timing/model': 1.5594608527608216, 'learning_rate/qwen_vl_interface': 8.382083423011472e-06, 'learning_rate/action_model': 8.382083423011471e-05, 'epoch': 0.58}) +[2026-06-16 00:16:03,460][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.051711585372686386, 'train/grad_norm_pre_clip': 3.612978458404541, 'timing/data': 0.011498555541038513, 'timing/model': 1.5681121270172298, 'learning_rate/qwen_vl_interface': 8.379793217637979e-06, 'learning_rate/action_model': 8.379793217637979e-05, 'epoch': 0.58}) +[2026-06-16 00:16:09,756][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.05323643982410431, 'train/grad_norm_pre_clip': 5.905776023864746, 'timing/data': 0.010970984119921923, 'timing/model': 1.5544487619772553, 'learning_rate/qwen_vl_interface': 8.377501725724719e-06, 'learning_rate/action_model': 8.377501725724719e-05, 'epoch': 0.58}) +[2026-06-16 00:16:16,061][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.046175070106983185, 'train/grad_norm_pre_clip': 4.99758768081665, 'timing/data': 0.01106372382491827, 'timing/model': 1.5156934252008796, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.58}) +[2026-06-16 00:16:22,386][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.05307331681251526, 'train/grad_norm_pre_clip': 3.9103281497955322, 'timing/data': 0.011197094805538654, 'timing/model': 1.7860072110779583, 'learning_rate/qwen_vl_interface': 8.372914886047203e-06, 'learning_rate/action_model': 8.372914886047203e-05, 'epoch': 0.58}) +[2026-06-16 00:16:28,455][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.2285403460264206, 'train/grad_norm_pre_clip': 4.568160057067871, 'timing/data': 0.011101133655756712, 'timing/model': 1.5570309893228114, 'learning_rate/qwen_vl_interface': 8.370619540168422e-06, 'learning_rate/action_model': 8.370619540168423e-05, 'epoch': 0.58}) +[2026-06-16 00:16:34,757][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.07428379356861115, 'train/grad_norm_pre_clip': 4.64171028137207, 'timing/data': 0.012321678921580315, 'timing/model': 1.5647412394173443, 'learning_rate/qwen_vl_interface': 8.368322911520826e-06, 'learning_rate/action_model': 8.368322911520826e-05, 'epoch': 0.58}) +[2026-06-16 00:16:40,956][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.07502887398004532, 'train/grad_norm_pre_clip': 5.901916980743408, 'timing/data': 0.010849292855709791, 'timing/model': 1.5647304309532046, 'learning_rate/qwen_vl_interface': 8.36602500104847e-06, 'learning_rate/action_model': 8.36602500104847e-05, 'epoch': 0.58}) +[2026-06-16 00:16:47,253][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.15891015529632568, 'train/grad_norm_pre_clip': 14.145952224731445, 'timing/data': 0.011477645952254534, 'timing/model': 1.5510082659311593, 'learning_rate/qwen_vl_interface': 8.36372580969594e-06, 'learning_rate/action_model': 8.36372580969594e-05, 'epoch': 0.58}) +[2026-06-16 00:16:53,552][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.10831177979707718, 'train/grad_norm_pre_clip': 3.7341725826263428, 'timing/data': 0.012121038977056742, 'timing/model': 1.5168913342058659, 'learning_rate/qwen_vl_interface': 8.361425338408344e-06, 'learning_rate/action_model': 8.361425338408343e-05, 'epoch': 0.58}) +[2026-06-16 00:16:59,885][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.2674907147884369, 'train/grad_norm_pre_clip': 14.401928901672363, 'timing/data': 0.011410464998334646, 'timing/model': 1.774847062304616, 'learning_rate/qwen_vl_interface': 8.359123588131319e-06, 'learning_rate/action_model': 8.359123588131319e-05, 'epoch': 0.58}) +[2026-06-16 00:17:05,958][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.13690997660160065, 'train/grad_norm_pre_clip': 8.724785804748535, 'timing/data': 0.01079638209193945, 'timing/model': 1.5669560409151018, 'learning_rate/qwen_vl_interface': 8.35682055981103e-06, 'learning_rate/action_model': 8.35682055981103e-05, 'epoch': 0.58}) +[2026-06-16 00:17:12,262][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.1153881624341011, 'train/grad_norm_pre_clip': 9.987651824951172, 'timing/data': 0.011894026771187782, 'timing/model': 1.5623252913355827, 'learning_rate/qwen_vl_interface': 8.354516254394163e-06, 'learning_rate/action_model': 8.354516254394162e-05, 'epoch': 0.58}) +[2026-06-16 00:17:18,636][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.047838013619184494, 'train/grad_norm_pre_clip': 6.2112908363342285, 'timing/data': 0.010895393323153257, 'timing/model': 1.6311081242747605, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 0.58}) +[2026-06-16 00:17:24,869][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.029238810762763023, 'train/grad_norm_pre_clip': 3.758380651473999, 'timing/data': 0.01198769686743617, 'timing/model': 1.560799009166658, 'learning_rate/qwen_vl_interface': 8.349903816060074e-06, 'learning_rate/action_model': 8.349903816060075e-05, 'epoch': 0.58}) +[2026-06-16 00:17:31,159][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.1724056601524353, 'train/grad_norm_pre_clip': 10.855321884155273, 'timing/data': 0.011387194972485304, 'timing/model': 1.5243927757255733, 'learning_rate/qwen_vl_interface': 8.34759568503885e-06, 'learning_rate/action_model': 8.347595685038851e-05, 'epoch': 0.58}) +[2026-06-16 00:17:37,572][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.0765841007232666, 'train/grad_norm_pre_clip': 8.728930473327637, 'timing/data': 0.012318628840148449, 'timing/model': 1.8696501371450722, 'learning_rate/qwen_vl_interface': 8.345286280713049e-06, 'learning_rate/action_model': 8.345286280713048e-05, 'epoch': 0.58}) +[2026-06-16 00:17:43,660][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.12522149085998535, 'train/grad_norm_pre_clip': 12.042682647705078, 'timing/data': 0.010954692028462887, 'timing/model': 1.559799951966852, 'learning_rate/qwen_vl_interface': 8.342975604031974e-06, 'learning_rate/action_model': 8.342975604031974e-05, 'epoch': 0.58}) +[2026-06-16 00:17:50,063][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.1282302737236023, 'train/grad_norm_pre_clip': 10.841891288757324, 'timing/data': 0.013535825069993734, 'timing/model': 1.5857830187305808, 'learning_rate/qwen_vl_interface': 8.340663655945462e-06, 'learning_rate/action_model': 8.340663655945462e-05, 'epoch': 0.58}) +[2026-06-16 00:17:56,360][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.15057967603206635, 'train/grad_norm_pre_clip': 16.614025115966797, 'timing/data': 0.01126320380717516, 'timing/model': 1.5543710007332265, 'learning_rate/qwen_vl_interface': 8.338350437403866e-06, 'learning_rate/action_model': 8.338350437403865e-05, 'epoch': 0.58}) +[2026-06-16 00:18:02,661][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.24378761649131775, 'train/grad_norm_pre_clip': 14.04020881652832, 'timing/data': 0.011896006762981415, 'timing/model': 1.5633344468660653, 'learning_rate/qwen_vl_interface': 8.336035949358063e-06, 'learning_rate/action_model': 8.336035949358062e-05, 'epoch': 0.59}) +[2026-06-16 00:18:09,227][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.13422757387161255, 'train/grad_norm_pre_clip': 17.907777786254883, 'timing/data': 0.011399654671549797, 'timing/model': 1.5203773556277156, 'learning_rate/qwen_vl_interface': 8.333720192759448e-06, 'learning_rate/action_model': 8.333720192759448e-05, 'epoch': 0.59}) +[2026-06-16 00:18:15,614][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.1791924685239792, 'train/grad_norm_pre_clip': 6.726291656494141, 'timing/data': 0.011263133957982063, 'timing/model': 1.8134693875908852, 'learning_rate/qwen_vl_interface': 8.331403168559946e-06, 'learning_rate/action_model': 8.331403168559945e-05, 'epoch': 0.59}) +[2026-06-16 00:18:21,821][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.030719149857759476, 'train/grad_norm_pre_clip': 2.0329368114471436, 'timing/data': 0.01191342668607831, 'timing/model': 1.7214145818725228, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 0.59}) +[2026-06-16 00:18:27,867][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.02328471653163433, 'train/grad_norm_pre_clip': 1.6170762777328491, 'timing/data': 0.010866901837289333, 'timing/model': 1.5636294432915747, 'learning_rate/qwen_vl_interface': 8.32676532116856e-06, 'learning_rate/action_model': 8.326765321168559e-05, 'epoch': 0.59}) +[2026-06-16 00:18:34,162][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.10908911377191544, 'train/grad_norm_pre_clip': 7.440324306488037, 'timing/data': 0.011170642916113138, 'timing/model': 1.5623660399578512, 'learning_rate/qwen_vl_interface': 8.324444499883119e-06, 'learning_rate/action_model': 8.324444499883119e-05, 'epoch': 0.59}) +[2026-06-16 00:18:40,461][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.06668955087661743, 'train/grad_norm_pre_clip': 18.062480926513672, 'timing/data': 0.010230508632957935, 'timing/model': 1.5562923541292548, 'learning_rate/qwen_vl_interface': 8.322122414809678e-06, 'learning_rate/action_model': 8.322122414809677e-05, 'epoch': 0.59}) +[2026-06-16 00:18:46,759][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.39718905091285706, 'train/grad_norm_pre_clip': 25.82525634765625, 'timing/data': 0.011173753067851067, 'timing/model': 1.5182999577373266, 'learning_rate/qwen_vl_interface': 8.319799066902755e-06, 'learning_rate/action_model': 8.319799066902755e-05, 'epoch': 0.59}) +[2026-06-16 00:18:53,090][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.2171398550271988, 'train/grad_norm_pre_clip': 12.105728149414062, 'timing/data': 0.011055022943764925, 'timing/model': 1.7841886249370873, 'learning_rate/qwen_vl_interface': 8.317474457117393e-06, 'learning_rate/action_model': 8.317474457117392e-05, 'epoch': 0.59}) +[2026-06-16 00:18:59,392][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.1570691466331482, 'train/grad_norm_pre_clip': 6.11232852935791, 'timing/data': 0.011262673884630203, 'timing/model': 1.7897432637400925, 'learning_rate/qwen_vl_interface': 8.315148586409148e-06, 'learning_rate/action_model': 8.315148586409148e-05, 'epoch': 0.59}) +[2026-06-16 00:19:05,474][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.22434638440608978, 'train/grad_norm_pre_clip': 14.674141883850098, 'timing/data': 0.01081263180822134, 'timing/model': 1.525183282326907, 'learning_rate/qwen_vl_interface': 8.312821455734104e-06, 'learning_rate/action_model': 8.312821455734103e-05, 'epoch': 0.59}) +[2026-06-16 00:19:11,761][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.018044957891106606, 'train/grad_norm_pre_clip': 3.5312416553497314, 'timing/data': 0.010964281857013702, 'timing/model': 1.560201252810657, 'learning_rate/qwen_vl_interface': 8.31049306604885e-06, 'learning_rate/action_model': 8.310493066048849e-05, 'epoch': 0.59}) +[2026-06-16 00:19:18,164][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.014407305046916008, 'train/grad_norm_pre_clip': 2.854764699935913, 'timing/data': 0.011148483026772738, 'timing/model': 1.5717575210146606, 'learning_rate/qwen_vl_interface': 8.3081634183105e-06, 'learning_rate/action_model': 8.308163418310501e-05, 'epoch': 0.59}) +[2026-06-16 00:19:24,459][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.23444467782974243, 'train/grad_norm_pre_clip': 16.673826217651367, 'timing/data': 0.010791881009936333, 'timing/model': 1.5151092279702425, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 0.59}) +[2026-06-16 00:19:30,809][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.3882586658000946, 'train/grad_norm_pre_clip': 31.153533935546875, 'timing/data': 0.010842381045222282, 'timing/model': 1.8080757167190313, 'learning_rate/qwen_vl_interface': 8.303500352505557e-06, 'learning_rate/action_model': 8.303500352505557e-05, 'epoch': 0.59}) +[2026-06-16 00:19:36,866][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.1668459177017212, 'train/grad_norm_pre_clip': 19.861806869506836, 'timing/data': 0.011766226030886173, 'timing/model': 1.5666383397765458, 'learning_rate/qwen_vl_interface': 8.301166936355773e-06, 'learning_rate/action_model': 8.301166936355773e-05, 'epoch': 0.59}) +[2026-06-16 00:19:43,262][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.12792719900608063, 'train/grad_norm_pre_clip': 16.023359298706055, 'timing/data': 0.011414464097470045, 'timing/model': 1.5572022390551865, 'learning_rate/qwen_vl_interface': 8.298832265986514e-06, 'learning_rate/action_model': 8.298832265986513e-05, 'epoch': 0.59}) +[2026-06-16 00:19:49,564][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.01648203656077385, 'train/grad_norm_pre_clip': 3.5809755325317383, 'timing/data': 0.01131980400532484, 'timing/model': 1.5609668716788292, 'learning_rate/qwen_vl_interface': 8.296496342357475e-06, 'learning_rate/action_model': 8.296496342357474e-05, 'epoch': 0.59}) +[2026-06-16 00:19:55,863][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.02430790662765503, 'train/grad_norm_pre_clip': 5.473166465759277, 'timing/data': 0.010777100920677185, 'timing/model': 1.5339645091444254, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 0.59}) +[2026-06-16 00:20:02,150][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.0574953630566597, 'train/grad_norm_pre_clip': 6.10756778717041, 'timing/data': 0.010464439168572426, 'timing/model': 1.588032758794725, 'learning_rate/qwen_vl_interface': 8.29182073916141e-06, 'learning_rate/action_model': 8.29182073916141e-05, 'epoch': 0.59}) +[2026-06-16 00:20:08,443][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.1951165348291397, 'train/grad_norm_pre_clip': 20.410261154174805, 'timing/data': 0.011491553857922554, 'timing/model': 1.7343671722337604, 'learning_rate/qwen_vl_interface': 8.28948106151635e-06, 'learning_rate/action_model': 8.28948106151635e-05, 'epoch': 0.59}) +[2026-06-16 00:20:14,762][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.3064860999584198, 'train/grad_norm_pre_clip': 11.127466201782227, 'timing/data': 0.014670859090983868, 'timing/model': 1.6357279508374631, 'learning_rate/qwen_vl_interface': 8.287140134455435e-06, 'learning_rate/action_model': 8.287140134455435e-05, 'epoch': 0.59}) +[2026-06-16 00:20:21,350][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.1187758594751358, 'train/grad_norm_pre_clip': 5.49000358581543, 'timing/data': 0.010872961021959782, 'timing/model': 1.6509663979522884, 'learning_rate/qwen_vl_interface': 8.284797958940934e-06, 'learning_rate/action_model': 8.284797958940935e-05, 'epoch': 0.59}) +[2026-06-16 00:20:27,669][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.026995768770575523, 'train/grad_norm_pre_clip': 7.503762245178223, 'timing/data': 0.01093765115365386, 'timing/model': 1.5608159289695323, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 0.59}) +[2026-06-16 00:20:34,061][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.2139921635389328, 'train/grad_norm_pre_clip': 11.043397903442383, 'timing/data': 0.010966112371534109, 'timing/model': 1.5211298670619726, 'learning_rate/qwen_vl_interface': 8.280109866402805e-06, 'learning_rate/action_model': 8.280109866402806e-05, 'epoch': 0.59}) +[2026-06-16 00:20:40,406][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.07403887808322906, 'train/grad_norm_pre_clip': 12.705138206481934, 'timing/data': 0.011276883073151112, 'timing/model': 1.5558377648703754, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.6}) +[2026-06-16 00:20:46,661][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.05539204180240631, 'train/grad_norm_pre_clip': 5.725341320037842, 'timing/data': 0.011777695268392563, 'timing/model': 1.760462668724358, 'learning_rate/qwen_vl_interface': 8.275416791610356e-06, 'learning_rate/action_model': 8.275416791610355e-05, 'epoch': 0.6}) +[2026-06-16 00:20:52,764][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.04469595104455948, 'train/grad_norm_pre_clip': 3.7746777534484863, 'timing/data': 0.0105279297567904, 'timing/model': 1.5622490351088345, 'learning_rate/qwen_vl_interface': 8.27306838827987e-06, 'learning_rate/action_model': 8.273068388279871e-05, 'epoch': 0.6}) +[2026-06-16 00:20:59,062][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.01750517450273037, 'train/grad_norm_pre_clip': 3.039544105529785, 'timing/data': 0.01135763293132186, 'timing/model': 1.5623731720261276, 'learning_rate/qwen_vl_interface': 8.270718742280162e-06, 'learning_rate/action_model': 8.270718742280162e-05, 'epoch': 0.6}) +[2026-06-16 00:21:05,379][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.5353313684463501, 'train/grad_norm_pre_clip': 17.355865478515625, 'timing/data': 0.012123536318540573, 'timing/model': 1.5713244481012225, 'learning_rate/qwen_vl_interface': 8.268367854577084e-06, 'learning_rate/action_model': 8.268367854577082e-05, 'epoch': 0.6}) +[2026-06-16 00:21:11,770][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.17224842309951782, 'train/grad_norm_pre_clip': 10.933115005493164, 'timing/data': 0.01089355070143938, 'timing/model': 1.5495340819470584, 'learning_rate/qwen_vl_interface': 8.266015726136991e-06, 'learning_rate/action_model': 8.266015726136991e-05, 'epoch': 0.6}) +[2026-06-16 00:21:18,149][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.05521417781710625, 'train/grad_norm_pre_clip': 4.75075101852417, 'timing/data': 0.01125608291476965, 'timing/model': 1.828061608131975, 'learning_rate/qwen_vl_interface': 8.26366235792676e-06, 'learning_rate/action_model': 8.26366235792676e-05, 'epoch': 0.6}) +[2026-06-16 00:21:24,265][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.07497429847717285, 'train/grad_norm_pre_clip': 11.440244674682617, 'timing/data': 0.011143742129206657, 'timing/model': 1.5637455880641937, 'learning_rate/qwen_vl_interface': 8.261307750913769e-06, 'learning_rate/action_model': 8.261307750913769e-05, 'epoch': 0.6}) +[2026-06-16 00:21:30,568][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.3715784251689911, 'train/grad_norm_pre_clip': 19.327289581298828, 'timing/data': 0.011305273044854403, 'timing/model': 1.5731070679612458, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 0.6}) +[2026-06-16 00:21:36,763][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.32622066140174866, 'train/grad_norm_pre_clip': 13.679398536682129, 'timing/data': 0.010696940124034882, 'timing/model': 1.56142554897815, 'learning_rate/qwen_vl_interface': 8.256594824351577e-06, 'learning_rate/action_model': 8.256594824351576e-05, 'epoch': 0.6}) +[2026-06-16 00:21:43,062][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.09920644760131836, 'train/grad_norm_pre_clip': 7.367765426635742, 'timing/data': 0.010578629095107317, 'timing/model': 1.5700767650268972, 'learning_rate/qwen_vl_interface': 8.254236506739683e-06, 'learning_rate/action_model': 8.254236506739681e-05, 'epoch': 0.6}) +[2026-06-16 00:21:49,363][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.0638359785079956, 'train/grad_norm_pre_clip': 5.849908828735352, 'timing/data': 0.010658379644155502, 'timing/model': 1.5692367162555456, 'learning_rate/qwen_vl_interface': 8.25187695419964e-06, 'learning_rate/action_model': 8.25187695419964e-05, 'epoch': 0.6}) +[2026-06-16 00:21:55,702][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.17624704539775848, 'train/grad_norm_pre_clip': 12.96951961517334, 'timing/data': 0.010881619993597269, 'timing/model': 1.7913446258753538, 'learning_rate/qwen_vl_interface': 8.249516167701373e-06, 'learning_rate/action_model': 8.249516167701374e-05, 'epoch': 0.6}) +[2026-06-16 00:22:01,869][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.11335165798664093, 'train/grad_norm_pre_clip': 20.277450561523438, 'timing/data': 0.011125612072646618, 'timing/model': 1.5650135879404843, 'learning_rate/qwen_vl_interface': 8.247154148215311e-06, 'learning_rate/action_model': 8.24715414821531e-05, 'epoch': 0.6}) +[2026-06-16 00:22:08,161][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.179742231965065, 'train/grad_norm_pre_clip': 8.651490211486816, 'timing/data': 0.01073272991925478, 'timing/model': 1.5557057010009885, 'learning_rate/qwen_vl_interface': 8.24479089671239e-06, 'learning_rate/action_model': 8.24479089671239e-05, 'epoch': 0.6}) +[2026-06-16 00:22:14,563][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.050959035754203796, 'train/grad_norm_pre_clip': 6.446276664733887, 'timing/data': 0.010928820818662643, 'timing/model': 1.5623764568008482, 'learning_rate/qwen_vl_interface': 8.242426414164056e-06, 'learning_rate/action_model': 8.242426414164056e-05, 'epoch': 0.6}) +[2026-06-16 00:22:20,973][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.05017849802970886, 'train/grad_norm_pre_clip': 6.657214641571045, 'timing/data': 0.010812491178512573, 'timing/model': 1.5686588818207383, 'learning_rate/qwen_vl_interface': 8.240060701542257e-06, 'learning_rate/action_model': 8.240060701542257e-05, 'epoch': 0.6}) +[2026-06-16 00:22:27,367][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.11164684593677521, 'train/grad_norm_pre_clip': 6.3801679611206055, 'timing/data': 0.01123523199930787, 'timing/model': 1.5439714738167822, 'learning_rate/qwen_vl_interface': 8.237693759819447e-06, 'learning_rate/action_model': 8.237693759819448e-05, 'epoch': 0.6}) +[2026-06-16 00:22:33,698][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.048715438693761826, 'train/grad_norm_pre_clip': 3.6316092014312744, 'timing/data': 0.011090090963989496, 'timing/model': 1.8006844902411103, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 0.6}) +[2026-06-16 00:22:39,765][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.04055919498205185, 'train/grad_norm_pre_clip': 3.0764992237091064, 'timing/data': 0.010574059095233679, 'timing/model': 1.5609789537265897, 'learning_rate/qwen_vl_interface': 8.232956192963143e-06, 'learning_rate/action_model': 8.232956192963144e-05, 'epoch': 0.6}) +[2026-06-16 00:22:46,067][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.09560373425483704, 'train/grad_norm_pre_clip': 9.824532508850098, 'timing/data': 0.011417662724852562, 'timing/model': 1.5621671960689127, 'learning_rate/qwen_vl_interface': 8.230585569777086e-06, 'learning_rate/action_model': 8.230585569777085e-05, 'epoch': 0.6}) +[2026-06-16 00:22:52,368][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.0996934249997139, 'train/grad_norm_pre_clip': 10.242420196533203, 'timing/data': 0.012179676908999681, 'timing/model': 1.5649908450432122, 'learning_rate/qwen_vl_interface': 8.228213721384885e-06, 'learning_rate/action_model': 8.228213721384885e-05, 'epoch': 0.6}) +[2026-06-16 00:22:58,665][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.07242511212825775, 'train/grad_norm_pre_clip': 10.342058181762695, 'timing/data': 0.010941330809146166, 'timing/model': 1.5256755794398487, 'learning_rate/qwen_vl_interface': 8.225840648761523e-06, 'learning_rate/action_model': 8.225840648761523e-05, 'epoch': 0.6}) +[2026-06-16 00:23:04,984][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.030067721381783485, 'train/grad_norm_pre_clip': 4.079859256744385, 'timing/data': 0.010946740861982107, 'timing/model': 1.5717815891839564, 'learning_rate/qwen_vl_interface': 8.223466352882476e-06, 'learning_rate/action_model': 8.223466352882475e-05, 'epoch': 0.6}) +[2026-06-16 00:23:11,301][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.07150895148515701, 'train/grad_norm_pre_clip': 7.924691200256348, 'timing/data': 0.01186366518959403, 'timing/model': 1.790275784675032, 'learning_rate/qwen_vl_interface': 8.22109083472373e-06, 'learning_rate/action_model': 8.22109083472373e-05, 'epoch': 0.6}) +[2026-06-16 00:23:17,374][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.07675615698099136, 'train/grad_norm_pre_clip': 4.8372297286987305, 'timing/data': 0.011984284967184067, 'timing/model': 1.5723439138382673, 'learning_rate/qwen_vl_interface': 8.218714095261769e-06, 'learning_rate/action_model': 8.218714095261768e-05, 'epoch': 0.61}) +[2026-06-16 00:23:23,664][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.16608217358589172, 'train/grad_norm_pre_clip': 12.17700481414795, 'timing/data': 0.010968161281198263, 'timing/model': 1.5633951392956078, 'learning_rate/qwen_vl_interface': 8.21633613547358e-06, 'learning_rate/action_model': 8.21633613547358e-05, 'epoch': 0.61}) +[2026-06-16 00:23:29,967][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.014994241297245026, 'train/grad_norm_pre_clip': 3.7921090126037598, 'timing/data': 0.01063064904883504, 'timing/model': 1.5388943417929113, 'learning_rate/qwen_vl_interface': 8.213956956336656e-06, 'learning_rate/action_model': 8.213956956336656e-05, 'epoch': 0.61}) +[2026-06-16 00:34:29,427][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.02635844610631466, 'train/grad_norm_pre_clip': 5.911069393157959, 'eval/loss': 0.07668149471282959, 'eval/flap_precision': 0.9076923076923077, 'eval/flap_recall': 0.965893587994543, 'eval/flap_f1': 0.9358889623265036, 'eval/flap_rate_model': 0.21666666666666667, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9358889623265035, 'timing/data': 0.010550938546657562, 'timing/model': 1.5605572988279164, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.61}) +[2026-06-16 00:34:29,428][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1500_state +[2026-06-16 00:34:29,434][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 00:34:38,367][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1500_state/model.safetensors +[2026-06-16 00:34:58,255][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1500_state/optimizer.bin +[2026-06-16 00:34:58,256][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1500_state/sampler.bin +[2026-06-16 00:34:58,256][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-16 00:34:58,257][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-16 00:34:58,451][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-16 00:34:58,452][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 00:34:58,468][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 00:35:04,839][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.1504812240600586, 'train/grad_norm_pre_clip': 8.1466646194458, 'timing/data': 0.011970683932304382, 'timing/model': 1.5596968908794224, 'learning_rate/qwen_vl_interface': 8.209194943929059e-06, 'learning_rate/action_model': 8.20919494392906e-05, 'epoch': 0.61}) +[2026-06-16 00:35:11,048][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.034127723425626755, 'train/grad_norm_pre_clip': 14.861329078674316, 'timing/data': 0.010610117111355066, 'timing/model': 1.5675008846446872, 'learning_rate/qwen_vl_interface': 8.20681211261587e-06, 'learning_rate/action_model': 8.20681211261587e-05, 'epoch': 0.61}) +[2026-06-16 00:35:17,231][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.1900310069322586, 'train/grad_norm_pre_clip': 3.2569706439971924, 'timing/data': 0.011436301283538342, 'timing/model': 1.5589724062010646, 'learning_rate/qwen_vl_interface': 8.204428065868911e-06, 'learning_rate/action_model': 8.204428065868911e-05, 'epoch': 0.61}) +[2026-06-16 00:35:23,399][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.13754728436470032, 'train/grad_norm_pre_clip': 23.271150588989258, 'timing/data': 0.011211350094527006, 'timing/model': 1.729100987315178, 'learning_rate/qwen_vl_interface': 8.202042804668173e-06, 'learning_rate/action_model': 8.202042804668172e-05, 'epoch': 0.61}) +[2026-06-16 00:35:29,427][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.07188691943883896, 'train/grad_norm_pre_clip': 9.519034385681152, 'timing/data': 0.010972368996590376, 'timing/model': 1.5692613800056279, 'learning_rate/qwen_vl_interface': 8.199656329994145e-06, 'learning_rate/action_model': 8.199656329994145e-05, 'epoch': 0.61}) +[2026-06-16 00:35:35,644][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.18174023926258087, 'train/grad_norm_pre_clip': 13.93352222442627, 'timing/data': 0.011484242044389248, 'timing/model': 1.5713019277900457, 'learning_rate/qwen_vl_interface': 8.19726864282782e-06, 'learning_rate/action_model': 8.197268642827818e-05, 'epoch': 0.61}) +[2026-06-16 00:35:41,845][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.2682960629463196, 'train/grad_norm_pre_clip': 29.449054718017578, 'timing/data': 0.011576081160455942, 'timing/model': 1.5635063727386296, 'learning_rate/qwen_vl_interface': 8.194879744150682e-06, 'learning_rate/action_model': 8.194879744150681e-05, 'epoch': 0.61}) +[2026-06-16 00:35:48,053][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.04523034393787384, 'train/grad_norm_pre_clip': 8.373441696166992, 'timing/data': 0.011793902609497309, 'timing/model': 1.5067600179463625, 'learning_rate/qwen_vl_interface': 8.192489634944718e-06, 'learning_rate/action_model': 8.192489634944718e-05, 'epoch': 0.61}) +[2026-06-16 00:35:54,287][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.026486963033676147, 'train/grad_norm_pre_clip': 3.5054314136505127, 'timing/data': 0.01132707018405199, 'timing/model': 1.7937227813526988, 'learning_rate/qwen_vl_interface': 8.190098316192416e-06, 'learning_rate/action_model': 8.190098316192416e-05, 'epoch': 0.61}) +[2026-06-16 00:36:00,347][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.18503724038600922, 'train/grad_norm_pre_clip': 7.966373443603516, 'timing/data': 0.011006579268723726, 'timing/model': 1.5629500281065702, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 0.61}) +[2026-06-16 00:36:06,674][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.11389897763729095, 'train/grad_norm_pre_clip': 13.308273315429688, 'timing/data': 0.011082179844379425, 'timing/model': 1.5890534832142293, 'learning_rate/qwen_vl_interface': 8.185312053981204e-06, 'learning_rate/action_model': 8.185312053981204e-05, 'epoch': 0.61}) +[2026-06-16 00:36:12,949][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.08746779710054398, 'train/grad_norm_pre_clip': 5.2819437980651855, 'timing/data': 0.011680551804602146, 'timing/model': 1.5318513470701873, 'learning_rate/qwen_vl_interface': 8.182917112489747e-06, 'learning_rate/action_model': 8.182917112489746e-05, 'epoch': 0.61}) +[2026-06-16 00:36:19,321][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.10972804576158524, 'train/grad_norm_pre_clip': 8.550971984863281, 'timing/data': 0.010760568082332611, 'timing/model': 1.6247633206658065, 'learning_rate/qwen_vl_interface': 8.18052096538685e-06, 'learning_rate/action_model': 8.18052096538685e-05, 'epoch': 0.61}) +[2026-06-16 00:36:25,654][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.057345036417245865, 'train/grad_norm_pre_clip': 7.150424003601074, 'timing/data': 0.011239910032600164, 'timing/model': 1.5046119829639792, 'learning_rate/qwen_vl_interface': 8.178123613657483e-06, 'learning_rate/action_model': 8.178123613657482e-05, 'epoch': 0.61}) +[2026-06-16 00:36:31,997][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.020145157352089882, 'train/grad_norm_pre_clip': 2.3579649925231934, 'timing/data': 0.011418171226978302, 'timing/model': 1.8080806890502572, 'learning_rate/qwen_vl_interface': 8.175725058287102e-06, 'learning_rate/action_model': 8.175725058287102e-05, 'epoch': 0.61}) +[2026-06-16 00:36:38,071][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.04051785543560982, 'train/grad_norm_pre_clip': 3.1499409675598145, 'timing/data': 0.011732612270861864, 'timing/model': 1.5779723599553108, 'learning_rate/qwen_vl_interface': 8.173325300261664e-06, 'learning_rate/action_model': 8.173325300261663e-05, 'epoch': 0.61}) +[2026-06-16 00:36:44,360][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.1402917355298996, 'train/grad_norm_pre_clip': 13.38190746307373, 'timing/data': 0.011089518666267395, 'timing/model': 1.563961436972022, 'learning_rate/qwen_vl_interface': 8.170924340567617e-06, 'learning_rate/action_model': 8.170924340567616e-05, 'epoch': 0.61}) +[2026-06-16 00:36:50,649][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.2960394620895386, 'train/grad_norm_pre_clip': 17.187366485595703, 'timing/data': 0.011310450732707977, 'timing/model': 1.5535060786642134, 'learning_rate/qwen_vl_interface': 8.168522180191909e-06, 'learning_rate/action_model': 8.168522180191909e-05, 'epoch': 0.61}) +[2026-06-16 00:36:57,029][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.04030301794409752, 'train/grad_norm_pre_clip': 3.8149561882019043, 'timing/data': 0.011192178819328547, 'timing/model': 1.5647273580543697, 'learning_rate/qwen_vl_interface': 8.166118820121972e-06, 'learning_rate/action_model': 8.166118820121973e-05, 'epoch': 0.61}) +[2026-06-16 00:37:03,248][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.05813038349151611, 'train/grad_norm_pre_clip': 5.619513511657715, 'timing/data': 0.010360846295952797, 'timing/model': 1.5108071267604828, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 0.61}) +[2026-06-16 00:37:09,766][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.09542958438396454, 'train/grad_norm_pre_clip': 22.30708122253418, 'timing/data': 0.011950334068387747, 'timing/model': 1.87524774344638, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.62}) +[2026-06-16 00:37:16,019][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.01815173029899597, 'train/grad_norm_pre_clip': 1.87279212474823, 'timing/data': 0.011624171864241362, 'timing/model': 1.7228685123845935, 'learning_rate/qwen_vl_interface': 8.158901551628579e-06, 'learning_rate/action_model': 8.158901551628579e-05, 'epoch': 0.62}) +[2026-06-16 00:37:22,057][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.32294607162475586, 'train/grad_norm_pre_clip': 11.428442001342773, 'timing/data': 0.010770617984235287, 'timing/model': 1.5596714532002807, 'learning_rate/qwen_vl_interface': 8.156493402665971e-06, 'learning_rate/action_model': 8.15649340266597e-05, 'epoch': 0.62}) +[2026-06-16 00:37:28,361][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.08092527091503143, 'train/grad_norm_pre_clip': 16.852134704589844, 'timing/data': 0.011356641072779894, 'timing/model': 1.5635225889272988, 'learning_rate/qwen_vl_interface': 8.154084058953713e-06, 'learning_rate/action_model': 8.154084058953713e-05, 'epoch': 0.62}) +[2026-06-16 00:37:34,860][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.17608734965324402, 'train/grad_norm_pre_clip': 10.453117370605469, 'timing/data': 0.011079329065978527, 'timing/model': 1.5582401263527572, 'learning_rate/qwen_vl_interface': 8.151673521482193e-06, 'learning_rate/action_model': 8.151673521482194e-05, 'epoch': 0.62}) +[2026-06-16 00:37:41,151][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.14519542455673218, 'train/grad_norm_pre_clip': 16.86275291442871, 'timing/data': 0.011329340748488903, 'timing/model': 1.5098791192285717, 'learning_rate/qwen_vl_interface': 8.149261791242297e-06, 'learning_rate/action_model': 8.149261791242297e-05, 'epoch': 0.62}) +[2026-06-16 00:37:47,583][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.041658561676740646, 'train/grad_norm_pre_clip': 8.075241088867188, 'timing/data': 0.010853628162294626, 'timing/model': 1.7914740247651935, 'learning_rate/qwen_vl_interface': 8.146848869225391e-06, 'learning_rate/action_model': 8.146848869225391e-05, 'epoch': 0.62}) +[2026-06-16 00:37:53,990][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.19854092597961426, 'train/grad_norm_pre_clip': 21.28046226501465, 'timing/data': 0.011657241731882095, 'timing/model': 1.8809707532636821, 'learning_rate/qwen_vl_interface': 8.14443475642334e-06, 'learning_rate/action_model': 8.144434756423339e-05, 'epoch': 0.62}) +[2026-06-16 00:38:00,057][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.2728998064994812, 'train/grad_norm_pre_clip': 15.876843452453613, 'timing/data': 0.011533021926879883, 'timing/model': 1.5412410362623632, 'learning_rate/qwen_vl_interface': 8.142019453828493e-06, 'learning_rate/action_model': 8.142019453828493e-05, 'epoch': 0.62}) +[2026-06-16 00:38:06,363][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.02997925877571106, 'train/grad_norm_pre_clip': 10.669882774353027, 'timing/data': 0.010686418041586876, 'timing/model': 1.5651098620146513, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 0.62}) +[2026-06-16 00:38:12,658][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.07528030127286911, 'train/grad_norm_pre_clip': 8.825265884399414, 'timing/data': 0.011184760369360447, 'timing/model': 1.5469440501183271, 'learning_rate/qwen_vl_interface': 8.137185283232255e-06, 'learning_rate/action_model': 8.137185283232256e-05, 'epoch': 0.62}) +[2026-06-16 00:38:18,957][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.2335028201341629, 'train/grad_norm_pre_clip': 13.941529273986816, 'timing/data': 0.010748847853392363, 'timing/model': 1.520846274215728, 'learning_rate/qwen_vl_interface': 8.134766417218014e-06, 'learning_rate/action_model': 8.134766417218012e-05, 'epoch': 0.62}) +[2026-06-16 00:38:25,278][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.10695487260818481, 'train/grad_norm_pre_clip': 12.300093650817871, 'timing/data': 0.010181244928389788, 'timing/model': 1.764060027897358, 'learning_rate/qwen_vl_interface': 8.132346365385262e-06, 'learning_rate/action_model': 8.132346365385261e-05, 'epoch': 0.62}) +[2026-06-16 00:38:31,369][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.5064534544944763, 'train/grad_norm_pre_clip': 31.219791412353516, 'timing/data': 0.01093471935018897, 'timing/model': 1.566620676778257, 'learning_rate/qwen_vl_interface': 8.129925128728796e-06, 'learning_rate/action_model': 8.129925128728796e-05, 'epoch': 0.62}) +[2026-06-16 00:38:37,664][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.041999947279691696, 'train/grad_norm_pre_clip': 10.297544479370117, 'timing/data': 0.011464901268482208, 'timing/model': 1.563989254180342, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 0.62}) +[2026-06-16 00:38:43,965][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.325824111700058, 'train/grad_norm_pre_clip': 12.4705228805542, 'timing/data': 0.01132114976644516, 'timing/model': 1.563698222860694, 'learning_rate/qwen_vl_interface': 8.125079104926324e-06, 'learning_rate/action_model': 8.125079104926323e-05, 'epoch': 0.62}) +[2026-06-16 00:38:50,265][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.5289472937583923, 'train/grad_norm_pre_clip': 25.356882095336914, 'timing/data': 0.01422240398824215, 'timing/model': 1.5273143886588514, 'learning_rate/qwen_vl_interface': 8.122654319772332e-06, 'learning_rate/action_model': 8.122654319772332e-05, 'epoch': 0.62}) +[2026-06-16 00:38:56,565][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.1549409031867981, 'train/grad_norm_pre_clip': 14.429780960083008, 'timing/data': 0.010972049087285995, 'timing/model': 1.5802272446453571, 'learning_rate/qwen_vl_interface': 8.120228353778661e-06, 'learning_rate/action_model': 8.12022835377866e-05, 'epoch': 0.62}) +[2026-06-16 00:39:02,853][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.09737614542245865, 'train/grad_norm_pre_clip': 12.27989673614502, 'timing/data': 0.011745333205908537, 'timing/model': 1.7456009611487389, 'learning_rate/qwen_vl_interface': 8.117801207942528e-06, 'learning_rate/action_model': 8.117801207942528e-05, 'epoch': 0.62}) +[2026-06-16 00:39:08,966][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.05924108624458313, 'train/grad_norm_pre_clip': 3.3104248046875, 'timing/data': 0.011106149759143591, 'timing/model': 1.561110437847674, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 0.62}) +[2026-06-16 00:39:15,447][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.06703581660985947, 'train/grad_norm_pre_clip': 5.241445064544678, 'timing/data': 0.011212710291147232, 'timing/model': 1.5776779018342495, 'learning_rate/qwen_vl_interface': 8.11294338073421e-06, 'learning_rate/action_model': 8.11294338073421e-05, 'epoch': 0.62}) +[2026-06-16 00:39:21,865][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.07727024704217911, 'train/grad_norm_pre_clip': 4.43425178527832, 'timing/data': 0.011615772265940905, 'timing/model': 1.5655192770063877, 'learning_rate/qwen_vl_interface': 8.110512701358892e-06, 'learning_rate/action_model': 8.110512701358891e-05, 'epoch': 0.62}) +[2026-06-16 00:39:28,160][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.1551276296377182, 'train/grad_norm_pre_clip': 11.901713371276855, 'timing/data': 0.011230109725147486, 'timing/model': 1.525509787723422, 'learning_rate/qwen_vl_interface': 8.108080846134852e-06, 'learning_rate/action_model': 8.108080846134852e-05, 'epoch': 0.62}) +[2026-06-16 00:39:34,542][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.12275208532810211, 'train/grad_norm_pre_clip': 7.119844436645508, 'timing/data': 0.011004868894815445, 'timing/model': 1.5867328313179314, 'learning_rate/qwen_vl_interface': 8.105647816061734e-06, 'learning_rate/action_model': 8.105647816061735e-05, 'epoch': 0.62}) +[2026-06-16 00:39:40,860][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.10306275635957718, 'train/grad_norm_pre_clip': 8.84328556060791, 'timing/data': 0.011441110633313656, 'timing/model': 1.7586430558003485, 'learning_rate/qwen_vl_interface': 8.103213612139667e-06, 'learning_rate/action_model': 8.103213612139667e-05, 'epoch': 0.62}) +[2026-06-16 00:39:46,964][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.2278643250465393, 'train/grad_norm_pre_clip': 3.892998456954956, 'timing/data': 0.011474240571260452, 'timing/model': 1.5608310448005795, 'learning_rate/qwen_vl_interface': 8.100778235369262e-06, 'learning_rate/action_model': 8.10077823536926e-05, 'epoch': 0.63}) +[2026-06-16 00:39:53,260][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.04325294494628906, 'train/grad_norm_pre_clip': 7.940802574157715, 'timing/data': 0.011743161827325821, 'timing/model': 1.5576212196610868, 'learning_rate/qwen_vl_interface': 8.098341686751605e-06, 'learning_rate/action_model': 8.098341686751605e-05, 'epoch': 0.63}) +[2026-06-16 00:39:59,560][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.23988738656044006, 'train/grad_norm_pre_clip': 6.929512977600098, 'timing/data': 0.010710407979786396, 'timing/model': 1.5539585631340742, 'learning_rate/qwen_vl_interface': 8.095903967288273e-06, 'learning_rate/action_model': 8.095903967288272e-05, 'epoch': 0.63}) +[2026-06-16 00:40:05,867][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.204533189535141, 'train/grad_norm_pre_clip': 12.087973594665527, 'timing/data': 0.011347400024533272, 'timing/model': 1.526595810893923, 'learning_rate/qwen_vl_interface': 8.093465077981318e-06, 'learning_rate/action_model': 8.093465077981318e-05, 'epoch': 0.63}) +[2026-06-16 00:40:12,198][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.11166197806596756, 'train/grad_norm_pre_clip': 4.810589790344238, 'timing/data': 0.01090435916557908, 'timing/model': 1.7906062863767147, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 0.63}) +[2026-06-16 00:40:18,272][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.07368764281272888, 'train/grad_norm_pre_clip': 3.211897850036621, 'timing/data': 0.011175920255482197, 'timing/model': 1.5720341512933373, 'learning_rate/qwen_vl_interface': 8.088583793847165e-06, 'learning_rate/action_model': 8.088583793847164e-05, 'epoch': 0.63}) +[2026-06-16 00:40:24,571][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.21988822519779205, 'train/grad_norm_pre_clip': 15.112102508544922, 'timing/data': 0.011120379902422428, 'timing/model': 1.5744683821685612, 'learning_rate/qwen_vl_interface': 8.086141401026478e-06, 'learning_rate/action_model': 8.086141401026478e-05, 'epoch': 0.63}) +[2026-06-16 00:40:30,870][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.03721827268600464, 'train/grad_norm_pre_clip': 2.3722622394561768, 'timing/data': 0.011189409997314215, 'timing/model': 1.5643869256600738, 'learning_rate/qwen_vl_interface': 8.083697842375193e-06, 'learning_rate/action_model': 8.083697842375193e-05, 'epoch': 0.63}) +[2026-06-16 00:40:37,165][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.0625179260969162, 'train/grad_norm_pre_clip': 3.5632662773132324, 'timing/data': 0.010603237897157669, 'timing/model': 1.5840457528829575, 'learning_rate/qwen_vl_interface': 8.081253118897762e-06, 'learning_rate/action_model': 8.081253118897761e-05, 'epoch': 0.63}) +[2026-06-16 00:40:43,462][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.24840208888053894, 'train/grad_norm_pre_clip': 18.43621826171875, 'timing/data': 0.011438220739364624, 'timing/model': 1.5591212823055685, 'learning_rate/qwen_vl_interface': 8.07880723159912e-06, 'learning_rate/action_model': 8.07880723159912e-05, 'epoch': 0.63}) +[2026-06-16 00:40:49,797][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.08478973060846329, 'train/grad_norm_pre_clip': 5.890947341918945, 'timing/data': 0.011362090241163969, 'timing/model': 1.7918008780106902, 'learning_rate/qwen_vl_interface': 8.076360181484676e-06, 'learning_rate/action_model': 8.076360181484676e-05, 'epoch': 0.63}) +[2026-06-16 00:40:55,961][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.19404909014701843, 'train/grad_norm_pre_clip': 8.028834342956543, 'timing/data': 0.011145679745823145, 'timing/model': 1.545546050183475, 'learning_rate/qwen_vl_interface': 8.073911969560326e-06, 'learning_rate/action_model': 8.073911969560326e-05, 'epoch': 0.63}) +[2026-06-16 00:41:02,271][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.06078311428427696, 'train/grad_norm_pre_clip': 5.512043476104736, 'timing/data': 0.010927839204668999, 'timing/model': 1.5650238571688533, 'learning_rate/qwen_vl_interface': 8.071462596832432e-06, 'learning_rate/action_model': 8.071462596832432e-05, 'epoch': 0.63}) +[2026-06-16 00:41:08,570][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.11607189476490021, 'train/grad_norm_pre_clip': 6.077553749084473, 'timing/data': 0.01084974780678749, 'timing/model': 1.5688811736181378, 'learning_rate/qwen_vl_interface': 8.06901206430784e-06, 'learning_rate/action_model': 8.06901206430784e-05, 'epoch': 0.63}) +[2026-06-16 00:41:14,866][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.07090385258197784, 'train/grad_norm_pre_clip': 4.142516136169434, 'timing/data': 0.010765037965029478, 'timing/model': 1.5686849332414567, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 0.63}) +[2026-06-16 00:41:21,159][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.15382786095142365, 'train/grad_norm_pre_clip': 14.297979354858398, 'timing/data': 0.010710377711802721, 'timing/model': 1.5670778537169099, 'learning_rate/qwen_vl_interface': 8.064107523898327e-06, 'learning_rate/action_model': 8.064107523898326e-05, 'epoch': 0.63}) +[2026-06-16 00:41:27,493][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.2788974940776825, 'train/grad_norm_pre_clip': 19.005605697631836, 'timing/data': 0.010816887952387333, 'timing/model': 1.7921907263807952, 'learning_rate/qwen_vl_interface': 8.061653518029476e-06, 'learning_rate/action_model': 8.061653518029476e-05, 'epoch': 0.63}) +[2026-06-16 00:41:33,760][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.08530836552381516, 'train/grad_norm_pre_clip': 5.679654598236084, 'timing/data': 0.011071734130382538, 'timing/model': 1.6592145431786776, 'learning_rate/qwen_vl_interface': 8.059198356396069e-06, 'learning_rate/action_model': 8.059198356396068e-05, 'epoch': 0.63}) +[2026-06-16 00:41:40,179][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.22812393307685852, 'train/grad_norm_pre_clip': 6.44927978515625, 'timing/data': 0.011659218929708004, 'timing/model': 1.596336665097624, 'learning_rate/qwen_vl_interface': 8.05674204000733e-06, 'learning_rate/action_model': 8.05674204000733e-05, 'epoch': 0.63}) +[2026-06-16 00:41:46,477][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.03515130281448364, 'train/grad_norm_pre_clip': 4.406391143798828, 'timing/data': 0.010785490274429321, 'timing/model': 1.5596496779471636, 'learning_rate/qwen_vl_interface': 8.05428456987296e-06, 'learning_rate/action_model': 8.05428456987296e-05, 'epoch': 0.63}) +[2026-06-16 00:41:52,762][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.05907658115029335, 'train/grad_norm_pre_clip': 5.6442790031433105, 'timing/data': 0.011584966909140348, 'timing/model': 1.5230620200745761, 'learning_rate/qwen_vl_interface': 8.05182594700313e-06, 'learning_rate/action_model': 8.05182594700313e-05, 'epoch': 0.63}) +[2026-06-16 00:41:59,079][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.21365506947040558, 'train/grad_norm_pre_clip': 10.609350204467773, 'timing/data': 0.011393625289201736, 'timing/model': 1.5677455812692642, 'learning_rate/qwen_vl_interface': 8.049366172408484e-06, 'learning_rate/action_model': 8.049366172408485e-05, 'epoch': 0.63}) +[2026-06-16 00:42:05,399][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.08380170911550522, 'train/grad_norm_pre_clip': 6.888222694396973, 'timing/data': 0.011347923893481493, 'timing/model': 1.7995706461369991, 'learning_rate/qwen_vl_interface': 8.04690524710015e-06, 'learning_rate/action_model': 8.046905247100151e-05, 'epoch': 0.63}) +[2026-06-16 00:42:11,532][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.36538320779800415, 'train/grad_norm_pre_clip': 18.026567459106445, 'timing/data': 0.010930909309536219, 'timing/model': 1.6342384456656873, 'learning_rate/qwen_vl_interface': 8.044443172089715e-06, 'learning_rate/action_model': 8.044443172089714e-05, 'epoch': 0.63}) +[2026-06-16 00:42:17,768][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.09655839204788208, 'train/grad_norm_pre_clip': 13.385641098022461, 'timing/data': 0.011461833957582712, 'timing/model': 1.569634672254324, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.63}) +[2026-06-16 00:42:24,072][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.2229810506105423, 'train/grad_norm_pre_clip': 15.863701820373535, 'timing/data': 0.010814987123012543, 'timing/model': 1.569677866064012, 'learning_rate/qwen_vl_interface': 8.039515577011287e-06, 'learning_rate/action_model': 8.039515577011288e-05, 'epoch': 0.64}) +[2026-06-16 00:42:30,368][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.0753367468714714, 'train/grad_norm_pre_clip': 4.9442033767700195, 'timing/data': 0.011310161091387272, 'timing/model': 1.5691191419027746, 'learning_rate/qwen_vl_interface': 8.037050058968841e-06, 'learning_rate/action_model': 8.03705005896884e-05, 'epoch': 0.64}) +[2026-06-16 00:42:36,743][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.1736307591199875, 'train/grad_norm_pre_clip': 12.570954322814941, 'timing/data': 0.011842534877359867, 'timing/model': 1.5869816346094012, 'learning_rate/qwen_vl_interface': 8.034583395275392e-06, 'learning_rate/action_model': 8.034583395275392e-05, 'epoch': 0.64}) +[2026-06-16 00:42:43,097][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.12844343483448029, 'train/grad_norm_pre_clip': 6.893743991851807, 'timing/data': 0.011373091023415327, 'timing/model': 1.7957867737859488, 'learning_rate/qwen_vl_interface': 8.032115586944893e-06, 'learning_rate/action_model': 8.032115586944892e-05, 'epoch': 0.64}) +[2026-06-16 00:42:49,270][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.021623361855745316, 'train/grad_norm_pre_clip': 2.196016788482666, 'timing/data': 0.011702952906489372, 'timing/model': 1.5774346920661628, 'learning_rate/qwen_vl_interface': 8.029646634991762e-06, 'learning_rate/action_model': 8.029646634991762e-05, 'epoch': 0.64}) +[2026-06-16 00:42:55,570][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.4316365122795105, 'train/grad_norm_pre_clip': 14.20108699798584, 'timing/data': 0.0114878099411726, 'timing/model': 1.5624901368282735, 'learning_rate/qwen_vl_interface': 8.0271765404309e-06, 'learning_rate/action_model': 8.0271765404309e-05, 'epoch': 0.64}) +[2026-06-16 00:43:01,866][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.20965592563152313, 'train/grad_norm_pre_clip': 10.053133964538574, 'timing/data': 0.011851403396576643, 'timing/model': 1.5618178914301097, 'learning_rate/qwen_vl_interface': 8.024705304277664e-06, 'learning_rate/action_model': 8.024705304277664e-05, 'epoch': 0.64}) +[2026-06-16 00:43:08,160][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.1298496425151825, 'train/grad_norm_pre_clip': 5.22172212600708, 'timing/data': 0.010602171998471022, 'timing/model': 1.55641366308555, 'learning_rate/qwen_vl_interface': 8.022232927547888e-06, 'learning_rate/action_model': 8.022232927547886e-05, 'epoch': 0.64}) +[2026-06-16 00:43:14,462][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.15539009869098663, 'train/grad_norm_pre_clip': 14.17403793334961, 'timing/data': 0.010734261944890022, 'timing/model': 1.5312816561199725, 'learning_rate/qwen_vl_interface': 8.01975941125787e-06, 'learning_rate/action_model': 8.01975941125787e-05, 'epoch': 0.64}) +[2026-06-16 00:43:20,820][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.021392423659563065, 'train/grad_norm_pre_clip': 2.778355598449707, 'timing/data': 0.012165403924882412, 'timing/model': 1.8010363769717515, 'learning_rate/qwen_vl_interface': 8.017284756424381e-06, 'learning_rate/action_model': 8.017284756424381e-05, 'epoch': 0.64}) +[2026-06-16 00:43:26,900][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.22640986740589142, 'train/grad_norm_pre_clip': 6.29462194442749, 'timing/data': 0.011819230858236551, 'timing/model': 1.5749063449911773, 'learning_rate/qwen_vl_interface': 8.01480896406466e-06, 'learning_rate/action_model': 8.01480896406466e-05, 'epoch': 0.64}) +[2026-06-16 00:43:33,180][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.22793048620224, 'train/grad_norm_pre_clip': 6.9104156494140625, 'timing/data': 0.011623157653957605, 'timing/model': 1.5749026681296527, 'learning_rate/qwen_vl_interface': 8.01233203519641e-06, 'learning_rate/action_model': 8.012332035196409e-05, 'epoch': 0.64}) +[2026-06-16 00:43:39,776][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.1416761726140976, 'train/grad_norm_pre_clip': 6.56813907623291, 'timing/data': 0.010945581831037998, 'timing/model': 1.5635149022564292, 'learning_rate/qwen_vl_interface': 8.009853970837801e-06, 'learning_rate/action_model': 8.009853970837801e-05, 'epoch': 0.64}) +[2026-06-16 00:43:46,067][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.14643295109272003, 'train/grad_norm_pre_clip': 8.384878158569336, 'timing/data': 0.010756829753518105, 'timing/model': 1.5637295679189265, 'learning_rate/qwen_vl_interface': 8.007374772007474e-06, 'learning_rate/action_model': 8.007374772007474e-05, 'epoch': 0.64}) +[2026-06-16 00:43:52,459][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.14140935242176056, 'train/grad_norm_pre_clip': 5.798267364501953, 'timing/data': 0.011558706406503916, 'timing/model': 1.5229458408430219, 'learning_rate/qwen_vl_interface': 8.004894439724533e-06, 'learning_rate/action_model': 8.004894439724533e-05, 'epoch': 0.64}) +[2026-06-16 00:43:58,987][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.036450278013944626, 'train/grad_norm_pre_clip': 3.4843251705169678, 'timing/data': 0.011160003021359444, 'timing/model': 1.8931262302212417, 'learning_rate/qwen_vl_interface': 8.002412975008552e-06, 'learning_rate/action_model': 8.002412975008552e-05, 'epoch': 0.64}) +[2026-06-16 00:44:05,059][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.05238226801156998, 'train/grad_norm_pre_clip': 3.788764238357544, 'timing/data': 0.010524396784603596, 'timing/model': 1.5619376096874475, 'learning_rate/qwen_vl_interface': 7.999930378879562e-06, 'learning_rate/action_model': 7.999930378879561e-05, 'epoch': 0.64}) +[2026-06-16 00:44:11,366][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.10020485520362854, 'train/grad_norm_pre_clip': 6.649034023284912, 'timing/data': 0.0104087358340621, 'timing/model': 1.5650859009474516, 'learning_rate/qwen_vl_interface': 7.997446652358067e-06, 'learning_rate/action_model': 7.997446652358068e-05, 'epoch': 0.64}) +[2026-06-16 00:44:17,699][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.0462455190718174, 'train/grad_norm_pre_clip': 2.782428503036499, 'timing/data': 0.010843878146260977, 'timing/model': 1.5978980078361928, 'learning_rate/qwen_vl_interface': 7.994961796465035e-06, 'learning_rate/action_model': 7.994961796465035e-05, 'epoch': 0.64}) +[2026-06-16 00:44:23,955][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.10219968110322952, 'train/grad_norm_pre_clip': 5.982397556304932, 'timing/data': 0.010064302012324333, 'timing/model': 1.5634204358793795, 'learning_rate/qwen_vl_interface': 7.992475812221894e-06, 'learning_rate/action_model': 7.992475812221894e-05, 'epoch': 0.64}) +[2026-06-16 00:44:30,361][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.3330574929714203, 'train/grad_norm_pre_clip': 10.282230377197266, 'timing/data': 0.011386253405362368, 'timing/model': 1.5894610141403973, 'learning_rate/qwen_vl_interface': 7.989988700650538e-06, 'learning_rate/action_model': 7.989988700650538e-05, 'epoch': 0.64}) +[2026-06-16 00:44:36,721][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.15622735023498535, 'train/grad_norm_pre_clip': 7.334008693695068, 'timing/data': 0.011087719351053238, 'timing/model': 1.8205215083435178, 'learning_rate/qwen_vl_interface': 7.987500462773326e-06, 'learning_rate/action_model': 7.987500462773326e-05, 'epoch': 0.64}) +[2026-06-16 00:44:43,043][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.2357645332813263, 'train/grad_norm_pre_clip': 6.983763694763184, 'timing/data': 0.011215319857001305, 'timing/model': 1.7395967589691281, 'learning_rate/qwen_vl_interface': 7.985011099613078e-06, 'learning_rate/action_model': 7.985011099613077e-05, 'epoch': 0.64}) +[2026-06-16 00:44:49,258][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.38999873399734497, 'train/grad_norm_pre_clip': 10.615113258361816, 'timing/data': 0.011245959904044867, 'timing/model': 1.5550245800986886, 'learning_rate/qwen_vl_interface': 7.98252061219308e-06, 'learning_rate/action_model': 7.982520612193078e-05, 'epoch': 0.64}) +[2026-06-16 00:44:55,630][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.06893637031316757, 'train/grad_norm_pre_clip': 7.607647895812988, 'timing/data': 0.011862555053085089, 'timing/model': 1.6204729010351002, 'learning_rate/qwen_vl_interface': 7.980029001537073e-06, 'learning_rate/action_model': 7.980029001537072e-05, 'epoch': 0.64}) +[2026-06-16 00:45:01,963][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.10941018164157867, 'train/grad_norm_pre_clip': 8.27125072479248, 'timing/data': 0.010757804848253727, 'timing/model': 1.5873751221224666, 'learning_rate/qwen_vl_interface': 7.977536268669264e-06, 'learning_rate/action_model': 7.977536268669264e-05, 'epoch': 0.65}) +[2026-06-16 00:45:08,382][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.03111838735640049, 'train/grad_norm_pre_clip': 4.506218910217285, 'timing/data': 0.011810463387519121, 'timing/model': 1.5694965817965567, 'learning_rate/qwen_vl_interface': 7.975042414614324e-06, 'learning_rate/action_model': 7.975042414614324e-05, 'epoch': 0.65}) +[2026-06-16 00:45:14,705][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.26352667808532715, 'train/grad_norm_pre_clip': 5.228062629699707, 'timing/data': 0.012447107583284378, 'timing/model': 1.8055587182752788, 'learning_rate/qwen_vl_interface': 7.972547440397382e-06, 'learning_rate/action_model': 7.972547440397383e-05, 'epoch': 0.65}) +[2026-06-16 00:45:21,005][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.0976829081773758, 'train/grad_norm_pre_clip': 9.746685981750488, 'timing/data': 0.011242818087339401, 'timing/model': 1.7971359002403915, 'learning_rate/qwen_vl_interface': 7.970051347044026e-06, 'learning_rate/action_model': 7.970051347044025e-05, 'epoch': 0.65}) +[2026-06-16 00:56:24,513][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.1470334827899933, 'train/grad_norm_pre_clip': 11.499366760253906, 'eval/loss': 0.08424514532089233, 'eval/flap_precision': 0.8817073170731707, 'eval/flap_recall': 0.9863574351978172, 'eval/flap_f1': 0.9311010946555055, 'eval/flap_rate_model': 0.22777777777777777, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9311010946555055, 'timing/data': 0.011047276202589273, 'timing/model': 1.6390601289458573, 'learning_rate/qwen_vl_interface': 7.967554135580307e-06, 'learning_rate/action_model': 7.967554135580307e-05, 'epoch': 0.65}) +[2026-06-16 00:56:24,514][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1600_state +[2026-06-16 00:56:24,520][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 00:56:33,168][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1600_state/model.safetensors +[2026-06-16 00:56:52,846][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1600_state/optimizer.bin +[2026-06-16 00:56:52,847][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1600_state/sampler.bin +[2026-06-16 00:56:52,847][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1600_state/sampler_1.bin +[2026-06-16 00:56:52,848][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1600_state/random_states_0.pkl +[2026-06-16 00:56:53,054][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1600 +[2026-06-16 00:56:53,054][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 00:56:53,069][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 00:56:59,163][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.0952458530664444, 'train/grad_norm_pre_clip': 5.258340835571289, 'timing/data': 0.011595176067203283, 'timing/model': 1.5108091980218887, 'learning_rate/qwen_vl_interface': 7.965055807032734e-06, 'learning_rate/action_model': 7.965055807032735e-05, 'epoch': 0.65}) +[2026-06-16 00:57:05,454][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.5034535527229309, 'train/grad_norm_pre_clip': 11.20433521270752, 'timing/data': 0.011835457757115364, 'timing/model': 1.5852820076979697, 'learning_rate/qwen_vl_interface': 7.962556362428275e-06, 'learning_rate/action_model': 7.962556362428274e-05, 'epoch': 0.65}) +[2026-06-16 00:57:11,746][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.1978921741247177, 'train/grad_norm_pre_clip': 16.278533935546875, 'timing/data': 0.0114222951233387, 'timing/model': 1.5668333787471056, 'learning_rate/qwen_vl_interface': 7.960055802794358e-06, 'learning_rate/action_model': 7.960055802794357e-05, 'epoch': 0.65}) +[2026-06-16 00:57:18,110][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.05632524564862251, 'train/grad_norm_pre_clip': 3.210019826889038, 'timing/data': 0.011515945196151733, 'timing/model': 1.6265411153435707, 'learning_rate/qwen_vl_interface': 7.957554129158866e-06, 'learning_rate/action_model': 7.957554129158866e-05, 'epoch': 0.65}) +[2026-06-16 00:57:24,436][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.07454819977283478, 'train/grad_norm_pre_clip': 5.398637294769287, 'timing/data': 0.01320277526974678, 'timing/model': 1.5394919780083, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 0.65}) +[2026-06-16 00:57:30,688][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.08720263093709946, 'train/grad_norm_pre_clip': 6.661605358123779, 'timing/data': 0.01054480904713273, 'timing/model': 1.8157296786084771, 'learning_rate/qwen_vl_interface': 7.952547443997e-06, 'learning_rate/action_model': 7.952547443997e-05, 'epoch': 0.65}) +[2026-06-16 00:57:36,749][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.10424885153770447, 'train/grad_norm_pre_clip': 5.174258232116699, 'timing/data': 0.011191833298653364, 'timing/model': 1.5675897519104183, 'learning_rate/qwen_vl_interface': 7.950042434528681e-06, 'learning_rate/action_model': 7.950042434528682e-05, 'epoch': 0.65}) +[2026-06-16 00:57:43,044][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.0894138514995575, 'train/grad_norm_pre_clip': 4.520160675048828, 'timing/data': 0.011469005141407251, 'timing/model': 1.5638637971132994, 'learning_rate/qwen_vl_interface': 7.94753631517491e-06, 'learning_rate/action_model': 7.947536315174909e-05, 'epoch': 0.65}) +[2026-06-16 00:57:49,568][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.13409972190856934, 'train/grad_norm_pre_clip': 9.539586067199707, 'timing/data': 0.01248809089884162, 'timing/model': 1.6593914022669196, 'learning_rate/qwen_vl_interface': 7.945029086965854e-06, 'learning_rate/action_model': 7.945029086965853e-05, 'epoch': 0.65}) +[2026-06-16 00:57:55,980][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.2385132759809494, 'train/grad_norm_pre_clip': 18.833168029785156, 'timing/data': 0.011770776007324457, 'timing/model': 1.5707584111951292, 'learning_rate/qwen_vl_interface': 7.942520750932138e-06, 'learning_rate/action_model': 7.942520750932138e-05, 'epoch': 0.65}) +[2026-06-16 00:58:02,249][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.11619919538497925, 'train/grad_norm_pre_clip': 8.466134071350098, 'timing/data': 0.010751830879598856, 'timing/model': 1.5399432042613626, 'learning_rate/qwen_vl_interface': 7.940011308104848e-06, 'learning_rate/action_model': 7.940011308104848e-05, 'epoch': 0.65}) +[2026-06-16 00:58:08,599][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.03271154314279556, 'train/grad_norm_pre_clip': 2.449618101119995, 'timing/data': 0.010983952321112156, 'timing/model': 1.81013015890494, 'learning_rate/qwen_vl_interface': 7.93750075951552e-06, 'learning_rate/action_model': 7.93750075951552e-05, 'epoch': 0.65}) +[2026-06-16 00:58:14,842][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.06472102552652359, 'train/grad_norm_pre_clip': 8.123176574707031, 'timing/data': 0.01108102174475789, 'timing/model': 1.6584556442685425, 'learning_rate/qwen_vl_interface': 7.934989106196145e-06, 'learning_rate/action_model': 7.934989106196146e-05, 'epoch': 0.65}) +[2026-06-16 00:58:21,158][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.1877230405807495, 'train/grad_norm_pre_clip': 9.859541893005371, 'timing/data': 0.012829022016376257, 'timing/model': 1.5807509077712893, 'learning_rate/qwen_vl_interface': 7.93247634917917e-06, 'learning_rate/action_model': 7.93247634917917e-05, 'epoch': 0.65}) +[2026-06-16 00:58:27,448][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.1251119077205658, 'train/grad_norm_pre_clip': 4.9670634269714355, 'timing/data': 0.011455753818154335, 'timing/model': 1.5574431158602238, 'learning_rate/qwen_vl_interface': 7.929962489497494e-06, 'learning_rate/action_model': 7.929962489497493e-05, 'epoch': 0.65}) +[2026-06-16 00:58:33,751][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.03817202150821686, 'train/grad_norm_pre_clip': 10.437416076660156, 'timing/data': 0.011362713761627674, 'timing/model': 1.5635124165564775, 'learning_rate/qwen_vl_interface': 7.927447528184469e-06, 'learning_rate/action_model': 7.927447528184468e-05, 'epoch': 0.65}) +[2026-06-16 00:58:40,061][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.08135679364204407, 'train/grad_norm_pre_clip': 6.076869964599609, 'timing/data': 0.010943610686808825, 'timing/model': 1.5548764681443572, 'learning_rate/qwen_vl_interface': 7.924931466273901e-06, 'learning_rate/action_model': 7.924931466273902e-05, 'epoch': 0.65}) +[2026-06-16 00:58:46,386][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.10600206255912781, 'train/grad_norm_pre_clip': 13.661629676818848, 'timing/data': 0.010857080109417439, 'timing/model': 1.7823830740526319, 'learning_rate/qwen_vl_interface': 7.92241430480005e-06, 'learning_rate/action_model': 7.922414304800049e-05, 'epoch': 0.65}) +[2026-06-16 00:58:52,465][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.2471545934677124, 'train/grad_norm_pre_clip': 8.513959884643555, 'timing/data': 0.011661345139145851, 'timing/model': 1.5755255413241684, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.65}) +[2026-06-16 00:58:58,748][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.09333117306232452, 'train/grad_norm_pre_clip': 3.183062791824341, 'timing/data': 0.011519604362547398, 'timing/model': 1.5505902511067688, 'learning_rate/qwen_vl_interface': 7.91737668730178e-06, 'learning_rate/action_model': 7.91737668730178e-05, 'epoch': 0.66}) +[2026-06-16 00:59:05,156][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.13033714890480042, 'train/grad_norm_pre_clip': 8.069733619689941, 'timing/data': 0.011038691271096468, 'timing/model': 1.5706618521362543, 'learning_rate/qwen_vl_interface': 7.91485623334814e-06, 'learning_rate/action_model': 7.91485623334814e-05, 'epoch': 0.66}) +[2026-06-16 00:59:11,451][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.09896117448806763, 'train/grad_norm_pre_clip': 5.566819190979004, 'timing/data': 0.01095876982435584, 'timing/model': 1.5209766458719969, 'learning_rate/qwen_vl_interface': 7.91233468397276e-06, 'learning_rate/action_model': 7.912334683972761e-05, 'epoch': 0.66}) +[2026-06-16 00:59:17,751][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.03775234520435333, 'train/grad_norm_pre_clip': 5.278347492218018, 'timing/data': 0.011341922916471958, 'timing/model': 1.5548270619474351, 'learning_rate/qwen_vl_interface': 7.90981204021216e-06, 'learning_rate/action_model': 7.90981204021216e-05, 'epoch': 0.66}) +[2026-06-16 00:59:24,092][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.020876610651612282, 'train/grad_norm_pre_clip': 2.0890908241271973, 'timing/data': 0.011155981104820967, 'timing/model': 1.8024835409596562, 'learning_rate/qwen_vl_interface': 7.907288303103298e-06, 'learning_rate/action_model': 7.907288303103298e-05, 'epoch': 0.66}) +[2026-06-16 00:59:30,157][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.03665551170706749, 'train/grad_norm_pre_clip': 9.757625579833984, 'timing/data': 0.011265481822192669, 'timing/model': 1.5621513817459345, 'learning_rate/qwen_vl_interface': 7.90476347368359e-06, 'learning_rate/action_model': 7.90476347368359e-05, 'epoch': 0.66}) +[2026-06-16 00:59:36,458][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.019109021872282028, 'train/grad_norm_pre_clip': 2.288278102874756, 'timing/data': 0.010486798360943794, 'timing/model': 1.5591225558891892, 'learning_rate/qwen_vl_interface': 7.902237552990897e-06, 'learning_rate/action_model': 7.902237552990897e-05, 'epoch': 0.66}) +[2026-06-16 00:59:42,747][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.29985713958740234, 'train/grad_norm_pre_clip': 28.21507453918457, 'timing/data': 0.010639699175953865, 'timing/model': 1.563223217613995, 'learning_rate/qwen_vl_interface': 7.89971054206353e-06, 'learning_rate/action_model': 7.89971054206353e-05, 'epoch': 0.66}) +[2026-06-16 00:59:49,044][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.35371798276901245, 'train/grad_norm_pre_clip': 24.125593185424805, 'timing/data': 0.010658598970621824, 'timing/model': 1.5750410319305956, 'learning_rate/qwen_vl_interface': 7.897182441940247e-06, 'learning_rate/action_model': 7.897182441940247e-05, 'epoch': 0.66}) +[2026-06-16 00:59:55,361][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.2318442016839981, 'train/grad_norm_pre_clip': 35.060325622558594, 'timing/data': 0.010012974962592125, 'timing/model': 1.5244726953096688, 'learning_rate/qwen_vl_interface': 7.894653253660256e-06, 'learning_rate/action_model': 7.894653253660256e-05, 'epoch': 0.66}) +[2026-06-16 01:00:01,685][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.02261752262711525, 'train/grad_norm_pre_clip': 4.234769344329834, 'timing/data': 0.011123801115900278, 'timing/model': 1.7964349081739783, 'learning_rate/qwen_vl_interface': 7.89212297826321e-06, 'learning_rate/action_model': 7.892122978263209e-05, 'epoch': 0.66}) +[2026-06-16 01:00:07,751][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.1756136268377304, 'train/grad_norm_pre_clip': 38.320308685302734, 'timing/data': 0.010292276740074158, 'timing/model': 1.5357344578951597, 'learning_rate/qwen_vl_interface': 7.889591616789206e-06, 'learning_rate/action_model': 7.889591616789206e-05, 'epoch': 0.66}) +[2026-06-16 01:00:14,054][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.11694366484880447, 'train/grad_norm_pre_clip': 10.701805114746094, 'timing/data': 0.010197126306593418, 'timing/model': 1.5592211172915995, 'learning_rate/qwen_vl_interface': 7.887059170278799e-06, 'learning_rate/action_model': 7.887059170278797e-05, 'epoch': 0.66}) +[2026-06-16 01:00:20,252][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.048793334513902664, 'train/grad_norm_pre_clip': 3.9274983406066895, 'timing/data': 0.010251956060528755, 'timing/model': 1.554860773961991, 'learning_rate/qwen_vl_interface': 7.884525639772974e-06, 'learning_rate/action_model': 7.884525639772974e-05, 'epoch': 0.66}) +[2026-06-16 01:00:26,553][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.18887445330619812, 'train/grad_norm_pre_clip': 6.4485368728637695, 'timing/data': 0.010337876621633768, 'timing/model': 1.5560896513052285, 'learning_rate/qwen_vl_interface': 7.881991026313176e-06, 'learning_rate/action_model': 7.881991026313175e-05, 'epoch': 0.66}) +[2026-06-16 01:00:32,852][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.07008934020996094, 'train/grad_norm_pre_clip': 5.7205891609191895, 'timing/data': 0.010197075083851814, 'timing/model': 1.509156976826489, 'learning_rate/qwen_vl_interface': 7.879455330941283e-06, 'learning_rate/action_model': 7.879455330941282e-05, 'epoch': 0.66}) +[2026-06-16 01:00:39,210][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.044707898050546646, 'train/grad_norm_pre_clip': 11.211281776428223, 'timing/data': 0.010116685181856155, 'timing/model': 1.8233006806112826, 'learning_rate/qwen_vl_interface': 7.876918554699627e-06, 'learning_rate/action_model': 7.876918554699627e-05, 'epoch': 0.66}) +[2026-06-16 01:00:45,265][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.13681484758853912, 'train/grad_norm_pre_clip': 10.150087356567383, 'timing/data': 0.01121848076581955, 'timing/model': 1.5532640498131514, 'learning_rate/qwen_vl_interface': 7.874380698630982e-06, 'learning_rate/action_model': 7.874380698630982e-05, 'epoch': 0.66}) +[2026-06-16 01:00:51,559][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.11116008460521698, 'train/grad_norm_pre_clip': 8.425209045410156, 'timing/data': 0.011390771716833115, 'timing/model': 1.5610373727977276, 'learning_rate/qwen_vl_interface': 7.871841763778562e-06, 'learning_rate/action_model': 7.871841763778563e-05, 'epoch': 0.66}) +[2026-06-16 01:00:57,886][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.09975829720497131, 'train/grad_norm_pre_clip': 13.615949630737305, 'timing/data': 0.011096670757979155, 'timing/model': 1.5369168329052627, 'learning_rate/qwen_vl_interface': 7.86930175118603e-06, 'learning_rate/action_model': 7.86930175118603e-05, 'epoch': 0.66}) +[2026-06-16 01:01:04,163][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.1688358336687088, 'train/grad_norm_pre_clip': 12.00385570526123, 'timing/data': 0.010367066133767366, 'timing/model': 1.556191018782556, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.66}) +[2026-06-16 01:01:10,481][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.16395729780197144, 'train/grad_norm_pre_clip': 9.722880363464355, 'timing/data': 0.012463157996535301, 'timing/model': 1.5451508210971951, 'learning_rate/qwen_vl_interface': 7.86421849695748e-06, 'learning_rate/action_model': 7.864218496957481e-05, 'epoch': 0.66}) +[2026-06-16 01:01:16,797][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.26355767250061035, 'train/grad_norm_pre_clip': 19.218570709228516, 'timing/data': 0.011082601267844439, 'timing/model': 1.7917449483647943, 'learning_rate/qwen_vl_interface': 7.861675257410998e-06, 'learning_rate/action_model': 7.861675257410997e-05, 'epoch': 0.66}) +[2026-06-16 01:01:22,866][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.34418001770973206, 'train/grad_norm_pre_clip': 14.33077335357666, 'timing/data': 0.011508493218570948, 'timing/model': 1.5698309480212629, 'learning_rate/qwen_vl_interface': 7.859130944303468e-06, 'learning_rate/action_model': 7.859130944303467e-05, 'epoch': 0.66}) +[2026-06-16 01:01:29,156][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.21292132139205933, 'train/grad_norm_pre_clip': 12.937196731567383, 'timing/data': 0.010601487010717392, 'timing/model': 1.5632359650917351, 'learning_rate/qwen_vl_interface': 7.856585558680761e-06, 'learning_rate/action_model': 7.856585558680761e-05, 'epoch': 0.66}) +[2026-06-16 01:01:35,464][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.0375799722969532, 'train/grad_norm_pre_clip': 5.0384368896484375, 'timing/data': 0.011313370894640684, 'timing/model': 1.5658478597179055, 'learning_rate/qwen_vl_interface': 7.854039101589191e-06, 'learning_rate/action_model': 7.854039101589191e-05, 'epoch': 0.67}) +[2026-06-16 01:01:41,762][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.07578648626804352, 'train/grad_norm_pre_clip': 5.9638190269470215, 'timing/data': 0.011566073168069124, 'timing/model': 1.5606178431771696, 'learning_rate/qwen_vl_interface': 7.851491574075509e-06, 'learning_rate/action_model': 7.851491574075509e-05, 'epoch': 0.67}) +[2026-06-16 01:01:48,063][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.050093889236450195, 'train/grad_norm_pre_clip': 7.480857849121094, 'timing/data': 0.010529477149248123, 'timing/model': 1.5302809323184192, 'learning_rate/qwen_vl_interface': 7.848942977186909e-06, 'learning_rate/action_model': 7.848942977186908e-05, 'epoch': 0.67}) +[2026-06-16 01:01:54,400][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.1762004792690277, 'train/grad_norm_pre_clip': 12.173056602478027, 'timing/data': 0.011213211342692375, 'timing/model': 1.8004482900723815, 'learning_rate/qwen_vl_interface': 7.84639331197102e-06, 'learning_rate/action_model': 7.84639331197102e-05, 'epoch': 0.67}) +[2026-06-16 01:02:00,638][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.26292720437049866, 'train/grad_norm_pre_clip': 19.520336151123047, 'timing/data': 0.010804789140820503, 'timing/model': 1.7342285043559968, 'learning_rate/qwen_vl_interface': 7.843842579475918e-06, 'learning_rate/action_model': 7.843842579475918e-05, 'epoch': 0.67}) +[2026-06-16 01:02:06,663][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.15707673132419586, 'train/grad_norm_pre_clip': 15.022114753723145, 'timing/data': 0.011109419632703066, 'timing/model': 1.5646189549006522, 'learning_rate/qwen_vl_interface': 7.841290780750107e-06, 'learning_rate/action_model': 7.841290780750107e-05, 'epoch': 0.67}) +[2026-06-16 01:02:12,965][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.08875299990177155, 'train/grad_norm_pre_clip': 10.756232261657715, 'timing/data': 0.011406222358345985, 'timing/model': 1.5612383089028299, 'learning_rate/qwen_vl_interface': 7.838737916842538e-06, 'learning_rate/action_model': 7.838737916842538e-05, 'epoch': 0.67}) +[2026-06-16 01:02:19,263][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.07828685641288757, 'train/grad_norm_pre_clip': 5.1901068687438965, 'timing/data': 0.011511291842907667, 'timing/model': 1.5614748420193791, 'learning_rate/qwen_vl_interface': 7.836183988802596e-06, 'learning_rate/action_model': 7.836183988802595e-05, 'epoch': 0.67}) +[2026-06-16 01:02:25,559][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.09996097534894943, 'train/grad_norm_pre_clip': 8.55184555053711, 'timing/data': 0.011173659935593605, 'timing/model': 1.5222146338783205, 'learning_rate/qwen_vl_interface': 7.833628997680107e-06, 'learning_rate/action_model': 7.833628997680107e-05, 'epoch': 0.67}) +[2026-06-16 01:02:31,896][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.022013291716575623, 'train/grad_norm_pre_clip': 1.1318879127502441, 'timing/data': 0.01149766193702817, 'timing/model': 1.7959412140771747, 'learning_rate/qwen_vl_interface': 7.831072944525329e-06, 'learning_rate/action_model': 7.831072944525328e-05, 'epoch': 0.67}) +[2026-06-16 01:02:38,208][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.0438300259411335, 'train/grad_norm_pre_clip': 2.9787158966064453, 'timing/data': 0.011601281818002462, 'timing/model': 1.762093654833734, 'learning_rate/qwen_vl_interface': 7.828515830388956e-06, 'learning_rate/action_model': 7.828515830388956e-05, 'epoch': 0.67}) +[2026-06-16 01:02:44,265][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.10002642124891281, 'train/grad_norm_pre_clip': 9.316753387451172, 'timing/data': 0.011116490233689547, 'timing/model': 1.5617043552920222, 'learning_rate/qwen_vl_interface': 7.825957656322129e-06, 'learning_rate/action_model': 7.825957656322128e-05, 'epoch': 0.67}) +[2026-06-16 01:02:50,563][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.06779517233371735, 'train/grad_norm_pre_clip': 11.79600715637207, 'timing/data': 0.011368961073458195, 'timing/model': 1.5634572682902217, 'learning_rate/qwen_vl_interface': 7.823398423376406e-06, 'learning_rate/action_model': 7.823398423376406e-05, 'epoch': 0.67}) +[2026-06-16 01:02:56,865][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.003998355474323034, 'train/grad_norm_pre_clip': 0.39451879262924194, 'timing/data': 0.011261549778282642, 'timing/model': 1.5611568177118897, 'learning_rate/qwen_vl_interface': 7.820838132603801e-06, 'learning_rate/action_model': 7.8208381326038e-05, 'epoch': 0.67}) +[2026-06-16 01:03:03,266][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.18319469690322876, 'train/grad_norm_pre_clip': 7.150665283203125, 'timing/data': 0.011230169795453548, 'timing/model': 1.5839129826053977, 'learning_rate/qwen_vl_interface': 7.818276785056748e-06, 'learning_rate/action_model': 7.818276785056749e-05, 'epoch': 0.67}) +[2026-06-16 01:03:09,605][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.023583538830280304, 'train/grad_norm_pre_clip': 2.2852327823638916, 'timing/data': 0.0114178410731256, 'timing/model': 1.7900227047502995, 'learning_rate/qwen_vl_interface': 7.815714381788123e-06, 'learning_rate/action_model': 7.815714381788122e-05, 'epoch': 0.67}) +[2026-06-16 01:03:15,671][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.012569837272167206, 'train/grad_norm_pre_clip': 0.9797554612159729, 'timing/data': 0.01144297095015645, 'timing/model': 1.5692893918603659, 'learning_rate/qwen_vl_interface': 7.813150923851227e-06, 'learning_rate/action_model': 7.813150923851227e-05, 'epoch': 0.67}) +[2026-06-16 01:03:21,978][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.12059425562620163, 'train/grad_norm_pre_clip': 9.629029273986816, 'timing/data': 0.012857288587838411, 'timing/model': 1.5764361317269504, 'learning_rate/qwen_vl_interface': 7.810586412299807e-06, 'learning_rate/action_model': 7.810586412299806e-05, 'epoch': 0.67}) +[2026-06-16 01:03:28,266][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.08731712400913239, 'train/grad_norm_pre_clip': 8.294303894042969, 'timing/data': 0.011326320003718138, 'timing/model': 1.5641944310627878, 'learning_rate/qwen_vl_interface': 7.808020848188034e-06, 'learning_rate/action_model': 7.808020848188034e-05, 'epoch': 0.67}) +[2026-06-16 01:03:34,592][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.06284601241350174, 'train/grad_norm_pre_clip': 8.854544639587402, 'timing/data': 0.012353376019746065, 'timing/model': 1.5580223919823766, 'learning_rate/qwen_vl_interface': 7.805454232570517e-06, 'learning_rate/action_model': 7.805454232570517e-05, 'epoch': 0.67}) +[2026-06-16 01:03:40,876][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.18807581067085266, 'train/grad_norm_pre_clip': 21.61849021911621, 'timing/data': 0.011299609672278166, 'timing/model': 1.5812335889786482, 'learning_rate/qwen_vl_interface': 7.80288656650229e-06, 'learning_rate/action_model': 7.80288656650229e-05, 'epoch': 0.67}) +[2026-06-16 01:03:47,280][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.09326599538326263, 'train/grad_norm_pre_clip': 13.034032821655273, 'timing/data': 0.011420050635933876, 'timing/model': 1.7773299627006054, 'learning_rate/qwen_vl_interface': 7.800317851038829e-06, 'learning_rate/action_model': 7.800317851038828e-05, 'epoch': 0.67}) +[2026-06-16 01:03:53,418][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.15779253840446472, 'train/grad_norm_pre_clip': 14.896004676818848, 'timing/data': 0.011756073217839003, 'timing/model': 1.6161615210585296, 'learning_rate/qwen_vl_interface': 7.797748087236033e-06, 'learning_rate/action_model': 7.797748087236033e-05, 'epoch': 0.67}) +[2026-06-16 01:03:59,679][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.015563299879431725, 'train/grad_norm_pre_clip': 7.400776386260986, 'timing/data': 0.011599711142480373, 'timing/model': 1.5670545548200607, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.67}) +[2026-06-16 01:04:06,070][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.08076605945825577, 'train/grad_norm_pre_clip': 13.229109764099121, 'timing/data': 0.011566771194338799, 'timing/model': 1.5764924078248441, 'learning_rate/qwen_vl_interface': 7.792605418838196e-06, 'learning_rate/action_model': 7.792605418838197e-05, 'epoch': 0.67}) +[2026-06-16 01:04:12,362][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.27645471692085266, 'train/grad_norm_pre_clip': 23.728116989135742, 'timing/data': 0.012992509175091982, 'timing/model': 1.5218123532831669, 'learning_rate/qwen_vl_interface': 7.790032516357116e-06, 'learning_rate/action_model': 7.790032516357117e-05, 'epoch': 0.68}) +[2026-06-16 01:04:18,659][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.011248145252466202, 'train/grad_norm_pre_clip': 3.5338101387023926, 'timing/data': 0.011458500754088163, 'timing/model': 1.5163801689632237, 'learning_rate/qwen_vl_interface': 7.787458569764613e-06, 'learning_rate/action_model': 7.787458569764612e-05, 'epoch': 0.68}) +[2026-06-16 01:04:24,979][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.23206274211406708, 'train/grad_norm_pre_clip': 25.035423278808594, 'timing/data': 0.009934533387422562, 'timing/model': 1.7768410188145936, 'learning_rate/qwen_vl_interface': 7.784883580118738e-06, 'learning_rate/action_model': 7.784883580118738e-05, 'epoch': 0.68}) +[2026-06-16 01:04:31,058][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.23759838938713074, 'train/grad_norm_pre_clip': 41.17525863647461, 'timing/data': 0.010970797855407, 'timing/model': 1.5199265661649406, 'learning_rate/qwen_vl_interface': 7.782307548477975e-06, 'learning_rate/action_model': 7.782307548477975e-05, 'epoch': 0.68}) +[2026-06-16 01:04:37,358][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.0018508594948798418, 'train/grad_norm_pre_clip': 0.22564473748207092, 'timing/data': 0.0098501518368721, 'timing/model': 1.5656970920972526, 'learning_rate/qwen_vl_interface': 7.779730475901233e-06, 'learning_rate/action_model': 7.779730475901232e-05, 'epoch': 0.68}) +[2026-06-16 01:04:43,663][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.18571607768535614, 'train/grad_norm_pre_clip': 12.411535263061523, 'timing/data': 0.011488580144941807, 'timing/model': 1.5630830829031765, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 0.68}) +[2026-06-16 01:04:50,043][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.215442955493927, 'train/grad_norm_pre_clip': 25.197751998901367, 'timing/data': 0.010987448040395975, 'timing/model': 1.582204488106072, 'learning_rate/qwen_vl_interface': 7.774573212177587e-06, 'learning_rate/action_model': 7.774573212177588e-05, 'epoch': 0.68}) +[2026-06-16 01:04:56,463][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.03313925117254257, 'train/grad_norm_pre_clip': 21.645286560058594, 'timing/data': 0.010253253858536482, 'timing/model': 1.8554813121445477, 'learning_rate/qwen_vl_interface': 7.77199302315064e-06, 'learning_rate/action_model': 7.77199302315064e-05, 'epoch': 0.68}) +[2026-06-16 01:05:02,573][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.07761604338884354, 'train/grad_norm_pre_clip': 8.755090713500977, 'timing/data': 0.01077156700193882, 'timing/model': 1.5713351098820567, 'learning_rate/qwen_vl_interface': 7.769411797427622e-06, 'learning_rate/action_model': 7.769411797427623e-05, 'epoch': 0.68}) +[2026-06-16 01:05:08,861][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.01532280258834362, 'train/grad_norm_pre_clip': 9.718865394592285, 'timing/data': 0.011675151064991951, 'timing/model': 1.562037305906415, 'learning_rate/qwen_vl_interface': 7.766829536069584e-06, 'learning_rate/action_model': 7.766829536069584e-05, 'epoch': 0.68}) +[2026-06-16 01:05:15,056][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.013980668969452381, 'train/grad_norm_pre_clip': 5.072261333465576, 'timing/data': 0.010273633990436792, 'timing/model': 1.5645198142156005, 'learning_rate/qwen_vl_interface': 7.764246240137991e-06, 'learning_rate/action_model': 7.764246240137991e-05, 'epoch': 0.68}) +[2026-06-16 01:05:21,366][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.2758137583732605, 'train/grad_norm_pre_clip': 14.082833290100098, 'timing/data': 0.009931182954460382, 'timing/model': 1.5743718389421701, 'learning_rate/qwen_vl_interface': 7.761661910694741e-06, 'learning_rate/action_model': 7.761661910694741e-05, 'epoch': 0.68}) +[2026-06-16 01:05:27,657][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.013876717537641525, 'train/grad_norm_pre_clip': 2.989885091781616, 'timing/data': 0.009831031318753958, 'timing/model': 1.5749940387904644, 'learning_rate/qwen_vl_interface': 7.759076548802153e-06, 'learning_rate/action_model': 7.759076548802152e-05, 'epoch': 0.68}) +[2026-06-16 01:05:33,996][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.13417920470237732, 'train/grad_norm_pre_clip': 13.276832580566406, 'timing/data': 0.010576915927231312, 'timing/model': 1.795036697294563, 'learning_rate/qwen_vl_interface': 7.75649015552297e-06, 'learning_rate/action_model': 7.75649015552297e-05, 'epoch': 0.68}) +[2026-06-16 01:05:40,065][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.0704481229186058, 'train/grad_norm_pre_clip': 21.001413345336914, 'timing/data': 0.011550419963896275, 'timing/model': 1.5563207701779902, 'learning_rate/qwen_vl_interface': 7.753902731920364e-06, 'learning_rate/action_model': 7.753902731920364e-05, 'epoch': 0.68}) +[2026-06-16 01:05:46,482][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.06086965277791023, 'train/grad_norm_pre_clip': 22.34694480895996, 'timing/data': 0.011591570917516947, 'timing/model': 1.5688696708530188, 'learning_rate/qwen_vl_interface': 7.751314279057925e-06, 'learning_rate/action_model': 7.751314279057923e-05, 'epoch': 0.68}) +[2026-06-16 01:05:52,881][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.012898274697363377, 'train/grad_norm_pre_clip': 3.2990570068359375, 'timing/data': 0.01202394301071763, 'timing/model': 1.563507127109915, 'learning_rate/qwen_vl_interface': 7.74872479799967e-06, 'learning_rate/action_model': 7.74872479799967e-05, 'epoch': 0.68}) +[2026-06-16 01:05:59,264][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.014013879001140594, 'train/grad_norm_pre_clip': 2.784850597381592, 'timing/data': 0.011363460216671228, 'timing/model': 1.538287980016321, 'learning_rate/qwen_vl_interface': 7.746134289810034e-06, 'learning_rate/action_model': 7.746134289810033e-05, 'epoch': 0.68}) +[2026-06-16 01:06:05,565][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.3336527943611145, 'train/grad_norm_pre_clip': 16.85057258605957, 'timing/data': 0.010809026658535004, 'timing/model': 1.570568025112152, 'learning_rate/qwen_vl_interface': 7.74354275555388e-06, 'learning_rate/action_model': 7.74354275555388e-05, 'epoch': 0.68}) +[2026-06-16 01:06:11,898][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.38901740312576294, 'train/grad_norm_pre_clip': 35.72426223754883, 'timing/data': 0.011569831054657698, 'timing/model': 1.752790803089738, 'learning_rate/qwen_vl_interface': 7.740950196296488e-06, 'learning_rate/action_model': 7.740950196296488e-05, 'epoch': 0.68}) +[2026-06-16 01:06:18,058][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.3194315433502197, 'train/grad_norm_pre_clip': 14.464530944824219, 'timing/data': 0.011269528418779373, 'timing/model': 1.5622522230260074, 'learning_rate/qwen_vl_interface': 7.738356613103562e-06, 'learning_rate/action_model': 7.73835661310356e-05, 'epoch': 0.68}) +[2026-06-16 01:06:24,371][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.23325206339359283, 'train/grad_norm_pre_clip': 11.927055358886719, 'timing/data': 0.011028727982193232, 'timing/model': 1.5725752408616245, 'learning_rate/qwen_vl_interface': 7.735762007041226e-06, 'learning_rate/action_model': 7.735762007041226e-05, 'epoch': 0.68}) +[2026-06-16 01:06:30,769][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.10953392088413239, 'train/grad_norm_pre_clip': 8.416625022888184, 'timing/data': 0.011771101970225573, 'timing/model': 1.5486960620619357, 'learning_rate/qwen_vl_interface': 7.733166379176023e-06, 'learning_rate/action_model': 7.733166379176023e-05, 'epoch': 0.68}) +[2026-06-16 01:06:37,059][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.00936141237616539, 'train/grad_norm_pre_clip': 1.9491337537765503, 'timing/data': 0.012287775054574013, 'timing/model': 1.5264433808624744, 'learning_rate/qwen_vl_interface': 7.730569730574921e-06, 'learning_rate/action_model': 7.73056973057492e-05, 'epoch': 0.68}) +[2026-06-16 01:06:43,458][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.19135703146457672, 'train/grad_norm_pre_clip': 14.876300811767578, 'timing/data': 0.010980247054249048, 'timing/model': 1.5838930769823492, 'learning_rate/qwen_vl_interface': 7.727972062305299e-06, 'learning_rate/action_model': 7.727972062305299e-05, 'epoch': 0.69}) +[2026-06-16 01:06:49,793][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.011679846793413162, 'train/grad_norm_pre_clip': 0.9931684136390686, 'timing/data': 0.010815076064318419, 'timing/model': 1.7979781366884708, 'learning_rate/qwen_vl_interface': 7.725373375434965e-06, 'learning_rate/action_model': 7.725373375434965e-05, 'epoch': 0.69}) +[2026-06-16 01:06:55,860][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.10910195112228394, 'train/grad_norm_pre_clip': 7.790571689605713, 'timing/data': 0.010723426006734371, 'timing/model': 1.5644906470552087, 'learning_rate/qwen_vl_interface': 7.722773671032141e-06, 'learning_rate/action_model': 7.722773671032142e-05, 'epoch': 0.69}) +[2026-06-16 01:07:02,251][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.25701823830604553, 'train/grad_norm_pre_clip': 21.714679718017578, 'timing/data': 0.010878665838390589, 'timing/model': 1.592031696345657, 'learning_rate/qwen_vl_interface': 7.720172950165466e-06, 'learning_rate/action_model': 7.720172950165465e-05, 'epoch': 0.69}) +[2026-06-16 01:07:08,584][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.4025071859359741, 'train/grad_norm_pre_clip': 21.985363006591797, 'timing/data': 0.010797266848385334, 'timing/model': 1.5890107052400708, 'learning_rate/qwen_vl_interface': 7.717571213903996e-06, 'learning_rate/action_model': 7.717571213903996e-05, 'epoch': 0.69}) +[2026-06-16 01:07:14,882][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.11067865788936615, 'train/grad_norm_pre_clip': 10.723905563354492, 'timing/data': 0.010678795166313648, 'timing/model': 1.5876374351792037, 'learning_rate/qwen_vl_interface': 7.714968463317207e-06, 'learning_rate/action_model': 7.714968463317207e-05, 'epoch': 0.69}) +[2026-06-16 01:18:15,121][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.2553725838661194, 'train/grad_norm_pre_clip': 15.74208927154541, 'eval/loss': 0.07739269733428955, 'eval/flap_precision': 0.9276315789473685, 'eval/flap_recall': 0.9618008185538881, 'eval/flap_f1': 0.9444072337575352, 'eval/flap_rate_model': 0.2111111111111111, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9444072337575352, 'timing/data': 0.011040756944566965, 'timing/model': 1.5244483724236488, 'learning_rate/qwen_vl_interface': 7.712364699474996e-06, 'learning_rate/action_model': 7.712364699474996e-05, 'epoch': 0.69}) +[2026-06-16 01:18:15,122][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1700_state +[2026-06-16 01:18:15,128][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 01:18:23,930][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1700_state/model.safetensors +[2026-06-16 01:18:43,571][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1700_state/optimizer.bin +[2026-06-16 01:18:43,572][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1700_state/sampler.bin +[2026-06-16 01:18:43,572][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1700_state/sampler_1.bin +[2026-06-16 01:18:43,573][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1700_state/random_states_0.pkl +[2026-06-16 01:18:43,763][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1700 +[2026-06-16 01:18:43,764][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 01:18:43,779][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 01:18:50,024][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.040576424449682236, 'train/grad_norm_pre_clip': 5.102075099945068, 'timing/data': 0.011153336148709059, 'timing/model': 1.5208431649953127, 'learning_rate/qwen_vl_interface': 7.709759923447667e-06, 'learning_rate/action_model': 7.709759923447665e-05, 'epoch': 0.69}) +[2026-06-16 01:18:56,234][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.004535955842584372, 'train/grad_norm_pre_clip': 0.4296617805957794, 'timing/data': 0.011513258330523968, 'timing/model': 1.549012394156307, 'learning_rate/qwen_vl_interface': 7.707154136305945e-06, 'learning_rate/action_model': 7.707154136305945e-05, 'epoch': 0.69}) +[2026-06-16 01:19:02,501][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.07575089484453201, 'train/grad_norm_pre_clip': 10.637255668640137, 'timing/data': 0.010539013892412186, 'timing/model': 1.8285520621575415, 'learning_rate/qwen_vl_interface': 7.704547339120974e-06, 'learning_rate/action_model': 7.704547339120974e-05, 'epoch': 0.69}) +[2026-06-16 01:19:08,518][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.1124514788389206, 'train/grad_norm_pre_clip': 5.768306732177734, 'timing/data': 0.010977495927363634, 'timing/model': 1.5580263449810445, 'learning_rate/qwen_vl_interface': 7.70193953296431e-06, 'learning_rate/action_model': 7.70193953296431e-05, 'epoch': 0.69}) +[2026-06-16 01:19:14,726][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.08304145187139511, 'train/grad_norm_pre_clip': 17.95698356628418, 'timing/data': 0.01094418577849865, 'timing/model': 1.5595480529591441, 'learning_rate/qwen_vl_interface': 7.69933071890792e-06, 'learning_rate/action_model': 7.699330718907919e-05, 'epoch': 0.69}) +[2026-06-16 01:19:20,939][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.11046898365020752, 'train/grad_norm_pre_clip': 11.845285415649414, 'timing/data': 0.011309847235679626, 'timing/model': 1.4953090748749673, 'learning_rate/qwen_vl_interface': 7.696720898024188e-06, 'learning_rate/action_model': 7.69672089802419e-05, 'epoch': 0.69}) +[2026-06-16 01:19:27,217][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.0054774293676018715, 'train/grad_norm_pre_clip': 0.536681592464447, 'timing/data': 0.011677269823849201, 'timing/model': 1.5603696191683412, 'learning_rate/qwen_vl_interface': 7.69411007138592e-06, 'learning_rate/action_model': 7.694110071385919e-05, 'epoch': 0.69}) +[2026-06-16 01:19:33,549][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.02189655601978302, 'train/grad_norm_pre_clip': 1.9275771379470825, 'timing/data': 0.011154856998473406, 'timing/model': 1.5502392589114606, 'learning_rate/qwen_vl_interface': 7.691498240066322e-06, 'learning_rate/action_model': 7.691498240066322e-05, 'epoch': 0.69}) +[2026-06-16 01:19:40,133][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.1364312469959259, 'train/grad_norm_pre_clip': 10.637043952941895, 'timing/data': 0.03573888074606657, 'timing/model': 1.8268033973872662, 'learning_rate/qwen_vl_interface': 7.688885405139022e-06, 'learning_rate/action_model': 7.688885405139021e-05, 'epoch': 0.69}) +[2026-06-16 01:19:46,145][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.014157055877149105, 'train/grad_norm_pre_clip': 2.1824376583099365, 'timing/data': 0.017466084100306034, 'timing/model': 1.5558476336300373, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.69}) +[2026-06-16 01:19:52,456][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.011264081113040447, 'train/grad_norm_pre_clip': 4.23000431060791, 'timing/data': 0.011259507853537798, 'timing/model': 1.5687315198592842, 'learning_rate/qwen_vl_interface': 7.683656728757873e-06, 'learning_rate/action_model': 7.683656728757872e-05, 'epoch': 0.69}) +[2026-06-16 01:19:58,653][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.03183837980031967, 'train/grad_norm_pre_clip': 5.486690998077393, 'timing/data': 0.011416198220103979, 'timing/model': 1.570692707784474, 'learning_rate/qwen_vl_interface': 7.681040889453337e-06, 'learning_rate/action_model': 7.681040889453336e-05, 'epoch': 0.69}) +[2026-06-16 01:20:04,954][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.01452871598303318, 'train/grad_norm_pre_clip': 2.5641322135925293, 'timing/data': 0.011240518186241388, 'timing/model': 1.5639845398254693, 'learning_rate/qwen_vl_interface': 7.678424050839722e-06, 'learning_rate/action_model': 7.67842405083972e-05, 'epoch': 0.69}) +[2026-06-16 01:20:11,263][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.11775925755500793, 'train/grad_norm_pre_clip': 17.693483352661133, 'timing/data': 0.011311338283121586, 'timing/model': 1.516300689894706, 'learning_rate/qwen_vl_interface': 7.675806213992706e-06, 'learning_rate/action_model': 7.675806213992707e-05, 'epoch': 0.69}) +[2026-06-16 01:20:17,580][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.1128506064414978, 'train/grad_norm_pre_clip': 11.355705261230469, 'timing/data': 0.010984356980770826, 'timing/model': 1.7849297942593694, 'learning_rate/qwen_vl_interface': 7.673187379988388e-06, 'learning_rate/action_model': 7.673187379988388e-05, 'epoch': 0.69}) +[2026-06-16 01:20:23,656][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.21145856380462646, 'train/grad_norm_pre_clip': 12.732856750488281, 'timing/data': 0.010778165888041258, 'timing/model': 1.558996242005378, 'learning_rate/qwen_vl_interface': 7.67056754990327e-06, 'learning_rate/action_model': 7.670567549903269e-05, 'epoch': 0.69}) +[2026-06-16 01:20:30,038][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.0656101405620575, 'train/grad_norm_pre_clip': 4.552816867828369, 'timing/data': 0.012002161238342524, 'timing/model': 1.6476849312894046, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.69}) +[2026-06-16 01:20:36,340][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.010762166231870651, 'train/grad_norm_pre_clip': 3.155989170074463, 'timing/data': 0.01084652729332447, 'timing/model': 1.5736619750969112, 'learning_rate/qwen_vl_interface': 7.665324905798694e-06, 'learning_rate/action_model': 7.665324905798695e-05, 'epoch': 0.69}) +[2026-06-16 01:20:42,724][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.042703378945589066, 'train/grad_norm_pre_clip': 20.752981185913086, 'timing/data': 0.01102830795571208, 'timing/model': 1.6392236058600247, 'learning_rate/qwen_vl_interface': 7.66270209393429e-06, 'learning_rate/action_model': 7.66270209393429e-05, 'epoch': 0.7}) +[2026-06-16 01:20:49,046][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.007079023402184248, 'train/grad_norm_pre_clip': 1.1674374341964722, 'timing/data': 0.015887347981333733, 'timing/model': 1.5325638391077518, 'learning_rate/qwen_vl_interface': 7.66007829029919e-06, 'learning_rate/action_model': 7.660078290299191e-05, 'epoch': 0.7}) +[2026-06-16 01:20:55,537][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.008534694090485573, 'train/grad_norm_pre_clip': 1.0034205913543701, 'timing/data': 0.010965866968035698, 'timing/model': 1.8534376788884401, 'learning_rate/qwen_vl_interface': 7.657453495971942e-06, 'learning_rate/action_model': 7.657453495971943e-05, 'epoch': 0.7}) +[2026-06-16 01:21:01,793][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.004578473046422005, 'train/grad_norm_pre_clip': 0.8204206824302673, 'timing/data': 0.01164992991834879, 'timing/model': 1.7914121430367231, 'learning_rate/qwen_vl_interface': 7.6548277120315e-06, 'learning_rate/action_model': 7.6548277120315e-05, 'epoch': 0.7}) +[2026-06-16 01:21:07,852][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.1965109258890152, 'train/grad_norm_pre_clip': 17.49564552307129, 'timing/data': 0.010715195909142494, 'timing/model': 1.558538669720292, 'learning_rate/qwen_vl_interface': 7.652200939557225e-06, 'learning_rate/action_model': 7.652200939557225e-05, 'epoch': 0.7}) +[2026-06-16 01:21:14,161][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.2532055079936981, 'train/grad_norm_pre_clip': 17.979690551757812, 'timing/data': 0.010991087183356285, 'timing/model': 1.5645532454364002, 'learning_rate/qwen_vl_interface': 7.649573179628881e-06, 'learning_rate/action_model': 7.649573179628881e-05, 'epoch': 0.7}) +[2026-06-16 01:21:20,463][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.0026670480147004128, 'train/grad_norm_pre_clip': 0.5556786060333252, 'timing/data': 0.01155993016436696, 'timing/model': 1.5432666521519423, 'learning_rate/qwen_vl_interface': 7.646944433326645e-06, 'learning_rate/action_model': 7.646944433326643e-05, 'epoch': 0.7}) +[2026-06-16 01:21:26,848][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.08714701980352402, 'train/grad_norm_pre_clip': 14.413776397705078, 'timing/data': 0.011076247785240412, 'timing/model': 1.5866199997253716, 'learning_rate/qwen_vl_interface': 7.64431470173109e-06, 'learning_rate/action_model': 7.64431470173109e-05, 'epoch': 0.7}) +[2026-06-16 01:21:33,198][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.10274504125118256, 'train/grad_norm_pre_clip': 7.178023815155029, 'timing/data': 0.01162515114992857, 'timing/model': 1.8028491362929344, 'learning_rate/qwen_vl_interface': 7.641683985923203e-06, 'learning_rate/action_model': 7.641683985923203e-05, 'epoch': 0.7}) +[2026-06-16 01:21:39,319][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.3324449062347412, 'train/grad_norm_pre_clip': 48.863365173339844, 'timing/data': 0.011294409167021513, 'timing/model': 1.622465563006699, 'learning_rate/qwen_vl_interface': 7.63905228698437e-06, 'learning_rate/action_model': 7.63905228698437e-05, 'epoch': 0.7}) +[2026-06-16 01:21:45,563][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.09830699861049652, 'train/grad_norm_pre_clip': 26.711183547973633, 'timing/data': 0.012549625243991613, 'timing/model': 1.5659394268877804, 'learning_rate/qwen_vl_interface': 7.636419605996385e-06, 'learning_rate/action_model': 7.636419605996384e-05, 'epoch': 0.7}) +[2026-06-16 01:21:51,861][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.12550172209739685, 'train/grad_norm_pre_clip': 21.13616943359375, 'timing/data': 0.012089232914149761, 'timing/model': 1.5654432228766382, 'learning_rate/qwen_vl_interface': 7.633785944041444e-06, 'learning_rate/action_model': 7.633785944041444e-05, 'epoch': 0.7}) +[2026-06-16 01:21:58,160][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.262842059135437, 'train/grad_norm_pre_clip': 17.104368209838867, 'timing/data': 0.010815986897796392, 'timing/model': 1.530311749316752, 'learning_rate/qwen_vl_interface': 7.631151302202145e-06, 'learning_rate/action_model': 7.631151302202144e-05, 'epoch': 0.7}) +[2026-06-16 01:22:04,464][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.06338150054216385, 'train/grad_norm_pre_clip': 6.630174160003662, 'timing/data': 0.011122237890958786, 'timing/model': 1.5902246492914855, 'learning_rate/qwen_vl_interface': 7.628515681561488e-06, 'learning_rate/action_model': 7.628515681561487e-05, 'epoch': 0.7}) +[2026-06-16 01:22:10,747][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.41620853543281555, 'train/grad_norm_pre_clip': 54.84462356567383, 'timing/data': 0.010820078197866678, 'timing/model': 1.7226616060361266, 'learning_rate/qwen_vl_interface': 7.62587908320288e-06, 'learning_rate/action_model': 7.62587908320288e-05, 'epoch': 0.7}) +[2026-06-16 01:22:16,865][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.18106907606124878, 'train/grad_norm_pre_clip': 38.22908401489258, 'timing/data': 0.010652096942067146, 'timing/model': 1.571228631772101, 'learning_rate/qwen_vl_interface': 7.6232415082101255e-06, 'learning_rate/action_model': 7.623241508210124e-05, 'epoch': 0.7}) +[2026-06-16 01:22:23,236][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.1438692957162857, 'train/grad_norm_pre_clip': 15.648527145385742, 'timing/data': 0.011495050974190235, 'timing/model': 1.5806059017777443, 'learning_rate/qwen_vl_interface': 7.620602957667432e-06, 'learning_rate/action_model': 7.620602957667432e-05, 'epoch': 0.7}) +[2026-06-16 01:22:29,459][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.11397357285022736, 'train/grad_norm_pre_clip': 33.946868896484375, 'timing/data': 0.011777892243117094, 'timing/model': 1.556758294813335, 'learning_rate/qwen_vl_interface': 7.617963432659408e-06, 'learning_rate/action_model': 7.617963432659408e-05, 'epoch': 0.7}) +[2026-06-16 01:22:35,756][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.09230726212263107, 'train/grad_norm_pre_clip': 8.521541595458984, 'timing/data': 0.011576800607144833, 'timing/model': 1.5224326527677476, 'learning_rate/qwen_vl_interface': 7.615322934271063e-06, 'learning_rate/action_model': 7.615322934271064e-05, 'epoch': 0.7}) +[2026-06-16 01:22:42,145][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.23426687717437744, 'train/grad_norm_pre_clip': 8.225481033325195, 'timing/data': 0.011319498997181654, 'timing/model': 1.5817188816145062, 'learning_rate/qwen_vl_interface': 7.6126814635878055e-06, 'learning_rate/action_model': 7.612681463587805e-05, 'epoch': 0.7}) +[2026-06-16 01:22:48,466][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.4130534529685974, 'train/grad_norm_pre_clip': 41.835445404052734, 'timing/data': 0.01083880802616477, 'timing/model': 1.7593285609036684, 'learning_rate/qwen_vl_interface': 7.6100390216954445e-06, 'learning_rate/action_model': 7.610039021695445e-05, 'epoch': 0.7}) +[2026-06-16 01:22:54,565][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.03292465582489967, 'train/grad_norm_pre_clip': 4.489924907684326, 'timing/data': 0.011762161739170551, 'timing/model': 1.5667704911902547, 'learning_rate/qwen_vl_interface': 7.607395609680189e-06, 'learning_rate/action_model': 7.60739560968019e-05, 'epoch': 0.7}) +[2026-06-16 01:23:00,863][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.2494245022535324, 'train/grad_norm_pre_clip': 1.9261999130249023, 'timing/data': 0.011573981028050184, 'timing/model': 1.5584852020256221, 'learning_rate/qwen_vl_interface': 7.604751228628644e-06, 'learning_rate/action_model': 7.604751228628644e-05, 'epoch': 0.7}) +[2026-06-16 01:23:07,157][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.03013974241912365, 'train/grad_norm_pre_clip': 19.30499839782715, 'timing/data': 0.011284450069069862, 'timing/model': 1.5614884332753718, 'learning_rate/qwen_vl_interface': 7.602105879627814e-06, 'learning_rate/action_model': 7.602105879627814e-05, 'epoch': 0.7}) +[2026-06-16 01:23:13,519][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.20216946303844452, 'train/grad_norm_pre_clip': 13.989904403686523, 'timing/data': 0.011525451205670834, 'timing/model': 1.5675521078519523, 'learning_rate/qwen_vl_interface': 7.599459563765107e-06, 'learning_rate/action_model': 7.599459563765106e-05, 'epoch': 0.7}) +[2026-06-16 01:23:19,810][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.004231826867908239, 'train/grad_norm_pre_clip': 0.9444169998168945, 'timing/data': 0.01092472905293107, 'timing/model': 1.8150093592703342, 'learning_rate/qwen_vl_interface': 7.596812282128319e-06, 'learning_rate/action_model': 7.596812282128319e-05, 'epoch': 0.71}) +[2026-06-16 01:23:25,871][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.01899975910782814, 'train/grad_norm_pre_clip': 5.368850231170654, 'timing/data': 0.011505910661071539, 'timing/model': 1.5728729059919715, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 0.71}) +[2026-06-16 01:23:32,161][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.20266225934028625, 'train/grad_norm_pre_clip': 12.72562026977539, 'timing/data': 0.01152269123122096, 'timing/model': 1.5583557281643152, 'learning_rate/qwen_vl_interface': 7.59151482588569e-06, 'learning_rate/action_model': 7.59151482588569e-05, 'epoch': 0.71}) +[2026-06-16 01:23:38,363][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.12115216255187988, 'train/grad_norm_pre_clip': 9.263127326965332, 'timing/data': 0.010670537129044533, 'timing/model': 1.5600082529708743, 'learning_rate/qwen_vl_interface': 7.588864653457433e-06, 'learning_rate/action_model': 7.588864653457432e-05, 'epoch': 0.71}) +[2026-06-16 01:23:44,661][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.1747959554195404, 'train/grad_norm_pre_clip': 30.220258712768555, 'timing/data': 0.011484381277114153, 'timing/model': 1.557045937050134, 'learning_rate/qwen_vl_interface': 7.5862135196102634e-06, 'learning_rate/action_model': 7.586213519610263e-05, 'epoch': 0.71}) +[2026-06-16 01:23:50,965][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.16441749036312103, 'train/grad_norm_pre_clip': 6.5874552726745605, 'timing/data': 0.013043458107858896, 'timing/model': 1.5551316449418664, 'learning_rate/qwen_vl_interface': 7.583561425433962e-06, 'learning_rate/action_model': 7.583561425433962e-05, 'epoch': 0.71}) +[2026-06-16 01:23:57,305][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.12745735049247742, 'train/grad_norm_pre_clip': 4.09262228012085, 'timing/data': 0.011265949811786413, 'timing/model': 1.7976351510733366, 'learning_rate/qwen_vl_interface': 7.580908372018706e-06, 'learning_rate/action_model': 7.580908372018705e-05, 'epoch': 0.71}) +[2026-06-16 01:24:03,366][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.25355252623558044, 'train/grad_norm_pre_clip': 9.760880470275879, 'timing/data': 0.011853322852402925, 'timing/model': 1.545832637231797, 'learning_rate/qwen_vl_interface': 7.5782543604550655e-06, 'learning_rate/action_model': 7.578254360455064e-05, 'epoch': 0.71}) +[2026-06-16 01:24:09,664][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.18571442365646362, 'train/grad_norm_pre_clip': 15.719054222106934, 'timing/data': 0.011598842218518257, 'timing/model': 1.555922248866409, 'learning_rate/qwen_vl_interface': 7.575599391834e-06, 'learning_rate/action_model': 7.575599391834e-05, 'epoch': 0.71}) +[2026-06-16 01:24:15,995][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.1500728577375412, 'train/grad_norm_pre_clip': 8.82734489440918, 'timing/data': 0.010968939401209354, 'timing/model': 1.5934117618016899, 'learning_rate/qwen_vl_interface': 7.572943467246871e-06, 'learning_rate/action_model': 7.57294346724687e-05, 'epoch': 0.71}) +[2026-06-16 01:24:22,346][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.13195380568504333, 'train/grad_norm_pre_clip': 4.688978672027588, 'timing/data': 0.011336901225149632, 'timing/model': 1.5869515999220312, 'learning_rate/qwen_vl_interface': 7.570286587785428e-06, 'learning_rate/action_model': 7.570286587785428e-05, 'epoch': 0.71}) +[2026-06-16 01:24:28,672][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.09130649268627167, 'train/grad_norm_pre_clip': 7.546969890594482, 'timing/data': 0.011307960376143456, 'timing/model': 1.5476109110750258, 'learning_rate/qwen_vl_interface': 7.567628754541815e-06, 'learning_rate/action_model': 7.567628754541814e-05, 'epoch': 0.71}) +[2026-06-16 01:24:35,002][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.0967286229133606, 'train/grad_norm_pre_clip': 5.731852054595947, 'timing/data': 0.010889098048210144, 'timing/model': 1.8030410907231271, 'learning_rate/qwen_vl_interface': 7.564969968608565e-06, 'learning_rate/action_model': 7.564969968608565e-05, 'epoch': 0.71}) +[2026-06-16 01:24:41,126][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.027252450585365295, 'train/grad_norm_pre_clip': 2.241931915283203, 'timing/data': 0.011855063028633595, 'timing/model': 1.6128308563493192, 'learning_rate/qwen_vl_interface': 7.562310231078605e-06, 'learning_rate/action_model': 7.562310231078604e-05, 'epoch': 0.71}) +[2026-06-16 01:24:47,361][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.1833345741033554, 'train/grad_norm_pre_clip': 5.391205787658691, 'timing/data': 0.011764123104512691, 'timing/model': 1.543355151079595, 'learning_rate/qwen_vl_interface': 7.559649543045252e-06, 'learning_rate/action_model': 7.559649543045252e-05, 'epoch': 0.71}) +[2026-06-16 01:24:53,750][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.07755422592163086, 'train/grad_norm_pre_clip': 13.885458946228027, 'timing/data': 0.011093129869550467, 'timing/model': 1.648699658922851, 'learning_rate/qwen_vl_interface': 7.556987905602215e-06, 'learning_rate/action_model': 7.556987905602215e-05, 'epoch': 0.71}) +[2026-06-16 01:25:00,067][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.07840878516435623, 'train/grad_norm_pre_clip': 4.208537578582764, 'timing/data': 0.011265891138464212, 'timing/model': 1.5298970523290336, 'learning_rate/qwen_vl_interface': 7.554325319843593e-06, 'learning_rate/action_model': 7.554325319843592e-05, 'epoch': 0.71}) +[2026-06-16 01:25:06,370][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.11447952687740326, 'train/grad_norm_pre_clip': 3.7176625728607178, 'timing/data': 0.010953038930892944, 'timing/model': 1.5592785207554698, 'learning_rate/qwen_vl_interface': 7.551661786863874e-06, 'learning_rate/action_model': 7.551661786863874e-05, 'epoch': 0.71}) +[2026-06-16 01:25:12,702][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.12118460237979889, 'train/grad_norm_pre_clip': 8.108159065246582, 'timing/data': 0.01092188898473978, 'timing/model': 1.797582388855517, 'learning_rate/qwen_vl_interface': 7.548997307757936e-06, 'learning_rate/action_model': 7.548997307757935e-05, 'epoch': 0.71}) +[2026-06-16 01:25:18,765][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.2571057379245758, 'train/grad_norm_pre_clip': 20.76243019104004, 'timing/data': 0.01182253286242485, 'timing/model': 1.5615487420000136, 'learning_rate/qwen_vl_interface': 7.546331883621044e-06, 'learning_rate/action_model': 7.546331883621044e-05, 'epoch': 0.71}) +[2026-06-16 01:25:25,066][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.0748043954372406, 'train/grad_norm_pre_clip': 7.507462978363037, 'timing/data': 0.011317361146211624, 'timing/model': 1.565174164250493, 'learning_rate/qwen_vl_interface': 7.543665515548857e-06, 'learning_rate/action_model': 7.543665515548856e-05, 'epoch': 0.71}) +[2026-06-16 01:25:31,366][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.14864131808280945, 'train/grad_norm_pre_clip': 19.223777770996094, 'timing/data': 0.011292401235550642, 'timing/model': 1.5612186333164573, 'learning_rate/qwen_vl_interface': 7.540998204637413e-06, 'learning_rate/action_model': 7.540998204637413e-05, 'epoch': 0.71}) +[2026-06-16 01:25:37,661][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.0639905333518982, 'train/grad_norm_pre_clip': 4.381036758422852, 'timing/data': 0.01110842963680625, 'timing/model': 1.5602967729791999, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.71}) +[2026-06-16 01:25:44,054][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.03451152890920639, 'train/grad_norm_pre_clip': 2.078953504562378, 'timing/data': 0.011556053068488836, 'timing/model': 1.5957549391314387, 'learning_rate/qwen_vl_interface': 7.535660758682877e-06, 'learning_rate/action_model': 7.535660758682876e-05, 'epoch': 0.71}) +[2026-06-16 01:25:50,479][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.13497139513492584, 'train/grad_norm_pre_clip': 16.69384002685547, 'timing/data': 0.01121803093701601, 'timing/model': 1.8700760449282825, 'learning_rate/qwen_vl_interface': 7.532990625833806e-06, 'learning_rate/action_model': 7.532990625833805e-05, 'epoch': 0.71}) +[2026-06-16 01:25:56,569][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.005730392411351204, 'train/grad_norm_pre_clip': 0.24924136698246002, 'timing/data': 0.01159261167049408, 'timing/model': 1.5238689063116908, 'learning_rate/qwen_vl_interface': 7.5303195545335225e-06, 'learning_rate/action_model': 7.530319554533522e-05, 'epoch': 0.72}) +[2026-06-16 01:26:02,869][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.07468650490045547, 'train/grad_norm_pre_clip': 12.252130508422852, 'timing/data': 0.01119556138291955, 'timing/model': 1.562801537103951, 'learning_rate/qwen_vl_interface': 7.527647545880007e-06, 'learning_rate/action_model': 7.527647545880007e-05, 'epoch': 0.72}) +[2026-06-16 01:26:09,169][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.19332650303840637, 'train/grad_norm_pre_clip': 16.880401611328125, 'timing/data': 0.011191420257091522, 'timing/model': 1.5603316207416356, 'learning_rate/qwen_vl_interface': 7.52497460097162e-06, 'learning_rate/action_model': 7.52497460097162e-05, 'epoch': 0.72}) +[2026-06-16 01:26:15,468][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.13058367371559143, 'train/grad_norm_pre_clip': 9.484872817993164, 'timing/data': 0.01136004226282239, 'timing/model': 1.563152709044516, 'learning_rate/qwen_vl_interface': 7.522300720907105e-06, 'learning_rate/action_model': 7.522300720907105e-05, 'epoch': 0.72}) +[2026-06-16 01:26:21,772][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.02116159163415432, 'train/grad_norm_pre_clip': 22.928470611572266, 'timing/data': 0.011617633048444986, 'timing/model': 1.5280868150293827, 'learning_rate/qwen_vl_interface': 7.519625906785601e-06, 'learning_rate/action_model': 7.5196259067856e-05, 'epoch': 0.72}) +[2026-06-16 01:26:28,105][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.02725113369524479, 'train/grad_norm_pre_clip': 4.954782962799072, 'timing/data': 0.011720533017069101, 'timing/model': 1.799553843215108, 'learning_rate/qwen_vl_interface': 7.516950159706614e-06, 'learning_rate/action_model': 7.516950159706614e-05, 'epoch': 0.72}) +[2026-06-16 01:26:34,182][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.01305883564054966, 'train/grad_norm_pre_clip': 1.8170558214187622, 'timing/data': 0.011157970875501633, 'timing/model': 1.5741641344502568, 'learning_rate/qwen_vl_interface': 7.514273480770047e-06, 'learning_rate/action_model': 7.514273480770047e-05, 'epoch': 0.72}) +[2026-06-16 01:26:40,606][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.08681407570838928, 'train/grad_norm_pre_clip': 8.486080169677734, 'timing/data': 0.011812652926892042, 'timing/model': 1.5861614239402115, 'learning_rate/qwen_vl_interface': 7.511595871076181e-06, 'learning_rate/action_model': 7.511595871076181e-05, 'epoch': 0.72}) +[2026-06-16 01:26:46,972][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.18946994841098785, 'train/grad_norm_pre_clip': 11.283751487731934, 'timing/data': 0.011784004047513008, 'timing/model': 1.5588652649894357, 'learning_rate/qwen_vl_interface': 7.50891733172568e-06, 'learning_rate/action_model': 7.50891733172568e-05, 'epoch': 0.72}) +[2026-06-16 01:26:53,266][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.13923989236354828, 'train/grad_norm_pre_clip': 7.613075256347656, 'timing/data': 0.011219000909477472, 'timing/model': 1.559106471017003, 'learning_rate/qwen_vl_interface': 7.506237863819593e-06, 'learning_rate/action_model': 7.506237863819593e-05, 'epoch': 0.72}) +[2026-06-16 01:26:59,565][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.09057598561048508, 'train/grad_norm_pre_clip': 14.40737247467041, 'timing/data': 0.011349550914019346, 'timing/model': 1.5267414380796254, 'learning_rate/qwen_vl_interface': 7.503557468459345e-06, 'learning_rate/action_model': 7.503557468459344e-05, 'epoch': 0.72}) +[2026-06-16 01:27:05,905][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.010545779019594193, 'train/grad_norm_pre_clip': 2.461653232574463, 'timing/data': 0.010972320102155209, 'timing/model': 1.8022834500297904, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.72}) +[2026-06-16 01:27:12,067][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.238754004240036, 'train/grad_norm_pre_clip': 11.852335929870605, 'timing/data': 0.010800458956509829, 'timing/model': 1.5631264727562666, 'learning_rate/qwen_vl_interface': 7.498193899783983e-06, 'learning_rate/action_model': 7.498193899783983e-05, 'epoch': 0.72}) +[2026-06-16 01:27:18,366][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.08976142853498459, 'train/grad_norm_pre_clip': 8.820429801940918, 'timing/data': 0.010784529149532318, 'timing/model': 1.5659784507006407, 'learning_rate/qwen_vl_interface': 7.495510728673631e-06, 'learning_rate/action_model': 7.495510728673632e-05, 'epoch': 0.72}) +[2026-06-16 01:27:24,692][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.1829918771982193, 'train/grad_norm_pre_clip': 7.841061592102051, 'timing/data': 0.011128040961921215, 'timing/model': 1.5916680530644953, 'learning_rate/qwen_vl_interface': 7.492826634518642e-06, 'learning_rate/action_model': 7.492826634518641e-05, 'epoch': 0.72}) +[2026-06-16 01:27:30,965][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.00871769804507494, 'train/grad_norm_pre_clip': 0.9595891237258911, 'timing/data': 0.011384362354874611, 'timing/model': 1.5430968659929931, 'learning_rate/qwen_vl_interface': 7.4901416184223404e-06, 'learning_rate/action_model': 7.49014161842234e-05, 'epoch': 0.72}) +[2026-06-16 01:27:37,300][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.00943720806390047, 'train/grad_norm_pre_clip': 2.1471047401428223, 'timing/data': 0.010892960242927074, 'timing/model': 1.5525724133476615, 'learning_rate/qwen_vl_interface': 7.487455681488437e-06, 'learning_rate/action_model': 7.487455681488436e-05, 'epoch': 0.72}) +[2026-06-16 01:27:43,600][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.18769681453704834, 'train/grad_norm_pre_clip': 19.642452239990234, 'timing/data': 0.011936844792217016, 'timing/model': 1.7858418356627226, 'learning_rate/qwen_vl_interface': 7.484768824821018e-06, 'learning_rate/action_model': 7.484768824821018e-05, 'epoch': 0.72}) +[2026-06-16 01:27:49,835][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.24219265580177307, 'train/grad_norm_pre_clip': 19.23150634765625, 'timing/data': 0.013054290320724249, 'timing/model': 1.727231802418828, 'learning_rate/qwen_vl_interface': 7.482081049524551e-06, 'learning_rate/action_model': 7.482081049524551e-05, 'epoch': 0.72}) +[2026-06-16 01:27:55,866][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.11837361752986908, 'train/grad_norm_pre_clip': 12.344865798950195, 'timing/data': 0.011193091981112957, 'timing/model': 1.5567072853446007, 'learning_rate/qwen_vl_interface': 7.479392356703878e-06, 'learning_rate/action_model': 7.479392356703878e-05, 'epoch': 0.72}) +[2026-06-16 01:28:02,171][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.08320145308971405, 'train/grad_norm_pre_clip': 14.134881973266602, 'timing/data': 0.011681562755256891, 'timing/model': 1.5630713691934943, 'learning_rate/qwen_vl_interface': 7.476702747464219e-06, 'learning_rate/action_model': 7.476702747464219e-05, 'epoch': 0.72}) +[2026-06-16 01:28:08,478][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.03949758782982826, 'train/grad_norm_pre_clip': 8.38828182220459, 'timing/data': 0.01163880294188857, 'timing/model': 1.5753879300318658, 'learning_rate/qwen_vl_interface': 7.474012222911169e-06, 'learning_rate/action_model': 7.474012222911169e-05, 'epoch': 0.72}) +[2026-06-16 01:28:14,854][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.01708272472023964, 'train/grad_norm_pre_clip': 20.947301864624023, 'timing/data': 0.011563072912395, 'timing/model': 1.584356274921447, 'learning_rate/qwen_vl_interface': 7.471320784150703e-06, 'learning_rate/action_model': 7.471320784150703e-05, 'epoch': 0.72}) +[2026-06-16 01:28:21,209][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.152217835187912, 'train/grad_norm_pre_clip': 18.386558532714844, 'timing/data': 0.01206819573417306, 'timing/model': 1.8085494078695774, 'learning_rate/qwen_vl_interface': 7.468628432289169e-06, 'learning_rate/action_model': 7.468628432289168e-05, 'epoch': 0.72}) +[2026-06-16 01:28:27,507][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.06111384555697441, 'train/grad_norm_pre_clip': 17.22583770751953, 'timing/data': 0.011412032879889011, 'timing/model': 1.810662193223834, 'learning_rate/qwen_vl_interface': 7.465935168433292e-06, 'learning_rate/action_model': 7.465935168433292e-05, 'epoch': 0.73}) +[2026-06-16 01:28:33,563][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.1480807662010193, 'train/grad_norm_pre_clip': 15.08401107788086, 'timing/data': 0.011700174305588007, 'timing/model': 1.5688063059933484, 'learning_rate/qwen_vl_interface': 7.463240993690168e-06, 'learning_rate/action_model': 7.463240993690168e-05, 'epoch': 0.73}) +[2026-06-16 01:28:39,856][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.1513720452785492, 'train/grad_norm_pre_clip': 9.7635498046875, 'timing/data': 0.011122960597276688, 'timing/model': 1.5602979199029505, 'learning_rate/qwen_vl_interface': 7.460545909167276e-06, 'learning_rate/action_model': 7.460545909167275e-05, 'epoch': 0.73}) +[2026-06-16 01:28:46,154][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.011231208220124245, 'train/grad_norm_pre_clip': 3.7554869651794434, 'timing/data': 0.01116899074986577, 'timing/model': 1.5583203248679638, 'learning_rate/qwen_vl_interface': 7.457849915972459e-06, 'learning_rate/action_model': 7.457849915972458e-05, 'epoch': 0.73}) +[2026-06-16 01:28:52,547][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.11326289921998978, 'train/grad_norm_pre_clip': 17.117141723632812, 'timing/data': 0.01159638399258256, 'timing/model': 1.5955389202572405, 'learning_rate/qwen_vl_interface': 7.455153015213941e-06, 'learning_rate/action_model': 7.45515301521394e-05, 'epoch': 0.73}) +[2026-06-16 01:28:58,938][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.16090001165866852, 'train/grad_norm_pre_clip': 6.162119388580322, 'timing/data': 0.010936470236629248, 'timing/model': 1.8386305449530482, 'learning_rate/qwen_vl_interface': 7.452455208000311e-06, 'learning_rate/action_model': 7.452455208000311e-05, 'epoch': 0.73}) +[2026-06-16 01:29:05,058][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.1198958158493042, 'train/grad_norm_pre_clip': 5.980578422546387, 'timing/data': 0.011365602724254131, 'timing/model': 1.5615593907423317, 'learning_rate/qwen_vl_interface': 7.449756495440541e-06, 'learning_rate/action_model': 7.44975649544054e-05, 'epoch': 0.73}) +[2026-06-16 01:39:53,522][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.16439677774906158, 'train/grad_norm_pre_clip': 42.58583450317383, 'eval/loss': 0.06157466396689415, 'eval/flap_precision': 0.9648382559774965, 'eval/flap_recall': 0.9358799454297408, 'eval/flap_f1': 0.9501385041551247, 'eval/flap_rate_model': 0.1975, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9501385041551247, 'timing/data': 0.011794925201684237, 'timing/model': 1.5701991552487016, 'learning_rate/qwen_vl_interface': 7.447056878643969e-06, 'learning_rate/action_model': 7.447056878643968e-05, 'epoch': 0.73}) +[2026-06-16 01:39:53,523][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1800_state +[2026-06-16 01:39:53,529][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 01:40:02,377][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1800_state/model.safetensors +[2026-06-16 01:40:14,276][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1800_state/optimizer.bin +[2026-06-16 01:40:14,276][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1800_state/sampler.bin +[2026-06-16 01:40:14,277][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1800_state/sampler_1.bin +[2026-06-16 01:40:14,278][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1800_state/random_states_0.pkl +[2026-06-16 01:40:14,480][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1800 +[2026-06-16 01:40:14,481][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 01:40:14,495][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 01:40:20,432][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.024624869227409363, 'train/grad_norm_pre_clip': 3.569239377975464, 'timing/data': 0.011895286850631237, 'timing/model': 1.4866353683173656, 'learning_rate/qwen_vl_interface': 7.444356358720303e-06, 'learning_rate/action_model': 7.444356358720302e-05, 'epoch': 0.73}) +[2026-06-16 01:40:26,623][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.09833309054374695, 'train/grad_norm_pre_clip': 9.647599220275879, 'timing/data': 0.011201583780348301, 'timing/model': 1.5548411030322313, 'learning_rate/qwen_vl_interface': 7.4416549367796255e-06, 'learning_rate/action_model': 7.441654936779625e-05, 'epoch': 0.73}) +[2026-06-16 01:40:32,828][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.17729198932647705, 'train/grad_norm_pre_clip': 11.043886184692383, 'timing/data': 0.014020957052707672, 'timing/model': 1.4979686229489744, 'learning_rate/qwen_vl_interface': 7.4389526139323895e-06, 'learning_rate/action_model': 7.438952613932389e-05, 'epoch': 0.73}) +[2026-06-16 01:40:39,038][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.24650101363658905, 'train/grad_norm_pre_clip': 11.714344024658203, 'timing/data': 0.011683325748890638, 'timing/model': 1.5189576540142298, 'learning_rate/qwen_vl_interface': 7.436249391289419e-06, 'learning_rate/action_model': 7.436249391289418e-05, 'epoch': 0.73}) +[2026-06-16 01:40:45,474][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.06827621906995773, 'train/grad_norm_pre_clip': 9.614051818847656, 'timing/data': 0.011412783991545439, 'timing/model': 1.9052434018813074, 'learning_rate/qwen_vl_interface': 7.433545269961906e-06, 'learning_rate/action_model': 7.433545269961906e-05, 'epoch': 0.73}) +[2026-06-16 01:40:51,553][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.05246078968048096, 'train/grad_norm_pre_clip': 11.604928970336914, 'timing/data': 0.011417083907872438, 'timing/model': 1.574315789155662, 'learning_rate/qwen_vl_interface': 7.430840251061411e-06, 'learning_rate/action_model': 7.430840251061411e-05, 'epoch': 0.73}) +[2026-06-16 01:40:57,925][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.05854133889079094, 'train/grad_norm_pre_clip': 8.520245552062988, 'timing/data': 0.011413604021072388, 'timing/model': 1.5626218151301146, 'learning_rate/qwen_vl_interface': 7.428134335699866e-06, 'learning_rate/action_model': 7.428134335699865e-05, 'epoch': 0.73}) +[2026-06-16 01:41:04,244][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.02694990672171116, 'train/grad_norm_pre_clip': 2.398118495941162, 'timing/data': 0.011428563855588436, 'timing/model': 1.552333986852318, 'learning_rate/qwen_vl_interface': 7.425427524989572e-06, 'learning_rate/action_model': 7.425427524989572e-05, 'epoch': 0.73}) +[2026-06-16 01:41:10,460][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.07406186312437057, 'train/grad_norm_pre_clip': 11.03094482421875, 'timing/data': 0.01633753813803196, 'timing/model': 1.575612388085574, 'learning_rate/qwen_vl_interface': 7.422719820043194e-06, 'learning_rate/action_model': 7.422719820043193e-05, 'epoch': 0.73}) +[2026-06-16 01:41:16,743][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.20708750188350677, 'train/grad_norm_pre_clip': 10.508901596069336, 'timing/data': 0.011078912299126387, 'timing/model': 1.558158087078482, 'learning_rate/qwen_vl_interface': 7.420011221973771e-06, 'learning_rate/action_model': 7.420011221973771e-05, 'epoch': 0.73}) +[2026-06-16 01:41:23,076][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.11915110796689987, 'train/grad_norm_pre_clip': 6.601175785064697, 'timing/data': 0.013868796173483133, 'timing/model': 1.774981729220599, 'learning_rate/qwen_vl_interface': 7.417301731894702e-06, 'learning_rate/action_model': 7.417301731894701e-05, 'epoch': 0.73}) +[2026-06-16 01:41:29,185][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.10449951142072678, 'train/grad_norm_pre_clip': 4.66041374206543, 'timing/data': 0.01153117511421442, 'timing/model': 1.5771192680113018, 'learning_rate/qwen_vl_interface': 7.414591350919756e-06, 'learning_rate/action_model': 7.414591350919756e-05, 'epoch': 0.73}) +[2026-06-16 01:41:35,453][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.2153192162513733, 'train/grad_norm_pre_clip': 12.256463050842285, 'timing/data': 0.011406044941395521, 'timing/model': 1.5532410247251391, 'learning_rate/qwen_vl_interface': 7.4118800801630715e-06, 'learning_rate/action_model': 7.411880080163072e-05, 'epoch': 0.73}) +[2026-06-16 01:41:41,755][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.13006767630577087, 'train/grad_norm_pre_clip': 7.005758285522461, 'timing/data': 0.011831476353108883, 'timing/model': 1.5597419370897114, 'learning_rate/qwen_vl_interface': 7.409167920739146e-06, 'learning_rate/action_model': 7.409167920739147e-05, 'epoch': 0.73}) +[2026-06-16 01:41:48,133][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.092213936150074, 'train/grad_norm_pre_clip': 8.096768379211426, 'timing/data': 0.011255354154855013, 'timing/model': 1.608341878745705, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.73}) +[2026-06-16 01:41:54,354][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.19736434519290924, 'train/grad_norm_pre_clip': 10.942767143249512, 'timing/data': 0.011412533931434155, 'timing/model': 1.5471316892653704, 'learning_rate/qwen_vl_interface': 7.403740940349408e-06, 'learning_rate/action_model': 7.403740940349408e-05, 'epoch': 0.73}) +[2026-06-16 01:42:00,778][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.12642988562583923, 'train/grad_norm_pre_clip': 9.04715633392334, 'timing/data': 0.011323594022542238, 'timing/model': 1.8087567957118154, 'learning_rate/qwen_vl_interface': 7.401026121614421e-06, 'learning_rate/action_model': 7.40102612161442e-05, 'epoch': 0.73}) +[2026-06-16 01:42:06,862][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.30675938725471497, 'train/grad_norm_pre_clip': 10.151973724365234, 'timing/data': 0.011556754820048809, 'timing/model': 1.5724745010957122, 'learning_rate/qwen_vl_interface': 7.398310418673846e-06, 'learning_rate/action_model': 7.398310418673846e-05, 'epoch': 0.74}) +[2026-06-16 01:42:13,155][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.10562396794557571, 'train/grad_norm_pre_clip': 9.283713340759277, 'timing/data': 0.011059843003749847, 'timing/model': 1.5647098440676928, 'learning_rate/qwen_vl_interface': 7.395593832644007e-06, 'learning_rate/action_model': 7.395593832644007e-05, 'epoch': 0.74}) +[2026-06-16 01:42:19,550][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.2813103497028351, 'train/grad_norm_pre_clip': 9.212882995605469, 'timing/data': 0.010759251192212105, 'timing/model': 1.55952792102471, 'learning_rate/qwen_vl_interface': 7.39287636464159e-06, 'learning_rate/action_model': 7.39287636464159e-05, 'epoch': 0.74}) +[2026-06-16 01:42:25,847][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.11891523003578186, 'train/grad_norm_pre_clip': 4.945292949676514, 'timing/data': 0.011292433831840754, 'timing/model': 1.560372435953468, 'learning_rate/qwen_vl_interface': 7.3901580157836425e-06, 'learning_rate/action_model': 7.390158015783643e-05, 'epoch': 0.74}) +[2026-06-16 01:42:32,149][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.2545194625854492, 'train/grad_norm_pre_clip': 13.125185012817383, 'timing/data': 0.011342575307935476, 'timing/model': 1.5426328219473362, 'learning_rate/qwen_vl_interface': 7.387438787187576e-06, 'learning_rate/action_model': 7.387438787187576e-05, 'epoch': 0.74}) +[2026-06-16 01:42:38,581][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.06273446977138519, 'train/grad_norm_pre_clip': 3.7371997833251953, 'timing/data': 0.01106985192745924, 'timing/model': 1.8824065928347409, 'learning_rate/qwen_vl_interface': 7.3847186799711625e-06, 'learning_rate/action_model': 7.384718679971162e-05, 'epoch': 0.74}) +[2026-06-16 01:42:44,682][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.17573875188827515, 'train/grad_norm_pre_clip': 13.93652057647705, 'timing/data': 0.01053493982180953, 'timing/model': 1.5887759439647198, 'learning_rate/qwen_vl_interface': 7.381997695252535e-06, 'learning_rate/action_model': 7.381997695252535e-05, 'epoch': 0.74}) +[2026-06-16 01:42:50,979][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.17770852148532867, 'train/grad_norm_pre_clip': 14.589126586914062, 'timing/data': 0.01157906511798501, 'timing/model': 1.5694961729459465, 'learning_rate/qwen_vl_interface': 7.379275834150186e-06, 'learning_rate/action_model': 7.379275834150186e-05, 'epoch': 0.74}) +[2026-06-16 01:42:57,273][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.08158373087644577, 'train/grad_norm_pre_clip': 7.578822612762451, 'timing/data': 0.01185858715325594, 'timing/model': 1.5458950302563608, 'learning_rate/qwen_vl_interface': 7.376553097782972e-06, 'learning_rate/action_model': 7.376553097782972e-05, 'epoch': 0.74}) +[2026-06-16 01:43:03,680][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.0429653637111187, 'train/grad_norm_pre_clip': 2.807927370071411, 'timing/data': 0.01163587486371398, 'timing/model': 1.547355939168483, 'learning_rate/qwen_vl_interface': 7.373829487270109e-06, 'learning_rate/action_model': 7.373829487270109e-05, 'epoch': 0.74}) +[2026-06-16 01:43:09,982][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.27615514397621155, 'train/grad_norm_pre_clip': 15.243000984191895, 'timing/data': 0.011882797349244356, 'timing/model': 1.535980734974146, 'learning_rate/qwen_vl_interface': 7.371105003731165e-06, 'learning_rate/action_model': 7.371105003731164e-05, 'epoch': 0.74}) +[2026-06-16 01:43:16,432][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.08713511377573013, 'train/grad_norm_pre_clip': 7.0478034019470215, 'timing/data': 0.011667085811495781, 'timing/model': 1.8436523429118097, 'learning_rate/qwen_vl_interface': 7.3683796482860755e-06, 'learning_rate/action_model': 7.368379648286075e-05, 'epoch': 0.74}) +[2026-06-16 01:43:22,477][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.10288865119218826, 'train/grad_norm_pre_clip': 7.9580078125, 'timing/data': 0.011488704942166805, 'timing/model': 1.5561197618953884, 'learning_rate/qwen_vl_interface': 7.365653422055131e-06, 'learning_rate/action_model': 7.365653422055131e-05, 'epoch': 0.74}) +[2026-06-16 01:43:28,878][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.041696544736623764, 'train/grad_norm_pre_clip': 2.5054023265838623, 'timing/data': 0.01162418583407998, 'timing/model': 1.559679829981178, 'learning_rate/qwen_vl_interface': 7.3629263261589816e-06, 'learning_rate/action_model': 7.362926326158981e-05, 'epoch': 0.74}) +[2026-06-16 01:43:35,323][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.15015725791454315, 'train/grad_norm_pre_clip': 15.865906715393066, 'timing/data': 0.011986547149717808, 'timing/model': 1.6369251292198896, 'learning_rate/qwen_vl_interface': 7.36019836171863e-06, 'learning_rate/action_model': 7.360198361718629e-05, 'epoch': 0.74}) +[2026-06-16 01:43:41,680][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.13096381723880768, 'train/grad_norm_pre_clip': 9.426962852478027, 'timing/data': 0.011724845971912146, 'timing/model': 1.56059987610206, 'learning_rate/qwen_vl_interface': 7.357469529855441e-06, 'learning_rate/action_model': 7.35746952985544e-05, 'epoch': 0.74}) +[2026-06-16 01:43:48,081][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.19291846454143524, 'train/grad_norm_pre_clip': 10.72102165222168, 'timing/data': 0.0112648936919868, 'timing/model': 1.5492610530927777, 'learning_rate/qwen_vl_interface': 7.354739831691134e-06, 'learning_rate/action_model': 7.354739831691135e-05, 'epoch': 0.74}) +[2026-06-16 01:43:54,443][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.13861890137195587, 'train/grad_norm_pre_clip': 6.4203948974609375, 'timing/data': 0.011366994120180607, 'timing/model': 1.8343522627837956, 'learning_rate/qwen_vl_interface': 7.352009268347786e-06, 'learning_rate/action_model': 7.352009268347786e-05, 'epoch': 0.74}) +[2026-06-16 01:44:00,684][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.024716462939977646, 'train/grad_norm_pre_clip': 2.244198799133301, 'timing/data': 0.011771585792303085, 'timing/model': 1.5686870557256043, 'learning_rate/qwen_vl_interface': 7.349277840947828e-06, 'learning_rate/action_model': 7.349277840947828e-05, 'epoch': 0.74}) +[2026-06-16 01:44:07,097][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.0631655901670456, 'train/grad_norm_pre_clip': 9.617147445678711, 'timing/data': 0.011712046340107918, 'timing/model': 1.5857011079788208, 'learning_rate/qwen_vl_interface': 7.346545550614046e-06, 'learning_rate/action_model': 7.346545550614046e-05, 'epoch': 0.74}) +[2026-06-16 01:44:13,483][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.1922357976436615, 'train/grad_norm_pre_clip': 11.389780044555664, 'timing/data': 0.012593119870871305, 'timing/model': 1.570753957144916, 'learning_rate/qwen_vl_interface': 7.343812398469582e-06, 'learning_rate/action_model': 7.343812398469581e-05, 'epoch': 0.74}) +[2026-06-16 01:44:19,779][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.24395808577537537, 'train/grad_norm_pre_clip': 11.083130836486816, 'timing/data': 0.011941967066377401, 'timing/model': 1.5465938518755138, 'learning_rate/qwen_vl_interface': 7.34107838563793e-06, 'learning_rate/action_model': 7.341078385637929e-05, 'epoch': 0.74}) +[2026-06-16 01:44:26,168][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.18916521966457367, 'train/grad_norm_pre_clip': 14.256253242492676, 'timing/data': 0.011344623751938343, 'timing/model': 1.6079138964414597, 'learning_rate/qwen_vl_interface': 7.338343513242941e-06, 'learning_rate/action_model': 7.338343513242941e-05, 'epoch': 0.74}) +[2026-06-16 01:44:32,543][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.10895093530416489, 'train/grad_norm_pre_clip': 11.697834968566895, 'timing/data': 0.010993773117661476, 'timing/model': 1.8175827781669796, 'learning_rate/qwen_vl_interface': 7.335607782408818e-06, 'learning_rate/action_model': 7.335607782408818e-05, 'epoch': 0.74}) +[2026-06-16 01:44:38,762][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.11700160801410675, 'train/grad_norm_pre_clip': 9.064657211303711, 'timing/data': 0.013148992788046598, 'timing/model': 1.6574020041152835, 'learning_rate/qwen_vl_interface': 7.3328711942601145e-06, 'learning_rate/action_model': 7.332871194260115e-05, 'epoch': 0.74}) +[2026-06-16 01:44:45,167][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.0717262402176857, 'train/grad_norm_pre_clip': 12.191969871520996, 'timing/data': 0.010838261805474758, 'timing/model': 1.614410348702222, 'learning_rate/qwen_vl_interface': 7.33013374992174e-06, 'learning_rate/action_model': 7.33013374992174e-05, 'epoch': 0.75}) +[2026-06-16 01:44:52,114][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.04880175366997719, 'train/grad_norm_pre_clip': 5.816528797149658, 'timing/data': 0.012084607034921646, 'timing/model': 1.572371309157461, 'learning_rate/qwen_vl_interface': 7.327395450518956e-06, 'learning_rate/action_model': 7.327395450518956e-05, 'epoch': 0.75}) +[2026-06-16 01:44:58,587][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.17274688184261322, 'train/grad_norm_pre_clip': 7.607193470001221, 'timing/data': 0.012023167219012976, 'timing/model': 1.5570115260779858, 'learning_rate/qwen_vl_interface': 7.324656297177372e-06, 'learning_rate/action_model': 7.324656297177371e-05, 'epoch': 0.75}) +[2026-06-16 01:45:05,080][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.1238052174448967, 'train/grad_norm_pre_clip': 11.426448822021484, 'timing/data': 0.01180695602670312, 'timing/model': 1.6144757713191211, 'learning_rate/qwen_vl_interface': 7.32191629102295e-06, 'learning_rate/action_model': 7.32191629102295e-05, 'epoch': 0.75}) +[2026-06-16 01:45:11,418][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.010860349982976913, 'train/grad_norm_pre_clip': 1.3917841911315918, 'timing/data': 0.01115705305710435, 'timing/model': 1.8041152781806886, 'learning_rate/qwen_vl_interface': 7.319175433182004e-06, 'learning_rate/action_model': 7.319175433182004e-05, 'epoch': 0.75}) +[2026-06-16 01:45:17,582][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.06041127070784569, 'train/grad_norm_pre_clip': 4.680544376373291, 'timing/data': 0.011307683773338795, 'timing/model': 1.5762612600810826, 'learning_rate/qwen_vl_interface': 7.316433724781198e-06, 'learning_rate/action_model': 7.316433724781197e-05, 'epoch': 0.75}) +[2026-06-16 01:45:23,998][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.0061609684489667416, 'train/grad_norm_pre_clip': 1.5010504722595215, 'timing/data': 0.0118216872215271, 'timing/model': 1.5741072301752865, 'learning_rate/qwen_vl_interface': 7.313691166947543e-06, 'learning_rate/action_model': 7.313691166947543e-05, 'epoch': 0.75}) +[2026-06-16 01:45:30,300][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.2784261703491211, 'train/grad_norm_pre_clip': 16.7500057220459, 'timing/data': 0.011986657045781612, 'timing/model': 1.577630186919123, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.75}) +[2026-06-16 01:45:36,681][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.14288489520549774, 'train/grad_norm_pre_clip': 14.044843673706055, 'timing/data': 0.011459283996373415, 'timing/model': 1.5858959881588817, 'learning_rate/qwen_vl_interface': 7.308203507491485e-06, 'learning_rate/action_model': 7.308203507491484e-05, 'epoch': 0.75}) +[2026-06-16 01:45:42,995][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.19189994037151337, 'train/grad_norm_pre_clip': 14.656990051269531, 'timing/data': 0.01084029208868742, 'timing/model': 1.5480337869375944, 'learning_rate/qwen_vl_interface': 7.3054584081248515e-06, 'learning_rate/action_model': 7.305458408124852e-05, 'epoch': 0.75}) +[2026-06-16 01:45:49,332][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.15488046407699585, 'train/grad_norm_pre_clip': 22.30915069580078, 'timing/data': 0.011690245941281319, 'timing/model': 1.814222919754684, 'learning_rate/qwen_vl_interface': 7.302712463836907e-06, 'learning_rate/action_model': 7.302712463836907e-05, 'epoch': 0.75}) +[2026-06-16 01:45:55,621][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.08770723640918732, 'train/grad_norm_pre_clip': 11.74071979522705, 'timing/data': 0.010950813069939613, 'timing/model': 1.8064072029665112, 'learning_rate/qwen_vl_interface': 7.2999656757564086e-06, 'learning_rate/action_model': 7.299965675756409e-05, 'epoch': 0.75}) +[2026-06-16 01:46:01,688][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.07901886105537415, 'train/grad_norm_pre_clip': 7.056692600250244, 'timing/data': 0.011477685067802668, 'timing/model': 1.5724677848629653, 'learning_rate/qwen_vl_interface': 7.297218045012453e-06, 'learning_rate/action_model': 7.297218045012453e-05, 'epoch': 0.75}) +[2026-06-16 01:46:07,989][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.026368122547864914, 'train/grad_norm_pre_clip': 2.4675405025482178, 'timing/data': 0.011803175788372755, 'timing/model': 1.5680837552063167, 'learning_rate/qwen_vl_interface': 7.29446957273449e-06, 'learning_rate/action_model': 7.294469572734489e-05, 'epoch': 0.75}) +[2026-06-16 01:46:14,482][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.2226494550704956, 'train/grad_norm_pre_clip': 8.35659122467041, 'timing/data': 0.012037968263030052, 'timing/model': 1.5811806782148778, 'learning_rate/qwen_vl_interface': 7.2917202600523104e-06, 'learning_rate/action_model': 7.29172026005231e-05, 'epoch': 0.75}) +[2026-06-16 01:46:20,881][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.3347289562225342, 'train/grad_norm_pre_clip': 10.462636947631836, 'timing/data': 0.011521025095134974, 'timing/model': 1.6049538017250597, 'learning_rate/qwen_vl_interface': 7.2889701080960565e-06, 'learning_rate/action_model': 7.288970108096056e-05, 'epoch': 0.75}) +[2026-06-16 01:46:27,225][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.01898183859884739, 'train/grad_norm_pre_clip': 4.828416347503662, 'timing/data': 0.010820151306688786, 'timing/model': 1.8079759441316128, 'learning_rate/qwen_vl_interface': 7.286219117996209e-06, 'learning_rate/action_model': 7.286219117996207e-05, 'epoch': 0.75}) +[2026-06-16 01:46:33,974][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.12452083826065063, 'train/grad_norm_pre_clip': 9.217819213867188, 'timing/data': 0.01075601065531373, 'timing/model': 2.2610213197767735, 'learning_rate/qwen_vl_interface': 7.283467290883598e-06, 'learning_rate/action_model': 7.283467290883598e-05, 'epoch': 0.75}) +[2026-06-16 01:46:40,069][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.1035829484462738, 'train/grad_norm_pre_clip': 10.848791122436523, 'timing/data': 0.01213418785482645, 'timing/model': 1.56309105316177, 'learning_rate/qwen_vl_interface': 7.2807146278893925e-06, 'learning_rate/action_model': 7.280714627889393e-05, 'epoch': 0.75}) +[2026-06-16 01:46:46,357][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.3654991090297699, 'train/grad_norm_pre_clip': 14.648277282714844, 'timing/data': 0.012259257957339287, 'timing/model': 1.5396354822441936, 'learning_rate/qwen_vl_interface': 7.277961130145111e-06, 'learning_rate/action_model': 7.277961130145111e-05, 'epoch': 0.75}) +[2026-06-16 01:46:52,758][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.11518211662769318, 'train/grad_norm_pre_clip': 18.866281509399414, 'timing/data': 0.011818645987659693, 'timing/model': 1.5416667917743325, 'learning_rate/qwen_vl_interface': 7.2752067987826116e-06, 'learning_rate/action_model': 7.275206798782611e-05, 'epoch': 0.75}) +[2026-06-16 01:46:59,162][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.013075042515993118, 'train/grad_norm_pre_clip': 1.4835684299468994, 'timing/data': 0.011681615840643644, 'timing/model': 1.5191344250924885, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.75}) +[2026-06-16 01:47:05,602][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.032007262110710144, 'train/grad_norm_pre_clip': 3.8452141284942627, 'timing/data': 0.011300474870949984, 'timing/model': 1.7914099171757698, 'learning_rate/qwen_vl_interface': 7.269695639732108e-06, 'learning_rate/action_model': 7.269695639732108e-05, 'epoch': 0.75}) +[2026-06-16 01:47:11,674][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.011745885014533997, 'train/grad_norm_pre_clip': 1.580597996711731, 'timing/data': 0.011062412988394499, 'timing/model': 1.565527378115803, 'learning_rate/qwen_vl_interface': 7.266938814309533e-06, 'learning_rate/action_model': 7.266938814309533e-05, 'epoch': 0.75}) +[2026-06-16 01:47:18,082][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.07365471869707108, 'train/grad_norm_pre_clip': 5.841622352600098, 'timing/data': 0.014554200228303671, 'timing/model': 1.5643113041296601, 'learning_rate/qwen_vl_interface': 7.264181159799598e-06, 'learning_rate/action_model': 7.264181159799597e-05, 'epoch': 0.75}) +[2026-06-16 01:47:24,470][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.06454633921384811, 'train/grad_norm_pre_clip': 5.177181243896484, 'timing/data': 0.01089024217799306, 'timing/model': 1.5521401949226856, 'learning_rate/qwen_vl_interface': 7.261422677335871e-06, 'learning_rate/action_model': 7.261422677335871e-05, 'epoch': 0.76}) +[2026-06-16 01:47:30,857][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.29315415024757385, 'train/grad_norm_pre_clip': 16.647912979125977, 'timing/data': 0.01138017512857914, 'timing/model': 1.5871749129146338, 'learning_rate/qwen_vl_interface': 7.258663368052259e-06, 'learning_rate/action_model': 7.258663368052259e-05, 'epoch': 0.76}) +[2026-06-16 01:47:37,162][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.02182784304022789, 'train/grad_norm_pre_clip': 4.491745948791504, 'timing/data': 0.010778351221233606, 'timing/model': 1.6040436141192913, 'learning_rate/qwen_vl_interface': 7.2559032330830105e-06, 'learning_rate/action_model': 7.25590323308301e-05, 'epoch': 0.76}) +[2026-06-16 01:47:43,455][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.006357192527502775, 'train/grad_norm_pre_clip': 3.0972743034362793, 'timing/data': 0.011394895147532225, 'timing/model': 1.7528509958647192, 'learning_rate/qwen_vl_interface': 7.253142273562712e-06, 'learning_rate/action_model': 7.253142273562712e-05, 'epoch': 0.76}) +[2026-06-16 01:47:49,568][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.18716193735599518, 'train/grad_norm_pre_clip': 10.735652923583984, 'timing/data': 0.010658331215381622, 'timing/model': 1.5684397439472377, 'learning_rate/qwen_vl_interface': 7.25038049062629e-06, 'learning_rate/action_model': 7.250380490626289e-05, 'epoch': 0.76}) +[2026-06-16 01:47:55,869][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.025634489953517914, 'train/grad_norm_pre_clip': 3.360051393508911, 'timing/data': 0.011491245124489069, 'timing/model': 1.5612745797261596, 'learning_rate/qwen_vl_interface': 7.24761788540901e-06, 'learning_rate/action_model': 7.24761788540901e-05, 'epoch': 0.76}) +[2026-06-16 01:48:02,165][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.4059964120388031, 'train/grad_norm_pre_clip': 13.060643196105957, 'timing/data': 0.010739061050117016, 'timing/model': 1.564217154867947, 'learning_rate/qwen_vl_interface': 7.244854459046473e-06, 'learning_rate/action_model': 7.244854459046473e-05, 'epoch': 0.76}) +[2026-06-16 01:48:08,471][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.14642807841300964, 'train/grad_norm_pre_clip': 11.209096908569336, 'timing/data': 0.010837141890078783, 'timing/model': 1.5332853067666292, 'learning_rate/qwen_vl_interface': 7.24209021267462e-06, 'learning_rate/action_model': 7.24209021267462e-05, 'epoch': 0.76}) +[2026-06-16 01:48:14,852][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.162580668926239, 'train/grad_norm_pre_clip': 11.983718872070312, 'timing/data': 0.011537745129317045, 'timing/model': 1.5997615563683212, 'learning_rate/qwen_vl_interface': 7.23932514742973e-06, 'learning_rate/action_model': 7.23932514742973e-05, 'epoch': 0.76}) +[2026-06-16 01:48:21,249][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.06214502826333046, 'train/grad_norm_pre_clip': 5.922023773193359, 'timing/data': 0.011184663977473974, 'timing/model': 1.763446168974042, 'learning_rate/qwen_vl_interface': 7.236559264448413e-06, 'learning_rate/action_model': 7.236559264448412e-05, 'epoch': 0.76}) +[2026-06-16 01:48:27,359][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.06986827403306961, 'train/grad_norm_pre_clip': 9.811356544494629, 'timing/data': 0.01084127090871334, 'timing/model': 1.5554952351376414, 'learning_rate/qwen_vl_interface': 7.233792564867623e-06, 'learning_rate/action_model': 7.233792564867623e-05, 'epoch': 0.76}) +[2026-06-16 01:48:33,667][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.010458078235387802, 'train/grad_norm_pre_clip': 0.9996457099914551, 'timing/data': 0.010979763232171535, 'timing/model': 1.559610694181174, 'learning_rate/qwen_vl_interface': 7.231025049824644e-06, 'learning_rate/action_model': 7.231025049824643e-05, 'epoch': 0.76}) +[2026-06-16 01:48:39,958][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.4160900413990021, 'train/grad_norm_pre_clip': 22.792686462402344, 'timing/data': 0.010828970931470394, 'timing/model': 1.5608655619435012, 'learning_rate/qwen_vl_interface': 7.228256720457097e-06, 'learning_rate/action_model': 7.228256720457096e-05, 'epoch': 0.76}) +[2026-06-16 01:48:46,261][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.24898794293403625, 'train/grad_norm_pre_clip': 22.283416748046875, 'timing/data': 0.01112636411562562, 'timing/model': 1.5300412839278579, 'learning_rate/qwen_vl_interface': 7.225487577902937e-06, 'learning_rate/action_model': 7.225487577902936e-05, 'epoch': 0.76}) +[2026-06-16 01:48:52,593][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.07041028141975403, 'train/grad_norm_pre_clip': 4.168012619018555, 'timing/data': 0.0111431828700006, 'timing/model': 1.7866493230685592, 'learning_rate/qwen_vl_interface': 7.222717623300455e-06, 'learning_rate/action_model': 7.222717623300455e-05, 'epoch': 0.76}) +[2026-06-16 01:48:58,662][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.4791907072067261, 'train/grad_norm_pre_clip': 17.617368698120117, 'timing/data': 0.011777325998991728, 'timing/model': 1.5626585609279573, 'learning_rate/qwen_vl_interface': 7.219946857788273e-06, 'learning_rate/action_model': 7.219946857788272e-05, 'epoch': 0.76}) +[2026-06-16 01:49:04,961][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.042519960552453995, 'train/grad_norm_pre_clip': 7.001577854156494, 'timing/data': 0.011637066025286913, 'timing/model': 1.5625651609152555, 'learning_rate/qwen_vl_interface': 7.21717528250535e-06, 'learning_rate/action_model': 7.21717528250535e-05, 'epoch': 0.76}) +[2026-06-16 01:49:11,159][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.3088909089565277, 'train/grad_norm_pre_clip': 6.774604797363281, 'timing/data': 0.011720926035195589, 'timing/model': 1.5586729827336967, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 0.76}) +[2026-06-16 01:49:17,464][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.30743327736854553, 'train/grad_norm_pre_clip': 20.905513763427734, 'timing/data': 0.011471064761281013, 'timing/model': 1.5672569437883794, 'learning_rate/qwen_vl_interface': 7.211629707184768e-06, 'learning_rate/action_model': 7.211629707184768e-05, 'epoch': 0.76}) +[2026-06-16 01:49:23,767][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.15870961546897888, 'train/grad_norm_pre_clip': 7.439631938934326, 'timing/data': 0.012870841659605503, 'timing/model': 1.5607701330445707, 'learning_rate/qwen_vl_interface': 7.2088557094266885e-06, 'learning_rate/action_model': 7.208855709426688e-05, 'epoch': 0.76}) +[2026-06-16 01:49:30,105][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.0620577298104763, 'train/grad_norm_pre_clip': 4.761075973510742, 'timing/data': 0.011321783997118473, 'timing/model': 1.8028715131804347, 'learning_rate/qwen_vl_interface': 7.2060809064570195e-06, 'learning_rate/action_model': 7.20608090645702e-05, 'epoch': 0.76}) +[2026-06-16 01:49:36,165][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.11635248363018036, 'train/grad_norm_pre_clip': 9.45042896270752, 'timing/data': 0.013473404105752707, 'timing/model': 1.560215131379664, 'learning_rate/qwen_vl_interface': 7.203305299416375e-06, 'learning_rate/action_model': 7.203305299416375e-05, 'epoch': 0.76}) +[2026-06-16 01:49:42,460][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.061616089195013046, 'train/grad_norm_pre_clip': 5.934929370880127, 'timing/data': 0.011875587049871683, 'timing/model': 1.5590358758345246, 'learning_rate/qwen_vl_interface': 7.200528889445707e-06, 'learning_rate/action_model': 7.200528889445706e-05, 'epoch': 0.76}) +[2026-06-16 01:49:48,787][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.1460304707288742, 'train/grad_norm_pre_clip': 10.361397743225098, 'timing/data': 0.010896832216531038, 'timing/model': 1.590531556867063, 'learning_rate/qwen_vl_interface': 7.19775167768629e-06, 'learning_rate/action_model': 7.19775167768629e-05, 'epoch': 0.76}) +[2026-06-16 01:49:55,134][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.22650142014026642, 'train/grad_norm_pre_clip': 12.713663101196289, 'timing/data': 0.01082921214401722, 'timing/model': 1.6356207691133022, 'learning_rate/qwen_vl_interface': 7.194973665279731e-06, 'learning_rate/action_model': 7.19497366527973e-05, 'epoch': 0.77}) +[2026-06-16 01:50:01,365][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.07440115511417389, 'train/grad_norm_pre_clip': 6.336853981018066, 'timing/data': 0.010184620972722769, 'timing/model': 1.5560477399267256, 'learning_rate/qwen_vl_interface': 7.1921948533679666e-06, 'learning_rate/action_model': 7.192194853367967e-05, 'epoch': 0.77}) +[2026-06-16 01:50:07,741][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.08744195848703384, 'train/grad_norm_pre_clip': 6.912188529968262, 'timing/data': 0.011053283233195543, 'timing/model': 1.8434450002387166, 'learning_rate/qwen_vl_interface': 7.1894152430932605e-06, 'learning_rate/action_model': 7.189415243093261e-05, 'epoch': 0.77}) +[2026-06-16 01:50:13,869][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.09494224935770035, 'train/grad_norm_pre_clip': 6.52371883392334, 'timing/data': 0.011339695192873478, 'timing/model': 1.5712610012851655, 'learning_rate/qwen_vl_interface': 7.186634835598205e-06, 'learning_rate/action_model': 7.186634835598205e-05, 'epoch': 0.77}) +[2026-06-16 01:50:20,160][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.09687361866235733, 'train/grad_norm_pre_clip': 5.700076580047607, 'timing/data': 0.011555205099284649, 'timing/model': 1.5610340228304267, 'learning_rate/qwen_vl_interface': 7.183853632025724e-06, 'learning_rate/action_model': 7.183853632025723e-05, 'epoch': 0.77}) +[2026-06-16 01:50:26,457][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.0779929831624031, 'train/grad_norm_pre_clip': 4.972272872924805, 'timing/data': 0.01156357629224658, 'timing/model': 1.56038155592978, 'learning_rate/qwen_vl_interface': 7.181071633519062e-06, 'learning_rate/action_model': 7.181071633519062e-05, 'epoch': 0.77}) +[2026-06-16 01:50:32,805][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.17838633060455322, 'train/grad_norm_pre_clip': 13.491150856018066, 'timing/data': 0.010864493902772665, 'timing/model': 1.556795062031597, 'learning_rate/qwen_vl_interface': 7.178288841221795e-06, 'learning_rate/action_model': 7.178288841221794e-05, 'epoch': 0.77}) +[2026-06-16 01:50:39,062][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.06486324965953827, 'train/grad_norm_pre_clip': 3.27587628364563, 'timing/data': 0.010480054188519716, 'timing/model': 1.5446054060012102, 'learning_rate/qwen_vl_interface': 7.1755052562778215e-06, 'learning_rate/action_model': 7.175505256277822e-05, 'epoch': 0.77}) +[2026-06-16 02:01:38,618][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.08549460768699646, 'train/grad_norm_pre_clip': 6.393196105957031, 'eval/loss': 0.07363995164632797, 'eval/flap_precision': 0.8848039215686274, 'eval/flap_recall': 0.984993178717599, 'eval/flap_f1': 0.9322143318269852, 'eval/flap_rate_model': 0.22666666666666666, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9322143318269852, 'timing/data': 0.01178697682917118, 'timing/model': 1.8022205331362784, 'learning_rate/qwen_vl_interface': 7.1727208798313704e-06, 'learning_rate/action_model': 7.172720879831371e-05, 'epoch': 0.77}) +[2026-06-16 02:01:38,619][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1900_state +[2026-06-16 02:01:38,625][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:01:47,496][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1900_state/model.safetensors +[2026-06-16 02:02:07,490][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1900_state/optimizer.bin +[2026-06-16 02:02:07,491][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1900_state/sampler.bin +[2026-06-16 02:02:07,491][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1900_state/sampler_1.bin +[2026-06-16 02:02:07,492][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_1900_state/random_states_0.pkl +[2026-06-16 02:02:07,678][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1900 +[2026-06-16 02:02:07,679][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:02:07,694][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:02:13,845][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.06862325966358185, 'train/grad_norm_pre_clip': 4.194349765777588, 'timing/data': 0.011614036280661821, 'timing/model': 1.7847959231585264, 'learning_rate/qwen_vl_interface': 7.169935713026994e-06, 'learning_rate/action_model': 7.169935713026994e-05, 'epoch': 0.77}) +[2026-06-16 02:02:19,863][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.09771686792373657, 'train/grad_norm_pre_clip': 13.222023963928223, 'timing/data': 0.010335159953683615, 'timing/model': 1.6256734090857208, 'learning_rate/qwen_vl_interface': 7.1671497570095684e-06, 'learning_rate/action_model': 7.167149757009568e-05, 'epoch': 0.77}) +[2026-06-16 02:02:25,922][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.10382159054279327, 'train/grad_norm_pre_clip': 6.880292892456055, 'timing/data': 0.010323050897568464, 'timing/model': 1.556401016190648, 'learning_rate/qwen_vl_interface': 7.164363012924295e-06, 'learning_rate/action_model': 7.164363012924294e-05, 'epoch': 0.77}) +[2026-06-16 02:02:32,030][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.27311059832572937, 'train/grad_norm_pre_clip': 12.920354843139648, 'timing/data': 0.009907179046422243, 'timing/model': 1.5588683555833995, 'learning_rate/qwen_vl_interface': 7.161575481916699e-06, 'learning_rate/action_model': 7.1615754819167e-05, 'epoch': 0.77}) +[2026-06-16 02:02:38,237][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.20505228638648987, 'train/grad_norm_pre_clip': 3.8533976078033447, 'timing/data': 0.011495576240122318, 'timing/model': 1.5709746340289712, 'learning_rate/qwen_vl_interface': 7.158787165132627e-06, 'learning_rate/action_model': 7.158787165132626e-05, 'epoch': 0.77}) +[2026-06-16 02:02:44,449][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.06692327558994293, 'train/grad_norm_pre_clip': 6.972192764282227, 'timing/data': 0.011030164081603289, 'timing/model': 1.5394934900105, 'learning_rate/qwen_vl_interface': 7.1559980637182535e-06, 'learning_rate/action_model': 7.155998063718254e-05, 'epoch': 0.77}) +[2026-06-16 02:02:50,773][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.1465727984905243, 'train/grad_norm_pre_clip': 10.554441452026367, 'timing/data': 0.011752427089959383, 'timing/model': 1.7997869527898729, 'learning_rate/qwen_vl_interface': 7.153208178820071e-06, 'learning_rate/action_model': 7.153208178820072e-05, 'epoch': 0.77}) +[2026-06-16 02:02:56,856][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.06447634845972061, 'train/grad_norm_pre_clip': 9.905539512634277, 'timing/data': 0.0117943468503654, 'timing/model': 1.5612052357755601, 'learning_rate/qwen_vl_interface': 7.150417511584898e-06, 'learning_rate/action_model': 7.150417511584898e-05, 'epoch': 0.77}) +[2026-06-16 02:03:03,152][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.2874835133552551, 'train/grad_norm_pre_clip': 14.374481201171875, 'timing/data': 0.011289224959909916, 'timing/model': 1.5667128777131438, 'learning_rate/qwen_vl_interface': 7.147626063159872e-06, 'learning_rate/action_model': 7.147626063159872e-05, 'epoch': 0.77}) +[2026-06-16 02:03:09,448][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.17936822772026062, 'train/grad_norm_pre_clip': 13.337091445922852, 'timing/data': 0.011446594726294279, 'timing/model': 1.5649791420437396, 'learning_rate/qwen_vl_interface': 7.14483383469245e-06, 'learning_rate/action_model': 7.144833834692449e-05, 'epoch': 0.77}) +[2026-06-16 02:03:15,748][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.11954601854085922, 'train/grad_norm_pre_clip': 11.263396263122559, 'timing/data': 0.011345565784722567, 'timing/model': 1.51030845195055, 'learning_rate/qwen_vl_interface': 7.142040827330411e-06, 'learning_rate/action_model': 7.142040827330411e-05, 'epoch': 0.77}) +[2026-06-16 02:03:22,043][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.03373337164521217, 'train/grad_norm_pre_clip': 4.507828235626221, 'timing/data': 0.010788143146783113, 'timing/model': 1.5168771897442639, 'learning_rate/qwen_vl_interface': 7.139247042221859e-06, 'learning_rate/action_model': 7.139247042221858e-05, 'epoch': 0.77}) +[2026-06-16 02:03:28,483][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.11785029619932175, 'train/grad_norm_pre_clip': 9.586627006530762, 'timing/data': 0.01098529389128089, 'timing/model': 1.8139733490534127, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.77}) +[2026-06-16 02:03:34,619][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.11247240006923676, 'train/grad_norm_pre_clip': 9.692049026489258, 'timing/data': 0.011122254189103842, 'timing/model': 1.6351573620922863, 'learning_rate/qwen_vl_interface': 7.133657143359203e-06, 'learning_rate/action_model': 7.133657143359202e-05, 'epoch': 0.77}) +[2026-06-16 02:03:40,855][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.09318003058433533, 'train/grad_norm_pre_clip': 7.587132453918457, 'timing/data': 0.011698626913130283, 'timing/model': 1.5682236021384597, 'learning_rate/qwen_vl_interface': 7.130861031902897e-06, 'learning_rate/action_model': 7.130861031902897e-05, 'epoch': 0.77}) +[2026-06-16 02:03:47,287][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.04500003531575203, 'train/grad_norm_pre_clip': 9.55142879486084, 'timing/data': 0.01566688297316432, 'timing/model': 1.5581812006421387, 'learning_rate/qwen_vl_interface': 7.128064147295667e-06, 'learning_rate/action_model': 7.128064147295667e-05, 'epoch': 0.77}) +[2026-06-16 02:03:54,052][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.045132286846637726, 'train/grad_norm_pre_clip': 6.958345890045166, 'timing/data': 0.010735314339399338, 'timing/model': 1.5513434722088277, 'learning_rate/qwen_vl_interface': 7.125266490687206e-06, 'learning_rate/action_model': 7.125266490687206e-05, 'epoch': 0.78}) +[2026-06-16 02:04:00,465][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.047959379851818085, 'train/grad_norm_pre_clip': 6.649383544921875, 'timing/data': 0.011416975874453783, 'timing/model': 1.5976982391439378, 'learning_rate/qwen_vl_interface': 7.1224680632275265e-06, 'learning_rate/action_model': 7.122468063227527e-05, 'epoch': 0.78}) +[2026-06-16 02:04:06,794][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.035177383571863174, 'train/grad_norm_pre_clip': 2.084303855895996, 'timing/data': 0.010840342845767736, 'timing/model': 1.7989708720706403, 'learning_rate/qwen_vl_interface': 7.1196688660669555e-06, 'learning_rate/action_model': 7.119668866066955e-05, 'epoch': 0.78}) +[2026-06-16 02:04:13,022][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.08805893361568451, 'train/grad_norm_pre_clip': 9.888042449951172, 'timing/data': 0.011344154831022024, 'timing/model': 1.7252630735747516, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.78}) +[2026-06-16 02:04:19,059][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.024250904098153114, 'train/grad_norm_pre_clip': 2.275057792663574, 'timing/data': 0.012612641789019108, 'timing/model': 1.5583021049387753, 'learning_rate/qwen_vl_interface': 7.114068167246033e-06, 'learning_rate/action_model': 7.114068167246033e-05, 'epoch': 0.78}) +[2026-06-16 02:04:25,350][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.040585100650787354, 'train/grad_norm_pre_clip': 8.518111228942871, 'timing/data': 0.010988124180585146, 'timing/model': 1.5632850648835301, 'learning_rate/qwen_vl_interface': 7.111266667887916e-06, 'learning_rate/action_model': 7.111266667887915e-05, 'epoch': 0.78}) +[2026-06-16 02:04:31,654][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.09930995106697083, 'train/grad_norm_pre_clip': 7.854583740234375, 'timing/data': 0.010761782992631197, 'timing/model': 1.5592398368753493, 'learning_rate/qwen_vl_interface': 7.108464403433378e-06, 'learning_rate/action_model': 7.108464403433378e-05, 'epoch': 0.78}) +[2026-06-16 02:04:38,041][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.28962820768356323, 'train/grad_norm_pre_clip': 12.797529220581055, 'timing/data': 0.011218845844268799, 'timing/model': 1.5804875809699297, 'learning_rate/qwen_vl_interface': 7.105661375034324e-06, 'learning_rate/action_model': 7.105661375034324e-05, 'epoch': 0.78}) +[2026-06-16 02:04:44,399][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.3269180357456207, 'train/grad_norm_pre_clip': 22.918344497680664, 'timing/data': 0.011557708028703928, 'timing/model': 1.7961544981226325, 'learning_rate/qwen_vl_interface': 7.102857583842973e-06, 'learning_rate/action_model': 7.102857583842974e-05, 'epoch': 0.78}) +[2026-06-16 02:04:50,689][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.22059662640094757, 'train/grad_norm_pre_clip': 17.2315616607666, 'timing/data': 0.011775649152696133, 'timing/model': 1.7903968542814255, 'learning_rate/qwen_vl_interface': 7.100053031011859e-06, 'learning_rate/action_model': 7.100053031011859e-05, 'epoch': 0.78}) +[2026-06-16 02:04:56,755][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.12278351187705994, 'train/grad_norm_pre_clip': 9.272215843200684, 'timing/data': 0.010827925056219101, 'timing/model': 1.559224628843367, 'learning_rate/qwen_vl_interface': 7.097247717693827e-06, 'learning_rate/action_model': 7.097247717693827e-05, 'epoch': 0.78}) +[2026-06-16 02:05:03,058][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.016048023477196693, 'train/grad_norm_pre_clip': 2.5016396045684814, 'timing/data': 0.011656908318400383, 'timing/model': 1.5587285049259663, 'learning_rate/qwen_vl_interface': 7.094441645042033e-06, 'learning_rate/action_model': 7.094441645042032e-05, 'epoch': 0.78}) +[2026-06-16 02:05:09,456][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.20787540078163147, 'train/grad_norm_pre_clip': 17.565397262573242, 'timing/data': 0.011273387353867292, 'timing/model': 1.556448323186487, 'learning_rate/qwen_vl_interface': 7.09163481420995e-06, 'learning_rate/action_model': 7.09163481420995e-05, 'epoch': 0.78}) +[2026-06-16 02:05:16,044][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.44568121433258057, 'train/grad_norm_pre_clip': 21.19086456298828, 'timing/data': 0.010895384009927511, 'timing/model': 1.673005229793489, 'learning_rate/qwen_vl_interface': 7.0888272263513595e-06, 'learning_rate/action_model': 7.088827226351359e-05, 'epoch': 0.78}) +[2026-06-16 02:05:22,504][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.17032554745674133, 'train/grad_norm_pre_clip': 11.908771514892578, 'timing/data': 0.011290875729173422, 'timing/model': 1.8278638538904488, 'learning_rate/qwen_vl_interface': 7.086018882620354e-06, 'learning_rate/action_model': 7.086018882620354e-05, 'epoch': 0.78}) +[2026-06-16 02:05:28,673][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.009389630518853664, 'train/grad_norm_pre_clip': 1.1849863529205322, 'timing/data': 0.011643728241324425, 'timing/model': 1.571380347944796, 'learning_rate/qwen_vl_interface': 7.083209784171341e-06, 'learning_rate/action_model': 7.083209784171341e-05, 'epoch': 0.78}) +[2026-06-16 02:05:34,969][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.12507909536361694, 'train/grad_norm_pre_clip': 15.042642593383789, 'timing/data': 0.01133138732984662, 'timing/model': 1.563088591210544, 'learning_rate/qwen_vl_interface': 7.080399932159029e-06, 'learning_rate/action_model': 7.080399932159029e-05, 'epoch': 0.78}) +[2026-06-16 02:05:41,369][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.004726686980575323, 'train/grad_norm_pre_clip': 0.5549342036247253, 'timing/data': 0.011296056676656008, 'timing/model': 1.5567815629765391, 'learning_rate/qwen_vl_interface': 7.077589327738448e-06, 'learning_rate/action_model': 7.077589327738447e-05, 'epoch': 0.78}) +[2026-06-16 02:05:47,756][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.2635917365550995, 'train/grad_norm_pre_clip': 13.669438362121582, 'timing/data': 0.011641837656497955, 'timing/model': 1.5963724530301988, 'learning_rate/qwen_vl_interface': 7.0747779720649265e-06, 'learning_rate/action_model': 7.074777972064927e-05, 'epoch': 0.78}) +[2026-06-16 02:05:54,064][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.2637863755226135, 'train/grad_norm_pre_clip': 17.301856994628906, 'timing/data': 0.011334887240082026, 'timing/model': 1.5217353231273592, 'learning_rate/qwen_vl_interface': 7.071965866294108e-06, 'learning_rate/action_model': 7.071965866294108e-05, 'epoch': 0.78}) +[2026-06-16 02:06:00,358][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.03768619894981384, 'train/grad_norm_pre_clip': 24.774988174438477, 'timing/data': 0.011128266341984272, 'timing/model': 1.7593638016842306, 'learning_rate/qwen_vl_interface': 7.0691530115819454e-06, 'learning_rate/action_model': 7.069153011581945e-05, 'epoch': 0.78}) +[2026-06-16 02:06:06,467][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.1740187406539917, 'train/grad_norm_pre_clip': 10.765613555908203, 'timing/data': 0.011430116835981607, 'timing/model': 1.5674468949437141, 'learning_rate/qwen_vl_interface': 7.066339409084694e-06, 'learning_rate/action_model': 7.066339409084694e-05, 'epoch': 0.78}) +[2026-06-16 02:06:12,860][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.0623258575797081, 'train/grad_norm_pre_clip': 14.449372291564941, 'timing/data': 0.011700968723744154, 'timing/model': 1.5603615823201835, 'learning_rate/qwen_vl_interface': 7.0635250599589215e-06, 'learning_rate/action_model': 7.06352505995892e-05, 'epoch': 0.78}) +[2026-06-16 02:06:19,168][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.1326548010110855, 'train/grad_norm_pre_clip': 18.712759017944336, 'timing/data': 0.012398841790854931, 'timing/model': 1.5574951684102416, 'learning_rate/qwen_vl_interface': 7.060709965361499e-06, 'learning_rate/action_model': 7.060709965361499e-05, 'epoch': 0.78}) +[2026-06-16 02:06:25,551][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.029433609917759895, 'train/grad_norm_pre_clip': 2.8082892894744873, 'timing/data': 0.011052554938942194, 'timing/model': 1.589670148678124, 'learning_rate/qwen_vl_interface': 7.057894126449606e-06, 'learning_rate/action_model': 7.057894126449606e-05, 'epoch': 0.78}) +[2026-06-16 02:06:31,945][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.15075822174549103, 'train/grad_norm_pre_clip': 13.513043403625488, 'timing/data': 0.010987516026943922, 'timing/model': 1.5868878928013146, 'learning_rate/qwen_vl_interface': 7.055077544380728e-06, 'learning_rate/action_model': 7.055077544380728e-05, 'epoch': 0.79}) +[2026-06-16 02:06:38,262][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.32366546988487244, 'train/grad_norm_pre_clip': 10.501758575439453, 'timing/data': 0.011065965984016657, 'timing/model': 1.7645810199901462, 'learning_rate/qwen_vl_interface': 7.0522602203126554e-06, 'learning_rate/action_model': 7.052260220312656e-05, 'epoch': 0.79}) +[2026-06-16 02:06:44,360][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.09540668874979019, 'train/grad_norm_pre_clip': 5.952816009521484, 'timing/data': 0.011056995950639248, 'timing/model': 1.55681117111817, 'learning_rate/qwen_vl_interface': 7.0494421554034845e-06, 'learning_rate/action_model': 7.049442155403484e-05, 'epoch': 0.79}) +[2026-06-16 02:06:50,661][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.13139474391937256, 'train/grad_norm_pre_clip': 15.146778106689453, 'timing/data': 0.011170405894517899, 'timing/model': 1.567041290923953, 'learning_rate/qwen_vl_interface': 7.046623350811612e-06, 'learning_rate/action_model': 7.046623350811612e-05, 'epoch': 0.79}) +[2026-06-16 02:06:56,961][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.2560781240463257, 'train/grad_norm_pre_clip': 18.460973739624023, 'timing/data': 0.011106956284493208, 'timing/model': 1.5617091236636043, 'learning_rate/qwen_vl_interface': 7.043803807695747e-06, 'learning_rate/action_model': 7.043803807695747e-05, 'epoch': 0.79}) +[2026-06-16 02:07:03,267][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.03943750634789467, 'train/grad_norm_pre_clip': 8.64147663116455, 'timing/data': 0.012964573688805103, 'timing/model': 1.5216233041137457, 'learning_rate/qwen_vl_interface': 7.040983527214893e-06, 'learning_rate/action_model': 7.040983527214893e-05, 'epoch': 0.79}) +[2026-06-16 02:07:09,638][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.05003533884882927, 'train/grad_norm_pre_clip': 6.685102462768555, 'timing/data': 0.010541954077780247, 'timing/model': 1.8348905988968909, 'learning_rate/qwen_vl_interface': 7.038162510528363e-06, 'learning_rate/action_model': 7.038162510528362e-05, 'epoch': 0.79}) +[2026-06-16 02:07:15,754][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.018592091277241707, 'train/grad_norm_pre_clip': 3.3044652938842773, 'timing/data': 0.010949076153337955, 'timing/model': 1.562897517811507, 'learning_rate/qwen_vl_interface': 7.03534075879577e-06, 'learning_rate/action_model': 7.03534075879577e-05, 'epoch': 0.79}) +[2026-06-16 02:07:22,057][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.09000926464796066, 'train/grad_norm_pre_clip': 12.649831771850586, 'timing/data': 0.012304010801017284, 'timing/model': 1.554884250741452, 'learning_rate/qwen_vl_interface': 7.032518273177028e-06, 'learning_rate/action_model': 7.032518273177028e-05, 'epoch': 0.79}) +[2026-06-16 02:07:28,375][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.1856498420238495, 'train/grad_norm_pre_clip': 13.830214500427246, 'timing/data': 0.010832564905285835, 'timing/model': 1.563140971120447, 'learning_rate/qwen_vl_interface': 7.029695054832357e-06, 'learning_rate/action_model': 7.029695054832357e-05, 'epoch': 0.79}) +[2026-06-16 02:07:34,671][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.1156626045703888, 'train/grad_norm_pre_clip': 7.495175838470459, 'timing/data': 0.011416947934776545, 'timing/model': 1.5632600570097566, 'learning_rate/qwen_vl_interface': 7.026871104922273e-06, 'learning_rate/action_model': 7.026871104922273e-05, 'epoch': 0.79}) +[2026-06-16 02:07:41,003][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.081688791513443, 'train/grad_norm_pre_clip': 8.150838851928711, 'timing/data': 0.01115724677219987, 'timing/model': 1.5993561437353492, 'learning_rate/qwen_vl_interface': 7.024046424607596e-06, 'learning_rate/action_model': 7.024046424607595e-05, 'epoch': 0.79}) +[2026-06-16 02:07:47,384][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.10270921140909195, 'train/grad_norm_pre_clip': 4.308535575866699, 'timing/data': 0.01135424803942442, 'timing/model': 1.8812951352447271, 'learning_rate/qwen_vl_interface': 7.021221015049445e-06, 'learning_rate/action_model': 7.021221015049446e-05, 'epoch': 0.79}) +[2026-06-16 02:07:53,466][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.05137595534324646, 'train/grad_norm_pre_clip': 5.144199371337891, 'timing/data': 0.010681775398552418, 'timing/model': 1.5662281066179276, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 0.79}) +[2026-06-16 02:07:59,849][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.11166024953126907, 'train/grad_norm_pre_clip': 15.314050674438477, 'timing/data': 0.011779939755797386, 'timing/model': 1.648372179362923, 'learning_rate/qwen_vl_interface': 7.015568012848699e-06, 'learning_rate/action_model': 7.015568012848699e-05, 'epoch': 0.79}) +[2026-06-16 02:08:06,159][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.16078393161296844, 'train/grad_norm_pre_clip': 9.48166561126709, 'timing/data': 0.011096916627138853, 'timing/model': 1.552753010764718, 'learning_rate/qwen_vl_interface': 7.01274042252984e-06, 'learning_rate/action_model': 7.01274042252984e-05, 'epoch': 0.79}) +[2026-06-16 02:08:12,563][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.109328493475914, 'train/grad_norm_pre_clip': 10.05446720123291, 'timing/data': 0.010957716032862663, 'timing/model': 1.5629595988430083, 'learning_rate/qwen_vl_interface': 7.009912107614978e-06, 'learning_rate/action_model': 7.009912107614978e-05, 'epoch': 0.79}) +[2026-06-16 02:08:18,859][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.05387033522129059, 'train/grad_norm_pre_clip': 6.381954669952393, 'timing/data': 0.01152871921658516, 'timing/model': 1.538732911925763, 'learning_rate/qwen_vl_interface': 7.0070830692667244e-06, 'learning_rate/action_model': 7.007083069266724e-05, 'epoch': 0.79}) +[2026-06-16 02:08:25,201][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.20518384873867035, 'train/grad_norm_pre_clip': 7.84395694732666, 'timing/data': 0.010077773127704859, 'timing/model': 1.7994065494276583, 'learning_rate/qwen_vl_interface': 7.004253308647993e-06, 'learning_rate/action_model': 7.004253308647994e-05, 'epoch': 0.79}) +[2026-06-16 02:08:31,274][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.009117953479290009, 'train/grad_norm_pre_clip': 0.848787248134613, 'timing/data': 0.011521088890731335, 'timing/model': 1.5695203337818384, 'learning_rate/qwen_vl_interface': 7.001422826921991e-06, 'learning_rate/action_model': 7.00142282692199e-05, 'epoch': 0.79}) +[2026-06-16 02:08:37,672][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.08011895418167114, 'train/grad_norm_pre_clip': 11.9918212890625, 'timing/data': 0.011389507912099361, 'timing/model': 1.5679928730241954, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 0.79}) +[2026-06-16 02:08:44,066][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.04566888138651848, 'train/grad_norm_pre_clip': 5.7498273849487305, 'timing/data': 0.011806870810687542, 'timing/model': 1.577595247887075, 'learning_rate/qwen_vl_interface': 6.995759704802485e-06, 'learning_rate/action_model': 6.995759704802485e-05, 'epoch': 0.79}) +[2026-06-16 02:08:50,439][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.1251721978187561, 'train/grad_norm_pre_clip': 6.569403648376465, 'timing/data': 0.012030900921672583, 'timing/model': 1.5918538640253246, 'learning_rate/qwen_vl_interface': 6.992927066736876e-06, 'learning_rate/action_model': 6.992927066736875e-05, 'epoch': 0.79}) +[2026-06-16 02:08:56,776][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.1515851765871048, 'train/grad_norm_pre_clip': 6.316780090332031, 'timing/data': 0.011084756813943386, 'timing/model': 1.5836817352101207, 'learning_rate/qwen_vl_interface': 6.990093712219787e-06, 'learning_rate/action_model': 6.990093712219787e-05, 'epoch': 0.79}) +[2026-06-16 02:09:03,206][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.35607272386550903, 'train/grad_norm_pre_clip': 12.252455711364746, 'timing/data': 0.011329437606036663, 'timing/model': 1.8146679601632059, 'learning_rate/qwen_vl_interface': 6.9872596424159e-06, 'learning_rate/action_model': 6.9872596424159e-05, 'epoch': 0.8}) +[2026-06-16 02:09:09,263][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.2569577097892761, 'train/grad_norm_pre_clip': 18.219751358032227, 'timing/data': 0.011637920048087835, 'timing/model': 1.556803209707141, 'learning_rate/qwen_vl_interface': 6.9844248584902e-06, 'learning_rate/action_model': 6.984424858490199e-05, 'epoch': 0.8}) +[2026-06-16 02:09:15,665][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.013051318004727364, 'train/grad_norm_pre_clip': 1.3774504661560059, 'timing/data': 0.011451919097453356, 'timing/model': 1.5668456079438329, 'learning_rate/qwen_vl_interface': 6.981589361607952e-06, 'learning_rate/action_model': 6.981589361607953e-05, 'epoch': 0.8}) +[2026-06-16 02:09:21,968][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.08673615753650665, 'train/grad_norm_pre_clip': 4.510090351104736, 'timing/data': 0.011847620829939842, 'timing/model': 1.556859659962356, 'learning_rate/qwen_vl_interface': 6.9787531529347274e-06, 'learning_rate/action_model': 6.978753152934727e-05, 'epoch': 0.8}) +[2026-06-16 02:09:28,273][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.06363866478204727, 'train/grad_norm_pre_clip': 7.132899761199951, 'timing/data': 0.011820310726761818, 'timing/model': 1.5614234544336796, 'learning_rate/qwen_vl_interface': 6.97591623363638e-06, 'learning_rate/action_model': 6.97591623363638e-05, 'epoch': 0.8}) +[2026-06-16 02:09:34,638][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.07962256669998169, 'train/grad_norm_pre_clip': 11.959826469421387, 'timing/data': 0.011124297976493835, 'timing/model': 1.5858715828508139, 'learning_rate/qwen_vl_interface': 6.9730786048790644e-06, 'learning_rate/action_model': 6.973078604879064e-05, 'epoch': 0.8}) +[2026-06-16 02:09:41,100][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.12107852101325989, 'train/grad_norm_pre_clip': 11.84578800201416, 'timing/data': 0.011579068843275309, 'timing/model': 1.8064157883636653, 'learning_rate/qwen_vl_interface': 6.970240267829221e-06, 'learning_rate/action_model': 6.970240267829222e-05, 'epoch': 0.8}) +[2026-06-16 02:09:47,168][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.05977490171790123, 'train/grad_norm_pre_clip': 12.388277053833008, 'timing/data': 0.010636144783347845, 'timing/model': 1.5657547670416534, 'learning_rate/qwen_vl_interface': 6.967401223653582e-06, 'learning_rate/action_model': 6.967401223653581e-05, 'epoch': 0.8}) +[2026-06-16 02:09:53,468][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.08686563372612, 'train/grad_norm_pre_clip': 7.098459243774414, 'timing/data': 0.01190247107297182, 'timing/model': 1.5686050131917, 'learning_rate/qwen_vl_interface': 6.9645614735191724e-06, 'learning_rate/action_model': 6.964561473519172e-05, 'epoch': 0.8}) +[2026-06-16 02:09:59,755][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.11116567999124527, 'train/grad_norm_pre_clip': 7.688276767730713, 'timing/data': 0.010652875062078238, 'timing/model': 1.5583835947327316, 'learning_rate/qwen_vl_interface': 6.961721018593305e-06, 'learning_rate/action_model': 6.961721018593305e-05, 'epoch': 0.8}) +[2026-06-16 02:10:06,062][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.013940342701971531, 'train/grad_norm_pre_clip': 1.0908626317977905, 'timing/data': 0.010708755813539028, 'timing/model': 1.5353965302929282, 'learning_rate/qwen_vl_interface': 6.958879860043585e-06, 'learning_rate/action_model': 6.958879860043585e-05, 'epoch': 0.8}) +[2026-06-16 02:10:12,462][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.09306125342845917, 'train/grad_norm_pre_clip': 8.336010932922363, 'timing/data': 0.011840650346130133, 'timing/model': 1.6079966546967626, 'learning_rate/qwen_vl_interface': 6.9560379990379025e-06, 'learning_rate/action_model': 6.956037999037902e-05, 'epoch': 0.8}) +[2026-06-16 02:10:18,798][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.4568940997123718, 'train/grad_norm_pre_clip': 22.700868606567383, 'timing/data': 0.01126392837613821, 'timing/model': 1.7949620196595788, 'learning_rate/qwen_vl_interface': 6.953195436744442e-06, 'learning_rate/action_model': 6.953195436744442e-05, 'epoch': 0.8}) +[2026-06-16 02:10:25,161][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.23863868415355682, 'train/grad_norm_pre_clip': 5.398137092590332, 'timing/data': 0.011931650806218386, 'timing/model': 1.5528141180984676, 'learning_rate/qwen_vl_interface': 6.950352174331676e-06, 'learning_rate/action_model': 6.950352174331675e-05, 'epoch': 0.8}) +[2026-06-16 02:10:31,482][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.1742231696844101, 'train/grad_norm_pre_clip': 9.25224781036377, 'timing/data': 0.011053938418626785, 'timing/model': 1.5814499836415052, 'learning_rate/qwen_vl_interface': 6.947508212968356e-06, 'learning_rate/action_model': 6.947508212968356e-05, 'epoch': 0.8}) +[2026-06-16 02:10:37,771][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.023521075025200844, 'train/grad_norm_pre_clip': 12.855984687805176, 'timing/data': 0.011341058183461428, 'timing/model': 1.5710680037736893, 'learning_rate/qwen_vl_interface': 6.9446635538235295e-06, 'learning_rate/action_model': 6.944663553823529e-05, 'epoch': 0.8}) +[2026-06-16 02:10:44,068][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.06199845299124718, 'train/grad_norm_pre_clip': 5.654228687286377, 'timing/data': 0.01141987880691886, 'timing/model': 1.5681284959428012, 'learning_rate/qwen_vl_interface': 6.94181819806653e-06, 'learning_rate/action_model': 6.94181819806653e-05, 'epoch': 0.8}) +[2026-06-16 02:10:50,395][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.061413880437612534, 'train/grad_norm_pre_clip': 9.739145278930664, 'timing/data': 0.01155798975378275, 'timing/model': 1.546183834783733, 'learning_rate/qwen_vl_interface': 6.938972146866975e-06, 'learning_rate/action_model': 6.938972146866974e-05, 'epoch': 0.8}) +[2026-06-16 02:10:56,712][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.010430656373500824, 'train/grad_norm_pre_clip': 1.092848777770996, 'timing/data': 0.011008827947080135, 'timing/model': 1.7960942820645869, 'learning_rate/qwen_vl_interface': 6.936125401394769e-06, 'learning_rate/action_model': 6.936125401394768e-05, 'epoch': 0.8}) +[2026-06-16 02:11:02,768][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.11426321417093277, 'train/grad_norm_pre_clip': 9.586782455444336, 'timing/data': 0.012010572012513876, 'timing/model': 1.5623102122917771, 'learning_rate/qwen_vl_interface': 6.933277962820101e-06, 'learning_rate/action_model': 6.9332779628201e-05, 'epoch': 0.8}) +[2026-06-16 02:11:09,063][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.015208860859274864, 'train/grad_norm_pre_clip': 1.3423482179641724, 'timing/data': 0.010878356639295816, 'timing/model': 1.5530693759210408, 'learning_rate/qwen_vl_interface': 6.930429832313444e-06, 'learning_rate/action_model': 6.930429832313444e-05, 'epoch': 0.8}) +[2026-06-16 02:11:15,363][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.10880886018276215, 'train/grad_norm_pre_clip': 8.33869457244873, 'timing/data': 0.011467468924820423, 'timing/model': 1.562735672108829, 'learning_rate/qwen_vl_interface': 6.92758101104556e-06, 'learning_rate/action_model': 6.92758101104556e-05, 'epoch': 0.8}) +[2026-06-16 02:11:21,662][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.0973939374089241, 'train/grad_norm_pre_clip': 9.138891220092773, 'timing/data': 0.011273497715592384, 'timing/model': 1.5614949106238782, 'learning_rate/qwen_vl_interface': 6.9247315001874885e-06, 'learning_rate/action_model': 6.924731500187488e-05, 'epoch': 0.8}) +[2026-06-16 02:11:28,048][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.031193677335977554, 'train/grad_norm_pre_clip': 7.654852867126465, 'timing/data': 0.013369877822697163, 'timing/model': 1.5879401778802276, 'learning_rate/qwen_vl_interface': 6.9218813009105605e-06, 'learning_rate/action_model': 6.92188130091056e-05, 'epoch': 0.8}) +[2026-06-16 02:11:34,418][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.05248246714472771, 'train/grad_norm_pre_clip': 14.216435432434082, 'timing/data': 0.011658149771392345, 'timing/model': 1.8105990909971297, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.8}) +[2026-06-16 02:11:40,635][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.028141142800450325, 'train/grad_norm_pre_clip': 4.99760627746582, 'timing/data': 0.01159422006458044, 'timing/model': 1.726682084146887, 'learning_rate/qwen_vl_interface': 6.916178841786839e-06, 'learning_rate/action_model': 6.916178841786838e-05, 'epoch': 0.81}) +[2026-06-16 02:11:46,862][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.0706004798412323, 'train/grad_norm_pre_clip': 8.131359100341797, 'timing/data': 0.010844886768609285, 'timing/model': 1.5628123399801552, 'learning_rate/qwen_vl_interface': 6.913326584284115e-06, 'learning_rate/action_model': 6.913326584284115e-05, 'epoch': 0.81}) +[2026-06-16 02:11:53,236][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.21605926752090454, 'train/grad_norm_pre_clip': 11.506094932556152, 'timing/data': 0.012845115270465612, 'timing/model': 1.6250144732184708, 'learning_rate/qwen_vl_interface': 6.910473643050662e-06, 'learning_rate/action_model': 6.910473643050662e-05, 'epoch': 0.81}) +[2026-06-16 02:11:59,567][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.1531410962343216, 'train/grad_norm_pre_clip': 14.412775993347168, 'timing/data': 0.010981307830661535, 'timing/model': 1.5652793082408607, 'learning_rate/qwen_vl_interface': 6.907620019259215e-06, 'learning_rate/action_model': 6.907620019259216e-05, 'epoch': 0.81}) +[2026-06-16 02:12:06,057][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.013104717247188091, 'train/grad_norm_pre_clip': 1.8280235528945923, 'timing/data': 0.0121081518009305, 'timing/model': 1.579356084112078, 'learning_rate/qwen_vl_interface': 6.9047657140827935e-06, 'learning_rate/action_model': 6.904765714082794e-05, 'epoch': 0.81}) +[2026-06-16 02:12:12,424][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.003363557392731309, 'train/grad_norm_pre_clip': 0.4936907887458801, 'timing/data': 0.012264783028513193, 'timing/model': 1.8243660246953368, 'learning_rate/qwen_vl_interface': 6.901910728694694e-06, 'learning_rate/action_model': 6.901910728694694e-05, 'epoch': 0.81}) +[2026-06-16 02:12:18,685][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.014166180044412613, 'train/grad_norm_pre_clip': 2.7336671352386475, 'timing/data': 0.011390399187803268, 'timing/model': 1.78515778016299, 'learning_rate/qwen_vl_interface': 6.899055064268489e-06, 'learning_rate/action_model': 6.899055064268489e-05, 'epoch': 0.81}) +[2026-06-16 02:12:24,758][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.09146952629089355, 'train/grad_norm_pre_clip': 12.3240385055542, 'timing/data': 0.01142706023529172, 'timing/model': 1.5560920620337129, 'learning_rate/qwen_vl_interface': 6.896198721978038e-06, 'learning_rate/action_model': 6.896198721978038e-05, 'epoch': 0.81}) +[2026-06-16 02:12:31,056][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.04169750213623047, 'train/grad_norm_pre_clip': 5.7754716873168945, 'timing/data': 0.011683660093694925, 'timing/model': 1.562366392929107, 'learning_rate/qwen_vl_interface': 6.893341702997475e-06, 'learning_rate/action_model': 6.893341702997475e-05, 'epoch': 0.81}) +[2026-06-16 02:23:27,920][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.03640956059098244, 'train/grad_norm_pre_clip': 7.611759185791016, 'eval/loss': 0.06292103230953217, 'eval/flap_precision': 0.9244558258642765, 'eval/flap_recall': 0.984993178717599, 'eval/flap_f1': 0.9537648612945839, 'eval/flap_rate_model': 0.21694444444444444, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.953764861294584, 'timing/data': 0.011447509750723839, 'timing/model': 1.5622676853090525, 'learning_rate/qwen_vl_interface': 6.890484008501211e-06, 'learning_rate/action_model': 6.890484008501211e-05, 'epoch': 0.81}) +[2026-06-16 02:23:27,921][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2000_state +[2026-06-16 02:23:27,928][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:23:36,451][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2000_state/model.safetensors +[2026-06-16 02:23:55,771][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2000_state/optimizer.bin +[2026-06-16 02:23:55,772][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2000_state/sampler.bin +[2026-06-16 02:23:55,773][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-16 02:23:55,774][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-16 02:23:55,965][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-16 02:23:55,965][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:23:55,981][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:24:02,323][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.05743253603577614, 'train/grad_norm_pre_clip': 5.122783660888672, 'timing/data': 0.01084444997832179, 'timing/model': 1.5678764386102557, 'learning_rate/qwen_vl_interface': 6.887625639663941e-06, 'learning_rate/action_model': 6.88762563966394e-05, 'epoch': 0.81}) +[2026-06-16 02:24:08,523][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.11632054299116135, 'train/grad_norm_pre_clip': 17.616899490356445, 'timing/data': 0.013837568927556276, 'timing/model': 1.5474524502642453, 'learning_rate/qwen_vl_interface': 6.884766597660626e-06, 'learning_rate/action_model': 6.884766597660626e-05, 'epoch': 0.81}) +[2026-06-16 02:24:14,630][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.1570860743522644, 'train/grad_norm_pre_clip': 26.792198181152344, 'timing/data': 0.0118963741697371, 'timing/model': 1.5526555515825748, 'learning_rate/qwen_vl_interface': 6.881906883666515e-06, 'learning_rate/action_model': 6.881906883666515e-05, 'epoch': 0.81}) +[2026-06-16 02:24:20,804][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.13761861622333527, 'train/grad_norm_pre_clip': 19.586692810058594, 'timing/data': 0.011077451054006815, 'timing/model': 1.749121729284525, 'learning_rate/qwen_vl_interface': 6.879046498857125e-06, 'learning_rate/action_model': 6.879046498857124e-05, 'epoch': 0.81}) +[2026-06-16 02:24:26,839][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.13995088636875153, 'train/grad_norm_pre_clip': 32.640926361083984, 'timing/data': 0.010457700118422508, 'timing/model': 1.5651324619539082, 'learning_rate/qwen_vl_interface': 6.876185444408255e-06, 'learning_rate/action_model': 6.876185444408255e-05, 'epoch': 0.81}) +[2026-06-16 02:24:33,030][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.10943122953176498, 'train/grad_norm_pre_clip': 17.490970611572266, 'timing/data': 0.011161942034959793, 'timing/model': 1.520419369917363, 'learning_rate/qwen_vl_interface': 6.873323721495976e-06, 'learning_rate/action_model': 6.873323721495976e-05, 'epoch': 0.81}) +[2026-06-16 02:24:39,238][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.325059175491333, 'train/grad_norm_pre_clip': 18.34485626220703, 'timing/data': 0.011033032089471817, 'timing/model': 1.5695483293384314, 'learning_rate/qwen_vl_interface': 6.870461331296632e-06, 'learning_rate/action_model': 6.870461331296632e-05, 'epoch': 0.81}) +[2026-06-16 02:24:45,449][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.43333154916763306, 'train/grad_norm_pre_clip': 25.26479148864746, 'timing/data': 0.010205020196735859, 'timing/model': 1.5477532222867012, 'learning_rate/qwen_vl_interface': 6.8675982749868455e-06, 'learning_rate/action_model': 6.867598274986845e-05, 'epoch': 0.81}) +[2026-06-16 02:24:51,751][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.27397027611732483, 'train/grad_norm_pre_clip': 15.827438354492188, 'timing/data': 0.010796781163662672, 'timing/model': 1.5626678923144937, 'learning_rate/qwen_vl_interface': 6.864734553743509e-06, 'learning_rate/action_model': 6.86473455374351e-05, 'epoch': 0.81}) +[2026-06-16 02:24:58,032][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.08611759543418884, 'train/grad_norm_pre_clip': 18.55262565612793, 'timing/data': 0.010959941893815994, 'timing/model': 1.7744821230880916, 'learning_rate/qwen_vl_interface': 6.861870168743793e-06, 'learning_rate/action_model': 6.861870168743793e-05, 'epoch': 0.81}) +[2026-06-16 02:25:04,037][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.1553369015455246, 'train/grad_norm_pre_clip': 11.410472869873047, 'timing/data': 0.009874900337308645, 'timing/model': 1.552241011057049, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 0.81}) +[2026-06-16 02:25:10,235][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.025888288393616676, 'train/grad_norm_pre_clip': 8.465438842773438, 'timing/data': 0.009816049132496119, 'timing/model': 1.5684167360886931, 'learning_rate/qwen_vl_interface': 6.85613941218525e-06, 'learning_rate/action_model': 6.85613941218525e-05, 'epoch': 0.81}) +[2026-06-16 02:25:16,445][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.09371873736381531, 'train/grad_norm_pre_clip': 15.971860885620117, 'timing/data': 0.009629708714783192, 'timing/model': 1.5595619538798928, 'learning_rate/qwen_vl_interface': 6.8532730429821215e-06, 'learning_rate/action_model': 6.853273042982122e-05, 'epoch': 0.81}) +[2026-06-16 02:25:22,741][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.10886020213365555, 'train/grad_norm_pre_clip': 4.837394714355469, 'timing/data': 0.011030713096261024, 'timing/model': 1.515524946153164, 'learning_rate/qwen_vl_interface': 6.850406014734005e-06, 'learning_rate/action_model': 6.850406014734005e-05, 'epoch': 0.81}) +[2026-06-16 02:25:28,990][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.01683162897825241, 'train/grad_norm_pre_clip': 4.191192150115967, 'timing/data': 0.011533655226230621, 'timing/model': 1.806056389119476, 'learning_rate/qwen_vl_interface': 6.8475383286194295e-06, 'learning_rate/action_model': 6.84753832861943e-05, 'epoch': 0.81}) +[2026-06-16 02:25:35,045][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.010532569140195847, 'train/grad_norm_pre_clip': 1.7961819171905518, 'timing/data': 0.010145410895347595, 'timing/model': 1.5592283247970045, 'learning_rate/qwen_vl_interface': 6.844669985817192e-06, 'learning_rate/action_model': 6.844669985817192e-05, 'epoch': 0.82}) +[2026-06-16 02:25:41,461][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.039109814912080765, 'train/grad_norm_pre_clip': 6.570675849914551, 'timing/data': 0.011412975378334522, 'timing/model': 1.5856549409218132, 'learning_rate/qwen_vl_interface': 6.8418009875063594e-06, 'learning_rate/action_model': 6.84180098750636e-05, 'epoch': 0.82}) +[2026-06-16 02:25:47,651][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.011917559430003166, 'train/grad_norm_pre_clip': 1.0415632724761963, 'timing/data': 0.011590436100959778, 'timing/model': 1.5577054000459611, 'learning_rate/qwen_vl_interface': 6.838931334866267e-06, 'learning_rate/action_model': 6.838931334866266e-05, 'epoch': 0.82}) +[2026-06-16 02:25:53,957][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.04748691990971565, 'train/grad_norm_pre_clip': 7.3042097091674805, 'timing/data': 0.01353670097887516, 'timing/model': 1.5607194020412862, 'learning_rate/qwen_vl_interface': 6.836061029076523e-06, 'learning_rate/action_model': 6.836061029076522e-05, 'epoch': 0.82}) +[2026-06-16 02:26:00,262][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.06755246967077255, 'train/grad_norm_pre_clip': 11.444937705993652, 'timing/data': 0.012678819242864847, 'timing/model': 1.5649992991238832, 'learning_rate/qwen_vl_interface': 6.833190071317e-06, 'learning_rate/action_model': 6.833190071316999e-05, 'epoch': 0.82}) +[2026-06-16 02:26:06,684][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.03080645389854908, 'train/grad_norm_pre_clip': 4.325528144836426, 'timing/data': 0.011537026148289442, 'timing/model': 1.8015919090248644, 'learning_rate/qwen_vl_interface': 6.830318462767841e-06, 'learning_rate/action_model': 6.830318462767841e-05, 'epoch': 0.82}) +[2026-06-16 02:26:12,758][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.011025979183614254, 'train/grad_norm_pre_clip': 2.2242743968963623, 'timing/data': 0.011493615806102753, 'timing/model': 1.5707935420796275, 'learning_rate/qwen_vl_interface': 6.827446204609456e-06, 'learning_rate/action_model': 6.827446204609456e-05, 'epoch': 0.82}) +[2026-06-16 02:26:19,053][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.11615912616252899, 'train/grad_norm_pre_clip': 15.376484870910645, 'timing/data': 0.011567067354917526, 'timing/model': 1.5489975158125162, 'learning_rate/qwen_vl_interface': 6.824573298022522e-06, 'learning_rate/action_model': 6.824573298022522e-05, 'epoch': 0.82}) +[2026-06-16 02:26:25,353][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.06747187674045563, 'train/grad_norm_pre_clip': 13.145394325256348, 'timing/data': 0.011796288192272186, 'timing/model': 1.5488919368945062, 'learning_rate/qwen_vl_interface': 6.8216997441879805e-06, 'learning_rate/action_model': 6.821699744187981e-05, 'epoch': 0.82}) +[2026-06-16 02:26:31,762][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.3738245368003845, 'train/grad_norm_pre_clip': 28.324655532836914, 'timing/data': 0.011570617090910673, 'timing/model': 1.5637712869793177, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 0.82}) +[2026-06-16 02:26:38,053][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.012864229269325733, 'train/grad_norm_pre_clip': 1.8119078874588013, 'timing/data': 0.011284566018730402, 'timing/model': 1.551869269926101, 'learning_rate/qwen_vl_interface': 6.815950699501186e-06, 'learning_rate/action_model': 6.815950699501186e-05, 'epoch': 0.82}) +[2026-06-16 02:26:44,402][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.19723278284072876, 'train/grad_norm_pre_clip': 21.445907592773438, 'timing/data': 0.01116838539019227, 'timing/model': 1.803587379399687, 'learning_rate/qwen_vl_interface': 6.813075211012148e-06, 'learning_rate/action_model': 6.813075211012147e-05, 'epoch': 0.82}) +[2026-06-16 02:26:50,557][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.08553481847047806, 'train/grad_norm_pre_clip': 7.148974418640137, 'timing/data': 0.011061484925448895, 'timing/model': 1.562494516838342, 'learning_rate/qwen_vl_interface': 6.81019908000193e-06, 'learning_rate/action_model': 6.81019908000193e-05, 'epoch': 0.82}) +[2026-06-16 02:26:56,861][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.05179827660322189, 'train/grad_norm_pre_clip': 9.197267532348633, 'timing/data': 0.011009796056896448, 'timing/model': 1.5439928169362247, 'learning_rate/qwen_vl_interface': 6.807322307652806e-06, 'learning_rate/action_model': 6.807322307652806e-05, 'epoch': 0.82}) +[2026-06-16 02:27:03,239][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.02356269210577011, 'train/grad_norm_pre_clip': 8.856621742248535, 'timing/data': 0.010920934844762087, 'timing/model': 1.6309632998891175, 'learning_rate/qwen_vl_interface': 6.804444895147306e-06, 'learning_rate/action_model': 6.804444895147306e-05, 'epoch': 0.82}) +[2026-06-16 02:27:09,577][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.2536031901836395, 'train/grad_norm_pre_clip': 20.93123435974121, 'timing/data': 0.010713335126638412, 'timing/model': 1.6020976179279387, 'learning_rate/qwen_vl_interface': 6.8015668436682255e-06, 'learning_rate/action_model': 6.801566843668226e-05, 'epoch': 0.82}) +[2026-06-16 02:27:15,871][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.015024986118078232, 'train/grad_norm_pre_clip': 2.8835060596466064, 'timing/data': 0.011349356733262539, 'timing/model': 1.5248521459288895, 'learning_rate/qwen_vl_interface': 6.7986881543986245e-06, 'learning_rate/action_model': 6.798688154398624e-05, 'epoch': 0.82}) +[2026-06-16 02:27:22,211][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.013687850907444954, 'train/grad_norm_pre_clip': 2.883052349090576, 'timing/data': 0.011439118068665266, 'timing/model': 1.804924094118178, 'learning_rate/qwen_vl_interface': 6.795808828521822e-06, 'learning_rate/action_model': 6.795808828521821e-05, 'epoch': 0.82}) +[2026-06-16 02:27:28,266][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.013588819652795792, 'train/grad_norm_pre_clip': 3.800963878631592, 'timing/data': 0.01125409733504057, 'timing/model': 1.5234371391125023, 'learning_rate/qwen_vl_interface': 6.792928867221401e-06, 'learning_rate/action_model': 6.792928867221401e-05, 'epoch': 0.82}) +[2026-06-16 02:27:34,569][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.006505438592284918, 'train/grad_norm_pre_clip': 3.0326969623565674, 'timing/data': 0.010724665131419897, 'timing/model': 1.5542229898273945, 'learning_rate/qwen_vl_interface': 6.790048271681202e-06, 'learning_rate/action_model': 6.790048271681202e-05, 'epoch': 0.82}) +[2026-06-16 02:27:40,934][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.398472398519516, 'train/grad_norm_pre_clip': 25.035551071166992, 'timing/data': 0.011529508046805859, 'timing/model': 1.6325090089812875, 'learning_rate/qwen_vl_interface': 6.7871670430853335e-06, 'learning_rate/action_model': 6.787167043085333e-05, 'epoch': 0.82}) +[2026-06-16 02:27:47,257][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.11431535333395004, 'train/grad_norm_pre_clip': 16.68286895751953, 'timing/data': 0.011138337198644876, 'timing/model': 1.5484887692146003, 'learning_rate/qwen_vl_interface': 6.784285182618156e-06, 'learning_rate/action_model': 6.784285182618156e-05, 'epoch': 0.82}) +[2026-06-16 02:27:53,590][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.20794172585010529, 'train/grad_norm_pre_clip': 13.02729606628418, 'timing/data': 0.011171686928719282, 'timing/model': 1.5536628048866987, 'learning_rate/qwen_vl_interface': 6.781402691464297e-06, 'learning_rate/action_model': 6.781402691464297e-05, 'epoch': 0.82}) +[2026-06-16 02:27:59,928][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.092892587184906, 'train/grad_norm_pre_clip': 15.480378150939941, 'timing/data': 0.011230848263949156, 'timing/model': 1.8342517144046724, 'learning_rate/qwen_vl_interface': 6.778519570808636e-06, 'learning_rate/action_model': 6.778519570808636e-05, 'epoch': 0.82}) +[2026-06-16 02:28:05,961][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.10621563345193863, 'train/grad_norm_pre_clip': 8.084111213684082, 'timing/data': 0.011164797935634851, 'timing/model': 1.5677760071121156, 'learning_rate/qwen_vl_interface': 6.775635821836317e-06, 'learning_rate/action_model': 6.775635821836317e-05, 'epoch': 0.82}) +[2026-06-16 02:28:12,269][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.06501911580562592, 'train/grad_norm_pre_clip': 6.209456920623779, 'timing/data': 0.011835789773613214, 'timing/model': 1.571744428947568, 'learning_rate/qwen_vl_interface': 6.77275144573274e-06, 'learning_rate/action_model': 6.77275144573274e-05, 'epoch': 0.83}) +[2026-06-16 02:28:18,556][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.19526515901088715, 'train/grad_norm_pre_clip': 14.853035926818848, 'timing/data': 0.01115430798381567, 'timing/model': 1.545617728959769, 'learning_rate/qwen_vl_interface': 6.769866443683564e-06, 'learning_rate/action_model': 6.769866443683563e-05, 'epoch': 0.83}) +[2026-06-16 02:28:24,866][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.26887017488479614, 'train/grad_norm_pre_clip': 38.780094146728516, 'timing/data': 0.010809587314724922, 'timing/model': 1.567623030859977, 'learning_rate/qwen_vl_interface': 6.766980816874701e-06, 'learning_rate/action_model': 6.7669808168747e-05, 'epoch': 0.83}) +[2026-06-16 02:28:31,173][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.1331373006105423, 'train/grad_norm_pre_clip': 19.662118911743164, 'timing/data': 0.010178785305470228, 'timing/model': 1.5278860330581665, 'learning_rate/qwen_vl_interface': 6.764094566492327e-06, 'learning_rate/action_model': 6.764094566492328e-05, 'epoch': 0.83}) +[2026-06-16 02:28:37,501][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.05601000413298607, 'train/grad_norm_pre_clip': 12.546180725097656, 'timing/data': 0.011468480341136456, 'timing/model': 1.799672769382596, 'learning_rate/qwen_vl_interface': 6.7612076937228685e-06, 'learning_rate/action_model': 6.761207693722869e-05, 'epoch': 0.83}) +[2026-06-16 02:28:43,581][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.009286623448133469, 'train/grad_norm_pre_clip': 3.97651743888855, 'timing/data': 0.012248822022229433, 'timing/model': 1.5756456600502133, 'learning_rate/qwen_vl_interface': 6.7583201997530126e-06, 'learning_rate/action_model': 6.758320199753012e-05, 'epoch': 0.83}) +[2026-06-16 02:28:50,072][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.020362766459584236, 'train/grad_norm_pre_clip': 3.9551260471343994, 'timing/data': 0.015150141902267933, 'timing/model': 1.6191878071986139, 'learning_rate/qwen_vl_interface': 6.755432085769693e-06, 'learning_rate/action_model': 6.755432085769694e-05, 'epoch': 0.83}) +[2026-06-16 02:28:56,574][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.015163196250796318, 'train/grad_norm_pre_clip': 4.94103479385376, 'timing/data': 0.011480509769171476, 'timing/model': 1.5667457426898181, 'learning_rate/qwen_vl_interface': 6.75254335296011e-06, 'learning_rate/action_model': 6.75254335296011e-05, 'epoch': 0.83}) +[2026-06-16 02:29:02,854][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.04753434285521507, 'train/grad_norm_pre_clip': 22.97454261779785, 'timing/data': 0.011059548240154982, 'timing/model': 1.5596923953853548, 'learning_rate/qwen_vl_interface': 6.749654002511712e-06, 'learning_rate/action_model': 6.749654002511712e-05, 'epoch': 0.83}) +[2026-06-16 02:29:09,160][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.051671892404556274, 'train/grad_norm_pre_clip': 16.31642723083496, 'timing/data': 0.011439749971032143, 'timing/model': 1.5182888070121408, 'learning_rate/qwen_vl_interface': 6.7467640356122e-06, 'learning_rate/action_model': 6.746764035612199e-05, 'epoch': 0.83}) +[2026-06-16 02:29:15,499][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.08170842379331589, 'train/grad_norm_pre_clip': 19.288915634155273, 'timing/data': 0.010884267278015614, 'timing/model': 1.8075335710309446, 'learning_rate/qwen_vl_interface': 6.74387345344953e-06, 'learning_rate/action_model': 6.74387345344953e-05, 'epoch': 0.83}) +[2026-06-16 02:29:21,566][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.2021251916885376, 'train/grad_norm_pre_clip': 20.783472061157227, 'timing/data': 0.010375486221164465, 'timing/model': 1.5757975387386978, 'learning_rate/qwen_vl_interface': 6.740982257211911e-06, 'learning_rate/action_model': 6.740982257211911e-05, 'epoch': 0.83}) +[2026-06-16 02:29:27,859][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.02433444745838642, 'train/grad_norm_pre_clip': 4.4015679359436035, 'timing/data': 0.012406453024595976, 'timing/model': 1.549371181987226, 'learning_rate/qwen_vl_interface': 6.738090448087806e-06, 'learning_rate/action_model': 6.738090448087806e-05, 'epoch': 0.83}) +[2026-06-16 02:29:34,157][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.003888145787641406, 'train/grad_norm_pre_clip': 0.8129186034202576, 'timing/data': 0.011373268906027079, 'timing/model': 1.5556972189806402, 'learning_rate/qwen_vl_interface': 6.735198027265928e-06, 'learning_rate/action_model': 6.735198027265928e-05, 'epoch': 0.83}) +[2026-06-16 02:29:40,462][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.16762736439704895, 'train/grad_norm_pre_clip': 21.099346160888672, 'timing/data': 0.011232038959860802, 'timing/model': 1.569565373007208, 'learning_rate/qwen_vl_interface': 6.7323049959352424e-06, 'learning_rate/action_model': 6.732304995935243e-05, 'epoch': 0.83}) +[2026-06-16 02:29:46,778][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.00656189676374197, 'train/grad_norm_pre_clip': 2.12526535987854, 'timing/data': 0.011485880240797997, 'timing/model': 1.546650956850499, 'learning_rate/qwen_vl_interface': 6.729411355284962e-06, 'learning_rate/action_model': 6.729411355284963e-05, 'epoch': 0.83}) +[2026-06-16 02:29:53,124][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.007454360835254192, 'train/grad_norm_pre_clip': 1.732688069343567, 'timing/data': 0.010893458966165781, 'timing/model': 1.829132772050798, 'learning_rate/qwen_vl_interface': 6.726517106504559e-06, 'learning_rate/action_model': 6.726517106504558e-05, 'epoch': 0.83}) +[2026-06-16 02:29:59,332][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.18444058299064636, 'train/grad_norm_pre_clip': 24.050758361816406, 'timing/data': 0.012912806123495102, 'timing/model': 1.7262183837592602, 'learning_rate/qwen_vl_interface': 6.723622250783745e-06, 'learning_rate/action_model': 6.723622250783744e-05, 'epoch': 0.83}) +[2026-06-16 02:30:05,359][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.028222907334566116, 'train/grad_norm_pre_clip': 8.220913887023926, 'timing/data': 0.011283509898930788, 'timing/model': 1.55856643198058, 'learning_rate/qwen_vl_interface': 6.720726789312487e-06, 'learning_rate/action_model': 6.720726789312487e-05, 'epoch': 0.83}) +[2026-06-16 02:30:11,662][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.008714281022548676, 'train/grad_norm_pre_clip': 2.3023123741149902, 'timing/data': 0.011099840048700571, 'timing/model': 1.5658914223313332, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.83}) +[2026-06-16 02:30:18,081][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.17854370176792145, 'train/grad_norm_pre_clip': 13.22470474243164, 'timing/data': 0.011455470696091652, 'timing/model': 1.539964143652469, 'learning_rate/qwen_vl_interface': 6.714934053879749e-06, 'learning_rate/action_model': 6.71493405387975e-05, 'epoch': 0.83}) +[2026-06-16 02:30:24,469][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.07228797674179077, 'train/grad_norm_pre_clip': 24.551206588745117, 'timing/data': 0.01122712017968297, 'timing/model': 1.5222329548560083, 'learning_rate/qwen_vl_interface': 6.712036782299442e-06, 'learning_rate/action_model': 6.712036782299443e-05, 'epoch': 0.83}) +[2026-06-16 02:30:30,803][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.10022244602441788, 'train/grad_norm_pre_clip': 26.993446350097656, 'timing/data': 0.011052889749407768, 'timing/model': 1.8027399759739637, 'learning_rate/qwen_vl_interface': 6.709138909731041e-06, 'learning_rate/action_model': 6.709138909731041e-05, 'epoch': 0.83}) +[2026-06-16 02:30:37,187][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.1299576759338379, 'train/grad_norm_pre_clip': 8.399450302124023, 'timing/data': 0.011982664000242949, 'timing/model': 1.8773675579577684, 'learning_rate/qwen_vl_interface': 6.706240437365751e-06, 'learning_rate/action_model': 6.70624043736575e-05, 'epoch': 0.83}) +[2026-06-16 02:30:43,284][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.2513881027698517, 'train/grad_norm_pre_clip': 15.82890510559082, 'timing/data': 0.014404372312128544, 'timing/model': 1.5550761711783707, 'learning_rate/qwen_vl_interface': 6.703341366395023e-06, 'learning_rate/action_model': 6.703341366395023e-05, 'epoch': 0.84}) +[2026-06-16 02:30:49,992][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.023628516122698784, 'train/grad_norm_pre_clip': 10.261874198913574, 'timing/data': 0.014581163413822651, 'timing/model': 1.56888232473284, 'learning_rate/qwen_vl_interface': 6.700441698010558e-06, 'learning_rate/action_model': 6.700441698010558e-05, 'epoch': 0.84}) +[2026-06-16 02:30:56,583][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.00892464816570282, 'train/grad_norm_pre_clip': 2.3875083923339844, 'timing/data': 0.014638903085142374, 'timing/model': 1.5946348742581904, 'learning_rate/qwen_vl_interface': 6.697541433404299e-06, 'learning_rate/action_model': 6.697541433404299e-05, 'epoch': 0.84}) +[2026-06-16 02:31:03,169][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.2703470289707184, 'train/grad_norm_pre_clip': 27.623844146728516, 'timing/data': 0.015359336044639349, 'timing/model': 1.6086033177562058, 'learning_rate/qwen_vl_interface': 6.6946405737684325e-06, 'learning_rate/action_model': 6.694640573768433e-05, 'epoch': 0.84}) +[2026-06-16 02:31:09,743][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.14936667680740356, 'train/grad_norm_pre_clip': 21.95547103881836, 'timing/data': 0.01570004690438509, 'timing/model': 2.0142347221262753, 'learning_rate/qwen_vl_interface': 6.691739120295397e-06, 'learning_rate/action_model': 6.691739120295397e-05, 'epoch': 0.84}) +[2026-06-16 02:31:15,902][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.20628587901592255, 'train/grad_norm_pre_clip': 28.546459197998047, 'timing/data': 0.015137654729187489, 'timing/model': 1.5577489077113569, 'learning_rate/qwen_vl_interface': 6.688837074177868e-06, 'learning_rate/action_model': 6.688837074177868e-05, 'epoch': 0.84}) +[2026-06-16 02:31:22,455][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.2094997763633728, 'train/grad_norm_pre_clip': 29.647354125976562, 'timing/data': 0.010748819913715124, 'timing/model': 1.5632765009067953, 'learning_rate/qwen_vl_interface': 6.68593443660877e-06, 'learning_rate/action_model': 6.685934436608769e-05, 'epoch': 0.84}) +[2026-06-16 02:31:28,754][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.0012004949385300279, 'train/grad_norm_pre_clip': 0.43374043703079224, 'timing/data': 0.010427968110889196, 'timing/model': 1.562665211968124, 'learning_rate/qwen_vl_interface': 6.6830312087812625e-06, 'learning_rate/action_model': 6.683031208781262e-05, 'epoch': 0.84}) +[2026-06-16 02:31:35,093][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.3054116368293762, 'train/grad_norm_pre_clip': 15.993358612060547, 'timing/data': 0.01131349103525281, 'timing/model': 1.5547723253257573, 'learning_rate/qwen_vl_interface': 6.680127391888754e-06, 'learning_rate/action_model': 6.680127391888754e-05, 'epoch': 0.84}) +[2026-06-16 02:31:41,359][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.1394234597682953, 'train/grad_norm_pre_clip': 18.041933059692383, 'timing/data': 0.01100624119862914, 'timing/model': 1.586109062191099, 'learning_rate/qwen_vl_interface': 6.677222987124897e-06, 'learning_rate/action_model': 6.677222987124897e-05, 'epoch': 0.84}) +[2026-06-16 02:31:47,665][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.010137146338820457, 'train/grad_norm_pre_clip': 1.6236375570297241, 'timing/data': 0.011710142716765404, 'timing/model': 1.749308757018298, 'learning_rate/qwen_vl_interface': 6.67431799568358e-06, 'learning_rate/action_model': 6.67431799568358e-05, 'epoch': 0.84}) +[2026-06-16 02:31:53,776][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.008554738946259022, 'train/grad_norm_pre_clip': 2.2592062950134277, 'timing/data': 0.011185541283339262, 'timing/model': 1.5756831471808255, 'learning_rate/qwen_vl_interface': 6.671412418758939e-06, 'learning_rate/action_model': 6.671412418758938e-05, 'epoch': 0.84}) +[2026-06-16 02:32:00,158][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.1844533234834671, 'train/grad_norm_pre_clip': 24.20311164855957, 'timing/data': 0.011335182003676891, 'timing/model': 1.5565315973944962, 'learning_rate/qwen_vl_interface': 6.668506257545341e-06, 'learning_rate/action_model': 6.668506257545341e-05, 'epoch': 0.84}) +[2026-06-16 02:32:06,473][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.26144877076148987, 'train/grad_norm_pre_clip': 18.14275550842285, 'timing/data': 0.012790197972208261, 'timing/model': 1.5717631950974464, 'learning_rate/qwen_vl_interface': 6.665599513237403e-06, 'learning_rate/action_model': 6.665599513237402e-05, 'epoch': 0.84}) +[2026-06-16 02:32:12,756][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.22500842809677124, 'train/grad_norm_pre_clip': 10.773030281066895, 'timing/data': 0.011113720946013927, 'timing/model': 1.5245877811685205, 'learning_rate/qwen_vl_interface': 6.662692187029978e-06, 'learning_rate/action_model': 6.662692187029978e-05, 'epoch': 0.84}) +[2026-06-16 02:32:19,054][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.38457268476486206, 'train/grad_norm_pre_clip': 35.129642486572266, 'timing/data': 0.011214261874556541, 'timing/model': 1.5094960550777614, 'learning_rate/qwen_vl_interface': 6.659784280118156e-06, 'learning_rate/action_model': 6.659784280118156e-05, 'epoch': 0.84}) +[2026-06-16 02:32:25,347][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.11620726436376572, 'train/grad_norm_pre_clip': 20.550987243652344, 'timing/data': 0.010849220212548971, 'timing/model': 1.7471732245758176, 'learning_rate/qwen_vl_interface': 6.65687579369727e-06, 'learning_rate/action_model': 6.656875793697271e-05, 'epoch': 0.84}) +[2026-06-16 02:32:31,462][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.16693197190761566, 'train/grad_norm_pre_clip': 10.046955108642578, 'timing/data': 0.011459253262728453, 'timing/model': 1.5702785239554942, 'learning_rate/qwen_vl_interface': 6.653966728962885e-06, 'learning_rate/action_model': 6.653966728962885e-05, 'epoch': 0.84}) +[2026-06-16 02:32:37,757][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.2018072009086609, 'train/grad_norm_pre_clip': 12.70401382446289, 'timing/data': 0.010737200267612934, 'timing/model': 1.5645333142019808, 'learning_rate/qwen_vl_interface': 6.651057087110813e-06, 'learning_rate/action_model': 6.651057087110813e-05, 'epoch': 0.84}) +[2026-06-16 02:32:44,052][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.006601063534617424, 'train/grad_norm_pre_clip': 0.9723396301269531, 'timing/data': 0.01129525201395154, 'timing/model': 1.5187999717891216, 'learning_rate/qwen_vl_interface': 6.6481468693370924e-06, 'learning_rate/action_model': 6.648146869337091e-05, 'epoch': 0.84}) +[2026-06-16 02:32:50,351][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.002684468636289239, 'train/grad_norm_pre_clip': 0.24370019137859344, 'timing/data': 0.011485784314572811, 'timing/model': 1.5155998701229692, 'learning_rate/qwen_vl_interface': 6.645236076838009e-06, 'learning_rate/action_model': 6.645236076838009e-05, 'epoch': 0.84}) +[2026-06-16 02:32:56,704][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.1885382980108261, 'train/grad_norm_pre_clip': 8.288029670715332, 'timing/data': 0.010793501045554876, 'timing/model': 1.811473117209971, 'learning_rate/qwen_vl_interface': 6.642324710810075e-06, 'learning_rate/action_model': 6.642324710810076e-05, 'epoch': 0.84}) +[2026-06-16 02:33:02,776][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.0035445557441562414, 'train/grad_norm_pre_clip': 0.18888741731643677, 'timing/data': 0.017477355897426605, 'timing/model': 1.573923490010202, 'learning_rate/qwen_vl_interface': 6.639412772450047e-06, 'learning_rate/action_model': 6.639412772450047e-05, 'epoch': 0.84}) +[2026-06-16 02:33:09,060][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.14334169030189514, 'train/grad_norm_pre_clip': 29.122695922851562, 'timing/data': 0.012459737714380026, 'timing/model': 1.5533914221450686, 'learning_rate/qwen_vl_interface': 6.6365002629549104e-06, 'learning_rate/action_model': 6.63650026295491e-05, 'epoch': 0.84}) +[2026-06-16 02:33:15,363][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.09496419131755829, 'train/grad_norm_pre_clip': 14.220311164855957, 'timing/data': 0.011198332998901606, 'timing/model': 1.553992375265807, 'learning_rate/qwen_vl_interface': 6.633587183521887e-06, 'learning_rate/action_model': 6.633587183521887e-05, 'epoch': 0.84}) +[2026-06-16 02:33:21,670][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.0067852758802473545, 'train/grad_norm_pre_clip': 0.5786146521568298, 'timing/data': 0.011344152968376875, 'timing/model': 1.5713930698111653, 'learning_rate/qwen_vl_interface': 6.6306735353484365e-06, 'learning_rate/action_model': 6.630673535348436e-05, 'epoch': 0.85}) +[2026-06-16 02:33:27,968][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.5388424396514893, 'train/grad_norm_pre_clip': 17.223472595214844, 'timing/data': 0.0110965920612216, 'timing/model': 1.5733773680403829, 'learning_rate/qwen_vl_interface': 6.627759319632246e-06, 'learning_rate/action_model': 6.627759319632247e-05, 'epoch': 0.85}) +[2026-06-16 02:33:34,321][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.045992497354745865, 'train/grad_norm_pre_clip': 13.381041526794434, 'timing/data': 0.011184512171894312, 'timing/model': 1.8165819849818945, 'learning_rate/qwen_vl_interface': 6.624844537571245e-06, 'learning_rate/action_model': 6.624844537571245e-05, 'epoch': 0.85}) +[2026-06-16 02:33:40,466][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.04618325084447861, 'train/grad_norm_pre_clip': 13.0364408493042, 'timing/data': 0.011650524102151394, 'timing/model': 1.5625739763490856, 'learning_rate/qwen_vl_interface': 6.621929190363585e-06, 'learning_rate/action_model': 6.621929190363585e-05, 'epoch': 0.85}) +[2026-06-16 02:33:46,760][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.17798395454883575, 'train/grad_norm_pre_clip': 11.013299942016602, 'timing/data': 0.01200598618015647, 'timing/model': 1.5495116650126874, 'learning_rate/qwen_vl_interface': 6.619013279207657e-06, 'learning_rate/action_model': 6.619013279207656e-05, 'epoch': 0.85}) +[2026-06-16 02:33:53,062][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.049769505858421326, 'train/grad_norm_pre_clip': 10.686570167541504, 'timing/data': 0.011045821942389011, 'timing/model': 1.5625927443616092, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 0.85}) +[2026-06-16 02:33:59,360][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.009486349299550056, 'train/grad_norm_pre_clip': 2.129746437072754, 'timing/data': 0.012050505727529526, 'timing/model': 1.5522178043611348, 'learning_rate/qwen_vl_interface': 6.6131797698457135e-06, 'learning_rate/action_model': 6.613179769845713e-05, 'epoch': 0.85}) +[2026-06-16 02:34:05,667][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.33431920409202576, 'train/grad_norm_pre_clip': 15.666882514953613, 'timing/data': 0.012715828139334917, 'timing/model': 1.5662144762463868, 'learning_rate/qwen_vl_interface': 6.610262174037633e-06, 'learning_rate/action_model': 6.610262174037633e-05, 'epoch': 0.85}) +[2026-06-16 02:34:12,034][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.06833487749099731, 'train/grad_norm_pre_clip': 4.149550914764404, 'timing/data': 0.011572734918445349, 'timing/model': 1.8342718863859773, 'learning_rate/qwen_vl_interface': 6.607344019077155e-06, 'learning_rate/action_model': 6.607344019077156e-05, 'epoch': 0.85}) +[2026-06-16 02:34:18,073][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.17199663817882538, 'train/grad_norm_pre_clip': 38.52848434448242, 'timing/data': 0.011324184015393257, 'timing/model': 1.570528110023588, 'learning_rate/qwen_vl_interface': 6.604425306163822e-06, 'learning_rate/action_model': 6.604425306163822e-05, 'epoch': 0.85}) +[2026-06-16 02:45:17,417][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.024365980178117752, 'train/grad_norm_pre_clip': 4.846207141876221, 'eval/loss': 0.04817753657698631, 'eval/flap_precision': 0.9400798934753661, 'eval/flap_recall': 0.9631650750341064, 'eval/flap_f1': 0.9514824797843666, 'eval/flap_rate_model': 0.2086111111111111, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9514824797843666, 'timing/data': 0.011200683191418648, 'timing/model': 1.563323681242764, 'learning_rate/qwen_vl_interface': 6.6015060364974045e-06, 'learning_rate/action_model': 6.601506036497404e-05, 'epoch': 0.85}) +[2026-06-16 02:45:17,418][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2100_state +[2026-06-16 02:45:17,424][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:45:25,969][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2100_state/model.safetensors +[2026-06-16 02:45:45,980][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2100_state/optimizer.bin +[2026-06-16 02:45:45,981][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2100_state/sampler.bin +[2026-06-16 02:45:45,981][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2100_state/sampler_1.bin +[2026-06-16 02:45:45,983][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2100_state/random_states_0.pkl +[2026-06-16 02:45:46,167][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2100 +[2026-06-16 02:45:46,167][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:45:46,182][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:45:52,125][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.05516887456178665, 'train/grad_norm_pre_clip': 7.8424506187438965, 'timing/data': 0.011151617392897606, 'timing/model': 1.4826018111780286, 'learning_rate/qwen_vl_interface': 6.598586211277907e-06, 'learning_rate/action_model': 6.598586211277907e-05, 'epoch': 0.85}) +[2026-06-16 02:45:58,332][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.0073086777701973915, 'train/grad_norm_pre_clip': 0.7233558893203735, 'timing/data': 0.010960804764181376, 'timing/model': 1.565922424197197, 'learning_rate/qwen_vl_interface': 6.59566583170556e-06, 'learning_rate/action_model': 6.59566583170556e-05, 'epoch': 0.85}) +[2026-06-16 02:46:04,536][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.16132275760173798, 'train/grad_norm_pre_clip': 11.676594734191895, 'timing/data': 0.010951746255159378, 'timing/model': 1.5526306191459298, 'learning_rate/qwen_vl_interface': 6.59274489898082e-06, 'learning_rate/action_model': 6.592744898980819e-05, 'epoch': 0.85}) +[2026-06-16 02:46:10,634][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.11362536251544952, 'train/grad_norm_pre_clip': 9.342548370361328, 'timing/data': 0.010967886075377464, 'timing/model': 1.564749431796372, 'learning_rate/qwen_vl_interface': 6.5898234143043716e-06, 'learning_rate/action_model': 6.589823414304371e-05, 'epoch': 0.85}) +[2026-06-16 02:46:17,032][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.15260343253612518, 'train/grad_norm_pre_clip': 9.17524528503418, 'timing/data': 0.011060746852308512, 'timing/model': 1.4870165684260428, 'learning_rate/qwen_vl_interface': 6.5869013788771264e-06, 'learning_rate/action_model': 6.586901378877126e-05, 'epoch': 0.85}) +[2026-06-16 02:46:23,244][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.05201531946659088, 'train/grad_norm_pre_clip': 9.113000869750977, 'timing/data': 0.01030789315700531, 'timing/model': 1.7885330556891859, 'learning_rate/qwen_vl_interface': 6.583978793900223e-06, 'learning_rate/action_model': 6.583978793900223e-05, 'epoch': 0.85}) +[2026-06-16 02:46:29,236][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.2225913554430008, 'train/grad_norm_pre_clip': 11.505139350891113, 'timing/data': 0.01051752408966422, 'timing/model': 1.566999128088355, 'learning_rate/qwen_vl_interface': 6.581055660575028e-06, 'learning_rate/action_model': 6.581055660575027e-05, 'epoch': 0.85}) +[2026-06-16 02:46:35,434][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.12653610110282898, 'train/grad_norm_pre_clip': 6.405491828918457, 'timing/data': 0.010500053875148296, 'timing/model': 1.5637443559244275, 'learning_rate/qwen_vl_interface': 6.578131980103128e-06, 'learning_rate/action_model': 6.578131980103128e-05, 'epoch': 0.85}) +[2026-06-16 02:46:41,539][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.09512711316347122, 'train/grad_norm_pre_clip': 14.068707466125488, 'timing/data': 0.011101706884801388, 'timing/model': 1.5605956548824906, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 0.85}) +[2026-06-16 02:46:47,775][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.01341207791119814, 'train/grad_norm_pre_clip': 2.1297976970672607, 'timing/data': 0.01163429906591773, 'timing/model': 1.592922033276409, 'learning_rate/qwen_vl_interface': 6.572282982526702e-06, 'learning_rate/action_model': 6.572282982526702e-05, 'epoch': 0.85}) +[2026-06-16 02:46:54,039][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.3155697286128998, 'train/grad_norm_pre_clip': 17.089149475097656, 'timing/data': 0.01085813483223319, 'timing/model': 1.5151027189567685, 'learning_rate/qwen_vl_interface': 6.5693576678264755e-06, 'learning_rate/action_model': 6.569357667826476e-05, 'epoch': 0.85}) +[2026-06-16 02:47:00,277][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.03838649019598961, 'train/grad_norm_pre_clip': 4.608141899108887, 'timing/data': 0.011195557191967964, 'timing/model': 1.7319077220745385, 'learning_rate/qwen_vl_interface': 6.566431810788148e-06, 'learning_rate/action_model': 6.566431810788148e-05, 'epoch': 0.85}) +[2026-06-16 02:47:06,447][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.00963091105222702, 'train/grad_norm_pre_clip': 1.4732304811477661, 'timing/data': 0.010763166006654501, 'timing/model': 1.5816539810039103, 'learning_rate/qwen_vl_interface': 6.5635054126144306e-06, 'learning_rate/action_model': 6.56350541261443e-05, 'epoch': 0.85}) +[2026-06-16 02:47:12,845][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.10402045398950577, 'train/grad_norm_pre_clip': 6.811679363250732, 'timing/data': 0.01234262203797698, 'timing/model': 1.557919223792851, 'learning_rate/qwen_vl_interface': 6.560578474508254e-06, 'learning_rate/action_model': 6.560578474508253e-05, 'epoch': 0.85}) +[2026-06-16 02:47:19,044][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.32569754123687744, 'train/grad_norm_pre_clip': 23.24863624572754, 'timing/data': 0.012049670796841383, 'timing/model': 1.5603188560344279, 'learning_rate/qwen_vl_interface': 6.5576509976727685e-06, 'learning_rate/action_model': 6.557650997672769e-05, 'epoch': 0.86}) +[2026-06-16 02:47:25,345][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.0473664291203022, 'train/grad_norm_pre_clip': 9.927987098693848, 'timing/data': 0.011740808840841055, 'timing/model': 1.556870514061302, 'learning_rate/qwen_vl_interface': 6.554722983311354e-06, 'learning_rate/action_model': 6.554722983311355e-05, 'epoch': 0.86}) +[2026-06-16 02:47:31,647][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.11286698281764984, 'train/grad_norm_pre_clip': 8.131741523742676, 'timing/data': 0.010456684045493603, 'timing/model': 1.522356039378792, 'learning_rate/qwen_vl_interface': 6.551794432627605e-06, 'learning_rate/action_model': 6.551794432627605e-05, 'epoch': 0.86}) +[2026-06-16 02:47:38,028][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.014284003525972366, 'train/grad_norm_pre_clip': 2.3026175498962402, 'timing/data': 0.011544599197804928, 'timing/model': 1.8434318527579308, 'learning_rate/qwen_vl_interface': 6.548865346825338e-06, 'learning_rate/action_model': 6.548865346825338e-05, 'epoch': 0.86}) +[2026-06-16 02:47:44,289][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.10304144024848938, 'train/grad_norm_pre_clip': 11.29178524017334, 'timing/data': 0.010519654024392366, 'timing/model': 1.8017696510069072, 'learning_rate/qwen_vl_interface': 6.545935727108589e-06, 'learning_rate/action_model': 6.545935727108588e-05, 'epoch': 0.86}) +[2026-06-16 02:47:50,436][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.00579623319208622, 'train/grad_norm_pre_clip': 0.6833482384681702, 'timing/data': 0.011197797022759914, 'timing/model': 1.6417661285959184, 'learning_rate/qwen_vl_interface': 6.543005574681613e-06, 'learning_rate/action_model': 6.543005574681613e-05, 'epoch': 0.86}) +[2026-06-16 02:47:56,738][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.006381798069924116, 'train/grad_norm_pre_clip': 0.6887688636779785, 'timing/data': 0.01141843805089593, 'timing/model': 1.581463186070323, 'learning_rate/qwen_vl_interface': 6.5400748907488886e-06, 'learning_rate/action_model': 6.540074890748888e-05, 'epoch': 0.86}) +[2026-06-16 02:48:03,058][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.2706140875816345, 'train/grad_norm_pre_clip': 22.059263229370117, 'timing/data': 0.011357698123902082, 'timing/model': 1.5639747739769518, 'learning_rate/qwen_vl_interface': 6.537143676515105e-06, 'learning_rate/action_model': 6.537143676515105e-05, 'epoch': 0.86}) +[2026-06-16 02:48:09,358][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.056816913187503815, 'train/grad_norm_pre_clip': 8.264579772949219, 'timing/data': 0.011014395859092474, 'timing/model': 1.5266524692997336, 'learning_rate/qwen_vl_interface': 6.534211933185176e-06, 'learning_rate/action_model': 6.534211933185176e-05, 'epoch': 0.86}) +[2026-06-16 02:48:15,693][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.1077425479888916, 'train/grad_norm_pre_clip': 13.091240882873535, 'timing/data': 0.010519484989345074, 'timing/model': 1.7993858801200986, 'learning_rate/qwen_vl_interface': 6.531279661964231e-06, 'learning_rate/action_model': 6.531279661964231e-05, 'epoch': 0.86}) +[2026-06-16 02:48:21,761][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.07516638189554214, 'train/grad_norm_pre_clip': 11.067535400390625, 'timing/data': 0.011117257177829742, 'timing/model': 1.5681422459892929, 'learning_rate/qwen_vl_interface': 6.528346864057616e-06, 'learning_rate/action_model': 6.528346864057615e-05, 'epoch': 0.86}) +[2026-06-16 02:48:28,168][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.01883021928369999, 'train/grad_norm_pre_clip': 4.673465251922607, 'timing/data': 0.01127439783886075, 'timing/model': 1.5695874537341297, 'learning_rate/qwen_vl_interface': 6.5254135406708906e-06, 'learning_rate/action_model': 6.52541354067089e-05, 'epoch': 0.86}) +[2026-06-16 02:48:34,469][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.30791792273521423, 'train/grad_norm_pre_clip': 22.29026985168457, 'timing/data': 0.010488905012607574, 'timing/model': 1.5708925910294056, 'learning_rate/qwen_vl_interface': 6.522479693009836e-06, 'learning_rate/action_model': 6.522479693009837e-05, 'epoch': 0.86}) +[2026-06-16 02:48:40,759][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.3608359396457672, 'train/grad_norm_pre_clip': 19.55446434020996, 'timing/data': 0.010678354650735855, 'timing/model': 1.5247846189886332, 'learning_rate/qwen_vl_interface': 6.519545322280446e-06, 'learning_rate/action_model': 6.519545322280446e-05, 'epoch': 0.86}) +[2026-06-16 02:48:47,053][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.06642556190490723, 'train/grad_norm_pre_clip': 9.022651672363281, 'timing/data': 0.011479248758405447, 'timing/model': 1.5897981049492955, 'learning_rate/qwen_vl_interface': 6.5166104296889275e-06, 'learning_rate/action_model': 6.516610429688927e-05, 'epoch': 0.86}) +[2026-06-16 02:48:53,349][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.027775641530752182, 'train/grad_norm_pre_clip': 6.336714267730713, 'timing/data': 0.010876016225665808, 'timing/model': 1.7542949807830155, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.86}) +[2026-06-16 02:48:59,460][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.08577211946249008, 'train/grad_norm_pre_clip': 24.683612823486328, 'timing/data': 0.010813705157488585, 'timing/model': 1.5635657077655196, 'learning_rate/qwen_vl_interface': 6.510739083745422e-06, 'learning_rate/action_model': 6.510739083745422e-05, 'epoch': 0.86}) +[2026-06-16 02:49:05,854][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.14695288240909576, 'train/grad_norm_pre_clip': 8.963828086853027, 'timing/data': 0.010377563070505857, 'timing/model': 1.5937786167487502, 'learning_rate/qwen_vl_interface': 6.507802632806919e-06, 'learning_rate/action_model': 6.507802632806919e-05, 'epoch': 0.86}) +[2026-06-16 02:49:12,158][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.06603670865297318, 'train/grad_norm_pre_clip': 5.140359878540039, 'timing/data': 0.010933686047792435, 'timing/model': 1.5676782173104584, 'learning_rate/qwen_vl_interface': 6.5048656648332646e-06, 'learning_rate/action_model': 6.504865664833264e-05, 'epoch': 0.86}) +[2026-06-16 02:49:18,456][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.05110983923077583, 'train/grad_norm_pre_clip': 4.924184799194336, 'timing/data': 0.01157256867736578, 'timing/model': 1.5217547649517655, 'learning_rate/qwen_vl_interface': 6.501928181031737e-06, 'learning_rate/action_model': 6.501928181031737e-05, 'epoch': 0.86}) +[2026-06-16 02:49:24,839][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.023466676473617554, 'train/grad_norm_pre_clip': 2.824982166290283, 'timing/data': 0.011168176773935556, 'timing/model': 1.5813998491503298, 'learning_rate/qwen_vl_interface': 6.498990182609823e-06, 'learning_rate/action_model': 6.498990182609823e-05, 'epoch': 0.86}) +[2026-06-16 02:49:31,077][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.044909946620464325, 'train/grad_norm_pre_clip': 8.658843040466309, 'timing/data': 0.011150356847792864, 'timing/model': 1.7772350879386067, 'learning_rate/qwen_vl_interface': 6.4960516707752255e-06, 'learning_rate/action_model': 6.496051670775226e-05, 'epoch': 0.86}) +[2026-06-16 02:49:37,158][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.05355196073651314, 'train/grad_norm_pre_clip': 8.23067569732666, 'timing/data': 0.010839395690709352, 'timing/model': 1.5564549257978797, 'learning_rate/qwen_vl_interface': 6.493112646735851e-06, 'learning_rate/action_model': 6.493112646735851e-05, 'epoch': 0.86}) +[2026-06-16 02:49:43,461][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.2349970042705536, 'train/grad_norm_pre_clip': 13.736270904541016, 'timing/data': 0.011387497652322054, 'timing/model': 1.5671934131532907, 'learning_rate/qwen_vl_interface': 6.490173111699824e-06, 'learning_rate/action_model': 6.490173111699825e-05, 'epoch': 0.86}) +[2026-06-16 02:49:49,836][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.03924621269106865, 'train/grad_norm_pre_clip': 8.30911922454834, 'timing/data': 0.011022467166185379, 'timing/model': 1.6326711010187864, 'learning_rate/qwen_vl_interface': 6.487233066875479e-06, 'learning_rate/action_model': 6.487233066875479e-05, 'epoch': 0.86}) +[2026-06-16 02:49:56,164][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.1134391650557518, 'train/grad_norm_pre_clip': 7.408069610595703, 'timing/data': 0.011104316916316748, 'timing/model': 1.5417487379163504, 'learning_rate/qwen_vl_interface': 6.484292513471352e-06, 'learning_rate/action_model': 6.484292513471352e-05, 'epoch': 0.87}) +[2026-06-16 02:50:02,491][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.0848410502076149, 'train/grad_norm_pre_clip': 9.256566047668457, 'timing/data': 0.011029046960175037, 'timing/model': 1.79540073312819, 'learning_rate/qwen_vl_interface': 6.4813514526961975e-06, 'learning_rate/action_model': 6.481351452696198e-05, 'epoch': 0.87}) +[2026-06-16 02:50:08,553][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.13681621849536896, 'train/grad_norm_pre_clip': 11.362910270690918, 'timing/data': 0.01099357707425952, 'timing/model': 1.5555775286629796, 'learning_rate/qwen_vl_interface': 6.478409885758972e-06, 'learning_rate/action_model': 6.478409885758971e-05, 'epoch': 0.87}) +[2026-06-16 02:50:14,858][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.2853081226348877, 'train/grad_norm_pre_clip': 9.442957878112793, 'timing/data': 0.010997076984494925, 'timing/model': 1.5507194697856903, 'learning_rate/qwen_vl_interface': 6.475467813868844e-06, 'learning_rate/action_model': 6.475467813868844e-05, 'epoch': 0.87}) +[2026-06-16 02:50:21,076][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.4652998447418213, 'train/grad_norm_pre_clip': 19.061737060546875, 'timing/data': 0.011157056782394648, 'timing/model': 1.5347436042502522, 'learning_rate/qwen_vl_interface': 6.472525238235187e-06, 'learning_rate/action_model': 6.472525238235186e-05, 'epoch': 0.87}) +[2026-06-16 02:50:27,268][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.016676636412739754, 'train/grad_norm_pre_clip': 2.6895804405212402, 'timing/data': 0.011166267096996307, 'timing/model': 1.5456807920709252, 'learning_rate/qwen_vl_interface': 6.4695821600675834e-06, 'learning_rate/action_model': 6.469582160067583e-05, 'epoch': 0.87}) +[2026-06-16 02:50:33,568][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.11710678786039352, 'train/grad_norm_pre_clip': 9.03366756439209, 'timing/data': 0.010764464735984802, 'timing/model': 1.5647215638309717, 'learning_rate/qwen_vl_interface': 6.466638580575824e-06, 'learning_rate/action_model': 6.466638580575824e-05, 'epoch': 0.87}) +[2026-06-16 02:50:39,994][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.051767390221357346, 'train/grad_norm_pre_clip': 11.725323677062988, 'timing/data': 0.010777626186609268, 'timing/model': 1.7991838078014553, 'learning_rate/qwen_vl_interface': 6.463694500969898e-06, 'learning_rate/action_model': 6.463694500969898e-05, 'epoch': 0.87}) +[2026-06-16 02:50:46,062][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.08077608793973923, 'train/grad_norm_pre_clip': 7.0415496826171875, 'timing/data': 0.010858345776796341, 'timing/model': 1.5304224700666964, 'learning_rate/qwen_vl_interface': 6.460749922460008e-06, 'learning_rate/action_model': 6.460749922460008e-05, 'epoch': 0.87}) +[2026-06-16 02:50:52,385][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.04947737604379654, 'train/grad_norm_pre_clip': 8.376617431640625, 'timing/data': 0.01478490186855197, 'timing/model': 1.5557557400316, 'learning_rate/qwen_vl_interface': 6.457804846256559e-06, 'learning_rate/action_model': 6.457804846256559e-05, 'epoch': 0.87}) +[2026-06-16 02:50:58,791][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.033810533583164215, 'train/grad_norm_pre_clip': 2.9454445838928223, 'timing/data': 0.011205767747014761, 'timing/model': 1.5944754551164806, 'learning_rate/qwen_vl_interface': 6.45485927357016e-06, 'learning_rate/action_model': 6.454859273570159e-05, 'epoch': 0.87}) +[2026-06-16 02:51:05,064][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.06524501740932465, 'train/grad_norm_pre_clip': 5.755706310272217, 'timing/data': 0.011089406907558441, 'timing/model': 1.5684211659245193, 'learning_rate/qwen_vl_interface': 6.451913205611626e-06, 'learning_rate/action_model': 6.451913205611626e-05, 'epoch': 0.87}) +[2026-06-16 02:51:11,368][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.08227416127920151, 'train/grad_norm_pre_clip': 9.990376472473145, 'timing/data': 0.010894356295466423, 'timing/model': 1.5819887341931462, 'learning_rate/qwen_vl_interface': 6.448966643591974e-06, 'learning_rate/action_model': 6.448966643591974e-05, 'epoch': 0.87}) +[2026-06-16 02:51:17,712][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.022076424211263657, 'train/grad_norm_pre_clip': 3.7838528156280518, 'timing/data': 0.011238717939704657, 'timing/model': 1.814242370892316, 'learning_rate/qwen_vl_interface': 6.446019588722423e-06, 'learning_rate/action_model': 6.446019588722422e-05, 'epoch': 0.87}) +[2026-06-16 02:51:23,876][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.07586050033569336, 'train/grad_norm_pre_clip': 10.456632614135742, 'timing/data': 0.010292013641446829, 'timing/model': 1.5884874039329588, 'learning_rate/qwen_vl_interface': 6.443072042214398e-06, 'learning_rate/action_model': 6.443072042214397e-05, 'epoch': 0.87}) +[2026-06-16 02:51:30,154][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.12455692142248154, 'train/grad_norm_pre_clip': 20.852293014526367, 'timing/data': 0.01128791831433773, 'timing/model': 1.5631560976617038, 'learning_rate/qwen_vl_interface': 6.440124005279521e-06, 'learning_rate/action_model': 6.440124005279521e-05, 'epoch': 0.87}) +[2026-06-16 02:51:36,454][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.1017504632472992, 'train/grad_norm_pre_clip': 6.452081203460693, 'timing/data': 0.011044536717236042, 'timing/model': 1.5560045489110053, 'learning_rate/qwen_vl_interface': 6.43717547912962e-06, 'learning_rate/action_model': 6.437175479129621e-05, 'epoch': 0.87}) +[2026-06-16 02:51:42,756][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.3239084780216217, 'train/grad_norm_pre_clip': 19.671546936035156, 'timing/data': 0.010566334705799818, 'timing/model': 1.5401773620396852, 'learning_rate/qwen_vl_interface': 6.434226464976725e-06, 'learning_rate/action_model': 6.434226464976724e-05, 'epoch': 0.87}) +[2026-06-16 02:51:49,077][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.2115476429462433, 'train/grad_norm_pre_clip': 13.834836959838867, 'timing/data': 0.01073267636820674, 'timing/model': 1.5711469850502908, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 0.87}) +[2026-06-16 02:51:55,407][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.018943769857287407, 'train/grad_norm_pre_clip': 1.9384524822235107, 'timing/data': 0.012213381938636303, 'timing/model': 1.8054034020751715, 'learning_rate/qwen_vl_interface': 6.428326977511059e-06, 'learning_rate/action_model': 6.428326977511058e-05, 'epoch': 0.87}) +[2026-06-16 02:52:01,460][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.07455816119909286, 'train/grad_norm_pre_clip': 8.345609664916992, 'timing/data': 0.011085457168519497, 'timing/model': 1.5597354592755437, 'learning_rate/qwen_vl_interface': 6.425376506623344e-06, 'learning_rate/action_model': 6.425376506623344e-05, 'epoch': 0.87}) +[2026-06-16 02:52:07,761][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.07523036748170853, 'train/grad_norm_pre_clip': 7.856550216674805, 'timing/data': 0.010968186892569065, 'timing/model': 1.5360868191346526, 'learning_rate/qwen_vl_interface': 6.422425552582747e-06, 'learning_rate/action_model': 6.422425552582747e-05, 'epoch': 0.87}) +[2026-06-16 02:52:14,073][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.02330215461552143, 'train/grad_norm_pre_clip': 2.74775767326355, 'timing/data': 0.0111180879175663, 'timing/model': 1.5735723790712655, 'learning_rate/qwen_vl_interface': 6.41947411660229e-06, 'learning_rate/action_model': 6.41947411660229e-05, 'epoch': 0.87}) +[2026-06-16 02:52:20,365][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.012429818511009216, 'train/grad_norm_pre_clip': 1.6219443082809448, 'timing/data': 0.011109618004411459, 'timing/model': 1.5580909652635455, 'learning_rate/qwen_vl_interface': 6.4165221998951995e-06, 'learning_rate/action_model': 6.4165221998952e-05, 'epoch': 0.87}) +[2026-06-16 02:52:26,659][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.02869396284222603, 'train/grad_norm_pre_clip': 3.8483145236968994, 'timing/data': 0.01110698701813817, 'timing/model': 1.527638356667012, 'learning_rate/qwen_vl_interface': 6.413569803674896e-06, 'learning_rate/action_model': 6.413569803674896e-05, 'epoch': 0.88}) +[2026-06-16 02:52:32,989][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.18687357008457184, 'train/grad_norm_pre_clip': 15.581145286560059, 'timing/data': 0.011019747238606215, 'timing/model': 1.7846741206012666, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 0.88}) +[2026-06-16 02:52:39,065][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.12112462520599365, 'train/grad_norm_pre_clip': 15.245186805725098, 'timing/data': 0.010792616289108992, 'timing/model': 1.5625920370221138, 'learning_rate/qwen_vl_interface': 6.40766357754932e-06, 'learning_rate/action_model': 6.407663577549319e-05, 'epoch': 0.88}) +[2026-06-16 02:52:45,364][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.30527263879776, 'train/grad_norm_pre_clip': 12.572779655456543, 'timing/data': 0.010600795038044453, 'timing/model': 1.5609428123570979, 'learning_rate/qwen_vl_interface': 6.404709750071875e-06, 'learning_rate/action_model': 6.404709750071874e-05, 'epoch': 0.88}) +[2026-06-16 02:52:51,663][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.17247174680233002, 'train/grad_norm_pre_clip': 12.904317855834961, 'timing/data': 0.01075707608833909, 'timing/model': 1.555777040310204, 'learning_rate/qwen_vl_interface': 6.401755447936869e-06, 'learning_rate/action_model': 6.401755447936869e-05, 'epoch': 0.88}) +[2026-06-16 02:52:57,962][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.2056393176317215, 'train/grad_norm_pre_clip': 13.075437545776367, 'timing/data': 0.01050158403813839, 'timing/model': 1.5588531349785626, 'learning_rate/qwen_vl_interface': 6.3988006723587036e-06, 'learning_rate/action_model': 6.398800672358703e-05, 'epoch': 0.88}) +[2026-06-16 02:53:04,299][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.09564671665430069, 'train/grad_norm_pre_clip': 12.919693946838379, 'timing/data': 0.010642414912581444, 'timing/model': 1.546616262756288, 'learning_rate/qwen_vl_interface': 6.395845424551975e-06, 'learning_rate/action_model': 6.395845424551975e-05, 'epoch': 0.88}) +[2026-06-16 02:53:10,598][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.012365884147584438, 'train/grad_norm_pre_clip': 1.7494977712631226, 'timing/data': 0.012682774104177952, 'timing/model': 1.7903516720980406, 'learning_rate/qwen_vl_interface': 6.392889705731475e-06, 'learning_rate/action_model': 6.392889705731476e-05, 'epoch': 0.88}) +[2026-06-16 02:53:16,665][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.0474688820540905, 'train/grad_norm_pre_clip': 7.8354597091674805, 'timing/data': 0.011501369066536427, 'timing/model': 1.5644246996380389, 'learning_rate/qwen_vl_interface': 6.389933517112188e-06, 'learning_rate/action_model': 6.389933517112187e-05, 'epoch': 0.88}) +[2026-06-16 02:53:22,961][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.11355803906917572, 'train/grad_norm_pre_clip': 15.230213165283203, 'timing/data': 0.01154686976224184, 'timing/model': 1.5595217109657824, 'learning_rate/qwen_vl_interface': 6.386976859909293e-06, 'learning_rate/action_model': 6.386976859909292e-05, 'epoch': 0.88}) +[2026-06-16 02:53:29,482][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.02371051348745823, 'train/grad_norm_pre_clip': 2.0788204669952393, 'timing/data': 0.011519568972289562, 'timing/model': 1.6045815940015018, 'learning_rate/qwen_vl_interface': 6.384019735338156e-06, 'learning_rate/action_model': 6.384019735338156e-05, 'epoch': 0.88}) +[2026-06-16 02:53:35,766][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.008049680851399899, 'train/grad_norm_pre_clip': 0.8813595771789551, 'timing/data': 0.011277478188276291, 'timing/model': 1.5588332801125944, 'learning_rate/qwen_vl_interface': 6.381062144614342e-06, 'learning_rate/action_model': 6.381062144614342e-05, 'epoch': 0.88}) +[2026-06-16 02:53:42,067][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.12196458131074905, 'train/grad_norm_pre_clip': 16.39488983154297, 'timing/data': 0.011878510005772114, 'timing/model': 1.5273001668974757, 'learning_rate/qwen_vl_interface': 6.378104088953605e-06, 'learning_rate/action_model': 6.378104088953605e-05, 'epoch': 0.88}) +[2026-06-16 02:53:48,416][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.18105731904506683, 'train/grad_norm_pre_clip': 16.333911895751953, 'timing/data': 0.010822285898029804, 'timing/model': 1.807788223028183, 'learning_rate/qwen_vl_interface': 6.375145569571889e-06, 'learning_rate/action_model': 6.375145569571889e-05, 'epoch': 0.88}) +[2026-06-16 02:53:54,466][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.014066222123801708, 'train/grad_norm_pre_clip': 1.8686124086380005, 'timing/data': 0.011124256998300552, 'timing/model': 1.5608855611644685, 'learning_rate/qwen_vl_interface': 6.372186587685329e-06, 'learning_rate/action_model': 6.372186587685329e-05, 'epoch': 0.88}) +[2026-06-16 02:54:00,862][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.14050275087356567, 'train/grad_norm_pre_clip': 6.970980644226074, 'timing/data': 0.01102025667205453, 'timing/model': 1.5559281911700964, 'learning_rate/qwen_vl_interface': 6.369227144510252e-06, 'learning_rate/action_model': 6.369227144510252e-05, 'epoch': 0.88}) +[2026-06-16 02:54:07,164][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.08360129594802856, 'train/grad_norm_pre_clip': 17.160076141357422, 'timing/data': 0.01094258576631546, 'timing/model': 1.5675562722608447, 'learning_rate/qwen_vl_interface': 6.3662672412631736e-06, 'learning_rate/action_model': 6.366267241263173e-05, 'epoch': 0.88}) +[2026-06-16 02:54:13,464][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.0010448332177475095, 'train/grad_norm_pre_clip': 0.11952558159828186, 'timing/data': 0.011111167259514332, 'timing/model': 1.5596097679808736, 'learning_rate/qwen_vl_interface': 6.3633068791607954e-06, 'learning_rate/action_model': 6.363306879160795e-05, 'epoch': 0.88}) +[2026-06-16 02:54:19,865][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.1525893211364746, 'train/grad_norm_pre_clip': 14.31653881072998, 'timing/data': 0.01224435307085514, 'timing/model': 1.6099185729399323, 'learning_rate/qwen_vl_interface': 6.360346059420014e-06, 'learning_rate/action_model': 6.360346059420014e-05, 'epoch': 0.88}) +[2026-06-16 02:54:26,190][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.056330833584070206, 'train/grad_norm_pre_clip': 9.687468528747559, 'timing/data': 0.011245088186115026, 'timing/model': 1.7839680570177734, 'learning_rate/qwen_vl_interface': 6.357384783257909e-06, 'learning_rate/action_model': 6.357384783257908e-05, 'epoch': 0.88}) +[2026-06-16 02:54:32,528][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.1229449138045311, 'train/grad_norm_pre_clip': 17.00527000427246, 'timing/data': 0.013065726030617952, 'timing/model': 1.7268173028714955, 'learning_rate/qwen_vl_interface': 6.354423051891748e-06, 'learning_rate/action_model': 6.354423051891749e-05, 'epoch': 0.88}) +[2026-06-16 02:54:38,565][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.08988343179225922, 'train/grad_norm_pre_clip': 6.209024429321289, 'timing/data': 0.011423207819461823, 'timing/model': 1.5667828628793359, 'learning_rate/qwen_vl_interface': 6.35146086653899e-06, 'learning_rate/action_model': 6.35146086653899e-05, 'epoch': 0.88}) +[2026-06-16 02:54:44,867][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.10848227143287659, 'train/grad_norm_pre_clip': 11.220549583435059, 'timing/data': 0.011486058589071035, 'timing/model': 1.566181710921228, 'learning_rate/qwen_vl_interface': 6.348498228417273e-06, 'learning_rate/action_model': 6.348498228417273e-05, 'epoch': 0.88}) +[2026-06-16 02:54:51,160][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.054181162267923355, 'train/grad_norm_pre_clip': 10.648906707763672, 'timing/data': 0.011339758988469839, 'timing/model': 1.5579391759820282, 'learning_rate/qwen_vl_interface': 6.34553513874443e-06, 'learning_rate/action_model': 6.34553513874443e-05, 'epoch': 0.88}) +[2026-06-16 02:54:57,479][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.002121484372764826, 'train/grad_norm_pre_clip': 0.3663170635700226, 'timing/data': 0.011479668784886599, 'timing/model': 1.5312664830125868, 'learning_rate/qwen_vl_interface': 6.342571598738473e-06, 'learning_rate/action_model': 6.342571598738472e-05, 'epoch': 0.88}) +[2026-06-16 02:55:03,802][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.3523982763290405, 'train/grad_norm_pre_clip': 25.300138473510742, 'timing/data': 0.01146435784175992, 'timing/model': 1.8023370113223791, 'learning_rate/qwen_vl_interface': 6.339607609617603e-06, 'learning_rate/action_model': 6.339607609617603e-05, 'epoch': 0.89}) +[2026-06-16 02:55:10,125][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.029131174087524414, 'train/grad_norm_pre_clip': 9.043044090270996, 'timing/data': 0.010957716964185238, 'timing/model': 1.7925371890887618, 'learning_rate/qwen_vl_interface': 6.336643172600202e-06, 'learning_rate/action_model': 6.336643172600202e-05, 'epoch': 0.89}) +[2026-06-16 02:55:16,173][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.02440042980015278, 'train/grad_norm_pre_clip': 6.642814636230469, 'timing/data': 0.012810664717108011, 'timing/model': 1.5621054070070386, 'learning_rate/qwen_vl_interface': 6.333678288904838e-06, 'learning_rate/action_model': 6.333678288904837e-05, 'epoch': 0.89}) +[2026-06-16 02:55:22,475][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.0401197224855423, 'train/grad_norm_pre_clip': 10.822482109069824, 'timing/data': 0.010549844708293676, 'timing/model': 1.5736196767538786, 'learning_rate/qwen_vl_interface': 6.330712959750264e-06, 'learning_rate/action_model': 6.330712959750263e-05, 'epoch': 0.89}) +[2026-06-16 02:55:28,876][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.08864790946245193, 'train/grad_norm_pre_clip': 8.712531089782715, 'timing/data': 0.010744736064225435, 'timing/model': 1.5664094872772694, 'learning_rate/qwen_vl_interface': 6.3277471863554145e-06, 'learning_rate/action_model': 6.327747186355414e-05, 'epoch': 0.89}) +[2026-06-16 02:55:35,233][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.04049842804670334, 'train/grad_norm_pre_clip': 26.92698097229004, 'timing/data': 0.012785424012690783, 'timing/model': 1.5745462528429925, 'learning_rate/qwen_vl_interface': 6.324780969939406e-06, 'learning_rate/action_model': 6.324780969939405e-05, 'epoch': 0.89}) +[2026-06-16 02:55:41,503][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.006290843710303307, 'train/grad_norm_pre_clip': 0.7044495344161987, 'timing/data': 0.010438524652272463, 'timing/model': 1.7810280141420662, 'learning_rate/qwen_vl_interface': 6.32181431172154e-06, 'learning_rate/action_model': 6.32181431172154e-05, 'epoch': 0.89}) +[2026-06-16 02:55:47,570][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.0034605946857482195, 'train/grad_norm_pre_clip': 1.1848433017730713, 'timing/data': 0.01152513874694705, 'timing/model': 1.5655501559376717, 'learning_rate/qwen_vl_interface': 6.318847212921295e-06, 'learning_rate/action_model': 6.318847212921295e-05, 'epoch': 0.89}) +[2026-06-16 02:55:53,865][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.0909765437245369, 'train/grad_norm_pre_clip': 16.842697143554688, 'timing/data': 0.011281328275799751, 'timing/model': 1.5651248139329255, 'learning_rate/qwen_vl_interface': 6.3158796747583364e-06, 'learning_rate/action_model': 6.315879674758336e-05, 'epoch': 0.89}) +[2026-06-16 02:56:00,264][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.014754114672541618, 'train/grad_norm_pre_clip': 2.300062417984009, 'timing/data': 0.01191697083413601, 'timing/model': 1.5678127161227167, 'learning_rate/qwen_vl_interface': 6.312911698452503e-06, 'learning_rate/action_model': 6.312911698452502e-05, 'epoch': 0.89}) +[2026-06-16 02:56:06,744][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.05297417938709259, 'train/grad_norm_pre_clip': 13.891680717468262, 'timing/data': 0.011222228407859802, 'timing/model': 1.6172335776500404, 'learning_rate/qwen_vl_interface': 6.309943285223824e-06, 'learning_rate/action_model': 6.309943285223824e-05, 'epoch': 0.89}) +[2026-06-16 03:06:57,812][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.03682216256856918, 'train/grad_norm_pre_clip': 12.017436027526855, 'eval/loss': 0.04921267554163933, 'eval/flap_precision': 0.9576502732240437, 'eval/flap_recall': 0.956343792633015, 'eval/flap_f1': 0.9569965870307168, 'eval/flap_rate_model': 0.20333333333333334, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9569965870307169, 'timing/data': 0.011931790970265865, 'timing/model': 1.5071421093307436, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.89}) +[2026-06-16 03:06:57,813][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2200_state +[2026-06-16 03:06:57,820][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 03:07:06,262][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2200_state/model.safetensors +[2026-06-16 03:07:25,737][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2200_state/optimizer.bin +[2026-06-16 03:07:25,738][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2200_state/sampler.bin +[2026-06-16 03:07:25,738][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2200_state/sampler_1.bin +[2026-06-16 03:07:25,739][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2200_state/random_states_0.pkl +[2026-06-16 03:07:25,930][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2200 +[2026-06-16 03:07:25,930][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 03:07:25,945][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 03:07:32,120][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.08359457552433014, 'train/grad_norm_pre_clip': 24.18324089050293, 'timing/data': 0.010240933392196894, 'timing/model': 1.509588435292244, 'learning_rate/qwen_vl_interface': 6.304005152878902e-06, 'learning_rate/action_model': 6.304005152878901e-05, 'epoch': 0.89}) +[2026-06-16 03:07:38,363][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.005081314593553543, 'train/grad_norm_pre_clip': 0.93549644947052, 'timing/data': 0.011397491209208965, 'timing/model': 1.8010656698606908, 'learning_rate/qwen_vl_interface': 6.301035436203602e-06, 'learning_rate/action_model': 6.301035436203602e-05, 'epoch': 0.89}) +[2026-06-16 03:07:44,623][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.13830265402793884, 'train/grad_norm_pre_clip': 39.15325164794922, 'timing/data': 0.011394931003451347, 'timing/model': 1.7599323098547757, 'learning_rate/qwen_vl_interface': 6.298065287487337e-06, 'learning_rate/action_model': 6.298065287487337e-05, 'epoch': 0.89}) +[2026-06-16 03:07:50,630][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.2765059173107147, 'train/grad_norm_pre_clip': 25.492877960205078, 'timing/data': 0.012191786896437407, 'timing/model': 1.5652294717729092, 'learning_rate/qwen_vl_interface': 6.295094707951019e-06, 'learning_rate/action_model': 6.295094707951019e-05, 'epoch': 0.89}) +[2026-06-16 03:07:56,833][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.010609044693410397, 'train/grad_norm_pre_clip': 4.808630466461182, 'timing/data': 0.012192105874419212, 'timing/model': 1.4954016478732228, 'learning_rate/qwen_vl_interface': 6.292123698815745e-06, 'learning_rate/action_model': 6.292123698815745e-05, 'epoch': 0.89}) +[2026-06-16 03:08:03,044][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.09115226566791534, 'train/grad_norm_pre_clip': 8.689931869506836, 'timing/data': 0.01147451065480709, 'timing/model': 1.5610775207169354, 'learning_rate/qwen_vl_interface': 6.289152261302783e-06, 'learning_rate/action_model': 6.289152261302784e-05, 'epoch': 0.89}) +[2026-06-16 03:08:09,351][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.08229663223028183, 'train/grad_norm_pre_clip': 5.747422218322754, 'timing/data': 0.010763055179268122, 'timing/model': 1.5511442827992141, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 0.89}) +[2026-06-16 03:08:15,641][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.3495570421218872, 'train/grad_norm_pre_clip': 17.669837951660156, 'timing/data': 0.011387988924980164, 'timing/model': 1.5129481349140406, 'learning_rate/qwen_vl_interface': 6.283208106029753e-06, 'learning_rate/action_model': 6.283208106029752e-05, 'epoch': 0.89}) +[2026-06-16 03:08:21,935][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.07416237145662308, 'train/grad_norm_pre_clip': 12.148270606994629, 'timing/data': 0.011056126095354557, 'timing/model': 1.7727883821353316, 'learning_rate/qwen_vl_interface': 6.280235390713102e-06, 'learning_rate/action_model': 6.280235390713102e-05, 'epoch': 0.89}) +[2026-06-16 03:08:28,085][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.04125974327325821, 'train/grad_norm_pre_clip': 22.323415756225586, 'timing/data': 0.011190507095307112, 'timing/model': 1.6861120839603245, 'learning_rate/qwen_vl_interface': 6.277262251905597e-06, 'learning_rate/action_model': 6.277262251905597e-05, 'epoch': 0.89}) +[2026-06-16 03:08:34,432][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.0013109204592183232, 'train/grad_norm_pre_clip': 0.24973006546497345, 'timing/data': 0.011218927334994078, 'timing/model': 1.658016124740243, 'learning_rate/qwen_vl_interface': 6.274288690829385e-06, 'learning_rate/action_model': 6.274288690829384e-05, 'epoch': 0.89}) +[2026-06-16 03:08:40,792][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.08457162231206894, 'train/grad_norm_pre_clip': 19.54523468017578, 'timing/data': 0.012125573586672544, 'timing/model': 1.600249899085611, 'learning_rate/qwen_vl_interface': 6.271314708706781e-06, 'learning_rate/action_model': 6.271314708706782e-05, 'epoch': 0.89}) +[2026-06-16 03:08:47,050][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.5075016617774963, 'train/grad_norm_pre_clip': 59.89545822143555, 'timing/data': 0.011620969045907259, 'timing/model': 1.506046556867659, 'learning_rate/qwen_vl_interface': 6.26834030676028e-06, 'learning_rate/action_model': 6.26834030676028e-05, 'epoch': 0.89}) +[2026-06-16 03:08:53,386][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.013432007282972336, 'train/grad_norm_pre_clip': 3.7206175327301025, 'timing/data': 0.011101584881544113, 'timing/model': 1.7887983829714358, 'learning_rate/qwen_vl_interface': 6.265365486212545e-06, 'learning_rate/action_model': 6.265365486212545e-05, 'epoch': 0.9}) +[2026-06-16 03:08:59,453][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.05212298780679703, 'train/grad_norm_pre_clip': 15.713859558105469, 'timing/data': 0.012005981989204884, 'timing/model': 1.5627460647374392, 'learning_rate/qwen_vl_interface': 6.26239024828641e-06, 'learning_rate/action_model': 6.26239024828641e-05, 'epoch': 0.9}) +[2026-06-16 03:09:05,753][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.021887777373194695, 'train/grad_norm_pre_clip': 7.067666053771973, 'timing/data': 0.010909863281995058, 'timing/model': 1.5642795329913497, 'learning_rate/qwen_vl_interface': 6.259414594204888e-06, 'learning_rate/action_model': 6.259414594204888e-05, 'epoch': 0.9}) +[2026-06-16 03:09:12,169][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.11146166175603867, 'train/grad_norm_pre_clip': 14.37696647644043, 'timing/data': 0.01076514134183526, 'timing/model': 1.651562862098217, 'learning_rate/qwen_vl_interface': 6.256438525191155e-06, 'learning_rate/action_model': 6.256438525191156e-05, 'epoch': 0.9}) +[2026-06-16 03:09:18,557][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.02882390096783638, 'train/grad_norm_pre_clip': 8.976243019104004, 'timing/data': 0.011860259808599949, 'timing/model': 1.63446383504197, 'learning_rate/qwen_vl_interface': 6.253462042468561e-06, 'learning_rate/action_model': 6.25346204246856e-05, 'epoch': 0.9}) +[2026-06-16 03:09:24,849][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.1695631891489029, 'train/grad_norm_pre_clip': 8.492792129516602, 'timing/data': 0.01162478793412447, 'timing/model': 1.5665481500327587, 'learning_rate/qwen_vl_interface': 6.2504851472606246e-06, 'learning_rate/action_model': 6.250485147260625e-05, 'epoch': 0.9}) +[2026-06-16 03:09:31,188][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.26442983746528625, 'train/grad_norm_pre_clip': 12.33499813079834, 'timing/data': 0.011355754919350147, 'timing/model': 1.7918906598351896, 'learning_rate/qwen_vl_interface': 6.247507840791035e-06, 'learning_rate/action_model': 6.247507840791035e-05, 'epoch': 0.9}) +[2026-06-16 03:09:37,256][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.11764854192733765, 'train/grad_norm_pre_clip': 16.611156463623047, 'timing/data': 0.01099019218236208, 'timing/model': 1.5542943156324327, 'learning_rate/qwen_vl_interface': 6.244530124283651e-06, 'learning_rate/action_model': 6.24453012428365e-05, 'epoch': 0.9}) +[2026-06-16 03:09:43,744][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.01535372156649828, 'train/grad_norm_pre_clip': 2.014997720718384, 'timing/data': 0.011883948929607868, 'timing/model': 1.5591271999292076, 'learning_rate/qwen_vl_interface': 6.2415519989625016e-06, 'learning_rate/action_model': 6.241551998962501e-05, 'epoch': 0.9}) +[2026-06-16 03:09:50,061][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.37720218300819397, 'train/grad_norm_pre_clip': 17.195636749267578, 'timing/data': 0.011079161893576384, 'timing/model': 1.5688898130320013, 'learning_rate/qwen_vl_interface': 6.238573466051781e-06, 'learning_rate/action_model': 6.23857346605178e-05, 'epoch': 0.9}) +[2026-06-16 03:09:56,364][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.0009660020004957914, 'train/grad_norm_pre_clip': 0.08570460975170135, 'timing/data': 0.010837451089173555, 'timing/model': 1.5768518708646297, 'learning_rate/qwen_vl_interface': 6.2355945267758465e-06, 'learning_rate/action_model': 6.235594526775846e-05, 'epoch': 0.9}) +[2026-06-16 03:10:02,658][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.03617022559046745, 'train/grad_norm_pre_clip': 9.66356372833252, 'timing/data': 0.010450287256389856, 'timing/model': 1.5557156358845532, 'learning_rate/qwen_vl_interface': 6.232615182359231e-06, 'learning_rate/action_model': 6.232615182359231e-05, 'epoch': 0.9}) +[2026-06-16 03:10:09,090][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.003362542251124978, 'train/grad_norm_pre_clip': 0.5432478785514832, 'timing/data': 0.01124656293541193, 'timing/model': 1.8038918310776353, 'learning_rate/qwen_vl_interface': 6.229635434026634e-06, 'learning_rate/action_model': 6.229635434026634e-05, 'epoch': 0.9}) +[2026-06-16 03:10:15,174][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.07740455865859985, 'train/grad_norm_pre_clip': 14.144659996032715, 'timing/data': 0.011295191943645477, 'timing/model': 1.577554080169648, 'learning_rate/qwen_vl_interface': 6.226655283002912e-06, 'learning_rate/action_model': 6.226655283002911e-05, 'epoch': 0.9}) +[2026-06-16 03:10:21,461][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.09729227423667908, 'train/grad_norm_pre_clip': 5.6736063957214355, 'timing/data': 0.011278931982815266, 'timing/model': 1.5647748382762074, 'learning_rate/qwen_vl_interface': 6.223674730513096e-06, 'learning_rate/action_model': 6.223674730513096e-05, 'epoch': 0.9}) +[2026-06-16 03:10:27,928][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.041002724319696426, 'train/grad_norm_pre_clip': 6.523136615753174, 'timing/data': 0.010828178841620684, 'timing/model': 1.636344050988555, 'learning_rate/qwen_vl_interface': 6.220693777782375e-06, 'learning_rate/action_model': 6.220693777782375e-05, 'epoch': 0.9}) +[2026-06-16 03:10:34,155][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.010937711223959923, 'train/grad_norm_pre_clip': 1.587099313735962, 'timing/data': 0.010773047804832458, 'timing/model': 1.5216553299687803, 'learning_rate/qwen_vl_interface': 6.21771242603611e-06, 'learning_rate/action_model': 6.21771242603611e-05, 'epoch': 0.9}) +[2026-06-16 03:10:40,644][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.014868137426674366, 'train/grad_norm_pre_clip': 5.039952278137207, 'timing/data': 0.0112755810841918, 'timing/model': 1.5543130110017955, 'learning_rate/qwen_vl_interface': 6.21473067649982e-06, 'learning_rate/action_model': 6.21473067649982e-05, 'epoch': 0.9}) +[2026-06-16 03:10:47,008][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.21043699979782104, 'train/grad_norm_pre_clip': 15.365224838256836, 'timing/data': 0.010534485802054405, 'timing/model': 1.814796351827681, 'learning_rate/qwen_vl_interface': 6.211748530399189e-06, 'learning_rate/action_model': 6.211748530399189e-05, 'epoch': 0.9}) +[2026-06-16 03:10:53,061][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.1365186721086502, 'train/grad_norm_pre_clip': 22.14047622680664, 'timing/data': 0.010925128124654293, 'timing/model': 1.5620008553378284, 'learning_rate/qwen_vl_interface': 6.208765988960067e-06, 'learning_rate/action_model': 6.208765988960066e-05, 'epoch': 0.9}) +[2026-06-16 03:10:59,470][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.07233302295207977, 'train/grad_norm_pre_clip': 15.3338623046875, 'timing/data': 0.010923569090664387, 'timing/model': 1.5637022997252643, 'learning_rate/qwen_vl_interface': 6.2057830534084615e-06, 'learning_rate/action_model': 6.205783053408461e-05, 'epoch': 0.9}) +[2026-06-16 03:11:05,765][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.11740166693925858, 'train/grad_norm_pre_clip': 8.656543731689453, 'timing/data': 0.010841317940503359, 'timing/model': 1.5553963058628142, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 0.9}) +[2026-06-16 03:11:12,185][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.003987311851233244, 'train/grad_norm_pre_clip': 0.38039353489875793, 'timing/data': 0.011103079188615084, 'timing/model': 1.5609336206689477, 'learning_rate/qwen_vl_interface': 6.199816004872655e-06, 'learning_rate/action_model': 6.199816004872655e-05, 'epoch': 0.9}) +[2026-06-16 03:11:18,731][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.13040395081043243, 'train/grad_norm_pre_clip': 11.289810180664062, 'timing/data': 0.012501108925789595, 'timing/model': 1.626448741182685, 'learning_rate/qwen_vl_interface': 6.196831894341281e-06, 'learning_rate/action_model': 6.19683189434128e-05, 'epoch': 0.9}) +[2026-06-16 03:11:24,997][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.07737469673156738, 'train/grad_norm_pre_clip': 19.77596092224121, 'timing/data': 0.010544575285166502, 'timing/model': 1.8038722621276975, 'learning_rate/qwen_vl_interface': 6.1938473946030795e-06, 'learning_rate/action_model': 6.193847394603079e-05, 'epoch': 0.91}) +[2026-06-16 03:11:31,058][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.151254341006279, 'train/grad_norm_pre_clip': 20.175161361694336, 'timing/data': 0.010706644970923662, 'timing/model': 1.5610248032025993, 'learning_rate/qwen_vl_interface': 6.1908625068848695e-06, 'learning_rate/action_model': 6.19086250688487e-05, 'epoch': 0.91}) +[2026-06-16 03:11:37,377][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.1308167576789856, 'train/grad_norm_pre_clip': 23.459278106689453, 'timing/data': 0.011104047764092684, 'timing/model': 1.5720404759049416, 'learning_rate/qwen_vl_interface': 6.187877232413619e-06, 'learning_rate/action_model': 6.18787723241362e-05, 'epoch': 0.91}) +[2026-06-16 03:11:43,665][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.03644238039851189, 'train/grad_norm_pre_clip': 9.350481033325195, 'timing/data': 0.01079353503882885, 'timing/model': 1.5619934741407633, 'learning_rate/qwen_vl_interface': 6.184891572416467e-06, 'learning_rate/action_model': 6.184891572416466e-05, 'epoch': 0.91}) +[2026-06-16 03:11:49,961][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.03175431862473488, 'train/grad_norm_pre_clip': 10.733526229858398, 'timing/data': 0.011672751046717167, 'timing/model': 1.5604242412373424, 'learning_rate/qwen_vl_interface': 6.181905528120703e-06, 'learning_rate/action_model': 6.181905528120703e-05, 'epoch': 0.91}) +[2026-06-16 03:11:56,270][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.0110922921448946, 'train/grad_norm_pre_clip': 1.2274971008300781, 'timing/data': 0.011178687680512667, 'timing/model': 1.521175836212933, 'learning_rate/qwen_vl_interface': 6.178919100753778e-06, 'learning_rate/action_model': 6.178919100753778e-05, 'epoch': 0.91}) +[2026-06-16 03:12:02,601][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.021449005231261253, 'train/grad_norm_pre_clip': 10.399349212646484, 'timing/data': 0.011425869073718786, 'timing/model': 1.8036267696879804, 'learning_rate/qwen_vl_interface': 6.175932291543299e-06, 'learning_rate/action_model': 6.175932291543298e-05, 'epoch': 0.91}) +[2026-06-16 03:12:08,656][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.20799776911735535, 'train/grad_norm_pre_clip': 4.172566890716553, 'timing/data': 0.010935526341199875, 'timing/model': 1.5538755431771278, 'learning_rate/qwen_vl_interface': 6.172945101717029e-06, 'learning_rate/action_model': 6.17294510171703e-05, 'epoch': 0.91}) +[2026-06-16 03:12:14,966][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.5046327710151672, 'train/grad_norm_pre_clip': 19.25513458251953, 'timing/data': 0.010974275879561901, 'timing/model': 1.564963508863002, 'learning_rate/qwen_vl_interface': 6.169957532502892e-06, 'learning_rate/action_model': 6.169957532502892e-05, 'epoch': 0.91}) +[2026-06-16 03:12:21,364][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.23778505623340607, 'train/grad_norm_pre_clip': 35.32445526123047, 'timing/data': 0.011969393119215965, 'timing/model': 1.5630129547789693, 'learning_rate/qwen_vl_interface': 6.166969585128962e-06, 'learning_rate/action_model': 6.166969585128962e-05, 'epoch': 0.91}) +[2026-06-16 03:12:27,666][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.0014416269259527326, 'train/grad_norm_pre_clip': 0.28335726261138916, 'timing/data': 0.013342971913516521, 'timing/model': 1.5635468089021742, 'learning_rate/qwen_vl_interface': 6.163981260823473e-06, 'learning_rate/action_model': 6.163981260823473e-05, 'epoch': 0.91}) +[2026-06-16 03:12:34,040][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.12298016995191574, 'train/grad_norm_pre_clip': 7.572927474975586, 'timing/data': 0.010804363992065191, 'timing/model': 1.5852908790111542, 'learning_rate/qwen_vl_interface': 6.16099256081481e-06, 'learning_rate/action_model': 6.16099256081481e-05, 'epoch': 0.91}) +[2026-06-16 03:12:40,301][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.01803719252347946, 'train/grad_norm_pre_clip': 4.942346572875977, 'timing/data': 0.011832240968942642, 'timing/model': 1.8011765670962632, 'learning_rate/qwen_vl_interface': 6.15800348633152e-06, 'learning_rate/action_model': 6.15800348633152e-05, 'epoch': 0.91}) +[2026-06-16 03:12:46,359][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.015931637957692146, 'train/grad_norm_pre_clip': 4.765203952789307, 'timing/data': 0.01053600199520588, 'timing/model': 1.5638520452193916, 'learning_rate/qwen_vl_interface': 6.155014038602293e-06, 'learning_rate/action_model': 6.155014038602292e-05, 'epoch': 0.91}) +[2026-06-16 03:12:52,662][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.08822604268789291, 'train/grad_norm_pre_clip': 6.312662601470947, 'timing/data': 0.011046535335481167, 'timing/model': 1.5653248480521142, 'learning_rate/qwen_vl_interface': 6.152024218855979e-06, 'learning_rate/action_model': 6.15202421885598e-05, 'epoch': 0.91}) +[2026-06-16 03:12:58,957][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.18730247020721436, 'train/grad_norm_pre_clip': 16.02704620361328, 'timing/data': 0.011058975011110306, 'timing/model': 1.564684795215726, 'learning_rate/qwen_vl_interface': 6.149034028321583e-06, 'learning_rate/action_model': 6.149034028321583e-05, 'epoch': 0.91}) +[2026-06-16 03:13:05,261][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.2685234546661377, 'train/grad_norm_pre_clip': 11.92182731628418, 'timing/data': 0.010589872021228075, 'timing/model': 1.566575480159372, 'learning_rate/qwen_vl_interface': 6.1460434682282566e-06, 'learning_rate/action_model': 6.146043468228256e-05, 'epoch': 0.91}) +[2026-06-16 03:13:11,563][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.1994396597146988, 'train/grad_norm_pre_clip': 24.59379768371582, 'timing/data': 0.012823806144297123, 'timing/model': 1.528529026079923, 'learning_rate/qwen_vl_interface': 6.143052539805308e-06, 'learning_rate/action_model': 6.143052539805308e-05, 'epoch': 0.91}) +[2026-06-16 03:13:17,884][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.17844818532466888, 'train/grad_norm_pre_clip': 20.965045928955078, 'timing/data': 0.012193852104246616, 'timing/model': 1.7614251687191427, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 0.91}) +[2026-06-16 03:13:24,119][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.00437224330380559, 'train/grad_norm_pre_clip': 0.3565676212310791, 'timing/data': 0.011470186989754438, 'timing/model': 1.715533060953021, 'learning_rate/qwen_vl_interface': 6.137069582888521e-06, 'learning_rate/action_model': 6.137069582888521e-05, 'epoch': 0.91}) +[2026-06-16 03:13:30,182][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.21156655251979828, 'train/grad_norm_pre_clip': 11.1192626953125, 'timing/data': 0.011773888021707535, 'timing/model': 1.583667229861021, 'learning_rate/qwen_vl_interface': 6.13407755685405e-06, 'learning_rate/action_model': 6.13407755685405e-05, 'epoch': 0.91}) +[2026-06-16 03:13:36,472][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.09632087498903275, 'train/grad_norm_pre_clip': 6.607309818267822, 'timing/data': 0.013238599058240652, 'timing/model': 1.5583671131171286, 'learning_rate/qwen_vl_interface': 6.131085167408692e-06, 'learning_rate/action_model': 6.131085167408692e-05, 'epoch': 0.91}) +[2026-06-16 03:13:42,925][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.02479693666100502, 'train/grad_norm_pre_clip': 3.430569887161255, 'timing/data': 0.01182265905663371, 'timing/model': 1.604728997219354, 'learning_rate/qwen_vl_interface': 6.128092415782502e-06, 'learning_rate/action_model': 6.128092415782501e-05, 'epoch': 0.91}) +[2026-06-16 03:13:49,271][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.01781909540295601, 'train/grad_norm_pre_clip': 3.1529860496520996, 'timing/data': 0.010611360892653465, 'timing/model': 1.6060157492756844, 'learning_rate/qwen_vl_interface': 6.125099303205688e-06, 'learning_rate/action_model': 6.125099303205688e-05, 'epoch': 0.91}) +[2026-06-16 03:13:55,701][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.06400979310274124, 'train/grad_norm_pre_clip': 12.17126178741455, 'timing/data': 0.011773478239774704, 'timing/model': 1.7891632081009448, 'learning_rate/qwen_vl_interface': 6.122105830908607e-06, 'learning_rate/action_model': 6.122105830908606e-05, 'epoch': 0.91}) +[2026-06-16 03:14:02,017][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.1553277224302292, 'train/grad_norm_pre_clip': 11.183246612548828, 'timing/data': 0.01103531289845705, 'timing/model': 1.815379774197936, 'learning_rate/qwen_vl_interface': 6.11911200012176e-06, 'learning_rate/action_model': 6.11911200012176e-05, 'epoch': 0.92}) +[2026-06-16 03:14:08,272][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.03901670128107071, 'train/grad_norm_pre_clip': 7.951599597930908, 'timing/data': 0.011187003925442696, 'timing/model': 1.5728008528240025, 'learning_rate/qwen_vl_interface': 6.116117812075799e-06, 'learning_rate/action_model': 6.116117812075799e-05, 'epoch': 0.92}) +[2026-06-16 03:14:14,666][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.0915146917104721, 'train/grad_norm_pre_clip': 16.329397201538086, 'timing/data': 0.0109856016933918, 'timing/model': 1.573386783245951, 'learning_rate/qwen_vl_interface': 6.113123268001524e-06, 'learning_rate/action_model': 6.113123268001524e-05, 'epoch': 0.92}) +[2026-06-16 03:14:21,069][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.014344045892357826, 'train/grad_norm_pre_clip': 4.9530487060546875, 'timing/data': 0.011784047819674015, 'timing/model': 1.5722406413406134, 'learning_rate/qwen_vl_interface': 6.110128369129873e-06, 'learning_rate/action_model': 6.110128369129873e-05, 'epoch': 0.92}) +[2026-06-16 03:14:27,369][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.02072836086153984, 'train/grad_norm_pre_clip': 2.048621416091919, 'timing/data': 0.011166352778673172, 'timing/model': 1.5181487118825316, 'learning_rate/qwen_vl_interface': 6.107133116691943e-06, 'learning_rate/action_model': 6.107133116691942e-05, 'epoch': 0.92}) +[2026-06-16 03:14:33,810][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.04300766438245773, 'train/grad_norm_pre_clip': 10.035289764404297, 'timing/data': 0.012282890267670155, 'timing/model': 1.8007091088220477, 'learning_rate/qwen_vl_interface': 6.104137511918963e-06, 'learning_rate/action_model': 6.104137511918963e-05, 'epoch': 0.92}) +[2026-06-16 03:14:39,867][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.06718605756759644, 'train/grad_norm_pre_clip': 9.59627914428711, 'timing/data': 0.010943461675196886, 'timing/model': 1.5641898992471397, 'learning_rate/qwen_vl_interface': 6.1011415560423155e-06, 'learning_rate/action_model': 6.101141556042316e-05, 'epoch': 0.92}) +[2026-06-16 03:14:46,163][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.17370446026325226, 'train/grad_norm_pre_clip': 15.409735679626465, 'timing/data': 0.010903680697083473, 'timing/model': 1.5511788721196353, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.92}) +[2026-06-16 03:14:52,462][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.019687727093696594, 'train/grad_norm_pre_clip': 4.014915466308594, 'timing/data': 0.013829180039465427, 'timing/model': 1.5610025930218399, 'learning_rate/qwen_vl_interface': 6.09514859590426e-06, 'learning_rate/action_model': 6.09514859590426e-05, 'epoch': 0.92}) +[2026-06-16 03:14:58,765][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.02878888137638569, 'train/grad_norm_pre_clip': 7.834902763366699, 'timing/data': 0.011847317218780518, 'timing/model': 1.5278162960894406, 'learning_rate/qwen_vl_interface': 6.09215159410633e-06, 'learning_rate/action_model': 6.09215159410633e-05, 'epoch': 0.92}) +[2026-06-16 03:15:05,055][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.028070639818906784, 'train/grad_norm_pre_clip': 4.498373985290527, 'timing/data': 0.011458024848252535, 'timing/model': 1.5891146310605109, 'learning_rate/qwen_vl_interface': 6.089154246131692e-06, 'learning_rate/action_model': 6.089154246131692e-05, 'epoch': 0.92}) +[2026-06-16 03:15:11,369][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.001684027723968029, 'train/grad_norm_pre_clip': 0.21459200978279114, 'timing/data': 0.011395062785595655, 'timing/model': 1.7613132619298995, 'learning_rate/qwen_vl_interface': 6.08615655321244e-06, 'learning_rate/action_model': 6.08615655321244e-05, 'epoch': 0.92}) +[2026-06-16 03:15:17,469][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.12728184461593628, 'train/grad_norm_pre_clip': 18.912431716918945, 'timing/data': 0.012055297382175922, 'timing/model': 1.561946993228048, 'learning_rate/qwen_vl_interface': 6.083158516580814e-06, 'learning_rate/action_model': 6.083158516580814e-05, 'epoch': 0.92}) +[2026-06-16 03:15:23,790][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.005693200509995222, 'train/grad_norm_pre_clip': 0.863113284111023, 'timing/data': 0.010990690905600786, 'timing/model': 1.586129818111658, 'learning_rate/qwen_vl_interface': 6.080160137469192e-06, 'learning_rate/action_model': 6.0801601374691916e-05, 'epoch': 0.92}) +[2026-06-16 03:15:30,158][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.07938817888498306, 'train/grad_norm_pre_clip': 15.891279220581055, 'timing/data': 0.011139401234686375, 'timing/model': 1.565412444062531, 'learning_rate/qwen_vl_interface': 6.077161417110097e-06, 'learning_rate/action_model': 6.0771614171100965e-05, 'epoch': 0.92}) +[2026-06-16 03:15:36,459][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.1681162714958191, 'train/grad_norm_pre_clip': 19.664772033691406, 'timing/data': 0.011226562317460775, 'timing/model': 1.5317824962548912, 'learning_rate/qwen_vl_interface': 6.0741623567361864e-06, 'learning_rate/action_model': 6.074162356736186e-05, 'epoch': 0.92}) +[2026-06-16 03:15:42,774][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.009421689435839653, 'train/grad_norm_pre_clip': 1.6136292219161987, 'timing/data': 0.011010739952325821, 'timing/model': 1.5280267633497715, 'learning_rate/qwen_vl_interface': 6.071162957580262e-06, 'learning_rate/action_model': 6.071162957580262e-05, 'epoch': 0.92}) +[2026-06-16 03:15:49,064][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.030131688341498375, 'train/grad_norm_pre_clip': 18.835498809814453, 'timing/data': 0.011195552069693804, 'timing/model': 1.7626134813763201, 'learning_rate/qwen_vl_interface': 6.068163220875266e-06, 'learning_rate/action_model': 6.068163220875266e-05, 'epoch': 0.92}) +[2026-06-16 03:15:55,170][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.00751685444265604, 'train/grad_norm_pre_clip': 2.596092700958252, 'timing/data': 0.013163814321160316, 'timing/model': 1.5663418900221586, 'learning_rate/qwen_vl_interface': 6.065163147854273e-06, 'learning_rate/action_model': 6.065163147854273e-05, 'epoch': 0.92}) +[2026-06-16 03:16:01,466][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.014566927216947079, 'train/grad_norm_pre_clip': 2.1949102878570557, 'timing/data': 0.010899438988417387, 'timing/model': 1.56529699312523, 'learning_rate/qwen_vl_interface': 6.062162739750502e-06, 'learning_rate/action_model': 6.062162739750502e-05, 'epoch': 0.92}) +[2026-06-16 03:16:07,769][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.07616981863975525, 'train/grad_norm_pre_clip': 8.088432312011719, 'timing/data': 0.011373471934348345, 'timing/model': 1.5694876709021628, 'learning_rate/qwen_vl_interface': 6.059161997797306e-06, 'learning_rate/action_model': 6.0591619977973056e-05, 'epoch': 0.92}) +[2026-06-16 03:16:14,063][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.13893017172813416, 'train/grad_norm_pre_clip': 8.885887145996094, 'timing/data': 0.010759907774627209, 'timing/model': 1.5180000360123813, 'learning_rate/qwen_vl_interface': 6.056160923228178e-06, 'learning_rate/action_model': 6.056160923228178e-05, 'epoch': 0.92}) +[2026-06-16 03:16:20,393][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.010810391046106815, 'train/grad_norm_pre_clip': 2.1170380115509033, 'timing/data': 0.011187410913407803, 'timing/model': 1.7891319100745022, 'learning_rate/qwen_vl_interface': 6.0531595172767455e-06, 'learning_rate/action_model': 6.0531595172767455e-05, 'epoch': 0.92}) +[2026-06-16 03:16:26,508][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.03508271649479866, 'train/grad_norm_pre_clip': 7.102457046508789, 'timing/data': 0.010227554012089968, 'timing/model': 1.601945429109037, 'learning_rate/qwen_vl_interface': 6.050157781176774e-06, 'learning_rate/action_model': 6.050157781176774e-05, 'epoch': 0.92}) +[2026-06-16 03:16:32,958][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.08636973798274994, 'train/grad_norm_pre_clip': 11.900470733642578, 'timing/data': 0.011969905346632004, 'timing/model': 1.573450441006571, 'learning_rate/qwen_vl_interface': 6.047155716162164e-06, 'learning_rate/action_model': 6.047155716162164e-05, 'epoch': 0.92}) +[2026-06-16 03:16:39,266][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.023701850324869156, 'train/grad_norm_pre_clip': 9.244514465332031, 'timing/data': 0.010645716916769743, 'timing/model': 1.579180058091879, 'learning_rate/qwen_vl_interface': 6.044153323466948e-06, 'learning_rate/action_model': 6.044153323466948e-05, 'epoch': 0.93}) +[2026-06-16 03:16:45,674][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.12199550122022629, 'train/grad_norm_pre_clip': 10.953499794006348, 'timing/data': 0.012325686868280172, 'timing/model': 1.6942118108272552, 'learning_rate/qwen_vl_interface': 6.0411506043253e-06, 'learning_rate/action_model': 6.0411506043253e-05, 'epoch': 0.93}) +[2026-06-16 03:16:51,960][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.22078825533390045, 'train/grad_norm_pre_clip': 10.090302467346191, 'timing/data': 0.010158264078199863, 'timing/model': 1.5490455944091082, 'learning_rate/qwen_vl_interface': 6.038147559971523e-06, 'learning_rate/action_model': 6.038147559971523e-05, 'epoch': 0.93}) +[2026-06-16 03:16:58,300][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.0959194004535675, 'train/grad_norm_pre_clip': 11.211760520935059, 'timing/data': 0.010183424223214388, 'timing/model': 1.7958809481933713, 'learning_rate/qwen_vl_interface': 6.035144191640055e-06, 'learning_rate/action_model': 6.035144191640055e-05, 'epoch': 0.93}) +[2026-06-16 03:17:04,384][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.04887375980615616, 'train/grad_norm_pre_clip': 22.88010025024414, 'timing/data': 0.011449820827692747, 'timing/model': 1.5887786480598152, 'learning_rate/qwen_vl_interface': 6.032140500565467e-06, 'learning_rate/action_model': 6.0321405005654665e-05, 'epoch': 0.93}) +[2026-06-16 03:17:10,680][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.1446482390165329, 'train/grad_norm_pre_clip': 7.4352335929870605, 'timing/data': 0.011493080761283636, 'timing/model': 1.5332696549594402, 'learning_rate/qwen_vl_interface': 6.029136487982463e-06, 'learning_rate/action_model': 6.029136487982463e-05, 'epoch': 0.93}) +[2026-06-16 03:17:16,962][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.012170243076980114, 'train/grad_norm_pre_clip': 12.330810546875, 'timing/data': 0.010779985692352057, 'timing/model': 1.5584673550911248, 'learning_rate/qwen_vl_interface': 6.02613215512588e-06, 'learning_rate/action_model': 6.0261321551258796e-05, 'epoch': 0.93}) +[2026-06-16 03:17:23,264][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.03465535119175911, 'train/grad_norm_pre_clip': 23.219863891601562, 'timing/data': 0.010565584991127253, 'timing/model': 1.5626965840347111, 'learning_rate/qwen_vl_interface': 6.023127503230684e-06, 'learning_rate/action_model': 6.023127503230683e-05, 'epoch': 0.93}) +[2026-06-16 03:17:29,565][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.2544723153114319, 'train/grad_norm_pre_clip': 15.819783210754395, 'timing/data': 0.011608502361923456, 'timing/model': 1.5608126162551343, 'learning_rate/qwen_vl_interface': 6.020122533531976e-06, 'learning_rate/action_model': 6.0201225335319755e-05, 'epoch': 0.93}) +[2026-06-16 03:17:35,899][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.08949100226163864, 'train/grad_norm_pre_clip': 14.059499740600586, 'timing/data': 0.011571021750569344, 'timing/model': 1.7879826151765883, 'learning_rate/qwen_vl_interface': 6.017117247264983e-06, 'learning_rate/action_model': 6.017117247264983e-05, 'epoch': 0.93}) +[2026-06-16 03:17:41,968][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.11942356824874878, 'train/grad_norm_pre_clip': 7.960186958312988, 'timing/data': 0.010696826037019491, 'timing/model': 1.568895183969289, 'learning_rate/qwen_vl_interface': 6.014111645665069e-06, 'learning_rate/action_model': 6.0141116456650684e-05, 'epoch': 0.93}) +[2026-06-16 03:17:48,263][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.11010821908712387, 'train/grad_norm_pre_clip': 7.086360454559326, 'timing/data': 0.011575421784073114, 'timing/model': 1.5589370266534388, 'learning_rate/qwen_vl_interface': 6.011105729967719e-06, 'learning_rate/action_model': 6.0111057299677184e-05, 'epoch': 0.93}) +[2026-06-16 03:28:40,216][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.11484996229410172, 'train/grad_norm_pre_clip': 17.96904754638672, 'eval/loss': 0.04290285333991051, 'eval/flap_precision': 0.9574175824175825, 'eval/flap_recall': 0.9508867667121419, 'eval/flap_f1': 0.9541409993155373, 'eval/flap_rate_model': 0.20222222222222222, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9541409993155373, 'timing/data': 0.01134506007656455, 'timing/model': 1.559495594818145, 'learning_rate/qwen_vl_interface': 6.008099501408552e-06, 'learning_rate/action_model': 6.008099501408552e-05, 'epoch': 0.93}) +[2026-06-16 03:28:40,217][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2300_state +[2026-06-16 03:28:40,223][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 03:28:48,634][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2300_state/model.safetensors +[2026-06-16 03:29:07,965][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2300_state/optimizer.bin +[2026-06-16 03:29:07,966][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2300_state/sampler.bin +[2026-06-16 03:29:07,967][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2300_state/sampler_1.bin +[2026-06-16 03:29:07,968][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2300_state/random_states_0.pkl +[2026-06-16 03:29:08,153][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2300 +[2026-06-16 03:29:08,154][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 03:29:08,169][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 03:29:14,325][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.006177516654133797, 'train/grad_norm_pre_clip': 0.5718179941177368, 'timing/data': 0.01135525293648243, 'timing/model': 1.5200188593007624, 'learning_rate/qwen_vl_interface': 6.0050929612233165e-06, 'learning_rate/action_model': 6.005092961223316e-05, 'epoch': 0.93}) +[2026-06-16 03:29:20,426][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.20701265335083008, 'train/grad_norm_pre_clip': 14.591962814331055, 'timing/data': 0.01213216781616211, 'timing/model': 1.5532535999082029, 'learning_rate/qwen_vl_interface': 6.002086110647888e-06, 'learning_rate/action_model': 6.002086110647888e-05, 'epoch': 0.93}) +[2026-06-16 03:29:26,594][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.00594481872394681, 'train/grad_norm_pre_clip': 0.6724024415016174, 'timing/data': 0.01159176416695118, 'timing/model': 1.546343837864697, 'learning_rate/qwen_vl_interface': 5.999078950918267e-06, 'learning_rate/action_model': 5.999078950918267e-05, 'epoch': 0.93}) +[2026-06-16 03:29:32,864][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.042915232479572296, 'train/grad_norm_pre_clip': 12.881149291992188, 'timing/data': 0.011393873021006584, 'timing/model': 1.8080486068502069, 'learning_rate/qwen_vl_interface': 5.996071483270585e-06, 'learning_rate/action_model': 5.996071483270585e-05, 'epoch': 0.93}) +[2026-06-16 03:29:38,945][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.008673341013491154, 'train/grad_norm_pre_clip': 4.334274768829346, 'timing/data': 0.011278052814304829, 'timing/model': 1.5743994158692658, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.93}) +[2026-06-16 03:29:45,241][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.3175550699234009, 'train/grad_norm_pre_clip': 19.191221237182617, 'timing/data': 0.010759939905256033, 'timing/model': 1.570152647793293, 'learning_rate/qwen_vl_interface': 5.990055629166187e-06, 'learning_rate/action_model': 5.990055629166187e-05, 'epoch': 0.93}) +[2026-06-16 03:29:51,444][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.2985486686229706, 'train/grad_norm_pre_clip': 8.4501953125, 'timing/data': 0.01131492294371128, 'timing/model': 1.564929106272757, 'learning_rate/qwen_vl_interface': 5.987047245182363e-06, 'learning_rate/action_model': 5.987047245182362e-05, 'epoch': 0.93}) +[2026-06-16 03:29:57,640][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.008136960677802563, 'train/grad_norm_pre_clip': 1.1252195835113525, 'timing/data': 0.01069822907447815, 'timing/model': 1.5811576819978654, 'learning_rate/qwen_vl_interface': 5.984038558226255e-06, 'learning_rate/action_model': 5.9840385582262546e-05, 'epoch': 0.93}) +[2026-06-16 03:30:03,938][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.048891082406044006, 'train/grad_norm_pre_clip': 11.543417930603027, 'timing/data': 0.010433177929371595, 'timing/model': 1.5363549897447228, 'learning_rate/qwen_vl_interface': 5.981029569534625e-06, 'learning_rate/action_model': 5.981029569534625e-05, 'epoch': 0.93}) +[2026-06-16 03:30:10,271][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.1071704849600792, 'train/grad_norm_pre_clip': 12.18143367767334, 'timing/data': 0.010726799257099628, 'timing/model': 1.7967432290315628, 'learning_rate/qwen_vl_interface': 5.978020280344351e-06, 'learning_rate/action_model': 5.9780202803443506e-05, 'epoch': 0.93}) +[2026-06-16 03:30:16,346][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.004328580107539892, 'train/grad_norm_pre_clip': 0.43397706747055054, 'timing/data': 0.01107995118945837, 'timing/model': 1.5622463962063193, 'learning_rate/qwen_vl_interface': 5.97501069189244e-06, 'learning_rate/action_model': 5.97501069189244e-05, 'epoch': 0.93}) +[2026-06-16 03:30:22,646][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.13236385583877563, 'train/grad_norm_pre_clip': 8.726716995239258, 'timing/data': 0.011642265133559704, 'timing/model': 1.5559952780604362, 'learning_rate/qwen_vl_interface': 5.97200080541602e-06, 'learning_rate/action_model': 5.97200080541602e-05, 'epoch': 0.93}) +[2026-06-16 03:30:28,854][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.07942932099103928, 'train/grad_norm_pre_clip': 4.753047466278076, 'timing/data': 0.00999260600656271, 'timing/model': 1.5220593358390033, 'learning_rate/qwen_vl_interface': 5.968990622152341e-06, 'learning_rate/action_model': 5.9689906221523404e-05, 'epoch': 0.94}) +[2026-06-16 03:30:35,151][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.008090409450232983, 'train/grad_norm_pre_clip': 0.9356772303581238, 'timing/data': 0.011212672106921673, 'timing/model': 1.5550683760084212, 'learning_rate/qwen_vl_interface': 5.965980143338774e-06, 'learning_rate/action_model': 5.965980143338774e-05, 'epoch': 0.94}) +[2026-06-16 03:30:41,443][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.01802896149456501, 'train/grad_norm_pre_clip': 14.350894927978516, 'timing/data': 0.012498068623244762, 'timing/model': 1.5180323980748653, 'learning_rate/qwen_vl_interface': 5.962969370212819e-06, 'learning_rate/action_model': 5.9629693702128185e-05, 'epoch': 0.94}) +[2026-06-16 03:30:47,785][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.015835454687476158, 'train/grad_norm_pre_clip': 8.774718284606934, 'timing/data': 0.012259208131581545, 'timing/model': 1.8031804892234504, 'learning_rate/qwen_vl_interface': 5.959958304012085e-06, 'learning_rate/action_model': 5.9599583040120845e-05, 'epoch': 0.94}) +[2026-06-16 03:30:53,860][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.20439158380031586, 'train/grad_norm_pre_clip': 11.89732837677002, 'timing/data': 0.010802968870848417, 'timing/model': 1.5710478341206908, 'learning_rate/qwen_vl_interface': 5.956946945974309e-06, 'learning_rate/action_model': 5.9569469459743085e-05, 'epoch': 0.94}) +[2026-06-16 03:31:00,152][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.12645934522151947, 'train/grad_norm_pre_clip': 11.66785717010498, 'timing/data': 0.010504518169909716, 'timing/model': 1.5565058710053563, 'learning_rate/qwen_vl_interface': 5.953935297337347e-06, 'learning_rate/action_model': 5.953935297337347e-05, 'epoch': 0.94}) +[2026-06-16 03:31:06,461][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.010038575157523155, 'train/grad_norm_pre_clip': 6.775396347045898, 'timing/data': 0.01109463069587946, 'timing/model': 1.5578698352910578, 'learning_rate/qwen_vl_interface': 5.950923359339175e-06, 'learning_rate/action_model': 5.950923359339174e-05, 'epoch': 0.94}) +[2026-06-16 03:31:12,760][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.4215179979801178, 'train/grad_norm_pre_clip': 26.86050796508789, 'timing/data': 0.010984661057591438, 'timing/model': 1.5619278829544783, 'learning_rate/qwen_vl_interface': 5.947911133217886e-06, 'learning_rate/action_model': 5.947911133217886e-05, 'epoch': 0.94}) +[2026-06-16 03:31:19,152][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.006914868950843811, 'train/grad_norm_pre_clip': 5.461627960205078, 'timing/data': 0.010619817767292261, 'timing/model': 1.5372677352279425, 'learning_rate/qwen_vl_interface': 5.944898620211693e-06, 'learning_rate/action_model': 5.944898620211693e-05, 'epoch': 0.94}) +[2026-06-16 03:31:25,486][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.02873701974749565, 'train/grad_norm_pre_clip': 9.114153861999512, 'timing/data': 0.010908900760114193, 'timing/model': 1.7888430268503726, 'learning_rate/qwen_vl_interface': 5.941885821558925e-06, 'learning_rate/action_model': 5.9418858215589245e-05, 'epoch': 0.94}) +[2026-06-16 03:31:31,553][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.16066594421863556, 'train/grad_norm_pre_clip': 10.643889427185059, 'timing/data': 0.010553847998380661, 'timing/model': 1.5638848352245986, 'learning_rate/qwen_vl_interface': 5.938872738498031e-06, 'learning_rate/action_model': 5.93887273849803e-05, 'epoch': 0.94}) +[2026-06-16 03:31:37,956][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.07771188765764236, 'train/grad_norm_pre_clip': 15.437189102172852, 'timing/data': 0.013024481013417244, 'timing/model': 1.5618351912125945, 'learning_rate/qwen_vl_interface': 5.935859372267575e-06, 'learning_rate/action_model': 5.935859372267575e-05, 'epoch': 0.94}) +[2026-06-16 03:31:44,255][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.26970136165618896, 'train/grad_norm_pre_clip': 27.10744285583496, 'timing/data': 0.011606373824179173, 'timing/model': 1.5646722228266299, 'learning_rate/qwen_vl_interface': 5.9328457241062396e-06, 'learning_rate/action_model': 5.9328457241062394e-05, 'epoch': 0.94}) +[2026-06-16 03:31:50,657][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.13447849452495575, 'train/grad_norm_pre_clip': 10.399016380310059, 'timing/data': 0.010831109248101711, 'timing/model': 1.5726145338267088, 'learning_rate/qwen_vl_interface': 5.929831795252819e-06, 'learning_rate/action_model': 5.9298317952528184e-05, 'epoch': 0.94}) +[2026-06-16 03:31:56,959][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.004908136557787657, 'train/grad_norm_pre_clip': 0.5169154405593872, 'timing/data': 0.010981750208884478, 'timing/model': 1.5177163491025567, 'learning_rate/qwen_vl_interface': 5.926817586946227e-06, 'learning_rate/action_model': 5.926817586946227e-05, 'epoch': 0.94}) +[2026-06-16 03:32:03,365][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.14775881171226501, 'train/grad_norm_pre_clip': 7.371831893920898, 'timing/data': 0.012084056157618761, 'timing/model': 1.8711840989999473, 'learning_rate/qwen_vl_interface': 5.923803100425491e-06, 'learning_rate/action_model': 5.92380310042549e-05, 'epoch': 0.94}) +[2026-06-16 03:32:09,641][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.010816880501806736, 'train/grad_norm_pre_clip': 2.5421345233917236, 'timing/data': 0.011432833038270473, 'timing/model': 1.7524487432092428, 'learning_rate/qwen_vl_interface': 5.92078833692975e-06, 'learning_rate/action_model': 5.92078833692975e-05, 'epoch': 0.94}) +[2026-06-16 03:32:15,764][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.004363361746072769, 'train/grad_norm_pre_clip': 0.5886292457580566, 'timing/data': 0.011218792293220758, 'timing/model': 1.5713195838034153, 'learning_rate/qwen_vl_interface': 5.917773297698264e-06, 'learning_rate/action_model': 5.9177732976982636e-05, 'epoch': 0.94}) +[2026-06-16 03:32:22,060][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.010876311920583248, 'train/grad_norm_pre_clip': 2.1560025215148926, 'timing/data': 0.011204470880329609, 'timing/model': 1.565308899153024, 'learning_rate/qwen_vl_interface': 5.9147579839704e-06, 'learning_rate/action_model': 5.9147579839704e-05, 'epoch': 0.94}) +[2026-06-16 03:32:28,452][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.12801112234592438, 'train/grad_norm_pre_clip': 12.900798797607422, 'timing/data': 0.01005415478721261, 'timing/model': 1.5733693400397897, 'learning_rate/qwen_vl_interface': 5.911742396985637e-06, 'learning_rate/action_model': 5.9117423969856364e-05, 'epoch': 0.94}) +[2026-06-16 03:32:34,794][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.19307871162891388, 'train/grad_norm_pre_clip': 39.47562026977539, 'timing/data': 0.010561488103121519, 'timing/model': 1.5633787522092462, 'learning_rate/qwen_vl_interface': 5.908726537983571e-06, 'learning_rate/action_model': 5.908726537983571e-05, 'epoch': 0.94}) +[2026-06-16 03:32:41,095][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.11578229069709778, 'train/grad_norm_pre_clip': 6.821996212005615, 'timing/data': 0.011517142876982689, 'timing/model': 1.800292570143938, 'learning_rate/qwen_vl_interface': 5.9057104082039085e-06, 'learning_rate/action_model': 5.905710408203908e-05, 'epoch': 0.94}) +[2026-06-16 03:32:47,407][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.005215756129473448, 'train/grad_norm_pre_clip': 2.382030963897705, 'timing/data': 0.009904864244163036, 'timing/model': 1.8020596471615136, 'learning_rate/qwen_vl_interface': 5.902694008886466e-06, 'learning_rate/action_model': 5.902694008886466e-05, 'epoch': 0.94}) +[2026-06-16 03:32:53,468][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.008221362717449665, 'train/grad_norm_pre_clip': 1.6341651678085327, 'timing/data': 0.012066405732184649, 'timing/model': 1.5804739482700825, 'learning_rate/qwen_vl_interface': 5.8996773412711694e-06, 'learning_rate/action_model': 5.899677341271169e-05, 'epoch': 0.94}) +[2026-06-16 03:32:59,786][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.005744713358581066, 'train/grad_norm_pre_clip': 1.1814148426055908, 'timing/data': 0.011103530880063772, 'timing/model': 1.5732750459574163, 'learning_rate/qwen_vl_interface': 5.896660406598059e-06, 'learning_rate/action_model': 5.896660406598059e-05, 'epoch': 0.95}) +[2026-06-16 03:33:06,189][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.1430206447839737, 'train/grad_norm_pre_clip': 53.978458404541016, 'timing/data': 0.010245295707136393, 'timing/model': 1.6908957678824663, 'learning_rate/qwen_vl_interface': 5.893643206107283e-06, 'learning_rate/action_model': 5.893643206107282e-05, 'epoch': 0.95}) +[2026-06-16 03:33:12,692][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.0046438477002084255, 'train/grad_norm_pre_clip': 1.5443387031555176, 'timing/data': 0.011057450901716948, 'timing/model': 1.5415827468968928, 'learning_rate/qwen_vl_interface': 5.890625741039096e-06, 'learning_rate/action_model': 5.890625741039095e-05, 'epoch': 0.95}) +[2026-06-16 03:33:19,137][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.1320383995771408, 'train/grad_norm_pre_clip': 20.31743049621582, 'timing/data': 0.010715838987380266, 'timing/model': 1.9386107390746474, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 0.95}) +[2026-06-16 03:33:25,167][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.001839580712839961, 'train/grad_norm_pre_clip': 0.26879292726516724, 'timing/data': 0.012020584661513567, 'timing/model': 1.5747202229686081, 'learning_rate/qwen_vl_interface': 5.884590022132064e-06, 'learning_rate/action_model': 5.884590022132064e-05, 'epoch': 0.95}) +[2026-06-16 03:33:31,558][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.004370952490717173, 'train/grad_norm_pre_clip': 1.2263805866241455, 'timing/data': 0.011470532976090908, 'timing/model': 1.5475258720107377, 'learning_rate/qwen_vl_interface': 5.881571770774278e-06, 'learning_rate/action_model': 5.881571770774278e-05, 'epoch': 0.95}) +[2026-06-16 03:33:37,861][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.00458157854154706, 'train/grad_norm_pre_clip': 0.8358923196792603, 'timing/data': 0.011159460991621017, 'timing/model': 1.563940588850528, 'learning_rate/qwen_vl_interface': 5.878553259801192e-06, 'learning_rate/action_model': 5.878553259801192e-05, 'epoch': 0.95}) +[2026-06-16 03:33:44,159][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.15315519273281097, 'train/grad_norm_pre_clip': 9.836737632751465, 'timing/data': 0.010941789951175451, 'timing/model': 1.5260517192073166, 'learning_rate/qwen_vl_interface': 5.8755344904536025e-06, 'learning_rate/action_model': 5.8755344904536025e-05, 'epoch': 0.95}) +[2026-06-16 03:33:50,443][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.00438087061047554, 'train/grad_norm_pre_clip': 0.9674227237701416, 'timing/data': 0.011088009923696518, 'timing/model': 1.5786023046821356, 'learning_rate/qwen_vl_interface': 5.872515463972415e-06, 'learning_rate/action_model': 5.8725154639724146e-05, 'epoch': 0.95}) +[2026-06-16 03:33:56,827][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.00861792080104351, 'train/grad_norm_pre_clip': 3.272160291671753, 'timing/data': 0.013637464959174395, 'timing/model': 1.8082485869526863, 'learning_rate/qwen_vl_interface': 5.869496181598633e-06, 'learning_rate/action_model': 5.869496181598633e-05, 'epoch': 0.95}) +[2026-06-16 03:34:02,883][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.11458458751440048, 'train/grad_norm_pre_clip': 7.038826942443848, 'timing/data': 0.011353723239153624, 'timing/model': 1.5836653071455657, 'learning_rate/qwen_vl_interface': 5.866476644573371e-06, 'learning_rate/action_model': 5.866476644573371e-05, 'epoch': 0.95}) +[2026-06-16 03:34:09,166][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.2024478018283844, 'train/grad_norm_pre_clip': 14.926427841186523, 'timing/data': 0.01178248506039381, 'timing/model': 1.5614609732292593, 'learning_rate/qwen_vl_interface': 5.863456854137848e-06, 'learning_rate/action_model': 5.8634568541378474e-05, 'epoch': 0.95}) +[2026-06-16 03:34:15,474][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.007567830383777618, 'train/grad_norm_pre_clip': 2.818324089050293, 'timing/data': 0.012304617092013359, 'timing/model': 1.5758808497339487, 'learning_rate/qwen_vl_interface': 5.860436811533383e-06, 'learning_rate/action_model': 5.860436811533383e-05, 'epoch': 0.95}) +[2026-06-16 03:34:21,761][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.07767584919929504, 'train/grad_norm_pre_clip': 21.027156829833984, 'timing/data': 0.010732018854469061, 'timing/model': 1.5285832099616528, 'learning_rate/qwen_vl_interface': 5.857416518001404e-06, 'learning_rate/action_model': 5.857416518001404e-05, 'epoch': 0.95}) +[2026-06-16 03:34:28,055][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.051820382475852966, 'train/grad_norm_pre_clip': 31.44215965270996, 'timing/data': 0.011443051975220442, 'timing/model': 1.518106210976839, 'learning_rate/qwen_vl_interface': 5.854395974783439e-06, 'learning_rate/action_model': 5.8543959747834386e-05, 'epoch': 0.95}) +[2026-06-16 03:34:34,352][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.001120316912420094, 'train/grad_norm_pre_clip': 0.1609077751636505, 'timing/data': 0.01097270892933011, 'timing/model': 1.7557224961929023, 'learning_rate/qwen_vl_interface': 5.851375183121119e-06, 'learning_rate/action_model': 5.8513751831211186e-05, 'epoch': 0.95}) +[2026-06-16 03:34:40,461][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.11712135374546051, 'train/grad_norm_pre_clip': 10.786192893981934, 'timing/data': 0.01061033783480525, 'timing/model': 1.5598344039171934, 'learning_rate/qwen_vl_interface': 5.848354144256175e-06, 'learning_rate/action_model': 5.8483541442561754e-05, 'epoch': 0.95}) +[2026-06-16 03:34:47,143][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.026237262412905693, 'train/grad_norm_pre_clip': 23.718917846679688, 'timing/data': 0.011042589787393808, 'timing/model': 1.583973383065313, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 0.95}) +[2026-06-16 03:34:53,472][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.0032655277755111456, 'train/grad_norm_pre_clip': 0.5282872915267944, 'timing/data': 0.010961929801851511, 'timing/model': 1.5694796722382307, 'learning_rate/qwen_vl_interface': 5.842311329885866e-06, 'learning_rate/action_model': 5.842311329885866e-05, 'epoch': 0.95}) +[2026-06-16 03:34:59,777][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.13273203372955322, 'train/grad_norm_pre_clip': 15.450274467468262, 'timing/data': 0.011107800994068384, 'timing/model': 1.5289534078910947, 'learning_rate/qwen_vl_interface': 5.839289556864472e-06, 'learning_rate/action_model': 5.8392895568644714e-05, 'epoch': 0.95}) +[2026-06-16 03:35:06,384][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.3099276125431061, 'train/grad_norm_pre_clip': 10.955687522888184, 'timing/data': 0.01104007102549076, 'timing/model': 1.8066683607175946, 'learning_rate/qwen_vl_interface': 5.836267541608398e-06, 'learning_rate/action_model': 5.836267541608398e-05, 'epoch': 0.95}) +[2026-06-16 03:35:12,456][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.1368594765663147, 'train/grad_norm_pre_clip': 38.01361846923828, 'timing/data': 0.010545818135142326, 'timing/model': 1.5552975279279053, 'learning_rate/qwen_vl_interface': 5.833245285359887e-06, 'learning_rate/action_model': 5.833245285359886e-05, 'epoch': 0.95}) +[2026-06-16 03:35:18,753][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.13847249746322632, 'train/grad_norm_pre_clip': 24.96628761291504, 'timing/data': 0.010286536067724228, 'timing/model': 1.546613928861916, 'learning_rate/qwen_vl_interface': 5.830222789361268e-06, 'learning_rate/action_model': 5.8302227893612684e-05, 'epoch': 0.95}) +[2026-06-16 03:35:24,963][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.13895021378993988, 'train/grad_norm_pre_clip': 16.06041717529297, 'timing/data': 0.010777769144624472, 'timing/model': 1.5691598397679627, 'learning_rate/qwen_vl_interface': 5.8272000548549776e-06, 'learning_rate/action_model': 5.827200054854978e-05, 'epoch': 0.95}) +[2026-06-16 03:35:31,257][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.0020010333973914385, 'train/grad_norm_pre_clip': 0.41890791058540344, 'timing/data': 0.010600507259368896, 'timing/model': 1.5844629621133208, 'learning_rate/qwen_vl_interface': 5.824177083083547e-06, 'learning_rate/action_model': 5.8241770830835475e-05, 'epoch': 0.95}) +[2026-06-16 03:35:37,566][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.25118857622146606, 'train/grad_norm_pre_clip': 3.699263572692871, 'timing/data': 0.011139260604977608, 'timing/model': 1.576550685800612, 'learning_rate/qwen_vl_interface': 5.8211538752896056e-06, 'learning_rate/action_model': 5.821153875289606e-05, 'epoch': 0.96}) +[2026-06-16 03:35:43,898][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.05639569088816643, 'train/grad_norm_pre_clip': 45.98173141479492, 'timing/data': 0.011147111188620329, 'timing/model': 1.799553737975657, 'learning_rate/qwen_vl_interface': 5.8181304327158814e-06, 'learning_rate/action_model': 5.8181304327158814e-05, 'epoch': 0.96}) +[2026-06-16 03:35:49,968][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.0022965618409216404, 'train/grad_norm_pre_clip': 0.3739047050476074, 'timing/data': 0.010452187154442072, 'timing/model': 1.5663963868282735, 'learning_rate/qwen_vl_interface': 5.8151067566051965e-06, 'learning_rate/action_model': 5.815106756605196e-05, 'epoch': 0.96}) +[2026-06-16 03:35:56,365][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.193240225315094, 'train/grad_norm_pre_clip': 50.86190414428711, 'timing/data': 0.01128193223848939, 'timing/model': 1.5714083323255181, 'learning_rate/qwen_vl_interface': 5.81208284820047e-06, 'learning_rate/action_model': 5.81208284820047e-05, 'epoch': 0.96}) +[2026-06-16 03:36:02,664][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.2706090807914734, 'train/grad_norm_pre_clip': 35.50762176513672, 'timing/data': 0.013091581873595715, 'timing/model': 1.553702054079622, 'learning_rate/qwen_vl_interface': 5.809058708744716e-06, 'learning_rate/action_model': 5.809058708744716e-05, 'epoch': 0.96}) +[2026-06-16 03:36:09,062][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.06339288502931595, 'train/grad_norm_pre_clip': 34.319881439208984, 'timing/data': 0.010450187139213085, 'timing/model': 1.569106416311115, 'learning_rate/qwen_vl_interface': 5.8060343394810445e-06, 'learning_rate/action_model': 5.806034339481045e-05, 'epoch': 0.96}) +[2026-06-16 03:36:15,483][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.002797553315758705, 'train/grad_norm_pre_clip': 0.5468034148216248, 'timing/data': 0.01110786059871316, 'timing/model': 1.5691713439300656, 'learning_rate/qwen_vl_interface': 5.803009741652659e-06, 'learning_rate/action_model': 5.8030097416526585e-05, 'epoch': 0.96}) +[2026-06-16 03:36:21,824][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.0010715211974456906, 'train/grad_norm_pre_clip': 0.13429388403892517, 'timing/data': 0.011006060056388378, 'timing/model': 1.8247626731172204, 'learning_rate/qwen_vl_interface': 5.799984916502859e-06, 'learning_rate/action_model': 5.799984916502859e-05, 'epoch': 0.96}) +[2026-06-16 03:36:27,905][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.0845019593834877, 'train/grad_norm_pre_clip': 17.95408821105957, 'timing/data': 0.010946409311145544, 'timing/model': 1.5961236981675029, 'learning_rate/qwen_vl_interface': 5.7969598652750334e-06, 'learning_rate/action_model': 5.796959865275033e-05, 'epoch': 0.96}) +[2026-06-16 03:36:34,189][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.13062983751296997, 'train/grad_norm_pre_clip': 29.41358184814453, 'timing/data': 0.011700463946908712, 'timing/model': 1.5780334272421896, 'learning_rate/qwen_vl_interface': 5.793934589212666e-06, 'learning_rate/action_model': 5.793934589212666e-05, 'epoch': 0.96}) +[2026-06-16 03:36:40,630][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.2486943155527115, 'train/grad_norm_pre_clip': 22.938915252685547, 'timing/data': 0.011323710903525352, 'timing/model': 1.6270005418919027, 'learning_rate/qwen_vl_interface': 5.7909090895593375e-06, 'learning_rate/action_model': 5.790909089559338e-05, 'epoch': 0.96}) +[2026-06-16 03:36:46,863][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.0027680443599820137, 'train/grad_norm_pre_clip': 0.24052534997463226, 'timing/data': 0.011019209865480661, 'timing/model': 1.5328086582012475, 'learning_rate/qwen_vl_interface': 5.7878833675587135e-06, 'learning_rate/action_model': 5.787883367558713e-05, 'epoch': 0.96}) +[2026-06-16 03:36:53,260][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.46136489510536194, 'train/grad_norm_pre_clip': 34.17446517944336, 'timing/data': 0.011901774909347296, 'timing/model': 1.5776255000382662, 'learning_rate/qwen_vl_interface': 5.784857424454556e-06, 'learning_rate/action_model': 5.7848574244545563e-05, 'epoch': 0.96}) +[2026-06-16 03:36:59,597][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.004808837082237005, 'train/grad_norm_pre_clip': 0.9164831042289734, 'timing/data': 0.010689117945730686, 'timing/model': 1.80246448982507, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 0.96}) +[2026-06-16 03:37:05,731][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.10008010268211365, 'train/grad_norm_pre_clip': 26.124670028686523, 'timing/data': 0.010660318192094564, 'timing/model': 1.6384352366439998, 'learning_rate/qwen_vl_interface': 5.778804879911129e-06, 'learning_rate/action_model': 5.778804879911128e-05, 'epoch': 0.96}) +[2026-06-16 03:37:11,967][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.27245819568634033, 'train/grad_norm_pre_clip': 10.475165367126465, 'timing/data': 0.010035275015980005, 'timing/model': 1.574359237216413, 'learning_rate/qwen_vl_interface': 5.775778280959832e-06, 'learning_rate/action_model': 5.7757782809598315e-05, 'epoch': 0.96}) +[2026-06-16 03:37:18,383][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.26564761996269226, 'train/grad_norm_pre_clip': 11.263340950012207, 'timing/data': 0.011393330991268158, 'timing/model': 1.587312466930598, 'learning_rate/qwen_vl_interface': 5.7727514658809435e-06, 'learning_rate/action_model': 5.772751465880944e-05, 'epoch': 0.96}) +[2026-06-16 03:37:24,738][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.13879215717315674, 'train/grad_norm_pre_clip': 6.736388206481934, 'timing/data': 0.01159965293481946, 'timing/model': 1.635290214791894, 'learning_rate/qwen_vl_interface': 5.769724435918676e-06, 'learning_rate/action_model': 5.7697244359186755e-05, 'epoch': 0.96}) +[2026-06-16 03:37:30,961][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.1321610063314438, 'train/grad_norm_pre_clip': 28.314929962158203, 'timing/data': 0.01129657169803977, 'timing/model': 1.5275866189040244, 'learning_rate/qwen_vl_interface': 5.766697192317323e-06, 'learning_rate/action_model': 5.766697192317323e-05, 'epoch': 0.96}) +[2026-06-16 03:37:37,298][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.11180254817008972, 'train/grad_norm_pre_clip': 5.079353332519531, 'timing/data': 0.01176230376586318, 'timing/model': 1.7963444851338863, 'learning_rate/qwen_vl_interface': 5.763669736321271e-06, 'learning_rate/action_model': 5.763669736321271e-05, 'epoch': 0.96}) +[2026-06-16 03:37:43,460][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.017546452581882477, 'train/grad_norm_pre_clip': 10.78451919555664, 'timing/data': 0.011147779878228903, 'timing/model': 1.5718231657519937, 'learning_rate/qwen_vl_interface': 5.760642069174994e-06, 'learning_rate/action_model': 5.760642069174994e-05, 'epoch': 0.96}) +[2026-06-16 03:37:49,761][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.003997439984232187, 'train/grad_norm_pre_clip': 0.5516270995140076, 'timing/data': 0.011604372877627611, 'timing/model': 1.5576148368418217, 'learning_rate/qwen_vl_interface': 5.757614192123051e-06, 'learning_rate/action_model': 5.757614192123051e-05, 'epoch': 0.96}) +[2026-06-16 03:37:56,059][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.16277267038822174, 'train/grad_norm_pre_clip': 22.676959991455078, 'timing/data': 0.011004029773175716, 'timing/model': 1.5587787018157542, 'learning_rate/qwen_vl_interface': 5.754586106410089e-06, 'learning_rate/action_model': 5.754586106410088e-05, 'epoch': 0.96}) +[2026-06-16 03:38:02,359][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.09721963107585907, 'train/grad_norm_pre_clip': 14.787217140197754, 'timing/data': 0.010980119928717613, 'timing/model': 1.5544296470470726, 'learning_rate/qwen_vl_interface': 5.751557813280838e-06, 'learning_rate/action_model': 5.751557813280838e-05, 'epoch': 0.96}) +[2026-06-16 03:38:08,662][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.10673743486404419, 'train/grad_norm_pre_clip': 17.881669998168945, 'timing/data': 0.01097462885081768, 'timing/model': 1.5289980778470635, 'learning_rate/qwen_vl_interface': 5.7485293139801146e-06, 'learning_rate/action_model': 5.7485293139801146e-05, 'epoch': 0.96}) +[2026-06-16 03:38:15,009][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.013713992200791836, 'train/grad_norm_pre_clip': 4.003671169281006, 'timing/data': 0.010856749024242163, 'timing/model': 1.8165304642170668, 'learning_rate/qwen_vl_interface': 5.745500609752821e-06, 'learning_rate/action_model': 5.745500609752821e-05, 'epoch': 0.97}) +[2026-06-16 03:38:21,064][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.009837079793214798, 'train/grad_norm_pre_clip': 0.9484198689460754, 'timing/data': 0.011506262701004744, 'timing/model': 1.5654523223638535, 'learning_rate/qwen_vl_interface': 5.742471701843943e-06, 'learning_rate/action_model': 5.742471701843943e-05, 'epoch': 0.97}) +[2026-06-16 03:38:27,370][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.03689272329211235, 'train/grad_norm_pre_clip': 20.12102508544922, 'timing/data': 0.011307911016047001, 'timing/model': 1.5659206039272249, 'learning_rate/qwen_vl_interface': 5.739442591498553e-06, 'learning_rate/action_model': 5.739442591498553e-05, 'epoch': 0.97}) +[2026-06-16 03:38:33,666][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.08232318609952927, 'train/grad_norm_pre_clip': 35.52741241455078, 'timing/data': 0.011943415272980928, 'timing/model': 1.5631316169165075, 'learning_rate/qwen_vl_interface': 5.7364132799617985e-06, 'learning_rate/action_model': 5.7364132799617984e-05, 'epoch': 0.97}) +[2026-06-16 03:38:39,968][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.015165295451879501, 'train/grad_norm_pre_clip': 1.5344324111938477, 'timing/data': 0.01257220795378089, 'timing/model': 1.5634028473868966, 'learning_rate/qwen_vl_interface': 5.733383768478923e-06, 'learning_rate/action_model': 5.7333837684789225e-05, 'epoch': 0.97}) +[2026-06-16 03:38:46,448][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.1027308776974678, 'train/grad_norm_pre_clip': 10.173437118530273, 'timing/data': 0.011239281389862299, 'timing/model': 1.5867583616636693, 'learning_rate/qwen_vl_interface': 5.730354058295234e-06, 'learning_rate/action_model': 5.730354058295234e-05, 'epoch': 0.97}) +[2026-06-16 03:38:52,902][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.050171859562397, 'train/grad_norm_pre_clip': 15.241284370422363, 'timing/data': 0.01135195093229413, 'timing/model': 1.80261512985453, 'learning_rate/qwen_vl_interface': 5.72732415065614e-06, 'learning_rate/action_model': 5.727324150656139e-05, 'epoch': 0.97}) +[2026-06-16 03:38:58,970][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.2269303947687149, 'train/grad_norm_pre_clip': 10.057238578796387, 'timing/data': 0.011531391646713018, 'timing/model': 1.5689945141784847, 'learning_rate/qwen_vl_interface': 5.724294046807115e-06, 'learning_rate/action_model': 5.724294046807115e-05, 'epoch': 0.97}) +[2026-06-16 03:39:05,261][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.0825837254524231, 'train/grad_norm_pre_clip': 29.463626861572266, 'timing/data': 0.011426452081650496, 'timing/model': 1.5642000259831548, 'learning_rate/qwen_vl_interface': 5.7212637479937234e-06, 'learning_rate/action_model': 5.721263747993724e-05, 'epoch': 0.97}) +[2026-06-16 03:39:11,576][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.1976192593574524, 'train/grad_norm_pre_clip': 23.681987762451172, 'timing/data': 0.010761687997728586, 'timing/model': 1.5808525853790343, 'learning_rate/qwen_vl_interface': 5.718233255461607e-06, 'learning_rate/action_model': 5.718233255461607e-05, 'epoch': 0.97}) +[2026-06-16 03:39:17,912][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.06896119564771652, 'train/grad_norm_pre_clip': 7.285862445831299, 'timing/data': 0.011731673963367939, 'timing/model': 1.6094232276082039, 'learning_rate/qwen_vl_interface': 5.715202570456486e-06, 'learning_rate/action_model': 5.715202570456486e-05, 'epoch': 0.97}) +[2026-06-16 03:39:24,169][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.08789970725774765, 'train/grad_norm_pre_clip': 15.061495780944824, 'timing/data': 0.012306636199355125, 'timing/model': 1.5217034425586462, 'learning_rate/qwen_vl_interface': 5.7121716942241615e-06, 'learning_rate/action_model': 5.712171694224161e-05, 'epoch': 0.97}) +[2026-06-16 03:39:30,501][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.01071663573384285, 'train/grad_norm_pre_clip': 4.33581018447876, 'timing/data': 0.011011078953742981, 'timing/model': 1.8020788170397282, 'learning_rate/qwen_vl_interface': 5.709140628010509e-06, 'learning_rate/action_model': 5.7091406280105085e-05, 'epoch': 0.97}) +[2026-06-16 03:50:25,010][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.06560289114713669, 'train/grad_norm_pre_clip': 4.33685302734375, 'eval/loss': 0.0523090735077858, 'eval/flap_precision': 0.9465954606141522, 'eval/flap_recall': 0.9672578444747613, 'eval/flap_f1': 0.9568151147098516, 'eval/flap_rate_model': 0.20805555555555555, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9568151147098516, 'timing/data': 0.011213991325348616, 'timing/model': 1.7248993888497353, 'learning_rate/qwen_vl_interface': 5.706109373061489e-06, 'learning_rate/action_model': 5.706109373061489e-05, 'epoch': 0.97}) +[2026-06-16 03:50:25,011][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2400_state +[2026-06-16 03:50:25,018][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 03:50:33,397][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2400_state/model.safetensors +[2026-06-16 03:50:52,253][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2400_state/optimizer.bin +[2026-06-16 03:50:52,254][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2400_state/sampler.bin +[2026-06-16 03:50:52,254][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2400_state/sampler_1.bin +[2026-06-16 03:50:52,255][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2400_state/random_states_0.pkl +[2026-06-16 03:50:52,442][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2400 +[2026-06-16 03:50:52,442][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 03:50:52,457][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 03:50:58,770][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.1945827752351761, 'train/grad_norm_pre_clip': 4.508657455444336, 'timing/data': 0.011205530259758234, 'timing/model': 1.8241121713072062, 'learning_rate/qwen_vl_interface': 5.703077930623134e-06, 'learning_rate/action_model': 5.7030779306231344e-05, 'epoch': 0.97}) +[2026-06-16 03:51:04,708][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.006498402450233698, 'train/grad_norm_pre_clip': 0.6380751729011536, 'timing/data': 0.010505856946110725, 'timing/model': 1.5498441089875996, 'learning_rate/qwen_vl_interface': 5.7000463019415565e-06, 'learning_rate/action_model': 5.7000463019415563e-05, 'epoch': 0.97}) +[2026-06-16 03:51:11,025][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.22768080234527588, 'train/grad_norm_pre_clip': 13.547924995422363, 'timing/data': 0.01201604399830103, 'timing/model': 1.6470889211632311, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 0.97}) +[2026-06-16 03:51:17,229][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.07001330703496933, 'train/grad_norm_pre_clip': 4.256092548370361, 'timing/data': 0.011452730745077133, 'timing/model': 1.6566214822232723, 'learning_rate/qwen_vl_interface': 5.693982490833561e-06, 'learning_rate/action_model': 5.693982490833561e-05, 'epoch': 0.97}) +[2026-06-16 03:51:23,447][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.02409045398235321, 'train/grad_norm_pre_clip': 7.494551658630371, 'timing/data': 0.012414196971803904, 'timing/model': 1.5689976308494806, 'learning_rate/qwen_vl_interface': 5.690950310899745e-06, 'learning_rate/action_model': 5.6909503108997456e-05, 'epoch': 0.97}) +[2026-06-16 03:51:29,733][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.06205875799059868, 'train/grad_norm_pre_clip': 9.952545166015625, 'timing/data': 0.009866962675005198, 'timing/model': 1.521887089125812, 'learning_rate/qwen_vl_interface': 5.687917949707913e-06, 'learning_rate/action_model': 5.687917949707913e-05, 'epoch': 0.97}) +[2026-06-16 03:51:35,972][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.08072492480278015, 'train/grad_norm_pre_clip': 6.969238758087158, 'timing/data': 0.011444300878793001, 'timing/model': 1.7807125756517053, 'learning_rate/qwen_vl_interface': 5.684885408504555e-06, 'learning_rate/action_model': 5.684885408504554e-05, 'epoch': 0.97}) +[2026-06-16 03:51:42,312][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.022097699344158173, 'train/grad_norm_pre_clip': 15.825942993164062, 'timing/data': 0.011514991987496614, 'timing/model': 1.834295412991196, 'learning_rate/qwen_vl_interface': 5.681852688536231e-06, 'learning_rate/action_model': 5.681852688536231e-05, 'epoch': 0.97}) +[2026-06-16 03:51:48,350][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.07993457466363907, 'train/grad_norm_pre_clip': 22.29532241821289, 'timing/data': 0.011402050964534283, 'timing/model': 1.5627211071550846, 'learning_rate/qwen_vl_interface': 5.67881979104958e-06, 'learning_rate/action_model': 5.67881979104958e-05, 'epoch': 0.97}) +[2026-06-16 03:51:54,637][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.0051002055406570435, 'train/grad_norm_pre_clip': 0.5377304553985596, 'timing/data': 0.010706188157200813, 'timing/model': 1.5603472930379212, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 0.97}) +[2026-06-16 03:52:00,858][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.006187950260937214, 'train/grad_norm_pre_clip': 0.6967650651931763, 'timing/data': 0.011484910733997822, 'timing/model': 1.575490654911846, 'learning_rate/qwen_vl_interface': 5.672753468508204e-06, 'learning_rate/action_model': 5.672753468508204e-05, 'epoch': 0.97}) +[2026-06-16 03:52:07,158][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.06541980057954788, 'train/grad_norm_pre_clip': 8.25466251373291, 'timing/data': 0.011193959973752499, 'timing/model': 1.5229352628812194, 'learning_rate/qwen_vl_interface': 5.6697200459471165e-06, 'learning_rate/action_model': 5.669720045947117e-05, 'epoch': 0.98}) +[2026-06-16 03:52:13,526][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.18151243031024933, 'train/grad_norm_pre_clip': 6.158843994140625, 'timing/data': 0.011224020272493362, 'timing/model': 1.833421106915921, 'learning_rate/qwen_vl_interface': 5.666686450854973e-06, 'learning_rate/action_model': 5.666686450854973e-05, 'epoch': 0.98}) +[2026-06-16 03:52:19,554][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.07407307624816895, 'train/grad_norm_pre_clip': 18.361183166503906, 'timing/data': 0.011134679894894361, 'timing/model': 1.561450257897377, 'learning_rate/qwen_vl_interface': 5.663652684478768e-06, 'learning_rate/action_model': 5.6636526844787676e-05, 'epoch': 0.98}) +[2026-06-16 03:52:25,847][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.008774461224675179, 'train/grad_norm_pre_clip': 1.0224063396453857, 'timing/data': 0.01043040631338954, 'timing/model': 1.5626316047273576, 'learning_rate/qwen_vl_interface': 5.660618748065569e-06, 'learning_rate/action_model': 5.6606187480655695e-05, 'epoch': 0.98}) +[2026-06-16 03:52:32,153][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.11684775352478027, 'train/grad_norm_pre_clip': 16.984375, 'timing/data': 0.011076158843934536, 'timing/model': 1.5646534748375416, 'learning_rate/qwen_vl_interface': 5.657584642862514e-06, 'learning_rate/action_model': 5.657584642862513e-05, 'epoch': 0.98}) +[2026-06-16 03:52:38,453][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.007745785638689995, 'train/grad_norm_pre_clip': 1.03741455078125, 'timing/data': 0.012689787894487381, 'timing/model': 1.52996273804456, 'learning_rate/qwen_vl_interface': 5.654550370116808e-06, 'learning_rate/action_model': 5.654550370116808e-05, 'epoch': 0.98}) +[2026-06-16 03:52:44,753][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.010410148650407791, 'train/grad_norm_pre_clip': 3.0016250610351562, 'timing/data': 0.0111512397415936, 'timing/model': 1.5692720301449299, 'learning_rate/qwen_vl_interface': 5.651515931075726e-06, 'learning_rate/action_model': 5.651515931075726e-05, 'epoch': 0.98}) +[2026-06-16 03:52:51,052][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.14795298874378204, 'train/grad_norm_pre_clip': 14.411050796508789, 'timing/data': 0.011719923000782728, 'timing/model': 1.7614539489150047, 'learning_rate/qwen_vl_interface': 5.6484813269866126e-06, 'learning_rate/action_model': 5.648481326986612e-05, 'epoch': 0.98}) +[2026-06-16 03:52:57,154][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.011478410102427006, 'train/grad_norm_pre_clip': 5.411961555480957, 'timing/data': 0.011218270286917686, 'timing/model': 1.5613931068219244, 'learning_rate/qwen_vl_interface': 5.645446559096875e-06, 'learning_rate/action_model': 5.645446559096875e-05, 'epoch': 0.98}) +[2026-06-16 03:53:03,463][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.005862521007657051, 'train/grad_norm_pre_clip': 0.8922812938690186, 'timing/data': 0.011427971068769693, 'timing/model': 1.5150704588741064, 'learning_rate/qwen_vl_interface': 5.6424116286539965e-06, 'learning_rate/action_model': 5.642411628653996e-05, 'epoch': 0.98}) +[2026-06-16 03:53:09,756][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.1431572437286377, 'train/grad_norm_pre_clip': 18.708032608032227, 'timing/data': 0.010755597148090601, 'timing/model': 1.5557644655928016, 'learning_rate/qwen_vl_interface': 5.639376536905517e-06, 'learning_rate/action_model': 5.6393765369055175e-05, 'epoch': 0.98}) +[2026-06-16 03:53:16,057][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.026857834309339523, 'train/grad_norm_pre_clip': 15.56371021270752, 'timing/data': 0.010913458187133074, 'timing/model': 1.5242305169813335, 'learning_rate/qwen_vl_interface': 5.636341285099055e-06, 'learning_rate/action_model': 5.636341285099055e-05, 'epoch': 0.98}) +[2026-06-16 03:53:22,375][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.24587050080299377, 'train/grad_norm_pre_clip': 19.319164276123047, 'timing/data': 0.0121076051145792, 'timing/model': 1.5640602400526404, 'learning_rate/qwen_vl_interface': 5.6333058744822835e-06, 'learning_rate/action_model': 5.633305874482283e-05, 'epoch': 0.98}) +[2026-06-16 03:53:28,701][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.07250609993934631, 'train/grad_norm_pre_clip': 11.184876441955566, 'timing/data': 0.011309809982776642, 'timing/model': 1.7922823717817664, 'learning_rate/qwen_vl_interface': 5.630270306302944e-06, 'learning_rate/action_model': 5.6302703063029436e-05, 'epoch': 0.98}) +[2026-06-16 03:53:34,765][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.2582899034023285, 'train/grad_norm_pre_clip': 1.7469288110733032, 'timing/data': 0.011485651135444641, 'timing/model': 1.5559766460210085, 'learning_rate/qwen_vl_interface': 5.627234581808848e-06, 'learning_rate/action_model': 5.627234581808848e-05, 'epoch': 0.98}) +[2026-06-16 03:53:41,071][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.183033287525177, 'train/grad_norm_pre_clip': 17.34678840637207, 'timing/data': 0.011648662853986025, 'timing/model': 1.5647298032417893, 'learning_rate/qwen_vl_interface': 5.624198702247863e-06, 'learning_rate/action_model': 5.624198702247863e-05, 'epoch': 0.98}) +[2026-06-16 03:53:47,361][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.20912878215312958, 'train/grad_norm_pre_clip': 17.530906677246094, 'timing/data': 0.010925738606601954, 'timing/model': 1.554618407972157, 'learning_rate/qwen_vl_interface': 5.621162668867925e-06, 'learning_rate/action_model': 5.621162668867925e-05, 'epoch': 0.98}) +[2026-06-16 03:53:53,661][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.005198052618652582, 'train/grad_norm_pre_clip': 0.8624594807624817, 'timing/data': 0.010608367156237364, 'timing/model': 1.5597661547362804, 'learning_rate/qwen_vl_interface': 5.618126482917037e-06, 'learning_rate/action_model': 5.618126482917036e-05, 'epoch': 0.98}) +[2026-06-16 03:53:59,957][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.007420995272696018, 'train/grad_norm_pre_clip': 1.9319905042648315, 'timing/data': 0.010692576877772808, 'timing/model': 1.5313504240475595, 'learning_rate/qwen_vl_interface': 5.6150901456432546e-06, 'learning_rate/action_model': 5.615090145643255e-05, 'epoch': 0.98}) +[2026-06-16 03:54:06,300][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.040587712079286575, 'train/grad_norm_pre_clip': 18.021379470825195, 'timing/data': 0.01093387883156538, 'timing/model': 1.7993482183665037, 'learning_rate/qwen_vl_interface': 5.612053658294705e-06, 'learning_rate/action_model': 5.6120536582947045e-05, 'epoch': 0.98}) +[2026-06-16 03:54:12,365][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.04854515194892883, 'train/grad_norm_pre_clip': 22.7845401763916, 'timing/data': 0.011128859128803015, 'timing/model': 1.5503630237653852, 'learning_rate/qwen_vl_interface': 5.609017022119572e-06, 'learning_rate/action_model': 5.609017022119573e-05, 'epoch': 0.98}) +[2026-06-16 03:54:18,733][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.1526825875043869, 'train/grad_norm_pre_clip': 6.962230205535889, 'timing/data': 0.011877032928168774, 'timing/model': 1.6205944707617164, 'learning_rate/qwen_vl_interface': 5.605980238366104e-06, 'learning_rate/action_model': 5.605980238366103e-05, 'epoch': 0.98}) +[2026-06-16 03:54:24,983][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.019770309329032898, 'train/grad_norm_pre_clip': 6.5735321044921875, 'timing/data': 0.01128936093300581, 'timing/model': 1.5735239977948368, 'learning_rate/qwen_vl_interface': 5.602943308282606e-06, 'learning_rate/action_model': 5.602943308282605e-05, 'epoch': 0.98}) +[2026-06-16 03:54:31,359][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.008078974671661854, 'train/grad_norm_pre_clip': 4.091248989105225, 'timing/data': 0.01099321898072958, 'timing/model': 1.5956733170896769, 'learning_rate/qwen_vl_interface': 5.599906233117445e-06, 'learning_rate/action_model': 5.5999062331174446e-05, 'epoch': 0.98}) +[2026-06-16 03:54:37,666][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.2830181121826172, 'train/grad_norm_pre_clip': 24.38530921936035, 'timing/data': 0.010736486874520779, 'timing/model': 1.5497919209301472, 'learning_rate/qwen_vl_interface': 5.596869014119048e-06, 'learning_rate/action_model': 5.596869014119048e-05, 'epoch': 0.99}) +[2026-06-16 03:54:43,995][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.13781340420246124, 'train/grad_norm_pre_clip': 18.81334114074707, 'timing/data': 0.010949289426207542, 'timing/model': 1.7869476811029017, 'learning_rate/qwen_vl_interface': 5.593831652535902e-06, 'learning_rate/action_model': 5.593831652535902e-05, 'epoch': 0.99}) +[2026-06-16 03:54:50,167][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.10480496287345886, 'train/grad_norm_pre_clip': 14.989067077636719, 'timing/data': 0.01020383508875966, 'timing/model': 1.6456094831228256, 'learning_rate/qwen_vl_interface': 5.590794149616552e-06, 'learning_rate/action_model': 5.590794149616552e-05, 'epoch': 0.99}) +[2026-06-16 03:54:56,585][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.015468930825591087, 'train/grad_norm_pre_clip': 8.089517593383789, 'timing/data': 0.010934658348560333, 'timing/model': 1.5667888820171356, 'learning_rate/qwen_vl_interface': 5.587756506609601e-06, 'learning_rate/action_model': 5.5877565066096006e-05, 'epoch': 0.99}) +[2026-06-16 03:55:02,862][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.2991337478160858, 'train/grad_norm_pre_clip': 15.63850212097168, 'timing/data': 0.011297120247036219, 'timing/model': 1.5622862470336258, 'learning_rate/qwen_vl_interface': 5.584718724763707e-06, 'learning_rate/action_model': 5.584718724763707e-05, 'epoch': 0.99}) +[2026-06-16 03:55:09,158][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.1560659259557724, 'train/grad_norm_pre_clip': 7.2630743980407715, 'timing/data': 0.01054762676358223, 'timing/model': 1.5455143167637289, 'learning_rate/qwen_vl_interface': 5.581680805327588e-06, 'learning_rate/action_model': 5.5816808053275885e-05, 'epoch': 0.99}) +[2026-06-16 03:55:15,469][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.00667942501604557, 'train/grad_norm_pre_clip': 1.090268611907959, 'timing/data': 0.010744337923824787, 'timing/model': 1.519374436698854, 'learning_rate/qwen_vl_interface': 5.578642749550019e-06, 'learning_rate/action_model': 5.578642749550019e-05, 'epoch': 0.99}) +[2026-06-16 03:55:21,788][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.08390899747610092, 'train/grad_norm_pre_clip': 9.047150611877441, 'timing/data': 0.010777988005429506, 'timing/model': 1.7938923160545528, 'learning_rate/qwen_vl_interface': 5.575604558679831e-06, 'learning_rate/action_model': 5.57560455867983e-05, 'epoch': 0.99}) +[2026-06-16 03:55:27,943][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.02165587805211544, 'train/grad_norm_pre_clip': 6.2193450927734375, 'timing/data': 0.010465526022017002, 'timing/model': 1.5469924439676106, 'learning_rate/qwen_vl_interface': 5.5725662339659085e-06, 'learning_rate/action_model': 5.572566233965908e-05, 'epoch': 0.99}) +[2026-06-16 03:55:34,255][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.011894571594893932, 'train/grad_norm_pre_clip': 1.2212355136871338, 'timing/data': 0.01045200601220131, 'timing/model': 1.5513743381015956, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 0.99}) +[2026-06-16 03:55:40,553][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.012106047011911869, 'train/grad_norm_pre_clip': 4.111724376678467, 'timing/data': 0.011046758852899075, 'timing/model': 1.5581184839829803, 'learning_rate/qwen_vl_interface': 5.566489188002671e-06, 'learning_rate/action_model': 5.56648918800267e-05, 'epoch': 0.99}) +[2026-06-16 03:55:46,856][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.16811326146125793, 'train/grad_norm_pre_clip': 32.10081100463867, 'timing/data': 0.01037154532968998, 'timing/model': 1.550614973064512, 'learning_rate/qwen_vl_interface': 5.563450469251403e-06, 'learning_rate/action_model': 5.563450469251402e-05, 'epoch': 0.99}) +[2026-06-16 03:55:53,185][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.13202685117721558, 'train/grad_norm_pre_clip': 6.86836576461792, 'timing/data': 0.011132598854601383, 'timing/model': 1.5540431309491396, 'learning_rate/qwen_vl_interface': 5.560411621652488e-06, 'learning_rate/action_model': 5.560411621652488e-05, 'epoch': 0.99}) +[2026-06-16 03:55:59,606][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.12189295142889023, 'train/grad_norm_pre_clip': 7.503633975982666, 'timing/data': 0.010449306108057499, 'timing/model': 1.8127550459466875, 'learning_rate/qwen_vl_interface': 5.557372646455081e-06, 'learning_rate/action_model': 5.55737264645508e-05, 'epoch': 0.99}) +[2026-06-16 03:56:05,671][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.10174694657325745, 'train/grad_norm_pre_clip': 10.846817970275879, 'timing/data': 0.010978349018841982, 'timing/model': 1.5662708161398768, 'learning_rate/qwen_vl_interface': 5.554333544908388e-06, 'learning_rate/action_model': 5.5543335449083875e-05, 'epoch': 0.99}) +[2026-06-16 03:56:11,963][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.07392442971467972, 'train/grad_norm_pre_clip': 29.88343620300293, 'timing/data': 0.01097851898521185, 'timing/model': 1.5436004349030554, 'learning_rate/qwen_vl_interface': 5.551294318261671e-06, 'learning_rate/action_model': 5.5512943182616705e-05, 'epoch': 0.99}) +[2026-06-16 03:56:18,362][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.010932431556284428, 'train/grad_norm_pre_clip': 2.7217113971710205, 'timing/data': 0.011041229125112295, 'timing/model': 1.565135289914906, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 0.99}) +[2026-06-16 03:56:24,663][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.007101121358573437, 'train/grad_norm_pre_clip': 0.6605294942855835, 'timing/data': 0.011373820714652538, 'timing/model': 1.5666532879695296, 'learning_rate/qwen_vl_interface': 5.545215494665458e-06, 'learning_rate/action_model': 5.545215494665458e-05, 'epoch': 0.99}) +[2026-06-16 03:56:30,971][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.009539336897432804, 'train/grad_norm_pre_clip': 2.0968728065490723, 'timing/data': 0.011331411078572273, 'timing/model': 1.5247490229085088, 'learning_rate/qwen_vl_interface': 5.5421759002147376e-06, 'learning_rate/action_model': 5.542175900214738e-05, 'epoch': 0.99}) +[2026-06-16 03:56:37,291][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.08166555315256119, 'train/grad_norm_pre_clip': 26.217153549194336, 'timing/data': 0.011393322143703699, 'timing/model': 1.7896489407867193, 'learning_rate/qwen_vl_interface': 5.53913618566154e-06, 'learning_rate/action_model': 5.53913618566154e-05, 'epoch': 0.99}) +[2026-06-16 03:56:43,371][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.2591768205165863, 'train/grad_norm_pre_clip': 13.711316108703613, 'timing/data': 0.011475261766463518, 'timing/model': 1.5756070450879633, 'learning_rate/qwen_vl_interface': 5.536096352255381e-06, 'learning_rate/action_model': 5.536096352255381e-05, 'epoch': 0.99}) +[2026-06-16 03:56:49,661][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.004867507144808769, 'train/grad_norm_pre_clip': 0.5773054957389832, 'timing/data': 0.011311699636280537, 'timing/model': 1.5743087190203369, 'learning_rate/qwen_vl_interface': 5.533056401245817e-06, 'learning_rate/action_model': 5.5330564012458164e-05, 'epoch': 0.99}) +[2026-06-16 03:56:55,956][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.1869339942932129, 'train/grad_norm_pre_clip': 16.242412567138672, 'timing/data': 0.010677086655050516, 'timing/model': 1.5482821390032768, 'learning_rate/qwen_vl_interface': 5.530016333882456e-06, 'learning_rate/action_model': 5.5300163338824565e-05, 'epoch': 0.99}) +[2026-06-16 03:57:02,352][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.20901675522327423, 'train/grad_norm_pre_clip': 18.517261505126953, 'timing/data': 0.01067848689854145, 'timing/model': 1.5808189217932522, 'learning_rate/qwen_vl_interface': 5.52697615141496e-06, 'learning_rate/action_model': 5.52697615141496e-05, 'epoch': 0.99}) +[2026-06-16 03:57:08,733][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.2713013291358948, 'train/grad_norm_pre_clip': 10.118456840515137, 'timing/data': 0.010852279141545296, 'timing/model': 1.588918604888022, 'learning_rate/qwen_vl_interface': 5.52393585509303e-06, 'learning_rate/action_model': 5.52393585509303e-05, 'epoch': 0.99}) +[2026-06-16 03:57:15,018][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.07291914522647858, 'train/grad_norm_pre_clip': 9.547789573669434, 'timing/data': 0.010726277250796556, 'timing/model': 1.8155411090701818, 'learning_rate/qwen_vl_interface': 5.520895446166418e-06, 'learning_rate/action_model': 5.5208954461664175e-05, 'epoch': 1.0}) +[2026-06-16 03:57:21,352][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.018702389672398567, 'train/grad_norm_pre_clip': 8.689881324768066, 'timing/data': 0.010902388952672482, 'timing/model': 1.77563147386536, 'learning_rate/qwen_vl_interface': 5.517854925884919e-06, 'learning_rate/action_model': 5.517854925884919e-05, 'epoch': 1.0}) +[2026-06-16 03:57:27,457][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.06459293514490128, 'train/grad_norm_pre_clip': 39.23537063598633, 'timing/data': 0.011640122625976801, 'timing/model': 1.5592083055526018, 'learning_rate/qwen_vl_interface': 5.514814295498379e-06, 'learning_rate/action_model': 5.514814295498378e-05, 'epoch': 1.0}) +[2026-06-16 03:57:33,763][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.004738404415547848, 'train/grad_norm_pre_clip': 1.6182293891906738, 'timing/data': 0.01127023110166192, 'timing/model': 1.5501739070750773, 'learning_rate/qwen_vl_interface': 5.5117735562566846e-06, 'learning_rate/action_model': 5.5117735562566844e-05, 'epoch': 1.0}) +[2026-06-16 03:57:40,062][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.13848434388637543, 'train/grad_norm_pre_clip': 22.589635848999023, 'timing/data': 0.011427110992372036, 'timing/model': 1.566620344761759, 'learning_rate/qwen_vl_interface': 5.5087327094097685e-06, 'learning_rate/action_model': 5.508732709409769e-05, 'epoch': 1.0}) +[2026-06-16 03:57:46,463][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.0785464495420456, 'train/grad_norm_pre_clip': 7.774256229400635, 'timing/data': 0.011374081019312143, 'timing/model': 1.514466725755483, 'learning_rate/qwen_vl_interface': 5.505691756207609e-06, 'learning_rate/action_model': 5.505691756207609e-05, 'epoch': 1.0}) +[2026-06-16 03:57:52,792][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.005537814926356077, 'train/grad_norm_pre_clip': 0.7143864035606384, 'timing/data': 0.011267570778727531, 'timing/model': 1.784911043010652, 'learning_rate/qwen_vl_interface': 5.502650697900228e-06, 'learning_rate/action_model': 5.502650697900228e-05, 'epoch': 1.0}) +[2026-06-16 03:57:59,108][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.003074145643040538, 'train/grad_norm_pre_clip': 0.4888198673725128, 'timing/data': 0.010830957908183336, 'timing/model': 1.808824551757425, 'learning_rate/qwen_vl_interface': 5.499609535737688e-06, 'learning_rate/action_model': 5.4996095357376875e-05, 'epoch': 1.0}) +[2026-06-16 03:58:05,164][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.0022287580650299788, 'train/grad_norm_pre_clip': 0.22887709736824036, 'timing/data': 0.01110598910599947, 'timing/model': 1.5656573111191392, 'learning_rate/qwen_vl_interface': 5.496568270970096e-06, 'learning_rate/action_model': 5.496568270970096e-05, 'epoch': 1.0}) +[2026-06-16 03:58:11,468][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.007143502589315176, 'train/grad_norm_pre_clip': 0.842441976070404, 'timing/data': 0.011602762155234814, 'timing/model': 1.572375196032226, 'learning_rate/qwen_vl_interface': 5.4935269048476015e-06, 'learning_rate/action_model': 5.493526904847601e-05, 'epoch': 1.0}) +[2026-06-16 03:58:17,759][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.00488013681024313, 'train/grad_norm_pre_clip': 0.5121228694915771, 'timing/data': 0.011324760969728231, 'timing/model': 1.5609831949695945, 'learning_rate/qwen_vl_interface': 5.490485438620397e-06, 'learning_rate/action_model': 5.4904854386203965e-05, 'epoch': 1.0}) +[2026-06-16 03:58:24,141][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.08270031213760376, 'train/grad_norm_pre_clip': 22.469482421875, 'timing/data': 0.011342980898916721, 'timing/model': 1.5885334517806768, 'learning_rate/qwen_vl_interface': 5.487443873538712e-06, 'learning_rate/action_model': 5.4874438735387115e-05, 'epoch': 1.0}) +[2026-06-16 03:58:29,584][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.29283976554870605, 'train/grad_norm_pre_clip': 12.643941879272461, 'timing/data': 0.00044254306703805923, 'timing/model': 0.8677233541384339, 'learning_rate/qwen_vl_interface': 5.484402210852819e-06, 'learning_rate/action_model': 5.484402210852818e-05, 'epoch': 1.0}) +[2026-06-16 03:58:37,681][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.08442151546478271, 'train/grad_norm_pre_clip': 18.865747451782227, 'timing/data': 0.011236289981752634, 'timing/model': 1.5818898961879313, 'learning_rate/qwen_vl_interface': 5.48136045181303e-06, 'learning_rate/action_model': 5.48136045181303e-05, 'epoch': 1.0}) +[2026-06-16 03:58:43,857][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.07621775567531586, 'train/grad_norm_pre_clip': 12.82633113861084, 'timing/data': 0.01141696097329259, 'timing/model': 1.5541217071004212, 'learning_rate/qwen_vl_interface': 5.478318597669699e-06, 'learning_rate/action_model': 5.4783185976696995e-05, 'epoch': 1.0}) +[2026-06-16 03:58:50,071][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.008399181999266148, 'train/grad_norm_pre_clip': 1.0732300281524658, 'timing/data': 0.012578017078340054, 'timing/model': 1.5730390488170087, 'learning_rate/qwen_vl_interface': 5.475276649673218e-06, 'learning_rate/action_model': 5.4752766496732174e-05, 'epoch': 1.0}) +[2026-06-16 03:58:56,359][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.12545964121818542, 'train/grad_norm_pre_clip': 8.086129188537598, 'timing/data': 0.010418074671179056, 'timing/model': 1.6004025852307677, 'learning_rate/qwen_vl_interface': 5.472234609074014e-06, 'learning_rate/action_model': 5.472234609074014e-05, 'epoch': 1.0}) +[2026-06-16 03:59:02,717][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.0015681016957387328, 'train/grad_norm_pre_clip': 0.1549832671880722, 'timing/data': 0.011770923156291246, 'timing/model': 1.8194937272928655, 'learning_rate/qwen_vl_interface': 5.46919247712256e-06, 'learning_rate/action_model': 5.46919247712256e-05, 'epoch': 1.0}) +[2026-06-16 03:59:08,768][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.07720083743333817, 'train/grad_norm_pre_clip': 12.011746406555176, 'timing/data': 0.011643972247838974, 'timing/model': 1.5682875630445778, 'learning_rate/qwen_vl_interface': 5.466150255069356e-06, 'learning_rate/action_model': 5.4661502550693555e-05, 'epoch': 1.0}) +[2026-06-16 03:59:15,064][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.0010605888674035668, 'train/grad_norm_pre_clip': 0.13640670478343964, 'timing/data': 0.01189116295427084, 'timing/model': 1.5648593846708536, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 1.0}) +[2026-06-16 03:59:21,362][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.006439647637307644, 'train/grad_norm_pre_clip': 4.690661430358887, 'timing/data': 0.011346179991960526, 'timing/model': 1.5569461616687477, 'learning_rate/qwen_vl_interface': 5.46006554565991e-06, 'learning_rate/action_model': 5.4600655456599105e-05, 'epoch': 1.0}) +[2026-06-16 03:59:27,673][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.18608446419239044, 'train/grad_norm_pre_clip': 12.788593292236328, 'timing/data': 0.011053419206291437, 'timing/model': 1.5399936810135841, 'learning_rate/qwen_vl_interface': 5.457023060804865e-06, 'learning_rate/action_model': 5.457023060804864e-05, 'epoch': 1.0}) +[2026-06-16 03:59:34,287][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.07745786011219025, 'train/grad_norm_pre_clip': 23.019559860229492, 'timing/data': 0.03677937714383006, 'timing/model': 1.7816472742706537, 'learning_rate/qwen_vl_interface': 5.453980490850459e-06, 'learning_rate/action_model': 5.453980490850459e-05, 'epoch': 1.0}) +[2026-06-16 03:59:40,548][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.26669174432754517, 'train/grad_norm_pre_clip': 35.47310256958008, 'timing/data': 0.011815734207630157, 'timing/model': 1.756669550202787, 'learning_rate/qwen_vl_interface': 5.4509378370473785e-06, 'learning_rate/action_model': 5.450937837047378e-05, 'epoch': 1.0}) +[2026-06-16 03:59:46,665][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.005504191853106022, 'train/grad_norm_pre_clip': 1.6421723365783691, 'timing/data': 0.01099987793713808, 'timing/model': 1.5684603923000395, 'learning_rate/qwen_vl_interface': 5.4478951006463424e-06, 'learning_rate/action_model': 5.447895100646342e-05, 'epoch': 1.0}) +[2026-06-16 03:59:52,958][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.12950663268566132, 'train/grad_norm_pre_clip': 5.725584506988525, 'timing/data': 0.010838137939572334, 'timing/model': 1.5653851563110948, 'learning_rate/qwen_vl_interface': 5.444852282898106e-06, 'learning_rate/action_model': 5.444852282898106e-05, 'epoch': 1.01}) +[2026-06-16 03:59:59,263][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.29132935404777527, 'train/grad_norm_pre_clip': 18.984941482543945, 'timing/data': 0.01034500589594245, 'timing/model': 1.5605188999325037, 'learning_rate/qwen_vl_interface': 5.441809385053456e-06, 'learning_rate/action_model': 5.4418093850534555e-05, 'epoch': 1.01}) +[2026-06-16 04:00:05,558][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.0033634603023529053, 'train/grad_norm_pre_clip': 0.4076794683933258, 'timing/data': 0.011117649730294943, 'timing/model': 1.5153165878728032, 'learning_rate/qwen_vl_interface': 5.438766408363214e-06, 'learning_rate/action_model': 5.438766408363214e-05, 'epoch': 1.01}) +[2026-06-16 04:00:11,951][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.018375875428318977, 'train/grad_norm_pre_clip': 13.467886924743652, 'timing/data': 0.01106951991096139, 'timing/model': 1.57843220513314, 'learning_rate/qwen_vl_interface': 5.43572335407823e-06, 'learning_rate/action_model': 5.43572335407823e-05, 'epoch': 1.01}) +[2026-06-16 04:00:18,252][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.032555751502513885, 'train/grad_norm_pre_clip': 24.184839248657227, 'timing/data': 0.010822248179465532, 'timing/model': 1.7514936518855393, 'learning_rate/qwen_vl_interface': 5.432680223449392e-06, 'learning_rate/action_model': 5.4326802234493924e-05, 'epoch': 1.01}) +[2026-06-16 04:00:24,360][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.16115660965442657, 'train/grad_norm_pre_clip': 17.986127853393555, 'timing/data': 0.011297639925032854, 'timing/model': 1.559444174170494, 'learning_rate/qwen_vl_interface': 5.429637017727614e-06, 'learning_rate/action_model': 5.4296370177276144e-05, 'epoch': 1.01}) +[2026-06-16 04:00:30,663][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.10368532687425613, 'train/grad_norm_pre_clip': 16.00499153137207, 'timing/data': 0.011232140008360147, 'timing/model': 1.5549162197858095, 'learning_rate/qwen_vl_interface': 5.426593738163845e-06, 'learning_rate/action_model': 5.4265937381638446e-05, 'epoch': 1.01}) +[2026-06-16 04:00:37,056][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.30729660391807556, 'train/grad_norm_pre_clip': 28.974712371826172, 'timing/data': 0.010943008121103048, 'timing/model': 1.5669037541374564, 'learning_rate/qwen_vl_interface': 5.42355038600906e-06, 'learning_rate/action_model': 5.42355038600906e-05, 'epoch': 1.01}) +[2026-06-16 04:00:43,469][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.36724865436553955, 'train/grad_norm_pre_clip': 24.586181640625, 'timing/data': 0.010979958809912205, 'timing/model': 1.5172890983521938, 'learning_rate/qwen_vl_interface': 5.420506962514267e-06, 'learning_rate/action_model': 5.420506962514267e-05, 'epoch': 1.01}) +[2026-06-16 04:00:49,898][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.11143603175878525, 'train/grad_norm_pre_clip': 10.933390617370605, 'timing/data': 0.01162881264463067, 'timing/model': 1.8269248041324317, 'learning_rate/qwen_vl_interface': 5.4174634689305015e-06, 'learning_rate/action_model': 5.417463468930501e-05, 'epoch': 1.01}) +[2026-06-16 04:00:55,965][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.10079175978899002, 'train/grad_norm_pre_clip': 4.847581386566162, 'timing/data': 0.010684006847441196, 'timing/model': 1.5557144731283188, 'learning_rate/qwen_vl_interface': 5.41441990650883e-06, 'learning_rate/action_model': 5.4144199065088295e-05, 'epoch': 1.01}) +[2026-06-16 04:01:02,380][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.13089200854301453, 'train/grad_norm_pre_clip': 14.230751037597656, 'timing/data': 0.011496070772409439, 'timing/model': 1.574200853239745, 'learning_rate/qwen_vl_interface': 5.411376276500345e-06, 'learning_rate/action_model': 5.411376276500345e-05, 'epoch': 1.01}) +[2026-06-16 04:01:08,561][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.12342392653226852, 'train/grad_norm_pre_clip': 24.784826278686523, 'timing/data': 0.010468415915966034, 'timing/model': 1.566976364236325, 'learning_rate/qwen_vl_interface': 5.408332580156168e-06, 'learning_rate/action_model': 5.408332580156168e-05, 'epoch': 1.01}) +[2026-06-16 04:01:14,868][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.0033002484124153852, 'train/grad_norm_pre_clip': 0.6416152119636536, 'timing/data': 0.011187789961695671, 'timing/model': 1.571695868857205, 'learning_rate/qwen_vl_interface': 5.40528881872745e-06, 'learning_rate/action_model': 5.40528881872745e-05, 'epoch': 1.01}) +[2026-06-16 04:12:27,213][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.015990879386663437, 'train/grad_norm_pre_clip': 10.893688201904297, 'eval/loss': 0.0456237718462944, 'eval/flap_precision': 0.9312581063553826, 'eval/flap_recall': 0.9795361527967258, 'eval/flap_f1': 0.9547872340425533, 'eval/flap_rate_model': 0.21416666666666667, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9547872340425533, 'timing/data': 0.01123828999698162, 'timing/model': 1.5718644191510975, 'learning_rate/qwen_vl_interface': 5.402244993465364e-06, 'learning_rate/action_model': 5.402244993465364e-05, 'epoch': 1.01}) +[2026-06-16 04:12:27,214][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2500_state +[2026-06-16 04:12:27,221][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 04:12:35,547][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2500_state/model.safetensors +[2026-06-16 04:12:54,793][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2500_state/optimizer.bin +[2026-06-16 04:12:54,794][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2500_state/sampler.bin +[2026-06-16 04:12:54,795][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-16 04:12:54,796][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-16 04:12:54,977][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-16 04:12:54,977][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 04:12:54,993][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 04:13:01,127][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.004953800700604916, 'train/grad_norm_pre_clip': 0.5562405586242676, 'timing/data': 0.01020429702475667, 'timing/model': 1.4795552911236882, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 1.01}) +[2026-06-16 04:13:07,259][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.028060516342520714, 'train/grad_norm_pre_clip': 8.751519203186035, 'timing/data': 0.010941033717244864, 'timing/model': 1.717453335877508, 'learning_rate/qwen_vl_interface': 5.396157156445918e-06, 'learning_rate/action_model': 5.396157156445918e-05, 'epoch': 1.01}) +[2026-06-16 04:13:13,448][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.041493963450193405, 'train/grad_norm_pre_clip': 11.21444320678711, 'timing/data': 0.01086917519569397, 'timing/model': 1.6820136969909072, 'learning_rate/qwen_vl_interface': 5.39311314719104e-06, 'learning_rate/action_model': 5.39311314719104e-05, 'epoch': 1.01}) +[2026-06-16 04:13:19,334][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.203827902674675, 'train/grad_norm_pre_clip': 13.06917953491211, 'timing/data': 0.010970095172524452, 'timing/model': 1.4859725986607373, 'learning_rate/qwen_vl_interface': 5.390069079107755e-06, 'learning_rate/action_model': 5.390069079107755e-05, 'epoch': 1.01}) +[2026-06-16 04:13:25,536][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.1508014053106308, 'train/grad_norm_pre_clip': 24.388702392578125, 'timing/data': 0.010986885987222195, 'timing/model': 1.5368177378550172, 'learning_rate/qwen_vl_interface': 5.387024953447361e-06, 'learning_rate/action_model': 5.387024953447361e-05, 'epoch': 1.01}) +[2026-06-16 04:13:31,722][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.025248989462852478, 'train/grad_norm_pre_clip': 6.119999408721924, 'timing/data': 0.011695833876729012, 'timing/model': 1.568530234042555, 'learning_rate/qwen_vl_interface': 5.383980771461184e-06, 'learning_rate/action_model': 5.3839807714611846e-05, 'epoch': 1.01}) +[2026-06-16 04:13:37,835][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.0636419877409935, 'train/grad_norm_pre_clip': 17.158058166503906, 'timing/data': 0.011229417286813259, 'timing/model': 1.5103586339391768, 'learning_rate/qwen_vl_interface': 5.380936534400572e-06, 'learning_rate/action_model': 5.3809365344005726e-05, 'epoch': 1.01}) +[2026-06-16 04:13:44,086][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.13402363657951355, 'train/grad_norm_pre_clip': 24.600767135620117, 'timing/data': 0.011673295870423317, 'timing/model': 1.8093193913809955, 'learning_rate/qwen_vl_interface': 5.377892243516899e-06, 'learning_rate/action_model': 5.3778922435168985e-05, 'epoch': 1.01}) +[2026-06-16 04:13:50,151][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.14465002715587616, 'train/grad_norm_pre_clip': 9.804909706115723, 'timing/data': 0.011651057284325361, 'timing/model': 1.5688488129526377, 'learning_rate/qwen_vl_interface': 5.3748479000615525e-06, 'learning_rate/action_model': 5.374847900061553e-05, 'epoch': 1.01}) +[2026-06-16 04:13:56,456][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.18816278874874115, 'train/grad_norm_pre_clip': 6.536886215209961, 'timing/data': 0.010713422205299139, 'timing/model': 1.576155339833349, 'learning_rate/qwen_vl_interface': 5.371803505285954e-06, 'learning_rate/action_model': 5.371803505285954e-05, 'epoch': 1.02}) +[2026-06-16 04:14:02,838][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.26326075196266174, 'train/grad_norm_pre_clip': 20.479948043823242, 'timing/data': 0.010291534941643476, 'timing/model': 1.5543409921228886, 'learning_rate/qwen_vl_interface': 5.368759060441534e-06, 'learning_rate/action_model': 5.368759060441534e-05, 'epoch': 1.02}) +[2026-06-16 04:14:09,049][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.041533347219228745, 'train/grad_norm_pre_clip': 22.232757568359375, 'timing/data': 0.011531351134181023, 'timing/model': 1.4966411371715367, 'learning_rate/qwen_vl_interface': 5.365714566779749e-06, 'learning_rate/action_model': 5.3657145667797484e-05, 'epoch': 1.02}) +[2026-06-16 04:14:15,347][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.08664422482252121, 'train/grad_norm_pre_clip': 9.031412124633789, 'timing/data': 0.01120651327073574, 'timing/model': 1.5563350259326398, 'learning_rate/qwen_vl_interface': 5.362670025552077e-06, 'learning_rate/action_model': 5.362670025552077e-05, 'epoch': 1.02}) +[2026-06-16 04:14:21,676][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.052753496915102005, 'train/grad_norm_pre_clip': 11.023179054260254, 'timing/data': 0.011165353935211897, 'timing/model': 1.8055788157507777, 'learning_rate/qwen_vl_interface': 5.359625438010013e-06, 'learning_rate/action_model': 5.359625438010013e-05, 'epoch': 1.02}) +[2026-06-16 04:14:27,752][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.003175409510731697, 'train/grad_norm_pre_clip': 0.5932285785675049, 'timing/data': 0.012665789108723402, 'timing/model': 1.5499034859240055, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 1.02}) +[2026-06-16 04:14:34,053][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.0075651369988918304, 'train/grad_norm_pre_clip': 1.3300949335098267, 'timing/data': 0.012514251284301281, 'timing/model': 1.5538358646444976, 'learning_rate/qwen_vl_interface': 5.353536128988791e-06, 'learning_rate/action_model': 5.353536128988791e-05, 'epoch': 1.02}) +[2026-06-16 04:14:40,348][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.014971399679780006, 'train/grad_norm_pre_clip': 7.906802177429199, 'timing/data': 0.011202066671103239, 'timing/model': 1.564788259100169, 'learning_rate/qwen_vl_interface': 5.350491410012717e-06, 'learning_rate/action_model': 5.350491410012717e-05, 'epoch': 1.02}) +[2026-06-16 04:14:46,650][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.006445464212447405, 'train/grad_norm_pre_clip': 1.3470146656036377, 'timing/data': 0.011419296730309725, 'timing/model': 1.5642264462076128, 'learning_rate/qwen_vl_interface': 5.347446649728419e-06, 'learning_rate/action_model': 5.347446649728418e-05, 'epoch': 1.02}) +[2026-06-16 04:14:52,948][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.01762300916016102, 'train/grad_norm_pre_clip': 10.030620574951172, 'timing/data': 0.011835135985165834, 'timing/model': 1.5193132627755404, 'learning_rate/qwen_vl_interface': 5.344401849387484e-06, 'learning_rate/action_model': 5.344401849387484e-05, 'epoch': 1.02}) +[2026-06-16 04:14:59,278][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.013496589846909046, 'train/grad_norm_pre_clip': 5.850233554840088, 'timing/data': 0.010369252879172564, 'timing/model': 1.7335413261316717, 'learning_rate/qwen_vl_interface': 5.3413570102415144e-06, 'learning_rate/action_model': 5.3413570102415144e-05, 'epoch': 1.02}) +[2026-06-16 04:15:05,364][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.04499463737010956, 'train/grad_norm_pre_clip': 13.322677612304688, 'timing/data': 0.011643088888376951, 'timing/model': 1.5733530838042498, 'learning_rate/qwen_vl_interface': 5.338312133542131e-06, 'learning_rate/action_model': 5.3383121335421305e-05, 'epoch': 1.02}) +[2026-06-16 04:15:11,656][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.019608842208981514, 'train/grad_norm_pre_clip': 5.25860071182251, 'timing/data': 0.010656373109668493, 'timing/model': 1.5666432157158852, 'learning_rate/qwen_vl_interface': 5.335267220540963e-06, 'learning_rate/action_model': 5.335267220540963e-05, 'epoch': 1.02}) +[2026-06-16 04:15:17,858][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.002545018680393696, 'train/grad_norm_pre_clip': 0.25158289074897766, 'timing/data': 0.013005906250327826, 'timing/model': 1.5677861161530018, 'learning_rate/qwen_vl_interface': 5.332222272489662e-06, 'learning_rate/action_model': 5.332222272489662e-05, 'epoch': 1.02}) +[2026-06-16 04:15:24,275][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.005625342484563589, 'train/grad_norm_pre_clip': 0.8408892750740051, 'timing/data': 0.01137852342799306, 'timing/model': 1.5757711669430137, 'learning_rate/qwen_vl_interface': 5.3291772906398895e-06, 'learning_rate/action_model': 5.32917729063989e-05, 'epoch': 1.02}) +[2026-06-16 04:15:30,633][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.01224913727492094, 'train/grad_norm_pre_clip': 2.007080316543579, 'timing/data': 0.011541572865098715, 'timing/model': 1.5818902947939932, 'learning_rate/qwen_vl_interface': 5.326132276243327e-06, 'learning_rate/action_model': 5.3261322762433266e-05, 'epoch': 1.02}) +[2026-06-16 04:15:37,127][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.10413636267185211, 'train/grad_norm_pre_clip': 7.758216857910156, 'timing/data': 0.010892705991864204, 'timing/model': 1.8498248411342502, 'learning_rate/qwen_vl_interface': 5.323087230551659e-06, 'learning_rate/action_model': 5.323087230551659e-05, 'epoch': 1.02}) +[2026-06-16 04:15:43,167][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.24495172500610352, 'train/grad_norm_pre_clip': 25.0347957611084, 'timing/data': 0.010813537053763866, 'timing/model': 1.566018037032336, 'learning_rate/qwen_vl_interface': 5.3200421548165916e-06, 'learning_rate/action_model': 5.3200421548165914e-05, 'epoch': 1.02}) +[2026-06-16 04:15:49,464][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.005161346402019262, 'train/grad_norm_pre_clip': 0.8518091440200806, 'timing/data': 0.011533294804394245, 'timing/model': 1.5722172097302973, 'learning_rate/qwen_vl_interface': 5.3169970502898405e-06, 'learning_rate/action_model': 5.31699705028984e-05, 'epoch': 1.02}) +[2026-06-16 04:15:55,767][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.007415121886879206, 'train/grad_norm_pre_clip': 1.2911152839660645, 'timing/data': 0.011591976042836905, 'timing/model': 1.5662774182856083, 'learning_rate/qwen_vl_interface': 5.31395191822313e-06, 'learning_rate/action_model': 5.31395191822313e-05, 'epoch': 1.02}) +[2026-06-16 04:16:02,156][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.19720304012298584, 'train/grad_norm_pre_clip': 48.502586364746094, 'timing/data': 0.01167757622897625, 'timing/model': 1.5812817206606269, 'learning_rate/qwen_vl_interface': 5.310906759868205e-06, 'learning_rate/action_model': 5.310906759868205e-05, 'epoch': 1.02}) +[2026-06-16 04:16:08,455][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.1857210248708725, 'train/grad_norm_pre_clip': 17.336713790893555, 'timing/data': 0.01157398708164692, 'timing/model': 1.5118862348608673, 'learning_rate/qwen_vl_interface': 5.30786157647681e-06, 'learning_rate/action_model': 5.30786157647681e-05, 'epoch': 1.02}) +[2026-06-16 04:16:14,921][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.16323626041412354, 'train/grad_norm_pre_clip': 8.891510009765625, 'timing/data': 0.010842330288141966, 'timing/model': 1.8291878229938447, 'learning_rate/qwen_vl_interface': 5.304816369300709e-06, 'learning_rate/action_model': 5.304816369300709e-05, 'epoch': 1.02}) +[2026-06-16 04:16:20,957][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.12849122285842896, 'train/grad_norm_pre_clip': 23.96034049987793, 'timing/data': 0.011483268812298775, 'timing/model': 1.5719120111316442, 'learning_rate/qwen_vl_interface': 5.301771139591666e-06, 'learning_rate/action_model': 5.301771139591666e-05, 'epoch': 1.02}) +[2026-06-16 04:16:27,581][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.005767846014350653, 'train/grad_norm_pre_clip': 3.476072072982788, 'timing/data': 0.011826438829302788, 'timing/model': 1.5463527478277683, 'learning_rate/qwen_vl_interface': 5.298725888601465e-06, 'learning_rate/action_model': 5.298725888601465e-05, 'epoch': 1.02}) +[2026-06-16 04:16:34,047][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.0005432555335573852, 'train/grad_norm_pre_clip': 0.06554678082466125, 'timing/data': 0.011479820590466261, 'timing/model': 1.6506870612502098, 'learning_rate/qwen_vl_interface': 5.295680617581891e-06, 'learning_rate/action_model': 5.2956806175818904e-05, 'epoch': 1.03}) +[2026-06-16 04:16:40,358][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.25646498799324036, 'train/grad_norm_pre_clip': 34.408260345458984, 'timing/data': 0.011037382762879133, 'timing/model': 1.5602913829497993, 'learning_rate/qwen_vl_interface': 5.2926353277847395e-06, 'learning_rate/action_model': 5.292635327784739e-05, 'epoch': 1.03}) +[2026-06-16 04:16:46,661][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.058991655707359314, 'train/grad_norm_pre_clip': 19.76545524597168, 'timing/data': 0.011268042959272861, 'timing/model': 1.533504196908325, 'learning_rate/qwen_vl_interface': 5.2895900204618175e-06, 'learning_rate/action_model': 5.289590020461818e-05, 'epoch': 1.03}) +[2026-06-16 04:16:52,989][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.1321394294500351, 'train/grad_norm_pre_clip': 14.543498992919922, 'timing/data': 0.010518065188080072, 'timing/model': 1.788047008216381, 'learning_rate/qwen_vl_interface': 5.286544696864933e-06, 'learning_rate/action_model': 5.2865446968649324e-05, 'epoch': 1.03}) +[2026-06-16 04:16:59,437][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.36529162526130676, 'train/grad_norm_pre_clip': 24.01885223388672, 'timing/data': 0.01018288591876626, 'timing/model': 1.794763735961169, 'learning_rate/qwen_vl_interface': 5.283499358245904e-06, 'learning_rate/action_model': 5.2834993582459034e-05, 'epoch': 1.03}) +[2026-06-16 04:17:05,545][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.20668669044971466, 'train/grad_norm_pre_clip': 17.04691505432129, 'timing/data': 0.010193027090281248, 'timing/model': 1.5446514836512506, 'learning_rate/qwen_vl_interface': 5.280454005856553e-06, 'learning_rate/action_model': 5.2804540058565534e-05, 'epoch': 1.03}) +[2026-06-16 04:17:11,849][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.0964103490114212, 'train/grad_norm_pre_clip': 22.03887367248535, 'timing/data': 0.010592696256935596, 'timing/model': 1.566104358062148, 'learning_rate/qwen_vl_interface': 5.277408640948711e-06, 'learning_rate/action_model': 5.277408640948711e-05, 'epoch': 1.03}) +[2026-06-16 04:17:18,170][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.01652447134256363, 'train/grad_norm_pre_clip': 15.63161563873291, 'timing/data': 0.01055649807676673, 'timing/model': 1.5646124328486621, 'learning_rate/qwen_vl_interface': 5.274363264774213e-06, 'learning_rate/action_model': 5.274363264774212e-05, 'epoch': 1.03}) +[2026-06-16 04:17:24,540][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.03736111894249916, 'train/grad_norm_pre_clip': 18.842649459838867, 'timing/data': 0.010959866922348738, 'timing/model': 1.5954579231329262, 'learning_rate/qwen_vl_interface': 5.271317878584897e-06, 'learning_rate/action_model': 5.2713178785848974e-05, 'epoch': 1.03}) +[2026-06-16 04:17:30,820][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.0027303409297019243, 'train/grad_norm_pre_clip': 0.2924959659576416, 'timing/data': 0.010978998150676489, 'timing/model': 1.8162587410770357, 'learning_rate/qwen_vl_interface': 5.268272483632606e-06, 'learning_rate/action_model': 5.268272483632606e-05, 'epoch': 1.03}) +[2026-06-16 04:17:37,299][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.06592768430709839, 'train/grad_norm_pre_clip': 5.360158443450928, 'timing/data': 0.011039359029382467, 'timing/model': 1.9032795960083604, 'learning_rate/qwen_vl_interface': 5.265227081169187e-06, 'learning_rate/action_model': 5.265227081169186e-05, 'epoch': 1.03}) +[2026-06-16 04:17:43,449][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.0020614508539438248, 'train/grad_norm_pre_clip': 0.3232061564922333, 'timing/data': 0.011135029140859842, 'timing/model': 1.6192274610511959, 'learning_rate/qwen_vl_interface': 5.2621816724464895e-06, 'learning_rate/action_model': 5.262181672446489e-05, 'epoch': 1.03}) +[2026-06-16 04:17:49,769][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.04879459738731384, 'train/grad_norm_pre_clip': 19.296560287475586, 'timing/data': 0.011230668984353542, 'timing/model': 1.5708233788609505, 'learning_rate/qwen_vl_interface': 5.259136258716367e-06, 'learning_rate/action_model': 5.2591362587163675e-05, 'epoch': 1.03}) +[2026-06-16 04:17:56,165][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.18809251487255096, 'train/grad_norm_pre_clip': 8.26777458190918, 'timing/data': 0.010683211032301188, 'timing/model': 1.5868545481935143, 'learning_rate/qwen_vl_interface': 5.256090841230675e-06, 'learning_rate/action_model': 5.256090841230674e-05, 'epoch': 1.03}) +[2026-06-16 04:18:02,644][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.07992963492870331, 'train/grad_norm_pre_clip': 20.82368278503418, 'timing/data': 0.01102446112781763, 'timing/model': 1.5844379309564829, 'learning_rate/qwen_vl_interface': 5.253045421241266e-06, 'learning_rate/action_model': 5.253045421241266e-05, 'epoch': 1.03}) +[2026-06-16 04:18:09,005][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.0858229324221611, 'train/grad_norm_pre_clip': 22.305755615234375, 'timing/data': 0.011414641048759222, 'timing/model': 1.8136988407932222, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 1.03}) +[2026-06-16 04:18:15,058][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.3111274540424347, 'train/grad_norm_pre_clip': 25.52910614013672, 'timing/data': 0.011097101960331202, 'timing/model': 1.5594804463908076, 'learning_rate/qwen_vl_interface': 5.246954578758735e-06, 'learning_rate/action_model': 5.246954578758735e-05, 'epoch': 1.03}) +[2026-06-16 04:18:21,459][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.16274966299533844, 'train/grad_norm_pre_clip': 15.109891891479492, 'timing/data': 0.01120607228949666, 'timing/model': 1.5472885398194194, 'learning_rate/qwen_vl_interface': 5.243909158769327e-06, 'learning_rate/action_model': 5.2439091587693264e-05, 'epoch': 1.03}) +[2026-06-16 04:18:27,758][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.04482385516166687, 'train/grad_norm_pre_clip': 23.471187591552734, 'timing/data': 0.011422412004321814, 'timing/model': 1.5605847192928195, 'learning_rate/qwen_vl_interface': 5.240863741283634e-06, 'learning_rate/action_model': 5.2408637412836336e-05, 'epoch': 1.03}) +[2026-06-16 04:18:34,160][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.03456959128379822, 'train/grad_norm_pre_clip': 12.633881568908691, 'timing/data': 0.010691334959119558, 'timing/model': 1.5583806219510734, 'learning_rate/qwen_vl_interface': 5.2378183275535125e-06, 'learning_rate/action_model': 5.237818327553512e-05, 'epoch': 1.03}) +[2026-06-16 04:18:40,453][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.013829614967107773, 'train/grad_norm_pre_clip': 5.832990646362305, 'timing/data': 0.01108601363375783, 'timing/model': 1.5788589320145547, 'learning_rate/qwen_vl_interface': 5.234772918830815e-06, 'learning_rate/action_model': 5.234772918830815e-05, 'epoch': 1.03}) +[2026-06-16 04:18:46,752][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.015611892566084862, 'train/grad_norm_pre_clip': 3.659303903579712, 'timing/data': 0.0104933250695467, 'timing/model': 1.7617488410323858, 'learning_rate/qwen_vl_interface': 5.231727516367395e-06, 'learning_rate/action_model': 5.231727516367395e-05, 'epoch': 1.03}) +[2026-06-16 04:18:52,871][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.009006354957818985, 'train/grad_norm_pre_clip': 3.794360399246216, 'timing/data': 0.010577996261417866, 'timing/model': 1.5795369539409876, 'learning_rate/qwen_vl_interface': 5.228682121415104e-06, 'learning_rate/action_model': 5.228682121415104e-05, 'epoch': 1.03}) +[2026-06-16 04:18:59,152][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.08328969776630402, 'train/grad_norm_pre_clip': 10.331753730773926, 'timing/data': 0.011337866075336933, 'timing/model': 1.5559099400416017, 'learning_rate/qwen_vl_interface': 5.225636735225789e-06, 'learning_rate/action_model': 5.225636735225788e-05, 'epoch': 1.03}) +[2026-06-16 04:19:05,459][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.10460274666547775, 'train/grad_norm_pre_clip': 30.268049240112305, 'timing/data': 0.011068947147578001, 'timing/model': 1.5681253508664668, 'learning_rate/qwen_vl_interface': 5.22259135905129e-06, 'learning_rate/action_model': 5.2225913590512895e-05, 'epoch': 1.03}) +[2026-06-16 04:19:11,752][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.019782327115535736, 'train/grad_norm_pre_clip': 3.949345827102661, 'timing/data': 0.011116167064756155, 'timing/model': 1.5296544702723622, 'learning_rate/qwen_vl_interface': 5.219545994143448e-06, 'learning_rate/action_model': 5.219545994143448e-05, 'epoch': 1.04}) +[2026-06-16 04:19:18,050][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.051321763545274734, 'train/grad_norm_pre_clip': 17.935495376586914, 'timing/data': 0.00975748896598816, 'timing/model': 1.5179183036088943, 'learning_rate/qwen_vl_interface': 5.216500641754097e-06, 'learning_rate/action_model': 5.216500641754097e-05, 'epoch': 1.04}) +[2026-06-16 04:19:24,358][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.04581980034708977, 'train/grad_norm_pre_clip': 18.90129852294922, 'timing/data': 0.010426218155771494, 'timing/model': 1.7581534036435187, 'learning_rate/qwen_vl_interface': 5.213455303135067e-06, 'learning_rate/action_model': 5.2134553031350673e-05, 'epoch': 1.04}) +[2026-06-16 04:19:30,464][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.1229088306427002, 'train/grad_norm_pre_clip': 21.740427017211914, 'timing/data': 0.01186758792027831, 'timing/model': 1.5692188879474998, 'learning_rate/qwen_vl_interface': 5.2104099795381836e-06, 'learning_rate/action_model': 5.210409979538183e-05, 'epoch': 1.04}) +[2026-06-16 04:19:36,789][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.21069201827049255, 'train/grad_norm_pre_clip': 4.112707138061523, 'timing/data': 0.011801069136708975, 'timing/model': 1.5926885660737753, 'learning_rate/qwen_vl_interface': 5.207364672215261e-06, 'learning_rate/action_model': 5.2073646722152604e-05, 'epoch': 1.04}) +[2026-06-16 04:19:43,156][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.036309678107500076, 'train/grad_norm_pre_clip': 11.509870529174805, 'timing/data': 0.01144554978236556, 'timing/model': 1.5766841522417963, 'learning_rate/qwen_vl_interface': 5.2043193824181095e-06, 'learning_rate/action_model': 5.2043193824181094e-05, 'epoch': 1.04}) +[2026-06-16 04:19:49,461][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.026375584304332733, 'train/grad_norm_pre_clip': 25.292728424072266, 'timing/data': 0.010969599708914757, 'timing/model': 1.5141050186939538, 'learning_rate/qwen_vl_interface': 5.201274111398535e-06, 'learning_rate/action_model': 5.2012741113985356e-05, 'epoch': 1.04}) +[2026-06-16 04:19:55,811][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.06068675220012665, 'train/grad_norm_pre_clip': 14.705443382263184, 'timing/data': 0.011362620163708925, 'timing/model': 1.8172472738660872, 'learning_rate/qwen_vl_interface': 5.198228860408334e-06, 'learning_rate/action_model': 5.198228860408334e-05, 'epoch': 1.04}) +[2026-06-16 04:20:01,857][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.07008669525384903, 'train/grad_norm_pre_clip': 24.32942008972168, 'timing/data': 0.01135740103200078, 'timing/model': 1.5634084763005376, 'learning_rate/qwen_vl_interface': 5.195183630699292e-06, 'learning_rate/action_model': 5.195183630699292e-05, 'epoch': 1.04}) +[2026-06-16 04:20:08,170][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.002164754318073392, 'train/grad_norm_pre_clip': 0.257425457239151, 'timing/data': 0.011458160821348429, 'timing/model': 1.5738679650239646, 'learning_rate/qwen_vl_interface': 5.192138423523189e-06, 'learning_rate/action_model': 5.1921384235231895e-05, 'epoch': 1.04}) +[2026-06-16 04:20:14,463][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.19114910066127777, 'train/grad_norm_pre_clip': 5.449047088623047, 'timing/data': 0.011291441973298788, 'timing/model': 1.5679037552326918, 'learning_rate/qwen_vl_interface': 5.189093240131795e-06, 'learning_rate/action_model': 5.189093240131795e-05, 'epoch': 1.04}) +[2026-06-16 04:20:20,764][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.007334896828979254, 'train/grad_norm_pre_clip': 1.3585585355758667, 'timing/data': 0.011116933077573776, 'timing/model': 1.5680377702228725, 'learning_rate/qwen_vl_interface': 5.186048081776869e-06, 'learning_rate/action_model': 5.1860480817768685e-05, 'epoch': 1.04}) +[2026-06-16 04:20:27,064][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.017508523538708687, 'train/grad_norm_pre_clip': 9.088359832763672, 'timing/data': 0.010881192982196808, 'timing/model': 1.5746423788368702, 'learning_rate/qwen_vl_interface': 5.183002949710161e-06, 'learning_rate/action_model': 5.18300294971016e-05, 'epoch': 1.04}) +[2026-06-16 04:20:33,407][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.007233080454170704, 'train/grad_norm_pre_clip': 1.4158859252929688, 'timing/data': 0.011757212691009045, 'timing/model': 1.8071871511638165, 'learning_rate/qwen_vl_interface': 5.179957845183411e-06, 'learning_rate/action_model': 5.179957845183411e-05, 'epoch': 1.04}) +[2026-06-16 04:20:39,489][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.06676880270242691, 'train/grad_norm_pre_clip': 9.678119659423828, 'timing/data': 0.013557122088968754, 'timing/model': 1.5849326187744737, 'learning_rate/qwen_vl_interface': 5.176912769448343e-06, 'learning_rate/action_model': 5.176912769448343e-05, 'epoch': 1.04}) +[2026-06-16 04:20:45,778][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.014870353043079376, 'train/grad_norm_pre_clip': 3.305461883544922, 'timing/data': 0.01100297411903739, 'timing/model': 1.5565107348375022, 'learning_rate/qwen_vl_interface': 5.173867723756676e-06, 'learning_rate/action_model': 5.173867723756676e-05, 'epoch': 1.04}) +[2026-06-16 04:20:52,069][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.5146892070770264, 'train/grad_norm_pre_clip': 15.831419944763184, 'timing/data': 0.012295933905988932, 'timing/model': 1.5726042366586626, 'learning_rate/qwen_vl_interface': 5.170822709360112e-06, 'learning_rate/action_model': 5.170822709360111e-05, 'epoch': 1.04}) +[2026-06-16 04:20:58,367][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.2132384181022644, 'train/grad_norm_pre_clip': 21.503787994384766, 'timing/data': 0.01109791500493884, 'timing/model': 1.570794939994812, 'learning_rate/qwen_vl_interface': 5.16777772751034e-06, 'learning_rate/action_model': 5.16777772751034e-05, 'epoch': 1.04}) +[2026-06-16 04:21:04,659][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.1986938714981079, 'train/grad_norm_pre_clip': 16.132123947143555, 'timing/data': 0.011868345085531473, 'timing/model': 1.556504005100578, 'learning_rate/qwen_vl_interface': 5.16473277945904e-06, 'learning_rate/action_model': 5.1647327794590396e-05, 'epoch': 1.04}) +[2026-06-16 04:21:11,018][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.02855675481259823, 'train/grad_norm_pre_clip': 34.18177032470703, 'timing/data': 0.011466084979474545, 'timing/model': 1.7924226531758904, 'learning_rate/qwen_vl_interface': 5.161687866457872e-06, 'learning_rate/action_model': 5.161687866457871e-05, 'epoch': 1.04}) +[2026-06-16 04:21:17,065][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.17466318607330322, 'train/grad_norm_pre_clip': 17.931522369384766, 'timing/data': 0.011414225678890944, 'timing/model': 1.5687376740388572, 'learning_rate/qwen_vl_interface': 5.158642989758487e-06, 'learning_rate/action_model': 5.158642989758487e-05, 'epoch': 1.04}) +[2026-06-16 04:21:23,457][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.1375509351491928, 'train/grad_norm_pre_clip': 14.297487258911133, 'timing/data': 0.011040356010198593, 'timing/model': 1.5636290349066257, 'learning_rate/qwen_vl_interface': 5.155598150612519e-06, 'learning_rate/action_model': 5.1555981506125185e-05, 'epoch': 1.04}) +[2026-06-16 04:21:29,764][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.012642603367567062, 'train/grad_norm_pre_clip': 2.0039222240448, 'timing/data': 0.011144217103719711, 'timing/model': 1.5626135421916842, 'learning_rate/qwen_vl_interface': 5.152553350271584e-06, 'learning_rate/action_model': 5.152553350271584e-05, 'epoch': 1.04}) +[2026-06-16 04:21:36,069][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.03875582292675972, 'train/grad_norm_pre_clip': 13.4940185546875, 'timing/data': 0.01114230789244175, 'timing/model': 1.5300890542566776, 'learning_rate/qwen_vl_interface': 5.149508589987286e-06, 'learning_rate/action_model': 5.1495085899872855e-05, 'epoch': 1.04}) +[2026-06-16 04:21:42,361][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.17074590921401978, 'train/grad_norm_pre_clip': 26.526309967041016, 'timing/data': 0.011189907789230347, 'timing/model': 1.5298516191542149, 'learning_rate/qwen_vl_interface': 5.14646387101121e-06, 'learning_rate/action_model': 5.14646387101121e-05, 'epoch': 1.04}) +[2026-06-16 04:21:48,716][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.029667140915989876, 'train/grad_norm_pre_clip': 11.534916877746582, 'timing/data': 0.011687127873301506, 'timing/model': 1.819026743993163, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 1.05}) +[2026-06-16 04:21:54,824][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.05342482775449753, 'train/grad_norm_pre_clip': 16.303245544433594, 'timing/data': 0.011514198035001755, 'timing/model': 1.5709622879512608, 'learning_rate/qwen_vl_interface': 5.140374561989988e-06, 'learning_rate/action_model': 5.140374561989988e-05, 'epoch': 1.05}) +[2026-06-16 04:22:01,062][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.10231180489063263, 'train/grad_norm_pre_clip': 20.286481857299805, 'timing/data': 0.011407459154725075, 'timing/model': 1.559438860975206, 'learning_rate/qwen_vl_interface': 5.137329974447924e-06, 'learning_rate/action_model': 5.137329974447924e-05, 'epoch': 1.05}) +[2026-06-16 04:22:07,365][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.05769415944814682, 'train/grad_norm_pre_clip': 7.2168169021606445, 'timing/data': 0.010513069108128548, 'timing/model': 1.5697965831495821, 'learning_rate/qwen_vl_interface': 5.1342854332202516e-06, 'learning_rate/action_model': 5.1342854332202514e-05, 'epoch': 1.05}) +[2026-06-16 04:22:13,671][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.005679790396243334, 'train/grad_norm_pre_clip': 0.5110284686088562, 'timing/data': 0.014017479959875345, 'timing/model': 1.5659397463314235, 'learning_rate/qwen_vl_interface': 5.131240939558467e-06, 'learning_rate/action_model': 5.131240939558467e-05, 'epoch': 1.05}) +[2026-06-16 04:22:20,015][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.02340809255838394, 'train/grad_norm_pre_clip': 4.21425724029541, 'timing/data': 0.012357749976217747, 'timing/model': 1.5659744669683278, 'learning_rate/qwen_vl_interface': 5.128196494714048e-06, 'learning_rate/action_model': 5.128196494714047e-05, 'epoch': 1.05}) +[2026-06-16 04:22:26,313][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.004295725841075182, 'train/grad_norm_pre_clip': 0.4192632734775543, 'timing/data': 0.013631660025566816, 'timing/model': 1.8052655048668385, 'learning_rate/qwen_vl_interface': 5.125152099938448e-06, 'learning_rate/action_model': 5.125152099938447e-05, 'epoch': 1.05}) +[2026-06-16 04:22:32,362][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.004376943223178387, 'train/grad_norm_pre_clip': 0.4229348301887512, 'timing/data': 0.011851971037685871, 'timing/model': 1.548300825059414, 'learning_rate/qwen_vl_interface': 5.122107756483104e-06, 'learning_rate/action_model': 5.122107756483103e-05, 'epoch': 1.05}) +[2026-06-16 04:22:38,657][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.001224155304953456, 'train/grad_norm_pre_clip': 0.10886778682470322, 'timing/data': 0.011166350916028023, 'timing/model': 1.5626634950749576, 'learning_rate/qwen_vl_interface': 5.119063465599429e-06, 'learning_rate/action_model': 5.1190634655994285e-05, 'epoch': 1.05}) +[2026-06-16 04:22:44,955][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.09090873599052429, 'train/grad_norm_pre_clip': 4.856353759765625, 'timing/data': 0.011062812060117722, 'timing/model': 1.561740973033011, 'learning_rate/qwen_vl_interface': 5.116019228538818e-06, 'learning_rate/action_model': 5.116019228538818e-05, 'epoch': 1.05}) +[2026-06-16 04:22:51,341][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.18390609323978424, 'train/grad_norm_pre_clip': 16.726064682006836, 'timing/data': 0.011888471897691488, 'timing/model': 1.6459792647510767, 'learning_rate/qwen_vl_interface': 5.112975046552641e-06, 'learning_rate/action_model': 5.1129750465526404e-05, 'epoch': 1.05}) +[2026-06-16 04:22:57,756][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.09974019974470139, 'train/grad_norm_pre_clip': 5.629543304443359, 'timing/data': 0.011476782150566578, 'timing/model': 1.508607161231339, 'learning_rate/qwen_vl_interface': 5.109930920892246e-06, 'learning_rate/action_model': 5.1099309208922464e-05, 'epoch': 1.05}) +[2026-06-16 04:23:04,115][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.0775623768568039, 'train/grad_norm_pre_clip': 13.019173622131348, 'timing/data': 0.011791332624852657, 'timing/model': 1.8126361020840704, 'learning_rate/qwen_vl_interface': 5.10688685280896e-06, 'learning_rate/action_model': 5.10688685280896e-05, 'epoch': 1.05}) +[2026-06-16 04:23:10,170][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.23654906451702118, 'train/grad_norm_pre_clip': 12.043289184570312, 'timing/data': 0.011344113387167454, 'timing/model': 1.5624397322535515, 'learning_rate/qwen_vl_interface': 5.103842843554082e-06, 'learning_rate/action_model': 5.103842843554082e-05, 'epoch': 1.05}) +[2026-06-16 04:23:16,587][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.006513943430036306, 'train/grad_norm_pre_clip': 0.5446966290473938, 'timing/data': 0.011381733231246471, 'timing/model': 1.547090856358409, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 1.05}) +[2026-06-16 04:34:18,899][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.007144290488213301, 'train/grad_norm_pre_clip': 0.8478558659553528, 'eval/loss': 0.04203726723790169, 'eval/flap_precision': 0.9590723055934516, 'eval/flap_recall': 0.9590723055934516, 'eval/flap_f1': 0.9590723055934516, 'eval/flap_rate_model': 0.2036111111111111, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9590723055934515, 'timing/data': 0.011123284231871367, 'timing/model': 1.5591028933413327, 'learning_rate/qwen_vl_interface': 5.097755006534638e-06, 'learning_rate/action_model': 5.097755006534638e-05, 'epoch': 1.05}) +[2026-06-16 04:34:18,900][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2600_state +[2026-06-16 04:34:18,906][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 04:34:27,214][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2600_state/model.safetensors +[2026-06-16 04:34:46,608][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2600_state/optimizer.bin +[2026-06-16 04:34:46,608][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2600_state/sampler.bin +[2026-06-16 04:34:46,609][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2600_state/sampler_1.bin +[2026-06-16 04:34:46,610][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2600_state/random_states_0.pkl +[2026-06-16 04:34:46,803][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2600 +[2026-06-16 04:34:47,040][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 04:34:47,055][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 04:34:53,312][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.1384180784225464, 'train/grad_norm_pre_clip': 10.309908866882324, 'timing/data': 0.012549246195703745, 'timing/model': 1.5510427742265165, 'learning_rate/qwen_vl_interface': 5.094711181272551e-06, 'learning_rate/action_model': 5.094711181272551e-05, 'epoch': 1.05}) +[2026-06-16 04:34:59,540][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.008427933789789677, 'train/grad_norm_pre_clip': 1.3279544115066528, 'timing/data': 0.011057662777602673, 'timing/model': 1.566178775858134, 'learning_rate/qwen_vl_interface': 5.091667419843833e-06, 'learning_rate/action_model': 5.091667419843833e-05, 'epoch': 1.05}) +[2026-06-16 04:35:05,802][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.006768302991986275, 'train/grad_norm_pre_clip': 1.2049663066864014, 'timing/data': 0.01080738240852952, 'timing/model': 1.5518981791101396, 'learning_rate/qwen_vl_interface': 5.088623723499657e-06, 'learning_rate/action_model': 5.088623723499657e-05, 'epoch': 1.05}) +[2026-06-16 04:35:12,038][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.2327800989151001, 'train/grad_norm_pre_clip': 17.098974227905273, 'timing/data': 0.011606994085013866, 'timing/model': 1.5125425988808274, 'learning_rate/qwen_vl_interface': 5.085580093491171e-06, 'learning_rate/action_model': 5.085580093491171e-05, 'epoch': 1.05}) +[2026-06-16 04:35:18,271][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.011930158361792564, 'train/grad_norm_pre_clip': 2.453190803527832, 'timing/data': 0.011251572985202074, 'timing/model': 1.7927768789231777, 'learning_rate/qwen_vl_interface': 5.0825365310695e-06, 'learning_rate/action_model': 5.082536531069499e-05, 'epoch': 1.05}) +[2026-06-16 04:35:24,443][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.19543342292308807, 'train/grad_norm_pre_clip': 10.541068077087402, 'timing/data': 0.012314234860241413, 'timing/model': 1.5662669511511922, 'learning_rate/qwen_vl_interface': 5.079493037485735e-06, 'learning_rate/action_model': 5.0794930374857344e-05, 'epoch': 1.05}) +[2026-06-16 04:35:30,744][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.21193356812000275, 'train/grad_norm_pre_clip': 17.2336368560791, 'timing/data': 0.011400722898542881, 'timing/model': 1.560837501194328, 'learning_rate/qwen_vl_interface': 5.076449613990941e-06, 'learning_rate/action_model': 5.076449613990941e-05, 'epoch': 1.05}) +[2026-06-16 04:35:37,050][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.10419217497110367, 'train/grad_norm_pre_clip': 10.458534240722656, 'timing/data': 0.0115235042758286, 'timing/model': 1.5751103721559048, 'learning_rate/qwen_vl_interface': 5.073406261836157e-06, 'learning_rate/action_model': 5.0734062618361565e-05, 'epoch': 1.05}) +[2026-06-16 04:35:43,348][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.00924235861748457, 'train/grad_norm_pre_clip': 0.9828118085861206, 'timing/data': 0.010868123266845942, 'timing/model': 1.562266655266285, 'learning_rate/qwen_vl_interface': 5.070362982272386e-06, 'learning_rate/action_model': 5.070362982272385e-05, 'epoch': 1.06}) +[2026-06-16 04:35:49,722][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.06033189967274666, 'train/grad_norm_pre_clip': 22.20257568359375, 'timing/data': 0.010231470689177513, 'timing/model': 1.5726869469508529, 'learning_rate/qwen_vl_interface': 5.067319776550607e-06, 'learning_rate/action_model': 5.067319776550607e-05, 'epoch': 1.06}) +[2026-06-16 04:35:56,092][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.1226005107164383, 'train/grad_norm_pre_clip': 6.907731056213379, 'timing/data': 0.011260522995144129, 'timing/model': 1.8030831152573228, 'learning_rate/qwen_vl_interface': 5.064276645921768e-06, 'learning_rate/action_model': 5.064276645921768e-05, 'epoch': 1.06}) +[2026-06-16 04:36:02,395][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.013656966388225555, 'train/grad_norm_pre_clip': 4.276030540466309, 'timing/data': 0.011380134150385857, 'timing/model': 1.7993324589915574, 'learning_rate/qwen_vl_interface': 5.061233591636787e-06, 'learning_rate/action_model': 5.061233591636787e-05, 'epoch': 1.06}) +[2026-06-16 04:36:08,459][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.016533061861991882, 'train/grad_norm_pre_clip': 6.1476898193359375, 'timing/data': 0.010890882927924395, 'timing/model': 1.5671764970757067, 'learning_rate/qwen_vl_interface': 5.058190614946544e-06, 'learning_rate/action_model': 5.058190614946544e-05, 'epoch': 1.06}) +[2026-06-16 04:36:14,848][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.25588110089302063, 'train/grad_norm_pre_clip': 18.941373825073242, 'timing/data': 0.010387391783297062, 'timing/model': 1.5573112359270453, 'learning_rate/qwen_vl_interface': 5.055147717101895e-06, 'learning_rate/action_model': 5.055147717101894e-05, 'epoch': 1.06}) +[2026-06-16 04:36:21,218][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.005362863652408123, 'train/grad_norm_pre_clip': 0.6340587735176086, 'timing/data': 0.011867984663695097, 'timing/model': 1.6256508021615446, 'learning_rate/qwen_vl_interface': 5.052104899353658e-06, 'learning_rate/action_model': 5.0521048993536577e-05, 'epoch': 1.06}) +[2026-06-16 04:36:27,459][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.1009206548333168, 'train/grad_norm_pre_clip': 11.585447311401367, 'timing/data': 0.011461264919489622, 'timing/model': 1.5176236401312053, 'learning_rate/qwen_vl_interface': 5.049062162952622e-06, 'learning_rate/action_model': 5.0490621629526214e-05, 'epoch': 1.06}) +[2026-06-16 04:36:33,824][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.00516722584143281, 'train/grad_norm_pre_clip': 1.825231909751892, 'timing/data': 0.011323394253849983, 'timing/model': 1.8193106283433735, 'learning_rate/qwen_vl_interface': 5.046019509149541e-06, 'learning_rate/action_model': 5.046019509149541e-05, 'epoch': 1.06}) +[2026-06-16 04:36:39,851][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.0028449201490730047, 'train/grad_norm_pre_clip': 0.32340216636657715, 'timing/data': 0.011042623315006495, 'timing/model': 1.5525175537914038, 'learning_rate/qwen_vl_interface': 5.042976939195135e-06, 'learning_rate/action_model': 5.0429769391951355e-05, 'epoch': 1.06}) +[2026-06-16 04:36:46,151][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.16047163307666779, 'train/grad_norm_pre_clip': 12.228177070617676, 'timing/data': 0.010291061829775572, 'timing/model': 1.5626172553747892, 'learning_rate/qwen_vl_interface': 5.039934454340091e-06, 'learning_rate/action_model': 5.0399344543400907e-05, 'epoch': 1.06}) +[2026-06-16 04:36:52,450][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.06710153818130493, 'train/grad_norm_pre_clip': 38.75761795043945, 'timing/data': 0.010464772116392851, 'timing/model': 1.5619119931943715, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 1.06}) +[2026-06-16 04:36:58,766][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.089365154504776, 'train/grad_norm_pre_clip': 11.330656051635742, 'timing/data': 0.011021433863788843, 'timing/model': 1.5451604458503425, 'learning_rate/qwen_vl_interface': 5.033849744930647e-06, 'learning_rate/action_model': 5.033849744930646e-05, 'epoch': 1.06}) +[2026-06-16 04:37:05,152][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.0022795340046286583, 'train/grad_norm_pre_clip': 0.48249921202659607, 'timing/data': 0.011137553956359625, 'timing/model': 1.5833662683144212, 'learning_rate/qwen_vl_interface': 5.030807522877443e-06, 'learning_rate/action_model': 5.030807522877443e-05, 'epoch': 1.06}) +[2026-06-16 04:37:11,442][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.013165200129151344, 'train/grad_norm_pre_clip': 4.343814373016357, 'timing/data': 0.01120682479813695, 'timing/model': 1.7468227231875062, 'learning_rate/qwen_vl_interface': 5.027765390925987e-06, 'learning_rate/action_model': 5.027765390925987e-05, 'epoch': 1.06}) +[2026-06-16 04:37:17,457][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.12712052464485168, 'train/grad_norm_pre_clip': 36.356510162353516, 'timing/data': 0.01050925301387906, 'timing/model': 1.5631599808111787, 'learning_rate/qwen_vl_interface': 5.024723350326783e-06, 'learning_rate/action_model': 5.024723350326783e-05, 'epoch': 1.06}) +[2026-06-16 04:37:23,753][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.009506484493613243, 'train/grad_norm_pre_clip': 4.447006702423096, 'timing/data': 0.010699413251131773, 'timing/model': 1.5618603266775608, 'learning_rate/qwen_vl_interface': 5.021681402330303e-06, 'learning_rate/action_model': 5.021681402330303e-05, 'epoch': 1.06}) +[2026-06-16 04:37:30,062][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.08845947682857513, 'train/grad_norm_pre_clip': 22.79100227355957, 'timing/data': 0.011232454795390368, 'timing/model': 1.5454675699584186, 'learning_rate/qwen_vl_interface': 5.018639548186972e-06, 'learning_rate/action_model': 5.0186395481869716e-05, 'epoch': 1.06}) +[2026-06-16 04:37:36,560][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.05091697722673416, 'train/grad_norm_pre_clip': 22.425010681152344, 'timing/data': 0.012483926955610514, 'timing/model': 1.6041108882054687, 'learning_rate/qwen_vl_interface': 5.015597789147184e-06, 'learning_rate/action_model': 5.015597789147184e-05, 'epoch': 1.06}) +[2026-06-16 04:37:42,854][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.32055631279945374, 'train/grad_norm_pre_clip': 41.075408935546875, 'timing/data': 0.01130508491769433, 'timing/model': 1.5216587139293551, 'learning_rate/qwen_vl_interface': 5.012556126461291e-06, 'learning_rate/action_model': 5.012556126461291e-05, 'epoch': 1.06}) +[2026-06-16 04:37:49,168][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.0056652650237083435, 'train/grad_norm_pre_clip': 0.9373920559883118, 'timing/data': 0.011218325234949589, 'timing/model': 1.7766693066805601, 'learning_rate/qwen_vl_interface': 5.0095145613796044e-06, 'learning_rate/action_model': 5.0095145613796046e-05, 'epoch': 1.06}) +[2026-06-16 04:37:55,258][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.00896009523421526, 'train/grad_norm_pre_clip': 4.748448848724365, 'timing/data': 0.010785223916172981, 'timing/model': 1.5708770188502967, 'learning_rate/qwen_vl_interface': 5.0064730951524e-06, 'learning_rate/action_model': 5.0064730951524e-05, 'epoch': 1.06}) +[2026-06-16 04:38:01,779][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.19823499023914337, 'train/grad_norm_pre_clip': 11.098053932189941, 'timing/data': 0.011078834999352694, 'timing/model': 1.5789116048254073, 'learning_rate/qwen_vl_interface': 5.003431729029905e-06, 'learning_rate/action_model': 5.003431729029905e-05, 'epoch': 1.06}) +[2026-06-16 04:38:08,158][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.002993086352944374, 'train/grad_norm_pre_clip': 1.4558422565460205, 'timing/data': 0.011163485236465931, 'timing/model': 1.5656232540495694, 'learning_rate/qwen_vl_interface': 5.000390464262314e-06, 'learning_rate/action_model': 5.0003904642623136e-05, 'epoch': 1.06}) +[2026-06-16 04:38:14,464][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.0009125130018219352, 'train/grad_norm_pre_clip': 0.2772817611694336, 'timing/data': 0.011043725069612265, 'timing/model': 1.531689906027168, 'learning_rate/qwen_vl_interface': 4.997349302099774e-06, 'learning_rate/action_model': 4.997349302099773e-05, 'epoch': 1.06}) +[2026-06-16 04:38:20,909][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.08181807398796082, 'train/grad_norm_pre_clip': 16.676855087280273, 'timing/data': 0.012393638025969267, 'timing/model': 1.8134553260169923, 'learning_rate/qwen_vl_interface': 4.994308243792391e-06, 'learning_rate/action_model': 4.994308243792391e-05, 'epoch': 1.07}) +[2026-06-16 04:38:26,962][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.13246503472328186, 'train/grad_norm_pre_clip': 21.982770919799805, 'timing/data': 0.010530882980674505, 'timing/model': 1.5706553892232478, 'learning_rate/qwen_vl_interface': 4.991267290590233e-06, 'learning_rate/action_model': 4.991267290590233e-05, 'epoch': 1.07}) +[2026-06-16 04:38:33,259][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.004955742042511702, 'train/grad_norm_pre_clip': 2.1699512004852295, 'timing/data': 0.010953165125101805, 'timing/model': 1.5639774920418859, 'learning_rate/qwen_vl_interface': 4.988226443743317e-06, 'learning_rate/action_model': 4.9882264437433174e-05, 'epoch': 1.07}) +[2026-06-16 04:38:39,458][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.11559507995843887, 'train/grad_norm_pre_clip': 8.741462707519531, 'timing/data': 0.01081008417531848, 'timing/model': 1.5609352621249855, 'learning_rate/qwen_vl_interface': 4.985185704501623e-06, 'learning_rate/action_model': 4.985185704501623e-05, 'epoch': 1.07}) +[2026-06-16 04:38:45,767][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.056484635919332504, 'train/grad_norm_pre_clip': 27.4925537109375, 'timing/data': 0.010572274215519428, 'timing/model': 1.5547320661135018, 'learning_rate/qwen_vl_interface': 4.982145074115082e-06, 'learning_rate/action_model': 4.982145074115082e-05, 'epoch': 1.07}) +[2026-06-16 04:38:52,075][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.2138979583978653, 'train/grad_norm_pre_clip': 17.645523071289062, 'timing/data': 0.010951634962111712, 'timing/model': 1.5960166067816317, 'learning_rate/qwen_vl_interface': 4.979104553833584e-06, 'learning_rate/action_model': 4.979104553833584e-05, 'epoch': 1.07}) +[2026-06-16 04:38:58,406][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.08272481709718704, 'train/grad_norm_pre_clip': 25.080305099487305, 'timing/data': 0.01154222572222352, 'timing/model': 1.8030385137535632, 'learning_rate/qwen_vl_interface': 4.9760641449069724e-06, 'learning_rate/action_model': 4.976064144906972e-05, 'epoch': 1.07}) +[2026-06-16 04:39:04,471][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.20510603487491608, 'train/grad_norm_pre_clip': 12.05217456817627, 'timing/data': 0.010648483876138926, 'timing/model': 1.565844161901623, 'learning_rate/qwen_vl_interface': 4.973023848585042e-06, 'learning_rate/action_model': 4.973023848585042e-05, 'epoch': 1.07}) +[2026-06-16 04:39:10,786][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.0013920285273343325, 'train/grad_norm_pre_clip': 0.25484660267829895, 'timing/data': 0.010691385250538588, 'timing/model': 1.583565500099212, 'learning_rate/qwen_vl_interface': 4.969983666117545e-06, 'learning_rate/action_model': 4.9699836661175446e-05, 'epoch': 1.07}) +[2026-06-16 04:39:17,059][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.17699067294597626, 'train/grad_norm_pre_clip': 20.606674194335938, 'timing/data': 0.011306546162813902, 'timing/model': 1.5579122677445412, 'learning_rate/qwen_vl_interface': 4.966943598754185e-06, 'learning_rate/action_model': 4.9669435987541854e-05, 'epoch': 1.07}) +[2026-06-16 04:39:23,366][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.0019737500697374344, 'train/grad_norm_pre_clip': 0.30009588599205017, 'timing/data': 0.012939898762851954, 'timing/model': 1.5657241428270936, 'learning_rate/qwen_vl_interface': 4.963903647744621e-06, 'learning_rate/action_model': 4.96390364774462e-05, 'epoch': 1.07}) +[2026-06-16 04:39:29,665][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.03705751523375511, 'train/grad_norm_pre_clip': 19.0571231842041, 'timing/data': 0.010361064225435257, 'timing/model': 1.5468589663505554, 'learning_rate/qwen_vl_interface': 4.9608638143384604e-06, 'learning_rate/action_model': 4.960863814338461e-05, 'epoch': 1.07}) +[2026-06-16 04:39:35,996][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.10108518600463867, 'train/grad_norm_pre_clip': 29.312904357910156, 'timing/data': 0.01079801470041275, 'timing/model': 1.7875819061882794, 'learning_rate/qwen_vl_interface': 4.9578240997852636e-06, 'learning_rate/action_model': 4.957824099785264e-05, 'epoch': 1.07}) +[2026-06-16 04:39:42,072][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.12156002968549728, 'train/grad_norm_pre_clip': 92.69945526123047, 'timing/data': 0.01091101486235857, 'timing/model': 1.570668593980372, 'learning_rate/qwen_vl_interface': 4.954784505334544e-06, 'learning_rate/action_model': 4.9547845053345434e-05, 'epoch': 1.07}) +[2026-06-16 04:39:48,362][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.08621019870042801, 'train/grad_norm_pre_clip': 13.19629192352295, 'timing/data': 0.010804925113916397, 'timing/model': 1.551208856049925, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 1.07}) +[2026-06-16 04:39:54,678][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.27245858311653137, 'train/grad_norm_pre_clip': 38.54769515991211, 'timing/data': 0.011860707774758339, 'timing/model': 1.5746442070230842, 'learning_rate/qwen_vl_interface': 4.94870568173833e-06, 'learning_rate/action_model': 4.9487056817383307e-05, 'epoch': 1.07}) +[2026-06-16 04:40:01,073][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.28989431262016296, 'train/grad_norm_pre_clip': 17.06589126586914, 'timing/data': 0.011215756181627512, 'timing/model': 1.5415351772680879, 'learning_rate/qwen_vl_interface': 4.945666455091614e-06, 'learning_rate/action_model': 4.9456664550916136e-05, 'epoch': 1.07}) +[2026-06-16 04:40:07,360][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.11137326061725616, 'train/grad_norm_pre_clip': 5.0966339111328125, 'timing/data': 0.010604224167764187, 'timing/model': 1.5512275258079171, 'learning_rate/qwen_vl_interface': 4.942627353544921e-06, 'learning_rate/action_model': 4.942627353544921e-05, 'epoch': 1.07}) +[2026-06-16 04:40:13,722][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.0058836499229073524, 'train/grad_norm_pre_clip': 1.6902166604995728, 'timing/data': 0.011793447192758322, 'timing/model': 1.8179770107381046, 'learning_rate/qwen_vl_interface': 4.939588378347513e-06, 'learning_rate/action_model': 4.9395883783475125e-05, 'epoch': 1.07}) +[2026-06-16 04:40:19,772][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.11335460096597672, 'train/grad_norm_pre_clip': 10.746710777282715, 'timing/data': 0.010003994219005108, 'timing/model': 1.563250444829464, 'learning_rate/qwen_vl_interface': 4.936549530748597e-06, 'learning_rate/action_model': 4.9365495307485963e-05, 'epoch': 1.07}) +[2026-06-16 04:40:26,068][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.10832952708005905, 'train/grad_norm_pre_clip': 7.493088245391846, 'timing/data': 0.010782044846564531, 'timing/model': 1.5765918232500553, 'learning_rate/qwen_vl_interface': 4.9335108119973286e-06, 'learning_rate/action_model': 4.933510811997328e-05, 'epoch': 1.07}) +[2026-06-16 04:40:32,451][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.03282657265663147, 'train/grad_norm_pre_clip': 20.421968460083008, 'timing/data': 0.01074211485683918, 'timing/model': 1.567355948034674, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 1.07}) +[2026-06-16 04:40:38,753][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.04689070209860802, 'train/grad_norm_pre_clip': 13.33823013305664, 'timing/data': 0.011311845853924751, 'timing/model': 1.557444900739938, 'learning_rate/qwen_vl_interface': 4.927433766034093e-06, 'learning_rate/action_model': 4.9274337660340934e-05, 'epoch': 1.07}) +[2026-06-16 04:40:45,148][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.006769424770027399, 'train/grad_norm_pre_clip': 1.3027515411376953, 'timing/data': 0.010512884240597486, 'timing/model': 1.515897758770734, 'learning_rate/qwen_vl_interface': 4.92439544132017e-06, 'learning_rate/action_model': 4.92439544132017e-05, 'epoch': 1.07}) +[2026-06-16 04:40:51,495][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.04222520813345909, 'train/grad_norm_pre_clip': 18.19939613342285, 'timing/data': 0.011129285674542189, 'timing/model': 1.8045896720141172, 'learning_rate/qwen_vl_interface': 4.92135725044998e-06, 'learning_rate/action_model': 4.92135725044998e-05, 'epoch': 1.07}) +[2026-06-16 04:40:57,601][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.17770598828792572, 'train/grad_norm_pre_clip': 17.872682571411133, 'timing/data': 0.010883285198360682, 'timing/model': 1.6038812268525362, 'learning_rate/qwen_vl_interface': 4.918319194672412e-06, 'learning_rate/action_model': 4.918319194672412e-05, 'epoch': 1.08}) +[2026-06-16 04:41:03,856][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.005887126550078392, 'train/grad_norm_pre_clip': 0.8545961976051331, 'timing/data': 0.01118660718202591, 'timing/model': 1.558215145021677, 'learning_rate/qwen_vl_interface': 4.9152812752362935e-06, 'learning_rate/action_model': 4.915281275236293e-05, 'epoch': 1.08}) +[2026-06-16 04:41:10,163][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.0039005111902952194, 'train/grad_norm_pre_clip': 0.44978347420692444, 'timing/data': 0.011611337773501873, 'timing/model': 1.5677607129327953, 'learning_rate/qwen_vl_interface': 4.9122434933904e-06, 'learning_rate/action_model': 4.9122434933904e-05, 'epoch': 1.08}) +[2026-06-16 04:41:16,468][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.13248181343078613, 'train/grad_norm_pre_clip': 6.926389217376709, 'timing/data': 0.011050215922296047, 'timing/model': 1.57219213899225, 'learning_rate/qwen_vl_interface': 4.909205850383449e-06, 'learning_rate/action_model': 4.9092058503834485e-05, 'epoch': 1.08}) +[2026-06-16 04:41:22,757][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.03506278991699219, 'train/grad_norm_pre_clip': 21.569387435913086, 'timing/data': 0.010962535627186298, 'timing/model': 1.5204082718119025, 'learning_rate/qwen_vl_interface': 4.906168347464098e-06, 'learning_rate/action_model': 4.906168347464098e-05, 'epoch': 1.08}) +[2026-06-16 04:41:29,104][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.0021762095857411623, 'train/grad_norm_pre_clip': 0.23863725364208221, 'timing/data': 0.01119955675676465, 'timing/model': 1.8042826289311051, 'learning_rate/qwen_vl_interface': 4.903130985880955e-06, 'learning_rate/action_model': 4.903130985880954e-05, 'epoch': 1.08}) +[2026-06-16 04:41:35,165][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.005387310869991779, 'train/grad_norm_pre_clip': 0.7364721298217773, 'timing/data': 0.011237055994570255, 'timing/model': 1.5700171012431383, 'learning_rate/qwen_vl_interface': 4.900093766882558e-06, 'learning_rate/action_model': 4.900093766882558e-05, 'epoch': 1.08}) +[2026-06-16 04:41:41,489][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.0036375222261995077, 'train/grad_norm_pre_clip': 0.4240926504135132, 'timing/data': 0.014334994368255138, 'timing/model': 1.5703830770216882, 'learning_rate/qwen_vl_interface': 4.897056691717397e-06, 'learning_rate/action_model': 4.8970566917173966e-05, 'epoch': 1.08}) +[2026-06-16 04:41:47,757][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.20294779539108276, 'train/grad_norm_pre_clip': 15.902381896972656, 'timing/data': 0.010995866730809212, 'timing/model': 1.5648479680530727, 'learning_rate/qwen_vl_interface': 4.894019761633898e-06, 'learning_rate/action_model': 4.894019761633898e-05, 'epoch': 1.08}) +[2026-06-16 04:41:54,060][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.1014055386185646, 'train/grad_norm_pre_clip': 9.436120986938477, 'timing/data': 0.011779727879911661, 'timing/model': 1.563825040590018, 'learning_rate/qwen_vl_interface': 4.8909829778804304e-06, 'learning_rate/action_model': 4.89098297788043e-05, 'epoch': 1.08}) +[2026-06-16 04:42:00,453][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.005963304080069065, 'train/grad_norm_pre_clip': 1.6086074113845825, 'timing/data': 0.011320036370307207, 'timing/model': 1.5995629718527198, 'learning_rate/qwen_vl_interface': 4.887946341705298e-06, 'learning_rate/action_model': 4.887946341705297e-05, 'epoch': 1.08}) +[2026-06-16 04:42:06,902][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.009063069708645344, 'train/grad_norm_pre_clip': 10.203178405761719, 'timing/data': 0.014342733658850193, 'timing/model': 1.8215028471313417, 'learning_rate/qwen_vl_interface': 4.8849098543567465e-06, 'learning_rate/action_model': 4.884909854356747e-05, 'epoch': 1.08}) +[2026-06-16 04:42:12,985][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.08125832676887512, 'train/grad_norm_pre_clip': 18.115190505981445, 'timing/data': 0.013196201995015144, 'timing/model': 1.5650758990086615, 'learning_rate/qwen_vl_interface': 4.881873517082965e-06, 'learning_rate/action_model': 4.8818735170829655e-05, 'epoch': 1.08}) +[2026-06-16 04:42:19,262][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.09187845885753632, 'train/grad_norm_pre_clip': 7.029526710510254, 'timing/data': 0.011226936243474483, 'timing/model': 1.5670673693530262, 'learning_rate/qwen_vl_interface': 4.878837331132076e-06, 'learning_rate/action_model': 4.8788373311320754e-05, 'epoch': 1.08}) +[2026-06-16 04:42:25,584][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.09777773171663284, 'train/grad_norm_pre_clip': 12.229127883911133, 'timing/data': 0.0117175979539752, 'timing/model': 1.5908789210952818, 'learning_rate/qwen_vl_interface': 4.87580129775214e-06, 'learning_rate/action_model': 4.87580129775214e-05, 'epoch': 1.08}) +[2026-06-16 04:42:31,948][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.017236461862921715, 'train/grad_norm_pre_clip': 8.359763145446777, 'timing/data': 0.010843626223504543, 'timing/model': 1.6489693950861692, 'learning_rate/qwen_vl_interface': 4.872765418191156e-06, 'learning_rate/action_model': 4.872765418191155e-05, 'epoch': 1.08}) +[2026-06-16 04:42:38,154][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.31038132309913635, 'train/grad_norm_pre_clip': 21.879302978515625, 'timing/data': 0.011155697051435709, 'timing/model': 1.5327931414358318, 'learning_rate/qwen_vl_interface': 4.8697296936970575e-06, 'learning_rate/action_model': 4.869729693697057e-05, 'epoch': 1.08}) +[2026-06-16 04:42:44,545][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.14214251935482025, 'train/grad_norm_pre_clip': 10.901363372802734, 'timing/data': 0.010710734874010086, 'timing/model': 1.840328244958073, 'learning_rate/qwen_vl_interface': 4.8666941255177185e-06, 'learning_rate/action_model': 4.8666941255177185e-05, 'epoch': 1.08}) +[2026-06-16 04:42:50,839][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.04182763770222664, 'train/grad_norm_pre_clip': 21.86735725402832, 'timing/data': 0.011416926980018616, 'timing/model': 1.7319621001370251, 'learning_rate/qwen_vl_interface': 4.863658714900945e-06, 'learning_rate/action_model': 4.863658714900945e-05, 'epoch': 1.08}) +[2026-06-16 04:42:56,964][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.14688092470169067, 'train/grad_norm_pre_clip': 17.921436309814453, 'timing/data': 0.011526728048920631, 'timing/model': 1.5561310909688473, 'learning_rate/qwen_vl_interface': 4.860623463094483e-06, 'learning_rate/action_model': 4.860623463094483e-05, 'epoch': 1.08}) +[2026-06-16 04:43:03,353][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.10739917308092117, 'train/grad_norm_pre_clip': 12.321279525756836, 'timing/data': 0.010238853748887777, 'timing/model': 1.5532042887061834, 'learning_rate/qwen_vl_interface': 4.857588371346006e-06, 'learning_rate/action_model': 4.8575883713460057e-05, 'epoch': 1.08}) +[2026-06-16 04:43:09,657][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.01671777293086052, 'train/grad_norm_pre_clip': 10.116654396057129, 'timing/data': 0.01054572593420744, 'timing/model': 1.5681473389267921, 'learning_rate/qwen_vl_interface': 4.854553440903127e-06, 'learning_rate/action_model': 4.854553440903126e-05, 'epoch': 1.08}) +[2026-06-16 04:43:15,962][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.0019673777278512716, 'train/grad_norm_pre_clip': 0.22715063393115997, 'timing/data': 0.011312717106193304, 'timing/model': 1.5217305412515998, 'learning_rate/qwen_vl_interface': 4.85151867301339e-06, 'learning_rate/action_model': 4.85151867301339e-05, 'epoch': 1.08}) +[2026-06-16 04:43:22,297][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.014651800505816936, 'train/grad_norm_pre_clip': 7.429864883422852, 'timing/data': 0.010661575943231583, 'timing/model': 1.7881415002048016, 'learning_rate/qwen_vl_interface': 4.848484068924274e-06, 'learning_rate/action_model': 4.848484068924275e-05, 'epoch': 1.08}) +[2026-06-16 04:43:28,597][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.002256031846627593, 'train/grad_norm_pre_clip': 0.3730117380619049, 'timing/data': 0.011205357033759356, 'timing/model': 1.7975453180260956, 'learning_rate/qwen_vl_interface': 4.845449629883193e-06, 'learning_rate/action_model': 4.8454496298831934e-05, 'epoch': 1.08}) +[2026-06-16 04:43:34,662][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.28109607100486755, 'train/grad_norm_pre_clip': 13.437514305114746, 'timing/data': 0.01109096733853221, 'timing/model': 1.5648911483585835, 'learning_rate/qwen_vl_interface': 4.842415357137487e-06, 'learning_rate/action_model': 4.842415357137487e-05, 'epoch': 1.09}) +[2026-06-16 04:43:41,041][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.05144592747092247, 'train/grad_norm_pre_clip': 20.929977416992188, 'timing/data': 0.011375876609236002, 'timing/model': 1.6456374693661928, 'learning_rate/qwen_vl_interface': 4.839381251934432e-06, 'learning_rate/action_model': 4.8393812519344316e-05, 'epoch': 1.09}) +[2026-06-16 04:43:47,469][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.03517293930053711, 'train/grad_norm_pre_clip': 11.994458198547363, 'timing/data': 0.01125528709962964, 'timing/model': 1.5567701067775488, 'learning_rate/qwen_vl_interface': 4.836347315521233e-06, 'learning_rate/action_model': 4.836347315521233e-05, 'epoch': 1.09}) +[2026-06-16 04:43:53,979][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.0032266227062791586, 'train/grad_norm_pre_clip': 0.45338740944862366, 'timing/data': 0.011443337891250849, 'timing/model': 1.6126155471429229, 'learning_rate/qwen_vl_interface': 4.833313549145029e-06, 'learning_rate/action_model': 4.833313549145029e-05, 'epoch': 1.09}) +[2026-06-16 04:44:00,311][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.06665254384279251, 'train/grad_norm_pre_clip': 13.882328987121582, 'timing/data': 0.011866508983075619, 'timing/model': 1.7917498582974076, 'learning_rate/qwen_vl_interface': 4.830279954052884e-06, 'learning_rate/action_model': 4.8302799540528836e-05, 'epoch': 1.09}) +[2026-06-16 04:44:06,378][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.013551702722907066, 'train/grad_norm_pre_clip': 9.628154754638672, 'timing/data': 0.010917896870523691, 'timing/model': 1.5729242777451873, 'learning_rate/qwen_vl_interface': 4.827246531491797e-06, 'learning_rate/action_model': 4.827246531491797e-05, 'epoch': 1.09}) +[2026-06-16 04:44:12,776][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.1109650507569313, 'train/grad_norm_pre_clip': 13.442109107971191, 'timing/data': 0.010955417063087225, 'timing/model': 1.5734791629947722, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 1.09}) +[2026-06-16 04:44:19,277][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.07463768869638443, 'train/grad_norm_pre_clip': 10.114642143249512, 'timing/data': 0.011145617347210646, 'timing/model': 1.6490702130831778, 'learning_rate/qwen_vl_interface': 4.821180208950422e-06, 'learning_rate/action_model': 4.821180208950421e-05, 'epoch': 1.09}) +[2026-06-16 04:44:25,579][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.2430376261472702, 'train/grad_norm_pre_clip': 15.905291557312012, 'timing/data': 0.010818866081535816, 'timing/model': 1.5358002800494432, 'learning_rate/qwen_vl_interface': 4.81814731146377e-06, 'learning_rate/action_model': 4.8181473114637695e-05, 'epoch': 1.09}) +[2026-06-16 04:44:31,885][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.05060604587197304, 'train/grad_norm_pre_clip': 11.873199462890625, 'timing/data': 0.01287509175017476, 'timing/model': 1.5950749278999865, 'learning_rate/qwen_vl_interface': 4.8151145914954465e-06, 'learning_rate/action_model': 4.815114591495446e-05, 'epoch': 1.09}) +[2026-06-16 04:44:38,178][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.04437919333577156, 'train/grad_norm_pre_clip': 19.68536376953125, 'timing/data': 0.011456487700343132, 'timing/model': 1.769065079279244, 'learning_rate/qwen_vl_interface': 4.812082050292088e-06, 'learning_rate/action_model': 4.8120820502920876e-05, 'epoch': 1.09}) +[2026-06-16 04:44:44,378][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.00269731180742383, 'train/grad_norm_pre_clip': 0.5135085582733154, 'timing/data': 0.010301865637302399, 'timing/model': 1.5719230282120407, 'learning_rate/qwen_vl_interface': 4.809049689100256e-06, 'learning_rate/action_model': 4.8090496891002555e-05, 'epoch': 1.09}) +[2026-06-16 04:44:50,766][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.07840926200151443, 'train/grad_norm_pre_clip': 21.172367095947266, 'timing/data': 0.010853226762264967, 'timing/model': 1.5641123428940773, 'learning_rate/qwen_vl_interface': 4.80601750916644e-06, 'learning_rate/action_model': 4.8060175091664404e-05, 'epoch': 1.09}) +[2026-06-16 04:44:57,171][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.015151974745094776, 'train/grad_norm_pre_clip': 3.6839473247528076, 'timing/data': 0.010581885930150747, 'timing/model': 1.5694233202375472, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 1.09}) +[2026-06-16 04:45:03,473][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.09004119038581848, 'train/grad_norm_pre_clip': 7.7413434982299805, 'timing/data': 0.014748947229236364, 'timing/model': 1.5300263166427612, 'learning_rate/qwen_vl_interface': 4.799953698058444e-06, 'learning_rate/action_model': 4.7999536980584434e-05, 'epoch': 1.09}) +[2026-06-16 04:45:09,781][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.252121239900589, 'train/grad_norm_pre_clip': 16.802053451538086, 'timing/data': 0.011117047164589167, 'timing/model': 1.5400017048232257, 'learning_rate/qwen_vl_interface': 4.796922069376867e-06, 'learning_rate/action_model': 4.796922069376867e-05, 'epoch': 1.09}) +[2026-06-16 04:56:09,423][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.09676654636859894, 'train/grad_norm_pre_clip': 7.752099990844727, 'eval/loss': 0.03757467865943909, 'eval/flap_precision': 0.9490616621983914, 'eval/flap_recall': 0.965893587994543, 'eval/flap_f1': 0.9574036511156186, 'eval/flap_rate_model': 0.20722222222222222, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9574036511156186, 'timing/data': 0.012264439836144447, 'timing/model': 1.7655667853541672, 'learning_rate/qwen_vl_interface': 4.793890626938511e-06, 'learning_rate/action_model': 4.793890626938511e-05, 'epoch': 1.09}) +[2026-06-16 04:56:09,424][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2700_state +[2026-06-16 04:56:09,430][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 04:56:17,796][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2700_state/model.safetensors +[2026-06-16 04:56:36,905][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2700_state/optimizer.bin +[2026-06-16 04:56:36,907][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2700_state/sampler.bin +[2026-06-16 04:56:36,907][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2700_state/sampler_1.bin +[2026-06-16 04:56:36,908][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2700_state/random_states_0.pkl +[2026-06-16 04:56:37,094][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2700 +[2026-06-16 04:56:37,094][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 04:56:37,110][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 04:56:43,372][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.016140149906277657, 'train/grad_norm_pre_clip': 3.193692922592163, 'timing/data': 0.011294509749859571, 'timing/model': 1.8123154798522592, 'learning_rate/qwen_vl_interface': 4.7908593719894914e-06, 'learning_rate/action_model': 4.790859371989491e-05, 'epoch': 1.09}) +[2026-06-16 04:56:49,514][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.005508176051080227, 'train/grad_norm_pre_clip': 0.6747173070907593, 'timing/data': 0.010639057029038668, 'timing/model': 1.6624259920790792, 'learning_rate/qwen_vl_interface': 4.78782830577584e-06, 'learning_rate/action_model': 4.7878283057758396e-05, 'epoch': 1.09}) +[2026-06-16 04:56:55,535][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.02129633165895939, 'train/grad_norm_pre_clip': 8.46349048614502, 'timing/data': 0.010808757971972227, 'timing/model': 1.5583715219981968, 'learning_rate/qwen_vl_interface': 4.784797429543513e-06, 'learning_rate/action_model': 4.784797429543513e-05, 'epoch': 1.09}) +[2026-06-16 04:57:01,738][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.0024802950210869312, 'train/grad_norm_pre_clip': 0.8109813332557678, 'timing/data': 0.01076866826042533, 'timing/model': 1.5665648239664733, 'learning_rate/qwen_vl_interface': 4.781766744538393e-06, 'learning_rate/action_model': 4.781766744538393e-05, 'epoch': 1.09}) +[2026-06-16 04:57:07,936][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.0952497348189354, 'train/grad_norm_pre_clip': 9.832376480102539, 'timing/data': 0.011381108779460192, 'timing/model': 1.5498956218361855, 'learning_rate/qwen_vl_interface': 4.778736252006277e-06, 'learning_rate/action_model': 4.7787362520062765e-05, 'epoch': 1.09}) +[2026-06-16 04:57:14,143][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.07411143183708191, 'train/grad_norm_pre_clip': 11.470633506774902, 'timing/data': 0.010717297904193401, 'timing/model': 1.5292554488405585, 'learning_rate/qwen_vl_interface': 4.775705953192885e-06, 'learning_rate/action_model': 4.7757059531928856e-05, 'epoch': 1.09}) +[2026-06-16 04:57:20,347][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.0665893480181694, 'train/grad_norm_pre_clip': 12.738626480102539, 'timing/data': 0.011148768942803144, 'timing/model': 1.499630912207067, 'learning_rate/qwen_vl_interface': 4.7726758493438615e-06, 'learning_rate/action_model': 4.7726758493438613e-05, 'epoch': 1.09}) +[2026-06-16 04:57:26,695][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.00521730026230216, 'train/grad_norm_pre_clip': 0.927311897277832, 'timing/data': 0.010380066931247711, 'timing/model': 1.8199164071120322, 'learning_rate/qwen_vl_interface': 4.769645941704766e-06, 'learning_rate/action_model': 4.769645941704766e-05, 'epoch': 1.1}) +[2026-06-16 04:57:32,746][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.05044345557689667, 'train/grad_norm_pre_clip': 13.966076850891113, 'timing/data': 0.011556090321391821, 'timing/model': 1.5026490720920265, 'learning_rate/qwen_vl_interface': 4.766616231521081e-06, 'learning_rate/action_model': 4.7666162315210806e-05, 'epoch': 1.1}) +[2026-06-16 04:57:39,057][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.0036371732130646706, 'train/grad_norm_pre_clip': 0.5278377532958984, 'timing/data': 0.01151362992823124, 'timing/model': 1.5756507818587124, 'learning_rate/qwen_vl_interface': 4.763586720038202e-06, 'learning_rate/action_model': 4.7635867200382014e-05, 'epoch': 1.1}) +[2026-06-16 04:57:45,379][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.006113962735980749, 'train/grad_norm_pre_clip': 0.7526600360870361, 'timing/data': 0.011034069117158651, 'timing/model': 1.587560093961656, 'learning_rate/qwen_vl_interface': 4.76055740850145e-06, 'learning_rate/action_model': 4.76055740850145e-05, 'epoch': 1.1}) +[2026-06-16 04:57:51,757][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.19922250509262085, 'train/grad_norm_pre_clip': 20.05118751525879, 'timing/data': 0.014121877029538155, 'timing/model': 1.5289953220635653, 'learning_rate/qwen_vl_interface': 4.757528298156058e-06, 'learning_rate/action_model': 4.757528298156058e-05, 'epoch': 1.1}) +[2026-06-16 04:57:58,043][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.006128114648163319, 'train/grad_norm_pre_clip': 0.8885789513587952, 'timing/data': 0.01104540890082717, 'timing/model': 1.5505611589178443, 'learning_rate/qwen_vl_interface': 4.754499390247181e-06, 'learning_rate/action_model': 4.754499390247181e-05, 'epoch': 1.1}) +[2026-06-16 04:58:04,382][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.2610085904598236, 'train/grad_norm_pre_clip': 24.402462005615234, 'timing/data': 0.010591157246381044, 'timing/model': 1.8087429031729698, 'learning_rate/qwen_vl_interface': 4.751470686019887e-06, 'learning_rate/action_model': 4.751470686019888e-05, 'epoch': 1.1}) +[2026-06-16 04:58:10,447][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.12589944899082184, 'train/grad_norm_pre_clip': 10.456634521484375, 'timing/data': 0.010779818054288626, 'timing/model': 1.5581308109685779, 'learning_rate/qwen_vl_interface': 4.7484421867191645e-06, 'learning_rate/action_model': 4.748442186719165e-05, 'epoch': 1.1}) +[2026-06-16 04:58:16,753][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.003662087954580784, 'train/grad_norm_pre_clip': 0.7670418620109558, 'timing/data': 0.011083218734711409, 'timing/model': 1.563184994738549, 'learning_rate/qwen_vl_interface': 4.745413893589913e-06, 'learning_rate/action_model': 4.7454138935899125e-05, 'epoch': 1.1}) +[2026-06-16 04:58:23,067][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.43427497148513794, 'train/grad_norm_pre_clip': 15.504987716674805, 'timing/data': 0.011100118979811668, 'timing/model': 1.5794228170998394, 'learning_rate/qwen_vl_interface': 4.7423858078769506e-06, 'learning_rate/action_model': 4.742385807876951e-05, 'epoch': 1.1}) +[2026-06-16 04:58:29,346][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.02019856683909893, 'train/grad_norm_pre_clip': 6.358919620513916, 'timing/data': 0.013528964947909117, 'timing/model': 1.5265469308942556, 'learning_rate/qwen_vl_interface': 4.739357930825008e-06, 'learning_rate/action_model': 4.739357930825008e-05, 'epoch': 1.1}) +[2026-06-16 04:58:35,650][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.0013597558718174696, 'train/grad_norm_pre_clip': 0.4452205300331116, 'timing/data': 0.010027776937931776, 'timing/model': 1.5632654363289475, 'learning_rate/qwen_vl_interface': 4.736330263678731e-06, 'learning_rate/action_model': 4.736330263678731e-05, 'epoch': 1.1}) +[2026-06-16 04:58:41,981][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.04598447307944298, 'train/grad_norm_pre_clip': 16.272489547729492, 'timing/data': 0.010583567898720503, 'timing/model': 1.782063317950815, 'learning_rate/qwen_vl_interface': 4.733302807682679e-06, 'learning_rate/action_model': 4.733302807682679e-05, 'epoch': 1.1}) +[2026-06-16 04:58:48,148][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.09079723060131073, 'train/grad_norm_pre_clip': 13.49507999420166, 'timing/data': 0.010844348929822445, 'timing/model': 1.6570068830624223, 'learning_rate/qwen_vl_interface': 4.7302755640813255e-06, 'learning_rate/action_model': 4.7302755640813256e-05, 'epoch': 1.1}) +[2026-06-16 04:58:54,447][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.3190005123615265, 'train/grad_norm_pre_clip': 18.823944091796875, 'timing/data': 0.01104010920971632, 'timing/model': 1.5537059949710965, 'learning_rate/qwen_vl_interface': 4.727248534119058e-06, 'learning_rate/action_model': 4.7272485341190574e-05, 'epoch': 1.1}) +[2026-06-16 04:59:00,762][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.013969997875392437, 'train/grad_norm_pre_clip': 1.986730933189392, 'timing/data': 0.010924058966338634, 'timing/model': 1.5773825268261135, 'learning_rate/qwen_vl_interface': 4.72422171904017e-06, 'learning_rate/action_model': 4.7242217190401697e-05, 'epoch': 1.1}) +[2026-06-16 04:59:07,057][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.1730508804321289, 'train/grad_norm_pre_clip': 14.792424201965332, 'timing/data': 0.010622826870530844, 'timing/model': 1.566204928793013, 'learning_rate/qwen_vl_interface': 4.721195120088873e-06, 'learning_rate/action_model': 4.721195120088873e-05, 'epoch': 1.1}) +[2026-06-16 04:59:13,359][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.0008978805271908641, 'train/grad_norm_pre_clip': 0.22247907519340515, 'timing/data': 0.010965189430862665, 'timing/model': 1.533788023982197, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 1.1}) +[2026-06-16 04:59:19,887][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.1458296924829483, 'train/grad_norm_pre_clip': 13.716922760009766, 'timing/data': 0.0112808789126575, 'timing/model': 1.796205730177462, 'learning_rate/qwen_vl_interface': 4.715142575545445e-06, 'learning_rate/action_model': 4.7151425755454454e-05, 'epoch': 1.1}) +[2026-06-16 04:59:25,992][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.013268798589706421, 'train/grad_norm_pre_clip': 5.449525833129883, 'timing/data': 0.011961451265960932, 'timing/model': 1.5904580228962004, 'learning_rate/qwen_vl_interface': 4.712116632441287e-06, 'learning_rate/action_model': 4.7121166324412866e-05, 'epoch': 1.1}) +[2026-06-16 04:59:32,296][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.0008239879971370101, 'train/grad_norm_pre_clip': 0.25600099563598633, 'timing/data': 0.011272819712758064, 'timing/model': 1.5883015291765332, 'learning_rate/qwen_vl_interface': 4.709090910440664e-06, 'learning_rate/action_model': 4.7090909104406634e-05, 'epoch': 1.1}) +[2026-06-16 04:59:38,659][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.23721592128276825, 'train/grad_norm_pre_clip': 20.144132614135742, 'timing/data': 0.011768191121518612, 'timing/model': 1.5652474793605506, 'learning_rate/qwen_vl_interface': 4.7060654107873335e-06, 'learning_rate/action_model': 4.7060654107873335e-05, 'epoch': 1.1}) +[2026-06-16 04:59:44,961][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.0024009165354073048, 'train/grad_norm_pre_clip': 1.3699644804000854, 'timing/data': 0.010702097788453102, 'timing/model': 1.561041769105941, 'learning_rate/qwen_vl_interface': 4.7030401347249685e-06, 'learning_rate/action_model': 4.703040134724968e-05, 'epoch': 1.1}) +[2026-06-16 04:59:51,253][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.09635975956916809, 'train/grad_norm_pre_clip': 64.29679870605469, 'timing/data': 0.010603998322039843, 'timing/model': 1.5266200699843466, 'learning_rate/qwen_vl_interface': 4.700015083497143e-06, 'learning_rate/action_model': 4.700015083497142e-05, 'epoch': 1.1}) +[2026-06-16 04:59:57,594][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.002946803579106927, 'train/grad_norm_pre_clip': 0.9597852230072021, 'timing/data': 0.011079038958996534, 'timing/model': 1.7955303331837058, 'learning_rate/qwen_vl_interface': 4.696990258347342e-06, 'learning_rate/action_model': 4.696990258347342e-05, 'epoch': 1.1}) +[2026-06-16 05:00:03,657][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.37025612592697144, 'train/grad_norm_pre_clip': 33.13657760620117, 'timing/data': 0.011115889064967632, 'timing/model': 1.5600737077184021, 'learning_rate/qwen_vl_interface': 4.693965660518957e-06, 'learning_rate/action_model': 4.693965660518957e-05, 'epoch': 1.11}) +[2026-06-16 05:00:09,961][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.051418840885162354, 'train/grad_norm_pre_clip': 24.805688858032227, 'timing/data': 0.011050978675484657, 'timing/model': 1.5675486782565713, 'learning_rate/qwen_vl_interface': 4.690941291255285e-06, 'learning_rate/action_model': 4.6909412912552844e-05, 'epoch': 1.11}) +[2026-06-16 05:00:16,251][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.17546482384204865, 'train/grad_norm_pre_clip': 16.300186157226562, 'timing/data': 0.011054549366235733, 'timing/model': 1.566507055889815, 'learning_rate/qwen_vl_interface': 4.68791715179953e-06, 'learning_rate/action_model': 4.68791715179953e-05, 'epoch': 1.11}) +[2026-06-16 05:00:22,559][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.002786946250125766, 'train/grad_norm_pre_clip': 0.5343531370162964, 'timing/data': 0.012250091880559921, 'timing/model': 1.5712550389580429, 'learning_rate/qwen_vl_interface': 4.684893243394804e-06, 'learning_rate/action_model': 4.6848932433948036e-05, 'epoch': 1.11}) +[2026-06-16 05:00:28,851][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.002652981784194708, 'train/grad_norm_pre_clip': 0.45794743299484253, 'timing/data': 0.011275520082563162, 'timing/model': 1.520102655980736, 'learning_rate/qwen_vl_interface': 4.68186956728412e-06, 'learning_rate/action_model': 4.68186956728412e-05, 'epoch': 1.11}) +[2026-06-16 05:00:35,181][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.1502406895160675, 'train/grad_norm_pre_clip': 19.53920555114746, 'timing/data': 0.010789147578179836, 'timing/model': 1.7948710448108613, 'learning_rate/qwen_vl_interface': 4.678846124710395e-06, 'learning_rate/action_model': 4.678846124710394e-05, 'epoch': 1.11}) +[2026-06-16 05:00:41,263][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.039091672748327255, 'train/grad_norm_pre_clip': 13.488941192626953, 'timing/data': 0.012320681940764189, 'timing/model': 1.5650077350437641, 'learning_rate/qwen_vl_interface': 4.675822916916454e-06, 'learning_rate/action_model': 4.6758229169164536e-05, 'epoch': 1.11}) +[2026-06-16 05:00:47,665][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.0003421947476454079, 'train/grad_norm_pre_clip': 0.0360579639673233, 'timing/data': 0.010826378129422665, 'timing/model': 1.5676555521786213, 'learning_rate/qwen_vl_interface': 4.6727999451450236e-06, 'learning_rate/action_model': 4.672799945145023e-05, 'epoch': 1.11}) +[2026-06-16 05:00:53,966][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.07738907635211945, 'train/grad_norm_pre_clip': 5.7231340408325195, 'timing/data': 0.010792688000947237, 'timing/model': 1.5678332829847932, 'learning_rate/qwen_vl_interface': 4.669777210638733e-06, 'learning_rate/action_model': 4.669777210638733e-05, 'epoch': 1.11}) +[2026-06-16 05:01:00,360][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.03141302242875099, 'train/grad_norm_pre_clip': 20.648433685302734, 'timing/data': 0.010424746666103601, 'timing/model': 1.567768534179777, 'learning_rate/qwen_vl_interface': 4.666754714640115e-06, 'learning_rate/action_model': 4.666754714640115e-05, 'epoch': 1.11}) +[2026-06-16 05:01:06,654][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.09207909554243088, 'train/grad_norm_pre_clip': 19.001651763916016, 'timing/data': 0.010558037087321281, 'timing/model': 1.5335939647629857, 'learning_rate/qwen_vl_interface': 4.663732458391602e-06, 'learning_rate/action_model': 4.663732458391601e-05, 'epoch': 1.11}) +[2026-06-16 05:01:12,993][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.2251654416322708, 'train/grad_norm_pre_clip': 25.825191497802734, 'timing/data': 0.010475647170096636, 'timing/model': 1.8004233939573169, 'learning_rate/qwen_vl_interface': 4.6607104431355295e-06, 'learning_rate/action_model': 4.66071044313553e-05, 'epoch': 1.11}) +[2026-06-16 05:01:19,318][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.00722108781337738, 'train/grad_norm_pre_clip': 0.9195191860198975, 'timing/data': 0.010430428199470043, 'timing/model': 1.732898537069559, 'learning_rate/qwen_vl_interface': 4.657688670114135e-06, 'learning_rate/action_model': 4.657688670114135e-05, 'epoch': 1.11}) +[2026-06-16 05:01:25,357][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.010788913816213608, 'train/grad_norm_pre_clip': 4.67050838470459, 'timing/data': 0.010335687082260847, 'timing/model': 1.5566781261004508, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 1.11}) +[2026-06-16 05:01:31,654][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.29857760667800903, 'train/grad_norm_pre_clip': 22.53285789489746, 'timing/data': 0.010337466839700937, 'timing/model': 1.565881400834769, 'learning_rate/qwen_vl_interface': 4.651645855743825e-06, 'learning_rate/action_model': 4.651645855743825e-05, 'epoch': 1.11}) +[2026-06-16 05:01:37,958][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.22966688871383667, 'train/grad_norm_pre_clip': 14.499409675598145, 'timing/data': 0.010862048249691725, 'timing/model': 1.5666502444073558, 'learning_rate/qwen_vl_interface': 4.648624816878883e-06, 'learning_rate/action_model': 4.6486248168788825e-05, 'epoch': 1.11}) +[2026-06-16 05:01:44,377][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.015039048157632351, 'train/grad_norm_pre_clip': 8.762303352355957, 'timing/data': 0.010654948186129332, 'timing/model': 1.6461217617616057, 'learning_rate/qwen_vl_interface': 4.6456040252165615e-06, 'learning_rate/action_model': 4.645604025216562e-05, 'epoch': 1.11}) +[2026-06-16 05:01:50,682][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.0052622645162045956, 'train/grad_norm_pre_clip': 0.8948531746864319, 'timing/data': 0.010450006928294897, 'timing/model': 1.7782731992192566, 'learning_rate/qwen_vl_interface': 4.642583481998596e-06, 'learning_rate/action_model': 4.642583481998596e-05, 'epoch': 1.11}) +[2026-06-16 05:01:56,991][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.11208584159612656, 'train/grad_norm_pre_clip': 6.065742015838623, 'timing/data': 0.01091544795781374, 'timing/model': 1.793948540929705, 'learning_rate/qwen_vl_interface': 4.639563188466616e-06, 'learning_rate/action_model': 4.6395631884666165e-05, 'epoch': 1.11}) +[2026-06-16 05:02:03,055][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.12233001738786697, 'train/grad_norm_pre_clip': 16.907405853271484, 'timing/data': 0.010234196670353413, 'timing/model': 1.5625645848922431, 'learning_rate/qwen_vl_interface': 4.636543145862152e-06, 'learning_rate/action_model': 4.6365431458621523e-05, 'epoch': 1.11}) +[2026-06-16 05:02:09,453][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.05847850441932678, 'train/grad_norm_pre_clip': 10.372556686401367, 'timing/data': 0.010386596899479628, 'timing/model': 1.5597348283044994, 'learning_rate/qwen_vl_interface': 4.633523355426629e-06, 'learning_rate/action_model': 4.6335233554266295e-05, 'epoch': 1.11}) +[2026-06-16 05:02:15,762][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.012127475813031197, 'train/grad_norm_pre_clip': 8.01665210723877, 'timing/data': 0.011283748783171177, 'timing/model': 1.570057176053524, 'learning_rate/qwen_vl_interface': 4.6305038184013675e-06, 'learning_rate/action_model': 4.6305038184013675e-05, 'epoch': 1.11}) +[2026-06-16 05:02:22,060][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.0029206702020019293, 'train/grad_norm_pre_clip': 0.6844347715377808, 'timing/data': 0.013348115142434835, 'timing/model': 1.5187366921454668, 'learning_rate/qwen_vl_interface': 4.627484536027589e-06, 'learning_rate/action_model': 4.627484536027588e-05, 'epoch': 1.11}) +[2026-06-16 05:02:28,400][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.0020113999489694834, 'train/grad_norm_pre_clip': 0.3008352816104889, 'timing/data': 0.01061524823307991, 'timing/model': 1.7995680277235806, 'learning_rate/qwen_vl_interface': 4.624465509546399e-06, 'learning_rate/action_model': 4.6244655095463986e-05, 'epoch': 1.11}) +[2026-06-16 05:02:34,460][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.06337758898735046, 'train/grad_norm_pre_clip': 12.902094841003418, 'timing/data': 0.011878462042659521, 'timing/model': 1.5588246169500053, 'learning_rate/qwen_vl_interface': 4.6214467401988106e-06, 'learning_rate/action_model': 4.621446740198811e-05, 'epoch': 1.11}) +[2026-06-16 05:02:40,855][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.09943610429763794, 'train/grad_norm_pre_clip': 46.739952087402344, 'timing/data': 0.010256136767566204, 'timing/model': 1.566057417076081, 'learning_rate/qwen_vl_interface': 4.6184282292257245e-06, 'learning_rate/action_model': 4.618428229225725e-05, 'epoch': 1.12}) +[2026-06-16 05:02:47,180][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.01220705360174179, 'train/grad_norm_pre_clip': 5.397213935852051, 'timing/data': 0.0111206891015172, 'timing/model': 1.5849395468831062, 'learning_rate/qwen_vl_interface': 4.615409977867938e-06, 'learning_rate/action_model': 4.615409977867937e-05, 'epoch': 1.12}) +[2026-06-16 05:02:53,636][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.1456746757030487, 'train/grad_norm_pre_clip': 29.872146606445312, 'timing/data': 0.0126327034085989, 'timing/model': 1.6068016537465155, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 1.12}) +[2026-06-16 05:02:59,862][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.003120392095297575, 'train/grad_norm_pre_clip': 0.8762797117233276, 'timing/data': 0.010944608598947525, 'timing/model': 1.575661103706807, 'learning_rate/qwen_vl_interface': 4.609374258960908e-06, 'learning_rate/action_model': 4.6093742589609074e-05, 'epoch': 1.12}) +[2026-06-16 05:03:06,146][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.13554994761943817, 'train/grad_norm_pre_clip': 8.485617637634277, 'timing/data': 0.010683577973395586, 'timing/model': 1.7553873951546848, 'learning_rate/qwen_vl_interface': 4.60635679389272e-06, 'learning_rate/action_model': 4.60635679389272e-05, 'epoch': 1.12}) +[2026-06-16 05:03:12,259][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.0816311314702034, 'train/grad_norm_pre_clip': 19.77212142944336, 'timing/data': 0.011340719647705555, 'timing/model': 1.5590387810952961, 'learning_rate/qwen_vl_interface': 4.603339593401943e-06, 'learning_rate/action_model': 4.603339593401942e-05, 'epoch': 1.12}) +[2026-06-16 05:03:18,564][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.16668492555618286, 'train/grad_norm_pre_clip': 21.112903594970703, 'timing/data': 0.011324409861117601, 'timing/model': 1.5664457301609218, 'learning_rate/qwen_vl_interface': 4.600322658728832e-06, 'learning_rate/action_model': 4.600322658728832e-05, 'epoch': 1.12}) +[2026-06-16 05:03:24,855][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.07332062721252441, 'train/grad_norm_pre_clip': 14.688814163208008, 'timing/data': 0.00985518516972661, 'timing/model': 1.5637045241892338, 'learning_rate/qwen_vl_interface': 4.597305991113537e-06, 'learning_rate/action_model': 4.597305991113537e-05, 'epoch': 1.12}) +[2026-06-16 05:03:31,232][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.10134056210517883, 'train/grad_norm_pre_clip': 11.823309898376465, 'timing/data': 0.010524658020585775, 'timing/model': 1.6097948751412332, 'learning_rate/qwen_vl_interface': 4.5942895917960935e-06, 'learning_rate/action_model': 4.5942895917960935e-05, 'epoch': 1.12}) +[2026-06-16 05:03:37,553][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.08154848963022232, 'train/grad_norm_pre_clip': 15.998866081237793, 'timing/data': 0.011717521119862795, 'timing/model': 1.5496628871187568, 'learning_rate/qwen_vl_interface': 4.59127346201643e-06, 'learning_rate/action_model': 4.59127346201643e-05, 'epoch': 1.12}) +[2026-06-16 05:03:43,858][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.0014324791263788939, 'train/grad_norm_pre_clip': 0.3034079074859619, 'timing/data': 0.011173549108207226, 'timing/model': 1.7601274908520281, 'learning_rate/qwen_vl_interface': 4.588257603014364e-06, 'learning_rate/action_model': 4.588257603014364e-05, 'epoch': 1.12}) +[2026-06-16 05:03:49,960][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.12128867954015732, 'train/grad_norm_pre_clip': 16.855388641357422, 'timing/data': 0.011098828632384539, 'timing/model': 1.5380992679856718, 'learning_rate/qwen_vl_interface': 4.585242016029602e-06, 'learning_rate/action_model': 4.5852420160296015e-05, 'epoch': 1.12}) +[2026-06-16 05:03:56,261][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.059976205229759216, 'train/grad_norm_pre_clip': 16.470888137817383, 'timing/data': 0.011384259443730116, 'timing/model': 1.5546075217425823, 'learning_rate/qwen_vl_interface': 4.582226702301737e-06, 'learning_rate/action_model': 4.5822267023017375e-05, 'epoch': 1.12}) +[2026-06-16 05:04:02,591][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.12102726846933365, 'train/grad_norm_pre_clip': 32.39345932006836, 'timing/data': 0.011207709088921547, 'timing/model': 1.5524932267144322, 'learning_rate/qwen_vl_interface': 4.57921166307025e-06, 'learning_rate/action_model': 4.57921166307025e-05, 'epoch': 1.12}) +[2026-06-16 05:04:08,859][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.18087583780288696, 'train/grad_norm_pre_clip': 28.012836456298828, 'timing/data': 0.011581419967114925, 'timing/model': 1.5273374007083476, 'learning_rate/qwen_vl_interface': 4.576196899574511e-06, 'learning_rate/action_model': 4.576196899574511e-05, 'epoch': 1.12}) +[2026-06-16 05:04:15,214][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.1381717324256897, 'train/grad_norm_pre_clip': 11.390524864196777, 'timing/data': 0.011731750797480345, 'timing/model': 1.8160868086852133, 'learning_rate/qwen_vl_interface': 4.573182413053775e-06, 'learning_rate/action_model': 4.5731824130537746e-05, 'epoch': 1.12}) +[2026-06-16 05:04:21,264][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.140680730342865, 'train/grad_norm_pre_clip': 18.43919563293457, 'timing/data': 0.013105044141411781, 'timing/model': 1.5602125977165997, 'learning_rate/qwen_vl_interface': 4.570168204747183e-06, 'learning_rate/action_model': 4.570168204747183e-05, 'epoch': 1.12}) +[2026-06-16 05:04:27,560][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.014061637222766876, 'train/grad_norm_pre_clip': 2.5854663848876953, 'timing/data': 0.010052626021206379, 'timing/model': 1.556018847040832, 'learning_rate/qwen_vl_interface': 4.567154275893762e-06, 'learning_rate/action_model': 4.5671542758937624e-05, 'epoch': 1.12}) +[2026-06-16 05:04:33,857][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.11069148033857346, 'train/grad_norm_pre_clip': 20.328960418701172, 'timing/data': 0.011158110108226538, 'timing/model': 1.5611284906044602, 'learning_rate/qwen_vl_interface': 4.564140627732426e-06, 'learning_rate/action_model': 4.564140627732426e-05, 'epoch': 1.12}) +[2026-06-16 05:04:40,163][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.2920648455619812, 'train/grad_norm_pre_clip': 16.757755279541016, 'timing/data': 0.010467207059264183, 'timing/model': 1.5720330500043929, 'learning_rate/qwen_vl_interface': 4.56112726150197e-06, 'learning_rate/action_model': 4.56112726150197e-05, 'epoch': 1.12}) +[2026-06-16 05:04:46,557][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.16476629674434662, 'train/grad_norm_pre_clip': 20.33523941040039, 'timing/data': 0.011846851091831923, 'timing/model': 1.6733189658261836, 'learning_rate/qwen_vl_interface': 4.558114178441076e-06, 'learning_rate/action_model': 4.558114178441076e-05, 'epoch': 1.12}) +[2026-06-16 05:04:52,894][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.011433500796556473, 'train/grad_norm_pre_clip': 4.83819580078125, 'timing/data': 0.010803997982293367, 'timing/model': 1.790890146046877, 'learning_rate/qwen_vl_interface': 4.555101379788308e-06, 'learning_rate/action_model': 4.555101379788308e-05, 'epoch': 1.12}) +[2026-06-16 05:04:59,063][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.12088391184806824, 'train/grad_norm_pre_clip': 12.245763778686523, 'timing/data': 0.011113439686596394, 'timing/model': 1.5677504688501358, 'learning_rate/qwen_vl_interface': 4.552088866782115e-06, 'learning_rate/action_model': 4.552088866782115e-05, 'epoch': 1.12}) +[2026-06-16 05:05:05,361][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.005415514577180147, 'train/grad_norm_pre_clip': 1.1279242038726807, 'timing/data': 0.01052823755890131, 'timing/model': 1.560295820236206, 'learning_rate/qwen_vl_interface': 4.5490766406608265e-06, 'learning_rate/action_model': 4.549076640660827e-05, 'epoch': 1.12}) +[2026-06-16 05:05:11,757][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.02108759805560112, 'train/grad_norm_pre_clip': 8.86742877960205, 'timing/data': 0.010758008807897568, 'timing/model': 1.5587980570271611, 'learning_rate/qwen_vl_interface': 4.546064702662655e-06, 'learning_rate/action_model': 4.546064702662655e-05, 'epoch': 1.13}) +[2026-06-16 05:05:18,166][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.001689717173576355, 'train/grad_norm_pre_clip': 0.1741228550672531, 'timing/data': 0.011121238581836224, 'timing/model': 1.568125631660223, 'learning_rate/qwen_vl_interface': 4.543053054025693e-06, 'learning_rate/action_model': 4.5430530540256926e-05, 'epoch': 1.13}) +[2026-06-16 05:05:24,454][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.29087182879447937, 'train/grad_norm_pre_clip': 21.205991744995117, 'timing/data': 0.010088296141475439, 'timing/model': 1.5572896637022495, 'learning_rate/qwen_vl_interface': 4.540041695987917e-06, 'learning_rate/action_model': 4.540041695987917e-05, 'epoch': 1.13}) +[2026-06-16 05:05:30,802][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.014771097339689732, 'train/grad_norm_pre_clip': 4.0061421394348145, 'timing/data': 0.011727631092071533, 'timing/model': 1.7983464878052473, 'learning_rate/qwen_vl_interface': 4.537030629787182e-06, 'learning_rate/action_model': 4.5370306297871826e-05, 'epoch': 1.13}) +[2026-06-16 05:05:36,865][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.1529424786567688, 'train/grad_norm_pre_clip': 16.172382354736328, 'timing/data': 0.011098169256001711, 'timing/model': 1.5607441430911422, 'learning_rate/qwen_vl_interface': 4.534019856661226e-06, 'learning_rate/action_model': 4.534019856661226e-05, 'epoch': 1.13}) +[2026-06-16 05:05:43,156][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.41297370195388794, 'train/grad_norm_pre_clip': 26.779176712036133, 'timing/data': 0.011302540078759193, 'timing/model': 1.5597324608825147, 'learning_rate/qwen_vl_interface': 4.531009377847661e-06, 'learning_rate/action_model': 4.531009377847661e-05, 'epoch': 1.13}) +[2026-06-16 05:05:49,462][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.1852898746728897, 'train/grad_norm_pre_clip': 20.863187789916992, 'timing/data': 0.011062158737331629, 'timing/model': 1.561096184886992, 'learning_rate/qwen_vl_interface': 4.527999194583982e-06, 'learning_rate/action_model': 4.527999194583981e-05, 'epoch': 1.13}) +[2026-06-16 05:05:55,948][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.011124859564006329, 'train/grad_norm_pre_clip': 5.159463882446289, 'timing/data': 0.011348818894475698, 'timing/model': 1.5955834458582103, 'learning_rate/qwen_vl_interface': 4.52498930810756e-06, 'learning_rate/action_model': 4.52498930810756e-05, 'epoch': 1.13}) +[2026-06-16 05:06:02,265][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.02717645652592182, 'train/grad_norm_pre_clip': 11.3004150390625, 'timing/data': 0.0110610187985003, 'timing/model': 1.5580590567551553, 'learning_rate/qwen_vl_interface': 4.521979719655648e-06, 'learning_rate/action_model': 4.5219797196556484e-05, 'epoch': 1.13}) +[2026-06-16 05:06:08,696][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.0009340920951217413, 'train/grad_norm_pre_clip': 0.2284017652273178, 'timing/data': 0.012508523184806108, 'timing/model': 1.8231613948009908, 'learning_rate/qwen_vl_interface': 4.518970430465377e-06, 'learning_rate/action_model': 4.518970430465376e-05, 'epoch': 1.13}) +[2026-06-16 05:06:14,769][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.11994973570108414, 'train/grad_norm_pre_clip': 17.79692268371582, 'timing/data': 0.01220637233927846, 'timing/model': 1.5707509219646454, 'learning_rate/qwen_vl_interface': 4.515961441773745e-06, 'learning_rate/action_model': 4.515961441773745e-05, 'epoch': 1.13}) +[2026-06-16 05:06:21,071][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.013181236572563648, 'train/grad_norm_pre_clip': 3.9243388175964355, 'timing/data': 0.013121074996888638, 'timing/model': 1.5725270258262753, 'learning_rate/qwen_vl_interface': 4.5129527548176385e-06, 'learning_rate/action_model': 4.512952754817638e-05, 'epoch': 1.13}) +[2026-06-16 05:06:27,365][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.050351109355688095, 'train/grad_norm_pre_clip': 6.229503631591797, 'timing/data': 0.01130646001547575, 'timing/model': 1.566061019897461, 'learning_rate/qwen_vl_interface': 4.509944370833813e-06, 'learning_rate/action_model': 4.509944370833813e-05, 'epoch': 1.13}) +[2026-06-16 05:06:33,694][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.166080504655838, 'train/grad_norm_pre_clip': 14.374625205993652, 'timing/data': 0.010846239048987627, 'timing/model': 1.600741972681135, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 1.13}) +[2026-06-16 05:06:39,987][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.15558893978595734, 'train/grad_norm_pre_clip': 13.951494216918945, 'timing/data': 0.012296893168240786, 'timing/model': 1.5485458839684725, 'learning_rate/qwen_vl_interface': 4.503928516729416e-06, 'learning_rate/action_model': 4.503928516729416e-05, 'epoch': 1.13}) +[2026-06-16 05:06:46,431][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.01038090605288744, 'train/grad_norm_pre_clip': 1.79655921459198, 'timing/data': 0.012105772271752357, 'timing/model': 1.8247861010022461, 'learning_rate/qwen_vl_interface': 4.500921049081734e-06, 'learning_rate/action_model': 4.500921049081734e-05, 'epoch': 1.13}) +[2026-06-16 05:06:52,565][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.27590158581733704, 'train/grad_norm_pre_clip': 17.82119369506836, 'timing/data': 0.010920988861471415, 'timing/model': 1.5504134390503168, 'learning_rate/qwen_vl_interface': 4.497913889352113e-06, 'learning_rate/action_model': 4.4979138893521126e-05, 'epoch': 1.13}) +[2026-06-16 05:06:58,870][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.043440502136945724, 'train/grad_norm_pre_clip': 12.079538345336914, 'timing/data': 0.011147669050842524, 'timing/model': 1.5693186703138053, 'learning_rate/qwen_vl_interface': 4.494907038776684e-06, 'learning_rate/action_model': 4.4949070387766836e-05, 'epoch': 1.13}) +[2026-06-16 05:18:09,229][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.15979722142219543, 'train/grad_norm_pre_clip': 16.385391235351562, 'eval/loss': 0.0371687151491642, 'eval/flap_precision': 0.9546061415220294, 'eval/flap_recall': 0.975443383356071, 'eval/flap_f1': 0.9649122807017543, 'eval/flap_rate_model': 0.20805555555555555, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9649122807017544, 'timing/data': 0.01099536893889308, 'timing/model': 1.55355224898085, 'learning_rate/qwen_vl_interface': 4.49190049859145e-06, 'learning_rate/action_model': 4.49190049859145e-05, 'epoch': 1.13}) +[2026-06-16 05:18:09,230][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2800_state +[2026-06-16 05:18:09,237][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 05:18:17,739][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2800_state/model.safetensors +[2026-06-16 05:18:37,649][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2800_state/optimizer.bin +[2026-06-16 05:18:37,650][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2800_state/sampler.bin +[2026-06-16 05:18:37,651][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2800_state/sampler_1.bin +[2026-06-16 05:18:37,652][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2800_state/random_states_0.pkl +[2026-06-16 05:18:37,835][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2800 +[2026-06-16 05:18:37,836][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 05:18:37,851][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 05:18:44,028][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.07142800837755203, 'train/grad_norm_pre_clip': 16.283823013305664, 'timing/data': 0.010556367225944996, 'timing/model': 1.4723091120831668, 'learning_rate/qwen_vl_interface': 4.488894270032284e-06, 'learning_rate/action_model': 4.4888942700322834e-05, 'epoch': 1.13}) +[2026-06-16 05:18:50,207][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.052356939762830734, 'train/grad_norm_pre_clip': 5.737957954406738, 'timing/data': 0.010966073721647263, 'timing/model': 1.551488425116986, 'learning_rate/qwen_vl_interface': 4.485888354334933e-06, 'learning_rate/action_model': 4.4858883543349334e-05, 'epoch': 1.13}) +[2026-06-16 05:18:56,345][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.017678875476121902, 'train/grad_norm_pre_clip': 7.334151744842529, 'timing/data': 0.011080374009907246, 'timing/model': 1.5775141092017293, 'learning_rate/qwen_vl_interface': 4.482882752735017e-06, 'learning_rate/action_model': 4.482882752735017e-05, 'epoch': 1.13}) +[2026-06-16 05:19:02,544][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.13313665986061096, 'train/grad_norm_pre_clip': 17.047887802124023, 'timing/data': 0.010754796210676432, 'timing/model': 1.5156687581911683, 'learning_rate/qwen_vl_interface': 4.479877466468027e-06, 'learning_rate/action_model': 4.479877466468026e-05, 'epoch': 1.13}) +[2026-06-16 05:19:08,788][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.1309822052717209, 'train/grad_norm_pre_clip': 15.669516563415527, 'timing/data': 0.011216673068702221, 'timing/model': 1.812791544944048, 'learning_rate/qwen_vl_interface': 4.476872496769318e-06, 'learning_rate/action_model': 4.476872496769318e-05, 'epoch': 1.13}) +[2026-06-16 05:19:14,861][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.010690838098526001, 'train/grad_norm_pre_clip': 2.374574899673462, 'timing/data': 0.012593396939337254, 'timing/model': 1.5684906379319727, 'learning_rate/qwen_vl_interface': 4.473867844874122e-06, 'learning_rate/action_model': 4.473867844874122e-05, 'epoch': 1.13}) +[2026-06-16 05:19:21,139][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.13162845373153687, 'train/grad_norm_pre_clip': 12.240840911865234, 'timing/data': 0.011185290291905403, 'timing/model': 1.5580870187841356, 'learning_rate/qwen_vl_interface': 4.470863512017539e-06, 'learning_rate/action_model': 4.470863512017539e-05, 'epoch': 1.14}) +[2026-06-16 05:19:27,341][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.10157771408557892, 'train/grad_norm_pre_clip': 7.929023742675781, 'timing/data': 0.011362091172486544, 'timing/model': 1.5660227676853538, 'learning_rate/qwen_vl_interface': 4.4678594994345344e-06, 'learning_rate/action_model': 4.467859499434534e-05, 'epoch': 1.14}) +[2026-06-16 05:19:33,686][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.007533191703259945, 'train/grad_norm_pre_clip': 1.2135266065597534, 'timing/data': 0.010878949891775846, 'timing/model': 1.6166260880418122, 'learning_rate/qwen_vl_interface': 4.4648558083599475e-06, 'learning_rate/action_model': 4.464855808359947e-05, 'epoch': 1.14}) +[2026-06-16 05:19:39,952][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.03672359883785248, 'train/grad_norm_pre_clip': 5.23537015914917, 'timing/data': 0.012551939114928246, 'timing/model': 1.517509879078716, 'learning_rate/qwen_vl_interface': 4.461852440028479e-06, 'learning_rate/action_model': 4.461852440028479e-05, 'epoch': 1.14}) +[2026-06-16 05:19:46,278][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.1482965648174286, 'train/grad_norm_pre_clip': 15.072189331054688, 'timing/data': 0.011503797955811024, 'timing/model': 1.7807479412294924, 'learning_rate/qwen_vl_interface': 4.458849395674702e-06, 'learning_rate/action_model': 4.458849395674701e-05, 'epoch': 1.14}) +[2026-06-16 05:19:52,353][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.012748192995786667, 'train/grad_norm_pre_clip': 1.777921438217163, 'timing/data': 0.011775622144341469, 'timing/model': 1.5634790295735002, 'learning_rate/qwen_vl_interface': 4.455846676533052e-06, 'learning_rate/action_model': 4.4558466765330525e-05, 'epoch': 1.14}) +[2026-06-16 05:19:58,644][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.0628385990858078, 'train/grad_norm_pre_clip': 6.819873809814453, 'timing/data': 0.011202279012650251, 'timing/model': 1.5537913097068667, 'learning_rate/qwen_vl_interface': 4.452844283837838e-06, 'learning_rate/action_model': 4.452844283837838e-05, 'epoch': 1.14}) +[2026-06-16 05:20:04,841][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.04096318408846855, 'train/grad_norm_pre_clip': 9.725726127624512, 'timing/data': 0.01034247176721692, 'timing/model': 1.5583448838442564, 'learning_rate/qwen_vl_interface': 4.449842218823228e-06, 'learning_rate/action_model': 4.449842218823228e-05, 'epoch': 1.14}) +[2026-06-16 05:20:11,143][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.11193355172872543, 'train/grad_norm_pre_clip': 10.042166709899902, 'timing/data': 0.010931692086160183, 'timing/model': 1.5548943420872092, 'learning_rate/qwen_vl_interface': 4.446840482723256e-06, 'learning_rate/action_model': 4.446840482723255e-05, 'epoch': 1.14}) +[2026-06-16 05:20:17,448][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.008987830020487309, 'train/grad_norm_pre_clip': 4.165111064910889, 'timing/data': 0.010503362864255905, 'timing/model': 1.5124514219351113, 'learning_rate/qwen_vl_interface': 4.443839076771823e-06, 'learning_rate/action_model': 4.443839076771823e-05, 'epoch': 1.14}) +[2026-06-16 05:20:23,788][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.11410847306251526, 'train/grad_norm_pre_clip': 6.551294326782227, 'timing/data': 0.01074430625885725, 'timing/model': 1.792179292999208, 'learning_rate/qwen_vl_interface': 4.4408380022026945e-06, 'learning_rate/action_model': 4.440838002202694e-05, 'epoch': 1.14}) +[2026-06-16 05:20:30,024][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.0013019067700952291, 'train/grad_norm_pre_clip': 0.26536035537719727, 'timing/data': 0.011634308844804764, 'timing/model': 1.734003187622875, 'learning_rate/qwen_vl_interface': 4.437837260249499e-06, 'learning_rate/action_model': 4.437837260249499e-05, 'epoch': 1.14}) +[2026-06-16 05:20:36,054][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.033239059150218964, 'train/grad_norm_pre_clip': 10.610478401184082, 'timing/data': 0.010668751318007708, 'timing/model': 1.559719430282712, 'learning_rate/qwen_vl_interface': 4.434836852145728e-06, 'learning_rate/action_model': 4.434836852145728e-05, 'epoch': 1.14}) +[2026-06-16 05:20:42,365][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.02531212568283081, 'train/grad_norm_pre_clip': 8.490421295166016, 'timing/data': 0.013242674060165882, 'timing/model': 1.5742306825704873, 'learning_rate/qwen_vl_interface': 4.431836779124736e-06, 'learning_rate/action_model': 4.431836779124735e-05, 'epoch': 1.14}) +[2026-06-16 05:20:48,774][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.012744463048875332, 'train/grad_norm_pre_clip': 2.883793592453003, 'timing/data': 0.011617145035415888, 'timing/model': 1.5952107552438974, 'learning_rate/qwen_vl_interface': 4.428837042419738e-06, 'learning_rate/action_model': 4.428837042419738e-05, 'epoch': 1.14}) +[2026-06-16 05:20:55,254][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.0872383639216423, 'train/grad_norm_pre_clip': 12.778203964233398, 'timing/data': 0.011464291252195835, 'timing/model': 1.6055269022472203, 'learning_rate/qwen_vl_interface': 4.425837643263815e-06, 'learning_rate/action_model': 4.425837643263814e-05, 'epoch': 1.14}) +[2026-06-16 05:21:01,624][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.20224659144878387, 'train/grad_norm_pre_clip': 12.856252670288086, 'timing/data': 0.011632950976490974, 'timing/model': 1.8132317550480366, 'learning_rate/qwen_vl_interface': 4.4228385828899045e-06, 'learning_rate/action_model': 4.4228385828899046e-05, 'epoch': 1.14}) +[2026-06-16 05:21:07,901][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.011179668828845024, 'train/grad_norm_pre_clip': 4.888327121734619, 'timing/data': 0.01113335182890296, 'timing/model': 1.7900467780418694, 'learning_rate/qwen_vl_interface': 4.419839862530809e-06, 'learning_rate/action_model': 4.419839862530809e-05, 'epoch': 1.14}) +[2026-06-16 05:21:13,965][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.001780958380550146, 'train/grad_norm_pre_clip': 0.8698869347572327, 'timing/data': 0.011137741152197123, 'timing/model': 1.564602225087583, 'learning_rate/qwen_vl_interface': 4.416841483419188e-06, 'learning_rate/action_model': 4.416841483419188e-05, 'epoch': 1.14}) +[2026-06-16 05:21:20,264][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.04215116798877716, 'train/grad_norm_pre_clip': 7.2985005378723145, 'timing/data': 0.011714939028024673, 'timing/model': 1.5581504250876606, 'learning_rate/qwen_vl_interface': 4.413843446787561e-06, 'learning_rate/action_model': 4.413843446787561e-05, 'epoch': 1.14}) +[2026-06-16 05:21:26,672][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.03188137710094452, 'train/grad_norm_pre_clip': 9.314369201660156, 'timing/data': 0.010931635275483131, 'timing/model': 1.5631512203253806, 'learning_rate/qwen_vl_interface': 4.41084575386831e-06, 'learning_rate/action_model': 4.4108457538683093e-05, 'epoch': 1.14}) +[2026-06-16 05:21:33,059][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.06794431060552597, 'train/grad_norm_pre_clip': 14.754788398742676, 'timing/data': 0.01210907194763422, 'timing/model': 1.5748101458884776, 'learning_rate/qwen_vl_interface': 4.40784840589367e-06, 'learning_rate/action_model': 4.4078484058936706e-05, 'epoch': 1.14}) +[2026-06-16 05:21:39,410][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.12165626883506775, 'train/grad_norm_pre_clip': 21.86488151550293, 'timing/data': 0.010952861979603767, 'timing/model': 1.8105377648025751, 'learning_rate/qwen_vl_interface': 4.404851404095742e-06, 'learning_rate/action_model': 4.404851404095742e-05, 'epoch': 1.14}) +[2026-06-16 05:21:45,459][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.00217142840847373, 'train/grad_norm_pre_clip': 0.47059980034828186, 'timing/data': 0.011191744823008776, 'timing/model': 1.5461635179817677, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 1.14}) +[2026-06-16 05:21:51,761][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.06358161568641663, 'train/grad_norm_pre_clip': 22.2882080078125, 'timing/data': 0.011032310780137777, 'timing/model': 1.5636713900603354, 'learning_rate/qwen_vl_interface': 4.398858443957686e-06, 'learning_rate/action_model': 4.398858443957686e-05, 'epoch': 1.14}) +[2026-06-16 05:21:58,159][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.10459963232278824, 'train/grad_norm_pre_clip': 13.130929946899414, 'timing/data': 0.01144695607945323, 'timing/model': 1.5762506928294897, 'learning_rate/qwen_vl_interface': 4.395862488081038e-06, 'learning_rate/action_model': 4.395862488081038e-05, 'epoch': 1.15}) +[2026-06-16 05:22:04,546][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.008430292829871178, 'train/grad_norm_pre_clip': 1.6765087842941284, 'timing/data': 0.01187179097905755, 'timing/model': 1.5840804977342486, 'learning_rate/qwen_vl_interface': 4.392866883308059e-06, 'learning_rate/action_model': 4.392866883308059e-05, 'epoch': 1.15}) +[2026-06-16 05:22:10,859][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.12108367681503296, 'train/grad_norm_pre_clip': 5.870664119720459, 'timing/data': 0.011370041873306036, 'timing/model': 1.5620161080732942, 'learning_rate/qwen_vl_interface': 4.389871630870127e-06, 'learning_rate/action_model': 4.389871630870127e-05, 'epoch': 1.15}) +[2026-06-16 05:22:17,171][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.05543549358844757, 'train/grad_norm_pre_clip': 12.529511451721191, 'timing/data': 0.011391422245651484, 'timing/model': 1.771968204062432, 'learning_rate/qwen_vl_interface': 4.386876731998478e-06, 'learning_rate/action_model': 4.386876731998478e-05, 'epoch': 1.15}) +[2026-06-16 05:22:23,257][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.03299977630376816, 'train/grad_norm_pre_clip': 7.784771919250488, 'timing/data': 0.011605442967265844, 'timing/model': 1.560584066901356, 'learning_rate/qwen_vl_interface': 4.3838821879242006e-06, 'learning_rate/action_model': 4.3838821879242e-05, 'epoch': 1.15}) +[2026-06-16 05:22:29,562][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.005398296285420656, 'train/grad_norm_pre_clip': 1.2673438787460327, 'timing/data': 0.010632427874952555, 'timing/model': 1.5607024282217026, 'learning_rate/qwen_vl_interface': 4.38088799987824e-06, 'learning_rate/action_model': 4.38088799987824e-05, 'epoch': 1.15}) +[2026-06-16 05:22:35,997][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.09583991020917892, 'train/grad_norm_pre_clip': 5.372459888458252, 'timing/data': 0.011414438020437956, 'timing/model': 1.5978934690356255, 'learning_rate/qwen_vl_interface': 4.3778941690913935e-06, 'learning_rate/action_model': 4.3778941690913934e-05, 'epoch': 1.15}) +[2026-06-16 05:22:42,332][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.015747496858239174, 'train/grad_norm_pre_clip': 5.368834972381592, 'timing/data': 0.010909960139542818, 'timing/model': 1.5780663979239762, 'learning_rate/qwen_vl_interface': 4.374900696794311e-06, 'learning_rate/action_model': 4.374900696794311e-05, 'epoch': 1.15}) +[2026-06-16 05:22:48,660][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.005333280190825462, 'train/grad_norm_pre_clip': 1.4859529733657837, 'timing/data': 0.012556721922010183, 'timing/model': 1.5790898269042373, 'learning_rate/qwen_vl_interface': 4.371907584217499e-06, 'learning_rate/action_model': 4.371907584217498e-05, 'epoch': 1.15}) +[2026-06-16 05:22:55,012][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.044851306825876236, 'train/grad_norm_pre_clip': 17.893102645874023, 'timing/data': 0.009778930805623531, 'timing/model': 1.7966663311235607, 'learning_rate/qwen_vl_interface': 4.368914832591309e-06, 'learning_rate/action_model': 4.368914832591309e-05, 'epoch': 1.15}) +[2026-06-16 05:23:01,063][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.31641286611557007, 'train/grad_norm_pre_clip': 16.152488708496094, 'timing/data': 0.011520045343786478, 'timing/model': 1.5539557309821248, 'learning_rate/qwen_vl_interface': 4.365922443145949e-06, 'learning_rate/action_model': 4.365922443145949e-05, 'epoch': 1.15}) +[2026-06-16 05:23:07,361][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.14956669509410858, 'train/grad_norm_pre_clip': 16.811586380004883, 'timing/data': 0.011164199095219374, 'timing/model': 1.5552885718643665, 'learning_rate/qwen_vl_interface': 4.362930417111479e-06, 'learning_rate/action_model': 4.362930417111479e-05, 'epoch': 1.15}) +[2026-06-16 05:23:13,661][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.0018129405798390508, 'train/grad_norm_pre_clip': 0.37044230103492737, 'timing/data': 0.010603619739413261, 'timing/model': 1.567823754157871, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 1.15}) +[2026-06-16 05:23:19,964][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.01626884937286377, 'train/grad_norm_pre_clip': 3.334505796432495, 'timing/data': 0.011245197150856256, 'timing/model': 1.5309395561926067, 'learning_rate/qwen_vl_interface': 4.3569474601946934e-06, 'learning_rate/action_model': 4.356947460194694e-05, 'epoch': 1.15}) +[2026-06-16 05:23:26,299][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.01041360478848219, 'train/grad_norm_pre_clip': 4.332146167755127, 'timing/data': 0.011258216109126806, 'timing/model': 1.7919068080373108, 'learning_rate/qwen_vl_interface': 4.3539565317717446e-06, 'learning_rate/action_model': 4.353956531771745e-05, 'epoch': 1.15}) +[2026-06-16 05:23:32,463][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.024487797170877457, 'train/grad_norm_pre_clip': 8.477425575256348, 'timing/data': 0.010979911778122187, 'timing/model': 1.5916497698053718, 'learning_rate/qwen_vl_interface': 4.350965971678419e-06, 'learning_rate/action_model': 4.3509659716784185e-05, 'epoch': 1.15}) +[2026-06-16 05:23:38,779][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.07521413266658783, 'train/grad_norm_pre_clip': 7.015859603881836, 'timing/data': 0.010762637946754694, 'timing/model': 1.5833088960498571, 'learning_rate/qwen_vl_interface': 4.3479757811440225e-06, 'learning_rate/action_model': 4.347975781144022e-05, 'epoch': 1.15}) +[2026-06-16 05:23:45,126][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.00982726737856865, 'train/grad_norm_pre_clip': 3.1462512016296387, 'timing/data': 0.011068190913647413, 'timing/model': 1.6217349618673325, 'learning_rate/qwen_vl_interface': 4.34498596139771e-06, 'learning_rate/action_model': 4.34498596139771e-05, 'epoch': 1.15}) +[2026-06-16 05:23:51,366][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.11139699071645737, 'train/grad_norm_pre_clip': 16.94234848022461, 'timing/data': 0.010997398756444454, 'timing/model': 1.536490358877927, 'learning_rate/qwen_vl_interface': 4.341996513668483e-06, 'learning_rate/action_model': 4.3419965136684824e-05, 'epoch': 1.15}) +[2026-06-16 05:23:57,662][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.08155371993780136, 'train/grad_norm_pre_clip': 12.841134071350098, 'timing/data': 0.010728843975812197, 'timing/model': 1.5548342550173402, 'learning_rate/qwen_vl_interface': 4.33900743918519e-06, 'learning_rate/action_model': 4.33900743918519e-05, 'epoch': 1.15}) +[2026-06-16 05:24:04,203][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.020654788240790367, 'train/grad_norm_pre_clip': 6.829733371734619, 'timing/data': 0.03841738309711218, 'timing/model': 1.7858021571300924, 'learning_rate/qwen_vl_interface': 4.336018739176529e-06, 'learning_rate/action_model': 4.33601873917653e-05, 'epoch': 1.15}) +[2026-06-16 05:24:10,368][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.15237192809581757, 'train/grad_norm_pre_clip': 9.7234525680542, 'timing/data': 0.011048627085983753, 'timing/model': 1.572136223781854, 'learning_rate/qwen_vl_interface': 4.333030414871041e-06, 'learning_rate/action_model': 4.33303041487104e-05, 'epoch': 1.15}) +[2026-06-16 05:24:16,763][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.04565895348787308, 'train/grad_norm_pre_clip': 7.999603271484375, 'timing/data': 0.010765412356704473, 'timing/model': 1.554868496954441, 'learning_rate/qwen_vl_interface': 4.330042467497111e-06, 'learning_rate/action_model': 4.330042467497111e-05, 'epoch': 1.15}) +[2026-06-16 05:24:23,163][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.10495760291814804, 'train/grad_norm_pre_clip': 18.953548431396484, 'timing/data': 0.010851382743567228, 'timing/model': 1.5636147549375892, 'learning_rate/qwen_vl_interface': 4.327054898282973e-06, 'learning_rate/action_model': 4.3270548982829727e-05, 'epoch': 1.15}) +[2026-06-16 05:24:29,460][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.005419099237769842, 'train/grad_norm_pre_clip': 0.8744074106216431, 'timing/data': 0.010876671876758337, 'timing/model': 1.5644731810316443, 'learning_rate/qwen_vl_interface': 4.324067708456703e-06, 'learning_rate/action_model': 4.3240677084567036e-05, 'epoch': 1.15}) +[2026-06-16 05:24:35,769][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.10131622850894928, 'train/grad_norm_pre_clip': 9.937751770019531, 'timing/data': 0.010763138998299837, 'timing/model': 1.5433471072465181, 'learning_rate/qwen_vl_interface': 4.321080899246225e-06, 'learning_rate/action_model': 4.321080899246225e-05, 'epoch': 1.16}) +[2026-06-16 05:24:42,112][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.0610140897333622, 'train/grad_norm_pre_clip': 34.50979995727539, 'timing/data': 0.011004650965332985, 'timing/model': 1.8140086429193616, 'learning_rate/qwen_vl_interface': 4.318094471879299e-06, 'learning_rate/action_model': 4.3180944718792985e-05, 'epoch': 1.16}) +[2026-06-16 05:24:48,165][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.038792550563812256, 'train/grad_norm_pre_clip': 26.469079971313477, 'timing/data': 0.01325845904648304, 'timing/model': 1.5628105849027634, 'learning_rate/qwen_vl_interface': 4.315108427583535e-06, 'learning_rate/action_model': 4.315108427583535e-05, 'epoch': 1.16}) +[2026-06-16 05:24:54,465][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.11806230247020721, 'train/grad_norm_pre_clip': 25.030649185180664, 'timing/data': 0.011538946069777012, 'timing/model': 1.5653008571825922, 'learning_rate/qwen_vl_interface': 4.312122767586381e-06, 'learning_rate/action_model': 4.3121227675863815e-05, 'epoch': 1.16}) +[2026-06-16 05:25:00,862][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.24496497213840485, 'train/grad_norm_pre_clip': 21.014982223510742, 'timing/data': 0.011108629871159792, 'timing/model': 1.5654334919527173, 'learning_rate/qwen_vl_interface': 4.3091374931151324e-06, 'learning_rate/action_model': 4.309137493115132e-05, 'epoch': 1.16}) +[2026-06-16 05:25:07,263][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.0008080311818048358, 'train/grad_norm_pre_clip': 0.12065281718969345, 'timing/data': 0.011033957824110985, 'timing/model': 1.610165714751929, 'learning_rate/qwen_vl_interface': 4.306152605396921e-06, 'learning_rate/action_model': 4.306152605396921e-05, 'epoch': 1.16}) +[2026-06-16 05:25:13,660][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.2812122702598572, 'train/grad_norm_pre_clip': 21.433162689208984, 'timing/data': 0.011194078717380762, 'timing/model': 1.5769695630297065, 'learning_rate/qwen_vl_interface': 4.30316810565872e-06, 'learning_rate/action_model': 4.3031681056587206e-05, 'epoch': 1.16}) +[2026-06-16 05:25:19,995][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.0016082191141322255, 'train/grad_norm_pre_clip': 0.6096758246421814, 'timing/data': 0.010770533233880997, 'timing/model': 1.7959492476657033, 'learning_rate/qwen_vl_interface': 4.300183995127347e-06, 'learning_rate/action_model': 4.300183995127346e-05, 'epoch': 1.16}) +[2026-06-16 05:25:26,058][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.009820924140512943, 'train/grad_norm_pre_clip': 2.169229507446289, 'timing/data': 0.010771742090582848, 'timing/model': 1.5609821886755526, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 1.16}) +[2026-06-16 05:25:32,385][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.0350714810192585, 'train/grad_norm_pre_clip': 6.833247661590576, 'timing/data': 0.011107734870165586, 'timing/model': 1.5859254742972553, 'learning_rate/qwen_vl_interface': 4.29421694659154e-06, 'learning_rate/action_model': 4.294216946591539e-05, 'epoch': 1.16}) +[2026-06-16 05:25:38,689][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.33149412274360657, 'train/grad_norm_pre_clip': 11.486937522888184, 'timing/data': 0.01118771592155099, 'timing/model': 1.5951009560376406, 'learning_rate/qwen_vl_interface': 4.2912340110399345e-06, 'learning_rate/action_model': 4.291234011039934e-05, 'epoch': 1.16}) +[2026-06-16 05:25:44,977][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.06330133229494095, 'train/grad_norm_pre_clip': 13.816901206970215, 'timing/data': 0.010664698202162981, 'timing/model': 1.5760989720001817, 'learning_rate/qwen_vl_interface': 4.2882514696008125e-06, 'learning_rate/action_model': 4.2882514696008125e-05, 'epoch': 1.16}) +[2026-06-16 05:25:51,467][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.012723858468234539, 'train/grad_norm_pre_clip': 3.350560426712036, 'timing/data': 0.010686788242310286, 'timing/model': 1.5468777609057724, 'learning_rate/qwen_vl_interface': 4.2852693235001815e-06, 'learning_rate/action_model': 4.2852693235001815e-05, 'epoch': 1.16}) +[2026-06-16 05:25:57,911][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.012635323219001293, 'train/grad_norm_pre_clip': 8.799574851989746, 'timing/data': 0.010160090867429972, 'timing/model': 1.818512013182044, 'learning_rate/qwen_vl_interface': 4.2822875739638915e-06, 'learning_rate/action_model': 4.282287573963891e-05, 'epoch': 1.16}) +[2026-06-16 05:26:03,961][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.10246328264474869, 'train/grad_norm_pre_clip': 4.221970081329346, 'timing/data': 0.010932379867881536, 'timing/model': 1.5593537408858538, 'learning_rate/qwen_vl_interface': 4.279306222217625e-06, 'learning_rate/action_model': 4.279306222217626e-05, 'epoch': 1.16}) +[2026-06-16 05:26:10,261][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.057759612798690796, 'train/grad_norm_pre_clip': 14.885720252990723, 'timing/data': 0.01116273133084178, 'timing/model': 1.5496309460140765, 'learning_rate/qwen_vl_interface': 4.276325269486906e-06, 'learning_rate/action_model': 4.276325269486906e-05, 'epoch': 1.16}) +[2026-06-16 05:26:16,565][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.11141042411327362, 'train/grad_norm_pre_clip': 23.4172420501709, 'timing/data': 0.011451404076069593, 'timing/model': 1.5594750721938908, 'learning_rate/qwen_vl_interface': 4.273344716997089e-06, 'learning_rate/action_model': 4.273344716997089e-05, 'epoch': 1.16}) +[2026-06-16 05:26:22,967][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.05239322409033775, 'train/grad_norm_pre_clip': 9.509747505187988, 'timing/data': 0.015457000117748976, 'timing/model': 1.537414148915559, 'learning_rate/qwen_vl_interface': 4.270364565973368e-06, 'learning_rate/action_model': 4.270364565973367e-05, 'epoch': 1.16}) +[2026-06-16 05:26:29,359][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.0026193037629127502, 'train/grad_norm_pre_clip': 0.3640734851360321, 'timing/data': 0.011148808989673853, 'timing/model': 1.5918378420174122, 'learning_rate/qwen_vl_interface': 4.267384817640768e-06, 'learning_rate/action_model': 4.2673848176407685e-05, 'epoch': 1.16}) +[2026-06-16 05:26:35,701][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.04377385601401329, 'train/grad_norm_pre_clip': 18.87017822265625, 'timing/data': 0.011720804031938314, 'timing/model': 1.7996227932162583, 'learning_rate/qwen_vl_interface': 4.264405473224155e-06, 'learning_rate/action_model': 4.2644054732241546e-05, 'epoch': 1.16}) +[2026-06-16 05:26:41,769][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.009751711040735245, 'train/grad_norm_pre_clip': 1.519354224205017, 'timing/data': 0.010820033960044384, 'timing/model': 1.552274971269071, 'learning_rate/qwen_vl_interface': 4.261426533948221e-06, 'learning_rate/action_model': 4.2614265339482213e-05, 'epoch': 1.16}) +[2026-06-16 05:26:48,153][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.30863314867019653, 'train/grad_norm_pre_clip': 34.87670135498047, 'timing/data': 0.01069201109930873, 'timing/model': 1.5593647602945566, 'learning_rate/qwen_vl_interface': 4.258448001037499e-06, 'learning_rate/action_model': 4.2584480010374985e-05, 'epoch': 1.16}) +[2026-06-16 05:26:54,457][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.10980679094791412, 'train/grad_norm_pre_clip': 19.550067901611328, 'timing/data': 0.010584230069071054, 'timing/model': 1.5471704099327326, 'learning_rate/qwen_vl_interface': 4.255469875716348e-06, 'learning_rate/action_model': 4.255469875716348e-05, 'epoch': 1.16}) +[2026-06-16 05:27:00,778][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.016616983339190483, 'train/grad_norm_pre_clip': 7.723067283630371, 'timing/data': 0.012100055813789368, 'timing/model': 1.5710829705931246, 'learning_rate/qwen_vl_interface': 4.252492159208966e-06, 'learning_rate/action_model': 4.252492159208966e-05, 'epoch': 1.16}) +[2026-06-16 05:27:07,075][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.2991010546684265, 'train/grad_norm_pre_clip': 12.167569160461426, 'timing/data': 0.011473007965832949, 'timing/model': 1.5326425838284194, 'learning_rate/qwen_vl_interface': 4.2495148527393766e-06, 'learning_rate/action_model': 4.2495148527393766e-05, 'epoch': 1.17}) +[2026-06-16 05:27:13,413][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.0059398082084953785, 'train/grad_norm_pre_clip': 0.9629793167114258, 'timing/data': 0.011363557074218988, 'timing/model': 1.818626904860139, 'learning_rate/qwen_vl_interface': 4.246537957531441e-06, 'learning_rate/action_model': 4.2465379575314404e-05, 'epoch': 1.17}) +[2026-06-16 05:27:19,464][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.011184143833816051, 'train/grad_norm_pre_clip': 3.3570871353149414, 'timing/data': 0.01100787241011858, 'timing/model': 1.5656953710131347, 'learning_rate/qwen_vl_interface': 4.243561474808846e-06, 'learning_rate/action_model': 4.2435614748088456e-05, 'epoch': 1.17}) +[2026-06-16 05:27:25,759][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.007557831238955259, 'train/grad_norm_pre_clip': 1.9171236753463745, 'timing/data': 0.010347994044423103, 'timing/model': 1.5567661290988326, 'learning_rate/qwen_vl_interface': 4.240585405795112e-06, 'learning_rate/action_model': 4.240585405795112e-05, 'epoch': 1.17}) +[2026-06-16 05:27:32,068][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.006525989156216383, 'train/grad_norm_pre_clip': 1.9087822437286377, 'timing/data': 0.013406545855104923, 'timing/model': 1.5636580819264054, 'learning_rate/qwen_vl_interface': 4.237609751713589e-06, 'learning_rate/action_model': 4.2376097517135896e-05, 'epoch': 1.17}) +[2026-06-16 05:27:38,367][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.010341363027691841, 'train/grad_norm_pre_clip': 2.8794994354248047, 'timing/data': 0.011235011741518974, 'timing/model': 1.572261971887201, 'learning_rate/qwen_vl_interface': 4.234634513787455e-06, 'learning_rate/action_model': 4.2346345137874556e-05, 'epoch': 1.17}) +[2026-06-16 05:27:44,662][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.07794712483882904, 'train/grad_norm_pre_clip': 7.666618824005127, 'timing/data': 0.010426212102174759, 'timing/model': 1.5198579728603363, 'learning_rate/qwen_vl_interface': 4.2316596932397195e-06, 'learning_rate/action_model': 4.2316596932397195e-05, 'epoch': 1.17}) +[2026-06-16 05:27:51,143][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.004683353938162327, 'train/grad_norm_pre_clip': 2.139085531234741, 'timing/data': 0.011144539806991816, 'timing/model': 1.829275342170149, 'learning_rate/qwen_vl_interface': 4.22868529129322e-06, 'learning_rate/action_model': 4.22868529129322e-05, 'epoch': 1.17}) +[2026-06-16 05:27:57,487][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.01717926748096943, 'train/grad_norm_pre_clip': 4.549673557281494, 'timing/data': 0.011631384026259184, 'timing/model': 1.7637128452770412, 'learning_rate/qwen_vl_interface': 4.2257113091706165e-06, 'learning_rate/action_model': 4.2257113091706165e-05, 'epoch': 1.17}) +[2026-06-16 05:28:03,574][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.1372971534729004, 'train/grad_norm_pre_clip': 23.164474487304688, 'timing/data': 0.011249999050050974, 'timing/model': 1.5612847139127553, 'learning_rate/qwen_vl_interface': 4.222737748094404e-06, 'learning_rate/action_model': 4.222737748094404e-05, 'epoch': 1.17}) +[2026-06-16 05:28:09,975][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.20123371481895447, 'train/grad_norm_pre_clip': 20.127887725830078, 'timing/data': 0.011593122966587543, 'timing/model': 1.5577750690281391, 'learning_rate/qwen_vl_interface': 4.219764609286901e-06, 'learning_rate/action_model': 4.219764609286901e-05, 'epoch': 1.17}) +[2026-06-16 05:28:16,367][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.003483504755422473, 'train/grad_norm_pre_clip': 2.1345419883728027, 'timing/data': 0.011438049841672182, 'timing/model': 1.5637500849552453, 'learning_rate/qwen_vl_interface': 4.21679189397025e-06, 'learning_rate/action_model': 4.21679189397025e-05, 'epoch': 1.17}) +[2026-06-16 05:28:22,757][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.01892789453268051, 'train/grad_norm_pre_clip': 6.069695949554443, 'timing/data': 0.011390369851142168, 'timing/model': 1.5918385731056333, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 1.17}) +[2026-06-16 05:28:29,106][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.0006994461873546243, 'train/grad_norm_pre_clip': 0.0996786430478096, 'timing/data': 0.011478928849101067, 'timing/model': 1.7998213660903275, 'learning_rate/qwen_vl_interface': 4.210847738697219e-06, 'learning_rate/action_model': 4.210847738697218e-05, 'epoch': 1.17}) +[2026-06-16 05:28:35,422][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.002548410790041089, 'train/grad_norm_pre_clip': 0.6479791402816772, 'timing/data': 0.010990214068442583, 'timing/model': 1.8242296930402517, 'learning_rate/qwen_vl_interface': 4.207876301184256e-06, 'learning_rate/action_model': 4.207876301184256e-05, 'epoch': 1.17}) +[2026-06-16 05:28:41,461][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.06380511820316315, 'train/grad_norm_pre_clip': 14.187949180603027, 'timing/data': 0.011309226043522358, 'timing/model': 1.5499715809710324, 'learning_rate/qwen_vl_interface': 4.2049052920489825e-06, 'learning_rate/action_model': 4.2049052920489825e-05, 'epoch': 1.17}) +[2026-06-16 05:28:47,768][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.08972073346376419, 'train/grad_norm_pre_clip': 6.797647953033447, 'timing/data': 0.01212982414290309, 'timing/model': 1.5670739160850644, 'learning_rate/qwen_vl_interface': 4.201934712512666e-06, 'learning_rate/action_model': 4.2019347125126656e-05, 'epoch': 1.17}) +[2026-06-16 05:28:54,060][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.11840283125638962, 'train/grad_norm_pre_clip': 13.397163391113281, 'timing/data': 0.010773949790745974, 'timing/model': 1.5608971249312162, 'learning_rate/qwen_vl_interface': 4.1989645637964e-06, 'learning_rate/action_model': 4.1989645637964e-05, 'epoch': 1.17}) +[2026-06-16 05:29:00,464][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.06082181632518768, 'train/grad_norm_pre_clip': 17.965545654296875, 'timing/data': 0.011379336006939411, 'timing/model': 1.6073360149748623, 'learning_rate/qwen_vl_interface': 4.195994847121101e-06, 'learning_rate/action_model': 4.195994847121101e-05, 'epoch': 1.17}) +[2026-06-16 05:40:14,618][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.00798211432993412, 'train/grad_norm_pre_clip': 2.4418540000915527, 'eval/loss': 0.04294262453913689, 'eval/flap_precision': 0.9572763684913218, 'eval/flap_recall': 0.9781718963165075, 'eval/flap_f1': 0.9676113360323887, 'eval/flap_rate_model': 0.20805555555555555, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9676113360323888, 'timing/data': 0.011466616299003363, 'timing/model': 1.7802287437953055, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 1.17}) +[2026-06-16 05:40:14,619][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2900_state +[2026-06-16 05:40:14,627][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 05:40:23,803][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2900_state/model.safetensors +[2026-06-16 05:40:43,929][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2900_state/optimizer.bin +[2026-06-16 05:40:43,931][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2900_state/sampler.bin +[2026-06-16 05:40:43,931][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2900_state/sampler_1.bin +[2026-06-16 05:40:43,932][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_2900_state/random_states_0.pkl +[2026-06-16 05:40:44,117][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2900 +[2026-06-16 05:40:44,118][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 05:40:44,133][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 05:40:50,376][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.09923668950796127, 'train/grad_norm_pre_clip': 13.333022117614746, 'timing/data': 0.011618242599070072, 'timing/model': 1.8007780779153109, 'learning_rate/qwen_vl_interface': 4.19005671477618e-06, 'learning_rate/action_model': 4.1900567147761794e-05, 'epoch': 1.17}) +[2026-06-16 05:40:56,564][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.04734574258327484, 'train/grad_norm_pre_clip': 14.206936836242676, 'timing/data': 0.010863007046282291, 'timing/model': 1.793351562693715, 'learning_rate/qwen_vl_interface': 4.1870883015474975e-06, 'learning_rate/action_model': 4.187088301547497e-05, 'epoch': 1.17}) +[2026-06-16 05:41:02,632][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.015525905415415764, 'train/grad_norm_pre_clip': 8.704431533813477, 'timing/data': 0.01196338515728712, 'timing/model': 1.5545120700262487, 'learning_rate/qwen_vl_interface': 4.1841203252416655e-06, 'learning_rate/action_model': 4.1841203252416655e-05, 'epoch': 1.17}) +[2026-06-16 05:41:08,834][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.13642042875289917, 'train/grad_norm_pre_clip': 24.786550521850586, 'timing/data': 0.012259416282176971, 'timing/model': 1.5674182130023837, 'learning_rate/qwen_vl_interface': 4.181152787078705e-06, 'learning_rate/action_model': 4.1811527870787056e-05, 'epoch': 1.17}) +[2026-06-16 05:41:15,056][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.10557682067155838, 'train/grad_norm_pre_clip': 17.746700286865234, 'timing/data': 0.011098587885499, 'timing/model': 1.5819970779120922, 'learning_rate/qwen_vl_interface': 4.178185688278462e-06, 'learning_rate/action_model': 4.1781856882784616e-05, 'epoch': 1.17}) +[2026-06-16 05:41:21,335][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.009017262607812881, 'train/grad_norm_pre_clip': 1.9597378969192505, 'timing/data': 0.011223628651350737, 'timing/model': 1.5175449941307306, 'learning_rate/qwen_vl_interface': 4.175219030060595e-06, 'learning_rate/action_model': 4.175219030060595e-05, 'epoch': 1.18}) +[2026-06-16 05:41:27,596][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.0013591699535027146, 'train/grad_norm_pre_clip': 0.23551508784294128, 'timing/data': 0.010786646045744419, 'timing/model': 1.815028970129788, 'learning_rate/qwen_vl_interface': 4.1722528136445875e-06, 'learning_rate/action_model': 4.172252813644588e-05, 'epoch': 1.18}) +[2026-06-16 05:41:33,858][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.08264729380607605, 'train/grad_norm_pre_clip': 6.325863838195801, 'timing/data': 0.011511111166328192, 'timing/model': 1.7590944259427488, 'learning_rate/qwen_vl_interface': 4.169287040249738e-06, 'learning_rate/action_model': 4.169287040249738e-05, 'epoch': 1.18}) +[2026-06-16 05:41:39,954][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.0013962846715003252, 'train/grad_norm_pre_clip': 0.2479342669248581, 'timing/data': 0.011413310188800097, 'timing/model': 1.5689705940894783, 'learning_rate/qwen_vl_interface': 4.166321711095164e-06, 'learning_rate/action_model': 4.166321711095164e-05, 'epoch': 1.18}) +[2026-06-16 05:41:46,247][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.0004554250626824796, 'train/grad_norm_pre_clip': 0.09889625757932663, 'timing/data': 0.010690675117075443, 'timing/model': 1.5707705179229379, 'learning_rate/qwen_vl_interface': 4.1633568273998e-06, 'learning_rate/action_model': 4.1633568273998e-05, 'epoch': 1.18}) +[2026-06-16 05:41:52,546][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.21748746931552887, 'train/grad_norm_pre_clip': 10.247908592224121, 'timing/data': 0.011138528119772673, 'timing/model': 1.560218766797334, 'learning_rate/qwen_vl_interface': 4.1603923903823995e-06, 'learning_rate/action_model': 4.160392390382399e-05, 'epoch': 1.18}) +[2026-06-16 05:41:58,844][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.4024352729320526, 'train/grad_norm_pre_clip': 27.16038703918457, 'timing/data': 0.011141006834805012, 'timing/model': 1.519507666118443, 'learning_rate/qwen_vl_interface': 4.1574284012615276e-06, 'learning_rate/action_model': 4.157428401261528e-05, 'epoch': 1.18}) +[2026-06-16 05:42:05,147][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.017581695690751076, 'train/grad_norm_pre_clip': 16.015825271606445, 'timing/data': 0.011311498004943132, 'timing/model': 1.517138612922281, 'learning_rate/qwen_vl_interface': 4.154464861255571e-06, 'learning_rate/action_model': 4.154464861255571e-05, 'epoch': 1.18}) +[2026-06-16 05:42:11,525][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.21585707366466522, 'train/grad_norm_pre_clip': 25.90420913696289, 'timing/data': 0.011137626599520445, 'timing/model': 1.7975401473231614, 'learning_rate/qwen_vl_interface': 4.151501771582728e-06, 'learning_rate/action_model': 4.151501771582728e-05, 'epoch': 1.18}) +[2026-06-16 05:42:17,698][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.0001292569504585117, 'train/grad_norm_pre_clip': 0.026728712022304535, 'timing/data': 0.011787142138928175, 'timing/model': 1.6302965772338212, 'learning_rate/qwen_vl_interface': 4.148539133461012e-06, 'learning_rate/action_model': 4.1485391334610116e-05, 'epoch': 1.18}) +[2026-06-16 05:42:24,063][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.015075786039233208, 'train/grad_norm_pre_clip': 10.504778861999512, 'timing/data': 0.01220813486725092, 'timing/model': 1.5505313659086823, 'learning_rate/qwen_vl_interface': 4.145576948108254e-06, 'learning_rate/action_model': 4.145576948108253e-05, 'epoch': 1.18}) +[2026-06-16 05:42:30,659][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.11133619397878647, 'train/grad_norm_pre_clip': 6.051074504852295, 'timing/data': 0.011443659663200378, 'timing/model': 1.6502604698762298, 'learning_rate/qwen_vl_interface': 4.142615216742093e-06, 'learning_rate/action_model': 4.142615216742093e-05, 'epoch': 1.18}) +[2026-06-16 05:42:37,171][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.0024632620625197887, 'train/grad_norm_pre_clip': 0.9741566777229309, 'timing/data': 0.01258683716878295, 'timing/model': 1.5948761799372733, 'learning_rate/qwen_vl_interface': 4.139653940579988e-06, 'learning_rate/action_model': 4.1396539405799874e-05, 'epoch': 1.18}) +[2026-06-16 05:42:43,485][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.005958084017038345, 'train/grad_norm_pre_clip': 1.5835431814193726, 'timing/data': 0.011223537847399712, 'timing/model': 1.785283898934722, 'learning_rate/qwen_vl_interface': 4.136693120839206e-06, 'learning_rate/action_model': 4.136693120839205e-05, 'epoch': 1.18}) +[2026-06-16 05:42:49,556][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.017636490985751152, 'train/grad_norm_pre_clip': 10.439703941345215, 'timing/data': 0.011050416622310877, 'timing/model': 1.5563549720682204, 'learning_rate/qwen_vl_interface': 4.133732758736828e-06, 'learning_rate/action_model': 4.133732758736828e-05, 'epoch': 1.18}) +[2026-06-16 05:42:55,864][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.0310457032173872, 'train/grad_norm_pre_clip': 15.156285285949707, 'timing/data': 0.010980014689266682, 'timing/model': 1.5655791708268225, 'learning_rate/qwen_vl_interface': 4.13077285548975e-06, 'learning_rate/action_model': 4.13077285548975e-05, 'epoch': 1.18}) +[2026-06-16 05:43:02,055][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.006011639256030321, 'train/grad_norm_pre_clip': 1.8697166442871094, 'timing/data': 0.010870815254747868, 'timing/model': 1.5520765231922269, 'learning_rate/qwen_vl_interface': 4.127813412314672e-06, 'learning_rate/action_model': 4.1278134123146725e-05, 'epoch': 1.18}) +[2026-06-16 05:43:08,359][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.01941772922873497, 'train/grad_norm_pre_clip': 15.403545379638672, 'timing/data': 0.011728739831596613, 'timing/model': 1.5774370119906962, 'learning_rate/qwen_vl_interface': 4.124854430428113e-06, 'learning_rate/action_model': 4.1248544304281126e-05, 'epoch': 1.18}) +[2026-06-16 05:43:14,662][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.010362697765231133, 'train/grad_norm_pre_clip': 4.265336990356445, 'timing/data': 0.013996736146509647, 'timing/model': 1.5513131180778146, 'learning_rate/qwen_vl_interface': 4.121895911046397e-06, 'learning_rate/action_model': 4.121895911046396e-05, 'epoch': 1.18}) +[2026-06-16 05:43:20,997][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.02029498666524887, 'train/grad_norm_pre_clip': 12.938993453979492, 'timing/data': 0.010967404581606388, 'timing/model': 1.797434588894248, 'learning_rate/qwen_vl_interface': 4.118937855385659e-06, 'learning_rate/action_model': 4.1189378553856584e-05, 'epoch': 1.18}) +[2026-06-16 05:43:27,053][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.19625931978225708, 'train/grad_norm_pre_clip': 18.32495880126953, 'timing/data': 0.01147740799933672, 'timing/model': 1.5569893489591777, 'learning_rate/qwen_vl_interface': 4.115980264661844e-06, 'learning_rate/action_model': 4.115980264661844e-05, 'epoch': 1.18}) +[2026-06-16 05:43:33,361][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.0008793957531452179, 'train/grad_norm_pre_clip': 0.3546255826950073, 'timing/data': 0.012224233709275723, 'timing/model': 1.5625216620974243, 'learning_rate/qwen_vl_interface': 4.113023140090709e-06, 'learning_rate/action_model': 4.113023140090709e-05, 'epoch': 1.18}) +[2026-06-16 05:43:39,668][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.013835754245519638, 'train/grad_norm_pre_clip': 4.783517837524414, 'timing/data': 0.012378893792629242, 'timing/model': 1.5666216360405087, 'learning_rate/qwen_vl_interface': 4.110066482887812e-06, 'learning_rate/action_model': 4.110066482887812e-05, 'epoch': 1.18}) +[2026-06-16 05:43:46,057][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.09940283000469208, 'train/grad_norm_pre_clip': 9.785914421081543, 'timing/data': 0.01084741298109293, 'timing/model': 1.570721719879657, 'learning_rate/qwen_vl_interface': 4.107110294268525e-06, 'learning_rate/action_model': 4.107110294268525e-05, 'epoch': 1.18}) +[2026-06-16 05:43:52,362][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.09208422154188156, 'train/grad_norm_pre_clip': 14.880642890930176, 'timing/data': 0.0116714290343225, 'timing/model': 1.5646094242110848, 'learning_rate/qwen_vl_interface': 4.104154575448026e-06, 'learning_rate/action_model': 4.104154575448025e-05, 'epoch': 1.18}) +[2026-06-16 05:43:58,691][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.0044558062218129635, 'train/grad_norm_pre_clip': 1.1767189502716064, 'timing/data': 0.0106468521989882, 'timing/model': 1.7914793831296265, 'learning_rate/qwen_vl_interface': 4.101199327641297e-06, 'learning_rate/action_model': 4.101199327641297e-05, 'epoch': 1.19}) +[2026-06-16 05:44:04,792][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.26068010926246643, 'train/grad_norm_pre_clip': 16.985027313232422, 'timing/data': 0.010904114227741957, 'timing/model': 1.5831065829843283, 'learning_rate/qwen_vl_interface': 4.098244552063133e-06, 'learning_rate/action_model': 4.098244552063132e-05, 'epoch': 1.19}) +[2026-06-16 05:44:11,065][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.08607612550258636, 'train/grad_norm_pre_clip': 9.573665618896484, 'timing/data': 0.011677159927785397, 'timing/model': 1.5555703230202198, 'learning_rate/qwen_vl_interface': 4.0952902499281254e-06, 'learning_rate/action_model': 4.095290249928125e-05, 'epoch': 1.19}) +[2026-06-16 05:44:17,364][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.005646729841828346, 'train/grad_norm_pre_clip': 1.672052025794983, 'timing/data': 0.010474060662090778, 'timing/model': 1.5625836965627968, 'learning_rate/qwen_vl_interface': 4.092336422450681e-06, 'learning_rate/action_model': 4.0923364224506804e-05, 'epoch': 1.19}) +[2026-06-16 05:44:23,656][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.002122994512319565, 'train/grad_norm_pre_clip': 0.2766702473163605, 'timing/data': 0.011505827773362398, 'timing/model': 1.528555313590914, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 1.19}) +[2026-06-16 05:44:30,068][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.25316375494003296, 'train/grad_norm_pre_clip': 23.117755889892578, 'timing/data': 0.010772302746772766, 'timing/model': 1.5606645187363029, 'learning_rate/qwen_vl_interface': 4.086430196325106e-06, 'learning_rate/action_model': 4.086430196325106e-05, 'epoch': 1.19}) +[2026-06-16 05:44:36,393][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.017810510471463203, 'train/grad_norm_pre_clip': 8.256402969360352, 'timing/data': 0.011512057855725288, 'timing/model': 1.7961913142353296, 'learning_rate/qwen_vl_interface': 4.083477800104802e-06, 'learning_rate/action_model': 4.083477800104802e-05, 'epoch': 1.19}) +[2026-06-16 05:44:42,548][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.08052735775709152, 'train/grad_norm_pre_clip': 6.5141921043396, 'timing/data': 0.012058300897479057, 'timing/model': 1.6483027641661465, 'learning_rate/qwen_vl_interface': 4.080525883397711e-06, 'learning_rate/action_model': 4.080525883397711e-05, 'epoch': 1.19}) +[2026-06-16 05:44:48,874][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.00354492524638772, 'train/grad_norm_pre_clip': 0.6013212203979492, 'timing/data': 0.010851733386516571, 'timing/model': 1.5669882437214255, 'learning_rate/qwen_vl_interface': 4.077574447417256e-06, 'learning_rate/action_model': 4.077574447417256e-05, 'epoch': 1.19}) +[2026-06-16 05:44:55,164][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.22347024083137512, 'train/grad_norm_pre_clip': 26.009464263916016, 'timing/data': 0.010846322868019342, 'timing/model': 1.5620285281911492, 'learning_rate/qwen_vl_interface': 4.074623493376657e-06, 'learning_rate/action_model': 4.074623493376657e-05, 'epoch': 1.19}) +[2026-06-16 05:45:01,467][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.0006646799156442285, 'train/grad_norm_pre_clip': 0.18756334483623505, 'timing/data': 0.010222998913377523, 'timing/model': 1.5703873629681766, 'learning_rate/qwen_vl_interface': 4.071673022488944e-06, 'learning_rate/action_model': 4.071673022488944e-05, 'epoch': 1.19}) +[2026-06-16 05:45:07,817][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.015095294453203678, 'train/grad_norm_pre_clip': 11.721284866333008, 'timing/data': 0.010602731257677078, 'timing/model': 1.5668401871807873, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 1.19}) +[2026-06-16 05:45:14,097][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.0005254168063402176, 'train/grad_norm_pre_clip': 0.07402884215116501, 'timing/data': 0.011258235201239586, 'timing/model': 1.7917122719809413, 'learning_rate/qwen_vl_interface': 4.065773535023277e-06, 'learning_rate/action_model': 4.0657735350232766e-05, 'epoch': 1.19}) +[2026-06-16 05:45:20,173][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.10869349539279938, 'train/grad_norm_pre_clip': 16.430465698242188, 'timing/data': 0.012537662871181965, 'timing/model': 1.5657759169116616, 'learning_rate/qwen_vl_interface': 4.062824520870382e-06, 'learning_rate/action_model': 4.062824520870381e-05, 'epoch': 1.19}) +[2026-06-16 05:45:26,462][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.008767419494688511, 'train/grad_norm_pre_clip': 1.6299192905426025, 'timing/data': 0.011829738970845938, 'timing/model': 1.5597869451157749, 'learning_rate/qwen_vl_interface': 4.059875994720481e-06, 'learning_rate/action_model': 4.059875994720481e-05, 'epoch': 1.19}) +[2026-06-16 05:45:32,765][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.0845189318060875, 'train/grad_norm_pre_clip': 22.79084014892578, 'timing/data': 0.011291536036878824, 'timing/model': 1.5577874099835753, 'learning_rate/qwen_vl_interface': 4.056927957785605e-06, 'learning_rate/action_model': 4.056927957785605e-05, 'epoch': 1.19}) +[2026-06-16 05:45:39,190][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.001737611135467887, 'train/grad_norm_pre_clip': 0.4573441743850708, 'timing/data': 0.011737148277461529, 'timing/model': 1.605384967289865, 'learning_rate/qwen_vl_interface': 4.0539804112775785e-06, 'learning_rate/action_model': 4.053980411277578e-05, 'epoch': 1.19}) +[2026-06-16 05:45:45,557][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.13141784071922302, 'train/grad_norm_pre_clip': 23.60169219970703, 'timing/data': 0.010392539668828249, 'timing/model': 1.5384314288385212, 'learning_rate/qwen_vl_interface': 4.051033356408027e-06, 'learning_rate/action_model': 4.0510333564080264e-05, 'epoch': 1.19}) +[2026-06-16 05:45:51,881][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.2963399589061737, 'train/grad_norm_pre_clip': 16.713951110839844, 'timing/data': 0.012856364715844393, 'timing/model': 1.7492406922392547, 'learning_rate/qwen_vl_interface': 4.048086794388375e-06, 'learning_rate/action_model': 4.048086794388375e-05, 'epoch': 1.19}) +[2026-06-16 05:45:57,978][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.13788332045078278, 'train/grad_norm_pre_clip': 20.40321922302246, 'timing/data': 0.011110254097729921, 'timing/model': 1.5777365788817406, 'learning_rate/qwen_vl_interface': 4.045140726429841e-06, 'learning_rate/action_model': 4.0451407264298406e-05, 'epoch': 1.19}) +[2026-06-16 05:46:04,254][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.00407022051513195, 'train/grad_norm_pre_clip': 0.9931496381759644, 'timing/data': 0.01087426207959652, 'timing/model': 1.5560984648764133, 'learning_rate/qwen_vl_interface': 4.042195153743443e-06, 'learning_rate/action_model': 4.042195153743443e-05, 'epoch': 1.19}) +[2026-06-16 05:46:10,662][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.06779583543539047, 'train/grad_norm_pre_clip': 20.550886154174805, 'timing/data': 0.010429208632558584, 'timing/model': 1.5686751082539558, 'learning_rate/qwen_vl_interface': 4.039250077539993e-06, 'learning_rate/action_model': 4.039250077539993e-05, 'epoch': 1.19}) +[2026-06-16 05:46:16,968][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.015207527205348015, 'train/grad_norm_pre_clip': 7.836061000823975, 'timing/data': 0.011029613204300404, 'timing/model': 1.5697097647935152, 'learning_rate/qwen_vl_interface': 4.036305499030104e-06, 'learning_rate/action_model': 4.0363054990301036e-05, 'epoch': 1.19}) +[2026-06-16 05:46:23,451][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.12818068265914917, 'train/grad_norm_pre_clip': 11.812564849853516, 'timing/data': 0.011090153828263283, 'timing/model': 1.622972205746919, 'learning_rate/qwen_vl_interface': 4.033361419424178e-06, 'learning_rate/action_model': 4.033361419424178e-05, 'epoch': 1.19}) +[2026-06-16 05:46:29,803][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.010308708995580673, 'train/grad_norm_pre_clip': 4.470751762390137, 'timing/data': 0.010823132004588842, 'timing/model': 1.8097294019535184, 'learning_rate/qwen_vl_interface': 4.030417839932418e-06, 'learning_rate/action_model': 4.030417839932417e-05, 'epoch': 1.19}) +[2026-06-16 05:46:35,957][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.23549805581569672, 'train/grad_norm_pre_clip': 18.636825561523438, 'timing/data': 0.011126082856208086, 'timing/model': 1.5605947938747704, 'learning_rate/qwen_vl_interface': 4.027474761764815e-06, 'learning_rate/action_model': 4.027474761764814e-05, 'epoch': 1.2}) +[2026-06-16 05:46:42,262][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.22635799646377563, 'train/grad_norm_pre_clip': 10.041579246520996, 'timing/data': 0.011309004854410887, 'timing/model': 1.5602236008271575, 'learning_rate/qwen_vl_interface': 4.024532186131158e-06, 'learning_rate/action_model': 4.0245321861311575e-05, 'epoch': 1.2}) +[2026-06-16 05:46:48,554][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.012141027487814426, 'train/grad_norm_pre_clip': 4.613246440887451, 'timing/data': 0.011330123990774155, 'timing/model': 1.558776631951332, 'learning_rate/qwen_vl_interface': 4.021590114241029e-06, 'learning_rate/action_model': 4.021590114241029e-05, 'epoch': 1.2}) +[2026-06-16 05:46:54,859][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.0002821152447722852, 'train/grad_norm_pre_clip': 0.06281349062919617, 'timing/data': 0.010716480668634176, 'timing/model': 1.5604365328326821, 'learning_rate/qwen_vl_interface': 4.018648547303804e-06, 'learning_rate/action_model': 4.0186485473038036e-05, 'epoch': 1.2}) +[2026-06-16 05:47:01,175][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.0012914687395095825, 'train/grad_norm_pre_clip': 0.3046013414859772, 'timing/data': 0.011016972828656435, 'timing/model': 1.5308114779181778, 'learning_rate/qwen_vl_interface': 4.015707486528649e-06, 'learning_rate/action_model': 4.015707486528649e-05, 'epoch': 1.2}) +[2026-06-16 05:47:07,500][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.1140735000371933, 'train/grad_norm_pre_clip': 18.522260665893555, 'timing/data': 0.012474091723561287, 'timing/model': 1.7969463700428605, 'learning_rate/qwen_vl_interface': 4.012766933124523e-06, 'learning_rate/action_model': 4.012766933124523e-05, 'epoch': 1.2}) +[2026-06-16 05:47:13,721][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.32030627131462097, 'train/grad_norm_pre_clip': 17.411449432373047, 'timing/data': 0.011265323031693697, 'timing/model': 1.723249426111579, 'learning_rate/qwen_vl_interface': 4.009826888300176e-06, 'learning_rate/action_model': 4.0098268883001755e-05, 'epoch': 1.2}) +[2026-06-16 05:47:19,763][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.003327759448438883, 'train/grad_norm_pre_clip': 0.4354560971260071, 'timing/data': 0.011886138003319502, 'timing/model': 1.5515352780930698, 'learning_rate/qwen_vl_interface': 4.00688735326415e-06, 'learning_rate/action_model': 4.0068873532641495e-05, 'epoch': 1.2}) +[2026-06-16 05:47:26,066][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.1525086760520935, 'train/grad_norm_pre_clip': 24.979591369628906, 'timing/data': 0.011324003804475069, 'timing/model': 1.5668831099756062, 'learning_rate/qwen_vl_interface': 4.0039483292247765e-06, 'learning_rate/action_model': 4.003948329224776e-05, 'epoch': 1.2}) +[2026-06-16 05:47:32,360][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.001357266679406166, 'train/grad_norm_pre_clip': 0.3295662999153137, 'timing/data': 0.011033462826162577, 'timing/model': 1.559421090874821, 'learning_rate/qwen_vl_interface': 4.001009817390178e-06, 'learning_rate/action_model': 4.001009817390178e-05, 'epoch': 1.2}) +[2026-06-16 05:47:38,663][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.003226066008210182, 'train/grad_norm_pre_clip': 0.8769345879554749, 'timing/data': 0.010511759202927351, 'timing/model': 1.5349444630555809, 'learning_rate/qwen_vl_interface': 3.998071818968264e-06, 'learning_rate/action_model': 3.998071818968264e-05, 'epoch': 1.2}) +[2026-06-16 05:47:44,996][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.011089797131717205, 'train/grad_norm_pre_clip': 3.8539042472839355, 'timing/data': 0.01094310125336051, 'timing/model': 1.7949980460107327, 'learning_rate/qwen_vl_interface': 3.995134335166736e-06, 'learning_rate/action_model': 3.995134335166736e-05, 'epoch': 1.2}) +[2026-06-16 05:47:51,296][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.07443904876708984, 'train/grad_norm_pre_clip': 13.293416976928711, 'timing/data': 0.010548439342528582, 'timing/model': 1.7916679540649056, 'learning_rate/qwen_vl_interface': 3.992197367193083e-06, 'learning_rate/action_model': 3.992197367193082e-05, 'epoch': 1.2}) +[2026-06-16 05:47:57,369][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.00670161098241806, 'train/grad_norm_pre_clip': 2.0727882385253906, 'timing/data': 0.010945871006697416, 'timing/model': 1.5678094220347703, 'learning_rate/qwen_vl_interface': 3.98926091625458e-06, 'learning_rate/action_model': 3.9892609162545795e-05, 'epoch': 1.2}) +[2026-06-16 05:48:03,867][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.0032088318839669228, 'train/grad_norm_pre_clip': 0.720507025718689, 'timing/data': 0.011244102846831083, 'timing/model': 1.6665815780870616, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 1.2}) +[2026-06-16 05:48:10,193][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.0021081070881336927, 'train/grad_norm_pre_clip': 0.8509340882301331, 'timing/data': 0.010166375897824764, 'timing/model': 1.5766754327341914, 'learning_rate/qwen_vl_interface': 3.983389570311073e-06, 'learning_rate/action_model': 3.9833895703110724e-05, 'epoch': 1.2}) +[2026-06-16 05:48:16,466][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.0013087522238492966, 'train/grad_norm_pre_clip': 0.7822842001914978, 'timing/data': 0.010454067960381508, 'timing/model': 1.515694806817919, 'learning_rate/qwen_vl_interface': 3.9804546777195546e-06, 'learning_rate/action_model': 3.980454677719554e-05, 'epoch': 1.2}) +[2026-06-16 05:48:22,793][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.20017819106578827, 'train/grad_norm_pre_clip': 14.043612480163574, 'timing/data': 0.011352873872965574, 'timing/model': 1.7984495600685477, 'learning_rate/qwen_vl_interface': 3.977520306990164e-06, 'learning_rate/action_model': 3.977520306990163e-05, 'epoch': 1.2}) +[2026-06-16 05:48:28,859][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.09619521349668503, 'train/grad_norm_pre_clip': 11.416250228881836, 'timing/data': 0.011402974370867014, 'timing/model': 1.5648346906527877, 'learning_rate/qwen_vl_interface': 3.974586459329109e-06, 'learning_rate/action_model': 3.9745864593291085e-05, 'epoch': 1.2}) +[2026-06-16 05:48:35,169][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.19966591894626617, 'train/grad_norm_pre_clip': 24.4425106048584, 'timing/data': 0.011008801870048046, 'timing/model': 1.571154914330691, 'learning_rate/qwen_vl_interface': 3.9716531359423855e-06, 'learning_rate/action_model': 3.971653135942385e-05, 'epoch': 1.2}) +[2026-06-16 05:48:41,465][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.0656193196773529, 'train/grad_norm_pre_clip': 9.751228332519531, 'timing/data': 0.010870940051972866, 'timing/model': 1.5658526429906487, 'learning_rate/qwen_vl_interface': 3.968720338035769e-06, 'learning_rate/action_model': 3.968720338035769e-05, 'epoch': 1.2}) +[2026-06-16 05:48:47,767][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.023478206247091293, 'train/grad_norm_pre_clip': 10.823195457458496, 'timing/data': 0.010954088997095823, 'timing/model': 1.521326644346118, 'learning_rate/qwen_vl_interface': 3.965788066814824e-06, 'learning_rate/action_model': 3.965788066814823e-05, 'epoch': 1.2}) +[2026-06-16 05:48:54,272][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.11637911945581436, 'train/grad_norm_pre_clip': 15.061600685119629, 'timing/data': 0.01234262716025114, 'timing/model': 1.5880241990089417, 'learning_rate/qwen_vl_interface': 3.962856323484895e-06, 'learning_rate/action_model': 3.9628563234848945e-05, 'epoch': 1.2}) +[2026-06-16 05:49:00,558][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.0053838808089494705, 'train/grad_norm_pre_clip': 2.7445881366729736, 'timing/data': 0.011426462326198816, 'timing/model': 1.7591196270659566, 'learning_rate/qwen_vl_interface': 3.959925109251112e-06, 'learning_rate/action_model': 3.959925109251112e-05, 'epoch': 1.2}) +[2026-06-16 05:49:06,668][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.0011724599171429873, 'train/grad_norm_pre_clip': 0.2848178446292877, 'timing/data': 0.012425288092345, 'timing/model': 1.564700578339398, 'learning_rate/qwen_vl_interface': 3.956994425318387e-06, 'learning_rate/action_model': 3.956994425318387e-05, 'epoch': 1.21}) +[2026-06-16 05:49:12,964][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.13247381150722504, 'train/grad_norm_pre_clip': 12.072566032409668, 'timing/data': 0.011153158731758595, 'timing/model': 1.5663895858451724, 'learning_rate/qwen_vl_interface': 3.9540642728914115e-06, 'learning_rate/action_model': 3.9540642728914115e-05, 'epoch': 1.21}) +[2026-06-16 05:49:19,267][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.2788565456867218, 'train/grad_norm_pre_clip': 28.284303665161133, 'timing/data': 0.012091316282749176, 'timing/model': 1.5678951730951667, 'learning_rate/qwen_vl_interface': 3.951134653174664e-06, 'learning_rate/action_model': 3.951134653174663e-05, 'epoch': 1.21}) +[2026-06-16 05:49:25,670][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.00045996211702004075, 'train/grad_norm_pre_clip': 0.05015667900443077, 'timing/data': 0.012086835689842701, 'timing/model': 1.6086592352949083, 'learning_rate/qwen_vl_interface': 3.948205567372397e-06, 'learning_rate/action_model': 3.948205567372397e-05, 'epoch': 1.21}) +[2026-06-16 05:49:31,969][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.32876577973365784, 'train/grad_norm_pre_clip': 20.093299865722656, 'timing/data': 0.010948978830128908, 'timing/model': 1.5202811588533223, 'learning_rate/qwen_vl_interface': 3.945277016688647e-06, 'learning_rate/action_model': 3.945277016688647e-05, 'epoch': 1.21}) +[2026-06-16 05:49:38,272][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.011995028704404831, 'train/grad_norm_pre_clip': 4.930110931396484, 'timing/data': 0.012079315260052681, 'timing/model': 1.7688649408519268, 'learning_rate/qwen_vl_interface': 3.9423490023272326e-06, 'learning_rate/action_model': 3.9423490023272324e-05, 'epoch': 1.21}) +[2026-06-16 05:49:44,377][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.0015412837965413928, 'train/grad_norm_pre_clip': 0.5162742137908936, 'timing/data': 0.013596665114164352, 'timing/model': 1.5722881900146604, 'learning_rate/qwen_vl_interface': 3.939421525491748e-06, 'learning_rate/action_model': 3.9394215254917483e-05, 'epoch': 1.21}) +[2026-06-16 05:49:50,849][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.010616406798362732, 'train/grad_norm_pre_clip': 2.414630174636841, 'timing/data': 0.012329396791756153, 'timing/model': 1.6359346061944962, 'learning_rate/qwen_vl_interface': 3.9364945873855705e-06, 'learning_rate/action_model': 3.93649458738557e-05, 'epoch': 1.21}) +[2026-06-16 05:49:57,275][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.0846838727593422, 'train/grad_norm_pre_clip': 26.382455825805664, 'timing/data': 0.012425456661731005, 'timing/model': 1.591816138010472, 'learning_rate/qwen_vl_interface': 3.9335681892118536e-06, 'learning_rate/action_model': 3.933568189211854e-05, 'epoch': 1.21}) +[2026-06-16 05:50:03,654][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.007078397553414106, 'train/grad_norm_pre_clip': 5.137726306915283, 'timing/data': 0.01172867277637124, 'timing/model': 1.5877280798740685, 'learning_rate/qwen_vl_interface': 3.930642332173527e-06, 'learning_rate/action_model': 3.930642332173527e-05, 'epoch': 1.21}) +[2026-06-16 05:50:10,032][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.01001711655408144, 'train/grad_norm_pre_clip': 1.8238049745559692, 'timing/data': 0.011125539895147085, 'timing/model': 1.8129754117690027, 'learning_rate/qwen_vl_interface': 3.927717017473301e-06, 'learning_rate/action_model': 3.9277170174733006e-05, 'epoch': 1.21}) +[2026-06-16 05:50:16,065][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.016601718962192535, 'train/grad_norm_pre_clip': 8.099894523620605, 'timing/data': 0.010795698035508394, 'timing/model': 1.547314171679318, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 1.21}) +[2026-06-16 05:50:22,474][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.0663013905286789, 'train/grad_norm_pre_clip': 6.279052734375, 'timing/data': 0.011957355309277773, 'timing/model': 1.5801287041977048, 'learning_rate/qwen_vl_interface': 3.921868019896873e-06, 'learning_rate/action_model': 3.921868019896873e-05, 'epoch': 1.21}) +[2026-06-16 05:50:28,672][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.0495782233774662, 'train/grad_norm_pre_clip': 13.214200973510742, 'timing/data': 0.010885677766054869, 'timing/model': 1.570406321901828, 'learning_rate/qwen_vl_interface': 3.918944339424974e-06, 'learning_rate/action_model': 3.9189443394249736e-05, 'epoch': 1.21}) +[2026-06-16 05:50:34,981][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.14613258838653564, 'train/grad_norm_pre_clip': 10.957685470581055, 'timing/data': 0.011288781184703112, 'timing/model': 1.603990619070828, 'learning_rate/qwen_vl_interface': 3.9160212060997774e-06, 'learning_rate/action_model': 3.9160212060997774e-05, 'epoch': 1.21}) +[2026-06-16 05:50:41,259][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.2148124873638153, 'train/grad_norm_pre_clip': 5.709625244140625, 'timing/data': 0.01173435291275382, 'timing/model': 1.554431748110801, 'learning_rate/qwen_vl_interface': 3.913098621122875e-06, 'learning_rate/action_model': 3.913098621122875e-05, 'epoch': 1.21}) +[2026-06-16 05:50:47,604][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.023410245776176453, 'train/grad_norm_pre_clip': 6.684384822845459, 'timing/data': 0.01097871782258153, 'timing/model': 1.798424475826323, 'learning_rate/qwen_vl_interface': 3.91017658569563e-06, 'learning_rate/action_model': 3.91017658569563e-05, 'epoch': 1.21}) +[2026-06-16 05:50:53,668][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.001504014478996396, 'train/grad_norm_pre_clip': 0.304948627948761, 'timing/data': 0.010483026038855314, 'timing/model': 1.5642959750257432, 'learning_rate/qwen_vl_interface': 3.907255101019181e-06, 'learning_rate/action_model': 3.907255101019181e-05, 'epoch': 1.21}) +[2026-06-16 05:50:59,961][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.012980841100215912, 'train/grad_norm_pre_clip': 3.104236364364624, 'timing/data': 0.010914558079093695, 'timing/model': 1.5645447047427297, 'learning_rate/qwen_vl_interface': 3.904334168294442e-06, 'learning_rate/action_model': 3.9043341682944415e-05, 'epoch': 1.21}) +[2026-06-16 05:51:06,269][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.0015586455119773746, 'train/grad_norm_pre_clip': 0.498600035905838, 'timing/data': 0.010919338092207909, 'timing/model': 1.5716348867863417, 'learning_rate/qwen_vl_interface': 3.901413788722093e-06, 'learning_rate/action_model': 3.901413788722093e-05, 'epoch': 1.21}) +[2026-06-16 06:02:20,431][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.013254960998892784, 'train/grad_norm_pre_clip': 3.223707675933838, 'eval/loss': 0.036789894104003906, 'eval/flap_precision': 0.9597855227882037, 'eval/flap_recall': 0.9768076398362893, 'eval/flap_f1': 0.9682217714672076, 'eval/flap_rate_model': 0.20722222222222222, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9682217714672076, 'timing/data': 0.010785786900669336, 'timing/model': 1.5485675111413002, 'learning_rate/qwen_vl_interface': 3.8984939635025966e-06, 'learning_rate/action_model': 3.898493963502596e-05, 'epoch': 1.21}) +[2026-06-16 06:02:20,432][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3000_state +[2026-06-16 06:02:20,439][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 06:02:29,491][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3000_state/model.safetensors +[2026-06-16 06:02:49,433][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3000_state/optimizer.bin +[2026-06-16 06:02:49,434][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3000_state/sampler.bin +[2026-06-16 06:02:49,434][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-16 06:02:49,435][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-16 06:02:49,632][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-16 06:02:49,633][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 06:02:49,648][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 06:02:55,917][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.0023825610987842083, 'train/grad_norm_pre_clip': 0.753922164440155, 'timing/data': 0.011701211798936129, 'timing/model': 1.5517129367217422, 'learning_rate/qwen_vl_interface': 3.895574693836181e-06, 'learning_rate/action_model': 3.8955746938361804e-05, 'epoch': 1.21}) +[2026-06-16 06:03:02,125][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.00852272193878889, 'train/grad_norm_pre_clip': 3.9177796840667725, 'timing/data': 0.01129171997308731, 'timing/model': 1.532064024824649, 'learning_rate/qwen_vl_interface': 3.892655980922847e-06, 'learning_rate/action_model': 3.8926559809228466e-05, 'epoch': 1.21}) +[2026-06-16 06:03:08,224][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.03744708001613617, 'train/grad_norm_pre_clip': 29.334197998046875, 'timing/data': 0.010291993152350187, 'timing/model': 1.5543693415820599, 'learning_rate/qwen_vl_interface': 3.889737825962368e-06, 'learning_rate/action_model': 3.889737825962368e-05, 'epoch': 1.21}) +[2026-06-16 06:03:14,534][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.0004712895315606147, 'train/grad_norm_pre_clip': 0.057537853717803955, 'timing/data': 0.011252879165112972, 'timing/model': 1.8081860719248652, 'learning_rate/qwen_vl_interface': 3.886820230154288e-06, 'learning_rate/action_model': 3.886820230154287e-05, 'epoch': 1.21}) +[2026-06-16 06:03:20,584][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.11240291595458984, 'train/grad_norm_pre_clip': 16.380931854248047, 'timing/data': 0.010835136752575636, 'timing/model': 1.60625009983778, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 1.22}) +[2026-06-16 06:03:26,917][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.01612653024494648, 'train/grad_norm_pre_clip': 13.484771728515625, 'timing/data': 0.010801136959344149, 'timing/model': 1.637547181919217, 'learning_rate/qwen_vl_interface': 3.880986720792344e-06, 'learning_rate/action_model': 3.880986720792344e-05, 'epoch': 1.22}) +[2026-06-16 06:03:33,135][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.001601558760739863, 'train/grad_norm_pre_clip': 0.49897798895835876, 'timing/data': 0.010681645944714546, 'timing/model': 1.5552576459012926, 'learning_rate/qwen_vl_interface': 3.878070809636416e-06, 'learning_rate/action_model': 3.878070809636416e-05, 'epoch': 1.22}) +[2026-06-16 06:03:39,235][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.022562574595212936, 'train/grad_norm_pre_clip': 11.178544044494629, 'timing/data': 0.010751746129244566, 'timing/model': 1.5558735788799822, 'learning_rate/qwen_vl_interface': 3.875155462428758e-06, 'learning_rate/action_model': 3.875155462428758e-05, 'epoch': 1.22}) +[2026-06-16 06:03:45,518][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.0028530783019959927, 'train/grad_norm_pre_clip': 0.44061970710754395, 'timing/data': 0.010665304958820343, 'timing/model': 1.6127176778391004, 'learning_rate/qwen_vl_interface': 3.872240680367754e-06, 'learning_rate/action_model': 3.872240680367754e-05, 'epoch': 1.22}) +[2026-06-16 06:03:51,778][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.21416135132312775, 'train/grad_norm_pre_clip': 25.735595703125, 'timing/data': 0.010679774917662144, 'timing/model': 1.798170377034694, 'learning_rate/qwen_vl_interface': 3.869326464651566e-06, 'learning_rate/action_model': 3.869326464651566e-05, 'epoch': 1.22}) +[2026-06-16 06:03:57,850][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.12208659201860428, 'train/grad_norm_pre_clip': 19.50249481201172, 'timing/data': 0.01108931889757514, 'timing/model': 1.5650040828622878, 'learning_rate/qwen_vl_interface': 3.866412816478114e-06, 'learning_rate/action_model': 3.8664128164781144e-05, 'epoch': 1.22}) +[2026-06-16 06:04:04,154][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.23454488813877106, 'train/grad_norm_pre_clip': 16.303754806518555, 'timing/data': 0.010797056369483471, 'timing/model': 1.5693499399349093, 'learning_rate/qwen_vl_interface': 3.8634997370450915e-06, 'learning_rate/action_model': 3.8634997370450914e-05, 'epoch': 1.22}) +[2026-06-16 06:04:10,472][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.005198098719120026, 'train/grad_norm_pre_clip': 3.6426944732666016, 'timing/data': 0.010376303922384977, 'timing/model': 1.5898181553930044, 'learning_rate/qwen_vl_interface': 3.860587227549956e-06, 'learning_rate/action_model': 3.860587227549955e-05, 'epoch': 1.22}) +[2026-06-16 06:04:16,756][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.0021820159163326025, 'train/grad_norm_pre_clip': 0.6965164542198181, 'timing/data': 0.01128156902268529, 'timing/model': 1.570853057783097, 'learning_rate/qwen_vl_interface': 3.857675289189926e-06, 'learning_rate/action_model': 3.857675289189926e-05, 'epoch': 1.22}) +[2026-06-16 06:04:23,050][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.002815249375998974, 'train/grad_norm_pre_clip': 0.865796685218811, 'timing/data': 0.011884063016623259, 'timing/model': 1.5171940489672124, 'learning_rate/qwen_vl_interface': 3.854763923161993e-06, 'learning_rate/action_model': 3.854763923161992e-05, 'epoch': 1.22}) +[2026-06-16 06:04:29,391][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.08924825489521027, 'train/grad_norm_pre_clip': 11.139625549316406, 'timing/data': 0.010930996853858232, 'timing/model': 1.7904534260742366, 'learning_rate/qwen_vl_interface': 3.851853130662908e-06, 'learning_rate/action_model': 3.8518531306629077e-05, 'epoch': 1.22}) +[2026-06-16 06:04:35,457][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.0012096492573618889, 'train/grad_norm_pre_clip': 0.17757362127304077, 'timing/data': 0.010760616045445204, 'timing/model': 1.5549453287385404, 'learning_rate/qwen_vl_interface': 3.8489429128891884e-06, 'learning_rate/action_model': 3.848942912889188e-05, 'epoch': 1.22}) +[2026-06-16 06:04:41,764][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.02316138707101345, 'train/grad_norm_pre_clip': 24.951915740966797, 'timing/data': 0.010411133989691734, 'timing/model': 1.5142155177891254, 'learning_rate/qwen_vl_interface': 3.846033271037114e-06, 'learning_rate/action_model': 3.8460332710371136e-05, 'epoch': 1.22}) +[2026-06-16 06:04:48,049][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.04135971888899803, 'train/grad_norm_pre_clip': 11.308391571044922, 'timing/data': 0.01097806729376316, 'timing/model': 1.5542276138439775, 'learning_rate/qwen_vl_interface': 3.843124206302732e-06, 'learning_rate/action_model': 3.8431242063027316e-05, 'epoch': 1.22}) +[2026-06-16 06:04:54,453][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.06773930042982101, 'train/grad_norm_pre_clip': 27.938663482666016, 'timing/data': 0.010573574807494879, 'timing/model': 1.5760690779425204, 'learning_rate/qwen_vl_interface': 3.840215719881844e-06, 'learning_rate/action_model': 3.8402157198818445e-05, 'epoch': 1.22}) +[2026-06-16 06:05:00,754][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.05693131685256958, 'train/grad_norm_pre_clip': 7.494884967803955, 'timing/data': 0.010890947189182043, 'timing/model': 1.5199526031501591, 'learning_rate/qwen_vl_interface': 3.837307812970022e-06, 'learning_rate/action_model': 3.837307812970022e-05, 'epoch': 1.22}) +[2026-06-16 06:05:07,099][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.3502415418624878, 'train/grad_norm_pre_clip': 18.814048767089844, 'timing/data': 0.011015737894922495, 'timing/model': 1.8049117722548544, 'learning_rate/qwen_vl_interface': 3.834400486762596e-06, 'learning_rate/action_model': 3.834400486762596e-05, 'epoch': 1.22}) +[2026-06-16 06:05:13,156][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.006696586962789297, 'train/grad_norm_pre_clip': 1.240841269493103, 'timing/data': 0.010889586992561817, 'timing/model': 1.5697678681463003, 'learning_rate/qwen_vl_interface': 3.8314937424546585e-06, 'learning_rate/action_model': 3.8314937424546584e-05, 'epoch': 1.22}) +[2026-06-16 06:05:19,461][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.014933585189282894, 'train/grad_norm_pre_clip': 7.800220489501953, 'timing/data': 0.010927686933428049, 'timing/model': 1.5706769628450274, 'learning_rate/qwen_vl_interface': 3.828587581241064e-06, 'learning_rate/action_model': 3.8285875812410635e-05, 'epoch': 1.22}) +[2026-06-16 06:05:25,758][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.13913492858409882, 'train/grad_norm_pre_clip': 26.90680503845215, 'timing/data': 0.011177167762070894, 'timing/model': 1.5656059719622135, 'learning_rate/qwen_vl_interface': 3.825682004316419e-06, 'learning_rate/action_model': 3.825682004316419e-05, 'epoch': 1.22}) +[2026-06-16 06:05:32,160][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.0014501166297122836, 'train/grad_norm_pre_clip': 0.1359884887933731, 'timing/data': 0.011100478935986757, 'timing/model': 1.5774782341904938, 'learning_rate/qwen_vl_interface': 3.822777012875104e-06, 'learning_rate/action_model': 3.822777012875103e-05, 'epoch': 1.22}) +[2026-06-16 06:05:38,457][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.23389655351638794, 'train/grad_norm_pre_clip': 32.4365234375, 'timing/data': 0.01096074702218175, 'timing/model': 1.5125696761533618, 'learning_rate/qwen_vl_interface': 3.819872608111248e-06, 'learning_rate/action_model': 3.819872608111248e-05, 'epoch': 1.22}) +[2026-06-16 06:05:44,880][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.001339579001069069, 'train/grad_norm_pre_clip': 0.3030063211917877, 'timing/data': 0.011325189843773842, 'timing/model': 1.8795598582364619, 'learning_rate/qwen_vl_interface': 3.816968791218741e-06, 'learning_rate/action_model': 3.816968791218741e-05, 'epoch': 1.22}) +[2026-06-16 06:05:51,126][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.0767674371600151, 'train/grad_norm_pre_clip': 11.956189155578613, 'timing/data': 0.011477449908852577, 'timing/model': 1.735165061429143, 'learning_rate/qwen_vl_interface': 3.8140655633912337e-06, 'learning_rate/action_model': 3.814065563391234e-05, 'epoch': 1.22}) +[2026-06-16 06:05:57,155][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.2044648677110672, 'train/grad_norm_pre_clip': 48.988094329833984, 'timing/data': 0.010698736179620028, 'timing/model': 1.555667287670076, 'learning_rate/qwen_vl_interface': 3.811162925822133e-06, 'learning_rate/action_model': 3.811162925822132e-05, 'epoch': 1.23}) +[2026-06-16 06:06:03,455][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.16569124162197113, 'train/grad_norm_pre_clip': 19.033451080322266, 'timing/data': 0.010597785003483295, 'timing/model': 1.5578029011376202, 'learning_rate/qwen_vl_interface': 3.8082608797046045e-06, 'learning_rate/action_model': 3.8082608797046045e-05, 'epoch': 1.23}) +[2026-06-16 06:06:09,853][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.389775812625885, 'train/grad_norm_pre_clip': 44.00934982299805, 'timing/data': 0.01127222878858447, 'timing/model': 1.5825371127575636, 'learning_rate/qwen_vl_interface': 3.8053594262315687e-06, 'learning_rate/action_model': 3.805359426231568e-05, 'epoch': 1.23}) +[2026-06-16 06:06:16,159][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.01330341026186943, 'train/grad_norm_pre_clip': 6.376672267913818, 'timing/data': 0.010926146991550922, 'timing/model': 1.5148191573098302, 'learning_rate/qwen_vl_interface': 3.802458566595704e-06, 'learning_rate/action_model': 3.802458566595704e-05, 'epoch': 1.23}) +[2026-06-16 06:06:22,486][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.05959903076291084, 'train/grad_norm_pre_clip': 18.706937789916992, 'timing/data': 0.01266002794727683, 'timing/model': 1.7853892068378627, 'learning_rate/qwen_vl_interface': 3.7995583019894444e-06, 'learning_rate/action_model': 3.7995583019894445e-05, 'epoch': 1.23}) +[2026-06-16 06:06:28,894][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.009465450420975685, 'train/grad_norm_pre_clip': 3.4110870361328125, 'timing/data': 0.011402439326047897, 'timing/model': 1.7942420518957078, 'learning_rate/qwen_vl_interface': 3.7966586336049782e-06, 'learning_rate/action_model': 3.796658633604978e-05, 'epoch': 1.23}) +[2026-06-16 06:06:35,074][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.10582303255796432, 'train/grad_norm_pre_clip': 11.824908256530762, 'timing/data': 0.014806091319769621, 'timing/model': 1.6171380537562072, 'learning_rate/qwen_vl_interface': 3.7937595626342516e-06, 'learning_rate/action_model': 3.793759562634252e-05, 'epoch': 1.23}) +[2026-06-16 06:06:41,455][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.0015094451373443007, 'train/grad_norm_pre_clip': 0.27129030227661133, 'timing/data': 0.01228748494759202, 'timing/model': 1.5583353429101408, 'learning_rate/qwen_vl_interface': 3.79086109026896e-06, 'learning_rate/action_model': 3.7908610902689604e-05, 'epoch': 1.23}) +[2026-06-16 06:06:47,767][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.06318534910678864, 'train/grad_norm_pre_clip': 7.465956211090088, 'timing/data': 0.011508690193295479, 'timing/model': 1.5593774691224098, 'learning_rate/qwen_vl_interface': 3.7879632177005584e-06, 'learning_rate/action_model': 3.787963217700558e-05, 'epoch': 1.23}) +[2026-06-16 06:06:54,162][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.2208363115787506, 'train/grad_norm_pre_clip': 10.922306060791016, 'timing/data': 0.010957826860249043, 'timing/model': 1.5894413627684116, 'learning_rate/qwen_vl_interface': 3.7850659461202523e-06, 'learning_rate/action_model': 3.785065946120252e-05, 'epoch': 1.23}) +[2026-06-16 06:07:00,504][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.002163028810173273, 'train/grad_norm_pre_clip': 0.34600839018821716, 'timing/data': 0.011608520988374949, 'timing/model': 1.802391008939594, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 1.23}) +[2026-06-16 06:07:06,562][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.1871674805879593, 'train/grad_norm_pre_clip': 9.51542854309082, 'timing/data': 0.011693071108311415, 'timing/model': 1.5606531761586666, 'learning_rate/qwen_vl_interface': 3.779273210687514e-06, 'learning_rate/action_model': 3.779273210687514e-05, 'epoch': 1.23}) +[2026-06-16 06:07:12,873][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.0027808693703264, 'train/grad_norm_pre_clip': 0.6770762801170349, 'timing/data': 0.01101301796734333, 'timing/model': 1.5752562740817666, 'learning_rate/qwen_vl_interface': 3.776377749216257e-06, 'learning_rate/action_model': 3.776377749216257e-05, 'epoch': 1.23}) +[2026-06-16 06:07:19,158][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.07691174745559692, 'train/grad_norm_pre_clip': 7.022790431976318, 'timing/data': 0.011840591672807932, 'timing/model': 1.533732830081135, 'learning_rate/qwen_vl_interface': 3.773482893495442e-06, 'learning_rate/action_model': 3.773482893495442e-05, 'epoch': 1.23}) +[2026-06-16 06:07:25,461][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.0010037784231826663, 'train/grad_norm_pre_clip': 0.27211761474609375, 'timing/data': 0.010749805252999067, 'timing/model': 1.5166300949640572, 'learning_rate/qwen_vl_interface': 3.7705886447150375e-06, 'learning_rate/action_model': 3.7705886447150376e-05, 'epoch': 1.23}) +[2026-06-16 06:07:31,944][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.03545882925391197, 'train/grad_norm_pre_clip': 19.46685218811035, 'timing/data': 0.010638876352459192, 'timing/model': 1.628195557743311, 'learning_rate/qwen_vl_interface': 3.767695004064759e-06, 'learning_rate/action_model': 3.767695004064759e-05, 'epoch': 1.23}) +[2026-06-16 06:07:38,161][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.0044517009519040585, 'train/grad_norm_pre_clip': 0.7348846197128296, 'timing/data': 0.010767645668238401, 'timing/model': 1.7652863790281117, 'learning_rate/qwen_vl_interface': 3.7648019727340725e-06, 'learning_rate/action_model': 3.764801972734072e-05, 'epoch': 1.23}) +[2026-06-16 06:07:44,276][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.0069622197188436985, 'train/grad_norm_pre_clip': 1.1062959432601929, 'timing/data': 0.01110349828377366, 'timing/model': 1.5842680991627276, 'learning_rate/qwen_vl_interface': 3.761909551912196e-06, 'learning_rate/action_model': 3.761909551912196e-05, 'epoch': 1.23}) +[2026-06-16 06:07:50,701][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.08659028261899948, 'train/grad_norm_pre_clip': 25.086416244506836, 'timing/data': 0.01115657901391387, 'timing/model': 1.6573311458341777, 'learning_rate/qwen_vl_interface': 3.75901774278809e-06, 'learning_rate/action_model': 3.7590177427880904e-05, 'epoch': 1.23}) +[2026-06-16 06:07:57,058][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.0005933876964263618, 'train/grad_norm_pre_clip': 0.08686523884534836, 'timing/data': 0.010692765936255455, 'timing/model': 1.5462174438871443, 'learning_rate/qwen_vl_interface': 3.756126546550472e-06, 'learning_rate/action_model': 3.756126546550472e-05, 'epoch': 1.23}) +[2026-06-16 06:08:03,447][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.0057393708266317844, 'train/grad_norm_pre_clip': 1.6284312009811401, 'timing/data': 0.011983374133706093, 'timing/model': 1.5981882032938302, 'learning_rate/qwen_vl_interface': 3.753235964387802e-06, 'learning_rate/action_model': 3.753235964387802e-05, 'epoch': 1.23}) +[2026-06-16 06:08:09,861][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.12473995238542557, 'train/grad_norm_pre_clip': 12.317412376403809, 'timing/data': 0.010501894168555737, 'timing/model': 1.5618230202235281, 'learning_rate/qwen_vl_interface': 3.750345997488289e-06, 'learning_rate/action_model': 3.7503459974882886e-05, 'epoch': 1.23}) +[2026-06-16 06:08:16,159][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.045018404722213745, 'train/grad_norm_pre_clip': 12.286455154418945, 'timing/data': 0.01177464285865426, 'timing/model': 1.7574282488785684, 'learning_rate/qwen_vl_interface': 3.7474566470398903e-06, 'learning_rate/action_model': 3.74745664703989e-05, 'epoch': 1.23}) +[2026-06-16 06:08:22,265][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.007378541864454746, 'train/grad_norm_pre_clip': 2.6521823406219482, 'timing/data': 0.011306359898298979, 'timing/model': 1.560653232038021, 'learning_rate/qwen_vl_interface': 3.744567914230307e-06, 'learning_rate/action_model': 3.744567914230307e-05, 'epoch': 1.23}) +[2026-06-16 06:08:28,567][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.22840139269828796, 'train/grad_norm_pre_clip': 28.907928466796875, 'timing/data': 0.011123528238385916, 'timing/model': 1.557085209991783, 'learning_rate/qwen_vl_interface': 3.7416798002469902e-06, 'learning_rate/action_model': 3.74167980024699e-05, 'epoch': 1.23}) +[2026-06-16 06:08:34,875][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.0015556514263153076, 'train/grad_norm_pre_clip': 0.26848679780960083, 'timing/data': 0.012891228776425123, 'timing/model': 1.5766216297633946, 'learning_rate/qwen_vl_interface': 3.7387923062771313e-06, 'learning_rate/action_model': 3.738792306277131e-05, 'epoch': 1.24}) +[2026-06-16 06:08:41,260][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.05927072837948799, 'train/grad_norm_pre_clip': 17.56882095336914, 'timing/data': 0.010690436232835054, 'timing/model': 1.5171339036896825, 'learning_rate/qwen_vl_interface': 3.7359054335076732e-06, 'learning_rate/action_model': 3.7359054335076734e-05, 'epoch': 1.24}) +[2026-06-16 06:08:47,608][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.000811598205473274, 'train/grad_norm_pre_clip': 0.1260300576686859, 'timing/data': 0.010571364779025316, 'timing/model': 1.8038598718121648, 'learning_rate/qwen_vl_interface': 3.7330191831253006e-06, 'learning_rate/action_model': 3.7330191831253e-05, 'epoch': 1.24}) +[2026-06-16 06:08:53,708][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.0008112454088404775, 'train/grad_norm_pre_clip': 0.14934521913528442, 'timing/data': 0.010895146988332272, 'timing/model': 1.6129790912382305, 'learning_rate/qwen_vl_interface': 3.7301335563164382e-06, 'learning_rate/action_model': 3.730133556316438e-05, 'epoch': 1.24}) +[2026-06-16 06:08:59,961][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.0918337032198906, 'train/grad_norm_pre_clip': 18.080533981323242, 'timing/data': 0.01125758932903409, 'timing/model': 1.555624180007726, 'learning_rate/qwen_vl_interface': 3.727248554267261e-06, 'learning_rate/action_model': 3.727248554267261e-05, 'epoch': 1.24}) +[2026-06-16 06:09:06,166][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.08725489675998688, 'train/grad_norm_pre_clip': 10.903651237487793, 'timing/data': 0.01187903294339776, 'timing/model': 1.564221081789583, 'learning_rate/qwen_vl_interface': 3.7243641781636838e-06, 'learning_rate/action_model': 3.7243641781636836e-05, 'epoch': 1.24}) +[2026-06-16 06:09:12,478][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.005191678646951914, 'train/grad_norm_pre_clip': 5.676097393035889, 'timing/data': 0.01136159896850586, 'timing/model': 1.5876057357527316, 'learning_rate/qwen_vl_interface': 3.721480429191364e-06, 'learning_rate/action_model': 3.721480429191364e-05, 'epoch': 1.24}) +[2026-06-16 06:09:18,781][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.0968359112739563, 'train/grad_norm_pre_clip': 16.6597843170166, 'timing/data': 0.012265994679182768, 'timing/model': 1.5827298648655415, 'learning_rate/qwen_vl_interface': 3.7185973085357045e-06, 'learning_rate/action_model': 3.718597308535705e-05, 'epoch': 1.24}) +[2026-06-16 06:09:25,114][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.010348360985517502, 'train/grad_norm_pre_clip': 12.964373588562012, 'timing/data': 0.011283169034868479, 'timing/model': 1.8030065661296248, 'learning_rate/qwen_vl_interface': 3.7157148173818443e-06, 'learning_rate/action_model': 3.715714817381844e-05, 'epoch': 1.24}) +[2026-06-16 06:09:31,172][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.09578843414783478, 'train/grad_norm_pre_clip': 32.731178283691406, 'timing/data': 0.011448380071669817, 'timing/model': 1.5669392379932106, 'learning_rate/qwen_vl_interface': 3.7128329569146672e-06, 'learning_rate/action_model': 3.7128329569146666e-05, 'epoch': 1.24}) +[2026-06-16 06:09:37,464][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.0024915854446589947, 'train/grad_norm_pre_clip': 0.8966973423957825, 'timing/data': 0.011417008936405182, 'timing/model': 1.5610076119191945, 'learning_rate/qwen_vl_interface': 3.7099517283187977e-06, 'learning_rate/action_model': 3.7099517283187976e-05, 'epoch': 1.24}) +[2026-06-16 06:09:43,761][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.19992558658123016, 'train/grad_norm_pre_clip': 28.772666931152344, 'timing/data': 0.012012324295938015, 'timing/model': 1.5577316209673882, 'learning_rate/qwen_vl_interface': 3.7070711327786e-06, 'learning_rate/action_model': 3.7070711327785997e-05, 'epoch': 1.24}) +[2026-06-16 06:09:50,061][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.1639612913131714, 'train/grad_norm_pre_clip': 23.779054641723633, 'timing/data': 0.011585201136767864, 'timing/model': 1.5195759171620011, 'learning_rate/qwen_vl_interface': 3.704191171478179e-06, 'learning_rate/action_model': 3.704191171478179e-05, 'epoch': 1.24}) +[2026-06-16 06:09:56,362][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.00524569908156991, 'train/grad_norm_pre_clip': 0.7990228533744812, 'timing/data': 0.01192090380936861, 'timing/model': 1.5559747889637947, 'learning_rate/qwen_vl_interface': 3.701311845601376e-06, 'learning_rate/action_model': 3.701311845601376e-05, 'epoch': 1.24}) +[2026-06-16 06:10:02,700][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.296952486038208, 'train/grad_norm_pre_clip': 178.54197692871094, 'timing/data': 0.0119851129129529, 'timing/model': 1.797038818243891, 'learning_rate/qwen_vl_interface': 3.6984331563317744e-06, 'learning_rate/action_model': 3.698433156331774e-05, 'epoch': 1.24}) +[2026-06-16 06:10:08,793][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.067092165350914, 'train/grad_norm_pre_clip': 17.47223472595215, 'timing/data': 0.011583961080759764, 'timing/model': 1.5510202292352915, 'learning_rate/qwen_vl_interface': 3.695555104852695e-06, 'learning_rate/action_model': 3.6955551048526946e-05, 'epoch': 1.24}) +[2026-06-16 06:10:15,190][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.0027514866087585688, 'train/grad_norm_pre_clip': 0.43385931849479675, 'timing/data': 0.011179838329553604, 'timing/model': 1.5691220699809492, 'learning_rate/qwen_vl_interface': 3.6926776923471948e-06, 'learning_rate/action_model': 3.6926776923471946e-05, 'epoch': 1.24}) +[2026-06-16 06:10:21,716][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.0016487677348777652, 'train/grad_norm_pre_clip': 0.3679139316082001, 'timing/data': 0.014888571109622717, 'timing/model': 1.5828817337751389, 'learning_rate/qwen_vl_interface': 3.6898009199980715e-06, 'learning_rate/action_model': 3.6898009199980714e-05, 'epoch': 1.24}) +[2026-06-16 06:10:28,099][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.0009032204397954047, 'train/grad_norm_pre_clip': 0.22327451407909393, 'timing/data': 0.011363939382135868, 'timing/model': 1.6267162719741464, 'learning_rate/qwen_vl_interface': 3.686924788987855e-06, 'learning_rate/action_model': 3.686924788987855e-05, 'epoch': 1.24}) +[2026-06-16 06:10:34,397][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.004879206418991089, 'train/grad_norm_pre_clip': 3.1657626628875732, 'timing/data': 0.011549150105565786, 'timing/model': 1.5521823461167514, 'learning_rate/qwen_vl_interface': 3.6840493004988155e-06, 'learning_rate/action_model': 3.684049300498815e-05, 'epoch': 1.24}) +[2026-06-16 06:10:40,696][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.0005790549330413342, 'train/grad_norm_pre_clip': 0.13550299406051636, 'timing/data': 0.010873006656765938, 'timing/model': 1.7993864198215306, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 1.24}) +[2026-06-16 06:10:46,762][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.002090762834995985, 'train/grad_norm_pre_clip': 0.645612359046936, 'timing/data': 0.011111208237707615, 'timing/model': 1.5637333751656115, 'learning_rate/qwen_vl_interface': 3.6783002558120206e-06, 'learning_rate/action_model': 3.67830025581202e-05, 'epoch': 1.24}) +[2026-06-16 06:10:53,074][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.08865705132484436, 'train/grad_norm_pre_clip': 26.935958862304688, 'timing/data': 0.012030254118144512, 'timing/model': 1.5667958441190422, 'learning_rate/qwen_vl_interface': 3.6754267019774804e-06, 'learning_rate/action_model': 3.67542670197748e-05, 'epoch': 1.24}) +[2026-06-16 06:10:59,363][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.16354621946811676, 'train/grad_norm_pre_clip': 11.667351722717285, 'timing/data': 0.011278869118541479, 'timing/model': 1.5664933822117746, 'learning_rate/qwen_vl_interface': 3.6725537953905453e-06, 'learning_rate/action_model': 3.672553795390545e-05, 'epoch': 1.24}) +[2026-06-16 06:11:05,662][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.07374480366706848, 'train/grad_norm_pre_clip': 25.554574966430664, 'timing/data': 0.011499519925564528, 'timing/model': 1.5615242905914783, 'learning_rate/qwen_vl_interface': 3.6696815372321603e-06, 'learning_rate/action_model': 3.6696815372321604e-05, 'epoch': 1.25}) +[2026-06-16 06:11:12,061][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.005914686247706413, 'train/grad_norm_pre_clip': 4.251303672790527, 'timing/data': 0.01121998904272914, 'timing/model': 1.5172353689558804, 'learning_rate/qwen_vl_interface': 3.6668099286830023e-06, 'learning_rate/action_model': 3.666809928683002e-05, 'epoch': 1.25}) +[2026-06-16 06:11:18,411][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.004241305869072676, 'train/grad_norm_pre_clip': 1.5355976819992065, 'timing/data': 0.011330740060657263, 'timing/model': 1.8085390236228704, 'learning_rate/qwen_vl_interface': 3.66393897092348e-06, 'learning_rate/action_model': 3.66393897092348e-05, 'epoch': 1.25}) +[2026-06-16 06:11:24,470][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.0013155783526599407, 'train/grad_norm_pre_clip': 0.38353973627090454, 'timing/data': 0.01331916218623519, 'timing/model': 1.566947814077139, 'learning_rate/qwen_vl_interface': 3.661068665133735e-06, 'learning_rate/action_model': 3.661068665133735e-05, 'epoch': 1.25}) +[2026-06-16 06:11:30,762][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.007894612848758698, 'train/grad_norm_pre_clip': 2.265507698059082, 'timing/data': 0.010659325402230024, 'timing/model': 1.5632388219237328, 'learning_rate/qwen_vl_interface': 3.6581990124936434e-06, 'learning_rate/action_model': 3.6581990124936434e-05, 'epoch': 1.25}) +[2026-06-16 06:11:37,080][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.0010440624319016933, 'train/grad_norm_pre_clip': 0.2647720277309418, 'timing/data': 0.011512950994074345, 'timing/model': 1.5712808999232948, 'learning_rate/qwen_vl_interface': 3.65533001418281e-06, 'learning_rate/action_model': 3.655330014182809e-05, 'epoch': 1.25}) +[2026-06-16 06:11:43,364][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.0009584373328834772, 'train/grad_norm_pre_clip': 0.2586422562599182, 'timing/data': 0.011860433034598827, 'timing/model': 1.5609808159060776, 'learning_rate/qwen_vl_interface': 3.652461671380572e-06, 'learning_rate/action_model': 3.652461671380572e-05, 'epoch': 1.25}) +[2026-06-16 06:11:49,756][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.001463850261643529, 'train/grad_norm_pre_clip': 0.7237731218338013, 'timing/data': 0.011206187773495913, 'timing/model': 1.5985431969165802, 'learning_rate/qwen_vl_interface': 3.6495939852659966e-06, 'learning_rate/action_model': 3.649593985265997e-05, 'epoch': 1.25}) +[2026-06-16 06:11:56,098][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.02411164529621601, 'train/grad_norm_pre_clip': 21.872894287109375, 'timing/data': 0.011304389219731092, 'timing/model': 1.7592423120513558, 'learning_rate/qwen_vl_interface': 3.646726957017881e-06, 'learning_rate/action_model': 3.646726957017881e-05, 'epoch': 1.25}) +[2026-06-16 06:12:02,164][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.1459309458732605, 'train/grad_norm_pre_clip': 53.242061614990234, 'timing/data': 0.010701585095375776, 'timing/model': 1.553709362167865, 'learning_rate/qwen_vl_interface': 3.643860587814752e-06, 'learning_rate/action_model': 3.643860587814752e-05, 'epoch': 1.25}) +[2026-06-16 06:12:08,460][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.11408382654190063, 'train/grad_norm_pre_clip': 21.302244186401367, 'timing/data': 0.011093018110841513, 'timing/model': 1.5580239570699632, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 1.25}) +[2026-06-16 06:12:14,763][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.0016381266759708524, 'train/grad_norm_pre_clip': 0.5311465859413147, 'timing/data': 0.011021947022527456, 'timing/model': 1.5600088997744024, 'learning_rate/qwen_vl_interface': 3.6381298312562095e-06, 'learning_rate/action_model': 3.6381298312562094e-05, 'epoch': 1.25}) +[2026-06-16 06:12:21,061][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.00016368957585655153, 'train/grad_norm_pre_clip': 0.07610167562961578, 'timing/data': 0.011562049854546785, 'timing/model': 1.556276727002114, 'learning_rate/qwen_vl_interface': 3.635265446256493e-06, 'learning_rate/action_model': 3.6352654462564926e-05, 'epoch': 1.25}) +[2026-06-16 06:12:27,369][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.20954179763793945, 'train/grad_norm_pre_clip': 12.38848876953125, 'timing/data': 0.011162857990711927, 'timing/model': 1.5220371568575501, 'learning_rate/qwen_vl_interface': 3.6324017250131565e-06, 'learning_rate/action_model': 3.6324017250131567e-05, 'epoch': 1.25}) +[2026-06-16 06:12:33,711][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.19535629451274872, 'train/grad_norm_pre_clip': 14.62488842010498, 'timing/data': 0.011636791285127401, 'timing/model': 1.8115894412621856, 'learning_rate/qwen_vl_interface': 3.6295386687033696e-06, 'learning_rate/action_model': 3.6295386687033695e-05, 'epoch': 1.25}) +[2026-06-16 06:12:39,775][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.000930843292735517, 'train/grad_norm_pre_clip': 0.6172029972076416, 'timing/data': 0.011545161250978708, 'timing/model': 1.5654636421240866, 'learning_rate/qwen_vl_interface': 3.6266762785040253e-06, 'learning_rate/action_model': 3.626676278504025e-05, 'epoch': 1.25}) +[2026-06-16 06:12:46,168][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.06353135406970978, 'train/grad_norm_pre_clip': 19.022754669189453, 'timing/data': 0.011401468887925148, 'timing/model': 1.5717871822416782, 'learning_rate/qwen_vl_interface': 3.6238145555917464e-06, 'learning_rate/action_model': 3.6238145555917465e-05, 'epoch': 1.25}) +[2026-06-16 06:12:52,471][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.00012560057803057134, 'train/grad_norm_pre_clip': 0.06767965108156204, 'timing/data': 0.011188407894223928, 'timing/model': 1.573282201308757, 'learning_rate/qwen_vl_interface': 3.6209535011428757e-06, 'learning_rate/action_model': 3.620953501142876e-05, 'epoch': 1.25}) +[2026-06-16 06:12:58,779][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 8.910726319300011e-05, 'train/grad_norm_pre_clip': 0.06587642431259155, 'timing/data': 0.01072163600474596, 'timing/model': 1.58065928472206, 'learning_rate/qwen_vl_interface': 3.6180931163334863e-06, 'learning_rate/action_model': 3.6180931163334863e-05, 'epoch': 1.25}) +[2026-06-16 06:13:05,163][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.01842748187482357, 'train/grad_norm_pre_clip': 17.404457092285156, 'timing/data': 0.01128858933225274, 'timing/model': 1.5298367040231824, 'learning_rate/qwen_vl_interface': 3.6152334023393747e-06, 'learning_rate/action_model': 3.6152334023393745e-05, 'epoch': 1.25}) +[2026-06-16 06:13:11,592][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.5640273094177246, 'train/grad_norm_pre_clip': 32.68559265136719, 'timing/data': 0.01119247917085886, 'timing/model': 1.7870955700054765, 'learning_rate/qwen_vl_interface': 3.6123743603360613e-06, 'learning_rate/action_model': 3.612374360336061e-05, 'epoch': 1.25}) +[2026-06-16 06:24:17,137][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.014951013959944248, 'train/grad_norm_pre_clip': 9.940021514892578, 'eval/loss': 0.07286837697029114, 'eval/flap_precision': 0.9296675191815856, 'eval/flap_recall': 0.9918144611186903, 'eval/flap_f1': 0.9597359735973596, 'eval/flap_rate_model': 0.21722222222222223, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9597359735973596, 'timing/data': 0.011794202961027622, 'timing/model': 1.7201857101172209, 'learning_rate/qwen_vl_interface': 3.6095159914987893e-06, 'learning_rate/action_model': 3.6095159914987894e-05, 'epoch': 1.25}) +[2026-06-16 06:24:17,139][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3100_state +[2026-06-16 06:24:17,145][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 06:24:26,175][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3100_state/model.safetensors +[2026-06-16 06:24:46,646][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3100_state/optimizer.bin +[2026-06-16 06:24:46,646][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3100_state/sampler.bin +[2026-06-16 06:24:46,647][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3100_state/sampler_1.bin +[2026-06-16 06:24:46,648][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3100_state/random_states_0.pkl +[2026-06-16 06:24:46,848][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3100 +[2026-06-16 06:24:46,848][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 06:24:46,864][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 06:24:53,317][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.15625730156898499, 'train/grad_norm_pre_clip': 52.69329071044922, 'timing/data': 0.011575673706829548, 'timing/model': 1.81634156499058, 'learning_rate/qwen_vl_interface': 3.6066582970025266e-06, 'learning_rate/action_model': 3.6066582970025264e-05, 'epoch': 1.25}) +[2026-06-16 06:24:59,602][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.00162781763356179, 'train/grad_norm_pre_clip': 0.48241668939590454, 'timing/data': 0.014116412028670311, 'timing/model': 1.7276454619131982, 'learning_rate/qwen_vl_interface': 3.603801278021964e-06, 'learning_rate/action_model': 3.6038012780219635e-05, 'epoch': 1.25}) +[2026-06-16 06:25:05,844][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.004094156436622143, 'train/grad_norm_pre_clip': 3.2639167308807373, 'timing/data': 0.011761595960706472, 'timing/model': 1.547846139408648, 'learning_rate/qwen_vl_interface': 3.6009449357315122e-06, 'learning_rate/action_model': 3.6009449357315125e-05, 'epoch': 1.25}) +[2026-06-16 06:25:12,042][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.2584262490272522, 'train/grad_norm_pre_clip': 23.270538330078125, 'timing/data': 0.012927244883030653, 'timing/model': 1.568022696301341, 'learning_rate/qwen_vl_interface': 3.5980892713053084e-06, 'learning_rate/action_model': 3.598089271305308e-05, 'epoch': 1.26}) +[2026-06-16 06:25:18,347][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.18621087074279785, 'train/grad_norm_pre_clip': 35.70448303222656, 'timing/data': 0.010758117772638798, 'timing/model': 1.5797674371860921, 'learning_rate/qwen_vl_interface': 3.5952342859172063e-06, 'learning_rate/action_model': 3.595234285917206e-05, 'epoch': 1.26}) +[2026-06-16 06:25:24,756][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.00022159104992169887, 'train/grad_norm_pre_clip': 0.048428744077682495, 'timing/data': 0.011327199172228575, 'timing/model': 1.5390353798866272, 'learning_rate/qwen_vl_interface': 3.592379980740785e-06, 'learning_rate/action_model': 3.5923799807407845e-05, 'epoch': 1.26}) +[2026-06-16 06:25:31,147][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.10411074757575989, 'train/grad_norm_pre_clip': 23.3356990814209, 'timing/data': 0.010710619855672121, 'timing/model': 1.864995234645903, 'learning_rate/qwen_vl_interface': 3.5895263569493384e-06, 'learning_rate/action_model': 3.589526356949338e-05, 'epoch': 1.26}) +[2026-06-16 06:25:37,593][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.0003447607159614563, 'train/grad_norm_pre_clip': 0.11448902636766434, 'timing/data': 0.009649872314184904, 'timing/model': 1.8130199392326176, 'learning_rate/qwen_vl_interface': 3.586673415715885e-06, 'learning_rate/action_model': 3.586673415715885e-05, 'epoch': 1.26}) +[2026-06-16 06:25:43,738][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.14547210931777954, 'train/grad_norm_pre_clip': 31.31662368774414, 'timing/data': 0.010667741764336824, 'timing/model': 1.5543271666392684, 'learning_rate/qwen_vl_interface': 3.5838211582131603e-06, 'learning_rate/action_model': 3.58382115821316e-05, 'epoch': 1.26}) +[2026-06-16 06:25:50,036][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.0007873161812312901, 'train/grad_norm_pre_clip': 0.2570212781429291, 'timing/data': 0.009623752906918526, 'timing/model': 1.5666439710184932, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 1.26}) +[2026-06-16 06:25:56,387][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.004910604562610388, 'train/grad_norm_pre_clip': 3.7429146766662598, 'timing/data': 0.009785054251551628, 'timing/model': 1.6023243349045515, 'learning_rate/qwen_vl_interface': 3.5781186990894423e-06, 'learning_rate/action_model': 3.578118699089442e-05, 'epoch': 1.26}) +[2026-06-16 06:26:02,668][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.4637662470340729, 'train/grad_norm_pre_clip': 35.8880729675293, 'timing/data': 0.011940914206206799, 'timing/model': 1.5300043798051775, 'learning_rate/qwen_vl_interface': 3.5752684998125113e-06, 'learning_rate/action_model': 3.5752684998125116e-05, 'epoch': 1.26}) +[2026-06-16 06:26:09,014][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.00025420449674129486, 'train/grad_norm_pre_clip': 0.07456454634666443, 'timing/data': 0.013251123949885368, 'timing/model': 1.812800016719848, 'learning_rate/qwen_vl_interface': 3.5724189889544407e-06, 'learning_rate/action_model': 3.572418988954441e-05, 'epoch': 1.26}) +[2026-06-16 06:26:15,046][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.1387932449579239, 'train/grad_norm_pre_clip': 30.864776611328125, 'timing/data': 0.011300595011562109, 'timing/model': 1.551550782751292, 'learning_rate/qwen_vl_interface': 3.569570167686556e-06, 'learning_rate/action_model': 3.569570167686556e-05, 'epoch': 1.26}) +[2026-06-16 06:26:21,379][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.041842348873615265, 'train/grad_norm_pre_clip': 39.30784225463867, 'timing/data': 0.01096941577270627, 'timing/model': 1.5821661623194814, 'learning_rate/qwen_vl_interface': 3.5667220371798995e-06, 'learning_rate/action_model': 3.566722037179899e-05, 'epoch': 1.26}) +[2026-06-16 06:26:27,754][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.001425791415385902, 'train/grad_norm_pre_clip': 0.796326220035553, 'timing/data': 0.011999226175248623, 'timing/model': 1.546067401766777, 'learning_rate/qwen_vl_interface': 3.563874598605232e-06, 'learning_rate/action_model': 3.5638745986052316e-05, 'epoch': 1.26}) +[2026-06-16 06:26:34,101][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.0008889238815754652, 'train/grad_norm_pre_clip': 0.15723161399364471, 'timing/data': 0.011340036988258362, 'timing/model': 1.5492326263338327, 'learning_rate/qwen_vl_interface': 3.561027853133025e-06, 'learning_rate/action_model': 3.561027853133025e-05, 'epoch': 1.26}) +[2026-06-16 06:26:40,355][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.42735904455184937, 'train/grad_norm_pre_clip': 20.496891021728516, 'timing/data': 0.011697757989168167, 'timing/model': 1.5609036390669644, 'learning_rate/qwen_vl_interface': 3.558181801933472e-06, 'learning_rate/action_model': 3.5581818019334714e-05, 'epoch': 1.26}) +[2026-06-16 06:26:46,675][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.005099643021821976, 'train/grad_norm_pre_clip': 2.5620789527893066, 'timing/data': 0.011543938890099525, 'timing/model': 1.7783821350894868, 'learning_rate/qwen_vl_interface': 3.555336446176473e-06, 'learning_rate/action_model': 3.5553364461764726e-05, 'epoch': 1.26}) +[2026-06-16 06:26:52,757][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.20200127363204956, 'train/grad_norm_pre_clip': 18.568239212036133, 'timing/data': 0.011194308754056692, 'timing/model': 1.5736759509891272, 'learning_rate/qwen_vl_interface': 3.552491787031647e-06, 'learning_rate/action_model': 3.552491787031647e-05, 'epoch': 1.26}) +[2026-06-16 06:26:59,154][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.0991113930940628, 'train/grad_norm_pre_clip': 22.768348693847656, 'timing/data': 0.011445300187915564, 'timing/model': 1.5753751546144485, 'learning_rate/qwen_vl_interface': 3.5496478256683277e-06, 'learning_rate/action_model': 3.549647825668327e-05, 'epoch': 1.26}) +[2026-06-16 06:27:05,560][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.00338515592738986, 'train/grad_norm_pre_clip': 1.3178547620773315, 'timing/data': 0.01130937086418271, 'timing/model': 1.5916872890666127, 'learning_rate/qwen_vl_interface': 3.5468045632555585e-06, 'learning_rate/action_model': 3.546804563255559e-05, 'epoch': 1.26}) +[2026-06-16 06:27:11,858][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.13299375772476196, 'train/grad_norm_pre_clip': 13.220648765563965, 'timing/data': 0.01131901191547513, 'timing/model': 1.5261210510507226, 'learning_rate/qwen_vl_interface': 3.5439620009620995e-06, 'learning_rate/action_model': 3.543962000962099e-05, 'epoch': 1.26}) +[2026-06-16 06:27:18,162][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.07612235099077225, 'train/grad_norm_pre_clip': 27.798433303833008, 'timing/data': 0.010888452176004648, 'timing/model': 1.5580665227025747, 'learning_rate/qwen_vl_interface': 3.541120139956419e-06, 'learning_rate/action_model': 3.5411201399564184e-05, 'epoch': 1.26}) +[2026-06-16 06:27:24,528][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.0027978222351521254, 'train/grad_norm_pre_clip': 1.656856656074524, 'timing/data': 0.011268292088061571, 'timing/model': 1.784653261769563, 'learning_rate/qwen_vl_interface': 3.538278981406697e-06, 'learning_rate/action_model': 3.5382789814066966e-05, 'epoch': 1.26}) +[2026-06-16 06:27:30,678][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.0028819304425269365, 'train/grad_norm_pre_clip': 0.7934181094169617, 'timing/data': 0.011058252770453691, 'timing/model': 1.5888412441127002, 'learning_rate/qwen_vl_interface': 3.5354385264808295e-06, 'learning_rate/action_model': 3.535438526480829e-05, 'epoch': 1.26}) +[2026-06-16 06:27:37,060][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.0270302202552557, 'train/grad_norm_pre_clip': 69.83804321289062, 'timing/data': 0.01140999374911189, 'timing/model': 1.569051994010806, 'learning_rate/qwen_vl_interface': 3.5325987763464196e-06, 'learning_rate/action_model': 3.5325987763464194e-05, 'epoch': 1.26}) +[2026-06-16 06:27:43,353][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.0042176698334515095, 'train/grad_norm_pre_clip': 1.2958617210388184, 'timing/data': 0.010391493793576956, 'timing/model': 1.5616840468719602, 'learning_rate/qwen_vl_interface': 3.529759732170781e-06, 'learning_rate/action_model': 3.529759732170781e-05, 'epoch': 1.26}) +[2026-06-16 06:27:49,653][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.001241442863829434, 'train/grad_norm_pre_clip': 0.3210379183292389, 'timing/data': 0.011627855245023966, 'timing/model': 1.5163629301823676, 'learning_rate/qwen_vl_interface': 3.5269213951209375e-06, 'learning_rate/action_model': 3.5269213951209374e-05, 'epoch': 1.27}) +[2026-06-16 06:27:55,952][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.001993704354390502, 'train/grad_norm_pre_clip': 0.9301720261573792, 'timing/data': 0.01078730495646596, 'timing/model': 1.5695409337058663, 'learning_rate/qwen_vl_interface': 3.5240837663636207e-06, 'learning_rate/action_model': 3.5240837663636204e-05, 'epoch': 1.27}) +[2026-06-16 06:28:02,295][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.4316749572753906, 'train/grad_norm_pre_clip': 20.08795738220215, 'timing/data': 0.010785476304590702, 'timing/model': 1.7973954598419368, 'learning_rate/qwen_vl_interface': 3.5212468470652745e-06, 'learning_rate/action_model': 3.521246847065274e-05, 'epoch': 1.27}) +[2026-06-16 06:28:08,357][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.050509653985500336, 'train/grad_norm_pre_clip': 32.139530181884766, 'timing/data': 0.010681636165827513, 'timing/model': 1.5571549669839442, 'learning_rate/qwen_vl_interface': 3.5184106383920487e-06, 'learning_rate/action_model': 3.5184106383920486e-05, 'epoch': 1.27}) +[2026-06-16 06:28:14,656][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.00022386251657735556, 'train/grad_norm_pre_clip': 0.037755634635686874, 'timing/data': 0.010413476265966892, 'timing/model': 1.565196971874684, 'learning_rate/qwen_vl_interface': 3.5155751415098022e-06, 'learning_rate/action_model': 3.515575141509802e-05, 'epoch': 1.27}) +[2026-06-16 06:28:20,962][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.0001359119196422398, 'train/grad_norm_pre_clip': 0.038891322910785675, 'timing/data': 0.01092112809419632, 'timing/model': 1.561621556058526, 'learning_rate/qwen_vl_interface': 3.5127403575841e-06, 'learning_rate/action_model': 3.5127403575841e-05, 'epoch': 1.27}) +[2026-06-16 06:28:27,257][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.001976308412849903, 'train/grad_norm_pre_clip': 0.2600020170211792, 'timing/data': 0.010857787914574146, 'timing/model': 1.5242905556224287, 'learning_rate/qwen_vl_interface': 3.509906287780215e-06, 'learning_rate/action_model': 3.5099062877802144e-05, 'epoch': 1.27}) +[2026-06-16 06:28:33,558][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.1563819944858551, 'train/grad_norm_pre_clip': 11.246950149536133, 'timing/data': 0.011525350157171488, 'timing/model': 1.5483396700583398, 'learning_rate/qwen_vl_interface': 3.5070729332631247e-06, 'learning_rate/action_model': 3.507072933263125e-05, 'epoch': 1.27}) +[2026-06-16 06:28:39,887][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.0917881652712822, 'train/grad_norm_pre_clip': 19.811813354492188, 'timing/data': 0.010758698917925358, 'timing/model': 1.7881293748505414, 'learning_rate/qwen_vl_interface': 3.5042402951975167e-06, 'learning_rate/action_model': 3.5042402951975167e-05, 'epoch': 1.27}) +[2026-06-16 06:28:46,062][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.007609471678733826, 'train/grad_norm_pre_clip': 3.063692092895508, 'timing/data': 0.010782628785818815, 'timing/model': 1.6034199339337647, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 1.27}) +[2026-06-16 06:28:52,366][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.0018579125171527267, 'train/grad_norm_pre_clip': 0.44549813866615295, 'timing/data': 0.011532421223819256, 'timing/model': 1.5650586653500795, 'learning_rate/qwen_vl_interface': 3.498577173078011e-06, 'learning_rate/action_model': 3.498577173078011e-05, 'epoch': 1.27}) +[2026-06-16 06:28:58,672][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.12258688360452652, 'train/grad_norm_pre_clip': 22.552648544311523, 'timing/data': 0.013014412019401789, 'timing/model': 1.5696346033364534, 'learning_rate/qwen_vl_interface': 3.495746691352008e-06, 'learning_rate/action_model': 3.495746691352008e-05, 'epoch': 1.27}) +[2026-06-16 06:29:05,059][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.12498699873685837, 'train/grad_norm_pre_clip': 17.186969757080078, 'timing/data': 0.011219171341508627, 'timing/model': 1.5556056499481201, 'learning_rate/qwen_vl_interface': 3.4929169307332763e-06, 'learning_rate/action_model': 3.492916930733276e-05, 'epoch': 1.27}) +[2026-06-16 06:29:11,461][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.0005526908207684755, 'train/grad_norm_pre_clip': 0.08218013495206833, 'timing/data': 0.011504321824759245, 'timing/model': 1.5152408787980676, 'learning_rate/qwen_vl_interface': 3.4900878923850234e-06, 'learning_rate/action_model': 3.490087892385023e-05, 'epoch': 1.27}) +[2026-06-16 06:29:17,801][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.07503058016300201, 'train/grad_norm_pre_clip': 36.265708923339844, 'timing/data': 0.010777371935546398, 'timing/model': 1.808352324180305, 'learning_rate/qwen_vl_interface': 3.4872595774701607e-06, 'learning_rate/action_model': 3.487259577470161e-05, 'epoch': 1.27}) +[2026-06-16 06:29:23,866][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.0020844575483351946, 'train/grad_norm_pre_clip': 1.7901619672775269, 'timing/data': 0.010697011835873127, 'timing/model': 1.568318892735988, 'learning_rate/qwen_vl_interface': 3.4844319871513014e-06, 'learning_rate/action_model': 3.484431987151301e-05, 'epoch': 1.27}) +[2026-06-16 06:29:30,160][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.0011108313919976354, 'train/grad_norm_pre_clip': 0.1992451548576355, 'timing/data': 0.01247991481795907, 'timing/model': 1.559958390891552, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 1.27}) +[2026-06-16 06:29:36,473][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.024166196584701538, 'train/grad_norm_pre_clip': 25.67835235595703, 'timing/data': 0.010906393639743328, 'timing/model': 1.5743724070489407, 'learning_rate/qwen_vl_interface': 3.478778984950556e-06, 'learning_rate/action_model': 3.478778984950556e-05, 'epoch': 1.27}) +[2026-06-16 06:29:42,768][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.0030518139246851206, 'train/grad_norm_pre_clip': 1.7098196744918823, 'timing/data': 0.01328786788508296, 'timing/model': 1.5700038177892566, 'learning_rate/qwen_vl_interface': 3.475953575392405e-06, 'learning_rate/action_model': 3.475953575392405e-05, 'epoch': 1.27}) +[2026-06-16 06:29:49,124][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.019906653091311455, 'train/grad_norm_pre_clip': 38.393558502197266, 'timing/data': 0.010956595186144114, 'timing/model': 1.5733811389654875, 'learning_rate/qwen_vl_interface': 3.473128895077728e-06, 'learning_rate/action_model': 3.473128895077728e-05, 'epoch': 1.27}) +[2026-06-16 06:29:55,389][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.13481062650680542, 'train/grad_norm_pre_clip': 46.12485122680664, 'timing/data': 0.01117920596152544, 'timing/model': 1.7496576854027808, 'learning_rate/qwen_vl_interface': 3.4703049451676453e-06, 'learning_rate/action_model': 3.4703049451676454e-05, 'epoch': 1.27}) +[2026-06-16 06:30:01,465][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.0022160103544592857, 'train/grad_norm_pre_clip': 1.010454773902893, 'timing/data': 0.010466154664754868, 'timing/model': 1.5695369113236666, 'learning_rate/qwen_vl_interface': 3.4674817268229732e-06, 'learning_rate/action_model': 3.467481726822973e-05, 'epoch': 1.27}) +[2026-06-16 06:30:07,757][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.1544538289308548, 'train/grad_norm_pre_clip': 18.095735549926758, 'timing/data': 0.010308695025742054, 'timing/model': 1.5617779330350459, 'learning_rate/qwen_vl_interface': 3.464659241204232e-06, 'learning_rate/action_model': 3.4646592412042314e-05, 'epoch': 1.27}) +[2026-06-16 06:30:14,057][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.16723738610744476, 'train/grad_norm_pre_clip': 21.29189682006836, 'timing/data': 0.011137116234749556, 'timing/model': 1.5589949409477413, 'learning_rate/qwen_vl_interface': 3.4618374894716377e-06, 'learning_rate/action_model': 3.461837489471638e-05, 'epoch': 1.27}) +[2026-06-16 06:30:20,373][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.13764813542366028, 'train/grad_norm_pre_clip': 5.500260353088379, 'timing/data': 0.01195001881569624, 'timing/model': 1.553886964917183, 'learning_rate/qwen_vl_interface': 3.459016472785107e-06, 'learning_rate/action_model': 3.459016472785107e-05, 'epoch': 1.28}) +[2026-06-16 06:30:26,754][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.0020112365018576384, 'train/grad_norm_pre_clip': 0.2995060980319977, 'timing/data': 0.011251138057559729, 'timing/model': 1.5727779972366989, 'learning_rate/qwen_vl_interface': 3.456196192304254e-06, 'learning_rate/action_model': 3.456196192304254e-05, 'epoch': 1.28}) +[2026-06-16 06:30:33,205][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.009045177139341831, 'train/grad_norm_pre_clip': 7.570000171661377, 'timing/data': 0.01638572569936514, 'timing/model': 1.8156658089719713, 'learning_rate/qwen_vl_interface': 3.4533766491883887e-06, 'learning_rate/action_model': 3.453376649188388e-05, 'epoch': 1.28}) +[2026-06-16 06:30:39,266][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.005325458478182554, 'train/grad_norm_pre_clip': 2.5891072750091553, 'timing/data': 0.011211188975721598, 'timing/model': 1.566750266123563, 'learning_rate/qwen_vl_interface': 3.450557844596517e-06, 'learning_rate/action_model': 3.450557844596517e-05, 'epoch': 1.28}) +[2026-06-16 06:30:45,663][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.09598449617624283, 'train/grad_norm_pre_clip': 8.43536376953125, 'timing/data': 0.009720996953547001, 'timing/model': 1.5433278665877879, 'learning_rate/qwen_vl_interface': 3.447739779687345e-06, 'learning_rate/action_model': 3.4477397796873446e-05, 'epoch': 1.28}) +[2026-06-16 06:30:51,962][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.11984678357839584, 'train/grad_norm_pre_clip': 18.454755783081055, 'timing/data': 0.011127229779958725, 'timing/model': 1.5420167311094701, 'learning_rate/qwen_vl_interface': 3.4449224556192722e-06, 'learning_rate/action_model': 3.444922455619272e-05, 'epoch': 1.28}) +[2026-06-16 06:30:58,386][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.19724173843860626, 'train/grad_norm_pre_clip': 34.92057800292969, 'timing/data': 0.010754988994449377, 'timing/model': 1.5821471503004432, 'learning_rate/qwen_vl_interface': 3.442105873550395e-06, 'learning_rate/action_model': 3.442105873550395e-05, 'epoch': 1.28}) +[2026-06-16 06:31:04,776][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.07158578932285309, 'train/grad_norm_pre_clip': 21.730615615844727, 'timing/data': 0.01410559518262744, 'timing/model': 1.5221434808336198, 'learning_rate/qwen_vl_interface': 3.4392900346385022e-06, 'learning_rate/action_model': 3.4392900346385024e-05, 'epoch': 1.28}) +[2026-06-16 06:31:11,153][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.002256411360576749, 'train/grad_norm_pre_clip': 0.8311737179756165, 'timing/data': 0.010266219265758991, 'timing/model': 1.8472055909223855, 'learning_rate/qwen_vl_interface': 3.4364749400410796e-06, 'learning_rate/action_model': 3.43647494004108e-05, 'epoch': 1.28}) +[2026-06-16 06:31:17,545][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.0008490895852446556, 'train/grad_norm_pre_clip': 0.11428899317979813, 'timing/data': 0.011080640368163586, 'timing/model': 1.7574480511248112, 'learning_rate/qwen_vl_interface': 3.4336605909153066e-06, 'learning_rate/action_model': 3.433660590915306e-05, 'epoch': 1.28}) +[2026-06-16 06:31:23,656][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.25916385650634766, 'train/grad_norm_pre_clip': 20.328636169433594, 'timing/data': 0.010912070982158184, 'timing/model': 1.5522590400651097, 'learning_rate/qwen_vl_interface': 3.430846988418057e-06, 'learning_rate/action_model': 3.430846988418057e-05, 'epoch': 1.28}) +[2026-06-16 06:31:29,969][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.23257191479206085, 'train/grad_norm_pre_clip': 31.792932510375977, 'timing/data': 0.01109823165461421, 'timing/model': 1.5676689418032765, 'learning_rate/qwen_vl_interface': 3.428034133705893e-06, 'learning_rate/action_model': 3.428034133705893e-05, 'epoch': 1.28}) +[2026-06-16 06:31:36,459][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.3187454342842102, 'train/grad_norm_pre_clip': 27.999300003051758, 'timing/data': 0.011257042177021503, 'timing/model': 1.6728249164298177, 'learning_rate/qwen_vl_interface': 3.4252220279350763e-06, 'learning_rate/action_model': 3.425222027935076e-05, 'epoch': 1.28}) +[2026-06-16 06:31:42,859][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.07256639003753662, 'train/grad_norm_pre_clip': 14.478546142578125, 'timing/data': 0.01100118225440383, 'timing/model': 1.5927214217372239, 'learning_rate/qwen_vl_interface': 3.4224106722615547e-06, 'learning_rate/action_model': 3.422410672261554e-05, 'epoch': 1.28}) +[2026-06-16 06:31:49,208][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.06085652485489845, 'train/grad_norm_pre_clip': 19.494747161865234, 'timing/data': 0.010544281918555498, 'timing/model': 1.8091109357774258, 'learning_rate/qwen_vl_interface': 3.419600067840973e-06, 'learning_rate/action_model': 3.419600067840973e-05, 'epoch': 1.28}) +[2026-06-16 06:31:55,501][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.3415258526802063, 'train/grad_norm_pre_clip': 23.570999145507812, 'timing/data': 0.010834253393113613, 'timing/model': 1.8028743332251906, 'learning_rate/qwen_vl_interface': 3.4167902158286627e-06, 'learning_rate/action_model': 3.4167902158286626e-05, 'epoch': 1.28}) +[2026-06-16 06:32:01,566][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.0835108757019043, 'train/grad_norm_pre_clip': 18.944358825683594, 'timing/data': 0.011891535017639399, 'timing/model': 1.5658428007736802, 'learning_rate/qwen_vl_interface': 3.4139811173796468e-06, 'learning_rate/action_model': 3.413981117379647e-05, 'epoch': 1.28}) +[2026-06-16 06:32:07,868][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.0030202146153897047, 'train/grad_norm_pre_clip': 1.3089107275009155, 'timing/data': 0.011337594129145145, 'timing/model': 1.561442210804671, 'learning_rate/qwen_vl_interface': 3.411172773648642e-06, 'learning_rate/action_model': 3.411172773648642e-05, 'epoch': 1.28}) +[2026-06-16 06:32:14,163][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.011330587789416313, 'train/grad_norm_pre_clip': 8.855412483215332, 'timing/data': 0.013385838828980923, 'timing/model': 1.552464481908828, 'learning_rate/qwen_vl_interface': 3.408365185790052e-06, 'learning_rate/action_model': 3.4083651857900515e-05, 'epoch': 1.28}) +[2026-06-16 06:32:20,461][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.023658525198698044, 'train/grad_norm_pre_clip': 35.2457160949707, 'timing/data': 0.010584814008325338, 'timing/model': 1.523598066996783, 'learning_rate/qwen_vl_interface': 3.40555835495797e-06, 'learning_rate/action_model': 3.40555835495797e-05, 'epoch': 1.28}) +[2026-06-16 06:32:26,915][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.004777639638632536, 'train/grad_norm_pre_clip': 0.9731894731521606, 'timing/data': 0.01183397602289915, 'timing/model': 1.8271564221940935, 'learning_rate/qwen_vl_interface': 3.4027522823061763e-06, 'learning_rate/action_model': 3.4027522823061765e-05, 'epoch': 1.28}) +[2026-06-16 06:32:32,960][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.1915232241153717, 'train/grad_norm_pre_clip': 19.140960693359375, 'timing/data': 0.01090410491451621, 'timing/model': 1.5549733601510525, 'learning_rate/qwen_vl_interface': 3.399946968988143e-06, 'learning_rate/action_model': 3.399946968988143e-05, 'epoch': 1.28}) +[2026-06-16 06:32:39,374][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.21414117515087128, 'train/grad_norm_pre_clip': 63.39588928222656, 'timing/data': 0.011681126896291971, 'timing/model': 1.5710690952837467, 'learning_rate/qwen_vl_interface': 3.397142416157028e-06, 'learning_rate/action_model': 3.397142416157028e-05, 'epoch': 1.28}) +[2026-06-16 06:32:45,869][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.03806913644075394, 'train/grad_norm_pre_clip': 36.00080490112305, 'timing/data': 0.011133416090160608, 'timing/model': 1.5688740476034582, 'learning_rate/qwen_vl_interface': 3.3943386249656766e-06, 'learning_rate/action_model': 3.394338624965677e-05, 'epoch': 1.28}) +[2026-06-16 06:32:52,270][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.17457082867622375, 'train/grad_norm_pre_clip': 15.863933563232422, 'timing/data': 0.0112971062771976, 'timing/model': 1.6159250256605446, 'learning_rate/qwen_vl_interface': 3.391535596566624e-06, 'learning_rate/action_model': 3.391535596566624e-05, 'epoch': 1.28}) +[2026-06-16 06:32:58,566][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.1269703060388565, 'train/grad_norm_pre_clip': 21.169511795043945, 'timing/data': 0.011401037219911814, 'timing/model': 1.5846971590071917, 'learning_rate/qwen_vl_interface': 3.388733332112086e-06, 'learning_rate/action_model': 3.3887333321120856e-05, 'epoch': 1.29}) +[2026-06-16 06:33:04,846][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.2285933792591095, 'train/grad_norm_pre_clip': 40.63413619995117, 'timing/data': 0.01071847602725029, 'timing/model': 1.749318619724363, 'learning_rate/qwen_vl_interface': 3.3859318327539693e-06, 'learning_rate/action_model': 3.385931832753969e-05, 'epoch': 1.29}) +[2026-06-16 06:33:10,959][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.004732381086796522, 'train/grad_norm_pre_clip': 1.1689715385437012, 'timing/data': 0.01088545611128211, 'timing/model': 1.5626031132414937, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 1.29}) +[2026-06-16 06:33:17,377][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.012021088972687721, 'train/grad_norm_pre_clip': 6.070200443267822, 'timing/data': 0.011102186981588602, 'timing/model': 1.55337970610708, 'learning_rate/qwen_vl_interface': 3.3803311339330447e-06, 'learning_rate/action_model': 3.3803311339330444e-05, 'epoch': 1.29}) +[2026-06-16 06:33:23,661][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.08141585439443588, 'train/grad_norm_pre_clip': 19.447406768798828, 'timing/data': 0.010802196804434061, 'timing/model': 1.555876323953271, 'learning_rate/qwen_vl_interface': 3.3775319367724746e-06, 'learning_rate/action_model': 3.3775319367724745e-05, 'epoch': 1.29}) +[2026-06-16 06:33:29,961][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.07196646183729172, 'train/grad_norm_pre_clip': 35.36235809326172, 'timing/data': 0.011465228628367186, 'timing/model': 1.5356465349905193, 'learning_rate/qwen_vl_interface': 3.374733509312795e-06, 'learning_rate/action_model': 3.374733509312795e-05, 'epoch': 1.29}) +[2026-06-16 06:33:36,258][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.0014559783739969134, 'train/grad_norm_pre_clip': 0.31633782386779785, 'timing/data': 0.011097298003733158, 'timing/model': 1.5158856571651995, 'learning_rate/qwen_vl_interface': 3.3719358527043345e-06, 'learning_rate/action_model': 3.371935852704334e-05, 'epoch': 1.29}) +[2026-06-16 06:33:42,570][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.14069703221321106, 'train/grad_norm_pre_clip': 6.828118801116943, 'timing/data': 0.011450369376689196, 'timing/model': 1.7692182096652687, 'learning_rate/qwen_vl_interface': 3.3691389680971047e-06, 'learning_rate/action_model': 3.3691389680971044e-05, 'epoch': 1.29}) +[2026-06-16 06:33:48,657][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.0026617010589689016, 'train/grad_norm_pre_clip': 0.9900217056274414, 'timing/data': 0.009706925135105848, 'timing/model': 1.5665997080504894, 'learning_rate/qwen_vl_interface': 3.3663428566407985e-06, 'learning_rate/action_model': 3.366342856640798e-05, 'epoch': 1.29}) +[2026-06-16 06:33:54,965][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.0036271223798394203, 'train/grad_norm_pre_clip': 0.43503496050834656, 'timing/data': 0.009948376100510359, 'timing/model': 1.5620422959327698, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 1.29}) +[2026-06-16 06:34:01,258][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.11000216007232666, 'train/grad_norm_pre_clip': 3.6210389137268066, 'timing/data': 0.010461167898029089, 'timing/model': 1.5560583001933992, 'learning_rate/qwen_vl_interface': 3.3607529577781433e-06, 'learning_rate/action_model': 3.360752957778143e-05, 'epoch': 1.29}) +[2026-06-16 06:34:07,555][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.0041282279416918755, 'train/grad_norm_pre_clip': 0.47774118185043335, 'timing/data': 0.010557537898421288, 'timing/model': 1.5244593257084489, 'learning_rate/qwen_vl_interface': 3.3579591726695906e-06, 'learning_rate/action_model': 3.35795917266959e-05, 'epoch': 1.29}) +[2026-06-16 06:34:14,074][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.17917758226394653, 'train/grad_norm_pre_clip': 27.65283966064453, 'timing/data': 0.01087794965133071, 'timing/model': 1.7889875089749694, 'learning_rate/qwen_vl_interface': 3.3551661653075522e-06, 'learning_rate/action_model': 3.355166165307552e-05, 'epoch': 1.29}) +[2026-06-16 06:34:20,162][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.09319918602705002, 'train/grad_norm_pre_clip': 5.596694469451904, 'timing/data': 0.011900961864739656, 'timing/model': 1.565682103857398, 'learning_rate/qwen_vl_interface': 3.35237393684013e-06, 'learning_rate/action_model': 3.35237393684013e-05, 'epoch': 1.29}) +[2026-06-16 06:34:26,651][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.0031357293482869864, 'train/grad_norm_pre_clip': 1.418047547340393, 'timing/data': 0.010477839037775993, 'timing/model': 1.5464015598408878, 'learning_rate/qwen_vl_interface': 3.349582488415103e-06, 'learning_rate/action_model': 3.3495824884151026e-05, 'epoch': 1.29}) +[2026-06-16 06:34:32,961][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.07204356789588928, 'train/grad_norm_pre_clip': 21.27267074584961, 'timing/data': 0.015152342151850462, 'timing/model': 1.5436338703148067, 'learning_rate/qwen_vl_interface': 3.3467918211799294e-06, 'learning_rate/action_model': 3.3467918211799295e-05, 'epoch': 1.29}) +[2026-06-16 06:34:39,366][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.10350960493087769, 'train/grad_norm_pre_clip': 20.79042625427246, 'timing/data': 0.01071021007373929, 'timing/model': 1.5635795542038977, 'learning_rate/qwen_vl_interface': 3.3440019362817476e-06, 'learning_rate/action_model': 3.344001936281747e-05, 'epoch': 1.29}) +[2026-06-16 06:34:45,667][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.00311639835126698, 'train/grad_norm_pre_clip': 0.7987683415412903, 'timing/data': 0.01099053118377924, 'timing/model': 1.5825437540188432, 'learning_rate/qwen_vl_interface': 3.3412128348673745e-06, 'learning_rate/action_model': 3.341212834867374e-05, 'epoch': 1.29}) +[2026-06-16 06:34:52,000][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.23719258606433868, 'train/grad_norm_pre_clip': 30.504207611083984, 'timing/data': 0.010987811256200075, 'timing/model': 1.7902450421825051, 'learning_rate/qwen_vl_interface': 3.338424518083303e-06, 'learning_rate/action_model': 3.338424518083303e-05, 'epoch': 1.29}) +[2026-06-16 06:34:58,148][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.1926363855600357, 'train/grad_norm_pre_clip': 29.841995239257812, 'timing/data': 0.01105028111487627, 'timing/model': 1.6475798431783915, 'learning_rate/qwen_vl_interface': 3.3356369870757056e-06, 'learning_rate/action_model': 3.335636987075705e-05, 'epoch': 1.29}) +[2026-06-16 06:35:04,461][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.08309830725193024, 'train/grad_norm_pre_clip': 15.165717124938965, 'timing/data': 0.010191909968852997, 'timing/model': 1.5515168416313827, 'learning_rate/qwen_vl_interface': 3.3328502429904327e-06, 'learning_rate/action_model': 3.3328502429904325e-05, 'epoch': 1.29}) +[2026-06-16 06:35:10,761][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.14445774257183075, 'train/grad_norm_pre_clip': 14.992884635925293, 'timing/data': 0.01096977200359106, 'timing/model': 1.5657115778885782, 'learning_rate/qwen_vl_interface': 3.3300642869730066e-06, 'learning_rate/action_model': 3.330064286973006e-05, 'epoch': 1.29}) +[2026-06-16 06:45:58,433][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.09420883655548096, 'train/grad_norm_pre_clip': 11.244449615478516, 'eval/loss': 0.03512299805879593, 'eval/flap_precision': 0.9595141700404858, 'eval/flap_recall': 0.9699863574351978, 'eval/flap_f1': 0.9647218453188602, 'eval/flap_rate_model': 0.20583333333333334, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9647218453188601, 'timing/data': 0.012394136283546686, 'timing/model': 1.5777956591919065, 'learning_rate/qwen_vl_interface': 3.3272791201686294e-06, 'learning_rate/action_model': 3.327279120168629e-05, 'epoch': 1.29}) +[2026-06-16 06:45:58,434][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3200_state +[2026-06-16 06:45:58,440][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 06:46:07,173][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3200_state/model.safetensors +[2026-06-16 06:46:27,839][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3200_state/optimizer.bin +[2026-06-16 06:46:27,840][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3200_state/sampler.bin +[2026-06-16 06:46:27,840][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3200_state/sampler_1.bin +[2026-06-16 06:46:27,842][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3200_state/random_states_0.pkl +[2026-06-16 06:46:28,033][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3200 +[2026-06-16 06:46:28,033][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 06:46:28,049][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 06:46:34,328][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.07463297247886658, 'train/grad_norm_pre_clip': 6.868381977081299, 'timing/data': 0.01092869695276022, 'timing/model': 1.602625923231244, 'learning_rate/qwen_vl_interface': 3.324494743722179e-06, 'learning_rate/action_model': 3.3244947437221785e-05, 'epoch': 1.29}) +[2026-06-16 06:46:40,523][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.3130037486553192, 'train/grad_norm_pre_clip': 15.990331649780273, 'timing/data': 0.011500809341669083, 'timing/model': 1.4936225791461766, 'learning_rate/qwen_vl_interface': 3.321711158778205e-06, 'learning_rate/action_model': 3.321711158778205e-05, 'epoch': 1.29}) +[2026-06-16 06:46:46,613][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.04222099855542183, 'train/grad_norm_pre_clip': 17.460466384887695, 'timing/data': 0.010075433179736137, 'timing/model': 1.5530397100374103, 'learning_rate/qwen_vl_interface': 3.3189283664809393e-06, 'learning_rate/action_model': 3.318928366480939e-05, 'epoch': 1.3}) +[2026-06-16 06:46:52,849][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.2028397172689438, 'train/grad_norm_pre_clip': 19.213115692138672, 'timing/data': 0.009968721773475409, 'timing/model': 1.7894353778101504, 'learning_rate/qwen_vl_interface': 3.316146367974276e-06, 'learning_rate/action_model': 3.316146367974276e-05, 'epoch': 1.3}) +[2026-06-16 06:46:58,835][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.2699708044528961, 'train/grad_norm_pre_clip': 19.914228439331055, 'timing/data': 0.012462603859603405, 'timing/model': 1.5599687271751463, 'learning_rate/qwen_vl_interface': 3.3133651644017945e-06, 'learning_rate/action_model': 3.3133651644017944e-05, 'epoch': 1.3}) +[2026-06-16 06:47:05,006][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.0036262525245547295, 'train/grad_norm_pre_clip': 0.5765687227249146, 'timing/data': 0.01058965502306819, 'timing/model': 1.55549741582945, 'learning_rate/qwen_vl_interface': 3.31058475690674e-06, 'learning_rate/action_model': 3.31058475690674e-05, 'epoch': 1.3}) +[2026-06-16 06:47:11,249][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.00604205671697855, 'train/grad_norm_pre_clip': 1.8413945436477661, 'timing/data': 0.01143893925473094, 'timing/model': 1.569847445935011, 'learning_rate/qwen_vl_interface': 3.3078051466320337e-06, 'learning_rate/action_model': 3.307805146632034e-05, 'epoch': 1.3}) +[2026-06-16 06:47:17,635][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.005436730105429888, 'train/grad_norm_pre_clip': 0.5706666111946106, 'timing/data': 0.011546219699084759, 'timing/model': 1.5602730442769825, 'learning_rate/qwen_vl_interface': 3.3050263347202712e-06, 'learning_rate/action_model': 3.305026334720271e-05, 'epoch': 1.3}) +[2026-06-16 06:47:23,843][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.15345215797424316, 'train/grad_norm_pre_clip': 26.857271194458008, 'timing/data': 0.011324258986860514, 'timing/model': 1.5510894330218434, 'learning_rate/qwen_vl_interface': 3.302248322313712e-06, 'learning_rate/action_model': 3.302248322313712e-05, 'epoch': 1.3}) +[2026-06-16 06:47:30,173][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.0014855070039629936, 'train/grad_norm_pre_clip': 0.24450093507766724, 'timing/data': 0.010632646270096302, 'timing/model': 1.7989784269593656, 'learning_rate/qwen_vl_interface': 3.299471110554295e-06, 'learning_rate/action_model': 3.299471110554295e-05, 'epoch': 1.3}) +[2026-06-16 06:47:36,235][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.07583969831466675, 'train/grad_norm_pre_clip': 25.38350486755371, 'timing/data': 0.010319485329091549, 'timing/model': 1.5507275867275894, 'learning_rate/qwen_vl_interface': 3.296694700583626e-06, 'learning_rate/action_model': 3.296694700583626e-05, 'epoch': 1.3}) +[2026-06-16 06:47:42,548][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.08180348575115204, 'train/grad_norm_pre_clip': 11.436564445495605, 'timing/data': 0.011194557882845402, 'timing/model': 1.5794273288920522, 'learning_rate/qwen_vl_interface': 3.2939190935429833e-06, 'learning_rate/action_model': 3.293919093542983e-05, 'epoch': 1.3}) +[2026-06-16 06:47:48,896][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.008090888150036335, 'train/grad_norm_pre_clip': 3.909621477127075, 'timing/data': 0.010498204734176397, 'timing/model': 1.5896957209333777, 'learning_rate/qwen_vl_interface': 3.2911442905733126e-06, 'learning_rate/action_model': 3.291144290573312e-05, 'epoch': 1.3}) +[2026-06-16 06:47:55,252][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.14524099230766296, 'train/grad_norm_pre_clip': 26.3782901763916, 'timing/data': 0.011108587961643934, 'timing/model': 1.5326876528561115, 'learning_rate/qwen_vl_interface': 3.2883702928152323e-06, 'learning_rate/action_model': 3.288370292815232e-05, 'epoch': 1.3}) +[2026-06-16 06:48:01,558][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.07409428060054779, 'train/grad_norm_pre_clip': 12.46007251739502, 'timing/data': 0.00996413268148899, 'timing/model': 1.5673829610459507, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 1.3}) +[2026-06-16 06:48:08,092][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.00332772103138268, 'train/grad_norm_pre_clip': 0.49170947074890137, 'timing/data': 0.011860281694680452, 'timing/model': 1.9073476302437484, 'learning_rate/qwen_vl_interface': 3.2828247174946528e-06, 'learning_rate/action_model': 3.2828247174946524e-05, 'epoch': 1.3}) +[2026-06-16 06:48:14,143][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.013279755599796772, 'train/grad_norm_pre_clip': 3.524829149246216, 'timing/data': 0.010390505194664001, 'timing/model': 1.5597682129591703, 'learning_rate/qwen_vl_interface': 3.280053142211729e-06, 'learning_rate/action_model': 3.280053142211729e-05, 'epoch': 1.3}) +[2026-06-16 06:48:20,520][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.002014649799093604, 'train/grad_norm_pre_clip': 0.2590799629688263, 'timing/data': 0.011011587921530008, 'timing/model': 1.5659957979805768, 'learning_rate/qwen_vl_interface': 3.2772823766995475e-06, 'learning_rate/action_model': 3.277282376699548e-05, 'epoch': 1.3}) +[2026-06-16 06:48:26,746][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.0538632906973362, 'train/grad_norm_pre_clip': 14.64294147491455, 'timing/data': 0.010486455634236336, 'timing/model': 1.563289362937212, 'learning_rate/qwen_vl_interface': 3.2745124220970646e-06, 'learning_rate/action_model': 3.274512422097065e-05, 'epoch': 1.3}) +[2026-06-16 06:48:33,055][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.06820535659790039, 'train/grad_norm_pre_clip': 11.955062866210938, 'timing/data': 0.01041347486898303, 'timing/model': 1.5676080491393805, 'learning_rate/qwen_vl_interface': 3.2717432795429057e-06, 'learning_rate/action_model': 3.271743279542905e-05, 'epoch': 1.3}) +[2026-06-16 06:48:39,645][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.0016154339537024498, 'train/grad_norm_pre_clip': 0.30985260009765625, 'timing/data': 0.011478550266474485, 'timing/model': 1.5065130898728967, 'learning_rate/qwen_vl_interface': 3.2689749501753578e-06, 'learning_rate/action_model': 3.2689749501753574e-05, 'epoch': 1.3}) +[2026-06-16 06:48:45,983][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.0031344005838036537, 'train/grad_norm_pre_clip': 0.5899409651756287, 'timing/data': 0.0102876047603786, 'timing/model': 1.803522352129221, 'learning_rate/qwen_vl_interface': 3.2662074351323774e-06, 'learning_rate/action_model': 3.266207435132378e-05, 'epoch': 1.3}) +[2026-06-16 06:48:52,053][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.03937476873397827, 'train/grad_norm_pre_clip': 15.447325706481934, 'timing/data': 0.012187683023512363, 'timing/model': 1.5562313008122146, 'learning_rate/qwen_vl_interface': 3.2634407355515875e-06, 'learning_rate/action_model': 3.2634407355515875e-05, 'epoch': 1.3}) +[2026-06-16 06:48:58,355][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.17825894057750702, 'train/grad_norm_pre_clip': 21.125228881835938, 'timing/data': 0.011133138090372086, 'timing/model': 1.5485789952799678, 'learning_rate/qwen_vl_interface': 3.2606748525702715e-06, 'learning_rate/action_model': 3.260674852570271e-05, 'epoch': 1.3}) +[2026-06-16 06:49:04,739][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.028552940115332603, 'train/grad_norm_pre_clip': 17.050689697265625, 'timing/data': 0.011225359048694372, 'timing/model': 1.638408973813057, 'learning_rate/qwen_vl_interface': 3.2579097873253803e-06, 'learning_rate/action_model': 3.25790978732538e-05, 'epoch': 1.3}) +[2026-06-16 06:49:11,082][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.0009482341702096164, 'train/grad_norm_pre_clip': 0.11655501276254654, 'timing/data': 0.011377288959920406, 'timing/model': 1.607505855616182, 'learning_rate/qwen_vl_interface': 3.2551455409535274e-06, 'learning_rate/action_model': 3.255145540953528e-05, 'epoch': 1.3}) +[2026-06-16 06:49:17,898][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.0015313950134441257, 'train/grad_norm_pre_clip': 0.23970478773117065, 'timing/data': 0.016312011051923037, 'timing/model': 1.556445799767971, 'learning_rate/qwen_vl_interface': 3.2523821145909907e-06, 'learning_rate/action_model': 3.252382114590991e-05, 'epoch': 1.3}) +[2026-06-16 06:49:24,230][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.001952023128978908, 'train/grad_norm_pre_clip': 0.46472182869911194, 'timing/data': 0.013890590984374285, 'timing/model': 1.8191231479868293, 'learning_rate/qwen_vl_interface': 3.249619509373712e-06, 'learning_rate/action_model': 3.249619509373712e-05, 'epoch': 1.31}) +[2026-06-16 06:49:30,247][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.03667544201016426, 'train/grad_norm_pre_clip': 13.520204544067383, 'timing/data': 0.010537966154515743, 'timing/model': 1.5487379292026162, 'learning_rate/qwen_vl_interface': 3.2468577264372905e-06, 'learning_rate/action_model': 3.24685772643729e-05, 'epoch': 1.31}) +[2026-06-16 06:49:36,553][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.0024526817724108696, 'train/grad_norm_pre_clip': 0.8101608157157898, 'timing/data': 0.010880507994443178, 'timing/model': 1.5638950718566775, 'learning_rate/qwen_vl_interface': 3.244096766916992e-06, 'learning_rate/action_model': 3.244096766916992e-05, 'epoch': 1.31}) +[2026-06-16 06:49:42,936][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.0003302693658042699, 'train/grad_norm_pre_clip': 0.08015758544206619, 'timing/data': 0.011331689078360796, 'timing/model': 1.581799556966871, 'learning_rate/qwen_vl_interface': 3.241336631947743e-06, 'learning_rate/action_model': 3.2413366319477427e-05, 'epoch': 1.31}) +[2026-06-16 06:49:49,229][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.009478690102696419, 'train/grad_norm_pre_clip': 3.3167717456817627, 'timing/data': 0.01039807591587305, 'timing/model': 1.6349366968497634, 'learning_rate/qwen_vl_interface': 3.238577322664131e-06, 'learning_rate/action_model': 3.238577322664131e-05, 'epoch': 1.31}) +[2026-06-16 06:49:55,451][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.12993943691253662, 'train/grad_norm_pre_clip': 39.822898864746094, 'timing/data': 0.010602606926113367, 'timing/model': 1.5186886810697615, 'learning_rate/qwen_vl_interface': 3.2358188402004026e-06, 'learning_rate/action_model': 3.2358188402004024e-05, 'epoch': 1.31}) +[2026-06-16 06:50:01,805][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.0014785646926611662, 'train/grad_norm_pre_clip': 0.27311423420906067, 'timing/data': 0.011376109905540943, 'timing/model': 1.804370907600969, 'learning_rate/qwen_vl_interface': 3.2330611856904675e-06, 'learning_rate/action_model': 3.233061185690468e-05, 'epoch': 1.31}) +[2026-06-16 06:50:07,855][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.00410179328173399, 'train/grad_norm_pre_clip': 0.7487626075744629, 'timing/data': 0.011299190111458302, 'timing/model': 1.5624213460832834, 'learning_rate/qwen_vl_interface': 3.2303043602678928e-06, 'learning_rate/action_model': 3.230304360267893e-05, 'epoch': 1.31}) +[2026-06-16 06:50:14,172][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.0006617878680117428, 'train/grad_norm_pre_clip': 0.10678685456514359, 'timing/data': 0.011531780939549208, 'timing/model': 1.5696101039648056, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 1.31}) +[2026-06-16 06:50:20,556][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.16929256916046143, 'train/grad_norm_pre_clip': 8.214818954467773, 'timing/data': 0.01231647515669465, 'timing/model': 1.5733487661927938, 'learning_rate/qwen_vl_interface': 3.2247932012173887e-06, 'learning_rate/action_model': 3.2247932012173884e-05, 'epoch': 1.31}) +[2026-06-16 06:50:26,868][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.0004209409817121923, 'train/grad_norm_pre_clip': 0.12038692831993103, 'timing/data': 0.012145673856139183, 'timing/model': 1.570452849380672, 'learning_rate/qwen_vl_interface': 3.2220388698548897e-06, 'learning_rate/action_model': 3.22203886985489e-05, 'epoch': 1.31}) +[2026-06-16 06:50:33,268][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.2709619104862213, 'train/grad_norm_pre_clip': 16.3438663482666, 'timing/data': 0.011443651281297207, 'timing/model': 1.5374583592638373, 'learning_rate/qwen_vl_interface': 3.21928537211061e-06, 'learning_rate/action_model': 3.21928537211061e-05, 'epoch': 1.31}) +[2026-06-16 06:50:39,646][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.24683760106563568, 'train/grad_norm_pre_clip': 19.160829544067383, 'timing/data': 0.011478250846266747, 'timing/model': 1.8303491631522775, 'learning_rate/qwen_vl_interface': 3.216532709116405e-06, 'learning_rate/action_model': 3.216532709116405e-05, 'epoch': 1.31}) +[2026-06-16 06:50:45,836][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.00028498732717707753, 'train/grad_norm_pre_clip': 0.05665424466133118, 'timing/data': 0.011733882129192352, 'timing/model': 1.7420683079399168, 'learning_rate/qwen_vl_interface': 3.2137808820037925e-06, 'learning_rate/action_model': 3.2137808820037924e-05, 'epoch': 1.31}) +[2026-06-16 06:50:51,884][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.0032243577297776937, 'train/grad_norm_pre_clip': 0.7344347238540649, 'timing/data': 0.010962508618831635, 'timing/model': 1.5873478772118688, 'learning_rate/qwen_vl_interface': 3.211029891903945e-06, 'learning_rate/action_model': 3.211029891903945e-05, 'epoch': 1.31}) +[2026-06-16 06:50:58,276][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.0006166913080960512, 'train/grad_norm_pre_clip': 0.13696667551994324, 'timing/data': 0.011142129078507423, 'timing/model': 1.5728150489740074, 'learning_rate/qwen_vl_interface': 3.2082797399476894e-06, 'learning_rate/action_model': 3.2082797399476895e-05, 'epoch': 1.31}) +[2026-06-16 06:51:04,662][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.01911555416882038, 'train/grad_norm_pre_clip': 10.433440208435059, 'timing/data': 0.010902578011155128, 'timing/model': 1.5696679698303342, 'learning_rate/qwen_vl_interface': 3.2055304272655114e-06, 'learning_rate/action_model': 3.205530427265511e-05, 'epoch': 1.31}) +[2026-06-16 06:51:10,976][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.0002338991325814277, 'train/grad_norm_pre_clip': 0.040717411786317825, 'timing/data': 0.011089660227298737, 'timing/model': 1.5464552319608629, 'learning_rate/qwen_vl_interface': 3.2027819549875477e-06, 'learning_rate/action_model': 3.202781954987548e-05, 'epoch': 1.31}) +[2026-06-16 06:51:17,306][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.07992223650217056, 'train/grad_norm_pre_clip': 13.895716667175293, 'timing/data': 0.011017078999429941, 'timing/model': 1.7894746786914766, 'learning_rate/qwen_vl_interface': 3.2000343242435926e-06, 'learning_rate/action_model': 3.2000343242435924e-05, 'epoch': 1.31}) +[2026-06-16 06:51:23,601][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.051839932799339294, 'train/grad_norm_pre_clip': 23.731523513793945, 'timing/data': 0.01089990884065628, 'timing/model': 1.7946499781683087, 'learning_rate/qwen_vl_interface': 3.1972875361630923e-06, 'learning_rate/action_model': 3.1972875361630923e-05, 'epoch': 1.31}) +[2026-06-16 06:51:29,663][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.005556959193199873, 'train/grad_norm_pre_clip': 2.6827197074890137, 'timing/data': 0.01139722066000104, 'timing/model': 1.5563530013896525, 'learning_rate/qwen_vl_interface': 3.1945415918751484e-06, 'learning_rate/action_model': 3.194541591875148e-05, 'epoch': 1.31}) +[2026-06-16 06:51:35,968][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.04611273482441902, 'train/grad_norm_pre_clip': 21.247556686401367, 'timing/data': 0.011407101061195135, 'timing/model': 1.5706234397366643, 'learning_rate/qwen_vl_interface': 3.191796492508516e-06, 'learning_rate/action_model': 3.191796492508516e-05, 'epoch': 1.31}) +[2026-06-16 06:51:42,456][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.07011677324771881, 'train/grad_norm_pre_clip': 16.123424530029297, 'timing/data': 0.011763561982661486, 'timing/model': 1.6620632843114436, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 1.31}) +[2026-06-16 06:51:48,754][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.0004379318852443248, 'train/grad_norm_pre_clip': 0.06131259724497795, 'timing/data': 0.011239039711654186, 'timing/model': 1.5138466572389007, 'learning_rate/qwen_vl_interface': 3.1863088330524577e-06, 'learning_rate/action_model': 3.186308833052458e-05, 'epoch': 1.31}) +[2026-06-16 06:51:55,092][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.10659240931272507, 'train/grad_norm_pre_clip': 10.631449699401855, 'timing/data': 0.011260840110480785, 'timing/model': 1.7898938558064401, 'learning_rate/qwen_vl_interface': 3.183566275218803e-06, 'learning_rate/action_model': 3.1835662752188026e-05, 'epoch': 1.32}) +[2026-06-16 06:52:01,166][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.11100996285676956, 'train/grad_norm_pre_clip': 22.885278701782227, 'timing/data': 0.011303901206701994, 'timing/model': 1.5610875780694187, 'learning_rate/qwen_vl_interface': 3.180824566817996e-06, 'learning_rate/action_model': 3.1808245668179956e-05, 'epoch': 1.32}) +[2026-06-16 06:52:07,455][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.13532431423664093, 'train/grad_norm_pre_clip': 30.035795211791992, 'timing/data': 0.011008559260517359, 'timing/model': 1.5532485372386873, 'learning_rate/qwen_vl_interface': 3.1780837089770524e-06, 'learning_rate/action_model': 3.1780837089770524e-05, 'epoch': 1.32}) +[2026-06-16 06:52:13,759][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.047143153846263885, 'train/grad_norm_pre_clip': 23.75872039794922, 'timing/data': 0.01140044117346406, 'timing/model': 1.5596392997540534, 'learning_rate/qwen_vl_interface': 3.17534370282263e-06, 'learning_rate/action_model': 3.1753437028226303e-05, 'epoch': 1.32}) +[2026-06-16 06:52:20,059][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.031739890575408936, 'train/grad_norm_pre_clip': 22.932863235473633, 'timing/data': 0.011694023385643959, 'timing/model': 1.5194585649296641, 'learning_rate/qwen_vl_interface': 3.172604549481046e-06, 'learning_rate/action_model': 3.1726045494810455e-05, 'epoch': 1.32}) +[2026-06-16 06:52:26,364][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.003244045889005065, 'train/grad_norm_pre_clip': 0.9000970125198364, 'timing/data': 0.01069797808304429, 'timing/model': 1.5601448323577642, 'learning_rate/qwen_vl_interface': 3.169866250078261e-06, 'learning_rate/action_model': 3.169866250078261e-05, 'epoch': 1.32}) +[2026-06-16 06:52:32,655][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.004877408966422081, 'train/grad_norm_pre_clip': 1.8835705518722534, 'timing/data': 0.011503021698445082, 'timing/model': 1.7609738879837096, 'learning_rate/qwen_vl_interface': 3.167128805739887e-06, 'learning_rate/action_model': 3.1671288057398866e-05, 'epoch': 1.32}) +[2026-06-16 06:52:38,762][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.05934043601155281, 'train/grad_norm_pre_clip': 12.499748229980469, 'timing/data': 0.01113324984908104, 'timing/model': 1.5632741847075522, 'learning_rate/qwen_vl_interface': 3.1643922175911847e-06, 'learning_rate/action_model': 3.164392217591185e-05, 'epoch': 1.32}) +[2026-06-16 06:52:45,056][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.21732953190803528, 'train/grad_norm_pre_clip': 28.146800994873047, 'timing/data': 0.011169360019266605, 'timing/model': 1.5585207678377628, 'learning_rate/qwen_vl_interface': 3.1616564867570603e-06, 'learning_rate/action_model': 3.16165648675706e-05, 'epoch': 1.32}) +[2026-06-16 06:52:51,359][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.020223841071128845, 'train/grad_norm_pre_clip': 10.237078666687012, 'timing/data': 0.010400217026472092, 'timing/model': 1.5662536770105362, 'learning_rate/qwen_vl_interface': 3.1589216143620733e-06, 'learning_rate/action_model': 3.1589216143620734e-05, 'epoch': 1.32}) +[2026-06-16 06:52:57,661][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.0016314955428242683, 'train/grad_norm_pre_clip': 0.3317737281322479, 'timing/data': 0.011143411044031382, 'timing/model': 1.5152120711281896, 'learning_rate/qwen_vl_interface': 3.1561876015304215e-06, 'learning_rate/action_model': 3.156187601530421e-05, 'epoch': 1.32}) +[2026-06-16 06:53:03,958][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.05011821910738945, 'train/grad_norm_pre_clip': 40.76371765136719, 'timing/data': 0.011766253039240837, 'timing/model': 1.5216311449185014, 'learning_rate/qwen_vl_interface': 3.1534544493859563e-06, 'learning_rate/action_model': 3.1534544493859566e-05, 'epoch': 1.32}) +[2026-06-16 06:53:10,250][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.002228174125775695, 'train/grad_norm_pre_clip': 0.6678060293197632, 'timing/data': 0.013580930884927511, 'timing/model': 1.75120271416381, 'learning_rate/qwen_vl_interface': 3.150722159052174e-06, 'learning_rate/action_model': 3.150722159052174e-05, 'epoch': 1.32}) +[2026-06-16 06:53:16,361][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.0035920448135584593, 'train/grad_norm_pre_clip': 1.5433346033096313, 'timing/data': 0.011423701886087656, 'timing/model': 1.561537662986666, 'learning_rate/qwen_vl_interface': 3.147990731652215e-06, 'learning_rate/action_model': 3.147990731652215e-05, 'epoch': 1.32}) +[2026-06-16 06:53:22,663][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.012673800811171532, 'train/grad_norm_pre_clip': 8.905519485473633, 'timing/data': 0.010929518844932318, 'timing/model': 1.5578085109591484, 'learning_rate/qwen_vl_interface': 3.1452601683088673e-06, 'learning_rate/action_model': 3.145260168308867e-05, 'epoch': 1.32}) +[2026-06-16 06:53:29,074][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.142433300614357, 'train/grad_norm_pre_clip': 17.51781463623047, 'timing/data': 0.014568735379725695, 'timing/model': 1.5817215330898762, 'learning_rate/qwen_vl_interface': 3.1425304701445608e-06, 'learning_rate/action_model': 3.1425304701445605e-05, 'epoch': 1.32}) +[2026-06-16 06:53:35,361][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.003825838677585125, 'train/grad_norm_pre_clip': 0.6802628040313721, 'timing/data': 0.011775013990700245, 'timing/model': 1.5189238130114973, 'learning_rate/qwen_vl_interface': 3.139801638281372e-06, 'learning_rate/action_model': 3.139801638281372e-05, 'epoch': 1.32}) +[2026-06-16 06:53:41,695][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.010239475406706333, 'train/grad_norm_pre_clip': 11.63575553894043, 'timing/data': 0.010365237016230822, 'timing/model': 1.7843508059158921, 'learning_rate/qwen_vl_interface': 3.13707367384102e-06, 'learning_rate/action_model': 3.13707367384102e-05, 'epoch': 1.32}) +[2026-06-16 06:53:47,954][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.19087417423725128, 'train/grad_norm_pre_clip': 15.738901138305664, 'timing/data': 0.010645827744156122, 'timing/model': 1.633595089893788, 'learning_rate/qwen_vl_interface': 3.134346577944869e-06, 'learning_rate/action_model': 3.134346577944869e-05, 'epoch': 1.32}) +[2026-06-16 06:53:54,280][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.047864656895399094, 'train/grad_norm_pre_clip': 20.270809173583984, 'timing/data': 0.01170720299705863, 'timing/model': 1.5784993930719793, 'learning_rate/qwen_vl_interface': 3.1316203517139243e-06, 'learning_rate/action_model': 3.131620351713924e-05, 'epoch': 1.32}) +[2026-06-16 06:54:00,461][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.001529018976725638, 'train/grad_norm_pre_clip': 0.8596380949020386, 'timing/data': 0.011224451009184122, 'timing/model': 1.5563732869923115, 'learning_rate/qwen_vl_interface': 3.128894996268837e-06, 'learning_rate/action_model': 3.128894996268837e-05, 'epoch': 1.32}) +[2026-06-16 06:54:06,770][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.13508650660514832, 'train/grad_norm_pre_clip': 13.481392860412598, 'timing/data': 0.011448310688138008, 'timing/model': 1.5887602791190147, 'learning_rate/qwen_vl_interface': 3.1261705127298934e-06, 'learning_rate/action_model': 3.1261705127298935e-05, 'epoch': 1.32}) +[2026-06-16 06:54:13,074][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.028256405144929886, 'train/grad_norm_pre_clip': 38.11775207519531, 'timing/data': 0.010901928879320621, 'timing/model': 1.5669703027233481, 'learning_rate/qwen_vl_interface': 3.123446902217028e-06, 'learning_rate/action_model': 3.1234469022170283e-05, 'epoch': 1.32}) +[2026-06-16 06:54:19,407][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.17841193079948425, 'train/grad_norm_pre_clip': 15.548155784606934, 'timing/data': 0.010887088719755411, 'timing/model': 1.8040923117659986, 'learning_rate/qwen_vl_interface': 3.1207241658498145e-06, 'learning_rate/action_model': 3.1207241658498145e-05, 'epoch': 1.32}) +[2026-06-16 06:54:25,459][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.18252119421958923, 'train/grad_norm_pre_clip': 19.169044494628906, 'timing/data': 0.010668127797544003, 'timing/model': 1.5563952433876693, 'learning_rate/qwen_vl_interface': 3.1180023047474667e-06, 'learning_rate/action_model': 3.1180023047474664e-05, 'epoch': 1.32}) +[2026-06-16 06:54:31,870][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.018098175525665283, 'train/grad_norm_pre_clip': 14.076583862304688, 'timing/data': 0.013081599026918411, 'timing/model': 1.579788991715759, 'learning_rate/qwen_vl_interface': 3.1152813200288386e-06, 'learning_rate/action_model': 3.115281320028838e-05, 'epoch': 1.33}) +[2026-06-16 06:54:38,258][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.0014485383871942759, 'train/grad_norm_pre_clip': 0.5752392411231995, 'timing/data': 0.010606089141219854, 'timing/model': 1.5655119610019028, 'learning_rate/qwen_vl_interface': 3.1125612128124244e-06, 'learning_rate/action_model': 3.112561212812424e-05, 'epoch': 1.33}) +[2026-06-16 06:54:44,723][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.001031368738040328, 'train/grad_norm_pre_clip': 0.4020216763019562, 'timing/data': 0.010377236641943455, 'timing/model': 1.6320177768357098, 'learning_rate/qwen_vl_interface': 3.1098419842163573e-06, 'learning_rate/action_model': 3.109841984216357e-05, 'epoch': 1.33}) +[2026-06-16 06:54:51,053][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.0011569615453481674, 'train/grad_norm_pre_clip': 0.23717834055423737, 'timing/data': 0.011559573002159595, 'timing/model': 1.5517799258232117, 'learning_rate/qwen_vl_interface': 3.10712363535841e-06, 'learning_rate/action_model': 3.10712363535841e-05, 'epoch': 1.33}) +[2026-06-16 06:54:57,505][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.0009760287939570844, 'train/grad_norm_pre_clip': 0.3819666802883148, 'timing/data': 0.011077039875090122, 'timing/model': 1.9038401213474572, 'learning_rate/qwen_vl_interface': 3.1044061673559926e-06, 'learning_rate/action_model': 3.1044061673559925e-05, 'epoch': 1.33}) +[2026-06-16 06:55:03,557][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.0015057969139888883, 'train/grad_norm_pre_clip': 0.8615086078643799, 'timing/data': 0.011088720057159662, 'timing/model': 1.5549199869856238, 'learning_rate/qwen_vl_interface': 3.1016895813261556e-06, 'learning_rate/action_model': 3.101689581326155e-05, 'epoch': 1.33}) +[2026-06-16 06:55:09,859][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.07331014424562454, 'train/grad_norm_pre_clip': 16.800222396850586, 'timing/data': 0.010724439285695553, 'timing/model': 1.5567111978307366, 'learning_rate/qwen_vl_interface': 3.098973878385581e-06, 'learning_rate/action_model': 3.098973878385581e-05, 'epoch': 1.33}) +[2026-06-16 06:55:16,162][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.07903778553009033, 'train/grad_norm_pre_clip': 23.46457862854004, 'timing/data': 0.010826549027115107, 'timing/model': 1.5617856648750603, 'learning_rate/qwen_vl_interface': 3.096259059650594e-06, 'learning_rate/action_model': 3.096259059650594e-05, 'epoch': 1.33}) +[2026-06-16 06:55:22,551][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.10791061818599701, 'train/grad_norm_pre_clip': 16.66802406311035, 'timing/data': 0.012354875914752483, 'timing/model': 1.6023454428650439, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 1.33}) +[2026-06-16 06:55:28,875][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.07850006967782974, 'train/grad_norm_pre_clip': 33.727699279785156, 'timing/data': 0.011453952174633741, 'timing/model': 1.5517622060142457, 'learning_rate/qwen_vl_interface': 3.090832079260855e-06, 'learning_rate/action_model': 3.0908320792608544e-05, 'epoch': 1.33}) +[2026-06-16 06:55:35,281][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.006437900476157665, 'train/grad_norm_pre_clip': 8.082069396972656, 'timing/data': 0.011136560700833797, 'timing/model': 1.888253384269774, 'learning_rate/qwen_vl_interface': 3.088119919836929e-06, 'learning_rate/action_model': 3.088119919836929e-05, 'epoch': 1.33}) +[2026-06-16 06:55:41,353][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.3055618405342102, 'train/grad_norm_pre_clip': 40.563819885253906, 'timing/data': 0.011070921085774899, 'timing/model': 1.5557655398733914, 'learning_rate/qwen_vl_interface': 3.0854086490802433e-06, 'learning_rate/action_model': 3.085408649080243e-05, 'epoch': 1.33}) +[2026-06-16 06:55:47,656][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.1889650970697403, 'train/grad_norm_pre_clip': 13.363899230957031, 'timing/data': 0.011114980094134808, 'timing/model': 1.5537244440056384, 'learning_rate/qwen_vl_interface': 3.082698268105299e-06, 'learning_rate/action_model': 3.082698268105299e-05, 'epoch': 1.33}) +[2026-06-16 06:55:53,962][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.00146795017644763, 'train/grad_norm_pre_clip': 0.40330424904823303, 'timing/data': 0.011005750391632318, 'timing/model': 1.5581179969012737, 'learning_rate/qwen_vl_interface': 3.0799887780262295e-06, 'learning_rate/action_model': 3.0799887780262296e-05, 'epoch': 1.33}) +[2026-06-16 06:56:00,266][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.027237508445978165, 'train/grad_norm_pre_clip': 11.558452606201172, 'timing/data': 0.01065404899418354, 'timing/model': 1.5716738519258797, 'learning_rate/qwen_vl_interface': 3.0772801799568054e-06, 'learning_rate/action_model': 3.0772801799568054e-05, 'epoch': 1.33}) +[2026-06-16 06:56:06,567][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.45163828134536743, 'train/grad_norm_pre_clip': 64.41533660888672, 'timing/data': 0.011538553051650524, 'timing/model': 1.5311436299234629, 'learning_rate/qwen_vl_interface': 3.07457247501043e-06, 'learning_rate/action_model': 3.0745724750104296e-05, 'epoch': 1.33}) +[2026-06-16 06:56:12,986][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.0012460898142307997, 'train/grad_norm_pre_clip': 0.6403185129165649, 'timing/data': 0.01146868197247386, 'timing/model': 1.8873143019154668, 'learning_rate/qwen_vl_interface': 3.0718656643001353e-06, 'learning_rate/action_model': 3.071865664300135e-05, 'epoch': 1.33}) +[2026-06-16 06:56:19,069][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.0002967216423712671, 'train/grad_norm_pre_clip': 0.04990142956376076, 'timing/data': 0.012031625024974346, 'timing/model': 1.5591834532096982, 'learning_rate/qwen_vl_interface': 3.069159748938591e-06, 'learning_rate/action_model': 3.069159748938591e-05, 'epoch': 1.33}) +[2026-06-16 06:56:25,361][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.07360509783029556, 'train/grad_norm_pre_clip': 30.291790008544922, 'timing/data': 0.011500122025609016, 'timing/model': 1.539195675868541, 'learning_rate/qwen_vl_interface': 3.066454730038096e-06, 'learning_rate/action_model': 3.0664547300380956e-05, 'epoch': 1.33}) +[2026-06-16 06:56:31,693][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.0004947537672705948, 'train/grad_norm_pre_clip': 0.16316407918930054, 'timing/data': 0.011530023068189621, 'timing/model': 1.5808520489372313, 'learning_rate/qwen_vl_interface': 3.0637506087105816e-06, 'learning_rate/action_model': 3.0637506087105815e-05, 'epoch': 1.33}) +[2026-06-16 06:56:37,966][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.03050556778907776, 'train/grad_norm_pre_clip': 32.31100845336914, 'timing/data': 0.01122900191694498, 'timing/model': 1.5624137469567358, 'learning_rate/qwen_vl_interface': 3.0610473860676103e-06, 'learning_rate/action_model': 3.06104738606761e-05, 'epoch': 1.33}) +[2026-06-16 06:56:44,261][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.13608525693416595, 'train/grad_norm_pre_clip': 36.26582336425781, 'timing/data': 0.011047190986573696, 'timing/model': 1.5221029352396727, 'learning_rate/qwen_vl_interface': 3.0583450632203748e-06, 'learning_rate/action_model': 3.058345063220375e-05, 'epoch': 1.33}) +[2026-06-16 06:56:50,598][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.01502866018563509, 'train/grad_norm_pre_clip': 12.053085327148438, 'timing/data': 0.010811078827828169, 'timing/model': 1.7966648871079087, 'learning_rate/qwen_vl_interface': 3.055643641279699e-06, 'learning_rate/action_model': 3.055643641279699e-05, 'epoch': 1.33}) +[2026-06-16 07:07:49,412][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.0010374828707426786, 'train/grad_norm_pre_clip': 0.4670018255710602, 'eval/loss': 0.04136509820818901, 'eval/flap_precision': 0.9587217043941412, 'eval/flap_recall': 0.9822646657571623, 'eval/flap_f1': 0.9703504043126684, 'eval/flap_rate_model': 0.2086111111111111, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9703504043126684, 'timing/data': 0.012810988817363977, 'timing/model': 1.557695402763784, 'learning_rate/qwen_vl_interface': 3.0529431213560334e-06, 'learning_rate/action_model': 3.0529431213560334e-05, 'epoch': 1.33}) +[2026-06-16 07:07:49,413][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3300_state +[2026-06-16 07:07:49,419][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 07:07:58,083][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3300_state/model.safetensors +[2026-06-16 07:08:18,030][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3300_state/optimizer.bin +[2026-06-16 07:08:18,031][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3300_state/sampler.bin +[2026-06-16 07:08:18,031][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3300_state/sampler_1.bin +[2026-06-16 07:08:18,033][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3300_state/random_states_0.pkl +[2026-06-16 07:08:18,236][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3300 +[2026-06-16 07:08:18,236][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 07:08:18,251][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 07:08:24,508][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.0020931148901581764, 'train/grad_norm_pre_clip': 0.42429158091545105, 'timing/data': 0.012386411894112825, 'timing/model': 1.8260162570513785, 'learning_rate/qwen_vl_interface': 3.0502435045594596e-06, 'learning_rate/action_model': 3.0502435045594595e-05, 'epoch': 1.33}) +[2026-06-16 07:08:30,531][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.1901329606771469, 'train/grad_norm_pre_clip': 30.608501434326172, 'timing/data': 0.039765825029462576, 'timing/model': 1.5305013391189277, 'learning_rate/qwen_vl_interface': 3.047544791999689e-06, 'learning_rate/action_model': 3.0475447919996892e-05, 'epoch': 1.34}) +[2026-06-16 07:08:36,730][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.048044607043266296, 'train/grad_norm_pre_clip': 22.580644607543945, 'timing/data': 0.031223736703395844, 'timing/model': 1.545336636248976, 'learning_rate/qwen_vl_interface': 3.044846984786061e-06, 'learning_rate/action_model': 3.044846984786061e-05, 'epoch': 1.34}) +[2026-06-16 07:08:42,935][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.29498183727264404, 'train/grad_norm_pre_clip': 31.340036392211914, 'timing/data': 0.030829626135528088, 'timing/model': 1.5378556768409908, 'learning_rate/qwen_vl_interface': 3.0421500840275413e-06, 'learning_rate/action_model': 3.0421500840275413e-05, 'epoch': 1.34}) +[2026-06-16 07:08:49,130][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.07446905970573425, 'train/grad_norm_pre_clip': 22.049423217773438, 'timing/data': 0.030709034763276577, 'timing/model': 1.5394291589036584, 'learning_rate/qwen_vl_interface': 3.039454090832727e-06, 'learning_rate/action_model': 3.0394540908327268e-05, 'epoch': 1.34}) +[2026-06-16 07:08:55,437][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.16921214759349823, 'train/grad_norm_pre_clip': 17.4982852935791, 'timing/data': 0.03156440891325474, 'timing/model': 1.5684618749655783, 'learning_rate/qwen_vl_interface': 3.0367590063098333e-06, 'learning_rate/action_model': 3.0367590063098333e-05, 'epoch': 1.34}) +[2026-06-16 07:09:01,691][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.11546679586172104, 'train/grad_norm_pre_clip': 14.408349990844727, 'timing/data': 0.030876806005835533, 'timing/model': 1.7797312526963651, 'learning_rate/qwen_vl_interface': 3.0340648315667115e-06, 'learning_rate/action_model': 3.0340648315667113e-05, 'epoch': 1.34}) +[2026-06-16 07:09:07,749][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.001468961127102375, 'train/grad_norm_pre_clip': 0.7820031642913818, 'timing/data': 0.03149269986897707, 'timing/model': 1.5398310869932175, 'learning_rate/qwen_vl_interface': 3.0313715677108336e-06, 'learning_rate/action_model': 3.0313715677108335e-05, 'epoch': 1.34}) +[2026-06-16 07:09:14,050][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.2556998133659363, 'train/grad_norm_pre_clip': 21.71998405456543, 'timing/data': 0.03600175632163882, 'timing/model': 1.5418071900494397, 'learning_rate/qwen_vl_interface': 3.0286792158492996e-06, 'learning_rate/action_model': 3.0286792158492994e-05, 'epoch': 1.34}) +[2026-06-16 07:09:20,351][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.028795382007956505, 'train/grad_norm_pre_clip': 20.90493392944336, 'timing/data': 0.03524280292913318, 'timing/model': 1.534305620007217, 'learning_rate/qwen_vl_interface': 3.0259877770888334e-06, 'learning_rate/action_model': 3.0259877770888333e-05, 'epoch': 1.34}) +[2026-06-16 07:09:26,734][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.00022462134074885398, 'train/grad_norm_pre_clip': 0.0278476532548666, 'timing/data': 0.031858121044933796, 'timing/model': 1.5801392360590398, 'learning_rate/qwen_vl_interface': 3.0232972525357836e-06, 'learning_rate/action_model': 3.0232972525357834e-05, 'epoch': 1.34}) +[2026-06-16 07:09:32,992][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.00195445423014462, 'train/grad_norm_pre_clip': 1.5083181858062744, 'timing/data': 0.032045742962509394, 'timing/model': 1.779245019890368, 'learning_rate/qwen_vl_interface': 3.020607643296123e-06, 'learning_rate/action_model': 3.0206076432961226e-05, 'epoch': 1.34}) +[2026-06-16 07:09:39,064][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.10728111863136292, 'train/grad_norm_pre_clip': 22.63517189025879, 'timing/data': 0.030580418184399605, 'timing/model': 1.5551391770131886, 'learning_rate/qwen_vl_interface': 3.017918950475449e-06, 'learning_rate/action_model': 3.017918950475449e-05, 'epoch': 1.34}) +[2026-06-16 07:09:45,352][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.25844255089759827, 'train/grad_norm_pre_clip': 21.419769287109375, 'timing/data': 0.031582691706717014, 'timing/model': 1.5375915998592973, 'learning_rate/qwen_vl_interface': 3.0152311751789815e-06, 'learning_rate/action_model': 3.0152311751789814e-05, 'epoch': 1.34}) +[2026-06-16 07:09:51,553][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.16359283030033112, 'train/grad_norm_pre_clip': 27.850526809692383, 'timing/data': 0.031249929685145617, 'timing/model': 1.5404889062047005, 'learning_rate/qwen_vl_interface': 3.0125443185115655e-06, 'learning_rate/action_model': 3.0125443185115655e-05, 'epoch': 1.34}) +[2026-06-16 07:09:57,853][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.00021481217117980123, 'train/grad_norm_pre_clip': 0.023986268788576126, 'timing/data': 0.030633077956736088, 'timing/model': 1.53245017118752, 'learning_rate/qwen_vl_interface': 3.0098583815776615e-06, 'learning_rate/action_model': 3.0098583815776615e-05, 'epoch': 1.34}) +[2026-06-16 07:10:04,238][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.000750868406612426, 'train/grad_norm_pre_clip': 0.11195450276136398, 'timing/data': 0.03474025288596749, 'timing/model': 1.574273953679949, 'learning_rate/qwen_vl_interface': 3.0071733654813595e-06, 'learning_rate/action_model': 3.0071733654813595e-05, 'epoch': 1.34}) +[2026-06-16 07:10:10,482][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.15488748252391815, 'train/grad_norm_pre_clip': 17.210872650146484, 'timing/data': 0.03054964914917946, 'timing/model': 1.7807111949659884, 'learning_rate/qwen_vl_interface': 3.0044892713263685e-06, 'learning_rate/action_model': 3.0044892713263682e-05, 'epoch': 1.34}) +[2026-06-16 07:10:16,792][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.2648939788341522, 'train/grad_norm_pre_clip': 10.739531517028809, 'timing/data': 0.030802059918642044, 'timing/model': 1.7810098016634583, 'learning_rate/qwen_vl_interface': 3.001806100216017e-06, 'learning_rate/action_model': 3.001806100216017e-05, 'epoch': 1.34}) +[2026-06-16 07:10:23,025][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.0007274673553183675, 'train/grad_norm_pre_clip': 0.17776937782764435, 'timing/data': 0.032056884840130806, 'timing/model': 1.606906773056835, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 1.34}) +[2026-06-16 07:10:29,255][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.3376767039299011, 'train/grad_norm_pre_clip': 18.456989288330078, 'timing/data': 0.03217356512323022, 'timing/model': 1.5351771097630262, 'learning_rate/qwen_vl_interface': 2.996442531540657e-06, 'learning_rate/action_model': 2.9964425315406568e-05, 'epoch': 1.34}) +[2026-06-16 07:10:35,568][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.054064057767391205, 'train/grad_norm_pre_clip': 21.480459213256836, 'timing/data': 0.029826417099684477, 'timing/model': 1.5797386639751494, 'learning_rate/qwen_vl_interface': 2.993762136180408e-06, 'learning_rate/action_model': 2.993762136180408e-05, 'epoch': 1.34}) +[2026-06-16 07:10:42,187][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.0009578977478668094, 'train/grad_norm_pre_clip': 0.25357094407081604, 'timing/data': 0.0316596943885088, 'timing/model': 1.5429734149947762, 'learning_rate/qwen_vl_interface': 2.9910826682743195e-06, 'learning_rate/action_model': 2.9910826682743192e-05, 'epoch': 1.34}) +[2026-06-16 07:10:48,611][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.4076218008995056, 'train/grad_norm_pre_clip': 7.305240631103516, 'timing/data': 0.0345790758728981, 'timing/model': 1.8872445751912892, 'learning_rate/qwen_vl_interface': 2.9884041289238192e-06, 'learning_rate/action_model': 2.988404128923819e-05, 'epoch': 1.34}) +[2026-06-16 07:10:54,661][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.04791721701622009, 'train/grad_norm_pre_clip': 23.251331329345703, 'timing/data': 0.0314681651070714, 'timing/model': 1.5345593602396548, 'learning_rate/qwen_vl_interface': 2.9857265192299557e-06, 'learning_rate/action_model': 2.9857265192299556e-05, 'epoch': 1.34}) +[2026-06-16 07:11:00,978][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.08338335156440735, 'train/grad_norm_pre_clip': 12.610881805419922, 'timing/data': 0.03295052004978061, 'timing/model': 1.5603993469849229, 'learning_rate/qwen_vl_interface': 2.9830498402933894e-06, 'learning_rate/action_model': 2.9830498402933894e-05, 'epoch': 1.34}) +[2026-06-16 07:11:07,271][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.06497013568878174, 'train/grad_norm_pre_clip': 10.160922050476074, 'timing/data': 0.03181596612557769, 'timing/model': 1.5447257831692696, 'learning_rate/qwen_vl_interface': 2.9803740932144025e-06, 'learning_rate/action_model': 2.9803740932144024e-05, 'epoch': 1.35}) +[2026-06-16 07:11:13,652][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.012992226518690586, 'train/grad_norm_pre_clip': 6.051804065704346, 'timing/data': 0.03185782628133893, 'timing/model': 1.596183073706925, 'learning_rate/qwen_vl_interface': 2.9776992790928953e-06, 'learning_rate/action_model': 2.9776992790928952e-05, 'epoch': 1.35}) +[2026-06-16 07:11:19,968][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.000267840368906036, 'train/grad_norm_pre_clip': 0.028131786733865738, 'timing/data': 0.03220973815768957, 'timing/model': 1.5579596771858633, 'learning_rate/qwen_vl_interface': 2.975025399028382e-06, 'learning_rate/action_model': 2.975025399028382e-05, 'epoch': 1.35}) +[2026-06-16 07:11:26,269][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.011147323995828629, 'train/grad_norm_pre_clip': 2.9721765518188477, 'timing/data': 0.031341145280748606, 'timing/model': 1.7559594851918519, 'learning_rate/qwen_vl_interface': 2.9723524541199943e-06, 'learning_rate/action_model': 2.972352454119994e-05, 'epoch': 1.35}) +[2026-06-16 07:11:32,452][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.00230508204549551, 'train/grad_norm_pre_clip': 0.41663119196891785, 'timing/data': 0.03219712944701314, 'timing/model': 1.5417440407909453, 'learning_rate/qwen_vl_interface': 2.969680445466478e-06, 'learning_rate/action_model': 2.969680445466478e-05, 'epoch': 1.35}) +[2026-06-16 07:11:38,659][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.11499841511249542, 'train/grad_norm_pre_clip': 11.3492431640625, 'timing/data': 0.030305661726742983, 'timing/model': 1.5363833433948457, 'learning_rate/qwen_vl_interface': 2.9670093741661955e-06, 'learning_rate/action_model': 2.9670093741661954e-05, 'epoch': 1.35}) +[2026-06-16 07:11:44,958][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.1576627790927887, 'train/grad_norm_pre_clip': 19.549556732177734, 'timing/data': 0.030505621805787086, 'timing/model': 1.5382667719386518, 'learning_rate/qwen_vl_interface': 2.964339241317123e-06, 'learning_rate/action_model': 2.964339241317123e-05, 'epoch': 1.35}) +[2026-06-16 07:11:51,348][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.11094344407320023, 'train/grad_norm_pre_clip': 8.791985511779785, 'timing/data': 0.032776761800050735, 'timing/model': 1.5987690268084407, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 1.35}) +[2026-06-16 07:11:57,566][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.09959039092063904, 'train/grad_norm_pre_clip': 12.831599235534668, 'timing/data': 0.030408713035285473, 'timing/model': 1.5474792490713298, 'learning_rate/qwen_vl_interface': 2.9590017953625854e-06, 'learning_rate/action_model': 2.9590017953625853e-05, 'epoch': 1.35}) +[2026-06-16 07:12:03,905][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.007338108494877815, 'train/grad_norm_pre_clip': 1.5903359651565552, 'timing/data': 0.03203576896339655, 'timing/model': 1.771990635432303, 'learning_rate/qwen_vl_interface': 2.9563344844511455e-06, 'learning_rate/action_model': 2.9563344844511455e-05, 'epoch': 1.35}) +[2026-06-16 07:12:10,066][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.0017685738857835531, 'train/grad_norm_pre_clip': 0.23333117365837097, 'timing/data': 0.031142255757004023, 'timing/model': 1.5440612379461527, 'learning_rate/qwen_vl_interface': 2.9536681163789568e-06, 'learning_rate/action_model': 2.9536681163789564e-05, 'epoch': 1.35}) +[2026-06-16 07:12:16,364][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.04741983115673065, 'train/grad_norm_pre_clip': 9.381604194641113, 'timing/data': 0.03130214707925916, 'timing/model': 1.5410082279704511, 'learning_rate/qwen_vl_interface': 2.9510026922420653e-06, 'learning_rate/action_model': 2.9510026922420653e-05, 'epoch': 1.35}) +[2026-06-16 07:12:22,767][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.0548473559319973, 'train/grad_norm_pre_clip': 16.925344467163086, 'timing/data': 0.030717664398252964, 'timing/model': 1.5408750688657165, 'learning_rate/qwen_vl_interface': 2.9483382131361273e-06, 'learning_rate/action_model': 2.948338213136127e-05, 'epoch': 1.35}) +[2026-06-16 07:12:29,182][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.09747938066720963, 'train/grad_norm_pre_clip': 7.40444803237915, 'timing/data': 0.028931397944688797, 'timing/model': 1.5565376179292798, 'learning_rate/qwen_vl_interface': 2.9456746801564072e-06, 'learning_rate/action_model': 2.945674680156407e-05, 'epoch': 1.35}) +[2026-06-16 07:12:35,765][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.010429282672703266, 'train/grad_norm_pre_clip': 4.053769111633301, 'timing/data': 0.02691311063244939, 'timing/model': 1.7339599840342999, 'learning_rate/qwen_vl_interface': 2.9430120943977854e-06, 'learning_rate/action_model': 2.9430120943977853e-05, 'epoch': 1.35}) +[2026-06-16 07:12:42,231][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.1394798457622528, 'train/grad_norm_pre_clip': 18.27630615234375, 'timing/data': 0.03143290802836418, 'timing/model': 1.806029297877103, 'learning_rate/qwen_vl_interface': 2.940350456954748e-06, 'learning_rate/action_model': 2.940350456954748e-05, 'epoch': 1.35}) +[2026-06-16 07:12:48,365][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.000748503371141851, 'train/grad_norm_pre_clip': 0.10883113741874695, 'timing/data': 0.030317884404212236, 'timing/model': 1.5616302196867764, 'learning_rate/qwen_vl_interface': 2.9376897689213952e-06, 'learning_rate/action_model': 2.937689768921395e-05, 'epoch': 1.35}) +[2026-06-16 07:12:54,671][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.00932812225073576, 'train/grad_norm_pre_clip': 5.625140190124512, 'timing/data': 0.031001536641269922, 'timing/model': 1.5442326739430428, 'learning_rate/qwen_vl_interface': 2.9350300313914364e-06, 'learning_rate/action_model': 2.9350300313914363e-05, 'epoch': 1.35}) +[2026-06-16 07:13:01,170][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.0033184366766363382, 'train/grad_norm_pre_clip': 0.6896606087684631, 'timing/data': 0.032402453012764454, 'timing/model': 1.5524685559794307, 'learning_rate/qwen_vl_interface': 2.9323712454581867e-06, 'learning_rate/action_model': 2.9323712454581866e-05, 'epoch': 1.35}) +[2026-06-16 07:13:07,780][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.1279546022415161, 'train/grad_norm_pre_clip': 19.188461303710938, 'timing/data': 0.032118350733071566, 'timing/model': 1.6034683091565967, 'learning_rate/qwen_vl_interface': 2.929713412214572e-06, 'learning_rate/action_model': 2.929713412214572e-05, 'epoch': 1.35}) +[2026-06-16 07:13:14,066][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.037280477583408356, 'train/grad_norm_pre_clip': 23.725976943969727, 'timing/data': 0.033118476159870625, 'timing/model': 1.5243107890710235, 'learning_rate/qwen_vl_interface': 2.9270565327531294e-06, 'learning_rate/action_model': 2.927056532753129e-05, 'epoch': 1.35}) +[2026-06-16 07:13:20,398][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.16911710798740387, 'train/grad_norm_pre_clip': 18.156978607177734, 'timing/data': 0.030883976723998785, 'timing/model': 1.7795821391046047, 'learning_rate/qwen_vl_interface': 2.924400608166003e-06, 'learning_rate/action_model': 2.924400608166003e-05, 'epoch': 1.35}) +[2026-06-16 07:13:26,464][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.11983169615268707, 'train/grad_norm_pre_clip': 19.906936645507812, 'timing/data': 0.030273614916950464, 'timing/model': 1.5298708598129451, 'learning_rate/qwen_vl_interface': 2.921745639544939e-06, 'learning_rate/action_model': 2.921745639544939e-05, 'epoch': 1.35}) +[2026-06-16 07:13:32,886][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.12484140694141388, 'train/grad_norm_pre_clip': 9.0819673538208, 'timing/data': 0.03122757887467742, 'timing/model': 1.66402177978307, 'learning_rate/qwen_vl_interface': 2.919091627981297e-06, 'learning_rate/action_model': 2.919091627981297e-05, 'epoch': 1.35}) +[2026-06-16 07:13:39,172][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.30340173840522766, 'train/grad_norm_pre_clip': 12.47880744934082, 'timing/data': 0.03038497595116496, 'timing/model': 1.5242988876998425, 'learning_rate/qwen_vl_interface': 2.9164385745660396e-06, 'learning_rate/action_model': 2.9164385745660394e-05, 'epoch': 1.36}) +[2026-06-16 07:13:45,570][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.0035090057644993067, 'train/grad_norm_pre_clip': 0.5369762182235718, 'timing/data': 0.030594157055020332, 'timing/model': 1.5193838779814541, 'learning_rate/qwen_vl_interface': 2.9137864803897385e-06, 'learning_rate/action_model': 2.9137864803897387e-05, 'epoch': 1.36}) +[2026-06-16 07:13:51,949][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.1474730223417282, 'train/grad_norm_pre_clip': 7.503653526306152, 'timing/data': 0.031468639615923166, 'timing/model': 1.5866444250568748, 'learning_rate/qwen_vl_interface': 2.9111353465425696e-06, 'learning_rate/action_model': 2.9111353465425694e-05, 'epoch': 1.36}) +[2026-06-16 07:13:58,305][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.21932867169380188, 'train/grad_norm_pre_clip': 3.94931697845459, 'timing/data': 0.03197651170194149, 'timing/model': 1.7718862439505756, 'learning_rate/qwen_vl_interface': 2.908485174114312e-06, 'learning_rate/action_model': 2.908485174114312e-05, 'epoch': 1.36}) +[2026-06-16 07:14:04,377][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.0022108994890004396, 'train/grad_norm_pre_clip': 0.34103694558143616, 'timing/data': 0.03175928210839629, 'timing/model': 1.556499577127397, 'learning_rate/qwen_vl_interface': 2.9058359641943523e-06, 'learning_rate/action_model': 2.9058359641943523e-05, 'epoch': 1.36}) +[2026-06-16 07:14:10,757][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.030653521418571472, 'train/grad_norm_pre_clip': 7.884881496429443, 'timing/data': 0.03412851132452488, 'timing/model': 1.6210498143918812, 'learning_rate/qwen_vl_interface': 2.9031877178716823e-06, 'learning_rate/action_model': 2.903187717871682e-05, 'epoch': 1.36}) +[2026-06-16 07:14:17,064][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.028356999158859253, 'train/grad_norm_pre_clip': 7.742593288421631, 'timing/data': 0.031315759755671024, 'timing/model': 1.579150621779263, 'learning_rate/qwen_vl_interface': 2.900540436234894e-06, 'learning_rate/action_model': 2.9005404362348937e-05, 'epoch': 1.36}) +[2026-06-16 07:14:23,451][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.008970879018306732, 'train/grad_norm_pre_clip': 2.157966136932373, 'timing/data': 0.03419370064511895, 'timing/model': 1.621304634027183, 'learning_rate/qwen_vl_interface': 2.897894120372185e-06, 'learning_rate/action_model': 2.897894120372185e-05, 'epoch': 1.36}) +[2026-06-16 07:14:29,763][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.019554901868104935, 'train/grad_norm_pre_clip': 6.714466094970703, 'timing/data': 0.031444841995835304, 'timing/model': 1.5276770908385515, 'learning_rate/qwen_vl_interface': 2.8952487713713585e-06, 'learning_rate/action_model': 2.8952487713713582e-05, 'epoch': 1.36}) +[2026-06-16 07:14:36,111][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.04241521656513214, 'train/grad_norm_pre_clip': 8.32943344116211, 'timing/data': 0.03084505908191204, 'timing/model': 1.7888652989640832, 'learning_rate/qwen_vl_interface': 2.8926043903198135e-06, 'learning_rate/action_model': 2.8926043903198135e-05, 'epoch': 1.36}) +[2026-06-16 07:14:42,160][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.0038867052644491196, 'train/grad_norm_pre_clip': 1.0994648933410645, 'timing/data': 0.030509857926517725, 'timing/model': 1.5432664481922984, 'learning_rate/qwen_vl_interface': 2.8899609783045574e-06, 'learning_rate/action_model': 2.8899609783045573e-05, 'epoch': 1.36}) +[2026-06-16 07:14:48,462][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.08265914022922516, 'train/grad_norm_pre_clip': 44.78111267089844, 'timing/data': 0.03206311399117112, 'timing/model': 1.5386446379125118, 'learning_rate/qwen_vl_interface': 2.8873185364121965e-06, 'learning_rate/action_model': 2.8873185364121963e-05, 'epoch': 1.36}) +[2026-06-16 07:14:54,763][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.046280067414045334, 'train/grad_norm_pre_clip': 19.758325576782227, 'timing/data': 0.03454780438914895, 'timing/model': 1.5441620768979192, 'learning_rate/qwen_vl_interface': 2.8846770657289386e-06, 'learning_rate/action_model': 2.8846770657289386e-05, 'epoch': 1.36}) +[2026-06-16 07:15:01,063][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.07960616052150726, 'train/grad_norm_pre_clip': 28.734607696533203, 'timing/data': 0.032010653987526894, 'timing/model': 1.541677765082568, 'learning_rate/qwen_vl_interface': 2.882036567340593e-06, 'learning_rate/action_model': 2.8820365673405932e-05, 'epoch': 1.36}) +[2026-06-16 07:15:07,370][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.07475965470075607, 'train/grad_norm_pre_clip': 19.890277862548828, 'timing/data': 0.030314987991005182, 'timing/model': 1.51875907368958, 'learning_rate/qwen_vl_interface': 2.8793970423325684e-06, 'learning_rate/action_model': 2.8793970423325685e-05, 'epoch': 1.36}) +[2026-06-16 07:15:13,753][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.011400178074836731, 'train/grad_norm_pre_clip': 9.270918846130371, 'timing/data': 0.030592039227485657, 'timing/model': 1.831793602090329, 'learning_rate/qwen_vl_interface': 2.8767584917898756e-06, 'learning_rate/action_model': 2.8767584917898754e-05, 'epoch': 1.36}) +[2026-06-16 07:15:19,868][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.08925282955169678, 'train/grad_norm_pre_clip': 10.153166770935059, 'timing/data': 0.02983742579817772, 'timing/model': 1.550241001881659, 'learning_rate/qwen_vl_interface': 2.874120916797121e-06, 'learning_rate/action_model': 2.8741209167971207e-05, 'epoch': 1.36}) +[2026-06-16 07:15:26,260][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.001440328313037753, 'train/grad_norm_pre_clip': 0.14325788617134094, 'timing/data': 0.03457762487232685, 'timing/model': 1.542660130187869, 'learning_rate/qwen_vl_interface': 2.8714843184385124e-06, 'learning_rate/action_model': 2.871484318438512e-05, 'epoch': 1.36}) +[2026-06-16 07:15:32,568][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.12220257520675659, 'train/grad_norm_pre_clip': 5.5506086349487305, 'timing/data': 0.031969984993338585, 'timing/model': 1.5372893973253667, 'learning_rate/qwen_vl_interface': 2.8688486977978575e-06, 'learning_rate/action_model': 2.8688486977978572e-05, 'epoch': 1.36}) +[2026-06-16 07:15:38,865][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.009325849823653698, 'train/grad_norm_pre_clip': 1.136111855506897, 'timing/data': 0.03197953524067998, 'timing/model': 1.5407506581395864, 'learning_rate/qwen_vl_interface': 2.8662140559585584e-06, 'learning_rate/action_model': 2.866214055958558e-05, 'epoch': 1.36}) +[2026-06-16 07:15:45,258][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.014651333913207054, 'train/grad_norm_pre_clip': 4.582727432250977, 'timing/data': 0.03158255387097597, 'timing/model': 1.5969556150957942, 'learning_rate/qwen_vl_interface': 2.8635803940036162e-06, 'learning_rate/action_model': 2.863580394003616e-05, 'epoch': 1.36}) +[2026-06-16 07:15:51,606][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.17943553626537323, 'train/grad_norm_pre_clip': 14.211847305297852, 'timing/data': 0.03160637430846691, 'timing/model': 1.7784889661706984, 'learning_rate/qwen_vl_interface': 2.8609477130156317e-06, 'learning_rate/action_model': 2.8609477130156314e-05, 'epoch': 1.36}) +[2026-06-16 07:15:57,852][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.029876604676246643, 'train/grad_norm_pre_clip': 8.59370231628418, 'timing/data': 0.03127827309072018, 'timing/model': 1.7261707088910043, 'learning_rate/qwen_vl_interface': 2.8583160140767997e-06, 'learning_rate/action_model': 2.8583160140767994e-05, 'epoch': 1.36}) +[2026-06-16 07:16:03,966][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.026319459080696106, 'train/grad_norm_pre_clip': 7.8293046951293945, 'timing/data': 0.03198010567575693, 'timing/model': 1.535601266194135, 'learning_rate/qwen_vl_interface': 2.8556852982689125e-06, 'learning_rate/action_model': 2.8556852982689126e-05, 'epoch': 1.36}) +[2026-06-16 07:16:10,450][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.005123714916408062, 'train/grad_norm_pre_clip': 1.1709553003311157, 'timing/data': 0.03192715626209974, 'timing/model': 1.6266597919166088, 'learning_rate/qwen_vl_interface': 2.8530555666733584e-06, 'learning_rate/action_model': 2.8530555666733583e-05, 'epoch': 1.36}) +[2026-06-16 07:16:16,761][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.09455891698598862, 'train/grad_norm_pre_clip': 11.901719093322754, 'timing/data': 0.03255501901730895, 'timing/model': 1.6376016018912196, 'learning_rate/qwen_vl_interface': 2.85042682037112e-06, 'learning_rate/action_model': 2.85042682037112e-05, 'epoch': 1.37}) +[2026-06-16 07:16:23,169][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.002366185886785388, 'train/grad_norm_pre_clip': 0.6832840442657471, 'timing/data': 0.03435358498245478, 'timing/model': 1.5973685709759593, 'learning_rate/qwen_vl_interface': 2.8477990604427764e-06, 'learning_rate/action_model': 2.8477990604427763e-05, 'epoch': 1.37}) +[2026-06-16 07:16:29,494][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.008621187880635262, 'train/grad_norm_pre_clip': 2.4799418449401855, 'timing/data': 0.03222721675410867, 'timing/model': 1.7674090308137238, 'learning_rate/qwen_vl_interface': 2.8451722879684994e-06, 'learning_rate/action_model': 2.8451722879684994e-05, 'epoch': 1.37}) +[2026-06-16 07:16:35,934][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.1866483986377716, 'train/grad_norm_pre_clip': 20.528087615966797, 'timing/data': 0.0359611427411437, 'timing/model': 1.803657972253859, 'learning_rate/qwen_vl_interface': 2.842546504028059e-06, 'learning_rate/action_model': 2.8425465040280586e-05, 'epoch': 1.37}) +[2026-06-16 07:16:41,974][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.09986463189125061, 'train/grad_norm_pre_clip': 12.358662605285645, 'timing/data': 0.032551608979701996, 'timing/model': 1.5492665711790323, 'learning_rate/qwen_vl_interface': 2.83992170970081e-06, 'learning_rate/action_model': 2.8399217097008102e-05, 'epoch': 1.37}) +[2026-06-16 07:16:48,368][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.0006246358389034867, 'train/grad_norm_pre_clip': 0.03576837480068207, 'timing/data': 0.03224567836150527, 'timing/model': 1.5480928532779217, 'learning_rate/qwen_vl_interface': 2.837297906065711e-06, 'learning_rate/action_model': 2.8372979060657107e-05, 'epoch': 1.37}) +[2026-06-16 07:16:54,664][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.06582923978567123, 'train/grad_norm_pre_clip': 12.298784255981445, 'timing/data': 0.03207194712013006, 'timing/model': 1.5346637857146561, 'learning_rate/qwen_vl_interface': 2.834675094201306e-06, 'learning_rate/action_model': 2.8346750942013063e-05, 'epoch': 1.37}) +[2026-06-16 07:17:01,161][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.09482485055923462, 'train/grad_norm_pre_clip': 14.718108177185059, 'timing/data': 0.03366133291274309, 'timing/model': 1.588781057856977, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 1.37}) +[2026-06-16 07:17:07,501][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.004743841476738453, 'train/grad_norm_pre_clip': 0.6010077595710754, 'timing/data': 0.030716721899807453, 'timing/model': 1.7765155211091042, 'learning_rate/qwen_vl_interface': 2.829432450096731e-06, 'learning_rate/action_model': 2.8294324500967306e-05, 'epoch': 1.37}) +[2026-06-16 07:17:13,667][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.008992979303002357, 'train/grad_norm_pre_clip': 3.332531452178955, 'timing/data': 0.031284044962376356, 'timing/model': 1.5470325341448188, 'learning_rate/qwen_vl_interface': 2.8268126200116118e-06, 'learning_rate/action_model': 2.8268126200116117e-05, 'epoch': 1.37}) +[2026-06-16 07:17:19,967][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.001561650075018406, 'train/grad_norm_pre_clip': 0.23816949129104614, 'timing/data': 0.030917454045265913, 'timing/model': 1.5477100629359484, 'learning_rate/qwen_vl_interface': 2.824193786007294e-06, 'learning_rate/action_model': 2.8241937860072937e-05, 'epoch': 1.37}) +[2026-06-16 07:17:26,276][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.1611139476299286, 'train/grad_norm_pre_clip': 18.71324920654297, 'timing/data': 0.0307206017896533, 'timing/model': 1.5497202672995627, 'learning_rate/qwen_vl_interface': 2.821575949160279e-06, 'learning_rate/action_model': 2.8215759491602787e-05, 'epoch': 1.37}) +[2026-06-16 07:17:32,663][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.012736033648252487, 'train/grad_norm_pre_clip': 3.877110481262207, 'timing/data': 0.03555069211870432, 'timing/model': 1.5810448178090155, 'learning_rate/qwen_vl_interface': 2.8189591105466625e-06, 'learning_rate/action_model': 2.8189591105466624e-05, 'epoch': 1.37}) +[2026-06-16 07:17:38,963][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.01389465481042862, 'train/grad_norm_pre_clip': 4.305385112762451, 'timing/data': 0.032810081262141466, 'timing/model': 1.5589564158581197, 'learning_rate/qwen_vl_interface': 2.816343271242126e-06, 'learning_rate/action_model': 2.816343271242126e-05, 'epoch': 1.37}) +[2026-06-16 07:17:45,255][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.010840433649718761, 'train/grad_norm_pre_clip': 3.4685285091400146, 'timing/data': 0.030591881833970547, 'timing/model': 1.7283747303299606, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 1.37}) +[2026-06-16 07:17:51,365][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.0056537967175245285, 'train/grad_norm_pre_clip': 0.9541035890579224, 'timing/data': 0.03149314597249031, 'timing/model': 1.5321661899797618, 'learning_rate/qwen_vl_interface': 2.81111459486098e-06, 'learning_rate/action_model': 2.81111459486098e-05, 'epoch': 1.37}) +[2026-06-16 07:17:57,670][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.16455437242984772, 'train/grad_norm_pre_clip': 20.29209327697754, 'timing/data': 0.03122839517891407, 'timing/model': 1.5333171812817454, 'learning_rate/qwen_vl_interface': 2.8085017599336806e-06, 'learning_rate/action_model': 2.8085017599336806e-05, 'epoch': 1.37}) +[2026-06-16 07:18:03,971][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.0010389898670837283, 'train/grad_norm_pre_clip': 0.17115919291973114, 'timing/data': 0.03071408299729228, 'timing/model': 1.5528905936516821, 'learning_rate/qwen_vl_interface': 2.8058899286140828e-06, 'learning_rate/action_model': 2.805889928614083e-05, 'epoch': 1.37}) +[2026-06-16 07:18:10,360][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.1854361593723297, 'train/grad_norm_pre_clip': 10.731244087219238, 'timing/data': 0.0315867867320776, 'timing/model': 1.6018109358847141, 'learning_rate/qwen_vl_interface': 2.8032791019758127e-06, 'learning_rate/action_model': 2.8032791019758127e-05, 'epoch': 1.37}) +[2026-06-16 07:18:16,775][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.009522262960672379, 'train/grad_norm_pre_clip': 2.791182518005371, 'timing/data': 0.030842794105410576, 'timing/model': 1.5662262486293912, 'learning_rate/qwen_vl_interface': 2.800669281092084e-06, 'learning_rate/action_model': 2.8006692810920838e-05, 'epoch': 1.37}) +[2026-06-16 07:18:23,242][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.2356310933828354, 'train/grad_norm_pre_clip': 13.691003799438477, 'timing/data': 0.030962234362959862, 'timing/model': 1.7865690616890788, 'learning_rate/qwen_vl_interface': 2.7980604670356938e-06, 'learning_rate/action_model': 2.7980604670356937e-05, 'epoch': 1.37}) +[2026-06-16 07:18:29,365][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.10389108210802078, 'train/grad_norm_pre_clip': 19.09612464904785, 'timing/data': 0.031516777351498604, 'timing/model': 1.5395274320617318, 'learning_rate/qwen_vl_interface': 2.795452660879027e-06, 'learning_rate/action_model': 2.7954526608790273e-05, 'epoch': 1.37}) +[2026-06-16 07:18:35,668][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.002451741835102439, 'train/grad_norm_pre_clip': 0.38499411940574646, 'timing/data': 0.03189521795138717, 'timing/model': 1.540879094041884, 'learning_rate/qwen_vl_interface': 2.7928458636940565e-06, 'learning_rate/action_model': 2.7928458636940564e-05, 'epoch': 1.37}) +[2026-06-16 07:18:42,068][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.11046727001667023, 'train/grad_norm_pre_clip': 11.025179862976074, 'timing/data': 0.03076639398932457, 'timing/model': 1.5369098922237754, 'learning_rate/qwen_vl_interface': 2.790240076552336e-06, 'learning_rate/action_model': 2.7902400765523355e-05, 'epoch': 1.37}) +[2026-06-16 07:29:47,036][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.004797214642167091, 'train/grad_norm_pre_clip': 1.6487432718276978, 'eval/loss': 0.03583775460720062, 'eval/flap_precision': 0.9549071618037135, 'eval/flap_recall': 0.9822646657571623, 'eval/flap_f1': 0.968392737054472, 'eval/flap_rate_model': 0.20944444444444443, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.968392737054472, 'timing/data': 0.03084303392097354, 'timing/model': 1.583337345160544, 'learning_rate/qwen_vl_interface': 2.787635300525006e-06, 'learning_rate/action_model': 2.787635300525006e-05, 'epoch': 1.37}) +[2026-06-16 07:29:47,037][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3400_state +[2026-06-16 07:29:47,043][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 07:29:55,720][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3400_state/model.safetensors +[2026-06-16 07:30:16,439][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3400_state/optimizer.bin +[2026-06-16 07:30:16,440][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3400_state/sampler.bin +[2026-06-16 07:30:16,440][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3400_state/sampler_1.bin +[2026-06-16 07:30:16,441][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3400_state/random_states_0.pkl +[2026-06-16 07:30:16,628][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3400 +[2026-06-16 07:30:16,629][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 07:30:16,644][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 07:30:22,845][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.14949430525302887, 'train/grad_norm_pre_clip': 31.553268432617188, 'timing/data': 0.030979950912296772, 'timing/model': 1.4957686788402498, 'learning_rate/qwen_vl_interface': 2.7850315366827933e-06, 'learning_rate/action_model': 2.7850315366827934e-05, 'epoch': 1.38}) +[2026-06-16 07:30:29,083][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.21564731001853943, 'train/grad_norm_pre_clip': 22.929508209228516, 'timing/data': 0.030580628663301468, 'timing/model': 1.7997276689857244, 'learning_rate/qwen_vl_interface': 2.7824287860960076e-06, 'learning_rate/action_model': 2.7824287860960075e-05, 'epoch': 1.38}) +[2026-06-16 07:30:35,346][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.0039917719550430775, 'train/grad_norm_pre_clip': 1.473024606704712, 'timing/data': 0.030800370033830404, 'timing/model': 1.7657629358582199, 'learning_rate/qwen_vl_interface': 2.7798270498345377e-06, 'learning_rate/action_model': 2.7798270498345376e-05, 'epoch': 1.38}) +[2026-06-16 07:30:41,335][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.1538175344467163, 'train/grad_norm_pre_clip': 22.87380599975586, 'timing/data': 0.03149882331490517, 'timing/model': 1.5390880559571087, 'learning_rate/qwen_vl_interface': 2.777226328967861e-06, 'learning_rate/action_model': 2.7772263289678607e-05, 'epoch': 1.38}) +[2026-06-16 07:30:47,557][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.12412326037883759, 'train/grad_norm_pre_clip': 21.959888458251953, 'timing/data': 0.03240339690819383, 'timing/model': 1.5557658588513732, 'learning_rate/qwen_vl_interface': 2.774626624565035e-06, 'learning_rate/action_model': 2.774626624565035e-05, 'epoch': 1.38}) +[2026-06-16 07:30:53,853][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.0012452792143449187, 'train/grad_norm_pre_clip': 0.16094040870666504, 'timing/data': 0.034203174989670515, 'timing/model': 1.6118810176849365, 'learning_rate/qwen_vl_interface': 2.772027937694701e-06, 'learning_rate/action_model': 2.772027937694701e-05, 'epoch': 1.38}) +[2026-06-16 07:31:00,237][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.0016272516222670674, 'train/grad_norm_pre_clip': 0.37607342004776, 'timing/data': 0.03275906899943948, 'timing/model': 1.5647737798281014, 'learning_rate/qwen_vl_interface': 2.7694302694250806e-06, 'learning_rate/action_model': 2.7694302694250808e-05, 'epoch': 1.38}) +[2026-06-16 07:31:06,436][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.001867957296781242, 'train/grad_norm_pre_clip': 0.31076866388320923, 'timing/data': 0.032480517867952585, 'timing/model': 1.5403384566307068, 'learning_rate/qwen_vl_interface': 2.7668336208239775e-06, 'learning_rate/action_model': 2.7668336208239776e-05, 'epoch': 1.38}) +[2026-06-16 07:31:12,678][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.014131234958767891, 'train/grad_norm_pre_clip': 4.844781398773193, 'timing/data': 0.03444246621802449, 'timing/model': 1.780647428240627, 'learning_rate/qwen_vl_interface': 2.7642379929587755e-06, 'learning_rate/action_model': 2.7642379929587754e-05, 'epoch': 1.38}) +[2026-06-16 07:31:18,823][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.3380796015262604, 'train/grad_norm_pre_clip': 20.18714714050293, 'timing/data': 0.03130757296457887, 'timing/model': 1.6185097410343587, 'learning_rate/qwen_vl_interface': 2.7616433868964387e-06, 'learning_rate/action_model': 2.7616433868964388e-05, 'epoch': 1.38}) +[2026-06-16 07:31:25,143][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.008618268184363842, 'train/grad_norm_pre_clip': 4.3279547691345215, 'timing/data': 0.03216363023966551, 'timing/model': 1.5590007146820426, 'learning_rate/qwen_vl_interface': 2.7590498037035122e-06, 'learning_rate/action_model': 2.7590498037035124e-05, 'epoch': 1.38}) +[2026-06-16 07:31:31,451][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.0008793030283413827, 'train/grad_norm_pre_clip': 0.25355783104896545, 'timing/data': 0.031294372864067554, 'timing/model': 1.564172666054219, 'learning_rate/qwen_vl_interface': 2.7564572444461218e-06, 'learning_rate/action_model': 2.7564572444461217e-05, 'epoch': 1.38}) +[2026-06-16 07:31:37,764][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.06852133572101593, 'train/grad_norm_pre_clip': 12.837982177734375, 'timing/data': 0.03169008484110236, 'timing/model': 1.5567896980792284, 'learning_rate/qwen_vl_interface': 2.7538657101899674e-06, 'learning_rate/action_model': 2.7538657101899673e-05, 'epoch': 1.38}) +[2026-06-16 07:31:44,154][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.050932373851537704, 'train/grad_norm_pre_clip': 12.11545467376709, 'timing/data': 0.03066709591075778, 'timing/model': 1.5776004898361862, 'learning_rate/qwen_vl_interface': 2.751275202000332e-06, 'learning_rate/action_model': 2.7512752020003317e-05, 'epoch': 1.38}) +[2026-06-16 07:31:50,494][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.04592500999569893, 'train/grad_norm_pre_clip': 17.04108428955078, 'timing/data': 0.03196964506059885, 'timing/model': 1.7821611459366977, 'learning_rate/qwen_vl_interface': 2.7486857209420763e-06, 'learning_rate/action_model': 2.7486857209420763e-05, 'epoch': 1.38}) +[2026-06-16 07:31:56,653][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.1684344857931137, 'train/grad_norm_pre_clip': 9.425704956054688, 'timing/data': 0.03251240774989128, 'timing/model': 1.545321995858103, 'learning_rate/qwen_vl_interface': 2.7460972680796377e-06, 'learning_rate/action_model': 2.7460972680796375e-05, 'epoch': 1.38}) +[2026-06-16 07:32:02,953][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.12060143053531647, 'train/grad_norm_pre_clip': 20.892799377441406, 'timing/data': 0.032355066388845444, 'timing/model': 1.5421176007948816, 'learning_rate/qwen_vl_interface': 2.7435098444770314e-06, 'learning_rate/action_model': 2.7435098444770315e-05, 'epoch': 1.38}) +[2026-06-16 07:32:09,372][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.010704203508794308, 'train/grad_norm_pre_clip': 7.382118225097656, 'timing/data': 0.03198395203799009, 'timing/model': 1.6449637338519096, 'learning_rate/qwen_vl_interface': 2.7409234511978495e-06, 'learning_rate/action_model': 2.7409234511978492e-05, 'epoch': 1.38}) +[2026-06-16 07:32:15,664][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.017279602587223053, 'train/grad_norm_pre_clip': 12.237868309020996, 'timing/data': 0.030682752840220928, 'timing/model': 1.5142574831843376, 'learning_rate/qwen_vl_interface': 2.7383380893052606e-06, 'learning_rate/action_model': 2.7383380893052608e-05, 'epoch': 1.38}) +[2026-06-16 07:32:21,955][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.0015990162501111627, 'train/grad_norm_pre_clip': 0.39501309394836426, 'timing/data': 0.03270030580461025, 'timing/model': 1.5253746500238776, 'learning_rate/qwen_vl_interface': 2.73575375986201e-06, 'learning_rate/action_model': 2.7357537598620096e-05, 'epoch': 1.38}) +[2026-06-16 07:32:28,314][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.005893148947507143, 'train/grad_norm_pre_clip': 2.013760805130005, 'timing/data': 0.0327087352052331, 'timing/model': 1.8013759176246822, 'learning_rate/qwen_vl_interface': 2.733170463930417e-06, 'learning_rate/action_model': 2.7331704639304167e-05, 'epoch': 1.38}) +[2026-06-16 07:32:34,424][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.17887461185455322, 'train/grad_norm_pre_clip': 18.953815460205078, 'timing/data': 0.032305391039699316, 'timing/model': 1.6115176207385957, 'learning_rate/qwen_vl_interface': 2.7305882025723774e-06, 'learning_rate/action_model': 2.7305882025723774e-05, 'epoch': 1.38}) +[2026-06-16 07:32:40,656][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.24141673743724823, 'train/grad_norm_pre_clip': 21.147869110107422, 'timing/data': 0.03236159076914191, 'timing/model': 1.5388709637336433, 'learning_rate/qwen_vl_interface': 2.7280069768493626e-06, 'learning_rate/action_model': 2.7280069768493628e-05, 'epoch': 1.38}) +[2026-06-16 07:32:47,043][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.019752733409404755, 'train/grad_norm_pre_clip': 6.022009372711182, 'timing/data': 0.031301562674343586, 'timing/model': 1.6191064813174307, 'learning_rate/qwen_vl_interface': 2.7254267878224144e-06, 'learning_rate/action_model': 2.7254267878224143e-05, 'epoch': 1.38}) +[2026-06-16 07:32:53,254][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.09767445176839828, 'train/grad_norm_pre_clip': 14.318085670471191, 'timing/data': 0.031148942187428474, 'timing/model': 1.5338791529648006, 'learning_rate/qwen_vl_interface': 2.7228476365521524e-06, 'learning_rate/action_model': 2.722847636552152e-05, 'epoch': 1.39}) +[2026-06-16 07:32:59,560][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.024815848097205162, 'train/grad_norm_pre_clip': 11.122940063476562, 'timing/data': 0.030179213266819715, 'timing/model': 1.511249671690166, 'learning_rate/qwen_vl_interface': 2.7202695240987674e-06, 'learning_rate/action_model': 2.7202695240987676e-05, 'epoch': 1.39}) +[2026-06-16 07:33:05,941][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.22930149734020233, 'train/grad_norm_pre_clip': 24.142704010009766, 'timing/data': 0.031623892951756716, 'timing/model': 1.8179752491414547, 'learning_rate/qwen_vl_interface': 2.717692451522025e-06, 'learning_rate/action_model': 2.7176924515220248e-05, 'epoch': 1.39}) +[2026-06-16 07:33:11,994][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.1399768739938736, 'train/grad_norm_pre_clip': 28.368253707885742, 'timing/data': 0.03222847590222955, 'timing/model': 1.5792612861841917, 'learning_rate/qwen_vl_interface': 2.7151164198812624e-06, 'learning_rate/action_model': 2.7151164198812622e-05, 'epoch': 1.39}) +[2026-06-16 07:33:18,255][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.0008020734530873597, 'train/grad_norm_pre_clip': 0.09996101260185242, 'timing/data': 0.030944816768169403, 'timing/model': 1.543692421168089, 'learning_rate/qwen_vl_interface': 2.712541430235388e-06, 'learning_rate/action_model': 2.712541430235388e-05, 'epoch': 1.39}) +[2026-06-16 07:33:24,452][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.04078448563814163, 'train/grad_norm_pre_clip': 19.139545440673828, 'timing/data': 0.03127288771793246, 'timing/model': 1.5428301990032196, 'learning_rate/qwen_vl_interface': 2.7099674836428842e-06, 'learning_rate/action_model': 2.709967483642884e-05, 'epoch': 1.39}) +[2026-06-16 07:33:30,758][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.002678698394447565, 'train/grad_norm_pre_clip': 0.49325820803642273, 'timing/data': 0.035793366841971874, 'timing/model': 1.5473722540773451, 'learning_rate/qwen_vl_interface': 2.7073945811618018e-06, 'learning_rate/action_model': 2.7073945811618017e-05, 'epoch': 1.39}) +[2026-06-16 07:33:37,153][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.015013610944151878, 'train/grad_norm_pre_clip': 10.591998100280762, 'timing/data': 0.032869177870452404, 'timing/model': 1.5387340602464974, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 1.39}) +[2026-06-16 07:33:43,495][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.004416024312376976, 'train/grad_norm_pre_clip': 2.311556100845337, 'timing/data': 0.030698182992637157, 'timing/model': 1.7730914689600468, 'learning_rate/qwen_vl_interface': 2.7022519127639687e-06, 'learning_rate/action_model': 2.7022519127639682e-05, 'epoch': 1.39}) +[2026-06-16 07:33:49,646][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.23519502580165863, 'train/grad_norm_pre_clip': 44.96883773803711, 'timing/data': 0.03070004191249609, 'timing/model': 1.6299554598517716, 'learning_rate/qwen_vl_interface': 2.6996821489611713e-06, 'learning_rate/action_model': 2.699682148961171e-05, 'epoch': 1.39}) +[2026-06-16 07:33:55,958][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.08280155807733536, 'train/grad_norm_pre_clip': 22.867029190063477, 'timing/data': 0.031643647234886885, 'timing/model': 1.6355253518559039, 'learning_rate/qwen_vl_interface': 2.697113433497712e-06, 'learning_rate/action_model': 2.6971134334977115e-05, 'epoch': 1.39}) +[2026-06-16 07:34:02,306][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.01437053456902504, 'train/grad_norm_pre_clip': 4.236374855041504, 'timing/data': 0.03251431183889508, 'timing/model': 1.5723655931651592, 'learning_rate/qwen_vl_interface': 2.6945457674294863e-06, 'learning_rate/action_model': 2.694545767429486e-05, 'epoch': 1.39}) +[2026-06-16 07:34:08,636][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.11491241306066513, 'train/grad_norm_pre_clip': 13.363869667053223, 'timing/data': 0.032318669836968184, 'timing/model': 1.6056892657652497, 'learning_rate/qwen_vl_interface': 2.6919791518119676e-06, 'learning_rate/action_model': 2.6919791518119674e-05, 'epoch': 1.39}) +[2026-06-16 07:34:15,081][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.04138782620429993, 'train/grad_norm_pre_clip': 10.896042823791504, 'timing/data': 0.032385210040956736, 'timing/model': 1.6136731621809304, 'learning_rate/qwen_vl_interface': 2.6894135877001946e-06, 'learning_rate/action_model': 2.689413587700195e-05, 'epoch': 1.39}) +[2026-06-16 07:34:21,463][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.0014794060261920094, 'train/grad_norm_pre_clip': 0.20504316687583923, 'timing/data': 0.03236160008236766, 'timing/model': 1.8140451721847057, 'learning_rate/qwen_vl_interface': 2.686849076148775e-06, 'learning_rate/action_model': 2.6868490761487747e-05, 'epoch': 1.39}) +[2026-06-16 07:34:27,565][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.09556101262569427, 'train/grad_norm_pre_clip': 22.16029930114746, 'timing/data': 0.03369251685217023, 'timing/model': 1.5437950361520052, 'learning_rate/qwen_vl_interface': 2.68428561821188e-06, 'learning_rate/action_model': 2.6842856182118802e-05, 'epoch': 1.39}) +[2026-06-16 07:34:33,866][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.0026485337875783443, 'train/grad_norm_pre_clip': 0.9037958383560181, 'timing/data': 0.031259980984032154, 'timing/model': 1.5353928804397583, 'learning_rate/qwen_vl_interface': 2.681723214943253e-06, 'learning_rate/action_model': 2.681723214943253e-05, 'epoch': 1.39}) +[2026-06-16 07:34:40,172][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.002781522460281849, 'train/grad_norm_pre_clip': 0.8463540077209473, 'timing/data': 0.030593776144087315, 'timing/model': 1.537441001739353, 'learning_rate/qwen_vl_interface': 2.6791618673961996e-06, 'learning_rate/action_model': 2.6791618673961994e-05, 'epoch': 1.39}) +[2026-06-16 07:34:46,483][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.07997371256351471, 'train/grad_norm_pre_clip': 17.731351852416992, 'timing/data': 0.03284400887787342, 'timing/model': 1.5536494641564786, 'learning_rate/qwen_vl_interface': 2.676601576623594e-06, 'learning_rate/action_model': 2.6766015766235937e-05, 'epoch': 1.39}) +[2026-06-16 07:34:52,977][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.1057375892996788, 'train/grad_norm_pre_clip': 25.34366798400879, 'timing/data': 0.03334488160908222, 'timing/model': 1.6208944213576615, 'learning_rate/qwen_vl_interface': 2.674042343677874e-06, 'learning_rate/action_model': 2.6740423436778738e-05, 'epoch': 1.39}) +[2026-06-16 07:34:59,295][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.01128941960632801, 'train/grad_norm_pre_clip': 4.282440662384033, 'timing/data': 0.03140287892892957, 'timing/model': 1.7700263732112944, 'learning_rate/qwen_vl_interface': 2.6714841696110456e-06, 'learning_rate/action_model': 2.6714841696110456e-05, 'epoch': 1.39}) +[2026-06-16 07:35:05,521][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.0059434156864881516, 'train/grad_norm_pre_clip': 1.8252530097961426, 'timing/data': 0.03363559162244201, 'timing/model': 1.6963213039562106, 'learning_rate/qwen_vl_interface': 2.6689270554746743e-06, 'learning_rate/action_model': 2.6689270554746743e-05, 'epoch': 1.39}) +[2026-06-16 07:35:11,580][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.003825940191745758, 'train/grad_norm_pre_clip': 1.175559401512146, 'timing/data': 0.03138103801757097, 'timing/model': 1.5544584980234504, 'learning_rate/qwen_vl_interface': 2.6663710023198956e-06, 'learning_rate/action_model': 2.6663710023198956e-05, 'epoch': 1.39}) +[2026-06-16 07:35:17,961][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.0075132716447114944, 'train/grad_norm_pre_clip': 3.6942033767700195, 'timing/data': 0.03068625181913376, 'timing/model': 1.553703434765339, 'learning_rate/qwen_vl_interface': 2.663816011197405e-06, 'learning_rate/action_model': 2.6638160111974048e-05, 'epoch': 1.39}) +[2026-06-16 07:35:24,268][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.0025915824808180332, 'train/grad_norm_pre_clip': 0.377130925655365, 'timing/data': 0.030682031996548176, 'timing/model': 1.5480387997813523, 'learning_rate/qwen_vl_interface': 2.661262083157463e-06, 'learning_rate/action_model': 2.661262083157463e-05, 'epoch': 1.39}) +[2026-06-16 07:35:30,664][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.06742875277996063, 'train/grad_norm_pre_clip': 14.015568733215332, 'timing/data': 0.03198269894346595, 'timing/model': 1.5770153529010713, 'learning_rate/qwen_vl_interface': 2.6587092192498943e-06, 'learning_rate/action_model': 2.658709219249894e-05, 'epoch': 1.4}) +[2026-06-16 07:35:37,099][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.12433110922574997, 'train/grad_norm_pre_clip': 12.087583541870117, 'timing/data': 0.03070987109094858, 'timing/model': 1.8225277406163514, 'learning_rate/qwen_vl_interface': 2.656157420524084e-06, 'learning_rate/action_model': 2.6561574205240842e-05, 'epoch': 1.4}) +[2026-06-16 07:35:43,433][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.1453743427991867, 'train/grad_norm_pre_clip': 19.50749969482422, 'timing/data': 0.030483338981866837, 'timing/model': 1.7991710333153605, 'learning_rate/qwen_vl_interface': 2.6536066880289795e-06, 'learning_rate/action_model': 2.6536066880289794e-05, 'epoch': 1.4}) +[2026-06-16 07:35:49,471][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.018076762557029724, 'train/grad_norm_pre_clip': 12.381531715393066, 'timing/data': 0.030941141303628683, 'timing/model': 1.5483601000159979, 'learning_rate/qwen_vl_interface': 2.6510570228130916e-06, 'learning_rate/action_model': 2.6510570228130914e-05, 'epoch': 1.4}) +[2026-06-16 07:35:55,768][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.0035998220555484295, 'train/grad_norm_pre_clip': 1.3490689992904663, 'timing/data': 0.03243653988465667, 'timing/model': 1.5336133623495698, 'learning_rate/qwen_vl_interface': 2.648508425924491e-06, 'learning_rate/action_model': 2.6485084259244907e-05, 'epoch': 1.4}) +[2026-06-16 07:36:02,065][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.07648738473653793, 'train/grad_norm_pre_clip': 42.10198974609375, 'timing/data': 0.03270821180194616, 'timing/model': 1.5390702588483691, 'learning_rate/qwen_vl_interface': 2.6459608984108093e-06, 'learning_rate/action_model': 2.645960898410809e-05, 'epoch': 1.4}) +[2026-06-16 07:36:08,465][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.008593854494392872, 'train/grad_norm_pre_clip': 5.651041030883789, 'timing/data': 0.031187190674245358, 'timing/model': 1.5943162417970598, 'learning_rate/qwen_vl_interface': 2.6434144413192414e-06, 'learning_rate/action_model': 2.6434144413192414e-05, 'epoch': 1.4}) +[2026-06-16 07:36:14,799][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.10360477864742279, 'train/grad_norm_pre_clip': 8.619945526123047, 'timing/data': 0.03230147669091821, 'timing/model': 1.7748824101872742, 'learning_rate/qwen_vl_interface': 2.6408690556965347e-06, 'learning_rate/action_model': 2.640869055696535e-05, 'epoch': 1.4}) +[2026-06-16 07:36:20,855][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.13481637835502625, 'train/grad_norm_pre_clip': 22.887773513793945, 'timing/data': 0.03184619499370456, 'timing/model': 1.5330858752131462, 'learning_rate/qwen_vl_interface': 2.6383247425890047e-06, 'learning_rate/action_model': 2.6383247425890047e-05, 'epoch': 1.4}) +[2026-06-16 07:36:27,265][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.02825557440519333, 'train/grad_norm_pre_clip': 16.762672424316406, 'timing/data': 0.03158777207136154, 'timing/model': 1.534799890127033, 'learning_rate/qwen_vl_interface': 2.6357815030425204e-06, 'learning_rate/action_model': 2.6357815030425205e-05, 'epoch': 1.4}) +[2026-06-16 07:36:33,559][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.14520646631717682, 'train/grad_norm_pre_clip': 28.36981201171875, 'timing/data': 0.03238588711246848, 'timing/model': 1.5298439436592162, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 1.4}) +[2026-06-16 07:36:39,945][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.012133973650634289, 'train/grad_norm_pre_clip': 6.294950008392334, 'timing/data': 0.03098022798076272, 'timing/model': 1.5846429620869458, 'learning_rate/qwen_vl_interface': 2.6306982488139704e-06, 'learning_rate/action_model': 2.6306982488139702e-05, 'epoch': 1.4}) +[2026-06-16 07:36:46,163][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.057685475796461105, 'train/grad_norm_pre_clip': 40.91964340209961, 'timing/data': 0.031595680862665176, 'timing/model': 1.574972698930651, 'learning_rate/qwen_vl_interface': 2.628158236221438e-06, 'learning_rate/action_model': 2.628158236221438e-05, 'epoch': 1.4}) +[2026-06-16 07:36:52,473][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.10323171317577362, 'train/grad_norm_pre_clip': 21.09661293029785, 'timing/data': 0.03154828026890755, 'timing/model': 1.7360341139137745, 'learning_rate/qwen_vl_interface': 2.6256193013690183e-06, 'learning_rate/action_model': 2.6256193013690182e-05, 'epoch': 1.4}) +[2026-06-16 07:36:58,784][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.0013283896259963512, 'train/grad_norm_pre_clip': 0.3271072208881378, 'timing/data': 0.034179185051470995, 'timing/model': 1.6654727240093052, 'learning_rate/qwen_vl_interface': 2.623081445300373e-06, 'learning_rate/action_model': 2.6230814453003728e-05, 'epoch': 1.4}) +[2026-06-16 07:37:05,073][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.003152982098981738, 'train/grad_norm_pre_clip': 1.036097526550293, 'timing/data': 0.03092717519029975, 'timing/model': 1.5472637098282576, 'learning_rate/qwen_vl_interface': 2.6205446690587183e-06, 'learning_rate/action_model': 2.6205446690587182e-05, 'epoch': 1.4}) +[2026-06-16 07:37:11,361][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.00045987419434823096, 'train/grad_norm_pre_clip': 0.06331407278776169, 'timing/data': 0.031283827032893896, 'timing/model': 1.5387830552645028, 'learning_rate/qwen_vl_interface': 2.618008973686828e-06, 'learning_rate/action_model': 2.618008973686828e-05, 'epoch': 1.4}) +[2026-06-16 07:37:17,667][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.010663824155926704, 'train/grad_norm_pre_clip': 9.454879760742188, 'timing/data': 0.032770656049251556, 'timing/model': 1.517534895800054, 'learning_rate/qwen_vl_interface': 2.615474360227028e-06, 'learning_rate/action_model': 2.615474360227028e-05, 'epoch': 1.4}) +[2026-06-16 07:37:24,034][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.045689575374126434, 'train/grad_norm_pre_clip': 12.130651473999023, 'timing/data': 0.03023487888276577, 'timing/model': 1.5785410767421126, 'learning_rate/qwen_vl_interface': 2.612940829721204e-06, 'learning_rate/action_model': 2.612940829721204e-05, 'epoch': 1.4}) +[2026-06-16 07:37:30,258][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.18168456852436066, 'train/grad_norm_pre_clip': 27.029170989990234, 'timing/data': 0.030487610027194023, 'timing/model': 1.7399137821048498, 'learning_rate/qwen_vl_interface': 2.6104083832107948e-06, 'learning_rate/action_model': 2.6104083832107945e-05, 'epoch': 1.4}) +[2026-06-16 07:37:36,361][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.12014321982860565, 'train/grad_norm_pre_clip': 17.470645904541016, 'timing/data': 0.029853237327188253, 'timing/model': 1.5454622660763562, 'learning_rate/qwen_vl_interface': 2.607877021736793e-06, 'learning_rate/action_model': 2.6078770217367926e-05, 'epoch': 1.4}) +[2026-06-16 07:37:42,660][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.0006085664499551058, 'train/grad_norm_pre_clip': 0.1546865999698639, 'timing/data': 0.03046874888241291, 'timing/model': 1.539589588996023, 'learning_rate/qwen_vl_interface': 2.605346746339745e-06, 'learning_rate/action_model': 2.605346746339745e-05, 'epoch': 1.4}) +[2026-06-16 07:37:49,146][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.02797297202050686, 'train/grad_norm_pre_clip': 9.726140022277832, 'timing/data': 0.03144297515973449, 'timing/model': 1.5400904510170221, 'learning_rate/qwen_vl_interface': 2.602817558059753e-06, 'learning_rate/action_model': 2.6028175580597532e-05, 'epoch': 1.4}) +[2026-06-16 07:37:55,487][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.03520287200808525, 'train/grad_norm_pre_clip': 17.237573623657227, 'timing/data': 0.03130788402631879, 'timing/model': 1.5362544157542288, 'learning_rate/qwen_vl_interface': 2.60028945793647e-06, 'learning_rate/action_model': 2.60028945793647e-05, 'epoch': 1.4}) +[2026-06-16 07:38:01,804][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.15388359129428864, 'train/grad_norm_pre_clip': 25.1438045501709, 'timing/data': 0.030292117036879063, 'timing/model': 1.780078524723649, 'learning_rate/qwen_vl_interface': 2.5977624470091024e-06, 'learning_rate/action_model': 2.5977624470091023e-05, 'epoch': 1.4}) +[2026-06-16 07:38:07,861][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.009916652925312519, 'train/grad_norm_pre_clip': 4.143416881561279, 'timing/data': 0.02990884520113468, 'timing/model': 1.5323927816934884, 'learning_rate/qwen_vl_interface': 2.59523652631641e-06, 'learning_rate/action_model': 2.5952365263164097e-05, 'epoch': 1.41}) +[2026-06-16 07:38:14,162][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.01466919481754303, 'train/grad_norm_pre_clip': 6.602574825286865, 'timing/data': 0.030268377158790827, 'timing/model': 1.540618719998747, 'learning_rate/qwen_vl_interface': 2.5927116968967024e-06, 'learning_rate/action_model': 2.5927116968967025e-05, 'epoch': 1.41}) +[2026-06-16 07:38:20,370][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.0006118572200648487, 'train/grad_norm_pre_clip': 0.11198969185352325, 'timing/data': 0.030961259733885527, 'timing/model': 1.5536677748896182, 'learning_rate/qwen_vl_interface': 2.5901879597878417e-06, 'learning_rate/action_model': 2.5901879597878415e-05, 'epoch': 1.41}) +[2026-06-16 07:38:26,763][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.14860299229621887, 'train/grad_norm_pre_clip': 20.294605255126953, 'timing/data': 0.03396992711350322, 'timing/model': 1.568879353813827, 'learning_rate/qwen_vl_interface': 2.5876653160272395e-06, 'learning_rate/action_model': 2.5876653160272397e-05, 'epoch': 1.41}) +[2026-06-16 07:38:33,072][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.037287648767232895, 'train/grad_norm_pre_clip': 15.23471450805664, 'timing/data': 0.03101216908544302, 'timing/model': 1.56363918306306, 'learning_rate/qwen_vl_interface': 2.5851437666518607e-06, 'learning_rate/action_model': 2.5851437666518607e-05, 'epoch': 1.41}) +[2026-06-16 07:38:39,441][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.07375500351190567, 'train/grad_norm_pre_clip': 47.794002532958984, 'timing/data': 0.029903032816946507, 'timing/model': 1.8238578578457236, 'learning_rate/qwen_vl_interface': 2.5826233126982198e-06, 'learning_rate/action_model': 2.5826233126982196e-05, 'epoch': 1.41}) +[2026-06-16 07:38:45,564][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.02296452969312668, 'train/grad_norm_pre_clip': 13.4053316116333, 'timing/data': 0.03173037292435765, 'timing/model': 1.5463540740311146, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 1.41}) +[2026-06-16 07:38:51,970][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.24128778278827667, 'train/grad_norm_pre_clip': 17.935422897338867, 'timing/data': 0.03412013594061136, 'timing/model': 1.5512349531054497, 'learning_rate/qwen_vl_interface': 2.577585695199954e-06, 'learning_rate/action_model': 2.5775856951999538e-05, 'epoch': 1.41}) +[2026-06-16 07:38:58,303][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.11107934266328812, 'train/grad_norm_pre_clip': 10.588659286499023, 'timing/data': 0.03203370375558734, 'timing/model': 1.555713628884405, 'learning_rate/qwen_vl_interface': 2.575068533726101e-06, 'learning_rate/action_model': 2.575068533726101e-05, 'epoch': 1.41}) +[2026-06-16 07:39:04,557][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.00042566374759189785, 'train/grad_norm_pre_clip': 0.1064520999789238, 'timing/data': 0.029615039005875587, 'timing/model': 1.5340867373161018, 'learning_rate/qwen_vl_interface': 2.5725524718155333e-06, 'learning_rate/action_model': 2.572552471815533e-05, 'epoch': 1.41}) +[2026-06-16 07:39:10,856][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.03524664416909218, 'train/grad_norm_pre_clip': 12.593650817871094, 'timing/data': 0.03031455399468541, 'timing/model': 1.5445834477432072, 'learning_rate/qwen_vl_interface': 2.570037510502508e-06, 'learning_rate/action_model': 2.5700375105025083e-05, 'epoch': 1.41}) +[2026-06-16 07:39:17,239][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.00013151646999176592, 'train/grad_norm_pre_clip': 0.01482757180929184, 'timing/data': 0.029222017154097557, 'timing/model': 1.8271625339984894, 'learning_rate/qwen_vl_interface': 2.5675236508208313e-06, 'learning_rate/action_model': 2.5675236508208313e-05, 'epoch': 1.41}) +[2026-06-16 07:39:23,257][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.251761794090271, 'train/grad_norm_pre_clip': 21.257823944091797, 'timing/data': 0.03042304376140237, 'timing/model': 1.5382419060915709, 'learning_rate/qwen_vl_interface': 2.565010893803856e-06, 'learning_rate/action_model': 2.565010893803856e-05, 'epoch': 1.41}) +[2026-06-16 07:39:29,560][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.01879074238240719, 'train/grad_norm_pre_clip': 12.945834159851074, 'timing/data': 0.02986529003828764, 'timing/model': 1.542531682178378, 'learning_rate/qwen_vl_interface': 2.562499240484482e-06, 'learning_rate/action_model': 2.5624992404844816e-05, 'epoch': 1.41}) +[2026-06-16 07:39:35,857][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.10097213089466095, 'train/grad_norm_pre_clip': 12.267312049865723, 'timing/data': 0.02920980518683791, 'timing/model': 1.5432186601683497, 'learning_rate/qwen_vl_interface': 2.559988691895155e-06, 'learning_rate/action_model': 2.559988691895155e-05, 'epoch': 1.41}) +[2026-06-16 07:39:42,250][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.028438720852136612, 'train/grad_norm_pre_clip': 15.763951301574707, 'timing/data': 0.03161604888737202, 'timing/model': 1.591175393667072, 'learning_rate/qwen_vl_interface': 2.5574792490678648e-06, 'learning_rate/action_model': 2.5574792490678644e-05, 'epoch': 1.41}) +[2026-06-16 07:39:48,657][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.005251738242805004, 'train/grad_norm_pre_clip': 1.9286259412765503, 'timing/data': 0.03256935393437743, 'timing/model': 1.5881256381981075, 'learning_rate/qwen_vl_interface': 2.5549709130341493e-06, 'learning_rate/action_model': 2.5549709130341493e-05, 'epoch': 1.41}) +[2026-06-16 07:39:55,042][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.0027064094319939613, 'train/grad_norm_pre_clip': 0.4946674108505249, 'timing/data': 0.03201559092849493, 'timing/model': 1.8192926649935544, 'learning_rate/qwen_vl_interface': 2.5524636848250924e-06, 'learning_rate/action_model': 2.552463684825092e-05, 'epoch': 1.41}) +[2026-06-16 07:40:01,183][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.1341821700334549, 'train/grad_norm_pre_clip': 13.064067840576172, 'timing/data': 0.03157193819060922, 'timing/model': 1.5613140827044845, 'learning_rate/qwen_vl_interface': 2.5499575654713186e-06, 'learning_rate/action_model': 2.5499575654713185e-05, 'epoch': 1.41}) +[2026-06-16 07:40:07,468][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.003266749670729041, 'train/grad_norm_pre_clip': 0.8918644785881042, 'timing/data': 0.03550875000655651, 'timing/model': 1.5430510328151286, 'learning_rate/qwen_vl_interface': 2.547452556003001e-06, 'learning_rate/action_model': 2.547452556003001e-05, 'epoch': 1.41}) +[2026-06-16 07:40:13,948][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.014118239283561707, 'train/grad_norm_pre_clip': 10.5523099899292, 'timing/data': 0.032375242095440626, 'timing/model': 1.6446955767460167, 'learning_rate/qwen_vl_interface': 2.5449486574498535e-06, 'learning_rate/action_model': 2.5449486574498534e-05, 'epoch': 1.41}) +[2026-06-16 07:40:20,273][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.058363258838653564, 'train/grad_norm_pre_clip': 11.298538208007812, 'timing/data': 0.03179551800712943, 'timing/model': 1.556605697143823, 'learning_rate/qwen_vl_interface': 2.5424458708411335e-06, 'learning_rate/action_model': 2.5424458708411335e-05, 'epoch': 1.41}) +[2026-06-16 07:40:26,668][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.0009739073575474322, 'train/grad_norm_pre_clip': 0.6937628984451294, 'timing/data': 0.03063212102279067, 'timing/model': 1.5982377431355417, 'learning_rate/qwen_vl_interface': 2.5399441972056437e-06, 'learning_rate/action_model': 2.5399441972056438e-05, 'epoch': 1.41}) +[2026-06-16 07:40:33,011][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.24100780487060547, 'train/grad_norm_pre_clip': 32.936954498291016, 'timing/data': 0.03064691089093685, 'timing/model': 1.7836270909756422, 'learning_rate/qwen_vl_interface': 2.537443637571726e-06, 'learning_rate/action_model': 2.537443637571726e-05, 'epoch': 1.41}) +[2026-06-16 07:40:39,223][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.004125025123357773, 'train/grad_norm_pre_clip': 0.6238450407981873, 'timing/data': 0.030411430168896914, 'timing/model': 1.6086977678351104, 'learning_rate/qwen_vl_interface': 2.534944192967269e-06, 'learning_rate/action_model': 2.534944192967269e-05, 'epoch': 1.41}) +[2026-06-16 07:51:42,924][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.18873298168182373, 'train/grad_norm_pre_clip': 10.831644058227539, 'eval/loss': 0.036059215664863586, 'eval/flap_precision': 0.9549668874172186, 'eval/flap_recall': 0.9836289222373806, 'eval/flap_f1': 0.9690860215053764, 'eval/flap_rate_model': 0.20972222222222223, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9690860215053764, 'timing/data': 0.03047025902196765, 'timing/model': 1.5484810280613601, 'learning_rate/qwen_vl_interface': 2.532445864419695e-06, 'learning_rate/action_model': 2.5324458644196952e-05, 'epoch': 1.42}) +[2026-06-16 07:51:42,925][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3500_state +[2026-06-16 07:51:42,931][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 07:51:51,575][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3500_state/model.safetensors +[2026-06-16 07:52:11,602][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3500_state/optimizer.bin +[2026-06-16 07:52:11,603][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3500_state/sampler.bin +[2026-06-16 07:52:11,603][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-16 07:52:11,604][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-16 07:52:11,792][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-16 07:52:11,793][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 07:52:11,807][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 07:52:17,727][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.1761859506368637, 'train/grad_norm_pre_clip': 29.740751266479492, 'timing/data': 0.02969309687614441, 'timing/model': 1.5468979398719966, 'learning_rate/qwen_vl_interface': 2.5299486529559758e-06, 'learning_rate/action_model': 2.5299486529559756e-05, 'epoch': 1.42}) +[2026-06-16 07:52:23,924][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.09338769316673279, 'train/grad_norm_pre_clip': 13.548046112060547, 'timing/data': 0.030104360077530146, 'timing/model': 1.5309914369136095, 'learning_rate/qwen_vl_interface': 2.5274525596026202e-06, 'learning_rate/action_model': 2.52745255960262e-05, 'epoch': 1.42}) +[2026-06-16 07:52:30,082][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.11770641803741455, 'train/grad_norm_pre_clip': 10.298908233642578, 'timing/data': 0.031338686123490334, 'timing/model': 1.587642283178866, 'learning_rate/qwen_vl_interface': 2.524957585385677e-06, 'learning_rate/action_model': 2.524957585385677e-05, 'epoch': 1.42}) +[2026-06-16 07:52:36,368][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.13759377598762512, 'train/grad_norm_pre_clip': 24.492530822753906, 'timing/data': 0.03564196592196822, 'timing/model': 1.5706313257105649, 'learning_rate/qwen_vl_interface': 2.522463731330738e-06, 'learning_rate/action_model': 2.5224637313307374e-05, 'epoch': 1.42}) +[2026-06-16 07:52:42,636][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.09204906970262527, 'train/grad_norm_pre_clip': 18.02696990966797, 'timing/data': 0.03117620525881648, 'timing/model': 1.497119047679007, 'learning_rate/qwen_vl_interface': 2.5199709984629294e-06, 'learning_rate/action_model': 2.5199709984629292e-05, 'epoch': 1.42}) +[2026-06-16 07:52:48,978][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.06382095813751221, 'train/grad_norm_pre_clip': 20.036880493164062, 'timing/data': 0.029688887298107147, 'timing/model': 1.7975188419222832, 'learning_rate/qwen_vl_interface': 2.517479387806922e-06, 'learning_rate/action_model': 2.517479387806922e-05, 'epoch': 1.42}) +[2026-06-16 07:52:55,301][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.023168757557868958, 'train/grad_norm_pre_clip': 9.304593086242676, 'timing/data': 0.03502753423526883, 'timing/model': 1.7853813068941236, 'learning_rate/qwen_vl_interface': 2.5149889003869217e-06, 'learning_rate/action_model': 2.5149889003869214e-05, 'epoch': 1.42}) +[2026-06-16 07:53:01,354][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.0030968894716352224, 'train/grad_norm_pre_clip': 0.8185445070266724, 'timing/data': 0.0303031699731946, 'timing/model': 1.5627096709795296, 'learning_rate/qwen_vl_interface': 2.5124995372266736e-06, 'learning_rate/action_model': 2.5124995372266735e-05, 'epoch': 1.42}) +[2026-06-16 07:53:07,660][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.023901810869574547, 'train/grad_norm_pre_clip': 11.019623756408691, 'timing/data': 0.03154622623696923, 'timing/model': 1.5556429331190884, 'learning_rate/qwen_vl_interface': 2.5100112993494626e-06, 'learning_rate/action_model': 2.5100112993494624e-05, 'epoch': 1.42}) +[2026-06-16 07:53:14,043][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.07724539190530777, 'train/grad_norm_pre_clip': 7.455533981323242, 'timing/data': 0.03059223061427474, 'timing/model': 1.5765353026799858, 'learning_rate/qwen_vl_interface': 2.5075241877781085e-06, 'learning_rate/action_model': 2.5075241877781086e-05, 'epoch': 1.42}) +[2026-06-16 07:53:20,340][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.07958860695362091, 'train/grad_norm_pre_clip': 7.800047874450684, 'timing/data': 0.03023583022877574, 'timing/model': 1.5226018517278135, 'learning_rate/qwen_vl_interface': 2.5050382035349673e-06, 'learning_rate/action_model': 2.5050382035349674e-05, 'epoch': 1.42}) +[2026-06-16 07:53:26,588][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.0007194423815235496, 'train/grad_norm_pre_clip': 0.15717504918575287, 'timing/data': 0.03119680378586054, 'timing/model': 1.777848708909005, 'learning_rate/qwen_vl_interface': 2.5025533476419343e-06, 'learning_rate/action_model': 2.5025533476419343e-05, 'epoch': 1.42}) +[2026-06-16 07:53:32,997][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.1782139390707016, 'train/grad_norm_pre_clip': 8.857597351074219, 'timing/data': 0.03389577800408006, 'timing/model': 1.8451083479449153, 'learning_rate/qwen_vl_interface': 2.5000696211204394e-06, 'learning_rate/action_model': 2.5000696211204394e-05, 'epoch': 1.42}) +[2026-06-16 07:53:39,063][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.029231341555714607, 'train/grad_norm_pre_clip': 17.76983070373535, 'timing/data': 0.03115177433937788, 'timing/model': 1.5562963392585516, 'learning_rate/qwen_vl_interface': 2.4975870249914495e-06, 'learning_rate/action_model': 2.4975870249914497e-05, 'epoch': 1.42}) +[2026-06-16 07:53:45,347][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.012009973637759686, 'train/grad_norm_pre_clip': 4.67092752456665, 'timing/data': 0.030363900121301413, 'timing/model': 1.5327176391147077, 'learning_rate/qwen_vl_interface': 2.4951055602754664e-06, 'learning_rate/action_model': 2.4951055602754665e-05, 'epoch': 1.42}) +[2026-06-16 07:53:51,552][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.038158029317855835, 'train/grad_norm_pre_clip': 11.973711967468262, 'timing/data': 0.030027518048882484, 'timing/model': 1.5345085049048066, 'learning_rate/qwen_vl_interface': 2.4926252279925255e-06, 'learning_rate/action_model': 2.4926252279925254e-05, 'epoch': 1.42}) +[2026-06-16 07:53:57,854][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.16226349771022797, 'train/grad_norm_pre_clip': 19.038185119628906, 'timing/data': 0.032164908945560455, 'timing/model': 1.5770142409019172, 'learning_rate/qwen_vl_interface': 2.490146029162199e-06, 'learning_rate/action_model': 2.4901460291621988e-05, 'epoch': 1.42}) +[2026-06-16 07:54:04,220][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.4280755817890167, 'train/grad_norm_pre_clip': 34.53841018676758, 'timing/data': 0.03179490612819791, 'timing/model': 1.7979358658194542, 'learning_rate/qwen_vl_interface': 2.4876679648035905e-06, 'learning_rate/action_model': 2.4876679648035903e-05, 'epoch': 1.42}) +[2026-06-16 07:54:10,404][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.00048361512017436326, 'train/grad_norm_pre_clip': 0.10904622077941895, 'timing/data': 0.03534221509471536, 'timing/model': 1.5856144363060594, 'learning_rate/qwen_vl_interface': 2.4851910359353392e-06, 'learning_rate/action_model': 2.485191035935339e-05, 'epoch': 1.42}) +[2026-06-16 07:54:16,696][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.023834476247429848, 'train/grad_norm_pre_clip': 18.54428482055664, 'timing/data': 0.03282537218183279, 'timing/model': 1.5654812538996339, 'learning_rate/qwen_vl_interface': 2.4827152435756185e-06, 'learning_rate/action_model': 2.4827152435756185e-05, 'epoch': 1.42}) +[2026-06-16 07:54:22,951][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.12595875561237335, 'train/grad_norm_pre_clip': 28.887929916381836, 'timing/data': 0.031335214618593454, 'timing/model': 1.534031176008284, 'learning_rate/qwen_vl_interface': 2.4802405887421307e-06, 'learning_rate/action_model': 2.4802405887421306e-05, 'epoch': 1.42}) +[2026-06-16 07:54:29,332][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.02895229123532772, 'train/grad_norm_pre_clip': 10.585905075073242, 'timing/data': 0.031046503223478794, 'timing/model': 1.5782448812387884, 'learning_rate/qwen_vl_interface': 2.4777670724521134e-06, 'learning_rate/action_model': 2.4777670724521134e-05, 'epoch': 1.42}) +[2026-06-16 07:54:35,561][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.0002784529933705926, 'train/grad_norm_pre_clip': 0.0524095855653286, 'timing/data': 0.03093382203951478, 'timing/model': 1.569188594352454, 'learning_rate/qwen_vl_interface': 2.475294695722337e-06, 'learning_rate/action_model': 2.475294695722337e-05, 'epoch': 1.42}) +[2026-06-16 07:54:41,953][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.01248998660594225, 'train/grad_norm_pre_clip': 2.0893890857696533, 'timing/data': 0.031719565857201815, 'timing/model': 1.7378160590305924, 'learning_rate/qwen_vl_interface': 2.4728234595691006e-06, 'learning_rate/action_model': 2.4728234595691005e-05, 'epoch': 1.43}) +[2026-06-16 07:54:48,129][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.03851202502846718, 'train/grad_norm_pre_clip': 15.501993179321289, 'timing/data': 0.03139359503984451, 'timing/model': 1.6136791151948273, 'learning_rate/qwen_vl_interface': 2.4703533650082368e-06, 'learning_rate/action_model': 2.4703533650082366e-05, 'epoch': 1.43}) +[2026-06-16 07:54:54,465][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.29045653343200684, 'train/grad_norm_pre_clip': 21.314382553100586, 'timing/data': 0.030756871215999126, 'timing/model': 1.5684544313699007, 'learning_rate/qwen_vl_interface': 2.4678844130551103e-06, 'learning_rate/action_model': 2.4678844130551103e-05, 'epoch': 1.43}) +[2026-06-16 07:55:00,765][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.0955405905842781, 'train/grad_norm_pre_clip': 10.884241104125977, 'timing/data': 0.031552935019135475, 'timing/model': 1.5398209877312183, 'learning_rate/qwen_vl_interface': 2.4654166047246102e-06, 'learning_rate/action_model': 2.46541660472461e-05, 'epoch': 1.43}) +[2026-06-16 07:55:07,269][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.007936663925647736, 'train/grad_norm_pre_clip': 3.1662840843200684, 'timing/data': 0.030753471422940493, 'timing/model': 1.7004981026984751, 'learning_rate/qwen_vl_interface': 2.4629499410311606e-06, 'learning_rate/action_model': 2.462949941031161e-05, 'epoch': 1.43}) +[2026-06-16 07:55:13,683][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.017298853024840355, 'train/grad_norm_pre_clip': 9.956557273864746, 'timing/data': 0.03122243331745267, 'timing/model': 1.6151362909004092, 'learning_rate/qwen_vl_interface': 2.460484422988715e-06, 'learning_rate/action_model': 2.4604844229887146e-05, 'epoch': 1.43}) +[2026-06-16 07:55:19,966][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.026719054207205772, 'train/grad_norm_pre_clip': 23.112211227416992, 'timing/data': 0.03084414079785347, 'timing/model': 1.7488728938624263, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 1.43}) +[2026-06-16 07:55:26,098][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.020148005336523056, 'train/grad_norm_pre_clip': 3.8401236534118652, 'timing/data': 0.031077782157808542, 'timing/model': 1.5766606871038675, 'learning_rate/qwen_vl_interface': 2.455556827910286e-06, 'learning_rate/action_model': 2.455556827910286e-05, 'epoch': 1.43}) +[2026-06-16 07:55:32,360][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.00026013332535512745, 'train/grad_norm_pre_clip': 0.03086475469172001, 'timing/data': 0.032489579170942307, 'timing/model': 1.5356913129799068, 'learning_rate/qwen_vl_interface': 2.453094752899853e-06, 'learning_rate/action_model': 2.453094752899853e-05, 'epoch': 1.43}) +[2026-06-16 07:55:38,661][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.2600104510784149, 'train/grad_norm_pre_clip': 7.894323348999023, 'timing/data': 0.03123872261494398, 'timing/model': 1.5370347183197737, 'learning_rate/qwen_vl_interface': 2.4506338275915173e-06, 'learning_rate/action_model': 2.4506338275915172e-05, 'epoch': 1.43}) +[2026-06-16 07:55:45,053][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.00903922226279974, 'train/grad_norm_pre_clip': 1.47601318359375, 'timing/data': 0.030493039172142744, 'timing/model': 1.5835026721470058, 'learning_rate/qwen_vl_interface': 2.4481740529968735e-06, 'learning_rate/action_model': 2.4481740529968734e-05, 'epoch': 1.43}) +[2026-06-16 07:55:51,403][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.014888049103319645, 'train/grad_norm_pre_clip': 4.339946746826172, 'timing/data': 0.03161719487980008, 'timing/model': 1.779558417852968, 'learning_rate/qwen_vl_interface': 2.4457154301270427e-06, 'learning_rate/action_model': 2.4457154301270427e-05, 'epoch': 1.43}) +[2026-06-16 07:55:57,462][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.18662390112876892, 'train/grad_norm_pre_clip': 12.892010688781738, 'timing/data': 0.0315497238188982, 'timing/model': 1.5354515723884106, 'learning_rate/qwen_vl_interface': 2.443257959992671e-06, 'learning_rate/action_model': 2.443257959992671e-05, 'epoch': 1.43}) +[2026-06-16 07:56:03,761][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.0750274807214737, 'train/grad_norm_pre_clip': 30.475425720214844, 'timing/data': 0.03183005517348647, 'timing/model': 1.5393509780988097, 'learning_rate/qwen_vl_interface': 2.440801643603933e-06, 'learning_rate/action_model': 2.440801643603933e-05, 'epoch': 1.43}) +[2026-06-16 07:56:10,139][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.15944674611091614, 'train/grad_norm_pre_clip': 30.487468719482422, 'timing/data': 0.030996460933238268, 'timing/model': 1.6193837272003293, 'learning_rate/qwen_vl_interface': 2.4383464819705252e-06, 'learning_rate/action_model': 2.438346481970525e-05, 'epoch': 1.43}) +[2026-06-16 07:56:16,464][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.013414888642728329, 'train/grad_norm_pre_clip': 2.399366855621338, 'timing/data': 0.03196932701393962, 'timing/model': 1.5548353581689298, 'learning_rate/qwen_vl_interface': 2.435892476101674e-06, 'learning_rate/action_model': 2.435892476101674e-05, 'epoch': 1.43}) +[2026-06-16 07:56:22,761][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.009873797185719013, 'train/grad_norm_pre_clip': 2.6872975826263428, 'timing/data': 0.03099064063280821, 'timing/model': 1.5375798419117928, 'learning_rate/qwen_vl_interface': 2.433439627006128e-06, 'learning_rate/action_model': 2.4334396270061277e-05, 'epoch': 1.43}) +[2026-06-16 07:56:29,115][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.02667897194623947, 'train/grad_norm_pre_clip': 24.656755447387695, 'timing/data': 0.031406912952661514, 'timing/model': 1.7868099357001483, 'learning_rate/qwen_vl_interface': 2.4309879356921602e-06, 'learning_rate/action_model': 2.43098793569216e-05, 'epoch': 1.43}) +[2026-06-16 07:56:35,169][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.07032763957977295, 'train/grad_norm_pre_clip': 11.620283126831055, 'timing/data': 0.031215331982821226, 'timing/model': 1.5475817960686982, 'learning_rate/qwen_vl_interface': 2.428537403167569e-06, 'learning_rate/action_model': 2.4285374031675687e-05, 'epoch': 1.43}) +[2026-06-16 07:56:41,479][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.03294096514582634, 'train/grad_norm_pre_clip': 14.9277982711792, 'timing/data': 0.030958910938352346, 'timing/model': 1.5473707527853549, 'learning_rate/qwen_vl_interface': 2.4260880304396765e-06, 'learning_rate/action_model': 2.4260880304396765e-05, 'epoch': 1.43}) +[2026-06-16 07:56:47,760][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.03214840218424797, 'train/grad_norm_pre_clip': 10.738280296325684, 'timing/data': 0.03179546492174268, 'timing/model': 1.53862579818815, 'learning_rate/qwen_vl_interface': 2.4236398185153245e-06, 'learning_rate/action_model': 2.4236398185153245e-05, 'epoch': 1.43}) +[2026-06-16 07:56:54,157][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.051784027367830276, 'train/grad_norm_pre_clip': 23.943883895874023, 'timing/data': 0.03217658679932356, 'timing/model': 1.5562276509590447, 'learning_rate/qwen_vl_interface': 2.4211927684008823e-06, 'learning_rate/action_model': 2.4211927684008824e-05, 'epoch': 1.43}) +[2026-06-16 07:57:00,466][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.197970911860466, 'train/grad_norm_pre_clip': 6.350866317749023, 'timing/data': 0.034604539163410664, 'timing/model': 1.543211195152253, 'learning_rate/qwen_vl_interface': 2.4187468811022398e-06, 'learning_rate/action_model': 2.4187468811022396e-05, 'epoch': 1.43}) +[2026-06-16 07:57:06,930][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.23785264790058136, 'train/grad_norm_pre_clip': 17.72218894958496, 'timing/data': 0.03207213617861271, 'timing/model': 1.8049246398732066, 'learning_rate/qwen_vl_interface': 2.416302157624809e-06, 'learning_rate/action_model': 2.4163021576248088e-05, 'epoch': 1.43}) +[2026-06-16 07:57:12,993][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.001993653131648898, 'train/grad_norm_pre_clip': 0.5980020761489868, 'timing/data': 0.03259764891117811, 'timing/model': 1.5585716958157718, 'learning_rate/qwen_vl_interface': 2.4138585989735227e-06, 'learning_rate/action_model': 2.4138585989735226e-05, 'epoch': 1.43}) +[2026-06-16 07:57:19,457][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.34152084589004517, 'train/grad_norm_pre_clip': 19.32076072692871, 'timing/data': 0.03400542587041855, 'timing/model': 1.625381238758564, 'learning_rate/qwen_vl_interface': 2.4114162061528357e-06, 'learning_rate/action_model': 2.4114162061528354e-05, 'epoch': 1.44}) +[2026-06-16 07:57:25,779][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.1411169320344925, 'train/grad_norm_pre_clip': 7.780299186706543, 'timing/data': 0.031737175304442644, 'timing/model': 1.5415983190760016, 'learning_rate/qwen_vl_interface': 2.408974980166724e-06, 'learning_rate/action_model': 2.408974980166724e-05, 'epoch': 1.44}) +[2026-06-16 07:57:32,178][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.20421288907527924, 'train/grad_norm_pre_clip': 25.97179412841797, 'timing/data': 0.03394665475934744, 'timing/model': 1.6229754728265107, 'learning_rate/qwen_vl_interface': 2.406534922018683e-06, 'learning_rate/action_model': 2.4065349220186827e-05, 'epoch': 1.44}) +[2026-06-16 07:57:38,530][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.13339602947235107, 'train/grad_norm_pre_clip': 31.51328468322754, 'timing/data': 0.03138020308688283, 'timing/model': 1.579503069166094, 'learning_rate/qwen_vl_interface': 2.4040960327117287e-06, 'learning_rate/action_model': 2.4040960327117283e-05, 'epoch': 1.44}) +[2026-06-16 07:57:44,991][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.0016469539841637015, 'train/grad_norm_pre_clip': 0.256305456161499, 'timing/data': 0.03192704590037465, 'timing/model': 1.8737673480063677, 'learning_rate/qwen_vl_interface': 2.4016583132483976e-06, 'learning_rate/action_model': 2.4016583132483976e-05, 'epoch': 1.44}) +[2026-06-16 07:57:51,066][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.004511302337050438, 'train/grad_norm_pre_clip': 0.6455716490745544, 'timing/data': 0.032261456828564405, 'timing/model': 1.538069193251431, 'learning_rate/qwen_vl_interface': 2.399221764630741e-06, 'learning_rate/action_model': 2.3992217646307406e-05, 'epoch': 1.44}) +[2026-06-16 07:57:57,360][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.004453086759895086, 'train/grad_norm_pre_clip': 1.2964195013046265, 'timing/data': 0.031865095254033804, 'timing/model': 1.5379777108319104, 'learning_rate/qwen_vl_interface': 2.3967863878603334e-06, 'learning_rate/action_model': 2.3967863878603332e-05, 'epoch': 1.44}) +[2026-06-16 07:58:03,667][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.18427903950214386, 'train/grad_norm_pre_clip': 22.023454666137695, 'timing/data': 0.031464092899113894, 'timing/model': 1.5443298402242362, 'learning_rate/qwen_vl_interface': 2.394352183938266e-06, 'learning_rate/action_model': 2.3943521839382664e-05, 'epoch': 1.44}) +[2026-06-16 07:58:09,963][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.06723795086145401, 'train/grad_norm_pre_clip': 20.65914535522461, 'timing/data': 0.031748744659125805, 'timing/model': 1.5429123602807522, 'learning_rate/qwen_vl_interface': 2.39191915386515e-06, 'learning_rate/action_model': 2.39191915386515e-05, 'epoch': 1.44}) +[2026-06-16 07:58:16,367][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.348711758852005, 'train/grad_norm_pre_clip': 21.294523239135742, 'timing/data': 0.03212917596101761, 'timing/model': 1.5819911290891469, 'learning_rate/qwen_vl_interface': 2.3894872986411095e-06, 'learning_rate/action_model': 2.3894872986411094e-05, 'epoch': 1.44}) +[2026-06-16 07:58:22,751][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.10385838150978088, 'train/grad_norm_pre_clip': 12.51671028137207, 'timing/data': 0.030792719218879938, 'timing/model': 1.8228203072212636, 'learning_rate/qwen_vl_interface': 2.3870566192657903e-06, 'learning_rate/action_model': 2.38705661926579e-05, 'epoch': 1.44}) +[2026-06-16 07:58:28,987][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.03622325509786606, 'train/grad_norm_pre_clip': 11.972244262695312, 'timing/data': 0.03222302719950676, 'timing/model': 1.5717677138745785, 'learning_rate/qwen_vl_interface': 2.3846271167383515e-06, 'learning_rate/action_model': 2.3846271167383515e-05, 'epoch': 1.44}) +[2026-06-16 07:58:35,348][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.1863965392112732, 'train/grad_norm_pre_clip': 7.5488996505737305, 'timing/data': 0.030429688282310963, 'timing/model': 1.5884219543077052, 'learning_rate/qwen_vl_interface': 2.382198792057471e-06, 'learning_rate/action_model': 2.3821987920574707e-05, 'epoch': 1.44}) +[2026-06-16 07:58:41,660][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.1649850755929947, 'train/grad_norm_pre_clip': 23.567100524902344, 'timing/data': 0.030088985804468393, 'timing/model': 1.519773688632995, 'learning_rate/qwen_vl_interface': 2.3797716462213402e-06, 'learning_rate/action_model': 2.3797716462213402e-05, 'epoch': 1.44}) +[2026-06-16 07:58:48,057][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.12861080467700958, 'train/grad_norm_pre_clip': 18.405487060546875, 'timing/data': 0.031633853912353516, 'timing/model': 1.637199536897242, 'learning_rate/qwen_vl_interface': 2.37734568022767e-06, 'learning_rate/action_model': 2.3773456802276697e-05, 'epoch': 1.44}) +[2026-06-16 07:58:54,449][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.17805595695972443, 'train/grad_norm_pre_clip': 14.169304847717285, 'timing/data': 0.03130523208528757, 'timing/model': 1.5860050460323691, 'learning_rate/qwen_vl_interface': 2.3749208950736783e-06, 'learning_rate/action_model': 2.374920895073678e-05, 'epoch': 1.44}) +[2026-06-16 07:59:00,895][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.03192847594618797, 'train/grad_norm_pre_clip': 15.574785232543945, 'timing/data': 0.031190301291644573, 'timing/model': 1.7808969100005925, 'learning_rate/qwen_vl_interface': 2.372497291756106e-06, 'learning_rate/action_model': 2.3724972917561058e-05, 'epoch': 1.44}) +[2026-06-16 07:59:06,967][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.2381467968225479, 'train/grad_norm_pre_clip': 15.122411727905273, 'timing/data': 0.01175520708784461, 'timing/model': 1.5558059755712748, 'learning_rate/qwen_vl_interface': 2.3700748712712043e-06, 'learning_rate/action_model': 2.370074871271204e-05, 'epoch': 1.44}) +[2026-06-16 07:59:13,262][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.02008603699505329, 'train/grad_norm_pre_clip': 4.260024547576904, 'timing/data': 0.02127573313191533, 'timing/model': 1.5510299312882125, 'learning_rate/qwen_vl_interface': 2.3676536346147388e-06, 'learning_rate/action_model': 2.3676536346147388e-05, 'epoch': 1.44}) +[2026-06-16 07:59:19,566][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.0010589759331196547, 'train/grad_norm_pre_clip': 0.11801441758871078, 'timing/data': 0.02002879697829485, 'timing/model': 1.5619862130843103, 'learning_rate/qwen_vl_interface': 2.365233582781988e-06, 'learning_rate/action_model': 2.365233582781988e-05, 'epoch': 1.44}) +[2026-06-16 07:59:25,856][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.007878663949668407, 'train/grad_norm_pre_clip': 1.0882762670516968, 'timing/data': 0.020595579873770475, 'timing/model': 1.5506763458251953, 'learning_rate/qwen_vl_interface': 2.3628147167677444e-06, 'learning_rate/action_model': 2.3628147167677443e-05, 'epoch': 1.44}) +[2026-06-16 07:59:32,159][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.052181489765644073, 'train/grad_norm_pre_clip': 20.91486167907715, 'timing/data': 0.02044603880494833, 'timing/model': 1.5182559373788536, 'learning_rate/qwen_vl_interface': 2.3603970375663123e-06, 'learning_rate/action_model': 2.3603970375663125e-05, 'epoch': 1.44}) +[2026-06-16 07:59:38,493][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.04103587940335274, 'train/grad_norm_pre_clip': 12.84164047241211, 'timing/data': 0.02089616097509861, 'timing/model': 1.791072517633438, 'learning_rate/qwen_vl_interface': 2.35798054617151e-06, 'learning_rate/action_model': 2.3579805461715096e-05, 'epoch': 1.44}) +[2026-06-16 07:59:44,654][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.007871091365814209, 'train/grad_norm_pre_clip': 2.476287841796875, 'timing/data': 0.020601768977940083, 'timing/model': 1.5551822930574417, 'learning_rate/qwen_vl_interface': 2.355565243576662e-06, 'learning_rate/action_model': 2.3555652435766617e-05, 'epoch': 1.44}) +[2026-06-16 07:59:51,058][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.003199611324816942, 'train/grad_norm_pre_clip': 0.5276113748550415, 'timing/data': 0.020846500992774963, 'timing/model': 1.547857805620879, 'learning_rate/qwen_vl_interface': 2.35315113077461e-06, 'learning_rate/action_model': 2.3531511307746098e-05, 'epoch': 1.44}) +[2026-06-16 07:59:57,415][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.0015737154753878713, 'train/grad_norm_pre_clip': 0.235401913523674, 'timing/data': 0.021828855387866497, 'timing/model': 1.6064097080379725, 'learning_rate/qwen_vl_interface': 2.3507382087577047e-06, 'learning_rate/action_model': 2.350738208757705e-05, 'epoch': 1.45}) +[2026-06-16 08:00:03,669][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.003087294055148959, 'train/grad_norm_pre_clip': 0.4732426404953003, 'timing/data': 0.022068137302994728, 'timing/model': 1.5504977852106094, 'learning_rate/qwen_vl_interface': 2.348326478517807e-06, 'learning_rate/action_model': 2.3483264785178072e-05, 'epoch': 1.45}) +[2026-06-16 08:00:10,051][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.24575671553611755, 'train/grad_norm_pre_clip': 11.552430152893066, 'timing/data': 0.02042686939239502, 'timing/model': 1.5909717488102615, 'learning_rate/qwen_vl_interface': 2.3459159410462907e-06, 'learning_rate/action_model': 2.3459159410462905e-05, 'epoch': 1.45}) +[2026-06-16 08:00:16,520][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.06380978226661682, 'train/grad_norm_pre_clip': 17.523338317871094, 'timing/data': 0.022455927915871143, 'timing/model': 1.870025489013642, 'learning_rate/qwen_vl_interface': 2.343506597334031e-06, 'learning_rate/action_model': 2.343506597334031e-05, 'epoch': 1.45}) +[2026-06-16 08:00:22,732][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.04575665295124054, 'train/grad_norm_pre_clip': 10.081470489501953, 'timing/data': 0.020655480213463306, 'timing/model': 1.6874872441403568, 'learning_rate/qwen_vl_interface': 2.341098448371423e-06, 'learning_rate/action_model': 2.3410984483714227e-05, 'epoch': 1.45}) +[2026-06-16 08:00:28,768][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.0010983485262840986, 'train/grad_norm_pre_clip': 0.11899526417255402, 'timing/data': 0.021195882000029087, 'timing/model': 1.5560572962276638, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 1.45}) +[2026-06-16 08:00:35,061][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.04945699870586395, 'train/grad_norm_pre_clip': 15.435714721679688, 'timing/data': 0.02073189988732338, 'timing/model': 1.554761027917266, 'learning_rate/qwen_vl_interface': 2.33628573865426e-06, 'learning_rate/action_model': 2.3362857386542596e-05, 'epoch': 1.45}) +[2026-06-16 08:00:41,471][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.083389513194561, 'train/grad_norm_pre_clip': 7.246001720428467, 'timing/data': 0.021029142197221518, 'timing/model': 1.576035588979721, 'learning_rate/qwen_vl_interface': 2.3338811798780282e-06, 'learning_rate/action_model': 2.333881179878028e-05, 'epoch': 1.45}) +[2026-06-16 08:00:47,864][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.03235946223139763, 'train/grad_norm_pre_clip': 25.565242767333984, 'timing/data': 0.021903115790337324, 'timing/model': 1.6124549540691078, 'learning_rate/qwen_vl_interface': 2.3314778198080933e-06, 'learning_rate/action_model': 2.331477819808093e-05, 'epoch': 1.45}) +[2026-06-16 08:00:54,229][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.03338566794991493, 'train/grad_norm_pre_clip': 17.51727294921875, 'timing/data': 0.020207997877150774, 'timing/model': 1.817264531739056, 'learning_rate/qwen_vl_interface': 2.3290756594323834e-06, 'learning_rate/action_model': 2.3290756594323836e-05, 'epoch': 1.45}) +[2026-06-16 08:01:00,599][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.007986238226294518, 'train/grad_norm_pre_clip': 2.3492395877838135, 'timing/data': 0.022750719916075468, 'timing/model': 1.7872948558069766, 'learning_rate/qwen_vl_interface': 2.3266746997383378e-06, 'learning_rate/action_model': 2.3266746997383377e-05, 'epoch': 1.45}) +[2026-06-16 08:01:06,674][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.25867828726768494, 'train/grad_norm_pre_clip': 16.30769920349121, 'timing/data': 0.020312519278377295, 'timing/model': 1.5613179723732173, 'learning_rate/qwen_vl_interface': 2.324274941712899e-06, 'learning_rate/action_model': 2.324274941712899e-05, 'epoch': 1.45}) +[2026-06-16 08:01:12,992][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.05434897541999817, 'train/grad_norm_pre_clip': 8.835714340209961, 'timing/data': 0.021467714104801416, 'timing/model': 1.5883119022473693, 'learning_rate/qwen_vl_interface': 2.321876386342519e-06, 'learning_rate/action_model': 2.3218763863425187e-05, 'epoch': 1.45}) +[2026-06-16 08:01:19,267][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.011867434717714787, 'train/grad_norm_pre_clip': 5.609001636505127, 'timing/data': 0.02067067986354232, 'timing/model': 1.5433013732545078, 'learning_rate/qwen_vl_interface': 2.31947903461315e-06, 'learning_rate/action_model': 2.3194790346131502e-05, 'epoch': 1.45}) +[2026-06-16 08:01:25,658][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.04017408564686775, 'train/grad_norm_pre_clip': 9.504735946655273, 'timing/data': 0.0203269780613482, 'timing/model': 1.5359569657593966, 'learning_rate/qwen_vl_interface': 2.3170828875102544e-06, 'learning_rate/action_model': 2.3170828875102543e-05, 'epoch': 1.45}) +[2026-06-16 08:01:32,106][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.003930530045181513, 'train/grad_norm_pre_clip': 0.7273927927017212, 'timing/data': 0.023933484684675932, 'timing/model': 1.8350059390068054, 'learning_rate/qwen_vl_interface': 2.3146879460187984e-06, 'learning_rate/action_model': 2.3146879460187982e-05, 'epoch': 1.45}) +[2026-06-16 08:01:38,168][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.013866259716451168, 'train/grad_norm_pre_clip': 5.04706335067749, 'timing/data': 0.02442538784816861, 'timing/model': 1.5593236959539354, 'learning_rate/qwen_vl_interface': 2.3122942111232506e-06, 'learning_rate/action_model': 2.3122942111232506e-05, 'epoch': 1.45}) +[2026-06-16 08:01:44,466][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.0025693841744214296, 'train/grad_norm_pre_clip': 0.3894951045513153, 'timing/data': 0.022088687401264906, 'timing/model': 1.5507858931086957, 'learning_rate/qwen_vl_interface': 2.3099016838075857e-06, 'learning_rate/action_model': 2.3099016838075856e-05, 'epoch': 1.45}) +[2026-06-16 08:01:50,861][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.0032855477184057236, 'train/grad_norm_pre_clip': 0.6626084446907043, 'timing/data': 0.020620359107851982, 'timing/model': 1.5648946412838995, 'learning_rate/qwen_vl_interface': 2.307510365055282e-06, 'learning_rate/action_model': 2.307510365055282e-05, 'epoch': 1.45}) +[2026-06-16 08:01:57,242][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.006399376317858696, 'train/grad_norm_pre_clip': 0.7054566740989685, 'timing/data': 0.020710680168122053, 'timing/model': 1.5897178589366376, 'learning_rate/qwen_vl_interface': 2.3051202558493196e-06, 'learning_rate/action_model': 2.3051202558493192e-05, 'epoch': 1.45}) +[2026-06-16 08:02:03,456][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.0027612170670181513, 'train/grad_norm_pre_clip': 0.681898832321167, 'timing/data': 0.020600079093128443, 'timing/model': 1.5724729038774967, 'learning_rate/qwen_vl_interface': 2.302731357172182e-06, 'learning_rate/action_model': 2.302731357172182e-05, 'epoch': 1.45}) +[2026-06-16 08:02:09,745][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.12183187156915665, 'train/grad_norm_pre_clip': 13.724200248718262, 'timing/data': 0.020167937967926264, 'timing/model': 1.736835225019604, 'learning_rate/qwen_vl_interface': 2.3003436700058556e-06, 'learning_rate/action_model': 2.3003436700058554e-05, 'epoch': 1.45}) +[2026-06-16 08:02:15,834][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.028865521773695946, 'train/grad_norm_pre_clip': 9.585021018981934, 'timing/data': 0.020444459281861782, 'timing/model': 1.6184413130395114, 'learning_rate/qwen_vl_interface': 2.297957195331828e-06, 'learning_rate/action_model': 2.2979571953318276e-05, 'epoch': 1.45}) +[2026-06-16 08:02:22,064][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.030702276155352592, 'train/grad_norm_pre_clip': 17.802452087402344, 'timing/data': 0.021984036080539227, 'timing/model': 1.5494593898765743, 'learning_rate/qwen_vl_interface': 2.2955719341310903e-06, 'learning_rate/action_model': 2.2955719341310903e-05, 'epoch': 1.45}) +[2026-06-16 08:02:28,662][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.0014435696648433805, 'train/grad_norm_pre_clip': 0.25235989689826965, 'timing/data': 0.025110931135714054, 'timing/model': 1.5594588359817863, 'learning_rate/qwen_vl_interface': 2.2931878873841304e-06, 'learning_rate/action_model': 2.2931878873841302e-05, 'epoch': 1.45}) +[2026-06-16 08:02:34,973][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.024005023762583733, 'train/grad_norm_pre_clip': 10.901651382446289, 'timing/data': 0.021340892184525728, 'timing/model': 1.5239507546648383, 'learning_rate/qwen_vl_interface': 2.290805056070942e-06, 'learning_rate/action_model': 2.290805056070942e-05, 'epoch': 1.46}) +[2026-06-16 08:13:38,916][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.0030087244231253862, 'train/grad_norm_pre_clip': 0.43448567390441895, 'eval/loss': 0.031117873266339302, 'eval/flap_precision': 0.9664429530201343, 'eval/flap_recall': 0.9822646657571623, 'eval/flap_f1': 0.9742895805142083, 'eval/flap_rate_model': 0.20694444444444443, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9742895805142082, 'timing/data': 0.02143531385809183, 'timing/model': 1.589961270801723, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 1.46}) +[2026-06-16 08:13:38,917][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3600_state +[2026-06-16 08:13:38,923][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 08:13:47,646][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3600_state/model.safetensors +[2026-06-16 08:14:08,351][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3600_state/optimizer.bin +[2026-06-16 08:14:08,352][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3600_state/sampler.bin +[2026-06-16 08:14:08,352][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3600_state/sampler_1.bin +[2026-06-16 08:14:08,353][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3600_state/random_states_0.pkl +[2026-06-16 08:14:08,546][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3600 +[2026-06-16 08:14:08,547][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 08:14:08,562][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 08:14:14,829][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.0018470261711627245, 'train/grad_norm_pre_clip': 0.26105713844299316, 'timing/data': 0.02464281115680933, 'timing/model': 1.5693275732919574, 'learning_rate/qwen_vl_interface': 2.2860430436633447e-06, 'learning_rate/action_model': 2.2860430436633447e-05, 'epoch': 1.46}) +[2026-06-16 08:14:21,066][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.006604163441807032, 'train/grad_norm_pre_clip': 1.4043041467666626, 'timing/data': 0.02154242806136608, 'timing/model': 1.7935804822482169, 'learning_rate/qwen_vl_interface': 2.28366386452642e-06, 'learning_rate/action_model': 2.28366386452642e-05, 'epoch': 1.46}) +[2026-06-16 08:14:27,351][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.0631614625453949, 'train/grad_norm_pre_clip': 20.280319213867188, 'timing/data': 0.019944800063967705, 'timing/model': 1.6584480931051075, 'learning_rate/qwen_vl_interface': 2.2812859047382326e-06, 'learning_rate/action_model': 2.2812859047382323e-05, 'epoch': 1.46}) +[2026-06-16 08:14:33,345][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.0004356247081886977, 'train/grad_norm_pre_clip': 0.10382206737995148, 'timing/data': 0.021599698346108198, 'timing/model': 1.5267893080599606, 'learning_rate/qwen_vl_interface': 2.2789091652762712e-06, 'learning_rate/action_model': 2.2789091652762712e-05, 'epoch': 1.46}) +[2026-06-16 08:14:39,642][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.005236996337771416, 'train/grad_norm_pre_clip': 2.283601999282837, 'timing/data': 0.02202027989551425, 'timing/model': 1.5583053547888994, 'learning_rate/qwen_vl_interface': 2.276533647117524e-06, 'learning_rate/action_model': 2.2765336471175242e-05, 'epoch': 1.46}) +[2026-06-16 08:14:45,946][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.1100846379995346, 'train/grad_norm_pre_clip': 16.656740188598633, 'timing/data': 0.022079110145568848, 'timing/model': 1.557113183196634, 'learning_rate/qwen_vl_interface': 2.2741593512384773e-06, 'learning_rate/action_model': 2.2741593512384772e-05, 'epoch': 1.46}) +[2026-06-16 08:14:52,150][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.006963209714740515, 'train/grad_norm_pre_clip': 4.514504909515381, 'timing/data': 0.020130150951445103, 'timing/model': 1.5271431696601212, 'learning_rate/qwen_vl_interface': 2.2717862786151153e-06, 'learning_rate/action_model': 2.271786278615115e-05, 'epoch': 1.46}) +[2026-06-16 08:14:58,435][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.06377556920051575, 'train/grad_norm_pre_clip': 15.636587142944336, 'timing/data': 0.02534914528951049, 'timing/model': 1.5771107892505825, 'learning_rate/qwen_vl_interface': 2.2694144302229152e-06, 'learning_rate/action_model': 2.2694144302229152e-05, 'epoch': 1.46}) +[2026-06-16 08:15:04,669][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.1281650811433792, 'train/grad_norm_pre_clip': 24.2887020111084, 'timing/data': 0.021852239966392517, 'timing/model': 1.7723164199851453, 'learning_rate/qwen_vl_interface': 2.2670438070368568e-06, 'learning_rate/action_model': 2.2670438070368566e-05, 'epoch': 1.46}) +[2026-06-16 08:15:10,749][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.0524735301733017, 'train/grad_norm_pre_clip': 13.436877250671387, 'timing/data': 0.021203855983912945, 'timing/model': 1.5502980728633702, 'learning_rate/qwen_vl_interface': 2.264674410031413e-06, 'learning_rate/action_model': 2.2646744100314127e-05, 'epoch': 1.46}) +[2026-06-16 08:15:17,046][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.13892090320587158, 'train/grad_norm_pre_clip': 16.720054626464844, 'timing/data': 0.02139071701094508, 'timing/model': 1.5198555765673518, 'learning_rate/qwen_vl_interface': 2.2623062401805537e-06, 'learning_rate/action_model': 2.2623062401805535e-05, 'epoch': 1.46}) +[2026-06-16 08:15:23,344][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.03256285563111305, 'train/grad_norm_pre_clip': 9.053410530090332, 'timing/data': 0.020519752986729145, 'timing/model': 1.6438563652336597, 'learning_rate/qwen_vl_interface': 2.2599392984577436e-06, 'learning_rate/action_model': 2.2599392984577438e-05, 'epoch': 1.46}) +[2026-06-16 08:15:29,555][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.0033092547673732042, 'train/grad_norm_pre_clip': 0.7660471796989441, 'timing/data': 0.021364507265388966, 'timing/model': 1.5194715410470963, 'learning_rate/qwen_vl_interface': 2.2575735858359436e-06, 'learning_rate/action_model': 2.257573585835944e-05, 'epoch': 1.46}) +[2026-06-16 08:15:35,892][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.00019003773923031986, 'train/grad_norm_pre_clip': 0.01988718844950199, 'timing/data': 0.020919545087963343, 'timing/model': 1.7891976730898023, 'learning_rate/qwen_vl_interface': 2.255209103287609e-06, 'learning_rate/action_model': 2.255209103287609e-05, 'epoch': 1.46}) +[2026-06-16 08:15:42,404][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.12539149820804596, 'train/grad_norm_pre_clip': 9.240779876708984, 'timing/data': 0.021085965912789106, 'timing/model': 1.9108147011138499, 'learning_rate/qwen_vl_interface': 2.2528458517846886e-06, 'learning_rate/action_model': 2.252845851784689e-05, 'epoch': 1.46}) +[2026-06-16 08:15:48,469][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.03037651628255844, 'train/grad_norm_pre_clip': 15.595071792602539, 'timing/data': 0.02025304175913334, 'timing/model': 1.5308850910514593, 'learning_rate/qwen_vl_interface': 2.250483832298627e-06, 'learning_rate/action_model': 2.250483832298627e-05, 'epoch': 1.46}) +[2026-06-16 08:15:54,853][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.1051456406712532, 'train/grad_norm_pre_clip': 13.319954872131348, 'timing/data': 0.021109815686941147, 'timing/model': 1.6478500771336257, 'learning_rate/qwen_vl_interface': 2.2481230458003608e-06, 'learning_rate/action_model': 2.2481230458003606e-05, 'epoch': 1.46}) +[2026-06-16 08:16:01,051][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.002708328189328313, 'train/grad_norm_pre_clip': 0.4389468729496002, 'timing/data': 0.02016417169943452, 'timing/model': 1.5508882571011782, 'learning_rate/qwen_vl_interface': 2.2457634932603184e-06, 'learning_rate/action_model': 2.2457634932603183e-05, 'epoch': 1.46}) +[2026-06-16 08:16:07,883][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.0037841869052499533, 'train/grad_norm_pre_clip': 1.060727834701538, 'timing/data': 0.029468134976923466, 'timing/model': 1.531998774036765, 'learning_rate/qwen_vl_interface': 2.2434051756484255e-06, 'learning_rate/action_model': 2.2434051756484253e-05, 'epoch': 1.46}) +[2026-06-16 08:16:14,260][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.0017665476771071553, 'train/grad_norm_pre_clip': 0.3577192425727844, 'timing/data': 0.020525552798062563, 'timing/model': 1.846442462876439, 'learning_rate/qwen_vl_interface': 2.2410480939340945e-06, 'learning_rate/action_model': 2.2410480939340945e-05, 'epoch': 1.46}) +[2026-06-16 08:16:20,593][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.009916896000504494, 'train/grad_norm_pre_clip': 3.604417562484741, 'timing/data': 0.02102688606828451, 'timing/model': 1.710973967332393, 'learning_rate/qwen_vl_interface': 2.238692249086234e-06, 'learning_rate/action_model': 2.2386922490862336e-05, 'epoch': 1.46}) +[2026-06-16 08:16:26,689][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.2651086151599884, 'train/grad_norm_pre_clip': 23.776649475097656, 'timing/data': 0.02081880532205105, 'timing/model': 1.5822453913278878, 'learning_rate/qwen_vl_interface': 2.2363376420732426e-06, 'learning_rate/action_model': 2.2363376420732423e-05, 'epoch': 1.46}) +[2026-06-16 08:16:32,972][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.0033436573576182127, 'train/grad_norm_pre_clip': 2.0009841918945312, 'timing/data': 0.021174476947635412, 'timing/model': 1.5672849551774561, 'learning_rate/qwen_vl_interface': 2.233984273863011e-06, 'learning_rate/action_model': 2.233984273863011e-05, 'epoch': 1.47}) +[2026-06-16 08:16:39,768][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.01737581193447113, 'train/grad_norm_pre_clip': 7.59700870513916, 'timing/data': 0.02222266187891364, 'timing/model': 1.551946667023003, 'learning_rate/qwen_vl_interface': 2.231632145422919e-06, 'learning_rate/action_model': 2.231632145422919e-05, 'epoch': 1.47}) +[2026-06-16 08:16:46,276][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.0003999182372353971, 'train/grad_norm_pre_clip': 0.06533485651016235, 'timing/data': 0.021435577888041735, 'timing/model': 1.7078358931466937, 'learning_rate/qwen_vl_interface': 2.2292812577198383e-06, 'learning_rate/action_model': 2.229281257719838e-05, 'epoch': 1.47}) +[2026-06-16 08:16:52,639][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.0006866516778245568, 'train/grad_norm_pre_clip': 0.2288265973329544, 'timing/data': 0.021625428926199675, 'timing/model': 1.8098488529212773, 'learning_rate/qwen_vl_interface': 2.2269316117201307e-06, 'learning_rate/action_model': 2.2269316117201307e-05, 'epoch': 1.47}) +[2026-06-16 08:16:58,964][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.07423313707113266, 'train/grad_norm_pre_clip': 12.617341041564941, 'timing/data': 0.021921990904957056, 'timing/model': 1.7937311511486769, 'learning_rate/qwen_vl_interface': 2.224583208389646e-06, 'learning_rate/action_model': 2.224583208389646e-05, 'epoch': 1.47}) +[2026-06-16 08:17:05,075][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.00679442286491394, 'train/grad_norm_pre_clip': 2.575817584991455, 'timing/data': 0.022277622949332, 'timing/model': 1.5451922761276364, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 1.47}) +[2026-06-16 08:17:11,484][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.00046911154640838504, 'train/grad_norm_pre_clip': 0.06748367100954056, 'timing/data': 0.021877790801227093, 'timing/model': 1.5656921518966556, 'learning_rate/qwen_vl_interface': 2.219890133597194e-06, 'learning_rate/action_model': 2.219890133597194e-05, 'epoch': 1.47}) +[2026-06-16 08:17:17,775][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.04875088110566139, 'train/grad_norm_pre_clip': 21.306682586669922, 'timing/data': 0.020796475932002068, 'timing/model': 1.5511105973273516, 'learning_rate/qwen_vl_interface': 2.217545464064374e-06, 'learning_rate/action_model': 2.217545464064374e-05, 'epoch': 1.47}) +[2026-06-16 08:17:24,076][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.022923413664102554, 'train/grad_norm_pre_clip': 18.128101348876953, 'timing/data': 0.021771919913589954, 'timing/model': 1.5205305498093367, 'learning_rate/qwen_vl_interface': 2.215202041059067e-06, 'learning_rate/action_model': 2.215202041059067e-05, 'epoch': 1.47}) +[2026-06-16 08:17:30,417][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.0148676298558712, 'train/grad_norm_pre_clip': 10.986209869384766, 'timing/data': 0.021626808680593967, 'timing/model': 1.8056495087221265, 'learning_rate/qwen_vl_interface': 2.212859865544566e-06, 'learning_rate/action_model': 2.212859865544566e-05, 'epoch': 1.47}) +[2026-06-16 08:17:36,470][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.015290303155779839, 'train/grad_norm_pre_clip': 10.478318214416504, 'timing/data': 0.021683999337255955, 'timing/model': 1.5560192572884262, 'learning_rate/qwen_vl_interface': 2.210518938483652e-06, 'learning_rate/action_model': 2.210518938483652e-05, 'epoch': 1.47}) +[2026-06-16 08:17:42,857][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.18998757004737854, 'train/grad_norm_pre_clip': 12.802574157714844, 'timing/data': 0.022150611970573664, 'timing/model': 1.5377032146789134, 'learning_rate/qwen_vl_interface': 2.20817926083859e-06, 'learning_rate/action_model': 2.20817926083859e-05, 'epoch': 1.47}) +[2026-06-16 08:17:49,273][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.2212822586297989, 'train/grad_norm_pre_clip': 30.011234283447266, 'timing/data': 0.019717200193554163, 'timing/model': 1.5649472624063492, 'learning_rate/qwen_vl_interface': 2.205840833571135e-06, 'learning_rate/action_model': 2.205840833571135e-05, 'epoch': 1.47}) +[2026-06-16 08:17:55,664][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.11400630325078964, 'train/grad_norm_pre_clip': 25.364728927612305, 'timing/data': 0.021028506103903055, 'timing/model': 1.5537431431002915, 'learning_rate/qwen_vl_interface': 2.2035036576425264e-06, 'learning_rate/action_model': 2.203503657642526e-05, 'epoch': 1.47}) +[2026-06-16 08:18:02,069][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.007327674888074398, 'train/grad_norm_pre_clip': 7.368442535400391, 'timing/data': 0.023600107990205288, 'timing/model': 1.5984299900010228, 'learning_rate/qwen_vl_interface': 2.2011677340134867e-06, 'learning_rate/action_model': 2.2011677340134867e-05, 'epoch': 1.47}) +[2026-06-16 08:18:08,394][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.13235501945018768, 'train/grad_norm_pre_clip': 12.502628326416016, 'timing/data': 0.020889736246317625, 'timing/model': 1.7824771371670067, 'learning_rate/qwen_vl_interface': 2.198833063644227e-06, 'learning_rate/action_model': 2.198833063644227e-05, 'epoch': 1.47}) +[2026-06-16 08:18:14,543][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.08873716741800308, 'train/grad_norm_pre_clip': 14.96947956085205, 'timing/data': 0.02033425262197852, 'timing/model': 1.637565053999424, 'learning_rate/qwen_vl_interface': 2.196499647494443e-06, 'learning_rate/action_model': 2.196499647494443e-05, 'epoch': 1.47}) +[2026-06-16 08:18:20,765][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.04951375722885132, 'train/grad_norm_pre_clip': 14.736591339111328, 'timing/data': 0.02084351610392332, 'timing/model': 1.5166660598479211, 'learning_rate/qwen_vl_interface': 2.1941674865233144e-06, 'learning_rate/action_model': 2.1941674865233143e-05, 'epoch': 1.47}) +[2026-06-16 08:18:27,215][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.05834323167800903, 'train/grad_norm_pre_clip': 20.446352005004883, 'timing/data': 0.02129605831578374, 'timing/model': 1.6177678969688714, 'learning_rate/qwen_vl_interface': 2.1918365816895013e-06, 'learning_rate/action_model': 2.191836581689501e-05, 'epoch': 1.47}) +[2026-06-16 08:18:33,561][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.005843556020408869, 'train/grad_norm_pre_clip': 2.7261505126953125, 'timing/data': 0.02003940287977457, 'timing/model': 1.5735722263343632, 'learning_rate/qwen_vl_interface': 2.189506933951153e-06, 'learning_rate/action_model': 2.1895069339511525e-05, 'epoch': 1.47}) +[2026-06-16 08:18:39,958][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.0013227193849161267, 'train/grad_norm_pre_clip': 0.4313197731971741, 'timing/data': 0.02175976987928152, 'timing/model': 1.5923810442909598, 'learning_rate/qwen_vl_interface': 2.187178544265898e-06, 'learning_rate/action_model': 2.1871785442658983e-05, 'epoch': 1.47}) +[2026-06-16 08:18:46,433][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.17219598591327667, 'train/grad_norm_pre_clip': 16.361740112304688, 'timing/data': 0.019857872277498245, 'timing/model': 1.8487171959131956, 'learning_rate/qwen_vl_interface': 2.184851413590852e-06, 'learning_rate/action_model': 2.1848514135908518e-05, 'epoch': 1.47}) +[2026-06-16 08:18:52,699][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.015076640993356705, 'train/grad_norm_pre_clip': 9.79019832611084, 'timing/data': 0.020938116125762463, 'timing/model': 1.79109588265419, 'learning_rate/qwen_vl_interface': 2.1825255428826087e-06, 'learning_rate/action_model': 2.182525542882609e-05, 'epoch': 1.47}) +[2026-06-16 08:18:58,861][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.00028389113140292466, 'train/grad_norm_pre_clip': 0.041778601706027985, 'timing/data': 0.02146188821643591, 'timing/model': 1.5544951558113098, 'learning_rate/qwen_vl_interface': 2.180200933097246e-06, 'learning_rate/action_model': 2.180200933097246e-05, 'epoch': 1.47}) +[2026-06-16 08:19:05,257][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.0674048587679863, 'train/grad_norm_pre_clip': 19.79486846923828, 'timing/data': 0.020272753201425076, 'timing/model': 1.5496212551370263, 'learning_rate/qwen_vl_interface': 2.1778775851903235e-06, 'learning_rate/action_model': 2.177877585190323e-05, 'epoch': 1.47}) +[2026-06-16 08:19:11,663][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.12330246716737747, 'train/grad_norm_pre_clip': 17.709354400634766, 'timing/data': 0.02080398565158248, 'timing/model': 1.5483682909980416, 'learning_rate/qwen_vl_interface': 2.175555500116881e-06, 'learning_rate/action_model': 2.175555500116881e-05, 'epoch': 1.48}) +[2026-06-16 08:19:17,958][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.015778593719005585, 'train/grad_norm_pre_clip': 7.394460201263428, 'timing/data': 0.021347337868064642, 'timing/model': 1.5156993228010833, 'learning_rate/qwen_vl_interface': 2.1732346788314404e-06, 'learning_rate/action_model': 2.1732346788314402e-05, 'epoch': 1.48}) +[2026-06-16 08:19:24,319][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.0024199485778808594, 'train/grad_norm_pre_clip': 1.0851305723190308, 'timing/data': 0.021204428281635046, 'timing/model': 1.8007997977547348, 'learning_rate/qwen_vl_interface': 2.170915122288006e-06, 'learning_rate/action_model': 2.170915122288006e-05, 'epoch': 1.48}) +[2026-06-16 08:19:30,698][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.14720585942268372, 'train/grad_norm_pre_clip': 12.776908874511719, 'timing/data': 0.020624535158276558, 'timing/model': 1.8071473389863968, 'learning_rate/qwen_vl_interface': 2.168596831440055e-06, 'learning_rate/action_model': 2.1685968314400552e-05, 'epoch': 1.48}) +[2026-06-16 08:19:36,956][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.019786356016993523, 'train/grad_norm_pre_clip': 6.493609428405762, 'timing/data': 0.020698274951428175, 'timing/model': 1.5474690338596702, 'learning_rate/qwen_vl_interface': 2.166279807240553e-06, 'learning_rate/action_model': 2.1662798072405528e-05, 'epoch': 1.48}) +[2026-06-16 08:19:43,456][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.02775179222226143, 'train/grad_norm_pre_clip': 13.626057624816895, 'timing/data': 0.021505328826606274, 'timing/model': 1.5616679219529033, 'learning_rate/qwen_vl_interface': 2.1639640506419394e-06, 'learning_rate/action_model': 2.163964050641939e-05, 'epoch': 1.48}) +[2026-06-16 08:19:49,880][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.0007068489212542772, 'train/grad_norm_pre_clip': 0.1199306845664978, 'timing/data': 0.02192460186779499, 'timing/model': 1.5840055458247662, 'learning_rate/qwen_vl_interface': 2.161649562596135e-06, 'learning_rate/action_model': 2.161649562596135e-05, 'epoch': 1.48}) +[2026-06-16 08:19:56,171][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.0004583988047670573, 'train/grad_norm_pre_clip': 0.061373960226774216, 'timing/data': 0.021454199217259884, 'timing/model': 1.5206605158746243, 'learning_rate/qwen_vl_interface': 2.1593363440545383e-06, 'learning_rate/action_model': 2.1593363440545384e-05, 'epoch': 1.48}) +[2026-06-16 08:20:02,633][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.0848584845662117, 'train/grad_norm_pre_clip': 28.76019287109375, 'timing/data': 0.020123583264648914, 'timing/model': 1.8299853149801493, 'learning_rate/qwen_vl_interface': 2.157024395968025e-06, 'learning_rate/action_model': 2.1570243959680252e-05, 'epoch': 1.48}) +[2026-06-16 08:20:08,672][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.0019030376570299268, 'train/grad_norm_pre_clip': 0.32232412695884705, 'timing/data': 0.02017382299527526, 'timing/model': 1.5698899258859456, 'learning_rate/qwen_vl_interface': 2.154713719286952e-06, 'learning_rate/action_model': 2.154713719286952e-05, 'epoch': 1.48}) +[2026-06-16 08:20:15,158][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.09072728455066681, 'train/grad_norm_pre_clip': 66.35101318359375, 'timing/data': 0.021946990862488747, 'timing/model': 1.5349574671126902, 'learning_rate/qwen_vl_interface': 2.152404314961149e-06, 'learning_rate/action_model': 2.152404314961149e-05, 'epoch': 1.48}) +[2026-06-16 08:20:21,669][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.002336810575798154, 'train/grad_norm_pre_clip': 2.0392956733703613, 'timing/data': 0.021491460036486387, 'timing/model': 1.5464535509236157, 'learning_rate/qwen_vl_interface': 2.1500961839399264e-06, 'learning_rate/action_model': 2.1500961839399262e-05, 'epoch': 1.48}) +[2026-06-16 08:20:28,065][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.045700352638959885, 'train/grad_norm_pre_clip': 9.655465126037598, 'timing/data': 0.019410418812185526, 'timing/model': 1.603370666038245, 'learning_rate/qwen_vl_interface': 2.1477893271720683e-06, 'learning_rate/action_model': 2.1477893271720682e-05, 'epoch': 1.48}) +[2026-06-16 08:20:34,424][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.005425516050308943, 'train/grad_norm_pre_clip': 4.087265491485596, 'timing/data': 0.021035027224570513, 'timing/model': 1.6078394982032478, 'learning_rate/qwen_vl_interface': 2.1454837456058385e-06, 'learning_rate/action_model': 2.1454837456058386e-05, 'epoch': 1.48}) +[2026-06-16 08:20:40,641][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.06946071982383728, 'train/grad_norm_pre_clip': 32.95115661621094, 'timing/data': 0.02134406939148903, 'timing/model': 1.7367526250891387, 'learning_rate/qwen_vl_interface': 2.1431794401889722e-06, 'learning_rate/action_model': 2.143179440188972e-05, 'epoch': 1.48}) +[2026-06-16 08:20:46,662][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.005067485384643078, 'train/grad_norm_pre_clip': 1.0643620491027832, 'timing/data': 0.020039701834321022, 'timing/model': 1.553272409364581, 'learning_rate/qwen_vl_interface': 2.1408764118686837e-06, 'learning_rate/action_model': 2.1408764118686838e-05, 'epoch': 1.48}) +[2026-06-16 08:20:53,270][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.09643704444169998, 'train/grad_norm_pre_clip': 9.263148307800293, 'timing/data': 0.02111119730398059, 'timing/model': 1.5549810281954706, 'learning_rate/qwen_vl_interface': 2.1385746615916598e-06, 'learning_rate/action_model': 2.1385746615916596e-05, 'epoch': 1.48}) +[2026-06-16 08:20:59,689][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.1160615086555481, 'train/grad_norm_pre_clip': 16.04610252380371, 'timing/data': 0.020394233986735344, 'timing/model': 1.5680355397053063, 'learning_rate/qwen_vl_interface': 2.1362741903040636e-06, 'learning_rate/action_model': 2.1362741903040633e-05, 'epoch': 1.48}) +[2026-06-16 08:21:05,974][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.00043604851816780865, 'train/grad_norm_pre_clip': 0.05060126632452011, 'timing/data': 0.020897946786135435, 'timing/model': 1.5245629898272455, 'learning_rate/qwen_vl_interface': 2.1339749989515318e-06, 'learning_rate/action_model': 2.1339749989515316e-05, 'epoch': 1.48}) +[2026-06-16 08:21:12,377][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.0024695273023098707, 'train/grad_norm_pre_clip': 0.5940017700195312, 'timing/data': 0.02335442788898945, 'timing/model': 1.6091250022873282, 'learning_rate/qwen_vl_interface': 2.131677088479175e-06, 'learning_rate/action_model': 2.1316770884791752e-05, 'epoch': 1.48}) +[2026-06-16 08:21:18,795][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.001328291604295373, 'train/grad_norm_pre_clip': 0.23743246495723724, 'timing/data': 0.021482359152287245, 'timing/model': 1.771367704961449, 'learning_rate/qwen_vl_interface': 2.1293804598315787e-06, 'learning_rate/action_model': 2.1293804598315787e-05, 'epoch': 1.48}) +[2026-06-16 08:21:24,966][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.004018447361886501, 'train/grad_norm_pre_clip': 1.1294869184494019, 'timing/data': 0.0199086032807827, 'timing/model': 1.5613905042409897, 'learning_rate/qwen_vl_interface': 2.1270851139527986e-06, 'learning_rate/action_model': 2.1270851139527983e-05, 'epoch': 1.48}) +[2026-06-16 08:21:31,267][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.01772860810160637, 'train/grad_norm_pre_clip': 14.376236915588379, 'timing/data': 0.019873072393238544, 'timing/model': 1.5631990339607, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 1.48}) +[2026-06-16 08:21:38,138][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.00011770202399929985, 'train/grad_norm_pre_clip': 0.012147515080869198, 'timing/data': 0.021731631364673376, 'timing/model': 1.6214426360093057, 'learning_rate/qwen_vl_interface': 2.1224982742752814e-06, 'learning_rate/action_model': 2.1224982742752812e-05, 'epoch': 1.48}) +[2026-06-16 08:21:44,437][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.00415234686806798, 'train/grad_norm_pre_clip': 4.787196636199951, 'timing/data': 0.019865102134644985, 'timing/model': 1.5836349120363593, 'learning_rate/qwen_vl_interface': 2.1202067823620213e-06, 'learning_rate/action_model': 2.1202067823620212e-05, 'epoch': 1.48}) +[2026-06-16 08:21:50,713][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.001655237632803619, 'train/grad_norm_pre_clip': 0.5969523191452026, 'timing/data': 0.021552840247750282, 'timing/model': 1.7999803149141371, 'learning_rate/qwen_vl_interface': 2.117916576988531e-06, 'learning_rate/action_model': 2.117916576988531e-05, 'epoch': 1.49}) +[2026-06-16 08:21:56,767][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.011145961470901966, 'train/grad_norm_pre_clip': 10.450791358947754, 'timing/data': 0.020839695818722248, 'timing/model': 1.544472083915025, 'learning_rate/qwen_vl_interface': 2.1156276590962257e-06, 'learning_rate/action_model': 2.1156276590962257e-05, 'epoch': 1.49}) +[2026-06-16 08:22:03,067][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.0019750851206481457, 'train/grad_norm_pre_clip': 0.8413175940513611, 'timing/data': 0.02193958219140768, 'timing/model': 1.545847271103412, 'learning_rate/qwen_vl_interface': 2.113340029625994e-06, 'learning_rate/action_model': 2.1133400296259938e-05, 'epoch': 1.49}) +[2026-06-16 08:22:09,358][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.027167676016688347, 'train/grad_norm_pre_clip': 22.175365447998047, 'timing/data': 0.02075761603191495, 'timing/model': 1.5552040049806237, 'learning_rate/qwen_vl_interface': 2.111053689518194e-06, 'learning_rate/action_model': 2.111053689518194e-05, 'epoch': 1.49}) +[2026-06-16 08:22:15,566][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.13817551732063293, 'train/grad_norm_pre_clip': 7.259385585784912, 'timing/data': 0.020821677055209875, 'timing/model': 1.5472402698360384, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 1.49}) +[2026-06-16 08:22:21,875][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.018455253913998604, 'train/grad_norm_pre_clip': 5.310871601104736, 'timing/data': 0.020889197010546923, 'timing/model': 1.5736355329863727, 'learning_rate/qwen_vl_interface': 2.106484881148672e-06, 'learning_rate/action_model': 2.1064848811486718e-05, 'epoch': 1.49}) +[2026-06-16 08:22:28,309][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.0005939932307228446, 'train/grad_norm_pre_clip': 0.166079580783844, 'timing/data': 0.01989051280543208, 'timing/model': 1.8988521019928157, 'learning_rate/qwen_vl_interface': 2.1042024147650137e-06, 'learning_rate/action_model': 2.1042024147650137e-05, 'epoch': 1.49}) +[2026-06-16 08:22:34,376][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.3153344690799713, 'train/grad_norm_pre_clip': 22.027143478393555, 'timing/data': 0.02072145603597164, 'timing/model': 1.5595815409906209, 'learning_rate/qwen_vl_interface': 2.1019212414999163e-06, 'learning_rate/action_model': 2.101921241499916e-05, 'epoch': 1.49}) +[2026-06-16 08:22:40,674][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.0026059714145958424, 'train/grad_norm_pre_clip': 0.8138851523399353, 'timing/data': 0.021505439653992653, 'timing/model': 1.55469423905015, 'learning_rate/qwen_vl_interface': 2.099641362291082e-06, 'learning_rate/action_model': 2.099641362291082e-05, 'epoch': 1.49}) +[2026-06-16 08:22:46,968][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.0008198645664379001, 'train/grad_norm_pre_clip': 0.3826422393321991, 'timing/data': 0.019992162939161062, 'timing/model': 1.547451146878302, 'learning_rate/qwen_vl_interface': 2.097362778075685e-06, 'learning_rate/action_model': 2.097362778075685e-05, 'epoch': 1.49}) +[2026-06-16 08:22:53,463][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.0005688203382305801, 'train/grad_norm_pre_clip': 0.24264977872371674, 'timing/data': 0.021383728832006454, 'timing/model': 1.5745356641709805, 'learning_rate/qwen_vl_interface': 2.0950854897903663e-06, 'learning_rate/action_model': 2.095085489790366e-05, 'epoch': 1.49}) +[2026-06-16 08:22:59,767][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.17360004782676697, 'train/grad_norm_pre_clip': 29.829118728637695, 'timing/data': 0.021158028859645128, 'timing/model': 1.5458174711093307, 'learning_rate/qwen_vl_interface': 2.092809498371229e-06, 'learning_rate/action_model': 2.0928094983712292e-05, 'epoch': 1.49}) +[2026-06-16 08:23:06,122][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.004036758095026016, 'train/grad_norm_pre_clip': 2.1877567768096924, 'timing/data': 0.026237691286951303, 'timing/model': 1.7988024367950857, 'learning_rate/qwen_vl_interface': 2.0905348047538498e-06, 'learning_rate/action_model': 2.0905348047538497e-05, 'epoch': 1.49}) +[2026-06-16 08:23:12,261][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.16005022823810577, 'train/grad_norm_pre_clip': 28.565629959106445, 'timing/data': 0.021034476812928915, 'timing/model': 1.5505511756055057, 'learning_rate/qwen_vl_interface': 2.0882614098732686e-06, 'learning_rate/action_model': 2.0882614098732685e-05, 'epoch': 1.49}) +[2026-06-16 08:23:18,561][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.002741785254329443, 'train/grad_norm_pre_clip': 0.9335430264472961, 'timing/data': 0.020573316141963005, 'timing/model': 1.5379119580611587, 'learning_rate/qwen_vl_interface': 2.085989314663993e-06, 'learning_rate/action_model': 2.0859893146639927e-05, 'epoch': 1.49}) +[2026-06-16 08:23:24,978][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.09832826256752014, 'train/grad_norm_pre_clip': 27.26421546936035, 'timing/data': 0.021071748342365026, 'timing/model': 1.5629629660397768, 'learning_rate/qwen_vl_interface': 2.083718520059995e-06, 'learning_rate/action_model': 2.0837185200599946e-05, 'epoch': 1.49}) +[2026-06-16 08:23:31,263][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.06597349047660828, 'train/grad_norm_pre_clip': 20.53473472595215, 'timing/data': 0.020690746139734983, 'timing/model': 1.5504872286692262, 'learning_rate/qwen_vl_interface': 2.0814490269947117e-06, 'learning_rate/action_model': 2.0814490269947118e-05, 'epoch': 1.49}) +[2026-06-16 08:23:37,662][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.0011191865196451545, 'train/grad_norm_pre_clip': 0.7559196949005127, 'timing/data': 0.021795001812279224, 'timing/model': 1.5595241216942668, 'learning_rate/qwen_vl_interface': 2.079180836401047e-06, 'learning_rate/action_model': 2.079180836401047e-05, 'epoch': 1.49}) +[2026-06-16 08:23:44,070][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.047923743724823, 'train/grad_norm_pre_clip': 18.94549560546875, 'timing/data': 0.021141158882528543, 'timing/model': 1.818496884778142, 'learning_rate/qwen_vl_interface': 2.0769139492113687e-06, 'learning_rate/action_model': 2.076913949211369e-05, 'epoch': 1.49}) +[2026-06-16 08:23:50,155][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.017051465809345245, 'train/grad_norm_pre_clip': 15.387948036193848, 'timing/data': 0.020651195663958788, 'timing/model': 1.5435967897064984, 'learning_rate/qwen_vl_interface': 2.0746483663575075e-06, 'learning_rate/action_model': 2.0746483663575076e-05, 'epoch': 1.49}) +[2026-06-16 08:23:56,455][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.020567674189805984, 'train/grad_norm_pre_clip': 11.457411766052246, 'timing/data': 0.02085801726207137, 'timing/model': 1.5456333528272808, 'learning_rate/qwen_vl_interface': 2.0723840887707595e-06, 'learning_rate/action_model': 2.0723840887707594e-05, 'epoch': 1.49}) +[2026-06-16 08:24:02,767][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.28561875224113464, 'train/grad_norm_pre_clip': 28.15400505065918, 'timing/data': 0.025997551158070564, 'timing/model': 1.6028208071365952, 'learning_rate/qwen_vl_interface': 2.0701211173818854e-06, 'learning_rate/action_model': 2.0701211173818853e-05, 'epoch': 1.49}) +[2026-06-16 08:24:09,079][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.0008592123631387949, 'train/grad_norm_pre_clip': 0.45295608043670654, 'timing/data': 0.02399722207337618, 'timing/model': 1.554543204139918, 'learning_rate/qwen_vl_interface': 2.0678594531211036e-06, 'learning_rate/action_model': 2.0678594531211033e-05, 'epoch': 1.49}) +[2026-06-16 08:24:15,391][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.0014245392521843314, 'train/grad_norm_pre_clip': 0.49201300740242004, 'timing/data': 0.02060498623177409, 'timing/model': 1.538620701059699, 'learning_rate/qwen_vl_interface': 2.0655990969181e-06, 'learning_rate/action_model': 2.0655990969181003e-05, 'epoch': 1.49}) +[2026-06-16 08:24:21,700][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.0011282401392236352, 'train/grad_norm_pre_clip': 0.20272426307201385, 'timing/data': 0.02344767889007926, 'timing/model': 1.7768042390234768, 'learning_rate/qwen_vl_interface': 2.0633400497020236e-06, 'learning_rate/action_model': 2.0633400497020232e-05, 'epoch': 1.5}) +[2026-06-16 08:24:27,860][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.00038012012373656034, 'train/grad_norm_pre_clip': 0.08460702002048492, 'timing/data': 0.0212619686499238, 'timing/model': 1.5494468626566231, 'learning_rate/qwen_vl_interface': 2.061082312401481e-06, 'learning_rate/action_model': 2.061082312401481e-05, 'epoch': 1.5}) +[2026-06-16 08:24:34,264][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.09913007915019989, 'train/grad_norm_pre_clip': 19.306604385375977, 'timing/data': 0.021360999904572964, 'timing/model': 1.5737117566168308, 'learning_rate/qwen_vl_interface': 2.058825885944544e-06, 'learning_rate/action_model': 2.0588258859445442e-05, 'epoch': 1.5}) +[2026-06-16 08:35:21,429][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.0010258284164592624, 'train/grad_norm_pre_clip': 0.7130929231643677, 'eval/loss': 0.034568965435028076, 'eval/flap_precision': 0.9793672627235214, 'eval/flap_recall': 0.9713506139154161, 'eval/flap_f1': 0.9753424657534246, 'eval/flap_rate_model': 0.20194444444444445, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9753424657534245, 'timing/data': 0.020813236944377422, 'timing/model': 1.5676404191181064, 'learning_rate/qwen_vl_interface': 2.0565707712587447e-06, 'learning_rate/action_model': 2.0565707712587444e-05, 'epoch': 1.5}) +[2026-06-16 08:35:21,431][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3700_state +[2026-06-16 08:35:21,437][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 08:35:29,862][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3700_state/model.safetensors +[2026-06-16 08:35:49,322][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3700_state/optimizer.bin +[2026-06-16 08:35:49,324][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3700_state/sampler.bin +[2026-06-16 08:35:49,324][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3700_state/sampler_1.bin +[2026-06-16 08:35:49,325][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3700_state/random_states_0.pkl +[2026-06-16 08:35:49,518][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3700 +[2026-06-16 08:35:49,518][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 08:35:49,533][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 08:35:55,716][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.0028834345284849405, 'train/grad_norm_pre_clip': 2.4683995246887207, 'timing/data': 0.020708086900413036, 'timing/model': 1.5528905168175697, 'learning_rate/qwen_vl_interface': 2.0543169692710747e-06, 'learning_rate/action_model': 2.054316969271075e-05, 'epoch': 1.5}) +[2026-06-16 08:36:01,929][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.08028847724199295, 'train/grad_norm_pre_clip': 35.12560272216797, 'timing/data': 0.020332645159214735, 'timing/model': 1.5493545709177852, 'learning_rate/qwen_vl_interface': 2.0520644809079873e-06, 'learning_rate/action_model': 2.0520644809079873e-05, 'epoch': 1.5}) +[2026-06-16 08:36:08,131][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.05658932775259018, 'train/grad_norm_pre_clip': 32.4907341003418, 'timing/data': 0.021787311881780624, 'timing/model': 1.5531009691767395, 'learning_rate/qwen_vl_interface': 2.049813307095395e-06, 'learning_rate/action_model': 2.0498133070953947e-05, 'epoch': 1.5}) +[2026-06-16 08:36:14,236][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.00019601878011599183, 'train/grad_norm_pre_clip': 0.04241502657532692, 'timing/data': 0.021148428786545992, 'timing/model': 1.502030709758401, 'learning_rate/qwen_vl_interface': 2.047563448758672e-06, 'learning_rate/action_model': 2.0475634487586717e-05, 'epoch': 1.5}) +[2026-06-16 08:36:20,485][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.07862364500761032, 'train/grad_norm_pre_clip': 17.179794311523438, 'timing/data': 0.020763187669217587, 'timing/model': 1.8017314742319286, 'learning_rate/qwen_vl_interface': 2.045314906822647e-06, 'learning_rate/action_model': 2.0453149068226467e-05, 'epoch': 1.5}) +[2026-06-16 08:36:26,541][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.014322920702397823, 'train/grad_norm_pre_clip': 9.731307983398438, 'timing/data': 0.020483594853430986, 'timing/model': 1.5510452701710165, 'learning_rate/qwen_vl_interface': 2.0430676822116124e-06, 'learning_rate/action_model': 2.0430676822116123e-05, 'epoch': 1.5}) +[2026-06-16 08:36:32,748][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.018037352710962296, 'train/grad_norm_pre_clip': 16.015167236328125, 'timing/data': 0.021316039841622114, 'timing/model': 1.521822181995958, 'learning_rate/qwen_vl_interface': 2.040821775849319e-06, 'learning_rate/action_model': 2.0408217758493188e-05, 'epoch': 1.5}) +[2026-06-16 08:36:39,025][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.20447508990764618, 'train/grad_norm_pre_clip': 56.70673370361328, 'timing/data': 0.0211529191583395, 'timing/model': 1.633114923723042, 'learning_rate/qwen_vl_interface': 2.03857718865897e-06, 'learning_rate/action_model': 2.0385771886589702e-05, 'epoch': 1.5}) +[2026-06-16 08:36:45,348][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.15046019852161407, 'train/grad_norm_pre_clip': 11.7308349609375, 'timing/data': 0.02101962873712182, 'timing/model': 1.5174553231336176, 'learning_rate/qwen_vl_interface': 2.0363339215632337e-06, 'learning_rate/action_model': 2.0363339215632334e-05, 'epoch': 1.5}) +[2026-06-16 08:36:51,579][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.0007228729082271457, 'train/grad_norm_pre_clip': 0.19389735162258148, 'timing/data': 0.01914948970079422, 'timing/model': 1.7446737377904356, 'learning_rate/qwen_vl_interface': 2.034091975484231e-06, 'learning_rate/action_model': 2.0340919754842307e-05, 'epoch': 1.5}) +[2026-06-16 08:36:57,839][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.1584227830171585, 'train/grad_norm_pre_clip': 13.924942016601562, 'timing/data': 0.0214874311350286, 'timing/model': 1.7465578969568014, 'learning_rate/qwen_vl_interface': 2.0318513513435415e-06, 'learning_rate/action_model': 2.0318513513435415e-05, 'epoch': 1.5}) +[2026-06-16 08:37:03,859][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.1657675951719284, 'train/grad_norm_pre_clip': 46.88029861450195, 'timing/data': 0.021233139093965292, 'timing/model': 1.5590359992347658, 'learning_rate/qwen_vl_interface': 2.0296120500622018e-06, 'learning_rate/action_model': 2.0296120500622017e-05, 'epoch': 1.5}) +[2026-06-16 08:37:10,243][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.00426109042018652, 'train/grad_norm_pre_clip': 1.8266578912734985, 'timing/data': 0.02127121016383171, 'timing/model': 1.5820583761669695, 'learning_rate/qwen_vl_interface': 2.0273740725607034e-06, 'learning_rate/action_model': 2.0273740725607032e-05, 'epoch': 1.5}) +[2026-06-16 08:37:16,635][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.1978740245103836, 'train/grad_norm_pre_clip': 15.078625679016113, 'timing/data': 0.022317003924399614, 'timing/model': 1.6620381209068, 'learning_rate/qwen_vl_interface': 2.0251374197589942e-06, 'learning_rate/action_model': 2.0251374197589943e-05, 'epoch': 1.5}) +[2026-06-16 08:37:22,874][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.030931437388062477, 'train/grad_norm_pre_clip': 21.46895408630371, 'timing/data': 0.020566487219184637, 'timing/model': 1.5275231311097741, 'learning_rate/qwen_vl_interface': 2.0229020925764785e-06, 'learning_rate/action_model': 2.0229020925764785e-05, 'epoch': 1.5}) +[2026-06-16 08:37:29,592][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.2200569361448288, 'train/grad_norm_pre_clip': 27.3299617767334, 'timing/data': 0.022889748215675354, 'timing/model': 1.6125362101010978, 'learning_rate/qwen_vl_interface': 2.0206680919320145e-06, 'learning_rate/action_model': 2.020668091932014e-05, 'epoch': 1.5}) +[2026-06-16 08:37:36,062][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.08479060977697372, 'train/grad_norm_pre_clip': 32.99943923950195, 'timing/data': 0.02568145003169775, 'timing/model': 1.7423179987818003, 'learning_rate/qwen_vl_interface': 2.0184354187439164e-06, 'learning_rate/action_model': 2.0184354187439162e-05, 'epoch': 1.5}) +[2026-06-16 08:37:42,237][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.15806186199188232, 'train/grad_norm_pre_clip': 20.7294921875, 'timing/data': 0.021194080356508493, 'timing/model': 1.6350394659675658, 'learning_rate/qwen_vl_interface': 2.016204073929949e-06, 'learning_rate/action_model': 2.016204073929949e-05, 'epoch': 1.5}) +[2026-06-16 08:37:48,466][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.11373849213123322, 'train/grad_norm_pre_clip': 13.007759094238281, 'timing/data': 0.022580505814403296, 'timing/model': 1.5353786488994956, 'learning_rate/qwen_vl_interface': 2.0139740584073363e-06, 'learning_rate/action_model': 2.0139740584073364e-05, 'epoch': 1.5}) +[2026-06-16 08:37:54,764][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.08039673417806625, 'train/grad_norm_pre_clip': 25.552568435668945, 'timing/data': 0.020999127998948097, 'timing/model': 1.5554188126698136, 'learning_rate/qwen_vl_interface': 2.011745373092753e-06, 'learning_rate/action_model': 2.011745373092753e-05, 'epoch': 1.5}) +[2026-06-16 08:38:01,095][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.0001115250852308236, 'train/grad_norm_pre_clip': 0.023985901847481728, 'timing/data': 0.020488956943154335, 'timing/model': 1.5484141297638416, 'learning_rate/qwen_vl_interface': 2.0095180189023267e-06, 'learning_rate/action_model': 2.0095180189023265e-05, 'epoch': 1.5}) +[2026-06-16 08:38:07,486][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.04246916621923447, 'train/grad_norm_pre_clip': 16.295324325561523, 'timing/data': 0.023001407273113728, 'timing/model': 1.8214806611649692, 'learning_rate/qwen_vl_interface': 2.0072919967516393e-06, 'learning_rate/action_model': 2.0072919967516392e-05, 'epoch': 1.51}) +[2026-06-16 08:38:13,561][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.0699094906449318, 'train/grad_norm_pre_clip': 14.643400192260742, 'timing/data': 0.02106086863204837, 'timing/model': 1.5526393009349704, 'learning_rate/qwen_vl_interface': 2.0050673075557245e-06, 'learning_rate/action_model': 2.0050673075557243e-05, 'epoch': 1.51}) +[2026-06-16 08:38:19,922][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.0011730898404493928, 'train/grad_norm_pre_clip': 0.643347978591919, 'timing/data': 0.020655847154557705, 'timing/model': 1.6142936600372195, 'learning_rate/qwen_vl_interface': 2.0028439522290677e-06, 'learning_rate/action_model': 2.0028439522290676e-05, 'epoch': 1.51}) +[2026-06-16 08:38:26,062][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.0021056931000202894, 'train/grad_norm_pre_clip': 0.7228154540061951, 'timing/data': 0.02197206299751997, 'timing/model': 1.5473713958635926, 'learning_rate/qwen_vl_interface': 2.000621931685606e-06, 'learning_rate/action_model': 2.000621931685606e-05, 'epoch': 1.51}) +[2026-06-16 08:38:32,361][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.060783177614212036, 'train/grad_norm_pre_clip': 24.417339324951172, 'timing/data': 0.021635981742292643, 'timing/model': 1.56176425376907, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 1.51}) +[2026-06-16 08:38:38,762][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.0028708726167678833, 'train/grad_norm_pre_clip': 1.0764211416244507, 'timing/data': 0.023864781018346548, 'timing/model': 1.568645247258246, 'learning_rate/qwen_vl_interface': 1.9961818986012767e-06, 'learning_rate/action_model': 1.9961818986012763e-05, 'epoch': 1.51}) +[2026-06-16 08:38:45,097][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.09591912478208542, 'train/grad_norm_pre_clip': 10.951550483703613, 'timing/data': 0.02143340976908803, 'timing/model': 1.7938922038301826, 'learning_rate/qwen_vl_interface': 1.9939638878855372e-06, 'learning_rate/action_model': 1.993963887885537e-05, 'epoch': 1.51}) +[2026-06-16 08:38:51,171][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.17752942442893982, 'train/grad_norm_pre_clip': 10.9854736328125, 'timing/data': 0.021654751151800156, 'timing/model': 1.5587983909063041, 'learning_rate/qwen_vl_interface': 1.991747215603252e-06, 'learning_rate/action_model': 1.991747215603252e-05, 'epoch': 1.51}) +[2026-06-16 08:38:57,463][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.29388195276260376, 'train/grad_norm_pre_clip': 26.433090209960938, 'timing/data': 0.02114316914230585, 'timing/model': 1.5572455641813576, 'learning_rate/qwen_vl_interface': 1.9895318826656116e-06, 'learning_rate/action_model': 1.9895318826656114e-05, 'epoch': 1.51}) +[2026-06-16 08:39:03,765][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.028752904385328293, 'train/grad_norm_pre_clip': 14.770941734313965, 'timing/data': 0.021267959848046303, 'timing/model': 1.5490021049045026, 'learning_rate/qwen_vl_interface': 1.9873178899832553e-06, 'learning_rate/action_model': 1.9873178899832554e-05, 'epoch': 1.51}) +[2026-06-16 08:39:10,165][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.0004396779404487461, 'train/grad_norm_pre_clip': 0.09853248298168182, 'timing/data': 0.02168147172778845, 'timing/model': 1.5620617168024182, 'learning_rate/qwen_vl_interface': 1.985105238466272e-06, 'learning_rate/action_model': 1.9851052384662722e-05, 'epoch': 1.51}) +[2026-06-16 08:39:16,461][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.04712917283177376, 'train/grad_norm_pre_clip': 19.567852020263672, 'timing/data': 0.02135063987225294, 'timing/model': 1.5392891699448228, 'learning_rate/qwen_vl_interface': 1.9828939290242e-06, 'learning_rate/action_model': 1.9828939290241998e-05, 'epoch': 1.51}) +[2026-06-16 08:39:22,800][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.03552670776844025, 'train/grad_norm_pre_clip': 10.853846549987793, 'timing/data': 0.029425637796521187, 'timing/model': 1.7789461240172386, 'learning_rate/qwen_vl_interface': 1.980683962566024e-06, 'learning_rate/action_model': 1.980683962566024e-05, 'epoch': 1.51}) +[2026-06-16 08:39:28,863][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.00022331405489239842, 'train/grad_norm_pre_clip': 0.031907182186841965, 'timing/data': 0.021711992099881172, 'timing/model': 1.545437796972692, 'learning_rate/qwen_vl_interface': 1.9784753400001784e-06, 'learning_rate/action_model': 1.978475340000178e-05, 'epoch': 1.51}) +[2026-06-16 08:39:35,170][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.0021254876628518105, 'train/grad_norm_pre_clip': 0.49966567754745483, 'timing/data': 0.020647906698286533, 'timing/model': 1.557763225864619, 'learning_rate/qwen_vl_interface': 1.9762680622345443e-06, 'learning_rate/action_model': 1.9762680622345442e-05, 'epoch': 1.51}) +[2026-06-16 08:39:41,467][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.00039749121060594916, 'train/grad_norm_pre_clip': 0.12502385675907135, 'timing/data': 0.023658969905227423, 'timing/model': 1.5378862698562443, 'learning_rate/qwen_vl_interface': 1.9740621301764513e-06, 'learning_rate/action_model': 1.9740621301764513e-05, 'epoch': 1.51}) +[2026-06-16 08:39:47,778][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.000346504122717306, 'train/grad_norm_pre_clip': 0.09615887701511383, 'timing/data': 0.021124400198459625, 'timing/model': 1.535125118214637, 'learning_rate/qwen_vl_interface': 1.971857544732672e-06, 'learning_rate/action_model': 1.971857544732672e-05, 'epoch': 1.51}) +[2026-06-16 08:39:54,155][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.10010716319084167, 'train/grad_norm_pre_clip': 29.641210556030273, 'timing/data': 0.023463531862944365, 'timing/model': 1.5915665701031685, 'learning_rate/qwen_vl_interface': 1.96965430680943e-06, 'learning_rate/action_model': 1.96965430680943e-05, 'epoch': 1.51}) +[2026-06-16 08:40:00,500][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.008131314069032669, 'train/grad_norm_pre_clip': 5.036364555358887, 'timing/data': 0.02081010304391384, 'timing/model': 1.7805805979296565, 'learning_rate/qwen_vl_interface': 1.967452417312393e-06, 'learning_rate/action_model': 1.9674524173123927e-05, 'epoch': 1.51}) +[2026-06-16 08:40:06,565][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.02737395279109478, 'train/grad_norm_pre_clip': 17.249622344970703, 'timing/data': 0.02151323389261961, 'timing/model': 1.5444228588603437, 'learning_rate/qwen_vl_interface': 1.9652518771466746e-06, 'learning_rate/action_model': 1.9652518771466747e-05, 'epoch': 1.51}) +[2026-06-16 08:40:12,959][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.014918780885636806, 'train/grad_norm_pre_clip': 9.86907958984375, 'timing/data': 0.021816925145685673, 'timing/model': 1.5296714906580746, 'learning_rate/qwen_vl_interface': 1.963052687216834e-06, 'learning_rate/action_model': 1.963052687216834e-05, 'epoch': 1.51}) +[2026-06-16 08:40:19,271][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.0008397154160775244, 'train/grad_norm_pre_clip': 0.19486695528030396, 'timing/data': 0.0217336961068213, 'timing/model': 1.5484451507218182, 'learning_rate/qwen_vl_interface': 1.960854848426874e-06, 'learning_rate/action_model': 1.960854848426874e-05, 'epoch': 1.51}) +[2026-06-16 08:40:25,583][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.0006826950702816248, 'train/grad_norm_pre_clip': 0.24641531705856323, 'timing/data': 0.021266616880893707, 'timing/model': 1.5594769190065563, 'learning_rate/qwen_vl_interface': 1.9586583616802446e-06, 'learning_rate/action_model': 1.9586583616802446e-05, 'epoch': 1.51}) +[2026-06-16 08:40:32,003][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.1765974909067154, 'train/grad_norm_pre_clip': 14.334628105163574, 'timing/data': 0.021578047890216112, 'timing/model': 1.6326515204273164, 'learning_rate/qwen_vl_interface': 1.956463227879838e-06, 'learning_rate/action_model': 1.9564632278798377e-05, 'epoch': 1.51}) +[2026-06-16 08:40:38,306][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.05415746569633484, 'train/grad_norm_pre_clip': 28.70920181274414, 'timing/data': 0.021853799000382423, 'timing/model': 1.787526581902057, 'learning_rate/qwen_vl_interface': 1.95426944792799e-06, 'learning_rate/action_model': 1.95426944792799e-05, 'epoch': 1.51}) +[2026-06-16 08:40:44,363][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.002856645965948701, 'train/grad_norm_pre_clip': 0.40950900316238403, 'timing/data': 0.021266140043735504, 'timing/model': 1.5501377498731017, 'learning_rate/qwen_vl_interface': 1.9520770227264817e-06, 'learning_rate/action_model': 1.9520770227264817e-05, 'epoch': 1.52}) +[2026-06-16 08:40:50,660][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.002925541950389743, 'train/grad_norm_pre_clip': 1.8303849697113037, 'timing/data': 0.021087090019136667, 'timing/model': 1.5528543828986585, 'learning_rate/qwen_vl_interface': 1.9498859531765373e-06, 'learning_rate/action_model': 1.9498859531765372e-05, 'epoch': 1.52}) +[2026-06-16 08:40:57,058][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.002004035748541355, 'train/grad_norm_pre_clip': 0.3610321879386902, 'timing/data': 0.02107356209307909, 'timing/model': 1.6421873611398041, 'learning_rate/qwen_vl_interface': 1.9476962401788204e-06, 'learning_rate/action_model': 1.9476962401788204e-05, 'epoch': 1.52}) +[2026-06-16 08:41:03,359][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.005333741195499897, 'train/grad_norm_pre_clip': 1.2949098348617554, 'timing/data': 0.021412422880530357, 'timing/model': 1.5363623220473528, 'learning_rate/qwen_vl_interface': 1.9455078846334397e-06, 'learning_rate/action_model': 1.9455078846334395e-05, 'epoch': 1.52}) +[2026-06-16 08:41:09,766][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.009492844343185425, 'train/grad_norm_pre_clip': 3.1079366207122803, 'timing/data': 0.02139181410893798, 'timing/model': 1.5672069992870092, 'learning_rate/qwen_vl_interface': 1.9433208874399465e-06, 'learning_rate/action_model': 1.9433208874399464e-05, 'epoch': 1.52}) +[2026-06-16 08:41:16,105][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.006510780658572912, 'train/grad_norm_pre_clip': 6.12396240234375, 'timing/data': 0.02158327493816614, 'timing/model': 1.7911530300043523, 'learning_rate/qwen_vl_interface': 1.9411352494973314e-06, 'learning_rate/action_model': 1.9411352494973315e-05, 'epoch': 1.52}) +[2026-06-16 08:41:22,206][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.00027489446802064776, 'train/grad_norm_pre_clip': 0.0515034981071949, 'timing/data': 0.02088971482589841, 'timing/model': 1.5945080919191241, 'learning_rate/qwen_vl_interface': 1.938950971704031e-06, 'learning_rate/action_model': 1.9389509717040308e-05, 'epoch': 1.52}) +[2026-06-16 08:41:28,466][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.26630300283432007, 'train/grad_norm_pre_clip': 11.770943641662598, 'timing/data': 0.02038175566121936, 'timing/model': 1.5536158001050353, 'learning_rate/qwen_vl_interface': 1.936768054957915e-06, 'learning_rate/action_model': 1.9367680549579147e-05, 'epoch': 1.52}) +[2026-06-16 08:41:34,766][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.00030663429060950875, 'train/grad_norm_pre_clip': 0.058590661734342575, 'timing/data': 0.020782505627721548, 'timing/model': 1.5466126361861825, 'learning_rate/qwen_vl_interface': 1.9345865001562994e-06, 'learning_rate/action_model': 1.9345865001562994e-05, 'epoch': 1.52}) +[2026-06-16 08:41:41,174][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.03169538825750351, 'train/grad_norm_pre_clip': 17.788694381713867, 'timing/data': 0.021659428719431162, 'timing/model': 1.5737489932216704, 'learning_rate/qwen_vl_interface': 1.9324063081959394e-06, 'learning_rate/action_model': 1.9324063081959395e-05, 'epoch': 1.52}) +[2026-06-16 08:41:47,529][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.007952269166707993, 'train/grad_norm_pre_clip': 4.683388710021973, 'timing/data': 0.021561119239777327, 'timing/model': 1.577227811794728, 'learning_rate/qwen_vl_interface': 1.9302274799730304e-06, 'learning_rate/action_model': 1.9302274799730306e-05, 'epoch': 1.52}) +[2026-06-16 08:41:53,792][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.06404245644807816, 'train/grad_norm_pre_clip': 13.404111862182617, 'timing/data': 0.019938188139349222, 'timing/model': 1.7717337869107723, 'learning_rate/qwen_vl_interface': 1.9280500163832055e-06, 'learning_rate/action_model': 1.9280500163832055e-05, 'epoch': 1.52}) +[2026-06-16 08:42:00,043][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.22616715729236603, 'train/grad_norm_pre_clip': 10.485647201538086, 'timing/data': 0.02161782095208764, 'timing/model': 1.729308282956481, 'learning_rate/qwen_vl_interface': 1.925873918321537e-06, 'learning_rate/action_model': 1.925873918321537e-05, 'epoch': 1.52}) +[2026-06-16 08:42:06,067][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.000374864786863327, 'train/grad_norm_pre_clip': 0.0861370712518692, 'timing/data': 0.02260095300152898, 'timing/model': 1.5561785041354597, 'learning_rate/qwen_vl_interface': 1.92369918668254e-06, 'learning_rate/action_model': 1.92369918668254e-05, 'epoch': 1.52}) +[2026-06-16 08:42:12,478][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.06971369683742523, 'train/grad_norm_pre_clip': 36.58717727661133, 'timing/data': 0.020839810837060213, 'timing/model': 1.568481178022921, 'learning_rate/qwen_vl_interface': 1.9215258223601605e-06, 'learning_rate/action_model': 1.9215258223601605e-05, 'epoch': 1.52}) +[2026-06-16 08:42:18,867][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.0009702657698653638, 'train/grad_norm_pre_clip': 0.1900559663772583, 'timing/data': 0.021930374205112457, 'timing/model': 1.5570232858881354, 'learning_rate/qwen_vl_interface': 1.9193538262477874e-06, 'learning_rate/action_model': 1.9193538262477874e-05, 'epoch': 1.52}) +[2026-06-16 08:42:25,274][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.00019816473650280386, 'train/grad_norm_pre_clip': 0.0191547479480505, 'timing/data': 0.02167041413486004, 'timing/model': 1.5997614939697087, 'learning_rate/qwen_vl_interface': 1.917183199238247e-06, 'learning_rate/action_model': 1.917183199238247e-05, 'epoch': 1.52}) +[2026-06-16 08:42:31,606][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.0015927141066640615, 'train/grad_norm_pre_clip': 0.5201606154441833, 'timing/data': 0.021779764909297228, 'timing/model': 1.7886954150162637, 'learning_rate/qwen_vl_interface': 1.9150139422238016e-06, 'learning_rate/action_model': 1.9150139422238018e-05, 'epoch': 1.52}) +[2026-06-16 08:42:37,908][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.007316575851291418, 'train/grad_norm_pre_clip': 3.4473652839660645, 'timing/data': 0.02138000587001443, 'timing/model': 1.790170330554247, 'learning_rate/qwen_vl_interface': 1.9128460560961513e-06, 'learning_rate/action_model': 1.9128460560961514e-05, 'epoch': 1.52}) +[2026-06-16 08:42:44,080][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.2917589545249939, 'train/grad_norm_pre_clip': 30.413463592529297, 'timing/data': 0.025126971304416656, 'timing/model': 1.5657824478112161, 'learning_rate/qwen_vl_interface': 1.910679541746432e-06, 'learning_rate/action_model': 1.9106795417464323e-05, 'epoch': 1.52}) +[2026-06-16 08:42:50,374][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 4.6226647100411355e-05, 'train/grad_norm_pre_clip': 0.004627286456525326, 'timing/data': 0.020547357387840748, 'timing/model': 1.555107249878347, 'learning_rate/qwen_vl_interface': 1.9085144000652166e-06, 'learning_rate/action_model': 1.9085144000652166e-05, 'epoch': 1.52}) +[2026-06-16 08:42:56,678][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.12416943907737732, 'train/grad_norm_pre_clip': 18.859203338623047, 'timing/data': 0.020359127316623926, 'timing/model': 1.5563132562674582, 'learning_rate/qwen_vl_interface': 1.9063506319425116e-06, 'learning_rate/action_model': 1.9063506319425114e-05, 'epoch': 1.52}) +[2026-06-16 08:43:03,068][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.10366354137659073, 'train/grad_norm_pre_clip': 18.526762008666992, 'timing/data': 0.021685588639229536, 'timing/model': 1.5236995983868837, 'learning_rate/qwen_vl_interface': 1.9041882382677613e-06, 'learning_rate/action_model': 1.9041882382677612e-05, 'epoch': 1.52}) +[2026-06-16 08:43:09,410][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.002227137330919504, 'train/grad_norm_pre_clip': 0.5915451049804688, 'timing/data': 0.021600060164928436, 'timing/model': 1.7892912980169058, 'learning_rate/qwen_vl_interface': 1.9020272199298452e-06, 'learning_rate/action_model': 1.902027219929845e-05, 'epoch': 1.52}) +[2026-06-16 08:43:15,474][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.07797425240278244, 'train/grad_norm_pre_clip': 38.059478759765625, 'timing/data': 0.020962608978152275, 'timing/model': 1.5573767493478954, 'learning_rate/qwen_vl_interface': 1.8998675778170736e-06, 'learning_rate/action_model': 1.8998675778170734e-05, 'epoch': 1.52}) +[2026-06-16 08:43:21,859][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.0647900402545929, 'train/grad_norm_pre_clip': 23.626615524291992, 'timing/data': 0.020304680336266756, 'timing/model': 1.6094871568493545, 'learning_rate/qwen_vl_interface': 1.8977093128171955e-06, 'learning_rate/action_model': 1.8977093128171954e-05, 'epoch': 1.53}) +[2026-06-16 08:43:28,270][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.003919520415365696, 'train/grad_norm_pre_clip': 1.5106967687606812, 'timing/data': 0.021121341735124588, 'timing/model': 1.5677683199755847, 'learning_rate/qwen_vl_interface': 1.8955524258173919e-06, 'learning_rate/action_model': 1.895552425817392e-05, 'epoch': 1.53}) +[2026-06-16 08:43:34,605][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.0002134371461579576, 'train/grad_norm_pre_clip': 0.026128413155674934, 'timing/data': 0.02140801213681698, 'timing/model': 1.5549963526427746, 'learning_rate/qwen_vl_interface': 1.8933969177042769e-06, 'learning_rate/action_model': 1.8933969177042768e-05, 'epoch': 1.53}) +[2026-06-16 08:43:41,040][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.014356986619532108, 'train/grad_norm_pre_clip': 5.715412616729736, 'timing/data': 0.020157441962510347, 'timing/model': 1.5849787699989974, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 1.53}) +[2026-06-16 08:43:47,388][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.11789663881063461, 'train/grad_norm_pre_clip': 12.854020118713379, 'timing/data': 0.02063578274101019, 'timing/model': 1.8152682860381901, 'learning_rate/qwen_vl_interface': 1.8890900416817398e-06, 'learning_rate/action_model': 1.8890900416817397e-05, 'epoch': 1.53}) +[2026-06-16 08:43:53,599][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.0005043027922511101, 'train/grad_norm_pre_clip': 0.09899259358644485, 'timing/data': 0.0194961610250175, 'timing/model': 1.590543411206454, 'learning_rate/qwen_vl_interface': 1.8869386755427117e-06, 'learning_rate/action_model': 1.8869386755427114e-05, 'epoch': 1.53}) +[2026-06-16 08:44:00,064][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.24552799761295319, 'train/grad_norm_pre_clip': 20.073978424072266, 'timing/data': 0.021243794821202755, 'timing/model': 1.5404073791578412, 'learning_rate/qwen_vl_interface': 1.8847886918311597e-06, 'learning_rate/action_model': 1.8847886918311598e-05, 'epoch': 1.53}) +[2026-06-16 08:44:06,461][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.06950859725475311, 'train/grad_norm_pre_clip': 28.38548469543457, 'timing/data': 0.020730855409055948, 'timing/model': 1.5464889579452574, 'learning_rate/qwen_vl_interface': 1.8826400914308616e-06, 'learning_rate/action_model': 1.8826400914308613e-05, 'epoch': 1.53}) +[2026-06-16 08:44:12,874][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.12783943116664886, 'train/grad_norm_pre_clip': 27.833362579345703, 'timing/data': 0.02095663594081998, 'timing/model': 1.5362107679247856, 'learning_rate/qwen_vl_interface': 1.880492875225025e-06, 'learning_rate/action_model': 1.8804928752250252e-05, 'epoch': 1.53}) +[2026-06-16 08:44:19,249][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.28942620754241943, 'train/grad_norm_pre_clip': 20.296913146972656, 'timing/data': 0.020824255887418985, 'timing/model': 1.5887578870169818, 'learning_rate/qwen_vl_interface': 1.8783470440962922e-06, 'learning_rate/action_model': 1.878347044096292e-05, 'epoch': 1.53}) +[2026-06-16 08:44:25,550][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.0031695265788584948, 'train/grad_norm_pre_clip': 3.565676689147949, 'timing/data': 0.024598454125225544, 'timing/model': 1.8334469459950924, 'learning_rate/qwen_vl_interface': 1.8762025989267293e-06, 'learning_rate/action_model': 1.876202598926729e-05, 'epoch': 1.53}) +[2026-06-16 08:44:31,567][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.00023865597904659808, 'train/grad_norm_pre_clip': 0.0487530380487442, 'timing/data': 0.02044701622799039, 'timing/model': 1.5562303708866239, 'learning_rate/qwen_vl_interface': 1.874059540597839e-06, 'learning_rate/action_model': 1.874059540597839e-05, 'epoch': 1.53}) +[2026-06-16 08:44:37,868][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.009088631719350815, 'train/grad_norm_pre_clip': 4.69027853012085, 'timing/data': 0.019329785834997892, 'timing/model': 1.5647763698361814, 'learning_rate/qwen_vl_interface': 1.8719178699905508e-06, 'learning_rate/action_model': 1.8719178699905508e-05, 'epoch': 1.53}) +[2026-06-16 08:44:44,363][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.00024218513863161206, 'train/grad_norm_pre_clip': 0.03283555433154106, 'timing/data': 0.024268925189971924, 'timing/model': 1.595115743111819, 'learning_rate/qwen_vl_interface': 1.8697775879852242e-06, 'learning_rate/action_model': 1.8697775879852242e-05, 'epoch': 1.53}) +[2026-06-16 08:44:50,764][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.03151150420308113, 'train/grad_norm_pre_clip': 20.252351760864258, 'timing/data': 0.021544411312788725, 'timing/model': 1.6046727770008147, 'learning_rate/qwen_vl_interface': 1.86763869546165e-06, 'learning_rate/action_model': 1.86763869546165e-05, 'epoch': 1.53}) +[2026-06-16 08:44:57,204][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.0032983548007905483, 'train/grad_norm_pre_clip': 0.6426163911819458, 'timing/data': 0.02335646515712142, 'timing/model': 1.7895146859809756, 'learning_rate/qwen_vl_interface': 1.865501193299045e-06, 'learning_rate/action_model': 1.8655011932990447e-05, 'epoch': 1.53}) +[2026-06-16 08:45:03,266][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.001864172867499292, 'train/grad_norm_pre_clip': 0.6944401264190674, 'timing/data': 0.021741041913628578, 'timing/model': 1.5606288211420178, 'learning_rate/qwen_vl_interface': 1.8633650823760555e-06, 'learning_rate/action_model': 1.8633650823760555e-05, 'epoch': 1.53}) +[2026-06-16 08:45:09,564][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.001150271506048739, 'train/grad_norm_pre_clip': 0.18204492330551147, 'timing/data': 0.020345690194517374, 'timing/model': 1.550055653322488, 'learning_rate/qwen_vl_interface': 1.8612303635707562e-06, 'learning_rate/action_model': 1.861230363570756e-05, 'epoch': 1.53}) +[2026-06-16 08:45:15,772][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.0005855868221260607, 'train/grad_norm_pre_clip': 0.21624501049518585, 'timing/data': 0.02126677194610238, 'timing/model': 1.535528075415641, 'learning_rate/qwen_vl_interface': 1.8590970377606494e-06, 'learning_rate/action_model': 1.859097037760649e-05, 'epoch': 1.53}) +[2026-06-16 08:45:22,169][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.0021555342245846987, 'train/grad_norm_pre_clip': 0.6600351929664612, 'timing/data': 0.021983654238283634, 'timing/model': 1.552587830927223, 'learning_rate/qwen_vl_interface': 1.8569651058226667e-06, 'learning_rate/action_model': 1.8569651058226666e-05, 'epoch': 1.53}) +[2026-06-16 08:45:28,554][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.21245242655277252, 'train/grad_norm_pre_clip': 11.609395980834961, 'timing/data': 0.020685612224042416, 'timing/model': 1.5523709417320788, 'learning_rate/qwen_vl_interface': 1.854834568633161e-06, 'learning_rate/action_model': 1.854834568633161e-05, 'epoch': 1.53}) +[2026-06-16 08:45:34,887][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.0009033679380081594, 'train/grad_norm_pre_clip': 0.31021279096603394, 'timing/data': 0.021654563955962658, 'timing/model': 1.7755785551853478, 'learning_rate/qwen_vl_interface': 1.8527054270679174e-06, 'learning_rate/action_model': 1.8527054270679172e-05, 'epoch': 1.53}) +[2026-06-16 08:45:41,025][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.14746585488319397, 'train/grad_norm_pre_clip': 25.42799186706543, 'timing/data': 0.020670874044299126, 'timing/model': 1.6163104549050331, 'learning_rate/qwen_vl_interface': 1.850577682002146e-06, 'learning_rate/action_model': 1.850577682002146e-05, 'epoch': 1.53}) +[2026-06-16 08:45:47,460][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.0006801224662922323, 'train/grad_norm_pre_clip': 0.09510726481676102, 'timing/data': 0.021964356303215027, 'timing/model': 1.5363001278601587, 'learning_rate/qwen_vl_interface': 1.848451334310482e-06, 'learning_rate/action_model': 1.848451334310482e-05, 'epoch': 1.53}) +[2026-06-16 08:45:53,864][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.0003481284948065877, 'train/grad_norm_pre_clip': 0.08191336691379547, 'timing/data': 0.021321066189557314, 'timing/model': 1.5613974900916219, 'learning_rate/qwen_vl_interface': 1.8463263848669865e-06, 'learning_rate/action_model': 1.8463263848669864e-05, 'epoch': 1.54}) +[2026-06-16 08:46:00,261][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.0010967125417664647, 'train/grad_norm_pre_clip': 0.6279066205024719, 'timing/data': 0.020479914732277393, 'timing/model': 1.558783964253962, 'learning_rate/qwen_vl_interface': 1.8442028345451468e-06, 'learning_rate/action_model': 1.8442028345451468e-05, 'epoch': 1.54}) +[2026-06-16 08:46:06,561][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.2060219645500183, 'train/grad_norm_pre_clip': 57.873958587646484, 'timing/data': 0.021562386769801378, 'timing/model': 1.5467114560306072, 'learning_rate/qwen_vl_interface': 1.8420806842178746e-06, 'learning_rate/action_model': 1.8420806842178745e-05, 'epoch': 1.54}) +[2026-06-16 08:46:12,922][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.001959947869181633, 'train/grad_norm_pre_clip': 0.46545878052711487, 'timing/data': 0.0200364850461483, 'timing/model': 1.8186274580657482, 'learning_rate/qwen_vl_interface': 1.8399599347575034e-06, 'learning_rate/action_model': 1.8399599347575033e-05, 'epoch': 1.54}) +[2026-06-16 08:57:00,723][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.1912146508693695, 'train/grad_norm_pre_clip': 34.442771911621094, 'eval/loss': 0.034812990576028824, 'eval/flap_precision': 0.9689189189189189, 'eval/flap_recall': 0.9781718963165075, 'eval/flap_f1': 0.9735234215885946, 'eval/flap_rate_model': 0.20555555555555555, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9735234215885945, 'timing/data': 0.02133524790406227, 'timing/model': 1.5778301688842475, 'learning_rate/qwen_vl_interface': 1.8378405870357943e-06, 'learning_rate/action_model': 1.837840587035794e-05, 'epoch': 1.54}) +[2026-06-16 08:57:00,724][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3800_state +[2026-06-16 08:57:00,730][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 08:57:09,329][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3800_state/model.safetensors +[2026-06-16 08:57:30,390][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3800_state/optimizer.bin +[2026-06-16 08:57:30,390][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3800_state/sampler.bin +[2026-06-16 08:57:30,391][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3800_state/sampler_1.bin +[2026-06-16 08:57:30,392][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3800_state/random_states_0.pkl +[2026-06-16 08:57:30,581][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3800 +[2026-06-16 08:57:30,581][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 08:57:30,597][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 08:57:36,924][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.004050521645694971, 'train/grad_norm_pre_clip': 4.221644401550293, 'timing/data': 0.021407008171081543, 'timing/model': 1.8445290359668434, 'learning_rate/qwen_vl_interface': 1.8357226419239316e-06, 'learning_rate/action_model': 1.8357226419239315e-05, 'epoch': 1.54}) +[2026-06-16 08:57:42,939][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 7.248074689414352e-05, 'train/grad_norm_pre_clip': 0.005481415428221226, 'timing/data': 0.022013980895280838, 'timing/model': 1.5587353589944541, 'learning_rate/qwen_vl_interface': 1.8336061002925224e-06, 'learning_rate/action_model': 1.8336061002925224e-05, 'epoch': 1.54}) +[2026-06-16 08:57:49,212][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.04019255191087723, 'train/grad_norm_pre_clip': 17.65033531188965, 'timing/data': 0.02073008567094803, 'timing/model': 1.6141702020540833, 'learning_rate/qwen_vl_interface': 1.8314909630115985e-06, 'learning_rate/action_model': 1.8314909630115983e-05, 'epoch': 1.54}) +[2026-06-16 08:57:55,444][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.0005856261705048382, 'train/grad_norm_pre_clip': 0.09861477464437485, 'timing/data': 0.02055291598662734, 'timing/model': 1.5022428599186242, 'learning_rate/qwen_vl_interface': 1.8293772309506092e-06, 'learning_rate/action_model': 1.829377230950609e-05, 'epoch': 1.54}) +[2026-06-16 08:58:01,643][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.0015474118990823627, 'train/grad_norm_pre_clip': 0.805576741695404, 'timing/data': 0.02084504719823599, 'timing/model': 1.5185152785852551, 'learning_rate/qwen_vl_interface': 1.8272649049784326e-06, 'learning_rate/action_model': 1.8272649049784323e-05, 'epoch': 1.54}) +[2026-06-16 08:58:07,868][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.0002475859655532986, 'train/grad_norm_pre_clip': 0.05547376349568367, 'timing/data': 0.020298545248806477, 'timing/model': 1.541617683134973, 'learning_rate/qwen_vl_interface': 1.8251539859633653e-06, 'learning_rate/action_model': 1.825153985963365e-05, 'epoch': 1.54}) +[2026-06-16 08:58:14,193][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.31586572527885437, 'train/grad_norm_pre_clip': 11.140621185302734, 'timing/data': 0.021017488092184067, 'timing/model': 1.796016143169254, 'learning_rate/qwen_vl_interface': 1.8230444747731265e-06, 'learning_rate/action_model': 1.8230444747731264e-05, 'epoch': 1.54}) +[2026-06-16 08:58:20,352][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.0006199198542162776, 'train/grad_norm_pre_clip': 0.20761582255363464, 'timing/data': 0.0193138113245368, 'timing/model': 1.6616290560923517, 'learning_rate/qwen_vl_interface': 1.820936372274857e-06, 'learning_rate/action_model': 1.820936372274857e-05, 'epoch': 1.54}) +[2026-06-16 08:58:26,552][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.4101501703262329, 'train/grad_norm_pre_clip': 18.403583526611328, 'timing/data': 0.02222101204097271, 'timing/model': 1.5485358061268926, 'learning_rate/qwen_vl_interface': 1.8188296793351178e-06, 'learning_rate/action_model': 1.818829679335118e-05, 'epoch': 1.54}) +[2026-06-16 08:58:32,750][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.18724973499774933, 'train/grad_norm_pre_clip': 13.544580459594727, 'timing/data': 0.020615926012396812, 'timing/model': 1.5468956576660275, 'learning_rate/qwen_vl_interface': 1.8167243968198905e-06, 'learning_rate/action_model': 1.8167243968198905e-05, 'epoch': 1.54}) +[2026-06-16 08:58:38,959][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.28291332721710205, 'train/grad_norm_pre_clip': 8.971391677856445, 'timing/data': 0.02059390628710389, 'timing/model': 1.5377165330573916, 'learning_rate/qwen_vl_interface': 1.8146205255945775e-06, 'learning_rate/action_model': 1.8146205255945777e-05, 'epoch': 1.54}) +[2026-06-16 08:58:45,247][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.09819946438074112, 'train/grad_norm_pre_clip': 16.859315872192383, 'timing/data': 0.02195514179766178, 'timing/model': 1.503012292087078, 'learning_rate/qwen_vl_interface': 1.8125180665240007e-06, 'learning_rate/action_model': 1.8125180665240008e-05, 'epoch': 1.54}) +[2026-06-16 08:58:51,750][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.0014447304420173168, 'train/grad_norm_pre_clip': 0.5039671659469604, 'timing/data': 0.020167855080217123, 'timing/model': 1.9549693469889462, 'learning_rate/qwen_vl_interface': 1.8104170204724025e-06, 'learning_rate/action_model': 1.8104170204724024e-05, 'epoch': 1.54}) +[2026-06-16 08:58:57,757][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.01585591956973076, 'train/grad_norm_pre_clip': 24.071077346801758, 'timing/data': 0.020304876379668713, 'timing/model': 1.5580645138397813, 'learning_rate/qwen_vl_interface': 1.8083173883034435e-06, 'learning_rate/action_model': 1.8083173883034435e-05, 'epoch': 1.54}) +[2026-06-16 08:59:04,058][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.12084751576185226, 'train/grad_norm_pre_clip': 16.598485946655273, 'timing/data': 0.02005466492846608, 'timing/model': 1.5094457496888936, 'learning_rate/qwen_vl_interface': 1.8062191708802013e-06, 'learning_rate/action_model': 1.8062191708802012e-05, 'epoch': 1.54}) +[2026-06-16 08:59:10,350][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.008444022387266159, 'train/grad_norm_pre_clip': 10.661903381347656, 'timing/data': 0.020205184817314148, 'timing/model': 1.5466311075724661, 'learning_rate/qwen_vl_interface': 1.804122369065175e-06, 'learning_rate/action_model': 1.8041223690651748e-05, 'epoch': 1.54}) +[2026-06-16 08:59:16,659][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.0013413673732429743, 'train/grad_norm_pre_clip': 0.42459025979042053, 'timing/data': 0.02020778600126505, 'timing/model': 1.5623158239759505, 'learning_rate/qwen_vl_interface': 1.8020269837202806e-06, 'learning_rate/action_model': 1.8020269837202804e-05, 'epoch': 1.54}) +[2026-06-16 08:59:23,045][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.11470452696084976, 'train/grad_norm_pre_clip': 28.276697158813477, 'timing/data': 0.019142170902341604, 'timing/model': 1.6120262499898672, 'learning_rate/qwen_vl_interface': 1.7999330157068517e-06, 'learning_rate/action_model': 1.7999330157068515e-05, 'epoch': 1.54}) +[2026-06-16 08:59:29,289][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.030280746519565582, 'train/grad_norm_pre_clip': 42.63714599609375, 'timing/data': 0.021100590005517006, 'timing/model': 1.7733431640081108, 'learning_rate/qwen_vl_interface': 1.7978404658856393e-06, 'learning_rate/action_model': 1.7978404658856394e-05, 'epoch': 1.54}) +[2026-06-16 08:59:35,696][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.00023681213497184217, 'train/grad_norm_pre_clip': 0.04081204533576965, 'timing/data': 0.02025621710345149, 'timing/model': 1.7943937322124839, 'learning_rate/qwen_vl_interface': 1.7957493351168126e-06, 'learning_rate/action_model': 1.7957493351168126e-05, 'epoch': 1.54}) +[2026-06-16 08:59:41,853][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.0002656960568856448, 'train/grad_norm_pre_clip': 0.05214131250977516, 'timing/data': 0.02051081694662571, 'timing/model': 1.5494532980956137, 'learning_rate/qwen_vl_interface': 1.7936596242599552e-06, 'learning_rate/action_model': 1.7936596242599553e-05, 'epoch': 1.55}) +[2026-06-16 08:59:48,157][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.0025941813364624977, 'train/grad_norm_pre_clip': 1.3452900648117065, 'timing/data': 0.020070555619895458, 'timing/model': 1.5470537547953427, 'learning_rate/qwen_vl_interface': 1.7915713341740692e-06, 'learning_rate/action_model': 1.791571334174069e-05, 'epoch': 1.55}) +[2026-06-16 08:59:54,561][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.0011701638577505946, 'train/grad_norm_pre_clip': 0.25660422444343567, 'timing/data': 0.019739725161343813, 'timing/model': 1.5521512730047107, 'learning_rate/qwen_vl_interface': 1.7894844657175716e-06, 'learning_rate/action_model': 1.7894844657175717e-05, 'epoch': 1.55}) +[2026-06-16 09:00:00,923][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.19968931376934052, 'train/grad_norm_pre_clip': 19.71090316772461, 'timing/data': 0.020543788094073534, 'timing/model': 1.5712958117946982, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 1.55}) +[2026-06-16 09:00:07,270][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.00013202961417846382, 'train/grad_norm_pre_clip': 0.011940882541239262, 'timing/data': 0.020096135791391134, 'timing/model': 1.8639162299223244, 'learning_rate/qwen_vl_interface': 1.7853149971234883e-06, 'learning_rate/action_model': 1.785314997123488e-05, 'epoch': 1.55}) +[2026-06-16 09:00:13,386][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.003483975538983941, 'train/grad_norm_pre_clip': 1.6675357818603516, 'timing/data': 0.01991531578823924, 'timing/model': 1.5834129331633449, 'learning_rate/qwen_vl_interface': 1.7832323986998128e-06, 'learning_rate/action_model': 1.783232398699813e-05, 'epoch': 1.55}) +[2026-06-16 09:00:19,671][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.007872185669839382, 'train/grad_norm_pre_clip': 4.856316566467285, 'timing/data': 0.02030110824853182, 'timing/model': 1.5607612738385797, 'learning_rate/qwen_vl_interface': 1.7811512253333468e-06, 'learning_rate/action_model': 1.7811512253333467e-05, 'epoch': 1.55}) +[2026-06-16 09:00:26,168][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.0019775230903178453, 'train/grad_norm_pre_clip': 0.529360830783844, 'timing/data': 0.020752498880028725, 'timing/model': 1.5550139201804996, 'learning_rate/qwen_vl_interface': 1.7790714778795823e-06, 'learning_rate/action_model': 1.779071477879582e-05, 'epoch': 1.55}) +[2026-06-16 09:00:32,542][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.0016788169741630554, 'train/grad_norm_pre_clip': 1.0818181037902832, 'timing/data': 0.021347220987081528, 'timing/model': 1.5799082019366324, 'learning_rate/qwen_vl_interface': 1.7769931571934246e-06, 'learning_rate/action_model': 1.7769931571934246e-05, 'epoch': 1.55}) +[2026-06-16 09:00:38,889][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.0005315644666552544, 'train/grad_norm_pre_clip': 0.07456528395414352, 'timing/data': 0.02454533288255334, 'timing/model': 1.604981835000217, 'learning_rate/qwen_vl_interface': 1.7749162641291933e-06, 'learning_rate/action_model': 1.7749162641291933e-05, 'epoch': 1.55}) +[2026-06-16 09:00:45,350][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.0018592088017612696, 'train/grad_norm_pre_clip': 0.49203088879585266, 'timing/data': 0.021844192873686552, 'timing/model': 1.740956881083548, 'learning_rate/qwen_vl_interface': 1.7728407995406206e-06, 'learning_rate/action_model': 1.7728407995406205e-05, 'epoch': 1.55}) +[2026-06-16 09:00:51,476][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.09953421354293823, 'train/grad_norm_pre_clip': 20.841188430786133, 'timing/data': 0.0207497994415462, 'timing/model': 1.5989884259179235, 'learning_rate/qwen_vl_interface': 1.7707667642808513e-06, 'learning_rate/action_model': 1.7707667642808512e-05, 'epoch': 1.55}) +[2026-06-16 09:00:57,769][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.16598279774188995, 'train/grad_norm_pre_clip': 15.548585891723633, 'timing/data': 0.0210752310231328, 'timing/model': 1.5564741990529, 'learning_rate/qwen_vl_interface': 1.7686941592024423e-06, 'learning_rate/action_model': 1.7686941592024423e-05, 'epoch': 1.55}) +[2026-06-16 09:01:04,165][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.00023782363859936595, 'train/grad_norm_pre_clip': 0.0184484776109457, 'timing/data': 0.020132347010076046, 'timing/model': 1.5624380498193204, 'learning_rate/qwen_vl_interface': 1.7666229851573662e-06, 'learning_rate/action_model': 1.7666229851573664e-05, 'epoch': 1.55}) +[2026-06-16 09:01:10,465][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.00023994865478016436, 'train/grad_norm_pre_clip': 0.0928695872426033, 'timing/data': 0.020098736975342035, 'timing/model': 1.5266421088017523, 'learning_rate/qwen_vl_interface': 1.7645532429970006e-06, 'learning_rate/action_model': 1.7645532429970004e-05, 'epoch': 1.55}) +[2026-06-16 09:01:16,781][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.10938922315835953, 'train/grad_norm_pre_clip': 16.50895881652832, 'timing/data': 0.02025275770574808, 'timing/model': 1.53542040893808, 'learning_rate/qwen_vl_interface': 1.7624849335721402e-06, 'learning_rate/action_model': 1.7624849335721402e-05, 'epoch': 1.55}) +[2026-06-16 09:01:23,063][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.024580854922533035, 'train/grad_norm_pre_clip': 20.728219985961914, 'timing/data': 0.0206358190625906, 'timing/model': 1.7464644657447934, 'learning_rate/qwen_vl_interface': 1.7604180577329883e-06, 'learning_rate/action_model': 1.7604180577329883e-05, 'epoch': 1.55}) +[2026-06-16 09:01:29,174][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.0013165418058633804, 'train/grad_norm_pre_clip': 0.3377319574356079, 'timing/data': 0.02183761401101947, 'timing/model': 1.5615386217832565, 'learning_rate/qwen_vl_interface': 1.7583526163291604e-06, 'learning_rate/action_model': 1.7583526163291604e-05, 'epoch': 1.55}) +[2026-06-16 09:01:35,469][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.011711983010172844, 'train/grad_norm_pre_clip': 6.999078750610352, 'timing/data': 0.020736809819936752, 'timing/model': 1.5503444839268923, 'learning_rate/qwen_vl_interface': 1.7562886102096807e-06, 'learning_rate/action_model': 1.7562886102096807e-05, 'epoch': 1.55}) +[2026-06-16 09:01:41,872][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.04568934813141823, 'train/grad_norm_pre_clip': 36.2032585144043, 'timing/data': 0.021692723967134953, 'timing/model': 1.570247707888484, 'learning_rate/qwen_vl_interface': 1.7542260402229838e-06, 'learning_rate/action_model': 1.7542260402229838e-05, 'epoch': 1.55}) +[2026-06-16 09:01:48,419][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.23818044364452362, 'train/grad_norm_pre_clip': 4.981975555419922, 'timing/data': 0.021046031266450882, 'timing/model': 1.5894619482569396, 'learning_rate/qwen_vl_interface': 1.7521649072169152e-06, 'learning_rate/action_model': 1.752164907216915e-05, 'epoch': 1.55}) +[2026-06-16 09:01:54,702][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.002230040030553937, 'train/grad_norm_pre_clip': 2.442974805831909, 'timing/data': 0.020069377962499857, 'timing/model': 1.782677510753274, 'learning_rate/qwen_vl_interface': 1.7501052120387285e-06, 'learning_rate/action_model': 1.7501052120387285e-05, 'epoch': 1.55}) +[2026-06-16 09:02:00,767][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.0012023330200463533, 'train/grad_norm_pre_clip': 0.4589290916919708, 'timing/data': 0.02060898020863533, 'timing/model': 1.5528551279567182, 'learning_rate/qwen_vl_interface': 1.7480469555350866e-06, 'learning_rate/action_model': 1.7480469555350865e-05, 'epoch': 1.55}) +[2026-06-16 09:02:07,070][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.000950911664403975, 'train/grad_norm_pre_clip': 0.30079856514930725, 'timing/data': 0.0201698481105268, 'timing/model': 1.554891511797905, 'learning_rate/qwen_vl_interface': 1.7459901385520619e-06, 'learning_rate/action_model': 1.745990138552062e-05, 'epoch': 1.55}) +[2026-06-16 09:02:13,271][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.0010553932515904307, 'train/grad_norm_pre_clip': 0.2723589539527893, 'timing/data': 0.02111010206863284, 'timing/model': 1.5568411648273468, 'learning_rate/qwen_vl_interface': 1.7439347619351324e-06, 'learning_rate/action_model': 1.7439347619351324e-05, 'epoch': 1.55}) +[2026-06-16 09:02:19,669][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.09882792830467224, 'train/grad_norm_pre_clip': 24.08132553100586, 'timing/data': 0.02038820879533887, 'timing/model': 1.572331220842898, 'learning_rate/qwen_vl_interface': 1.741880826529186e-06, 'learning_rate/action_model': 1.741880826529186e-05, 'epoch': 1.56}) +[2026-06-16 09:02:26,008][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.003083050949499011, 'train/grad_norm_pre_clip': 1.1162444353103638, 'timing/data': 0.02099409094080329, 'timing/model': 1.6080586747266352, 'learning_rate/qwen_vl_interface': 1.7398283331785205e-06, 'learning_rate/action_model': 1.7398283331785203e-05, 'epoch': 1.56}) +[2026-06-16 09:02:32,323][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.03649141639471054, 'train/grad_norm_pre_clip': 34.41787338256836, 'timing/data': 0.020129508804529905, 'timing/model': 1.8041558330878615, 'learning_rate/qwen_vl_interface': 1.737777282726835e-06, 'learning_rate/action_model': 1.737777282726835e-05, 'epoch': 1.56}) +[2026-06-16 09:02:38,384][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.0018266537226736546, 'train/grad_norm_pre_clip': 0.5844408273696899, 'timing/data': 0.021862345281988382, 'timing/model': 1.5593865672126412, 'learning_rate/qwen_vl_interface': 1.7357276760172403e-06, 'learning_rate/action_model': 1.7357276760172404e-05, 'epoch': 1.56}) +[2026-06-16 09:02:44,772][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.0005592667730525136, 'train/grad_norm_pre_clip': 0.0767611712217331, 'timing/data': 0.021554564125835896, 'timing/model': 1.5549301048740745, 'learning_rate/qwen_vl_interface': 1.733679513892252e-06, 'learning_rate/action_model': 1.7336795138922517e-05, 'epoch': 1.56}) +[2026-06-16 09:02:51,061][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.14581774175167084, 'train/grad_norm_pre_clip': 53.153892517089844, 'timing/data': 0.019751866813749075, 'timing/model': 1.5483851861208677, 'learning_rate/qwen_vl_interface': 1.7316327971937926e-06, 'learning_rate/action_model': 1.7316327971937926e-05, 'epoch': 1.56}) +[2026-06-16 09:02:57,409][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.0023950119502842426, 'train/grad_norm_pre_clip': 1.4642012119293213, 'timing/data': 0.02090106112882495, 'timing/model': 1.594000609125942, 'learning_rate/qwen_vl_interface': 1.7295875267631904e-06, 'learning_rate/action_model': 1.7295875267631904e-05, 'epoch': 1.56}) +[2026-06-16 09:03:03,761][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.0003246032865718007, 'train/grad_norm_pre_clip': 0.039484575390815735, 'timing/data': 0.021461782976984978, 'timing/model': 1.6067773941904306, 'learning_rate/qwen_vl_interface': 1.7275437034411778e-06, 'learning_rate/action_model': 1.7275437034411776e-05, 'epoch': 1.56}) +[2026-06-16 09:03:10,119][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.10652708262205124, 'train/grad_norm_pre_clip': 25.83541488647461, 'timing/data': 0.021413233131170273, 'timing/model': 1.8058666470460594, 'learning_rate/qwen_vl_interface': 1.7255013280678941e-06, 'learning_rate/action_model': 1.725501328067894e-05, 'epoch': 1.56}) +[2026-06-16 09:03:16,268][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.00881869811564684, 'train/grad_norm_pre_clip': 9.181605339050293, 'timing/data': 0.021946175955235958, 'timing/model': 1.5516146086156368, 'learning_rate/qwen_vl_interface': 1.7234604014828824e-06, 'learning_rate/action_model': 1.7234604014828823e-05, 'epoch': 1.56}) +[2026-06-16 09:03:22,638][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.0005407438729889691, 'train/grad_norm_pre_clip': 0.15363003313541412, 'timing/data': 0.02136453427374363, 'timing/model': 1.6189508670940995, 'learning_rate/qwen_vl_interface': 1.7214209245250906e-06, 'learning_rate/action_model': 1.7214209245250905e-05, 'epoch': 1.56}) +[2026-06-16 09:03:29,068][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.010024399496614933, 'train/grad_norm_pre_clip': 5.316384315490723, 'timing/data': 0.019202584866434336, 'timing/model': 1.5510862269438803, 'learning_rate/qwen_vl_interface': 1.7193828980328723e-06, 'learning_rate/action_model': 1.719382898032872e-05, 'epoch': 1.56}) +[2026-06-16 09:03:35,473][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.015827840194106102, 'train/grad_norm_pre_clip': 6.539095878601074, 'timing/data': 0.022392827086150646, 'timing/model': 1.64415770675987, 'learning_rate/qwen_vl_interface': 1.71734632284398e-06, 'learning_rate/action_model': 1.7173463228439798e-05, 'epoch': 1.56}) +[2026-06-16 09:03:41,867][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.01412245910614729, 'train/grad_norm_pre_clip': 16.218427658081055, 'timing/data': 0.022073566913604736, 'timing/model': 1.5893520461395383, 'learning_rate/qwen_vl_interface': 1.7153111997955744e-06, 'learning_rate/action_model': 1.7153111997955743e-05, 'epoch': 1.56}) +[2026-06-16 09:03:48,257][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.078444704413414, 'train/grad_norm_pre_clip': 26.358556747436523, 'timing/data': 0.020820672158151865, 'timing/model': 1.8434359589591622, 'learning_rate/qwen_vl_interface': 1.7132775297242175e-06, 'learning_rate/action_model': 1.7132775297242172e-05, 'epoch': 1.56}) +[2026-06-16 09:03:54,364][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.0013588258298113942, 'train/grad_norm_pre_clip': 0.5490291118621826, 'timing/data': 0.01897599408403039, 'timing/model': 1.5361114768311381, 'learning_rate/qwen_vl_interface': 1.7112453134658747e-06, 'learning_rate/action_model': 1.7112453134658747e-05, 'epoch': 1.56}) +[2026-06-16 09:04:00,677][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.15631729364395142, 'train/grad_norm_pre_clip': 38.77579116821289, 'timing/data': 0.020875901915133, 'timing/model': 1.5647070240229368, 'learning_rate/qwen_vl_interface': 1.7092145518559129e-06, 'learning_rate/action_model': 1.709214551855913e-05, 'epoch': 1.56}) +[2026-06-16 09:04:06,970][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.08764400333166122, 'train/grad_norm_pre_clip': 17.60639190673828, 'timing/data': 0.021138963289558887, 'timing/model': 1.5574341397732496, 'learning_rate/qwen_vl_interface': 1.7071852457291013e-06, 'learning_rate/action_model': 1.707185245729101e-05, 'epoch': 1.56}) +[2026-06-16 09:04:13,265][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.039312731474637985, 'train/grad_norm_pre_clip': 23.34040641784668, 'timing/data': 0.020860271994024515, 'timing/model': 1.5393029530532658, 'learning_rate/qwen_vl_interface': 1.7051573959196122e-06, 'learning_rate/action_model': 1.7051573959196122e-05, 'epoch': 1.56}) +[2026-06-16 09:04:19,569][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.06699348986148834, 'train/grad_norm_pre_clip': 51.075416564941406, 'timing/data': 0.020191019866615534, 'timing/model': 1.5234676753170788, 'learning_rate/qwen_vl_interface': 1.7031310032610166e-06, 'learning_rate/action_model': 1.7031310032610163e-05, 'epoch': 1.56}) +[2026-06-16 09:04:25,900][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.0007370322127826512, 'train/grad_norm_pre_clip': 0.10836239904165268, 'timing/data': 0.019842257723212242, 'timing/model': 1.7894861092790961, 'learning_rate/qwen_vl_interface': 1.701106068586289e-06, 'learning_rate/action_model': 1.701106068586289e-05, 'epoch': 1.56}) +[2026-06-16 09:04:31,963][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.0023230428341776133, 'train/grad_norm_pre_clip': 0.6124252080917358, 'timing/data': 0.020328930113464594, 'timing/model': 1.554649277124554, 'learning_rate/qwen_vl_interface': 1.6990825927278032e-06, 'learning_rate/action_model': 1.699082592727803e-05, 'epoch': 1.56}) +[2026-06-16 09:04:38,358][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.0003749788156710565, 'train/grad_norm_pre_clip': 0.0821327492594719, 'timing/data': 0.020526340696960688, 'timing/model': 1.5907933339476585, 'learning_rate/qwen_vl_interface': 1.6970605765173344e-06, 'learning_rate/action_model': 1.6970605765173346e-05, 'epoch': 1.56}) +[2026-06-16 09:04:44,766][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.0007157152867875993, 'train/grad_norm_pre_clip': 0.26746314764022827, 'timing/data': 0.02047315100207925, 'timing/model': 1.5927066151052713, 'learning_rate/qwen_vl_interface': 1.6950400207860557e-06, 'learning_rate/action_model': 1.6950400207860557e-05, 'epoch': 1.56}) +[2026-06-16 09:04:51,159][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.000657362979836762, 'train/grad_norm_pre_clip': 0.11436692625284195, 'timing/data': 0.019745048135519028, 'timing/model': 1.5494230287149549, 'learning_rate/qwen_vl_interface': 1.6930209263645412e-06, 'learning_rate/action_model': 1.6930209263645412e-05, 'epoch': 1.56}) +[2026-06-16 09:04:57,474][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.18985861539840698, 'train/grad_norm_pre_clip': 9.227324485778809, 'timing/data': 0.020178699865937233, 'timing/model': 1.5267179631628096, 'learning_rate/qwen_vl_interface': 1.6910032940827656e-06, 'learning_rate/action_model': 1.6910032940827657e-05, 'epoch': 1.57}) +[2026-06-16 09:05:03,812][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.04628937691450119, 'train/grad_norm_pre_clip': 31.108539581298828, 'timing/data': 0.019191026221960783, 'timing/model': 1.7953397310338914, 'learning_rate/qwen_vl_interface': 1.6889871247701003e-06, 'learning_rate/action_model': 1.6889871247701002e-05, 'epoch': 1.57}) +[2026-06-16 09:05:09,917][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.2924608588218689, 'train/grad_norm_pre_clip': 23.437198638916016, 'timing/data': 0.0270638270303607, 'timing/model': 1.5844534998759627, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 1.57}) +[2026-06-16 09:05:16,268][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.04917850345373154, 'train/grad_norm_pre_clip': 46.166683197021484, 'timing/data': 0.02064437186345458, 'timing/model': 1.5508776209317148, 'learning_rate/qwen_vl_interface': 1.6849591783665847e-06, 'learning_rate/action_model': 1.6849591783665847e-05, 'epoch': 1.57}) +[2026-06-16 09:05:22,571][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.37054207921028137, 'train/grad_norm_pre_clip': 29.710243225097656, 'timing/data': 0.02134829480201006, 'timing/model': 1.5478633628226817, 'learning_rate/qwen_vl_interface': 1.6829474029314713e-06, 'learning_rate/action_model': 1.6829474029314713e-05, 'epoch': 1.57}) +[2026-06-16 09:05:29,093][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.028255779296159744, 'train/grad_norm_pre_clip': 19.98825454711914, 'timing/data': 0.026988856960088015, 'timing/model': 1.5506981192156672, 'learning_rate/qwen_vl_interface': 1.6809370937769408e-06, 'learning_rate/action_model': 1.680937093776941e-05, 'epoch': 1.57}) +[2026-06-16 09:05:35,450][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.2885138988494873, 'train/grad_norm_pre_clip': 20.732730865478516, 'timing/data': 0.020967022981494665, 'timing/model': 1.5938095529563725, 'learning_rate/qwen_vl_interface': 1.678928251729356e-06, 'learning_rate/action_model': 1.6789282517293557e-05, 'epoch': 1.57}) +[2026-06-16 09:05:41,791][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.05424243211746216, 'train/grad_norm_pre_clip': 42.64888000488281, 'timing/data': 0.019643127918243408, 'timing/model': 1.7940476527437568, 'learning_rate/qwen_vl_interface': 1.6769208776144767e-06, 'learning_rate/action_model': 1.6769208776144764e-05, 'epoch': 1.57}) +[2026-06-16 09:05:48,023][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.23448403179645538, 'train/grad_norm_pre_clip': 4.106137752532959, 'timing/data': 0.020912223029881716, 'timing/model': 1.7076727650128305, 'learning_rate/qwen_vl_interface': 1.6749149722574564e-06, 'learning_rate/action_model': 1.6749149722574564e-05, 'epoch': 1.57}) +[2026-06-16 09:05:54,161][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.16149233281612396, 'train/grad_norm_pre_clip': 14.096151351928711, 'timing/data': 0.023663065396249294, 'timing/model': 1.6442338447086513, 'learning_rate/qwen_vl_interface': 1.6729105364828486e-06, 'learning_rate/action_model': 1.6729105364828484e-05, 'epoch': 1.57}) +[2026-06-16 09:06:00,484][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.027521489188075066, 'train/grad_norm_pre_clip': 31.893495559692383, 'timing/data': 0.02105884440243244, 'timing/model': 1.5568989110179245, 'learning_rate/qwen_vl_interface': 1.6709075711146e-06, 'learning_rate/action_model': 1.6709075711146e-05, 'epoch': 1.57}) +[2026-06-16 09:06:06,875][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.0024177967570722103, 'train/grad_norm_pre_clip': 1.1768709421157837, 'timing/data': 0.02269922010600567, 'timing/model': 1.5697783874347806, 'learning_rate/qwen_vl_interface': 1.6689060769760556e-06, 'learning_rate/action_model': 1.6689060769760557e-05, 'epoch': 1.57}) +[2026-06-16 09:06:13,183][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.04973206669092178, 'train/grad_norm_pre_clip': 26.65741729736328, 'timing/data': 0.020219300873577595, 'timing/model': 1.5362092568539083, 'learning_rate/qwen_vl_interface': 1.6669060548899528e-06, 'learning_rate/action_model': 1.6669060548899525e-05, 'epoch': 1.57}) +[2026-06-16 09:06:19,504][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.002460853196680546, 'train/grad_norm_pre_clip': 0.6507894396781921, 'timing/data': 0.023172233253717422, 'timing/model': 1.7884004260413349, 'learning_rate/qwen_vl_interface': 1.6649075056784253e-06, 'learning_rate/action_model': 1.6649075056784253e-05, 'epoch': 1.57}) +[2026-06-16 09:06:25,905][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.0005083458381704986, 'train/grad_norm_pre_clip': 0.08178737014532089, 'timing/data': 0.023625934962183237, 'timing/model': 1.7870535049587488, 'learning_rate/qwen_vl_interface': 1.6629104301630019e-06, 'learning_rate/action_model': 1.662910430163002e-05, 'epoch': 1.57}) +[2026-06-16 09:06:32,128][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.0702703446149826, 'train/grad_norm_pre_clip': 24.416584014892578, 'timing/data': 0.020440562162548304, 'timing/model': 1.607893972657621, 'learning_rate/qwen_vl_interface': 1.6609148291646043e-06, 'learning_rate/action_model': 1.6609148291646042e-05, 'epoch': 1.57}) +[2026-06-16 09:06:38,375][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.07167822867631912, 'train/grad_norm_pre_clip': 18.231870651245117, 'timing/data': 0.020284471102058887, 'timing/model': 1.5482239588163793, 'learning_rate/qwen_vl_interface': 1.658920703503549e-06, 'learning_rate/action_model': 1.6589207035035488e-05, 'epoch': 1.57}) +[2026-06-16 09:06:44,678][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.14366647601127625, 'train/grad_norm_pre_clip': 23.3876953125, 'timing/data': 0.023880945052951574, 'timing/model': 1.5417600059881806, 'learning_rate/qwen_vl_interface': 1.6569280539995467e-06, 'learning_rate/action_model': 1.6569280539995467e-05, 'epoch': 1.57}) +[2026-06-16 09:06:51,149][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.0013934760354459286, 'train/grad_norm_pre_clip': 0.4111160933971405, 'timing/data': 0.020371512044221163, 'timing/model': 1.595164233352989, 'learning_rate/qwen_vl_interface': 1.6549368814716977e-06, 'learning_rate/action_model': 1.6549368814716977e-05, 'epoch': 1.57}) +[2026-06-16 09:06:57,507][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.008691645227372646, 'train/grad_norm_pre_clip': 3.6526451110839844, 'timing/data': 0.02045683190226555, 'timing/model': 1.7973656356334686, 'learning_rate/qwen_vl_interface': 1.6529471867385009e-06, 'learning_rate/action_model': 1.6529471867385007e-05, 'epoch': 1.57}) +[2026-06-16 09:07:03,575][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.0028928606770932674, 'train/grad_norm_pre_clip': 2.208451986312866, 'timing/data': 0.020564781967550516, 'timing/model': 1.5610631727613509, 'learning_rate/qwen_vl_interface': 1.6509589706178417e-06, 'learning_rate/action_model': 1.6509589706178417e-05, 'epoch': 1.57}) +[2026-06-16 09:07:09,897][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.06527312099933624, 'train/grad_norm_pre_clip': 13.371541976928711, 'timing/data': 0.02172328718006611, 'timing/model': 1.5821793698705733, 'learning_rate/qwen_vl_interface': 1.648972233927002e-06, 'learning_rate/action_model': 1.648972233927002e-05, 'epoch': 1.57}) +[2026-06-16 09:07:16,163][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.15733124315738678, 'train/grad_norm_pre_clip': 16.888242721557617, 'timing/data': 0.02021003095433116, 'timing/model': 1.5391000709496439, 'learning_rate/qwen_vl_interface': 1.646986977482654e-06, 'learning_rate/action_model': 1.6469869774826538e-05, 'epoch': 1.57}) +[2026-06-16 09:07:22,571][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.04800059273838997, 'train/grad_norm_pre_clip': 23.656208038330078, 'timing/data': 0.019087936263531446, 'timing/model': 1.5388720538467169, 'learning_rate/qwen_vl_interface': 1.6450032021008618e-06, 'learning_rate/action_model': 1.6450032021008617e-05, 'epoch': 1.57}) +[2026-06-16 09:07:28,873][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.0031970080453902483, 'train/grad_norm_pre_clip': 0.8991293907165527, 'timing/data': 0.0217922767624259, 'timing/model': 1.579295578878373, 'learning_rate/qwen_vl_interface': 1.6430209085970812e-06, 'learning_rate/action_model': 1.643020908597081e-05, 'epoch': 1.58}) +[2026-06-16 09:07:35,198][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.13520683348178864, 'train/grad_norm_pre_clip': 7.6616058349609375, 'timing/data': 0.020920863840728998, 'timing/model': 1.778755501843989, 'learning_rate/qwen_vl_interface': 1.641040097786157e-06, 'learning_rate/action_model': 1.641040097786157e-05, 'epoch': 1.58}) +[2026-06-16 09:07:41,268][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.24089734256267548, 'train/grad_norm_pre_clip': 21.975841522216797, 'timing/data': 0.02105189487338066, 'timing/model': 1.5507779498584569, 'learning_rate/qwen_vl_interface': 1.6390607704823267e-06, 'learning_rate/action_model': 1.6390607704823265e-05, 'epoch': 1.58}) +[2026-06-16 09:07:47,854][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.18692824244499207, 'train/grad_norm_pre_clip': 24.978633880615234, 'timing/data': 0.021822467911988497, 'timing/model': 1.737034282181412, 'learning_rate/qwen_vl_interface': 1.6370829274992162e-06, 'learning_rate/action_model': 1.6370829274992162e-05, 'epoch': 1.58}) +[2026-06-16 09:07:54,308][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.0010271741775795817, 'train/grad_norm_pre_clip': 0.1732696294784546, 'timing/data': 0.02829870395362377, 'timing/model': 1.5644849818199873, 'learning_rate/qwen_vl_interface': 1.6351065696498422e-06, 'learning_rate/action_model': 1.6351065696498422e-05, 'epoch': 1.58}) +[2026-06-16 09:18:48,816][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.006497563794255257, 'train/grad_norm_pre_clip': 3.2769975662231445, 'eval/loss': 0.031478818506002426, 'eval/flap_precision': 0.9701492537313433, 'eval/flap_recall': 0.975443383356071, 'eval/flap_f1': 0.9727891156462585, 'eval/flap_rate_model': 0.20472222222222222, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9727891156462586, 'timing/data': 0.021725227124989033, 'timing/model': 1.5905409096740186, 'learning_rate/qwen_vl_interface': 1.6331316977466117e-06, 'learning_rate/action_model': 1.6331316977466115e-05, 'epoch': 1.58}) +[2026-06-16 09:18:48,817][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3900_state +[2026-06-16 09:18:48,823][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 09:18:57,383][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3900_state/model.safetensors +[2026-06-16 09:19:17,597][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3900_state/optimizer.bin +[2026-06-16 09:19:17,598][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3900_state/sampler.bin +[2026-06-16 09:19:17,598][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3900_state/sampler_1.bin +[2026-06-16 09:19:17,600][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_3900_state/random_states_0.pkl +[2026-06-16 09:19:17,792][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3900 +[2026-06-16 09:19:17,792][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 09:19:17,808][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 09:19:24,040][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.025667037814855576, 'train/grad_norm_pre_clip': 17.079727172851562, 'timing/data': 0.02051254129037261, 'timing/model': 1.5257683899253607, 'learning_rate/qwen_vl_interface': 1.6311583126013198e-06, 'learning_rate/action_model': 1.6311583126013198e-05, 'epoch': 1.58}) +[2026-06-16 09:19:30,237][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.016216721385717392, 'train/grad_norm_pre_clip': 14.046712875366211, 'timing/data': 0.0314700766466558, 'timing/model': 1.5117017650045455, 'learning_rate/qwen_vl_interface': 1.6291864150251488e-06, 'learning_rate/action_model': 1.629186415025149e-05, 'epoch': 1.58}) +[2026-06-16 09:19:36,475][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.00697678979486227, 'train/grad_norm_pre_clip': 7.460887908935547, 'timing/data': 0.020767943002283573, 'timing/model': 1.7970087202265859, 'learning_rate/qwen_vl_interface': 1.6272160058286727e-06, 'learning_rate/action_model': 1.6272160058286726e-05, 'epoch': 1.58}) +[2026-06-16 09:19:42,763][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.0034198928624391556, 'train/grad_norm_pre_clip': 0.7212235331535339, 'timing/data': 0.021841321140527725, 'timing/model': 1.6893108692020178, 'learning_rate/qwen_vl_interface': 1.6252470858218519e-06, 'learning_rate/action_model': 1.6252470858218517e-05, 'epoch': 1.58}) +[2026-06-16 09:19:48,747][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.25306588411331177, 'train/grad_norm_pre_clip': 2.6642582416534424, 'timing/data': 0.020317406859248877, 'timing/model': 1.5629855589941144, 'learning_rate/qwen_vl_interface': 1.6232796558140357e-06, 'learning_rate/action_model': 1.6232796558140357e-05, 'epoch': 1.58}) +[2026-06-16 09:19:54,949][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.002717006718739867, 'train/grad_norm_pre_clip': 0.9272621870040894, 'timing/data': 0.020137473940849304, 'timing/model': 1.5565856979228556, 'learning_rate/qwen_vl_interface': 1.6213137166139592e-06, 'learning_rate/action_model': 1.621313716613959e-05, 'epoch': 1.58}) +[2026-06-16 09:20:01,152][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.0026816779281944036, 'train/grad_norm_pre_clip': 0.39837998151779175, 'timing/data': 0.021291083190590143, 'timing/model': 1.5519960508681834, 'learning_rate/qwen_vl_interface': 1.6193492690297463e-06, 'learning_rate/action_model': 1.6193492690297464e-05, 'epoch': 1.58}) +[2026-06-16 09:20:07,446][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.001261338940821588, 'train/grad_norm_pre_clip': 0.15157213807106018, 'timing/data': 0.021055479999631643, 'timing/model': 1.5088308351114392, 'learning_rate/qwen_vl_interface': 1.6173863138689078e-06, 'learning_rate/action_model': 1.617386313868908e-05, 'epoch': 1.58}) +[2026-06-16 09:20:13,666][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.0012400448322296143, 'train/grad_norm_pre_clip': 0.2107703983783722, 'timing/data': 0.021763944998383522, 'timing/model': 1.5322354207746685, 'learning_rate/qwen_vl_interface': 1.6154248519383396e-06, 'learning_rate/action_model': 1.6154248519383395e-05, 'epoch': 1.58}) +[2026-06-16 09:20:19,984][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.000940723461098969, 'train/grad_norm_pre_clip': 0.18692025542259216, 'timing/data': 0.02242112997919321, 'timing/model': 1.780145918019116, 'learning_rate/qwen_vl_interface': 1.6134648840443248e-06, 'learning_rate/action_model': 1.613464884044325e-05, 'epoch': 1.58}) +[2026-06-16 09:20:26,138][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.14879150688648224, 'train/grad_norm_pre_clip': 7.413321018218994, 'timing/data': 0.02205281611531973, 'timing/model': 1.630525424145162, 'learning_rate/qwen_vl_interface': 1.6115064109925334e-06, 'learning_rate/action_model': 1.6115064109925336e-05, 'epoch': 1.58}) +[2026-06-16 09:20:32,361][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.0013682101853191853, 'train/grad_norm_pre_clip': 0.325535386800766, 'timing/data': 0.024968248791992664, 'timing/model': 1.5514786336570978, 'learning_rate/qwen_vl_interface': 1.6095494335880173e-06, 'learning_rate/action_model': 1.6095494335880174e-05, 'epoch': 1.58}) +[2026-06-16 09:20:38,684][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.0016754321986809373, 'train/grad_norm_pre_clip': 0.7605917453765869, 'timing/data': 0.02084936434403062, 'timing/model': 1.5844558756798506, 'learning_rate/qwen_vl_interface': 1.6075939526352175e-06, 'learning_rate/action_model': 1.6075939526352175e-05, 'epoch': 1.58}) +[2026-06-16 09:20:44,955][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.05197504162788391, 'train/grad_norm_pre_clip': 22.96615219116211, 'timing/data': 0.022234205156564713, 'timing/model': 1.5224190377630293, 'learning_rate/qwen_vl_interface': 1.6056399689379576e-06, 'learning_rate/action_model': 1.6056399689379575e-05, 'epoch': 1.58}) +[2026-06-16 09:20:51,346][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.027458902448415756, 'train/grad_norm_pre_clip': 15.094193458557129, 'timing/data': 0.021740880329161882, 'timing/model': 1.6194468601606786, 'learning_rate/qwen_vl_interface': 1.6036874832994473e-06, 'learning_rate/action_model': 1.603687483299447e-05, 'epoch': 1.58}) +[2026-06-16 09:20:57,690][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.001418944215402007, 'train/grad_norm_pre_clip': 0.24475550651550293, 'timing/data': 0.021124353632330894, 'timing/model': 1.8587635760195553, 'learning_rate/qwen_vl_interface': 1.6017364965222796e-06, 'learning_rate/action_model': 1.6017364965222793e-05, 'epoch': 1.58}) +[2026-06-16 09:21:03,952][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.11723251640796661, 'train/grad_norm_pre_clip': 8.480052947998047, 'timing/data': 0.021413704846054316, 'timing/model': 1.5810924302786589, 'learning_rate/qwen_vl_interface': 1.5997870094084305e-06, 'learning_rate/action_model': 1.5997870094084306e-05, 'epoch': 1.58}) +[2026-06-16 09:21:10,254][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.11122605204582214, 'train/grad_norm_pre_clip': 23.00970458984375, 'timing/data': 0.022406102623790503, 'timing/model': 1.5357812498696148, 'learning_rate/qwen_vl_interface': 1.5978390227592616e-06, 'learning_rate/action_model': 1.5978390227592615e-05, 'epoch': 1.58}) +[2026-06-16 09:21:16,701][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.0013089848216623068, 'train/grad_norm_pre_clip': 0.1718018352985382, 'timing/data': 0.026510724797844887, 'timing/model': 1.5862387800589204, 'learning_rate/qwen_vl_interface': 1.5958925373755164e-06, 'learning_rate/action_model': 1.5958925373755164e-05, 'epoch': 1.58}) +[2026-06-16 09:21:23,177][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.25583815574645996, 'train/grad_norm_pre_clip': 17.05107879638672, 'timing/data': 0.021099461242556572, 'timing/model': 1.5325653972104192, 'learning_rate/qwen_vl_interface': 1.5939475540573213e-06, 'learning_rate/action_model': 1.5939475540573212e-05, 'epoch': 1.59}) +[2026-06-16 09:21:29,531][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.0010707444744184613, 'train/grad_norm_pre_clip': 0.19580437242984772, 'timing/data': 0.020669776014983654, 'timing/model': 1.5966503852978349, 'learning_rate/qwen_vl_interface': 1.5920040736041867e-06, 'learning_rate/action_model': 1.5920040736041866e-05, 'epoch': 1.59}) +[2026-06-16 09:21:35,805][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.01726970449090004, 'train/grad_norm_pre_clip': 12.797212600708008, 'timing/data': 0.02129728114232421, 'timing/model': 1.7852354971691966, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 1.59}) +[2026-06-16 09:21:41,865][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.0006111768889240921, 'train/grad_norm_pre_clip': 0.11307818442583084, 'timing/data': 0.02110431808978319, 'timing/model': 1.5537159708328545, 'learning_rate/qwen_vl_interface': 1.5881216244880407e-06, 'learning_rate/action_model': 1.5881216244880407e-05, 'epoch': 1.59}) +[2026-06-16 09:21:48,159][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.00208524358458817, 'train/grad_norm_pre_clip': 1.3110575675964355, 'timing/data': 0.020307801198214293, 'timing/model': 1.551855101250112, 'learning_rate/qwen_vl_interface': 1.5861826574209587e-06, 'learning_rate/action_model': 1.5861826574209587e-05, 'epoch': 1.59}) +[2026-06-16 09:21:54,456][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.0006913365796208382, 'train/grad_norm_pre_clip': 0.10098690539598465, 'timing/data': 0.02042618067935109, 'timing/model': 1.5328557561151683, 'learning_rate/qwen_vl_interface': 1.5842451964107916e-06, 'learning_rate/action_model': 1.5842451964107916e-05, 'epoch': 1.59}) +[2026-06-16 09:22:00,769][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.0006671671289950609, 'train/grad_norm_pre_clip': 0.09396325796842575, 'timing/data': 0.021517399232834578, 'timing/model': 1.5562480478547513, 'learning_rate/qwen_vl_interface': 1.5823092422539567e-06, 'learning_rate/action_model': 1.582309242253957e-05, 'epoch': 1.59}) +[2026-06-16 09:22:07,066][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.08699396252632141, 'train/grad_norm_pre_clip': 33.974998474121094, 'timing/data': 0.02100168401375413, 'timing/model': 1.5225403308868408, 'learning_rate/qwen_vl_interface': 1.5803747957462514e-06, 'learning_rate/action_model': 1.5803747957462514e-05, 'epoch': 1.59}) +[2026-06-16 09:22:13,400][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.0006220238283276558, 'train/grad_norm_pre_clip': 0.05472799763083458, 'timing/data': 0.020576481241732836, 'timing/model': 1.7846476240083575, 'learning_rate/qwen_vl_interface': 1.5784418576828541e-06, 'learning_rate/action_model': 1.5784418576828543e-05, 'epoch': 1.59}) +[2026-06-16 09:22:19,487][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.005111628212034702, 'train/grad_norm_pre_clip': 2.9383819103240967, 'timing/data': 0.018799856305122375, 'timing/model': 1.5672846478410065, 'learning_rate/qwen_vl_interface': 1.576510428858322e-06, 'learning_rate/action_model': 1.5765104288583217e-05, 'epoch': 1.59}) +[2026-06-16 09:22:25,869][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.012119321152567863, 'train/grad_norm_pre_clip': 6.780371189117432, 'timing/data': 0.020095644984394312, 'timing/model': 1.5647070370614529, 'learning_rate/qwen_vl_interface': 1.5745805100665922e-06, 'learning_rate/action_model': 1.5745805100665923e-05, 'epoch': 1.59}) +[2026-06-16 09:22:32,173][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.002240227535367012, 'train/grad_norm_pre_clip': 0.6594289541244507, 'timing/data': 0.021653286647051573, 'timing/model': 1.547158321365714, 'learning_rate/qwen_vl_interface': 1.572652102100983e-06, 'learning_rate/action_model': 1.572652102100983e-05, 'epoch': 1.59}) +[2026-06-16 09:22:38,623][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.001098492182791233, 'train/grad_norm_pre_clip': 0.14803773164749146, 'timing/data': 0.021563454996794462, 'timing/model': 1.7064124410972, 'learning_rate/qwen_vl_interface': 1.5707252057541898e-06, 'learning_rate/action_model': 1.5707252057541896e-05, 'epoch': 1.59}) +[2026-06-16 09:22:44,922][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.012968233786523342, 'train/grad_norm_pre_clip': 10.5130615234375, 'timing/data': 0.020521226804703474, 'timing/model': 1.5725845051929355, 'learning_rate/qwen_vl_interface': 1.5687998218182849e-06, 'learning_rate/action_model': 1.5687998218182847e-05, 'epoch': 1.59}) +[2026-06-16 09:22:51,304][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.04463975504040718, 'train/grad_norm_pre_clip': 19.98881721496582, 'timing/data': 0.020908358972519636, 'timing/model': 1.8142071631737053, 'learning_rate/qwen_vl_interface': 1.5668759510847235e-06, 'learning_rate/action_model': 1.5668759510847233e-05, 'epoch': 1.59}) +[2026-06-16 09:22:57,371][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.0003308356099296361, 'train/grad_norm_pre_clip': 0.0293691698461771, 'timing/data': 0.02300897380337119, 'timing/model': 1.547709729988128, 'learning_rate/qwen_vl_interface': 1.5649535943443331e-06, 'learning_rate/action_model': 1.564953594344333e-05, 'epoch': 1.59}) +[2026-06-16 09:23:03,674][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.13997232913970947, 'train/grad_norm_pre_clip': 10.613122940063477, 'timing/data': 0.021913915872573853, 'timing/model': 1.5547835063189268, 'learning_rate/qwen_vl_interface': 1.5630327523873244e-06, 'learning_rate/action_model': 1.563032752387324e-05, 'epoch': 1.59}) +[2026-06-16 09:23:09,971][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.00048061239067465067, 'train/grad_norm_pre_clip': 0.11326870322227478, 'timing/data': 0.021900255233049393, 'timing/model': 1.547139462083578, 'learning_rate/qwen_vl_interface': 1.561113426003282e-06, 'learning_rate/action_model': 1.561113426003282e-05, 'epoch': 1.59}) +[2026-06-16 09:23:16,321][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.002312283730134368, 'train/grad_norm_pre_clip': 0.8767020106315613, 'timing/data': 0.020691784098744392, 'timing/model': 1.607062414754182, 'learning_rate/qwen_vl_interface': 1.5591956159811686e-06, 'learning_rate/action_model': 1.5591956159811687e-05, 'epoch': 1.59}) +[2026-06-16 09:23:22,578][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.011968249455094337, 'train/grad_norm_pre_clip': 15.301661491394043, 'timing/data': 0.02135949907824397, 'timing/model': 1.5279168272390962, 'learning_rate/qwen_vl_interface': 1.5572793231093235e-06, 'learning_rate/action_model': 1.5572793231093236e-05, 'epoch': 1.59}) +[2026-06-16 09:23:28,903][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.08753184229135513, 'train/grad_norm_pre_clip': 39.168983459472656, 'timing/data': 0.022010592743754387, 'timing/model': 1.7864565243944526, 'learning_rate/qwen_vl_interface': 1.5553645481754626e-06, 'learning_rate/action_model': 1.5553645481754626e-05, 'epoch': 1.59}) +[2026-06-16 09:23:34,982][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.004356949124485254, 'train/grad_norm_pre_clip': 1.552103042602539, 'timing/data': 0.0214793779887259, 'timing/model': 1.5472465152852237, 'learning_rate/qwen_vl_interface': 1.5534512919666775e-06, 'learning_rate/action_model': 1.5534512919666776e-05, 'epoch': 1.59}) +[2026-06-16 09:23:41,364][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.16996179521083832, 'train/grad_norm_pre_clip': 15.394048690795898, 'timing/data': 0.020991035271435976, 'timing/model': 1.55672677885741, 'learning_rate/qwen_vl_interface': 1.551539555269435e-06, 'learning_rate/action_model': 1.5515395552694348e-05, 'epoch': 1.59}) +[2026-06-16 09:23:47,666][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.0003979528555646539, 'train/grad_norm_pre_clip': 0.09187000244855881, 'timing/data': 0.021432026755064726, 'timing/model': 1.5544250486418605, 'learning_rate/qwen_vl_interface': 1.5496293388695787e-06, 'learning_rate/action_model': 1.5496293388695786e-05, 'epoch': 1.59}) +[2026-06-16 09:23:54,068][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.10535836219787598, 'train/grad_norm_pre_clip': 13.715519905090332, 'timing/data': 0.021804688964039087, 'timing/model': 1.5572217958979309, 'learning_rate/qwen_vl_interface': 1.547720643552327e-06, 'learning_rate/action_model': 1.547720643552327e-05, 'epoch': 1.59}) +[2026-06-16 09:24:00,639][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.16245459020137787, 'train/grad_norm_pre_clip': 10.705636978149414, 'timing/data': 0.021621836815029383, 'timing/model': 1.585964831057936, 'learning_rate/qwen_vl_interface': 1.5458134701022706e-06, 'learning_rate/action_model': 1.5458134701022705e-05, 'epoch': 1.6}) +[2026-06-16 09:24:07,018][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.0004981489037163556, 'train/grad_norm_pre_clip': 0.08987346291542053, 'timing/data': 0.02118267398327589, 'timing/model': 1.7926390878856182, 'learning_rate/qwen_vl_interface': 1.5439078193033765e-06, 'learning_rate/action_model': 1.5439078193033766e-05, 'epoch': 1.6}) +[2026-06-16 09:24:13,251][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.07335515320301056, 'train/grad_norm_pre_clip': 26.824329376220703, 'timing/data': 0.022214919794350863, 'timing/model': 1.7343181688338518, 'learning_rate/qwen_vl_interface': 1.5420036919389872e-06, 'learning_rate/action_model': 1.5420036919389873e-05, 'epoch': 1.6}) +[2026-06-16 09:24:19,365][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.00024116407439578325, 'train/grad_norm_pre_clip': 0.016649741679430008, 'timing/data': 0.02273496426641941, 'timing/model': 1.5517295082099736, 'learning_rate/qwen_vl_interface': 1.540101088791817e-06, 'learning_rate/action_model': 1.5401010887918168e-05, 'epoch': 1.6}) +[2026-06-16 09:24:25,671][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.24980862438678741, 'train/grad_norm_pre_clip': 22.121706008911133, 'timing/data': 0.02134115295484662, 'timing/model': 1.5582477450370789, 'learning_rate/qwen_vl_interface': 1.5382000106439532e-06, 'learning_rate/action_model': 1.5382000106439532e-05, 'epoch': 1.6}) +[2026-06-16 09:24:32,074][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.07145440578460693, 'train/grad_norm_pre_clip': 22.019424438476562, 'timing/data': 0.021607233677059412, 'timing/model': 1.5840866398066282, 'learning_rate/qwen_vl_interface': 1.5363004582768584e-06, 'learning_rate/action_model': 1.5363004582768584e-05, 'epoch': 1.6}) +[2026-06-16 09:24:38,425][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.028611356392502785, 'train/grad_norm_pre_clip': 21.83765411376953, 'timing/data': 0.021363232750445604, 'timing/model': 1.5712453168816864, 'learning_rate/qwen_vl_interface': 1.534402432471367e-06, 'learning_rate/action_model': 1.534402432471367e-05, 'epoch': 1.6}) +[2026-06-16 09:24:44,708][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.00026582260034047067, 'train/grad_norm_pre_clip': 0.02245975472033024, 'timing/data': 0.021026560105383396, 'timing/model': 1.787338873371482, 'learning_rate/qwen_vl_interface': 1.532505934007685e-06, 'learning_rate/action_model': 1.532505934007685e-05, 'epoch': 1.6}) +[2026-06-16 09:24:51,198][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.17867843806743622, 'train/grad_norm_pre_clip': 8.726384162902832, 'timing/data': 0.01953612780198455, 'timing/model': 1.8916541766375303, 'learning_rate/qwen_vl_interface': 1.5306109636653915e-06, 'learning_rate/action_model': 1.5306109636653914e-05, 'epoch': 1.6}) +[2026-06-16 09:24:57,271][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.09494803100824356, 'train/grad_norm_pre_clip': 5.986231803894043, 'timing/data': 0.01957825804129243, 'timing/model': 1.56167900422588, 'learning_rate/qwen_vl_interface': 1.5287175222234383e-06, 'learning_rate/action_model': 1.528717522223438e-05, 'epoch': 1.6}) +[2026-06-16 09:25:03,664][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.1063285619020462, 'train/grad_norm_pre_clip': 12.938664436340332, 'timing/data': 0.02121366886422038, 'timing/model': 1.5810945723205805, 'learning_rate/qwen_vl_interface': 1.5268256104601457e-06, 'learning_rate/action_model': 1.5268256104601454e-05, 'epoch': 1.6}) +[2026-06-16 09:25:09,968][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.002942480146884918, 'train/grad_norm_pre_clip': 2.8877992630004883, 'timing/data': 0.020579954143613577, 'timing/model': 1.549570303875953, 'learning_rate/qwen_vl_interface': 1.5249352291532075e-06, 'learning_rate/action_model': 1.5249352291532073e-05, 'epoch': 1.6}) +[2026-06-16 09:25:16,270][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.001692894846200943, 'train/grad_norm_pre_clip': 0.3263908624649048, 'timing/data': 0.020240191835910082, 'timing/model': 1.520237582270056, 'learning_rate/qwen_vl_interface': 1.523046379079689e-06, 'learning_rate/action_model': 1.523046379079689e-05, 'epoch': 1.6}) +[2026-06-16 09:25:22,651][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.001324929646216333, 'train/grad_norm_pre_clip': 0.2809399962425232, 'timing/data': 0.02160634007304907, 'timing/model': 1.822326370049268, 'learning_rate/qwen_vl_interface': 1.5211590610160244e-06, 'learning_rate/action_model': 1.5211590610160242e-05, 'epoch': 1.6}) +[2026-06-16 09:25:28,675][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.08684425055980682, 'train/grad_norm_pre_clip': 85.85054779052734, 'timing/data': 0.025310156401246786, 'timing/model': 1.550696334335953, 'learning_rate/qwen_vl_interface': 1.5192732757380193e-06, 'learning_rate/action_model': 1.5192732757380194e-05, 'epoch': 1.6}) +[2026-06-16 09:25:34,980][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.0009663653909228742, 'train/grad_norm_pre_clip': 0.10309017449617386, 'timing/data': 0.020764393266290426, 'timing/model': 1.5654052519239485, 'learning_rate/qwen_vl_interface': 1.517389024020849e-06, 'learning_rate/action_model': 1.5173890240208488e-05, 'epoch': 1.6}) +[2026-06-16 09:25:41,448][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.0007562226383015513, 'train/grad_norm_pre_clip': 0.13451410830020905, 'timing/data': 0.02020915923640132, 'timing/model': 1.6487556099891663, 'learning_rate/qwen_vl_interface': 1.515506306639057e-06, 'learning_rate/action_model': 1.515506306639057e-05, 'epoch': 1.6}) +[2026-06-16 09:25:47,682][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.0003635203465819359, 'train/grad_norm_pre_clip': 0.040713150054216385, 'timing/data': 0.020498510915786028, 'timing/model': 1.5342666869983077, 'learning_rate/qwen_vl_interface': 1.5136251243665589e-06, 'learning_rate/action_model': 1.5136251243665589e-05, 'epoch': 1.6}) +[2026-06-16 09:25:53,976][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.0010531526058912277, 'train/grad_norm_pre_clip': 0.17554090917110443, 'timing/data': 0.02088875323534012, 'timing/model': 1.594844828825444, 'learning_rate/qwen_vl_interface': 1.5117454779766372e-06, 'learning_rate/action_model': 1.5117454779766371e-05, 'epoch': 1.6}) +[2026-06-16 09:26:00,260][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.0010865035001188517, 'train/grad_norm_pre_clip': 0.23604141175746918, 'timing/data': 0.020663180854171515, 'timing/model': 1.7419703188352287, 'learning_rate/qwen_vl_interface': 1.5098673682419428e-06, 'learning_rate/action_model': 1.5098673682419429e-05, 'epoch': 1.6}) +[2026-06-16 09:26:06,364][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.0011499911779537797, 'train/grad_norm_pre_clip': 0.2249305099248886, 'timing/data': 0.020965981762856245, 'timing/model': 1.5393418339081109, 'learning_rate/qwen_vl_interface': 1.5079907959344977e-06, 'learning_rate/action_model': 1.5079907959344978e-05, 'epoch': 1.6}) +[2026-06-16 09:26:12,672][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.010210041888058186, 'train/grad_norm_pre_clip': 7.414939880371094, 'timing/data': 0.020061655901372433, 'timing/model': 1.5555279911495745, 'learning_rate/qwen_vl_interface': 1.5061157618256873e-06, 'learning_rate/action_model': 1.5061157618256871e-05, 'epoch': 1.6}) +[2026-06-16 09:26:19,272][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.014964661560952663, 'train/grad_norm_pre_clip': 12.323872566223145, 'timing/data': 0.0205903397873044, 'timing/model': 1.5502550220116973, 'learning_rate/qwen_vl_interface': 1.5042422666862675e-06, 'learning_rate/action_model': 1.5042422666862673e-05, 'epoch': 1.6}) +[2026-06-16 09:26:25,663][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.0007264238083735108, 'train/grad_norm_pre_clip': 0.12438329309225082, 'timing/data': 0.021247003227472305, 'timing/model': 1.5965560409240425, 'learning_rate/qwen_vl_interface': 1.5023703112863628e-06, 'learning_rate/action_model': 1.5023703112863626e-05, 'epoch': 1.6}) +[2026-06-16 09:26:32,051][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.008735873736441135, 'train/grad_norm_pre_clip': 5.49958610534668, 'timing/data': 0.020917960908263922, 'timing/model': 1.589397210162133, 'learning_rate/qwen_vl_interface': 1.5004998963954629e-06, 'learning_rate/action_model': 1.5004998963954628e-05, 'epoch': 1.61}) +[2026-06-16 09:26:38,362][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.037621475756168365, 'train/grad_norm_pre_clip': 30.463274002075195, 'timing/data': 0.021735535003244877, 'timing/model': 1.7408985840156674, 'learning_rate/qwen_vl_interface': 1.498631022782424e-06, 'learning_rate/action_model': 1.4986310227824241e-05, 'epoch': 1.61}) +[2026-06-16 09:26:44,479][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.053972046822309494, 'train/grad_norm_pre_clip': 18.865312576293945, 'timing/data': 0.027968465816229582, 'timing/model': 1.533589675091207, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 1.61}) +[2026-06-16 09:26:50,780][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.022985853254795074, 'train/grad_norm_pre_clip': 13.094953536987305, 'timing/data': 0.023551735561341047, 'timing/model': 1.5436133919283748, 'learning_rate/qwen_vl_interface': 1.4948979024621915e-06, 'learning_rate/action_model': 1.4948979024621914e-05, 'epoch': 1.61}) +[2026-06-16 09:26:57,185][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.0013309059431776404, 'train/grad_norm_pre_clip': 0.42411190271377563, 'timing/data': 0.020195605233311653, 'timing/model': 1.5746568371541798, 'learning_rate/qwen_vl_interface': 1.493033657289542e-06, 'learning_rate/action_model': 1.493033657289542e-05, 'epoch': 1.61}) +[2026-06-16 09:27:03,550][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.05249175801873207, 'train/grad_norm_pre_clip': 25.774080276489258, 'timing/data': 0.021455871872603893, 'timing/model': 1.597620918881148, 'learning_rate/qwen_vl_interface': 1.4911709564638428e-06, 'learning_rate/action_model': 1.4911709564638427e-05, 'epoch': 1.61}) +[2026-06-16 09:27:09,910][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.0010430625407025218, 'train/grad_norm_pre_clip': 0.2922157943248749, 'timing/data': 0.02172737382352352, 'timing/model': 1.789307655300945, 'learning_rate/qwen_vl_interface': 1.4893098007507817e-06, 'learning_rate/action_model': 1.4893098007507816e-05, 'epoch': 1.61}) +[2026-06-16 09:27:15,979][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.05965391919016838, 'train/grad_norm_pre_clip': 18.472780227661133, 'timing/data': 0.022003005258738995, 'timing/model': 1.553933483082801, 'learning_rate/qwen_vl_interface': 1.487450190915406e-06, 'learning_rate/action_model': 1.487450190915406e-05, 'epoch': 1.61}) +[2026-06-16 09:27:22,273][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.19680269062519073, 'train/grad_norm_pre_clip': 23.181724548339844, 'timing/data': 0.021663932129740715, 'timing/model': 1.5528281070291996, 'learning_rate/qwen_vl_interface': 1.4855921277221336e-06, 'learning_rate/action_model': 1.4855921277221336e-05, 'epoch': 1.61}) +[2026-06-16 09:27:28,471][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.0009986776858568192, 'train/grad_norm_pre_clip': 0.15764401853084564, 'timing/data': 0.02010349091142416, 'timing/model': 1.5547534320503473, 'learning_rate/qwen_vl_interface': 1.4837356119347423e-06, 'learning_rate/action_model': 1.4837356119347423e-05, 'epoch': 1.61}) +[2026-06-16 09:27:34,770][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.00084399722982198, 'train/grad_norm_pre_clip': 0.2421494722366333, 'timing/data': 0.023296141996979713, 'timing/model': 1.5553547982126474, 'learning_rate/qwen_vl_interface': 1.4818806443163774e-06, 'learning_rate/action_model': 1.4818806443163775e-05, 'epoch': 1.61}) +[2026-06-16 09:27:41,069][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.006834974512457848, 'train/grad_norm_pre_clip': 3.16634202003479, 'timing/data': 0.021192587912082672, 'timing/model': 1.5509425420314074, 'learning_rate/qwen_vl_interface': 1.480027225629547e-06, 'learning_rate/action_model': 1.480027225629547e-05, 'epoch': 1.61}) +[2026-06-16 09:27:47,431][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.0013535164762288332, 'train/grad_norm_pre_clip': 0.2551635801792145, 'timing/data': 0.02197302272543311, 'timing/model': 1.8115790276788175, 'learning_rate/qwen_vl_interface': 1.4781753566361182e-06, 'learning_rate/action_model': 1.4781753566361182e-05, 'epoch': 1.61}) +[2026-06-16 09:27:53,560][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.00046719456440769136, 'train/grad_norm_pre_clip': 0.04555324465036392, 'timing/data': 0.021542279981076717, 'timing/model': 1.5970303970389068, 'learning_rate/qwen_vl_interface': 1.4763250380973255e-06, 'learning_rate/action_model': 1.4763250380973254e-05, 'epoch': 1.61}) +[2026-06-16 09:27:59,875][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.1891130656003952, 'train/grad_norm_pre_clip': 4.846441268920898, 'timing/data': 0.020681113004684448, 'timing/model': 1.5573902539908886, 'learning_rate/qwen_vl_interface': 1.4744762707737662e-06, 'learning_rate/action_model': 1.4744762707737662e-05, 'epoch': 1.61}) +[2026-06-16 09:28:06,261][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.0036725515965372324, 'train/grad_norm_pre_clip': 0.6287506818771362, 'timing/data': 0.020421931985765696, 'timing/model': 1.5991732962429523, 'learning_rate/qwen_vl_interface': 1.4726290554253977e-06, 'learning_rate/action_model': 1.4726290554253977e-05, 'epoch': 1.61}) +[2026-06-16 09:28:12,651][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.0004275436222087592, 'train/grad_norm_pre_clip': 0.046724606305360794, 'timing/data': 0.021093985997140408, 'timing/model': 1.5950304819270968, 'learning_rate/qwen_vl_interface': 1.4707833928115406e-06, 'learning_rate/action_model': 1.4707833928115405e-05, 'epoch': 1.61}) +[2026-06-16 09:28:18,967][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.0005100804264657199, 'train/grad_norm_pre_clip': 0.07391148805618286, 'timing/data': 0.020666381809860468, 'timing/model': 1.544461403042078, 'learning_rate/qwen_vl_interface': 1.4689392836908775e-06, 'learning_rate/action_model': 1.4689392836908774e-05, 'epoch': 1.61}) +[2026-06-16 09:28:25,308][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.006925174500793219, 'train/grad_norm_pre_clip': 6.19000768661499, 'timing/data': 0.019652626011520624, 'timing/model': 1.7961424556560814, 'learning_rate/qwen_vl_interface': 1.4670967288214514e-06, 'learning_rate/action_model': 1.4670967288214513e-05, 'epoch': 1.61}) +[2026-06-16 09:28:31,362][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.09541603922843933, 'train/grad_norm_pre_clip': 29.667713165283203, 'timing/data': 0.02088370407000184, 'timing/model': 1.5492398031055927, 'learning_rate/qwen_vl_interface': 1.4652557289606687e-06, 'learning_rate/action_model': 1.4652557289606685e-05, 'epoch': 1.61}) +[2026-06-16 09:28:37,684][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.0010632261401042342, 'train/grad_norm_pre_clip': 0.14833155274391174, 'timing/data': 0.02393532218411565, 'timing/model': 1.5663825976662338, 'learning_rate/qwen_vl_interface': 1.4634162848652918e-06, 'learning_rate/action_model': 1.4634162848652919e-05, 'epoch': 1.61}) +[2026-06-16 09:28:44,088][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.03463895246386528, 'train/grad_norm_pre_clip': 22.646358489990234, 'timing/data': 0.020938172936439514, 'timing/model': 1.5866616019047797, 'learning_rate/qwen_vl_interface': 1.4615783972914494e-06, 'learning_rate/action_model': 1.4615783972914493e-05, 'epoch': 1.61}) +[2026-06-16 09:28:50,385][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.06551282107830048, 'train/grad_norm_pre_clip': 13.954900741577148, 'timing/data': 0.022346701938658953, 'timing/model': 1.5602307301014662, 'learning_rate/qwen_vl_interface': 1.459742066994626e-06, 'learning_rate/action_model': 1.4597420669946261e-05, 'epoch': 1.61}) +[2026-06-16 09:28:56,772][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.05176354944705963, 'train/grad_norm_pre_clip': 20.303314208984375, 'timing/data': 0.022968764882534742, 'timing/model': 1.583531622774899, 'learning_rate/qwen_vl_interface': 1.4579072947296687e-06, 'learning_rate/action_model': 1.4579072947296687e-05, 'epoch': 1.61}) +[2026-06-16 09:29:03,120][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.12802375853061676, 'train/grad_norm_pre_clip': 20.637516021728516, 'timing/data': 0.0221573393791914, 'timing/model': 1.8001361736096442, 'learning_rate/qwen_vl_interface': 1.4560740812507823e-06, 'learning_rate/action_model': 1.4560740812507822e-05, 'epoch': 1.61}) +[2026-06-16 09:29:09,365][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.056810714304447174, 'train/grad_norm_pre_clip': 18.899303436279297, 'timing/data': 0.021889006718993187, 'timing/model': 1.6491111079230905, 'learning_rate/qwen_vl_interface': 1.4542424273115322e-06, 'learning_rate/action_model': 1.454242427311532e-05, 'epoch': 1.62}) +[2026-06-16 09:29:15,676][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.0025700770784169436, 'train/grad_norm_pre_clip': 0.7349087595939636, 'timing/data': 0.02062657894566655, 'timing/model': 1.5554783758707345, 'learning_rate/qwen_vl_interface': 1.4524123336648419e-06, 'learning_rate/action_model': 1.4524123336648418e-05, 'epoch': 1.62}) +[2026-06-16 09:29:21,972][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.033809952437877655, 'train/grad_norm_pre_clip': 18.138656616210938, 'timing/data': 0.02100292220711708, 'timing/model': 1.5548731600865722, 'learning_rate/qwen_vl_interface': 1.4505838010629942e-06, 'learning_rate/action_model': 1.4505838010629941e-05, 'epoch': 1.62}) +[2026-06-16 09:29:28,372][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.003275723662227392, 'train/grad_norm_pre_clip': 2.1854774951934814, 'timing/data': 0.020783530082553625, 'timing/model': 1.5747112110257149, 'learning_rate/qwen_vl_interface': 1.4487568302576284e-06, 'learning_rate/action_model': 1.4487568302576282e-05, 'epoch': 1.62}) +[2026-06-16 09:29:34,770][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.0022669085301458836, 'train/grad_norm_pre_clip': 0.856453537940979, 'timing/data': 0.021418442949652672, 'timing/model': 1.614482435863465, 'learning_rate/qwen_vl_interface': 1.446931421999746e-06, 'learning_rate/action_model': 1.4469314219997459e-05, 'epoch': 1.62}) +[2026-06-16 09:29:41,204][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.09325002133846283, 'train/grad_norm_pre_clip': 24.925539016723633, 'timing/data': 0.021205341909080744, 'timing/model': 1.7931265328079462, 'learning_rate/qwen_vl_interface': 1.4451075770397004e-06, 'learning_rate/action_model': 1.4451075770397004e-05, 'epoch': 1.62}) +[2026-06-16 09:40:36,521][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.0012372200144454837, 'train/grad_norm_pre_clip': 0.28972551226615906, 'eval/loss': 0.031729649752378464, 'eval/flap_precision': 0.9690860215053764, 'eval/flap_recall': 0.9836289222373806, 'eval/flap_f1': 0.976303317535545, 'eval/flap_rate_model': 0.20666666666666667, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.976303317535545, 'timing/data': 0.020532446913421154, 'timing/model': 1.5517610227689147, 'learning_rate/qwen_vl_interface': 1.4432852961272068e-06, 'learning_rate/action_model': 1.4432852961272066e-05, 'epoch': 1.62}) +[2026-06-16 09:40:36,522][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4000_state +[2026-06-16 09:40:36,529][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 09:40:45,120][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4000_state/model.safetensors +[2026-06-16 09:41:06,197][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4000_state/optimizer.bin +[2026-06-16 09:41:06,198][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4000_state/sampler.bin +[2026-06-16 09:41:06,198][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-16 09:41:06,199][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-16 09:41:06,391][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-16 09:41:06,391][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 09:41:06,406][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 09:41:12,792][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.11926814913749695, 'train/grad_norm_pre_clip': 13.231376647949219, 'timing/data': 0.02212341921404004, 'timing/model': 1.8995769140310585, 'learning_rate/qwen_vl_interface': 1.4414645800113353e-06, 'learning_rate/action_model': 1.4414645800113352e-05, 'epoch': 1.62}) +[2026-06-16 09:41:18,856][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.19591644406318665, 'train/grad_norm_pre_clip': 14.21249008178711, 'timing/data': 0.02176412707194686, 'timing/model': 1.5466141500510275, 'learning_rate/qwen_vl_interface': 1.4396454294405138e-06, 'learning_rate/action_model': 1.4396454294405136e-05, 'epoch': 1.62}) +[2026-06-16 09:41:25,170][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 0.0019050930859521031, 'train/grad_norm_pre_clip': 1.1522979736328125, 'timing/data': 0.022014138288795948, 'timing/model': 1.5577059839852154, 'learning_rate/qwen_vl_interface': 1.4378278451625271e-06, 'learning_rate/action_model': 1.4378278451625271e-05, 'epoch': 1.62}) +[2026-06-16 09:41:31,466][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.004699165467172861, 'train/grad_norm_pre_clip': 5.329345703125, 'timing/data': 0.022494201082736254, 'timing/model': 1.5573384570889175, 'learning_rate/qwen_vl_interface': 1.4360118279245143e-06, 'learning_rate/action_model': 1.4360118279245144e-05, 'epoch': 1.62}) +[2026-06-16 09:41:37,748][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.08109433948993683, 'train/grad_norm_pre_clip': 40.942623138427734, 'timing/data': 0.02464051218703389, 'timing/model': 1.5326824500225484, 'learning_rate/qwen_vl_interface': 1.434197378472973e-06, 'learning_rate/action_model': 1.4341973784729729e-05, 'epoch': 1.62}) +[2026-06-16 09:41:43,952][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.004622952081263065, 'train/grad_norm_pre_clip': 3.8823893070220947, 'timing/data': 0.02135643409565091, 'timing/model': 1.5127995880320668, 'learning_rate/qwen_vl_interface': 1.4323844975537535e-06, 'learning_rate/action_model': 1.4323844975537534e-05, 'epoch': 1.62}) +[2026-06-16 09:41:50,315][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.31057193875312805, 'train/grad_norm_pre_clip': 19.724735260009766, 'timing/data': 0.023429416120052338, 'timing/model': 1.8106264467351139, 'learning_rate/qwen_vl_interface': 1.4305731859120628e-06, 'learning_rate/action_model': 1.4305731859120628e-05, 'epoch': 1.62}) +[2026-06-16 09:41:56,538][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.0016206904547289014, 'train/grad_norm_pre_clip': 0.28522491455078125, 'timing/data': 0.020317778922617435, 'timing/model': 1.7188224103301764, 'learning_rate/qwen_vl_interface': 1.4287634442924647e-06, 'learning_rate/action_model': 1.4287634442924646e-05, 'epoch': 1.62}) +[2026-06-16 09:42:02,572][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.0028401287272572517, 'train/grad_norm_pre_clip': 1.2758293151855469, 'timing/data': 0.022904062177985907, 'timing/model': 1.567895787768066, 'learning_rate/qwen_vl_interface': 1.4269552734388725e-06, 'learning_rate/action_model': 1.4269552734388724e-05, 'epoch': 1.62}) +[2026-06-16 09:42:08,868][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.008062364533543587, 'train/grad_norm_pre_clip': 3.2048025131225586, 'timing/data': 0.02089767064899206, 'timing/model': 1.5653622089885175, 'learning_rate/qwen_vl_interface': 1.4251486740945584e-06, 'learning_rate/action_model': 1.4251486740945583e-05, 'epoch': 1.62}) +[2026-06-16 09:42:15,378][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.05851958692073822, 'train/grad_norm_pre_clip': 27.987655639648438, 'timing/data': 0.021274504251778126, 'timing/model': 1.7002881146036088, 'learning_rate/qwen_vl_interface': 1.423343647002147e-06, 'learning_rate/action_model': 1.423343647002147e-05, 'epoch': 1.62}) +[2026-06-16 09:42:21,736][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.0017748380778357387, 'train/grad_norm_pre_clip': 0.4591479003429413, 'timing/data': 0.0206189495511353, 'timing/model': 1.581921768374741, 'learning_rate/qwen_vl_interface': 1.421540192903617e-06, 'learning_rate/action_model': 1.421540192903617e-05, 'epoch': 1.62}) +[2026-06-16 09:42:28,090][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 0.013038599863648415, 'train/grad_norm_pre_clip': 10.812788009643555, 'timing/data': 0.021303643006831408, 'timing/model': 1.7847392447292805, 'learning_rate/qwen_vl_interface': 1.4197383125402997e-06, 'learning_rate/action_model': 1.4197383125402997e-05, 'epoch': 1.62}) +[2026-06-16 09:42:34,410][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.0010313454549759626, 'train/grad_norm_pre_clip': 0.1923704892396927, 'timing/data': 0.020755010657012463, 'timing/model': 1.7964707217179239, 'learning_rate/qwen_vl_interface': 1.4179380066528807e-06, 'learning_rate/action_model': 1.4179380066528807e-05, 'epoch': 1.62}) +[2026-06-16 09:42:40,460][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.005096198990941048, 'train/grad_norm_pre_clip': 4.8344197273254395, 'timing/data': 0.021451635286211967, 'timing/model': 1.5442186910659075, 'learning_rate/qwen_vl_interface': 1.4161392759813984e-06, 'learning_rate/action_model': 1.4161392759813983e-05, 'epoch': 1.62}) +[2026-06-16 09:42:46,771][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.09952044486999512, 'train/grad_norm_pre_clip': 26.91695785522461, 'timing/data': 0.02124880300834775, 'timing/model': 1.5624653440900147, 'learning_rate/qwen_vl_interface': 1.4143421212652411e-06, 'learning_rate/action_model': 1.414342121265241e-05, 'epoch': 1.62}) +[2026-06-16 09:42:53,071][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.0006932724500074983, 'train/grad_norm_pre_clip': 0.09477084130048752, 'timing/data': 0.020971951074898243, 'timing/model': 1.5522144846618176, 'learning_rate/qwen_vl_interface': 1.4125465432431529e-06, 'learning_rate/action_model': 1.4125465432431529e-05, 'epoch': 1.62}) +[2026-06-16 09:42:59,463][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.0004715384275186807, 'train/grad_norm_pre_clip': 0.04557495936751366, 'timing/data': 0.02129138307645917, 'timing/model': 1.559241317678243, 'learning_rate/qwen_vl_interface': 1.4107525426532273e-06, 'learning_rate/action_model': 1.4107525426532272e-05, 'epoch': 1.62}) +[2026-06-16 09:43:05,809][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.2476484477519989, 'train/grad_norm_pre_clip': 22.5574893951416, 'timing/data': 0.020842989906668663, 'timing/model': 1.791854450944811, 'learning_rate/qwen_vl_interface': 1.4089601202329116e-06, 'learning_rate/action_model': 1.4089601202329114e-05, 'epoch': 1.63}) +[2026-06-16 09:43:11,865][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.0004966568085364997, 'train/grad_norm_pre_clip': 0.051369693130254745, 'timing/data': 0.021383554209023714, 'timing/model': 1.5548599460162222, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 1.63}) +[2026-06-16 09:43:18,266][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.1074310839176178, 'train/grad_norm_pre_clip': 11.57668685913086, 'timing/data': 0.020654200110584497, 'timing/model': 1.5497834640555084, 'learning_rate/qwen_vl_interface': 1.4053800128476426e-06, 'learning_rate/action_model': 1.4053800128476427e-05, 'epoch': 1.63}) +[2026-06-16 09:43:24,762][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.035794440656900406, 'train/grad_norm_pre_clip': 24.605764389038086, 'timing/data': 0.019654934760183096, 'timing/model': 1.5612936899997294, 'learning_rate/qwen_vl_interface': 1.403592329354338e-06, 'learning_rate/action_model': 1.403592329354338e-05, 'epoch': 1.63}) +[2026-06-16 09:43:31,148][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.08695665001869202, 'train/grad_norm_pre_clip': 19.597307205200195, 'timing/data': 0.021334114018827677, 'timing/model': 1.5905046612024307, 'learning_rate/qwen_vl_interface': 1.4018062269739354e-06, 'learning_rate/action_model': 1.4018062269739355e-05, 'epoch': 1.63}) +[2026-06-16 09:43:37,472][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.10972613096237183, 'train/grad_norm_pre_clip': 13.10601806640625, 'timing/data': 0.024410868994891644, 'timing/model': 1.572082320228219, 'learning_rate/qwen_vl_interface': 1.4000217064406339e-06, 'learning_rate/action_model': 1.4000217064406337e-05, 'epoch': 1.63}) +[2026-06-16 09:43:43,755][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.19540852308273315, 'train/grad_norm_pre_clip': 11.563777923583984, 'timing/data': 0.022787950932979584, 'timing/model': 1.7358889640308917, 'learning_rate/qwen_vl_interface': 1.3982387684879836e-06, 'learning_rate/action_model': 1.3982387684879836e-05, 'epoch': 1.63}) +[2026-06-16 09:43:49,865][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.00017603355809114873, 'train/grad_norm_pre_clip': 0.009426608681678772, 'timing/data': 0.02058911882340908, 'timing/model': 1.552102664951235, 'learning_rate/qwen_vl_interface': 1.3964574138488804e-06, 'learning_rate/action_model': 1.3964574138488803e-05, 'epoch': 1.63}) +[2026-06-16 09:43:56,162][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.008019871078431606, 'train/grad_norm_pre_clip': 3.6139190196990967, 'timing/data': 0.021740364842116833, 'timing/model': 1.5390163217671216, 'learning_rate/qwen_vl_interface': 1.3946776432555743e-06, 'learning_rate/action_model': 1.3946776432555741e-05, 'epoch': 1.63}) +[2026-06-16 09:44:02,468][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.0025352949742227793, 'train/grad_norm_pre_clip': 0.6257642507553101, 'timing/data': 0.02208919683471322, 'timing/model': 1.5441439049318433, 'learning_rate/qwen_vl_interface': 1.3928994574396595e-06, 'learning_rate/action_model': 1.3928994574396594e-05, 'epoch': 1.63}) +[2026-06-16 09:44:08,959][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.0020914683118462563, 'train/grad_norm_pre_clip': 0.8145715594291687, 'timing/data': 0.02085907105356455, 'timing/model': 1.5933243809267879, 'learning_rate/qwen_vl_interface': 1.391122857132083e-06, 'learning_rate/action_model': 1.3911228571320831e-05, 'epoch': 1.63}) +[2026-06-16 09:44:15,365][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.0011791768483817577, 'train/grad_norm_pre_clip': 0.2728610634803772, 'timing/data': 0.021289012394845486, 'timing/model': 1.5937098301947117, 'learning_rate/qwen_vl_interface': 1.3893478430631376e-06, 'learning_rate/action_model': 1.3893478430631376e-05, 'epoch': 1.63}) +[2026-06-16 09:44:21,679][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.00033171664108522236, 'train/grad_norm_pre_clip': 0.02577570639550686, 'timing/data': 0.020615828689187765, 'timing/model': 1.7650649207644165, 'learning_rate/qwen_vl_interface': 1.3875744159624652e-06, 'learning_rate/action_model': 1.3875744159624652e-05, 'epoch': 1.63}) +[2026-06-16 09:44:27,780][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.007794382981956005, 'train/grad_norm_pre_clip': 8.331683158874512, 'timing/data': 0.020297357346862555, 'timing/model': 1.5636920728720725, 'learning_rate/qwen_vl_interface': 1.3858025765590525e-06, 'learning_rate/action_model': 1.3858025765590525e-05, 'epoch': 1.63}) +[2026-06-16 09:44:34,073][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.0479920394718647, 'train/grad_norm_pre_clip': 24.53350067138672, 'timing/data': 0.021446792874485254, 'timing/model': 1.5405274759978056, 'learning_rate/qwen_vl_interface': 1.384032325581237e-06, 'learning_rate/action_model': 1.3840323255812371e-05, 'epoch': 1.63}) +[2026-06-16 09:44:40,474][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.0994790643453598, 'train/grad_norm_pre_clip': 29.957012176513672, 'timing/data': 0.02365649538114667, 'timing/model': 1.573237325064838, 'learning_rate/qwen_vl_interface': 1.3822636637567025e-06, 'learning_rate/action_model': 1.3822636637567024e-05, 'epoch': 1.63}) +[2026-06-16 09:44:46,865][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.001953165978193283, 'train/grad_norm_pre_clip': 0.8666130304336548, 'timing/data': 0.020427908282727003, 'timing/model': 1.5780592169612646, 'learning_rate/qwen_vl_interface': 1.3804965918124788e-06, 'learning_rate/action_model': 1.3804965918124787e-05, 'epoch': 1.63}) +[2026-06-16 09:44:53,282][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.012091984041035175, 'train/grad_norm_pre_clip': 5.9446587562561035, 'timing/data': 0.0230784323066473, 'timing/model': 1.8599620801396668, 'learning_rate/qwen_vl_interface': 1.3787311104749418e-06, 'learning_rate/action_model': 1.3787311104749418e-05, 'epoch': 1.63}) +[2026-06-16 09:44:59,461][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.16236434876918793, 'train/grad_norm_pre_clip': 45.08100509643555, 'timing/data': 0.02098770998418331, 'timing/model': 1.5544551056809723, 'learning_rate/qwen_vl_interface': 1.3769672204698149e-06, 'learning_rate/action_model': 1.3769672204698147e-05, 'epoch': 1.63}) +[2026-06-16 09:45:05,766][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.0005576605326496065, 'train/grad_norm_pre_clip': 0.07323025912046432, 'timing/data': 0.020497838035225868, 'timing/model': 1.5489693731069565, 'learning_rate/qwen_vl_interface': 1.3752049225221664e-06, 'learning_rate/action_model': 1.3752049225221663e-05, 'epoch': 1.63}) +[2026-06-16 09:45:12,064][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.049781206995248795, 'train/grad_norm_pre_clip': 23.305978775024414, 'timing/data': 0.02014812594279647, 'timing/model': 1.5401650229468942, 'learning_rate/qwen_vl_interface': 1.3734442173564103e-06, 'learning_rate/action_model': 1.3734442173564102e-05, 'epoch': 1.63}) +[2026-06-16 09:45:18,363][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.06445836275815964, 'train/grad_norm_pre_clip': 28.800519943237305, 'timing/data': 0.02093350002542138, 'timing/model': 1.54576110932976, 'learning_rate/qwen_vl_interface': 1.3716851056963062e-06, 'learning_rate/action_model': 1.3716851056963062e-05, 'epoch': 1.63}) +[2026-06-16 09:45:24,668][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.13195744156837463, 'train/grad_norm_pre_clip': 11.31501579284668, 'timing/data': 0.021375752054154873, 'timing/model': 1.554152260068804, 'learning_rate/qwen_vl_interface': 1.3699275882649597e-06, 'learning_rate/action_model': 1.3699275882649596e-05, 'epoch': 1.63}) +[2026-06-16 09:45:31,001][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.00043025327613577247, 'train/grad_norm_pre_clip': 0.10437613725662231, 'timing/data': 0.020747979171574116, 'timing/model': 1.7902650376781821, 'learning_rate/qwen_vl_interface': 1.368171665784817e-06, 'learning_rate/action_model': 1.3681716657848171e-05, 'epoch': 1.63}) +[2026-06-16 09:45:37,069][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.1265331357717514, 'train/grad_norm_pre_clip': 7.933725833892822, 'timing/data': 0.020723809022456408, 'timing/model': 1.5457482188940048, 'learning_rate/qwen_vl_interface': 1.3664173389776735e-06, 'learning_rate/action_model': 1.3664173389776735e-05, 'epoch': 1.63}) +[2026-06-16 09:45:43,369][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.07319506257772446, 'train/grad_norm_pre_clip': 22.204713821411133, 'timing/data': 0.02116207079961896, 'timing/model': 1.5504789408296347, 'learning_rate/qwen_vl_interface': 1.3646646085646658e-06, 'learning_rate/action_model': 1.3646646085646659e-05, 'epoch': 1.64}) +[2026-06-16 09:45:49,806][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.0012852053623646498, 'train/grad_norm_pre_clip': 0.4003431499004364, 'timing/data': 0.020511386916041374, 'timing/model': 1.5867887982167304, 'learning_rate/qwen_vl_interface': 1.3629134752662756e-06, 'learning_rate/action_model': 1.3629134752662756e-05, 'epoch': 1.64}) +[2026-06-16 09:45:56,192][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 0.037984494119882584, 'train/grad_norm_pre_clip': 22.509138107299805, 'timing/data': 0.020960839930921793, 'timing/model': 1.5826247530058026, 'learning_rate/qwen_vl_interface': 1.3611639398023282e-06, 'learning_rate/action_model': 1.3611639398023283e-05, 'epoch': 1.64}) +[2026-06-16 09:46:02,486][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.0013103685341775417, 'train/grad_norm_pre_clip': 0.375923216342926, 'timing/data': 0.02139746304601431, 'timing/model': 1.5445972592569888, 'learning_rate/qwen_vl_interface': 1.35941600289199e-06, 'learning_rate/action_model': 1.35941600289199e-05, 'epoch': 1.64}) +[2026-06-16 09:46:08,809][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.0011628300417214632, 'train/grad_norm_pre_clip': 0.4288589358329773, 'timing/data': 0.02131118206307292, 'timing/model': 1.7724355217069387, 'learning_rate/qwen_vl_interface': 1.3576696652537735e-06, 'learning_rate/action_model': 1.3576696652537736e-05, 'epoch': 1.64}) +[2026-06-16 09:46:14,950][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.1392957866191864, 'train/grad_norm_pre_clip': 39.987972259521484, 'timing/data': 0.021153120789676905, 'timing/model': 1.6263885330408812, 'learning_rate/qwen_vl_interface': 1.355924927605531e-06, 'learning_rate/action_model': 1.3559249276055311e-05, 'epoch': 1.64}) +[2026-06-16 09:46:21,186][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.007634901907294989, 'train/grad_norm_pre_clip': 7.844425201416016, 'timing/data': 0.021727824117988348, 'timing/model': 1.547240043990314, 'learning_rate/qwen_vl_interface': 1.3541817906644591e-06, 'learning_rate/action_model': 1.3541817906644591e-05, 'epoch': 1.64}) +[2026-06-16 09:46:27,668][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.020399445667862892, 'train/grad_norm_pre_clip': 14.66054630279541, 'timing/data': 0.020927020348608494, 'timing/model': 1.562566771171987, 'learning_rate/qwen_vl_interface': 1.3524402551470955e-06, 'learning_rate/action_model': 1.3524402551470956e-05, 'epoch': 1.64}) +[2026-06-16 09:46:33,966][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.1119561418890953, 'train/grad_norm_pre_clip': 15.066704750061035, 'timing/data': 0.020152725744992495, 'timing/model': 1.5339120081625879, 'learning_rate/qwen_vl_interface': 1.3507003217693196e-06, 'learning_rate/action_model': 1.3507003217693196e-05, 'epoch': 1.64}) +[2026-06-16 09:46:40,346][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.0018929492216557264, 'train/grad_norm_pre_clip': 0.6657176613807678, 'timing/data': 0.021647694054991007, 'timing/model': 1.5945106726139784, 'learning_rate/qwen_vl_interface': 1.3489619912463515e-06, 'learning_rate/action_model': 1.3489619912463514e-05, 'epoch': 1.64}) +[2026-06-16 09:46:46,826][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.026425717398524284, 'train/grad_norm_pre_clip': 14.451482772827148, 'timing/data': 0.021019499748945236, 'timing/model': 1.8360545141622424, 'learning_rate/qwen_vl_interface': 1.3472252642927534e-06, 'learning_rate/action_model': 1.3472252642927533e-05, 'epoch': 1.64}) +[2026-06-16 09:46:52,872][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 0.0017722126794978976, 'train/grad_norm_pre_clip': 1.9752111434936523, 'timing/data': 0.02119701122865081, 'timing/model': 1.5503993649035692, 'learning_rate/qwen_vl_interface': 1.3454901416224284e-06, 'learning_rate/action_model': 1.3454901416224284e-05, 'epoch': 1.64}) +[2026-06-16 09:46:59,276][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 0.038185205310583115, 'train/grad_norm_pre_clip': 29.59358024597168, 'timing/data': 0.030415148939937353, 'timing/model': 1.5685972678475082, 'learning_rate/qwen_vl_interface': 1.34375662394862e-06, 'learning_rate/action_model': 1.34375662394862e-05, 'epoch': 1.64}) +[2026-06-16 09:47:05,570][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.13494600355625153, 'train/grad_norm_pre_clip': 14.11506462097168, 'timing/data': 0.020409796852618456, 'timing/model': 1.5556685770861804, 'learning_rate/qwen_vl_interface': 1.3420247119839116e-06, 'learning_rate/action_model': 1.3420247119839117e-05, 'epoch': 1.64}) +[2026-06-16 09:47:11,869][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.10822661966085434, 'train/grad_norm_pre_clip': 16.338842391967773, 'timing/data': 0.020477548241615295, 'timing/model': 1.5531247309409082, 'learning_rate/qwen_vl_interface': 1.340294406440227e-06, 'learning_rate/action_model': 1.340294406440227e-05, 'epoch': 1.64}) +[2026-06-16 09:47:18,263][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.0006383205763995647, 'train/grad_norm_pre_clip': 0.21458148956298828, 'timing/data': 0.021712752990424633, 'timing/model': 1.6086700786836445, 'learning_rate/qwen_vl_interface': 1.3385657080288297e-06, 'learning_rate/action_model': 1.3385657080288297e-05, 'epoch': 1.64}) +[2026-06-16 09:47:24,606][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 6.296934589045122e-05, 'train/grad_norm_pre_clip': 0.0049135684967041016, 'timing/data': 0.019817824009805918, 'timing/model': 1.7875543511472642, 'learning_rate/qwen_vl_interface': 1.336838617460322e-06, 'learning_rate/action_model': 1.336838617460322e-05, 'epoch': 1.64}) +[2026-06-16 09:47:30,808][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.0023350147530436516, 'train/grad_norm_pre_clip': 1.1141173839569092, 'timing/data': 0.019420522265136242, 'timing/model': 1.5811949400231242, 'learning_rate/qwen_vl_interface': 1.335113135444646e-06, 'learning_rate/action_model': 1.335113135444646e-05, 'epoch': 1.64}) +[2026-06-16 09:47:37,167][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 0.0010389756644144654, 'train/grad_norm_pre_clip': 0.6739192605018616, 'timing/data': 0.020814718678593636, 'timing/model': 1.5547435292974114, 'learning_rate/qwen_vl_interface': 1.3333892626910823e-06, 'learning_rate/action_model': 1.3333892626910823e-05, 'epoch': 1.64}) +[2026-06-16 09:47:43,474][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.0007770743104629219, 'train/grad_norm_pre_clip': 0.1240861788392067, 'timing/data': 0.020970790181308985, 'timing/model': 1.5486249439418316, 'learning_rate/qwen_vl_interface': 1.3316669999082486e-06, 'learning_rate/action_model': 1.3316669999082484e-05, 'epoch': 1.64}) +[2026-06-16 09:47:49,777][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.11462431401014328, 'train/grad_norm_pre_clip': 51.36890411376953, 'timing/data': 0.020893048960715532, 'timing/model': 1.5656196000054479, 'learning_rate/qwen_vl_interface': 1.3299463478041024e-06, 'learning_rate/action_model': 1.3299463478041024e-05, 'epoch': 1.64}) +[2026-06-16 09:47:56,279][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.06705528497695923, 'train/grad_norm_pre_clip': 30.16962432861328, 'timing/data': 0.021583253052085638, 'timing/model': 1.6132594989612699, 'learning_rate/qwen_vl_interface': 1.3282273070859396e-06, 'learning_rate/action_model': 1.3282273070859396e-05, 'epoch': 1.64}) +[2026-06-16 09:48:02,728][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 0.09037590771913528, 'train/grad_norm_pre_clip': 19.239370346069336, 'timing/data': 0.021568284835666418, 'timing/model': 1.8068987629376352, 'learning_rate/qwen_vl_interface': 1.3265098784603924e-06, 'learning_rate/action_model': 1.3265098784603924e-05, 'epoch': 1.64}) +[2026-06-16 09:48:08,780][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.31698986887931824, 'train/grad_norm_pre_clip': 36.66630935668945, 'timing/data': 0.0212415331043303, 'timing/model': 1.5595641741529107, 'learning_rate/qwen_vl_interface': 1.324794062633431e-06, 'learning_rate/action_model': 1.324794062633431e-05, 'epoch': 1.64}) +[2026-06-16 09:48:15,083][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.1471518576145172, 'train/grad_norm_pre_clip': 15.390321731567383, 'timing/data': 0.021373604889959097, 'timing/model': 1.5625566677190363, 'learning_rate/qwen_vl_interface': 1.323079860310363e-06, 'learning_rate/action_model': 1.3230798603103628e-05, 'epoch': 1.65}) +[2026-06-16 09:48:21,370][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.0006544798379763961, 'train/grad_norm_pre_clip': 0.19398227334022522, 'timing/data': 0.020124739035964012, 'timing/model': 1.550421095918864, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 1.65}) +[2026-06-16 09:48:27,772][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.0009479180444031954, 'train/grad_norm_pre_clip': 0.3304841220378876, 'timing/data': 0.020560540724545717, 'timing/model': 1.5753592066466808, 'learning_rate/qwen_vl_interface': 1.3196562989938175e-06, 'learning_rate/action_model': 1.3196562989938174e-05, 'epoch': 1.65}) +[2026-06-16 09:48:34,161][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.20241433382034302, 'train/grad_norm_pre_clip': 21.06658935546875, 'timing/data': 0.021208374295383692, 'timing/model': 1.605630700942129, 'learning_rate/qwen_vl_interface': 1.3179469414076357e-06, 'learning_rate/action_model': 1.3179469414076356e-05, 'epoch': 1.65}) +[2026-06-16 09:48:40,539][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.00014108046889305115, 'train/grad_norm_pre_clip': 0.04185500368475914, 'timing/data': 0.020627671852707863, 'timing/model': 1.8224744349718094, 'learning_rate/qwen_vl_interface': 1.3162392001399401e-06, 'learning_rate/action_model': 1.3162392001399401e-05, 'epoch': 1.65}) +[2026-06-16 09:48:46,736][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.023860689252614975, 'train/grad_norm_pre_clip': 16.73403549194336, 'timing/data': 0.020389751065522432, 'timing/model': 1.7219734597019851, 'learning_rate/qwen_vl_interface': 1.3145330758927182e-06, 'learning_rate/action_model': 1.3145330758927182e-05, 'epoch': 1.65}) +[2026-06-16 09:48:52,769][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.0016363277100026608, 'train/grad_norm_pre_clip': 1.3314237594604492, 'timing/data': 0.020474720746278763, 'timing/model': 1.5530686690472066, 'learning_rate/qwen_vl_interface': 1.3128285693672935e-06, 'learning_rate/action_model': 1.3128285693672934e-05, 'epoch': 1.65}) +[2026-06-16 09:48:59,089][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.20129522681236267, 'train/grad_norm_pre_clip': 51.79662322998047, 'timing/data': 0.021210354287177324, 'timing/model': 1.5715827508829534, 'learning_rate/qwen_vl_interface': 1.311125681264325e-06, 'learning_rate/action_model': 1.311125681264325e-05, 'epoch': 1.65}) +[2026-06-16 09:49:05,471][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.0016927991528064013, 'train/grad_norm_pre_clip': 0.3020213842391968, 'timing/data': 0.020635182037949562, 'timing/model': 1.6107675759121776, 'learning_rate/qwen_vl_interface': 1.309424412283803e-06, 'learning_rate/action_model': 1.3094244122838029e-05, 'epoch': 1.65}) +[2026-06-16 09:49:11,861][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 0.0001681497087702155, 'train/grad_norm_pre_clip': 0.025774797424674034, 'timing/data': 0.021349266171455383, 'timing/model': 1.6021321271546185, 'learning_rate/qwen_vl_interface': 1.307724763125057e-06, 'learning_rate/action_model': 1.307724763125057e-05, 'epoch': 1.65}) +[2026-06-16 09:49:18,205][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.07512107491493225, 'train/grad_norm_pre_clip': 15.058622360229492, 'timing/data': 0.021876487880945206, 'timing/model': 1.7817482119426131, 'learning_rate/qwen_vl_interface': 1.3060267344867478e-06, 'learning_rate/action_model': 1.3060267344867478e-05, 'epoch': 1.65}) +[2026-06-16 09:49:24,517][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.015132011845707893, 'train/grad_norm_pre_clip': 10.680306434631348, 'timing/data': 0.020313750952482224, 'timing/model': 1.7910274537280202, 'learning_rate/qwen_vl_interface': 1.3043303270668715e-06, 'learning_rate/action_model': 1.3043303270668713e-05, 'epoch': 1.65}) +[2026-06-16 09:49:30,573][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.0004422989732120186, 'train/grad_norm_pre_clip': 0.12957380712032318, 'timing/data': 0.021474025677889585, 'timing/model': 1.5488031669519842, 'learning_rate/qwen_vl_interface': 1.3026355415627565e-06, 'learning_rate/action_model': 1.3026355415627564e-05, 'epoch': 1.65}) +[2026-06-16 09:49:36,869][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.0010155793279409409, 'train/grad_norm_pre_clip': 0.21216164529323578, 'timing/data': 0.02230140008032322, 'timing/model': 1.5451638400554657, 'learning_rate/qwen_vl_interface': 1.3009423786710653e-06, 'learning_rate/action_model': 1.3009423786710653e-05, 'epoch': 1.65}) +[2026-06-16 09:49:43,279][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.00039317546179518104, 'train/grad_norm_pre_clip': 0.08960294723510742, 'timing/data': 0.021366215776652098, 'timing/model': 1.5641995542682707, 'learning_rate/qwen_vl_interface': 1.2992508390877933e-06, 'learning_rate/action_model': 1.2992508390877933e-05, 'epoch': 1.65}) +[2026-06-16 09:49:49,659][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.00033238399191759527, 'train/grad_norm_pre_clip': 0.08135396987199783, 'timing/data': 0.021499237045645714, 'timing/model': 1.5875707888044417, 'learning_rate/qwen_vl_interface': 1.297560923508269e-06, 'learning_rate/action_model': 1.297560923508269e-05, 'epoch': 1.65}) +[2026-06-16 09:49:56,109][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.0006101547623984516, 'train/grad_norm_pre_clip': 0.08920379728078842, 'timing/data': 0.02162836818024516, 'timing/model': 1.8898144820705056, 'learning_rate/qwen_vl_interface': 1.295872632627152e-06, 'learning_rate/action_model': 1.2958726326271519e-05, 'epoch': 1.65}) +[2026-06-16 09:50:02,170][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 0.0008688821690157056, 'train/grad_norm_pre_clip': 0.5083653926849365, 'timing/data': 0.020951685030013323, 'timing/model': 1.556430745869875, 'learning_rate/qwen_vl_interface': 1.2941859671384364e-06, 'learning_rate/action_model': 1.2941859671384362e-05, 'epoch': 1.65}) +[2026-06-16 09:50:08,467][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.0014027238357812166, 'train/grad_norm_pre_clip': 0.7587220668792725, 'timing/data': 0.021061065141111612, 'timing/model': 1.5519508039578795, 'learning_rate/qwen_vl_interface': 1.2925009277354444e-06, 'learning_rate/action_model': 1.2925009277354444e-05, 'epoch': 1.65}) +[2026-06-16 09:50:14,865][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.21380995213985443, 'train/grad_norm_pre_clip': 14.319480895996094, 'timing/data': 0.021096546202898026, 'timing/model': 1.5686770277097821, 'learning_rate/qwen_vl_interface': 1.2908175151108335e-06, 'learning_rate/action_model': 1.2908175151108334e-05, 'epoch': 1.65}) +[2026-06-16 09:50:21,286][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.011722286231815815, 'train/grad_norm_pre_clip': 5.962978363037109, 'timing/data': 0.021228287369012833, 'timing/model': 1.5398689769208431, 'learning_rate/qwen_vl_interface': 1.2891357299565899e-06, 'learning_rate/action_model': 1.2891357299565898e-05, 'epoch': 1.65}) +[2026-06-16 09:50:27,570][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.004399423021823168, 'train/grad_norm_pre_clip': 1.490529179573059, 'timing/data': 0.02156265825033188, 'timing/model': 1.5886206571012735, 'learning_rate/qwen_vl_interface': 1.2874555729640333e-06, 'learning_rate/action_model': 1.2874555729640334e-05, 'epoch': 1.65}) +[2026-06-16 09:50:33,854][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.0011792879085987806, 'train/grad_norm_pre_clip': 0.8112975358963013, 'timing/data': 0.019497098866850138, 'timing/model': 1.7421991531737149, 'learning_rate/qwen_vl_interface': 1.2857770448238124e-06, 'learning_rate/action_model': 1.2857770448238123e-05, 'epoch': 1.65}) +[2026-06-16 09:50:40,086][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.08823183923959732, 'train/grad_norm_pre_clip': 27.530826568603516, 'timing/data': 0.019719319883733988, 'timing/model': 1.5824085548520088, 'learning_rate/qwen_vl_interface': 1.2841001462259067e-06, 'learning_rate/action_model': 1.2841001462259067e-05, 'epoch': 1.65}) +[2026-06-16 09:50:46,374][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.031892869621515274, 'train/grad_norm_pre_clip': 25.711193084716797, 'timing/data': 0.02106135617941618, 'timing/model': 1.5498669249936938, 'learning_rate/qwen_vl_interface': 1.282424877859626e-06, 'learning_rate/action_model': 1.282424877859626e-05, 'epoch': 1.65}) +[2026-06-16 09:50:52,959][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.037311945110559464, 'train/grad_norm_pre_clip': 19.526662826538086, 'timing/data': 0.04827342135831714, 'timing/model': 1.5975821898318827, 'learning_rate/qwen_vl_interface': 1.2807512404136088e-06, 'learning_rate/action_model': 1.2807512404136087e-05, 'epoch': 1.66}) +[2026-06-16 09:50:59,270][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.13145025074481964, 'train/grad_norm_pre_clip': 10.328158378601074, 'timing/data': 0.021471898071467876, 'timing/model': 1.5250352169387043, 'learning_rate/qwen_vl_interface': 1.2790792345758255e-06, 'learning_rate/action_model': 1.2790792345758254e-05, 'epoch': 1.66}) +[2026-06-16 09:51:05,652][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.14961935579776764, 'train/grad_norm_pre_clip': 9.736373901367188, 'timing/data': 0.021001335699111223, 'timing/model': 1.5992874852381647, 'learning_rate/qwen_vl_interface': 1.2774088610335752e-06, 'learning_rate/action_model': 1.2774088610335752e-05, 'epoch': 1.66}) +[2026-06-16 09:51:11,973][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 0.00015941180754452944, 'train/grad_norm_pre_clip': 0.027940748259425163, 'timing/data': 0.02063734456896782, 'timing/model': 1.7619183771312237, 'learning_rate/qwen_vl_interface': 1.2757401204734827e-06, 'learning_rate/action_model': 1.2757401204734826e-05, 'epoch': 1.66}) +[2026-06-16 09:51:18,163][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.0003344618889968842, 'train/grad_norm_pre_clip': 0.14706969261169434, 'timing/data': 0.021303668152540922, 'timing/model': 1.6390743223018944, 'learning_rate/qwen_vl_interface': 1.2740730135815064e-06, 'learning_rate/action_model': 1.2740730135815062e-05, 'epoch': 1.66}) +[2026-06-16 09:51:24,575][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 0.00014280746108852327, 'train/grad_norm_pre_clip': 0.04665859416127205, 'timing/data': 0.01985607109963894, 'timing/model': 1.5684181260876358, 'learning_rate/qwen_vl_interface': 1.2724075410429298e-06, 'learning_rate/action_model': 1.2724075410429299e-05, 'epoch': 1.66}) +[2026-06-16 09:51:30,873][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.0003205202519893646, 'train/grad_norm_pre_clip': 0.06959137320518494, 'timing/data': 0.02143160765990615, 'timing/model': 1.549082834739238, 'learning_rate/qwen_vl_interface': 1.2707437035423664e-06, 'learning_rate/action_model': 1.2707437035423664e-05, 'epoch': 1.66}) +[2026-06-16 10:02:13,545][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.033287715166807175, 'train/grad_norm_pre_clip': 22.555986404418945, 'eval/loss': 0.032740432769060135, 'eval/flap_precision': 0.9689608636977058, 'eval/flap_recall': 0.9795361527967258, 'eval/flap_f1': 0.9742198100407056, 'eval/flap_rate_model': 0.20583333333333334, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9742198100407056, 'timing/data': 0.02155502885580063, 'timing/model': 1.5994829442352057, 'learning_rate/qwen_vl_interface': 1.2690815017637573e-06, 'learning_rate/action_model': 1.2690815017637572e-05, 'epoch': 1.66}) +[2026-06-16 10:02:13,546][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4100_state +[2026-06-16 10:02:13,554][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 10:02:22,052][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4100_state/model.safetensors +[2026-06-16 10:02:41,821][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4100_state/optimizer.bin +[2026-06-16 10:02:41,822][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4100_state/sampler.bin +[2026-06-16 10:02:41,822][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4100_state/sampler_1.bin +[2026-06-16 10:02:41,824][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4100_state/random_states_0.pkl +[2026-06-16 10:02:42,014][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4100 +[2026-06-16 10:02:42,015][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 10:02:42,030][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 10:02:48,245][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.0010423308704048395, 'train/grad_norm_pre_clip': 0.30391451716423035, 'timing/data': 0.021110632922500372, 'timing/model': 1.497948010917753, 'learning_rate/qwen_vl_interface': 1.2674209363903694e-06, 'learning_rate/action_model': 1.2674209363903693e-05, 'epoch': 1.66}) +[2026-06-16 10:02:54,561][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.17334212362766266, 'train/grad_norm_pre_clip': 9.251333236694336, 'timing/data': 0.02179992711171508, 'timing/model': 1.8785538179799914, 'learning_rate/qwen_vl_interface': 1.2657620081048005e-06, 'learning_rate/action_model': 1.2657620081048006e-05, 'epoch': 1.66}) +[2026-06-16 10:03:00,926][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.00010076176113216206, 'train/grad_norm_pre_clip': 0.010786714032292366, 'timing/data': 0.023033431731164455, 'timing/model': 1.740749878808856, 'learning_rate/qwen_vl_interface': 1.2641047175889714e-06, 'learning_rate/action_model': 1.2641047175889714e-05, 'epoch': 1.66}) +[2026-06-16 10:03:06,915][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.07115385681390762, 'train/grad_norm_pre_clip': 17.200712203979492, 'timing/data': 0.02072579087689519, 'timing/model': 1.5627017123624682, 'learning_rate/qwen_vl_interface': 1.2624490655241332e-06, 'learning_rate/action_model': 1.262449065524133e-05, 'epoch': 1.66}) +[2026-06-16 10:03:13,138][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.0006911076488904655, 'train/grad_norm_pre_clip': 0.15068496763706207, 'timing/data': 0.019368924666196108, 'timing/model': 1.5496115400455892, 'learning_rate/qwen_vl_interface': 1.2607950525908602e-06, 'learning_rate/action_model': 1.2607950525908601e-05, 'epoch': 1.66}) +[2026-06-16 10:03:19,400][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.00020858057541772723, 'train/grad_norm_pre_clip': 0.05747828260064125, 'timing/data': 0.01989913685247302, 'timing/model': 1.6179160550236702, 'learning_rate/qwen_vl_interface': 1.2591426794690575e-06, 'learning_rate/action_model': 1.2591426794690576e-05, 'epoch': 1.66}) +[2026-06-16 10:03:25,641][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.002406864892691374, 'train/grad_norm_pre_clip': 1.5387095212936401, 'timing/data': 0.019649306312203407, 'timing/model': 1.5205455711111426, 'learning_rate/qwen_vl_interface': 1.25749194683795e-06, 'learning_rate/action_model': 1.25749194683795e-05, 'epoch': 1.66}) +[2026-06-16 10:03:31,842][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.00018091767560690641, 'train/grad_norm_pre_clip': 0.02904706448316574, 'timing/data': 0.020058977883309126, 'timing/model': 1.5499650076963007, 'learning_rate/qwen_vl_interface': 1.2558428553760933e-06, 'learning_rate/action_model': 1.2558428553760932e-05, 'epoch': 1.66}) +[2026-06-16 10:03:38,091][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.0003952606348320842, 'train/grad_norm_pre_clip': 0.07993029057979584, 'timing/data': 0.025858457200229168, 'timing/model': 1.7911290861666203, 'learning_rate/qwen_vl_interface': 1.2541954057613659e-06, 'learning_rate/action_model': 1.2541954057613658e-05, 'epoch': 1.66}) +[2026-06-16 10:03:44,147][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.004304667469114065, 'train/grad_norm_pre_clip': 4.222180366516113, 'timing/data': 0.020466089714318514, 'timing/model': 1.553789979312569, 'learning_rate/qwen_vl_interface': 1.2525495986709727e-06, 'learning_rate/action_model': 1.2525495986709726e-05, 'epoch': 1.66}) +[2026-06-16 10:03:50,349][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.2176007330417633, 'train/grad_norm_pre_clip': 17.75567054748535, 'timing/data': 0.021023642737418413, 'timing/model': 1.549691610969603, 'learning_rate/qwen_vl_interface': 1.2509054347814423e-06, 'learning_rate/action_model': 1.2509054347814423e-05, 'epoch': 1.66}) +[2026-06-16 10:03:56,564][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.09745433181524277, 'train/grad_norm_pre_clip': 14.734138488769531, 'timing/data': 0.0216060858219862, 'timing/model': 1.5544427460990846, 'learning_rate/qwen_vl_interface': 1.2492629147686288e-06, 'learning_rate/action_model': 1.2492629147686289e-05, 'epoch': 1.66}) +[2026-06-16 10:04:02,855][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.014055089093744755, 'train/grad_norm_pre_clip': 5.825143814086914, 'timing/data': 0.02123633399605751, 'timing/model': 1.55249851802364, 'learning_rate/qwen_vl_interface': 1.2476220393077098e-06, 'learning_rate/action_model': 1.2476220393077098e-05, 'epoch': 1.66}) +[2026-06-16 10:04:09,164][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.12304095178842545, 'train/grad_norm_pre_clip': 48.262752532958984, 'timing/data': 0.020394870080053806, 'timing/model': 1.5528657222166657, 'learning_rate/qwen_vl_interface': 1.2459828090731877e-06, 'learning_rate/action_model': 1.2459828090731876e-05, 'epoch': 1.66}) +[2026-06-16 10:04:15,490][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.09054931253194809, 'train/grad_norm_pre_clip': 12.275522232055664, 'timing/data': 0.020825651939958334, 'timing/model': 1.7818533331155777, 'learning_rate/qwen_vl_interface': 1.2443452247388876e-06, 'learning_rate/action_model': 1.2443452247388877e-05, 'epoch': 1.66}) +[2026-06-16 10:04:21,556][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.06764406710863113, 'train/grad_norm_pre_clip': 26.70672035217285, 'timing/data': 0.0197025160305202, 'timing/model': 1.5440234215930104, 'learning_rate/qwen_vl_interface': 1.2427092869779603e-06, 'learning_rate/action_model': 1.2427092869779602e-05, 'epoch': 1.66}) +[2026-06-16 10:04:27,855][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.005926797166466713, 'train/grad_norm_pre_clip': 4.900188446044922, 'timing/data': 0.02059660106897354, 'timing/model': 1.5558272111229599, 'learning_rate/qwen_vl_interface': 1.2410749964628751e-06, 'learning_rate/action_model': 1.2410749964628751e-05, 'epoch': 1.66}) +[2026-06-16 10:04:34,167][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.0008156130788847804, 'train/grad_norm_pre_clip': 0.216851145029068, 'timing/data': 0.019648775923997164, 'timing/model': 1.5691153472289443, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 1.67}) +[2026-06-16 10:04:40,470][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.0005944147706031799, 'train/grad_norm_pre_clip': 0.11606255918741226, 'timing/data': 0.02087409235537052, 'timing/model': 1.5644557257182896, 'learning_rate/qwen_vl_interface': 1.2378113598567394e-06, 'learning_rate/action_model': 1.2378113598567392e-05, 'epoch': 1.67}) +[2026-06-16 10:04:46,967][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.0007189968600869179, 'train/grad_norm_pre_clip': 0.14003926515579224, 'timing/data': 0.02780313603579998, 'timing/model': 1.5646690991707146, 'learning_rate/qwen_vl_interface': 1.2361820151072454e-06, 'learning_rate/action_model': 1.2361820151072453e-05, 'epoch': 1.67}) +[2026-06-16 10:04:53,301][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.035677745938301086, 'train/grad_norm_pre_clip': 16.329999923706055, 'timing/data': 0.023043453227728605, 'timing/model': 1.7806028467603028, 'learning_rate/qwen_vl_interface': 1.23455432028671e-06, 'learning_rate/action_model': 1.23455432028671e-05, 'epoch': 1.67}) +[2026-06-16 10:04:59,364][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.022665325552225113, 'train/grad_norm_pre_clip': 19.526771545410156, 'timing/data': 0.022208468057215214, 'timing/model': 1.5539150098338723, 'learning_rate/qwen_vl_interface': 1.2329282760642167e-06, 'learning_rate/action_model': 1.2329282760642166e-05, 'epoch': 1.67}) +[2026-06-16 10:05:05,668][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.009687269106507301, 'train/grad_norm_pre_clip': 8.549054145812988, 'timing/data': 0.022234358824789524, 'timing/model': 1.5628124941140413, 'learning_rate/qwen_vl_interface': 1.2313038831081716e-06, 'learning_rate/action_model': 1.2313038831081717e-05, 'epoch': 1.67}) +[2026-06-16 10:05:11,999][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.056361887603998184, 'train/grad_norm_pre_clip': 15.931573867797852, 'timing/data': 0.020533119793981314, 'timing/model': 1.567779250908643, 'learning_rate/qwen_vl_interface': 1.2296811420863002e-06, 'learning_rate/action_model': 1.2296811420863003e-05, 'epoch': 1.67}) +[2026-06-16 10:05:18,265][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.0003812771465163678, 'train/grad_norm_pre_clip': 0.13705536723136902, 'timing/data': 0.02182691590860486, 'timing/model': 1.5523469857871532, 'learning_rate/qwen_vl_interface': 1.2280600536656504e-06, 'learning_rate/action_model': 1.2280600536656503e-05, 'epoch': 1.67}) +[2026-06-16 10:05:24,648][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.000710216467268765, 'train/grad_norm_pre_clip': 0.34497538208961487, 'timing/data': 0.021258643828332424, 'timing/model': 1.5946040251292288, 'learning_rate/qwen_vl_interface': 1.2264406185125907e-06, 'learning_rate/action_model': 1.2264406185125907e-05, 'epoch': 1.67}) +[2026-06-16 10:05:31,019][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.0002138668205589056, 'train/grad_norm_pre_clip': 0.020703166723251343, 'timing/data': 0.02094746334478259, 'timing/model': 1.8094218280166388, 'learning_rate/qwen_vl_interface': 1.2248228372928098e-06, 'learning_rate/action_model': 1.2248228372928097e-05, 'epoch': 1.67}) +[2026-06-16 10:05:37,059][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.00044871901627629995, 'train/grad_norm_pre_clip': 0.08095387369394302, 'timing/data': 0.02053937129676342, 'timing/model': 1.5439730887301266, 'learning_rate/qwen_vl_interface': 1.2232067106713174e-06, 'learning_rate/action_model': 1.2232067106713174e-05, 'epoch': 1.67}) +[2026-06-16 10:05:43,465][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.002192811109125614, 'train/grad_norm_pre_clip': 0.5720093250274658, 'timing/data': 0.020964722614735365, 'timing/model': 1.5504813520237803, 'learning_rate/qwen_vl_interface': 1.22159223931244e-06, 'learning_rate/action_model': 1.2215922393124399e-05, 'epoch': 1.67}) +[2026-06-16 10:05:49,868][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.02552287094295025, 'train/grad_norm_pre_clip': 23.133371353149414, 'timing/data': 0.025000932160764933, 'timing/model': 1.5537463100627065, 'learning_rate/qwen_vl_interface': 1.2199794238798266e-06, 'learning_rate/action_model': 1.2199794238798264e-05, 'epoch': 1.67}) +[2026-06-16 10:05:56,165][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.0002416256902506575, 'train/grad_norm_pre_clip': 0.040954720228910446, 'timing/data': 0.020378150045871735, 'timing/model': 1.554052762221545, 'learning_rate/qwen_vl_interface': 1.2183682650364447e-06, 'learning_rate/action_model': 1.2183682650364447e-05, 'epoch': 1.67}) +[2026-06-16 10:06:02,470][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.09269170463085175, 'train/grad_norm_pre_clip': 15.008709907531738, 'timing/data': 0.02057009097188711, 'timing/model': 1.52008045790717, 'learning_rate/qwen_vl_interface': 1.216758763444581e-06, 'learning_rate/action_model': 1.2167587634445809e-05, 'epoch': 1.67}) +[2026-06-16 10:06:08,797][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.0004238386463839561, 'train/grad_norm_pre_clip': 0.08351762592792511, 'timing/data': 0.022065878845751286, 'timing/model': 1.7774289911612868, 'learning_rate/qwen_vl_interface': 1.215150919765841e-06, 'learning_rate/action_model': 1.215150919765841e-05, 'epoch': 1.67}) +[2026-06-16 10:06:14,874][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.0007142004906199872, 'train/grad_norm_pre_clip': 0.34730586409568787, 'timing/data': 0.020365689881145954, 'timing/model': 1.56308988109231, 'learning_rate/qwen_vl_interface': 1.213544734661148e-06, 'learning_rate/action_model': 1.2135447346611478e-05, 'epoch': 1.67}) +[2026-06-16 10:06:21,166][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.03884756937623024, 'train/grad_norm_pre_clip': 17.425752639770508, 'timing/data': 0.022388409823179245, 'timing/model': 1.5461180987767875, 'learning_rate/qwen_vl_interface': 1.2119402087907448e-06, 'learning_rate/action_model': 1.2119402087907448e-05, 'epoch': 1.67}) +[2026-06-16 10:06:27,554][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.0010649641044437885, 'train/grad_norm_pre_clip': 0.3380383849143982, 'timing/data': 0.021534415893256664, 'timing/model': 1.6361184120178223, 'learning_rate/qwen_vl_interface': 1.2103373428141903e-06, 'learning_rate/action_model': 1.2103373428141903e-05, 'epoch': 1.67}) +[2026-06-16 10:06:33,880][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.10025615990161896, 'train/grad_norm_pre_clip': 19.60965919494629, 'timing/data': 0.021835577208548784, 'timing/model': 1.5882187588140368, 'learning_rate/qwen_vl_interface': 1.2087361373903633e-06, 'learning_rate/action_model': 1.2087361373903633e-05, 'epoch': 1.67}) +[2026-06-16 10:06:40,276][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.23840591311454773, 'train/grad_norm_pre_clip': 32.9302978515625, 'timing/data': 0.021125344093888998, 'timing/model': 1.5928109916858375, 'learning_rate/qwen_vl_interface': 1.2071365931774584e-06, 'learning_rate/action_model': 1.2071365931774584e-05, 'epoch': 1.67}) +[2026-06-16 10:06:46,702][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.040774378925561905, 'train/grad_norm_pre_clip': 20.103271484375, 'timing/data': 0.02148289605975151, 'timing/model': 1.7963950727134943, 'learning_rate/qwen_vl_interface': 1.2055387108329885e-06, 'learning_rate/action_model': 1.2055387108329886e-05, 'epoch': 1.67}) +[2026-06-16 10:06:52,930][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.002376871882006526, 'train/grad_norm_pre_clip': 1.7144964933395386, 'timing/data': 0.021086793392896652, 'timing/model': 1.7183399321511388, 'learning_rate/qwen_vl_interface': 1.2039424910137807e-06, 'learning_rate/action_model': 1.2039424910137806e-05, 'epoch': 1.67}) +[2026-06-16 10:06:58,968][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.00017380168719682842, 'train/grad_norm_pre_clip': 0.028091201558709145, 'timing/data': 0.021141574252396822, 'timing/model': 1.5555084221996367, 'learning_rate/qwen_vl_interface': 1.2023479343759822e-06, 'learning_rate/action_model': 1.2023479343759823e-05, 'epoch': 1.67}) +[2026-06-16 10:07:05,269][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.11129208654165268, 'train/grad_norm_pre_clip': 18.115415573120117, 'timing/data': 0.02140018530189991, 'timing/model': 1.554361438844353, 'learning_rate/qwen_vl_interface': 1.2007550415750546e-06, 'learning_rate/action_model': 1.2007550415750545e-05, 'epoch': 1.67}) +[2026-06-16 10:07:11,586][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.09343952685594559, 'train/grad_norm_pre_clip': 14.941974639892578, 'timing/data': 0.020610192324966192, 'timing/model': 1.5687269899062812, 'learning_rate/qwen_vl_interface': 1.1991638132657759e-06, 'learning_rate/action_model': 1.199163813265776e-05, 'epoch': 1.68}) +[2026-06-16 10:07:17,863][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.0268105436116457, 'train/grad_norm_pre_clip': 25.239212036132812, 'timing/data': 0.02358596632257104, 'timing/model': 1.5205328152514994, 'learning_rate/qwen_vl_interface': 1.1975742501022406e-06, 'learning_rate/action_model': 1.1975742501022405e-05, 'epoch': 1.68}) +[2026-06-16 10:07:24,222][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.1375114619731903, 'train/grad_norm_pre_clip': 9.135795593261719, 'timing/data': 0.021948177833110094, 'timing/model': 1.8008062141016126, 'learning_rate/qwen_vl_interface': 1.1959863527378569e-06, 'learning_rate/action_model': 1.1959863527378567e-05, 'epoch': 1.68}) +[2026-06-16 10:07:30,703][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.07899145036935806, 'train/grad_norm_pre_clip': 37.25646209716797, 'timing/data': 0.021504845935851336, 'timing/model': 1.81823163991794, 'learning_rate/qwen_vl_interface': 1.1944001218253505e-06, 'learning_rate/action_model': 1.1944001218253503e-05, 'epoch': 1.68}) +[2026-06-16 10:07:36,865][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.0007378854788839817, 'train/grad_norm_pre_clip': 0.17492449283599854, 'timing/data': 0.021391124930232763, 'timing/model': 1.5463782651349902, 'learning_rate/qwen_vl_interface': 1.1928155580167608e-06, 'learning_rate/action_model': 1.1928155580167608e-05, 'epoch': 1.68}) +[2026-06-16 10:07:43,161][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.12972316145896912, 'train/grad_norm_pre_clip': 14.892943382263184, 'timing/data': 0.021953987888991833, 'timing/model': 1.547612362075597, 'learning_rate/qwen_vl_interface': 1.1912326619634412e-06, 'learning_rate/action_model': 1.1912326619634412e-05, 'epoch': 1.68}) +[2026-06-16 10:07:49,573][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.009239518083631992, 'train/grad_norm_pre_clip': 10.59732437133789, 'timing/data': 0.021535745821893215, 'timing/model': 1.5564619358628988, 'learning_rate/qwen_vl_interface': 1.1896514343160632e-06, 'learning_rate/action_model': 1.189651434316063e-05, 'epoch': 1.68}) +[2026-06-16 10:07:55,962][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.00038885732647031546, 'train/grad_norm_pre_clip': 0.05182915925979614, 'timing/data': 0.020998393651098013, 'timing/model': 1.6031406773254275, 'learning_rate/qwen_vl_interface': 1.1880718757246062e-06, 'learning_rate/action_model': 1.1880718757246063e-05, 'epoch': 1.68}) +[2026-06-16 10:08:02,307][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.012677948921918869, 'train/grad_norm_pre_clip': 11.476339340209961, 'timing/data': 0.02149636624380946, 'timing/model': 1.7911624629050493, 'learning_rate/qwen_vl_interface': 1.186493986838369e-06, 'learning_rate/action_model': 1.186493986838369e-05, 'epoch': 1.68}) +[2026-06-16 10:08:08,366][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.0031290214974433184, 'train/grad_norm_pre_clip': 2.544412612915039, 'timing/data': 0.021990858018398285, 'timing/model': 1.5498282271437347, 'learning_rate/qwen_vl_interface': 1.1849177683059612e-06, 'learning_rate/action_model': 1.1849177683059612e-05, 'epoch': 1.68}) +[2026-06-16 10:08:14,670][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.14787538349628448, 'train/grad_norm_pre_clip': 53.43608093261719, 'timing/data': 0.021401495207101107, 'timing/model': 1.5526649723760784, 'learning_rate/qwen_vl_interface': 1.1833432207753075e-06, 'learning_rate/action_model': 1.1833432207753076e-05, 'epoch': 1.68}) +[2026-06-16 10:08:20,970][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.0011026777792721987, 'train/grad_norm_pre_clip': 0.2503286898136139, 'timing/data': 0.021109833847731352, 'timing/model': 1.5489448448643088, 'learning_rate/qwen_vl_interface': 1.1817703448936448e-06, 'learning_rate/action_model': 1.181770344893645e-05, 'epoch': 1.68}) +[2026-06-16 10:08:27,282][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.00195604027248919, 'train/grad_norm_pre_clip': 1.3183667659759521, 'timing/data': 0.021912848111242056, 'timing/model': 1.5358511907979846, 'learning_rate/qwen_vl_interface': 1.1801991413075228e-06, 'learning_rate/action_model': 1.1801991413075227e-05, 'epoch': 1.68}) +[2026-06-16 10:08:33,570][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.0006421045982278883, 'train/grad_norm_pre_clip': 0.12966902554035187, 'timing/data': 0.021185044199228287, 'timing/model': 1.573721569031477, 'learning_rate/qwen_vl_interface': 1.1786296106628038e-06, 'learning_rate/action_model': 1.1786296106628036e-05, 'epoch': 1.68}) +[2026-06-16 10:08:39,858][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.0007024264195933938, 'train/grad_norm_pre_clip': 0.19566960632801056, 'timing/data': 0.0201953100040555, 'timing/model': 1.739051674026996, 'learning_rate/qwen_vl_interface': 1.1770617536046633e-06, 'learning_rate/action_model': 1.1770617536046632e-05, 'epoch': 1.68}) +[2026-06-16 10:08:45,966][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.001168696559034288, 'train/grad_norm_pre_clip': 1.032434105873108, 'timing/data': 0.020787391811609268, 'timing/model': 1.553456830792129, 'learning_rate/qwen_vl_interface': 1.1754955707775873e-06, 'learning_rate/action_model': 1.1754955707775872e-05, 'epoch': 1.68}) +[2026-06-16 10:08:52,368][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.0475233756005764, 'train/grad_norm_pre_clip': 21.970603942871094, 'timing/data': 0.021345585118979216, 'timing/model': 1.5456601828336716, 'learning_rate/qwen_vl_interface': 1.1739310628253753e-06, 'learning_rate/action_model': 1.1739310628253753e-05, 'epoch': 1.68}) +[2026-06-16 10:08:58,782][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.012133430689573288, 'train/grad_norm_pre_clip': 18.188623428344727, 'timing/data': 0.020781132392585278, 'timing/model': 1.66132649499923, 'learning_rate/qwen_vl_interface': 1.172368230391136e-06, 'learning_rate/action_model': 1.1723682303911359e-05, 'epoch': 1.68}) +[2026-06-16 10:09:05,588][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.4025140106678009, 'train/grad_norm_pre_clip': 25.08243179321289, 'timing/data': 0.028758531901985407, 'timing/model': 1.6264559631235898, 'learning_rate/qwen_vl_interface': 1.1708070741172928e-06, 'learning_rate/action_model': 1.1708070741172928e-05, 'epoch': 1.68}) +[2026-06-16 10:09:12,096][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.23035749793052673, 'train/grad_norm_pre_clip': 22.559961318969727, 'timing/data': 0.029681857209652662, 'timing/model': 1.6137534617446363, 'learning_rate/qwen_vl_interface': 1.1692475946455755e-06, 'learning_rate/action_model': 1.1692475946455755e-05, 'epoch': 1.68}) +[2026-06-16 10:09:18,668][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.19607199728488922, 'train/grad_norm_pre_clip': 76.2537841796875, 'timing/data': 0.021645197179168463, 'timing/model': 2.0270232488401234, 'learning_rate/qwen_vl_interface': 1.1676897926170289e-06, 'learning_rate/action_model': 1.1676897926170289e-05, 'epoch': 1.68}) +[2026-06-16 10:09:24,900][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.1467738300561905, 'train/grad_norm_pre_clip': 11.230554580688477, 'timing/data': 0.027601836249232292, 'timing/model': 1.6022036168724298, 'learning_rate/qwen_vl_interface': 1.166133668672006e-06, 'learning_rate/action_model': 1.166133668672006e-05, 'epoch': 1.68}) +[2026-06-16 10:09:31,468][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.0017686160281300545, 'train/grad_norm_pre_clip': 0.4298919141292572, 'timing/data': 0.021279385313391685, 'timing/model': 1.5338142709806561, 'learning_rate/qwen_vl_interface': 1.1645792234501706e-06, 'learning_rate/action_model': 1.1645792234501706e-05, 'epoch': 1.68}) +[2026-06-16 10:09:38,195][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.0005926418234594166, 'train/grad_norm_pre_clip': 0.11340699344873428, 'timing/data': 0.02805380802601576, 'timing/model': 1.6203645090572536, 'learning_rate/qwen_vl_interface': 1.1630264575904972e-06, 'learning_rate/action_model': 1.1630264575904973e-05, 'epoch': 1.68}) +[2026-06-16 10:09:44,883][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.004957317374646664, 'train/grad_norm_pre_clip': 2.254866361618042, 'timing/data': 0.02679105196148157, 'timing/model': 1.6234345040284097, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 1.69}) +[2026-06-16 10:09:51,510][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.0010938311461359262, 'train/grad_norm_pre_clip': 0.20213638246059418, 'timing/data': 0.026024858932942152, 'timing/model': 1.988418925087899, 'learning_rate/qwen_vl_interface': 1.1599259665100767e-06, 'learning_rate/action_model': 1.1599259665100767e-05, 'epoch': 1.69}) +[2026-06-16 10:09:57,561][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.04319166764616966, 'train/grad_norm_pre_clip': 30.96274185180664, 'timing/data': 0.021247105207294226, 'timing/model': 1.547242479864508, 'learning_rate/qwen_vl_interface': 1.1583782425638252e-06, 'learning_rate/action_model': 1.158378242563825e-05, 'epoch': 1.69}) +[2026-06-16 10:10:03,864][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.00493831979110837, 'train/grad_norm_pre_clip': 4.294095039367676, 'timing/data': 0.02093214262276888, 'timing/model': 1.5459171049296856, 'learning_rate/qwen_vl_interface': 1.1568322005287237e-06, 'learning_rate/action_model': 1.1568322005287235e-05, 'epoch': 1.69}) +[2026-06-16 10:10:10,161][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.0480055995285511, 'train/grad_norm_pre_clip': 24.91617774963379, 'timing/data': 0.020400790963321924, 'timing/model': 1.570241984911263, 'learning_rate/qwen_vl_interface': 1.1552878410402923e-06, 'learning_rate/action_model': 1.1552878410402924e-05, 'epoch': 1.69}) +[2026-06-16 10:10:16,502][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.16441085934638977, 'train/grad_norm_pre_clip': 11.265514373779297, 'timing/data': 0.021588307339698076, 'timing/model': 1.5520613356493413, 'learning_rate/qwen_vl_interface': 1.1537451647333594e-06, 'learning_rate/action_model': 1.1537451647333593e-05, 'epoch': 1.69}) +[2026-06-16 10:10:22,874][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.0013549493160098791, 'train/grad_norm_pre_clip': 0.5811579823493958, 'timing/data': 0.02192621771246195, 'timing/model': 1.57191980490461, 'learning_rate/qwen_vl_interface': 1.1522041722420594e-06, 'learning_rate/action_model': 1.1522041722420594e-05, 'epoch': 1.69}) +[2026-06-16 10:10:29,246][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.3033433258533478, 'train/grad_norm_pre_clip': 28.47903823852539, 'timing/data': 0.021543155889958143, 'timing/model': 1.8229538649320602, 'learning_rate/qwen_vl_interface': 1.1506648641998365e-06, 'learning_rate/action_model': 1.1506648641998365e-05, 'epoch': 1.69}) +[2026-06-16 10:10:35,279][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 0.00012979585153516382, 'train/grad_norm_pre_clip': 0.017835024744272232, 'timing/data': 0.022067878860980272, 'timing/model': 1.553250775206834, 'learning_rate/qwen_vl_interface': 1.1491272412394433e-06, 'learning_rate/action_model': 1.1491272412394433e-05, 'epoch': 1.69}) +[2026-06-16 10:10:41,683][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.0005964503507129848, 'train/grad_norm_pre_clip': 0.1601821333169937, 'timing/data': 0.02006204891949892, 'timing/model': 1.5361653612926602, 'learning_rate/qwen_vl_interface': 1.1475913039929373e-06, 'learning_rate/action_model': 1.1475913039929372e-05, 'epoch': 1.69}) +[2026-06-16 10:10:48,078][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.04428991302847862, 'train/grad_norm_pre_clip': 23.70071029663086, 'timing/data': 0.02204918907955289, 'timing/model': 1.5646413932554424, 'learning_rate/qwen_vl_interface': 1.1460570530916844e-06, 'learning_rate/action_model': 1.1460570530916844e-05, 'epoch': 1.69}) +[2026-06-16 10:10:54,480][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.0005830464651808143, 'train/grad_norm_pre_clip': 0.18249809741973877, 'timing/data': 0.021794287022203207, 'timing/model': 1.562708543613553, 'learning_rate/qwen_vl_interface': 1.1445244891663581e-06, 'learning_rate/action_model': 1.1445244891663581e-05, 'epoch': 1.69}) +[2026-06-16 10:11:00,763][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 0.06100376322865486, 'train/grad_norm_pre_clip': 24.60886001586914, 'timing/data': 0.019509215839207172, 'timing/model': 1.5472093881107867, 'learning_rate/qwen_vl_interface': 1.1429936128469366e-06, 'learning_rate/action_model': 1.1429936128469366e-05, 'epoch': 1.69}) +[2026-06-16 10:11:07,102][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.00041124733979813755, 'train/grad_norm_pre_clip': 0.08672560751438141, 'timing/data': 0.021334254648536444, 'timing/model': 1.7925531212240458, 'learning_rate/qwen_vl_interface': 1.1414644247627056e-06, 'learning_rate/action_model': 1.1414644247627058e-05, 'epoch': 1.69}) +[2026-06-16 10:11:13,230][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.0005093691870570183, 'train/grad_norm_pre_clip': 0.1551671326160431, 'timing/data': 0.01974675804376602, 'timing/model': 1.6243755011819303, 'learning_rate/qwen_vl_interface': 1.1399369255422568e-06, 'learning_rate/action_model': 1.1399369255422568e-05, 'epoch': 1.69}) +[2026-06-16 10:11:19,461][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.06476829200983047, 'train/grad_norm_pre_clip': 23.642122268676758, 'timing/data': 0.020476261153817177, 'timing/model': 1.5512311910279095, 'learning_rate/qwen_vl_interface': 1.138411115813488e-06, 'learning_rate/action_model': 1.138411115813488e-05, 'epoch': 1.69}) +[2026-06-16 10:11:25,763][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.007240224163979292, 'train/grad_norm_pre_clip': 9.520502090454102, 'timing/data': 0.02171485684812069, 'timing/model': 1.5448263389989734, 'learning_rate/qwen_vl_interface': 1.1368869962036006e-06, 'learning_rate/action_model': 1.1368869962036005e-05, 'epoch': 1.69}) +[2026-06-16 10:11:32,067][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.0009524990455247462, 'train/grad_norm_pre_clip': 0.41168299317359924, 'timing/data': 0.020438520703464746, 'timing/model': 1.5611298303119838, 'learning_rate/qwen_vl_interface': 1.1353645673391025e-06, 'learning_rate/action_model': 1.1353645673391026e-05, 'epoch': 1.69}) +[2026-06-16 10:11:38,441][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.0005011907196603715, 'train/grad_norm_pre_clip': 0.18743927776813507, 'timing/data': 0.021024613175541162, 'timing/model': 1.5862498856149614, 'learning_rate/qwen_vl_interface': 1.1338438298458078e-06, 'learning_rate/action_model': 1.1338438298458078e-05, 'epoch': 1.69}) +[2026-06-16 10:11:44,702][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.000641757040284574, 'train/grad_norm_pre_clip': 0.1361112892627716, 'timing/data': 0.02025997033342719, 'timing/model': 1.7907462967559695, 'learning_rate/qwen_vl_interface': 1.1323247843488345e-06, 'learning_rate/action_model': 1.1323247843488344e-05, 'epoch': 1.69}) +[2026-06-16 10:11:50,764][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.09179595857858658, 'train/grad_norm_pre_clip': 32.471309661865234, 'timing/data': 0.021003372967243195, 'timing/model': 1.5541737889871001, 'learning_rate/qwen_vl_interface': 1.130807431472603e-06, 'learning_rate/action_model': 1.130807431472603e-05, 'epoch': 1.69}) +[2026-06-16 10:11:57,157][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.0006708002183586359, 'train/grad_norm_pre_clip': 0.1649487316608429, 'timing/data': 0.02012581005692482, 'timing/model': 1.6434157588519156, 'learning_rate/qwen_vl_interface': 1.1292917718408416e-06, 'learning_rate/action_model': 1.1292917718408416e-05, 'epoch': 1.69}) +[2026-06-16 10:12:03,493][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.0004598059749696404, 'train/grad_norm_pre_clip': 0.13154344260692596, 'timing/data': 0.021576936822384596, 'timing/model': 1.5702313189394772, 'learning_rate/qwen_vl_interface': 1.12777780607658e-06, 'learning_rate/action_model': 1.12777780607658e-05, 'epoch': 1.69}) +[2026-06-16 10:12:09,770][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.008573641069233418, 'train/grad_norm_pre_clip': 7.55881929397583, 'timing/data': 0.02085872320458293, 'timing/model': 1.55186577886343, 'learning_rate/qwen_vl_interface': 1.1262655348021523e-06, 'learning_rate/action_model': 1.1262655348021523e-05, 'epoch': 1.69}) +[2026-06-16 10:12:16,274][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.0003276521747466177, 'train/grad_norm_pre_clip': 0.07346685230731964, 'timing/data': 0.021079764235764742, 'timing/model': 1.5947998519986868, 'learning_rate/qwen_vl_interface': 1.1247549586391966e-06, 'learning_rate/action_model': 1.1247549586391964e-05, 'epoch': 1.69}) +[2026-06-16 10:12:22,713][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.07580358535051346, 'train/grad_norm_pre_clip': 52.15835952758789, 'timing/data': 0.021571396850049496, 'timing/model': 1.81909618107602, 'learning_rate/qwen_vl_interface': 1.1232460782086534e-06, 'learning_rate/action_model': 1.1232460782086534e-05, 'epoch': 1.7}) +[2026-06-16 10:12:28,770][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.007525227032601833, 'train/grad_norm_pre_clip': 7.285906791687012, 'timing/data': 0.02141970582306385, 'timing/model': 1.5487881847657263, 'learning_rate/qwen_vl_interface': 1.1217388941307658e-06, 'learning_rate/action_model': 1.1217388941307658e-05, 'epoch': 1.7}) +[2026-06-16 10:12:35,064][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.005730350501835346, 'train/grad_norm_pre_clip': 3.9656503200531006, 'timing/data': 0.021523957140743732, 'timing/model': 1.5517410798929632, 'learning_rate/qwen_vl_interface': 1.1202334070250807e-06, 'learning_rate/action_model': 1.1202334070250806e-05, 'epoch': 1.7}) +[2026-06-16 10:12:41,368][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.004034463781863451, 'train/grad_norm_pre_clip': 4.714249610900879, 'timing/data': 0.019230665173381567, 'timing/model': 1.528692967724055, 'learning_rate/qwen_vl_interface': 1.1187296175104472e-06, 'learning_rate/action_model': 1.1187296175104471e-05, 'epoch': 1.7}) +[2026-06-16 10:12:47,762][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.0018311669118702412, 'train/grad_norm_pre_clip': 1.2662441730499268, 'timing/data': 0.020680482033640146, 'timing/model': 1.5485698557458818, 'learning_rate/qwen_vl_interface': 1.1172275262050158e-06, 'learning_rate/action_model': 1.1172275262050159e-05, 'epoch': 1.7}) +[2026-06-16 10:12:54,171][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.12297868728637695, 'train/grad_norm_pre_clip': 36.2620735168457, 'timing/data': 0.021232045255601406, 'timing/model': 1.5692273187451065, 'learning_rate/qwen_vl_interface': 1.1157271337262401e-06, 'learning_rate/action_model': 1.11572713372624e-05, 'epoch': 1.7}) +[2026-06-16 10:13:00,519][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 7.679629197809845e-05, 'train/grad_norm_pre_clip': 0.030035318806767464, 'timing/data': 0.020485761109739542, 'timing/model': 1.7938016401603818, 'learning_rate/qwen_vl_interface': 1.1142284406908751e-06, 'learning_rate/action_model': 1.114228440690875e-05, 'epoch': 1.7}) +[2026-06-16 10:13:06,664][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.01626863330602646, 'train/grad_norm_pre_clip': 11.479632377624512, 'timing/data': 0.020636542234569788, 'timing/model': 1.5375423938967288, 'learning_rate/qwen_vl_interface': 1.1127314477149758e-06, 'learning_rate/action_model': 1.1127314477149759e-05, 'epoch': 1.7}) +[2026-06-16 10:24:02,813][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.008139965124428272, 'train/grad_norm_pre_clip': 4.160092353820801, 'eval/loss': 0.0372663289308548, 'eval/flap_precision': 0.9651006711409396, 'eval/flap_recall': 0.9809004092769441, 'eval/flap_f1': 0.972936400541272, 'eval/flap_rate_model': 0.20694444444444443, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.972936400541272, 'timing/data': 0.02037770999595523, 'timing/model': 1.5547754196450114, 'learning_rate/qwen_vl_interface': 1.1112361554139003e-06, 'learning_rate/action_model': 1.1112361554139004e-05, 'epoch': 1.7}) +[2026-06-16 10:24:02,814][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4200_state +[2026-06-16 10:24:02,837][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 10:24:11,244][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4200_state/model.safetensors +[2026-06-16 10:24:21,996][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4200_state/optimizer.bin +[2026-06-16 10:24:21,997][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4200_state/sampler.bin +[2026-06-16 10:24:21,997][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4200_state/sampler_1.bin +[2026-06-16 10:24:21,999][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4200_state/random_states_0.pkl +[2026-06-16 10:24:22,421][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4200 +[2026-06-16 10:24:22,421][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 10:24:22,436][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 10:24:28,440][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.0009937732247635722, 'train/grad_norm_pre_clip': 0.5409221053123474, 'timing/data': 0.02159538958221674, 'timing/model': 1.555407544132322, 'learning_rate/qwen_vl_interface': 1.1097425644023065e-06, 'learning_rate/action_model': 1.1097425644023065e-05, 'epoch': 1.7}) +[2026-06-16 10:24:34,648][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 0.001347068347968161, 'train/grad_norm_pre_clip': 0.7846601605415344, 'timing/data': 0.02146653924137354, 'timing/model': 1.5537292049266398, 'learning_rate/qwen_vl_interface': 1.1082506752941529e-06, 'learning_rate/action_model': 1.1082506752941528e-05, 'epoch': 1.7}) +[2026-06-16 10:24:40,855][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.0052570863626897335, 'train/grad_norm_pre_clip': 40.519588470458984, 'timing/data': 0.021155265159904957, 'timing/model': 1.522137759719044, 'learning_rate/qwen_vl_interface': 1.1067604887027004e-06, 'learning_rate/action_model': 1.1067604887027004e-05, 'epoch': 1.7}) +[2026-06-16 10:24:47,086][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.001808888977393508, 'train/grad_norm_pre_clip': 0.9184505343437195, 'timing/data': 0.020625710021704435, 'timing/model': 1.5933524160645902, 'learning_rate/qwen_vl_interface': 1.1052720052405066e-06, 'learning_rate/action_model': 1.1052720052405064e-05, 'epoch': 1.7}) +[2026-06-16 10:24:53,349][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.018520165234804153, 'train/grad_norm_pre_clip': 11.52144718170166, 'timing/data': 0.021921539679169655, 'timing/model': 1.5562051562592387, 'learning_rate/qwen_vl_interface': 1.1037852255194309e-06, 'learning_rate/action_model': 1.1037852255194308e-05, 'epoch': 1.7}) +[2026-06-16 10:24:59,626][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 0.012195128947496414, 'train/grad_norm_pre_clip': 5.223080158233643, 'timing/data': 0.022645905148237944, 'timing/model': 1.8258611797355115, 'learning_rate/qwen_vl_interface': 1.1023001501506329e-06, 'learning_rate/action_model': 1.102300150150633e-05, 'epoch': 1.7}) +[2026-06-16 10:25:05,655][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.000468828366138041, 'train/grad_norm_pre_clip': 0.16585363447666168, 'timing/data': 0.020884240046143532, 'timing/model': 1.5567059782333672, 'learning_rate/qwen_vl_interface': 1.1008167797445717e-06, 'learning_rate/action_model': 1.1008167797445718e-05, 'epoch': 1.7}) +[2026-06-16 10:25:11,959][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.00025718577671796083, 'train/grad_norm_pre_clip': 0.05902917683124542, 'timing/data': 0.022223851177841425, 'timing/model': 1.5564705249853432, 'learning_rate/qwen_vl_interface': 1.099335114911003e-06, 'learning_rate/action_model': 1.0993351149110031e-05, 'epoch': 1.7}) +[2026-06-16 10:25:18,358][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.006195264868438244, 'train/grad_norm_pre_clip': 6.286880016326904, 'timing/data': 0.021314541809260845, 'timing/model': 1.564852821175009, 'learning_rate/qwen_vl_interface': 1.0978551562589846e-06, 'learning_rate/action_model': 1.0978551562589845e-05, 'epoch': 1.7}) +[2026-06-16 10:25:24,785][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.00020761154883075505, 'train/grad_norm_pre_clip': 0.07516888529062271, 'timing/data': 0.02063025627285242, 'timing/model': 1.5928450552746654, 'learning_rate/qwen_vl_interface': 1.0963769043968708e-06, 'learning_rate/action_model': 1.0963769043968708e-05, 'epoch': 1.7}) +[2026-06-16 10:25:31,311][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.0008506884914822876, 'train/grad_norm_pre_clip': 0.45507675409317017, 'timing/data': 0.02478384878486395, 'timing/model': 1.799793042242527, 'learning_rate/qwen_vl_interface': 1.0949003599323158e-06, 'learning_rate/action_model': 1.0949003599323157e-05, 'epoch': 1.7}) +[2026-06-16 10:25:37,580][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 0.11191040277481079, 'train/grad_norm_pre_clip': 11.203378677368164, 'timing/data': 0.021945504937320948, 'timing/model': 1.7743534930050373, 'learning_rate/qwen_vl_interface': 1.0934255234722708e-06, 'learning_rate/action_model': 1.0934255234722706e-05, 'epoch': 1.7}) +[2026-06-16 10:25:43,657][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.01037128921598196, 'train/grad_norm_pre_clip': 7.890796184539795, 'timing/data': 0.02111826790496707, 'timing/model': 1.5543651687912643, 'learning_rate/qwen_vl_interface': 1.0919523956229858e-06, 'learning_rate/action_model': 1.0919523956229857e-05, 'epoch': 1.7}) +[2026-06-16 10:25:50,061][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.21835868060588837, 'train/grad_norm_pre_clip': 9.092309951782227, 'timing/data': 0.023655917961150408, 'timing/model': 1.5514222090132535, 'learning_rate/qwen_vl_interface': 1.0904809769900075e-06, 'learning_rate/action_model': 1.0904809769900073e-05, 'epoch': 1.7}) +[2026-06-16 10:25:56,456][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.232938751578331, 'train/grad_norm_pre_clip': 7.1083550453186035, 'timing/data': 0.02173721231520176, 'timing/model': 1.554796386975795, 'learning_rate/qwen_vl_interface': 1.0890112681781821e-06, 'learning_rate/action_model': 1.089011268178182e-05, 'epoch': 1.7}) +[2026-06-16 10:26:02,863][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.00859813578426838, 'train/grad_norm_pre_clip': 11.360220909118652, 'timing/data': 0.0206488324329257, 'timing/model': 1.5898007061332464, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 1.7}) +[2026-06-16 10:26:09,272][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 3.9867616578703746e-05, 'train/grad_norm_pre_clip': 0.007175232749432325, 'timing/data': 0.0212767468765378, 'timing/model': 1.578023524954915, 'learning_rate/qwen_vl_interface': 1.0860769824338477e-06, 'learning_rate/action_model': 1.0860769824338478e-05, 'epoch': 1.71}) +[2026-06-16 10:26:15,601][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.0018408790929242969, 'train/grad_norm_pre_clip': 1.0773842334747314, 'timing/data': 0.021517067681998014, 'timing/model': 1.7851247959770262, 'learning_rate/qwen_vl_interface': 1.084612406707514e-06, 'learning_rate/action_model': 1.084612406707514e-05, 'epoch': 1.71}) +[2026-06-16 10:26:21,661][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.2685638964176178, 'train/grad_norm_pre_clip': 27.08323860168457, 'timing/data': 0.020353087689727545, 'timing/model': 1.553795237094164, 'learning_rate/qwen_vl_interface': 1.0831495432146795e-06, 'learning_rate/action_model': 1.0831495432146795e-05, 'epoch': 1.71}) +[2026-06-16 10:26:27,965][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.10464116185903549, 'train/grad_norm_pre_clip': 22.772642135620117, 'timing/data': 0.02044368814677, 'timing/model': 1.5558902299962938, 'learning_rate/qwen_vl_interface': 1.0816883925566717e-06, 'learning_rate/action_model': 1.0816883925566716e-05, 'epoch': 1.71}) +[2026-06-16 10:26:34,362][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.00019842457550112158, 'train/grad_norm_pre_clip': 0.0281374491751194, 'timing/data': 0.02125089382752776, 'timing/model': 1.5516775622963905, 'learning_rate/qwen_vl_interface': 1.0802289553341153e-06, 'learning_rate/action_model': 1.0802289553341152e-05, 'epoch': 1.71}) +[2026-06-16 10:26:40,760][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.11831672489643097, 'train/grad_norm_pre_clip': 37.56060028076172, 'timing/data': 0.021159403026103973, 'timing/model': 1.5814464879222214, 'learning_rate/qwen_vl_interface': 1.0787712321469289e-06, 'learning_rate/action_model': 1.0787712321469288e-05, 'epoch': 1.71}) +[2026-06-16 10:26:47,127][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.03700483962893486, 'train/grad_norm_pre_clip': 26.279170989990234, 'timing/data': 0.021148572210222483, 'timing/model': 1.8187838438898325, 'learning_rate/qwen_vl_interface': 1.077315223594328e-06, 'learning_rate/action_model': 1.077315223594328e-05, 'epoch': 1.71}) +[2026-06-16 10:26:53,167][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.2937748432159424, 'train/grad_norm_pre_clip': 6.657001972198486, 'timing/data': 0.021276012994349003, 'timing/model': 1.5493195578455925, 'learning_rate/qwen_vl_interface': 1.0758609302748227e-06, 'learning_rate/action_model': 1.0758609302748227e-05, 'epoch': 1.71}) +[2026-06-16 10:26:59,467][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.16691666841506958, 'train/grad_norm_pre_clip': 33.22517776489258, 'timing/data': 0.02140529314056039, 'timing/model': 1.549548908136785, 'learning_rate/qwen_vl_interface': 1.0744083527862182e-06, 'learning_rate/action_model': 1.074408352786218e-05, 'epoch': 1.71}) +[2026-06-16 10:27:05,768][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.046817220747470856, 'train/grad_norm_pre_clip': 22.371570587158203, 'timing/data': 0.021423263009637594, 'timing/model': 1.5302773513831198, 'learning_rate/qwen_vl_interface': 1.0729574917256144e-06, 'learning_rate/action_model': 1.0729574917256145e-05, 'epoch': 1.71}) +[2026-06-16 10:27:12,061][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.058609750121831894, 'train/grad_norm_pre_clip': 36.19645309448242, 'timing/data': 0.02052462473511696, 'timing/model': 1.5541161722503603, 'learning_rate/qwen_vl_interface': 1.0715083476894047e-06, 'learning_rate/action_model': 1.0715083476894047e-05, 'epoch': 1.71}) +[2026-06-16 10:27:18,480][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 0.0003476366982795298, 'train/grad_norm_pre_clip': 0.08674030750989914, 'timing/data': 0.021024707704782486, 'timing/model': 1.537328664213419, 'learning_rate/qwen_vl_interface': 1.0700609212732773e-06, 'learning_rate/action_model': 1.0700609212732774e-05, 'epoch': 1.71}) +[2026-06-16 10:27:24,840][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.0001411059347447008, 'train/grad_norm_pre_clip': 0.020401114597916603, 'timing/data': 0.02179446443915367, 'timing/model': 1.8158945362083614, 'learning_rate/qwen_vl_interface': 1.0686152130722155e-06, 'learning_rate/action_model': 1.0686152130722154e-05, 'epoch': 1.71}) +[2026-06-16 10:27:30,866][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.0009374848450534046, 'train/grad_norm_pre_clip': 0.6552987694740295, 'timing/data': 0.022205696906894445, 'timing/model': 1.5435024737380445, 'learning_rate/qwen_vl_interface': 1.0671712236804947e-06, 'learning_rate/action_model': 1.0671712236804946e-05, 'epoch': 1.71}) +[2026-06-16 10:27:37,171][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 0.003932396415621042, 'train/grad_norm_pre_clip': 1.9274451732635498, 'timing/data': 0.02031186083331704, 'timing/model': 1.5244841952808201, 'learning_rate/qwen_vl_interface': 1.0657289536916845e-06, 'learning_rate/action_model': 1.0657289536916845e-05, 'epoch': 1.71}) +[2026-06-16 10:27:43,567][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.0003585806116461754, 'train/grad_norm_pre_clip': 0.06495099514722824, 'timing/data': 0.0210712063126266, 'timing/model': 1.5453376811929047, 'learning_rate/qwen_vl_interface': 1.0642884036986481e-06, 'learning_rate/action_model': 1.064288403698648e-05, 'epoch': 1.71}) +[2026-06-16 10:27:50,065][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.008882803842425346, 'train/grad_norm_pre_clip': 5.647251129150391, 'timing/data': 0.02202622313052416, 'timing/model': 1.5504430113360286, 'learning_rate/qwen_vl_interface': 1.0628495742935404e-06, 'learning_rate/action_model': 1.0628495742935405e-05, 'epoch': 1.71}) +[2026-06-16 10:27:56,462][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 0.08936980366706848, 'train/grad_norm_pre_clip': 18.545392990112305, 'timing/data': 0.021240506786853075, 'timing/model': 1.599222432821989, 'learning_rate/qwen_vl_interface': 1.0614124660678103e-06, 'learning_rate/action_model': 1.0614124660678102e-05, 'epoch': 1.71}) +[2026-06-16 10:28:02,843][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.0038593201898038387, 'train/grad_norm_pre_clip': 3.0566036701202393, 'timing/data': 0.02148431772366166, 'timing/model': 1.8153144656680524, 'learning_rate/qwen_vl_interface': 1.0599770796121988e-06, 'learning_rate/action_model': 1.0599770796121987e-05, 'epoch': 1.71}) +[2026-06-16 10:28:08,872][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.048479076474905014, 'train/grad_norm_pre_clip': 39.41799545288086, 'timing/data': 0.02115293499082327, 'timing/model': 1.5466834641993046, 'learning_rate/qwen_vl_interface': 1.0585434155167394e-06, 'learning_rate/action_model': 1.0585434155167393e-05, 'epoch': 1.71}) +[2026-06-16 10:28:15,283][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.049765899777412415, 'train/grad_norm_pre_clip': 60.29631805419922, 'timing/data': 0.021168155129998922, 'timing/model': 1.5627225949428976, 'learning_rate/qwen_vl_interface': 1.0571114743707556e-06, 'learning_rate/action_model': 1.0571114743707555e-05, 'epoch': 1.71}) +[2026-06-16 10:28:21,678][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.005064710974693298, 'train/grad_norm_pre_clip': 2.4673092365264893, 'timing/data': 0.020484389271587133, 'timing/model': 1.5636790259741247, 'learning_rate/qwen_vl_interface': 1.0556812567628661e-06, 'learning_rate/action_model': 1.055681256762866e-05, 'epoch': 1.71}) +[2026-06-16 10:28:28,029][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.025051839649677277, 'train/grad_norm_pre_clip': 24.156436920166016, 'timing/data': 0.023213298991322517, 'timing/model': 1.5985600450076163, 'learning_rate/qwen_vl_interface': 1.0542527632809783e-06, 'learning_rate/action_model': 1.0542527632809784e-05, 'epoch': 1.71}) +[2026-06-16 10:28:34,363][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.0311521477997303, 'train/grad_norm_pre_clip': 19.86623764038086, 'timing/data': 0.023007526993751526, 'timing/model': 1.5534301022998989, 'learning_rate/qwen_vl_interface': 1.0528259945122925e-06, 'learning_rate/action_model': 1.0528259945122925e-05, 'epoch': 1.71}) +[2026-06-16 10:28:40,828][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.0059615690261125565, 'train/grad_norm_pre_clip': 3.5752906799316406, 'timing/data': 0.021737967152148485, 'timing/model': 1.8529616300947964, 'learning_rate/qwen_vl_interface': 1.0514009510432988e-06, 'learning_rate/action_model': 1.0514009510432988e-05, 'epoch': 1.72}) +[2026-06-16 10:28:46,866][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 4.853490463574417e-05, 'train/grad_norm_pre_clip': 0.005248070694506168, 'timing/data': 0.020789459347724915, 'timing/model': 1.555727700702846, 'learning_rate/qwen_vl_interface': 1.0499776334597796e-06, 'learning_rate/action_model': 1.0499776334597797e-05, 'epoch': 1.72}) +[2026-06-16 10:28:53,265][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.00041088179568760097, 'train/grad_norm_pre_clip': 0.1276043802499771, 'timing/data': 0.020193685311824083, 'timing/model': 1.5536270220763981, 'learning_rate/qwen_vl_interface': 1.0485560423468064e-06, 'learning_rate/action_model': 1.0485560423468063e-05, 'epoch': 1.72}) +[2026-06-16 10:28:59,583][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.00023435233742929995, 'train/grad_norm_pre_clip': 0.04748659208416939, 'timing/data': 0.022833464201539755, 'timing/model': 1.567727998830378, 'learning_rate/qwen_vl_interface': 1.0471361782887421e-06, 'learning_rate/action_model': 1.0471361782887421e-05, 'epoch': 1.72}) +[2026-06-16 10:29:05,886][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.002150774234905839, 'train/grad_norm_pre_clip': 0.5200149416923523, 'timing/data': 0.021397052332758904, 'timing/model': 1.5728878350928426, 'learning_rate/qwen_vl_interface': 1.0457180418692385e-06, 'learning_rate/action_model': 1.0457180418692383e-05, 'epoch': 1.72}) +[2026-06-16 10:29:12,378][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.0015399125404655933, 'train/grad_norm_pre_clip': 0.5899545550346375, 'timing/data': 0.02121647074818611, 'timing/model': 1.6016751849092543, 'learning_rate/qwen_vl_interface': 1.0443016336712395e-06, 'learning_rate/action_model': 1.0443016336712393e-05, 'epoch': 1.72}) +[2026-06-16 10:29:18,759][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.00261073955334723, 'train/grad_norm_pre_clip': 1.8737592697143555, 'timing/data': 0.02120622107759118, 'timing/model': 1.839362339116633, 'learning_rate/qwen_vl_interface': 1.0428869542769744e-06, 'learning_rate/action_model': 1.0428869542769744e-05, 'epoch': 1.72}) +[2026-06-16 10:29:24,870][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 6.733807094860822e-05, 'train/grad_norm_pre_clip': 0.009790790267288685, 'timing/data': 0.02090057823807001, 'timing/model': 1.5504740891046822, 'learning_rate/qwen_vl_interface': 1.0414740042679663e-06, 'learning_rate/action_model': 1.0414740042679662e-05, 'epoch': 1.72}) +[2026-06-16 10:29:31,268][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.13156816363334656, 'train/grad_norm_pre_clip': 33.574745178222656, 'timing/data': 0.021552452351897955, 'timing/model': 1.550403056666255, 'learning_rate/qwen_vl_interface': 1.040062784225025e-06, 'learning_rate/action_model': 1.0400627842250249e-05, 'epoch': 1.72}) +[2026-06-16 10:29:37,673][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.13324002921581268, 'train/grad_norm_pre_clip': 19.581239700317383, 'timing/data': 0.021733513101935387, 'timing/model': 1.5501449941657484, 'learning_rate/qwen_vl_interface': 1.0386532947282496e-06, 'learning_rate/action_model': 1.0386532947282495e-05, 'epoch': 1.72}) +[2026-06-16 10:29:43,992][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.0036407040897756815, 'train/grad_norm_pre_clip': 4.102034568786621, 'timing/data': 0.020119770895689726, 'timing/model': 1.5721958098001778, 'learning_rate/qwen_vl_interface': 1.0372455363570282e-06, 'learning_rate/action_model': 1.0372455363570283e-05, 'epoch': 1.72}) +[2026-06-16 10:29:50,478][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.0005446247523650527, 'train/grad_norm_pre_clip': 0.19392946362495422, 'timing/data': 0.01842963881790638, 'timing/model': 1.6799455522559583, 'learning_rate/qwen_vl_interface': 1.035839509690039e-06, 'learning_rate/action_model': 1.0358395096900388e-05, 'epoch': 1.72}) +[2026-06-16 10:29:56,853][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.0001541112142149359, 'train/grad_norm_pre_clip': 0.037110812962055206, 'timing/data': 0.021458689589053392, 'timing/model': 1.8254502662457526, 'learning_rate/qwen_vl_interface': 1.0344352153052429e-06, 'learning_rate/action_model': 1.0344352153052427e-05, 'epoch': 1.72}) +[2026-06-16 10:30:03,071][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.00028091942658647895, 'train/grad_norm_pre_clip': 0.0501575767993927, 'timing/data': 0.021939392667263746, 'timing/model': 1.5706912386231124, 'learning_rate/qwen_vl_interface': 1.033032653779894e-06, 'learning_rate/action_model': 1.033032653779894e-05, 'epoch': 1.72}) +[2026-06-16 10:30:09,476][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.04843955114483833, 'train/grad_norm_pre_clip': 33.3100471496582, 'timing/data': 0.021011896431446075, 'timing/model': 1.5569649538956583, 'learning_rate/qwen_vl_interface': 1.0316318256905325e-06, 'learning_rate/action_model': 1.0316318256905324e-05, 'epoch': 1.72}) +[2026-06-16 10:30:15,773][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.21402251720428467, 'train/grad_norm_pre_clip': 9.327713966369629, 'timing/data': 0.0202498696744442, 'timing/model': 1.550951394252479, 'learning_rate/qwen_vl_interface': 1.0302327316129856e-06, 'learning_rate/action_model': 1.0302327316129855e-05, 'epoch': 1.72}) +[2026-06-16 10:30:22,179][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.0006667044945061207, 'train/grad_norm_pre_clip': 0.19713617861270905, 'timing/data': 0.02152331918478012, 'timing/model': 1.5474173524416983, 'learning_rate/qwen_vl_interface': 1.028835372122368e-06, 'learning_rate/action_model': 1.0288353721223682e-05, 'epoch': 1.72}) +[2026-06-16 10:30:28,590][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 0.0002638478472363204, 'train/grad_norm_pre_clip': 0.07707573473453522, 'timing/data': 0.022296832874417305, 'timing/model': 1.5976610998623073, 'learning_rate/qwen_vl_interface': 1.0274397477930822e-06, 'learning_rate/action_model': 1.0274397477930822e-05, 'epoch': 1.72}) +[2026-06-16 10:30:34,938][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.1395273357629776, 'train/grad_norm_pre_clip': 37.47960662841797, 'timing/data': 0.021399426739662886, 'timing/model': 1.817603612318635, 'learning_rate/qwen_vl_interface': 1.0260458591988164e-06, 'learning_rate/action_model': 1.0260458591988165e-05, 'epoch': 1.72}) +[2026-06-16 10:30:41,169][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.2954038977622986, 'train/grad_norm_pre_clip': 47.35348129272461, 'timing/data': 0.02102523483335972, 'timing/model': 1.7509598638862371, 'learning_rate/qwen_vl_interface': 1.0246537069125445e-06, 'learning_rate/action_model': 1.0246537069125444e-05, 'epoch': 1.72}) +[2026-06-16 10:30:47,373][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.00060239527374506, 'train/grad_norm_pre_clip': 0.16458645462989807, 'timing/data': 0.021015324164181948, 'timing/model': 1.559864188078791, 'learning_rate/qwen_vl_interface': 1.0232632915065286e-06, 'learning_rate/action_model': 1.0232632915065285e-05, 'epoch': 1.72}) +[2026-06-16 10:30:53,670][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.041516728699207306, 'train/grad_norm_pre_clip': 16.06212615966797, 'timing/data': 0.020509020425379276, 'timing/model': 1.5539866629987955, 'learning_rate/qwen_vl_interface': 1.021874613552316e-06, 'learning_rate/action_model': 1.0218746135523157e-05, 'epoch': 1.72}) +[2026-06-16 10:31:00,072][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.00035884376848116517, 'train/grad_norm_pre_clip': 0.08623024821281433, 'timing/data': 0.025562464259564877, 'timing/model': 1.548967911861837, 'learning_rate/qwen_vl_interface': 1.0204876736207384e-06, 'learning_rate/action_model': 1.0204876736207384e-05, 'epoch': 1.72}) +[2026-06-16 10:31:06,469][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.0006860337452962995, 'train/grad_norm_pre_clip': 0.24240708351135254, 'timing/data': 0.021712317131459713, 'timing/model': 1.5817602891474962, 'learning_rate/qwen_vl_interface': 1.0191024722819163e-06, 'learning_rate/action_model': 1.0191024722819162e-05, 'epoch': 1.72}) +[2026-06-16 10:31:12,844][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.002323324792087078, 'train/grad_norm_pre_clip': 2.740725040435791, 'timing/data': 0.0209697219543159, 'timing/model': 1.8225066727027297, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 1.72}) +[2026-06-16 10:31:19,147][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.0006826166063547134, 'train/grad_norm_pre_clip': 0.2728237807750702, 'timing/data': 0.022448941133916378, 'timing/model': 1.8246009778231382, 'learning_rate/qwen_vl_interface': 1.016337287659437e-06, 'learning_rate/action_model': 1.0163372876594368e-05, 'epoch': 1.73}) +[2026-06-16 10:31:25,244][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.0003017291019205004, 'train/grad_norm_pre_clip': 0.06889079511165619, 'timing/data': 0.021210042759776115, 'timing/model': 1.616578347980976, 'learning_rate/qwen_vl_interface': 1.0149573055124427e-06, 'learning_rate/action_model': 1.0149573055124426e-05, 'epoch': 1.73}) +[2026-06-16 10:31:31,576][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.0031615919433534145, 'train/grad_norm_pre_clip': 1.9718513488769531, 'timing/data': 0.021308393217623234, 'timing/model': 1.5543743842281401, 'learning_rate/qwen_vl_interface': 1.013579064231529e-06, 'learning_rate/action_model': 1.013579064231529e-05, 'epoch': 1.73}) +[2026-06-16 10:31:38,071][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.0003846710023935884, 'train/grad_norm_pre_clip': 0.17968115210533142, 'timing/data': 0.021765645127743483, 'timing/model': 1.5589051800779998, 'learning_rate/qwen_vl_interface': 1.012202564383239e-06, 'learning_rate/action_model': 1.0122025643832388e-05, 'epoch': 1.73}) +[2026-06-16 10:31:44,477][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.0003745172289200127, 'train/grad_norm_pre_clip': 0.1754225194454193, 'timing/data': 0.021376872900873423, 'timing/model': 1.576047679875046, 'learning_rate/qwen_vl_interface': 1.0108278065333978e-06, 'learning_rate/action_model': 1.0108278065333978e-05, 'epoch': 1.73}) +[2026-06-16 10:31:50,859][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.006097532343119383, 'train/grad_norm_pre_clip': 6.727921485900879, 'timing/data': 0.02187022613361478, 'timing/model': 1.8129396671429276, 'learning_rate/qwen_vl_interface': 1.0094547912471183e-06, 'learning_rate/action_model': 1.0094547912471184e-05, 'epoch': 1.73}) +[2026-06-16 10:31:56,973][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.23663540184497833, 'train/grad_norm_pre_clip': 12.001561164855957, 'timing/data': 0.021536303218454123, 'timing/model': 1.5482537178322673, 'learning_rate/qwen_vl_interface': 1.0080835190887949e-06, 'learning_rate/action_model': 1.0080835190887947e-05, 'epoch': 1.73}) +[2026-06-16 10:32:03,373][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.04410811886191368, 'train/grad_norm_pre_clip': 34.73103713989258, 'timing/data': 0.021640073973685503, 'timing/model': 1.5529562160372734, 'learning_rate/qwen_vl_interface': 1.0067139906221052e-06, 'learning_rate/action_model': 1.0067139906221052e-05, 'epoch': 1.73}) +[2026-06-16 10:32:09,774][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 0.00013807322829961777, 'train/grad_norm_pre_clip': 0.01957072503864765, 'timing/data': 0.020927688106894493, 'timing/model': 1.5550982081331313, 'learning_rate/qwen_vl_interface': 1.0053462064100112e-06, 'learning_rate/action_model': 1.0053462064100112e-05, 'epoch': 1.73}) +[2026-06-16 10:32:16,177][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.002112767891958356, 'train/grad_norm_pre_clip': 0.6291599273681641, 'timing/data': 0.020657426677644253, 'timing/model': 1.5936214397661388, 'learning_rate/qwen_vl_interface': 1.003980167014757e-06, 'learning_rate/action_model': 1.003980167014757e-05, 'epoch': 1.73}) +[2026-06-16 10:32:22,471][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.15901070833206177, 'train/grad_norm_pre_clip': 14.683271408081055, 'timing/data': 0.021110610105097294, 'timing/model': 1.5304856300354004, 'learning_rate/qwen_vl_interface': 1.0026158729978702e-06, 'learning_rate/action_model': 1.0026158729978702e-05, 'epoch': 1.73}) +[2026-06-16 10:32:28,884][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.16542552411556244, 'train/grad_norm_pre_clip': 46.34003448486328, 'timing/data': 0.02205897495150566, 'timing/model': 1.7619043532758951, 'learning_rate/qwen_vl_interface': 1.0012533249201604e-06, 'learning_rate/action_model': 1.0012533249201602e-05, 'epoch': 1.73}) +[2026-06-16 10:32:34,970][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.010659228079020977, 'train/grad_norm_pre_clip': 7.766676425933838, 'timing/data': 0.02123402990400791, 'timing/model': 1.5472838077694178, 'learning_rate/qwen_vl_interface': 9.99892523341719e-07, 'learning_rate/action_model': 9.99892523341719e-06, 'epoch': 1.73}) +[2026-06-16 10:32:41,471][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.00017467164434492588, 'train/grad_norm_pre_clip': 0.032609470188617706, 'timing/data': 0.02181411301717162, 'timing/model': 1.5560454931110144, 'learning_rate/qwen_vl_interface': 9.98533468821922e-07, 'learning_rate/action_model': 9.98533468821922e-06, 'epoch': 1.73}) +[2026-06-16 10:32:47,873][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.0012489293003454804, 'train/grad_norm_pre_clip': 0.37008360028266907, 'timing/data': 0.02108172792941332, 'timing/model': 1.5520335459150374, 'learning_rate/qwen_vl_interface': 9.971761619194231e-07, 'learning_rate/action_model': 9.97176161919423e-06, 'epoch': 1.73}) +[2026-06-16 10:32:54,276][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 5.662896728608757e-05, 'train/grad_norm_pre_clip': 0.004888560622930527, 'timing/data': 0.021634611766785383, 'timing/model': 1.573191842995584, 'learning_rate/qwen_vl_interface': 9.958206031921606e-07, 'learning_rate/action_model': 9.958206031921604e-06, 'epoch': 1.73}) +[2026-06-16 10:33:00,672][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.0921773761510849, 'train/grad_norm_pre_clip': 37.1411018371582, 'timing/data': 0.021452730055898428, 'timing/model': 1.5507018859498203, 'learning_rate/qwen_vl_interface': 9.944667931973537e-07, 'learning_rate/action_model': 9.944667931973536e-06, 'epoch': 1.73}) +[2026-06-16 10:33:06,994][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.1698182076215744, 'train/grad_norm_pre_clip': 19.640235900878906, 'timing/data': 0.022340605035424232, 'timing/model': 1.761092253960669, 'learning_rate/qwen_vl_interface': 9.93114732491503e-07, 'learning_rate/action_model': 9.93114732491503e-06, 'epoch': 1.73}) +[2026-06-16 10:33:13,083][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.003218571888282895, 'train/grad_norm_pre_clip': 1.7463501691818237, 'timing/data': 0.021126926876604557, 'timing/model': 1.5642812340520322, 'learning_rate/qwen_vl_interface': 9.91764421630389e-07, 'learning_rate/action_model': 9.91764421630389e-06, 'epoch': 1.73}) +[2026-06-16 10:33:19,471][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.0027442239224910736, 'train/grad_norm_pre_clip': 1.1082922220230103, 'timing/data': 0.021277808118611574, 'timing/model': 1.550579464994371, 'learning_rate/qwen_vl_interface': 9.904158611690732e-07, 'learning_rate/action_model': 9.90415861169073e-06, 'epoch': 1.73}) +[2026-06-16 10:33:25,874][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.3965473175048828, 'train/grad_norm_pre_clip': 11.247474670410156, 'timing/data': 0.021683810278773308, 'timing/model': 1.533775218296796, 'learning_rate/qwen_vl_interface': 9.890690516618987e-07, 'learning_rate/action_model': 9.890690516618985e-06, 'epoch': 1.73}) +[2026-06-16 10:33:32,265][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.010951808653771877, 'train/grad_norm_pre_clip': 5.845937728881836, 'timing/data': 0.019857358187437057, 'timing/model': 1.5766626838594675, 'learning_rate/qwen_vl_interface': 9.877239936624875e-07, 'learning_rate/action_model': 9.877239936624875e-06, 'epoch': 1.73}) +[2026-06-16 10:33:38,753][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.0006019941065460443, 'train/grad_norm_pre_clip': 0.1315719336271286, 'timing/data': 0.021675749216228724, 'timing/model': 1.8359092250466347, 'learning_rate/qwen_vl_interface': 9.863806877237421e-07, 'learning_rate/action_model': 9.863806877237422e-06, 'epoch': 1.73}) +[2026-06-16 10:33:44,978][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.0885547623038292, 'train/grad_norm_pre_clip': 21.856334686279297, 'timing/data': 0.02119559608399868, 'timing/model': 1.565779035910964, 'learning_rate/qwen_vl_interface': 9.850391343978454e-07, 'learning_rate/action_model': 9.850391343978454e-06, 'epoch': 1.73}) +[2026-06-16 10:33:51,445][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.0007225346053019166, 'train/grad_norm_pre_clip': 0.1996854692697525, 'timing/data': 0.020732413046061993, 'timing/model': 1.5934992441907525, 'learning_rate/qwen_vl_interface': 9.836993342362596e-07, 'learning_rate/action_model': 9.836993342362596e-06, 'epoch': 1.73}) +[2026-06-16 10:33:57,724][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.08784055709838867, 'train/grad_norm_pre_clip': 29.408443450927734, 'timing/data': 0.019124672282487154, 'timing/model': 1.5937215080484748, 'learning_rate/qwen_vl_interface': 9.823612877897246e-07, 'learning_rate/action_model': 9.823612877897245e-06, 'epoch': 1.74}) +[2026-06-16 10:34:03,977][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.0015085602644830942, 'train/grad_norm_pre_clip': 1.300713300704956, 'timing/data': 0.01964867627248168, 'timing/model': 1.557735628914088, 'learning_rate/qwen_vl_interface': 9.81024995608261e-07, 'learning_rate/action_model': 9.810249956082609e-06, 'epoch': 1.74}) +[2026-06-16 10:34:10,269][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.52890944480896, 'train/grad_norm_pre_clip': 29.652812957763672, 'timing/data': 0.021528717130422592, 'timing/model': 1.5270990058779716, 'learning_rate/qwen_vl_interface': 9.796904582411685e-07, 'learning_rate/action_model': 9.796904582411684e-06, 'epoch': 1.74}) +[2026-06-16 10:34:16,632][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.001671084319241345, 'train/grad_norm_pre_clip': 1.9151227474212646, 'timing/data': 0.021543496754020452, 'timing/model': 1.8175348690710962, 'learning_rate/qwen_vl_interface': 9.78357676237025e-07, 'learning_rate/action_model': 9.78357676237025e-06, 'epoch': 1.74}) +[2026-06-16 10:34:22,681][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 0.166961669921875, 'train/grad_norm_pre_clip': 33.73619079589844, 'timing/data': 0.021763738244771957, 'timing/model': 1.5621821051463485, 'learning_rate/qwen_vl_interface': 9.77026650143687e-07, 'learning_rate/action_model': 9.77026650143687e-06, 'epoch': 1.74}) +[2026-06-16 10:34:29,070][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.0032812519930303097, 'train/grad_norm_pre_clip': 2.7973601818084717, 'timing/data': 0.019997597206383944, 'timing/model': 1.552599126007408, 'learning_rate/qwen_vl_interface': 9.756973805082882e-07, 'learning_rate/action_model': 9.756973805082883e-06, 'epoch': 1.74}) +[2026-06-16 10:34:35,469][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.0015390367479994893, 'train/grad_norm_pre_clip': 0.8591204285621643, 'timing/data': 0.020886393263936043, 'timing/model': 1.5489268749952316, 'learning_rate/qwen_vl_interface': 9.743698678772434e-07, 'learning_rate/action_model': 9.743698678772433e-06, 'epoch': 1.74}) +[2026-06-16 10:34:41,872][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.0204472579061985, 'train/grad_norm_pre_clip': 14.628782272338867, 'timing/data': 0.021497475914657116, 'timing/model': 1.557018040213734, 'learning_rate/qwen_vl_interface': 9.730441127962395e-07, 'learning_rate/action_model': 9.730441127962395e-06, 'epoch': 1.74}) +[2026-06-16 10:34:48,261][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.00267050601541996, 'train/grad_norm_pre_clip': 2.7043449878692627, 'timing/data': 0.021194854285568, 'timing/model': 1.6019954169169068, 'learning_rate/qwen_vl_interface': 9.717201158102465e-07, 'learning_rate/action_model': 9.717201158102466e-06, 'epoch': 1.74}) +[2026-06-16 10:45:41,640][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.000295613135676831, 'train/grad_norm_pre_clip': 0.1159340888261795, 'eval/loss': 0.03645709529519081, 'eval/flap_precision': 0.9637096774193549, 'eval/flap_recall': 0.9781718963165075, 'eval/flap_f1': 0.970886932972241, 'eval/flap_rate_model': 0.20666666666666667, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9708869329722409, 'timing/data': 0.020452748984098434, 'timing/model': 1.8599799163639545, 'learning_rate/qwen_vl_interface': 9.703978774635094e-07, 'learning_rate/action_model': 9.703978774635095e-06, 'epoch': 1.74}) +[2026-06-16 10:45:41,641][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4300_state +[2026-06-16 10:45:41,648][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 10:45:50,248][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4300_state/model.safetensors +[2026-06-16 10:46:10,523][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4300_state/optimizer.bin +[2026-06-16 10:46:10,524][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4300_state/sampler.bin +[2026-06-16 10:46:10,525][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4300_state/sampler_1.bin +[2026-06-16 10:46:10,526][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4300_state/random_states_0.pkl +[2026-06-16 10:46:10,715][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4300 +[2026-06-16 10:46:10,716][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 10:46:10,732][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 10:46:17,009][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.10461286455392838, 'train/grad_norm_pre_clip': 37.13973617553711, 'timing/data': 0.022042005322873592, 'timing/model': 1.804153123870492, 'learning_rate/qwen_vl_interface': 9.6907739829955e-07, 'learning_rate/action_model': 9.6907739829955e-06, 'epoch': 1.74}) +[2026-06-16 10:46:23,211][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.004715492017567158, 'train/grad_norm_pre_clip': 5.909336566925049, 'timing/data': 0.02031169505789876, 'timing/model': 1.7129465723410249, 'learning_rate/qwen_vl_interface': 9.677586788611681e-07, 'learning_rate/action_model': 9.677586788611681e-06, 'epoch': 1.74}) +[2026-06-16 10:46:29,230][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.0007084727403707802, 'train/grad_norm_pre_clip': 0.23494568467140198, 'timing/data': 0.021178700029850006, 'timing/model': 1.5562536283396184, 'learning_rate/qwen_vl_interface': 9.664417196904382e-07, 'learning_rate/action_model': 9.664417196904382e-06, 'epoch': 1.74}) +[2026-06-16 10:46:35,444][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 9.105602657655254e-05, 'train/grad_norm_pre_clip': 0.009959632530808449, 'timing/data': 0.021572961937636137, 'timing/model': 1.5562515738420188, 'learning_rate/qwen_vl_interface': 9.651265213287134e-07, 'learning_rate/action_model': 9.651265213287134e-06, 'epoch': 1.74}) +[2026-06-16 10:46:41,638][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.12818442285060883, 'train/grad_norm_pre_clip': 19.902467727661133, 'timing/data': 0.01934746978804469, 'timing/model': 1.5507287289947271, 'learning_rate/qwen_vl_interface': 9.638130843166218e-07, 'learning_rate/action_model': 9.638130843166219e-06, 'epoch': 1.74}) +[2026-06-16 10:46:47,849][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.0004420358163770288, 'train/grad_norm_pre_clip': 0.15065692365169525, 'timing/data': 0.022734678350389004, 'timing/model': 1.503060624934733, 'learning_rate/qwen_vl_interface': 9.62501409194068e-07, 'learning_rate/action_model': 9.625014091940678e-06, 'epoch': 1.74}) +[2026-06-16 10:46:54,117][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.2845260500907898, 'train/grad_norm_pre_clip': 7.074019908905029, 'timing/data': 0.022504157852381468, 'timing/model': 1.8225179398432374, 'learning_rate/qwen_vl_interface': 9.611914965002317e-07, 'learning_rate/action_model': 9.611914965002316e-06, 'epoch': 1.74}) +[2026-06-16 10:47:00,155][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.01432112604379654, 'train/grad_norm_pre_clip': 8.185117721557617, 'timing/data': 0.020976759027689695, 'timing/model': 1.5597362578846514, 'learning_rate/qwen_vl_interface': 9.598833467735686e-07, 'learning_rate/action_model': 9.598833467735685e-06, 'epoch': 1.74}) +[2026-06-16 10:47:06,451][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.00011274301505181938, 'train/grad_norm_pre_clip': 0.01726531609892845, 'timing/data': 0.021068648900836706, 'timing/model': 1.5525248926132917, 'learning_rate/qwen_vl_interface': 9.585769605518097e-07, 'learning_rate/action_model': 9.585769605518097e-06, 'epoch': 1.74}) +[2026-06-16 10:47:12,759][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.004970722831785679, 'train/grad_norm_pre_clip': 4.052702903747559, 'timing/data': 0.024178697261959314, 'timing/model': 1.5549925127997994, 'learning_rate/qwen_vl_interface': 9.572723383719611e-07, 'learning_rate/action_model': 9.572723383719612e-06, 'epoch': 1.74}) +[2026-06-16 10:47:19,055][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.1298040747642517, 'train/grad_norm_pre_clip': 22.760196685791016, 'timing/data': 0.02191066276282072, 'timing/model': 1.549275548197329, 'learning_rate/qwen_vl_interface': 9.559694807703041e-07, 'learning_rate/action_model': 9.55969480770304e-06, 'epoch': 1.74}) +[2026-06-16 10:47:25,450][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.34863677620887756, 'train/grad_norm_pre_clip': 22.42314910888672, 'timing/data': 0.020585847087204456, 'timing/model': 1.581321633886546, 'learning_rate/qwen_vl_interface': 9.546683882823945e-07, 'learning_rate/action_model': 9.546683882823945e-06, 'epoch': 1.74}) +[2026-06-16 10:47:31,821][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.07924997806549072, 'train/grad_norm_pre_clip': 36.39665603637695, 'timing/data': 0.020774567034095526, 'timing/model': 1.8128575989976525, 'learning_rate/qwen_vl_interface': 9.533690614430608e-07, 'learning_rate/action_model': 9.533690614430608e-06, 'epoch': 1.74}) +[2026-06-16 10:47:37,867][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.005337823182344437, 'train/grad_norm_pre_clip': 4.760510444641113, 'timing/data': 0.024393267929553986, 'timing/model': 1.5514604891650379, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 1.74}) +[2026-06-16 10:47:44,260][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.02817714586853981, 'train/grad_norm_pre_clip': 32.998355865478516, 'timing/data': 0.021199729293584824, 'timing/model': 1.5639834757894278, 'learning_rate/qwen_vl_interface': 9.50775706845814e-07, 'learning_rate/action_model': 9.50775706845814e-06, 'epoch': 1.74}) +[2026-06-16 10:47:50,562][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.0009539779275655746, 'train/grad_norm_pre_clip': 0.31079694628715515, 'timing/data': 0.021496101282536983, 'timing/model': 1.5526932077482343, 'learning_rate/qwen_vl_interface': 9.49481680153931e-07, 'learning_rate/action_model': 9.494816801539309e-06, 'epoch': 1.75}) +[2026-06-16 10:47:56,967][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.00044015864841639996, 'train/grad_norm_pre_clip': 0.07292183488607407, 'timing/data': 0.021246950142085552, 'timing/model': 1.550947463605553, 'learning_rate/qwen_vl_interface': 9.481894212426848e-07, 'learning_rate/action_model': 9.481894212426847e-06, 'epoch': 1.75}) +[2026-06-16 10:48:03,364][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.002139048185199499, 'train/grad_norm_pre_clip': 1.8730555772781372, 'timing/data': 0.021294590085744858, 'timing/model': 1.5918199908919632, 'learning_rate/qwen_vl_interface': 9.468989306432738e-07, 'learning_rate/action_model': 9.468989306432738e-06, 'epoch': 1.75}) +[2026-06-16 10:48:09,727][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.10914717614650726, 'train/grad_norm_pre_clip': 20.539710998535156, 'timing/data': 0.02856660122051835, 'timing/model': 1.80381592316553, 'learning_rate/qwen_vl_interface': 9.456102088861702e-07, 'learning_rate/action_model': 9.456102088861701e-06, 'epoch': 1.75}) +[2026-06-16 10:48:15,966][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.0030650144908577204, 'train/grad_norm_pre_clip': 2.5296120643615723, 'timing/data': 0.0215727211907506, 'timing/model': 1.7575421477667987, 'learning_rate/qwen_vl_interface': 9.443232565011188e-07, 'learning_rate/action_model': 9.443232565011187e-06, 'epoch': 1.75}) +[2026-06-16 10:48:22,070][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.21604366600513458, 'train/grad_norm_pre_clip': 17.27103042602539, 'timing/data': 0.020814067218452692, 'timing/model': 1.5555105973035097, 'learning_rate/qwen_vl_interface': 9.430380740171369e-07, 'learning_rate/action_model': 9.430380740171368e-06, 'epoch': 1.75}) +[2026-06-16 10:48:28,366][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.0002941102138720453, 'train/grad_norm_pre_clip': 0.09292639046907425, 'timing/data': 0.021043768618255854, 'timing/model': 1.5533672492019832, 'learning_rate/qwen_vl_interface': 9.417546619625143e-07, 'learning_rate/action_model': 9.417546619625143e-06, 'epoch': 1.75}) +[2026-06-16 10:48:34,763][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.12569476664066315, 'train/grad_norm_pre_clip': 16.44927978515625, 'timing/data': 0.02039680490270257, 'timing/model': 1.5447472101077437, 'learning_rate/qwen_vl_interface': 9.404730208648144e-07, 'learning_rate/action_model': 9.404730208648144e-06, 'epoch': 1.75}) +[2026-06-16 10:48:41,227][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.010625870898365974, 'train/grad_norm_pre_clip': 7.019269943237305, 'timing/data': 0.021486470010131598, 'timing/model': 1.5773944202810526, 'learning_rate/qwen_vl_interface': 9.3919315125087e-07, 'learning_rate/action_model': 9.3919315125087e-06, 'epoch': 1.75}) +[2026-06-16 10:48:47,523][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.00562714459374547, 'train/grad_norm_pre_clip': 5.847460746765137, 'timing/data': 0.02039344422519207, 'timing/model': 1.8127632099203765, 'learning_rate/qwen_vl_interface': 9.379150536467879e-07, 'learning_rate/action_model': 9.379150536467878e-06, 'epoch': 1.75}) +[2026-06-16 10:48:53,836][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.00026763114146888256, 'train/grad_norm_pre_clip': 0.07168169319629669, 'timing/data': 0.020481355022639036, 'timing/model': 1.8231806848198175, 'learning_rate/qwen_vl_interface': 9.366387285779453e-07, 'learning_rate/action_model': 9.366387285779453e-06, 'epoch': 1.75}) +[2026-06-16 10:48:59,863][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.023025639355182648, 'train/grad_norm_pre_clip': 16.82060432434082, 'timing/data': 0.021705851890146732, 'timing/model': 1.5566178541630507, 'learning_rate/qwen_vl_interface': 9.353641765689915e-07, 'learning_rate/action_model': 9.353641765689915e-06, 'epoch': 1.75}) +[2026-06-16 10:49:06,159][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.005190238356590271, 'train/grad_norm_pre_clip': 6.3098883628845215, 'timing/data': 0.02027311362326145, 'timing/model': 1.55273204902187, 'learning_rate/qwen_vl_interface': 9.340913981438462e-07, 'learning_rate/action_model': 9.340913981438463e-06, 'epoch': 1.75}) +[2026-06-16 10:49:12,566][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.01121517363935709, 'train/grad_norm_pre_clip': 7.810233116149902, 'timing/data': 0.02066820627078414, 'timing/model': 1.5480909091420472, 'learning_rate/qwen_vl_interface': 9.328203938257019e-07, 'learning_rate/action_model': 9.328203938257018e-06, 'epoch': 1.75}) +[2026-06-16 10:49:18,969][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.22333985567092896, 'train/grad_norm_pre_clip': 31.35708236694336, 'timing/data': 0.021177138667553663, 'timing/model': 1.5903885941952467, 'learning_rate/qwen_vl_interface': 9.315511641370191e-07, 'learning_rate/action_model': 9.315511641370191e-06, 'epoch': 1.75}) +[2026-06-16 10:49:25,330][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.32500967383384705, 'train/grad_norm_pre_clip': 26.486751556396484, 'timing/data': 0.022620366886258125, 'timing/model': 1.818863001652062, 'learning_rate/qwen_vl_interface': 9.302837095995306e-07, 'learning_rate/action_model': 9.302837095995306e-06, 'epoch': 1.75}) +[2026-06-16 10:49:31,364][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.001045624609105289, 'train/grad_norm_pre_clip': 0.7745676636695862, 'timing/data': 0.019368098117411137, 'timing/model': 1.5571102611720562, 'learning_rate/qwen_vl_interface': 9.290180307342392e-07, 'learning_rate/action_model': 9.290180307342391e-06, 'epoch': 1.75}) +[2026-06-16 10:49:37,662][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 0.010741320438683033, 'train/grad_norm_pre_clip': 7.556731700897217, 'timing/data': 0.020773676224052906, 'timing/model': 1.5525752431713045, 'learning_rate/qwen_vl_interface': 9.277541280614183e-07, 'learning_rate/action_model': 9.277541280614183e-06, 'epoch': 1.75}) +[2026-06-16 10:49:44,131][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.005561792757362127, 'train/grad_norm_pre_clip': 4.894924640655518, 'timing/data': 0.01939973933622241, 'timing/model': 1.6260273307561874, 'learning_rate/qwen_vl_interface': 9.264920021006093e-07, 'learning_rate/action_model': 9.264920021006093e-06, 'epoch': 1.75}) +[2026-06-16 10:49:50,466][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.20158769190311432, 'train/grad_norm_pre_clip': 35.803585052490234, 'timing/data': 0.01942671788856387, 'timing/model': 1.522885381244123, 'learning_rate/qwen_vl_interface': 9.252316533706246e-07, 'learning_rate/action_model': 9.252316533706246e-06, 'epoch': 1.75}) +[2026-06-16 10:49:56,765][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.0008560651913285255, 'train/grad_norm_pre_clip': 0.18983079493045807, 'timing/data': 0.020028683356940746, 'timing/model': 1.5695188390091062, 'learning_rate/qwen_vl_interface': 9.239730823895469e-07, 'learning_rate/action_model': 9.23973082389547e-06, 'epoch': 1.75}) +[2026-06-16 10:50:03,281][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.07470891624689102, 'train/grad_norm_pre_clip': 40.0428466796875, 'timing/data': 0.02258595684543252, 'timing/model': 1.7703285696916282, 'learning_rate/qwen_vl_interface': 9.227162896747264e-07, 'learning_rate/action_model': 9.227162896747264e-06, 'epoch': 1.75}) +[2026-06-16 10:50:09,363][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.13546942174434662, 'train/grad_norm_pre_clip': 28.590682983398438, 'timing/data': 0.020844706799834967, 'timing/model': 1.5452760066837072, 'learning_rate/qwen_vl_interface': 9.214612757427833e-07, 'learning_rate/action_model': 9.214612757427833e-06, 'epoch': 1.75}) +[2026-06-16 10:50:15,664][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.00019854433776345104, 'train/grad_norm_pre_clip': 0.03806333243846893, 'timing/data': 0.021829372737556696, 'timing/model': 1.5560411247424781, 'learning_rate/qwen_vl_interface': 9.202080411096071e-07, 'learning_rate/action_model': 9.20208041109607e-06, 'epoch': 1.75}) +[2026-06-16 10:50:22,073][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.0003548525273799896, 'train/grad_norm_pre_clip': 0.09989246726036072, 'timing/data': 0.021946042776107788, 'timing/model': 1.5509356539696455, 'learning_rate/qwen_vl_interface': 9.189565862903543e-07, 'learning_rate/action_model': 9.189565862903543e-06, 'epoch': 1.76}) +[2026-06-16 10:50:28,502][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.00042644923087209463, 'train/grad_norm_pre_clip': 0.16842903196811676, 'timing/data': 0.019869551062583923, 'timing/model': 1.6137918829917908, 'learning_rate/qwen_vl_interface': 9.17706911799451e-07, 'learning_rate/action_model': 9.17706911799451e-06, 'epoch': 1.76}) +[2026-06-16 10:50:34,892][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.03001333959400654, 'train/grad_norm_pre_clip': 23.7088623046875, 'timing/data': 0.020573065150529146, 'timing/model': 1.5736280563287437, 'learning_rate/qwen_vl_interface': 9.164590181505912e-07, 'learning_rate/action_model': 9.164590181505911e-06, 'epoch': 1.76}) +[2026-06-16 10:50:41,220][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.0007291696383617818, 'train/grad_norm_pre_clip': 0.17637957632541656, 'timing/data': 0.020350662991404533, 'timing/model': 1.8031007759273052, 'learning_rate/qwen_vl_interface': 9.15212905856738e-07, 'learning_rate/action_model': 9.15212905856738e-06, 'epoch': 1.76}) +[2026-06-16 10:50:47,269][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.13566498458385468, 'train/grad_norm_pre_clip': 27.550579071044922, 'timing/data': 0.02203332306817174, 'timing/model': 1.546654960140586, 'learning_rate/qwen_vl_interface': 9.139685754301189e-07, 'learning_rate/action_model': 9.139685754301189e-06, 'epoch': 1.76}) +[2026-06-16 10:50:53,780][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.0009357495000585914, 'train/grad_norm_pre_clip': 0.35521426796913147, 'timing/data': 0.021838832180947065, 'timing/model': 1.602198837324977, 'learning_rate/qwen_vl_interface': 9.127260273822324e-07, 'learning_rate/action_model': 9.127260273822325e-06, 'epoch': 1.76}) +[2026-06-16 10:51:00,185][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 0.0009698644280433655, 'train/grad_norm_pre_clip': 0.24000051617622375, 'timing/data': 0.021515939850360155, 'timing/model': 1.5590927442535758, 'learning_rate/qwen_vl_interface': 9.114852622238435e-07, 'learning_rate/action_model': 9.114852622238435e-06, 'epoch': 1.76}) +[2026-06-16 10:51:06,582][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.526300311088562, 'train/grad_norm_pre_clip': 31.29786491394043, 'timing/data': 0.021674200892448425, 'timing/model': 1.587649680674076, 'learning_rate/qwen_vl_interface': 9.102462804649827e-07, 'learning_rate/action_model': 9.102462804649827e-06, 'epoch': 1.76}) +[2026-06-16 10:51:12,942][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.0003174009034410119, 'train/grad_norm_pre_clip': 0.04808393493294716, 'timing/data': 0.021557420026510954, 'timing/model': 1.821244001854211, 'learning_rate/qwen_vl_interface': 9.090090826149489e-07, 'learning_rate/action_model': 9.09009082614949e-06, 'epoch': 1.76}) +[2026-06-16 10:51:18,974][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.02610885351896286, 'train/grad_norm_pre_clip': 22.57388687133789, 'timing/data': 0.022068593185395002, 'timing/model': 1.5472188009880483, 'learning_rate/qwen_vl_interface': 9.077736691823081e-07, 'learning_rate/action_model': 9.07773669182308e-06, 'epoch': 1.76}) +[2026-06-16 10:51:25,370][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.08472321927547455, 'train/grad_norm_pre_clip': 30.48766326904297, 'timing/data': 0.021801641676574945, 'timing/model': 1.5465237339958549, 'learning_rate/qwen_vl_interface': 9.065400406748912e-07, 'learning_rate/action_model': 9.06540040674891e-06, 'epoch': 1.76}) +[2026-06-16 10:51:31,674][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.0008657615398988128, 'train/grad_norm_pre_clip': 0.30138564109802246, 'timing/data': 0.022422165144234896, 'timing/model': 1.5533285504207015, 'learning_rate/qwen_vl_interface': 9.053081975997969e-07, 'learning_rate/action_model': 9.053081975997968e-06, 'epoch': 1.76}) +[2026-06-16 10:51:38,079][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.00011405764962546527, 'train/grad_norm_pre_clip': 0.013723746873438358, 'timing/data': 0.022770986892282963, 'timing/model': 1.5646050898358226, 'learning_rate/qwen_vl_interface': 9.040781404633894e-07, 'learning_rate/action_model': 9.040781404633895e-06, 'epoch': 1.76}) +[2026-06-16 10:51:44,454][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 0.05010784789919853, 'train/grad_norm_pre_clip': 24.157060623168945, 'timing/data': 0.022238844074308872, 'timing/model': 1.6075643980875611, 'learning_rate/qwen_vl_interface': 9.028498697712982e-07, 'learning_rate/action_model': 9.028498697712981e-06, 'epoch': 1.76}) +[2026-06-16 10:51:50,850][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.13018561899662018, 'train/grad_norm_pre_clip': 17.01289176940918, 'timing/data': 0.021452218759804964, 'timing/model': 1.8285277062095702, 'learning_rate/qwen_vl_interface': 9.016233860284195e-07, 'learning_rate/action_model': 9.016233860284195e-06, 'epoch': 1.76}) +[2026-06-16 10:51:56,970][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.08438170701265335, 'train/grad_norm_pre_clip': 29.12835121154785, 'timing/data': 0.021217719186097383, 'timing/model': 1.5498869908042252, 'learning_rate/qwen_vl_interface': 9.003986897389142e-07, 'learning_rate/action_model': 9.003986897389141e-06, 'epoch': 1.76}) +[2026-06-16 10:52:03,374][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.37926679849624634, 'train/grad_norm_pre_clip': 28.309036254882812, 'timing/data': 0.021778022404760122, 'timing/model': 1.5478177280165255, 'learning_rate/qwen_vl_interface': 8.991757814062094e-07, 'learning_rate/action_model': 8.991757814062093e-06, 'epoch': 1.76}) +[2026-06-16 10:52:09,778][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.01740010641515255, 'train/grad_norm_pre_clip': 22.412782669067383, 'timing/data': 0.021305848378688097, 'timing/model': 1.560038574039936, 'learning_rate/qwen_vl_interface': 8.979546615329954e-07, 'learning_rate/action_model': 8.979546615329953e-06, 'epoch': 1.76}) +[2026-06-16 10:52:16,176][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.2911735773086548, 'train/grad_norm_pre_clip': 18.110475540161133, 'timing/data': 0.02280151704326272, 'timing/model': 1.54023521207273, 'learning_rate/qwen_vl_interface': 8.967353306212286e-07, 'learning_rate/action_model': 8.967353306212286e-06, 'epoch': 1.76}) +[2026-06-16 10:52:22,568][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.244315505027771, 'train/grad_norm_pre_clip': 23.616485595703125, 'timing/data': 0.02237832499668002, 'timing/model': 1.573756926227361, 'learning_rate/qwen_vl_interface': 8.955177891721303e-07, 'learning_rate/action_model': 8.955177891721303e-06, 'epoch': 1.76}) +[2026-06-16 10:52:28,965][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.15049627423286438, 'train/grad_norm_pre_clip': 14.068347930908203, 'timing/data': 0.022210313007235527, 'timing/model': 1.8218040140345693, 'learning_rate/qwen_vl_interface': 8.943020376861858e-07, 'learning_rate/action_model': 8.943020376861857e-06, 'epoch': 1.76}) +[2026-06-16 10:52:35,072][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.007277416065335274, 'train/grad_norm_pre_clip': 5.410900592803955, 'timing/data': 0.021222347859293222, 'timing/model': 1.553459418937564, 'learning_rate/qwen_vl_interface': 8.930880766631444e-07, 'learning_rate/action_model': 8.930880766631445e-06, 'epoch': 1.76}) +[2026-06-16 10:52:41,587][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.008046356029808521, 'train/grad_norm_pre_clip': 10.999249458312988, 'timing/data': 0.021886262111365795, 'timing/model': 1.6087379641830921, 'learning_rate/qwen_vl_interface': 8.918759066020201e-07, 'learning_rate/action_model': 8.9187590660202e-06, 'epoch': 1.76}) +[2026-06-16 10:52:48,167][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 0.12696917355060577, 'train/grad_norm_pre_clip': 13.991440773010254, 'timing/data': 0.022488355170935392, 'timing/model': 1.5362770091742277, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 1.76}) +[2026-06-16 10:52:54,458][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.00049588578986004, 'train/grad_norm_pre_clip': 0.08792921900749207, 'timing/data': 0.01983034983277321, 'timing/model': 1.5459999511949718, 'learning_rate/qwen_vl_interface': 8.894569413578938e-07, 'learning_rate/action_model': 8.894569413578937e-06, 'epoch': 1.76}) +[2026-06-16 10:53:00,955][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.1206502765417099, 'train/grad_norm_pre_clip': 15.281866073608398, 'timing/data': 0.020604954101145267, 'timing/model': 1.5673138978891075, 'learning_rate/qwen_vl_interface': 8.882501471692379e-07, 'learning_rate/action_model': 8.882501471692378e-06, 'epoch': 1.77}) +[2026-06-16 10:53:07,327][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.007789099123328924, 'train/grad_norm_pre_clip': 8.03357219696045, 'timing/data': 0.02470455691218376, 'timing/model': 1.8084233147092164, 'learning_rate/qwen_vl_interface': 8.870451459311894e-07, 'learning_rate/action_model': 8.870451459311893e-06, 'epoch': 1.77}) +[2026-06-16 10:53:13,371][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.0009045089827850461, 'train/grad_norm_pre_clip': 0.2712738811969757, 'timing/data': 0.022227694280445576, 'timing/model': 1.5520548289641738, 'learning_rate/qwen_vl_interface': 8.85841938139077e-07, 'learning_rate/action_model': 8.85841938139077e-06, 'epoch': 1.77}) +[2026-06-16 10:53:19,778][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.0005873790360055864, 'train/grad_norm_pre_clip': 0.18977530300617218, 'timing/data': 0.02216443419456482, 'timing/model': 1.556187140289694, 'learning_rate/qwen_vl_interface': 8.846405242874955e-07, 'learning_rate/action_model': 8.846405242874954e-06, 'epoch': 1.77}) +[2026-06-16 10:53:26,072][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.008926990441977978, 'train/grad_norm_pre_clip': 7.044774532318115, 'timing/data': 0.02265039598569274, 'timing/model': 1.5521034649573267, 'learning_rate/qwen_vl_interface': 8.834409048703004e-07, 'learning_rate/action_model': 8.834409048703005e-06, 'epoch': 1.77}) +[2026-06-16 10:53:32,489][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.02318282052874565, 'train/grad_norm_pre_clip': 17.411502838134766, 'timing/data': 0.025056928861886263, 'timing/model': 1.5618615876883268, 'learning_rate/qwen_vl_interface': 8.822430803806098e-07, 'learning_rate/action_model': 8.822430803806097e-06, 'epoch': 1.77}) +[2026-06-16 10:53:39,016][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.0018509651999920607, 'train/grad_norm_pre_clip': 0.868460476398468, 'timing/data': 0.02075743582099676, 'timing/model': 1.6930523831397295, 'learning_rate/qwen_vl_interface': 8.810470513108041e-07, 'learning_rate/action_model': 8.81047051310804e-06, 'epoch': 1.77}) +[2026-06-16 10:53:45,353][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 0.0024982085451483727, 'train/grad_norm_pre_clip': 0.9426600337028503, 'timing/data': 0.021781891118735075, 'timing/model': 1.8232879438437521, 'learning_rate/qwen_vl_interface': 8.798528181525262e-07, 'learning_rate/action_model': 8.798528181525261e-06, 'epoch': 1.77}) +[2026-06-16 10:53:51,496][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.039545733481645584, 'train/grad_norm_pre_clip': 31.576894760131836, 'timing/data': 0.02060574386268854, 'timing/model': 1.5768288150429726, 'learning_rate/qwen_vl_interface': 8.786603813966798e-07, 'learning_rate/action_model': 8.786603813966798e-06, 'epoch': 1.77}) +[2026-06-16 10:53:57,871][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 0.00032259896397590637, 'train/grad_norm_pre_clip': 0.038729194551706314, 'timing/data': 0.02129408810287714, 'timing/model': 1.5580500299111009, 'learning_rate/qwen_vl_interface': 8.774697415334304e-07, 'learning_rate/action_model': 8.774697415334304e-06, 'epoch': 1.77}) +[2026-06-16 10:54:04,268][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.0037394396495074034, 'train/grad_norm_pre_clip': 2.054131507873535, 'timing/data': 0.020694424863904715, 'timing/model': 1.5510449092835188, 'learning_rate/qwen_vl_interface': 8.762808990522056e-07, 'learning_rate/action_model': 8.762808990522057e-06, 'epoch': 1.77}) +[2026-06-16 10:54:10,677][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.033491313457489014, 'train/grad_norm_pre_clip': 19.263587951660156, 'timing/data': 0.02101597608998418, 'timing/model': 1.576987011358142, 'learning_rate/qwen_vl_interface': 8.750938544416934e-07, 'learning_rate/action_model': 8.750938544416934e-06, 'epoch': 1.77}) +[2026-06-16 10:54:17,073][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.0006060645682737231, 'train/grad_norm_pre_clip': 0.15533161163330078, 'timing/data': 0.020970036275684834, 'timing/model': 1.586590362712741, 'learning_rate/qwen_vl_interface': 8.739086081898437e-07, 'learning_rate/action_model': 8.739086081898437e-06, 'epoch': 1.77}) +[2026-06-16 10:54:23,453][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 0.1666637659072876, 'train/grad_norm_pre_clip': 23.84042739868164, 'timing/data': 0.02207844192162156, 'timing/model': 1.8273220751434565, 'learning_rate/qwen_vl_interface': 8.727251607838647e-07, 'learning_rate/action_model': 8.727251607838647e-06, 'epoch': 1.77}) +[2026-06-16 10:54:29,607][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.18748965859413147, 'train/grad_norm_pre_clip': 28.234777450561523, 'timing/data': 0.020515493117272854, 'timing/model': 1.5801944439299405, 'learning_rate/qwen_vl_interface': 8.715435127102281e-07, 'learning_rate/action_model': 8.715435127102281e-06, 'epoch': 1.77}) +[2026-06-16 10:54:35,972][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.12476417422294617, 'train/grad_norm_pre_clip': 33.56900405883789, 'timing/data': 0.02157530002295971, 'timing/model': 1.5488801482133567, 'learning_rate/qwen_vl_interface': 8.703636644546651e-07, 'learning_rate/action_model': 8.703636644546651e-06, 'epoch': 1.77}) +[2026-06-16 10:54:42,272][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.008702829480171204, 'train/grad_norm_pre_clip': 5.590808391571045, 'timing/data': 0.023978902958333492, 'timing/model': 1.548758045770228, 'learning_rate/qwen_vl_interface': 8.69185616502166e-07, 'learning_rate/action_model': 8.69185616502166e-06, 'epoch': 1.77}) +[2026-06-16 10:54:48,774][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.17825843393802643, 'train/grad_norm_pre_clip': 59.688472747802734, 'timing/data': 0.02213580207899213, 'timing/model': 1.5542126949876547, 'learning_rate/qwen_vl_interface': 8.680093693369821e-07, 'learning_rate/action_model': 8.68009369336982e-06, 'epoch': 1.77}) +[2026-06-16 10:54:55,174][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.0030389593448489904, 'train/grad_norm_pre_clip': 2.2916760444641113, 'timing/data': 0.021662369836121798, 'timing/model': 1.5316514782607555, 'learning_rate/qwen_vl_interface': 8.66834923442624e-07, 'learning_rate/action_model': 8.66834923442624e-06, 'epoch': 1.77}) +[2026-06-16 10:55:01,637][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.0014341776259243488, 'train/grad_norm_pre_clip': 0.4455134868621826, 'timing/data': 0.021898641251027584, 'timing/model': 1.8049474298022687, 'learning_rate/qwen_vl_interface': 8.65662279301863e-07, 'learning_rate/action_model': 8.656622793018628e-06, 'epoch': 1.77}) +[2026-06-16 10:55:07,876][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.0018193467985838652, 'train/grad_norm_pre_clip': 1.483633279800415, 'timing/data': 0.021938591729849577, 'timing/model': 1.7499965531751513, 'learning_rate/qwen_vl_interface': 8.644914373967274e-07, 'learning_rate/action_model': 8.644914373967273e-06, 'epoch': 1.77}) +[2026-06-16 10:55:13,977][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 0.10898460447788239, 'train/grad_norm_pre_clip': 27.477718353271484, 'timing/data': 0.023193038068711758, 'timing/model': 1.5474555706605315, 'learning_rate/qwen_vl_interface': 8.633223982085068e-07, 'learning_rate/action_model': 8.633223982085068e-06, 'epoch': 1.77}) +[2026-06-16 10:55:20,363][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.002309642732143402, 'train/grad_norm_pre_clip': 1.4280822277069092, 'timing/data': 0.02141360891982913, 'timing/model': 1.5954726450145245, 'learning_rate/qwen_vl_interface': 8.621551622177499e-07, 'learning_rate/action_model': 8.621551622177499e-06, 'epoch': 1.77}) +[2026-06-16 10:55:26,870][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 0.04528624191880226, 'train/grad_norm_pre_clip': 23.5081729888916, 'timing/data': 0.022391163744032383, 'timing/model': 1.5701752942986786, 'learning_rate/qwen_vl_interface': 8.609897299042615e-07, 'learning_rate/action_model': 8.609897299042614e-06, 'epoch': 1.77}) +[2026-06-16 10:55:33,273][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.19191665947437286, 'train/grad_norm_pre_clip': 22.301347732543945, 'timing/data': 0.020447032991796732, 'timing/model': 1.5642230799421668, 'learning_rate/qwen_vl_interface': 8.598261017471088e-07, 'learning_rate/action_model': 8.598261017471088e-06, 'epoch': 1.77}) +[2026-06-16 10:55:39,658][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.07609586417675018, 'train/grad_norm_pre_clip': 23.783748626708984, 'timing/data': 0.022353743202984333, 'timing/model': 1.8280873470939696, 'learning_rate/qwen_vl_interface': 8.586642782246133e-07, 'learning_rate/action_model': 8.586642782246131e-06, 'epoch': 1.78}) +[2026-06-16 10:55:46,062][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.00394823681563139, 'train/grad_norm_pre_clip': 2.610990047454834, 'timing/data': 0.022700326051563025, 'timing/model': 1.8447035667486489, 'learning_rate/qwen_vl_interface': 8.57504259814358e-07, 'learning_rate/action_model': 8.57504259814358e-06, 'epoch': 1.78}) +[2026-06-16 10:55:52,167][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.07927432656288147, 'train/grad_norm_pre_clip': 26.950855255126953, 'timing/data': 0.020510092843323946, 'timing/model': 1.5527297616936266, 'learning_rate/qwen_vl_interface': 8.563460469931823e-07, 'learning_rate/action_model': 8.563460469931822e-06, 'epoch': 1.78}) +[2026-06-16 10:55:58,466][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.0016291897045448422, 'train/grad_norm_pre_clip': 0.7755729556083679, 'timing/data': 0.020588044077157974, 'timing/model': 1.554913121741265, 'learning_rate/qwen_vl_interface': 8.551896402371834e-07, 'learning_rate/action_model': 8.551896402371834e-06, 'epoch': 1.78}) +[2026-06-16 10:56:04,865][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.009749260731041431, 'train/grad_norm_pre_clip': 11.06177806854248, 'timing/data': 0.021884590853005648, 'timing/model': 1.544649193994701, 'learning_rate/qwen_vl_interface': 8.540350400217177e-07, 'learning_rate/action_model': 8.540350400217176e-06, 'epoch': 1.78}) +[2026-06-16 10:56:11,355][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 0.0022227170411497355, 'train/grad_norm_pre_clip': 0.6759927272796631, 'timing/data': 0.020703504793345928, 'timing/model': 1.5910153090953827, 'learning_rate/qwen_vl_interface': 8.528822468213964e-07, 'learning_rate/action_model': 8.528822468213964e-06, 'epoch': 1.78}) +[2026-06-16 10:56:17,732][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.0011939264368265867, 'train/grad_norm_pre_clip': 0.35717883706092834, 'timing/data': 0.020794674288481474, 'timing/model': 1.7868294031359255, 'learning_rate/qwen_vl_interface': 8.517312611100898e-07, 'learning_rate/action_model': 8.517312611100897e-06, 'epoch': 1.78}) +[2026-06-16 10:56:23,886][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.01116188708692789, 'train/grad_norm_pre_clip': 8.664142608642578, 'timing/data': 0.020976365078240633, 'timing/model': 1.573159079067409, 'learning_rate/qwen_vl_interface': 8.505820833609249e-07, 'learning_rate/action_model': 8.50582083360925e-06, 'epoch': 1.78}) +[2026-06-16 10:56:30,170][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.09968251734972, 'train/grad_norm_pre_clip': 15.23823356628418, 'timing/data': 0.021378007251769304, 'timing/model': 1.5560340122319758, 'learning_rate/qwen_vl_interface': 8.494347140462847e-07, 'learning_rate/action_model': 8.494347140462846e-06, 'epoch': 1.78}) +[2026-06-16 10:56:36,570][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.07856641709804535, 'train/grad_norm_pre_clip': 25.38560676574707, 'timing/data': 0.02090710634365678, 'timing/model': 1.5615006079897285, 'learning_rate/qwen_vl_interface': 8.482891536378105e-07, 'learning_rate/action_model': 8.482891536378105e-06, 'epoch': 1.78}) +[2026-06-16 11:07:19,040][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.08155353367328644, 'train/grad_norm_pre_clip': 23.456676483154297, 'eval/loss': 0.034050822257995605, 'eval/flap_precision': 0.9714285714285714, 'eval/flap_recall': 0.9740791268758526, 'eval/flap_f1': 0.9727520435967301, 'eval/flap_rate_model': 0.20416666666666666, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9727520435967302, 'timing/data': 0.02123412676155567, 'timing/model': 1.5869581359438598, 'learning_rate/qwen_vl_interface': 8.471454026063972e-07, 'learning_rate/action_model': 8.47145402606397e-06, 'epoch': 1.78}) +[2026-06-16 11:07:19,041][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4400_state +[2026-06-16 11:07:19,047][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 11:07:27,422][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4400_state/model.safetensors +[2026-06-16 11:07:47,260][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4400_state/optimizer.bin +[2026-06-16 11:07:47,261][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4400_state/sampler.bin +[2026-06-16 11:07:47,261][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4400_state/sampler_1.bin +[2026-06-16 11:07:47,263][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4400_state/random_states_0.pkl +[2026-06-16 11:07:47,449][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4400 +[2026-06-16 11:07:47,450][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 11:07:47,465][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 11:07:53,741][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.022545799612998962, 'train/grad_norm_pre_clip': 19.3637638092041, 'timing/data': 0.021715089213103056, 'timing/model': 1.5588141027837992, 'learning_rate/qwen_vl_interface': 8.460034614221988e-07, 'learning_rate/action_model': 8.460034614221988e-06, 'epoch': 1.78}) +[2026-06-16 11:08:00,037][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.001940208370797336, 'train/grad_norm_pre_clip': 0.4965006113052368, 'timing/data': 0.021379776764661074, 'timing/model': 1.57548351213336, 'learning_rate/qwen_vl_interface': 8.448633305546236e-07, 'learning_rate/action_model': 8.448633305546235e-06, 'epoch': 1.78}) +[2026-06-16 11:08:06,420][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.08645519614219666, 'train/grad_norm_pre_clip': 26.79343605041504, 'timing/data': 0.019554506987333298, 'timing/model': 1.8967325598932803, 'learning_rate/qwen_vl_interface': 8.437250104723369e-07, 'learning_rate/action_model': 8.437250104723369e-06, 'epoch': 1.78}) +[2026-06-16 11:08:12,709][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.028586823493242264, 'train/grad_norm_pre_clip': 24.96805763244629, 'timing/data': 0.02211947087198496, 'timing/model': 1.8144461270421743, 'learning_rate/qwen_vl_interface': 8.42588501643259e-07, 'learning_rate/action_model': 8.42588501643259e-06, 'epoch': 1.78}) +[2026-06-16 11:08:18,749][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.00043286706204526126, 'train/grad_norm_pre_clip': 0.09513270109891891, 'timing/data': 0.021533807273954153, 'timing/model': 1.533197688870132, 'learning_rate/qwen_vl_interface': 8.414538045345655e-07, 'learning_rate/action_model': 8.414538045345654e-06, 'epoch': 1.78}) +[2026-06-16 11:08:25,050][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.020632263273000717, 'train/grad_norm_pre_clip': 18.981096267700195, 'timing/data': 0.02170564839616418, 'timing/model': 1.551479266025126, 'learning_rate/qwen_vl_interface': 8.403209196126871e-07, 'learning_rate/action_model': 8.403209196126871e-06, 'epoch': 1.78}) +[2026-06-16 11:08:31,449][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.0015925413463264704, 'train/grad_norm_pre_clip': 0.6867159605026245, 'timing/data': 0.02205935027450323, 'timing/model': 1.5621390049345791, 'learning_rate/qwen_vl_interface': 8.391898473433111e-07, 'learning_rate/action_model': 8.39189847343311e-06, 'epoch': 1.78}) +[2026-06-16 11:08:37,834][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.02063964121043682, 'train/grad_norm_pre_clip': 15.494651794433594, 'timing/data': 0.02123338682577014, 'timing/model': 1.5802726512774825, 'learning_rate/qwen_vl_interface': 8.380605881913778e-07, 'learning_rate/action_model': 8.380605881913779e-06, 'epoch': 1.78}) +[2026-06-16 11:08:44,156][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 0.40031811594963074, 'train/grad_norm_pre_clip': 18.36544418334961, 'timing/data': 0.021656048018485308, 'timing/model': 1.5728744696825743, 'learning_rate/qwen_vl_interface': 8.369331426210832e-07, 'learning_rate/action_model': 8.369331426210832e-06, 'epoch': 1.78}) +[2026-06-16 11:08:50,482][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.0001835639268392697, 'train/grad_norm_pre_clip': 0.04192521423101425, 'timing/data': 0.020438421983271837, 'timing/model': 1.778873898088932, 'learning_rate/qwen_vl_interface': 8.358075110958785e-07, 'learning_rate/action_model': 8.358075110958784e-06, 'epoch': 1.78}) +[2026-06-16 11:08:56,561][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.01299525611102581, 'train/grad_norm_pre_clip': 9.279670715332031, 'timing/data': 0.02108140615746379, 'timing/model': 1.5536445737816393, 'learning_rate/qwen_vl_interface': 8.346836940784661e-07, 'learning_rate/action_model': 8.346836940784661e-06, 'epoch': 1.78}) +[2026-06-16 11:09:02,860][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.00030071043875068426, 'train/grad_norm_pre_clip': 0.10208673775196075, 'timing/data': 0.021376207005232573, 'timing/model': 1.5524056036956608, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 1.78}) +[2026-06-16 11:09:09,162][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.023863298818469048, 'train/grad_norm_pre_clip': 17.52573585510254, 'timing/data': 0.02064072387292981, 'timing/model': 1.521525353193283, 'learning_rate/qwen_vl_interface': 8.324415054141111e-07, 'learning_rate/action_model': 8.32441505414111e-06, 'epoch': 1.78}) +[2026-06-16 11:09:15,561][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.018527351319789886, 'train/grad_norm_pre_clip': 16.75994300842285, 'timing/data': 0.02246371330693364, 'timing/model': 1.588289933744818, 'learning_rate/qwen_vl_interface': 8.313231346888476e-07, 'learning_rate/action_model': 8.313231346888476e-06, 'epoch': 1.78}) +[2026-06-16 11:09:21,926][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.023216797038912773, 'train/grad_norm_pre_clip': 16.615219116210938, 'timing/data': 0.021698468830436468, 'timing/model': 1.7775649018585682, 'learning_rate/qwen_vl_interface': 8.302065803147353e-07, 'learning_rate/action_model': 8.302065803147353e-06, 'epoch': 1.79}) +[2026-06-16 11:09:27,962][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.07278972864151001, 'train/grad_norm_pre_clip': 32.7586669921875, 'timing/data': 0.021911160089075565, 'timing/model': 1.5300396820530295, 'learning_rate/qwen_vl_interface': 8.29091842750748e-07, 'learning_rate/action_model': 8.29091842750748e-06, 'epoch': 1.79}) +[2026-06-16 11:09:34,255][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 0.0010804837802425027, 'train/grad_norm_pre_clip': 0.6609586477279663, 'timing/data': 0.020903345197439194, 'timing/model': 1.5519110863097012, 'learning_rate/qwen_vl_interface': 8.279789224551116e-07, 'learning_rate/action_model': 8.279789224551116e-06, 'epoch': 1.79}) +[2026-06-16 11:09:40,544][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.0011464098934084177, 'train/grad_norm_pre_clip': 0.4089387059211731, 'timing/data': 0.01962897926568985, 'timing/model': 1.630698368884623, 'learning_rate/qwen_vl_interface': 8.268678198853058e-07, 'learning_rate/action_model': 8.268678198853059e-06, 'epoch': 1.79}) +[2026-06-16 11:09:46,862][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.2074485868215561, 'train/grad_norm_pre_clip': 50.22298812866211, 'timing/data': 0.02211870113387704, 'timing/model': 1.5452935029752553, 'learning_rate/qwen_vl_interface': 8.257585354980628e-07, 'learning_rate/action_model': 8.257585354980628e-06, 'epoch': 1.79}) +[2026-06-16 11:09:53,246][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.0008757324540056288, 'train/grad_norm_pre_clip': 0.315218985080719, 'timing/data': 0.0254151183180511, 'timing/model': 1.5997517732903361, 'learning_rate/qwen_vl_interface': 8.246510697493681e-07, 'learning_rate/action_model': 8.246510697493682e-06, 'epoch': 1.79}) +[2026-06-16 11:09:59,649][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.0004855816368944943, 'train/grad_norm_pre_clip': 0.08856922388076782, 'timing/data': 0.021947890054434538, 'timing/model': 1.8374292878434062, 'learning_rate/qwen_vl_interface': 8.235454230944584e-07, 'learning_rate/action_model': 8.235454230944584e-06, 'epoch': 1.79}) +[2026-06-16 11:10:05,671][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.0002458849339745939, 'train/grad_norm_pre_clip': 0.04517845809459686, 'timing/data': 0.02101863594725728, 'timing/model': 1.5541039421223104, 'learning_rate/qwen_vl_interface': 8.224415959878234e-07, 'learning_rate/action_model': 8.224415959878234e-06, 'epoch': 1.79}) +[2026-06-16 11:10:12,066][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.004576848354190588, 'train/grad_norm_pre_clip': 5.215670108795166, 'timing/data': 0.02149056876078248, 'timing/model': 1.550019915215671, 'learning_rate/qwen_vl_interface': 8.213395888832052e-07, 'learning_rate/action_model': 8.213395888832051e-06, 'epoch': 1.79}) +[2026-06-16 11:10:18,480][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.014912240207195282, 'train/grad_norm_pre_clip': 10.073385238647461, 'timing/data': 0.022975215688347816, 'timing/model': 1.5493362178094685, 'learning_rate/qwen_vl_interface': 8.202394022335958e-07, 'learning_rate/action_model': 8.202394022335957e-06, 'epoch': 1.79}) +[2026-06-16 11:10:24,962][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 0.1277141124010086, 'train/grad_norm_pre_clip': 13.598675727844238, 'timing/data': 0.02127220807597041, 'timing/model': 1.5432647010311484, 'learning_rate/qwen_vl_interface': 8.191410364912424e-07, 'learning_rate/action_model': 8.191410364912424e-06, 'epoch': 1.79}) +[2026-06-16 11:10:31,438][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.0011116709792986512, 'train/grad_norm_pre_clip': 0.18709202110767365, 'timing/data': 0.023000186774879694, 'timing/model': 1.6462405268102884, 'learning_rate/qwen_vl_interface': 8.1804449210764e-07, 'learning_rate/action_model': 8.1804449210764e-06, 'epoch': 1.79}) +[2026-06-16 11:10:37,765][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.0593172051012516, 'train/grad_norm_pre_clip': 21.514381408691406, 'timing/data': 0.021714990958571434, 'timing/model': 1.8434560210444033, 'learning_rate/qwen_vl_interface': 8.169497695335377e-07, 'learning_rate/action_model': 8.169497695335376e-06, 'epoch': 1.79}) +[2026-06-16 11:10:43,873][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 0.0003804208245128393, 'train/grad_norm_pre_clip': 0.15073664486408234, 'timing/data': 0.021197957918047905, 'timing/model': 1.5545669412240386, 'learning_rate/qwen_vl_interface': 8.158568692189342e-07, 'learning_rate/action_model': 8.158568692189342e-06, 'epoch': 1.79}) +[2026-06-16 11:10:50,299][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.06782456487417221, 'train/grad_norm_pre_clip': 30.810245513916016, 'timing/data': 0.021081247366964817, 'timing/model': 1.580343646928668, 'learning_rate/qwen_vl_interface': 8.147657916130797e-07, 'learning_rate/action_model': 8.147657916130796e-06, 'epoch': 1.79}) +[2026-06-16 11:10:56,677][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.16655713319778442, 'train/grad_norm_pre_clip': 15.728872299194336, 'timing/data': 0.02098354697227478, 'timing/model': 1.569021264091134, 'learning_rate/qwen_vl_interface': 8.136765371644753e-07, 'learning_rate/action_model': 8.136765371644753e-06, 'epoch': 1.79}) +[2026-06-16 11:11:02,974][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.06246381625533104, 'train/grad_norm_pre_clip': 39.29638671875, 'timing/data': 0.02243393426761031, 'timing/model': 1.5309383147396147, 'learning_rate/qwen_vl_interface': 8.125891063208713e-07, 'learning_rate/action_model': 8.125891063208712e-06, 'epoch': 1.79}) +[2026-06-16 11:11:09,467][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.00030301036895252764, 'train/grad_norm_pre_clip': 0.15000863373279572, 'timing/data': 0.020536443684250116, 'timing/model': 1.5776481456123292, 'learning_rate/qwen_vl_interface': 8.115034995292698e-07, 'learning_rate/action_model': 8.115034995292698e-06, 'epoch': 1.79}) +[2026-06-16 11:11:15,851][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.14383642375469208, 'train/grad_norm_pre_clip': 15.029376029968262, 'timing/data': 0.021836011204868555, 'timing/model': 1.825173539109528, 'learning_rate/qwen_vl_interface': 8.104197172359235e-07, 'learning_rate/action_model': 8.104197172359234e-06, 'epoch': 1.79}) +[2026-06-16 11:11:22,043][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 9.30495880311355e-05, 'train/grad_norm_pre_clip': 0.00680723087862134, 'timing/data': 0.02058095484972, 'timing/model': 1.622488962020725, 'learning_rate/qwen_vl_interface': 8.093377598863334e-07, 'learning_rate/action_model': 8.093377598863334e-06, 'epoch': 1.79}) +[2026-06-16 11:11:28,382][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.04746764898300171, 'train/grad_norm_pre_clip': 30.488662719726562, 'timing/data': 0.021860350854694843, 'timing/model': 1.5632872162386775, 'learning_rate/qwen_vl_interface': 8.082576279252519e-07, 'learning_rate/action_model': 8.082576279252518e-06, 'epoch': 1.79}) +[2026-06-16 11:11:34,678][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.0005718540633097291, 'train/grad_norm_pre_clip': 0.12994052469730377, 'timing/data': 0.021166508086025715, 'timing/model': 1.5714795207604766, 'learning_rate/qwen_vl_interface': 8.071793217966803e-07, 'learning_rate/action_model': 8.071793217966802e-06, 'epoch': 1.79}) +[2026-06-16 11:11:41,059][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.11642993986606598, 'train/grad_norm_pre_clip': 9.757415771484375, 'timing/data': 0.02151074865832925, 'timing/model': 1.5497855748981237, 'learning_rate/qwen_vl_interface': 8.061028419438691e-07, 'learning_rate/action_model': 8.06102841943869e-06, 'epoch': 1.79}) +[2026-06-16 11:11:47,473][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 0.003180937608703971, 'train/grad_norm_pre_clip': 1.1987948417663574, 'timing/data': 0.01997071085497737, 'timing/model': 1.5180925778113306, 'learning_rate/qwen_vl_interface': 8.050281888093193e-07, 'learning_rate/action_model': 8.050281888093192e-06, 'epoch': 1.79}) +[2026-06-16 11:11:53,817][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.3080615699291229, 'train/grad_norm_pre_clip': 48.16352462768555, 'timing/data': 0.022162492852658033, 'timing/model': 1.7869690218940377, 'learning_rate/qwen_vl_interface': 8.039553628347794e-07, 'learning_rate/action_model': 8.039553628347793e-06, 'epoch': 1.8}) +[2026-06-16 11:11:59,870][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.012666748836636543, 'train/grad_norm_pre_clip': 8.903714179992676, 'timing/data': 0.02048915484920144, 'timing/model': 1.5617128666490316, 'learning_rate/qwen_vl_interface': 8.028843644612483e-07, 'learning_rate/action_model': 8.028843644612482e-06, 'epoch': 1.8}) +[2026-06-16 11:12:06,358][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.00015445153985638171, 'train/grad_norm_pre_clip': 0.03231240063905716, 'timing/data': 0.02112609799951315, 'timing/model': 1.6438584201969206, 'learning_rate/qwen_vl_interface': 8.01815194128973e-07, 'learning_rate/action_model': 8.01815194128973e-06, 'epoch': 1.8}) +[2026-06-16 11:12:12,668][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.018594980239868164, 'train/grad_norm_pre_clip': 16.519264221191406, 'timing/data': 0.021723641082644463, 'timing/model': 1.5509961303323507, 'learning_rate/qwen_vl_interface': 8.007478522774481e-07, 'learning_rate/action_model': 8.00747852277448e-06, 'epoch': 1.8}) +[2026-06-16 11:12:18,970][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.0009038385469466448, 'train/grad_norm_pre_clip': 0.22277699410915375, 'timing/data': 0.02153006987646222, 'timing/model': 1.5463993102312088, 'learning_rate/qwen_vl_interface': 7.996823393454197e-07, 'learning_rate/action_model': 7.996823393454196e-06, 'epoch': 1.8}) +[2026-06-16 11:12:25,468][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.000397965544834733, 'train/grad_norm_pre_clip': 0.08465571701526642, 'timing/data': 0.02149038016796112, 'timing/model': 1.5855181133374572, 'learning_rate/qwen_vl_interface': 7.986186557708773e-07, 'learning_rate/action_model': 7.986186557708771e-06, 'epoch': 1.8}) +[2026-06-16 11:12:31,916][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.12138602882623672, 'train/grad_norm_pre_clip': 41.97543716430664, 'timing/data': 0.020724565722048283, 'timing/model': 1.8206405988894403, 'learning_rate/qwen_vl_interface': 7.975568019910617e-07, 'learning_rate/action_model': 7.975568019910617e-06, 'epoch': 1.8}) +[2026-06-16 11:12:38,075][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.0045522041618824005, 'train/grad_norm_pre_clip': 5.368054389953613, 'timing/data': 0.020601075142621994, 'timing/model': 1.5712836682796478, 'learning_rate/qwen_vl_interface': 7.964967784424618e-07, 'learning_rate/action_model': 7.964967784424618e-06, 'epoch': 1.8}) +[2026-06-16 11:12:44,383][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.12482669204473495, 'train/grad_norm_pre_clip': 19.91787338256836, 'timing/data': 0.021281349007040262, 'timing/model': 1.5577727230265737, 'learning_rate/qwen_vl_interface': 7.954385855608129e-07, 'learning_rate/action_model': 7.95438585560813e-06, 'epoch': 1.8}) +[2026-06-16 11:12:50,672][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.1643461436033249, 'train/grad_norm_pre_clip': 24.313016891479492, 'timing/data': 0.020588174927979708, 'timing/model': 1.5655591948889196, 'learning_rate/qwen_vl_interface': 7.943822237810974e-07, 'learning_rate/action_model': 7.943822237810975e-06, 'epoch': 1.8}) +[2026-06-16 11:12:57,067][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.00015511797391809523, 'train/grad_norm_pre_clip': 0.015390250831842422, 'timing/data': 0.01935957930982113, 'timing/model': 1.5660861721262336, 'learning_rate/qwen_vl_interface': 7.93327693537547e-07, 'learning_rate/action_model': 7.93327693537547e-06, 'epoch': 1.8}) +[2026-06-16 11:13:03,466][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.005359877832233906, 'train/grad_norm_pre_clip': 4.439257621765137, 'timing/data': 0.01887773582711816, 'timing/model': 1.589212004095316, 'learning_rate/qwen_vl_interface': 7.92274995263638e-07, 'learning_rate/action_model': 7.92274995263638e-06, 'epoch': 1.8}) +[2026-06-16 11:13:09,913][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.27953827381134033, 'train/grad_norm_pre_clip': 52.223392486572266, 'timing/data': 0.024926996789872646, 'timing/model': 1.9005357641726732, 'learning_rate/qwen_vl_interface': 7.912241293920958e-07, 'learning_rate/action_model': 7.912241293920959e-06, 'epoch': 1.8}) +[2026-06-16 11:13:16,144][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.2550205588340759, 'train/grad_norm_pre_clip': 12.663211822509766, 'timing/data': 0.021371349226683378, 'timing/model': 1.7303289650008082, 'learning_rate/qwen_vl_interface': 7.90175096354891e-07, 'learning_rate/action_model': 7.90175096354891e-06, 'epoch': 1.8}) +[2026-06-16 11:13:22,171][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.21752026677131653, 'train/grad_norm_pre_clip': 30.059728622436523, 'timing/data': 0.021537779830396175, 'timing/model': 1.5500715430825949, 'learning_rate/qwen_vl_interface': 7.891278965832431e-07, 'learning_rate/action_model': 7.891278965832432e-06, 'epoch': 1.8}) +[2026-06-16 11:13:28,477][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.0015338505618274212, 'train/grad_norm_pre_clip': 0.6344071626663208, 'timing/data': 0.022214633878320456, 'timing/model': 1.5599727970547974, 'learning_rate/qwen_vl_interface': 7.880825305076138e-07, 'learning_rate/action_model': 7.880825305076139e-06, 'epoch': 1.8}) +[2026-06-16 11:13:34,974][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.000515759747941047, 'train/grad_norm_pre_clip': 0.15812116861343384, 'timing/data': 0.020342704840004444, 'timing/model': 1.567532307934016, 'learning_rate/qwen_vl_interface': 7.870389985577148e-07, 'learning_rate/action_model': 7.870389985577148e-06, 'epoch': 1.8}) +[2026-06-16 11:13:41,371][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.00258965534158051, 'train/grad_norm_pre_clip': 4.262359142303467, 'timing/data': 0.02141069993376732, 'timing/model': 1.5943518388085067, 'learning_rate/qwen_vl_interface': 7.859973011625028e-07, 'learning_rate/action_model': 7.859973011625026e-06, 'epoch': 1.8}) +[2026-06-16 11:13:47,837][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.0005171921802684665, 'train/grad_norm_pre_clip': 0.11180275678634644, 'timing/data': 0.020614145789295435, 'timing/model': 1.8233828288502991, 'learning_rate/qwen_vl_interface': 7.849574387501801e-07, 'learning_rate/action_model': 7.8495743875018e-06, 'epoch': 1.8}) +[2026-06-16 11:13:54,133][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.0007713919621892273, 'train/grad_norm_pre_clip': 0.18497894704341888, 'timing/data': 0.021035168319940567, 'timing/model': 1.7979349219240248, 'learning_rate/qwen_vl_interface': 7.839194117481952e-07, 'learning_rate/action_model': 7.839194117481952e-06, 'epoch': 1.8}) +[2026-06-16 11:14:00,175][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.0002523012226447463, 'train/grad_norm_pre_clip': 0.04147498682141304, 'timing/data': 0.02111057797446847, 'timing/model': 1.5522215622477233, 'learning_rate/qwen_vl_interface': 7.828832205832418e-07, 'learning_rate/action_model': 7.828832205832418e-06, 'epoch': 1.8}) +[2026-06-16 11:14:06,571][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.00512759480625391, 'train/grad_norm_pre_clip': 2.0006210803985596, 'timing/data': 0.021500419825315475, 'timing/model': 1.5516336830332875, 'learning_rate/qwen_vl_interface': 7.818488656812581e-07, 'learning_rate/action_model': 7.818488656812582e-06, 'epoch': 1.8}) +[2026-06-16 11:14:13,076][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 0.15991689264774323, 'train/grad_norm_pre_clip': 15.679366111755371, 'timing/data': 0.026078274939209223, 'timing/model': 1.5630422439426184, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 1.8}) +[2026-06-16 11:14:19,752][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.0015671496512368321, 'train/grad_norm_pre_clip': 0.43794938921928406, 'timing/data': 0.021099197678267956, 'timing/model': 1.589026470668614, 'learning_rate/qwen_vl_interface': 7.797856663661836e-07, 'learning_rate/action_model': 7.797856663661835e-06, 'epoch': 1.8}) +[2026-06-16 11:14:26,132][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.006846722215414047, 'train/grad_norm_pre_clip': 4.2068023681640625, 'timing/data': 0.022193993907421827, 'timing/model': 1.818900268059224, 'learning_rate/qwen_vl_interface': 7.787568228011959e-07, 'learning_rate/action_model': 7.78756822801196e-06, 'epoch': 1.8}) +[2026-06-16 11:14:32,170][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.03571658581495285, 'train/grad_norm_pre_clip': 23.21432876586914, 'timing/data': 0.02061032596975565, 'timing/model': 1.5568758710287511, 'learning_rate/qwen_vl_interface': 7.777298171953858e-07, 'learning_rate/action_model': 7.777298171953857e-06, 'epoch': 1.81}) +[2026-06-16 11:14:38,563][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.04523133113980293, 'train/grad_norm_pre_clip': 24.85972785949707, 'timing/data': 0.020921348128467798, 'timing/model': 1.5908119487576187, 'learning_rate/qwen_vl_interface': 7.767046499709144e-07, 'learning_rate/action_model': 7.767046499709143e-06, 'epoch': 1.81}) +[2026-06-16 11:14:45,045][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.19710110127925873, 'train/grad_norm_pre_clip': 7.835766315460205, 'timing/data': 0.020520695950835943, 'timing/model': 1.6364525873214006, 'learning_rate/qwen_vl_interface': 7.756813215491904e-07, 'learning_rate/action_model': 7.756813215491904e-06, 'epoch': 1.81}) +[2026-06-16 11:14:51,356][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.12240216881036758, 'train/grad_norm_pre_clip': 26.393796920776367, 'timing/data': 0.020008062943816185, 'timing/model': 1.591636598110199, 'learning_rate/qwen_vl_interface': 7.746598323508655e-07, 'learning_rate/action_model': 7.746598323508655e-06, 'epoch': 1.81}) +[2026-06-16 11:14:57,667][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.00048360932851210237, 'train/grad_norm_pre_clip': 0.08433938771486282, 'timing/data': 0.020481355022639036, 'timing/model': 1.5573521251790226, 'learning_rate/qwen_vl_interface': 7.736401827958351e-07, 'learning_rate/action_model': 7.73640182795835e-06, 'epoch': 1.81}) +[2026-06-16 11:15:03,980][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 9.4402494141832e-05, 'train/grad_norm_pre_clip': 0.009831879287958145, 'timing/data': 0.02088588709011674, 'timing/model': 1.7598263323307037, 'learning_rate/qwen_vl_interface': 7.726223733032382e-07, 'learning_rate/action_model': 7.72622373303238e-06, 'epoch': 1.81}) +[2026-06-16 11:15:10,087][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.0027139296289533377, 'train/grad_norm_pre_clip': 2.2239832878112793, 'timing/data': 0.021892522927373648, 'timing/model': 1.563884184230119, 'learning_rate/qwen_vl_interface': 7.716064042914584e-07, 'learning_rate/action_model': 7.716064042914584e-06, 'epoch': 1.81}) +[2026-06-16 11:15:16,471][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.027169600129127502, 'train/grad_norm_pre_clip': 26.36568832397461, 'timing/data': 0.021996074356138706, 'timing/model': 1.5363962538540363, 'learning_rate/qwen_vl_interface': 7.705922761781226e-07, 'learning_rate/action_model': 7.705922761781226e-06, 'epoch': 1.81}) +[2026-06-16 11:15:22,873][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.06688052415847778, 'train/grad_norm_pre_clip': 38.19816207885742, 'timing/data': 0.022632896900177002, 'timing/model': 1.5484501384198666, 'learning_rate/qwen_vl_interface': 7.695799893800996e-07, 'learning_rate/action_model': 7.695799893800996e-06, 'epoch': 1.81}) +[2026-06-16 11:15:29,270][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.0066448720172047615, 'train/grad_norm_pre_clip': 3.396336317062378, 'timing/data': 0.021806063130497932, 'timing/model': 1.5301741366274655, 'learning_rate/qwen_vl_interface': 7.685695443135028e-07, 'learning_rate/action_model': 7.685695443135027e-06, 'epoch': 1.81}) +[2026-06-16 11:15:35,667][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 0.009851682931184769, 'train/grad_norm_pre_clip': 6.123751640319824, 'timing/data': 0.021394700277596712, 'timing/model': 1.5666105379350483, 'learning_rate/qwen_vl_interface': 7.675609413936888e-07, 'learning_rate/action_model': 7.675609413936888e-06, 'epoch': 1.81}) +[2026-06-16 11:15:42,084][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.12445326894521713, 'train/grad_norm_pre_clip': 24.442293167114258, 'timing/data': 0.021834342274814844, 'timing/model': 1.771550820209086, 'learning_rate/qwen_vl_interface': 7.665541810352541e-07, 'learning_rate/action_model': 7.66554181035254e-06, 'epoch': 1.81}) +[2026-06-16 11:15:48,171][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.1317916065454483, 'train/grad_norm_pre_clip': 24.615541458129883, 'timing/data': 0.021611901931464672, 'timing/model': 1.5505400099791586, 'learning_rate/qwen_vl_interface': 7.655492636520415e-07, 'learning_rate/action_model': 7.655492636520414e-06, 'epoch': 1.81}) +[2026-06-16 11:15:54,473][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.000268420233624056, 'train/grad_norm_pre_clip': 0.04777248203754425, 'timing/data': 0.02168176183477044, 'timing/model': 1.5497146379202604, 'learning_rate/qwen_vl_interface': 7.64546189657134e-07, 'learning_rate/action_model': 7.645461896571339e-06, 'epoch': 1.81}) +[2026-06-16 11:16:00,872][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.0003946576325688511, 'train/grad_norm_pre_clip': 0.32153645157814026, 'timing/data': 0.021380089689046144, 'timing/model': 1.5570543082430959, 'learning_rate/qwen_vl_interface': 7.635449594628571e-07, 'learning_rate/action_model': 7.635449594628571e-06, 'epoch': 1.81}) +[2026-06-16 11:16:07,266][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.19306530058383942, 'train/grad_norm_pre_clip': 7.31685209274292, 'timing/data': 0.020440205931663513, 'timing/model': 1.5991179193370044, 'learning_rate/qwen_vl_interface': 7.625455734807803e-07, 'learning_rate/action_model': 7.625455734807803e-06, 'epoch': 1.81}) +[2026-06-16 11:16:13,633][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.15332895517349243, 'train/grad_norm_pre_clip': 17.781038284301758, 'timing/data': 0.020382234826683998, 'timing/model': 1.8174355709925294, 'learning_rate/qwen_vl_interface': 7.615480321217116e-07, 'learning_rate/action_model': 7.615480321217117e-06, 'epoch': 1.81}) +[2026-06-16 11:16:19,766][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.13167326152324677, 'train/grad_norm_pre_clip': 14.893152236938477, 'timing/data': 0.02068868698552251, 'timing/model': 1.5840412736870348, 'learning_rate/qwen_vl_interface': 7.605523357957028e-07, 'learning_rate/action_model': 7.605523357957028e-06, 'epoch': 1.81}) +[2026-06-16 11:16:26,090][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.02128618396818638, 'train/grad_norm_pre_clip': 18.093786239624023, 'timing/data': 0.02119617024436593, 'timing/model': 1.5677275722846389, 'learning_rate/qwen_vl_interface': 7.59558484912048e-07, 'learning_rate/action_model': 7.59558484912048e-06, 'epoch': 1.81}) +[2026-06-16 11:16:32,373][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.0014753093710169196, 'train/grad_norm_pre_clip': 0.2988380491733551, 'timing/data': 0.02140062116086483, 'timing/model': 1.556330546271056, 'learning_rate/qwen_vl_interface': 7.585664798792807e-07, 'learning_rate/action_model': 7.585664798792807e-06, 'epoch': 1.81}) +[2026-06-16 11:16:38,913][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.14646032452583313, 'train/grad_norm_pre_clip': 11.278643608093262, 'timing/data': 0.01967411208897829, 'timing/model': 1.6048596501350403, 'learning_rate/qwen_vl_interface': 7.575763211051779e-07, 'learning_rate/action_model': 7.575763211051778e-06, 'epoch': 1.81}) +[2026-06-16 11:16:45,178][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.0012992650736123323, 'train/grad_norm_pre_clip': 0.5559572577476501, 'timing/data': 0.021806603763252497, 'timing/model': 1.5386845581233501, 'learning_rate/qwen_vl_interface': 7.565880089967553e-07, 'learning_rate/action_model': 7.565880089967553e-06, 'epoch': 1.81}) +[2026-06-16 11:16:51,529][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.006463085301220417, 'train/grad_norm_pre_clip': 4.57060432434082, 'timing/data': 0.02172047272324562, 'timing/model': 1.8101777778938413, 'learning_rate/qwen_vl_interface': 7.556015439602725e-07, 'learning_rate/action_model': 7.556015439602725e-06, 'epoch': 1.81}) +[2026-06-16 11:16:57,653][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.02871728502213955, 'train/grad_norm_pre_clip': 14.947233200073242, 'timing/data': 0.024052404798567295, 'timing/model': 1.6306707602925599, 'learning_rate/qwen_vl_interface': 7.546169264012252e-07, 'learning_rate/action_model': 7.5461692640122525e-06, 'epoch': 1.81}) +[2026-06-16 11:17:04,070][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.0010067387484014034, 'train/grad_norm_pre_clip': 0.3655070662498474, 'timing/data': 0.02110600983723998, 'timing/model': 1.5490321381948888, 'learning_rate/qwen_vl_interface': 7.536341567243544e-07, 'learning_rate/action_model': 7.536341567243543e-06, 'epoch': 1.81}) +[2026-06-16 11:17:10,483][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.00117271370254457, 'train/grad_norm_pre_clip': 0.3209497034549713, 'timing/data': 0.021730753127485514, 'timing/model': 1.558859771117568, 'learning_rate/qwen_vl_interface': 7.526532353336395e-07, 'learning_rate/action_model': 7.526532353336395e-06, 'epoch': 1.82}) +[2026-06-16 11:17:16,873][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.0009827966568991542, 'train/grad_norm_pre_clip': 0.16588656604290009, 'timing/data': 0.02419322496280074, 'timing/model': 1.5521042090840638, 'learning_rate/qwen_vl_interface': 7.516741626322994e-07, 'learning_rate/action_model': 7.516741626322994e-06, 'epoch': 1.82}) +[2026-06-16 11:17:23,374][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.0002682227932382375, 'train/grad_norm_pre_clip': 0.03513224795460701, 'timing/data': 0.021383370738476515, 'timing/model': 1.5801401063799858, 'learning_rate/qwen_vl_interface': 7.506969390227945e-07, 'learning_rate/action_model': 7.506969390227945e-06, 'epoch': 1.82}) +[2026-06-16 11:17:29,765][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.003945223055779934, 'train/grad_norm_pre_clip': 2.0981366634368896, 'timing/data': 0.021639653015881777, 'timing/model': 1.848899483680725, 'learning_rate/qwen_vl_interface': 7.497215649068244e-07, 'learning_rate/action_model': 7.497215649068244e-06, 'epoch': 1.82}) +[2026-06-16 11:17:35,870][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.00045500585110858083, 'train/grad_norm_pre_clip': 0.11912640184164047, 'timing/data': 0.021114539355039597, 'timing/model': 1.558660828974098, 'learning_rate/qwen_vl_interface': 7.48748040685329e-07, 'learning_rate/action_model': 7.487480406853289e-06, 'epoch': 1.82}) +[2026-06-16 11:17:42,162][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.000660349556710571, 'train/grad_norm_pre_clip': 0.14594770967960358, 'timing/data': 0.020090604200959206, 'timing/model': 1.5550988116301596, 'learning_rate/qwen_vl_interface': 7.477763667584867e-07, 'learning_rate/action_model': 7.477763667584866e-06, 'epoch': 1.82}) +[2026-06-16 11:17:48,565][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.10673869401216507, 'train/grad_norm_pre_clip': 19.65513801574707, 'timing/data': 0.019903854001313448, 'timing/model': 1.5472747129388154, 'learning_rate/qwen_vl_interface': 7.468065435257159e-07, 'learning_rate/action_model': 7.468065435257159e-06, 'epoch': 1.82}) +[2026-06-16 11:17:54,890][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.00977710634469986, 'train/grad_norm_pre_clip': 6.138924598693848, 'timing/data': 0.019970314111560583, 'timing/model': 1.5659764832817018, 'learning_rate/qwen_vl_interface': 7.458385713856745e-07, 'learning_rate/action_model': 7.458385713856744e-06, 'epoch': 1.82}) +[2026-06-16 11:18:01,290][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.004128770902752876, 'train/grad_norm_pre_clip': 1.7463401556015015, 'timing/data': 0.02159995213150978, 'timing/model': 1.6183350970968604, 'learning_rate/qwen_vl_interface': 7.448724507362605e-07, 'learning_rate/action_model': 7.448724507362605e-06, 'epoch': 1.82}) +[2026-06-16 11:18:07,631][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.03637779504060745, 'train/grad_norm_pre_clip': 57.190670013427734, 'timing/data': 0.02227217610925436, 'timing/model': 1.8142550280317664, 'learning_rate/qwen_vl_interface': 7.439081819746078e-07, 'learning_rate/action_model': 7.4390818197460775e-06, 'epoch': 1.82}) +[2026-06-16 11:18:13,735][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.00017445377307012677, 'train/grad_norm_pre_clip': 0.039016108959913254, 'timing/data': 0.02088780887424946, 'timing/model': 1.6176886167377234, 'learning_rate/qwen_vl_interface': 7.42945765497091e-07, 'learning_rate/action_model': 7.42945765497091e-06, 'epoch': 1.82}) +[2026-06-16 11:28:58,235][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.1424768716096878, 'train/grad_norm_pre_clip': 20.873140335083008, 'eval/loss': 0.03322194516658783, 'eval/flap_precision': 0.9702300405953992, 'eval/flap_recall': 0.9781718963165075, 'eval/flap_f1': 0.9741847826086956, 'eval/flap_rate_model': 0.20527777777777778, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9741847826086957, 'timing/data': 0.023470141924917698, 'timing/model': 1.5463087176904082, 'learning_rate/qwen_vl_interface': 7.41985201699324e-07, 'learning_rate/action_model': 7.41985201699324e-06, 'epoch': 1.82}) +[2026-06-16 11:28:58,236][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4500_state +[2026-06-16 11:28:58,242][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 11:29:06,566][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4500_state/model.safetensors +[2026-06-16 11:29:25,882][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4500_state/optimizer.bin +[2026-06-16 11:29:25,882][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4500_state/sampler.bin +[2026-06-16 11:29:25,883][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4500_state/sampler_1.bin +[2026-06-16 11:29:25,884][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4500_state/random_states_0.pkl +[2026-06-16 11:29:26,066][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-06-16 11:29:26,067][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 11:29:26,082][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 11:29:32,222][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.008620456792414188, 'train/grad_norm_pre_clip': 3.816823959350586, 'timing/data': 0.02587334578856826, 'timing/model': 1.624853567685932, 'learning_rate/qwen_vl_interface': 7.410264909761581e-07, 'learning_rate/action_model': 7.41026490976158e-06, 'epoch': 1.82}) +[2026-06-16 11:29:38,533][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.006707151420414448, 'train/grad_norm_pre_clip': 4.619536399841309, 'timing/data': 0.02822253806516528, 'timing/model': 1.537519657984376, 'learning_rate/qwen_vl_interface': 7.40069633721683e-07, 'learning_rate/action_model': 7.400696337216829e-06, 'epoch': 1.82}) +[2026-06-16 11:29:44,727][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.00029204916791059077, 'train/grad_norm_pre_clip': 0.046348001807928085, 'timing/data': 0.020240527112036943, 'timing/model': 1.5296221757307649, 'learning_rate/qwen_vl_interface': 7.391146303292262e-07, 'learning_rate/action_model': 7.391146303292262e-06, 'epoch': 1.82}) +[2026-06-16 11:29:50,837][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.07037397474050522, 'train/grad_norm_pre_clip': 31.59803581237793, 'timing/data': 0.021474513690918684, 'timing/model': 1.5504563059657812, 'learning_rate/qwen_vl_interface': 7.381614811913538e-07, 'learning_rate/action_model': 7.381614811913537e-06, 'epoch': 1.82}) +[2026-06-16 11:29:57,029][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.0011007495922967792, 'train/grad_norm_pre_clip': 0.32599297165870667, 'timing/data': 0.02136028278619051, 'timing/model': 1.56933604599908, 'learning_rate/qwen_vl_interface': 7.372101866998693e-07, 'learning_rate/action_model': 7.372101866998693e-06, 'epoch': 1.82}) +[2026-06-16 11:30:03,296][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.12976102530956268, 'train/grad_norm_pre_clip': 19.168224334716797, 'timing/data': 0.020700988825410604, 'timing/model': 1.8089763610623777, 'learning_rate/qwen_vl_interface': 7.362607472458147e-07, 'learning_rate/action_model': 7.362607472458147e-06, 'epoch': 1.82}) +[2026-06-16 11:30:09,354][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.0007801791653037071, 'train/grad_norm_pre_clip': 0.1932855099439621, 'timing/data': 0.021665093954652548, 'timing/model': 1.5595870460383594, 'learning_rate/qwen_vl_interface': 7.353131632194682e-07, 'learning_rate/action_model': 7.3531316321946816e-06, 'epoch': 1.82}) +[2026-06-16 11:30:15,653][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.05271170660853386, 'train/grad_norm_pre_clip': 31.513120651245117, 'timing/data': 0.0221549472771585, 'timing/model': 1.5550387720577419, 'learning_rate/qwen_vl_interface': 7.343674350103455e-07, 'learning_rate/action_model': 7.343674350103455e-06, 'epoch': 1.82}) +[2026-06-16 11:30:21,952][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.04416719451546669, 'train/grad_norm_pre_clip': 23.647859573364258, 'timing/data': 0.022217700257897377, 'timing/model': 1.548063388094306, 'learning_rate/qwen_vl_interface': 7.334235630071995e-07, 'learning_rate/action_model': 7.334235630071995e-06, 'epoch': 1.82}) +[2026-06-16 11:30:28,357][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.0465971864759922, 'train/grad_norm_pre_clip': 97.47110748291016, 'timing/data': 0.025217453949153423, 'timing/model': 1.5950067942030728, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 1.82}) +[2026-06-16 11:30:34,725][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 0.02900751866400242, 'train/grad_norm_pre_clip': 27.5152530670166, 'timing/data': 0.020652583334594965, 'timing/model': 1.8222450651228428, 'learning_rate/qwen_vl_interface': 7.315413891700363e-07, 'learning_rate/action_model': 7.315413891700362e-06, 'epoch': 1.82}) +[2026-06-16 11:30:40,964][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.0031272652558982372, 'train/grad_norm_pre_clip': 1.5778319835662842, 'timing/data': 0.02179943909868598, 'timing/model': 1.7601773492060602, 'learning_rate/qwen_vl_interface': 7.306030881097092e-07, 'learning_rate/action_model': 7.3060308810970925e-06, 'epoch': 1.82}) +[2026-06-16 11:30:47,054][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.0031159687787294388, 'train/grad_norm_pre_clip': 2.2677834033966064, 'timing/data': 0.021345356944948435, 'timing/model': 1.5471206968650222, 'learning_rate/qwen_vl_interface': 7.296666448027406e-07, 'learning_rate/action_model': 7.296666448027406e-06, 'epoch': 1.83}) +[2026-06-16 11:30:53,355][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.002253566635772586, 'train/grad_norm_pre_clip': 1.2572152614593506, 'timing/data': 0.020634453743696213, 'timing/model': 1.5558934290893376, 'learning_rate/qwen_vl_interface': 7.287320596340664e-07, 'learning_rate/action_model': 7.287320596340663e-06, 'epoch': 1.83}) +[2026-06-16 11:30:59,658][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.11579495668411255, 'train/grad_norm_pre_clip': 11.973824501037598, 'timing/data': 0.02087649516761303, 'timing/model': 1.5274824909865856, 'learning_rate/qwen_vl_interface': 7.277993329878593e-07, 'learning_rate/action_model': 7.277993329878594e-06, 'epoch': 1.83}) +[2026-06-16 11:31:06,081][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.2985726296901703, 'train/grad_norm_pre_clip': 24.953989028930664, 'timing/data': 0.020618894137442112, 'timing/model': 1.59033437166363, 'learning_rate/qwen_vl_interface': 7.268684652475277e-07, 'learning_rate/action_model': 7.268684652475277e-06, 'epoch': 1.83}) +[2026-06-16 11:31:12,458][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.08132065087556839, 'train/grad_norm_pre_clip': 34.037845611572266, 'timing/data': 0.021190527360886335, 'timing/model': 1.586628207936883, 'learning_rate/qwen_vl_interface': 7.259394567957169e-07, 'learning_rate/action_model': 7.259394567957169e-06, 'epoch': 1.83}) +[2026-06-16 11:31:18,887][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.0003178283222950995, 'train/grad_norm_pre_clip': 0.05244377255439758, 'timing/data': 0.02046111412346363, 'timing/model': 1.8700460861437023, 'learning_rate/qwen_vl_interface': 7.250123080143066e-07, 'learning_rate/action_model': 7.2501230801430665e-06, 'epoch': 1.83}) +[2026-06-16 11:31:25,054][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.0010800062445923686, 'train/grad_norm_pre_clip': 0.24948514997959137, 'timing/data': 0.020996006205677986, 'timing/model': 1.623177650384605, 'learning_rate/qwen_vl_interface': 7.240870192844127e-07, 'learning_rate/action_model': 7.240870192844127e-06, 'epoch': 1.83}) +[2026-06-16 11:31:31,471][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.002397902077063918, 'train/grad_norm_pre_clip': 2.187300443649292, 'timing/data': 0.02443571202456951, 'timing/model': 1.5739366868510842, 'learning_rate/qwen_vl_interface': 7.231635909863858e-07, 'learning_rate/action_model': 7.2316359098638585e-06, 'epoch': 1.83}) +[2026-06-16 11:31:37,868][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 0.06691611558198929, 'train/grad_norm_pre_clip': 38.53755569458008, 'timing/data': 0.02067434461787343, 'timing/model': 1.5590090998448431, 'learning_rate/qwen_vl_interface': 7.222420234998138e-07, 'learning_rate/action_model': 7.2224202349981385e-06, 'epoch': 1.83}) +[2026-06-16 11:31:44,259][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.0011514503275975585, 'train/grad_norm_pre_clip': 0.5176029801368713, 'timing/data': 0.02169707976281643, 'timing/model': 1.5829294021241367, 'learning_rate/qwen_vl_interface': 7.213223172035175e-07, 'learning_rate/action_model': 7.213223172035174e-06, 'epoch': 1.83}) +[2026-06-16 11:31:50,682][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.11360055208206177, 'train/grad_norm_pre_clip': 23.045209884643555, 'timing/data': 0.021618609316647053, 'timing/model': 1.8725637500174344, 'learning_rate/qwen_vl_interface': 7.204044724755528e-07, 'learning_rate/action_model': 7.204044724755528e-06, 'epoch': 1.83}) +[2026-06-16 11:31:56,765][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.0026495326310396194, 'train/grad_norm_pre_clip': 3.6181976795196533, 'timing/data': 0.02057866519317031, 'timing/model': 1.5179456137120724, 'learning_rate/qwen_vl_interface': 7.194884896932132e-07, 'learning_rate/action_model': 7.194884896932132e-06, 'epoch': 1.83}) +[2026-06-16 11:32:03,061][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.001778084202669561, 'train/grad_norm_pre_clip': 0.36118489503860474, 'timing/data': 0.019795960746705532, 'timing/model': 1.5495527517050505, 'learning_rate/qwen_vl_interface': 7.185743692330221e-07, 'learning_rate/action_model': 7.185743692330221e-06, 'epoch': 1.83}) +[2026-06-16 11:32:09,259][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 0.0015477644046768546, 'train/grad_norm_pre_clip': 0.37510743737220764, 'timing/data': 0.020801916252821684, 'timing/model': 1.5472036022692919, 'learning_rate/qwen_vl_interface': 7.176621114707408e-07, 'learning_rate/action_model': 7.176621114707407e-06, 'epoch': 1.83}) +[2026-06-16 11:32:15,567][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.00012035577674396336, 'train/grad_norm_pre_clip': 0.012685881927609444, 'timing/data': 0.02129364898428321, 'timing/model': 1.5509333098307252, 'learning_rate/qwen_vl_interface': 7.167517167813642e-07, 'learning_rate/action_model': 7.167517167813642e-06, 'epoch': 1.83}) +[2026-06-16 11:32:21,872][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.13213317096233368, 'train/grad_norm_pre_clip': 29.837095260620117, 'timing/data': 0.020371634978801012, 'timing/model': 1.5242729671299458, 'learning_rate/qwen_vl_interface': 7.15843185539121e-07, 'learning_rate/action_model': 7.15843185539121e-06, 'epoch': 1.83}) +[2026-06-16 11:32:28,327][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.00039576386916451156, 'train/grad_norm_pre_clip': 0.1272633969783783, 'timing/data': 0.030675882939249277, 'timing/model': 1.836918513290584, 'learning_rate/qwen_vl_interface': 7.149365181174739e-07, 'learning_rate/action_model': 7.149365181174739e-06, 'epoch': 1.83}) +[2026-06-16 11:32:34,369][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.0002070838090730831, 'train/grad_norm_pre_clip': 0.028291331604123116, 'timing/data': 0.03180488804355264, 'timing/model': 1.5473437048494816, 'learning_rate/qwen_vl_interface': 7.14031714889121e-07, 'learning_rate/action_model': 7.14031714889121e-06, 'epoch': 1.83}) +[2026-06-16 11:32:40,863][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.0009392091888003051, 'train/grad_norm_pre_clip': 0.15896032750606537, 'timing/data': 0.03462142124772072, 'timing/model': 1.5339658930897713, 'learning_rate/qwen_vl_interface': 7.131287762259907e-07, 'learning_rate/action_model': 7.131287762259907e-06, 'epoch': 1.83}) +[2026-06-16 11:32:47,165][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.00631756754592061, 'train/grad_norm_pre_clip': 5.400765895843506, 'timing/data': 0.031194965820759535, 'timing/model': 1.5460722702555358, 'learning_rate/qwen_vl_interface': 7.12227702499248e-07, 'learning_rate/action_model': 7.12227702499248e-06, 'epoch': 1.83}) +[2026-06-16 11:32:53,563][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.000760417606215924, 'train/grad_norm_pre_clip': 0.21572327613830566, 'timing/data': 0.029084445908665657, 'timing/model': 1.5532681439071894, 'learning_rate/qwen_vl_interface': 7.113284940792903e-07, 'learning_rate/action_model': 7.113284940792903e-06, 'epoch': 1.83}) +[2026-06-16 11:32:59,955][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.2252388745546341, 'train/grad_norm_pre_clip': 28.37677001953125, 'timing/data': 0.03007112117484212, 'timing/model': 1.524060966912657, 'learning_rate/qwen_vl_interface': 7.104311513357476e-07, 'learning_rate/action_model': 7.1043115133574755e-06, 'epoch': 1.83}) +[2026-06-16 11:33:06,441][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.12807519733905792, 'train/grad_norm_pre_clip': 26.6341609954834, 'timing/data': 0.03957286570221186, 'timing/model': 1.8939690329134464, 'learning_rate/qwen_vl_interface': 7.095356746374845e-07, 'learning_rate/action_model': 7.095356746374844e-06, 'epoch': 1.83}) +[2026-06-16 11:33:12,467][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.0007310595829039812, 'train/grad_norm_pre_clip': 0.1799246072769165, 'timing/data': 0.030529854353517294, 'timing/model': 1.5456689079292119, 'learning_rate/qwen_vl_interface': 7.086420643525973e-07, 'learning_rate/action_model': 7.086420643525972e-06, 'epoch': 1.83}) +[2026-06-16 11:33:18,768][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.053544554859399796, 'train/grad_norm_pre_clip': 40.24435043334961, 'timing/data': 0.031635778956115246, 'timing/model': 1.5351492380723357, 'learning_rate/qwen_vl_interface': 7.077503208484151e-07, 'learning_rate/action_model': 7.0775032084841515e-06, 'epoch': 1.83}) +[2026-06-16 11:33:25,175][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.253248929977417, 'train/grad_norm_pre_clip': 37.88954544067383, 'timing/data': 0.03192747989669442, 'timing/model': 1.5416251881979406, 'learning_rate/qwen_vl_interface': 7.068604444915003e-07, 'learning_rate/action_model': 7.068604444915003e-06, 'epoch': 1.84}) +[2026-06-16 11:33:31,577][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.04786057770252228, 'train/grad_norm_pre_clip': 39.25312042236328, 'timing/data': 0.029531630221754313, 'timing/model': 1.6042501432821155, 'learning_rate/qwen_vl_interface': 7.059724356476475e-07, 'learning_rate/action_model': 7.059724356476474e-06, 'epoch': 1.84}) +[2026-06-16 11:33:37,879][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.0009395164088346064, 'train/grad_norm_pre_clip': 0.5407264232635498, 'timing/data': 0.030829485040158033, 'timing/model': 1.6029924470931292, 'learning_rate/qwen_vl_interface': 7.050862946818836e-07, 'learning_rate/action_model': 7.050862946818836e-06, 'epoch': 1.84}) +[2026-06-16 11:33:44,225][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.0010051996214315295, 'train/grad_norm_pre_clip': 0.2172759473323822, 'timing/data': 0.03160499036312103, 'timing/model': 1.8056925651617348, 'learning_rate/qwen_vl_interface': 7.042020219584673e-07, 'learning_rate/action_model': 7.042020219584673e-06, 'epoch': 1.84}) +[2026-06-16 11:33:50,274][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.0006281841197051108, 'train/grad_norm_pre_clip': 0.16598933935165405, 'timing/data': 0.02886719675734639, 'timing/model': 1.54945264197886, 'learning_rate/qwen_vl_interface': 7.0331961784089e-07, 'learning_rate/action_model': 7.033196178408899e-06, 'epoch': 1.84}) +[2026-06-16 11:33:56,687][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.008621984161436558, 'train/grad_norm_pre_clip': 10.859487533569336, 'timing/data': 0.03472983418032527, 'timing/model': 1.5511085586622357, 'learning_rate/qwen_vl_interface': 7.024390826918748e-07, 'learning_rate/action_model': 7.024390826918747e-06, 'epoch': 1.84}) +[2026-06-16 11:34:03,112][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.006567257922142744, 'train/grad_norm_pre_clip': 2.7623329162597656, 'timing/data': 0.03456673538312316, 'timing/model': 1.5624899407848716, 'learning_rate/qwen_vl_interface': 7.01560416873376e-07, 'learning_rate/action_model': 7.015604168733759e-06, 'epoch': 1.84}) +[2026-06-16 11:34:09,388][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.0002039257378783077, 'train/grad_norm_pre_clip': 0.12155387550592422, 'timing/data': 0.03039705380797386, 'timing/model': 1.563187263906002, 'learning_rate/qwen_vl_interface': 7.006836207465803e-07, 'learning_rate/action_model': 7.006836207465803e-06, 'epoch': 1.84}) +[2026-06-16 11:34:15,857][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 0.27538347244262695, 'train/grad_norm_pre_clip': 35.67163848876953, 'timing/data': 0.03086120681837201, 'timing/model': 1.5978793278336525, 'learning_rate/qwen_vl_interface': 6.998086946719058e-07, 'learning_rate/action_model': 6.998086946719057e-06, 'epoch': 1.84}) +[2026-06-16 11:34:22,363][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.0006905152113176882, 'train/grad_norm_pre_clip': 0.3455328941345215, 'timing/data': 0.031236029230058193, 'timing/model': 1.8391426466405392, 'learning_rate/qwen_vl_interface': 6.989356390090008e-07, 'learning_rate/action_model': 6.989356390090007e-06, 'epoch': 1.84}) +[2026-06-16 11:34:28,469][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.0013521261280402541, 'train/grad_norm_pre_clip': 0.46616876125335693, 'timing/data': 0.03123852889984846, 'timing/model': 1.537369676399976, 'learning_rate/qwen_vl_interface': 6.980644541167459e-07, 'learning_rate/action_model': 6.980644541167458e-06, 'epoch': 1.84}) +[2026-06-16 11:34:34,963][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.0008797912742011249, 'train/grad_norm_pre_clip': 0.31246450543403625, 'timing/data': 0.03084462694823742, 'timing/model': 1.5366148110479116, 'learning_rate/qwen_vl_interface': 6.971951403532524e-07, 'learning_rate/action_model': 6.971951403532523e-06, 'epoch': 1.84}) +[2026-06-16 11:34:41,366][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.15564501285552979, 'train/grad_norm_pre_clip': 29.166399002075195, 'timing/data': 0.030609247274696827, 'timing/model': 1.549571461044252, 'learning_rate/qwen_vl_interface': 6.963276980758622e-07, 'learning_rate/action_model': 6.963276980758622e-06, 'epoch': 1.84}) +[2026-06-16 11:34:47,660][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.02428564615547657, 'train/grad_norm_pre_clip': 16.468368530273438, 'timing/data': 0.03064083680510521, 'timing/model': 1.5247750412672758, 'learning_rate/qwen_vl_interface': 6.954621276411485e-07, 'learning_rate/action_model': 6.954621276411484e-06, 'epoch': 1.84}) +[2026-06-16 11:34:54,065][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 0.15951578319072723, 'train/grad_norm_pre_clip': 7.193077087402344, 'timing/data': 0.03132485086098313, 'timing/model': 1.572636139113456, 'learning_rate/qwen_vl_interface': 6.945984294049136e-07, 'learning_rate/action_model': 6.9459842940491364e-06, 'epoch': 1.84}) +[2026-06-16 11:35:00,431][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.23023395240306854, 'train/grad_norm_pre_clip': 9.159733772277832, 'timing/data': 0.030435557011514902, 'timing/model': 1.8005294590257108, 'learning_rate/qwen_vl_interface': 6.937366037221919e-07, 'learning_rate/action_model': 6.937366037221918e-06, 'epoch': 1.84}) +[2026-06-16 11:35:06,470][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.0010030217235907912, 'train/grad_norm_pre_clip': 0.24892829358577728, 'timing/data': 0.031650782097131014, 'timing/model': 1.5386270708404481, 'learning_rate/qwen_vl_interface': 6.928766509472471e-07, 'learning_rate/action_model': 6.92876650947247e-06, 'epoch': 1.84}) +[2026-06-16 11:35:12,773][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 0.004310190211981535, 'train/grad_norm_pre_clip': 4.852665901184082, 'timing/data': 0.031362590845674276, 'timing/model': 1.54415971506387, 'learning_rate/qwen_vl_interface': 6.920185714335737e-07, 'learning_rate/action_model': 6.9201857143357375e-06, 'epoch': 1.84}) +[2026-06-16 11:35:19,169][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.1765957474708557, 'train/grad_norm_pre_clip': 11.994906425476074, 'timing/data': 0.03160576242953539, 'timing/model': 1.5422776844352484, 'learning_rate/qwen_vl_interface': 6.911623655338959e-07, 'learning_rate/action_model': 6.911623655338959e-06, 'epoch': 1.84}) +[2026-06-16 11:35:25,576][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.00028437431319616735, 'train/grad_norm_pre_clip': 0.040877580642700195, 'timing/data': 0.03129297122359276, 'timing/model': 1.5636385739780962, 'learning_rate/qwen_vl_interface': 6.903080336001666e-07, 'learning_rate/action_model': 6.903080336001666e-06, 'epoch': 1.84}) +[2026-06-16 11:35:31,978][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 6.632594886468723e-05, 'train/grad_norm_pre_clip': 0.009011607617139816, 'timing/data': 0.03177084308117628, 'timing/model': 1.5841819802299142, 'learning_rate/qwen_vl_interface': 6.894555759835707e-07, 'learning_rate/action_model': 6.894555759835707e-06, 'epoch': 1.84}) +[2026-06-16 11:35:38,333][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.013314129784703255, 'train/grad_norm_pre_clip': 11.450091361999512, 'timing/data': 0.030917108990252018, 'timing/model': 1.8100818190723658, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 1.84}) +[2026-06-16 11:35:44,377][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.00019031377451028675, 'train/grad_norm_pre_clip': 0.028055602684617043, 'timing/data': 0.02966862265020609, 'timing/model': 1.5467117452062666, 'learning_rate/qwen_vl_interface': 6.877562851026584e-07, 'learning_rate/action_model': 6.877562851026584e-06, 'epoch': 1.84}) +[2026-06-16 11:35:50,775][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.000742787029594183, 'train/grad_norm_pre_clip': 0.5364890694618225, 'timing/data': 0.030737799126654863, 'timing/model': 1.5305028432048857, 'learning_rate/qwen_vl_interface': 6.869094525368572e-07, 'learning_rate/action_model': 6.8690945253685715e-06, 'epoch': 1.84}) +[2026-06-16 11:35:57,090][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.0005530552007257938, 'train/grad_norm_pre_clip': 0.32952994108200073, 'timing/data': 0.029677844140678644, 'timing/model': 1.543647853191942, 'learning_rate/qwen_vl_interface': 6.860644956852168e-07, 'learning_rate/action_model': 6.860644956852168e-06, 'epoch': 1.84}) +[2026-06-16 11:36:03,474][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.0003244566614739597, 'train/grad_norm_pre_clip': 0.10165691375732422, 'timing/data': 0.029799323994666338, 'timing/model': 1.5551479472778738, 'learning_rate/qwen_vl_interface': 6.852214148950667e-07, 'learning_rate/action_model': 6.852214148950667e-06, 'epoch': 1.85}) +[2026-06-16 11:36:09,866][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.4803226888179779, 'train/grad_norm_pre_clip': 51.132362365722656, 'timing/data': 0.029287601821124554, 'timing/model': 1.5949895470403135, 'learning_rate/qwen_vl_interface': 6.843802105129667e-07, 'learning_rate/action_model': 6.843802105129666e-06, 'epoch': 1.85}) +[2026-06-16 11:36:16,245][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.0006970029789954424, 'train/grad_norm_pre_clip': 0.13664820790290833, 'timing/data': 0.0321932858787477, 'timing/model': 1.8085255711339414, 'learning_rate/qwen_vl_interface': 6.835408828847035e-07, 'learning_rate/action_model': 6.8354088288470356e-06, 'epoch': 1.85}) +[2026-06-16 11:36:22,443][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.0003397814289201051, 'train/grad_norm_pre_clip': 0.07849794626235962, 'timing/data': 0.03017452685162425, 'timing/model': 1.7050127852708101, 'learning_rate/qwen_vl_interface': 6.827034323552929e-07, 'learning_rate/action_model': 6.827034323552929e-06, 'epoch': 1.85}) +[2026-06-16 11:36:28,471][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.00160691374912858, 'train/grad_norm_pre_clip': 0.6764941215515137, 'timing/data': 0.02927507320418954, 'timing/model': 1.537845069076866, 'learning_rate/qwen_vl_interface': 6.818678592689794e-07, 'learning_rate/action_model': 6.818678592689794e-06, 'epoch': 1.85}) +[2026-06-16 11:36:34,771][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.0023333157878369093, 'train/grad_norm_pre_clip': 1.6063185930252075, 'timing/data': 0.031407253351062536, 'timing/model': 1.5361483981832862, 'learning_rate/qwen_vl_interface': 6.810341639692357e-07, 'learning_rate/action_model': 6.810341639692357e-06, 'epoch': 1.85}) +[2026-06-16 11:36:41,171][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 0.00014469894813373685, 'train/grad_norm_pre_clip': 0.02158743515610695, 'timing/data': 0.03031005710363388, 'timing/model': 1.530665148049593, 'learning_rate/qwen_vl_interface': 6.802023467987631e-07, 'learning_rate/action_model': 6.802023467987631e-06, 'epoch': 1.85}) +[2026-06-16 11:36:47,575][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.36771160364151, 'train/grad_norm_pre_clip': 23.22235679626465, 'timing/data': 0.030418417882174253, 'timing/model': 1.530388182029128, 'learning_rate/qwen_vl_interface': 6.793724080994884e-07, 'learning_rate/action_model': 6.793724080994884e-06, 'epoch': 1.85}) +[2026-06-16 11:36:54,020][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.0016084996750578284, 'train/grad_norm_pre_clip': 0.8942729234695435, 'timing/data': 0.029898486100137234, 'timing/model': 1.8525466811843216, 'learning_rate/qwen_vl_interface': 6.785443482125693e-07, 'learning_rate/action_model': 6.7854434821256915e-06, 'epoch': 1.85}) +[2026-06-16 11:37:00,349][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.00038029273855499923, 'train/grad_norm_pre_clip': 0.12213923037052155, 'timing/data': 0.03189973579719663, 'timing/model': 1.8108478640206158, 'learning_rate/qwen_vl_interface': 6.777181674783895e-07, 'learning_rate/action_model': 6.777181674783895e-06, 'epoch': 1.85}) +[2026-06-16 11:37:06,429][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.0006629479466937482, 'train/grad_norm_pre_clip': 1.2029950618743896, 'timing/data': 0.03143344400450587, 'timing/model': 1.5811275281012058, 'learning_rate/qwen_vl_interface': 6.76893866236561e-07, 'learning_rate/action_model': 6.7689386623656094e-06, 'epoch': 1.85}) +[2026-06-16 11:37:12,781][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.0624174065887928, 'train/grad_norm_pre_clip': 20.486852645874023, 'timing/data': 0.030467869248241186, 'timing/model': 1.5568556259386241, 'learning_rate/qwen_vl_interface': 6.76071444825923e-07, 'learning_rate/action_model': 6.76071444825923e-06, 'epoch': 1.85}) +[2026-06-16 11:37:19,285][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 0.00013698116526938975, 'train/grad_norm_pre_clip': 0.016154266893863678, 'timing/data': 0.032078546937555075, 'timing/model': 1.5779426558874547, 'learning_rate/qwen_vl_interface': 6.752509035845413e-07, 'learning_rate/action_model': 6.7525090358454126e-06, 'epoch': 1.85}) +[2026-06-16 11:37:25,672][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.14945247769355774, 'train/grad_norm_pre_clip': 17.256826400756836, 'timing/data': 0.031026911921799183, 'timing/model': 1.5728969778865576, 'learning_rate/qwen_vl_interface': 6.744322428497095e-07, 'learning_rate/action_model': 6.744322428497095e-06, 'epoch': 1.85}) +[2026-06-16 11:37:32,057][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.02179880626499653, 'train/grad_norm_pre_clip': 10.217412948608398, 'timing/data': 0.0315805752761662, 'timing/model': 1.82777030300349, 'learning_rate/qwen_vl_interface': 6.736154629579487e-07, 'learning_rate/action_model': 6.736154629579487e-06, 'epoch': 1.85}) +[2026-06-16 11:37:38,180][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.062361039221286774, 'train/grad_norm_pre_clip': 28.354095458984375, 'timing/data': 0.030458549968898296, 'timing/model': 1.5533540332689881, 'learning_rate/qwen_vl_interface': 6.728005642450055e-07, 'learning_rate/action_model': 6.7280056424500554e-06, 'epoch': 1.85}) +[2026-06-16 11:37:44,575][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 0.0008413003524765372, 'train/grad_norm_pre_clip': 0.28344979882240295, 'timing/data': 0.031924257054924965, 'timing/model': 1.5500193131156266, 'learning_rate/qwen_vl_interface': 6.719875470458553e-07, 'learning_rate/action_model': 6.719875470458553e-06, 'epoch': 1.85}) +[2026-06-16 11:37:50,967][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.16535121202468872, 'train/grad_norm_pre_clip': 30.99062728881836, 'timing/data': 0.03231391776353121, 'timing/model': 1.539238405879587, 'learning_rate/qwen_vl_interface': 6.711764116946972e-07, 'learning_rate/action_model': 6.7117641169469715e-06, 'epoch': 1.85}) +[2026-06-16 11:37:57,365][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.08587907254695892, 'train/grad_norm_pre_clip': 29.05118179321289, 'timing/data': 0.03157979529350996, 'timing/model': 1.5510819791816175, 'learning_rate/qwen_vl_interface': 6.703671585249587e-07, 'learning_rate/action_model': 6.703671585249586e-06, 'epoch': 1.85}) +[2026-06-16 11:38:03,758][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.0025116363540291786, 'train/grad_norm_pre_clip': 0.7835703492164612, 'timing/data': 0.03141874494031072, 'timing/model': 1.5916563039645553, 'learning_rate/qwen_vl_interface': 6.69559787869294e-07, 'learning_rate/action_model': 6.69559787869294e-06, 'epoch': 1.85}) +[2026-06-16 11:38:10,096][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.00012876525579486042, 'train/grad_norm_pre_clip': 0.01836242713034153, 'timing/data': 0.030116449110209942, 'timing/model': 1.7707424024119973, 'learning_rate/qwen_vl_interface': 6.68754300059582e-07, 'learning_rate/action_model': 6.687543000595819e-06, 'epoch': 1.85}) +[2026-06-16 11:38:16,168][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.419191837310791, 'train/grad_norm_pre_clip': 25.715394973754883, 'timing/data': 0.03032153006643057, 'timing/model': 1.5422568824142218, 'learning_rate/qwen_vl_interface': 6.67950695426929e-07, 'learning_rate/action_model': 6.679506954269291e-06, 'epoch': 1.85}) +[2026-06-16 11:38:22,467][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.0009325255523435771, 'train/grad_norm_pre_clip': 0.1407196819782257, 'timing/data': 0.030625680927187204, 'timing/model': 1.5386693989858031, 'learning_rate/qwen_vl_interface': 6.671489743016651e-07, 'learning_rate/action_model': 6.671489743016651e-06, 'epoch': 1.85}) +[2026-06-16 11:38:28,870][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.0013996560592204332, 'train/grad_norm_pre_clip': 1.0484775304794312, 'timing/data': 0.03136610519140959, 'timing/model': 1.5680148391984403, 'learning_rate/qwen_vl_interface': 6.663491370133487e-07, 'learning_rate/action_model': 6.663491370133487e-06, 'epoch': 1.85}) +[2026-06-16 11:38:35,280][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.0016183566767722368, 'train/grad_norm_pre_clip': 0.7482594847679138, 'timing/data': 0.03213780885562301, 'timing/model': 1.6027590483427048, 'learning_rate/qwen_vl_interface': 6.655511838907629e-07, 'learning_rate/action_model': 6.655511838907629e-06, 'epoch': 1.85}) +[2026-06-16 11:38:41,848][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.0031031016260385513, 'train/grad_norm_pre_clip': 0.7024662494659424, 'timing/data': 0.03076785197481513, 'timing/model': 1.550395314116031, 'learning_rate/qwen_vl_interface': 6.647551152619149e-07, 'learning_rate/action_model': 6.647551152619148e-06, 'epoch': 1.86}) +[2026-06-16 11:38:48,183][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.00033997040009126067, 'train/grad_norm_pre_clip': 0.04325730353593826, 'timing/data': 0.03303599311038852, 'timing/model': 1.75469174887985, 'learning_rate/qwen_vl_interface': 6.639609314540396e-07, 'learning_rate/action_model': 6.639609314540396e-06, 'epoch': 1.86}) +[2026-06-16 11:38:54,265][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.12513552606105804, 'train/grad_norm_pre_clip': 11.583959579467773, 'timing/data': 0.03159084729850292, 'timing/model': 1.5364250936545432, 'learning_rate/qwen_vl_interface': 6.631686327935954e-07, 'learning_rate/action_model': 6.631686327935954e-06, 'epoch': 1.86}) +[2026-06-16 11:39:00,560][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.005445228889584541, 'train/grad_norm_pre_clip': 7.969403266906738, 'timing/data': 0.03132804483175278, 'timing/model': 1.5304456502199173, 'learning_rate/qwen_vl_interface': 6.623782196062665e-07, 'learning_rate/action_model': 6.623782196062664e-06, 'epoch': 1.86}) +[2026-06-16 11:39:06,976][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.0008832828025333583, 'train/grad_norm_pre_clip': 0.541877269744873, 'timing/data': 0.03208915889263153, 'timing/model': 1.53064100490883, 'learning_rate/qwen_vl_interface': 6.615896922169622e-07, 'learning_rate/action_model': 6.615896922169622e-06, 'epoch': 1.86}) +[2026-06-16 11:39:13,469][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.002326228655874729, 'train/grad_norm_pre_clip': 1.4717034101486206, 'timing/data': 0.031196414958685637, 'timing/model': 1.5739597142674029, 'learning_rate/qwen_vl_interface': 6.608030509498156e-07, 'learning_rate/action_model': 6.608030509498156e-06, 'epoch': 1.86}) +[2026-06-16 11:39:19,836][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.000453394022770226, 'train/grad_norm_pre_clip': 0.1699068695306778, 'timing/data': 0.030957352835685015, 'timing/model': 1.8120071510784328, 'learning_rate/qwen_vl_interface': 6.600182961281865e-07, 'learning_rate/action_model': 6.600182961281864e-06, 'epoch': 1.86}) +[2026-06-16 11:39:25,859][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.2722264230251312, 'train/grad_norm_pre_clip': 43.49551010131836, 'timing/data': 0.03136572614312172, 'timing/model': 1.543562503065914, 'learning_rate/qwen_vl_interface': 6.592354280746561e-07, 'learning_rate/action_model': 6.5923542807465595e-06, 'epoch': 1.86}) +[2026-06-16 11:39:32,172][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 0.031016139313578606, 'train/grad_norm_pre_clip': 22.527158737182617, 'timing/data': 0.029107084963470697, 'timing/model': 1.5442234710790217, 'learning_rate/qwen_vl_interface': 6.584544471110335e-07, 'learning_rate/action_model': 6.584544471110335e-06, 'epoch': 1.86}) +[2026-06-16 11:39:38,464][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.06817341595888138, 'train/grad_norm_pre_clip': 18.45024299621582, 'timing/data': 0.03054745215922594, 'timing/model': 1.603049328085035, 'learning_rate/qwen_vl_interface': 6.576753535583495e-07, 'learning_rate/action_model': 6.576753535583495e-06, 'epoch': 1.86}) +[2026-06-16 11:39:44,872][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.025104230269789696, 'train/grad_norm_pre_clip': 28.090456008911133, 'timing/data': 0.03272966295480728, 'timing/model': 1.6195549853146076, 'learning_rate/qwen_vl_interface': 6.5689814773686e-07, 'learning_rate/action_model': 6.568981477368599e-06, 'epoch': 1.86}) +[2026-06-16 11:39:51,169][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.02909672074019909, 'train/grad_norm_pre_clip': 28.002193450927734, 'timing/data': 0.03054469171911478, 'timing/model': 1.5310400994494557, 'learning_rate/qwen_vl_interface': 6.56122829966046e-07, 'learning_rate/action_model': 6.56122829966046e-06, 'epoch': 1.86}) +[2026-06-16 11:50:34,516][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.28036952018737793, 'train/grad_norm_pre_clip': 23.704381942749023, 'eval/loss': 0.034478917717933655, 'eval/flap_precision': 0.9689608636977058, 'eval/flap_recall': 0.9795361527967258, 'eval/flap_f1': 0.9742198100407056, 'eval/flap_rate_model': 0.20583333333333334, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9742198100407056, 'timing/data': 0.03217240981757641, 'timing/model': 1.8199072424322367, 'learning_rate/qwen_vl_interface': 6.553494005646101e-07, 'learning_rate/action_model': 6.553494005646101e-06, 'epoch': 1.86}) +[2026-06-16 11:50:34,517][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4600_state +[2026-06-16 11:50:34,524][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 11:50:42,894][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4600_state/model.safetensors +[2026-06-16 11:51:02,806][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4600_state/optimizer.bin +[2026-06-16 11:51:02,807][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4600_state/sampler.bin +[2026-06-16 11:51:02,807][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4600_state/sampler_1.bin +[2026-06-16 11:51:02,808][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4600_state/random_states_0.pkl +[2026-06-16 11:51:02,998][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4600 +[2026-06-16 11:51:02,999][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 11:51:03,013][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 11:51:09,271][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.0077674416825175285, 'train/grad_norm_pre_clip': 6.1389007568359375, 'timing/data': 0.03094367915764451, 'timing/model': 1.7847060970962048, 'learning_rate/qwen_vl_interface': 6.545778598504807e-07, 'learning_rate/action_model': 6.545778598504807e-06, 'epoch': 1.86}) +[2026-06-16 11:51:15,636][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.00044530394370667636, 'train/grad_norm_pre_clip': 0.11540635675191879, 'timing/data': 0.0357377240434289, 'timing/model': 1.7667758059687912, 'learning_rate/qwen_vl_interface': 6.538082081408093e-07, 'learning_rate/action_model': 6.538082081408093e-06, 'epoch': 1.86}) +[2026-06-16 11:51:21,653][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 0.08769289404153824, 'train/grad_norm_pre_clip': 29.1275691986084, 'timing/data': 0.030087974853813648, 'timing/model': 1.5509207653813064, 'learning_rate/qwen_vl_interface': 6.530404457519697e-07, 'learning_rate/action_model': 6.530404457519697e-06, 'epoch': 1.86}) +[2026-06-16 11:51:27,846][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.0017673058900982141, 'train/grad_norm_pre_clip': 0.5413691401481628, 'timing/data': 0.031175870914012194, 'timing/model': 1.5428714752197266, 'learning_rate/qwen_vl_interface': 6.522745729995614e-07, 'learning_rate/action_model': 6.522745729995613e-06, 'epoch': 1.86}) +[2026-06-16 11:51:34,046][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.01194524671882391, 'train/grad_norm_pre_clip': 11.260873794555664, 'timing/data': 0.03017834573984146, 'timing/model': 1.5448341760784388, 'learning_rate/qwen_vl_interface': 6.515105901984037e-07, 'learning_rate/action_model': 6.515105901984037e-06, 'epoch': 1.86}) +[2026-06-16 11:51:40,342][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.070371612906456, 'train/grad_norm_pre_clip': 50.48320770263672, 'timing/data': 0.031415962148457766, 'timing/model': 1.5286190039478242, 'learning_rate/qwen_vl_interface': 6.50748497662543e-07, 'learning_rate/action_model': 6.507484976625429e-06, 'epoch': 1.86}) +[2026-06-16 11:51:46,591][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.01164907868951559, 'train/grad_norm_pre_clip': 9.727453231811523, 'timing/data': 0.03145925188437104, 'timing/model': 1.7810569945722818, 'learning_rate/qwen_vl_interface': 6.499882957052459e-07, 'learning_rate/action_model': 6.4998829570524586e-06, 'epoch': 1.86}) +[2026-06-16 11:51:52,885][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.07428310066461563, 'train/grad_norm_pre_clip': 48.08700180053711, 'timing/data': 0.032223026268184185, 'timing/model': 1.7760941213928163, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 1.86}) +[2026-06-16 11:51:58,949][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.00015369878383353353, 'train/grad_norm_pre_clip': 0.035469185560941696, 'timing/data': 0.03072086814790964, 'timing/model': 1.5311578107066453, 'learning_rate/qwen_vl_interface': 6.484735647755259e-07, 'learning_rate/action_model': 6.48473564775526e-06, 'epoch': 1.86}) +[2026-06-16 11:52:05,248][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.0006090163951739669, 'train/grad_norm_pre_clip': 0.11743300408124924, 'timing/data': 0.02932562120258808, 'timing/model': 1.5493492647074163, 'learning_rate/qwen_vl_interface': 6.477190364257524e-07, 'learning_rate/action_model': 6.477190364257524e-06, 'epoch': 1.86}) +[2026-06-16 11:52:11,446][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.21320600807666779, 'train/grad_norm_pre_clip': 24.800479888916016, 'timing/data': 0.02978616300970316, 'timing/model': 1.5389503329060972, 'learning_rate/qwen_vl_interface': 6.469663998998393e-07, 'learning_rate/action_model': 6.469663998998393e-06, 'epoch': 1.86}) +[2026-06-16 11:52:17,656][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.00012528622755780816, 'train/grad_norm_pre_clip': 0.011341534554958344, 'timing/data': 0.030810998752713203, 'timing/model': 1.5131586911156774, 'learning_rate/qwen_vl_interface': 6.462156555071681e-07, 'learning_rate/action_model': 6.46215655507168e-06, 'epoch': 1.87}) +[2026-06-16 11:52:23,985][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 0.016927465796470642, 'train/grad_norm_pre_clip': 15.885591506958008, 'timing/data': 0.03098271880298853, 'timing/model': 1.7690144614316523, 'learning_rate/qwen_vl_interface': 6.454668035563401e-07, 'learning_rate/action_model': 6.454668035563401e-06, 'epoch': 1.87}) +[2026-06-16 11:52:30,325][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.07156933099031448, 'train/grad_norm_pre_clip': 36.084598541259766, 'timing/data': 0.030265355948358774, 'timing/model': 1.7713443539105356, 'learning_rate/qwen_vl_interface': 6.447198443551809e-07, 'learning_rate/action_model': 6.4471984435518075e-06, 'epoch': 1.87}) +[2026-06-16 11:52:36,353][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.15320155024528503, 'train/grad_norm_pre_clip': 14.845723152160645, 'timing/data': 0.03388015506789088, 'timing/model': 1.5391803779639304, 'learning_rate/qwen_vl_interface': 6.439747782107373e-07, 'learning_rate/action_model': 6.4397477821073735e-06, 'epoch': 1.87}) +[2026-06-16 11:52:42,655][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 0.1941070258617401, 'train/grad_norm_pre_clip': 23.265790939331055, 'timing/data': 0.031006770208477974, 'timing/model': 1.5433982112444937, 'learning_rate/qwen_vl_interface': 6.432316054292779e-07, 'learning_rate/action_model': 6.43231605429278e-06, 'epoch': 1.87}) +[2026-06-16 11:52:48,965][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.22956687211990356, 'train/grad_norm_pre_clip': 26.835914611816406, 'timing/data': 0.037008400075137615, 'timing/model': 1.549386822618544, 'learning_rate/qwen_vl_interface': 6.424903263162924e-07, 'learning_rate/action_model': 6.424903263162923e-06, 'epoch': 1.87}) +[2026-06-16 11:52:55,156][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.00035022859810851514, 'train/grad_norm_pre_clip': 0.0682561844587326, 'timing/data': 0.03109905030578375, 'timing/model': 1.5429241303354502, 'learning_rate/qwen_vl_interface': 6.417509411764924e-07, 'learning_rate/action_model': 6.4175094117649244e-06, 'epoch': 1.87}) +[2026-06-16 11:53:01,463][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.0027345961425453424, 'train/grad_norm_pre_clip': 2.122589349746704, 'timing/data': 0.030538388062268496, 'timing/model': 1.534249095711857, 'learning_rate/qwen_vl_interface': 6.410134503138122e-07, 'learning_rate/action_model': 6.4101345031381224e-06, 'epoch': 1.87}) +[2026-06-16 11:53:07,806][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.0011867706198245287, 'train/grad_norm_pre_clip': 0.27115756273269653, 'timing/data': 0.031922254245728254, 'timing/model': 1.7871813317760825, 'learning_rate/qwen_vl_interface': 6.402778540314054e-07, 'learning_rate/action_model': 6.402778540314054e-06, 'epoch': 1.87}) +[2026-06-16 11:53:13,864][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.0010237616952508688, 'train/grad_norm_pre_clip': 0.17359884083271027, 'timing/data': 0.03165912302210927, 'timing/model': 1.5405378509312868, 'learning_rate/qwen_vl_interface': 6.39544152631649e-07, 'learning_rate/action_model': 6.3954415263164895e-06, 'epoch': 1.87}) +[2026-06-16 11:53:20,165][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 0.0006541272741742432, 'train/grad_norm_pre_clip': 0.38670068979263306, 'timing/data': 0.031337752006947994, 'timing/model': 1.5386737519875169, 'learning_rate/qwen_vl_interface': 6.388123464161388e-07, 'learning_rate/action_model': 6.388123464161389e-06, 'epoch': 1.87}) +[2026-06-16 11:53:26,463][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.044236090034246445, 'train/grad_norm_pre_clip': 24.39012908935547, 'timing/data': 0.0313539719209075, 'timing/model': 1.5400301199406385, 'learning_rate/qwen_vl_interface': 6.380824356856936e-07, 'learning_rate/action_model': 6.380824356856936e-06, 'epoch': 1.87}) +[2026-06-16 11:53:32,862][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.044282279908657074, 'train/grad_norm_pre_clip': 32.21285629272461, 'timing/data': 0.030904799234122038, 'timing/model': 1.5798921259120107, 'learning_rate/qwen_vl_interface': 6.373544207403513e-07, 'learning_rate/action_model': 6.373544207403513e-06, 'epoch': 1.87}) +[2026-06-16 11:53:39,326][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 0.0011807854752987623, 'train/grad_norm_pre_clip': 0.5970649123191833, 'timing/data': 0.035694925114512444, 'timing/model': 1.5707875499501824, 'learning_rate/qwen_vl_interface': 6.366283018793723e-07, 'learning_rate/action_model': 6.366283018793723e-06, 'epoch': 1.87}) +[2026-06-16 11:53:46,116][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.0005975489621050656, 'train/grad_norm_pre_clip': 0.12200448662042618, 'timing/data': 0.042063386645168066, 'timing/model': 2.1647337381727993, 'learning_rate/qwen_vl_interface': 6.359040794012358e-07, 'learning_rate/action_model': 6.359040794012358e-06, 'epoch': 1.87}) +[2026-06-16 11:53:52,170][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.002386814448982477, 'train/grad_norm_pre_clip': 2.751938581466675, 'timing/data': 0.03301231982186437, 'timing/model': 1.5374743700958788, 'learning_rate/qwen_vl_interface': 6.351817536036431e-07, 'learning_rate/action_model': 6.351817536036431e-06, 'epoch': 1.87}) +[2026-06-16 11:53:58,473][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.00026786464150063694, 'train/grad_norm_pre_clip': 0.037479743361473083, 'timing/data': 0.03329045185819268, 'timing/model': 1.5421197451651096, 'learning_rate/qwen_vl_interface': 6.344613247835135e-07, 'learning_rate/action_model': 6.344613247835135e-06, 'epoch': 1.87}) +[2026-06-16 11:54:04,772][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.11523900926113129, 'train/grad_norm_pre_clip': 7.385799884796143, 'timing/data': 0.03181675495579839, 'timing/model': 1.5413574208505452, 'learning_rate/qwen_vl_interface': 6.337427932369898e-07, 'learning_rate/action_model': 6.337427932369898e-06, 'epoch': 1.87}) +[2026-06-16 11:54:11,299][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.12522056698799133, 'train/grad_norm_pre_clip': 19.709169387817383, 'timing/data': 0.03164639323949814, 'timing/model': 1.6080815070308745, 'learning_rate/qwen_vl_interface': 6.330261592594328e-07, 'learning_rate/action_model': 6.330261592594327e-06, 'epoch': 1.87}) +[2026-06-16 11:54:17,675][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.19530916213989258, 'train/grad_norm_pre_clip': 28.455785751342773, 'timing/data': 0.03564178431406617, 'timing/model': 1.6003881180658937, 'learning_rate/qwen_vl_interface': 6.323114231454229e-07, 'learning_rate/action_model': 6.323114231454229e-06, 'epoch': 1.87}) +[2026-06-16 11:54:23,998][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.0004826845251955092, 'train/grad_norm_pre_clip': 0.14328676462173462, 'timing/data': 0.030246306210756302, 'timing/model': 1.7757193450815976, 'learning_rate/qwen_vl_interface': 6.315985851887618e-07, 'learning_rate/action_model': 6.315985851887618e-06, 'epoch': 1.87}) +[2026-06-16 11:54:30,162][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.06549955159425735, 'train/grad_norm_pre_clip': 28.231489181518555, 'timing/data': 0.0316965039819479, 'timing/model': 1.533400953747332, 'learning_rate/qwen_vl_interface': 6.308876456824697e-07, 'learning_rate/action_model': 6.3088764568246975e-06, 'epoch': 1.87}) +[2026-06-16 11:54:36,462][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.003509272588416934, 'train/grad_norm_pre_clip': 1.575226902961731, 'timing/data': 0.0314763430505991, 'timing/model': 1.535327693913132, 'learning_rate/qwen_vl_interface': 6.301786049187875e-07, 'learning_rate/action_model': 6.3017860491878745e-06, 'epoch': 1.87}) +[2026-06-16 11:54:42,792][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.11671949177980423, 'train/grad_norm_pre_clip': 25.34925079345703, 'timing/data': 0.03186140488833189, 'timing/model': 1.568993838969618, 'learning_rate/qwen_vl_interface': 6.294714631891755e-07, 'learning_rate/action_model': 6.294714631891754e-06, 'epoch': 1.87}) +[2026-06-16 11:54:49,168][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.0005289441905915737, 'train/grad_norm_pre_clip': 0.13193199038505554, 'timing/data': 0.031237630639225245, 'timing/model': 1.5331944962963462, 'learning_rate/qwen_vl_interface': 6.287662207843115e-07, 'learning_rate/action_model': 6.287662207843114e-06, 'epoch': 1.87}) +[2026-06-16 11:54:55,575][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.0018433756195008755, 'train/grad_norm_pre_clip': 1.1681808233261108, 'timing/data': 0.03050979832187295, 'timing/model': 1.6078394809737802, 'learning_rate/qwen_vl_interface': 6.280628779940956e-07, 'learning_rate/action_model': 6.280628779940956e-06, 'epoch': 1.88}) +[2026-06-16 11:55:01,915][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.006253038067370653, 'train/grad_norm_pre_clip': 7.2158894538879395, 'timing/data': 0.03378788381814957, 'timing/model': 1.7836885121650994, 'learning_rate/qwen_vl_interface': 6.273614351076445e-07, 'learning_rate/action_model': 6.273614351076445e-06, 'epoch': 1.88}) +[2026-06-16 11:55:08,138][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.0002096520911436528, 'train/grad_norm_pre_clip': 0.04718508571386337, 'timing/data': 0.030933409929275513, 'timing/model': 1.7145539452321827, 'learning_rate/qwen_vl_interface': 6.266618924132951e-07, 'learning_rate/action_model': 6.266618924132951e-06, 'epoch': 1.88}) +[2026-06-16 11:55:14,268][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.006582229398190975, 'train/grad_norm_pre_clip': 4.74345588684082, 'timing/data': 0.030827718786895275, 'timing/model': 1.5501857963390648, 'learning_rate/qwen_vl_interface': 6.25964250198603e-07, 'learning_rate/action_model': 6.2596425019860294e-06, 'epoch': 1.88}) +[2026-06-16 11:55:20,584][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.23880262672901154, 'train/grad_norm_pre_clip': 35.9980583190918, 'timing/data': 0.033143230713903904, 'timing/model': 1.5473275817930698, 'learning_rate/qwen_vl_interface': 6.25268508750342e-07, 'learning_rate/action_model': 6.25268508750342e-06, 'epoch': 1.88}) +[2026-06-16 11:55:26,873][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.08798419684171677, 'train/grad_norm_pre_clip': 20.60116195678711, 'timing/data': 0.029389342293143272, 'timing/model': 1.5399415153078735, 'learning_rate/qwen_vl_interface': 6.245746683545062e-07, 'learning_rate/action_model': 6.245746683545062e-06, 'epoch': 1.88}) +[2026-06-16 11:55:33,266][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.005333972163498402, 'train/grad_norm_pre_clip': 7.647970199584961, 'timing/data': 0.03140159184113145, 'timing/model': 1.5768438358791173, 'learning_rate/qwen_vl_interface': 6.238827292963063e-07, 'learning_rate/action_model': 6.238827292963063e-06, 'epoch': 1.88}) +[2026-06-16 11:55:39,637][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 0.0002703686768654734, 'train/grad_norm_pre_clip': 0.06126789748668671, 'timing/data': 0.029549253173172474, 'timing/model': 1.7987321936525404, 'learning_rate/qwen_vl_interface': 6.231926918601731e-07, 'learning_rate/action_model': 6.2319269186017305e-06, 'epoch': 1.88}) +[2026-06-16 11:55:46,034][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 0.0008797773625701666, 'train/grad_norm_pre_clip': 0.2790117561817169, 'timing/data': 0.03135500196367502, 'timing/model': 1.8009755769744515, 'learning_rate/qwen_vl_interface': 6.22504556329754e-07, 'learning_rate/action_model': 6.22504556329754e-06, 'epoch': 1.88}) +[2026-06-16 11:55:52,088][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.04231042414903641, 'train/grad_norm_pre_clip': 29.884429931640625, 'timing/data': 0.03087335964664817, 'timing/model': 1.5516124269925058, 'learning_rate/qwen_vl_interface': 6.218183229879164e-07, 'learning_rate/action_model': 6.218183229879164e-06, 'epoch': 1.88}) +[2026-06-16 11:55:58,465][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.014459982514381409, 'train/grad_norm_pre_clip': 14.306480407714844, 'timing/data': 0.03171646408736706, 'timing/model': 1.6155505790375173, 'learning_rate/qwen_vl_interface': 6.211339921167447e-07, 'learning_rate/action_model': 6.211339921167446e-06, 'epoch': 1.88}) +[2026-06-16 11:56:04,888][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.00023098303063306957, 'train/grad_norm_pre_clip': 0.028023652732372284, 'timing/data': 0.03032127721235156, 'timing/model': 1.5593860801309347, 'learning_rate/qwen_vl_interface': 6.204515639975415e-07, 'learning_rate/action_model': 6.204515639975414e-06, 'epoch': 1.88}) +[2026-06-16 11:56:11,283][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.01724392920732498, 'train/grad_norm_pre_clip': 25.100753784179688, 'timing/data': 0.03125244099646807, 'timing/model': 1.6079076412133873, 'learning_rate/qwen_vl_interface': 6.197710389108262e-07, 'learning_rate/action_model': 6.197710389108262e-06, 'epoch': 1.88}) +[2026-06-16 11:56:17,609][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.166616752743721, 'train/grad_norm_pre_clip': 8.83757495880127, 'timing/data': 0.028707159217447042, 'timing/model': 1.7703178622759879, 'learning_rate/qwen_vl_interface': 6.190924171363378e-07, 'learning_rate/action_model': 6.190924171363378e-06, 'epoch': 1.88}) +[2026-06-16 11:56:23,787][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.00015496132255066186, 'train/grad_norm_pre_clip': 0.019253550097346306, 'timing/data': 0.029920225031673908, 'timing/model': 1.5300474800169468, 'learning_rate/qwen_vl_interface': 6.184156989530326e-07, 'learning_rate/action_model': 6.184156989530325e-06, 'epoch': 1.88}) +[2026-06-16 11:56:30,074][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.012551074847579002, 'train/grad_norm_pre_clip': 17.597558975219727, 'timing/data': 0.028275606222450733, 'timing/model': 1.530166720971465, 'learning_rate/qwen_vl_interface': 6.17740884639083e-07, 'learning_rate/action_model': 6.177408846390829e-06, 'epoch': 1.88}) +[2026-06-16 11:56:36,479][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.03152330592274666, 'train/grad_norm_pre_clip': 37.019771575927734, 'timing/data': 0.03563885483890772, 'timing/model': 1.5642252871766686, 'learning_rate/qwen_vl_interface': 6.170679744718798e-07, 'learning_rate/action_model': 6.1706797447187975e-06, 'epoch': 1.88}) +[2026-06-16 11:56:42,876][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.00016815787239465863, 'train/grad_norm_pre_clip': 0.03434760123491287, 'timing/data': 0.031837995164096355, 'timing/model': 1.5918656811118126, 'learning_rate/qwen_vl_interface': 6.163969687280311e-07, 'learning_rate/action_model': 6.163969687280311e-06, 'epoch': 1.88}) +[2026-06-16 11:56:49,284][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.12956121563911438, 'train/grad_norm_pre_clip': 23.86419105529785, 'timing/data': 0.031465152744203806, 'timing/model': 1.5803287322632968, 'learning_rate/qwen_vl_interface': 6.157278676833627e-07, 'learning_rate/action_model': 6.157278676833627e-06, 'epoch': 1.88}) +[2026-06-16 11:56:55,562][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.00415016757324338, 'train/grad_norm_pre_clip': 4.319546222686768, 'timing/data': 0.03357033431529999, 'timing/model': 1.7349962438456714, 'learning_rate/qwen_vl_interface': 6.150606716129155e-07, 'learning_rate/action_model': 6.150606716129154e-06, 'epoch': 1.88}) +[2026-06-16 11:57:01,705][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.00922329816967249, 'train/grad_norm_pre_clip': 4.818502426147461, 'timing/data': 0.031034220941364765, 'timing/model': 1.5705303228460252, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 1.88}) +[2026-06-16 11:57:08,075][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 0.04230302944779396, 'train/grad_norm_pre_clip': 26.720182418823242, 'timing/data': 0.0348503808490932, 'timing/model': 1.6154551152139902, 'learning_rate/qwen_vl_interface': 6.137319954909407e-07, 'learning_rate/action_model': 6.137319954909406e-06, 'epoch': 1.88}) +[2026-06-16 11:57:14,475][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.0006125480285845697, 'train/grad_norm_pre_clip': 0.11798077821731567, 'timing/data': 0.03191369539126754, 'timing/model': 1.548470820300281, 'learning_rate/qwen_vl_interface': 6.130705159855815e-07, 'learning_rate/action_model': 6.130705159855814e-06, 'epoch': 1.88}) +[2026-06-16 11:57:20,873][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.004866508301347494, 'train/grad_norm_pre_clip': 4.773477077484131, 'timing/data': 0.03154927305877209, 'timing/model': 1.5817696726880968, 'learning_rate/qwen_vl_interface': 6.124109425467805e-07, 'learning_rate/action_model': 6.124109425467805e-06, 'epoch': 1.88}) +[2026-06-16 11:57:27,374][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.00047122687101364136, 'train/grad_norm_pre_clip': 0.13567161560058594, 'timing/data': 0.03255443926900625, 'timing/model': 1.5777801419608295, 'learning_rate/qwen_vl_interface': 6.117532754456642e-07, 'learning_rate/action_model': 6.117532754456642e-06, 'epoch': 1.88}) +[2026-06-16 11:57:33,695][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.3412413001060486, 'train/grad_norm_pre_clip': 31.087024688720703, 'timing/data': 0.03357753390446305, 'timing/model': 1.7581791570410132, 'learning_rate/qwen_vl_interface': 6.110975149525746e-07, 'learning_rate/action_model': 6.110975149525746e-06, 'epoch': 1.89}) +[2026-06-16 11:57:39,877][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.0007980707450769842, 'train/grad_norm_pre_clip': 0.1971682459115982, 'timing/data': 0.03225849708542228, 'timing/model': 1.5641837627626956, 'learning_rate/qwen_vl_interface': 6.104436613370692e-07, 'learning_rate/action_model': 6.104436613370692e-06, 'epoch': 1.89}) +[2026-06-16 11:57:46,271][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.038026727735996246, 'train/grad_norm_pre_clip': 30.128681182861328, 'timing/data': 0.03134998260065913, 'timing/model': 1.5414505652152002, 'learning_rate/qwen_vl_interface': 6.09791714867924e-07, 'learning_rate/action_model': 6.0979171486792396e-06, 'epoch': 1.89}) +[2026-06-16 11:57:52,578][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.00027428471366874874, 'train/grad_norm_pre_clip': 0.08880941569805145, 'timing/data': 0.03147611301392317, 'timing/model': 1.5506337038241327, 'learning_rate/qwen_vl_interface': 6.091416758131287e-07, 'learning_rate/action_model': 6.091416758131286e-06, 'epoch': 1.89}) +[2026-06-16 11:57:59,074][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.07198075950145721, 'train/grad_norm_pre_clip': 29.2784423828125, 'timing/data': 0.03607043670490384, 'timing/model': 1.5898710982874036, 'learning_rate/qwen_vl_interface': 6.084935444398895e-07, 'learning_rate/action_model': 6.084935444398894e-06, 'epoch': 1.89}) +[2026-06-16 11:58:05,421][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.0001935276377480477, 'train/grad_norm_pre_clip': 0.05046296492218971, 'timing/data': 0.030852030031383038, 'timing/model': 1.7975058751180768, 'learning_rate/qwen_vl_interface': 6.078473210146294e-07, 'learning_rate/action_model': 6.078473210146294e-06, 'epoch': 1.89}) +[2026-06-16 11:58:11,468][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 0.002541881985962391, 'train/grad_norm_pre_clip': 1.8325856924057007, 'timing/data': 0.0315094031393528, 'timing/model': 1.5369500149972737, 'learning_rate/qwen_vl_interface': 6.072030058029862e-07, 'learning_rate/action_model': 6.072030058029861e-06, 'epoch': 1.89}) +[2026-06-16 11:58:17,789][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 2.5222498152288608e-05, 'train/grad_norm_pre_clip': 0.003060170216485858, 'timing/data': 0.031578314024955034, 'timing/model': 1.5566311269067228, 'learning_rate/qwen_vl_interface': 6.065605990698136e-07, 'learning_rate/action_model': 6.065605990698136e-06, 'epoch': 1.89}) +[2026-06-16 11:58:24,071][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.0010778738651424646, 'train/grad_norm_pre_clip': 0.7552730441093445, 'timing/data': 0.03402571566402912, 'timing/model': 1.5370808569714427, 'learning_rate/qwen_vl_interface': 6.059201010791811e-07, 'learning_rate/action_model': 6.059201010791811e-06, 'epoch': 1.89}) +[2026-06-16 11:58:30,372][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.0008457085350528359, 'train/grad_norm_pre_clip': 0.2614440619945526, 'timing/data': 0.030308318324387074, 'timing/model': 1.5393903679214418, 'learning_rate/qwen_vl_interface': 6.052815120943721e-07, 'learning_rate/action_model': 6.052815120943721e-06, 'epoch': 1.89}) +[2026-06-16 11:58:36,674][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.0011153585510328412, 'train/grad_norm_pre_clip': 0.28499606251716614, 'timing/data': 0.031060210894793272, 'timing/model': 1.5455375909805298, 'learning_rate/qwen_vl_interface': 6.046448323778874e-07, 'learning_rate/action_model': 6.046448323778874e-06, 'epoch': 1.89}) +[2026-06-16 11:58:43,013][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.0007553918985649943, 'train/grad_norm_pre_clip': 0.1448092758655548, 'timing/data': 0.03122973069548607, 'timing/model': 1.7814622251316905, 'learning_rate/qwen_vl_interface': 6.040100621914419e-07, 'learning_rate/action_model': 6.040100621914418e-06, 'epoch': 1.89}) +[2026-06-16 11:58:49,074][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.00038725847844034433, 'train/grad_norm_pre_clip': 0.07746872305870056, 'timing/data': 0.03191386489197612, 'timing/model': 1.536987588275224, 'learning_rate/qwen_vl_interface': 6.033772017959652e-07, 'learning_rate/action_model': 6.033772017959652e-06, 'epoch': 1.89}) +[2026-06-16 11:58:55,378][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.0009734449558891356, 'train/grad_norm_pre_clip': 0.38864561915397644, 'timing/data': 0.03144202288240194, 'timing/model': 1.5446845176629722, 'learning_rate/qwen_vl_interface': 6.027462514516022e-07, 'learning_rate/action_model': 6.0274625145160214e-06, 'epoch': 1.89}) +[2026-06-16 11:59:01,673][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.0006566665251739323, 'train/grad_norm_pre_clip': 0.15071594715118408, 'timing/data': 0.03187893470749259, 'timing/model': 1.5430587497539818, 'learning_rate/qwen_vl_interface': 6.021172114177139e-07, 'learning_rate/action_model': 6.021172114177139e-06, 'epoch': 1.89}) +[2026-06-16 11:59:08,075][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.3193540871143341, 'train/grad_norm_pre_clip': 29.74458122253418, 'timing/data': 0.0352939129807055, 'timing/model': 1.5622764900326729, 'learning_rate/qwen_vl_interface': 6.014900819528734e-07, 'learning_rate/action_model': 6.0149008195287335e-06, 'epoch': 1.89}) +[2026-06-16 11:59:14,372][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.0005224765045568347, 'train/grad_norm_pre_clip': 0.11291221529245377, 'timing/data': 0.03032400691881776, 'timing/model': 1.5464112581685185, 'learning_rate/qwen_vl_interface': 6.008648633148707e-07, 'learning_rate/action_model': 6.008648633148707e-06, 'epoch': 1.89}) +[2026-06-16 11:59:20,743][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.00029848385020159185, 'train/grad_norm_pre_clip': 0.05117001757025719, 'timing/data': 0.031560963951051235, 'timing/model': 1.8089173398911953, 'learning_rate/qwen_vl_interface': 6.002415557607097e-07, 'learning_rate/action_model': 6.002415557607096e-06, 'epoch': 1.89}) +[2026-06-16 11:59:26,771][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.00205545243807137, 'train/grad_norm_pre_clip': 3.1457607746124268, 'timing/data': 0.031147710978984833, 'timing/model': 1.5435321647673845, 'learning_rate/qwen_vl_interface': 5.996201595466088e-07, 'learning_rate/action_model': 5.996201595466088e-06, 'epoch': 1.89}) +[2026-06-16 11:59:33,070][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.28530624508857727, 'train/grad_norm_pre_clip': 24.813215255737305, 'timing/data': 0.031091590877622366, 'timing/model': 1.543844607193023, 'learning_rate/qwen_vl_interface': 5.990006749279998e-07, 'learning_rate/action_model': 5.990006749279998e-06, 'epoch': 1.89}) +[2026-06-16 11:59:39,448][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.05595143139362335, 'train/grad_norm_pre_clip': 29.22170639038086, 'timing/data': 0.030180826783180237, 'timing/model': 1.6206664349883795, 'learning_rate/qwen_vl_interface': 5.983831021595296e-07, 'learning_rate/action_model': 5.9838310215952956e-06, 'epoch': 1.89}) +[2026-06-16 11:59:45,871][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.006280412431806326, 'train/grad_norm_pre_clip': 3.8875482082366943, 'timing/data': 0.033035640604794025, 'timing/model': 1.5840620552189648, 'learning_rate/qwen_vl_interface': 5.977674414950595e-07, 'learning_rate/action_model': 5.9776744149505955e-06, 'epoch': 1.89}) +[2026-06-16 11:59:52,171][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.0005297200987115502, 'train/grad_norm_pre_clip': 0.18289123475551605, 'timing/data': 0.03173670405521989, 'timing/model': 1.5298728463239968, 'learning_rate/qwen_vl_interface': 5.971536931876635e-07, 'learning_rate/action_model': 5.971536931876634e-06, 'epoch': 1.89}) +[2026-06-16 11:59:58,601][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.001968044089153409, 'train/grad_norm_pre_clip': 0.7636694312095642, 'timing/data': 0.03172249486669898, 'timing/model': 1.868807383812964, 'learning_rate/qwen_vl_interface': 5.965418574896315e-07, 'learning_rate/action_model': 5.9654185748963145e-06, 'epoch': 1.89}) +[2026-06-16 12:00:04,661][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.09994868189096451, 'train/grad_norm_pre_clip': 79.66769409179688, 'timing/data': 0.03132783202454448, 'timing/model': 1.538955524098128, 'learning_rate/qwen_vl_interface': 5.959319346524646e-07, 'learning_rate/action_model': 5.959319346524646e-06, 'epoch': 1.89}) +[2026-06-16 12:00:10,965][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.010140417143702507, 'train/grad_norm_pre_clip': 17.8178653717041, 'timing/data': 0.03260724898427725, 'timing/model': 1.5388521528802812, 'learning_rate/qwen_vl_interface': 5.953239249268804e-07, 'learning_rate/action_model': 5.953239249268804e-06, 'epoch': 1.9}) +[2026-06-16 12:00:17,270][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.05313500016927719, 'train/grad_norm_pre_clip': 31.663257598876953, 'timing/data': 0.030197015963494778, 'timing/model': 1.543684198986739, 'learning_rate/qwen_vl_interface': 5.947178285628081e-07, 'learning_rate/action_model': 5.94717828562808e-06, 'epoch': 1.9}) +[2026-06-16 12:00:23,569][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.00027488640625961125, 'train/grad_norm_pre_clip': 0.050516802817583084, 'timing/data': 0.03151397313922644, 'timing/model': 1.5394860277883708, 'learning_rate/qwen_vl_interface': 5.941136458093907e-07, 'learning_rate/action_model': 5.941136458093907e-06, 'epoch': 1.9}) +[2026-06-16 12:00:29,993][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.047668375074863434, 'train/grad_norm_pre_clip': 16.981504440307617, 'timing/data': 0.03143655275925994, 'timing/model': 1.6036030910909176, 'learning_rate/qwen_vl_interface': 5.935113769149855e-07, 'learning_rate/action_model': 5.935113769149856e-06, 'epoch': 1.9}) +[2026-06-16 12:00:36,312][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.025032125413417816, 'train/grad_norm_pre_clip': 26.27594566345215, 'timing/data': 0.029983926098793745, 'timing/model': 1.7886236403137445, 'learning_rate/qwen_vl_interface': 5.92911022127163e-07, 'learning_rate/action_model': 5.929110221271629e-06, 'epoch': 1.9}) +[2026-06-16 12:00:42,465][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.08776787668466568, 'train/grad_norm_pre_clip': 20.611797332763672, 'timing/data': 0.03071597032248974, 'timing/model': 1.544988377019763, 'learning_rate/qwen_vl_interface': 5.92312581692705e-07, 'learning_rate/action_model': 5.923125816927049e-06, 'epoch': 1.9}) +[2026-06-16 12:00:48,765][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.03283972293138504, 'train/grad_norm_pre_clip': 28.71162986755371, 'timing/data': 0.02982092509046197, 'timing/model': 1.5421933229081333, 'learning_rate/qwen_vl_interface': 5.917160558576088e-07, 'learning_rate/action_model': 5.917160558576088e-06, 'epoch': 1.9}) +[2026-06-16 12:00:54,965][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.0007630424806848168, 'train/grad_norm_pre_clip': 0.19185279309749603, 'timing/data': 0.03041598806157708, 'timing/model': 1.543969361577183, 'learning_rate/qwen_vl_interface': 5.911214448670829e-07, 'learning_rate/action_model': 5.911214448670828e-06, 'epoch': 1.9}) +[2026-06-16 12:01:01,299][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.005700609181076288, 'train/grad_norm_pre_clip': 6.599802017211914, 'timing/data': 0.03177087381482124, 'timing/model': 1.5607292600907385, 'learning_rate/qwen_vl_interface': 5.905287489655499e-07, 'learning_rate/action_model': 5.905287489655499e-06, 'epoch': 1.9}) +[2026-06-16 12:01:07,684][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 0.24540568888187408, 'train/grad_norm_pre_clip': 23.711210250854492, 'timing/data': 0.030732998624444008, 'timing/model': 1.6121528069488704, 'learning_rate/qwen_vl_interface': 5.899379683966445e-07, 'learning_rate/action_model': 5.899379683966445e-06, 'epoch': 1.9}) +[2026-06-16 12:01:14,009][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 7.495252066291869e-05, 'train/grad_norm_pre_clip': 0.005226996727287769, 'timing/data': 0.03156138397753239, 'timing/model': 1.7737224362790585, 'learning_rate/qwen_vl_interface': 5.893491034032142e-07, 'learning_rate/action_model': 5.893491034032142e-06, 'epoch': 1.9}) +[2026-06-16 12:01:20,074][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.00017162636504508555, 'train/grad_norm_pre_clip': 0.026351632550358772, 'timing/data': 0.03129394259303808, 'timing/model': 1.5426578568294644, 'learning_rate/qwen_vl_interface': 5.887621542273196e-07, 'learning_rate/action_model': 5.887621542273196e-06, 'epoch': 1.9}) +[2026-06-16 12:01:26,477][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.0005265666404739022, 'train/grad_norm_pre_clip': 0.2094404399394989, 'timing/data': 0.03151331329718232, 'timing/model': 1.5493590431287885, 'learning_rate/qwen_vl_interface': 5.881771211102327e-07, 'learning_rate/action_model': 5.881771211102327e-06, 'epoch': 1.9}) +[2026-06-16 12:12:12,211][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.0019659188110381365, 'train/grad_norm_pre_clip': 1.781571388244629, 'eval/loss': 0.03515880927443504, 'eval/flap_precision': 0.9715061058344641, 'eval/flap_recall': 0.9768076398362893, 'eval/flap_f1': 0.9741496598639455, 'eval/flap_rate_model': 0.20472222222222222, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9741496598639455, 'timing/data': 0.03093566093593836, 'timing/model': 1.5573021941818297, 'learning_rate/qwen_vl_interface': 5.875940042924391e-07, 'learning_rate/action_model': 5.875940042924391e-06, 'epoch': 1.9}) +[2026-06-16 12:12:12,213][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4700_state +[2026-06-16 12:12:12,219][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 12:12:20,526][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4700_state/model.safetensors +[2026-06-16 12:12:39,456][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4700_state/optimizer.bin +[2026-06-16 12:12:39,457][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4700_state/sampler.bin +[2026-06-16 12:12:39,457][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4700_state/sampler_1.bin +[2026-06-16 12:12:39,458][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4700_state/random_states_0.pkl +[2026-06-16 12:12:39,641][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4700 +[2026-06-16 12:12:39,641][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 12:12:39,656][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 12:12:45,929][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.008256650529801846, 'train/grad_norm_pre_clip': 7.339668273925781, 'timing/data': 0.0316970469430089, 'timing/model': 1.549052372109145, 'learning_rate/qwen_vl_interface': 5.870128040136368e-07, 'learning_rate/action_model': 5.870128040136367e-06, 'epoch': 1.9}) +[2026-06-16 12:12:52,136][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.0002143817109754309, 'train/grad_norm_pre_clip': 0.039010971784591675, 'timing/data': 0.0312717380002141, 'timing/model': 1.5303925001062453, 'learning_rate/qwen_vl_interface': 5.864335205127332e-07, 'learning_rate/action_model': 5.864335205127332e-06, 'epoch': 1.9}) +[2026-06-16 12:12:58,338][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.0008578192791901529, 'train/grad_norm_pre_clip': 0.4995512068271637, 'timing/data': 0.031020336784422398, 'timing/model': 1.544512689113617, 'learning_rate/qwen_vl_interface': 5.858561540278513e-07, 'learning_rate/action_model': 5.858561540278513e-06, 'epoch': 1.9}) +[2026-06-16 12:13:04,545][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 0.00020272891561035067, 'train/grad_norm_pre_clip': 0.05644148588180542, 'timing/data': 0.03152428986504674, 'timing/model': 1.5214196448214352, 'learning_rate/qwen_vl_interface': 5.852807047963244e-07, 'learning_rate/action_model': 5.852807047963243e-06, 'epoch': 1.9}) +[2026-06-16 12:13:10,866][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.0012687480775639415, 'train/grad_norm_pre_clip': 0.8810746073722839, 'timing/data': 0.029887896962463856, 'timing/model': 1.8492590938694775, 'learning_rate/qwen_vl_interface': 5.847071730546974e-07, 'learning_rate/action_model': 5.847071730546974e-06, 'epoch': 1.9}) +[2026-06-16 12:13:17,136][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.0027866356540471315, 'train/grad_norm_pre_clip': 1.9052430391311646, 'timing/data': 0.031026230193674564, 'timing/model': 1.534869397059083, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 1.9}) +[2026-06-16 12:13:23,371][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 0.003502040635794401, 'train/grad_norm_pre_clip': 3.5059902667999268, 'timing/data': 0.033629216719418764, 'timing/model': 1.559860761743039, 'learning_rate/qwen_vl_interface': 5.835658629833866e-07, 'learning_rate/action_model': 5.8356586298338654e-06, 'epoch': 1.9}) +[2026-06-16 12:13:29,652][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.07113063335418701, 'train/grad_norm_pre_clip': 55.823997497558594, 'timing/data': 0.030649160966277122, 'timing/model': 1.54509585397318, 'learning_rate/qwen_vl_interface': 5.829980851228512e-07, 'learning_rate/action_model': 5.829980851228512e-06, 'epoch': 1.9}) +[2026-06-16 12:13:35,954][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.00027188562671653926, 'train/grad_norm_pre_clip': 0.1040925532579422, 'timing/data': 0.03272007638588548, 'timing/model': 1.534544873982668, 'learning_rate/qwen_vl_interface': 5.824322256905155e-07, 'learning_rate/action_model': 5.824322256905154e-06, 'epoch': 1.9}) +[2026-06-16 12:13:42,251][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.004661388695240021, 'train/grad_norm_pre_clip': 4.293875217437744, 'timing/data': 0.029832350090146065, 'timing/model': 1.5035434272140265, 'learning_rate/qwen_vl_interface': 5.818682849189818e-07, 'learning_rate/action_model': 5.818682849189817e-06, 'epoch': 1.9}) +[2026-06-16 12:13:48,588][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.0003566288505680859, 'train/grad_norm_pre_clip': 0.05707614868879318, 'timing/data': 0.03521351283416152, 'timing/model': 1.7821699432097375, 'learning_rate/qwen_vl_interface': 5.813062630400653e-07, 'learning_rate/action_model': 5.813062630400653e-06, 'epoch': 1.91}) +[2026-06-16 12:13:54,903][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 0.19495858252048492, 'train/grad_norm_pre_clip': 8.462904930114746, 'timing/data': 0.031226494815200567, 'timing/model': 1.7871315432712436, 'learning_rate/qwen_vl_interface': 5.807461602847922e-07, 'learning_rate/action_model': 5.807461602847922e-06, 'epoch': 1.91}) +[2026-06-16 12:14:00,948][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.0005298979813233018, 'train/grad_norm_pre_clip': 0.22567525506019592, 'timing/data': 0.03052217373624444, 'timing/model': 1.5333556379191577, 'learning_rate/qwen_vl_interface': 5.801879768833998e-07, 'learning_rate/action_model': 5.801879768833998e-06, 'epoch': 1.91}) +[2026-06-16 12:14:07,261][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.003177554812282324, 'train/grad_norm_pre_clip': 4.214419841766357, 'timing/data': 0.031284525990486145, 'timing/model': 1.5370602309703827, 'learning_rate/qwen_vl_interface': 5.796317130653363e-07, 'learning_rate/action_model': 5.796317130653363e-06, 'epoch': 1.91}) +[2026-06-16 12:14:13,558][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.0016762400045990944, 'train/grad_norm_pre_clip': 1.2206246852874756, 'timing/data': 0.030468905344605446, 'timing/model': 1.5418895659968257, 'learning_rate/qwen_vl_interface': 5.790773690592592e-07, 'learning_rate/action_model': 5.7907736905925925e-06, 'epoch': 1.91}) +[2026-06-16 12:14:19,965][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.00023049412993714213, 'train/grad_norm_pre_clip': 0.028116391971707344, 'timing/data': 0.03166200779378414, 'timing/model': 1.6005542953498662, 'learning_rate/qwen_vl_interface': 5.7852494509304e-07, 'learning_rate/action_model': 5.7852494509304e-06, 'epoch': 1.91}) +[2026-06-16 12:14:26,318][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.0007965601398609579, 'train/grad_norm_pre_clip': 0.18502451479434967, 'timing/data': 0.030305475927889347, 'timing/model': 1.7873767861165106, 'learning_rate/qwen_vl_interface': 5.779744413937592e-07, 'learning_rate/action_model': 5.779744413937592e-06, 'epoch': 1.91}) +[2026-06-16 12:14:32,380][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.3100241720676422, 'train/grad_norm_pre_clip': 36.6890754699707, 'timing/data': 0.03171862009912729, 'timing/model': 1.5624797539785504, 'learning_rate/qwen_vl_interface': 5.774258581877073e-07, 'learning_rate/action_model': 5.774258581877073e-06, 'epoch': 1.91}) +[2026-06-16 12:14:38,774][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.11548113077878952, 'train/grad_norm_pre_clip': 22.264453887939453, 'timing/data': 0.03049309691414237, 'timing/model': 1.5453009232878685, 'learning_rate/qwen_vl_interface': 5.768791957003866e-07, 'learning_rate/action_model': 5.768791957003865e-06, 'epoch': 1.91}) +[2026-06-16 12:14:45,071][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.08938447386026382, 'train/grad_norm_pre_clip': 63.79236602783203, 'timing/data': 0.03179252101108432, 'timing/model': 1.5243570301681757, 'learning_rate/qwen_vl_interface': 5.763344541565091e-07, 'learning_rate/action_model': 5.7633445415650905e-06, 'epoch': 1.91}) +[2026-06-16 12:14:51,437][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 6.259384099394083e-05, 'train/grad_norm_pre_clip': 0.017906274646520615, 'timing/data': 0.029700187034904957, 'timing/model': 1.582591142039746, 'learning_rate/qwen_vl_interface': 5.75791633779998e-07, 'learning_rate/action_model': 5.75791633779998e-06, 'epoch': 1.91}) +[2026-06-16 12:14:57,661][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 0.018095124512910843, 'train/grad_norm_pre_clip': 9.505188941955566, 'timing/data': 0.030952800065279007, 'timing/model': 1.582238371949643, 'learning_rate/qwen_vl_interface': 5.752507347939857e-07, 'learning_rate/action_model': 5.752507347939858e-06, 'epoch': 1.91}) +[2026-06-16 12:15:03,948][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.0005595219554379582, 'train/grad_norm_pre_clip': 0.24264292418956757, 'timing/data': 0.028946456033736467, 'timing/model': 1.7332289880141616, 'learning_rate/qwen_vl_interface': 5.74711757420816e-07, 'learning_rate/action_model': 5.747117574208159e-06, 'epoch': 1.91}) +[2026-06-16 12:15:09,966][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.10664491355419159, 'train/grad_norm_pre_clip': 14.838194847106934, 'timing/data': 0.031504442915320396, 'timing/model': 1.537385503295809, 'learning_rate/qwen_vl_interface': 5.74174701882042e-07, 'learning_rate/action_model': 5.74174701882042e-06, 'epoch': 1.91}) +[2026-06-16 12:15:16,269][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.21986250579357147, 'train/grad_norm_pre_clip': 31.611188888549805, 'timing/data': 0.03205511532723904, 'timing/model': 1.539727597963065, 'learning_rate/qwen_vl_interface': 5.736395683984262e-07, 'learning_rate/action_model': 5.736395683984261e-06, 'epoch': 1.91}) +[2026-06-16 12:15:22,667][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.1394025981426239, 'train/grad_norm_pre_clip': 18.083974838256836, 'timing/data': 0.03101291274651885, 'timing/model': 1.5603198129683733, 'learning_rate/qwen_vl_interface': 5.731063571899421e-07, 'learning_rate/action_model': 5.731063571899421e-06, 'epoch': 1.91}) +[2026-06-16 12:15:29,058][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.14039170742034912, 'train/grad_norm_pre_clip': 57.576942443847656, 'timing/data': 0.02914360910654068, 'timing/model': 1.5956826470792294, 'learning_rate/qwen_vl_interface': 5.72575068475773e-07, 'learning_rate/action_model': 5.72575068475773e-06, 'epoch': 1.91}) +[2026-06-16 12:15:35,474][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.00048233280540443957, 'train/grad_norm_pre_clip': 0.14638079702854156, 'timing/data': 0.03049116302281618, 'timing/model': 1.605149972718209, 'learning_rate/qwen_vl_interface': 5.720457024743115e-07, 'learning_rate/action_model': 5.720457024743115e-06, 'epoch': 1.91}) +[2026-06-16 12:15:41,784][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.19568802416324615, 'train/grad_norm_pre_clip': 43.43601989746094, 'timing/data': 0.03049031412228942, 'timing/model': 1.7402200708165765, 'learning_rate/qwen_vl_interface': 5.715182594031601e-07, 'learning_rate/action_model': 5.7151825940316e-06, 'epoch': 1.91}) +[2026-06-16 12:15:47,875][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.00016772608796600252, 'train/grad_norm_pre_clip': 0.03661974146962166, 'timing/data': 0.030307534150779247, 'timing/model': 1.556211344897747, 'learning_rate/qwen_vl_interface': 5.709927394791305e-07, 'learning_rate/action_model': 5.709927394791305e-06, 'epoch': 1.91}) +[2026-06-16 12:15:54,167][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.13091498613357544, 'train/grad_norm_pre_clip': 38.20444107055664, 'timing/data': 0.03210771922022104, 'timing/model': 1.5288356062956154, 'learning_rate/qwen_vl_interface': 5.70469142918244e-07, 'learning_rate/action_model': 5.704691429182441e-06, 'epoch': 1.91}) +[2026-06-16 12:16:00,470][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.0006231896113604307, 'train/grad_norm_pre_clip': 0.12025388330221176, 'timing/data': 0.030905886087566614, 'timing/model': 1.5419046599417925, 'learning_rate/qwen_vl_interface': 5.699474699357317e-07, 'learning_rate/action_model': 5.699474699357317e-06, 'epoch': 1.91}) +[2026-06-16 12:16:06,872][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.0005266857915557921, 'train/grad_norm_pre_clip': 0.27445188164711, 'timing/data': 0.03202119003981352, 'timing/model': 1.5929625793360174, 'learning_rate/qwen_vl_interface': 5.694277207460331e-07, 'learning_rate/action_model': 5.694277207460332e-06, 'epoch': 1.91}) +[2026-06-16 12:16:13,206][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.0004314081452321261, 'train/grad_norm_pre_clip': 0.11670629680156708, 'timing/data': 0.029997865203768015, 'timing/model': 1.7792847719974816, 'learning_rate/qwen_vl_interface': 5.689098955627982e-07, 'learning_rate/action_model': 5.689098955627982e-06, 'epoch': 1.91}) +[2026-06-16 12:16:19,276][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.0068093412555754185, 'train/grad_norm_pre_clip': 5.592228889465332, 'timing/data': 0.03118953760713339, 'timing/model': 1.5476735550910234, 'learning_rate/qwen_vl_interface': 5.683939945988853e-07, 'learning_rate/action_model': 5.683939945988853e-06, 'epoch': 1.91}) +[2026-06-16 12:16:25,585][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.11526503413915634, 'train/grad_norm_pre_clip': 25.816099166870117, 'timing/data': 0.030195216182619333, 'timing/model': 1.5481993020512164, 'learning_rate/qwen_vl_interface': 5.678800180663608e-07, 'learning_rate/action_model': 5.678800180663609e-06, 'epoch': 1.92}) +[2026-06-16 12:16:31,966][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.06137550249695778, 'train/grad_norm_pre_clip': 39.78712844848633, 'timing/data': 0.03145475033670664, 'timing/model': 1.5592052889987826, 'learning_rate/qwen_vl_interface': 5.673679661765018e-07, 'learning_rate/action_model': 5.6736796617650175e-06, 'epoch': 1.92}) +[2026-06-16 12:16:38,284][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.0042760008946061134, 'train/grad_norm_pre_clip': 4.3591389656066895, 'timing/data': 0.031011899933218956, 'timing/model': 1.5370732299052179, 'learning_rate/qwen_vl_interface': 5.668578391397938e-07, 'learning_rate/action_model': 5.668578391397937e-06, 'epoch': 1.92}) +[2026-06-16 12:16:44,565][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 6.10005808994174e-05, 'train/grad_norm_pre_clip': 0.00813709944486618, 'timing/data': 0.030675499234348536, 'timing/model': 1.5347726591862738, 'learning_rate/qwen_vl_interface': 5.663496371659303e-07, 'learning_rate/action_model': 5.663496371659303e-06, 'epoch': 1.92}) +[2026-06-16 12:16:50,916][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.0003143912472296506, 'train/grad_norm_pre_clip': 0.06533107906579971, 'timing/data': 0.031220521312206984, 'timing/model': 1.7841248298063874, 'learning_rate/qwen_vl_interface': 5.658433604638141e-07, 'learning_rate/action_model': 5.65843360463814e-06, 'epoch': 1.92}) +[2026-06-16 12:16:56,976][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.00015785422874614596, 'train/grad_norm_pre_clip': 0.023292824625968933, 'timing/data': 0.030978929717093706, 'timing/model': 1.5414465046487749, 'learning_rate/qwen_vl_interface': 5.653390092415559e-07, 'learning_rate/action_model': 5.653390092415559e-06, 'epoch': 1.92}) +[2026-06-16 12:17:03,266][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 0.08598349243402481, 'train/grad_norm_pre_clip': 31.792810440063477, 'timing/data': 0.031195451971143484, 'timing/model': 1.54120192816481, 'learning_rate/qwen_vl_interface': 5.648365837064759e-07, 'learning_rate/action_model': 5.648365837064759e-06, 'epoch': 1.92}) +[2026-06-16 12:17:09,566][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.21045225858688354, 'train/grad_norm_pre_clip': 8.52510929107666, 'timing/data': 0.02817470394074917, 'timing/model': 1.547043717931956, 'learning_rate/qwen_vl_interface': 5.643360840651025e-07, 'learning_rate/action_model': 5.643360840651025e-06, 'epoch': 1.92}) +[2026-06-16 12:17:16,045][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.0009689012076705694, 'train/grad_norm_pre_clip': 0.47089025378227234, 'timing/data': 0.0315023441798985, 'timing/model': 1.6182689550332725, 'learning_rate/qwen_vl_interface': 5.638375105231719e-07, 'learning_rate/action_model': 5.638375105231719e-06, 'epoch': 1.92}) +[2026-06-16 12:17:22,365][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.001036856323480606, 'train/grad_norm_pre_clip': 0.26293984055519104, 'timing/data': 0.03011688031256199, 'timing/model': 1.556281863246113, 'learning_rate/qwen_vl_interface': 5.633408632856294e-07, 'learning_rate/action_model': 5.6334086328562945e-06, 'epoch': 1.92}) +[2026-06-16 12:17:28,701][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.00046953107812441885, 'train/grad_norm_pre_clip': 0.19248418509960175, 'timing/data': 0.03044834081083536, 'timing/model': 1.776224018074572, 'learning_rate/qwen_vl_interface': 5.628461425566268e-07, 'learning_rate/action_model': 5.6284614255662675e-06, 'epoch': 1.92}) +[2026-06-16 12:17:34,762][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 2.200800008722581e-05, 'train/grad_norm_pre_clip': 0.0020043847616761923, 'timing/data': 0.034648384898900986, 'timing/model': 1.5355725451372564, 'learning_rate/qwen_vl_interface': 5.62353348539526e-07, 'learning_rate/action_model': 5.623533485395259e-06, 'epoch': 1.92}) +[2026-06-16 12:17:41,169][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 0.05661527067422867, 'train/grad_norm_pre_clip': 33.36431884765625, 'timing/data': 0.033367801923304796, 'timing/model': 1.540212631225586, 'learning_rate/qwen_vl_interface': 5.618624814368956e-07, 'learning_rate/action_model': 5.618624814368956e-06, 'epoch': 1.92}) +[2026-06-16 12:17:47,469][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.00015890975191723555, 'train/grad_norm_pre_clip': 0.03106091171503067, 'timing/data': 0.029519291128963232, 'timing/model': 1.5470511340536177, 'learning_rate/qwen_vl_interface': 5.613735414505126e-07, 'learning_rate/action_model': 5.613735414505125e-06, 'epoch': 1.92}) +[2026-06-16 12:17:53,847][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.021510951220989227, 'train/grad_norm_pre_clip': 17.31391716003418, 'timing/data': 0.029970532283186913, 'timing/model': 1.5901043280027807, 'learning_rate/qwen_vl_interface': 5.608865287813617e-07, 'learning_rate/action_model': 5.608865287813616e-06, 'epoch': 1.92}) +[2026-06-16 12:18:00,073][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.11583666503429413, 'train/grad_norm_pre_clip': 51.6860466003418, 'timing/data': 0.033384403213858604, 'timing/model': 1.5292669022455812, 'learning_rate/qwen_vl_interface': 5.604014436296362e-07, 'learning_rate/action_model': 5.604014436296362e-06, 'epoch': 1.92}) +[2026-06-16 12:18:06,471][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.0012750716414302588, 'train/grad_norm_pre_clip': 0.9022775292396545, 'timing/data': 0.031293577048927546, 'timing/model': 1.8431213237345219, 'learning_rate/qwen_vl_interface': 5.599182861947348e-07, 'learning_rate/action_model': 5.599182861947347e-06, 'epoch': 1.92}) +[2026-06-16 12:18:12,666][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.0018890888895839453, 'train/grad_norm_pre_clip': 0.8184370398521423, 'timing/data': 0.03191971918568015, 'timing/model': 1.5462776967324317, 'learning_rate/qwen_vl_interface': 5.594370566752665e-07, 'learning_rate/action_model': 5.5943705667526645e-06, 'epoch': 1.92}) +[2026-06-16 12:18:18,971][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.0002666980726644397, 'train/grad_norm_pre_clip': 0.060121938586235046, 'timing/data': 0.03261564113199711, 'timing/model': 1.5350049119442701, 'learning_rate/qwen_vl_interface': 5.589577552690456e-07, 'learning_rate/action_model': 5.589577552690456e-06, 'epoch': 1.92}) +[2026-06-16 12:18:25,267][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.0006289551965892315, 'train/grad_norm_pre_clip': 0.14951667189598083, 'timing/data': 0.03128233831375837, 'timing/model': 1.5427590170875192, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 1.92}) +[2026-06-16 12:18:31,569][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.00839769747108221, 'train/grad_norm_pre_clip': 6.269076824188232, 'timing/data': 0.030178575310856104, 'timing/model': 1.536016325931996, 'learning_rate/qwen_vl_interface': 5.580049375836466e-07, 'learning_rate/action_model': 5.580049375836466e-06, 'epoch': 1.92}) +[2026-06-16 12:18:38,144][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.10782959312200546, 'train/grad_norm_pre_clip': 15.873334884643555, 'timing/data': 0.03168315999209881, 'timing/model': 1.6619072328321636, 'learning_rate/qwen_vl_interface': 5.575314216961356e-07, 'learning_rate/action_model': 5.575314216961356e-06, 'epoch': 1.92}) +[2026-06-16 12:18:44,401][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.0165396761149168, 'train/grad_norm_pre_clip': 31.13828468322754, 'timing/data': 0.031073818914592266, 'timing/model': 1.77498962264508, 'learning_rate/qwen_vl_interface': 5.570598347052073e-07, 'learning_rate/action_model': 5.570598347052072e-06, 'epoch': 1.92}) +[2026-06-16 12:18:50,484][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.011063341051340103, 'train/grad_norm_pre_clip': 15.050065994262695, 'timing/data': 0.030890418216586113, 'timing/model': 1.555160925257951, 'learning_rate/qwen_vl_interface': 5.565901768047126e-07, 'learning_rate/action_model': 5.5659017680471255e-06, 'epoch': 1.92}) +[2026-06-16 12:18:56,768][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.0009927665814757347, 'train/grad_norm_pre_clip': 0.5552616119384766, 'timing/data': 0.03175620222464204, 'timing/model': 1.5391444116830826, 'learning_rate/qwen_vl_interface': 5.561224481877105e-07, 'learning_rate/action_model': 5.561224481877105e-06, 'epoch': 1.92}) +[2026-06-16 12:19:03,181][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.011078745126724243, 'train/grad_norm_pre_clip': 10.65696907043457, 'timing/data': 0.030716908164322376, 'timing/model': 1.6020426130853593, 'learning_rate/qwen_vl_interface': 5.55656649046467e-07, 'learning_rate/action_model': 5.556566490464669e-06, 'epoch': 1.93}) +[2026-06-16 12:19:09,595][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 0.1564376801252365, 'train/grad_norm_pre_clip': 10.999079704284668, 'timing/data': 0.030289977323263884, 'timing/model': 1.5899533946067095, 'learning_rate/qwen_vl_interface': 5.551927795724544e-07, 'learning_rate/action_model': 5.551927795724543e-06, 'epoch': 1.93}) +[2026-06-16 12:19:15,991][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.013934383168816566, 'train/grad_norm_pre_clip': 12.149062156677246, 'timing/data': 0.03205141285434365, 'timing/model': 1.6034820773638785, 'learning_rate/qwen_vl_interface': 5.547308399563515e-07, 'learning_rate/action_model': 5.547308399563514e-06, 'epoch': 1.93}) +[2026-06-16 12:19:22,410][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.14244522154331207, 'train/grad_norm_pre_clip': 23.806222915649414, 'timing/data': 0.03162671206519008, 'timing/model': 1.7710266401991248, 'learning_rate/qwen_vl_interface': 5.542708303880442e-07, 'learning_rate/action_model': 5.542708303880442e-06, 'epoch': 1.93}) +[2026-06-16 12:19:28,475][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.002806606236845255, 'train/grad_norm_pre_clip': 3.9310951232910156, 'timing/data': 0.03176575293764472, 'timing/model': 1.545165783725679, 'learning_rate/qwen_vl_interface': 5.538127510566258e-07, 'learning_rate/action_model': 5.538127510566258e-06, 'epoch': 1.93}) +[2026-06-16 12:19:34,871][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.002366820815950632, 'train/grad_norm_pre_clip': 1.949367880821228, 'timing/data': 0.03079140977934003, 'timing/model': 1.5435897968709469, 'learning_rate/qwen_vl_interface': 5.533566021503944e-07, 'learning_rate/action_model': 5.533566021503943e-06, 'epoch': 1.93}) +[2026-06-16 12:19:41,171][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 5.0589136662893e-05, 'train/grad_norm_pre_clip': 0.013646431267261505, 'timing/data': 0.03120914287865162, 'timing/model': 1.5371441030874848, 'learning_rate/qwen_vl_interface': 5.529023838568566e-07, 'learning_rate/action_model': 5.529023838568566e-06, 'epoch': 1.93}) +[2026-06-16 12:19:47,473][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.0004135750059504062, 'train/grad_norm_pre_clip': 0.09137837588787079, 'timing/data': 0.029725397936999798, 'timing/model': 1.5496947215870023, 'learning_rate/qwen_vl_interface': 5.524500963627239e-07, 'learning_rate/action_model': 5.524500963627239e-06, 'epoch': 1.93}) +[2026-06-16 12:19:53,866][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 0.0008500211406499147, 'train/grad_norm_pre_clip': 0.40372422337532043, 'timing/data': 0.030492120888084173, 'timing/model': 1.6001511462964118, 'learning_rate/qwen_vl_interface': 5.519997398539145e-07, 'learning_rate/action_model': 5.519997398539145e-06, 'epoch': 1.93}) +[2026-06-16 12:20:00,403][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.0013894648291170597, 'train/grad_norm_pre_clip': 1.5781663656234741, 'timing/data': 0.030139280017465353, 'timing/model': 1.8704813988879323, 'learning_rate/qwen_vl_interface': 5.515513145155531e-07, 'learning_rate/action_model': 5.515513145155531e-06, 'epoch': 1.93}) +[2026-06-16 12:20:06,473][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.001156975980848074, 'train/grad_norm_pre_clip': 0.3338625133037567, 'timing/data': 0.03136462438851595, 'timing/model': 1.5469970935955644, 'learning_rate/qwen_vl_interface': 5.511048205319702e-07, 'learning_rate/action_model': 5.511048205319701e-06, 'epoch': 1.93}) +[2026-06-16 12:20:12,879][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 0.01266686711460352, 'train/grad_norm_pre_clip': 10.816786766052246, 'timing/data': 0.032384087797254324, 'timing/model': 1.545775645878166, 'learning_rate/qwen_vl_interface': 5.506602580867034e-07, 'learning_rate/action_model': 5.506602580867034e-06, 'epoch': 1.93}) +[2026-06-16 12:20:19,232][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.03781410679221153, 'train/grad_norm_pre_clip': 25.05398178100586, 'timing/data': 0.031936067156493664, 'timing/model': 1.5922837378457189, 'learning_rate/qwen_vl_interface': 5.502176273624942e-07, 'learning_rate/action_model': 5.502176273624942e-06, 'epoch': 1.93}) +[2026-06-16 12:20:25,573][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.0007189925527200103, 'train/grad_norm_pre_clip': 0.16721147298812866, 'timing/data': 0.02977703884243965, 'timing/model': 1.5602310081012547, 'learning_rate/qwen_vl_interface': 5.497769285412921e-07, 'learning_rate/action_model': 5.497769285412921e-06, 'epoch': 1.93}) +[2026-06-16 12:20:31,965][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.2687547206878662, 'train/grad_norm_pre_clip': 24.797874450683594, 'timing/data': 0.030886254273355007, 'timing/model': 1.6007795496843755, 'learning_rate/qwen_vl_interface': 5.493381618042519e-07, 'learning_rate/action_model': 5.493381618042518e-06, 'epoch': 1.93}) +[2026-06-16 12:20:38,339][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.00023196006077341735, 'train/grad_norm_pre_clip': 0.05810830742120743, 'timing/data': 0.0302263917401433, 'timing/model': 1.8044300340116024, 'learning_rate/qwen_vl_interface': 5.489013273317343e-07, 'learning_rate/action_model': 5.489013273317343e-06, 'epoch': 1.93}) +[2026-06-16 12:20:44,640][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.0007055437308736145, 'train/grad_norm_pre_clip': 0.12317272275686264, 'timing/data': 0.03101722477003932, 'timing/model': 1.8003929280675948, 'learning_rate/qwen_vl_interface': 5.484664253033042e-07, 'learning_rate/action_model': 5.484664253033042e-06, 'epoch': 1.93}) +[2026-06-16 12:20:50,670][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 9.419964044354856e-05, 'train/grad_norm_pre_clip': 0.038503117859363556, 'timing/data': 0.03009664174169302, 'timing/model': 1.5434784130193293, 'learning_rate/qwen_vl_interface': 5.480334558977351e-07, 'learning_rate/action_model': 5.480334558977351e-06, 'epoch': 1.93}) +[2026-06-16 12:20:57,052][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.0869433730840683, 'train/grad_norm_pre_clip': 51.9864501953125, 'timing/data': 0.03077018540352583, 'timing/model': 1.6194823272526264, 'learning_rate/qwen_vl_interface': 5.476024192930029e-07, 'learning_rate/action_model': 5.47602419293003e-06, 'epoch': 1.93}) +[2026-06-16 12:21:03,276][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.317341685295105, 'train/grad_norm_pre_clip': 2.5750091075897217, 'timing/data': 0.031103435903787613, 'timing/model': 1.5478873578831553, 'learning_rate/qwen_vl_interface': 5.471733156662914e-07, 'learning_rate/action_model': 5.4717331566629135e-06, 'epoch': 1.93}) +[2026-06-16 12:21:09,650][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 4.206282028462738e-05, 'train/grad_norm_pre_clip': 0.0030056005343794823, 'timing/data': 0.02931339107453823, 'timing/model': 1.5875620390288532, 'learning_rate/qwen_vl_interface': 5.467461451939882e-07, 'learning_rate/action_model': 5.467461451939882e-06, 'epoch': 1.93}) +[2026-06-16 12:21:15,900][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.0005694410065189004, 'train/grad_norm_pre_clip': 0.2854979932308197, 'timing/data': 0.029247931204736233, 'timing/model': 1.771187933627516, 'learning_rate/qwen_vl_interface': 5.463209080516873e-07, 'learning_rate/action_model': 5.463209080516873e-06, 'epoch': 1.93}) +[2026-06-16 12:21:22,207][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.02272525057196617, 'train/grad_norm_pre_clip': 24.838083267211914, 'timing/data': 0.030524394940584898, 'timing/model': 1.78073290316388, 'learning_rate/qwen_vl_interface': 5.458976044141875e-07, 'learning_rate/action_model': 5.458976044141875e-06, 'epoch': 1.93}) +[2026-06-16 12:21:28,265][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.00030410458566620946, 'train/grad_norm_pre_clip': 0.07232465595006943, 'timing/data': 0.030403415206819773, 'timing/model': 1.5329877138137817, 'learning_rate/qwen_vl_interface': 5.454762344554931e-07, 'learning_rate/action_model': 5.4547623445549316e-06, 'epoch': 1.93}) +[2026-06-16 12:21:34,567][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 0.0006977802840992808, 'train/grad_norm_pre_clip': 0.1673959493637085, 'timing/data': 0.028917270246893167, 'timing/model': 1.5313556720502675, 'learning_rate/qwen_vl_interface': 5.450567983488137e-07, 'learning_rate/action_model': 5.450567983488137e-06, 'epoch': 1.93}) +[2026-06-16 12:21:40,963][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.12061833590269089, 'train/grad_norm_pre_clip': 18.786861419677734, 'timing/data': 0.0315793096087873, 'timing/model': 1.5455196551047266, 'learning_rate/qwen_vl_interface': 5.446392962665629e-07, 'learning_rate/action_model': 5.4463929626656296e-06, 'epoch': 1.94}) +[2026-06-16 12:21:47,354][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.00048008898738771677, 'train/grad_norm_pre_clip': 0.10172943025827408, 'timing/data': 0.0291447420604527, 'timing/model': 1.5893219923600554, 'learning_rate/qwen_vl_interface': 5.442237283803605e-07, 'learning_rate/action_model': 5.442237283803604e-06, 'epoch': 1.94}) +[2026-06-16 12:21:53,708][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.041632894426584244, 'train/grad_norm_pre_clip': 48.189605712890625, 'timing/data': 0.03198079112917185, 'timing/model': 1.7728040562942624, 'learning_rate/qwen_vl_interface': 5.43810094861031e-07, 'learning_rate/action_model': 5.43810094861031e-06, 'epoch': 1.94}) +[2026-06-16 12:21:59,767][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.0004218778049107641, 'train/grad_norm_pre_clip': 0.09730199724435806, 'timing/data': 0.02994268387556076, 'timing/model': 1.5450070137158036, 'learning_rate/qwen_vl_interface': 5.433983958786027e-07, 'learning_rate/action_model': 5.433983958786027e-06, 'epoch': 1.94}) +[2026-06-16 12:22:06,118][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.08231353014707565, 'train/grad_norm_pre_clip': 38.50172805786133, 'timing/data': 0.03253079392015934, 'timing/model': 1.5792033369652927, 'learning_rate/qwen_vl_interface': 5.429886316023102e-07, 'learning_rate/action_model': 5.429886316023101e-06, 'epoch': 1.94}) +[2026-06-16 12:22:12,486][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.001957810251042247, 'train/grad_norm_pre_clip': 1.3751472234725952, 'timing/data': 0.030718276742845774, 'timing/model': 1.5560247390531003, 'learning_rate/qwen_vl_interface': 5.425808022005916e-07, 'learning_rate/action_model': 5.425808022005916e-06, 'epoch': 1.94}) +[2026-06-16 12:22:18,872][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.00019301891734357923, 'train/grad_norm_pre_clip': 0.03215749189257622, 'timing/data': 0.030050174798816442, 'timing/model': 1.604999974835664, 'learning_rate/qwen_vl_interface': 5.421749078410903e-07, 'learning_rate/action_model': 5.421749078410903e-06, 'epoch': 1.94}) +[2026-06-16 12:22:25,168][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.06935467571020126, 'train/grad_norm_pre_clip': 41.14583969116211, 'timing/data': 0.030457267072051764, 'timing/model': 1.5583708826452494, 'learning_rate/qwen_vl_interface': 5.417709486906547e-07, 'learning_rate/action_model': 5.417709486906547e-06, 'epoch': 1.94}) +[2026-06-16 12:22:31,478][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.0032538825180381536, 'train/grad_norm_pre_clip': 2.5449137687683105, 'timing/data': 0.030482888221740723, 'timing/model': 1.73943893564865, 'learning_rate/qwen_vl_interface': 5.413689249153367e-07, 'learning_rate/action_model': 5.413689249153366e-06, 'epoch': 1.94}) +[2026-06-16 12:22:37,580][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.0007401927723549306, 'train/grad_norm_pre_clip': 0.3356825113296509, 'timing/data': 0.035098003689199686, 'timing/model': 1.5424508322030306, 'learning_rate/qwen_vl_interface': 5.409688366803931e-07, 'learning_rate/action_model': 5.409688366803931e-06, 'epoch': 1.94}) +[2026-06-16 12:22:43,882][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.14201980829238892, 'train/grad_norm_pre_clip': 15.962409973144531, 'timing/data': 0.03091843007132411, 'timing/model': 1.5500842100009322, 'learning_rate/qwen_vl_interface': 5.405706841502853e-07, 'learning_rate/action_model': 5.4057068415028525e-06, 'epoch': 1.94}) +[2026-06-16 12:22:50,269][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 0.014730610884726048, 'train/grad_norm_pre_clip': 11.077585220336914, 'timing/data': 0.029363885056227446, 'timing/model': 1.5716051501221955, 'learning_rate/qwen_vl_interface': 5.401744674886781e-07, 'learning_rate/action_model': 5.40174467488678e-06, 'epoch': 1.94}) +[2026-06-16 12:22:56,669][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.0202688779681921, 'train/grad_norm_pre_clip': 19.892932891845703, 'timing/data': 0.030493948608636856, 'timing/model': 1.5939205209724605, 'learning_rate/qwen_vl_interface': 5.397801868584418e-07, 'learning_rate/action_model': 5.397801868584417e-06, 'epoch': 1.94}) +[2026-06-16 12:23:03,073][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.0007832329720258713, 'train/grad_norm_pre_clip': 0.15432222187519073, 'timing/data': 0.030997311230748892, 'timing/model': 1.5901457709260285, 'learning_rate/qwen_vl_interface': 5.393878424216502e-07, 'learning_rate/action_model': 5.393878424216501e-06, 'epoch': 1.94}) +[2026-06-16 12:33:40,035][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.00033553101820871234, 'train/grad_norm_pre_clip': 0.1021108478307724, 'eval/loss': 0.034816961735486984, 'eval/flap_precision': 0.9715447154471545, 'eval/flap_recall': 0.9781718963165075, 'eval/flap_f1': 0.974847042828008, 'eval/flap_rate_model': 0.205, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.974847042828008, 'timing/data': 0.03416654234752059, 'timing/model': 1.7196202366612852, 'learning_rate/qwen_vl_interface': 5.389974343395813e-07, 'learning_rate/action_model': 5.389974343395812e-06, 'epoch': 1.94}) +[2026-06-16 12:33:40,036][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4800_state +[2026-06-16 12:33:40,042][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 12:33:48,382][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4800_state/model.safetensors +[2026-06-16 12:34:07,425][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4800_state/optimizer.bin +[2026-06-16 12:34:07,427][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4800_state/sampler.bin +[2026-06-16 12:34:07,427][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4800_state/sampler_1.bin +[2026-06-16 12:34:07,429][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4800_state/random_states_0.pkl +[2026-06-16 12:34:07,615][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4800 +[2026-06-16 12:34:07,616][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 12:34:07,632][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 12:34:13,895][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 0.03922504559159279, 'train/grad_norm_pre_clip': 19.34540367126465, 'timing/data': 0.032177799846976995, 'timing/model': 1.792742501012981, 'learning_rate/qwen_vl_interface': 5.38608962772717e-07, 'learning_rate/action_model': 5.38608962772717e-06, 'epoch': 1.94}) +[2026-06-16 12:34:19,940][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.008994425646960735, 'train/grad_norm_pre_clip': 7.277216911315918, 'timing/data': 0.03167403815314174, 'timing/model': 1.5618724138475955, 'learning_rate/qwen_vl_interface': 5.382224278807429e-07, 'learning_rate/action_model': 5.382224278807429e-06, 'epoch': 1.94}) +[2026-06-16 12:34:26,134][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.0018197351600974798, 'train/grad_norm_pre_clip': 1.2371807098388672, 'timing/data': 0.03208862990140915, 'timing/model': 1.5360850328579545, 'learning_rate/qwen_vl_interface': 5.378378298225496e-07, 'learning_rate/action_model': 5.378378298225496e-06, 'epoch': 1.94}) +[2026-06-16 12:34:32,388][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.0011138245463371277, 'train/grad_norm_pre_clip': 0.3492870330810547, 'timing/data': 0.029693119693547487, 'timing/model': 1.5774441729299724, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 1.94}) +[2026-06-16 12:34:38,540][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.14558543264865875, 'train/grad_norm_pre_clip': 43.0573616027832, 'timing/data': 0.030591643881052732, 'timing/model': 1.5329742468893528, 'learning_rate/qwen_vl_interface': 5.37074444839083e-07, 'learning_rate/action_model': 5.370744448390829e-06, 'epoch': 1.94}) +[2026-06-16 12:34:44,746][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 0.01806204952299595, 'train/grad_norm_pre_clip': 20.5695743560791, 'timing/data': 0.030474423430860043, 'timing/model': 1.5490787751041353, 'learning_rate/qwen_vl_interface': 5.366956582276085e-07, 'learning_rate/action_model': 5.366956582276084e-06, 'epoch': 1.94}) +[2026-06-16 12:34:51,010][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.0014495672658085823, 'train/grad_norm_pre_clip': 0.8088374733924866, 'timing/data': 0.03120932588353753, 'timing/model': 1.7914243671111763, 'learning_rate/qwen_vl_interface': 5.363188090775122e-07, 'learning_rate/action_model': 5.363188090775122e-06, 'epoch': 1.94}) +[2026-06-16 12:34:57,048][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.00014998321421444416, 'train/grad_norm_pre_clip': 0.02990960329771042, 'timing/data': 0.030114241410046816, 'timing/model': 1.5363475787453353, 'learning_rate/qwen_vl_interface': 5.359438975437018e-07, 'learning_rate/action_model': 5.359438975437017e-06, 'epoch': 1.94}) +[2026-06-16 12:35:03,345][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.0008649517549201846, 'train/grad_norm_pre_clip': 0.18952584266662598, 'timing/data': 0.030874984338879585, 'timing/model': 1.5333555000834167, 'learning_rate/qwen_vl_interface': 5.355709237802895e-07, 'learning_rate/action_model': 5.3557092378028955e-06, 'epoch': 1.94}) +[2026-06-16 12:35:09,555][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.00012730715388897806, 'train/grad_norm_pre_clip': 0.015840765088796616, 'timing/data': 0.03093255590647459, 'timing/model': 1.5441184961237013, 'learning_rate/qwen_vl_interface': 5.351998879405911e-07, 'learning_rate/action_model': 5.351998879405911e-06, 'epoch': 1.95}) +[2026-06-16 12:35:15,890][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.0012048360658809543, 'train/grad_norm_pre_clip': 1.1175956726074219, 'timing/data': 0.0306499139405787, 'timing/model': 1.5380158540792763, 'learning_rate/qwen_vl_interface': 5.348307901771256e-07, 'learning_rate/action_model': 5.348307901771256e-06, 'epoch': 1.95}) +[2026-06-16 12:35:22,151][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.00013614626368507743, 'train/grad_norm_pre_clip': 0.021963080391287804, 'timing/data': 0.0291061969473958, 'timing/model': 1.553900911938399, 'learning_rate/qwen_vl_interface': 5.344636306416147e-07, 'learning_rate/action_model': 5.344636306416146e-06, 'epoch': 1.95}) +[2026-06-16 12:35:28,486][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.00017381079669576138, 'train/grad_norm_pre_clip': 0.019857337698340416, 'timing/data': 0.03020630218088627, 'timing/model': 1.774419796653092, 'learning_rate/qwen_vl_interface': 5.340984094849836e-07, 'learning_rate/action_model': 5.340984094849836e-06, 'epoch': 1.95}) +[2026-06-16 12:35:34,549][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.0007510986179113388, 'train/grad_norm_pre_clip': 0.5079560875892639, 'timing/data': 0.029757529962807894, 'timing/model': 1.5409039440564811, 'learning_rate/qwen_vl_interface': 5.337351268573606e-07, 'learning_rate/action_model': 5.337351268573606e-06, 'epoch': 1.95}) +[2026-06-16 12:35:40,855][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.00021763583936262876, 'train/grad_norm_pre_clip': 0.06684403866529465, 'timing/data': 0.02991111110895872, 'timing/model': 1.5432287156581879, 'learning_rate/qwen_vl_interface': 5.33373782908079e-07, 'learning_rate/action_model': 5.333737829080789e-06, 'epoch': 1.95}) +[2026-06-16 12:35:47,153][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.00029067721334286034, 'train/grad_norm_pre_clip': 0.06909514963626862, 'timing/data': 0.029726309701800346, 'timing/model': 1.5413504620082676, 'learning_rate/qwen_vl_interface': 5.330143777856718e-07, 'learning_rate/action_model': 5.330143777856718e-06, 'epoch': 1.95}) +[2026-06-16 12:35:53,454][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 0.1558809131383896, 'train/grad_norm_pre_clip': 51.720726013183594, 'timing/data': 0.02897508814930916, 'timing/model': 1.544749406632036, 'learning_rate/qwen_vl_interface': 5.326569116378786e-07, 'learning_rate/action_model': 5.326569116378786e-06, 'epoch': 1.95}) +[2026-06-16 12:35:59,770][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.03673141822218895, 'train/grad_norm_pre_clip': 34.510154724121094, 'timing/data': 0.03045969409868121, 'timing/model': 1.5555994440801442, 'learning_rate/qwen_vl_interface': 5.323013846116387e-07, 'learning_rate/action_model': 5.323013846116386e-06, 'epoch': 1.95}) +[2026-06-16 12:36:06,091][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.0007503155502490699, 'train/grad_norm_pre_clip': 0.17754633724689484, 'timing/data': 0.029817011207342148, 'timing/model': 1.7738574827089906, 'learning_rate/qwen_vl_interface': 5.31947796853097e-07, 'learning_rate/action_model': 5.3194779685309696e-06, 'epoch': 1.95}) +[2026-06-16 12:36:12,157][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.008640785701572895, 'train/grad_norm_pre_clip': 10.348725318908691, 'timing/data': 0.030660293996334076, 'timing/model': 1.530913928989321, 'learning_rate/qwen_vl_interface': 5.315961485075992e-07, 'learning_rate/action_model': 5.315961485075993e-06, 'epoch': 1.95}) +[2026-06-16 12:36:18,504][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.013760819099843502, 'train/grad_norm_pre_clip': 17.009735107421875, 'timing/data': 0.03823428601026535, 'timing/model': 1.5731203532777727, 'learning_rate/qwen_vl_interface': 5.312464397196958e-07, 'learning_rate/action_model': 5.312464397196958e-06, 'epoch': 1.95}) +[2026-06-16 12:36:24,876][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.00013955560280010104, 'train/grad_norm_pre_clip': 0.026511920616030693, 'timing/data': 0.029458390083163977, 'timing/model': 1.5471791187301278, 'learning_rate/qwen_vl_interface': 5.308986706331377e-07, 'learning_rate/action_model': 5.308986706331376e-06, 'epoch': 1.95}) +[2026-06-16 12:36:31,174][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.008968341164290905, 'train/grad_norm_pre_clip': 5.399693965911865, 'timing/data': 0.030156472697854042, 'timing/model': 1.5308983461000025, 'learning_rate/qwen_vl_interface': 5.305528413908797e-07, 'learning_rate/action_model': 5.305528413908798e-06, 'epoch': 1.95}) +[2026-06-16 12:36:37,459][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.00021413106878753752, 'train/grad_norm_pre_clip': 0.07724609225988388, 'timing/data': 0.029228538740426302, 'timing/model': 1.530901727732271, 'learning_rate/qwen_vl_interface': 5.302089521350799e-07, 'learning_rate/action_model': 5.3020895213507996e-06, 'epoch': 1.95}) +[2026-06-16 12:36:43,802][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.00013104345998726785, 'train/grad_norm_pre_clip': 0.020407317206263542, 'timing/data': 0.029604410752654076, 'timing/model': 1.769578142091632, 'learning_rate/qwen_vl_interface': 5.298670030070977e-07, 'learning_rate/action_model': 5.298670030070976e-06, 'epoch': 1.95}) +[2026-06-16 12:36:49,874][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.08018683642148972, 'train/grad_norm_pre_clip': 23.25558090209961, 'timing/data': 0.0314801880158484, 'timing/model': 1.5515638077631593, 'learning_rate/qwen_vl_interface': 5.295269941474957e-07, 'learning_rate/action_model': 5.295269941474957e-06, 'epoch': 1.95}) +[2026-06-16 12:36:56,168][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.0005648548249155283, 'train/grad_norm_pre_clip': 0.11420930922031403, 'timing/data': 0.031081646215170622, 'timing/model': 1.5382161550223827, 'learning_rate/qwen_vl_interface': 5.291889256960384e-07, 'learning_rate/action_model': 5.291889256960384e-06, 'epoch': 1.95}) +[2026-06-16 12:37:02,476][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.011701169423758984, 'train/grad_norm_pre_clip': 6.23010778427124, 'timing/data': 0.03351618675515056, 'timing/model': 1.55124615970999, 'learning_rate/qwen_vl_interface': 5.288527977916929e-07, 'learning_rate/action_model': 5.288527977916929e-06, 'epoch': 1.95}) +[2026-06-16 12:37:08,764][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.003042350523173809, 'train/grad_norm_pre_clip': 1.5542446374893188, 'timing/data': 0.030700325034558773, 'timing/model': 1.5381419388577342, 'learning_rate/qwen_vl_interface': 5.28518610572629e-07, 'learning_rate/action_model': 5.28518610572629e-06, 'epoch': 1.95}) +[2026-06-16 12:37:15,153][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.10925990343093872, 'train/grad_norm_pre_clip': 22.222599029541016, 'timing/data': 0.03211151110008359, 'timing/model': 1.5891178101301193, 'learning_rate/qwen_vl_interface': 5.281863641762184e-07, 'learning_rate/action_model': 5.281863641762184e-06, 'epoch': 1.95}) +[2026-06-16 12:37:21,500][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 0.0010504706297069788, 'train/grad_norm_pre_clip': 0.3968428373336792, 'timing/data': 0.03146326821297407, 'timing/model': 1.7774402000941336, 'learning_rate/qwen_vl_interface': 5.278560587390355e-07, 'learning_rate/action_model': 5.278560587390355e-06, 'epoch': 1.95}) +[2026-06-16 12:37:27,751][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.010497311130166054, 'train/grad_norm_pre_clip': 11.83304214477539, 'timing/data': 0.029121977742761374, 'timing/model': 1.5386541783809662, 'learning_rate/qwen_vl_interface': 5.275276943968557e-07, 'learning_rate/action_model': 5.275276943968557e-06, 'epoch': 1.95}) +[2026-06-16 12:37:34,253][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.0025125169195234776, 'train/grad_norm_pre_clip': 3.189319372177124, 'timing/data': 0.03535775374621153, 'timing/model': 1.6283585862256587, 'learning_rate/qwen_vl_interface': 5.272012712846566e-07, 'learning_rate/action_model': 5.2720127128465654e-06, 'epoch': 1.95}) +[2026-06-16 12:37:40,483][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.002922798739746213, 'train/grad_norm_pre_clip': 2.869694232940674, 'timing/data': 0.030174083076417446, 'timing/model': 1.5425832886248827, 'learning_rate/qwen_vl_interface': 5.268767895366198e-07, 'learning_rate/action_model': 5.268767895366198e-06, 'epoch': 1.95}) +[2026-06-16 12:37:46,871][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.15044143795967102, 'train/grad_norm_pre_clip': 13.469079971313477, 'timing/data': 0.02949209976941347, 'timing/model': 1.5497160400263965, 'learning_rate/qwen_vl_interface': 5.265542492861268e-07, 'learning_rate/action_model': 5.2655424928612676e-06, 'epoch': 1.96}) +[2026-06-16 12:37:53,249][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 0.02016127109527588, 'train/grad_norm_pre_clip': 20.72551155090332, 'timing/data': 0.02914272015914321, 'timing/model': 1.5960973212495446, 'learning_rate/qwen_vl_interface': 5.262336506657617e-07, 'learning_rate/action_model': 5.2623365066576165e-06, 'epoch': 1.96}) +[2026-06-16 12:37:59,595][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.00820157490670681, 'train/grad_norm_pre_clip': 17.669246673583984, 'timing/data': 0.03160950867459178, 'timing/model': 1.766705029644072, 'learning_rate/qwen_vl_interface': 5.259149938073103e-07, 'learning_rate/action_model': 5.259149938073102e-06, 'epoch': 1.96}) +[2026-06-16 12:38:05,774][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.2175700068473816, 'train/grad_norm_pre_clip': 8.555014610290527, 'timing/data': 0.03188205091282725, 'timing/model': 1.537042303942144, 'learning_rate/qwen_vl_interface': 5.255982788417609e-07, 'learning_rate/action_model': 5.255982788417608e-06, 'epoch': 1.96}) +[2026-06-16 12:38:12,174][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.10327103734016418, 'train/grad_norm_pre_clip': 72.23303985595703, 'timing/data': 0.030774585902690887, 'timing/model': 1.5493941460736096, 'learning_rate/qwen_vl_interface': 5.252835058993029e-07, 'learning_rate/action_model': 5.252835058993029e-06, 'epoch': 1.96}) +[2026-06-16 12:38:18,472][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.00533326854929328, 'train/grad_norm_pre_clip': 6.710999011993408, 'timing/data': 0.02978669200092554, 'timing/model': 1.5442467080429196, 'learning_rate/qwen_vl_interface': 5.249706751093274e-07, 'learning_rate/action_model': 5.249706751093273e-06, 'epoch': 1.96}) +[2026-06-16 12:38:24,772][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 0.03369741886854172, 'train/grad_norm_pre_clip': 28.24899673461914, 'timing/data': 0.0289286975748837, 'timing/model': 1.5480709848925471, 'learning_rate/qwen_vl_interface': 5.246597866004273e-07, 'learning_rate/action_model': 5.246597866004273e-06, 'epoch': 1.96}) +[2026-06-16 12:38:31,274][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.00022008191444911063, 'train/grad_norm_pre_clip': 0.07359582930803299, 'timing/data': 0.030370404943823814, 'timing/model': 1.5807375609874725, 'learning_rate/qwen_vl_interface': 5.243508405003975e-07, 'learning_rate/action_model': 5.243508405003975e-06, 'epoch': 1.96}) +[2026-06-16 12:38:37,733][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.001517820986919105, 'train/grad_norm_pre_clip': 1.1766332387924194, 'timing/data': 0.030355724040418863, 'timing/model': 1.8095071092247963, 'learning_rate/qwen_vl_interface': 5.240438369362336e-07, 'learning_rate/action_model': 5.2404383693623366e-06, 'epoch': 1.96}) +[2026-06-16 12:38:43,871][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.21457821130752563, 'train/grad_norm_pre_clip': 38.60531997680664, 'timing/data': 0.030744384974241257, 'timing/model': 1.529685806017369, 'learning_rate/qwen_vl_interface': 5.237387760341335e-07, 'learning_rate/action_model': 5.237387760341335e-06, 'epoch': 1.96}) +[2026-06-16 12:38:50,198][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.004701741971075535, 'train/grad_norm_pre_clip': 3.7362804412841797, 'timing/data': 0.03103006724268198, 'timing/model': 1.5719185299240053, 'learning_rate/qwen_vl_interface': 5.234356579194963e-07, 'learning_rate/action_model': 5.234356579194963e-06, 'epoch': 1.96}) +[2026-06-16 12:38:56,481][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.003364155301824212, 'train/grad_norm_pre_clip': 1.5592799186706543, 'timing/data': 0.03206649189814925, 'timing/model': 1.545130123384297, 'learning_rate/qwen_vl_interface': 5.231344827169218e-07, 'learning_rate/action_model': 5.231344827169217e-06, 'epoch': 1.96}) +[2026-06-16 12:39:02,869][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.0705585926771164, 'train/grad_norm_pre_clip': 55.820960998535156, 'timing/data': 0.0316386129707098, 'timing/model': 1.5299956682138145, 'learning_rate/qwen_vl_interface': 5.228352505502119e-07, 'learning_rate/action_model': 5.228352505502119e-06, 'epoch': 1.96}) +[2026-06-16 12:39:09,272][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.000514486338943243, 'train/grad_norm_pre_clip': 0.07139294594526291, 'timing/data': 0.031117713078856468, 'timing/model': 1.579449811950326, 'learning_rate/qwen_vl_interface': 5.225379615423701e-07, 'learning_rate/action_model': 5.2253796154237e-06, 'epoch': 1.96}) +[2026-06-16 12:39:15,625][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.21825221180915833, 'train/grad_norm_pre_clip': 54.30305862426758, 'timing/data': 0.03135626297444105, 'timing/model': 1.7929593874141574, 'learning_rate/qwen_vl_interface': 5.222426158156008e-07, 'learning_rate/action_model': 5.222426158156008e-06, 'epoch': 1.96}) +[2026-06-16 12:39:21,881][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.1000981479883194, 'train/grad_norm_pre_clip': 14.052103042602539, 'timing/data': 0.03086753375828266, 'timing/model': 1.75015272712335, 'learning_rate/qwen_vl_interface': 5.219492134913084e-07, 'learning_rate/action_model': 5.219492134913084e-06, 'epoch': 1.96}) +[2026-06-16 12:39:27,972][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.08323963731527328, 'train/grad_norm_pre_clip': 36.14960861206055, 'timing/data': 0.029436721932142973, 'timing/model': 1.5373765439726412, 'learning_rate/qwen_vl_interface': 5.216577546901008e-07, 'learning_rate/action_model': 5.216577546901007e-06, 'epoch': 1.96}) +[2026-06-16 12:39:34,290][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 0.1355224996805191, 'train/grad_norm_pre_clip': 14.682045936584473, 'timing/data': 0.03101588599383831, 'timing/model': 1.5388061599805951, 'learning_rate/qwen_vl_interface': 5.213682395317852e-07, 'learning_rate/action_model': 5.213682395317851e-06, 'epoch': 1.96}) +[2026-06-16 12:39:40,765][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 0.0427238866686821, 'train/grad_norm_pre_clip': 22.273208618164062, 'timing/data': 0.030009846203029156, 'timing/model': 1.6296753203496337, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 1.96}) +[2026-06-16 12:39:47,146][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.2899586260318756, 'train/grad_norm_pre_clip': 43.33382034301758, 'timing/data': 0.03027384588494897, 'timing/model': 1.591356955934316, 'learning_rate/qwen_vl_interface': 5.207950406190654e-07, 'learning_rate/action_model': 5.207950406190653e-06, 'epoch': 1.96}) +[2026-06-16 12:39:53,407][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.00030970474472269416, 'train/grad_norm_pre_clip': 0.10579574108123779, 'timing/data': 0.030187548138201237, 'timing/model': 1.7838821900077164, 'learning_rate/qwen_vl_interface': 5.205113571002818e-07, 'learning_rate/action_model': 5.205113571002818e-06, 'epoch': 1.96}) +[2026-06-16 12:39:59,702][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 9.810969640966505e-05, 'train/grad_norm_pre_clip': 0.018676305189728737, 'timing/data': 0.034657774958759546, 'timing/model': 1.7689566239714622, 'learning_rate/qwen_vl_interface': 5.202296176956308e-07, 'learning_rate/action_model': 5.202296176956308e-06, 'epoch': 1.96}) +[2026-06-16 12:40:05,782][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 0.001533976523205638, 'train/grad_norm_pre_clip': 0.32876479625701904, 'timing/data': 0.030147959012538195, 'timing/model': 1.5555565627291799, 'learning_rate/qwen_vl_interface': 5.19949822520925e-07, 'learning_rate/action_model': 5.19949822520925e-06, 'epoch': 1.96}) +[2026-06-16 12:40:12,163][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.03298778086900711, 'train/grad_norm_pre_clip': 31.21449089050293, 'timing/data': 0.02972804894670844, 'timing/model': 1.5318979951553047, 'learning_rate/qwen_vl_interface': 5.196719716911777e-07, 'learning_rate/action_model': 5.1967197169117764e-06, 'epoch': 1.96}) +[2026-06-16 12:40:18,466][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.024478517472743988, 'train/grad_norm_pre_clip': 14.140674591064453, 'timing/data': 0.030284580774605274, 'timing/model': 1.5348306386731565, 'learning_rate/qwen_vl_interface': 5.193960653206029e-07, 'learning_rate/action_model': 5.193960653206029e-06, 'epoch': 1.96}) +[2026-06-16 12:40:24,854][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.009641998447477818, 'train/grad_norm_pre_clip': 10.450386047363281, 'timing/data': 0.029312558937817812, 'timing/model': 1.5912807951681316, 'learning_rate/qwen_vl_interface': 5.191221035226144e-07, 'learning_rate/action_model': 5.191221035226144e-06, 'epoch': 1.97}) +[2026-06-16 12:40:31,211][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.010595234110951424, 'train/grad_norm_pre_clip': 22.07191276550293, 'timing/data': 0.03090647328644991, 'timing/model': 1.7865952411666512, 'learning_rate/qwen_vl_interface': 5.188500864098292e-07, 'learning_rate/action_model': 5.1885008640982915e-06, 'epoch': 1.97}) +[2026-06-16 12:40:37,272][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.0003345968434587121, 'train/grad_norm_pre_clip': 0.08726245164871216, 'timing/data': 0.03186678607016802, 'timing/model': 1.5408143638633192, 'learning_rate/qwen_vl_interface': 5.185800140940613e-07, 'learning_rate/action_model': 5.185800140940613e-06, 'epoch': 1.97}) +[2026-06-16 12:40:43,672][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 7.494273449992761e-05, 'train/grad_norm_pre_clip': 0.01697264239192009, 'timing/data': 0.032199567183852196, 'timing/model': 1.5429996466264129, 'learning_rate/qwen_vl_interface': 5.183118866863288e-07, 'learning_rate/action_model': 5.1831188668632875e-06, 'epoch': 1.97}) +[2026-06-16 12:40:50,068][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.00019652448827400804, 'train/grad_norm_pre_clip': 0.03298584371805191, 'timing/data': 0.029418762773275375, 'timing/model': 1.5359693812206388, 'learning_rate/qwen_vl_interface': 5.180457042968484e-07, 'learning_rate/action_model': 5.1804570429684835e-06, 'epoch': 1.97}) +[2026-06-16 12:40:56,468][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.019101131707429886, 'train/grad_norm_pre_clip': 21.829322814941406, 'timing/data': 0.033160889986902475, 'timing/model': 1.560537772718817, 'learning_rate/qwen_vl_interface': 5.17781467035037e-07, 'learning_rate/action_model': 5.177814670350371e-06, 'epoch': 1.97}) +[2026-06-16 12:41:02,869][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.0007053918670862913, 'train/grad_norm_pre_clip': 0.13947643339633942, 'timing/data': 0.03447050368413329, 'timing/model': 1.5978349382057786, 'learning_rate/qwen_vl_interface': 5.175191750095133e-07, 'learning_rate/action_model': 5.175191750095132e-06, 'epoch': 1.97}) +[2026-06-16 12:41:09,151][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.0584878996014595, 'train/grad_norm_pre_clip': 14.181107521057129, 'timing/data': 0.031231768894940615, 'timing/model': 1.7244226806797087, 'learning_rate/qwen_vl_interface': 5.172588283280953e-07, 'learning_rate/action_model': 5.172588283280953e-06, 'epoch': 1.97}) +[2026-06-16 12:41:15,347][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.16399900615215302, 'train/grad_norm_pre_clip': 23.67450523376465, 'timing/data': 0.030282527208328247, 'timing/model': 1.5817624521441758, 'learning_rate/qwen_vl_interface': 5.170004270978019e-07, 'learning_rate/action_model': 5.170004270978018e-06, 'epoch': 1.97}) +[2026-06-16 12:41:21,666][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 0.08374771475791931, 'train/grad_norm_pre_clip': 34.922027587890625, 'timing/data': 0.030541528947651386, 'timing/model': 1.5429680431261659, 'learning_rate/qwen_vl_interface': 5.167439714248512e-07, 'learning_rate/action_model': 5.1674397142485116e-06, 'epoch': 1.97}) +[2026-06-16 12:41:28,062][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.020681438967585564, 'train/grad_norm_pre_clip': 9.939043045043945, 'timing/data': 0.029816027265042067, 'timing/model': 1.5450572823174298, 'learning_rate/qwen_vl_interface': 5.164894614146633e-07, 'learning_rate/action_model': 5.164894614146634e-06, 'epoch': 1.97}) +[2026-06-16 12:41:34,450][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.00036666501546278596, 'train/grad_norm_pre_clip': 0.054510775953531265, 'timing/data': 0.030173048842698336, 'timing/model': 1.590437576174736, 'learning_rate/qwen_vl_interface': 5.16236897171858e-07, 'learning_rate/action_model': 5.16236897171858e-06, 'epoch': 1.97}) +[2026-06-16 12:41:40,852][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.00034434563713148236, 'train/grad_norm_pre_clip': 0.06373920291662216, 'timing/data': 0.029197268187999725, 'timing/model': 1.5931004858575761, 'learning_rate/qwen_vl_interface': 5.15986278800254e-07, 'learning_rate/action_model': 5.15986278800254e-06, 'epoch': 1.97}) +[2026-06-16 12:41:47,168][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.022749511525034904, 'train/grad_norm_pre_clip': 18.494293212890625, 'timing/data': 0.03158887289464474, 'timing/model': 1.7323019681498408, 'learning_rate/qwen_vl_interface': 5.157376064028719e-07, 'learning_rate/action_model': 5.157376064028718e-06, 'epoch': 1.97}) +[2026-06-16 12:41:53,375][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.16951465606689453, 'train/grad_norm_pre_clip': 5.745818138122559, 'timing/data': 0.03269565664231777, 'timing/model': 1.5421655899845064, 'learning_rate/qwen_vl_interface': 5.154908800819307e-07, 'learning_rate/action_model': 5.154908800819306e-06, 'epoch': 1.97}) +[2026-06-16 12:41:59,671][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.0014094759244471788, 'train/grad_norm_pre_clip': 0.38302716612815857, 'timing/data': 0.030839283019304276, 'timing/model': 1.5333217452280223, 'learning_rate/qwen_vl_interface': 5.152460999388504e-07, 'learning_rate/action_model': 5.152460999388503e-06, 'epoch': 1.97}) +[2026-06-16 12:42:06,046][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.24551761150360107, 'train/grad_norm_pre_clip': 25.283924102783203, 'timing/data': 0.03068184293806553, 'timing/model': 1.6186433131806552, 'learning_rate/qwen_vl_interface': 5.150032660742517e-07, 'learning_rate/action_model': 5.150032660742517e-06, 'epoch': 1.97}) +[2026-06-16 12:42:12,354][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.04198428988456726, 'train/grad_norm_pre_clip': 17.404056549072266, 'timing/data': 0.031144034583121538, 'timing/model': 1.5946899871341884, 'learning_rate/qwen_vl_interface': 5.147623785879535e-07, 'learning_rate/action_model': 5.147623785879534e-06, 'epoch': 1.97}) +[2026-06-16 12:42:18,738][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.001737923943437636, 'train/grad_norm_pre_clip': 1.1544232368469238, 'timing/data': 0.035473695024847984, 'timing/model': 1.8085953891277313, 'learning_rate/qwen_vl_interface': 5.145234375789758e-07, 'learning_rate/action_model': 5.145234375789758e-06, 'epoch': 1.97}) +[2026-06-16 12:42:24,773][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.0003217121702618897, 'train/grad_norm_pre_clip': 0.12082308530807495, 'timing/data': 0.030130014289170504, 'timing/model': 1.5420103180222213, 'learning_rate/qwen_vl_interface': 5.142864431455387e-07, 'learning_rate/action_model': 5.1428644314553875e-06, 'epoch': 1.97}) +[2026-06-16 12:42:31,166][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.0005619439180009067, 'train/grad_norm_pre_clip': 0.14904706180095673, 'timing/data': 0.031750207766890526, 'timing/model': 1.527836258057505, 'learning_rate/qwen_vl_interface': 5.140513953850612e-07, 'learning_rate/action_model': 5.140513953850612e-06, 'epoch': 1.97}) +[2026-06-16 12:42:37,373][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.0542311891913414, 'train/grad_norm_pre_clip': 41.99212646484375, 'timing/data': 0.03163775894790888, 'timing/model': 1.5355883440934122, 'learning_rate/qwen_vl_interface': 5.138182943941623e-07, 'learning_rate/action_model': 5.138182943941623e-06, 'epoch': 1.97}) +[2026-06-16 12:42:43,676][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.0010441712802276015, 'train/grad_norm_pre_clip': 0.28655141592025757, 'timing/data': 0.03140373807400465, 'timing/model': 1.5477998410351574, 'learning_rate/qwen_vl_interface': 5.135871402686615e-07, 'learning_rate/action_model': 5.135871402686615e-06, 'epoch': 1.97}) +[2026-06-16 12:42:50,093][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.0010208843741565943, 'train/grad_norm_pre_clip': 0.8768176436424255, 'timing/data': 0.032757492270320654, 'timing/model': 1.5732346358709037, 'learning_rate/qwen_vl_interface': 5.133579331035777e-07, 'learning_rate/action_model': 5.133579331035777e-06, 'epoch': 1.97}) +[2026-06-16 12:42:56,612][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.011315912008285522, 'train/grad_norm_pre_clip': 9.64101505279541, 'timing/data': 0.030973148997873068, 'timing/model': 1.7814535298384726, 'learning_rate/qwen_vl_interface': 5.131306729931284e-07, 'learning_rate/action_model': 5.1313067299312836e-06, 'epoch': 1.98}) +[2026-06-16 12:43:02,675][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.0003719878732226789, 'train/grad_norm_pre_clip': 0.13380570709705353, 'timing/data': 0.038110786117613316, 'timing/model': 1.5337309371680021, 'learning_rate/qwen_vl_interface': 5.129053600307326e-07, 'learning_rate/action_model': 5.129053600307326e-06, 'epoch': 1.98}) +[2026-06-16 12:43:08,964][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.16625282168388367, 'train/grad_norm_pre_clip': 34.56935501098633, 'timing/data': 0.02951468713581562, 'timing/model': 1.5395847377367318, 'learning_rate/qwen_vl_interface': 5.126819943090078e-07, 'learning_rate/action_model': 5.1268199430900775e-06, 'epoch': 1.98}) +[2026-06-16 12:43:15,277][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.145934596657753, 'train/grad_norm_pre_clip': 22.253568649291992, 'timing/data': 0.031148271169513464, 'timing/model': 1.5404540970921516, 'learning_rate/qwen_vl_interface': 5.124605759197706e-07, 'learning_rate/action_model': 5.124605759197706e-06, 'epoch': 1.98}) +[2026-06-16 12:43:21,767][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.0002453060878906399, 'train/grad_norm_pre_clip': 0.04017116129398346, 'timing/data': 0.03031929023563862, 'timing/model': 1.6401494047604501, 'learning_rate/qwen_vl_interface': 5.122411049540383e-07, 'learning_rate/action_model': 5.1224110495403825e-06, 'epoch': 1.98}) +[2026-06-16 12:43:28,194][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.03171053156256676, 'train/grad_norm_pre_clip': 24.516481399536133, 'timing/data': 0.03057741094380617, 'timing/model': 1.5610440713353455, 'learning_rate/qwen_vl_interface': 5.120235815020269e-07, 'learning_rate/action_model': 5.120235815020269e-06, 'epoch': 1.98}) +[2026-06-16 12:43:34,580][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.000240929817664437, 'train/grad_norm_pre_clip': 0.1348164677619934, 'timing/data': 0.03253811597824097, 'timing/model': 1.8368933210149407, 'learning_rate/qwen_vl_interface': 5.118080056531524e-07, 'learning_rate/action_model': 5.118080056531524e-06, 'epoch': 1.98}) +[2026-06-16 12:43:40,798][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.0005145440227352083, 'train/grad_norm_pre_clip': 0.16499538719654083, 'timing/data': 0.03286010725423694, 'timing/model': 1.6294700959697366, 'learning_rate/qwen_vl_interface': 5.115943774960295e-07, 'learning_rate/action_model': 5.115943774960295e-06, 'epoch': 1.98}) +[2026-06-16 12:43:47,373][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.0023631788790225983, 'train/grad_norm_pre_clip': 1.9368268251419067, 'timing/data': 0.032539966981858015, 'timing/model': 1.6638144361786544, 'learning_rate/qwen_vl_interface': 5.113826971184727e-07, 'learning_rate/action_model': 5.113826971184727e-06, 'epoch': 1.98}) +[2026-06-16 12:43:53,898][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.0005251584225334227, 'train/grad_norm_pre_clip': 0.15415987372398376, 'timing/data': 0.03411138104274869, 'timing/model': 1.62106972374022, 'learning_rate/qwen_vl_interface': 5.111729646074954e-07, 'learning_rate/action_model': 5.111729646074954e-06, 'epoch': 1.98}) +[2026-06-16 12:44:00,294][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.00024823378771543503, 'train/grad_norm_pre_clip': 0.03603657707571983, 'timing/data': 0.02981261210516095, 'timing/model': 1.5809625941328704, 'learning_rate/qwen_vl_interface': 5.10965180049312e-07, 'learning_rate/action_model': 5.10965180049312e-06, 'epoch': 1.98}) +[2026-06-16 12:44:06,571][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.00046271682367660105, 'train/grad_norm_pre_clip': 0.09645510464906693, 'timing/data': 0.031276965979486704, 'timing/model': 1.5373260248452425, 'learning_rate/qwen_vl_interface': 5.107593435293336e-07, 'learning_rate/action_model': 5.107593435293336e-06, 'epoch': 1.98}) +[2026-06-16 12:44:12,907][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.10852153599262238, 'train/grad_norm_pre_clip': 44.22765350341797, 'timing/data': 0.03295913012698293, 'timing/model': 1.7747209668159485, 'learning_rate/qwen_vl_interface': 5.105554551321724e-07, 'learning_rate/action_model': 5.105554551321723e-06, 'epoch': 1.98}) +[2026-06-16 12:44:19,072][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.07591136544942856, 'train/grad_norm_pre_clip': 38.20249557495117, 'timing/data': 0.03032977506518364, 'timing/model': 1.5423320159316063, 'learning_rate/qwen_vl_interface': 5.103535149416389e-07, 'learning_rate/action_model': 5.1035351494163885e-06, 'epoch': 1.98}) +[2026-06-16 12:44:25,449][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.007427021395415068, 'train/grad_norm_pre_clip': 8.208207130432129, 'timing/data': 0.030552485957741737, 'timing/model': 1.620374172925949, 'learning_rate/qwen_vl_interface': 5.101535230407436e-07, 'learning_rate/action_model': 5.101535230407436e-06, 'epoch': 1.98}) +[2026-06-16 12:44:31,763][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.23502779006958008, 'train/grad_norm_pre_clip': 8.676104545593262, 'timing/data': 0.032210099045187235, 'timing/model': 1.5951781971380115, 'learning_rate/qwen_vl_interface': 5.099554795116956e-07, 'learning_rate/action_model': 5.0995547951169565e-06, 'epoch': 1.98}) +[2026-06-16 12:55:17,398][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.0035039258655160666, 'train/grad_norm_pre_clip': 3.6641435623168945, 'eval/loss': 0.0359373576939106, 'eval/flap_precision': 0.9728260869565217, 'eval/flap_recall': 0.9768076398362893, 'eval/flap_f1': 0.9748127978216474, 'eval/flap_rate_model': 0.20444444444444446, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9748127978216474, 'timing/data': 0.021959105040878057, 'timing/model': 1.5429553599096835, 'learning_rate/qwen_vl_interface': 5.097593844359024e-07, 'learning_rate/action_model': 5.0975938443590245e-06, 'epoch': 1.98}) +[2026-06-16 12:55:17,399][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4900_state +[2026-06-16 12:55:17,405][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 12:55:25,789][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4900_state/model.safetensors +[2026-06-16 12:55:45,226][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4900_state/optimizer.bin +[2026-06-16 12:55:45,227][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4900_state/sampler.bin +[2026-06-16 12:55:45,228][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4900_state/sampler_1.bin +[2026-06-16 12:55:45,229][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_4900_state/random_states_0.pkl +[2026-06-16 12:55:45,420][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4900 +[2026-06-16 12:55:45,421][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 12:55:45,436][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 12:55:51,733][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.006974559277296066, 'train/grad_norm_pre_clip': 5.505948543548584, 'timing/data': 0.020939111709594727, 'timing/model': 1.5519317439757288, 'learning_rate/qwen_vl_interface': 5.095652378939718e-07, 'learning_rate/action_model': 5.095652378939717e-06, 'epoch': 1.98}) +[2026-06-16 12:55:57,935][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.06050359830260277, 'train/grad_norm_pre_clip': 37.29598617553711, 'timing/data': 0.022204655688256025, 'timing/model': 1.5107521992176771, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 1.98}) +[2026-06-16 12:56:04,043][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.00046787544852122664, 'train/grad_norm_pre_clip': 0.13285937905311584, 'timing/data': 0.023648120928555727, 'timing/model': 1.5279283039271832, 'learning_rate/qwen_vl_interface': 5.091827907301227e-07, 'learning_rate/action_model': 5.091827907301227e-06, 'epoch': 1.98}) +[2026-06-16 12:56:10,193][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.0934925526380539, 'train/grad_norm_pre_clip': 54.8184700012207, 'timing/data': 0.022632916923612356, 'timing/model': 1.6923485919833183, 'learning_rate/qwen_vl_interface': 5.089944902654134e-07, 'learning_rate/action_model': 5.089944902654134e-06, 'epoch': 1.98}) +[2026-06-16 12:56:16,250][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.0003874994581565261, 'train/grad_norm_pre_clip': 0.06419149041175842, 'timing/data': 0.02072140108793974, 'timing/model': 1.5519334091804922, 'learning_rate/qwen_vl_interface': 5.088081386489856e-07, 'learning_rate/action_model': 5.088081386489857e-06, 'epoch': 1.98}) +[2026-06-16 12:56:22,447][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.06900179386138916, 'train/grad_norm_pre_clip': 20.36736488342285, 'timing/data': 0.021677594166249037, 'timing/model': 1.555903207976371, 'learning_rate/qwen_vl_interface': 5.086237359574419e-07, 'learning_rate/action_model': 5.086237359574419e-06, 'epoch': 1.98}) +[2026-06-16 12:56:28,649][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.0005261262995190918, 'train/grad_norm_pre_clip': 0.09194598346948624, 'timing/data': 0.021557574160397053, 'timing/model': 1.5448899357579648, 'learning_rate/qwen_vl_interface': 5.084412822665825e-07, 'learning_rate/action_model': 5.084412822665825e-06, 'epoch': 1.98}) +[2026-06-16 12:56:34,963][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.0017393114976584911, 'train/grad_norm_pre_clip': 1.3108950853347778, 'timing/data': 0.020143439061939716, 'timing/model': 1.520748799201101, 'learning_rate/qwen_vl_interface': 5.082607776514081e-07, 'learning_rate/action_model': 5.082607776514081e-06, 'epoch': 1.98}) +[2026-06-16 12:56:41,297][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.23046429455280304, 'train/grad_norm_pre_clip': 9.941483497619629, 'timing/data': 0.021082763094455004, 'timing/model': 1.7946108053438365, 'learning_rate/qwen_vl_interface': 5.080822221861167e-07, 'learning_rate/action_model': 5.080822221861166e-06, 'epoch': 1.99}) +[2026-06-16 12:56:47,353][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.02831300161778927, 'train/grad_norm_pre_clip': 19.076900482177734, 'timing/data': 0.0203581303358078, 'timing/model': 1.548972137272358, 'learning_rate/qwen_vl_interface': 5.07905615944106e-07, 'learning_rate/action_model': 5.0790561594410595e-06, 'epoch': 1.99}) +[2026-06-16 12:56:53,657][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.17512401938438416, 'train/grad_norm_pre_clip': 9.490740776062012, 'timing/data': 0.02232901705428958, 'timing/model': 1.5370388408191502, 'learning_rate/qwen_vl_interface': 5.077309589979716e-07, 'learning_rate/action_model': 5.077309589979715e-06, 'epoch': 1.99}) +[2026-06-16 12:56:59,959][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.0012380662374198437, 'train/grad_norm_pre_clip': 0.4541151523590088, 'timing/data': 0.021824246272444725, 'timing/model': 1.5549558787606657, 'learning_rate/qwen_vl_interface': 5.075582514195089e-07, 'learning_rate/action_model': 5.0755825141950895e-06, 'epoch': 1.99}) +[2026-06-16 12:57:06,314][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 0.2989204525947571, 'train/grad_norm_pre_clip': 32.061485290527344, 'timing/data': 0.021136163268238306, 'timing/model': 1.602056527044624, 'learning_rate/qwen_vl_interface': 5.073874932797115e-07, 'learning_rate/action_model': 5.0738749327971155e-06, 'epoch': 1.99}) +[2026-06-16 12:57:12,558][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.15101242065429688, 'train/grad_norm_pre_clip': 25.796688079833984, 'timing/data': 0.021152352914214134, 'timing/model': 1.517610712442547, 'learning_rate/qwen_vl_interface': 5.072186846487713e-07, 'learning_rate/action_model': 5.072186846487713e-06, 'epoch': 1.99}) +[2026-06-16 12:57:18,890][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 0.00012451433576643467, 'train/grad_norm_pre_clip': 0.01886104606091976, 'timing/data': 0.019803498405963182, 'timing/model': 1.772635277826339, 'learning_rate/qwen_vl_interface': 5.070518255960796e-07, 'learning_rate/action_model': 5.070518255960795e-06, 'epoch': 1.99}) +[2026-06-16 12:57:25,066][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.07564376294612885, 'train/grad_norm_pre_clip': 26.27366828918457, 'timing/data': 0.021319673862308264, 'timing/model': 1.594265356194228, 'learning_rate/qwen_vl_interface': 5.068869161902256e-07, 'learning_rate/action_model': 5.068869161902256e-06, 'epoch': 1.99}) +[2026-06-16 12:57:31,388][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.000794519844930619, 'train/grad_norm_pre_clip': 0.1918254792690277, 'timing/data': 0.022072196938097477, 'timing/model': 1.5666020261123776, 'learning_rate/qwen_vl_interface': 5.067239564989968e-07, 'learning_rate/action_model': 5.067239564989968e-06, 'epoch': 1.99}) +[2026-06-16 12:57:37,660][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 0.0006964518688619137, 'train/grad_norm_pre_clip': 0.1525450050830841, 'timing/data': 0.020720040891319513, 'timing/model': 1.5460816007107496, 'learning_rate/qwen_vl_interface': 5.065629465893808e-07, 'learning_rate/action_model': 5.065629465893807e-06, 'epoch': 1.99}) +[2026-06-16 12:57:43,960][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.0767657458782196, 'train/grad_norm_pre_clip': 61.0947380065918, 'timing/data': 0.021122763864696026, 'timing/model': 1.5525587880983949, 'learning_rate/qwen_vl_interface': 5.064038865275625e-07, 'learning_rate/action_model': 5.0640388652756245e-06, 'epoch': 1.99}) +[2026-06-16 12:57:50,370][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.06539097428321838, 'train/grad_norm_pre_clip': 28.826160430908203, 'timing/data': 0.021186084020882845, 'timing/model': 1.5441295336931944, 'learning_rate/qwen_vl_interface': 5.062467763789242e-07, 'learning_rate/action_model': 5.062467763789242e-06, 'epoch': 1.99}) +[2026-06-16 12:57:56,703][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.12816135585308075, 'train/grad_norm_pre_clip': 11.85941219329834, 'timing/data': 0.022341866977512836, 'timing/model': 1.760830412618816, 'learning_rate/qwen_vl_interface': 5.060916162080496e-07, 'learning_rate/action_model': 5.060916162080496e-06, 'epoch': 1.99}) +[2026-06-16 12:58:02,936][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.03643794730305672, 'train/grad_norm_pre_clip': 15.123324394226074, 'timing/data': 0.02231440832838416, 'timing/model': 1.7166454042308033, 'learning_rate/qwen_vl_interface': 5.059384060787182e-07, 'learning_rate/action_model': 5.059384060787183e-06, 'epoch': 1.99}) +[2026-06-16 12:58:08,974][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 0.042826130986213684, 'train/grad_norm_pre_clip': 62.41044235229492, 'timing/data': 0.021470805164426565, 'timing/model': 1.5542571637779474, 'learning_rate/qwen_vl_interface': 5.05787146053909e-07, 'learning_rate/action_model': 5.0578714605390904e-06, 'epoch': 1.99}) +[2026-06-16 12:58:15,270][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.026616433635354042, 'train/grad_norm_pre_clip': 16.84172248840332, 'timing/data': 0.021828826051205397, 'timing/model': 1.5440583629533648, 'learning_rate/qwen_vl_interface': 5.056378361957998e-07, 'learning_rate/action_model': 5.056378361957998e-06, 'epoch': 1.99}) +[2026-06-16 12:58:21,601][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.00010799087613122538, 'train/grad_norm_pre_clip': 0.011438390240073204, 'timing/data': 0.02083472302183509, 'timing/model': 1.585477361921221, 'learning_rate/qwen_vl_interface': 5.05490476565766e-07, 'learning_rate/action_model': 5.054904765657659e-06, 'epoch': 1.99}) +[2026-06-16 12:58:27,951][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.00047038894263096154, 'train/grad_norm_pre_clip': 0.12688754498958588, 'timing/data': 0.022479648236185312, 'timing/model': 1.598050591070205, 'learning_rate/qwen_vl_interface': 5.053450672243811e-07, 'learning_rate/action_model': 5.0534506722438114e-06, 'epoch': 1.99}) +[2026-06-16 12:58:34,295][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 0.00045379684888757765, 'train/grad_norm_pre_clip': 0.08870711922645569, 'timing/data': 0.02169715566560626, 'timing/model': 1.7777080740779638, 'learning_rate/qwen_vl_interface': 5.052016082314184e-07, 'learning_rate/action_model': 5.052016082314184e-06, 'epoch': 1.99}) +[2026-06-16 12:58:40,773][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.00015278707724064589, 'train/grad_norm_pre_clip': 0.04781319573521614, 'timing/data': 0.020774643402546644, 'timing/model': 1.9596636560745537, 'learning_rate/qwen_vl_interface': 5.050600996458473e-07, 'learning_rate/action_model': 5.050600996458473e-06, 'epoch': 1.99}) +[2026-06-16 12:58:46,869][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 0.2755276560783386, 'train/grad_norm_pre_clip': 49.242000579833984, 'timing/data': 0.02187325619161129, 'timing/model': 1.5427845418453217, 'learning_rate/qwen_vl_interface': 5.049205415258373e-07, 'learning_rate/action_model': 5.049205415258373e-06, 'epoch': 1.99}) +[2026-06-16 12:58:53,276][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.0004583202535286546, 'train/grad_norm_pre_clip': 0.12478554248809814, 'timing/data': 0.020896323025226593, 'timing/model': 1.5620048041455448, 'learning_rate/qwen_vl_interface': 5.047829339287557e-07, 'learning_rate/action_model': 5.047829339287557e-06, 'epoch': 1.99}) +[2026-06-16 12:58:59,671][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.12519937753677368, 'train/grad_norm_pre_clip': 12.574843406677246, 'timing/data': 0.021387834567576647, 'timing/model': 1.5792892891913652, 'learning_rate/qwen_vl_interface': 5.046472769111668e-07, 'learning_rate/action_model': 5.046472769111667e-06, 'epoch': 1.99}) +[2026-06-16 12:59:06,072][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.004831775091588497, 'train/grad_norm_pre_clip': 4.765260696411133, 'timing/data': 0.021328995004296303, 'timing/model': 1.6020620130002499, 'learning_rate/qwen_vl_interface': 5.045135705288353e-07, 'learning_rate/action_model': 5.0451357052883525e-06, 'epoch': 1.99}) +[2026-06-16 12:59:12,430][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.2636963129043579, 'train/grad_norm_pre_clip': 5.301715850830078, 'timing/data': 0.020592952147126198, 'timing/model': 1.806303545832634, 'learning_rate/qwen_vl_interface': 5.043818148367221e-07, 'learning_rate/action_model': 5.043818148367221e-06, 'epoch': 1.99}) +[2026-06-16 12:59:18,482][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.014031920582056046, 'train/grad_norm_pre_clip': 25.48695945739746, 'timing/data': 0.02065509231761098, 'timing/model': 1.5644431100226939, 'learning_rate/qwen_vl_interface': 5.042520098889869e-07, 'learning_rate/action_model': 5.0425200988898685e-06, 'epoch': 2.0}) +[2026-06-16 12:59:24,782][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.0040163579396903515, 'train/grad_norm_pre_clip': 3.3220953941345215, 'timing/data': 0.021029003895819187, 'timing/model': 1.5626341179013252, 'learning_rate/qwen_vl_interface': 5.041241557389879e-07, 'learning_rate/action_model': 5.041241557389879e-06, 'epoch': 2.0}) +[2026-06-16 12:59:31,386][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.052004482597112656, 'train/grad_norm_pre_clip': 43.68963623046875, 'timing/data': 0.020934113301336765, 'timing/model': 1.65309051098302, 'learning_rate/qwen_vl_interface': 5.039982524392814e-07, 'learning_rate/action_model': 5.039982524392814e-06, 'epoch': 2.0}) +[2026-06-16 12:59:37,735][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.0029670896474272013, 'train/grad_norm_pre_clip': 4.567659378051758, 'timing/data': 0.02021896094083786, 'timing/model': 1.5746710789389908, 'learning_rate/qwen_vl_interface': 5.038743000416207e-07, 'learning_rate/action_model': 5.0387430004162064e-06, 'epoch': 2.0}) +[2026-06-16 12:59:44,071][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.1712968349456787, 'train/grad_norm_pre_clip': 18.91914939880371, 'timing/data': 0.020469681825488806, 'timing/model': 1.570489508099854, 'learning_rate/qwen_vl_interface': 5.037522985969588e-07, 'learning_rate/action_model': 5.037522985969587e-06, 'epoch': 2.0}) +[2026-06-16 12:59:50,387][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.0025183167308568954, 'train/grad_norm_pre_clip': 1.0691875219345093, 'timing/data': 0.021123383659869432, 'timing/model': 1.7560291071422398, 'learning_rate/qwen_vl_interface': 5.036322481554451e-07, 'learning_rate/action_model': 5.036322481554451e-06, 'epoch': 2.0}) +[2026-06-16 12:59:56,468][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.0005879646050743759, 'train/grad_norm_pre_clip': 0.16570959985256195, 'timing/data': 0.020186661276966333, 'timing/model': 1.5295385322533548, 'learning_rate/qwen_vl_interface': 5.035141487664285e-07, 'learning_rate/action_model': 5.035141487664285e-06, 'epoch': 2.0}) +[2026-06-16 13:00:02,766][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.000148052436998114, 'train/grad_norm_pre_clip': 0.01915307156741619, 'timing/data': 0.020599721930921078, 'timing/model': 1.5422966508194804, 'learning_rate/qwen_vl_interface': 5.03398000478455e-07, 'learning_rate/action_model': 5.03398000478455e-06, 'epoch': 2.0}) +[2026-06-16 13:00:09,170][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.0001824979408411309, 'train/grad_norm_pre_clip': 0.03382200375199318, 'timing/data': 0.02077366318553686, 'timing/model': 1.5523697300814092, 'learning_rate/qwen_vl_interface': 5.032838033392686e-07, 'learning_rate/action_model': 5.032838033392685e-06, 'epoch': 2.0}) +[2026-06-16 13:00:15,503][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.001152162323705852, 'train/grad_norm_pre_clip': 0.551173210144043, 'timing/data': 0.02105099381878972, 'timing/model': 1.5526367351412773, 'learning_rate/qwen_vl_interface': 5.031715573958111e-07, 'learning_rate/action_model': 5.031715573958111e-06, 'epoch': 2.0}) +[2026-06-16 13:00:21,874][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.0012174899457022548, 'train/grad_norm_pre_clip': 0.8730621933937073, 'timing/data': 0.019510508980602026, 'timing/model': 1.5765037131495774, 'learning_rate/qwen_vl_interface': 5.030612626942235e-07, 'learning_rate/action_model': 5.030612626942235e-06, 'epoch': 2.0}) +[2026-06-16 13:00:28,163][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.008972207084298134, 'train/grad_norm_pre_clip': 4.648138999938965, 'timing/data': 0.019976239651441574, 'timing/model': 1.7476727138273418, 'learning_rate/qwen_vl_interface': 5.029529192798433e-07, 'learning_rate/action_model': 5.029529192798433e-06, 'epoch': 2.0}) +[2026-06-16 13:00:33,552][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.01421396154910326, 'train/grad_norm_pre_clip': 11.825152397155762, 'timing/data': 0.0004957523196935654, 'timing/model': 0.8490494913421571, 'learning_rate/qwen_vl_interface': 5.028465271972058e-07, 'learning_rate/action_model': 5.028465271972058e-06, 'epoch': 2.0}) +[2026-06-16 13:00:41,869][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.0006033893441781402, 'train/grad_norm_pre_clip': 0.13277438282966614, 'timing/data': 0.02942095370963216, 'timing/model': 1.6687287529930472, 'learning_rate/qwen_vl_interface': 5.027420864900457e-07, 'learning_rate/action_model': 5.027420864900456e-06, 'epoch': 2.0}) +[2026-06-16 13:00:48,180][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.009382731281220913, 'train/grad_norm_pre_clip': 9.265792846679688, 'timing/data': 0.021666217129677534, 'timing/model': 1.5575017109513283, 'learning_rate/qwen_vl_interface': 5.026395972012943e-07, 'learning_rate/action_model': 5.0263959720129436e-06, 'epoch': 2.0}) +[2026-06-16 13:00:54,574][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.0014476074138656259, 'train/grad_norm_pre_clip': 0.5718469023704529, 'timing/data': 0.02313655288890004, 'timing/model': 1.5919690071605146, 'learning_rate/qwen_vl_interface': 5.025390593730809e-07, 'learning_rate/action_model': 5.025390593730809e-06, 'epoch': 2.0}) +[2026-06-16 13:01:00,908][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.07716912031173706, 'train/grad_norm_pre_clip': 17.739395141601562, 'timing/data': 0.02121755573898554, 'timing/model': 1.781680354848504, 'learning_rate/qwen_vl_interface': 5.024404730467325e-07, 'learning_rate/action_model': 5.024404730467325e-06, 'epoch': 2.0}) +[2026-06-16 13:01:07,071][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.01626443676650524, 'train/grad_norm_pre_clip': 18.88127326965332, 'timing/data': 0.020916124805808067, 'timing/model': 1.5518848709762096, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 2.0}) +[2026-06-16 13:01:13,367][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.001822758000344038, 'train/grad_norm_pre_clip': 1.4940327405929565, 'timing/data': 0.019530849996954203, 'timing/model': 1.543992237187922, 'learning_rate/qwen_vl_interface': 5.022491550609311e-07, 'learning_rate/action_model': 5.022491550609311e-06, 'epoch': 2.0}) +[2026-06-16 13:01:19,666][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.00036743152304552495, 'train/grad_norm_pre_clip': 0.14789460599422455, 'timing/data': 0.0205958541482687, 'timing/model': 1.5480767940171063, 'learning_rate/qwen_vl_interface': 5.021564234801213e-07, 'learning_rate/action_model': 5.021564234801213e-06, 'epoch': 2.0}) +[2026-06-16 13:01:26,056][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.0009217455517500639, 'train/grad_norm_pre_clip': 0.2339828610420227, 'timing/data': 0.01963756000623107, 'timing/model': 1.603809637017548, 'learning_rate/qwen_vl_interface': 5.020656435584636e-07, 'learning_rate/action_model': 5.020656435584635e-06, 'epoch': 2.0}) +[2026-06-16 13:01:32,374][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.09528356045484543, 'train/grad_norm_pre_clip': 22.569961547851562, 'timing/data': 0.020917855203151703, 'timing/model': 1.5601882743649185, 'learning_rate/qwen_vl_interface': 5.019768153332752e-07, 'learning_rate/action_model': 5.019768153332752e-06, 'epoch': 2.0}) +[2026-06-16 13:01:38,688][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.005430480930954218, 'train/grad_norm_pre_clip': 4.861518383026123, 'timing/data': 0.0235665338113904, 'timing/model': 1.7567197997123003, 'learning_rate/qwen_vl_interface': 5.018899388410693e-07, 'learning_rate/action_model': 5.0188993884106935e-06, 'epoch': 2.0}) +[2026-06-16 13:01:44,890][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 0.04596324637532234, 'train/grad_norm_pre_clip': 28.234153747558594, 'timing/data': 0.02429249696433544, 'timing/model': 1.6026238831691444, 'learning_rate/qwen_vl_interface': 5.018050141175576e-07, 'learning_rate/action_model': 5.018050141175575e-06, 'epoch': 2.0}) +[2026-06-16 13:01:51,564][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.00042150181252509356, 'train/grad_norm_pre_clip': 0.07688095420598984, 'timing/data': 0.020723843947052956, 'timing/model': 1.6240238631144166, 'learning_rate/qwen_vl_interface': 5.017220411976496e-07, 'learning_rate/action_model': 5.0172204119764965e-06, 'epoch': 2.0}) +[2026-06-16 13:01:57,879][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.0008901875698938966, 'train/grad_norm_pre_clip': 0.18998989462852478, 'timing/data': 0.021696058101952076, 'timing/model': 1.543808288872242, 'learning_rate/qwen_vl_interface': 5.016410201154523e-07, 'learning_rate/action_model': 5.016410201154523e-06, 'epoch': 2.01}) +[2026-06-16 13:02:04,369][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.06212981417775154, 'train/grad_norm_pre_clip': 38.03474044799805, 'timing/data': 0.021767017897218466, 'timing/model': 1.6888735508546233, 'learning_rate/qwen_vl_interface': 5.015619509042709e-07, 'learning_rate/action_model': 5.015619509042709e-06, 'epoch': 2.01}) +[2026-06-16 13:02:10,773][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.0008753441506996751, 'train/grad_norm_pre_clip': 0.43820855021476746, 'timing/data': 0.021539567038416862, 'timing/model': 1.5687483050860465, 'learning_rate/qwen_vl_interface': 5.014848335966067e-07, 'learning_rate/action_model': 5.014848335966067e-06, 'epoch': 2.01}) +[2026-06-16 13:02:17,213][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 0.0004223040596116334, 'train/grad_norm_pre_clip': 0.08240624517202377, 'timing/data': 0.02116995630785823, 'timing/model': 1.8433238407596946, 'learning_rate/qwen_vl_interface': 5.014096682241608e-07, 'learning_rate/action_model': 5.014096682241608e-06, 'epoch': 2.01}) +[2026-06-16 13:02:23,374][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.058817058801651, 'train/grad_norm_pre_clip': 65.95206451416016, 'timing/data': 0.020267142914235592, 'timing/model': 1.552084210794419, 'learning_rate/qwen_vl_interface': 5.013364548178299e-07, 'learning_rate/action_model': 5.013364548178299e-06, 'epoch': 2.01}) +[2026-06-16 13:02:29,888][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.060585856437683105, 'train/grad_norm_pre_clip': 40.8275260925293, 'timing/data': 0.021701788064092398, 'timing/model': 1.5558841181918979, 'learning_rate/qwen_vl_interface': 5.012651934077101e-07, 'learning_rate/action_model': 5.0126519340771e-06, 'epoch': 2.01}) +[2026-06-16 13:02:36,171][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 0.0022206821013242006, 'train/grad_norm_pre_clip': 1.465705156326294, 'timing/data': 0.022198909893631935, 'timing/model': 1.541545479092747, 'learning_rate/qwen_vl_interface': 5.01195884023094e-07, 'learning_rate/action_model': 5.0119588402309395e-06, 'epoch': 2.01}) +[2026-06-16 13:02:42,524][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.14707069098949432, 'train/grad_norm_pre_clip': 43.734622955322266, 'timing/data': 0.020351143088191748, 'timing/model': 1.5912708011455834, 'learning_rate/qwen_vl_interface': 5.011285266924714e-07, 'learning_rate/action_model': 5.011285266924714e-06, 'epoch': 2.01}) +[2026-06-16 13:02:48,871][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.03058118373155594, 'train/grad_norm_pre_clip': 19.63686752319336, 'timing/data': 0.02108949515968561, 'timing/model': 1.64699278306216, 'learning_rate/qwen_vl_interface': 5.010631214435312e-07, 'learning_rate/action_model': 5.010631214435312e-06, 'epoch': 2.01}) +[2026-06-16 13:02:55,418][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 0.0007967619458213449, 'train/grad_norm_pre_clip': 0.2712112069129944, 'timing/data': 0.02176331914961338, 'timing/model': 1.7948680957779288, 'learning_rate/qwen_vl_interface': 5.009996683031586e-07, 'learning_rate/action_model': 5.009996683031586e-06, 'epoch': 2.01}) +[2026-06-16 13:03:01,473][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.0010706502944231033, 'train/grad_norm_pre_clip': 0.2626440227031708, 'timing/data': 0.021035024896264076, 'timing/model': 1.54307972593233, 'learning_rate/qwen_vl_interface': 5.009381672974375e-07, 'learning_rate/action_model': 5.009381672974375e-06, 'epoch': 2.01}) +[2026-06-16 13:03:07,876][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.007811621762812138, 'train/grad_norm_pre_clip': 4.818723201751709, 'timing/data': 0.021409007254987955, 'timing/model': 1.5446232138201594, 'learning_rate/qwen_vl_interface': 5.008786184516475e-07, 'learning_rate/action_model': 5.008786184516474e-06, 'epoch': 2.01}) +[2026-06-16 13:03:14,265][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 0.00011213331890758127, 'train/grad_norm_pre_clip': 0.04442475363612175, 'timing/data': 0.021642387844622135, 'timing/model': 1.635255672968924, 'learning_rate/qwen_vl_interface': 5.008210217902676e-07, 'learning_rate/action_model': 5.008210217902676e-06, 'epoch': 2.01}) +[2026-06-16 13:03:20,658][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.00022886489750817418, 'train/grad_norm_pre_clip': 0.045891329646110535, 'timing/data': 0.020429584197700024, 'timing/model': 1.6028385790996253, 'learning_rate/qwen_vl_interface': 5.007653773369737e-07, 'learning_rate/action_model': 5.007653773369736e-06, 'epoch': 2.01}) +[2026-06-16 13:03:27,087][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.0011575849493965507, 'train/grad_norm_pre_clip': 0.49218857288360596, 'timing/data': 0.021561467088758945, 'timing/model': 1.611868015024811, 'learning_rate/qwen_vl_interface': 5.007116851146386e-07, 'learning_rate/action_model': 5.007116851146386e-06, 'epoch': 2.01}) +[2026-06-16 13:03:33,417][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.00011663891928037629, 'train/grad_norm_pre_clip': 0.012186978943645954, 'timing/data': 0.021390406880527735, 'timing/model': 1.787719190120697, 'learning_rate/qwen_vl_interface': 5.006599451453334e-07, 'learning_rate/action_model': 5.006599451453334e-06, 'epoch': 2.01}) +[2026-06-16 13:03:39,488][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.00097948987968266, 'train/grad_norm_pre_clip': 0.31004583835601807, 'timing/data': 0.020364344120025635, 'timing/model': 1.5626754621043801, 'learning_rate/qwen_vl_interface': 5.006101574503268e-07, 'learning_rate/action_model': 5.006101574503267e-06, 'epoch': 2.01}) +[2026-06-16 13:03:45,873][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.00043126108357682824, 'train/grad_norm_pre_clip': 0.23592321574687958, 'timing/data': 0.021747038699686527, 'timing/model': 1.5375357740558684, 'learning_rate/qwen_vl_interface': 5.005623220500839e-07, 'learning_rate/action_model': 5.005623220500839e-06, 'epoch': 2.01}) +[2026-06-16 13:03:52,274][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.12449096143245697, 'train/grad_norm_pre_clip': 7.806989669799805, 'timing/data': 0.020803075283765793, 'timing/model': 1.5821805279701948, 'learning_rate/qwen_vl_interface': 5.005164389642687e-07, 'learning_rate/action_model': 5.005164389642687e-06, 'epoch': 2.01}) +[2026-06-16 13:03:58,592][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.1076078787446022, 'train/grad_norm_pre_clip': 27.038427352905273, 'timing/data': 0.0204707239754498, 'timing/model': 1.5586423049680889, 'learning_rate/qwen_vl_interface': 5.004725082117423e-07, 'learning_rate/action_model': 5.004725082117422e-06, 'epoch': 2.01}) +[2026-06-16 13:04:05,079][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.001151996199041605, 'train/grad_norm_pre_clip': 0.47329986095428467, 'timing/data': 0.024647768586874008, 'timing/model': 1.5904535530135036, 'learning_rate/qwen_vl_interface': 5.004305298105616e-07, 'learning_rate/action_model': 5.004305298105616e-06, 'epoch': 2.01}) +[2026-06-16 13:04:11,400][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 9.489658259553835e-05, 'train/grad_norm_pre_clip': 0.00882474984973669, 'timing/data': 0.019981661811470985, 'timing/model': 1.7795925592072308, 'learning_rate/qwen_vl_interface': 5.003905037779834e-07, 'learning_rate/action_model': 5.0039050377798345e-06, 'epoch': 2.01}) +[2026-06-16 13:04:17,496][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.08745727688074112, 'train/grad_norm_pre_clip': 67.72454071044922, 'timing/data': 0.021913249976933002, 'timing/model': 1.5553348502144217, 'learning_rate/qwen_vl_interface': 5.003524301304608e-07, 'learning_rate/action_model': 5.003524301304608e-06, 'epoch': 2.01}) +[2026-06-16 13:04:23,960][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.07066464424133301, 'train/grad_norm_pre_clip': 26.457138061523438, 'timing/data': 0.028157182037830353, 'timing/model': 1.6080593536607921, 'learning_rate/qwen_vl_interface': 5.003163088836445e-07, 'learning_rate/action_model': 5.003163088836446e-06, 'epoch': 2.01}) +[2026-06-16 13:04:30,292][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.007530865725129843, 'train/grad_norm_pre_clip': 4.584184169769287, 'timing/data': 0.021901220083236694, 'timing/model': 1.5589934079907835, 'learning_rate/qwen_vl_interface': 5.002821400523818e-07, 'learning_rate/action_model': 5.002821400523818e-06, 'epoch': 2.02}) +[2026-06-16 13:04:36,617][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 0.1557592749595642, 'train/grad_norm_pre_clip': 16.733314514160156, 'timing/data': 0.021059016697108746, 'timing/model': 1.5941141578368843, 'learning_rate/qwen_vl_interface': 5.002499236507192e-07, 'learning_rate/action_model': 5.0024992365071925e-06, 'epoch': 2.02}) +[2026-06-16 13:04:42,975][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.001009551458992064, 'train/grad_norm_pre_clip': 0.6707437634468079, 'timing/data': 0.020711515098810196, 'timing/model': 1.6201647450216115, 'learning_rate/qwen_vl_interface': 5.002196596918993e-07, 'learning_rate/action_model': 5.002196596918993e-06, 'epoch': 2.02}) +[2026-06-16 13:04:49,380][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.01891835778951645, 'train/grad_norm_pre_clip': 19.43158721923828, 'timing/data': 0.02229012083262205, 'timing/model': 1.8512229439802468, 'learning_rate/qwen_vl_interface': 5.001913481883623e-07, 'learning_rate/action_model': 5.001913481883623e-06, 'epoch': 2.02}) +[2026-06-16 13:04:55,479][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.000849507050588727, 'train/grad_norm_pre_clip': 0.23435205221176147, 'timing/data': 0.021290977019816637, 'timing/model': 1.551510120742023, 'learning_rate/qwen_vl_interface': 5.001649891517459e-07, 'learning_rate/action_model': 5.001649891517459e-06, 'epoch': 2.02}) +[2026-06-16 13:05:01,886][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.00011601591540966183, 'train/grad_norm_pre_clip': 0.024093439802527428, 'timing/data': 0.021248566918075085, 'timing/model': 1.5561485597863793, 'learning_rate/qwen_vl_interface': 5.001405825928858e-07, 'learning_rate/action_model': 5.0014058259288575e-06, 'epoch': 2.02}) +[2026-06-16 13:05:08,289][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.0002967868058476597, 'train/grad_norm_pre_clip': 0.07115058600902557, 'timing/data': 0.02147539798170328, 'timing/model': 1.5598724051378667, 'learning_rate/qwen_vl_interface': 5.001181285218143e-07, 'learning_rate/action_model': 5.0011812852181434e-06, 'epoch': 2.02}) +[2026-06-16 13:05:14,580][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.00015985322534106672, 'train/grad_norm_pre_clip': 0.03602396696805954, 'timing/data': 0.02178904926404357, 'timing/model': 1.5451628351584077, 'learning_rate/qwen_vl_interface': 5.000976269477613e-07, 'learning_rate/action_model': 5.0009762694776125e-06, 'epoch': 2.02}) +[2026-06-16 13:05:20,975][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.00014073277998249978, 'train/grad_norm_pre_clip': 0.03584009408950806, 'timing/data': 0.021704048849642277, 'timing/model': 1.6049552243202925, 'learning_rate/qwen_vl_interface': 5.000790778791542e-07, 'learning_rate/action_model': 5.000790778791543e-06, 'epoch': 2.02}) +[2026-06-16 13:05:27,518][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.0001982538087759167, 'train/grad_norm_pre_clip': 0.05485442653298378, 'timing/data': 0.021700249053537846, 'timing/model': 1.872055934742093, 'learning_rate/qwen_vl_interface': 5.000624813236185e-07, 'learning_rate/action_model': 5.000624813236185e-06, 'epoch': 2.02}) +[2026-06-16 13:05:33,594][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.1799931526184082, 'train/grad_norm_pre_clip': 18.985410690307617, 'timing/data': 0.021583358757197857, 'timing/model': 1.565877494867891, 'learning_rate/qwen_vl_interface': 5.000478372879752e-07, 'learning_rate/action_model': 5.000478372879752e-06, 'epoch': 2.02}) +[2026-06-16 13:05:39,877][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.0020356301683932543, 'train/grad_norm_pre_clip': 1.261377215385437, 'timing/data': 0.020202092826366425, 'timing/model': 1.5595562448725104, 'learning_rate/qwen_vl_interface': 5.000351457782452e-07, 'learning_rate/action_model': 5.000351457782451e-06, 'epoch': 2.02}) +[2026-06-16 13:05:46,182][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 0.0002546378527767956, 'train/grad_norm_pre_clip': 0.08744579553604126, 'timing/data': 0.020137323066592216, 'timing/model': 1.557396579068154, 'learning_rate/qwen_vl_interface': 5.000244067996445e-07, 'learning_rate/action_model': 5.000244067996445e-06, 'epoch': 2.02}) +[2026-06-16 13:05:52,477][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.00019696024537552148, 'train/grad_norm_pre_clip': 0.030916692689061165, 'timing/data': 0.021214998327195644, 'timing/model': 1.5538697480224073, 'learning_rate/qwen_vl_interface': 5.000156203565884e-07, 'learning_rate/action_model': 5.000156203565883e-06, 'epoch': 2.02}) +[2026-06-16 13:05:58,909][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.00052443309687078, 'train/grad_norm_pre_clip': 0.1527971774339676, 'timing/data': 0.02180474903434515, 'timing/model': 1.6371986926533282, 'learning_rate/qwen_vl_interface': 5.000087864526878e-07, 'learning_rate/action_model': 5.000087864526878e-06, 'epoch': 2.02}) +[2026-06-16 13:06:05,226][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.12942683696746826, 'train/grad_norm_pre_clip': 23.14043426513672, 'timing/data': 0.02119081700220704, 'timing/model': 1.7914190259762108, 'learning_rate/qwen_vl_interface': 5.000039050907523e-07, 'learning_rate/action_model': 5.000039050907523e-06, 'epoch': 2.02}) +[2026-06-16 13:06:11,436][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.0010619162349030375, 'train/grad_norm_pre_clip': 0.5374038219451904, 'timing/data': 0.020075692795217037, 'timing/model': 1.7081871456466615, 'learning_rate/qwen_vl_interface': 5.000009762727885e-07, 'learning_rate/action_model': 5.000009762727884e-06, 'epoch': 2.02}) +[2026-06-16 13:16:58,518][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.03169330209493637, 'train/grad_norm_pre_clip': 38.698875427246094, 'eval/loss': 0.03644631430506706, 'eval/flap_precision': 0.9727520435967303, 'eval/flap_recall': 0.9740791268758526, 'eval/flap_f1': 0.9734151329243353, 'eval/flap_rate_model': 0.2038888888888889, 'eval/flap_rate_teacher': 0.2036111111111111, 'eval/flap_cc_f1': 0.9734151329243352, 'timing/data': 0.021141186822205782, 'timing/model': 1.5464327582158148, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 2.02}) +[2026-06-16 13:16:58,519][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_5000_state +[2026-06-16 13:16:58,525][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 13:17:06,838][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_5000_state/model.safetensors +[2026-06-16 13:17:26,463][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_5000_state/optimizer.bin +[2026-06-16 13:17:26,464][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_5000_state/sampler.bin +[2026-06-16 13:17:26,464][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_5000_state/sampler_1.bin +[2026-06-16 13:17:26,465][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints/steps_5000_state/random_states_0.pkl +[2026-06-16 13:17:26,675][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-06-16 13:17:26,675][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 13:17:26,691][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 13:17:26,691][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /workspace/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep/checkpoints +[2026-06-16 13:17:28,522][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..ffa1ed16fa7cd08f0c83739d659e9d2a8bd4e827 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_3_200ep/summary.jsonl @@ -0,0 +1,50 @@ +{"steps": 100} +{"steps": 200} +{"steps": 300} +{"steps": 400} +{"steps": 500} +{"steps": 600} +{"steps": 700} +{"steps": 800} +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500} +{"steps": 1600} +{"steps": 1700} +{"steps": 1800} +{"steps": 1900} +{"steps": 2000} +{"steps": 2100} +{"steps": 2200} +{"steps": 2300} +{"steps": 2400} +{"steps": 2500} +{"steps": 2600} +{"steps": 2700} +{"steps": 2800} +{"steps": 2900} +{"steps": 3000} +{"steps": 3100} +{"steps": 3200} +{"steps": 3300} +{"steps": 3400} +{"steps": 3500} +{"steps": 3600} +{"steps": 3700} +{"steps": 3800} +{"steps": 3900} +{"steps": 4000} +{"steps": 4100} +{"steps": 4200} +{"steps": 4300} +{"steps": 4400} +{"steps": 4500} +{"steps": 4600} +{"steps": 4700} +{"steps": 4800} +{"steps": 4900} +{"steps": 5000} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/checkpoints/steps_5000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/checkpoints/steps_5000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d991c61f8f69fdfbe7fde7d087305f6d438dcde8 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/checkpoints/steps_5000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:99f4d57787efbfd68e5fc68789f72e7e77060c0061568006a8ce8c721c73b305 +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..7c530911e592e87649aa660c5344dd4e0d8ec974 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/config.full.yaml @@ -0,0 +1,199 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_fix_latency_4_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: null + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_4_200ep +output_dir: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..29bbdaa05ac02fca5b3388dd7ae3d53feec080ed --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/config.yaml @@ -0,0 +1,71 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep +rl_games: + env_eval: + enabled: true + mid_train: + enabled: false + interval_steps: 100 + task: flappy +run_id: flappy_fix_latency_4_200ep +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_action_classification: true + eval_interval: 100 + eval_num_batches: 100 + freeze_modules: '' + gradient_accumulation_steps: 4 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + pretrained_checkpoint: null + reload_modules: null + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..b516d17b79357a0336dcde8aa3881059b7d5eb90 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.6249502301216125, + 0.37504974007606506, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4841352105140686, + 0.4841352105140686, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 311588, + "num_trajectories": 180 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..0bc1a6f4dcb5ef9fca0be2e7bf6de46e465a6a42 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8188333511352539, + 0.1811666637659073, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.38520029187202454, + 0.38520029187202454, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 72000, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/hydra/.hydra/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..32f4174ac1c96aa7d68419dec119a05e4df2d69b --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/hydra/.hydra/config.yaml @@ -0,0 +1,197 @@ +framework: + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_fix_latency_4_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + config_name: null + source_subdir: null + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: null + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_4_200ep +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/hydra/.hydra/hydra.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..54a448f1ab07b3a769d1d10da41f3aeed1d098c4 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/hydra/.hydra/hydra.yaml @@ -0,0 +1,252 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=single + - run_id=flappy_fix_latency_4_200ep + - run_root_dir=/workspace/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=false + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=5000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=100 + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=4 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9, 0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=2 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.shuffle=true + - datasets.vla_data.action_balance.enabled=false + - datasets.vla_data.action_balance.strategy=balanced_epoch + - datasets.vla_data.action_balance.action_key=action_id + - datasets.vla_data.action_balance.target_flap_fraction=0.3 + - datasets.vla_data.action_balance.noop_id=0 + - datasets.vla_data.action_balance.flap_id=1 + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_4_200ep + - datasets.vla_data.data_mix=flappy_train__bridge + - datasets.vla_data.eval_data_mix=flappy_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=flappy + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=1 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.mid_train.enabled=false + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[0] + - rl_games.env_eval.post_train.enabled=false + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=false,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_balance.action_key=action_id,datasets.vla_data.action_balance.enabled=false,datasets.vla_data.action_balance.flap_id=1,datasets.vla_data.action_balance.noop_id=0,datasets.vla_data.action_balance.strategy=balanced_epoch,datasets.vla_data.action_balance.target_flap_fraction=0.3,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=flappy_train__bridge,datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_4_200ep,datasets.vla_data.eval_data_mix=flappy_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.shuffle=true,datasets.vla_data.video_backend=torchvision_av,env=flappy,framework.action_model.action_dim=7,framework.action_model.action_env_dim=2,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=1,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=false,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[0],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=false,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=flappy,run_id=flappy_fix_latency_4_200ep,run_root_dir=/workspace/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=100,trainer.gradient_accumulation_steps=4,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=5000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9, + 0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.3 + version_base: '1.1' + cwd: /workspace/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/hydra/.hydra/overrides.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..9c60db0284b27f97d26f40083c11c447f3a8754a --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/hydra/.hydra/overrides.yaml @@ -0,0 +1,91 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=single +- run_id=flappy_fix_latency_4_200ep +- run_root_dir=/workspace/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=false +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=5000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=100 +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=4 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9, 0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=2 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.shuffle=true +- datasets.vla_data.action_balance.enabled=false +- datasets.vla_data.action_balance.strategy=balanced_epoch +- datasets.vla_data.action_balance.action_key=action_id +- datasets.vla_data.action_balance.target_flap_fraction=0.3 +- datasets.vla_data.action_balance.noop_id=0 +- datasets.vla_data.action_balance.flap_id=1 +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_4_200ep +- datasets.vla_data.data_mix=flappy_train__bridge +- datasets.vla_data.eval_data_mix=flappy_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=flappy +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=1 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.mid_train.enabled=false +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[0] +- rl_games.env_eval.post_train.enabled=false +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/hydra/train_starvla_hydra.log b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..670f0ecde1b578986e66587dd39a6f77af6b7f98 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/hydra/train_starvla_hydra.log @@ -0,0 +1,5552 @@ +[2026-06-15 18:53:47,916][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-15 18:53:47,925][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-15 18:53:52,361][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-15 18:53:52,388][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-15 18:53:52,952][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-15 18:53:52,952][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-15 18:53:52,953][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-15 18:53:52,953][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-15 18:53:52,953][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-15 18:53:52,953][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-15 18:53:52,953][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-15 18:53:52,953][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-15 18:53:52,953][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-15 18:53:52,953][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-15 18:53:52,953][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-15 18:53:52,953][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-15 18:53:52,953][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-15 18:53:52,953][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-15 18:53:52,953][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-15 18:53:52,953][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-15 18:53:52,953][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-15 18:53:52,953][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-15 18:53:52,953][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-15 18:53:52,954][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-15 18:53:52,954][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-15 18:53:52,954][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-15 18:53:52,954][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-15 18:53:52,954][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-15 18:53:52,954][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-15 18:53:52,954][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-15 18:53:52,954][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-15 18:53:52,954][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-15 18:53:52,954][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-15 18:53:52,954][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-15 18:53:52,954][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-15 18:53:52,980][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-15 18:53:52,999][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/config.full.yaml +[2026-06-15 18:53:53,004][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/config.yaml +[2026-06-15 18:53:53,004][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-15 19:00:57,545][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-15 19:00:57,545][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-06-15 19:00:57,546][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-15 19:00:57,546][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 4 +[2026-06-15 19:00:57,546][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-15 19:01:13,796][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 0.6955085396766663, 'train/grad_norm_pre_clip': 8.103461265563965, 'timing/data': 0.009557304030749947, 'timing/model': 0.9058086180011742, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-15 19:01:17,226][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 0.6946386098861694, 'train/grad_norm_pre_clip': 2.5028083324432373, 'timing/data': 0.00901514501310885, 'timing/model': 0.8935349829844199, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-15 19:01:20,636][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 0.6906237602233887, 'train/grad_norm_pre_clip': 3.880314588546753, 'timing/data': 0.007873269962146878, 'timing/model': 0.8887834329507314, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-06-15 19:01:24,041][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 0.6803893446922302, 'train/grad_norm_pre_clip': 4.338570594787598, 'timing/data': 0.007747141004074365, 'timing/model': 0.8948111709905788, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-06-15 19:01:27,466][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 0.6568842530250549, 'train/grad_norm_pre_clip': 5.409146308898926, 'timing/data': 0.009672201005741954, 'timing/model': 0.8898348579532467, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-06-15 19:01:30,883][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 0.6772836446762085, 'train/grad_norm_pre_clip': 0.7047368288040161, 'timing/data': 0.0095020110020414, 'timing/model': 0.8941849939874373, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-06-15 19:01:34,308][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 0.6502873301506042, 'train/grad_norm_pre_clip': 1.4461857080459595, 'timing/data': 0.008049666997976601, 'timing/model': 0.9026806050096639, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-06-15 19:01:37,753][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.6795910596847534, 'train/grad_norm_pre_clip': 1.1672319173812866, 'timing/data': 0.008107996021863073, 'timing/model': 0.896871495002415, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-06-15 19:01:41,189][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 0.6439646482467651, 'train/grad_norm_pre_clip': 0.3854997158050537, 'timing/data': 0.009835768025368452, 'timing/model': 0.8990413210121915, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-06-15 19:01:44,633][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.7459221482276917, 'train/grad_norm_pre_clip': 5.018887042999268, 'timing/data': 0.007985405973158777, 'timing/model': 0.9029809279600158, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-15 19:01:48,236][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.7034953832626343, 'train/grad_norm_pre_clip': 2.863957405090332, 'timing/data': 0.009310344001278281, 'timing/model': 0.901919080002699, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-06-15 19:01:51,696][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.7049027681350708, 'train/grad_norm_pre_clip': 2.9912257194519043, 'timing/data': 0.010068616014905274, 'timing/model': 0.9078567770193331, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0}) +[2026-06-15 19:01:55,156][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.5933234095573425, 'train/grad_norm_pre_clip': 0.38611382246017456, 'timing/data': 0.009478243999183178, 'timing/model': 0.9038560659973882, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.01}) +[2026-06-15 19:01:58,601][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.8080387115478516, 'train/grad_norm_pre_clip': 6.050872802734375, 'timing/data': 0.009418350004125386, 'timing/model': 0.9022555100382306, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.01}) +[2026-06-15 19:02:02,055][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.6966424584388733, 'train/grad_norm_pre_clip': 2.5483100414276123, 'timing/data': 0.00878706097137183, 'timing/model': 0.9062867130269296, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.01}) +[2026-06-15 19:02:05,516][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.6615900993347168, 'train/grad_norm_pre_clip': 1.2245779037475586, 'timing/data': 0.009913780959323049, 'timing/model': 0.907314496987965, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 19:02:09,003][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.5887283086776733, 'train/grad_norm_pre_clip': 2.9947659969329834, 'timing/data': 0.009750324999913573, 'timing/model': 0.9082177249947563, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 19:02:12,458][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.7035321593284607, 'train/grad_norm_pre_clip': 2.1855621337890625, 'timing/data': 0.009071574022527784, 'timing/model': 0.9053974440321326, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.01}) +[2026-06-15 19:02:15,924][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.6323629021644592, 'train/grad_norm_pre_clip': 3.5391156673431396, 'timing/data': 0.009779030981007963, 'timing/model': 0.9090012360247783, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.01}) +[2026-06-15 19:02:19,388][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.6328970789909363, 'train/grad_norm_pre_clip': 4.713605880737305, 'timing/data': 0.007884048041887581, 'timing/model': 0.9065215759910643, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-15 19:02:22,855][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.6641322374343872, 'train/grad_norm_pre_clip': 1.5832794904708862, 'timing/data': 0.008771970053203404, 'timing/model': 0.9048975220066495, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.01}) +[2026-06-15 19:02:26,321][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.6912835836410522, 'train/grad_norm_pre_clip': 1.426522135734558, 'timing/data': 0.009656567010097206, 'timing/model': 0.9095141509897076, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 19:02:29,793][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.6503074169158936, 'train/grad_norm_pre_clip': 2.0307581424713135, 'timing/data': 0.00974627595860511, 'timing/model': 0.90736549498979, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 19:02:33,266][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.6962518692016602, 'train/grad_norm_pre_clip': 1.7342455387115479, 'timing/data': 0.008359060040675104, 'timing/model': 0.9054719469859265, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.01}) +[2026-06-15 19:02:36,745][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.6656790375709534, 'train/grad_norm_pre_clip': 0.6657341718673706, 'timing/data': 0.00959560100454837, 'timing/model': 0.9071494190138765, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.01}) +[2026-06-15 19:02:40,213][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.6057483553886414, 'train/grad_norm_pre_clip': 4.110410213470459, 'timing/data': 0.009398589027114213, 'timing/model': 0.9049223340116441, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.01}) +[2026-06-15 19:02:43,686][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.6300223469734192, 'train/grad_norm_pre_clip': 2.117731809616089, 'timing/data': 0.009041476994752884, 'timing/model': 0.9102734650368802, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.01}) +[2026-06-15 19:02:47,170][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.677175760269165, 'train/grad_norm_pre_clip': 1.4444431066513062, 'timing/data': 0.009829891030676663, 'timing/model': 0.907734762004111, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 19:02:50,650][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.7847064733505249, 'train/grad_norm_pre_clip': 6.557800769805908, 'timing/data': 0.009349061001557857, 'timing/model': 0.9144688859814778, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.01}) +[2026-06-15 19:02:54,136][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.6435302495956421, 'train/grad_norm_pre_clip': 0.9313087463378906, 'timing/data': 0.010111116978805512, 'timing/model': 0.9122722059837542, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-15 19:02:57,628][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.6143822073936462, 'train/grad_norm_pre_clip': 1.8751509189605713, 'timing/data': 0.009071955981198698, 'timing/model': 0.909230949007906, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.01}) +[2026-06-15 19:03:01,122][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.6461894512176514, 'train/grad_norm_pre_clip': 1.1635637283325195, 'timing/data': 0.009665947989560664, 'timing/model': 0.9106856930302456, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.01}) +[2026-06-15 19:03:04,619][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.6312053203582764, 'train/grad_norm_pre_clip': 0.9336298108100891, 'timing/data': 0.009027655003592372, 'timing/model': 0.9095020820386708, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.01}) +[2026-06-15 19:03:08,095][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.6863242983818054, 'train/grad_norm_pre_clip': 3.070188522338867, 'timing/data': 0.008524065953679383, 'timing/model': 0.9111536029959098, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.01}) +[2026-06-15 19:03:11,576][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.6610103249549866, 'train/grad_norm_pre_clip': 3.878824234008789, 'timing/data': 0.007894659007433802, 'timing/model': 0.9118764479644597, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.01}) +[2026-06-15 19:03:15,062][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.6833410859107971, 'train/grad_norm_pre_clip': 2.0232808589935303, 'timing/data': 0.008428316039498895, 'timing/model': 0.9155611449969001, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.01}) +[2026-06-15 19:03:18,543][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.6204583644866943, 'train/grad_norm_pre_clip': 6.6811628341674805, 'timing/data': 0.008140583988279104, 'timing/model': 0.9103910909616388, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.02}) +[2026-06-15 19:03:22,087][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.6054292917251587, 'train/grad_norm_pre_clip': 8.131797790527344, 'timing/data': 0.008406740031205118, 'timing/model': 0.9082683169981465, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.02}) +[2026-06-15 19:03:25,567][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.6781052947044373, 'train/grad_norm_pre_clip': 9.695724487304688, 'timing/data': 0.010019665991421789, 'timing/model': 0.9086202189791948, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 19:03:29,052][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.5908752083778381, 'train/grad_norm_pre_clip': 5.440011024475098, 'timing/data': 0.0080356239923276, 'timing/model': 0.9121464779600501, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.02}) +[2026-06-15 19:03:32,533][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.5861214399337769, 'train/grad_norm_pre_clip': 3.381772994995117, 'timing/data': 0.009171672980301082, 'timing/model': 0.9100535080069676, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.02}) +[2026-06-15 19:03:36,026][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.6317911148071289, 'train/grad_norm_pre_clip': 3.4599273204803467, 'timing/data': 0.008487800019793212, 'timing/model': 0.9103464650106616, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.02}) +[2026-06-15 19:03:39,538][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.6413402557373047, 'train/grad_norm_pre_clip': 10.07348346710205, 'timing/data': 0.008421252947300673, 'timing/model': 0.9185933289700188, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.02}) +[2026-06-15 19:03:43,037][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.5362833142280579, 'train/grad_norm_pre_clip': 12.070600509643555, 'timing/data': 0.009294947958551347, 'timing/model': 0.9185321879922412, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 19:03:46,521][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.5561083555221558, 'train/grad_norm_pre_clip': 3.9485480785369873, 'timing/data': 0.00988038198556751, 'timing/model': 0.9127721060067415, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.02}) +[2026-06-15 19:03:50,018][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.48639601469039917, 'train/grad_norm_pre_clip': 8.512493133544922, 'timing/data': 0.008570291043724865, 'timing/model': 0.9222469929954968, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.02}) +[2026-06-15 19:03:53,517][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.35905370116233826, 'train/grad_norm_pre_clip': 8.069856643676758, 'timing/data': 0.009238684957381338, 'timing/model': 0.917617094994057, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.02}) +[2026-06-15 19:03:57,025][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.5831936001777649, 'train/grad_norm_pre_clip': 5.187798023223877, 'timing/data': 0.008433953975327313, 'timing/model': 0.916856920986902, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.02}) +[2026-06-15 19:04:00,520][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.43464240431785583, 'train/grad_norm_pre_clip': 3.9558122158050537, 'timing/data': 0.00948876200709492, 'timing/model': 0.9144969179760665, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.02}) +[2026-06-15 19:04:04,029][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.5487074255943298, 'train/grad_norm_pre_clip': 6.25313663482666, 'timing/data': 0.009636785020120442, 'timing/model': 0.9173185729887336, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.02}) +[2026-06-15 19:04:07,533][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.5301651358604431, 'train/grad_norm_pre_clip': 10.929885864257812, 'timing/data': 0.010194535949267447, 'timing/model': 0.9149287120089866, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 19:04:11,033][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.3715798258781433, 'train/grad_norm_pre_clip': 13.19903564453125, 'timing/data': 0.009871923015452921, 'timing/model': 0.916380101989489, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.02}) +[2026-06-15 19:04:14,531][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.2968083620071411, 'train/grad_norm_pre_clip': 7.058311462402344, 'timing/data': 0.008947986003477126, 'timing/model': 0.9135028710006736, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.02}) +[2026-06-15 19:04:18,042][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.3688967823982239, 'train/grad_norm_pre_clip': 6.861915588378906, 'timing/data': 0.008858854998834431, 'timing/model': 0.9163585539790802, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.02}) +[2026-06-15 19:04:21,571][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.46628594398498535, 'train/grad_norm_pre_clip': 7.4873433113098145, 'timing/data': 0.009730527992360294, 'timing/model': 0.9310720919747837, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.02}) +[2026-06-15 19:04:25,086][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.45987457036972046, 'train/grad_norm_pre_clip': 10.55474853515625, 'timing/data': 0.009504709974862635, 'timing/model': 0.921027508971747, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 19:04:28,606][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.6848193407058716, 'train/grad_norm_pre_clip': 26.450775146484375, 'timing/data': 0.00891981297172606, 'timing/model': 0.917772774992045, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.02}) +[2026-06-15 19:04:32,099][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.4949456453323364, 'train/grad_norm_pre_clip': 9.517253875732422, 'timing/data': 0.009454126004129648, 'timing/model': 0.9102353340131231, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.02}) +[2026-06-15 19:04:35,598][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.42098233103752136, 'train/grad_norm_pre_clip': 6.209900379180908, 'timing/data': 0.009345520986244082, 'timing/model': 0.9159049730515108, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.02}) +[2026-06-15 19:04:39,110][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.5273237228393555, 'train/grad_norm_pre_clip': 4.900149345397949, 'timing/data': 0.009892035974189639, 'timing/model': 0.9143790960079059, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.02}) +[2026-06-15 19:04:42,630][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.7238432168960571, 'train/grad_norm_pre_clip': 20.66408920288086, 'timing/data': 0.009180123044643551, 'timing/model': 0.9217192240175791, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.03}) +[2026-06-15 19:04:46,146][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.4906455874443054, 'train/grad_norm_pre_clip': 6.104318618774414, 'timing/data': 0.009609857050236315, 'timing/model': 0.9147096520173363, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.03}) +[2026-06-15 19:04:49,646][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.6087301969528198, 'train/grad_norm_pre_clip': 18.1630802154541, 'timing/data': 0.009032503992784768, 'timing/model': 0.9137097009806894, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.03}) +[2026-06-15 19:04:53,231][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.5128008723258972, 'train/grad_norm_pre_clip': 5.852586269378662, 'timing/data': 0.009758497995790094, 'timing/model': 0.9182427630294114, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.03}) +[2026-06-15 19:04:56,745][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.4676252007484436, 'train/grad_norm_pre_clip': 6.467687129974365, 'timing/data': 0.007773938006721437, 'timing/model': 0.9260631300276145, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.03}) +[2026-06-15 19:05:00,239][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.47120657563209534, 'train/grad_norm_pre_clip': 10.766463279724121, 'timing/data': 0.009537296020425856, 'timing/model': 0.9162792449933477, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.03}) +[2026-06-15 19:05:03,733][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.5453054308891296, 'train/grad_norm_pre_clip': 9.888710021972656, 'timing/data': 0.008077095029875636, 'timing/model': 0.9191765980212949, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.03}) +[2026-06-15 19:05:07,251][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.44665437936782837, 'train/grad_norm_pre_clip': 6.727838516235352, 'timing/data': 0.009950450039468706, 'timing/model': 0.9246413300279528, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.03}) +[2026-06-15 19:05:10,794][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.432975172996521, 'train/grad_norm_pre_clip': 6.361504077911377, 'timing/data': 0.00826382200466469, 'timing/model': 0.936625259986613, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.03}) +[2026-06-15 19:05:14,297][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.514622688293457, 'train/grad_norm_pre_clip': 17.937685012817383, 'timing/data': 0.008185760001651943, 'timing/model': 0.9156906810239889, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.03}) +[2026-06-15 19:05:17,823][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.37284865975379944, 'train/grad_norm_pre_clip': 10.712440490722656, 'timing/data': 0.009460380009841174, 'timing/model': 0.9336182679980993, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.03}) +[2026-06-15 19:05:21,314][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.4243448078632355, 'train/grad_norm_pre_clip': 10.467925071716309, 'timing/data': 0.009842218016274273, 'timing/model': 0.9130686349817552, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.03}) +[2026-06-15 19:05:24,830][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.5961555242538452, 'train/grad_norm_pre_clip': 13.61270809173584, 'timing/data': 0.009178805979900062, 'timing/model': 0.9183338049915619, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.03}) +[2026-06-15 19:05:28,347][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.6850391626358032, 'train/grad_norm_pre_clip': 21.845016479492188, 'timing/data': 0.00937265099491924, 'timing/model': 0.9144802250084467, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.03}) +[2026-06-15 19:05:31,858][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.34986379742622375, 'train/grad_norm_pre_clip': 7.324368953704834, 'timing/data': 0.008474586007650942, 'timing/model': 0.924174927000422, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.03}) +[2026-06-15 19:05:35,375][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 0.5567992925643921, 'train/grad_norm_pre_clip': 17.05506134033203, 'timing/data': 0.009641581040341407, 'timing/model': 0.9227790409931913, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.03}) +[2026-06-15 19:05:38,887][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.5959217548370361, 'train/grad_norm_pre_clip': 6.772936820983887, 'timing/data': 0.009531045972835273, 'timing/model': 0.9199710440007038, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.03}) +[2026-06-15 19:05:42,401][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.28982511162757874, 'train/grad_norm_pre_clip': 10.155532836914062, 'timing/data': 0.009469517972320318, 'timing/model': 0.9200601750053465, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.03}) +[2026-06-15 19:05:45,949][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.37969791889190674, 'train/grad_norm_pre_clip': 87.04230499267578, 'timing/data': 0.012299559952225536, 'timing/model': 0.9345610470045358, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.03}) +[2026-06-15 19:05:49,482][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.5582709312438965, 'train/grad_norm_pre_clip': 25.70890235900879, 'timing/data': 0.00955606799107045, 'timing/model': 0.9220109489979222, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.03}) +[2026-06-15 19:05:53,018][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.4197440445423126, 'train/grad_norm_pre_clip': 7.6888227462768555, 'timing/data': 0.009938226023223251, 'timing/model': 0.9203664530068636, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.03}) +[2026-06-15 19:05:56,559][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.5065476894378662, 'train/grad_norm_pre_clip': 10.0657320022583, 'timing/data': 0.00982169300550595, 'timing/model': 0.9324456140166149, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.03}) +[2026-06-15 19:06:00,091][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.5402456521987915, 'train/grad_norm_pre_clip': 10.807493209838867, 'timing/data': 0.009009492991026491, 'timing/model': 0.9201623609988019, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.03}) +[2026-06-15 19:06:03,620][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.5188331007957458, 'train/grad_norm_pre_clip': 4.3967061042785645, 'timing/data': 0.008162109006661922, 'timing/model': 0.9290373579715379, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.03}) +[2026-06-15 19:06:07,143][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.3905932903289795, 'train/grad_norm_pre_clip': 5.34047794342041, 'timing/data': 0.00924892199691385, 'timing/model': 0.9180513640167192, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.03}) +[2026-06-15 19:06:10,666][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.3734208047389984, 'train/grad_norm_pre_clip': 7.5557942390441895, 'timing/data': 0.009391630010213703, 'timing/model': 0.9160124220070429, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.04}) +[2026-06-15 19:06:14,203][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.7792285680770874, 'train/grad_norm_pre_clip': 15.331257820129395, 'timing/data': 0.009473676967900246, 'timing/model': 0.9309691580128856, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.04}) +[2026-06-15 19:06:17,733][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.45890867710113525, 'train/grad_norm_pre_clip': 9.813037872314453, 'timing/data': 0.00958280498161912, 'timing/model': 0.930656274955254, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.04}) +[2026-06-15 19:06:21,261][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.39628204703330994, 'train/grad_norm_pre_clip': 4.507465839385986, 'timing/data': 0.009625129983760417, 'timing/model': 0.9174681740114465, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.04}) +[2026-06-15 19:06:24,773][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.7177024483680725, 'train/grad_norm_pre_clip': 12.020139694213867, 'timing/data': 0.008215038978960365, 'timing/model': 0.9194468079949729, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.04}) +[2026-06-15 19:06:28,315][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.34590718150138855, 'train/grad_norm_pre_clip': 8.211822509765625, 'timing/data': 0.009461490029934794, 'timing/model': 0.9279078799881972, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.04}) +[2026-06-15 19:06:31,829][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.5059298276901245, 'train/grad_norm_pre_clip': 4.034139156341553, 'timing/data': 0.009627221035771072, 'timing/model': 0.9169725240208209, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.04}) +[2026-06-15 19:06:35,356][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.3767511248588562, 'train/grad_norm_pre_clip': 4.931389331817627, 'timing/data': 0.009755962993949652, 'timing/model': 0.9255550270318054, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.04}) +[2026-06-15 19:06:38,875][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.5525810122489929, 'train/grad_norm_pre_clip': 6.7688984870910645, 'timing/data': 0.009345311962533742, 'timing/model': 0.9154310050071217, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.04}) +[2026-06-15 19:06:42,383][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.40136590600013733, 'train/grad_norm_pre_clip': 7.151098728179932, 'timing/data': 0.008203260018490255, 'timing/model': 0.9126973139937036, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.04}) +[2026-06-15 19:06:45,905][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.4046165347099304, 'train/grad_norm_pre_clip': 6.941243648529053, 'timing/data': 0.009374543034937233, 'timing/model': 0.9220033949823119, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.04}) +[2026-06-15 19:06:49,418][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.3908560276031494, 'train/grad_norm_pre_clip': 11.572733879089355, 'timing/data': 0.009620453987736255, 'timing/model': 0.9170836220146157, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.04}) +[2026-06-15 19:06:52,931][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.4076884686946869, 'train/grad_norm_pre_clip': 8.052474021911621, 'timing/data': 0.008628306037280709, 'timing/model': 0.9203495450201444, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.04}) +[2026-06-15 19:06:56,440][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.3186592161655426, 'train/grad_norm_pre_clip': 17.917856216430664, 'timing/data': 0.009824134001974016, 'timing/model': 0.918462355970405, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.04}) +[2026-06-15 19:08:15,296][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp6mq907ut/test.c -o /tmp/tmp6mq907ut/test.o +[2026-06-15 19:08:15,309][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp6mq907ut/test.o -laio -o /tmp/tmp6mq907ut/a.out +[2026-06-15 19:08:15,766][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpzdt6n7l4/test.c -o /tmp/tmpzdt6n7l4/test.o +[2026-06-15 19:08:15,782][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpzdt6n7l4/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpzdt6n7l4/a.out +[2026-06-15 19:08:15,814][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp2nn15o19/test.c -o /tmp/tmp2nn15o19/test.o +[2026-06-15 19:08:15,829][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp2nn15o19/test.o -laio -o /tmp/tmp2nn15o19/a.out +[2026-06-15 19:11:25,960][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.6545184254646301, 'train/grad_norm_pre_clip': 23.896987915039062, 'eval/loss': 0.388754278421402, 'eval/flap_precision': 0.5450236966824644, 'eval/flap_recall': 0.8873456790123457, 'eval/flap_f1': 0.6752789195537287, 'eval/flap_rate_model': 0.29305555555555557, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.6752789195537287, 'timing/data': 0.00846083799842745, 'timing/model': 0.9215452750213444, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.04}) +[2026-06-15 19:11:25,960][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_100_state +[2026-06-15 19:11:25,963][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 19:11:29,495][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_100_state/model.safetensors +[2026-06-15 19:11:37,101][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_100_state/optimizer.bin +[2026-06-15 19:11:37,102][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_100_state/sampler.bin +[2026-06-15 19:11:37,102][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_100_state/sampler_1.bin +[2026-06-15 19:11:37,103][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-15 19:11:37,105][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-15 19:11:37,105][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 19:11:37,141][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 19:11:41,523][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.6358760595321655, 'train/grad_norm_pre_clip': 24.758220672607422, 'timing/data': 0.011685646953992546, 'timing/model': 1.1241061260225251, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 0.04}) +[2026-06-15 19:11:45,968][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.36403220891952515, 'train/grad_norm_pre_clip': 4.783939361572266, 'timing/data': 0.011959836992900819, 'timing/model': 1.1095170860062353, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 0.04}) +[2026-06-15 19:11:50,433][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.5839147567749023, 'train/grad_norm_pre_clip': 8.148846626281738, 'timing/data': 0.011826553964056075, 'timing/model': 1.296492538007442, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 0.04}) +[2026-06-15 19:11:54,753][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.4860811233520508, 'train/grad_norm_pre_clip': 14.895654678344727, 'timing/data': 0.011922462028451264, 'timing/model': 1.1222481959848665, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 0.04}) +[2026-06-15 19:11:59,243][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.47510001063346863, 'train/grad_norm_pre_clip': 9.138419151306152, 'timing/data': 0.011896588956005871, 'timing/model': 1.1202477890183218, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 0.04}) +[2026-06-15 19:12:03,696][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.7609439492225647, 'train/grad_norm_pre_clip': 25.759166717529297, 'timing/data': 0.012116594007238746, 'timing/model': 1.1240054490044713, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 0.04}) +[2026-06-15 19:12:08,189][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.8043355345726013, 'train/grad_norm_pre_clip': 27.446025848388672, 'timing/data': 0.011984876997303218, 'timing/model': 1.1336489850073121, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 0.04}) +[2026-06-15 19:12:12,704][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.5762365460395813, 'train/grad_norm_pre_clip': 16.39287757873535, 'timing/data': 0.01311999000608921, 'timing/model': 1.130727379990276, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 0.04}) +[2026-06-15 19:12:17,241][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.48415014147758484, 'train/grad_norm_pre_clip': 4.781063079833984, 'timing/data': 0.012477940996177495, 'timing/model': 1.3170234440476634, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 0.04}) +[2026-06-15 19:12:21,598][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.47769343852996826, 'train/grad_norm_pre_clip': 9.82817268371582, 'timing/data': 0.012632752011995763, 'timing/model': 1.1357874139794149, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.05}) +[2026-06-15 19:12:26,114][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.6009124517440796, 'train/grad_norm_pre_clip': 11.831466674804688, 'timing/data': 0.012006344040855765, 'timing/model': 1.1300046010292135, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 0.05}) +[2026-06-15 19:12:30,636][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.5489479899406433, 'train/grad_norm_pre_clip': 9.73698902130127, 'timing/data': 0.011963407043367624, 'timing/model': 1.1335706859827042, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 0.05}) +[2026-06-15 19:12:35,126][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.3948695659637451, 'train/grad_norm_pre_clip': 5.146848201751709, 'timing/data': 0.012574228981975466, 'timing/model': 1.1241907719522715, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 0.05}) +[2026-06-15 19:12:39,620][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.3911511301994324, 'train/grad_norm_pre_clip': 5.561732769012451, 'timing/data': 0.012680230953264982, 'timing/model': 1.2796149829518981, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 0.05}) +[2026-06-15 19:12:43,954][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.5035369396209717, 'train/grad_norm_pre_clip': 6.665677070617676, 'timing/data': 0.012663219997193664, 'timing/model': 1.1243807849823497, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 0.05}) +[2026-06-15 19:12:48,456][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.480929970741272, 'train/grad_norm_pre_clip': 15.72899341583252, 'timing/data': 0.011934312991797924, 'timing/model': 1.1245536009664647, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 0.05}) +[2026-06-15 19:12:52,851][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.5360724925994873, 'train/grad_norm_pre_clip': 15.362689018249512, 'timing/data': 0.012077526014763862, 'timing/model': 1.120461524988059, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 0.05}) +[2026-06-15 19:12:57,368][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.46836766600608826, 'train/grad_norm_pre_clip': 7.625107765197754, 'timing/data': 0.011715984961483628, 'timing/model': 1.1207287310389802, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 0.05}) +[2026-06-15 19:13:01,873][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.37977007031440735, 'train/grad_norm_pre_clip': 3.492750644683838, 'timing/data': 0.012251259002368897, 'timing/model': 1.122434289951343, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 0.05}) +[2026-06-15 19:13:06,371][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.6550471186637878, 'train/grad_norm_pre_clip': 19.835186004638672, 'timing/data': 0.011958254035562277, 'timing/model': 1.2845748259569518, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.05}) +[2026-06-15 19:13:10,824][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.5699307322502136, 'train/grad_norm_pre_clip': 10.786992073059082, 'timing/data': 0.011731748003512621, 'timing/model': 1.2388397680479102, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 0.05}) +[2026-06-15 19:13:15,170][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.8885695338249207, 'train/grad_norm_pre_clip': 15.922079086303711, 'timing/data': 0.012721542036160827, 'timing/model': 1.1244102870114148, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 0.05}) +[2026-06-15 19:13:19,657][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.4627280831336975, 'train/grad_norm_pre_clip': 10.674637794494629, 'timing/data': 0.012432696996256709, 'timing/model': 1.1235251619946212, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 0.05}) +[2026-06-15 19:13:24,169][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.4829820990562439, 'train/grad_norm_pre_clip': 7.299475193023682, 'timing/data': 0.012667233007960021, 'timing/model': 1.1300283159944229, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 0.05}) +[2026-06-15 19:13:28,842][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.4732478857040405, 'train/grad_norm_pre_clip': 5.659974098205566, 'timing/data': 0.011609409004449844, 'timing/model': 1.132630038948264, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 0.05}) +[2026-06-15 19:13:33,364][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.39098307490348816, 'train/grad_norm_pre_clip': 5.459453105926514, 'timing/data': 0.012147061992436647, 'timing/model': 1.2806230469723232, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 0.05}) +[2026-06-15 19:13:37,738][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.4689745604991913, 'train/grad_norm_pre_clip': 11.399663925170898, 'timing/data': 0.011908655986189842, 'timing/model': 1.134407926001586, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 0.05}) +[2026-06-15 19:13:42,271][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.4373583495616913, 'train/grad_norm_pre_clip': 4.939530849456787, 'timing/data': 0.011957721959333867, 'timing/model': 1.126243742997758, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 0.05}) +[2026-06-15 19:13:46,793][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.3239096999168396, 'train/grad_norm_pre_clip': 5.118288993835449, 'timing/data': 0.012039981957059354, 'timing/model': 1.127259270986542, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 0.05}) +[2026-06-15 19:13:51,314][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.3734322488307953, 'train/grad_norm_pre_clip': 6.320639133453369, 'timing/data': 0.011795864964369684, 'timing/model': 1.1317813890054822, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.05}) +[2026-06-15 19:13:55,769][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.35319069027900696, 'train/grad_norm_pre_clip': 12.778609275817871, 'timing/data': 0.011869757028762251, 'timing/model': 1.1331397920148447, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 0.05}) +[2026-06-15 19:14:00,266][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.2070581614971161, 'train/grad_norm_pre_clip': 43.41925811767578, 'timing/data': 0.012116612982936203, 'timing/model': 1.269079119025264, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 0.05}) +[2026-06-15 19:14:04,640][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.3620378077030182, 'train/grad_norm_pre_clip': 19.222930908203125, 'timing/data': 0.011449262034147978, 'timing/model': 1.1470794180058874, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 0.05}) +[2026-06-15 19:14:09,158][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 1.092014193534851, 'train/grad_norm_pre_clip': 40.640499114990234, 'timing/data': 0.012191957968752831, 'timing/model': 1.1279950229800306, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 0.06}) +[2026-06-15 19:14:13,725][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.5713964104652405, 'train/grad_norm_pre_clip': 31.97382926940918, 'timing/data': 0.011817327002063394, 'timing/model': 1.1458729750011116, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.06}) +[2026-06-15 19:14:18,280][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.34973376989364624, 'train/grad_norm_pre_clip': 5.32400369644165, 'timing/data': 0.012238577008247375, 'timing/model': 1.1347546930192038, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 0.06}) +[2026-06-15 19:14:22,789][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.17838649451732635, 'train/grad_norm_pre_clip': 4.959160804748535, 'timing/data': 0.012258740025572479, 'timing/model': 1.1307730929693207, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 0.06}) +[2026-06-15 19:14:27,347][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.6442296504974365, 'train/grad_norm_pre_clip': 16.380578994750977, 'timing/data': 0.015316225995775312, 'timing/model': 1.2997501469799317, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 0.06}) +[2026-06-15 19:14:31,719][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.4881136417388916, 'train/grad_norm_pre_clip': 20.1566219329834, 'timing/data': 0.011993002030067146, 'timing/model': 1.1304114169906825, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 0.06}) +[2026-06-15 19:14:36,251][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.6296311020851135, 'train/grad_norm_pre_clip': 12.450592041015625, 'timing/data': 0.01208498899359256, 'timing/model': 1.1300394730060361, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.06}) +[2026-06-15 19:14:40,798][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.5314635634422302, 'train/grad_norm_pre_clip': 12.940485000610352, 'timing/data': 0.01233746501384303, 'timing/model': 1.1302340300171636, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 0.06}) +[2026-06-15 19:14:45,298][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.44592857360839844, 'train/grad_norm_pre_clip': 18.233617782592773, 'timing/data': 0.015425521996803582, 'timing/model': 1.1431127410032786, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 0.06}) +[2026-06-15 19:14:49,779][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.27560532093048096, 'train/grad_norm_pre_clip': 4.404649257659912, 'timing/data': 0.012074447993654758, 'timing/model': 1.127595294965431, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 0.06}) +[2026-06-15 19:14:54,311][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.7153148651123047, 'train/grad_norm_pre_clip': 8.643448829650879, 'timing/data': 0.012512600980699062, 'timing/model': 1.2922443689894862, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 0.06}) +[2026-06-15 19:14:58,750][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.40256786346435547, 'train/grad_norm_pre_clip': 6.986015796661377, 'timing/data': 0.012257054971996695, 'timing/model': 1.1424286770052277, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 0.06}) +[2026-06-15 19:15:03,286][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.7067126035690308, 'train/grad_norm_pre_clip': 17.923255920410156, 'timing/data': 0.01208718796260655, 'timing/model': 1.1337377000018023, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 0.06}) +[2026-06-15 19:15:07,860][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.6098881363868713, 'train/grad_norm_pre_clip': 10.984179496765137, 'timing/data': 0.012236472044605762, 'timing/model': 1.130495288001839, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 0.06}) +[2026-06-15 19:15:12,396][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.5450693368911743, 'train/grad_norm_pre_clip': 10.286579132080078, 'timing/data': 0.012251498003024608, 'timing/model': 1.1354339949903078, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 0.06}) +[2026-06-15 19:15:16,894][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.41662392020225525, 'train/grad_norm_pre_clip': 5.521152496337891, 'timing/data': 0.012344359012786299, 'timing/model': 1.127517575980164, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.06}) +[2026-06-15 19:15:21,446][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.47160613536834717, 'train/grad_norm_pre_clip': 6.761393070220947, 'timing/data': 0.012294976972043514, 'timing/model': 1.3023339019855484, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.06}) +[2026-06-15 19:15:25,829][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.4592283070087433, 'train/grad_norm_pre_clip': 7.326261520385742, 'timing/data': 0.012338739994447678, 'timing/model': 1.1389859900227748, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 0.06}) +[2026-06-15 19:15:30,380][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.4415542781352997, 'train/grad_norm_pre_clip': 5.433050155639648, 'timing/data': 0.012116378988139331, 'timing/model': 1.1441597900120541, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 0.06}) +[2026-06-15 19:15:34,924][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.3844885528087616, 'train/grad_norm_pre_clip': 4.754485130310059, 'timing/data': 0.012569812999572605, 'timing/model': 1.1312239060061984, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 0.06}) +[2026-06-15 19:15:39,443][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.4369409680366516, 'train/grad_norm_pre_clip': 9.200505256652832, 'timing/data': 0.011729877965990454, 'timing/model': 1.1385146850370802, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 0.06}) +[2026-06-15 19:15:43,982][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.4498627185821533, 'train/grad_norm_pre_clip': 6.702317237854004, 'timing/data': 0.012475563969928771, 'timing/model': 1.1394724390120246, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 0.06}) +[2026-06-15 19:15:48,527][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.7768929600715637, 'train/grad_norm_pre_clip': 10.748350143432617, 'timing/data': 0.011692378029692918, 'timing/model': 1.306323416007217, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 0.06}) +[2026-06-15 19:15:52,888][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.4571904242038727, 'train/grad_norm_pre_clip': 9.097811698913574, 'timing/data': 0.012454344949219376, 'timing/model': 1.1336388049530797, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 0.06}) +[2026-06-15 19:15:57,404][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.6166260838508606, 'train/grad_norm_pre_clip': 7.660653591156006, 'timing/data': 0.011958578950725496, 'timing/model': 1.1304416039492935, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 0.06}) +[2026-06-15 19:16:01,902][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.6425540447235107, 'train/grad_norm_pre_clip': 10.973628044128418, 'timing/data': 0.012014028034172952, 'timing/model': 1.1333184160175733, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 0.07}) +[2026-06-15 19:16:06,405][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.3683231472969055, 'train/grad_norm_pre_clip': 4.33763313293457, 'timing/data': 0.01208821905311197, 'timing/model': 1.1315630080061965, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.07}) +[2026-06-15 19:16:10,931][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.5632463693618774, 'train/grad_norm_pre_clip': 20.153043746948242, 'timing/data': 0.012630105018615723, 'timing/model': 1.1298959160340019, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 0.07}) +[2026-06-15 19:16:15,458][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.43568721413612366, 'train/grad_norm_pre_clip': 7.017539024353027, 'timing/data': 0.011955264024436474, 'timing/model': 1.2947212010039948, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 0.07}) +[2026-06-15 19:16:19,816][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.39225807785987854, 'train/grad_norm_pre_clip': 5.219000816345215, 'timing/data': 0.012447192973922938, 'timing/model': 1.1283130419906229, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 0.07}) +[2026-06-15 19:16:24,342][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.21574902534484863, 'train/grad_norm_pre_clip': 5.275245189666748, 'timing/data': 0.011860399041324854, 'timing/model': 1.1408663190086372, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 0.07}) +[2026-06-15 19:16:28,847][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.35660189390182495, 'train/grad_norm_pre_clip': 9.458490371704102, 'timing/data': 0.011799273954238743, 'timing/model': 1.1430232910206541, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 0.07}) +[2026-06-15 19:16:33,386][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.36894917488098145, 'train/grad_norm_pre_clip': 10.441764831542969, 'timing/data': 0.01206028403248638, 'timing/model': 1.1506882610265166, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 0.07}) +[2026-06-15 19:16:37,915][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.38867321610450745, 'train/grad_norm_pre_clip': 5.659822463989258, 'timing/data': 0.012291110993828624, 'timing/model': 1.1319578349939547, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 0.07}) +[2026-06-15 19:16:42,436][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.43410685658454895, 'train/grad_norm_pre_clip': 11.348915100097656, 'timing/data': 0.01226175797637552, 'timing/model': 1.2851217599818483, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 0.07}) +[2026-06-15 19:16:46,818][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.6842793226242065, 'train/grad_norm_pre_clip': 15.18535327911377, 'timing/data': 0.012343550974037498, 'timing/model': 1.134259395999834, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 0.07}) +[2026-06-15 19:16:51,340][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.5398716926574707, 'train/grad_norm_pre_clip': 15.106383323669434, 'timing/data': 0.011859917955007404, 'timing/model': 1.1302347490563989, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.07}) +[2026-06-15 19:16:55,842][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.32520797848701477, 'train/grad_norm_pre_clip': 8.579219818115234, 'timing/data': 0.012246896978467703, 'timing/model': 1.1363868780317716, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 0.07}) +[2026-06-15 19:17:00,387][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.5716501474380493, 'train/grad_norm_pre_clip': 11.219585418701172, 'timing/data': 0.012149557005614042, 'timing/model': 1.1336234720074572, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 0.07}) +[2026-06-15 19:17:04,912][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.455737829208374, 'train/grad_norm_pre_clip': 7.97301721572876, 'timing/data': 0.011918978008907288, 'timing/model': 1.1358556670020334, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 0.07}) +[2026-06-15 19:17:09,446][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.3433653712272644, 'train/grad_norm_pre_clip': 7.670039176940918, 'timing/data': 0.012749122979585081, 'timing/model': 1.2908919689944014, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 0.07}) +[2026-06-15 19:17:13,926][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.2975674271583557, 'train/grad_norm_pre_clip': 4.672088623046875, 'timing/data': 0.012412448006216437, 'timing/model': 1.2307591370190494, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 0.07}) +[2026-06-15 19:17:18,288][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.35595738887786865, 'train/grad_norm_pre_clip': 4.662730693817139, 'timing/data': 0.01185550499940291, 'timing/model': 1.1311767940060236, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 0.07}) +[2026-06-15 19:17:22,787][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.6431837677955627, 'train/grad_norm_pre_clip': 15.958767890930176, 'timing/data': 0.011730987986084074, 'timing/model': 1.1497549509513192, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 0.07}) +[2026-06-15 19:17:27,354][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.5827953219413757, 'train/grad_norm_pre_clip': 9.047311782836914, 'timing/data': 0.012473954004235566, 'timing/model': 1.1423351209959947, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 0.07}) +[2026-06-15 19:17:31,882][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.41661763191223145, 'train/grad_norm_pre_clip': 7.713787078857422, 'timing/data': 0.011938058014493436, 'timing/model': 1.12919934297679, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 0.07}) +[2026-06-15 19:17:36,433][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.4167662560939789, 'train/grad_norm_pre_clip': 8.280416488647461, 'timing/data': 0.012170091969892383, 'timing/model': 1.3062416960019618, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.07}) +[2026-06-15 19:17:40,919][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.27614983916282654, 'train/grad_norm_pre_clip': 6.909753322601318, 'timing/data': 0.011869342997670174, 'timing/model': 1.246859967999626, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 0.07}) +[2026-06-15 19:17:45,295][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.5448032021522522, 'train/grad_norm_pre_clip': 16.17151641845703, 'timing/data': 0.012014937994536012, 'timing/model': 1.1337410440319218, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 0.07}) +[2026-06-15 19:17:49,794][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.38971272110939026, 'train/grad_norm_pre_clip': 16.080053329467773, 'timing/data': 0.012366398994345218, 'timing/model': 1.129573407000862, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 0.08}) +[2026-06-15 19:17:54,325][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.4410497546195984, 'train/grad_norm_pre_clip': 10.36805248260498, 'timing/data': 0.012065695016644895, 'timing/model': 1.135044110007584, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 0.08}) +[2026-06-15 19:17:58,831][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.3677738904953003, 'train/grad_norm_pre_clip': 12.72467041015625, 'timing/data': 0.012426922970917076, 'timing/model': 1.126027807011269, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 0.08}) +[2026-06-15 19:18:03,372][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.19525843858718872, 'train/grad_norm_pre_clip': 6.971582889556885, 'timing/data': 0.01236679800786078, 'timing/model': 1.29300293099368, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 0.08}) +[2026-06-15 19:18:07,762][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.49792802333831787, 'train/grad_norm_pre_clip': 10.719504356384277, 'timing/data': 0.012160056037828326, 'timing/model': 1.1357645809766836, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 0.08}) +[2026-06-15 19:18:12,290][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.3499194383621216, 'train/grad_norm_pre_clip': 7.96036958694458, 'timing/data': 0.01317214802838862, 'timing/model': 1.1294409550027922, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 0.08}) +[2026-06-15 19:18:16,849][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.7649730443954468, 'train/grad_norm_pre_clip': 22.650476455688477, 'timing/data': 0.012608936987817287, 'timing/model': 1.1310445580165833, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 0.08}) +[2026-06-15 19:18:21,376][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.7825148701667786, 'train/grad_norm_pre_clip': 14.625147819519043, 'timing/data': 0.012073741003405303, 'timing/model': 1.1397611770080402, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.08}) +[2026-06-15 19:18:25,825][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.7643425464630127, 'train/grad_norm_pre_clip': 21.950353622436523, 'timing/data': 0.01200916402740404, 'timing/model': 1.1320424689911306, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 0.08}) +[2026-06-15 19:18:30,335][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 1.0642824172973633, 'train/grad_norm_pre_clip': 29.550769805908203, 'timing/data': 0.012291107035707682, 'timing/model': 1.265710404026322, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 0.08}) +[2026-06-15 19:18:34,708][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.528581976890564, 'train/grad_norm_pre_clip': 11.570631980895996, 'timing/data': 0.012577619985677302, 'timing/model': 1.1312564029940404, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 0.08}) +[2026-06-15 19:18:39,253][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.4735671877861023, 'train/grad_norm_pre_clip': 11.821351051330566, 'timing/data': 0.012160687998402864, 'timing/model': 1.134531222982332, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 0.08}) +[2026-06-15 19:18:43,795][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.44195812940597534, 'train/grad_norm_pre_clip': 8.51085376739502, 'timing/data': 0.01217905996600166, 'timing/model': 1.1385998129844666, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 0.08}) +[2026-06-15 19:18:48,376][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.43867602944374084, 'train/grad_norm_pre_clip': 5.039959907531738, 'timing/data': 0.030925991013646126, 'timing/model': 1.1445954219670966, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 0.08}) +[2026-06-15 19:18:52,854][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.5756250619888306, 'train/grad_norm_pre_clip': 8.67234992980957, 'timing/data': 0.01821471704170108, 'timing/model': 1.1316339200129732, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 0.08}) +[2026-06-15 19:18:57,360][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.4751759171485901, 'train/grad_norm_pre_clip': 11.378103256225586, 'timing/data': 0.0241338629857637, 'timing/model': 1.269663744024001, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.08}) +[2026-06-15 19:19:01,731][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.5291887521743774, 'train/grad_norm_pre_clip': 6.040280818939209, 'timing/data': 0.02365210000425577, 'timing/model': 1.139023725001607, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 0.08}) +[2026-06-15 19:23:18,674][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.39080214500427246, 'train/grad_norm_pre_clip': 5.598989963531494, 'eval/loss': 0.3023933470249176, 'eval/flap_precision': 0.5757009345794393, 'eval/flap_recall': 0.9506172839506173, 'eval/flap_f1': 0.7171129220023282, 'eval/flap_rate_model': 0.2972222222222222, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.7171129220023282, 'timing/data': 0.023866677016485482, 'timing/model': 1.1399524789885618, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.08}) +[2026-06-15 19:23:18,674][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_200_state +[2026-06-15 19:23:18,677][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 19:23:22,340][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_200_state/model.safetensors +[2026-06-15 19:23:32,594][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_200_state/optimizer.bin +[2026-06-15 19:23:32,595][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_200_state/sampler.bin +[2026-06-15 19:23:32,595][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_200_state/sampler_1.bin +[2026-06-15 19:23:32,596][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-15 19:23:32,636][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-15 19:23:32,636][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 19:23:32,642][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 19:23:37,095][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.47297582030296326, 'train/grad_norm_pre_clip': 7.759657382965088, 'timing/data': 0.024141625966876745, 'timing/model': 1.1107918950146995, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 0.08}) +[2026-06-15 19:23:41,561][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.4661557376384735, 'train/grad_norm_pre_clip': 3.024401903152466, 'timing/data': 0.022915514011401683, 'timing/model': 1.1229810700169764, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 0.08}) +[2026-06-15 19:23:46,032][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.5200526714324951, 'train/grad_norm_pre_clip': 5.736915588378906, 'timing/data': 0.023040953965391964, 'timing/model': 1.1144575540092774, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 0.08}) +[2026-06-15 19:23:50,521][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.38322627544403076, 'train/grad_norm_pre_clip': 4.425160884857178, 'timing/data': 0.022579971002414823, 'timing/model': 1.1303510940051638, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 0.08}) +[2026-06-15 19:23:54,985][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.30277571082115173, 'train/grad_norm_pre_clip': 4.605000019073486, 'timing/data': 0.02275152096990496, 'timing/model': 1.1440068270312622, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.08}) +[2026-06-15 19:23:59,482][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.4530162513256073, 'train/grad_norm_pre_clip': 4.2594313621521, 'timing/data': 0.023000010987743735, 'timing/model': 1.251738930994179, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 0.08}) +[2026-06-15 19:24:03,837][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.4238799214363098, 'train/grad_norm_pre_clip': 4.4404473304748535, 'timing/data': 0.028249747992958874, 'timing/model': 1.128315577050671, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 0.09}) +[2026-06-15 19:24:08,325][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.2846236228942871, 'train/grad_norm_pre_clip': 11.098730087280273, 'timing/data': 0.022633116983342916, 'timing/model': 1.1276730070239864, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 0.09}) +[2026-06-15 19:24:12,859][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.5832899808883667, 'train/grad_norm_pre_clip': 19.590688705444336, 'timing/data': 0.023187963000964373, 'timing/model': 1.1346628100145608, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 0.09}) +[2026-06-15 19:24:17,405][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.570177435874939, 'train/grad_norm_pre_clip': 14.874870300292969, 'timing/data': 0.022671039041597396, 'timing/model': 1.1362863929825835, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.09}) +[2026-06-15 19:24:21,931][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.488390177488327, 'train/grad_norm_pre_clip': 13.417884826660156, 'timing/data': 0.02372377400752157, 'timing/model': 1.2853607350261882, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 0.09}) +[2026-06-15 19:24:26,283][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.4208182692527771, 'train/grad_norm_pre_clip': 7.849125385284424, 'timing/data': 0.022713913989719003, 'timing/model': 1.1351569790276699, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 0.09}) +[2026-06-15 19:24:30,801][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.711679220199585, 'train/grad_norm_pre_clip': 11.80579662322998, 'timing/data': 0.023176143993623555, 'timing/model': 1.1248266139882617, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 0.09}) +[2026-06-15 19:24:35,297][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.34183499217033386, 'train/grad_norm_pre_clip': 5.3501763343811035, 'timing/data': 0.024360545037779957, 'timing/model': 1.1301009660237469, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 0.09}) +[2026-06-15 19:24:39,795][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.6009730696678162, 'train/grad_norm_pre_clip': 12.623943328857422, 'timing/data': 0.023899842984974384, 'timing/model': 1.1339277270017192, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 0.09}) +[2026-06-15 19:24:44,325][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.4145592153072357, 'train/grad_norm_pre_clip': 8.749068260192871, 'timing/data': 0.024036778951995075, 'timing/model': 1.129516684974078, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 0.09}) +[2026-06-15 19:24:48,865][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.3128347396850586, 'train/grad_norm_pre_clip': 4.570591449737549, 'timing/data': 0.02349485398735851, 'timing/model': 1.2962130699888803, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 0.09}) +[2026-06-15 19:24:53,241][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.6101634502410889, 'train/grad_norm_pre_clip': 9.345605850219727, 'timing/data': 0.023869798984378576, 'timing/model': 1.1284873560070992, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 0.09}) +[2026-06-15 19:24:57,759][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.4907354712486267, 'train/grad_norm_pre_clip': 6.913914680480957, 'timing/data': 0.0232324959943071, 'timing/model': 1.1290440859738737, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 0.09}) +[2026-06-15 19:25:02,265][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.4614328145980835, 'train/grad_norm_pre_clip': 7.395129203796387, 'timing/data': 0.022748004004824907, 'timing/model': 1.1413201480172575, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.09}) +[2026-06-15 19:25:06,802][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.4267398715019226, 'train/grad_norm_pre_clip': 6.237827301025391, 'timing/data': 0.023370384995359927, 'timing/model': 1.1343826889642514, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 0.09}) +[2026-06-15 19:25:11,347][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.3993796110153198, 'train/grad_norm_pre_clip': 9.463556289672852, 'timing/data': 0.024198197992518544, 'timing/model': 1.1308112609549426, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 0.09}) +[2026-06-15 19:25:15,900][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.3083972930908203, 'train/grad_norm_pre_clip': 4.148934364318848, 'timing/data': 0.02320302795851603, 'timing/model': 1.3116942969500087, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 0.09}) +[2026-06-15 19:25:20,276][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.3800959885120392, 'train/grad_norm_pre_clip': 31.05448341369629, 'timing/data': 0.022976976993959397, 'timing/model': 1.1291074830223806, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 0.09}) +[2026-06-15 19:25:24,816][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.4362610876560211, 'train/grad_norm_pre_clip': 6.710793972015381, 'timing/data': 0.02328704099636525, 'timing/model': 1.1404054050217383, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 0.09}) +[2026-06-15 19:25:29,340][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.35412317514419556, 'train/grad_norm_pre_clip': 6.811214447021484, 'timing/data': 0.023063159023877233, 'timing/model': 1.1322005719994195, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 0.09}) +[2026-06-15 19:25:33,880][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.4571574330329895, 'train/grad_norm_pre_clip': 6.205987453460693, 'timing/data': 0.023766424972563982, 'timing/model': 1.1355057960026897, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 0.09}) +[2026-06-15 19:25:38,428][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.3098042607307434, 'train/grad_norm_pre_clip': 5.869733810424805, 'timing/data': 0.02902501297648996, 'timing/model': 1.138226751005277, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 0.09}) +[2026-06-15 19:25:42,966][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.29947373270988464, 'train/grad_norm_pre_clip': 5.162734031677246, 'timing/data': 0.022873739944770932, 'timing/model': 1.3013929690350778, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 0.09}) +[2026-06-15 19:25:47,416][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.5364313721656799, 'train/grad_norm_pre_clip': 13.293678283691406, 'timing/data': 0.02377886400790885, 'timing/model': 1.213294790009968, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.09}) +[2026-06-15 19:25:51,789][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.22575025260448456, 'train/grad_norm_pre_clip': 9.853401184082031, 'timing/data': 0.023577608983032405, 'timing/model': 1.1318888050154783, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 0.09}) +[2026-06-15 19:25:56,252][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.22434470057487488, 'train/grad_norm_pre_clip': 5.352961540222168, 'timing/data': 0.023474697954952717, 'timing/model': 1.131256405962631, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 0.1}) +[2026-06-15 19:26:00,785][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.2928123474121094, 'train/grad_norm_pre_clip': 7.097064018249512, 'timing/data': 0.023511012957897037, 'timing/model': 1.1320387499872595, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 0.1}) +[2026-06-15 19:26:05,308][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.3934932351112366, 'train/grad_norm_pre_clip': 9.323163986206055, 'timing/data': 0.02359332301421091, 'timing/model': 1.1306018710019998, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 0.1}) +[2026-06-15 19:26:09,884][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.4115797281265259, 'train/grad_norm_pre_clip': 6.429213047027588, 'timing/data': 0.02336232695961371, 'timing/model': 1.320995418005623, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 0.1}) +[2026-06-15 19:26:14,382][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.40291163325309753, 'train/grad_norm_pre_clip': 30.956283569335938, 'timing/data': 0.022879665018990636, 'timing/model': 1.2575658169807866, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 0.1}) +[2026-06-15 19:26:18,783][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.6097838282585144, 'train/grad_norm_pre_clip': 19.879690170288086, 'timing/data': 0.022993305989075452, 'timing/model': 1.1457439879886806, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 0.1}) +[2026-06-15 19:26:23,355][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.23576873540878296, 'train/grad_norm_pre_clip': 10.87639331817627, 'timing/data': 0.02314573002513498, 'timing/model': 1.1439612889662385, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 0.1}) +[2026-06-15 19:26:27,898][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.587967574596405, 'train/grad_norm_pre_clip': 13.402287483215332, 'timing/data': 0.023823686002288014, 'timing/model': 1.1349191530025564, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 0.1}) +[2026-06-15 19:26:32,444][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.26898977160453796, 'train/grad_norm_pre_clip': 7.181480407714844, 'timing/data': 0.023110262991394848, 'timing/model': 1.1411453419714235, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.1}) +[2026-06-15 19:26:36,992][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.39996108412742615, 'train/grad_norm_pre_clip': 15.065465927124023, 'timing/data': 0.02319127699593082, 'timing/model': 1.302977911022026, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 0.1}) +[2026-06-15 19:26:41,534][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.2771780788898468, 'train/grad_norm_pre_clip': 6.373749732971191, 'timing/data': 0.022561029996722937, 'timing/model': 1.2889294340275228, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 0.1}) +[2026-06-15 19:26:45,910][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.30705195665359497, 'train/grad_norm_pre_clip': 11.052702903747559, 'timing/data': 0.024240533006377518, 'timing/model': 1.1321144620305859, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 0.1}) +[2026-06-15 19:26:50,452][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.3497799336910248, 'train/grad_norm_pre_clip': 16.108320236206055, 'timing/data': 0.022767818009015173, 'timing/model': 1.1409176909946837, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 0.1}) +[2026-06-15 19:26:54,992][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.32082420587539673, 'train/grad_norm_pre_clip': 4.151593208312988, 'timing/data': 0.024413464998360723, 'timing/model': 1.1345552499988116, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 0.1}) +[2026-06-15 19:26:59,567][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.3250791132450104, 'train/grad_norm_pre_clip': 4.917685031890869, 'timing/data': 0.023462492041289806, 'timing/model': 1.139384931011591, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 0.1}) +[2026-06-15 19:27:04,082][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.26881980895996094, 'train/grad_norm_pre_clip': 5.27540922164917, 'timing/data': 0.02258084900677204, 'timing/model': 1.1450670380145311, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.1}) +[2026-06-15 19:27:08,642][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.24714332818984985, 'train/grad_norm_pre_clip': 9.170324325561523, 'timing/data': 0.022925610013771802, 'timing/model': 1.3033783829887398, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 0.1}) +[2026-06-15 19:27:13,038][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.35389402508735657, 'train/grad_norm_pre_clip': 8.70277214050293, 'timing/data': 0.02297439204994589, 'timing/model': 1.1307947500026785, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 0.1}) +[2026-06-15 19:27:17,588][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.2525802552700043, 'train/grad_norm_pre_clip': 5.418824195861816, 'timing/data': 0.023799583956133574, 'timing/model': 1.136960509000346, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.1}) +[2026-06-15 19:27:22,137][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.26216360926628113, 'train/grad_norm_pre_clip': 5.510926723480225, 'timing/data': 0.023590954020619392, 'timing/model': 1.138135073997546, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 0.1}) +[2026-06-15 19:27:26,686][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.337709903717041, 'train/grad_norm_pre_clip': 8.96716594696045, 'timing/data': 0.022925617988221347, 'timing/model': 1.133322064008098, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 0.1}) +[2026-06-15 19:27:31,251][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.4114319384098053, 'train/grad_norm_pre_clip': 11.273028373718262, 'timing/data': 0.023688393994234502, 'timing/model': 1.13621773396153, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 0.1}) +[2026-06-15 19:27:35,835][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.2825016379356384, 'train/grad_norm_pre_clip': 12.269290924072266, 'timing/data': 0.02313122694613412, 'timing/model': 1.3057747170096263, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 0.1}) +[2026-06-15 19:27:40,229][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.3858083486557007, 'train/grad_norm_pre_clip': 16.582632064819336, 'timing/data': 0.023101877013687044, 'timing/model': 1.1499657629756257, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 0.1}) +[2026-06-15 19:27:44,795][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.313523530960083, 'train/grad_norm_pre_clip': 12.478389739990234, 'timing/data': 0.022783385997172445, 'timing/model': 1.1401242520078085, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 0.11}) +[2026-06-15 19:27:49,359][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.2445954531431198, 'train/grad_norm_pre_clip': 15.202712059020996, 'timing/data': 0.023191828979179263, 'timing/model': 1.1377268949872814, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 0.11}) +[2026-06-15 19:27:53,929][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.2684679925441742, 'train/grad_norm_pre_clip': 7.667707920074463, 'timing/data': 0.024410173995420337, 'timing/model': 1.1336882639443502, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 0.11}) +[2026-06-15 19:27:58,464][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.4072398841381073, 'train/grad_norm_pre_clip': 6.748543739318848, 'timing/data': 0.023226235003676265, 'timing/model': 1.1494763279915787, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 0.11}) +[2026-06-15 19:28:03,005][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.2946159243583679, 'train/grad_norm_pre_clip': 10.978157997131348, 'timing/data': 0.022992854996118695, 'timing/model': 1.2850475599989295, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.11}) +[2026-06-15 19:28:07,407][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.4909518361091614, 'train/grad_norm_pre_clip': 13.423565864562988, 'timing/data': 0.023032110009808093, 'timing/model': 1.1502673689974472, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 0.11}) +[2026-06-15 19:28:11,958][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.2817939519882202, 'train/grad_norm_pre_clip': 7.556732177734375, 'timing/data': 0.02376079495297745, 'timing/model': 1.1397510259994306, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 0.11}) +[2026-06-15 19:28:16,510][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.4103928804397583, 'train/grad_norm_pre_clip': 6.7655348777771, 'timing/data': 0.023737516021355987, 'timing/model': 1.1319980130065233, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 0.11}) +[2026-06-15 19:28:21,066][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.4060429632663727, 'train/grad_norm_pre_clip': 13.37820053100586, 'timing/data': 0.02269380702637136, 'timing/model': 1.137186201987788, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 0.11}) +[2026-06-15 19:28:25,608][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.320999413728714, 'train/grad_norm_pre_clip': 3.9980642795562744, 'timing/data': 0.023185114027000964, 'timing/model': 1.1288761629839428, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 0.11}) +[2026-06-15 19:28:30,149][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.4014292061328888, 'train/grad_norm_pre_clip': 12.31539249420166, 'timing/data': 0.023076120007317513, 'timing/model': 1.296033227990847, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 0.11}) +[2026-06-15 19:28:34,528][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.34607434272766113, 'train/grad_norm_pre_clip': 8.46554183959961, 'timing/data': 0.023329029965680093, 'timing/model': 1.1334172009956092, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 0.11}) +[2026-06-15 19:28:39,079][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.352967768907547, 'train/grad_norm_pre_clip': 6.521797180175781, 'timing/data': 0.022850664041470736, 'timing/model': 1.1413114400347695, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 0.11}) +[2026-06-15 19:28:43,612][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.44102150201797485, 'train/grad_norm_pre_clip': 9.114155769348145, 'timing/data': 0.023914338962640613, 'timing/model': 1.1334393300348893, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 0.11}) +[2026-06-15 19:28:48,175][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.324835866689682, 'train/grad_norm_pre_clip': 7.700491428375244, 'timing/data': 0.02257484197616577, 'timing/model': 1.1373874029959552, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.11}) +[2026-06-15 19:28:52,712][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.298984169960022, 'train/grad_norm_pre_clip': 6.379990577697754, 'timing/data': 0.02322909201029688, 'timing/model': 1.2902447230299003, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 0.11}) +[2026-06-15 19:28:57,240][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.23720405995845795, 'train/grad_norm_pre_clip': 4.742696762084961, 'timing/data': 0.02247899997746572, 'timing/model': 1.2811374699813314, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 0.11}) +[2026-06-15 19:29:01,630][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.6127395033836365, 'train/grad_norm_pre_clip': 11.148289680480957, 'timing/data': 0.022684048977680504, 'timing/model': 1.1455999889876693, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 0.11}) +[2026-06-15 19:29:06,159][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.35594478249549866, 'train/grad_norm_pre_clip': 12.169540405273438, 'timing/data': 0.023526044038590044, 'timing/model': 1.130743005021941, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 0.11}) +[2026-06-15 19:29:10,725][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.23818513751029968, 'train/grad_norm_pre_clip': 9.619993209838867, 'timing/data': 0.023309532029088587, 'timing/model': 1.139540461008437, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.11}) +[2026-06-15 19:29:15,255][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.3936554789543152, 'train/grad_norm_pre_clip': 15.280169486999512, 'timing/data': 0.02394847699906677, 'timing/model': 1.137240919051692, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 0.11}) +[2026-06-15 19:29:19,783][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.22082625329494476, 'train/grad_norm_pre_clip': 12.866455078125, 'timing/data': 0.023517421970609576, 'timing/model': 1.130589021020569, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 0.11}) +[2026-06-15 19:29:24,339][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.3112246096134186, 'train/grad_norm_pre_clip': 5.900060653686523, 'timing/data': 0.030696733039803803, 'timing/model': 1.2963310700142756, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 0.11}) +[2026-06-15 19:29:28,724][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.24732805788516998, 'train/grad_norm_pre_clip': 8.384711265563965, 'timing/data': 0.023744625970721245, 'timing/model': 1.1374950369936414, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 0.11}) +[2026-06-15 19:29:33,254][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.4847652018070221, 'train/grad_norm_pre_clip': 14.522171974182129, 'timing/data': 0.02303755795583129, 'timing/model': 1.1344253930146806, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.12}) +[2026-06-15 19:29:37,806][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.2843339741230011, 'train/grad_norm_pre_clip': 15.910666465759277, 'timing/data': 0.023682672006543726, 'timing/model': 1.1358401449979283, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 0.12}) +[2026-06-15 19:29:42,345][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.4409472942352295, 'train/grad_norm_pre_clip': 9.534097671508789, 'timing/data': 0.023785436002071947, 'timing/model': 1.1348590099951252, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 0.12}) +[2026-06-15 19:29:46,904][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.3692532479763031, 'train/grad_norm_pre_clip': 12.84489917755127, 'timing/data': 0.02334225398954004, 'timing/model': 1.1383466779952869, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 0.12}) +[2026-06-15 19:29:51,467][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.2703438997268677, 'train/grad_norm_pre_clip': 3.194877862930298, 'timing/data': 0.023521360009908676, 'timing/model': 1.3134613479487598, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 0.12}) +[2026-06-15 19:29:55,861][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.2150547057390213, 'train/grad_norm_pre_clip': 5.618858337402344, 'timing/data': 0.022691681049764156, 'timing/model': 1.1405851789750159, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 0.12}) +[2026-06-15 19:30:00,396][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.19516213238239288, 'train/grad_norm_pre_clip': 5.392667293548584, 'timing/data': 0.023401003039907664, 'timing/model': 1.1359390059951693, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 0.12}) +[2026-06-15 19:30:04,951][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.15355657041072845, 'train/grad_norm_pre_clip': 6.580595016479492, 'timing/data': 0.023735104012303054, 'timing/model': 1.1366314730257727, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 0.12}) +[2026-06-15 19:30:09,498][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.3538880944252014, 'train/grad_norm_pre_clip': 14.955595016479492, 'timing/data': 0.022951984021347016, 'timing/model': 1.1315311749931425, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 0.12}) +[2026-06-15 19:30:14,032][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.6352633237838745, 'train/grad_norm_pre_clip': 16.740034103393555, 'timing/data': 0.02447223901981488, 'timing/model': 1.148631032963749, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 0.12}) +[2026-06-15 19:30:18,608][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.273638516664505, 'train/grad_norm_pre_clip': 12.294818878173828, 'timing/data': 0.024092057021334767, 'timing/model': 1.3120199400000274, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.12}) +[2026-06-15 19:30:22,999][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.5172269344329834, 'train/grad_norm_pre_clip': 20.281394958496094, 'timing/data': 0.023431574983987957, 'timing/model': 1.1323168530361727, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 0.12}) +[2026-06-15 19:30:27,552][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.5400877594947815, 'train/grad_norm_pre_clip': 9.078885078430176, 'timing/data': 0.023688271001446992, 'timing/model': 1.135234032990411, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 0.12}) +[2026-06-15 19:30:32,104][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.2783733606338501, 'train/grad_norm_pre_clip': 9.589174270629883, 'timing/data': 0.023722273006569594, 'timing/model': 1.1360957929864526, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 0.12}) +[2026-06-15 19:30:36,647][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.43914464116096497, 'train/grad_norm_pre_clip': 10.783615112304688, 'timing/data': 0.02294535201508552, 'timing/model': 1.1388014599797316, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 0.12}) +[2026-06-15 19:30:41,214][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.45286184549331665, 'train/grad_norm_pre_clip': 4.500397205352783, 'timing/data': 0.022512123978231102, 'timing/model': 1.1496856149751693, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 0.12}) +[2026-06-15 19:30:45,772][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.2913667559623718, 'train/grad_norm_pre_clip': 13.145132064819336, 'timing/data': 0.02349198202136904, 'timing/model': 1.2975722390110604, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.12}) +[2026-06-15 19:30:50,165][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.3905739486217499, 'train/grad_norm_pre_clip': 6.354169845581055, 'timing/data': 0.023293133999686688, 'timing/model': 1.1432440070202574, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 0.12}) +[2026-06-15 19:30:54,699][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.4335496127605438, 'train/grad_norm_pre_clip': 9.76367473602295, 'timing/data': 0.023517622030340135, 'timing/model': 1.1315789130167104, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 0.12}) +[2026-06-15 19:30:59,246][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.3009428381919861, 'train/grad_norm_pre_clip': 6.161128520965576, 'timing/data': 0.022835991985630244, 'timing/model': 1.1323525420157239, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 0.12}) +[2026-06-15 19:35:21,762][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.3534623384475708, 'train/grad_norm_pre_clip': 4.473878383636475, 'eval/loss': 0.18346908688545227, 'eval/flap_precision': 0.823327615780446, 'eval/flap_recall': 0.7407407407407407, 'eval/flap_f1': 0.7798537774167343, 'eval/flap_rate_model': 0.16194444444444445, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.7798537774167345, 'timing/data': 0.02431398897897452, 'timing/model': 1.1369212749996223, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.12}) +[2026-06-15 19:35:21,762][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_300_state +[2026-06-15 19:35:21,765][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 19:35:25,301][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_300_state/model.safetensors +[2026-06-15 19:35:35,255][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_300_state/optimizer.bin +[2026-06-15 19:35:35,256][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_300_state/sampler.bin +[2026-06-15 19:35:35,256][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_300_state/sampler_1.bin +[2026-06-15 19:35:35,257][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-15 19:35:35,299][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-15 19:35:35,299][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 19:35:35,306][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 19:35:39,733][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.33593857288360596, 'train/grad_norm_pre_clip': 7.032608509063721, 'timing/data': 0.02631053200457245, 'timing/model': 1.1196001520147547, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 0.12}) +[2026-06-15 19:35:44,177][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.353861004114151, 'train/grad_norm_pre_clip': 5.747613430023193, 'timing/data': 0.023190205974970013, 'timing/model': 1.117802970984485, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 0.12}) +[2026-06-15 19:35:48,667][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.2851869761943817, 'train/grad_norm_pre_clip': 3.2410407066345215, 'timing/data': 0.02329685795120895, 'timing/model': 1.2930508549907245, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 0.12}) +[2026-06-15 19:35:52,998][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.2921498417854309, 'train/grad_norm_pre_clip': 5.493826389312744, 'timing/data': 0.023050770978443325, 'timing/model': 1.1284399540163577, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 0.12}) +[2026-06-15 19:35:57,496][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.42594316601753235, 'train/grad_norm_pre_clip': 3.7517142295837402, 'timing/data': 0.023068166978191584, 'timing/model': 1.1221417839988135, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 0.13}) +[2026-06-15 19:36:01,994][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.42822858691215515, 'train/grad_norm_pre_clip': 14.576922416687012, 'timing/data': 0.02355164400069043, 'timing/model': 1.1210668229614384, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 0.13}) +[2026-06-15 19:36:06,478][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.18301169574260712, 'train/grad_norm_pre_clip': 6.91732120513916, 'timing/data': 0.023732244037091732, 'timing/model': 1.1357726730057038, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 0.13}) +[2026-06-15 19:36:10,976][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.2825489342212677, 'train/grad_norm_pre_clip': 7.1789021492004395, 'timing/data': 0.023569872020743787, 'timing/model': 1.1251160249812528, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 0.13}) +[2026-06-15 19:36:15,477][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.3282679617404938, 'train/grad_norm_pre_clip': 12.144980430603027, 'timing/data': 0.023012457008007914, 'timing/model': 1.2850244959699921, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 0.13}) +[2026-06-15 19:36:19,818][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.29483363032341003, 'train/grad_norm_pre_clip': 6.936757564544678, 'timing/data': 0.024064351979177445, 'timing/model': 1.1223304799641483, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.13}) +[2026-06-15 19:36:24,342][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.20415356755256653, 'train/grad_norm_pre_clip': 3.9575705528259277, 'timing/data': 0.024133636965416372, 'timing/model': 1.1242942049866542, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 0.13}) +[2026-06-15 19:36:28,874][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.864215612411499, 'train/grad_norm_pre_clip': 19.763202667236328, 'timing/data': 0.0233676609932445, 'timing/model': 1.1357604030054063, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 0.13}) +[2026-06-15 19:36:33,408][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.38188645243644714, 'train/grad_norm_pre_clip': 11.525151252746582, 'timing/data': 0.023404633975587785, 'timing/model': 1.1315276070381515, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 0.13}) +[2026-06-15 19:36:37,942][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.3632071614265442, 'train/grad_norm_pre_clip': 9.507280349731445, 'timing/data': 0.023972726019565016, 'timing/model': 1.300970583979506, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 0.13}) +[2026-06-15 19:36:42,452][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.3842736482620239, 'train/grad_norm_pre_clip': 17.458240509033203, 'timing/data': 0.023989716020878404, 'timing/model': 1.2734452330041677, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 0.13}) +[2026-06-15 19:36:46,811][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.15811102092266083, 'train/grad_norm_pre_clip': 4.018698692321777, 'timing/data': 0.024596730014309287, 'timing/model': 1.1289155089762062, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 0.13}) +[2026-06-15 19:36:51,337][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.26616349816322327, 'train/grad_norm_pre_clip': 6.03248405456543, 'timing/data': 0.02353770495392382, 'timing/model': 1.1304307639948092, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 0.13}) +[2026-06-15 19:36:55,880][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.2222064882516861, 'train/grad_norm_pre_clip': 10.777477264404297, 'timing/data': 0.024045807018410414, 'timing/model': 1.1301551000215113, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 0.13}) +[2026-06-15 19:37:00,434][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.14387090504169464, 'train/grad_norm_pre_clip': 30.32168960571289, 'timing/data': 0.02334913198137656, 'timing/model': 1.1458907260093838, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 0.13}) +[2026-06-15 19:37:04,899][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.35688358545303345, 'train/grad_norm_pre_clip': 10.007204055786133, 'timing/data': 0.023866589996032417, 'timing/model': 1.1293408310157247, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.13}) +[2026-06-15 19:37:09,450][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.5139622688293457, 'train/grad_norm_pre_clip': 20.489347457885742, 'timing/data': 0.023531635000836104, 'timing/model': 1.3152300579822622, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 0.13}) +[2026-06-15 19:37:13,838][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.3511524796485901, 'train/grad_norm_pre_clip': 8.619543075561523, 'timing/data': 0.0259534859796986, 'timing/model': 1.1330301080015488, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 0.13}) +[2026-06-15 19:37:18,365][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.24063362181186676, 'train/grad_norm_pre_clip': 13.21536922454834, 'timing/data': 0.024362774041946977, 'timing/model': 1.132712529040873, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 0.13}) +[2026-06-15 19:37:22,906][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.38079696893692017, 'train/grad_norm_pre_clip': 11.372467041015625, 'timing/data': 0.02384855499258265, 'timing/model': 1.1441872430150397, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 0.13}) +[2026-06-15 19:37:27,450][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.45789194107055664, 'train/grad_norm_pre_clip': 14.31399917602539, 'timing/data': 0.023694822972174734, 'timing/model': 1.1360255519975908, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 0.13}) +[2026-06-15 19:37:31,985][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.2715970277786255, 'train/grad_norm_pre_clip': 6.362433910369873, 'timing/data': 0.024043238023296, 'timing/model': 1.1281347009935416, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 0.13}) +[2026-06-15 19:37:36,519][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.3974606990814209, 'train/grad_norm_pre_clip': 14.199393272399902, 'timing/data': 0.023898362996988, 'timing/model': 1.2939944219542667, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 0.13}) +[2026-06-15 19:37:40,904][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.2740970551967621, 'train/grad_norm_pre_clip': 9.852205276489258, 'timing/data': 0.02311693999217823, 'timing/model': 1.143894724023994, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 0.13}) +[2026-06-15 19:37:45,442][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.3558363914489746, 'train/grad_norm_pre_clip': 18.9353084564209, 'timing/data': 0.023525572032667696, 'timing/model': 1.1331987970042974, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 0.14}) +[2026-06-15 19:37:49,968][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.2917000651359558, 'train/grad_norm_pre_clip': 13.147663116455078, 'timing/data': 0.023793629021383822, 'timing/model': 1.1308655649772845, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.14}) +[2026-06-15 19:37:54,510][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.20625588297843933, 'train/grad_norm_pre_clip': 6.290316581726074, 'timing/data': 0.023724725004285574, 'timing/model': 1.1331522180116735, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 0.14}) +[2026-06-15 19:37:59,043][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.26074451208114624, 'train/grad_norm_pre_clip': 5.605112552642822, 'timing/data': 0.023881175962742418, 'timing/model': 1.1538181910291314, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 0.14}) +[2026-06-15 19:38:03,570][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.24892032146453857, 'train/grad_norm_pre_clip': 5.913094997406006, 'timing/data': 0.02386375202331692, 'timing/model': 1.2934498030226678, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 0.14}) +[2026-06-15 19:38:07,945][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.16615238785743713, 'train/grad_norm_pre_clip': 5.754549980163574, 'timing/data': 0.03048192401183769, 'timing/model': 1.1329016499803402, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 0.14}) +[2026-06-15 19:38:12,484][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.21051806211471558, 'train/grad_norm_pre_clip': 6.863712310791016, 'timing/data': 0.023006393981631845, 'timing/model': 1.132970465987455, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 0.14}) +[2026-06-15 19:38:17,045][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.14453935623168945, 'train/grad_norm_pre_clip': 9.356608390808105, 'timing/data': 0.023867103969678283, 'timing/model': 1.1346794359851629, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 0.14}) +[2026-06-15 19:38:21,559][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.4478380084037781, 'train/grad_norm_pre_clip': 21.14887237548828, 'timing/data': 0.02373027498833835, 'timing/model': 1.130589222011622, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 0.14}) +[2026-06-15 19:38:26,116][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.2901678681373596, 'train/grad_norm_pre_clip': 7.67498254776001, 'timing/data': 0.023442674020770937, 'timing/model': 1.1378342200187035, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 0.14}) +[2026-06-15 19:38:30,678][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.12225319445133209, 'train/grad_norm_pre_clip': 5.516791343688965, 'timing/data': 0.023789042024873197, 'timing/model': 1.3201117109856568, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 0.14}) +[2026-06-15 19:38:35,054][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.1821698397397995, 'train/grad_norm_pre_clip': 3.367435932159424, 'timing/data': 0.02359165798407048, 'timing/model': 1.1316142590367235, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.14}) +[2026-06-15 19:38:39,567][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.07263939082622528, 'train/grad_norm_pre_clip': 4.955699920654297, 'timing/data': 0.023734294984024018, 'timing/model': 1.1317843740107492, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 0.14}) +[2026-06-15 19:38:44,103][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.5926657915115356, 'train/grad_norm_pre_clip': 9.614590644836426, 'timing/data': 0.023021436005365103, 'timing/model': 1.1403853769879788, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 0.14}) +[2026-06-15 19:38:48,650][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.3321554958820343, 'train/grad_norm_pre_clip': 19.309785842895508, 'timing/data': 0.02336080401437357, 'timing/model': 1.145872225984931, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 0.14}) +[2026-06-15 19:38:53,201][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.21214494109153748, 'train/grad_norm_pre_clip': 6.644163608551025, 'timing/data': 0.023427920998074114, 'timing/model': 1.3041044540004805, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 0.14}) +[2026-06-15 19:38:57,725][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.39052680134773254, 'train/grad_norm_pre_clip': 17.235042572021484, 'timing/data': 0.02287943399278447, 'timing/model': 1.2828854880062863, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.14}) +[2026-06-15 19:39:02,110][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.41504672169685364, 'train/grad_norm_pre_clip': 15.878058433532715, 'timing/data': 0.02359333197819069, 'timing/model': 1.143462860956788, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 0.14}) +[2026-06-15 19:39:06,656][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.1829311102628708, 'train/grad_norm_pre_clip': 12.667037963867188, 'timing/data': 0.02347623696550727, 'timing/model': 1.140580042032525, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 0.14}) +[2026-06-15 19:39:11,203][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.3378593623638153, 'train/grad_norm_pre_clip': 7.853592395782471, 'timing/data': 0.023845236049965024, 'timing/model': 1.1393900490365922, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 0.14}) +[2026-06-15 19:39:15,731][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.13969925045967102, 'train/grad_norm_pre_clip': 8.374195098876953, 'timing/data': 0.023512220999691635, 'timing/model': 1.1402107549947686, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 0.14}) +[2026-06-15 19:39:20,236][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.1281045526266098, 'train/grad_norm_pre_clip': 6.246283054351807, 'timing/data': 0.023256308981217444, 'timing/model': 1.136543166008778, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.14}) +[2026-06-15 19:39:24,808][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.11826839298009872, 'train/grad_norm_pre_clip': 4.260036468505859, 'timing/data': 0.02374653099104762, 'timing/model': 1.3212955059716478, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 0.14}) +[2026-06-15 19:39:29,214][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.29122617840766907, 'train/grad_norm_pre_clip': 7.6817240715026855, 'timing/data': 0.03181243600556627, 'timing/model': 1.1378599909949116, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 0.14}) +[2026-06-15 19:39:33,761][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.42270147800445557, 'train/grad_norm_pre_clip': 9.018855094909668, 'timing/data': 0.023866927018389106, 'timing/model': 1.1374353850260377, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 0.15}) +[2026-06-15 19:39:38,313][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.15465743839740753, 'train/grad_norm_pre_clip': 8.085509300231934, 'timing/data': 0.02381850901292637, 'timing/model': 1.138946792052593, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 0.15}) +[2026-06-15 19:39:42,876][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.26987501978874207, 'train/grad_norm_pre_clip': 5.900773048400879, 'timing/data': 0.023266361968126148, 'timing/model': 1.1462456620065495, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 0.15}) +[2026-06-15 19:39:47,439][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.29927825927734375, 'train/grad_norm_pre_clip': 13.3329496383667, 'timing/data': 0.024059972958639264, 'timing/model': 1.140368521970231, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 0.15}) +[2026-06-15 19:39:51,977][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.16165819764137268, 'train/grad_norm_pre_clip': 7.8053717613220215, 'timing/data': 0.02348291000816971, 'timing/model': 1.2939796339487657, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 0.15}) +[2026-06-15 19:39:56,359][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.32986578345298767, 'train/grad_norm_pre_clip': 22.751018524169922, 'timing/data': 0.02412970596924424, 'timing/model': 1.1296580049674958, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 0.15}) +[2026-06-15 19:40:00,896][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.19581268727779388, 'train/grad_norm_pre_clip': 10.484136581420898, 'timing/data': 0.022754071978852153, 'timing/model': 1.1368030979647301, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 0.15}) +[2026-06-15 19:40:05,455][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.22562718391418457, 'train/grad_norm_pre_clip': 11.375611305236816, 'timing/data': 0.02391359000466764, 'timing/model': 1.1371813800069503, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.15}) +[2026-06-15 19:40:09,987][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.2882470190525055, 'train/grad_norm_pre_clip': 9.507034301757812, 'timing/data': 0.023029926989693195, 'timing/model': 1.1332201900077052, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 0.15}) +[2026-06-15 19:40:14,498][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.17060890793800354, 'train/grad_norm_pre_clip': 13.372186660766602, 'timing/data': 0.02456857997458428, 'timing/model': 1.1341597170103341, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 0.15}) +[2026-06-15 19:40:19,066][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.2588263750076294, 'train/grad_norm_pre_clip': 8.518011093139648, 'timing/data': 0.023086024972144514, 'timing/model': 1.3211157369660214, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 0.15}) +[2026-06-15 19:40:23,462][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.173682302236557, 'train/grad_norm_pre_clip': 7.905846118927002, 'timing/data': 0.02331183897331357, 'timing/model': 1.1398680330021307, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 0.15}) +[2026-06-15 19:40:28,020][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.3008057773113251, 'train/grad_norm_pre_clip': 17.195693969726562, 'timing/data': 0.0234279939904809, 'timing/model': 1.131989844958298, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 0.15}) +[2026-06-15 19:40:32,571][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.15028363466262817, 'train/grad_norm_pre_clip': 6.73301887512207, 'timing/data': 0.02409936999902129, 'timing/model': 1.1367235999787226, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 0.15}) +[2026-06-15 19:40:37,111][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.3401740491390228, 'train/grad_norm_pre_clip': 7.009488582611084, 'timing/data': 0.023567558964714408, 'timing/model': 1.1380504419794306, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 0.15}) +[2026-06-15 19:40:41,652][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.2725997865200043, 'train/grad_norm_pre_clip': 8.484762191772461, 'timing/data': 0.023584646987728775, 'timing/model': 1.1351660459768027, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 0.15}) +[2026-06-15 19:40:46,200][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.48743247985839844, 'train/grad_norm_pre_clip': 18.19733428955078, 'timing/data': 0.023289237986318767, 'timing/model': 1.2953104299958795, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 0.15}) +[2026-06-15 19:40:50,587][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.18195541203022003, 'train/grad_norm_pre_clip': 8.422986030578613, 'timing/data': 0.02325477998238057, 'timing/model': 1.1435438360203989, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.15}) +[2026-06-15 19:40:55,121][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.2046283483505249, 'train/grad_norm_pre_clip': 13.73695182800293, 'timing/data': 0.022589541040360928, 'timing/model': 1.13997575402027, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 0.15}) +[2026-06-15 19:40:59,677][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.18424920737743378, 'train/grad_norm_pre_clip': 11.215330123901367, 'timing/data': 0.02303637401200831, 'timing/model': 1.1339422360179015, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 0.15}) +[2026-06-15 19:41:04,246][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.42202144861221313, 'train/grad_norm_pre_clip': 12.575284957885742, 'timing/data': 0.024121563008520752, 'timing/model': 1.1428442809847184, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 0.15}) +[2026-06-15 19:41:08,796][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.34801918268203735, 'train/grad_norm_pre_clip': 10.503262519836426, 'timing/data': 0.023549436009489, 'timing/model': 1.305228098994121, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 0.15}) +[2026-06-15 19:41:13,336][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.17972898483276367, 'train/grad_norm_pre_clip': 7.177713394165039, 'timing/data': 0.02349770796718076, 'timing/model': 1.2888367280247621, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 0.15}) +[2026-06-15 19:41:17,715][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.20908716320991516, 'train/grad_norm_pre_clip': 6.558377265930176, 'timing/data': 0.023645703971851617, 'timing/model': 1.129366512002889, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 0.15}) +[2026-06-15 19:41:22,261][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.39465516805648804, 'train/grad_norm_pre_clip': 7.7983012199401855, 'timing/data': 0.02334959403378889, 'timing/model': 1.1338000929681584, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 0.15}) +[2026-06-15 19:41:26,802][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.1964070349931717, 'train/grad_norm_pre_clip': 7.3652663230896, 'timing/data': 0.024565852945670485, 'timing/model': 1.1340235659736209, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 0.16}) +[2026-06-15 19:41:31,349][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.4231966733932495, 'train/grad_norm_pre_clip': 11.959150314331055, 'timing/data': 0.023353900003712624, 'timing/model': 1.1476800460368395, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 0.16}) +[2026-06-15 19:41:35,849][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.18129712343215942, 'train/grad_norm_pre_clip': 8.919329643249512, 'timing/data': 0.023274426988791674, 'timing/model': 1.132079319038894, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.16}) +[2026-06-15 19:41:40,429][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.3433433175086975, 'train/grad_norm_pre_clip': 13.65727710723877, 'timing/data': 0.023282790964003652, 'timing/model': 1.3298736090073362, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 0.16}) +[2026-06-15 19:41:44,831][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.26675885915756226, 'train/grad_norm_pre_clip': 13.467244148254395, 'timing/data': 0.02345110004534945, 'timing/model': 1.144433162000496, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 0.16}) +[2026-06-15 19:41:49,372][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.4072285592556, 'train/grad_norm_pre_clip': 15.885478973388672, 'timing/data': 0.022971829981543124, 'timing/model': 1.1327652839827351, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 0.16}) +[2026-06-15 19:41:53,908][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.1955982744693756, 'train/grad_norm_pre_clip': 8.46564769744873, 'timing/data': 0.023192015010863543, 'timing/model': 1.135579542024061, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 0.16}) +[2026-06-15 19:41:58,451][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.30202215909957886, 'train/grad_norm_pre_clip': 8.930853843688965, 'timing/data': 0.026434899016749114, 'timing/model': 1.1434220719966106, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 0.16}) +[2026-06-15 19:42:02,989][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.14775662124156952, 'train/grad_norm_pre_clip': 9.81964111328125, 'timing/data': 0.02375429798848927, 'timing/model': 1.1345211399602704, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 0.16}) +[2026-06-15 19:42:07,553][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.22057117521762848, 'train/grad_norm_pre_clip': 21.578224182128906, 'timing/data': 0.02309043798595667, 'timing/model': 1.303554967045784, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 0.16}) +[2026-06-15 19:42:11,956][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.2927035391330719, 'train/grad_norm_pre_clip': 12.231996536254883, 'timing/data': 0.02355996100232005, 'timing/model': 1.1422017519944347, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 0.16}) +[2026-06-15 19:42:16,521][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.1821756511926651, 'train/grad_norm_pre_clip': 2.103503704071045, 'timing/data': 0.02391676901606843, 'timing/model': 1.1390990419895388, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 0.16}) +[2026-06-15 19:42:21,095][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.4009607434272766, 'train/grad_norm_pre_clip': 19.710168838500977, 'timing/data': 0.023559297958854586, 'timing/model': 1.13629256200511, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.16}) +[2026-06-15 19:42:25,651][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.19772742688655853, 'train/grad_norm_pre_clip': 4.019079685211182, 'timing/data': 0.023014207021333277, 'timing/model': 1.1387774369795807, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 0.16}) +[2026-06-15 19:42:30,183][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.16936679184436798, 'train/grad_norm_pre_clip': 14.079385757446289, 'timing/data': 0.02352254302240908, 'timing/model': 1.1373440920142457, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 0.16}) +[2026-06-15 19:42:34,774][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.24534247815608978, 'train/grad_norm_pre_clip': 35.600521087646484, 'timing/data': 0.023453061003237963, 'timing/model': 1.315315060026478, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 0.16}) +[2026-06-15 19:42:39,171][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.1296716183423996, 'train/grad_norm_pre_clip': 7.91859245300293, 'timing/data': 0.02315261197509244, 'timing/model': 1.1398323280154727, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.16}) +[2026-06-15 19:42:43,728][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.18671856820583344, 'train/grad_norm_pre_clip': 9.169062614440918, 'timing/data': 0.02364406897686422, 'timing/model': 1.1394904779735953, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 0.16}) +[2026-06-15 19:42:48,301][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.4892288148403168, 'train/grad_norm_pre_clip': 9.425734519958496, 'timing/data': 0.02322006900794804, 'timing/model': 1.1379256230429746, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 0.16}) +[2026-06-15 19:42:52,836][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.25218766927719116, 'train/grad_norm_pre_clip': 35.5313720703125, 'timing/data': 0.022928465041331947, 'timing/model': 1.1342208749847487, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 0.16}) +[2026-06-15 19:42:57,371][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.12200924009084702, 'train/grad_norm_pre_clip': 8.692056655883789, 'timing/data': 0.02339584002038464, 'timing/model': 1.1346201350097544, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 0.16}) +[2026-06-15 19:43:01,916][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.216408371925354, 'train/grad_norm_pre_clip': 11.930781364440918, 'timing/data': 0.023361689993180335, 'timing/model': 1.3000147030106746, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 0.16}) +[2026-06-15 19:47:23,852][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.22440847754478455, 'train/grad_norm_pre_clip': 13.997377395629883, 'eval/loss': 0.1530444175004959, 'eval/flap_precision': 0.768944099378882, 'eval/flap_recall': 0.9552469135802469, 'eval/flap_f1': 0.8520302821748107, 'eval/flap_rate_model': 0.22361111111111112, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.8520302821748108, 'timing/data': 0.02354190102778375, 'timing/model': 1.1359961359994486, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.16}) +[2026-06-15 19:47:23,853][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_400_state +[2026-06-15 19:47:23,860][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 19:47:27,420][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_400_state/model.safetensors +[2026-06-15 19:47:37,536][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_400_state/optimizer.bin +[2026-06-15 19:47:37,536][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_400_state/sampler.bin +[2026-06-15 19:47:37,536][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_400_state/sampler_1.bin +[2026-06-15 19:47:37,537][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-15 19:47:37,577][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-15 19:47:37,577][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 19:47:37,584][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 19:47:41,978][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.23007814586162567, 'train/grad_norm_pre_clip': 17.68445587158203, 'timing/data': 0.027083961002063006, 'timing/model': 1.1981242049951106, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 0.16}) +[2026-06-15 19:47:46,286][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.3521348834037781, 'train/grad_norm_pre_clip': 13.507608413696289, 'timing/data': 0.02269951201742515, 'timing/model': 1.1192842780146748, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 0.17}) +[2026-06-15 19:47:50,681][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.16057442128658295, 'train/grad_norm_pre_clip': 3.7902731895446777, 'timing/data': 0.02251145796617493, 'timing/model': 1.1221976589877158, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 0.17}) +[2026-06-15 19:47:55,139][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.30154815316200256, 'train/grad_norm_pre_clip': 11.059480667114258, 'timing/data': 0.023074996017385274, 'timing/model': 1.1216067370260134, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 0.17}) +[2026-06-15 19:47:59,562][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.31672441959381104, 'train/grad_norm_pre_clip': 12.269759178161621, 'timing/data': 0.02335760701680556, 'timing/model': 1.1160288330283947, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 0.17}) +[2026-06-15 19:48:04,035][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.16466251015663147, 'train/grad_norm_pre_clip': 8.721145629882812, 'timing/data': 0.022428172000218183, 'timing/model': 1.2773990120040253, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 0.17}) +[2026-06-15 19:48:08,375][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.11457611620426178, 'train/grad_norm_pre_clip': 8.735800743103027, 'timing/data': 0.023151261964812875, 'timing/model': 1.1250917119905353, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 0.17}) +[2026-06-15 19:48:12,890][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.15273866057395935, 'train/grad_norm_pre_clip': 9.27005386352539, 'timing/data': 0.02299454400781542, 'timing/model': 1.1335701229982078, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 0.17}) +[2026-06-15 19:48:17,395][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.3674391806125641, 'train/grad_norm_pre_clip': 13.862561225891113, 'timing/data': 0.023307709023356438, 'timing/model': 1.1198293879861012, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 0.17}) +[2026-06-15 19:48:21,877][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.46541276574134827, 'train/grad_norm_pre_clip': 13.635519027709961, 'timing/data': 0.023172843968495727, 'timing/model': 1.129697971045971, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.17}) +[2026-06-15 19:48:26,393][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.2516863942146301, 'train/grad_norm_pre_clip': 9.7926025390625, 'timing/data': 0.022664031013846397, 'timing/model': 1.127600005012937, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 0.17}) +[2026-06-15 19:48:30,917][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.19463491439819336, 'train/grad_norm_pre_clip': 9.82991886138916, 'timing/data': 0.022931448998861015, 'timing/model': 1.2925542650045827, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 0.17}) +[2026-06-15 19:48:35,269][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.23504045605659485, 'train/grad_norm_pre_clip': 8.002059936523438, 'timing/data': 0.023102595994714648, 'timing/model': 1.1253459440195002, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 0.17}) +[2026-06-15 19:48:39,783][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.5397944450378418, 'train/grad_norm_pre_clip': 12.987695693969727, 'timing/data': 0.023130391957238317, 'timing/model': 1.1291274469695054, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 0.17}) +[2026-06-15 19:48:44,343][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.395327091217041, 'train/grad_norm_pre_clip': 9.590424537658691, 'timing/data': 0.029879682988394052, 'timing/model': 1.126701930013951, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.17}) +[2026-06-15 19:48:48,818][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.22176086902618408, 'train/grad_norm_pre_clip': 11.704665184020996, 'timing/data': 0.0231547350413166, 'timing/model': 1.1265441330033354, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 0.17}) +[2026-06-15 19:48:53,308][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.15734383463859558, 'train/grad_norm_pre_clip': 4.629329204559326, 'timing/data': 0.023477278009522706, 'timing/model': 1.1354593510041013, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 0.17}) +[2026-06-15 19:48:57,804][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.27571097016334534, 'train/grad_norm_pre_clip': 69.94007110595703, 'timing/data': 0.023363099026028067, 'timing/model': 1.2686679949983954, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 0.17}) +[2026-06-15 19:49:02,186][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.2685045003890991, 'train/grad_norm_pre_clip': 6.071589469909668, 'timing/data': 0.022793764946982265, 'timing/model': 1.137789600994438, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 0.17}) +[2026-06-15 19:49:06,730][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.44233739376068115, 'train/grad_norm_pre_clip': 35.557525634765625, 'timing/data': 0.025276182044763118, 'timing/model': 1.1407595300115645, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.17}) +[2026-06-15 19:49:11,266][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.3125368356704712, 'train/grad_norm_pre_clip': 9.769781112670898, 'timing/data': 0.023140577017329633, 'timing/model': 1.134341113967821, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 0.17}) +[2026-06-15 19:49:15,823][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.2830771207809448, 'train/grad_norm_pre_clip': 9.682329177856445, 'timing/data': 0.022917048016097397, 'timing/model': 1.143217004952021, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 0.17}) +[2026-06-15 19:49:20,320][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.21719203889369965, 'train/grad_norm_pre_clip': 7.6483354568481445, 'timing/data': 0.022907016042154282, 'timing/model': 1.129805964010302, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 0.17}) +[2026-06-15 19:49:24,850][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.29404839873313904, 'train/grad_norm_pre_clip': 9.44670581817627, 'timing/data': 0.02292573801241815, 'timing/model': 1.2959366780123673, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 0.17}) +[2026-06-15 19:49:29,222][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.17591094970703125, 'train/grad_norm_pre_clip': 13.865249633789062, 'timing/data': 0.022788607981055975, 'timing/model': 1.1308649990241975, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 0.17}) +[2026-06-15 19:49:33,742][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.14630314707756042, 'train/grad_norm_pre_clip': 5.674290180206299, 'timing/data': 0.023078166996128857, 'timing/model': 1.1310912640183233, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 0.18}) +[2026-06-15 19:49:38,275][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.21262823045253754, 'train/grad_norm_pre_clip': 12.922523498535156, 'timing/data': 0.0225375079899095, 'timing/model': 1.1288603599532507, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 0.18}) +[2026-06-15 19:49:42,791][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.23213720321655273, 'train/grad_norm_pre_clip': 10.019487380981445, 'timing/data': 0.023803536023478955, 'timing/model': 1.1317527390201576, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 0.18}) +[2026-06-15 19:49:47,306][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.20577244460582733, 'train/grad_norm_pre_clip': 13.296052932739258, 'timing/data': 0.023300700006075203, 'timing/model': 1.128357056004461, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 0.18}) +[2026-06-15 19:49:51,832][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.2575436532497406, 'train/grad_norm_pre_clip': 13.425505638122559, 'timing/data': 0.023533813015092164, 'timing/model': 1.2866929050069302, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.18}) +[2026-06-15 19:49:56,200][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.2472468465566635, 'train/grad_norm_pre_clip': 11.948572158813477, 'timing/data': 0.02474659401923418, 'timing/model': 1.128327024052851, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 0.18}) +[2026-06-15 19:50:00,734][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.06259965151548386, 'train/grad_norm_pre_clip': 7.273561954498291, 'timing/data': 0.02340339997317642, 'timing/model': 1.1293235900229774, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 0.18}) +[2026-06-15 19:50:05,338][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.2670342028141022, 'train/grad_norm_pre_clip': 8.4506254196167, 'timing/data': 0.02320625400170684, 'timing/model': 1.1495282420073636, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 0.18}) +[2026-06-15 19:50:09,916][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.2847961485385895, 'train/grad_norm_pre_clip': 15.170745849609375, 'timing/data': 0.023103464976884425, 'timing/model': 1.1616098820231855, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 0.18}) +[2026-06-15 19:50:14,450][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.23660705983638763, 'train/grad_norm_pre_clip': 22.16499137878418, 'timing/data': 0.02399968501413241, 'timing/model': 1.138517437968403, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 0.18}) +[2026-06-15 19:50:18,985][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.05911295861005783, 'train/grad_norm_pre_clip': 6.376194477081299, 'timing/data': 0.023145951970946044, 'timing/model': 1.2976819180184975, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 0.18}) +[2026-06-15 19:50:23,359][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.1797989159822464, 'train/grad_norm_pre_clip': 13.193357467651367, 'timing/data': 0.022636025038082153, 'timing/model': 1.13553663197672, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 0.18}) +[2026-06-15 19:50:27,904][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.13431532680988312, 'train/grad_norm_pre_clip': 7.118007659912109, 'timing/data': 0.023156461014878005, 'timing/model': 1.1393667040392756, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 0.18}) +[2026-06-15 19:50:32,456][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.09685597568750381, 'train/grad_norm_pre_clip': 20.835636138916016, 'timing/data': 0.023833143000956625, 'timing/model': 1.1328603269648738, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 0.18}) +[2026-06-15 19:50:36,971][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.28191107511520386, 'train/grad_norm_pre_clip': 7.7206807136535645, 'timing/data': 0.02375299797859043, 'timing/model': 1.1309057240141556, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.18}) +[2026-06-15 19:50:41,510][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.29494136571884155, 'train/grad_norm_pre_clip': 13.931224822998047, 'timing/data': 0.022922751028090715, 'timing/model': 1.1401640819967724, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 0.18}) +[2026-06-15 19:50:46,070][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.11471302062273026, 'train/grad_norm_pre_clip': 11.554883003234863, 'timing/data': 0.02295060601318255, 'timing/model': 1.3138279739650898, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 0.18}) +[2026-06-15 19:50:50,448][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.10025931149721146, 'train/grad_norm_pre_clip': 10.273547172546387, 'timing/data': 0.023284088994842023, 'timing/model': 1.1336437120335177, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.18}) +[2026-06-15 19:50:55,013][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.18437515199184418, 'train/grad_norm_pre_clip': 6.223361968994141, 'timing/data': 0.02382360096089542, 'timing/model': 1.1406866789911874, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 0.18}) +[2026-06-15 19:50:59,556][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.503315269947052, 'train/grad_norm_pre_clip': 21.05546760559082, 'timing/data': 0.024149052973371, 'timing/model': 1.1301081650308333, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 0.18}) +[2026-06-15 19:51:04,107][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.33418092131614685, 'train/grad_norm_pre_clip': 12.011322975158691, 'timing/data': 0.0232762600062415, 'timing/model': 1.142456070985645, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 0.18}) +[2026-06-15 19:51:08,658][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.39873215556144714, 'train/grad_norm_pre_clip': 12.507650375366211, 'timing/data': 0.023607852985151112, 'timing/model': 1.3106809029704891, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 0.18}) +[2026-06-15 19:51:13,164][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.18545514345169067, 'train/grad_norm_pre_clip': 14.808842658996582, 'timing/data': 0.023716577037703246, 'timing/model': 1.268362333998084, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 0.18}) +[2026-06-15 19:51:17,539][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.2776970863342285, 'train/grad_norm_pre_clip': 13.311335563659668, 'timing/data': 0.024074211018159986, 'timing/model': 1.1309693529619835, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 0.18}) +[2026-06-15 19:51:22,081][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.12344053387641907, 'train/grad_norm_pre_clip': 3.275632858276367, 'timing/data': 0.02363942301599309, 'timing/model': 1.141314771026373, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.18}) +[2026-06-15 19:51:26,631][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.2565125524997711, 'train/grad_norm_pre_clip': 10.392199516296387, 'timing/data': 0.024287476029712707, 'timing/model': 1.1461797810043208, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 0.19}) +[2026-06-15 19:51:31,156][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.44091689586639404, 'train/grad_norm_pre_clip': 11.695538520812988, 'timing/data': 0.023449303000234067, 'timing/model': 1.128041520016268, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 0.19}) +[2026-06-15 19:51:35,635][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.33999696373939514, 'train/grad_norm_pre_clip': 11.56104564666748, 'timing/data': 0.0228667400078848, 'timing/model': 1.1296211410081014, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 0.19}) +[2026-06-15 19:51:40,200][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.18789459764957428, 'train/grad_norm_pre_clip': 10.292492866516113, 'timing/data': 0.02285507699707523, 'timing/model': 1.3142170690116473, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 0.19}) +[2026-06-15 19:51:44,586][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.44956687092781067, 'train/grad_norm_pre_clip': 18.524654388427734, 'timing/data': 0.02338711899938062, 'timing/model': 1.1404741189908236, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 0.19}) +[2026-06-15 19:51:49,146][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.231199711561203, 'train/grad_norm_pre_clip': 8.719747543334961, 'timing/data': 0.0230660040397197, 'timing/model': 1.145389882964082, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 0.19}) +[2026-06-15 19:51:53,714][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.45625537633895874, 'train/grad_norm_pre_clip': 15.168895721435547, 'timing/data': 0.023485590994823724, 'timing/model': 1.1427850839681923, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 0.19}) +[2026-06-15 19:51:58,256][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.597054660320282, 'train/grad_norm_pre_clip': 16.262983322143555, 'timing/data': 0.023430561006534845, 'timing/model': 1.1348026750492863, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 0.19}) +[2026-06-15 19:52:02,825][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.8445395231246948, 'train/grad_norm_pre_clip': 21.74485206604004, 'timing/data': 0.023119791003409773, 'timing/model': 1.1489701649989001, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 0.19}) +[2026-06-15 19:52:07,387][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.284371942281723, 'train/grad_norm_pre_clip': 9.857746124267578, 'timing/data': 0.023455661023035645, 'timing/model': 1.3040929079870693, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.19}) +[2026-06-15 19:52:11,774][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.282795250415802, 'train/grad_norm_pre_clip': 14.76682186126709, 'timing/data': 0.023722232959698886, 'timing/model': 1.1328108300222084, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 0.19}) +[2026-06-15 19:52:16,317][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.23202650249004364, 'train/grad_norm_pre_clip': 9.342248916625977, 'timing/data': 0.024249481968581676, 'timing/model': 1.132339552976191, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 0.19}) +[2026-06-15 19:52:20,888][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.2505306601524353, 'train/grad_norm_pre_clip': 6.869845390319824, 'timing/data': 0.02364712196867913, 'timing/model': 1.1429949310258962, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 0.19}) +[2026-06-15 19:52:25,456][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.3130026161670685, 'train/grad_norm_pre_clip': 12.202513694763184, 'timing/data': 0.022834570962004364, 'timing/model': 1.1449045190238394, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 0.19}) +[2026-06-15 19:52:30,001][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.13672477006912231, 'train/grad_norm_pre_clip': 9.086787223815918, 'timing/data': 0.02363987098215148, 'timing/model': 1.150320953980554, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 0.19}) +[2026-06-15 19:52:34,569][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.376577228307724, 'train/grad_norm_pre_clip': 25.629512786865234, 'timing/data': 0.02346440003020689, 'timing/model': 1.3038806010154076, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 0.19}) +[2026-06-15 19:52:38,959][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.2627177834510803, 'train/grad_norm_pre_clip': 8.812173843383789, 'timing/data': 0.023574489983730018, 'timing/model': 1.1442910029436462, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 0.19}) +[2026-06-15 19:52:43,497][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.222134530544281, 'train/grad_norm_pre_clip': 11.735350608825684, 'timing/data': 0.023248954967129976, 'timing/model': 1.1444458679761738, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 0.19}) +[2026-06-15 19:52:48,042][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.4184017479419708, 'train/grad_norm_pre_clip': 10.21378231048584, 'timing/data': 0.023339029983617365, 'timing/model': 1.1361977430060506, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 0.19}) +[2026-06-15 19:52:52,577][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.27989834547042847, 'train/grad_norm_pre_clip': 10.416901588439941, 'timing/data': 0.023922449036035687, 'timing/model': 1.1308284970000386, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.19}) +[2026-06-15 19:52:57,126][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.19198137521743774, 'train/grad_norm_pre_clip': 13.151787757873535, 'timing/data': 0.023837675980757922, 'timing/model': 1.1374603990116157, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 0.19}) +[2026-06-15 19:53:01,688][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.24725212156772614, 'train/grad_norm_pre_clip': 7.430857181549072, 'timing/data': 0.023595455975737423, 'timing/model': 1.3159500870388001, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 0.19}) +[2026-06-15 19:53:06,073][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.10646898299455643, 'train/grad_norm_pre_clip': 9.280061721801758, 'timing/data': 0.02446567901642993, 'timing/model': 1.1320099050062709, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 0.19}) +[2026-06-15 19:53:10,627][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.2219819277524948, 'train/grad_norm_pre_clip': 15.064093589782715, 'timing/data': 0.022850946988910437, 'timing/model': 1.135035949992016, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 0.19}) +[2026-06-15 19:53:15,185][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.23304498195648193, 'train/grad_norm_pre_clip': 19.2536563873291, 'timing/data': 0.02301817201077938, 'timing/model': 1.1511124219978228, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 0.2}) +[2026-06-15 19:53:19,711][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.2728409171104431, 'train/grad_norm_pre_clip': 13.786544799804688, 'timing/data': 0.023865699011366814, 'timing/model': 1.1330831099767238, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 0.2}) +[2026-06-15 19:53:24,267][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.11995729058980942, 'train/grad_norm_pre_clip': 3.913938045501709, 'timing/data': 0.023377228993922472, 'timing/model': 1.29016135801794, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 0.2}) +[2026-06-15 19:53:28,815][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.3645080029964447, 'train/grad_norm_pre_clip': 11.245588302612305, 'timing/data': 0.023723872029222548, 'timing/model': 1.2875045469845645, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 0.2}) +[2026-06-15 19:53:33,188][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.15429626405239105, 'train/grad_norm_pre_clip': 19.302814483642578, 'timing/data': 0.02417641202919185, 'timing/model': 1.1272720959968865, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 0.2}) +[2026-06-15 19:53:37,724][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.09994466602802277, 'train/grad_norm_pre_clip': 21.899904251098633, 'timing/data': 0.023343157023191452, 'timing/model': 1.132766402035486, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.2}) +[2026-06-15 19:53:42,289][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.37481558322906494, 'train/grad_norm_pre_clip': 20.12615203857422, 'timing/data': 0.023943710955791175, 'timing/model': 1.1316398360067979, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 0.2}) +[2026-06-15 19:53:46,820][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.2681036591529846, 'train/grad_norm_pre_clip': 21.183134078979492, 'timing/data': 0.023854769009631127, 'timing/model': 1.1297191680059768, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 0.2}) +[2026-06-15 19:53:51,302][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.18657058477401733, 'train/grad_norm_pre_clip': 6.565978527069092, 'timing/data': 0.024248422007076442, 'timing/model': 1.1293654430191964, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 0.2}) +[2026-06-15 19:53:55,843][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.1991363763809204, 'train/grad_norm_pre_clip': 15.423686981201172, 'timing/data': 0.023936823999974877, 'timing/model': 1.2908231539768167, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 0.2}) +[2026-06-15 19:54:00,227][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.10190719366073608, 'train/grad_norm_pre_clip': 10.826282501220703, 'timing/data': 0.023553004022687674, 'timing/model': 1.1348768780007958, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.2}) +[2026-06-15 19:54:04,770][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.3320760428905487, 'train/grad_norm_pre_clip': 25.32834815979004, 'timing/data': 0.024039336945861578, 'timing/model': 1.1321763269952498, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 0.2}) +[2026-06-15 19:54:09,336][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.34903860092163086, 'train/grad_norm_pre_clip': 20.70318603515625, 'timing/data': 0.02363406802760437, 'timing/model': 1.1451095609809272, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 0.2}) +[2026-06-15 19:54:13,870][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.25022733211517334, 'train/grad_norm_pre_clip': 27.43531608581543, 'timing/data': 0.023287413001526147, 'timing/model': 1.1345399969723076, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 0.2}) +[2026-06-15 19:54:18,412][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.2941017150878906, 'train/grad_norm_pre_clip': 9.870936393737793, 'timing/data': 0.024076010973658413, 'timing/model': 1.1376973820151761, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 0.2}) +[2026-06-15 19:54:22,969][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.15652257204055786, 'train/grad_norm_pre_clip': 12.696961402893066, 'timing/data': 0.02348645799793303, 'timing/model': 1.3072263249778189, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 0.2}) +[2026-06-15 19:54:27,361][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.1396872103214264, 'train/grad_norm_pre_clip': 7.731014728546143, 'timing/data': 0.02318258601007983, 'timing/model': 1.141192416951526, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 0.2}) +[2026-06-15 19:54:31,913][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.4087654650211334, 'train/grad_norm_pre_clip': 9.999442100524902, 'timing/data': 0.023291790974326432, 'timing/model': 1.137928269978147, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.2}) +[2026-06-15 19:54:36,473][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.34741833806037903, 'train/grad_norm_pre_clip': 23.807058334350586, 'timing/data': 0.023905681038741022, 'timing/model': 1.1355195060023107, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 0.2}) +[2026-06-15 19:54:40,999][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.17174425721168518, 'train/grad_norm_pre_clip': 7.146265029907227, 'timing/data': 0.022550318972207606, 'timing/model': 1.1347618230502121, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 0.2}) +[2026-06-15 19:54:45,510][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.21875038743019104, 'train/grad_norm_pre_clip': 7.680849552154541, 'timing/data': 0.023738249961752445, 'timing/model': 1.1296304800198413, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 0.2}) +[2026-06-15 19:54:50,065][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.1929347962141037, 'train/grad_norm_pre_clip': 7.7322821617126465, 'timing/data': 0.024394358973950148, 'timing/model': 1.2963738440303132, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 0.2}) +[2026-06-15 19:54:54,447][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.23434855043888092, 'train/grad_norm_pre_clip': 9.5479097366333, 'timing/data': 0.023630911950021982, 'timing/model': 1.1396112790098414, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 0.2}) +[2026-06-15 19:54:58,999][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.2388630509376526, 'train/grad_norm_pre_clip': 9.336666107177734, 'timing/data': 0.023140105011407286, 'timing/model': 1.133698484976776, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 0.2}) +[2026-06-15 19:55:03,553][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.21310892701148987, 'train/grad_norm_pre_clip': 6.094557762145996, 'timing/data': 0.023560546978842467, 'timing/model': 1.1321479640319012, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 0.2}) +[2026-06-15 19:59:20,930][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.2473788559436798, 'train/grad_norm_pre_clip': 18.47303009033203, 'eval/loss': 0.11396338790655136, 'eval/flap_precision': 0.8595744680851064, 'eval/flap_recall': 0.9351851851851852, 'eval/flap_f1': 0.8957871396895788, 'eval/flap_rate_model': 0.19583333333333333, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.8957871396895788, 'timing/data': 0.02272979501867667, 'timing/model': 1.1341401970130391, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 0.21}) +[2026-06-15 19:59:20,931][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_500_state +[2026-06-15 19:59:20,933][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 19:59:24,482][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_500_state/model.safetensors +[2026-06-15 19:59:34,518][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_500_state/optimizer.bin +[2026-06-15 19:59:34,518][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_500_state/sampler.bin +[2026-06-15 19:59:34,518][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_500_state/sampler_1.bin +[2026-06-15 19:59:34,519][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-15 19:59:34,565][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-15 19:59:34,565][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 19:59:34,572][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 19:59:39,028][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.07062177360057831, 'train/grad_norm_pre_clip': 2.7335691452026367, 'timing/data': 0.026326161983888596, 'timing/model': 1.1147262189770117, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 0.21}) +[2026-06-15 19:59:43,403][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.12126170098781586, 'train/grad_norm_pre_clip': 6.511258125305176, 'timing/data': 0.023927956994157284, 'timing/model': 1.1111987570184283, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 0.21}) +[2026-06-15 19:59:47,806][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.27747756242752075, 'train/grad_norm_pre_clip': 36.57475280761719, 'timing/data': 0.023964609019458294, 'timing/model': 1.1127887470065616, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 0.21}) +[2026-06-15 19:59:52,252][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.25345510244369507, 'train/grad_norm_pre_clip': 9.79344654083252, 'timing/data': 0.023526917037088424, 'timing/model': 1.248616925033275, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 0.21}) +[2026-06-15 19:59:56,564][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.32377368211746216, 'train/grad_norm_pre_clip': 9.992652893066406, 'timing/data': 0.024091039958875626, 'timing/model': 1.1171879599569365, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 0.21}) +[2026-06-15 20:00:01,054][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.4565883278846741, 'train/grad_norm_pre_clip': 20.251798629760742, 'timing/data': 0.02329779596766457, 'timing/model': 1.1208579260273837, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 0.21}) +[2026-06-15 20:00:05,567][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.0997343361377716, 'train/grad_norm_pre_clip': 7.022364139556885, 'timing/data': 0.022959112015087157, 'timing/model': 1.12101993302349, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 0.21}) +[2026-06-15 20:00:10,041][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.21353811025619507, 'train/grad_norm_pre_clip': 21.693653106689453, 'timing/data': 0.023246584984008223, 'timing/model': 1.116634468024131, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 0.21}) +[2026-06-15 20:00:14,523][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.4035888910293579, 'train/grad_norm_pre_clip': 6.823553562164307, 'timing/data': 0.02322336402721703, 'timing/model': 1.1241193269961514, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 0.21}) +[2026-06-15 20:00:19,042][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.4411005675792694, 'train/grad_norm_pre_clip': 12.950056076049805, 'timing/data': 0.022646309982519597, 'timing/model': 1.2816328720073216, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 0.21}) +[2026-06-15 20:00:23,387][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.26225078105926514, 'train/grad_norm_pre_clip': 9.67424488067627, 'timing/data': 0.023337623977568, 'timing/model': 1.1281291110208258, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 0.21}) +[2026-06-15 20:00:27,877][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.23998679220676422, 'train/grad_norm_pre_clip': 9.720466613769531, 'timing/data': 0.02302622696151957, 'timing/model': 1.122487088025082, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 0.21}) +[2026-06-15 20:00:32,381][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.3306814134120941, 'train/grad_norm_pre_clip': 12.83255386352539, 'timing/data': 0.023213290027342737, 'timing/model': 1.1286059889825992, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 0.21}) +[2026-06-15 20:00:36,906][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.2242911458015442, 'train/grad_norm_pre_clip': 14.795186042785645, 'timing/data': 0.022659829002805054, 'timing/model': 1.133793493034318, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 0.21}) +[2026-06-15 20:00:41,451][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.2587728202342987, 'train/grad_norm_pre_clip': 7.830872058868408, 'timing/data': 0.02303814998595044, 'timing/model': 1.2933422799687833, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 0.21}) +[2026-06-15 20:00:45,800][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.14162322878837585, 'train/grad_norm_pre_clip': 4.493852615356445, 'timing/data': 0.02321049402235076, 'timing/model': 1.1234476810204796, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 0.21}) +[2026-06-15 20:00:50,300][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.21150259673595428, 'train/grad_norm_pre_clip': 5.825085163116455, 'timing/data': 0.023123714025132358, 'timing/model': 1.126046746969223, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 0.21}) +[2026-06-15 20:00:54,802][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.31960350275039673, 'train/grad_norm_pre_clip': 10.303396224975586, 'timing/data': 0.02300326299155131, 'timing/model': 1.1249918160028756, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 0.21}) +[2026-06-15 20:00:59,328][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.23698674142360687, 'train/grad_norm_pre_clip': 10.149840354919434, 'timing/data': 0.023729959968477488, 'timing/model': 1.135156100965105, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 0.21}) +[2026-06-15 20:01:03,840][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.21638774871826172, 'train/grad_norm_pre_clip': 9.645040512084961, 'timing/data': 0.023078283003997058, 'timing/model': 1.128019119030796, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.21}) +[2026-06-15 20:01:08,356][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.22212575376033783, 'train/grad_norm_pre_clip': 7.600390911102295, 'timing/data': 0.02307594899320975, 'timing/model': 1.28515490796417, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 0.21}) +[2026-06-15 20:01:12,722][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.3138924837112427, 'train/grad_norm_pre_clip': 9.05462646484375, 'timing/data': 0.02331102400785312, 'timing/model': 1.1295272120041773, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 0.21}) +[2026-06-15 20:01:17,248][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.3282769024372101, 'train/grad_norm_pre_clip': 20.602617263793945, 'timing/data': 0.024600750009994954, 'timing/model': 1.1304187509813346, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 0.21}) +[2026-06-15 20:01:21,780][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.2060760259628296, 'train/grad_norm_pre_clip': 7.386874675750732, 'timing/data': 0.023841143993195146, 'timing/model': 1.1293738060048781, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 0.22}) +[2026-06-15 20:01:26,277][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.27900639176368713, 'train/grad_norm_pre_clip': 13.099897384643555, 'timing/data': 0.024118822999298573, 'timing/model': 1.1292947370093316, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 0.22}) +[2026-06-15 20:01:30,832][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.14794921875, 'train/grad_norm_pre_clip': 13.431612968444824, 'timing/data': 0.02308801602339372, 'timing/model': 1.1286925959866494, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 0.22}) +[2026-06-15 20:01:35,387][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.08871540427207947, 'train/grad_norm_pre_clip': 15.058276176452637, 'timing/data': 0.023733116046059877, 'timing/model': 1.305960048979614, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 0.22}) +[2026-06-15 20:01:39,775][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.20760677754878998, 'train/grad_norm_pre_clip': 6.501071929931641, 'timing/data': 0.022532375005539507, 'timing/model': 1.1409718820359558, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 0.22}) +[2026-06-15 20:01:44,297][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.3286028206348419, 'train/grad_norm_pre_clip': 9.718762397766113, 'timing/data': 0.023163308040238917, 'timing/model': 1.1292908189934678, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 0.22}) +[2026-06-15 20:01:48,751][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.2116132527589798, 'train/grad_norm_pre_clip': 6.367806434631348, 'timing/data': 0.023652484989725053, 'timing/model': 1.152806078956928, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 0.22}) +[2026-06-15 20:01:53,326][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.2869013845920563, 'train/grad_norm_pre_clip': 12.485953330993652, 'timing/data': 0.023310774995479733, 'timing/model': 1.1413305479800329, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 0.22}) +[2026-06-15 20:01:57,877][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.41098856925964355, 'train/grad_norm_pre_clip': 21.235549926757812, 'timing/data': 0.023574542021378875, 'timing/model': 1.1318679960095324, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 0.22}) +[2026-06-15 20:02:02,437][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.3191508650779724, 'train/grad_norm_pre_clip': 23.537240982055664, 'timing/data': 0.023895149002783, 'timing/model': 1.3082687940332107, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 0.22}) +[2026-06-15 20:02:06,822][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.15781404078006744, 'train/grad_norm_pre_clip': 5.261622905731201, 'timing/data': 0.02337080001598224, 'timing/model': 1.1397388819605112, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 0.22}) +[2026-06-15 20:02:11,370][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.28819945454597473, 'train/grad_norm_pre_clip': 13.330757141113281, 'timing/data': 0.023153426009230316, 'timing/model': 1.1415702719823457, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 0.22}) +[2026-06-15 20:02:15,919][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.4455479383468628, 'train/grad_norm_pre_clip': 27.612375259399414, 'timing/data': 0.023885642003733665, 'timing/model': 1.1368007129640318, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 0.22}) +[2026-06-15 20:02:20,418][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.28702521324157715, 'train/grad_norm_pre_clip': 10.339752197265625, 'timing/data': 0.023647028021514416, 'timing/model': 1.128602373006288, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 0.22}) +[2026-06-15 20:02:24,932][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.1489754319190979, 'train/grad_norm_pre_clip': 11.459786415100098, 'timing/data': 0.023558225017040968, 'timing/model': 1.1303391349501908, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 0.22}) +[2026-06-15 20:02:29,491][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.24196746945381165, 'train/grad_norm_pre_clip': 11.178788185119629, 'timing/data': 0.023760681971907616, 'timing/model': 1.3048828419996426, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 0.22}) +[2026-06-15 20:02:33,884][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.09168782830238342, 'train/grad_norm_pre_clip': 5.4797844886779785, 'timing/data': 0.02306199399754405, 'timing/model': 1.1333160279900767, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 0.22}) +[2026-06-15 20:02:38,435][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.08660050481557846, 'train/grad_norm_pre_clip': 4.395547389984131, 'timing/data': 0.023216759029310197, 'timing/model': 1.133980699989479, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.22}) +[2026-06-15 20:02:43,000][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.26064175367355347, 'train/grad_norm_pre_clip': 16.557682037353516, 'timing/data': 0.024724716029595584, 'timing/model': 1.1464959069853649, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 0.22}) +[2026-06-15 20:02:47,545][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.08418900519609451, 'train/grad_norm_pre_clip': 23.03671646118164, 'timing/data': 0.023334579018410295, 'timing/model': 1.1283642229973339, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 0.22}) +[2026-06-15 20:02:52,094][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.2369026392698288, 'train/grad_norm_pre_clip': 14.30082893371582, 'timing/data': 0.02420642098877579, 'timing/model': 1.2948197629884817, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 0.22}) +[2026-06-15 20:02:56,476][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.17213396728038788, 'train/grad_norm_pre_clip': 13.70228099822998, 'timing/data': 0.0235801039962098, 'timing/model': 1.1369779270025901, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 0.22}) +[2026-06-15 20:03:01,005][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.06807467341423035, 'train/grad_norm_pre_clip': 6.909569263458252, 'timing/data': 0.023295253980904818, 'timing/model': 1.13532798999222, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 0.22}) +[2026-06-15 20:03:05,555][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.29475468397140503, 'train/grad_norm_pre_clip': 18.84016990661621, 'timing/data': 0.023615079000592232, 'timing/model': 1.1402920130058192, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 0.22}) +[2026-06-15 20:03:10,139][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.44008445739746094, 'train/grad_norm_pre_clip': 13.830362319946289, 'timing/data': 0.024173879995942116, 'timing/model': 1.1332781129749492, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 0.23}) +[2026-06-15 20:03:14,673][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.13345077633857727, 'train/grad_norm_pre_clip': 26.75077247619629, 'timing/data': 0.02345255803084001, 'timing/model': 1.1265373679925688, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 0.23}) +[2026-06-15 20:03:19,218][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.23156097531318665, 'train/grad_norm_pre_clip': 13.827963829040527, 'timing/data': 0.023380158992949873, 'timing/model': 1.2950826580054127, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 0.23}) +[2026-06-15 20:03:23,585][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.040128499269485474, 'train/grad_norm_pre_clip': 6.883756160736084, 'timing/data': 0.02355852397158742, 'timing/model': 1.1318620989914052, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 0.23}) +[2026-06-15 20:03:28,117][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.3177468478679657, 'train/grad_norm_pre_clip': 29.39767074584961, 'timing/data': 0.023969011032022536, 'timing/model': 1.1370470529654995, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 0.23}) +[2026-06-15 20:03:32,652][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.5325936079025269, 'train/grad_norm_pre_clip': 25.880022048950195, 'timing/data': 0.02403253800002858, 'timing/model': 1.1303111860179342, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 0.23}) +[2026-06-15 20:03:37,140][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.30803072452545166, 'train/grad_norm_pre_clip': 16.22262954711914, 'timing/data': 0.023519025999121368, 'timing/model': 1.130846759013366, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 0.23}) +[2026-06-15 20:03:41,691][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.23176419734954834, 'train/grad_norm_pre_clip': 9.812143325805664, 'timing/data': 0.02291426999727264, 'timing/model': 1.1400330609758385, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.23}) +[2026-06-15 20:03:46,271][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.3364390730857849, 'train/grad_norm_pre_clip': 23.9943904876709, 'timing/data': 0.023410814988892525, 'timing/model': 1.291006567014847, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 0.23}) +[2026-06-15 20:03:50,656][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.20249809324741364, 'train/grad_norm_pre_clip': 7.977942943572998, 'timing/data': 0.03226139099569991, 'timing/model': 1.1325337769812904, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 0.23}) +[2026-06-15 20:03:55,212][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.2104196697473526, 'train/grad_norm_pre_clip': 8.448715209960938, 'timing/data': 0.02335521997883916, 'timing/model': 1.14046820695512, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 0.23}) +[2026-06-15 20:03:59,681][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.13571451604366302, 'train/grad_norm_pre_clip': 16.256898880004883, 'timing/data': 0.023300142027437687, 'timing/model': 1.135553554049693, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 0.23}) +[2026-06-15 20:04:04,239][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.2765563428401947, 'train/grad_norm_pre_clip': 10.861233711242676, 'timing/data': 0.023763165983837098, 'timing/model': 1.131329684983939, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 0.23}) +[2026-06-15 20:04:08,791][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.21666303277015686, 'train/grad_norm_pre_clip': 9.258106231689453, 'timing/data': 0.02359467203496024, 'timing/model': 1.128146507020574, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 0.23}) +[2026-06-15 20:04:13,307][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.23594030737876892, 'train/grad_norm_pre_clip': 8.096903800964355, 'timing/data': 0.024219303973950446, 'timing/model': 1.2796346570248716, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 0.23}) +[2026-06-15 20:04:17,686][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.30607160925865173, 'train/grad_norm_pre_clip': 9.435281753540039, 'timing/data': 0.023696820018813014, 'timing/model': 1.1375224700313993, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 0.23}) +[2026-06-15 20:04:22,219][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.1926758736371994, 'train/grad_norm_pre_clip': 7.380899906158447, 'timing/data': 0.023729328997433186, 'timing/model': 1.1345248150173575, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 0.23}) +[2026-06-15 20:04:26,775][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.13670076429843903, 'train/grad_norm_pre_clip': 7.015565872192383, 'timing/data': 0.02335440303431824, 'timing/model': 1.1383194029913284, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 0.23}) +[2026-06-15 20:04:31,282][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.3994794189929962, 'train/grad_norm_pre_clip': 16.53459358215332, 'timing/data': 0.02340034901862964, 'timing/model': 1.1478753549745306, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 0.23}) +[2026-06-15 20:04:35,822][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.2738681435585022, 'train/grad_norm_pre_clip': 10.982539176940918, 'timing/data': 0.023969800968188792, 'timing/model': 1.138390353997238, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 0.23}) +[2026-06-15 20:04:40,377][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.17510633170604706, 'train/grad_norm_pre_clip': 6.254821300506592, 'timing/data': 0.02389876398956403, 'timing/model': 1.305884787987452, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 0.23}) +[2026-06-15 20:04:44,756][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.1306828111410141, 'train/grad_norm_pre_clip': 10.761049270629883, 'timing/data': 0.023403503990266472, 'timing/model': 1.130803701991681, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 0.23}) +[2026-06-15 20:04:49,312][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.22937056422233582, 'train/grad_norm_pre_clip': 15.417309761047363, 'timing/data': 0.02305564604466781, 'timing/model': 1.1342559769982472, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 0.23}) +[2026-06-15 20:04:53,864][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.09236157685518265, 'train/grad_norm_pre_clip': 7.992326259613037, 'timing/data': 0.02403299603611231, 'timing/model': 1.12935160897905, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 0.23}) +[2026-06-15 20:04:58,438][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.3676449656486511, 'train/grad_norm_pre_clip': 54.300228118896484, 'timing/data': 0.023489664017688483, 'timing/model': 1.1502442099736072, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 0.23}) +[2026-06-15 20:05:03,004][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.0645909532904625, 'train/grad_norm_pre_clip': 4.973995208740234, 'timing/data': 0.026600485027302057, 'timing/model': 1.3065784290083684, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 0.24}) +[2026-06-15 20:05:07,395][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.13649554550647736, 'train/grad_norm_pre_clip': 8.576025009155273, 'timing/data': 0.02381687000161037, 'timing/model': 1.1322512610349804, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 0.24}) +[2026-06-15 20:05:11,972][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.0593310222029686, 'train/grad_norm_pre_clip': 5.1591949462890625, 'timing/data': 0.02347945503424853, 'timing/model': 1.1647279360331595, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 0.24}) +[2026-06-15 20:05:16,509][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.3715461194515228, 'train/grad_norm_pre_clip': 12.319304466247559, 'timing/data': 0.0234176570083946, 'timing/model': 1.1410842780023813, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 0.24}) +[2026-06-15 20:05:21,049][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.4121243953704834, 'train/grad_norm_pre_clip': 17.194950103759766, 'timing/data': 0.023803341027814895, 'timing/model': 1.1315456979791634, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 0.24}) +[2026-06-15 20:05:25,611][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.2693137526512146, 'train/grad_norm_pre_clip': 14.690108299255371, 'timing/data': 0.02386757096974179, 'timing/model': 1.1466100459801964, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 0.24}) +[2026-06-15 20:05:30,204][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.36433395743370056, 'train/grad_norm_pre_clip': 39.29509353637695, 'timing/data': 0.024400763970334083, 'timing/model': 1.330513452005107, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 0.24}) +[2026-06-15 20:05:34,611][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.07169635593891144, 'train/grad_norm_pre_clip': 8.537169456481934, 'timing/data': 0.023278188018593937, 'timing/model': 1.1477313269861042, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 0.24}) +[2026-06-15 20:05:39,162][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.14480674266815186, 'train/grad_norm_pre_clip': 12.179604530334473, 'timing/data': 0.023588381009176373, 'timing/model': 1.1312871979898773, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 0.24}) +[2026-06-15 20:05:43,720][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.2655973732471466, 'train/grad_norm_pre_clip': 15.827776908874512, 'timing/data': 0.024257721030153334, 'timing/model': 1.1370936590246856, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 0.24}) +[2026-06-15 20:05:48,213][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.17272396385669708, 'train/grad_norm_pre_clip': 10.52418041229248, 'timing/data': 0.023608982970472425, 'timing/model': 1.1309792840038426, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 0.24}) +[2026-06-15 20:05:52,771][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.1718757301568985, 'train/grad_norm_pre_clip': 11.959263801574707, 'timing/data': 0.02390894095879048, 'timing/model': 1.1332473690272309, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 0.24}) +[2026-06-15 20:05:57,324][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.1772276908159256, 'train/grad_norm_pre_clip': 11.20285701751709, 'timing/data': 0.023315584985539317, 'timing/model': 1.3047946710139513, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 0.24}) +[2026-06-15 20:06:01,700][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.20604819059371948, 'train/grad_norm_pre_clip': 11.417450904846191, 'timing/data': 0.023324960959143937, 'timing/model': 1.1341801389935426, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 0.24}) +[2026-06-15 20:06:06,237][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.13323962688446045, 'train/grad_norm_pre_clip': 7.551214694976807, 'timing/data': 0.0236649620346725, 'timing/model': 1.1339245970011689, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 0.24}) +[2026-06-15 20:06:10,672][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.0892002061009407, 'train/grad_norm_pre_clip': 7.0364274978637695, 'timing/data': 0.02384014701237902, 'timing/model': 1.134272956987843, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 0.24}) +[2026-06-15 20:06:15,222][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.3895995616912842, 'train/grad_norm_pre_clip': 14.316216468811035, 'timing/data': 0.023809716978576034, 'timing/model': 1.130575972958468, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 0.24}) +[2026-06-15 20:06:19,765][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.06086934730410576, 'train/grad_norm_pre_clip': 5.852949619293213, 'timing/data': 0.02350742998532951, 'timing/model': 1.1330443160259165, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.24}) +[2026-06-15 20:06:24,319][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.09160185605287552, 'train/grad_norm_pre_clip': 5.5632100105285645, 'timing/data': 0.02338239998789504, 'timing/model': 1.297204053029418, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 0.24}) +[2026-06-15 20:06:28,706][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.5061514377593994, 'train/grad_norm_pre_clip': 12.436445236206055, 'timing/data': 0.02344677795190364, 'timing/model': 1.1460799289634451, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 0.24}) +[2026-06-15 20:06:33,251][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.210636168718338, 'train/grad_norm_pre_clip': 6.477320671081543, 'timing/data': 0.024001498997677118, 'timing/model': 1.1324841800378636, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 0.24}) +[2026-06-15 20:06:37,799][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.13298660516738892, 'train/grad_norm_pre_clip': 8.185009956359863, 'timing/data': 0.023748878971673548, 'timing/model': 1.1321529970155098, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 0.24}) +[2026-06-15 20:06:42,286][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.13163812458515167, 'train/grad_norm_pre_clip': 7.347540855407715, 'timing/data': 0.023617705039214343, 'timing/model': 1.1301500559784472, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 0.24}) +[2026-06-15 20:06:46,812][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.1770506650209427, 'train/grad_norm_pre_clip': 7.3881378173828125, 'timing/data': 0.023610146017745137, 'timing/model': 1.1350913959904574, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 0.24}) +[2026-06-15 20:06:51,374][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.5278252363204956, 'train/grad_norm_pre_clip': 17.132369995117188, 'timing/data': 0.023021390021312982, 'timing/model': 1.3147589490399696, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 0.25}) +[2026-06-15 20:06:55,750][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.544700562953949, 'train/grad_norm_pre_clip': 11.683121681213379, 'timing/data': 0.02305475703906268, 'timing/model': 1.1287416170234792, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 0.25}) +[2026-06-15 20:07:00,297][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.4809972643852234, 'train/grad_norm_pre_clip': 13.409945487976074, 'timing/data': 0.023285005998332053, 'timing/model': 1.1328324839705601, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 0.25}) +[2026-06-15 20:11:23,773][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.3700460195541382, 'train/grad_norm_pre_clip': 16.10568618774414, 'eval/loss': 0.1435386836528778, 'eval/flap_precision': 0.8918406072106262, 'eval/flap_recall': 0.7253086419753086, 'eval/flap_f1': 0.7999999999999999, 'eval/flap_rate_model': 0.1463888888888889, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.8, 'timing/data': 0.023262066999450326, 'timing/model': 1.1345747379818931, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 0.25}) +[2026-06-15 20:11:23,773][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_600_state +[2026-06-15 20:11:23,776][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:11:27,344][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_600_state/model.safetensors +[2026-06-15 20:11:37,298][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_600_state/optimizer.bin +[2026-06-15 20:11:37,298][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_600_state/sampler.bin +[2026-06-15 20:11:37,298][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_600_state/sampler_1.bin +[2026-06-15 20:11:37,299][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_600_state/random_states_0.pkl +[2026-06-15 20:11:37,345][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600 +[2026-06-15 20:11:37,345][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:11:37,351][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:11:41,776][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.30430641770362854, 'train/grad_norm_pre_clip': 10.423962593078613, 'timing/data': 0.02328176802257076, 'timing/model': 1.1101334140403196, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 0.25}) +[2026-06-15 20:11:46,258][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.17558366060256958, 'train/grad_norm_pre_clip': 4.453216075897217, 'timing/data': 0.022812067007180303, 'timing/model': 1.1125975760514848, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 0.25}) +[2026-06-15 20:11:50,725][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.24088412523269653, 'train/grad_norm_pre_clip': 15.812355995178223, 'timing/data': 0.023349890019744635, 'timing/model': 1.1164674370083958, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 0.25}) +[2026-06-15 20:11:55,200][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.1776355803012848, 'train/grad_norm_pre_clip': 23.549409866333008, 'timing/data': 0.02307430299697444, 'timing/model': 1.2781275830348022, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 0.25}) +[2026-06-15 20:11:59,534][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.494340181350708, 'train/grad_norm_pre_clip': 10.87520980834961, 'timing/data': 0.02283292298670858, 'timing/model': 1.1246058769756928, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 0.25}) +[2026-06-15 20:12:04,053][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.4301585555076599, 'train/grad_norm_pre_clip': 13.340063095092773, 'timing/data': 0.023118261015042663, 'timing/model': 1.1273240899899974, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 0.25}) +[2026-06-15 20:12:08,536][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.11086512356996536, 'train/grad_norm_pre_clip': 4.485771656036377, 'timing/data': 0.023277517990209162, 'timing/model': 1.1187139229732566, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 0.25}) +[2026-06-15 20:12:13,019][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.22864775359630585, 'train/grad_norm_pre_clip': 8.35901927947998, 'timing/data': 0.02307145099621266, 'timing/model': 1.1220114040188491, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 0.25}) +[2026-06-15 20:12:17,529][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.20505370199680328, 'train/grad_norm_pre_clip': 8.73313045501709, 'timing/data': 0.02376596798421815, 'timing/model': 1.1220034410362132, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 0.25}) +[2026-06-15 20:12:22,049][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.28948304057121277, 'train/grad_norm_pre_clip': 9.394491195678711, 'timing/data': 0.0234866319806315, 'timing/model': 1.283335025946144, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 0.25}) +[2026-06-15 20:12:26,423][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.11020611226558685, 'train/grad_norm_pre_clip': 4.930091381072998, 'timing/data': 0.02353844401659444, 'timing/model': 1.129575987055432, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 0.25}) +[2026-06-15 20:12:30,953][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.13676060736179352, 'train/grad_norm_pre_clip': 9.071731567382812, 'timing/data': 0.02497764799045399, 'timing/model': 1.1349833459826186, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 0.25}) +[2026-06-15 20:12:35,376][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.1059560552239418, 'train/grad_norm_pre_clip': 4.258868217468262, 'timing/data': 0.023505009012296796, 'timing/model': 1.1271065659821033, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 0.25}) +[2026-06-15 20:12:39,805][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.13406522572040558, 'train/grad_norm_pre_clip': 7.393576622009277, 'timing/data': 0.023426575993653387, 'timing/model': 1.1262059269938618, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 0.25}) +[2026-06-15 20:12:44,267][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.22914621233940125, 'train/grad_norm_pre_clip': 14.752388000488281, 'timing/data': 0.023687546025030315, 'timing/model': 1.1289856439689174, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 0.25}) +[2026-06-15 20:12:48,763][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.19058452546596527, 'train/grad_norm_pre_clip': 13.523749351501465, 'timing/data': 0.023130147019401193, 'timing/model': 1.270749268005602, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 0.25}) +[2026-06-15 20:12:53,148][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.07362330704927444, 'train/grad_norm_pre_clip': 18.694679260253906, 'timing/data': 0.02343935996759683, 'timing/model': 1.1303432119893841, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 0.25}) +[2026-06-15 20:12:57,684][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.027880650013685226, 'train/grad_norm_pre_clip': 4.870663166046143, 'timing/data': 0.023735793016385287, 'timing/model': 1.1313905990100466, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 0.25}) +[2026-06-15 20:13:02,215][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.10915319621562958, 'train/grad_norm_pre_clip': 11.431418418884277, 'timing/data': 0.023533916973974556, 'timing/model': 1.1386217499966733, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 0.25}) +[2026-06-15 20:13:06,743][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.1822982281446457, 'train/grad_norm_pre_clip': 16.438888549804688, 'timing/data': 0.024217395985033363, 'timing/model': 1.1329498490085825, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 0.25}) +[2026-06-15 20:13:11,288][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.2213270515203476, 'train/grad_norm_pre_clip': 8.385225296020508, 'timing/data': 0.022989908989984542, 'timing/model': 1.3029559439746663, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 0.26}) +[2026-06-15 20:13:15,661][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.28358253836631775, 'train/grad_norm_pre_clip': 18.3079833984375, 'timing/data': 0.023623779008630663, 'timing/model': 1.131373910990078, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 0.26}) +[2026-06-15 20:13:20,212][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.13358427584171295, 'train/grad_norm_pre_clip': 11.53541374206543, 'timing/data': 0.023571742000058293, 'timing/model': 1.1449907070491463, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 0.26}) +[2026-06-15 20:13:24,711][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.07596700638532639, 'train/grad_norm_pre_clip': 13.03229808807373, 'timing/data': 0.023526641947682947, 'timing/model': 1.1384695879532956, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 0.26}) +[2026-06-15 20:13:29,250][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.30882954597473145, 'train/grad_norm_pre_clip': 17.152456283569336, 'timing/data': 0.023093905998393893, 'timing/model': 1.1313541019917466, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.26}) +[2026-06-15 20:13:33,808][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.27077561616897583, 'train/grad_norm_pre_clip': 21.794681549072266, 'timing/data': 0.02302339399466291, 'timing/model': 1.1466020680381916, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 0.26}) +[2026-06-15 20:13:38,350][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.12212329357862473, 'train/grad_norm_pre_clip': 9.387377738952637, 'timing/data': 0.023441484023351222, 'timing/model': 1.289997199026402, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 0.26}) +[2026-06-15 20:13:42,868][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.21918985247612, 'train/grad_norm_pre_clip': 27.295616149902344, 'timing/data': 0.023012085002847016, 'timing/model': 1.2546667360002175, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 0.26}) +[2026-06-15 20:13:47,250][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.21451692283153534, 'train/grad_norm_pre_clip': 14.468316078186035, 'timing/data': 0.023304122034460306, 'timing/model': 1.133835025015287, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 0.26}) +[2026-06-15 20:13:51,799][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.014524252153933048, 'train/grad_norm_pre_clip': 1.5438544750213623, 'timing/data': 0.023271936981473118, 'timing/model': 1.1336597700137645, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 0.26}) +[2026-06-15 20:13:56,339][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.15651287138462067, 'train/grad_norm_pre_clip': 11.251015663146973, 'timing/data': 0.022934446984436363, 'timing/model': 1.1432395449955948, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 0.26}) +[2026-06-15 20:14:00,903][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.2320091873407364, 'train/grad_norm_pre_clip': 17.617416381835938, 'timing/data': 0.02321832801681012, 'timing/model': 1.130203357024584, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 0.26}) +[2026-06-15 20:14:05,439][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.2759721279144287, 'train/grad_norm_pre_clip': 13.999258995056152, 'timing/data': 0.02262654999503866, 'timing/model': 1.2930102410027757, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 0.26}) +[2026-06-15 20:14:09,961][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.20567482709884644, 'train/grad_norm_pre_clip': 21.44197654724121, 'timing/data': 0.02299184800358489, 'timing/model': 1.2840476279961877, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 0.26}) +[2026-06-15 20:14:14,339][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.531912088394165, 'train/grad_norm_pre_clip': 16.700122833251953, 'timing/data': 0.023215875029563904, 'timing/model': 1.1357689449796453, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 0.26}) +[2026-06-15 20:14:18,882][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.12979954481124878, 'train/grad_norm_pre_clip': 13.134943962097168, 'timing/data': 0.022557156975381076, 'timing/model': 1.1376437520375475, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 0.26}) +[2026-06-15 20:14:23,429][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.2226855754852295, 'train/grad_norm_pre_clip': 12.321974754333496, 'timing/data': 0.022884086007252336, 'timing/model': 1.1327985430252738, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 0.26}) +[2026-06-15 20:14:27,971][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.09966576099395752, 'train/grad_norm_pre_clip': 6.452968120574951, 'timing/data': 0.02304978301981464, 'timing/model': 1.1353583589661866, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 0.26}) +[2026-06-15 20:14:32,458][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.13384166359901428, 'train/grad_norm_pre_clip': 5.392153739929199, 'timing/data': 0.023550727986730635, 'timing/model': 1.131359977007378, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.26}) +[2026-06-15 20:14:37,008][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.09980939328670502, 'train/grad_norm_pre_clip': 4.730041027069092, 'timing/data': 0.022952789964620024, 'timing/model': 1.294009773992002, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 0.26}) +[2026-06-15 20:14:41,407][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.2431069165468216, 'train/grad_norm_pre_clip': 14.233858108520508, 'timing/data': 0.023182766046375036, 'timing/model': 1.1421767930150963, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 0.26}) +[2026-06-15 20:14:45,956][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.3700857162475586, 'train/grad_norm_pre_clip': 19.203577041625977, 'timing/data': 0.023706367006525397, 'timing/model': 1.132188268005848, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 0.26}) +[2026-06-15 20:14:50,518][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.20209109783172607, 'train/grad_norm_pre_clip': 8.59909725189209, 'timing/data': 0.023658512043766677, 'timing/model': 1.133255279972218, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 0.26}) +[2026-06-15 20:14:55,098][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.27825313806533813, 'train/grad_norm_pre_clip': 13.269083023071289, 'timing/data': 0.02330888604046777, 'timing/model': 1.138987582991831, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 0.26}) +[2026-06-15 20:14:59,681][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.09326540678739548, 'train/grad_norm_pre_clip': 7.565943717956543, 'timing/data': 0.023755162954330444, 'timing/model': 1.1426170600461774, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 0.26}) +[2026-06-15 20:15:04,224][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.1998271346092224, 'train/grad_norm_pre_clip': 9.787569046020508, 'timing/data': 0.02465964894508943, 'timing/model': 1.2906146100140177, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 0.27}) +[2026-06-15 20:15:08,621][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.12274821847677231, 'train/grad_norm_pre_clip': 9.77751350402832, 'timing/data': 0.02385435503674671, 'timing/model': 1.1318290199851617, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 0.27}) +[2026-06-15 20:15:13,162][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.12746776640415192, 'train/grad_norm_pre_clip': 12.008298873901367, 'timing/data': 0.023518852016422898, 'timing/model': 1.1371555609512143, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 0.27}) +[2026-06-15 20:15:17,709][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.230522021651268, 'train/grad_norm_pre_clip': 15.137126922607422, 'timing/data': 0.023138059012126178, 'timing/model': 1.1327005799976178, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 0.27}) +[2026-06-15 20:15:22,250][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.23293665051460266, 'train/grad_norm_pre_clip': 16.186288833618164, 'timing/data': 0.023117651988286525, 'timing/model': 1.1344355510082096, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 0.27}) +[2026-06-15 20:15:26,954][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.41537779569625854, 'train/grad_norm_pre_clip': 15.037751197814941, 'timing/data': 0.024136940017342567, 'timing/model': 1.1352072900044732, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 0.27}) +[2026-06-15 20:15:31,504][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.34574416279792786, 'train/grad_norm_pre_clip': 14.476232528686523, 'timing/data': 0.023118878016248345, 'timing/model': 1.3019807240343653, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 0.27}) +[2026-06-15 20:15:35,920][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.15550419688224792, 'train/grad_norm_pre_clip': 6.674253463745117, 'timing/data': 0.023431801004335284, 'timing/model': 1.1474732930073515, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 0.27}) +[2026-06-15 20:15:40,467][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.18198838829994202, 'train/grad_norm_pre_clip': 7.158205032348633, 'timing/data': 0.024805577006191015, 'timing/model': 1.1350165869807824, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 0.27}) +[2026-06-15 20:15:45,004][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.27580103278160095, 'train/grad_norm_pre_clip': 11.84693431854248, 'timing/data': 0.02351235703099519, 'timing/model': 1.133971133036539, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 0.27}) +[2026-06-15 20:15:49,559][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.13332216441631317, 'train/grad_norm_pre_clip': 4.041162490844727, 'timing/data': 0.024384830030612648, 'timing/model': 1.1330443270271644, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 0.27}) +[2026-06-15 20:15:54,108][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.248625248670578, 'train/grad_norm_pre_clip': 8.726210594177246, 'timing/data': 0.023494789958931506, 'timing/model': 1.1341725130332634, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 0.27}) +[2026-06-15 20:15:58,674][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.26308003067970276, 'train/grad_norm_pre_clip': 11.6216402053833, 'timing/data': 0.02784839499508962, 'timing/model': 1.306783965032082, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 0.27}) +[2026-06-15 20:16:03,068][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.5238587856292725, 'train/grad_norm_pre_clip': 9.775877952575684, 'timing/data': 0.023580235021654516, 'timing/model': 1.1412869560299441, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 0.27}) +[2026-06-15 20:16:07,619][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.3331372141838074, 'train/grad_norm_pre_clip': 12.808563232421875, 'timing/data': 0.024139006971381605, 'timing/model': 1.13218097895151, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.27}) +[2026-06-15 20:16:12,178][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.38240155577659607, 'train/grad_norm_pre_clip': 17.9283504486084, 'timing/data': 0.02263056702213362, 'timing/model': 1.1318374979891814, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 0.27}) +[2026-06-15 20:16:16,753][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.2616044580936432, 'train/grad_norm_pre_clip': 5.285012722015381, 'timing/data': 0.02293853205628693, 'timing/model': 1.1429197089746594, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 0.27}) +[2026-06-15 20:16:21,308][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.14659741520881653, 'train/grad_norm_pre_clip': 5.07282018661499, 'timing/data': 0.02451053896220401, 'timing/model': 1.296470258035697, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 0.27}) +[2026-06-15 20:16:25,845][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.12240700423717499, 'train/grad_norm_pre_clip': 3.698631763458252, 'timing/data': 0.023495595960412174, 'timing/model': 1.273659748956561, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 0.27}) +[2026-06-15 20:16:30,235][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.34317880868911743, 'train/grad_norm_pre_clip': 11.595412254333496, 'timing/data': 0.023726879036985338, 'timing/model': 1.1396350889699534, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 0.27}) +[2026-06-15 20:16:34,782][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.2502032220363617, 'train/grad_norm_pre_clip': 8.335134506225586, 'timing/data': 0.02299915801268071, 'timing/model': 1.136134943983052, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 0.27}) +[2026-06-15 20:16:39,320][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.2193734347820282, 'train/grad_norm_pre_clip': 6.79986572265625, 'timing/data': 0.023379947000648826, 'timing/model': 1.1333847899804823, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 0.27}) +[2026-06-15 20:16:43,848][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.20541928708553314, 'train/grad_norm_pre_clip': 5.295292377471924, 'timing/data': 0.02421255299123004, 'timing/model': 1.1336708050221205, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 0.27}) +[2026-06-15 20:16:48,350][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.20154160261154175, 'train/grad_norm_pre_clip': 6.658435821533203, 'timing/data': 0.02350760600529611, 'timing/model': 1.1326431619818322, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 0.27}) +[2026-06-15 20:16:52,904][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.17505253851413727, 'train/grad_norm_pre_clip': 6.987712860107422, 'timing/data': 0.022364788979757577, 'timing/model': 1.3116491290275007, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 0.28}) +[2026-06-15 20:16:57,306][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.15364715456962585, 'train/grad_norm_pre_clip': 9.743614196777344, 'timing/data': 0.023157729010563344, 'timing/model': 1.1320088290376589, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 0.28}) +[2026-06-15 20:17:01,839][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.4142588973045349, 'train/grad_norm_pre_clip': 23.614892959594727, 'timing/data': 0.023533781000878662, 'timing/model': 1.134570301976055, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 0.28}) +[2026-06-15 20:17:06,369][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.4426782727241516, 'train/grad_norm_pre_clip': 13.300339698791504, 'timing/data': 0.023668818990699947, 'timing/model': 1.134716500993818, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 0.28}) +[2026-06-15 20:17:10,911][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.38756945729255676, 'train/grad_norm_pre_clip': 15.078649520874023, 'timing/data': 0.02396757301175967, 'timing/model': 1.1324788769707084, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 0.28}) +[2026-06-15 20:17:15,462][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.33155837655067444, 'train/grad_norm_pre_clip': 11.567940711975098, 'timing/data': 0.023799384012818336, 'timing/model': 1.129640668979846, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 0.28}) +[2026-06-15 20:17:20,000][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.09452132135629654, 'train/grad_norm_pre_clip': 4.625380039215088, 'timing/data': 0.033849844010546803, 'timing/model': 1.2863096620421857, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 0.28}) +[2026-06-15 20:17:24,379][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.10515782237052917, 'train/grad_norm_pre_clip': 48.747955322265625, 'timing/data': 0.023296747996937484, 'timing/model': 1.132250987982843, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 0.28}) +[2026-06-15 20:17:28,911][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.17529882490634918, 'train/grad_norm_pre_clip': 10.238100051879883, 'timing/data': 0.02355318597983569, 'timing/model': 1.1423857259796932, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 0.28}) +[2026-06-15 20:17:33,459][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.18755429983139038, 'train/grad_norm_pre_clip': 6.8257856369018555, 'timing/data': 0.023533456958830357, 'timing/model': 1.1342155910097063, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 0.28}) +[2026-06-15 20:17:38,002][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.3419116139411926, 'train/grad_norm_pre_clip': 11.605545997619629, 'timing/data': 0.023332029988523573, 'timing/model': 1.1338575039990246, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 0.28}) +[2026-06-15 20:17:42,517][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.34731319546699524, 'train/grad_norm_pre_clip': 12.029866218566895, 'timing/data': 0.024040480027906597, 'timing/model': 1.1371297789737582, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 0.28}) +[2026-06-15 20:17:47,049][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.4656132459640503, 'train/grad_norm_pre_clip': 26.152475357055664, 'timing/data': 0.023857101041357964, 'timing/model': 1.2869528270093724, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 0.28}) +[2026-06-15 20:17:51,430][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.19096891582012177, 'train/grad_norm_pre_clip': 17.069063186645508, 'timing/data': 0.023409485002048314, 'timing/model': 1.138942986028269, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 0.28}) +[2026-06-15 20:17:55,969][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.3524538576602936, 'train/grad_norm_pre_clip': 11.743732452392578, 'timing/data': 0.023813486972358078, 'timing/model': 1.1317427519825287, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 0.28}) +[2026-06-15 20:18:00,503][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.22867347300052643, 'train/grad_norm_pre_clip': 10.175792694091797, 'timing/data': 0.024302611011080444, 'timing/model': 1.131313982012216, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 0.28}) +[2026-06-15 20:18:05,028][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.3676309585571289, 'train/grad_norm_pre_clip': 17.05229377746582, 'timing/data': 0.023491828003898263, 'timing/model': 1.1392191990162246, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 0.28}) +[2026-06-15 20:18:09,576][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.16310197114944458, 'train/grad_norm_pre_clip': 5.829883575439453, 'timing/data': 0.02361957705579698, 'timing/model': 1.13542453100672, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 0.28}) +[2026-06-15 20:18:14,125][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.23676258325576782, 'train/grad_norm_pre_clip': 10.294194221496582, 'timing/data': 0.030201072979252785, 'timing/model': 1.2897788939881139, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.28}) +[2026-06-15 20:18:18,513][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.29742369055747986, 'train/grad_norm_pre_clip': 10.392168998718262, 'timing/data': 0.022878326999489218, 'timing/model': 1.135352680052165, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 0.28}) +[2026-06-15 20:18:23,053][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.19619137048721313, 'train/grad_norm_pre_clip': 8.4671049118042, 'timing/data': 0.023167614999692887, 'timing/model': 1.1296322199632414, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 0.28}) +[2026-06-15 20:18:27,589][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.2677099108695984, 'train/grad_norm_pre_clip': 6.7919745445251465, 'timing/data': 0.022822498984169215, 'timing/model': 1.1347670539980754, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 0.28}) +[2026-06-15 20:18:32,131][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.16938674449920654, 'train/grad_norm_pre_clip': 6.550967693328857, 'timing/data': 0.022596349008381367, 'timing/model': 1.1359570909989998, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 0.28}) +[2026-06-15 20:18:36,660][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.1911429613828659, 'train/grad_norm_pre_clip': 8.44829273223877, 'timing/data': 0.02229414798785001, 'timing/model': 1.2884324899641797, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 0.28}) +[2026-06-15 20:18:41,178][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.3145107328891754, 'train/grad_norm_pre_clip': 11.452352523803711, 'timing/data': 0.023337893013376743, 'timing/model': 1.2739582679932937, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 0.29}) +[2026-06-15 20:18:45,559][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.17019979655742645, 'train/grad_norm_pre_clip': 6.426336288452148, 'timing/data': 0.02320381999015808, 'timing/model': 1.1327118320041336, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.29}) +[2026-06-15 20:18:50,109][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.20620505511760712, 'train/grad_norm_pre_clip': 8.43862533569336, 'timing/data': 0.023320755979511887, 'timing/model': 1.138704084965866, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 0.29}) +[2026-06-15 20:18:54,643][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.19290012121200562, 'train/grad_norm_pre_clip': 12.123306274414062, 'timing/data': 0.023375760996714234, 'timing/model': 1.1343646359746344, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 0.29}) +[2026-06-15 20:18:59,225][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.2559051215648651, 'train/grad_norm_pre_clip': 8.762345314025879, 'timing/data': 0.022634102031588554, 'timing/model': 1.1291945020202547, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 0.29}) +[2026-06-15 20:19:03,764][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.21256525814533234, 'train/grad_norm_pre_clip': 11.83846378326416, 'timing/data': 0.034028851019684225, 'timing/model': 1.1358767360216007, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 0.29}) +[2026-06-15 20:23:23,960][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.06261750310659409, 'train/grad_norm_pre_clip': 2.0014233589172363, 'eval/loss': 0.08913690596818924, 'eval/flap_precision': 0.9240710823909531, 'eval/flap_recall': 0.8827160493827161, 'eval/flap_f1': 0.9029202841357538, 'eval/flap_rate_model': 0.17194444444444446, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9029202841357538, 'timing/data': 0.02359859400894493, 'timing/model': 1.2946434330078773, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 0.29}) +[2026-06-15 20:23:23,960][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_700_state +[2026-06-15 20:23:23,963][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:23:27,532][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_700_state/model.safetensors +[2026-06-15 20:23:37,605][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_700_state/optimizer.bin +[2026-06-15 20:23:37,605][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_700_state/sampler.bin +[2026-06-15 20:23:37,606][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_700_state/sampler_1.bin +[2026-06-15 20:23:37,606][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_700_state/random_states_0.pkl +[2026-06-15 20:23:37,648][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700 +[2026-06-15 20:23:37,648][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:23:37,655][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:23:42,124][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.2102881371974945, 'train/grad_norm_pre_clip': 5.850841045379639, 'timing/data': 0.02610106102656573, 'timing/model': 1.3015814760001376, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 0.29}) +[2026-06-15 20:23:46,511][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.18277451395988464, 'train/grad_norm_pre_clip': 9.196760177612305, 'timing/data': 0.022933518048375845, 'timing/model': 1.2161866020178422, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 0.29}) +[2026-06-15 20:23:50,822][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.12916286289691925, 'train/grad_norm_pre_clip': 5.120945930480957, 'timing/data': 0.022950732964091003, 'timing/model': 1.1168580359662883, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 0.29}) +[2026-06-15 20:23:55,238][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.15245646238327026, 'train/grad_norm_pre_clip': 5.805329322814941, 'timing/data': 0.023201093019451946, 'timing/model': 1.114519814029336, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 0.29}) +[2026-06-15 20:23:59,714][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.17538686096668243, 'train/grad_norm_pre_clip': 60.71383285522461, 'timing/data': 0.023778586997650564, 'timing/model': 1.1177743279840797, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 0.29}) +[2026-06-15 20:24:04,199][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.20555727183818817, 'train/grad_norm_pre_clip': 7.859796524047852, 'timing/data': 0.023259410983882844, 'timing/model': 1.1170454630046152, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 0.29}) +[2026-06-15 20:24:08,715][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.03693133220076561, 'train/grad_norm_pre_clip': 4.831536769866943, 'timing/data': 0.023971871996764094, 'timing/model': 1.2971053249784745, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 0.29}) +[2026-06-15 20:24:13,176][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.12067152559757233, 'train/grad_norm_pre_clip': 8.111080169677734, 'timing/data': 0.022989109973423183, 'timing/model': 1.2378400180023164, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 0.29}) +[2026-06-15 20:24:17,519][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.060360945761203766, 'train/grad_norm_pre_clip': 5.645979881286621, 'timing/data': 0.02302580000832677, 'timing/model': 1.122964726993814, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 0.29}) +[2026-06-15 20:24:22,030][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.258108526468277, 'train/grad_norm_pre_clip': 10.41111946105957, 'timing/data': 0.023342318017967045, 'timing/model': 1.1326314019970596, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 0.29}) +[2026-06-15 20:24:26,564][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.10595256090164185, 'train/grad_norm_pre_clip': 9.980484008789062, 'timing/data': 0.023328879033215344, 'timing/model': 1.122742899984587, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 0.29}) +[2026-06-15 20:24:31,091][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.21822941303253174, 'train/grad_norm_pre_clip': 7.57533597946167, 'timing/data': 0.023592846002429724, 'timing/model': 1.118414290016517, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 0.29}) +[2026-06-15 20:24:35,617][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.04988443851470947, 'train/grad_norm_pre_clip': 24.574127197265625, 'timing/data': 0.023826894001103938, 'timing/model': 1.2977940919809043, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 0.29}) +[2026-06-15 20:24:39,962][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.5325561165809631, 'train/grad_norm_pre_clip': 14.362546920776367, 'timing/data': 0.023852697981055826, 'timing/model': 1.125986853032373, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 0.29}) +[2026-06-15 20:24:44,473][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.17077013850212097, 'train/grad_norm_pre_clip': 13.06637954711914, 'timing/data': 0.0240300860023126, 'timing/model': 1.1267396610346623, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 0.29}) +[2026-06-15 20:24:48,937][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.05482768639922142, 'train/grad_norm_pre_clip': 3.9253346920013428, 'timing/data': 0.024232055991888046, 'timing/model': 1.1262612760183401, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 0.29}) +[2026-06-15 20:24:53,475][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.2304229438304901, 'train/grad_norm_pre_clip': 9.447524070739746, 'timing/data': 0.024009598011616617, 'timing/model': 1.1288769189850427, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 0.29}) +[2026-06-15 20:24:58,006][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.16039882600307465, 'train/grad_norm_pre_clip': 10.102718353271484, 'timing/data': 0.02466406801249832, 'timing/model': 1.125987139006611, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 0.29}) +[2026-06-15 20:25:02,524][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.10056114196777344, 'train/grad_norm_pre_clip': 6.723422050476074, 'timing/data': 0.0285836550174281, 'timing/model': 1.2808320579933934, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 0.3}) +[2026-06-15 20:25:07,006][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.2628863751888275, 'train/grad_norm_pre_clip': 15.706908226013184, 'timing/data': 0.023371581977698952, 'timing/model': 1.2525653449702077, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 0.3}) +[2026-06-15 20:25:11,375][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.608924388885498, 'train/grad_norm_pre_clip': 14.596588134765625, 'timing/data': 0.023181075986940414, 'timing/model': 1.127378596982453, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 0.3}) +[2026-06-15 20:25:15,915][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.2163073867559433, 'train/grad_norm_pre_clip': 8.822239875793457, 'timing/data': 0.023756981012411416, 'timing/model': 1.126437596976757, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 0.3}) +[2026-06-15 20:25:20,455][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.0689464658498764, 'train/grad_norm_pre_clip': 11.363943099975586, 'timing/data': 0.022967324999626726, 'timing/model': 1.1427612120169215, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 0.3}) +[2026-06-15 20:25:24,990][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.09355302900075912, 'train/grad_norm_pre_clip': 11.248388290405273, 'timing/data': 0.02364195801783353, 'timing/model': 1.1328082230174914, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 0.3}) +[2026-06-15 20:25:29,508][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.14377491176128387, 'train/grad_norm_pre_clip': 10.569913864135742, 'timing/data': 0.023358621983788908, 'timing/model': 1.2879119380377233, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 0.3}) +[2026-06-15 20:25:34,023][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.18252424895763397, 'train/grad_norm_pre_clip': 4.855463027954102, 'timing/data': 0.023340978019405156, 'timing/model': 1.2722220499999821, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 0.3}) +[2026-06-15 20:25:38,391][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.3347145915031433, 'train/grad_norm_pre_clip': 9.86421012878418, 'timing/data': 0.023428876942489296, 'timing/model': 1.1297729589859955, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 0.3}) +[2026-06-15 20:25:42,921][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.3401094675064087, 'train/grad_norm_pre_clip': 15.480074882507324, 'timing/data': 0.02403264300664887, 'timing/model': 1.1320319339865819, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 0.3}) +[2026-06-15 20:25:47,453][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.08894234895706177, 'train/grad_norm_pre_clip': 5.891602039337158, 'timing/data': 0.023365632980130613, 'timing/model': 1.1307474479544908, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 0.3}) +[2026-06-15 20:25:51,978][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.1970539391040802, 'train/grad_norm_pre_clip': 9.614312171936035, 'timing/data': 0.023360936029348522, 'timing/model': 1.131554841005709, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.3}) +[2026-06-15 20:25:56,467][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.15834450721740723, 'train/grad_norm_pre_clip': 6.132779598236084, 'timing/data': 0.02276318904478103, 'timing/model': 1.129371731018182, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 0.3}) +[2026-06-15 20:26:01,022][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.16484269499778748, 'train/grad_norm_pre_clip': 11.21013069152832, 'timing/data': 0.023276899999473244, 'timing/model': 1.3022460950305685, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 0.3}) +[2026-06-15 20:26:05,414][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.14798176288604736, 'train/grad_norm_pre_clip': 9.992782592773438, 'timing/data': 0.02280396001879126, 'timing/model': 1.1303705539903603, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 0.3}) +[2026-06-15 20:26:09,944][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.14845451712608337, 'train/grad_norm_pre_clip': 7.892894268035889, 'timing/data': 0.023113906965591013, 'timing/model': 1.1366798449889757, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 0.3}) +[2026-06-15 20:26:14,392][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.3414941132068634, 'train/grad_norm_pre_clip': 7.098298072814941, 'timing/data': 0.023449996020644903, 'timing/model': 1.1334921350353397, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 0.3}) +[2026-06-15 20:26:18,806][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.2339196652173996, 'train/grad_norm_pre_clip': 7.311585903167725, 'timing/data': 0.02377588296076283, 'timing/model': 1.131347926973831, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 0.3}) +[2026-06-15 20:26:23,267][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.4870185852050781, 'train/grad_norm_pre_clip': 22.359521865844727, 'timing/data': 0.023499935981817544, 'timing/model': 1.1324903400382027, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.3}) +[2026-06-15 20:26:27,764][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.09338349848985672, 'train/grad_norm_pre_clip': 5.1137375831604, 'timing/data': 0.02353230200242251, 'timing/model': 1.2559033540310338, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 0.3}) +[2026-06-15 20:26:32,148][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.1304713934659958, 'train/grad_norm_pre_clip': 3.1319186687469482, 'timing/data': 0.023850032011978328, 'timing/model': 1.1308644409873523, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 0.3}) +[2026-06-15 20:26:36,680][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.19814333319664001, 'train/grad_norm_pre_clip': 5.188828468322754, 'timing/data': 0.023227357014548033, 'timing/model': 1.130644921970088, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 0.3}) +[2026-06-15 20:26:41,227][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.06385170668363571, 'train/grad_norm_pre_clip': 3.9217967987060547, 'timing/data': 0.023560677014756948, 'timing/model': 1.137664143985603, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 0.3}) +[2026-06-15 20:26:45,764][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.24896124005317688, 'train/grad_norm_pre_clip': 7.510410785675049, 'timing/data': 0.023538014967925847, 'timing/model': 1.1365432619932108, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 0.3}) +[2026-06-15 20:26:50,308][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.1585702747106552, 'train/grad_norm_pre_clip': 9.241012573242188, 'timing/data': 0.02421505900565535, 'timing/model': 1.2919238670147024, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 0.31}) +[2026-06-15 20:26:54,692][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.1756933331489563, 'train/grad_norm_pre_clip': 7.436269283294678, 'timing/data': 0.02296462800586596, 'timing/model': 1.1370315449894406, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 0.31}) +[2026-06-15 20:26:59,229][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.2558238208293915, 'train/grad_norm_pre_clip': 8.290386199951172, 'timing/data': 0.023382955987472087, 'timing/model': 1.1306135449558496, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 0.31}) +[2026-06-15 20:27:03,694][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.5331538915634155, 'train/grad_norm_pre_clip': 29.202747344970703, 'timing/data': 0.022897474991623312, 'timing/model': 1.1296045579947531, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 0.31}) +[2026-06-15 20:27:08,248][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.1282242089509964, 'train/grad_norm_pre_clip': 8.423574447631836, 'timing/data': 0.02488041896140203, 'timing/model': 1.1294048550189473, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 0.31}) +[2026-06-15 20:27:12,806][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.14444582164287567, 'train/grad_norm_pre_clip': 3.795930862426758, 'timing/data': 0.023535308020655066, 'timing/model': 1.1337127460283227, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 0.31}) +[2026-06-15 20:27:17,327][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.17746031284332275, 'train/grad_norm_pre_clip': 10.95181655883789, 'timing/data': 0.023316966951824725, 'timing/model': 1.278540410043206, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 0.31}) +[2026-06-15 20:27:21,828][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.11822616308927536, 'train/grad_norm_pre_clip': 7.09555721282959, 'timing/data': 0.023656904988456517, 'timing/model': 1.253018177987542, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 0.31}) +[2026-06-15 20:27:26,208][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.1346295326948166, 'train/grad_norm_pre_clip': 12.224374771118164, 'timing/data': 0.023644230037461966, 'timing/model': 1.1340643169824034, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 0.31}) +[2026-06-15 20:27:30,755][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.1101013794541359, 'train/grad_norm_pre_clip': 16.36831283569336, 'timing/data': 0.023428558954037726, 'timing/model': 1.1432573169586249, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 0.31}) +[2026-06-15 20:27:35,283][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.10720467567443848, 'train/grad_norm_pre_clip': 6.155259609222412, 'timing/data': 0.02366163901751861, 'timing/model': 1.132127165037673, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 0.31}) +[2026-06-15 20:27:39,802][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.3504733741283417, 'train/grad_norm_pre_clip': 11.393546104431152, 'timing/data': 0.023954438976943493, 'timing/model': 1.129974739975296, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 0.31}) +[2026-06-15 20:27:44,355][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.15600568056106567, 'train/grad_norm_pre_clip': 16.058401107788086, 'timing/data': 0.0238278970355168, 'timing/model': 1.2916383119882084, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 0.31}) +[2026-06-15 20:27:48,875][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.4052030146121979, 'train/grad_norm_pre_clip': 15.505725860595703, 'timing/data': 0.02316388295730576, 'timing/model': 1.2786259930580854, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 0.31}) +[2026-06-15 20:27:53,243][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.19099366664886475, 'train/grad_norm_pre_clip': 15.463726997375488, 'timing/data': 0.023342339030932635, 'timing/model': 1.132181812019553, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 0.31}) +[2026-06-15 20:27:57,802][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.13498827815055847, 'train/grad_norm_pre_clip': 14.753358840942383, 'timing/data': 0.023708866036031395, 'timing/model': 1.1304007269791327, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 0.31}) +[2026-06-15 20:28:02,357][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.3326962888240814, 'train/grad_norm_pre_clip': 20.82343101501465, 'timing/data': 0.02346634300192818, 'timing/model': 1.1341209019883536, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 0.31}) +[2026-06-15 20:28:06,888][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.05413932353258133, 'train/grad_norm_pre_clip': 4.710882186889648, 'timing/data': 0.024195413978304714, 'timing/model': 1.1299692149623297, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 0.31}) +[2026-06-15 20:28:11,387][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.1232493668794632, 'train/grad_norm_pre_clip': 11.090615272521973, 'timing/data': 0.023103467014152557, 'timing/model': 1.131288675009273, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 0.31}) +[2026-06-15 20:28:15,937][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.16551324725151062, 'train/grad_norm_pre_clip': 7.267627239227295, 'timing/data': 0.022897164046298712, 'timing/model': 1.3031849859980866, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 0.31}) +[2026-06-15 20:28:20,314][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.03447100147604942, 'train/grad_norm_pre_clip': 3.7088630199432373, 'timing/data': 0.02348024502862245, 'timing/model': 1.1323523939936422, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 0.31}) +[2026-06-15 20:28:24,841][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.24636146426200867, 'train/grad_norm_pre_clip': 13.124878883361816, 'timing/data': 0.024029100954066962, 'timing/model': 1.1315380879677832, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 0.31}) +[2026-06-15 20:28:29,291][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.4952549636363983, 'train/grad_norm_pre_clip': 18.435266494750977, 'timing/data': 0.024010627006646246, 'timing/model': 1.1323036099784076, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.31}) +[2026-06-15 20:28:33,721][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.03978844732046127, 'train/grad_norm_pre_clip': 4.929126262664795, 'timing/data': 0.023938382044434547, 'timing/model': 1.1297017260221764, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 0.31}) +[2026-06-15 20:28:38,195][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.08087471127510071, 'train/grad_norm_pre_clip': 6.122481346130371, 'timing/data': 0.02340044901939109, 'timing/model': 1.1309215459623374, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 0.32}) +[2026-06-15 20:28:42,690][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.10568209737539291, 'train/grad_norm_pre_clip': 10.230522155761719, 'timing/data': 0.023884030990302563, 'timing/model': 1.2552260560332797, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 0.32}) +[2026-06-15 20:28:47,068][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.1403372436761856, 'train/grad_norm_pre_clip': 5.374893665313721, 'timing/data': 0.023084186017513275, 'timing/model': 1.1349621340050362, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 0.32}) +[2026-06-15 20:28:51,606][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.28083911538124084, 'train/grad_norm_pre_clip': 42.629451751708984, 'timing/data': 0.023948943009600043, 'timing/model': 1.135969639988616, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 0.32}) +[2026-06-15 20:28:56,140][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.315797358751297, 'train/grad_norm_pre_clip': 17.62967872619629, 'timing/data': 0.02321612398372963, 'timing/model': 1.1325673930114135, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 0.32}) +[2026-06-15 20:29:00,676][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.39034709334373474, 'train/grad_norm_pre_clip': 21.54727554321289, 'timing/data': 0.023878083971794695, 'timing/model': 1.1304638580186293, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 0.32}) +[2026-06-15 20:29:05,209][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.10650889575481415, 'train/grad_norm_pre_clip': 24.938953399658203, 'timing/data': 0.023637454956769943, 'timing/model': 1.2908442330081016, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 0.32}) +[2026-06-15 20:29:09,578][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.07050830870866776, 'train/grad_norm_pre_clip': 9.371416091918945, 'timing/data': 0.023579197993967682, 'timing/model': 1.1295422340044752, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 0.32}) +[2026-06-15 20:29:14,118][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.1188313364982605, 'train/grad_norm_pre_clip': 12.758223533630371, 'timing/data': 0.023360673047136515, 'timing/model': 1.1332581950118765, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 0.32}) +[2026-06-15 20:29:18,616][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.6476057171821594, 'train/grad_norm_pre_clip': 23.06081199645996, 'timing/data': 0.02361178898718208, 'timing/model': 1.137791440996807, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 0.32}) +[2026-06-15 20:29:23,143][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.40961700677871704, 'train/grad_norm_pre_clip': 17.875289916992188, 'timing/data': 0.024205535009969026, 'timing/model': 1.130064401018899, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 0.32}) +[2026-06-15 20:29:27,671][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.3155764937400818, 'train/grad_norm_pre_clip': 17.518651962280273, 'timing/data': 0.02384370402432978, 'timing/model': 1.1315337530104443, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 0.32}) +[2026-06-15 20:29:32,191][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.41342034935951233, 'train/grad_norm_pre_clip': 13.7460298538208, 'timing/data': 0.023888888012152165, 'timing/model': 1.2744358020136133, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 0.32}) +[2026-06-15 20:29:36,694][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.4002939760684967, 'train/grad_norm_pre_clip': 14.785441398620605, 'timing/data': 0.02335216203937307, 'timing/model': 1.2481592170079239, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 0.32}) +[2026-06-15 20:29:41,080][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.2620220482349396, 'train/grad_norm_pre_clip': 20.217588424682617, 'timing/data': 0.02293240203289315, 'timing/model': 1.1352135429624468, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 0.32}) +[2026-06-15 20:29:45,623][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.11918697506189346, 'train/grad_norm_pre_clip': 6.669914245605469, 'timing/data': 0.02342941699316725, 'timing/model': 1.1335825719870627, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 0.32}) +[2026-06-15 20:29:50,156][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.2775620222091675, 'train/grad_norm_pre_clip': 10.163379669189453, 'timing/data': 0.02363872400019318, 'timing/model': 1.132973059953656, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 0.32}) +[2026-06-15 20:29:54,726][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.1694324016571045, 'train/grad_norm_pre_clip': 6.43080997467041, 'timing/data': 0.023367704008705914, 'timing/model': 1.1389882119838148, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 0.32}) +[2026-06-15 20:29:59,296][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.2066737860441208, 'train/grad_norm_pre_clip': 7.626716613769531, 'timing/data': 0.023828756995499134, 'timing/model': 1.2942144349799491, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 0.32}) +[2026-06-15 20:30:03,847][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.28802046179771423, 'train/grad_norm_pre_clip': 12.35588264465332, 'timing/data': 0.023562495014630258, 'timing/model': 1.2958969059982337, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.32}) +[2026-06-15 20:30:08,233][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.17121611535549164, 'train/grad_norm_pre_clip': 5.579517841339111, 'timing/data': 0.023180166957899928, 'timing/model': 1.1304221210302785, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 0.32}) +[2026-06-15 20:30:12,770][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.15309345722198486, 'train/grad_norm_pre_clip': 5.145315170288086, 'timing/data': 0.023265796015039086, 'timing/model': 1.1304141510045156, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 0.32}) +[2026-06-15 20:30:17,301][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.09581887722015381, 'train/grad_norm_pre_clip': 3.1622543334960938, 'timing/data': 0.023355118988547474, 'timing/model': 1.1323302650125697, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 0.32}) +[2026-06-15 20:30:21,843][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.21039506793022156, 'train/grad_norm_pre_clip': 5.0306925773620605, 'timing/data': 0.023554490006063133, 'timing/model': 1.1347760710050352, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 0.32}) +[2026-06-15 20:30:26,349][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.21832773089408875, 'train/grad_norm_pre_clip': 11.295171737670898, 'timing/data': 0.02291734900791198, 'timing/model': 1.1348621009965427, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 0.32}) +[2026-06-15 20:30:30,886][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.12576359510421753, 'train/grad_norm_pre_clip': 7.8986358642578125, 'timing/data': 0.023565891024190933, 'timing/model': 1.2886216430342756, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 0.33}) +[2026-06-15 20:30:35,260][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.3383135199546814, 'train/grad_norm_pre_clip': 21.158418655395508, 'timing/data': 0.02304207201814279, 'timing/model': 1.1322891800082289, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 0.33}) +[2026-06-15 20:30:39,786][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.3231525719165802, 'train/grad_norm_pre_clip': 13.578441619873047, 'timing/data': 0.02325684600509703, 'timing/model': 1.1347714319708757, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 0.33}) +[2026-06-15 20:30:44,209][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.19423086941242218, 'train/grad_norm_pre_clip': 11.884931564331055, 'timing/data': 0.023087047040462494, 'timing/model': 1.1318561579682864, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 0.33}) +[2026-06-15 20:30:48,647][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.28617310523986816, 'train/grad_norm_pre_clip': 17.191143035888672, 'timing/data': 0.023905774985905737, 'timing/model': 1.1326715070172213, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 0.33}) +[2026-06-15 20:30:53,149][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.38371604681015015, 'train/grad_norm_pre_clip': 29.812795639038086, 'timing/data': 0.023369078990072012, 'timing/model': 1.1474097879836336, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 0.33}) +[2026-06-15 20:30:57,678][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.05786944553256035, 'train/grad_norm_pre_clip': 7.586812496185303, 'timing/data': 0.023562956019304693, 'timing/model': 1.2673870280268602, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 0.33}) +[2026-06-15 20:31:02,052][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.3473956882953644, 'train/grad_norm_pre_clip': 3.147840976715088, 'timing/data': 0.022546770982444286, 'timing/model': 1.1339283269480802, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 0.33}) +[2026-06-15 20:35:21,848][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.4935433268547058, 'train/grad_norm_pre_clip': 13.607207298278809, 'eval/loss': 0.1360902041196823, 'eval/flap_precision': 0.8260292164674635, 'eval/flap_recall': 0.9598765432098766, 'eval/flap_f1': 0.887937187723055, 'eval/flap_rate_model': 0.20916666666666667, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.887937187723055, 'timing/data': 0.02290931297466159, 'timing/model': 1.1321921939961612, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.33}) +[2026-06-15 20:35:21,849][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_800_state +[2026-06-15 20:35:21,851][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:35:25,420][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_800_state/model.safetensors +[2026-06-15 20:35:33,012][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_800_state/optimizer.bin +[2026-06-15 20:35:33,012][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_800_state/sampler.bin +[2026-06-15 20:35:33,012][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_800_state/sampler_1.bin +[2026-06-15 20:35:33,013][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_800_state/random_states_0.pkl +[2026-06-15 20:35:33,056][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800 +[2026-06-15 20:35:33,056][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:35:33,062][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:35:37,534][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.03448747098445892, 'train/grad_norm_pre_clip': 3.7778687477111816, 'timing/data': 0.029446075030136853, 'timing/model': 1.1139804519480094, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 0.33}) +[2026-06-15 20:35:41,995][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.07627641409635544, 'train/grad_norm_pre_clip': 6.216226100921631, 'timing/data': 0.02348902396624908, 'timing/model': 1.1111041859840043, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 0.33}) +[2026-06-15 20:35:46,376][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.016894519329071045, 'train/grad_norm_pre_clip': 0.7276615500450134, 'timing/data': 0.0236169439740479, 'timing/model': 1.1211956350016408, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 0.33}) +[2026-06-15 20:35:50,767][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.2849649488925934, 'train/grad_norm_pre_clip': 8.553471565246582, 'timing/data': 0.023178723989985883, 'timing/model': 1.1193599759717472, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 0.33}) +[2026-06-15 20:35:55,252][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.07969272136688232, 'train/grad_norm_pre_clip': 8.93645191192627, 'timing/data': 0.02306552295340225, 'timing/model': 1.2837771740159951, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 0.33}) +[2026-06-15 20:35:59,566][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.06113161891698837, 'train/grad_norm_pre_clip': 4.529302597045898, 'timing/data': 0.023401355952955782, 'timing/model': 1.1177854080451652, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 0.33}) +[2026-06-15 20:36:04,081][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.24849064648151398, 'train/grad_norm_pre_clip': 16.473649978637695, 'timing/data': 0.023052253993228078, 'timing/model': 1.1323166500078514, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 0.33}) +[2026-06-15 20:36:08,583][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.25699713826179504, 'train/grad_norm_pre_clip': 12.468572616577148, 'timing/data': 0.022893415007274598, 'timing/model': 1.1249651570105925, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 0.33}) +[2026-06-15 20:36:13,090][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.3109002709388733, 'train/grad_norm_pre_clip': 5.824127197265625, 'timing/data': 0.023296315979678184, 'timing/model': 1.1239644649904221, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 0.33}) +[2026-06-15 20:36:17,611][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.7930793166160583, 'train/grad_norm_pre_clip': 21.391014099121094, 'timing/data': 0.023013608006294817, 'timing/model': 1.1339550039847381, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 0.33}) +[2026-06-15 20:36:22,126][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.3757050335407257, 'train/grad_norm_pre_clip': 21.051137924194336, 'timing/data': 0.02283776004333049, 'timing/model': 1.29215385398129, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 0.33}) +[2026-06-15 20:36:26,468][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.7949702143669128, 'train/grad_norm_pre_clip': 25.19797134399414, 'timing/data': 0.023335418954957277, 'timing/model': 1.1289932450163178, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 0.33}) +[2026-06-15 20:36:30,963][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.2902906835079193, 'train/grad_norm_pre_clip': 22.78230857849121, 'timing/data': 0.022747149982023984, 'timing/model': 1.1214861489715986, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 0.33}) +[2026-06-15 20:36:35,493][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.210054412484169, 'train/grad_norm_pre_clip': 6.996417999267578, 'timing/data': 0.02284411696018651, 'timing/model': 1.124352473008912, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 0.33}) +[2026-06-15 20:36:40,015][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.3434400260448456, 'train/grad_norm_pre_clip': 9.906390190124512, 'timing/data': 0.023530091042630374, 'timing/model': 1.1321062400238588, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 0.33}) +[2026-06-15 20:36:44,522][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.21215471625328064, 'train/grad_norm_pre_clip': 4.450055122375488, 'timing/data': 0.0232442730339244, 'timing/model': 1.1241697590448894, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 0.34}) +[2026-06-15 20:36:49,056][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.14202873408794403, 'train/grad_norm_pre_clip': 3.3178930282592773, 'timing/data': 0.02305068203713745, 'timing/model': 1.2852598389727063, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 0.34}) +[2026-06-15 20:36:53,421][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.08271966874599457, 'train/grad_norm_pre_clip': 8.05103588104248, 'timing/data': 0.02309504896402359, 'timing/model': 1.1282362339552492, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 0.34}) +[2026-06-15 20:36:57,936][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.35093674063682556, 'train/grad_norm_pre_clip': 24.240089416503906, 'timing/data': 0.022796155011747032, 'timing/model': 1.1271265810355544, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 0.34}) +[2026-06-15 20:37:02,458][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.14926420152187347, 'train/grad_norm_pre_clip': 8.953634262084961, 'timing/data': 0.02256400801707059, 'timing/model': 1.1307918719830923, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 0.34}) +[2026-06-15 20:37:06,967][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.30810317397117615, 'train/grad_norm_pre_clip': 5.623640060424805, 'timing/data': 0.02302653802325949, 'timing/model': 1.1266234859940596, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 0.34}) +[2026-06-15 20:37:11,506][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.1631680727005005, 'train/grad_norm_pre_clip': 4.9868268966674805, 'timing/data': 0.023336720012594014, 'timing/model': 1.136915698996745, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 0.34}) +[2026-06-15 20:37:16,054][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.25423723459243774, 'train/grad_norm_pre_clip': 7.51078462600708, 'timing/data': 0.026532243995461613, 'timing/model': 1.2949997709947638, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 0.34}) +[2026-06-15 20:37:20,422][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.14839032292366028, 'train/grad_norm_pre_clip': 4.632774829864502, 'timing/data': 0.030658885021694005, 'timing/model': 1.1301765000098385, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 0.34}) +[2026-06-15 20:37:24,961][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.07187794893980026, 'train/grad_norm_pre_clip': 1.831532597541809, 'timing/data': 0.02323517296463251, 'timing/model': 1.1290834320243448, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 0.34}) +[2026-06-15 20:37:29,480][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.13948595523834229, 'train/grad_norm_pre_clip': 5.15097713470459, 'timing/data': 0.02300013997592032, 'timing/model': 1.1285875100293197, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 0.34}) +[2026-06-15 20:37:33,996][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.07578379660844803, 'train/grad_norm_pre_clip': 4.477922439575195, 'timing/data': 0.022822909988462925, 'timing/model': 1.1308200489729643, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 0.34}) +[2026-06-15 20:37:38,543][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.24042262136936188, 'train/grad_norm_pre_clip': 11.314935684204102, 'timing/data': 0.022481666004750878, 'timing/model': 1.2931805559783243, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 0.34}) +[2026-06-15 20:37:43,058][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.2579582631587982, 'train/grad_norm_pre_clip': 4.110434055328369, 'timing/data': 0.02336365997325629, 'timing/model': 1.2702870120410807, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 0.34}) +[2026-06-15 20:37:47,449][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.1069721132516861, 'train/grad_norm_pre_clip': 9.696378707885742, 'timing/data': 0.023216558969579637, 'timing/model': 1.130994061997626, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 0.34}) +[2026-06-15 20:37:52,001][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.08791347593069077, 'train/grad_norm_pre_clip': 5.189974308013916, 'timing/data': 0.023643115011509508, 'timing/model': 1.1332345300470479, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 0.34}) +[2026-06-15 20:37:56,562][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.0384129136800766, 'train/grad_norm_pre_clip': 18.90607452392578, 'timing/data': 0.023095109965652227, 'timing/model': 1.138927738997154, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 0.34}) +[2026-06-15 20:38:01,108][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.07479424774646759, 'train/grad_norm_pre_clip': 8.503090858459473, 'timing/data': 0.023605365015100688, 'timing/model': 1.1400587979587726, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 0.34}) +[2026-06-15 20:38:05,588][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.11016402393579483, 'train/grad_norm_pre_clip': 10.23271656036377, 'timing/data': 0.02362749195890501, 'timing/model': 1.1312782810418867, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 0.34}) +[2026-06-15 20:38:10,183][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.15701304376125336, 'train/grad_norm_pre_clip': 4.355259895324707, 'timing/data': 0.023268683988135308, 'timing/model': 1.297125260985922, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.34}) +[2026-06-15 20:38:14,562][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.5140463709831238, 'train/grad_norm_pre_clip': 15.29175090789795, 'timing/data': 0.03036079596495256, 'timing/model': 1.1283768030116335, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 0.34}) +[2026-06-15 20:38:19,098][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.15806283056735992, 'train/grad_norm_pre_clip': 9.165342330932617, 'timing/data': 0.023647290014196187, 'timing/model': 1.132713757979218, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 0.34}) +[2026-06-15 20:38:23,635][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.03873017057776451, 'train/grad_norm_pre_clip': 3.540992498397827, 'timing/data': 0.02312311698915437, 'timing/model': 1.132541568018496, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 0.34}) +[2026-06-15 20:38:28,145][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.2941502034664154, 'train/grad_norm_pre_clip': 11.481022834777832, 'timing/data': 0.024127760960254818, 'timing/model': 1.1299792529898696, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 0.34}) +[2026-06-15 20:38:32,698][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.6023861169815063, 'train/grad_norm_pre_clip': 16.21183204650879, 'timing/data': 0.023572826001327485, 'timing/model': 1.1461372239864431, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.35}) +[2026-06-15 20:38:37,242][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.34945014119148254, 'train/grad_norm_pre_clip': 16.98558235168457, 'timing/data': 0.0233368210028857, 'timing/model': 1.2951499609625898, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 0.35}) +[2026-06-15 20:38:41,622][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.062175869941711426, 'train/grad_norm_pre_clip': 11.73523235321045, 'timing/data': 0.023297881009057164, 'timing/model': 1.1322619640268385, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 0.35}) +[2026-06-15 20:38:46,166][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.3287721276283264, 'train/grad_norm_pre_clip': 8.096680641174316, 'timing/data': 0.022903133009094745, 'timing/model': 1.1301788889686577, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 0.35}) +[2026-06-15 20:38:50,701][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.29932156205177307, 'train/grad_norm_pre_clip': 46.25124740600586, 'timing/data': 0.022320433985441923, 'timing/model': 1.1293815650278702, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 0.35}) +[2026-06-15 20:38:55,242][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.39816156029701233, 'train/grad_norm_pre_clip': 15.750386238098145, 'timing/data': 0.02201047498965636, 'timing/model': 1.1368126919842325, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 0.35}) +[2026-06-15 20:38:59,752][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.16348162293434143, 'train/grad_norm_pre_clip': 5.275509834289551, 'timing/data': 0.02274291403591633, 'timing/model': 1.129225177981425, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 0.35}) +[2026-06-15 20:39:04,288][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.0665494054555893, 'train/grad_norm_pre_clip': 5.438105583190918, 'timing/data': 0.023266879958100617, 'timing/model': 1.2937527970061637, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 0.35}) +[2026-06-15 20:39:08,672][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.1999436467885971, 'train/grad_norm_pre_clip': 8.23209285736084, 'timing/data': 0.023272231977898628, 'timing/model': 1.1303780039888807, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 0.35}) +[2026-06-15 20:39:13,215][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.16804653406143188, 'train/grad_norm_pre_clip': 9.61487865447998, 'timing/data': 0.022887182014528662, 'timing/model': 1.1327369050122797, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 0.35}) +[2026-06-15 20:39:17,751][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.24756799638271332, 'train/grad_norm_pre_clip': 16.17180824279785, 'timing/data': 0.023095365962944925, 'timing/model': 1.1286424660356715, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.35}) +[2026-06-15 20:39:22,256][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.06032118573784828, 'train/grad_norm_pre_clip': 3.4764645099639893, 'timing/data': 0.022976769017986953, 'timing/model': 1.1306670380290598, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 0.35}) +[2026-06-15 20:39:26,801][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.041895925998687744, 'train/grad_norm_pre_clip': 3.346088171005249, 'timing/data': 0.022695910010952502, 'timing/model': 1.139927821001038, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 0.35}) +[2026-06-15 20:39:31,356][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.1668565571308136, 'train/grad_norm_pre_clip': 6.2144269943237305, 'timing/data': 0.023005969007499516, 'timing/model': 1.298508221982047, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 0.35}) +[2026-06-15 20:39:35,752][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.20728150010108948, 'train/grad_norm_pre_clip': 6.8244547843933105, 'timing/data': 0.023806901997886598, 'timing/model': 1.1319225049810484, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 0.35}) +[2026-06-15 20:39:40,301][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.09051782637834549, 'train/grad_norm_pre_clip': 5.299773693084717, 'timing/data': 0.023233184008859098, 'timing/model': 1.1329823239939287, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 0.35}) +[2026-06-15 20:39:44,841][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.06558119505643845, 'train/grad_norm_pre_clip': 4.49444055557251, 'timing/data': 0.023192825028672814, 'timing/model': 1.1289746779948473, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 0.35}) +[2026-06-15 20:39:49,406][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.10071282088756561, 'train/grad_norm_pre_clip': 5.451405048370361, 'timing/data': 0.023152610985562205, 'timing/model': 1.137326768017374, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 0.35}) +[2026-06-15 20:39:53,964][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.23611123859882355, 'train/grad_norm_pre_clip': 7.9416399002075195, 'timing/data': 0.022744592977687716, 'timing/model': 1.2899222419946454, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 0.35}) +[2026-06-15 20:39:58,481][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.1270069181919098, 'train/grad_norm_pre_clip': 15.778624534606934, 'timing/data': 0.0220681470236741, 'timing/model': 1.2712179810041562, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 0.35}) +[2026-06-15 20:40:02,891][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.20035026967525482, 'train/grad_norm_pre_clip': 11.36477279663086, 'timing/data': 0.025315193983260542, 'timing/model': 1.1417001860099845, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.35}) +[2026-06-15 20:40:07,469][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.13007521629333496, 'train/grad_norm_pre_clip': 2.22652006149292, 'timing/data': 0.022845972969662398, 'timing/model': 1.147015628986992, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 0.35}) +[2026-06-15 20:40:12,028][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.15142038464546204, 'train/grad_norm_pre_clip': 15.113441467285156, 'timing/data': 0.024028334009926766, 'timing/model': 1.134716780972667, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 0.35}) +[2026-06-15 20:40:16,564][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.030255991965532303, 'train/grad_norm_pre_clip': 1.676538348197937, 'timing/data': 0.02321881998796016, 'timing/model': 1.1371611589565873, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 0.35}) +[2026-06-15 20:40:21,083][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.2531029284000397, 'train/grad_norm_pre_clip': 7.527444839477539, 'timing/data': 0.023475766007322818, 'timing/model': 1.1428429529769346, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 0.35}) +[2026-06-15 20:40:25,621][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.3403704762458801, 'train/grad_norm_pre_clip': 19.064395904541016, 'timing/data': 0.02301659103250131, 'timing/model': 1.2986687539960258, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 0.36}) +[2026-06-15 20:40:30,000][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.16481617093086243, 'train/grad_norm_pre_clip': 9.378750801086426, 'timing/data': 0.0232467669993639, 'timing/model': 1.1331321930047125, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 0.36}) +[2026-06-15 20:40:34,520][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.0941293016076088, 'train/grad_norm_pre_clip': 11.911388397216797, 'timing/data': 0.022758296981919557, 'timing/model': 1.1319175219978206, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 0.36}) +[2026-06-15 20:40:39,066][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.5194629430770874, 'train/grad_norm_pre_clip': 19.860841751098633, 'timing/data': 0.023627291026059538, 'timing/model': 1.133431012975052, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 0.36}) +[2026-06-15 20:40:43,602][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.09463897347450256, 'train/grad_norm_pre_clip': 10.45389461517334, 'timing/data': 0.023323882021941245, 'timing/model': 1.1311445910250768, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 0.36}) +[2026-06-15 20:40:48,161][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.13734622299671173, 'train/grad_norm_pre_clip': 8.664192199707031, 'timing/data': 0.02855860802810639, 'timing/model': 1.134730543999467, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.36}) +[2026-06-15 20:40:52,720][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.09606009721755981, 'train/grad_norm_pre_clip': 8.793309211730957, 'timing/data': 0.027509587991517037, 'timing/model': 1.2932245470001362, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 0.36}) +[2026-06-15 20:40:57,139][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.06395550072193146, 'train/grad_norm_pre_clip': 8.144909858703613, 'timing/data': 0.022144694987218827, 'timing/model': 1.1441350260283798, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 0.36}) +[2026-06-15 20:41:01,692][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.04540421441197395, 'train/grad_norm_pre_clip': 3.1535279750823975, 'timing/data': 0.02282415801892057, 'timing/model': 1.1389692349475808, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 0.36}) +[2026-06-15 20:41:06,265][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.04673067480325699, 'train/grad_norm_pre_clip': 3.03745698928833, 'timing/data': 0.023427958018146455, 'timing/model': 1.1419244619901292, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 0.36}) +[2026-06-15 20:41:10,843][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.13475818932056427, 'train/grad_norm_pre_clip': 24.121341705322266, 'timing/data': 0.022836384014226496, 'timing/model': 1.144943967985455, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 0.36}) +[2026-06-15 20:41:15,372][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.09351750463247299, 'train/grad_norm_pre_clip': 6.910368919372559, 'timing/data': 0.025951858959160745, 'timing/model': 1.1325482910033315, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 0.36}) +[2026-06-15 20:41:19,900][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.07220020890235901, 'train/grad_norm_pre_clip': 5.521430969238281, 'timing/data': 0.023391978989820927, 'timing/model': 1.2815683490480296, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 0.36}) +[2026-06-15 20:41:24,296][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.3737339973449707, 'train/grad_norm_pre_clip': 8.581307411193848, 'timing/data': 0.023363158979918808, 'timing/model': 1.1366308990400285, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 0.36}) +[2026-06-15 20:41:28,837][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.3344177305698395, 'train/grad_norm_pre_clip': 11.138164520263672, 'timing/data': 0.023280894965864718, 'timing/model': 1.130447256960906, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 0.36}) +[2026-06-15 20:41:33,371][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.1865268051624298, 'train/grad_norm_pre_clip': 8.44948959350586, 'timing/data': 0.023186122009065002, 'timing/model': 1.1299666980048642, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 0.36}) +[2026-06-15 20:41:37,882][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.1412227600812912, 'train/grad_norm_pre_clip': 12.831883430480957, 'timing/data': 0.023993832000996917, 'timing/model': 1.1290777430403978, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 0.36}) +[2026-06-15 20:41:42,414][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.13599877059459686, 'train/grad_norm_pre_clip': 9.652103424072266, 'timing/data': 0.02339743496850133, 'timing/model': 1.132599914970342, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 0.36}) +[2026-06-15 20:41:46,980][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.3124752640724182, 'train/grad_norm_pre_clip': 12.17481517791748, 'timing/data': 0.023317445011343807, 'timing/model': 1.3025153769995086, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 0.36}) +[2026-06-15 20:41:51,374][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.4064837694168091, 'train/grad_norm_pre_clip': 10.005651473999023, 'timing/data': 0.023466209997422993, 'timing/model': 1.1354326659929939, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.36}) +[2026-06-15 20:41:55,920][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.12113426625728607, 'train/grad_norm_pre_clip': 10.385998725891113, 'timing/data': 0.02346311998553574, 'timing/model': 1.1420904669794254, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 0.36}) +[2026-06-15 20:42:00,492][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.3803344964981079, 'train/grad_norm_pre_clip': 12.40652084350586, 'timing/data': 0.023131952970288694, 'timing/model': 1.13542258401867, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 0.36}) +[2026-06-15 20:42:05,056][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.02757839299738407, 'train/grad_norm_pre_clip': 1.7173367738723755, 'timing/data': 0.023474565998185426, 'timing/model': 1.1341790439910255, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 0.36}) +[2026-06-15 20:42:09,613][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.10217055678367615, 'train/grad_norm_pre_clip': 5.4998650550842285, 'timing/data': 0.023551777994725853, 'timing/model': 1.2983759240014479, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 0.36}) +[2026-06-15 20:42:14,166][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.3891910910606384, 'train/grad_norm_pre_clip': 17.64668846130371, 'timing/data': 0.028514760022517294, 'timing/model': 1.2989622040186077, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 0.37}) +[2026-06-15 20:42:18,564][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.3714638948440552, 'train/grad_norm_pre_clip': 10.7844820022583, 'timing/data': 0.023163890989962965, 'timing/model': 1.1342872260138392, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 0.37}) +[2026-06-15 20:42:23,134][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.07933894544839859, 'train/grad_norm_pre_clip': 5.116021633148193, 'timing/data': 0.022929253987967968, 'timing/model': 1.1375915789976716, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 0.37}) +[2026-06-15 20:42:27,690][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.1876610815525055, 'train/grad_norm_pre_clip': 7.639973163604736, 'timing/data': 0.026611556997522712, 'timing/model': 1.1337945659761317, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 0.37}) +[2026-06-15 20:42:32,226][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.16655246913433075, 'train/grad_norm_pre_clip': 8.864760398864746, 'timing/data': 0.022729693970177323, 'timing/model': 1.1329445040319115, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 0.37}) +[2026-06-15 20:42:36,724][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.2829604148864746, 'train/grad_norm_pre_clip': 10.951457023620605, 'timing/data': 0.023064925975631922, 'timing/model': 1.1319484899868257, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 0.37}) +[2026-06-15 20:42:41,285][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.21393240988254547, 'train/grad_norm_pre_clip': 8.465706825256348, 'timing/data': 0.023366779962088913, 'timing/model': 1.307840263005346, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 0.37}) +[2026-06-15 20:42:45,672][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.30092835426330566, 'train/grad_norm_pre_clip': 12.509293556213379, 'timing/data': 0.02296047902200371, 'timing/model': 1.1374277340364642, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 0.37}) +[2026-06-15 20:42:50,232][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.29184821248054504, 'train/grad_norm_pre_clip': 16.048954010009766, 'timing/data': 0.023126765969209373, 'timing/model': 1.1375364029663615, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 0.37}) +[2026-06-15 20:42:54,787][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.3526850640773773, 'train/grad_norm_pre_clip': 18.619287490844727, 'timing/data': 0.023558660002890974, 'timing/model': 1.1354151839623228, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 0.37}) +[2026-06-15 20:42:59,315][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.3102247714996338, 'train/grad_norm_pre_clip': 7.9915995597839355, 'timing/data': 0.023358275997452438, 'timing/model': 1.132240177015774, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 0.37}) +[2026-06-15 20:47:19,286][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.10549899190664291, 'train/grad_norm_pre_clip': 6.675410270690918, 'eval/loss': 0.10228010267019272, 'eval/flap_precision': 0.8299866131191432, 'eval/flap_recall': 0.9567901234567902, 'eval/flap_f1': 0.888888888888889, 'eval/flap_rate_model': 0.2075, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.8888888888888888, 'timing/data': 0.023014690028503537, 'timing/model': 1.1315822620526887, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 0.37}) +[2026-06-15 20:47:19,287][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_900_state +[2026-06-15 20:47:19,289][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:47:22,859][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_900_state/model.safetensors +[2026-06-15 20:47:32,863][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_900_state/optimizer.bin +[2026-06-15 20:47:32,864][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_900_state/sampler.bin +[2026-06-15 20:47:32,864][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_900_state/sampler_1.bin +[2026-06-15 20:47:32,864][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_900_state/random_states_0.pkl +[2026-06-15 20:47:32,911][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-06-15 20:47:32,911][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:47:32,917][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:47:37,373][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.04996507242321968, 'train/grad_norm_pre_clip': 3.672516107559204, 'timing/data': 0.02627742098411545, 'timing/model': 1.124280916992575, 'learning_rate/qwen_vl_interface': 9.387268552285024e-06, 'learning_rate/action_model': 9.387268552285025e-05, 'epoch': 0.37}) +[2026-06-15 20:47:41,830][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.14726310968399048, 'train/grad_norm_pre_clip': 13.10815143585205, 'timing/data': 0.025429548986721784, 'timing/model': 1.2763552279793657, 'learning_rate/qwen_vl_interface': 9.385771559309125e-06, 'learning_rate/action_model': 9.385771559309125e-05, 'epoch': 0.37}) +[2026-06-15 20:47:46,267][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.36761337518692017, 'train/grad_norm_pre_clip': 14.5993013381958, 'timing/data': 0.02318829100113362, 'timing/model': 1.2387084960355423, 'learning_rate/qwen_vl_interface': 9.38427286627376e-06, 'learning_rate/action_model': 9.38427286627376e-05, 'epoch': 0.37}) +[2026-06-15 20:47:50,574][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.3381451368331909, 'train/grad_norm_pre_clip': 5.9001312255859375, 'timing/data': 0.023271129990462214, 'timing/model': 1.1161139300093055, 'learning_rate/qwen_vl_interface': 9.382772473794985e-06, 'learning_rate/action_model': 9.382772473794985e-05, 'epoch': 0.37}) +[2026-06-15 20:47:55,049][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.2803317606449127, 'train/grad_norm_pre_clip': 21.596046447753906, 'timing/data': 0.023130607034545392, 'timing/model': 1.1194696000311524, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.37}) +[2026-06-15 20:47:59,530][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.10308240354061127, 'train/grad_norm_pre_clip': 3.556997060775757, 'timing/data': 0.023945965978782624, 'timing/model': 1.117246211040765, 'learning_rate/qwen_vl_interface': 9.37976659297492e-06, 'learning_rate/action_model': 9.379766592974921e-05, 'epoch': 0.37}) +[2026-06-15 20:48:04,143][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.389060378074646, 'train/grad_norm_pre_clip': 10.959769248962402, 'timing/data': 0.024129752011504024, 'timing/model': 1.2635800479911268, 'learning_rate/qwen_vl_interface': 9.378261105869237e-06, 'learning_rate/action_model': 9.378261105869237e-05, 'epoch': 0.37}) +[2026-06-15 20:48:08,624][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.2993418276309967, 'train/grad_norm_pre_clip': 9.446873664855957, 'timing/data': 0.023815011023543775, 'timing/model': 1.1214665289735422, 'learning_rate/qwen_vl_interface': 9.376753921791348e-06, 'learning_rate/action_model': 9.376753921791348e-05, 'epoch': 0.37}) +[2026-06-15 20:48:13,123][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.1987558752298355, 'train/grad_norm_pre_clip': 16.988567352294922, 'timing/data': 0.023122488986700773, 'timing/model': 1.2886793880024925, 'learning_rate/qwen_vl_interface': 9.375245041360805e-06, 'learning_rate/action_model': 9.375245041360804e-05, 'epoch': 0.37}) +[2026-06-15 20:48:17,491][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.2836341857910156, 'train/grad_norm_pre_clip': 10.51132583618164, 'timing/data': 0.023367876012343913, 'timing/model': 1.128484663029667, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 0.37}) +[2026-06-15 20:48:22,010][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.32380712032318115, 'train/grad_norm_pre_clip': 10.590421676635742, 'timing/data': 0.02312795602483675, 'timing/model': 1.1266711800126359, 'learning_rate/qwen_vl_interface': 9.372222193923421e-06, 'learning_rate/action_model': 9.372222193923421e-05, 'epoch': 0.37}) +[2026-06-15 20:48:26,525][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.26042863726615906, 'train/grad_norm_pre_clip': 9.536185264587402, 'timing/data': 0.023281722038518637, 'timing/model': 1.1333355420501903, 'learning_rate/qwen_vl_interface': 9.370708228159158e-06, 'learning_rate/action_model': 9.370708228159159e-05, 'epoch': 0.37}) +[2026-06-15 20:48:31,041][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.09781604260206223, 'train/grad_norm_pre_clip': 5.401226997375488, 'timing/data': 0.023511729028541595, 'timing/model': 1.1258422440150753, 'learning_rate/qwen_vl_interface': 9.369192568527398e-06, 'learning_rate/action_model': 9.369192568527398e-05, 'epoch': 0.38}) +[2026-06-15 20:48:35,559][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.16476842761039734, 'train/grad_norm_pre_clip': 12.938018798828125, 'timing/data': 0.02374058699933812, 'timing/model': 1.2857615990214981, 'learning_rate/qwen_vl_interface': 9.367675215651168e-06, 'learning_rate/action_model': 9.367675215651168e-05, 'epoch': 0.38}) +[2026-06-15 20:48:39,920][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.15681850910186768, 'train/grad_norm_pre_clip': 7.316561222076416, 'timing/data': 0.023927978007122874, 'timing/model': 1.1303308139904402, 'learning_rate/qwen_vl_interface': 9.366156170154194e-06, 'learning_rate/action_model': 9.366156170154194e-05, 'epoch': 0.38}) +[2026-06-15 20:48:44,436][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.22847557067871094, 'train/grad_norm_pre_clip': 9.255178451538086, 'timing/data': 0.023880826018285006, 'timing/model': 1.1250213580206037, 'learning_rate/qwen_vl_interface': 9.3646354326609e-06, 'learning_rate/action_model': 9.364635432660899e-05, 'epoch': 0.38}) +[2026-06-15 20:48:48,955][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.2004469335079193, 'train/grad_norm_pre_clip': 8.936070442199707, 'timing/data': 0.023507211997639388, 'timing/model': 1.1296794979716651, 'learning_rate/qwen_vl_interface': 9.3631130037964e-06, 'learning_rate/action_model': 9.363113003796401e-05, 'epoch': 0.38}) +[2026-06-15 20:48:53,467][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.12375576049089432, 'train/grad_norm_pre_clip': 7.413007736206055, 'timing/data': 0.023497752961702645, 'timing/model': 1.1295448470045812, 'learning_rate/qwen_vl_interface': 9.361588884186513e-06, 'learning_rate/action_model': 9.361588884186513e-05, 'epoch': 0.38}) +[2026-06-15 20:48:57,994][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.10093578696250916, 'train/grad_norm_pre_clip': 4.863018035888672, 'timing/data': 0.02322892501251772, 'timing/model': 1.1300861510098912, 'learning_rate/qwen_vl_interface': 9.360063074457744e-06, 'learning_rate/action_model': 9.360063074457743e-05, 'epoch': 0.38}) +[2026-06-15 20:49:02,544][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.23582643270492554, 'train/grad_norm_pre_clip': 12.410244941711426, 'timing/data': 0.023102636041585356, 'timing/model': 1.3044271689723246, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 0.38}) +[2026-06-15 20:49:06,898][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.09614988416433334, 'train/grad_norm_pre_clip': 3.1350972652435303, 'timing/data': 0.02299294399563223, 'timing/model': 1.127763670985587, 'learning_rate/qwen_vl_interface': 9.357006387153065e-06, 'learning_rate/action_model': 9.357006387153065e-05, 'epoch': 0.38}) +[2026-06-15 20:49:11,410][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.10333296656608582, 'train/grad_norm_pre_clip': 5.755840301513672, 'timing/data': 0.02344383898889646, 'timing/model': 1.1307890510070138, 'learning_rate/qwen_vl_interface': 9.355475510833643e-06, 'learning_rate/action_model': 9.355475510833643e-05, 'epoch': 0.38}) +[2026-06-15 20:49:15,937][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.15884919464588165, 'train/grad_norm_pre_clip': 4.8656086921691895, 'timing/data': 0.02326014603022486, 'timing/model': 1.1277592310216278, 'learning_rate/qwen_vl_interface': 9.353942946908318e-06, 'learning_rate/action_model': 9.353942946908317e-05, 'epoch': 0.38}) +[2026-06-15 20:49:20,447][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.1038493886590004, 'train/grad_norm_pre_clip': 7.726832866668701, 'timing/data': 0.024165646987967193, 'timing/model': 1.131634117977228, 'learning_rate/qwen_vl_interface': 9.352408696007065e-06, 'learning_rate/action_model': 9.352408696007065e-05, 'epoch': 0.38}) +[2026-06-15 20:49:24,984][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.032380420714616776, 'train/grad_norm_pre_clip': 3.3647663593292236, 'timing/data': 0.023402790015097708, 'timing/model': 1.1336447250214405, 'learning_rate/qwen_vl_interface': 9.350872758760558e-06, 'learning_rate/action_model': 9.350872758760559e-05, 'epoch': 0.38}) +[2026-06-15 20:49:29,518][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.12717236578464508, 'train/grad_norm_pre_clip': 6.772150039672852, 'timing/data': 0.0227719449903816, 'timing/model': 1.2820329520036466, 'learning_rate/qwen_vl_interface': 9.349335135800164e-06, 'learning_rate/action_model': 9.349335135800163e-05, 'epoch': 0.38}) +[2026-06-15 20:49:33,879][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.13623535633087158, 'train/grad_norm_pre_clip': 7.302170753479004, 'timing/data': 0.0227781769935973, 'timing/model': 1.1327710159821436, 'learning_rate/qwen_vl_interface': 9.347795827757942e-06, 'learning_rate/action_model': 9.347795827757941e-05, 'epoch': 0.38}) +[2026-06-15 20:49:38,417][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.23105721175670624, 'train/grad_norm_pre_clip': 10.80405044555664, 'timing/data': 0.023035719990730286, 'timing/model': 1.1339131190325134, 'learning_rate/qwen_vl_interface': 9.346254835266643e-06, 'learning_rate/action_model': 9.346254835266642e-05, 'epoch': 0.38}) +[2026-06-15 20:49:42,858][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.12130370736122131, 'train/grad_norm_pre_clip': 7.821047306060791, 'timing/data': 0.024200326995924115, 'timing/model': 1.1326235230080783, 'learning_rate/qwen_vl_interface': 9.344712158959709e-06, 'learning_rate/action_model': 9.344712158959709e-05, 'epoch': 0.38}) +[2026-06-15 20:49:47,376][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.28384727239608765, 'train/grad_norm_pre_clip': 5.926215648651123, 'timing/data': 0.023309297976084054, 'timing/model': 1.129149247019086, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 0.38}) +[2026-06-15 20:49:51,903][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.1460091471672058, 'train/grad_norm_pre_clip': 2.9559762477874756, 'timing/data': 0.023513432999607176, 'timing/model': 1.1333039589808322, 'learning_rate/qwen_vl_interface': 9.341621757436177e-06, 'learning_rate/action_model': 9.341621757436177e-05, 'epoch': 0.38}) +[2026-06-15 20:49:56,457][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.11566244065761566, 'train/grad_norm_pre_clip': 11.880134582519531, 'timing/data': 0.02293752902187407, 'timing/model': 1.305736661015544, 'learning_rate/qwen_vl_interface': 9.340074033489924e-06, 'learning_rate/action_model': 9.340074033489923e-05, 'epoch': 0.38}) +[2026-06-15 20:50:00,835][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.13556358218193054, 'train/grad_norm_pre_clip': 7.55517578125, 'timing/data': 0.023716196010354906, 'timing/model': 1.1361921470379457, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.38}) +[2026-06-15 20:50:05,375][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.11839362233877182, 'train/grad_norm_pre_clip': 6.0086894035339355, 'timing/data': 0.023081557999830693, 'timing/model': 1.1300762590253726, 'learning_rate/qwen_vl_interface': 9.336973542409505e-06, 'learning_rate/action_model': 9.336973542409504e-05, 'epoch': 0.38}) +[2026-06-15 20:50:09,901][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.28716138005256653, 'train/grad_norm_pre_clip': 19.79208755493164, 'timing/data': 0.023707651009317487, 'timing/model': 1.1298695940058678, 'learning_rate/qwen_vl_interface': 9.33542077654983e-06, 'learning_rate/action_model': 9.33542077654983e-05, 'epoch': 0.38}) +[2026-06-15 20:50:14,385][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.022822296246886253, 'train/grad_norm_pre_clip': 1.196489930152893, 'timing/data': 0.023013023019302636, 'timing/model': 1.1314479790162295, 'learning_rate/qwen_vl_interface': 9.333866331327995e-06, 'learning_rate/action_model': 9.333866331327995e-05, 'epoch': 0.38}) +[2026-06-15 20:50:18,894][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.0852544829249382, 'train/grad_norm_pre_clip': 14.54229736328125, 'timing/data': 0.02305242500733584, 'timing/model': 1.127977388969157, 'learning_rate/qwen_vl_interface': 9.332310207382972e-06, 'learning_rate/action_model': 9.332310207382972e-05, 'epoch': 0.38}) +[2026-06-15 20:50:23,432][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.04842780902981758, 'train/grad_norm_pre_clip': 3.7030787467956543, 'timing/data': 0.023227552010212094, 'timing/model': 1.287097177002579, 'learning_rate/qwen_vl_interface': 9.330752405354426e-06, 'learning_rate/action_model': 9.330752405354426e-05, 'epoch': 0.39}) +[2026-06-15 20:50:27,801][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.29113295674324036, 'train/grad_norm_pre_clip': 18.719051361083984, 'timing/data': 0.022907275997567922, 'timing/model': 1.134797106962651, 'learning_rate/qwen_vl_interface': 9.32919292588271e-06, 'learning_rate/action_model': 9.32919292588271e-05, 'epoch': 0.39}) +[2026-06-15 20:50:32,352][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.36513766646385193, 'train/grad_norm_pre_clip': 12.491172790527344, 'timing/data': 0.023255295993294567, 'timing/model': 1.1289825410349295, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.39}) +[2026-06-15 20:50:36,875][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.1699124574661255, 'train/grad_norm_pre_clip': 11.161955833435059, 'timing/data': 0.023160012962762266, 'timing/model': 1.1314324869890697, 'learning_rate/qwen_vl_interface': 9.326068937174627e-06, 'learning_rate/action_model': 9.326068937174626e-05, 'epoch': 0.39}) +[2026-06-15 20:50:41,397][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.22163233160972595, 'train/grad_norm_pre_clip': 9.02861499786377, 'timing/data': 0.02323804801562801, 'timing/model': 1.1299310069880448, 'learning_rate/qwen_vl_interface': 9.324504429222412e-06, 'learning_rate/action_model': 9.324504429222412e-05, 'epoch': 0.39}) +[2026-06-15 20:50:45,930][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.2214852273464203, 'train/grad_norm_pre_clip': 8.843880653381348, 'timing/data': 0.023329593997914344, 'timing/model': 1.289563488040585, 'learning_rate/qwen_vl_interface': 9.322938246395336e-06, 'learning_rate/action_model': 9.322938246395337e-05, 'epoch': 0.39}) +[2026-06-15 20:50:50,307][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.09285136312246323, 'train/grad_norm_pre_clip': 17.612071990966797, 'timing/data': 0.02326233801431954, 'timing/model': 1.134402906987816, 'learning_rate/qwen_vl_interface': 9.321370389337196e-06, 'learning_rate/action_model': 9.321370389337196e-05, 'epoch': 0.39}) +[2026-06-15 20:50:54,883][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.21842887997627258, 'train/grad_norm_pre_clip': 5.268860340118408, 'timing/data': 0.023605868977028877, 'timing/model': 1.1326343860127963, 'learning_rate/qwen_vl_interface': 9.319800858692478e-06, 'learning_rate/action_model': 9.319800858692476e-05, 'epoch': 0.39}) +[2026-06-15 20:50:59,420][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.11130319535732269, 'train/grad_norm_pre_clip': 14.302483558654785, 'timing/data': 0.02358999295393005, 'timing/model': 1.1346450669807382, 'learning_rate/qwen_vl_interface': 9.318229655106357e-06, 'learning_rate/action_model': 9.318229655106356e-05, 'epoch': 0.39}) +[2026-06-15 20:51:03,961][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.22976095974445343, 'train/grad_norm_pre_clip': 8.772072792053223, 'timing/data': 0.023454536974895746, 'timing/model': 1.1353804029640742, 'learning_rate/qwen_vl_interface': 9.316656779224693e-06, 'learning_rate/action_model': 9.316656779224693e-05, 'epoch': 0.39}) +[2026-06-15 20:51:08,508][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.027747921645641327, 'train/grad_norm_pre_clip': 1.4825159311294556, 'timing/data': 0.02315344801172614, 'timing/model': 1.1294787289807573, 'learning_rate/qwen_vl_interface': 9.31508223169404e-06, 'learning_rate/action_model': 9.31508223169404e-05, 'epoch': 0.39}) +[2026-06-15 20:51:13,047][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.1825069636106491, 'train/grad_norm_pre_clip': 12.597753524780273, 'timing/data': 0.02314422302879393, 'timing/model': 1.2880459630396217, 'learning_rate/qwen_vl_interface': 9.313506013161633e-06, 'learning_rate/action_model': 9.313506013161633e-05, 'epoch': 0.39}) +[2026-06-15 20:51:17,422][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.05375334620475769, 'train/grad_norm_pre_clip': 3.4534807205200195, 'timing/data': 0.022942568990401924, 'timing/model': 1.1309570270241238, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 0.39}) +[2026-06-15 20:51:21,960][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.15368270874023438, 'train/grad_norm_pre_clip': 10.250722885131836, 'timing/data': 0.02329768700292334, 'timing/model': 1.132074568013195, 'learning_rate/qwen_vl_interface': 9.310348565683938e-06, 'learning_rate/action_model': 9.310348565683939e-05, 'epoch': 0.39}) +[2026-06-15 20:51:26,525][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.1480378359556198, 'train/grad_norm_pre_clip': 15.14183235168457, 'timing/data': 0.023042136046569794, 'timing/model': 1.1437457019928843, 'learning_rate/qwen_vl_interface': 9.30876733803656e-06, 'learning_rate/action_model': 9.308767338036559e-05, 'epoch': 0.39}) +[2026-06-15 20:51:31,043][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.041447293013334274, 'train/grad_norm_pre_clip': 17.96993637084961, 'timing/data': 0.023484828998334706, 'timing/model': 1.1399842699756846, 'learning_rate/qwen_vl_interface': 9.30718444198324e-06, 'learning_rate/action_model': 9.307184441983239e-05, 'epoch': 0.39}) +[2026-06-15 20:51:35,598][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.027524380013346672, 'train/grad_norm_pre_clip': 3.195744037628174, 'timing/data': 0.023283245973289013, 'timing/model': 1.1371700349845923, 'learning_rate/qwen_vl_interface': 9.30559987817465e-06, 'learning_rate/action_model': 9.30559987817465e-05, 'epoch': 0.39}) +[2026-06-15 20:51:40,136][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.09998134523630142, 'train/grad_norm_pre_clip': 3.0988898277282715, 'timing/data': 0.023100229038391262, 'timing/model': 1.293291951995343, 'learning_rate/qwen_vl_interface': 9.304013647262143e-06, 'learning_rate/action_model': 9.304013647262143e-05, 'epoch': 0.39}) +[2026-06-15 20:51:44,517][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.5607224702835083, 'train/grad_norm_pre_clip': 16.592605590820312, 'timing/data': 0.022802870953455567, 'timing/model': 1.1327759240521118, 'learning_rate/qwen_vl_interface': 9.302425749897761e-06, 'learning_rate/action_model': 9.302425749897761e-05, 'epoch': 0.39}) +[2026-06-15 20:51:49,051][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.06219103932380676, 'train/grad_norm_pre_clip': 8.574492454528809, 'timing/data': 0.023537859960924834, 'timing/model': 1.1332521290169097, 'learning_rate/qwen_vl_interface': 9.300836186734225e-06, 'learning_rate/action_model': 9.300836186734226e-05, 'epoch': 0.39}) +[2026-06-15 20:51:53,495][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.07730040699243546, 'train/grad_norm_pre_clip': 15.958714485168457, 'timing/data': 0.023281442990992218, 'timing/model': 1.13189555198187, 'learning_rate/qwen_vl_interface': 9.299244958424947e-06, 'learning_rate/action_model': 9.299244958424946e-05, 'epoch': 0.39}) +[2026-06-15 20:51:58,047][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.09778057783842087, 'train/grad_norm_pre_clip': 15.684896469116211, 'timing/data': 0.023175619018729776, 'timing/model': 1.133966055989731, 'learning_rate/qwen_vl_interface': 9.29765206562402e-06, 'learning_rate/action_model': 9.297652065624018e-05, 'epoch': 0.39}) +[2026-06-15 20:52:02,578][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.22405268251895905, 'train/grad_norm_pre_clip': 13.488495826721191, 'timing/data': 0.02374982397304848, 'timing/model': 1.1313341280329041, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 0.39}) +[2026-06-15 20:52:07,114][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.2450525015592575, 'train/grad_norm_pre_clip': 22.351558685302734, 'timing/data': 0.024009670014493167, 'timing/model': 1.2852444720338099, 'learning_rate/qwen_vl_interface': 9.294461289167013e-06, 'learning_rate/action_model': 9.294461289167014e-05, 'epoch': 0.39}) +[2026-06-15 20:52:11,492][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.15699836611747742, 'train/grad_norm_pre_clip': 12.110894203186035, 'timing/data': 0.023360633000265807, 'timing/model': 1.1347421359969303, 'learning_rate/qwen_vl_interface': 9.292863406822543e-06, 'learning_rate/action_model': 9.292863406822542e-05, 'epoch': 0.4}) +[2026-06-15 20:52:16,032][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.09612716734409332, 'train/grad_norm_pre_clip': 12.675200462341309, 'timing/data': 0.024620603013318032, 'timing/model': 1.1330834430409595, 'learning_rate/qwen_vl_interface': 9.291263862609638e-06, 'learning_rate/action_model': 9.291263862609638e-05, 'epoch': 0.4}) +[2026-06-15 20:52:20,594][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.04186427965760231, 'train/grad_norm_pre_clip': 3.6727118492126465, 'timing/data': 0.023616249964106828, 'timing/model': 1.1496366629726253, 'learning_rate/qwen_vl_interface': 9.28966265718581e-06, 'learning_rate/action_model': 9.28966265718581e-05, 'epoch': 0.4}) +[2026-06-15 20:52:25,102][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.22550754249095917, 'train/grad_norm_pre_clip': 28.11931610107422, 'timing/data': 0.02352858299855143, 'timing/model': 1.13059593597427, 'learning_rate/qwen_vl_interface': 9.288059791209256e-06, 'learning_rate/action_model': 9.288059791209256e-05, 'epoch': 0.4}) +[2026-06-15 20:52:29,641][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.3342733383178711, 'train/grad_norm_pre_clip': 12.313202857971191, 'timing/data': 0.023389972979202867, 'timing/model': 1.1500416260096245, 'learning_rate/qwen_vl_interface': 9.286455265338853e-06, 'learning_rate/action_model': 9.286455265338853e-05, 'epoch': 0.4}) +[2026-06-15 20:52:34,186][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.34532636404037476, 'train/grad_norm_pre_clip': 22.29522705078125, 'timing/data': 0.02363842399790883, 'timing/model': 1.29540216102032, 'learning_rate/qwen_vl_interface': 9.284849080234161e-06, 'learning_rate/action_model': 9.28484908023416e-05, 'epoch': 0.4}) +[2026-06-15 20:52:38,565][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.0833389088511467, 'train/grad_norm_pre_clip': 3.3487038612365723, 'timing/data': 0.02369412302505225, 'timing/model': 1.1349915320170112, 'learning_rate/qwen_vl_interface': 9.283241236555421e-06, 'learning_rate/action_model': 9.283241236555421e-05, 'epoch': 0.4}) +[2026-06-15 20:52:43,118][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.013853214681148529, 'train/grad_norm_pre_clip': 1.6065186262130737, 'timing/data': 0.023730862012598664, 'timing/model': 1.1403680910007097, 'learning_rate/qwen_vl_interface': 9.281631734963557e-06, 'learning_rate/action_model': 9.281631734963557e-05, 'epoch': 0.4}) +[2026-06-15 20:52:47,658][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.15048228204250336, 'train/grad_norm_pre_clip': 10.433972358703613, 'timing/data': 0.023523478012066334, 'timing/model': 1.1340774800046347, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 0.4}) +[2026-06-15 20:52:52,201][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.06310801208019257, 'train/grad_norm_pre_clip': 9.3731050491333, 'timing/data': 0.023618365987204015, 'timing/model': 1.138335408992134, 'learning_rate/qwen_vl_interface': 9.278407760687562e-06, 'learning_rate/action_model': 9.278407760687562e-05, 'epoch': 0.4}) +[2026-06-15 20:52:56,755][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.1801910400390625, 'train/grad_norm_pre_clip': 9.470032691955566, 'timing/data': 0.02439269796013832, 'timing/model': 1.287919873953797, 'learning_rate/qwen_vl_interface': 9.276793289328686e-06, 'learning_rate/action_model': 9.276793289328685e-05, 'epoch': 0.4}) +[2026-06-15 20:53:01,258][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.31333962082862854, 'train/grad_norm_pre_clip': 35.49094009399414, 'timing/data': 0.0244144779862836, 'timing/model': 1.253528407018166, 'learning_rate/qwen_vl_interface': 9.27517716270719e-06, 'learning_rate/action_model': 9.275177162707191e-05, 'epoch': 0.4}) +[2026-06-15 20:53:05,641][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.37634655833244324, 'train/grad_norm_pre_clip': 7.939722537994385, 'timing/data': 0.024017805000767112, 'timing/model': 1.1356535580125637, 'learning_rate/qwen_vl_interface': 9.273559381487409e-06, 'learning_rate/action_model': 9.273559381487409e-05, 'epoch': 0.4}) +[2026-06-15 20:53:10,186][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.13932371139526367, 'train/grad_norm_pre_clip': 7.487597465515137, 'timing/data': 0.02353003399912268, 'timing/model': 1.1345917870057747, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.4}) +[2026-06-15 20:53:14,758][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.3231866657733917, 'train/grad_norm_pre_clip': 16.654563903808594, 'timing/data': 0.023270778998266906, 'timing/model': 1.1319146010209806, 'learning_rate/qwen_vl_interface': 9.2703188579137e-06, 'learning_rate/action_model': 9.2703188579137e-05, 'epoch': 0.4}) +[2026-06-15 20:53:19,359][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.23693408071994781, 'train/grad_norm_pre_clip': 7.638220310211182, 'timing/data': 0.022889771033078432, 'timing/model': 1.1382524030050263, 'learning_rate/qwen_vl_interface': 9.26869611689183e-06, 'learning_rate/action_model': 9.26869611689183e-05, 'epoch': 0.4}) +[2026-06-15 20:53:23,926][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.22774595022201538, 'train/grad_norm_pre_clip': 7.985596656799316, 'timing/data': 0.03156742104329169, 'timing/model': 1.297078656964004, 'learning_rate/qwen_vl_interface': 9.267071723935784e-06, 'learning_rate/action_model': 9.267071723935784e-05, 'epoch': 0.4}) +[2026-06-15 20:53:28,327][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.04739388823509216, 'train/grad_norm_pre_clip': 2.865077257156372, 'timing/data': 0.02280361601151526, 'timing/model': 1.1399120609858073, 'learning_rate/qwen_vl_interface': 9.26544567971329e-06, 'learning_rate/action_model': 9.26544567971329e-05, 'epoch': 0.4}) +[2026-06-15 20:53:32,888][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.15983106195926666, 'train/grad_norm_pre_clip': 11.060954093933105, 'timing/data': 0.023173970985226333, 'timing/model': 1.1363132820115425, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 0.4}) +[2026-06-15 20:53:37,442][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.14831559360027313, 'train/grad_norm_pre_clip': 9.231770515441895, 'timing/data': 0.023429324966855347, 'timing/model': 1.1404782839817926, 'learning_rate/qwen_vl_interface': 9.262188640143262e-06, 'learning_rate/action_model': 9.262188640143262e-05, 'epoch': 0.4}) +[2026-06-15 20:53:41,983][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.17627492547035217, 'train/grad_norm_pre_clip': 8.149561882019043, 'timing/data': 0.02318314102012664, 'timing/model': 1.1385688549489714, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.4}) +[2026-06-15 20:53:46,538][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.25117790699005127, 'train/grad_norm_pre_clip': 12.805398941040039, 'timing/data': 0.0234721060260199, 'timing/model': 1.1406577290035784, 'learning_rate/qwen_vl_interface': 9.258925003537127e-06, 'learning_rate/action_model': 9.258925003537126e-05, 'epoch': 0.4}) +[2026-06-15 20:53:51,108][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.13527415692806244, 'train/grad_norm_pre_clip': 7.294391632080078, 'timing/data': 0.024382450035773218, 'timing/model': 1.3032750499551184, 'learning_rate/qwen_vl_interface': 9.257290713022041e-06, 'learning_rate/action_model': 9.257290713022041e-05, 'epoch': 0.4}) +[2026-06-15 20:53:55,505][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.07968972623348236, 'train/grad_norm_pre_clip': 5.616488933563232, 'timing/data': 0.0234156459919177, 'timing/model': 1.1385006420314312, 'learning_rate/qwen_vl_interface': 9.255654775261112e-06, 'learning_rate/action_model': 9.255654775261112e-05, 'epoch': 0.4}) +[2026-06-15 20:54:00,073][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.04936908558011055, 'train/grad_norm_pre_clip': 5.435062408447266, 'timing/data': 0.02349056303501129, 'timing/model': 1.1352725310134701, 'learning_rate/qwen_vl_interface': 9.254017190926813e-06, 'learning_rate/action_model': 9.254017190926813e-05, 'epoch': 0.41}) +[2026-06-15 20:54:04,543][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.08467661589384079, 'train/grad_norm_pre_clip': 9.071295738220215, 'timing/data': 0.02456711302511394, 'timing/model': 1.1477133629960008, 'learning_rate/qwen_vl_interface': 9.25237796069229e-06, 'learning_rate/action_model': 9.25237796069229e-05, 'epoch': 0.41}) +[2026-06-15 20:54:09,088][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.15190957486629486, 'train/grad_norm_pre_clip': 6.733367443084717, 'timing/data': 0.024162288988009095, 'timing/model': 1.132346639991738, 'learning_rate/qwen_vl_interface': 9.250737085231371e-06, 'learning_rate/action_model': 9.250737085231372e-05, 'epoch': 0.41}) +[2026-06-15 20:54:13,639][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.023986855521798134, 'train/grad_norm_pre_clip': 1.5698487758636475, 'timing/data': 0.024607097962871194, 'timing/model': 1.1343462630175054, 'learning_rate/qwen_vl_interface': 9.24909456521856e-06, 'learning_rate/action_model': 9.249094565218559e-05, 'epoch': 0.41}) +[2026-06-15 20:54:18,192][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.03692423924803734, 'train/grad_norm_pre_clip': 3.9691569805145264, 'timing/data': 0.024072909960523248, 'timing/model': 1.300174864998553, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 0.41}) +[2026-06-15 20:54:22,575][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.15362665057182312, 'train/grad_norm_pre_clip': 5.857100963592529, 'timing/data': 0.023990682035218924, 'timing/model': 1.1356791089638136, 'learning_rate/qwen_vl_interface': 9.245804594238637e-06, 'learning_rate/action_model': 9.245804594238636e-05, 'epoch': 0.41}) +[2026-06-15 20:54:27,129][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.12298751622438431, 'train/grad_norm_pre_clip': 12.663437843322754, 'timing/data': 0.023783285985700786, 'timing/model': 1.1365657550049946, 'learning_rate/qwen_vl_interface': 9.244157144623909e-06, 'learning_rate/action_model': 9.244157144623909e-05, 'epoch': 0.41}) +[2026-06-15 20:54:31,685][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.0968133881688118, 'train/grad_norm_pre_clip': 15.980574607849121, 'timing/data': 0.02314806601498276, 'timing/model': 1.1342451919917949, 'learning_rate/qwen_vl_interface': 9.242508053162051e-06, 'learning_rate/action_model': 9.242508053162051e-05, 'epoch': 0.41}) +[2026-06-15 20:54:36,214][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.11907999217510223, 'train/grad_norm_pre_clip': 2.628225803375244, 'timing/data': 0.02439414703985676, 'timing/model': 1.1457347469986416, 'learning_rate/qwen_vl_interface': 9.240857320530944e-06, 'learning_rate/action_model': 9.240857320530943e-05, 'epoch': 0.41}) +[2026-06-15 20:54:40,749][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.1795220971107483, 'train/grad_norm_pre_clip': 17.60093116760254, 'timing/data': 0.023552183993160725, 'timing/model': 1.1346352909458801, 'learning_rate/qwen_vl_interface': 9.23920494740914e-06, 'learning_rate/action_model': 9.23920494740914e-05, 'epoch': 0.41}) +[2026-06-15 20:54:45,295][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.4768578112125397, 'train/grad_norm_pre_clip': 15.964421272277832, 'timing/data': 0.024052868015132844, 'timing/model': 1.2958635269897059, 'learning_rate/qwen_vl_interface': 9.237550934475868e-06, 'learning_rate/action_model': 9.237550934475867e-05, 'epoch': 0.41}) +[2026-06-15 20:54:49,695][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.1601613610982895, 'train/grad_norm_pre_clip': 11.845760345458984, 'timing/data': 0.023298938001971692, 'timing/model': 1.1367503060027957, 'learning_rate/qwen_vl_interface': 9.235895282411029e-06, 'learning_rate/action_model': 9.235895282411029e-05, 'epoch': 0.41}) +[2026-06-15 20:54:54,243][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.2553001046180725, 'train/grad_norm_pre_clip': 8.552035331726074, 'timing/data': 0.023768027022015303, 'timing/model': 1.1372038880363107, 'learning_rate/qwen_vl_interface': 9.2342379918952e-06, 'learning_rate/action_model': 9.2342379918952e-05, 'epoch': 0.41}) +[2026-06-15 20:54:58,786][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.18994902074337006, 'train/grad_norm_pre_clip': 13.139753341674805, 'timing/data': 0.023594553000293672, 'timing/model': 1.1387106869951822, 'learning_rate/qwen_vl_interface': 9.232579063609631e-06, 'learning_rate/action_model': 9.232579063609631e-05, 'epoch': 0.41}) +[2026-06-15 20:59:20,283][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.3444620966911316, 'train/grad_norm_pre_clip': 16.845317840576172, 'eval/loss': 0.08215475082397461, 'eval/flap_precision': 0.9269406392694064, 'eval/flap_recall': 0.9398148148148148, 'eval/flap_f1': 0.9333333333333332, 'eval/flap_rate_model': 0.1825, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9333333333333331, 'timing/data': 0.024143917951732874, 'timing/model': 1.1360346130095422, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 0.41}) +[2026-06-15 20:59:20,283][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1000_state +[2026-06-15 20:59:20,286][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:59:23,848][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1000_state/model.safetensors +[2026-06-15 20:59:33,855][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1000_state/optimizer.bin +[2026-06-15 20:59:33,855][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1000_state/sampler.bin +[2026-06-15 20:59:33,856][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-15 20:59:33,856][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-15 20:59:33,901][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-15 20:59:33,902][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:59:33,908][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:59:38,253][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.3319348692893982, 'train/grad_norm_pre_clip': 6.876135349273682, 'timing/data': 0.023692003975156695, 'timing/model': 1.114377613994293, 'learning_rate/qwen_vl_interface': 9.229256296457635e-06, 'learning_rate/action_model': 9.229256296457634e-05, 'epoch': 0.41}) +[2026-06-15 20:59:42,713][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.3068894147872925, 'train/grad_norm_pre_clip': 11.985509872436523, 'timing/data': 0.023090569011401385, 'timing/model': 1.1084425479639322, 'learning_rate/qwen_vl_interface': 9.22759245895707e-06, 'learning_rate/action_model': 9.227592458957071e-05, 'epoch': 0.41}) +[2026-06-15 20:59:47,165][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.32268378138542175, 'train/grad_norm_pre_clip': 15.594164848327637, 'timing/data': 0.02392292598960921, 'timing/model': 1.2734618610120378, 'learning_rate/qwen_vl_interface': 9.225926986418494e-06, 'learning_rate/action_model': 9.225926986418494e-05, 'epoch': 0.41}) +[2026-06-15 20:59:51,486][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.2149544209241867, 'train/grad_norm_pre_clip': 10.05159854888916, 'timing/data': 0.023805752978660166, 'timing/model': 1.118772424000781, 'learning_rate/qwen_vl_interface': 9.224259879526519e-06, 'learning_rate/action_model': 9.224259879526519e-05, 'epoch': 0.41}) +[2026-06-15 20:59:55,976][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.1280180662870407, 'train/grad_norm_pre_clip': 10.679234504699707, 'timing/data': 0.023486052989028394, 'timing/model': 1.1177346900221892, 'learning_rate/qwen_vl_interface': 9.222591138966427e-06, 'learning_rate/action_model': 9.222591138966426e-05, 'epoch': 0.41}) +[2026-06-15 21:00:00,461][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.02739577926695347, 'train/grad_norm_pre_clip': 1.6351152658462524, 'timing/data': 0.02416277094744146, 'timing/model': 1.1349433189607225, 'learning_rate/qwen_vl_interface': 9.220920765424175e-06, 'learning_rate/action_model': 9.220920765424174e-05, 'epoch': 0.41}) +[2026-06-15 21:00:04,935][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.08217184245586395, 'train/grad_norm_pre_clip': 5.027292251586914, 'timing/data': 0.024364837037865072, 'timing/model': 1.1153435650048777, 'learning_rate/qwen_vl_interface': 9.219248759586392e-06, 'learning_rate/action_model': 9.219248759586391e-05, 'epoch': 0.41}) +[2026-06-15 21:00:09,396][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.24398168921470642, 'train/grad_norm_pre_clip': 11.517448425292969, 'timing/data': 0.02472655102610588, 'timing/model': 1.1147932690219022, 'learning_rate/qwen_vl_interface': 9.217575122140375e-06, 'learning_rate/action_model': 9.217575122140375e-05, 'epoch': 0.41}) +[2026-06-15 21:00:13,886][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.23069652915000916, 'train/grad_norm_pre_clip': 5.435591220855713, 'timing/data': 0.023873461002949625, 'timing/model': 1.2841137319919653, 'learning_rate/qwen_vl_interface': 9.215899853774094e-06, 'learning_rate/action_model': 9.215899853774093e-05, 'epoch': 0.41}) +[2026-06-15 21:00:18,217][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.2442532479763031, 'train/grad_norm_pre_clip': 19.005922317504883, 'timing/data': 0.023620018968358636, 'timing/model': 1.1204326639999636, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.41}) +[2026-06-15 21:00:22,703][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.47946712374687195, 'train/grad_norm_pre_clip': 32.20649337768555, 'timing/data': 0.023985162959434092, 'timing/model': 1.120050268014893, 'learning_rate/qwen_vl_interface': 9.212544427035967e-06, 'learning_rate/action_model': 9.212544427035967e-05, 'epoch': 0.42}) +[2026-06-15 21:00:27,179][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.13127928972244263, 'train/grad_norm_pre_clip': 10.872918128967285, 'timing/data': 0.023472693981602788, 'timing/model': 1.11916157201631, 'learning_rate/qwen_vl_interface': 9.21086427004341e-06, 'learning_rate/action_model': 9.210864270043411e-05, 'epoch': 0.42}) +[2026-06-15 21:00:31,658][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.27383139729499817, 'train/grad_norm_pre_clip': 10.138418197631836, 'timing/data': 0.023666347027756274, 'timing/model': 1.1258205029880628, 'learning_rate/qwen_vl_interface': 9.209182484889168e-06, 'learning_rate/action_model': 9.209182484889167e-05, 'epoch': 0.42}) +[2026-06-15 21:00:36,170][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.23340310156345367, 'train/grad_norm_pre_clip': 21.80659294128418, 'timing/data': 0.022886894992552698, 'timing/model': 1.1254993220209144, 'learning_rate/qwen_vl_interface': 9.207499072264557e-06, 'learning_rate/action_model': 9.207499072264558e-05, 'epoch': 0.42}) +[2026-06-15 21:00:40,677][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.16878727078437805, 'train/grad_norm_pre_clip': 11.992079734802246, 'timing/data': 0.02365254395408556, 'timing/model': 1.2936686970060691, 'learning_rate/qwen_vl_interface': 9.205814032861566e-06, 'learning_rate/action_model': 9.205814032861566e-05, 'epoch': 0.42}) +[2026-06-15 21:00:45,022][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.2096254676580429, 'train/grad_norm_pre_clip': 7.680517673492432, 'timing/data': 0.02446487097768113, 'timing/model': 1.1270057180081494, 'learning_rate/qwen_vl_interface': 9.204127367372849e-06, 'learning_rate/action_model': 9.204127367372849e-05, 'epoch': 0.42}) +[2026-06-15 21:00:49,527][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.04561952129006386, 'train/grad_norm_pre_clip': 4.111833095550537, 'timing/data': 0.023811958031728864, 'timing/model': 1.1316248580114916, 'learning_rate/qwen_vl_interface': 9.202439076491732e-06, 'learning_rate/action_model': 9.202439076491732e-05, 'epoch': 0.42}) +[2026-06-15 21:00:54,047][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.06424236297607422, 'train/grad_norm_pre_clip': 9.122859954833984, 'timing/data': 0.023733808018732816, 'timing/model': 1.1328434940078296, 'learning_rate/qwen_vl_interface': 9.200749160912208e-06, 'learning_rate/action_model': 9.200749160912208e-05, 'epoch': 0.42}) +[2026-06-15 21:00:58,550][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.346187561750412, 'train/grad_norm_pre_clip': 15.447369575500488, 'timing/data': 0.02332313498482108, 'timing/model': 1.1308590829721652, 'learning_rate/qwen_vl_interface': 9.199057621328937e-06, 'learning_rate/action_model': 9.199057621328935e-05, 'epoch': 0.42}) +[2026-06-15 21:01:03,058][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.3156714141368866, 'train/grad_norm_pre_clip': 7.000646114349365, 'timing/data': 0.023932960990350693, 'timing/model': 1.270681458991021, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 0.42}) +[2026-06-15 21:01:07,577][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.17001794278621674, 'train/grad_norm_pre_clip': 16.887447357177734, 'timing/data': 0.024218614969868213, 'timing/model': 1.2857690600212663, 'learning_rate/qwen_vl_interface': 9.19566967293313e-06, 'learning_rate/action_model': 9.19566967293313e-05, 'epoch': 0.42}) +[2026-06-15 21:01:11,928][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.13683384656906128, 'train/grad_norm_pre_clip': 16.973316192626953, 'timing/data': 0.029702137981075794, 'timing/model': 1.1243202119949274, 'learning_rate/qwen_vl_interface': 9.193973265513253e-06, 'learning_rate/action_model': 9.193973265513253e-05, 'epoch': 0.42}) +[2026-06-15 21:01:16,434][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.09074877947568893, 'train/grad_norm_pre_clip': 5.478693962097168, 'timing/data': 0.023950957984197885, 'timing/model': 1.1264954200014472, 'learning_rate/qwen_vl_interface': 9.192275236874946e-06, 'learning_rate/action_model': 9.192275236874946e-05, 'epoch': 0.42}) +[2026-06-15 21:01:20,976][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.3423662483692169, 'train/grad_norm_pre_clip': 11.309760093688965, 'timing/data': 0.02349759399658069, 'timing/model': 1.1326648960239254, 'learning_rate/qwen_vl_interface': 9.190575587716199e-06, 'learning_rate/action_model': 9.190575587716199e-05, 'epoch': 0.42}) +[2026-06-15 21:01:25,536][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.29396727681159973, 'train/grad_norm_pre_clip': 9.762389183044434, 'timing/data': 0.029389341012574732, 'timing/model': 1.1354576160083525, 'learning_rate/qwen_vl_interface': 9.188874318735677e-06, 'learning_rate/action_model': 9.188874318735676e-05, 'epoch': 0.42}) +[2026-06-15 21:01:30,019][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.09464361518621445, 'train/grad_norm_pre_clip': 5.028741836547852, 'timing/data': 0.02364081802079454, 'timing/model': 1.1347871180041693, 'learning_rate/qwen_vl_interface': 9.187171430632707e-06, 'learning_rate/action_model': 9.187171430632707e-05, 'epoch': 0.42}) +[2026-06-15 21:01:34,552][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.13861973583698273, 'train/grad_norm_pre_clip': 27.34366798400879, 'timing/data': 0.022928787977434695, 'timing/model': 1.2829366330406629, 'learning_rate/qwen_vl_interface': 9.185466924107283e-06, 'learning_rate/action_model': 9.185466924107282e-05, 'epoch': 0.42}) +[2026-06-15 21:01:38,915][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.06301449239253998, 'train/grad_norm_pre_clip': 4.565736293792725, 'timing/data': 0.022518218960613012, 'timing/model': 1.1277129179798067, 'learning_rate/qwen_vl_interface': 9.18376079986006e-06, 'learning_rate/action_model': 9.183760799860061e-05, 'epoch': 0.42}) +[2026-06-15 21:01:43,438][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.17244118452072144, 'train/grad_norm_pre_clip': 6.891911029815674, 'timing/data': 0.02358971198555082, 'timing/model': 1.1284352660295554, 'learning_rate/qwen_vl_interface': 9.182053058592365e-06, 'learning_rate/action_model': 9.182053058592365e-05, 'epoch': 0.42}) +[2026-06-15 21:01:47,856][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.07151448726654053, 'train/grad_norm_pre_clip': 5.438051700592041, 'timing/data': 0.02349628898082301, 'timing/model': 1.128741730004549, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 0.42}) +[2026-06-15 21:01:52,314][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.17766380310058594, 'train/grad_norm_pre_clip': 6.705193996429443, 'timing/data': 0.023844520968850702, 'timing/model': 1.126966871030163, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.42}) +[2026-06-15 21:01:56,776][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.09494587779045105, 'train/grad_norm_pre_clip': 7.051276206970215, 'timing/data': 0.024336094968020916, 'timing/model': 1.126085015013814, 'learning_rate/qwen_vl_interface': 9.176920139689638e-06, 'learning_rate/action_model': 9.176920139689637e-05, 'epoch': 0.42}) +[2026-06-15 21:02:01,289][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.08899915963411331, 'train/grad_norm_pre_clip': 9.180891036987305, 'timing/data': 0.023218373011332005, 'timing/model': 1.2544991690083407, 'learning_rate/qwen_vl_interface': 9.175205937366568e-06, 'learning_rate/action_model': 9.175205937366569e-05, 'epoch': 0.42}) +[2026-06-15 21:02:05,665][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.22190779447555542, 'train/grad_norm_pre_clip': 7.103418350219727, 'timing/data': 0.023641887004487216, 'timing/model': 1.1331795370206237, 'learning_rate/qwen_vl_interface': 9.173490121539608e-06, 'learning_rate/action_model': 9.173490121539608e-05, 'epoch': 0.42}) +[2026-06-15 21:02:10,184][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.21487697958946228, 'train/grad_norm_pre_clip': 12.675762176513672, 'timing/data': 0.023448524007108063, 'timing/model': 1.1333143700030632, 'learning_rate/qwen_vl_interface': 9.171772692914062e-06, 'learning_rate/action_model': 9.171772692914061e-05, 'epoch': 0.43}) +[2026-06-15 21:02:14,739][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.2564794719219208, 'train/grad_norm_pre_clip': 15.916814804077148, 'timing/data': 0.023324545996729285, 'timing/model': 1.135168349952437, 'learning_rate/qwen_vl_interface': 9.1700536521959e-06, 'learning_rate/action_model': 9.170053652195899e-05, 'epoch': 0.43}) +[2026-06-15 21:02:19,268][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.1850472241640091, 'train/grad_norm_pre_clip': 15.709259986877441, 'timing/data': 0.024520692008081824, 'timing/model': 1.1270241610473022, 'learning_rate/qwen_vl_interface': 9.168333000091754e-06, 'learning_rate/action_model': 9.168333000091754e-05, 'epoch': 0.43}) +[2026-06-15 21:02:23,789][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.3738764524459839, 'train/grad_norm_pre_clip': 14.092576026916504, 'timing/data': 0.024019445991143584, 'timing/model': 1.2875706399790943, 'learning_rate/qwen_vl_interface': 9.16661073730892e-06, 'learning_rate/action_model': 9.166610737308919e-05, 'epoch': 0.43}) +[2026-06-15 21:02:28,151][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.17781934142112732, 'train/grad_norm_pre_clip': 7.699611186981201, 'timing/data': 0.02494740200927481, 'timing/model': 1.1286540320143104, 'learning_rate/qwen_vl_interface': 9.164886864555355e-06, 'learning_rate/action_model': 9.164886864555355e-05, 'epoch': 0.43}) +[2026-06-15 21:02:32,661][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.13542021811008453, 'train/grad_norm_pre_clip': 5.288971900939941, 'timing/data': 0.02380088996142149, 'timing/model': 1.1283755840267986, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 0.43}) +[2026-06-15 21:02:37,131][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.15842117369174957, 'train/grad_norm_pre_clip': 5.654223442077637, 'timing/data': 0.0235412809997797, 'timing/model': 1.144481202994939, 'learning_rate/qwen_vl_interface': 9.16143429197117e-06, 'learning_rate/action_model': 9.16143429197117e-05, 'epoch': 0.43}) +[2026-06-15 21:02:41,681][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.092164546251297, 'train/grad_norm_pre_clip': 6.532527923583984, 'timing/data': 0.02472303598187864, 'timing/model': 1.1390883199637756, 'learning_rate/qwen_vl_interface': 9.159705593559773e-06, 'learning_rate/action_model': 9.159705593559773e-05, 'epoch': 0.43}) +[2026-06-15 21:02:46,218][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.1712130308151245, 'train/grad_norm_pre_clip': 7.767162799835205, 'timing/data': 0.02372976200422272, 'timing/model': 1.1320962290046737, 'learning_rate/qwen_vl_interface': 9.157975288016089e-06, 'learning_rate/action_model': 9.157975288016088e-05, 'epoch': 0.43}) +[2026-06-15 21:02:50,743][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.160598024725914, 'train/grad_norm_pre_clip': 5.2200822830200195, 'timing/data': 0.023739661963190883, 'timing/model': 1.2843985239742324, 'learning_rate/qwen_vl_interface': 9.156243376051381e-06, 'learning_rate/action_model': 9.156243376051381e-05, 'epoch': 0.43}) +[2026-06-15 21:02:55,226][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.17419490218162537, 'train/grad_norm_pre_clip': 6.894225597381592, 'timing/data': 0.023958716017659754, 'timing/model': 1.2434301109751686, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.43}) +[2026-06-15 21:02:59,603][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.2557961642742157, 'train/grad_norm_pre_clip': 10.443238258361816, 'timing/data': 0.024487876042257994, 'timing/model': 1.1332940669963136, 'learning_rate/qwen_vl_interface': 9.152774735707248e-06, 'learning_rate/action_model': 9.152774735707247e-05, 'epoch': 0.43}) +[2026-06-15 21:03:04,129][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.16163715720176697, 'train/grad_norm_pre_clip': 5.449812412261963, 'timing/data': 0.023641579959075898, 'timing/model': 1.133636933984235, 'learning_rate/qwen_vl_interface': 9.15103800875365e-06, 'learning_rate/action_model': 9.15103800875365e-05, 'epoch': 0.43}) +[2026-06-15 21:03:08,665][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.08801233023405075, 'train/grad_norm_pre_clip': 6.35499382019043, 'timing/data': 0.023120453988667578, 'timing/model': 1.1301085980376229, 'learning_rate/qwen_vl_interface': 9.149299678230681e-06, 'learning_rate/action_model': 9.149299678230681e-05, 'epoch': 0.43}) +[2026-06-15 21:03:13,205][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.20699521899223328, 'train/grad_norm_pre_clip': 6.55153751373291, 'timing/data': 0.022841717000119388, 'timing/model': 1.1405885870335624, 'learning_rate/qwen_vl_interface': 9.147559744852905e-06, 'learning_rate/action_model': 9.147559744852905e-05, 'epoch': 0.43}) +[2026-06-15 21:03:17,754][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.08155757933855057, 'train/grad_norm_pre_clip': 7.607632637023926, 'timing/data': 0.023438881966285408, 'timing/model': 1.3010956100188196, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 0.43}) +[2026-06-15 21:03:22,255][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.12220057845115662, 'train/grad_norm_pre_clip': 5.091341018676758, 'timing/data': 0.024327390012331307, 'timing/model': 1.2648224789882079, 'learning_rate/qwen_vl_interface': 9.14407507239447e-06, 'learning_rate/action_model': 9.14407507239447e-05, 'epoch': 0.43}) +[2026-06-15 21:03:26,634][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.21157008409500122, 'train/grad_norm_pre_clip': 8.590505599975586, 'timing/data': 0.023572098987642676, 'timing/model': 1.1335308330017142, 'learning_rate/qwen_vl_interface': 9.142330334746227e-06, 'learning_rate/action_model': 9.142330334746227e-05, 'epoch': 0.43}) +[2026-06-15 21:03:31,174][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.14545020461082458, 'train/grad_norm_pre_clip': 10.023283004760742, 'timing/data': 0.024013603979256004, 'timing/model': 1.1289474749937654, 'learning_rate/qwen_vl_interface': 9.14058399710801e-06, 'learning_rate/action_model': 9.14058399710801e-05, 'epoch': 0.43}) +[2026-06-15 21:03:35,719][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.10054376721382141, 'train/grad_norm_pre_clip': 6.12189245223999, 'timing/data': 0.023491321015171707, 'timing/model': 1.1348514709970914, 'learning_rate/qwen_vl_interface': 9.138836060197673e-06, 'learning_rate/action_model': 9.138836060197672e-05, 'epoch': 0.43}) +[2026-06-15 21:03:40,258][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.308956503868103, 'train/grad_norm_pre_clip': 12.360432624816895, 'timing/data': 0.02384800504660234, 'timing/model': 1.1321642479742877, 'learning_rate/qwen_vl_interface': 9.137086524733725e-06, 'learning_rate/action_model': 9.137086524733724e-05, 'epoch': 0.43}) +[2026-06-15 21:03:44,744][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.17875270545482635, 'train/grad_norm_pre_clip': 15.169787406921387, 'timing/data': 0.02381044701905921, 'timing/model': 1.1260124099790119, 'learning_rate/qwen_vl_interface': 9.135335391435336e-06, 'learning_rate/action_model': 9.135335391435335e-05, 'epoch': 0.43}) +[2026-06-15 21:03:49,256][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.2062024623155594, 'train/grad_norm_pre_clip': 10.215510368347168, 'timing/data': 0.024604528967756778, 'timing/model': 1.2863166669849306, 'learning_rate/qwen_vl_interface': 9.13358266102233e-06, 'learning_rate/action_model': 9.133582661022328e-05, 'epoch': 0.43}) +[2026-06-15 21:03:53,616][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.12467248737812042, 'train/grad_norm_pre_clip': 7.844451904296875, 'timing/data': 0.023365169996395707, 'timing/model': 1.1289000369724818, 'learning_rate/qwen_vl_interface': 9.131828334215185e-06, 'learning_rate/action_model': 9.131828334215185e-05, 'epoch': 0.43}) +[2026-06-15 21:03:58,148][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.06651318818330765, 'train/grad_norm_pre_clip': 5.2417192459106445, 'timing/data': 0.023818752029910684, 'timing/model': 1.1297474629827775, 'learning_rate/qwen_vl_interface': 9.130072411735043e-06, 'learning_rate/action_model': 9.130072411735043e-05, 'epoch': 0.44}) +[2026-06-15 21:04:02,585][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.13751180469989777, 'train/grad_norm_pre_clip': 7.317421913146973, 'timing/data': 0.023127202002797276, 'timing/model': 1.1384682530188002, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 0.44}) +[2026-06-15 21:04:07,037][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.31302908062934875, 'train/grad_norm_pre_clip': 13.636269569396973, 'timing/data': 0.0241188210202381, 'timing/model': 1.1340538010117598, 'learning_rate/qwen_vl_interface': 9.126555782643592e-06, 'learning_rate/action_model': 9.126555782643591e-05, 'epoch': 0.44}) +[2026-06-15 21:04:11,524][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.19160883128643036, 'train/grad_norm_pre_clip': 12.029104232788086, 'timing/data': 0.022882252000272274, 'timing/model': 1.143151124008, 'learning_rate/qwen_vl_interface': 9.124795077477835e-06, 'learning_rate/action_model': 9.124795077477834e-05, 'epoch': 0.44}) +[2026-06-15 21:04:16,010][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.30327853560447693, 'train/grad_norm_pre_clip': 14.21969985961914, 'timing/data': 0.022893534041941166, 'timing/model': 1.2517628229688853, 'learning_rate/qwen_vl_interface': 9.123032779530186e-06, 'learning_rate/action_model': 9.123032779530187e-05, 'epoch': 0.44}) +[2026-06-15 21:04:20,367][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.16486632823944092, 'train/grad_norm_pre_clip': 7.373392581939697, 'timing/data': 0.023701293976046145, 'timing/model': 1.1301993450033478, 'learning_rate/qwen_vl_interface': 9.12126888952506e-06, 'learning_rate/action_model': 9.12126888952506e-05, 'epoch': 0.44}) +[2026-06-15 21:04:24,875][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.16221365332603455, 'train/grad_norm_pre_clip': 7.107572555541992, 'timing/data': 0.023558522982057184, 'timing/model': 1.128911264007911, 'learning_rate/qwen_vl_interface': 9.119503408187521e-06, 'learning_rate/action_model': 9.119503408187522e-05, 'epoch': 0.44}) +[2026-06-15 21:04:29,402][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.019985130056738853, 'train/grad_norm_pre_clip': 1.4444043636322021, 'timing/data': 0.02348490501753986, 'timing/model': 1.1288998880190775, 'learning_rate/qwen_vl_interface': 9.117736336243299e-06, 'learning_rate/action_model': 9.117736336243298e-05, 'epoch': 0.44}) +[2026-06-15 21:04:33,926][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.398283988237381, 'train/grad_norm_pre_clip': 33.22173309326172, 'timing/data': 0.023495519999414682, 'timing/model': 1.1289282190264203, 'learning_rate/qwen_vl_interface': 9.115967674418765e-06, 'learning_rate/action_model': 9.115967674418765e-05, 'epoch': 0.44}) +[2026-06-15 21:04:38,478][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.2223169356584549, 'train/grad_norm_pre_clip': 7.7683329582214355, 'timing/data': 0.023226796998642385, 'timing/model': 1.302664010960143, 'learning_rate/qwen_vl_interface': 9.114197423440949e-06, 'learning_rate/action_model': 9.11419742344095e-05, 'epoch': 0.44}) +[2026-06-15 21:04:42,863][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.32829052209854126, 'train/grad_norm_pre_clip': 12.330021858215332, 'timing/data': 0.02424048795364797, 'timing/model': 1.13161531399237, 'learning_rate/qwen_vl_interface': 9.112425584037536e-06, 'learning_rate/action_model': 9.112425584037536e-05, 'epoch': 0.44}) +[2026-06-15 21:04:47,367][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.19774863123893738, 'train/grad_norm_pre_clip': 18.071956634521484, 'timing/data': 0.023768936982378364, 'timing/model': 1.1314104269840755, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 0.44}) +[2026-06-15 21:04:51,849][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.18053516745567322, 'train/grad_norm_pre_clip': 29.550884246826172, 'timing/data': 0.024175327969715, 'timing/model': 1.1332749550347216, 'learning_rate/qwen_vl_interface': 9.108877142867918e-06, 'learning_rate/action_model': 9.108877142867918e-05, 'epoch': 0.44}) +[2026-06-15 21:04:56,407][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.12947142124176025, 'train/grad_norm_pre_clip': 9.76879596710205, 'timing/data': 0.023828639008570462, 'timing/model': 1.1310538209509104, 'learning_rate/qwen_vl_interface': 9.107100542560342e-06, 'learning_rate/action_model': 9.107100542560342e-05, 'epoch': 0.44}) +[2026-06-15 21:05:00,943][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.1896725445985794, 'train/grad_norm_pre_clip': 17.221263885498047, 'timing/data': 0.023311536002438515, 'timing/model': 1.130353817017749, 'learning_rate/qwen_vl_interface': 9.105322356744429e-06, 'learning_rate/action_model': 9.105322356744429e-05, 'epoch': 0.44}) +[2026-06-15 21:05:05,501][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.1094476655125618, 'train/grad_norm_pre_clip': 5.864569187164307, 'timing/data': 0.022628747974522412, 'timing/model': 1.3062182019930333, 'learning_rate/qwen_vl_interface': 9.103542586151121e-06, 'learning_rate/action_model': 9.10354258615112e-05, 'epoch': 0.44}) +[2026-06-15 21:05:09,993][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.1396554559469223, 'train/grad_norm_pre_clip': 9.014249801635742, 'timing/data': 0.023654588032513857, 'timing/model': 1.2560654220287688, 'learning_rate/qwen_vl_interface': 9.101761231512017e-06, 'learning_rate/action_model': 9.101761231512017e-05, 'epoch': 0.44}) +[2026-06-15 21:05:14,356][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.22909913957118988, 'train/grad_norm_pre_clip': 8.291848182678223, 'timing/data': 0.02340709703275934, 'timing/model': 1.1272660219692625, 'learning_rate/qwen_vl_interface': 9.099978293559365e-06, 'learning_rate/action_model': 9.099978293559366e-05, 'epoch': 0.44}) +[2026-06-15 21:05:18,895][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.22502106428146362, 'train/grad_norm_pre_clip': 7.2424092292785645, 'timing/data': 0.02564326795982197, 'timing/model': 1.1349443680373952, 'learning_rate/qwen_vl_interface': 9.098193773026065e-06, 'learning_rate/action_model': 9.098193773026064e-05, 'epoch': 0.44}) +[2026-06-15 21:05:23,425][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.40798214077949524, 'train/grad_norm_pre_clip': 14.818479537963867, 'timing/data': 0.023492030974011868, 'timing/model': 1.1335810520104133, 'learning_rate/qwen_vl_interface': 9.096407670645663e-06, 'learning_rate/action_model': 9.096407670645662e-05, 'epoch': 0.44}) +[2026-06-15 21:05:27,958][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.038619264960289, 'train/grad_norm_pre_clip': 1.7253540754318237, 'timing/data': 0.023437263967934996, 'timing/model': 1.1308063509641215, 'learning_rate/qwen_vl_interface': 9.094619987152357e-06, 'learning_rate/action_model': 9.094619987152358e-05, 'epoch': 0.44}) +[2026-06-15 21:05:32,498][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.035407036542892456, 'train/grad_norm_pre_clip': 2.182776689529419, 'timing/data': 0.022579415002837777, 'timing/model': 1.2922184580238536, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.44}) +[2026-06-15 21:05:37,020][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.1640823632478714, 'train/grad_norm_pre_clip': 11.767369270324707, 'timing/data': 0.02363560802768916, 'timing/model': 1.2855617220047861, 'learning_rate/qwen_vl_interface': 9.09103987976709e-06, 'learning_rate/action_model': 9.09103987976709e-05, 'epoch': 0.44}) +[2026-06-15 21:05:41,376][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.17621701955795288, 'train/grad_norm_pre_clip': 4.329326629638672, 'timing/data': 0.023358603997621685, 'timing/model': 1.1313869439763948, 'learning_rate/qwen_vl_interface': 9.089247457346774e-06, 'learning_rate/action_model': 9.089247457346773e-05, 'epoch': 0.44}) +[2026-06-15 21:05:45,913][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.3693200349807739, 'train/grad_norm_pre_clip': 11.15900707244873, 'timing/data': 0.023452584980987012, 'timing/model': 1.130165856971871, 'learning_rate/qwen_vl_interface': 9.087453456756849e-06, 'learning_rate/action_model': 9.087453456756849e-05, 'epoch': 0.44}) +[2026-06-15 21:05:50,426][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.47534847259521484, 'train/grad_norm_pre_clip': 25.250228881835938, 'timing/data': 0.02362809400074184, 'timing/model': 1.1287265029968694, 'learning_rate/qwen_vl_interface': 9.085657878734759e-06, 'learning_rate/action_model': 9.085657878734758e-05, 'epoch': 0.45}) +[2026-06-15 21:05:54,982][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.12883031368255615, 'train/grad_norm_pre_clip': 13.521927833557129, 'timing/data': 0.02366688201436773, 'timing/model': 1.1302675739862025, 'learning_rate/qwen_vl_interface': 9.083860724018603e-06, 'learning_rate/action_model': 9.083860724018602e-05, 'epoch': 0.45}) +[2026-06-15 21:05:59,473][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.3044441342353821, 'train/grad_norm_pre_clip': 13.294917106628418, 'timing/data': 0.023823030991479754, 'timing/model': 1.131434507027734, 'learning_rate/qwen_vl_interface': 9.08206199334712e-06, 'learning_rate/action_model': 9.082061993347119e-05, 'epoch': 0.45}) +[2026-06-15 21:06:04,002][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.41606438159942627, 'train/grad_norm_pre_clip': 11.011425971984863, 'timing/data': 0.023307311988901347, 'timing/model': 1.29595929000061, 'learning_rate/qwen_vl_interface': 9.0802616874597e-06, 'learning_rate/action_model': 9.0802616874597e-05, 'epoch': 0.45}) +[2026-06-15 21:06:08,389][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.12704390287399292, 'train/grad_norm_pre_clip': 9.685900688171387, 'timing/data': 0.023795255983714014, 'timing/model': 1.1356247559888288, 'learning_rate/qwen_vl_interface': 9.078459807096385e-06, 'learning_rate/action_model': 9.078459807096385e-05, 'epoch': 0.45}) +[2026-06-15 21:06:12,921][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.2737221419811249, 'train/grad_norm_pre_clip': 12.302511215209961, 'timing/data': 0.0237127419677563, 'timing/model': 1.1328633049852215, 'learning_rate/qwen_vl_interface': 9.076656352997854e-06, 'learning_rate/action_model': 9.076656352997853e-05, 'epoch': 0.45}) +[2026-06-15 21:06:17,366][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.02378515899181366, 'train/grad_norm_pre_clip': 2.920144557952881, 'timing/data': 0.023695332987699658, 'timing/model': 1.1342086960212328, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 0.45}) +[2026-06-15 21:06:21,809][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.1161525547504425, 'train/grad_norm_pre_clip': 6.480813026428223, 'timing/data': 0.02354560198727995, 'timing/model': 1.1311898549902253, 'learning_rate/qwen_vl_interface': 9.073044726561128e-06, 'learning_rate/action_model': 9.073044726561129e-05, 'epoch': 0.45}) +[2026-06-15 21:06:26,286][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.1792384535074234, 'train/grad_norm_pre_clip': 18.255311965942383, 'timing/data': 0.023555960971862078, 'timing/model': 1.1311191249988042, 'learning_rate/qwen_vl_interface': 9.071236555707536e-06, 'learning_rate/action_model': 9.071236555707537e-05, 'epoch': 0.45}) +[2026-06-15 21:06:30,770][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.12344557791948318, 'train/grad_norm_pre_clip': 8.624065399169922, 'timing/data': 0.023725782986730337, 'timing/model': 1.2477825420210138, 'learning_rate/qwen_vl_interface': 9.069426814087937e-06, 'learning_rate/action_model': 9.069426814087937e-05, 'epoch': 0.45}) +[2026-06-15 21:06:35,138][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.053483255207538605, 'train/grad_norm_pre_clip': 4.509520053863525, 'timing/data': 0.02433973696315661, 'timing/model': 1.1306921530049294, 'learning_rate/qwen_vl_interface': 9.067615502446247e-06, 'learning_rate/action_model': 9.067615502446246e-05, 'epoch': 0.45}) +[2026-06-15 21:06:39,657][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.2227419912815094, 'train/grad_norm_pre_clip': 11.0510892868042, 'timing/data': 0.024020084005314857, 'timing/model': 1.127838381973561, 'learning_rate/qwen_vl_interface': 9.06580262152703e-06, 'learning_rate/action_model': 9.065802621527028e-05, 'epoch': 0.45}) +[2026-06-15 21:06:44,177][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.2871910631656647, 'train/grad_norm_pre_clip': 9.937370300292969, 'timing/data': 0.02329364197794348, 'timing/model': 1.129937008023262, 'learning_rate/qwen_vl_interface': 9.063988172075486e-06, 'learning_rate/action_model': 9.063988172075486e-05, 'epoch': 0.45}) +[2026-06-15 21:06:48,687][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.025802211835980415, 'train/grad_norm_pre_clip': 2.074315071105957, 'timing/data': 0.023509659979026765, 'timing/model': 1.1279202660080045, 'learning_rate/qwen_vl_interface': 9.062172154837475e-06, 'learning_rate/action_model': 9.062172154837475e-05, 'epoch': 0.45}) +[2026-06-15 21:06:53,203][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.15779879689216614, 'train/grad_norm_pre_clip': 13.900585174560547, 'timing/data': 0.023096886987332255, 'timing/model': 1.2862372760428116, 'learning_rate/qwen_vl_interface': 9.060354570559487e-06, 'learning_rate/action_model': 9.060354570559487e-05, 'epoch': 0.45}) +[2026-06-15 21:06:57,576][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.07792797684669495, 'train/grad_norm_pre_clip': 5.240233898162842, 'timing/data': 0.023472497996408492, 'timing/model': 1.1286527269985527, 'learning_rate/qwen_vl_interface': 9.058535419988667e-06, 'learning_rate/action_model': 9.058535419988667e-05, 'epoch': 0.45}) +[2026-06-15 21:11:19,167][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.062150947749614716, 'train/grad_norm_pre_clip': 4.770283222198486, 'eval/loss': 0.07767155021429062, 'eval/flap_precision': 0.9162929745889388, 'eval/flap_recall': 0.9459876543209876, 'eval/flap_f1': 0.9309035687167806, 'eval/flap_rate_model': 0.18583333333333332, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9309035687167805, 'timing/data': 0.023222610994707793, 'timing/model': 1.13166104699485, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 0.45}) +[2026-06-15 21:11:19,168][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1100_state +[2026-06-15 21:11:19,170][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 21:11:22,757][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1100_state/model.safetensors +[2026-06-15 21:11:32,849][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1100_state/optimizer.bin +[2026-06-15 21:11:32,849][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1100_state/sampler.bin +[2026-06-15 21:11:32,849][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1100_state/sampler_1.bin +[2026-06-15 21:11:32,850][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1100_state/random_states_0.pkl +[2026-06-15 21:11:32,891][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-06-15 21:11:32,892][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 21:11:32,898][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 21:11:37,204][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.06625974178314209, 'train/grad_norm_pre_clip': 5.760649681091309, 'timing/data': 0.026494466001167893, 'timing/model': 1.1319859560462646, 'learning_rate/qwen_vl_interface': 9.0548924229603e-06, 'learning_rate/action_model': 9.054892422960301e-05, 'epoch': 0.45}) +[2026-06-15 21:11:41,648][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.10134689509868622, 'train/grad_norm_pre_clip': 6.361509323120117, 'timing/data': 0.023472252010833472, 'timing/model': 1.1118316469946876, 'learning_rate/qwen_vl_interface': 9.053068578000256e-06, 'learning_rate/action_model': 9.053068578000256e-05, 'epoch': 0.45}) +[2026-06-15 21:11:46,078][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.057148855179548264, 'train/grad_norm_pre_clip': 8.891555786132812, 'timing/data': 0.024146291019860655, 'timing/model': 1.1118088900111616, 'learning_rate/qwen_vl_interface': 9.051243169742372e-06, 'learning_rate/action_model': 9.051243169742372e-05, 'epoch': 0.45}) +[2026-06-15 21:11:50,552][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.2044050246477127, 'train/grad_norm_pre_clip': 7.8896894454956055, 'timing/data': 0.023053824028465897, 'timing/model': 1.1199094980256632, 'learning_rate/qwen_vl_interface': 9.049416198937008e-06, 'learning_rate/action_model': 9.049416198937007e-05, 'epoch': 0.45}) +[2026-06-15 21:11:55,030][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.14826560020446777, 'train/grad_norm_pre_clip': 6.279329776763916, 'timing/data': 0.02397374901920557, 'timing/model': 1.1201083830092102, 'learning_rate/qwen_vl_interface': 9.047587666335159e-06, 'learning_rate/action_model': 9.047587666335159e-05, 'epoch': 0.45}) +[2026-06-15 21:11:59,531][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.2742970585823059, 'train/grad_norm_pre_clip': 13.221043586730957, 'timing/data': 0.024398012028541416, 'timing/model': 1.2725522240507416, 'learning_rate/qwen_vl_interface': 9.045757572688469e-06, 'learning_rate/action_model': 9.04575757268847e-05, 'epoch': 0.45}) +[2026-06-15 21:12:03,924][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.04501275718212128, 'train/grad_norm_pre_clip': 6.622699737548828, 'timing/data': 0.023593826044816524, 'timing/model': 1.1801829459727742, 'learning_rate/qwen_vl_interface': 9.043925918749218e-06, 'learning_rate/action_model': 9.043925918749218e-05, 'epoch': 0.45}) +[2026-06-15 21:12:08,253][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.19184276461601257, 'train/grad_norm_pre_clip': 8.207810401916504, 'timing/data': 0.024333306006155908, 'timing/model': 1.1208061440265737, 'learning_rate/qwen_vl_interface': 9.042092705270333e-06, 'learning_rate/action_model': 9.042092705270333e-05, 'epoch': 0.46}) +[2026-06-15 21:12:12,729][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.16982363164424896, 'train/grad_norm_pre_clip': 13.01709270477295, 'timing/data': 0.02449364395579323, 'timing/model': 1.1198339199763723, 'learning_rate/qwen_vl_interface': 9.040257933005375e-06, 'learning_rate/action_model': 9.040257933005374e-05, 'epoch': 0.46}) +[2026-06-15 21:12:17,242][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.18178722262382507, 'train/grad_norm_pre_clip': 5.860800743103027, 'timing/data': 0.0240395040018484, 'timing/model': 1.1267702570185065, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 0.46}) +[2026-06-15 21:12:21,757][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.15314431488513947, 'train/grad_norm_pre_clip': 12.001930236816406, 'timing/data': 0.024113280000165105, 'timing/model': 1.1232555740280077, 'learning_rate/qwen_vl_interface': 9.036583715134709e-06, 'learning_rate/action_model': 9.036583715134708e-05, 'epoch': 0.46}) +[2026-06-15 21:12:26,265][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.13033823668956757, 'train/grad_norm_pre_clip': 9.081754684448242, 'timing/data': 0.024026061000768095, 'timing/model': 1.2828498000162654, 'learning_rate/qwen_vl_interface': 9.034744271039332e-06, 'learning_rate/action_model': 9.034744271039332e-05, 'epoch': 0.46}) +[2026-06-15 21:12:30,707][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.18377715349197388, 'train/grad_norm_pre_clip': 6.446730613708496, 'timing/data': 0.02353017800487578, 'timing/model': 1.2148715740186162, 'learning_rate/qwen_vl_interface': 9.03290327117855e-06, 'learning_rate/action_model': 9.032903271178549e-05, 'epoch': 0.46}) +[2026-06-15 21:12:35,105][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.16115149855613708, 'train/grad_norm_pre_clip': 7.463656902313232, 'timing/data': 0.024001477984711528, 'timing/model': 1.1349880970083177, 'learning_rate/qwen_vl_interface': 9.031060716309122e-06, 'learning_rate/action_model': 9.031060716309121e-05, 'epoch': 0.46}) +[2026-06-15 21:12:39,597][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.12005717307329178, 'train/grad_norm_pre_clip': 4.873891353607178, 'timing/data': 0.03144731104839593, 'timing/model': 1.1377880300278775, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.46}) +[2026-06-15 21:12:44,130][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.21849168837070465, 'train/grad_norm_pre_clip': 13.570577621459961, 'timing/data': 0.023163369041867554, 'timing/model': 1.1319664939655922, 'learning_rate/qwen_vl_interface': 9.027370944574604e-06, 'learning_rate/action_model': 9.027370944574604e-05, 'epoch': 0.46}) +[2026-06-15 21:12:48,650][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.19517509639263153, 'train/grad_norm_pre_clip': 7.857376575469971, 'timing/data': 0.02444851998006925, 'timing/model': 1.134151263977401, 'learning_rate/qwen_vl_interface': 9.025523729226236e-06, 'learning_rate/action_model': 9.025523729226236e-05, 'epoch': 0.46}) +[2026-06-15 21:12:53,158][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.2053212672472, 'train/grad_norm_pre_clip': 9.743640899658203, 'timing/data': 0.024640256946440786, 'timing/model': 1.2803720739902928, 'learning_rate/qwen_vl_interface': 9.023674961902677e-06, 'learning_rate/action_model': 9.023674961902676e-05, 'epoch': 0.46}) +[2026-06-15 21:12:57,648][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.09870307147502899, 'train/grad_norm_pre_clip': 4.691661357879639, 'timing/data': 0.023582989000715315, 'timing/model': 1.2515620560152456, 'learning_rate/qwen_vl_interface': 9.021824643363883e-06, 'learning_rate/action_model': 9.021824643363883e-05, 'epoch': 0.46}) +[2026-06-15 21:13:02,006][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.2754446864128113, 'train/grad_norm_pre_clip': 13.510984420776367, 'timing/data': 0.0229846200090833, 'timing/model': 1.128572579997126, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 0.46}) +[2026-06-15 21:13:06,531][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.054332878440618515, 'train/grad_norm_pre_clip': 4.384704113006592, 'timing/data': 0.023510083032306284, 'timing/model': 1.137395032972563, 'learning_rate/qwen_vl_interface': 9.018119355683623e-06, 'learning_rate/action_model': 9.018119355683623e-05, 'epoch': 0.46}) +[2026-06-15 21:13:11,071][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.030434101819992065, 'train/grad_norm_pre_clip': 2.249542236328125, 'timing/data': 0.02341833896934986, 'timing/model': 1.1292402449762449, 'learning_rate/qwen_vl_interface': 9.016264388065258e-06, 'learning_rate/action_model': 9.016264388065258e-05, 'epoch': 0.46}) +[2026-06-15 21:13:15,599][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.11038818210363388, 'train/grad_norm_pre_clip': 7.404304504394531, 'timing/data': 0.02363267698092386, 'timing/model': 1.1275504170334898, 'learning_rate/qwen_vl_interface': 9.014407872277869e-06, 'learning_rate/action_model': 9.014407872277869e-05, 'epoch': 0.46}) +[2026-06-15 21:13:20,127][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.2299192249774933, 'train/grad_norm_pre_clip': 7.062971115112305, 'timing/data': 0.023330774041824043, 'timing/model': 1.2891389059950598, 'learning_rate/qwen_vl_interface': 9.012549809084595e-06, 'learning_rate/action_model': 9.012549809084595e-05, 'epoch': 0.46}) +[2026-06-15 21:13:24,656][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.09482896327972412, 'train/grad_norm_pre_clip': 6.657948017120361, 'timing/data': 0.023549730016384274, 'timing/model': 1.2669993169838563, 'learning_rate/qwen_vl_interface': 9.01069019924922e-06, 'learning_rate/action_model': 9.010690199249219e-05, 'epoch': 0.46}) +[2026-06-15 21:13:29,046][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.13681356608867645, 'train/grad_norm_pre_clip': 6.761501312255859, 'timing/data': 0.02348614000948146, 'timing/model': 1.130417710985057, 'learning_rate/qwen_vl_interface': 9.008829043536157e-06, 'learning_rate/action_model': 9.008829043536156e-05, 'epoch': 0.46}) +[2026-06-15 21:13:33,595][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.035286061465740204, 'train/grad_norm_pre_clip': 4.283596038818359, 'timing/data': 0.023235569009557366, 'timing/model': 1.1451234749983996, 'learning_rate/qwen_vl_interface': 9.006966342710458e-06, 'learning_rate/action_model': 9.006966342710457e-05, 'epoch': 0.46}) +[2026-06-15 21:13:38,138][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.12266617268323898, 'train/grad_norm_pre_clip': 4.425831317901611, 'timing/data': 0.023248434998095036, 'timing/model': 1.1353390979929827, 'learning_rate/qwen_vl_interface': 9.005102097537809e-06, 'learning_rate/action_model': 9.005102097537809e-05, 'epoch': 0.46}) +[2026-06-15 21:13:42,665][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.25227972865104675, 'train/grad_norm_pre_clip': 16.444997787475586, 'timing/data': 0.023109629983082414, 'timing/model': 1.131186140992213, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.46}) +[2026-06-15 21:13:47,154][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.02134786918759346, 'train/grad_norm_pre_clip': 2.5085556507110596, 'timing/data': 0.023553716018795967, 'timing/model': 1.1395773600088432, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 0.46}) +[2026-06-15 21:13:51,690][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.0803343877196312, 'train/grad_norm_pre_clip': 8.352676391601562, 'timing/data': 0.02316953201079741, 'timing/model': 1.2859854989801534, 'learning_rate/qwen_vl_interface': 8.999500103604538e-06, 'learning_rate/action_model': 8.999500103604536e-05, 'epoch': 0.46}) +[2026-06-15 21:13:56,065][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.17048999667167664, 'train/grad_norm_pre_clip': 9.554387092590332, 'timing/data': 0.023221719020511955, 'timing/model': 1.1420269949594513, 'learning_rate/qwen_vl_interface': 8.997629688713639e-06, 'learning_rate/action_model': 8.997629688713638e-05, 'epoch': 0.47}) +[2026-06-15 21:14:00,598][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.1327563375234604, 'train/grad_norm_pre_clip': 26.788232803344727, 'timing/data': 0.023136198986321688, 'timing/model': 1.1369387080194429, 'learning_rate/qwen_vl_interface': 8.995757733313733e-06, 'learning_rate/action_model': 8.995757733313732e-05, 'epoch': 0.47}) +[2026-06-15 21:14:05,060][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.14165405929088593, 'train/grad_norm_pre_clip': 8.841917037963867, 'timing/data': 0.02366479398915544, 'timing/model': 1.1442444659769535, 'learning_rate/qwen_vl_interface': 8.993884238174314e-06, 'learning_rate/action_model': 8.993884238174314e-05, 'epoch': 0.47}) +[2026-06-15 21:14:09,515][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.23107728362083435, 'train/grad_norm_pre_clip': 26.90411376953125, 'timing/data': 0.023855726001784205, 'timing/model': 1.1348490280215628, 'learning_rate/qwen_vl_interface': 8.992009204065505e-06, 'learning_rate/action_model': 8.992009204065505e-05, 'epoch': 0.47}) +[2026-06-15 21:14:13,996][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.3585244417190552, 'train/grad_norm_pre_clip': 4.849506378173828, 'timing/data': 0.023479892988689244, 'timing/model': 1.1351625180104747, 'learning_rate/qwen_vl_interface': 8.990132631758058e-06, 'learning_rate/action_model': 8.990132631758059e-05, 'epoch': 0.47}) +[2026-06-15 21:14:18,531][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.08007865399122238, 'train/grad_norm_pre_clip': 4.059097766876221, 'timing/data': 0.02290984697174281, 'timing/model': 1.262155479984358, 'learning_rate/qwen_vl_interface': 8.988254522023364e-06, 'learning_rate/action_model': 8.988254522023363e-05, 'epoch': 0.47}) +[2026-06-15 21:14:22,914][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.17433416843414307, 'train/grad_norm_pre_clip': 11.420830726623535, 'timing/data': 0.02352174004772678, 'timing/model': 1.1338767280103639, 'learning_rate/qwen_vl_interface': 8.986374875633441e-06, 'learning_rate/action_model': 8.98637487563344e-05, 'epoch': 0.47}) +[2026-06-15 21:14:27,488][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.04472345486283302, 'train/grad_norm_pre_clip': 6.476784706115723, 'timing/data': 0.023500481969676912, 'timing/model': 1.1442285669618286, 'learning_rate/qwen_vl_interface': 8.984493693360944e-06, 'learning_rate/action_model': 8.984493693360943e-05, 'epoch': 0.47}) +[2026-06-15 21:14:32,050][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.22230179607868195, 'train/grad_norm_pre_clip': 13.414399147033691, 'timing/data': 0.02294675202574581, 'timing/model': 1.1312952210428193, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 0.47}) +[2026-06-15 21:14:36,618][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.16101881861686707, 'train/grad_norm_pre_clip': 14.33130168914795, 'timing/data': 0.02426149504026398, 'timing/model': 1.1436856350046583, 'learning_rate/qwen_vl_interface': 8.98072672426198e-06, 'learning_rate/action_model': 8.980726724261981e-05, 'epoch': 0.47}) +[2026-06-15 21:14:41,173][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.044984765350818634, 'train/grad_norm_pre_clip': 4.189035892486572, 'timing/data': 0.02297625399660319, 'timing/model': 1.2964544359710999, 'learning_rate/qwen_vl_interface': 8.978840938983977e-06, 'learning_rate/action_model': 8.978840938983977e-05, 'epoch': 0.47}) +[2026-06-15 21:14:45,552][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.025255056098103523, 'train/grad_norm_pre_clip': 1.601616382598877, 'timing/data': 0.023337355989497155, 'timing/model': 1.1313411659793928, 'learning_rate/qwen_vl_interface': 8.976953620920313e-06, 'learning_rate/action_model': 8.976953620920312e-05, 'epoch': 0.47}) +[2026-06-15 21:14:50,074][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.12761248648166656, 'train/grad_norm_pre_clip': 2.9254326820373535, 'timing/data': 0.023577929998282343, 'timing/model': 1.1285268450155854, 'learning_rate/qwen_vl_interface': 8.975064770846794e-06, 'learning_rate/action_model': 8.975064770846794e-05, 'epoch': 0.47}) +[2026-06-15 21:14:54,562][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.37646666169166565, 'train/grad_norm_pre_clip': 12.249043464660645, 'timing/data': 0.023477900947909802, 'timing/model': 1.1476674430305138, 'learning_rate/qwen_vl_interface': 8.973174389539856e-06, 'learning_rate/action_model': 8.973174389539855e-05, 'epoch': 0.47}) +[2026-06-15 21:14:59,106][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.14877276122570038, 'train/grad_norm_pre_clip': 10.893070220947266, 'timing/data': 0.02340484200976789, 'timing/model': 1.1403886200278066, 'learning_rate/qwen_vl_interface': 8.971282477776563e-06, 'learning_rate/action_model': 8.971282477776563e-05, 'epoch': 0.47}) +[2026-06-15 21:15:03,675][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.20165640115737915, 'train/grad_norm_pre_clip': 13.278738975524902, 'timing/data': 0.023442003992386162, 'timing/model': 1.137089496012777, 'learning_rate/qwen_vl_interface': 8.969389036334609e-06, 'learning_rate/action_model': 8.969389036334609e-05, 'epoch': 0.47}) +[2026-06-15 21:15:08,205][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.15250010788440704, 'train/grad_norm_pre_clip': 7.2042012214660645, 'timing/data': 0.02349971595685929, 'timing/model': 1.2905644600396045, 'learning_rate/qwen_vl_interface': 8.967494065992315e-06, 'learning_rate/action_model': 8.967494065992315e-05, 'epoch': 0.47}) +[2026-06-15 21:15:12,720][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.16179867088794708, 'train/grad_norm_pre_clip': 11.295334815979004, 'timing/data': 0.027993425028398633, 'timing/model': 1.2561311550089158, 'learning_rate/qwen_vl_interface': 8.965597567528634e-06, 'learning_rate/action_model': 8.965597567528633e-05, 'epoch': 0.47}) +[2026-06-15 21:15:17,114][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.0725683718919754, 'train/grad_norm_pre_clip': 9.990776062011719, 'timing/data': 0.023299061052966863, 'timing/model': 1.150358613987919, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.47}) +[2026-06-15 21:15:21,652][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.06497140228748322, 'train/grad_norm_pre_clip': 2.9731204509735107, 'timing/data': 0.0231056860066019, 'timing/model': 1.1376280770055018, 'learning_rate/qwen_vl_interface': 8.961799989356048e-06, 'learning_rate/action_model': 8.961799989356048e-05, 'epoch': 0.47}) +[2026-06-15 21:15:26,215][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.3141498863697052, 'train/grad_norm_pre_clip': 15.905572891235352, 'timing/data': 0.02343817299697548, 'timing/model': 1.1384487919858657, 'learning_rate/qwen_vl_interface': 8.959898911208185e-06, 'learning_rate/action_model': 8.959898911208185e-05, 'epoch': 0.47}) +[2026-06-15 21:15:30,790][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.15883402526378632, 'train/grad_norm_pre_clip': 3.1966679096221924, 'timing/data': 0.02358623401960358, 'timing/model': 1.1448422770481557, 'learning_rate/qwen_vl_interface': 8.957996308061015e-06, 'learning_rate/action_model': 8.957996308061014e-05, 'epoch': 0.47}) +[2026-06-15 21:15:35,304][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.10973133146762848, 'train/grad_norm_pre_clip': 7.7185773849487305, 'timing/data': 0.023472995031625032, 'timing/model': 1.276988905039616, 'learning_rate/qwen_vl_interface': 8.956092180696626e-06, 'learning_rate/action_model': 8.956092180696625e-05, 'epoch': 0.47}) +[2026-06-15 21:15:39,858][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.3189443349838257, 'train/grad_norm_pre_clip': 4.4063849449157715, 'timing/data': 0.023346834001131356, 'timing/model': 1.28089794999687, 'learning_rate/qwen_vl_interface': 8.95418652989773e-06, 'learning_rate/action_model': 8.95418652989773e-05, 'epoch': 0.47}) +[2026-06-15 21:15:44,241][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.13489441573619843, 'train/grad_norm_pre_clip': 18.534936904907227, 'timing/data': 0.022727639006916434, 'timing/model': 1.1319154250086285, 'learning_rate/qwen_vl_interface': 8.952279356447675e-06, 'learning_rate/action_model': 8.952279356447675e-05, 'epoch': 0.47}) +[2026-06-15 21:15:48,785][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.17374224960803986, 'train/grad_norm_pre_clip': 8.030576705932617, 'timing/data': 0.023320077045354992, 'timing/model': 1.135206995997578, 'learning_rate/qwen_vl_interface': 8.950370661130421e-06, 'learning_rate/action_model': 8.950370661130422e-05, 'epoch': 0.48}) +[2026-06-15 21:15:53,336][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.11128908395767212, 'train/grad_norm_pre_clip': 3.029296875, 'timing/data': 0.023489431012421846, 'timing/model': 1.1365781229687855, 'learning_rate/qwen_vl_interface': 8.948460444730566e-06, 'learning_rate/action_model': 8.948460444730566e-05, 'epoch': 0.48}) +[2026-06-15 21:15:57,874][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.3449205160140991, 'train/grad_norm_pre_clip': 10.229135513305664, 'timing/data': 0.02387949201511219, 'timing/model': 1.133878119988367, 'learning_rate/qwen_vl_interface': 8.946548708033324e-06, 'learning_rate/action_model': 8.946548708033323e-05, 'epoch': 0.48}) +[2026-06-15 21:16:02,372][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.3107661008834839, 'train/grad_norm_pre_clip': 19.763668060302734, 'timing/data': 0.023073474992997944, 'timing/model': 1.1354758540401235, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 0.48}) +[2026-06-15 21:16:06,915][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.10981111973524094, 'train/grad_norm_pre_clip': 8.876368522644043, 'timing/data': 0.023112800961825997, 'timing/model': 1.2913644489599392, 'learning_rate/qwen_vl_interface': 8.942720676890677e-06, 'learning_rate/action_model': 8.942720676890677e-05, 'epoch': 0.48}) +[2026-06-15 21:16:11,308][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.47453776001930237, 'train/grad_norm_pre_clip': 13.717060089111328, 'timing/data': 0.023249255027621984, 'timing/model': 1.1378712879959494, 'learning_rate/qwen_vl_interface': 8.940804384018833e-06, 'learning_rate/action_model': 8.940804384018832e-05, 'epoch': 0.48}) +[2026-06-15 21:16:15,875][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.1609259843826294, 'train/grad_norm_pre_clip': 7.466950416564941, 'timing/data': 0.026752579025924206, 'timing/model': 1.1384291040012613, 'learning_rate/qwen_vl_interface': 8.93888657399672e-06, 'learning_rate/action_model': 8.93888657399672e-05, 'epoch': 0.48}) +[2026-06-15 21:16:20,303][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.20018891990184784, 'train/grad_norm_pre_clip': 17.36100196838379, 'timing/data': 0.02345624699955806, 'timing/model': 1.13088215602329, 'learning_rate/qwen_vl_interface': 8.936967247612678e-06, 'learning_rate/action_model': 8.936967247612678e-05, 'epoch': 0.48}) +[2026-06-15 21:16:24,748][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.1252373605966568, 'train/grad_norm_pre_clip': 5.870023727416992, 'timing/data': 0.03139358601765707, 'timing/model': 1.1358431410044432, 'learning_rate/qwen_vl_interface': 8.93504640565567e-06, 'learning_rate/action_model': 8.93504640565567e-05, 'epoch': 0.48}) +[2026-06-15 21:16:29,224][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.3966364860534668, 'train/grad_norm_pre_clip': 11.526451110839844, 'timing/data': 0.02405185700627044, 'timing/model': 1.1459419730235822, 'learning_rate/qwen_vl_interface': 8.933124048915278e-06, 'learning_rate/action_model': 8.933124048915278e-05, 'epoch': 0.48}) +[2026-06-15 21:16:33,760][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.10729570686817169, 'train/grad_norm_pre_clip': 6.768152713775635, 'timing/data': 0.023161719960626215, 'timing/model': 1.2777707349741831, 'learning_rate/qwen_vl_interface': 8.931200178181715e-06, 'learning_rate/action_model': 8.931200178181715e-05, 'epoch': 0.48}) +[2026-06-15 21:16:38,166][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.14742237329483032, 'train/grad_norm_pre_clip': 7.520225524902344, 'timing/data': 0.024183458008337766, 'timing/model': 1.1405325249652378, 'learning_rate/qwen_vl_interface': 8.929274794245812e-06, 'learning_rate/action_model': 8.929274794245812e-05, 'epoch': 0.48}) +[2026-06-15 21:16:42,742][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.22304292023181915, 'train/grad_norm_pre_clip': 10.30691146850586, 'timing/data': 0.023001905996352434, 'timing/model': 1.1331849329872057, 'learning_rate/qwen_vl_interface': 8.927347897899018e-06, 'learning_rate/action_model': 8.927347897899016e-05, 'epoch': 0.48}) +[2026-06-15 21:16:47,296][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.15950481593608856, 'train/grad_norm_pre_clip': 8.776713371276855, 'timing/data': 0.02337423199787736, 'timing/model': 1.1325658130226657, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 0.48}) +[2026-06-15 21:16:51,865][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.13479256629943848, 'train/grad_norm_pre_clip': 7.2163777351379395, 'timing/data': 0.023435507027897984, 'timing/model': 1.1413310570060275, 'learning_rate/qwen_vl_interface': 8.923489571141679e-06, 'learning_rate/action_model': 8.923489571141679e-05, 'epoch': 0.48}) +[2026-06-15 21:16:56,433][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.17179806530475616, 'train/grad_norm_pre_clip': 5.895420551300049, 'timing/data': 0.023359594983048737, 'timing/model': 1.2964228549972177, 'learning_rate/qwen_vl_interface': 8.921558142317147e-06, 'learning_rate/action_model': 8.921558142317147e-05, 'epoch': 0.48}) +[2026-06-15 21:17:00,813][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.23598653078079224, 'train/grad_norm_pre_clip': 9.80435848236084, 'timing/data': 0.023910182004328817, 'timing/model': 1.1373085820232518, 'learning_rate/qwen_vl_interface': 8.91962520425375e-06, 'learning_rate/action_model': 8.91962520425375e-05, 'epoch': 0.48}) +[2026-06-15 21:17:05,371][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.09644220769405365, 'train/grad_norm_pre_clip': 5.617591381072998, 'timing/data': 0.023426997009664774, 'timing/model': 1.1374195190146565, 'learning_rate/qwen_vl_interface': 8.917690757746044e-06, 'learning_rate/action_model': 8.917690757746044e-05, 'epoch': 0.48}) +[2026-06-15 21:17:09,857][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.13709832727909088, 'train/grad_norm_pre_clip': 6.681721210479736, 'timing/data': 0.024603719997685403, 'timing/model': 1.1382857819553465, 'learning_rate/qwen_vl_interface': 8.915754803589208e-06, 'learning_rate/action_model': 8.915754803589209e-05, 'epoch': 0.48}) +[2026-06-15 21:17:14,431][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.05663415417075157, 'train/grad_norm_pre_clip': 3.9545767307281494, 'timing/data': 0.02359731198521331, 'timing/model': 1.1315749380155466, 'learning_rate/qwen_vl_interface': 8.913817342579042e-06, 'learning_rate/action_model': 8.913817342579042e-05, 'epoch': 0.48}) +[2026-06-15 21:17:18,981][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.067485012114048, 'train/grad_norm_pre_clip': 8.466525077819824, 'timing/data': 0.023378348036203533, 'timing/model': 1.1301740680355579, 'learning_rate/qwen_vl_interface': 8.91187837551196e-06, 'learning_rate/action_model': 8.911878375511961e-05, 'epoch': 0.48}) +[2026-06-15 21:17:23,518][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.204940065741539, 'train/grad_norm_pre_clip': 7.5099945068359375, 'timing/data': 0.02290983498096466, 'timing/model': 1.2870238489704207, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.48}) +[2026-06-15 21:17:28,021][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.25774985551834106, 'train/grad_norm_pre_clip': 7.819265842437744, 'timing/data': 0.023233456013258547, 'timing/model': 1.2533771460293792, 'learning_rate/qwen_vl_interface': 8.907995926395814e-06, 'learning_rate/action_model': 8.907995926395815e-05, 'epoch': 0.48}) +[2026-06-15 21:17:32,417][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.0590381845831871, 'train/grad_norm_pre_clip': 8.036056518554688, 'timing/data': 0.023166165046859533, 'timing/model': 1.133644252025988, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 0.48}) +[2026-06-15 21:17:36,970][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.07661403715610504, 'train/grad_norm_pre_clip': 11.624999046325684, 'timing/data': 0.023117502976674587, 'timing/model': 1.1328033109894022, 'learning_rate/qwen_vl_interface': 8.904107462624484e-06, 'learning_rate/action_model': 8.904107462624483e-05, 'epoch': 0.49}) +[2026-06-15 21:17:41,532][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.15579073131084442, 'train/grad_norm_pre_clip': 7.9361186027526855, 'timing/data': 0.023373767966404557, 'timing/model': 1.1390310870483518, 'learning_rate/qwen_vl_interface': 8.902160977240739e-06, 'learning_rate/action_model': 8.902160977240739e-05, 'epoch': 0.49}) +[2026-06-15 21:17:46,097][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.21918021142482758, 'train/grad_norm_pre_clip': 8.879541397094727, 'timing/data': 0.024054294975940138, 'timing/model': 1.1289769429713488, 'learning_rate/qwen_vl_interface': 8.90021299059157e-06, 'learning_rate/action_model': 8.90021299059157e-05, 'epoch': 0.49}) +[2026-06-15 21:17:50,647][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.2249486744403839, 'train/grad_norm_pre_clip': 9.074491500854492, 'timing/data': 0.023220703995320946, 'timing/model': 1.2917807600460947, 'learning_rate/qwen_vl_interface': 8.898263503477722e-06, 'learning_rate/action_model': 8.898263503477721e-05, 'epoch': 0.49}) +[2026-06-15 21:17:55,194][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.12530602514743805, 'train/grad_norm_pre_clip': 11.382469177246094, 'timing/data': 0.02325138502055779, 'timing/model': 1.29112550400896, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.49}) +[2026-06-15 21:17:59,594][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.38623782992362976, 'train/grad_norm_pre_clip': 18.574342727661133, 'timing/data': 0.02285146398935467, 'timing/model': 1.1331206550239585, 'learning_rate/qwen_vl_interface': 8.894360031062043e-06, 'learning_rate/action_model': 8.894360031062043e-05, 'epoch': 0.49}) +[2026-06-15 21:18:04,150][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.27143585681915283, 'train/grad_norm_pre_clip': 12.323273658752441, 'timing/data': 0.023472739034332335, 'timing/model': 1.1356154109816998, 'learning_rate/qwen_vl_interface': 8.892406047364783e-06, 'learning_rate/action_model': 8.892406047364784e-05, 'epoch': 0.49}) +[2026-06-15 21:18:08,701][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.3764727711677551, 'train/grad_norm_pre_clip': 11.261748313903809, 'timing/data': 0.023113807023037225, 'timing/model': 1.1307379089994356, 'learning_rate/qwen_vl_interface': 8.890450566411984e-06, 'learning_rate/action_model': 8.890450566411984e-05, 'epoch': 0.49}) +[2026-06-15 21:18:13,249][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.13938705623149872, 'train/grad_norm_pre_clip': 8.067959785461426, 'timing/data': 0.02316638402407989, 'timing/model': 1.1403614589944482, 'learning_rate/qwen_vl_interface': 8.888493589007467e-06, 'learning_rate/action_model': 8.888493589007468e-05, 'epoch': 0.49}) +[2026-06-15 21:18:17,758][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.1125480905175209, 'train/grad_norm_pre_clip': 4.705432891845703, 'timing/data': 0.02341908897506073, 'timing/model': 1.1371173689840361, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 0.49}) +[2026-06-15 21:18:22,284][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.09662164747714996, 'train/grad_norm_pre_clip': 3.2803146839141846, 'timing/data': 0.02370783797232434, 'timing/model': 1.2827910259948112, 'learning_rate/qwen_vl_interface': 8.884575148061662e-06, 'learning_rate/action_model': 8.88457514806166e-05, 'epoch': 0.49}) +[2026-06-15 21:18:26,673][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.07036891579627991, 'train/grad_norm_pre_clip': 2.216726303100586, 'timing/data': 0.023534200037829578, 'timing/model': 1.1312162060057744, 'learning_rate/qwen_vl_interface': 8.882613686131095e-06, 'learning_rate/action_model': 8.882613686131093e-05, 'epoch': 0.49}) +[2026-06-15 21:18:31,235][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.054444320499897, 'train/grad_norm_pre_clip': 2.967007637023926, 'timing/data': 0.022654858010355383, 'timing/model': 1.1426001460058615, 'learning_rate/qwen_vl_interface': 8.880650730970254e-06, 'learning_rate/action_model': 8.880650730970254e-05, 'epoch': 0.49}) +[2026-06-15 21:18:35,688][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.221986785531044, 'train/grad_norm_pre_clip': 12.801176071166992, 'timing/data': 0.022889245999976993, 'timing/model': 1.1398588080191985, 'learning_rate/qwen_vl_interface': 8.878686283386042e-06, 'learning_rate/action_model': 8.878686283386042e-05, 'epoch': 0.49}) +[2026-06-15 21:18:40,174][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.041250504553318024, 'train/grad_norm_pre_clip': 1.932419776916504, 'timing/data': 0.02333751501282677, 'timing/model': 1.1324978390475735, 'learning_rate/qwen_vl_interface': 8.876720344185967e-06, 'learning_rate/action_model': 8.876720344185967e-05, 'epoch': 0.49}) +[2026-06-15 21:18:44,656][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.08397474884986877, 'train/grad_norm_pre_clip': 5.103782653808594, 'timing/data': 0.02303144003963098, 'timing/model': 1.1361105190007947, 'learning_rate/qwen_vl_interface': 8.874752914178149e-06, 'learning_rate/action_model': 8.87475291417815e-05, 'epoch': 0.49}) +[2026-06-15 21:18:49,169][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.07152270525693893, 'train/grad_norm_pre_clip': 8.432655334472656, 'timing/data': 0.023234213993418962, 'timing/model': 1.268929619982373, 'learning_rate/qwen_vl_interface': 8.87278399417133e-06, 'learning_rate/action_model': 8.872783994171329e-05, 'epoch': 0.49}) +[2026-06-15 21:18:53,566][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.0729648768901825, 'train/grad_norm_pre_clip': 3.733849048614502, 'timing/data': 0.023152974026743323, 'timing/model': 1.1346147559816018, 'learning_rate/qwen_vl_interface': 8.870813584974852e-06, 'learning_rate/action_model': 8.870813584974852e-05, 'epoch': 0.49}) +[2026-06-15 21:18:58,104][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.08528847247362137, 'train/grad_norm_pre_clip': 6.9170942306518555, 'timing/data': 0.02327613701345399, 'timing/model': 1.1357351900078356, 'learning_rate/qwen_vl_interface': 8.868841687398683e-06, 'learning_rate/action_model': 8.868841687398683e-05, 'epoch': 0.49}) +[2026-06-15 21:23:20,410][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.018788082525134087, 'train/grad_norm_pre_clip': 1.6781235933303833, 'eval/loss': 0.06726726144552231, 'eval/flap_precision': 0.9172821270310192, 'eval/flap_recall': 0.9583333333333334, 'eval/flap_f1': 0.9373584905660377, 'eval/flap_rate_model': 0.18805555555555556, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9373584905660377, 'timing/data': 0.02298402797896415, 'timing/model': 1.1329310089931823, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 0.49}) +[2026-06-15 21:23:20,410][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1200_state +[2026-06-15 21:23:20,413][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 21:23:24,004][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1200_state/model.safetensors +[2026-06-15 21:23:34,061][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1200_state/optimizer.bin +[2026-06-15 21:23:34,062][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1200_state/sampler.bin +[2026-06-15 21:23:34,062][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1200_state/sampler_1.bin +[2026-06-15 21:23:34,062][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1200_state/random_states_0.pkl +[2026-06-15 21:23:34,108][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-06-15 21:23:34,108][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 21:23:34,115][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 21:23:38,585][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.05432881787419319, 'train/grad_norm_pre_clip': 7.713785648345947, 'timing/data': 0.02992643998004496, 'timing/model': 1.124797120981384, 'learning_rate/qwen_vl_interface': 8.864893430350159e-06, 'learning_rate/action_model': 8.864893430350159e-05, 'epoch': 0.49}) +[2026-06-15 21:23:43,036][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.2286670058965683, 'train/grad_norm_pre_clip': 10.50073528289795, 'timing/data': 0.023591215955093503, 'timing/model': 1.1240444229915738, 'learning_rate/qwen_vl_interface': 8.862917072500785e-06, 'learning_rate/action_model': 8.862917072500784e-05, 'epoch': 0.49}) +[2026-06-15 21:23:47,479][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.028661683201789856, 'train/grad_norm_pre_clip': 2.004730224609375, 'timing/data': 0.02314967301208526, 'timing/model': 1.11693290801486, 'learning_rate/qwen_vl_interface': 8.860939229517674e-06, 'learning_rate/action_model': 8.860939229517674e-05, 'epoch': 0.49}) +[2026-06-15 21:23:51,967][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.2355072945356369, 'train/grad_norm_pre_clip': 11.46768856048584, 'timing/data': 0.02287451404845342, 'timing/model': 1.288595940975938, 'learning_rate/qwen_vl_interface': 8.858959902213844e-06, 'learning_rate/action_model': 8.858959902213845e-05, 'epoch': 0.49}) +[2026-06-15 21:23:56,279][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.08308390527963638, 'train/grad_norm_pre_clip': 14.626687049865723, 'timing/data': 0.022958431974984705, 'timing/model': 1.1190591350314207, 'learning_rate/qwen_vl_interface': 8.856979091402921e-06, 'learning_rate/action_model': 8.85697909140292e-05, 'epoch': 0.5}) +[2026-06-15 21:24:00,769][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.01293218694627285, 'train/grad_norm_pre_clip': 1.3380261659622192, 'timing/data': 0.02323460904881358, 'timing/model': 1.1205911010038108, 'learning_rate/qwen_vl_interface': 8.854996797899139e-06, 'learning_rate/action_model': 8.854996797899139e-05, 'epoch': 0.5}) +[2026-06-15 21:24:05,287][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.04840776324272156, 'train/grad_norm_pre_clip': 4.467007160186768, 'timing/data': 0.022991415986325592, 'timing/model': 1.1356341789942235, 'learning_rate/qwen_vl_interface': 8.853013022517348e-06, 'learning_rate/action_model': 8.853013022517348e-05, 'epoch': 0.5}) +[2026-06-15 21:24:09,768][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.3363393545150757, 'train/grad_norm_pre_clip': 8.972620964050293, 'timing/data': 0.023073946009390056, 'timing/model': 1.1211286240140907, 'learning_rate/qwen_vl_interface': 8.851027766073002e-06, 'learning_rate/action_model': 8.851027766073001e-05, 'epoch': 0.5}) +[2026-06-15 21:24:14,289][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.014719159342348576, 'train/grad_norm_pre_clip': 2.9653823375701904, 'timing/data': 0.022881161014083773, 'timing/model': 1.1299044609768316, 'learning_rate/qwen_vl_interface': 8.84904102938216e-06, 'learning_rate/action_model': 8.84904102938216e-05, 'epoch': 0.5}) +[2026-06-15 21:24:18,810][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.012554211542010307, 'train/grad_norm_pre_clip': 2.729210615158081, 'timing/data': 0.023351551964879036, 'timing/model': 1.294820429990068, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 0.5}) +[2026-06-15 21:24:23,143][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.09062647074460983, 'train/grad_norm_pre_clip': 17.293792724609375, 'timing/data': 0.023924254986923188, 'timing/model': 1.1221751010161825, 'learning_rate/qwen_vl_interface': 8.845063118528303e-06, 'learning_rate/action_model': 8.845063118528304e-05, 'epoch': 0.5}) +[2026-06-15 21:24:27,631][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.2546994686126709, 'train/grad_norm_pre_clip': 22.309595108032227, 'timing/data': 0.023362809035461396, 'timing/model': 1.1228113920078613, 'learning_rate/qwen_vl_interface': 8.843071946000455e-06, 'learning_rate/action_model': 8.843071946000456e-05, 'epoch': 0.5}) +[2026-06-15 21:24:32,155][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.10247118771076202, 'train/grad_norm_pre_clip': 23.616973876953125, 'timing/data': 0.023778172966558486, 'timing/model': 1.122978902014438, 'learning_rate/qwen_vl_interface': 8.841079296496451e-06, 'learning_rate/action_model': 8.84107929649645e-05, 'epoch': 0.5}) +[2026-06-15 21:24:36,659][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.4227122366428375, 'train/grad_norm_pre_clip': 22.037397384643555, 'timing/data': 0.023008949996437877, 'timing/model': 1.1246622640173882, 'learning_rate/qwen_vl_interface': 8.839085170835396e-06, 'learning_rate/action_model': 8.839085170835395e-05, 'epoch': 0.5}) +[2026-06-15 21:24:41,168][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.4993291199207306, 'train/grad_norm_pre_clip': 34.389801025390625, 'timing/data': 0.02325411100173369, 'timing/model': 1.2764222259866074, 'learning_rate/qwen_vl_interface': 8.837089569837e-06, 'learning_rate/action_model': 8.837089569836998e-05, 'epoch': 0.5}) +[2026-06-15 21:24:45,683][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.2967391908168793, 'train/grad_norm_pre_clip': 17.931339263916016, 'timing/data': 0.023566499003209174, 'timing/model': 1.2849807469756342, 'learning_rate/qwen_vl_interface': 8.835092494321575e-06, 'learning_rate/action_model': 8.835092494321575e-05, 'epoch': 0.5}) +[2026-06-15 21:24:50,034][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.3512510359287262, 'train/grad_norm_pre_clip': 20.219322204589844, 'timing/data': 0.023689327004831284, 'timing/model': 1.1269115970353596, 'learning_rate/qwen_vl_interface': 8.833093945110048e-06, 'learning_rate/action_model': 8.833093945110047e-05, 'epoch': 0.5}) +[2026-06-15 21:24:54,551][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.15309125185012817, 'train/grad_norm_pre_clip': 9.241822242736816, 'timing/data': 0.02320444700308144, 'timing/model': 1.128463383007329, 'learning_rate/qwen_vl_interface': 8.831093923023946e-06, 'learning_rate/action_model': 8.831093923023945e-05, 'epoch': 0.5}) +[2026-06-15 21:24:59,068][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.029374653473496437, 'train/grad_norm_pre_clip': 9.272010803222656, 'timing/data': 0.023339906008914113, 'timing/model': 1.1245172939961776, 'learning_rate/qwen_vl_interface': 8.8290924288854e-06, 'learning_rate/action_model': 8.8290924288854e-05, 'epoch': 0.5}) +[2026-06-15 21:25:03,586][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.026339232921600342, 'train/grad_norm_pre_clip': 9.523760795593262, 'timing/data': 0.023960585007444024, 'timing/model': 1.1285123840207234, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.5}) +[2026-06-15 21:25:08,068][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.02821851335465908, 'train/grad_norm_pre_clip': 4.945338726043701, 'timing/data': 0.02376193698728457, 'timing/model': 1.1316295640426688, 'learning_rate/qwen_vl_interface': 8.825085027742545e-06, 'learning_rate/action_model': 8.825085027742544e-05, 'epoch': 0.5}) +[2026-06-15 21:25:12,597][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.005888903979212046, 'train/grad_norm_pre_clip': 1.041817545890808, 'timing/data': 0.023711839981842786, 'timing/model': 1.2910687400144525, 'learning_rate/qwen_vl_interface': 8.823079122385523e-06, 'learning_rate/action_model': 8.823079122385523e-05, 'epoch': 0.5}) +[2026-06-15 21:25:16,965][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.21039965748786926, 'train/grad_norm_pre_clip': 17.799015045166016, 'timing/data': 0.023265103984158486, 'timing/model': 1.1291085630073212, 'learning_rate/qwen_vl_interface': 8.821071748270644e-06, 'learning_rate/action_model': 8.821071748270644e-05, 'epoch': 0.5}) +[2026-06-15 21:25:21,487][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.006487758364528418, 'train/grad_norm_pre_clip': 0.5909765362739563, 'timing/data': 0.023515080974902958, 'timing/model': 1.1294570079771802, 'learning_rate/qwen_vl_interface': 8.81906290622306e-06, 'learning_rate/action_model': 8.81906290622306e-05, 'epoch': 0.5}) +[2026-06-15 21:25:25,908][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.11100570112466812, 'train/grad_norm_pre_clip': 9.578989028930664, 'timing/data': 0.02333100902615115, 'timing/model': 1.1276796740130521, 'learning_rate/qwen_vl_interface': 8.81705259706853e-06, 'learning_rate/action_model': 8.817052597068529e-05, 'epoch': 0.5}) +[2026-06-15 21:25:30,344][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.1916746348142624, 'train/grad_norm_pre_clip': 16.590774536132812, 'timing/data': 0.023608823015820235, 'timing/model': 1.130333321983926, 'learning_rate/qwen_vl_interface': 8.815040821633416e-06, 'learning_rate/action_model': 8.815040821633416e-05, 'epoch': 0.5}) +[2026-06-15 21:25:34,804][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.013651985675096512, 'train/grad_norm_pre_clip': 6.483089447021484, 'timing/data': 0.02366153697948903, 'timing/model': 1.1277078380226158, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.5}) +[2026-06-15 21:25:39,284][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.20733016729354858, 'train/grad_norm_pre_clip': 13.201029777526855, 'timing/data': 0.024208205053582788, 'timing/model': 1.2518098430009559, 'learning_rate/qwen_vl_interface': 8.8110128752299e-06, 'learning_rate/action_model': 8.8110128752299e-05, 'epoch': 0.5}) +[2026-06-15 21:25:43,647][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.12036892026662827, 'train/grad_norm_pre_clip': 14.55574893951416, 'timing/data': 0.02331247099209577, 'timing/model': 1.1305202529765666, 'learning_rate/qwen_vl_interface': 8.808996705917235e-06, 'learning_rate/action_model': 8.808996705917235e-05, 'epoch': 0.5}) +[2026-06-15 21:25:48,197][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.07556197047233582, 'train/grad_norm_pre_clip': 11.168675422668457, 'timing/data': 0.024022622033953667, 'timing/model': 1.1299338730168529, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 0.51}) +[2026-06-15 21:25:52,728][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.4158182144165039, 'train/grad_norm_pre_clip': 20.18855094909668, 'timing/data': 0.023034594021737576, 'timing/model': 1.1279713800176978, 'learning_rate/qwen_vl_interface': 8.804959979213945e-06, 'learning_rate/action_model': 8.804959979213945e-05, 'epoch': 0.51}) +[2026-06-15 21:25:57,265][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.011166740208864212, 'train/grad_norm_pre_clip': 2.9979591369628906, 'timing/data': 0.023157689021900296, 'timing/model': 1.1306272660149261, 'learning_rate/qwen_vl_interface': 8.802939423482667e-06, 'learning_rate/action_model': 8.802939423482666e-05, 'epoch': 0.51}) +[2026-06-15 21:26:01,802][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.1192878857254982, 'train/grad_norm_pre_clip': 12.306741714477539, 'timing/data': 0.02333041501697153, 'timing/model': 1.2984008549829014, 'learning_rate/qwen_vl_interface': 8.800917407272198e-06, 'learning_rate/action_model': 8.800917407272198e-05, 'epoch': 0.51}) +[2026-06-15 21:26:06,176][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.14072105288505554, 'train/grad_norm_pre_clip': 14.472480773925781, 'timing/data': 0.023319162020925432, 'timing/model': 1.130403529969044, 'learning_rate/qwen_vl_interface': 8.798893931413711e-06, 'learning_rate/action_model': 8.79889393141371e-05, 'epoch': 0.51}) +[2026-06-15 21:26:10,696][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.38318949937820435, 'train/grad_norm_pre_clip': 37.729549407958984, 'timing/data': 0.023407245986163616, 'timing/model': 1.1305864929454401, 'learning_rate/qwen_vl_interface': 8.796868996738985e-06, 'learning_rate/action_model': 8.796868996738984e-05, 'epoch': 0.51}) +[2026-06-15 21:26:15,186][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.040599722415208817, 'train/grad_norm_pre_clip': 10.005084991455078, 'timing/data': 0.02288383897393942, 'timing/model': 1.129771486041136, 'learning_rate/qwen_vl_interface': 8.794842604080389e-06, 'learning_rate/action_model': 8.794842604080389e-05, 'epoch': 0.51}) +[2026-06-15 21:26:19,700][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.21122179925441742, 'train/grad_norm_pre_clip': 74.63009643554688, 'timing/data': 0.02413183997850865, 'timing/model': 1.12944535195129, 'learning_rate/qwen_vl_interface': 8.7928147542709e-06, 'learning_rate/action_model': 8.7928147542709e-05, 'epoch': 0.51}) +[2026-06-15 21:26:24,205][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.13462020456790924, 'train/grad_norm_pre_clip': 21.177379608154297, 'timing/data': 0.023459438991267234, 'timing/model': 1.1281564650125802, 'learning_rate/qwen_vl_interface': 8.790785448144088e-06, 'learning_rate/action_model': 8.790785448144088e-05, 'epoch': 0.51}) +[2026-06-15 21:26:28,756][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.2803218364715576, 'train/grad_norm_pre_clip': 21.861177444458008, 'timing/data': 0.023317001992836595, 'timing/model': 1.3099726460059173, 'learning_rate/qwen_vl_interface': 8.788754686534126e-06, 'learning_rate/action_model': 8.788754686534127e-05, 'epoch': 0.51}) +[2026-06-15 21:26:33,241][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.25370851159095764, 'train/grad_norm_pre_clip': 15.20865535736084, 'timing/data': 0.023349345021415502, 'timing/model': 1.249096315994393, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 0.51}) +[2026-06-15 21:26:37,613][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.19670578837394714, 'train/grad_norm_pre_clip': 17.79279136657715, 'timing/data': 0.023125206993427128, 'timing/model': 1.138681388052646, 'learning_rate/qwen_vl_interface': 8.784688800204427e-06, 'learning_rate/action_model': 8.784688800204426e-05, 'epoch': 0.51}) +[2026-06-15 21:26:42,147][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.12968869507312775, 'train/grad_norm_pre_clip': 6.268303394317627, 'timing/data': 0.02383640396874398, 'timing/model': 1.1311105659697205, 'learning_rate/qwen_vl_interface': 8.782653677156022e-06, 'learning_rate/action_model': 8.782653677156022e-05, 'epoch': 0.51}) +[2026-06-15 21:26:46,667][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.3679577708244324, 'train/grad_norm_pre_clip': 16.084924697875977, 'timing/data': 0.023397757962811738, 'timing/model': 1.12986947002355, 'learning_rate/qwen_vl_interface': 8.780617101967129e-06, 'learning_rate/action_model': 8.780617101967129e-05, 'epoch': 0.51}) +[2026-06-15 21:26:51,202][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.19565442204475403, 'train/grad_norm_pre_clip': 10.99206256866455, 'timing/data': 0.026346710976213217, 'timing/model': 1.1353219759766944, 'learning_rate/qwen_vl_interface': 8.77857907547491e-06, 'learning_rate/action_model': 8.77857907547491e-05, 'epoch': 0.51}) +[2026-06-15 21:26:55,740][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.18293432891368866, 'train/grad_norm_pre_clip': 14.90255069732666, 'timing/data': 0.023141773010138422, 'timing/model': 1.2967908030259423, 'learning_rate/qwen_vl_interface': 8.776539598517119e-06, 'learning_rate/action_model': 8.776539598517118e-05, 'epoch': 0.51}) +[2026-06-15 21:27:00,248][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.3167855143547058, 'train/grad_norm_pre_clip': 6.816258430480957, 'timing/data': 0.023655873956158757, 'timing/model': 1.2714757090434432, 'learning_rate/qwen_vl_interface': 8.774498671932108e-06, 'learning_rate/action_model': 8.774498671932107e-05, 'epoch': 0.51}) +[2026-06-15 21:27:04,641][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.17282970249652863, 'train/grad_norm_pre_clip': 12.714552879333496, 'timing/data': 0.023126309970393777, 'timing/model': 1.1314028719789349, 'learning_rate/qwen_vl_interface': 8.772456296558824e-06, 'learning_rate/action_model': 8.772456296558824e-05, 'epoch': 0.51}) +[2026-06-15 21:27:09,207][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.21788763999938965, 'train/grad_norm_pre_clip': 8.50584888458252, 'timing/data': 0.023120445024687797, 'timing/model': 1.131239233945962, 'learning_rate/qwen_vl_interface': 8.77041247323681e-06, 'learning_rate/action_model': 8.770412473236811e-05, 'epoch': 0.51}) +[2026-06-15 21:27:13,777][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.16363006830215454, 'train/grad_norm_pre_clip': 6.010021209716797, 'timing/data': 0.024047132988926023, 'timing/model': 1.1298557799891569, 'learning_rate/qwen_vl_interface': 8.76836720280621e-06, 'learning_rate/action_model': 8.768367202806209e-05, 'epoch': 0.51}) +[2026-06-15 21:27:18,336][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.1114315390586853, 'train/grad_norm_pre_clip': 4.948185443878174, 'timing/data': 0.029056742030661553, 'timing/model': 1.131111717026215, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 0.51}) +[2026-06-15 21:27:22,871][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.16852320730686188, 'train/grad_norm_pre_clip': 9.25915241241455, 'timing/data': 0.023042979009915143, 'timing/model': 1.1361556440242566, 'learning_rate/qwen_vl_interface': 8.76427232398276e-06, 'learning_rate/action_model': 8.76427232398276e-05, 'epoch': 0.51}) +[2026-06-15 21:27:27,409][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.07317491620779037, 'train/grad_norm_pre_clip': 3.520613431930542, 'timing/data': 0.023132645001169294, 'timing/model': 1.293802540982142, 'learning_rate/qwen_vl_interface': 8.762222717273167e-06, 'learning_rate/action_model': 8.762222717273167e-05, 'epoch': 0.51}) +[2026-06-15 21:27:31,779][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.18295595049858093, 'train/grad_norm_pre_clip': 7.127509117126465, 'timing/data': 0.02437908702995628, 'timing/model': 1.1307094959774986, 'learning_rate/qwen_vl_interface': 8.76017166682148e-06, 'learning_rate/action_model': 8.760171666821481e-05, 'epoch': 0.51}) +[2026-06-15 21:27:36,321][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.1693277508020401, 'train/grad_norm_pre_clip': 24.362192153930664, 'timing/data': 0.023138483986258507, 'timing/model': 1.148694041010458, 'learning_rate/qwen_vl_interface': 8.758119173470815e-06, 'learning_rate/action_model': 8.758119173470815e-05, 'epoch': 0.52}) +[2026-06-15 21:27:40,762][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.22680936753749847, 'train/grad_norm_pre_clip': 12.972107887268066, 'timing/data': 0.022981146990787238, 'timing/model': 1.1321242160047404, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 0.52}) +[2026-06-15 21:27:45,249][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.23531973361968994, 'train/grad_norm_pre_clip': 7.637470722198486, 'timing/data': 0.024972278042696416, 'timing/model': 1.1296961309853941, 'learning_rate/qwen_vl_interface': 8.75400986144794e-06, 'learning_rate/action_model': 8.75400986144794e-05, 'epoch': 0.52}) +[2026-06-15 21:27:49,747][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.13866078853607178, 'train/grad_norm_pre_clip': 8.337240219116211, 'timing/data': 0.0309114339761436, 'timing/model': 1.1316404650569893, 'learning_rate/qwen_vl_interface': 8.751953044464913e-06, 'learning_rate/action_model': 8.751953044464913e-05, 'epoch': 0.52}) +[2026-06-15 21:27:54,274][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.11904703825712204, 'train/grad_norm_pre_clip': 5.862706661224365, 'timing/data': 0.02346396102802828, 'timing/model': 1.2632350440253504, 'learning_rate/qwen_vl_interface': 8.749894787961272e-06, 'learning_rate/action_model': 8.749894787961271e-05, 'epoch': 0.52}) +[2026-06-15 21:27:58,684][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.07187260687351227, 'train/grad_norm_pre_clip': 3.9200189113616943, 'timing/data': 0.023259873036295176, 'timing/model': 1.1478973010089248, 'learning_rate/qwen_vl_interface': 8.747835092783086e-06, 'learning_rate/action_model': 8.747835092783085e-05, 'epoch': 0.52}) +[2026-06-15 21:28:03,228][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.27998510003089905, 'train/grad_norm_pre_clip': 9.185644149780273, 'timing/data': 0.023354445991571993, 'timing/model': 1.1320015719975345, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 0.52}) +[2026-06-15 21:28:07,787][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.15320372581481934, 'train/grad_norm_pre_clip': 7.0372772216796875, 'timing/data': 0.02309441100805998, 'timing/model': 1.131890273944009, 'learning_rate/qwen_vl_interface': 8.743711389790321e-06, 'learning_rate/action_model': 8.74371138979032e-05, 'epoch': 0.52}) +[2026-06-15 21:28:12,342][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.2528311610221863, 'train/grad_norm_pre_clip': 10.926835060119629, 'timing/data': 0.02442044799681753, 'timing/model': 1.1347659540479071, 'learning_rate/qwen_vl_interface': 8.741647383670839e-06, 'learning_rate/action_model': 8.741647383670839e-05, 'epoch': 0.52}) +[2026-06-15 21:28:16,905][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.06815072894096375, 'train/grad_norm_pre_clip': 3.6164138317108154, 'timing/data': 0.02337350498419255, 'timing/model': 1.300319436006248, 'learning_rate/qwen_vl_interface': 8.739581942267011e-06, 'learning_rate/action_model': 8.739581942267011e-05, 'epoch': 0.52}) +[2026-06-15 21:28:21,305][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.03683764114975929, 'train/grad_norm_pre_clip': 3.124481201171875, 'timing/data': 0.023589959950186312, 'timing/model': 1.136564565997105, 'learning_rate/qwen_vl_interface': 8.737515066427861e-06, 'learning_rate/action_model': 8.73751506642786e-05, 'epoch': 0.52}) +[2026-06-15 21:28:25,848][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.02431255578994751, 'train/grad_norm_pre_clip': 1.545401692390442, 'timing/data': 0.022935642977245152, 'timing/model': 1.1316098920069635, 'learning_rate/qwen_vl_interface': 8.735446757003e-06, 'learning_rate/action_model': 8.735446757003e-05, 'epoch': 0.52}) +[2026-06-15 21:28:30,344][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.020315904170274734, 'train/grad_norm_pre_clip': 1.3537992238998413, 'timing/data': 0.02415682398714125, 'timing/model': 1.1383787909871899, 'learning_rate/qwen_vl_interface': 8.733377014842635e-06, 'learning_rate/action_model': 8.733377014842635e-05, 'epoch': 0.52}) +[2026-06-15 21:28:34,884][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.05210177227854729, 'train/grad_norm_pre_clip': 7.295164585113525, 'timing/data': 0.02426130202366039, 'timing/model': 1.1319970279582776, 'learning_rate/qwen_vl_interface': 8.731305840797557e-06, 'learning_rate/action_model': 8.731305840797558e-05, 'epoch': 0.52}) +[2026-06-15 21:28:39,417][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.05055651441216469, 'train/grad_norm_pre_clip': 5.470596790313721, 'timing/data': 0.023303923022467643, 'timing/model': 1.1316046319552697, 'learning_rate/qwen_vl_interface': 8.72923323571915e-06, 'learning_rate/action_model': 8.72923323571915e-05, 'epoch': 0.52}) +[2026-06-15 21:28:43,967][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.14283372461795807, 'train/grad_norm_pre_clip': 9.916281700134277, 'timing/data': 0.02302467601839453, 'timing/model': 1.2894896999932826, 'learning_rate/qwen_vl_interface': 8.727159200459381e-06, 'learning_rate/action_model': 8.727159200459381e-05, 'epoch': 0.52}) +[2026-06-15 21:28:48,472][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.21066150069236755, 'train/grad_norm_pre_clip': 2.967731237411499, 'timing/data': 0.023621693020686507, 'timing/model': 1.2530372919864021, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 0.52}) +[2026-06-15 21:28:52,853][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.019704949110746384, 'train/grad_norm_pre_clip': 3.4518253803253174, 'timing/data': 0.02386893203947693, 'timing/model': 1.1308857789845206, 'learning_rate/qwen_vl_interface': 8.723006842806575e-06, 'learning_rate/action_model': 8.723006842806576e-05, 'epoch': 0.52}) +[2026-06-15 21:28:57,408][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.46939897537231445, 'train/grad_norm_pre_clip': 21.096752166748047, 'timing/data': 0.025284854986239225, 'timing/model': 1.1428223160328344, 'learning_rate/qwen_vl_interface': 8.720928522120419e-06, 'learning_rate/action_model': 8.720928522120418e-05, 'epoch': 0.52}) +[2026-06-15 21:29:01,958][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.22186210751533508, 'train/grad_norm_pre_clip': 12.254656791687012, 'timing/data': 0.023715463001281023, 'timing/model': 1.1344024679856375, 'learning_rate/qwen_vl_interface': 8.718848774666655e-06, 'learning_rate/action_model': 8.718848774666655e-05, 'epoch': 0.52}) +[2026-06-15 21:29:06,496][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.5112432837486267, 'train/grad_norm_pre_clip': 25.79267120361328, 'timing/data': 0.02308769803494215, 'timing/model': 1.1325710419914685, 'learning_rate/qwen_vl_interface': 8.716767601300189e-06, 'learning_rate/action_model': 8.716767601300188e-05, 'epoch': 0.52}) +[2026-06-15 21:29:11,054][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.15442948043346405, 'train/grad_norm_pre_clip': 22.289844512939453, 'timing/data': 0.02338327601319179, 'timing/model': 1.2977090129861608, 'learning_rate/qwen_vl_interface': 8.714685002876514e-06, 'learning_rate/action_model': 8.714685002876513e-05, 'epoch': 0.52}) +[2026-06-15 21:29:15,594][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.22204150259494781, 'train/grad_norm_pre_clip': 15.649243354797363, 'timing/data': 0.02277953695738688, 'timing/model': 1.2963846839847974, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.52}) +[2026-06-15 21:29:19,983][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.05502321943640709, 'train/grad_norm_pre_clip': 11.272104263305664, 'timing/data': 0.02377868996700272, 'timing/model': 1.138186156982556, 'learning_rate/qwen_vl_interface': 8.710515534282428e-06, 'learning_rate/action_model': 8.710515534282428e-05, 'epoch': 0.52}) +[2026-06-15 21:29:24,528][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.13519540429115295, 'train/grad_norm_pre_clip': 4.846628665924072, 'timing/data': 0.023098118021152914, 'timing/model': 1.1323513250099495, 'learning_rate/qwen_vl_interface': 8.708428665825931e-06, 'learning_rate/action_model': 8.708428665825931e-05, 'epoch': 0.53}) +[2026-06-15 21:29:29,108][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.044652342796325684, 'train/grad_norm_pre_clip': 12.921072959899902, 'timing/data': 0.022854459995869547, 'timing/model': 1.132922260032501, 'learning_rate/qwen_vl_interface': 8.706340375740045e-06, 'learning_rate/action_model': 8.706340375740045e-05, 'epoch': 0.53}) +[2026-06-15 21:29:33,633][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.1305719017982483, 'train/grad_norm_pre_clip': 11.199017524719238, 'timing/data': 0.02321387000847608, 'timing/model': 1.1317668810370378, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 0.53}) +[2026-06-15 21:29:38,118][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.6982231736183167, 'train/grad_norm_pre_clip': 14.07681655883789, 'timing/data': 0.02306787099223584, 'timing/model': 1.133522720017936, 'learning_rate/qwen_vl_interface': 8.70215953411436e-06, 'learning_rate/action_model': 8.70215953411436e-05, 'epoch': 0.53}) +[2026-06-15 21:29:42,660][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.027001814916729927, 'train/grad_norm_pre_clip': 4.719782829284668, 'timing/data': 0.023307369032409042, 'timing/model': 1.2922799779917113, 'learning_rate/qwen_vl_interface': 8.70006698429315e-06, 'learning_rate/action_model': 8.700066984293149e-05, 'epoch': 0.53}) +[2026-06-15 21:29:47,051][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.10840247571468353, 'train/grad_norm_pre_clip': 14.709266662597656, 'timing/data': 0.02334852598141879, 'timing/model': 1.135864672018215, 'learning_rate/qwen_vl_interface': 8.697973016279722e-06, 'learning_rate/action_model': 8.697973016279722e-05, 'epoch': 0.53}) +[2026-06-15 21:29:51,605][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.022113636136054993, 'train/grad_norm_pre_clip': 6.895388603210449, 'timing/data': 0.023707389016635716, 'timing/model': 1.1352148620062508, 'learning_rate/qwen_vl_interface': 8.695877630934827e-06, 'learning_rate/action_model': 8.695877630934826e-05, 'epoch': 0.53}) +[2026-06-15 21:29:56,032][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.08489613234996796, 'train/grad_norm_pre_clip': 10.61622142791748, 'timing/data': 0.023633536999113858, 'timing/model': 1.13422210898716, 'learning_rate/qwen_vl_interface': 8.693780829119801e-06, 'learning_rate/action_model': 8.693780829119801e-05, 'epoch': 0.53}) +[2026-06-15 21:30:00,467][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.20638900995254517, 'train/grad_norm_pre_clip': 12.513984680175781, 'timing/data': 0.0233685519779101, 'timing/model': 1.1311302380054258, 'learning_rate/qwen_vl_interface': 8.69168261169656e-06, 'learning_rate/action_model': 8.691682611696559e-05, 'epoch': 0.53}) +[2026-06-15 21:30:04,926][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.35869869589805603, 'train/grad_norm_pre_clip': 19.310903549194336, 'timing/data': 0.02393135998863727, 'timing/model': 1.1328778130118735, 'learning_rate/qwen_vl_interface': 8.689582979527599e-06, 'learning_rate/action_model': 8.689582979527599e-05, 'epoch': 0.53}) +[2026-06-15 21:30:09,468][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.3939186930656433, 'train/grad_norm_pre_clip': 23.750532150268555, 'timing/data': 0.02347437699791044, 'timing/model': 1.2716089290333912, 'learning_rate/qwen_vl_interface': 8.687481933476e-06, 'learning_rate/action_model': 8.687481933476e-05, 'epoch': 0.53}) +[2026-06-15 21:30:13,880][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.16443505883216858, 'train/grad_norm_pre_clip': 6.2005696296691895, 'timing/data': 0.023415728996042162, 'timing/model': 1.1404163489933126, 'learning_rate/qwen_vl_interface': 8.685379474405423e-06, 'learning_rate/action_model': 8.685379474405423e-05, 'epoch': 0.53}) +[2026-06-15 21:30:18,448][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.013133443892002106, 'train/grad_norm_pre_clip': 0.5100030303001404, 'timing/data': 0.02285707596456632, 'timing/model': 1.143863266974222, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.53}) +[2026-06-15 21:30:23,015][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.1148529052734375, 'train/grad_norm_pre_clip': 10.138874053955078, 'timing/data': 0.023540574009530246, 'timing/model': 1.1437043059850112, 'learning_rate/qwen_vl_interface': 8.681170320664883e-06, 'learning_rate/action_model': 8.681170320664883e-05, 'epoch': 0.53}) +[2026-06-15 21:30:27,585][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.30469179153442383, 'train/grad_norm_pre_clip': 15.59327220916748, 'timing/data': 0.023550842015538365, 'timing/model': 1.1346651029889472, 'learning_rate/qwen_vl_interface': 8.679063627725143e-06, 'learning_rate/action_model': 8.679063627725144e-05, 'epoch': 0.53}) +[2026-06-15 21:30:32,138][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.09308107197284698, 'train/grad_norm_pre_clip': 2.4890365600585938, 'timing/data': 0.023467694001737982, 'timing/model': 1.2963726519956253, 'learning_rate/qwen_vl_interface': 8.676955525226876e-06, 'learning_rate/action_model': 8.676955525226875e-05, 'epoch': 0.53}) +[2026-06-15 21:30:36,521][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.1565217822790146, 'train/grad_norm_pre_clip': 11.189579010009766, 'timing/data': 0.02312834101030603, 'timing/model': 1.1336772340000607, 'learning_rate/qwen_vl_interface': 8.674846014036636e-06, 'learning_rate/action_model': 8.674846014036636e-05, 'epoch': 0.53}) +[2026-06-15 21:30:41,075][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.12727338075637817, 'train/grad_norm_pre_clip': 1.719437837600708, 'timing/data': 0.02317659603431821, 'timing/model': 1.1373835270060226, 'learning_rate/qwen_vl_interface': 8.672735095021569e-06, 'learning_rate/action_model': 8.672735095021569e-05, 'epoch': 0.53}) +[2026-06-15 21:30:45,574][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.22194483876228333, 'train/grad_norm_pre_clip': 10.033366203308105, 'timing/data': 0.022951471037231386, 'timing/model': 1.1470522279851139, 'learning_rate/qwen_vl_interface': 8.670622769049393e-06, 'learning_rate/action_model': 8.670622769049392e-05, 'epoch': 0.53}) +[2026-06-15 21:30:50,136][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.08958951383829117, 'train/grad_norm_pre_clip': 12.068278312683105, 'timing/data': 0.023051358002703637, 'timing/model': 1.1356265470385551, 'learning_rate/qwen_vl_interface': 8.668509036988406e-06, 'learning_rate/action_model': 8.668509036988405e-05, 'epoch': 0.53}) +[2026-06-15 21:30:54,667][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.19037483632564545, 'train/grad_norm_pre_clip': 18.258373260498047, 'timing/data': 0.02351002802606672, 'timing/model': 1.1347254990250804, 'learning_rate/qwen_vl_interface': 8.66639389970748e-06, 'learning_rate/action_model': 8.666393899707479e-05, 'epoch': 0.53}) +[2026-06-15 21:30:59,199][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.25332218408584595, 'train/grad_norm_pre_clip': 9.933051109313965, 'timing/data': 0.023216189991217107, 'timing/model': 1.2884302259772085, 'learning_rate/qwen_vl_interface': 8.66427735807607e-06, 'learning_rate/action_model': 8.664277358076069e-05, 'epoch': 0.53}) +[2026-06-15 21:35:18,821][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.0987522304058075, 'train/grad_norm_pre_clip': 3.0323522090911865, 'eval/loss': 0.06670200824737549, 'eval/flap_precision': 0.9304084720121029, 'eval/flap_recall': 0.9490740740740741, 'eval/flap_f1': 0.9396485867074102, 'eval/flap_rate_model': 0.1836111111111111, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9396485867074102, 'timing/data': 0.0232629650272429, 'timing/model': 1.2517543609719723, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 0.53}) +[2026-06-15 21:35:18,822][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1300_state +[2026-06-15 21:35:18,824][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 21:35:22,396][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1300_state/model.safetensors +[2026-06-15 21:35:30,103][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1300_state/optimizer.bin +[2026-06-15 21:35:30,103][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1300_state/sampler.bin +[2026-06-15 21:35:30,103][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1300_state/sampler_1.bin +[2026-06-15 21:35:30,104][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1300_state/random_states_0.pkl +[2026-06-15 21:35:30,158][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-06-15 21:35:30,159][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 21:35:30,179][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 21:35:34,641][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.2551557123661041, 'train/grad_norm_pre_clip': 10.975712776184082, 'timing/data': 0.023384261992760003, 'timing/model': 1.274244951957371, 'learning_rate/qwen_vl_interface': 8.660040065242499e-06, 'learning_rate/action_model': 8.660040065242498e-05, 'epoch': 0.53}) +[2026-06-15 21:35:38,968][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.0531034991145134, 'train/grad_norm_pre_clip': 9.41042423248291, 'timing/data': 0.023324464971665293, 'timing/model': 1.125269837968517, 'learning_rate/qwen_vl_interface': 8.657919315782127e-06, 'learning_rate/action_model': 8.657919315782127e-05, 'epoch': 0.53}) +[2026-06-15 21:35:43,445][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.19926688075065613, 'train/grad_norm_pre_clip': 10.22059440612793, 'timing/data': 0.023098614008631557, 'timing/model': 1.1228451700299047, 'learning_rate/qwen_vl_interface': 8.655797165454853e-06, 'learning_rate/action_model': 8.655797165454853e-05, 'epoch': 0.54}) +[2026-06-15 21:35:47,925][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.239053413271904, 'train/grad_norm_pre_clip': 15.428177833557129, 'timing/data': 0.02405074395937845, 'timing/model': 1.1182012759963982, 'learning_rate/qwen_vl_interface': 8.653673615133013e-06, 'learning_rate/action_model': 8.653673615133013e-05, 'epoch': 0.54}) +[2026-06-15 21:35:52,430][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.1784677803516388, 'train/grad_norm_pre_clip': 11.23066234588623, 'timing/data': 0.023094295989722013, 'timing/model': 1.1189864890184253, 'learning_rate/qwen_vl_interface': 8.651548665689519e-06, 'learning_rate/action_model': 8.651548665689518e-05, 'epoch': 0.54}) +[2026-06-15 21:35:56,898][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.09644925594329834, 'train/grad_norm_pre_clip': 3.75622820854187, 'timing/data': 0.02358609699876979, 'timing/model': 1.1220431239926256, 'learning_rate/qwen_vl_interface': 8.649422317997855e-06, 'learning_rate/action_model': 8.649422317997856e-05, 'epoch': 0.54}) +[2026-06-15 21:36:01,412][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.2251196801662445, 'train/grad_norm_pre_clip': 5.064657211303711, 'timing/data': 0.028485052986070514, 'timing/model': 1.2889756949734874, 'learning_rate/qwen_vl_interface': 8.647294572932082e-06, 'learning_rate/action_model': 8.647294572932082e-05, 'epoch': 0.54}) +[2026-06-15 21:36:05,789][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.19085140526294708, 'train/grad_norm_pre_clip': 11.439410209655762, 'timing/data': 0.023129259003326297, 'timing/model': 1.1334674709942192, 'learning_rate/qwen_vl_interface': 8.64516543136684e-06, 'learning_rate/action_model': 8.64516543136684e-05, 'epoch': 0.54}) +[2026-06-15 21:36:10,303][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.29265570640563965, 'train/grad_norm_pre_clip': 44.98744583129883, 'timing/data': 0.023157430987339467, 'timing/model': 1.126218379009515, 'learning_rate/qwen_vl_interface': 8.643034894177335e-06, 'learning_rate/action_model': 8.643034894177335e-05, 'epoch': 0.54}) +[2026-06-15 21:36:14,801][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.189022034406662, 'train/grad_norm_pre_clip': 6.33595085144043, 'timing/data': 0.02366492000874132, 'timing/model': 1.124082603957504, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 0.54}) +[2026-06-15 21:36:19,314][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.20052999258041382, 'train/grad_norm_pre_clip': 14.914153099060059, 'timing/data': 0.024047449987847358, 'timing/model': 1.1271278909989633, 'learning_rate/qwen_vl_interface': 8.638769636429245e-06, 'learning_rate/action_model': 8.638769636429245e-05, 'epoch': 0.54}) +[2026-06-15 21:36:23,859][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.21360109746456146, 'train/grad_norm_pre_clip': 10.15897274017334, 'timing/data': 0.02357604302233085, 'timing/model': 1.137709311034996, 'learning_rate/qwen_vl_interface': 8.636634917623946e-06, 'learning_rate/action_model': 8.636634917623944e-05, 'epoch': 0.54}) +[2026-06-15 21:36:28,393][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.18314038217067719, 'train/grad_norm_pre_clip': 4.1223273277282715, 'timing/data': 0.02332536003086716, 'timing/model': 1.29056811099872, 'learning_rate/qwen_vl_interface': 8.634498806700955e-06, 'learning_rate/action_model': 8.634498806700954e-05, 'epoch': 0.54}) +[2026-06-15 21:36:32,769][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.1105143204331398, 'train/grad_norm_pre_clip': 4.259667873382568, 'timing/data': 0.023997862997930497, 'timing/model': 1.1273854059982114, 'learning_rate/qwen_vl_interface': 8.63236130453835e-06, 'learning_rate/action_model': 8.63236130453835e-05, 'epoch': 0.54}) +[2026-06-15 21:36:37,303][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.11514168977737427, 'train/grad_norm_pre_clip': 6.915484428405762, 'timing/data': 0.023011256009340286, 'timing/model': 1.1279440950020216, 'learning_rate/qwen_vl_interface': 8.630222412014777e-06, 'learning_rate/action_model': 8.630222412014776e-05, 'epoch': 0.54}) +[2026-06-15 21:36:41,839][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.1608269363641739, 'train/grad_norm_pre_clip': 3.852663516998291, 'timing/data': 0.023032222990877926, 'timing/model': 1.1290406709886156, 'learning_rate/qwen_vl_interface': 8.628082130009451e-06, 'learning_rate/action_model': 8.628082130009451e-05, 'epoch': 0.54}) +[2026-06-15 21:36:46,359][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.05979321151971817, 'train/grad_norm_pre_clip': 3.714047431945801, 'timing/data': 0.023464418947696686, 'timing/model': 1.1288763930206187, 'learning_rate/qwen_vl_interface': 8.62594045940216e-06, 'learning_rate/action_model': 8.62594045940216e-05, 'epoch': 0.54}) +[2026-06-15 21:36:50,905][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.15910281240940094, 'train/grad_norm_pre_clip': 6.424247741699219, 'timing/data': 0.023062204010784626, 'timing/model': 1.2883911230019294, 'learning_rate/qwen_vl_interface': 8.623797401073272e-06, 'learning_rate/action_model': 8.623797401073271e-05, 'epoch': 0.54}) +[2026-06-15 21:36:55,423][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.23184682428836823, 'train/grad_norm_pre_clip': 14.66680908203125, 'timing/data': 0.02322422800352797, 'timing/model': 1.2840471219969913, 'learning_rate/qwen_vl_interface': 8.62165295590371e-06, 'learning_rate/action_model': 8.621652955903709e-05, 'epoch': 0.54}) +[2026-06-15 21:36:59,803][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.16695213317871094, 'train/grad_norm_pre_clip': 11.199155807495117, 'timing/data': 0.022942707990296185, 'timing/model': 1.1429593240027316, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 0.54}) +[2026-06-15 21:37:04,338][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.07830218225717545, 'train/grad_norm_pre_clip': 7.03666353225708, 'timing/data': 0.022988372016698122, 'timing/model': 1.1369502479792573, 'learning_rate/qwen_vl_interface': 8.61735990856914e-06, 'learning_rate/action_model': 8.617359908569139e-05, 'epoch': 0.54}) +[2026-06-15 21:37:08,874][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.1475483477115631, 'train/grad_norm_pre_clip': 6.952286243438721, 'timing/data': 0.02358681702753529, 'timing/model': 1.1358625919674523, 'learning_rate/qwen_vl_interface': 8.615211308168842e-06, 'learning_rate/action_model': 8.61521130816884e-05, 'epoch': 0.54}) +[2026-06-15 21:37:13,413][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.2541826665401459, 'train/grad_norm_pre_clip': 12.183682441711426, 'timing/data': 0.023276116990018636, 'timing/model': 1.1324650879832916, 'learning_rate/qwen_vl_interface': 8.613061324457289e-06, 'learning_rate/action_model': 8.613061324457288e-05, 'epoch': 0.54}) +[2026-06-15 21:37:17,908][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.29880475997924805, 'train/grad_norm_pre_clip': 16.303197860717773, 'timing/data': 0.023986704007256776, 'timing/model': 1.1359170919749886, 'learning_rate/qwen_vl_interface': 8.610909958318262e-06, 'learning_rate/action_model': 8.610909958318261e-05, 'epoch': 0.54}) +[2026-06-15 21:37:22,460][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.27837973833084106, 'train/grad_norm_pre_clip': 7.162642955780029, 'timing/data': 0.022792204981669784, 'timing/model': 1.2958711679675616, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.54}) +[2026-06-15 21:37:26,845][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.09181488305330276, 'train/grad_norm_pre_clip': 7.358994483947754, 'timing/data': 0.023006264003925025, 'timing/model': 1.1301542249857448, 'learning_rate/qwen_vl_interface': 8.606603082295724e-06, 'learning_rate/action_model': 8.606603082295724e-05, 'epoch': 0.54}) +[2026-06-15 21:37:31,373][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.13652366399765015, 'train/grad_norm_pre_clip': 11.500085830688477, 'timing/data': 0.02315825899131596, 'timing/model': 1.1295512979850173, 'learning_rate/qwen_vl_interface': 8.60444757418261e-06, 'learning_rate/action_model': 8.604447574182609e-05, 'epoch': 0.55}) +[2026-06-15 21:37:35,803][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.2809004783630371, 'train/grad_norm_pre_clip': 18.932958602905273, 'timing/data': 0.022466999012976885, 'timing/model': 1.128378622001037, 'learning_rate/qwen_vl_interface': 8.602290687182805e-06, 'learning_rate/action_model': 8.602290687182806e-05, 'epoch': 0.55}) +[2026-06-15 21:37:40,248][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.10020631551742554, 'train/grad_norm_pre_clip': 9.04311752319336, 'timing/data': 0.022878756979480386, 'timing/model': 1.1364382389583625, 'learning_rate/qwen_vl_interface': 8.600132422182929e-06, 'learning_rate/action_model': 8.600132422182928e-05, 'epoch': 0.55}) +[2026-06-15 21:37:44,725][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.1908256858587265, 'train/grad_norm_pre_clip': 10.473618507385254, 'timing/data': 0.02312882395926863, 'timing/model': 1.137032511003781, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 0.55}) +[2026-06-15 21:37:49,221][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.046066366136074066, 'train/grad_norm_pre_clip': 6.134218692779541, 'timing/data': 0.022844448045361787, 'timing/model': 1.2548537299735472, 'learning_rate/qwen_vl_interface': 8.595811761732239e-06, 'learning_rate/action_model': 8.595811761732238e-05, 'epoch': 0.55}) +[2026-06-15 21:37:53,593][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.11567695438861847, 'train/grad_norm_pre_clip': 13.373895645141602, 'timing/data': 0.022910561005119234, 'timing/model': 1.1316654559923336, 'learning_rate/qwen_vl_interface': 8.593649368057489e-06, 'learning_rate/action_model': 8.59364936805749e-05, 'epoch': 0.55}) +[2026-06-15 21:37:58,138][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.012216062285006046, 'train/grad_norm_pre_clip': 0.8870137333869934, 'timing/data': 0.02290456398623064, 'timing/model': 1.1407083150115795, 'learning_rate/qwen_vl_interface': 8.591485599934786e-06, 'learning_rate/action_model': 8.591485599934785e-05, 'epoch': 0.55}) +[2026-06-15 21:38:02,740][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.02092583477497101, 'train/grad_norm_pre_clip': 6.146676063537598, 'timing/data': 0.02359578898176551, 'timing/model': 1.1493661620188504, 'learning_rate/qwen_vl_interface': 8.58932045825357e-06, 'learning_rate/action_model': 8.589320458253569e-05, 'epoch': 0.55}) +[2026-06-15 21:38:07,279][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.13425157964229584, 'train/grad_norm_pre_clip': 9.462494850158691, 'timing/data': 0.02289507701061666, 'timing/model': 1.1370221399702132, 'learning_rate/qwen_vl_interface': 8.58715394390385e-06, 'learning_rate/action_model': 8.58715394390385e-05, 'epoch': 0.55}) +[2026-06-15 21:38:11,826][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.5936106443405151, 'train/grad_norm_pre_clip': 21.639802932739258, 'timing/data': 0.022694555984344333, 'timing/model': 1.2908555300091393, 'learning_rate/qwen_vl_interface': 8.5849860577762e-06, 'learning_rate/action_model': 8.5849860577762e-05, 'epoch': 0.55}) +[2026-06-15 21:38:16,211][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.42924997210502625, 'train/grad_norm_pre_clip': 9.48551082611084, 'timing/data': 0.022980983019806445, 'timing/model': 1.143875627021771, 'learning_rate/qwen_vl_interface': 8.582816800761756e-06, 'learning_rate/action_model': 8.582816800761755e-05, 'epoch': 0.55}) +[2026-06-15 21:38:20,740][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.24126631021499634, 'train/grad_norm_pre_clip': 9.686629295349121, 'timing/data': 0.02365803101565689, 'timing/model': 1.132733887003269, 'learning_rate/qwen_vl_interface': 8.580646173752214e-06, 'learning_rate/action_model': 8.580646173752213e-05, 'epoch': 0.55}) +[2026-06-15 21:38:25,242][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.0749485194683075, 'train/grad_norm_pre_clip': 7.242636203765869, 'timing/data': 0.022692530998028815, 'timing/model': 1.139567904989235, 'learning_rate/qwen_vl_interface': 8.578474177639841e-06, 'learning_rate/action_model': 8.578474177639842e-05, 'epoch': 0.55}) +[2026-06-15 21:38:29,789][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.19703343510627747, 'train/grad_norm_pre_clip': 15.397950172424316, 'timing/data': 0.02345853397855535, 'timing/model': 1.1347319270134903, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 0.55}) +[2026-06-15 21:38:34,318][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.2286868393421173, 'train/grad_norm_pre_clip': 6.340252876281738, 'timing/data': 0.024092591018415987, 'timing/model': 1.131583887967281, 'learning_rate/qwen_vl_interface': 8.574126081678463e-06, 'learning_rate/action_model': 8.574126081678462e-05, 'epoch': 0.55}) +[2026-06-15 21:38:38,854][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.21587339043617249, 'train/grad_norm_pre_clip': 7.510183334350586, 'timing/data': 0.023144054983276874, 'timing/model': 1.2850490559940226, 'learning_rate/qwen_vl_interface': 8.571949983616796e-06, 'learning_rate/action_model': 8.571949983616795e-05, 'epoch': 0.55}) +[2026-06-15 21:38:43,364][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.052444811910390854, 'train/grad_norm_pre_clip': 8.139827728271484, 'timing/data': 0.023374594980850816, 'timing/model': 1.26151680998737, 'learning_rate/qwen_vl_interface': 8.569772520026971e-06, 'learning_rate/action_model': 8.569772520026971e-05, 'epoch': 0.55}) +[2026-06-15 21:38:47,744][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.13353416323661804, 'train/grad_norm_pre_clip': 12.737403869628906, 'timing/data': 0.023057049023918808, 'timing/model': 1.133003334980458, 'learning_rate/qwen_vl_interface': 8.567593691804061e-06, 'learning_rate/action_model': 8.567593691804061e-05, 'epoch': 0.55}) +[2026-06-15 21:38:52,286][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.07447658479213715, 'train/grad_norm_pre_clip': 10.376310348510742, 'timing/data': 0.023085106979124248, 'timing/model': 1.1327998669585213, 'learning_rate/qwen_vl_interface': 8.565413499843703e-06, 'learning_rate/action_model': 8.565413499843703e-05, 'epoch': 0.55}) +[2026-06-15 21:38:56,840][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.02583250030875206, 'train/grad_norm_pre_clip': 1.5332280397415161, 'timing/data': 0.02326495898887515, 'timing/model': 1.1314296529744752, 'learning_rate/qwen_vl_interface': 8.563231945042087e-06, 'learning_rate/action_model': 8.563231945042088e-05, 'epoch': 0.55}) +[2026-06-15 21:39:01,391][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.21825356781482697, 'train/grad_norm_pre_clip': 9.876352310180664, 'timing/data': 0.023254434985574335, 'timing/model': 1.1341228560195304, 'learning_rate/qwen_vl_interface': 8.56104902829597e-06, 'learning_rate/action_model': 8.56104902829597e-05, 'epoch': 0.55}) +[2026-06-15 21:39:05,950][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.36041906476020813, 'train/grad_norm_pre_clip': 11.370542526245117, 'timing/data': 0.02348163799615577, 'timing/model': 1.3116390549694188, 'learning_rate/qwen_vl_interface': 8.558864750502669e-06, 'learning_rate/action_model': 8.558864750502668e-05, 'epoch': 0.55}) +[2026-06-15 21:39:10,507][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.362032949924469, 'train/grad_norm_pre_clip': 16.62076187133789, 'timing/data': 0.02362013002857566, 'timing/model': 1.290446739003528, 'learning_rate/qwen_vl_interface': 8.556679112560055e-06, 'learning_rate/action_model': 8.556679112560054e-05, 'epoch': 0.55}) +[2026-06-15 21:39:14,894][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.29368895292282104, 'train/grad_norm_pre_clip': 8.781892776489258, 'timing/data': 0.023091797018423676, 'timing/model': 1.135334359947592, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 0.55}) +[2026-06-15 21:39:19,436][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.18250484764575958, 'train/grad_norm_pre_clip': 8.311928749084473, 'timing/data': 0.023628336028195918, 'timing/model': 1.132058632967528, 'learning_rate/qwen_vl_interface': 8.55230375982118e-06, 'learning_rate/action_model': 8.55230375982118e-05, 'epoch': 0.55}) +[2026-06-15 21:39:23,976][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.16229248046875, 'train/grad_norm_pre_clip': 10.592377662658691, 'timing/data': 0.02335533400764689, 'timing/model': 1.1316377189941704, 'learning_rate/qwen_vl_interface': 8.550114046823463e-06, 'learning_rate/action_model': 8.550114046823463e-05, 'epoch': 0.56}) +[2026-06-15 21:39:28,516][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.06435326486825943, 'train/grad_norm_pre_clip': 2.576535940170288, 'timing/data': 0.023694564995821565, 'timing/model': 1.1324603399843909, 'learning_rate/qwen_vl_interface': 8.547922977273518e-06, 'learning_rate/action_model': 8.547922977273518e-05, 'epoch': 0.56}) +[2026-06-15 21:39:33,015][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.09007958322763443, 'train/grad_norm_pre_clip': 3.449437141418457, 'timing/data': 0.02330312301637605, 'timing/model': 1.1338788439752534, 'learning_rate/qwen_vl_interface': 8.54573055207201e-06, 'learning_rate/action_model': 8.545730552072008e-05, 'epoch': 0.56}) +[2026-06-15 21:39:37,577][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.09846706688404083, 'train/grad_norm_pre_clip': 5.720550060272217, 'timing/data': 0.023440369986929, 'timing/model': 1.2958437759662047, 'learning_rate/qwen_vl_interface': 8.543536772120161e-06, 'learning_rate/action_model': 8.54353677212016e-05, 'epoch': 0.56}) +[2026-06-15 21:39:41,975][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.2837526202201843, 'train/grad_norm_pre_clip': 8.417683601379395, 'timing/data': 0.02386843296699226, 'timing/model': 1.1471301940036938, 'learning_rate/qwen_vl_interface': 8.541341638319754e-06, 'learning_rate/action_model': 8.541341638319754e-05, 'epoch': 0.56}) +[2026-06-15 21:39:46,532][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.05386301875114441, 'train/grad_norm_pre_clip': 10.373143196105957, 'timing/data': 0.023710807028692216, 'timing/model': 1.136460647045169, 'learning_rate/qwen_vl_interface': 8.539145151573126e-06, 'learning_rate/action_model': 8.539145151573125e-05, 'epoch': 0.56}) +[2026-06-15 21:39:50,979][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.0925726369023323, 'train/grad_norm_pre_clip': 7.632320404052734, 'timing/data': 0.02317411103285849, 'timing/model': 1.1348801080021076, 'learning_rate/qwen_vl_interface': 8.536947312783167e-06, 'learning_rate/action_model': 8.536947312783166e-05, 'epoch': 0.56}) +[2026-06-15 21:39:55,448][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.04170544445514679, 'train/grad_norm_pre_clip': 2.6151390075683594, 'timing/data': 0.026006972009781748, 'timing/model': 1.1383925179834478, 'learning_rate/qwen_vl_interface': 8.534748122853326e-06, 'learning_rate/action_model': 8.534748122853325e-05, 'epoch': 0.56}) +[2026-06-15 21:39:59,940][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.031536854803562164, 'train/grad_norm_pre_clip': 2.1873650550842285, 'timing/data': 0.023036855971440673, 'timing/model': 1.1370777660049498, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.56}) +[2026-06-15 21:40:04,482][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.082139752805233, 'train/grad_norm_pre_clip': 7.363587379455566, 'timing/data': 0.023568322998471558, 'timing/model': 1.267478282970842, 'learning_rate/qwen_vl_interface': 8.530345693190573e-06, 'learning_rate/action_model': 8.530345693190572e-05, 'epoch': 0.56}) +[2026-06-15 21:40:08,895][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.32476213574409485, 'train/grad_norm_pre_clip': 14.870030403137207, 'timing/data': 0.02317868499085307, 'timing/model': 1.142908214009367, 'learning_rate/qwen_vl_interface': 8.528142455267328e-06, 'learning_rate/action_model': 8.528142455267329e-05, 'epoch': 0.56}) +[2026-06-15 21:40:13,488][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.3391049802303314, 'train/grad_norm_pre_clip': 14.493489265441895, 'timing/data': 0.023640794039238244, 'timing/model': 1.1374170379713178, 'learning_rate/qwen_vl_interface': 8.52593786982355e-06, 'learning_rate/action_model': 8.525937869823551e-05, 'epoch': 0.56}) +[2026-06-15 21:40:18,063][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.3637996315956116, 'train/grad_norm_pre_clip': 14.464296340942383, 'timing/data': 0.024063249002210796, 'timing/model': 1.133833845029585, 'learning_rate/qwen_vl_interface': 8.523731937765456e-06, 'learning_rate/action_model': 8.523731937765456e-05, 'epoch': 0.56}) +[2026-06-15 21:40:22,601][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.019612206146121025, 'train/grad_norm_pre_clip': 1.5602205991744995, 'timing/data': 0.02342936396598816, 'timing/model': 1.1330019640154205, 'learning_rate/qwen_vl_interface': 8.521524659999823e-06, 'learning_rate/action_model': 8.521524659999822e-05, 'epoch': 0.56}) +[2026-06-15 21:40:27,149][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.12456799298524857, 'train/grad_norm_pre_clip': 11.55353832244873, 'timing/data': 0.023513608030043542, 'timing/model': 1.2888519060215913, 'learning_rate/qwen_vl_interface': 8.519316037433976e-06, 'learning_rate/action_model': 8.519316037433977e-05, 'epoch': 0.56}) +[2026-06-15 21:40:31,555][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.3259449601173401, 'train/grad_norm_pre_clip': 18.935914993286133, 'timing/data': 0.023354822013061494, 'timing/model': 1.1355944339884445, 'learning_rate/qwen_vl_interface': 8.517106070975801e-06, 'learning_rate/action_model': 8.5171060709758e-05, 'epoch': 0.56}) +[2026-06-15 21:40:36,153][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.17569497227668762, 'train/grad_norm_pre_clip': 14.340493202209473, 'timing/data': 0.02358497300883755, 'timing/model': 1.150294458027929, 'learning_rate/qwen_vl_interface': 8.514894761533729e-06, 'learning_rate/action_model': 8.514894761533728e-05, 'epoch': 0.56}) +[2026-06-15 21:40:40,659][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.09646515548229218, 'train/grad_norm_pre_clip': 10.701369285583496, 'timing/data': 0.023371043032966554, 'timing/model': 1.1465646789874882, 'learning_rate/qwen_vl_interface': 8.512682110016747e-06, 'learning_rate/action_model': 8.512682110016747e-05, 'epoch': 0.56}) +[2026-06-15 21:40:45,220][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.3053489625453949, 'train/grad_norm_pre_clip': 15.627350807189941, 'timing/data': 0.02340977796120569, 'timing/model': 1.135323511029128, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 0.56}) +[2026-06-15 21:40:49,783][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.03003711812198162, 'train/grad_norm_pre_clip': 8.660906791687012, 'timing/data': 0.02369840198662132, 'timing/model': 1.138589026988484, 'learning_rate/qwen_vl_interface': 8.50825278439675e-06, 'learning_rate/action_model': 8.508252784396749e-05, 'epoch': 0.56}) +[2026-06-15 21:40:54,353][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.09672223776578903, 'train/grad_norm_pre_clip': 6.851855754852295, 'timing/data': 0.02198088902514428, 'timing/model': 1.3067858690046705, 'learning_rate/qwen_vl_interface': 8.506036112114465e-06, 'learning_rate/action_model': 8.506036112114464e-05, 'epoch': 0.56}) +[2026-06-15 21:40:58,880][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.06285630166530609, 'train/grad_norm_pre_clip': 11.935954093933105, 'timing/data': 0.023174551024567336, 'timing/model': 1.2484230059781112, 'learning_rate/qwen_vl_interface': 8.503818101398725e-06, 'learning_rate/action_model': 8.503818101398725e-05, 'epoch': 0.56}) +[2026-06-15 21:41:03,279][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.23200832307338715, 'train/grad_norm_pre_clip': 11.051115989685059, 'timing/data': 0.023474352026823908, 'timing/model': 1.141499819001183, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.56}) +[2026-06-15 21:41:07,830][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.24807490408420563, 'train/grad_norm_pre_clip': 6.952451705932617, 'timing/data': 0.02349760598735884, 'timing/model': 1.1348304039565846, 'learning_rate/qwen_vl_interface': 8.499378068314395e-06, 'learning_rate/action_model': 8.499378068314395e-05, 'epoch': 0.56}) +[2026-06-15 21:41:12,392][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.2723468244075775, 'train/grad_norm_pre_clip': 15.518139839172363, 'timing/data': 0.023362006002571434, 'timing/model': 1.132561734993942, 'learning_rate/qwen_vl_interface': 8.497156047770932e-06, 'learning_rate/action_model': 8.497156047770933e-05, 'epoch': 0.57}) +[2026-06-15 21:41:16,951][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.14055047929286957, 'train/grad_norm_pre_clip': 11.926326751708984, 'timing/data': 0.023722872021608055, 'timing/model': 1.1310427730204538, 'learning_rate/qwen_vl_interface': 8.494932692444276e-06, 'learning_rate/action_model': 8.494932692444276e-05, 'epoch': 0.57}) +[2026-06-15 21:41:21,502][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.1730836182832718, 'train/grad_norm_pre_clip': 23.875667572021484, 'timing/data': 0.023619098006747663, 'timing/model': 1.3014400689862669, 'learning_rate/qwen_vl_interface': 8.492708003248362e-06, 'learning_rate/action_model': 8.492708003248362e-05, 'epoch': 0.57}) +[2026-06-15 21:41:26,033][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.19696706533432007, 'train/grad_norm_pre_clip': 7.550431728363037, 'timing/data': 0.023468373983632773, 'timing/model': 1.2758697580429725, 'learning_rate/qwen_vl_interface': 8.490481981097674e-06, 'learning_rate/action_model': 8.490481981097674e-05, 'epoch': 0.57}) +[2026-06-15 21:41:30,428][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.12489598244428635, 'train/grad_norm_pre_clip': 15.846847534179688, 'timing/data': 0.023057006997987628, 'timing/model': 1.1343255509855226, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 0.57}) +[2026-06-15 21:41:34,986][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.24431052803993225, 'train/grad_norm_pre_clip': 11.69710922241211, 'timing/data': 0.02378328301711008, 'timing/model': 1.1336435239645652, 'learning_rate/qwen_vl_interface': 8.486025941592664e-06, 'learning_rate/action_model': 8.486025941592664e-05, 'epoch': 0.57}) +[2026-06-15 21:41:39,533][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.24566569924354553, 'train/grad_norm_pre_clip': 14.49305534362793, 'timing/data': 0.0234134029597044, 'timing/model': 1.1335039179539308, 'learning_rate/qwen_vl_interface': 8.483795926070054e-06, 'learning_rate/action_model': 8.483795926070052e-05, 'epoch': 0.57}) +[2026-06-15 21:41:44,073][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.20906193554401398, 'train/grad_norm_pre_clip': 13.556846618652344, 'timing/data': 0.024226273002568632, 'timing/model': 1.1359198300051503, 'learning_rate/qwen_vl_interface': 8.481564581256087e-06, 'learning_rate/action_model': 8.481564581256087e-05, 'epoch': 0.57}) +[2026-06-15 21:41:48,567][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.14407509565353394, 'train/grad_norm_pre_clip': 5.104087829589844, 'timing/data': 0.023602386994753033, 'timing/model': 1.1313664460321888, 'learning_rate/qwen_vl_interface': 8.479331908067987e-06, 'learning_rate/action_model': 8.479331908067987e-05, 'epoch': 0.57}) +[2026-06-15 21:41:53,097][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.0834110751748085, 'train/grad_norm_pre_clip': 4.669304370880127, 'timing/data': 0.024363273987546563, 'timing/model': 1.2884980889502913, 'learning_rate/qwen_vl_interface': 8.477097907423524e-06, 'learning_rate/action_model': 8.477097907423524e-05, 'epoch': 0.57}) +[2026-06-15 21:41:57,491][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.10241836309432983, 'train/grad_norm_pre_clip': 5.8420891761779785, 'timing/data': 0.023006226983852684, 'timing/model': 1.1408106469898485, 'learning_rate/qwen_vl_interface': 8.474862580241006e-06, 'learning_rate/action_model': 8.474862580241006e-05, 'epoch': 0.57}) +[2026-06-15 21:42:02,040][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.09840423613786697, 'train/grad_norm_pre_clip': 10.687421798706055, 'timing/data': 0.023180784948635846, 'timing/model': 1.1493416199809872, 'learning_rate/qwen_vl_interface': 8.472625927439299e-06, 'learning_rate/action_model': 8.472625927439299e-05, 'epoch': 0.57}) +[2026-06-15 21:42:06,496][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.0854581817984581, 'train/grad_norm_pre_clip': 3.978546142578125, 'timing/data': 0.022373105981387198, 'timing/model': 1.1363166429800913, 'learning_rate/qwen_vl_interface': 8.4703879499378e-06, 'learning_rate/action_model': 8.4703879499378e-05, 'epoch': 0.57}) +[2026-06-15 21:42:10,972][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.05927453190088272, 'train/grad_norm_pre_clip': 3.2022387981414795, 'timing/data': 0.023012196004856378, 'timing/model': 1.136769409000408, 'learning_rate/qwen_vl_interface': 8.46814864865646e-06, 'learning_rate/action_model': 8.46814864865646e-05, 'epoch': 0.57}) +[2026-06-15 21:42:15,467][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.0304123405367136, 'train/grad_norm_pre_clip': 2.021693706512451, 'timing/data': 0.02326967497356236, 'timing/model': 1.1365177180268802, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 0.57}) +[2026-06-15 21:42:19,989][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.07483143359422684, 'train/grad_norm_pre_clip': 5.124410152435303, 'timing/data': 0.024189179996028543, 'timing/model': 1.2557221610331908, 'learning_rate/qwen_vl_interface': 8.463666078436769e-06, 'learning_rate/action_model': 8.463666078436769e-05, 'epoch': 0.57}) +[2026-06-15 21:42:24,382][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.08808908611536026, 'train/grad_norm_pre_clip': 4.762134552001953, 'timing/data': 0.02333794900914654, 'timing/model': 1.135939800995402, 'learning_rate/qwen_vl_interface': 8.46142281134103e-06, 'learning_rate/action_model': 8.461422811341031e-05, 'epoch': 0.57}) +[2026-06-15 21:42:28,944][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.32034197449684143, 'train/grad_norm_pre_clip': 9.653685569763184, 'timing/data': 0.023306997027248144, 'timing/model': 1.1353230960085057, 'learning_rate/qwen_vl_interface': 8.459178224150683e-06, 'learning_rate/action_model': 8.459178224150683e-05, 'epoch': 0.57}) +[2026-06-15 21:42:33,511][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.31601089239120483, 'train/grad_norm_pre_clip': 20.995054244995117, 'timing/data': 0.022642876021564007, 'timing/model': 1.1350362590164877, 'learning_rate/qwen_vl_interface': 8.456932317788389e-06, 'learning_rate/action_model': 8.456932317788389e-05, 'epoch': 0.57}) +[2026-06-15 21:42:38,062][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.07108350843191147, 'train/grad_norm_pre_clip': 6.129380702972412, 'timing/data': 0.02346939197741449, 'timing/model': 1.1339542610221542, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 0.57}) +[2026-06-15 21:42:42,624][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.2388582080602646, 'train/grad_norm_pre_clip': 15.932374954223633, 'timing/data': 0.0233004130423069, 'timing/model': 1.3002904370077886, 'learning_rate/qwen_vl_interface': 8.452436551241328e-06, 'learning_rate/action_model': 8.452436551241328e-05, 'epoch': 0.57}) +[2026-06-15 21:42:47,024][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.1788455843925476, 'train/grad_norm_pre_clip': 8.120814323425293, 'timing/data': 0.023200338997412473, 'timing/model': 1.1420906130224466, 'learning_rate/qwen_vl_interface': 8.450186692904605e-06, 'learning_rate/action_model': 8.450186692904606e-05, 'epoch': 0.57}) +[2026-06-15 21:42:51,627][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.04867646098136902, 'train/grad_norm_pre_clip': 8.96768856048584, 'timing/data': 0.02367497602244839, 'timing/model': 1.1351876560365781, 'learning_rate/qwen_vl_interface': 8.447935519092014e-06, 'learning_rate/action_model': 8.447935519092014e-05, 'epoch': 0.57}) +[2026-06-15 21:42:56,134][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.05920998752117157, 'train/grad_norm_pre_clip': 5.1949005126953125, 'timing/data': 0.031365522008854896, 'timing/model': 1.13418544101296, 'learning_rate/qwen_vl_interface': 8.445683030728925e-06, 'learning_rate/action_model': 8.445683030728925e-05, 'epoch': 0.57}) +[2026-06-15 21:47:19,183][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.10120633989572525, 'train/grad_norm_pre_clip': 12.427388191223145, 'eval/loss': 0.06957605481147766, 'eval/flap_precision': 0.9354838709677419, 'eval/flap_recall': 0.9398148148148148, 'eval/flap_f1': 0.9376443418013857, 'eval/flap_rate_model': 0.18083333333333335, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9376443418013856, 'timing/data': 0.023272287973668426, 'timing/model': 1.1376307990285568, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 0.58}) +[2026-06-15 21:47:19,183][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1400_state +[2026-06-15 21:47:19,186][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 21:47:22,784][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1400_state/model.safetensors +[2026-06-15 21:47:32,762][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1400_state/optimizer.bin +[2026-06-15 21:47:32,762][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1400_state/sampler.bin +[2026-06-15 21:47:32,762][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1400_state/sampler_1.bin +[2026-06-15 21:47:32,763][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1400_state/random_states_0.pkl +[2026-06-15 21:47:32,807][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-06-15 21:47:32,807][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 21:47:32,856][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 21:47:37,327][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.04517827555537224, 'train/grad_norm_pre_clip': 5.681849479675293, 'timing/data': 0.02302318101283163, 'timing/model': 1.109655751963146, 'learning_rate/qwen_vl_interface': 8.441174114055457e-06, 'learning_rate/action_model': 8.441174114055456e-05, 'epoch': 0.58}) +[2026-06-15 21:47:41,790][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.0123980101197958, 'train/grad_norm_pre_clip': 1.1418111324310303, 'timing/data': 0.022708451026119292, 'timing/model': 1.1132149839540944, 'learning_rate/qwen_vl_interface': 8.43891768759852e-06, 'learning_rate/action_model': 8.43891768759852e-05, 'epoch': 0.58}) +[2026-06-15 21:47:46,251][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.14905233681201935, 'train/grad_norm_pre_clip': 10.328789710998535, 'timing/data': 0.023675161006394774, 'timing/model': 1.1176712980377488, 'learning_rate/qwen_vl_interface': 8.436659950297978e-06, 'learning_rate/action_model': 8.436659950297977e-05, 'epoch': 0.58}) +[2026-06-15 21:47:50,719][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.019822189584374428, 'train/grad_norm_pre_clip': 2.190941095352173, 'timing/data': 0.02397233503870666, 'timing/model': 1.2678509239922278, 'learning_rate/qwen_vl_interface': 8.4344009030819e-06, 'learning_rate/action_model': 8.4344009030819e-05, 'epoch': 0.58}) +[2026-06-15 21:47:55,044][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.13708578050136566, 'train/grad_norm_pre_clip': 7.123571395874023, 'timing/data': 0.023813695006538182, 'timing/model': 1.1193400360061787, 'learning_rate/qwen_vl_interface': 8.432140546878899e-06, 'learning_rate/action_model': 8.432140546878898e-05, 'epoch': 0.58}) +[2026-06-15 21:47:59,533][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.08085343986749649, 'train/grad_norm_pre_clip': 11.733063697814941, 'timing/data': 0.023447780986316502, 'timing/model': 1.1193063940154389, 'learning_rate/qwen_vl_interface': 8.429878882618116e-06, 'learning_rate/action_model': 8.429878882618116e-05, 'epoch': 0.58}) +[2026-06-15 21:48:04,020][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.19798016548156738, 'train/grad_norm_pre_clip': 4.2918853759765625, 'timing/data': 0.02398113498929888, 'timing/model': 1.1170626999810338, 'learning_rate/qwen_vl_interface': 8.42761591122924e-06, 'learning_rate/action_model': 8.427615911229241e-05, 'epoch': 0.58}) +[2026-06-15 21:48:08,532][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.134161576628685, 'train/grad_norm_pre_clip': 7.952973365783691, 'timing/data': 0.0239965480286628, 'timing/model': 1.119140027964022, 'learning_rate/qwen_vl_interface': 8.425351633642491e-06, 'learning_rate/action_model': 8.425351633642492e-05, 'epoch': 0.58}) +[2026-06-15 21:48:12,990][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.11063458025455475, 'train/grad_norm_pre_clip': 6.646157741546631, 'timing/data': 0.02308655303204432, 'timing/model': 1.1207906819763593, 'learning_rate/qwen_vl_interface': 8.423086050788632e-06, 'learning_rate/action_model': 8.423086050788631e-05, 'epoch': 0.58}) +[2026-06-15 21:48:17,480][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.2704717814922333, 'train/grad_norm_pre_clip': 31.009641647338867, 'timing/data': 0.022264619008637965, 'timing/model': 1.2817267130012624, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 0.58}) +[2026-06-15 21:48:21,820][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.29545682668685913, 'train/grad_norm_pre_clip': 7.442471504211426, 'timing/data': 0.02268514601746574, 'timing/model': 1.1205950980074704, 'learning_rate/qwen_vl_interface': 8.418550973005289e-06, 'learning_rate/action_model': 8.418550973005289e-05, 'epoch': 0.58}) +[2026-06-15 21:48:26,311][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.07236385345458984, 'train/grad_norm_pre_clip': 15.003759384155273, 'timing/data': 0.022733743011485785, 'timing/model': 1.1215888379956596, 'learning_rate/qwen_vl_interface': 8.416281479940005e-06, 'learning_rate/action_model': 8.416281479940005e-05, 'epoch': 0.58}) +[2026-06-15 21:48:30,810][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.04292764142155647, 'train/grad_norm_pre_clip': 5.989605903625488, 'timing/data': 0.022248708992265165, 'timing/model': 1.1295418019872159, 'learning_rate/qwen_vl_interface': 8.414010685336008e-06, 'learning_rate/action_model': 8.414010685336008e-05, 'epoch': 0.58}) +[2026-06-15 21:48:35,280][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.05278662592172623, 'train/grad_norm_pre_clip': 10.271363258361816, 'timing/data': 0.023473077977541834, 'timing/model': 1.124941899033729, 'learning_rate/qwen_vl_interface': 8.411738590126733e-06, 'learning_rate/action_model': 8.411738590126732e-05, 'epoch': 0.58}) +[2026-06-15 21:48:39,782][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.14864593744277954, 'train/grad_norm_pre_clip': 16.80559539794922, 'timing/data': 0.02338011801475659, 'timing/model': 1.122703882982023, 'learning_rate/qwen_vl_interface': 8.409465195246152e-06, 'learning_rate/action_model': 8.409465195246152e-05, 'epoch': 0.58}) +[2026-06-15 21:48:44,309][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.06528082489967346, 'train/grad_norm_pre_clip': 13.963822364807129, 'timing/data': 0.023516301007475704, 'timing/model': 1.2965549189830199, 'learning_rate/qwen_vl_interface': 8.407190501628772e-06, 'learning_rate/action_model': 8.407190501628772e-05, 'epoch': 0.58}) +[2026-06-15 21:48:48,654][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.036720216274261475, 'train/grad_norm_pre_clip': 4.476935863494873, 'timing/data': 0.023709049972239882, 'timing/model': 1.1248221070272848, 'learning_rate/qwen_vl_interface': 8.404914510209635e-06, 'learning_rate/action_model': 8.404914510209635e-05, 'epoch': 0.58}) +[2026-06-15 21:48:53,171][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.07201322913169861, 'train/grad_norm_pre_clip': 3.414262056350708, 'timing/data': 0.02377617999445647, 'timing/model': 1.1283627380034886, 'learning_rate/qwen_vl_interface': 8.402637221924315e-06, 'learning_rate/action_model': 8.402637221924314e-05, 'epoch': 0.58}) +[2026-06-15 21:48:57,683][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.06075574457645416, 'train/grad_norm_pre_clip': 11.803613662719727, 'timing/data': 0.02285953500540927, 'timing/model': 1.1271922420128249, 'learning_rate/qwen_vl_interface': 8.400358637708919e-06, 'learning_rate/action_model': 8.400358637708918e-05, 'epoch': 0.58}) +[2026-06-15 21:49:02,151][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.44615140557289124, 'train/grad_norm_pre_clip': 15.005535125732422, 'timing/data': 0.023813486041035503, 'timing/model': 1.1345348129980266, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 0.58}) +[2026-06-15 21:49:06,620][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.1488218456506729, 'train/grad_norm_pre_clip': 14.148449897766113, 'timing/data': 0.023091300041414797, 'timing/model': 1.1271745529957116, 'learning_rate/qwen_vl_interface': 8.395797585234987e-06, 'learning_rate/action_model': 8.395797585234988e-05, 'epoch': 0.58}) +[2026-06-15 21:49:11,136][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.07103274017572403, 'train/grad_norm_pre_clip': 8.406009674072266, 'timing/data': 0.023332705022767186, 'timing/model': 1.2833741609938443, 'learning_rate/qwen_vl_interface': 8.39351511885133e-06, 'learning_rate/action_model': 8.393515118851329e-05, 'epoch': 0.58}) +[2026-06-15 21:49:15,488][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.018488941714167595, 'train/grad_norm_pre_clip': 1.3786150217056274, 'timing/data': 0.023473781009670347, 'timing/model': 1.127586633025203, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.58}) +[2026-06-15 21:49:20,020][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.06082872301340103, 'train/grad_norm_pre_clip': 8.572872161865234, 'timing/data': 0.023608027026057243, 'timing/model': 1.1413109929999337, 'learning_rate/qwen_vl_interface': 8.388946310481807e-06, 'learning_rate/action_model': 8.388946310481807e-05, 'epoch': 0.58}) +[2026-06-15 21:49:24,589][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.1295078992843628, 'train/grad_norm_pre_clip': 16.712953567504883, 'timing/data': 0.02341100398916751, 'timing/model': 1.1392112690373324, 'learning_rate/qwen_vl_interface': 8.386659970374007e-06, 'learning_rate/action_model': 8.386659970374006e-05, 'epoch': 0.59}) +[2026-06-15 21:49:29,133][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.0566275492310524, 'train/grad_norm_pre_clip': 12.628018379211426, 'timing/data': 0.02367387100821361, 'timing/model': 1.1362478930386715, 'learning_rate/qwen_vl_interface': 8.384372340903775e-06, 'learning_rate/action_model': 8.384372340903775e-05, 'epoch': 0.59}) +[2026-06-15 21:49:33,627][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.2745586633682251, 'train/grad_norm_pre_clip': 22.10887336730957, 'timing/data': 0.023752338020130992, 'timing/model': 1.1371637339470908, 'learning_rate/qwen_vl_interface': 8.382083423011472e-06, 'learning_rate/action_model': 8.382083423011471e-05, 'epoch': 0.59}) +[2026-06-15 21:49:38,154][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.06225941330194473, 'train/grad_norm_pre_clip': 6.5149664878845215, 'timing/data': 0.023061226995196193, 'timing/model': 1.2925920060370117, 'learning_rate/qwen_vl_interface': 8.379793217637979e-06, 'learning_rate/action_model': 8.379793217637979e-05, 'epoch': 0.59}) +[2026-06-15 21:49:42,530][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.11034724116325378, 'train/grad_norm_pre_clip': 8.907410621643066, 'timing/data': 0.022790772025473416, 'timing/model': 1.1364041860215366, 'learning_rate/qwen_vl_interface': 8.377501725724719e-06, 'learning_rate/action_model': 8.377501725724719e-05, 'epoch': 0.59}) +[2026-06-15 21:49:47,055][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.2447950690984726, 'train/grad_norm_pre_clip': 20.6024227142334, 'timing/data': 0.023578643973451108, 'timing/model': 1.129307494033128, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.59}) +[2026-06-15 21:49:51,476][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.0223220381885767, 'train/grad_norm_pre_clip': 2.9150097370147705, 'timing/data': 0.023090481001418084, 'timing/model': 1.1295884430292062, 'learning_rate/qwen_vl_interface': 8.372914886047203e-06, 'learning_rate/action_model': 8.372914886047203e-05, 'epoch': 0.59}) +[2026-06-15 21:49:55,904][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.019715454429388046, 'train/grad_norm_pre_clip': 2.9580307006835938, 'timing/data': 0.02402471797540784, 'timing/model': 1.1322231870144606, 'learning_rate/qwen_vl_interface': 8.370619540168422e-06, 'learning_rate/action_model': 8.370619540168423e-05, 'epoch': 0.59}) +[2026-06-15 21:50:00,392][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.10774879902601242, 'train/grad_norm_pre_clip': 16.837018966674805, 'timing/data': 0.02370749699184671, 'timing/model': 1.131203394033946, 'learning_rate/qwen_vl_interface': 8.368322911520826e-06, 'learning_rate/action_model': 8.368322911520826e-05, 'epoch': 0.59}) +[2026-06-15 21:50:04,904][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.06639526784420013, 'train/grad_norm_pre_clip': 7.060551643371582, 'timing/data': 0.022787668043747544, 'timing/model': 1.2732104280148633, 'learning_rate/qwen_vl_interface': 8.36602500104847e-06, 'learning_rate/action_model': 8.36602500104847e-05, 'epoch': 0.59}) +[2026-06-15 21:50:09,291][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.014178065583109856, 'train/grad_norm_pre_clip': 1.6476808786392212, 'timing/data': 0.023112093040253967, 'timing/model': 1.1321109010023065, 'learning_rate/qwen_vl_interface': 8.36372580969594e-06, 'learning_rate/action_model': 8.36372580969594e-05, 'epoch': 0.59}) +[2026-06-15 21:50:13,841][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.1435948759317398, 'train/grad_norm_pre_clip': 10.370433807373047, 'timing/data': 0.023179242969490588, 'timing/model': 1.130725489987526, 'learning_rate/qwen_vl_interface': 8.361425338408344e-06, 'learning_rate/action_model': 8.361425338408343e-05, 'epoch': 0.59}) +[2026-06-15 21:50:18,371][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.2072298377752304, 'train/grad_norm_pre_clip': 12.027326583862305, 'timing/data': 0.02300504600862041, 'timing/model': 1.1313456820207648, 'learning_rate/qwen_vl_interface': 8.359123588131319e-06, 'learning_rate/action_model': 8.359123588131319e-05, 'epoch': 0.59}) +[2026-06-15 21:50:22,896][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.04482939466834068, 'train/grad_norm_pre_clip': 8.343931198120117, 'timing/data': 0.02309831598540768, 'timing/model': 1.128512267023325, 'learning_rate/qwen_vl_interface': 8.35682055981103e-06, 'learning_rate/action_model': 8.35682055981103e-05, 'epoch': 0.59}) +[2026-06-15 21:50:27,436][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.11473413556814194, 'train/grad_norm_pre_clip': 9.115653038024902, 'timing/data': 0.02325132698751986, 'timing/model': 1.2965707309776917, 'learning_rate/qwen_vl_interface': 8.354516254394163e-06, 'learning_rate/action_model': 8.354516254394162e-05, 'epoch': 0.59}) +[2026-06-15 21:50:31,819][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.17607417702674866, 'train/grad_norm_pre_clip': 5.6135663986206055, 'timing/data': 0.023016666993498802, 'timing/model': 1.1315086710383184, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 0.59}) +[2026-06-15 21:50:36,363][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.3028193414211273, 'train/grad_norm_pre_clip': 22.814577102661133, 'timing/data': 0.022969554993323982, 'timing/model': 1.1331440199865028, 'learning_rate/qwen_vl_interface': 8.349903816060074e-06, 'learning_rate/action_model': 8.349903816060075e-05, 'epoch': 0.59}) +[2026-06-15 21:50:40,830][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.0714782178401947, 'train/grad_norm_pre_clip': 9.16144847869873, 'timing/data': 0.023464648984372616, 'timing/model': 1.1314692230080254, 'learning_rate/qwen_vl_interface': 8.34759568503885e-06, 'learning_rate/action_model': 8.347595685038851e-05, 'epoch': 0.59}) +[2026-06-15 21:50:45,368][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.10188750177621841, 'train/grad_norm_pre_clip': 6.020577907562256, 'timing/data': 0.023522677016444504, 'timing/model': 1.1347600679728203, 'learning_rate/qwen_vl_interface': 8.345286280713049e-06, 'learning_rate/action_model': 8.345286280713048e-05, 'epoch': 0.59}) +[2026-06-15 21:50:49,924][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.3972690999507904, 'train/grad_norm_pre_clip': 16.714086532592773, 'timing/data': 0.023443370999302715, 'timing/model': 1.1355495900497772, 'learning_rate/qwen_vl_interface': 8.342975604031974e-06, 'learning_rate/action_model': 8.342975604031974e-05, 'epoch': 0.59}) +[2026-06-15 21:50:54,472][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.2070469856262207, 'train/grad_norm_pre_clip': 13.22743034362793, 'timing/data': 0.023445306986104697, 'timing/model': 1.2869240620057099, 'learning_rate/qwen_vl_interface': 8.340663655945462e-06, 'learning_rate/action_model': 8.340663655945462e-05, 'epoch': 0.59}) +[2026-06-15 21:50:58,967][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.37602710723876953, 'train/grad_norm_pre_clip': 22.745637893676758, 'timing/data': 0.02261000103317201, 'timing/model': 1.2504878650070168, 'learning_rate/qwen_vl_interface': 8.338350437403866e-06, 'learning_rate/action_model': 8.338350437403865e-05, 'epoch': 0.59}) +[2026-06-15 21:51:03,357][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.1438264548778534, 'train/grad_norm_pre_clip': 11.399042129516602, 'timing/data': 0.022859785007312894, 'timing/model': 1.1403992010164075, 'learning_rate/qwen_vl_interface': 8.336035949358063e-06, 'learning_rate/action_model': 8.336035949358062e-05, 'epoch': 0.59}) +[2026-06-15 21:51:07,905][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.18347562849521637, 'train/grad_norm_pre_clip': 6.5781965255737305, 'timing/data': 0.022995278995949775, 'timing/model': 1.132322904013563, 'learning_rate/qwen_vl_interface': 8.333720192759448e-06, 'learning_rate/action_model': 8.333720192759448e-05, 'epoch': 0.59}) +[2026-06-15 21:51:12,456][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.22443699836730957, 'train/grad_norm_pre_clip': 7.430655479431152, 'timing/data': 0.02310739498352632, 'timing/model': 1.1325642049778253, 'learning_rate/qwen_vl_interface': 8.331403168559946e-06, 'learning_rate/action_model': 8.331403168559945e-05, 'epoch': 0.6}) +[2026-06-15 21:51:16,989][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.1325809359550476, 'train/grad_norm_pre_clip': 12.591582298278809, 'timing/data': 0.023309221025556326, 'timing/model': 1.1315279979608022, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 0.6}) +[2026-06-15 21:51:21,508][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.03363916277885437, 'train/grad_norm_pre_clip': 6.184963226318359, 'timing/data': 0.023042836983222514, 'timing/model': 1.2763838419923559, 'learning_rate/qwen_vl_interface': 8.32676532116856e-06, 'learning_rate/action_model': 8.326765321168559e-05, 'epoch': 0.6}) +[2026-06-15 21:51:26,019][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.2075057327747345, 'train/grad_norm_pre_clip': 6.514972686767578, 'timing/data': 0.023603848996572196, 'timing/model': 1.2686847819713876, 'learning_rate/qwen_vl_interface': 8.324444499883119e-06, 'learning_rate/action_model': 8.324444499883119e-05, 'epoch': 0.6}) +[2026-06-15 21:51:30,390][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.03390885517001152, 'train/grad_norm_pre_clip': 3.702453851699829, 'timing/data': 0.023478576040361077, 'timing/model': 1.1320802320260555, 'learning_rate/qwen_vl_interface': 8.322122414809678e-06, 'learning_rate/action_model': 8.322122414809677e-05, 'epoch': 0.6}) +[2026-06-15 21:51:34,912][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.12135668843984604, 'train/grad_norm_pre_clip': 5.162679672241211, 'timing/data': 0.02372717799153179, 'timing/model': 1.1307925159926526, 'learning_rate/qwen_vl_interface': 8.319799066902755e-06, 'learning_rate/action_model': 8.319799066902755e-05, 'epoch': 0.6}) +[2026-06-15 21:51:39,446][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.018158281221985817, 'train/grad_norm_pre_clip': 1.00712251663208, 'timing/data': 0.023724407015834004, 'timing/model': 1.1311627890099771, 'learning_rate/qwen_vl_interface': 8.317474457117393e-06, 'learning_rate/action_model': 8.317474457117392e-05, 'epoch': 0.6}) +[2026-06-15 21:51:43,983][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.2515849471092224, 'train/grad_norm_pre_clip': 9.3702392578125, 'timing/data': 0.024070152023341507, 'timing/model': 1.129370259004645, 'learning_rate/qwen_vl_interface': 8.315148586409148e-06, 'learning_rate/action_model': 8.315148586409148e-05, 'epoch': 0.6}) +[2026-06-15 21:51:48,478][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.11016643792390823, 'train/grad_norm_pre_clip': 7.291976451873779, 'timing/data': 0.02333561098203063, 'timing/model': 1.1351543960045092, 'learning_rate/qwen_vl_interface': 8.312821455734104e-06, 'learning_rate/action_model': 8.312821455734103e-05, 'epoch': 0.6}) +[2026-06-15 21:51:53,040][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.0990530326962471, 'train/grad_norm_pre_clip': 11.907804489135742, 'timing/data': 0.02373032202012837, 'timing/model': 1.3178114509792067, 'learning_rate/qwen_vl_interface': 8.31049306604885e-06, 'learning_rate/action_model': 8.310493066048849e-05, 'epoch': 0.6}) +[2026-06-15 21:51:57,427][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.06459898501634598, 'train/grad_norm_pre_clip': 3.7580859661102295, 'timing/data': 0.023506491037551314, 'timing/model': 1.1331438350025564, 'learning_rate/qwen_vl_interface': 8.3081634183105e-06, 'learning_rate/action_model': 8.308163418310501e-05, 'epoch': 0.6}) +[2026-06-15 21:52:01,992][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.10701866447925568, 'train/grad_norm_pre_clip': 5.561337947845459, 'timing/data': 0.02299547498114407, 'timing/model': 1.1328469339641742, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 0.6}) +[2026-06-15 21:52:06,434][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.19924962520599365, 'train/grad_norm_pre_clip': 16.261672973632812, 'timing/data': 0.022824051964562386, 'timing/model': 1.1324470430263318, 'learning_rate/qwen_vl_interface': 8.303500352505557e-06, 'learning_rate/action_model': 8.303500352505557e-05, 'epoch': 0.6}) +[2026-06-15 21:52:10,925][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.405962198972702, 'train/grad_norm_pre_clip': 18.540943145751953, 'timing/data': 0.023950717004481703, 'timing/model': 1.1391305400175042, 'learning_rate/qwen_vl_interface': 8.301166936355773e-06, 'learning_rate/action_model': 8.301166936355773e-05, 'epoch': 0.6}) +[2026-06-15 21:52:15,424][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.010594951920211315, 'train/grad_norm_pre_clip': 0.7591516375541687, 'timing/data': 0.023641874024178833, 'timing/model': 1.1383216200047173, 'learning_rate/qwen_vl_interface': 8.298832265986514e-06, 'learning_rate/action_model': 8.298832265986513e-05, 'epoch': 0.6}) +[2026-06-15 21:52:19,975][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.10772334039211273, 'train/grad_norm_pre_clip': 9.745728492736816, 'timing/data': 0.031548231025226414, 'timing/model': 1.2722400309867226, 'learning_rate/qwen_vl_interface': 8.296496342357475e-06, 'learning_rate/action_model': 8.296496342357474e-05, 'epoch': 0.6}) +[2026-06-15 21:52:24,362][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.20179428160190582, 'train/grad_norm_pre_clip': 7.530941963195801, 'timing/data': 0.02874734502984211, 'timing/model': 1.1362715360010043, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 0.6}) +[2026-06-15 21:52:28,924][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.07423361390829086, 'train/grad_norm_pre_clip': 7.10674524307251, 'timing/data': 0.02320137497736141, 'timing/model': 1.1390165039920248, 'learning_rate/qwen_vl_interface': 8.29182073916141e-06, 'learning_rate/action_model': 8.29182073916141e-05, 'epoch': 0.6}) +[2026-06-15 21:52:33,484][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.06881294399499893, 'train/grad_norm_pre_clip': 13.2507905960083, 'timing/data': 0.02259876800235361, 'timing/model': 1.1346334649715573, 'learning_rate/qwen_vl_interface': 8.28948106151635e-06, 'learning_rate/action_model': 8.28948106151635e-05, 'epoch': 0.6}) +[2026-06-15 21:52:38,020][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.2258986532688141, 'train/grad_norm_pre_clip': 8.840380668640137, 'timing/data': 0.028351678978651762, 'timing/model': 1.1318942829966545, 'learning_rate/qwen_vl_interface': 8.287140134455435e-06, 'learning_rate/action_model': 8.287140134455435e-05, 'epoch': 0.6}) +[2026-06-15 21:52:42,554][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.015755753964185715, 'train/grad_norm_pre_clip': 1.6387211084365845, 'timing/data': 0.029074030986521393, 'timing/model': 1.2873681249911897, 'learning_rate/qwen_vl_interface': 8.284797958940934e-06, 'learning_rate/action_model': 8.284797958940935e-05, 'epoch': 0.6}) +[2026-06-15 21:52:46,933][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.28400135040283203, 'train/grad_norm_pre_clip': 16.823463439941406, 'timing/data': 0.022711913974490017, 'timing/model': 1.1334298399742693, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 0.6}) +[2026-06-15 21:52:51,515][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.17606253921985626, 'train/grad_norm_pre_clip': 12.807136535644531, 'timing/data': 0.023018286970909685, 'timing/model': 1.1461905900505371, 'learning_rate/qwen_vl_interface': 8.280109866402805e-06, 'learning_rate/action_model': 8.280109866402806e-05, 'epoch': 0.6}) +[2026-06-15 21:52:56,014][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.12285216897726059, 'train/grad_norm_pre_clip': 6.592892169952393, 'timing/data': 0.023089146008715034, 'timing/model': 1.1445443640113808, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.6}) +[2026-06-15 21:53:00,568][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.08166545629501343, 'train/grad_norm_pre_clip': 17.087020874023438, 'timing/data': 0.02378365903859958, 'timing/model': 1.1398786189965904, 'learning_rate/qwen_vl_interface': 8.275416791610356e-06, 'learning_rate/action_model': 8.275416791610355e-05, 'epoch': 0.61}) +[2026-06-15 21:53:05,116][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.06045788154006004, 'train/grad_norm_pre_clip': 6.957275390625, 'timing/data': 0.022783762018661946, 'timing/model': 1.1378815020434558, 'learning_rate/qwen_vl_interface': 8.27306838827987e-06, 'learning_rate/action_model': 8.273068388279871e-05, 'epoch': 0.61}) +[2026-06-15 21:53:09,678][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.2042384147644043, 'train/grad_norm_pre_clip': 9.432647705078125, 'timing/data': 0.023264034010935575, 'timing/model': 1.2892994260182604, 'learning_rate/qwen_vl_interface': 8.270718742280162e-06, 'learning_rate/action_model': 8.270718742280162e-05, 'epoch': 0.61}) +[2026-06-15 21:53:14,195][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.10885640978813171, 'train/grad_norm_pre_clip': 13.778800964355469, 'timing/data': 0.023313882004003972, 'timing/model': 1.2667902290122584, 'learning_rate/qwen_vl_interface': 8.268367854577084e-06, 'learning_rate/action_model': 8.268367854577082e-05, 'epoch': 0.61}) +[2026-06-15 21:53:18,593][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.3239525258541107, 'train/grad_norm_pre_clip': 10.45201587677002, 'timing/data': 0.02347304404247552, 'timing/model': 1.13569012604421, 'learning_rate/qwen_vl_interface': 8.266015726136991e-06, 'learning_rate/action_model': 8.266015726136991e-05, 'epoch': 0.61}) +[2026-06-15 21:53:23,144][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.03784511983394623, 'train/grad_norm_pre_clip': 5.217644691467285, 'timing/data': 0.02319951303070411, 'timing/model': 1.1337286969646811, 'learning_rate/qwen_vl_interface': 8.26366235792676e-06, 'learning_rate/action_model': 8.26366235792676e-05, 'epoch': 0.61}) +[2026-06-15 21:53:27,702][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.08388814330101013, 'train/grad_norm_pre_clip': 10.871500015258789, 'timing/data': 0.023026266018860042, 'timing/model': 1.1343124759732746, 'learning_rate/qwen_vl_interface': 8.261307750913769e-06, 'learning_rate/action_model': 8.261307750913769e-05, 'epoch': 0.61}) +[2026-06-15 21:53:32,246][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.1852625012397766, 'train/grad_norm_pre_clip': 12.699980735778809, 'timing/data': 0.0231940450030379, 'timing/model': 1.1277288140263408, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 0.61}) +[2026-06-15 21:53:36,780][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.22062760591506958, 'train/grad_norm_pre_clip': 21.952701568603516, 'timing/data': 0.0244817030034028, 'timing/model': 1.2968293650192209, 'learning_rate/qwen_vl_interface': 8.256594824351577e-06, 'learning_rate/action_model': 8.256594824351576e-05, 'epoch': 0.61}) +[2026-06-15 21:53:41,309][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.08397939056158066, 'train/grad_norm_pre_clip': 8.427267074584961, 'timing/data': 0.022991483041550964, 'timing/model': 1.2785890569793992, 'learning_rate/qwen_vl_interface': 8.254236506739683e-06, 'learning_rate/action_model': 8.254236506739681e-05, 'epoch': 0.61}) +[2026-06-15 21:53:45,680][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.009100762195885181, 'train/grad_norm_pre_clip': 0.6995114684104919, 'timing/data': 0.023200272989924997, 'timing/model': 1.1316224149777554, 'learning_rate/qwen_vl_interface': 8.25187695419964e-06, 'learning_rate/action_model': 8.25187695419964e-05, 'epoch': 0.61}) +[2026-06-15 21:53:50,211][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.0525253489613533, 'train/grad_norm_pre_clip': 4.62924337387085, 'timing/data': 0.022782300016842782, 'timing/model': 1.1343111679889262, 'learning_rate/qwen_vl_interface': 8.249516167701373e-06, 'learning_rate/action_model': 8.249516167701374e-05, 'epoch': 0.61}) +[2026-06-15 21:53:54,764][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.199076309800148, 'train/grad_norm_pre_clip': 9.878602027893066, 'timing/data': 0.023467960010748357, 'timing/model': 1.14488189102849, 'learning_rate/qwen_vl_interface': 8.247154148215311e-06, 'learning_rate/action_model': 8.24715414821531e-05, 'epoch': 0.61}) +[2026-06-15 21:53:59,320][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.07872243970632553, 'train/grad_norm_pre_clip': 5.819389820098877, 'timing/data': 0.024021015968173742, 'timing/model': 1.1342135079903528, 'learning_rate/qwen_vl_interface': 8.24479089671239e-06, 'learning_rate/action_model': 8.24479089671239e-05, 'epoch': 0.61}) +[2026-06-15 21:54:03,825][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.032186008989810944, 'train/grad_norm_pre_clip': 2.6262497901916504, 'timing/data': 0.023250458005350083, 'timing/model': 1.1362430619774386, 'learning_rate/qwen_vl_interface': 8.242426414164056e-06, 'learning_rate/action_model': 8.242426414164056e-05, 'epoch': 0.61}) +[2026-06-15 21:54:08,371][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.07627033442258835, 'train/grad_norm_pre_clip': 7.868139743804932, 'timing/data': 0.022432275000028312, 'timing/model': 1.2973973240004852, 'learning_rate/qwen_vl_interface': 8.240060701542257e-06, 'learning_rate/action_model': 8.240060701542257e-05, 'epoch': 0.61}) +[2026-06-15 21:54:12,755][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.04705306887626648, 'train/grad_norm_pre_clip': 12.966543197631836, 'timing/data': 0.023587514006067067, 'timing/model': 1.134638545976486, 'learning_rate/qwen_vl_interface': 8.237693759819447e-06, 'learning_rate/action_model': 8.237693759819448e-05, 'epoch': 0.61}) +[2026-06-15 21:54:17,314][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.10361282527446747, 'train/grad_norm_pre_clip': 9.015913963317871, 'timing/data': 0.022936051012948155, 'timing/model': 1.1334573440253735, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 0.61}) +[2026-06-15 21:54:21,759][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.02958638034760952, 'train/grad_norm_pre_clip': 3.057133436203003, 'timing/data': 0.023063801985699683, 'timing/model': 1.1372475679963827, 'learning_rate/qwen_vl_interface': 8.232956192963143e-06, 'learning_rate/action_model': 8.232956192963144e-05, 'epoch': 0.61}) +[2026-06-15 21:54:26,221][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.032880015671253204, 'train/grad_norm_pre_clip': 4.2425432205200195, 'timing/data': 0.02256038400810212, 'timing/model': 1.1357043780153617, 'learning_rate/qwen_vl_interface': 8.230585569777086e-06, 'learning_rate/action_model': 8.230585569777085e-05, 'epoch': 0.61}) +[2026-06-15 21:54:30,711][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.21781699359416962, 'train/grad_norm_pre_clip': 9.963408470153809, 'timing/data': 0.023629323986824602, 'timing/model': 1.1418329940061085, 'learning_rate/qwen_vl_interface': 8.228213721384885e-06, 'learning_rate/action_model': 8.228213721384885e-05, 'epoch': 0.61}) +[2026-06-15 21:54:35,222][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.24541661143302917, 'train/grad_norm_pre_clip': 16.024240493774414, 'timing/data': 0.023694359988439828, 'timing/model': 1.2605241519631818, 'learning_rate/qwen_vl_interface': 8.225840648761523e-06, 'learning_rate/action_model': 8.225840648761523e-05, 'epoch': 0.61}) +[2026-06-15 21:54:39,602][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.28735411167144775, 'train/grad_norm_pre_clip': 8.119565963745117, 'timing/data': 0.022451871016528457, 'timing/model': 1.1386461880174465, 'learning_rate/qwen_vl_interface': 8.223466352882476e-06, 'learning_rate/action_model': 8.223466352882475e-05, 'epoch': 0.61}) +[2026-06-15 21:54:44,136][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.13360103964805603, 'train/grad_norm_pre_clip': 7.1275410652160645, 'timing/data': 0.023694484028965235, 'timing/model': 1.1309355210396461, 'learning_rate/qwen_vl_interface': 8.22109083472373e-06, 'learning_rate/action_model': 8.22109083472373e-05, 'epoch': 0.61}) +[2026-06-15 21:54:48,660][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.18952593207359314, 'train/grad_norm_pre_clip': 11.93439769744873, 'timing/data': 0.023510503000579774, 'timing/model': 1.1303329559741542, 'learning_rate/qwen_vl_interface': 8.218714095261769e-06, 'learning_rate/action_model': 8.218714095261768e-05, 'epoch': 0.61}) +[2026-06-15 21:54:53,214][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.11778685450553894, 'train/grad_norm_pre_clip': 11.832444190979004, 'timing/data': 0.023437474970705807, 'timing/model': 1.1315368510549888, 'learning_rate/qwen_vl_interface': 8.21633613547358e-06, 'learning_rate/action_model': 8.21633613547358e-05, 'epoch': 0.62}) +[2026-06-15 21:54:57,732][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.21356216073036194, 'train/grad_norm_pre_clip': 10.960129737854004, 'timing/data': 0.023709219007287174, 'timing/model': 1.2803666719701141, 'learning_rate/qwen_vl_interface': 8.213956956336656e-06, 'learning_rate/action_model': 8.213956956336656e-05, 'epoch': 0.62}) +[2026-06-15 21:59:16,883][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.09130275249481201, 'train/grad_norm_pre_clip': 7.114031791687012, 'eval/loss': 0.07007701694965363, 'eval/flap_precision': 0.9447004608294931, 'eval/flap_recall': 0.9490740740740741, 'eval/flap_f1': 0.9468822170900693, 'eval/flap_rate_model': 0.18083333333333335, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9468822170900693, 'timing/data': 0.023484675970394164, 'timing/model': 1.1411617529811338, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.62}) +[2026-06-15 21:59:16,883][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1500_state +[2026-06-15 21:59:16,886][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 21:59:20,477][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1500_state/model.safetensors +[2026-06-15 21:59:28,018][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1500_state/optimizer.bin +[2026-06-15 21:59:28,019][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1500_state/sampler.bin +[2026-06-15 21:59:28,019][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-15 21:59:28,019][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-15 21:59:28,067][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-15 21:59:28,067][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 21:59:28,073][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 21:59:32,594][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.07874396443367004, 'train/grad_norm_pre_clip': 4.9230427742004395, 'timing/data': 0.026261769991833717, 'timing/model': 1.314236981968861, 'learning_rate/qwen_vl_interface': 8.209194943929059e-06, 'learning_rate/action_model': 8.20919494392906e-05, 'epoch': 0.62}) +[2026-06-15 21:59:36,915][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.12979742884635925, 'train/grad_norm_pre_clip': 8.351768493652344, 'timing/data': 0.02310015296097845, 'timing/model': 1.11506747698877, 'learning_rate/qwen_vl_interface': 8.20681211261587e-06, 'learning_rate/action_model': 8.20681211261587e-05, 'epoch': 0.62}) +[2026-06-15 21:59:41,419][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.2193101942539215, 'train/grad_norm_pre_clip': 9.450094223022461, 'timing/data': 0.023270706995390356, 'timing/model': 1.1232682609697804, 'learning_rate/qwen_vl_interface': 8.204428065868911e-06, 'learning_rate/action_model': 8.204428065868911e-05, 'epoch': 0.62}) +[2026-06-15 21:59:45,873][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.17665685713291168, 'train/grad_norm_pre_clip': 11.540201187133789, 'timing/data': 0.0230067900265567, 'timing/model': 1.115079186973162, 'learning_rate/qwen_vl_interface': 8.202042804668173e-06, 'learning_rate/action_model': 8.202042804668172e-05, 'epoch': 0.62}) +[2026-06-15 21:59:50,374][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.13240762054920197, 'train/grad_norm_pre_clip': 5.86594295501709, 'timing/data': 0.023434937000274658, 'timing/model': 1.1280240489868447, 'learning_rate/qwen_vl_interface': 8.199656329994145e-06, 'learning_rate/action_model': 8.199656329994145e-05, 'epoch': 0.62}) +[2026-06-15 21:59:54,863][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.2567291557788849, 'train/grad_norm_pre_clip': 10.013800621032715, 'timing/data': 0.022975286992732435, 'timing/model': 1.12130535999313, 'learning_rate/qwen_vl_interface': 8.19726864282782e-06, 'learning_rate/action_model': 8.197268642827818e-05, 'epoch': 0.62}) +[2026-06-15 21:59:59,388][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.14621931314468384, 'train/grad_norm_pre_clip': 44.15351867675781, 'timing/data': 0.023011497978586704, 'timing/model': 1.3007968320162036, 'learning_rate/qwen_vl_interface': 8.194879744150682e-06, 'learning_rate/action_model': 8.194879744150681e-05, 'epoch': 0.62}) +[2026-06-15 22:00:03,737][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.026630569249391556, 'train/grad_norm_pre_clip': 4.149905204772949, 'timing/data': 0.022884805977810174, 'timing/model': 1.1240462140413001, 'learning_rate/qwen_vl_interface': 8.192489634944718e-06, 'learning_rate/action_model': 8.192489634944718e-05, 'epoch': 0.62}) +[2026-06-15 22:00:08,225][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.13876639306545258, 'train/grad_norm_pre_clip': 8.108187675476074, 'timing/data': 0.023130844987463206, 'timing/model': 1.1249228590168059, 'learning_rate/qwen_vl_interface': 8.190098316192416e-06, 'learning_rate/action_model': 8.190098316192416e-05, 'epoch': 0.62}) +[2026-06-15 22:00:12,740][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.135275200009346, 'train/grad_norm_pre_clip': 6.494763374328613, 'timing/data': 0.02353270899038762, 'timing/model': 1.1240972289815545, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 0.62}) +[2026-06-15 22:00:17,249][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.05281761661171913, 'train/grad_norm_pre_clip': 5.28680944442749, 'timing/data': 0.02344496501609683, 'timing/model': 1.1281634920160286, 'learning_rate/qwen_vl_interface': 8.185312053981204e-06, 'learning_rate/action_model': 8.185312053981204e-05, 'epoch': 0.62}) +[2026-06-15 22:00:21,749][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.1948844939470291, 'train/grad_norm_pre_clip': 11.840591430664062, 'timing/data': 0.023218824993819, 'timing/model': 1.12559227400925, 'learning_rate/qwen_vl_interface': 8.182917112489747e-06, 'learning_rate/action_model': 8.182917112489746e-05, 'epoch': 0.62}) +[2026-06-15 22:00:26,280][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.3748758137226105, 'train/grad_norm_pre_clip': 13.909558296203613, 'timing/data': 0.02356642100494355, 'timing/model': 1.2973097130306996, 'learning_rate/qwen_vl_interface': 8.18052096538685e-06, 'learning_rate/action_model': 8.18052096538685e-05, 'epoch': 0.62}) +[2026-06-15 22:00:30,713][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.10337314009666443, 'train/grad_norm_pre_clip': 6.631866931915283, 'timing/data': 0.02321986702736467, 'timing/model': 1.2056276249932125, 'learning_rate/qwen_vl_interface': 8.178123613657483e-06, 'learning_rate/action_model': 8.178123613657482e-05, 'epoch': 0.62}) +[2026-06-15 22:00:35,069][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.19472570717334747, 'train/grad_norm_pre_clip': 6.1308441162109375, 'timing/data': 0.02376778202597052, 'timing/model': 1.1291101499809884, 'learning_rate/qwen_vl_interface': 8.175725058287102e-06, 'learning_rate/action_model': 8.175725058287102e-05, 'epoch': 0.62}) +[2026-06-15 22:00:39,531][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.12726156413555145, 'train/grad_norm_pre_clip': 14.090381622314453, 'timing/data': 0.023321470012888312, 'timing/model': 1.1303571099997498, 'learning_rate/qwen_vl_interface': 8.173325300261664e-06, 'learning_rate/action_model': 8.173325300261663e-05, 'epoch': 0.62}) +[2026-06-15 22:00:44,059][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.16661614179611206, 'train/grad_norm_pre_clip': 10.944064140319824, 'timing/data': 0.023782045987900347, 'timing/model': 1.1326005039736629, 'learning_rate/qwen_vl_interface': 8.170924340567617e-06, 'learning_rate/action_model': 8.170924340567616e-05, 'epoch': 0.62}) +[2026-06-15 22:00:48,586][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.17084340751171112, 'train/grad_norm_pre_clip': 7.081628799438477, 'timing/data': 0.02317656099330634, 'timing/model': 1.1286435299552977, 'learning_rate/qwen_vl_interface': 8.168522180191909e-06, 'learning_rate/action_model': 8.168522180191909e-05, 'epoch': 0.62}) +[2026-06-15 22:00:53,092][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.09638495743274689, 'train/grad_norm_pre_clip': 4.545809745788574, 'timing/data': 0.023688288987614214, 'timing/model': 1.280257894017268, 'learning_rate/qwen_vl_interface': 8.166118820121972e-06, 'learning_rate/action_model': 8.166118820121973e-05, 'epoch': 0.62}) +[2026-06-15 22:00:57,572][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.12692032754421234, 'train/grad_norm_pre_clip': 8.849933624267578, 'timing/data': 0.023322632012423128, 'timing/model': 1.24246670701541, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 0.62}) +[2026-06-15 22:01:01,933][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.06430772691965103, 'train/grad_norm_pre_clip': 3.625392436981201, 'timing/data': 0.023691842041444033, 'timing/model': 1.126252576999832, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.62}) +[2026-06-15 22:01:06,485][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.14782118797302246, 'train/grad_norm_pre_clip': 6.428240776062012, 'timing/data': 0.022650694998446852, 'timing/model': 1.1325117009691894, 'learning_rate/qwen_vl_interface': 8.158901551628579e-06, 'learning_rate/action_model': 8.158901551628579e-05, 'epoch': 0.63}) +[2026-06-15 22:01:11,026][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.15183795988559723, 'train/grad_norm_pre_clip': 12.970324516296387, 'timing/data': 0.023885199043434113, 'timing/model': 1.1326987610082142, 'learning_rate/qwen_vl_interface': 8.156493402665971e-06, 'learning_rate/action_model': 8.15649340266597e-05, 'epoch': 0.63}) +[2026-06-15 22:01:15,558][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.1693604290485382, 'train/grad_norm_pre_clip': 15.012706756591797, 'timing/data': 0.023505656979978085, 'timing/model': 1.1296424599713646, 'learning_rate/qwen_vl_interface': 8.154084058953713e-06, 'learning_rate/action_model': 8.154084058953713e-05, 'epoch': 0.63}) +[2026-06-15 22:01:20,076][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.06330312788486481, 'train/grad_norm_pre_clip': 2.2689857482910156, 'timing/data': 0.023588698008097708, 'timing/model': 1.2875566330039874, 'learning_rate/qwen_vl_interface': 8.151673521482193e-06, 'learning_rate/action_model': 8.151673521482194e-05, 'epoch': 0.63}) +[2026-06-15 22:01:24,607][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.2700658142566681, 'train/grad_norm_pre_clip': 15.212854385375977, 'timing/data': 0.02348624001024291, 'timing/model': 1.2961307669756934, 'learning_rate/qwen_vl_interface': 8.149261791242297e-06, 'learning_rate/action_model': 8.149261791242297e-05, 'epoch': 0.63}) +[2026-06-15 22:01:28,974][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.1129564642906189, 'train/grad_norm_pre_clip': 7.4262542724609375, 'timing/data': 0.023691170033998787, 'timing/model': 1.1303542929817922, 'learning_rate/qwen_vl_interface': 8.146848869225391e-06, 'learning_rate/action_model': 8.146848869225391e-05, 'epoch': 0.63}) +[2026-06-15 22:01:33,518][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.13365299999713898, 'train/grad_norm_pre_clip': 21.333229064941406, 'timing/data': 0.02389764500549063, 'timing/model': 1.1352904940140434, 'learning_rate/qwen_vl_interface': 8.14443475642334e-06, 'learning_rate/action_model': 8.144434756423339e-05, 'epoch': 0.63}) +[2026-06-15 22:01:38,045][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.0447622612118721, 'train/grad_norm_pre_clip': 4.997239589691162, 'timing/data': 0.02340857102535665, 'timing/model': 1.1294975170167163, 'learning_rate/qwen_vl_interface': 8.142019453828493e-06, 'learning_rate/action_model': 8.142019453828493e-05, 'epoch': 0.63}) +[2026-06-15 22:01:42,550][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.044484104961156845, 'train/grad_norm_pre_clip': 5.006842136383057, 'timing/data': 0.023425267019774765, 'timing/model': 1.1309973189490847, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 0.63}) +[2026-06-15 22:01:47,036][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.1792185753583908, 'train/grad_norm_pre_clip': 7.9924726486206055, 'timing/data': 0.023643676016945392, 'timing/model': 1.1290101769845933, 'learning_rate/qwen_vl_interface': 8.137185283232255e-06, 'learning_rate/action_model': 8.137185283232256e-05, 'epoch': 0.63}) +[2026-06-15 22:01:51,550][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.10806255042552948, 'train/grad_norm_pre_clip': 7.890342712402344, 'timing/data': 0.02339056896744296, 'timing/model': 1.2860359820188023, 'learning_rate/qwen_vl_interface': 8.134766417218014e-06, 'learning_rate/action_model': 8.134766417218012e-05, 'epoch': 0.63}) +[2026-06-15 22:01:55,927][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.10433219373226166, 'train/grad_norm_pre_clip': 5.928925037384033, 'timing/data': 0.023305307026021183, 'timing/model': 1.1313411010196432, 'learning_rate/qwen_vl_interface': 8.132346365385262e-06, 'learning_rate/action_model': 8.132346365385261e-05, 'epoch': 0.63}) +[2026-06-15 22:02:00,463][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.11967719346284866, 'train/grad_norm_pre_clip': 14.281140327453613, 'timing/data': 0.023569981975015253, 'timing/model': 1.1367731990176253, 'learning_rate/qwen_vl_interface': 8.129925128728796e-06, 'learning_rate/action_model': 8.129925128728796e-05, 'epoch': 0.63}) +[2026-06-15 22:02:04,949][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.010008408688008785, 'train/grad_norm_pre_clip': 1.0244765281677246, 'timing/data': 0.023116610012948513, 'timing/model': 1.142720700998325, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 0.63}) +[2026-06-15 22:02:09,399][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.08044963330030441, 'train/grad_norm_pre_clip': 5.718358039855957, 'timing/data': 0.03179642400937155, 'timing/model': 1.1349276910186745, 'learning_rate/qwen_vl_interface': 8.125079104926324e-06, 'learning_rate/action_model': 8.125079104926323e-05, 'epoch': 0.63}) +[2026-06-15 22:02:13,870][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.18904916942119598, 'train/grad_norm_pre_clip': 13.10212516784668, 'timing/data': 0.02428447501733899, 'timing/model': 1.13552431500284, 'learning_rate/qwen_vl_interface': 8.122654319772332e-06, 'learning_rate/action_model': 8.122654319772332e-05, 'epoch': 0.63}) +[2026-06-15 22:02:18,378][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.17435471713542938, 'train/grad_norm_pre_clip': 8.725767135620117, 'timing/data': 0.024140856985468417, 'timing/model': 1.2589306040317751, 'learning_rate/qwen_vl_interface': 8.120228353778661e-06, 'learning_rate/action_model': 8.12022835377866e-05, 'epoch': 0.63}) +[2026-06-15 22:02:22,756][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.03222072869539261, 'train/grad_norm_pre_clip': 4.909025192260742, 'timing/data': 0.023646215966437012, 'timing/model': 1.1304063689894974, 'learning_rate/qwen_vl_interface': 8.117801207942528e-06, 'learning_rate/action_model': 8.117801207942528e-05, 'epoch': 0.63}) +[2026-06-15 22:02:27,311][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.06811996549367905, 'train/grad_norm_pre_clip': 4.84825325012207, 'timing/data': 0.023313455982133746, 'timing/model': 1.1294270840007812, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 0.63}) +[2026-06-15 22:02:31,875][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.05818131938576698, 'train/grad_norm_pre_clip': 7.545319557189941, 'timing/data': 0.024573447997681797, 'timing/model': 1.1320533769903705, 'learning_rate/qwen_vl_interface': 8.11294338073421e-06, 'learning_rate/action_model': 8.11294338073421e-05, 'epoch': 0.63}) +[2026-06-15 22:02:36,417][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.18926724791526794, 'train/grad_norm_pre_clip': 16.964527130126953, 'timing/data': 0.02411373401992023, 'timing/model': 1.1336426479974762, 'learning_rate/qwen_vl_interface': 8.110512701358892e-06, 'learning_rate/action_model': 8.110512701358891e-05, 'epoch': 0.63}) +[2026-06-15 22:02:40,961][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.04994070529937744, 'train/grad_norm_pre_clip': 10.679800987243652, 'timing/data': 0.0235300519852899, 'timing/model': 1.2931263049831614, 'learning_rate/qwen_vl_interface': 8.108080846134852e-06, 'learning_rate/action_model': 8.108080846134852e-05, 'epoch': 0.63}) +[2026-06-15 22:02:45,337][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.11585589498281479, 'train/grad_norm_pre_clip': 10.487281799316406, 'timing/data': 0.02356972696725279, 'timing/model': 1.128715883998666, 'learning_rate/qwen_vl_interface': 8.105647816061734e-06, 'learning_rate/action_model': 8.105647816061735e-05, 'epoch': 0.63}) +[2026-06-15 22:02:49,869][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.0841313973069191, 'train/grad_norm_pre_clip': 9.570460319519043, 'timing/data': 0.02365552802802995, 'timing/model': 1.1310882789548486, 'learning_rate/qwen_vl_interface': 8.103213612139667e-06, 'learning_rate/action_model': 8.103213612139667e-05, 'epoch': 0.63}) +[2026-06-15 22:02:54,352][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.2073037475347519, 'train/grad_norm_pre_clip': 21.456228256225586, 'timing/data': 0.023148157983087003, 'timing/model': 1.135811781976372, 'learning_rate/qwen_vl_interface': 8.100778235369262e-06, 'learning_rate/action_model': 8.10077823536926e-05, 'epoch': 0.64}) +[2026-06-15 22:02:58,916][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.04421943053603172, 'train/grad_norm_pre_clip': 6.40413236618042, 'timing/data': 0.022762003005482256, 'timing/model': 1.1346164440037683, 'learning_rate/qwen_vl_interface': 8.098341686751605e-06, 'learning_rate/action_model': 8.098341686751605e-05, 'epoch': 0.64}) +[2026-06-15 22:03:03,460][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.29379042983055115, 'train/grad_norm_pre_clip': 6.6795806884765625, 'timing/data': 0.02334387303562835, 'timing/model': 1.1381637309677899, 'learning_rate/qwen_vl_interface': 8.095903967288273e-06, 'learning_rate/action_model': 8.095903967288272e-05, 'epoch': 0.64}) +[2026-06-15 22:03:08,000][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.08073300868272781, 'train/grad_norm_pre_clip': 8.52851390838623, 'timing/data': 0.023205704987049103, 'timing/model': 1.292567900964059, 'learning_rate/qwen_vl_interface': 8.093465077981318e-06, 'learning_rate/action_model': 8.093465077981318e-05, 'epoch': 0.64}) +[2026-06-15 22:03:12,525][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.16267888247966766, 'train/grad_norm_pre_clip': 13.192342758178711, 'timing/data': 0.023132795991841704, 'timing/model': 1.2557740909978747, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 0.64}) +[2026-06-15 22:03:16,952][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.04956509917974472, 'train/grad_norm_pre_clip': 4.922442436218262, 'timing/data': 0.024390746955759823, 'timing/model': 1.1522333250031807, 'learning_rate/qwen_vl_interface': 8.088583793847165e-06, 'learning_rate/action_model': 8.088583793847164e-05, 'epoch': 0.64}) +[2026-06-15 22:03:21,505][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.25820448994636536, 'train/grad_norm_pre_clip': 23.12991714477539, 'timing/data': 0.023885306029114872, 'timing/model': 1.1422832089592703, 'learning_rate/qwen_vl_interface': 8.086141401026478e-06, 'learning_rate/action_model': 8.086141401026478e-05, 'epoch': 0.64}) +[2026-06-15 22:03:26,074][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.41744688153266907, 'train/grad_norm_pre_clip': 21.03360939025879, 'timing/data': 0.023496976005844772, 'timing/model': 1.1346409490215592, 'learning_rate/qwen_vl_interface': 8.083697842375193e-06, 'learning_rate/action_model': 8.083697842375193e-05, 'epoch': 0.64}) +[2026-06-15 22:03:30,616][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.06772509217262268, 'train/grad_norm_pre_clip': 8.017776489257812, 'timing/data': 0.024040821997914463, 'timing/model': 1.1349834520369768, 'learning_rate/qwen_vl_interface': 8.081253118897762e-06, 'learning_rate/action_model': 8.081253118897761e-05, 'epoch': 0.64}) +[2026-06-15 22:03:35,155][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.28493866324424744, 'train/grad_norm_pre_clip': 10.916440963745117, 'timing/data': 0.023305795970372856, 'timing/model': 1.2964383810176514, 'learning_rate/qwen_vl_interface': 8.07880723159912e-06, 'learning_rate/action_model': 8.07880723159912e-05, 'epoch': 0.64}) +[2026-06-15 22:03:39,680][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.05363029986619949, 'train/grad_norm_pre_clip': 4.539234638214111, 'timing/data': 0.024637716996949166, 'timing/model': 1.2762003760435618, 'learning_rate/qwen_vl_interface': 8.076360181484676e-06, 'learning_rate/action_model': 8.076360181484676e-05, 'epoch': 0.64}) +[2026-06-15 22:03:44,063][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.054939817637205124, 'train/grad_norm_pre_clip': 5.177423000335693, 'timing/data': 0.023971465998329222, 'timing/model': 1.1358240160043351, 'learning_rate/qwen_vl_interface': 8.073911969560326e-06, 'learning_rate/action_model': 8.073911969560326e-05, 'epoch': 0.64}) +[2026-06-15 22:03:48,616][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.0368824303150177, 'train/grad_norm_pre_clip': 2.6057863235473633, 'timing/data': 0.023982605955097824, 'timing/model': 1.1327849180088378, 'learning_rate/qwen_vl_interface': 8.071462596832432e-06, 'learning_rate/action_model': 8.071462596832432e-05, 'epoch': 0.64}) +[2026-06-15 22:03:53,177][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.04066458344459534, 'train/grad_norm_pre_clip': 3.079718589782715, 'timing/data': 0.023211523017380387, 'timing/model': 1.1310866159619763, 'learning_rate/qwen_vl_interface': 8.06901206430784e-06, 'learning_rate/action_model': 8.06901206430784e-05, 'epoch': 0.64}) +[2026-06-15 22:03:57,722][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.10218046605587006, 'train/grad_norm_pre_clip': 5.159820556640625, 'timing/data': 0.023330616997554898, 'timing/model': 1.1334741659811698, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 0.64}) +[2026-06-15 22:04:02,226][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.18514612317085266, 'train/grad_norm_pre_clip': 10.058028221130371, 'timing/data': 0.023411582980770618, 'timing/model': 1.1350701579940505, 'learning_rate/qwen_vl_interface': 8.064107523898327e-06, 'learning_rate/action_model': 8.064107523898326e-05, 'epoch': 0.64}) +[2026-06-15 22:04:06,795][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.036002106964588165, 'train/grad_norm_pre_clip': 6.299410820007324, 'timing/data': 0.023242261027917266, 'timing/model': 1.2946180219878443, 'learning_rate/qwen_vl_interface': 8.061653518029476e-06, 'learning_rate/action_model': 8.061653518029476e-05, 'epoch': 0.64}) +[2026-06-15 22:04:11,178][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.24188938736915588, 'train/grad_norm_pre_clip': 15.926557540893555, 'timing/data': 0.02407824096735567, 'timing/model': 1.1334567760350183, 'learning_rate/qwen_vl_interface': 8.059198356396069e-06, 'learning_rate/action_model': 8.059198356396068e-05, 'epoch': 0.64}) +[2026-06-15 22:04:15,707][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.022122742608189583, 'train/grad_norm_pre_clip': 3.5524535179138184, 'timing/data': 0.02373939595418051, 'timing/model': 1.130451868986711, 'learning_rate/qwen_vl_interface': 8.05674204000733e-06, 'learning_rate/action_model': 8.05674204000733e-05, 'epoch': 0.64}) +[2026-06-15 22:04:20,155][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.06768158078193665, 'train/grad_norm_pre_clip': 5.333209037780762, 'timing/data': 0.024776291975285858, 'timing/model': 1.145796044031158, 'learning_rate/qwen_vl_interface': 8.05428456987296e-06, 'learning_rate/action_model': 8.05428456987296e-05, 'epoch': 0.64}) +[2026-06-15 22:04:24,613][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.01828913390636444, 'train/grad_norm_pre_clip': 6.770251750946045, 'timing/data': 0.023931364994496107, 'timing/model': 1.1343009229749441, 'learning_rate/qwen_vl_interface': 8.05182594700313e-06, 'learning_rate/action_model': 8.05182594700313e-05, 'epoch': 0.64}) +[2026-06-15 22:04:29,093][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.04097437486052513, 'train/grad_norm_pre_clip': 4.161377429962158, 'timing/data': 0.023962225997820497, 'timing/model': 1.1311423730221577, 'learning_rate/qwen_vl_interface': 8.049366172408484e-06, 'learning_rate/action_model': 8.049366172408485e-05, 'epoch': 0.64}) +[2026-06-15 22:04:33,616][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.11424481868743896, 'train/grad_norm_pre_clip': 9.103602409362793, 'timing/data': 0.02339514798950404, 'timing/model': 1.2734501009690575, 'learning_rate/qwen_vl_interface': 8.04690524710015e-06, 'learning_rate/action_model': 8.046905247100151e-05, 'epoch': 0.64}) +[2026-06-15 22:04:38,005][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.1347658336162567, 'train/grad_norm_pre_clip': 15.90929889678955, 'timing/data': 0.022967139026150107, 'timing/model': 1.1370643260306679, 'learning_rate/qwen_vl_interface': 8.044443172089715e-06, 'learning_rate/action_model': 8.044443172089714e-05, 'epoch': 0.64}) +[2026-06-15 22:04:42,541][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.025543566793203354, 'train/grad_norm_pre_clip': 4.890552043914795, 'timing/data': 0.022915954003110528, 'timing/model': 1.1312102500232868, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.64}) +[2026-06-15 22:04:47,107][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.003123184898868203, 'train/grad_norm_pre_clip': 0.8252868056297302, 'timing/data': 0.023681946040596813, 'timing/model': 1.1369155619759113, 'learning_rate/qwen_vl_interface': 8.039515577011287e-06, 'learning_rate/action_model': 8.039515577011288e-05, 'epoch': 0.65}) +[2026-06-15 22:04:51,662][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.250793993473053, 'train/grad_norm_pre_clip': 13.685869216918945, 'timing/data': 0.02384842495666817, 'timing/model': 1.1364789319923148, 'learning_rate/qwen_vl_interface': 8.037050058968841e-06, 'learning_rate/action_model': 8.03705005896884e-05, 'epoch': 0.65}) +[2026-06-15 22:04:56,236][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.20067402720451355, 'train/grad_norm_pre_clip': 8.871153831481934, 'timing/data': 0.02866429800633341, 'timing/model': 1.3239331240183674, 'learning_rate/qwen_vl_interface': 8.034583395275392e-06, 'learning_rate/action_model': 8.034583395275392e-05, 'epoch': 0.65}) +[2026-06-15 22:05:00,621][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.17390970885753632, 'train/grad_norm_pre_clip': 17.85704231262207, 'timing/data': 0.02334731799783185, 'timing/model': 1.1389289500075392, 'learning_rate/qwen_vl_interface': 8.032115586944893e-06, 'learning_rate/action_model': 8.032115586944892e-05, 'epoch': 0.65}) +[2026-06-15 22:05:05,166][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.16172106564044952, 'train/grad_norm_pre_clip': 14.101097106933594, 'timing/data': 0.02306822797982022, 'timing/model': 1.137663497007452, 'learning_rate/qwen_vl_interface': 8.029646634991762e-06, 'learning_rate/action_model': 8.029646634991762e-05, 'epoch': 0.65}) +[2026-06-15 22:05:09,645][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.1280662566423416, 'train/grad_norm_pre_clip': 4.287319183349609, 'timing/data': 0.023299255000893027, 'timing/model': 1.1296330739860423, 'learning_rate/qwen_vl_interface': 8.0271765404309e-06, 'learning_rate/action_model': 8.0271765404309e-05, 'epoch': 0.65}) +[2026-06-15 22:05:14,197][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.002805421594530344, 'train/grad_norm_pre_clip': 0.26776617765426636, 'timing/data': 0.023404882987961173, 'timing/model': 1.1380443829693832, 'learning_rate/qwen_vl_interface': 8.024705304277664e-06, 'learning_rate/action_model': 8.024705304277664e-05, 'epoch': 0.65}) +[2026-06-15 22:05:18,748][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.067839615046978, 'train/grad_norm_pre_clip': 8.712565422058105, 'timing/data': 0.023678296012803912, 'timing/model': 1.1352135349879973, 'learning_rate/qwen_vl_interface': 8.022232927547888e-06, 'learning_rate/action_model': 8.022232927547886e-05, 'epoch': 0.65}) +[2026-06-15 22:05:23,312][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.012047129683196545, 'train/grad_norm_pre_clip': 2.367676019668579, 'timing/data': 0.023510438040830195, 'timing/model': 1.315584137046244, 'learning_rate/qwen_vl_interface': 8.01975941125787e-06, 'learning_rate/action_model': 8.01975941125787e-05, 'epoch': 0.65}) +[2026-06-15 22:05:27,814][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.11686772108078003, 'train/grad_norm_pre_clip': 13.838106155395508, 'timing/data': 0.02294729504501447, 'timing/model': 1.2525134909665212, 'learning_rate/qwen_vl_interface': 8.017284756424381e-06, 'learning_rate/action_model': 8.017284756424381e-05, 'epoch': 0.65}) +[2026-06-15 22:05:32,206][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.39805611968040466, 'train/grad_norm_pre_clip': 23.344688415527344, 'timing/data': 0.023932177980896086, 'timing/model': 1.1329320389777422, 'learning_rate/qwen_vl_interface': 8.01480896406466e-06, 'learning_rate/action_model': 8.01480896406466e-05, 'epoch': 0.65}) +[2026-06-15 22:05:36,757][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.1867476999759674, 'train/grad_norm_pre_clip': 10.813892364501953, 'timing/data': 0.024009981018025428, 'timing/model': 1.1305497490102425, 'learning_rate/qwen_vl_interface': 8.01233203519641e-06, 'learning_rate/action_model': 8.012332035196409e-05, 'epoch': 0.65}) +[2026-06-15 22:05:41,289][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.15421205759048462, 'train/grad_norm_pre_clip': 12.112512588500977, 'timing/data': 0.02320637699449435, 'timing/model': 1.135894792038016, 'learning_rate/qwen_vl_interface': 8.009853970837801e-06, 'learning_rate/action_model': 8.009853970837801e-05, 'epoch': 0.65}) +[2026-06-15 22:05:45,829][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.26085689663887024, 'train/grad_norm_pre_clip': 26.860933303833008, 'timing/data': 0.022998692002147436, 'timing/model': 1.133854480984155, 'learning_rate/qwen_vl_interface': 8.007374772007474e-06, 'learning_rate/action_model': 8.007374772007474e-05, 'epoch': 0.65}) +[2026-06-15 22:05:50,378][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.08702093362808228, 'train/grad_norm_pre_clip': 25.717973709106445, 'timing/data': 0.022846136009320617, 'timing/model': 1.2919463320286013, 'learning_rate/qwen_vl_interface': 8.004894439724533e-06, 'learning_rate/action_model': 8.004894439724533e-05, 'epoch': 0.65}) +[2026-06-15 22:05:54,903][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.153172105550766, 'train/grad_norm_pre_clip': 19.925575256347656, 'timing/data': 0.02299404301447794, 'timing/model': 1.2804290820495225, 'learning_rate/qwen_vl_interface': 8.002412975008552e-06, 'learning_rate/action_model': 8.002412975008552e-05, 'epoch': 0.65}) +[2026-06-15 22:05:59,306][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.013993524946272373, 'train/grad_norm_pre_clip': 2.0851528644561768, 'timing/data': 0.0228229439817369, 'timing/model': 1.132767926028464, 'learning_rate/qwen_vl_interface': 7.999930378879562e-06, 'learning_rate/action_model': 7.999930378879561e-05, 'epoch': 0.65}) +[2026-06-15 22:06:03,839][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.039480242878198624, 'train/grad_norm_pre_clip': 3.2839109897613525, 'timing/data': 0.024020164972171187, 'timing/model': 1.1322855170001276, 'learning_rate/qwen_vl_interface': 7.997446652358067e-06, 'learning_rate/action_model': 7.997446652358068e-05, 'epoch': 0.65}) +[2026-06-15 22:06:08,381][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.013560594990849495, 'train/grad_norm_pre_clip': 1.0483038425445557, 'timing/data': 0.02411625999957323, 'timing/model': 1.131521616014652, 'learning_rate/qwen_vl_interface': 7.994961796465035e-06, 'learning_rate/action_model': 7.994961796465035e-05, 'epoch': 0.65}) +[2026-06-15 22:06:12,909][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.012720859609544277, 'train/grad_norm_pre_clip': 1.3091439008712769, 'timing/data': 0.02402183401864022, 'timing/model': 1.1303186640143394, 'learning_rate/qwen_vl_interface': 7.992475812221894e-06, 'learning_rate/action_model': 7.992475812221894e-05, 'epoch': 0.65}) +[2026-06-15 22:06:17,392][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.08860807865858078, 'train/grad_norm_pre_clip': 8.186128616333008, 'timing/data': 0.023273516970220953, 'timing/model': 1.1320139830349945, 'learning_rate/qwen_vl_interface': 7.989988700650538e-06, 'learning_rate/action_model': 7.989988700650538e-05, 'epoch': 0.65}) +[2026-06-15 22:06:21,944][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.10800328850746155, 'train/grad_norm_pre_clip': 6.480846881866455, 'timing/data': 0.02370278403395787, 'timing/model': 1.2964842380024493, 'learning_rate/qwen_vl_interface': 7.987500462773326e-06, 'learning_rate/action_model': 7.987500462773326e-05, 'epoch': 0.65}) +[2026-06-15 22:06:26,317][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.30572500824928284, 'train/grad_norm_pre_clip': 10.344267845153809, 'timing/data': 0.023356819001492113, 'timing/model': 1.1327046559890732, 'learning_rate/qwen_vl_interface': 7.985011099613078e-06, 'learning_rate/action_model': 7.985011099613077e-05, 'epoch': 0.65}) +[2026-06-15 22:06:30,868][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.15718325972557068, 'train/grad_norm_pre_clip': 22.47364044189453, 'timing/data': 0.023655649973079562, 'timing/model': 1.1395398799795657, 'learning_rate/qwen_vl_interface': 7.98252061219308e-06, 'learning_rate/action_model': 7.982520612193078e-05, 'epoch': 0.65}) +[2026-06-15 22:06:35,315][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.3700619637966156, 'train/grad_norm_pre_clip': 30.853479385375977, 'timing/data': 0.02329080901108682, 'timing/model': 1.1350060349795967, 'learning_rate/qwen_vl_interface': 7.980029001537073e-06, 'learning_rate/action_model': 7.980029001537072e-05, 'epoch': 0.66}) +[2026-06-15 22:06:39,753][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.11232425272464752, 'train/grad_norm_pre_clip': 17.963123321533203, 'timing/data': 0.023774113971740007, 'timing/model': 1.1392710970249027, 'learning_rate/qwen_vl_interface': 7.977536268669264e-06, 'learning_rate/action_model': 7.977536268669264e-05, 'epoch': 0.66}) +[2026-06-15 22:06:44,214][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.35098397731781006, 'train/grad_norm_pre_clip': 12.18838119506836, 'timing/data': 0.023494993045460433, 'timing/model': 1.131719172000885, 'learning_rate/qwen_vl_interface': 7.975042414614324e-06, 'learning_rate/action_model': 7.975042414614324e-05, 'epoch': 0.66}) +[2026-06-15 22:06:48,718][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.04238142818212509, 'train/grad_norm_pre_clip': 37.7275390625, 'timing/data': 0.024209425027947873, 'timing/model': 1.2566168479970656, 'learning_rate/qwen_vl_interface': 7.972547440397382e-06, 'learning_rate/action_model': 7.972547440397383e-05, 'epoch': 0.66}) +[2026-06-15 22:06:53,094][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.02394411340355873, 'train/grad_norm_pre_clip': 4.778940200805664, 'timing/data': 0.02311158599331975, 'timing/model': 1.1300442240317352, 'learning_rate/qwen_vl_interface': 7.970051347044026e-06, 'learning_rate/action_model': 7.970051347044025e-05, 'epoch': 0.66}) +[2026-06-15 22:11:15,191][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.08842092752456665, 'train/grad_norm_pre_clip': 9.874654769897461, 'eval/loss': 0.09083976596593857, 'eval/flap_precision': 0.8890429958391124, 'eval/flap_recall': 0.9891975308641975, 'eval/flap_f1': 0.9364499634769906, 'eval/flap_rate_model': 0.20027777777777778, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9364499634769906, 'timing/data': 0.023503228032495826, 'timing/model': 1.1297270340146497, 'learning_rate/qwen_vl_interface': 7.967554135580307e-06, 'learning_rate/action_model': 7.967554135580307e-05, 'epoch': 0.66}) +[2026-06-15 22:11:15,191][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1600_state +[2026-06-15 22:11:15,194][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 22:11:18,770][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1600_state/model.safetensors +[2026-06-15 22:11:28,843][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1600_state/optimizer.bin +[2026-06-15 22:11:28,843][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1600_state/sampler.bin +[2026-06-15 22:11:28,844][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1600_state/sampler_1.bin +[2026-06-15 22:11:28,844][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1600_state/random_states_0.pkl +[2026-06-15 22:11:28,893][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1600 +[2026-06-15 22:11:28,893][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 22:11:28,900][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 22:11:33,322][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.01234141830354929, 'train/grad_norm_pre_clip': 3.489948272705078, 'timing/data': 0.02351880300557241, 'timing/model': 1.1094126669922844, 'learning_rate/qwen_vl_interface': 7.965055807032734e-06, 'learning_rate/action_model': 7.965055807032735e-05, 'epoch': 0.66}) +[2026-06-15 22:11:37,790][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.018212806433439255, 'train/grad_norm_pre_clip': 3.31351637840271, 'timing/data': 0.0234708430361934, 'timing/model': 1.1118789149913937, 'learning_rate/qwen_vl_interface': 7.962556362428275e-06, 'learning_rate/action_model': 7.962556362428274e-05, 'epoch': 0.66}) +[2026-06-15 22:11:42,208][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.008880696259438992, 'train/grad_norm_pre_clip': 1.6350616216659546, 'timing/data': 0.023709608998615295, 'timing/model': 1.1142357920180075, 'learning_rate/qwen_vl_interface': 7.960055802794358e-06, 'learning_rate/action_model': 7.960055802794357e-05, 'epoch': 0.66}) +[2026-06-15 22:11:46,652][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.05835244804620743, 'train/grad_norm_pre_clip': 8.062134742736816, 'timing/data': 0.023856921994592994, 'timing/model': 1.1146867420175113, 'learning_rate/qwen_vl_interface': 7.957554129158866e-06, 'learning_rate/action_model': 7.957554129158866e-05, 'epoch': 0.66}) +[2026-06-15 22:11:51,124][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.4141250252723694, 'train/grad_norm_pre_clip': 27.273468017578125, 'timing/data': 0.023386331973597407, 'timing/model': 1.2706098550115712, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 0.66}) +[2026-06-15 22:11:55,448][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.1934727132320404, 'train/grad_norm_pre_clip': 9.57563304901123, 'timing/data': 0.023426355968695134, 'timing/model': 1.1198704519774765, 'learning_rate/qwen_vl_interface': 7.952547443997e-06, 'learning_rate/action_model': 7.952547443997e-05, 'epoch': 0.66}) +[2026-06-15 22:11:59,936][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.09119737148284912, 'train/grad_norm_pre_clip': 21.31646728515625, 'timing/data': 0.02320312400115654, 'timing/model': 1.1219314689515159, 'learning_rate/qwen_vl_interface': 7.950042434528681e-06, 'learning_rate/action_model': 7.950042434528682e-05, 'epoch': 0.66}) +[2026-06-15 22:12:04,428][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.20235677063465118, 'train/grad_norm_pre_clip': 10.8417329788208, 'timing/data': 0.023475457972381264, 'timing/model': 1.127593084005639, 'learning_rate/qwen_vl_interface': 7.94753631517491e-06, 'learning_rate/action_model': 7.947536315174909e-05, 'epoch': 0.66}) +[2026-06-15 22:12:08,920][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.1805216372013092, 'train/grad_norm_pre_clip': 6.569712162017822, 'timing/data': 0.024001070007216185, 'timing/model': 1.1211021419730969, 'learning_rate/qwen_vl_interface': 7.945029086965854e-06, 'learning_rate/action_model': 7.945029086965853e-05, 'epoch': 0.66}) +[2026-06-15 22:12:13,422][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.019436044618487358, 'train/grad_norm_pre_clip': 3.500087022781372, 'timing/data': 0.02406687702750787, 'timing/model': 1.276968638994731, 'learning_rate/qwen_vl_interface': 7.942520750932138e-06, 'learning_rate/action_model': 7.942520750932138e-05, 'epoch': 0.66}) +[2026-06-15 22:12:17,776][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.42466142773628235, 'train/grad_norm_pre_clip': 33.5840950012207, 'timing/data': 0.023081701016053557, 'timing/model': 1.1274447389878333, 'learning_rate/qwen_vl_interface': 7.940011308104848e-06, 'learning_rate/action_model': 7.940011308104848e-05, 'epoch': 0.66}) +[2026-06-15 22:12:22,282][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.22162017226219177, 'train/grad_norm_pre_clip': 12.976306915283203, 'timing/data': 0.02359002799494192, 'timing/model': 1.123898497025948, 'learning_rate/qwen_vl_interface': 7.93750075951552e-06, 'learning_rate/action_model': 7.93750075951552e-05, 'epoch': 0.66}) +[2026-06-15 22:12:26,804][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.17542517185211182, 'train/grad_norm_pre_clip': 11.1781587600708, 'timing/data': 0.022892577049788088, 'timing/model': 1.1372558229486458, 'learning_rate/qwen_vl_interface': 7.934989106196145e-06, 'learning_rate/action_model': 7.934989106196146e-05, 'epoch': 0.66}) +[2026-06-15 22:12:31,340][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.08803137391805649, 'train/grad_norm_pre_clip': 23.463876724243164, 'timing/data': 0.023223422991577536, 'timing/model': 1.139006984012667, 'learning_rate/qwen_vl_interface': 7.93247634917917e-06, 'learning_rate/action_model': 7.93247634917917e-05, 'epoch': 0.66}) +[2026-06-15 22:12:35,877][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.1940007358789444, 'train/grad_norm_pre_clip': 10.864107131958008, 'timing/data': 0.023549258999992162, 'timing/model': 1.1323802109691314, 'learning_rate/qwen_vl_interface': 7.929962489497494e-06, 'learning_rate/action_model': 7.929962489497493e-05, 'epoch': 0.66}) +[2026-06-15 22:12:40,389][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.021346768364310265, 'train/grad_norm_pre_clip': 4.6083807945251465, 'timing/data': 0.02350197301711887, 'timing/model': 1.2862162049859762, 'learning_rate/qwen_vl_interface': 7.927447528184469e-06, 'learning_rate/action_model': 7.927447528184468e-05, 'epoch': 0.66}) +[2026-06-15 22:12:44,735][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.040222782641649246, 'train/grad_norm_pre_clip': 6.171562671661377, 'timing/data': 0.02344740799162537, 'timing/model': 1.1248507370473817, 'learning_rate/qwen_vl_interface': 7.924931466273901e-06, 'learning_rate/action_model': 7.924931466273902e-05, 'epoch': 0.66}) +[2026-06-15 22:12:49,240][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.01605970785021782, 'train/grad_norm_pre_clip': 1.4507770538330078, 'timing/data': 0.023103090992663056, 'timing/model': 1.1254050720017403, 'learning_rate/qwen_vl_interface': 7.92241430480005e-06, 'learning_rate/action_model': 7.922414304800049e-05, 'epoch': 0.66}) +[2026-06-15 22:12:53,765][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.07222698628902435, 'train/grad_norm_pre_clip': 7.75499963760376, 'timing/data': 0.023238681023940444, 'timing/model': 1.1262138529564254, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.67}) +[2026-06-15 22:12:58,261][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.1376485824584961, 'train/grad_norm_pre_clip': 10.09953498840332, 'timing/data': 0.0231826159870252, 'timing/model': 1.129702055011876, 'learning_rate/qwen_vl_interface': 7.91737668730178e-06, 'learning_rate/action_model': 7.91737668730178e-05, 'epoch': 0.67}) +[2026-06-15 22:13:02,776][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.11693167686462402, 'train/grad_norm_pre_clip': 7.977363109588623, 'timing/data': 0.0234006829559803, 'timing/model': 1.1280402580159716, 'learning_rate/qwen_vl_interface': 7.91485623334814e-06, 'learning_rate/action_model': 7.91485623334814e-05, 'epoch': 0.67}) +[2026-06-15 22:13:07,322][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.1664847433567047, 'train/grad_norm_pre_clip': 2.546128511428833, 'timing/data': 0.023430582019500434, 'timing/model': 1.3003505069646053, 'learning_rate/qwen_vl_interface': 7.91233468397276e-06, 'learning_rate/action_model': 7.912334683972761e-05, 'epoch': 0.67}) +[2026-06-15 22:13:11,696][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.015679679811000824, 'train/grad_norm_pre_clip': 2.5324227809906006, 'timing/data': 0.023285869043320417, 'timing/model': 1.1385526389931329, 'learning_rate/qwen_vl_interface': 7.90981204021216e-06, 'learning_rate/action_model': 7.90981204021216e-05, 'epoch': 0.67}) +[2026-06-15 22:13:16,208][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.286335289478302, 'train/grad_norm_pre_clip': 32.48786163330078, 'timing/data': 0.022978354012593627, 'timing/model': 1.1271201490308158, 'learning_rate/qwen_vl_interface': 7.907288303103298e-06, 'learning_rate/action_model': 7.907288303103298e-05, 'epoch': 0.67}) +[2026-06-15 22:13:20,621][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.19935359060764313, 'train/grad_norm_pre_clip': 14.815876960754395, 'timing/data': 0.024965747026726604, 'timing/model': 1.1252499229740351, 'learning_rate/qwen_vl_interface': 7.90476347368359e-06, 'learning_rate/action_model': 7.90476347368359e-05, 'epoch': 0.67}) +[2026-06-15 22:13:25,178][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.03456271067261696, 'train/grad_norm_pre_clip': 7.522562503814697, 'timing/data': 0.0232432649936527, 'timing/model': 1.131266816984862, 'learning_rate/qwen_vl_interface': 7.902237552990897e-06, 'learning_rate/action_model': 7.902237552990897e-05, 'epoch': 0.67}) +[2026-06-15 22:13:29,685][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.1918971687555313, 'train/grad_norm_pre_clip': 31.68054962158203, 'timing/data': 0.02356354595394805, 'timing/model': 1.1274235749733634, 'learning_rate/qwen_vl_interface': 7.89971054206353e-06, 'learning_rate/action_model': 7.89971054206353e-05, 'epoch': 0.67}) +[2026-06-15 22:13:34,215][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.029105020686984062, 'train/grad_norm_pre_clip': 7.038716793060303, 'timing/data': 0.022912506014108658, 'timing/model': 1.2962735840119421, 'learning_rate/qwen_vl_interface': 7.897182441940247e-06, 'learning_rate/action_model': 7.897182441940247e-05, 'epoch': 0.67}) +[2026-06-15 22:13:38,591][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.013047060929238796, 'train/grad_norm_pre_clip': 0.9774639010429382, 'timing/data': 0.022939964022953063, 'timing/model': 1.1318471599952318, 'learning_rate/qwen_vl_interface': 7.894653253660256e-06, 'learning_rate/action_model': 7.894653253660256e-05, 'epoch': 0.67}) +[2026-06-15 22:13:43,140][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.3424157500267029, 'train/grad_norm_pre_clip': 17.54483985900879, 'timing/data': 0.02347359899431467, 'timing/model': 1.144730374973733, 'learning_rate/qwen_vl_interface': 7.89212297826321e-06, 'learning_rate/action_model': 7.892122978263209e-05, 'epoch': 0.67}) +[2026-06-15 22:13:47,703][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.08510775864124298, 'train/grad_norm_pre_clip': 9.138022422790527, 'timing/data': 0.023451530025340617, 'timing/model': 1.1386918000061996, 'learning_rate/qwen_vl_interface': 7.889591616789206e-06, 'learning_rate/action_model': 7.889591616789206e-05, 'epoch': 0.67}) +[2026-06-15 22:13:52,199][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.02827487513422966, 'train/grad_norm_pre_clip': 3.2921676635742188, 'timing/data': 0.027753485017456114, 'timing/model': 1.1378722049994394, 'learning_rate/qwen_vl_interface': 7.887059170278799e-06, 'learning_rate/action_model': 7.887059170278797e-05, 'epoch': 0.67}) +[2026-06-15 22:13:56,701][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.024001773446798325, 'train/grad_norm_pre_clip': 2.2498178482055664, 'timing/data': 0.02432588703231886, 'timing/model': 1.1294764230260625, 'learning_rate/qwen_vl_interface': 7.884525639772974e-06, 'learning_rate/action_model': 7.884525639772974e-05, 'epoch': 0.67}) +[2026-06-15 22:14:01,231][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.180365651845932, 'train/grad_norm_pre_clip': 8.159831047058105, 'timing/data': 0.024032622983213514, 'timing/model': 1.292608616000507, 'learning_rate/qwen_vl_interface': 7.881991026313176e-06, 'learning_rate/action_model': 7.881991026313175e-05, 'epoch': 0.67}) +[2026-06-15 22:14:05,595][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.006963028106838465, 'train/grad_norm_pre_clip': 0.5150421857833862, 'timing/data': 0.023012317018583417, 'timing/model': 1.1333493129932322, 'learning_rate/qwen_vl_interface': 7.879455330941283e-06, 'learning_rate/action_model': 7.879455330941282e-05, 'epoch': 0.67}) +[2026-06-15 22:14:10,144][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.13023239374160767, 'train/grad_norm_pre_clip': 5.781709671020508, 'timing/data': 0.024401943024713546, 'timing/model': 1.129583663016092, 'learning_rate/qwen_vl_interface': 7.876918554699627e-06, 'learning_rate/action_model': 7.876918554699627e-05, 'epoch': 0.67}) +[2026-06-15 22:14:14,693][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.3256734311580658, 'train/grad_norm_pre_clip': 29.325956344604492, 'timing/data': 0.02339798997854814, 'timing/model': 1.131292910024058, 'learning_rate/qwen_vl_interface': 7.874380698630982e-06, 'learning_rate/action_model': 7.874380698630982e-05, 'epoch': 0.67}) +[2026-06-15 22:14:19,225][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.13696518540382385, 'train/grad_norm_pre_clip': 10.418132781982422, 'timing/data': 0.023313128971494734, 'timing/model': 1.1296299999812618, 'learning_rate/qwen_vl_interface': 7.871841763778562e-06, 'learning_rate/action_model': 7.871841763778563e-05, 'epoch': 0.67}) +[2026-06-15 22:14:23,755][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.33867257833480835, 'train/grad_norm_pre_clip': 22.34764862060547, 'timing/data': 0.02280582202365622, 'timing/model': 1.2878984710550867, 'learning_rate/qwen_vl_interface': 7.86930175118603e-06, 'learning_rate/action_model': 7.86930175118603e-05, 'epoch': 0.67}) +[2026-06-15 22:14:28,140][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.16159449517726898, 'train/grad_norm_pre_clip': 12.501763343811035, 'timing/data': 0.02299976401263848, 'timing/model': 1.136463523027487, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.67}) +[2026-06-15 22:14:32,668][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.03530997037887573, 'train/grad_norm_pre_clip': 10.81127643585205, 'timing/data': 0.022897772025316954, 'timing/model': 1.1334114100318402, 'learning_rate/qwen_vl_interface': 7.86421849695748e-06, 'learning_rate/action_model': 7.864218496957481e-05, 'epoch': 0.67}) +[2026-06-15 22:14:37,185][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.05779285356402397, 'train/grad_norm_pre_clip': 10.422533988952637, 'timing/data': 0.023219318012706935, 'timing/model': 1.1289678280008957, 'learning_rate/qwen_vl_interface': 7.861675257410998e-06, 'learning_rate/action_model': 7.861675257410997e-05, 'epoch': 0.67}) +[2026-06-15 22:14:41,760][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.36488136649131775, 'train/grad_norm_pre_clip': 26.031803131103516, 'timing/data': 0.02329317497787997, 'timing/model': 1.1360492060193792, 'learning_rate/qwen_vl_interface': 7.859130944303468e-06, 'learning_rate/action_model': 7.859130944303467e-05, 'epoch': 0.67}) +[2026-06-15 22:14:46,312][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.006015135440975428, 'train/grad_norm_pre_clip': 0.42954134941101074, 'timing/data': 0.023120584024582058, 'timing/model': 1.1340845709783025, 'learning_rate/qwen_vl_interface': 7.856585558680761e-06, 'learning_rate/action_model': 7.856585558680761e-05, 'epoch': 0.68}) +[2026-06-15 22:14:50,844][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.16802462935447693, 'train/grad_norm_pre_clip': 10.270514488220215, 'timing/data': 0.023762843979056925, 'timing/model': 1.283151785028167, 'learning_rate/qwen_vl_interface': 7.854039101589191e-06, 'learning_rate/action_model': 7.854039101589191e-05, 'epoch': 0.68}) +[2026-06-15 22:14:55,204][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.2981880009174347, 'train/grad_norm_pre_clip': 9.931014060974121, 'timing/data': 0.023097875993698835, 'timing/model': 1.1269247259479016, 'learning_rate/qwen_vl_interface': 7.851491574075509e-06, 'learning_rate/action_model': 7.851491574075509e-05, 'epoch': 0.68}) +[2026-06-15 22:14:59,732][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.06243935227394104, 'train/grad_norm_pre_clip': 10.804390907287598, 'timing/data': 0.02298147603869438, 'timing/model': 1.1303942059748806, 'learning_rate/qwen_vl_interface': 7.848942977186909e-06, 'learning_rate/action_model': 7.848942977186908e-05, 'epoch': 0.68}) +[2026-06-15 22:15:04,292][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.171024888753891, 'train/grad_norm_pre_clip': 7.865821361541748, 'timing/data': 0.023625153000466526, 'timing/model': 1.1374224310275167, 'learning_rate/qwen_vl_interface': 7.84639331197102e-06, 'learning_rate/action_model': 7.84639331197102e-05, 'epoch': 0.68}) +[2026-06-15 22:15:08,803][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.1752878576517105, 'train/grad_norm_pre_clip': 10.659137725830078, 'timing/data': 0.02333699498558417, 'timing/model': 1.1340673330123536, 'learning_rate/qwen_vl_interface': 7.843842579475918e-06, 'learning_rate/action_model': 7.843842579475918e-05, 'epoch': 0.68}) +[2026-06-15 22:15:13,405][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.11304806172847748, 'train/grad_norm_pre_clip': 5.947685241699219, 'timing/data': 0.023862523026764393, 'timing/model': 1.14093617600156, 'learning_rate/qwen_vl_interface': 7.841290780750107e-06, 'learning_rate/action_model': 7.841290780750107e-05, 'epoch': 0.68}) +[2026-06-15 22:15:17,960][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.2827519178390503, 'train/grad_norm_pre_clip': 20.579944610595703, 'timing/data': 0.031430445960722864, 'timing/model': 1.2964479830116034, 'learning_rate/qwen_vl_interface': 7.838737916842538e-06, 'learning_rate/action_model': 7.838737916842538e-05, 'epoch': 0.68}) +[2026-06-15 22:15:22,354][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.049887001514434814, 'train/grad_norm_pre_clip': 2.9593801498413086, 'timing/data': 0.023620485968422145, 'timing/model': 1.1444147839793004, 'learning_rate/qwen_vl_interface': 7.836183988802596e-06, 'learning_rate/action_model': 7.836183988802595e-05, 'epoch': 0.68}) +[2026-06-15 22:15:26,906][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.168748140335083, 'train/grad_norm_pre_clip': 10.233329772949219, 'timing/data': 0.023323790985159576, 'timing/model': 1.1368627900374122, 'learning_rate/qwen_vl_interface': 7.833628997680107e-06, 'learning_rate/action_model': 7.833628997680107e-05, 'epoch': 0.68}) +[2026-06-15 22:15:31,365][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.10739126801490784, 'train/grad_norm_pre_clip': 17.15784454345703, 'timing/data': 0.02349648904055357, 'timing/model': 1.138059155025985, 'learning_rate/qwen_vl_interface': 7.831072944525329e-06, 'learning_rate/action_model': 7.831072944525328e-05, 'epoch': 0.68}) +[2026-06-15 22:15:35,904][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.0444231815636158, 'train/grad_norm_pre_clip': 2.2524538040161133, 'timing/data': 0.023175523965619504, 'timing/model': 1.1426015910110436, 'learning_rate/qwen_vl_interface': 7.828515830388956e-06, 'learning_rate/action_model': 7.828515830388956e-05, 'epoch': 0.68}) +[2026-06-15 22:15:40,474][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.041187021881341934, 'train/grad_norm_pre_clip': 3.8447952270507812, 'timing/data': 0.028412867977749556, 'timing/model': 1.1370232129702345, 'learning_rate/qwen_vl_interface': 7.825957656322129e-06, 'learning_rate/action_model': 7.825957656322128e-05, 'epoch': 0.68}) +[2026-06-15 22:15:45,030][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.1079709529876709, 'train/grad_norm_pre_clip': 5.660069465637207, 'timing/data': 0.023393896990455687, 'timing/model': 1.3045455139945261, 'learning_rate/qwen_vl_interface': 7.823398423376406e-06, 'learning_rate/action_model': 7.823398423376406e-05, 'epoch': 0.68}) +[2026-06-15 22:15:49,398][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.007324996870011091, 'train/grad_norm_pre_clip': 1.400911569595337, 'timing/data': 0.0240918819908984, 'timing/model': 1.130753679026384, 'learning_rate/qwen_vl_interface': 7.820838132603801e-06, 'learning_rate/action_model': 7.8208381326038e-05, 'epoch': 0.68}) +[2026-06-15 22:15:53,952][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.22012731432914734, 'train/grad_norm_pre_clip': 12.27085018157959, 'timing/data': 0.024106845958158374, 'timing/model': 1.1344114260282367, 'learning_rate/qwen_vl_interface': 7.818276785056748e-06, 'learning_rate/action_model': 7.818276785056749e-05, 'epoch': 0.68}) +[2026-06-15 22:15:58,477][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.02407972887158394, 'train/grad_norm_pre_clip': 2.80252742767334, 'timing/data': 0.023735221999231726, 'timing/model': 1.1305582959903404, 'learning_rate/qwen_vl_interface': 7.815714381788123e-06, 'learning_rate/action_model': 7.815714381788122e-05, 'epoch': 0.68}) +[2026-06-15 22:16:02,967][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.22294361889362335, 'train/grad_norm_pre_clip': 14.549593925476074, 'timing/data': 0.022916912974324077, 'timing/model': 1.1285118760424666, 'learning_rate/qwen_vl_interface': 7.813150923851227e-06, 'learning_rate/action_model': 7.813150923851227e-05, 'epoch': 0.68}) +[2026-06-15 22:16:07,503][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.11753130704164505, 'train/grad_norm_pre_clip': 8.586052894592285, 'timing/data': 0.0233625530381687, 'timing/model': 1.1329942329903133, 'learning_rate/qwen_vl_interface': 7.810586412299807e-06, 'learning_rate/action_model': 7.810586412299806e-05, 'epoch': 0.68}) +[2026-06-15 22:16:12,069][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.045882027596235275, 'train/grad_norm_pre_clip': 10.316091537475586, 'timing/data': 0.023120899975765496, 'timing/model': 1.30404987197835, 'learning_rate/qwen_vl_interface': 7.808020848188034e-06, 'learning_rate/action_model': 7.808020848188034e-05, 'epoch': 0.68}) +[2026-06-15 22:16:16,455][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.017940063029527664, 'train/grad_norm_pre_clip': 2.079096555709839, 'timing/data': 0.0234905980178155, 'timing/model': 1.1337927299900912, 'learning_rate/qwen_vl_interface': 7.805454232570517e-06, 'learning_rate/action_model': 7.805454232570517e-05, 'epoch': 0.68}) +[2026-06-15 22:16:20,975][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.05378813296556473, 'train/grad_norm_pre_clip': 7.555394172668457, 'timing/data': 0.023022470995783806, 'timing/model': 1.1311993629788049, 'learning_rate/qwen_vl_interface': 7.80288656650229e-06, 'learning_rate/action_model': 7.80288656650229e-05, 'epoch': 0.68}) +[2026-06-15 22:16:25,515][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.007829994894564152, 'train/grad_norm_pre_clip': 0.6465619802474976, 'timing/data': 0.02310754096833989, 'timing/model': 1.130973068997264, 'learning_rate/qwen_vl_interface': 7.800317851038829e-06, 'learning_rate/action_model': 7.800317851038828e-05, 'epoch': 0.68}) +[2026-06-15 22:16:30,049][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.046333424746990204, 'train/grad_norm_pre_clip': 7.580240249633789, 'timing/data': 0.023351911979261786, 'timing/model': 1.131068104004953, 'learning_rate/qwen_vl_interface': 7.797748087236033e-06, 'learning_rate/action_model': 7.797748087236033e-05, 'epoch': 0.68}) +[2026-06-15 22:16:34,591][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.03682885318994522, 'train/grad_norm_pre_clip': 4.917507171630859, 'timing/data': 0.02325835300143808, 'timing/model': 1.3034394250134937, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.69}) +[2026-06-15 22:16:38,971][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.3313634395599365, 'train/grad_norm_pre_clip': 13.421341896057129, 'timing/data': 0.02269123197766021, 'timing/model': 1.1314590790425427, 'learning_rate/qwen_vl_interface': 7.792605418838196e-06, 'learning_rate/action_model': 7.792605418838197e-05, 'epoch': 0.69}) +[2026-06-15 22:16:43,519][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.05824773386120796, 'train/grad_norm_pre_clip': 7.179236888885498, 'timing/data': 0.02357535099145025, 'timing/model': 1.134053991001565, 'learning_rate/qwen_vl_interface': 7.790032516357116e-06, 'learning_rate/action_model': 7.790032516357117e-05, 'epoch': 0.69}) +[2026-06-15 22:16:48,072][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.11415667831897736, 'train/grad_norm_pre_clip': 13.00787353515625, 'timing/data': 0.024022718018386513, 'timing/model': 1.1331251839874312, 'learning_rate/qwen_vl_interface': 7.787458569764613e-06, 'learning_rate/action_model': 7.787458569764612e-05, 'epoch': 0.69}) +[2026-06-15 22:16:52,615][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.2739599347114563, 'train/grad_norm_pre_clip': 14.445539474487305, 'timing/data': 0.0231317889993079, 'timing/model': 1.1310909279854968, 'learning_rate/qwen_vl_interface': 7.784883580118738e-06, 'learning_rate/action_model': 7.784883580118738e-05, 'epoch': 0.69}) +[2026-06-15 22:16:57,188][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.021631984040141106, 'train/grad_norm_pre_clip': 9.074789047241211, 'timing/data': 0.02333807898685336, 'timing/model': 1.1324363150051795, 'learning_rate/qwen_vl_interface': 7.782307548477975e-06, 'learning_rate/action_model': 7.782307548477975e-05, 'epoch': 0.69}) +[2026-06-15 22:17:01,722][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.22258535027503967, 'train/grad_norm_pre_clip': 15.51785945892334, 'timing/data': 0.02417085401248187, 'timing/model': 1.2911883359774947, 'learning_rate/qwen_vl_interface': 7.779730475901233e-06, 'learning_rate/action_model': 7.779730475901232e-05, 'epoch': 0.69}) +[2026-06-15 22:17:06,105][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.009495334699749947, 'train/grad_norm_pre_clip': 2.3574249744415283, 'timing/data': 0.023692222021054476, 'timing/model': 1.135207156999968, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 0.69}) +[2026-06-15 22:17:10,651][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.2115325778722763, 'train/grad_norm_pre_clip': 41.3411750793457, 'timing/data': 0.02358524495502934, 'timing/model': 1.133963202999439, 'learning_rate/qwen_vl_interface': 7.774573212177587e-06, 'learning_rate/action_model': 7.774573212177588e-05, 'epoch': 0.69}) +[2026-06-15 22:17:15,251][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.08322322368621826, 'train/grad_norm_pre_clip': 13.559100151062012, 'timing/data': 0.02334813302149996, 'timing/model': 1.1414727560477331, 'learning_rate/qwen_vl_interface': 7.77199302315064e-06, 'learning_rate/action_model': 7.77199302315064e-05, 'epoch': 0.69}) +[2026-06-15 22:17:19,757][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.18530067801475525, 'train/grad_norm_pre_clip': 13.741106033325195, 'timing/data': 0.030781507026404142, 'timing/model': 1.1327521089697257, 'learning_rate/qwen_vl_interface': 7.769411797427622e-06, 'learning_rate/action_model': 7.769411797427623e-05, 'epoch': 0.69}) +[2026-06-15 22:17:24,317][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.007711481768637896, 'train/grad_norm_pre_clip': 2.2870852947235107, 'timing/data': 0.023753479996230453, 'timing/model': 1.145368050027173, 'learning_rate/qwen_vl_interface': 7.766829536069584e-06, 'learning_rate/action_model': 7.766829536069584e-05, 'epoch': 0.69}) +[2026-06-15 22:17:28,853][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.012925060465931892, 'train/grad_norm_pre_clip': 4.3094635009765625, 'timing/data': 0.023198824957944453, 'timing/model': 1.2930221860297024, 'learning_rate/qwen_vl_interface': 7.764246240137991e-06, 'learning_rate/action_model': 7.764246240137991e-05, 'epoch': 0.69}) +[2026-06-15 22:17:33,248][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.055999755859375, 'train/grad_norm_pre_clip': 14.081754684448242, 'timing/data': 0.023397414945065975, 'timing/model': 1.1346307269996032, 'learning_rate/qwen_vl_interface': 7.761661910694741e-06, 'learning_rate/action_model': 7.761661910694741e-05, 'epoch': 0.69}) +[2026-06-15 22:17:37,802][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.07136417180299759, 'train/grad_norm_pre_clip': 14.919540405273438, 'timing/data': 0.02282062399899587, 'timing/model': 1.143034214968793, 'learning_rate/qwen_vl_interface': 7.759076548802153e-06, 'learning_rate/action_model': 7.759076548802152e-05, 'epoch': 0.69}) +[2026-06-15 22:17:42,253][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.3085455596446991, 'train/grad_norm_pre_clip': 25.577917098999023, 'timing/data': 0.02401247003581375, 'timing/model': 1.132185130030848, 'learning_rate/qwen_vl_interface': 7.75649015552297e-06, 'learning_rate/action_model': 7.75649015552297e-05, 'epoch': 0.69}) +[2026-06-15 22:17:46,790][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.0018134299898520112, 'train/grad_norm_pre_clip': 0.17409838736057281, 'timing/data': 0.022863852966111153, 'timing/model': 1.1307073819916695, 'learning_rate/qwen_vl_interface': 7.753902731920364e-06, 'learning_rate/action_model': 7.753902731920364e-05, 'epoch': 0.69}) +[2026-06-15 22:17:51,326][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.006980353035032749, 'train/grad_norm_pre_clip': 1.4697496891021729, 'timing/data': 0.02418694895459339, 'timing/model': 1.1314842270221561, 'learning_rate/qwen_vl_interface': 7.751314279057925e-06, 'learning_rate/action_model': 7.751314279057923e-05, 'epoch': 0.69}) +[2026-06-15 22:17:55,865][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.381041020154953, 'train/grad_norm_pre_clip': 13.768538475036621, 'timing/data': 0.023224356002174318, 'timing/model': 1.2973507479764521, 'learning_rate/qwen_vl_interface': 7.74872479799967e-06, 'learning_rate/action_model': 7.74872479799967e-05, 'epoch': 0.69}) +[2026-06-15 22:18:00,260][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.08470844477415085, 'train/grad_norm_pre_clip': 9.714917182922363, 'timing/data': 0.023531546001322567, 'timing/model': 1.1350412360043265, 'learning_rate/qwen_vl_interface': 7.746134289810034e-06, 'learning_rate/action_model': 7.746134289810033e-05, 'epoch': 0.69}) +[2026-06-15 22:18:04,809][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.24478018283843994, 'train/grad_norm_pre_clip': 12.201406478881836, 'timing/data': 0.02271638798993081, 'timing/model': 1.1299419169663452, 'learning_rate/qwen_vl_interface': 7.74354275555388e-06, 'learning_rate/action_model': 7.74354275555388e-05, 'epoch': 0.69}) +[2026-06-15 22:18:09,369][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.054994210600852966, 'train/grad_norm_pre_clip': 9.303926467895508, 'timing/data': 0.02376148698385805, 'timing/model': 1.1333035869756714, 'learning_rate/qwen_vl_interface': 7.740950196296488e-06, 'learning_rate/action_model': 7.740950196296488e-05, 'epoch': 0.69}) +[2026-06-15 22:18:13,875][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.1111949011683464, 'train/grad_norm_pre_clip': 19.7788028717041, 'timing/data': 0.024073657987173647, 'timing/model': 1.1320343849947676, 'learning_rate/qwen_vl_interface': 7.738356613103562e-06, 'learning_rate/action_model': 7.73835661310356e-05, 'epoch': 0.69}) +[2026-06-15 22:18:18,400][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.21365268528461456, 'train/grad_norm_pre_clip': 16.329727172851562, 'timing/data': 0.023321660002693534, 'timing/model': 1.1356975340167992, 'learning_rate/qwen_vl_interface': 7.735762007041226e-06, 'learning_rate/action_model': 7.735762007041226e-05, 'epoch': 0.69}) +[2026-06-15 22:18:22,954][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.0835682824254036, 'train/grad_norm_pre_clip': 22.31548309326172, 'timing/data': 0.023053637996781617, 'timing/model': 1.2974572929670103, 'learning_rate/qwen_vl_interface': 7.733166379176023e-06, 'learning_rate/action_model': 7.733166379176023e-05, 'epoch': 0.7}) +[2026-06-15 22:18:27,337][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.20310844480991364, 'train/grad_norm_pre_clip': 7.818749904632568, 'timing/data': 0.024258601013571024, 'timing/model': 1.1336758839897811, 'learning_rate/qwen_vl_interface': 7.730569730574921e-06, 'learning_rate/action_model': 7.73056973057492e-05, 'epoch': 0.7}) +[2026-06-15 22:18:31,862][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.02385556325316429, 'train/grad_norm_pre_clip': 11.758999824523926, 'timing/data': 0.023392681032419205, 'timing/model': 1.133141283004079, 'learning_rate/qwen_vl_interface': 7.727972062305299e-06, 'learning_rate/action_model': 7.727972062305299e-05, 'epoch': 0.7}) +[2026-06-15 22:18:36,393][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.03404349833726883, 'train/grad_norm_pre_clip': 2.487236738204956, 'timing/data': 0.02368067397037521, 'timing/model': 1.131935610959772, 'learning_rate/qwen_vl_interface': 7.725373375434965e-06, 'learning_rate/action_model': 7.725373375434965e-05, 'epoch': 0.7}) +[2026-06-15 22:18:40,920][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.168190136551857, 'train/grad_norm_pre_clip': 11.387242317199707, 'timing/data': 0.023758053954225034, 'timing/model': 1.1319075970095582, 'learning_rate/qwen_vl_interface': 7.722773671032141e-06, 'learning_rate/action_model': 7.722773671032142e-05, 'epoch': 0.7}) +[2026-06-15 22:18:45,463][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.09614304453134537, 'train/grad_norm_pre_clip': 6.878602504730225, 'timing/data': 0.023420837009325624, 'timing/model': 1.2916723599773832, 'learning_rate/qwen_vl_interface': 7.720172950165466e-06, 'learning_rate/action_model': 7.720172950165465e-05, 'epoch': 0.7}) +[2026-06-15 22:18:49,837][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.022464772686362267, 'train/grad_norm_pre_clip': 4.213348865509033, 'timing/data': 0.023591875040438026, 'timing/model': 1.132884526043199, 'learning_rate/qwen_vl_interface': 7.717571213903996e-06, 'learning_rate/action_model': 7.717571213903996e-05, 'epoch': 0.7}) +[2026-06-15 22:18:54,370][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.2634144425392151, 'train/grad_norm_pre_clip': 14.625659942626953, 'timing/data': 0.02376599097624421, 'timing/model': 1.134666939964518, 'learning_rate/qwen_vl_interface': 7.714968463317207e-06, 'learning_rate/action_model': 7.714968463317207e-05, 'epoch': 0.7}) +[2026-06-15 22:23:18,597][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.14873655140399933, 'train/grad_norm_pre_clip': 10.973343849182129, 'eval/loss': 0.07771673053503036, 'eval/flap_precision': 0.917037037037037, 'eval/flap_recall': 0.9552469135802469, 'eval/flap_f1': 0.9357520786092214, 'eval/flap_rate_model': 0.1875, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9357520786092214, 'timing/data': 0.02341415995033458, 'timing/model': 1.1326672990107909, 'learning_rate/qwen_vl_interface': 7.712364699474996e-06, 'learning_rate/action_model': 7.712364699474996e-05, 'epoch': 0.7}) +[2026-06-15 22:23:18,597][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1700_state +[2026-06-15 22:23:18,600][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 22:23:22,183][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1700_state/model.safetensors +[2026-06-15 22:23:32,033][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1700_state/optimizer.bin +[2026-06-15 22:23:32,034][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1700_state/sampler.bin +[2026-06-15 22:23:32,034][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1700_state/sampler_1.bin +[2026-06-15 22:23:32,035][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1700_state/random_states_0.pkl +[2026-06-15 22:23:32,084][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1700 +[2026-06-15 22:23:32,084][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 22:23:32,090][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 22:23:36,523][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.11022761464118958, 'train/grad_norm_pre_clip': 18.636919021606445, 'timing/data': 0.02332925598602742, 'timing/model': 1.1133283750386909, 'learning_rate/qwen_vl_interface': 7.709759923447667e-06, 'learning_rate/action_model': 7.709759923447665e-05, 'epoch': 0.7}) +[2026-06-15 22:23:41,022][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.22653110325336456, 'train/grad_norm_pre_clip': 15.568135261535645, 'timing/data': 0.0240149189485237, 'timing/model': 1.1187444839742966, 'learning_rate/qwen_vl_interface': 7.707154136305945e-06, 'learning_rate/action_model': 7.707154136305945e-05, 'epoch': 0.7}) +[2026-06-15 22:23:45,455][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.22136344015598297, 'train/grad_norm_pre_clip': 15.153173446655273, 'timing/data': 0.024006057996302843, 'timing/model': 1.124690094962716, 'learning_rate/qwen_vl_interface': 7.704547339120974e-06, 'learning_rate/action_model': 7.704547339120974e-05, 'epoch': 0.7}) +[2026-06-15 22:23:49,932][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.21078607439994812, 'train/grad_norm_pre_clip': 7.894423007965088, 'timing/data': 0.02344114298466593, 'timing/model': 1.1269741429714486, 'learning_rate/qwen_vl_interface': 7.70193953296431e-06, 'learning_rate/action_model': 7.70193953296431e-05, 'epoch': 0.7}) +[2026-06-15 22:23:54,417][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.02316761203110218, 'train/grad_norm_pre_clip': 2.8515448570251465, 'timing/data': 0.024176951963454485, 'timing/model': 1.2856483759824187, 'learning_rate/qwen_vl_interface': 7.69933071890792e-06, 'learning_rate/action_model': 7.699330718907919e-05, 'epoch': 0.7}) +[2026-06-15 22:23:58,740][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.33860015869140625, 'train/grad_norm_pre_clip': 11.505705833435059, 'timing/data': 0.023005350958555937, 'timing/model': 1.1168458400061354, 'learning_rate/qwen_vl_interface': 7.696720898024188e-06, 'learning_rate/action_model': 7.69672089802419e-05, 'epoch': 0.7}) +[2026-06-15 22:24:03,215][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.11526092141866684, 'train/grad_norm_pre_clip': 10.986405372619629, 'timing/data': 0.02378762001171708, 'timing/model': 1.1214391909888946, 'learning_rate/qwen_vl_interface': 7.69411007138592e-06, 'learning_rate/action_model': 7.694110071385919e-05, 'epoch': 0.7}) +[2026-06-15 22:24:07,711][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.16660527884960175, 'train/grad_norm_pre_clip': 9.922956466674805, 'timing/data': 0.024678249028511345, 'timing/model': 1.12399801495485, 'learning_rate/qwen_vl_interface': 7.691498240066322e-06, 'learning_rate/action_model': 7.691498240066322e-05, 'epoch': 0.7}) +[2026-06-15 22:24:12,224][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.010547551326453686, 'train/grad_norm_pre_clip': 2.164583444595337, 'timing/data': 0.023277653963305056, 'timing/model': 1.1301569549832493, 'learning_rate/qwen_vl_interface': 7.688885405139022e-06, 'learning_rate/action_model': 7.688885405139021e-05, 'epoch': 0.7}) +[2026-06-15 22:24:16,710][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.0787431076169014, 'train/grad_norm_pre_clip': 8.757880210876465, 'timing/data': 0.023662313004024327, 'timing/model': 1.2824291659635492, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.7}) +[2026-06-15 22:24:21,180][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.03690207004547119, 'train/grad_norm_pre_clip': 10.017301559448242, 'timing/data': 0.022969426005147398, 'timing/model': 1.247772568021901, 'learning_rate/qwen_vl_interface': 7.683656728757873e-06, 'learning_rate/action_model': 7.683656728757872e-05, 'epoch': 0.7}) +[2026-06-15 22:24:25,526][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.10731060802936554, 'train/grad_norm_pre_clip': 7.958625316619873, 'timing/data': 0.023088293964974582, 'timing/model': 1.1245466119726188, 'learning_rate/qwen_vl_interface': 7.681040889453337e-06, 'learning_rate/action_model': 7.681040889453336e-05, 'epoch': 0.7}) +[2026-06-15 22:24:30,043][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.1201823502779007, 'train/grad_norm_pre_clip': 7.544241905212402, 'timing/data': 0.023567679978441447, 'timing/model': 1.1320248609990813, 'learning_rate/qwen_vl_interface': 7.678424050839722e-06, 'learning_rate/action_model': 7.67842405083972e-05, 'epoch': 0.7}) +[2026-06-15 22:24:34,551][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.07986588776111603, 'train/grad_norm_pre_clip': 4.085024833679199, 'timing/data': 0.02344342099968344, 'timing/model': 1.1223431650432758, 'learning_rate/qwen_vl_interface': 7.675806213992706e-06, 'learning_rate/action_model': 7.675806213992707e-05, 'epoch': 0.7}) +[2026-06-15 22:24:39,070][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.0892796665430069, 'train/grad_norm_pre_clip': 6.446007251739502, 'timing/data': 0.022981443034950644, 'timing/model': 1.127399304008577, 'learning_rate/qwen_vl_interface': 7.673187379988388e-06, 'learning_rate/action_model': 7.673187379988388e-05, 'epoch': 0.7}) +[2026-06-15 22:24:43,605][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.3144841492176056, 'train/grad_norm_pre_clip': 11.846124649047852, 'timing/data': 0.023371249029878527, 'timing/model': 1.2889333209604956, 'learning_rate/qwen_vl_interface': 7.67056754990327e-06, 'learning_rate/action_model': 7.670567549903269e-05, 'epoch': 0.7}) +[2026-06-15 22:24:47,964][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.3371785879135132, 'train/grad_norm_pre_clip': 35.30160903930664, 'timing/data': 0.023400067002512515, 'timing/model': 1.1333520989865065, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.71}) +[2026-06-15 22:24:52,481][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.039977338165044785, 'train/grad_norm_pre_clip': 6.17501974105835, 'timing/data': 0.023024479974992573, 'timing/model': 1.1380913240136579, 'learning_rate/qwen_vl_interface': 7.665324905798694e-06, 'learning_rate/action_model': 7.665324905798695e-05, 'epoch': 0.71}) +[2026-06-15 22:24:57,000][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.21652619540691376, 'train/grad_norm_pre_clip': 11.148140907287598, 'timing/data': 0.02257036801893264, 'timing/model': 1.1357647569966502, 'learning_rate/qwen_vl_interface': 7.66270209393429e-06, 'learning_rate/action_model': 7.66270209393429e-05, 'epoch': 0.71}) +[2026-06-15 22:25:01,484][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.12584778666496277, 'train/grad_norm_pre_clip': 7.67959451675415, 'timing/data': 0.023497964022681117, 'timing/model': 1.1290003009489737, 'learning_rate/qwen_vl_interface': 7.66007829029919e-06, 'learning_rate/action_model': 7.660078290299191e-05, 'epoch': 0.71}) +[2026-06-15 22:25:06,000][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.07666756212711334, 'train/grad_norm_pre_clip': 4.865799427032471, 'timing/data': 0.023623726970981807, 'timing/model': 1.1251771739916876, 'learning_rate/qwen_vl_interface': 7.657453495971942e-06, 'learning_rate/action_model': 7.657453495971943e-05, 'epoch': 0.71}) +[2026-06-15 22:25:10,542][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.27071547508239746, 'train/grad_norm_pre_clip': 14.544293403625488, 'timing/data': 0.02373839303618297, 'timing/model': 1.284608269983437, 'learning_rate/qwen_vl_interface': 7.6548277120315e-06, 'learning_rate/action_model': 7.6548277120315e-05, 'epoch': 0.71}) +[2026-06-15 22:25:14,909][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.060642801225185394, 'train/grad_norm_pre_clip': 3.8778574466705322, 'timing/data': 0.023371969000436366, 'timing/model': 1.1338149000075646, 'learning_rate/qwen_vl_interface': 7.652200939557225e-06, 'learning_rate/action_model': 7.652200939557225e-05, 'epoch': 0.71}) +[2026-06-15 22:25:19,432][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.12292493879795074, 'train/grad_norm_pre_clip': 6.936280727386475, 'timing/data': 0.02311121200909838, 'timing/model': 1.1307100979611278, 'learning_rate/qwen_vl_interface': 7.649573179628881e-06, 'learning_rate/action_model': 7.649573179628881e-05, 'epoch': 0.71}) +[2026-06-15 22:25:23,880][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.056167926639318466, 'train/grad_norm_pre_clip': 3.6404216289520264, 'timing/data': 0.022933573985937983, 'timing/model': 1.1363717920030467, 'learning_rate/qwen_vl_interface': 7.646944433326645e-06, 'learning_rate/action_model': 7.646944433326643e-05, 'epoch': 0.71}) +[2026-06-15 22:25:28,436][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.04069467633962631, 'train/grad_norm_pre_clip': 3.396542549133301, 'timing/data': 0.022900568030308932, 'timing/model': 1.1392028100090101, 'learning_rate/qwen_vl_interface': 7.64431470173109e-06, 'learning_rate/action_model': 7.64431470173109e-05, 'epoch': 0.71}) +[2026-06-15 22:25:32,958][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.06150477007031441, 'train/grad_norm_pre_clip': 9.915154457092285, 'timing/data': 0.024910059990361333, 'timing/model': 1.1264695090358146, 'learning_rate/qwen_vl_interface': 7.641683985923203e-06, 'learning_rate/action_model': 7.641683985923203e-05, 'epoch': 0.71}) +[2026-06-15 22:25:37,483][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.06828178465366364, 'train/grad_norm_pre_clip': 6.28460168838501, 'timing/data': 0.0228804400539957, 'timing/model': 1.2919021190027706, 'learning_rate/qwen_vl_interface': 7.63905228698437e-06, 'learning_rate/action_model': 7.63905228698437e-05, 'epoch': 0.71}) +[2026-06-15 22:25:41,871][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.3901914358139038, 'train/grad_norm_pre_clip': 12.07218074798584, 'timing/data': 0.023485728015657514, 'timing/model': 1.1311230880091898, 'learning_rate/qwen_vl_interface': 7.636419605996385e-06, 'learning_rate/action_model': 7.636419605996384e-05, 'epoch': 0.71}) +[2026-06-15 22:25:46,392][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.1474943459033966, 'train/grad_norm_pre_clip': 9.126969337463379, 'timing/data': 0.02388021699152887, 'timing/model': 1.1294354459969327, 'learning_rate/qwen_vl_interface': 7.633785944041444e-06, 'learning_rate/action_model': 7.633785944041444e-05, 'epoch': 0.71}) +[2026-06-15 22:25:50,901][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.04832236096262932, 'train/grad_norm_pre_clip': 8.8103666305542, 'timing/data': 0.024310745007824153, 'timing/model': 1.1273184579913504, 'learning_rate/qwen_vl_interface': 7.631151302202145e-06, 'learning_rate/action_model': 7.631151302202144e-05, 'epoch': 0.71}) +[2026-06-15 22:25:55,377][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.1182546392083168, 'train/grad_norm_pre_clip': 11.520930290222168, 'timing/data': 0.023884077032562345, 'timing/model': 1.128920064016711, 'learning_rate/qwen_vl_interface': 7.628515681561488e-06, 'learning_rate/action_model': 7.628515681561487e-05, 'epoch': 0.71}) +[2026-06-15 22:25:59,902][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.09428007900714874, 'train/grad_norm_pre_clip': 9.280837059020996, 'timing/data': 0.02345907798735425, 'timing/model': 1.1310457039508037, 'learning_rate/qwen_vl_interface': 7.62587908320288e-06, 'learning_rate/action_model': 7.62587908320288e-05, 'epoch': 0.71}) +[2026-06-15 22:26:04,454][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.24114680290222168, 'train/grad_norm_pre_clip': 17.687294006347656, 'timing/data': 0.02350824198219925, 'timing/model': 1.3059282390167937, 'learning_rate/qwen_vl_interface': 7.6232415082101255e-06, 'learning_rate/action_model': 7.623241508210124e-05, 'epoch': 0.71}) +[2026-06-15 22:26:08,834][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.023671627044677734, 'train/grad_norm_pre_clip': 2.71644926071167, 'timing/data': 0.024443683971185237, 'timing/model': 1.1380328360246494, 'learning_rate/qwen_vl_interface': 7.620602957667432e-06, 'learning_rate/action_model': 7.620602957667432e-05, 'epoch': 0.71}) +[2026-06-15 22:26:13,362][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.13078393042087555, 'train/grad_norm_pre_clip': 4.827732086181641, 'timing/data': 0.023682436964008957, 'timing/model': 1.1372804240090773, 'learning_rate/qwen_vl_interface': 7.617963432659408e-06, 'learning_rate/action_model': 7.617963432659408e-05, 'epoch': 0.71}) +[2026-06-15 22:26:17,890][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.12796561419963837, 'train/grad_norm_pre_clip': 10.472463607788086, 'timing/data': 0.0241972419898957, 'timing/model': 1.1291217550169677, 'learning_rate/qwen_vl_interface': 7.615322934271063e-06, 'learning_rate/action_model': 7.615322934271064e-05, 'epoch': 0.71}) +[2026-06-15 22:26:22,408][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.13128222525119781, 'train/grad_norm_pre_clip': 6.677433490753174, 'timing/data': 0.024024298996664584, 'timing/model': 1.1277063260204159, 'learning_rate/qwen_vl_interface': 7.6126814635878055e-06, 'learning_rate/action_model': 7.612681463587805e-05, 'epoch': 0.71}) +[2026-06-15 22:26:26,947][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.07320156693458557, 'train/grad_norm_pre_clip': 7.026316165924072, 'timing/data': 0.023217480978928506, 'timing/model': 1.2985117970383726, 'learning_rate/qwen_vl_interface': 7.6100390216954445e-06, 'learning_rate/action_model': 7.610039021695445e-05, 'epoch': 0.71}) +[2026-06-15 22:26:31,323][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.06645678728818893, 'train/grad_norm_pre_clip': 7.6624274253845215, 'timing/data': 0.02352542500011623, 'timing/model': 1.136293949035462, 'learning_rate/qwen_vl_interface': 7.607395609680189e-06, 'learning_rate/action_model': 7.60739560968019e-05, 'epoch': 0.71}) +[2026-06-15 22:26:35,850][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.14566031098365784, 'train/grad_norm_pre_clip': 12.057493209838867, 'timing/data': 0.023947798006702214, 'timing/model': 1.1280380130046979, 'learning_rate/qwen_vl_interface': 7.604751228628644e-06, 'learning_rate/action_model': 7.604751228628644e-05, 'epoch': 0.72}) +[2026-06-15 22:26:40,386][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.2783595025539398, 'train/grad_norm_pre_clip': 21.20231056213379, 'timing/data': 0.02423003502190113, 'timing/model': 1.1280313589959405, 'learning_rate/qwen_vl_interface': 7.602105879627814e-06, 'learning_rate/action_model': 7.602105879627814e-05, 'epoch': 0.72}) +[2026-06-15 22:26:44,925][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.080055832862854, 'train/grad_norm_pre_clip': 7.314713478088379, 'timing/data': 0.02363001898629591, 'timing/model': 1.1296548409736715, 'learning_rate/qwen_vl_interface': 7.599459563765107e-06, 'learning_rate/action_model': 7.599459563765106e-05, 'epoch': 0.72}) +[2026-06-15 22:26:49,453][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.07667538523674011, 'train/grad_norm_pre_clip': 8.941956520080566, 'timing/data': 0.024365437973756343, 'timing/model': 1.1315969140268862, 'learning_rate/qwen_vl_interface': 7.596812282128319e-06, 'learning_rate/action_model': 7.596812282128319e-05, 'epoch': 0.72}) +[2026-06-15 22:26:53,988][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.09153429418802261, 'train/grad_norm_pre_clip': 21.955074310302734, 'timing/data': 0.02444314898457378, 'timing/model': 1.2884791960241273, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 0.72}) +[2026-06-15 22:26:58,368][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.11912021785974503, 'train/grad_norm_pre_clip': 10.30300235748291, 'timing/data': 0.029272107989527285, 'timing/model': 1.128041571995709, 'learning_rate/qwen_vl_interface': 7.59151482588569e-06, 'learning_rate/action_model': 7.59151482588569e-05, 'epoch': 0.72}) +[2026-06-15 22:27:02,895][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.3637810945510864, 'train/grad_norm_pre_clip': 30.713232040405273, 'timing/data': 0.023378194018732756, 'timing/model': 1.1303410779801197, 'learning_rate/qwen_vl_interface': 7.588864653457433e-06, 'learning_rate/action_model': 7.588864653457432e-05, 'epoch': 0.72}) +[2026-06-15 22:27:07,429][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.24321573972702026, 'train/grad_norm_pre_clip': 15.17781925201416, 'timing/data': 0.02381876896834001, 'timing/model': 1.1373669799650088, 'learning_rate/qwen_vl_interface': 7.5862135196102634e-06, 'learning_rate/action_model': 7.586213519610263e-05, 'epoch': 0.72}) +[2026-06-15 22:27:11,920][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.08501055091619492, 'train/grad_norm_pre_clip': 5.28513240814209, 'timing/data': 0.023375999997369945, 'timing/model': 1.1282324850326404, 'learning_rate/qwen_vl_interface': 7.583561425433962e-06, 'learning_rate/action_model': 7.583561425433962e-05, 'epoch': 0.72}) +[2026-06-15 22:27:16,477][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.16866201162338257, 'train/grad_norm_pre_clip': 12.891642570495605, 'timing/data': 0.023369242961052805, 'timing/model': 1.1441644149599597, 'learning_rate/qwen_vl_interface': 7.580908372018706e-06, 'learning_rate/action_model': 7.580908372018705e-05, 'epoch': 0.72}) +[2026-06-15 22:27:21,012][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.05516136437654495, 'train/grad_norm_pre_clip': 9.359580039978027, 'timing/data': 0.023468678002245724, 'timing/model': 1.2951305959722959, 'learning_rate/qwen_vl_interface': 7.5782543604550655e-06, 'learning_rate/action_model': 7.578254360455064e-05, 'epoch': 0.72}) +[2026-06-15 22:27:25,372][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.17598165571689606, 'train/grad_norm_pre_clip': 12.122477531433105, 'timing/data': 0.023713839997071773, 'timing/model': 1.1334379860199988, 'learning_rate/qwen_vl_interface': 7.575599391834e-06, 'learning_rate/action_model': 7.575599391834e-05, 'epoch': 0.72}) +[2026-06-15 22:27:29,897][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.14330805838108063, 'train/grad_norm_pre_clip': 5.648210525512695, 'timing/data': 0.023089858004823327, 'timing/model': 1.1332429900066927, 'learning_rate/qwen_vl_interface': 7.572943467246871e-06, 'learning_rate/action_model': 7.57294346724687e-05, 'epoch': 0.72}) +[2026-06-15 22:27:34,371][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.14162631332874298, 'train/grad_norm_pre_clip': 15.99484634399414, 'timing/data': 0.02339206001488492, 'timing/model': 1.1387334190076217, 'learning_rate/qwen_vl_interface': 7.570286587785428e-06, 'learning_rate/action_model': 7.570286587785428e-05, 'epoch': 0.72}) +[2026-06-15 22:27:38,933][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.02739296294748783, 'train/grad_norm_pre_clip': 1.844329595565796, 'timing/data': 0.023206468031276017, 'timing/model': 1.1363290550070815, 'learning_rate/qwen_vl_interface': 7.567628754541815e-06, 'learning_rate/action_model': 7.567628754541814e-05, 'epoch': 0.72}) +[2026-06-15 22:27:43,482][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.09001687914133072, 'train/grad_norm_pre_clip': 12.960601806640625, 'timing/data': 0.023567476018797606, 'timing/model': 1.1295197189901955, 'learning_rate/qwen_vl_interface': 7.564969968608565e-06, 'learning_rate/action_model': 7.564969968608565e-05, 'epoch': 0.72}) +[2026-06-15 22:27:48,039][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.34134137630462646, 'train/grad_norm_pre_clip': 26.234432220458984, 'timing/data': 0.023017955012619495, 'timing/model': 1.2953761279932223, 'learning_rate/qwen_vl_interface': 7.562310231078605e-06, 'learning_rate/action_model': 7.562310231078604e-05, 'epoch': 0.72}) +[2026-06-15 22:27:52,407][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.12223200500011444, 'train/grad_norm_pre_clip': 8.777413368225098, 'timing/data': 0.02341802802402526, 'timing/model': 1.1299372569774278, 'learning_rate/qwen_vl_interface': 7.559649543045252e-06, 'learning_rate/action_model': 7.559649543045252e-05, 'epoch': 0.72}) +[2026-06-15 22:27:56,941][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.22145365178585052, 'train/grad_norm_pre_clip': 26.804012298583984, 'timing/data': 0.022680172987747937, 'timing/model': 1.1375700210337527, 'learning_rate/qwen_vl_interface': 7.556987905602215e-06, 'learning_rate/action_model': 7.556987905602215e-05, 'epoch': 0.72}) +[2026-06-15 22:28:01,499][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.0954052209854126, 'train/grad_norm_pre_clip': 11.425902366638184, 'timing/data': 0.023964348947629333, 'timing/model': 1.1314838210237212, 'learning_rate/qwen_vl_interface': 7.554325319843593e-06, 'learning_rate/action_model': 7.554325319843592e-05, 'epoch': 0.72}) +[2026-06-15 22:28:05,989][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.08869650214910507, 'train/grad_norm_pre_clip': 5.231722831726074, 'timing/data': 0.02329438499873504, 'timing/model': 1.1335463920258917, 'learning_rate/qwen_vl_interface': 7.551661786863874e-06, 'learning_rate/action_model': 7.551661786863874e-05, 'epoch': 0.72}) +[2026-06-15 22:28:10,543][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.16974478960037231, 'train/grad_norm_pre_clip': 10.336875915527344, 'timing/data': 0.02333670703228563, 'timing/model': 1.1358903599902987, 'learning_rate/qwen_vl_interface': 7.548997307757936e-06, 'learning_rate/action_model': 7.548997307757935e-05, 'epoch': 0.72}) +[2026-06-15 22:28:15,096][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.33334895968437195, 'train/grad_norm_pre_clip': 7.691719055175781, 'timing/data': 0.023575051978696138, 'timing/model': 1.3085066479980014, 'learning_rate/qwen_vl_interface': 7.546331883621044e-06, 'learning_rate/action_model': 7.546331883621044e-05, 'epoch': 0.72}) +[2026-06-15 22:28:19,477][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.027749311178922653, 'train/grad_norm_pre_clip': 2.110412120819092, 'timing/data': 0.023863229958806187, 'timing/model': 1.12942206999287, 'learning_rate/qwen_vl_interface': 7.543665515548857e-06, 'learning_rate/action_model': 7.543665515548856e-05, 'epoch': 0.72}) +[2026-06-15 22:28:24,051][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.032656095921993256, 'train/grad_norm_pre_clip': 2.93333101272583, 'timing/data': 0.023049345996696502, 'timing/model': 1.1451497129746713, 'learning_rate/qwen_vl_interface': 7.540998204637413e-06, 'learning_rate/action_model': 7.540998204637413e-05, 'epoch': 0.73}) +[2026-06-15 22:28:28,593][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.06138290837407112, 'train/grad_norm_pre_clip': 5.426699638366699, 'timing/data': 0.023459211050067097, 'timing/model': 1.1376481609768234, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.73}) +[2026-06-15 22:28:33,136][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.17402850091457367, 'train/grad_norm_pre_clip': 15.250397682189941, 'timing/data': 0.02394619002006948, 'timing/model': 1.135153238021303, 'learning_rate/qwen_vl_interface': 7.535660758682877e-06, 'learning_rate/action_model': 7.535660758682876e-05, 'epoch': 0.73}) +[2026-06-15 22:28:37,678][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.014881937764585018, 'train/grad_norm_pre_clip': 1.870282769203186, 'timing/data': 0.023498577007558197, 'timing/model': 1.2941651399596594, 'learning_rate/qwen_vl_interface': 7.532990625833806e-06, 'learning_rate/action_model': 7.532990625833805e-05, 'epoch': 0.73}) +[2026-06-15 22:28:42,061][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.10798228532075882, 'train/grad_norm_pre_clip': 6.189100742340088, 'timing/data': 0.023632274009287357, 'timing/model': 1.130784949986264, 'learning_rate/qwen_vl_interface': 7.5303195545335225e-06, 'learning_rate/action_model': 7.530319554533522e-05, 'epoch': 0.73}) +[2026-06-15 22:28:46,588][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.1580830216407776, 'train/grad_norm_pre_clip': 11.170669555664062, 'timing/data': 0.024235315970145166, 'timing/model': 1.1326763139804825, 'learning_rate/qwen_vl_interface': 7.527647545880007e-06, 'learning_rate/action_model': 7.527647545880007e-05, 'epoch': 0.73}) +[2026-06-15 22:28:51,132][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.16818569600582123, 'train/grad_norm_pre_clip': 8.86379623413086, 'timing/data': 0.023489287996198982, 'timing/model': 1.1368613769882359, 'learning_rate/qwen_vl_interface': 7.52497460097162e-06, 'learning_rate/action_model': 7.52497460097162e-05, 'epoch': 0.73}) +[2026-06-15 22:28:55,664][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.13573016226291656, 'train/grad_norm_pre_clip': 15.606772422790527, 'timing/data': 0.02431292401161045, 'timing/model': 1.1313898300286382, 'learning_rate/qwen_vl_interface': 7.522300720907105e-06, 'learning_rate/action_model': 7.522300720907105e-05, 'epoch': 0.73}) +[2026-06-15 22:29:00,228][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.09077122807502747, 'train/grad_norm_pre_clip': 10.375570297241211, 'timing/data': 0.023738673015031964, 'timing/model': 1.1415682599763386, 'learning_rate/qwen_vl_interface': 7.519625906785601e-06, 'learning_rate/action_model': 7.5196259067856e-05, 'epoch': 0.73}) +[2026-06-15 22:29:04,757][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.19956836104393005, 'train/grad_norm_pre_clip': 52.149600982666016, 'timing/data': 0.02486557699739933, 'timing/model': 1.29320880305022, 'learning_rate/qwen_vl_interface': 7.516950159706614e-06, 'learning_rate/action_model': 7.516950159706614e-05, 'epoch': 0.73}) +[2026-06-15 22:29:09,155][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.13970063626766205, 'train/grad_norm_pre_clip': 28.190696716308594, 'timing/data': 0.02407945302547887, 'timing/model': 1.137622844951693, 'learning_rate/qwen_vl_interface': 7.514273480770047e-06, 'learning_rate/action_model': 7.514273480770047e-05, 'epoch': 0.73}) +[2026-06-15 22:29:13,705][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.10742244869470596, 'train/grad_norm_pre_clip': 5.577701091766357, 'timing/data': 0.023720061988569796, 'timing/model': 1.1302033890387975, 'learning_rate/qwen_vl_interface': 7.511595871076181e-06, 'learning_rate/action_model': 7.511595871076181e-05, 'epoch': 0.73}) +[2026-06-15 22:29:18,232][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.03812802582979202, 'train/grad_norm_pre_clip': 6.966533184051514, 'timing/data': 0.02370197098935023, 'timing/model': 1.1290310180047527, 'learning_rate/qwen_vl_interface': 7.50891733172568e-06, 'learning_rate/action_model': 7.50891733172568e-05, 'epoch': 0.73}) +[2026-06-15 22:29:22,747][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.12825115025043488, 'train/grad_norm_pre_clip': 10.148225784301758, 'timing/data': 0.024591109016910195, 'timing/model': 1.1320232750149444, 'learning_rate/qwen_vl_interface': 7.506237863819593e-06, 'learning_rate/action_model': 7.506237863819593e-05, 'epoch': 0.73}) +[2026-06-15 22:29:27,281][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.13176728785037994, 'train/grad_norm_pre_clip': 7.991209030151367, 'timing/data': 0.024622269032988697, 'timing/model': 1.1301330460119061, 'learning_rate/qwen_vl_interface': 7.503557468459345e-06, 'learning_rate/action_model': 7.503557468459344e-05, 'epoch': 0.73}) +[2026-06-15 22:29:31,839][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.09971493482589722, 'train/grad_norm_pre_clip': 11.013273239135742, 'timing/data': 0.023239677015226334, 'timing/model': 1.3061540509806946, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.73}) +[2026-06-15 22:29:36,213][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.0614878311753273, 'train/grad_norm_pre_clip': 4.996769428253174, 'timing/data': 0.023988879984244704, 'timing/model': 1.1310904949787073, 'learning_rate/qwen_vl_interface': 7.498193899783983e-06, 'learning_rate/action_model': 7.498193899783983e-05, 'epoch': 0.73}) +[2026-06-15 22:29:40,730][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.1694914549589157, 'train/grad_norm_pre_clip': 14.457265853881836, 'timing/data': 0.02383134700357914, 'timing/model': 1.1324654149939306, 'learning_rate/qwen_vl_interface': 7.495510728673631e-06, 'learning_rate/action_model': 7.495510728673632e-05, 'epoch': 0.73}) +[2026-06-15 22:29:45,165][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.29281026124954224, 'train/grad_norm_pre_clip': 12.2330904006958, 'timing/data': 0.024473304045386612, 'timing/model': 1.1313887289725244, 'learning_rate/qwen_vl_interface': 7.492826634518642e-06, 'learning_rate/action_model': 7.492826634518641e-05, 'epoch': 0.73}) +[2026-06-15 22:29:49,710][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.08132904767990112, 'train/grad_norm_pre_clip': 7.281271934509277, 'timing/data': 0.023282042995560914, 'timing/model': 1.136980537965428, 'learning_rate/qwen_vl_interface': 7.4901416184223404e-06, 'learning_rate/action_model': 7.49014161842234e-05, 'epoch': 0.73}) +[2026-06-15 22:29:54,250][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.2506510019302368, 'train/grad_norm_pre_clip': 11.07174015045166, 'timing/data': 0.024102484981995076, 'timing/model': 1.1351693969918415, 'learning_rate/qwen_vl_interface': 7.487455681488437e-06, 'learning_rate/action_model': 7.487455681488436e-05, 'epoch': 0.73}) +[2026-06-15 22:29:58,791][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.07080504298210144, 'train/grad_norm_pre_clip': 4.7535080909729, 'timing/data': 0.02393034700071439, 'timing/model': 1.2942182759870775, 'learning_rate/qwen_vl_interface': 7.484768824821018e-06, 'learning_rate/action_model': 7.484768824821018e-05, 'epoch': 0.73}) +[2026-06-15 22:30:03,194][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.10249438881874084, 'train/grad_norm_pre_clip': 34.21220779418945, 'timing/data': 0.0278264670050703, 'timing/model': 1.1563412029645406, 'learning_rate/qwen_vl_interface': 7.482081049524551e-06, 'learning_rate/action_model': 7.482081049524551e-05, 'epoch': 0.73}) +[2026-06-15 22:30:07,719][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.06341033428907394, 'train/grad_norm_pre_clip': 6.1062846183776855, 'timing/data': 0.02369397896109149, 'timing/model': 1.1322202410083264, 'learning_rate/qwen_vl_interface': 7.479392356703878e-06, 'learning_rate/action_model': 7.479392356703878e-05, 'epoch': 0.73}) +[2026-06-15 22:30:12,250][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.2200479805469513, 'train/grad_norm_pre_clip': 8.91760540008545, 'timing/data': 0.023256010026670992, 'timing/model': 1.1312402740004472, 'learning_rate/qwen_vl_interface': 7.476702747464219e-06, 'learning_rate/action_model': 7.476702747464219e-05, 'epoch': 0.73}) +[2026-06-15 22:30:16,757][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.11241738498210907, 'train/grad_norm_pre_clip': 7.621449947357178, 'timing/data': 0.02443512697936967, 'timing/model': 1.135403860011138, 'learning_rate/qwen_vl_interface': 7.474012222911169e-06, 'learning_rate/action_model': 7.474012222911169e-05, 'epoch': 0.74}) +[2026-06-15 22:30:21,299][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.10052842646837234, 'train/grad_norm_pre_clip': 6.139079570770264, 'timing/data': 0.024521589977666736, 'timing/model': 1.1305984340142459, 'learning_rate/qwen_vl_interface': 7.471320784150703e-06, 'learning_rate/action_model': 7.471320784150703e-05, 'epoch': 0.74}) +[2026-06-15 22:30:25,863][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.1281173974275589, 'train/grad_norm_pre_clip': 11.805927276611328, 'timing/data': 0.02318766398821026, 'timing/model': 1.2940701500046998, 'learning_rate/qwen_vl_interface': 7.468628432289169e-06, 'learning_rate/action_model': 7.468628432289168e-05, 'epoch': 0.74}) +[2026-06-15 22:30:30,259][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.1079210489988327, 'train/grad_norm_pre_clip': 6.271223545074463, 'timing/data': 0.023728947970084846, 'timing/model': 1.1377413489972241, 'learning_rate/qwen_vl_interface': 7.465935168433292e-06, 'learning_rate/action_model': 7.465935168433292e-05, 'epoch': 0.74}) +[2026-06-15 22:30:34,805][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.22921524941921234, 'train/grad_norm_pre_clip': 13.758879661560059, 'timing/data': 0.02353517903247848, 'timing/model': 1.1326228780089878, 'learning_rate/qwen_vl_interface': 7.463240993690168e-06, 'learning_rate/action_model': 7.463240993690168e-05, 'epoch': 0.74}) +[2026-06-15 22:30:39,354][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.2667178511619568, 'train/grad_norm_pre_clip': 13.107126235961914, 'timing/data': 0.02381434600101784, 'timing/model': 1.130958363006357, 'learning_rate/qwen_vl_interface': 7.460545909167276e-06, 'learning_rate/action_model': 7.460545909167275e-05, 'epoch': 0.74}) +[2026-06-15 22:30:43,914][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.10321265459060669, 'train/grad_norm_pre_clip': 5.409291744232178, 'timing/data': 0.024742635956499726, 'timing/model': 1.1309030170086771, 'learning_rate/qwen_vl_interface': 7.457849915972459e-06, 'learning_rate/action_model': 7.457849915972458e-05, 'epoch': 0.74}) +[2026-06-15 22:30:48,458][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.08073585480451584, 'train/grad_norm_pre_clip': 7.572106838226318, 'timing/data': 0.02351676905527711, 'timing/model': 1.2880673939944245, 'learning_rate/qwen_vl_interface': 7.455153015213941e-06, 'learning_rate/action_model': 7.45515301521394e-05, 'epoch': 0.74}) +[2026-06-15 22:30:52,829][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.08844206482172012, 'train/grad_norm_pre_clip': 8.91308307647705, 'timing/data': 0.023728704953100532, 'timing/model': 1.129507987992838, 'learning_rate/qwen_vl_interface': 7.452455208000311e-06, 'learning_rate/action_model': 7.452455208000311e-05, 'epoch': 0.74}) +[2026-06-15 22:30:57,363][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.1465889811515808, 'train/grad_norm_pre_clip': 8.353948593139648, 'timing/data': 0.023999350029043853, 'timing/model': 1.1308446599869058, 'learning_rate/qwen_vl_interface': 7.449756495440541e-06, 'learning_rate/action_model': 7.44975649544054e-05, 'epoch': 0.74}) +[2026-06-15 22:35:18,362][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.07425543665885925, 'train/grad_norm_pre_clip': 3.968994617462158, 'eval/loss': 0.05934939906001091, 'eval/flap_precision': 0.9275147928994083, 'eval/flap_recall': 0.9675925925925926, 'eval/flap_f1': 0.9471299093655589, 'eval/flap_rate_model': 0.18777777777777777, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9471299093655589, 'timing/data': 0.02521189901744947, 'timing/model': 1.1345215549808927, 'learning_rate/qwen_vl_interface': 7.447056878643969e-06, 'learning_rate/action_model': 7.447056878643968e-05, 'epoch': 0.74}) +[2026-06-15 22:35:18,363][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1800_state +[2026-06-15 22:35:18,365][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 22:35:21,936][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1800_state/model.safetensors +[2026-06-15 22:35:31,958][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1800_state/optimizer.bin +[2026-06-15 22:35:31,959][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1800_state/sampler.bin +[2026-06-15 22:35:31,959][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1800_state/sampler_1.bin +[2026-06-15 22:35:31,960][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1800_state/random_states_0.pkl +[2026-06-15 22:35:32,008][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1800 +[2026-06-15 22:35:32,008][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 22:35:32,014][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 22:35:36,497][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.12972240149974823, 'train/grad_norm_pre_clip': 6.384721279144287, 'timing/data': 0.02296358800958842, 'timing/model': 1.1176019010017626, 'learning_rate/qwen_vl_interface': 7.444356358720303e-06, 'learning_rate/action_model': 7.444356358720302e-05, 'epoch': 0.74}) +[2026-06-15 22:35:40,997][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.043146710842847824, 'train/grad_norm_pre_clip': 2.7913718223571777, 'timing/data': 0.023985305975656956, 'timing/model': 1.1113232970237732, 'learning_rate/qwen_vl_interface': 7.4416549367796255e-06, 'learning_rate/action_model': 7.441654936779625e-05, 'epoch': 0.74}) +[2026-06-15 22:35:45,467][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.18124710023403168, 'train/grad_norm_pre_clip': 9.014972686767578, 'timing/data': 0.023429167980793864, 'timing/model': 1.1123151240171865, 'learning_rate/qwen_vl_interface': 7.4389526139323895e-06, 'learning_rate/action_model': 7.438952613932389e-05, 'epoch': 0.74}) +[2026-06-15 22:35:49,937][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.08827216178178787, 'train/grad_norm_pre_clip': 5.446157455444336, 'timing/data': 0.023945342982187867, 'timing/model': 1.1262731490423903, 'learning_rate/qwen_vl_interface': 7.436249391289419e-06, 'learning_rate/action_model': 7.436249391289418e-05, 'epoch': 0.74}) +[2026-06-15 22:35:54,424][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.017335576936602592, 'train/grad_norm_pre_clip': 2.9667611122131348, 'timing/data': 0.023837335989810526, 'timing/model': 1.2847612079931423, 'learning_rate/qwen_vl_interface': 7.433545269961906e-06, 'learning_rate/action_model': 7.433545269961906e-05, 'epoch': 0.74}) +[2026-06-15 22:35:58,748][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.02503368817269802, 'train/grad_norm_pre_clip': 2.9346067905426025, 'timing/data': 0.02376435202313587, 'timing/model': 1.1181210200302303, 'learning_rate/qwen_vl_interface': 7.430840251061411e-06, 'learning_rate/action_model': 7.430840251061411e-05, 'epoch': 0.74}) +[2026-06-15 22:36:03,222][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.08632459491491318, 'train/grad_norm_pre_clip': 8.908485412597656, 'timing/data': 0.02324596798280254, 'timing/model': 1.120231622015126, 'learning_rate/qwen_vl_interface': 7.428134335699866e-06, 'learning_rate/action_model': 7.428134335699865e-05, 'epoch': 0.74}) +[2026-06-15 22:36:07,710][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.08785773068666458, 'train/grad_norm_pre_clip': 13.408465385437012, 'timing/data': 0.023301081964746118, 'timing/model': 1.128168462018948, 'learning_rate/qwen_vl_interface': 7.425427524989572e-06, 'learning_rate/action_model': 7.425427524989572e-05, 'epoch': 0.74}) +[2026-06-15 22:36:12,214][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.036693885922431946, 'train/grad_norm_pre_clip': 4.247655391693115, 'timing/data': 0.023667481960728765, 'timing/model': 1.1285065669799224, 'learning_rate/qwen_vl_interface': 7.422719820043194e-06, 'learning_rate/action_model': 7.422719820043193e-05, 'epoch': 0.74}) +[2026-06-15 22:36:16,711][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.08576042205095291, 'train/grad_norm_pre_clip': 11.62682819366455, 'timing/data': 0.023689966998063028, 'timing/model': 1.1180454069981351, 'learning_rate/qwen_vl_interface': 7.420011221973771e-06, 'learning_rate/action_model': 7.420011221973771e-05, 'epoch': 0.74}) +[2026-06-15 22:36:21,211][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.005284760147333145, 'train/grad_norm_pre_clip': 0.5475513339042664, 'timing/data': 0.024597192998044193, 'timing/model': 1.285325566015672, 'learning_rate/qwen_vl_interface': 7.417301731894702e-06, 'learning_rate/action_model': 7.417301731894701e-05, 'epoch': 0.74}) +[2026-06-15 22:36:25,548][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.11732635647058487, 'train/grad_norm_pre_clip': 7.629217147827148, 'timing/data': 0.0238980469875969, 'timing/model': 1.1247818600386381, 'learning_rate/qwen_vl_interface': 7.414591350919756e-06, 'learning_rate/action_model': 7.414591350919756e-05, 'epoch': 0.74}) +[2026-06-15 22:36:30,060][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.1387498378753662, 'train/grad_norm_pre_clip': 18.602170944213867, 'timing/data': 0.024651726009324193, 'timing/model': 1.1211200220277533, 'learning_rate/qwen_vl_interface': 7.4118800801630715e-06, 'learning_rate/action_model': 7.411880080163072e-05, 'epoch': 0.74}) +[2026-06-15 22:36:34,607][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.06088806688785553, 'train/grad_norm_pre_clip': 13.003385543823242, 'timing/data': 0.023872088000644, 'timing/model': 1.1229153479798697, 'learning_rate/qwen_vl_interface': 7.409167920739146e-06, 'learning_rate/action_model': 7.409167920739147e-05, 'epoch': 0.75}) +[2026-06-15 22:36:39,145][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.030294235795736313, 'train/grad_norm_pre_clip': 5.9079484939575195, 'timing/data': 0.03189024602761492, 'timing/model': 1.1251014120061882, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.75}) +[2026-06-15 22:36:43,677][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.03401011973619461, 'train/grad_norm_pre_clip': 5.775481224060059, 'timing/data': 0.02298026898643002, 'timing/model': 1.2825772920041345, 'learning_rate/qwen_vl_interface': 7.403740940349408e-06, 'learning_rate/action_model': 7.403740940349408e-05, 'epoch': 0.75}) +[2026-06-15 22:36:48,166][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.020691201090812683, 'train/grad_norm_pre_clip': 4.453629970550537, 'timing/data': 0.022860964003484696, 'timing/model': 1.2647510769893415, 'learning_rate/qwen_vl_interface': 7.401026121614421e-06, 'learning_rate/action_model': 7.40102612161442e-05, 'epoch': 0.75}) +[2026-06-15 22:36:52,522][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.0693872720003128, 'train/grad_norm_pre_clip': 9.445535659790039, 'timing/data': 0.02329737297259271, 'timing/model': 1.1278231409960426, 'learning_rate/qwen_vl_interface': 7.398310418673846e-06, 'learning_rate/action_model': 7.398310418673846e-05, 'epoch': 0.75}) +[2026-06-15 22:36:57,059][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.05531151965260506, 'train/grad_norm_pre_clip': 5.868686676025391, 'timing/data': 0.023711126996204257, 'timing/model': 1.1339299209648743, 'learning_rate/qwen_vl_interface': 7.395593832644007e-06, 'learning_rate/action_model': 7.395593832644007e-05, 'epoch': 0.75}) +[2026-06-15 22:37:01,593][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.013595164753496647, 'train/grad_norm_pre_clip': 2.426328182220459, 'timing/data': 0.0239470349624753, 'timing/model': 1.1287431670352817, 'learning_rate/qwen_vl_interface': 7.39287636464159e-06, 'learning_rate/action_model': 7.39287636464159e-05, 'epoch': 0.75}) +[2026-06-15 22:37:06,120][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.31699687242507935, 'train/grad_norm_pre_clip': 10.056092262268066, 'timing/data': 0.024313297006301582, 'timing/model': 1.1272706729942001, 'learning_rate/qwen_vl_interface': 7.3901580157836425e-06, 'learning_rate/action_model': 7.390158015783643e-05, 'epoch': 0.75}) +[2026-06-15 22:37:10,604][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.18223777413368225, 'train/grad_norm_pre_clip': 8.900858879089355, 'timing/data': 0.0233057209989056, 'timing/model': 1.1238519449834712, 'learning_rate/qwen_vl_interface': 7.387438787187576e-06, 'learning_rate/action_model': 7.387438787187576e-05, 'epoch': 0.75}) +[2026-06-15 22:37:15,122][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.010145436972379684, 'train/grad_norm_pre_clip': 1.142324686050415, 'timing/data': 0.024042233999352902, 'timing/model': 1.2835396230220795, 'learning_rate/qwen_vl_interface': 7.3847186799711625e-06, 'learning_rate/action_model': 7.384718679971162e-05, 'epoch': 0.75}) +[2026-06-15 22:37:19,477][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.16954085230827332, 'train/grad_norm_pre_clip': 15.190254211425781, 'timing/data': 0.02400067902635783, 'timing/model': 1.1279279850423336, 'learning_rate/qwen_vl_interface': 7.381997695252535e-06, 'learning_rate/action_model': 7.381997695252535e-05, 'epoch': 0.75}) +[2026-06-15 22:37:23,996][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.1979527473449707, 'train/grad_norm_pre_clip': 11.768623352050781, 'timing/data': 0.023495704983361065, 'timing/model': 1.1302535959985107, 'learning_rate/qwen_vl_interface': 7.379275834150186e-06, 'learning_rate/action_model': 7.379275834150186e-05, 'epoch': 0.75}) +[2026-06-15 22:37:28,438][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.13823333382606506, 'train/grad_norm_pre_clip': 19.355146408081055, 'timing/data': 0.022747421055100858, 'timing/model': 1.1318139609647915, 'learning_rate/qwen_vl_interface': 7.376553097782972e-06, 'learning_rate/action_model': 7.376553097782972e-05, 'epoch': 0.75}) +[2026-06-15 22:37:32,880][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.09198304265737534, 'train/grad_norm_pre_clip': 18.3721923828125, 'timing/data': 0.02289524901425466, 'timing/model': 1.1276327000232413, 'learning_rate/qwen_vl_interface': 7.373829487270109e-06, 'learning_rate/action_model': 7.373829487270109e-05, 'epoch': 0.75}) +[2026-06-15 22:37:37,338][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.012557975016534328, 'train/grad_norm_pre_clip': 1.9474669694900513, 'timing/data': 0.023075227974914014, 'timing/model': 1.130732592951972, 'learning_rate/qwen_vl_interface': 7.371105003731165e-06, 'learning_rate/action_model': 7.371105003731164e-05, 'epoch': 0.75}) +[2026-06-15 22:37:41,852][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.23798008263111115, 'train/grad_norm_pre_clip': 15.79484748840332, 'timing/data': 0.022717583982739598, 'timing/model': 1.266983728040941, 'learning_rate/qwen_vl_interface': 7.3683796482860755e-06, 'learning_rate/action_model': 7.368379648286075e-05, 'epoch': 0.75}) +[2026-06-15 22:37:46,218][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.23206685483455658, 'train/grad_norm_pre_clip': 15.325644493103027, 'timing/data': 0.023759836971294135, 'timing/model': 1.1308964809868485, 'learning_rate/qwen_vl_interface': 7.365653422055131e-06, 'learning_rate/action_model': 7.365653422055131e-05, 'epoch': 0.75}) +[2026-06-15 22:37:50,746][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.030322251841425896, 'train/grad_norm_pre_clip': 6.287487506866455, 'timing/data': 0.022638705966528505, 'timing/model': 1.1311579960165545, 'learning_rate/qwen_vl_interface': 7.3629263261589816e-06, 'learning_rate/action_model': 7.362926326158981e-05, 'epoch': 0.75}) +[2026-06-15 22:37:55,314][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.18393278121948242, 'train/grad_norm_pre_clip': 8.751508712768555, 'timing/data': 0.02360785595374182, 'timing/model': 1.139505919010844, 'learning_rate/qwen_vl_interface': 7.36019836171863e-06, 'learning_rate/action_model': 7.360198361718629e-05, 'epoch': 0.75}) +[2026-06-15 22:37:59,865][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.05200350284576416, 'train/grad_norm_pre_clip': 8.960681915283203, 'timing/data': 0.022708289965521544, 'timing/model': 1.130698986002244, 'learning_rate/qwen_vl_interface': 7.357469529855441e-06, 'learning_rate/action_model': 7.35746952985544e-05, 'epoch': 0.75}) +[2026-06-15 22:38:04,403][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.23667411506175995, 'train/grad_norm_pre_clip': 13.682710647583008, 'timing/data': 0.022511538991238922, 'timing/model': 1.3014643960050307, 'learning_rate/qwen_vl_interface': 7.354739831691134e-06, 'learning_rate/action_model': 7.354739831691135e-05, 'epoch': 0.75}) +[2026-06-15 22:38:08,780][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.17436562478542328, 'train/grad_norm_pre_clip': 7.801860809326172, 'timing/data': 0.023397279961500317, 'timing/model': 1.1331269249785691, 'learning_rate/qwen_vl_interface': 7.352009268347786e-06, 'learning_rate/action_model': 7.352009268347786e-05, 'epoch': 0.75}) +[2026-06-15 22:38:13,327][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.1178668737411499, 'train/grad_norm_pre_clip': 9.5099458694458, 'timing/data': 0.02390891802497208, 'timing/model': 1.130674674990587, 'learning_rate/qwen_vl_interface': 7.349277840947828e-06, 'learning_rate/action_model': 7.349277840947828e-05, 'epoch': 0.75}) +[2026-06-15 22:38:17,784][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.05829450488090515, 'train/grad_norm_pre_clip': 11.228300094604492, 'timing/data': 0.02276963397162035, 'timing/model': 1.1316281729959883, 'learning_rate/qwen_vl_interface': 7.346545550614046e-06, 'learning_rate/action_model': 7.346545550614046e-05, 'epoch': 0.75}) +[2026-06-15 22:38:22,315][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.3797227740287781, 'train/grad_norm_pre_clip': 11.027487754821777, 'timing/data': 0.022792283969465643, 'timing/model': 1.1297585800057277, 'learning_rate/qwen_vl_interface': 7.343812398469582e-06, 'learning_rate/action_model': 7.343812398469581e-05, 'epoch': 0.76}) +[2026-06-15 22:38:26,861][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.036072395741939545, 'train/grad_norm_pre_clip': 3.7144482135772705, 'timing/data': 0.02254123002057895, 'timing/model': 1.1340307290083729, 'learning_rate/qwen_vl_interface': 7.34107838563793e-06, 'learning_rate/action_model': 7.341078385637929e-05, 'epoch': 0.76}) +[2026-06-15 22:38:31,405][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.13797755539417267, 'train/grad_norm_pre_clip': 5.246040344238281, 'timing/data': 0.022033604967873544, 'timing/model': 1.2873907599714585, 'learning_rate/qwen_vl_interface': 7.338343513242941e-06, 'learning_rate/action_model': 7.338343513242941e-05, 'epoch': 0.76}) +[2026-06-15 22:38:35,899][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.14091120660305023, 'train/grad_norm_pre_clip': 6.177683353424072, 'timing/data': 0.0225504920235835, 'timing/model': 1.2512310680467635, 'learning_rate/qwen_vl_interface': 7.335607782408818e-06, 'learning_rate/action_model': 7.335607782408818e-05, 'epoch': 0.76}) +[2026-06-15 22:38:40,291][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.10480339825153351, 'train/grad_norm_pre_clip': 18.44123077392578, 'timing/data': 0.023197049042209983, 'timing/model': 1.1414194710087031, 'learning_rate/qwen_vl_interface': 7.3328711942601145e-06, 'learning_rate/action_model': 7.332871194260115e-05, 'epoch': 0.76}) +[2026-06-15 22:38:44,836][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.045411668717861176, 'train/grad_norm_pre_clip': 2.3247411251068115, 'timing/data': 0.024275981006212533, 'timing/model': 1.1362390770227648, 'learning_rate/qwen_vl_interface': 7.33013374992174e-06, 'learning_rate/action_model': 7.33013374992174e-05, 'epoch': 0.76}) +[2026-06-15 22:38:49,386][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.09791966527700424, 'train/grad_norm_pre_clip': 8.345010757446289, 'timing/data': 0.023986222979146987, 'timing/model': 1.1345696370117366, 'learning_rate/qwen_vl_interface': 7.327395450518956e-06, 'learning_rate/action_model': 7.327395450518956e-05, 'epoch': 0.76}) +[2026-06-15 22:38:53,941][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.17432288825511932, 'train/grad_norm_pre_clip': 11.41829776763916, 'timing/data': 0.023737679002806544, 'timing/model': 1.1304604400065728, 'learning_rate/qwen_vl_interface': 7.324656297177372e-06, 'learning_rate/action_model': 7.324656297177371e-05, 'epoch': 0.76}) +[2026-06-15 22:38:58,483][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.2855340838432312, 'train/grad_norm_pre_clip': 15.941579818725586, 'timing/data': 0.02389103901805356, 'timing/model': 1.2821001379634254, 'learning_rate/qwen_vl_interface': 7.32191629102295e-06, 'learning_rate/action_model': 7.32191629102295e-05, 'epoch': 0.76}) +[2026-06-15 22:39:03,002][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.23130223155021667, 'train/grad_norm_pre_clip': 17.03005027770996, 'timing/data': 0.02396116399904713, 'timing/model': 1.2715967390104197, 'learning_rate/qwen_vl_interface': 7.319175433182004e-06, 'learning_rate/action_model': 7.319175433182004e-05, 'epoch': 0.76}) +[2026-06-15 22:39:07,382][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.16313594579696655, 'train/grad_norm_pre_clip': 17.080005645751953, 'timing/data': 0.023899163003079593, 'timing/model': 1.1370118660270236, 'learning_rate/qwen_vl_interface': 7.316433724781198e-06, 'learning_rate/action_model': 7.316433724781197e-05, 'epoch': 0.76}) +[2026-06-15 22:39:11,944][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.058017242699861526, 'train/grad_norm_pre_clip': 12.620033264160156, 'timing/data': 0.02316734497435391, 'timing/model': 1.1308053610264324, 'learning_rate/qwen_vl_interface': 7.313691166947543e-06, 'learning_rate/action_model': 7.313691166947543e-05, 'epoch': 0.76}) +[2026-06-15 22:39:16,492][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.09461409598588943, 'train/grad_norm_pre_clip': 7.664668560028076, 'timing/data': 0.02407251700060442, 'timing/model': 1.1456266660243273, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.76}) +[2026-06-15 22:39:21,050][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.05154241621494293, 'train/grad_norm_pre_clip': 2.7406089305877686, 'timing/data': 0.023219009046442807, 'timing/model': 1.1396946430322714, 'learning_rate/qwen_vl_interface': 7.308203507491485e-06, 'learning_rate/action_model': 7.308203507491484e-05, 'epoch': 0.76}) +[2026-06-15 22:39:25,566][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.027067260816693306, 'train/grad_norm_pre_clip': 1.5350409746170044, 'timing/data': 0.023454543959815055, 'timing/model': 1.1418472590157762, 'learning_rate/qwen_vl_interface': 7.3054584081248515e-06, 'learning_rate/action_model': 7.305458408124852e-05, 'epoch': 0.76}) +[2026-06-15 22:39:30,138][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.05683998018503189, 'train/grad_norm_pre_clip': 3.391331434249878, 'timing/data': 0.023134315968491137, 'timing/model': 1.3194624789757654, 'learning_rate/qwen_vl_interface': 7.302712463836907e-06, 'learning_rate/action_model': 7.302712463836907e-05, 'epoch': 0.76}) +[2026-06-15 22:39:34,540][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.07569938153028488, 'train/grad_norm_pre_clip': 7.796489238739014, 'timing/data': 0.023267853015568107, 'timing/model': 1.1351122230407782, 'learning_rate/qwen_vl_interface': 7.2999656757564086e-06, 'learning_rate/action_model': 7.299965675756409e-05, 'epoch': 0.76}) +[2026-06-15 22:39:39,087][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.06877194344997406, 'train/grad_norm_pre_clip': 5.932623863220215, 'timing/data': 0.023417214979417622, 'timing/model': 1.1390760189970024, 'learning_rate/qwen_vl_interface': 7.297218045012453e-06, 'learning_rate/action_model': 7.297218045012453e-05, 'epoch': 0.76}) +[2026-06-15 22:39:43,568][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.15621648728847504, 'train/grad_norm_pre_clip': 12.435091972351074, 'timing/data': 0.02321433601900935, 'timing/model': 1.1387354619801044, 'learning_rate/qwen_vl_interface': 7.29446957273449e-06, 'learning_rate/action_model': 7.294469572734489e-05, 'epoch': 0.76}) +[2026-06-15 22:39:48,021][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.1434611827135086, 'train/grad_norm_pre_clip': 7.774564743041992, 'timing/data': 0.023343070992268622, 'timing/model': 1.1442702119820751, 'learning_rate/qwen_vl_interface': 7.2917202600523104e-06, 'learning_rate/action_model': 7.29172026005231e-05, 'epoch': 0.76}) +[2026-06-15 22:39:52,499][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.17187224328517914, 'train/grad_norm_pre_clip': 6.300373554229736, 'timing/data': 0.02333953103516251, 'timing/model': 1.1358442909549922, 'learning_rate/qwen_vl_interface': 7.2889701080960565e-06, 'learning_rate/action_model': 7.288970108096056e-05, 'epoch': 0.76}) +[2026-06-15 22:39:57,017][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.018041079863905907, 'train/grad_norm_pre_clip': 2.2637081146240234, 'timing/data': 0.023392139992211014, 'timing/model': 1.2690036300336942, 'learning_rate/qwen_vl_interface': 7.286219117996209e-06, 'learning_rate/action_model': 7.286219117996207e-05, 'epoch': 0.76}) +[2026-06-15 22:40:01,403][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.026011431589722633, 'train/grad_norm_pre_clip': 7.555608749389648, 'timing/data': 0.023731919995043427, 'timing/model': 1.1346058290218934, 'learning_rate/qwen_vl_interface': 7.283467290883598e-06, 'learning_rate/action_model': 7.283467290883598e-05, 'epoch': 0.76}) +[2026-06-15 22:40:05,955][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.02641942910850048, 'train/grad_norm_pre_clip': 2.3372769355773926, 'timing/data': 0.02340200904291123, 'timing/model': 1.132010283996351, 'learning_rate/qwen_vl_interface': 7.2807146278893925e-06, 'learning_rate/action_model': 7.280714627889393e-05, 'epoch': 0.76}) +[2026-06-15 22:40:10,501][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.28219521045684814, 'train/grad_norm_pre_clip': 18.570068359375, 'timing/data': 0.023475288995541632, 'timing/model': 1.1338596710120328, 'learning_rate/qwen_vl_interface': 7.277961130145111e-06, 'learning_rate/action_model': 7.277961130145111e-05, 'epoch': 0.76}) +[2026-06-15 22:40:15,061][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.09461396187543869, 'train/grad_norm_pre_clip': 8.936774253845215, 'timing/data': 0.024396867025643587, 'timing/model': 1.1357816269737668, 'learning_rate/qwen_vl_interface': 7.2752067987826116e-06, 'learning_rate/action_model': 7.275206798782611e-05, 'epoch': 0.77}) +[2026-06-15 22:40:19,622][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.024641435593366623, 'train/grad_norm_pre_clip': 3.298605442047119, 'timing/data': 0.023770475992932916, 'timing/model': 1.290877811028622, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.77}) +[2026-06-15 22:40:24,008][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.019826684147119522, 'train/grad_norm_pre_clip': 3.211651086807251, 'timing/data': 0.02358654804993421, 'timing/model': 1.1324839179869741, 'learning_rate/qwen_vl_interface': 7.269695639732108e-06, 'learning_rate/action_model': 7.269695639732108e-05, 'epoch': 0.77}) +[2026-06-15 22:40:28,558][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.015800582244992256, 'train/grad_norm_pre_clip': 2.1553876399993896, 'timing/data': 0.02362271398305893, 'timing/model': 1.1358306059846655, 'learning_rate/qwen_vl_interface': 7.266938814309533e-06, 'learning_rate/action_model': 7.266938814309533e-05, 'epoch': 0.77}) +[2026-06-15 22:40:33,060][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.024829139932990074, 'train/grad_norm_pre_clip': 12.73647403717041, 'timing/data': 0.023825126991141587, 'timing/model': 1.1432358039892279, 'learning_rate/qwen_vl_interface': 7.264181159799598e-06, 'learning_rate/action_model': 7.264181159799597e-05, 'epoch': 0.77}) +[2026-06-15 22:40:37,641][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.5140204429626465, 'train/grad_norm_pre_clip': 36.869991302490234, 'timing/data': 0.023488058999646455, 'timing/model': 1.1520568450214341, 'learning_rate/qwen_vl_interface': 7.261422677335871e-06, 'learning_rate/action_model': 7.261422677335871e-05, 'epoch': 0.77}) +[2026-06-15 22:40:42,198][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.17884020507335663, 'train/grad_norm_pre_clip': 14.257965087890625, 'timing/data': 0.023861841997131705, 'timing/model': 1.1386253929813392, 'learning_rate/qwen_vl_interface': 7.258663368052259e-06, 'learning_rate/action_model': 7.258663368052259e-05, 'epoch': 0.77}) +[2026-06-15 22:40:46,736][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.054474227130413055, 'train/grad_norm_pre_clip': 16.426515579223633, 'timing/data': 0.023713990987744182, 'timing/model': 1.2906635259860195, 'learning_rate/qwen_vl_interface': 7.2559032330830105e-06, 'learning_rate/action_model': 7.25590323308301e-05, 'epoch': 0.77}) +[2026-06-15 22:40:51,251][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.005556007847189903, 'train/grad_norm_pre_clip': 0.536191463470459, 'timing/data': 0.023184810008388013, 'timing/model': 1.264854904031381, 'learning_rate/qwen_vl_interface': 7.253142273562712e-06, 'learning_rate/action_model': 7.253142273562712e-05, 'epoch': 0.77}) +[2026-06-15 22:40:55,634][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.044341083616018295, 'train/grad_norm_pre_clip': 6.082284450531006, 'timing/data': 0.02405654900940135, 'timing/model': 1.1343864079681225, 'learning_rate/qwen_vl_interface': 7.25038049062629e-06, 'learning_rate/action_model': 7.250380490626289e-05, 'epoch': 0.77}) +[2026-06-15 22:41:00,195][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.31008821725845337, 'train/grad_norm_pre_clip': 12.687453269958496, 'timing/data': 0.023022877983748913, 'timing/model': 1.1310235079727136, 'learning_rate/qwen_vl_interface': 7.24761788540901e-06, 'learning_rate/action_model': 7.24761788540901e-05, 'epoch': 0.77}) +[2026-06-15 22:41:04,746][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.15901315212249756, 'train/grad_norm_pre_clip': 19.682876586914062, 'timing/data': 0.023543959017843008, 'timing/model': 1.146655542019289, 'learning_rate/qwen_vl_interface': 7.244854459046473e-06, 'learning_rate/action_model': 7.244854459046473e-05, 'epoch': 0.77}) +[2026-06-15 22:41:09,288][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.028121355921030045, 'train/grad_norm_pre_clip': 3.4190616607666016, 'timing/data': 0.02359052002429962, 'timing/model': 1.1335016619996168, 'learning_rate/qwen_vl_interface': 7.24209021267462e-06, 'learning_rate/action_model': 7.24209021267462e-05, 'epoch': 0.77}) +[2026-06-15 22:41:13,841][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.007706720381975174, 'train/grad_norm_pre_clip': 1.637188196182251, 'timing/data': 0.023115435033105314, 'timing/model': 1.2879606849746779, 'learning_rate/qwen_vl_interface': 7.23932514742973e-06, 'learning_rate/action_model': 7.23932514742973e-05, 'epoch': 0.77}) +[2026-06-15 22:41:18,375][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.07498051971197128, 'train/grad_norm_pre_clip': 13.056137084960938, 'timing/data': 0.024446640978567302, 'timing/model': 1.283706501009874, 'learning_rate/qwen_vl_interface': 7.236559264448413e-06, 'learning_rate/action_model': 7.236559264448412e-05, 'epoch': 0.77}) +[2026-06-15 22:41:22,776][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.0017580429557710886, 'train/grad_norm_pre_clip': 0.2125261425971985, 'timing/data': 0.0226604730123654, 'timing/model': 1.1386845459928736, 'learning_rate/qwen_vl_interface': 7.233792564867623e-06, 'learning_rate/action_model': 7.233792564867623e-05, 'epoch': 0.77}) +[2026-06-15 22:41:27,313][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.05748436599969864, 'train/grad_norm_pre_clip': 11.9747314453125, 'timing/data': 0.022920750023331493, 'timing/model': 1.1441808369709179, 'learning_rate/qwen_vl_interface': 7.231025049824644e-06, 'learning_rate/action_model': 7.231025049824643e-05, 'epoch': 0.77}) +[2026-06-15 22:41:31,853][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.03451051563024521, 'train/grad_norm_pre_clip': 9.438106536865234, 'timing/data': 0.023603735957294703, 'timing/model': 1.1324244510033168, 'learning_rate/qwen_vl_interface': 7.228256720457097e-06, 'learning_rate/action_model': 7.228256720457096e-05, 'epoch': 0.77}) +[2026-06-15 22:41:36,380][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.2711079716682434, 'train/grad_norm_pre_clip': 13.186184883117676, 'timing/data': 0.023050313990097493, 'timing/model': 1.130095837987028, 'learning_rate/qwen_vl_interface': 7.225487577902937e-06, 'learning_rate/action_model': 7.225487577902936e-05, 'epoch': 0.77}) +[2026-06-15 22:41:40,869][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.1384652554988861, 'train/grad_norm_pre_clip': 27.645673751831055, 'timing/data': 0.02348175796214491, 'timing/model': 1.1342999460175633, 'learning_rate/qwen_vl_interface': 7.222717623300455e-06, 'learning_rate/action_model': 7.222717623300455e-05, 'epoch': 0.77}) +[2026-06-15 22:41:45,413][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.22517162561416626, 'train/grad_norm_pre_clip': 22.24514389038086, 'timing/data': 0.022749714960809797, 'timing/model': 1.2926070260000415, 'learning_rate/qwen_vl_interface': 7.219946857788273e-06, 'learning_rate/action_model': 7.219946857788272e-05, 'epoch': 0.77}) +[2026-06-15 22:41:49,798][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.22229589521884918, 'train/grad_norm_pre_clip': 14.434231758117676, 'timing/data': 0.022869943000841886, 'timing/model': 1.132721172994934, 'learning_rate/qwen_vl_interface': 7.21717528250535e-06, 'learning_rate/action_model': 7.21717528250535e-05, 'epoch': 0.77}) +[2026-06-15 22:41:54,325][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.003730796743184328, 'train/grad_norm_pre_clip': 0.5129895806312561, 'timing/data': 0.023214576998725533, 'timing/model': 1.1390510050114244, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 0.77}) +[2026-06-15 22:41:58,761][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.14243657886981964, 'train/grad_norm_pre_clip': 8.986546516418457, 'timing/data': 0.023409070970956236, 'timing/model': 1.131933267985005, 'learning_rate/qwen_vl_interface': 7.211629707184768e-06, 'learning_rate/action_model': 7.211629707184768e-05, 'epoch': 0.77}) +[2026-06-15 22:42:03,215][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.17164361476898193, 'train/grad_norm_pre_clip': 20.779260635375977, 'timing/data': 0.023305526992771775, 'timing/model': 1.1343920590006746, 'learning_rate/qwen_vl_interface': 7.2088557094266885e-06, 'learning_rate/action_model': 7.208855709426688e-05, 'epoch': 0.78}) +[2026-06-15 22:42:07,694][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.13387998938560486, 'train/grad_norm_pre_clip': 7.271656513214111, 'timing/data': 0.023096688964869827, 'timing/model': 1.1331967499572784, 'learning_rate/qwen_vl_interface': 7.2060809064570195e-06, 'learning_rate/action_model': 7.20608090645702e-05, 'epoch': 0.78}) +[2026-06-15 22:42:12,221][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.0197064857929945, 'train/grad_norm_pre_clip': 7.0418701171875, 'timing/data': 0.023094676027540118, 'timing/model': 1.275437849049922, 'learning_rate/qwen_vl_interface': 7.203305299416375e-06, 'learning_rate/action_model': 7.203305299416375e-05, 'epoch': 0.78}) +[2026-06-15 22:42:16,611][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.1378878653049469, 'train/grad_norm_pre_clip': 16.115673065185547, 'timing/data': 0.02274789399234578, 'timing/model': 1.1340618599788286, 'learning_rate/qwen_vl_interface': 7.200528889445707e-06, 'learning_rate/action_model': 7.200528889445706e-05, 'epoch': 0.78}) +[2026-06-15 22:42:21,148][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.017120951786637306, 'train/grad_norm_pre_clip': 14.854704856872559, 'timing/data': 0.02355724899098277, 'timing/model': 1.1289624189957976, 'learning_rate/qwen_vl_interface': 7.19775167768629e-06, 'learning_rate/action_model': 7.19775167768629e-05, 'epoch': 0.78}) +[2026-06-15 22:42:25,685][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.2163766771554947, 'train/grad_norm_pre_clip': 18.608890533447266, 'timing/data': 0.023464591009542346, 'timing/model': 1.1293012829846703, 'learning_rate/qwen_vl_interface': 7.194973665279731e-06, 'learning_rate/action_model': 7.19497366527973e-05, 'epoch': 0.78}) +[2026-06-15 22:42:30,233][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.38550442457199097, 'train/grad_norm_pre_clip': 12.203667640686035, 'timing/data': 0.0239448850043118, 'timing/model': 1.1303475319873542, 'learning_rate/qwen_vl_interface': 7.1921948533679666e-06, 'learning_rate/action_model': 7.192194853367967e-05, 'epoch': 0.78}) +[2026-06-15 22:42:34,781][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.16901439428329468, 'train/grad_norm_pre_clip': 17.90843963623047, 'timing/data': 0.02283642499241978, 'timing/model': 1.2958897330099717, 'learning_rate/qwen_vl_interface': 7.1894152430932605e-06, 'learning_rate/action_model': 7.189415243093261e-05, 'epoch': 0.78}) +[2026-06-15 22:42:39,160][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.03128178045153618, 'train/grad_norm_pre_clip': 8.149686813354492, 'timing/data': 0.023501075978856534, 'timing/model': 1.140103690966498, 'learning_rate/qwen_vl_interface': 7.186634835598205e-06, 'learning_rate/action_model': 7.186634835598205e-05, 'epoch': 0.78}) +[2026-06-15 22:42:43,727][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.014947295188903809, 'train/grad_norm_pre_clip': 4.159446716308594, 'timing/data': 0.02281554799992591, 'timing/model': 1.1335452980129048, 'learning_rate/qwen_vl_interface': 7.183853632025724e-06, 'learning_rate/action_model': 7.183853632025723e-05, 'epoch': 0.78}) +[2026-06-15 22:42:48,211][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.019627511501312256, 'train/grad_norm_pre_clip': 6.132379055023193, 'timing/data': 0.0234478049678728, 'timing/model': 1.1366357480292208, 'learning_rate/qwen_vl_interface': 7.181071633519062e-06, 'learning_rate/action_model': 7.181071633519062e-05, 'epoch': 0.78}) +[2026-06-15 22:42:52,761][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.007926375605165958, 'train/grad_norm_pre_clip': 1.7660530805587769, 'timing/data': 0.02312648599036038, 'timing/model': 1.1327905789949, 'learning_rate/qwen_vl_interface': 7.178288841221795e-06, 'learning_rate/action_model': 7.178288841221794e-05, 'epoch': 0.78}) +[2026-06-15 22:42:57,302][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.03564320132136345, 'train/grad_norm_pre_clip': 4.630810737609863, 'timing/data': 0.023313443001825362, 'timing/model': 1.1345614889869466, 'learning_rate/qwen_vl_interface': 7.1755052562778215e-06, 'learning_rate/action_model': 7.175505256277822e-05, 'epoch': 0.78}) +[2026-06-15 22:47:17,605][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.05473052337765694, 'train/grad_norm_pre_clip': 4.445443630218506, 'eval/loss': 0.06857369095087051, 'eval/flap_precision': 0.9499241274658573, 'eval/flap_recall': 0.9660493827160493, 'eval/flap_f1': 0.9579188982402448, 'eval/flap_rate_model': 0.18305555555555555, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9579188982402448, 'timing/data': 0.022558867989573628, 'timing/model': 1.2950202350039035, 'learning_rate/qwen_vl_interface': 7.1727208798313704e-06, 'learning_rate/action_model': 7.172720879831371e-05, 'epoch': 0.78}) +[2026-06-15 22:47:17,605][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1900_state +[2026-06-15 22:47:17,608][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 22:47:21,204][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1900_state/model.safetensors +[2026-06-15 22:47:31,290][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1900_state/optimizer.bin +[2026-06-15 22:47:31,291][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1900_state/sampler.bin +[2026-06-15 22:47:31,291][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1900_state/sampler_1.bin +[2026-06-15 22:47:31,291][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_1900_state/random_states_0.pkl +[2026-06-15 22:47:31,333][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1900 +[2026-06-15 22:47:31,334][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 22:47:31,340][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 22:47:35,800][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.015822524204850197, 'train/grad_norm_pre_clip': 2.311061382293701, 'timing/data': 0.02352958597475663, 'timing/model': 1.1096017739619128, 'learning_rate/qwen_vl_interface': 7.169935713026994e-06, 'learning_rate/action_model': 7.169935713026994e-05, 'epoch': 0.78}) +[2026-06-15 22:47:40,272][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.245494082570076, 'train/grad_norm_pre_clip': 11.018253326416016, 'timing/data': 0.022974151012022048, 'timing/model': 1.2776329559856094, 'learning_rate/qwen_vl_interface': 7.1671497570095684e-06, 'learning_rate/action_model': 7.167149757009568e-05, 'epoch': 0.78}) +[2026-06-15 22:47:44,583][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.6214887499809265, 'train/grad_norm_pre_clip': 21.956235885620117, 'timing/data': 0.023537219036370516, 'timing/model': 1.1132234549731947, 'learning_rate/qwen_vl_interface': 7.164363012924295e-06, 'learning_rate/action_model': 7.164363012924294e-05, 'epoch': 0.78}) +[2026-06-15 22:47:49,068][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.052251193672418594, 'train/grad_norm_pre_clip': 15.577451705932617, 'timing/data': 0.023122806975152344, 'timing/model': 1.1175808050320484, 'learning_rate/qwen_vl_interface': 7.161575481916699e-06, 'learning_rate/action_model': 7.1615754819167e-05, 'epoch': 0.78}) +[2026-06-15 22:47:53,551][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.06467626988887787, 'train/grad_norm_pre_clip': 9.266031265258789, 'timing/data': 0.02346399601083249, 'timing/model': 1.1255816979683004, 'learning_rate/qwen_vl_interface': 7.158787165132627e-06, 'learning_rate/action_model': 7.158787165132626e-05, 'epoch': 0.78}) +[2026-06-15 22:47:58,031][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.16474002599716187, 'train/grad_norm_pre_clip': 11.404769897460938, 'timing/data': 0.023779935960192233, 'timing/model': 1.1226359980064444, 'learning_rate/qwen_vl_interface': 7.1559980637182535e-06, 'learning_rate/action_model': 7.155998063718254e-05, 'epoch': 0.78}) +[2026-06-15 22:48:02,489][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.21691405773162842, 'train/grad_norm_pre_clip': 12.546144485473633, 'timing/data': 0.023237733985297382, 'timing/model': 1.1242573740310036, 'learning_rate/qwen_vl_interface': 7.153208178820071e-06, 'learning_rate/action_model': 7.153208178820072e-05, 'epoch': 0.78}) +[2026-06-15 22:48:06,991][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.0746043249964714, 'train/grad_norm_pre_clip': 20.729795455932617, 'timing/data': 0.022869044041726738, 'timing/model': 1.2908041430055164, 'learning_rate/qwen_vl_interface': 7.150417511584898e-06, 'learning_rate/action_model': 7.150417511584898e-05, 'epoch': 0.78}) +[2026-06-15 22:48:11,336][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.12457337230443954, 'train/grad_norm_pre_clip': 7.988036632537842, 'timing/data': 0.023165077960584313, 'timing/model': 1.1260494810412638, 'learning_rate/qwen_vl_interface': 7.147626063159872e-06, 'learning_rate/action_model': 7.147626063159872e-05, 'epoch': 0.78}) +[2026-06-15 22:48:15,842][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.0417213961482048, 'train/grad_norm_pre_clip': 13.465373992919922, 'timing/data': 0.023659045982640237, 'timing/model': 1.124883524957113, 'learning_rate/qwen_vl_interface': 7.14483383469245e-06, 'learning_rate/action_model': 7.144833834692449e-05, 'epoch': 0.78}) +[2026-06-15 22:48:20,369][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.016564136371016502, 'train/grad_norm_pre_clip': 1.055203914642334, 'timing/data': 0.023052469012327492, 'timing/model': 1.1372763850376941, 'learning_rate/qwen_vl_interface': 7.142040827330411e-06, 'learning_rate/action_model': 7.142040827330411e-05, 'epoch': 0.79}) +[2026-06-15 22:48:24,878][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.14489199221134186, 'train/grad_norm_pre_clip': 11.996740341186523, 'timing/data': 0.0313462259946391, 'timing/model': 1.1267329070251435, 'learning_rate/qwen_vl_interface': 7.139247042221859e-06, 'learning_rate/action_model': 7.139247042221858e-05, 'epoch': 0.79}) +[2026-06-15 22:48:29,410][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.14145362377166748, 'train/grad_norm_pre_clip': 8.412364959716797, 'timing/data': 0.02306010405300185, 'timing/model': 1.129645696026273, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.79}) +[2026-06-15 22:48:33,938][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.05228210240602493, 'train/grad_norm_pre_clip': 15.552258491516113, 'timing/data': 0.023157306015491486, 'timing/model': 1.2774431960424408, 'learning_rate/qwen_vl_interface': 7.133657143359203e-06, 'learning_rate/action_model': 7.133657143359202e-05, 'epoch': 0.79}) +[2026-06-15 22:48:38,301][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.09091020375490189, 'train/grad_norm_pre_clip': 10.162280082702637, 'timing/data': 0.023104760970454663, 'timing/model': 1.129774141998496, 'learning_rate/qwen_vl_interface': 7.130861031902897e-06, 'learning_rate/action_model': 7.130861031902897e-05, 'epoch': 0.79}) +[2026-06-15 22:48:42,861][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.13069307804107666, 'train/grad_norm_pre_clip': 8.675180435180664, 'timing/data': 0.023450906039215624, 'timing/model': 1.1263275789679028, 'learning_rate/qwen_vl_interface': 7.128064147295667e-06, 'learning_rate/action_model': 7.128064147295667e-05, 'epoch': 0.79}) +[2026-06-15 22:48:47,370][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.027561726048588753, 'train/grad_norm_pre_clip': 9.941244125366211, 'timing/data': 0.02376347000245005, 'timing/model': 1.1295554820098914, 'learning_rate/qwen_vl_interface': 7.125266490687206e-06, 'learning_rate/action_model': 7.125266490687206e-05, 'epoch': 0.79}) +[2026-06-15 22:48:51,801][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.04669974744319916, 'train/grad_norm_pre_clip': 10.546616554260254, 'timing/data': 0.0238021650002338, 'timing/model': 1.1276865749969147, 'learning_rate/qwen_vl_interface': 7.1224680632275265e-06, 'learning_rate/action_model': 7.122468063227527e-05, 'epoch': 0.79}) +[2026-06-15 22:48:56,284][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.22141899168491364, 'train/grad_norm_pre_clip': 9.929972648620605, 'timing/data': 0.02350521698826924, 'timing/model': 1.1328606369788758, 'learning_rate/qwen_vl_interface': 7.1196688660669555e-06, 'learning_rate/action_model': 7.119668866066955e-05, 'epoch': 0.79}) +[2026-06-15 22:49:00,831][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.09575816988945007, 'train/grad_norm_pre_clip': 12.324694633483887, 'timing/data': 0.023874105012509972, 'timing/model': 1.2970871559809893, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.79}) +[2026-06-15 22:49:05,215][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.0563487708568573, 'train/grad_norm_pre_clip': 8.678536415100098, 'timing/data': 0.023393757990561426, 'timing/model': 1.1326076670084149, 'learning_rate/qwen_vl_interface': 7.114068167246033e-06, 'learning_rate/action_model': 7.114068167246033e-05, 'epoch': 0.79}) +[2026-06-15 22:49:09,749][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.009571585804224014, 'train/grad_norm_pre_clip': 1.0085927248001099, 'timing/data': 0.023352022981271148, 'timing/model': 1.1317815619986504, 'learning_rate/qwen_vl_interface': 7.111266667887916e-06, 'learning_rate/action_model': 7.111266667887915e-05, 'epoch': 0.79}) +[2026-06-15 22:49:14,308][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.08239635825157166, 'train/grad_norm_pre_clip': 16.98652458190918, 'timing/data': 0.023671959002967924, 'timing/model': 1.1532708409940824, 'learning_rate/qwen_vl_interface': 7.108464403433378e-06, 'learning_rate/action_model': 7.108464403433378e-05, 'epoch': 0.79}) +[2026-06-15 22:49:18,869][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.23581023514270782, 'train/grad_norm_pre_clip': 13.136130332946777, 'timing/data': 0.023148801003117114, 'timing/model': 1.1314816800295375, 'learning_rate/qwen_vl_interface': 7.105661375034324e-06, 'learning_rate/action_model': 7.105661375034324e-05, 'epoch': 0.79}) +[2026-06-15 22:49:23,368][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.02041327953338623, 'train/grad_norm_pre_clip': 3.0449764728546143, 'timing/data': 0.023607812996488065, 'timing/model': 1.1301590540097095, 'learning_rate/qwen_vl_interface': 7.102857583842973e-06, 'learning_rate/action_model': 7.102857583842974e-05, 'epoch': 0.79}) +[2026-06-15 22:49:27,924][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.02329253777861595, 'train/grad_norm_pre_clip': 8.42114543914795, 'timing/data': 0.024305257014930248, 'timing/model': 1.3061223510303535, 'learning_rate/qwen_vl_interface': 7.100053031011859e-06, 'learning_rate/action_model': 7.100053031011859e-05, 'epoch': 0.79}) +[2026-06-15 22:49:32,299][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.17318294942378998, 'train/grad_norm_pre_clip': 14.801916122436523, 'timing/data': 0.02341648197034374, 'timing/model': 1.136704474047292, 'learning_rate/qwen_vl_interface': 7.097247717693827e-06, 'learning_rate/action_model': 7.097247717693827e-05, 'epoch': 0.79}) +[2026-06-15 22:49:36,839][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.012358924373984337, 'train/grad_norm_pre_clip': 3.611920118331909, 'timing/data': 0.02333394601009786, 'timing/model': 1.1336221699602902, 'learning_rate/qwen_vl_interface': 7.094441645042033e-06, 'learning_rate/action_model': 7.094441645042032e-05, 'epoch': 0.79}) +[2026-06-15 22:49:41,279][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.2889244258403778, 'train/grad_norm_pre_clip': 11.731685638427734, 'timing/data': 0.023351312964223325, 'timing/model': 1.1377038760110736, 'learning_rate/qwen_vl_interface': 7.09163481420995e-06, 'learning_rate/action_model': 7.09163481420995e-05, 'epoch': 0.79}) +[2026-06-15 22:49:45,729][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.14724017679691315, 'train/grad_norm_pre_clip': 13.532946586608887, 'timing/data': 0.023453668982256204, 'timing/model': 1.1358948089764453, 'learning_rate/qwen_vl_interface': 7.0888272263513595e-06, 'learning_rate/action_model': 7.088827226351359e-05, 'epoch': 0.79}) +[2026-06-15 22:49:50,214][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.0110938036814332, 'train/grad_norm_pre_clip': 4.040971755981445, 'timing/data': 0.02333783300127834, 'timing/model': 1.135109142982401, 'learning_rate/qwen_vl_interface': 7.086018882620354e-06, 'learning_rate/action_model': 7.086018882620354e-05, 'epoch': 0.79}) +[2026-06-15 22:49:54,716][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.08173097670078278, 'train/grad_norm_pre_clip': 27.06617546081543, 'timing/data': 0.023374218959361315, 'timing/model': 1.258566819014959, 'learning_rate/qwen_vl_interface': 7.083209784171341e-06, 'learning_rate/action_model': 7.083209784171341e-05, 'epoch': 0.79}) +[2026-06-15 22:49:59,105][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.13433490693569183, 'train/grad_norm_pre_clip': 10.199387550354004, 'timing/data': 0.024229595030192286, 'timing/model': 1.138446543016471, 'learning_rate/qwen_vl_interface': 7.080399932159029e-06, 'learning_rate/action_model': 7.080399932159029e-05, 'epoch': 0.79}) +[2026-06-15 22:50:03,649][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.02519233524799347, 'train/grad_norm_pre_clip': 4.648251533508301, 'timing/data': 0.023696406977251172, 'timing/model': 1.133701684011612, 'learning_rate/qwen_vl_interface': 7.077589327738448e-06, 'learning_rate/action_model': 7.077589327738447e-05, 'epoch': 0.79}) +[2026-06-15 22:50:08,197][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.1333671659231186, 'train/grad_norm_pre_clip': 14.941258430480957, 'timing/data': 0.023159576987382025, 'timing/model': 1.1311602480127476, 'learning_rate/qwen_vl_interface': 7.0747779720649265e-06, 'learning_rate/action_model': 7.074777972064927e-05, 'epoch': 0.79}) +[2026-06-15 22:50:12,720][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.20657508075237274, 'train/grad_norm_pre_clip': 19.119266510009766, 'timing/data': 0.02336572902277112, 'timing/model': 1.1325237390119582, 'learning_rate/qwen_vl_interface': 7.071965866294108e-06, 'learning_rate/action_model': 7.071965866294108e-05, 'epoch': 0.8}) +[2026-06-15 22:50:17,267][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.11587509512901306, 'train/grad_norm_pre_clip': 19.8780574798584, 'timing/data': 0.023854942002799362, 'timing/model': 1.2980250879772939, 'learning_rate/qwen_vl_interface': 7.0691530115819454e-06, 'learning_rate/action_model': 7.069153011581945e-05, 'epoch': 0.8}) +[2026-06-15 22:50:21,658][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.0050445483066141605, 'train/grad_norm_pre_clip': 0.5439889430999756, 'timing/data': 0.02430598804494366, 'timing/model': 1.1403342399862595, 'learning_rate/qwen_vl_interface': 7.066339409084694e-06, 'learning_rate/action_model': 7.066339409084694e-05, 'epoch': 0.8}) +[2026-06-15 22:50:26,180][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.006390373222529888, 'train/grad_norm_pre_clip': 0.7341881990432739, 'timing/data': 0.0234692040248774, 'timing/model': 1.1314747250289656, 'learning_rate/qwen_vl_interface': 7.0635250599589215e-06, 'learning_rate/action_model': 7.06352505995892e-05, 'epoch': 0.8}) +[2026-06-15 22:50:30,669][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.09530066698789597, 'train/grad_norm_pre_clip': 12.686308860778809, 'timing/data': 0.023390441958326846, 'timing/model': 1.1323823219863698, 'learning_rate/qwen_vl_interface': 7.060709965361499e-06, 'learning_rate/action_model': 7.060709965361499e-05, 'epoch': 0.8}) +[2026-06-15 22:50:35,212][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.00642910273745656, 'train/grad_norm_pre_clip': 1.5719268321990967, 'timing/data': 0.023198153998237103, 'timing/model': 1.1332592409453355, 'learning_rate/qwen_vl_interface': 7.057894126449606e-06, 'learning_rate/action_model': 7.057894126449606e-05, 'epoch': 0.8}) +[2026-06-15 22:50:39,753][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.2598366141319275, 'train/grad_norm_pre_clip': 36.16913604736328, 'timing/data': 0.023470939020626247, 'timing/model': 1.1393991409568116, 'learning_rate/qwen_vl_interface': 7.055077544380728e-06, 'learning_rate/action_model': 7.055077544380728e-05, 'epoch': 0.8}) +[2026-06-15 22:50:44,304][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.018212629482150078, 'train/grad_norm_pre_clip': 4.442198753356934, 'timing/data': 0.023430683999322355, 'timing/model': 1.2990900960285217, 'learning_rate/qwen_vl_interface': 7.0522602203126554e-06, 'learning_rate/action_model': 7.052260220312656e-05, 'epoch': 0.8}) +[2026-06-15 22:50:48,804][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.03058886155486107, 'train/grad_norm_pre_clip': 7.027292728424072, 'timing/data': 0.02431465999688953, 'timing/model': 1.2487250749836676, 'learning_rate/qwen_vl_interface': 7.0494421554034845e-06, 'learning_rate/action_model': 7.049442155403484e-05, 'epoch': 0.8}) +[2026-06-15 22:50:53,188][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.01537727564573288, 'train/grad_norm_pre_clip': 4.363612651824951, 'timing/data': 0.022740676999092102, 'timing/model': 1.1334000600036234, 'learning_rate/qwen_vl_interface': 7.046623350811612e-06, 'learning_rate/action_model': 7.046623350811612e-05, 'epoch': 0.8}) +[2026-06-15 22:50:57,742][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.005452908109873533, 'train/grad_norm_pre_clip': 0.7170346975326538, 'timing/data': 0.02375767903868109, 'timing/model': 1.134106001991313, 'learning_rate/qwen_vl_interface': 7.043803807695747e-06, 'learning_rate/action_model': 7.043803807695747e-05, 'epoch': 0.8}) +[2026-06-15 22:51:02,303][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.14086352288722992, 'train/grad_norm_pre_clip': 17.81112289428711, 'timing/data': 0.022626321006100625, 'timing/model': 1.1367363010067493, 'learning_rate/qwen_vl_interface': 7.040983527214893e-06, 'learning_rate/action_model': 7.040983527214893e-05, 'epoch': 0.8}) +[2026-06-15 22:51:06,853][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.056572332978248596, 'train/grad_norm_pre_clip': 12.638711929321289, 'timing/data': 0.022611792024690658, 'timing/model': 1.1369472370133735, 'learning_rate/qwen_vl_interface': 7.038162510528363e-06, 'learning_rate/action_model': 7.038162510528362e-05, 'epoch': 0.8}) +[2026-06-15 22:51:11,395][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.15437771379947662, 'train/grad_norm_pre_clip': 14.65811538696289, 'timing/data': 0.023041075037326664, 'timing/model': 1.3005235270247795, 'learning_rate/qwen_vl_interface': 7.03534075879577e-06, 'learning_rate/action_model': 7.03534075879577e-05, 'epoch': 0.8}) +[2026-06-15 22:51:15,914][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.2796635329723358, 'train/grad_norm_pre_clip': 9.845842361450195, 'timing/data': 0.022879974974785, 'timing/model': 1.2711939479922876, 'learning_rate/qwen_vl_interface': 7.032518273177028e-06, 'learning_rate/action_model': 7.032518273177028e-05, 'epoch': 0.8}) +[2026-06-15 22:51:20,320][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.5154845118522644, 'train/grad_norm_pre_clip': 17.258554458618164, 'timing/data': 0.022844125982373953, 'timing/model': 1.152191964967642, 'learning_rate/qwen_vl_interface': 7.029695054832357e-06, 'learning_rate/action_model': 7.029695054832357e-05, 'epoch': 0.8}) +[2026-06-15 22:51:24,853][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.47895103693008423, 'train/grad_norm_pre_clip': 14.590483665466309, 'timing/data': 0.023054812045302242, 'timing/model': 1.1325175979873165, 'learning_rate/qwen_vl_interface': 7.026871104922273e-06, 'learning_rate/action_model': 7.026871104922273e-05, 'epoch': 0.8}) +[2026-06-15 22:51:29,393][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.02206277661025524, 'train/grad_norm_pre_clip': 13.902301788330078, 'timing/data': 0.022888657986186445, 'timing/model': 1.1328479300136678, 'learning_rate/qwen_vl_interface': 7.024046424607596e-06, 'learning_rate/action_model': 7.024046424607595e-05, 'epoch': 0.8}) +[2026-06-15 22:51:33,917][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.029605351388454437, 'train/grad_norm_pre_clip': 13.698281288146973, 'timing/data': 0.022762170992791653, 'timing/model': 1.139789336011745, 'learning_rate/qwen_vl_interface': 7.021221015049445e-06, 'learning_rate/action_model': 7.021221015049446e-05, 'epoch': 0.8}) +[2026-06-15 22:51:38,440][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.19300299882888794, 'train/grad_norm_pre_clip': 17.26393699645996, 'timing/data': 0.023138102958910167, 'timing/model': 1.1479784530238248, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 0.8}) +[2026-06-15 22:51:43,016][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.11663486063480377, 'train/grad_norm_pre_clip': 16.699190139770508, 'timing/data': 0.02336402499349788, 'timing/model': 1.284018529986497, 'learning_rate/qwen_vl_interface': 7.015568012848699e-06, 'learning_rate/action_model': 7.015568012848699e-05, 'epoch': 0.8}) +[2026-06-15 22:51:47,423][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.034869544208049774, 'train/grad_norm_pre_clip': 15.788537979125977, 'timing/data': 0.030880464997608215, 'timing/model': 1.1386339060263708, 'learning_rate/qwen_vl_interface': 7.01274042252984e-06, 'learning_rate/action_model': 7.01274042252984e-05, 'epoch': 0.8}) +[2026-06-15 22:51:51,962][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.0014651634264737368, 'train/grad_norm_pre_clip': 0.12295787781476974, 'timing/data': 0.02349389501614496, 'timing/model': 1.1364471850101836, 'learning_rate/qwen_vl_interface': 7.009912107614978e-06, 'learning_rate/action_model': 7.009912107614978e-05, 'epoch': 0.8}) +[2026-06-15 22:51:56,409][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.2364611029624939, 'train/grad_norm_pre_clip': 13.1624755859375, 'timing/data': 0.023753917950671166, 'timing/model': 1.1358415950089693, 'learning_rate/qwen_vl_interface': 7.0070830692667244e-06, 'learning_rate/action_model': 7.007083069266724e-05, 'epoch': 0.8}) +[2026-06-15 22:52:00,865][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.1305752396583557, 'train/grad_norm_pre_clip': 23.36011505126953, 'timing/data': 0.023357069003395736, 'timing/model': 1.1376432269462384, 'learning_rate/qwen_vl_interface': 7.004253308647993e-06, 'learning_rate/action_model': 7.004253308647994e-05, 'epoch': 0.81}) +[2026-06-15 22:52:05,395][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.13252760469913483, 'train/grad_norm_pre_clip': 17.61566925048828, 'timing/data': 0.030571261944714934, 'timing/model': 1.1420969349564984, 'learning_rate/qwen_vl_interface': 7.001422826921991e-06, 'learning_rate/action_model': 7.00142282692199e-05, 'epoch': 0.81}) +[2026-06-15 22:52:09,918][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.12660205364227295, 'train/grad_norm_pre_clip': 13.980860710144043, 'timing/data': 0.023771229025442153, 'timing/model': 1.2642111749737523, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 0.81}) +[2026-06-15 22:52:14,311][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.09087309241294861, 'train/grad_norm_pre_clip': 11.310171127319336, 'timing/data': 0.02327934099594131, 'timing/model': 1.13755545305321, 'learning_rate/qwen_vl_interface': 6.995759704802485e-06, 'learning_rate/action_model': 6.995759704802485e-05, 'epoch': 0.81}) +[2026-06-15 22:52:18,878][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.04396579787135124, 'train/grad_norm_pre_clip': 8.33335018157959, 'timing/data': 0.02350831701187417, 'timing/model': 1.149020564975217, 'learning_rate/qwen_vl_interface': 6.992927066736876e-06, 'learning_rate/action_model': 6.992927066736875e-05, 'epoch': 0.81}) +[2026-06-15 22:52:23,432][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.40037912130355835, 'train/grad_norm_pre_clip': 19.72947120666504, 'timing/data': 0.02354174799984321, 'timing/model': 1.1355161810060963, 'learning_rate/qwen_vl_interface': 6.990093712219787e-06, 'learning_rate/action_model': 6.990093712219787e-05, 'epoch': 0.81}) +[2026-06-15 22:52:28,017][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.2721661925315857, 'train/grad_norm_pre_clip': 11.44880485534668, 'timing/data': 0.023059880011714995, 'timing/model': 1.1389606299926527, 'learning_rate/qwen_vl_interface': 6.9872596424159e-06, 'learning_rate/action_model': 6.9872596424159e-05, 'epoch': 0.81}) +[2026-06-15 22:52:32,584][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.12085936218500137, 'train/grad_norm_pre_clip': 6.000768661499023, 'timing/data': 0.022901082003954798, 'timing/model': 1.306153326004278, 'learning_rate/qwen_vl_interface': 6.9844248584902e-06, 'learning_rate/action_model': 6.984424858490199e-05, 'epoch': 0.81}) +[2026-06-15 22:52:36,968][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.07762216031551361, 'train/grad_norm_pre_clip': 7.097909450531006, 'timing/data': 0.02328498795395717, 'timing/model': 1.132935864967294, 'learning_rate/qwen_vl_interface': 6.981589361607952e-06, 'learning_rate/action_model': 6.981589361607953e-05, 'epoch': 0.81}) +[2026-06-15 22:52:41,525][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.10724171251058578, 'train/grad_norm_pre_clip': 9.666487693786621, 'timing/data': 0.02329548803390935, 'timing/model': 1.147804902982898, 'learning_rate/qwen_vl_interface': 6.9787531529347274e-06, 'learning_rate/action_model': 6.978753152934727e-05, 'epoch': 0.81}) +[2026-06-15 22:52:46,028][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.24218708276748657, 'train/grad_norm_pre_clip': 11.808158874511719, 'timing/data': 0.023518765985500067, 'timing/model': 1.1362787749967538, 'learning_rate/qwen_vl_interface': 6.97591623363638e-06, 'learning_rate/action_model': 6.97591623363638e-05, 'epoch': 0.81}) +[2026-06-15 22:52:50,588][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.0151039008051157, 'train/grad_norm_pre_clip': 2.3578879833221436, 'timing/data': 0.023136673029512167, 'timing/model': 1.1389272600063123, 'learning_rate/qwen_vl_interface': 6.9730786048790644e-06, 'learning_rate/action_model': 6.973078604879064e-05, 'epoch': 0.81}) +[2026-06-15 22:52:55,153][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.17215383052825928, 'train/grad_norm_pre_clip': 7.048165321350098, 'timing/data': 0.023611871001776308, 'timing/model': 1.1417159180273302, 'learning_rate/qwen_vl_interface': 6.970240267829221e-06, 'learning_rate/action_model': 6.970240267829222e-05, 'epoch': 0.81}) +[2026-06-15 22:52:59,714][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.2216041088104248, 'train/grad_norm_pre_clip': 15.715951919555664, 'timing/data': 0.023463828023523092, 'timing/model': 1.2979655300150625, 'learning_rate/qwen_vl_interface': 6.967401223653582e-06, 'learning_rate/action_model': 6.967401223653581e-05, 'epoch': 0.81}) +[2026-06-15 22:53:04,217][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.07840946316719055, 'train/grad_norm_pre_clip': 8.038824081420898, 'timing/data': 0.023196569993160665, 'timing/model': 1.2546850009821355, 'learning_rate/qwen_vl_interface': 6.9645614735191724e-06, 'learning_rate/action_model': 6.964561473519172e-05, 'epoch': 0.81}) +[2026-06-15 22:53:08,598][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.0549851693212986, 'train/grad_norm_pre_clip': 4.063310146331787, 'timing/data': 0.023319763015024364, 'timing/model': 1.137055879982654, 'learning_rate/qwen_vl_interface': 6.961721018593305e-06, 'learning_rate/action_model': 6.961721018593305e-05, 'epoch': 0.81}) +[2026-06-15 22:53:13,165][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.09391278028488159, 'train/grad_norm_pre_clip': 5.419262886047363, 'timing/data': 0.024030927976127714, 'timing/model': 1.135853947023861, 'learning_rate/qwen_vl_interface': 6.958879860043585e-06, 'learning_rate/action_model': 6.958879860043585e-05, 'epoch': 0.81}) +[2026-06-15 22:53:17,723][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.02912631630897522, 'train/grad_norm_pre_clip': 2.707451581954956, 'timing/data': 0.02267097099684179, 'timing/model': 1.1386835309676826, 'learning_rate/qwen_vl_interface': 6.9560379990379025e-06, 'learning_rate/action_model': 6.956037999037902e-05, 'epoch': 0.81}) +[2026-06-15 22:53:22,284][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.0829584077000618, 'train/grad_norm_pre_clip': 3.819289207458496, 'timing/data': 0.023118794022593647, 'timing/model': 1.143011242034845, 'learning_rate/qwen_vl_interface': 6.953195436744442e-06, 'learning_rate/action_model': 6.953195436744442e-05, 'epoch': 0.81}) +[2026-06-15 22:53:26,828][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.07142333686351776, 'train/grad_norm_pre_clip': 8.285621643066406, 'timing/data': 0.023006767965853214, 'timing/model': 1.291588607011363, 'learning_rate/qwen_vl_interface': 6.950352174331676e-06, 'learning_rate/action_model': 6.950352174331675e-05, 'epoch': 0.81}) +[2026-06-15 22:53:31,368][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.3827550411224365, 'train/grad_norm_pre_clip': 11.941201210021973, 'timing/data': 0.023042962013278157, 'timing/model': 1.2861638540052809, 'learning_rate/qwen_vl_interface': 6.947508212968356e-06, 'learning_rate/action_model': 6.947508212968356e-05, 'epoch': 0.81}) +[2026-06-15 22:53:35,757][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.05484209582209587, 'train/grad_norm_pre_clip': 5.420950412750244, 'timing/data': 0.02311949897557497, 'timing/model': 1.1380035469774157, 'learning_rate/qwen_vl_interface': 6.9446635538235295e-06, 'learning_rate/action_model': 6.944663553823529e-05, 'epoch': 0.81}) +[2026-06-15 22:53:40,332][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.02877565659582615, 'train/grad_norm_pre_clip': 2.6868631839752197, 'timing/data': 0.027590194018557668, 'timing/model': 1.1435475650359876, 'learning_rate/qwen_vl_interface': 6.94181819806653e-06, 'learning_rate/action_model': 6.94181819806653e-05, 'epoch': 0.81}) +[2026-06-15 22:53:44,881][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.0396270714700222, 'train/grad_norm_pre_clip': 20.44324493408203, 'timing/data': 0.023190340027213097, 'timing/model': 1.1349676639656536, 'learning_rate/qwen_vl_interface': 6.938972146866975e-06, 'learning_rate/action_model': 6.938972146866974e-05, 'epoch': 0.81}) +[2026-06-15 22:53:49,430][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.08658760040998459, 'train/grad_norm_pre_clip': 4.556253910064697, 'timing/data': 0.023235887987539172, 'timing/model': 1.141099824977573, 'learning_rate/qwen_vl_interface': 6.936125401394769e-06, 'learning_rate/action_model': 6.936125401394768e-05, 'epoch': 0.82}) +[2026-06-15 22:53:53,955][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.04674714431166649, 'train/grad_norm_pre_clip': 10.585152626037598, 'timing/data': 0.023259546025656164, 'timing/model': 1.1326881669810973, 'learning_rate/qwen_vl_interface': 6.933277962820101e-06, 'learning_rate/action_model': 6.9332779628201e-05, 'epoch': 0.82}) +[2026-06-15 22:53:58,509][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.20613262057304382, 'train/grad_norm_pre_clip': 10.542140007019043, 'timing/data': 0.023418421973474324, 'timing/model': 1.2950373810017481, 'learning_rate/qwen_vl_interface': 6.930429832313444e-06, 'learning_rate/action_model': 6.930429832313444e-05, 'epoch': 0.82}) +[2026-06-15 22:54:02,899][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.0453643836081028, 'train/grad_norm_pre_clip': 9.699543952941895, 'timing/data': 0.02324666001368314, 'timing/model': 1.1354929709923454, 'learning_rate/qwen_vl_interface': 6.92758101104556e-06, 'learning_rate/action_model': 6.92758101104556e-05, 'epoch': 0.82}) +[2026-06-15 22:54:07,438][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.008383008651435375, 'train/grad_norm_pre_clip': 0.8463812470436096, 'timing/data': 0.023353038006462157, 'timing/model': 1.1335381680401042, 'learning_rate/qwen_vl_interface': 6.9247315001874885e-06, 'learning_rate/action_model': 6.924731500187488e-05, 'epoch': 0.82}) +[2026-06-15 22:54:11,878][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.12009750306606293, 'train/grad_norm_pre_clip': 7.280306816101074, 'timing/data': 0.023126514977775514, 'timing/model': 1.1333875909913331, 'learning_rate/qwen_vl_interface': 6.9218813009105605e-06, 'learning_rate/action_model': 6.92188130091056e-05, 'epoch': 0.82}) +[2026-06-15 22:54:16,312][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.19795018434524536, 'train/grad_norm_pre_clip': 10.720987319946289, 'timing/data': 0.023487734026275575, 'timing/model': 1.133065125963185, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.82}) +[2026-06-15 22:54:20,791][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.24100732803344727, 'train/grad_norm_pre_clip': 11.967418670654297, 'timing/data': 0.023418333032168448, 'timing/model': 1.1372909510391764, 'learning_rate/qwen_vl_interface': 6.916178841786839e-06, 'learning_rate/action_model': 6.916178841786838e-05, 'epoch': 0.82}) +[2026-06-15 22:54:25,328][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.008533459156751633, 'train/grad_norm_pre_clip': 1.487785816192627, 'timing/data': 0.022909399995114654, 'timing/model': 1.2785940920002759, 'learning_rate/qwen_vl_interface': 6.913326584284115e-06, 'learning_rate/action_model': 6.913326584284115e-05, 'epoch': 0.82}) +[2026-06-15 22:54:29,735][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.0855662003159523, 'train/grad_norm_pre_clip': 5.500477313995361, 'timing/data': 0.023364883032627404, 'timing/model': 1.1389589530299418, 'learning_rate/qwen_vl_interface': 6.910473643050662e-06, 'learning_rate/action_model': 6.910473643050662e-05, 'epoch': 0.82}) +[2026-06-15 22:54:34,286][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.19438882172107697, 'train/grad_norm_pre_clip': 8.117071151733398, 'timing/data': 0.023230477003380656, 'timing/model': 1.1356553630321287, 'learning_rate/qwen_vl_interface': 6.907620019259215e-06, 'learning_rate/action_model': 6.907620019259216e-05, 'epoch': 0.82}) +[2026-06-15 22:54:38,841][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.15904206037521362, 'train/grad_norm_pre_clip': 13.298666000366211, 'timing/data': 0.024002896971069276, 'timing/model': 1.1381040450069122, 'learning_rate/qwen_vl_interface': 6.9047657140827935e-06, 'learning_rate/action_model': 6.904765714082794e-05, 'epoch': 0.82}) +[2026-06-15 22:54:43,445][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.007407620549201965, 'train/grad_norm_pre_clip': 0.769222617149353, 'timing/data': 0.028801398002542555, 'timing/model': 1.1503400760120712, 'learning_rate/qwen_vl_interface': 6.901910728694694e-06, 'learning_rate/action_model': 6.901910728694694e-05, 'epoch': 0.82}) +[2026-06-15 22:54:48,012][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.02771846391260624, 'train/grad_norm_pre_clip': 10.868387222290039, 'timing/data': 0.023446374048944563, 'timing/model': 1.3085864130407572, 'learning_rate/qwen_vl_interface': 6.899055064268489e-06, 'learning_rate/action_model': 6.899055064268489e-05, 'epoch': 0.82}) +[2026-06-15 22:54:52,425][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.09695709496736526, 'train/grad_norm_pre_clip': 7.5579047203063965, 'timing/data': 0.02322899899445474, 'timing/model': 1.1390844279667363, 'learning_rate/qwen_vl_interface': 6.896198721978038e-06, 'learning_rate/action_model': 6.896198721978038e-05, 'epoch': 0.82}) +[2026-06-15 22:54:56,968][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.007776207756251097, 'train/grad_norm_pre_clip': 1.371918797492981, 'timing/data': 0.023342171974945813, 'timing/model': 1.1368266819627024, 'learning_rate/qwen_vl_interface': 6.893341702997475e-06, 'learning_rate/action_model': 6.893341702997475e-05, 'epoch': 0.82}) +[2026-06-15 22:59:18,848][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.10822239518165588, 'train/grad_norm_pre_clip': 4.7788004875183105, 'eval/loss': 0.04989256709814072, 'eval/flap_precision': 0.9587155963302753, 'eval/flap_recall': 0.9675925925925926, 'eval/flap_f1': 0.9631336405529954, 'eval/flap_rate_model': 0.18166666666666667, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9631336405529954, 'timing/data': 0.02289946604287252, 'timing/model': 1.1381818739464507, 'learning_rate/qwen_vl_interface': 6.890484008501211e-06, 'learning_rate/action_model': 6.890484008501211e-05, 'epoch': 0.82}) +[2026-06-15 22:59:18,848][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2000_state +[2026-06-15 22:59:18,851][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 22:59:22,433][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2000_state/model.safetensors +[2026-06-15 22:59:32,436][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2000_state/optimizer.bin +[2026-06-15 22:59:32,436][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2000_state/sampler.bin +[2026-06-15 22:59:32,436][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-15 22:59:32,437][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-15 22:59:32,484][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-15 22:59:32,484][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 22:59:32,490][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 22:59:36,950][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.0064881532453000546, 'train/grad_norm_pre_clip': 1.0581504106521606, 'timing/data': 0.02590095199411735, 'timing/model': 1.1166808269917965, 'learning_rate/qwen_vl_interface': 6.887625639663941e-06, 'learning_rate/action_model': 6.88762563966394e-05, 'epoch': 0.82}) +[2026-06-15 22:59:41,404][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.13221116364002228, 'train/grad_norm_pre_clip': 17.950546264648438, 'timing/data': 0.023464337980840355, 'timing/model': 1.1137720359838568, 'learning_rate/qwen_vl_interface': 6.884766597660626e-06, 'learning_rate/action_model': 6.884766597660626e-05, 'epoch': 0.82}) +[2026-06-15 22:59:45,828][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.08551833778619766, 'train/grad_norm_pre_clip': 9.345128059387207, 'timing/data': 0.02380989002995193, 'timing/model': 1.1132938520167954, 'learning_rate/qwen_vl_interface': 6.881906883666515e-06, 'learning_rate/action_model': 6.881906883666515e-05, 'epoch': 0.82}) +[2026-06-15 22:59:50,230][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.08343902230262756, 'train/grad_norm_pre_clip': 8.74957275390625, 'timing/data': 0.023861549969296902, 'timing/model': 1.1139215139555745, 'learning_rate/qwen_vl_interface': 6.879046498857125e-06, 'learning_rate/action_model': 6.879046498857124e-05, 'epoch': 0.82}) +[2026-06-15 22:59:54,676][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.010449329391121864, 'train/grad_norm_pre_clip': 3.0887129306793213, 'timing/data': 0.02364555699750781, 'timing/model': 1.2425215420080349, 'learning_rate/qwen_vl_interface': 6.876185444408255e-06, 'learning_rate/action_model': 6.876185444408255e-05, 'epoch': 0.82}) +[2026-06-15 22:59:58,997][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.006558428052812815, 'train/grad_norm_pre_clip': 0.7969397306442261, 'timing/data': 0.023181261029094458, 'timing/model': 1.1180367189808749, 'learning_rate/qwen_vl_interface': 6.873323721495976e-06, 'learning_rate/action_model': 6.873323721495976e-05, 'epoch': 0.82}) +[2026-06-15 23:00:03,485][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.004735477734357119, 'train/grad_norm_pre_clip': 0.4612264633178711, 'timing/data': 0.023591827019117773, 'timing/model': 1.1217382539762184, 'learning_rate/qwen_vl_interface': 6.870461331296632e-06, 'learning_rate/action_model': 6.870461331296632e-05, 'epoch': 0.82}) +[2026-06-15 23:00:07,974][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.08574748039245605, 'train/grad_norm_pre_clip': 13.129159927368164, 'timing/data': 0.02339342300547287, 'timing/model': 1.1213631719583645, 'learning_rate/qwen_vl_interface': 6.8675982749868455e-06, 'learning_rate/action_model': 6.867598274986845e-05, 'epoch': 0.82}) +[2026-06-15 23:00:12,488][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.05370749533176422, 'train/grad_norm_pre_clip': 7.839311122894287, 'timing/data': 0.023907767958007753, 'timing/model': 1.1253727419534698, 'learning_rate/qwen_vl_interface': 6.864734553743509e-06, 'learning_rate/action_model': 6.86473455374351e-05, 'epoch': 0.83}) +[2026-06-15 23:00:16,991][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.07352574914693832, 'train/grad_norm_pre_clip': 9.271785736083984, 'timing/data': 0.023778124013915658, 'timing/model': 1.2786744610057212, 'learning_rate/qwen_vl_interface': 6.861870168743793e-06, 'learning_rate/action_model': 6.861870168743793e-05, 'epoch': 0.83}) +[2026-06-15 23:00:21,448][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.18563953042030334, 'train/grad_norm_pre_clip': 6.842244625091553, 'timing/data': 0.02375346899498254, 'timing/model': 1.239214107976295, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 0.83}) +[2026-06-15 23:00:25,814][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.30091753602027893, 'train/grad_norm_pre_clip': 18.62434196472168, 'timing/data': 0.02416091598570347, 'timing/model': 1.1366118509904481, 'learning_rate/qwen_vl_interface': 6.85613941218525e-06, 'learning_rate/action_model': 6.85613941218525e-05, 'epoch': 0.83}) +[2026-06-15 23:00:30,351][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.09362050145864487, 'train/grad_norm_pre_clip': 16.89708137512207, 'timing/data': 0.022789434005971998, 'timing/model': 1.139155528973788, 'learning_rate/qwen_vl_interface': 6.8532730429821215e-06, 'learning_rate/action_model': 6.853273042982122e-05, 'epoch': 0.83}) +[2026-06-15 23:00:34,858][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.4353044331073761, 'train/grad_norm_pre_clip': 21.77203941345215, 'timing/data': 0.023018614971078932, 'timing/model': 1.1284476770088077, 'learning_rate/qwen_vl_interface': 6.850406014734005e-06, 'learning_rate/action_model': 6.850406014734005e-05, 'epoch': 0.83}) +[2026-06-15 23:00:39,368][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.06704693287611008, 'train/grad_norm_pre_clip': 10.10009479522705, 'timing/data': 0.023515373992267996, 'timing/model': 1.127273691992741, 'learning_rate/qwen_vl_interface': 6.8475383286194295e-06, 'learning_rate/action_model': 6.84753832861943e-05, 'epoch': 0.83}) +[2026-06-15 23:00:43,919][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.004435775801539421, 'train/grad_norm_pre_clip': 0.660686194896698, 'timing/data': 0.023228010977618396, 'timing/model': 1.30475672299508, 'learning_rate/qwen_vl_interface': 6.844669985817192e-06, 'learning_rate/action_model': 6.844669985817192e-05, 'epoch': 0.83}) +[2026-06-15 23:00:48,371][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.12121398746967316, 'train/grad_norm_pre_clip': 16.12777328491211, 'timing/data': 0.02369715500390157, 'timing/model': 1.217354846012313, 'learning_rate/qwen_vl_interface': 6.8418009875063594e-06, 'learning_rate/action_model': 6.84180098750636e-05, 'epoch': 0.83}) +[2026-06-15 23:00:52,738][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.2968963086605072, 'train/grad_norm_pre_clip': 18.005775451660156, 'timing/data': 0.023721434001345187, 'timing/model': 1.1302592830033973, 'learning_rate/qwen_vl_interface': 6.838931334866267e-06, 'learning_rate/action_model': 6.838931334866266e-05, 'epoch': 0.83}) +[2026-06-15 23:00:57,233][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.03199600428342819, 'train/grad_norm_pre_clip': 10.49902629852295, 'timing/data': 0.023606361995916814, 'timing/model': 1.1300424700020812, 'learning_rate/qwen_vl_interface': 6.836061029076523e-06, 'learning_rate/action_model': 6.836061029076522e-05, 'epoch': 0.83}) +[2026-06-15 23:01:01,765][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.2756853401660919, 'train/grad_norm_pre_clip': 11.458642959594727, 'timing/data': 0.02398923598229885, 'timing/model': 1.1357521619647741, 'learning_rate/qwen_vl_interface': 6.833190071317e-06, 'learning_rate/action_model': 6.833190071316999e-05, 'epoch': 0.83}) +[2026-06-15 23:01:06,312][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.11961670964956284, 'train/grad_norm_pre_clip': 34.77229690551758, 'timing/data': 0.022693692008033395, 'timing/model': 1.1356311139534228, 'learning_rate/qwen_vl_interface': 6.830318462767841e-06, 'learning_rate/action_model': 6.830318462767841e-05, 'epoch': 0.83}) +[2026-06-15 23:01:10,869][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.11068272590637207, 'train/grad_norm_pre_clip': 12.513012886047363, 'timing/data': 0.0263485120376572, 'timing/model': 1.3079537660232745, 'learning_rate/qwen_vl_interface': 6.827446204609456e-06, 'learning_rate/action_model': 6.827446204609456e-05, 'epoch': 0.83}) +[2026-06-15 23:01:15,257][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.11912229657173157, 'train/grad_norm_pre_clip': 13.489592552185059, 'timing/data': 0.023319496016483754, 'timing/model': 1.1362377970363013, 'learning_rate/qwen_vl_interface': 6.824573298022522e-06, 'learning_rate/action_model': 6.824573298022522e-05, 'epoch': 0.83}) +[2026-06-15 23:01:19,789][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.08728884160518646, 'train/grad_norm_pre_clip': 24.076335906982422, 'timing/data': 0.023281260044313967, 'timing/model': 1.1389499510405585, 'learning_rate/qwen_vl_interface': 6.8216997441879805e-06, 'learning_rate/action_model': 6.821699744187981e-05, 'epoch': 0.83}) +[2026-06-15 23:01:24,317][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.230828195810318, 'train/grad_norm_pre_clip': 15.654632568359375, 'timing/data': 0.022949867008719593, 'timing/model': 1.1305349600152113, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 0.83}) +[2026-06-15 23:01:28,863][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.09387417882680893, 'train/grad_norm_pre_clip': 11.260904312133789, 'timing/data': 0.02312985301250592, 'timing/model': 1.1324227919685654, 'learning_rate/qwen_vl_interface': 6.815950699501186e-06, 'learning_rate/action_model': 6.815950699501186e-05, 'epoch': 0.83}) +[2026-06-15 23:01:33,417][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.061563022434711456, 'train/grad_norm_pre_clip': 7.239164352416992, 'timing/data': 0.023258948989678174, 'timing/model': 1.132942745985929, 'learning_rate/qwen_vl_interface': 6.813075211012148e-06, 'learning_rate/action_model': 6.813075211012147e-05, 'epoch': 0.83}) +[2026-06-15 23:01:37,950][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.02549915574491024, 'train/grad_norm_pre_clip': 1.8357677459716797, 'timing/data': 0.022952868952415884, 'timing/model': 1.2918882869998924, 'learning_rate/qwen_vl_interface': 6.81019908000193e-06, 'learning_rate/action_model': 6.81019908000193e-05, 'epoch': 0.83}) +[2026-06-15 23:01:42,383][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.03972235321998596, 'train/grad_norm_pre_clip': 4.660727024078369, 'timing/data': 0.023667752975597978, 'timing/model': 1.1885373019613326, 'learning_rate/qwen_vl_interface': 6.807322307652806e-06, 'learning_rate/action_model': 6.807322307652806e-05, 'epoch': 0.83}) +[2026-06-15 23:01:46,754][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.10991410166025162, 'train/grad_norm_pre_clip': 8.252656936645508, 'timing/data': 0.023158927040640265, 'timing/model': 1.13558716699481, 'learning_rate/qwen_vl_interface': 6.804444895147306e-06, 'learning_rate/action_model': 6.804444895147306e-05, 'epoch': 0.83}) +[2026-06-15 23:01:51,245][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.10386630892753601, 'train/grad_norm_pre_clip': 7.145829200744629, 'timing/data': 0.023271865968126804, 'timing/model': 1.1373375130351633, 'learning_rate/qwen_vl_interface': 6.8015668436682255e-06, 'learning_rate/action_model': 6.801566843668226e-05, 'epoch': 0.83}) +[2026-06-15 23:01:55,778][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.14903585612773895, 'train/grad_norm_pre_clip': 10.016317367553711, 'timing/data': 0.023511273029726, 'timing/model': 1.132217358041089, 'learning_rate/qwen_vl_interface': 6.7986881543986245e-06, 'learning_rate/action_model': 6.798688154398624e-05, 'epoch': 0.83}) +[2026-06-15 23:02:00,327][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.2203141152858734, 'train/grad_norm_pre_clip': 6.9418044090271, 'timing/data': 0.024061829026322812, 'timing/model': 1.1349631529883482, 'learning_rate/qwen_vl_interface': 6.795808828521822e-06, 'learning_rate/action_model': 6.795808828521821e-05, 'epoch': 0.84}) +[2026-06-15 23:02:04,875][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.08385487645864487, 'train/grad_norm_pre_clip': 22.197721481323242, 'timing/data': 0.023151939967647195, 'timing/model': 1.3078135669929907, 'learning_rate/qwen_vl_interface': 6.792928867221401e-06, 'learning_rate/action_model': 6.792928867221401e-05, 'epoch': 0.84}) +[2026-06-15 23:02:09,370][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.08598439395427704, 'train/grad_norm_pre_clip': 12.541600227355957, 'timing/data': 0.02377878996776417, 'timing/model': 1.250202739960514, 'learning_rate/qwen_vl_interface': 6.790048271681202e-06, 'learning_rate/action_model': 6.790048271681202e-05, 'epoch': 0.84}) +[2026-06-15 23:02:13,740][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.1390012949705124, 'train/grad_norm_pre_clip': 12.378506660461426, 'timing/data': 0.02359233098104596, 'timing/model': 1.1295919630210847, 'learning_rate/qwen_vl_interface': 6.7871670430853335e-06, 'learning_rate/action_model': 6.787167043085333e-05, 'epoch': 0.84}) +[2026-06-15 23:02:18,295][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.01761237531900406, 'train/grad_norm_pre_clip': 1.9413944482803345, 'timing/data': 0.023569712007883936, 'timing/model': 1.1387910530320369, 'learning_rate/qwen_vl_interface': 6.784285182618156e-06, 'learning_rate/action_model': 6.784285182618156e-05, 'epoch': 0.84}) +[2026-06-15 23:02:22,864][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.05777081847190857, 'train/grad_norm_pre_clip': 5.2151408195495605, 'timing/data': 0.023419531993567944, 'timing/model': 1.141665083996486, 'learning_rate/qwen_vl_interface': 6.781402691464297e-06, 'learning_rate/action_model': 6.781402691464297e-05, 'epoch': 0.84}) +[2026-06-15 23:02:27,400][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.03336505964398384, 'train/grad_norm_pre_clip': 4.6642351150512695, 'timing/data': 0.02313829498598352, 'timing/model': 1.1328180360142142, 'learning_rate/qwen_vl_interface': 6.778519570808636e-06, 'learning_rate/action_model': 6.778519570808636e-05, 'epoch': 0.84}) +[2026-06-15 23:02:31,922][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.1057891696691513, 'train/grad_norm_pre_clip': 7.145190715789795, 'timing/data': 0.022699352994095534, 'timing/model': 1.2837397880502976, 'learning_rate/qwen_vl_interface': 6.775635821836317e-06, 'learning_rate/action_model': 6.775635821836317e-05, 'epoch': 0.84}) +[2026-06-15 23:02:36,453][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.06769682466983795, 'train/grad_norm_pre_clip': 9.723522186279297, 'timing/data': 0.023223423049785197, 'timing/model': 1.27761140599614, 'learning_rate/qwen_vl_interface': 6.77275144573274e-06, 'learning_rate/action_model': 6.77275144573274e-05, 'epoch': 0.84}) +[2026-06-15 23:02:40,828][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.03875396028161049, 'train/grad_norm_pre_clip': 3.0627896785736084, 'timing/data': 0.023685595020651817, 'timing/model': 1.1316782050416805, 'learning_rate/qwen_vl_interface': 6.769866443683564e-06, 'learning_rate/action_model': 6.769866443683563e-05, 'epoch': 0.84}) +[2026-06-15 23:02:45,368][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.21693295240402222, 'train/grad_norm_pre_clip': 9.290322303771973, 'timing/data': 0.02268513001035899, 'timing/model': 1.1310868830187246, 'learning_rate/qwen_vl_interface': 6.766980816874701e-06, 'learning_rate/action_model': 6.7669808168747e-05, 'epoch': 0.84}) +[2026-06-15 23:02:49,915][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.3597463071346283, 'train/grad_norm_pre_clip': 12.710962295532227, 'timing/data': 0.023208903963677585, 'timing/model': 1.139007463993039, 'learning_rate/qwen_vl_interface': 6.764094566492327e-06, 'learning_rate/action_model': 6.764094566492328e-05, 'epoch': 0.84}) +[2026-06-15 23:02:54,455][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.30470243096351624, 'train/grad_norm_pre_clip': 12.907368659973145, 'timing/data': 0.023602328961715102, 'timing/model': 1.1369486679905094, 'learning_rate/qwen_vl_interface': 6.7612076937228685e-06, 'learning_rate/action_model': 6.761207693722869e-05, 'epoch': 0.84}) +[2026-06-15 23:02:58,951][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.1396467238664627, 'train/grad_norm_pre_clip': 10.600903511047363, 'timing/data': 0.024471658980473876, 'timing/model': 1.1324669549940154, 'learning_rate/qwen_vl_interface': 6.7583201997530126e-06, 'learning_rate/action_model': 6.758320199753012e-05, 'epoch': 0.84}) +[2026-06-15 23:03:03,490][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.03991193696856499, 'train/grad_norm_pre_clip': 3.75179386138916, 'timing/data': 0.023404615989420563, 'timing/model': 1.2942335460102186, 'learning_rate/qwen_vl_interface': 6.755432085769693e-06, 'learning_rate/action_model': 6.755432085769694e-05, 'epoch': 0.84}) +[2026-06-15 23:03:07,872][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.07862383127212524, 'train/grad_norm_pre_clip': 4.883868217468262, 'timing/data': 0.023692814982496202, 'timing/model': 1.1389452260336839, 'learning_rate/qwen_vl_interface': 6.75254335296011e-06, 'learning_rate/action_model': 6.75254335296011e-05, 'epoch': 0.84}) +[2026-06-15 23:03:12,418][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.12576229870319366, 'train/grad_norm_pre_clip': 17.55558967590332, 'timing/data': 0.025046416965778917, 'timing/model': 1.135517962044105, 'learning_rate/qwen_vl_interface': 6.749654002511712e-06, 'learning_rate/action_model': 6.749654002511712e-05, 'epoch': 0.84}) +[2026-06-15 23:03:16,852][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.14791011810302734, 'train/grad_norm_pre_clip': 16.179256439208984, 'timing/data': 0.02375819004373625, 'timing/model': 1.130954112042673, 'learning_rate/qwen_vl_interface': 6.7467640356122e-06, 'learning_rate/action_model': 6.746764035612199e-05, 'epoch': 0.84}) +[2026-06-15 23:03:21,294][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.31553855538368225, 'train/grad_norm_pre_clip': 11.502171516418457, 'timing/data': 0.02398408396402374, 'timing/model': 1.1376602469827048, 'learning_rate/qwen_vl_interface': 6.74387345344953e-06, 'learning_rate/action_model': 6.74387345344953e-05, 'epoch': 0.84}) +[2026-06-15 23:03:25,762][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.26862677931785583, 'train/grad_norm_pre_clip': 18.05972671508789, 'timing/data': 0.023754280991852283, 'timing/model': 1.1312895369483158, 'learning_rate/qwen_vl_interface': 6.740982257211911e-06, 'learning_rate/action_model': 6.740982257211911e-05, 'epoch': 0.84}) +[2026-06-15 23:03:30,270][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.15730547904968262, 'train/grad_norm_pre_clip': 13.81959342956543, 'timing/data': 0.024197368999011815, 'timing/model': 1.2535017660120502, 'learning_rate/qwen_vl_interface': 6.738090448087806e-06, 'learning_rate/action_model': 6.738090448087806e-05, 'epoch': 0.84}) +[2026-06-15 23:03:34,663][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.14987559616565704, 'train/grad_norm_pre_clip': 27.643848419189453, 'timing/data': 0.023678216035477817, 'timing/model': 1.138708727026824, 'learning_rate/qwen_vl_interface': 6.735198027265928e-06, 'learning_rate/action_model': 6.735198027265928e-05, 'epoch': 0.84}) +[2026-06-15 23:03:39,227][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.07976736128330231, 'train/grad_norm_pre_clip': 8.673829078674316, 'timing/data': 0.02356424101162702, 'timing/model': 1.1389009790145792, 'learning_rate/qwen_vl_interface': 6.7323049959352424e-06, 'learning_rate/action_model': 6.732304995935243e-05, 'epoch': 0.84}) +[2026-06-15 23:03:43,776][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.04971972852945328, 'train/grad_norm_pre_clip': 5.614648818969727, 'timing/data': 0.02337772201281041, 'timing/model': 1.1371327860397287, 'learning_rate/qwen_vl_interface': 6.729411355284962e-06, 'learning_rate/action_model': 6.729411355284963e-05, 'epoch': 0.84}) +[2026-06-15 23:03:48,311][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.07434578239917755, 'train/grad_norm_pre_clip': 3.899991989135742, 'timing/data': 0.023845051997341216, 'timing/model': 1.1330110490089282, 'learning_rate/qwen_vl_interface': 6.726517106504559e-06, 'learning_rate/action_model': 6.726517106504558e-05, 'epoch': 0.85}) +[2026-06-15 23:03:52,862][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.054759953171014786, 'train/grad_norm_pre_clip': 6.304679870605469, 'timing/data': 0.023913414974231273, 'timing/model': 1.2977191910031252, 'learning_rate/qwen_vl_interface': 6.723622250783745e-06, 'learning_rate/action_model': 6.723622250783744e-05, 'epoch': 0.85}) +[2026-06-15 23:03:57,249][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.02255682833492756, 'train/grad_norm_pre_clip': 2.0311219692230225, 'timing/data': 0.023981117003131658, 'timing/model': 1.1363492290256545, 'learning_rate/qwen_vl_interface': 6.720726789312487e-06, 'learning_rate/action_model': 6.720726789312487e-05, 'epoch': 0.85}) +[2026-06-15 23:04:01,789][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.11957215517759323, 'train/grad_norm_pre_clip': 8.605237007141113, 'timing/data': 0.023480887000914663, 'timing/model': 1.1342010429943912, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.85}) +[2026-06-15 23:04:06,287][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.22529850900173187, 'train/grad_norm_pre_clip': 12.346148490905762, 'timing/data': 0.023372280003968626, 'timing/model': 1.1398946440313011, 'learning_rate/qwen_vl_interface': 6.714934053879749e-06, 'learning_rate/action_model': 6.71493405387975e-05, 'epoch': 0.85}) +[2026-06-15 23:04:10,828][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.16977021098136902, 'train/grad_norm_pre_clip': 14.835253715515137, 'timing/data': 0.023558506043627858, 'timing/model': 1.1346287719788961, 'learning_rate/qwen_vl_interface': 6.712036782299442e-06, 'learning_rate/action_model': 6.712036782299443e-05, 'epoch': 0.85}) +[2026-06-15 23:04:15,376][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.06163082271814346, 'train/grad_norm_pre_clip': 6.242151260375977, 'timing/data': 0.02335185499396175, 'timing/model': 1.1407449859543703, 'learning_rate/qwen_vl_interface': 6.709138909731041e-06, 'learning_rate/action_model': 6.709138909731041e-05, 'epoch': 0.85}) +[2026-06-15 23:04:19,927][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.08222179114818573, 'train/grad_norm_pre_clip': 12.567941665649414, 'timing/data': 0.02352919796248898, 'timing/model': 1.2998016609926708, 'learning_rate/qwen_vl_interface': 6.706240437365751e-06, 'learning_rate/action_model': 6.70624043736575e-05, 'epoch': 0.85}) +[2026-06-15 23:04:24,445][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.033988215029239655, 'train/grad_norm_pre_clip': 6.44489049911499, 'timing/data': 0.023345601046457887, 'timing/model': 1.2527321829693392, 'learning_rate/qwen_vl_interface': 6.703341366395023e-06, 'learning_rate/action_model': 6.703341366395023e-05, 'epoch': 0.85}) +[2026-06-15 23:04:28,846][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.1245010644197464, 'train/grad_norm_pre_clip': 16.174654006958008, 'timing/data': 0.0230908389785327, 'timing/model': 1.1346547059947625, 'learning_rate/qwen_vl_interface': 6.700441698010558e-06, 'learning_rate/action_model': 6.700441698010558e-05, 'epoch': 0.85}) +[2026-06-15 23:04:33,387][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.03874863311648369, 'train/grad_norm_pre_clip': 8.385125160217285, 'timing/data': 0.023606522008776665, 'timing/model': 1.1346689150086604, 'learning_rate/qwen_vl_interface': 6.697541433404299e-06, 'learning_rate/action_model': 6.697541433404299e-05, 'epoch': 0.85}) +[2026-06-15 23:04:37,924][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.12177271395921707, 'train/grad_norm_pre_clip': 10.018051147460938, 'timing/data': 0.02391935500781983, 'timing/model': 1.1343037849874236, 'learning_rate/qwen_vl_interface': 6.6946405737684325e-06, 'learning_rate/action_model': 6.694640573768433e-05, 'epoch': 0.85}) +[2026-06-15 23:04:42,464][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.15927009284496307, 'train/grad_norm_pre_clip': 11.439549446105957, 'timing/data': 0.02283043903298676, 'timing/model': 1.141004856035579, 'learning_rate/qwen_vl_interface': 6.691739120295397e-06, 'learning_rate/action_model': 6.691739120295397e-05, 'epoch': 0.85}) +[2026-06-15 23:04:47,011][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.05188624560832977, 'train/grad_norm_pre_clip': 6.029556751251221, 'timing/data': 0.022964171948842704, 'timing/model': 1.2887192639755085, 'learning_rate/qwen_vl_interface': 6.688837074177868e-06, 'learning_rate/action_model': 6.688837074177868e-05, 'epoch': 0.85}) +[2026-06-15 23:04:51,564][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.0295211523771286, 'train/grad_norm_pre_clip': 14.149927139282227, 'timing/data': 0.023687429958954453, 'timing/model': 1.284158673021011, 'learning_rate/qwen_vl_interface': 6.68593443660877e-06, 'learning_rate/action_model': 6.685934436608769e-05, 'epoch': 0.85}) +[2026-06-15 23:04:55,957][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.0076098572462797165, 'train/grad_norm_pre_clip': 0.7919066548347473, 'timing/data': 0.02311890199780464, 'timing/model': 1.1442101409775205, 'learning_rate/qwen_vl_interface': 6.6830312087812625e-06, 'learning_rate/action_model': 6.683031208781262e-05, 'epoch': 0.85}) +[2026-06-15 23:05:00,514][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.019861232489347458, 'train/grad_norm_pre_clip': 2.7086706161499023, 'timing/data': 0.023098379955627024, 'timing/model': 1.1331637079711072, 'learning_rate/qwen_vl_interface': 6.680127391888754e-06, 'learning_rate/action_model': 6.680127391888754e-05, 'epoch': 0.85}) +[2026-06-15 23:05:05,065][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.12289867550134659, 'train/grad_norm_pre_clip': 18.14131736755371, 'timing/data': 0.024133680039085448, 'timing/model': 1.1398471930297092, 'learning_rate/qwen_vl_interface': 6.677222987124897e-06, 'learning_rate/action_model': 6.677222987124897e-05, 'epoch': 0.85}) +[2026-06-15 23:05:09,610][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.024836456403136253, 'train/grad_norm_pre_clip': 4.898569583892822, 'timing/data': 0.022961754992138594, 'timing/model': 1.1358824000344612, 'learning_rate/qwen_vl_interface': 6.67431799568358e-06, 'learning_rate/action_model': 6.67431799568358e-05, 'epoch': 0.85}) +[2026-06-15 23:05:14,111][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.19046534597873688, 'train/grad_norm_pre_clip': 10.879403114318848, 'timing/data': 0.02373659802833572, 'timing/model': 1.1376698430394754, 'learning_rate/qwen_vl_interface': 6.671412418758939e-06, 'learning_rate/action_model': 6.671412418758938e-05, 'epoch': 0.85}) +[2026-06-15 23:05:18,655][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.2774371802806854, 'train/grad_norm_pre_clip': 21.715099334716797, 'timing/data': 0.023697776021435857, 'timing/model': 1.2908313299994916, 'learning_rate/qwen_vl_interface': 6.668506257545341e-06, 'learning_rate/action_model': 6.668506257545341e-05, 'epoch': 0.85}) +[2026-06-15 23:05:23,033][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.25335532426834106, 'train/grad_norm_pre_clip': 16.301822662353516, 'timing/data': 0.02461403503548354, 'timing/model': 1.1333994530141354, 'learning_rate/qwen_vl_interface': 6.665599513237403e-06, 'learning_rate/action_model': 6.665599513237402e-05, 'epoch': 0.85}) +[2026-06-15 23:05:27,576][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.0772421583533287, 'train/grad_norm_pre_clip': 9.630831718444824, 'timing/data': 0.023283357033506036, 'timing/model': 1.1406926049967296, 'learning_rate/qwen_vl_interface': 6.662692187029978e-06, 'learning_rate/action_model': 6.662692187029978e-05, 'epoch': 0.85}) +[2026-06-15 23:05:32,019][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.12486429512500763, 'train/grad_norm_pre_clip': 12.266854286193848, 'timing/data': 0.023563958995509893, 'timing/model': 1.135860284033697, 'learning_rate/qwen_vl_interface': 6.659784280118156e-06, 'learning_rate/action_model': 6.659784280118156e-05, 'epoch': 0.85}) +[2026-06-15 23:05:36,467][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.007414147723466158, 'train/grad_norm_pre_clip': 1.0173397064208984, 'timing/data': 0.023675389005802572, 'timing/model': 1.1357300089439377, 'learning_rate/qwen_vl_interface': 6.65687579369727e-06, 'learning_rate/action_model': 6.656875793697271e-05, 'epoch': 0.85}) +[2026-06-15 23:05:40,934][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.14013570547103882, 'train/grad_norm_pre_clip': 6.5406646728515625, 'timing/data': 0.024496043974068016, 'timing/model': 1.130296618968714, 'learning_rate/qwen_vl_interface': 6.653966728962885e-06, 'learning_rate/action_model': 6.653966728962885e-05, 'epoch': 0.86}) +[2026-06-15 23:05:45,440][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.326387882232666, 'train/grad_norm_pre_clip': 13.035682678222656, 'timing/data': 0.024353731016162783, 'timing/model': 1.2584798349998891, 'learning_rate/qwen_vl_interface': 6.651057087110813e-06, 'learning_rate/action_model': 6.651057087110813e-05, 'epoch': 0.86}) +[2026-06-15 23:05:49,834][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.145835742354393, 'train/grad_norm_pre_clip': 11.254887580871582, 'timing/data': 0.02365407400066033, 'timing/model': 1.1336987109971233, 'learning_rate/qwen_vl_interface': 6.6481468693370924e-06, 'learning_rate/action_model': 6.648146869337091e-05, 'epoch': 0.86}) +[2026-06-15 23:05:54,382][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.009061465971171856, 'train/grad_norm_pre_clip': 1.2831792831420898, 'timing/data': 0.023456826980691403, 'timing/model': 1.1436093730153516, 'learning_rate/qwen_vl_interface': 6.645236076838009e-06, 'learning_rate/action_model': 6.645236076838009e-05, 'epoch': 0.86}) +[2026-06-15 23:05:58,939][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.03503023833036423, 'train/grad_norm_pre_clip': 10.586893081665039, 'timing/data': 0.02343901601852849, 'timing/model': 1.1371360140037723, 'learning_rate/qwen_vl_interface': 6.642324710810075e-06, 'learning_rate/action_model': 6.642324710810076e-05, 'epoch': 0.86}) +[2026-06-15 23:06:03,505][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.036369144916534424, 'train/grad_norm_pre_clip': 10.605212211608887, 'timing/data': 0.023335859004873782, 'timing/model': 1.134842285013292, 'learning_rate/qwen_vl_interface': 6.639412772450047e-06, 'learning_rate/action_model': 6.639412772450047e-05, 'epoch': 0.86}) +[2026-06-15 23:06:08,049][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.011821718886494637, 'train/grad_norm_pre_clip': 2.8175411224365234, 'timing/data': 0.023100753023754805, 'timing/model': 1.292451395012904, 'learning_rate/qwen_vl_interface': 6.6365002629549104e-06, 'learning_rate/action_model': 6.63650026295491e-05, 'epoch': 0.86}) +[2026-06-15 23:06:12,430][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.1061214730143547, 'train/grad_norm_pre_clip': 24.97732162475586, 'timing/data': 0.02395481598796323, 'timing/model': 1.1301292069838382, 'learning_rate/qwen_vl_interface': 6.633587183521887e-06, 'learning_rate/action_model': 6.633587183521887e-05, 'epoch': 0.86}) +[2026-06-15 23:06:16,978][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.151293084025383, 'train/grad_norm_pre_clip': 5.456202983856201, 'timing/data': 0.02340640500187874, 'timing/model': 1.135177846008446, 'learning_rate/qwen_vl_interface': 6.6306735353484365e-06, 'learning_rate/action_model': 6.630673535348436e-05, 'epoch': 0.86}) +[2026-06-15 23:06:21,471][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.1596354842185974, 'train/grad_norm_pre_clip': 11.900222778320312, 'timing/data': 0.023415105999447405, 'timing/model': 1.1345884010079317, 'learning_rate/qwen_vl_interface': 6.627759319632246e-06, 'learning_rate/action_model': 6.627759319632247e-05, 'epoch': 0.86}) +[2026-06-15 23:06:26,005][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.09414408355951309, 'train/grad_norm_pre_clip': 6.569515705108643, 'timing/data': 0.023292934987694025, 'timing/model': 1.1305778659880161, 'learning_rate/qwen_vl_interface': 6.624844537571245e-06, 'learning_rate/action_model': 6.624844537571245e-05, 'epoch': 0.86}) +[2026-06-15 23:06:30,545][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.003242780454456806, 'train/grad_norm_pre_clip': 0.4108014702796936, 'timing/data': 0.023035351012367755, 'timing/model': 1.1361402990296483, 'learning_rate/qwen_vl_interface': 6.621929190363585e-06, 'learning_rate/action_model': 6.621929190363585e-05, 'epoch': 0.86}) +[2026-06-15 23:06:35,095][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.010399151593446732, 'train/grad_norm_pre_clip': 3.302126884460449, 'timing/data': 0.02973050798755139, 'timing/model': 1.3011150459642522, 'learning_rate/qwen_vl_interface': 6.619013279207657e-06, 'learning_rate/action_model': 6.619013279207656e-05, 'epoch': 0.86}) +[2026-06-15 23:06:39,608][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.027308551594614983, 'train/grad_norm_pre_clip': 4.454722881317139, 'timing/data': 0.022567386040464044, 'timing/model': 1.256353773002047, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 0.86}) +[2026-06-15 23:06:44,001][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.21691788733005524, 'train/grad_norm_pre_clip': 5.1880106925964355, 'timing/data': 0.023576218984089792, 'timing/model': 1.1333337139803916, 'learning_rate/qwen_vl_interface': 6.6131797698457135e-06, 'learning_rate/action_model': 6.613179769845713e-05, 'epoch': 0.86}) +[2026-06-15 23:06:48,550][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.05550708249211311, 'train/grad_norm_pre_clip': 16.170780181884766, 'timing/data': 0.02342370100086555, 'timing/model': 1.1316732590203173, 'learning_rate/qwen_vl_interface': 6.610262174037633e-06, 'learning_rate/action_model': 6.610262174037633e-05, 'epoch': 0.86}) +[2026-06-15 23:06:53,179][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.18411456048488617, 'train/grad_norm_pre_clip': 10.512382507324219, 'timing/data': 0.04625398002099246, 'timing/model': 1.1445858420338482, 'learning_rate/qwen_vl_interface': 6.607344019077155e-06, 'learning_rate/action_model': 6.607344019077156e-05, 'epoch': 0.86}) +[2026-06-15 23:06:57,734][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.014591556042432785, 'train/grad_norm_pre_clip': 3.1428966522216797, 'timing/data': 0.0225023910170421, 'timing/model': 1.1363998699816875, 'learning_rate/qwen_vl_interface': 6.604425306163822e-06, 'learning_rate/action_model': 6.604425306163822e-05, 'epoch': 0.86}) +[2026-06-15 23:11:20,168][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.1680954396724701, 'train/grad_norm_pre_clip': 10.915291786193848, 'eval/loss': 0.06222466379404068, 'eval/flap_precision': 0.947289156626506, 'eval/flap_recall': 0.970679012345679, 'eval/flap_f1': 0.9588414634146342, 'eval/flap_rate_model': 0.18444444444444444, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9588414634146342, 'timing/data': 0.022833461000118405, 'timing/model': 1.2869835590245202, 'learning_rate/qwen_vl_interface': 6.6015060364974045e-06, 'learning_rate/action_model': 6.601506036497404e-05, 'epoch': 0.86}) +[2026-06-15 23:11:20,169][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2100_state +[2026-06-15 23:11:20,171][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:11:23,753][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2100_state/model.safetensors +[2026-06-15 23:11:33,819][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2100_state/optimizer.bin +[2026-06-15 23:11:33,820][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2100_state/sampler.bin +[2026-06-15 23:11:33,820][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2100_state/sampler_1.bin +[2026-06-15 23:11:33,821][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2100_state/random_states_0.pkl +[2026-06-15 23:11:33,870][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2100 +[2026-06-15 23:11:33,870][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:11:33,876][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:11:38,329][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.032447438687086105, 'train/grad_norm_pre_clip': 5.5211639404296875, 'timing/data': 0.024277114018332213, 'timing/model': 1.1152910679811612, 'learning_rate/qwen_vl_interface': 6.598586211277907e-06, 'learning_rate/action_model': 6.598586211277907e-05, 'epoch': 0.86}) +[2026-06-15 23:11:42,789][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.09941689670085907, 'train/grad_norm_pre_clip': 10.213549613952637, 'timing/data': 0.023550828045699745, 'timing/model': 1.2722748000524007, 'learning_rate/qwen_vl_interface': 6.59566583170556e-06, 'learning_rate/action_model': 6.59566583170556e-05, 'epoch': 0.86}) +[2026-06-15 23:11:47,094][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.08379700034856796, 'train/grad_norm_pre_clip': 16.091552734375, 'timing/data': 0.023099270998500288, 'timing/model': 1.116435302013997, 'learning_rate/qwen_vl_interface': 6.59274489898082e-06, 'learning_rate/action_model': 6.592744898980819e-05, 'epoch': 0.86}) +[2026-06-15 23:11:51,545][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.3250373303890228, 'train/grad_norm_pre_clip': 10.527178764343262, 'timing/data': 0.022986272000707686, 'timing/model': 1.1129242070019245, 'learning_rate/qwen_vl_interface': 6.5898234143043716e-06, 'learning_rate/action_model': 6.589823414304371e-05, 'epoch': 0.86}) +[2026-06-15 23:11:55,968][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.033003393560647964, 'train/grad_norm_pre_clip': 5.559201717376709, 'timing/data': 0.023532858001999557, 'timing/model': 1.1168389869853854, 'learning_rate/qwen_vl_interface': 6.5869013788771264e-06, 'learning_rate/action_model': 6.586901378877126e-05, 'epoch': 0.86}) +[2026-06-15 23:12:00,446][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.005100654438138008, 'train/grad_norm_pre_clip': 0.7190341353416443, 'timing/data': 0.02341960195917636, 'timing/model': 1.1250173730077222, 'learning_rate/qwen_vl_interface': 6.583978793900223e-06, 'learning_rate/action_model': 6.583978793900223e-05, 'epoch': 0.87}) +[2026-06-15 23:12:04,925][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.021184787154197693, 'train/grad_norm_pre_clip': 4.525963306427002, 'timing/data': 0.023543893999885768, 'timing/model': 1.1215702510089613, 'learning_rate/qwen_vl_interface': 6.581055660575028e-06, 'learning_rate/action_model': 6.581055660575027e-05, 'epoch': 0.87}) +[2026-06-15 23:12:09,438][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.01636991836130619, 'train/grad_norm_pre_clip': 3.831512689590454, 'timing/data': 0.02330115798395127, 'timing/model': 1.2924814549623989, 'learning_rate/qwen_vl_interface': 6.578131980103128e-06, 'learning_rate/action_model': 6.578131980103128e-05, 'epoch': 0.87}) +[2026-06-15 23:12:13,881][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.34758713841438293, 'train/grad_norm_pre_clip': 24.31120491027832, 'timing/data': 0.02323444001376629, 'timing/model': 1.2388300059828907, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 0.87}) +[2026-06-15 23:12:18,216][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.010533934459090233, 'train/grad_norm_pre_clip': 1.8457566499710083, 'timing/data': 0.023863868962507695, 'timing/model': 1.1262828729813918, 'learning_rate/qwen_vl_interface': 6.572282982526702e-06, 'learning_rate/action_model': 6.572282982526702e-05, 'epoch': 0.87}) +[2026-06-15 23:12:22,724][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.07170713692903519, 'train/grad_norm_pre_clip': 11.501068115234375, 'timing/data': 0.023735588998533785, 'timing/model': 1.126839631004259, 'learning_rate/qwen_vl_interface': 6.5693576678264755e-06, 'learning_rate/action_model': 6.569357667826476e-05, 'epoch': 0.87}) +[2026-06-15 23:12:27,239][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.20384378731250763, 'train/grad_norm_pre_clip': 15.98005485534668, 'timing/data': 0.02270533802220598, 'timing/model': 1.1319792240392417, 'learning_rate/qwen_vl_interface': 6.566431810788148e-06, 'learning_rate/action_model': 6.566431810788148e-05, 'epoch': 0.87}) +[2026-06-15 23:12:31,749][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.006683184765279293, 'train/grad_norm_pre_clip': 0.5594202876091003, 'timing/data': 0.022739661973901093, 'timing/model': 1.1263940850039944, 'learning_rate/qwen_vl_interface': 6.5635054126144306e-06, 'learning_rate/action_model': 6.56350541261443e-05, 'epoch': 0.87}) +[2026-06-15 23:12:36,272][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.016699466854333878, 'train/grad_norm_pre_clip': 8.236570358276367, 'timing/data': 0.024331662978511304, 'timing/model': 1.2884681849973276, 'learning_rate/qwen_vl_interface': 6.560578474508254e-06, 'learning_rate/action_model': 6.560578474508253e-05, 'epoch': 0.87}) +[2026-06-15 23:12:40,808][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.10663211345672607, 'train/grad_norm_pre_clip': 14.66422176361084, 'timing/data': 0.02407710999250412, 'timing/model': 1.2827807710273191, 'learning_rate/qwen_vl_interface': 6.5576509976727685e-06, 'learning_rate/action_model': 6.557650997672769e-05, 'epoch': 0.87}) +[2026-06-15 23:12:45,154][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.14456500113010406, 'train/grad_norm_pre_clip': 14.86015796661377, 'timing/data': 0.030978611030150205, 'timing/model': 1.1275612420286052, 'learning_rate/qwen_vl_interface': 6.554722983311354e-06, 'learning_rate/action_model': 6.554722983311355e-05, 'epoch': 0.87}) +[2026-06-15 23:12:49,679][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.19777370989322662, 'train/grad_norm_pre_clip': 13.940778732299805, 'timing/data': 0.023152024019509554, 'timing/model': 1.1281821390148252, 'learning_rate/qwen_vl_interface': 6.551794432627605e-06, 'learning_rate/action_model': 6.551794432627605e-05, 'epoch': 0.87}) +[2026-06-15 23:12:54,200][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.050058092921972275, 'train/grad_norm_pre_clip': 14.585411071777344, 'timing/data': 0.023711178044322878, 'timing/model': 1.133379814971704, 'learning_rate/qwen_vl_interface': 6.548865346825338e-06, 'learning_rate/action_model': 6.548865346825338e-05, 'epoch': 0.87}) +[2026-06-15 23:12:58,731][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.1657755970954895, 'train/grad_norm_pre_clip': 13.88365650177002, 'timing/data': 0.023445423983503133, 'timing/model': 1.126675006991718, 'learning_rate/qwen_vl_interface': 6.545935727108589e-06, 'learning_rate/action_model': 6.545935727108588e-05, 'epoch': 0.87}) +[2026-06-15 23:13:03,199][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.002544249640777707, 'train/grad_norm_pre_clip': 0.34034764766693115, 'timing/data': 0.023376600991468877, 'timing/model': 1.1242763349437155, 'learning_rate/qwen_vl_interface': 6.543005574681613e-06, 'learning_rate/action_model': 6.543005574681613e-05, 'epoch': 0.87}) +[2026-06-15 23:13:07,702][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.015137194655835629, 'train/grad_norm_pre_clip': 3.2229673862457275, 'timing/data': 0.023924854991491884, 'timing/model': 1.2819046250078827, 'learning_rate/qwen_vl_interface': 6.5400748907488886e-06, 'learning_rate/action_model': 6.540074890748888e-05, 'epoch': 0.87}) +[2026-06-15 23:13:12,051][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.11603366583585739, 'train/grad_norm_pre_clip': 6.4206671714782715, 'timing/data': 0.023782250005751848, 'timing/model': 1.1262088710209355, 'learning_rate/qwen_vl_interface': 6.537143676515105e-06, 'learning_rate/action_model': 6.537143676515105e-05, 'epoch': 0.87}) +[2026-06-15 23:13:16,567][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.20666316151618958, 'train/grad_norm_pre_clip': 17.5921573638916, 'timing/data': 0.023605415015481412, 'timing/model': 1.1265476839616895, 'learning_rate/qwen_vl_interface': 6.534211933185176e-06, 'learning_rate/action_model': 6.534211933185176e-05, 'epoch': 0.87}) +[2026-06-15 23:13:20,981][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.18342918157577515, 'train/grad_norm_pre_clip': 10.580028533935547, 'timing/data': 0.024250498972833157, 'timing/model': 1.1294267969788052, 'learning_rate/qwen_vl_interface': 6.531279661964231e-06, 'learning_rate/action_model': 6.531279661964231e-05, 'epoch': 0.87}) +[2026-06-15 23:13:25,401][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.003197639947757125, 'train/grad_norm_pre_clip': 0.3218391239643097, 'timing/data': 0.022700943984091282, 'timing/model': 1.1311867270269431, 'learning_rate/qwen_vl_interface': 6.528346864057616e-06, 'learning_rate/action_model': 6.528346864057615e-05, 'epoch': 0.87}) +[2026-06-15 23:13:29,862][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.31385740637779236, 'train/grad_norm_pre_clip': 19.23038101196289, 'timing/data': 0.023295671970117837, 'timing/model': 1.1435890090069734, 'learning_rate/qwen_vl_interface': 6.5254135406708906e-06, 'learning_rate/action_model': 6.52541354067089e-05, 'epoch': 0.87}) +[2026-06-15 23:13:34,378][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.00584696140140295, 'train/grad_norm_pre_clip': 3.41388201713562, 'timing/data': 0.024082378018647432, 'timing/model': 1.262288925005123, 'learning_rate/qwen_vl_interface': 6.522479693009836e-06, 'learning_rate/action_model': 6.522479693009837e-05, 'epoch': 0.87}) +[2026-06-15 23:13:38,751][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.18253150582313538, 'train/grad_norm_pre_clip': 16.425495147705078, 'timing/data': 0.02371741202659905, 'timing/model': 1.1325501509709284, 'learning_rate/qwen_vl_interface': 6.519545322280446e-06, 'learning_rate/action_model': 6.519545322280446e-05, 'epoch': 0.87}) +[2026-06-15 23:13:43,289][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.160811185836792, 'train/grad_norm_pre_clip': 10.755524635314941, 'timing/data': 0.023865599010605365, 'timing/model': 1.1326677469769493, 'learning_rate/qwen_vl_interface': 6.5166104296889275e-06, 'learning_rate/action_model': 6.516610429688927e-05, 'epoch': 0.87}) +[2026-06-15 23:13:47,793][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.0635468140244484, 'train/grad_norm_pre_clip': 10.533872604370117, 'timing/data': 0.023674455005675554, 'timing/model': 1.1303658910328522, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.88}) +[2026-06-15 23:13:52,296][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.29078635573387146, 'train/grad_norm_pre_clip': 20.559228897094727, 'timing/data': 0.023019056010525674, 'timing/model': 1.1259624480153434, 'learning_rate/qwen_vl_interface': 6.510739083745422e-06, 'learning_rate/action_model': 6.510739083745422e-05, 'epoch': 0.88}) +[2026-06-15 23:13:56,824][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.14885658025741577, 'train/grad_norm_pre_clip': 9.202461242675781, 'timing/data': 0.023783749027643353, 'timing/model': 1.2824739739880897, 'learning_rate/qwen_vl_interface': 6.507802632806919e-06, 'learning_rate/action_model': 6.507802632806919e-05, 'epoch': 0.88}) +[2026-06-15 23:14:01,190][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.169723242521286, 'train/grad_norm_pre_clip': 14.332359313964844, 'timing/data': 0.022946409008000046, 'timing/model': 1.1310245030326769, 'learning_rate/qwen_vl_interface': 6.5048656648332646e-06, 'learning_rate/action_model': 6.504865664833264e-05, 'epoch': 0.88}) +[2026-06-15 23:14:05,716][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.1941804438829422, 'train/grad_norm_pre_clip': 3.8592607975006104, 'timing/data': 0.023155254020821303, 'timing/model': 1.1343959859805182, 'learning_rate/qwen_vl_interface': 6.501928181031737e-06, 'learning_rate/action_model': 6.501928181031737e-05, 'epoch': 0.88}) +[2026-06-15 23:14:10,205][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.1772845983505249, 'train/grad_norm_pre_clip': 16.747713088989258, 'timing/data': 0.029648119991179556, 'timing/model': 1.1345592419966124, 'learning_rate/qwen_vl_interface': 6.498990182609823e-06, 'learning_rate/action_model': 6.498990182609823e-05, 'epoch': 0.88}) +[2026-06-15 23:14:14,761][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.06347798556089401, 'train/grad_norm_pre_clip': 12.222411155700684, 'timing/data': 0.027625449001789093, 'timing/model': 1.139439216000028, 'learning_rate/qwen_vl_interface': 6.4960516707752255e-06, 'learning_rate/action_model': 6.496051670775226e-05, 'epoch': 0.88}) +[2026-06-15 23:14:19,291][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.01201492827385664, 'train/grad_norm_pre_clip': 1.897971510887146, 'timing/data': 0.024146543990354985, 'timing/model': 1.132779284962453, 'learning_rate/qwen_vl_interface': 6.493112646735851e-06, 'learning_rate/action_model': 6.493112646735851e-05, 'epoch': 0.88}) +[2026-06-15 23:14:23,821][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.125804141163826, 'train/grad_norm_pre_clip': 13.207917213439941, 'timing/data': 0.02404976199613884, 'timing/model': 1.291447577998042, 'learning_rate/qwen_vl_interface': 6.490173111699824e-06, 'learning_rate/action_model': 6.490173111699825e-05, 'epoch': 0.88}) +[2026-06-15 23:14:28,322][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.12206514924764633, 'train/grad_norm_pre_clip': 14.46345043182373, 'timing/data': 0.023320562031585723, 'timing/model': 1.2443977530347183, 'learning_rate/qwen_vl_interface': 6.487233066875479e-06, 'learning_rate/action_model': 6.487233066875479e-05, 'epoch': 0.88}) +[2026-06-15 23:14:32,698][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.02270643785595894, 'train/grad_norm_pre_clip': 2.0830180644989014, 'timing/data': 0.023767331964336336, 'timing/model': 1.1282532640034333, 'learning_rate/qwen_vl_interface': 6.484292513471352e-06, 'learning_rate/action_model': 6.484292513471352e-05, 'epoch': 0.88}) +[2026-06-15 23:14:37,219][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.3330942988395691, 'train/grad_norm_pre_clip': 20.63242530822754, 'timing/data': 0.023759723000694066, 'timing/model': 1.129957580997143, 'learning_rate/qwen_vl_interface': 6.4813514526961975e-06, 'learning_rate/action_model': 6.481351452696198e-05, 'epoch': 0.88}) +[2026-06-15 23:14:41,749][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.17465831339359283, 'train/grad_norm_pre_clip': 16.457117080688477, 'timing/data': 0.023179899028036743, 'timing/model': 1.1337805920047686, 'learning_rate/qwen_vl_interface': 6.478409885758972e-06, 'learning_rate/action_model': 6.478409885758971e-05, 'epoch': 0.88}) +[2026-06-15 23:14:46,276][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.026352016255259514, 'train/grad_norm_pre_clip': 5.927875995635986, 'timing/data': 0.02315486903535202, 'timing/model': 1.1294136489741504, 'learning_rate/qwen_vl_interface': 6.475467813868844e-06, 'learning_rate/action_model': 6.475467813868844e-05, 'epoch': 0.88}) +[2026-06-15 23:14:50,804][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.0841347798705101, 'train/grad_norm_pre_clip': 6.674304962158203, 'timing/data': 0.022907753009349108, 'timing/model': 1.2834370689815842, 'learning_rate/qwen_vl_interface': 6.472525238235187e-06, 'learning_rate/action_model': 6.472525238235186e-05, 'epoch': 0.88}) +[2026-06-15 23:14:55,309][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.030950626358389854, 'train/grad_norm_pre_clip': 3.8559610843658447, 'timing/data': 0.02286125300452113, 'timing/model': 1.2714010559720919, 'learning_rate/qwen_vl_interface': 6.4695821600675834e-06, 'learning_rate/action_model': 6.469582160067583e-05, 'epoch': 0.88}) +[2026-06-15 23:14:59,684][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.05748114734888077, 'train/grad_norm_pre_clip': 4.720080852508545, 'timing/data': 0.023086520028300583, 'timing/model': 1.129331964009907, 'learning_rate/qwen_vl_interface': 6.466638580575824e-06, 'learning_rate/action_model': 6.466638580575824e-05, 'epoch': 0.88}) +[2026-06-15 23:15:04,237][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.11545907706022263, 'train/grad_norm_pre_clip': 7.041679859161377, 'timing/data': 0.023939924023579806, 'timing/model': 1.1300936660263687, 'learning_rate/qwen_vl_interface': 6.463694500969898e-06, 'learning_rate/action_model': 6.463694500969898e-05, 'epoch': 0.88}) +[2026-06-15 23:15:08,778][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.10344269871711731, 'train/grad_norm_pre_clip': 7.994625091552734, 'timing/data': 0.02339837799081579, 'timing/model': 1.1382238889927976, 'learning_rate/qwen_vl_interface': 6.460749922460008e-06, 'learning_rate/action_model': 6.460749922460008e-05, 'epoch': 0.88}) +[2026-06-15 23:15:13,318][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.06932839751243591, 'train/grad_norm_pre_clip': 9.696843147277832, 'timing/data': 0.023788928985595703, 'timing/model': 1.1318243540008552, 'learning_rate/qwen_vl_interface': 6.457804846256559e-06, 'learning_rate/action_model': 6.457804846256559e-05, 'epoch': 0.88}) +[2026-06-15 23:15:17,816][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.021619245409965515, 'train/grad_norm_pre_clip': 1.8377400636672974, 'timing/data': 0.023409468005411327, 'timing/model': 1.132228871982079, 'learning_rate/qwen_vl_interface': 6.45485927357016e-06, 'learning_rate/action_model': 6.454859273570159e-05, 'epoch': 0.88}) +[2026-06-15 23:15:22,355][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.04520729184150696, 'train/grad_norm_pre_clip': 3.4720771312713623, 'timing/data': 0.023223501979373395, 'timing/model': 1.2900075610377826, 'learning_rate/qwen_vl_interface': 6.451913205611626e-06, 'learning_rate/action_model': 6.451913205611626e-05, 'epoch': 0.88}) +[2026-06-15 23:15:26,731][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.14963285624980927, 'train/grad_norm_pre_clip': 18.404613494873047, 'timing/data': 0.02390253299381584, 'timing/model': 1.1324256449588574, 'learning_rate/qwen_vl_interface': 6.448966643591974e-06, 'learning_rate/action_model': 6.448966643591974e-05, 'epoch': 0.88}) +[2026-06-15 23:15:31,276][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.10219576209783554, 'train/grad_norm_pre_clip': 8.11437702178955, 'timing/data': 0.02418173901969567, 'timing/model': 1.1343536300119013, 'learning_rate/qwen_vl_interface': 6.446019588722423e-06, 'learning_rate/action_model': 6.446019588722422e-05, 'epoch': 0.88}) +[2026-06-15 23:15:35,728][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.09109790623188019, 'train/grad_norm_pre_clip': 10.885466575622559, 'timing/data': 0.022919501003343612, 'timing/model': 1.1375896410318092, 'learning_rate/qwen_vl_interface': 6.443072042214398e-06, 'learning_rate/action_model': 6.443072042214397e-05, 'epoch': 0.88}) +[2026-06-15 23:15:40,191][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.17132332921028137, 'train/grad_norm_pre_clip': 11.417893409729004, 'timing/data': 0.023042681044898927, 'timing/model': 1.13464002602268, 'learning_rate/qwen_vl_interface': 6.440124005279521e-06, 'learning_rate/action_model': 6.440124005279521e-05, 'epoch': 0.89}) +[2026-06-15 23:15:44,677][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.05926775932312012, 'train/grad_norm_pre_clip': 14.04524040222168, 'timing/data': 0.023170463973656297, 'timing/model': 1.1292618539882824, 'learning_rate/qwen_vl_interface': 6.43717547912962e-06, 'learning_rate/action_model': 6.437175479129621e-05, 'epoch': 0.89}) +[2026-06-15 23:15:49,193][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.07247724384069443, 'train/grad_norm_pre_clip': 10.377066612243652, 'timing/data': 0.02288561500608921, 'timing/model': 1.2637134109972976, 'learning_rate/qwen_vl_interface': 6.434226464976725e-06, 'learning_rate/action_model': 6.434226464976724e-05, 'epoch': 0.89}) +[2026-06-15 23:15:53,569][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.019082237035036087, 'train/grad_norm_pre_clip': 2.8807456493377686, 'timing/data': 0.022694454004522413, 'timing/model': 1.131373829965014, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 0.89}) +[2026-06-15 23:15:58,110][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.11873321235179901, 'train/grad_norm_pre_clip': 8.336509704589844, 'timing/data': 0.023568422999233007, 'timing/model': 1.1319080439861864, 'learning_rate/qwen_vl_interface': 6.428326977511059e-06, 'learning_rate/action_model': 6.428326977511058e-05, 'epoch': 0.89}) +[2026-06-15 23:16:02,647][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.19119995832443237, 'train/grad_norm_pre_clip': 11.77855396270752, 'timing/data': 0.02317782904719934, 'timing/model': 1.1360444549936801, 'learning_rate/qwen_vl_interface': 6.425376506623344e-06, 'learning_rate/action_model': 6.425376506623344e-05, 'epoch': 0.89}) +[2026-06-15 23:16:07,196][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.01575295813381672, 'train/grad_norm_pre_clip': 1.6546751260757446, 'timing/data': 0.023193700006231666, 'timing/model': 1.1359145199530758, 'learning_rate/qwen_vl_interface': 6.422425552582747e-06, 'learning_rate/action_model': 6.422425552582747e-05, 'epoch': 0.89}) +[2026-06-15 23:16:11,751][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.012968342751264572, 'train/grad_norm_pre_clip': 1.7413631677627563, 'timing/data': 0.022987990989349782, 'timing/model': 1.2969116110471077, 'learning_rate/qwen_vl_interface': 6.41947411660229e-06, 'learning_rate/action_model': 6.41947411660229e-05, 'epoch': 0.89}) +[2026-06-15 23:16:16,128][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.18437087535858154, 'train/grad_norm_pre_clip': 10.3728609085083, 'timing/data': 0.023625319998245686, 'timing/model': 1.13233518198831, 'learning_rate/qwen_vl_interface': 6.4165221998951995e-06, 'learning_rate/action_model': 6.4165221998952e-05, 'epoch': 0.89}) +[2026-06-15 23:16:20,658][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.05135411024093628, 'train/grad_norm_pre_clip': 6.355038166046143, 'timing/data': 0.023695575015153736, 'timing/model': 1.1305708980071358, 'learning_rate/qwen_vl_interface': 6.413569803674896e-06, 'learning_rate/action_model': 6.413569803674896e-05, 'epoch': 0.89}) +[2026-06-15 23:16:25,134][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.004109497647732496, 'train/grad_norm_pre_clip': 0.3151710629463196, 'timing/data': 0.02320148696890101, 'timing/model': 1.1363399549736641, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 0.89}) +[2026-06-15 23:16:29,685][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.05901490896940231, 'train/grad_norm_pre_clip': 5.332192420959473, 'timing/data': 0.023268574033863842, 'timing/model': 1.1297320390003733, 'learning_rate/qwen_vl_interface': 6.40766357754932e-06, 'learning_rate/action_model': 6.407663577549319e-05, 'epoch': 0.89}) +[2026-06-15 23:16:34,238][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.03489508107304573, 'train/grad_norm_pre_clip': 4.2025628089904785, 'timing/data': 0.023492584004998207, 'timing/model': 1.1332348180003464, 'learning_rate/qwen_vl_interface': 6.404709750071875e-06, 'learning_rate/action_model': 6.404709750071874e-05, 'epoch': 0.89}) +[2026-06-15 23:16:38,789][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.2197590172290802, 'train/grad_norm_pre_clip': 15.874692916870117, 'timing/data': 0.03062877699267119, 'timing/model': 1.298882816976402, 'learning_rate/qwen_vl_interface': 6.401755447936869e-06, 'learning_rate/action_model': 6.401755447936869e-05, 'epoch': 0.89}) +[2026-06-15 23:16:43,302][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.005085506942123175, 'train/grad_norm_pre_clip': 0.8371262550354004, 'timing/data': 0.02261971397092566, 'timing/model': 1.2591114210081287, 'learning_rate/qwen_vl_interface': 6.3988006723587036e-06, 'learning_rate/action_model': 6.398800672358703e-05, 'epoch': 0.89}) +[2026-06-15 23:16:47,677][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.011343131773173809, 'train/grad_norm_pre_clip': 3.350768804550171, 'timing/data': 0.02309185703052208, 'timing/model': 1.1371587129542604, 'learning_rate/qwen_vl_interface': 6.395845424551975e-06, 'learning_rate/action_model': 6.395845424551975e-05, 'epoch': 0.89}) +[2026-06-15 23:16:52,223][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.17114295065402985, 'train/grad_norm_pre_clip': 28.810178756713867, 'timing/data': 0.022956603032071143, 'timing/model': 1.1363898019772023, 'learning_rate/qwen_vl_interface': 6.392889705731475e-06, 'learning_rate/action_model': 6.392889705731476e-05, 'epoch': 0.89}) +[2026-06-15 23:16:56,742][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.010692920535802841, 'train/grad_norm_pre_clip': 2.054669141769409, 'timing/data': 0.022962191957049072, 'timing/model': 1.1315097120241262, 'learning_rate/qwen_vl_interface': 6.389933517112188e-06, 'learning_rate/action_model': 6.389933517112187e-05, 'epoch': 0.89}) +[2026-06-15 23:17:01,296][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.00515562342479825, 'train/grad_norm_pre_clip': 0.714198887348175, 'timing/data': 0.025584243005141616, 'timing/model': 1.1403805650188588, 'learning_rate/qwen_vl_interface': 6.386976859909293e-06, 'learning_rate/action_model': 6.386976859909292e-05, 'epoch': 0.89}) +[2026-06-15 23:17:05,811][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.12148350477218628, 'train/grad_norm_pre_clip': 10.711287498474121, 'timing/data': 0.023412392009049654, 'timing/model': 1.2804044359945692, 'learning_rate/qwen_vl_interface': 6.384019735338156e-06, 'learning_rate/action_model': 6.384019735338156e-05, 'epoch': 0.89}) +[2026-06-15 23:17:10,329][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.005536752752959728, 'train/grad_norm_pre_clip': 0.9861577749252319, 'timing/data': 0.023304781992919743, 'timing/model': 1.280052557005547, 'learning_rate/qwen_vl_interface': 6.381062144614342e-06, 'learning_rate/action_model': 6.381062144614342e-05, 'epoch': 0.89}) +[2026-06-15 23:17:14,709][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.00209642737172544, 'train/grad_norm_pre_clip': 0.4270994961261749, 'timing/data': 0.02280263596912846, 'timing/model': 1.1399250049726106, 'learning_rate/qwen_vl_interface': 6.378104088953605e-06, 'learning_rate/action_model': 6.378104088953605e-05, 'epoch': 0.89}) +[2026-06-15 23:17:19,241][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.034576352685689926, 'train/grad_norm_pre_clip': 12.150877952575684, 'timing/data': 0.0232264639926143, 'timing/model': 1.1301710809930228, 'learning_rate/qwen_vl_interface': 6.375145569571889e-06, 'learning_rate/action_model': 6.375145569571889e-05, 'epoch': 0.89}) +[2026-06-15 23:17:23,772][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.0336725190281868, 'train/grad_norm_pre_clip': 5.778100490570068, 'timing/data': 0.023485544021241367, 'timing/model': 1.1300955979968421, 'learning_rate/qwen_vl_interface': 6.372186587685329e-06, 'learning_rate/action_model': 6.372186587685329e-05, 'epoch': 0.89}) +[2026-06-15 23:17:28,301][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.04381959140300751, 'train/grad_norm_pre_clip': 16.005029678344727, 'timing/data': 0.023633131000678986, 'timing/model': 1.1297962369862944, 'learning_rate/qwen_vl_interface': 6.369227144510252e-06, 'learning_rate/action_model': 6.369227144510252e-05, 'epoch': 0.9}) +[2026-06-15 23:17:32,792][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.2818579077720642, 'train/grad_norm_pre_clip': 26.85358428955078, 'timing/data': 0.023104593972675502, 'timing/model': 1.1304107289761305, 'learning_rate/qwen_vl_interface': 6.3662672412631736e-06, 'learning_rate/action_model': 6.366267241263173e-05, 'epoch': 0.9}) +[2026-06-15 23:17:37,344][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.11048632115125656, 'train/grad_norm_pre_clip': 18.540014266967773, 'timing/data': 0.022599229007028043, 'timing/model': 1.3066658289753832, 'learning_rate/qwen_vl_interface': 6.3633068791607954e-06, 'learning_rate/action_model': 6.363306879160795e-05, 'epoch': 0.9}) +[2026-06-15 23:17:41,710][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.06398145109415054, 'train/grad_norm_pre_clip': 13.41346549987793, 'timing/data': 0.023154522001277655, 'timing/model': 1.131574837025255, 'learning_rate/qwen_vl_interface': 6.360346059420014e-06, 'learning_rate/action_model': 6.360346059420014e-05, 'epoch': 0.9}) +[2026-06-15 23:17:46,235][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.1410607546567917, 'train/grad_norm_pre_clip': 6.045877933502197, 'timing/data': 0.02319922496099025, 'timing/model': 1.1327307219617069, 'learning_rate/qwen_vl_interface': 6.357384783257909e-06, 'learning_rate/action_model': 6.357384783257908e-05, 'epoch': 0.9}) +[2026-06-15 23:17:50,664][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.013481993228197098, 'train/grad_norm_pre_clip': 5.077042102813721, 'timing/data': 0.023718482989352196, 'timing/model': 1.1297177050146274, 'learning_rate/qwen_vl_interface': 6.354423051891748e-06, 'learning_rate/action_model': 6.354423051891749e-05, 'epoch': 0.9}) +[2026-06-15 23:17:55,099][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.3856965899467468, 'train/grad_norm_pre_clip': 27.617877960205078, 'timing/data': 0.023838235007133335, 'timing/model': 1.1313082359847613, 'learning_rate/qwen_vl_interface': 6.35146086653899e-06, 'learning_rate/action_model': 6.35146086653899e-05, 'epoch': 0.9}) +[2026-06-15 23:17:59,578][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.04521622508764267, 'train/grad_norm_pre_clip': 8.436528205871582, 'timing/data': 0.02358857198851183, 'timing/model': 1.130320475960616, 'learning_rate/qwen_vl_interface': 6.348498228417273e-06, 'learning_rate/action_model': 6.348498228417273e-05, 'epoch': 0.9}) +[2026-06-15 23:18:04,086][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.0016409656964242458, 'train/grad_norm_pre_clip': 0.20724532008171082, 'timing/data': 0.023348606016952544, 'timing/model': 1.2645027490216307, 'learning_rate/qwen_vl_interface': 6.34553513874443e-06, 'learning_rate/action_model': 6.34553513874443e-05, 'epoch': 0.9}) +[2026-06-15 23:18:08,473][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.20077629387378693, 'train/grad_norm_pre_clip': 35.081905364990234, 'timing/data': 0.02262233302462846, 'timing/model': 1.1371906470158137, 'learning_rate/qwen_vl_interface': 6.342571598738473e-06, 'learning_rate/action_model': 6.342571598738472e-05, 'epoch': 0.9}) +[2026-06-15 23:18:13,025][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.0017491626786068082, 'train/grad_norm_pre_clip': 0.2872278392314911, 'timing/data': 0.023534714011475444, 'timing/model': 1.130719868990127, 'learning_rate/qwen_vl_interface': 6.339607609617603e-06, 'learning_rate/action_model': 6.339607609617603e-05, 'epoch': 0.9}) +[2026-06-15 23:18:17,574][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.0015068815555423498, 'train/grad_norm_pre_clip': 0.198708176612854, 'timing/data': 0.023797706002369523, 'timing/model': 1.1380313550471328, 'learning_rate/qwen_vl_interface': 6.336643172600202e-06, 'learning_rate/action_model': 6.336643172600202e-05, 'epoch': 0.9}) +[2026-06-15 23:18:22,125][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.007666163146495819, 'train/grad_norm_pre_clip': 6.299756050109863, 'timing/data': 0.02354000398190692, 'timing/model': 1.131627855997067, 'learning_rate/qwen_vl_interface': 6.333678288904838e-06, 'learning_rate/action_model': 6.333678288904837e-05, 'epoch': 0.9}) +[2026-06-15 23:18:26,653][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.005561360158026218, 'train/grad_norm_pre_clip': 1.1003738641738892, 'timing/data': 0.0234523739782162, 'timing/model': 1.2938836549874395, 'learning_rate/qwen_vl_interface': 6.330712959750264e-06, 'learning_rate/action_model': 6.330712959750263e-05, 'epoch': 0.9}) +[2026-06-15 23:18:31,029][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.04979019612073898, 'train/grad_norm_pre_clip': 20.273151397705078, 'timing/data': 0.023183561977930367, 'timing/model': 1.1299508360098116, 'learning_rate/qwen_vl_interface': 6.3277471863554145e-06, 'learning_rate/action_model': 6.327747186355414e-05, 'epoch': 0.9}) +[2026-06-15 23:18:35,573][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.01111860666424036, 'train/grad_norm_pre_clip': 7.030988693237305, 'timing/data': 0.023727794003207237, 'timing/model': 1.1391919130110182, 'learning_rate/qwen_vl_interface': 6.324780969939406e-06, 'learning_rate/action_model': 6.324780969939405e-05, 'epoch': 0.9}) +[2026-06-15 23:18:40,052][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.3884836435317993, 'train/grad_norm_pre_clip': 25.125621795654297, 'timing/data': 0.023303427034989, 'timing/model': 1.1440347540192306, 'learning_rate/qwen_vl_interface': 6.32181431172154e-06, 'learning_rate/action_model': 6.32181431172154e-05, 'epoch': 0.9}) +[2026-06-15 23:18:44,581][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.12017790228128433, 'train/grad_norm_pre_clip': 11.19921875, 'timing/data': 0.023526759003289044, 'timing/model': 1.1336543879588135, 'learning_rate/qwen_vl_interface': 6.318847212921295e-06, 'learning_rate/action_model': 6.318847212921295e-05, 'epoch': 0.9}) +[2026-06-15 23:18:49,131][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.10810928046703339, 'train/grad_norm_pre_clip': 14.92266845703125, 'timing/data': 0.02374425297603011, 'timing/model': 1.1340449430281296, 'learning_rate/qwen_vl_interface': 6.3158796747583364e-06, 'learning_rate/action_model': 6.315879674758336e-05, 'epoch': 0.9}) +[2026-06-15 23:18:53,655][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.12302616238594055, 'train/grad_norm_pre_clip': 13.612168312072754, 'timing/data': 0.0231893919990398, 'timing/model': 1.2841931990114972, 'learning_rate/qwen_vl_interface': 6.312911698452503e-06, 'learning_rate/action_model': 6.312911698452502e-05, 'epoch': 0.9}) +[2026-06-15 23:18:58,156][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.0589311458170414, 'train/grad_norm_pre_clip': 23.156517028808594, 'timing/data': 0.023008330026641488, 'timing/model': 1.2464169079903513, 'learning_rate/qwen_vl_interface': 6.309943285223824e-06, 'learning_rate/action_model': 6.309943285223824e-05, 'epoch': 0.9}) +[2026-06-15 23:23:21,324][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.2107299268245697, 'train/grad_norm_pre_clip': 28.044471740722656, 'eval/loss': 0.07568547129631042, 'eval/flap_precision': 0.9487951807228916, 'eval/flap_recall': 0.9722222222222222, 'eval/flap_f1': 0.9603658536585367, 'eval/flap_rate_model': 0.18444444444444444, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9603658536585368, 'timing/data': 0.023747290018945932, 'timing/model': 1.1306687989854254, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.9}) +[2026-06-15 23:23:21,324][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2200_state +[2026-06-15 23:23:21,326][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:23:24,916][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2200_state/model.safetensors +[2026-06-15 23:23:34,959][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2200_state/optimizer.bin +[2026-06-15 23:23:34,959][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2200_state/sampler.bin +[2026-06-15 23:23:34,959][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2200_state/sampler_1.bin +[2026-06-15 23:23:34,960][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2200_state/random_states_0.pkl +[2026-06-15 23:23:35,000][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2200 +[2026-06-15 23:23:35,001][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:23:35,007][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:23:39,284][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.05805548280477524, 'train/grad_norm_pre_clip': 15.955926895141602, 'timing/data': 0.02619246597168967, 'timing/model': 1.1114652850083075, 'learning_rate/qwen_vl_interface': 6.304005152878902e-06, 'learning_rate/action_model': 6.304005152878901e-05, 'epoch': 0.9}) +[2026-06-15 23:23:43,750][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.3354346752166748, 'train/grad_norm_pre_clip': 13.497969627380371, 'timing/data': 0.02348297496791929, 'timing/model': 1.1151751870056614, 'learning_rate/qwen_vl_interface': 6.301035436203602e-06, 'learning_rate/action_model': 6.301035436203602e-05, 'epoch': 0.9}) +[2026-06-15 23:23:48,137][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.15466850996017456, 'train/grad_norm_pre_clip': 20.450511932373047, 'timing/data': 0.02380995498970151, 'timing/model': 1.114183537021745, 'learning_rate/qwen_vl_interface': 6.298065287487337e-06, 'learning_rate/action_model': 6.298065287487337e-05, 'epoch': 0.9}) +[2026-06-15 23:23:52,567][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.10649368911981583, 'train/grad_norm_pre_clip': 20.89203643798828, 'timing/data': 0.023556831991299987, 'timing/model': 1.125299362000078, 'learning_rate/qwen_vl_interface': 6.295094707951019e-06, 'learning_rate/action_model': 6.295094707951019e-05, 'epoch': 0.91}) +[2026-06-15 23:23:57,039][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.002588780364021659, 'train/grad_norm_pre_clip': 0.45673519372940063, 'timing/data': 0.0236871269880794, 'timing/model': 1.1273859349894337, 'learning_rate/qwen_vl_interface': 6.292123698815745e-06, 'learning_rate/action_model': 6.292123698815745e-05, 'epoch': 0.91}) +[2026-06-15 23:24:01,521][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.0033851899206638336, 'train/grad_norm_pre_clip': 0.5938820838928223, 'timing/data': 0.023427272972185165, 'timing/model': 1.2804586120182648, 'learning_rate/qwen_vl_interface': 6.289152261302783e-06, 'learning_rate/action_model': 6.289152261302784e-05, 'epoch': 0.91}) +[2026-06-15 23:24:05,856][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.07701897621154785, 'train/grad_norm_pre_clip': 13.619307518005371, 'timing/data': 0.02374417899409309, 'timing/model': 1.1259300829842687, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 0.91}) +[2026-06-15 23:24:10,359][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.007687224540859461, 'train/grad_norm_pre_clip': 5.5544233322143555, 'timing/data': 0.02377165894722566, 'timing/model': 1.1266571449814364, 'learning_rate/qwen_vl_interface': 6.283208106029753e-06, 'learning_rate/action_model': 6.283208106029752e-05, 'epoch': 0.91}) +[2026-06-15 23:24:14,848][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.004872309975326061, 'train/grad_norm_pre_clip': 0.8374790549278259, 'timing/data': 0.023712504014838487, 'timing/model': 1.1267322169733234, 'learning_rate/qwen_vl_interface': 6.280235390713102e-06, 'learning_rate/action_model': 6.280235390713102e-05, 'epoch': 0.91}) +[2026-06-15 23:24:19,331][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.2806268036365509, 'train/grad_norm_pre_clip': 20.561389923095703, 'timing/data': 0.023367425019387156, 'timing/model': 1.1270951859769411, 'learning_rate/qwen_vl_interface': 6.277262251905597e-06, 'learning_rate/action_model': 6.277262251905597e-05, 'epoch': 0.91}) +[2026-06-15 23:24:23,846][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.08735281974077225, 'train/grad_norm_pre_clip': 10.241531372070312, 'timing/data': 0.024038459989242256, 'timing/model': 1.1252949530025944, 'learning_rate/qwen_vl_interface': 6.274288690829385e-06, 'learning_rate/action_model': 6.274288690829384e-05, 'epoch': 0.91}) +[2026-06-15 23:24:28,385][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.0846104621887207, 'train/grad_norm_pre_clip': 12.477906227111816, 'timing/data': 0.031740384991280735, 'timing/model': 1.300164325046353, 'learning_rate/qwen_vl_interface': 6.271314708706781e-06, 'learning_rate/action_model': 6.271314708706782e-05, 'epoch': 0.91}) +[2026-06-15 23:24:32,747][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.055128443986177444, 'train/grad_norm_pre_clip': 20.56426429748535, 'timing/data': 0.024932708009146154, 'timing/model': 1.1309666949673556, 'learning_rate/qwen_vl_interface': 6.26834030676028e-06, 'learning_rate/action_model': 6.26834030676028e-05, 'epoch': 0.91}) +[2026-06-15 23:24:37,253][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.007597491145133972, 'train/grad_norm_pre_clip': 1.894800066947937, 'timing/data': 0.023845007992349565, 'timing/model': 1.1239573209895752, 'learning_rate/qwen_vl_interface': 6.265365486212545e-06, 'learning_rate/action_model': 6.265365486212545e-05, 'epoch': 0.91}) +[2026-06-15 23:24:41,766][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.0006772666238248348, 'train/grad_norm_pre_clip': 0.09992429614067078, 'timing/data': 0.0235928239999339, 'timing/model': 1.1287161629879847, 'learning_rate/qwen_vl_interface': 6.26239024828641e-06, 'learning_rate/action_model': 6.26239024828641e-05, 'epoch': 0.91}) +[2026-06-15 23:24:46,286][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.07807614654302597, 'train/grad_norm_pre_clip': 20.404085159301758, 'timing/data': 0.023405536951031536, 'timing/model': 1.1291732520330697, 'learning_rate/qwen_vl_interface': 6.259414594204888e-06, 'learning_rate/action_model': 6.259414594204888e-05, 'epoch': 0.91}) +[2026-06-15 23:24:50,801][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.08116414397954941, 'train/grad_norm_pre_clip': 13.386920928955078, 'timing/data': 0.02289256500080228, 'timing/model': 1.284030698996503, 'learning_rate/qwen_vl_interface': 6.256438525191155e-06, 'learning_rate/action_model': 6.256438525191156e-05, 'epoch': 0.91}) +[2026-06-15 23:24:55,301][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.008368798531591892, 'train/grad_norm_pre_clip': 2.0701870918273926, 'timing/data': 0.023574861988890916, 'timing/model': 1.2672594059840776, 'learning_rate/qwen_vl_interface': 6.253462042468561e-06, 'learning_rate/action_model': 6.25346204246856e-05, 'epoch': 0.91}) +[2026-06-15 23:24:59,660][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.00015501574671361595, 'train/grad_norm_pre_clip': 0.07591918855905533, 'timing/data': 0.02294248901307583, 'timing/model': 1.1293648889986798, 'learning_rate/qwen_vl_interface': 6.2504851472606246e-06, 'learning_rate/action_model': 6.250485147260625e-05, 'epoch': 0.91}) +[2026-06-15 23:25:04,196][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.1765080988407135, 'train/grad_norm_pre_clip': 28.90616226196289, 'timing/data': 0.023146634979639202, 'timing/model': 1.1344364239485003, 'learning_rate/qwen_vl_interface': 6.247507840791035e-06, 'learning_rate/action_model': 6.247507840791035e-05, 'epoch': 0.91}) +[2026-06-15 23:25:08,726][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.009675754234194756, 'train/grad_norm_pre_clip': 1.7810546159744263, 'timing/data': 0.024057906004600227, 'timing/model': 1.126465042005293, 'learning_rate/qwen_vl_interface': 6.244530124283651e-06, 'learning_rate/action_model': 6.24453012428365e-05, 'epoch': 0.91}) +[2026-06-15 23:25:13,250][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.007022785022854805, 'train/grad_norm_pre_clip': 1.4610518217086792, 'timing/data': 0.024118292960338295, 'timing/model': 1.128206213004887, 'learning_rate/qwen_vl_interface': 6.2415519989625016e-06, 'learning_rate/action_model': 6.241551998962501e-05, 'epoch': 0.91}) +[2026-06-15 23:25:17,731][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.022693468257784843, 'train/grad_norm_pre_clip': 15.616229057312012, 'timing/data': 0.02317546895937994, 'timing/model': 1.1362410489819013, 'learning_rate/qwen_vl_interface': 6.238573466051781e-06, 'learning_rate/action_model': 6.23857346605178e-05, 'epoch': 0.91}) +[2026-06-15 23:25:22,272][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.36854732036590576, 'train/grad_norm_pre_clip': 11.57955265045166, 'timing/data': 0.02313412498915568, 'timing/model': 1.2995337700122036, 'learning_rate/qwen_vl_interface': 6.2355945267758465e-06, 'learning_rate/action_model': 6.235594526775846e-05, 'epoch': 0.91}) +[2026-06-15 23:25:26,654][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.2237001210451126, 'train/grad_norm_pre_clip': 20.863574981689453, 'timing/data': 0.022475623001810163, 'timing/model': 1.1340986419818364, 'learning_rate/qwen_vl_interface': 6.232615182359231e-06, 'learning_rate/action_model': 6.232615182359231e-05, 'epoch': 0.91}) +[2026-06-15 23:25:31,212][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.013708293437957764, 'train/grad_norm_pre_clip': 4.725215911865234, 'timing/data': 0.024037856957875192, 'timing/model': 1.1355136989732273, 'learning_rate/qwen_vl_interface': 6.229635434026634e-06, 'learning_rate/action_model': 6.229635434026634e-05, 'epoch': 0.91}) +[2026-06-15 23:25:35,715][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.096527598798275, 'train/grad_norm_pre_clip': 21.258562088012695, 'timing/data': 0.031047877040691674, 'timing/model': 1.1498535170103423, 'learning_rate/qwen_vl_interface': 6.226655283002912e-06, 'learning_rate/action_model': 6.226655283002911e-05, 'epoch': 0.91}) +[2026-06-15 23:25:40,174][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.0034633397590368986, 'train/grad_norm_pre_clip': 1.7567362785339355, 'timing/data': 0.023873039986938238, 'timing/model': 1.1385881659807637, 'learning_rate/qwen_vl_interface': 6.223674730513096e-06, 'learning_rate/action_model': 6.223674730513096e-05, 'epoch': 0.92}) +[2026-06-15 23:25:44,637][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.15042506158351898, 'train/grad_norm_pre_clip': 20.83295440673828, 'timing/data': 0.023866361007094383, 'timing/model': 1.1316429380094633, 'learning_rate/qwen_vl_interface': 6.220693777782375e-06, 'learning_rate/action_model': 6.220693777782375e-05, 'epoch': 0.92}) +[2026-06-15 23:25:49,154][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.10474947839975357, 'train/grad_norm_pre_clip': 11.284320831298828, 'timing/data': 0.023200655996333808, 'timing/model': 1.2629566229879856, 'learning_rate/qwen_vl_interface': 6.21771242603611e-06, 'learning_rate/action_model': 6.21771242603611e-05, 'epoch': 0.92}) +[2026-06-15 23:25:53,548][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.00803343765437603, 'train/grad_norm_pre_clip': 4.360905647277832, 'timing/data': 0.023461017000954598, 'timing/model': 1.1441251349751838, 'learning_rate/qwen_vl_interface': 6.21473067649982e-06, 'learning_rate/action_model': 6.21473067649982e-05, 'epoch': 0.92}) +[2026-06-15 23:25:58,095][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.09483614563941956, 'train/grad_norm_pre_clip': 13.940074920654297, 'timing/data': 0.023330531956162304, 'timing/model': 1.1323371269972995, 'learning_rate/qwen_vl_interface': 6.211748530399189e-06, 'learning_rate/action_model': 6.211748530399189e-05, 'epoch': 0.92}) +[2026-06-15 23:26:02,631][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.13053588569164276, 'train/grad_norm_pre_clip': 28.221263885498047, 'timing/data': 0.022180503990966827, 'timing/model': 1.1329607180086896, 'learning_rate/qwen_vl_interface': 6.208765988960067e-06, 'learning_rate/action_model': 6.208765988960066e-05, 'epoch': 0.92}) +[2026-06-15 23:26:07,217][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.5450672507286072, 'train/grad_norm_pre_clip': 29.48357391357422, 'timing/data': 0.023832530016079545, 'timing/model': 1.1433969689533114, 'learning_rate/qwen_vl_interface': 6.2057830534084615e-06, 'learning_rate/action_model': 6.205783053408461e-05, 'epoch': 0.92}) +[2026-06-15 23:26:11,762][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.10428286343812943, 'train/grad_norm_pre_clip': 12.517014503479004, 'timing/data': 0.023180365038570017, 'timing/model': 1.3005083670141175, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 0.92}) +[2026-06-15 23:26:16,137][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.06131814420223236, 'train/grad_norm_pre_clip': 8.516279220581055, 'timing/data': 0.02394758799346164, 'timing/model': 1.132730143959634, 'learning_rate/qwen_vl_interface': 6.199816004872655e-06, 'learning_rate/action_model': 6.199816004872655e-05, 'epoch': 0.92}) +[2026-06-15 23:26:20,672][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.004005043301731348, 'train/grad_norm_pre_clip': 1.2714875936508179, 'timing/data': 0.02327756502199918, 'timing/model': 1.1318340720026754, 'learning_rate/qwen_vl_interface': 6.196831894341281e-06, 'learning_rate/action_model': 6.19683189434128e-05, 'epoch': 0.92}) +[2026-06-15 23:26:25,154][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.018188312649726868, 'train/grad_norm_pre_clip': 5.240559101104736, 'timing/data': 0.024278702039737254, 'timing/model': 1.136405367986299, 'learning_rate/qwen_vl_interface': 6.1938473946030795e-06, 'learning_rate/action_model': 6.193847394603079e-05, 'epoch': 0.92}) +[2026-06-15 23:26:29,704][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.14666421711444855, 'train/grad_norm_pre_clip': 20.487802505493164, 'timing/data': 0.02370422496460378, 'timing/model': 1.1385344720329158, 'learning_rate/qwen_vl_interface': 6.1908625068848695e-06, 'learning_rate/action_model': 6.19086250688487e-05, 'epoch': 0.92}) +[2026-06-15 23:26:34,260][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.11785241961479187, 'train/grad_norm_pre_clip': 16.483896255493164, 'timing/data': 0.023477170965634286, 'timing/model': 1.142908118024934, 'learning_rate/qwen_vl_interface': 6.187877232413619e-06, 'learning_rate/action_model': 6.18787723241362e-05, 'epoch': 0.92}) +[2026-06-15 23:26:38,807][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.03213847801089287, 'train/grad_norm_pre_clip': 7.239622116088867, 'timing/data': 0.024005656014196575, 'timing/model': 1.2967535490170121, 'learning_rate/qwen_vl_interface': 6.184891572416467e-06, 'learning_rate/action_model': 6.184891572416466e-05, 'epoch': 0.92}) +[2026-06-15 23:26:43,305][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.07493267208337784, 'train/grad_norm_pre_clip': 18.353187561035156, 'timing/data': 0.02399698196677491, 'timing/model': 1.2496370080043562, 'learning_rate/qwen_vl_interface': 6.181905528120703e-06, 'learning_rate/action_model': 6.181905528120703e-05, 'epoch': 0.92}) +[2026-06-15 23:26:47,687][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.1725941002368927, 'train/grad_norm_pre_clip': 9.07080078125, 'timing/data': 0.024093357962556183, 'timing/model': 1.1361777349957265, 'learning_rate/qwen_vl_interface': 6.178919100753778e-06, 'learning_rate/action_model': 6.178919100753778e-05, 'epoch': 0.92}) +[2026-06-15 23:26:52,241][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.013560082763433456, 'train/grad_norm_pre_clip': 4.859674453735352, 'timing/data': 0.02427750499919057, 'timing/model': 1.1407909879926592, 'learning_rate/qwen_vl_interface': 6.175932291543299e-06, 'learning_rate/action_model': 6.175932291543298e-05, 'epoch': 0.92}) +[2026-06-15 23:26:56,805][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.09766899794340134, 'train/grad_norm_pre_clip': 10.050257682800293, 'timing/data': 0.02387416997225955, 'timing/model': 1.1421435810043477, 'learning_rate/qwen_vl_interface': 6.172945101717029e-06, 'learning_rate/action_model': 6.17294510171703e-05, 'epoch': 0.92}) +[2026-06-15 23:27:01,356][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.34114113450050354, 'train/grad_norm_pre_clip': 29.357982635498047, 'timing/data': 0.023788560007233173, 'timing/model': 1.130094358988572, 'learning_rate/qwen_vl_interface': 6.169957532502892e-06, 'learning_rate/action_model': 6.169957532502892e-05, 'epoch': 0.92}) +[2026-06-15 23:27:05,897][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.007837194949388504, 'train/grad_norm_pre_clip': 1.8851886987686157, 'timing/data': 0.023239904025103897, 'timing/model': 1.2872734509874135, 'learning_rate/qwen_vl_interface': 6.166969585128962e-06, 'learning_rate/action_model': 6.166969585128962e-05, 'epoch': 0.92}) +[2026-06-15 23:27:10,442][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.12043030560016632, 'train/grad_norm_pre_clip': 9.375299453735352, 'timing/data': 0.023369916947558522, 'timing/model': 1.2660818789736368, 'learning_rate/qwen_vl_interface': 6.163981260823473e-06, 'learning_rate/action_model': 6.163981260823473e-05, 'epoch': 0.92}) +[2026-06-15 23:27:14,951][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.10616671293973923, 'train/grad_norm_pre_clip': 12.14887809753418, 'timing/data': 0.023536198015790433, 'timing/model': 1.1323205099906772, 'learning_rate/qwen_vl_interface': 6.16099256081481e-06, 'learning_rate/action_model': 6.16099256081481e-05, 'epoch': 0.92}) +[2026-06-15 23:27:19,489][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.01771271415054798, 'train/grad_norm_pre_clip': 8.657184600830078, 'timing/data': 0.023665194981731474, 'timing/model': 1.139998344995547, 'learning_rate/qwen_vl_interface': 6.15800348633152e-06, 'learning_rate/action_model': 6.15800348633152e-05, 'epoch': 0.92}) +[2026-06-15 23:27:24,011][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.17476586997509003, 'train/grad_norm_pre_clip': 17.347618103027344, 'timing/data': 0.024464312009513378, 'timing/model': 1.1323603839846328, 'learning_rate/qwen_vl_interface': 6.155014038602293e-06, 'learning_rate/action_model': 6.155014038602292e-05, 'epoch': 0.92}) +[2026-06-15 23:27:28,563][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.12523901462554932, 'train/grad_norm_pre_clip': 5.634679794311523, 'timing/data': 0.02321629400830716, 'timing/model': 1.1400163380312733, 'learning_rate/qwen_vl_interface': 6.152024218855979e-06, 'learning_rate/action_model': 6.15202421885598e-05, 'epoch': 0.93}) +[2026-06-15 23:27:33,063][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.3001382350921631, 'train/grad_norm_pre_clip': 20.351520538330078, 'timing/data': 0.0231112539768219, 'timing/model': 1.138305479020346, 'learning_rate/qwen_vl_interface': 6.149034028321583e-06, 'learning_rate/action_model': 6.149034028321583e-05, 'epoch': 0.93}) +[2026-06-15 23:27:37,609][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.013850802555680275, 'train/grad_norm_pre_clip': 2.59112286567688, 'timing/data': 0.02336806704988703, 'timing/model': 1.2958036619820632, 'learning_rate/qwen_vl_interface': 6.1460434682282566e-06, 'learning_rate/action_model': 6.146043468228256e-05, 'epoch': 0.93}) +[2026-06-15 23:27:42,023][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.04399444907903671, 'train/grad_norm_pre_clip': 8.837197303771973, 'timing/data': 0.023701000958681107, 'timing/model': 1.139849261031486, 'learning_rate/qwen_vl_interface': 6.143052539805308e-06, 'learning_rate/action_model': 6.143052539805308e-05, 'epoch': 0.93}) +[2026-06-15 23:27:46,585][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.008330650627613068, 'train/grad_norm_pre_clip': 0.7390857338905334, 'timing/data': 0.023487384023610502, 'timing/model': 1.1342488970258273, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 0.93}) +[2026-06-15 23:27:51,017][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.08078699558973312, 'train/grad_norm_pre_clip': 28.10650062561035, 'timing/data': 0.02404312405269593, 'timing/model': 1.1319267300423235, 'learning_rate/qwen_vl_interface': 6.137069582888521e-06, 'learning_rate/action_model': 6.137069582888521e-05, 'epoch': 0.93}) +[2026-06-15 23:27:55,452][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.12353266775608063, 'train/grad_norm_pre_clip': 11.642838478088379, 'timing/data': 0.02387248503509909, 'timing/model': 1.133172478992492, 'learning_rate/qwen_vl_interface': 6.13407755685405e-06, 'learning_rate/action_model': 6.13407755685405e-05, 'epoch': 0.93}) +[2026-06-15 23:27:59,937][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.20546667277812958, 'train/grad_norm_pre_clip': 15.797294616699219, 'timing/data': 0.02372479101177305, 'timing/model': 1.1327264650026336, 'learning_rate/qwen_vl_interface': 6.131085167408692e-06, 'learning_rate/action_model': 6.131085167408692e-05, 'epoch': 0.93}) +[2026-06-15 23:28:04,479][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.08674805611371994, 'train/grad_norm_pre_clip': 12.065641403198242, 'timing/data': 0.023554558982141316, 'timing/model': 1.2635223559918813, 'learning_rate/qwen_vl_interface': 6.128092415782502e-06, 'learning_rate/action_model': 6.128092415782501e-05, 'epoch': 0.93}) +[2026-06-15 23:28:08,895][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.255325049161911, 'train/grad_norm_pre_clip': 14.910187721252441, 'timing/data': 0.023208538012113422, 'timing/model': 1.1457283460185863, 'learning_rate/qwen_vl_interface': 6.125099303205688e-06, 'learning_rate/action_model': 6.125099303205688e-05, 'epoch': 0.93}) +[2026-06-15 23:28:13,442][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.26622137427330017, 'train/grad_norm_pre_clip': 13.6170072555542, 'timing/data': 0.023589311982505023, 'timing/model': 1.1444782660109922, 'learning_rate/qwen_vl_interface': 6.122105830908607e-06, 'learning_rate/action_model': 6.122105830908606e-05, 'epoch': 0.93}) +[2026-06-15 23:28:18,000][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.10709860175848007, 'train/grad_norm_pre_clip': 14.28688907623291, 'timing/data': 0.023961508995853364, 'timing/model': 1.134462098008953, 'learning_rate/qwen_vl_interface': 6.11911200012176e-06, 'learning_rate/action_model': 6.11911200012176e-05, 'epoch': 0.93}) +[2026-06-15 23:28:22,561][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.06349533051252365, 'train/grad_norm_pre_clip': 5.961186408996582, 'timing/data': 0.024202075961511582, 'timing/model': 1.1356835089973174, 'learning_rate/qwen_vl_interface': 6.116117812075799e-06, 'learning_rate/action_model': 6.116117812075799e-05, 'epoch': 0.93}) +[2026-06-15 23:28:27,125][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.014249410480260849, 'train/grad_norm_pre_clip': 2.3411357402801514, 'timing/data': 0.023735692026093602, 'timing/model': 1.3088633169536479, 'learning_rate/qwen_vl_interface': 6.113123268001524e-06, 'learning_rate/action_model': 6.113123268001524e-05, 'epoch': 0.93}) +[2026-06-15 23:28:31,511][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.10700178146362305, 'train/grad_norm_pre_clip': 11.193609237670898, 'timing/data': 0.02411177899921313, 'timing/model': 1.1334283290198073, 'learning_rate/qwen_vl_interface': 6.110128369129873e-06, 'learning_rate/action_model': 6.110128369129873e-05, 'epoch': 0.93}) +[2026-06-15 23:28:36,054][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.011737838387489319, 'train/grad_norm_pre_clip': 1.922345519065857, 'timing/data': 0.0236490840325132, 'timing/model': 1.1468795830151066, 'learning_rate/qwen_vl_interface': 6.107133116691943e-06, 'learning_rate/action_model': 6.107133116691942e-05, 'epoch': 0.93}) +[2026-06-15 23:28:40,544][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.08674056082963943, 'train/grad_norm_pre_clip': 11.733241081237793, 'timing/data': 0.023307970026507974, 'timing/model': 1.139436202996876, 'learning_rate/qwen_vl_interface': 6.104137511918963e-06, 'learning_rate/action_model': 6.104137511918963e-05, 'epoch': 0.93}) +[2026-06-15 23:28:45,095][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.21622946858406067, 'train/grad_norm_pre_clip': 19.04550552368164, 'timing/data': 0.023648371046874672, 'timing/model': 1.1355263069854118, 'learning_rate/qwen_vl_interface': 6.1011415560423155e-06, 'learning_rate/action_model': 6.101141556042316e-05, 'epoch': 0.93}) +[2026-06-15 23:28:49,680][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.08438117057085037, 'train/grad_norm_pre_clip': 10.37093448638916, 'timing/data': 0.02366408600937575, 'timing/model': 1.1373135920148343, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.93}) +[2026-06-15 23:28:54,237][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.16650882363319397, 'train/grad_norm_pre_clip': 20.063879013061523, 'timing/data': 0.02328746096463874, 'timing/model': 1.3058050069957972, 'learning_rate/qwen_vl_interface': 6.09514859590426e-06, 'learning_rate/action_model': 6.09514859590426e-05, 'epoch': 0.93}) +[2026-06-15 23:28:58,757][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.1494656652212143, 'train/grad_norm_pre_clip': 7.860283851623535, 'timing/data': 0.024936173984315246, 'timing/model': 1.263531206001062, 'learning_rate/qwen_vl_interface': 6.09215159410633e-06, 'learning_rate/action_model': 6.09215159410633e-05, 'epoch': 0.93}) +[2026-06-15 23:29:03,148][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.058385543525218964, 'train/grad_norm_pre_clip': 7.264878273010254, 'timing/data': 0.02411087800282985, 'timing/model': 1.135185207996983, 'learning_rate/qwen_vl_interface': 6.089154246131692e-06, 'learning_rate/action_model': 6.089154246131692e-05, 'epoch': 0.93}) +[2026-06-15 23:29:07,700][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.08871444314718246, 'train/grad_norm_pre_clip': 6.657136917114258, 'timing/data': 0.023412090027704835, 'timing/model': 1.1419873590348288, 'learning_rate/qwen_vl_interface': 6.08615655321244e-06, 'learning_rate/action_model': 6.08615655321244e-05, 'epoch': 0.93}) +[2026-06-15 23:29:12,242][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.003564896062016487, 'train/grad_norm_pre_clip': 0.3551146388053894, 'timing/data': 0.023407675966154784, 'timing/model': 1.1383230719948187, 'learning_rate/qwen_vl_interface': 6.083158516580814e-06, 'learning_rate/action_model': 6.083158516580814e-05, 'epoch': 0.93}) +[2026-06-15 23:29:16,808][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.0075169336050748825, 'train/grad_norm_pre_clip': 0.596831202507019, 'timing/data': 0.02428179298294708, 'timing/model': 1.1401298100245185, 'learning_rate/qwen_vl_interface': 6.080160137469192e-06, 'learning_rate/action_model': 6.0801601374691916e-05, 'epoch': 0.93}) +[2026-06-15 23:29:21,364][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.0023303613997995853, 'train/grad_norm_pre_clip': 0.3105185627937317, 'timing/data': 0.02341644000262022, 'timing/model': 1.3005139169981703, 'learning_rate/qwen_vl_interface': 6.077161417110097e-06, 'learning_rate/action_model': 6.0771614171100965e-05, 'epoch': 0.94}) +[2026-06-15 23:29:25,897][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.013548709452152252, 'train/grad_norm_pre_clip': 5.2327423095703125, 'timing/data': 0.02311543602263555, 'timing/model': 1.275025925016962, 'learning_rate/qwen_vl_interface': 6.0741623567361864e-06, 'learning_rate/action_model': 6.074162356736186e-05, 'epoch': 0.94}) +[2026-06-15 23:29:30,289][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.3745886981487274, 'train/grad_norm_pre_clip': 21.299516677856445, 'timing/data': 0.023252922983374447, 'timing/model': 1.1366680380306207, 'learning_rate/qwen_vl_interface': 6.071162957580262e-06, 'learning_rate/action_model': 6.071162957580262e-05, 'epoch': 0.94}) +[2026-06-15 23:29:34,816][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.006519720423966646, 'train/grad_norm_pre_clip': 0.6495442390441895, 'timing/data': 0.02325078600551933, 'timing/model': 1.1377794060390443, 'learning_rate/qwen_vl_interface': 6.068163220875266e-06, 'learning_rate/action_model': 6.068163220875266e-05, 'epoch': 0.94}) +[2026-06-15 23:29:39,366][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.007474180776625872, 'train/grad_norm_pre_clip': 1.3027935028076172, 'timing/data': 0.02391575399087742, 'timing/model': 1.1357364790164866, 'learning_rate/qwen_vl_interface': 6.065163147854273e-06, 'learning_rate/action_model': 6.065163147854273e-05, 'epoch': 0.94}) +[2026-06-15 23:29:43,903][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.0041840351186692715, 'train/grad_norm_pre_clip': 0.5451472401618958, 'timing/data': 0.02331056003458798, 'timing/model': 1.1344555199611932, 'learning_rate/qwen_vl_interface': 6.062162739750502e-06, 'learning_rate/action_model': 6.062162739750502e-05, 'epoch': 0.94}) +[2026-06-15 23:29:48,407][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.3505057096481323, 'train/grad_norm_pre_clip': 12.949504852294922, 'timing/data': 0.024064821016509086, 'timing/model': 1.1375314350007102, 'learning_rate/qwen_vl_interface': 6.059161997797306e-06, 'learning_rate/action_model': 6.0591619977973056e-05, 'epoch': 0.94}) +[2026-06-15 23:29:52,955][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.029477033764123917, 'train/grad_norm_pre_clip': 10.647828102111816, 'timing/data': 0.02359798998804763, 'timing/model': 1.2966424830374308, 'learning_rate/qwen_vl_interface': 6.056160923228178e-06, 'learning_rate/action_model': 6.056160923228178e-05, 'epoch': 0.94}) +[2026-06-15 23:29:57,350][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.038099586963653564, 'train/grad_norm_pre_clip': 9.459650993347168, 'timing/data': 0.023204701952636242, 'timing/model': 1.138136513007339, 'learning_rate/qwen_vl_interface': 6.0531595172767455e-06, 'learning_rate/action_model': 6.0531595172767455e-05, 'epoch': 0.94}) +[2026-06-15 23:30:01,886][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.0791468620300293, 'train/grad_norm_pre_clip': 20.49094009399414, 'timing/data': 0.02362481300951913, 'timing/model': 1.1368536649970338, 'learning_rate/qwen_vl_interface': 6.050157781176774e-06, 'learning_rate/action_model': 6.050157781176774e-05, 'epoch': 0.94}) +[2026-06-15 23:30:06,335][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.13968296349048615, 'train/grad_norm_pre_clip': 4.160025596618652, 'timing/data': 0.024117086024489254, 'timing/model': 1.1335372379980981, 'learning_rate/qwen_vl_interface': 6.047155716162164e-06, 'learning_rate/action_model': 6.047155716162164e-05, 'epoch': 0.94}) +[2026-06-15 23:30:10,779][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.08839165419340134, 'train/grad_norm_pre_clip': 13.909625053405762, 'timing/data': 0.02390045003267005, 'timing/model': 1.136901844991371, 'learning_rate/qwen_vl_interface': 6.044153323466948e-06, 'learning_rate/action_model': 6.044153323466948e-05, 'epoch': 0.94}) +[2026-06-15 23:30:15,261][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.1360359638929367, 'train/grad_norm_pre_clip': 15.28384780883789, 'timing/data': 0.023176540969870985, 'timing/model': 1.1352635489893146, 'learning_rate/qwen_vl_interface': 6.0411506043253e-06, 'learning_rate/action_model': 6.0411506043253e-05, 'epoch': 0.94}) +[2026-06-15 23:30:19,793][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.12402588129043579, 'train/grad_norm_pre_clip': 9.150453567504883, 'timing/data': 0.02387183503014967, 'timing/model': 1.2639997559599578, 'learning_rate/qwen_vl_interface': 6.038147559971523e-06, 'learning_rate/action_model': 6.038147559971523e-05, 'epoch': 0.94}) +[2026-06-15 23:30:24,180][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.17986097931861877, 'train/grad_norm_pre_clip': 13.235182762145996, 'timing/data': 0.023224916018079966, 'timing/model': 1.1338587430072948, 'learning_rate/qwen_vl_interface': 6.035144191640055e-06, 'learning_rate/action_model': 6.035144191640055e-05, 'epoch': 0.94}) +[2026-06-15 23:30:28,745][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.0077346013858914375, 'train/grad_norm_pre_clip': 1.1996017694473267, 'timing/data': 0.024426762014627457, 'timing/model': 1.1396322619984858, 'learning_rate/qwen_vl_interface': 6.032140500565467e-06, 'learning_rate/action_model': 6.0321405005654665e-05, 'epoch': 0.94}) +[2026-06-15 23:30:33,295][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.21412910521030426, 'train/grad_norm_pre_clip': 18.43691062927246, 'timing/data': 0.023151959001552314, 'timing/model': 1.1409207690157928, 'learning_rate/qwen_vl_interface': 6.029136487982463e-06, 'learning_rate/action_model': 6.029136487982463e-05, 'epoch': 0.94}) +[2026-06-15 23:30:37,859][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.015198391862213612, 'train/grad_norm_pre_clip': 2.8692526817321777, 'timing/data': 0.024371932027861476, 'timing/model': 1.1461948239593767, 'learning_rate/qwen_vl_interface': 6.02613215512588e-06, 'learning_rate/action_model': 6.0261321551258796e-05, 'epoch': 0.94}) +[2026-06-15 23:30:42,419][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.3039983808994293, 'train/grad_norm_pre_clip': 25.524072647094727, 'timing/data': 0.02338539797347039, 'timing/model': 1.2925889259786345, 'learning_rate/qwen_vl_interface': 6.023127503230684e-06, 'learning_rate/action_model': 6.023127503230683e-05, 'epoch': 0.94}) +[2026-06-15 23:30:46,817][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.00480663264170289, 'train/grad_norm_pre_clip': 0.5483770370483398, 'timing/data': 0.02406952896853909, 'timing/model': 1.1403882869635709, 'learning_rate/qwen_vl_interface': 6.020122533531976e-06, 'learning_rate/action_model': 6.0201225335319755e-05, 'epoch': 0.94}) +[2026-06-15 23:30:51,372][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.003961487207561731, 'train/grad_norm_pre_clip': 0.4392540454864502, 'timing/data': 0.023405020008794963, 'timing/model': 1.1406487789936364, 'learning_rate/qwen_vl_interface': 6.017117247264983e-06, 'learning_rate/action_model': 6.017117247264983e-05, 'epoch': 0.94}) +[2026-06-15 23:30:55,861][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.19324736297130585, 'train/grad_norm_pre_clip': 18.540925979614258, 'timing/data': 0.024041165015660226, 'timing/model': 1.1335483909933828, 'learning_rate/qwen_vl_interface': 6.014111645665069e-06, 'learning_rate/action_model': 6.0141116456650684e-05, 'epoch': 0.94}) +[2026-06-15 23:31:00,412][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.05569900572299957, 'train/grad_norm_pre_clip': 8.997361183166504, 'timing/data': 0.023288463009521365, 'timing/model': 1.1356705219950527, 'learning_rate/qwen_vl_interface': 6.011105729967719e-06, 'learning_rate/action_model': 6.0111057299677184e-05, 'epoch': 0.94}) +[2026-06-15 23:35:23,141][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.010600247420370579, 'train/grad_norm_pre_clip': 1.2164840698242188, 'eval/loss': 0.05959569662809372, 'eval/flap_precision': 0.9454545454545454, 'eval/flap_recall': 0.9629629629629629, 'eval/flap_f1': 0.9541284403669724, 'eval/flap_rate_model': 0.18333333333333332, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9541284403669724, 'timing/data': 0.025130680005531758, 'timing/model': 1.1337716419948265, 'learning_rate/qwen_vl_interface': 6.008099501408552e-06, 'learning_rate/action_model': 6.008099501408552e-05, 'epoch': 0.94}) +[2026-06-15 23:35:23,142][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2300_state +[2026-06-15 23:35:23,144][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:35:26,698][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2300_state/model.safetensors +[2026-06-15 23:35:36,615][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2300_state/optimizer.bin +[2026-06-15 23:35:36,615][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2300_state/sampler.bin +[2026-06-15 23:35:36,615][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2300_state/sampler_1.bin +[2026-06-15 23:35:36,616][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2300_state/random_states_0.pkl +[2026-06-15 23:35:36,660][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2300 +[2026-06-15 23:35:36,661][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:35:36,667][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:35:41,016][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.14365330338478088, 'train/grad_norm_pre_clip': 10.26668930053711, 'timing/data': 0.027097946032881737, 'timing/model': 1.1167241560178809, 'learning_rate/qwen_vl_interface': 6.0050929612233165e-06, 'learning_rate/action_model': 6.005092961223316e-05, 'epoch': 0.95}) +[2026-06-15 23:35:45,419][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.03230729326605797, 'train/grad_norm_pre_clip': 16.866029739379883, 'timing/data': 0.023918877996038646, 'timing/model': 1.1213490959489718, 'learning_rate/qwen_vl_interface': 6.002086110647888e-06, 'learning_rate/action_model': 6.002086110647888e-05, 'epoch': 0.95}) +[2026-06-15 23:35:49,856][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.029669297859072685, 'train/grad_norm_pre_clip': 10.94806957244873, 'timing/data': 0.02510032500140369, 'timing/model': 1.2440999369719066, 'learning_rate/qwen_vl_interface': 5.999078950918267e-06, 'learning_rate/action_model': 5.999078950918267e-05, 'epoch': 0.95}) +[2026-06-15 23:35:54,179][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.003540593199431896, 'train/grad_norm_pre_clip': 0.3259608745574951, 'timing/data': 0.025771654036361724, 'timing/model': 1.1172901690006256, 'learning_rate/qwen_vl_interface': 5.996071483270585e-06, 'learning_rate/action_model': 5.996071483270585e-05, 'epoch': 0.95}) +[2026-06-15 23:35:58,666][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.003316952148452401, 'train/grad_norm_pre_clip': 0.4616638422012329, 'timing/data': 0.02332907705567777, 'timing/model': 1.1235227250144817, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.95}) +[2026-06-15 23:36:03,151][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.02066769078373909, 'train/grad_norm_pre_clip': 10.555736541748047, 'timing/data': 0.02368534402921796, 'timing/model': 1.1186196639901027, 'learning_rate/qwen_vl_interface': 5.990055629166187e-06, 'learning_rate/action_model': 5.990055629166187e-05, 'epoch': 0.95}) +[2026-06-15 23:36:07,645][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.05815492197871208, 'train/grad_norm_pre_clip': 12.504202842712402, 'timing/data': 0.023721020959783345, 'timing/model': 1.1210189529811032, 'learning_rate/qwen_vl_interface': 5.987047245182363e-06, 'learning_rate/action_model': 5.987047245182362e-05, 'epoch': 0.95}) +[2026-06-15 23:36:12,150][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.005557405762374401, 'train/grad_norm_pre_clip': 1.1226744651794434, 'timing/data': 0.02346740203211084, 'timing/model': 1.284117438015528, 'learning_rate/qwen_vl_interface': 5.984038558226255e-06, 'learning_rate/action_model': 5.9840385582262546e-05, 'epoch': 0.95}) +[2026-06-15 23:36:16,498][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.14525319635868073, 'train/grad_norm_pre_clip': 26.436237335205078, 'timing/data': 0.023674035968724638, 'timing/model': 1.12471252202522, 'learning_rate/qwen_vl_interface': 5.981029569534625e-06, 'learning_rate/action_model': 5.981029569534625e-05, 'epoch': 0.95}) +[2026-06-15 23:36:21,000][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.016718851402401924, 'train/grad_norm_pre_clip': 9.141851425170898, 'timing/data': 0.023203326971270144, 'timing/model': 1.1284922520280816, 'learning_rate/qwen_vl_interface': 5.978020280344351e-06, 'learning_rate/action_model': 5.9780202803443506e-05, 'epoch': 0.95}) +[2026-06-15 23:36:25,459][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.024356063455343246, 'train/grad_norm_pre_clip': 15.275961875915527, 'timing/data': 0.024133030034136027, 'timing/model': 1.1320921400329098, 'learning_rate/qwen_vl_interface': 5.97501069189244e-06, 'learning_rate/action_model': 5.97501069189244e-05, 'epoch': 0.95}) +[2026-06-15 23:36:29,981][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.014099705033004284, 'train/grad_norm_pre_clip': 4.820091724395752, 'timing/data': 0.024094538995996118, 'timing/model': 1.1324648130103014, 'learning_rate/qwen_vl_interface': 5.97200080541602e-06, 'learning_rate/action_model': 5.97200080541602e-05, 'epoch': 0.95}) +[2026-06-15 23:36:34,499][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.23396199941635132, 'train/grad_norm_pre_clip': 11.571853637695312, 'timing/data': 0.02406477404292673, 'timing/model': 1.1262270340230316, 'learning_rate/qwen_vl_interface': 5.968990622152341e-06, 'learning_rate/action_model': 5.9689906221523404e-05, 'epoch': 0.95}) +[2026-06-15 23:36:39,012][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.18730545043945312, 'train/grad_norm_pre_clip': 12.713949203491211, 'timing/data': 0.02426216599997133, 'timing/model': 1.2808530789916404, 'learning_rate/qwen_vl_interface': 5.965980143338774e-06, 'learning_rate/action_model': 5.965980143338774e-05, 'epoch': 0.95}) +[2026-06-15 23:36:43,495][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.15671515464782715, 'train/grad_norm_pre_clip': 20.598800659179688, 'timing/data': 0.023457949981093407, 'timing/model': 1.2464761629817076, 'learning_rate/qwen_vl_interface': 5.962969370212819e-06, 'learning_rate/action_model': 5.9629693702128185e-05, 'epoch': 0.95}) +[2026-06-15 23:36:47,860][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.007352753076702356, 'train/grad_norm_pre_clip': 3.285038471221924, 'timing/data': 0.022448953008279204, 'timing/model': 1.1314437090186402, 'learning_rate/qwen_vl_interface': 5.959958304012085e-06, 'learning_rate/action_model': 5.9599583040120845e-05, 'epoch': 0.95}) +[2026-06-15 23:36:52,382][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.0285494364798069, 'train/grad_norm_pre_clip': 8.535006523132324, 'timing/data': 0.023931166040711105, 'timing/model': 1.1264533780049533, 'learning_rate/qwen_vl_interface': 5.956946945974309e-06, 'learning_rate/action_model': 5.9569469459743085e-05, 'epoch': 0.95}) +[2026-06-15 23:36:56,902][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.0786428153514862, 'train/grad_norm_pre_clip': 6.2704997062683105, 'timing/data': 0.02334923599846661, 'timing/model': 1.1297907869447954, 'learning_rate/qwen_vl_interface': 5.953935297337347e-06, 'learning_rate/action_model': 5.953935297337347e-05, 'epoch': 0.95}) +[2026-06-15 23:37:01,429][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.0009351545013487339, 'train/grad_norm_pre_clip': 0.1761854737997055, 'timing/data': 0.02324005705304444, 'timing/model': 1.131645555025898, 'learning_rate/qwen_vl_interface': 5.950923359339175e-06, 'learning_rate/action_model': 5.950923359339174e-05, 'epoch': 0.95}) +[2026-06-15 23:37:05,979][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.22306552529335022, 'train/grad_norm_pre_clip': 19.655553817749023, 'timing/data': 0.024378982023335993, 'timing/model': 1.306333254033234, 'learning_rate/qwen_vl_interface': 5.947911133217886e-06, 'learning_rate/action_model': 5.947911133217886e-05, 'epoch': 0.95}) +[2026-06-15 23:37:10,492][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.0965724065899849, 'train/grad_norm_pre_clip': 17.651601791381836, 'timing/data': 0.02367421699455008, 'timing/model': 1.2708305270061828, 'learning_rate/qwen_vl_interface': 5.944898620211693e-06, 'learning_rate/action_model': 5.944898620211693e-05, 'epoch': 0.95}) +[2026-06-15 23:37:14,877][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.20142106711864471, 'train/grad_norm_pre_clip': 20.547531127929688, 'timing/data': 0.023780257964972407, 'timing/model': 1.1301750350394286, 'learning_rate/qwen_vl_interface': 5.941885821558925e-06, 'learning_rate/action_model': 5.9418858215589245e-05, 'epoch': 0.95}) +[2026-06-15 23:37:19,420][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.02752748504281044, 'train/grad_norm_pre_clip': 15.44481086730957, 'timing/data': 0.023638986982405186, 'timing/model': 1.1392302520107478, 'learning_rate/qwen_vl_interface': 5.938872738498031e-06, 'learning_rate/action_model': 5.93887273849803e-05, 'epoch': 0.95}) +[2026-06-15 23:37:23,962][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.1333201378583908, 'train/grad_norm_pre_clip': 16.966508865356445, 'timing/data': 0.02367122599389404, 'timing/model': 1.1352460799971595, 'learning_rate/qwen_vl_interface': 5.935859372267575e-06, 'learning_rate/action_model': 5.935859372267575e-05, 'epoch': 0.95}) +[2026-06-15 23:37:28,509][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.2139384150505066, 'train/grad_norm_pre_clip': 28.898338317871094, 'timing/data': 0.02317970700096339, 'timing/model': 1.1277651769923978, 'learning_rate/qwen_vl_interface': 5.9328457241062396e-06, 'learning_rate/action_model': 5.9328457241062394e-05, 'epoch': 0.96}) +[2026-06-15 23:37:32,995][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.07251925766468048, 'train/grad_norm_pre_clip': 21.795265197753906, 'timing/data': 0.022906443977262825, 'timing/model': 1.1324771950021386, 'learning_rate/qwen_vl_interface': 5.929831795252819e-06, 'learning_rate/action_model': 5.9298317952528184e-05, 'epoch': 0.96}) +[2026-06-15 23:37:37,527][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.15685278177261353, 'train/grad_norm_pre_clip': 25.857372283935547, 'timing/data': 0.022903551987838, 'timing/model': 1.289362731971778, 'learning_rate/qwen_vl_interface': 5.926817586946227e-06, 'learning_rate/action_model': 5.926817586946227e-05, 'epoch': 0.96}) +[2026-06-15 23:37:41,905][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.08657053858041763, 'train/grad_norm_pre_clip': 18.290985107421875, 'timing/data': 0.023151002009399235, 'timing/model': 1.1340968500007875, 'learning_rate/qwen_vl_interface': 5.923803100425491e-06, 'learning_rate/action_model': 5.92380310042549e-05, 'epoch': 0.96}) +[2026-06-15 23:37:46,443][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.13016879558563232, 'train/grad_norm_pre_clip': 12.009820938110352, 'timing/data': 0.023612686956766993, 'timing/model': 1.1304329390404746, 'learning_rate/qwen_vl_interface': 5.92078833692975e-06, 'learning_rate/action_model': 5.92078833692975e-05, 'epoch': 0.96}) +[2026-06-15 23:37:50,879][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.23654936254024506, 'train/grad_norm_pre_clip': 14.615453720092773, 'timing/data': 0.023795219021849334, 'timing/model': 1.1382778289844282, 'learning_rate/qwen_vl_interface': 5.917773297698264e-06, 'learning_rate/action_model': 5.9177732976982636e-05, 'epoch': 0.96}) +[2026-06-15 23:37:55,317][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.06265521794557571, 'train/grad_norm_pre_clip': 25.253541946411133, 'timing/data': 0.02359152300050482, 'timing/model': 1.1364516480243765, 'learning_rate/qwen_vl_interface': 5.9147579839704e-06, 'learning_rate/action_model': 5.9147579839704e-05, 'epoch': 0.96}) +[2026-06-15 23:37:59,793][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.15768501162528992, 'train/grad_norm_pre_clip': 18.303075790405273, 'timing/data': 0.023589877993799746, 'timing/model': 1.1345865639741533, 'learning_rate/qwen_vl_interface': 5.911742396985637e-06, 'learning_rate/action_model': 5.9117423969856364e-05, 'epoch': 0.96}) +[2026-06-15 23:38:04,310][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.008022054098546505, 'train/grad_norm_pre_clip': 1.505566954612732, 'timing/data': 0.023316528997384012, 'timing/model': 1.274947885016445, 'learning_rate/qwen_vl_interface': 5.908726537983571e-06, 'learning_rate/action_model': 5.908726537983571e-05, 'epoch': 0.96}) +[2026-06-15 23:38:08,685][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.012707705609500408, 'train/grad_norm_pre_clip': 2.5317232608795166, 'timing/data': 0.02360077202320099, 'timing/model': 1.131928225979209, 'learning_rate/qwen_vl_interface': 5.9057104082039085e-06, 'learning_rate/action_model': 5.905710408203908e-05, 'epoch': 0.96}) +[2026-06-15 23:38:13,224][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.019845841452479362, 'train/grad_norm_pre_clip': 7.53951358795166, 'timing/data': 0.023545422998722643, 'timing/model': 1.1340306050260551, 'learning_rate/qwen_vl_interface': 5.902694008886466e-06, 'learning_rate/action_model': 5.902694008886466e-05, 'epoch': 0.96}) +[2026-06-15 23:38:17,765][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.015321144834160805, 'train/grad_norm_pre_clip': 4.788922309875488, 'timing/data': 0.02358115103561431, 'timing/model': 1.1373806230258197, 'learning_rate/qwen_vl_interface': 5.8996773412711694e-06, 'learning_rate/action_model': 5.899677341271169e-05, 'epoch': 0.96}) +[2026-06-15 23:38:22,318][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.1700129508972168, 'train/grad_norm_pre_clip': 13.255311012268066, 'timing/data': 0.023383844003546983, 'timing/model': 1.1303469499689527, 'learning_rate/qwen_vl_interface': 5.896660406598059e-06, 'learning_rate/action_model': 5.896660406598059e-05, 'epoch': 0.96}) +[2026-06-15 23:38:26,866][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.0021096961572766304, 'train/grad_norm_pre_clip': 0.4166528284549713, 'timing/data': 0.023322255990933627, 'timing/model': 1.2944491190137342, 'learning_rate/qwen_vl_interface': 5.893643206107283e-06, 'learning_rate/action_model': 5.893643206107282e-05, 'epoch': 0.96}) +[2026-06-15 23:38:31,254][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.024465149268507957, 'train/grad_norm_pre_clip': 4.513277530670166, 'timing/data': 0.023167202016338706, 'timing/model': 1.13545255898498, 'learning_rate/qwen_vl_interface': 5.890625741039096e-06, 'learning_rate/action_model': 5.890625741039095e-05, 'epoch': 0.96}) +[2026-06-15 23:38:35,836][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.35709938406944275, 'train/grad_norm_pre_clip': 15.241558074951172, 'timing/data': 0.022746584028936923, 'timing/model': 1.1359148529591039, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 0.96}) +[2026-06-15 23:38:40,314][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.011435188353061676, 'train/grad_norm_pre_clip': 7.405745506286621, 'timing/data': 0.022864531027153134, 'timing/model': 1.134467089024838, 'learning_rate/qwen_vl_interface': 5.884590022132064e-06, 'learning_rate/action_model': 5.884590022132064e-05, 'epoch': 0.96}) +[2026-06-15 23:38:44,854][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.2323036640882492, 'train/grad_norm_pre_clip': 11.708789825439453, 'timing/data': 0.023264348972588778, 'timing/model': 1.1344378929934464, 'learning_rate/qwen_vl_interface': 5.881571770774278e-06, 'learning_rate/action_model': 5.881571770774278e-05, 'epoch': 0.96}) +[2026-06-15 23:38:49,405][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.0768350288271904, 'train/grad_norm_pre_clip': 9.651947021484375, 'timing/data': 0.02464232302736491, 'timing/model': 1.1368072840268724, 'learning_rate/qwen_vl_interface': 5.878553259801192e-06, 'learning_rate/action_model': 5.878553259801192e-05, 'epoch': 0.96}) +[2026-06-15 23:38:53,955][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.015219403430819511, 'train/grad_norm_pre_clip': 3.3287148475646973, 'timing/data': 0.023891828022897243, 'timing/model': 1.3019201490096748, 'learning_rate/qwen_vl_interface': 5.8755344904536025e-06, 'learning_rate/action_model': 5.8755344904536025e-05, 'epoch': 0.96}) +[2026-06-15 23:38:58,460][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.008501757867634296, 'train/grad_norm_pre_clip': 1.3881617784500122, 'timing/data': 0.023634925950318575, 'timing/model': 1.2533633100101724, 'learning_rate/qwen_vl_interface': 5.872515463972415e-06, 'learning_rate/action_model': 5.8725154639724146e-05, 'epoch': 0.96}) +[2026-06-15 23:39:02,835][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.004108493681997061, 'train/grad_norm_pre_clip': 0.38580840826034546, 'timing/data': 0.02297295699827373, 'timing/model': 1.1352171929902397, 'learning_rate/qwen_vl_interface': 5.869496181598633e-06, 'learning_rate/action_model': 5.869496181598633e-05, 'epoch': 0.96}) +[2026-06-15 23:39:07,397][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.03859809413552284, 'train/grad_norm_pre_clip': 8.506696701049805, 'timing/data': 0.02429153601406142, 'timing/model': 1.1409524139598943, 'learning_rate/qwen_vl_interface': 5.866476644573371e-06, 'learning_rate/action_model': 5.866476644573371e-05, 'epoch': 0.96}) +[2026-06-15 23:39:11,949][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.12550105154514313, 'train/grad_norm_pre_clip': 31.288169860839844, 'timing/data': 0.023317619983572513, 'timing/model': 1.1318364839535207, 'learning_rate/qwen_vl_interface': 5.863456854137848e-06, 'learning_rate/action_model': 5.8634568541378474e-05, 'epoch': 0.96}) +[2026-06-15 23:39:16,507][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.01101221889257431, 'train/grad_norm_pre_clip': 1.3817850351333618, 'timing/data': 0.022593956033233553, 'timing/model': 1.1344311269931495, 'learning_rate/qwen_vl_interface': 5.860436811533383e-06, 'learning_rate/action_model': 5.860436811533383e-05, 'epoch': 0.96}) +[2026-06-15 23:39:21,062][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.11944222450256348, 'train/grad_norm_pre_clip': 13.90200138092041, 'timing/data': 0.023976904980372638, 'timing/model': 1.2982359749730676, 'learning_rate/qwen_vl_interface': 5.857416518001404e-06, 'learning_rate/action_model': 5.857416518001404e-05, 'epoch': 0.97}) +[2026-06-15 23:39:25,600][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.05594343692064285, 'train/grad_norm_pre_clip': 20.503847122192383, 'timing/data': 0.023526908014900982, 'timing/model': 1.2863452939782292, 'learning_rate/qwen_vl_interface': 5.854395974783439e-06, 'learning_rate/action_model': 5.8543959747834386e-05, 'epoch': 0.97}) +[2026-06-15 23:39:29,980][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.181523859500885, 'train/grad_norm_pre_clip': 18.28800392150879, 'timing/data': 0.023931126983370632, 'timing/model': 1.1311160019831732, 'learning_rate/qwen_vl_interface': 5.851375183121119e-06, 'learning_rate/action_model': 5.8513751831211186e-05, 'epoch': 0.97}) +[2026-06-15 23:39:34,509][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.12076990306377411, 'train/grad_norm_pre_clip': 14.89689826965332, 'timing/data': 0.024041911005042493, 'timing/model': 1.129258774977643, 'learning_rate/qwen_vl_interface': 5.848354144256175e-06, 'learning_rate/action_model': 5.8483541442561754e-05, 'epoch': 0.97}) +[2026-06-15 23:39:39,055][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.23102907836437225, 'train/grad_norm_pre_clip': 25.983007431030273, 'timing/data': 0.023363763990346342, 'timing/model': 1.132897187024355, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 0.97}) +[2026-06-15 23:39:43,617][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.029200967401266098, 'train/grad_norm_pre_clip': 10.46200180053711, 'timing/data': 0.02326702297432348, 'timing/model': 1.1353635170380585, 'learning_rate/qwen_vl_interface': 5.842311329885866e-06, 'learning_rate/action_model': 5.842311329885866e-05, 'epoch': 0.97}) +[2026-06-15 23:39:48,131][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.10441496968269348, 'train/grad_norm_pre_clip': 7.62012243270874, 'timing/data': 0.024177259008865803, 'timing/model': 1.1317028839839622, 'learning_rate/qwen_vl_interface': 5.839289556864472e-06, 'learning_rate/action_model': 5.8392895568644714e-05, 'epoch': 0.97}) +[2026-06-15 23:39:52,677][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.20331266522407532, 'train/grad_norm_pre_clip': 47.2301139831543, 'timing/data': 0.02357237902469933, 'timing/model': 1.2914297410170548, 'learning_rate/qwen_vl_interface': 5.836267541608398e-06, 'learning_rate/action_model': 5.836267541608398e-05, 'epoch': 0.97}) +[2026-06-15 23:39:57,082][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.12678873538970947, 'train/grad_norm_pre_clip': 8.528033256530762, 'timing/data': 0.023158377967774868, 'timing/model': 1.1340226500178687, 'learning_rate/qwen_vl_interface': 5.833245285359887e-06, 'learning_rate/action_model': 5.833245285359886e-05, 'epoch': 0.97}) +[2026-06-15 23:40:01,685][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.003968348726630211, 'train/grad_norm_pre_clip': 0.5303786396980286, 'timing/data': 0.03173063800204545, 'timing/model': 1.1350862690014765, 'learning_rate/qwen_vl_interface': 5.830222789361268e-06, 'learning_rate/action_model': 5.8302227893612684e-05, 'epoch': 0.97}) +[2026-06-15 23:40:06,160][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.050058942288160324, 'train/grad_norm_pre_clip': 15.611456871032715, 'timing/data': 0.02406877197790891, 'timing/model': 1.141170485992916, 'learning_rate/qwen_vl_interface': 5.8272000548549776e-06, 'learning_rate/action_model': 5.827200054854978e-05, 'epoch': 0.97}) +[2026-06-15 23:40:10,607][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.06137991324067116, 'train/grad_norm_pre_clip': 26.49917221069336, 'timing/data': 0.023390455986373127, 'timing/model': 1.1343392200069502, 'learning_rate/qwen_vl_interface': 5.824177083083547e-06, 'learning_rate/action_model': 5.8241770830835475e-05, 'epoch': 0.97}) +[2026-06-15 23:40:15,113][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.020874083042144775, 'train/grad_norm_pre_clip': 6.393665790557861, 'timing/data': 0.023406739986967295, 'timing/model': 1.1443762929993682, 'learning_rate/qwen_vl_interface': 5.8211538752896056e-06, 'learning_rate/action_model': 5.821153875289606e-05, 'epoch': 0.97}) +[2026-06-15 23:40:19,617][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.16570840775966644, 'train/grad_norm_pre_clip': 18.254491806030273, 'timing/data': 0.02346864197170362, 'timing/model': 1.2505589820211753, 'learning_rate/qwen_vl_interface': 5.8181304327158814e-06, 'learning_rate/action_model': 5.8181304327158814e-05, 'epoch': 0.97}) +[2026-06-15 23:40:24,006][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.07789827138185501, 'train/grad_norm_pre_clip': 11.615781784057617, 'timing/data': 0.023667750996537507, 'timing/model': 1.138003917003516, 'learning_rate/qwen_vl_interface': 5.8151067566051965e-06, 'learning_rate/action_model': 5.815106756605196e-05, 'epoch': 0.97}) +[2026-06-15 23:40:28,575][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.12299734354019165, 'train/grad_norm_pre_clip': 29.549135208129883, 'timing/data': 0.022920624003745615, 'timing/model': 1.1419626259594224, 'learning_rate/qwen_vl_interface': 5.81208284820047e-06, 'learning_rate/action_model': 5.81208284820047e-05, 'epoch': 0.97}) +[2026-06-15 23:40:33,142][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.08980941027402878, 'train/grad_norm_pre_clip': 13.786028861999512, 'timing/data': 0.023151003988459706, 'timing/model': 1.1390183279872872, 'learning_rate/qwen_vl_interface': 5.809058708744716e-06, 'learning_rate/action_model': 5.809058708744716e-05, 'epoch': 0.97}) +[2026-06-15 23:40:37,699][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.014843693003058434, 'train/grad_norm_pre_clip': 6.064180850982666, 'timing/data': 0.022721693967469037, 'timing/model': 1.13795977801783, 'learning_rate/qwen_vl_interface': 5.8060343394810445e-06, 'learning_rate/action_model': 5.806034339481045e-05, 'epoch': 0.97}) +[2026-06-15 23:40:42,250][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.04561843350529671, 'train/grad_norm_pre_clip': 11.06486988067627, 'timing/data': 0.022306923056021333, 'timing/model': 1.3010972799966112, 'learning_rate/qwen_vl_interface': 5.803009741652659e-06, 'learning_rate/action_model': 5.8030097416526585e-05, 'epoch': 0.97}) +[2026-06-15 23:40:46,645][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.019376255571842194, 'train/grad_norm_pre_clip': 6.216534614562988, 'timing/data': 0.022891152009833604, 'timing/model': 1.1439393240143545, 'learning_rate/qwen_vl_interface': 5.799984916502859e-06, 'learning_rate/action_model': 5.799984916502859e-05, 'epoch': 0.97}) +[2026-06-15 23:40:51,197][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.08486387133598328, 'train/grad_norm_pre_clip': 15.250240325927734, 'timing/data': 0.023138020012993366, 'timing/model': 1.1364806290366687, 'learning_rate/qwen_vl_interface': 5.7969598652750334e-06, 'learning_rate/action_model': 5.796959865275033e-05, 'epoch': 0.97}) +[2026-06-15 23:40:55,700][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.1203438937664032, 'train/grad_norm_pre_clip': 23.93699836730957, 'timing/data': 0.023039522988256067, 'timing/model': 1.1432825200026855, 'learning_rate/qwen_vl_interface': 5.793934589212666e-06, 'learning_rate/action_model': 5.793934589212666e-05, 'epoch': 0.97}) +[2026-06-15 23:41:00,255][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.13837306201457977, 'train/grad_norm_pre_clip': 12.340974807739258, 'timing/data': 0.02324561995919794, 'timing/model': 1.1442504830192775, 'learning_rate/qwen_vl_interface': 5.7909090895593375e-06, 'learning_rate/action_model': 5.790909089559338e-05, 'epoch': 0.97}) +[2026-06-15 23:41:04,820][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.018067607656121254, 'train/grad_norm_pre_clip': 8.434969902038574, 'timing/data': 0.023079556005541235, 'timing/model': 1.143931464990601, 'learning_rate/qwen_vl_interface': 5.7878833675587135e-06, 'learning_rate/action_model': 5.787883367558713e-05, 'epoch': 0.97}) +[2026-06-15 23:41:09,407][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.15210406482219696, 'train/grad_norm_pre_clip': 20.627452850341797, 'timing/data': 0.028800881002098322, 'timing/model': 1.312026061990764, 'learning_rate/qwen_vl_interface': 5.784857424454556e-06, 'learning_rate/action_model': 5.7848574244545563e-05, 'epoch': 0.98}) +[2026-06-15 23:41:13,931][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.14836440980434418, 'train/grad_norm_pre_clip': 9.763689994812012, 'timing/data': 0.02299634099472314, 'timing/model': 1.2616315959603526, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 0.98}) +[2026-06-15 23:41:18,332][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.005503412336111069, 'train/grad_norm_pre_clip': 0.5360779166221619, 'timing/data': 0.023666154011152685, 'timing/model': 1.1459791980450973, 'learning_rate/qwen_vl_interface': 5.778804879911129e-06, 'learning_rate/action_model': 5.778804879911128e-05, 'epoch': 0.98}) +[2026-06-15 23:41:22,878][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.07541368901729584, 'train/grad_norm_pre_clip': 12.426167488098145, 'timing/data': 0.023311040014959872, 'timing/model': 1.134782116045244, 'learning_rate/qwen_vl_interface': 5.775778280959832e-06, 'learning_rate/action_model': 5.7757782809598315e-05, 'epoch': 0.98}) +[2026-06-15 23:41:27,423][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.0016388498479500413, 'train/grad_norm_pre_clip': 0.22251565754413605, 'timing/data': 0.022978244989644736, 'timing/model': 1.1347144360188395, 'learning_rate/qwen_vl_interface': 5.7727514658809435e-06, 'learning_rate/action_model': 5.772751465880944e-05, 'epoch': 0.98}) +[2026-06-15 23:41:31,977][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.09335104376077652, 'train/grad_norm_pre_clip': 7.592660427093506, 'timing/data': 0.0229016300290823, 'timing/model': 1.1361991919693537, 'learning_rate/qwen_vl_interface': 5.769724435918676e-06, 'learning_rate/action_model': 5.7697244359186755e-05, 'epoch': 0.98}) +[2026-06-15 23:41:36,519][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.1416958123445511, 'train/grad_norm_pre_clip': 15.581414222717285, 'timing/data': 0.023112549970392138, 'timing/model': 1.2910060100257397, 'learning_rate/qwen_vl_interface': 5.766697192317323e-06, 'learning_rate/action_model': 5.766697192317323e-05, 'epoch': 0.98}) +[2026-06-15 23:41:41,056][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.3892443776130676, 'train/grad_norm_pre_clip': 16.8868350982666, 'timing/data': 0.022242504986934364, 'timing/model': 1.2825837390264496, 'learning_rate/qwen_vl_interface': 5.763669736321271e-06, 'learning_rate/action_model': 5.763669736321271e-05, 'epoch': 0.98}) +[2026-06-15 23:41:45,460][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.14075133204460144, 'train/grad_norm_pre_clip': 6.851310729980469, 'timing/data': 0.022814778028987348, 'timing/model': 1.1451068019960076, 'learning_rate/qwen_vl_interface': 5.760642069174994e-06, 'learning_rate/action_model': 5.760642069174994e-05, 'epoch': 0.98}) +[2026-06-15 23:41:50,025][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.19707337021827698, 'train/grad_norm_pre_clip': 36.274559020996094, 'timing/data': 0.02771332097472623, 'timing/model': 1.1419233019696549, 'learning_rate/qwen_vl_interface': 5.757614192123051e-06, 'learning_rate/action_model': 5.757614192123051e-05, 'epoch': 0.98}) +[2026-06-15 23:41:54,593][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.27800238132476807, 'train/grad_norm_pre_clip': 17.74056625366211, 'timing/data': 0.023284460010472685, 'timing/model': 1.1339480030001141, 'learning_rate/qwen_vl_interface': 5.754586106410089e-06, 'learning_rate/action_model': 5.754586106410088e-05, 'epoch': 0.98}) +[2026-06-15 23:41:59,144][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.009776366874575615, 'train/grad_norm_pre_clip': 0.9791107177734375, 'timing/data': 0.023949114023707807, 'timing/model': 1.1346375870052725, 'learning_rate/qwen_vl_interface': 5.751557813280838e-06, 'learning_rate/action_model': 5.751557813280838e-05, 'epoch': 0.98}) +[2026-06-15 23:42:03,649][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.02126159518957138, 'train/grad_norm_pre_clip': 11.094640731811523, 'timing/data': 0.0230659669614397, 'timing/model': 1.1364353009848855, 'learning_rate/qwen_vl_interface': 5.7485293139801146e-06, 'learning_rate/action_model': 5.7485293139801146e-05, 'epoch': 0.98}) +[2026-06-15 23:42:08,216][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.07429356873035431, 'train/grad_norm_pre_clip': 10.880043983459473, 'timing/data': 0.02237465197686106, 'timing/model': 1.2968742569792084, 'learning_rate/qwen_vl_interface': 5.745500609752821e-06, 'learning_rate/action_model': 5.745500609752821e-05, 'epoch': 0.98}) +[2026-06-15 23:42:12,610][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.08745016902685165, 'train/grad_norm_pre_clip': 5.168925762176514, 'timing/data': 0.02353991096606478, 'timing/model': 1.1344105540192686, 'learning_rate/qwen_vl_interface': 5.742471701843943e-06, 'learning_rate/action_model': 5.742471701843943e-05, 'epoch': 0.98}) +[2026-06-15 23:42:17,172][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.06028006225824356, 'train/grad_norm_pre_clip': 10.259284019470215, 'timing/data': 0.023505781951826066, 'timing/model': 1.1453728380147368, 'learning_rate/qwen_vl_interface': 5.739442591498553e-06, 'learning_rate/action_model': 5.739442591498553e-05, 'epoch': 0.98}) +[2026-06-15 23:42:21,644][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.02936587855219841, 'train/grad_norm_pre_clip': 4.909391403198242, 'timing/data': 0.022838260978460312, 'timing/model': 1.1463318889727816, 'learning_rate/qwen_vl_interface': 5.7364132799617985e-06, 'learning_rate/action_model': 5.7364132799617984e-05, 'epoch': 0.98}) +[2026-06-15 23:42:26,110][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.003656200598925352, 'train/grad_norm_pre_clip': 0.4075658321380615, 'timing/data': 0.02306747797410935, 'timing/model': 1.1386652949731797, 'learning_rate/qwen_vl_interface': 5.733383768478923e-06, 'learning_rate/action_model': 5.7333837684789225e-05, 'epoch': 0.98}) +[2026-06-15 23:42:30,599][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.14715301990509033, 'train/grad_norm_pre_clip': 6.942436695098877, 'timing/data': 0.027912198973353952, 'timing/model': 1.1456179430242628, 'learning_rate/qwen_vl_interface': 5.730354058295234e-06, 'learning_rate/action_model': 5.730354058295234e-05, 'epoch': 0.98}) +[2026-06-15 23:42:35,121][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.05976497381925583, 'train/grad_norm_pre_clip': 23.292871475219727, 'timing/data': 0.022554077033419162, 'timing/model': 1.267264914989937, 'learning_rate/qwen_vl_interface': 5.72732415065614e-06, 'learning_rate/action_model': 5.727324150656139e-05, 'epoch': 0.98}) +[2026-06-15 23:42:39,527][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.0240242388099432, 'train/grad_norm_pre_clip': 4.5422258377075195, 'timing/data': 0.022767793969251215, 'timing/model': 1.139987413014751, 'learning_rate/qwen_vl_interface': 5.724294046807115e-06, 'learning_rate/action_model': 5.724294046807115e-05, 'epoch': 0.98}) +[2026-06-15 23:42:44,085][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.011502165347337723, 'train/grad_norm_pre_clip': 0.9044094085693359, 'timing/data': 0.023026812996249646, 'timing/model': 1.139263178978581, 'learning_rate/qwen_vl_interface': 5.7212637479937234e-06, 'learning_rate/action_model': 5.721263747993724e-05, 'epoch': 0.98}) +[2026-06-15 23:42:48,637][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.0887564942240715, 'train/grad_norm_pre_clip': 11.328279495239258, 'timing/data': 0.02371571899857372, 'timing/model': 1.1395027780090459, 'learning_rate/qwen_vl_interface': 5.718233255461607e-06, 'learning_rate/action_model': 5.718233255461607e-05, 'epoch': 0.98}) +[2026-06-15 23:42:53,200][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.16153118014335632, 'train/grad_norm_pre_clip': 15.06108570098877, 'timing/data': 0.02303267299430445, 'timing/model': 1.1385070569813251, 'learning_rate/qwen_vl_interface': 5.715202570456486e-06, 'learning_rate/action_model': 5.715202570456486e-05, 'epoch': 0.98}) +[2026-06-15 23:42:57,742][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.024656934663653374, 'train/grad_norm_pre_clip': 5.422330856323242, 'timing/data': 0.02406737598357722, 'timing/model': 1.2923406069749035, 'learning_rate/qwen_vl_interface': 5.7121716942241615e-06, 'learning_rate/action_model': 5.712171694224161e-05, 'epoch': 0.99}) +[2026-06-15 23:43:02,160][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.003964494448155165, 'train/grad_norm_pre_clip': 0.5882805585861206, 'timing/data': 0.024117335968185216, 'timing/model': 1.1434485260397196, 'learning_rate/qwen_vl_interface': 5.709140628010509e-06, 'learning_rate/action_model': 5.7091406280105085e-05, 'epoch': 0.99}) +[2026-06-15 23:47:23,111][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.009693071246147156, 'train/grad_norm_pre_clip': 0.9951975345611572, 'eval/loss': 0.05440467223525047, 'eval/flap_precision': 0.9498480243161094, 'eval/flap_recall': 0.9645061728395061, 'eval/flap_f1': 0.9571209800918835, 'eval/flap_rate_model': 0.1827777777777778, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9571209800918835, 'timing/data': 0.022967166965827346, 'timing/model': 1.1338771359878592, 'learning_rate/qwen_vl_interface': 5.706109373061489e-06, 'learning_rate/action_model': 5.706109373061489e-05, 'epoch': 0.99}) +[2026-06-15 23:47:23,112][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2400_state +[2026-06-15 23:47:23,114][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:47:26,689][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2400_state/model.safetensors +[2026-06-15 23:47:36,679][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2400_state/optimizer.bin +[2026-06-15 23:47:36,680][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2400_state/sampler.bin +[2026-06-15 23:47:36,680][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2400_state/sampler_1.bin +[2026-06-15 23:47:36,681][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2400_state/random_states_0.pkl +[2026-06-15 23:47:36,728][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2400 +[2026-06-15 23:47:36,728][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:47:36,734][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:47:41,017][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.04357493668794632, 'train/grad_norm_pre_clip': 16.287967681884766, 'timing/data': 0.02591945300810039, 'timing/model': 1.111855152004864, 'learning_rate/qwen_vl_interface': 5.703077930623134e-06, 'learning_rate/action_model': 5.7030779306231344e-05, 'epoch': 0.99}) +[2026-06-15 23:47:45,478][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.0032246229238808155, 'train/grad_norm_pre_clip': 0.28294357657432556, 'timing/data': 0.023824995034374297, 'timing/model': 1.1208971380256116, 'learning_rate/qwen_vl_interface': 5.7000463019415565e-06, 'learning_rate/action_model': 5.7000463019415563e-05, 'epoch': 0.99}) +[2026-06-15 23:47:49,914][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.0043024541810154915, 'train/grad_norm_pre_clip': 0.7830502390861511, 'timing/data': 0.02361880597891286, 'timing/model': 1.1247707729926333, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 0.99}) +[2026-06-15 23:47:54,392][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.1303020715713501, 'train/grad_norm_pre_clip': 27.030197143554688, 'timing/data': 0.02353629795834422, 'timing/model': 1.115181760978885, 'learning_rate/qwen_vl_interface': 5.693982490833561e-06, 'learning_rate/action_model': 5.693982490833561e-05, 'epoch': 0.99}) +[2026-06-15 23:47:58,874][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.0047080544754862785, 'train/grad_norm_pre_clip': 0.9879904985427856, 'timing/data': 0.023804931959602982, 'timing/model': 1.121755508997012, 'learning_rate/qwen_vl_interface': 5.690950310899745e-06, 'learning_rate/action_model': 5.6909503108997456e-05, 'epoch': 0.99}) +[2026-06-15 23:48:03,372][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.32051533460617065, 'train/grad_norm_pre_clip': 13.324023246765137, 'timing/data': 0.024010221008211374, 'timing/model': 1.2775889949989505, 'learning_rate/qwen_vl_interface': 5.687917949707913e-06, 'learning_rate/action_model': 5.687917949707913e-05, 'epoch': 0.99}) +[2026-06-15 23:48:07,777][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.00019878963939845562, 'train/grad_norm_pre_clip': 0.051830749958753586, 'timing/data': 0.023301731038372964, 'timing/model': 1.1918194580357522, 'learning_rate/qwen_vl_interface': 5.684885408504555e-06, 'learning_rate/action_model': 5.684885408504554e-05, 'epoch': 0.99}) +[2026-06-15 23:48:12,131][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.11039009690284729, 'train/grad_norm_pre_clip': 21.88152313232422, 'timing/data': 0.02356269903248176, 'timing/model': 1.1270430940203369, 'learning_rate/qwen_vl_interface': 5.681852688536231e-06, 'learning_rate/action_model': 5.681852688536231e-05, 'epoch': 0.99}) +[2026-06-15 23:48:16,648][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.04884162172675133, 'train/grad_norm_pre_clip': 22.49095344543457, 'timing/data': 0.02380897494731471, 'timing/model': 1.123350131965708, 'learning_rate/qwen_vl_interface': 5.67881979104958e-06, 'learning_rate/action_model': 5.67881979104958e-05, 'epoch': 0.99}) +[2026-06-15 23:48:21,162][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.17262712121009827, 'train/grad_norm_pre_clip': 13.814106941223145, 'timing/data': 0.023350834962911904, 'timing/model': 1.1219601080520079, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 0.99}) +[2026-06-15 23:48:25,678][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.0008704443462193012, 'train/grad_norm_pre_clip': 0.19086945056915283, 'timing/data': 0.0234384469804354, 'timing/model': 1.1328644679742865, 'learning_rate/qwen_vl_interface': 5.672753468508204e-06, 'learning_rate/action_model': 5.672753468508204e-05, 'epoch': 0.99}) +[2026-06-15 23:48:30,187][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.1258077472448349, 'train/grad_norm_pre_clip': 32.47087478637695, 'timing/data': 0.023964768974110484, 'timing/model': 1.2857031890307553, 'learning_rate/qwen_vl_interface': 5.6697200459471165e-06, 'learning_rate/action_model': 5.669720045947117e-05, 'epoch': 0.99}) +[2026-06-15 23:48:34,605][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.003926022443920374, 'train/grad_norm_pre_clip': 1.7552679777145386, 'timing/data': 0.02395469998009503, 'timing/model': 1.185992548998911, 'learning_rate/qwen_vl_interface': 5.666686450854973e-06, 'learning_rate/action_model': 5.666686450854973e-05, 'epoch': 0.99}) +[2026-06-15 23:48:38,950][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.2648080885410309, 'train/grad_norm_pre_clip': 12.146312713623047, 'timing/data': 0.02349755697650835, 'timing/model': 1.1233982630074024, 'learning_rate/qwen_vl_interface': 5.663652684478768e-06, 'learning_rate/action_model': 5.6636526844787676e-05, 'epoch': 0.99}) +[2026-06-15 23:48:43,401][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.12366288900375366, 'train/grad_norm_pre_clip': 13.425471305847168, 'timing/data': 0.022812429000623524, 'timing/model': 1.125085621024482, 'learning_rate/qwen_vl_interface': 5.660618748065569e-06, 'learning_rate/action_model': 5.6606187480655695e-05, 'epoch': 0.99}) +[2026-06-15 23:48:47,919][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.003964606206864119, 'train/grad_norm_pre_clip': 0.7828192710876465, 'timing/data': 0.02294988400535658, 'timing/model': 1.1293562049977481, 'learning_rate/qwen_vl_interface': 5.657584642862514e-06, 'learning_rate/action_model': 5.657584642862513e-05, 'epoch': 0.99}) +[2026-06-15 23:48:52,455][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.2143288105726242, 'train/grad_norm_pre_clip': 36.417945861816406, 'timing/data': 0.0233357289689593, 'timing/model': 1.132807378016878, 'learning_rate/qwen_vl_interface': 5.654550370116808e-06, 'learning_rate/action_model': 5.654550370116808e-05, 'epoch': 0.99}) +[2026-06-15 23:48:56,974][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.04295121133327484, 'train/grad_norm_pre_clip': 45.70661544799805, 'timing/data': 0.0228742029867135, 'timing/model': 1.2856581120286137, 'learning_rate/qwen_vl_interface': 5.651515931075726e-06, 'learning_rate/action_model': 5.651515931075726e-05, 'epoch': 0.99}) +[2026-06-15 23:49:01,476][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.01238316111266613, 'train/grad_norm_pre_clip': 6.191591739654541, 'timing/data': 0.022488599992357194, 'timing/model': 1.251487257948611, 'learning_rate/qwen_vl_interface': 5.6484813269866126e-06, 'learning_rate/action_model': 5.648481326986612e-05, 'epoch': 0.99}) +[2026-06-15 23:49:05,839][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.05780129134654999, 'train/grad_norm_pre_clip': 23.79781150817871, 'timing/data': 0.02324692899128422, 'timing/model': 1.1266688739997335, 'learning_rate/qwen_vl_interface': 5.645446559096875e-06, 'learning_rate/action_model': 5.645446559096875e-05, 'epoch': 0.99}) +[2026-06-15 23:49:10,355][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.1262110471725464, 'train/grad_norm_pre_clip': 9.626283645629883, 'timing/data': 0.02332483499776572, 'timing/model': 1.1284651400055736, 'learning_rate/qwen_vl_interface': 5.6424116286539965e-06, 'learning_rate/action_model': 5.642411628653996e-05, 'epoch': 0.99}) +[2026-06-15 23:49:14,884][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.006567258853465319, 'train/grad_norm_pre_clip': 2.036776065826416, 'timing/data': 0.023673278046771884, 'timing/model': 1.1314243450178765, 'learning_rate/qwen_vl_interface': 5.639376536905517e-06, 'learning_rate/action_model': 5.6393765369055175e-05, 'epoch': 0.99}) +[2026-06-15 23:49:19,401][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.2918506860733032, 'train/grad_norm_pre_clip': 47.00508117675781, 'timing/data': 0.023198666982352734, 'timing/model': 1.1271520649897866, 'learning_rate/qwen_vl_interface': 5.636341285099055e-06, 'learning_rate/action_model': 5.636341285099055e-05, 'epoch': 1.0}) +[2026-06-15 23:49:23,919][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.23529687523841858, 'train/grad_norm_pre_clip': 25.608505249023438, 'timing/data': 0.023407578002661467, 'timing/model': 1.2826450009597465, 'learning_rate/qwen_vl_interface': 5.6333058744822835e-06, 'learning_rate/action_model': 5.633305874482283e-05, 'epoch': 1.0}) +[2026-06-15 23:49:28,423][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.12642702460289001, 'train/grad_norm_pre_clip': 20.771167755126953, 'timing/data': 0.02336702699540183, 'timing/model': 1.2687808769987896, 'learning_rate/qwen_vl_interface': 5.630270306302944e-06, 'learning_rate/action_model': 5.6302703063029436e-05, 'epoch': 1.0}) +[2026-06-15 23:49:32,789][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.2223379909992218, 'train/grad_norm_pre_clip': 20.492847442626953, 'timing/data': 0.02292408497305587, 'timing/model': 1.1313656370039098, 'learning_rate/qwen_vl_interface': 5.627234581808848e-06, 'learning_rate/action_model': 5.627234581808848e-05, 'epoch': 1.0}) +[2026-06-15 23:49:37,320][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.09708501398563385, 'train/grad_norm_pre_clip': 11.954961776733398, 'timing/data': 0.023653423006180674, 'timing/model': 1.1334089940064587, 'learning_rate/qwen_vl_interface': 5.624198702247863e-06, 'learning_rate/action_model': 5.624198702247863e-05, 'epoch': 1.0}) +[2026-06-15 23:49:41,847][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.008402133360505104, 'train/grad_norm_pre_clip': 2.266812562942505, 'timing/data': 0.023391566006466746, 'timing/model': 1.1344018729869276, 'learning_rate/qwen_vl_interface': 5.621162668867925e-06, 'learning_rate/action_model': 5.621162668867925e-05, 'epoch': 1.0}) +[2026-06-15 23:49:46,379][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.0443585067987442, 'train/grad_norm_pre_clip': 19.397541046142578, 'timing/data': 0.023548586992546916, 'timing/model': 1.1344158470164984, 'learning_rate/qwen_vl_interface': 5.618126482917037e-06, 'learning_rate/action_model': 5.618126482917036e-05, 'epoch': 1.0}) +[2026-06-15 23:49:50,858][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.10129164159297943, 'train/grad_norm_pre_clip': 10.355640411376953, 'timing/data': 0.02340414101490751, 'timing/model': 1.134439248999115, 'learning_rate/qwen_vl_interface': 5.6150901456432546e-06, 'learning_rate/action_model': 5.615090145643255e-05, 'epoch': 1.0}) +[2026-06-15 23:49:55,387][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.12300274521112442, 'train/grad_norm_pre_clip': 14.033060073852539, 'timing/data': 0.023998624994419515, 'timing/model': 1.2908650080207735, 'learning_rate/qwen_vl_interface': 5.612053658294705e-06, 'learning_rate/action_model': 5.6120536582947045e-05, 'epoch': 1.0}) +[2026-06-15 23:49:59,765][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.21504409611225128, 'train/grad_norm_pre_clip': 22.8275203704834, 'timing/data': 0.023447197978384793, 'timing/model': 1.1351782000274397, 'learning_rate/qwen_vl_interface': 5.609017022119572e-06, 'learning_rate/action_model': 5.609017022119573e-05, 'epoch': 1.0}) +[2026-06-15 23:50:04,199][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.03693749010562897, 'train/grad_norm_pre_clip': 21.162168502807617, 'timing/data': 0.021662131999619305, 'timing/model': 1.1185155580169521, 'learning_rate/qwen_vl_interface': 5.605980238366104e-06, 'learning_rate/action_model': 5.605980238366103e-05, 'epoch': 1.0}) +[2026-06-15 23:50:08,498][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.05227721855044365, 'train/grad_norm_pre_clip': 17.5083065032959, 'timing/data': 0.0066888039582408965, 'timing/model': 1.114569577970542, 'learning_rate/qwen_vl_interface': 5.602943308282606e-06, 'learning_rate/action_model': 5.602943308282605e-05, 'epoch': 1.0}) +[2026-06-15 23:50:13,809][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.18797768652439117, 'train/grad_norm_pre_clip': 10.824579238891602, 'timing/data': 0.001239929988514632, 'timing/model': 1.1599110539536923, 'learning_rate/qwen_vl_interface': 5.599906233117445e-06, 'learning_rate/action_model': 5.5999062331174446e-05, 'epoch': 1.0}) +[2026-06-15 23:50:18,341][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.000828164629638195, 'train/grad_norm_pre_clip': 0.23087280988693237, 'timing/data': 0.006832418031990528, 'timing/model': 1.2874897990259342, 'learning_rate/qwen_vl_interface': 5.596869014119048e-06, 'learning_rate/action_model': 5.596869014119048e-05, 'epoch': 1.0}) +[2026-06-15 23:50:22,721][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.35920250415802, 'train/grad_norm_pre_clip': 16.485271453857422, 'timing/data': 0.006305462040472776, 'timing/model': 1.1319662659661844, 'learning_rate/qwen_vl_interface': 5.593831652535902e-06, 'learning_rate/action_model': 5.593831652535902e-05, 'epoch': 1.0}) +[2026-06-15 23:50:27,262][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.19982415437698364, 'train/grad_norm_pre_clip': 18.138381958007812, 'timing/data': 0.006327101029455662, 'timing/model': 1.1348240899969824, 'learning_rate/qwen_vl_interface': 5.590794149616552e-06, 'learning_rate/action_model': 5.590794149616552e-05, 'epoch': 1.0}) +[2026-06-15 23:50:31,758][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.07799020409584045, 'train/grad_norm_pre_clip': 18.87177276611328, 'timing/data': 0.006230336031876504, 'timing/model': 1.1353078550309874, 'learning_rate/qwen_vl_interface': 5.587756506609601e-06, 'learning_rate/action_model': 5.5877565066096006e-05, 'epoch': 1.0}) +[2026-06-15 23:50:36,245][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.14552080631256104, 'train/grad_norm_pre_clip': 13.534950256347656, 'timing/data': 0.006395953008905053, 'timing/model': 1.1308317619841546, 'learning_rate/qwen_vl_interface': 5.584718724763707e-06, 'learning_rate/action_model': 5.584718724763707e-05, 'epoch': 1.0}) +[2026-06-15 23:50:40,748][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.1388024091720581, 'train/grad_norm_pre_clip': 22.614667892456055, 'timing/data': 0.006480359006673098, 'timing/model': 1.1323155169957317, 'learning_rate/qwen_vl_interface': 5.581680805327588e-06, 'learning_rate/action_model': 5.5816808053275885e-05, 'epoch': 1.0}) +[2026-06-15 23:50:45,291][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.20170265436172485, 'train/grad_norm_pre_clip': 16.722373962402344, 'timing/data': 0.006537588022183627, 'timing/model': 1.3022248809575103, 'learning_rate/qwen_vl_interface': 5.578642749550019e-06, 'learning_rate/action_model': 5.578642749550019e-05, 'epoch': 1.0}) +[2026-06-15 23:50:49,675][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.08184569329023361, 'train/grad_norm_pre_clip': 16.71547508239746, 'timing/data': 0.0070309670409187675, 'timing/model': 1.1327996419859119, 'learning_rate/qwen_vl_interface': 5.575604558679831e-06, 'learning_rate/action_model': 5.57560455867983e-05, 'epoch': 1.0}) +[2026-06-15 23:50:54,204][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.04651550576090813, 'train/grad_norm_pre_clip': 24.196212768554688, 'timing/data': 0.0069606429897248745, 'timing/model': 1.1293889500084333, 'learning_rate/qwen_vl_interface': 5.5725662339659085e-06, 'learning_rate/action_model': 5.572566233965908e-05, 'epoch': 1.0}) +[2026-06-15 23:50:58,733][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.022157320752739906, 'train/grad_norm_pre_clip': 15.485209465026855, 'timing/data': 0.006360771018080413, 'timing/model': 1.1341889390023425, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 1.0}) +[2026-06-15 23:51:03,248][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.13151557743549347, 'train/grad_norm_pre_clip': 12.299736976623535, 'timing/data': 0.006617437000386417, 'timing/model': 1.135603845003061, 'learning_rate/qwen_vl_interface': 5.566489188002671e-06, 'learning_rate/action_model': 5.56648918800267e-05, 'epoch': 1.0}) +[2026-06-15 23:51:07,818][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.01665423810482025, 'train/grad_norm_pre_clip': 5.862690448760986, 'timing/data': 0.006199197960086167, 'timing/model': 1.1273613610537723, 'learning_rate/qwen_vl_interface': 5.563450469251403e-06, 'learning_rate/action_model': 5.563450469251402e-05, 'epoch': 1.01}) +[2026-06-15 23:51:12,394][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.03258679434657097, 'train/grad_norm_pre_clip': 12.367818832397461, 'timing/data': 0.006352949945721775, 'timing/model': 1.319891421997454, 'learning_rate/qwen_vl_interface': 5.560411621652488e-06, 'learning_rate/action_model': 5.560411621652488e-05, 'epoch': 1.01}) +[2026-06-15 23:51:16,765][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.05417357385158539, 'train/grad_norm_pre_clip': 12.807442665100098, 'timing/data': 0.00697642401792109, 'timing/model': 1.1356931239715777, 'learning_rate/qwen_vl_interface': 5.557372646455081e-06, 'learning_rate/action_model': 5.55737264645508e-05, 'epoch': 1.01}) +[2026-06-15 23:51:21,294][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.2596655488014221, 'train/grad_norm_pre_clip': 13.409653663635254, 'timing/data': 0.006488826998975128, 'timing/model': 1.1314873259980232, 'learning_rate/qwen_vl_interface': 5.554333544908388e-06, 'learning_rate/action_model': 5.5543335449083875e-05, 'epoch': 1.01}) +[2026-06-15 23:51:25,757][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.00949257705360651, 'train/grad_norm_pre_clip': 3.9830117225646973, 'timing/data': 0.006232245999854058, 'timing/model': 1.131292921025306, 'learning_rate/qwen_vl_interface': 5.551294318261671e-06, 'learning_rate/action_model': 5.5512943182616705e-05, 'epoch': 1.01}) +[2026-06-15 23:51:30,258][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.15398000180721283, 'train/grad_norm_pre_clip': 11.58737564086914, 'timing/data': 0.006400963000487536, 'timing/model': 1.1395116680068895, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 1.01}) +[2026-06-15 23:51:34,784][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.13371779024600983, 'train/grad_norm_pre_clip': 14.09692668914795, 'timing/data': 0.006502615986391902, 'timing/model': 1.132177947962191, 'learning_rate/qwen_vl_interface': 5.545215494665458e-06, 'learning_rate/action_model': 5.545215494665458e-05, 'epoch': 1.01}) +[2026-06-15 23:51:39,313][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.023907411843538284, 'train/grad_norm_pre_clip': 5.578165531158447, 'timing/data': 0.006539169000461698, 'timing/model': 1.2860020440421067, 'learning_rate/qwen_vl_interface': 5.5421759002147376e-06, 'learning_rate/action_model': 5.542175900214738e-05, 'epoch': 1.01}) +[2026-06-15 23:51:43,710][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.03315158560872078, 'train/grad_norm_pre_clip': 8.204030990600586, 'timing/data': 0.006220625014975667, 'timing/model': 1.141783081984613, 'learning_rate/qwen_vl_interface': 5.53913618566154e-06, 'learning_rate/action_model': 5.53913618566154e-05, 'epoch': 1.01}) +[2026-06-15 23:51:48,244][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.03485870361328125, 'train/grad_norm_pre_clip': 12.086596488952637, 'timing/data': 0.006748037994839251, 'timing/model': 1.1371207620250061, 'learning_rate/qwen_vl_interface': 5.536096352255381e-06, 'learning_rate/action_model': 5.536096352255381e-05, 'epoch': 1.01}) +[2026-06-15 23:51:52,776][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.03749222680926323, 'train/grad_norm_pre_clip': 11.897477149963379, 'timing/data': 0.007534786011092365, 'timing/model': 1.1510299970395863, 'learning_rate/qwen_vl_interface': 5.533056401245817e-06, 'learning_rate/action_model': 5.5330564012458164e-05, 'epoch': 1.01}) +[2026-06-15 23:51:57,283][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.08410114794969559, 'train/grad_norm_pre_clip': 5.456294059753418, 'timing/data': 0.006288568954914808, 'timing/model': 1.1343560349778272, 'learning_rate/qwen_vl_interface': 5.530016333882456e-06, 'learning_rate/action_model': 5.5300163338824565e-05, 'epoch': 1.01}) +[2026-06-15 23:52:01,821][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.06344284117221832, 'train/grad_norm_pre_clip': 5.343366622924805, 'timing/data': 0.006224046985153109, 'timing/model': 1.131441708013881, 'learning_rate/qwen_vl_interface': 5.52697615141496e-06, 'learning_rate/action_model': 5.52697615141496e-05, 'epoch': 1.01}) +[2026-06-15 23:52:06,384][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.07145307958126068, 'train/grad_norm_pre_clip': 11.704876899719238, 'timing/data': 0.006351100979372859, 'timing/model': 1.3034923410159536, 'learning_rate/qwen_vl_interface': 5.52393585509303e-06, 'learning_rate/action_model': 5.52393585509303e-05, 'epoch': 1.01}) +[2026-06-15 23:52:10,781][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.021507686004042625, 'train/grad_norm_pre_clip': 7.984722137451172, 'timing/data': 0.006282267975620925, 'timing/model': 1.1411105020088144, 'learning_rate/qwen_vl_interface': 5.520895446166418e-06, 'learning_rate/action_model': 5.5208954461664175e-05, 'epoch': 1.01}) +[2026-06-15 23:52:15,314][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.020441511645913124, 'train/grad_norm_pre_clip': 4.0461626052856445, 'timing/data': 0.006263007991947234, 'timing/model': 1.1303296879632398, 'learning_rate/qwen_vl_interface': 5.517854925884919e-06, 'learning_rate/action_model': 5.517854925884919e-05, 'epoch': 1.01}) +[2026-06-15 23:52:19,792][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.1634521633386612, 'train/grad_norm_pre_clip': 11.583874702453613, 'timing/data': 0.0066158759873360395, 'timing/model': 1.1314868920017034, 'learning_rate/qwen_vl_interface': 5.514814295498379e-06, 'learning_rate/action_model': 5.514814295498378e-05, 'epoch': 1.01}) +[2026-06-15 23:52:24,355][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.13116727769374847, 'train/grad_norm_pre_clip': 4.197386264801025, 'timing/data': 0.006243917043320835, 'timing/model': 1.1349062930094078, 'learning_rate/qwen_vl_interface': 5.5117735562566846e-06, 'learning_rate/action_model': 5.5117735562566844e-05, 'epoch': 1.01}) +[2026-06-15 23:52:28,899][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.0953737273812294, 'train/grad_norm_pre_clip': 23.093223571777344, 'timing/data': 0.006289074022788554, 'timing/model': 1.137155877018813, 'learning_rate/qwen_vl_interface': 5.5087327094097685e-06, 'learning_rate/action_model': 5.508732709409769e-05, 'epoch': 1.01}) +[2026-06-15 23:52:33,440][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.13781599700450897, 'train/grad_norm_pre_clip': 12.227997779846191, 'timing/data': 0.006708805973175913, 'timing/model': 1.2866632720106281, 'learning_rate/qwen_vl_interface': 5.505691756207609e-06, 'learning_rate/action_model': 5.505691756207609e-05, 'epoch': 1.01}) +[2026-06-15 23:52:37,902][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.08652283251285553, 'train/grad_norm_pre_clip': 10.149944305419922, 'timing/data': 0.00650220102397725, 'timing/model': 1.2176574429613538, 'learning_rate/qwen_vl_interface': 5.502650697900228e-06, 'learning_rate/action_model': 5.502650697900228e-05, 'epoch': 1.01}) +[2026-06-15 23:52:42,279][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.1814032644033432, 'train/grad_norm_pre_clip': 11.715654373168945, 'timing/data': 0.006396298995241523, 'timing/model': 1.1354671780136414, 'learning_rate/qwen_vl_interface': 5.499609535737688e-06, 'learning_rate/action_model': 5.4996095357376875e-05, 'epoch': 1.01}) +[2026-06-15 23:52:46,778][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.023473918437957764, 'train/grad_norm_pre_clip': 2.5603013038635254, 'timing/data': 0.0066021980019286275, 'timing/model': 1.1331741819740273, 'learning_rate/qwen_vl_interface': 5.496568270970096e-06, 'learning_rate/action_model': 5.496568270970096e-05, 'epoch': 1.01}) +[2026-06-15 23:52:51,356][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.03146286681294441, 'train/grad_norm_pre_clip': 5.382939338684082, 'timing/data': 0.006434041017200798, 'timing/model': 1.1424969950458035, 'learning_rate/qwen_vl_interface': 5.4935269048476015e-06, 'learning_rate/action_model': 5.493526904847601e-05, 'epoch': 1.01}) +[2026-06-15 23:52:55,900][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.07933732867240906, 'train/grad_norm_pre_clip': 9.024307250976562, 'timing/data': 0.0065685000154189765, 'timing/model': 1.133036051003728, 'learning_rate/qwen_vl_interface': 5.490485438620397e-06, 'learning_rate/action_model': 5.4904854386203965e-05, 'epoch': 1.02}) +[2026-06-15 23:53:00,435][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.13603100180625916, 'train/grad_norm_pre_clip': 6.209803104400635, 'timing/data': 0.00674114003777504, 'timing/model': 1.2895711840246804, 'learning_rate/qwen_vl_interface': 5.487443873538712e-06, 'learning_rate/action_model': 5.4874438735387115e-05, 'epoch': 1.02}) +[2026-06-15 23:53:04,956][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.10365619510412216, 'train/grad_norm_pre_clip': 8.959375381469727, 'timing/data': 0.006230611994396895, 'timing/model': 1.2669247390003875, 'learning_rate/qwen_vl_interface': 5.484402210852819e-06, 'learning_rate/action_model': 5.484402210852818e-05, 'epoch': 1.02}) +[2026-06-15 23:53:09,345][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.1013462245464325, 'train/grad_norm_pre_clip': 13.214787483215332, 'timing/data': 0.006493855966255069, 'timing/model': 1.1391292799962685, 'learning_rate/qwen_vl_interface': 5.48136045181303e-06, 'learning_rate/action_model': 5.48136045181303e-05, 'epoch': 1.02}) +[2026-06-15 23:53:13,866][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.12430215626955032, 'train/grad_norm_pre_clip': 19.023475646972656, 'timing/data': 0.006530440005008131, 'timing/model': 1.1362066179863177, 'learning_rate/qwen_vl_interface': 5.478318597669699e-06, 'learning_rate/action_model': 5.4783185976696995e-05, 'epoch': 1.02}) +[2026-06-15 23:53:18,413][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.01518798153847456, 'train/grad_norm_pre_clip': 1.2108606100082397, 'timing/data': 0.006174829963129014, 'timing/model': 1.1342164349625818, 'learning_rate/qwen_vl_interface': 5.475276649673218e-06, 'learning_rate/action_model': 5.4752766496732174e-05, 'epoch': 1.02}) +[2026-06-15 23:53:22,957][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.24373534321784973, 'train/grad_norm_pre_clip': 11.359806060791016, 'timing/data': 0.00665081397164613, 'timing/model': 1.1346074359607883, 'learning_rate/qwen_vl_interface': 5.472234609074014e-06, 'learning_rate/action_model': 5.472234609074014e-05, 'epoch': 1.02}) +[2026-06-15 23:53:27,505][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.02767614834010601, 'train/grad_norm_pre_clip': 4.837761878967285, 'timing/data': 0.006415980984456837, 'timing/model': 1.2949683099868707, 'learning_rate/qwen_vl_interface': 5.46919247712256e-06, 'learning_rate/action_model': 5.46919247712256e-05, 'epoch': 1.02}) +[2026-06-15 23:53:31,876][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.05724184587597847, 'train/grad_norm_pre_clip': 4.605615139007568, 'timing/data': 0.006744565034750849, 'timing/model': 1.1323924290481955, 'learning_rate/qwen_vl_interface': 5.466150255069356e-06, 'learning_rate/action_model': 5.4661502550693555e-05, 'epoch': 1.02}) +[2026-06-15 23:53:36,439][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.03186386823654175, 'train/grad_norm_pre_clip': 2.9200942516326904, 'timing/data': 0.0063639209838584065, 'timing/model': 1.1416597610223107, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 1.02}) +[2026-06-15 23:53:40,987][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.010596511885523796, 'train/grad_norm_pre_clip': 2.478790521621704, 'timing/data': 0.00627375696785748, 'timing/model': 1.1382625609985553, 'learning_rate/qwen_vl_interface': 5.46006554565991e-06, 'learning_rate/action_model': 5.4600655456599105e-05, 'epoch': 1.02}) +[2026-06-15 23:53:45,561][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.08288770169019699, 'train/grad_norm_pre_clip': 7.504626274108887, 'timing/data': 0.0061852060025557876, 'timing/model': 1.1431841349694878, 'learning_rate/qwen_vl_interface': 5.457023060804865e-06, 'learning_rate/action_model': 5.457023060804864e-05, 'epoch': 1.02}) +[2026-06-15 23:53:50,063][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.049442145973443985, 'train/grad_norm_pre_clip': 7.944342613220215, 'timing/data': 0.0067156360018998384, 'timing/model': 1.1370300740236416, 'learning_rate/qwen_vl_interface': 5.453980490850459e-06, 'learning_rate/action_model': 5.453980490850459e-05, 'epoch': 1.02}) +[2026-06-15 23:53:54,565][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.11880331486463547, 'train/grad_norm_pre_clip': 22.53339958190918, 'timing/data': 0.006131678004749119, 'timing/model': 1.242710991005879, 'learning_rate/qwen_vl_interface': 5.4509378370473785e-06, 'learning_rate/action_model': 5.450937837047378e-05, 'epoch': 1.02}) +[2026-06-15 23:53:58,960][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.20332475006580353, 'train/grad_norm_pre_clip': 13.035693168640137, 'timing/data': 0.006701864011120051, 'timing/model': 1.1363533000112511, 'learning_rate/qwen_vl_interface': 5.4478951006463424e-06, 'learning_rate/action_model': 5.447895100646342e-05, 'epoch': 1.02}) +[2026-06-15 23:54:03,502][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.18440191447734833, 'train/grad_norm_pre_clip': 15.57751750946045, 'timing/data': 0.00650843494804576, 'timing/model': 1.1343131539761089, 'learning_rate/qwen_vl_interface': 5.444852282898106e-06, 'learning_rate/action_model': 5.444852282898106e-05, 'epoch': 1.02}) +[2026-06-15 23:54:08,053][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.0535355880856514, 'train/grad_norm_pre_clip': 24.431516647338867, 'timing/data': 0.006644286040682346, 'timing/model': 1.1388109190156683, 'learning_rate/qwen_vl_interface': 5.441809385053456e-06, 'learning_rate/action_model': 5.4418093850534555e-05, 'epoch': 1.02}) +[2026-06-15 23:54:12,589][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.10764729231595993, 'train/grad_norm_pre_clip': 9.299307823181152, 'timing/data': 0.006658258964307606, 'timing/model': 1.1317898590350524, 'learning_rate/qwen_vl_interface': 5.438766408363214e-06, 'learning_rate/action_model': 5.438766408363214e-05, 'epoch': 1.02}) +[2026-06-15 23:54:17,105][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.28676748275756836, 'train/grad_norm_pre_clip': 12.948212623596191, 'timing/data': 0.006294219987466931, 'timing/model': 1.138000706967432, 'learning_rate/qwen_vl_interface': 5.43572335407823e-06, 'learning_rate/action_model': 5.43572335407823e-05, 'epoch': 1.02}) +[2026-06-15 23:54:21,606][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.10368820279836655, 'train/grad_norm_pre_clip': 8.159525871276855, 'timing/data': 0.006120403995737433, 'timing/model': 1.254754204011988, 'learning_rate/qwen_vl_interface': 5.432680223449392e-06, 'learning_rate/action_model': 5.4326802234493924e-05, 'epoch': 1.02}) +[2026-06-15 23:54:26,076][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.06636353582143784, 'train/grad_norm_pre_clip': 12.047152519226074, 'timing/data': 0.007104454038199037, 'timing/model': 1.1482687009847723, 'learning_rate/qwen_vl_interface': 5.429637017727614e-06, 'learning_rate/action_model': 5.4296370177276144e-05, 'epoch': 1.02}) +[2026-06-15 23:54:30,650][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.06766092777252197, 'train/grad_norm_pre_clip': 9.787551879882812, 'timing/data': 0.006445222010370344, 'timing/model': 1.1413182130199857, 'learning_rate/qwen_vl_interface': 5.426593738163845e-06, 'learning_rate/action_model': 5.4265937381638446e-05, 'epoch': 1.02}) +[2026-06-15 23:54:35,202][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.13611315190792084, 'train/grad_norm_pre_clip': 13.69604206085205, 'timing/data': 0.006425593979656696, 'timing/model': 1.1412674960447475, 'learning_rate/qwen_vl_interface': 5.42355038600906e-06, 'learning_rate/action_model': 5.42355038600906e-05, 'epoch': 1.02}) +[2026-06-15 23:54:39,760][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.2173876017332077, 'train/grad_norm_pre_clip': 19.572547912597656, 'timing/data': 0.006380538980010897, 'timing/model': 1.1390337769989856, 'learning_rate/qwen_vl_interface': 5.420506962514267e-06, 'learning_rate/action_model': 5.420506962514267e-05, 'epoch': 1.02}) +[2026-06-15 23:54:44,329][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.06787101179361343, 'train/grad_norm_pre_clip': 5.66900634765625, 'timing/data': 0.0064266580156981945, 'timing/model': 1.305039405997377, 'learning_rate/qwen_vl_interface': 5.4174634689305015e-06, 'learning_rate/action_model': 5.417463468930501e-05, 'epoch': 1.02}) +[2026-06-15 23:54:48,735][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.11163090914487839, 'train/grad_norm_pre_clip': 7.604173183441162, 'timing/data': 0.006306022987700999, 'timing/model': 1.1523550450219773, 'learning_rate/qwen_vl_interface': 5.41441990650883e-06, 'learning_rate/action_model': 5.4144199065088295e-05, 'epoch': 1.03}) +[2026-06-15 23:54:53,294][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.2563001811504364, 'train/grad_norm_pre_clip': 9.71446704864502, 'timing/data': 0.006326404050923884, 'timing/model': 1.1370971239521168, 'learning_rate/qwen_vl_interface': 5.411376276500345e-06, 'learning_rate/action_model': 5.411376276500345e-05, 'epoch': 1.03}) +[2026-06-15 23:54:57,770][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.04175500199198723, 'train/grad_norm_pre_clip': 8.457502365112305, 'timing/data': 0.006528348021674901, 'timing/model': 1.1414237690041773, 'learning_rate/qwen_vl_interface': 5.408332580156168e-06, 'learning_rate/action_model': 5.408332580156168e-05, 'epoch': 1.03}) +[2026-06-15 23:55:02,268][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.023207856342196465, 'train/grad_norm_pre_clip': 5.336203575134277, 'timing/data': 0.006359336024615914, 'timing/model': 1.1403168480028398, 'learning_rate/qwen_vl_interface': 5.40528881872745e-06, 'learning_rate/action_model': 5.40528881872745e-05, 'epoch': 1.03}) +[2026-06-15 23:59:23,172][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.051454074680805206, 'train/grad_norm_pre_clip': 6.725720405578613, 'eval/loss': 0.04058695212006569, 'eval/flap_precision': 0.951734539969834, 'eval/flap_recall': 0.9737654320987654, 'eval/flap_f1': 0.9626239511823036, 'eval/flap_rate_model': 0.18416666666666667, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9626239511823036, 'timing/data': 0.006514703971333802, 'timing/model': 1.135613978956826, 'learning_rate/qwen_vl_interface': 5.402244993465364e-06, 'learning_rate/action_model': 5.402244993465364e-05, 'epoch': 1.03}) +[2026-06-15 23:59:23,172][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2500_state +[2026-06-15 23:59:23,175][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:59:26,756][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2500_state/model.safetensors +[2026-06-15 23:59:36,938][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2500_state/optimizer.bin +[2026-06-15 23:59:36,938][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2500_state/sampler.bin +[2026-06-15 23:59:36,939][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-15 23:59:36,939][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-15 23:59:36,981][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-15 23:59:36,981][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:59:36,987][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:59:41,458][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.1926475465297699, 'train/grad_norm_pre_clip': 10.089006423950195, 'timing/data': 0.006815451022703201, 'timing/model': 1.1228576470166445, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 1.03}) +[2026-06-15 23:59:45,918][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.035353753715753555, 'train/grad_norm_pre_clip': 6.080319881439209, 'timing/data': 0.0064655819442123175, 'timing/model': 1.2720622820197605, 'learning_rate/qwen_vl_interface': 5.396157156445918e-06, 'learning_rate/action_model': 5.396157156445918e-05, 'epoch': 1.03}) +[2026-06-15 23:59:50,284][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.08852125704288483, 'train/grad_norm_pre_clip': 7.866262912750244, 'timing/data': 0.006298841966781765, 'timing/model': 1.175719385035336, 'learning_rate/qwen_vl_interface': 5.39311314719104e-06, 'learning_rate/action_model': 5.39311314719104e-05, 'epoch': 1.03}) +[2026-06-15 23:59:54,613][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.19472472369670868, 'train/grad_norm_pre_clip': 12.997274398803711, 'timing/data': 0.006320038984995335, 'timing/model': 1.1325098419911228, 'learning_rate/qwen_vl_interface': 5.390069079107755e-06, 'learning_rate/action_model': 5.390069079107755e-05, 'epoch': 1.03}) +[2026-06-15 23:59:59,028][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.08700068295001984, 'train/grad_norm_pre_clip': 8.186882019042969, 'timing/data': 0.00622880196897313, 'timing/model': 1.1292076059617102, 'learning_rate/qwen_vl_interface': 5.387024953447361e-06, 'learning_rate/action_model': 5.387024953447361e-05, 'epoch': 1.03}) +[2026-06-16 00:00:03,471][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.10166390240192413, 'train/grad_norm_pre_clip': 4.525020599365234, 'timing/data': 0.006297339976299554, 'timing/model': 1.1243298499612138, 'learning_rate/qwen_vl_interface': 5.383980771461184e-06, 'learning_rate/action_model': 5.3839807714611846e-05, 'epoch': 1.03}) +[2026-06-16 00:00:07,930][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.021765802055597305, 'train/grad_norm_pre_clip': 2.0248782634735107, 'timing/data': 0.006626805989071727, 'timing/model': 1.1227184340241365, 'learning_rate/qwen_vl_interface': 5.380936534400572e-06, 'learning_rate/action_model': 5.3809365344005726e-05, 'epoch': 1.03}) +[2026-06-16 00:00:12,444][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.0488528236746788, 'train/grad_norm_pre_clip': 5.900049209594727, 'timing/data': 0.006481799995526671, 'timing/model': 1.286360933969263, 'learning_rate/qwen_vl_interface': 5.377892243516899e-06, 'learning_rate/action_model': 5.3778922435168985e-05, 'epoch': 1.03}) +[2026-06-16 00:00:16,797][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.03895113989710808, 'train/grad_norm_pre_clip': 5.6643452644348145, 'timing/data': 0.006273620994761586, 'timing/model': 1.1277289809659123, 'learning_rate/qwen_vl_interface': 5.3748479000615525e-06, 'learning_rate/action_model': 5.374847900061553e-05, 'epoch': 1.03}) +[2026-06-16 00:00:21,318][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.0816425159573555, 'train/grad_norm_pre_clip': 6.070176124572754, 'timing/data': 0.0064154009451158345, 'timing/model': 1.124956008978188, 'learning_rate/qwen_vl_interface': 5.371803505285954e-06, 'learning_rate/action_model': 5.371803505285954e-05, 'epoch': 1.03}) +[2026-06-16 00:00:25,839][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.08809453248977661, 'train/grad_norm_pre_clip': 6.732693195343018, 'timing/data': 0.006459579977672547, 'timing/model': 1.1265365009894595, 'learning_rate/qwen_vl_interface': 5.368759060441534e-06, 'learning_rate/action_model': 5.368759060441534e-05, 'epoch': 1.03}) +[2026-06-16 00:00:30,357][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.09006655961275101, 'train/grad_norm_pre_clip': 5.3209075927734375, 'timing/data': 0.006589149008505046, 'timing/model': 1.1276409959536977, 'learning_rate/qwen_vl_interface': 5.365714566779749e-06, 'learning_rate/action_model': 5.3657145667797484e-05, 'epoch': 1.03}) +[2026-06-16 00:00:34,851][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.15003816783428192, 'train/grad_norm_pre_clip': 23.09694480895996, 'timing/data': 0.006232280982658267, 'timing/model': 1.1258690769900568, 'learning_rate/qwen_vl_interface': 5.362670025552077e-06, 'learning_rate/action_model': 5.362670025552077e-05, 'epoch': 1.03}) +[2026-06-16 00:00:39,365][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.04982340708374977, 'train/grad_norm_pre_clip': 10.189936637878418, 'timing/data': 0.006241268012672663, 'timing/model': 1.2841699449927546, 'learning_rate/qwen_vl_interface': 5.359625438010013e-06, 'learning_rate/action_model': 5.359625438010013e-05, 'epoch': 1.03}) +[2026-06-16 00:00:43,723][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.041636016219854355, 'train/grad_norm_pre_clip': 5.184956073760986, 'timing/data': 0.006271042046137154, 'timing/model': 1.1271820570109412, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 1.03}) +[2026-06-16 00:00:48,240][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.23260101675987244, 'train/grad_norm_pre_clip': 9.896679878234863, 'timing/data': 0.00663102901307866, 'timing/model': 1.1281381759908982, 'learning_rate/qwen_vl_interface': 5.353536128988791e-06, 'learning_rate/action_model': 5.353536128988791e-05, 'epoch': 1.03}) +[2026-06-16 00:00:52,773][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.041575439274311066, 'train/grad_norm_pre_clip': 8.0551176071167, 'timing/data': 0.006849444005638361, 'timing/model': 1.1350524550070986, 'learning_rate/qwen_vl_interface': 5.350491410012717e-06, 'learning_rate/action_model': 5.350491410012717e-05, 'epoch': 1.03}) +[2026-06-16 00:00:57,285][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.1506132036447525, 'train/grad_norm_pre_clip': 17.20993995666504, 'timing/data': 0.006516457011457533, 'timing/model': 1.139453287993092, 'learning_rate/qwen_vl_interface': 5.347446649728419e-06, 'learning_rate/action_model': 5.347446649728418e-05, 'epoch': 1.03}) +[2026-06-16 00:01:01,839][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.030181560665369034, 'train/grad_norm_pre_clip': 2.9656057357788086, 'timing/data': 0.006084962980821729, 'timing/model': 1.131188456027303, 'learning_rate/qwen_vl_interface': 5.344401849387484e-06, 'learning_rate/action_model': 5.344401849387484e-05, 'epoch': 1.03}) +[2026-06-16 00:01:06,385][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.01174099836498499, 'train/grad_norm_pre_clip': 1.5083388090133667, 'timing/data': 0.0072972659836523235, 'timing/model': 1.316591325972695, 'learning_rate/qwen_vl_interface': 5.3413570102415144e-06, 'learning_rate/action_model': 5.3413570102415144e-05, 'epoch': 1.04}) +[2026-06-16 00:01:10,753][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.10146449506282806, 'train/grad_norm_pre_clip': 8.588522911071777, 'timing/data': 0.00635731196962297, 'timing/model': 1.1305387640022673, 'learning_rate/qwen_vl_interface': 5.338312133542131e-06, 'learning_rate/action_model': 5.3383121335421305e-05, 'epoch': 1.04}) +[2026-06-16 00:01:15,278][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.010494302958250046, 'train/grad_norm_pre_clip': 1.1130136251449585, 'timing/data': 0.006681501981802285, 'timing/model': 1.1301355810137466, 'learning_rate/qwen_vl_interface': 5.335267220540963e-06, 'learning_rate/action_model': 5.335267220540963e-05, 'epoch': 1.04}) +[2026-06-16 00:01:19,761][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.04284200817346573, 'train/grad_norm_pre_clip': 6.813480377197266, 'timing/data': 0.006702265003696084, 'timing/model': 1.1286396690411493, 'learning_rate/qwen_vl_interface': 5.332222272489662e-06, 'learning_rate/action_model': 5.332222272489662e-05, 'epoch': 1.04}) +[2026-06-16 00:01:24,256][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.008720606565475464, 'train/grad_norm_pre_clip': 0.9291229248046875, 'timing/data': 0.0064867770415730774, 'timing/model': 1.1341581910382956, 'learning_rate/qwen_vl_interface': 5.3291772906398895e-06, 'learning_rate/action_model': 5.32917729063989e-05, 'epoch': 1.04}) +[2026-06-16 00:01:28,786][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.17640335857868195, 'train/grad_norm_pre_clip': 11.619210243225098, 'timing/data': 0.006277317996136844, 'timing/model': 1.1318336700205691, 'learning_rate/qwen_vl_interface': 5.326132276243327e-06, 'learning_rate/action_model': 5.3261322762433266e-05, 'epoch': 1.04}) +[2026-06-16 00:01:33,343][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.42890432476997375, 'train/grad_norm_pre_clip': 15.79190444946289, 'timing/data': 0.00650187797145918, 'timing/model': 1.3032007130095735, 'learning_rate/qwen_vl_interface': 5.323087230551659e-06, 'learning_rate/action_model': 5.323087230551659e-05, 'epoch': 1.04}) +[2026-06-16 00:01:37,712][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.0794144868850708, 'train/grad_norm_pre_clip': 11.583218574523926, 'timing/data': 0.00622112798737362, 'timing/model': 1.1297828980023041, 'learning_rate/qwen_vl_interface': 5.3200421548165916e-06, 'learning_rate/action_model': 5.3200421548165914e-05, 'epoch': 1.04}) +[2026-06-16 00:01:42,255][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.003882231190800667, 'train/grad_norm_pre_clip': 0.5845962166786194, 'timing/data': 0.0064052880043163896, 'timing/model': 1.1358124010148458, 'learning_rate/qwen_vl_interface': 5.3169970502898405e-06, 'learning_rate/action_model': 5.31699705028984e-05, 'epoch': 1.04}) +[2026-06-16 00:01:46,757][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.125467911362648, 'train/grad_norm_pre_clip': 11.862129211425781, 'timing/data': 0.006479001021943986, 'timing/model': 1.13332788500702, 'learning_rate/qwen_vl_interface': 5.31395191822313e-06, 'learning_rate/action_model': 5.31395191822313e-05, 'epoch': 1.04}) +[2026-06-16 00:01:51,273][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.13186481595039368, 'train/grad_norm_pre_clip': 20.911958694458008, 'timing/data': 0.006526398006826639, 'timing/model': 1.136978346039541, 'learning_rate/qwen_vl_interface': 5.310906759868205e-06, 'learning_rate/action_model': 5.310906759868205e-05, 'epoch': 1.04}) +[2026-06-16 00:01:55,835][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.006197178270667791, 'train/grad_norm_pre_clip': 1.4497437477111816, 'timing/data': 0.006291993020568043, 'timing/model': 1.134138272027485, 'learning_rate/qwen_vl_interface': 5.30786157647681e-06, 'learning_rate/action_model': 5.30786157647681e-05, 'epoch': 1.04}) +[2026-06-16 00:02:00,395][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.10089996457099915, 'train/grad_norm_pre_clip': 17.106821060180664, 'timing/data': 0.0064498219871893525, 'timing/model': 1.3025879790075123, 'learning_rate/qwen_vl_interface': 5.304816369300709e-06, 'learning_rate/action_model': 5.304816369300709e-05, 'epoch': 1.04}) +[2026-06-16 00:02:04,778][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.2707703411579132, 'train/grad_norm_pre_clip': 34.67949676513672, 'timing/data': 0.006561786984093487, 'timing/model': 1.1351807989994995, 'learning_rate/qwen_vl_interface': 5.301771139591666e-06, 'learning_rate/action_model': 5.301771139591666e-05, 'epoch': 1.04}) +[2026-06-16 00:02:09,325][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.18725688755512238, 'train/grad_norm_pre_clip': 7.553483486175537, 'timing/data': 0.006732242007274181, 'timing/model': 1.1382777050021105, 'learning_rate/qwen_vl_interface': 5.298725888601465e-06, 'learning_rate/action_model': 5.298725888601465e-05, 'epoch': 1.04}) +[2026-06-16 00:02:13,820][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.24094638228416443, 'train/grad_norm_pre_clip': 11.946885108947754, 'timing/data': 0.006401039019692689, 'timing/model': 1.1327089400147088, 'learning_rate/qwen_vl_interface': 5.295680617581891e-06, 'learning_rate/action_model': 5.2956806175818904e-05, 'epoch': 1.04}) +[2026-06-16 00:02:18,364][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.04815920442342758, 'train/grad_norm_pre_clip': 14.063285827636719, 'timing/data': 0.006475204019807279, 'timing/model': 1.1371284130145796, 'learning_rate/qwen_vl_interface': 5.2926353277847395e-06, 'learning_rate/action_model': 5.292635327784739e-05, 'epoch': 1.04}) +[2026-06-16 00:02:22,914][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.055099260061979294, 'train/grad_norm_pre_clip': 9.631621360778809, 'timing/data': 0.006565709016285837, 'timing/model': 1.1369126819772646, 'learning_rate/qwen_vl_interface': 5.2895900204618175e-06, 'learning_rate/action_model': 5.289590020461818e-05, 'epoch': 1.04}) +[2026-06-16 00:02:27,458][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.2652856111526489, 'train/grad_norm_pre_clip': 21.00042724609375, 'timing/data': 0.006615661026444286, 'timing/model': 1.284112227964215, 'learning_rate/qwen_vl_interface': 5.286544696864933e-06, 'learning_rate/action_model': 5.2865446968649324e-05, 'epoch': 1.04}) +[2026-06-16 00:02:31,965][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.20233675837516785, 'train/grad_norm_pre_clip': 19.825708389282227, 'timing/data': 0.006042099033948034, 'timing/model': 1.242018733988516, 'learning_rate/qwen_vl_interface': 5.283499358245904e-06, 'learning_rate/action_model': 5.2834993582459034e-05, 'epoch': 1.04}) +[2026-06-16 00:02:36,342][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.014268695376813412, 'train/grad_norm_pre_clip': 2.130601406097412, 'timing/data': 0.0064297899953089654, 'timing/model': 1.1340792690170929, 'learning_rate/qwen_vl_interface': 5.280454005856553e-06, 'learning_rate/action_model': 5.2804540058565534e-05, 'epoch': 1.04}) +[2026-06-16 00:02:40,860][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.064141646027565, 'train/grad_norm_pre_clip': 8.816658020019531, 'timing/data': 0.0067083199974149466, 'timing/model': 1.1369294920004904, 'learning_rate/qwen_vl_interface': 5.277408640948711e-06, 'learning_rate/action_model': 5.277408640948711e-05, 'epoch': 1.04}) +[2026-06-16 00:02:45,404][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.09226895123720169, 'train/grad_norm_pre_clip': 8.22889232635498, 'timing/data': 0.006595269951503724, 'timing/model': 1.1365052930195816, 'learning_rate/qwen_vl_interface': 5.274363264774213e-06, 'learning_rate/action_model': 5.274363264774212e-05, 'epoch': 1.04}) +[2026-06-16 00:02:49,938][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.2574630677700043, 'train/grad_norm_pre_clip': 25.0605525970459, 'timing/data': 0.007218638027552515, 'timing/model': 1.1311468509957194, 'learning_rate/qwen_vl_interface': 5.271317878584897e-06, 'learning_rate/action_model': 5.2713178785848974e-05, 'epoch': 1.04}) +[2026-06-16 00:02:54,508][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.017824748530983925, 'train/grad_norm_pre_clip': 5.511233329772949, 'timing/data': 0.006240555027034134, 'timing/model': 1.3083497800398618, 'learning_rate/qwen_vl_interface': 5.268272483632606e-06, 'learning_rate/action_model': 5.268272483632606e-05, 'epoch': 1.05}) +[2026-06-16 00:02:59,017][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.006969565991312265, 'train/grad_norm_pre_clip': 0.6230394840240479, 'timing/data': 0.006639991013798863, 'timing/model': 1.2485039279563352, 'learning_rate/qwen_vl_interface': 5.265227081169187e-06, 'learning_rate/action_model': 5.265227081169186e-05, 'epoch': 1.05}) +[2026-06-16 00:03:03,406][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.33344393968582153, 'train/grad_norm_pre_clip': 29.320459365844727, 'timing/data': 0.006220049981493503, 'timing/model': 1.1363625879748724, 'learning_rate/qwen_vl_interface': 5.2621816724464895e-06, 'learning_rate/action_model': 5.262181672446489e-05, 'epoch': 1.05}) +[2026-06-16 00:03:07,929][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.11505171656608582, 'train/grad_norm_pre_clip': 16.066036224365234, 'timing/data': 0.00642153003718704, 'timing/model': 1.1411493790219538, 'learning_rate/qwen_vl_interface': 5.259136258716367e-06, 'learning_rate/action_model': 5.2591362587163675e-05, 'epoch': 1.05}) +[2026-06-16 00:03:12,474][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.2389850914478302, 'train/grad_norm_pre_clip': 8.523300170898438, 'timing/data': 0.006461694021709263, 'timing/model': 1.1348899789736606, 'learning_rate/qwen_vl_interface': 5.256090841230675e-06, 'learning_rate/action_model': 5.256090841230674e-05, 'epoch': 1.05}) +[2026-06-16 00:03:17,025][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.04270855709910393, 'train/grad_norm_pre_clip': 5.704343795776367, 'timing/data': 0.006274865998420864, 'timing/model': 1.1322887160349637, 'learning_rate/qwen_vl_interface': 5.253045421241266e-06, 'learning_rate/action_model': 5.253045421241266e-05, 'epoch': 1.05}) +[2026-06-16 00:03:21,574][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.2394612431526184, 'train/grad_norm_pre_clip': 14.428231239318848, 'timing/data': 0.006214646971784532, 'timing/model': 1.2936616140068509, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 1.05}) +[2026-06-16 00:03:25,960][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.017477279528975487, 'train/grad_norm_pre_clip': 2.5214602947235107, 'timing/data': 0.006777499977033585, 'timing/model': 1.134068302984815, 'learning_rate/qwen_vl_interface': 5.246954578758735e-06, 'learning_rate/action_model': 5.246954578758735e-05, 'epoch': 1.05}) +[2026-06-16 00:03:30,495][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.022890741005539894, 'train/grad_norm_pre_clip': 5.07313871383667, 'timing/data': 0.006462383957114071, 'timing/model': 1.1319668639916927, 'learning_rate/qwen_vl_interface': 5.243909158769327e-06, 'learning_rate/action_model': 5.2439091587693264e-05, 'epoch': 1.05}) +[2026-06-16 00:03:35,039][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.032010290771722794, 'train/grad_norm_pre_clip': 8.304929733276367, 'timing/data': 0.006647692993283272, 'timing/model': 1.137203906022478, 'learning_rate/qwen_vl_interface': 5.240863741283634e-06, 'learning_rate/action_model': 5.2408637412836336e-05, 'epoch': 1.05}) +[2026-06-16 00:03:39,586][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.1522231549024582, 'train/grad_norm_pre_clip': 6.961250305175781, 'timing/data': 0.006704377999994904, 'timing/model': 1.1432301700115204, 'learning_rate/qwen_vl_interface': 5.2378183275535125e-06, 'learning_rate/action_model': 5.237818327553512e-05, 'epoch': 1.05}) +[2026-06-16 00:03:44,052][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.0026452266611158848, 'train/grad_norm_pre_clip': 0.27719441056251526, 'timing/data': 0.006696280965115875, 'timing/model': 1.13500758795999, 'learning_rate/qwen_vl_interface': 5.234772918830815e-06, 'learning_rate/action_model': 5.234772918830815e-05, 'epoch': 1.05}) +[2026-06-16 00:03:48,543][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.1368132084608078, 'train/grad_norm_pre_clip': 15.022476196289062, 'timing/data': 0.006708341999910772, 'timing/model': 1.2379843299859203, 'learning_rate/qwen_vl_interface': 5.231727516367395e-06, 'learning_rate/action_model': 5.231727516367395e-05, 'epoch': 1.05}) +[2026-06-16 00:03:52,938][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.01809055544435978, 'train/grad_norm_pre_clip': 7.934577465057373, 'timing/data': 0.00618418597150594, 'timing/model': 1.1392574859783053, 'learning_rate/qwen_vl_interface': 5.228682121415104e-06, 'learning_rate/action_model': 5.228682121415104e-05, 'epoch': 1.05}) +[2026-06-16 00:03:57,463][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.16065222024917603, 'train/grad_norm_pre_clip': 50.6242561340332, 'timing/data': 0.00645208201603964, 'timing/model': 1.1291151079931296, 'learning_rate/qwen_vl_interface': 5.225636735225789e-06, 'learning_rate/action_model': 5.225636735225788e-05, 'epoch': 1.05}) +[2026-06-16 00:04:02,018][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.2046317309141159, 'train/grad_norm_pre_clip': 20.337947845458984, 'timing/data': 0.006564397015608847, 'timing/model': 1.144868009025231, 'learning_rate/qwen_vl_interface': 5.22259135905129e-06, 'learning_rate/action_model': 5.2225913590512895e-05, 'epoch': 1.05}) +[2026-06-16 00:04:06,575][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.15089993178844452, 'train/grad_norm_pre_clip': 6.493190288543701, 'timing/data': 0.0068702169810421765, 'timing/model': 1.1375362139660865, 'learning_rate/qwen_vl_interface': 5.219545994143448e-06, 'learning_rate/action_model': 5.219545994143448e-05, 'epoch': 1.05}) +[2026-06-16 00:04:11,119][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.23841921985149384, 'train/grad_norm_pre_clip': 9.09105110168457, 'timing/data': 0.006387548986822367, 'timing/model': 1.1419556970358826, 'learning_rate/qwen_vl_interface': 5.216500641754097e-06, 'learning_rate/action_model': 5.216500641754097e-05, 'epoch': 1.05}) +[2026-06-16 00:04:15,644][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.03044975735247135, 'train/grad_norm_pre_clip': 3.875924587249756, 'timing/data': 0.006531091989018023, 'timing/model': 1.2698556349496357, 'learning_rate/qwen_vl_interface': 5.213455303135067e-06, 'learning_rate/action_model': 5.2134553031350673e-05, 'epoch': 1.05}) +[2026-06-16 00:04:20,023][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.050044141709804535, 'train/grad_norm_pre_clip': 4.9989013671875, 'timing/data': 0.006324164045508951, 'timing/model': 1.1353614069521427, 'learning_rate/qwen_vl_interface': 5.2104099795381836e-06, 'learning_rate/action_model': 5.210409979538183e-05, 'epoch': 1.05}) +[2026-06-16 00:04:24,575][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.043392617255449295, 'train/grad_norm_pre_clip': 14.345934867858887, 'timing/data': 0.006295943981967866, 'timing/model': 1.1408752290299162, 'learning_rate/qwen_vl_interface': 5.207364672215261e-06, 'learning_rate/action_model': 5.2073646722152604e-05, 'epoch': 1.05}) +[2026-06-16 00:04:29,105][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.088715560734272, 'train/grad_norm_pre_clip': 7.272619247436523, 'timing/data': 0.006556606967933476, 'timing/model': 1.1323442629654892, 'learning_rate/qwen_vl_interface': 5.2043193824181095e-06, 'learning_rate/action_model': 5.2043193824181094e-05, 'epoch': 1.05}) +[2026-06-16 00:04:33,644][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.11559207737445831, 'train/grad_norm_pre_clip': 13.301051139831543, 'timing/data': 0.00640222302172333, 'timing/model': 1.1337043329840526, 'learning_rate/qwen_vl_interface': 5.201274111398535e-06, 'learning_rate/action_model': 5.2012741113985356e-05, 'epoch': 1.05}) +[2026-06-16 00:04:38,176][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.03215308114886284, 'train/grad_norm_pre_clip': 5.457200050354004, 'timing/data': 0.006434756040107459, 'timing/model': 1.2883246450219303, 'learning_rate/qwen_vl_interface': 5.198228860408334e-06, 'learning_rate/action_model': 5.198228860408334e-05, 'epoch': 1.05}) +[2026-06-16 00:04:42,559][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.25175803899765015, 'train/grad_norm_pre_clip': 10.02689266204834, 'timing/data': 0.006308550015091896, 'timing/model': 1.133310839009937, 'learning_rate/qwen_vl_interface': 5.195183630699292e-06, 'learning_rate/action_model': 5.195183630699292e-05, 'epoch': 1.05}) +[2026-06-16 00:04:47,104][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.20171362161636353, 'train/grad_norm_pre_clip': 22.696168899536133, 'timing/data': 0.0060203089960850775, 'timing/model': 1.1347543679876253, 'learning_rate/qwen_vl_interface': 5.192138423523189e-06, 'learning_rate/action_model': 5.1921384235231895e-05, 'epoch': 1.06}) +[2026-06-16 00:04:51,554][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.005846663843840361, 'train/grad_norm_pre_clip': 1.2885991334915161, 'timing/data': 0.0064733310136944056, 'timing/model': 1.1315565960248932, 'learning_rate/qwen_vl_interface': 5.189093240131795e-06, 'learning_rate/action_model': 5.189093240131795e-05, 'epoch': 1.06}) +[2026-06-16 00:04:56,022][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.04985145106911659, 'train/grad_norm_pre_clip': 7.96432638168335, 'timing/data': 0.00640133599517867, 'timing/model': 1.1339659460354596, 'learning_rate/qwen_vl_interface': 5.186048081776869e-06, 'learning_rate/action_model': 5.1860480817768685e-05, 'epoch': 1.06}) +[2026-06-16 00:05:00,514][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.01745876669883728, 'train/grad_norm_pre_clip': 2.1462976932525635, 'timing/data': 0.006606325972825289, 'timing/model': 1.1344761060317978, 'learning_rate/qwen_vl_interface': 5.183002949710161e-06, 'learning_rate/action_model': 5.18300294971016e-05, 'epoch': 1.06}) +[2026-06-16 00:05:05,063][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.16240139305591583, 'train/grad_norm_pre_clip': 11.085466384887695, 'timing/data': 0.006562129012309015, 'timing/model': 1.2945507389958948, 'learning_rate/qwen_vl_interface': 5.179957845183411e-06, 'learning_rate/action_model': 5.179957845183411e-05, 'epoch': 1.06}) +[2026-06-16 00:05:09,443][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.010203846730291843, 'train/grad_norm_pre_clip': 1.2956873178482056, 'timing/data': 0.006429218978155404, 'timing/model': 1.1353170960210264, 'learning_rate/qwen_vl_interface': 5.176912769448343e-06, 'learning_rate/action_model': 5.176912769448343e-05, 'epoch': 1.06}) +[2026-06-16 00:05:13,987][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.10763437300920486, 'train/grad_norm_pre_clip': 14.169042587280273, 'timing/data': 0.006774690991733223, 'timing/model': 1.1348619009950198, 'learning_rate/qwen_vl_interface': 5.173867723756676e-06, 'learning_rate/action_model': 5.173867723756676e-05, 'epoch': 1.06}) +[2026-06-16 00:05:18,528][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.3934713900089264, 'train/grad_norm_pre_clip': 14.775934219360352, 'timing/data': 0.006373537005856633, 'timing/model': 1.1307751119602472, 'learning_rate/qwen_vl_interface': 5.170822709360112e-06, 'learning_rate/action_model': 5.170822709360111e-05, 'epoch': 1.06}) +[2026-06-16 00:05:23,046][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.010688496753573418, 'train/grad_norm_pre_clip': 1.9442875385284424, 'timing/data': 0.006418815988581628, 'timing/model': 1.1367147179553285, 'learning_rate/qwen_vl_interface': 5.16777772751034e-06, 'learning_rate/action_model': 5.16777772751034e-05, 'epoch': 1.06}) +[2026-06-16 00:05:27,579][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.2161472588777542, 'train/grad_norm_pre_clip': 20.55095863342285, 'timing/data': 0.0064541620085947216, 'timing/model': 1.135286297998391, 'learning_rate/qwen_vl_interface': 5.16473277945904e-06, 'learning_rate/action_model': 5.1647327794590396e-05, 'epoch': 1.06}) +[2026-06-16 00:05:32,133][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.008456436917185783, 'train/grad_norm_pre_clip': 1.151699423789978, 'timing/data': 0.006295997009146959, 'timing/model': 1.3047200739965774, 'learning_rate/qwen_vl_interface': 5.161687866457872e-06, 'learning_rate/action_model': 5.161687866457871e-05, 'epoch': 1.06}) +[2026-06-16 00:05:36,517][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.008466904051601887, 'train/grad_norm_pre_clip': 0.804244875907898, 'timing/data': 0.006587939977180213, 'timing/model': 1.13482751295669, 'learning_rate/qwen_vl_interface': 5.158642989758487e-06, 'learning_rate/action_model': 5.158642989758487e-05, 'epoch': 1.06}) +[2026-06-16 00:05:41,051][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.040183912962675095, 'train/grad_norm_pre_clip': 9.937952041625977, 'timing/data': 0.006575019971933216, 'timing/model': 1.1328908579889685, 'learning_rate/qwen_vl_interface': 5.155598150612519e-06, 'learning_rate/action_model': 5.1555981506125185e-05, 'epoch': 1.06}) +[2026-06-16 00:05:45,587][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.07481873035430908, 'train/grad_norm_pre_clip': 7.060817241668701, 'timing/data': 0.006297920946963131, 'timing/model': 1.1323931149672717, 'learning_rate/qwen_vl_interface': 5.152553350271584e-06, 'learning_rate/action_model': 5.152553350271584e-05, 'epoch': 1.06}) +[2026-06-16 00:05:50,102][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.01696261577308178, 'train/grad_norm_pre_clip': 1.933495283126831, 'timing/data': 0.006398758967407048, 'timing/model': 1.1315841950126924, 'learning_rate/qwen_vl_interface': 5.149508589987286e-06, 'learning_rate/action_model': 5.1495085899872855e-05, 'epoch': 1.06}) +[2026-06-16 00:05:54,637][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.07923411577939987, 'train/grad_norm_pre_clip': 7.349503040313721, 'timing/data': 0.006008716009091586, 'timing/model': 1.1378165080095641, 'learning_rate/qwen_vl_interface': 5.14646387101121e-06, 'learning_rate/action_model': 5.14646387101121e-05, 'epoch': 1.06}) +[2026-06-16 00:05:59,189][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.00916078221052885, 'train/grad_norm_pre_clip': 1.5419944524765015, 'timing/data': 0.00627067603636533, 'timing/model': 1.2846350719919428, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 1.06}) +[2026-06-16 00:06:03,578][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.010356112383306026, 'train/grad_norm_pre_clip': 2.034072160720825, 'timing/data': 0.006177786970511079, 'timing/model': 1.141803961014375, 'learning_rate/qwen_vl_interface': 5.140374561989988e-06, 'learning_rate/action_model': 5.140374561989988e-05, 'epoch': 1.06}) +[2026-06-16 00:06:08,107][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.056891538202762604, 'train/grad_norm_pre_clip': 9.333403587341309, 'timing/data': 0.006626762973610312, 'timing/model': 1.1340945860138163, 'learning_rate/qwen_vl_interface': 5.137329974447924e-06, 'learning_rate/action_model': 5.137329974447924e-05, 'epoch': 1.06}) +[2026-06-16 00:06:12,571][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.024726906791329384, 'train/grad_norm_pre_clip': 4.9588751792907715, 'timing/data': 0.006655366043560207, 'timing/model': 1.1318149919970892, 'learning_rate/qwen_vl_interface': 5.1342854332202516e-06, 'learning_rate/action_model': 5.1342854332202514e-05, 'epoch': 1.06}) +[2026-06-16 00:06:17,091][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.016029195860028267, 'train/grad_norm_pre_clip': 4.462193489074707, 'timing/data': 0.006596553954295814, 'timing/model': 1.136710274964571, 'learning_rate/qwen_vl_interface': 5.131240939558467e-06, 'learning_rate/action_model': 5.131240939558467e-05, 'epoch': 1.06}) +[2026-06-16 00:06:21,642][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.20255854725837708, 'train/grad_norm_pre_clip': 28.231664657592773, 'timing/data': 0.006344611989334226, 'timing/model': 1.1448705309885554, 'learning_rate/qwen_vl_interface': 5.128196494714048e-06, 'learning_rate/action_model': 5.128196494714047e-05, 'epoch': 1.06}) +[2026-06-16 00:06:26,191][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.013009415939450264, 'train/grad_norm_pre_clip': 2.77189302444458, 'timing/data': 0.006134916970040649, 'timing/model': 1.3011622519697994, 'learning_rate/qwen_vl_interface': 5.125152099938448e-06, 'learning_rate/action_model': 5.125152099938447e-05, 'epoch': 1.06}) +[2026-06-16 00:06:30,571][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.004735782742500305, 'train/grad_norm_pre_clip': 0.6575965285301208, 'timing/data': 0.0065298599656671286, 'timing/model': 1.1378126870258711, 'learning_rate/qwen_vl_interface': 5.122107756483104e-06, 'learning_rate/action_model': 5.122107756483103e-05, 'epoch': 1.06}) +[2026-06-16 00:06:35,118][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.1544754058122635, 'train/grad_norm_pre_clip': 12.162896156311035, 'timing/data': 0.006581577006727457, 'timing/model': 1.136609984037932, 'learning_rate/qwen_vl_interface': 5.119063465599429e-06, 'learning_rate/action_model': 5.1190634655994285e-05, 'epoch': 1.07}) +[2026-06-16 00:06:39,625][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.2198474258184433, 'train/grad_norm_pre_clip': 32.67771911621094, 'timing/data': 0.006747564999386668, 'timing/model': 1.1412147189839743, 'learning_rate/qwen_vl_interface': 5.116019228538818e-06, 'learning_rate/action_model': 5.116019228538818e-05, 'epoch': 1.07}) +[2026-06-16 00:06:44,159][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.101485475897789, 'train/grad_norm_pre_clip': 9.015738487243652, 'timing/data': 0.007202015025541186, 'timing/model': 1.1404973930330016, 'learning_rate/qwen_vl_interface': 5.112975046552641e-06, 'learning_rate/action_model': 5.1129750465526404e-05, 'epoch': 1.07}) +[2026-06-16 00:06:48,696][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.051672548055648804, 'train/grad_norm_pre_clip': 5.248638153076172, 'timing/data': 0.00672735198168084, 'timing/model': 1.1320041610160843, 'learning_rate/qwen_vl_interface': 5.109930920892246e-06, 'learning_rate/action_model': 5.1099309208922464e-05, 'epoch': 1.07}) +[2026-06-16 00:06:53,229][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.07978308945894241, 'train/grad_norm_pre_clip': 20.675222396850586, 'timing/data': 0.006081723025999963, 'timing/model': 1.2875610109767877, 'learning_rate/qwen_vl_interface': 5.10688685280896e-06, 'learning_rate/action_model': 5.10688685280896e-05, 'epoch': 1.07}) +[2026-06-16 00:06:57,612][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.23436160385608673, 'train/grad_norm_pre_clip': 17.09070587158203, 'timing/data': 0.006667143956292421, 'timing/model': 1.1357434010133147, 'learning_rate/qwen_vl_interface': 5.103842843554082e-06, 'learning_rate/action_model': 5.103842843554082e-05, 'epoch': 1.07}) +[2026-06-16 00:07:02,157][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.04940812289714813, 'train/grad_norm_pre_clip': 17.947486877441406, 'timing/data': 0.006131869042292237, 'timing/model': 1.135424455977045, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 1.07}) +[2026-06-16 00:11:21,383][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.07357832789421082, 'train/grad_norm_pre_clip': 6.214162826538086, 'eval/loss': 0.042875729501247406, 'eval/flap_precision': 0.9662058371735791, 'eval/flap_recall': 0.970679012345679, 'eval/flap_f1': 0.968437259430331, 'eval/flap_rate_model': 0.18083333333333335, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9684372594303309, 'timing/data': 0.006565665011294186, 'timing/model': 1.137752206006553, 'learning_rate/qwen_vl_interface': 5.097755006534638e-06, 'learning_rate/action_model': 5.097755006534638e-05, 'epoch': 1.07}) +[2026-06-16 00:11:21,383][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2600_state +[2026-06-16 00:11:21,386][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 00:11:24,964][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2600_state/model.safetensors +[2026-06-16 00:11:35,048][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2600_state/optimizer.bin +[2026-06-16 00:11:35,048][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2600_state/sampler.bin +[2026-06-16 00:11:35,048][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2600_state/sampler_1.bin +[2026-06-16 00:11:35,049][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2600_state/random_states_0.pkl +[2026-06-16 00:11:35,095][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2600 +[2026-06-16 00:11:35,096][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 00:11:35,102][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 00:11:39,575][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.2715659439563751, 'train/grad_norm_pre_clip': 33.199092864990234, 'timing/data': 0.006731269997544587, 'timing/model': 1.1243975079851225, 'learning_rate/qwen_vl_interface': 5.094711181272551e-06, 'learning_rate/action_model': 5.094711181272551e-05, 'epoch': 1.07}) +[2026-06-16 00:11:44,026][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.30870339274406433, 'train/grad_norm_pre_clip': 8.584484100341797, 'timing/data': 0.006399164965841919, 'timing/model': 1.1153536509955302, 'learning_rate/qwen_vl_interface': 5.091667419843833e-06, 'learning_rate/action_model': 5.091667419843833e-05, 'epoch': 1.07}) +[2026-06-16 00:11:48,448][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.007350227329879999, 'train/grad_norm_pre_clip': 2.9289278984069824, 'timing/data': 0.006609175004996359, 'timing/model': 1.1167140589677729, 'learning_rate/qwen_vl_interface': 5.088623723499657e-06, 'learning_rate/action_model': 5.088623723499657e-05, 'epoch': 1.07}) +[2026-06-16 00:11:52,862][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.07649344950914383, 'train/grad_norm_pre_clip': 10.83636474609375, 'timing/data': 0.006648126000072807, 'timing/model': 1.1195173229789361, 'learning_rate/qwen_vl_interface': 5.085580093491171e-06, 'learning_rate/action_model': 5.085580093491171e-05, 'epoch': 1.07}) +[2026-06-16 00:11:57,308][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.1253810077905655, 'train/grad_norm_pre_clip': 14.276914596557617, 'timing/data': 0.006187794031575322, 'timing/model': 1.2404664239729755, 'learning_rate/qwen_vl_interface': 5.0825365310695e-06, 'learning_rate/action_model': 5.082536531069499e-05, 'epoch': 1.07}) +[2026-06-16 00:12:01,644][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.01745683327317238, 'train/grad_norm_pre_clip': 8.455953598022461, 'timing/data': 0.0062084749806672335, 'timing/model': 1.1186281339614652, 'learning_rate/qwen_vl_interface': 5.079493037485735e-06, 'learning_rate/action_model': 5.0794930374857344e-05, 'epoch': 1.07}) +[2026-06-16 00:12:06,140][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.4344307780265808, 'train/grad_norm_pre_clip': 26.41187858581543, 'timing/data': 0.006806342978961766, 'timing/model': 1.121033635979984, 'learning_rate/qwen_vl_interface': 5.076449613990941e-06, 'learning_rate/action_model': 5.076449613990941e-05, 'epoch': 1.07}) +[2026-06-16 00:12:10,657][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.08539630472660065, 'train/grad_norm_pre_clip': 11.091707229614258, 'timing/data': 0.006518022040836513, 'timing/model': 1.133339143008925, 'learning_rate/qwen_vl_interface': 5.073406261836157e-06, 'learning_rate/action_model': 5.0734062618361565e-05, 'epoch': 1.07}) +[2026-06-16 00:12:15,157][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.02989467978477478, 'train/grad_norm_pre_clip': 8.728178024291992, 'timing/data': 0.006286889023613185, 'timing/model': 1.1255932900239713, 'learning_rate/qwen_vl_interface': 5.070362982272386e-06, 'learning_rate/action_model': 5.070362982272385e-05, 'epoch': 1.07}) +[2026-06-16 00:12:19,673][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.1301952451467514, 'train/grad_norm_pre_clip': 8.53532886505127, 'timing/data': 0.006696599011775106, 'timing/model': 1.2896178219816647, 'learning_rate/qwen_vl_interface': 5.067319776550607e-06, 'learning_rate/action_model': 5.067319776550607e-05, 'epoch': 1.07}) +[2026-06-16 00:12:24,164][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.004336062353104353, 'train/grad_norm_pre_clip': 0.5867230892181396, 'timing/data': 0.005906520993448794, 'timing/model': 1.263095315953251, 'learning_rate/qwen_vl_interface': 5.064276645921768e-06, 'learning_rate/action_model': 5.064276645921768e-05, 'epoch': 1.07}) +[2026-06-16 00:12:28,529][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.17476235330104828, 'train/grad_norm_pre_clip': 11.648406982421875, 'timing/data': 0.006295425002463162, 'timing/model': 1.131701564998366, 'learning_rate/qwen_vl_interface': 5.061233591636787e-06, 'learning_rate/action_model': 5.061233591636787e-05, 'epoch': 1.07}) +[2026-06-16 00:12:33,048][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.09747136384248734, 'train/grad_norm_pre_clip': 14.635437965393066, 'timing/data': 0.006929885013960302, 'timing/model': 1.1341873449855484, 'learning_rate/qwen_vl_interface': 5.058190614946544e-06, 'learning_rate/action_model': 5.058190614946544e-05, 'epoch': 1.07}) +[2026-06-16 00:12:37,576][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.1081119254231453, 'train/grad_norm_pre_clip': 7.829172611236572, 'timing/data': 0.006550993013661355, 'timing/model': 1.1292064120061696, 'learning_rate/qwen_vl_interface': 5.055147717101895e-06, 'learning_rate/action_model': 5.055147717101894e-05, 'epoch': 1.07}) +[2026-06-16 00:12:42,107][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.12402133643627167, 'train/grad_norm_pre_clip': 13.359243392944336, 'timing/data': 0.006481389980763197, 'timing/model': 1.1323945420444943, 'learning_rate/qwen_vl_interface': 5.052104899353658e-06, 'learning_rate/action_model': 5.0521048993536577e-05, 'epoch': 1.07}) +[2026-06-16 00:12:46,604][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.09502065181732178, 'train/grad_norm_pre_clip': 8.437164306640625, 'timing/data': 0.006721929996274412, 'timing/model': 1.1298247859813273, 'learning_rate/qwen_vl_interface': 5.049062162952622e-06, 'learning_rate/action_model': 5.0490621629526214e-05, 'epoch': 1.07}) +[2026-06-16 00:12:51,089][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.18278929591178894, 'train/grad_norm_pre_clip': 13.63640308380127, 'timing/data': 0.0066623620223253965, 'timing/model': 1.2489494999754243, 'learning_rate/qwen_vl_interface': 5.046019509149541e-06, 'learning_rate/action_model': 5.046019509149541e-05, 'epoch': 1.08}) +[2026-06-16 00:12:55,456][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.2312031090259552, 'train/grad_norm_pre_clip': 22.301151275634766, 'timing/data': 0.006362524989526719, 'timing/model': 1.130418243992608, 'learning_rate/qwen_vl_interface': 5.042976939195135e-06, 'learning_rate/action_model': 5.0429769391951355e-05, 'epoch': 1.08}) +[2026-06-16 00:13:00,044][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.0728750005364418, 'train/grad_norm_pre_clip': 12.043292045593262, 'timing/data': 0.006888745003379881, 'timing/model': 1.1409473909880035, 'learning_rate/qwen_vl_interface': 5.039934454340091e-06, 'learning_rate/action_model': 5.0399344543400907e-05, 'epoch': 1.08}) +[2026-06-16 00:13:04,569][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.0180742759257555, 'train/grad_norm_pre_clip': 5.013848781585693, 'timing/data': 0.0063959810067899525, 'timing/model': 1.1345590409473516, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 1.08}) +[2026-06-16 00:13:09,098][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.02221956104040146, 'train/grad_norm_pre_clip': 8.506933212280273, 'timing/data': 0.006772872002329677, 'timing/model': 1.1304613100364804, 'learning_rate/qwen_vl_interface': 5.033849744930647e-06, 'learning_rate/action_model': 5.033849744930646e-05, 'epoch': 1.08}) +[2026-06-16 00:13:13,636][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.09460248053073883, 'train/grad_norm_pre_clip': 12.518229484558105, 'timing/data': 0.006597576022613794, 'timing/model': 1.291927571000997, 'learning_rate/qwen_vl_interface': 5.030807522877443e-06, 'learning_rate/action_model': 5.030807522877443e-05, 'epoch': 1.08}) +[2026-06-16 00:13:18,030][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.2238973081111908, 'train/grad_norm_pre_clip': 16.374711990356445, 'timing/data': 0.006555993983056396, 'timing/model': 1.139726671972312, 'learning_rate/qwen_vl_interface': 5.027765390925987e-06, 'learning_rate/action_model': 5.027765390925987e-05, 'epoch': 1.08}) +[2026-06-16 00:13:22,589][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.009907367639243603, 'train/grad_norm_pre_clip': 0.9470611214637756, 'timing/data': 0.006292149017099291, 'timing/model': 1.1442483540158719, 'learning_rate/qwen_vl_interface': 5.024723350326783e-06, 'learning_rate/action_model': 5.024723350326783e-05, 'epoch': 1.08}) +[2026-06-16 00:13:27,059][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.05020925775170326, 'train/grad_norm_pre_clip': 7.65329647064209, 'timing/data': 0.006569446006324142, 'timing/model': 1.1329598979791626, 'learning_rate/qwen_vl_interface': 5.021681402330303e-06, 'learning_rate/action_model': 5.021681402330303e-05, 'epoch': 1.08}) +[2026-06-16 00:13:31,552][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.07940502464771271, 'train/grad_norm_pre_clip': 15.643696784973145, 'timing/data': 0.006847443000879139, 'timing/model': 1.1357193189905956, 'learning_rate/qwen_vl_interface': 5.018639548186972e-06, 'learning_rate/action_model': 5.0186395481869716e-05, 'epoch': 1.08}) +[2026-06-16 00:13:36,049][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.02405383251607418, 'train/grad_norm_pre_clip': 10.427882194519043, 'timing/data': 0.006664933986030519, 'timing/model': 1.1299185599782504, 'learning_rate/qwen_vl_interface': 5.015597789147184e-06, 'learning_rate/action_model': 5.015597789147184e-05, 'epoch': 1.08}) +[2026-06-16 00:13:40,596][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.0404445081949234, 'train/grad_norm_pre_clip': 18.606313705444336, 'timing/data': 0.006362220970913768, 'timing/model': 1.300815552007407, 'learning_rate/qwen_vl_interface': 5.012556126461291e-06, 'learning_rate/action_model': 5.012556126461291e-05, 'epoch': 1.08}) +[2026-06-16 00:13:44,982][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.14568579196929932, 'train/grad_norm_pre_clip': 11.063483238220215, 'timing/data': 0.006910508032888174, 'timing/model': 1.1310822769883089, 'learning_rate/qwen_vl_interface': 5.0095145613796044e-06, 'learning_rate/action_model': 5.0095145613796046e-05, 'epoch': 1.08}) +[2026-06-16 00:13:49,534][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.018285663798451424, 'train/grad_norm_pre_clip': 1.808241367340088, 'timing/data': 0.006495822977740318, 'timing/model': 1.1346329979714938, 'learning_rate/qwen_vl_interface': 5.0064730951524e-06, 'learning_rate/action_model': 5.0064730951524e-05, 'epoch': 1.08}) +[2026-06-16 00:13:54,065][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.23060376942157745, 'train/grad_norm_pre_clip': 4.823803901672363, 'timing/data': 0.0062679279944859445, 'timing/model': 1.1346863660146482, 'learning_rate/qwen_vl_interface': 5.003431729029905e-06, 'learning_rate/action_model': 5.003431729029905e-05, 'epoch': 1.08}) +[2026-06-16 00:13:58,580][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.017071032896637917, 'train/grad_norm_pre_clip': 2.011624574661255, 'timing/data': 0.005832305003423244, 'timing/model': 1.1411634589894675, 'learning_rate/qwen_vl_interface': 5.000390464262314e-06, 'learning_rate/action_model': 5.0003904642623136e-05, 'epoch': 1.08}) +[2026-06-16 00:14:03,097][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.013382107019424438, 'train/grad_norm_pre_clip': 2.2619714736938477, 'timing/data': 0.007070313964504749, 'timing/model': 1.1367556669865735, 'learning_rate/qwen_vl_interface': 4.997349302099774e-06, 'learning_rate/action_model': 4.997349302099773e-05, 'epoch': 1.08}) +[2026-06-16 00:14:07,644][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.010631856508553028, 'train/grad_norm_pre_clip': 2.4656858444213867, 'timing/data': 0.006474132009316236, 'timing/model': 1.305576418992132, 'learning_rate/qwen_vl_interface': 4.994308243792391e-06, 'learning_rate/action_model': 4.994308243792391e-05, 'epoch': 1.08}) +[2026-06-16 00:14:12,043][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.40641242265701294, 'train/grad_norm_pre_clip': 21.232772827148438, 'timing/data': 0.006509045953862369, 'timing/model': 1.1324506090022624, 'learning_rate/qwen_vl_interface': 4.991267290590233e-06, 'learning_rate/action_model': 4.991267290590233e-05, 'epoch': 1.08}) +[2026-06-16 00:14:16,592][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.013748464174568653, 'train/grad_norm_pre_clip': 2.349968194961548, 'timing/data': 0.00661545799812302, 'timing/model': 1.1328681519953534, 'learning_rate/qwen_vl_interface': 4.988226443743317e-06, 'learning_rate/action_model': 4.9882264437433174e-05, 'epoch': 1.08}) +[2026-06-16 00:14:21,132][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.001879264134913683, 'train/grad_norm_pre_clip': 0.1589069813489914, 'timing/data': 0.006566266994923353, 'timing/model': 1.1321841059834696, 'learning_rate/qwen_vl_interface': 4.985185704501623e-06, 'learning_rate/action_model': 4.985185704501623e-05, 'epoch': 1.08}) +[2026-06-16 00:14:25,622][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.12435171008110046, 'train/grad_norm_pre_clip': 6.05257511138916, 'timing/data': 0.006432776048313826, 'timing/model': 1.134380790987052, 'learning_rate/qwen_vl_interface': 4.982145074115082e-06, 'learning_rate/action_model': 4.982145074115082e-05, 'epoch': 1.08}) +[2026-06-16 00:14:30,186][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.1062651127576828, 'train/grad_norm_pre_clip': 6.151413440704346, 'timing/data': 0.006254765030462295, 'timing/model': 1.1427778489887714, 'learning_rate/qwen_vl_interface': 4.979104553833584e-06, 'learning_rate/action_model': 4.979104553833584e-05, 'epoch': 1.08}) +[2026-06-16 00:14:34,760][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.2704576551914215, 'train/grad_norm_pre_clip': 47.54735565185547, 'timing/data': 0.006526691955514252, 'timing/model': 1.3058567160042003, 'learning_rate/qwen_vl_interface': 4.9760641449069724e-06, 'learning_rate/action_model': 4.976064144906972e-05, 'epoch': 1.08}) +[2026-06-16 00:14:39,148][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.08463193476200104, 'train/grad_norm_pre_clip': 3.476926565170288, 'timing/data': 0.006802503019571304, 'timing/model': 1.132615280977916, 'learning_rate/qwen_vl_interface': 4.973023848585042e-06, 'learning_rate/action_model': 4.973023848585042e-05, 'epoch': 1.08}) +[2026-06-16 00:14:43,695][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.11912152171134949, 'train/grad_norm_pre_clip': 8.895256042480469, 'timing/data': 0.0067012119689024985, 'timing/model': 1.1406541730393656, 'learning_rate/qwen_vl_interface': 4.969983666117545e-06, 'learning_rate/action_model': 4.9699836661175446e-05, 'epoch': 1.09}) +[2026-06-16 00:14:48,205][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.10859070718288422, 'train/grad_norm_pre_clip': 14.671273231506348, 'timing/data': 0.006456589966546744, 'timing/model': 1.1365580019773915, 'learning_rate/qwen_vl_interface': 4.966943598754185e-06, 'learning_rate/action_model': 4.9669435987541854e-05, 'epoch': 1.09}) +[2026-06-16 00:14:52,718][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.20431849360466003, 'train/grad_norm_pre_clip': 6.410880088806152, 'timing/data': 0.006382394989486784, 'timing/model': 1.1345966949593276, 'learning_rate/qwen_vl_interface': 4.963903647744621e-06, 'learning_rate/action_model': 4.96390364774462e-05, 'epoch': 1.09}) +[2026-06-16 00:14:57,287][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.28825807571411133, 'train/grad_norm_pre_clip': 25.282901763916016, 'timing/data': 0.006543898023664951, 'timing/model': 1.1343933870084584, 'learning_rate/qwen_vl_interface': 4.9608638143384604e-06, 'learning_rate/action_model': 4.960863814338461e-05, 'epoch': 1.09}) +[2026-06-16 00:15:01,872][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.04398947209119797, 'train/grad_norm_pre_clip': 6.210872173309326, 'timing/data': 0.00643762102117762, 'timing/model': 1.322146493010223, 'learning_rate/qwen_vl_interface': 4.9578240997852636e-06, 'learning_rate/action_model': 4.957824099785264e-05, 'epoch': 1.09}) +[2026-06-16 00:15:06,263][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.05898115411400795, 'train/grad_norm_pre_clip': 11.106586456298828, 'timing/data': 0.00657374900765717, 'timing/model': 1.1323053929954767, 'learning_rate/qwen_vl_interface': 4.954784505334544e-06, 'learning_rate/action_model': 4.9547845053345434e-05, 'epoch': 1.09}) +[2026-06-16 00:15:10,836][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.048543743789196014, 'train/grad_norm_pre_clip': 9.635208129882812, 'timing/data': 0.006248581048566848, 'timing/model': 1.1439418960362673, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 1.09}) +[2026-06-16 00:15:15,358][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.03889966011047363, 'train/grad_norm_pre_clip': 7.7338666915893555, 'timing/data': 0.0063719520112499595, 'timing/model': 1.1463017260539345, 'learning_rate/qwen_vl_interface': 4.94870568173833e-06, 'learning_rate/action_model': 4.9487056817383307e-05, 'epoch': 1.09}) +[2026-06-16 00:15:19,890][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.11759932339191437, 'train/grad_norm_pre_clip': 9.619705200195312, 'timing/data': 0.006322542962152511, 'timing/model': 1.137580902024638, 'learning_rate/qwen_vl_interface': 4.945666455091614e-06, 'learning_rate/action_model': 4.9456664550916136e-05, 'epoch': 1.09}) +[2026-06-16 00:15:24,461][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.10441698879003525, 'train/grad_norm_pre_clip': 14.68571949005127, 'timing/data': 0.006510513019748032, 'timing/model': 1.1448744380031712, 'learning_rate/qwen_vl_interface': 4.942627353544921e-06, 'learning_rate/action_model': 4.942627353544921e-05, 'epoch': 1.09}) +[2026-06-16 00:15:29,020][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.03481142967939377, 'train/grad_norm_pre_clip': 6.157726287841797, 'timing/data': 0.00621381402015686, 'timing/model': 1.2998589589842595, 'learning_rate/qwen_vl_interface': 4.939588378347513e-06, 'learning_rate/action_model': 4.9395883783475125e-05, 'epoch': 1.09}) +[2026-06-16 00:15:33,440][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.11165159195661545, 'train/grad_norm_pre_clip': 10.417424201965332, 'timing/data': 0.007089653983712196, 'timing/model': 1.146828182041645, 'learning_rate/qwen_vl_interface': 4.936549530748597e-06, 'learning_rate/action_model': 4.9365495307485963e-05, 'epoch': 1.09}) +[2026-06-16 00:15:38,010][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.031184310093522072, 'train/grad_norm_pre_clip': 8.611248016357422, 'timing/data': 0.006485674995929003, 'timing/model': 1.1378965259646066, 'learning_rate/qwen_vl_interface': 4.9335108119973286e-06, 'learning_rate/action_model': 4.933510811997328e-05, 'epoch': 1.09}) +[2026-06-16 00:15:42,528][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.02708558551967144, 'train/grad_norm_pre_clip': 4.242574691772461, 'timing/data': 0.006361940992064774, 'timing/model': 1.1411783050280064, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 1.09}) +[2026-06-16 00:15:47,079][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.21434178948402405, 'train/grad_norm_pre_clip': 6.381392478942871, 'timing/data': 0.006665601045824587, 'timing/model': 1.1403728649602272, 'learning_rate/qwen_vl_interface': 4.927433766034093e-06, 'learning_rate/action_model': 4.9274337660340934e-05, 'epoch': 1.09}) +[2026-06-16 00:15:51,633][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.3179296553134918, 'train/grad_norm_pre_clip': 17.351709365844727, 'timing/data': 0.006352546974085271, 'timing/model': 1.1452204080414958, 'learning_rate/qwen_vl_interface': 4.92439544132017e-06, 'learning_rate/action_model': 4.92439544132017e-05, 'epoch': 1.09}) +[2026-06-16 00:15:56,178][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.07900529354810715, 'train/grad_norm_pre_clip': 5.227150917053223, 'timing/data': 0.00634605303639546, 'timing/model': 1.2853603840339929, 'learning_rate/qwen_vl_interface': 4.92135725044998e-06, 'learning_rate/action_model': 4.92135725044998e-05, 'epoch': 1.09}) +[2026-06-16 00:16:00,663][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.30191823840141296, 'train/grad_norm_pre_clip': 13.24842357635498, 'timing/data': 0.006282293994445354, 'timing/model': 1.2305961590027437, 'learning_rate/qwen_vl_interface': 4.918319194672412e-06, 'learning_rate/action_model': 4.918319194672412e-05, 'epoch': 1.09}) +[2026-06-16 00:16:05,058][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.08968319743871689, 'train/grad_norm_pre_clip': 31.31902313232422, 'timing/data': 0.006521289993543178, 'timing/model': 1.1392648599576205, 'learning_rate/qwen_vl_interface': 4.9152812752362935e-06, 'learning_rate/action_model': 4.915281275236293e-05, 'epoch': 1.09}) +[2026-06-16 00:16:09,576][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.03507572412490845, 'train/grad_norm_pre_clip': 4.192192077636719, 'timing/data': 0.006294696999248117, 'timing/model': 1.141110058000777, 'learning_rate/qwen_vl_interface': 4.9122434933904e-06, 'learning_rate/action_model': 4.9122434933904e-05, 'epoch': 1.09}) +[2026-06-16 00:16:14,119][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.2686711847782135, 'train/grad_norm_pre_clip': 13.836770057678223, 'timing/data': 0.0064124909695237875, 'timing/model': 1.1360909789800644, 'learning_rate/qwen_vl_interface': 4.909205850383449e-06, 'learning_rate/action_model': 4.9092058503834485e-05, 'epoch': 1.09}) +[2026-06-16 00:16:18,678][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.2105184942483902, 'train/grad_norm_pre_clip': 10.870851516723633, 'timing/data': 0.006239065027330071, 'timing/model': 1.141817451978568, 'learning_rate/qwen_vl_interface': 4.906168347464098e-06, 'learning_rate/action_model': 4.906168347464098e-05, 'epoch': 1.09}) +[2026-06-16 00:16:23,240][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.20550012588500977, 'train/grad_norm_pre_clip': 5.376800537109375, 'timing/data': 0.006877718027681112, 'timing/model': 1.2933807239751332, 'learning_rate/qwen_vl_interface': 4.903130985880955e-06, 'learning_rate/action_model': 4.903130985880954e-05, 'epoch': 1.09}) +[2026-06-16 00:16:27,748][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.017911069095134735, 'train/grad_norm_pre_clip': 2.3509674072265625, 'timing/data': 0.006529184989631176, 'timing/model': 1.2649612259701826, 'learning_rate/qwen_vl_interface': 4.900093766882558e-06, 'learning_rate/action_model': 4.900093766882558e-05, 'epoch': 1.09}) +[2026-06-16 00:16:32,143][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.10833998769521713, 'train/grad_norm_pre_clip': 14.041980743408203, 'timing/data': 0.006700689031276852, 'timing/model': 1.1421793270274065, 'learning_rate/qwen_vl_interface': 4.897056691717397e-06, 'learning_rate/action_model': 4.8970566917173966e-05, 'epoch': 1.1}) +[2026-06-16 00:16:36,672][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.023749733343720436, 'train/grad_norm_pre_clip': 3.4980428218841553, 'timing/data': 0.006716871983371675, 'timing/model': 1.1325773329590447, 'learning_rate/qwen_vl_interface': 4.894019761633898e-06, 'learning_rate/action_model': 4.894019761633898e-05, 'epoch': 1.1}) +[2026-06-16 00:16:41,217][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.10204360634088516, 'train/grad_norm_pre_clip': 7.351815223693848, 'timing/data': 0.006454018992371857, 'timing/model': 1.1378510119975545, 'learning_rate/qwen_vl_interface': 4.8909829778804304e-06, 'learning_rate/action_model': 4.89098297788043e-05, 'epoch': 1.1}) +[2026-06-16 00:16:45,780][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.13161678612232208, 'train/grad_norm_pre_clip': 7.7069268226623535, 'timing/data': 0.006199637020472437, 'timing/model': 1.1365297060110606, 'learning_rate/qwen_vl_interface': 4.887946341705298e-06, 'learning_rate/action_model': 4.887946341705297e-05, 'epoch': 1.1}) +[2026-06-16 00:16:50,332][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.17375105619430542, 'train/grad_norm_pre_clip': 18.17787742614746, 'timing/data': 0.006510984036140144, 'timing/model': 1.295648134953808, 'learning_rate/qwen_vl_interface': 4.8849098543567465e-06, 'learning_rate/action_model': 4.884909854356747e-05, 'epoch': 1.1}) +[2026-06-16 00:16:54,726][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.15240727365016937, 'train/grad_norm_pre_clip': 9.737555503845215, 'timing/data': 0.006521689996588975, 'timing/model': 1.1376589440042153, 'learning_rate/qwen_vl_interface': 4.881873517082965e-06, 'learning_rate/action_model': 4.8818735170829655e-05, 'epoch': 1.1}) +[2026-06-16 00:16:59,283][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.14294415712356567, 'train/grad_norm_pre_clip': 9.69275951385498, 'timing/data': 0.006384483014699072, 'timing/model': 1.1389386639930308, 'learning_rate/qwen_vl_interface': 4.878837331132076e-06, 'learning_rate/action_model': 4.8788373311320754e-05, 'epoch': 1.1}) +[2026-06-16 00:17:03,833][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.14265833795070648, 'train/grad_norm_pre_clip': 13.93370532989502, 'timing/data': 0.006208501989021897, 'timing/model': 1.1359524910221808, 'learning_rate/qwen_vl_interface': 4.87580129775214e-06, 'learning_rate/action_model': 4.87580129775214e-05, 'epoch': 1.1}) +[2026-06-16 00:17:08,390][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.15863488614559174, 'train/grad_norm_pre_clip': 15.2404146194458, 'timing/data': 0.006361818988807499, 'timing/model': 1.1364104350213893, 'learning_rate/qwen_vl_interface': 4.872765418191156e-06, 'learning_rate/action_model': 4.872765418191155e-05, 'epoch': 1.1}) +[2026-06-16 00:17:12,899][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.02987930364906788, 'train/grad_norm_pre_clip': 2.662724256515503, 'timing/data': 0.00652535300469026, 'timing/model': 1.1406726469867863, 'learning_rate/qwen_vl_interface': 4.8697296936970575e-06, 'learning_rate/action_model': 4.869729693697057e-05, 'epoch': 1.1}) +[2026-06-16 00:17:17,383][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.1999015212059021, 'train/grad_norm_pre_clip': 9.555468559265137, 'timing/data': 0.006519649003166705, 'timing/model': 1.2392788410070352, 'learning_rate/qwen_vl_interface': 4.8666941255177185e-06, 'learning_rate/action_model': 4.8666941255177185e-05, 'epoch': 1.1}) +[2026-06-16 00:17:21,771][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.057271771132946014, 'train/grad_norm_pre_clip': 6.9599528312683105, 'timing/data': 0.006128547014668584, 'timing/model': 1.13729925500229, 'learning_rate/qwen_vl_interface': 4.863658714900945e-06, 'learning_rate/action_model': 4.863658714900945e-05, 'epoch': 1.1}) +[2026-06-16 00:17:26,332][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.05670039728283882, 'train/grad_norm_pre_clip': 4.180047512054443, 'timing/data': 0.0066804729867726564, 'timing/model': 1.1415263500530273, 'learning_rate/qwen_vl_interface': 4.860623463094483e-06, 'learning_rate/action_model': 4.860623463094483e-05, 'epoch': 1.1}) +[2026-06-16 00:17:30,880][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.05569558963179588, 'train/grad_norm_pre_clip': 4.50025749206543, 'timing/data': 0.00627968501066789, 'timing/model': 1.1438363149645738, 'learning_rate/qwen_vl_interface': 4.857588371346006e-06, 'learning_rate/action_model': 4.8575883713460057e-05, 'epoch': 1.1}) +[2026-06-16 00:17:35,431][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.08516836911439896, 'train/grad_norm_pre_clip': 4.939908981323242, 'timing/data': 0.0065365819609723985, 'timing/model': 1.1359666630160064, 'learning_rate/qwen_vl_interface': 4.854553440903127e-06, 'learning_rate/action_model': 4.854553440903126e-05, 'epoch': 1.1}) +[2026-06-16 00:17:39,945][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.01528233103454113, 'train/grad_norm_pre_clip': 0.8583272695541382, 'timing/data': 0.006241275987122208, 'timing/model': 1.142400314041879, 'learning_rate/qwen_vl_interface': 4.85151867301339e-06, 'learning_rate/action_model': 4.85151867301339e-05, 'epoch': 1.1}) +[2026-06-16 00:17:44,455][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.15157848596572876, 'train/grad_norm_pre_clip': 10.573655128479004, 'timing/data': 0.0064376419759355485, 'timing/model': 1.2639165800064802, 'learning_rate/qwen_vl_interface': 4.848484068924274e-06, 'learning_rate/action_model': 4.848484068924275e-05, 'epoch': 1.1}) +[2026-06-16 00:17:48,852][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.05410851538181305, 'train/grad_norm_pre_clip': 10.545101165771484, 'timing/data': 0.006229288992471993, 'timing/model': 1.137683951004874, 'learning_rate/qwen_vl_interface': 4.845449629883193e-06, 'learning_rate/action_model': 4.8454496298831934e-05, 'epoch': 1.1}) +[2026-06-16 00:17:53,395][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.04221503436565399, 'train/grad_norm_pre_clip': 5.210216045379639, 'timing/data': 0.006534072977956384, 'timing/model': 1.1425027310033329, 'learning_rate/qwen_vl_interface': 4.842415357137487e-06, 'learning_rate/action_model': 4.842415357137487e-05, 'epoch': 1.1}) +[2026-06-16 00:17:57,960][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.20568016171455383, 'train/grad_norm_pre_clip': 9.374527931213379, 'timing/data': 0.006440323020797223, 'timing/model': 1.142482105991803, 'learning_rate/qwen_vl_interface': 4.839381251934432e-06, 'learning_rate/action_model': 4.8393812519344316e-05, 'epoch': 1.1}) +[2026-06-16 00:18:02,505][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.028077132999897003, 'train/grad_norm_pre_clip': 4.4901204109191895, 'timing/data': 0.006891179014928639, 'timing/model': 1.1379980989731848, 'learning_rate/qwen_vl_interface': 4.836347315521233e-06, 'learning_rate/action_model': 4.836347315521233e-05, 'epoch': 1.1}) +[2026-06-16 00:18:07,072][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.007748886942863464, 'train/grad_norm_pre_clip': 0.7880163788795471, 'timing/data': 0.006052631011698395, 'timing/model': 1.3180792899802327, 'learning_rate/qwen_vl_interface': 4.833313549145029e-06, 'learning_rate/action_model': 4.833313549145029e-05, 'epoch': 1.1}) +[2026-06-16 00:18:11,478][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.007412911392748356, 'train/grad_norm_pre_clip': 1.113433599472046, 'timing/data': 0.006424087972845882, 'timing/model': 1.1433527609915473, 'learning_rate/qwen_vl_interface': 4.830279954052884e-06, 'learning_rate/action_model': 4.8302799540528836e-05, 'epoch': 1.1}) +[2026-06-16 00:18:16,035][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.062126100063323975, 'train/grad_norm_pre_clip': 4.0109710693359375, 'timing/data': 0.006614761950913817, 'timing/model': 1.1478464860119857, 'learning_rate/qwen_vl_interface': 4.827246531491797e-06, 'learning_rate/action_model': 4.827246531491797e-05, 'epoch': 1.1}) +[2026-06-16 00:18:20,487][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.20867888629436493, 'train/grad_norm_pre_clip': 3.706743001937866, 'timing/data': 0.006628380971960723, 'timing/model': 1.1367800449952483, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 1.11}) +[2026-06-16 00:18:24,962][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.04956920072436333, 'train/grad_norm_pre_clip': 5.425497531890869, 'timing/data': 0.006763370998669416, 'timing/model': 1.1350379830109887, 'learning_rate/qwen_vl_interface': 4.821180208950422e-06, 'learning_rate/action_model': 4.821180208950421e-05, 'epoch': 1.11}) +[2026-06-16 00:18:29,500][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.11320916563272476, 'train/grad_norm_pre_clip': 20.88469696044922, 'timing/data': 0.006095573015045375, 'timing/model': 1.1354956310242414, 'learning_rate/qwen_vl_interface': 4.81814731146377e-06, 'learning_rate/action_model': 4.8181473114637695e-05, 'epoch': 1.11}) +[2026-06-16 00:18:34,057][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.03747555613517761, 'train/grad_norm_pre_clip': 11.503625869750977, 'timing/data': 0.006732962967362255, 'timing/model': 1.2990557189914398, 'learning_rate/qwen_vl_interface': 4.8151145914954465e-06, 'learning_rate/action_model': 4.815114591495446e-05, 'epoch': 1.11}) +[2026-06-16 00:18:38,453][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.050272196531295776, 'train/grad_norm_pre_clip': 18.11580467224121, 'timing/data': 0.006174360984005034, 'timing/model': 1.1352745559997857, 'learning_rate/qwen_vl_interface': 4.812082050292088e-06, 'learning_rate/action_model': 4.8120820502920876e-05, 'epoch': 1.11}) +[2026-06-16 00:18:43,003][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.014890344813466072, 'train/grad_norm_pre_clip': 3.466222047805786, 'timing/data': 0.006852997001260519, 'timing/model': 1.1310386160039343, 'learning_rate/qwen_vl_interface': 4.809049689100256e-06, 'learning_rate/action_model': 4.8090496891002555e-05, 'epoch': 1.11}) +[2026-06-16 00:18:47,552][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.034330178052186966, 'train/grad_norm_pre_clip': 9.446527481079102, 'timing/data': 0.006473751971498132, 'timing/model': 1.144341099949088, 'learning_rate/qwen_vl_interface': 4.80601750916644e-06, 'learning_rate/action_model': 4.8060175091664404e-05, 'epoch': 1.11}) +[2026-06-16 00:18:52,065][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.02753022313117981, 'train/grad_norm_pre_clip': 4.337568283081055, 'timing/data': 0.006636952981352806, 'timing/model': 1.1355679049738683, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 1.11}) +[2026-06-16 00:18:56,594][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.3032935559749603, 'train/grad_norm_pre_clip': 13.357579231262207, 'timing/data': 0.0064383940189145505, 'timing/model': 1.1377123430138454, 'learning_rate/qwen_vl_interface': 4.799953698058444e-06, 'learning_rate/action_model': 4.7999536980584434e-05, 'epoch': 1.11}) +[2026-06-16 00:19:01,148][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.13560418784618378, 'train/grad_norm_pre_clip': 6.290812015533447, 'timing/data': 0.006610839976929128, 'timing/model': 1.3051924610044807, 'learning_rate/qwen_vl_interface': 4.796922069376867e-06, 'learning_rate/action_model': 4.796922069376867e-05, 'epoch': 1.11}) +[2026-06-16 00:23:22,695][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.08979624509811401, 'train/grad_norm_pre_clip': 7.287497520446777, 'eval/loss': 0.04128611832857132, 'eval/flap_precision': 0.9328467153284672, 'eval/flap_recall': 0.9861111111111112, 'eval/flap_f1': 0.9587396849212303, 'eval/flap_rate_model': 0.19027777777777777, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9587396849212303, 'timing/data': 0.006385935004800558, 'timing/model': 1.1320748220314272, 'learning_rate/qwen_vl_interface': 4.793890626938511e-06, 'learning_rate/action_model': 4.793890626938511e-05, 'epoch': 1.11}) +[2026-06-16 00:23:22,695][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2700_state +[2026-06-16 00:23:22,698][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 00:23:26,283][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2700_state/model.safetensors +[2026-06-16 00:23:36,305][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2700_state/optimizer.bin +[2026-06-16 00:23:36,306][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2700_state/sampler.bin +[2026-06-16 00:23:36,306][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2700_state/sampler_1.bin +[2026-06-16 00:23:36,306][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2700_state/random_states_0.pkl +[2026-06-16 00:23:36,353][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2700 +[2026-06-16 00:23:36,353][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 00:23:36,359][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 00:23:40,816][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.04527886584401131, 'train/grad_norm_pre_clip': 6.715295314788818, 'timing/data': 0.0065374120022170246, 'timing/model': 1.2933675109525211, 'learning_rate/qwen_vl_interface': 4.7908593719894914e-06, 'learning_rate/action_model': 4.790859371989491e-05, 'epoch': 1.11}) +[2026-06-16 00:23:45,134][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.0612330436706543, 'train/grad_norm_pre_clip': 14.98491382598877, 'timing/data': 0.006653537042438984, 'timing/model': 1.1171513230074197, 'learning_rate/qwen_vl_interface': 4.78782830577584e-06, 'learning_rate/action_model': 4.7878283057758396e-05, 'epoch': 1.11}) +[2026-06-16 00:23:49,616][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.030215641483664513, 'train/grad_norm_pre_clip': 6.356927871704102, 'timing/data': 0.006252902967389673, 'timing/model': 1.1164808130124584, 'learning_rate/qwen_vl_interface': 4.784797429543513e-06, 'learning_rate/action_model': 4.784797429543513e-05, 'epoch': 1.11}) +[2026-06-16 00:23:54,096][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.09627863019704819, 'train/grad_norm_pre_clip': 13.483939170837402, 'timing/data': 0.00638874399010092, 'timing/model': 1.1135589000186883, 'learning_rate/qwen_vl_interface': 4.781766744538393e-06, 'learning_rate/action_model': 4.781766744538393e-05, 'epoch': 1.11}) +[2026-06-16 00:23:58,568][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.07660709321498871, 'train/grad_norm_pre_clip': 6.970455646514893, 'timing/data': 0.00619436998385936, 'timing/model': 1.1202415940351784, 'learning_rate/qwen_vl_interface': 4.778736252006277e-06, 'learning_rate/action_model': 4.7787362520062765e-05, 'epoch': 1.11}) +[2026-06-16 00:24:03,002][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.01869015023112297, 'train/grad_norm_pre_clip': 1.654470682144165, 'timing/data': 0.006350332987494767, 'timing/model': 1.118503329984378, 'learning_rate/qwen_vl_interface': 4.775705953192885e-06, 'learning_rate/action_model': 4.7757059531928856e-05, 'epoch': 1.11}) +[2026-06-16 00:24:07,499][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.01060505025088787, 'train/grad_norm_pre_clip': 3.5354719161987305, 'timing/data': 0.006068452028557658, 'timing/model': 1.2852186260279268, 'learning_rate/qwen_vl_interface': 4.7726758493438615e-06, 'learning_rate/action_model': 4.7726758493438613e-05, 'epoch': 1.11}) +[2026-06-16 00:24:11,844][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.17533519864082336, 'train/grad_norm_pre_clip': 4.501883506774902, 'timing/data': 0.00631809898186475, 'timing/model': 1.1219731520395726, 'learning_rate/qwen_vl_interface': 4.769645941704766e-06, 'learning_rate/action_model': 4.769645941704766e-05, 'epoch': 1.11}) +[2026-06-16 00:24:16,356][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.012328297831118107, 'train/grad_norm_pre_clip': 1.78954017162323, 'timing/data': 0.006612233002670109, 'timing/model': 1.12853854597779, 'learning_rate/qwen_vl_interface': 4.766616231521081e-06, 'learning_rate/action_model': 4.7666162315210806e-05, 'epoch': 1.11}) +[2026-06-16 00:24:20,859][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.12560997903347015, 'train/grad_norm_pre_clip': 18.626388549804688, 'timing/data': 0.006356840021908283, 'timing/model': 1.1282990960171446, 'learning_rate/qwen_vl_interface': 4.763586720038202e-06, 'learning_rate/action_model': 4.7635867200382014e-05, 'epoch': 1.11}) +[2026-06-16 00:24:25,373][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.10568229854106903, 'train/grad_norm_pre_clip': 8.26955795288086, 'timing/data': 0.006056557991541922, 'timing/model': 1.1234877750393935, 'learning_rate/qwen_vl_interface': 4.76055740850145e-06, 'learning_rate/action_model': 4.76055740850145e-05, 'epoch': 1.11}) +[2026-06-16 00:24:29,870][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.039153628051280975, 'train/grad_norm_pre_clip': 8.71133041381836, 'timing/data': 0.006289887998718768, 'timing/model': 1.1228178340243176, 'learning_rate/qwen_vl_interface': 4.757528298156058e-06, 'learning_rate/action_model': 4.757528298156058e-05, 'epoch': 1.11}) +[2026-06-16 00:24:34,384][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.12463492900133133, 'train/grad_norm_pre_clip': 19.202003479003906, 'timing/data': 0.006084602966438979, 'timing/model': 1.2797972140251659, 'learning_rate/qwen_vl_interface': 4.754499390247181e-06, 'learning_rate/action_model': 4.754499390247181e-05, 'epoch': 1.11}) +[2026-06-16 00:24:38,739][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.07636810839176178, 'train/grad_norm_pre_clip': 4.1591901779174805, 'timing/data': 0.006482659024186432, 'timing/model': 1.1306819649762474, 'learning_rate/qwen_vl_interface': 4.751470686019887e-06, 'learning_rate/action_model': 4.751470686019888e-05, 'epoch': 1.11}) +[2026-06-16 00:24:43,240][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.005038488656282425, 'train/grad_norm_pre_clip': 0.4665515422821045, 'timing/data': 0.006424175982829183, 'timing/model': 1.1324698959942907, 'learning_rate/qwen_vl_interface': 4.7484421867191645e-06, 'learning_rate/action_model': 4.748442186719165e-05, 'epoch': 1.12}) +[2026-06-16 00:24:47,652][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.05985609069466591, 'train/grad_norm_pre_clip': 6.506133079528809, 'timing/data': 0.006344435038045049, 'timing/model': 1.1237901920103468, 'learning_rate/qwen_vl_interface': 4.745413893589913e-06, 'learning_rate/action_model': 4.7454138935899125e-05, 'epoch': 1.12}) +[2026-06-16 00:24:52,139][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.13133317232131958, 'train/grad_norm_pre_clip': 9.113579750061035, 'timing/data': 0.0064742169925011694, 'timing/model': 1.1264596999972127, 'learning_rate/qwen_vl_interface': 4.7423858078769506e-06, 'learning_rate/action_model': 4.742385807876951e-05, 'epoch': 1.12}) +[2026-06-16 00:24:56,667][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.009059318341314793, 'train/grad_norm_pre_clip': 1.4848181009292603, 'timing/data': 0.006068473972845823, 'timing/model': 1.1227353169815615, 'learning_rate/qwen_vl_interface': 4.739357930825008e-06, 'learning_rate/action_model': 4.739357930825008e-05, 'epoch': 1.12}) +[2026-06-16 00:25:01,209][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.09706054627895355, 'train/grad_norm_pre_clip': 5.503283977508545, 'timing/data': 0.006333503988571465, 'timing/model': 1.3048234949819744, 'learning_rate/qwen_vl_interface': 4.736330263678731e-06, 'learning_rate/action_model': 4.736330263678731e-05, 'epoch': 1.12}) +[2026-06-16 00:25:05,590][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.0024601230397820473, 'train/grad_norm_pre_clip': 0.14401793479919434, 'timing/data': 0.006260152964387089, 'timing/model': 1.1318628279841505, 'learning_rate/qwen_vl_interface': 4.733302807682679e-06, 'learning_rate/action_model': 4.733302807682679e-05, 'epoch': 1.12}) +[2026-06-16 00:25:10,126][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.2255149781703949, 'train/grad_norm_pre_clip': 12.425822257995605, 'timing/data': 0.0063489029998891056, 'timing/model': 1.1324799919966608, 'learning_rate/qwen_vl_interface': 4.7302755640813255e-06, 'learning_rate/action_model': 4.7302755640813256e-05, 'epoch': 1.12}) +[2026-06-16 00:25:14,625][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.10764563083648682, 'train/grad_norm_pre_clip': 8.859580993652344, 'timing/data': 0.0064339039963670075, 'timing/model': 1.1367398889851756, 'learning_rate/qwen_vl_interface': 4.727248534119058e-06, 'learning_rate/action_model': 4.7272485341190574e-05, 'epoch': 1.12}) +[2026-06-16 00:25:19,150][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.16304446756839752, 'train/grad_norm_pre_clip': 21.524499893188477, 'timing/data': 0.006156478018965572, 'timing/model': 1.127106812025886, 'learning_rate/qwen_vl_interface': 4.72422171904017e-06, 'learning_rate/action_model': 4.7242217190401697e-05, 'epoch': 1.12}) +[2026-06-16 00:25:23,668][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.061739686876535416, 'train/grad_norm_pre_clip': 8.707462310791016, 'timing/data': 0.006136420997790992, 'timing/model': 1.1328866980038583, 'learning_rate/qwen_vl_interface': 4.721195120088873e-06, 'learning_rate/action_model': 4.721195120088873e-05, 'epoch': 1.12}) +[2026-06-16 00:25:28,172][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.10051648318767548, 'train/grad_norm_pre_clip': 7.724021911621094, 'timing/data': 0.006466631020884961, 'timing/model': 1.2718560390057974, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 1.12}) +[2026-06-16 00:25:32,686][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.09637077897787094, 'train/grad_norm_pre_clip': 17.022064208984375, 'timing/data': 0.0064757909858599305, 'timing/model': 1.2464749609935097, 'learning_rate/qwen_vl_interface': 4.715142575545445e-06, 'learning_rate/action_model': 4.7151425755454454e-05, 'epoch': 1.12}) +[2026-06-16 00:25:37,047][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.157907634973526, 'train/grad_norm_pre_clip': 10.311320304870605, 'timing/data': 0.006568673998117447, 'timing/model': 1.1289207969675772, 'learning_rate/qwen_vl_interface': 4.712116632441287e-06, 'learning_rate/action_model': 4.7121166324412866e-05, 'epoch': 1.12}) +[2026-06-16 00:25:41,557][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.009352266788482666, 'train/grad_norm_pre_clip': 0.6158310770988464, 'timing/data': 0.006439562945161015, 'timing/model': 1.1326446749735624, 'learning_rate/qwen_vl_interface': 4.709090910440664e-06, 'learning_rate/action_model': 4.7090909104406634e-05, 'epoch': 1.12}) +[2026-06-16 00:25:46,092][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.025642063468694687, 'train/grad_norm_pre_clip': 5.285621166229248, 'timing/data': 0.006495882000308484, 'timing/model': 1.136573154013604, 'learning_rate/qwen_vl_interface': 4.7060654107873335e-06, 'learning_rate/action_model': 4.7060654107873335e-05, 'epoch': 1.12}) +[2026-06-16 00:25:50,633][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.01641703210771084, 'train/grad_norm_pre_clip': 2.287220001220703, 'timing/data': 0.006294147053267807, 'timing/model': 1.1311328419833444, 'learning_rate/qwen_vl_interface': 4.7030401347249685e-06, 'learning_rate/action_model': 4.703040134724968e-05, 'epoch': 1.12}) +[2026-06-16 00:25:55,186][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.11929860711097717, 'train/grad_norm_pre_clip': 12.569037437438965, 'timing/data': 0.00624733604490757, 'timing/model': 1.3051205850206316, 'learning_rate/qwen_vl_interface': 4.700015083497143e-06, 'learning_rate/action_model': 4.700015083497142e-05, 'epoch': 1.12}) +[2026-06-16 00:25:59,569][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.08578341454267502, 'train/grad_norm_pre_clip': 5.535277366638184, 'timing/data': 0.006029921991284937, 'timing/model': 1.1351067149662413, 'learning_rate/qwen_vl_interface': 4.696990258347342e-06, 'learning_rate/action_model': 4.696990258347342e-05, 'epoch': 1.12}) +[2026-06-16 00:26:04,129][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.15826132893562317, 'train/grad_norm_pre_clip': 7.880591869354248, 'timing/data': 0.006653479009401053, 'timing/model': 1.1394039670121856, 'learning_rate/qwen_vl_interface': 4.693965660518957e-06, 'learning_rate/action_model': 4.693965660518957e-05, 'epoch': 1.12}) +[2026-06-16 00:26:08,672][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.13155780732631683, 'train/grad_norm_pre_clip': 11.03198528289795, 'timing/data': 0.006462869001552463, 'timing/model': 1.135807921004016, 'learning_rate/qwen_vl_interface': 4.690941291255285e-06, 'learning_rate/action_model': 4.6909412912552844e-05, 'epoch': 1.12}) +[2026-06-16 00:26:13,212][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.06660416722297668, 'train/grad_norm_pre_clip': 10.543052673339844, 'timing/data': 0.006748130021151155, 'timing/model': 1.1293052810360678, 'learning_rate/qwen_vl_interface': 4.68791715179953e-06, 'learning_rate/action_model': 4.68791715179953e-05, 'epoch': 1.12}) +[2026-06-16 00:26:17,686][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.35106831789016724, 'train/grad_norm_pre_clip': 21.56622314453125, 'timing/data': 0.00606164603959769, 'timing/model': 1.1314685679972172, 'learning_rate/qwen_vl_interface': 4.684893243394804e-06, 'learning_rate/action_model': 4.6848932433948036e-05, 'epoch': 1.12}) +[2026-06-16 00:26:22,173][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.011890971101820469, 'train/grad_norm_pre_clip': 1.5283085107803345, 'timing/data': 0.006678519013803452, 'timing/model': 1.236638070025947, 'learning_rate/qwen_vl_interface': 4.68186956728412e-06, 'learning_rate/action_model': 4.68186956728412e-05, 'epoch': 1.12}) +[2026-06-16 00:26:26,546][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.004468047991394997, 'train/grad_norm_pre_clip': 0.6184051036834717, 'timing/data': 0.006674067990388721, 'timing/model': 1.130679014022462, 'learning_rate/qwen_vl_interface': 4.678846124710395e-06, 'learning_rate/action_model': 4.678846124710394e-05, 'epoch': 1.12}) +[2026-06-16 00:26:31,073][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.2590373754501343, 'train/grad_norm_pre_clip': 20.406841278076172, 'timing/data': 0.006188492989167571, 'timing/model': 1.131050128955394, 'learning_rate/qwen_vl_interface': 4.675822916916454e-06, 'learning_rate/action_model': 4.6758229169164536e-05, 'epoch': 1.13}) +[2026-06-16 00:26:35,616][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.11214158684015274, 'train/grad_norm_pre_clip': 11.904972076416016, 'timing/data': 0.006504151038825512, 'timing/model': 1.1316954789799638, 'learning_rate/qwen_vl_interface': 4.6727999451450236e-06, 'learning_rate/action_model': 4.672799945145023e-05, 'epoch': 1.13}) +[2026-06-16 00:26:40,170][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.06665507704019547, 'train/grad_norm_pre_clip': 5.36833381652832, 'timing/data': 0.006544485047925264, 'timing/model': 1.144414046953898, 'learning_rate/qwen_vl_interface': 4.669777210638733e-06, 'learning_rate/action_model': 4.669777210638733e-05, 'epoch': 1.13}) +[2026-06-16 00:26:44,664][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.025555942207574844, 'train/grad_norm_pre_clip': 5.7612714767456055, 'timing/data': 0.006118627963587642, 'timing/model': 1.1306532949674875, 'learning_rate/qwen_vl_interface': 4.666754714640115e-06, 'learning_rate/action_model': 4.666754714640115e-05, 'epoch': 1.13}) +[2026-06-16 00:26:49,166][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.06089327484369278, 'train/grad_norm_pre_clip': 7.03741979598999, 'timing/data': 0.0064199710031971335, 'timing/model': 1.2571708349860273, 'learning_rate/qwen_vl_interface': 4.663732458391602e-06, 'learning_rate/action_model': 4.663732458391601e-05, 'epoch': 1.13}) +[2026-06-16 00:26:53,534][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.003995456267148256, 'train/grad_norm_pre_clip': 0.5376565456390381, 'timing/data': 0.006473347020801157, 'timing/model': 1.1301268989918754, 'learning_rate/qwen_vl_interface': 4.6607104431355295e-06, 'learning_rate/action_model': 4.66071044313553e-05, 'epoch': 1.13}) +[2026-06-16 00:26:58,063][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.01618378609418869, 'train/grad_norm_pre_clip': 5.744747161865234, 'timing/data': 0.006572714017238468, 'timing/model': 1.132793259981554, 'learning_rate/qwen_vl_interface': 4.657688670114135e-06, 'learning_rate/action_model': 4.657688670114135e-05, 'epoch': 1.13}) +[2026-06-16 00:27:02,597][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.08797310292720795, 'train/grad_norm_pre_clip': 15.63364028930664, 'timing/data': 0.0064375679939985275, 'timing/model': 1.1309217790258117, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 1.13}) +[2026-06-16 00:27:07,133][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.07307136058807373, 'train/grad_norm_pre_clip': 8.079645156860352, 'timing/data': 0.006277511012740433, 'timing/model': 1.134776177990716, 'learning_rate/qwen_vl_interface': 4.651645855743825e-06, 'learning_rate/action_model': 4.651645855743825e-05, 'epoch': 1.13}) +[2026-06-16 00:27:11,671][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.17731061577796936, 'train/grad_norm_pre_clip': 15.564597129821777, 'timing/data': 0.006544537027366459, 'timing/model': 1.2863794610020705, 'learning_rate/qwen_vl_interface': 4.648624816878883e-06, 'learning_rate/action_model': 4.6486248168788825e-05, 'epoch': 1.13}) +[2026-06-16 00:27:16,106][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.0346313901245594, 'train/grad_norm_pre_clip': 4.448459148406982, 'timing/data': 0.006441239034757018, 'timing/model': 1.1892973720096052, 'learning_rate/qwen_vl_interface': 4.6456040252165615e-06, 'learning_rate/action_model': 4.645604025216562e-05, 'epoch': 1.13}) +[2026-06-16 00:27:20,490][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.14027711749076843, 'train/grad_norm_pre_clip': 8.135600090026855, 'timing/data': 0.006363757012877613, 'timing/model': 1.1321862500044517, 'learning_rate/qwen_vl_interface': 4.642583481998596e-06, 'learning_rate/action_model': 4.642583481998596e-05, 'epoch': 1.13}) +[2026-06-16 00:27:25,036][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.10823145508766174, 'train/grad_norm_pre_clip': 11.960105895996094, 'timing/data': 0.006801665003877133, 'timing/model': 1.1291358800372109, 'learning_rate/qwen_vl_interface': 4.639563188466616e-06, 'learning_rate/action_model': 4.6395631884666165e-05, 'epoch': 1.13}) +[2026-06-16 00:27:29,559][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.14001591503620148, 'train/grad_norm_pre_clip': 8.332340240478516, 'timing/data': 0.006541471986565739, 'timing/model': 1.1323097039712593, 'learning_rate/qwen_vl_interface': 4.636543145862152e-06, 'learning_rate/action_model': 4.6365431458621523e-05, 'epoch': 1.13}) +[2026-06-16 00:27:34,093][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.29303738474845886, 'train/grad_norm_pre_clip': 12.62136459350586, 'timing/data': 0.006904293957632035, 'timing/model': 1.1369932509842329, 'learning_rate/qwen_vl_interface': 4.633523355426629e-06, 'learning_rate/action_model': 4.6335233554266295e-05, 'epoch': 1.13}) +[2026-06-16 00:27:38,590][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.1688718944787979, 'train/grad_norm_pre_clip': 13.870479583740234, 'timing/data': 0.006239201000425965, 'timing/model': 1.1334248849889264, 'learning_rate/qwen_vl_interface': 4.6305038184013675e-06, 'learning_rate/action_model': 4.6305038184013675e-05, 'epoch': 1.13}) +[2026-06-16 00:27:43,108][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.047560930252075195, 'train/grad_norm_pre_clip': 5.655628681182861, 'timing/data': 0.0063876110361889005, 'timing/model': 1.2766742670210078, 'learning_rate/qwen_vl_interface': 4.627484536027589e-06, 'learning_rate/action_model': 4.627484536027588e-05, 'epoch': 1.13}) +[2026-06-16 00:27:47,489][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.05853761360049248, 'train/grad_norm_pre_clip': 9.81885051727295, 'timing/data': 0.0057496330118738115, 'timing/model': 1.1322863569948822, 'learning_rate/qwen_vl_interface': 4.624465509546399e-06, 'learning_rate/action_model': 4.6244655095463986e-05, 'epoch': 1.13}) +[2026-06-16 00:27:52,019][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.05578266829252243, 'train/grad_norm_pre_clip': 13.603577613830566, 'timing/data': 0.00638148799771443, 'timing/model': 1.137790907989256, 'learning_rate/qwen_vl_interface': 4.6214467401988106e-06, 'learning_rate/action_model': 4.621446740198811e-05, 'epoch': 1.13}) +[2026-06-16 00:27:56,555][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.10968805849552155, 'train/grad_norm_pre_clip': 8.534711837768555, 'timing/data': 0.006181251956149936, 'timing/model': 1.1293453740072437, 'learning_rate/qwen_vl_interface': 4.6184282292257245e-06, 'learning_rate/action_model': 4.618428229225725e-05, 'epoch': 1.13}) +[2026-06-16 00:28:01,092][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.08338866382837296, 'train/grad_norm_pre_clip': 10.690351486206055, 'timing/data': 0.006696992029901594, 'timing/model': 1.1425998389604501, 'learning_rate/qwen_vl_interface': 4.615409977867938e-06, 'learning_rate/action_model': 4.615409977867937e-05, 'epoch': 1.13}) +[2026-06-16 00:28:05,617][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.020608339458703995, 'train/grad_norm_pre_clip': 2.9833950996398926, 'timing/data': 0.006978921010158956, 'timing/model': 1.1406077310093679, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 1.13}) +[2026-06-16 00:28:10,156][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.2840210199356079, 'train/grad_norm_pre_clip': 24.04059410095215, 'timing/data': 0.006546640011947602, 'timing/model': 1.293736525985878, 'learning_rate/qwen_vl_interface': 4.609374258960908e-06, 'learning_rate/action_model': 4.6093742589609074e-05, 'epoch': 1.13}) +[2026-06-16 00:28:14,542][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.03371202573180199, 'train/grad_norm_pre_clip': 4.834240436553955, 'timing/data': 0.006286908988840878, 'timing/model': 1.1380548710003495, 'learning_rate/qwen_vl_interface': 4.60635679389272e-06, 'learning_rate/action_model': 4.60635679389272e-05, 'epoch': 1.13}) +[2026-06-16 00:28:19,088][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.16457587480545044, 'train/grad_norm_pre_clip': 13.805137634277344, 'timing/data': 0.006573883001692593, 'timing/model': 1.129884630034212, 'learning_rate/qwen_vl_interface': 4.603339593401943e-06, 'learning_rate/action_model': 4.603339593401942e-05, 'epoch': 1.14}) +[2026-06-16 00:28:23,629][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.03622165694832802, 'train/grad_norm_pre_clip': 6.464591026306152, 'timing/data': 0.006512050982564688, 'timing/model': 1.1384781980304979, 'learning_rate/qwen_vl_interface': 4.600322658728832e-06, 'learning_rate/action_model': 4.600322658728832e-05, 'epoch': 1.14}) +[2026-06-16 00:28:28,167][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.07682786881923676, 'train/grad_norm_pre_clip': 5.9416022300720215, 'timing/data': 0.00616559898480773, 'timing/model': 1.1439199869637378, 'learning_rate/qwen_vl_interface': 4.597305991113537e-06, 'learning_rate/action_model': 4.597305991113537e-05, 'epoch': 1.14}) +[2026-06-16 00:28:32,684][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.028572142124176025, 'train/grad_norm_pre_clip': 9.893808364868164, 'timing/data': 0.00648732902482152, 'timing/model': 1.130430362012703, 'learning_rate/qwen_vl_interface': 4.5942895917960935e-06, 'learning_rate/action_model': 4.5942895917960935e-05, 'epoch': 1.14}) +[2026-06-16 00:28:37,219][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.20358535647392273, 'train/grad_norm_pre_clip': 12.880034446716309, 'timing/data': 0.005919904040638357, 'timing/model': 1.3004303440102376, 'learning_rate/qwen_vl_interface': 4.59127346201643e-06, 'learning_rate/action_model': 4.59127346201643e-05, 'epoch': 1.14}) +[2026-06-16 00:28:41,598][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.1673164665699005, 'train/grad_norm_pre_clip': 14.791549682617188, 'timing/data': 0.006601546017918736, 'timing/model': 1.1327680800459348, 'learning_rate/qwen_vl_interface': 4.588257603014364e-06, 'learning_rate/action_model': 4.588257603014364e-05, 'epoch': 1.14}) +[2026-06-16 00:28:46,136][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.11432167142629623, 'train/grad_norm_pre_clip': 9.687691688537598, 'timing/data': 0.006415397976525128, 'timing/model': 1.1338649149984121, 'learning_rate/qwen_vl_interface': 4.585242016029602e-06, 'learning_rate/action_model': 4.5852420160296015e-05, 'epoch': 1.14}) +[2026-06-16 00:28:50,667][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.03205953538417816, 'train/grad_norm_pre_clip': 4.864104270935059, 'timing/data': 0.006285119976382703, 'timing/model': 1.133603852998931, 'learning_rate/qwen_vl_interface': 4.582226702301737e-06, 'learning_rate/action_model': 4.5822267023017375e-05, 'epoch': 1.14}) +[2026-06-16 00:28:55,154][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.03650461882352829, 'train/grad_norm_pre_clip': 6.259861469268799, 'timing/data': 0.0066810319549404085, 'timing/model': 1.1347847260185517, 'learning_rate/qwen_vl_interface': 4.57921166307025e-06, 'learning_rate/action_model': 4.57921166307025e-05, 'epoch': 1.14}) +[2026-06-16 00:28:59,682][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.018136220052838326, 'train/grad_norm_pre_clip': 4.23115873336792, 'timing/data': 0.006335359998047352, 'timing/model': 1.128379660018254, 'learning_rate/qwen_vl_interface': 4.576196899574511e-06, 'learning_rate/action_model': 4.576196899574511e-05, 'epoch': 1.14}) +[2026-06-16 00:29:04,218][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.009567895904183388, 'train/grad_norm_pre_clip': 1.129151463508606, 'timing/data': 0.006042451015673578, 'timing/model': 1.2937227920047008, 'learning_rate/qwen_vl_interface': 4.573182413053775e-06, 'learning_rate/action_model': 4.5731824130537746e-05, 'epoch': 1.14}) +[2026-06-16 00:29:08,597][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.11168210953474045, 'train/grad_norm_pre_clip': 5.445433139801025, 'timing/data': 0.00642888096626848, 'timing/model': 1.1333400130388327, 'learning_rate/qwen_vl_interface': 4.570168204747183e-06, 'learning_rate/action_model': 4.570168204747183e-05, 'epoch': 1.14}) +[2026-06-16 00:29:13,116][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.06991493701934814, 'train/grad_norm_pre_clip': 15.257733345031738, 'timing/data': 0.006441327976062894, 'timing/model': 1.1368373789591715, 'learning_rate/qwen_vl_interface': 4.567154275893762e-06, 'learning_rate/action_model': 4.5671542758937624e-05, 'epoch': 1.14}) +[2026-06-16 00:29:17,589][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.04690258949995041, 'train/grad_norm_pre_clip': 6.596973896026611, 'timing/data': 0.006381251034326851, 'timing/model': 1.1304173050448298, 'learning_rate/qwen_vl_interface': 4.564140627732426e-06, 'learning_rate/action_model': 4.564140627732426e-05, 'epoch': 1.14}) +[2026-06-16 00:29:22,088][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.09934971481561661, 'train/grad_norm_pre_clip': 6.5818891525268555, 'timing/data': 0.006763115990906954, 'timing/model': 1.1324963319930248, 'learning_rate/qwen_vl_interface': 4.56112726150197e-06, 'learning_rate/action_model': 4.56112726150197e-05, 'epoch': 1.14}) +[2026-06-16 00:29:26,630][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.12116982042789459, 'train/grad_norm_pre_clip': 10.50001335144043, 'timing/data': 0.006700776983052492, 'timing/model': 1.1327707070158795, 'learning_rate/qwen_vl_interface': 4.558114178441076e-06, 'learning_rate/action_model': 4.558114178441076e-05, 'epoch': 1.14}) +[2026-06-16 00:29:31,175][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.022291598841547966, 'train/grad_norm_pre_clip': 6.231765270233154, 'timing/data': 0.006223308970220387, 'timing/model': 1.292250205995515, 'learning_rate/qwen_vl_interface': 4.555101379788308e-06, 'learning_rate/action_model': 4.555101379788308e-05, 'epoch': 1.14}) +[2026-06-16 00:29:35,552][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.03105580434203148, 'train/grad_norm_pre_clip': 7.7057013511657715, 'timing/data': 0.006592935009393841, 'timing/model': 1.130623513017781, 'learning_rate/qwen_vl_interface': 4.552088866782115e-06, 'learning_rate/action_model': 4.552088866782115e-05, 'epoch': 1.14}) +[2026-06-16 00:29:40,080][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.10374698787927628, 'train/grad_norm_pre_clip': 11.903298377990723, 'timing/data': 0.006340748048387468, 'timing/model': 1.1324523360235617, 'learning_rate/qwen_vl_interface': 4.5490766406608265e-06, 'learning_rate/action_model': 4.549076640660827e-05, 'epoch': 1.14}) +[2026-06-16 00:29:44,575][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.01444022823125124, 'train/grad_norm_pre_clip': 3.069843053817749, 'timing/data': 0.006323545007035136, 'timing/model': 1.1295693219872192, 'learning_rate/qwen_vl_interface': 4.546064702662655e-06, 'learning_rate/action_model': 4.546064702662655e-05, 'epoch': 1.14}) +[2026-06-16 00:29:49,070][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.1902112513780594, 'train/grad_norm_pre_clip': 17.453800201416016, 'timing/data': 0.005915050976909697, 'timing/model': 1.1328457790077664, 'learning_rate/qwen_vl_interface': 4.543053054025693e-06, 'learning_rate/action_model': 4.5430530540256926e-05, 'epoch': 1.14}) +[2026-06-16 00:29:53,597][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.011961003765463829, 'train/grad_norm_pre_clip': 1.3684031963348389, 'timing/data': 0.006470295018516481, 'timing/model': 1.1349303380120546, 'learning_rate/qwen_vl_interface': 4.540041695987917e-06, 'learning_rate/action_model': 4.540041695987917e-05, 'epoch': 1.14}) +[2026-06-16 00:29:58,128][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.022019384428858757, 'train/grad_norm_pre_clip': 3.0406205654144287, 'timing/data': 0.006572771992068738, 'timing/model': 1.2879086249740794, 'learning_rate/qwen_vl_interface': 4.537030629787182e-06, 'learning_rate/action_model': 4.5370306297871826e-05, 'epoch': 1.14}) +[2026-06-16 00:30:02,521][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.023959770798683167, 'train/grad_norm_pre_clip': 2.846567153930664, 'timing/data': 0.008024031994864345, 'timing/model': 1.1370733880321495, 'learning_rate/qwen_vl_interface': 4.534019856661226e-06, 'learning_rate/action_model': 4.534019856661226e-05, 'epoch': 1.14}) +[2026-06-16 00:30:07,062][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.01799892447888851, 'train/grad_norm_pre_clip': 5.177070617675781, 'timing/data': 0.006692980998195708, 'timing/model': 1.1331412050058134, 'learning_rate/qwen_vl_interface': 4.531009377847661e-06, 'learning_rate/action_model': 4.531009377847661e-05, 'epoch': 1.14}) +[2026-06-16 00:30:11,553][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.16618941724300385, 'train/grad_norm_pre_clip': 9.11007308959961, 'timing/data': 0.006413250986952335, 'timing/model': 1.1317551330430433, 'learning_rate/qwen_vl_interface': 4.527999194583982e-06, 'learning_rate/action_model': 4.527999194583981e-05, 'epoch': 1.15}) +[2026-06-16 00:30:16,089][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.03356998786330223, 'train/grad_norm_pre_clip': 4.677145957946777, 'timing/data': 0.006233593972865492, 'timing/model': 1.131804485048633, 'learning_rate/qwen_vl_interface': 4.52498930810756e-06, 'learning_rate/action_model': 4.52498930810756e-05, 'epoch': 1.15}) +[2026-06-16 00:30:20,636][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.22339481115341187, 'train/grad_norm_pre_clip': 17.520668029785156, 'timing/data': 0.006872831028886139, 'timing/model': 1.12989582400769, 'learning_rate/qwen_vl_interface': 4.521979719655648e-06, 'learning_rate/action_model': 4.5219797196556484e-05, 'epoch': 1.15}) +[2026-06-16 00:30:25,172][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.028164615854620934, 'train/grad_norm_pre_clip': 4.128340244293213, 'timing/data': 0.006476476031821221, 'timing/model': 1.2889147329842672, 'learning_rate/qwen_vl_interface': 4.518970430465377e-06, 'learning_rate/action_model': 4.518970430465376e-05, 'epoch': 1.15}) +[2026-06-16 00:30:29,648][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.15932387113571167, 'train/grad_norm_pre_clip': 15.01064395904541, 'timing/data': 0.0061600200133398175, 'timing/model': 1.222079731989652, 'learning_rate/qwen_vl_interface': 4.515961441773745e-06, 'learning_rate/action_model': 4.515961441773745e-05, 'epoch': 1.15}) +[2026-06-16 00:30:34,040][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.030501533299684525, 'train/grad_norm_pre_clip': 6.678060531616211, 'timing/data': 0.006392805022187531, 'timing/model': 1.1406109139788896, 'learning_rate/qwen_vl_interface': 4.5129527548176385e-06, 'learning_rate/action_model': 4.512952754817638e-05, 'epoch': 1.15}) +[2026-06-16 00:30:38,562][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.1563580185174942, 'train/grad_norm_pre_clip': 11.093594551086426, 'timing/data': 0.006054620957002044, 'timing/model': 1.1395942329545505, 'learning_rate/qwen_vl_interface': 4.509944370833813e-06, 'learning_rate/action_model': 4.509944370833813e-05, 'epoch': 1.15}) +[2026-06-16 00:30:43,123][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.003734264289960265, 'train/grad_norm_pre_clip': 0.47032085061073303, 'timing/data': 0.006533072970341891, 'timing/model': 1.1304948199540377, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 1.15}) +[2026-06-16 00:30:47,662][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.039731841534376144, 'train/grad_norm_pre_clip': 11.652400970458984, 'timing/data': 0.006533283973112702, 'timing/model': 1.1336460470338352, 'learning_rate/qwen_vl_interface': 4.503928516729416e-06, 'learning_rate/action_model': 4.503928516729416e-05, 'epoch': 1.15}) +[2026-06-16 00:30:52,216][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.011647159233689308, 'train/grad_norm_pre_clip': 1.848548412322998, 'timing/data': 0.00643981498433277, 'timing/model': 1.300424747983925, 'learning_rate/qwen_vl_interface': 4.500921049081734e-06, 'learning_rate/action_model': 4.500921049081734e-05, 'epoch': 1.15}) +[2026-06-16 00:30:56,722][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.009186161682009697, 'train/grad_norm_pre_clip': 1.1240904331207275, 'timing/data': 0.006068607035558671, 'timing/model': 1.2438371939933859, 'learning_rate/qwen_vl_interface': 4.497913889352113e-06, 'learning_rate/action_model': 4.4979138893521126e-05, 'epoch': 1.15}) +[2026-06-16 00:31:01,115][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.1242688000202179, 'train/grad_norm_pre_clip': 11.098627090454102, 'timing/data': 0.00626980303786695, 'timing/model': 1.1385683209518902, 'learning_rate/qwen_vl_interface': 4.494907038776684e-06, 'learning_rate/action_model': 4.4949070387766836e-05, 'epoch': 1.15}) +[2026-06-16 00:35:16,449][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.07361911982297897, 'train/grad_norm_pre_clip': 12.73766040802002, 'eval/loss': 0.04312891513109207, 'eval/flap_precision': 0.9662058371735791, 'eval/flap_recall': 0.970679012345679, 'eval/flap_f1': 0.968437259430331, 'eval/flap_rate_model': 0.18083333333333335, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9684372594303309, 'timing/data': 0.006470797990914434, 'timing/model': 1.1282757690059952, 'learning_rate/qwen_vl_interface': 4.49190049859145e-06, 'learning_rate/action_model': 4.49190049859145e-05, 'epoch': 1.15}) +[2026-06-16 00:35:16,450][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2800_state +[2026-06-16 00:35:16,452][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 00:35:20,023][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2800_state/model.safetensors +[2026-06-16 00:35:30,065][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2800_state/optimizer.bin +[2026-06-16 00:35:30,065][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2800_state/sampler.bin +[2026-06-16 00:35:30,066][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2800_state/sampler_1.bin +[2026-06-16 00:35:30,066][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2800_state/random_states_0.pkl +[2026-06-16 00:35:30,114][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2800 +[2026-06-16 00:35:30,115][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 00:35:30,121][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 00:35:34,571][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.09071384370326996, 'train/grad_norm_pre_clip': 11.116089820861816, 'timing/data': 0.006532502011395991, 'timing/model': 1.1092503279796802, 'learning_rate/qwen_vl_interface': 4.488894270032284e-06, 'learning_rate/action_model': 4.4888942700322834e-05, 'epoch': 1.15}) +[2026-06-16 00:35:39,035][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.006404959596693516, 'train/grad_norm_pre_clip': 0.8335036635398865, 'timing/data': 0.006415283016394824, 'timing/model': 1.112409915018361, 'learning_rate/qwen_vl_interface': 4.485888354334933e-06, 'learning_rate/action_model': 4.4858883543349334e-05, 'epoch': 1.15}) +[2026-06-16 00:35:43,434][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.11517678201198578, 'train/grad_norm_pre_clip': 30.506338119506836, 'timing/data': 0.006566727010067552, 'timing/model': 1.1248810910037719, 'learning_rate/qwen_vl_interface': 4.482882752735017e-06, 'learning_rate/action_model': 4.482882752735017e-05, 'epoch': 1.15}) +[2026-06-16 00:35:47,841][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.0762878879904747, 'train/grad_norm_pre_clip': 19.112319946289062, 'timing/data': 0.006516663008369505, 'timing/model': 1.1289716560277157, 'learning_rate/qwen_vl_interface': 4.479877466468027e-06, 'learning_rate/action_model': 4.479877466468026e-05, 'epoch': 1.15}) +[2026-06-16 00:35:52,294][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.03817391023039818, 'train/grad_norm_pre_clip': 13.137935638427734, 'timing/data': 0.005912230990361422, 'timing/model': 1.2408834759844467, 'learning_rate/qwen_vl_interface': 4.476872496769318e-06, 'learning_rate/action_model': 4.476872496769318e-05, 'epoch': 1.15}) +[2026-06-16 00:35:56,625][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.02646496705710888, 'train/grad_norm_pre_clip': 12.239097595214844, 'timing/data': 0.006590868986677378, 'timing/model': 1.1215495700016618, 'learning_rate/qwen_vl_interface': 4.473867844874122e-06, 'learning_rate/action_model': 4.473867844874122e-05, 'epoch': 1.15}) +[2026-06-16 00:36:01,096][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.22887298464775085, 'train/grad_norm_pre_clip': 26.225677490234375, 'timing/data': 0.006227898993529379, 'timing/model': 1.1230138699756935, 'learning_rate/qwen_vl_interface': 4.470863512017539e-06, 'learning_rate/action_model': 4.470863512017539e-05, 'epoch': 1.15}) +[2026-06-16 00:36:05,632][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.057726528495550156, 'train/grad_norm_pre_clip': 11.557511329650879, 'timing/data': 0.006191569031216204, 'timing/model': 1.1346790310344659, 'learning_rate/qwen_vl_interface': 4.4678594994345344e-06, 'learning_rate/action_model': 4.467859499434534e-05, 'epoch': 1.15}) +[2026-06-16 00:36:10,127][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.045031629502773285, 'train/grad_norm_pre_clip': 10.740715980529785, 'timing/data': 0.006355360965244472, 'timing/model': 1.1226997959893197, 'learning_rate/qwen_vl_interface': 4.4648558083599475e-06, 'learning_rate/action_model': 4.464855808359947e-05, 'epoch': 1.15}) +[2026-06-16 00:36:14,618][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.07993060350418091, 'train/grad_norm_pre_clip': 10.394415855407715, 'timing/data': 0.0062112449668347836, 'timing/model': 1.278965211007744, 'learning_rate/qwen_vl_interface': 4.461852440028479e-06, 'learning_rate/action_model': 4.461852440028479e-05, 'epoch': 1.15}) +[2026-06-16 00:36:19,093][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.006537445355206728, 'train/grad_norm_pre_clip': 1.7617191076278687, 'timing/data': 0.006134369992651045, 'timing/model': 1.2517461869865656, 'learning_rate/qwen_vl_interface': 4.458849395674702e-06, 'learning_rate/action_model': 4.458849395674701e-05, 'epoch': 1.15}) +[2026-06-16 00:36:23,445][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.07610628008842468, 'train/grad_norm_pre_clip': 31.74820899963379, 'timing/data': 0.006141951016616076, 'timing/model': 1.1281959699699655, 'learning_rate/qwen_vl_interface': 4.455846676533052e-06, 'learning_rate/action_model': 4.4558466765330525e-05, 'epoch': 1.16}) +[2026-06-16 00:36:27,980][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.008585942909121513, 'train/grad_norm_pre_clip': 4.145317077636719, 'timing/data': 0.006318205967545509, 'timing/model': 1.1399084950098768, 'learning_rate/qwen_vl_interface': 4.452844283837838e-06, 'learning_rate/action_model': 4.452844283837838e-05, 'epoch': 1.16}) +[2026-06-16 00:36:32,490][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.01158731710165739, 'train/grad_norm_pre_clip': 1.7516028881072998, 'timing/data': 0.006162729987408966, 'timing/model': 1.130413757986389, 'learning_rate/qwen_vl_interface': 4.449842218823228e-06, 'learning_rate/action_model': 4.449842218823228e-05, 'epoch': 1.16}) +[2026-06-16 00:36:37,019][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.01013818196952343, 'train/grad_norm_pre_clip': 3.619205951690674, 'timing/data': 0.006229726015590131, 'timing/model': 1.1404330449877307, 'learning_rate/qwen_vl_interface': 4.446840482723256e-06, 'learning_rate/action_model': 4.446840482723255e-05, 'epoch': 1.16}) +[2026-06-16 00:36:41,497][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.027108825743198395, 'train/grad_norm_pre_clip': 8.158106803894043, 'timing/data': 0.006279597990214825, 'timing/model': 1.1258895060163923, 'learning_rate/qwen_vl_interface': 4.443839076771823e-06, 'learning_rate/action_model': 4.443839076771823e-05, 'epoch': 1.16}) +[2026-06-16 00:36:45,985][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.09852635115385056, 'train/grad_norm_pre_clip': 18.6595516204834, 'timing/data': 0.006177594943437725, 'timing/model': 1.2567285770201124, 'learning_rate/qwen_vl_interface': 4.4408380022026945e-06, 'learning_rate/action_model': 4.440838002202694e-05, 'epoch': 1.16}) +[2026-06-16 00:36:50,346][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.06647901237010956, 'train/grad_norm_pre_clip': 8.298707962036133, 'timing/data': 0.006201373995281756, 'timing/model': 1.1274363989941776, 'learning_rate/qwen_vl_interface': 4.437837260249499e-06, 'learning_rate/action_model': 4.437837260249499e-05, 'epoch': 1.16}) +[2026-06-16 00:36:54,855][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.003966003656387329, 'train/grad_norm_pre_clip': 0.618305504322052, 'timing/data': 0.006509314000140876, 'timing/model': 1.1323205820517614, 'learning_rate/qwen_vl_interface': 4.434836852145728e-06, 'learning_rate/action_model': 4.434836852145728e-05, 'epoch': 1.16}) +[2026-06-16 00:36:59,394][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.07783215492963791, 'train/grad_norm_pre_clip': 13.900249481201172, 'timing/data': 0.006575658975634724, 'timing/model': 1.1327765970490873, 'learning_rate/qwen_vl_interface': 4.431836779124736e-06, 'learning_rate/action_model': 4.431836779124735e-05, 'epoch': 1.16}) +[2026-06-16 00:37:03,927][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.0369219109416008, 'train/grad_norm_pre_clip': 16.040241241455078, 'timing/data': 0.006079401005990803, 'timing/model': 1.1294193679932505, 'learning_rate/qwen_vl_interface': 4.428837042419738e-06, 'learning_rate/action_model': 4.428837042419738e-05, 'epoch': 1.16}) +[2026-06-16 00:37:08,467][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.0761028304696083, 'train/grad_norm_pre_clip': 13.675284385681152, 'timing/data': 0.006666002969723195, 'timing/model': 1.3003091160207987, 'learning_rate/qwen_vl_interface': 4.425837643263815e-06, 'learning_rate/action_model': 4.425837643263814e-05, 'epoch': 1.16}) +[2026-06-16 00:37:12,826][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.03799234703183174, 'train/grad_norm_pre_clip': 16.57166862487793, 'timing/data': 0.006471180007793009, 'timing/model': 1.131052056036424, 'learning_rate/qwen_vl_interface': 4.4228385828899045e-06, 'learning_rate/action_model': 4.4228385828899046e-05, 'epoch': 1.16}) +[2026-06-16 00:37:17,359][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.09507230669260025, 'train/grad_norm_pre_clip': 9.564044952392578, 'timing/data': 0.006596168037503958, 'timing/model': 1.1394785449956544, 'learning_rate/qwen_vl_interface': 4.419839862530809e-06, 'learning_rate/action_model': 4.419839862530809e-05, 'epoch': 1.16}) +[2026-06-16 00:37:21,797][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.17483104765415192, 'train/grad_norm_pre_clip': 15.668182373046875, 'timing/data': 0.006390049995388836, 'timing/model': 1.1353957160026766, 'learning_rate/qwen_vl_interface': 4.416841483419188e-06, 'learning_rate/action_model': 4.416841483419188e-05, 'epoch': 1.16}) +[2026-06-16 00:37:26,262][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.042909637093544006, 'train/grad_norm_pre_clip': 19.441638946533203, 'timing/data': 0.006578488042578101, 'timing/model': 1.1346452460275032, 'learning_rate/qwen_vl_interface': 4.413843446787561e-06, 'learning_rate/action_model': 4.413843446787561e-05, 'epoch': 1.16}) +[2026-06-16 00:37:30,746][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.005085056647658348, 'train/grad_norm_pre_clip': 1.4907019138336182, 'timing/data': 0.0064481430454179645, 'timing/model': 1.1330832789535634, 'learning_rate/qwen_vl_interface': 4.41084575386831e-06, 'learning_rate/action_model': 4.4108457538683093e-05, 'epoch': 1.16}) +[2026-06-16 00:37:35,277][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.06271357834339142, 'train/grad_norm_pre_clip': 7.058898448944092, 'timing/data': 0.006116607983130962, 'timing/model': 1.2930450529675, 'learning_rate/qwen_vl_interface': 4.40784840589367e-06, 'learning_rate/action_model': 4.4078484058936706e-05, 'epoch': 1.16}) +[2026-06-16 00:37:39,663][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.0053703091107308865, 'train/grad_norm_pre_clip': 0.929719090461731, 'timing/data': 0.006670447997748852, 'timing/model': 1.1361593600013293, 'learning_rate/qwen_vl_interface': 4.404851404095742e-06, 'learning_rate/action_model': 4.404851404095742e-05, 'epoch': 1.16}) +[2026-06-16 00:37:44,214][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.003206831170246005, 'train/grad_norm_pre_clip': 0.811947762966156, 'timing/data': 0.006384795007761568, 'timing/model': 1.1430710509885103, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 1.16}) +[2026-06-16 00:37:48,749][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.2637563645839691, 'train/grad_norm_pre_clip': 22.361045837402344, 'timing/data': 0.00657484499970451, 'timing/model': 1.130733951984439, 'learning_rate/qwen_vl_interface': 4.398858443957686e-06, 'learning_rate/action_model': 4.398858443957686e-05, 'epoch': 1.16}) +[2026-06-16 00:37:53,261][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.0942085012793541, 'train/grad_norm_pre_clip': 18.2834415435791, 'timing/data': 0.006532662024255842, 'timing/model': 1.1363833179930225, 'learning_rate/qwen_vl_interface': 4.395862488081038e-06, 'learning_rate/action_model': 4.395862488081038e-05, 'epoch': 1.16}) +[2026-06-16 00:37:57,781][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.00637420266866684, 'train/grad_norm_pre_clip': 3.337486982345581, 'timing/data': 0.006470205960795283, 'timing/model': 1.1398143940023147, 'learning_rate/qwen_vl_interface': 4.392866883308059e-06, 'learning_rate/action_model': 4.392866883308059e-05, 'epoch': 1.16}) +[2026-06-16 00:38:02,336][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.0793713629245758, 'train/grad_norm_pre_clip': 8.273322105407715, 'timing/data': 0.006310922966804355, 'timing/model': 1.302664156013634, 'learning_rate/qwen_vl_interface': 4.389871630870127e-06, 'learning_rate/action_model': 4.389871630870127e-05, 'epoch': 1.16}) +[2026-06-16 00:38:06,724][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.01328966673463583, 'train/grad_norm_pre_clip': 9.387228012084961, 'timing/data': 0.006366843997966498, 'timing/model': 1.1337705079931766, 'learning_rate/qwen_vl_interface': 4.386876731998478e-06, 'learning_rate/action_model': 4.386876731998478e-05, 'epoch': 1.16}) +[2026-06-16 00:38:11,254][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.11720574647188187, 'train/grad_norm_pre_clip': 20.111751556396484, 'timing/data': 0.006044787005521357, 'timing/model': 1.1354501419700682, 'learning_rate/qwen_vl_interface': 4.3838821879242006e-06, 'learning_rate/action_model': 4.3838821879242e-05, 'epoch': 1.17}) +[2026-06-16 00:38:15,796][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.006118494085967541, 'train/grad_norm_pre_clip': 2.47446608543396, 'timing/data': 0.006337599013932049, 'timing/model': 1.1334358869935386, 'learning_rate/qwen_vl_interface': 4.38088799987824e-06, 'learning_rate/action_model': 4.38088799987824e-05, 'epoch': 1.17}) +[2026-06-16 00:38:20,325][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.11931443959474564, 'train/grad_norm_pre_clip': 43.47907638549805, 'timing/data': 0.006717369018588215, 'timing/model': 1.143441060965415, 'learning_rate/qwen_vl_interface': 4.3778941690913935e-06, 'learning_rate/action_model': 4.3778941690913934e-05, 'epoch': 1.17}) +[2026-06-16 00:38:24,889][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.2054322361946106, 'train/grad_norm_pre_clip': 8.863495826721191, 'timing/data': 0.006529720965772867, 'timing/model': 1.1348257310455665, 'learning_rate/qwen_vl_interface': 4.374900696794311e-06, 'learning_rate/action_model': 4.374900696794311e-05, 'epoch': 1.17}) +[2026-06-16 00:38:29,423][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.3197883069515228, 'train/grad_norm_pre_clip': 5.21262264251709, 'timing/data': 0.006437878008000553, 'timing/model': 1.2810841550235637, 'learning_rate/qwen_vl_interface': 4.371907584217499e-06, 'learning_rate/action_model': 4.371907584217498e-05, 'epoch': 1.17}) +[2026-06-16 00:38:33,817][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.30517125129699707, 'train/grad_norm_pre_clip': 31.3692626953125, 'timing/data': 0.006118824006989598, 'timing/model': 1.1434868890210055, 'learning_rate/qwen_vl_interface': 4.368914832591309e-06, 'learning_rate/action_model': 4.368914832591309e-05, 'epoch': 1.17}) +[2026-06-16 00:38:38,370][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.5235108733177185, 'train/grad_norm_pre_clip': 37.19834899902344, 'timing/data': 0.0057579969870857894, 'timing/model': 1.1344964120071381, 'learning_rate/qwen_vl_interface': 4.365922443145949e-06, 'learning_rate/action_model': 4.365922443145949e-05, 'epoch': 1.17}) +[2026-06-16 00:38:42,857][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.363538533449173, 'train/grad_norm_pre_clip': 24.90215492248535, 'timing/data': 0.00642817304469645, 'timing/model': 1.1345678030047566, 'learning_rate/qwen_vl_interface': 4.362930417111479e-06, 'learning_rate/action_model': 4.362930417111479e-05, 'epoch': 1.17}) +[2026-06-16 00:38:47,351][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.2321608066558838, 'train/grad_norm_pre_clip': 20.02847671508789, 'timing/data': 0.006449434964451939, 'timing/model': 1.129981096019037, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 1.17}) +[2026-06-16 00:38:51,907][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.3840775489807129, 'train/grad_norm_pre_clip': 15.385825157165527, 'timing/data': 0.006664145039394498, 'timing/model': 1.137649928976316, 'learning_rate/qwen_vl_interface': 4.3569474601946934e-06, 'learning_rate/action_model': 4.356947460194694e-05, 'epoch': 1.17}) +[2026-06-16 00:38:56,449][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.14665015041828156, 'train/grad_norm_pre_clip': 13.593801498413086, 'timing/data': 0.006728940992616117, 'timing/model': 1.2940605290350504, 'learning_rate/qwen_vl_interface': 4.3539565317717446e-06, 'learning_rate/action_model': 4.353956531771745e-05, 'epoch': 1.17}) +[2026-06-16 00:39:00,838][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.05107845366001129, 'train/grad_norm_pre_clip': 17.05213737487793, 'timing/data': 0.006407357985153794, 'timing/model': 1.129201004980132, 'learning_rate/qwen_vl_interface': 4.350965971678419e-06, 'learning_rate/action_model': 4.3509659716784185e-05, 'epoch': 1.17}) +[2026-06-16 00:39:05,386][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.009241103194653988, 'train/grad_norm_pre_clip': 2.910269260406494, 'timing/data': 0.006325460039079189, 'timing/model': 1.1380720150191337, 'learning_rate/qwen_vl_interface': 4.3479757811440225e-06, 'learning_rate/action_model': 4.347975781144022e-05, 'epoch': 1.17}) +[2026-06-16 00:39:09,887][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.008685179986059666, 'train/grad_norm_pre_clip': 1.2329216003417969, 'timing/data': 0.006341874017380178, 'timing/model': 1.1370161699596792, 'learning_rate/qwen_vl_interface': 4.34498596139771e-06, 'learning_rate/action_model': 4.34498596139771e-05, 'epoch': 1.17}) +[2026-06-16 00:39:14,392][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.10199513286352158, 'train/grad_norm_pre_clip': 9.34068489074707, 'timing/data': 0.006247292971238494, 'timing/model': 1.136915980023332, 'learning_rate/qwen_vl_interface': 4.341996513668483e-06, 'learning_rate/action_model': 4.3419965136684824e-05, 'epoch': 1.17}) +[2026-06-16 00:39:18,932][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.08204536139965057, 'train/grad_norm_pre_clip': 12.46607494354248, 'timing/data': 0.006493185006547719, 'timing/model': 1.1306400430039503, 'learning_rate/qwen_vl_interface': 4.33900743918519e-06, 'learning_rate/action_model': 4.33900743918519e-05, 'epoch': 1.17}) +[2026-06-16 00:39:23,497][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.005988292396068573, 'train/grad_norm_pre_clip': 1.2194914817810059, 'timing/data': 0.006529013975523412, 'timing/model': 1.3134083860204555, 'learning_rate/qwen_vl_interface': 4.336018739176529e-06, 'learning_rate/action_model': 4.33601873917653e-05, 'epoch': 1.17}) +[2026-06-16 00:39:27,885][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.180576890707016, 'train/grad_norm_pre_clip': 16.426063537597656, 'timing/data': 0.006437079980969429, 'timing/model': 1.1461056109983474, 'learning_rate/qwen_vl_interface': 4.333030414871041e-06, 'learning_rate/action_model': 4.33303041487104e-05, 'epoch': 1.17}) +[2026-06-16 00:39:32,441][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.31038495898246765, 'train/grad_norm_pre_clip': 13.421625137329102, 'timing/data': 0.0063187420018948615, 'timing/model': 1.1377496090135537, 'learning_rate/qwen_vl_interface': 4.330042467497111e-06, 'learning_rate/action_model': 4.330042467497111e-05, 'epoch': 1.17}) +[2026-06-16 00:39:36,962][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.13991929590702057, 'train/grad_norm_pre_clip': 16.935136795043945, 'timing/data': 0.006420362973585725, 'timing/model': 1.142037692014128, 'learning_rate/qwen_vl_interface': 4.327054898282973e-06, 'learning_rate/action_model': 4.3270548982829727e-05, 'epoch': 1.17}) +[2026-06-16 00:39:41,514][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.006974251940846443, 'train/grad_norm_pre_clip': 0.7310086488723755, 'timing/data': 0.006473077984992415, 'timing/model': 1.1360325329587795, 'learning_rate/qwen_vl_interface': 4.324067708456703e-06, 'learning_rate/action_model': 4.3240677084567036e-05, 'epoch': 1.17}) +[2026-06-16 00:39:46,086][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.09673517197370529, 'train/grad_norm_pre_clip': 9.370898246765137, 'timing/data': 0.006884332979097962, 'timing/model': 1.1380990620236844, 'learning_rate/qwen_vl_interface': 4.321080899246225e-06, 'learning_rate/action_model': 4.321080899246225e-05, 'epoch': 1.17}) +[2026-06-16 00:39:50,666][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.044626638293266296, 'train/grad_norm_pre_clip': 7.181785583496094, 'timing/data': 0.006256981985643506, 'timing/model': 1.3018969260156155, 'learning_rate/qwen_vl_interface': 4.318094471879299e-06, 'learning_rate/action_model': 4.3180944718792985e-05, 'epoch': 1.17}) +[2026-06-16 00:39:55,164][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.04796266928315163, 'train/grad_norm_pre_clip': 19.907175064086914, 'timing/data': 0.006334020989015698, 'timing/model': 1.2370850519509986, 'learning_rate/qwen_vl_interface': 4.315108427583535e-06, 'learning_rate/action_model': 4.315108427583535e-05, 'epoch': 1.17}) +[2026-06-16 00:39:59,547][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.019453100860118866, 'train/grad_norm_pre_clip': 15.620125770568848, 'timing/data': 0.006772619963157922, 'timing/model': 1.1350055229850113, 'learning_rate/qwen_vl_interface': 4.312122767586381e-06, 'learning_rate/action_model': 4.3121227675863815e-05, 'epoch': 1.17}) +[2026-06-16 00:40:04,067][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.08453256636857986, 'train/grad_norm_pre_clip': 11.058995246887207, 'timing/data': 0.006555308005772531, 'timing/model': 1.1318626109859906, 'learning_rate/qwen_vl_interface': 4.3091374931151324e-06, 'learning_rate/action_model': 4.309137493115132e-05, 'epoch': 1.18}) +[2026-06-16 00:40:08,622][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.17836612462997437, 'train/grad_norm_pre_clip': 18.858274459838867, 'timing/data': 0.006654314987827092, 'timing/model': 1.1349846899975091, 'learning_rate/qwen_vl_interface': 4.306152605396921e-06, 'learning_rate/action_model': 4.306152605396921e-05, 'epoch': 1.18}) +[2026-06-16 00:40:13,161][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.08607348054647446, 'train/grad_norm_pre_clip': 15.056477546691895, 'timing/data': 0.006314589001704007, 'timing/model': 1.13782727799844, 'learning_rate/qwen_vl_interface': 4.30316810565872e-06, 'learning_rate/action_model': 4.3031681056587206e-05, 'epoch': 1.18}) +[2026-06-16 00:40:17,727][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.07113786041736603, 'train/grad_norm_pre_clip': 23.266876220703125, 'timing/data': 0.006335139973089099, 'timing/model': 1.3003389940131456, 'learning_rate/qwen_vl_interface': 4.300183995127347e-06, 'learning_rate/action_model': 4.300183995127346e-05, 'epoch': 1.18}) +[2026-06-16 00:40:22,252][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.08025158941745758, 'train/grad_norm_pre_clip': 8.642373085021973, 'timing/data': 0.006226568948477507, 'timing/model': 1.2573335750494152, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 1.18}) +[2026-06-16 00:40:26,641][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.07261456549167633, 'train/grad_norm_pre_clip': 20.580480575561523, 'timing/data': 0.006711400987114757, 'timing/model': 1.1382522200001404, 'learning_rate/qwen_vl_interface': 4.29421694659154e-06, 'learning_rate/action_model': 4.294216946591539e-05, 'epoch': 1.18}) +[2026-06-16 00:40:31,180][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.010200168937444687, 'train/grad_norm_pre_clip': 2.157693862915039, 'timing/data': 0.006562628026586026, 'timing/model': 1.1374787479871884, 'learning_rate/qwen_vl_interface': 4.2912340110399345e-06, 'learning_rate/action_model': 4.291234011039934e-05, 'epoch': 1.18}) +[2026-06-16 00:40:35,755][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.029002834111452103, 'train/grad_norm_pre_clip': 7.89401912689209, 'timing/data': 0.006217694957740605, 'timing/model': 1.1451329269912094, 'learning_rate/qwen_vl_interface': 4.2882514696008125e-06, 'learning_rate/action_model': 4.2882514696008125e-05, 'epoch': 1.18}) +[2026-06-16 00:40:40,322][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.10327316075563431, 'train/grad_norm_pre_clip': 8.657607078552246, 'timing/data': 0.006303740956354886, 'timing/model': 1.1402557410183363, 'learning_rate/qwen_vl_interface': 4.2852693235001815e-06, 'learning_rate/action_model': 4.2852693235001815e-05, 'epoch': 1.18}) +[2026-06-16 00:40:44,887][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.16029398143291473, 'train/grad_norm_pre_clip': 14.687103271484375, 'timing/data': 0.006497546972241253, 'timing/model': 1.303314620978199, 'learning_rate/qwen_vl_interface': 4.2822875739638915e-06, 'learning_rate/action_model': 4.282287573963891e-05, 'epoch': 1.18}) +[2026-06-16 00:40:49,293][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.047198303043842316, 'train/grad_norm_pre_clip': 9.944212913513184, 'timing/data': 0.00639723998028785, 'timing/model': 1.140880164981354, 'learning_rate/qwen_vl_interface': 4.279306222217625e-06, 'learning_rate/action_model': 4.279306222217626e-05, 'epoch': 1.18}) +[2026-06-16 00:40:53,844][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.016353202983736992, 'train/grad_norm_pre_clip': 7.1703267097473145, 'timing/data': 0.006445145001634955, 'timing/model': 1.1442531059728935, 'learning_rate/qwen_vl_interface': 4.276325269486906e-06, 'learning_rate/action_model': 4.276325269486906e-05, 'epoch': 1.18}) +[2026-06-16 00:40:58,370][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.21776504814624786, 'train/grad_norm_pre_clip': 13.025727272033691, 'timing/data': 0.006657977995928377, 'timing/model': 1.128068638034165, 'learning_rate/qwen_vl_interface': 4.273344716997089e-06, 'learning_rate/action_model': 4.273344716997089e-05, 'epoch': 1.18}) +[2026-06-16 00:41:02,912][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.04521031677722931, 'train/grad_norm_pre_clip': 6.878324508666992, 'timing/data': 0.006382949010003358, 'timing/model': 1.135165275016334, 'learning_rate/qwen_vl_interface': 4.270364565973368e-06, 'learning_rate/action_model': 4.270364565973367e-05, 'epoch': 1.18}) +[2026-06-16 00:41:07,380][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.054361794143915176, 'train/grad_norm_pre_clip': 8.047937393188477, 'timing/data': 0.0064289000001735985, 'timing/model': 1.1333453379920684, 'learning_rate/qwen_vl_interface': 4.267384817640768e-06, 'learning_rate/action_model': 4.2673848176407685e-05, 'epoch': 1.18}) +[2026-06-16 00:41:11,872][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.014757663942873478, 'train/grad_norm_pre_clip': 3.7250211238861084, 'timing/data': 0.006313018966466188, 'timing/model': 1.2499303090153262, 'learning_rate/qwen_vl_interface': 4.264405473224155e-06, 'learning_rate/action_model': 4.2644054732241546e-05, 'epoch': 1.18}) +[2026-06-16 00:41:16,246][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.05483391508460045, 'train/grad_norm_pre_clip': 13.275004386901855, 'timing/data': 0.006378104037139565, 'timing/model': 1.1309984200051986, 'learning_rate/qwen_vl_interface': 4.261426533948221e-06, 'learning_rate/action_model': 4.2614265339482213e-05, 'epoch': 1.18}) +[2026-06-16 00:41:20,799][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.0705142468214035, 'train/grad_norm_pre_clip': 12.524596214294434, 'timing/data': 0.00631411699578166, 'timing/model': 1.1430560580338351, 'learning_rate/qwen_vl_interface': 4.258448001037499e-06, 'learning_rate/action_model': 4.2584480010374985e-05, 'epoch': 1.18}) +[2026-06-16 00:41:25,325][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.1878012716770172, 'train/grad_norm_pre_clip': 11.159807205200195, 'timing/data': 0.006568782962858677, 'timing/model': 1.1329074690002017, 'learning_rate/qwen_vl_interface': 4.255469875716348e-06, 'learning_rate/action_model': 4.255469875716348e-05, 'epoch': 1.18}) +[2026-06-16 00:41:29,871][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.0005281938938423991, 'train/grad_norm_pre_clip': 0.052416879683732986, 'timing/data': 0.0065297510009258986, 'timing/model': 1.1354066059575416, 'learning_rate/qwen_vl_interface': 4.252492159208966e-06, 'learning_rate/action_model': 4.252492159208966e-05, 'epoch': 1.18}) +[2026-06-16 00:41:34,367][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.014358749613165855, 'train/grad_norm_pre_clip': 1.593091607093811, 'timing/data': 0.006347138958517462, 'timing/model': 1.1311084710177965, 'learning_rate/qwen_vl_interface': 4.2495148527393766e-06, 'learning_rate/action_model': 4.2495148527393766e-05, 'epoch': 1.18}) +[2026-06-16 00:41:38,881][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.0055182576179504395, 'train/grad_norm_pre_clip': 0.883962869644165, 'timing/data': 0.006554110033903271, 'timing/model': 1.267558210005518, 'learning_rate/qwen_vl_interface': 4.246537957531441e-06, 'learning_rate/action_model': 4.2465379575314404e-05, 'epoch': 1.18}) +[2026-06-16 00:41:43,308][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.012471416033804417, 'train/grad_norm_pre_clip': 2.8357343673706055, 'timing/data': 0.0159404770238325, 'timing/model': 1.1353531279601157, 'learning_rate/qwen_vl_interface': 4.243561474808846e-06, 'learning_rate/action_model': 4.2435614748088456e-05, 'epoch': 1.18}) +[2026-06-16 00:41:47,866][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.09988073259592056, 'train/grad_norm_pre_clip': 9.383194923400879, 'timing/data': 0.006626566988416016, 'timing/model': 1.1351310279569589, 'learning_rate/qwen_vl_interface': 4.240585405795112e-06, 'learning_rate/action_model': 4.240585405795112e-05, 'epoch': 1.18}) +[2026-06-16 00:41:52,412][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.11191820353269577, 'train/grad_norm_pre_clip': 6.070096492767334, 'timing/data': 0.006555002008099109, 'timing/model': 1.1337012630538084, 'learning_rate/qwen_vl_interface': 4.237609751713589e-06, 'learning_rate/action_model': 4.2376097517135896e-05, 'epoch': 1.19}) +[2026-06-16 00:41:56,958][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.013538358733057976, 'train/grad_norm_pre_clip': 2.088346242904663, 'timing/data': 0.006847211974672973, 'timing/model': 1.1368903939728625, 'learning_rate/qwen_vl_interface': 4.234634513787455e-06, 'learning_rate/action_model': 4.2346345137874556e-05, 'epoch': 1.19}) +[2026-06-16 00:42:01,518][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.004774454515427351, 'train/grad_norm_pre_clip': 0.5051713585853577, 'timing/data': 0.006838275003246963, 'timing/model': 1.2979423080105335, 'learning_rate/qwen_vl_interface': 4.2316596932397195e-06, 'learning_rate/action_model': 4.2316596932397195e-05, 'epoch': 1.19}) +[2026-06-16 00:42:05,923][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.05612115561962128, 'train/grad_norm_pre_clip': 7.275620460510254, 'timing/data': 0.006361744017340243, 'timing/model': 1.1443968510138802, 'learning_rate/qwen_vl_interface': 4.22868529129322e-06, 'learning_rate/action_model': 4.22868529129322e-05, 'epoch': 1.19}) +[2026-06-16 00:42:10,498][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.000780638016294688, 'train/grad_norm_pre_clip': 0.08651746064424515, 'timing/data': 0.006484418991021812, 'timing/model': 1.1400692940223962, 'learning_rate/qwen_vl_interface': 4.2257113091706165e-06, 'learning_rate/action_model': 4.2257113091706165e-05, 'epoch': 1.19}) +[2026-06-16 00:42:14,985][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.33171725273132324, 'train/grad_norm_pre_clip': 15.360638618469238, 'timing/data': 0.005897447990719229, 'timing/model': 1.143377872009296, 'learning_rate/qwen_vl_interface': 4.222737748094404e-06, 'learning_rate/action_model': 4.222737748094404e-05, 'epoch': 1.19}) +[2026-06-16 00:42:19,470][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.042497146874666214, 'train/grad_norm_pre_clip': 8.934812545776367, 'timing/data': 0.00644184899283573, 'timing/model': 1.1308528360095806, 'learning_rate/qwen_vl_interface': 4.219764609286901e-06, 'learning_rate/action_model': 4.219764609286901e-05, 'epoch': 1.19}) +[2026-06-16 00:42:23,962][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.014372305944561958, 'train/grad_norm_pre_clip': 4.550325393676758, 'timing/data': 0.006741428980603814, 'timing/model': 1.1333107309765182, 'learning_rate/qwen_vl_interface': 4.21679189397025e-06, 'learning_rate/action_model': 4.21679189397025e-05, 'epoch': 1.19}) +[2026-06-16 00:42:28,496][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.01091737486422062, 'train/grad_norm_pre_clip': 1.4078810214996338, 'timing/data': 0.006545837968587875, 'timing/model': 1.2885314309969544, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 1.19}) +[2026-06-16 00:42:32,882][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.006031960714608431, 'train/grad_norm_pre_clip': 1.040902018547058, 'timing/data': 0.006433248985558748, 'timing/model': 1.132197484956123, 'learning_rate/qwen_vl_interface': 4.210847738697219e-06, 'learning_rate/action_model': 4.210847738697218e-05, 'epoch': 1.19}) +[2026-06-16 00:42:37,425][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.06147139519453049, 'train/grad_norm_pre_clip': 28.043893814086914, 'timing/data': 0.006125928019173443, 'timing/model': 1.1387842610129155, 'learning_rate/qwen_vl_interface': 4.207876301184256e-06, 'learning_rate/action_model': 4.207876301184256e-05, 'epoch': 1.19}) +[2026-06-16 00:42:41,951][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.09587999433279037, 'train/grad_norm_pre_clip': 9.613420486450195, 'timing/data': 0.006580971006769687, 'timing/model': 1.1297013069852255, 'learning_rate/qwen_vl_interface': 4.2049052920489825e-06, 'learning_rate/action_model': 4.2049052920489825e-05, 'epoch': 1.19}) +[2026-06-16 00:42:46,447][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.007184357848018408, 'train/grad_norm_pre_clip': 1.2160884141921997, 'timing/data': 0.006284539005719125, 'timing/model': 1.1361871439730749, 'learning_rate/qwen_vl_interface': 4.201934712512666e-06, 'learning_rate/action_model': 4.2019347125126656e-05, 'epoch': 1.19}) +[2026-06-16 00:42:50,970][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.008139846846461296, 'train/grad_norm_pre_clip': 1.3258328437805176, 'timing/data': 0.0062969260034151375, 'timing/model': 1.14242878096411, 'learning_rate/qwen_vl_interface': 4.1989645637964e-06, 'learning_rate/action_model': 4.1989645637964e-05, 'epoch': 1.19}) +[2026-06-16 00:42:55,535][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.0003791952331084758, 'train/grad_norm_pre_clip': 0.06016397103667259, 'timing/data': 0.006308011012151837, 'timing/model': 1.299501845962368, 'learning_rate/qwen_vl_interface': 4.195994847121101e-06, 'learning_rate/action_model': 4.195994847121101e-05, 'epoch': 1.19}) +[2026-06-16 00:47:18,566][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.000876494450494647, 'train/grad_norm_pre_clip': 0.16555756330490112, 'eval/loss': 0.04799095168709755, 'eval/flap_precision': 0.9678899082568807, 'eval/flap_recall': 0.9768518518518519, 'eval/flap_f1': 0.9723502304147466, 'eval/flap_rate_model': 0.18166666666666667, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9723502304147464, 'timing/data': 0.006497811991721392, 'timing/model': 1.1338976150145754, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 1.19}) +[2026-06-16 00:47:18,566][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2900_state +[2026-06-16 00:47:18,569][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 00:47:22,159][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2900_state/model.safetensors +[2026-06-16 00:47:32,202][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2900_state/optimizer.bin +[2026-06-16 00:47:32,202][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2900_state/sampler.bin +[2026-06-16 00:47:32,202][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2900_state/sampler_1.bin +[2026-06-16 00:47:32,203][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_2900_state/random_states_0.pkl +[2026-06-16 00:47:32,249][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2900 +[2026-06-16 00:47:32,250][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 00:47:32,256][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 00:47:36,694][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.007988817058503628, 'train/grad_norm_pre_clip': 1.822011947631836, 'timing/data': 0.006618775019887835, 'timing/model': 1.2650773199857213, 'learning_rate/qwen_vl_interface': 4.19005671477618e-06, 'learning_rate/action_model': 4.1900567147761794e-05, 'epoch': 1.19}) +[2026-06-16 00:47:41,014][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.006617164704948664, 'train/grad_norm_pre_clip': 2.1920645236968994, 'timing/data': 0.006534689979162067, 'timing/model': 1.1169141209684312, 'learning_rate/qwen_vl_interface': 4.1870883015474975e-06, 'learning_rate/action_model': 4.187088301547497e-05, 'epoch': 1.19}) +[2026-06-16 00:47:45,485][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.06845561414957047, 'train/grad_norm_pre_clip': 10.323542594909668, 'timing/data': 0.006522126961499453, 'timing/model': 1.1174143739626743, 'learning_rate/qwen_vl_interface': 4.1841203252416655e-06, 'learning_rate/action_model': 4.1841203252416655e-05, 'epoch': 1.19}) +[2026-06-16 00:47:49,965][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.11850224435329437, 'train/grad_norm_pre_clip': 9.661745071411133, 'timing/data': 0.006214879977051169, 'timing/model': 1.1257520610233769, 'learning_rate/qwen_vl_interface': 4.181152787078705e-06, 'learning_rate/action_model': 4.1811527870787056e-05, 'epoch': 1.19}) +[2026-06-16 00:47:54,408][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.11084672808647156, 'train/grad_norm_pre_clip': 9.862356185913086, 'timing/data': 0.006283296970650554, 'timing/model': 1.1203850570018403, 'learning_rate/qwen_vl_interface': 4.178185688278462e-06, 'learning_rate/action_model': 4.1781856882784616e-05, 'epoch': 1.19}) +[2026-06-16 00:47:58,842][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.26605573296546936, 'train/grad_norm_pre_clip': 15.51380729675293, 'timing/data': 0.006269213976338506, 'timing/model': 1.1167137710144743, 'learning_rate/qwen_vl_interface': 4.175219030060595e-06, 'learning_rate/action_model': 4.175219030060595e-05, 'epoch': 1.19}) +[2026-06-16 00:48:03,343][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.004107892047613859, 'train/grad_norm_pre_clip': 0.6834565997123718, 'timing/data': 0.0062232709606178105, 'timing/model': 1.2916333100292832, 'learning_rate/qwen_vl_interface': 4.1722528136445875e-06, 'learning_rate/action_model': 4.172252813644588e-05, 'epoch': 1.19}) +[2026-06-16 00:48:07,690][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.0022521240171045065, 'train/grad_norm_pre_clip': 0.4145347774028778, 'timing/data': 0.006344946974422783, 'timing/model': 1.128180391038768, 'learning_rate/qwen_vl_interface': 4.169287040249738e-06, 'learning_rate/action_model': 4.169287040249738e-05, 'epoch': 1.19}) +[2026-06-16 00:48:12,187][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.0005142170703038573, 'train/grad_norm_pre_clip': 0.1771939992904663, 'timing/data': 0.006061413965653628, 'timing/model': 1.1251434439909644, 'learning_rate/qwen_vl_interface': 4.166321711095164e-06, 'learning_rate/action_model': 4.166321711095164e-05, 'epoch': 1.2}) +[2026-06-16 00:48:16,680][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.01196585688740015, 'train/grad_norm_pre_clip': 5.112028121948242, 'timing/data': 0.00643100799061358, 'timing/model': 1.124364343995694, 'learning_rate/qwen_vl_interface': 4.1633568273998e-06, 'learning_rate/action_model': 4.1633568273998e-05, 'epoch': 1.2}) +[2026-06-16 00:48:21,175][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.00018575646390672773, 'train/grad_norm_pre_clip': 0.020893730223178864, 'timing/data': 0.006378029007464647, 'timing/model': 1.1236853229929693, 'learning_rate/qwen_vl_interface': 4.1603923903823995e-06, 'learning_rate/action_model': 4.160392390382399e-05, 'epoch': 1.2}) +[2026-06-16 00:48:25,686][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.2038286328315735, 'train/grad_norm_pre_clip': 21.793001174926758, 'timing/data': 0.006636600010097027, 'timing/model': 1.1310768480179831, 'learning_rate/qwen_vl_interface': 4.1574284012615276e-06, 'learning_rate/action_model': 4.157428401261528e-05, 'epoch': 1.2}) +[2026-06-16 00:48:30,185][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.09151951968669891, 'train/grad_norm_pre_clip': 12.13623332977295, 'timing/data': 0.006049460964277387, 'timing/model': 1.279077596962452, 'learning_rate/qwen_vl_interface': 4.154464861255571e-06, 'learning_rate/action_model': 4.154464861255571e-05, 'epoch': 1.2}) +[2026-06-16 00:48:34,552][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.08132532984018326, 'train/grad_norm_pre_clip': 11.808609962463379, 'timing/data': 0.006131014029961079, 'timing/model': 1.130198366008699, 'learning_rate/qwen_vl_interface': 4.151501771582728e-06, 'learning_rate/action_model': 4.151501771582728e-05, 'epoch': 1.2}) +[2026-06-16 00:48:39,082][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.003390833269804716, 'train/grad_norm_pre_clip': 0.786835253238678, 'timing/data': 0.006504399992991239, 'timing/model': 1.1355057220207527, 'learning_rate/qwen_vl_interface': 4.148539133461012e-06, 'learning_rate/action_model': 4.1485391334610116e-05, 'epoch': 1.2}) +[2026-06-16 00:48:43,487][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.14164823293685913, 'train/grad_norm_pre_clip': 16.926502227783203, 'timing/data': 0.006251576996874064, 'timing/model': 1.1247324779978953, 'learning_rate/qwen_vl_interface': 4.145576948108254e-06, 'learning_rate/action_model': 4.145576948108253e-05, 'epoch': 1.2}) +[2026-06-16 00:48:47,999][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.11006210744380951, 'train/grad_norm_pre_clip': 11.220645904541016, 'timing/data': 0.006296422972809523, 'timing/model': 1.1277112669777125, 'learning_rate/qwen_vl_interface': 4.142615216742093e-06, 'learning_rate/action_model': 4.142615216742093e-05, 'epoch': 1.2}) +[2026-06-16 00:48:52,535][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.14724846184253693, 'train/grad_norm_pre_clip': 9.584456443786621, 'timing/data': 0.006828974990639836, 'timing/model': 1.128857811971102, 'learning_rate/qwen_vl_interface': 4.139653940579988e-06, 'learning_rate/action_model': 4.1396539405799874e-05, 'epoch': 1.2}) +[2026-06-16 00:48:57,073][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.13292744755744934, 'train/grad_norm_pre_clip': 16.70883560180664, 'timing/data': 0.0063362669898197055, 'timing/model': 1.2988484199740924, 'learning_rate/qwen_vl_interface': 4.136693120839206e-06, 'learning_rate/action_model': 4.136693120839205e-05, 'epoch': 1.2}) +[2026-06-16 00:49:01,454][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.003430828219279647, 'train/grad_norm_pre_clip': 0.6545153260231018, 'timing/data': 0.00634094700217247, 'timing/model': 1.132744891976472, 'learning_rate/qwen_vl_interface': 4.133732758736828e-06, 'learning_rate/action_model': 4.133732758736828e-05, 'epoch': 1.2}) +[2026-06-16 00:49:05,992][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.00976178701967001, 'train/grad_norm_pre_clip': 2.290604829788208, 'timing/data': 0.006428555003367364, 'timing/model': 1.1336403689929284, 'learning_rate/qwen_vl_interface': 4.13077285548975e-06, 'learning_rate/action_model': 4.13077285548975e-05, 'epoch': 1.2}) +[2026-06-16 00:49:10,479][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.24046596884727478, 'train/grad_norm_pre_clip': 6.18678092956543, 'timing/data': 0.006455751019529998, 'timing/model': 1.1336246529826894, 'learning_rate/qwen_vl_interface': 4.127813412314672e-06, 'learning_rate/action_model': 4.1278134123146725e-05, 'epoch': 1.2}) +[2026-06-16 00:49:15,017][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.0009992832783609629, 'train/grad_norm_pre_clip': 0.18877530097961426, 'timing/data': 0.006642536958679557, 'timing/model': 1.1317240300122648, 'learning_rate/qwen_vl_interface': 4.124854430428113e-06, 'learning_rate/action_model': 4.1248544304281126e-05, 'epoch': 1.2}) +[2026-06-16 00:49:19,560][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.12310103327035904, 'train/grad_norm_pre_clip': 14.57966423034668, 'timing/data': 0.00623472104780376, 'timing/model': 1.1388284760178067, 'learning_rate/qwen_vl_interface': 4.121895911046397e-06, 'learning_rate/action_model': 4.121895911046396e-05, 'epoch': 1.2}) +[2026-06-16 00:49:24,074][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.11411638557910919, 'train/grad_norm_pre_clip': 20.151399612426758, 'timing/data': 0.006656448007561266, 'timing/model': 1.2804456719895825, 'learning_rate/qwen_vl_interface': 4.118937855385659e-06, 'learning_rate/action_model': 4.1189378553856584e-05, 'epoch': 1.2}) +[2026-06-16 00:49:28,554][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.01811252534389496, 'train/grad_norm_pre_clip': 11.333480834960938, 'timing/data': 0.006246409029699862, 'timing/model': 1.246215991966892, 'learning_rate/qwen_vl_interface': 4.115980264661844e-06, 'learning_rate/action_model': 4.115980264661844e-05, 'epoch': 1.2}) +[2026-06-16 00:49:32,916][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.04400354251265526, 'train/grad_norm_pre_clip': 10.193007469177246, 'timing/data': 0.006417091004550457, 'timing/model': 1.13270721497247, 'learning_rate/qwen_vl_interface': 4.113023140090709e-06, 'learning_rate/action_model': 4.113023140090709e-05, 'epoch': 1.2}) +[2026-06-16 00:49:37,430][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.08870632201433182, 'train/grad_norm_pre_clip': 12.941068649291992, 'timing/data': 0.006320704007521272, 'timing/model': 1.1295576260308735, 'learning_rate/qwen_vl_interface': 4.110066482887812e-06, 'learning_rate/action_model': 4.110066482887812e-05, 'epoch': 1.2}) +[2026-06-16 00:49:41,984][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.3505837619304657, 'train/grad_norm_pre_clip': 21.122655868530273, 'timing/data': 0.006088856025598943, 'timing/model': 1.1324676030199043, 'learning_rate/qwen_vl_interface': 4.107110294268525e-06, 'learning_rate/action_model': 4.107110294268525e-05, 'epoch': 1.2}) +[2026-06-16 00:49:46,531][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.0002923710271716118, 'train/grad_norm_pre_clip': 0.04357890784740448, 'timing/data': 0.006616023951210082, 'timing/model': 1.1401435009902343, 'learning_rate/qwen_vl_interface': 4.104154575448026e-06, 'learning_rate/action_model': 4.104154575448025e-05, 'epoch': 1.2}) +[2026-06-16 00:49:51,084][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.23561108112335205, 'train/grad_norm_pre_clip': 19.35492706298828, 'timing/data': 0.006602393987122923, 'timing/model': 1.3033587969839573, 'learning_rate/qwen_vl_interface': 4.101199327641297e-06, 'learning_rate/action_model': 4.101199327641297e-05, 'epoch': 1.2}) +[2026-06-16 00:49:55,457][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.006286981515586376, 'train/grad_norm_pre_clip': 2.8439838886260986, 'timing/data': 0.006513757980428636, 'timing/model': 1.1362342670327052, 'learning_rate/qwen_vl_interface': 4.098244552063133e-06, 'learning_rate/action_model': 4.098244552063132e-05, 'epoch': 1.2}) +[2026-06-16 00:50:00,009][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.08654104918241501, 'train/grad_norm_pre_clip': 20.34686279296875, 'timing/data': 0.006341176049318165, 'timing/model': 1.1340673989616334, 'learning_rate/qwen_vl_interface': 4.0952902499281254e-06, 'learning_rate/action_model': 4.095290249928125e-05, 'epoch': 1.2}) +[2026-06-16 00:50:04,558][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.07683392614126205, 'train/grad_norm_pre_clip': 14.04776382446289, 'timing/data': 0.0059885880327783525, 'timing/model': 1.1311931569944136, 'learning_rate/qwen_vl_interface': 4.092336422450681e-06, 'learning_rate/action_model': 4.0923364224506804e-05, 'epoch': 1.21}) +[2026-06-16 00:50:09,101][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.08509334921836853, 'train/grad_norm_pre_clip': 5.699882984161377, 'timing/data': 0.006600317021366209, 'timing/model': 1.1405635100090876, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 1.21}) +[2026-06-16 00:50:13,572][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.09883660823106766, 'train/grad_norm_pre_clip': 15.36221981048584, 'timing/data': 0.006380726990755647, 'timing/model': 1.1382090949919075, 'learning_rate/qwen_vl_interface': 4.086430196325106e-06, 'learning_rate/action_model': 4.086430196325106e-05, 'epoch': 1.21}) +[2026-06-16 00:50:18,066][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.08917313814163208, 'train/grad_norm_pre_clip': 10.930041313171387, 'timing/data': 0.006323868990875781, 'timing/model': 1.2501347460201941, 'learning_rate/qwen_vl_interface': 4.083477800104802e-06, 'learning_rate/action_model': 4.083477800104802e-05, 'epoch': 1.21}) +[2026-06-16 00:50:22,457][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.0905652865767479, 'train/grad_norm_pre_clip': 11.200631141662598, 'timing/data': 0.006196851027198136, 'timing/model': 1.1384366459678859, 'learning_rate/qwen_vl_interface': 4.080525883397711e-06, 'learning_rate/action_model': 4.080525883397711e-05, 'epoch': 1.21}) +[2026-06-16 00:50:26,994][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.09559401124715805, 'train/grad_norm_pre_clip': 23.10391616821289, 'timing/data': 0.006246323988307267, 'timing/model': 1.1320964699843898, 'learning_rate/qwen_vl_interface': 4.077574447417256e-06, 'learning_rate/action_model': 4.077574447417256e-05, 'epoch': 1.21}) +[2026-06-16 00:50:31,525][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.11919985711574554, 'train/grad_norm_pre_clip': 15.452324867248535, 'timing/data': 0.006585559051018208, 'timing/model': 1.1306769350194372, 'learning_rate/qwen_vl_interface': 4.074623493376657e-06, 'learning_rate/action_model': 4.074623493376657e-05, 'epoch': 1.21}) +[2026-06-16 00:50:36,065][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.0048035853542387486, 'train/grad_norm_pre_clip': 0.9543775320053101, 'timing/data': 0.006504732999019325, 'timing/model': 1.1305294309859164, 'learning_rate/qwen_vl_interface': 4.071673022488944e-06, 'learning_rate/action_model': 4.071673022488944e-05, 'epoch': 1.21}) +[2026-06-16 00:50:40,567][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.05390714481472969, 'train/grad_norm_pre_clip': 4.4130682945251465, 'timing/data': 0.006439705030061305, 'timing/model': 1.1309418359887786, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 1.21}) +[2026-06-16 00:50:45,078][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.3967534005641937, 'train/grad_norm_pre_clip': 21.627178192138672, 'timing/data': 0.006467678002081811, 'timing/model': 1.2637040929985233, 'learning_rate/qwen_vl_interface': 4.065773535023277e-06, 'learning_rate/action_model': 4.0657735350232766e-05, 'epoch': 1.21}) +[2026-06-16 00:50:49,469][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.009461176581680775, 'train/grad_norm_pre_clip': 1.631387710571289, 'timing/data': 0.00612966000335291, 'timing/model': 1.1330137330223806, 'learning_rate/qwen_vl_interface': 4.062824520870382e-06, 'learning_rate/action_model': 4.062824520870381e-05, 'epoch': 1.21}) +[2026-06-16 00:50:54,025][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.029658207669854164, 'train/grad_norm_pre_clip': 7.396618366241455, 'timing/data': 0.006048985989764333, 'timing/model': 1.1387147959903814, 'learning_rate/qwen_vl_interface': 4.059875994720481e-06, 'learning_rate/action_model': 4.059875994720481e-05, 'epoch': 1.21}) +[2026-06-16 00:50:58,582][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.08538878709077835, 'train/grad_norm_pre_clip': 10.113165855407715, 'timing/data': 0.005891574022825807, 'timing/model': 1.1392688759951852, 'learning_rate/qwen_vl_interface': 4.056927957785605e-06, 'learning_rate/action_model': 4.056927957785605e-05, 'epoch': 1.21}) +[2026-06-16 00:51:03,125][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.09754467010498047, 'train/grad_norm_pre_clip': 10.321595191955566, 'timing/data': 0.006458471994847059, 'timing/model': 1.1405060869874433, 'learning_rate/qwen_vl_interface': 4.0539804112775785e-06, 'learning_rate/action_model': 4.053980411277578e-05, 'epoch': 1.21}) +[2026-06-16 00:51:07,683][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.00979217141866684, 'train/grad_norm_pre_clip': 1.9304327964782715, 'timing/data': 0.00609353999607265, 'timing/model': 1.3115727739641443, 'learning_rate/qwen_vl_interface': 4.051033356408027e-06, 'learning_rate/action_model': 4.0510333564080264e-05, 'epoch': 1.21}) +[2026-06-16 00:51:12,074][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.24403290450572968, 'train/grad_norm_pre_clip': 21.61705207824707, 'timing/data': 0.006392399955075234, 'timing/model': 1.1378238510224037, 'learning_rate/qwen_vl_interface': 4.048086794388375e-06, 'learning_rate/action_model': 4.048086794388375e-05, 'epoch': 1.21}) +[2026-06-16 00:51:16,627][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.09619396924972534, 'train/grad_norm_pre_clip': 8.101421356201172, 'timing/data': 0.00620233902009204, 'timing/model': 1.1309309370117262, 'learning_rate/qwen_vl_interface': 4.045140726429841e-06, 'learning_rate/action_model': 4.0451407264298406e-05, 'epoch': 1.21}) +[2026-06-16 00:51:21,064][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.047652408480644226, 'train/grad_norm_pre_clip': 6.247107028961182, 'timing/data': 0.006696176016703248, 'timing/model': 1.1324726279708557, 'learning_rate/qwen_vl_interface': 4.042195153743443e-06, 'learning_rate/action_model': 4.042195153743443e-05, 'epoch': 1.21}) +[2026-06-16 00:51:25,546][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.007578766439110041, 'train/grad_norm_pre_clip': 1.549230933189392, 'timing/data': 0.005909128987696022, 'timing/model': 1.1414077510125935, 'learning_rate/qwen_vl_interface': 4.039250077539993e-06, 'learning_rate/action_model': 4.039250077539993e-05, 'epoch': 1.21}) +[2026-06-16 00:51:30,058][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.0007701658178120852, 'train/grad_norm_pre_clip': 0.0794462189078331, 'timing/data': 0.006224587035831064, 'timing/model': 1.1338717520120554, 'learning_rate/qwen_vl_interface': 4.036305499030104e-06, 'learning_rate/action_model': 4.0363054990301036e-05, 'epoch': 1.21}) +[2026-06-16 00:51:34,608][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.1328788846731186, 'train/grad_norm_pre_clip': 7.265593528747559, 'timing/data': 0.0064445529715158045, 'timing/model': 1.2955341979977675, 'learning_rate/qwen_vl_interface': 4.033361419424178e-06, 'learning_rate/action_model': 4.033361419424178e-05, 'epoch': 1.21}) +[2026-06-16 00:51:39,000][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.004604138433933258, 'train/grad_norm_pre_clip': 0.49023157358169556, 'timing/data': 0.006729885993991047, 'timing/model': 1.13267044397071, 'learning_rate/qwen_vl_interface': 4.030417839932418e-06, 'learning_rate/action_model': 4.030417839932417e-05, 'epoch': 1.21}) +[2026-06-16 00:51:43,550][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.013653287664055824, 'train/grad_norm_pre_clip': 2.6140856742858887, 'timing/data': 0.006567884003743529, 'timing/model': 1.1453248579637147, 'learning_rate/qwen_vl_interface': 4.027474761764815e-06, 'learning_rate/action_model': 4.027474761764814e-05, 'epoch': 1.21}) +[2026-06-16 00:51:48,090][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.015112710185348988, 'train/grad_norm_pre_clip': 1.574305772781372, 'timing/data': 0.006568366021383554, 'timing/model': 1.1323098160210066, 'learning_rate/qwen_vl_interface': 4.024532186131158e-06, 'learning_rate/action_model': 4.0245321861311575e-05, 'epoch': 1.21}) +[2026-06-16 00:51:52,578][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.009253261610865593, 'train/grad_norm_pre_clip': 1.2058031558990479, 'timing/data': 0.006348501017782837, 'timing/model': 1.1340763159678318, 'learning_rate/qwen_vl_interface': 4.021590114241029e-06, 'learning_rate/action_model': 4.021590114241029e-05, 'epoch': 1.22}) +[2026-06-16 00:51:57,086][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.2028314769268036, 'train/grad_norm_pre_clip': 9.359784126281738, 'timing/data': 0.006229271995835006, 'timing/model': 1.1321540380013175, 'learning_rate/qwen_vl_interface': 4.018648547303804e-06, 'learning_rate/action_model': 4.0186485473038036e-05, 'epoch': 1.22}) +[2026-06-16 00:52:01,630][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.02687937580049038, 'train/grad_norm_pre_clip': 7.2678351402282715, 'timing/data': 0.0071866589714773, 'timing/model': 1.2974686660454609, 'learning_rate/qwen_vl_interface': 4.015707486528649e-06, 'learning_rate/action_model': 4.015707486528649e-05, 'epoch': 1.22}) +[2026-06-16 00:52:06,029][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.006196653936058283, 'train/grad_norm_pre_clip': 0.7051816582679749, 'timing/data': 0.006759429001249373, 'timing/model': 1.1424157580477186, 'learning_rate/qwen_vl_interface': 4.012766933124523e-06, 'learning_rate/action_model': 4.012766933124523e-05, 'epoch': 1.22}) +[2026-06-16 00:52:10,596][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.23458576202392578, 'train/grad_norm_pre_clip': 21.388042449951172, 'timing/data': 0.00674947997322306, 'timing/model': 1.1418880530400202, 'learning_rate/qwen_vl_interface': 4.009826888300176e-06, 'learning_rate/action_model': 4.0098268883001755e-05, 'epoch': 1.22}) +[2026-06-16 00:52:15,155][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.2112329751253128, 'train/grad_norm_pre_clip': 16.129199981689453, 'timing/data': 0.0067824049619957805, 'timing/model': 1.1404740969883278, 'learning_rate/qwen_vl_interface': 4.00688735326415e-06, 'learning_rate/action_model': 4.0068873532641495e-05, 'epoch': 1.22}) +[2026-06-16 00:52:19,695][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.08915956318378448, 'train/grad_norm_pre_clip': 10.009157180786133, 'timing/data': 0.006247544020880014, 'timing/model': 1.142255419981666, 'learning_rate/qwen_vl_interface': 4.0039483292247765e-06, 'learning_rate/action_model': 4.003948329224776e-05, 'epoch': 1.22}) +[2026-06-16 00:52:24,247][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.11211606860160828, 'train/grad_norm_pre_clip': 9.610265731811523, 'timing/data': 0.0061742610414512455, 'timing/model': 1.1417469329899177, 'learning_rate/qwen_vl_interface': 4.001009817390178e-06, 'learning_rate/action_model': 4.001009817390178e-05, 'epoch': 1.22}) +[2026-06-16 00:52:28,797][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.007699284236878157, 'train/grad_norm_pre_clip': 1.324607253074646, 'timing/data': 0.006533468025736511, 'timing/model': 1.296295200008899, 'learning_rate/qwen_vl_interface': 3.998071818968264e-06, 'learning_rate/action_model': 3.998071818968264e-05, 'epoch': 1.22}) +[2026-06-16 00:52:33,188][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.05743405222892761, 'train/grad_norm_pre_clip': 6.5267229080200195, 'timing/data': 0.006406921020243317, 'timing/model': 1.1449996840092354, 'learning_rate/qwen_vl_interface': 3.995134335166736e-06, 'learning_rate/action_model': 3.995134335166736e-05, 'epoch': 1.22}) +[2026-06-16 00:52:37,753][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.056537140160799026, 'train/grad_norm_pre_clip': 10.14936637878418, 'timing/data': 0.006348178954795003, 'timing/model': 1.1393898529931903, 'learning_rate/qwen_vl_interface': 3.992197367193083e-06, 'learning_rate/action_model': 3.992197367193082e-05, 'epoch': 1.22}) +[2026-06-16 00:52:42,250][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.003957644104957581, 'train/grad_norm_pre_clip': 0.6401412487030029, 'timing/data': 0.006388327979948372, 'timing/model': 1.1330519309849478, 'learning_rate/qwen_vl_interface': 3.98926091625458e-06, 'learning_rate/action_model': 3.9892609162545795e-05, 'epoch': 1.22}) +[2026-06-16 00:52:46,757][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.009417872875928879, 'train/grad_norm_pre_clip': 3.1165850162506104, 'timing/data': 0.0064250860013999045, 'timing/model': 1.1353567609912716, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 1.22}) +[2026-06-16 00:52:51,314][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.006338191218674183, 'train/grad_norm_pre_clip': 1.114203691482544, 'timing/data': 0.00620989000890404, 'timing/model': 1.146592908015009, 'learning_rate/qwen_vl_interface': 3.983389570311073e-06, 'learning_rate/action_model': 3.9833895703110724e-05, 'epoch': 1.22}) +[2026-06-16 00:52:55,858][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.17836514115333557, 'train/grad_norm_pre_clip': 22.796064376831055, 'timing/data': 0.006643192959018052, 'timing/model': 1.2920116479508579, 'learning_rate/qwen_vl_interface': 3.9804546777195546e-06, 'learning_rate/action_model': 3.980454677719554e-05, 'epoch': 1.22}) +[2026-06-16 00:53:00,238][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.08623126149177551, 'train/grad_norm_pre_clip': 8.271492958068848, 'timing/data': 0.006502520991489291, 'timing/model': 1.135968315997161, 'learning_rate/qwen_vl_interface': 3.977520306990164e-06, 'learning_rate/action_model': 3.977520306990163e-05, 'epoch': 1.22}) +[2026-06-16 00:53:04,775][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.11214947700500488, 'train/grad_norm_pre_clip': 15.118062973022461, 'timing/data': 0.006303736008703709, 'timing/model': 1.131035232974682, 'learning_rate/qwen_vl_interface': 3.974586459329109e-06, 'learning_rate/action_model': 3.9745864593291085e-05, 'epoch': 1.22}) +[2026-06-16 00:53:09,280][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.01509484276175499, 'train/grad_norm_pre_clip': 3.2159252166748047, 'timing/data': 0.006367901980411261, 'timing/model': 1.1375728830462322, 'learning_rate/qwen_vl_interface': 3.9716531359423855e-06, 'learning_rate/action_model': 3.971653135942385e-05, 'epoch': 1.22}) +[2026-06-16 00:53:13,801][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.013958984054625034, 'train/grad_norm_pre_clip': 7.799604415893555, 'timing/data': 0.006414458970539272, 'timing/model': 1.1385080960462801, 'learning_rate/qwen_vl_interface': 3.968720338035769e-06, 'learning_rate/action_model': 3.968720338035769e-05, 'epoch': 1.22}) +[2026-06-16 00:53:18,355][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.013872785493731499, 'train/grad_norm_pre_clip': 1.3602585792541504, 'timing/data': 0.006564446026459336, 'timing/model': 1.1346621119882911, 'learning_rate/qwen_vl_interface': 3.965788066814824e-06, 'learning_rate/action_model': 3.965788066814823e-05, 'epoch': 1.22}) +[2026-06-16 00:53:22,920][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.10545720160007477, 'train/grad_norm_pre_clip': 14.505422592163086, 'timing/data': 0.006441424950025976, 'timing/model': 1.3124197970028035, 'learning_rate/qwen_vl_interface': 3.962856323484895e-06, 'learning_rate/action_model': 3.9628563234848945e-05, 'epoch': 1.22}) +[2026-06-16 00:53:27,322][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.005019108299165964, 'train/grad_norm_pre_clip': 1.4215261936187744, 'timing/data': 0.0060334159643389285, 'timing/model': 1.1455670280265622, 'learning_rate/qwen_vl_interface': 3.959925109251112e-06, 'learning_rate/action_model': 3.959925109251112e-05, 'epoch': 1.22}) +[2026-06-16 00:53:31,863][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.22472114861011505, 'train/grad_norm_pre_clip': 15.934921264648438, 'timing/data': 0.008644776011351496, 'timing/model': 1.1324996079783887, 'learning_rate/qwen_vl_interface': 3.956994425318387e-06, 'learning_rate/action_model': 3.956994425318387e-05, 'epoch': 1.22}) +[2026-06-16 00:53:36,367][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.020535100251436234, 'train/grad_norm_pre_clip': 2.73311710357666, 'timing/data': 0.0064419719856232405, 'timing/model': 1.1355252800276503, 'learning_rate/qwen_vl_interface': 3.9540642728914115e-06, 'learning_rate/action_model': 3.9540642728914115e-05, 'epoch': 1.22}) +[2026-06-16 00:53:40,907][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.07260610163211823, 'train/grad_norm_pre_clip': 5.248191833496094, 'timing/data': 0.006141976977232844, 'timing/model': 1.1315400990424678, 'learning_rate/qwen_vl_interface': 3.951134653174664e-06, 'learning_rate/action_model': 3.951134653174663e-05, 'epoch': 1.23}) +[2026-06-16 00:53:45,485][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.013403456658124924, 'train/grad_norm_pre_clip': 3.490692377090454, 'timing/data': 0.006519264017697424, 'timing/model': 1.1391146950190887, 'learning_rate/qwen_vl_interface': 3.948205567372397e-06, 'learning_rate/action_model': 3.948205567372397e-05, 'epoch': 1.23}) +[2026-06-16 00:53:50,038][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.005828543566167355, 'train/grad_norm_pre_clip': 1.2022463083267212, 'timing/data': 0.006252526014577597, 'timing/model': 1.3080015939776786, 'learning_rate/qwen_vl_interface': 3.945277016688647e-06, 'learning_rate/action_model': 3.945277016688647e-05, 'epoch': 1.23}) +[2026-06-16 00:53:54,515][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.0012033128878101707, 'train/grad_norm_pre_clip': 0.23644395172595978, 'timing/data': 0.006335618032608181, 'timing/model': 1.2269477850059047, 'learning_rate/qwen_vl_interface': 3.9423490023272326e-06, 'learning_rate/action_model': 3.9423490023272324e-05, 'epoch': 1.23}) +[2026-06-16 00:53:58,901][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.5074280500411987, 'train/grad_norm_pre_clip': 28.52881622314453, 'timing/data': 0.0066577980178408325, 'timing/model': 1.1333627179847099, 'learning_rate/qwen_vl_interface': 3.939421525491748e-06, 'learning_rate/action_model': 3.9394215254917483e-05, 'epoch': 1.23}) +[2026-06-16 00:54:03,430][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.034804556518793106, 'train/grad_norm_pre_clip': 9.596232414245605, 'timing/data': 0.006781526957638562, 'timing/model': 1.1345308190211654, 'learning_rate/qwen_vl_interface': 3.9364945873855705e-06, 'learning_rate/action_model': 3.93649458738557e-05, 'epoch': 1.23}) +[2026-06-16 00:54:07,984][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.012438391335308552, 'train/grad_norm_pre_clip': 1.701711654663086, 'timing/data': 0.006247295008506626, 'timing/model': 1.1350426419521682, 'learning_rate/qwen_vl_interface': 3.9335681892118536e-06, 'learning_rate/action_model': 3.933568189211854e-05, 'epoch': 1.23}) +[2026-06-16 00:54:12,522][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.08487765491008759, 'train/grad_norm_pre_clip': 6.770920276641846, 'timing/data': 0.006613475037738681, 'timing/model': 1.134665960969869, 'learning_rate/qwen_vl_interface': 3.930642332173527e-06, 'learning_rate/action_model': 3.930642332173527e-05, 'epoch': 1.23}) +[2026-06-16 00:54:17,068][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.02231486327946186, 'train/grad_norm_pre_clip': 18.80499267578125, 'timing/data': 0.00651695899432525, 'timing/model': 1.2937585159670562, 'learning_rate/qwen_vl_interface': 3.927717017473301e-06, 'learning_rate/action_model': 3.9277170174733006e-05, 'epoch': 1.23}) +[2026-06-16 00:54:21,587][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.10737530887126923, 'train/grad_norm_pre_clip': 12.570401191711426, 'timing/data': 0.006284240982495248, 'timing/model': 1.2544280790025368, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 1.23}) +[2026-06-16 00:54:25,969][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.0198237095028162, 'train/grad_norm_pre_clip': 15.766949653625488, 'timing/data': 0.005960340960882604, 'timing/model': 1.137601608003024, 'learning_rate/qwen_vl_interface': 3.921868019896873e-06, 'learning_rate/action_model': 3.921868019896873e-05, 'epoch': 1.23}) +[2026-06-16 00:54:30,498][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.0003470711235422641, 'train/grad_norm_pre_clip': 0.08242127299308777, 'timing/data': 0.006339199957437813, 'timing/model': 1.1475039679789916, 'learning_rate/qwen_vl_interface': 3.918944339424974e-06, 'learning_rate/action_model': 3.9189443394249736e-05, 'epoch': 1.23}) +[2026-06-16 00:54:35,043][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.016983117908239365, 'train/grad_norm_pre_clip': 3.777625799179077, 'timing/data': 0.006273253005929291, 'timing/model': 1.1326106060296297, 'learning_rate/qwen_vl_interface': 3.9160212060997774e-06, 'learning_rate/action_model': 3.9160212060997774e-05, 'epoch': 1.23}) +[2026-06-16 00:54:39,574][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.01957184635102749, 'train/grad_norm_pre_clip': 3.4872443675994873, 'timing/data': 0.006336127989925444, 'timing/model': 1.1322063070256263, 'learning_rate/qwen_vl_interface': 3.913098621122875e-06, 'learning_rate/action_model': 3.913098621122875e-05, 'epoch': 1.23}) +[2026-06-16 00:54:44,101][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.05334842577576637, 'train/grad_norm_pre_clip': 14.16183090209961, 'timing/data': 0.006468948035035282, 'timing/model': 1.2859924810472876, 'learning_rate/qwen_vl_interface': 3.91017658569563e-06, 'learning_rate/action_model': 3.91017658569563e-05, 'epoch': 1.23}) +[2026-06-16 00:54:48,486][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.002475421642884612, 'train/grad_norm_pre_clip': 0.4934646189212799, 'timing/data': 0.006661360035650432, 'timing/model': 1.1336913640261628, 'learning_rate/qwen_vl_interface': 3.907255101019181e-06, 'learning_rate/action_model': 3.907255101019181e-05, 'epoch': 1.23}) +[2026-06-16 00:54:53,034][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.30439847707748413, 'train/grad_norm_pre_clip': 42.12793731689453, 'timing/data': 0.006670169997960329, 'timing/model': 1.1461577959707938, 'learning_rate/qwen_vl_interface': 3.904334168294442e-06, 'learning_rate/action_model': 3.9043341682944415e-05, 'epoch': 1.23}) +[2026-06-16 00:54:57,571][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.0038089915178716183, 'train/grad_norm_pre_clip': 1.3282116651535034, 'timing/data': 0.006631409982219338, 'timing/model': 1.136840881954413, 'learning_rate/qwen_vl_interface': 3.901413788722093e-06, 'learning_rate/action_model': 3.901413788722093e-05, 'epoch': 1.23}) +[2026-06-16 00:59:23,545][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.10543698817491531, 'train/grad_norm_pre_clip': 10.668495178222656, 'eval/loss': 0.06220685690641403, 'eval/flap_precision': 0.9747634069400631, 'eval/flap_recall': 0.9537037037037037, 'eval/flap_f1': 0.9641185647425896, 'eval/flap_rate_model': 0.1761111111111111, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9641185647425896, 'timing/data': 0.006267393997404724, 'timing/model': 1.1318426909856498, 'learning_rate/qwen_vl_interface': 3.8984939635025966e-06, 'learning_rate/action_model': 3.898493963502596e-05, 'epoch': 1.23}) +[2026-06-16 00:59:23,546][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3000_state +[2026-06-16 00:59:23,548][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 00:59:27,114][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3000_state/model.safetensors +[2026-06-16 00:59:37,054][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3000_state/optimizer.bin +[2026-06-16 00:59:37,055][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3000_state/sampler.bin +[2026-06-16 00:59:37,055][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-16 00:59:37,056][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-16 00:59:37,098][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-16 00:59:37,099][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 00:59:37,105][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 00:59:41,553][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.07229249179363251, 'train/grad_norm_pre_clip': 19.744789123535156, 'timing/data': 0.0069800050114281476, 'timing/model': 1.1210359740070999, 'learning_rate/qwen_vl_interface': 3.895574693836181e-06, 'learning_rate/action_model': 3.8955746938361804e-05, 'epoch': 1.23}) +[2026-06-16 00:59:46,018][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.03139558434486389, 'train/grad_norm_pre_clip': 15.362898826599121, 'timing/data': 0.006527426012326032, 'timing/model': 1.1189400770235807, 'learning_rate/qwen_vl_interface': 3.892655980922847e-06, 'learning_rate/action_model': 3.8926559809228466e-05, 'epoch': 1.23}) +[2026-06-16 00:59:50,503][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.05231631547212601, 'train/grad_norm_pre_clip': 15.73497200012207, 'timing/data': 0.006901214015670121, 'timing/model': 1.2819804340251721, 'learning_rate/qwen_vl_interface': 3.889737825962368e-06, 'learning_rate/action_model': 3.889737825962368e-05, 'epoch': 1.23}) +[2026-06-16 00:59:54,940][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.003710723714902997, 'train/grad_norm_pre_clip': 0.9098035097122192, 'timing/data': 0.006393020041286945, 'timing/model': 1.230304222961422, 'learning_rate/qwen_vl_interface': 3.886820230154288e-06, 'learning_rate/action_model': 3.886820230154287e-05, 'epoch': 1.23}) +[2026-06-16 00:59:59,273][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.011831345967948437, 'train/grad_norm_pre_clip': 5.344768047332764, 'timing/data': 0.006572726997546852, 'timing/model': 1.119638149975799, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 1.23}) +[2026-06-16 01:00:03,750][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.08456142246723175, 'train/grad_norm_pre_clip': 12.746051788330078, 'timing/data': 0.006656887009739876, 'timing/model': 1.1189267259906046, 'learning_rate/qwen_vl_interface': 3.880986720792344e-06, 'learning_rate/action_model': 3.880986720792344e-05, 'epoch': 1.23}) +[2026-06-16 01:00:08,231][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.0016625733114778996, 'train/grad_norm_pre_clip': 0.5070801973342896, 'timing/data': 0.00631290499586612, 'timing/model': 1.1202790329698473, 'learning_rate/qwen_vl_interface': 3.878070809636416e-06, 'learning_rate/action_model': 3.878070809636416e-05, 'epoch': 1.24}) +[2026-06-16 01:00:12,728][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.3497988283634186, 'train/grad_norm_pre_clip': 25.683650970458984, 'timing/data': 0.00624538300326094, 'timing/model': 1.1345898319850676, 'learning_rate/qwen_vl_interface': 3.875155462428758e-06, 'learning_rate/action_model': 3.875155462428758e-05, 'epoch': 1.24}) +[2026-06-16 01:00:17,236][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.25552937388420105, 'train/grad_norm_pre_clip': 19.929262161254883, 'timing/data': 0.006344199005980045, 'timing/model': 1.288130472006742, 'learning_rate/qwen_vl_interface': 3.872240680367754e-06, 'learning_rate/action_model': 3.872240680367754e-05, 'epoch': 1.24}) +[2026-06-16 01:00:21,574][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.002811641898006201, 'train/grad_norm_pre_clip': 1.017220377922058, 'timing/data': 0.006273848004639149, 'timing/model': 1.1249392780009657, 'learning_rate/qwen_vl_interface': 3.869326464651566e-06, 'learning_rate/action_model': 3.869326464651566e-05, 'epoch': 1.24}) +[2026-06-16 01:00:26,104][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.04521750286221504, 'train/grad_norm_pre_clip': 16.684001922607422, 'timing/data': 0.006843521958217025, 'timing/model': 1.1268137070001103, 'learning_rate/qwen_vl_interface': 3.866412816478114e-06, 'learning_rate/action_model': 3.8664128164781144e-05, 'epoch': 1.24}) +[2026-06-16 01:00:30,540][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.0003444594331085682, 'train/grad_norm_pre_clip': 0.13612692058086395, 'timing/data': 0.009476531005930156, 'timing/model': 1.1480975240119733, 'learning_rate/qwen_vl_interface': 3.8634997370450915e-06, 'learning_rate/action_model': 3.8634997370450914e-05, 'epoch': 1.24}) +[2026-06-16 01:00:34,983][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.015255516394972801, 'train/grad_norm_pre_clip': 5.80682373046875, 'timing/data': 0.006596209015697241, 'timing/model': 1.124641329050064, 'learning_rate/qwen_vl_interface': 3.860587227549956e-06, 'learning_rate/action_model': 3.860587227549955e-05, 'epoch': 1.24}) +[2026-06-16 01:00:39,471][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.008607665076851845, 'train/grad_norm_pre_clip': 5.893820285797119, 'timing/data': 0.006367928988765925, 'timing/model': 1.135669093055185, 'learning_rate/qwen_vl_interface': 3.857675289189926e-06, 'learning_rate/action_model': 3.857675289189926e-05, 'epoch': 1.24}) +[2026-06-16 01:00:44,095][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.001048605190590024, 'train/grad_norm_pre_clip': 0.29546859860420227, 'timing/data': 0.006262744020204991, 'timing/model': 1.3062048580031842, 'learning_rate/qwen_vl_interface': 3.854763923161993e-06, 'learning_rate/action_model': 3.854763923161992e-05, 'epoch': 1.24}) +[2026-06-16 01:00:48,469][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.16755157709121704, 'train/grad_norm_pre_clip': 19.566425323486328, 'timing/data': 0.006807302008382976, 'timing/model': 1.1334227009792812, 'learning_rate/qwen_vl_interface': 3.851853130662908e-06, 'learning_rate/action_model': 3.8518531306629077e-05, 'epoch': 1.24}) +[2026-06-16 01:00:52,989][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.008958225138485432, 'train/grad_norm_pre_clip': 4.365091323852539, 'timing/data': 0.006502064992673695, 'timing/model': 1.1316420669900253, 'learning_rate/qwen_vl_interface': 3.8489429128891884e-06, 'learning_rate/action_model': 3.848942912889188e-05, 'epoch': 1.24}) +[2026-06-16 01:00:57,530][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.12637662887573242, 'train/grad_norm_pre_clip': 17.21495246887207, 'timing/data': 0.0061675619799643755, 'timing/model': 1.1327230199822225, 'learning_rate/qwen_vl_interface': 3.846033271037114e-06, 'learning_rate/action_model': 3.8460332710371136e-05, 'epoch': 1.24}) +[2026-06-16 01:01:02,007][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.31197190284729004, 'train/grad_norm_pre_clip': 25.765810012817383, 'timing/data': 0.0063631319790147245, 'timing/model': 1.128190741990693, 'learning_rate/qwen_vl_interface': 3.843124206302732e-06, 'learning_rate/action_model': 3.8431242063027316e-05, 'epoch': 1.24}) +[2026-06-16 01:01:06,516][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.2063010334968567, 'train/grad_norm_pre_clip': 29.900588989257812, 'timing/data': 0.006299842963926494, 'timing/model': 1.1285518170334399, 'learning_rate/qwen_vl_interface': 3.840215719881844e-06, 'learning_rate/action_model': 3.8402157198818445e-05, 'epoch': 1.24}) +[2026-06-16 01:01:11,044][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.22462402284145355, 'train/grad_norm_pre_clip': 24.87489128112793, 'timing/data': 0.0059503650409169495, 'timing/model': 1.2913542179740034, 'learning_rate/qwen_vl_interface': 3.837307812970022e-06, 'learning_rate/action_model': 3.837307812970022e-05, 'epoch': 1.24}) +[2026-06-16 01:01:15,414][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.056693993508815765, 'train/grad_norm_pre_clip': 22.244049072265625, 'timing/data': 0.006115381955169141, 'timing/model': 1.1289198879967444, 'learning_rate/qwen_vl_interface': 3.834400486762596e-06, 'learning_rate/action_model': 3.834400486762596e-05, 'epoch': 1.24}) +[2026-06-16 01:01:19,959][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.0080189174041152, 'train/grad_norm_pre_clip': 4.43398380279541, 'timing/data': 0.006583831971511245, 'timing/model': 1.1301788869895972, 'learning_rate/qwen_vl_interface': 3.8314937424546585e-06, 'learning_rate/action_model': 3.8314937424546584e-05, 'epoch': 1.24}) +[2026-06-16 01:01:24,493][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.09327986091375351, 'train/grad_norm_pre_clip': 29.260650634765625, 'timing/data': 0.006166228966321796, 'timing/model': 1.1315624389681034, 'learning_rate/qwen_vl_interface': 3.828587581241064e-06, 'learning_rate/action_model': 3.8285875812410635e-05, 'epoch': 1.24}) +[2026-06-16 01:01:28,981][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.016662834212183952, 'train/grad_norm_pre_clip': 16.242467880249023, 'timing/data': 0.006733774032909423, 'timing/model': 1.147343076998368, 'learning_rate/qwen_vl_interface': 3.825682004316419e-06, 'learning_rate/action_model': 3.825682004316419e-05, 'epoch': 1.24}) +[2026-06-16 01:01:33,520][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.1014445573091507, 'train/grad_norm_pre_clip': 15.09769344329834, 'timing/data': 0.006415399024263024, 'timing/model': 1.1367154339677654, 'learning_rate/qwen_vl_interface': 3.822777012875104e-06, 'learning_rate/action_model': 3.822777012875103e-05, 'epoch': 1.24}) +[2026-06-16 01:01:38,077][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.09995687752962112, 'train/grad_norm_pre_clip': 10.093777656555176, 'timing/data': 0.006159815995488316, 'timing/model': 1.3028514119796455, 'learning_rate/qwen_vl_interface': 3.819872608111248e-06, 'learning_rate/action_model': 3.819872608111248e-05, 'epoch': 1.24}) +[2026-06-16 01:01:42,474][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.13213138282299042, 'train/grad_norm_pre_clip': 17.611572265625, 'timing/data': 0.0063927240553312, 'timing/model': 1.1327087720273994, 'learning_rate/qwen_vl_interface': 3.816968791218741e-06, 'learning_rate/action_model': 3.816968791218741e-05, 'epoch': 1.24}) +[2026-06-16 01:01:47,009][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.007725038565695286, 'train/grad_norm_pre_clip': 3.8368563652038574, 'timing/data': 0.006252787017729133, 'timing/model': 1.1325002890080214, 'learning_rate/qwen_vl_interface': 3.8140655633912337e-06, 'learning_rate/action_model': 3.814065563391234e-05, 'epoch': 1.24}) +[2026-06-16 01:01:51,498][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.5256896615028381, 'train/grad_norm_pre_clip': 17.298431396484375, 'timing/data': 0.006236212968360633, 'timing/model': 1.1316027939901687, 'learning_rate/qwen_vl_interface': 3.811162925822133e-06, 'learning_rate/action_model': 3.811162925822132e-05, 'epoch': 1.24}) +[2026-06-16 01:01:55,988][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.16642510890960693, 'train/grad_norm_pre_clip': 19.86754035949707, 'timing/data': 0.006542159011587501, 'timing/model': 1.131101390987169, 'learning_rate/qwen_vl_interface': 3.8082608797046045e-06, 'learning_rate/action_model': 3.8082608797046045e-05, 'epoch': 1.25}) +[2026-06-16 01:02:00,521][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.004368781112134457, 'train/grad_norm_pre_clip': 1.0197291374206543, 'timing/data': 0.00631789502222091, 'timing/model': 1.129669115995057, 'learning_rate/qwen_vl_interface': 3.8053594262315687e-06, 'learning_rate/action_model': 3.805359426231568e-05, 'epoch': 1.25}) +[2026-06-16 01:02:05,058][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.0029092805925756693, 'train/grad_norm_pre_clip': 0.4689061641693115, 'timing/data': 0.00645011302549392, 'timing/model': 1.2907210810226388, 'learning_rate/qwen_vl_interface': 3.802458566595704e-06, 'learning_rate/action_model': 3.802458566595704e-05, 'epoch': 1.25}) +[2026-06-16 01:02:09,434][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.0026572917122393847, 'train/grad_norm_pre_clip': 1.5414938926696777, 'timing/data': 0.006694853014778346, 'timing/model': 1.1388086169608869, 'learning_rate/qwen_vl_interface': 3.7995583019894444e-06, 'learning_rate/action_model': 3.7995583019894445e-05, 'epoch': 1.25}) +[2026-06-16 01:02:13,983][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.004157846327871084, 'train/grad_norm_pre_clip': 1.2183398008346558, 'timing/data': 0.007013168011326343, 'timing/model': 1.132965440978296, 'learning_rate/qwen_vl_interface': 3.7966586336049782e-06, 'learning_rate/action_model': 3.796658633604978e-05, 'epoch': 1.25}) +[2026-06-16 01:02:18,495][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.15735043585300446, 'train/grad_norm_pre_clip': 18.73576545715332, 'timing/data': 0.006181031989399344, 'timing/model': 1.135089791030623, 'learning_rate/qwen_vl_interface': 3.7937595626342516e-06, 'learning_rate/action_model': 3.793759562634252e-05, 'epoch': 1.25}) +[2026-06-16 01:02:23,022][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.02315709739923477, 'train/grad_norm_pre_clip': 10.06435489654541, 'timing/data': 0.006498991046100855, 'timing/model': 1.1395461010397412, 'learning_rate/qwen_vl_interface': 3.79086109026896e-06, 'learning_rate/action_model': 3.7908610902689604e-05, 'epoch': 1.25}) +[2026-06-16 01:02:27,573][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.09264874458312988, 'train/grad_norm_pre_clip': 11.928946495056152, 'timing/data': 0.006096920988056809, 'timing/model': 1.144537997955922, 'learning_rate/qwen_vl_interface': 3.7879632177005584e-06, 'learning_rate/action_model': 3.787963217700558e-05, 'epoch': 1.25}) +[2026-06-16 01:02:32,130][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.0124974325299263, 'train/grad_norm_pre_clip': 5.871533393859863, 'timing/data': 0.006705542968120426, 'timing/model': 1.2993892169906758, 'learning_rate/qwen_vl_interface': 3.7850659461202523e-06, 'learning_rate/action_model': 3.785065946120252e-05, 'epoch': 1.25}) +[2026-06-16 01:02:36,507][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.1170707643032074, 'train/grad_norm_pre_clip': 10.665650367736816, 'timing/data': 0.006404451967682689, 'timing/model': 1.1355482609942555, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 1.25}) +[2026-06-16 01:02:41,049][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.009256536141037941, 'train/grad_norm_pre_clip': 2.207545280456543, 'timing/data': 0.006025462003890425, 'timing/model': 1.1304711720440537, 'learning_rate/qwen_vl_interface': 3.779273210687514e-06, 'learning_rate/action_model': 3.779273210687514e-05, 'epoch': 1.25}) +[2026-06-16 01:02:45,558][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.012366393581032753, 'train/grad_norm_pre_clip': 5.734675884246826, 'timing/data': 0.006174769951030612, 'timing/model': 1.138973054010421, 'learning_rate/qwen_vl_interface': 3.776377749216257e-06, 'learning_rate/action_model': 3.776377749216257e-05, 'epoch': 1.25}) +[2026-06-16 01:02:50,107][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.04400201886892319, 'train/grad_norm_pre_clip': 17.029977798461914, 'timing/data': 0.006456362025346607, 'timing/model': 1.1389151089824736, 'learning_rate/qwen_vl_interface': 3.773482893495442e-06, 'learning_rate/action_model': 3.773482893495442e-05, 'epoch': 1.25}) +[2026-06-16 01:02:54,646][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.007132749073207378, 'train/grad_norm_pre_clip': 1.7776315212249756, 'timing/data': 0.006589506985619664, 'timing/model': 1.1330528469989076, 'learning_rate/qwen_vl_interface': 3.7705886447150375e-06, 'learning_rate/action_model': 3.7705886447150376e-05, 'epoch': 1.25}) +[2026-06-16 01:02:59,171][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.016164444386959076, 'train/grad_norm_pre_clip': 6.9410810470581055, 'timing/data': 0.006421199010219425, 'timing/model': 1.2880091789993457, 'learning_rate/qwen_vl_interface': 3.767695004064759e-06, 'learning_rate/action_model': 3.767695004064759e-05, 'epoch': 1.25}) +[2026-06-16 01:03:03,642][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.11116880923509598, 'train/grad_norm_pre_clip': 11.26069450378418, 'timing/data': 0.006403581006452441, 'timing/model': 1.2222468799445778, 'learning_rate/qwen_vl_interface': 3.7648019727340725e-06, 'learning_rate/action_model': 3.764801972734072e-05, 'epoch': 1.25}) +[2026-06-16 01:03:08,011][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.18695032596588135, 'train/grad_norm_pre_clip': 7.827737808227539, 'timing/data': 0.006225693970918655, 'timing/model': 1.1320037649711594, 'learning_rate/qwen_vl_interface': 3.761909551912196e-06, 'learning_rate/action_model': 3.761909551912196e-05, 'epoch': 1.25}) +[2026-06-16 01:03:12,519][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.2976425588130951, 'train/grad_norm_pre_clip': 31.22319984436035, 'timing/data': 0.0066441940143704414, 'timing/model': 1.1343142639962025, 'learning_rate/qwen_vl_interface': 3.75901774278809e-06, 'learning_rate/action_model': 3.7590177427880904e-05, 'epoch': 1.25}) +[2026-06-16 01:03:17,053][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.30926641821861267, 'train/grad_norm_pre_clip': 24.749847412109375, 'timing/data': 0.006396716984454542, 'timing/model': 1.1307267210213467, 'learning_rate/qwen_vl_interface': 3.756126546550472e-06, 'learning_rate/action_model': 3.756126546550472e-05, 'epoch': 1.25}) +[2026-06-16 01:03:21,585][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.0036445490550249815, 'train/grad_norm_pre_clip': 0.6054701209068298, 'timing/data': 0.006200054951477796, 'timing/model': 1.1310127089964226, 'learning_rate/qwen_vl_interface': 3.753235964387802e-06, 'learning_rate/action_model': 3.753235964387802e-05, 'epoch': 1.25}) +[2026-06-16 01:03:26,124][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.21944968402385712, 'train/grad_norm_pre_clip': 7.96707010269165, 'timing/data': 0.00661663303617388, 'timing/model': 1.292009694967419, 'learning_rate/qwen_vl_interface': 3.750345997488289e-06, 'learning_rate/action_model': 3.7503459974882886e-05, 'epoch': 1.25}) +[2026-06-16 01:03:30,620][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.15815050899982452, 'train/grad_norm_pre_clip': 9.589076042175293, 'timing/data': 0.006545834999997169, 'timing/model': 1.2531783119775355, 'learning_rate/qwen_vl_interface': 3.7474566470398903e-06, 'learning_rate/action_model': 3.74745664703989e-05, 'epoch': 1.25}) +[2026-06-16 01:03:35,007][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.018840119242668152, 'train/grad_norm_pre_clip': 4.815467357635498, 'timing/data': 0.0064121949835680425, 'timing/model': 1.1350651940447278, 'learning_rate/qwen_vl_interface': 3.744567914230307e-06, 'learning_rate/action_model': 3.744567914230307e-05, 'epoch': 1.25}) +[2026-06-16 01:03:39,527][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.0021576553117483854, 'train/grad_norm_pre_clip': 0.3625062108039856, 'timing/data': 0.006633340031839907, 'timing/model': 1.1362506939913146, 'learning_rate/qwen_vl_interface': 3.7416798002469902e-06, 'learning_rate/action_model': 3.74167980024699e-05, 'epoch': 1.25}) +[2026-06-16 01:03:44,054][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.009524955414235592, 'train/grad_norm_pre_clip': 2.324420690536499, 'timing/data': 0.007207205984741449, 'timing/model': 1.1317658589687198, 'learning_rate/qwen_vl_interface': 3.7387923062771313e-06, 'learning_rate/action_model': 3.738792306277131e-05, 'epoch': 1.25}) +[2026-06-16 01:03:48,595][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.11588714271783829, 'train/grad_norm_pre_clip': 15.27771282196045, 'timing/data': 0.006096491008065641, 'timing/model': 1.1316135129891336, 'learning_rate/qwen_vl_interface': 3.7359054335076732e-06, 'learning_rate/action_model': 3.7359054335076734e-05, 'epoch': 1.26}) +[2026-06-16 01:03:53,131][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.033536478877067566, 'train/grad_norm_pre_clip': 18.586658477783203, 'timing/data': 0.0065604610135778785, 'timing/model': 1.2968718699994497, 'learning_rate/qwen_vl_interface': 3.7330191831253006e-06, 'learning_rate/action_model': 3.7330191831253e-05, 'epoch': 1.26}) +[2026-06-16 01:03:57,531][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.0633196160197258, 'train/grad_norm_pre_clip': 7.251199245452881, 'timing/data': 0.006548242992721498, 'timing/model': 1.1403938659932464, 'learning_rate/qwen_vl_interface': 3.7301335563164382e-06, 'learning_rate/action_model': 3.730133556316438e-05, 'epoch': 1.26}) +[2026-06-16 01:04:02,089][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.0002770521095953882, 'train/grad_norm_pre_clip': 0.040336642414331436, 'timing/data': 0.00628590804990381, 'timing/model': 1.1429614129592665, 'learning_rate/qwen_vl_interface': 3.727248554267261e-06, 'learning_rate/action_model': 3.727248554267261e-05, 'epoch': 1.26}) +[2026-06-16 01:04:06,627][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.10882126539945602, 'train/grad_norm_pre_clip': 18.627511978149414, 'timing/data': 0.006671904993709177, 'timing/model': 1.1338251080014743, 'learning_rate/qwen_vl_interface': 3.7243641781636838e-06, 'learning_rate/action_model': 3.7243641781636836e-05, 'epoch': 1.26}) +[2026-06-16 01:04:11,166][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.08368604630231857, 'train/grad_norm_pre_clip': 7.631890773773193, 'timing/data': 0.006394771975465119, 'timing/model': 1.1358833700069226, 'learning_rate/qwen_vl_interface': 3.721480429191364e-06, 'learning_rate/action_model': 3.721480429191364e-05, 'epoch': 1.26}) +[2026-06-16 01:04:15,649][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.0010644090361893177, 'train/grad_norm_pre_clip': 0.18410377204418182, 'timing/data': 0.006792218948248774, 'timing/model': 1.1344214350101538, 'learning_rate/qwen_vl_interface': 3.7185973085357045e-06, 'learning_rate/action_model': 3.718597308535705e-05, 'epoch': 1.26}) +[2026-06-16 01:04:20,148][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.16237445175647736, 'train/grad_norm_pre_clip': 14.611485481262207, 'timing/data': 0.006518818961922079, 'timing/model': 1.2546789920306765, 'learning_rate/qwen_vl_interface': 3.7157148173818443e-06, 'learning_rate/action_model': 3.715714817381844e-05, 'epoch': 1.26}) +[2026-06-16 01:04:24,530][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.019162660464644432, 'train/grad_norm_pre_clip': 5.912120342254639, 'timing/data': 0.006450246029999107, 'timing/model': 1.1344566600164399, 'learning_rate/qwen_vl_interface': 3.7128329569146672e-06, 'learning_rate/action_model': 3.7128329569146666e-05, 'epoch': 1.26}) +[2026-06-16 01:04:29,078][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.08134229481220245, 'train/grad_norm_pre_clip': 13.281036376953125, 'timing/data': 0.006210403051227331, 'timing/model': 1.1348610459826887, 'learning_rate/qwen_vl_interface': 3.7099517283187977e-06, 'learning_rate/action_model': 3.7099517283187976e-05, 'epoch': 1.26}) +[2026-06-16 01:04:33,614][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.04413006454706192, 'train/grad_norm_pre_clip': 12.473074913024902, 'timing/data': 0.006290896970313042, 'timing/model': 1.1369558950536884, 'learning_rate/qwen_vl_interface': 3.7070711327786e-06, 'learning_rate/action_model': 3.7070711327785997e-05, 'epoch': 1.26}) +[2026-06-16 01:04:38,150][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.052250124514102936, 'train/grad_norm_pre_clip': 13.08995246887207, 'timing/data': 0.006559371016919613, 'timing/model': 1.1321760909631848, 'learning_rate/qwen_vl_interface': 3.704191171478179e-06, 'learning_rate/action_model': 3.704191171478179e-05, 'epoch': 1.26}) +[2026-06-16 01:04:42,650][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.08063451200723648, 'train/grad_norm_pre_clip': 5.948819637298584, 'timing/data': 0.006993272982072085, 'timing/model': 1.1430570799857378, 'learning_rate/qwen_vl_interface': 3.701311845601376e-06, 'learning_rate/action_model': 3.701311845601376e-05, 'epoch': 1.26}) +[2026-06-16 01:04:47,185][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.16040754318237305, 'train/grad_norm_pre_clip': 20.39393424987793, 'timing/data': 0.006114977994002402, 'timing/model': 1.2701732149580494, 'learning_rate/qwen_vl_interface': 3.6984331563317744e-06, 'learning_rate/action_model': 3.698433156331774e-05, 'epoch': 1.26}) +[2026-06-16 01:04:51,556][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.08588562905788422, 'train/grad_norm_pre_clip': 8.439851760864258, 'timing/data': 0.005866941006388515, 'timing/model': 1.1312856239965186, 'learning_rate/qwen_vl_interface': 3.695555104852695e-06, 'learning_rate/action_model': 3.6955551048526946e-05, 'epoch': 1.26}) +[2026-06-16 01:04:56,093][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.19495971500873566, 'train/grad_norm_pre_clip': 29.40169906616211, 'timing/data': 0.006368684000335634, 'timing/model': 1.1314566230284981, 'learning_rate/qwen_vl_interface': 3.6926776923471948e-06, 'learning_rate/action_model': 3.6926776923471946e-05, 'epoch': 1.26}) +[2026-06-16 01:05:00,623][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.0930599793791771, 'train/grad_norm_pre_clip': 10.87219524383545, 'timing/data': 0.006543712050188333, 'timing/model': 1.1332179290475324, 'learning_rate/qwen_vl_interface': 3.6898009199980715e-06, 'learning_rate/action_model': 3.6898009199980714e-05, 'epoch': 1.26}) +[2026-06-16 01:05:05,175][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.0019950098358094692, 'train/grad_norm_pre_clip': 0.3318280875682831, 'timing/data': 0.006340251013170928, 'timing/model': 1.1348164060036652, 'learning_rate/qwen_vl_interface': 3.686924788987855e-06, 'learning_rate/action_model': 3.686924788987855e-05, 'epoch': 1.26}) +[2026-06-16 01:05:09,718][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.004834848456084728, 'train/grad_norm_pre_clip': 0.7545426487922668, 'timing/data': 0.006645080982707441, 'timing/model': 1.2860007359995507, 'learning_rate/qwen_vl_interface': 3.6840493004988155e-06, 'learning_rate/action_model': 3.684049300498815e-05, 'epoch': 1.26}) +[2026-06-16 01:05:14,110][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.11461688578128815, 'train/grad_norm_pre_clip': 7.865298748016357, 'timing/data': 0.00623718200949952, 'timing/model': 1.1346838170429692, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 1.26}) +[2026-06-16 01:05:18,646][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.10211972147226334, 'train/grad_norm_pre_clip': 8.651558876037598, 'timing/data': 0.006451033987104893, 'timing/model': 1.1328061410458758, 'learning_rate/qwen_vl_interface': 3.6783002558120206e-06, 'learning_rate/action_model': 3.67830025581202e-05, 'epoch': 1.26}) +[2026-06-16 01:05:23,095][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.051836516708135605, 'train/grad_norm_pre_clip': 12.211000442504883, 'timing/data': 0.006126387044787407, 'timing/model': 1.1386376000009477, 'learning_rate/qwen_vl_interface': 3.6754267019774804e-06, 'learning_rate/action_model': 3.67542670197748e-05, 'epoch': 1.26}) +[2026-06-16 01:05:27,576][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.07159144431352615, 'train/grad_norm_pre_clip': 20.607160568237305, 'timing/data': 0.006248447985854, 'timing/model': 1.1363166800001636, 'learning_rate/qwen_vl_interface': 3.6725537953905453e-06, 'learning_rate/action_model': 3.672553795390545e-05, 'epoch': 1.26}) +[2026-06-16 01:05:32,072][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.025318406522274017, 'train/grad_norm_pre_clip': 10.096336364746094, 'timing/data': 0.0064372740453109145, 'timing/model': 1.133391882001888, 'learning_rate/qwen_vl_interface': 3.6696815372321603e-06, 'learning_rate/action_model': 3.6696815372321604e-05, 'epoch': 1.26}) +[2026-06-16 01:05:36,630][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.0021113737020641565, 'train/grad_norm_pre_clip': 0.40978115797042847, 'timing/data': 0.006650105002336204, 'timing/model': 1.3004403339582495, 'learning_rate/qwen_vl_interface': 3.6668099286830023e-06, 'learning_rate/action_model': 3.666809928683002e-05, 'epoch': 1.27}) +[2026-06-16 01:05:41,011][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.10809588432312012, 'train/grad_norm_pre_clip': 8.77493953704834, 'timing/data': 0.006417039956431836, 'timing/model': 1.1350509069743566, 'learning_rate/qwen_vl_interface': 3.66393897092348e-06, 'learning_rate/action_model': 3.66393897092348e-05, 'epoch': 1.27}) +[2026-06-16 01:05:45,570][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.15320250391960144, 'train/grad_norm_pre_clip': 19.421239852905273, 'timing/data': 0.009150194004178047, 'timing/model': 1.154651315009687, 'learning_rate/qwen_vl_interface': 3.661068665133735e-06, 'learning_rate/action_model': 3.661068665133735e-05, 'epoch': 1.27}) +[2026-06-16 01:05:50,117][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.002390841953456402, 'train/grad_norm_pre_clip': 0.3690825402736664, 'timing/data': 0.00660492799943313, 'timing/model': 1.1350032369955443, 'learning_rate/qwen_vl_interface': 3.6581990124936434e-06, 'learning_rate/action_model': 3.6581990124936434e-05, 'epoch': 1.27}) +[2026-06-16 01:05:54,624][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.07435104250907898, 'train/grad_norm_pre_clip': 9.959075927734375, 'timing/data': 0.006140033015981317, 'timing/model': 1.137726078974083, 'learning_rate/qwen_vl_interface': 3.65533001418281e-06, 'learning_rate/action_model': 3.655330014182809e-05, 'epoch': 1.27}) +[2026-06-16 01:05:59,145][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.08738153427839279, 'train/grad_norm_pre_clip': 18.4234676361084, 'timing/data': 0.006481641961727291, 'timing/model': 1.141118308994919, 'learning_rate/qwen_vl_interface': 3.652461671380572e-06, 'learning_rate/action_model': 3.652461671380572e-05, 'epoch': 1.27}) +[2026-06-16 01:06:03,732][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.016802603378891945, 'train/grad_norm_pre_clip': 4.131275653839111, 'timing/data': 0.006460289005190134, 'timing/model': 1.3022593920468353, 'learning_rate/qwen_vl_interface': 3.6495939852659966e-06, 'learning_rate/action_model': 3.649593985265997e-05, 'epoch': 1.27}) +[2026-06-16 01:06:08,115][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.3609699606895447, 'train/grad_norm_pre_clip': 21.701650619506836, 'timing/data': 0.007892285997513682, 'timing/model': 1.1358388049993664, 'learning_rate/qwen_vl_interface': 3.646726957017881e-06, 'learning_rate/action_model': 3.646726957017881e-05, 'epoch': 1.27}) +[2026-06-16 01:06:12,659][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.25951528549194336, 'train/grad_norm_pre_clip': 10.803145408630371, 'timing/data': 0.006492639018688351, 'timing/model': 1.138126061996445, 'learning_rate/qwen_vl_interface': 3.643860587814752e-06, 'learning_rate/action_model': 3.643860587814752e-05, 'epoch': 1.27}) +[2026-06-16 01:06:17,213][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.1490299105644226, 'train/grad_norm_pre_clip': 21.532203674316406, 'timing/data': 0.006635237019509077, 'timing/model': 1.1404990929877385, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 1.27}) +[2026-06-16 01:06:21,728][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.3551347851753235, 'train/grad_norm_pre_clip': 14.111674308776855, 'timing/data': 0.006012938043568283, 'timing/model': 1.132260704995133, 'learning_rate/qwen_vl_interface': 3.6381298312562095e-06, 'learning_rate/action_model': 3.6381298312562094e-05, 'epoch': 1.27}) +[2026-06-16 01:06:26,276][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.0534081757068634, 'train/grad_norm_pre_clip': 7.25015926361084, 'timing/data': 0.006274159008171409, 'timing/model': 1.1414636669796892, 'learning_rate/qwen_vl_interface': 3.635265446256493e-06, 'learning_rate/action_model': 3.6352654462564926e-05, 'epoch': 1.27}) +[2026-06-16 01:06:30,798][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.029488876461982727, 'train/grad_norm_pre_clip': 12.381256103515625, 'timing/data': 0.006385525979567319, 'timing/model': 1.2787282089702785, 'learning_rate/qwen_vl_interface': 3.6324017250131565e-06, 'learning_rate/action_model': 3.6324017250131567e-05, 'epoch': 1.27}) +[2026-06-16 01:06:35,176][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.04146496206521988, 'train/grad_norm_pre_clip': 9.53432559967041, 'timing/data': 0.006464793987106532, 'timing/model': 1.1390186219941825, 'learning_rate/qwen_vl_interface': 3.6295386687033696e-06, 'learning_rate/action_model': 3.6295386687033695e-05, 'epoch': 1.27}) +[2026-06-16 01:06:39,735][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.010726082138717175, 'train/grad_norm_pre_clip': 1.353737711906433, 'timing/data': 0.006364489032421261, 'timing/model': 1.1427726719994098, 'learning_rate/qwen_vl_interface': 3.6266762785040253e-06, 'learning_rate/action_model': 3.626676278504025e-05, 'epoch': 1.27}) +[2026-06-16 01:06:44,234][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.010599397122859955, 'train/grad_norm_pre_clip': 1.806335210800171, 'timing/data': 0.006024172005709261, 'timing/model': 1.1434630180010572, 'learning_rate/qwen_vl_interface': 3.6238145555917464e-06, 'learning_rate/action_model': 3.6238145555917465e-05, 'epoch': 1.27}) +[2026-06-16 01:06:48,767][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.011252855882048607, 'train/grad_norm_pre_clip': 1.029201626777649, 'timing/data': 0.006271305028349161, 'timing/model': 1.139125319023151, 'learning_rate/qwen_vl_interface': 3.6209535011428757e-06, 'learning_rate/action_model': 3.620953501142876e-05, 'epoch': 1.27}) +[2026-06-16 01:06:53,305][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.006402676459401846, 'train/grad_norm_pre_clip': 1.0881695747375488, 'timing/data': 0.00636756798485294, 'timing/model': 1.1296408429625444, 'learning_rate/qwen_vl_interface': 3.6180931163334863e-06, 'learning_rate/action_model': 3.6180931163334863e-05, 'epoch': 1.27}) +[2026-06-16 01:06:57,868][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.19084666669368744, 'train/grad_norm_pre_clip': 13.110559463500977, 'timing/data': 0.0064796010265126824, 'timing/model': 1.29436313000042, 'learning_rate/qwen_vl_interface': 3.6152334023393747e-06, 'learning_rate/action_model': 3.6152334023393745e-05, 'epoch': 1.27}) +[2026-06-16 01:07:02,269][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.03813014551997185, 'train/grad_norm_pre_clip': 6.756414890289307, 'timing/data': 0.006205583980772644, 'timing/model': 1.1360680580255575, 'learning_rate/qwen_vl_interface': 3.6123743603360613e-06, 'learning_rate/action_model': 3.612374360336061e-05, 'epoch': 1.27}) +[2026-06-16 01:11:20,466][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.0023133268114179373, 'train/grad_norm_pre_clip': 0.4072670638561249, 'eval/loss': 0.04406972602009773, 'eval/flap_precision': 0.9665653495440729, 'eval/flap_recall': 0.9814814814814815, 'eval/flap_f1': 0.9739663093415009, 'eval/flap_rate_model': 0.1827777777777778, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9739663093415009, 'timing/data': 0.006237286026589572, 'timing/model': 1.1386688149650581, 'learning_rate/qwen_vl_interface': 3.6095159914987893e-06, 'learning_rate/action_model': 3.6095159914987894e-05, 'epoch': 1.27}) +[2026-06-16 01:11:20,467][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3100_state +[2026-06-16 01:11:20,469][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 01:11:24,053][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3100_state/model.safetensors +[2026-06-16 01:11:34,025][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3100_state/optimizer.bin +[2026-06-16 01:11:34,025][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3100_state/sampler.bin +[2026-06-16 01:11:34,025][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3100_state/sampler_1.bin +[2026-06-16 01:11:34,026][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3100_state/random_states_0.pkl +[2026-06-16 01:11:34,071][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3100 +[2026-06-16 01:11:34,072][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 01:11:34,078][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 01:11:38,399][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.005647353362292051, 'train/grad_norm_pre_clip': 0.8520508408546448, 'timing/data': 0.006297771004028618, 'timing/model': 1.1118114499840885, 'learning_rate/qwen_vl_interface': 3.6066582970025266e-06, 'learning_rate/action_model': 3.6066582970025264e-05, 'epoch': 1.27}) +[2026-06-16 01:11:42,887][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.0053686234168708324, 'train/grad_norm_pre_clip': 1.0950156450271606, 'timing/data': 0.006459567986894399, 'timing/model': 1.1258107739849947, 'learning_rate/qwen_vl_interface': 3.603801278021964e-06, 'learning_rate/action_model': 3.6038012780219635e-05, 'epoch': 1.27}) +[2026-06-16 01:11:47,350][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.24312059581279755, 'train/grad_norm_pre_clip': 14.011847496032715, 'timing/data': 0.006613564968574792, 'timing/model': 1.115494119003415, 'learning_rate/qwen_vl_interface': 3.6009449357315122e-06, 'learning_rate/action_model': 3.6009449357315125e-05, 'epoch': 1.27}) +[2026-06-16 01:11:51,891][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.008901691064238548, 'train/grad_norm_pre_clip': 1.3701696395874023, 'timing/data': 0.006489981024060398, 'timing/model': 1.1216151739936322, 'learning_rate/qwen_vl_interface': 3.5980892713053084e-06, 'learning_rate/action_model': 3.598089271305308e-05, 'epoch': 1.28}) +[2026-06-16 01:11:56,370][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.0040844217874109745, 'train/grad_norm_pre_clip': 1.2699108123779297, 'timing/data': 0.006393413001205772, 'timing/model': 1.1167688900022767, 'learning_rate/qwen_vl_interface': 3.5952342859172063e-06, 'learning_rate/action_model': 3.595234285917206e-05, 'epoch': 1.28}) +[2026-06-16 01:12:00,862][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.01806376315653324, 'train/grad_norm_pre_clip': 4.261240005493164, 'timing/data': 0.006147405016236007, 'timing/model': 1.2712797980057076, 'learning_rate/qwen_vl_interface': 3.592379980740785e-06, 'learning_rate/action_model': 3.5923799807407845e-05, 'epoch': 1.28}) +[2026-06-16 01:12:05,325][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.02746085450053215, 'train/grad_norm_pre_clip': 7.474740505218506, 'timing/data': 0.006457887007854879, 'timing/model': 1.2453961950377561, 'learning_rate/qwen_vl_interface': 3.5895263569493384e-06, 'learning_rate/action_model': 3.589526356949338e-05, 'epoch': 1.28}) +[2026-06-16 01:12:09,652][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.030080646276474, 'train/grad_norm_pre_clip': 16.80998992919922, 'timing/data': 0.006070832023397088, 'timing/model': 1.123128586972598, 'learning_rate/qwen_vl_interface': 3.586673415715885e-06, 'learning_rate/action_model': 3.586673415715885e-05, 'epoch': 1.28}) +[2026-06-16 01:12:14,153][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.10098101943731308, 'train/grad_norm_pre_clip': 25.380449295043945, 'timing/data': 0.006077241967432201, 'timing/model': 1.1213287620339543, 'learning_rate/qwen_vl_interface': 3.5838211582131603e-06, 'learning_rate/action_model': 3.58382115821316e-05, 'epoch': 1.28}) +[2026-06-16 01:12:18,639][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.004109527915716171, 'train/grad_norm_pre_clip': 0.5679422616958618, 'timing/data': 0.006698028999380767, 'timing/model': 1.121165785996709, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 1.28}) +[2026-06-16 01:12:23,158][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.010533884167671204, 'train/grad_norm_pre_clip': 2.395601749420166, 'timing/data': 0.006409434019587934, 'timing/model': 1.1281123220105655, 'learning_rate/qwen_vl_interface': 3.5781186990894423e-06, 'learning_rate/action_model': 3.578118699089442e-05, 'epoch': 1.28}) +[2026-06-16 01:12:27,624][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.08790908008813858, 'train/grad_norm_pre_clip': 12.511445999145508, 'timing/data': 0.006179140007589012, 'timing/model': 1.1269015750149265, 'learning_rate/qwen_vl_interface': 3.5752684998125113e-06, 'learning_rate/action_model': 3.5752684998125116e-05, 'epoch': 1.28}) +[2026-06-16 01:12:32,101][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.22697249054908752, 'train/grad_norm_pre_clip': 15.769423484802246, 'timing/data': 0.006180069991387427, 'timing/model': 1.2532881229999475, 'learning_rate/qwen_vl_interface': 3.5724189889544407e-06, 'learning_rate/action_model': 3.572418988954441e-05, 'epoch': 1.28}) +[2026-06-16 01:12:36,449][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.000422529352363199, 'train/grad_norm_pre_clip': 0.11287801712751389, 'timing/data': 0.0061956619611009955, 'timing/model': 1.1264945319853723, 'learning_rate/qwen_vl_interface': 3.569570167686556e-06, 'learning_rate/action_model': 3.569570167686556e-05, 'epoch': 1.28}) +[2026-06-16 01:12:40,960][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.00043280338286422193, 'train/grad_norm_pre_clip': 0.13806970417499542, 'timing/data': 0.0065431149560026824, 'timing/model': 1.1309762329910882, 'learning_rate/qwen_vl_interface': 3.5667220371798995e-06, 'learning_rate/action_model': 3.566722037179899e-05, 'epoch': 1.28}) +[2026-06-16 01:12:45,466][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.0030229343101382256, 'train/grad_norm_pre_clip': 1.609508752822876, 'timing/data': 0.006279828026890755, 'timing/model': 1.1207986150402576, 'learning_rate/qwen_vl_interface': 3.563874598605232e-06, 'learning_rate/action_model': 3.5638745986052316e-05, 'epoch': 1.28}) +[2026-06-16 01:12:49,964][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.003313834546133876, 'train/grad_norm_pre_clip': 0.9235268235206604, 'timing/data': 0.006549088982865214, 'timing/model': 1.1241705419961363, 'learning_rate/qwen_vl_interface': 3.561027853133025e-06, 'learning_rate/action_model': 3.561027853133025e-05, 'epoch': 1.28}) +[2026-06-16 01:12:54,473][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.13252808153629303, 'train/grad_norm_pre_clip': 14.806035995483398, 'timing/data': 0.006601611035875976, 'timing/model': 1.2817874809843488, 'learning_rate/qwen_vl_interface': 3.558181801933472e-06, 'learning_rate/action_model': 3.5581818019334714e-05, 'epoch': 1.28}) +[2026-06-16 01:12:58,830][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.07011102139949799, 'train/grad_norm_pre_clip': 14.685473442077637, 'timing/data': 0.006296236999332905, 'timing/model': 1.1288098159711808, 'learning_rate/qwen_vl_interface': 3.555336446176473e-06, 'learning_rate/action_model': 3.5553364461764726e-05, 'epoch': 1.28}) +[2026-06-16 01:13:03,345][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.07684744894504547, 'train/grad_norm_pre_clip': 11.138726234436035, 'timing/data': 0.00660975300706923, 'timing/model': 1.1317363360431045, 'learning_rate/qwen_vl_interface': 3.552491787031647e-06, 'learning_rate/action_model': 3.552491787031647e-05, 'epoch': 1.28}) +[2026-06-16 01:13:07,776][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.05854257941246033, 'train/grad_norm_pre_clip': 14.343217849731445, 'timing/data': 0.006006354989949614, 'timing/model': 1.1282609039917588, 'learning_rate/qwen_vl_interface': 3.5496478256683277e-06, 'learning_rate/action_model': 3.549647825668327e-05, 'epoch': 1.28}) +[2026-06-16 01:13:12,245][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.15258871018886566, 'train/grad_norm_pre_clip': 16.523576736450195, 'timing/data': 0.006218505033757538, 'timing/model': 1.1334748679655604, 'learning_rate/qwen_vl_interface': 3.5468045632555585e-06, 'learning_rate/action_model': 3.546804563255559e-05, 'epoch': 1.28}) +[2026-06-16 01:13:16,720][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.004693000577390194, 'train/grad_norm_pre_clip': 1.216712236404419, 'timing/data': 0.0061617709579877555, 'timing/model': 1.1326386400032789, 'learning_rate/qwen_vl_interface': 3.5439620009620995e-06, 'learning_rate/action_model': 3.543962000962099e-05, 'epoch': 1.28}) +[2026-06-16 01:13:21,259][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.14390970766544342, 'train/grad_norm_pre_clip': 32.17374801635742, 'timing/data': 0.006343094981275499, 'timing/model': 1.303981303004548, 'learning_rate/qwen_vl_interface': 3.541120139956419e-06, 'learning_rate/action_model': 3.5411201399564184e-05, 'epoch': 1.28}) +[2026-06-16 01:13:25,627][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.25343751907348633, 'train/grad_norm_pre_clip': 27.498502731323242, 'timing/data': 0.006308770971372724, 'timing/model': 1.135118134028744, 'learning_rate/qwen_vl_interface': 3.538278981406697e-06, 'learning_rate/action_model': 3.5382789814066966e-05, 'epoch': 1.28}) +[2026-06-16 01:13:30,138][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.07841920107603073, 'train/grad_norm_pre_clip': 19.019865036010742, 'timing/data': 0.006532801024150103, 'timing/model': 1.1292439989629202, 'learning_rate/qwen_vl_interface': 3.5354385264808295e-06, 'learning_rate/action_model': 3.535438526480829e-05, 'epoch': 1.28}) +[2026-06-16 01:13:34,653][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.1359575092792511, 'train/grad_norm_pre_clip': 10.45336627960205, 'timing/data': 0.005899585026782006, 'timing/model': 1.1317060299916193, 'learning_rate/qwen_vl_interface': 3.5325987763464196e-06, 'learning_rate/action_model': 3.5325987763464194e-05, 'epoch': 1.28}) +[2026-06-16 01:13:39,135][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.2357347011566162, 'train/grad_norm_pre_clip': 31.209760665893555, 'timing/data': 0.006198453018441796, 'timing/model': 1.131825749995187, 'learning_rate/qwen_vl_interface': 3.529759732170781e-06, 'learning_rate/action_model': 3.529759732170781e-05, 'epoch': 1.28}) +[2026-06-16 01:13:43,637][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.20306192338466644, 'train/grad_norm_pre_clip': 9.29535961151123, 'timing/data': 0.006250564008951187, 'timing/model': 1.1296499340096489, 'learning_rate/qwen_vl_interface': 3.5269213951209375e-06, 'learning_rate/action_model': 3.5269213951209374e-05, 'epoch': 1.29}) +[2026-06-16 01:13:48,162][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.29050886631011963, 'train/grad_norm_pre_clip': 16.915462493896484, 'timing/data': 0.006469384010415524, 'timing/model': 1.2918898400384933, 'learning_rate/qwen_vl_interface': 3.5240837663636207e-06, 'learning_rate/action_model': 3.5240837663636204e-05, 'epoch': 1.29}) +[2026-06-16 01:13:52,535][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.009189738892018795, 'train/grad_norm_pre_clip': 3.6907598972320557, 'timing/data': 0.0064662290387786925, 'timing/model': 1.130287275009323, 'learning_rate/qwen_vl_interface': 3.5212468470652745e-06, 'learning_rate/action_model': 3.521246847065274e-05, 'epoch': 1.29}) +[2026-06-16 01:13:57,070][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.07987120747566223, 'train/grad_norm_pre_clip': 11.352663040161133, 'timing/data': 0.006616075988858938, 'timing/model': 1.1321049769758247, 'learning_rate/qwen_vl_interface': 3.5184106383920487e-06, 'learning_rate/action_model': 3.5184106383920486e-05, 'epoch': 1.29}) +[2026-06-16 01:14:01,592][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.04013187065720558, 'train/grad_norm_pre_clip': 15.230988502502441, 'timing/data': 0.006386257999110967, 'timing/model': 1.1302091929828748, 'learning_rate/qwen_vl_interface': 3.5155751415098022e-06, 'learning_rate/action_model': 3.515575141509802e-05, 'epoch': 1.29}) +[2026-06-16 01:14:06,085][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.03471292182803154, 'train/grad_norm_pre_clip': 28.30424690246582, 'timing/data': 0.006355502991937101, 'timing/model': 1.1339393070084043, 'learning_rate/qwen_vl_interface': 3.5127403575841e-06, 'learning_rate/action_model': 3.5127403575841e-05, 'epoch': 1.29}) +[2026-06-16 01:14:10,613][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.06027350202202797, 'train/grad_norm_pre_clip': 7.367184638977051, 'timing/data': 0.006346231035422534, 'timing/model': 1.1299786680028774, 'learning_rate/qwen_vl_interface': 3.509906287780215e-06, 'learning_rate/action_model': 3.5099062877802144e-05, 'epoch': 1.29}) +[2026-06-16 01:14:15,135][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.04039473459124565, 'train/grad_norm_pre_clip': 20.039169311523438, 'timing/data': 0.006068119022529572, 'timing/model': 1.2830102650332265, 'learning_rate/qwen_vl_interface': 3.5070729332631247e-06, 'learning_rate/action_model': 3.507072933263125e-05, 'epoch': 1.29}) +[2026-06-16 01:14:19,510][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.05097239464521408, 'train/grad_norm_pre_clip': 15.710630416870117, 'timing/data': 0.006059592007659376, 'timing/model': 1.131537067994941, 'learning_rate/qwen_vl_interface': 3.5042402951975167e-06, 'learning_rate/action_model': 3.5042402951975167e-05, 'epoch': 1.29}) +[2026-06-16 01:14:24,053][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.1940298229455948, 'train/grad_norm_pre_clip': 24.546131134033203, 'timing/data': 0.006586383038666099, 'timing/model': 1.1325286270002834, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 1.29}) +[2026-06-16 01:14:28,516][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.05017644166946411, 'train/grad_norm_pre_clip': 9.088765144348145, 'timing/data': 0.006371940020471811, 'timing/model': 1.131073620985262, 'learning_rate/qwen_vl_interface': 3.498577173078011e-06, 'learning_rate/action_model': 3.498577173078011e-05, 'epoch': 1.29}) +[2026-06-16 01:14:33,014][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.01160678081214428, 'train/grad_norm_pre_clip': 4.166231632232666, 'timing/data': 0.005925244011450559, 'timing/model': 1.1325997490203008, 'learning_rate/qwen_vl_interface': 3.495746691352008e-06, 'learning_rate/action_model': 3.495746691352008e-05, 'epoch': 1.29}) +[2026-06-16 01:14:37,539][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.4219996929168701, 'train/grad_norm_pre_clip': 24.34172248840332, 'timing/data': 0.00639949698233977, 'timing/model': 1.1284321690327488, 'learning_rate/qwen_vl_interface': 3.4929169307332763e-06, 'learning_rate/action_model': 3.492916930733276e-05, 'epoch': 1.29}) +[2026-06-16 01:14:42,058][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.030636368319392204, 'train/grad_norm_pre_clip': 12.34766674041748, 'timing/data': 0.00634696998167783, 'timing/model': 1.2796614050166681, 'learning_rate/qwen_vl_interface': 3.4900878923850234e-06, 'learning_rate/action_model': 3.490087892385023e-05, 'epoch': 1.29}) +[2026-06-16 01:14:46,495][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.07065446674823761, 'train/grad_norm_pre_clip': 13.44826602935791, 'timing/data': 0.015260339016094804, 'timing/model': 1.1330354629899375, 'learning_rate/qwen_vl_interface': 3.4872595774701607e-06, 'learning_rate/action_model': 3.487259577470161e-05, 'epoch': 1.29}) +[2026-06-16 01:14:51,038][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.2172103226184845, 'train/grad_norm_pre_clip': 21.674877166748047, 'timing/data': 0.006083667045459151, 'timing/model': 1.1382329459884204, 'learning_rate/qwen_vl_interface': 3.4844319871513014e-06, 'learning_rate/action_model': 3.484431987151301e-05, 'epoch': 1.29}) +[2026-06-16 01:14:55,534][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.04318665340542793, 'train/grad_norm_pre_clip': 17.659961700439453, 'timing/data': 0.006482806988060474, 'timing/model': 1.1333311059861444, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 1.29}) +[2026-06-16 01:15:00,073][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.08228714764118195, 'train/grad_norm_pre_clip': 9.124467849731445, 'timing/data': 0.00595848704688251, 'timing/model': 1.139554388995748, 'learning_rate/qwen_vl_interface': 3.478778984950556e-06, 'learning_rate/action_model': 3.478778984950556e-05, 'epoch': 1.29}) +[2026-06-16 01:15:04,628][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.007048928644508123, 'train/grad_norm_pre_clip': 0.8790135979652405, 'timing/data': 0.006331376964226365, 'timing/model': 1.1377605389570817, 'learning_rate/qwen_vl_interface': 3.475953575392405e-06, 'learning_rate/action_model': 3.475953575392405e-05, 'epoch': 1.29}) +[2026-06-16 01:15:09,184][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.0820990577340126, 'train/grad_norm_pre_clip': 6.4543657302856445, 'timing/data': 0.006307352043222636, 'timing/model': 1.3038693880080245, 'learning_rate/qwen_vl_interface': 3.473128895077728e-06, 'learning_rate/action_model': 3.473128895077728e-05, 'epoch': 1.29}) +[2026-06-16 01:15:13,595][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.008755900897085667, 'train/grad_norm_pre_clip': 4.484072208404541, 'timing/data': 0.005709529970772564, 'timing/model': 1.145767103007529, 'learning_rate/qwen_vl_interface': 3.4703049451676453e-06, 'learning_rate/action_model': 3.4703049451676454e-05, 'epoch': 1.29}) +[2026-06-16 01:15:18,145][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.3064550459384918, 'train/grad_norm_pre_clip': 8.49608039855957, 'timing/data': 0.0067436639801599085, 'timing/model': 1.1318976180045865, 'learning_rate/qwen_vl_interface': 3.4674817268229732e-06, 'learning_rate/action_model': 3.467481726822973e-05, 'epoch': 1.29}) +[2026-06-16 01:15:22,637][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.05627802386879921, 'train/grad_norm_pre_clip': 11.140491485595703, 'timing/data': 0.006318494968581945, 'timing/model': 1.1375685140374117, 'learning_rate/qwen_vl_interface': 3.464659241204232e-06, 'learning_rate/action_model': 3.4646592412042314e-05, 'epoch': 1.29}) +[2026-06-16 01:15:27,208][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.0045753768645226955, 'train/grad_norm_pre_clip': 1.0414550304412842, 'timing/data': 0.006238093017600477, 'timing/model': 1.1453832660336047, 'learning_rate/qwen_vl_interface': 3.4618374894716377e-06, 'learning_rate/action_model': 3.461837489471638e-05, 'epoch': 1.29}) +[2026-06-16 01:15:31,749][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.047400813549757004, 'train/grad_norm_pre_clip': 10.104766845703125, 'timing/data': 0.008114223019219935, 'timing/model': 1.1300103630055673, 'learning_rate/qwen_vl_interface': 3.459016472785107e-06, 'learning_rate/action_model': 3.459016472785107e-05, 'epoch': 1.3}) +[2026-06-16 01:15:36,291][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.05191730335354805, 'train/grad_norm_pre_clip': 12.028406143188477, 'timing/data': 0.006537468987517059, 'timing/model': 1.294847542012576, 'learning_rate/qwen_vl_interface': 3.456196192304254e-06, 'learning_rate/action_model': 3.456196192304254e-05, 'epoch': 1.3}) +[2026-06-16 01:15:40,755][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.061599332839250565, 'train/grad_norm_pre_clip': 7.152088165283203, 'timing/data': 0.005977427994366735, 'timing/model': 1.2053141600335948, 'learning_rate/qwen_vl_interface': 3.4533766491883887e-06, 'learning_rate/action_model': 3.453376649188388e-05, 'epoch': 1.3}) +[2026-06-16 01:15:45,142][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.021804269403219223, 'train/grad_norm_pre_clip': 4.004792213439941, 'timing/data': 0.0068438369780778885, 'timing/model': 1.1332957330159843, 'learning_rate/qwen_vl_interface': 3.450557844596517e-06, 'learning_rate/action_model': 3.450557844596517e-05, 'epoch': 1.3}) +[2026-06-16 01:15:49,671][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.01542246900498867, 'train/grad_norm_pre_clip': 3.359527826309204, 'timing/data': 0.006456497998442501, 'timing/model': 1.1389313949621283, 'learning_rate/qwen_vl_interface': 3.447739779687345e-06, 'learning_rate/action_model': 3.4477397796873446e-05, 'epoch': 1.3}) +[2026-06-16 01:15:54,198][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.1195388063788414, 'train/grad_norm_pre_clip': 14.197002410888672, 'timing/data': 0.0061550450045615435, 'timing/model': 1.1303714469540864, 'learning_rate/qwen_vl_interface': 3.4449224556192722e-06, 'learning_rate/action_model': 3.444922455619272e-05, 'epoch': 1.3}) +[2026-06-16 01:15:58,760][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.05575649067759514, 'train/grad_norm_pre_clip': 12.626310348510742, 'timing/data': 0.006509297003503889, 'timing/model': 1.144049473979976, 'learning_rate/qwen_vl_interface': 3.442105873550395e-06, 'learning_rate/action_model': 3.442105873550395e-05, 'epoch': 1.3}) +[2026-06-16 01:16:03,311][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.022773290053009987, 'train/grad_norm_pre_clip': 3.5517356395721436, 'timing/data': 0.006443869031500071, 'timing/model': 1.3036455530091189, 'learning_rate/qwen_vl_interface': 3.4392900346385022e-06, 'learning_rate/action_model': 3.4392900346385024e-05, 'epoch': 1.3}) +[2026-06-16 01:16:07,809][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.062234699726104736, 'train/grad_norm_pre_clip': 8.320521354675293, 'timing/data': 0.00609888497274369, 'timing/model': 1.2523681699531153, 'learning_rate/qwen_vl_interface': 3.4364749400410796e-06, 'learning_rate/action_model': 3.43647494004108e-05, 'epoch': 1.3}) +[2026-06-16 01:16:12,206][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.00685508269816637, 'train/grad_norm_pre_clip': 1.6477402448654175, 'timing/data': 0.006456527975387871, 'timing/model': 1.135508923965972, 'learning_rate/qwen_vl_interface': 3.4336605909153066e-06, 'learning_rate/action_model': 3.433660590915306e-05, 'epoch': 1.3}) +[2026-06-16 01:16:16,744][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.07798176258802414, 'train/grad_norm_pre_clip': 6.256887912750244, 'timing/data': 0.0068406600039452314, 'timing/model': 1.137579783971887, 'learning_rate/qwen_vl_interface': 3.430846988418057e-06, 'learning_rate/action_model': 3.430846988418057e-05, 'epoch': 1.3}) +[2026-06-16 01:16:21,291][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.022996259853243828, 'train/grad_norm_pre_clip': 4.979255676269531, 'timing/data': 0.007092888990882784, 'timing/model': 1.1350483340211213, 'learning_rate/qwen_vl_interface': 3.428034133705893e-06, 'learning_rate/action_model': 3.428034133705893e-05, 'epoch': 1.3}) +[2026-06-16 01:16:25,848][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.0762433111667633, 'train/grad_norm_pre_clip': 13.049210548400879, 'timing/data': 0.00694436498451978, 'timing/model': 1.1463946249568835, 'learning_rate/qwen_vl_interface': 3.4252220279350763e-06, 'learning_rate/action_model': 3.425222027935076e-05, 'epoch': 1.3}) +[2026-06-16 01:16:30,391][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.09637214243412018, 'train/grad_norm_pre_clip': 8.460844993591309, 'timing/data': 0.006449819018598646, 'timing/model': 1.287526672007516, 'learning_rate/qwen_vl_interface': 3.4224106722615547e-06, 'learning_rate/action_model': 3.422410672261554e-05, 'epoch': 1.3}) +[2026-06-16 01:16:34,771][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.12067626416683197, 'train/grad_norm_pre_clip': 26.991241455078125, 'timing/data': 0.006255068001337349, 'timing/model': 1.1329214950092137, 'learning_rate/qwen_vl_interface': 3.419600067840973e-06, 'learning_rate/action_model': 3.419600067840973e-05, 'epoch': 1.3}) +[2026-06-16 01:16:39,311][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.005242315586656332, 'train/grad_norm_pre_clip': 0.998237669467926, 'timing/data': 0.006274389976169914, 'timing/model': 1.1304482640116476, 'learning_rate/qwen_vl_interface': 3.4167902158286627e-06, 'learning_rate/action_model': 3.4167902158286626e-05, 'epoch': 1.3}) +[2026-06-16 01:16:43,858][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.0789697915315628, 'train/grad_norm_pre_clip': 9.864819526672363, 'timing/data': 0.006128906039521098, 'timing/model': 1.137812961009331, 'learning_rate/qwen_vl_interface': 3.4139811173796468e-06, 'learning_rate/action_model': 3.413981117379647e-05, 'epoch': 1.3}) +[2026-06-16 01:16:48,413][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.07073187083005905, 'train/grad_norm_pre_clip': 7.956919193267822, 'timing/data': 0.006456230999901891, 'timing/model': 1.1378072829684243, 'learning_rate/qwen_vl_interface': 3.411172773648642e-06, 'learning_rate/action_model': 3.411172773648642e-05, 'epoch': 1.3}) +[2026-06-16 01:16:52,907][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.09136642515659332, 'train/grad_norm_pre_clip': 14.708943367004395, 'timing/data': 0.00617342401528731, 'timing/model': 1.142340172023978, 'learning_rate/qwen_vl_interface': 3.408365185790052e-06, 'learning_rate/action_model': 3.4083651857900515e-05, 'epoch': 1.3}) +[2026-06-16 01:16:57,428][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.015841957181692123, 'train/grad_norm_pre_clip': 4.44649076461792, 'timing/data': 0.00643692200537771, 'timing/model': 1.2519984520040452, 'learning_rate/qwen_vl_interface': 3.40555835495797e-06, 'learning_rate/action_model': 3.40555835495797e-05, 'epoch': 1.3}) +[2026-06-16 01:17:01,810][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.004536884371191263, 'train/grad_norm_pre_clip': 1.0595824718475342, 'timing/data': 0.006402819009963423, 'timing/model': 1.1309616640210152, 'learning_rate/qwen_vl_interface': 3.4027522823061763e-06, 'learning_rate/action_model': 3.4027522823061765e-05, 'epoch': 1.3}) +[2026-06-16 01:17:06,369][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.13866955041885376, 'train/grad_norm_pre_clip': 14.159655570983887, 'timing/data': 0.006091776012908667, 'timing/model': 1.1409307569847442, 'learning_rate/qwen_vl_interface': 3.399946968988143e-06, 'learning_rate/action_model': 3.399946968988143e-05, 'epoch': 1.3}) +[2026-06-16 01:17:10,928][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.02051437273621559, 'train/grad_norm_pre_clip': 2.964398145675659, 'timing/data': 0.006049887975677848, 'timing/model': 1.1426601380226202, 'learning_rate/qwen_vl_interface': 3.397142416157028e-06, 'learning_rate/action_model': 3.397142416157028e-05, 'epoch': 1.3}) +[2026-06-16 01:17:15,501][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.0014400999061763287, 'train/grad_norm_pre_clip': 0.21409069001674652, 'timing/data': 0.006232094950973988, 'timing/model': 1.132522764033638, 'learning_rate/qwen_vl_interface': 3.3943386249656766e-06, 'learning_rate/action_model': 3.394338624965677e-05, 'epoch': 1.3}) +[2026-06-16 01:17:20,005][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.013093098066747189, 'train/grad_norm_pre_clip': 6.46713399887085, 'timing/data': 0.00613245100248605, 'timing/model': 1.1352780089946464, 'learning_rate/qwen_vl_interface': 3.391535596566624e-06, 'learning_rate/action_model': 3.391535596566624e-05, 'epoch': 1.31}) +[2026-06-16 01:17:24,510][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.050323162227869034, 'train/grad_norm_pre_clip': 13.114030838012695, 'timing/data': 0.006185550999362022, 'timing/model': 1.260525186022278, 'learning_rate/qwen_vl_interface': 3.388733332112086e-06, 'learning_rate/action_model': 3.3887333321120856e-05, 'epoch': 1.31}) +[2026-06-16 01:17:28,883][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.022330446168780327, 'train/grad_norm_pre_clip': 5.525290012359619, 'timing/data': 0.006164402002468705, 'timing/model': 1.130531988979783, 'learning_rate/qwen_vl_interface': 3.3859318327539693e-06, 'learning_rate/action_model': 3.385931832753969e-05, 'epoch': 1.31}) +[2026-06-16 01:17:33,421][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.0060631330125033855, 'train/grad_norm_pre_clip': 0.7852804064750671, 'timing/data': 0.00583329499932006, 'timing/model': 1.1380034189787693, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 1.31}) +[2026-06-16 01:17:37,966][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.06674189865589142, 'train/grad_norm_pre_clip': 16.667198181152344, 'timing/data': 0.006667046982329339, 'timing/model': 1.1447419610340148, 'learning_rate/qwen_vl_interface': 3.3803311339330447e-06, 'learning_rate/action_model': 3.3803311339330444e-05, 'epoch': 1.31}) +[2026-06-16 01:17:42,507][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.004861017223447561, 'train/grad_norm_pre_clip': 1.2011593580245972, 'timing/data': 0.006463533034548163, 'timing/model': 1.129773507011123, 'learning_rate/qwen_vl_interface': 3.3775319367724746e-06, 'learning_rate/action_model': 3.3775319367724745e-05, 'epoch': 1.31}) +[2026-06-16 01:17:47,059][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.06788387149572372, 'train/grad_norm_pre_clip': 10.777727127075195, 'timing/data': 0.006152579968329519, 'timing/model': 1.3032014000345953, 'learning_rate/qwen_vl_interface': 3.374733509312795e-06, 'learning_rate/action_model': 3.374733509312795e-05, 'epoch': 1.31}) +[2026-06-16 01:17:51,436][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.0031639989465475082, 'train/grad_norm_pre_clip': 0.5006783604621887, 'timing/data': 0.005944692995399237, 'timing/model': 1.135011049977038, 'learning_rate/qwen_vl_interface': 3.3719358527043345e-06, 'learning_rate/action_model': 3.371935852704334e-05, 'epoch': 1.31}) +[2026-06-16 01:17:55,982][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.15998035669326782, 'train/grad_norm_pre_clip': 18.770950317382812, 'timing/data': 0.006467439990956336, 'timing/model': 1.137196745024994, 'learning_rate/qwen_vl_interface': 3.3691389680971047e-06, 'learning_rate/action_model': 3.3691389680971044e-05, 'epoch': 1.31}) +[2026-06-16 01:18:00,451][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.12325836718082428, 'train/grad_norm_pre_clip': 10.019124031066895, 'timing/data': 0.006575939012691379, 'timing/model': 1.1349774459959008, 'learning_rate/qwen_vl_interface': 3.3663428566407985e-06, 'learning_rate/action_model': 3.366342856640798e-05, 'epoch': 1.31}) +[2026-06-16 01:18:04,933][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.19195958971977234, 'train/grad_norm_pre_clip': 6.995878219604492, 'timing/data': 0.006481121003162116, 'timing/model': 1.1403031810186803, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 1.31}) +[2026-06-16 01:18:09,454][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.005626662168651819, 'train/grad_norm_pre_clip': 1.8996156454086304, 'timing/data': 0.00614099099766463, 'timing/model': 1.133254542015493, 'learning_rate/qwen_vl_interface': 3.3607529577781433e-06, 'learning_rate/action_model': 3.360752957778143e-05, 'epoch': 1.31}) +[2026-06-16 01:18:14,001][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.02504623681306839, 'train/grad_norm_pre_clip': 4.615784645080566, 'timing/data': 0.006328594987280667, 'timing/model': 1.3043567119748332, 'learning_rate/qwen_vl_interface': 3.3579591726695906e-06, 'learning_rate/action_model': 3.35795917266959e-05, 'epoch': 1.31}) +[2026-06-16 01:18:18,381][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.007128587458282709, 'train/grad_norm_pre_clip': 1.7568535804748535, 'timing/data': 0.0059296959661878645, 'timing/model': 1.1343973450129852, 'learning_rate/qwen_vl_interface': 3.3551661653075522e-06, 'learning_rate/action_model': 3.355166165307552e-05, 'epoch': 1.31}) +[2026-06-16 01:18:22,919][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.19836291670799255, 'train/grad_norm_pre_clip': 10.618049621582031, 'timing/data': 0.006566086027305573, 'timing/model': 1.131167521001771, 'learning_rate/qwen_vl_interface': 3.35237393684013e-06, 'learning_rate/action_model': 3.35237393684013e-05, 'epoch': 1.31}) +[2026-06-16 01:18:27,466][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.13148266077041626, 'train/grad_norm_pre_clip': 17.17752456665039, 'timing/data': 0.006171991000883281, 'timing/model': 1.1311774390051141, 'learning_rate/qwen_vl_interface': 3.349582488415103e-06, 'learning_rate/action_model': 3.3495824884151026e-05, 'epoch': 1.31}) +[2026-06-16 01:18:31,954][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.22954678535461426, 'train/grad_norm_pre_clip': 25.633533477783203, 'timing/data': 0.006864160008262843, 'timing/model': 1.1291713980026543, 'learning_rate/qwen_vl_interface': 3.3467918211799294e-06, 'learning_rate/action_model': 3.3467918211799295e-05, 'epoch': 1.31}) +[2026-06-16 01:18:36,456][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.0713035985827446, 'train/grad_norm_pre_clip': 16.913719177246094, 'timing/data': 0.006481631950009614, 'timing/model': 1.1307559360284358, 'learning_rate/qwen_vl_interface': 3.3440019362817476e-06, 'learning_rate/action_model': 3.344001936281747e-05, 'epoch': 1.31}) +[2026-06-16 01:18:41,003][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.007630030624568462, 'train/grad_norm_pre_clip': 3.062919855117798, 'timing/data': 0.006234515982214361, 'timing/model': 1.2972293479833752, 'learning_rate/qwen_vl_interface': 3.3412128348673745e-06, 'learning_rate/action_model': 3.341212834867374e-05, 'epoch': 1.31}) +[2026-06-16 01:18:45,404][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.018050797283649445, 'train/grad_norm_pre_clip': 3.6497228145599365, 'timing/data': 0.006599613989237696, 'timing/model': 1.1382969659753144, 'learning_rate/qwen_vl_interface': 3.338424518083303e-06, 'learning_rate/action_model': 3.338424518083303e-05, 'epoch': 1.31}) +[2026-06-16 01:18:49,948][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.04409168288111687, 'train/grad_norm_pre_clip': 14.28087043762207, 'timing/data': 0.0064352170447818935, 'timing/model': 1.1393264060025103, 'learning_rate/qwen_vl_interface': 3.3356369870757056e-06, 'learning_rate/action_model': 3.335636987075705e-05, 'epoch': 1.31}) +[2026-06-16 01:18:54,482][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.02216414175927639, 'train/grad_norm_pre_clip': 10.456269264221191, 'timing/data': 0.006563946953974664, 'timing/model': 1.1337583789718337, 'learning_rate/qwen_vl_interface': 3.3328502429904327e-06, 'learning_rate/action_model': 3.3328502429904325e-05, 'epoch': 1.31}) +[2026-06-16 01:18:58,981][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.008986036293208599, 'train/grad_norm_pre_clip': 1.620734691619873, 'timing/data': 0.006799089023843408, 'timing/model': 1.1435076820198447, 'learning_rate/qwen_vl_interface': 3.3300642869730066e-06, 'learning_rate/action_model': 3.330064286973006e-05, 'epoch': 1.31}) +[2026-06-16 01:23:19,495][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.02373448759317398, 'train/grad_norm_pre_clip': 9.471931457519531, 'eval/loss': 0.04449179396033287, 'eval/flap_precision': 0.9623493975903614, 'eval/flap_recall': 0.9861111111111112, 'eval/flap_f1': 0.9740853658536586, 'eval/flap_rate_model': 0.18444444444444444, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9740853658536586, 'timing/data': 0.00633646099595353, 'timing/model': 1.1330331380013376, 'learning_rate/qwen_vl_interface': 3.3272791201686294e-06, 'learning_rate/action_model': 3.327279120168629e-05, 'epoch': 1.31}) +[2026-06-16 01:23:19,496][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3200_state +[2026-06-16 01:23:19,498][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 01:23:23,151][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3200_state/model.safetensors +[2026-06-16 01:23:33,140][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3200_state/optimizer.bin +[2026-06-16 01:23:33,141][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3200_state/sampler.bin +[2026-06-16 01:23:33,141][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3200_state/sampler_1.bin +[2026-06-16 01:23:33,141][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3200_state/random_states_0.pkl +[2026-06-16 01:23:33,184][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3200 +[2026-06-16 01:23:33,184][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 01:23:33,190][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 01:23:37,627][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.0023031330201774836, 'train/grad_norm_pre_clip': 0.9498157501220703, 'timing/data': 0.006320732994936407, 'timing/model': 1.1092050800216384, 'learning_rate/qwen_vl_interface': 3.324494743722179e-06, 'learning_rate/action_model': 3.3244947437221785e-05, 'epoch': 1.31}) +[2026-06-16 01:23:42,024][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.056514933705329895, 'train/grad_norm_pre_clip': 9.655694961547852, 'timing/data': 0.006253340048715472, 'timing/model': 1.1129439629730769, 'learning_rate/qwen_vl_interface': 3.321711158778205e-06, 'learning_rate/action_model': 3.321711158778205e-05, 'epoch': 1.32}) +[2026-06-16 01:23:46,428][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.15105833113193512, 'train/grad_norm_pre_clip': 10.687500953674316, 'timing/data': 0.006205979036167264, 'timing/model': 1.2056928419624455, 'learning_rate/qwen_vl_interface': 3.3189283664809393e-06, 'learning_rate/action_model': 3.318928366480939e-05, 'epoch': 1.32}) +[2026-06-16 01:23:50,729][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.002543045673519373, 'train/grad_norm_pre_clip': 0.38701608777046204, 'timing/data': 0.006577308988198638, 'timing/model': 1.115164156013634, 'learning_rate/qwen_vl_interface': 3.316146367974276e-06, 'learning_rate/action_model': 3.316146367974276e-05, 'epoch': 1.32}) +[2026-06-16 01:23:55,146][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.12768560647964478, 'train/grad_norm_pre_clip': 15.021110534667969, 'timing/data': 0.006116315955296159, 'timing/model': 1.1190643750014715, 'learning_rate/qwen_vl_interface': 3.3133651644017945e-06, 'learning_rate/action_model': 3.3133651644017944e-05, 'epoch': 1.32}) +[2026-06-16 01:23:59,609][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.10670169442892075, 'train/grad_norm_pre_clip': 21.02410888671875, 'timing/data': 0.006228951970115304, 'timing/model': 1.1202798539889045, 'learning_rate/qwen_vl_interface': 3.31058475690674e-06, 'learning_rate/action_model': 3.31058475690674e-05, 'epoch': 1.32}) +[2026-06-16 01:24:04,129][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.003910391591489315, 'train/grad_norm_pre_clip': 1.2387123107910156, 'timing/data': 0.00682154402602464, 'timing/model': 1.119829190953169, 'learning_rate/qwen_vl_interface': 3.3078051466320337e-06, 'learning_rate/action_model': 3.307805146632034e-05, 'epoch': 1.32}) +[2026-06-16 01:24:08,622][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.008728597313165665, 'train/grad_norm_pre_clip': 2.1246354579925537, 'timing/data': 0.006570667959749699, 'timing/model': 1.2799114070367068, 'learning_rate/qwen_vl_interface': 3.3050263347202712e-06, 'learning_rate/action_model': 3.305026334720271e-05, 'epoch': 1.32}) +[2026-06-16 01:24:12,958][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.0010188245214521885, 'train/grad_norm_pre_clip': 0.2250390499830246, 'timing/data': 0.006576990999747068, 'timing/model': 1.1256877569830976, 'learning_rate/qwen_vl_interface': 3.302248322313712e-06, 'learning_rate/action_model': 3.302248322313712e-05, 'epoch': 1.32}) +[2026-06-16 01:24:17,439][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.0010870250407606363, 'train/grad_norm_pre_clip': 0.5990694761276245, 'timing/data': 0.007050230051390827, 'timing/model': 1.121787769952789, 'learning_rate/qwen_vl_interface': 3.299471110554295e-06, 'learning_rate/action_model': 3.299471110554295e-05, 'epoch': 1.32}) +[2026-06-16 01:24:21,894][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.2591113746166229, 'train/grad_norm_pre_clip': 13.40183162689209, 'timing/data': 0.006375675031449646, 'timing/model': 1.1217224710271694, 'learning_rate/qwen_vl_interface': 3.296694700583626e-06, 'learning_rate/action_model': 3.296694700583626e-05, 'epoch': 1.32}) +[2026-06-16 01:24:26,404][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.08397440612316132, 'train/grad_norm_pre_clip': 10.833600044250488, 'timing/data': 0.00625330803450197, 'timing/model': 1.1389213650254533, 'learning_rate/qwen_vl_interface': 3.2939190935429833e-06, 'learning_rate/action_model': 3.293919093542983e-05, 'epoch': 1.32}) +[2026-06-16 01:24:30,924][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.0006895033293403685, 'train/grad_norm_pre_clip': 0.31425952911376953, 'timing/data': 0.006047642964404076, 'timing/model': 1.1294395739678293, 'learning_rate/qwen_vl_interface': 3.2911442905733126e-06, 'learning_rate/action_model': 3.291144290573312e-05, 'epoch': 1.32}) +[2026-06-16 01:24:35,447][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.0006378836114890873, 'train/grad_norm_pre_clip': 0.23051717877388, 'timing/data': 0.006454057991504669, 'timing/model': 1.2938489009975456, 'learning_rate/qwen_vl_interface': 3.2883702928152323e-06, 'learning_rate/action_model': 3.288370292815232e-05, 'epoch': 1.32}) +[2026-06-16 01:24:39,801][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.061655186116695404, 'train/grad_norm_pre_clip': 20.090213775634766, 'timing/data': 0.006449086009524763, 'timing/model': 1.1280020730337128, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 1.32}) +[2026-06-16 01:24:44,336][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.08259350061416626, 'train/grad_norm_pre_clip': 29.170093536376953, 'timing/data': 0.006466175022069365, 'timing/model': 1.1404852450359613, 'learning_rate/qwen_vl_interface': 3.2828247174946528e-06, 'learning_rate/action_model': 3.2828247174946524e-05, 'epoch': 1.32}) +[2026-06-16 01:24:48,821][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.39589723944664, 'train/grad_norm_pre_clip': 24.9162540435791, 'timing/data': 0.006871998019050807, 'timing/model': 1.1306916250032373, 'learning_rate/qwen_vl_interface': 3.280053142211729e-06, 'learning_rate/action_model': 3.280053142211729e-05, 'epoch': 1.32}) +[2026-06-16 01:24:53,346][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.1493159979581833, 'train/grad_norm_pre_clip': 36.988525390625, 'timing/data': 0.006561335001606494, 'timing/model': 1.1329158719745465, 'learning_rate/qwen_vl_interface': 3.2772823766995475e-06, 'learning_rate/action_model': 3.277282376699548e-05, 'epoch': 1.32}) +[2026-06-16 01:24:57,887][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.0033947874326258898, 'train/grad_norm_pre_clip': 0.6313076019287109, 'timing/data': 0.006436661991756409, 'timing/model': 1.1350526409805752, 'learning_rate/qwen_vl_interface': 3.2745124220970646e-06, 'learning_rate/action_model': 3.274512422097065e-05, 'epoch': 1.32}) +[2026-06-16 01:25:02,404][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.08422756940126419, 'train/grad_norm_pre_clip': 14.336220741271973, 'timing/data': 0.006522805022541434, 'timing/model': 1.2922821319662035, 'learning_rate/qwen_vl_interface': 3.2717432795429057e-06, 'learning_rate/action_model': 3.271743279542905e-05, 'epoch': 1.32}) +[2026-06-16 01:25:06,861][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.3410046696662903, 'train/grad_norm_pre_clip': 40.02487564086914, 'timing/data': 0.006729633023496717, 'timing/model': 1.2227529310039245, 'learning_rate/qwen_vl_interface': 3.2689749501753578e-06, 'learning_rate/action_model': 3.2689749501753574e-05, 'epoch': 1.32}) +[2026-06-16 01:25:11,224][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.0009298024233430624, 'train/grad_norm_pre_clip': 0.16751444339752197, 'timing/data': 0.006553039012942463, 'timing/model': 1.1268399989930913, 'learning_rate/qwen_vl_interface': 3.2662074351323774e-06, 'learning_rate/action_model': 3.266207435132378e-05, 'epoch': 1.32}) +[2026-06-16 01:25:15,711][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.0647163912653923, 'train/grad_norm_pre_clip': 28.1896915435791, 'timing/data': 0.006191377993673086, 'timing/model': 1.129266062984243, 'learning_rate/qwen_vl_interface': 3.2634407355515875e-06, 'learning_rate/action_model': 3.2634407355515875e-05, 'epoch': 1.32}) +[2026-06-16 01:25:20,250][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.04590405151247978, 'train/grad_norm_pre_clip': 24.180133819580078, 'timing/data': 0.006632426055148244, 'timing/model': 1.1286329140421003, 'learning_rate/qwen_vl_interface': 3.2606748525702715e-06, 'learning_rate/action_model': 3.260674852570271e-05, 'epoch': 1.32}) +[2026-06-16 01:25:24,785][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.0005856851348653436, 'train/grad_norm_pre_clip': 0.09365779906511307, 'timing/data': 0.00622609700076282, 'timing/model': 1.1363355810171925, 'learning_rate/qwen_vl_interface': 3.2579097873253803e-06, 'learning_rate/action_model': 3.25790978732538e-05, 'epoch': 1.32}) +[2026-06-16 01:25:29,331][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.09325986355543137, 'train/grad_norm_pre_clip': 21.610084533691406, 'timing/data': 0.005995948042254895, 'timing/model': 1.3002652950235642, 'learning_rate/qwen_vl_interface': 3.2551455409535274e-06, 'learning_rate/action_model': 3.255145540953528e-05, 'epoch': 1.33}) +[2026-06-16 01:25:33,820][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.004757912829518318, 'train/grad_norm_pre_clip': 0.93719482421875, 'timing/data': 0.006310877972282469, 'timing/model': 1.2469962949980982, 'learning_rate/qwen_vl_interface': 3.2523821145909907e-06, 'learning_rate/action_model': 3.252382114590991e-05, 'epoch': 1.33}) +[2026-06-16 01:25:38,199][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.019702011719346046, 'train/grad_norm_pre_clip': 10.675374031066895, 'timing/data': 0.0062064569792710245, 'timing/model': 1.1297350529930554, 'learning_rate/qwen_vl_interface': 3.249619509373712e-06, 'learning_rate/action_model': 3.249619509373712e-05, 'epoch': 1.33}) +[2026-06-16 01:25:42,727][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.13906729221343994, 'train/grad_norm_pre_clip': 25.155044555664062, 'timing/data': 0.0064726099953986704, 'timing/model': 1.1342894149711356, 'learning_rate/qwen_vl_interface': 3.2468577264372905e-06, 'learning_rate/action_model': 3.24685772643729e-05, 'epoch': 1.33}) +[2026-06-16 01:25:47,297][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.01890779659152031, 'train/grad_norm_pre_clip': 11.354887962341309, 'timing/data': 0.006554495019372553, 'timing/model': 1.140523235022556, 'learning_rate/qwen_vl_interface': 3.244096766916992e-06, 'learning_rate/action_model': 3.244096766916992e-05, 'epoch': 1.33}) +[2026-06-16 01:25:51,831][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.1814802885055542, 'train/grad_norm_pre_clip': 8.355403900146484, 'timing/data': 0.006116461998317391, 'timing/model': 1.130497056990862, 'learning_rate/qwen_vl_interface': 3.241336631947743e-06, 'learning_rate/action_model': 3.2413366319477427e-05, 'epoch': 1.33}) +[2026-06-16 01:25:56,379][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.0013529964489862323, 'train/grad_norm_pre_clip': 0.3404255509376526, 'timing/data': 0.0062445030198432505, 'timing/model': 1.294035780010745, 'learning_rate/qwen_vl_interface': 3.238577322664131e-06, 'learning_rate/action_model': 3.238577322664131e-05, 'epoch': 1.33}) +[2026-06-16 01:26:00,787][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.07422443479299545, 'train/grad_norm_pre_clip': 10.57436466217041, 'timing/data': 0.006400799960829318, 'timing/model': 1.1436586389900185, 'learning_rate/qwen_vl_interface': 3.2358188402004026e-06, 'learning_rate/action_model': 3.2358188402004024e-05, 'epoch': 1.33}) +[2026-06-16 01:26:05,336][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.0035064462572336197, 'train/grad_norm_pre_clip': 0.6774126887321472, 'timing/data': 0.0064453810336999595, 'timing/model': 1.1370059689506888, 'learning_rate/qwen_vl_interface': 3.2330611856904675e-06, 'learning_rate/action_model': 3.233061185690468e-05, 'epoch': 1.33}) +[2026-06-16 01:26:09,876][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.0008039207896217704, 'train/grad_norm_pre_clip': 0.21298788487911224, 'timing/data': 0.0065744739840738475, 'timing/model': 1.1370640479726717, 'learning_rate/qwen_vl_interface': 3.2303043602678928e-06, 'learning_rate/action_model': 3.230304360267893e-05, 'epoch': 1.33}) +[2026-06-16 01:26:14,411][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.30300888419151306, 'train/grad_norm_pre_clip': 24.110889434814453, 'timing/data': 0.006301969988271594, 'timing/model': 1.1369213549769484, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 1.33}) +[2026-06-16 01:26:18,876][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.09137271344661713, 'train/grad_norm_pre_clip': 10.704773902893066, 'timing/data': 0.006259334972128272, 'timing/model': 1.1368251579697244, 'learning_rate/qwen_vl_interface': 3.2247932012173887e-06, 'learning_rate/action_model': 3.2247932012173884e-05, 'epoch': 1.33}) +[2026-06-16 01:26:23,373][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.007330912631005049, 'train/grad_norm_pre_clip': 5.665933609008789, 'timing/data': 0.006309980992227793, 'timing/model': 1.2535216949763708, 'learning_rate/qwen_vl_interface': 3.2220388698548897e-06, 'learning_rate/action_model': 3.22203886985489e-05, 'epoch': 1.33}) +[2026-06-16 01:26:27,755][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.012676678597927094, 'train/grad_norm_pre_clip': 4.670429706573486, 'timing/data': 0.0065397549769841135, 'timing/model': 1.1352527939598076, 'learning_rate/qwen_vl_interface': 3.21928537211061e-06, 'learning_rate/action_model': 3.21928537211061e-05, 'epoch': 1.33}) +[2026-06-16 01:26:32,285][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.010885111056268215, 'train/grad_norm_pre_clip': 3.6024746894836426, 'timing/data': 0.00625239999499172, 'timing/model': 1.1361414570128545, 'learning_rate/qwen_vl_interface': 3.216532709116405e-06, 'learning_rate/action_model': 3.216532709116405e-05, 'epoch': 1.33}) +[2026-06-16 01:26:36,813][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.2519685626029968, 'train/grad_norm_pre_clip': 10.923883438110352, 'timing/data': 0.0061982800252735615, 'timing/model': 1.1277709159767255, 'learning_rate/qwen_vl_interface': 3.2137808820037925e-06, 'learning_rate/action_model': 3.2137808820037924e-05, 'epoch': 1.33}) +[2026-06-16 01:26:41,348][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.11913834512233734, 'train/grad_norm_pre_clip': 33.55363845825195, 'timing/data': 0.00649843696737662, 'timing/model': 1.135417448997032, 'learning_rate/qwen_vl_interface': 3.211029891903945e-06, 'learning_rate/action_model': 3.211029891903945e-05, 'epoch': 1.33}) +[2026-06-16 01:26:45,842][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.00029540318064391613, 'train/grad_norm_pre_clip': 0.05605585500597954, 'timing/data': 0.006322919973172247, 'timing/model': 1.1323725950205699, 'learning_rate/qwen_vl_interface': 3.2082797399476894e-06, 'learning_rate/action_model': 3.2082797399476895e-05, 'epoch': 1.33}) +[2026-06-16 01:26:50,332][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.13634918630123138, 'train/grad_norm_pre_clip': 7.583462238311768, 'timing/data': 0.006554435007274151, 'timing/model': 1.252023592009209, 'learning_rate/qwen_vl_interface': 3.2055304272655114e-06, 'learning_rate/action_model': 3.205530427265511e-05, 'epoch': 1.33}) +[2026-06-16 01:26:54,710][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.16209781169891357, 'train/grad_norm_pre_clip': 29.0860538482666, 'timing/data': 0.006472584966104478, 'timing/model': 1.1328478410141543, 'learning_rate/qwen_vl_interface': 3.2027819549875477e-06, 'learning_rate/action_model': 3.202781954987548e-05, 'epoch': 1.33}) +[2026-06-16 01:26:59,266][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.04364471137523651, 'train/grad_norm_pre_clip': 18.363252639770508, 'timing/data': 0.006636823003645986, 'timing/model': 1.1493665640009567, 'learning_rate/qwen_vl_interface': 3.2000343242435926e-06, 'learning_rate/action_model': 3.2000343242435924e-05, 'epoch': 1.33}) +[2026-06-16 01:27:03,839][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.0003651517035905272, 'train/grad_norm_pre_clip': 0.048852939158678055, 'timing/data': 0.006573894992470741, 'timing/model': 1.1495933199767023, 'learning_rate/qwen_vl_interface': 3.1972875361630923e-06, 'learning_rate/action_model': 3.1972875361630923e-05, 'epoch': 1.33}) +[2026-06-16 01:27:08,384][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.015028882771730423, 'train/grad_norm_pre_clip': 12.508523941040039, 'timing/data': 0.006717543990816921, 'timing/model': 1.132295374001842, 'learning_rate/qwen_vl_interface': 3.1945415918751484e-06, 'learning_rate/action_model': 3.194541591875148e-05, 'epoch': 1.33}) +[2026-06-16 01:27:12,909][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.36484646797180176, 'train/grad_norm_pre_clip': 45.25843048095703, 'timing/data': 0.006310248048976064, 'timing/model': 1.2875286920461804, 'learning_rate/qwen_vl_interface': 3.191796492508516e-06, 'learning_rate/action_model': 3.191796492508516e-05, 'epoch': 1.33}) +[2026-06-16 01:27:17,281][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.012629210948944092, 'train/grad_norm_pre_clip': 4.37091064453125, 'timing/data': 0.0062753219972364604, 'timing/model': 1.1301645179628395, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 1.34}) +[2026-06-16 01:27:21,804][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.0004937663325108588, 'train/grad_norm_pre_clip': 0.04760270193219185, 'timing/data': 0.0062344399630092084, 'timing/model': 1.129575937986374, 'learning_rate/qwen_vl_interface': 3.1863088330524577e-06, 'learning_rate/action_model': 3.186308833052458e-05, 'epoch': 1.34}) +[2026-06-16 01:27:26,262][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.11166632175445557, 'train/grad_norm_pre_clip': 21.155994415283203, 'timing/data': 0.006307285977527499, 'timing/model': 1.134722476999741, 'learning_rate/qwen_vl_interface': 3.183566275218803e-06, 'learning_rate/action_model': 3.1835662752188026e-05, 'epoch': 1.34}) +[2026-06-16 01:27:30,771][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.004447726998478174, 'train/grad_norm_pre_clip': 1.7777639627456665, 'timing/data': 0.006314155005384237, 'timing/model': 1.134201826993376, 'learning_rate/qwen_vl_interface': 3.180824566817996e-06, 'learning_rate/action_model': 3.1808245668179956e-05, 'epoch': 1.34}) +[2026-06-16 01:27:35,274][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.03968532383441925, 'train/grad_norm_pre_clip': 15.624650001525879, 'timing/data': 0.0060022499528713524, 'timing/model': 1.1355047329561785, 'learning_rate/qwen_vl_interface': 3.1780837089770524e-06, 'learning_rate/action_model': 3.1780837089770524e-05, 'epoch': 1.34}) +[2026-06-16 01:27:39,816][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.013544882647693157, 'train/grad_norm_pre_clip': 10.437859535217285, 'timing/data': 0.006459669035393745, 'timing/model': 1.2922397240181454, 'learning_rate/qwen_vl_interface': 3.17534370282263e-06, 'learning_rate/action_model': 3.1753437028226303e-05, 'epoch': 1.34}) +[2026-06-16 01:27:44,197][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.0023667532950639725, 'train/grad_norm_pre_clip': 0.4638320207595825, 'timing/data': 0.00624903803691268, 'timing/model': 1.1327079559559934, 'learning_rate/qwen_vl_interface': 3.172604549481046e-06, 'learning_rate/action_model': 3.1726045494810455e-05, 'epoch': 1.34}) +[2026-06-16 01:27:48,736][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.006189826875925064, 'train/grad_norm_pre_clip': 1.6213923692703247, 'timing/data': 0.006379450031090528, 'timing/model': 1.132634176989086, 'learning_rate/qwen_vl_interface': 3.169866250078261e-06, 'learning_rate/action_model': 3.169866250078261e-05, 'epoch': 1.34}) +[2026-06-16 01:27:53,278][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.001640540431253612, 'train/grad_norm_pre_clip': 0.4533499777317047, 'timing/data': 0.006023827008903027, 'timing/model': 1.1321612870087847, 'learning_rate/qwen_vl_interface': 3.167128805739887e-06, 'learning_rate/action_model': 3.1671288057398866e-05, 'epoch': 1.34}) +[2026-06-16 01:27:57,781][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.0434487983584404, 'train/grad_norm_pre_clip': 11.440155029296875, 'timing/data': 0.006792906031478196, 'timing/model': 1.1389518940122798, 'learning_rate/qwen_vl_interface': 3.1643922175911847e-06, 'learning_rate/action_model': 3.164392217591185e-05, 'epoch': 1.34}) +[2026-06-16 01:28:02,281][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.0010248972102999687, 'train/grad_norm_pre_clip': 0.19706343114376068, 'timing/data': 0.005982036993373185, 'timing/model': 1.1342063680058345, 'learning_rate/qwen_vl_interface': 3.1616564867570603e-06, 'learning_rate/action_model': 3.16165648675706e-05, 'epoch': 1.34}) +[2026-06-16 01:28:06,822][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.0014710307586938143, 'train/grad_norm_pre_clip': 0.33850395679473877, 'timing/data': 0.006314545986242592, 'timing/model': 1.2893826839863323, 'learning_rate/qwen_vl_interface': 3.1589216143620733e-06, 'learning_rate/action_model': 3.1589216143620734e-05, 'epoch': 1.34}) +[2026-06-16 01:28:11,222][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.357308030128479, 'train/grad_norm_pre_clip': 33.44304275512695, 'timing/data': 0.006380298000294715, 'timing/model': 1.1384157939464785, 'learning_rate/qwen_vl_interface': 3.1561876015304215e-06, 'learning_rate/action_model': 3.156187601530421e-05, 'epoch': 1.34}) +[2026-06-16 01:28:15,779][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.010863653384149075, 'train/grad_norm_pre_clip': 5.437536716461182, 'timing/data': 0.006281437003053725, 'timing/model': 1.1302214179886505, 'learning_rate/qwen_vl_interface': 3.1534544493859563e-06, 'learning_rate/action_model': 3.1534544493859566e-05, 'epoch': 1.34}) +[2026-06-16 01:28:20,299][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.006333295721560717, 'train/grad_norm_pre_clip': 1.464327335357666, 'timing/data': 0.0065383840119466186, 'timing/model': 1.1346368470112793, 'learning_rate/qwen_vl_interface': 3.150722159052174e-06, 'learning_rate/action_model': 3.150722159052174e-05, 'epoch': 1.34}) +[2026-06-16 01:28:24,783][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.07283403724431992, 'train/grad_norm_pre_clip': 18.839130401611328, 'timing/data': 0.006244633987080306, 'timing/model': 1.1378029250190593, 'learning_rate/qwen_vl_interface': 3.147990731652215e-06, 'learning_rate/action_model': 3.147990731652215e-05, 'epoch': 1.34}) +[2026-06-16 01:28:29,329][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.19904141128063202, 'train/grad_norm_pre_clip': 21.140851974487305, 'timing/data': 0.0065051590208895504, 'timing/model': 1.1323819999815896, 'learning_rate/qwen_vl_interface': 3.1452601683088673e-06, 'learning_rate/action_model': 3.145260168308867e-05, 'epoch': 1.34}) +[2026-06-16 01:28:33,863][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.2446286976337433, 'train/grad_norm_pre_clip': 34.02924728393555, 'timing/data': 0.006766929989680648, 'timing/model': 1.2899281170102768, 'learning_rate/qwen_vl_interface': 3.1425304701445608e-06, 'learning_rate/action_model': 3.1425304701445605e-05, 'epoch': 1.34}) +[2026-06-16 01:28:38,249][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.08689698576927185, 'train/grad_norm_pre_clip': 12.700663566589355, 'timing/data': 0.006028970004990697, 'timing/model': 1.1407776940031908, 'learning_rate/qwen_vl_interface': 3.139801638281372e-06, 'learning_rate/action_model': 3.139801638281372e-05, 'epoch': 1.34}) +[2026-06-16 01:28:42,795][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.03327683359384537, 'train/grad_norm_pre_clip': 12.593117713928223, 'timing/data': 0.006262595008593053, 'timing/model': 1.1362875320483, 'learning_rate/qwen_vl_interface': 3.13707367384102e-06, 'learning_rate/action_model': 3.13707367384102e-05, 'epoch': 1.34}) +[2026-06-16 01:28:47,277][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.13902032375335693, 'train/grad_norm_pre_clip': 15.337312698364258, 'timing/data': 0.006465065001975745, 'timing/model': 1.1348183539812453, 'learning_rate/qwen_vl_interface': 3.134346577944869e-06, 'learning_rate/action_model': 3.134346577944869e-05, 'epoch': 1.34}) +[2026-06-16 01:28:51,775][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.07454083859920502, 'train/grad_norm_pre_clip': 18.695070266723633, 'timing/data': 0.006250659003853798, 'timing/model': 1.133050873002503, 'learning_rate/qwen_vl_interface': 3.1316203517139243e-06, 'learning_rate/action_model': 3.131620351713924e-05, 'epoch': 1.34}) +[2026-06-16 01:28:56,309][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.08998576551675797, 'train/grad_norm_pre_clip': 16.909997940063477, 'timing/data': 0.006423384009394795, 'timing/model': 1.1371889400179498, 'learning_rate/qwen_vl_interface': 3.128894996268837e-06, 'learning_rate/action_model': 3.128894996268837e-05, 'epoch': 1.34}) +[2026-06-16 01:29:00,839][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.04326383396983147, 'train/grad_norm_pre_clip': 12.69289493560791, 'timing/data': 0.006350024952553213, 'timing/model': 1.2907142160111107, 'learning_rate/qwen_vl_interface': 3.1261705127298934e-06, 'learning_rate/action_model': 3.1261705127298935e-05, 'epoch': 1.34}) +[2026-06-16 01:29:05,228][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.004550030454993248, 'train/grad_norm_pre_clip': 0.5103846788406372, 'timing/data': 0.0065792910172604024, 'timing/model': 1.1407266229507513, 'learning_rate/qwen_vl_interface': 3.123446902217028e-06, 'learning_rate/action_model': 3.1234469022170283e-05, 'epoch': 1.34}) +[2026-06-16 01:29:09,796][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.01459748949855566, 'train/grad_norm_pre_clip': 5.604487895965576, 'timing/data': 0.006298406049609184, 'timing/model': 1.1309844719944522, 'learning_rate/qwen_vl_interface': 3.1207241658498145e-06, 'learning_rate/action_model': 3.1207241658498145e-05, 'epoch': 1.35}) +[2026-06-16 01:29:14,332][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.044792164117097855, 'train/grad_norm_pre_clip': 9.83602523803711, 'timing/data': 0.007010890985839069, 'timing/model': 1.1354897490236908, 'learning_rate/qwen_vl_interface': 3.1180023047474667e-06, 'learning_rate/action_model': 3.1180023047474664e-05, 'epoch': 1.35}) +[2026-06-16 01:29:18,846][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.00488729914650321, 'train/grad_norm_pre_clip': 4.069458961486816, 'timing/data': 0.006349926989059895, 'timing/model': 1.1319985809968784, 'learning_rate/qwen_vl_interface': 3.1152813200288386e-06, 'learning_rate/action_model': 3.115281320028838e-05, 'epoch': 1.35}) +[2026-06-16 01:29:23,394][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.18591979146003723, 'train/grad_norm_pre_clip': 14.1187744140625, 'timing/data': 0.006445436971262097, 'timing/model': 1.1343757070135325, 'learning_rate/qwen_vl_interface': 3.1125612128124244e-06, 'learning_rate/action_model': 3.112561212812424e-05, 'epoch': 1.35}) +[2026-06-16 01:29:27,956][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.0029659876599907875, 'train/grad_norm_pre_clip': 0.8853901028633118, 'timing/data': 0.0068631519679911435, 'timing/model': 1.3068440959905274, 'learning_rate/qwen_vl_interface': 3.1098419842163573e-06, 'learning_rate/action_model': 3.109841984216357e-05, 'epoch': 1.35}) +[2026-06-16 01:29:32,352][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.0067137558944523335, 'train/grad_norm_pre_clip': 2.7011096477508545, 'timing/data': 0.006960782979149371, 'timing/model': 1.1400536819710396, 'learning_rate/qwen_vl_interface': 3.10712363535841e-06, 'learning_rate/action_model': 3.10712363535841e-05, 'epoch': 1.35}) +[2026-06-16 01:29:36,882][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.16106456518173218, 'train/grad_norm_pre_clip': 60.65866470336914, 'timing/data': 0.006565113959368318, 'timing/model': 1.1359553590300493, 'learning_rate/qwen_vl_interface': 3.1044061673559926e-06, 'learning_rate/action_model': 3.1044061673559925e-05, 'epoch': 1.35}) +[2026-06-16 01:29:41,372][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.15483157336711884, 'train/grad_norm_pre_clip': 21.485584259033203, 'timing/data': 0.006617205974180251, 'timing/model': 1.1302989270188846, 'learning_rate/qwen_vl_interface': 3.1016895813261556e-06, 'learning_rate/action_model': 3.101689581326155e-05, 'epoch': 1.35}) +[2026-06-16 01:29:45,906][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.03182712197303772, 'train/grad_norm_pre_clip': 19.355180740356445, 'timing/data': 0.006086879991926253, 'timing/model': 1.138201585970819, 'learning_rate/qwen_vl_interface': 3.098973878385581e-06, 'learning_rate/action_model': 3.098973878385581e-05, 'epoch': 1.35}) +[2026-06-16 01:29:50,443][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.010808276012539864, 'train/grad_norm_pre_clip': 4.549100399017334, 'timing/data': 0.006703853025101125, 'timing/model': 1.1369287679553963, 'learning_rate/qwen_vl_interface': 3.096259059650594e-06, 'learning_rate/action_model': 3.096259059650594e-05, 'epoch': 1.35}) +[2026-06-16 01:29:54,976][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.04935909062623978, 'train/grad_norm_pre_clip': 11.670340538024902, 'timing/data': 0.006021436012815684, 'timing/model': 1.2803767750156112, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 1.35}) +[2026-06-16 01:29:59,442][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.1263650357723236, 'train/grad_norm_pre_clip': 18.522777557373047, 'timing/data': 0.006228336016647518, 'timing/model': 1.2198596830130555, 'learning_rate/qwen_vl_interface': 3.090832079260855e-06, 'learning_rate/action_model': 3.0908320792608544e-05, 'epoch': 1.35}) +[2026-06-16 01:30:03,828][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.0038303034380078316, 'train/grad_norm_pre_clip': 2.1089534759521484, 'timing/data': 0.006706101004965603, 'timing/model': 1.1323711749864742, 'learning_rate/qwen_vl_interface': 3.088119919836929e-06, 'learning_rate/action_model': 3.088119919836929e-05, 'epoch': 1.35}) +[2026-06-16 01:30:08,335][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.014689836651086807, 'train/grad_norm_pre_clip': 4.873835563659668, 'timing/data': 0.006395716045517474, 'timing/model': 1.1324954110314138, 'learning_rate/qwen_vl_interface': 3.0854086490802433e-06, 'learning_rate/action_model': 3.085408649080243e-05, 'epoch': 1.35}) +[2026-06-16 01:30:12,898][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.03903155401349068, 'train/grad_norm_pre_clip': 13.0393648147583, 'timing/data': 0.006709584966301918, 'timing/model': 1.1361672619823366, 'learning_rate/qwen_vl_interface': 3.082698268105299e-06, 'learning_rate/action_model': 3.082698268105299e-05, 'epoch': 1.35}) +[2026-06-16 01:30:17,457][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.0014507933519780636, 'train/grad_norm_pre_clip': 0.2899109423160553, 'timing/data': 0.006321833992842585, 'timing/model': 1.1414408690179698, 'learning_rate/qwen_vl_interface': 3.0799887780262295e-06, 'learning_rate/action_model': 3.0799887780262296e-05, 'epoch': 1.35}) +[2026-06-16 01:30:21,997][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.004500300623476505, 'train/grad_norm_pre_clip': 1.368224024772644, 'timing/data': 0.006509367027319968, 'timing/model': 1.2960854739649221, 'learning_rate/qwen_vl_interface': 3.0772801799568054e-06, 'learning_rate/action_model': 3.0772801799568054e-05, 'epoch': 1.35}) +[2026-06-16 01:30:26,506][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.058817241340875626, 'train/grad_norm_pre_clip': 20.422290802001953, 'timing/data': 0.0061880749999545515, 'timing/model': 1.2625908859772608, 'learning_rate/qwen_vl_interface': 3.07457247501043e-06, 'learning_rate/action_model': 3.0745724750104296e-05, 'epoch': 1.35}) +[2026-06-16 01:30:30,909][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.002376507967710495, 'train/grad_norm_pre_clip': 0.6027248501777649, 'timing/data': 0.006377522018738091, 'timing/model': 1.1330273750354536, 'learning_rate/qwen_vl_interface': 3.0718656643001353e-06, 'learning_rate/action_model': 3.071865664300135e-05, 'epoch': 1.35}) +[2026-06-16 01:30:35,450][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.021192654967308044, 'train/grad_norm_pre_clip': 8.789910316467285, 'timing/data': 0.00630278600146994, 'timing/model': 1.136111151019577, 'learning_rate/qwen_vl_interface': 3.069159748938591e-06, 'learning_rate/action_model': 3.069159748938591e-05, 'epoch': 1.35}) +[2026-06-16 01:30:39,989][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.122231125831604, 'train/grad_norm_pre_clip': 35.232330322265625, 'timing/data': 0.006519723043311387, 'timing/model': 1.1357993420097046, 'learning_rate/qwen_vl_interface': 3.066454730038096e-06, 'learning_rate/action_model': 3.0664547300380956e-05, 'epoch': 1.35}) +[2026-06-16 01:30:44,541][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.11351727694272995, 'train/grad_norm_pre_clip': 9.268163681030273, 'timing/data': 0.006399321020580828, 'timing/model': 1.1329699070192873, 'learning_rate/qwen_vl_interface': 3.0637506087105816e-06, 'learning_rate/action_model': 3.0637506087105815e-05, 'epoch': 1.35}) +[2026-06-16 01:30:49,086][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.13104569911956787, 'train/grad_norm_pre_clip': 23.868392944335938, 'timing/data': 0.006340286985505372, 'timing/model': 1.3063466459861957, 'learning_rate/qwen_vl_interface': 3.0610473860676103e-06, 'learning_rate/action_model': 3.06104738606761e-05, 'epoch': 1.35}) +[2026-06-16 01:30:53,481][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.04051988199353218, 'train/grad_norm_pre_clip': 16.04055404663086, 'timing/data': 0.006560404028277844, 'timing/model': 1.140016987978015, 'learning_rate/qwen_vl_interface': 3.0583450632203748e-06, 'learning_rate/action_model': 3.058345063220375e-05, 'epoch': 1.35}) +[2026-06-16 01:30:58,016][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.00034131098072975874, 'train/grad_norm_pre_clip': 0.16986891627311707, 'timing/data': 0.006243811978492886, 'timing/model': 1.1362358090118505, 'learning_rate/qwen_vl_interface': 3.055643641279699e-06, 'learning_rate/action_model': 3.055643641279699e-05, 'epoch': 1.36}) +[2026-06-16 01:35:21,120][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.0013094223104417324, 'train/grad_norm_pre_clip': 0.46290123462677, 'eval/loss': 0.051200274378061295, 'eval/flap_precision': 0.9508928571428571, 'eval/flap_recall': 0.9861111111111112, 'eval/flap_f1': 0.968181818181818, 'eval/flap_rate_model': 0.18666666666666668, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9681818181818179, 'timing/data': 0.0064234049641527236, 'timing/model': 1.1375058430130593, 'learning_rate/qwen_vl_interface': 3.0529431213560334e-06, 'learning_rate/action_model': 3.0529431213560334e-05, 'epoch': 1.36}) +[2026-06-16 01:35:21,120][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3300_state +[2026-06-16 01:35:21,123][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 01:35:24,707][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3300_state/model.safetensors +[2026-06-16 01:35:34,728][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3300_state/optimizer.bin +[2026-06-16 01:35:34,729][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3300_state/sampler.bin +[2026-06-16 01:35:34,729][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3300_state/sampler_1.bin +[2026-06-16 01:35:34,729][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3300_state/random_states_0.pkl +[2026-06-16 01:35:34,775][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3300 +[2026-06-16 01:35:34,776][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 01:35:34,782][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 01:35:39,150][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.0011037243530154228, 'train/grad_norm_pre_clip': 0.302948534488678, 'timing/data': 0.006298647029325366, 'timing/model': 1.111965925025288, 'learning_rate/qwen_vl_interface': 3.0502435045594596e-06, 'learning_rate/action_model': 3.0502435045594595e-05, 'epoch': 1.36}) +[2026-06-16 01:35:43,508][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.00038402227801270783, 'train/grad_norm_pre_clip': 0.11224820464849472, 'timing/data': 0.006436327006667852, 'timing/model': 1.1118400460109115, 'learning_rate/qwen_vl_interface': 3.047544791999689e-06, 'learning_rate/action_model': 3.0475447919996892e-05, 'epoch': 1.36}) +[2026-06-16 01:35:47,907][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.00011930399341508746, 'train/grad_norm_pre_clip': 0.021879741922020912, 'timing/data': 0.006414859963115305, 'timing/model': 1.112734146998264, 'learning_rate/qwen_vl_interface': 3.044846984786061e-06, 'learning_rate/action_model': 3.044846984786061e-05, 'epoch': 1.36}) +[2026-06-16 01:35:52,339][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.013285873457789421, 'train/grad_norm_pre_clip': 8.038734436035156, 'timing/data': 0.006317712017334998, 'timing/model': 1.2457927459618077, 'learning_rate/qwen_vl_interface': 3.0421500840275413e-06, 'learning_rate/action_model': 3.0421500840275413e-05, 'epoch': 1.36}) +[2026-06-16 01:35:56,676][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.0004530130245257169, 'train/grad_norm_pre_clip': 0.09029686450958252, 'timing/data': 0.006320106971543282, 'timing/model': 1.1339428360224701, 'learning_rate/qwen_vl_interface': 3.039454090832727e-06, 'learning_rate/action_model': 3.0394540908327268e-05, 'epoch': 1.36}) +[2026-06-16 01:36:01,164][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.4349663257598877, 'train/grad_norm_pre_clip': 22.24330711364746, 'timing/data': 0.000462135998532176, 'timing/model': 1.1205737070413306, 'learning_rate/qwen_vl_interface': 3.0367590063098333e-06, 'learning_rate/action_model': 3.0367590063098333e-05, 'epoch': 1.36}) +[2026-06-16 01:36:05,641][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.21432676911354065, 'train/grad_norm_pre_clip': 26.757122039794922, 'timing/data': 0.0003408449701964855, 'timing/model': 1.1189142260118388, 'learning_rate/qwen_vl_interface': 3.0340648315667115e-06, 'learning_rate/action_model': 3.0340648315667113e-05, 'epoch': 1.36}) +[2026-06-16 01:36:10,149][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.007439688313752413, 'train/grad_norm_pre_clip': 2.952235221862793, 'timing/data': 0.00036919600097462535, 'timing/model': 1.138174068997614, 'learning_rate/qwen_vl_interface': 3.0313715677108336e-06, 'learning_rate/action_model': 3.0313715677108335e-05, 'epoch': 1.36}) +[2026-06-16 01:36:14,551][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.019505072385072708, 'train/grad_norm_pre_clip': 8.626738548278809, 'timing/data': 0.00033267197431996465, 'timing/model': 1.1352132909814827, 'learning_rate/qwen_vl_interface': 3.0286792158492996e-06, 'learning_rate/action_model': 3.0286792158492994e-05, 'epoch': 1.36}) +[2026-06-16 01:36:19,040][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.0009947054786607623, 'train/grad_norm_pre_clip': 0.3021693527698517, 'timing/data': 0.000327391957398504, 'timing/model': 1.2752233149949461, 'learning_rate/qwen_vl_interface': 3.0259877770888334e-06, 'learning_rate/action_model': 3.0259877770888333e-05, 'epoch': 1.36}) +[2026-06-16 01:36:23,381][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 7.871744310250506e-05, 'train/grad_norm_pre_clip': 0.00805856566876173, 'timing/data': 0.00040611496660858393, 'timing/model': 1.1229550199932419, 'learning_rate/qwen_vl_interface': 3.0232972525357836e-06, 'learning_rate/action_model': 3.0232972525357834e-05, 'epoch': 1.36}) +[2026-06-16 01:36:27,897][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.015428158454596996, 'train/grad_norm_pre_clip': 15.391672134399414, 'timing/data': 0.0003613170119933784, 'timing/model': 1.1278692730120383, 'learning_rate/qwen_vl_interface': 3.020607643296123e-06, 'learning_rate/action_model': 3.0206076432961226e-05, 'epoch': 1.36}) +[2026-06-16 01:36:32,419][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.047948990017175674, 'train/grad_norm_pre_clip': 22.499963760375977, 'timing/data': 0.00034129200503230095, 'timing/model': 1.1240905089653097, 'learning_rate/qwen_vl_interface': 3.017918950475449e-06, 'learning_rate/action_model': 3.017918950475449e-05, 'epoch': 1.36}) +[2026-06-16 01:36:36,917][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.007131663151085377, 'train/grad_norm_pre_clip': 5.4885993003845215, 'timing/data': 0.0005886159487999976, 'timing/model': 1.123708366008941, 'learning_rate/qwen_vl_interface': 3.0152311751789815e-06, 'learning_rate/action_model': 3.0152311751789814e-05, 'epoch': 1.36}) +[2026-06-16 01:36:41,423][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.026790646836161613, 'train/grad_norm_pre_clip': 9.371132850646973, 'timing/data': 0.00033855903893709183, 'timing/model': 1.283653519989457, 'learning_rate/qwen_vl_interface': 3.0125443185115655e-06, 'learning_rate/action_model': 3.0125443185115655e-05, 'epoch': 1.36}) +[2026-06-16 01:36:45,766][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.12152613699436188, 'train/grad_norm_pre_clip': 17.757225036621094, 'timing/data': 0.0003480369923636317, 'timing/model': 1.1250636110198684, 'learning_rate/qwen_vl_interface': 3.0098583815776615e-06, 'learning_rate/action_model': 3.0098583815776615e-05, 'epoch': 1.36}) +[2026-06-16 01:36:50,264][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.00261473236605525, 'train/grad_norm_pre_clip': 0.8063349723815918, 'timing/data': 0.0004248340264894068, 'timing/model': 1.1253084969939664, 'learning_rate/qwen_vl_interface': 3.0071733654813595e-06, 'learning_rate/action_model': 3.0071733654813595e-05, 'epoch': 1.36}) +[2026-06-16 01:36:54,691][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.05612790584564209, 'train/grad_norm_pre_clip': 25.121829986572266, 'timing/data': 0.0005585970357060432, 'timing/model': 1.1358758339774795, 'learning_rate/qwen_vl_interface': 3.0044892713263685e-06, 'learning_rate/action_model': 3.0044892713263682e-05, 'epoch': 1.36}) +[2026-06-16 01:36:59,144][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.07830742001533508, 'train/grad_norm_pre_clip': 26.4876766204834, 'timing/data': 0.0005645659985020757, 'timing/model': 1.1252637199941091, 'learning_rate/qwen_vl_interface': 3.001806100216017e-06, 'learning_rate/action_model': 3.001806100216017e-05, 'epoch': 1.36}) +[2026-06-16 01:37:03,648][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.24070660769939423, 'train/grad_norm_pre_clip': 22.804048538208008, 'timing/data': 0.0003463120083324611, 'timing/model': 1.1274712749873288, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 1.36}) +[2026-06-16 01:37:08,186][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.00065594632178545, 'train/grad_norm_pre_clip': 0.2779315710067749, 'timing/data': 0.000323692976962775, 'timing/model': 1.2997494430164807, 'learning_rate/qwen_vl_interface': 2.996442531540657e-06, 'learning_rate/action_model': 2.9964425315406568e-05, 'epoch': 1.36}) +[2026-06-16 01:37:12,542][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.12993137538433075, 'train/grad_norm_pre_clip': 15.312676429748535, 'timing/data': 0.00035365403164178133, 'timing/model': 1.1303398839663714, 'learning_rate/qwen_vl_interface': 2.993762136180408e-06, 'learning_rate/action_model': 2.993762136180408e-05, 'epoch': 1.36}) +[2026-06-16 01:37:17,091][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.24469904601573944, 'train/grad_norm_pre_clip': 20.17206573486328, 'timing/data': 0.00034479895839467645, 'timing/model': 1.1284559210180305, 'learning_rate/qwen_vl_interface': 2.9910826682743195e-06, 'learning_rate/action_model': 2.9910826682743192e-05, 'epoch': 1.37}) +[2026-06-16 01:37:21,623][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.011253736913204193, 'train/grad_norm_pre_clip': 5.782055854797363, 'timing/data': 0.00039619399467483163, 'timing/model': 1.1357560830074362, 'learning_rate/qwen_vl_interface': 2.9884041289238192e-06, 'learning_rate/action_model': 2.988404128923819e-05, 'epoch': 1.37}) +[2026-06-16 01:37:26,142][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.08958360552787781, 'train/grad_norm_pre_clip': 7.699244022369385, 'timing/data': 0.00035418698098510504, 'timing/model': 1.129126011976041, 'learning_rate/qwen_vl_interface': 2.9857265192299557e-06, 'learning_rate/action_model': 2.9857265192299556e-05, 'epoch': 1.37}) +[2026-06-16 01:37:30,674][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.0036933831870555878, 'train/grad_norm_pre_clip': 1.828184723854065, 'timing/data': 0.00033758097561076283, 'timing/model': 1.1310682100011036, 'learning_rate/qwen_vl_interface': 2.9830498402933894e-06, 'learning_rate/action_model': 2.9830498402933894e-05, 'epoch': 1.37}) +[2026-06-16 01:37:35,222][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.0136374207213521, 'train/grad_norm_pre_clip': 6.8281731605529785, 'timing/data': 0.0007447800016961992, 'timing/model': 1.3022923959651962, 'learning_rate/qwen_vl_interface': 2.9803740932144025e-06, 'learning_rate/action_model': 2.9803740932144024e-05, 'epoch': 1.37}) +[2026-06-16 01:37:39,590][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.04998075217008591, 'train/grad_norm_pre_clip': 32.43069839477539, 'timing/data': 0.00034072500420734286, 'timing/model': 1.1251000400516205, 'learning_rate/qwen_vl_interface': 2.9776992790928953e-06, 'learning_rate/action_model': 2.9776992790928952e-05, 'epoch': 1.37}) +[2026-06-16 01:37:44,116][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.00016278719704132527, 'train/grad_norm_pre_clip': 0.03473774716258049, 'timing/data': 0.00033098895801231265, 'timing/model': 1.136562257015612, 'learning_rate/qwen_vl_interface': 2.975025399028382e-06, 'learning_rate/action_model': 2.975025399028382e-05, 'epoch': 1.37}) +[2026-06-16 01:37:48,669][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.006732365116477013, 'train/grad_norm_pre_clip': 2.044724225997925, 'timing/data': 0.00033697596518322825, 'timing/model': 1.1309664850123227, 'learning_rate/qwen_vl_interface': 2.9723524541199943e-06, 'learning_rate/action_model': 2.972352454119994e-05, 'epoch': 1.37}) +[2026-06-16 01:37:53,190][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.017886903136968613, 'train/grad_norm_pre_clip': 15.352639198303223, 'timing/data': 0.00033496704418212175, 'timing/model': 1.1314484179602005, 'learning_rate/qwen_vl_interface': 2.969680445466478e-06, 'learning_rate/action_model': 2.969680445466478e-05, 'epoch': 1.37}) +[2026-06-16 01:37:57,672][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.08083244413137436, 'train/grad_norm_pre_clip': 30.4541072845459, 'timing/data': 0.00035179901169613004, 'timing/model': 1.1301094349473715, 'learning_rate/qwen_vl_interface': 2.9670093741661955e-06, 'learning_rate/action_model': 2.9670093741661954e-05, 'epoch': 1.37}) +[2026-06-16 01:38:02,192][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.1441127061843872, 'train/grad_norm_pre_clip': 20.80451011657715, 'timing/data': 0.0003501080209389329, 'timing/model': 1.2896143940160982, 'learning_rate/qwen_vl_interface': 2.964339241317123e-06, 'learning_rate/action_model': 2.964339241317123e-05, 'epoch': 1.37}) +[2026-06-16 01:38:06,552][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.003955720923841, 'train/grad_norm_pre_clip': 0.7448657751083374, 'timing/data': 0.0005245120264589787, 'timing/model': 1.1280036669922993, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 1.37}) +[2026-06-16 01:38:11,079][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.1988428384065628, 'train/grad_norm_pre_clip': 29.83460807800293, 'timing/data': 0.00039656803710386157, 'timing/model': 1.1285352220293134, 'learning_rate/qwen_vl_interface': 2.9590017953625854e-06, 'learning_rate/action_model': 2.9590017953625853e-05, 'epoch': 1.37}) +[2026-06-16 01:38:15,605][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.21573488414287567, 'train/grad_norm_pre_clip': 9.225960731506348, 'timing/data': 0.0003490810049697757, 'timing/model': 1.1351690579904243, 'learning_rate/qwen_vl_interface': 2.9563344844511455e-06, 'learning_rate/action_model': 2.9563344844511455e-05, 'epoch': 1.37}) +[2026-06-16 01:38:20,132][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.0005254486459307373, 'train/grad_norm_pre_clip': 0.17140823602676392, 'timing/data': 0.0003453290555626154, 'timing/model': 1.1408975729718804, 'learning_rate/qwen_vl_interface': 2.9536681163789568e-06, 'learning_rate/action_model': 2.9536681163789564e-05, 'epoch': 1.37}) +[2026-06-16 01:38:24,659][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.1606450080871582, 'train/grad_norm_pre_clip': 7.3644866943359375, 'timing/data': 0.00033108802745118737, 'timing/model': 1.1284767640172504, 'learning_rate/qwen_vl_interface': 2.9510026922420653e-06, 'learning_rate/action_model': 2.9510026922420653e-05, 'epoch': 1.37}) +[2026-06-16 01:38:29,177][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.00017028034199029207, 'train/grad_norm_pre_clip': 0.06367070972919464, 'timing/data': 0.00035287003265693784, 'timing/model': 1.2855729650473222, 'learning_rate/qwen_vl_interface': 2.9483382131361273e-06, 'learning_rate/action_model': 2.948338213136127e-05, 'epoch': 1.37}) +[2026-06-16 01:38:33,545][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.22349002957344055, 'train/grad_norm_pre_clip': 22.353666305541992, 'timing/data': 0.0003408319898881018, 'timing/model': 1.129793944011908, 'learning_rate/qwen_vl_interface': 2.9456746801564072e-06, 'learning_rate/action_model': 2.945674680156407e-05, 'epoch': 1.37}) +[2026-06-16 01:38:38,080][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.09567058086395264, 'train/grad_norm_pre_clip': 11.222116470336914, 'timing/data': 0.00036024197470396757, 'timing/model': 1.1346713040256873, 'learning_rate/qwen_vl_interface': 2.9430120943977854e-06, 'learning_rate/action_model': 2.9430120943977853e-05, 'epoch': 1.37}) +[2026-06-16 01:38:42,603][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.02170114777982235, 'train/grad_norm_pre_clip': 15.574969291687012, 'timing/data': 0.0003146230010315776, 'timing/model': 1.1294536350178532, 'learning_rate/qwen_vl_interface': 2.940350456954748e-06, 'learning_rate/action_model': 2.940350456954748e-05, 'epoch': 1.37}) +[2026-06-16 01:38:47,134][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.014473751187324524, 'train/grad_norm_pre_clip': 6.534656047821045, 'timing/data': 0.0003620879724621773, 'timing/model': 1.1371001480147243, 'learning_rate/qwen_vl_interface': 2.9376897689213952e-06, 'learning_rate/action_model': 2.937689768921395e-05, 'epoch': 1.37}) +[2026-06-16 01:38:51,632][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.33939242362976074, 'train/grad_norm_pre_clip': 33.91399002075195, 'timing/data': 0.0003237659693695605, 'timing/model': 1.1334699399885722, 'learning_rate/qwen_vl_interface': 2.9350300313914364e-06, 'learning_rate/action_model': 2.9350300313914363e-05, 'epoch': 1.37}) +[2026-06-16 01:38:56,145][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.003020680509507656, 'train/grad_norm_pre_clip': 0.8660205602645874, 'timing/data': 0.0003356800298206508, 'timing/model': 1.285705232003238, 'learning_rate/qwen_vl_interface': 2.9323712454581867e-06, 'learning_rate/action_model': 2.9323712454581866e-05, 'epoch': 1.37}) +[2026-06-16 01:39:00,509][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.0011774217709898949, 'train/grad_norm_pre_clip': 0.4377458095550537, 'timing/data': 0.00036538299173116684, 'timing/model': 1.1303354020346887, 'learning_rate/qwen_vl_interface': 2.929713412214572e-06, 'learning_rate/action_model': 2.929713412214572e-05, 'epoch': 1.37}) +[2026-06-16 01:39:05,043][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.0019737365655601025, 'train/grad_norm_pre_clip': 0.3232406675815582, 'timing/data': 0.00032091804314404726, 'timing/model': 1.1303965219995007, 'learning_rate/qwen_vl_interface': 2.9270565327531294e-06, 'learning_rate/action_model': 2.927056532753129e-05, 'epoch': 1.37}) +[2026-06-16 01:39:09,570][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.0001571009197505191, 'train/grad_norm_pre_clip': 0.025657327845692635, 'timing/data': 0.0003672530292533338, 'timing/model': 1.1315470759873278, 'learning_rate/qwen_vl_interface': 2.924400608166003e-06, 'learning_rate/action_model': 2.924400608166003e-05, 'epoch': 1.38}) +[2026-06-16 01:39:14,093][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.01722283847630024, 'train/grad_norm_pre_clip': 7.432015419006348, 'timing/data': 0.000352540984749794, 'timing/model': 1.1309871470439248, 'learning_rate/qwen_vl_interface': 2.921745639544939e-06, 'learning_rate/action_model': 2.921745639544939e-05, 'epoch': 1.38}) +[2026-06-16 01:39:18,630][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.07804376631975174, 'train/grad_norm_pre_clip': 10.658391952514648, 'timing/data': 0.0003178790211677551, 'timing/model': 1.1303766870405525, 'learning_rate/qwen_vl_interface': 2.919091627981297e-06, 'learning_rate/action_model': 2.919091627981297e-05, 'epoch': 1.38}) +[2026-06-16 01:39:23,168][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.0009199969354085624, 'train/grad_norm_pre_clip': 0.1418447196483612, 'timing/data': 0.00042919005500152707, 'timing/model': 1.2910253300215118, 'learning_rate/qwen_vl_interface': 2.9164385745660396e-06, 'learning_rate/action_model': 2.9164385745660394e-05, 'epoch': 1.38}) +[2026-06-16 01:39:27,542][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.1461133360862732, 'train/grad_norm_pre_clip': 26.466371536254883, 'timing/data': 0.00032408098923042417, 'timing/model': 1.1316219249856658, 'learning_rate/qwen_vl_interface': 2.9137864803897385e-06, 'learning_rate/action_model': 2.9137864803897387e-05, 'epoch': 1.38}) +[2026-06-16 01:39:32,065][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.13094840943813324, 'train/grad_norm_pre_clip': 42.88417053222656, 'timing/data': 0.0003497029538266361, 'timing/model': 1.1336640800000168, 'learning_rate/qwen_vl_interface': 2.9111353465425696e-06, 'learning_rate/action_model': 2.9111353465425694e-05, 'epoch': 1.38}) +[2026-06-16 01:39:36,619][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.05515628680586815, 'train/grad_norm_pre_clip': 20.994447708129883, 'timing/data': 0.00035677599953487515, 'timing/model': 1.1406773580238223, 'learning_rate/qwen_vl_interface': 2.908485174114312e-06, 'learning_rate/action_model': 2.908485174114312e-05, 'epoch': 1.38}) +[2026-06-16 01:39:41,154][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.1829298734664917, 'train/grad_norm_pre_clip': 27.22239112854004, 'timing/data': 0.0005011609755456448, 'timing/model': 1.1298512320499867, 'learning_rate/qwen_vl_interface': 2.9058359641943523e-06, 'learning_rate/action_model': 2.9058359641943523e-05, 'epoch': 1.38}) +[2026-06-16 01:39:45,651][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.0013891214039176702, 'train/grad_norm_pre_clip': 0.31867071986198425, 'timing/data': 0.0005473719793371856, 'timing/model': 1.1325553419883363, 'learning_rate/qwen_vl_interface': 2.9031877178716823e-06, 'learning_rate/action_model': 2.903187717871682e-05, 'epoch': 1.38}) +[2026-06-16 01:39:50,185][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.00363472243770957, 'train/grad_norm_pre_clip': 0.876594066619873, 'timing/data': 0.00045181100722402334, 'timing/model': 1.288277227024082, 'learning_rate/qwen_vl_interface': 2.900540436234894e-06, 'learning_rate/action_model': 2.9005404362348937e-05, 'epoch': 1.38}) +[2026-06-16 01:39:54,564][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.003940167371183634, 'train/grad_norm_pre_clip': 1.1128227710723877, 'timing/data': 0.00039466703310608864, 'timing/model': 1.1320736550260335, 'learning_rate/qwen_vl_interface': 2.897894120372185e-06, 'learning_rate/action_model': 2.897894120372185e-05, 'epoch': 1.38}) +[2026-06-16 01:39:59,074][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.03391152620315552, 'train/grad_norm_pre_clip': 10.11362075805664, 'timing/data': 0.0004031750140711665, 'timing/model': 1.1300237950053997, 'learning_rate/qwen_vl_interface': 2.8952487713713585e-06, 'learning_rate/action_model': 2.8952487713713582e-05, 'epoch': 1.38}) +[2026-06-16 01:40:03,604][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.0026803722139447927, 'train/grad_norm_pre_clip': 0.7448339462280273, 'timing/data': 0.00034619501093402505, 'timing/model': 1.1341261169873178, 'learning_rate/qwen_vl_interface': 2.8926043903198135e-06, 'learning_rate/action_model': 2.8926043903198135e-05, 'epoch': 1.38}) +[2026-06-16 01:40:08,120][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.15229956805706024, 'train/grad_norm_pre_clip': 30.24114227294922, 'timing/data': 0.00032887503039091825, 'timing/model': 1.1318274610093795, 'learning_rate/qwen_vl_interface': 2.8899609783045574e-06, 'learning_rate/action_model': 2.8899609783045573e-05, 'epoch': 1.38}) +[2026-06-16 01:40:12,677][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.0008351256255991757, 'train/grad_norm_pre_clip': 0.27529048919677734, 'timing/data': 0.0006545679643750191, 'timing/model': 1.1318236060324125, 'learning_rate/qwen_vl_interface': 2.8873185364121965e-06, 'learning_rate/action_model': 2.8873185364121963e-05, 'epoch': 1.38}) +[2026-06-16 01:40:17,230][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.016965219751000404, 'train/grad_norm_pre_clip': 17.37611961364746, 'timing/data': 0.0003232500166632235, 'timing/model': 1.3039860210265033, 'learning_rate/qwen_vl_interface': 2.8846770657289386e-06, 'learning_rate/action_model': 2.8846770657289386e-05, 'epoch': 1.38}) +[2026-06-16 01:40:21,619][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.00026812514988705516, 'train/grad_norm_pre_clip': 0.05299755558371544, 'timing/data': 0.00036004697903990746, 'timing/model': 1.1337797950254753, 'learning_rate/qwen_vl_interface': 2.882036567340593e-06, 'learning_rate/action_model': 2.8820365673405932e-05, 'epoch': 1.38}) +[2026-06-16 01:40:26,163][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.2557515799999237, 'train/grad_norm_pre_clip': 29.385318756103516, 'timing/data': 0.000539782049600035, 'timing/model': 1.1273936469806358, 'learning_rate/qwen_vl_interface': 2.8793970423325684e-06, 'learning_rate/action_model': 2.8793970423325685e-05, 'epoch': 1.38}) +[2026-06-16 01:40:30,703][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.0002428705629426986, 'train/grad_norm_pre_clip': 0.042482130229473114, 'timing/data': 0.0003589490079320967, 'timing/model': 1.1376688279560767, 'learning_rate/qwen_vl_interface': 2.8767584917898756e-06, 'learning_rate/action_model': 2.8767584917898754e-05, 'epoch': 1.38}) +[2026-06-16 01:40:35,241][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.2321212887763977, 'train/grad_norm_pre_clip': 22.7784481048584, 'timing/data': 0.0004465020028874278, 'timing/model': 1.129672899027355, 'learning_rate/qwen_vl_interface': 2.874120916797121e-06, 'learning_rate/action_model': 2.8741209167971207e-05, 'epoch': 1.38}) +[2026-06-16 01:40:39,737][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.21001815795898438, 'train/grad_norm_pre_clip': 15.35411262512207, 'timing/data': 0.00033576402347534895, 'timing/model': 1.1309224179713055, 'learning_rate/qwen_vl_interface': 2.8714843184385124e-06, 'learning_rate/action_model': 2.871484318438512e-05, 'epoch': 1.38}) +[2026-06-16 01:40:44,278][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.015385592356324196, 'train/grad_norm_pre_clip': 7.785471439361572, 'timing/data': 0.00040841702139005065, 'timing/model': 1.2966394940158352, 'learning_rate/qwen_vl_interface': 2.8688486977978575e-06, 'learning_rate/action_model': 2.8688486977978572e-05, 'epoch': 1.38}) +[2026-06-16 01:40:48,676][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.0013379761949181557, 'train/grad_norm_pre_clip': 0.4196253716945648, 'timing/data': 0.0003467610222287476, 'timing/model': 1.140283768007066, 'learning_rate/qwen_vl_interface': 2.8662140559585584e-06, 'learning_rate/action_model': 2.866214055958558e-05, 'epoch': 1.38}) +[2026-06-16 01:40:53,199][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.21109755337238312, 'train/grad_norm_pre_clip': 26.734718322753906, 'timing/data': 0.00040895299753174186, 'timing/model': 1.1335849390015937, 'learning_rate/qwen_vl_interface': 2.8635803940036162e-06, 'learning_rate/action_model': 2.863580394003616e-05, 'epoch': 1.38}) +[2026-06-16 01:40:57,723][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.008215157315135002, 'train/grad_norm_pre_clip': 1.8742516040802002, 'timing/data': 0.0004142040270380676, 'timing/model': 1.1315114250173792, 'learning_rate/qwen_vl_interface': 2.8609477130156317e-06, 'learning_rate/action_model': 2.8609477130156314e-05, 'epoch': 1.39}) +[2026-06-16 01:41:02,228][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.01156847644597292, 'train/grad_norm_pre_clip': 9.276266098022461, 'timing/data': 0.0004439899930730462, 'timing/model': 1.1316411949810572, 'learning_rate/qwen_vl_interface': 2.8583160140767997e-06, 'learning_rate/action_model': 2.8583160140767994e-05, 'epoch': 1.39}) +[2026-06-16 01:41:06,785][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.01379077136516571, 'train/grad_norm_pre_clip': 3.430027723312378, 'timing/data': 0.0003529640380293131, 'timing/model': 1.1524937869980931, 'learning_rate/qwen_vl_interface': 2.8556852982689125e-06, 'learning_rate/action_model': 2.8556852982689126e-05, 'epoch': 1.39}) +[2026-06-16 01:41:11,317][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.16488613188266754, 'train/grad_norm_pre_clip': 15.436671257019043, 'timing/data': 0.0005333320004865527, 'timing/model': 1.2956738579669036, 'learning_rate/qwen_vl_interface': 2.8530555666733584e-06, 'learning_rate/action_model': 2.8530555666733583e-05, 'epoch': 1.39}) +[2026-06-16 01:41:15,687][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.11943742632865906, 'train/grad_norm_pre_clip': 17.144634246826172, 'timing/data': 0.00034532195422798395, 'timing/model': 1.1318657899973914, 'learning_rate/qwen_vl_interface': 2.85042682037112e-06, 'learning_rate/action_model': 2.85042682037112e-05, 'epoch': 1.39}) +[2026-06-16 01:41:20,241][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.001484746695496142, 'train/grad_norm_pre_clip': 0.39881008863449097, 'timing/data': 0.0003458270221017301, 'timing/model': 1.142619108024519, 'learning_rate/qwen_vl_interface': 2.8477990604427764e-06, 'learning_rate/action_model': 2.8477990604427763e-05, 'epoch': 1.39}) +[2026-06-16 01:41:24,789][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.45326268672943115, 'train/grad_norm_pre_clip': 30.210906982421875, 'timing/data': 0.00031432800460606813, 'timing/model': 1.1334315279964358, 'learning_rate/qwen_vl_interface': 2.8451722879684994e-06, 'learning_rate/action_model': 2.8451722879684994e-05, 'epoch': 1.39}) +[2026-06-16 01:41:29,326][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.003666582517325878, 'train/grad_norm_pre_clip': 1.1896040439605713, 'timing/data': 0.00036370399175211787, 'timing/model': 1.1295923659927212, 'learning_rate/qwen_vl_interface': 2.842546504028059e-06, 'learning_rate/action_model': 2.8425465040280586e-05, 'epoch': 1.39}) +[2026-06-16 01:41:33,823][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.007045468781143427, 'train/grad_norm_pre_clip': 2.5043325424194336, 'timing/data': 0.00034982897341251373, 'timing/model': 1.1305998230236582, 'learning_rate/qwen_vl_interface': 2.83992170970081e-06, 'learning_rate/action_model': 2.8399217097008102e-05, 'epoch': 1.39}) +[2026-06-16 01:41:38,351][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.0006446018232963979, 'train/grad_norm_pre_clip': 0.1508081704378128, 'timing/data': 0.00035501003731042147, 'timing/model': 1.291031158005353, 'learning_rate/qwen_vl_interface': 2.837297906065711e-06, 'learning_rate/action_model': 2.8372979060657107e-05, 'epoch': 1.39}) +[2026-06-16 01:41:42,738][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.15982520580291748, 'train/grad_norm_pre_clip': 17.99370574951172, 'timing/data': 0.00033055501990020275, 'timing/model': 1.139046723023057, 'learning_rate/qwen_vl_interface': 2.834675094201306e-06, 'learning_rate/action_model': 2.8346750942013063e-05, 'epoch': 1.39}) +[2026-06-16 01:41:47,282][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.12530969083309174, 'train/grad_norm_pre_clip': 19.56976890563965, 'timing/data': 0.00036372803151607513, 'timing/model': 1.1348492730176076, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 1.39}) +[2026-06-16 01:41:51,831][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.23532947897911072, 'train/grad_norm_pre_clip': 26.53541374206543, 'timing/data': 0.0003595770103856921, 'timing/model': 1.1317701519583352, 'learning_rate/qwen_vl_interface': 2.829432450096731e-06, 'learning_rate/action_model': 2.8294324500967306e-05, 'epoch': 1.39}) +[2026-06-16 01:41:56,352][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.06873339414596558, 'train/grad_norm_pre_clip': 10.587953567504883, 'timing/data': 0.0003568760002963245, 'timing/model': 1.1342177729820833, 'learning_rate/qwen_vl_interface': 2.8268126200116118e-06, 'learning_rate/action_model': 2.8268126200116117e-05, 'epoch': 1.39}) +[2026-06-16 01:42:00,887][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.06443746387958527, 'train/grad_norm_pre_clip': 33.288089752197266, 'timing/data': 0.00035556498914957047, 'timing/model': 1.1293169230339117, 'learning_rate/qwen_vl_interface': 2.824193786007294e-06, 'learning_rate/action_model': 2.8241937860072937e-05, 'epoch': 1.39}) +[2026-06-16 01:42:05,434][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.0012488587526604533, 'train/grad_norm_pre_clip': 0.3162142336368561, 'timing/data': 0.0003394889645278454, 'timing/model': 1.2842268510139547, 'learning_rate/qwen_vl_interface': 2.821575949160279e-06, 'learning_rate/action_model': 2.8215759491602787e-05, 'epoch': 1.39}) +[2026-06-16 01:42:09,794][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.06319799274206161, 'train/grad_norm_pre_clip': 22.46354866027832, 'timing/data': 0.0003308629966340959, 'timing/model': 1.1304884770070203, 'learning_rate/qwen_vl_interface': 2.8189591105466625e-06, 'learning_rate/action_model': 2.8189591105466624e-05, 'epoch': 1.39}) +[2026-06-16 01:42:14,320][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.021337203681468964, 'train/grad_norm_pre_clip': 12.647517204284668, 'timing/data': 0.00037420098669826984, 'timing/model': 1.1327370490180328, 'learning_rate/qwen_vl_interface': 2.816343271242126e-06, 'learning_rate/action_model': 2.816343271242126e-05, 'epoch': 1.39}) +[2026-06-16 01:42:18,841][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.028630925342440605, 'train/grad_norm_pre_clip': 16.072574615478516, 'timing/data': 0.00031177897471934557, 'timing/model': 1.1294199309777468, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 1.39}) +[2026-06-16 01:42:23,374][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.11009177565574646, 'train/grad_norm_pre_clip': 14.592254638671875, 'timing/data': 0.0003394189989194274, 'timing/model': 1.1306427259696648, 'learning_rate/qwen_vl_interface': 2.81111459486098e-06, 'learning_rate/action_model': 2.81111459486098e-05, 'epoch': 1.39}) +[2026-06-16 01:42:27,876][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.09140171110630035, 'train/grad_norm_pre_clip': 28.083786010742188, 'timing/data': 0.000360839010681957, 'timing/model': 1.137460956000723, 'learning_rate/qwen_vl_interface': 2.8085017599336806e-06, 'learning_rate/action_model': 2.8085017599336806e-05, 'epoch': 1.39}) +[2026-06-16 01:42:32,420][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.07495221495628357, 'train/grad_norm_pre_clip': 13.416055679321289, 'timing/data': 0.00034882104955613613, 'timing/model': 1.2858573439880274, 'learning_rate/qwen_vl_interface': 2.8058899286140828e-06, 'learning_rate/action_model': 2.805889928614083e-05, 'epoch': 1.39}) +[2026-06-16 01:42:36,808][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.055261388421058655, 'train/grad_norm_pre_clip': 27.652572631835938, 'timing/data': 0.0003273870097473264, 'timing/model': 1.1354684039833955, 'learning_rate/qwen_vl_interface': 2.8032791019758127e-06, 'learning_rate/action_model': 2.8032791019758127e-05, 'epoch': 1.39}) +[2026-06-16 01:42:41,354][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.0007667432655580342, 'train/grad_norm_pre_clip': 0.17370019853115082, 'timing/data': 0.00035550998291000724, 'timing/model': 1.133527665049769, 'learning_rate/qwen_vl_interface': 2.800669281092084e-06, 'learning_rate/action_model': 2.8006692810920838e-05, 'epoch': 1.39}) +[2026-06-16 01:42:45,894][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.25976189970970154, 'train/grad_norm_pre_clip': 14.696024894714355, 'timing/data': 0.00033197400625795126, 'timing/model': 1.1358039579936303, 'learning_rate/qwen_vl_interface': 2.7980604670356938e-06, 'learning_rate/action_model': 2.7980604670356937e-05, 'epoch': 1.4}) +[2026-06-16 01:42:50,441][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.0014900090172886848, 'train/grad_norm_pre_clip': 0.41476884484291077, 'timing/data': 0.000347086985129863, 'timing/model': 1.1318006289657205, 'learning_rate/qwen_vl_interface': 2.795452660879027e-06, 'learning_rate/action_model': 2.7954526608790273e-05, 'epoch': 1.4}) +[2026-06-16 01:42:54,982][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.0002277855819556862, 'train/grad_norm_pre_clip': 0.0315636470913887, 'timing/data': 0.00033497001277282834, 'timing/model': 1.1327662179828621, 'learning_rate/qwen_vl_interface': 2.7928458636940565e-06, 'learning_rate/action_model': 2.7928458636940564e-05, 'epoch': 1.4}) +[2026-06-16 01:42:59,531][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.1716533601284027, 'train/grad_norm_pre_clip': 8.127862930297852, 'timing/data': 0.0005124160088598728, 'timing/model': 1.3033595390152186, 'learning_rate/qwen_vl_interface': 2.790240076552336e-06, 'learning_rate/action_model': 2.7902400765523355e-05, 'epoch': 1.4}) +[2026-06-16 01:47:19,520][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.1861097663640976, 'train/grad_norm_pre_clip': 20.296127319335938, 'eval/loss': 0.041273318231105804, 'eval/flap_precision': 0.9678407350689127, 'eval/flap_recall': 0.9753086419753086, 'eval/flap_f1': 0.9715603382013835, 'eval/flap_rate_model': 0.18138888888888888, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9715603382013835, 'timing/data': 0.00039836898213252425, 'timing/model': 1.1357134520076215, 'learning_rate/qwen_vl_interface': 2.787635300525006e-06, 'learning_rate/action_model': 2.787635300525006e-05, 'epoch': 1.4}) +[2026-06-16 01:47:19,520][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3400_state +[2026-06-16 01:47:19,523][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 01:47:23,108][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3400_state/model.safetensors +[2026-06-16 01:47:33,218][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3400_state/optimizer.bin +[2026-06-16 01:47:33,218][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3400_state/sampler.bin +[2026-06-16 01:47:33,218][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3400_state/sampler_1.bin +[2026-06-16 01:47:33,219][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3400_state/random_states_0.pkl +[2026-06-16 01:47:33,265][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3400 +[2026-06-16 01:47:33,266][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 01:47:33,272][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 01:47:37,645][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.11665452271699905, 'train/grad_norm_pre_clip': 27.725658416748047, 'timing/data': 0.0003374640364199877, 'timing/model': 1.1800893630133942, 'learning_rate/qwen_vl_interface': 2.7850315366827933e-06, 'learning_rate/action_model': 2.7850315366827934e-05, 'epoch': 1.4}) +[2026-06-16 01:47:41,940][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.12724654376506805, 'train/grad_norm_pre_clip': 11.602184295654297, 'timing/data': 0.0003757579834200442, 'timing/model': 1.118958481005393, 'learning_rate/qwen_vl_interface': 2.7824287860960076e-06, 'learning_rate/action_model': 2.7824287860960075e-05, 'epoch': 1.4}) +[2026-06-16 01:47:46,486][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.006577400956302881, 'train/grad_norm_pre_clip': 2.9442107677459717, 'timing/data': 0.00037290301406756043, 'timing/model': 1.1365463720285334, 'learning_rate/qwen_vl_interface': 2.7798270498345377e-06, 'learning_rate/action_model': 2.7798270498345376e-05, 'epoch': 1.4}) +[2026-06-16 01:47:50,973][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.3385350704193115, 'train/grad_norm_pre_clip': 13.632587432861328, 'timing/data': 0.00034546595998108387, 'timing/model': 1.1175550249754451, 'learning_rate/qwen_vl_interface': 2.777226328967861e-06, 'learning_rate/action_model': 2.7772263289678607e-05, 'epoch': 1.4}) +[2026-06-16 01:47:55,463][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.2584332823753357, 'train/grad_norm_pre_clip': 17.979074478149414, 'timing/data': 0.00037147902185097337, 'timing/model': 1.1215075820218772, 'learning_rate/qwen_vl_interface': 2.774626624565035e-06, 'learning_rate/action_model': 2.774626624565035e-05, 'epoch': 1.4}) +[2026-06-16 01:47:59,965][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.26934343576431274, 'train/grad_norm_pre_clip': 17.20623207092285, 'timing/data': 0.0003158409963361919, 'timing/model': 1.2821961470181122, 'learning_rate/qwen_vl_interface': 2.772027937694701e-06, 'learning_rate/action_model': 2.772027937694701e-05, 'epoch': 1.4}) +[2026-06-16 01:48:04,307][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.0027775284834206104, 'train/grad_norm_pre_clip': 0.5566675662994385, 'timing/data': 0.0003798609832301736, 'timing/model': 1.1248443549848162, 'learning_rate/qwen_vl_interface': 2.7694302694250806e-06, 'learning_rate/action_model': 2.7694302694250808e-05, 'epoch': 1.4}) +[2026-06-16 01:48:08,798][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.18255263566970825, 'train/grad_norm_pre_clip': 30.893043518066406, 'timing/data': 0.0003364000003784895, 'timing/model': 1.1201431169756688, 'learning_rate/qwen_vl_interface': 2.7668336208239775e-06, 'learning_rate/action_model': 2.7668336208239776e-05, 'epoch': 1.4}) +[2026-06-16 01:48:13,305][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.03872774913907051, 'train/grad_norm_pre_clip': 13.257563591003418, 'timing/data': 0.00033347500720992684, 'timing/model': 1.1274874830269255, 'learning_rate/qwen_vl_interface': 2.7642379929587755e-06, 'learning_rate/action_model': 2.7642379929587754e-05, 'epoch': 1.4}) +[2026-06-16 01:48:17,843][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.06881231814622879, 'train/grad_norm_pre_clip': 7.027073860168457, 'timing/data': 0.0004254649975337088, 'timing/model': 1.1366292599705048, 'learning_rate/qwen_vl_interface': 2.7616433868964387e-06, 'learning_rate/action_model': 2.7616433868964388e-05, 'epoch': 1.4}) +[2026-06-16 01:48:22,378][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.1566038429737091, 'train/grad_norm_pre_clip': 7.138817310333252, 'timing/data': 0.00036772899329662323, 'timing/model': 1.1452211840078235, 'learning_rate/qwen_vl_interface': 2.7590498037035122e-06, 'learning_rate/action_model': 2.7590498037035124e-05, 'epoch': 1.4}) +[2026-06-16 01:48:26,882][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.015713654458522797, 'train/grad_norm_pre_clip': 7.90688943862915, 'timing/data': 0.0003296039649285376, 'timing/model': 1.2784530519857071, 'learning_rate/qwen_vl_interface': 2.7564572444461218e-06, 'learning_rate/action_model': 2.7564572444461217e-05, 'epoch': 1.4}) +[2026-06-16 01:48:31,365][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.003982993308454752, 'train/grad_norm_pre_clip': 0.6324965357780457, 'timing/data': 0.00032110995380207896, 'timing/model': 1.2532463049865328, 'learning_rate/qwen_vl_interface': 2.7538657101899674e-06, 'learning_rate/action_model': 2.7538657101899673e-05, 'epoch': 1.4}) +[2026-06-16 01:48:35,728][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.2981547713279724, 'train/grad_norm_pre_clip': 12.714797973632812, 'timing/data': 0.00034249399323016405, 'timing/model': 1.127924283966422, 'learning_rate/qwen_vl_interface': 2.751275202000332e-06, 'learning_rate/action_model': 2.7512752020003317e-05, 'epoch': 1.4}) +[2026-06-16 01:48:40,259][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.002628731308504939, 'train/grad_norm_pre_clip': 0.33625221252441406, 'timing/data': 0.0003619100316427648, 'timing/model': 1.13441308302572, 'learning_rate/qwen_vl_interface': 2.7486857209420763e-06, 'learning_rate/action_model': 2.7486857209420763e-05, 'epoch': 1.4}) +[2026-06-16 01:48:44,797][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.07426624000072479, 'train/grad_norm_pre_clip': 10.616146087646484, 'timing/data': 0.0003154560108669102, 'timing/model': 1.1354505629860796, 'learning_rate/qwen_vl_interface': 2.7460972680796377e-06, 'learning_rate/action_model': 2.7460972680796375e-05, 'epoch': 1.4}) +[2026-06-16 01:48:49,340][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.007482781540602446, 'train/grad_norm_pre_clip': 0.9838257431983948, 'timing/data': 0.00035732798278331757, 'timing/model': 1.130541201971937, 'learning_rate/qwen_vl_interface': 2.7435098444770314e-06, 'learning_rate/action_model': 2.7435098444770315e-05, 'epoch': 1.4}) +[2026-06-16 01:48:53,819][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.1912298947572708, 'train/grad_norm_pre_clip': 22.70741081237793, 'timing/data': 0.0003382200375199318, 'timing/model': 1.1306658170069568, 'learning_rate/qwen_vl_interface': 2.7409234511978495e-06, 'learning_rate/action_model': 2.7409234511978492e-05, 'epoch': 1.4}) +[2026-06-16 01:48:58,358][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.008568649180233479, 'train/grad_norm_pre_clip': 2.702704429626465, 'timing/data': 0.00039901997661218047, 'timing/model': 1.301517737971153, 'learning_rate/qwen_vl_interface': 2.7383380893052606e-06, 'learning_rate/action_model': 2.7383380893052608e-05, 'epoch': 1.4}) +[2026-06-16 01:49:02,753][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.028811683878302574, 'train/grad_norm_pre_clip': 5.442689895629883, 'timing/data': 0.0003683720133267343, 'timing/model': 1.1514167159912176, 'learning_rate/qwen_vl_interface': 2.73575375986201e-06, 'learning_rate/action_model': 2.7357537598620096e-05, 'epoch': 1.4}) +[2026-06-16 01:49:07,275][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.061904627829790115, 'train/grad_norm_pre_clip': 9.681118965148926, 'timing/data': 0.00035481899976730347, 'timing/model': 1.131999208999332, 'learning_rate/qwen_vl_interface': 2.733170463930417e-06, 'learning_rate/action_model': 2.7331704639304167e-05, 'epoch': 1.41}) +[2026-06-16 01:49:11,808][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.16420303285121918, 'train/grad_norm_pre_clip': 12.250995635986328, 'timing/data': 0.0003672119928523898, 'timing/model': 1.129341354011558, 'learning_rate/qwen_vl_interface': 2.7305882025723774e-06, 'learning_rate/action_model': 2.7305882025723774e-05, 'epoch': 1.41}) +[2026-06-16 01:49:16,338][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.0207444429397583, 'train/grad_norm_pre_clip': 8.317180633544922, 'timing/data': 0.0003684160183183849, 'timing/model': 1.1258587169577368, 'learning_rate/qwen_vl_interface': 2.7280069768493626e-06, 'learning_rate/action_model': 2.7280069768493628e-05, 'epoch': 1.41}) +[2026-06-16 01:49:20,867][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.008319037035107613, 'train/grad_norm_pre_clip': 3.6662216186523438, 'timing/data': 0.00032327097142115235, 'timing/model': 1.2858345799613744, 'learning_rate/qwen_vl_interface': 2.7254267878224144e-06, 'learning_rate/action_model': 2.7254267878224143e-05, 'epoch': 1.41}) +[2026-06-16 01:49:25,381][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.06099142134189606, 'train/grad_norm_pre_clip': 11.10533618927002, 'timing/data': 0.0003296509967185557, 'timing/model': 1.2666197229991667, 'learning_rate/qwen_vl_interface': 2.7228476365521524e-06, 'learning_rate/action_model': 2.722847636552152e-05, 'epoch': 1.41}) +[2026-06-16 01:49:29,765][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.0018438085680827498, 'train/grad_norm_pre_clip': 0.29784512519836426, 'timing/data': 0.00035198096884414554, 'timing/model': 1.140978146053385, 'learning_rate/qwen_vl_interface': 2.7202695240987674e-06, 'learning_rate/action_model': 2.7202695240987676e-05, 'epoch': 1.41}) +[2026-06-16 01:49:34,289][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.14566025137901306, 'train/grad_norm_pre_clip': 14.265460968017578, 'timing/data': 0.00041936803609132767, 'timing/model': 1.135614420985803, 'learning_rate/qwen_vl_interface': 2.717692451522025e-06, 'learning_rate/action_model': 2.7176924515220248e-05, 'epoch': 1.41}) +[2026-06-16 01:49:38,835][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.09997125715017319, 'train/grad_norm_pre_clip': 16.10227394104004, 'timing/data': 0.000353067007381469, 'timing/model': 1.134542251995299, 'learning_rate/qwen_vl_interface': 2.7151164198812624e-06, 'learning_rate/action_model': 2.7151164198812622e-05, 'epoch': 1.41}) +[2026-06-16 01:49:43,372][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.2763729691505432, 'train/grad_norm_pre_clip': 20.398893356323242, 'timing/data': 0.0003444300382398069, 'timing/model': 1.1315757780103013, 'learning_rate/qwen_vl_interface': 2.712541430235388e-06, 'learning_rate/action_model': 2.712541430235388e-05, 'epoch': 1.41}) +[2026-06-16 01:49:47,860][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.10340452194213867, 'train/grad_norm_pre_clip': 11.1940336227417, 'timing/data': 0.00040708802407607436, 'timing/model': 1.1355612039915286, 'learning_rate/qwen_vl_interface': 2.7099674836428842e-06, 'learning_rate/action_model': 2.709967483642884e-05, 'epoch': 1.41}) +[2026-06-16 01:49:52,390][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.12462787330150604, 'train/grad_norm_pre_clip': 19.955873489379883, 'timing/data': 0.0003348180325701833, 'timing/model': 1.2842213910189457, 'learning_rate/qwen_vl_interface': 2.7073945811618018e-06, 'learning_rate/action_model': 2.7073945811618017e-05, 'epoch': 1.41}) +[2026-06-16 01:49:56,771][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.013691961765289307, 'train/grad_norm_pre_clip': 5.865885257720947, 'timing/data': 0.0003501179744489491, 'timing/model': 1.133897129970137, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 1.41}) +[2026-06-16 01:50:01,301][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.31956154108047485, 'train/grad_norm_pre_clip': 13.941795349121094, 'timing/data': 0.00040725397411733866, 'timing/model': 1.1339097389718518, 'learning_rate/qwen_vl_interface': 2.7022519127639687e-06, 'learning_rate/action_model': 2.7022519127639682e-05, 'epoch': 1.41}) +[2026-06-16 01:50:05,840][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.2421906739473343, 'train/grad_norm_pre_clip': 19.42091178894043, 'timing/data': 0.0003450680524110794, 'timing/model': 1.129907256050501, 'learning_rate/qwen_vl_interface': 2.6996821489611713e-06, 'learning_rate/action_model': 2.699682148961171e-05, 'epoch': 1.41}) +[2026-06-16 01:50:10,373][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.04309678077697754, 'train/grad_norm_pre_clip': 20.342182159423828, 'timing/data': 0.00036908098263666034, 'timing/model': 1.136943571968004, 'learning_rate/qwen_vl_interface': 2.697113433497712e-06, 'learning_rate/action_model': 2.6971134334977115e-05, 'epoch': 1.41}) +[2026-06-16 01:50:14,907][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.30560463666915894, 'train/grad_norm_pre_clip': 13.183354377746582, 'timing/data': 0.000428021012339741, 'timing/model': 1.2875324600026943, 'learning_rate/qwen_vl_interface': 2.6945457674294863e-06, 'learning_rate/action_model': 2.694545767429486e-05, 'epoch': 1.41}) +[2026-06-16 01:50:19,418][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.08705328404903412, 'train/grad_norm_pre_clip': 10.138943672180176, 'timing/data': 0.00032561697298660874, 'timing/model': 1.2638515579747036, 'learning_rate/qwen_vl_interface': 2.6919791518119676e-06, 'learning_rate/action_model': 2.6919791518119674e-05, 'epoch': 1.41}) +[2026-06-16 01:50:23,792][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.04442115128040314, 'train/grad_norm_pre_clip': 11.915407180786133, 'timing/data': 0.0003643110394477844, 'timing/model': 1.134002276987303, 'learning_rate/qwen_vl_interface': 2.6894135877001946e-06, 'learning_rate/action_model': 2.689413587700195e-05, 'epoch': 1.41}) +[2026-06-16 01:50:28,328][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.01855376735329628, 'train/grad_norm_pre_clip': 6.098669052124023, 'timing/data': 0.00034323096042498946, 'timing/model': 1.1326882260036655, 'learning_rate/qwen_vl_interface': 2.686849076148775e-06, 'learning_rate/action_model': 2.6868490761487747e-05, 'epoch': 1.41}) +[2026-06-16 01:50:32,886][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.004342294298112392, 'train/grad_norm_pre_clip': 0.8174388408660889, 'timing/data': 0.0003498850273899734, 'timing/model': 1.1339371090289205, 'learning_rate/qwen_vl_interface': 2.68428561821188e-06, 'learning_rate/action_model': 2.6842856182118802e-05, 'epoch': 1.41}) +[2026-06-16 01:50:37,423][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.16783417761325836, 'train/grad_norm_pre_clip': 22.19811248779297, 'timing/data': 0.00035406003007665277, 'timing/model': 1.1381822780240327, 'learning_rate/qwen_vl_interface': 2.681723214943253e-06, 'learning_rate/action_model': 2.681723214943253e-05, 'epoch': 1.41}) +[2026-06-16 01:50:41,917][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.017070453613996506, 'train/grad_norm_pre_clip': 4.24005651473999, 'timing/data': 0.00033962499583140016, 'timing/model': 1.1310289870016277, 'learning_rate/qwen_vl_interface': 2.6791618673961996e-06, 'learning_rate/action_model': 2.6791618673961994e-05, 'epoch': 1.41}) +[2026-06-16 01:50:46,435][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.1865001618862152, 'train/grad_norm_pre_clip': 10.292533874511719, 'timing/data': 0.00033419299870729446, 'timing/model': 1.2887495110044256, 'learning_rate/qwen_vl_interface': 2.676601576623594e-06, 'learning_rate/action_model': 2.6766015766235937e-05, 'epoch': 1.41}) +[2026-06-16 01:50:50,817][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.08858034014701843, 'train/grad_norm_pre_clip': 9.94815444946289, 'timing/data': 0.000360086967702955, 'timing/model': 1.134090073988773, 'learning_rate/qwen_vl_interface': 2.674042343677874e-06, 'learning_rate/action_model': 2.6740423436778738e-05, 'epoch': 1.41}) +[2026-06-16 01:50:55,352][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.23071515560150146, 'train/grad_norm_pre_clip': 12.937849998474121, 'timing/data': 0.0003983090282417834, 'timing/model': 1.1300869130063802, 'learning_rate/qwen_vl_interface': 2.6714841696110456e-06, 'learning_rate/action_model': 2.6714841696110456e-05, 'epoch': 1.42}) +[2026-06-16 01:50:59,888][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.013419277034699917, 'train/grad_norm_pre_clip': 3.1636924743652344, 'timing/data': 0.0003274239716120064, 'timing/model': 1.136676313006319, 'learning_rate/qwen_vl_interface': 2.6689270554746743e-06, 'learning_rate/action_model': 2.6689270554746743e-05, 'epoch': 1.42}) +[2026-06-16 01:51:04,429][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.0029789393302053213, 'train/grad_norm_pre_clip': 0.34700319170951843, 'timing/data': 0.0003662510425783694, 'timing/model': 1.1346418210305274, 'learning_rate/qwen_vl_interface': 2.6663710023198956e-06, 'learning_rate/action_model': 2.6663710023198956e-05, 'epoch': 1.42}) +[2026-06-16 01:51:08,957][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.0565345324575901, 'train/grad_norm_pre_clip': 13.812731742858887, 'timing/data': 0.0003410649951547384, 'timing/model': 1.2892164360382594, 'learning_rate/qwen_vl_interface': 2.663816011197405e-06, 'learning_rate/action_model': 2.6638160111974048e-05, 'epoch': 1.42}) +[2026-06-16 01:51:13,456][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.024546466767787933, 'train/grad_norm_pre_clip': 4.721185207366943, 'timing/data': 0.00032481999369338155, 'timing/model': 1.258138613018673, 'learning_rate/qwen_vl_interface': 2.661262083157463e-06, 'learning_rate/action_model': 2.661262083157463e-05, 'epoch': 1.42}) +[2026-06-16 01:51:17,848][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.012387127615511417, 'train/grad_norm_pre_clip': 1.6401913166046143, 'timing/data': 0.00032061501406133175, 'timing/model': 1.135568408004474, 'learning_rate/qwen_vl_interface': 2.6587092192498943e-06, 'learning_rate/action_model': 2.658709219249894e-05, 'epoch': 1.42}) +[2026-06-16 01:51:22,397][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.0053904857486486435, 'train/grad_norm_pre_clip': 0.7085791230201721, 'timing/data': 0.00033923203591257334, 'timing/model': 1.1320237580221146, 'learning_rate/qwen_vl_interface': 2.656157420524084e-06, 'learning_rate/action_model': 2.6561574205240842e-05, 'epoch': 1.42}) +[2026-06-16 01:51:26,934][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.11469275504350662, 'train/grad_norm_pre_clip': 11.11617660522461, 'timing/data': 0.00042312603909522295, 'timing/model': 1.1346914379973896, 'learning_rate/qwen_vl_interface': 2.6536066880289795e-06, 'learning_rate/action_model': 2.6536066880289794e-05, 'epoch': 1.42}) +[2026-06-16 01:51:31,483][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.01751771941781044, 'train/grad_norm_pre_clip': 3.676028251647949, 'timing/data': 0.0003565740189515054, 'timing/model': 1.1337967590079643, 'learning_rate/qwen_vl_interface': 2.6510570228130916e-06, 'learning_rate/action_model': 2.6510570228130914e-05, 'epoch': 1.42}) +[2026-06-16 01:51:35,991][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.018214091658592224, 'train/grad_norm_pre_clip': 7.867063045501709, 'timing/data': 0.00038740801392123103, 'timing/model': 1.136328763968777, 'learning_rate/qwen_vl_interface': 2.648508425924491e-06, 'learning_rate/action_model': 2.6485084259244907e-05, 'epoch': 1.42}) +[2026-06-16 01:51:40,536][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.16643026471138, 'train/grad_norm_pre_clip': 16.02531623840332, 'timing/data': 0.0004053959855809808, 'timing/model': 1.2906570279737934, 'learning_rate/qwen_vl_interface': 2.6459608984108093e-06, 'learning_rate/action_model': 2.645960898410809e-05, 'epoch': 1.42}) +[2026-06-16 01:51:44,918][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.11858312785625458, 'train/grad_norm_pre_clip': 8.096221923828125, 'timing/data': 0.00037440803134813905, 'timing/model': 1.135960421001073, 'learning_rate/qwen_vl_interface': 2.6434144413192414e-06, 'learning_rate/action_model': 2.6434144413192414e-05, 'epoch': 1.42}) +[2026-06-16 01:51:49,463][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.15911707282066345, 'train/grad_norm_pre_clip': 26.11345672607422, 'timing/data': 0.0003384590381756425, 'timing/model': 1.1341155350091867, 'learning_rate/qwen_vl_interface': 2.6408690556965347e-06, 'learning_rate/action_model': 2.640869055696535e-05, 'epoch': 1.42}) +[2026-06-16 01:51:54,021][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.14718303084373474, 'train/grad_norm_pre_clip': 16.564300537109375, 'timing/data': 0.0003366839955560863, 'timing/model': 1.130482573993504, 'learning_rate/qwen_vl_interface': 2.6383247425890047e-06, 'learning_rate/action_model': 2.6383247425890047e-05, 'epoch': 1.42}) +[2026-06-16 01:51:58,546][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.1301775872707367, 'train/grad_norm_pre_clip': 23.95618438720703, 'timing/data': 0.00034879602026194334, 'timing/model': 1.1338249380351044, 'learning_rate/qwen_vl_interface': 2.6357815030425204e-06, 'learning_rate/action_model': 2.6357815030425205e-05, 'epoch': 1.42}) +[2026-06-16 01:52:03,090][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.011157695204019547, 'train/grad_norm_pre_clip': 2.0676939487457275, 'timing/data': 0.00031635299092158675, 'timing/model': 1.2951129709836096, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 1.42}) +[2026-06-16 01:52:07,604][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.007930554449558258, 'train/grad_norm_pre_clip': 1.0891178846359253, 'timing/data': 0.0003423590096645057, 'timing/model': 1.2674152330146171, 'learning_rate/qwen_vl_interface': 2.6306982488139704e-06, 'learning_rate/action_model': 2.6306982488139702e-05, 'epoch': 1.42}) +[2026-06-16 01:52:11,999][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.029552709311246872, 'train/grad_norm_pre_clip': 9.407950401306152, 'timing/data': 0.00039112899685278535, 'timing/model': 1.152187234023586, 'learning_rate/qwen_vl_interface': 2.628158236221438e-06, 'learning_rate/action_model': 2.628158236221438e-05, 'epoch': 1.42}) +[2026-06-16 01:52:16,541][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.15319181978702545, 'train/grad_norm_pre_clip': 8.835512161254883, 'timing/data': 0.0004390929825603962, 'timing/model': 1.1310432949685492, 'learning_rate/qwen_vl_interface': 2.6256193013690183e-06, 'learning_rate/action_model': 2.6256193013690182e-05, 'epoch': 1.42}) +[2026-06-16 01:52:21,081][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.13265885412693024, 'train/grad_norm_pre_clip': 13.775002479553223, 'timing/data': 0.0004603119450621307, 'timing/model': 1.133343315043021, 'learning_rate/qwen_vl_interface': 2.623081445300373e-06, 'learning_rate/action_model': 2.6230814453003728e-05, 'epoch': 1.42}) +[2026-06-16 01:52:25,624][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.0005100834532640874, 'train/grad_norm_pre_clip': 0.04774169251322746, 'timing/data': 0.00034972402499988675, 'timing/model': 1.1356856399797834, 'learning_rate/qwen_vl_interface': 2.6205446690587183e-06, 'learning_rate/action_model': 2.6205446690587182e-05, 'epoch': 1.42}) +[2026-06-16 01:52:30,121][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.004186364356428385, 'train/grad_norm_pre_clip': 0.37091320753097534, 'timing/data': 0.00038124300772324204, 'timing/model': 1.1334109929739498, 'learning_rate/qwen_vl_interface': 2.618008973686828e-06, 'learning_rate/action_model': 2.618008973686828e-05, 'epoch': 1.42}) +[2026-06-16 01:52:34,675][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.18727648258209229, 'train/grad_norm_pre_clip': 8.88060188293457, 'timing/data': 0.0003446699702180922, 'timing/model': 1.3094908199855126, 'learning_rate/qwen_vl_interface': 2.615474360227028e-06, 'learning_rate/action_model': 2.615474360227028e-05, 'epoch': 1.42}) +[2026-06-16 01:52:39,080][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.19929377734661102, 'train/grad_norm_pre_clip': 5.589777946472168, 'timing/data': 0.0004270420176908374, 'timing/model': 1.1455877629923634, 'learning_rate/qwen_vl_interface': 2.612940829721204e-06, 'learning_rate/action_model': 2.612940829721204e-05, 'epoch': 1.42}) +[2026-06-16 01:52:43,648][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.30081307888031006, 'train/grad_norm_pre_clip': 21.10828971862793, 'timing/data': 0.0003745279973372817, 'timing/model': 1.1459331280202605, 'learning_rate/qwen_vl_interface': 2.6104083832107948e-06, 'learning_rate/action_model': 2.6104083832107945e-05, 'epoch': 1.43}) +[2026-06-16 01:52:48,203][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.3288809657096863, 'train/grad_norm_pre_clip': 17.919485092163086, 'timing/data': 0.0003292459878139198, 'timing/model': 1.1315799719886854, 'learning_rate/qwen_vl_interface': 2.607877021736793e-06, 'learning_rate/action_model': 2.6078770217367926e-05, 'epoch': 1.43}) +[2026-06-16 01:52:52,739][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.07334746420383453, 'train/grad_norm_pre_clip': 14.984721183776855, 'timing/data': 0.0003556750016286969, 'timing/model': 1.129841926973313, 'learning_rate/qwen_vl_interface': 2.605346746339745e-06, 'learning_rate/action_model': 2.605346746339745e-05, 'epoch': 1.43}) +[2026-06-16 01:52:57,273][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.14142751693725586, 'train/grad_norm_pre_clip': 15.397350311279297, 'timing/data': 0.00034917000448331237, 'timing/model': 1.2889752379851416, 'learning_rate/qwen_vl_interface': 2.602817558059753e-06, 'learning_rate/action_model': 2.6028175580597532e-05, 'epoch': 1.43}) +[2026-06-16 01:53:01,786][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.03857434540987015, 'train/grad_norm_pre_clip': 6.90008020401001, 'timing/data': 0.00038797000888735056, 'timing/model': 1.264835252019111, 'learning_rate/qwen_vl_interface': 2.60028945793647e-06, 'learning_rate/action_model': 2.60028945793647e-05, 'epoch': 1.43}) +[2026-06-16 01:53:06,173][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.1651199758052826, 'train/grad_norm_pre_clip': 8.034109115600586, 'timing/data': 0.00038123398553580046, 'timing/model': 1.13495590997627, 'learning_rate/qwen_vl_interface': 2.5977624470091024e-06, 'learning_rate/action_model': 2.5977624470091023e-05, 'epoch': 1.43}) +[2026-06-16 01:53:10,724][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.18845435976982117, 'train/grad_norm_pre_clip': 6.34616231918335, 'timing/data': 0.000366182008292526, 'timing/model': 1.132773556979373, 'learning_rate/qwen_vl_interface': 2.59523652631641e-06, 'learning_rate/action_model': 2.5952365263164097e-05, 'epoch': 1.43}) +[2026-06-16 01:53:15,267][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.12801378965377808, 'train/grad_norm_pre_clip': 15.181792259216309, 'timing/data': 0.0003592929570004344, 'timing/model': 1.1351104489876889, 'learning_rate/qwen_vl_interface': 2.5927116968967024e-06, 'learning_rate/action_model': 2.5927116968967025e-05, 'epoch': 1.43}) +[2026-06-16 01:53:19,809][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.20095160603523254, 'train/grad_norm_pre_clip': 9.119900703430176, 'timing/data': 0.00033386697759851813, 'timing/model': 1.133243439020589, 'learning_rate/qwen_vl_interface': 2.5901879597878417e-06, 'learning_rate/action_model': 2.5901879597878415e-05, 'epoch': 1.43}) +[2026-06-16 01:53:24,330][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.03624168410897255, 'train/grad_norm_pre_clip': 13.095069885253906, 'timing/data': 0.0003487219801172614, 'timing/model': 1.1332729150308296, 'learning_rate/qwen_vl_interface': 2.5876653160272395e-06, 'learning_rate/action_model': 2.5876653160272397e-05, 'epoch': 1.43}) +[2026-06-16 01:53:28,875][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.094480499625206, 'train/grad_norm_pre_clip': 12.631433486938477, 'timing/data': 0.00032587500754743814, 'timing/model': 1.2959964629844762, 'learning_rate/qwen_vl_interface': 2.5851437666518607e-06, 'learning_rate/action_model': 2.5851437666518607e-05, 'epoch': 1.43}) +[2026-06-16 01:53:33,266][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.14601854979991913, 'train/grad_norm_pre_clip': 15.704145431518555, 'timing/data': 0.0003441850421950221, 'timing/model': 1.1328456129995175, 'learning_rate/qwen_vl_interface': 2.5826233126982198e-06, 'learning_rate/action_model': 2.5826233126982196e-05, 'epoch': 1.43}) +[2026-06-16 01:53:37,819][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.13290530443191528, 'train/grad_norm_pre_clip': 11.124159812927246, 'timing/data': 0.0006777290254831314, 'timing/model': 1.1314979729941115, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 1.43}) +[2026-06-16 01:53:42,360][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.03492024168372154, 'train/grad_norm_pre_clip': 13.38992977142334, 'timing/data': 0.0003488369984552264, 'timing/model': 1.1359927600133233, 'learning_rate/qwen_vl_interface': 2.577585695199954e-06, 'learning_rate/action_model': 2.5775856951999538e-05, 'epoch': 1.43}) +[2026-06-16 01:53:46,886][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.07352954894304276, 'train/grad_norm_pre_clip': 16.2160587310791, 'timing/data': 0.00032901798840612173, 'timing/model': 1.1393255920265801, 'learning_rate/qwen_vl_interface': 2.575068533726101e-06, 'learning_rate/action_model': 2.575068533726101e-05, 'epoch': 1.43}) +[2026-06-16 01:53:51,450][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.20066729187965393, 'train/grad_norm_pre_clip': 8.534439086914062, 'timing/data': 0.0003484599874354899, 'timing/model': 1.3088698239880614, 'learning_rate/qwen_vl_interface': 2.5725524718155333e-06, 'learning_rate/action_model': 2.572552471815533e-05, 'epoch': 1.43}) +[2026-06-16 01:53:55,965][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.02771281637251377, 'train/grad_norm_pre_clip': 4.477609157562256, 'timing/data': 0.00035965198185294867, 'timing/model': 1.2604912219685502, 'learning_rate/qwen_vl_interface': 2.570037510502508e-06, 'learning_rate/action_model': 2.5700375105025083e-05, 'epoch': 1.43}) +[2026-06-16 01:54:00,347][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.11793491989374161, 'train/grad_norm_pre_clip': 6.128012180328369, 'timing/data': 0.0005749919801019132, 'timing/model': 1.1312644979916513, 'learning_rate/qwen_vl_interface': 2.5675236508208313e-06, 'learning_rate/action_model': 2.5675236508208313e-05, 'epoch': 1.43}) +[2026-06-16 01:54:04,886][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.01465942244976759, 'train/grad_norm_pre_clip': 2.1745407581329346, 'timing/data': 0.0006443340098485351, 'timing/model': 1.1316833759774454, 'learning_rate/qwen_vl_interface': 2.565010893803856e-06, 'learning_rate/action_model': 2.565010893803856e-05, 'epoch': 1.43}) +[2026-06-16 01:54:09,441][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.015002163127064705, 'train/grad_norm_pre_clip': 2.0045833587646484, 'timing/data': 0.0004217469831928611, 'timing/model': 1.1357483299798332, 'learning_rate/qwen_vl_interface': 2.562499240484482e-06, 'learning_rate/action_model': 2.5624992404844816e-05, 'epoch': 1.43}) +[2026-06-16 01:54:13,982][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.013376814313232899, 'train/grad_norm_pre_clip': 1.2266942262649536, 'timing/data': 0.00042195001151412725, 'timing/model': 1.1358928530244157, 'learning_rate/qwen_vl_interface': 2.559988691895155e-06, 'learning_rate/action_model': 2.559988691895155e-05, 'epoch': 1.43}) +[2026-06-16 01:54:18,481][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.04271167889237404, 'train/grad_norm_pre_clip': 9.172307014465332, 'timing/data': 0.0003298520459793508, 'timing/model': 1.1325373240397312, 'learning_rate/qwen_vl_interface': 2.5574792490678648e-06, 'learning_rate/action_model': 2.5574792490678644e-05, 'epoch': 1.43}) +[2026-06-16 01:54:23,018][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.15909262001514435, 'train/grad_norm_pre_clip': 9.624690055847168, 'timing/data': 0.0003500860184431076, 'timing/model': 1.2974627270014025, 'learning_rate/qwen_vl_interface': 2.5549709130341493e-06, 'learning_rate/action_model': 2.5549709130341493e-05, 'epoch': 1.43}) +[2026-06-16 01:54:27,418][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.03426123410463333, 'train/grad_norm_pre_clip': 9.326305389404297, 'timing/data': 0.0003357460373081267, 'timing/model': 1.143900138034951, 'learning_rate/qwen_vl_interface': 2.5524636848250924e-06, 'learning_rate/action_model': 2.552463684825092e-05, 'epoch': 1.43}) +[2026-06-16 01:54:31,977][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.02743157558143139, 'train/grad_norm_pre_clip': 5.816983222961426, 'timing/data': 0.00031586800469085574, 'timing/model': 1.1411210150108673, 'learning_rate/qwen_vl_interface': 2.5499575654713186e-06, 'learning_rate/action_model': 2.5499575654713185e-05, 'epoch': 1.43}) +[2026-06-16 01:54:36,535][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.017841925844550133, 'train/grad_norm_pre_clip': 2.4770734310150146, 'timing/data': 0.00033405900467187166, 'timing/model': 1.1410935549647547, 'learning_rate/qwen_vl_interface': 2.547452556003001e-06, 'learning_rate/action_model': 2.547452556003001e-05, 'epoch': 1.44}) +[2026-06-16 01:54:41,096][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.010727813467383385, 'train/grad_norm_pre_clip': 3.598475456237793, 'timing/data': 0.0003423389862291515, 'timing/model': 1.1375073690433055, 'learning_rate/qwen_vl_interface': 2.5449486574498535e-06, 'learning_rate/action_model': 2.5449486574498534e-05, 'epoch': 1.44}) +[2026-06-16 01:54:45,623][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.011213677003979683, 'train/grad_norm_pre_clip': 1.9863381385803223, 'timing/data': 0.0003692650352604687, 'timing/model': 1.2820059960358776, 'learning_rate/qwen_vl_interface': 2.5424458708411335e-06, 'learning_rate/action_model': 2.5424458708411335e-05, 'epoch': 1.44}) +[2026-06-16 01:54:50,143][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.050804901868104935, 'train/grad_norm_pre_clip': 14.711898803710938, 'timing/data': 0.00034802197478711605, 'timing/model': 1.2656558550079353, 'learning_rate/qwen_vl_interface': 2.5399441972056437e-06, 'learning_rate/action_model': 2.5399441972056438e-05, 'epoch': 1.44}) +[2026-06-16 01:54:54,519][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.012301802635192871, 'train/grad_norm_pre_clip': 1.3069978952407837, 'timing/data': 0.0003584030200727284, 'timing/model': 1.1322913359617814, 'learning_rate/qwen_vl_interface': 2.537443637571726e-06, 'learning_rate/action_model': 2.537443637571726e-05, 'epoch': 1.44}) +[2026-06-16 01:54:59,062][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.13438713550567627, 'train/grad_norm_pre_clip': 5.584107875823975, 'timing/data': 0.0003494650009088218, 'timing/model': 1.1355746710323729, 'learning_rate/qwen_vl_interface': 2.534944192967269e-06, 'learning_rate/action_model': 2.534944192967269e-05, 'epoch': 1.44}) +[2026-06-16 01:59:21,527][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.1565341353416443, 'train/grad_norm_pre_clip': 9.181063652038574, 'eval/loss': 0.03262608125805855, 'eval/flap_precision': 0.9710365853658537, 'eval/flap_recall': 0.9830246913580247, 'eval/flap_f1': 0.976993865030675, 'eval/flap_rate_model': 0.18222222222222223, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.976993865030675, 'timing/data': 0.0003424130263738334, 'timing/model': 1.1357675030012615, 'learning_rate/qwen_vl_interface': 2.532445864419695e-06, 'learning_rate/action_model': 2.5324458644196952e-05, 'epoch': 1.44}) +[2026-06-16 01:59:21,528][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3500_state +[2026-06-16 01:59:21,530][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 01:59:25,116][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3500_state/model.safetensors +[2026-06-16 01:59:35,139][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3500_state/optimizer.bin +[2026-06-16 01:59:35,140][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3500_state/sampler.bin +[2026-06-16 01:59:35,140][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-16 01:59:35,141][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-16 01:59:35,188][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-16 01:59:35,188][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 01:59:35,195][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 01:59:39,689][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.14974673092365265, 'train/grad_norm_pre_clip': 9.25426197052002, 'timing/data': 0.00032892602030187845, 'timing/model': 1.1112750730244443, 'learning_rate/qwen_vl_interface': 2.5299486529559758e-06, 'learning_rate/action_model': 2.5299486529559756e-05, 'epoch': 1.44}) +[2026-06-16 01:59:44,150][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.0067900484427809715, 'train/grad_norm_pre_clip': 0.8487406373023987, 'timing/data': 0.000334159005433321, 'timing/model': 1.1150477469782345, 'learning_rate/qwen_vl_interface': 2.5274525596026202e-06, 'learning_rate/action_model': 2.52745255960262e-05, 'epoch': 1.44}) +[2026-06-16 01:59:48,600][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.00306190038099885, 'train/grad_norm_pre_clip': 0.32833337783813477, 'timing/data': 0.0003546960069797933, 'timing/model': 1.1126823240192607, 'learning_rate/qwen_vl_interface': 2.524957585385677e-06, 'learning_rate/action_model': 2.524957585385677e-05, 'epoch': 1.44}) +[2026-06-16 01:59:53,073][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.05943881720304489, 'train/grad_norm_pre_clip': 12.566059112548828, 'timing/data': 0.0003192800213582814, 'timing/model': 1.2806041459552944, 'learning_rate/qwen_vl_interface': 2.522463731330738e-06, 'learning_rate/action_model': 2.5224637313307374e-05, 'epoch': 1.44}) +[2026-06-16 01:59:57,544][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.08278486877679825, 'train/grad_norm_pre_clip': 12.293905258178711, 'timing/data': 0.00039047194877639413, 'timing/model': 1.2483466140110977, 'learning_rate/qwen_vl_interface': 2.5199709984629294e-06, 'learning_rate/action_model': 2.5199709984629292e-05, 'epoch': 1.44}) +[2026-06-16 02:00:01,886][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.008023969829082489, 'train/grad_norm_pre_clip': 5.309108257293701, 'timing/data': 0.000329977017827332, 'timing/model': 1.1317681110231206, 'learning_rate/qwen_vl_interface': 2.517479387806922e-06, 'learning_rate/action_model': 2.517479387806922e-05, 'epoch': 1.44}) +[2026-06-16 02:00:06,370][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.13097915053367615, 'train/grad_norm_pre_clip': 10.235884666442871, 'timing/data': 0.0003940309979952872, 'timing/model': 1.1304940329864621, 'learning_rate/qwen_vl_interface': 2.5149889003869217e-06, 'learning_rate/action_model': 2.5149889003869214e-05, 'epoch': 1.44}) +[2026-06-16 02:00:10,855][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.0011703497730195522, 'train/grad_norm_pre_clip': 0.09539703279733658, 'timing/data': 0.00035772804403677583, 'timing/model': 1.1243710150010884, 'learning_rate/qwen_vl_interface': 2.5124995372266736e-06, 'learning_rate/action_model': 2.5124995372266735e-05, 'epoch': 1.44}) +[2026-06-16 02:00:15,358][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.012635361403226852, 'train/grad_norm_pre_clip': 1.4193265438079834, 'timing/data': 0.0003356990055181086, 'timing/model': 1.1291371010011062, 'learning_rate/qwen_vl_interface': 2.5100112993494626e-06, 'learning_rate/action_model': 2.5100112993494624e-05, 'epoch': 1.44}) +[2026-06-16 02:00:19,857][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.03841136768460274, 'train/grad_norm_pre_clip': 5.124493598937988, 'timing/data': 0.0003209710121154785, 'timing/model': 1.2789440130000003, 'learning_rate/qwen_vl_interface': 2.5075241877781085e-06, 'learning_rate/action_model': 2.5075241877781086e-05, 'epoch': 1.44}) +[2026-06-16 02:00:24,194][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.09755228459835052, 'train/grad_norm_pre_clip': 10.030817985534668, 'timing/data': 0.00036753597669303417, 'timing/model': 1.122234321956057, 'learning_rate/qwen_vl_interface': 2.5050382035349673e-06, 'learning_rate/action_model': 2.5050382035349674e-05, 'epoch': 1.44}) +[2026-06-16 02:00:28,684][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.025482626631855965, 'train/grad_norm_pre_clip': 4.238770484924316, 'timing/data': 0.0003564489888958633, 'timing/model': 1.1230792989954352, 'learning_rate/qwen_vl_interface': 2.5025533476419343e-06, 'learning_rate/action_model': 2.5025533476419343e-05, 'epoch': 1.44}) +[2026-06-16 02:00:33,190][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.004676554352045059, 'train/grad_norm_pre_clip': 0.5470960140228271, 'timing/data': 0.000367639004252851, 'timing/model': 1.1325084150303155, 'learning_rate/qwen_vl_interface': 2.5000696211204394e-06, 'learning_rate/action_model': 2.5000696211204394e-05, 'epoch': 1.44}) +[2026-06-16 02:00:37,704][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.20115657150745392, 'train/grad_norm_pre_clip': 9.639182090759277, 'timing/data': 0.00033859204268082976, 'timing/model': 1.1350450950558297, 'learning_rate/qwen_vl_interface': 2.4975870249914495e-06, 'learning_rate/action_model': 2.4975870249914497e-05, 'epoch': 1.44}) +[2026-06-16 02:00:42,228][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.15232957899570465, 'train/grad_norm_pre_clip': 11.181512832641602, 'timing/data': 0.0003808409674093127, 'timing/model': 1.1258327080286108, 'learning_rate/qwen_vl_interface': 2.4951055602754664e-06, 'learning_rate/action_model': 2.4951055602754665e-05, 'epoch': 1.44}) +[2026-06-16 02:00:46,746][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.020847991108894348, 'train/grad_norm_pre_clip': 8.870062828063965, 'timing/data': 0.0003673169994726777, 'timing/model': 1.2887697339756414, 'learning_rate/qwen_vl_interface': 2.4926252279925255e-06, 'learning_rate/action_model': 2.4926252279925254e-05, 'epoch': 1.44}) +[2026-06-16 02:00:51,106][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.01810906082391739, 'train/grad_norm_pre_clip': 5.858626842498779, 'timing/data': 0.0003843819722533226, 'timing/model': 1.1273872579913586, 'learning_rate/qwen_vl_interface': 2.490146029162199e-06, 'learning_rate/action_model': 2.4901460291621988e-05, 'epoch': 1.44}) +[2026-06-16 02:00:55,624][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.004571151454001665, 'train/grad_norm_pre_clip': 0.6191578507423401, 'timing/data': 0.00036229396937415004, 'timing/model': 1.126320953015238, 'learning_rate/qwen_vl_interface': 2.4876679648035905e-06, 'learning_rate/action_model': 2.4876679648035903e-05, 'epoch': 1.45}) +[2026-06-16 02:01:00,060][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.0041809240356087685, 'train/grad_norm_pre_clip': 1.5343997478485107, 'timing/data': 0.00034416397102177143, 'timing/model': 1.1326377709629014, 'learning_rate/qwen_vl_interface': 2.4851910359353392e-06, 'learning_rate/action_model': 2.485191035935339e-05, 'epoch': 1.45}) +[2026-06-16 02:01:04,490][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.01387734990566969, 'train/grad_norm_pre_clip': 2.936809778213501, 'timing/data': 0.00036677299067378044, 'timing/model': 1.129218728048727, 'learning_rate/qwen_vl_interface': 2.4827152435756185e-06, 'learning_rate/action_model': 2.4827152435756185e-05, 'epoch': 1.45}) +[2026-06-16 02:01:08,968][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.052468039095401764, 'train/grad_norm_pre_clip': 13.020120620727539, 'timing/data': 0.0005725629744119942, 'timing/model': 1.125883337983396, 'learning_rate/qwen_vl_interface': 2.4802405887421307e-06, 'learning_rate/action_model': 2.4802405887421306e-05, 'epoch': 1.45}) +[2026-06-16 02:01:13,477][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.009628635831177235, 'train/grad_norm_pre_clip': 1.776405930519104, 'timing/data': 0.00041272700764238834, 'timing/model': 1.2772450000047684, 'learning_rate/qwen_vl_interface': 2.4777670724521134e-06, 'learning_rate/action_model': 2.4777670724521134e-05, 'epoch': 1.45}) +[2026-06-16 02:01:17,842][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.16901905834674835, 'train/grad_norm_pre_clip': 12.151973724365234, 'timing/data': 0.0003418619744479656, 'timing/model': 1.1278789700008929, 'learning_rate/qwen_vl_interface': 2.475294695722337e-06, 'learning_rate/action_model': 2.475294695722337e-05, 'epoch': 1.45}) +[2026-06-16 02:01:22,376][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.1596619039773941, 'train/grad_norm_pre_clip': 6.899138450622559, 'timing/data': 0.00047124404227361083, 'timing/model': 1.1350336690084077, 'learning_rate/qwen_vl_interface': 2.4728234595691006e-06, 'learning_rate/action_model': 2.4728234595691005e-05, 'epoch': 1.45}) +[2026-06-16 02:01:26,934][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.06558723002672195, 'train/grad_norm_pre_clip': 8.230274200439453, 'timing/data': 0.0003448890056461096, 'timing/model': 1.1458920300356112, 'learning_rate/qwen_vl_interface': 2.4703533650082368e-06, 'learning_rate/action_model': 2.4703533650082366e-05, 'epoch': 1.45}) +[2026-06-16 02:01:31,465][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.027411798015236855, 'train/grad_norm_pre_clip': 13.363122940063477, 'timing/data': 0.00037479400634765625, 'timing/model': 1.1378792489995249, 'learning_rate/qwen_vl_interface': 2.4678844130551103e-06, 'learning_rate/action_model': 2.4678844130551103e-05, 'epoch': 1.45}) +[2026-06-16 02:01:35,984][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.004707423038780689, 'train/grad_norm_pre_clip': 0.5283797383308411, 'timing/data': 0.00034290103940293193, 'timing/model': 1.2816063239588402, 'learning_rate/qwen_vl_interface': 2.4654166047246102e-06, 'learning_rate/action_model': 2.46541660472461e-05, 'epoch': 1.45}) +[2026-06-16 02:01:40,454][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.08565371483564377, 'train/grad_norm_pre_clip': 5.808585166931152, 'timing/data': 0.0005615020054392517, 'timing/model': 1.230996414029505, 'learning_rate/qwen_vl_interface': 2.4629499410311606e-06, 'learning_rate/action_model': 2.462949941031161e-05, 'epoch': 1.45}) +[2026-06-16 02:01:44,826][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.08268686383962631, 'train/grad_norm_pre_clip': 8.20333480834961, 'timing/data': 0.0003652130253612995, 'timing/model': 1.1302034980035387, 'learning_rate/qwen_vl_interface': 2.460484422988715e-06, 'learning_rate/action_model': 2.4604844229887146e-05, 'epoch': 1.45}) +[2026-06-16 02:01:49,357][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.12725357711315155, 'train/grad_norm_pre_clip': 14.492777824401855, 'timing/data': 0.0004715180257335305, 'timing/model': 1.1312434960273094, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 1.45}) +[2026-06-16 02:01:53,891][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.12404245883226395, 'train/grad_norm_pre_clip': 11.008495330810547, 'timing/data': 0.0007715540123172104, 'timing/model': 1.1278734860243276, 'learning_rate/qwen_vl_interface': 2.455556827910286e-06, 'learning_rate/action_model': 2.455556827910286e-05, 'epoch': 1.45}) +[2026-06-16 02:01:58,429][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.0047550811432302, 'train/grad_norm_pre_clip': 0.5246677398681641, 'timing/data': 0.0004010029952041805, 'timing/model': 1.1336699669482186, 'learning_rate/qwen_vl_interface': 2.453094752899853e-06, 'learning_rate/action_model': 2.453094752899853e-05, 'epoch': 1.45}) +[2026-06-16 02:02:02,953][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.005111371632665396, 'train/grad_norm_pre_clip': 0.7572348713874817, 'timing/data': 0.0004292429657652974, 'timing/model': 1.288237727014348, 'learning_rate/qwen_vl_interface': 2.4506338275915173e-06, 'learning_rate/action_model': 2.4506338275915172e-05, 'epoch': 1.45}) +[2026-06-16 02:02:07,378][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.04981404170393944, 'train/grad_norm_pre_clip': 12.21363353729248, 'timing/data': 0.000421595002990216, 'timing/model': 1.1885445090010762, 'learning_rate/qwen_vl_interface': 2.4481740529968735e-06, 'learning_rate/action_model': 2.4481740529968734e-05, 'epoch': 1.45}) +[2026-06-16 02:02:11,798][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.02613488957285881, 'train/grad_norm_pre_clip': 5.141311168670654, 'timing/data': 0.00040208298014476895, 'timing/model': 1.1345297719817609, 'learning_rate/qwen_vl_interface': 2.4457154301270427e-06, 'learning_rate/action_model': 2.4457154301270427e-05, 'epoch': 1.45}) +[2026-06-16 02:02:16,283][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.21216100454330444, 'train/grad_norm_pre_clip': 25.77328872680664, 'timing/data': 0.00034460797905921936, 'timing/model': 1.142679657030385, 'learning_rate/qwen_vl_interface': 2.443257959992671e-06, 'learning_rate/action_model': 2.443257959992671e-05, 'epoch': 1.45}) +[2026-06-16 02:02:20,794][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.011475996114313602, 'train/grad_norm_pre_clip': 2.3534634113311768, 'timing/data': 0.0003812870127148926, 'timing/model': 1.1409412670182064, 'learning_rate/qwen_vl_interface': 2.440801643603933e-06, 'learning_rate/action_model': 2.440801643603933e-05, 'epoch': 1.45}) +[2026-06-16 02:02:25,329][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.015710825100541115, 'train/grad_norm_pre_clip': 1.902368426322937, 'timing/data': 0.0005710679688490927, 'timing/model': 1.129394406976644, 'learning_rate/qwen_vl_interface': 2.4383464819705252e-06, 'learning_rate/action_model': 2.438346481970525e-05, 'epoch': 1.45}) +[2026-06-16 02:02:29,862][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.0033900635316967964, 'train/grad_norm_pre_clip': 0.5447105765342712, 'timing/data': 0.0003355429507791996, 'timing/model': 1.2891944600269198, 'learning_rate/qwen_vl_interface': 2.435892476101674e-06, 'learning_rate/action_model': 2.435892476101674e-05, 'epoch': 1.45}) +[2026-06-16 02:02:34,237][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.013019475154578686, 'train/grad_norm_pre_clip': 2.1825597286224365, 'timing/data': 0.0003407869953662157, 'timing/model': 1.132900252006948, 'learning_rate/qwen_vl_interface': 2.433439627006128e-06, 'learning_rate/action_model': 2.4334396270061277e-05, 'epoch': 1.45}) +[2026-06-16 02:02:38,776][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.009909245185554028, 'train/grad_norm_pre_clip': 2.793177604675293, 'timing/data': 0.0006125550135038793, 'timing/model': 1.1304797159973532, 'learning_rate/qwen_vl_interface': 2.4309879356921602e-06, 'learning_rate/action_model': 2.43098793569216e-05, 'epoch': 1.45}) +[2026-06-16 02:02:43,304][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.0028187218122184277, 'train/grad_norm_pre_clip': 0.6731618642807007, 'timing/data': 0.00035976595245301723, 'timing/model': 1.1324472969863564, 'learning_rate/qwen_vl_interface': 2.428537403167569e-06, 'learning_rate/action_model': 2.4285374031675687e-05, 'epoch': 1.46}) +[2026-06-16 02:02:47,834][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.0162239708006382, 'train/grad_norm_pre_clip': 3.5898122787475586, 'timing/data': 0.00038606702582910657, 'timing/model': 1.1321732160286047, 'learning_rate/qwen_vl_interface': 2.4260880304396765e-06, 'learning_rate/action_model': 2.4260880304396765e-05, 'epoch': 1.46}) +[2026-06-16 02:02:52,382][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.0028645277488976717, 'train/grad_norm_pre_clip': 0.4943859279155731, 'timing/data': 0.0003737689694389701, 'timing/model': 1.129247902019415, 'learning_rate/qwen_vl_interface': 2.4236398185153245e-06, 'learning_rate/action_model': 2.4236398185153245e-05, 'epoch': 1.46}) +[2026-06-16 02:02:56,938][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.21329469978809357, 'train/grad_norm_pre_clip': 14.256353378295898, 'timing/data': 0.00031934300204738975, 'timing/model': 1.3004256560234353, 'learning_rate/qwen_vl_interface': 2.4211927684008823e-06, 'learning_rate/action_model': 2.4211927684008824e-05, 'epoch': 1.46}) +[2026-06-16 02:03:01,326][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.0013528153067454696, 'train/grad_norm_pre_clip': 0.2971259355545044, 'timing/data': 0.00038424599915742874, 'timing/model': 1.132255450997036, 'learning_rate/qwen_vl_interface': 2.4187468811022398e-06, 'learning_rate/action_model': 2.4187468811022396e-05, 'epoch': 1.46}) +[2026-06-16 02:03:05,857][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.008064872585237026, 'train/grad_norm_pre_clip': 2.7240676879882812, 'timing/data': 0.00036503502633422613, 'timing/model': 1.1300162189872935, 'learning_rate/qwen_vl_interface': 2.416302157624809e-06, 'learning_rate/action_model': 2.4163021576248088e-05, 'epoch': 1.46}) +[2026-06-16 02:03:10,345][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.15370452404022217, 'train/grad_norm_pre_clip': 26.782148361206055, 'timing/data': 0.00035777996527031064, 'timing/model': 1.1348392720101401, 'learning_rate/qwen_vl_interface': 2.4138585989735227e-06, 'learning_rate/action_model': 2.4138585989735226e-05, 'epoch': 1.46}) +[2026-06-16 02:03:14,879][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.09717295318841934, 'train/grad_norm_pre_clip': 19.033952713012695, 'timing/data': 0.0003215129836462438, 'timing/model': 1.132292332011275, 'learning_rate/qwen_vl_interface': 2.4114162061528357e-06, 'learning_rate/action_model': 2.4114162061528354e-05, 'epoch': 1.46}) +[2026-06-16 02:03:19,438][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.016810860484838486, 'train/grad_norm_pre_clip': 7.372742176055908, 'timing/data': 0.0003388129989616573, 'timing/model': 1.1477386559708975, 'learning_rate/qwen_vl_interface': 2.408974980166724e-06, 'learning_rate/action_model': 2.408974980166724e-05, 'epoch': 1.46}) +[2026-06-16 02:03:23,974][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.0013916633324697614, 'train/grad_norm_pre_clip': 0.25797757506370544, 'timing/data': 0.00031916802981868386, 'timing/model': 1.2949703999911435, 'learning_rate/qwen_vl_interface': 2.406534922018683e-06, 'learning_rate/action_model': 2.4065349220186827e-05, 'epoch': 1.46}) +[2026-06-16 02:03:28,440][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.004627774935215712, 'train/grad_norm_pre_clip': 0.8933271169662476, 'timing/data': 0.0004913410521112382, 'timing/model': 1.2217971610371023, 'learning_rate/qwen_vl_interface': 2.4040960327117287e-06, 'learning_rate/action_model': 2.4040960327117283e-05, 'epoch': 1.46}) +[2026-06-16 02:03:32,823][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.001610148698091507, 'train/grad_norm_pre_clip': 0.26227712631225586, 'timing/data': 0.0003737059887498617, 'timing/model': 1.1378212010022253, 'learning_rate/qwen_vl_interface': 2.4016583132483976e-06, 'learning_rate/action_model': 2.4016583132483976e-05, 'epoch': 1.46}) +[2026-06-16 02:03:37,342][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.007413636893033981, 'train/grad_norm_pre_clip': 2.9332969188690186, 'timing/data': 0.00035425496753305197, 'timing/model': 1.134292267030105, 'learning_rate/qwen_vl_interface': 2.399221764630741e-06, 'learning_rate/action_model': 2.3992217646307406e-05, 'epoch': 1.46}) +[2026-06-16 02:03:41,940][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.06334159523248672, 'train/grad_norm_pre_clip': 14.976621627807617, 'timing/data': 0.00037308596074581146, 'timing/model': 1.1312190419994295, 'learning_rate/qwen_vl_interface': 2.3967863878603334e-06, 'learning_rate/action_model': 2.3967863878603332e-05, 'epoch': 1.46}) +[2026-06-16 02:03:46,478][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.005095114000141621, 'train/grad_norm_pre_clip': 1.7676928043365479, 'timing/data': 0.00035951397148892283, 'timing/model': 1.1351698269718327, 'learning_rate/qwen_vl_interface': 2.394352183938266e-06, 'learning_rate/action_model': 2.3943521839382664e-05, 'epoch': 1.46}) +[2026-06-16 02:03:51,017][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.01679467409849167, 'train/grad_norm_pre_clip': 8.616100311279297, 'timing/data': 0.00036228098906576633, 'timing/model': 1.285620896029286, 'learning_rate/qwen_vl_interface': 2.39191915386515e-06, 'learning_rate/action_model': 2.39191915386515e-05, 'epoch': 1.46}) +[2026-06-16 02:03:55,443][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.107426717877388, 'train/grad_norm_pre_clip': 13.911920547485352, 'timing/data': 0.00034050398971885443, 'timing/model': 1.1857949530240148, 'learning_rate/qwen_vl_interface': 2.3894872986411095e-06, 'learning_rate/action_model': 2.3894872986411094e-05, 'epoch': 1.46}) +[2026-06-16 02:03:59,845][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.32580822706222534, 'train/grad_norm_pre_clip': 27.313749313354492, 'timing/data': 0.0004175269859842956, 'timing/model': 1.135896057006903, 'learning_rate/qwen_vl_interface': 2.3870566192657903e-06, 'learning_rate/action_model': 2.38705661926579e-05, 'epoch': 1.46}) +[2026-06-16 02:04:04,315][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.00457591051235795, 'train/grad_norm_pre_clip': 1.2721459865570068, 'timing/data': 0.00035398598993197083, 'timing/model': 1.139349223987665, 'learning_rate/qwen_vl_interface': 2.3846271167383515e-06, 'learning_rate/action_model': 2.3846271167383515e-05, 'epoch': 1.46}) +[2026-06-16 02:04:08,826][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.18033987283706665, 'train/grad_norm_pre_clip': 12.971122741699219, 'timing/data': 0.00031939498148858547, 'timing/model': 1.1414063469856046, 'learning_rate/qwen_vl_interface': 2.382198792057471e-06, 'learning_rate/action_model': 2.3821987920574707e-05, 'epoch': 1.46}) +[2026-06-16 02:04:13,398][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.04503413662314415, 'train/grad_norm_pre_clip': 30.85732078552246, 'timing/data': 0.0003616319736465812, 'timing/model': 1.1334608100005426, 'learning_rate/qwen_vl_interface': 2.3797716462213402e-06, 'learning_rate/action_model': 2.3797716462213402e-05, 'epoch': 1.46}) +[2026-06-16 02:04:17,945][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.003280460136011243, 'train/grad_norm_pre_clip': 0.6211713552474976, 'timing/data': 0.0004145199782215059, 'timing/model': 1.298431817034725, 'learning_rate/qwen_vl_interface': 2.37734568022767e-06, 'learning_rate/action_model': 2.3773456802276697e-05, 'epoch': 1.46}) +[2026-06-16 02:04:22,324][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.1688155084848404, 'train/grad_norm_pre_clip': 20.40340232849121, 'timing/data': 0.00038067600689828396, 'timing/model': 1.1323871790082194, 'learning_rate/qwen_vl_interface': 2.3749208950736783e-06, 'learning_rate/action_model': 2.374920895073678e-05, 'epoch': 1.46}) +[2026-06-16 02:04:26,874][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.12386517971754074, 'train/grad_norm_pre_clip': 9.844237327575684, 'timing/data': 0.00035685603506863117, 'timing/model': 1.135236811998766, 'learning_rate/qwen_vl_interface': 2.372497291756106e-06, 'learning_rate/action_model': 2.3724972917561058e-05, 'epoch': 1.46}) +[2026-06-16 02:04:31,416][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.05742708593606949, 'train/grad_norm_pre_clip': 24.963850021362305, 'timing/data': 0.0003473609685897827, 'timing/model': 1.1309854400460608, 'learning_rate/qwen_vl_interface': 2.3700748712712043e-06, 'learning_rate/action_model': 2.370074871271204e-05, 'epoch': 1.46}) +[2026-06-16 02:04:35,960][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.0022090368438512087, 'train/grad_norm_pre_clip': 0.5060045719146729, 'timing/data': 0.00035135599318891764, 'timing/model': 1.1334320969763212, 'learning_rate/qwen_vl_interface': 2.3676536346147388e-06, 'learning_rate/action_model': 2.3676536346147388e-05, 'epoch': 1.47}) +[2026-06-16 02:04:40,506][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.2165004462003708, 'train/grad_norm_pre_clip': 9.62542724609375, 'timing/data': 0.00032786402152851224, 'timing/model': 1.129029718984384, 'learning_rate/qwen_vl_interface': 2.365233582781988e-06, 'learning_rate/action_model': 2.365233582781988e-05, 'epoch': 1.47}) +[2026-06-16 02:04:45,048][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.0009679339709691703, 'train/grad_norm_pre_clip': 0.12808077037334442, 'timing/data': 0.00038412801222875714, 'timing/model': 1.2996601989725605, 'learning_rate/qwen_vl_interface': 2.3628147167677444e-06, 'learning_rate/action_model': 2.3628147167677443e-05, 'epoch': 1.47}) +[2026-06-16 02:04:49,439][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.11189143359661102, 'train/grad_norm_pre_clip': 6.548097133636475, 'timing/data': 0.000347008986864239, 'timing/model': 1.1334772470290773, 'learning_rate/qwen_vl_interface': 2.3603970375663123e-06, 'learning_rate/action_model': 2.3603970375663125e-05, 'epoch': 1.47}) +[2026-06-16 02:04:53,982][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.0031860133167356253, 'train/grad_norm_pre_clip': 0.5086032748222351, 'timing/data': 0.00038212904473766685, 'timing/model': 1.1349505460239016, 'learning_rate/qwen_vl_interface': 2.35798054617151e-06, 'learning_rate/action_model': 2.3579805461715096e-05, 'epoch': 1.47}) +[2026-06-16 02:04:58,493][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.039491940289735794, 'train/grad_norm_pre_clip': 20.26984405517578, 'timing/data': 0.00035204197047278285, 'timing/model': 1.1317891249782406, 'learning_rate/qwen_vl_interface': 2.355565243576662e-06, 'learning_rate/action_model': 2.3555652435766617e-05, 'epoch': 1.47}) +[2026-06-16 02:05:03,047][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.1449078768491745, 'train/grad_norm_pre_clip': 18.051971435546875, 'timing/data': 0.0003258420038037002, 'timing/model': 1.134223388042301, 'learning_rate/qwen_vl_interface': 2.35315113077461e-06, 'learning_rate/action_model': 2.3531511307746098e-05, 'epoch': 1.47}) +[2026-06-16 02:05:07,582][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.0643630102276802, 'train/grad_norm_pre_clip': 17.565515518188477, 'timing/data': 0.00040778302354738116, 'timing/model': 1.1368491350440308, 'learning_rate/qwen_vl_interface': 2.3507382087577047e-06, 'learning_rate/action_model': 2.350738208757705e-05, 'epoch': 1.47}) +[2026-06-16 02:05:12,116][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.10412249714136124, 'train/grad_norm_pre_clip': 20.74383544921875, 'timing/data': 0.0003455020487308502, 'timing/model': 1.291017171985004, 'learning_rate/qwen_vl_interface': 2.348326478517807e-06, 'learning_rate/action_model': 2.3483264785178072e-05, 'epoch': 1.47}) +[2026-06-16 02:05:16,593][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.025759581476449966, 'train/grad_norm_pre_clip': 17.51234245300293, 'timing/data': 0.0003289220039732754, 'timing/model': 1.2257838570512831, 'learning_rate/qwen_vl_interface': 2.3459159410462907e-06, 'learning_rate/action_model': 2.3459159410462905e-05, 'epoch': 1.47}) +[2026-06-16 02:05:20,966][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.027422158047556877, 'train/grad_norm_pre_clip': 10.030682563781738, 'timing/data': 0.00033534999238327146, 'timing/model': 1.13225181796588, 'learning_rate/qwen_vl_interface': 2.343506597334031e-06, 'learning_rate/action_model': 2.343506597334031e-05, 'epoch': 1.47}) +[2026-06-16 02:05:25,470][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.05222185328602791, 'train/grad_norm_pre_clip': 21.15874671936035, 'timing/data': 0.0003523330087773502, 'timing/model': 1.1333898910088465, 'learning_rate/qwen_vl_interface': 2.341098448371423e-06, 'learning_rate/action_model': 2.3410984483714227e-05, 'epoch': 1.47}) +[2026-06-16 02:05:30,021][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.0007493399316444993, 'train/grad_norm_pre_clip': 0.16156603395938873, 'timing/data': 0.0003420530119910836, 'timing/model': 1.1305444100289606, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 1.47}) +[2026-06-16 02:05:34,558][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.22189293801784515, 'train/grad_norm_pre_clip': 27.425880432128906, 'timing/data': 0.000359151978045702, 'timing/model': 1.1352855010190979, 'learning_rate/qwen_vl_interface': 2.33628573865426e-06, 'learning_rate/action_model': 2.3362857386542596e-05, 'epoch': 1.47}) +[2026-06-16 02:05:39,093][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.0054472461342811584, 'train/grad_norm_pre_clip': 0.9433526396751404, 'timing/data': 0.00037367199547588825, 'timing/model': 1.288815625011921, 'learning_rate/qwen_vl_interface': 2.3338811798780282e-06, 'learning_rate/action_model': 2.333881179878028e-05, 'epoch': 1.47}) +[2026-06-16 02:05:43,551][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.008188659325242043, 'train/grad_norm_pre_clip': 3.469552755355835, 'timing/data': 0.0003149389522150159, 'timing/model': 1.2056409770157188, 'learning_rate/qwen_vl_interface': 2.3314778198080933e-06, 'learning_rate/action_model': 2.331477819808093e-05, 'epoch': 1.47}) +[2026-06-16 02:05:47,956][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.0017770774429664016, 'train/grad_norm_pre_clip': 0.7749367952346802, 'timing/data': 0.0003113329876214266, 'timing/model': 1.1407190529862419, 'learning_rate/qwen_vl_interface': 2.3290756594323834e-06, 'learning_rate/action_model': 2.3290756594323836e-05, 'epoch': 1.47}) +[2026-06-16 02:05:52,424][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.01413660403341055, 'train/grad_norm_pre_clip': 5.712730884552002, 'timing/data': 0.0005479030078276992, 'timing/model': 1.1359858760260977, 'learning_rate/qwen_vl_interface': 2.3266746997383378e-06, 'learning_rate/action_model': 2.3266746997383377e-05, 'epoch': 1.47}) +[2026-06-16 02:05:56,935][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.09139591455459595, 'train/grad_norm_pre_clip': 16.068618774414062, 'timing/data': 0.00042750901775434613, 'timing/model': 1.1327891469700262, 'learning_rate/qwen_vl_interface': 2.324274941712899e-06, 'learning_rate/action_model': 2.324274941712899e-05, 'epoch': 1.47}) +[2026-06-16 02:06:01,471][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.42171379923820496, 'train/grad_norm_pre_clip': 32.899784088134766, 'timing/data': 0.00035464001120999455, 'timing/model': 1.132686167024076, 'learning_rate/qwen_vl_interface': 2.321876386342519e-06, 'learning_rate/action_model': 2.3218763863425187e-05, 'epoch': 1.47}) +[2026-06-16 02:06:06,009][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.0021016611717641354, 'train/grad_norm_pre_clip': 0.531874418258667, 'timing/data': 0.0006804990116506815, 'timing/model': 1.2936918430496007, 'learning_rate/qwen_vl_interface': 2.31947903461315e-06, 'learning_rate/action_model': 2.3194790346131502e-05, 'epoch': 1.47}) +[2026-06-16 02:06:10,394][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.0004451912536751479, 'train/grad_norm_pre_clip': 0.09532104432582855, 'timing/data': 0.00031188700813800097, 'timing/model': 1.139276301022619, 'learning_rate/qwen_vl_interface': 2.3170828875102544e-06, 'learning_rate/action_model': 2.3170828875102543e-05, 'epoch': 1.47}) +[2026-06-16 02:06:14,926][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.008460947312414646, 'train/grad_norm_pre_clip': 7.7514777183532715, 'timing/data': 0.00033348402939736843, 'timing/model': 1.1328407680266537, 'learning_rate/qwen_vl_interface': 2.3146879460187984e-06, 'learning_rate/action_model': 2.3146879460187982e-05, 'epoch': 1.47}) +[2026-06-16 02:06:19,470][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.10793768614530563, 'train/grad_norm_pre_clip': 13.528239250183105, 'timing/data': 0.00033257598988711834, 'timing/model': 1.1411192149971612, 'learning_rate/qwen_vl_interface': 2.3122942111232506e-06, 'learning_rate/action_model': 2.3122942111232506e-05, 'epoch': 1.47}) +[2026-06-16 02:06:24,002][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.2489083856344223, 'train/grad_norm_pre_clip': 29.797412872314453, 'timing/data': 0.00036194699350744486, 'timing/model': 1.1352075549657457, 'learning_rate/qwen_vl_interface': 2.3099016838075857e-06, 'learning_rate/action_model': 2.3099016838075856e-05, 'epoch': 1.48}) +[2026-06-16 02:06:28,538][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.03202242776751518, 'train/grad_norm_pre_clip': 12.897021293640137, 'timing/data': 0.0003358870162628591, 'timing/model': 1.1357113699777983, 'learning_rate/qwen_vl_interface': 2.307510365055282e-06, 'learning_rate/action_model': 2.307510365055282e-05, 'epoch': 1.48}) +[2026-06-16 02:06:33,090][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.006308373995125294, 'train/grad_norm_pre_clip': 3.26826810836792, 'timing/data': 0.0003859309945255518, 'timing/model': 1.2880408319761045, 'learning_rate/qwen_vl_interface': 2.3051202558493196e-06, 'learning_rate/action_model': 2.3051202558493192e-05, 'epoch': 1.48}) +[2026-06-16 02:06:37,479][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.09207164496183395, 'train/grad_norm_pre_clip': 13.382597923278809, 'timing/data': 0.00038733897963538766, 'timing/model': 1.1317715239711106, 'learning_rate/qwen_vl_interface': 2.302731357172182e-06, 'learning_rate/action_model': 2.302731357172182e-05, 'epoch': 1.48}) +[2026-06-16 02:06:42,027][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.10880782455205917, 'train/grad_norm_pre_clip': 12.349935531616211, 'timing/data': 0.00037036201683804393, 'timing/model': 1.1338785190018825, 'learning_rate/qwen_vl_interface': 2.3003436700058556e-06, 'learning_rate/action_model': 2.3003436700058554e-05, 'epoch': 1.48}) +[2026-06-16 02:06:46,531][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.04998677596449852, 'train/grad_norm_pre_clip': 12.726367950439453, 'timing/data': 0.0003862339654006064, 'timing/model': 1.133334042970091, 'learning_rate/qwen_vl_interface': 2.297957195331828e-06, 'learning_rate/action_model': 2.2979571953318276e-05, 'epoch': 1.48}) +[2026-06-16 02:06:51,095][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.003435122314840555, 'train/grad_norm_pre_clip': 0.5393369793891907, 'timing/data': 0.00033484300365671515, 'timing/model': 1.1345724189886823, 'learning_rate/qwen_vl_interface': 2.2955719341310903e-06, 'learning_rate/action_model': 2.2955719341310903e-05, 'epoch': 1.48}) +[2026-06-16 02:06:55,639][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.14035995304584503, 'train/grad_norm_pre_clip': 33.411949157714844, 'timing/data': 0.000345399952493608, 'timing/model': 1.1338274820009246, 'learning_rate/qwen_vl_interface': 2.2931878873841304e-06, 'learning_rate/action_model': 2.2931878873841302e-05, 'epoch': 1.48}) +[2026-06-16 02:07:00,195][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.0039846329018473625, 'train/grad_norm_pre_clip': 0.9847497940063477, 'timing/data': 0.00035899796057492495, 'timing/model': 1.2919132959796116, 'learning_rate/qwen_vl_interface': 2.290805056070942e-06, 'learning_rate/action_model': 2.290805056070942e-05, 'epoch': 1.48}) +[2026-06-16 02:11:23,983][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.10550859570503235, 'train/grad_norm_pre_clip': 8.653693199157715, 'eval/loss': 0.0402284599840641, 'eval/flap_precision': 0.9740061162079511, 'eval/flap_recall': 0.9830246913580247, 'eval/flap_f1': 0.978494623655914, 'eval/flap_rate_model': 0.18166666666666667, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9784946236559141, 'timing/data': 0.00034584803506731987, 'timing/model': 1.2315622989553958, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 1.48}) +[2026-06-16 02:11:23,983][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3600_state +[2026-06-16 02:11:23,985][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:11:27,564][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3600_state/model.safetensors +[2026-06-16 02:11:37,499][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3600_state/optimizer.bin +[2026-06-16 02:11:37,500][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3600_state/sampler.bin +[2026-06-16 02:11:37,500][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3600_state/sampler_1.bin +[2026-06-16 02:11:37,501][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3600_state/random_states_0.pkl +[2026-06-16 02:11:37,548][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3600 +[2026-06-16 02:11:37,548][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:11:37,554][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:11:41,992][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.0021571656689047813, 'train/grad_norm_pre_clip': 0.4296712577342987, 'timing/data': 0.0003568860120140016, 'timing/model': 1.2701892260229215, 'learning_rate/qwen_vl_interface': 2.2860430436633447e-06, 'learning_rate/action_model': 2.2860430436633447e-05, 'epoch': 1.48}) +[2026-06-16 02:11:46,301][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.0646936371922493, 'train/grad_norm_pre_clip': 23.815942764282227, 'timing/data': 0.00034016004065051675, 'timing/model': 1.1190832289867103, 'learning_rate/qwen_vl_interface': 2.28366386452642e-06, 'learning_rate/action_model': 2.28366386452642e-05, 'epoch': 1.48}) +[2026-06-16 02:11:50,752][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.022796157747507095, 'train/grad_norm_pre_clip': 9.379807472229004, 'timing/data': 0.00031879497691988945, 'timing/model': 1.1136313540046103, 'learning_rate/qwen_vl_interface': 2.2812859047382326e-06, 'learning_rate/action_model': 2.2812859047382323e-05, 'epoch': 1.48}) +[2026-06-16 02:11:55,176][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.19786348938941956, 'train/grad_norm_pre_clip': 19.737638473510742, 'timing/data': 0.0003343310090713203, 'timing/model': 1.1174232229823247, 'learning_rate/qwen_vl_interface': 2.2789091652762712e-06, 'learning_rate/action_model': 2.2789091652762712e-05, 'epoch': 1.48}) +[2026-06-16 02:11:59,584][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.2618553042411804, 'train/grad_norm_pre_clip': 25.33757972717285, 'timing/data': 0.0003114709979854524, 'timing/model': 1.1196278689894825, 'learning_rate/qwen_vl_interface': 2.276533647117524e-06, 'learning_rate/action_model': 2.2765336471175242e-05, 'epoch': 1.48}) +[2026-06-16 02:12:04,047][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.22544579207897186, 'train/grad_norm_pre_clip': 40.132415771484375, 'timing/data': 0.00035708199720829725, 'timing/model': 1.1239663640153594, 'learning_rate/qwen_vl_interface': 2.2741593512384773e-06, 'learning_rate/action_model': 2.2741593512384772e-05, 'epoch': 1.48}) +[2026-06-16 02:12:08,539][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.018844973295927048, 'train/grad_norm_pre_clip': 11.113797187805176, 'timing/data': 0.0005455259815789759, 'timing/model': 1.2814991769846529, 'learning_rate/qwen_vl_interface': 2.2717862786151153e-06, 'learning_rate/action_model': 2.271786278615115e-05, 'epoch': 1.48}) +[2026-06-16 02:12:12,884][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.24879251420497894, 'train/grad_norm_pre_clip': 9.709539413452148, 'timing/data': 0.000358782010152936, 'timing/model': 1.1249943880247883, 'learning_rate/qwen_vl_interface': 2.2694144302229152e-06, 'learning_rate/action_model': 2.2694144302229152e-05, 'epoch': 1.48}) +[2026-06-16 02:12:17,386][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.0011765911476686597, 'train/grad_norm_pre_clip': 0.20315949618816376, 'timing/data': 0.0003489519585855305, 'timing/model': 1.1313817190239206, 'learning_rate/qwen_vl_interface': 2.2670438070368568e-06, 'learning_rate/action_model': 2.2670438070368566e-05, 'epoch': 1.48}) +[2026-06-16 02:12:21,898][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.11320770531892776, 'train/grad_norm_pre_clip': 17.373910903930664, 'timing/data': 0.0003393269726075232, 'timing/model': 1.124289650004357, 'learning_rate/qwen_vl_interface': 2.264674410031413e-06, 'learning_rate/action_model': 2.2646744100314127e-05, 'epoch': 1.48}) +[2026-06-16 02:12:26,380][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.0022568674758076668, 'train/grad_norm_pre_clip': 0.39741480350494385, 'timing/data': 0.0006143649807199836, 'timing/model': 1.1257112890016288, 'learning_rate/qwen_vl_interface': 2.2623062401805537e-06, 'learning_rate/action_model': 2.2623062401805535e-05, 'epoch': 1.48}) +[2026-06-16 02:12:30,884][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.0019861285109072924, 'train/grad_norm_pre_clip': 0.4915318489074707, 'timing/data': 0.0004016470047645271, 'timing/model': 1.1261422900133766, 'learning_rate/qwen_vl_interface': 2.2599392984577436e-06, 'learning_rate/action_model': 2.2599392984577438e-05, 'epoch': 1.48}) +[2026-06-16 02:12:35,393][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.0024657377507537603, 'train/grad_norm_pre_clip': 0.3115028440952301, 'timing/data': 0.00040368898771703243, 'timing/model': 1.28985936002573, 'learning_rate/qwen_vl_interface': 2.2575735858359436e-06, 'learning_rate/action_model': 2.257573585835944e-05, 'epoch': 1.48}) +[2026-06-16 02:12:39,753][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.0077183968387544155, 'train/grad_norm_pre_clip': 2.1316540241241455, 'timing/data': 0.00034528400283306837, 'timing/model': 1.1275458159507252, 'learning_rate/qwen_vl_interface': 2.255209103287609e-06, 'learning_rate/action_model': 2.255209103287609e-05, 'epoch': 1.48}) +[2026-06-16 02:12:44,265][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.0027043374720960855, 'train/grad_norm_pre_clip': 0.4329174757003784, 'timing/data': 0.00032105896389111876, 'timing/model': 1.1271494510001503, 'learning_rate/qwen_vl_interface': 2.2528458517846886e-06, 'learning_rate/action_model': 2.252845851784689e-05, 'epoch': 1.49}) +[2026-06-16 02:12:48,757][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.08318805694580078, 'train/grad_norm_pre_clip': 7.112540245056152, 'timing/data': 0.00035610300255939364, 'timing/model': 1.130461769993417, 'learning_rate/qwen_vl_interface': 2.250483832298627e-06, 'learning_rate/action_model': 2.250483832298627e-05, 'epoch': 1.49}) +[2026-06-16 02:12:53,225][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.07652337849140167, 'train/grad_norm_pre_clip': 15.79967975616455, 'timing/data': 0.0006103469640947878, 'timing/model': 1.1377335719880648, 'learning_rate/qwen_vl_interface': 2.2481230458003608e-06, 'learning_rate/action_model': 2.2481230458003606e-05, 'epoch': 1.49}) +[2026-06-16 02:12:57,706][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.0032846261747181416, 'train/grad_norm_pre_clip': 1.7868515253067017, 'timing/data': 0.00038011500146239996, 'timing/model': 1.1251043360098265, 'learning_rate/qwen_vl_interface': 2.2457634932603184e-06, 'learning_rate/action_model': 2.2457634932603183e-05, 'epoch': 1.49}) +[2026-06-16 02:13:02,230][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.11728550493717194, 'train/grad_norm_pre_clip': 72.71719360351562, 'timing/data': 0.0003121360205113888, 'timing/model': 1.2818757129716687, 'learning_rate/qwen_vl_interface': 2.2434051756484255e-06, 'learning_rate/action_model': 2.2434051756484253e-05, 'epoch': 1.49}) +[2026-06-16 02:13:06,591][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.10833173245191574, 'train/grad_norm_pre_clip': 11.590139389038086, 'timing/data': 0.00037186499685049057, 'timing/model': 1.1257676649838686, 'learning_rate/qwen_vl_interface': 2.2410480939340945e-06, 'learning_rate/action_model': 2.2410480939340945e-05, 'epoch': 1.49}) +[2026-06-16 02:13:11,115][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.017416682094335556, 'train/grad_norm_pre_clip': 8.585177421569824, 'timing/data': 0.00035300198942422867, 'timing/model': 1.1278935769805685, 'learning_rate/qwen_vl_interface': 2.238692249086234e-06, 'learning_rate/action_model': 2.2386922490862336e-05, 'epoch': 1.49}) +[2026-06-16 02:13:15,637][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.0011076154187321663, 'train/grad_norm_pre_clip': 0.2909388840198517, 'timing/data': 0.00040119094774127007, 'timing/model': 1.1337856550235301, 'learning_rate/qwen_vl_interface': 2.2363376420732426e-06, 'learning_rate/action_model': 2.2363376420732423e-05, 'epoch': 1.49}) +[2026-06-16 02:13:20,148][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.16407836973667145, 'train/grad_norm_pre_clip': 25.22237205505371, 'timing/data': 0.0003639319911599159, 'timing/model': 1.1368728879606351, 'learning_rate/qwen_vl_interface': 2.233984273863011e-06, 'learning_rate/action_model': 2.233984273863011e-05, 'epoch': 1.49}) +[2026-06-16 02:13:24,682][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.002968428423628211, 'train/grad_norm_pre_clip': 0.5207842588424683, 'timing/data': 0.00035152799682691693, 'timing/model': 1.132800139952451, 'learning_rate/qwen_vl_interface': 2.231632145422919e-06, 'learning_rate/action_model': 2.231632145422919e-05, 'epoch': 1.49}) +[2026-06-16 02:13:29,207][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.06674245744943619, 'train/grad_norm_pre_clip': 78.55757141113281, 'timing/data': 0.00034754001535475254, 'timing/model': 1.2926467939978465, 'learning_rate/qwen_vl_interface': 2.2292812577198383e-06, 'learning_rate/action_model': 2.229281257719838e-05, 'epoch': 1.49}) +[2026-06-16 02:13:33,580][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.026337841525673866, 'train/grad_norm_pre_clip': 5.414331436157227, 'timing/data': 0.00035169697366654873, 'timing/model': 1.1367764929891564, 'learning_rate/qwen_vl_interface': 2.2269316117201307e-06, 'learning_rate/action_model': 2.2269316117201307e-05, 'epoch': 1.49}) +[2026-06-16 02:13:38,109][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.07537723332643509, 'train/grad_norm_pre_clip': 15.685084342956543, 'timing/data': 0.0004040870117023587, 'timing/model': 1.131815223954618, 'learning_rate/qwen_vl_interface': 2.224583208389646e-06, 'learning_rate/action_model': 2.224583208389646e-05, 'epoch': 1.49}) +[2026-06-16 02:13:42,593][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.0437091626226902, 'train/grad_norm_pre_clip': 10.755826950073242, 'timing/data': 0.0003343859571032226, 'timing/model': 1.1344095839885995, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 1.49}) +[2026-06-16 02:13:47,035][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.00041749689262360334, 'train/grad_norm_pre_clip': 0.07719697803258896, 'timing/data': 0.0003591200220398605, 'timing/model': 1.1281191569869407, 'learning_rate/qwen_vl_interface': 2.219890133597194e-06, 'learning_rate/action_model': 2.219890133597194e-05, 'epoch': 1.49}) +[2026-06-16 02:13:51,517][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.0012062735622748733, 'train/grad_norm_pre_clip': 0.390688955783844, 'timing/data': 0.00035186298191547394, 'timing/model': 1.1332516969996504, 'learning_rate/qwen_vl_interface': 2.217545464064374e-06, 'learning_rate/action_model': 2.217545464064374e-05, 'epoch': 1.49}) +[2026-06-16 02:13:56,058][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.009179282933473587, 'train/grad_norm_pre_clip': 4.28203821182251, 'timing/data': 0.00033434503711760044, 'timing/model': 1.2987144750077277, 'learning_rate/qwen_vl_interface': 2.215202041059067e-06, 'learning_rate/action_model': 2.215202041059067e-05, 'epoch': 1.49}) +[2026-06-16 02:14:00,434][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.012668821960687637, 'train/grad_norm_pre_clip': 7.111721992492676, 'timing/data': 0.0003249240107834339, 'timing/model': 1.1357623310177587, 'learning_rate/qwen_vl_interface': 2.212859865544566e-06, 'learning_rate/action_model': 2.212859865544566e-05, 'epoch': 1.49}) +[2026-06-16 02:14:04,979][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.02058752067387104, 'train/grad_norm_pre_clip': 4.612029552459717, 'timing/data': 0.0003525629872456193, 'timing/model': 1.1336741860141046, 'learning_rate/qwen_vl_interface': 2.210518938483652e-06, 'learning_rate/action_model': 2.210518938483652e-05, 'epoch': 1.49}) +[2026-06-16 02:14:09,552][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.0013378390576690435, 'train/grad_norm_pre_clip': 0.29223760962486267, 'timing/data': 0.00036124204052612185, 'timing/model': 1.135916626022663, 'learning_rate/qwen_vl_interface': 2.20817926083859e-06, 'learning_rate/action_model': 2.20817926083859e-05, 'epoch': 1.49}) +[2026-06-16 02:14:14,092][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.0008203920442610979, 'train/grad_norm_pre_clip': 0.2118663340806961, 'timing/data': 0.00036285194801166654, 'timing/model': 1.1335998080321588, 'learning_rate/qwen_vl_interface': 2.205840833571135e-06, 'learning_rate/action_model': 2.205840833571135e-05, 'epoch': 1.49}) +[2026-06-16 02:14:18,638][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.1104973703622818, 'train/grad_norm_pre_clip': 6.956053733825684, 'timing/data': 0.0003438569838181138, 'timing/model': 1.145490026974585, 'learning_rate/qwen_vl_interface': 2.2035036576425264e-06, 'learning_rate/action_model': 2.203503657642526e-05, 'epoch': 1.49}) +[2026-06-16 02:14:23,170][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 8.777494804235175e-05, 'train/grad_norm_pre_clip': 0.01921379752457142, 'timing/data': 0.0003506530192680657, 'timing/model': 1.287671099009458, 'learning_rate/qwen_vl_interface': 2.2011677340134867e-06, 'learning_rate/action_model': 2.2011677340134867e-05, 'epoch': 1.49}) +[2026-06-16 02:14:27,533][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.021455779671669006, 'train/grad_norm_pre_clip': 12.529250144958496, 'timing/data': 0.00033267802791669965, 'timing/model': 1.1301409070147201, 'learning_rate/qwen_vl_interface': 2.198833063644227e-06, 'learning_rate/action_model': 2.198833063644227e-05, 'epoch': 1.49}) +[2026-06-16 02:14:32,087][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.01094797533005476, 'train/grad_norm_pre_clip': 3.8850433826446533, 'timing/data': 0.00033663498470559716, 'timing/model': 1.1479030809714459, 'learning_rate/qwen_vl_interface': 2.196499647494443e-06, 'learning_rate/action_model': 2.196499647494443e-05, 'epoch': 1.49}) +[2026-06-16 02:14:36,615][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.005607730709016323, 'train/grad_norm_pre_clip': 1.860315203666687, 'timing/data': 0.00035830301931127906, 'timing/model': 1.1360733889741823, 'learning_rate/qwen_vl_interface': 2.1941674865233144e-06, 'learning_rate/action_model': 2.1941674865233143e-05, 'epoch': 1.5}) +[2026-06-16 02:14:41,083][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.16477122902870178, 'train/grad_norm_pre_clip': 54.5725212097168, 'timing/data': 0.00034825096372514963, 'timing/model': 1.1439216840080917, 'learning_rate/qwen_vl_interface': 2.1918365816895013e-06, 'learning_rate/action_model': 2.191836581689501e-05, 'epoch': 1.5}) +[2026-06-16 02:14:45,586][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.020505258813500404, 'train/grad_norm_pre_clip': 7.294312477111816, 'timing/data': 0.00034214695915579796, 'timing/model': 1.1324196869973093, 'learning_rate/qwen_vl_interface': 2.189506933951153e-06, 'learning_rate/action_model': 2.1895069339511525e-05, 'epoch': 1.5}) +[2026-06-16 02:14:50,134][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.04586998000741005, 'train/grad_norm_pre_clip': 24.432132720947266, 'timing/data': 0.00035253504756838083, 'timing/model': 1.3019455220201053, 'learning_rate/qwen_vl_interface': 2.187178544265898e-06, 'learning_rate/action_model': 2.1871785442658983e-05, 'epoch': 1.5}) +[2026-06-16 02:14:54,527][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.41702866554260254, 'train/grad_norm_pre_clip': 27.48222541809082, 'timing/data': 0.00031061703339219093, 'timing/model': 1.1344312730361708, 'learning_rate/qwen_vl_interface': 2.184851413590852e-06, 'learning_rate/action_model': 2.1848514135908518e-05, 'epoch': 1.5}) +[2026-06-16 02:14:59,079][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.0032510636374354362, 'train/grad_norm_pre_clip': 3.2241663932800293, 'timing/data': 0.0003529610112309456, 'timing/model': 1.133494983019773, 'learning_rate/qwen_vl_interface': 2.1825255428826087e-06, 'learning_rate/action_model': 2.182525542882609e-05, 'epoch': 1.5}) +[2026-06-16 02:15:03,640][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.09140937775373459, 'train/grad_norm_pre_clip': 16.741336822509766, 'timing/data': 0.0005438869702629745, 'timing/model': 1.1449181939824484, 'learning_rate/qwen_vl_interface': 2.180200933097246e-06, 'learning_rate/action_model': 2.180200933097246e-05, 'epoch': 1.5}) +[2026-06-16 02:15:08,170][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.028331516310572624, 'train/grad_norm_pre_clip': 19.417421340942383, 'timing/data': 0.00031875300919637084, 'timing/model': 1.1361555919866078, 'learning_rate/qwen_vl_interface': 2.1778775851903235e-06, 'learning_rate/action_model': 2.177877585190323e-05, 'epoch': 1.5}) +[2026-06-16 02:15:12,719][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.19271980226039886, 'train/grad_norm_pre_clip': 16.44629669189453, 'timing/data': 0.00036270800046622753, 'timing/model': 1.1389584239805117, 'learning_rate/qwen_vl_interface': 2.175555500116881e-06, 'learning_rate/action_model': 2.175555500116881e-05, 'epoch': 1.5}) +[2026-06-16 02:15:17,275][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.001229248708114028, 'train/grad_norm_pre_clip': 0.37239864468574524, 'timing/data': 0.0003988639800809324, 'timing/model': 1.2978727279696614, 'learning_rate/qwen_vl_interface': 2.1732346788314404e-06, 'learning_rate/action_model': 2.1732346788314402e-05, 'epoch': 1.5}) +[2026-06-16 02:15:21,657][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.01857026480138302, 'train/grad_norm_pre_clip': 11.77573013305664, 'timing/data': 0.00034492596751078963, 'timing/model': 1.137713578005787, 'learning_rate/qwen_vl_interface': 2.170915122288006e-06, 'learning_rate/action_model': 2.170915122288006e-05, 'epoch': 1.5}) +[2026-06-16 02:15:26,200][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.11656378954648972, 'train/grad_norm_pre_clip': 13.152010917663574, 'timing/data': 0.0004175299545750022, 'timing/model': 1.1344251380069181, 'learning_rate/qwen_vl_interface': 2.168596831440055e-06, 'learning_rate/action_model': 2.1685968314400552e-05, 'epoch': 1.5}) +[2026-06-16 02:15:30,713][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.6234413981437683, 'train/grad_norm_pre_clip': 37.355918884277344, 'timing/data': 0.0003400759887881577, 'timing/model': 1.143164862995036, 'learning_rate/qwen_vl_interface': 2.166279807240553e-06, 'learning_rate/action_model': 2.1662798072405528e-05, 'epoch': 1.5}) +[2026-06-16 02:15:35,188][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.009966426528990269, 'train/grad_norm_pre_clip': 1.9545360803604126, 'timing/data': 0.000331053975969553, 'timing/model': 1.1326951049850322, 'learning_rate/qwen_vl_interface': 2.1639640506419394e-06, 'learning_rate/action_model': 2.163964050641939e-05, 'epoch': 1.5}) +[2026-06-16 02:15:39,724][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.053799230605363846, 'train/grad_norm_pre_clip': 12.915228843688965, 'timing/data': 0.0003224870306439698, 'timing/model': 1.1334862629882991, 'learning_rate/qwen_vl_interface': 2.161649562596135e-06, 'learning_rate/action_model': 2.161649562596135e-05, 'epoch': 1.5}) +[2026-06-16 02:15:44,236][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.0005487900343723595, 'train/grad_norm_pre_clip': 0.0912201777100563, 'timing/data': 0.0003480740124359727, 'timing/model': 1.268872543994803, 'learning_rate/qwen_vl_interface': 2.1593363440545383e-06, 'learning_rate/action_model': 2.1593363440545384e-05, 'epoch': 1.5}) +[2026-06-16 02:15:48,626][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.03252723067998886, 'train/grad_norm_pre_clip': 11.967331886291504, 'timing/data': 0.0003361969720572233, 'timing/model': 1.138005782966502, 'learning_rate/qwen_vl_interface': 2.157024395968025e-06, 'learning_rate/action_model': 2.1570243959680252e-05, 'epoch': 1.5}) +[2026-06-16 02:15:53,157][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.05597329139709473, 'train/grad_norm_pre_clip': 24.21073341369629, 'timing/data': 0.00048363901441916823, 'timing/model': 1.130633202032186, 'learning_rate/qwen_vl_interface': 2.154713719286952e-06, 'learning_rate/action_model': 2.154713719286952e-05, 'epoch': 1.5}) +[2026-06-16 02:15:57,676][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.1723432093858719, 'train/grad_norm_pre_clip': 29.652450561523438, 'timing/data': 0.00042005296563729644, 'timing/model': 1.1304435179918073, 'learning_rate/qwen_vl_interface': 2.152404314961149e-06, 'learning_rate/action_model': 2.152404314961149e-05, 'epoch': 1.5}) +[2026-06-16 02:16:02,183][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.20848239958286285, 'train/grad_norm_pre_clip': 12.83871841430664, 'timing/data': 0.0003480320447124541, 'timing/model': 1.131169594998937, 'learning_rate/qwen_vl_interface': 2.1500961839399264e-06, 'learning_rate/action_model': 2.1500961839399262e-05, 'epoch': 1.5}) +[2026-06-16 02:16:06,725][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.00016069547564256936, 'train/grad_norm_pre_clip': 0.022867020219564438, 'timing/data': 0.00038725894410163164, 'timing/model': 1.134038133022841, 'learning_rate/qwen_vl_interface': 2.1477893271720683e-06, 'learning_rate/action_model': 2.1477893271720682e-05, 'epoch': 1.5}) +[2026-06-16 02:16:11,275][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.0053375717252492905, 'train/grad_norm_pre_clip': 1.0191857814788818, 'timing/data': 0.00037350301863625646, 'timing/model': 1.3079083939664997, 'learning_rate/qwen_vl_interface': 2.1454837456058385e-06, 'learning_rate/action_model': 2.1454837456058386e-05, 'epoch': 1.5}) +[2026-06-16 02:16:15,647][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.24659861624240875, 'train/grad_norm_pre_clip': 12.274580001831055, 'timing/data': 0.000364833977073431, 'timing/model': 1.1308384419535287, 'learning_rate/qwen_vl_interface': 2.1431794401889722e-06, 'learning_rate/action_model': 2.143179440188972e-05, 'epoch': 1.5}) +[2026-06-16 02:16:20,189][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.24805231392383575, 'train/grad_norm_pre_clip': 16.54288673400879, 'timing/data': 0.00034413899993523955, 'timing/model': 1.133426774002146, 'learning_rate/qwen_vl_interface': 2.1408764118686837e-06, 'learning_rate/action_model': 2.1408764118686838e-05, 'epoch': 1.5}) +[2026-06-16 02:16:24,692][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.004972703289240599, 'train/grad_norm_pre_clip': 1.1925410032272339, 'timing/data': 0.0003330979961901903, 'timing/model': 1.1460110530024394, 'learning_rate/qwen_vl_interface': 2.1385746615916598e-06, 'learning_rate/action_model': 2.1385746615916596e-05, 'epoch': 1.51}) +[2026-06-16 02:16:29,175][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.009732477366924286, 'train/grad_norm_pre_clip': 3.550530195236206, 'timing/data': 0.0005742720095440745, 'timing/model': 1.1438189059845172, 'learning_rate/qwen_vl_interface': 2.1362741903040636e-06, 'learning_rate/action_model': 2.1362741903040633e-05, 'epoch': 1.51}) +[2026-06-16 02:16:33,674][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.3352183997631073, 'train/grad_norm_pre_clip': 17.281862258911133, 'timing/data': 0.0003562709898687899, 'timing/model': 1.1338325050310232, 'learning_rate/qwen_vl_interface': 2.1339749989515318e-06, 'learning_rate/action_model': 2.1339749989515316e-05, 'epoch': 1.51}) +[2026-06-16 02:16:38,258][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.08060334622859955, 'train/grad_norm_pre_clip': 19.921335220336914, 'timing/data': 0.00036084698513150215, 'timing/model': 1.2937970319762826, 'learning_rate/qwen_vl_interface': 2.131677088479175e-06, 'learning_rate/action_model': 2.1316770884791752e-05, 'epoch': 1.51}) +[2026-06-16 02:16:42,673][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.0003536761796567589, 'train/grad_norm_pre_clip': 0.06076318398118019, 'timing/data': 0.0003285710117779672, 'timing/model': 1.142395329021383, 'learning_rate/qwen_vl_interface': 2.1293804598315787e-06, 'learning_rate/action_model': 2.1293804598315787e-05, 'epoch': 1.51}) +[2026-06-16 02:16:47,227][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.015619074925780296, 'train/grad_norm_pre_clip': 6.393510341644287, 'timing/data': 0.00033346202690154314, 'timing/model': 1.1358313880045898, 'learning_rate/qwen_vl_interface': 2.1270851139527986e-06, 'learning_rate/action_model': 2.1270851139527983e-05, 'epoch': 1.51}) +[2026-06-16 02:16:51,767][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.04835442453622818, 'train/grad_norm_pre_clip': 14.557929992675781, 'timing/data': 0.00039574201218783855, 'timing/model': 1.1418371069594286, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 1.51}) +[2026-06-16 02:16:56,312][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.00798056274652481, 'train/grad_norm_pre_clip': 3.647362470626831, 'timing/data': 0.0003637719783000648, 'timing/model': 1.1418464750167914, 'learning_rate/qwen_vl_interface': 2.1224982742752814e-06, 'learning_rate/action_model': 2.1224982742752812e-05, 'epoch': 1.51}) +[2026-06-16 02:17:00,864][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.11040938645601273, 'train/grad_norm_pre_clip': 13.95754337310791, 'timing/data': 0.0005876810173504055, 'timing/model': 1.1328137399978004, 'learning_rate/qwen_vl_interface': 2.1202067823620213e-06, 'learning_rate/action_model': 2.1202067823620212e-05, 'epoch': 1.51}) +[2026-06-16 02:17:05,431][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.002591957338154316, 'train/grad_norm_pre_clip': 0.7709567546844482, 'timing/data': 0.0003444210160523653, 'timing/model': 1.3113813669770025, 'learning_rate/qwen_vl_interface': 2.117916576988531e-06, 'learning_rate/action_model': 2.117916576988531e-05, 'epoch': 1.51}) +[2026-06-16 02:17:09,821][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.11461775749921799, 'train/grad_norm_pre_clip': 44.49475860595703, 'timing/data': 0.00035032996674999595, 'timing/model': 1.146306311013177, 'learning_rate/qwen_vl_interface': 2.1156276590962257e-06, 'learning_rate/action_model': 2.1156276590962257e-05, 'epoch': 1.51}) +[2026-06-16 02:17:14,380][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.17856821417808533, 'train/grad_norm_pre_clip': 20.79141616821289, 'timing/data': 0.0003757999511435628, 'timing/model': 1.1383763320045546, 'learning_rate/qwen_vl_interface': 2.113340029625994e-06, 'learning_rate/action_model': 2.1133400296259938e-05, 'epoch': 1.51}) +[2026-06-16 02:17:18,884][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.2727523148059845, 'train/grad_norm_pre_clip': 43.782318115234375, 'timing/data': 0.00047891098074615, 'timing/model': 1.142412101966329, 'learning_rate/qwen_vl_interface': 2.111053689518194e-06, 'learning_rate/action_model': 2.111053689518194e-05, 'epoch': 1.51}) +[2026-06-16 02:17:23,357][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.004361254163086414, 'train/grad_norm_pre_clip': 2.145317554473877, 'timing/data': 0.00036865100264549255, 'timing/model': 1.141309430997353, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 1.51}) +[2026-06-16 02:17:27,887][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.10760850459337234, 'train/grad_norm_pre_clip': 34.95500564575195, 'timing/data': 0.0003404839662835002, 'timing/model': 1.1392940490040928, 'learning_rate/qwen_vl_interface': 2.106484881148672e-06, 'learning_rate/action_model': 2.1064848811486718e-05, 'epoch': 1.51}) +[2026-06-16 02:17:32,454][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.13188685476779938, 'train/grad_norm_pre_clip': 13.266410827636719, 'timing/data': 0.0004211419727653265, 'timing/model': 1.3040376210119575, 'learning_rate/qwen_vl_interface': 2.1042024147650137e-06, 'learning_rate/action_model': 2.1042024147650137e-05, 'epoch': 1.51}) +[2026-06-16 02:17:36,855][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.06488204747438431, 'train/grad_norm_pre_clip': 13.696639060974121, 'timing/data': 0.00037466298090294003, 'timing/model': 1.142817335028667, 'learning_rate/qwen_vl_interface': 2.1019212414999163e-06, 'learning_rate/action_model': 2.101921241499916e-05, 'epoch': 1.51}) +[2026-06-16 02:17:41,408][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.03436766564846039, 'train/grad_norm_pre_clip': 8.966695785522461, 'timing/data': 0.0005286590312607586, 'timing/model': 1.1433327700360678, 'learning_rate/qwen_vl_interface': 2.099641362291082e-06, 'learning_rate/action_model': 2.099641362291082e-05, 'epoch': 1.51}) +[2026-06-16 02:17:45,949][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.002074012765660882, 'train/grad_norm_pre_clip': 0.36878079175949097, 'timing/data': 0.0005807969719171524, 'timing/model': 1.1355909620178863, 'learning_rate/qwen_vl_interface': 2.097362778075685e-06, 'learning_rate/action_model': 2.097362778075685e-05, 'epoch': 1.51}) +[2026-06-16 02:17:50,481][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.0031418006401509047, 'train/grad_norm_pre_clip': 0.8257646560668945, 'timing/data': 0.00034265604335814714, 'timing/model': 1.135334595979657, 'learning_rate/qwen_vl_interface': 2.0950854897903663e-06, 'learning_rate/action_model': 2.095085489790366e-05, 'epoch': 1.51}) +[2026-06-16 02:17:55,033][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.004893554840236902, 'train/grad_norm_pre_clip': 1.7264736890792847, 'timing/data': 0.00036589900264516473, 'timing/model': 1.1396365700056776, 'learning_rate/qwen_vl_interface': 2.092809498371229e-06, 'learning_rate/action_model': 2.0928094983712292e-05, 'epoch': 1.51}) +[2026-06-16 02:17:59,577][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.0653458908200264, 'train/grad_norm_pre_clip': 22.878721237182617, 'timing/data': 0.00034288299502804875, 'timing/model': 1.2929328309837729, 'learning_rate/qwen_vl_interface': 2.0905348047538498e-06, 'learning_rate/action_model': 2.0905348047538497e-05, 'epoch': 1.51}) +[2026-06-16 02:18:03,982][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.003948064520955086, 'train/grad_norm_pre_clip': 1.0292887687683105, 'timing/data': 0.0003380870330147445, 'timing/model': 1.1518455159966834, 'learning_rate/qwen_vl_interface': 2.0882614098732686e-06, 'learning_rate/action_model': 2.0882614098732685e-05, 'epoch': 1.51}) +[2026-06-16 02:18:08,567][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.037734851241111755, 'train/grad_norm_pre_clip': 22.486371994018555, 'timing/data': 0.0005925120203755796, 'timing/model': 1.142592761956621, 'learning_rate/qwen_vl_interface': 2.085989314663993e-06, 'learning_rate/action_model': 2.0859893146639927e-05, 'epoch': 1.51}) +[2026-06-16 02:18:13,082][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.012922143563628197, 'train/grad_norm_pre_clip': 6.6222052574157715, 'timing/data': 0.0004045229870826006, 'timing/model': 1.13319548900472, 'learning_rate/qwen_vl_interface': 2.083718520059995e-06, 'learning_rate/action_model': 2.0837185200599946e-05, 'epoch': 1.52}) +[2026-06-16 02:18:17,574][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.05053205043077469, 'train/grad_norm_pre_clip': 15.683282852172852, 'timing/data': 0.0003319790121167898, 'timing/model': 1.135634598962497, 'learning_rate/qwen_vl_interface': 2.0814490269947117e-06, 'learning_rate/action_model': 2.0814490269947118e-05, 'epoch': 1.52}) +[2026-06-16 02:18:22,071][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.1371094137430191, 'train/grad_norm_pre_clip': 10.551587104797363, 'timing/data': 0.0006100399768911302, 'timing/model': 1.1357327839941718, 'learning_rate/qwen_vl_interface': 2.079180836401047e-06, 'learning_rate/action_model': 2.079180836401047e-05, 'epoch': 1.52}) +[2026-06-16 02:18:26,624][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.022901251912117004, 'train/grad_norm_pre_clip': 14.128332138061523, 'timing/data': 0.0003430240321904421, 'timing/model': 1.3015594569733366, 'learning_rate/qwen_vl_interface': 2.0769139492113687e-06, 'learning_rate/action_model': 2.076913949211369e-05, 'epoch': 1.52}) +[2026-06-16 02:18:31,014][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.005434501450508833, 'train/grad_norm_pre_clip': 1.256452202796936, 'timing/data': 0.00035604502772912383, 'timing/model': 1.1348112500272691, 'learning_rate/qwen_vl_interface': 2.0746483663575075e-06, 'learning_rate/action_model': 2.0746483663575076e-05, 'epoch': 1.52}) +[2026-06-16 02:18:35,569][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.014100863598287106, 'train/grad_norm_pre_clip': 5.526172161102295, 'timing/data': 0.00032764702336862683, 'timing/model': 1.13960453402251, 'learning_rate/qwen_vl_interface': 2.0723840887707595e-06, 'learning_rate/action_model': 2.0723840887707594e-05, 'epoch': 1.52}) +[2026-06-16 02:18:40,105][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.15051224827766418, 'train/grad_norm_pre_clip': 8.46086597442627, 'timing/data': 0.0004914940218441188, 'timing/model': 1.1350691360421479, 'learning_rate/qwen_vl_interface': 2.0701211173818854e-06, 'learning_rate/action_model': 2.0701211173818853e-05, 'epoch': 1.52}) +[2026-06-16 02:18:44,636][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.12473427504301071, 'train/grad_norm_pre_clip': 14.427331924438477, 'timing/data': 0.000351605995092541, 'timing/model': 1.1367560410290025, 'learning_rate/qwen_vl_interface': 2.0678594531211036e-06, 'learning_rate/action_model': 2.0678594531211033e-05, 'epoch': 1.52}) +[2026-06-16 02:18:49,194][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.1885678768157959, 'train/grad_norm_pre_clip': 28.363967895507812, 'timing/data': 0.0004036119789816439, 'timing/model': 1.1436062749708071, 'learning_rate/qwen_vl_interface': 2.0655990969181e-06, 'learning_rate/action_model': 2.0655990969181003e-05, 'epoch': 1.52}) +[2026-06-16 02:18:53,743][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.0006745490245521069, 'train/grad_norm_pre_clip': 0.12094218283891678, 'timing/data': 0.0004036749596707523, 'timing/model': 1.2910281550139189, 'learning_rate/qwen_vl_interface': 2.0633400497020236e-06, 'learning_rate/action_model': 2.0633400497020232e-05, 'epoch': 1.52}) +[2026-06-16 02:18:58,128][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.005074894055724144, 'train/grad_norm_pre_clip': 1.542910099029541, 'timing/data': 0.0003114520222879946, 'timing/model': 1.1391234539914876, 'learning_rate/qwen_vl_interface': 2.061082312401481e-06, 'learning_rate/action_model': 2.061082312401481e-05, 'epoch': 1.52}) +[2026-06-16 02:19:02,664][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.04097232222557068, 'train/grad_norm_pre_clip': 16.573841094970703, 'timing/data': 0.0003957279841415584, 'timing/model': 1.1348992480197921, 'learning_rate/qwen_vl_interface': 2.058825885944544e-06, 'learning_rate/action_model': 2.0588258859445442e-05, 'epoch': 1.52}) +[2026-06-16 02:23:18,661][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.0007698168628849089, 'train/grad_norm_pre_clip': 0.19286616146564484, 'eval/loss': 0.04108669236302376, 'eval/flap_precision': 0.9723926380368099, 'eval/flap_recall': 0.9783950617283951, 'eval/flap_f1': 0.9753846153846154, 'eval/flap_rate_model': 0.1811111111111111, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9753846153846153, 'timing/data': 0.0004274850361980498, 'timing/model': 1.1355581209645607, 'learning_rate/qwen_vl_interface': 2.0565707712587447e-06, 'learning_rate/action_model': 2.0565707712587444e-05, 'epoch': 1.52}) +[2026-06-16 02:23:18,662][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3700_state +[2026-06-16 02:23:18,664][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:23:22,360][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3700_state/model.safetensors +[2026-06-16 02:23:32,407][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3700_state/optimizer.bin +[2026-06-16 02:23:32,407][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3700_state/sampler.bin +[2026-06-16 02:23:32,407][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3700_state/sampler_1.bin +[2026-06-16 02:23:32,408][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3700_state/random_states_0.pkl +[2026-06-16 02:23:32,457][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3700 +[2026-06-16 02:23:32,457][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:23:32,464][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:23:36,907][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.11360827088356018, 'train/grad_norm_pre_clip': 16.275428771972656, 'timing/data': 0.00036662997445091605, 'timing/model': 1.110036766040139, 'learning_rate/qwen_vl_interface': 2.0543169692710747e-06, 'learning_rate/action_model': 2.054316969271075e-05, 'epoch': 1.52}) +[2026-06-16 02:23:41,393][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.00014385490794666111, 'train/grad_norm_pre_clip': 0.029137050732970238, 'timing/data': 0.0003375000087544322, 'timing/model': 1.1178454490145668, 'learning_rate/qwen_vl_interface': 2.0520644809079873e-06, 'learning_rate/action_model': 2.0520644809079873e-05, 'epoch': 1.52}) +[2026-06-16 02:23:45,867][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.05707526579499245, 'train/grad_norm_pre_clip': 15.242691993713379, 'timing/data': 0.000354824005626142, 'timing/model': 1.1190877439803444, 'learning_rate/qwen_vl_interface': 2.049813307095395e-06, 'learning_rate/action_model': 2.0498133070953947e-05, 'epoch': 1.52}) +[2026-06-16 02:23:50,332][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.00555596686899662, 'train/grad_norm_pre_clip': 3.679269313812256, 'timing/data': 0.00034869101364165545, 'timing/model': 1.1167490049847402, 'learning_rate/qwen_vl_interface': 2.047563448758672e-06, 'learning_rate/action_model': 2.0475634487586717e-05, 'epoch': 1.52}) +[2026-06-16 02:23:54,822][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.016290482133626938, 'train/grad_norm_pre_clip': 8.9539794921875, 'timing/data': 0.0003452720120549202, 'timing/model': 1.2909049679874443, 'learning_rate/qwen_vl_interface': 2.045314906822647e-06, 'learning_rate/action_model': 2.0453149068226467e-05, 'epoch': 1.52}) +[2026-06-16 02:23:59,141][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.20878258347511292, 'train/grad_norm_pre_clip': 30.878969192504883, 'timing/data': 0.00032277003629133105, 'timing/model': 1.1194311860017478, 'learning_rate/qwen_vl_interface': 2.0430676822116124e-06, 'learning_rate/action_model': 2.0430676822116123e-05, 'epoch': 1.52}) +[2026-06-16 02:24:03,626][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.002470346400514245, 'train/grad_norm_pre_clip': 0.5600546598434448, 'timing/data': 0.00036227796226739883, 'timing/model': 1.1214447790407576, 'learning_rate/qwen_vl_interface': 2.040821775849319e-06, 'learning_rate/action_model': 2.0408217758493188e-05, 'epoch': 1.52}) +[2026-06-16 02:24:08,132][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.027068311348557472, 'train/grad_norm_pre_clip': 17.410972595214844, 'timing/data': 0.0003897949936799705, 'timing/model': 1.1198667280259542, 'learning_rate/qwen_vl_interface': 2.03857718865897e-06, 'learning_rate/action_model': 2.0385771886589702e-05, 'epoch': 1.52}) +[2026-06-16 02:24:12,641][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.006631049327552319, 'train/grad_norm_pre_clip': 2.7587804794311523, 'timing/data': 0.00035773799754679203, 'timing/model': 1.128569591033738, 'learning_rate/qwen_vl_interface': 2.0363339215632337e-06, 'learning_rate/action_model': 2.0363339215632334e-05, 'epoch': 1.52}) +[2026-06-16 02:24:17,139][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.0062262327410280704, 'train/grad_norm_pre_clip': 4.127353191375732, 'timing/data': 0.0005373209714889526, 'timing/model': 1.2786189769976772, 'learning_rate/qwen_vl_interface': 2.034091975484231e-06, 'learning_rate/action_model': 2.0340919754842307e-05, 'epoch': 1.52}) +[2026-06-16 02:24:21,603][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.12887999415397644, 'train/grad_norm_pre_clip': 17.38507080078125, 'timing/data': 0.0003605420351959765, 'timing/model': 1.2372557910275646, 'learning_rate/qwen_vl_interface': 2.0318513513435415e-06, 'learning_rate/action_model': 2.0318513513435415e-05, 'epoch': 1.52}) +[2026-06-16 02:24:25,958][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.07954256236553192, 'train/grad_norm_pre_clip': 16.616029739379883, 'timing/data': 0.0003204129752703011, 'timing/model': 1.1242593610077165, 'learning_rate/qwen_vl_interface': 2.0296120500622018e-06, 'learning_rate/action_model': 2.0296120500622017e-05, 'epoch': 1.52}) +[2026-06-16 02:24:30,480][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.11801959574222565, 'train/grad_norm_pre_clip': 13.8043794631958, 'timing/data': 0.00031748501351103187, 'timing/model': 1.1266885690274648, 'learning_rate/qwen_vl_interface': 2.0273740725607034e-06, 'learning_rate/action_model': 2.0273740725607032e-05, 'epoch': 1.53}) +[2026-06-16 02:24:34,983][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.08341751247644424, 'train/grad_norm_pre_clip': 12.70370864868164, 'timing/data': 0.00036574999103322625, 'timing/model': 1.125629863992799, 'learning_rate/qwen_vl_interface': 2.0251374197589942e-06, 'learning_rate/action_model': 2.0251374197589943e-05, 'epoch': 1.53}) +[2026-06-16 02:24:39,494][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.013366006314754486, 'train/grad_norm_pre_clip': 8.319615364074707, 'timing/data': 0.00037283397978171706, 'timing/model': 1.1296284130075946, 'learning_rate/qwen_vl_interface': 2.0229020925764785e-06, 'learning_rate/action_model': 2.0229020925764785e-05, 'epoch': 1.53}) +[2026-06-16 02:24:43,913][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.23481272161006927, 'train/grad_norm_pre_clip': 23.595117568969727, 'timing/data': 0.00034267903538420796, 'timing/model': 1.1318124980316497, 'learning_rate/qwen_vl_interface': 2.0206680919320145e-06, 'learning_rate/action_model': 2.020668091932014e-05, 'epoch': 1.53}) +[2026-06-16 02:24:48,355][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.1301071047782898, 'train/grad_norm_pre_clip': 7.6583571434021, 'timing/data': 0.00035045400727540255, 'timing/model': 1.214509682962671, 'learning_rate/qwen_vl_interface': 2.0184354187439164e-06, 'learning_rate/action_model': 2.0184354187439162e-05, 'epoch': 1.53}) +[2026-06-16 02:24:52,732][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.053893763571977615, 'train/grad_norm_pre_clip': 18.126018524169922, 'timing/data': 0.00034740997944027185, 'timing/model': 1.1389752120012417, 'learning_rate/qwen_vl_interface': 2.016204073929949e-06, 'learning_rate/action_model': 2.016204073929949e-05, 'epoch': 1.53}) +[2026-06-16 02:24:57,216][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.005315673071891069, 'train/grad_norm_pre_clip': 2.3997926712036133, 'timing/data': 0.0003320989781059325, 'timing/model': 1.1305035410332493, 'learning_rate/qwen_vl_interface': 2.0139740584073363e-06, 'learning_rate/action_model': 2.0139740584073364e-05, 'epoch': 1.53}) +[2026-06-16 02:25:01,767][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.08140730857849121, 'train/grad_norm_pre_clip': 8.30740737915039, 'timing/data': 0.00034873600816354156, 'timing/model': 1.132855371979531, 'learning_rate/qwen_vl_interface': 2.011745373092753e-06, 'learning_rate/action_model': 2.011745373092753e-05, 'epoch': 1.53}) +[2026-06-16 02:25:06,289][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.036167729645967484, 'train/grad_norm_pre_clip': 27.2495059967041, 'timing/data': 0.0003478130092844367, 'timing/model': 1.1321508600376546, 'learning_rate/qwen_vl_interface': 2.0095180189023267e-06, 'learning_rate/action_model': 2.0095180189023265e-05, 'epoch': 1.53}) +[2026-06-16 02:25:10,812][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.08893847465515137, 'train/grad_norm_pre_clip': 11.421733856201172, 'timing/data': 0.0003453160170465708, 'timing/model': 1.2866467940039001, 'learning_rate/qwen_vl_interface': 2.0072919967516393e-06, 'learning_rate/action_model': 2.0072919967516392e-05, 'epoch': 1.53}) +[2026-06-16 02:25:15,178][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.0005083164433017373, 'train/grad_norm_pre_clip': 0.13799352943897247, 'timing/data': 0.0004060089704580605, 'timing/model': 1.1308256160118617, 'learning_rate/qwen_vl_interface': 2.0050673075557245e-06, 'learning_rate/action_model': 2.0050673075557243e-05, 'epoch': 1.53}) +[2026-06-16 02:25:19,732][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.012482834048569202, 'train/grad_norm_pre_clip': 4.923721790313721, 'timing/data': 0.00034865603083744645, 'timing/model': 1.1391158869955689, 'learning_rate/qwen_vl_interface': 2.0028439522290677e-06, 'learning_rate/action_model': 2.0028439522290676e-05, 'epoch': 1.53}) +[2026-06-16 02:25:24,175][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.2522871792316437, 'train/grad_norm_pre_clip': 16.623037338256836, 'timing/data': 0.00033665099181234837, 'timing/model': 1.131176113965921, 'learning_rate/qwen_vl_interface': 2.000621931685606e-06, 'learning_rate/action_model': 2.000621931685606e-05, 'epoch': 1.53}) +[2026-06-16 02:25:28,668][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.2195698767900467, 'train/grad_norm_pre_clip': 22.671337127685547, 'timing/data': 0.0003327979939058423, 'timing/model': 1.1458188730175607, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 1.53}) +[2026-06-16 02:25:33,184][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.13797374069690704, 'train/grad_norm_pre_clip': 21.030216217041016, 'timing/data': 0.00033749797148630023, 'timing/model': 1.1363627889659256, 'learning_rate/qwen_vl_interface': 1.9961818986012767e-06, 'learning_rate/action_model': 1.9961818986012763e-05, 'epoch': 1.53}) +[2026-06-16 02:25:37,726][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.35057055950164795, 'train/grad_norm_pre_clip': 27.046052932739258, 'timing/data': 0.00039901904528960586, 'timing/model': 1.2952249600202776, 'learning_rate/qwen_vl_interface': 1.9939638878855372e-06, 'learning_rate/action_model': 1.993963887885537e-05, 'epoch': 1.53}) +[2026-06-16 02:25:42,125][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.04610046371817589, 'train/grad_norm_pre_clip': 17.57159423828125, 'timing/data': 0.000341220002155751, 'timing/model': 1.1429488849826157, 'learning_rate/qwen_vl_interface': 1.991747215603252e-06, 'learning_rate/action_model': 1.991747215603252e-05, 'epoch': 1.53}) +[2026-06-16 02:25:46,670][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.09943929314613342, 'train/grad_norm_pre_clip': 7.260515213012695, 'timing/data': 0.00039305101381614804, 'timing/model': 1.130862201040145, 'learning_rate/qwen_vl_interface': 1.9895318826656116e-06, 'learning_rate/action_model': 1.9895318826656114e-05, 'epoch': 1.53}) +[2026-06-16 02:25:51,201][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.039444711059331894, 'train/grad_norm_pre_clip': 11.633627891540527, 'timing/data': 0.0003434930113144219, 'timing/model': 1.1300127460272051, 'learning_rate/qwen_vl_interface': 1.9873178899832553e-06, 'learning_rate/action_model': 1.9873178899832554e-05, 'epoch': 1.53}) +[2026-06-16 02:25:55,706][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.04719634726643562, 'train/grad_norm_pre_clip': 12.067584991455078, 'timing/data': 0.0003385329619050026, 'timing/model': 1.1345356160309166, 'learning_rate/qwen_vl_interface': 1.985105238466272e-06, 'learning_rate/action_model': 1.9851052384662722e-05, 'epoch': 1.53}) +[2026-06-16 02:26:00,252][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.02342124655842781, 'train/grad_norm_pre_clip': 11.01692008972168, 'timing/data': 0.0003471749951131642, 'timing/model': 1.1340598199749365, 'learning_rate/qwen_vl_interface': 1.9828939290242e-06, 'learning_rate/action_model': 1.9828939290241998e-05, 'epoch': 1.53}) +[2026-06-16 02:26:04,754][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.21490640938282013, 'train/grad_norm_pre_clip': 25.58631134033203, 'timing/data': 0.0003369039623066783, 'timing/model': 1.2565829049563035, 'learning_rate/qwen_vl_interface': 1.980683962566024e-06, 'learning_rate/action_model': 1.980683962566024e-05, 'epoch': 1.53}) +[2026-06-16 02:26:09,131][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.33734092116355896, 'train/grad_norm_pre_clip': 19.83382797241211, 'timing/data': 0.0003416690160520375, 'timing/model': 1.1347754820017144, 'learning_rate/qwen_vl_interface': 1.9784753400001784e-06, 'learning_rate/action_model': 1.978475340000178e-05, 'epoch': 1.53}) +[2026-06-16 02:26:13,663][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.0035781548358500004, 'train/grad_norm_pre_clip': 1.6513988971710205, 'timing/data': 0.00036542501766234636, 'timing/model': 1.1353659289889038, 'learning_rate/qwen_vl_interface': 1.9762680622345443e-06, 'learning_rate/action_model': 1.9762680622345442e-05, 'epoch': 1.53}) +[2026-06-16 02:26:18,208][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.004508956801146269, 'train/grad_norm_pre_clip': 0.7811059951782227, 'timing/data': 0.00036549801006913185, 'timing/model': 1.13182670698734, 'learning_rate/qwen_vl_interface': 1.9740621301764513e-06, 'learning_rate/action_model': 1.9740621301764513e-05, 'epoch': 1.54}) +[2026-06-16 02:26:22,781][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.018770160153508186, 'train/grad_norm_pre_clip': 4.527406692504883, 'timing/data': 0.0003272760077379644, 'timing/model': 1.1376086010131985, 'learning_rate/qwen_vl_interface': 1.971857544732672e-06, 'learning_rate/action_model': 1.971857544732672e-05, 'epoch': 1.54}) +[2026-06-16 02:26:27,233][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.018993182107806206, 'train/grad_norm_pre_clip': 4.999328136444092, 'timing/data': 0.00034906796645373106, 'timing/model': 1.1487058530328795, 'learning_rate/qwen_vl_interface': 1.96965430680943e-06, 'learning_rate/action_model': 1.96965430680943e-05, 'epoch': 1.54}) +[2026-06-16 02:26:31,778][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.14033910632133484, 'train/grad_norm_pre_clip': 18.40225601196289, 'timing/data': 0.00046328798634931445, 'timing/model': 1.2997082360088825, 'learning_rate/qwen_vl_interface': 1.967452417312393e-06, 'learning_rate/action_model': 1.9674524173123927e-05, 'epoch': 1.54}) +[2026-06-16 02:26:36,182][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.06479345262050629, 'train/grad_norm_pre_clip': 7.7822675704956055, 'timing/data': 0.00035014102468267083, 'timing/model': 1.1348189730197191, 'learning_rate/qwen_vl_interface': 1.9652518771466746e-06, 'learning_rate/action_model': 1.9652518771466747e-05, 'epoch': 1.54}) +[2026-06-16 02:26:40,726][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.002023177221417427, 'train/grad_norm_pre_clip': 0.4916623532772064, 'timing/data': 0.00034499098546803, 'timing/model': 1.1334271280211397, 'learning_rate/qwen_vl_interface': 1.963052687216834e-06, 'learning_rate/action_model': 1.963052687216834e-05, 'epoch': 1.54}) +[2026-06-16 02:26:45,280][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.13685037195682526, 'train/grad_norm_pre_clip': 10.495844841003418, 'timing/data': 0.0005659809685312212, 'timing/model': 1.1395696150138974, 'learning_rate/qwen_vl_interface': 1.960854848426874e-06, 'learning_rate/action_model': 1.960854848426874e-05, 'epoch': 1.54}) +[2026-06-16 02:26:49,817][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.14214588701725006, 'train/grad_norm_pre_clip': 15.691609382629395, 'timing/data': 0.0005607700441032648, 'timing/model': 1.1330114000011235, 'learning_rate/qwen_vl_interface': 1.9586583616802446e-06, 'learning_rate/action_model': 1.9586583616802446e-05, 'epoch': 1.54}) +[2026-06-16 02:26:54,360][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.0774916410446167, 'train/grad_norm_pre_clip': 6.113902568817139, 'timing/data': 0.00033655104925855994, 'timing/model': 1.2997795789851807, 'learning_rate/qwen_vl_interface': 1.956463227879838e-06, 'learning_rate/action_model': 1.9564632278798377e-05, 'epoch': 1.54}) +[2026-06-16 02:26:58,749][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.040123626589775085, 'train/grad_norm_pre_clip': 13.672184944152832, 'timing/data': 0.0003675140324048698, 'timing/model': 1.1364312790101394, 'learning_rate/qwen_vl_interface': 1.95426944792799e-06, 'learning_rate/action_model': 1.95426944792799e-05, 'epoch': 1.54}) +[2026-06-16 02:27:03,302][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.004892583005130291, 'train/grad_norm_pre_clip': 0.8066279292106628, 'timing/data': 0.0003284040139988065, 'timing/model': 1.1409600659972057, 'learning_rate/qwen_vl_interface': 1.9520770227264817e-06, 'learning_rate/action_model': 1.9520770227264817e-05, 'epoch': 1.54}) +[2026-06-16 02:27:07,748][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.004412869922816753, 'train/grad_norm_pre_clip': 0.8254330158233643, 'timing/data': 0.0003656249609775841, 'timing/model': 1.1296681269886903, 'learning_rate/qwen_vl_interface': 1.9498859531765373e-06, 'learning_rate/action_model': 1.9498859531765372e-05, 'epoch': 1.54}) +[2026-06-16 02:27:12,228][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.1422370821237564, 'train/grad_norm_pre_clip': 10.955512046813965, 'timing/data': 0.00033831497421488166, 'timing/model': 1.134856752003543, 'learning_rate/qwen_vl_interface': 1.9476962401788204e-06, 'learning_rate/action_model': 1.9476962401788204e-05, 'epoch': 1.54}) +[2026-06-16 02:27:16,734][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.019692614674568176, 'train/grad_norm_pre_clip': 9.900540351867676, 'timing/data': 0.00033850199542939663, 'timing/model': 1.1354708920116536, 'learning_rate/qwen_vl_interface': 1.9455078846334397e-06, 'learning_rate/action_model': 1.9455078846334395e-05, 'epoch': 1.54}) +[2026-06-16 02:27:21,281][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.10055431723594666, 'train/grad_norm_pre_clip': 21.331581115722656, 'timing/data': 0.0003960569738410413, 'timing/model': 1.2918034049798734, 'learning_rate/qwen_vl_interface': 1.9433208874399465e-06, 'learning_rate/action_model': 1.9433208874399464e-05, 'epoch': 1.54}) +[2026-06-16 02:27:25,685][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.002638868987560272, 'train/grad_norm_pre_clip': 0.3962137997150421, 'timing/data': 0.0006117550074122846, 'timing/model': 1.1504020239808597, 'learning_rate/qwen_vl_interface': 1.9411352494973314e-06, 'learning_rate/action_model': 1.9411352494973315e-05, 'epoch': 1.54}) +[2026-06-16 02:27:30,246][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.03063570335507393, 'train/grad_norm_pre_clip': 9.22206974029541, 'timing/data': 0.0006175019661895931, 'timing/model': 1.1423038229695521, 'learning_rate/qwen_vl_interface': 1.938950971704031e-06, 'learning_rate/action_model': 1.9389509717040308e-05, 'epoch': 1.54}) +[2026-06-16 02:27:34,794][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.0021675548050552607, 'train/grad_norm_pre_clip': 0.39817655086517334, 'timing/data': 0.0003534370334818959, 'timing/model': 1.1343679550336674, 'learning_rate/qwen_vl_interface': 1.936768054957915e-06, 'learning_rate/action_model': 1.9367680549579147e-05, 'epoch': 1.54}) +[2026-06-16 02:27:39,311][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.0706542357802391, 'train/grad_norm_pre_clip': 24.76332664489746, 'timing/data': 0.00031638500513508916, 'timing/model': 1.130639348004479, 'learning_rate/qwen_vl_interface': 1.9345865001562994e-06, 'learning_rate/action_model': 1.9345865001562994e-05, 'epoch': 1.54}) +[2026-06-16 02:27:43,847][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.005743005778640509, 'train/grad_norm_pre_clip': 1.3625221252441406, 'timing/data': 0.0006953770061954856, 'timing/model': 1.1320580779574811, 'learning_rate/qwen_vl_interface': 1.9324063081959394e-06, 'learning_rate/action_model': 1.9324063081959395e-05, 'epoch': 1.54}) +[2026-06-16 02:27:48,406][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.10132473707199097, 'train/grad_norm_pre_clip': 7.994956016540527, 'timing/data': 0.0004151199827902019, 'timing/model': 1.31033359200228, 'learning_rate/qwen_vl_interface': 1.9302274799730304e-06, 'learning_rate/action_model': 1.9302274799730306e-05, 'epoch': 1.54}) +[2026-06-16 02:27:52,806][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.12779787182807922, 'train/grad_norm_pre_clip': 10.016886711120605, 'timing/data': 0.0003372859791852534, 'timing/model': 1.1374587560421787, 'learning_rate/qwen_vl_interface': 1.9280500163832055e-06, 'learning_rate/action_model': 1.9280500163832055e-05, 'epoch': 1.54}) +[2026-06-16 02:27:57,351][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.007076981943100691, 'train/grad_norm_pre_clip': 3.068833351135254, 'timing/data': 0.0003192040021531284, 'timing/model': 1.1388222749810666, 'learning_rate/qwen_vl_interface': 1.925873918321537e-06, 'learning_rate/action_model': 1.925873918321537e-05, 'epoch': 1.54}) +[2026-06-16 02:28:01,918][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.006302082445472479, 'train/grad_norm_pre_clip': 2.244636058807373, 'timing/data': 0.0003254050388932228, 'timing/model': 1.1374454150209203, 'learning_rate/qwen_vl_interface': 1.92369918668254e-06, 'learning_rate/action_model': 1.92369918668254e-05, 'epoch': 1.54}) +[2026-06-16 02:28:06,476][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.004003709182143211, 'train/grad_norm_pre_clip': 0.7157647013664246, 'timing/data': 0.00037969701224938035, 'timing/model': 1.1347300910274498, 'learning_rate/qwen_vl_interface': 1.9215258223601605e-06, 'learning_rate/action_model': 1.9215258223601605e-05, 'epoch': 1.55}) +[2026-06-16 02:28:10,985][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.022810863330960274, 'train/grad_norm_pre_clip': 12.515786170959473, 'timing/data': 0.0003285320126451552, 'timing/model': 1.1350944630103186, 'learning_rate/qwen_vl_interface': 1.9193538262477874e-06, 'learning_rate/action_model': 1.9193538262477874e-05, 'epoch': 1.55}) +[2026-06-16 02:28:15,534][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.0026829587295651436, 'train/grad_norm_pre_clip': 0.5415170192718506, 'timing/data': 0.0003208779962733388, 'timing/model': 1.2863758970052004, 'learning_rate/qwen_vl_interface': 1.917183199238247e-06, 'learning_rate/action_model': 1.917183199238247e-05, 'epoch': 1.55}) +[2026-06-16 02:28:19,910][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.042480163276195526, 'train/grad_norm_pre_clip': 12.132181167602539, 'timing/data': 0.00033258303301408887, 'timing/model': 1.13094452599762, 'learning_rate/qwen_vl_interface': 1.9150139422238016e-06, 'learning_rate/action_model': 1.9150139422238018e-05, 'epoch': 1.55}) +[2026-06-16 02:28:24,467][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.005108860321342945, 'train/grad_norm_pre_clip': 0.893848180770874, 'timing/data': 0.0005879359669052064, 'timing/model': 1.1326820559916086, 'learning_rate/qwen_vl_interface': 1.9128460560961513e-06, 'learning_rate/action_model': 1.9128460560961514e-05, 'epoch': 1.55}) +[2026-06-16 02:28:28,991][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.0014513686764985323, 'train/grad_norm_pre_clip': 0.34260857105255127, 'timing/data': 0.0003609919804148376, 'timing/model': 1.129862652043812, 'learning_rate/qwen_vl_interface': 1.910679541746432e-06, 'learning_rate/action_model': 1.9106795417464323e-05, 'epoch': 1.55}) +[2026-06-16 02:28:33,525][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.0007667217287234962, 'train/grad_norm_pre_clip': 0.14354929327964783, 'timing/data': 0.00032762502087280154, 'timing/model': 1.1486294920323417, 'learning_rate/qwen_vl_interface': 1.9085144000652166e-06, 'learning_rate/action_model': 1.9085144000652166e-05, 'epoch': 1.55}) +[2026-06-16 02:28:38,061][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.006405194289982319, 'train/grad_norm_pre_clip': 1.7747756242752075, 'timing/data': 0.00035564997233450413, 'timing/model': 1.1368917750078253, 'learning_rate/qwen_vl_interface': 1.9063506319425116e-06, 'learning_rate/action_model': 1.9063506319425114e-05, 'epoch': 1.55}) +[2026-06-16 02:28:42,598][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.001006596488878131, 'train/grad_norm_pre_clip': 0.11658783257007599, 'timing/data': 0.000405077007599175, 'timing/model': 1.2914166080299765, 'learning_rate/qwen_vl_interface': 1.9041882382677613e-06, 'learning_rate/action_model': 1.9041882382677612e-05, 'epoch': 1.55}) +[2026-06-16 02:28:46,977][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.02877715416252613, 'train/grad_norm_pre_clip': 50.734981536865234, 'timing/data': 0.0003592540160752833, 'timing/model': 1.136916906980332, 'learning_rate/qwen_vl_interface': 1.9020272199298452e-06, 'learning_rate/action_model': 1.902027219929845e-05, 'epoch': 1.55}) +[2026-06-16 02:28:51,520][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.011197821237146854, 'train/grad_norm_pre_clip': 6.897371292114258, 'timing/data': 0.00033712503500282764, 'timing/model': 1.1393699549953453, 'learning_rate/qwen_vl_interface': 1.8998675778170736e-06, 'learning_rate/action_model': 1.8998675778170734e-05, 'epoch': 1.55}) +[2026-06-16 02:28:56,069][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.0003440575092099607, 'train/grad_norm_pre_clip': 0.05637644603848457, 'timing/data': 0.0003307830193080008, 'timing/model': 1.138189455959946, 'learning_rate/qwen_vl_interface': 1.8977093128171955e-06, 'learning_rate/action_model': 1.8977093128171954e-05, 'epoch': 1.55}) +[2026-06-16 02:29:00,612][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.02270994335412979, 'train/grad_norm_pre_clip': 10.54100513458252, 'timing/data': 0.0007221089908853173, 'timing/model': 1.1343489529681392, 'learning_rate/qwen_vl_interface': 1.8955524258173919e-06, 'learning_rate/action_model': 1.895552425817392e-05, 'epoch': 1.55}) +[2026-06-16 02:29:05,115][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.0005475255893543363, 'train/grad_norm_pre_clip': 0.1059081181883812, 'timing/data': 0.0003489700029604137, 'timing/model': 1.1373941799975, 'learning_rate/qwen_vl_interface': 1.8933969177042769e-06, 'learning_rate/action_model': 1.8933969177042768e-05, 'epoch': 1.55}) +[2026-06-16 02:29:09,667][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.02939034067094326, 'train/grad_norm_pre_clip': 11.54548454284668, 'timing/data': 0.000341032980941236, 'timing/model': 1.2888616070267744, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 1.55}) +[2026-06-16 02:29:14,048][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.08624182641506195, 'train/grad_norm_pre_clip': 6.330860614776611, 'timing/data': 0.0003844299935735762, 'timing/model': 1.1331452049780637, 'learning_rate/qwen_vl_interface': 1.8890900416817398e-06, 'learning_rate/action_model': 1.8890900416817397e-05, 'epoch': 1.55}) +[2026-06-16 02:29:18,612][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.0013746318873018026, 'train/grad_norm_pre_clip': 0.3080168068408966, 'timing/data': 0.0003453640383668244, 'timing/model': 1.1376657179789618, 'learning_rate/qwen_vl_interface': 1.8869386755427117e-06, 'learning_rate/action_model': 1.8869386755427114e-05, 'epoch': 1.55}) +[2026-06-16 02:29:23,167][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.3013651669025421, 'train/grad_norm_pre_clip': 20.23997688293457, 'timing/data': 0.0004923680098727345, 'timing/model': 1.1368327480158769, 'learning_rate/qwen_vl_interface': 1.8847886918311597e-06, 'learning_rate/action_model': 1.8847886918311598e-05, 'epoch': 1.55}) +[2026-06-16 02:29:27,665][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.00375833036378026, 'train/grad_norm_pre_clip': 1.0414152145385742, 'timing/data': 0.0003618279588408768, 'timing/model': 1.1318236869992688, 'learning_rate/qwen_vl_interface': 1.8826400914308616e-06, 'learning_rate/action_model': 1.8826400914308613e-05, 'epoch': 1.55}) +[2026-06-16 02:29:32,196][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.014828711748123169, 'train/grad_norm_pre_clip': 14.434804916381836, 'timing/data': 0.00036253000143915415, 'timing/model': 1.135383273998741, 'learning_rate/qwen_vl_interface': 1.880492875225025e-06, 'learning_rate/action_model': 1.8804928752250252e-05, 'epoch': 1.55}) +[2026-06-16 02:29:36,737][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.24972467124462128, 'train/grad_norm_pre_clip': 16.681764602661133, 'timing/data': 0.00034741597482934594, 'timing/model': 1.2920658910297789, 'learning_rate/qwen_vl_interface': 1.8783470440962922e-06, 'learning_rate/action_model': 1.878347044096292e-05, 'epoch': 1.55}) +[2026-06-16 02:29:41,123][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.021819135174155235, 'train/grad_norm_pre_clip': 11.514753341674805, 'timing/data': 0.0003530289977788925, 'timing/model': 1.136334765003994, 'learning_rate/qwen_vl_interface': 1.8762025989267293e-06, 'learning_rate/action_model': 1.876202598926729e-05, 'epoch': 1.55}) +[2026-06-16 02:29:45,652][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.019988667219877243, 'train/grad_norm_pre_clip': 11.074170112609863, 'timing/data': 0.0003304939600639045, 'timing/model': 1.1352206229930744, 'learning_rate/qwen_vl_interface': 1.874059540597839e-06, 'learning_rate/action_model': 1.874059540597839e-05, 'epoch': 1.55}) +[2026-06-16 02:29:50,201][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.22807471454143524, 'train/grad_norm_pre_clip': 10.114053726196289, 'timing/data': 0.0003432929515838623, 'timing/model': 1.1302450189832598, 'learning_rate/qwen_vl_interface': 1.8719178699905508e-06, 'learning_rate/action_model': 1.8719178699905508e-05, 'epoch': 1.55}) +[2026-06-16 02:29:54,768][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.0008147305925376713, 'train/grad_norm_pre_clip': 0.17319203913211823, 'timing/data': 0.00036249496042728424, 'timing/model': 1.1458014659583569, 'learning_rate/qwen_vl_interface': 1.8697775879852242e-06, 'learning_rate/action_model': 1.8697775879852242e-05, 'epoch': 1.55}) +[2026-06-16 02:29:59,284][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.2700316309928894, 'train/grad_norm_pre_clip': 19.13477897644043, 'timing/data': 0.0005438199732452631, 'timing/model': 1.1323624849901535, 'learning_rate/qwen_vl_interface': 1.86763869546165e-06, 'learning_rate/action_model': 1.86763869546165e-05, 'epoch': 1.56}) +[2026-06-16 02:30:03,838][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.11582617461681366, 'train/grad_norm_pre_clip': 13.062997817993164, 'timing/data': 0.000354708987288177, 'timing/model': 1.29776150098769, 'learning_rate/qwen_vl_interface': 1.865501193299045e-06, 'learning_rate/action_model': 1.8655011932990447e-05, 'epoch': 1.56}) +[2026-06-16 02:30:08,219][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.0031979898922145367, 'train/grad_norm_pre_clip': 0.5078079700469971, 'timing/data': 0.00035454099997878075, 'timing/model': 1.1327532120049, 'learning_rate/qwen_vl_interface': 1.8633650823760555e-06, 'learning_rate/action_model': 1.8633650823760555e-05, 'epoch': 1.56}) +[2026-06-16 02:30:12,772][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.08065248280763626, 'train/grad_norm_pre_clip': 20.65090560913086, 'timing/data': 0.0003785740118473768, 'timing/model': 1.1376802019658498, 'learning_rate/qwen_vl_interface': 1.8612303635707562e-06, 'learning_rate/action_model': 1.861230363570756e-05, 'epoch': 1.56}) +[2026-06-16 02:30:17,310][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.03146684542298317, 'train/grad_norm_pre_clip': 9.004049301147461, 'timing/data': 0.0004872879944741726, 'timing/model': 1.132379918009974, 'learning_rate/qwen_vl_interface': 1.8590970377606494e-06, 'learning_rate/action_model': 1.859097037760649e-05, 'epoch': 1.56}) +[2026-06-16 02:30:21,853][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.03993331640958786, 'train/grad_norm_pre_clip': 29.714933395385742, 'timing/data': 0.00036621798062697053, 'timing/model': 1.1368098490056582, 'learning_rate/qwen_vl_interface': 1.8569651058226667e-06, 'learning_rate/action_model': 1.8569651058226666e-05, 'epoch': 1.56}) +[2026-06-16 02:30:26,408][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.33368831872940063, 'train/grad_norm_pre_clip': 30.4030704498291, 'timing/data': 0.0003021170268766582, 'timing/model': 1.1318904369836673, 'learning_rate/qwen_vl_interface': 1.854834568633161e-06, 'learning_rate/action_model': 1.854834568633161e-05, 'epoch': 1.56}) +[2026-06-16 02:30:30,963][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.535296618938446, 'train/grad_norm_pre_clip': 7.70012092590332, 'timing/data': 0.00035509897861629725, 'timing/model': 1.3013808319810778, 'learning_rate/qwen_vl_interface': 1.8527054270679174e-06, 'learning_rate/action_model': 1.8527054270679172e-05, 'epoch': 1.56}) +[2026-06-16 02:30:35,354][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.04610249027609825, 'train/grad_norm_pre_clip': 41.18806457519531, 'timing/data': 0.00036849599564448, 'timing/model': 1.133507818041835, 'learning_rate/qwen_vl_interface': 1.850577682002146e-06, 'learning_rate/action_model': 1.850577682002146e-05, 'epoch': 1.56}) +[2026-06-16 02:30:39,902][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.0017861274536699057, 'train/grad_norm_pre_clip': 0.3281713128089905, 'timing/data': 0.00037174602039158344, 'timing/model': 1.1349351949756965, 'learning_rate/qwen_vl_interface': 1.848451334310482e-06, 'learning_rate/action_model': 1.848451334310482e-05, 'epoch': 1.56}) +[2026-06-16 02:30:44,455][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.10135851055383682, 'train/grad_norm_pre_clip': 11.661224365234375, 'timing/data': 0.00035680097062140703, 'timing/model': 1.134277815988753, 'learning_rate/qwen_vl_interface': 1.8463263848669865e-06, 'learning_rate/action_model': 1.8463263848669864e-05, 'epoch': 1.56}) +[2026-06-16 02:30:49,014][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.006261239293962717, 'train/grad_norm_pre_clip': 1.8647701740264893, 'timing/data': 0.0003534049610607326, 'timing/model': 1.1427165149943903, 'learning_rate/qwen_vl_interface': 1.8442028345451468e-06, 'learning_rate/action_model': 1.8442028345451468e-05, 'epoch': 1.56}) +[2026-06-16 02:30:53,530][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.1787670999765396, 'train/grad_norm_pre_clip': 38.02421951293945, 'timing/data': 0.0003470990341156721, 'timing/model': 1.1383917520288378, 'learning_rate/qwen_vl_interface': 1.8420806842178746e-06, 'learning_rate/action_model': 1.8420806842178745e-05, 'epoch': 1.56}) +[2026-06-16 02:30:58,112][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.002510363468900323, 'train/grad_norm_pre_clip': 0.4321657717227936, 'timing/data': 0.00036366196582093835, 'timing/model': 1.294746917032171, 'learning_rate/qwen_vl_interface': 1.8399599347575034e-06, 'learning_rate/action_model': 1.8399599347575033e-05, 'epoch': 1.56}) +[2026-06-16 02:35:19,811][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.000519651104696095, 'train/grad_norm_pre_clip': 0.19591869413852692, 'eval/loss': 0.03885293006896973, 'eval/flap_precision': 0.9753086419753086, 'eval/flap_recall': 0.9753086419753086, 'eval/flap_f1': 0.9753086419753086, 'eval/flap_rate_model': 0.18, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9753086419753088, 'timing/data': 0.00034435297129675746, 'timing/model': 1.1450698539847508, 'learning_rate/qwen_vl_interface': 1.8378405870357943e-06, 'learning_rate/action_model': 1.837840587035794e-05, 'epoch': 1.56}) +[2026-06-16 02:35:19,811][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3800_state +[2026-06-16 02:35:19,814][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:35:23,397][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3800_state/model.safetensors +[2026-06-16 02:35:33,399][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3800_state/optimizer.bin +[2026-06-16 02:35:33,399][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3800_state/sampler.bin +[2026-06-16 02:35:33,399][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3800_state/sampler_1.bin +[2026-06-16 02:35:33,400][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3800_state/random_states_0.pkl +[2026-06-16 02:35:33,443][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3800 +[2026-06-16 02:35:33,444][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:35:33,450][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:35:37,890][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.11372344195842743, 'train/grad_norm_pre_clip': 22.2495174407959, 'timing/data': 0.00037933100247755647, 'timing/model': 1.2659440439892933, 'learning_rate/qwen_vl_interface': 1.8357226419239316e-06, 'learning_rate/action_model': 1.8357226419239315e-05, 'epoch': 1.56}) +[2026-06-16 02:35:42,190][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.09386925399303436, 'train/grad_norm_pre_clip': 10.790581703186035, 'timing/data': 0.00033692200668156147, 'timing/model': 1.11490409495309, 'learning_rate/qwen_vl_interface': 1.8336061002925224e-06, 'learning_rate/action_model': 1.8336061002925224e-05, 'epoch': 1.56}) +[2026-06-16 02:35:46,678][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.005346805788576603, 'train/grad_norm_pre_clip': 0.7163718938827515, 'timing/data': 0.0003223190433345735, 'timing/model': 1.121956049988512, 'learning_rate/qwen_vl_interface': 1.8314909630115985e-06, 'learning_rate/action_model': 1.8314909630115983e-05, 'epoch': 1.56}) +[2026-06-16 02:35:51,201][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.12137851864099503, 'train/grad_norm_pre_clip': 9.562602043151855, 'timing/data': 0.0006456609698943794, 'timing/model': 1.1229034699499607, 'learning_rate/qwen_vl_interface': 1.8293772309506092e-06, 'learning_rate/action_model': 1.829377230950609e-05, 'epoch': 1.56}) +[2026-06-16 02:35:55,679][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.23816706240177155, 'train/grad_norm_pre_clip': 21.051450729370117, 'timing/data': 0.0003164439694955945, 'timing/model': 1.1281238910160027, 'learning_rate/qwen_vl_interface': 1.8272649049784326e-06, 'learning_rate/action_model': 1.8272649049784323e-05, 'epoch': 1.56}) +[2026-06-16 02:36:00,177][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.006197384092956781, 'train/grad_norm_pre_clip': 1.3713738918304443, 'timing/data': 0.0004923309898003936, 'timing/model': 1.2765599699923769, 'learning_rate/qwen_vl_interface': 1.8251539859633653e-06, 'learning_rate/action_model': 1.825153985963365e-05, 'epoch': 1.56}) +[2026-06-16 02:36:04,616][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.0010398945305496454, 'train/grad_norm_pre_clip': 0.1385546177625656, 'timing/data': 0.00036474602529779077, 'timing/model': 1.2355690699769184, 'learning_rate/qwen_vl_interface': 1.8230444747731265e-06, 'learning_rate/action_model': 1.8230444747731264e-05, 'epoch': 1.56}) +[2026-06-16 02:36:08,953][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.001269880449399352, 'train/grad_norm_pre_clip': 0.42406120896339417, 'timing/data': 0.00035327603109180927, 'timing/model': 1.12513363402104, 'learning_rate/qwen_vl_interface': 1.820936372274857e-06, 'learning_rate/action_model': 1.820936372274857e-05, 'epoch': 1.56}) +[2026-06-16 02:36:13,465][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.36437395215034485, 'train/grad_norm_pre_clip': 21.839717864990234, 'timing/data': 0.0003656079643405974, 'timing/model': 1.1247155890450813, 'learning_rate/qwen_vl_interface': 1.8188296793351178e-06, 'learning_rate/action_model': 1.818829679335118e-05, 'epoch': 1.56}) +[2026-06-16 02:36:17,971][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.002924011554569006, 'train/grad_norm_pre_clip': 1.7878482341766357, 'timing/data': 0.0003336939844302833, 'timing/model': 1.1203728999826126, 'learning_rate/qwen_vl_interface': 1.8167243968198905e-06, 'learning_rate/action_model': 1.8167243968198905e-05, 'epoch': 1.57}) +[2026-06-16 02:36:22,485][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.07948441058397293, 'train/grad_norm_pre_clip': 6.939085483551025, 'timing/data': 0.0003657420165836811, 'timing/model': 1.1221156700048596, 'learning_rate/qwen_vl_interface': 1.8146205255945775e-06, 'learning_rate/action_model': 1.8146205255945777e-05, 'epoch': 1.57}) +[2026-06-16 02:36:26,881][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.11410818248987198, 'train/grad_norm_pre_clip': 27.468122482299805, 'timing/data': 0.00034467497607693076, 'timing/model': 1.1234005060396157, 'learning_rate/qwen_vl_interface': 1.8125180665240007e-06, 'learning_rate/action_model': 1.8125180665240008e-05, 'epoch': 1.57}) +[2026-06-16 02:36:31,303][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.22052700817584991, 'train/grad_norm_pre_clip': 16.849044799804688, 'timing/data': 0.0003353250212967396, 'timing/model': 1.2041390140075237, 'learning_rate/qwen_vl_interface': 1.8104170204724025e-06, 'learning_rate/action_model': 1.8104170204724024e-05, 'epoch': 1.57}) +[2026-06-16 02:36:35,664][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.03862922638654709, 'train/grad_norm_pre_clip': 11.456965446472168, 'timing/data': 0.00040003599133342505, 'timing/model': 1.1271273539750837, 'learning_rate/qwen_vl_interface': 1.8083173883034435e-06, 'learning_rate/action_model': 1.8083173883034435e-05, 'epoch': 1.57}) +[2026-06-16 02:36:40,134][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.028510047122836113, 'train/grad_norm_pre_clip': 6.8246965408325195, 'timing/data': 0.0003562929923646152, 'timing/model': 1.1271800210233778, 'learning_rate/qwen_vl_interface': 1.8062191708802013e-06, 'learning_rate/action_model': 1.8062191708802012e-05, 'epoch': 1.57}) +[2026-06-16 02:36:44,654][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.0019495454616844654, 'train/grad_norm_pre_clip': 0.574542760848999, 'timing/data': 0.0006001369911246002, 'timing/model': 1.1269478680333123, 'learning_rate/qwen_vl_interface': 1.804122369065175e-06, 'learning_rate/action_model': 1.8041223690651748e-05, 'epoch': 1.57}) +[2026-06-16 02:36:49,176][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.05146024376153946, 'train/grad_norm_pre_clip': 15.079147338867188, 'timing/data': 0.00036402198020368814, 'timing/model': 1.1294575789943337, 'learning_rate/qwen_vl_interface': 1.8020269837202806e-06, 'learning_rate/action_model': 1.8020269837202804e-05, 'epoch': 1.57}) +[2026-06-16 02:36:53,701][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.0022435076534748077, 'train/grad_norm_pre_clip': 0.7268713116645813, 'timing/data': 0.000341104983817786, 'timing/model': 1.2863594819791615, 'learning_rate/qwen_vl_interface': 1.7999330157068517e-06, 'learning_rate/action_model': 1.7999330157068515e-05, 'epoch': 1.57}) +[2026-06-16 02:36:58,068][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.10973770916461945, 'train/grad_norm_pre_clip': 8.329357147216797, 'timing/data': 0.00033166399225592613, 'timing/model': 1.1249267290113494, 'learning_rate/qwen_vl_interface': 1.7978404658856393e-06, 'learning_rate/action_model': 1.7978404658856394e-05, 'epoch': 1.57}) +[2026-06-16 02:37:02,589][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.015827439725399017, 'train/grad_norm_pre_clip': 8.040234565734863, 'timing/data': 0.0003233359893783927, 'timing/model': 1.1327237769728526, 'learning_rate/qwen_vl_interface': 1.7957493351168126e-06, 'learning_rate/action_model': 1.7957493351168126e-05, 'epoch': 1.57}) +[2026-06-16 02:37:07,025][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.09575021266937256, 'train/grad_norm_pre_clip': 13.027935981750488, 'timing/data': 0.0003842510050162673, 'timing/model': 1.1293564079678617, 'learning_rate/qwen_vl_interface': 1.7936596242599552e-06, 'learning_rate/action_model': 1.7936596242599553e-05, 'epoch': 1.57}) +[2026-06-16 02:37:11,507][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.0017790429992601275, 'train/grad_norm_pre_clip': 0.4467278718948364, 'timing/data': 0.0003969270037487149, 'timing/model': 1.1285044879768975, 'learning_rate/qwen_vl_interface': 1.7915713341740692e-06, 'learning_rate/action_model': 1.791571334174069e-05, 'epoch': 1.57}) +[2026-06-16 02:37:16,005][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.006245730444788933, 'train/grad_norm_pre_clip': 2.702453136444092, 'timing/data': 0.00040528399404138327, 'timing/model': 1.1280611929832958, 'learning_rate/qwen_vl_interface': 1.7894844657175716e-06, 'learning_rate/action_model': 1.7894844657175717e-05, 'epoch': 1.57}) +[2026-06-16 02:37:20,534][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.004928078502416611, 'train/grad_norm_pre_clip': 0.903268039226532, 'timing/data': 0.0003450080403126776, 'timing/model': 1.2879128300119191, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 1.57}) +[2026-06-16 02:37:24,919][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.0031377628911286592, 'train/grad_norm_pre_clip': 1.1667001247406006, 'timing/data': 0.000575315032619983, 'timing/model': 1.1351485009654425, 'learning_rate/qwen_vl_interface': 1.7853149971234883e-06, 'learning_rate/action_model': 1.785314997123488e-05, 'epoch': 1.57}) +[2026-06-16 02:37:29,454][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.0021942369639873505, 'train/grad_norm_pre_clip': 0.46042394638061523, 'timing/data': 0.0005732669960707426, 'timing/model': 1.1313141389982775, 'learning_rate/qwen_vl_interface': 1.7832323986998128e-06, 'learning_rate/action_model': 1.783232398699813e-05, 'epoch': 1.57}) +[2026-06-16 02:37:33,993][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.05559537932276726, 'train/grad_norm_pre_clip': 28.341533660888672, 'timing/data': 0.00033782003447413445, 'timing/model': 1.132414148945827, 'learning_rate/qwen_vl_interface': 1.7811512253333468e-06, 'learning_rate/action_model': 1.7811512253333467e-05, 'epoch': 1.57}) +[2026-06-16 02:37:38,502][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.0691126361489296, 'train/grad_norm_pre_clip': 19.990018844604492, 'timing/data': 0.00036470399936661124, 'timing/model': 1.13155586598441, 'learning_rate/qwen_vl_interface': 1.7790714778795823e-06, 'learning_rate/action_model': 1.779071477879582e-05, 'epoch': 1.57}) +[2026-06-16 02:37:42,952][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.2943580150604248, 'train/grad_norm_pre_clip': 15.571467399597168, 'timing/data': 0.0003253419999964535, 'timing/model': 1.1315306940232404, 'learning_rate/qwen_vl_interface': 1.7769931571934246e-06, 'learning_rate/action_model': 1.7769931571934246e-05, 'epoch': 1.57}) +[2026-06-16 02:37:47,388][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.0007979046204127371, 'train/grad_norm_pre_clip': 0.09580681473016739, 'timing/data': 0.0003323960117995739, 'timing/model': 1.2001340340357274, 'learning_rate/qwen_vl_interface': 1.7749162641291933e-06, 'learning_rate/action_model': 1.7749162641291933e-05, 'epoch': 1.57}) +[2026-06-16 02:37:51,764][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.0004293878737371415, 'train/grad_norm_pre_clip': 0.04795341566205025, 'timing/data': 0.0003457989660091698, 'timing/model': 1.129699406970758, 'learning_rate/qwen_vl_interface': 1.7728407995406206e-06, 'learning_rate/action_model': 1.7728407995406205e-05, 'epoch': 1.57}) +[2026-06-16 02:37:56,196][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.19499219954013824, 'train/grad_norm_pre_clip': 15.678961753845215, 'timing/data': 0.0003569719847291708, 'timing/model': 1.1327215420315042, 'learning_rate/qwen_vl_interface': 1.7707667642808513e-06, 'learning_rate/action_model': 1.7707667642808512e-05, 'epoch': 1.57}) +[2026-06-16 02:38:00,667][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.024323634803295135, 'train/grad_norm_pre_clip': 14.821978569030762, 'timing/data': 0.0004277820116840303, 'timing/model': 1.1308118979795836, 'learning_rate/qwen_vl_interface': 1.7686941592024423e-06, 'learning_rate/action_model': 1.7686941592024423e-05, 'epoch': 1.57}) +[2026-06-16 02:38:05,175][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.0010960170766338706, 'train/grad_norm_pre_clip': 0.17740961909294128, 'timing/data': 0.0005291230045258999, 'timing/model': 1.142210037971381, 'learning_rate/qwen_vl_interface': 1.7666229851573662e-06, 'learning_rate/action_model': 1.7666229851573664e-05, 'epoch': 1.58}) +[2026-06-16 02:38:09,740][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.004785767290741205, 'train/grad_norm_pre_clip': 3.2267324924468994, 'timing/data': 0.0003848859923891723, 'timing/model': 1.2888967839535326, 'learning_rate/qwen_vl_interface': 1.7645532429970006e-06, 'learning_rate/action_model': 1.7645532429970004e-05, 'epoch': 1.58}) +[2026-06-16 02:38:14,114][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.004428629297763109, 'train/grad_norm_pre_clip': 1.9765058755874634, 'timing/data': 0.00034317601239308715, 'timing/model': 1.1346012610010803, 'learning_rate/qwen_vl_interface': 1.7624849335721402e-06, 'learning_rate/action_model': 1.7624849335721402e-05, 'epoch': 1.58}) +[2026-06-16 02:38:18,648][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.04201120510697365, 'train/grad_norm_pre_clip': 13.58497428894043, 'timing/data': 0.0003279679804109037, 'timing/model': 1.1306538400240242, 'learning_rate/qwen_vl_interface': 1.7604180577329883e-06, 'learning_rate/action_model': 1.7604180577329883e-05, 'epoch': 1.58}) +[2026-06-16 02:38:23,170][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.018654605373740196, 'train/grad_norm_pre_clip': 8.506292343139648, 'timing/data': 0.0004584909765981138, 'timing/model': 1.129794681968633, 'learning_rate/qwen_vl_interface': 1.7583526163291604e-06, 'learning_rate/action_model': 1.7583526163291604e-05, 'epoch': 1.58}) +[2026-06-16 02:38:27,694][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.060243159532547, 'train/grad_norm_pre_clip': 20.94890594482422, 'timing/data': 0.00033615302527323365, 'timing/model': 1.1312622109544463, 'learning_rate/qwen_vl_interface': 1.7562886102096807e-06, 'learning_rate/action_model': 1.7562886102096807e-05, 'epoch': 1.58}) +[2026-06-16 02:38:32,251][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.005282053258270025, 'train/grad_norm_pre_clip': 2.9111971855163574, 'timing/data': 0.0004400589969009161, 'timing/model': 1.1339182449737564, 'learning_rate/qwen_vl_interface': 1.7542260402229838e-06, 'learning_rate/action_model': 1.7542260402229838e-05, 'epoch': 1.58}) +[2026-06-16 02:38:36,803][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.006670279428362846, 'train/grad_norm_pre_clip': 4.472670078277588, 'timing/data': 0.0005229460075497627, 'timing/model': 1.2994505860260688, 'learning_rate/qwen_vl_interface': 1.7521649072169152e-06, 'learning_rate/action_model': 1.752164907216915e-05, 'epoch': 1.58}) +[2026-06-16 02:38:41,200][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.00784341711550951, 'train/grad_norm_pre_clip': 2.5032126903533936, 'timing/data': 0.00036122900201007724, 'timing/model': 1.1422603669925593, 'learning_rate/qwen_vl_interface': 1.7501052120387285e-06, 'learning_rate/action_model': 1.7501052120387285e-05, 'epoch': 1.58}) +[2026-06-16 02:38:45,732][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.0066309161484241486, 'train/grad_norm_pre_clip': 3.0718436241149902, 'timing/data': 0.00040976598393172026, 'timing/model': 1.132387858000584, 'learning_rate/qwen_vl_interface': 1.7480469555350866e-06, 'learning_rate/action_model': 1.7480469555350865e-05, 'epoch': 1.58}) +[2026-06-16 02:38:50,277][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.003895574016496539, 'train/grad_norm_pre_clip': 0.5892033576965332, 'timing/data': 0.00044302496826276183, 'timing/model': 1.1459130219882354, 'learning_rate/qwen_vl_interface': 1.7459901385520619e-06, 'learning_rate/action_model': 1.745990138552062e-05, 'epoch': 1.58}) +[2026-06-16 02:38:54,832][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.00599248381331563, 'train/grad_norm_pre_clip': 5.20608377456665, 'timing/data': 0.0006019660504534841, 'timing/model': 1.1323003599536605, 'learning_rate/qwen_vl_interface': 1.7439347619351324e-06, 'learning_rate/action_model': 1.7439347619351324e-05, 'epoch': 1.58}) +[2026-06-16 02:38:59,334][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.018894137814641, 'train/grad_norm_pre_clip': 4.847188949584961, 'timing/data': 0.0003447619965299964, 'timing/model': 1.1337596309604123, 'learning_rate/qwen_vl_interface': 1.741880826529186e-06, 'learning_rate/action_model': 1.741880826529186e-05, 'epoch': 1.58}) +[2026-06-16 02:39:03,888][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.049788009375333786, 'train/grad_norm_pre_clip': 11.149510383605957, 'timing/data': 0.00033239502226933837, 'timing/model': 1.3022533900220878, 'learning_rate/qwen_vl_interface': 1.7398283331785205e-06, 'learning_rate/action_model': 1.7398283331785203e-05, 'epoch': 1.58}) +[2026-06-16 02:39:08,270][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.11455486714839935, 'train/grad_norm_pre_clip': 12.336575508117676, 'timing/data': 0.0003514129784889519, 'timing/model': 1.1372335510095581, 'learning_rate/qwen_vl_interface': 1.737777282726835e-06, 'learning_rate/action_model': 1.737777282726835e-05, 'epoch': 1.58}) +[2026-06-16 02:39:12,813][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.0025257442612200975, 'train/grad_norm_pre_clip': 0.42393675446510315, 'timing/data': 0.00033091596560552716, 'timing/model': 1.142269688018132, 'learning_rate/qwen_vl_interface': 1.7357276760172403e-06, 'learning_rate/action_model': 1.7357276760172404e-05, 'epoch': 1.58}) +[2026-06-16 02:39:17,346][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.29071423411369324, 'train/grad_norm_pre_clip': 21.18373680114746, 'timing/data': 0.0003705680137500167, 'timing/model': 1.1296105759683996, 'learning_rate/qwen_vl_interface': 1.733679513892252e-06, 'learning_rate/action_model': 1.7336795138922517e-05, 'epoch': 1.58}) +[2026-06-16 02:39:21,885][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.00024949811631813645, 'train/grad_norm_pre_clip': 0.057899996638298035, 'timing/data': 0.0003611610154621303, 'timing/model': 1.1318309299531393, 'learning_rate/qwen_vl_interface': 1.7316327971937926e-06, 'learning_rate/action_model': 1.7316327971937926e-05, 'epoch': 1.58}) +[2026-06-16 02:39:26,465][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.2364129275083542, 'train/grad_norm_pre_clip': 18.991622924804688, 'timing/data': 0.000336135970428586, 'timing/model': 1.1372331580496393, 'learning_rate/qwen_vl_interface': 1.7295875267631904e-06, 'learning_rate/action_model': 1.7295875267631904e-05, 'epoch': 1.58}) +[2026-06-16 02:39:31,002][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.003374371211975813, 'train/grad_norm_pre_clip': 0.5500953793525696, 'timing/data': 0.00040851201629266143, 'timing/model': 1.2797829890041612, 'learning_rate/qwen_vl_interface': 1.7275437034411778e-06, 'learning_rate/action_model': 1.7275437034411776e-05, 'epoch': 1.58}) +[2026-06-16 02:39:35,382][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.011711983010172844, 'train/grad_norm_pre_clip': 4.431823253631592, 'timing/data': 0.00033034803345799446, 'timing/model': 1.1322006410337053, 'learning_rate/qwen_vl_interface': 1.7255013280678941e-06, 'learning_rate/action_model': 1.725501328067894e-05, 'epoch': 1.58}) +[2026-06-16 02:39:39,929][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.0008984475280158222, 'train/grad_norm_pre_clip': 0.13788114488124847, 'timing/data': 0.00033291097497567534, 'timing/model': 1.141380517045036, 'learning_rate/qwen_vl_interface': 1.7234604014828824e-06, 'learning_rate/action_model': 1.7234604014828823e-05, 'epoch': 1.58}) +[2026-06-16 02:39:44,463][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.15666437149047852, 'train/grad_norm_pre_clip': 23.09359359741211, 'timing/data': 0.0006555370055139065, 'timing/model': 1.1326869150507264, 'learning_rate/qwen_vl_interface': 1.7214209245250906e-06, 'learning_rate/action_model': 1.7214209245250905e-05, 'epoch': 1.58}) +[2026-06-16 02:39:49,021][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.02998296730220318, 'train/grad_norm_pre_clip': 23.88518714904785, 'timing/data': 0.0006265760166570544, 'timing/model': 1.1455597580061294, 'learning_rate/qwen_vl_interface': 1.7193828980328723e-06, 'learning_rate/action_model': 1.719382898032872e-05, 'epoch': 1.58}) +[2026-06-16 02:39:53,525][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.0017615474062040448, 'train/grad_norm_pre_clip': 0.3803253471851349, 'timing/data': 0.00040429201908409595, 'timing/model': 1.1356598919956014, 'learning_rate/qwen_vl_interface': 1.71734632284398e-06, 'learning_rate/action_model': 1.7173463228439798e-05, 'epoch': 1.58}) +[2026-06-16 02:39:58,068][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.20529688894748688, 'train/grad_norm_pre_clip': 21.461578369140625, 'timing/data': 0.0003513920237310231, 'timing/model': 1.3011432830244303, 'learning_rate/qwen_vl_interface': 1.7153111997955744e-06, 'learning_rate/action_model': 1.7153111997955743e-05, 'epoch': 1.59}) +[2026-06-16 02:40:02,462][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.1269376426935196, 'train/grad_norm_pre_clip': 22.928659439086914, 'timing/data': 0.00036618701415136456, 'timing/model': 1.1391697600483894, 'learning_rate/qwen_vl_interface': 1.7132775297242175e-06, 'learning_rate/action_model': 1.7132775297242172e-05, 'epoch': 1.59}) +[2026-06-16 02:40:06,990][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.07384276390075684, 'train/grad_norm_pre_clip': 22.021270751953125, 'timing/data': 0.0003506190259940922, 'timing/model': 1.1331967019941658, 'learning_rate/qwen_vl_interface': 1.7112453134658747e-06, 'learning_rate/action_model': 1.7112453134658747e-05, 'epoch': 1.59}) +[2026-06-16 02:40:11,556][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.002751316875219345, 'train/grad_norm_pre_clip': 0.9954448342323303, 'timing/data': 0.00034269498428329825, 'timing/model': 1.14578184398124, 'learning_rate/qwen_vl_interface': 1.7092145518559129e-06, 'learning_rate/action_model': 1.709214551855913e-05, 'epoch': 1.59}) +[2026-06-16 02:40:16,087][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.04983621463179588, 'train/grad_norm_pre_clip': 26.022281646728516, 'timing/data': 0.00031415000557899475, 'timing/model': 1.1309165640268475, 'learning_rate/qwen_vl_interface': 1.7071852457291013e-06, 'learning_rate/action_model': 1.707185245729101e-05, 'epoch': 1.59}) +[2026-06-16 02:40:20,631][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.00199635187163949, 'train/grad_norm_pre_clip': 0.5678265690803528, 'timing/data': 0.00039457104867324233, 'timing/model': 1.1326445649610832, 'learning_rate/qwen_vl_interface': 1.7051573959196122e-06, 'learning_rate/action_model': 1.7051573959196122e-05, 'epoch': 1.59}) +[2026-06-16 02:40:25,176][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.006140545941889286, 'train/grad_norm_pre_clip': 7.56203031539917, 'timing/data': 0.0003731860197149217, 'timing/model': 1.2923745720181614, 'learning_rate/qwen_vl_interface': 1.7031310032610166e-06, 'learning_rate/action_model': 1.7031310032610163e-05, 'epoch': 1.59}) +[2026-06-16 02:40:29,558][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.06307344883680344, 'train/grad_norm_pre_clip': 25.923873901367188, 'timing/data': 0.0003316630027256906, 'timing/model': 1.131703088991344, 'learning_rate/qwen_vl_interface': 1.701106068586289e-06, 'learning_rate/action_model': 1.701106068586289e-05, 'epoch': 1.59}) +[2026-06-16 02:40:34,100][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.04423638433218002, 'train/grad_norm_pre_clip': 16.175806045532227, 'timing/data': 0.0004103000392206013, 'timing/model': 1.1314560010214336, 'learning_rate/qwen_vl_interface': 1.6990825927278032e-06, 'learning_rate/action_model': 1.699082592727803e-05, 'epoch': 1.59}) +[2026-06-16 02:40:38,651][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.015497767366468906, 'train/grad_norm_pre_clip': 16.473955154418945, 'timing/data': 0.0004065379616804421, 'timing/model': 1.1408658659784123, 'learning_rate/qwen_vl_interface': 1.6970605765173344e-06, 'learning_rate/action_model': 1.6970605765173346e-05, 'epoch': 1.59}) +[2026-06-16 02:40:43,217][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.00029725555214099586, 'train/grad_norm_pre_clip': 0.0689581036567688, 'timing/data': 0.0003325360012240708, 'timing/model': 1.1349693690426648, 'learning_rate/qwen_vl_interface': 1.6950400207860557e-06, 'learning_rate/action_model': 1.6950400207860557e-05, 'epoch': 1.59}) +[2026-06-16 02:40:47,722][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.003521244041621685, 'train/grad_norm_pre_clip': 1.323423147201538, 'timing/data': 0.0003469529910944402, 'timing/model': 1.1385527909733355, 'learning_rate/qwen_vl_interface': 1.6930209263645412e-06, 'learning_rate/action_model': 1.6930209263645412e-05, 'epoch': 1.59}) +[2026-06-16 02:40:52,260][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.27084478735923767, 'train/grad_norm_pre_clip': 18.214168548583984, 'timing/data': 0.00034282699925825, 'timing/model': 1.2842936280067079, 'learning_rate/qwen_vl_interface': 1.6910032940827656e-06, 'learning_rate/action_model': 1.6910032940827657e-05, 'epoch': 1.59}) +[2026-06-16 02:40:56,650][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.0012227448169142008, 'train/grad_norm_pre_clip': 0.5891431570053101, 'timing/data': 0.000318241014610976, 'timing/model': 1.1331607670290396, 'learning_rate/qwen_vl_interface': 1.6889871247701003e-06, 'learning_rate/action_model': 1.6889871247701002e-05, 'epoch': 1.59}) +[2026-06-16 02:41:01,213][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 8.373281889362261e-05, 'train/grad_norm_pre_clip': 0.009218073450028896, 'timing/data': 0.00034619803773239255, 'timing/model': 1.1414302089833654, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 1.59}) +[2026-06-16 02:41:05,761][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.0004188156744930893, 'train/grad_norm_pre_clip': 0.20150911808013916, 'timing/data': 0.0005511310300789773, 'timing/model': 1.1428715639631264, 'learning_rate/qwen_vl_interface': 1.6849591783665847e-06, 'learning_rate/action_model': 1.6849591783665847e-05, 'epoch': 1.59}) +[2026-06-16 02:41:10,310][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.002108536195009947, 'train/grad_norm_pre_clip': 0.6479674577713013, 'timing/data': 0.0004060500068590045, 'timing/model': 1.1350194030092098, 'learning_rate/qwen_vl_interface': 1.6829474029314713e-06, 'learning_rate/action_model': 1.6829474029314713e-05, 'epoch': 1.59}) +[2026-06-16 02:41:14,863][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.00821158941835165, 'train/grad_norm_pre_clip': 2.7503440380096436, 'timing/data': 0.00035612803185358644, 'timing/model': 1.1369206449599005, 'learning_rate/qwen_vl_interface': 1.6809370937769408e-06, 'learning_rate/action_model': 1.680937093776941e-05, 'epoch': 1.59}) +[2026-06-16 02:41:19,457][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.12452778220176697, 'train/grad_norm_pre_clip': 14.575456619262695, 'timing/data': 0.00034874503035098314, 'timing/model': 1.3102485840208828, 'learning_rate/qwen_vl_interface': 1.678928251729356e-06, 'learning_rate/action_model': 1.6789282517293557e-05, 'epoch': 1.59}) +[2026-06-16 02:41:23,844][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.01737590879201889, 'train/grad_norm_pre_clip': 10.22868824005127, 'timing/data': 0.00035818497417494655, 'timing/model': 1.1345590320415795, 'learning_rate/qwen_vl_interface': 1.6769208776144767e-06, 'learning_rate/action_model': 1.6769208776144764e-05, 'epoch': 1.59}) +[2026-06-16 02:41:28,386][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.0037845054175704718, 'train/grad_norm_pre_clip': 1.460340976715088, 'timing/data': 0.0006597559549845755, 'timing/model': 1.1348309089662507, 'learning_rate/qwen_vl_interface': 1.6749149722574564e-06, 'learning_rate/action_model': 1.6749149722574564e-05, 'epoch': 1.59}) +[2026-06-16 02:41:32,933][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.010885434225201607, 'train/grad_norm_pre_clip': 6.584822177886963, 'timing/data': 0.0003374239895492792, 'timing/model': 1.1361357439891435, 'learning_rate/qwen_vl_interface': 1.6729105364828486e-06, 'learning_rate/action_model': 1.6729105364828484e-05, 'epoch': 1.59}) +[2026-06-16 02:41:37,462][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.0015287664718925953, 'train/grad_norm_pre_clip': 0.27849656343460083, 'timing/data': 0.0003499399754218757, 'timing/model': 1.1334370709955692, 'learning_rate/qwen_vl_interface': 1.6709075711146e-06, 'learning_rate/action_model': 1.6709075711146e-05, 'epoch': 1.59}) +[2026-06-16 02:41:41,958][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.10181679576635361, 'train/grad_norm_pre_clip': 35.0966796875, 'timing/data': 0.00031851395033299923, 'timing/model': 1.1345395380049013, 'learning_rate/qwen_vl_interface': 1.6689060769760556e-06, 'learning_rate/action_model': 1.6689060769760557e-05, 'epoch': 1.59}) +[2026-06-16 02:41:46,509][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.0019813901744782925, 'train/grad_norm_pre_clip': 0.5311592817306519, 'timing/data': 0.0003363169962540269, 'timing/model': 1.2938580799964257, 'learning_rate/qwen_vl_interface': 1.6669060548899528e-06, 'learning_rate/action_model': 1.6669060548899525e-05, 'epoch': 1.6}) +[2026-06-16 02:41:50,896][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.010591935366392136, 'train/grad_norm_pre_clip': 4.561461448669434, 'timing/data': 0.0003603100194595754, 'timing/model': 1.1411066389991902, 'learning_rate/qwen_vl_interface': 1.6649075056784253e-06, 'learning_rate/action_model': 1.6649075056784253e-05, 'epoch': 1.6}) +[2026-06-16 02:41:55,436][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.050630997866392136, 'train/grad_norm_pre_clip': 17.828264236450195, 'timing/data': 0.00035265303449705243, 'timing/model': 1.138847717025783, 'learning_rate/qwen_vl_interface': 1.6629104301630019e-06, 'learning_rate/action_model': 1.662910430163002e-05, 'epoch': 1.6}) +[2026-06-16 02:41:59,986][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.06083991378545761, 'train/grad_norm_pre_clip': 22.84185218811035, 'timing/data': 0.0003657100023701787, 'timing/model': 1.13303733401699, 'learning_rate/qwen_vl_interface': 1.6609148291646043e-06, 'learning_rate/action_model': 1.6609148291646042e-05, 'epoch': 1.6}) +[2026-06-16 02:42:04,540][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.13838844001293182, 'train/grad_norm_pre_clip': 12.100208282470703, 'timing/data': 0.0003494840348139405, 'timing/model': 1.1413096699980088, 'learning_rate/qwen_vl_interface': 1.658920703503549e-06, 'learning_rate/action_model': 1.6589207035035488e-05, 'epoch': 1.6}) +[2026-06-16 02:42:09,101][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.002142020734027028, 'train/grad_norm_pre_clip': 0.5120121836662292, 'timing/data': 0.00034447095822542906, 'timing/model': 1.1502889029798098, 'learning_rate/qwen_vl_interface': 1.6569280539995467e-06, 'learning_rate/action_model': 1.6569280539995467e-05, 'epoch': 1.6}) +[2026-06-16 02:42:13,636][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.2726456820964813, 'train/grad_norm_pre_clip': 6.369100570678711, 'timing/data': 0.0003551029949449003, 'timing/model': 1.2889918250148185, 'learning_rate/qwen_vl_interface': 1.6549368814716977e-06, 'learning_rate/action_model': 1.6549368814716977e-05, 'epoch': 1.6}) +[2026-06-16 02:42:18,033][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.2533149719238281, 'train/grad_norm_pre_clip': 26.324670791625977, 'timing/data': 0.0004017690080218017, 'timing/model': 1.140918707009405, 'learning_rate/qwen_vl_interface': 1.6529471867385009e-06, 'learning_rate/action_model': 1.6529471867385007e-05, 'epoch': 1.6}) +[2026-06-16 02:42:22,612][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.0009798124665394425, 'train/grad_norm_pre_clip': 0.22723965346813202, 'timing/data': 0.0003334070206619799, 'timing/model': 1.1356160899740644, 'learning_rate/qwen_vl_interface': 1.6509589706178417e-06, 'learning_rate/action_model': 1.6509589706178417e-05, 'epoch': 1.6}) +[2026-06-16 02:42:27,158][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.00011320783232804388, 'train/grad_norm_pre_clip': 0.015874363481998444, 'timing/data': 0.0005596030387096107, 'timing/model': 1.1348276539938524, 'learning_rate/qwen_vl_interface': 1.648972233927002e-06, 'learning_rate/action_model': 1.648972233927002e-05, 'epoch': 1.6}) +[2026-06-16 02:42:31,703][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.0005280571058392525, 'train/grad_norm_pre_clip': 0.11592540144920349, 'timing/data': 0.00037677999353036284, 'timing/model': 1.1334231350338086, 'learning_rate/qwen_vl_interface': 1.646986977482654e-06, 'learning_rate/action_model': 1.6469869774826538e-05, 'epoch': 1.6}) +[2026-06-16 02:42:36,222][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.1071891188621521, 'train/grad_norm_pre_clip': 34.41167068481445, 'timing/data': 0.0003492439864203334, 'timing/model': 1.1380040170042776, 'learning_rate/qwen_vl_interface': 1.6450032021008618e-06, 'learning_rate/action_model': 1.6450032021008617e-05, 'epoch': 1.6}) +[2026-06-16 02:42:40,770][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.04801059514284134, 'train/grad_norm_pre_clip': 16.244403839111328, 'timing/data': 0.00037615798646584153, 'timing/model': 1.2947308900183998, 'learning_rate/qwen_vl_interface': 1.6430209085970812e-06, 'learning_rate/action_model': 1.643020908597081e-05, 'epoch': 1.6}) +[2026-06-16 02:42:45,148][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.10617638379335403, 'train/grad_norm_pre_clip': 18.57698631286621, 'timing/data': 0.000516757951118052, 'timing/model': 1.1331289989757352, 'learning_rate/qwen_vl_interface': 1.641040097786157e-06, 'learning_rate/action_model': 1.641040097786157e-05, 'epoch': 1.6}) +[2026-06-16 02:42:49,690][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.000504171708598733, 'train/grad_norm_pre_clip': 0.07629027962684631, 'timing/data': 0.0003784899599850178, 'timing/model': 1.1338765539694577, 'learning_rate/qwen_vl_interface': 1.6390607704823267e-06, 'learning_rate/action_model': 1.6390607704823265e-05, 'epoch': 1.6}) +[2026-06-16 02:42:54,244][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.12493358552455902, 'train/grad_norm_pre_clip': 30.23834991455078, 'timing/data': 0.00032713700784370303, 'timing/model': 1.1391644320101477, 'learning_rate/qwen_vl_interface': 1.6370829274992162e-06, 'learning_rate/action_model': 1.6370829274992162e-05, 'epoch': 1.6}) +[2026-06-16 02:42:58,787][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.0008623542380519211, 'train/grad_norm_pre_clip': 0.22563165426254272, 'timing/data': 0.00034721201518550515, 'timing/model': 1.1434117329772562, 'learning_rate/qwen_vl_interface': 1.6351065696498422e-06, 'learning_rate/action_model': 1.6351065696498422e-05, 'epoch': 1.6}) +[2026-06-16 02:47:20,796][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.0013346391497179866, 'train/grad_norm_pre_clip': 0.2969958782196045, 'eval/loss': 0.04548685625195503, 'eval/flap_precision': 0.9681335356600911, 'eval/flap_recall': 0.9845679012345679, 'eval/flap_f1': 0.9762815608263199, 'eval/flap_rate_model': 0.18305555555555555, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9762815608263199, 'timing/data': 0.0003259279765188694, 'timing/model': 1.1418611200060695, 'learning_rate/qwen_vl_interface': 1.6331316977466117e-06, 'learning_rate/action_model': 1.6331316977466115e-05, 'epoch': 1.6}) +[2026-06-16 02:47:20,797][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3900_state +[2026-06-16 02:47:20,799][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:47:24,401][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3900_state/model.safetensors +[2026-06-16 02:47:34,531][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3900_state/optimizer.bin +[2026-06-16 02:47:34,531][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3900_state/sampler.bin +[2026-06-16 02:47:34,531][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3900_state/sampler_1.bin +[2026-06-16 02:47:34,532][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_3900_state/random_states_0.pkl +[2026-06-16 02:47:34,575][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3900 +[2026-06-16 02:47:34,575][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:47:34,582][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:47:39,085][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.0026743346825242043, 'train/grad_norm_pre_clip': 1.4228850603103638, 'timing/data': 0.00035788503009825945, 'timing/model': 1.129989253997337, 'learning_rate/qwen_vl_interface': 1.6311583126013198e-06, 'learning_rate/action_model': 1.6311583126013198e-05, 'epoch': 1.6}) +[2026-06-16 02:47:43,557][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.28117287158966064, 'train/grad_norm_pre_clip': 26.71860122680664, 'timing/data': 0.0003383579896762967, 'timing/model': 1.2823258260032162, 'learning_rate/qwen_vl_interface': 1.6291864150251488e-06, 'learning_rate/action_model': 1.629186415025149e-05, 'epoch': 1.6}) +[2026-06-16 02:47:47,886][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.0019083517836406827, 'train/grad_norm_pre_clip': 0.8746424913406372, 'timing/data': 0.0003637589979916811, 'timing/model': 1.129911882977467, 'learning_rate/qwen_vl_interface': 1.6272160058286727e-06, 'learning_rate/action_model': 1.6272160058286726e-05, 'epoch': 1.6}) +[2026-06-16 02:47:52,375][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.0021027792245149612, 'train/grad_norm_pre_clip': 1.0945255756378174, 'timing/data': 0.00033588899532333016, 'timing/model': 1.122880287002772, 'learning_rate/qwen_vl_interface': 1.6252470858218519e-06, 'learning_rate/action_model': 1.6252470858218517e-05, 'epoch': 1.6}) +[2026-06-16 02:47:56,762][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.12042134255170822, 'train/grad_norm_pre_clip': 7.758575439453125, 'timing/data': 0.00035055901389569044, 'timing/model': 1.1224454249604605, 'learning_rate/qwen_vl_interface': 1.6232796558140357e-06, 'learning_rate/action_model': 1.6232796558140357e-05, 'epoch': 1.6}) +[2026-06-16 02:48:01,213][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.15992587804794312, 'train/grad_norm_pre_clip': 15.565143585205078, 'timing/data': 0.00031599804060533643, 'timing/model': 1.1208238970139064, 'learning_rate/qwen_vl_interface': 1.6213137166139592e-06, 'learning_rate/action_model': 1.621313716613959e-05, 'epoch': 1.6}) +[2026-06-16 02:48:05,690][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.20073091983795166, 'train/grad_norm_pre_clip': 34.849464416503906, 'timing/data': 0.00034746696474030614, 'timing/model': 1.1400297729996964, 'learning_rate/qwen_vl_interface': 1.6193492690297463e-06, 'learning_rate/action_model': 1.6193492690297464e-05, 'epoch': 1.6}) +[2026-06-16 02:48:10,198][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.11951936036348343, 'train/grad_norm_pre_clip': 17.297945022583008, 'timing/data': 0.00039814197225496173, 'timing/model': 1.2827890350017697, 'learning_rate/qwen_vl_interface': 1.6173863138689078e-06, 'learning_rate/action_model': 1.617386313868908e-05, 'epoch': 1.61}) +[2026-06-16 02:48:14,560][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.25695955753326416, 'train/grad_norm_pre_clip': 18.452808380126953, 'timing/data': 0.000363414001185447, 'timing/model': 1.1263349860091694, 'learning_rate/qwen_vl_interface': 1.6154248519383396e-06, 'learning_rate/action_model': 1.6154248519383395e-05, 'epoch': 1.61}) +[2026-06-16 02:48:19,081][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.1768021434545517, 'train/grad_norm_pre_clip': 15.589361190795898, 'timing/data': 0.00035403697984293103, 'timing/model': 1.1219574409769848, 'learning_rate/qwen_vl_interface': 1.6134648840443248e-06, 'learning_rate/action_model': 1.613464884044325e-05, 'epoch': 1.61}) +[2026-06-16 02:48:23,615][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.003342432202771306, 'train/grad_norm_pre_clip': 0.8824959397315979, 'timing/data': 0.0003707219730131328, 'timing/model': 1.1425606199773028, 'learning_rate/qwen_vl_interface': 1.6115064109925334e-06, 'learning_rate/action_model': 1.6115064109925336e-05, 'epoch': 1.61}) +[2026-06-16 02:48:28,144][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.046636756509542465, 'train/grad_norm_pre_clip': 21.529502868652344, 'timing/data': 0.0003227999550290406, 'timing/model': 1.1373586289701052, 'learning_rate/qwen_vl_interface': 1.6095494335880173e-06, 'learning_rate/action_model': 1.6095494335880174e-05, 'epoch': 1.61}) +[2026-06-16 02:48:32,661][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.0006970354588702321, 'train/grad_norm_pre_clip': 0.17026624083518982, 'timing/data': 0.00034348503686487675, 'timing/model': 1.1333504090434872, 'learning_rate/qwen_vl_interface': 1.6075939526352175e-06, 'learning_rate/action_model': 1.6075939526352175e-05, 'epoch': 1.61}) +[2026-06-16 02:48:37,184][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.12389499694108963, 'train/grad_norm_pre_clip': 14.724358558654785, 'timing/data': 0.0003346379962749779, 'timing/model': 1.293892154993955, 'learning_rate/qwen_vl_interface': 1.6056399689379576e-06, 'learning_rate/action_model': 1.6056399689379575e-05, 'epoch': 1.61}) +[2026-06-16 02:48:41,573][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.20261718332767487, 'train/grad_norm_pre_clip': 14.393217086791992, 'timing/data': 0.0003395539824850857, 'timing/model': 1.1313605249742977, 'learning_rate/qwen_vl_interface': 1.6036874832994473e-06, 'learning_rate/action_model': 1.603687483299447e-05, 'epoch': 1.61}) +[2026-06-16 02:48:46,088][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.1082197055220604, 'train/grad_norm_pre_clip': 17.389419555664062, 'timing/data': 0.0003580779884941876, 'timing/model': 1.1311726839630865, 'learning_rate/qwen_vl_interface': 1.6017364965222796e-06, 'learning_rate/action_model': 1.6017364965222793e-05, 'epoch': 1.61}) +[2026-06-16 02:48:50,590][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.06922729313373566, 'train/grad_norm_pre_clip': 8.53796672821045, 'timing/data': 0.00036006595473736525, 'timing/model': 1.126465899986215, 'learning_rate/qwen_vl_interface': 1.5997870094084305e-06, 'learning_rate/action_model': 1.5997870094084306e-05, 'epoch': 1.61}) +[2026-06-16 02:48:55,105][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.002821394708007574, 'train/grad_norm_pre_clip': 0.995740532875061, 'timing/data': 0.00035358499735593796, 'timing/model': 1.1276018189964816, 'learning_rate/qwen_vl_interface': 1.5978390227592616e-06, 'learning_rate/action_model': 1.5978390227592615e-05, 'epoch': 1.61}) +[2026-06-16 02:48:59,580][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.0021676397882401943, 'train/grad_norm_pre_clip': 0.49277734756469727, 'timing/data': 0.0004646959714591503, 'timing/model': 1.128194196964614, 'learning_rate/qwen_vl_interface': 1.5958925373755164e-06, 'learning_rate/action_model': 1.5958925373755164e-05, 'epoch': 1.61}) +[2026-06-16 02:49:04,101][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.002848817966878414, 'train/grad_norm_pre_clip': 0.44083067774772644, 'timing/data': 0.0003369190380908549, 'timing/model': 1.2901929350337014, 'learning_rate/qwen_vl_interface': 1.5939475540573213e-06, 'learning_rate/action_model': 1.5939475540573212e-05, 'epoch': 1.61}) +[2026-06-16 02:49:08,480][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.0008065262227319181, 'train/grad_norm_pre_clip': 0.26873666048049927, 'timing/data': 0.00034017901634797454, 'timing/model': 1.1340733150136657, 'learning_rate/qwen_vl_interface': 1.5920040736041867e-06, 'learning_rate/action_model': 1.5920040736041866e-05, 'epoch': 1.61}) +[2026-06-16 02:49:12,992][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.07697658985853195, 'train/grad_norm_pre_clip': 9.72276782989502, 'timing/data': 0.00034428900107741356, 'timing/model': 1.1328559149987996, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 1.61}) +[2026-06-16 02:49:17,532][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.0007578668300993741, 'train/grad_norm_pre_clip': 0.18962717056274414, 'timing/data': 0.0003553869901224971, 'timing/model': 1.13125451101223, 'learning_rate/qwen_vl_interface': 1.5881216244880407e-06, 'learning_rate/action_model': 1.5881216244880407e-05, 'epoch': 1.61}) +[2026-06-16 02:49:22,064][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.0041831792332232, 'train/grad_norm_pre_clip': 0.6229716539382935, 'timing/data': 0.0003341719857417047, 'timing/model': 1.129627613001503, 'learning_rate/qwen_vl_interface': 1.5861826574209587e-06, 'learning_rate/action_model': 1.5861826574209587e-05, 'epoch': 1.61}) +[2026-06-16 02:49:26,597][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.12182185053825378, 'train/grad_norm_pre_clip': 14.697305679321289, 'timing/data': 0.0005132409860379994, 'timing/model': 1.1360384619911201, 'learning_rate/qwen_vl_interface': 1.5842451964107916e-06, 'learning_rate/action_model': 1.5842451964107916e-05, 'epoch': 1.61}) +[2026-06-16 02:49:31,125][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.001234940136782825, 'train/grad_norm_pre_clip': 0.2029157131910324, 'timing/data': 0.0003975379513576627, 'timing/model': 1.285822620964609, 'learning_rate/qwen_vl_interface': 1.5823092422539567e-06, 'learning_rate/action_model': 1.582309242253957e-05, 'epoch': 1.61}) +[2026-06-16 02:49:35,550][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.05403700843453407, 'train/grad_norm_pre_clip': 19.07242774963379, 'timing/data': 0.0004767579957842827, 'timing/model': 1.1375436859671026, 'learning_rate/qwen_vl_interface': 1.5803747957462514e-06, 'learning_rate/action_model': 1.5803747957462514e-05, 'epoch': 1.61}) +[2026-06-16 02:49:40,130][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.026725444942712784, 'train/grad_norm_pre_clip': 22.221708297729492, 'timing/data': 0.0003693830221891403, 'timing/model': 1.1435179690015502, 'learning_rate/qwen_vl_interface': 1.5784418576828541e-06, 'learning_rate/action_model': 1.5784418576828543e-05, 'epoch': 1.61}) +[2026-06-16 02:49:44,686][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.0009994750143960118, 'train/grad_norm_pre_clip': 0.38174811005592346, 'timing/data': 0.0003390500205568969, 'timing/model': 1.1385122329811566, 'learning_rate/qwen_vl_interface': 1.576510428858322e-06, 'learning_rate/action_model': 1.5765104288583217e-05, 'epoch': 1.61}) +[2026-06-16 02:49:49,249][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.01857621967792511, 'train/grad_norm_pre_clip': 17.030839920043945, 'timing/data': 0.00034350802889093757, 'timing/model': 1.1378639340400696, 'learning_rate/qwen_vl_interface': 1.5745805100665922e-06, 'learning_rate/action_model': 1.5745805100665923e-05, 'epoch': 1.61}) +[2026-06-16 02:49:53,745][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.15655744075775146, 'train/grad_norm_pre_clip': 18.358722686767578, 'timing/data': 0.00038303103065118194, 'timing/model': 1.1356963099678978, 'learning_rate/qwen_vl_interface': 1.572652102100983e-06, 'learning_rate/action_model': 1.572652102100983e-05, 'epoch': 1.61}) +[2026-06-16 02:49:58,302][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.0010591312311589718, 'train/grad_norm_pre_clip': 0.4059459865093231, 'timing/data': 0.00047133094631135464, 'timing/model': 1.2958922749967314, 'learning_rate/qwen_vl_interface': 1.5707252057541898e-06, 'learning_rate/action_model': 1.5707252057541896e-05, 'epoch': 1.62}) +[2026-06-16 02:50:02,684][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.3494932949542999, 'train/grad_norm_pre_clip': 15.54111099243164, 'timing/data': 0.0003402549773454666, 'timing/model': 1.1345966139924712, 'learning_rate/qwen_vl_interface': 1.5687998218182849e-06, 'learning_rate/action_model': 1.5687998218182847e-05, 'epoch': 1.62}) +[2026-06-16 02:50:07,244][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.003534831339493394, 'train/grad_norm_pre_clip': 1.882790207862854, 'timing/data': 0.00040908996015787125, 'timing/model': 1.1372509780339897, 'learning_rate/qwen_vl_interface': 1.5668759510847235e-06, 'learning_rate/action_model': 1.5668759510847233e-05, 'epoch': 1.62}) +[2026-06-16 02:50:11,780][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.21987979114055634, 'train/grad_norm_pre_clip': 29.768939971923828, 'timing/data': 0.0004664060543291271, 'timing/model': 1.133418631972745, 'learning_rate/qwen_vl_interface': 1.5649535943443331e-06, 'learning_rate/action_model': 1.564953594344333e-05, 'epoch': 1.62}) +[2026-06-16 02:50:16,307][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.012117130681872368, 'train/grad_norm_pre_clip': 5.347959041595459, 'timing/data': 0.00036114000249654055, 'timing/model': 1.1326843639835715, 'learning_rate/qwen_vl_interface': 1.5630327523873244e-06, 'learning_rate/action_model': 1.563032752387324e-05, 'epoch': 1.62}) +[2026-06-16 02:50:20,865][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.13850641250610352, 'train/grad_norm_pre_clip': 7.189762115478516, 'timing/data': 0.0003337160451337695, 'timing/model': 1.131662082974799, 'learning_rate/qwen_vl_interface': 1.561113426003282e-06, 'learning_rate/action_model': 1.561113426003282e-05, 'epoch': 1.62}) +[2026-06-16 02:50:25,429][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.030450042337179184, 'train/grad_norm_pre_clip': 11.654375076293945, 'timing/data': 0.00043619098141789436, 'timing/model': 1.3101513549918309, 'learning_rate/qwen_vl_interface': 1.5591956159811686e-06, 'learning_rate/action_model': 1.5591956159811687e-05, 'epoch': 1.62}) +[2026-06-16 02:50:29,845][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.00014519803517032415, 'train/grad_norm_pre_clip': 0.028109949082136154, 'timing/data': 0.00035289995139464736, 'timing/model': 1.1372003130381927, 'learning_rate/qwen_vl_interface': 1.5572793231093235e-06, 'learning_rate/action_model': 1.5572793231093236e-05, 'epoch': 1.62}) +[2026-06-16 02:50:34,391][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.01845189370214939, 'train/grad_norm_pre_clip': 6.846093654632568, 'timing/data': 0.0004083589883521199, 'timing/model': 1.1329550520167686, 'learning_rate/qwen_vl_interface': 1.5553645481754626e-06, 'learning_rate/action_model': 1.5553645481754626e-05, 'epoch': 1.62}) +[2026-06-16 02:50:38,963][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.08607496321201324, 'train/grad_norm_pre_clip': 22.505807876586914, 'timing/data': 0.0003647530102171004, 'timing/model': 1.1413447600207292, 'learning_rate/qwen_vl_interface': 1.5534512919666775e-06, 'learning_rate/action_model': 1.5534512919666776e-05, 'epoch': 1.62}) +[2026-06-16 02:50:43,512][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.020412635058164597, 'train/grad_norm_pre_clip': 10.899724960327148, 'timing/data': 0.0003707599826157093, 'timing/model': 1.1300359269953333, 'learning_rate/qwen_vl_interface': 1.551539555269435e-06, 'learning_rate/action_model': 1.5515395552694348e-05, 'epoch': 1.62}) +[2026-06-16 02:50:48,022][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.11435884237289429, 'train/grad_norm_pre_clip': 36.072235107421875, 'timing/data': 0.0003299249801784754, 'timing/model': 1.1384348009596579, 'learning_rate/qwen_vl_interface': 1.5496293388695787e-06, 'learning_rate/action_model': 1.5496293388695786e-05, 'epoch': 1.62}) +[2026-06-16 02:50:52,573][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.016180027276277542, 'train/grad_norm_pre_clip': 4.968280792236328, 'timing/data': 0.00034224503906443715, 'timing/model': 1.293170958000701, 'learning_rate/qwen_vl_interface': 1.547720643552327e-06, 'learning_rate/action_model': 1.547720643552327e-05, 'epoch': 1.62}) +[2026-06-16 02:50:56,961][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.0028834010008722544, 'train/grad_norm_pre_clip': 0.8445191383361816, 'timing/data': 0.00033686798997223377, 'timing/model': 1.138308770954609, 'learning_rate/qwen_vl_interface': 1.5458134701022706e-06, 'learning_rate/action_model': 1.5458134701022705e-05, 'epoch': 1.62}) +[2026-06-16 02:51:01,512][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.002056400990113616, 'train/grad_norm_pre_clip': 0.6205880045890808, 'timing/data': 0.00039415701758116484, 'timing/model': 1.1366047599585727, 'learning_rate/qwen_vl_interface': 1.5439078193033765e-06, 'learning_rate/action_model': 1.5439078193033766e-05, 'epoch': 1.62}) +[2026-06-16 02:51:06,080][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.0038296228740364313, 'train/grad_norm_pre_clip': 1.7947320938110352, 'timing/data': 0.00035163998836651444, 'timing/model': 1.1520335850073025, 'learning_rate/qwen_vl_interface': 1.5420036919389872e-06, 'learning_rate/action_model': 1.5420036919389873e-05, 'epoch': 1.62}) +[2026-06-16 02:51:10,627][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.0991610661149025, 'train/grad_norm_pre_clip': 12.451192855834961, 'timing/data': 0.000326982990372926, 'timing/model': 1.1345497189904563, 'learning_rate/qwen_vl_interface': 1.540101088791817e-06, 'learning_rate/action_model': 1.5401010887918168e-05, 'epoch': 1.62}) +[2026-06-16 02:51:15,168][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.09784092754125595, 'train/grad_norm_pre_clip': 18.394136428833008, 'timing/data': 0.00034562195651233196, 'timing/model': 1.1405212270328775, 'learning_rate/qwen_vl_interface': 1.5382000106439532e-06, 'learning_rate/action_model': 1.5382000106439532e-05, 'epoch': 1.62}) +[2026-06-16 02:51:19,745][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.017144465819001198, 'train/grad_norm_pre_clip': 13.585189819335938, 'timing/data': 0.0003867539926432073, 'timing/model': 1.3325002860510722, 'learning_rate/qwen_vl_interface': 1.5363004582768584e-06, 'learning_rate/action_model': 1.5363004582768584e-05, 'epoch': 1.62}) +[2026-06-16 02:51:24,124][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.07373174279928207, 'train/grad_norm_pre_clip': 18.86129379272461, 'timing/data': 0.0003692359896376729, 'timing/model': 1.1371302820043638, 'learning_rate/qwen_vl_interface': 1.534402432471367e-06, 'learning_rate/action_model': 1.534402432471367e-05, 'epoch': 1.62}) +[2026-06-16 02:51:28,677][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.035386379808187485, 'train/grad_norm_pre_clip': 29.43949317932129, 'timing/data': 0.00033044000156223774, 'timing/model': 1.1536466089892201, 'learning_rate/qwen_vl_interface': 1.532505934007685e-06, 'learning_rate/action_model': 1.532505934007685e-05, 'epoch': 1.62}) +[2026-06-16 02:51:33,224][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.0003923124459106475, 'train/grad_norm_pre_clip': 0.044133227318525314, 'timing/data': 0.0003537419834174216, 'timing/model': 1.1445834820042364, 'learning_rate/qwen_vl_interface': 1.5306109636653915e-06, 'learning_rate/action_model': 1.5306109636653914e-05, 'epoch': 1.62}) +[2026-06-16 02:51:37,762][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.4276266098022461, 'train/grad_norm_pre_clip': 28.81580352783203, 'timing/data': 0.00033799302764236927, 'timing/model': 1.1315624049748294, 'learning_rate/qwen_vl_interface': 1.5287175222234383e-06, 'learning_rate/action_model': 1.528717522223438e-05, 'epoch': 1.62}) +[2026-06-16 02:51:42,265][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.006823193281888962, 'train/grad_norm_pre_clip': 2.340376377105713, 'timing/data': 0.00036036496749147773, 'timing/model': 1.1355422439519316, 'learning_rate/qwen_vl_interface': 1.5268256104601457e-06, 'learning_rate/action_model': 1.5268256104601454e-05, 'epoch': 1.62}) +[2026-06-16 02:51:46,803][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.10491741448640823, 'train/grad_norm_pre_clip': 7.978515148162842, 'timing/data': 0.0003940850147046149, 'timing/model': 1.2847996720229276, 'learning_rate/qwen_vl_interface': 1.5249352291532075e-06, 'learning_rate/action_model': 1.5249352291532073e-05, 'epoch': 1.63}) +[2026-06-16 02:51:51,182][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.08252406120300293, 'train/grad_norm_pre_clip': 10.095422744750977, 'timing/data': 0.0004712340305559337, 'timing/model': 1.1337459679925814, 'learning_rate/qwen_vl_interface': 1.523046379079689e-06, 'learning_rate/action_model': 1.523046379079689e-05, 'epoch': 1.63}) +[2026-06-16 02:51:55,718][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.003929327707737684, 'train/grad_norm_pre_clip': 2.1366400718688965, 'timing/data': 0.00040165201062336564, 'timing/model': 1.1337150130420923, 'learning_rate/qwen_vl_interface': 1.5211590610160244e-06, 'learning_rate/action_model': 1.5211590610160242e-05, 'epoch': 1.63}) +[2026-06-16 02:52:00,267][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.0003156313905492425, 'train/grad_norm_pre_clip': 0.061458900570869446, 'timing/data': 0.0003878339775837958, 'timing/model': 1.135071863012854, 'learning_rate/qwen_vl_interface': 1.5192732757380193e-06, 'learning_rate/action_model': 1.5192732757380194e-05, 'epoch': 1.63}) +[2026-06-16 02:52:04,806][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.23145543038845062, 'train/grad_norm_pre_clip': 24.295398712158203, 'timing/data': 0.0003241740050725639, 'timing/model': 1.1368515700451098, 'learning_rate/qwen_vl_interface': 1.517389024020849e-06, 'learning_rate/action_model': 1.5173890240208488e-05, 'epoch': 1.63}) +[2026-06-16 02:52:09,345][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.00313782156445086, 'train/grad_norm_pre_clip': 1.0016578435897827, 'timing/data': 0.0003415039973333478, 'timing/model': 1.1342758300015703, 'learning_rate/qwen_vl_interface': 1.515506306639057e-06, 'learning_rate/action_model': 1.515506306639057e-05, 'epoch': 1.63}) +[2026-06-16 02:52:13,887][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.01403056737035513, 'train/grad_norm_pre_clip': 6.597569465637207, 'timing/data': 0.00036423298297449946, 'timing/model': 1.2929008249775507, 'learning_rate/qwen_vl_interface': 1.5136251243665589e-06, 'learning_rate/action_model': 1.5136251243665589e-05, 'epoch': 1.63}) +[2026-06-16 02:52:18,286][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.0977875217795372, 'train/grad_norm_pre_clip': 21.535648345947266, 'timing/data': 0.00032254098914563656, 'timing/model': 1.1322372969589196, 'learning_rate/qwen_vl_interface': 1.5117454779766372e-06, 'learning_rate/action_model': 1.5117454779766371e-05, 'epoch': 1.63}) +[2026-06-16 02:52:22,832][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.007437991909682751, 'train/grad_norm_pre_clip': 4.079287052154541, 'timing/data': 0.0003974270075559616, 'timing/model': 1.1345213510212488, 'learning_rate/qwen_vl_interface': 1.5098673682419428e-06, 'learning_rate/action_model': 1.5098673682419429e-05, 'epoch': 1.63}) +[2026-06-16 02:52:27,377][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.16124902665615082, 'train/grad_norm_pre_clip': 17.032331466674805, 'timing/data': 0.00033951696241274476, 'timing/model': 1.1353993049706332, 'learning_rate/qwen_vl_interface': 1.5079907959344977e-06, 'learning_rate/action_model': 1.5079907959344978e-05, 'epoch': 1.63}) +[2026-06-16 02:52:31,930][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.010294810868799686, 'train/grad_norm_pre_clip': 3.2495548725128174, 'timing/data': 0.0006383989821188152, 'timing/model': 1.1347484769648872, 'learning_rate/qwen_vl_interface': 1.5061157618256873e-06, 'learning_rate/action_model': 1.5061157618256871e-05, 'epoch': 1.63}) +[2026-06-16 02:52:36,427][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.004049060400575399, 'train/grad_norm_pre_clip': 1.7256382703781128, 'timing/data': 0.00033222499769181013, 'timing/model': 1.1350115140085109, 'learning_rate/qwen_vl_interface': 1.5042422666862675e-06, 'learning_rate/action_model': 1.5042422666862673e-05, 'epoch': 1.63}) +[2026-06-16 02:52:40,983][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.17449156939983368, 'train/grad_norm_pre_clip': 35.14300537109375, 'timing/data': 0.0003138420288451016, 'timing/model': 1.304216006945353, 'learning_rate/qwen_vl_interface': 1.5023703112863628e-06, 'learning_rate/action_model': 1.5023703112863626e-05, 'epoch': 1.63}) +[2026-06-16 02:52:45,380][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.003356502391397953, 'train/grad_norm_pre_clip': 0.9472838640213013, 'timing/data': 0.0003394859959371388, 'timing/model': 1.1345683690160513, 'learning_rate/qwen_vl_interface': 1.5004998963954629e-06, 'learning_rate/action_model': 1.5004998963954628e-05, 'epoch': 1.63}) +[2026-06-16 02:52:49,934][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.0019331932999193668, 'train/grad_norm_pre_clip': 0.2720772624015808, 'timing/data': 0.0003650469589047134, 'timing/model': 1.1392765660420991, 'learning_rate/qwen_vl_interface': 1.498631022782424e-06, 'learning_rate/action_model': 1.4986310227824241e-05, 'epoch': 1.63}) +[2026-06-16 02:52:54,492][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.022219112142920494, 'train/grad_norm_pre_clip': 7.9766716957092285, 'timing/data': 0.00034475699067115784, 'timing/model': 1.1446964250062592, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 1.63}) +[2026-06-16 02:52:59,021][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.08909690380096436, 'train/grad_norm_pre_clip': 9.671957015991211, 'timing/data': 0.00039631297113373876, 'timing/model': 1.1324666989967227, 'learning_rate/qwen_vl_interface': 1.4948979024621915e-06, 'learning_rate/action_model': 1.4948979024621914e-05, 'epoch': 1.63}) +[2026-06-16 02:53:03,589][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.016140272840857506, 'train/grad_norm_pre_clip': 9.928119659423828, 'timing/data': 0.00033851596526801586, 'timing/model': 1.1437675430206582, 'learning_rate/qwen_vl_interface': 1.493033657289542e-06, 'learning_rate/action_model': 1.493033657289542e-05, 'epoch': 1.63}) +[2026-06-16 02:53:08,149][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.003493737895041704, 'train/grad_norm_pre_clip': 1.6250981092453003, 'timing/data': 0.0005598719581030309, 'timing/model': 1.3061583369853906, 'learning_rate/qwen_vl_interface': 1.4911709564638428e-06, 'learning_rate/action_model': 1.4911709564638427e-05, 'epoch': 1.63}) +[2026-06-16 02:53:12,549][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.09485939890146255, 'train/grad_norm_pre_clip': 19.5400447845459, 'timing/data': 0.0003657989436760545, 'timing/model': 1.1403057469869964, 'learning_rate/qwen_vl_interface': 1.4893098007507817e-06, 'learning_rate/action_model': 1.4893098007507816e-05, 'epoch': 1.63}) +[2026-06-16 02:53:17,111][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.006572658661752939, 'train/grad_norm_pre_clip': 2.0979366302490234, 'timing/data': 0.0003677440108731389, 'timing/model': 1.139234016998671, 'learning_rate/qwen_vl_interface': 1.487450190915406e-06, 'learning_rate/action_model': 1.487450190915406e-05, 'epoch': 1.63}) +[2026-06-16 02:53:21,663][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.17278920114040375, 'train/grad_norm_pre_clip': 13.185651779174805, 'timing/data': 0.00042221701005473733, 'timing/model': 1.1358276560204104, 'learning_rate/qwen_vl_interface': 1.4855921277221336e-06, 'learning_rate/action_model': 1.4855921277221336e-05, 'epoch': 1.63}) +[2026-06-16 02:53:26,215][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.009140239097177982, 'train/grad_norm_pre_clip': 4.975454807281494, 'timing/data': 0.00034679396776482463, 'timing/model': 1.1390418469673023, 'learning_rate/qwen_vl_interface': 1.4837356119347423e-06, 'learning_rate/action_model': 1.4837356119347423e-05, 'epoch': 1.63}) +[2026-06-16 02:53:30,710][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.12805500626564026, 'train/grad_norm_pre_clip': 8.388888359069824, 'timing/data': 0.000709985033608973, 'timing/model': 1.1315771179506555, 'learning_rate/qwen_vl_interface': 1.4818806443163774e-06, 'learning_rate/action_model': 1.4818806443163775e-05, 'epoch': 1.63}) +[2026-06-16 02:53:35,271][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.12496036291122437, 'train/grad_norm_pre_clip': 9.79024600982666, 'timing/data': 0.0003593569854274392, 'timing/model': 1.2954482209752314, 'learning_rate/qwen_vl_interface': 1.480027225629547e-06, 'learning_rate/action_model': 1.480027225629547e-05, 'epoch': 1.63}) +[2026-06-16 02:53:39,663][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.004735273774713278, 'train/grad_norm_pre_clip': 3.373098611831665, 'timing/data': 0.00033849896863102913, 'timing/model': 1.1439027409651317, 'learning_rate/qwen_vl_interface': 1.4781753566361182e-06, 'learning_rate/action_model': 1.4781753566361182e-05, 'epoch': 1.64}) +[2026-06-16 02:53:44,215][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.15270666778087616, 'train/grad_norm_pre_clip': 17.987266540527344, 'timing/data': 0.0003308110171929002, 'timing/model': 1.1362972279894166, 'learning_rate/qwen_vl_interface': 1.4763250380973255e-06, 'learning_rate/action_model': 1.4763250380973254e-05, 'epoch': 1.64}) +[2026-06-16 02:53:48,779][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.0018363145645707846, 'train/grad_norm_pre_clip': 0.2493256777524948, 'timing/data': 0.00033887196332216263, 'timing/model': 1.1333839489961974, 'learning_rate/qwen_vl_interface': 1.4744762707737662e-06, 'learning_rate/action_model': 1.4744762707737662e-05, 'epoch': 1.64}) +[2026-06-16 02:53:53,341][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.1932453215122223, 'train/grad_norm_pre_clip': 24.84287452697754, 'timing/data': 0.0004206449957564473, 'timing/model': 1.1332104729954153, 'learning_rate/qwen_vl_interface': 1.4726290554253977e-06, 'learning_rate/action_model': 1.4726290554253977e-05, 'epoch': 1.64}) +[2026-06-16 02:53:57,895][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.004769599065184593, 'train/grad_norm_pre_clip': 3.6730990409851074, 'timing/data': 0.0003812870127148926, 'timing/model': 1.1342341609997675, 'learning_rate/qwen_vl_interface': 1.4707833928115406e-06, 'learning_rate/action_model': 1.4707833928115405e-05, 'epoch': 1.64}) +[2026-06-16 02:54:02,431][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.0012960991589352489, 'train/grad_norm_pre_clip': 0.3485231399536133, 'timing/data': 0.0003287229919806123, 'timing/model': 1.2946009269799106, 'learning_rate/qwen_vl_interface': 1.4689392836908775e-06, 'learning_rate/action_model': 1.4689392836908774e-05, 'epoch': 1.64}) +[2026-06-16 02:54:06,822][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.11011376976966858, 'train/grad_norm_pre_clip': 16.846010208129883, 'timing/data': 0.000369918008800596, 'timing/model': 1.1363759330124594, 'learning_rate/qwen_vl_interface': 1.4670967288214514e-06, 'learning_rate/action_model': 1.4670967288214513e-05, 'epoch': 1.64}) +[2026-06-16 02:54:11,372][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.0035213616210967302, 'train/grad_norm_pre_clip': 0.732327938079834, 'timing/data': 0.0003481319872662425, 'timing/model': 1.1309098349884152, 'learning_rate/qwen_vl_interface': 1.4652557289606687e-06, 'learning_rate/action_model': 1.4652557289606685e-05, 'epoch': 1.64}) +[2026-06-16 02:54:15,940][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.23284786939620972, 'train/grad_norm_pre_clip': 14.216591835021973, 'timing/data': 0.00034204096300527453, 'timing/model': 1.13627690798603, 'learning_rate/qwen_vl_interface': 1.4634162848652918e-06, 'learning_rate/action_model': 1.4634162848652919e-05, 'epoch': 1.64}) +[2026-06-16 02:54:20,477][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.1888846904039383, 'train/grad_norm_pre_clip': 22.745765686035156, 'timing/data': 0.00033221300691366196, 'timing/model': 1.1328728700173087, 'learning_rate/qwen_vl_interface': 1.4615783972914494e-06, 'learning_rate/action_model': 1.4615783972914493e-05, 'epoch': 1.64}) +[2026-06-16 02:54:24,984][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.029648974537849426, 'train/grad_norm_pre_clip': 11.226115226745605, 'timing/data': 0.000346112996339798, 'timing/model': 1.1350200150045566, 'learning_rate/qwen_vl_interface': 1.459742066994626e-06, 'learning_rate/action_model': 1.4597420669946261e-05, 'epoch': 1.64}) +[2026-06-16 02:54:29,540][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.11627226322889328, 'train/grad_norm_pre_clip': 19.839351654052734, 'timing/data': 0.00034034199779853225, 'timing/model': 1.2903274900163524, 'learning_rate/qwen_vl_interface': 1.4579072947296687e-06, 'learning_rate/action_model': 1.4579072947296687e-05, 'epoch': 1.64}) +[2026-06-16 02:54:33,932][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.21740023791790009, 'train/grad_norm_pre_clip': 5.722703456878662, 'timing/data': 0.00035925995325669646, 'timing/model': 1.1356208949582651, 'learning_rate/qwen_vl_interface': 1.4560740812507823e-06, 'learning_rate/action_model': 1.4560740812507822e-05, 'epoch': 1.64}) +[2026-06-16 02:54:38,478][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.00251830299384892, 'train/grad_norm_pre_clip': 1.151993751525879, 'timing/data': 0.0003544659703038633, 'timing/model': 1.1365670600207523, 'learning_rate/qwen_vl_interface': 1.4542424273115322e-06, 'learning_rate/action_model': 1.454242427311532e-05, 'epoch': 1.64}) +[2026-06-16 02:54:43,019][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.0714208111166954, 'train/grad_norm_pre_clip': 23.961936950683594, 'timing/data': 0.00034495798172429204, 'timing/model': 1.1405184339964762, 'learning_rate/qwen_vl_interface': 1.4524123336648419e-06, 'learning_rate/action_model': 1.4524123336648418e-05, 'epoch': 1.64}) +[2026-06-16 02:54:47,562][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.0005943342694081366, 'train/grad_norm_pre_clip': 0.4390925168991089, 'timing/data': 0.0003503920161165297, 'timing/model': 1.1342353309737518, 'learning_rate/qwen_vl_interface': 1.4505838010629942e-06, 'learning_rate/action_model': 1.4505838010629941e-05, 'epoch': 1.64}) +[2026-06-16 02:54:52,116][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.23187090456485748, 'train/grad_norm_pre_clip': 42.89720916748047, 'timing/data': 0.0003374529769644141, 'timing/model': 1.1340106669813395, 'learning_rate/qwen_vl_interface': 1.4487568302576284e-06, 'learning_rate/action_model': 1.4487568302576282e-05, 'epoch': 1.64}) +[2026-06-16 02:54:56,685][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.010656234808266163, 'train/grad_norm_pre_clip': 3.1157641410827637, 'timing/data': 0.0003692740574479103, 'timing/model': 1.2968345579574816, 'learning_rate/qwen_vl_interface': 1.446931421999746e-06, 'learning_rate/action_model': 1.4469314219997459e-05, 'epoch': 1.64}) +[2026-06-16 02:55:01,080][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.0006171267596073449, 'train/grad_norm_pre_clip': 0.10203936696052551, 'timing/data': 0.00033889198675751686, 'timing/model': 1.1386611789930612, 'learning_rate/qwen_vl_interface': 1.4451075770397004e-06, 'learning_rate/action_model': 1.4451075770397004e-05, 'epoch': 1.64}) +[2026-06-16 02:59:23,320][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.19253110885620117, 'train/grad_norm_pre_clip': 9.297383308410645, 'eval/loss': 0.04228421300649643, 'eval/flap_precision': 0.9636913767019667, 'eval/flap_recall': 0.9830246913580247, 'eval/flap_f1': 0.9732620320855616, 'eval/flap_rate_model': 0.1836111111111111, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9732620320855616, 'timing/data': 0.00035887298872694373, 'timing/model': 1.137353734986391, 'learning_rate/qwen_vl_interface': 1.4432852961272068e-06, 'learning_rate/action_model': 1.4432852961272066e-05, 'epoch': 1.64}) +[2026-06-16 02:59:23,321][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4000_state +[2026-06-16 02:59:23,323][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:59:26,896][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4000_state/model.safetensors +[2026-06-16 02:59:36,990][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4000_state/optimizer.bin +[2026-06-16 02:59:36,991][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4000_state/sampler.bin +[2026-06-16 02:59:36,991][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-16 02:59:36,992][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-16 02:59:37,033][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-16 02:59:37,033][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:59:37,039][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:59:41,500][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.14804185926914215, 'train/grad_norm_pre_clip': 8.73948860168457, 'timing/data': 0.0005894400528632104, 'timing/model': 1.117442119983025, 'learning_rate/qwen_vl_interface': 1.4414645800113353e-06, 'learning_rate/action_model': 1.4414645800113352e-05, 'epoch': 1.64}) +[2026-06-16 02:59:45,883][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.002767479745671153, 'train/grad_norm_pre_clip': 1.6502587795257568, 'timing/data': 0.00035112601472064853, 'timing/model': 1.1231493990053423, 'learning_rate/qwen_vl_interface': 1.4396454294405138e-06, 'learning_rate/action_model': 1.4396454294405136e-05, 'epoch': 1.64}) +[2026-06-16 02:59:50,300][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 0.0025958209298551083, 'train/grad_norm_pre_clip': 0.9322364330291748, 'timing/data': 0.000771598017308861, 'timing/model': 1.11531693598954, 'learning_rate/qwen_vl_interface': 1.4378278451625271e-06, 'learning_rate/action_model': 1.4378278451625271e-05, 'epoch': 1.64}) +[2026-06-16 02:59:54,751][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.06387093663215637, 'train/grad_norm_pre_clip': 18.319936752319336, 'timing/data': 0.0003558180178515613, 'timing/model': 1.1197257839958183, 'learning_rate/qwen_vl_interface': 1.4360118279245143e-06, 'learning_rate/action_model': 1.4360118279245144e-05, 'epoch': 1.64}) +[2026-06-16 02:59:59,221][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.06336143612861633, 'train/grad_norm_pre_clip': 22.605348587036133, 'timing/data': 0.00040298502426594496, 'timing/model': 1.2749101970111951, 'learning_rate/qwen_vl_interface': 1.434197378472973e-06, 'learning_rate/action_model': 1.4341973784729729e-05, 'epoch': 1.65}) +[2026-06-16 03:00:03,570][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.024310335516929626, 'train/grad_norm_pre_clip': 25.456602096557617, 'timing/data': 0.0003609189880080521, 'timing/model': 1.1359391969745047, 'learning_rate/qwen_vl_interface': 1.4323844975537535e-06, 'learning_rate/action_model': 1.4323844975537534e-05, 'epoch': 1.65}) +[2026-06-16 03:00:08,084][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.003293040907010436, 'train/grad_norm_pre_clip': 0.45790714025497437, 'timing/data': 0.0003662079689092934, 'timing/model': 1.1219258269993588, 'learning_rate/qwen_vl_interface': 1.4305731859120628e-06, 'learning_rate/action_model': 1.4305731859120628e-05, 'epoch': 1.65}) +[2026-06-16 03:00:12,595][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.14833398163318634, 'train/grad_norm_pre_clip': 8.67187786102295, 'timing/data': 0.00033480097772553563, 'timing/model': 1.1218339300248772, 'learning_rate/qwen_vl_interface': 1.4287634442924647e-06, 'learning_rate/action_model': 1.4287634442924646e-05, 'epoch': 1.65}) +[2026-06-16 03:00:17,047][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.0004124965635128319, 'train/grad_norm_pre_clip': 0.16387483477592468, 'timing/data': 0.00034604198299348354, 'timing/model': 1.1214701000135392, 'learning_rate/qwen_vl_interface': 1.4269552734388725e-06, 'learning_rate/action_model': 1.4269552734388724e-05, 'epoch': 1.65}) +[2026-06-16 03:00:21,463][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.000846625305712223, 'train/grad_norm_pre_clip': 0.22355270385742188, 'timing/data': 0.0003784839645959437, 'timing/model': 1.1230296529829502, 'learning_rate/qwen_vl_interface': 1.4251486740945584e-06, 'learning_rate/action_model': 1.4251486740945583e-05, 'epoch': 1.65}) +[2026-06-16 03:00:25,898][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.000246500683715567, 'train/grad_norm_pre_clip': 0.04281032085418701, 'timing/data': 0.0005387419951148331, 'timing/model': 1.21049284102628, 'learning_rate/qwen_vl_interface': 1.423343647002147e-06, 'learning_rate/action_model': 1.423343647002147e-05, 'epoch': 1.65}) +[2026-06-16 03:00:30,253][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.16842743754386902, 'train/grad_norm_pre_clip': 25.341455459594727, 'timing/data': 0.00039353902684524655, 'timing/model': 1.1301405630074441, 'learning_rate/qwen_vl_interface': 1.421540192903617e-06, 'learning_rate/action_model': 1.421540192903617e-05, 'epoch': 1.65}) +[2026-06-16 03:00:34,679][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 0.09593559056520462, 'train/grad_norm_pre_clip': 13.30395221710205, 'timing/data': 0.00036040303530171514, 'timing/model': 1.1277528229984455, 'learning_rate/qwen_vl_interface': 1.4197383125402997e-06, 'learning_rate/action_model': 1.4197383125402997e-05, 'epoch': 1.65}) +[2026-06-16 03:00:39,146][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.00018527393694967031, 'train/grad_norm_pre_clip': 0.019546374678611755, 'timing/data': 0.00033362500835210085, 'timing/model': 1.126718357030768, 'learning_rate/qwen_vl_interface': 1.4179380066528807e-06, 'learning_rate/action_model': 1.4179380066528807e-05, 'epoch': 1.65}) +[2026-06-16 03:00:43,630][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.03489501774311066, 'train/grad_norm_pre_clip': 16.435405731201172, 'timing/data': 0.0005095769884064794, 'timing/model': 1.1274642649805173, 'learning_rate/qwen_vl_interface': 1.4161392759813984e-06, 'learning_rate/action_model': 1.4161392759813983e-05, 'epoch': 1.65}) +[2026-06-16 03:00:48,170][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.0875421017408371, 'train/grad_norm_pre_clip': 19.34577178955078, 'timing/data': 0.0003978089662268758, 'timing/model': 1.2960357189876959, 'learning_rate/qwen_vl_interface': 1.4143421212652411e-06, 'learning_rate/action_model': 1.414342121265241e-05, 'epoch': 1.65}) +[2026-06-16 03:00:52,535][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.11142295598983765, 'train/grad_norm_pre_clip': 6.462207794189453, 'timing/data': 0.0003698350046761334, 'timing/model': 1.12976569699822, 'learning_rate/qwen_vl_interface': 1.4125465432431529e-06, 'learning_rate/action_model': 1.4125465432431529e-05, 'epoch': 1.65}) +[2026-06-16 03:00:57,041][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.15485382080078125, 'train/grad_norm_pre_clip': 11.025102615356445, 'timing/data': 0.0005470019532367587, 'timing/model': 1.1283227669773623, 'learning_rate/qwen_vl_interface': 1.4107525426532273e-06, 'learning_rate/action_model': 1.4107525426532272e-05, 'epoch': 1.65}) +[2026-06-16 03:01:01,560][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.0015924567123875022, 'train/grad_norm_pre_clip': 0.5232899785041809, 'timing/data': 0.0003192930016666651, 'timing/model': 1.1269011049880646, 'learning_rate/qwen_vl_interface': 1.4089601202329116e-06, 'learning_rate/action_model': 1.4089601202329114e-05, 'epoch': 1.65}) +[2026-06-16 03:01:06,074][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.08700353652238846, 'train/grad_norm_pre_clip': 12.418265342712402, 'timing/data': 0.00031812599627301097, 'timing/model': 1.1306505859829485, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 1.65}) +[2026-06-16 03:01:10,597][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.17162767052650452, 'train/grad_norm_pre_clip': 17.425556182861328, 'timing/data': 0.00036823295522481203, 'timing/model': 1.127206510049291, 'learning_rate/qwen_vl_interface': 1.4053800128476426e-06, 'learning_rate/action_model': 1.4053800128476427e-05, 'epoch': 1.65}) +[2026-06-16 03:01:15,126][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.0036977960262447596, 'train/grad_norm_pre_clip': 1.1587074995040894, 'timing/data': 0.0003599550109356642, 'timing/model': 1.298419237951748, 'learning_rate/qwen_vl_interface': 1.403592329354338e-06, 'learning_rate/action_model': 1.403592329354338e-05, 'epoch': 1.65}) +[2026-06-16 03:01:19,504][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.07390107214450836, 'train/grad_norm_pre_clip': 56.043556213378906, 'timing/data': 0.00034822302404791117, 'timing/model': 1.1364177389768884, 'learning_rate/qwen_vl_interface': 1.4018062269739354e-06, 'learning_rate/action_model': 1.4018062269739355e-05, 'epoch': 1.65}) +[2026-06-16 03:01:24,043][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.017281316220760345, 'train/grad_norm_pre_clip': 6.683680057525635, 'timing/data': 0.00036702200304716825, 'timing/model': 1.1303219750407152, 'learning_rate/qwen_vl_interface': 1.4000217064406339e-06, 'learning_rate/action_model': 1.4000217064406337e-05, 'epoch': 1.65}) +[2026-06-16 03:01:28,565][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.051092587411403656, 'train/grad_norm_pre_clip': 18.86460304260254, 'timing/data': 0.00039129203651100397, 'timing/model': 1.1278703169664368, 'learning_rate/qwen_vl_interface': 1.3982387684879836e-06, 'learning_rate/action_model': 1.3982387684879836e-05, 'epoch': 1.65}) +[2026-06-16 03:01:33,103][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.12460236996412277, 'train/grad_norm_pre_clip': 12.310701370239258, 'timing/data': 0.00032001896761357784, 'timing/model': 1.1330892760306597, 'learning_rate/qwen_vl_interface': 1.3964574138488804e-06, 'learning_rate/action_model': 1.3964574138488803e-05, 'epoch': 1.65}) +[2026-06-16 03:01:37,590][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.0847165510058403, 'train/grad_norm_pre_clip': 13.636675834655762, 'timing/data': 0.0005244520143605769, 'timing/model': 1.130506492976565, 'learning_rate/qwen_vl_interface': 1.3946776432555743e-06, 'learning_rate/action_model': 1.3946776432555741e-05, 'epoch': 1.65}) +[2026-06-16 03:01:42,134][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.0014102654531598091, 'train/grad_norm_pre_clip': 0.8424304127693176, 'timing/data': 0.0003313919878564775, 'timing/model': 1.2960278579848818, 'learning_rate/qwen_vl_interface': 1.3928994574396595e-06, 'learning_rate/action_model': 1.3928994574396594e-05, 'epoch': 1.65}) +[2026-06-16 03:01:46,521][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.14576967060565948, 'train/grad_norm_pre_clip': 31.365568161010742, 'timing/data': 0.0003730549942702055, 'timing/model': 1.1422363719902933, 'learning_rate/qwen_vl_interface': 1.391122857132083e-06, 'learning_rate/action_model': 1.3911228571320831e-05, 'epoch': 1.66}) +[2026-06-16 03:01:51,048][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.014490384608507156, 'train/grad_norm_pre_clip': 4.459549427032471, 'timing/data': 0.00035354297142475843, 'timing/model': 1.1361281250137836, 'learning_rate/qwen_vl_interface': 1.3893478430631376e-06, 'learning_rate/action_model': 1.3893478430631376e-05, 'epoch': 1.66}) +[2026-06-16 03:01:55,615][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.21989798545837402, 'train/grad_norm_pre_clip': 33.250267028808594, 'timing/data': 0.0003647439880296588, 'timing/model': 1.1478179180412553, 'learning_rate/qwen_vl_interface': 1.3875744159624652e-06, 'learning_rate/action_model': 1.3875744159624652e-05, 'epoch': 1.66}) +[2026-06-16 03:02:00,153][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.11739605665206909, 'train/grad_norm_pre_clip': 12.134746551513672, 'timing/data': 0.00036093697417527437, 'timing/model': 1.1342662110109814, 'learning_rate/qwen_vl_interface': 1.3858025765590525e-06, 'learning_rate/action_model': 1.3858025765590525e-05, 'epoch': 1.66}) +[2026-06-16 03:02:04,702][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.00475228950381279, 'train/grad_norm_pre_clip': 1.5127273797988892, 'timing/data': 0.00033983797766268253, 'timing/model': 1.1338420999818482, 'learning_rate/qwen_vl_interface': 1.384032325581237e-06, 'learning_rate/action_model': 1.3840323255812371e-05, 'epoch': 1.66}) +[2026-06-16 03:02:09,232][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.0067299469374120235, 'train/grad_norm_pre_clip': 3.327129364013672, 'timing/data': 0.0003540899488143623, 'timing/model': 1.2953429329791106, 'learning_rate/qwen_vl_interface': 1.3822636637567025e-06, 'learning_rate/action_model': 1.3822636637567024e-05, 'epoch': 1.66}) +[2026-06-16 03:02:13,602][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.01856733113527298, 'train/grad_norm_pre_clip': 11.142127990722656, 'timing/data': 0.0003739820094779134, 'timing/model': 1.133070404001046, 'learning_rate/qwen_vl_interface': 1.3804965918124788e-06, 'learning_rate/action_model': 1.3804965918124787e-05, 'epoch': 1.66}) +[2026-06-16 03:02:18,131][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.21800315380096436, 'train/grad_norm_pre_clip': 15.340326309204102, 'timing/data': 0.00033539097057655454, 'timing/model': 1.1312773330137134, 'learning_rate/qwen_vl_interface': 1.3787311104749418e-06, 'learning_rate/action_model': 1.3787311104749418e-05, 'epoch': 1.66}) +[2026-06-16 03:02:22,669][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.009499563835561275, 'train/grad_norm_pre_clip': 5.740284442901611, 'timing/data': 0.0003718429943546653, 'timing/model': 1.1378453409997746, 'learning_rate/qwen_vl_interface': 1.3769672204698149e-06, 'learning_rate/action_model': 1.3769672204698147e-05, 'epoch': 1.66}) +[2026-06-16 03:02:27,205][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.17362071573734283, 'train/grad_norm_pre_clip': 9.401957511901855, 'timing/data': 0.00034182396484538913, 'timing/model': 1.1438614000217058, 'learning_rate/qwen_vl_interface': 1.3752049225221664e-06, 'learning_rate/action_model': 1.3752049225221663e-05, 'epoch': 1.66}) +[2026-06-16 03:02:31,713][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.020167626440525055, 'train/grad_norm_pre_clip': 6.216841220855713, 'timing/data': 0.00033223204081878066, 'timing/model': 1.1311434620292857, 'learning_rate/qwen_vl_interface': 1.3734442173564103e-06, 'learning_rate/action_model': 1.3734442173564102e-05, 'epoch': 1.66}) +[2026-06-16 03:02:36,260][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.003302029799669981, 'train/grad_norm_pre_clip': 1.2582851648330688, 'timing/data': 0.0003631020081229508, 'timing/model': 1.2839708059909754, 'learning_rate/qwen_vl_interface': 1.3716851056963062e-06, 'learning_rate/action_model': 1.3716851056963062e-05, 'epoch': 1.66}) +[2026-06-16 03:02:40,633][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.14359433948993683, 'train/grad_norm_pre_clip': 17.734027862548828, 'timing/data': 0.00037452997639775276, 'timing/model': 1.1327370699727908, 'learning_rate/qwen_vl_interface': 1.3699275882649597e-06, 'learning_rate/action_model': 1.3699275882649596e-05, 'epoch': 1.66}) +[2026-06-16 03:02:45,187][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.01279864739626646, 'train/grad_norm_pre_clip': 3.591843843460083, 'timing/data': 0.00037312298081815243, 'timing/model': 1.1375942259910516, 'learning_rate/qwen_vl_interface': 1.368171665784817e-06, 'learning_rate/action_model': 1.3681716657848171e-05, 'epoch': 1.66}) +[2026-06-16 03:02:49,720][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.0013943883823230863, 'train/grad_norm_pre_clip': 0.23953770101070404, 'timing/data': 0.00032754300627857447, 'timing/model': 1.1315037220483646, 'learning_rate/qwen_vl_interface': 1.3664173389776735e-06, 'learning_rate/action_model': 1.3664173389776735e-05, 'epoch': 1.66}) +[2026-06-16 03:02:54,249][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.00782695971429348, 'train/grad_norm_pre_clip': 4.909033298492432, 'timing/data': 0.0005954320076853037, 'timing/model': 1.1383176679955795, 'learning_rate/qwen_vl_interface': 1.3646646085646658e-06, 'learning_rate/action_model': 1.3646646085646659e-05, 'epoch': 1.66}) +[2026-06-16 03:02:58,808][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.11130500584840775, 'train/grad_norm_pre_clip': 6.024033069610596, 'timing/data': 0.0003257269854657352, 'timing/model': 1.1332957640406676, 'learning_rate/qwen_vl_interface': 1.3629134752662756e-06, 'learning_rate/action_model': 1.3629134752662756e-05, 'epoch': 1.66}) +[2026-06-16 03:03:03,358][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 0.0015407168539240956, 'train/grad_norm_pre_clip': 0.25466281175613403, 'timing/data': 0.0003555759903974831, 'timing/model': 1.3022048579878174, 'learning_rate/qwen_vl_interface': 1.3611639398023282e-06, 'learning_rate/action_model': 1.3611639398023283e-05, 'epoch': 1.66}) +[2026-06-16 03:03:07,758][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.0019554628524929285, 'train/grad_norm_pre_clip': 0.5085040926933289, 'timing/data': 0.00038773997221142054, 'timing/model': 1.1370470409747213, 'learning_rate/qwen_vl_interface': 1.35941600289199e-06, 'learning_rate/action_model': 1.35941600289199e-05, 'epoch': 1.66}) +[2026-06-16 03:03:12,299][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.002767932368442416, 'train/grad_norm_pre_clip': 0.8509492874145508, 'timing/data': 0.0004209880135022104, 'timing/model': 1.1340689749922603, 'learning_rate/qwen_vl_interface': 1.3576696652537735e-06, 'learning_rate/action_model': 1.3576696652537736e-05, 'epoch': 1.66}) +[2026-06-16 03:03:16,857][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.005311762448400259, 'train/grad_norm_pre_clip': 1.2139135599136353, 'timing/data': 0.00039065699093043804, 'timing/model': 1.1320651659625582, 'learning_rate/qwen_vl_interface': 1.355924927605531e-06, 'learning_rate/action_model': 1.3559249276055311e-05, 'epoch': 1.66}) +[2026-06-16 03:03:21,398][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.2481020838022232, 'train/grad_norm_pre_clip': 23.83409881591797, 'timing/data': 0.00031531002605333924, 'timing/model': 1.1352924249949865, 'learning_rate/qwen_vl_interface': 1.3541817906644591e-06, 'learning_rate/action_model': 1.3541817906644591e-05, 'epoch': 1.66}) +[2026-06-16 03:03:25,914][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.0024178018793463707, 'train/grad_norm_pre_clip': 0.4195156395435333, 'timing/data': 0.0003643139498308301, 'timing/model': 1.137492106994614, 'learning_rate/qwen_vl_interface': 1.3524402551470955e-06, 'learning_rate/action_model': 1.3524402551470956e-05, 'epoch': 1.66}) +[2026-06-16 03:03:30,453][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.003344966098666191, 'train/grad_norm_pre_clip': 0.9361987709999084, 'timing/data': 0.0003262520185671747, 'timing/model': 1.2828983989893459, 'learning_rate/qwen_vl_interface': 1.3507003217693196e-06, 'learning_rate/action_model': 1.3507003217693196e-05, 'epoch': 1.66}) +[2026-06-16 03:03:34,851][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.034741174429655075, 'train/grad_norm_pre_clip': 8.461360931396484, 'timing/data': 0.00037802295992150903, 'timing/model': 1.137066246010363, 'learning_rate/qwen_vl_interface': 1.3489619912463515e-06, 'learning_rate/action_model': 1.3489619912463514e-05, 'epoch': 1.66}) +[2026-06-16 03:03:39,396][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.005584260914474726, 'train/grad_norm_pre_clip': 2.536604404449463, 'timing/data': 0.00039312097942456603, 'timing/model': 1.1397943159681745, 'learning_rate/qwen_vl_interface': 1.3472252642927534e-06, 'learning_rate/action_model': 1.3472252642927533e-05, 'epoch': 1.67}) +[2026-06-16 03:03:43,950][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 0.09812112152576447, 'train/grad_norm_pre_clip': 14.123115539550781, 'timing/data': 0.00034256401704624295, 'timing/model': 1.1344632699619979, 'learning_rate/qwen_vl_interface': 1.3454901416224284e-06, 'learning_rate/action_model': 1.3454901416224284e-05, 'epoch': 1.67}) +[2026-06-16 03:03:48,487][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 0.009528440423309803, 'train/grad_norm_pre_clip': 2.50007963180542, 'timing/data': 0.00034447602229192853, 'timing/model': 1.1419397569843568, 'learning_rate/qwen_vl_interface': 1.34375662394862e-06, 'learning_rate/action_model': 1.34375662394862e-05, 'epoch': 1.67}) +[2026-06-16 03:03:53,037][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.001615028944797814, 'train/grad_norm_pre_clip': 0.5654962062835693, 'timing/data': 0.000359382014721632, 'timing/model': 1.139445477980189, 'learning_rate/qwen_vl_interface': 1.3420247119839116e-06, 'learning_rate/action_model': 1.3420247119839117e-05, 'epoch': 1.67}) +[2026-06-16 03:03:57,573][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.024161789566278458, 'train/grad_norm_pre_clip': 11.93087100982666, 'timing/data': 0.0003506569773890078, 'timing/model': 1.2871977420290932, 'learning_rate/qwen_vl_interface': 1.340294406440227e-06, 'learning_rate/action_model': 1.340294406440227e-05, 'epoch': 1.67}) +[2026-06-16 03:04:01,999][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.0025914106518030167, 'train/grad_norm_pre_clip': 0.3914938271045685, 'timing/data': 0.000342411978635937, 'timing/model': 1.1333042229525745, 'learning_rate/qwen_vl_interface': 1.3385657080288297e-06, 'learning_rate/action_model': 1.3385657080288297e-05, 'epoch': 1.67}) +[2026-06-16 03:04:06,582][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.0013110630679875612, 'train/grad_norm_pre_clip': 0.34089383482933044, 'timing/data': 0.000355011026840657, 'timing/model': 1.139053207996767, 'learning_rate/qwen_vl_interface': 1.336838617460322e-06, 'learning_rate/action_model': 1.336838617460322e-05, 'epoch': 1.67}) +[2026-06-16 03:04:11,136][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.01316704973578453, 'train/grad_norm_pre_clip': 5.419403076171875, 'timing/data': 0.00036139803705736995, 'timing/model': 1.1356824740068987, 'learning_rate/qwen_vl_interface': 1.335113135444646e-06, 'learning_rate/action_model': 1.335113135444646e-05, 'epoch': 1.67}) +[2026-06-16 03:04:15,694][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 0.0010358706349506974, 'train/grad_norm_pre_clip': 0.2078067511320114, 'timing/data': 0.00035864097299054265, 'timing/model': 1.1403487949864939, 'learning_rate/qwen_vl_interface': 1.3333892626910823e-06, 'learning_rate/action_model': 1.3333892626910823e-05, 'epoch': 1.67}) +[2026-06-16 03:04:20,208][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.11025846004486084, 'train/grad_norm_pre_clip': 10.185325622558594, 'timing/data': 0.0003216049517504871, 'timing/model': 1.1422227009898052, 'learning_rate/qwen_vl_interface': 1.3316669999082486e-06, 'learning_rate/action_model': 1.3316669999082484e-05, 'epoch': 1.67}) +[2026-06-16 03:04:24,775][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.034536197781562805, 'train/grad_norm_pre_clip': 17.42371940612793, 'timing/data': 0.00033430295297876, 'timing/model': 1.3046608220320195, 'learning_rate/qwen_vl_interface': 1.3299463478041024e-06, 'learning_rate/action_model': 1.3299463478041024e-05, 'epoch': 1.67}) +[2026-06-16 03:04:29,168][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.014030721969902515, 'train/grad_norm_pre_clip': 6.311553955078125, 'timing/data': 0.0006038600113242865, 'timing/model': 1.1462226809817366, 'learning_rate/qwen_vl_interface': 1.3282273070859396e-06, 'learning_rate/action_model': 1.3282273070859396e-05, 'epoch': 1.67}) +[2026-06-16 03:04:33,702][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 0.001475691795349121, 'train/grad_norm_pre_clip': 0.31508415937423706, 'timing/data': 0.0005909220199100673, 'timing/model': 1.1324527590186335, 'learning_rate/qwen_vl_interface': 1.3265098784603924e-06, 'learning_rate/action_model': 1.3265098784603924e-05, 'epoch': 1.67}) +[2026-06-16 03:04:38,234][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.19692005217075348, 'train/grad_norm_pre_clip': 18.481592178344727, 'timing/data': 0.0003475600387901068, 'timing/model': 1.1387910440098494, 'learning_rate/qwen_vl_interface': 1.324794062633431e-06, 'learning_rate/action_model': 1.324794062633431e-05, 'epoch': 1.67}) +[2026-06-16 03:04:42,760][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.00347942765802145, 'train/grad_norm_pre_clip': 1.5452016592025757, 'timing/data': 0.000387699983548373, 'timing/model': 1.132031811052002, 'learning_rate/qwen_vl_interface': 1.323079860310363e-06, 'learning_rate/action_model': 1.3230798603103628e-05, 'epoch': 1.67}) +[2026-06-16 03:04:47,327][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.06376567482948303, 'train/grad_norm_pre_clip': 31.215167999267578, 'timing/data': 0.00037780997809022665, 'timing/model': 1.135691425006371, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 1.67}) +[2026-06-16 03:04:51,857][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.09163211286067963, 'train/grad_norm_pre_clip': 14.105040550231934, 'timing/data': 0.0004674720112234354, 'timing/model': 1.284087935986463, 'learning_rate/qwen_vl_interface': 1.3196562989938175e-06, 'learning_rate/action_model': 1.3196562989938174e-05, 'epoch': 1.67}) +[2026-06-16 03:04:56,242][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.0054410770535469055, 'train/grad_norm_pre_clip': 1.3352028131484985, 'timing/data': 0.0003538990276865661, 'timing/model': 1.1325231810333207, 'learning_rate/qwen_vl_interface': 1.3179469414076357e-06, 'learning_rate/action_model': 1.3179469414076356e-05, 'epoch': 1.67}) +[2026-06-16 03:05:00,809][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.19082844257354736, 'train/grad_norm_pre_clip': 8.623861312866211, 'timing/data': 0.00035804498475044966, 'timing/model': 1.143034620967228, 'learning_rate/qwen_vl_interface': 1.3162392001399401e-06, 'learning_rate/action_model': 1.3162392001399401e-05, 'epoch': 1.67}) +[2026-06-16 03:05:05,388][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.06255751103162766, 'train/grad_norm_pre_clip': 34.70865249633789, 'timing/data': 0.00035129301249980927, 'timing/model': 1.1367256960365921, 'learning_rate/qwen_vl_interface': 1.3145330758927182e-06, 'learning_rate/action_model': 1.3145330758927182e-05, 'epoch': 1.67}) +[2026-06-16 03:05:09,937][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.0028648434672504663, 'train/grad_norm_pre_clip': 0.5104945302009583, 'timing/data': 0.00034473196137696505, 'timing/model': 1.1403224910027348, 'learning_rate/qwen_vl_interface': 1.3128285693672935e-06, 'learning_rate/action_model': 1.3128285693672934e-05, 'epoch': 1.67}) +[2026-06-16 03:05:14,458][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.00018582645861897618, 'train/grad_norm_pre_clip': 0.031927645206451416, 'timing/data': 0.0003491929965093732, 'timing/model': 1.137122096028179, 'learning_rate/qwen_vl_interface': 1.311125681264325e-06, 'learning_rate/action_model': 1.311125681264325e-05, 'epoch': 1.67}) +[2026-06-16 03:05:19,024][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.07577656954526901, 'train/grad_norm_pre_clip': 15.080910682678223, 'timing/data': 0.00032833300065249205, 'timing/model': 1.3010414890013635, 'learning_rate/qwen_vl_interface': 1.309424412283803e-06, 'learning_rate/action_model': 1.3094244122838029e-05, 'epoch': 1.67}) +[2026-06-16 03:05:23,415][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 0.012278441339731216, 'train/grad_norm_pre_clip': 5.100744247436523, 'timing/data': 0.00048396200872957706, 'timing/model': 1.136867158988025, 'learning_rate/qwen_vl_interface': 1.307724763125057e-06, 'learning_rate/action_model': 1.307724763125057e-05, 'epoch': 1.67}) +[2026-06-16 03:05:27,970][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.005671469960361719, 'train/grad_norm_pre_clip': 4.056695461273193, 'timing/data': 0.00034659402444958687, 'timing/model': 1.132542872976046, 'learning_rate/qwen_vl_interface': 1.3060267344867478e-06, 'learning_rate/action_model': 1.3060267344867478e-05, 'epoch': 1.68}) +[2026-06-16 03:05:32,550][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.0102917505428195, 'train/grad_norm_pre_clip': 3.6432387828826904, 'timing/data': 0.00033981603337451816, 'timing/model': 1.1422724919975735, 'learning_rate/qwen_vl_interface': 1.3043303270668715e-06, 'learning_rate/action_model': 1.3043303270668713e-05, 'epoch': 1.68}) +[2026-06-16 03:05:37,105][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.0013647826854139566, 'train/grad_norm_pre_clip': 0.29021087288856506, 'timing/data': 0.0003597459872253239, 'timing/model': 1.1378913469961844, 'learning_rate/qwen_vl_interface': 1.3026355415627565e-06, 'learning_rate/action_model': 1.3026355415627564e-05, 'epoch': 1.68}) +[2026-06-16 03:05:41,654][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.0003791091439779848, 'train/grad_norm_pre_clip': 0.12258128821849823, 'timing/data': 0.0003368179895915091, 'timing/model': 1.1353490439942107, 'learning_rate/qwen_vl_interface': 1.3009423786710653e-06, 'learning_rate/action_model': 1.3009423786710653e-05, 'epoch': 1.68}) +[2026-06-16 03:05:46,222][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.011753113940358162, 'train/grad_norm_pre_clip': 3.913851261138916, 'timing/data': 0.00036936497781425714, 'timing/model': 1.3151195020182058, 'learning_rate/qwen_vl_interface': 1.2992508390877933e-06, 'learning_rate/action_model': 1.2992508390877933e-05, 'epoch': 1.68}) +[2026-06-16 03:05:50,616][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.007608279585838318, 'train/grad_norm_pre_clip': 6.619737148284912, 'timing/data': 0.00036217900924384594, 'timing/model': 1.1409596829907969, 'learning_rate/qwen_vl_interface': 1.297560923508269e-06, 'learning_rate/action_model': 1.297560923508269e-05, 'epoch': 1.68}) +[2026-06-16 03:05:55,165][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.025887850672006607, 'train/grad_norm_pre_clip': 14.097413063049316, 'timing/data': 0.00059040094492957, 'timing/model': 1.133309417986311, 'learning_rate/qwen_vl_interface': 1.295872632627152e-06, 'learning_rate/action_model': 1.2958726326271519e-05, 'epoch': 1.68}) +[2026-06-16 03:05:59,727][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 0.36274290084838867, 'train/grad_norm_pre_clip': 17.718568801879883, 'timing/data': 0.0003435180406086147, 'timing/model': 1.1478202909929678, 'learning_rate/qwen_vl_interface': 1.2941859671384364e-06, 'learning_rate/action_model': 1.2941859671384362e-05, 'epoch': 1.68}) +[2026-06-16 03:06:04,281][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.0005747874383814633, 'train/grad_norm_pre_clip': 0.08361034840345383, 'timing/data': 0.0003756469814106822, 'timing/model': 1.1386282820021734, 'learning_rate/qwen_vl_interface': 1.2925009277354444e-06, 'learning_rate/action_model': 1.2925009277354444e-05, 'epoch': 1.68}) +[2026-06-16 03:06:08,782][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.00043497278238646686, 'train/grad_norm_pre_clip': 0.06344816088676453, 'timing/data': 0.0003599789924919605, 'timing/model': 1.1374447470298037, 'learning_rate/qwen_vl_interface': 1.2908175151108335e-06, 'learning_rate/action_model': 1.2908175151108334e-05, 'epoch': 1.68}) +[2026-06-16 03:06:13,324][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.10106894373893738, 'train/grad_norm_pre_clip': 17.05155372619629, 'timing/data': 0.00048758601769804955, 'timing/model': 1.2815190429682843, 'learning_rate/qwen_vl_interface': 1.2891357299565899e-06, 'learning_rate/action_model': 1.2891357299565898e-05, 'epoch': 1.68}) +[2026-06-16 03:06:17,721][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.005433567799627781, 'train/grad_norm_pre_clip': 1.149375081062317, 'timing/data': 0.0003881249576807022, 'timing/model': 1.142387019994203, 'learning_rate/qwen_vl_interface': 1.2874555729640333e-06, 'learning_rate/action_model': 1.2874555729640334e-05, 'epoch': 1.68}) +[2026-06-16 03:06:22,269][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.002745826495811343, 'train/grad_norm_pre_clip': 0.36249834299087524, 'timing/data': 0.00038531102472916245, 'timing/model': 1.1335738699999638, 'learning_rate/qwen_vl_interface': 1.2857770448238124e-06, 'learning_rate/action_model': 1.2857770448238123e-05, 'epoch': 1.68}) +[2026-06-16 03:06:26,820][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.00039421525434590876, 'train/grad_norm_pre_clip': 0.07537834346294403, 'timing/data': 0.00035249500069767237, 'timing/model': 1.147610038984567, 'learning_rate/qwen_vl_interface': 1.2841001462259067e-06, 'learning_rate/action_model': 1.2841001462259067e-05, 'epoch': 1.68}) +[2026-06-16 03:06:31,348][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.4812896251678467, 'train/grad_norm_pre_clip': 25.219558715820312, 'timing/data': 0.000575690995901823, 'timing/model': 1.134981672046706, 'learning_rate/qwen_vl_interface': 1.282424877859626e-06, 'learning_rate/action_model': 1.282424877859626e-05, 'epoch': 1.68}) +[2026-06-16 03:06:35,899][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.03784874454140663, 'train/grad_norm_pre_clip': 17.26974868774414, 'timing/data': 0.000602623971644789, 'timing/model': 1.1356541460263543, 'learning_rate/qwen_vl_interface': 1.2807512404136088e-06, 'learning_rate/action_model': 1.2807512404136087e-05, 'epoch': 1.68}) +[2026-06-16 03:06:40,463][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.29813694953918457, 'train/grad_norm_pre_clip': 19.066020965576172, 'timing/data': 0.0004067099653184414, 'timing/model': 1.3011076740222052, 'learning_rate/qwen_vl_interface': 1.2790792345758255e-06, 'learning_rate/action_model': 1.2790792345758254e-05, 'epoch': 1.68}) +[2026-06-16 03:06:44,861][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.04241865128278732, 'train/grad_norm_pre_clip': 18.607433319091797, 'timing/data': 0.0005535679520107806, 'timing/model': 1.1353577570407651, 'learning_rate/qwen_vl_interface': 1.2774088610335752e-06, 'learning_rate/action_model': 1.2774088610335752e-05, 'epoch': 1.68}) +[2026-06-16 03:06:49,423][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 0.018808364868164062, 'train/grad_norm_pre_clip': 15.390421867370605, 'timing/data': 0.00036144600016996264, 'timing/model': 1.137662066030316, 'learning_rate/qwen_vl_interface': 1.2757401204734827e-06, 'learning_rate/action_model': 1.2757401204734826e-05, 'epoch': 1.68}) +[2026-06-16 03:06:53,981][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.007927384227514267, 'train/grad_norm_pre_clip': 2.1618001461029053, 'timing/data': 0.000544116017408669, 'timing/model': 1.1337743950425647, 'learning_rate/qwen_vl_interface': 1.2740730135815064e-06, 'learning_rate/action_model': 1.2740730135815062e-05, 'epoch': 1.68}) +[2026-06-16 03:06:58,523][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 0.0018539474112913013, 'train/grad_norm_pre_clip': 0.35287871956825256, 'timing/data': 0.0004352540127001703, 'timing/model': 1.1327654650085606, 'learning_rate/qwen_vl_interface': 1.2724075410429298e-06, 'learning_rate/action_model': 1.2724075410429299e-05, 'epoch': 1.68}) +[2026-06-16 03:07:03,019][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.00045102756121195853, 'train/grad_norm_pre_clip': 0.09824415296316147, 'timing/data': 0.0003227589768357575, 'timing/model': 1.1320915829855949, 'learning_rate/qwen_vl_interface': 1.2707437035423664e-06, 'learning_rate/action_model': 1.2707437035423664e-05, 'epoch': 1.68}) +[2026-06-16 03:11:26,661][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.15861393511295319, 'train/grad_norm_pre_clip': 21.997812271118164, 'eval/loss': 0.0428277850151062, 'eval/flap_precision': 0.9636363636363636, 'eval/flap_recall': 0.9814814814814815, 'eval/flap_f1': 0.9724770642201834, 'eval/flap_rate_model': 0.18333333333333332, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9724770642201834, 'timing/data': 0.00034474098356440663, 'timing/model': 1.2905864840140566, 'learning_rate/qwen_vl_interface': 1.2690815017637573e-06, 'learning_rate/action_model': 1.2690815017637572e-05, 'epoch': 1.68}) +[2026-06-16 03:11:26,661][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4100_state +[2026-06-16 03:11:26,665][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 03:11:30,254][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4100_state/model.safetensors +[2026-06-16 03:11:40,321][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4100_state/optimizer.bin +[2026-06-16 03:11:40,321][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4100_state/sampler.bin +[2026-06-16 03:11:40,322][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4100_state/sampler_1.bin +[2026-06-16 03:11:40,322][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4100_state/random_states_0.pkl +[2026-06-16 03:11:40,362][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4100 +[2026-06-16 03:11:40,362][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 03:11:40,369][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 03:11:44,848][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.0002974103845190257, 'train/grad_norm_pre_clip': 0.03616558387875557, 'timing/data': 0.0006272579776123166, 'timing/model': 1.3011373790213838, 'learning_rate/qwen_vl_interface': 1.2674209363903694e-06, 'learning_rate/action_model': 1.2674209363903693e-05, 'epoch': 1.68}) +[2026-06-16 03:11:49,246][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.13365033268928528, 'train/grad_norm_pre_clip': 21.2059383392334, 'timing/data': 0.00035027501871809363, 'timing/model': 1.2052404689602554, 'learning_rate/qwen_vl_interface': 1.2657620081048005e-06, 'learning_rate/action_model': 1.2657620081048006e-05, 'epoch': 1.69}) +[2026-06-16 03:11:53,568][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.0041608624160289764, 'train/grad_norm_pre_clip': 1.8685760498046875, 'timing/data': 0.0003371649654582143, 'timing/model': 1.1150948769645765, 'learning_rate/qwen_vl_interface': 1.2641047175889714e-06, 'learning_rate/action_model': 1.2641047175889714e-05, 'epoch': 1.69}) +[2026-06-16 03:11:58,002][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.00013997616770211607, 'train/grad_norm_pre_clip': 0.022412296384572983, 'timing/data': 0.00031782896257936954, 'timing/model': 1.1174557650228962, 'learning_rate/qwen_vl_interface': 1.2624490655241332e-06, 'learning_rate/action_model': 1.262449065524133e-05, 'epoch': 1.69}) +[2026-06-16 03:12:02,519][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.0008263170020654798, 'train/grad_norm_pre_clip': 0.3014504909515381, 'timing/data': 0.00034439703449606895, 'timing/model': 1.1266504899831489, 'learning_rate/qwen_vl_interface': 1.2607950525908602e-06, 'learning_rate/action_model': 1.2607950525908601e-05, 'epoch': 1.69}) +[2026-06-16 03:12:06,999][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.00026586398598738015, 'train/grad_norm_pre_clip': 0.038033246994018555, 'timing/data': 0.00047422200441360474, 'timing/model': 1.1194742710213177, 'learning_rate/qwen_vl_interface': 1.2591426794690575e-06, 'learning_rate/action_model': 1.2591426794690576e-05, 'epoch': 1.69}) +[2026-06-16 03:12:11,511][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.0425490066409111, 'train/grad_norm_pre_clip': 39.73494338989258, 'timing/data': 0.0003354120417498052, 'timing/model': 1.2964001679793, 'learning_rate/qwen_vl_interface': 1.25749194683795e-06, 'learning_rate/action_model': 1.25749194683795e-05, 'epoch': 1.69}) +[2026-06-16 03:12:15,864][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.10395563393831253, 'train/grad_norm_pre_clip': 9.056571960449219, 'timing/data': 0.0003385640447959304, 'timing/model': 1.123217503016349, 'learning_rate/qwen_vl_interface': 1.2558428553760933e-06, 'learning_rate/action_model': 1.2558428553760932e-05, 'epoch': 1.69}) +[2026-06-16 03:12:20,386][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.0004014952573925257, 'train/grad_norm_pre_clip': 0.08938481658697128, 'timing/data': 0.000343983992934227, 'timing/model': 1.125679777993355, 'learning_rate/qwen_vl_interface': 1.2541954057613659e-06, 'learning_rate/action_model': 1.2541954057613658e-05, 'epoch': 1.69}) +[2026-06-16 03:12:24,876][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.03825366124510765, 'train/grad_norm_pre_clip': 13.210430145263672, 'timing/data': 0.00033890397753566504, 'timing/model': 1.1238382610026747, 'learning_rate/qwen_vl_interface': 1.2525495986709727e-06, 'learning_rate/action_model': 1.2525495986709726e-05, 'epoch': 1.69}) +[2026-06-16 03:12:29,347][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.0011740520130842924, 'train/grad_norm_pre_clip': 0.265216201543808, 'timing/data': 0.0003530130488798022, 'timing/model': 1.1375251990393735, 'learning_rate/qwen_vl_interface': 1.2509054347814423e-06, 'learning_rate/action_model': 1.2509054347814423e-05, 'epoch': 1.69}) +[2026-06-16 03:12:33,881][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.0005486856680363417, 'train/grad_norm_pre_clip': 0.089137963950634, 'timing/data': 0.0004048860282637179, 'timing/model': 1.1337779519963078, 'learning_rate/qwen_vl_interface': 1.2492629147686288e-06, 'learning_rate/action_model': 1.2492629147686289e-05, 'epoch': 1.69}) +[2026-06-16 03:12:38,403][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.2177141010761261, 'train/grad_norm_pre_clip': 26.324352264404297, 'timing/data': 0.0003852100344374776, 'timing/model': 1.2967752490076236, 'learning_rate/qwen_vl_interface': 1.2476220393077098e-06, 'learning_rate/action_model': 1.2476220393077098e-05, 'epoch': 1.69}) +[2026-06-16 03:12:42,766][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.17385461926460266, 'train/grad_norm_pre_clip': 23.746253967285156, 'timing/data': 0.00035240798024460673, 'timing/model': 1.128502355015371, 'learning_rate/qwen_vl_interface': 1.2459828090731877e-06, 'learning_rate/action_model': 1.2459828090731876e-05, 'epoch': 1.69}) +[2026-06-16 03:12:47,317][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.01990353688597679, 'train/grad_norm_pre_clip': 13.191325187683105, 'timing/data': 0.00034790096106007695, 'timing/model': 1.1346641959971748, 'learning_rate/qwen_vl_interface': 1.2443452247388876e-06, 'learning_rate/action_model': 1.2443452247388877e-05, 'epoch': 1.69}) +[2026-06-16 03:12:51,837][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.0011160633293911815, 'train/grad_norm_pre_clip': 0.3572631776332855, 'timing/data': 0.00034833798417821527, 'timing/model': 1.1299057219875976, 'learning_rate/qwen_vl_interface': 1.2427092869779603e-06, 'learning_rate/action_model': 1.2427092869779602e-05, 'epoch': 1.69}) +[2026-06-16 03:12:56,378][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.23646754026412964, 'train/grad_norm_pre_clip': 12.356907844543457, 'timing/data': 0.000337668985594064, 'timing/model': 1.1305673979804851, 'learning_rate/qwen_vl_interface': 1.2410749964628751e-06, 'learning_rate/action_model': 1.2410749964628751e-05, 'epoch': 1.69}) +[2026-06-16 03:13:00,904][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.1497669368982315, 'train/grad_norm_pre_clip': 16.49956512451172, 'timing/data': 0.00036676402669399977, 'timing/model': 1.28596429299796, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 1.69}) +[2026-06-16 03:13:05,378][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.20104660093784332, 'train/grad_norm_pre_clip': 12.311847686767578, 'timing/data': 0.0003329210449010134, 'timing/model': 1.2400077400379814, 'learning_rate/qwen_vl_interface': 1.2378113598567394e-06, 'learning_rate/action_model': 1.2378113598567392e-05, 'epoch': 1.69}) +[2026-06-16 03:13:09,733][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.0005154192331247032, 'train/grad_norm_pre_clip': 0.111545629799366, 'timing/data': 0.00040614401223137975, 'timing/model': 1.129568260977976, 'learning_rate/qwen_vl_interface': 1.2361820151072454e-06, 'learning_rate/action_model': 1.2361820151072453e-05, 'epoch': 1.69}) +[2026-06-16 03:13:14,260][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.0035670942161232233, 'train/grad_norm_pre_clip': 1.6057981252670288, 'timing/data': 0.00032826996175572276, 'timing/model': 1.1301609590300359, 'learning_rate/qwen_vl_interface': 1.23455432028671e-06, 'learning_rate/action_model': 1.23455432028671e-05, 'epoch': 1.69}) +[2026-06-16 03:13:18,792][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.08181589841842651, 'train/grad_norm_pre_clip': 31.235334396362305, 'timing/data': 0.0003592480206862092, 'timing/model': 1.1284098449978046, 'learning_rate/qwen_vl_interface': 1.2329282760642167e-06, 'learning_rate/action_model': 1.2329282760642166e-05, 'epoch': 1.69}) +[2026-06-16 03:13:23,336][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.0008136933902278543, 'train/grad_norm_pre_clip': 0.5435426235198975, 'timing/data': 0.00038480700459331274, 'timing/model': 1.133921513974201, 'learning_rate/qwen_vl_interface': 1.2313038831081716e-06, 'learning_rate/action_model': 1.2313038831081717e-05, 'epoch': 1.69}) +[2026-06-16 03:13:27,764][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.11035381257534027, 'train/grad_norm_pre_clip': 23.021371841430664, 'timing/data': 0.0006775110377930105, 'timing/model': 1.1382845109910704, 'learning_rate/qwen_vl_interface': 1.2296811420863002e-06, 'learning_rate/action_model': 1.2296811420863003e-05, 'epoch': 1.69}) +[2026-06-16 03:13:32,205][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.21696686744689941, 'train/grad_norm_pre_clip': 26.797962188720703, 'timing/data': 0.00034168100683018565, 'timing/model': 1.2088198070414364, 'learning_rate/qwen_vl_interface': 1.2280600536656504e-06, 'learning_rate/action_model': 1.2280600536656503e-05, 'epoch': 1.69}) +[2026-06-16 03:13:36,574][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.0029284318443387747, 'train/grad_norm_pre_clip': 0.8636564016342163, 'timing/data': 0.0003413290251046419, 'timing/model': 1.135221430042293, 'learning_rate/qwen_vl_interface': 1.2264406185125907e-06, 'learning_rate/action_model': 1.2264406185125907e-05, 'epoch': 1.69}) +[2026-06-16 03:13:41,064][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.0002070589835057035, 'train/grad_norm_pre_clip': 0.056725699454545975, 'timing/data': 0.0003446500049903989, 'timing/model': 1.134050605003722, 'learning_rate/qwen_vl_interface': 1.2248228372928098e-06, 'learning_rate/action_model': 1.2248228372928097e-05, 'epoch': 1.7}) +[2026-06-16 03:13:45,586][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.10053970664739609, 'train/grad_norm_pre_clip': 18.69855499267578, 'timing/data': 0.00035055394982919097, 'timing/model': 1.1312600430101156, 'learning_rate/qwen_vl_interface': 1.2232067106713174e-06, 'learning_rate/action_model': 1.2232067106713174e-05, 'epoch': 1.7}) +[2026-06-16 03:13:50,124][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.3435938358306885, 'train/grad_norm_pre_clip': 52.3477783203125, 'timing/data': 0.000346256943885237, 'timing/model': 1.1362732450361364, 'learning_rate/qwen_vl_interface': 1.22159223931244e-06, 'learning_rate/action_model': 1.2215922393124399e-05, 'epoch': 1.7}) +[2026-06-16 03:13:54,667][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.05525358021259308, 'train/grad_norm_pre_clip': 21.147130966186523, 'timing/data': 0.0003443799796514213, 'timing/model': 1.2906674310215749, 'learning_rate/qwen_vl_interface': 1.2199794238798266e-06, 'learning_rate/action_model': 1.2199794238798264e-05, 'epoch': 1.7}) +[2026-06-16 03:13:59,050][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.0016658810200169683, 'train/grad_norm_pre_clip': 0.4528869092464447, 'timing/data': 0.00033570401137694716, 'timing/model': 1.130073345033452, 'learning_rate/qwen_vl_interface': 1.2183682650364447e-06, 'learning_rate/action_model': 1.2183682650364447e-05, 'epoch': 1.7}) +[2026-06-16 03:14:03,600][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.17725028097629547, 'train/grad_norm_pre_clip': 12.496365547180176, 'timing/data': 0.00036605203058570623, 'timing/model': 1.1381814840133302, 'learning_rate/qwen_vl_interface': 1.216758763444581e-06, 'learning_rate/action_model': 1.2167587634445809e-05, 'epoch': 1.7}) +[2026-06-16 03:14:08,033][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.0007619624957442284, 'train/grad_norm_pre_clip': 0.21386933326721191, 'timing/data': 0.0003750820178538561, 'timing/model': 1.1296681300154887, 'learning_rate/qwen_vl_interface': 1.215150919765841e-06, 'learning_rate/action_model': 1.215150919765841e-05, 'epoch': 1.7}) +[2026-06-16 03:14:12,526][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.006952498108148575, 'train/grad_norm_pre_clip': 3.0519356727600098, 'timing/data': 0.00036643195198848844, 'timing/model': 1.1315082740038633, 'learning_rate/qwen_vl_interface': 1.213544734661148e-06, 'learning_rate/action_model': 1.2135447346611478e-05, 'epoch': 1.7}) +[2026-06-16 03:14:17,037][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.03613121062517166, 'train/grad_norm_pre_clip': 15.30086612701416, 'timing/data': 0.00037850101944059134, 'timing/model': 1.1287275809445418, 'learning_rate/qwen_vl_interface': 1.2119402087907448e-06, 'learning_rate/action_model': 1.2119402087907448e-05, 'epoch': 1.7}) +[2026-06-16 03:14:21,576][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.001361014787107706, 'train/grad_norm_pre_clip': 0.4167141616344452, 'timing/data': 0.0007233640062622726, 'timing/model': 1.2928129869978875, 'learning_rate/qwen_vl_interface': 1.2103373428141903e-06, 'learning_rate/action_model': 1.2103373428141903e-05, 'epoch': 1.7}) +[2026-06-16 03:14:25,958][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.006693221628665924, 'train/grad_norm_pre_clip': 2.9861278533935547, 'timing/data': 0.0006402219878509641, 'timing/model': 1.1341874380013905, 'learning_rate/qwen_vl_interface': 1.2087361373903633e-06, 'learning_rate/action_model': 1.2087361373903633e-05, 'epoch': 1.7}) +[2026-06-16 03:14:30,506][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.07946449518203735, 'train/grad_norm_pre_clip': 8.602381706237793, 'timing/data': 0.0003243149840272963, 'timing/model': 1.1386917040217668, 'learning_rate/qwen_vl_interface': 1.2071365931774584e-06, 'learning_rate/action_model': 1.2071365931774584e-05, 'epoch': 1.7}) +[2026-06-16 03:14:35,065][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.04996175318956375, 'train/grad_norm_pre_clip': 33.155704498291016, 'timing/data': 0.000355489959474653, 'timing/model': 1.1347686459776014, 'learning_rate/qwen_vl_interface': 1.2055387108329885e-06, 'learning_rate/action_model': 1.2055387108329886e-05, 'epoch': 1.7}) +[2026-06-16 03:14:39,574][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.0019318413687869906, 'train/grad_norm_pre_clip': 0.5062135457992554, 'timing/data': 0.0005897979717701674, 'timing/model': 1.1368471879977733, 'learning_rate/qwen_vl_interface': 1.2039424910137807e-06, 'learning_rate/action_model': 1.2039424910137806e-05, 'epoch': 1.7}) +[2026-06-16 03:14:44,132][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.004644094966351986, 'train/grad_norm_pre_clip': 2.0259921550750732, 'timing/data': 0.0003686770214699209, 'timing/model': 1.1406483700266108, 'learning_rate/qwen_vl_interface': 1.2023479343759822e-06, 'learning_rate/action_model': 1.2023479343759823e-05, 'epoch': 1.7}) +[2026-06-16 03:14:48,646][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.008345336653292179, 'train/grad_norm_pre_clip': 3.996307611465454, 'timing/data': 0.0003344239667057991, 'timing/model': 1.269929724978283, 'learning_rate/qwen_vl_interface': 1.2007550415750546e-06, 'learning_rate/action_model': 1.2007550415750545e-05, 'epoch': 1.7}) +[2026-06-16 03:14:53,046][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.0005864933482371271, 'train/grad_norm_pre_clip': 0.08066680282354355, 'timing/data': 0.0004225089796818793, 'timing/model': 1.1355105090187863, 'learning_rate/qwen_vl_interface': 1.1991638132657759e-06, 'learning_rate/action_model': 1.199163813265776e-05, 'epoch': 1.7}) +[2026-06-16 03:14:57,604][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.002691027708351612, 'train/grad_norm_pre_clip': 1.3524805307388306, 'timing/data': 0.00036423903657123446, 'timing/model': 1.1404907800024375, 'learning_rate/qwen_vl_interface': 1.1975742501022406e-06, 'learning_rate/action_model': 1.1975742501022405e-05, 'epoch': 1.7}) +[2026-06-16 03:15:02,167][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.00045990251237526536, 'train/grad_norm_pre_clip': 0.07771928608417511, 'timing/data': 0.0003445009933784604, 'timing/model': 1.136585101019591, 'learning_rate/qwen_vl_interface': 1.1959863527378569e-06, 'learning_rate/action_model': 1.1959863527378567e-05, 'epoch': 1.7}) +[2026-06-16 03:15:06,752][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.1106766015291214, 'train/grad_norm_pre_clip': 17.551273345947266, 'timing/data': 0.00035560602555051446, 'timing/model': 1.1571796049829572, 'learning_rate/qwen_vl_interface': 1.1944001218253505e-06, 'learning_rate/action_model': 1.1944001218253503e-05, 'epoch': 1.7}) +[2026-06-16 03:15:11,205][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.002461800817400217, 'train/grad_norm_pre_clip': 0.7495533227920532, 'timing/data': 0.0003470699884928763, 'timing/model': 1.135252305015456, 'learning_rate/qwen_vl_interface': 1.1928155580167608e-06, 'learning_rate/action_model': 1.1928155580167608e-05, 'epoch': 1.7}) +[2026-06-16 03:15:15,797][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.0009435510146431625, 'train/grad_norm_pre_clip': 0.18831437826156616, 'timing/data': 0.00035353697603568435, 'timing/model': 1.3302198310266249, 'learning_rate/qwen_vl_interface': 1.1912326619634412e-06, 'learning_rate/action_model': 1.1912326619634412e-05, 'epoch': 1.7}) +[2026-06-16 03:15:20,183][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.12597878277301788, 'train/grad_norm_pre_clip': 25.311368942260742, 'timing/data': 0.0003536499571055174, 'timing/model': 1.1354254460311495, 'learning_rate/qwen_vl_interface': 1.1896514343160632e-06, 'learning_rate/action_model': 1.189651434316063e-05, 'epoch': 1.7}) +[2026-06-16 03:15:24,730][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.0004946814733557403, 'train/grad_norm_pre_clip': 0.06710314005613327, 'timing/data': 0.00046788895269855857, 'timing/model': 1.1323118510190398, 'learning_rate/qwen_vl_interface': 1.1880718757246062e-06, 'learning_rate/action_model': 1.1880718757246063e-05, 'epoch': 1.7}) +[2026-06-16 03:15:29,258][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.005151534453034401, 'train/grad_norm_pre_clip': 2.233339309692383, 'timing/data': 0.0005500969709828496, 'timing/model': 1.1323559089796618, 'learning_rate/qwen_vl_interface': 1.186493986838369e-06, 'learning_rate/action_model': 1.186493986838369e-05, 'epoch': 1.71}) +[2026-06-16 03:15:33,793][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.0011055630166083574, 'train/grad_norm_pre_clip': 0.3123941123485565, 'timing/data': 0.00041762302862480283, 'timing/model': 1.1334999480168335, 'learning_rate/qwen_vl_interface': 1.1849177683059612e-06, 'learning_rate/action_model': 1.1849177683059612e-05, 'epoch': 1.71}) +[2026-06-16 03:15:38,357][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.0010806580539792776, 'train/grad_norm_pre_clip': 0.2590247690677643, 'timing/data': 0.0003339550457894802, 'timing/model': 1.3036056470009498, 'learning_rate/qwen_vl_interface': 1.1833432207753075e-06, 'learning_rate/action_model': 1.1833432207753076e-05, 'epoch': 1.71}) +[2026-06-16 03:15:42,741][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.0019013165729120374, 'train/grad_norm_pre_clip': 0.28949594497680664, 'timing/data': 0.0003317719674669206, 'timing/model': 1.135111307958141, 'learning_rate/qwen_vl_interface': 1.1817703448936448e-06, 'learning_rate/action_model': 1.181770344893645e-05, 'epoch': 1.71}) +[2026-06-16 03:15:47,283][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.03984671086072922, 'train/grad_norm_pre_clip': 11.682129859924316, 'timing/data': 0.0003589860280044377, 'timing/model': 1.1427076580002904, 'learning_rate/qwen_vl_interface': 1.1801991413075228e-06, 'learning_rate/action_model': 1.1801991413075227e-05, 'epoch': 1.71}) +[2026-06-16 03:15:51,711][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.002521752379834652, 'train/grad_norm_pre_clip': 0.6617957353591919, 'timing/data': 0.0003306239959783852, 'timing/model': 1.1383980379905552, 'learning_rate/qwen_vl_interface': 1.1786296106628038e-06, 'learning_rate/action_model': 1.1786296106628036e-05, 'epoch': 1.71}) +[2026-06-16 03:15:56,212][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.043652646243572235, 'train/grad_norm_pre_clip': 64.24976348876953, 'timing/data': 0.00036522699519991875, 'timing/model': 1.1333511139964685, 'learning_rate/qwen_vl_interface': 1.1770617536046633e-06, 'learning_rate/action_model': 1.1770617536046632e-05, 'epoch': 1.71}) +[2026-06-16 03:16:00,710][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.1924012452363968, 'train/grad_norm_pre_clip': 19.2200984954834, 'timing/data': 0.0003640939830802381, 'timing/model': 1.1317281300434843, 'learning_rate/qwen_vl_interface': 1.1754955707775873e-06, 'learning_rate/action_model': 1.1754955707775872e-05, 'epoch': 1.71}) +[2026-06-16 03:16:05,265][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.0007957989582791924, 'train/grad_norm_pre_clip': 0.16648933291435242, 'timing/data': 0.00035988097079098225, 'timing/model': 1.2960281109553762, 'learning_rate/qwen_vl_interface': 1.1739310628253753e-06, 'learning_rate/action_model': 1.1739310628253753e-05, 'epoch': 1.71}) +[2026-06-16 03:16:09,668][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.009427518583834171, 'train/grad_norm_pre_clip': 9.207147598266602, 'timing/data': 0.00034721597330644727, 'timing/model': 1.1393729359842837, 'learning_rate/qwen_vl_interface': 1.172368230391136e-06, 'learning_rate/action_model': 1.1723682303911359e-05, 'epoch': 1.71}) +[2026-06-16 03:16:14,210][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.006424473132938147, 'train/grad_norm_pre_clip': 1.5595139265060425, 'timing/data': 0.00038636696990579367, 'timing/model': 1.138098708994221, 'learning_rate/qwen_vl_interface': 1.1708070741172928e-06, 'learning_rate/action_model': 1.1708070741172928e-05, 'epoch': 1.71}) +[2026-06-16 03:16:18,729][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.0400247648358345, 'train/grad_norm_pre_clip': 50.14932632446289, 'timing/data': 0.000344510015565902, 'timing/model': 1.1327488230308518, 'learning_rate/qwen_vl_interface': 1.1692475946455755e-06, 'learning_rate/action_model': 1.1692475946455755e-05, 'epoch': 1.71}) +[2026-06-16 03:16:23,261][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.0014533032663166523, 'train/grad_norm_pre_clip': 0.49587228894233704, 'timing/data': 0.00030985299963504076, 'timing/model': 1.137894409999717, 'learning_rate/qwen_vl_interface': 1.1676897926170289e-06, 'learning_rate/action_model': 1.1676897926170289e-05, 'epoch': 1.71}) +[2026-06-16 03:16:27,805][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.1401652991771698, 'train/grad_norm_pre_clip': 12.4705171585083, 'timing/data': 0.0003497989964671433, 'timing/model': 1.132923536002636, 'learning_rate/qwen_vl_interface': 1.166133668672006e-06, 'learning_rate/action_model': 1.166133668672006e-05, 'epoch': 1.71}) +[2026-06-16 03:16:32,340][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.01300050038844347, 'train/grad_norm_pre_clip': 7.191221714019775, 'timing/data': 0.00032367301173508167, 'timing/model': 1.296630628989078, 'learning_rate/qwen_vl_interface': 1.1645792234501706e-06, 'learning_rate/action_model': 1.1645792234501706e-05, 'epoch': 1.71}) +[2026-06-16 03:16:36,727][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.00021661596838384867, 'train/grad_norm_pre_clip': 0.04170715808868408, 'timing/data': 0.00031580403447151184, 'timing/model': 1.1346878719632514, 'learning_rate/qwen_vl_interface': 1.1630264575904972e-06, 'learning_rate/action_model': 1.1630264575904973e-05, 'epoch': 1.71}) +[2026-06-16 03:16:41,268][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.0002207343786722049, 'train/grad_norm_pre_clip': 0.05464087054133415, 'timing/data': 0.00033763598185032606, 'timing/model': 1.1356176679837517, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 1.71}) +[2026-06-16 03:16:45,832][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.04926940053701401, 'train/grad_norm_pre_clip': 43.1169548034668, 'timing/data': 0.00036755495239049196, 'timing/model': 1.145343686977867, 'learning_rate/qwen_vl_interface': 1.1599259665100767e-06, 'learning_rate/action_model': 1.1599259665100767e-05, 'epoch': 1.71}) +[2026-06-16 03:16:50,394][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.04030425846576691, 'train/grad_norm_pre_clip': 30.80352020263672, 'timing/data': 0.0003722609835676849, 'timing/model': 1.1390931059722789, 'learning_rate/qwen_vl_interface': 1.1583782425638252e-06, 'learning_rate/action_model': 1.158378242563825e-05, 'epoch': 1.71}) +[2026-06-16 03:16:54,902][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.22652003169059753, 'train/grad_norm_pre_clip': 23.12547492980957, 'timing/data': 0.0003555219736881554, 'timing/model': 1.1340458529884927, 'learning_rate/qwen_vl_interface': 1.1568322005287237e-06, 'learning_rate/action_model': 1.1568322005287235e-05, 'epoch': 1.71}) +[2026-06-16 03:16:59,444][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.21803490817546844, 'train/grad_norm_pre_clip': 8.920999526977539, 'timing/data': 0.0004775460110977292, 'timing/model': 1.2963974379817955, 'learning_rate/qwen_vl_interface': 1.1552878410402923e-06, 'learning_rate/action_model': 1.1552878410402924e-05, 'epoch': 1.71}) +[2026-06-16 03:17:03,828][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.0010529827559366822, 'train/grad_norm_pre_clip': 0.5147997736930847, 'timing/data': 0.00041454797610640526, 'timing/model': 1.1335150249651633, 'learning_rate/qwen_vl_interface': 1.1537451647333594e-06, 'learning_rate/action_model': 1.1537451647333593e-05, 'epoch': 1.71}) +[2026-06-16 03:17:08,365][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.04215451702475548, 'train/grad_norm_pre_clip': 23.385562896728516, 'timing/data': 0.0003483740147203207, 'timing/model': 1.1414479960221797, 'learning_rate/qwen_vl_interface': 1.1522041722420594e-06, 'learning_rate/action_model': 1.1522041722420594e-05, 'epoch': 1.71}) +[2026-06-16 03:17:12,912][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.061810776591300964, 'train/grad_norm_pre_clip': 23.336376190185547, 'timing/data': 0.0003558690077625215, 'timing/model': 1.1410046410164796, 'learning_rate/qwen_vl_interface': 1.1506648641998365e-06, 'learning_rate/action_model': 1.1506648641998365e-05, 'epoch': 1.71}) +[2026-06-16 03:17:17,451][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 0.0006236364715732634, 'train/grad_norm_pre_clip': 0.332924485206604, 'timing/data': 0.0003705580020323396, 'timing/model': 1.1357958890148439, 'learning_rate/qwen_vl_interface': 1.1491272412394433e-06, 'learning_rate/action_model': 1.1491272412394433e-05, 'epoch': 1.72}) +[2026-06-16 03:17:22,015][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.0002152392262360081, 'train/grad_norm_pre_clip': 0.04980695992708206, 'timing/data': 0.00036351598100736737, 'timing/model': 1.1365664409822784, 'learning_rate/qwen_vl_interface': 1.1475913039929373e-06, 'learning_rate/action_model': 1.1475913039929372e-05, 'epoch': 1.72}) +[2026-06-16 03:17:26,577][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.00017063012637663633, 'train/grad_norm_pre_clip': 0.04814235866069794, 'timing/data': 0.00033800594974309206, 'timing/model': 1.3017907839966938, 'learning_rate/qwen_vl_interface': 1.1460570530916844e-06, 'learning_rate/action_model': 1.1460570530916844e-05, 'epoch': 1.72}) +[2026-06-16 03:17:30,974][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.20965027809143066, 'train/grad_norm_pre_clip': 53.5281982421875, 'timing/data': 0.0003678039647638798, 'timing/model': 1.1507486659684218, 'learning_rate/qwen_vl_interface': 1.1445244891663581e-06, 'learning_rate/action_model': 1.1445244891663581e-05, 'epoch': 1.72}) +[2026-06-16 03:17:35,511][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 0.003940790891647339, 'train/grad_norm_pre_clip': 1.689726710319519, 'timing/data': 0.000363522965926677, 'timing/model': 1.1332470970228314, 'learning_rate/qwen_vl_interface': 1.1429936128469366e-06, 'learning_rate/action_model': 1.1429936128469366e-05, 'epoch': 1.72}) +[2026-06-16 03:17:40,060][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.0031165885739028454, 'train/grad_norm_pre_clip': 1.3242915868759155, 'timing/data': 0.0005141160218045115, 'timing/model': 1.1372401230037212, 'learning_rate/qwen_vl_interface': 1.1414644247627056e-06, 'learning_rate/action_model': 1.1414644247627058e-05, 'epoch': 1.72}) +[2026-06-16 03:17:44,596][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.00016100115317385644, 'train/grad_norm_pre_clip': 0.02612820267677307, 'timing/data': 0.00036379502853378654, 'timing/model': 1.1430048960028216, 'learning_rate/qwen_vl_interface': 1.1399369255422568e-06, 'learning_rate/action_model': 1.1399369255422568e-05, 'epoch': 1.72}) +[2026-06-16 03:17:49,093][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.0040124645456671715, 'train/grad_norm_pre_clip': 2.6077706813812256, 'timing/data': 0.0003554639988578856, 'timing/model': 1.1337300089653581, 'learning_rate/qwen_vl_interface': 1.138411115813488e-06, 'learning_rate/action_model': 1.138411115813488e-05, 'epoch': 1.72}) +[2026-06-16 03:17:53,648][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.15853820741176605, 'train/grad_norm_pre_clip': 61.736759185791016, 'timing/data': 0.0003495080163702369, 'timing/model': 1.3025411100243218, 'learning_rate/qwen_vl_interface': 1.1368869962036006e-06, 'learning_rate/action_model': 1.1368869962036005e-05, 'epoch': 1.72}) +[2026-06-16 03:17:58,031][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.07097085565328598, 'train/grad_norm_pre_clip': 33.013221740722656, 'timing/data': 0.00033478502882644534, 'timing/model': 1.1350863879779354, 'learning_rate/qwen_vl_interface': 1.1353645673391025e-06, 'learning_rate/action_model': 1.1353645673391026e-05, 'epoch': 1.72}) +[2026-06-16 03:18:02,561][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.009769361466169357, 'train/grad_norm_pre_clip': 5.990302085876465, 'timing/data': 0.0003289239830337465, 'timing/model': 1.132452848018147, 'learning_rate/qwen_vl_interface': 1.1338438298458078e-06, 'learning_rate/action_model': 1.1338438298458078e-05, 'epoch': 1.72}) +[2026-06-16 03:18:07,116][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.07677526772022247, 'train/grad_norm_pre_clip': 48.88861083984375, 'timing/data': 0.00036845699651166797, 'timing/model': 1.1400263870018534, 'learning_rate/qwen_vl_interface': 1.1323247843488345e-06, 'learning_rate/action_model': 1.1323247843488344e-05, 'epoch': 1.72}) +[2026-06-16 03:18:11,656][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.11974970251321793, 'train/grad_norm_pre_clip': 26.35434341430664, 'timing/data': 0.0003415770479477942, 'timing/model': 1.1342573179863393, 'learning_rate/qwen_vl_interface': 1.130807431472603e-06, 'learning_rate/action_model': 1.130807431472603e-05, 'epoch': 1.72}) +[2026-06-16 03:18:16,212][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.12277018278837204, 'train/grad_norm_pre_clip': 16.43773078918457, 'timing/data': 0.00039501499850302935, 'timing/model': 1.1348561170161702, 'learning_rate/qwen_vl_interface': 1.1292917718408416e-06, 'learning_rate/action_model': 1.1292917718408416e-05, 'epoch': 1.72}) +[2026-06-16 03:18:20,774][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.008799185045063496, 'train/grad_norm_pre_clip': 4.187897205352783, 'timing/data': 0.00034787104232236743, 'timing/model': 1.303694483009167, 'learning_rate/qwen_vl_interface': 1.12777780607658e-06, 'learning_rate/action_model': 1.12777780607658e-05, 'epoch': 1.72}) +[2026-06-16 03:18:25,146][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.015388330444693565, 'train/grad_norm_pre_clip': 10.293996810913086, 'timing/data': 0.00043757999083027244, 'timing/model': 1.133509737032, 'learning_rate/qwen_vl_interface': 1.1262655348021523e-06, 'learning_rate/action_model': 1.1262655348021523e-05, 'epoch': 1.72}) +[2026-06-16 03:18:29,732][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.010724782012403011, 'train/grad_norm_pre_clip': 5.9921650886535645, 'timing/data': 0.00047413300490006804, 'timing/model': 1.1346691059879959, 'learning_rate/qwen_vl_interface': 1.1247549586391966e-06, 'learning_rate/action_model': 1.1247549586391964e-05, 'epoch': 1.72}) +[2026-06-16 03:18:34,305][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.0478842668235302, 'train/grad_norm_pre_clip': 28.651857376098633, 'timing/data': 0.00036033301148563623, 'timing/model': 1.1334981370018795, 'learning_rate/qwen_vl_interface': 1.1232460782086534e-06, 'learning_rate/action_model': 1.1232460782086534e-05, 'epoch': 1.72}) +[2026-06-16 03:18:38,858][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.035938479006290436, 'train/grad_norm_pre_clip': 20.435941696166992, 'timing/data': 0.00035390100674703717, 'timing/model': 1.13209516601637, 'learning_rate/qwen_vl_interface': 1.1217388941307658e-06, 'learning_rate/action_model': 1.1217388941307658e-05, 'epoch': 1.72}) +[2026-06-16 03:18:43,378][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.0006162999197840691, 'train/grad_norm_pre_clip': 0.12240324914455414, 'timing/data': 0.0003134759608656168, 'timing/model': 1.1457424209802411, 'learning_rate/qwen_vl_interface': 1.1202334070250807e-06, 'learning_rate/action_model': 1.1202334070250806e-05, 'epoch': 1.72}) +[2026-06-16 03:18:47,952][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.0012428720947355032, 'train/grad_norm_pre_clip': 0.710964024066925, 'timing/data': 0.00032392999855801463, 'timing/model': 1.3069968009949662, 'learning_rate/qwen_vl_interface': 1.1187296175104472e-06, 'learning_rate/action_model': 1.1187296175104471e-05, 'epoch': 1.72}) +[2026-06-16 03:18:52,365][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.0011627784697338939, 'train/grad_norm_pre_clip': 0.5734642148017883, 'timing/data': 0.00035634904634207487, 'timing/model': 1.1369961379677989, 'learning_rate/qwen_vl_interface': 1.1172275262050158e-06, 'learning_rate/action_model': 1.1172275262050159e-05, 'epoch': 1.72}) +[2026-06-16 03:18:56,919][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.19826702773571014, 'train/grad_norm_pre_clip': 13.745984077453613, 'timing/data': 0.00033656798768788576, 'timing/model': 1.1346075420151465, 'learning_rate/qwen_vl_interface': 1.1157271337262401e-06, 'learning_rate/action_model': 1.11572713372624e-05, 'epoch': 1.72}) +[2026-06-16 03:19:01,455][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 0.033299386501312256, 'train/grad_norm_pre_clip': 19.53153419494629, 'timing/data': 0.0003450429649092257, 'timing/model': 1.1332723579835147, 'learning_rate/qwen_vl_interface': 1.1142284406908751e-06, 'learning_rate/action_model': 1.114228440690875e-05, 'epoch': 1.72}) +[2026-06-16 03:19:05,995][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.00011076070950366557, 'train/grad_norm_pre_clip': 0.033240918070077896, 'timing/data': 0.00032180798007175326, 'timing/model': 1.1327051309635863, 'learning_rate/qwen_vl_interface': 1.1127314477149758e-06, 'learning_rate/action_model': 1.1127314477149759e-05, 'epoch': 1.72}) +[2026-06-16 03:23:29,002][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.003538883989676833, 'train/grad_norm_pre_clip': 0.8427136540412903, 'eval/loss': 0.04380495101213455, 'eval/flap_precision': 0.9666160849772383, 'eval/flap_recall': 0.9830246913580247, 'eval/flap_f1': 0.9747513389441469, 'eval/flap_rate_model': 0.18305555555555555, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9747513389441469, 'timing/data': 0.00035766500514000654, 'timing/model': 1.1413024060311727, 'learning_rate/qwen_vl_interface': 1.1112361554139003e-06, 'learning_rate/action_model': 1.1112361554139004e-05, 'epoch': 1.73}) +[2026-06-16 03:23:29,002][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4200_state +[2026-06-16 03:23:29,005][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 03:23:32,587][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4200_state/model.safetensors +[2026-06-16 03:23:42,647][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4200_state/optimizer.bin +[2026-06-16 03:23:42,647][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4200_state/sampler.bin +[2026-06-16 03:23:42,647][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4200_state/sampler_1.bin +[2026-06-16 03:23:42,648][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4200_state/random_states_0.pkl +[2026-06-16 03:23:42,695][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4200 +[2026-06-16 03:23:42,695][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 03:23:42,702][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 03:23:47,153][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.011936917901039124, 'train/grad_norm_pre_clip': 5.683610439300537, 'timing/data': 0.00033730495488271117, 'timing/model': 1.1207370199845172, 'learning_rate/qwen_vl_interface': 1.1097425644023065e-06, 'learning_rate/action_model': 1.1097425644023065e-05, 'epoch': 1.73}) +[2026-06-16 03:23:51,635][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 0.0005295701557770371, 'train/grad_norm_pre_clip': 0.17042051255702972, 'timing/data': 0.0003543890197761357, 'timing/model': 1.2816017100121826, 'learning_rate/qwen_vl_interface': 1.1082506752941529e-06, 'learning_rate/action_model': 1.1082506752941528e-05, 'epoch': 1.73}) +[2026-06-16 03:23:55,970][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.0006188297993503511, 'train/grad_norm_pre_clip': 0.18603353202342987, 'timing/data': 0.0003119480097666383, 'timing/model': 1.1229688549647108, 'learning_rate/qwen_vl_interface': 1.1067604887027004e-06, 'learning_rate/action_model': 1.1067604887027004e-05, 'epoch': 1.73}) +[2026-06-16 03:24:00,454][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.14212040603160858, 'train/grad_norm_pre_clip': 20.291276931762695, 'timing/data': 0.00032298703445121646, 'timing/model': 1.1206590610090643, 'learning_rate/qwen_vl_interface': 1.1052720052405066e-06, 'learning_rate/action_model': 1.1052720052405064e-05, 'epoch': 1.73}) +[2026-06-16 03:24:04,858][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.12434626370668411, 'train/grad_norm_pre_clip': 15.877265930175781, 'timing/data': 0.00034085800871253014, 'timing/model': 1.1282883770181797, 'learning_rate/qwen_vl_interface': 1.1037852255194309e-06, 'learning_rate/action_model': 1.1037852255194308e-05, 'epoch': 1.73}) +[2026-06-16 03:24:09,288][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 0.15216299891471863, 'train/grad_norm_pre_clip': 15.676626205444336, 'timing/data': 0.0003210349823348224, 'timing/model': 1.118701747036539, 'learning_rate/qwen_vl_interface': 1.1023001501506329e-06, 'learning_rate/action_model': 1.102300150150633e-05, 'epoch': 1.73}) +[2026-06-16 03:24:13,753][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.013732785359025002, 'train/grad_norm_pre_clip': 6.570760250091553, 'timing/data': 0.00036440801341086626, 'timing/model': 1.1296643249806948, 'learning_rate/qwen_vl_interface': 1.1008167797445717e-06, 'learning_rate/action_model': 1.1008167797445718e-05, 'epoch': 1.73}) +[2026-06-16 03:24:18,248][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.003304904093965888, 'train/grad_norm_pre_clip': 1.3284698724746704, 'timing/data': 0.000329185975715518, 'timing/model': 1.2712587699643336, 'learning_rate/qwen_vl_interface': 1.099335114911003e-06, 'learning_rate/action_model': 1.0993351149110031e-05, 'epoch': 1.73}) +[2026-06-16 03:24:22,603][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.16747763752937317, 'train/grad_norm_pre_clip': 6.388166904449463, 'timing/data': 0.0005957520334050059, 'timing/model': 1.1289423549897037, 'learning_rate/qwen_vl_interface': 1.0978551562589846e-06, 'learning_rate/action_model': 1.0978551562589845e-05, 'epoch': 1.73}) +[2026-06-16 03:24:27,105][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.00027439440600574017, 'train/grad_norm_pre_clip': 0.034548576921224594, 'timing/data': 0.0003571310080587864, 'timing/model': 1.1277157639851794, 'learning_rate/qwen_vl_interface': 1.0963769043968708e-06, 'learning_rate/action_model': 1.0963769043968708e-05, 'epoch': 1.73}) +[2026-06-16 03:24:31,595][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.10831805318593979, 'train/grad_norm_pre_clip': 13.697314262390137, 'timing/data': 0.0004300200380384922, 'timing/model': 1.1227674820111133, 'learning_rate/qwen_vl_interface': 1.0949003599323158e-06, 'learning_rate/action_model': 1.0949003599323157e-05, 'epoch': 1.73}) +[2026-06-16 03:24:36,126][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 9.593317372491583e-05, 'train/grad_norm_pre_clip': 0.017442801967263222, 'timing/data': 0.00035776698496192694, 'timing/model': 1.1483451660024002, 'learning_rate/qwen_vl_interface': 1.0934255234722708e-06, 'learning_rate/action_model': 1.0934255234722706e-05, 'epoch': 1.73}) +[2026-06-16 03:24:40,639][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.2283545732498169, 'train/grad_norm_pre_clip': 10.781039237976074, 'timing/data': 0.0006145630031824112, 'timing/model': 1.136160047026351, 'learning_rate/qwen_vl_interface': 1.0919523956229858e-06, 'learning_rate/action_model': 1.0919523956229857e-05, 'epoch': 1.73}) +[2026-06-16 03:24:45,170][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.010665989480912685, 'train/grad_norm_pre_clip': 10.644726753234863, 'timing/data': 0.00035721296444535255, 'timing/model': 1.289308138017077, 'learning_rate/qwen_vl_interface': 1.0904809769900075e-06, 'learning_rate/action_model': 1.0904809769900073e-05, 'epoch': 1.73}) +[2026-06-16 03:24:49,530][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.0003358706599101424, 'train/grad_norm_pre_clip': 0.06556617468595505, 'timing/data': 0.00036667700624093413, 'timing/model': 1.133558715984691, 'learning_rate/qwen_vl_interface': 1.0890112681781821e-06, 'learning_rate/action_model': 1.089011268178182e-05, 'epoch': 1.73}) +[2026-06-16 03:24:54,060][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.2629994750022888, 'train/grad_norm_pre_clip': 32.96856689453125, 'timing/data': 0.0003273659967817366, 'timing/model': 1.1287459330051206, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 1.73}) +[2026-06-16 03:24:58,591][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.02664840966463089, 'train/grad_norm_pre_clip': 17.16202163696289, 'timing/data': 0.0003609880222938955, 'timing/model': 1.133899056003429, 'learning_rate/qwen_vl_interface': 1.0860769824338477e-06, 'learning_rate/action_model': 1.0860769824338478e-05, 'epoch': 1.73}) +[2026-06-16 03:25:03,099][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.00036146858474239707, 'train/grad_norm_pre_clip': 0.10900773108005524, 'timing/data': 0.000542176014278084, 'timing/model': 1.1293790300260298, 'learning_rate/qwen_vl_interface': 1.084612406707514e-06, 'learning_rate/action_model': 1.084612406707514e-05, 'epoch': 1.73}) +[2026-06-16 03:25:07,589][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.08114959299564362, 'train/grad_norm_pre_clip': 35.84030532836914, 'timing/data': 0.0003192899748682976, 'timing/model': 1.1271047099726275, 'learning_rate/qwen_vl_interface': 1.0831495432146795e-06, 'learning_rate/action_model': 1.0831495432146795e-05, 'epoch': 1.73}) +[2026-06-16 03:25:12,135][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.003204144537448883, 'train/grad_norm_pre_clip': 7.1083984375, 'timing/data': 0.0005588180501945317, 'timing/model': 1.3081497170496732, 'learning_rate/qwen_vl_interface': 1.0816883925566717e-06, 'learning_rate/action_model': 1.0816883925566716e-05, 'epoch': 1.73}) +[2026-06-16 03:25:16,533][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.00016615630011074245, 'train/grad_norm_pre_clip': 0.046124499291181564, 'timing/data': 0.00032773002749308944, 'timing/model': 1.1438637140090577, 'learning_rate/qwen_vl_interface': 1.0802289553341153e-06, 'learning_rate/action_model': 1.0802289553341152e-05, 'epoch': 1.73}) +[2026-06-16 03:25:21,050][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.3619203269481659, 'train/grad_norm_pre_clip': 13.28954792022705, 'timing/data': 0.0003954159910790622, 'timing/model': 1.1294240609859116, 'learning_rate/qwen_vl_interface': 1.0787712321469289e-06, 'learning_rate/action_model': 1.0787712321469288e-05, 'epoch': 1.73}) +[2026-06-16 03:25:25,597][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.23507028818130493, 'train/grad_norm_pre_clip': 6.184031963348389, 'timing/data': 0.00036091002402827144, 'timing/model': 1.1304723599459976, 'learning_rate/qwen_vl_interface': 1.077315223594328e-06, 'learning_rate/action_model': 1.077315223594328e-05, 'epoch': 1.73}) +[2026-06-16 03:25:30,131][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.2197534590959549, 'train/grad_norm_pre_clip': 22.63684844970703, 'timing/data': 0.0003402379807084799, 'timing/model': 1.1410721049760468, 'learning_rate/qwen_vl_interface': 1.0758609302748227e-06, 'learning_rate/action_model': 1.0758609302748227e-05, 'epoch': 1.74}) +[2026-06-16 03:25:34,649][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.1810649186372757, 'train/grad_norm_pre_clip': 25.62730598449707, 'timing/data': 0.00032863504020497203, 'timing/model': 1.1277442729915492, 'learning_rate/qwen_vl_interface': 1.0744083527862182e-06, 'learning_rate/action_model': 1.074408352786218e-05, 'epoch': 1.74}) +[2026-06-16 03:25:39,204][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.21193775534629822, 'train/grad_norm_pre_clip': 31.705936431884766, 'timing/data': 0.0003291610046289861, 'timing/model': 1.3015300540137105, 'learning_rate/qwen_vl_interface': 1.0729574917256144e-06, 'learning_rate/action_model': 1.0729574917256145e-05, 'epoch': 1.74}) +[2026-06-16 03:25:43,577][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.11707957088947296, 'train/grad_norm_pre_clip': 27.510839462280273, 'timing/data': 0.0003573889844119549, 'timing/model': 1.1333989769918844, 'learning_rate/qwen_vl_interface': 1.0715083476894047e-06, 'learning_rate/action_model': 1.0715083476894047e-05, 'epoch': 1.74}) +[2026-06-16 03:25:48,117][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 0.07079198956489563, 'train/grad_norm_pre_clip': 14.490310668945312, 'timing/data': 0.00036165700294077396, 'timing/model': 1.1327593880123459, 'learning_rate/qwen_vl_interface': 1.0700609212732773e-06, 'learning_rate/action_model': 1.0700609212732774e-05, 'epoch': 1.74}) +[2026-06-16 03:25:52,638][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.026871846988797188, 'train/grad_norm_pre_clip': 16.99172592163086, 'timing/data': 0.0003545529907569289, 'timing/model': 1.1317513379617594, 'learning_rate/qwen_vl_interface': 1.0686152130722155e-06, 'learning_rate/action_model': 1.0686152130722154e-05, 'epoch': 1.74}) +[2026-06-16 03:25:57,166][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.00021597016893792897, 'train/grad_norm_pre_clip': 0.029531313106417656, 'timing/data': 0.00039299100171774626, 'timing/model': 1.133549497986678, 'learning_rate/qwen_vl_interface': 1.0671712236804947e-06, 'learning_rate/action_model': 1.0671712236804946e-05, 'epoch': 1.74}) +[2026-06-16 03:26:01,657][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 0.015296589583158493, 'train/grad_norm_pre_clip': 8.359757423400879, 'timing/data': 0.0003427910269238055, 'timing/model': 1.1291827380191535, 'learning_rate/qwen_vl_interface': 1.0657289536916845e-06, 'learning_rate/action_model': 1.0657289536916845e-05, 'epoch': 1.74}) +[2026-06-16 03:26:06,198][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.002260511042550206, 'train/grad_norm_pre_clip': 1.1957069635391235, 'timing/data': 0.0003626049729064107, 'timing/model': 1.2908884100033902, 'learning_rate/qwen_vl_interface': 1.0642884036986481e-06, 'learning_rate/action_model': 1.064288403698648e-05, 'epoch': 1.74}) +[2026-06-16 03:26:10,578][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.037038907408714294, 'train/grad_norm_pre_clip': 53.66073989868164, 'timing/data': 0.0003834889503195882, 'timing/model': 1.1346236490062438, 'learning_rate/qwen_vl_interface': 1.0628495742935404e-06, 'learning_rate/action_model': 1.0628495742935405e-05, 'epoch': 1.74}) +[2026-06-16 03:26:15,119][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 0.0002537713444326073, 'train/grad_norm_pre_clip': 0.070052869617939, 'timing/data': 0.00035408197436481714, 'timing/model': 1.135483457997907, 'learning_rate/qwen_vl_interface': 1.0614124660678103e-06, 'learning_rate/action_model': 1.0614124660678102e-05, 'epoch': 1.74}) +[2026-06-16 03:26:19,664][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.12197203934192657, 'train/grad_norm_pre_clip': 6.307433128356934, 'timing/data': 0.00034138699993491173, 'timing/model': 1.1348023969912902, 'learning_rate/qwen_vl_interface': 1.0599770796121988e-06, 'learning_rate/action_model': 1.0599770796121987e-05, 'epoch': 1.74}) +[2026-06-16 03:26:24,185][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.2361573874950409, 'train/grad_norm_pre_clip': 26.371503829956055, 'timing/data': 0.00046042900066822767, 'timing/model': 1.1308098360314034, 'learning_rate/qwen_vl_interface': 1.0585434155167394e-06, 'learning_rate/action_model': 1.0585434155167393e-05, 'epoch': 1.74}) +[2026-06-16 03:26:28,716][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.18104015290737152, 'train/grad_norm_pre_clip': 9.129889488220215, 'timing/data': 0.0003967279917560518, 'timing/model': 1.1334031599690206, 'learning_rate/qwen_vl_interface': 1.0571114743707556e-06, 'learning_rate/action_model': 1.0571114743707555e-05, 'epoch': 1.74}) +[2026-06-16 03:26:33,248][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.029348785057663918, 'train/grad_norm_pre_clip': 11.971351623535156, 'timing/data': 0.00032557803206145763, 'timing/model': 1.2855519659933634, 'learning_rate/qwen_vl_interface': 1.0556812567628661e-06, 'learning_rate/action_model': 1.055681256762866e-05, 'epoch': 1.74}) +[2026-06-16 03:26:37,625][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.16822127997875214, 'train/grad_norm_pre_clip': 15.63941478729248, 'timing/data': 0.0003435090184211731, 'timing/model': 1.1324206010322087, 'learning_rate/qwen_vl_interface': 1.0542527632809783e-06, 'learning_rate/action_model': 1.0542527632809784e-05, 'epoch': 1.74}) +[2026-06-16 03:26:42,152][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.12678025662899017, 'train/grad_norm_pre_clip': 15.052482604980469, 'timing/data': 0.0003331669722683728, 'timing/model': 1.1305636310135014, 'learning_rate/qwen_vl_interface': 1.0528259945122925e-06, 'learning_rate/action_model': 1.0528259945122925e-05, 'epoch': 1.74}) +[2026-06-16 03:26:46,714][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.0064210654236376286, 'train/grad_norm_pre_clip': 5.970911502838135, 'timing/data': 0.00039591401582583785, 'timing/model': 1.1430062029976398, 'learning_rate/qwen_vl_interface': 1.0514009510432988e-06, 'learning_rate/action_model': 1.0514009510432988e-05, 'epoch': 1.74}) +[2026-06-16 03:26:51,281][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.0338352732360363, 'train/grad_norm_pre_clip': 17.884843826293945, 'timing/data': 0.0003495349665172398, 'timing/model': 1.1355609800084494, 'learning_rate/qwen_vl_interface': 1.0499776334597796e-06, 'learning_rate/action_model': 1.0499776334597797e-05, 'epoch': 1.74}) +[2026-06-16 03:26:55,781][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.0467539019882679, 'train/grad_norm_pre_clip': 12.4122314453125, 'timing/data': 0.00033512397203594446, 'timing/model': 1.1311969539965503, 'learning_rate/qwen_vl_interface': 1.0485560423468064e-06, 'learning_rate/action_model': 1.0485560423468063e-05, 'epoch': 1.74}) +[2026-06-16 03:27:00,345][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.0006479925941675901, 'train/grad_norm_pre_clip': 0.09347375482320786, 'timing/data': 0.0004141890094615519, 'timing/model': 1.2994690429768525, 'learning_rate/qwen_vl_interface': 1.0471361782887421e-06, 'learning_rate/action_model': 1.0471361782887421e-05, 'epoch': 1.74}) +[2026-06-16 03:27:04,743][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.1285526305437088, 'train/grad_norm_pre_clip': 26.189720153808594, 'timing/data': 0.0003170489799231291, 'timing/model': 1.1457619550055824, 'learning_rate/qwen_vl_interface': 1.0457180418692385e-06, 'learning_rate/action_model': 1.0457180418692383e-05, 'epoch': 1.74}) +[2026-06-16 03:27:09,293][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.15565340220928192, 'train/grad_norm_pre_clip': 14.80624771118164, 'timing/data': 0.0003905160119757056, 'timing/model': 1.1426646889885888, 'learning_rate/qwen_vl_interface': 1.0443016336712395e-06, 'learning_rate/action_model': 1.0443016336712393e-05, 'epoch': 1.74}) +[2026-06-16 03:27:13,852][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.1610139161348343, 'train/grad_norm_pre_clip': 20.4012451171875, 'timing/data': 0.000364648993127048, 'timing/model': 1.1600138990324922, 'learning_rate/qwen_vl_interface': 1.0428869542769744e-06, 'learning_rate/action_model': 1.0428869542769744e-05, 'epoch': 1.74}) +[2026-06-16 03:27:18,385][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.25492000579833984, 'train/grad_norm_pre_clip': 16.6772518157959, 'timing/data': 0.00040511501720175147, 'timing/model': 1.1341767910053022, 'learning_rate/qwen_vl_interface': 1.0414740042679663e-06, 'learning_rate/action_model': 1.0414740042679662e-05, 'epoch': 1.75}) +[2026-06-16 03:27:22,926][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.2533144950866699, 'train/grad_norm_pre_clip': 21.151058197021484, 'timing/data': 0.0005554300150834024, 'timing/model': 1.1382035280112177, 'learning_rate/qwen_vl_interface': 1.040062784225025e-06, 'learning_rate/action_model': 1.0400627842250249e-05, 'epoch': 1.75}) +[2026-06-16 03:27:27,486][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.286832332611084, 'train/grad_norm_pre_clip': 16.606998443603516, 'timing/data': 0.0003675789921544492, 'timing/model': 1.3094466680195183, 'learning_rate/qwen_vl_interface': 1.0386532947282496e-06, 'learning_rate/action_model': 1.0386532947282495e-05, 'epoch': 1.75}) +[2026-06-16 03:27:31,894][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.007223095279186964, 'train/grad_norm_pre_clip': 3.78287935256958, 'timing/data': 0.00035229400964453816, 'timing/model': 1.1425791790243238, 'learning_rate/qwen_vl_interface': 1.0372455363570282e-06, 'learning_rate/action_model': 1.0372455363570283e-05, 'epoch': 1.75}) +[2026-06-16 03:27:36,452][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.08240386843681335, 'train/grad_norm_pre_clip': 21.472076416015625, 'timing/data': 0.0006375089869834483, 'timing/model': 1.138771791011095, 'learning_rate/qwen_vl_interface': 1.035839509690039e-06, 'learning_rate/action_model': 1.0358395096900388e-05, 'epoch': 1.75}) +[2026-06-16 03:27:41,007][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.12560142576694489, 'train/grad_norm_pre_clip': 10.751322746276855, 'timing/data': 0.000332510971929878, 'timing/model': 1.1344750989810564, 'learning_rate/qwen_vl_interface': 1.0344352153052429e-06, 'learning_rate/action_model': 1.0344352153052427e-05, 'epoch': 1.75}) +[2026-06-16 03:27:45,539][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.16743914783000946, 'train/grad_norm_pre_clip': 17.391311645507812, 'timing/data': 0.0003582090139389038, 'timing/model': 1.133927904011216, 'learning_rate/qwen_vl_interface': 1.033032653779894e-06, 'learning_rate/action_model': 1.033032653779894e-05, 'epoch': 1.75}) +[2026-06-16 03:27:50,046][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.09560247510671616, 'train/grad_norm_pre_clip': 22.250577926635742, 'timing/data': 0.0003952429979108274, 'timing/model': 1.1400276380009018, 'learning_rate/qwen_vl_interface': 1.0316318256905325e-06, 'learning_rate/action_model': 1.0316318256905324e-05, 'epoch': 1.75}) +[2026-06-16 03:27:54,613][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.0010044738883152604, 'train/grad_norm_pre_clip': 0.18980608880519867, 'timing/data': 0.00035664200549945235, 'timing/model': 1.291718927968759, 'learning_rate/qwen_vl_interface': 1.0302327316129856e-06, 'learning_rate/action_model': 1.0302327316129855e-05, 'epoch': 1.75}) +[2026-06-16 03:27:58,985][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.05957576259970665, 'train/grad_norm_pre_clip': 9.922013282775879, 'timing/data': 0.00039595196722075343, 'timing/model': 1.1306207259767689, 'learning_rate/qwen_vl_interface': 1.028835372122368e-06, 'learning_rate/action_model': 1.0288353721223682e-05, 'epoch': 1.75}) +[2026-06-16 03:28:03,520][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 0.0678735077381134, 'train/grad_norm_pre_clip': 8.260663032531738, 'timing/data': 0.00035589002072811127, 'timing/model': 1.1382730570039712, 'learning_rate/qwen_vl_interface': 1.0274397477930822e-06, 'learning_rate/action_model': 1.0274397477930822e-05, 'epoch': 1.75}) +[2026-06-16 03:28:08,062][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.11778957396745682, 'train/grad_norm_pre_clip': 12.956351280212402, 'timing/data': 0.00036136299604550004, 'timing/model': 1.1316199650173075, 'learning_rate/qwen_vl_interface': 1.0260458591988164e-06, 'learning_rate/action_model': 1.0260458591988165e-05, 'epoch': 1.75}) +[2026-06-16 03:28:12,626][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.05216648429632187, 'train/grad_norm_pre_clip': 20.781078338623047, 'timing/data': 0.0003341110423207283, 'timing/model': 1.1421558890142478, 'learning_rate/qwen_vl_interface': 1.0246537069125445e-06, 'learning_rate/action_model': 1.0246537069125444e-05, 'epoch': 1.75}) +[2026-06-16 03:28:17,173][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.21907854080200195, 'train/grad_norm_pre_clip': 11.675524711608887, 'timing/data': 0.00035702099557965994, 'timing/model': 1.1342717530205846, 'learning_rate/qwen_vl_interface': 1.0232632915065286e-06, 'learning_rate/action_model': 1.0232632915065285e-05, 'epoch': 1.75}) +[2026-06-16 03:28:21,729][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.0005349087296053767, 'train/grad_norm_pre_clip': 0.1828446239233017, 'timing/data': 0.00035108899464830756, 'timing/model': 1.307189943967387, 'learning_rate/qwen_vl_interface': 1.021874613552316e-06, 'learning_rate/action_model': 1.0218746135523157e-05, 'epoch': 1.75}) +[2026-06-16 03:28:26,105][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.003666744101792574, 'train/grad_norm_pre_clip': 0.6803082823753357, 'timing/data': 0.0003544409992173314, 'timing/model': 1.1338275450398214, 'learning_rate/qwen_vl_interface': 1.0204876736207384e-06, 'learning_rate/action_model': 1.0204876736207384e-05, 'epoch': 1.75}) +[2026-06-16 03:28:30,658][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.005862527992576361, 'train/grad_norm_pre_clip': 5.5849761962890625, 'timing/data': 0.0003606069949455559, 'timing/model': 1.1362093110219575, 'learning_rate/qwen_vl_interface': 1.0191024722819163e-06, 'learning_rate/action_model': 1.0191024722819162e-05, 'epoch': 1.75}) +[2026-06-16 03:28:35,192][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.012542052194476128, 'train/grad_norm_pre_clip': 7.6867265701293945, 'timing/data': 0.0003439899883233011, 'timing/model': 1.1342333580250852, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 1.75}) +[2026-06-16 03:28:39,727][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.15069244801998138, 'train/grad_norm_pre_clip': 21.67922019958496, 'timing/data': 0.0003502429462969303, 'timing/model': 1.1332162009784952, 'learning_rate/qwen_vl_interface': 1.016337287659437e-06, 'learning_rate/action_model': 1.0163372876594368e-05, 'epoch': 1.75}) +[2026-06-16 03:28:44,238][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.09349260479211807, 'train/grad_norm_pre_clip': 28.424795150756836, 'timing/data': 0.0003473880351521075, 'timing/model': 1.1377336700097658, 'learning_rate/qwen_vl_interface': 1.0149573055124427e-06, 'learning_rate/action_model': 1.0149573055124426e-05, 'epoch': 1.75}) +[2026-06-16 03:28:48,799][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.0041488781571388245, 'train/grad_norm_pre_clip': 1.6706836223602295, 'timing/data': 0.0003962499904446304, 'timing/model': 1.3042090149829164, 'learning_rate/qwen_vl_interface': 1.013579064231529e-06, 'learning_rate/action_model': 1.013579064231529e-05, 'epoch': 1.75}) +[2026-06-16 03:28:53,184][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.003911801613867283, 'train/grad_norm_pre_clip': 1.1739368438720703, 'timing/data': 0.00034293701173737645, 'timing/model': 1.131240513001103, 'learning_rate/qwen_vl_interface': 1.012202564383239e-06, 'learning_rate/action_model': 1.0122025643832388e-05, 'epoch': 1.75}) +[2026-06-16 03:28:57,746][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.1643199920654297, 'train/grad_norm_pre_clip': 18.031862258911133, 'timing/data': 0.0003708960139192641, 'timing/model': 1.13758435100317, 'learning_rate/qwen_vl_interface': 1.0108278065333978e-06, 'learning_rate/action_model': 1.0108278065333978e-05, 'epoch': 1.75}) +[2026-06-16 03:29:02,310][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.17567041516304016, 'train/grad_norm_pre_clip': 10.323385238647461, 'timing/data': 0.0003499559825286269, 'timing/model': 1.132875052979216, 'learning_rate/qwen_vl_interface': 1.0094547912471183e-06, 'learning_rate/action_model': 1.0094547912471184e-05, 'epoch': 1.75}) +[2026-06-16 03:29:06,861][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.005444834474474192, 'train/grad_norm_pre_clip': 1.9251843690872192, 'timing/data': 0.0005618000286631286, 'timing/model': 1.137498204014264, 'learning_rate/qwen_vl_interface': 1.0080835190887949e-06, 'learning_rate/action_model': 1.0080835190887947e-05, 'epoch': 1.75}) +[2026-06-16 03:29:11,425][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.0014387641567736864, 'train/grad_norm_pre_clip': 0.35740283131599426, 'timing/data': 0.0004219039692543447, 'timing/model': 1.138619514007587, 'learning_rate/qwen_vl_interface': 1.0067139906221052e-06, 'learning_rate/action_model': 1.0067139906221052e-05, 'epoch': 1.76}) +[2026-06-16 03:29:15,982][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 0.26913437247276306, 'train/grad_norm_pre_clip': 28.94944953918457, 'timing/data': 0.00034752098144963384, 'timing/model': 1.3036569600226358, 'learning_rate/qwen_vl_interface': 1.0053462064100112e-06, 'learning_rate/action_model': 1.0053462064100112e-05, 'epoch': 1.76}) +[2026-06-16 03:29:20,372][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.07710966467857361, 'train/grad_norm_pre_clip': 30.32779884338379, 'timing/data': 0.00031532597495242953, 'timing/model': 1.1395693459780887, 'learning_rate/qwen_vl_interface': 1.003980167014757e-06, 'learning_rate/action_model': 1.003980167014757e-05, 'epoch': 1.76}) +[2026-06-16 03:29:24,920][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.030165381729602814, 'train/grad_norm_pre_clip': 9.205012321472168, 'timing/data': 0.0003283059922978282, 'timing/model': 1.1342989190015942, 'learning_rate/qwen_vl_interface': 1.0026158729978702e-06, 'learning_rate/action_model': 1.0026158729978702e-05, 'epoch': 1.76}) +[2026-06-16 03:29:29,454][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.07990915328264236, 'train/grad_norm_pre_clip': 17.894580841064453, 'timing/data': 0.0003367000026628375, 'timing/model': 1.135467747983057, 'learning_rate/qwen_vl_interface': 1.0012533249201604e-06, 'learning_rate/action_model': 1.0012533249201602e-05, 'epoch': 1.76}) +[2026-06-16 03:29:33,982][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.04669773951172829, 'train/grad_norm_pre_clip': 9.86955451965332, 'timing/data': 0.0003614129964262247, 'timing/model': 1.1348452510428615, 'learning_rate/qwen_vl_interface': 9.99892523341719e-07, 'learning_rate/action_model': 9.99892523341719e-06, 'epoch': 1.76}) +[2026-06-16 03:29:38,501][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.008314833976328373, 'train/grad_norm_pre_clip': 3.8549623489379883, 'timing/data': 0.00033521605655550957, 'timing/model': 1.1373198119690642, 'learning_rate/qwen_vl_interface': 9.98533468821922e-07, 'learning_rate/action_model': 9.98533468821922e-06, 'epoch': 1.76}) +[2026-06-16 03:29:43,017][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.002432997105643153, 'train/grad_norm_pre_clip': 0.6986226439476013, 'timing/data': 0.0004195190267637372, 'timing/model': 1.270028559025377, 'learning_rate/qwen_vl_interface': 9.971761619194231e-07, 'learning_rate/action_model': 9.97176161919423e-06, 'epoch': 1.76}) +[2026-06-16 03:29:47,392][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 0.01909307949244976, 'train/grad_norm_pre_clip': 18.27151870727539, 'timing/data': 0.00039742799708619714, 'timing/model': 1.1345044720219448, 'learning_rate/qwen_vl_interface': 9.958206031921606e-07, 'learning_rate/action_model': 9.958206031921604e-06, 'epoch': 1.76}) +[2026-06-16 03:29:51,912][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.008818000555038452, 'train/grad_norm_pre_clip': 1.953275203704834, 'timing/data': 0.0003777379752136767, 'timing/model': 1.1304692080011591, 'learning_rate/qwen_vl_interface': 9.944667931973537e-07, 'learning_rate/action_model': 9.944667931973536e-06, 'epoch': 1.76}) +[2026-06-16 03:29:56,440][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.004226624500006437, 'train/grad_norm_pre_clip': 2.624136447906494, 'timing/data': 0.0004959410289302468, 'timing/model': 1.1307565830065869, 'learning_rate/qwen_vl_interface': 9.93114732491503e-07, 'learning_rate/action_model': 9.93114732491503e-06, 'epoch': 1.76}) +[2026-06-16 03:30:00,992][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.014002406038343906, 'train/grad_norm_pre_clip': 3.321134090423584, 'timing/data': 0.00031951296841725707, 'timing/model': 1.1349496410111897, 'learning_rate/qwen_vl_interface': 9.91764421630389e-07, 'learning_rate/action_model': 9.91764421630389e-06, 'epoch': 1.76}) +[2026-06-16 03:30:05,537][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.01486851554363966, 'train/grad_norm_pre_clip': 2.903430700302124, 'timing/data': 0.000364390027243644, 'timing/model': 1.1307507259771228, 'learning_rate/qwen_vl_interface': 9.904158611690732e-07, 'learning_rate/action_model': 9.90415861169073e-06, 'epoch': 1.76}) +[2026-06-16 03:30:10,089][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.1264929473400116, 'train/grad_norm_pre_clip': 13.102907180786133, 'timing/data': 0.00036399904638528824, 'timing/model': 1.3048984250053763, 'learning_rate/qwen_vl_interface': 9.890690516618987e-07, 'learning_rate/action_model': 9.890690516618985e-06, 'epoch': 1.76}) +[2026-06-16 03:30:14,489][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.008886984549462795, 'train/grad_norm_pre_clip': 4.161419868469238, 'timing/data': 0.0003678019857034087, 'timing/model': 1.1396685070358217, 'learning_rate/qwen_vl_interface': 9.877239936624875e-07, 'learning_rate/action_model': 9.877239936624875e-06, 'epoch': 1.76}) +[2026-06-16 03:30:19,034][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.07459776848554611, 'train/grad_norm_pre_clip': 14.506704330444336, 'timing/data': 0.00042413000483065844, 'timing/model': 1.1344702349742875, 'learning_rate/qwen_vl_interface': 9.863806877237421e-07, 'learning_rate/action_model': 9.863806877237422e-06, 'epoch': 1.76}) +[2026-06-16 03:30:23,595][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.001607511192560196, 'train/grad_norm_pre_clip': 0.44966909289360046, 'timing/data': 0.00034675601636990905, 'timing/model': 1.1427527079940774, 'learning_rate/qwen_vl_interface': 9.850391343978454e-07, 'learning_rate/action_model': 9.850391343978454e-06, 'epoch': 1.76}) +[2026-06-16 03:30:28,143][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.0044919163919985294, 'train/grad_norm_pre_clip': 0.7581849694252014, 'timing/data': 0.00041158002568408847, 'timing/model': 1.135908466996625, 'learning_rate/qwen_vl_interface': 9.836993342362596e-07, 'learning_rate/action_model': 9.836993342362596e-06, 'epoch': 1.76}) +[2026-06-16 03:30:32,654][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.0008158386335708201, 'train/grad_norm_pre_clip': 0.17545771598815918, 'timing/data': 0.0003523459890857339, 'timing/model': 1.1312294799718075, 'learning_rate/qwen_vl_interface': 9.823612877897246e-07, 'learning_rate/action_model': 9.823612877897245e-06, 'epoch': 1.76}) +[2026-06-16 03:30:37,223][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.04942501708865166, 'train/grad_norm_pre_clip': 15.692220687866211, 'timing/data': 0.0003420680295675993, 'timing/model': 1.3185765050002374, 'learning_rate/qwen_vl_interface': 9.81024995608261e-07, 'learning_rate/action_model': 9.810249956082609e-06, 'epoch': 1.76}) +[2026-06-16 03:30:41,619][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.2897866368293762, 'train/grad_norm_pre_clip': 29.070802688598633, 'timing/data': 0.00034437800059095025, 'timing/model': 1.139939011016395, 'learning_rate/qwen_vl_interface': 9.796904582411685e-07, 'learning_rate/action_model': 9.796904582411684e-06, 'epoch': 1.76}) +[2026-06-16 03:30:46,167][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.0005375095643103123, 'train/grad_norm_pre_clip': 0.07843077182769775, 'timing/data': 0.00038936600321903825, 'timing/model': 1.1345788959879428, 'learning_rate/qwen_vl_interface': 9.78357676237025e-07, 'learning_rate/action_model': 9.78357676237025e-06, 'epoch': 1.76}) +[2026-06-16 03:30:50,707][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 0.009233069606125355, 'train/grad_norm_pre_clip': 2.339540958404541, 'timing/data': 0.0003544309874996543, 'timing/model': 1.1391217810451053, 'learning_rate/qwen_vl_interface': 9.77026650143687e-07, 'learning_rate/action_model': 9.77026650143687e-06, 'epoch': 1.76}) +[2026-06-16 03:30:55,246][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.04729069396853447, 'train/grad_norm_pre_clip': 20.465980529785156, 'timing/data': 0.0003472540411166847, 'timing/model': 1.1317580959876068, 'learning_rate/qwen_vl_interface': 9.756973805082882e-07, 'learning_rate/action_model': 9.756973805082883e-06, 'epoch': 1.76}) +[2026-06-16 03:30:59,783][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.0019787861965596676, 'train/grad_norm_pre_clip': 0.2774527668952942, 'timing/data': 0.00033158704172819853, 'timing/model': 1.1317024150048383, 'learning_rate/qwen_vl_interface': 9.743698678772434e-07, 'learning_rate/action_model': 9.743698678772433e-06, 'epoch': 1.77}) +[2026-06-16 03:31:04,341][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.17013980448246002, 'train/grad_norm_pre_clip': 24.60247230529785, 'timing/data': 0.0004030340351164341, 'timing/model': 1.3002480730065145, 'learning_rate/qwen_vl_interface': 9.730441127962395e-07, 'learning_rate/action_model': 9.730441127962395e-06, 'epoch': 1.77}) +[2026-06-16 03:31:08,718][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.019277049228549004, 'train/grad_norm_pre_clip': 7.814090251922607, 'timing/data': 0.00031624099938198924, 'timing/model': 1.1328966990113258, 'learning_rate/qwen_vl_interface': 9.717201158102465e-07, 'learning_rate/action_model': 9.717201158102466e-06, 'epoch': 1.77}) +[2026-06-16 03:35:32,300][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.0004563335969578475, 'train/grad_norm_pre_clip': 0.11608676612377167, 'eval/loss': 0.0386514738202095, 'eval/flap_precision': 0.9754224270353302, 'eval/flap_recall': 0.9799382716049383, 'eval/flap_f1': 0.9776751347190146, 'eval/flap_rate_model': 0.18083333333333335, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9776751347190146, 'timing/data': 0.0003883899771608412, 'timing/model': 1.1362031860044226, 'learning_rate/qwen_vl_interface': 9.703978774635094e-07, 'learning_rate/action_model': 9.703978774635095e-06, 'epoch': 1.77}) +[2026-06-16 03:35:32,301][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4300_state +[2026-06-16 03:35:32,303][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 03:35:35,876][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4300_state/model.safetensors +[2026-06-16 03:35:45,862][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4300_state/optimizer.bin +[2026-06-16 03:35:45,863][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4300_state/sampler.bin +[2026-06-16 03:35:45,863][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4300_state/sampler_1.bin +[2026-06-16 03:35:45,863][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4300_state/random_states_0.pkl +[2026-06-16 03:35:45,910][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4300 +[2026-06-16 03:35:45,910][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 03:35:45,916][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 03:35:50,367][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.25833794474601746, 'train/grad_norm_pre_clip': 33.111568450927734, 'timing/data': 0.00044135504867881536, 'timing/model': 1.1151415669592097, 'learning_rate/qwen_vl_interface': 9.6907739829955e-07, 'learning_rate/action_model': 9.6907739829955e-06, 'epoch': 1.77}) +[2026-06-16 03:35:54,748][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.20632950961589813, 'train/grad_norm_pre_clip': 27.251209259033203, 'timing/data': 0.0003406759933568537, 'timing/model': 1.1141926719574258, 'learning_rate/qwen_vl_interface': 9.677586788611681e-07, 'learning_rate/action_model': 9.677586788611681e-06, 'epoch': 1.77}) +[2026-06-16 03:35:59,176][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.09688672423362732, 'train/grad_norm_pre_clip': 31.53447151184082, 'timing/data': 0.000352351984474808, 'timing/model': 1.11697638197802, 'learning_rate/qwen_vl_interface': 9.664417196904382e-07, 'learning_rate/action_model': 9.664417196904382e-06, 'epoch': 1.77}) +[2026-06-16 03:36:03,657][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 0.005673639010637999, 'train/grad_norm_pre_clip': 2.817117214202881, 'timing/data': 0.0003992759739048779, 'timing/model': 1.1161460650037043, 'learning_rate/qwen_vl_interface': 9.651265213287134e-07, 'learning_rate/action_model': 9.651265213287134e-06, 'epoch': 1.77}) +[2026-06-16 03:36:08,142][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.19573158025741577, 'train/grad_norm_pre_clip': 22.287586212158203, 'timing/data': 0.0003596440074034035, 'timing/model': 1.2810271250200458, 'learning_rate/qwen_vl_interface': 9.638130843166218e-07, 'learning_rate/action_model': 9.638130843166219e-06, 'epoch': 1.77}) +[2026-06-16 03:36:12,467][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.003873014124110341, 'train/grad_norm_pre_clip': 1.163014531135559, 'timing/data': 0.00036857102531939745, 'timing/model': 1.1198035139823332, 'learning_rate/qwen_vl_interface': 9.62501409194068e-07, 'learning_rate/action_model': 9.625014091940678e-06, 'epoch': 1.77}) +[2026-06-16 03:36:16,956][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.018104614689946175, 'train/grad_norm_pre_clip': 27.971017837524414, 'timing/data': 0.0003354339860379696, 'timing/model': 1.11878943501506, 'learning_rate/qwen_vl_interface': 9.611914965002317e-07, 'learning_rate/action_model': 9.611914965002316e-06, 'epoch': 1.77}) +[2026-06-16 03:36:21,410][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.00032578466925770044, 'train/grad_norm_pre_clip': 0.07767610251903534, 'timing/data': 0.0003447109484113753, 'timing/model': 1.1243320570210926, 'learning_rate/qwen_vl_interface': 9.598833467735686e-07, 'learning_rate/action_model': 9.598833467735685e-06, 'epoch': 1.77}) +[2026-06-16 03:36:25,906][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.0037037667352706194, 'train/grad_norm_pre_clip': 1.2630581855773926, 'timing/data': 0.00036464404547587037, 'timing/model': 1.1228342730319127, 'learning_rate/qwen_vl_interface': 9.585769605518097e-07, 'learning_rate/action_model': 9.585769605518097e-06, 'epoch': 1.77}) +[2026-06-16 03:36:30,424][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.0017952207708731294, 'train/grad_norm_pre_clip': 0.4994617998600006, 'timing/data': 0.00036733102751895785, 'timing/model': 1.1252462100237608, 'learning_rate/qwen_vl_interface': 9.572723383719611e-07, 'learning_rate/action_model': 9.572723383719612e-06, 'epoch': 1.77}) +[2026-06-16 03:36:34,930][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.10323970764875412, 'train/grad_norm_pre_clip': 39.280975341796875, 'timing/data': 0.00035628798650577664, 'timing/model': 1.2938510020030662, 'learning_rate/qwen_vl_interface': 9.559694807703041e-07, 'learning_rate/action_model': 9.55969480770304e-06, 'epoch': 1.77}) +[2026-06-16 03:36:39,278][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.001595560577698052, 'train/grad_norm_pre_clip': 0.32245758175849915, 'timing/data': 0.0006993219722062349, 'timing/model': 1.1259346429724246, 'learning_rate/qwen_vl_interface': 9.546683882823945e-07, 'learning_rate/action_model': 9.546683882823945e-06, 'epoch': 1.77}) +[2026-06-16 03:36:43,797][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.0002565559698268771, 'train/grad_norm_pre_clip': 0.036099836230278015, 'timing/data': 0.0004025720409117639, 'timing/model': 1.1261171119986102, 'learning_rate/qwen_vl_interface': 9.533690614430608e-07, 'learning_rate/action_model': 9.533690614430608e-06, 'epoch': 1.77}) +[2026-06-16 03:36:48,321][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.018979065120220184, 'train/grad_norm_pre_clip': 9.201169967651367, 'timing/data': 0.00041257700650021434, 'timing/model': 1.1360099029843695, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 1.77}) +[2026-06-16 03:36:52,793][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.03015635348856449, 'train/grad_norm_pre_clip': 22.977100372314453, 'timing/data': 0.00037311995401978493, 'timing/model': 1.1277674720040523, 'learning_rate/qwen_vl_interface': 9.50775706845814e-07, 'learning_rate/action_model': 9.50775706845814e-06, 'epoch': 1.77}) +[2026-06-16 03:36:57,307][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.0027814817149192095, 'train/grad_norm_pre_clip': 1.0605237483978271, 'timing/data': 0.00033541902666911483, 'timing/model': 1.13182237703586, 'learning_rate/qwen_vl_interface': 9.49481680153931e-07, 'learning_rate/action_model': 9.494816801539309e-06, 'epoch': 1.77}) +[2026-06-16 03:37:01,842][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.039809975773096085, 'train/grad_norm_pre_clip': 11.867897033691406, 'timing/data': 0.0003536699805408716, 'timing/model': 1.2996791360201314, 'learning_rate/qwen_vl_interface': 9.481894212426848e-07, 'learning_rate/action_model': 9.481894212426847e-06, 'epoch': 1.77}) +[2026-06-16 03:37:06,211][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.02989359200000763, 'train/grad_norm_pre_clip': 11.035130500793457, 'timing/data': 0.0003363900468684733, 'timing/model': 1.1323519359575585, 'learning_rate/qwen_vl_interface': 9.468989306432738e-07, 'learning_rate/action_model': 9.468989306432738e-06, 'epoch': 1.77}) +[2026-06-16 03:37:10,792][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.00876796804368496, 'train/grad_norm_pre_clip': 2.8221144676208496, 'timing/data': 0.0003342389827594161, 'timing/model': 1.1273600459680893, 'learning_rate/qwen_vl_interface': 9.456102088861702e-07, 'learning_rate/action_model': 9.456102088861701e-06, 'epoch': 1.77}) +[2026-06-16 03:37:15,336][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.08880339562892914, 'train/grad_norm_pre_clip': 11.218216896057129, 'timing/data': 0.00036292203003540635, 'timing/model': 1.1307255399879068, 'learning_rate/qwen_vl_interface': 9.443232565011188e-07, 'learning_rate/action_model': 9.443232565011187e-06, 'epoch': 1.77}) +[2026-06-16 03:37:19,861][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.0007762087625451386, 'train/grad_norm_pre_clip': 0.1385190486907959, 'timing/data': 0.0003441500011831522, 'timing/model': 1.134772911958862, 'learning_rate/qwen_vl_interface': 9.430380740171369e-07, 'learning_rate/action_model': 9.430380740171368e-06, 'epoch': 1.78}) +[2026-06-16 03:37:24,400][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.21280497312545776, 'train/grad_norm_pre_clip': 7.742455005645752, 'timing/data': 0.000368754961527884, 'timing/model': 1.2885113169904798, 'learning_rate/qwen_vl_interface': 9.417546619625143e-07, 'learning_rate/action_model': 9.417546619625143e-06, 'epoch': 1.78}) +[2026-06-16 03:37:28,892][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.0010654248762875795, 'train/grad_norm_pre_clip': 0.2095244973897934, 'timing/data': 0.0003593949950300157, 'timing/model': 1.24598937202245, 'learning_rate/qwen_vl_interface': 9.404730208648144e-07, 'learning_rate/action_model': 9.404730208648144e-06, 'epoch': 1.78}) +[2026-06-16 03:37:33,272][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.02761347033083439, 'train/grad_norm_pre_clip': 14.586603164672852, 'timing/data': 0.0003610079875215888, 'timing/model': 1.1329829139867797, 'learning_rate/qwen_vl_interface': 9.3919315125087e-07, 'learning_rate/action_model': 9.3919315125087e-06, 'epoch': 1.78}) +[2026-06-16 03:37:37,817][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.01050239335745573, 'train/grad_norm_pre_clip': 9.287818908691406, 'timing/data': 0.0003561970079317689, 'timing/model': 1.1373636919888668, 'learning_rate/qwen_vl_interface': 9.379150536467879e-07, 'learning_rate/action_model': 9.379150536467878e-06, 'epoch': 1.78}) +[2026-06-16 03:37:42,362][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.0038841499481350183, 'train/grad_norm_pre_clip': 0.8251146078109741, 'timing/data': 0.000394688977394253, 'timing/model': 1.1385904740309343, 'learning_rate/qwen_vl_interface': 9.366387285779453e-07, 'learning_rate/action_model': 9.366387285779453e-06, 'epoch': 1.78}) +[2026-06-16 03:37:46,895][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.0076105049811303616, 'train/grad_norm_pre_clip': 6.9608306884765625, 'timing/data': 0.0003337380476295948, 'timing/model': 1.133362088992726, 'learning_rate/qwen_vl_interface': 9.353641765689915e-07, 'learning_rate/action_model': 9.353641765689915e-06, 'epoch': 1.78}) +[2026-06-16 03:37:51,350][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.0009113193373195827, 'train/grad_norm_pre_clip': 0.2675730586051941, 'timing/data': 0.0003750280011445284, 'timing/model': 1.1367401839816011, 'learning_rate/qwen_vl_interface': 9.340913981438462e-07, 'learning_rate/action_model': 9.340913981438463e-06, 'epoch': 1.78}) +[2026-06-16 03:37:55,796][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.24220483005046844, 'train/grad_norm_pre_clip': 16.28083038330078, 'timing/data': 0.00036090402863919735, 'timing/model': 1.2063392559648491, 'learning_rate/qwen_vl_interface': 9.328203938257019e-07, 'learning_rate/action_model': 9.328203938257018e-06, 'epoch': 1.78}) +[2026-06-16 03:38:00,184][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.263822466135025, 'train/grad_norm_pre_clip': 12.414752006530762, 'timing/data': 0.0003600099589675665, 'timing/model': 1.1325483069522306, 'learning_rate/qwen_vl_interface': 9.315511641370191e-07, 'learning_rate/action_model': 9.315511641370191e-06, 'epoch': 1.78}) +[2026-06-16 03:38:04,706][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.000712464505340904, 'train/grad_norm_pre_clip': 0.12227324396371841, 'timing/data': 0.0007061060168780386, 'timing/model': 1.1416460840264335, 'learning_rate/qwen_vl_interface': 9.302837095995306e-07, 'learning_rate/action_model': 9.302837095995306e-06, 'epoch': 1.78}) +[2026-06-16 03:38:09,274][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.1274544894695282, 'train/grad_norm_pre_clip': 9.06029224395752, 'timing/data': 0.00036262901267036796, 'timing/model': 1.137834577995818, 'learning_rate/qwen_vl_interface': 9.290180307342392e-07, 'learning_rate/action_model': 9.290180307342391e-06, 'epoch': 1.78}) +[2026-06-16 03:38:13,822][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 0.21665552258491516, 'train/grad_norm_pre_clip': 11.59410572052002, 'timing/data': 0.0003537230077199638, 'timing/model': 1.1373055489966646, 'learning_rate/qwen_vl_interface': 9.277541280614183e-07, 'learning_rate/action_model': 9.277541280614183e-06, 'epoch': 1.78}) +[2026-06-16 03:38:18,362][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.10022435337305069, 'train/grad_norm_pre_clip': 10.944842338562012, 'timing/data': 0.00035075104096904397, 'timing/model': 1.2902953199809417, 'learning_rate/qwen_vl_interface': 9.264920021006093e-07, 'learning_rate/action_model': 9.264920021006093e-06, 'epoch': 1.78}) +[2026-06-16 03:38:22,762][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.00014701299369335175, 'train/grad_norm_pre_clip': 0.01748151332139969, 'timing/data': 0.000326313020195812, 'timing/model': 1.1319722150219604, 'learning_rate/qwen_vl_interface': 9.252316533706246e-07, 'learning_rate/action_model': 9.252316533706246e-06, 'epoch': 1.78}) +[2026-06-16 03:38:27,306][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.02866310626268387, 'train/grad_norm_pre_clip': 13.807644844055176, 'timing/data': 0.00039366999408230186, 'timing/model': 1.1356792739825323, 'learning_rate/qwen_vl_interface': 9.239730823895469e-07, 'learning_rate/action_model': 9.23973082389547e-06, 'epoch': 1.78}) +[2026-06-16 03:38:31,766][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.00747063010931015, 'train/grad_norm_pre_clip': 1.7625648975372314, 'timing/data': 0.000402536999899894, 'timing/model': 1.1357547350344248, 'learning_rate/qwen_vl_interface': 9.227162896747264e-07, 'learning_rate/action_model': 9.227162896747264e-06, 'epoch': 1.78}) +[2026-06-16 03:38:36,257][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.013616062700748444, 'train/grad_norm_pre_clip': 17.641582489013672, 'timing/data': 0.00034251698525622487, 'timing/model': 1.1325413410086185, 'learning_rate/qwen_vl_interface': 9.214612757427833e-07, 'learning_rate/action_model': 9.214612757427833e-06, 'epoch': 1.78}) +[2026-06-16 03:38:40,799][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.10509979724884033, 'train/grad_norm_pre_clip': 18.01543617248535, 'timing/data': 0.0005996430409140885, 'timing/model': 1.146463979035616, 'learning_rate/qwen_vl_interface': 9.202080411096071e-07, 'learning_rate/action_model': 9.20208041109607e-06, 'epoch': 1.78}) +[2026-06-16 03:38:45,341][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.0022255051881074905, 'train/grad_norm_pre_clip': 0.6357609033584595, 'timing/data': 0.00035245600156486034, 'timing/model': 1.2901559500023723, 'learning_rate/qwen_vl_interface': 9.189565862903543e-07, 'learning_rate/action_model': 9.189565862903543e-06, 'epoch': 1.78}) +[2026-06-16 03:38:49,750][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.030957598239183426, 'train/grad_norm_pre_clip': 9.565045356750488, 'timing/data': 0.0003520820173434913, 'timing/model': 1.1410239610122517, 'learning_rate/qwen_vl_interface': 9.17706911799451e-07, 'learning_rate/action_model': 9.17706911799451e-06, 'epoch': 1.78}) +[2026-06-16 03:38:54,287][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.03729545697569847, 'train/grad_norm_pre_clip': 8.721342086791992, 'timing/data': 0.0003519509918987751, 'timing/model': 1.1338374760234728, 'learning_rate/qwen_vl_interface': 9.164590181505912e-07, 'learning_rate/action_model': 9.164590181505911e-06, 'epoch': 1.78}) +[2026-06-16 03:38:58,816][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.004387767985463142, 'train/grad_norm_pre_clip': 2.093719959259033, 'timing/data': 0.0003811590140685439, 'timing/model': 1.1305412590154447, 'learning_rate/qwen_vl_interface': 9.15212905856738e-07, 'learning_rate/action_model': 9.15212905856738e-06, 'epoch': 1.78}) +[2026-06-16 03:39:03,343][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.09665565937757492, 'train/grad_norm_pre_clip': 6.731908798217773, 'timing/data': 0.00036450999323278666, 'timing/model': 1.138055583985988, 'learning_rate/qwen_vl_interface': 9.139685754301189e-07, 'learning_rate/action_model': 9.139685754301189e-06, 'epoch': 1.78}) +[2026-06-16 03:39:07,823][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.35125723481178284, 'train/grad_norm_pre_clip': 20.840003967285156, 'timing/data': 0.00037584599340334535, 'timing/model': 1.1422443359624594, 'learning_rate/qwen_vl_interface': 9.127260273822324e-07, 'learning_rate/action_model': 9.127260273822325e-06, 'epoch': 1.78}) +[2026-06-16 03:39:12,305][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 0.03722749277949333, 'train/grad_norm_pre_clip': 12.158486366271973, 'timing/data': 0.0003425460308790207, 'timing/model': 1.2050751619972289, 'learning_rate/qwen_vl_interface': 9.114852622238435e-07, 'learning_rate/action_model': 9.114852622238435e-06, 'epoch': 1.79}) +[2026-06-16 03:39:16,714][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.01294488925486803, 'train/grad_norm_pre_clip': 8.260839462280273, 'timing/data': 0.0003168300027027726, 'timing/model': 1.143476793949958, 'learning_rate/qwen_vl_interface': 9.102462804649827e-07, 'learning_rate/action_model': 9.102462804649827e-06, 'epoch': 1.79}) +[2026-06-16 03:39:21,156][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.16140876710414886, 'train/grad_norm_pre_clip': 10.846177101135254, 'timing/data': 0.00035482202656567097, 'timing/model': 1.134369029023219, 'learning_rate/qwen_vl_interface': 9.090090826149489e-07, 'learning_rate/action_model': 9.09009082614949e-06, 'epoch': 1.79}) +[2026-06-16 03:39:25,653][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.004497589077800512, 'train/grad_norm_pre_clip': 1.180037498474121, 'timing/data': 0.0003517530276440084, 'timing/model': 1.1356749800033867, 'learning_rate/qwen_vl_interface': 9.077736691823081e-07, 'learning_rate/action_model': 9.07773669182308e-06, 'epoch': 1.79}) +[2026-06-16 03:39:30,161][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.013434348627924919, 'train/grad_norm_pre_clip': 9.71141242980957, 'timing/data': 0.00045577797573059797, 'timing/model': 1.1390498509863392, 'learning_rate/qwen_vl_interface': 9.065400406748912e-07, 'learning_rate/action_model': 9.06540040674891e-06, 'epoch': 1.79}) +[2026-06-16 03:39:34,697][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.13566149771213531, 'train/grad_norm_pre_clip': 24.916950225830078, 'timing/data': 0.0003272769972681999, 'timing/model': 1.286498032975942, 'learning_rate/qwen_vl_interface': 9.053081975997969e-07, 'learning_rate/action_model': 9.053081975997968e-06, 'epoch': 1.79}) +[2026-06-16 03:39:39,088][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.001631357823498547, 'train/grad_norm_pre_clip': 0.2875625491142273, 'timing/data': 0.00036238995380699635, 'timing/model': 1.1368194539682008, 'learning_rate/qwen_vl_interface': 9.040781404633894e-07, 'learning_rate/action_model': 9.040781404633895e-06, 'epoch': 1.79}) +[2026-06-16 03:39:43,632][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 0.08127506077289581, 'train/grad_norm_pre_clip': 16.434980392456055, 'timing/data': 0.00038028700510039926, 'timing/model': 1.1442667759838514, 'learning_rate/qwen_vl_interface': 9.028498697712982e-07, 'learning_rate/action_model': 9.028498697712981e-06, 'epoch': 1.79}) +[2026-06-16 03:39:48,169][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.27867263555526733, 'train/grad_norm_pre_clip': 12.034503936767578, 'timing/data': 0.0003418389824219048, 'timing/model': 1.136990389029961, 'learning_rate/qwen_vl_interface': 9.016233860284195e-07, 'learning_rate/action_model': 9.016233860284195e-06, 'epoch': 1.79}) +[2026-06-16 03:39:52,702][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.1270386129617691, 'train/grad_norm_pre_clip': 17.405580520629883, 'timing/data': 0.0003754360368475318, 'timing/model': 1.1338016590452753, 'learning_rate/qwen_vl_interface': 9.003986897389142e-07, 'learning_rate/action_model': 9.003986897389141e-06, 'epoch': 1.79}) +[2026-06-16 03:39:57,259][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.006081502418965101, 'train/grad_norm_pre_clip': 1.4021328687667847, 'timing/data': 0.0005828400026075542, 'timing/model': 1.137629576958716, 'learning_rate/qwen_vl_interface': 8.991757814062094e-07, 'learning_rate/action_model': 8.991757814062093e-06, 'epoch': 1.79}) +[2026-06-16 03:40:01,829][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.0038066068664193153, 'train/grad_norm_pre_clip': 1.5854088068008423, 'timing/data': 0.0003564649960026145, 'timing/model': 1.3248419449664652, 'learning_rate/qwen_vl_interface': 8.979546615329954e-07, 'learning_rate/action_model': 8.979546615329953e-06, 'epoch': 1.79}) +[2026-06-16 03:40:06,215][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.322207510471344, 'train/grad_norm_pre_clip': 51.105079650878906, 'timing/data': 0.00036053499206900597, 'timing/model': 1.1347988470224664, 'learning_rate/qwen_vl_interface': 8.967353306212286e-07, 'learning_rate/action_model': 8.967353306212286e-06, 'epoch': 1.79}) +[2026-06-16 03:40:10,747][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.0007591155590489507, 'train/grad_norm_pre_clip': 0.23731310665607452, 'timing/data': 0.0004831849946640432, 'timing/model': 1.1333796399994753, 'learning_rate/qwen_vl_interface': 8.955177891721303e-07, 'learning_rate/action_model': 8.955177891721303e-06, 'epoch': 1.79}) +[2026-06-16 03:40:15,290][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.11194025725126266, 'train/grad_norm_pre_clip': 9.73650074005127, 'timing/data': 0.0004128760192543268, 'timing/model': 1.12804529001005, 'learning_rate/qwen_vl_interface': 8.943020376861858e-07, 'learning_rate/action_model': 8.943020376861857e-06, 'epoch': 1.79}) +[2026-06-16 03:40:19,822][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.006691598799079657, 'train/grad_norm_pre_clip': 3.8777008056640625, 'timing/data': 0.0003693340113386512, 'timing/model': 1.1331302829785272, 'learning_rate/qwen_vl_interface': 8.930880766631444e-07, 'learning_rate/action_model': 8.930880766631445e-06, 'epoch': 1.79}) +[2026-06-16 03:40:24,319][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.026042655110359192, 'train/grad_norm_pre_clip': 12.806717872619629, 'timing/data': 0.00037181697553023696, 'timing/model': 1.1331289049703628, 'learning_rate/qwen_vl_interface': 8.918759066020201e-07, 'learning_rate/action_model': 8.9187590660202e-06, 'epoch': 1.79}) +[2026-06-16 03:40:28,886][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 0.0008640022715553641, 'train/grad_norm_pre_clip': 0.14290498197078705, 'timing/data': 0.00035305804340168834, 'timing/model': 1.2958744670031592, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 1.79}) +[2026-06-16 03:40:33,302][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.002707325853407383, 'train/grad_norm_pre_clip': 0.73984694480896, 'timing/data': 0.00038201402639970183, 'timing/model': 1.1366091199452057, 'learning_rate/qwen_vl_interface': 8.894569413578938e-07, 'learning_rate/action_model': 8.894569413578937e-06, 'epoch': 1.79}) +[2026-06-16 03:40:37,851][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.01089676283299923, 'train/grad_norm_pre_clip': 5.59163236618042, 'timing/data': 0.0003398910048417747, 'timing/model': 1.1367985630058683, 'learning_rate/qwen_vl_interface': 8.882501471692379e-07, 'learning_rate/action_model': 8.882501471692378e-06, 'epoch': 1.79}) +[2026-06-16 03:40:42,425][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.11028499901294708, 'train/grad_norm_pre_clip': 6.530758857727051, 'timing/data': 0.0003434710088185966, 'timing/model': 1.137871588987764, 'learning_rate/qwen_vl_interface': 8.870451459311894e-07, 'learning_rate/action_model': 8.870451459311893e-06, 'epoch': 1.79}) +[2026-06-16 03:40:46,963][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.11790085583925247, 'train/grad_norm_pre_clip': 11.98106575012207, 'timing/data': 0.000803205999545753, 'timing/model': 1.140774043975398, 'learning_rate/qwen_vl_interface': 8.85841938139077e-07, 'learning_rate/action_model': 8.85841938139077e-06, 'epoch': 1.79}) +[2026-06-16 03:40:51,520][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.00195358251221478, 'train/grad_norm_pre_clip': 0.4302397072315216, 'timing/data': 0.0003435579710640013, 'timing/model': 1.1362774840090424, 'learning_rate/qwen_vl_interface': 8.846405242874955e-07, 'learning_rate/action_model': 8.846405242874954e-06, 'epoch': 1.79}) +[2026-06-16 03:40:56,083][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.0026308316737413406, 'train/grad_norm_pre_clip': 0.3442431688308716, 'timing/data': 0.00034176604822278023, 'timing/model': 1.3125909059890546, 'learning_rate/qwen_vl_interface': 8.834409048703004e-07, 'learning_rate/action_model': 8.834409048703005e-06, 'epoch': 1.79}) +[2026-06-16 03:41:00,462][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.011596545577049255, 'train/grad_norm_pre_clip': 6.0838518142700195, 'timing/data': 0.0003409649943932891, 'timing/model': 1.1365549490437843, 'learning_rate/qwen_vl_interface': 8.822430803806098e-07, 'learning_rate/action_model': 8.822430803806097e-06, 'epoch': 1.8}) +[2026-06-16 03:41:05,024][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.0017203025054186583, 'train/grad_norm_pre_clip': 0.39379337430000305, 'timing/data': 0.0004086489789187908, 'timing/model': 1.1388726620352827, 'learning_rate/qwen_vl_interface': 8.810470513108041e-07, 'learning_rate/action_model': 8.81047051310804e-06, 'epoch': 1.8}) +[2026-06-16 03:41:09,562][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 0.16226449608802795, 'train/grad_norm_pre_clip': 11.745185852050781, 'timing/data': 0.00034566898830235004, 'timing/model': 1.1331300350138918, 'learning_rate/qwen_vl_interface': 8.798528181525262e-07, 'learning_rate/action_model': 8.798528181525261e-06, 'epoch': 1.8}) +[2026-06-16 03:41:14,099][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.11037339270114899, 'train/grad_norm_pre_clip': 6.836073875427246, 'timing/data': 0.00033044704468920827, 'timing/model': 1.1335996859706938, 'learning_rate/qwen_vl_interface': 8.786603813966798e-07, 'learning_rate/action_model': 8.786603813966798e-06, 'epoch': 1.8}) +[2026-06-16 03:41:18,608][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 0.010691373609006405, 'train/grad_norm_pre_clip': 8.201916694641113, 'timing/data': 0.00035621802089735866, 'timing/model': 1.1334403419750743, 'learning_rate/qwen_vl_interface': 8.774697415334304e-07, 'learning_rate/action_model': 8.774697415334304e-06, 'epoch': 1.8}) +[2026-06-16 03:41:23,169][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.00024494953686371446, 'train/grad_norm_pre_clip': 0.04865920543670654, 'timing/data': 0.0005625949706882238, 'timing/model': 1.3012991689611226, 'learning_rate/qwen_vl_interface': 8.762808990522056e-07, 'learning_rate/action_model': 8.762808990522057e-06, 'epoch': 1.8}) +[2026-06-16 03:41:27,582][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.034800875931978226, 'train/grad_norm_pre_clip': 11.18828296661377, 'timing/data': 0.0003399169654585421, 'timing/model': 1.142337670957204, 'learning_rate/qwen_vl_interface': 8.750938544416934e-07, 'learning_rate/action_model': 8.750938544416934e-06, 'epoch': 1.8}) +[2026-06-16 03:41:32,143][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.000227659271331504, 'train/grad_norm_pre_clip': 0.028720492497086525, 'timing/data': 0.000312295975163579, 'timing/model': 1.144057678990066, 'learning_rate/qwen_vl_interface': 8.739086081898437e-07, 'learning_rate/action_model': 8.739086081898437e-06, 'epoch': 1.8}) +[2026-06-16 03:41:36,693][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 0.022848887369036674, 'train/grad_norm_pre_clip': 12.8689546585083, 'timing/data': 0.00039588194340467453, 'timing/model': 1.1361294010421261, 'learning_rate/qwen_vl_interface': 8.727251607838647e-07, 'learning_rate/action_model': 8.727251607838647e-06, 'epoch': 1.8}) +[2026-06-16 03:41:41,223][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.027849765494465828, 'train/grad_norm_pre_clip': 11.173876762390137, 'timing/data': 0.0003578030155040324, 'timing/model': 1.1379567150142975, 'learning_rate/qwen_vl_interface': 8.715435127102281e-07, 'learning_rate/action_model': 8.715435127102281e-06, 'epoch': 1.8}) +[2026-06-16 03:41:45,798][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.0011438807705417275, 'train/grad_norm_pre_clip': 0.3210655748844147, 'timing/data': 0.0003746389993466437, 'timing/model': 1.1466598649858497, 'learning_rate/qwen_vl_interface': 8.703636644546651e-07, 'learning_rate/action_model': 8.703636644546651e-06, 'epoch': 1.8}) +[2026-06-16 03:41:50,342][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.02112184464931488, 'train/grad_norm_pre_clip': 15.17384147644043, 'timing/data': 0.0003459930303506553, 'timing/model': 1.2965024940203875, 'learning_rate/qwen_vl_interface': 8.69185616502166e-07, 'learning_rate/action_model': 8.69185616502166e-06, 'epoch': 1.8}) +[2026-06-16 03:41:54,736][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.020484471693634987, 'train/grad_norm_pre_clip': 11.644234657287598, 'timing/data': 0.00035816896706819534, 'timing/model': 1.1355218579992652, 'learning_rate/qwen_vl_interface': 8.680093693369821e-07, 'learning_rate/action_model': 8.68009369336982e-06, 'epoch': 1.8}) +[2026-06-16 03:41:59,290][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.01328673493117094, 'train/grad_norm_pre_clip': 13.025531768798828, 'timing/data': 0.00036052102223038673, 'timing/model': 1.1346286800107919, 'learning_rate/qwen_vl_interface': 8.66834923442624e-07, 'learning_rate/action_model': 8.66834923442624e-06, 'epoch': 1.8}) +[2026-06-16 03:42:03,842][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.0007873639115132391, 'train/grad_norm_pre_clip': 0.12668243050575256, 'timing/data': 0.0003472800017334521, 'timing/model': 1.1388566950336099, 'learning_rate/qwen_vl_interface': 8.65662279301863e-07, 'learning_rate/action_model': 8.656622793018628e-06, 'epoch': 1.8}) +[2026-06-16 03:42:08,424][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.14129342138767242, 'train/grad_norm_pre_clip': 26.032970428466797, 'timing/data': 0.0003161370404995978, 'timing/model': 1.1314359499956481, 'learning_rate/qwen_vl_interface': 8.644914373967274e-07, 'learning_rate/action_model': 8.644914373967273e-06, 'epoch': 1.8}) +[2026-06-16 03:42:12,926][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 0.00649668276309967, 'train/grad_norm_pre_clip': 1.6147172451019287, 'timing/data': 0.00032262603053823113, 'timing/model': 1.1419014220009558, 'learning_rate/qwen_vl_interface': 8.633223982085068e-07, 'learning_rate/action_model': 8.633223982085068e-06, 'epoch': 1.8}) +[2026-06-16 03:42:17,482][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.016532722860574722, 'train/grad_norm_pre_clip': 10.982770919799805, 'timing/data': 0.0003135110018774867, 'timing/model': 1.310380605980754, 'learning_rate/qwen_vl_interface': 8.621551622177499e-07, 'learning_rate/action_model': 8.621551622177499e-06, 'epoch': 1.8}) +[2026-06-16 03:42:21,883][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 0.1020655632019043, 'train/grad_norm_pre_clip': 8.698360443115234, 'timing/data': 0.0003186010289937258, 'timing/model': 1.1463223149767146, 'learning_rate/qwen_vl_interface': 8.609897299042615e-07, 'learning_rate/action_model': 8.609897299042614e-06, 'epoch': 1.8}) +[2026-06-16 03:42:26,428][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.040820252150297165, 'train/grad_norm_pre_clip': 17.359832763671875, 'timing/data': 0.00036920298589393497, 'timing/model': 1.1391056479769759, 'learning_rate/qwen_vl_interface': 8.598261017471088e-07, 'learning_rate/action_model': 8.598261017471088e-06, 'epoch': 1.8}) +[2026-06-16 03:42:30,978][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.23991182446479797, 'train/grad_norm_pre_clip': 16.153995513916016, 'timing/data': 0.0005476020160131156, 'timing/model': 1.1399387959972955, 'learning_rate/qwen_vl_interface': 8.586642782246133e-07, 'learning_rate/action_model': 8.586642782246131e-06, 'epoch': 1.8}) +[2026-06-16 03:42:35,545][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.1385371834039688, 'train/grad_norm_pre_clip': 30.088272094726562, 'timing/data': 0.00032258802093565464, 'timing/model': 1.1400945280329324, 'learning_rate/qwen_vl_interface': 8.57504259814358e-07, 'learning_rate/action_model': 8.57504259814358e-06, 'epoch': 1.8}) +[2026-06-16 03:42:40,108][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.0006316140061244369, 'train/grad_norm_pre_clip': 0.12226654589176178, 'timing/data': 0.0005130000063218176, 'timing/model': 1.1394258430227637, 'learning_rate/qwen_vl_interface': 8.563460469931823e-07, 'learning_rate/action_model': 8.563460469931822e-06, 'epoch': 1.8}) +[2026-06-16 03:42:44,668][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.0012938614236190915, 'train/grad_norm_pre_clip': 1.1458483934402466, 'timing/data': 0.0003412719815969467, 'timing/model': 1.3057951999944635, 'learning_rate/qwen_vl_interface': 8.551896402371834e-07, 'learning_rate/action_model': 8.551896402371834e-06, 'epoch': 1.8}) +[2026-06-16 03:42:49,055][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.013274847529828548, 'train/grad_norm_pre_clip': 9.195372581481934, 'timing/data': 0.000325906032230705, 'timing/model': 1.1340605469886214, 'learning_rate/qwen_vl_interface': 8.540350400217177e-07, 'learning_rate/action_model': 8.540350400217176e-06, 'epoch': 1.81}) +[2026-06-16 03:42:53,590][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 0.09965959936380386, 'train/grad_norm_pre_clip': 10.587638854980469, 'timing/data': 0.00037433201214298606, 'timing/model': 1.1349757289863192, 'learning_rate/qwen_vl_interface': 8.528822468213964e-07, 'learning_rate/action_model': 8.528822468213964e-06, 'epoch': 1.81}) +[2026-06-16 03:42:58,146][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.19724057614803314, 'train/grad_norm_pre_clip': 49.68501281738281, 'timing/data': 0.00047119701048359275, 'timing/model': 1.134920051030349, 'learning_rate/qwen_vl_interface': 8.517312611100898e-07, 'learning_rate/action_model': 8.517312611100897e-06, 'epoch': 1.81}) +[2026-06-16 03:43:02,703][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.0009463655296713114, 'train/grad_norm_pre_clip': 0.09790869802236557, 'timing/data': 0.00034710101317614317, 'timing/model': 1.1347757009789348, 'learning_rate/qwen_vl_interface': 8.505820833609249e-07, 'learning_rate/action_model': 8.50582083360925e-06, 'epoch': 1.81}) +[2026-06-16 03:43:07,212][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.00971080083400011, 'train/grad_norm_pre_clip': 5.504824161529541, 'timing/data': 0.00034448603400960565, 'timing/model': 1.140649494016543, 'learning_rate/qwen_vl_interface': 8.494347140462847e-07, 'learning_rate/action_model': 8.494347140462846e-06, 'epoch': 1.81}) +[2026-06-16 03:43:11,790][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.002766847610473633, 'train/grad_norm_pre_clip': 0.7125975489616394, 'timing/data': 0.0003633960150182247, 'timing/model': 1.3183498539729044, 'learning_rate/qwen_vl_interface': 8.482891536378105e-07, 'learning_rate/action_model': 8.482891536378105e-06, 'epoch': 1.81}) +[2026-06-16 03:47:30,043][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.016961894929409027, 'train/grad_norm_pre_clip': 6.454977512359619, 'eval/loss': 0.039143964648246765, 'eval/flap_precision': 0.9666666666666667, 'eval/flap_recall': 0.9845679012345679, 'eval/flap_f1': 0.9755351681957187, 'eval/flap_rate_model': 0.18333333333333332, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9755351681957187, 'timing/data': 0.0003423520247451961, 'timing/model': 1.1424149279482663, 'learning_rate/qwen_vl_interface': 8.471454026063972e-07, 'learning_rate/action_model': 8.47145402606397e-06, 'epoch': 1.81}) +[2026-06-16 03:47:30,043][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4400_state +[2026-06-16 03:47:30,046][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 03:47:33,619][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4400_state/model.safetensors +[2026-06-16 03:47:43,606][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4400_state/optimizer.bin +[2026-06-16 03:47:43,607][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4400_state/sampler.bin +[2026-06-16 03:47:43,607][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4400_state/sampler_1.bin +[2026-06-16 03:47:43,608][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4400_state/random_states_0.pkl +[2026-06-16 03:47:43,651][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4400 +[2026-06-16 03:47:43,651][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 03:47:43,658][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 03:47:48,106][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.0016835839487612247, 'train/grad_norm_pre_clip': 0.32708051800727844, 'timing/data': 0.0006870690267533064, 'timing/model': 1.2793902729754336, 'learning_rate/qwen_vl_interface': 8.460034614221988e-07, 'learning_rate/action_model': 8.460034614221988e-06, 'epoch': 1.81}) +[2026-06-16 03:47:52,394][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.24386781454086304, 'train/grad_norm_pre_clip': 4.416893482208252, 'timing/data': 0.0003807860193774104, 'timing/model': 1.1088162219966762, 'learning_rate/qwen_vl_interface': 8.448633305546236e-07, 'learning_rate/action_model': 8.448633305546235e-06, 'epoch': 1.81}) +[2026-06-16 03:47:56,862][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.017663611099123955, 'train/grad_norm_pre_clip': 8.406222343444824, 'timing/data': 0.000698066025506705, 'timing/model': 1.1251532949972898, 'learning_rate/qwen_vl_interface': 8.437250104723369e-07, 'learning_rate/action_model': 8.437250104723369e-06, 'epoch': 1.81}) +[2026-06-16 03:48:01,334][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.13475705683231354, 'train/grad_norm_pre_clip': 9.971892356872559, 'timing/data': 0.0005778350168839097, 'timing/model': 1.1186078119790182, 'learning_rate/qwen_vl_interface': 8.42588501643259e-07, 'learning_rate/action_model': 8.42588501643259e-06, 'epoch': 1.81}) +[2026-06-16 03:48:05,809][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.16310158371925354, 'train/grad_norm_pre_clip': 15.870587348937988, 'timing/data': 0.0004946389817632735, 'timing/model': 1.1259950840030797, 'learning_rate/qwen_vl_interface': 8.414538045345655e-07, 'learning_rate/action_model': 8.414538045345654e-06, 'epoch': 1.81}) +[2026-06-16 03:48:10,301][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.0022612852044403553, 'train/grad_norm_pre_clip': 0.5191896557807922, 'timing/data': 0.0005086519522592425, 'timing/model': 1.281553637003526, 'learning_rate/qwen_vl_interface': 8.403209196126871e-07, 'learning_rate/action_model': 8.403209196126871e-06, 'epoch': 1.81}) +[2026-06-16 03:48:14,786][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.0008564154850319028, 'train/grad_norm_pre_clip': 0.18656840920448303, 'timing/data': 0.0004056540201418102, 'timing/model': 1.2580591270234436, 'learning_rate/qwen_vl_interface': 8.391898473433111e-07, 'learning_rate/action_model': 8.39189847343311e-06, 'epoch': 1.81}) +[2026-06-16 03:48:19,121][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.23019938170909882, 'train/grad_norm_pre_clip': 14.208765983581543, 'timing/data': 0.0003793410141952336, 'timing/model': 1.1228360739769414, 'learning_rate/qwen_vl_interface': 8.380605881913778e-07, 'learning_rate/action_model': 8.380605881913779e-06, 'epoch': 1.81}) +[2026-06-16 03:48:23,617][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 0.005538774188607931, 'train/grad_norm_pre_clip': 2.326890707015991, 'timing/data': 0.0004062579828314483, 'timing/model': 1.1229459550231695, 'learning_rate/qwen_vl_interface': 8.369331426210832e-07, 'learning_rate/action_model': 8.369331426210832e-06, 'epoch': 1.81}) +[2026-06-16 03:48:28,126][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.0011683149496093392, 'train/grad_norm_pre_clip': 0.19546450674533844, 'timing/data': 0.00035276199923828244, 'timing/model': 1.1254701780271716, 'learning_rate/qwen_vl_interface': 8.358075110958785e-07, 'learning_rate/action_model': 8.358075110958784e-06, 'epoch': 1.81}) +[2026-06-16 03:48:32,632][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.2341909557580948, 'train/grad_norm_pre_clip': 19.00798797607422, 'timing/data': 0.00037013698602095246, 'timing/model': 1.1250489969970658, 'learning_rate/qwen_vl_interface': 8.346836940784661e-07, 'learning_rate/action_model': 8.346836940784661e-06, 'epoch': 1.81}) +[2026-06-16 03:48:37,105][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.0030451666098088026, 'train/grad_norm_pre_clip': 0.440106600522995, 'timing/data': 0.00034904503263533115, 'timing/model': 1.129516334971413, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 1.81}) +[2026-06-16 03:48:41,628][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.046396490186452866, 'train/grad_norm_pre_clip': 14.853447914123535, 'timing/data': 0.0003476340207271278, 'timing/model': 1.2979764920310117, 'learning_rate/qwen_vl_interface': 8.324415054141111e-07, 'learning_rate/action_model': 8.32441505414111e-06, 'epoch': 1.81}) +[2026-06-16 03:48:45,981][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.0009096063440665603, 'train/grad_norm_pre_clip': 0.19521386921405792, 'timing/data': 0.0003666890552267432, 'timing/model': 1.1284681789693423, 'learning_rate/qwen_vl_interface': 8.313231346888476e-07, 'learning_rate/action_model': 8.313231346888476e-06, 'epoch': 1.81}) +[2026-06-16 03:48:50,500][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.18288417160511017, 'train/grad_norm_pre_clip': 41.53502655029297, 'timing/data': 0.00037720799446105957, 'timing/model': 1.1321524960221723, 'learning_rate/qwen_vl_interface': 8.302065803147353e-07, 'learning_rate/action_model': 8.302065803147353e-06, 'epoch': 1.81}) +[2026-06-16 03:48:55,027][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.0016120522050186992, 'train/grad_norm_pre_clip': 0.7644767165184021, 'timing/data': 0.00035715801641345024, 'timing/model': 1.1311933409888297, 'learning_rate/qwen_vl_interface': 8.29091842750748e-07, 'learning_rate/action_model': 8.29091842750748e-06, 'epoch': 1.81}) +[2026-06-16 03:48:59,554][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 0.11330293864011765, 'train/grad_norm_pre_clip': 20.044647216796875, 'timing/data': 0.0003284839913249016, 'timing/model': 1.1272428160300478, 'learning_rate/qwen_vl_interface': 8.279789224551116e-07, 'learning_rate/action_model': 8.279789224551116e-06, 'epoch': 1.81}) +[2026-06-16 03:49:04,074][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.020948311313986778, 'train/grad_norm_pre_clip': 10.523187637329102, 'timing/data': 0.0003756060032173991, 'timing/model': 1.2876433500205167, 'learning_rate/qwen_vl_interface': 8.268678198853058e-07, 'learning_rate/action_model': 8.268678198853059e-06, 'epoch': 1.81}) +[2026-06-16 03:49:08,595][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.06396488845348358, 'train/grad_norm_pre_clip': 36.15974807739258, 'timing/data': 0.0004297110135667026, 'timing/model': 1.284889884002041, 'learning_rate/qwen_vl_interface': 8.257585354980628e-07, 'learning_rate/action_model': 8.257585354980628e-06, 'epoch': 1.82}) +[2026-06-16 03:49:12,974][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.0627342015504837, 'train/grad_norm_pre_clip': 14.661934852600098, 'timing/data': 0.00036387197906151414, 'timing/model': 1.128697580017615, 'learning_rate/qwen_vl_interface': 8.246510697493681e-07, 'learning_rate/action_model': 8.246510697493682e-06, 'epoch': 1.82}) +[2026-06-16 03:49:17,493][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.10392872244119644, 'train/grad_norm_pre_clip': 16.004911422729492, 'timing/data': 0.0003211579751223326, 'timing/model': 1.127704165992327, 'learning_rate/qwen_vl_interface': 8.235454230944584e-07, 'learning_rate/action_model': 8.235454230944584e-06, 'epoch': 1.82}) +[2026-06-16 03:49:22,014][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.00041594915091991425, 'train/grad_norm_pre_clip': 0.14895401895046234, 'timing/data': 0.00035006296820938587, 'timing/model': 1.1304642430040985, 'learning_rate/qwen_vl_interface': 8.224415959878234e-07, 'learning_rate/action_model': 8.224415959878234e-06, 'epoch': 1.82}) +[2026-06-16 03:49:26,557][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.036398597061634064, 'train/grad_norm_pre_clip': 20.380264282226562, 'timing/data': 0.00036475900560617447, 'timing/model': 1.1320647000102326, 'learning_rate/qwen_vl_interface': 8.213395888832052e-07, 'learning_rate/action_model': 8.213395888832051e-06, 'epoch': 1.82}) +[2026-06-16 03:49:31,045][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.207130566239357, 'train/grad_norm_pre_clip': 36.97056198120117, 'timing/data': 0.0003445069887675345, 'timing/model': 1.1297547910362482, 'learning_rate/qwen_vl_interface': 8.202394022335958e-07, 'learning_rate/action_model': 8.202394022335957e-06, 'epoch': 1.82}) +[2026-06-16 03:49:35,582][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 0.27368253469467163, 'train/grad_norm_pre_clip': 27.383176803588867, 'timing/data': 0.0003451519878581166, 'timing/model': 1.2940847220015712, 'learning_rate/qwen_vl_interface': 8.191410364912424e-07, 'learning_rate/action_model': 8.191410364912424e-06, 'epoch': 1.82}) +[2026-06-16 03:49:39,956][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.0011885359417647123, 'train/grad_norm_pre_clip': 0.22588978707790375, 'timing/data': 0.0004176769871264696, 'timing/model': 1.1351327780284919, 'learning_rate/qwen_vl_interface': 8.1804449210764e-07, 'learning_rate/action_model': 8.1804449210764e-06, 'epoch': 1.82}) +[2026-06-16 03:49:44,494][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.0054734498262405396, 'train/grad_norm_pre_clip': 2.296403408050537, 'timing/data': 0.0004186619771644473, 'timing/model': 1.1317341570393182, 'learning_rate/qwen_vl_interface': 8.169497695335377e-07, 'learning_rate/action_model': 8.169497695335376e-06, 'epoch': 1.82}) +[2026-06-16 03:49:49,022][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 0.001120621687732637, 'train/grad_norm_pre_clip': 0.20855635404586792, 'timing/data': 0.00042215100256726146, 'timing/model': 1.1304823670070618, 'learning_rate/qwen_vl_interface': 8.158568692189342e-07, 'learning_rate/action_model': 8.158568692189342e-06, 'epoch': 1.82}) +[2026-06-16 03:49:53,576][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.0007867813692428172, 'train/grad_norm_pre_clip': 0.10985866189002991, 'timing/data': 0.0003591679851524532, 'timing/model': 1.136188694043085, 'learning_rate/qwen_vl_interface': 8.147657916130797e-07, 'learning_rate/action_model': 8.147657916130796e-06, 'epoch': 1.82}) +[2026-06-16 03:49:58,128][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.0632355585694313, 'train/grad_norm_pre_clip': 16.726377487182617, 'timing/data': 0.00037260999670252204, 'timing/model': 1.2967183649889193, 'learning_rate/qwen_vl_interface': 8.136765371644753e-07, 'learning_rate/action_model': 8.136765371644753e-06, 'epoch': 1.82}) +[2026-06-16 03:50:02,654][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.004853133577853441, 'train/grad_norm_pre_clip': 3.2012014389038086, 'timing/data': 0.00034606701228767633, 'timing/model': 1.2624005930265412, 'learning_rate/qwen_vl_interface': 8.125891063208713e-07, 'learning_rate/action_model': 8.125891063208712e-06, 'epoch': 1.82}) +[2026-06-16 03:50:07,058][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.09292126446962357, 'train/grad_norm_pre_clip': 17.612926483154297, 'timing/data': 0.000356237986125052, 'timing/model': 1.1446754460339434, 'learning_rate/qwen_vl_interface': 8.115034995292698e-07, 'learning_rate/action_model': 8.115034995292698e-06, 'epoch': 1.82}) +[2026-06-16 03:50:11,599][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.000637202407233417, 'train/grad_norm_pre_clip': 0.19769547879695892, 'timing/data': 0.0003687450080178678, 'timing/model': 1.139088018971961, 'learning_rate/qwen_vl_interface': 8.104197172359235e-07, 'learning_rate/action_model': 8.104197172359234e-06, 'epoch': 1.82}) +[2026-06-16 03:50:16,155][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.02708173170685768, 'train/grad_norm_pre_clip': 19.410924911499023, 'timing/data': 0.0003408250049687922, 'timing/model': 1.1340871180291288, 'learning_rate/qwen_vl_interface': 8.093377598863334e-07, 'learning_rate/action_model': 8.093377598863334e-06, 'epoch': 1.82}) +[2026-06-16 03:50:20,697][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.04874216765165329, 'train/grad_norm_pre_clip': 15.337004661560059, 'timing/data': 0.00033036305103451014, 'timing/model': 1.1330864049959928, 'learning_rate/qwen_vl_interface': 8.082576279252519e-07, 'learning_rate/action_model': 8.082576279252518e-06, 'epoch': 1.82}) +[2026-06-16 03:50:25,195][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.11127565801143646, 'train/grad_norm_pre_clip': 7.407897472381592, 'timing/data': 0.0003354120417498052, 'timing/model': 1.1340793099952862, 'learning_rate/qwen_vl_interface': 8.071793217966803e-07, 'learning_rate/action_model': 8.071793217966802e-06, 'epoch': 1.82}) +[2026-06-16 03:50:29,734][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.0021481274161487818, 'train/grad_norm_pre_clip': 0.47896572947502136, 'timing/data': 0.0003339860122650862, 'timing/model': 1.2942218949901871, 'learning_rate/qwen_vl_interface': 8.061028419438691e-07, 'learning_rate/action_model': 8.06102841943869e-06, 'epoch': 1.82}) +[2026-06-16 03:50:34,123][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 0.010090842843055725, 'train/grad_norm_pre_clip': 4.147958278656006, 'timing/data': 0.00033818697556853294, 'timing/model': 1.1347447439911775, 'learning_rate/qwen_vl_interface': 8.050281888093193e-07, 'learning_rate/action_model': 8.050281888093192e-06, 'epoch': 1.82}) +[2026-06-16 03:50:38,661][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.0012536721769720316, 'train/grad_norm_pre_clip': 0.3521445393562317, 'timing/data': 0.0003228060086257756, 'timing/model': 1.1357194969896227, 'learning_rate/qwen_vl_interface': 8.039553628347794e-07, 'learning_rate/action_model': 8.039553628347793e-06, 'epoch': 1.82}) +[2026-06-16 03:50:43,227][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.2075701802968979, 'train/grad_norm_pre_clip': 16.395118713378906, 'timing/data': 0.00045633799163624644, 'timing/model': 1.1333745429874398, 'learning_rate/qwen_vl_interface': 8.028843644612483e-07, 'learning_rate/action_model': 8.028843644612482e-06, 'epoch': 1.82}) +[2026-06-16 03:50:47,775][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.016891397535800934, 'train/grad_norm_pre_clip': 9.432307243347168, 'timing/data': 0.0006164660444483161, 'timing/model': 1.134117181005422, 'learning_rate/qwen_vl_interface': 8.01815194128973e-07, 'learning_rate/action_model': 8.01815194128973e-06, 'epoch': 1.82}) +[2026-06-16 03:50:52,314][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.03682177886366844, 'train/grad_norm_pre_clip': 13.950943946838379, 'timing/data': 0.00040065200300887227, 'timing/model': 1.2884101049858145, 'learning_rate/qwen_vl_interface': 8.007478522774481e-07, 'learning_rate/action_model': 8.00747852277448e-06, 'epoch': 1.82}) +[2026-06-16 03:50:56,834][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.0227053239941597, 'train/grad_norm_pre_clip': 7.9373345375061035, 'timing/data': 0.0004845529911108315, 'timing/model': 1.269159730989486, 'learning_rate/qwen_vl_interface': 7.996823393454197e-07, 'learning_rate/action_model': 7.996823393454196e-06, 'epoch': 1.83}) +[2026-06-16 03:51:01,216][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.000875045545399189, 'train/grad_norm_pre_clip': 0.13186350464820862, 'timing/data': 0.0003773740027099848, 'timing/model': 1.1340646109892987, 'learning_rate/qwen_vl_interface': 7.986186557708773e-07, 'learning_rate/action_model': 7.986186557708771e-06, 'epoch': 1.83}) +[2026-06-16 03:51:05,745][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.3168899714946747, 'train/grad_norm_pre_clip': 19.641664505004883, 'timing/data': 0.0003548579989001155, 'timing/model': 1.1347335560130887, 'learning_rate/qwen_vl_interface': 7.975568019910617e-07, 'learning_rate/action_model': 7.975568019910617e-06, 'epoch': 1.83}) +[2026-06-16 03:51:10,297][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.0012239417992532253, 'train/grad_norm_pre_clip': 0.22471266984939575, 'timing/data': 0.0003446349874138832, 'timing/model': 1.1447287139599212, 'learning_rate/qwen_vl_interface': 7.964967784424618e-07, 'learning_rate/action_model': 7.964967784424618e-06, 'epoch': 1.83}) +[2026-06-16 03:51:14,855][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.1387249231338501, 'train/grad_norm_pre_clip': 17.01112937927246, 'timing/data': 0.0003690060111694038, 'timing/model': 1.1363644110388122, 'learning_rate/qwen_vl_interface': 7.954385855608129e-07, 'learning_rate/action_model': 7.95438585560813e-06, 'epoch': 1.83}) +[2026-06-16 03:51:19,378][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.011367269791662693, 'train/grad_norm_pre_clip': 10.063607215881348, 'timing/data': 0.0006999199977144599, 'timing/model': 1.1393866920261644, 'learning_rate/qwen_vl_interface': 7.943822237810974e-07, 'learning_rate/action_model': 7.943822237810975e-06, 'epoch': 1.83}) +[2026-06-16 03:51:23,933][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.10486140102148056, 'train/grad_norm_pre_clip': 9.488554954528809, 'timing/data': 0.0003949980018660426, 'timing/model': 1.3027509470121004, 'learning_rate/qwen_vl_interface': 7.93327693537547e-07, 'learning_rate/action_model': 7.93327693537547e-06, 'epoch': 1.83}) +[2026-06-16 03:51:28,311][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.0440310575067997, 'train/grad_norm_pre_clip': 20.533767700195312, 'timing/data': 0.00037994200829416513, 'timing/model': 1.1337597990059294, 'learning_rate/qwen_vl_interface': 7.92274995263638e-07, 'learning_rate/action_model': 7.92274995263638e-06, 'epoch': 1.83}) +[2026-06-16 03:51:32,894][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.0001441649510525167, 'train/grad_norm_pre_clip': 0.030832504853606224, 'timing/data': 0.0003210600116290152, 'timing/model': 1.1342419810243882, 'learning_rate/qwen_vl_interface': 7.912241293920958e-07, 'learning_rate/action_model': 7.912241293920959e-06, 'epoch': 1.83}) +[2026-06-16 03:51:37,465][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.14693443477153778, 'train/grad_norm_pre_clip': 11.170016288757324, 'timing/data': 0.00033767003333196044, 'timing/model': 1.1357109579839744, 'learning_rate/qwen_vl_interface': 7.90175096354891e-07, 'learning_rate/action_model': 7.90175096354891e-06, 'epoch': 1.83}) +[2026-06-16 03:51:42,023][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.0015439708949998021, 'train/grad_norm_pre_clip': 0.19488006830215454, 'timing/data': 0.0003694170154631138, 'timing/model': 1.1352027989923954, 'learning_rate/qwen_vl_interface': 7.891278965832431e-07, 'learning_rate/action_model': 7.891278965832432e-06, 'epoch': 1.83}) +[2026-06-16 03:51:46,585][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.16295860707759857, 'train/grad_norm_pre_clip': 18.695898056030273, 'timing/data': 0.0003500480088405311, 'timing/model': 1.3037452739663422, 'learning_rate/qwen_vl_interface': 7.880825305076138e-07, 'learning_rate/action_model': 7.880825305076139e-06, 'epoch': 1.83}) +[2026-06-16 03:51:51,127][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.0007052810396999121, 'train/grad_norm_pre_clip': 0.40183737874031067, 'timing/data': 0.0004060339997522533, 'timing/model': 1.279502903984394, 'learning_rate/qwen_vl_interface': 7.870389985577148e-07, 'learning_rate/action_model': 7.870389985577148e-06, 'epoch': 1.83}) +[2026-06-16 03:51:55,528][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.06676527112722397, 'train/grad_norm_pre_clip': 14.320648193359375, 'timing/data': 0.0003472460084594786, 'timing/model': 1.1364240770344622, 'learning_rate/qwen_vl_interface': 7.859973011625028e-07, 'learning_rate/action_model': 7.859973011625026e-06, 'epoch': 1.83}) +[2026-06-16 03:52:00,070][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.0005568298511207104, 'train/grad_norm_pre_clip': 0.14741697907447815, 'timing/data': 0.0003206589608453214, 'timing/model': 1.134340308024548, 'learning_rate/qwen_vl_interface': 7.849574387501801e-07, 'learning_rate/action_model': 7.8495743875018e-06, 'epoch': 1.83}) +[2026-06-16 03:52:04,656][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.002774071181192994, 'train/grad_norm_pre_clip': 1.4182528257369995, 'timing/data': 0.00033693999284878373, 'timing/model': 1.141121248016134, 'learning_rate/qwen_vl_interface': 7.839194117481952e-07, 'learning_rate/action_model': 7.839194117481952e-06, 'epoch': 1.83}) +[2026-06-16 03:52:09,230][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.09198671579360962, 'train/grad_norm_pre_clip': 3.9647722244262695, 'timing/data': 0.00041609699837863445, 'timing/model': 1.1370856369612738, 'learning_rate/qwen_vl_interface': 7.828832205832418e-07, 'learning_rate/action_model': 7.828832205832418e-06, 'epoch': 1.83}) +[2026-06-16 03:52:13,739][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.0003660226066131145, 'train/grad_norm_pre_clip': 0.1060936450958252, 'timing/data': 0.00032191199716180563, 'timing/model': 1.13715023198165, 'learning_rate/qwen_vl_interface': 7.818488656812581e-07, 'learning_rate/action_model': 7.818488656812582e-06, 'epoch': 1.83}) +[2026-06-16 03:52:18,288][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 0.0037434687837958336, 'train/grad_norm_pre_clip': 0.6621962785720825, 'timing/data': 0.00031828798819333315, 'timing/model': 1.288379104982596, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 1.83}) +[2026-06-16 03:52:22,690][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.002076231874525547, 'train/grad_norm_pre_clip': 0.3705710470676422, 'timing/data': 0.00035605899756774306, 'timing/model': 1.1408707140362822, 'learning_rate/qwen_vl_interface': 7.797856663661836e-07, 'learning_rate/action_model': 7.797856663661835e-06, 'epoch': 1.83}) +[2026-06-16 03:52:27,269][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.002214373555034399, 'train/grad_norm_pre_clip': 0.29829689860343933, 'timing/data': 0.0003293929621577263, 'timing/model': 1.142196403001435, 'learning_rate/qwen_vl_interface': 7.787568228011959e-07, 'learning_rate/action_model': 7.78756822801196e-06, 'epoch': 1.83}) +[2026-06-16 03:52:31,823][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.03664245083928108, 'train/grad_norm_pre_clip': 29.492412567138672, 'timing/data': 0.00038198998663574457, 'timing/model': 1.135726810025517, 'learning_rate/qwen_vl_interface': 7.777298171953858e-07, 'learning_rate/action_model': 7.777298171953857e-06, 'epoch': 1.83}) +[2026-06-16 03:52:36,387][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.18237502872943878, 'train/grad_norm_pre_clip': 21.731487274169922, 'timing/data': 0.00033184600761160254, 'timing/model': 1.1359036510111764, 'learning_rate/qwen_vl_interface': 7.767046499709144e-07, 'learning_rate/action_model': 7.767046499709143e-06, 'epoch': 1.83}) +[2026-06-16 03:52:40,968][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.09684520214796066, 'train/grad_norm_pre_clip': 23.583147048950195, 'timing/data': 0.0005539290141314268, 'timing/model': 1.3200713540427387, 'learning_rate/qwen_vl_interface': 7.756813215491904e-07, 'learning_rate/action_model': 7.756813215491904e-06, 'epoch': 1.83}) +[2026-06-16 03:52:45,502][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.30642372369766235, 'train/grad_norm_pre_clip': 20.90599250793457, 'timing/data': 0.00035037094494327903, 'timing/model': 1.2733614110038616, 'learning_rate/qwen_vl_interface': 7.746598323508655e-07, 'learning_rate/action_model': 7.746598323508655e-06, 'epoch': 1.84}) +[2026-06-16 03:52:49,892][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.012747456319630146, 'train/grad_norm_pre_clip': 5.490203380584717, 'timing/data': 0.00031663599656894803, 'timing/model': 1.1377329800161533, 'learning_rate/qwen_vl_interface': 7.736401827958351e-07, 'learning_rate/action_model': 7.73640182795835e-06, 'epoch': 1.84}) +[2026-06-16 03:52:54,440][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 0.15237659215927124, 'train/grad_norm_pre_clip': 17.33430290222168, 'timing/data': 0.0003670809674076736, 'timing/model': 1.1383283789618872, 'learning_rate/qwen_vl_interface': 7.726223733032382e-07, 'learning_rate/action_model': 7.72622373303238e-06, 'epoch': 1.84}) +[2026-06-16 03:52:58,991][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.17966219782829285, 'train/grad_norm_pre_clip': 17.039878845214844, 'timing/data': 0.00034998898627236485, 'timing/model': 1.1360502379829995, 'learning_rate/qwen_vl_interface': 7.716064042914584e-07, 'learning_rate/action_model': 7.716064042914584e-06, 'epoch': 1.84}) +[2026-06-16 03:53:03,558][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.0001887386606540531, 'train/grad_norm_pre_clip': 0.01862809620797634, 'timing/data': 0.00036231399280950427, 'timing/model': 1.137028853991069, 'learning_rate/qwen_vl_interface': 7.705922761781226e-07, 'learning_rate/action_model': 7.705922761781226e-06, 'epoch': 1.84}) +[2026-06-16 03:53:08,068][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.02943282574415207, 'train/grad_norm_pre_clip': 13.155997276306152, 'timing/data': 0.0003583240322768688, 'timing/model': 1.1350298119941726, 'learning_rate/qwen_vl_interface': 7.695799893800996e-07, 'learning_rate/action_model': 7.695799893800996e-06, 'epoch': 1.84}) +[2026-06-16 03:53:12,640][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.003465508809313178, 'train/grad_norm_pre_clip': 2.351078748703003, 'timing/data': 0.0003372519859112799, 'timing/model': 1.3049837529542856, 'learning_rate/qwen_vl_interface': 7.685695443135028e-07, 'learning_rate/action_model': 7.685695443135027e-06, 'epoch': 1.84}) +[2026-06-16 03:53:17,056][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 0.018129199743270874, 'train/grad_norm_pre_clip': 8.571023941040039, 'timing/data': 0.00034689996391534805, 'timing/model': 1.1411428229766898, 'learning_rate/qwen_vl_interface': 7.675609413936888e-07, 'learning_rate/action_model': 7.675609413936888e-06, 'epoch': 1.84}) +[2026-06-16 03:53:21,596][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.006411219947040081, 'train/grad_norm_pre_clip': 3.3224117755889893, 'timing/data': 0.0003293129848316312, 'timing/model': 1.1376914079883136, 'learning_rate/qwen_vl_interface': 7.665541810352541e-07, 'learning_rate/action_model': 7.66554181035254e-06, 'epoch': 1.84}) +[2026-06-16 03:53:26,176][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.00403883121907711, 'train/grad_norm_pre_clip': 1.073808193206787, 'timing/data': 0.0003510810201987624, 'timing/model': 1.151876761985477, 'learning_rate/qwen_vl_interface': 7.655492636520415e-07, 'learning_rate/action_model': 7.655492636520414e-06, 'epoch': 1.84}) +[2026-06-16 03:53:30,722][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.5424270033836365, 'train/grad_norm_pre_clip': 14.269373893737793, 'timing/data': 0.0005938619724474847, 'timing/model': 1.1375359650119208, 'learning_rate/qwen_vl_interface': 7.64546189657134e-07, 'learning_rate/action_model': 7.645461896571339e-06, 'epoch': 1.84}) +[2026-06-16 03:53:35,266][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.004964457359164953, 'train/grad_norm_pre_clip': 3.364757776260376, 'timing/data': 0.0003770270268432796, 'timing/model': 1.2833326890249737, 'learning_rate/qwen_vl_interface': 7.635449594628571e-07, 'learning_rate/action_model': 7.635449594628571e-06, 'epoch': 1.84}) +[2026-06-16 03:53:39,802][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.15797676146030426, 'train/grad_norm_pre_clip': 16.67811393737793, 'timing/data': 0.00036585202906280756, 'timing/model': 1.2728918009670451, 'learning_rate/qwen_vl_interface': 7.625455734807803e-07, 'learning_rate/action_model': 7.625455734807803e-06, 'epoch': 1.84}) +[2026-06-16 03:53:44,211][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.023591281846165657, 'train/grad_norm_pre_clip': 8.804332733154297, 'timing/data': 0.0003346939920447767, 'timing/model': 1.1362362899817526, 'learning_rate/qwen_vl_interface': 7.615480321217116e-07, 'learning_rate/action_model': 7.615480321217117e-06, 'epoch': 1.84}) +[2026-06-16 03:53:48,752][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.0007217596285045147, 'train/grad_norm_pre_clip': 0.09358616173267365, 'timing/data': 0.0006176549941301346, 'timing/model': 1.1396117529948242, 'learning_rate/qwen_vl_interface': 7.605523357957028e-07, 'learning_rate/action_model': 7.605523357957028e-06, 'epoch': 1.84}) +[2026-06-16 03:53:53,295][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.0012171313865110278, 'train/grad_norm_pre_clip': 0.17244984209537506, 'timing/data': 0.00033229298423975706, 'timing/model': 1.1365209220093675, 'learning_rate/qwen_vl_interface': 7.59558484912048e-07, 'learning_rate/action_model': 7.59558484912048e-06, 'epoch': 1.84}) +[2026-06-16 03:53:57,861][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.00039812171598896384, 'train/grad_norm_pre_clip': 0.07942886650562286, 'timing/data': 0.000353328010533005, 'timing/model': 1.1462559540523216, 'learning_rate/qwen_vl_interface': 7.585664798792807e-07, 'learning_rate/action_model': 7.585664798792807e-06, 'epoch': 1.84}) +[2026-06-16 03:54:02,369][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.0005588394124060869, 'train/grad_norm_pre_clip': 0.2584556043148041, 'timing/data': 0.0005000710370950401, 'timing/model': 1.1386982320109382, 'learning_rate/qwen_vl_interface': 7.575763211051779e-07, 'learning_rate/action_model': 7.575763211051778e-06, 'epoch': 1.84}) +[2026-06-16 03:54:06,930][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.005078677088022232, 'train/grad_norm_pre_clip': 1.5943214893341064, 'timing/data': 0.0003499449812807143, 'timing/model': 1.307952531031333, 'learning_rate/qwen_vl_interface': 7.565880089967553e-07, 'learning_rate/action_model': 7.565880089967553e-06, 'epoch': 1.84}) +[2026-06-16 03:54:11,334][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.01609787717461586, 'train/grad_norm_pre_clip': 10.11250114440918, 'timing/data': 0.000386201951187104, 'timing/model': 1.1355100210057572, 'learning_rate/qwen_vl_interface': 7.556015439602725e-07, 'learning_rate/action_model': 7.556015439602725e-06, 'epoch': 1.84}) +[2026-06-16 03:54:15,898][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.0021601158659905195, 'train/grad_norm_pre_clip': 0.49337685108184814, 'timing/data': 0.0003473180113360286, 'timing/model': 1.1452900790027343, 'learning_rate/qwen_vl_interface': 7.546169264012252e-07, 'learning_rate/action_model': 7.5461692640122525e-06, 'epoch': 1.84}) +[2026-06-16 03:54:20,465][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.01617884822189808, 'train/grad_norm_pre_clip': 18.165790557861328, 'timing/data': 0.0003134380094707012, 'timing/model': 1.141862439981196, 'learning_rate/qwen_vl_interface': 7.536341567243544e-07, 'learning_rate/action_model': 7.536341567243543e-06, 'epoch': 1.84}) +[2026-06-16 03:54:25,017][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.04740500450134277, 'train/grad_norm_pre_clip': 35.25459671020508, 'timing/data': 0.0003423520247451961, 'timing/model': 1.1361534459865652, 'learning_rate/qwen_vl_interface': 7.526532353336395e-07, 'learning_rate/action_model': 7.526532353336395e-06, 'epoch': 1.84}) +[2026-06-16 03:54:29,575][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.00024342320102732629, 'train/grad_norm_pre_clip': 0.04953701049089432, 'timing/data': 0.00035117799416184425, 'timing/model': 1.3049521999782883, 'learning_rate/qwen_vl_interface': 7.516741626322994e-07, 'learning_rate/action_model': 7.516741626322994e-06, 'epoch': 1.84}) +[2026-06-16 03:54:34,103][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.019012076780200005, 'train/grad_norm_pre_clip': 6.529776096343994, 'timing/data': 0.0003485149936750531, 'timing/model': 1.2718348890193738, 'learning_rate/qwen_vl_interface': 7.506969390227945e-07, 'learning_rate/action_model': 7.506969390227945e-06, 'epoch': 1.84}) +[2026-06-16 03:54:38,489][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.002461581025272608, 'train/grad_norm_pre_clip': 0.7583792805671692, 'timing/data': 0.00036112999077886343, 'timing/model': 1.1358021199703217, 'learning_rate/qwen_vl_interface': 7.497215649068244e-07, 'learning_rate/action_model': 7.497215649068244e-06, 'epoch': 1.85}) +[2026-06-16 03:54:43,053][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.12023390084505081, 'train/grad_norm_pre_clip': 29.70216941833496, 'timing/data': 0.0005840039812028408, 'timing/model': 1.137957250000909, 'learning_rate/qwen_vl_interface': 7.48748040685329e-07, 'learning_rate/action_model': 7.487480406853289e-06, 'epoch': 1.85}) +[2026-06-16 03:54:47,626][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.0020695915445685387, 'train/grad_norm_pre_clip': 0.5229538679122925, 'timing/data': 0.00033865898149088025, 'timing/model': 1.1361917270114645, 'learning_rate/qwen_vl_interface': 7.477763667584867e-07, 'learning_rate/action_model': 7.477763667584866e-06, 'epoch': 1.85}) +[2026-06-16 03:54:52,199][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.005792444106191397, 'train/grad_norm_pre_clip': 5.188638210296631, 'timing/data': 0.0003578959731385112, 'timing/model': 1.1444453609874472, 'learning_rate/qwen_vl_interface': 7.468065435257159e-07, 'learning_rate/action_model': 7.468065435257159e-06, 'epoch': 1.85}) +[2026-06-16 03:54:56,721][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.0007731700898148119, 'train/grad_norm_pre_clip': 0.14398512244224548, 'timing/data': 0.0003635960165411234, 'timing/model': 1.1410459650214761, 'learning_rate/qwen_vl_interface': 7.458385713856745e-07, 'learning_rate/action_model': 7.458385713856744e-06, 'epoch': 1.85}) +[2026-06-16 03:55:01,262][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.008542497642338276, 'train/grad_norm_pre_clip': 8.135576248168945, 'timing/data': 0.0004011570126749575, 'timing/model': 1.295473696955014, 'learning_rate/qwen_vl_interface': 7.448724507362605e-07, 'learning_rate/action_model': 7.448724507362605e-06, 'epoch': 1.85}) +[2026-06-16 03:55:05,672][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.0019356405828148127, 'train/grad_norm_pre_clip': 0.740172803401947, 'timing/data': 0.00037487404188141227, 'timing/model': 1.1406010570353828, 'learning_rate/qwen_vl_interface': 7.439081819746078e-07, 'learning_rate/action_model': 7.4390818197460775e-06, 'epoch': 1.85}) +[2026-06-16 03:55:10,224][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.12736018002033234, 'train/grad_norm_pre_clip': 11.341134071350098, 'timing/data': 0.00035395805025473237, 'timing/model': 1.1410987350391224, 'learning_rate/qwen_vl_interface': 7.42945765497091e-07, 'learning_rate/action_model': 7.42945765497091e-06, 'epoch': 1.85}) +[2026-06-16 03:59:31,856][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.05825686827301979, 'train/grad_norm_pre_clip': 18.751123428344727, 'eval/loss': 0.03843902796506882, 'eval/flap_precision': 0.9681335356600911, 'eval/flap_recall': 0.9845679012345679, 'eval/flap_f1': 0.9762815608263199, 'eval/flap_rate_model': 0.18305555555555555, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9762815608263199, 'timing/data': 0.0003544189967215061, 'timing/model': 1.140704450022895, 'learning_rate/qwen_vl_interface': 7.41985201699324e-07, 'learning_rate/action_model': 7.41985201699324e-06, 'epoch': 1.85}) +[2026-06-16 03:59:31,856][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4500_state +[2026-06-16 03:59:31,859][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 03:59:35,481][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4500_state/model.safetensors +[2026-06-16 03:59:43,071][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4500_state/optimizer.bin +[2026-06-16 03:59:43,071][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4500_state/sampler.bin +[2026-06-16 03:59:43,071][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4500_state/sampler_1.bin +[2026-06-16 03:59:43,072][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4500_state/random_states_0.pkl +[2026-06-16 03:59:43,118][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-06-16 03:59:43,119][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 03:59:43,125][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 03:59:47,552][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.0006549388635903597, 'train/grad_norm_pre_clip': 0.13864842057228088, 'timing/data': 0.00035273999674245715, 'timing/model': 1.119462819013279, 'learning_rate/qwen_vl_interface': 7.410264909761581e-07, 'learning_rate/action_model': 7.41026490976158e-06, 'epoch': 1.85}) +[2026-06-16 03:59:52,060][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.0033707914408296347, 'train/grad_norm_pre_clip': 1.2186332941055298, 'timing/data': 0.0003412080113776028, 'timing/model': 1.1183889859821647, 'learning_rate/qwen_vl_interface': 7.40069633721683e-07, 'learning_rate/action_model': 7.400696337216829e-06, 'epoch': 1.85}) +[2026-06-16 03:59:56,543][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.003401861060410738, 'train/grad_norm_pre_clip': 0.8251967430114746, 'timing/data': 0.00037650298327207565, 'timing/model': 1.1168292600195855, 'learning_rate/qwen_vl_interface': 7.391146303292262e-07, 'learning_rate/action_model': 7.391146303292262e-06, 'epoch': 1.85}) +[2026-06-16 04:00:01,025][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.3144053518772125, 'train/grad_norm_pre_clip': 20.94719886779785, 'timing/data': 0.00034300796687602997, 'timing/model': 1.2743443340295926, 'learning_rate/qwen_vl_interface': 7.381614811913538e-07, 'learning_rate/action_model': 7.381614811913537e-06, 'epoch': 1.85}) +[2026-06-16 04:00:05,378][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.056744568049907684, 'train/grad_norm_pre_clip': 13.828624725341797, 'timing/data': 0.0003970760153606534, 'timing/model': 1.1361370920203626, 'learning_rate/qwen_vl_interface': 7.372101866998693e-07, 'learning_rate/action_model': 7.372101866998693e-06, 'epoch': 1.85}) +[2026-06-16 04:00:09,873][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.00024576237774454057, 'train/grad_norm_pre_clip': 0.03484446927905083, 'timing/data': 0.00036302400985732675, 'timing/model': 1.129488039005082, 'learning_rate/qwen_vl_interface': 7.362607472458147e-07, 'learning_rate/action_model': 7.362607472458147e-06, 'epoch': 1.85}) +[2026-06-16 04:00:14,317][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.0007173760095611215, 'train/grad_norm_pre_clip': 5.63575553894043, 'timing/data': 0.00035446096444502473, 'timing/model': 1.121268407034222, 'learning_rate/qwen_vl_interface': 7.353131632194682e-07, 'learning_rate/action_model': 7.3531316321946816e-06, 'epoch': 1.85}) +[2026-06-16 04:00:18,818][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.00598635571077466, 'train/grad_norm_pre_clip': 1.3594945669174194, 'timing/data': 0.0003388330223970115, 'timing/model': 1.1270447800052352, 'learning_rate/qwen_vl_interface': 7.343674350103455e-07, 'learning_rate/action_model': 7.343674350103455e-06, 'epoch': 1.85}) +[2026-06-16 04:00:23,324][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.08936218172311783, 'train/grad_norm_pre_clip': 8.985947608947754, 'timing/data': 0.0003344910219311714, 'timing/model': 1.1259316420182586, 'learning_rate/qwen_vl_interface': 7.334235630071995e-07, 'learning_rate/action_model': 7.334235630071995e-06, 'epoch': 1.85}) +[2026-06-16 04:00:27,827][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.16614724695682526, 'train/grad_norm_pre_clip': 35.65069580078125, 'timing/data': 0.00037644500844180584, 'timing/model': 1.28297078498872, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 1.85}) +[2026-06-16 04:00:32,270][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 0.007437916472554207, 'train/grad_norm_pre_clip': 5.8208842277526855, 'timing/data': 0.0003621329669840634, 'timing/model': 1.2130579840159044, 'learning_rate/qwen_vl_interface': 7.315413891700363e-07, 'learning_rate/action_model': 7.315413891700362e-06, 'epoch': 1.85}) +[2026-06-16 04:00:36,630][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.0033998903818428516, 'train/grad_norm_pre_clip': 1.923771858215332, 'timing/data': 0.0003807320026680827, 'timing/model': 1.1318526139948517, 'learning_rate/qwen_vl_interface': 7.306030881097092e-07, 'learning_rate/action_model': 7.3060308810970925e-06, 'epoch': 1.85}) +[2026-06-16 04:00:41,133][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.07509718090295792, 'train/grad_norm_pre_clip': 9.52686882019043, 'timing/data': 0.00034146796679124236, 'timing/model': 1.13039269100409, 'learning_rate/qwen_vl_interface': 7.296666448027406e-07, 'learning_rate/action_model': 7.296666448027406e-06, 'epoch': 1.85}) +[2026-06-16 04:00:45,657][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.0005318645853549242, 'train/grad_norm_pre_clip': 0.17724129557609558, 'timing/data': 0.00034596899058669806, 'timing/model': 1.1395112639875151, 'learning_rate/qwen_vl_interface': 7.287320596340664e-07, 'learning_rate/action_model': 7.287320596340663e-06, 'epoch': 1.85}) +[2026-06-16 04:00:50,169][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.15344057977199554, 'train/grad_norm_pre_clip': 8.50525951385498, 'timing/data': 0.0003748360322788358, 'timing/model': 1.1308553360286169, 'learning_rate/qwen_vl_interface': 7.277993329878593e-07, 'learning_rate/action_model': 7.277993329878594e-06, 'epoch': 1.85}) +[2026-06-16 04:00:54,704][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.26945704221725464, 'train/grad_norm_pre_clip': 18.906108856201172, 'timing/data': 0.0003827270120382309, 'timing/model': 1.3035946959862486, 'learning_rate/qwen_vl_interface': 7.268684652475277e-07, 'learning_rate/action_model': 7.268684652475277e-06, 'epoch': 1.86}) +[2026-06-16 04:00:59,128][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.03011021949350834, 'train/grad_norm_pre_clip': 26.24137306213379, 'timing/data': 0.00047376001020893455, 'timing/model': 1.1937761009903625, 'learning_rate/qwen_vl_interface': 7.259394567957169e-07, 'learning_rate/action_model': 7.259394567957169e-06, 'epoch': 1.86}) +[2026-06-16 04:01:03,515][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.1488248109817505, 'train/grad_norm_pre_clip': 19.505329132080078, 'timing/data': 0.0003570220433175564, 'timing/model': 1.1343684899620712, 'learning_rate/qwen_vl_interface': 7.250123080143066e-07, 'learning_rate/action_model': 7.2501230801430665e-06, 'epoch': 1.86}) +[2026-06-16 04:01:07,958][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.01280486211180687, 'train/grad_norm_pre_clip': 4.785356044769287, 'timing/data': 0.0003818259574472904, 'timing/model': 1.1315355710103177, 'learning_rate/qwen_vl_interface': 7.240870192844127e-07, 'learning_rate/action_model': 7.240870192844127e-06, 'epoch': 1.86}) +[2026-06-16 04:01:12,451][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.005768870934844017, 'train/grad_norm_pre_clip': 1.5444450378417969, 'timing/data': 0.00040098000317811966, 'timing/model': 1.132838724006433, 'learning_rate/qwen_vl_interface': 7.231635909863858e-07, 'learning_rate/action_model': 7.2316359098638585e-06, 'epoch': 1.86}) +[2026-06-16 04:01:16,980][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 0.00033578957663848996, 'train/grad_norm_pre_clip': 0.04313403367996216, 'timing/data': 0.00032475998159497976, 'timing/model': 1.1300597380031832, 'learning_rate/qwen_vl_interface': 7.222420234998138e-07, 'learning_rate/action_model': 7.2224202349981385e-06, 'epoch': 1.86}) +[2026-06-16 04:01:21,504][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.03713150694966316, 'train/grad_norm_pre_clip': 18.609886169433594, 'timing/data': 0.00035784696228802204, 'timing/model': 1.2819284229772165, 'learning_rate/qwen_vl_interface': 7.213223172035175e-07, 'learning_rate/action_model': 7.213223172035174e-06, 'epoch': 1.86}) +[2026-06-16 04:01:25,881][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.003122192109003663, 'train/grad_norm_pre_clip': 2.8428733348846436, 'timing/data': 0.0004069010028615594, 'timing/model': 1.1290167950210162, 'learning_rate/qwen_vl_interface': 7.204044724755528e-07, 'learning_rate/action_model': 7.204044724755528e-06, 'epoch': 1.86}) +[2026-06-16 04:01:30,410][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.0017876848578453064, 'train/grad_norm_pre_clip': 0.3762889504432678, 'timing/data': 0.00035079900408163667, 'timing/model': 1.1299876190023497, 'learning_rate/qwen_vl_interface': 7.194884896932132e-07, 'learning_rate/action_model': 7.194884896932132e-06, 'epoch': 1.86}) +[2026-06-16 04:01:34,941][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.20014730095863342, 'train/grad_norm_pre_clip': 11.118033409118652, 'timing/data': 0.00034774100640788674, 'timing/model': 1.1347651279647835, 'learning_rate/qwen_vl_interface': 7.185743692330221e-07, 'learning_rate/action_model': 7.185743692330221e-06, 'epoch': 1.86}) +[2026-06-16 04:01:39,508][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 0.2918924391269684, 'train/grad_norm_pre_clip': 9.908219337463379, 'timing/data': 0.00035133701749145985, 'timing/model': 1.1466952479677275, 'learning_rate/qwen_vl_interface': 7.176621114707408e-07, 'learning_rate/action_model': 7.176621114707407e-06, 'epoch': 1.86}) +[2026-06-16 04:01:44,045][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.005041345022618771, 'train/grad_norm_pre_clip': 1.6864843368530273, 'timing/data': 0.0003388900076970458, 'timing/model': 1.1407358479918912, 'learning_rate/qwen_vl_interface': 7.167517167813642e-07, 'learning_rate/action_model': 7.167517167813642e-06, 'epoch': 1.86}) +[2026-06-16 04:01:48,570][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.042677268385887146, 'train/grad_norm_pre_clip': 34.363033294677734, 'timing/data': 0.0003321919939480722, 'timing/model': 1.2870658580213785, 'learning_rate/qwen_vl_interface': 7.15843185539121e-07, 'learning_rate/action_model': 7.15843185539121e-06, 'epoch': 1.86}) +[2026-06-16 04:01:52,944][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.037664007395505905, 'train/grad_norm_pre_clip': 12.547768592834473, 'timing/data': 0.0004009550320915878, 'timing/model': 1.1324326999601908, 'learning_rate/qwen_vl_interface': 7.149365181174739e-07, 'learning_rate/action_model': 7.149365181174739e-06, 'epoch': 1.86}) +[2026-06-16 04:01:57,467][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.2956201732158661, 'train/grad_norm_pre_clip': 19.87335777282715, 'timing/data': 0.0003245229599997401, 'timing/model': 1.131311110977549, 'learning_rate/qwen_vl_interface': 7.14031714889121e-07, 'learning_rate/action_model': 7.14031714889121e-06, 'epoch': 1.86}) +[2026-06-16 04:02:01,969][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.14428691565990448, 'train/grad_norm_pre_clip': 24.06450653076172, 'timing/data': 0.0004048349801450968, 'timing/model': 1.1327634140034206, 'learning_rate/qwen_vl_interface': 7.131287762259907e-07, 'learning_rate/action_model': 7.131287762259907e-06, 'epoch': 1.86}) +[2026-06-16 04:02:06,517][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.0015344349667429924, 'train/grad_norm_pre_clip': 0.648104727268219, 'timing/data': 0.0006600169581361115, 'timing/model': 1.1373717909445986, 'learning_rate/qwen_vl_interface': 7.12227702499248e-07, 'learning_rate/action_model': 7.12227702499248e-06, 'epoch': 1.86}) +[2026-06-16 04:02:11,059][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.045515913516283035, 'train/grad_norm_pre_clip': 20.202777862548828, 'timing/data': 0.0004496550536714494, 'timing/model': 1.1345536700100638, 'learning_rate/qwen_vl_interface': 7.113284940792903e-07, 'learning_rate/action_model': 7.113284940792903e-06, 'epoch': 1.86}) +[2026-06-16 04:02:15,606][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 8.162062295014039e-05, 'train/grad_norm_pre_clip': 0.00916686374694109, 'timing/data': 0.0003296859795227647, 'timing/model': 1.3030728460289538, 'learning_rate/qwen_vl_interface': 7.104311513357476e-07, 'learning_rate/action_model': 7.1043115133574755e-06, 'epoch': 1.86}) +[2026-06-16 04:02:20,066][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.32819801568984985, 'train/grad_norm_pre_clip': 13.043570518493652, 'timing/data': 0.0003994100261479616, 'timing/model': 1.217996290011797, 'learning_rate/qwen_vl_interface': 7.095356746374845e-07, 'learning_rate/action_model': 7.095356746374844e-06, 'epoch': 1.86}) +[2026-06-16 04:02:24,458][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.0012861810391768813, 'train/grad_norm_pre_clip': 0.3658807575702667, 'timing/data': 0.0006417889962904155, 'timing/model': 1.1358552309684455, 'learning_rate/qwen_vl_interface': 7.086420643525973e-07, 'learning_rate/action_model': 7.086420643525972e-06, 'epoch': 1.86}) +[2026-06-16 04:02:28,963][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.001998102990910411, 'train/grad_norm_pre_clip': 0.36019909381866455, 'timing/data': 0.00033143803011626005, 'timing/model': 1.137329782010056, 'learning_rate/qwen_vl_interface': 7.077503208484151e-07, 'learning_rate/action_model': 7.0775032084841515e-06, 'epoch': 1.86}) +[2026-06-16 04:02:33,502][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.13875040411949158, 'train/grad_norm_pre_clip': 9.384389877319336, 'timing/data': 0.0003224309766665101, 'timing/model': 1.1334592750063166, 'learning_rate/qwen_vl_interface': 7.068604444915003e-07, 'learning_rate/action_model': 7.068604444915003e-06, 'epoch': 1.86}) +[2026-06-16 04:02:38,057][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.15080860257148743, 'train/grad_norm_pre_clip': 27.412805557250977, 'timing/data': 0.0009412000072188675, 'timing/model': 1.1500820579822175, 'learning_rate/qwen_vl_interface': 7.059724356476475e-07, 'learning_rate/action_model': 7.059724356476474e-06, 'epoch': 1.86}) +[2026-06-16 04:02:42,586][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.045544736087322235, 'train/grad_norm_pre_clip': 25.494070053100586, 'timing/data': 0.0003928419901058078, 'timing/model': 1.2811571640195325, 'learning_rate/qwen_vl_interface': 7.050862946818836e-07, 'learning_rate/action_model': 7.050862946818836e-06, 'epoch': 1.87}) +[2026-06-16 04:02:47,020][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.0012284287950024009, 'train/grad_norm_pre_clip': 0.3002815842628479, 'timing/data': 0.0003032139502465725, 'timing/model': 1.1859181199688464, 'learning_rate/qwen_vl_interface': 7.042020219584673e-07, 'learning_rate/action_model': 7.042020219584673e-06, 'epoch': 1.87}) +[2026-06-16 04:02:51,415][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.0019368256907910109, 'train/grad_norm_pre_clip': 0.77392578125, 'timing/data': 0.00036048702895641327, 'timing/model': 1.1392058820347302, 'learning_rate/qwen_vl_interface': 7.0331961784089e-07, 'learning_rate/action_model': 7.033196178408899e-06, 'epoch': 1.87}) +[2026-06-16 04:02:55,879][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.053770218044519424, 'train/grad_norm_pre_clip': 17.553361892700195, 'timing/data': 0.00031829404179006815, 'timing/model': 1.1389023910160176, 'learning_rate/qwen_vl_interface': 7.024390826918748e-07, 'learning_rate/action_model': 7.024390826918747e-06, 'epoch': 1.87}) +[2026-06-16 04:03:00,386][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.002467525890097022, 'train/grad_norm_pre_clip': 0.33865150809288025, 'timing/data': 0.00042907503666356206, 'timing/model': 1.1322806249954738, 'learning_rate/qwen_vl_interface': 7.01560416873376e-07, 'learning_rate/action_model': 7.015604168733759e-06, 'epoch': 1.87}) +[2026-06-16 04:03:04,924][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.16417941451072693, 'train/grad_norm_pre_clip': 15.136223793029785, 'timing/data': 0.0003504789783619344, 'timing/model': 1.1331358820316382, 'learning_rate/qwen_vl_interface': 7.006836207465803e-07, 'learning_rate/action_model': 7.006836207465803e-06, 'epoch': 1.87}) +[2026-06-16 04:03:09,467][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 0.08806723356246948, 'train/grad_norm_pre_clip': 21.026437759399414, 'timing/data': 0.00040861498564481735, 'timing/model': 1.2939674709923565, 'learning_rate/qwen_vl_interface': 6.998086946719058e-07, 'learning_rate/action_model': 6.998086946719057e-06, 'epoch': 1.87}) +[2026-06-16 04:03:13,872][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.003174107987433672, 'train/grad_norm_pre_clip': 0.5318617820739746, 'timing/data': 0.0003830960486084223, 'timing/model': 1.1413278330001049, 'learning_rate/qwen_vl_interface': 6.989356390090008e-07, 'learning_rate/action_model': 6.989356390090007e-06, 'epoch': 1.87}) +[2026-06-16 04:03:18,417][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.009882497601211071, 'train/grad_norm_pre_clip': 5.0339436531066895, 'timing/data': 0.0003490819945000112, 'timing/model': 1.1368844530079514, 'learning_rate/qwen_vl_interface': 6.980644541167459e-07, 'learning_rate/action_model': 6.980644541167458e-06, 'epoch': 1.87}) +[2026-06-16 04:03:22,959][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.04180235043168068, 'train/grad_norm_pre_clip': 16.701717376708984, 'timing/data': 0.00036315101897343993, 'timing/model': 1.136810522992164, 'learning_rate/qwen_vl_interface': 6.971951403532524e-07, 'learning_rate/action_model': 6.971951403532523e-06, 'epoch': 1.87}) +[2026-06-16 04:03:27,501][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.08365125209093094, 'train/grad_norm_pre_clip': 21.27781867980957, 'timing/data': 0.00032910500885918736, 'timing/model': 1.1382800349965692, 'learning_rate/qwen_vl_interface': 6.963276980758622e-07, 'learning_rate/action_model': 6.963276980758622e-06, 'epoch': 1.87}) +[2026-06-16 04:03:32,050][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.00022645655553787947, 'train/grad_norm_pre_clip': 0.018153022974729538, 'timing/data': 0.0005007979925721884, 'timing/model': 1.1334832819993608, 'learning_rate/qwen_vl_interface': 6.954621276411485e-07, 'learning_rate/action_model': 6.954621276411484e-06, 'epoch': 1.87}) +[2026-06-16 04:03:36,600][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 0.013685754500329494, 'train/grad_norm_pre_clip': 8.958230018615723, 'timing/data': 0.0005808250280097127, 'timing/model': 1.3048297230270691, 'learning_rate/qwen_vl_interface': 6.945984294049136e-07, 'learning_rate/action_model': 6.9459842940491364e-06, 'epoch': 1.87}) +[2026-06-16 04:03:40,977][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.000549683696590364, 'train/grad_norm_pre_clip': 0.18252262473106384, 'timing/data': 0.00035600701812654734, 'timing/model': 1.1333366859471425, 'learning_rate/qwen_vl_interface': 6.937366037221919e-07, 'learning_rate/action_model': 6.937366037221918e-06, 'epoch': 1.87}) +[2026-06-16 04:03:45,523][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.005442136432975531, 'train/grad_norm_pre_clip': 0.7788451313972473, 'timing/data': 0.0003475029952824116, 'timing/model': 1.1342767349560745, 'learning_rate/qwen_vl_interface': 6.928766509472471e-07, 'learning_rate/action_model': 6.92876650947247e-06, 'epoch': 1.87}) +[2026-06-16 04:03:50,015][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 0.06507076323032379, 'train/grad_norm_pre_clip': 16.316129684448242, 'timing/data': 0.0003200299688614905, 'timing/model': 1.1326741789816879, 'learning_rate/qwen_vl_interface': 6.920185714335737e-07, 'learning_rate/action_model': 6.9201857143357375e-06, 'epoch': 1.87}) +[2026-06-16 04:03:54,558][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.0072677829302847385, 'train/grad_norm_pre_clip': 3.381422758102417, 'timing/data': 0.0003032960230484605, 'timing/model': 1.1328978869714774, 'learning_rate/qwen_vl_interface': 6.911623655338959e-07, 'learning_rate/action_model': 6.911623655338959e-06, 'epoch': 1.87}) +[2026-06-16 04:03:59,086][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.0016577424248680472, 'train/grad_norm_pre_clip': 0.5746263265609741, 'timing/data': 0.0005894609494134784, 'timing/model': 1.131810299004428, 'learning_rate/qwen_vl_interface': 6.903080336001666e-07, 'learning_rate/action_model': 6.903080336001666e-06, 'epoch': 1.87}) +[2026-06-16 04:04:03,626][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.006850596517324448, 'train/grad_norm_pre_clip': 3.4111721515655518, 'timing/data': 0.0005122600123286247, 'timing/model': 1.2848827919806354, 'learning_rate/qwen_vl_interface': 6.894555759835707e-07, 'learning_rate/action_model': 6.894555759835707e-06, 'epoch': 1.87}) +[2026-06-16 04:04:08,076][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.08495327830314636, 'train/grad_norm_pre_clip': 19.8502254486084, 'timing/data': 0.0004028240218758583, 'timing/model': 1.2105438569560647, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 1.87}) +[2026-06-16 04:04:12,454][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.10698781162500381, 'train/grad_norm_pre_clip': 7.646637439727783, 'timing/data': 0.0003209770075045526, 'timing/model': 1.133124608953949, 'learning_rate/qwen_vl_interface': 6.877562851026584e-07, 'learning_rate/action_model': 6.877562851026584e-06, 'epoch': 1.87}) +[2026-06-16 04:04:16,971][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.016918949782848358, 'train/grad_norm_pre_clip': 7.832826614379883, 'timing/data': 0.000631395960226655, 'timing/model': 1.1346420389600098, 'learning_rate/qwen_vl_interface': 6.869094525368572e-07, 'learning_rate/action_model': 6.8690945253685715e-06, 'epoch': 1.87}) +[2026-06-16 04:04:21,513][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.10691429674625397, 'train/grad_norm_pre_clip': 4.634570121765137, 'timing/data': 0.0003505889908410609, 'timing/model': 1.136771475954447, 'learning_rate/qwen_vl_interface': 6.860644956852168e-07, 'learning_rate/action_model': 6.860644956852168e-06, 'epoch': 1.87}) +[2026-06-16 04:04:26,042][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.0029048691503703594, 'train/grad_norm_pre_clip': 0.554344892501831, 'timing/data': 0.00033196399454027414, 'timing/model': 1.1336646450217813, 'learning_rate/qwen_vl_interface': 6.852214148950667e-07, 'learning_rate/action_model': 6.852214148950667e-06, 'epoch': 1.87}) +[2026-06-16 04:04:30,579][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.019004549831151962, 'train/grad_norm_pre_clip': 13.640076637268066, 'timing/data': 0.0003674289910122752, 'timing/model': 1.2844923459924757, 'learning_rate/qwen_vl_interface': 6.843802105129667e-07, 'learning_rate/action_model': 6.843802105129666e-06, 'epoch': 1.87}) +[2026-06-16 04:04:35,020][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.0018055171240121126, 'train/grad_norm_pre_clip': 0.5984590649604797, 'timing/data': 0.00036511599319055676, 'timing/model': 1.1853500080178492, 'learning_rate/qwen_vl_interface': 6.835408828847035e-07, 'learning_rate/action_model': 6.8354088288470356e-06, 'epoch': 1.88}) +[2026-06-16 04:04:39,414][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.0316266231238842, 'train/grad_norm_pre_clip': 21.35746955871582, 'timing/data': 0.00035582901909947395, 'timing/model': 1.1379967969842255, 'learning_rate/qwen_vl_interface': 6.827034323552929e-07, 'learning_rate/action_model': 6.827034323552929e-06, 'epoch': 1.88}) +[2026-06-16 04:04:43,875][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.002452349755913019, 'train/grad_norm_pre_clip': 0.8836508393287659, 'timing/data': 0.00039695302257314324, 'timing/model': 1.1391230069566518, 'learning_rate/qwen_vl_interface': 6.818678592689794e-07, 'learning_rate/action_model': 6.818678592689794e-06, 'epoch': 1.88}) +[2026-06-16 04:04:48,409][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.0008363305241800845, 'train/grad_norm_pre_clip': 0.23203590512275696, 'timing/data': 0.0004750140360556543, 'timing/model': 1.1359180560102686, 'learning_rate/qwen_vl_interface': 6.810341639692357e-07, 'learning_rate/action_model': 6.810341639692357e-06, 'epoch': 1.88}) +[2026-06-16 04:04:52,953][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 0.00034793169470503926, 'train/grad_norm_pre_clip': 0.052135828882455826, 'timing/data': 0.0003368170000612736, 'timing/model': 1.136004068015609, 'learning_rate/qwen_vl_interface': 6.802023467987631e-07, 'learning_rate/action_model': 6.802023467987631e-06, 'epoch': 1.88}) +[2026-06-16 04:04:57,512][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.006068886257708073, 'train/grad_norm_pre_clip': 1.3501296043395996, 'timing/data': 0.00034111301647499204, 'timing/model': 1.3084679610328749, 'learning_rate/qwen_vl_interface': 6.793724080994884e-07, 'learning_rate/action_model': 6.793724080994884e-06, 'epoch': 1.88}) +[2026-06-16 04:05:01,901][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.0016353202518075705, 'train/grad_norm_pre_clip': 1.2507656812667847, 'timing/data': 0.00032004498643800616, 'timing/model': 1.1390461540431716, 'learning_rate/qwen_vl_interface': 6.785443482125693e-07, 'learning_rate/action_model': 6.7854434821256915e-06, 'epoch': 1.88}) +[2026-06-16 04:05:06,455][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.007950644940137863, 'train/grad_norm_pre_clip': 2.0276107788085938, 'timing/data': 0.0003447660128585994, 'timing/model': 1.1326135119888932, 'learning_rate/qwen_vl_interface': 6.777181674783895e-07, 'learning_rate/action_model': 6.777181674783895e-06, 'epoch': 1.88}) +[2026-06-16 04:05:10,996][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.023601703345775604, 'train/grad_norm_pre_clip': 10.043680191040039, 'timing/data': 0.0003597299801185727, 'timing/model': 1.1386332360561937, 'learning_rate/qwen_vl_interface': 6.76893866236561e-07, 'learning_rate/action_model': 6.7689386623656094e-06, 'epoch': 1.88}) +[2026-06-16 04:05:15,537][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.20227345824241638, 'train/grad_norm_pre_clip': 16.63934898376465, 'timing/data': 0.00034325598971918225, 'timing/model': 1.1393945220042951, 'learning_rate/qwen_vl_interface': 6.76071444825923e-07, 'learning_rate/action_model': 6.76071444825923e-06, 'epoch': 1.88}) +[2026-06-16 04:05:20,090][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 0.0029135074000805616, 'train/grad_norm_pre_clip': 0.4886893928050995, 'timing/data': 0.0004061419749632478, 'timing/model': 1.1442970319767483, 'learning_rate/qwen_vl_interface': 6.752509035845413e-07, 'learning_rate/action_model': 6.7525090358454126e-06, 'epoch': 1.88}) +[2026-06-16 04:05:24,640][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.0002945891465060413, 'train/grad_norm_pre_clip': 0.057028159499168396, 'timing/data': 0.00034206698182970285, 'timing/model': 1.299386750033591, 'learning_rate/qwen_vl_interface': 6.744322428497095e-07, 'learning_rate/action_model': 6.744322428497095e-06, 'epoch': 1.88}) +[2026-06-16 04:05:29,032][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.2635853886604309, 'train/grad_norm_pre_clip': 12.176077842712402, 'timing/data': 0.00035159499384462833, 'timing/model': 1.1399717979948036, 'learning_rate/qwen_vl_interface': 6.736154629579487e-07, 'learning_rate/action_model': 6.736154629579487e-06, 'epoch': 1.88}) +[2026-06-16 04:05:33,590][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.03202515468001366, 'train/grad_norm_pre_clip': 6.968498229980469, 'timing/data': 0.0005630750092677772, 'timing/model': 1.133951326017268, 'learning_rate/qwen_vl_interface': 6.728005642450055e-07, 'learning_rate/action_model': 6.7280056424500554e-06, 'epoch': 1.88}) +[2026-06-16 04:05:38,095][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 0.010771047323942184, 'train/grad_norm_pre_clip': 4.931957244873047, 'timing/data': 0.0005219519953243434, 'timing/model': 1.1334615090163425, 'learning_rate/qwen_vl_interface': 6.719875470458553e-07, 'learning_rate/action_model': 6.719875470458553e-06, 'epoch': 1.88}) +[2026-06-16 04:05:42,652][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.11483778059482574, 'train/grad_norm_pre_clip': 8.83076286315918, 'timing/data': 0.0007819810416549444, 'timing/model': 1.1338255800073966, 'learning_rate/qwen_vl_interface': 6.711764116946972e-07, 'learning_rate/action_model': 6.7117641169469715e-06, 'epoch': 1.88}) +[2026-06-16 04:05:47,203][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.0004281447909306735, 'train/grad_norm_pre_clip': 0.05889071896672249, 'timing/data': 0.00033730699215084314, 'timing/model': 1.134525773988571, 'learning_rate/qwen_vl_interface': 6.703671585249587e-07, 'learning_rate/action_model': 6.703671585249586e-06, 'epoch': 1.88}) +[2026-06-16 04:05:51,762][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.0016752146184444427, 'train/grad_norm_pre_clip': 0.3487761318683624, 'timing/data': 0.0003260999801568687, 'timing/model': 1.2960755229578353, 'learning_rate/qwen_vl_interface': 6.69559787869294e-07, 'learning_rate/action_model': 6.69559787869294e-06, 'epoch': 1.88}) +[2026-06-16 04:05:56,228][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.13189387321472168, 'train/grad_norm_pre_clip': 20.959888458251953, 'timing/data': 0.0003319079987704754, 'timing/model': 1.2254343749955297, 'learning_rate/qwen_vl_interface': 6.68754300059582e-07, 'learning_rate/action_model': 6.687543000595819e-06, 'epoch': 1.88}) +[2026-06-16 04:06:00,691][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.1138341873884201, 'train/grad_norm_pre_clip': 7.532435894012451, 'timing/data': 0.0003399389679543674, 'timing/model': 1.1462843799963593, 'learning_rate/qwen_vl_interface': 6.67950695426929e-07, 'learning_rate/action_model': 6.679506954269291e-06, 'epoch': 1.88}) +[2026-06-16 04:06:05,228][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.00359905743971467, 'train/grad_norm_pre_clip': 1.2878538370132446, 'timing/data': 0.0003689369768835604, 'timing/model': 1.1322436010232195, 'learning_rate/qwen_vl_interface': 6.671489743016651e-07, 'learning_rate/action_model': 6.671489743016651e-06, 'epoch': 1.88}) +[2026-06-16 04:06:09,780][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.02269425243139267, 'train/grad_norm_pre_clip': 13.575117111206055, 'timing/data': 0.0003660799702629447, 'timing/model': 1.1356956799863838, 'learning_rate/qwen_vl_interface': 6.663491370133487e-07, 'learning_rate/action_model': 6.663491370133487e-06, 'epoch': 1.88}) +[2026-06-16 04:06:14,338][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.02705845609307289, 'train/grad_norm_pre_clip': 8.502826690673828, 'timing/data': 0.0003159240004606545, 'timing/model': 1.1372914590174332, 'learning_rate/qwen_vl_interface': 6.655511838907629e-07, 'learning_rate/action_model': 6.655511838907629e-06, 'epoch': 1.88}) +[2026-06-16 04:06:18,882][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.0901675596833229, 'train/grad_norm_pre_clip': 12.042737007141113, 'timing/data': 0.0005282479687593877, 'timing/model': 1.2962921970174648, 'learning_rate/qwen_vl_interface': 6.647551152619149e-07, 'learning_rate/action_model': 6.647551152619148e-06, 'epoch': 1.88}) +[2026-06-16 04:06:23,323][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.00016562720702495426, 'train/grad_norm_pre_clip': 0.020488258451223373, 'timing/data': 0.0005773750017397106, 'timing/model': 1.1786194969899952, 'learning_rate/qwen_vl_interface': 6.639609314540396e-07, 'learning_rate/action_model': 6.639609314540396e-06, 'epoch': 1.89}) +[2026-06-16 04:06:27,726][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.09987131506204605, 'train/grad_norm_pre_clip': 28.90679931640625, 'timing/data': 0.00040633499156683683, 'timing/model': 1.1373016610159539, 'learning_rate/qwen_vl_interface': 6.631686327935954e-07, 'learning_rate/action_model': 6.631686327935954e-06, 'epoch': 1.89}) +[2026-06-16 04:06:32,194][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.005576077848672867, 'train/grad_norm_pre_clip': 5.276876926422119, 'timing/data': 0.0003896330017596483, 'timing/model': 1.1337814020225778, 'learning_rate/qwen_vl_interface': 6.623782196062665e-07, 'learning_rate/action_model': 6.623782196062664e-06, 'epoch': 1.89}) +[2026-06-16 04:06:36,691][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.2587892711162567, 'train/grad_norm_pre_clip': 21.503305435180664, 'timing/data': 0.00036445300793275237, 'timing/model': 1.1311970539973117, 'learning_rate/qwen_vl_interface': 6.615896922169622e-07, 'learning_rate/action_model': 6.615896922169622e-06, 'epoch': 1.89}) +[2026-06-16 04:06:41,256][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.05929958075284958, 'train/grad_norm_pre_clip': 18.39997673034668, 'timing/data': 0.000420318974647671, 'timing/model': 1.1460532890050672, 'learning_rate/qwen_vl_interface': 6.608030509498156e-07, 'learning_rate/action_model': 6.608030509498156e-06, 'epoch': 1.89}) +[2026-06-16 04:06:45,805][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.0007163601694628596, 'train/grad_norm_pre_clip': 0.20540381968021393, 'timing/data': 0.00036804500268772244, 'timing/model': 1.2976963249966502, 'learning_rate/qwen_vl_interface': 6.600182961281865e-07, 'learning_rate/action_model': 6.600182961281864e-06, 'epoch': 1.89}) +[2026-06-16 04:06:50,191][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.0012972354888916016, 'train/grad_norm_pre_clip': 0.6046702861785889, 'timing/data': 0.00038704503094777465, 'timing/model': 1.139124935027212, 'learning_rate/qwen_vl_interface': 6.592354280746561e-07, 'learning_rate/action_model': 6.5923542807465595e-06, 'epoch': 1.89}) +[2026-06-16 04:06:54,757][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 0.0009008045308291912, 'train/grad_norm_pre_clip': 0.13998106122016907, 'timing/data': 0.000510144978761673, 'timing/model': 1.1367767520132475, 'learning_rate/qwen_vl_interface': 6.584544471110335e-07, 'learning_rate/action_model': 6.584544471110335e-06, 'epoch': 1.89}) +[2026-06-16 04:06:59,294][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.0011569453636184335, 'train/grad_norm_pre_clip': 0.7044566869735718, 'timing/data': 0.00042317097540944815, 'timing/model': 1.1371197800035588, 'learning_rate/qwen_vl_interface': 6.576753535583495e-07, 'learning_rate/action_model': 6.576753535583495e-06, 'epoch': 1.89}) +[2026-06-16 04:07:03,856][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.015574107877910137, 'train/grad_norm_pre_clip': 7.338186740875244, 'timing/data': 0.00036056601675227284, 'timing/model': 1.1377456419868395, 'learning_rate/qwen_vl_interface': 6.5689814773686e-07, 'learning_rate/action_model': 6.568981477368599e-06, 'epoch': 1.89}) +[2026-06-16 04:07:08,420][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.000407818122766912, 'train/grad_norm_pre_clip': 0.0736529603600502, 'timing/data': 0.00041400501504540443, 'timing/model': 1.1420427029952407, 'learning_rate/qwen_vl_interface': 6.56122829966046e-07, 'learning_rate/action_model': 6.56122829966046e-06, 'epoch': 1.89}) +[2026-06-16 04:11:32,272][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.0008007250726222992, 'train/grad_norm_pre_clip': 0.1595582664012909, 'eval/loss': 0.040021803230047226, 'eval/flap_precision': 0.9666666666666667, 'eval/flap_recall': 0.9845679012345679, 'eval/flap_f1': 0.9755351681957187, 'eval/flap_rate_model': 0.18333333333333332, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9755351681957187, 'timing/data': 0.0004648190224543214, 'timing/model': 1.2974735309835523, 'learning_rate/qwen_vl_interface': 6.553494005646101e-07, 'learning_rate/action_model': 6.553494005646101e-06, 'epoch': 1.89}) +[2026-06-16 04:11:32,272][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4600_state +[2026-06-16 04:11:32,275][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 04:11:35,860][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4600_state/model.safetensors +[2026-06-16 04:11:45,910][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4600_state/optimizer.bin +[2026-06-16 04:11:45,911][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4600_state/sampler.bin +[2026-06-16 04:11:45,911][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4600_state/sampler_1.bin +[2026-06-16 04:11:45,912][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4600_state/random_states_0.pkl +[2026-06-16 04:11:45,962][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4600 +[2026-06-16 04:11:45,962][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 04:11:45,983][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 04:11:50,411][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.16238000988960266, 'train/grad_norm_pre_clip': 18.987823486328125, 'timing/data': 0.0003429320058785379, 'timing/model': 1.1234247490065172, 'learning_rate/qwen_vl_interface': 6.545778598504807e-07, 'learning_rate/action_model': 6.545778598504807e-06, 'epoch': 1.89}) +[2026-06-16 04:11:54,883][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.20608949661254883, 'train/grad_norm_pre_clip': 22.75417137145996, 'timing/data': 0.0003284109989181161, 'timing/model': 1.2558768540038727, 'learning_rate/qwen_vl_interface': 6.538082081408093e-07, 'learning_rate/action_model': 6.538082081408093e-06, 'epoch': 1.89}) +[2026-06-16 04:11:59,209][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 0.11163423955440521, 'train/grad_norm_pre_clip': 15.177530288696289, 'timing/data': 0.0003525940119288862, 'timing/model': 1.1189851259696297, 'learning_rate/qwen_vl_interface': 6.530404457519697e-07, 'learning_rate/action_model': 6.530404457519697e-06, 'epoch': 1.89}) +[2026-06-16 04:12:03,693][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.1056353747844696, 'train/grad_norm_pre_clip': 6.980866432189941, 'timing/data': 0.0003557379823178053, 'timing/model': 1.1226656920043752, 'learning_rate/qwen_vl_interface': 6.522745729995614e-07, 'learning_rate/action_model': 6.522745729995613e-06, 'epoch': 1.89}) +[2026-06-16 04:12:08,186][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.14071179926395416, 'train/grad_norm_pre_clip': 13.930838584899902, 'timing/data': 0.0003317659720778465, 'timing/model': 1.1256179620395415, 'learning_rate/qwen_vl_interface': 6.515105901984037e-07, 'learning_rate/action_model': 6.515105901984037e-06, 'epoch': 1.89}) +[2026-06-16 04:12:12,669][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.0014194526011124253, 'train/grad_norm_pre_clip': 0.178771510720253, 'timing/data': 0.000343200983479619, 'timing/model': 1.125128678046167, 'learning_rate/qwen_vl_interface': 6.50748497662543e-07, 'learning_rate/action_model': 6.507484976625429e-06, 'epoch': 1.89}) +[2026-06-16 04:12:17,151][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.14998559653759003, 'train/grad_norm_pre_clip': 13.92544937133789, 'timing/data': 0.00036673102295026183, 'timing/model': 1.273482239979785, 'learning_rate/qwen_vl_interface': 6.499882957052459e-07, 'learning_rate/action_model': 6.4998829570524586e-06, 'epoch': 1.89}) +[2026-06-16 04:12:21,598][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.16621655225753784, 'train/grad_norm_pre_clip': 38.566349029541016, 'timing/data': 0.0005774129531346262, 'timing/model': 1.219369366997853, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 1.89}) +[2026-06-16 04:12:25,960][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.09615623205900192, 'train/grad_norm_pre_clip': 22.953672409057617, 'timing/data': 0.000342942017596215, 'timing/model': 1.1379980960045941, 'learning_rate/qwen_vl_interface': 6.484735647755259e-07, 'learning_rate/action_model': 6.48473564775526e-06, 'epoch': 1.89}) +[2026-06-16 04:12:30,467][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.002937874523922801, 'train/grad_norm_pre_clip': 1.924074649810791, 'timing/data': 0.0005993249942548573, 'timing/model': 1.1308731330209412, 'learning_rate/qwen_vl_interface': 6.477190364257524e-07, 'learning_rate/action_model': 6.477190364257524e-06, 'epoch': 1.89}) +[2026-06-16 04:12:34,993][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.2475982904434204, 'train/grad_norm_pre_clip': 26.931575775146484, 'timing/data': 0.0003683079848997295, 'timing/model': 1.1378372209728695, 'learning_rate/qwen_vl_interface': 6.469663998998393e-07, 'learning_rate/action_model': 6.469663998998393e-06, 'epoch': 1.89}) +[2026-06-16 04:12:39,503][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.018136756494641304, 'train/grad_norm_pre_clip': 15.900006294250488, 'timing/data': 0.00035122997360304, 'timing/model': 1.1239768200321123, 'learning_rate/qwen_vl_interface': 6.462156555071681e-07, 'learning_rate/action_model': 6.46215655507168e-06, 'epoch': 1.89}) +[2026-06-16 04:12:44,008][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 0.34072381258010864, 'train/grad_norm_pre_clip': 34.62394714355469, 'timing/data': 0.00036859099054709077, 'timing/model': 1.2772480730200186, 'learning_rate/qwen_vl_interface': 6.454668035563401e-07, 'learning_rate/action_model': 6.454668035563401e-06, 'epoch': 1.9}) +[2026-06-16 04:12:48,414][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.15175504982471466, 'train/grad_norm_pre_clip': 31.98129653930664, 'timing/data': 0.00036231300327926874, 'timing/model': 1.1892338990001008, 'learning_rate/qwen_vl_interface': 6.447198443551809e-07, 'learning_rate/action_model': 6.4471984435518075e-06, 'epoch': 1.9}) +[2026-06-16 04:12:52,776][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.0010441516060382128, 'train/grad_norm_pre_clip': 0.2527676820755005, 'timing/data': 0.00036187696969136596, 'timing/model': 1.1326552429818548, 'learning_rate/qwen_vl_interface': 6.439747782107373e-07, 'learning_rate/action_model': 6.4397477821073735e-06, 'epoch': 1.9}) +[2026-06-16 04:12:57,239][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 0.06820138543844223, 'train/grad_norm_pre_clip': 14.58596134185791, 'timing/data': 0.0007107100100256503, 'timing/model': 1.1268799160025083, 'learning_rate/qwen_vl_interface': 6.432316054292779e-07, 'learning_rate/action_model': 6.43231605429278e-06, 'epoch': 1.9}) +[2026-06-16 04:13:01,726][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.06371015310287476, 'train/grad_norm_pre_clip': 17.073545455932617, 'timing/data': 0.00035664503229781985, 'timing/model': 1.1264944929862395, 'learning_rate/qwen_vl_interface': 6.424903263162924e-07, 'learning_rate/action_model': 6.424903263162923e-06, 'epoch': 1.9}) +[2026-06-16 04:13:06,246][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.0019763996824622154, 'train/grad_norm_pre_clip': 0.5931801795959473, 'timing/data': 0.0003246529959142208, 'timing/model': 1.128822031954769, 'learning_rate/qwen_vl_interface': 6.417509411764924e-07, 'learning_rate/action_model': 6.4175094117649244e-06, 'epoch': 1.9}) +[2026-06-16 04:13:10,766][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.0010375466663390398, 'train/grad_norm_pre_clip': 0.18938930332660675, 'timing/data': 0.000366593012586236, 'timing/model': 1.290349667018745, 'learning_rate/qwen_vl_interface': 6.410134503138122e-07, 'learning_rate/action_model': 6.4101345031381224e-06, 'epoch': 1.9}) +[2026-06-16 04:13:15,147][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.21643124520778656, 'train/grad_norm_pre_clip': 15.55535888671875, 'timing/data': 0.00048731197603046894, 'timing/model': 1.132728020020295, 'learning_rate/qwen_vl_interface': 6.402778540314054e-07, 'learning_rate/action_model': 6.402778540314054e-06, 'epoch': 1.9}) +[2026-06-16 04:13:19,685][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.002146095270290971, 'train/grad_norm_pre_clip': 0.9020654559135437, 'timing/data': 0.0003380609559826553, 'timing/model': 1.1297561079845764, 'learning_rate/qwen_vl_interface': 6.39544152631649e-07, 'learning_rate/action_model': 6.3954415263164895e-06, 'epoch': 1.9}) +[2026-06-16 04:13:24,214][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 0.0008952947682701051, 'train/grad_norm_pre_clip': 0.20502759516239166, 'timing/data': 0.0003402759903110564, 'timing/model': 1.1355853650020435, 'learning_rate/qwen_vl_interface': 6.388123464161388e-07, 'learning_rate/action_model': 6.388123464161389e-06, 'epoch': 1.9}) +[2026-06-16 04:13:28,728][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.18012288212776184, 'train/grad_norm_pre_clip': 18.97042465209961, 'timing/data': 0.00032142200507223606, 'timing/model': 1.1282475260086358, 'learning_rate/qwen_vl_interface': 6.380824356856936e-07, 'learning_rate/action_model': 6.380824356856936e-06, 'epoch': 1.9}) +[2026-06-16 04:13:33,266][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.00042467177263461053, 'train/grad_norm_pre_clip': 0.06581470370292664, 'timing/data': 0.0003553779679350555, 'timing/model': 1.1438717700075358, 'learning_rate/qwen_vl_interface': 6.373544207403513e-07, 'learning_rate/action_model': 6.373544207403513e-06, 'epoch': 1.9}) +[2026-06-16 04:13:37,804][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 0.19999265670776367, 'train/grad_norm_pre_clip': 19.807802200317383, 'timing/data': 0.00042531301733106375, 'timing/model': 1.2961148289614357, 'learning_rate/qwen_vl_interface': 6.366283018793723e-07, 'learning_rate/action_model': 6.366283018793723e-06, 'epoch': 1.9}) +[2026-06-16 04:13:42,185][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.07342595607042313, 'train/grad_norm_pre_clip': 23.629295349121094, 'timing/data': 0.0003479879815131426, 'timing/model': 1.1334563250420615, 'learning_rate/qwen_vl_interface': 6.359040794012358e-07, 'learning_rate/action_model': 6.359040794012358e-06, 'epoch': 1.9}) +[2026-06-16 04:13:46,713][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.00827744696289301, 'train/grad_norm_pre_clip': 2.7148187160491943, 'timing/data': 0.0003383990260772407, 'timing/model': 1.1304644820047542, 'learning_rate/qwen_vl_interface': 6.351817536036431e-07, 'learning_rate/action_model': 6.351817536036431e-06, 'epoch': 1.9}) +[2026-06-16 04:13:51,199][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.0061419568955898285, 'train/grad_norm_pre_clip': 2.1369211673736572, 'timing/data': 0.00034360698191449046, 'timing/model': 1.1318995239562355, 'learning_rate/qwen_vl_interface': 6.344613247835135e-07, 'learning_rate/action_model': 6.344613247835135e-06, 'epoch': 1.9}) +[2026-06-16 04:13:55,736][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.0019328594207763672, 'train/grad_norm_pre_clip': 0.8913452625274658, 'timing/data': 0.0003259209915995598, 'timing/model': 1.1349425060325302, 'learning_rate/qwen_vl_interface': 6.337427932369898e-07, 'learning_rate/action_model': 6.337427932369898e-06, 'epoch': 1.9}) +[2026-06-16 04:14:00,266][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.0030920132994651794, 'train/grad_norm_pre_clip': 0.5027357935905457, 'timing/data': 0.000351644994225353, 'timing/model': 1.1361972260056064, 'learning_rate/qwen_vl_interface': 6.330261592594328e-07, 'learning_rate/action_model': 6.330261592594327e-06, 'epoch': 1.9}) +[2026-06-16 04:14:04,807][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.10373793542385101, 'train/grad_norm_pre_clip': 7.341602325439453, 'timing/data': 0.0003581299679353833, 'timing/model': 1.2936661770218052, 'learning_rate/qwen_vl_interface': 6.323114231454229e-07, 'learning_rate/action_model': 6.323114231454229e-06, 'epoch': 1.9}) +[2026-06-16 04:14:09,270][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.02517642080783844, 'train/grad_norm_pre_clip': 7.660770893096924, 'timing/data': 0.0003707370487973094, 'timing/model': 1.2297962820157409, 'learning_rate/qwen_vl_interface': 6.315985851887618e-07, 'learning_rate/action_model': 6.315985851887618e-06, 'epoch': 1.9}) +[2026-06-16 04:14:13,646][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.006082843989133835, 'train/grad_norm_pre_clip': 1.7613537311553955, 'timing/data': 0.0003339520189911127, 'timing/model': 1.1337463379604742, 'learning_rate/qwen_vl_interface': 6.308876456824697e-07, 'learning_rate/action_model': 6.3088764568246975e-06, 'epoch': 1.9}) +[2026-06-16 04:14:18,159][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.0030074375681579113, 'train/grad_norm_pre_clip': 0.35579320788383484, 'timing/data': 0.0003271939931437373, 'timing/model': 1.140965118946042, 'learning_rate/qwen_vl_interface': 6.301786049187875e-07, 'learning_rate/action_model': 6.3017860491878745e-06, 'epoch': 1.9}) +[2026-06-16 04:14:22,704][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.0008002225076779723, 'train/grad_norm_pre_clip': 0.17887380719184875, 'timing/data': 0.00033895199885591865, 'timing/model': 1.1352935849572532, 'learning_rate/qwen_vl_interface': 6.294714631891755e-07, 'learning_rate/action_model': 6.294714631891754e-06, 'epoch': 1.9}) +[2026-06-16 04:14:27,234][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.13587971031665802, 'train/grad_norm_pre_clip': 12.511205673217773, 'timing/data': 0.0003435949911363423, 'timing/model': 1.134414289961569, 'learning_rate/qwen_vl_interface': 6.287662207843115e-07, 'learning_rate/action_model': 6.287662207843114e-06, 'epoch': 1.9}) +[2026-06-16 04:14:31,767][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.22663244605064392, 'train/grad_norm_pre_clip': 27.11168670654297, 'timing/data': 0.0007055349997244775, 'timing/model': 1.2816950990236364, 'learning_rate/qwen_vl_interface': 6.280628779940956e-07, 'learning_rate/action_model': 6.280628779940956e-06, 'epoch': 1.9}) +[2026-06-16 04:14:36,194][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.00017661652236711234, 'train/grad_norm_pre_clip': 0.031236067414283752, 'timing/data': 0.00032532401382923126, 'timing/model': 1.1885081680375151, 'learning_rate/qwen_vl_interface': 6.273614351076445e-07, 'learning_rate/action_model': 6.273614351076445e-06, 'epoch': 1.91}) +[2026-06-16 04:14:40,562][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.3238949179649353, 'train/grad_norm_pre_clip': 24.909873962402344, 'timing/data': 0.00031943898648023605, 'timing/model': 1.1316552099888213, 'learning_rate/qwen_vl_interface': 6.266618924132951e-07, 'learning_rate/action_model': 6.266618924132951e-06, 'epoch': 1.91}) +[2026-06-16 04:14:45,028][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.00032855375320650637, 'train/grad_norm_pre_clip': 0.04421977698802948, 'timing/data': 0.00037616799818351865, 'timing/model': 1.1374400219647214, 'learning_rate/qwen_vl_interface': 6.25964250198603e-07, 'learning_rate/action_model': 6.2596425019860294e-06, 'epoch': 1.91}) +[2026-06-16 04:14:49,534][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.09075828641653061, 'train/grad_norm_pre_clip': 8.180487632751465, 'timing/data': 0.00037772103678435087, 'timing/model': 1.130639779032208, 'learning_rate/qwen_vl_interface': 6.25268508750342e-07, 'learning_rate/action_model': 6.25268508750342e-06, 'epoch': 1.91}) +[2026-06-16 04:14:54,104][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.06378234922885895, 'train/grad_norm_pre_clip': 12.668272018432617, 'timing/data': 0.00031792197842150927, 'timing/model': 1.1462768699857406, 'learning_rate/qwen_vl_interface': 6.245746683545062e-07, 'learning_rate/action_model': 6.245746683545062e-06, 'epoch': 1.91}) +[2026-06-16 04:14:58,653][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.09189103543758392, 'train/grad_norm_pre_clip': 8.772849082946777, 'timing/data': 0.00036495603853836656, 'timing/model': 1.3015372880036011, 'learning_rate/qwen_vl_interface': 6.238827292963063e-07, 'learning_rate/action_model': 6.238827292963063e-06, 'epoch': 1.91}) +[2026-06-16 04:15:03,029][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 0.014807070605456829, 'train/grad_norm_pre_clip': 8.364151954650879, 'timing/data': 0.0003989660181105137, 'timing/model': 1.1309450209955685, 'learning_rate/qwen_vl_interface': 6.231926918601731e-07, 'learning_rate/action_model': 6.2319269186017305e-06, 'epoch': 1.91}) +[2026-06-16 04:15:07,559][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 0.0011433372274041176, 'train/grad_norm_pre_clip': 0.3562992215156555, 'timing/data': 0.0003236399497836828, 'timing/model': 1.1337419369956478, 'learning_rate/qwen_vl_interface': 6.22504556329754e-07, 'learning_rate/action_model': 6.22504556329754e-06, 'epoch': 1.91}) +[2026-06-16 04:15:12,091][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.0005490431794896722, 'train/grad_norm_pre_clip': 0.12311173230409622, 'timing/data': 0.00035390997072681785, 'timing/model': 1.1321139439824037, 'learning_rate/qwen_vl_interface': 6.218183229879164e-07, 'learning_rate/action_model': 6.218183229879164e-06, 'epoch': 1.91}) +[2026-06-16 04:15:16,652][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.08672463893890381, 'train/grad_norm_pre_clip': 14.143412590026855, 'timing/data': 0.0003431179793551564, 'timing/model': 1.131655581004452, 'learning_rate/qwen_vl_interface': 6.211339921167447e-07, 'learning_rate/action_model': 6.211339921167446e-06, 'epoch': 1.91}) +[2026-06-16 04:15:21,202][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.20056764781475067, 'train/grad_norm_pre_clip': 17.65622329711914, 'timing/data': 0.00035093299811705947, 'timing/model': 1.1380083459662274, 'learning_rate/qwen_vl_interface': 6.204515639975415e-07, 'learning_rate/action_model': 6.204515639975414e-06, 'epoch': 1.91}) +[2026-06-16 04:15:25,748][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.1280682235956192, 'train/grad_norm_pre_clip': 18.67336654663086, 'timing/data': 0.0007362670148722827, 'timing/model': 1.2970655029639602, 'learning_rate/qwen_vl_interface': 6.197710389108262e-07, 'learning_rate/action_model': 6.197710389108262e-06, 'epoch': 1.91}) +[2026-06-16 04:15:30,141][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.0007026721723377705, 'train/grad_norm_pre_clip': 0.07776929438114166, 'timing/data': 0.0003301770193502307, 'timing/model': 1.1342990839621052, 'learning_rate/qwen_vl_interface': 6.190924171363378e-07, 'learning_rate/action_model': 6.190924171363378e-06, 'epoch': 1.91}) +[2026-06-16 04:15:34,686][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.15647560358047485, 'train/grad_norm_pre_clip': 22.931617736816406, 'timing/data': 0.0003304900019429624, 'timing/model': 1.1371486730058677, 'learning_rate/qwen_vl_interface': 6.184156989530326e-07, 'learning_rate/action_model': 6.184156989530325e-06, 'epoch': 1.91}) +[2026-06-16 04:15:39,193][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.0021723012905567884, 'train/grad_norm_pre_clip': 0.5533419847488403, 'timing/data': 0.0003516729921102524, 'timing/model': 1.1408203819883056, 'learning_rate/qwen_vl_interface': 6.17740884639083e-07, 'learning_rate/action_model': 6.177408846390829e-06, 'epoch': 1.91}) +[2026-06-16 04:15:43,733][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.0009256441262550652, 'train/grad_norm_pre_clip': 0.135712668299675, 'timing/data': 0.0004974960465915501, 'timing/model': 1.1332159300218336, 'learning_rate/qwen_vl_interface': 6.170679744718798e-07, 'learning_rate/action_model': 6.1706797447187975e-06, 'epoch': 1.91}) +[2026-06-16 04:15:48,278][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.2891954481601715, 'train/grad_norm_pre_clip': 32.285560607910156, 'timing/data': 0.00036955997347831726, 'timing/model': 1.1338500590063632, 'learning_rate/qwen_vl_interface': 6.163969687280311e-07, 'learning_rate/action_model': 6.163969687280311e-06, 'epoch': 1.91}) +[2026-06-16 04:15:52,810][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.07228776067495346, 'train/grad_norm_pre_clip': 28.804351806640625, 'timing/data': 0.0005339829949662089, 'timing/model': 1.2824799860245548, 'learning_rate/qwen_vl_interface': 6.157278676833627e-07, 'learning_rate/action_model': 6.157278676833627e-06, 'epoch': 1.91}) +[2026-06-16 04:15:57,312][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.0004551043384708464, 'train/grad_norm_pre_clip': 0.05904925614595413, 'timing/data': 0.00037819298449903727, 'timing/model': 1.2345076219644397, 'learning_rate/qwen_vl_interface': 6.150606716129155e-07, 'learning_rate/action_model': 6.150606716129154e-06, 'epoch': 1.91}) +[2026-06-16 04:16:01,697][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.0002408241416560486, 'train/grad_norm_pre_clip': 0.03551900014281273, 'timing/data': 0.0005913180066272616, 'timing/model': 1.1319530989858322, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 1.91}) +[2026-06-16 04:16:06,254][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 0.0023056487552821636, 'train/grad_norm_pre_clip': 0.8204377293586731, 'timing/data': 0.0003194689634256065, 'timing/model': 1.1392234939849004, 'learning_rate/qwen_vl_interface': 6.137319954909407e-07, 'learning_rate/action_model': 6.137319954909406e-06, 'epoch': 1.91}) +[2026-06-16 04:16:10,805][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.10331771522760391, 'train/grad_norm_pre_clip': 29.165193557739258, 'timing/data': 0.0003181209904141724, 'timing/model': 1.131715969007928, 'learning_rate/qwen_vl_interface': 6.130705159855815e-07, 'learning_rate/action_model': 6.130705159855814e-06, 'epoch': 1.91}) +[2026-06-16 04:16:15,382][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.003821699181571603, 'train/grad_norm_pre_clip': 1.7350002527236938, 'timing/data': 0.0003881970187649131, 'timing/model': 1.1412759360391647, 'learning_rate/qwen_vl_interface': 6.124109425467805e-07, 'learning_rate/action_model': 6.124109425467805e-06, 'epoch': 1.91}) +[2026-06-16 04:16:19,924][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.038200493901968, 'train/grad_norm_pre_clip': 16.832862854003906, 'timing/data': 0.00035107502480968833, 'timing/model': 1.2797417739639059, 'learning_rate/qwen_vl_interface': 6.117532754456642e-07, 'learning_rate/action_model': 6.117532754456642e-06, 'epoch': 1.91}) +[2026-06-16 04:16:24,363][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.046359311789274216, 'train/grad_norm_pre_clip': 28.748676300048828, 'timing/data': 0.0003526749787852168, 'timing/model': 1.1930745409918018, 'learning_rate/qwen_vl_interface': 6.110975149525746e-07, 'learning_rate/action_model': 6.110975149525746e-06, 'epoch': 1.92}) +[2026-06-16 04:16:28,759][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.07202218472957611, 'train/grad_norm_pre_clip': 27.599763870239258, 'timing/data': 0.00034300703555345535, 'timing/model': 1.136427138990257, 'learning_rate/qwen_vl_interface': 6.104436613370692e-07, 'learning_rate/action_model': 6.104436613370692e-06, 'epoch': 1.92}) +[2026-06-16 04:16:33,225][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.04932255297899246, 'train/grad_norm_pre_clip': 14.558296203613281, 'timing/data': 0.0003854779643006623, 'timing/model': 1.138437816989608, 'learning_rate/qwen_vl_interface': 6.09791714867924e-07, 'learning_rate/action_model': 6.0979171486792396e-06, 'epoch': 1.92}) +[2026-06-16 04:16:37,727][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.07473798096179962, 'train/grad_norm_pre_clip': 15.771620750427246, 'timing/data': 0.00035036797635257244, 'timing/model': 1.1345981690101326, 'learning_rate/qwen_vl_interface': 6.091416758131287e-07, 'learning_rate/action_model': 6.091416758131286e-06, 'epoch': 1.92}) +[2026-06-16 04:16:42,300][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.016352077946066856, 'train/grad_norm_pre_clip': 6.575841903686523, 'timing/data': 0.0003242000238969922, 'timing/model': 1.1380375170265324, 'learning_rate/qwen_vl_interface': 6.084935444398895e-07, 'learning_rate/action_model': 6.084935444398894e-06, 'epoch': 1.92}) +[2026-06-16 04:16:46,904][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.022246304899454117, 'train/grad_norm_pre_clip': 21.025903701782227, 'timing/data': 0.00044066604459658265, 'timing/model': 1.3080641300184652, 'learning_rate/qwen_vl_interface': 6.078473210146294e-07, 'learning_rate/action_model': 6.078473210146294e-06, 'epoch': 1.92}) +[2026-06-16 04:16:51,298][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 0.001531975925900042, 'train/grad_norm_pre_clip': 0.3186245262622833, 'timing/data': 0.0003385969903320074, 'timing/model': 1.137201770034153, 'learning_rate/qwen_vl_interface': 6.072030058029862e-07, 'learning_rate/action_model': 6.072030058029861e-06, 'epoch': 1.92}) +[2026-06-16 04:16:55,855][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 0.0022548227570950985, 'train/grad_norm_pre_clip': 0.9107038378715515, 'timing/data': 0.00033919198904186487, 'timing/model': 1.1328381639905274, 'learning_rate/qwen_vl_interface': 6.065605990698136e-07, 'learning_rate/action_model': 6.065605990698136e-06, 'epoch': 1.92}) +[2026-06-16 04:17:00,413][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.07856664806604385, 'train/grad_norm_pre_clip': 63.406253814697266, 'timing/data': 0.00034539803164079785, 'timing/model': 1.1392832139972597, 'learning_rate/qwen_vl_interface': 6.059201010791811e-07, 'learning_rate/action_model': 6.059201010791811e-06, 'epoch': 1.92}) +[2026-06-16 04:17:04,981][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.10407643020153046, 'train/grad_norm_pre_clip': 14.25716781616211, 'timing/data': 0.0003788929898291826, 'timing/model': 1.1521489460137673, 'learning_rate/qwen_vl_interface': 6.052815120943721e-07, 'learning_rate/action_model': 6.052815120943721e-06, 'epoch': 1.92}) +[2026-06-16 04:17:09,553][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.07783772051334381, 'train/grad_norm_pre_clip': 25.773569107055664, 'timing/data': 0.00039482302963733673, 'timing/model': 1.134119049005676, 'learning_rate/qwen_vl_interface': 6.046448323778874e-07, 'learning_rate/action_model': 6.046448323778874e-06, 'epoch': 1.92}) +[2026-06-16 04:17:14,094][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.0005517212557606399, 'train/grad_norm_pre_clip': 0.1248731017112732, 'timing/data': 0.0003639959613792598, 'timing/model': 1.291378662979696, 'learning_rate/qwen_vl_interface': 6.040100621914419e-07, 'learning_rate/action_model': 6.040100621914418e-06, 'epoch': 1.92}) +[2026-06-16 04:17:18,481][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.09917104989290237, 'train/grad_norm_pre_clip': 11.640979766845703, 'timing/data': 0.0003253409522585571, 'timing/model': 1.13141207001172, 'learning_rate/qwen_vl_interface': 6.033772017959652e-07, 'learning_rate/action_model': 6.033772017959652e-06, 'epoch': 1.92}) +[2026-06-16 04:17:23,033][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.1498875617980957, 'train/grad_norm_pre_clip': 9.87569522857666, 'timing/data': 0.00036778696812689304, 'timing/model': 1.1353751660208218, 'learning_rate/qwen_vl_interface': 6.027462514516022e-07, 'learning_rate/action_model': 6.0274625145160214e-06, 'epoch': 1.92}) +[2026-06-16 04:17:27,538][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.0021671708673238754, 'train/grad_norm_pre_clip': 0.7280252575874329, 'timing/data': 0.00037782202707603574, 'timing/model': 1.1321581119555049, 'learning_rate/qwen_vl_interface': 6.021172114177139e-07, 'learning_rate/action_model': 6.021172114177139e-06, 'epoch': 1.92}) +[2026-06-16 04:17:32,116][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.10564298927783966, 'train/grad_norm_pre_clip': 16.322677612304688, 'timing/data': 0.00033475301461294293, 'timing/model': 1.1461331879836507, 'learning_rate/qwen_vl_interface': 6.014900819528734e-07, 'learning_rate/action_model': 6.0149008195287335e-06, 'epoch': 1.92}) +[2026-06-16 04:17:36,683][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.0009063644683919847, 'train/grad_norm_pre_clip': 0.2957139015197754, 'timing/data': 0.0005393929895944893, 'timing/model': 1.1387271200073883, 'learning_rate/qwen_vl_interface': 6.008648633148707e-07, 'learning_rate/action_model': 6.008648633148707e-06, 'epoch': 1.92}) +[2026-06-16 04:17:41,242][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.008173765614628792, 'train/grad_norm_pre_clip': 2.3992807865142822, 'timing/data': 0.0003277640207670629, 'timing/model': 1.3031092900200747, 'learning_rate/qwen_vl_interface': 6.002415557607097e-07, 'learning_rate/action_model': 6.002415557607096e-06, 'epoch': 1.92}) +[2026-06-16 04:17:45,730][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.11967556923627853, 'train/grad_norm_pre_clip': 16.314411163330078, 'timing/data': 0.00037456501740962267, 'timing/model': 1.235293867008295, 'learning_rate/qwen_vl_interface': 5.996201595466088e-07, 'learning_rate/action_model': 5.996201595466088e-06, 'epoch': 1.92}) +[2026-06-16 04:17:50,121][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.025554940104484558, 'train/grad_norm_pre_clip': 23.029129028320312, 'timing/data': 0.00036129297222942114, 'timing/model': 1.1432887759874575, 'learning_rate/qwen_vl_interface': 5.990006749279998e-07, 'learning_rate/action_model': 5.990006749279998e-06, 'epoch': 1.92}) +[2026-06-16 04:17:54,649][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.024291515350341797, 'train/grad_norm_pre_clip': 9.382547378540039, 'timing/data': 0.00034824199974536896, 'timing/model': 1.1395401809713803, 'learning_rate/qwen_vl_interface': 5.983831021595296e-07, 'learning_rate/action_model': 5.9838310215952956e-06, 'epoch': 1.92}) +[2026-06-16 04:17:59,210][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.002894691191613674, 'train/grad_norm_pre_clip': 1.3486747741699219, 'timing/data': 0.0003641929943114519, 'timing/model': 1.137135682045482, 'learning_rate/qwen_vl_interface': 5.977674414950595e-07, 'learning_rate/action_model': 5.9776744149505955e-06, 'epoch': 1.92}) +[2026-06-16 04:18:03,769][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.0002539076376706362, 'train/grad_norm_pre_clip': 0.02933008223772049, 'timing/data': 0.0003447390045039356, 'timing/model': 1.1368142209830694, 'learning_rate/qwen_vl_interface': 5.971536931876635e-07, 'learning_rate/action_model': 5.971536931876634e-06, 'epoch': 1.92}) +[2026-06-16 04:18:08,317][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.2805638313293457, 'train/grad_norm_pre_clip': 11.10009765625, 'timing/data': 0.00031819200376048684, 'timing/model': 1.2958966320147738, 'learning_rate/qwen_vl_interface': 5.965418574896315e-07, 'learning_rate/action_model': 5.9654185748963145e-06, 'epoch': 1.92}) +[2026-06-16 04:18:12,752][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.00027817205409519374, 'train/grad_norm_pre_clip': 0.12739840149879456, 'timing/data': 0.00036026304587721825, 'timing/model': 1.1863511410192586, 'learning_rate/qwen_vl_interface': 5.959319346524646e-07, 'learning_rate/action_model': 5.959319346524646e-06, 'epoch': 1.93}) +[2026-06-16 04:18:17,137][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.06607750058174133, 'train/grad_norm_pre_clip': 21.867050170898438, 'timing/data': 0.0003685950068756938, 'timing/model': 1.137111549032852, 'learning_rate/qwen_vl_interface': 5.953239249268804e-07, 'learning_rate/action_model': 5.953239249268804e-06, 'epoch': 1.93}) +[2026-06-16 04:18:21,638][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.23127081990242004, 'train/grad_norm_pre_clip': 24.487375259399414, 'timing/data': 0.00032692798413336277, 'timing/model': 1.1342870880034752, 'learning_rate/qwen_vl_interface': 5.947178285628081e-07, 'learning_rate/action_model': 5.94717828562808e-06, 'epoch': 1.93}) +[2026-06-16 04:18:26,158][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.08285925537347794, 'train/grad_norm_pre_clip': 9.871505737304688, 'timing/data': 0.0003439229913055897, 'timing/model': 1.1365904349950142, 'learning_rate/qwen_vl_interface': 5.941136458093907e-07, 'learning_rate/action_model': 5.941136458093907e-06, 'epoch': 1.93}) +[2026-06-16 04:18:30,704][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.017347661778330803, 'train/grad_norm_pre_clip': 8.645750045776367, 'timing/data': 0.00037416903069242835, 'timing/model': 1.1356682250043377, 'learning_rate/qwen_vl_interface': 5.935113769149855e-07, 'learning_rate/action_model': 5.935113769149856e-06, 'epoch': 1.93}) +[2026-06-16 04:18:35,241][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.11368507146835327, 'train/grad_norm_pre_clip': 10.665325164794922, 'timing/data': 0.00036491500213742256, 'timing/model': 1.2914737450191751, 'learning_rate/qwen_vl_interface': 5.92911022127163e-07, 'learning_rate/action_model': 5.929110221271629e-06, 'epoch': 1.93}) +[2026-06-16 04:18:39,631][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.011976389214396477, 'train/grad_norm_pre_clip': 6.466847896575928, 'timing/data': 0.0003549069515429437, 'timing/model': 1.134188980970066, 'learning_rate/qwen_vl_interface': 5.92312581692705e-07, 'learning_rate/action_model': 5.923125816927049e-06, 'epoch': 1.93}) +[2026-06-16 04:18:44,168][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.03933696448802948, 'train/grad_norm_pre_clip': 12.972881317138672, 'timing/data': 0.0005128119955770671, 'timing/model': 1.1369547260110267, 'learning_rate/qwen_vl_interface': 5.917160558576088e-07, 'learning_rate/action_model': 5.917160558576088e-06, 'epoch': 1.93}) +[2026-06-16 04:18:48,712][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.0009895901894196868, 'train/grad_norm_pre_clip': 0.5972346663475037, 'timing/data': 0.00039865198777988553, 'timing/model': 1.1384425180149265, 'learning_rate/qwen_vl_interface': 5.911214448670829e-07, 'learning_rate/action_model': 5.911214448670828e-06, 'epoch': 1.93}) +[2026-06-16 04:18:53,293][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.033059198409318924, 'train/grad_norm_pre_clip': 13.421476364135742, 'timing/data': 0.0003495669807307422, 'timing/model': 1.1410687409806997, 'learning_rate/qwen_vl_interface': 5.905287489655499e-07, 'learning_rate/action_model': 5.905287489655499e-06, 'epoch': 1.93}) +[2026-06-16 04:18:57,851][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 0.002514120191335678, 'train/grad_norm_pre_clip': 0.46574369072914124, 'timing/data': 0.00034319498809054494, 'timing/model': 1.1346592780319043, 'learning_rate/qwen_vl_interface': 5.899379683966445e-07, 'learning_rate/action_model': 5.899379683966445e-06, 'epoch': 1.93}) +[2026-06-16 04:19:02,413][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.020493779331445694, 'train/grad_norm_pre_clip': 17.047109603881836, 'timing/data': 0.00036350800655782223, 'timing/model': 1.3058512050192803, 'learning_rate/qwen_vl_interface': 5.893491034032142e-07, 'learning_rate/action_model': 5.893491034032142e-06, 'epoch': 1.93}) +[2026-06-16 04:19:06,825][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.04252561181783676, 'train/grad_norm_pre_clip': 19.38772201538086, 'timing/data': 0.00035720004234462976, 'timing/model': 1.148263273993507, 'learning_rate/qwen_vl_interface': 5.887621542273196e-07, 'learning_rate/action_model': 5.887621542273196e-06, 'epoch': 1.93}) +[2026-06-16 04:19:11,390][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.07493113726377487, 'train/grad_norm_pre_clip': 21.96800994873047, 'timing/data': 0.0003860910073854029, 'timing/model': 1.1322373010334559, 'learning_rate/qwen_vl_interface': 5.881771211102327e-07, 'learning_rate/action_model': 5.881771211102327e-06, 'epoch': 1.93}) +[2026-06-16 04:23:32,193][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.0901852622628212, 'train/grad_norm_pre_clip': 23.190601348876953, 'eval/loss': 0.038262441754341125, 'eval/flap_precision': 0.9710806697108066, 'eval/flap_recall': 0.9845679012345679, 'eval/flap_f1': 0.9777777777777777, 'eval/flap_rate_model': 0.1825, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9777777777777777, 'timing/data': 0.0003818540135398507, 'timing/model': 1.1400620799977332, 'learning_rate/qwen_vl_interface': 5.875940042924391e-07, 'learning_rate/action_model': 5.875940042924391e-06, 'epoch': 1.93}) +[2026-06-16 04:23:32,193][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4700_state +[2026-06-16 04:23:32,196][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 04:23:35,789][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4700_state/model.safetensors +[2026-06-16 04:23:45,804][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4700_state/optimizer.bin +[2026-06-16 04:23:45,804][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4700_state/sampler.bin +[2026-06-16 04:23:45,804][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4700_state/sampler_1.bin +[2026-06-16 04:23:45,805][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4700_state/random_states_0.pkl +[2026-06-16 04:23:45,849][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4700 +[2026-06-16 04:23:45,849][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 04:23:45,856][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 04:23:50,321][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.00045461891568265855, 'train/grad_norm_pre_clip': 0.09326473623514175, 'timing/data': 0.000340436992701143, 'timing/model': 1.1199493440217339, 'learning_rate/qwen_vl_interface': 5.870128040136368e-07, 'learning_rate/action_model': 5.870128040136367e-06, 'epoch': 1.93}) +[2026-06-16 04:23:54,815][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.0004660903359763324, 'train/grad_norm_pre_clip': 0.07758462429046631, 'timing/data': 0.0003347449819557369, 'timing/model': 1.1150717670097947, 'learning_rate/qwen_vl_interface': 5.864335205127332e-07, 'learning_rate/action_model': 5.864335205127332e-06, 'epoch': 1.93}) +[2026-06-16 04:23:59,298][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.03374188765883446, 'train/grad_norm_pre_clip': 19.6719970703125, 'timing/data': 0.0003791220369748771, 'timing/model': 1.1166251830291003, 'learning_rate/qwen_vl_interface': 5.858561540278513e-07, 'learning_rate/action_model': 5.858561540278513e-06, 'epoch': 1.93}) +[2026-06-16 04:24:03,791][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 0.004634130746126175, 'train/grad_norm_pre_clip': 0.9547391533851624, 'timing/data': 0.0003763139829970896, 'timing/model': 1.118727035005577, 'learning_rate/qwen_vl_interface': 5.852807047963244e-07, 'learning_rate/action_model': 5.852807047963243e-06, 'epoch': 1.93}) +[2026-06-16 04:24:08,274][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.10100268572568893, 'train/grad_norm_pre_clip': 10.554118156433105, 'timing/data': 0.0005553779774345458, 'timing/model': 1.2797830330091529, 'learning_rate/qwen_vl_interface': 5.847071730546974e-07, 'learning_rate/action_model': 5.847071730546974e-06, 'epoch': 1.93}) +[2026-06-16 04:24:12,611][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.0015855260426178575, 'train/grad_norm_pre_clip': 0.6609496474266052, 'timing/data': 0.000347985012922436, 'timing/model': 1.1216649860143661, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 1.93}) +[2026-06-16 04:24:17,122][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 0.15175005793571472, 'train/grad_norm_pre_clip': 24.05687141418457, 'timing/data': 0.00033471599454060197, 'timing/model': 1.1309254959924147, 'learning_rate/qwen_vl_interface': 5.835658629833866e-07, 'learning_rate/action_model': 5.8356586298338654e-06, 'epoch': 1.93}) +[2026-06-16 04:24:21,649][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.13259842991828918, 'train/grad_norm_pre_clip': 38.41450500488281, 'timing/data': 0.0003342730342410505, 'timing/model': 1.1347576499683782, 'learning_rate/qwen_vl_interface': 5.829980851228512e-07, 'learning_rate/action_model': 5.829980851228512e-06, 'epoch': 1.93}) +[2026-06-16 04:24:26,165][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.0022433619014918804, 'train/grad_norm_pre_clip': 0.5075666904449463, 'timing/data': 0.00033378798980265856, 'timing/model': 1.1321197060169652, 'learning_rate/qwen_vl_interface': 5.824322256905155e-07, 'learning_rate/action_model': 5.824322256905154e-06, 'epoch': 1.93}) +[2026-06-16 04:24:30,645][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.08439133316278458, 'train/grad_norm_pre_clip': 5.501811504364014, 'timing/data': 0.0005529950140044093, 'timing/model': 1.1260240229894407, 'learning_rate/qwen_vl_interface': 5.818682849189818e-07, 'learning_rate/action_model': 5.818682849189817e-06, 'epoch': 1.93}) +[2026-06-16 04:24:35,231][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.004534763749688864, 'train/grad_norm_pre_clip': 1.418021321296692, 'timing/data': 0.0003942479961551726, 'timing/model': 1.3083036899915896, 'learning_rate/qwen_vl_interface': 5.813062630400653e-07, 'learning_rate/action_model': 5.813062630400653e-06, 'epoch': 1.94}) +[2026-06-16 04:24:39,620][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 0.05108784884214401, 'train/grad_norm_pre_clip': 15.576372146606445, 'timing/data': 0.00042832898907363415, 'timing/model': 1.131551610014867, 'learning_rate/qwen_vl_interface': 5.807461602847922e-07, 'learning_rate/action_model': 5.807461602847922e-06, 'epoch': 1.94}) +[2026-06-16 04:24:44,146][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.1403452605009079, 'train/grad_norm_pre_clip': 26.249235153198242, 'timing/data': 0.0006725980201736093, 'timing/model': 1.128957151027862, 'learning_rate/qwen_vl_interface': 5.801879768833998e-07, 'learning_rate/action_model': 5.801879768833998e-06, 'epoch': 1.94}) +[2026-06-16 04:24:48,688][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.0075541106052696705, 'train/grad_norm_pre_clip': 2.6005139350891113, 'timing/data': 0.00033844797872006893, 'timing/model': 1.1319489169982262, 'learning_rate/qwen_vl_interface': 5.796317130653363e-07, 'learning_rate/action_model': 5.796317130653363e-06, 'epoch': 1.94}) +[2026-06-16 04:24:53,239][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.23991534113883972, 'train/grad_norm_pre_clip': 4.0013322830200195, 'timing/data': 0.0003746559959836304, 'timing/model': 1.1369065980543382, 'learning_rate/qwen_vl_interface': 5.790773690592592e-07, 'learning_rate/action_model': 5.7907736905925925e-06, 'epoch': 1.94}) +[2026-06-16 04:24:57,765][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.0003290998574811965, 'train/grad_norm_pre_clip': 0.06001083552837372, 'timing/data': 0.0003323600394651294, 'timing/model': 1.1319632250233553, 'learning_rate/qwen_vl_interface': 5.7852494509304e-07, 'learning_rate/action_model': 5.7852494509304e-06, 'epoch': 1.94}) +[2026-06-16 04:25:02,311][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.0008695069700479507, 'train/grad_norm_pre_clip': 0.12392111867666245, 'timing/data': 0.00033692095894366503, 'timing/model': 1.305263342044782, 'learning_rate/qwen_vl_interface': 5.779744413937592e-07, 'learning_rate/action_model': 5.779744413937592e-06, 'epoch': 1.94}) +[2026-06-16 04:25:06,703][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.04611114412546158, 'train/grad_norm_pre_clip': 25.249738693237305, 'timing/data': 0.0003315029898658395, 'timing/model': 1.1351894440012984, 'learning_rate/qwen_vl_interface': 5.774258581877073e-07, 'learning_rate/action_model': 5.774258581877073e-06, 'epoch': 1.94}) +[2026-06-16 04:25:11,259][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.10518358647823334, 'train/grad_norm_pre_clip': 23.661293029785156, 'timing/data': 0.0003769479808397591, 'timing/model': 1.1371568519971333, 'learning_rate/qwen_vl_interface': 5.768791957003866e-07, 'learning_rate/action_model': 5.768791957003865e-06, 'epoch': 1.94}) +[2026-06-16 04:25:15,808][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.1537022590637207, 'train/grad_norm_pre_clip': 17.63519859313965, 'timing/data': 0.0003576450399123132, 'timing/model': 1.1426833929726854, 'learning_rate/qwen_vl_interface': 5.763344541565091e-07, 'learning_rate/action_model': 5.7633445415650905e-06, 'epoch': 1.94}) +[2026-06-16 04:25:20,368][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.009261141531169415, 'train/grad_norm_pre_clip': 3.2606048583984375, 'timing/data': 0.000337449018843472, 'timing/model': 1.1404199909884483, 'learning_rate/qwen_vl_interface': 5.75791633779998e-07, 'learning_rate/action_model': 5.75791633779998e-06, 'epoch': 1.94}) +[2026-06-16 04:25:24,876][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 0.0006646302063018084, 'train/grad_norm_pre_clip': 0.13924244046211243, 'timing/data': 0.0003657169872894883, 'timing/model': 1.1397382050054148, 'learning_rate/qwen_vl_interface': 5.752507347939857e-07, 'learning_rate/action_model': 5.752507347939858e-06, 'epoch': 1.94}) +[2026-06-16 04:25:29,423][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.37421709299087524, 'train/grad_norm_pre_clip': 37.88762283325195, 'timing/data': 0.00035961897810921073, 'timing/model': 1.2931788290152326, 'learning_rate/qwen_vl_interface': 5.74711757420816e-07, 'learning_rate/action_model': 5.747117574208159e-06, 'epoch': 1.94}) +[2026-06-16 04:25:33,799][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.005073842592537403, 'train/grad_norm_pre_clip': 1.0219390392303467, 'timing/data': 0.00040376902325078845, 'timing/model': 1.1421354620251805, 'learning_rate/qwen_vl_interface': 5.74174701882042e-07, 'learning_rate/action_model': 5.74174701882042e-06, 'epoch': 1.94}) +[2026-06-16 04:25:38,344][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.018337421119213104, 'train/grad_norm_pre_clip': 9.737454414367676, 'timing/data': 0.00040357501711696386, 'timing/model': 1.1330486059887335, 'learning_rate/qwen_vl_interface': 5.736395683984262e-07, 'learning_rate/action_model': 5.736395683984261e-06, 'epoch': 1.94}) +[2026-06-16 04:25:42,884][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.007181400433182716, 'train/grad_norm_pre_clip': 1.57590651512146, 'timing/data': 0.0003685409901663661, 'timing/model': 1.1408727059606463, 'learning_rate/qwen_vl_interface': 5.731063571899421e-07, 'learning_rate/action_model': 5.731063571899421e-06, 'epoch': 1.94}) +[2026-06-16 04:25:47,450][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.006909013260155916, 'train/grad_norm_pre_clip': 3.591456174850464, 'timing/data': 0.0004782089963555336, 'timing/model': 1.1313029400189407, 'learning_rate/qwen_vl_interface': 5.72575068475773e-07, 'learning_rate/action_model': 5.72575068475773e-06, 'epoch': 1.94}) +[2026-06-16 04:25:51,996][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.04284572973847389, 'train/grad_norm_pre_clip': 30.058589935302734, 'timing/data': 0.0003768100286833942, 'timing/model': 1.1461342929978855, 'learning_rate/qwen_vl_interface': 5.720457024743115e-07, 'learning_rate/action_model': 5.720457024743115e-06, 'epoch': 1.94}) +[2026-06-16 04:25:56,546][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.12786667048931122, 'train/grad_norm_pre_clip': 9.998486518859863, 'timing/data': 0.00036443601129576564, 'timing/model': 1.3043727679760195, 'learning_rate/qwen_vl_interface': 5.715182594031601e-07, 'learning_rate/action_model': 5.7151825940316e-06, 'epoch': 1.94}) +[2026-06-16 04:26:00,924][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.0017238185973837972, 'train/grad_norm_pre_clip': 0.45279279351234436, 'timing/data': 0.00034275499638170004, 'timing/model': 1.1333481480251066, 'learning_rate/qwen_vl_interface': 5.709927394791305e-07, 'learning_rate/action_model': 5.709927394791305e-06, 'epoch': 1.94}) +[2026-06-16 04:26:05,480][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.002806914271786809, 'train/grad_norm_pre_clip': 0.9541265964508057, 'timing/data': 0.00036896299570798874, 'timing/model': 1.1347789600258693, 'learning_rate/qwen_vl_interface': 5.70469142918244e-07, 'learning_rate/action_model': 5.704691429182441e-06, 'epoch': 1.94}) +[2026-06-16 04:26:10,037][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.00047495862236246467, 'train/grad_norm_pre_clip': 0.10023760795593262, 'timing/data': 0.0004390640533529222, 'timing/model': 1.1374958720407449, 'learning_rate/qwen_vl_interface': 5.699474699357317e-07, 'learning_rate/action_model': 5.699474699357317e-06, 'epoch': 1.94}) +[2026-06-16 04:26:14,590][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.008831167593598366, 'train/grad_norm_pre_clip': 4.083723068237305, 'timing/data': 0.0006754820351488888, 'timing/model': 1.1317574640270323, 'learning_rate/qwen_vl_interface': 5.694277207460331e-07, 'learning_rate/action_model': 5.694277207460332e-06, 'epoch': 1.94}) +[2026-06-16 04:26:19,097][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.0013316837139427662, 'train/grad_norm_pre_clip': 0.2768070101737976, 'timing/data': 0.00034440599847584963, 'timing/model': 1.1348702419782057, 'learning_rate/qwen_vl_interface': 5.689098955627982e-07, 'learning_rate/action_model': 5.689098955627982e-06, 'epoch': 1.94}) +[2026-06-16 04:26:23,647][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.021711667999625206, 'train/grad_norm_pre_clip': 7.896039962768555, 'timing/data': 0.0004728090134449303, 'timing/model': 1.3041538369725458, 'learning_rate/qwen_vl_interface': 5.683939945988853e-07, 'learning_rate/action_model': 5.683939945988853e-06, 'epoch': 1.95}) +[2026-06-16 04:26:28,031][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.08260294795036316, 'train/grad_norm_pre_clip': 20.10413360595703, 'timing/data': 0.0003570990520529449, 'timing/model': 1.1415209760307334, 'learning_rate/qwen_vl_interface': 5.678800180663608e-07, 'learning_rate/action_model': 5.678800180663609e-06, 'epoch': 1.95}) +[2026-06-16 04:26:32,564][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.10788533091545105, 'train/grad_norm_pre_clip': 24.757404327392578, 'timing/data': 0.00037902500480413437, 'timing/model': 1.1310016009956598, 'learning_rate/qwen_vl_interface': 5.673679661765018e-07, 'learning_rate/action_model': 5.6736796617650175e-06, 'epoch': 1.95}) +[2026-06-16 04:26:37,105][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.21340885758399963, 'train/grad_norm_pre_clip': 9.024280548095703, 'timing/data': 0.0003368230536580086, 'timing/model': 1.139363777008839, 'learning_rate/qwen_vl_interface': 5.668578391397938e-07, 'learning_rate/action_model': 5.668578391397937e-06, 'epoch': 1.95}) +[2026-06-16 04:26:41,652][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 0.0039020997937768698, 'train/grad_norm_pre_clip': 1.1961023807525635, 'timing/data': 0.0003754229983314872, 'timing/model': 1.1404743940220214, 'learning_rate/qwen_vl_interface': 5.663496371659303e-07, 'learning_rate/action_model': 5.663496371659303e-06, 'epoch': 1.95}) +[2026-06-16 04:26:46,195][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.04437540844082832, 'train/grad_norm_pre_clip': 18.24048614501953, 'timing/data': 0.00039257598109543324, 'timing/model': 1.1338020720286295, 'learning_rate/qwen_vl_interface': 5.658433604638141e-07, 'learning_rate/action_model': 5.65843360463814e-06, 'epoch': 1.95}) +[2026-06-16 04:26:50,744][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.023646144196391106, 'train/grad_norm_pre_clip': 9.023677825927734, 'timing/data': 0.0007360309828072786, 'timing/model': 1.297721517039463, 'learning_rate/qwen_vl_interface': 5.653390092415559e-07, 'learning_rate/action_model': 5.653390092415559e-06, 'epoch': 1.95}) +[2026-06-16 04:26:55,139][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 0.08283299952745438, 'train/grad_norm_pre_clip': 4.744690418243408, 'timing/data': 0.00038778100861236453, 'timing/model': 1.137335899984464, 'learning_rate/qwen_vl_interface': 5.648365837064759e-07, 'learning_rate/action_model': 5.648365837064759e-06, 'epoch': 1.95}) +[2026-06-16 04:26:59,710][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.22008347511291504, 'train/grad_norm_pre_clip': 26.05717658996582, 'timing/data': 0.000558088009711355, 'timing/model': 1.1482565829646774, 'learning_rate/qwen_vl_interface': 5.643360840651025e-07, 'learning_rate/action_model': 5.643360840651025e-06, 'epoch': 1.95}) +[2026-06-16 04:27:04,266][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.008137131109833717, 'train/grad_norm_pre_clip': 5.596432685852051, 'timing/data': 0.00032067502615973353, 'timing/model': 1.1413008910021745, 'learning_rate/qwen_vl_interface': 5.638375105231719e-07, 'learning_rate/action_model': 5.638375105231719e-06, 'epoch': 1.95}) +[2026-06-16 04:27:08,808][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.02581094391644001, 'train/grad_norm_pre_clip': 14.564436912536621, 'timing/data': 0.00032071000896394253, 'timing/model': 1.1351664260146208, 'learning_rate/qwen_vl_interface': 5.633408632856294e-07, 'learning_rate/action_model': 5.6334086328562945e-06, 'epoch': 1.95}) +[2026-06-16 04:27:13,321][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.06477532535791397, 'train/grad_norm_pre_clip': 13.56563949584961, 'timing/data': 0.00035159901017323136, 'timing/model': 1.1352273300290108, 'learning_rate/qwen_vl_interface': 5.628461425566268e-07, 'learning_rate/action_model': 5.6284614255662675e-06, 'epoch': 1.95}) +[2026-06-16 04:27:17,859][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.09528132528066635, 'train/grad_norm_pre_clip': 8.009042739868164, 'timing/data': 0.0003200580249540508, 'timing/model': 1.2889918900327757, 'learning_rate/qwen_vl_interface': 5.62353348539526e-07, 'learning_rate/action_model': 5.623533485395259e-06, 'epoch': 1.95}) +[2026-06-16 04:27:22,252][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 0.06564953923225403, 'train/grad_norm_pre_clip': 12.628785133361816, 'timing/data': 0.00044043001253157854, 'timing/model': 1.1431378760025837, 'learning_rate/qwen_vl_interface': 5.618624814368956e-07, 'learning_rate/action_model': 5.618624814368956e-06, 'epoch': 1.95}) +[2026-06-16 04:27:26,812][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.004337213933467865, 'train/grad_norm_pre_clip': 1.2200978994369507, 'timing/data': 0.00036788597935810685, 'timing/model': 1.133989499998279, 'learning_rate/qwen_vl_interface': 5.613735414505126e-07, 'learning_rate/action_model': 5.613735414505125e-06, 'epoch': 1.95}) +[2026-06-16 04:27:31,376][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.08303868770599365, 'train/grad_norm_pre_clip': 34.23908996582031, 'timing/data': 0.0003730399766936898, 'timing/model': 1.1374992040218785, 'learning_rate/qwen_vl_interface': 5.608865287813617e-07, 'learning_rate/action_model': 5.608865287813616e-06, 'epoch': 1.95}) +[2026-06-16 04:27:35,908][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.048977140337228775, 'train/grad_norm_pre_clip': 31.89194679260254, 'timing/data': 0.0003206860274076462, 'timing/model': 1.1359576560207643, 'learning_rate/qwen_vl_interface': 5.604014436296362e-07, 'learning_rate/action_model': 5.604014436296362e-06, 'epoch': 1.95}) +[2026-06-16 04:27:40,466][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.24787579476833344, 'train/grad_norm_pre_clip': 14.627656936645508, 'timing/data': 0.00040390301728621125, 'timing/model': 1.1344975919928402, 'learning_rate/qwen_vl_interface': 5.599182861947348e-07, 'learning_rate/action_model': 5.599182861947347e-06, 'epoch': 1.95}) +[2026-06-16 04:27:45,029][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.0002512072096578777, 'train/grad_norm_pre_clip': 0.06373568624258041, 'timing/data': 0.00031658599618822336, 'timing/model': 1.3028651810018346, 'learning_rate/qwen_vl_interface': 5.594370566752665e-07, 'learning_rate/action_model': 5.5943705667526645e-06, 'epoch': 1.95}) +[2026-06-16 04:27:49,437][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.01840572990477085, 'train/grad_norm_pre_clip': 15.227866172790527, 'timing/data': 0.000366177991963923, 'timing/model': 1.1375593199627474, 'learning_rate/qwen_vl_interface': 5.589577552690456e-07, 'learning_rate/action_model': 5.589577552690456e-06, 'epoch': 1.95}) +[2026-06-16 04:27:53,987][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.0006139923934824765, 'train/grad_norm_pre_clip': 0.10973668098449707, 'timing/data': 0.00039049802580848336, 'timing/model': 1.141027662961278, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 1.95}) +[2026-06-16 04:27:58,683][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.2299921214580536, 'train/grad_norm_pre_clip': 25.97498321533203, 'timing/data': 0.00038582400884479284, 'timing/model': 1.135537480993662, 'learning_rate/qwen_vl_interface': 5.580049375836466e-07, 'learning_rate/action_model': 5.580049375836466e-06, 'epoch': 1.95}) +[2026-06-16 04:28:03,251][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.004132436588406563, 'train/grad_norm_pre_clip': 0.9015671610832214, 'timing/data': 0.00042720098281279206, 'timing/model': 1.140133686014451, 'learning_rate/qwen_vl_interface': 5.575314216961356e-07, 'learning_rate/action_model': 5.575314216961356e-06, 'epoch': 1.95}) +[2026-06-16 04:28:07,748][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.0003101398760918528, 'train/grad_norm_pre_clip': 0.09382470697164536, 'timing/data': 0.0003328719758428633, 'timing/model': 1.1364207159494981, 'learning_rate/qwen_vl_interface': 5.570598347052073e-07, 'learning_rate/action_model': 5.570598347052072e-06, 'epoch': 1.95}) +[2026-06-16 04:28:12,291][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.0006648071575909853, 'train/grad_norm_pre_clip': 0.1265660524368286, 'timing/data': 0.0004742559976875782, 'timing/model': 1.2990604729857296, 'learning_rate/qwen_vl_interface': 5.565901768047126e-07, 'learning_rate/action_model': 5.5659017680471255e-06, 'epoch': 1.95}) +[2026-06-16 04:28:16,673][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.09868289530277252, 'train/grad_norm_pre_clip': 8.204237937927246, 'timing/data': 0.0003495379933156073, 'timing/model': 1.1326889609917998, 'learning_rate/qwen_vl_interface': 5.561224481877105e-07, 'learning_rate/action_model': 5.561224481877105e-06, 'epoch': 1.96}) +[2026-06-16 04:28:21,220][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.0011233811965212226, 'train/grad_norm_pre_clip': 0.14036114513874054, 'timing/data': 0.0004016069578938186, 'timing/model': 1.136289909016341, 'learning_rate/qwen_vl_interface': 5.55656649046467e-07, 'learning_rate/action_model': 5.556566490464669e-06, 'epoch': 1.96}) +[2026-06-16 04:28:25,769][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 0.014504696242511272, 'train/grad_norm_pre_clip': 11.630834579467773, 'timing/data': 0.0003300590324215591, 'timing/model': 1.1396236679865979, 'learning_rate/qwen_vl_interface': 5.551927795724544e-07, 'learning_rate/action_model': 5.551927795724543e-06, 'epoch': 1.96}) +[2026-06-16 04:28:30,299][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.0005384745309129357, 'train/grad_norm_pre_clip': 0.30058714747428894, 'timing/data': 0.000360608974006027, 'timing/model': 1.134378722985275, 'learning_rate/qwen_vl_interface': 5.547308399563515e-07, 'learning_rate/action_model': 5.547308399563514e-06, 'epoch': 1.96}) +[2026-06-16 04:28:34,855][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.0026584297884255648, 'train/grad_norm_pre_clip': 0.6889027953147888, 'timing/data': 0.0003181020147167146, 'timing/model': 1.1354829469928518, 'learning_rate/qwen_vl_interface': 5.542708303880442e-07, 'learning_rate/action_model': 5.542708303880442e-06, 'epoch': 1.96}) +[2026-06-16 04:28:39,396][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.005366249941289425, 'train/grad_norm_pre_clip': 2.1753792762756348, 'timing/data': 0.0005632259999401867, 'timing/model': 1.2938925789785571, 'learning_rate/qwen_vl_interface': 5.538127510566258e-07, 'learning_rate/action_model': 5.538127510566258e-06, 'epoch': 1.96}) +[2026-06-16 04:28:43,779][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.13512061536312103, 'train/grad_norm_pre_clip': 29.10385513305664, 'timing/data': 0.0003317910013720393, 'timing/model': 1.1357600730261765, 'learning_rate/qwen_vl_interface': 5.533566021503944e-07, 'learning_rate/action_model': 5.533566021503943e-06, 'epoch': 1.96}) +[2026-06-16 04:28:48,325][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 0.017451029270887375, 'train/grad_norm_pre_clip': 7.716835021972656, 'timing/data': 0.0003291739849373698, 'timing/model': 1.1334090559976175, 'learning_rate/qwen_vl_interface': 5.529023838568566e-07, 'learning_rate/action_model': 5.529023838568566e-06, 'epoch': 1.96}) +[2026-06-16 04:28:52,879][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.0011525721056386828, 'train/grad_norm_pre_clip': 0.20411325991153717, 'timing/data': 0.0003373019862920046, 'timing/model': 1.137987024965696, 'learning_rate/qwen_vl_interface': 5.524500963627239e-07, 'learning_rate/action_model': 5.524500963627239e-06, 'epoch': 1.96}) +[2026-06-16 04:28:57,408][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 0.07297312468290329, 'train/grad_norm_pre_clip': 17.89325714111328, 'timing/data': 0.00038315000711008906, 'timing/model': 1.133214033965487, 'learning_rate/qwen_vl_interface': 5.519997398539145e-07, 'learning_rate/action_model': 5.519997398539145e-06, 'epoch': 1.96}) +[2026-06-16 04:29:01,907][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.008404838852584362, 'train/grad_norm_pre_clip': 5.0827789306640625, 'timing/data': 0.00046637700870633125, 'timing/model': 1.1360498369904235, 'learning_rate/qwen_vl_interface': 5.515513145155531e-07, 'learning_rate/action_model': 5.515513145155531e-06, 'epoch': 1.96}) +[2026-06-16 04:29:06,444][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.0003042189928237349, 'train/grad_norm_pre_clip': 0.041274335235357285, 'timing/data': 0.00035146099980920553, 'timing/model': 1.2951776389963925, 'learning_rate/qwen_vl_interface': 5.511048205319702e-07, 'learning_rate/action_model': 5.511048205319701e-06, 'epoch': 1.96}) +[2026-06-16 04:29:10,832][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 0.0032942264806479216, 'train/grad_norm_pre_clip': 2.1447322368621826, 'timing/data': 0.00036776799242943525, 'timing/model': 1.132837268000003, 'learning_rate/qwen_vl_interface': 5.506602580867034e-07, 'learning_rate/action_model': 5.506602580867034e-06, 'epoch': 1.96}) +[2026-06-16 04:29:15,389][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.004441849421709776, 'train/grad_norm_pre_clip': 0.9236701130867004, 'timing/data': 0.00034353899536654353, 'timing/model': 1.1397807619650848, 'learning_rate/qwen_vl_interface': 5.502176273624942e-07, 'learning_rate/action_model': 5.502176273624942e-06, 'epoch': 1.96}) +[2026-06-16 04:29:19,935][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.015754299238324165, 'train/grad_norm_pre_clip': 9.345746040344238, 'timing/data': 0.0003809329937212169, 'timing/model': 1.1348777880193666, 'learning_rate/qwen_vl_interface': 5.497769285412921e-07, 'learning_rate/action_model': 5.497769285412921e-06, 'epoch': 1.96}) +[2026-06-16 04:29:24,480][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.10443711280822754, 'train/grad_norm_pre_clip': 13.62971305847168, 'timing/data': 0.0004411420086398721, 'timing/model': 1.1417452620225959, 'learning_rate/qwen_vl_interface': 5.493381618042519e-07, 'learning_rate/action_model': 5.493381618042518e-06, 'epoch': 1.96}) +[2026-06-16 04:29:29,025][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.0002665939973667264, 'train/grad_norm_pre_clip': 0.03250349685549736, 'timing/data': 0.00033882196294143796, 'timing/model': 1.1356645429623313, 'learning_rate/qwen_vl_interface': 5.489013273317343e-07, 'learning_rate/action_model': 5.489013273317343e-06, 'epoch': 1.96}) +[2026-06-16 04:29:33,570][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.0005305132945068181, 'train/grad_norm_pre_clip': 0.06597863137722015, 'timing/data': 0.00032151397317647934, 'timing/model': 1.2861384630086832, 'learning_rate/qwen_vl_interface': 5.484664253033042e-07, 'learning_rate/action_model': 5.484664253033042e-06, 'epoch': 1.96}) +[2026-06-16 04:29:37,993][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.17186148464679718, 'train/grad_norm_pre_clip': 12.047979354858398, 'timing/data': 0.0006330729811452329, 'timing/model': 1.143184213025961, 'learning_rate/qwen_vl_interface': 5.480334558977351e-07, 'learning_rate/action_model': 5.480334558977351e-06, 'epoch': 1.96}) +[2026-06-16 04:29:42,548][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.18057064712047577, 'train/grad_norm_pre_clip': 7.552646160125732, 'timing/data': 0.00045918201794847846, 'timing/model': 1.138975647976622, 'learning_rate/qwen_vl_interface': 5.476024192930029e-07, 'learning_rate/action_model': 5.47602419293003e-06, 'epoch': 1.96}) +[2026-06-16 04:29:47,094][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.00031426246277987957, 'train/grad_norm_pre_clip': 0.06140635162591934, 'timing/data': 0.0006466310005635023, 'timing/model': 1.1432228560443036, 'learning_rate/qwen_vl_interface': 5.471733156662914e-07, 'learning_rate/action_model': 5.4717331566629135e-06, 'epoch': 1.96}) +[2026-06-16 04:29:51,643][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 0.23796667158603668, 'train/grad_norm_pre_clip': 9.539094924926758, 'timing/data': 0.0005082339630462229, 'timing/model': 1.136602143989876, 'learning_rate/qwen_vl_interface': 5.467461451939882e-07, 'learning_rate/action_model': 5.467461451939882e-06, 'epoch': 1.96}) +[2026-06-16 04:29:56,148][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.1115039810538292, 'train/grad_norm_pre_clip': 48.316436767578125, 'timing/data': 0.00041347899241372943, 'timing/model': 1.136370514985174, 'learning_rate/qwen_vl_interface': 5.463209080516873e-07, 'learning_rate/action_model': 5.463209080516873e-06, 'epoch': 1.96}) +[2026-06-16 04:30:00,698][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.09725429862737656, 'train/grad_norm_pre_clip': 13.182580947875977, 'timing/data': 0.00040232803439721465, 'timing/model': 1.2984275090275332, 'learning_rate/qwen_vl_interface': 5.458976044141875e-07, 'learning_rate/action_model': 5.458976044141875e-06, 'epoch': 1.96}) +[2026-06-16 04:30:05,111][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.383224219083786, 'train/grad_norm_pre_clip': 31.343435287475586, 'timing/data': 0.00035540101816877723, 'timing/model': 1.143228916975204, 'learning_rate/qwen_vl_interface': 5.454762344554931e-07, 'learning_rate/action_model': 5.4547623445549316e-06, 'epoch': 1.97}) +[2026-06-16 04:30:09,646][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 0.015011991374194622, 'train/grad_norm_pre_clip': 12.182931900024414, 'timing/data': 0.00047981395618990064, 'timing/model': 1.1407958979834802, 'learning_rate/qwen_vl_interface': 5.450567983488137e-07, 'learning_rate/action_model': 5.450567983488137e-06, 'epoch': 1.97}) +[2026-06-16 04:30:14,221][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.07948622852563858, 'train/grad_norm_pre_clip': 10.64297103881836, 'timing/data': 0.00032368500251322985, 'timing/model': 1.1372031309874728, 'learning_rate/qwen_vl_interface': 5.446392962665629e-07, 'learning_rate/action_model': 5.4463929626656296e-06, 'epoch': 1.97}) +[2026-06-16 04:30:18,782][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.000312925229081884, 'train/grad_norm_pre_clip': 0.07990485429763794, 'timing/data': 0.0004147370345890522, 'timing/model': 1.1422774799866602, 'learning_rate/qwen_vl_interface': 5.442237283803605e-07, 'learning_rate/action_model': 5.442237283803604e-06, 'epoch': 1.97}) +[2026-06-16 04:30:23,327][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.0010902610374614596, 'train/grad_norm_pre_clip': 0.28731629252433777, 'timing/data': 0.0003832079819403589, 'timing/model': 1.144842914014589, 'learning_rate/qwen_vl_interface': 5.43810094861031e-07, 'learning_rate/action_model': 5.43810094861031e-06, 'epoch': 1.97}) +[2026-06-16 04:30:27,871][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.14992494881153107, 'train/grad_norm_pre_clip': 36.83442306518555, 'timing/data': 0.000399323005694896, 'timing/model': 1.2937612470122986, 'learning_rate/qwen_vl_interface': 5.433983958786027e-07, 'learning_rate/action_model': 5.433983958786027e-06, 'epoch': 1.97}) +[2026-06-16 04:30:32,256][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.26511821150779724, 'train/grad_norm_pre_clip': 16.043785095214844, 'timing/data': 0.00033905997406691313, 'timing/model': 1.1346182309789583, 'learning_rate/qwen_vl_interface': 5.429886316023102e-07, 'learning_rate/action_model': 5.429886316023101e-06, 'epoch': 1.97}) +[2026-06-16 04:30:36,799][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.001437365892343223, 'train/grad_norm_pre_clip': 0.9886431694030762, 'timing/data': 0.00033132301177829504, 'timing/model': 1.1396384990075603, 'learning_rate/qwen_vl_interface': 5.425808022005916e-07, 'learning_rate/action_model': 5.425808022005916e-06, 'epoch': 1.97}) +[2026-06-16 04:30:41,369][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.002955300034955144, 'train/grad_norm_pre_clip': 0.6662964224815369, 'timing/data': 0.0003204919630661607, 'timing/model': 1.1418948310310952, 'learning_rate/qwen_vl_interface': 5.421749078410903e-07, 'learning_rate/action_model': 5.421749078410903e-06, 'epoch': 1.97}) +[2026-06-16 04:30:45,907][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.0034529988188296556, 'train/grad_norm_pre_clip': 1.2277765274047852, 'timing/data': 0.0003427790361456573, 'timing/model': 1.1332601659814827, 'learning_rate/qwen_vl_interface': 5.417709486906547e-07, 'learning_rate/action_model': 5.417709486906547e-06, 'epoch': 1.97}) +[2026-06-16 04:30:50,418][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.1005135104060173, 'train/grad_norm_pre_clip': 27.079782485961914, 'timing/data': 0.0005205980269238353, 'timing/model': 1.1342792400391772, 'learning_rate/qwen_vl_interface': 5.413689249153367e-07, 'learning_rate/action_model': 5.413689249153366e-06, 'epoch': 1.97}) +[2026-06-16 04:30:54,991][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.1278596669435501, 'train/grad_norm_pre_clip': 21.631675720214844, 'timing/data': 0.00038707099156454206, 'timing/model': 1.3113671360188164, 'learning_rate/qwen_vl_interface': 5.409688366803931e-07, 'learning_rate/action_model': 5.409688366803931e-06, 'epoch': 1.97}) +[2026-06-16 04:30:59,392][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.002443847944959998, 'train/grad_norm_pre_clip': 1.030076265335083, 'timing/data': 0.0003327919985167682, 'timing/model': 1.1335587999783456, 'learning_rate/qwen_vl_interface': 5.405706841502853e-07, 'learning_rate/action_model': 5.4057068415028525e-06, 'epoch': 1.97}) +[2026-06-16 04:31:03,956][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 0.0404258593916893, 'train/grad_norm_pre_clip': 13.259747505187988, 'timing/data': 0.0006707729771733284, 'timing/model': 1.1367390420055017, 'learning_rate/qwen_vl_interface': 5.401744674886781e-07, 'learning_rate/action_model': 5.40174467488678e-06, 'epoch': 1.97}) +[2026-06-16 04:31:08,504][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.002261667512357235, 'train/grad_norm_pre_clip': 0.36468032002449036, 'timing/data': 0.0003520969767123461, 'timing/model': 1.1340453129960224, 'learning_rate/qwen_vl_interface': 5.397801868584418e-07, 'learning_rate/action_model': 5.397801868584417e-06, 'epoch': 1.97}) +[2026-06-16 04:31:13,046][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.0006132979760877788, 'train/grad_norm_pre_clip': 0.10885878652334213, 'timing/data': 0.00035945302806794643, 'timing/model': 1.1420507770380937, 'learning_rate/qwen_vl_interface': 5.393878424216502e-07, 'learning_rate/action_model': 5.393878424216501e-06, 'epoch': 1.97}) +[2026-06-16 04:35:35,325][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.2732393443584442, 'train/grad_norm_pre_clip': 23.633859634399414, 'eval/loss': 0.03870199993252754, 'eval/flap_precision': 0.9695121951219512, 'eval/flap_recall': 0.9814814814814815, 'eval/flap_f1': 0.9754601226993865, 'eval/flap_rate_model': 0.18222222222222223, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9754601226993865, 'timing/data': 0.0004107049899175763, 'timing/model': 1.1343068219721317, 'learning_rate/qwen_vl_interface': 5.389974343395813e-07, 'learning_rate/action_model': 5.389974343395812e-06, 'epoch': 1.97}) +[2026-06-16 04:35:35,325][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4800_state +[2026-06-16 04:35:35,328][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 04:35:38,913][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4800_state/model.safetensors +[2026-06-16 04:35:49,018][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4800_state/optimizer.bin +[2026-06-16 04:35:49,018][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4800_state/sampler.bin +[2026-06-16 04:35:49,019][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4800_state/sampler_1.bin +[2026-06-16 04:35:49,019][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4800_state/random_states_0.pkl +[2026-06-16 04:35:49,064][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4800 +[2026-06-16 04:35:49,065][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 04:35:49,071][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 04:35:53,512][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 0.029182294383645058, 'train/grad_norm_pre_clip': 20.407777786254883, 'timing/data': 0.0005788689595647156, 'timing/model': 1.1169488209998235, 'learning_rate/qwen_vl_interface': 5.38608962772717e-07, 'learning_rate/action_model': 5.38608962772717e-06, 'epoch': 1.97}) +[2026-06-16 04:35:57,957][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.021709101274609566, 'train/grad_norm_pre_clip': 12.665928840637207, 'timing/data': 0.00038556603249162436, 'timing/model': 1.2639385689981282, 'learning_rate/qwen_vl_interface': 5.382224278807429e-07, 'learning_rate/action_model': 5.382224278807429e-06, 'epoch': 1.97}) +[2026-06-16 04:36:02,255][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.1913888156414032, 'train/grad_norm_pre_clip': 15.032437324523926, 'timing/data': 0.00034054997377097607, 'timing/model': 1.1157395660411566, 'learning_rate/qwen_vl_interface': 5.378378298225496e-07, 'learning_rate/action_model': 5.378378298225496e-06, 'epoch': 1.97}) +[2026-06-16 04:36:06,753][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.027995798736810684, 'train/grad_norm_pre_clip': 8.746513366699219, 'timing/data': 0.00033926800824701786, 'timing/model': 1.116819561982993, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 1.97}) +[2026-06-16 04:36:11,142][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.002358329016715288, 'train/grad_norm_pre_clip': 0.4657071828842163, 'timing/data': 0.00033192295813933015, 'timing/model': 1.1189763729926199, 'learning_rate/qwen_vl_interface': 5.37074444839083e-07, 'learning_rate/action_model': 5.370744448390829e-06, 'epoch': 1.97}) +[2026-06-16 04:36:15,577][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 0.001108867465518415, 'train/grad_norm_pre_clip': 0.2266707569360733, 'timing/data': 0.0003296699724160135, 'timing/model': 1.124576103000436, 'learning_rate/qwen_vl_interface': 5.366956582276085e-07, 'learning_rate/action_model': 5.366956582276084e-06, 'epoch': 1.97}) +[2026-06-16 04:36:20,028][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.005113169085234404, 'train/grad_norm_pre_clip': 3.134852647781372, 'timing/data': 0.0003390979836694896, 'timing/model': 1.122902608010918, 'learning_rate/qwen_vl_interface': 5.363188090775122e-07, 'learning_rate/action_model': 5.363188090775122e-06, 'epoch': 1.97}) +[2026-06-16 04:36:24,527][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.078031025826931, 'train/grad_norm_pre_clip': 11.427948951721191, 'timing/data': 0.0005313719739206135, 'timing/model': 1.2817038869834505, 'learning_rate/qwen_vl_interface': 5.359438975437018e-07, 'learning_rate/action_model': 5.359438975437017e-06, 'epoch': 1.98}) +[2026-06-16 04:36:28,872][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.0034159566275775433, 'train/grad_norm_pre_clip': 0.6409866809844971, 'timing/data': 0.00032553798519074917, 'timing/model': 1.1359212459647097, 'learning_rate/qwen_vl_interface': 5.355709237802895e-07, 'learning_rate/action_model': 5.3557092378028955e-06, 'epoch': 1.98}) +[2026-06-16 04:36:33,375][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.04213340952992439, 'train/grad_norm_pre_clip': 17.259611129760742, 'timing/data': 0.0003607369726523757, 'timing/model': 1.1268010889762081, 'learning_rate/qwen_vl_interface': 5.351998879405911e-07, 'learning_rate/action_model': 5.351998879405911e-06, 'epoch': 1.98}) +[2026-06-16 04:36:37,875][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.0798545777797699, 'train/grad_norm_pre_clip': 9.868833541870117, 'timing/data': 0.00032765796640887856, 'timing/model': 1.1218978759716265, 'learning_rate/qwen_vl_interface': 5.348307901771256e-07, 'learning_rate/action_model': 5.348307901771256e-06, 'epoch': 1.98}) +[2026-06-16 04:36:42,364][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.09836864471435547, 'train/grad_norm_pre_clip': 26.264724731445312, 'timing/data': 0.0003339580143801868, 'timing/model': 1.1252269490505569, 'learning_rate/qwen_vl_interface': 5.344636306416147e-07, 'learning_rate/action_model': 5.344636306416146e-06, 'epoch': 1.98}) +[2026-06-16 04:36:46,880][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.00402872683480382, 'train/grad_norm_pre_clip': 3.1958441734313965, 'timing/data': 0.00035348295932635665, 'timing/model': 1.1308019710122608, 'learning_rate/qwen_vl_interface': 5.340984094849836e-07, 'learning_rate/action_model': 5.340984094849836e-06, 'epoch': 1.98}) +[2026-06-16 04:36:51,396][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.10931085795164108, 'train/grad_norm_pre_clip': 33.54280090332031, 'timing/data': 0.0003843269660137594, 'timing/model': 1.2818964659818448, 'learning_rate/qwen_vl_interface': 5.337351268573606e-07, 'learning_rate/action_model': 5.337351268573606e-06, 'epoch': 1.98}) +[2026-06-16 04:36:55,758][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.03798161447048187, 'train/grad_norm_pre_clip': 16.638713836669922, 'timing/data': 0.00034848402719944715, 'timing/model': 1.1273545370204374, 'learning_rate/qwen_vl_interface': 5.33373782908079e-07, 'learning_rate/action_model': 5.333737829080789e-06, 'epoch': 1.98}) +[2026-06-16 04:37:00,288][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.016928603872656822, 'train/grad_norm_pre_clip': 15.569954872131348, 'timing/data': 0.0003555280272848904, 'timing/model': 1.126582883996889, 'learning_rate/qwen_vl_interface': 5.330143777856718e-07, 'learning_rate/action_model': 5.330143777856718e-06, 'epoch': 1.98}) +[2026-06-16 04:37:04,800][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 0.01334469299763441, 'train/grad_norm_pre_clip': 6.840704917907715, 'timing/data': 0.00039181101601570845, 'timing/model': 1.127741297998, 'learning_rate/qwen_vl_interface': 5.326569116378786e-07, 'learning_rate/action_model': 5.326569116378786e-06, 'epoch': 1.98}) +[2026-06-16 04:37:09,314][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.021120261400938034, 'train/grad_norm_pre_clip': 9.487690925598145, 'timing/data': 0.0003348820027895272, 'timing/model': 1.1272830520174466, 'learning_rate/qwen_vl_interface': 5.323013846116387e-07, 'learning_rate/action_model': 5.323013846116386e-06, 'epoch': 1.98}) +[2026-06-16 04:37:13,787][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.0054414100013673306, 'train/grad_norm_pre_clip': 2.5110907554626465, 'timing/data': 0.00037133798468858004, 'timing/model': 1.1277341929962859, 'learning_rate/qwen_vl_interface': 5.31947796853097e-07, 'learning_rate/action_model': 5.3194779685309696e-06, 'epoch': 1.98}) +[2026-06-16 04:37:18,301][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.04562488570809364, 'train/grad_norm_pre_clip': 20.074562072753906, 'timing/data': 0.00034485303331166506, 'timing/model': 1.2863279940211214, 'learning_rate/qwen_vl_interface': 5.315961485075992e-07, 'learning_rate/action_model': 5.315961485075993e-06, 'epoch': 1.98}) +[2026-06-16 04:37:22,668][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.09555434435606003, 'train/grad_norm_pre_clip': 7.75192403793335, 'timing/data': 0.0005095309461466968, 'timing/model': 1.129930812981911, 'learning_rate/qwen_vl_interface': 5.312464397196958e-07, 'learning_rate/action_model': 5.312464397196958e-06, 'epoch': 1.98}) +[2026-06-16 04:37:27,202][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.00029823437216691673, 'train/grad_norm_pre_clip': 0.05500739440321922, 'timing/data': 0.0003628979902714491, 'timing/model': 1.1365905120037496, 'learning_rate/qwen_vl_interface': 5.308986706331377e-07, 'learning_rate/action_model': 5.308986706331376e-06, 'epoch': 1.98}) +[2026-06-16 04:37:31,761][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.004000811837613583, 'train/grad_norm_pre_clip': 1.2892305850982666, 'timing/data': 0.00035448395647108555, 'timing/model': 1.1412227459950373, 'learning_rate/qwen_vl_interface': 5.305528413908797e-07, 'learning_rate/action_model': 5.305528413908798e-06, 'epoch': 1.98}) +[2026-06-16 04:37:36,310][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.05737428739666939, 'train/grad_norm_pre_clip': 26.73910140991211, 'timing/data': 0.00034752703504636884, 'timing/model': 1.1386575289652683, 'learning_rate/qwen_vl_interface': 5.302089521350799e-07, 'learning_rate/action_model': 5.3020895213507996e-06, 'epoch': 1.98}) +[2026-06-16 04:37:40,883][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.004393028095364571, 'train/grad_norm_pre_clip': 1.2809264659881592, 'timing/data': 0.0003644769894890487, 'timing/model': 1.1397506479988806, 'learning_rate/qwen_vl_interface': 5.298670030070977e-07, 'learning_rate/action_model': 5.298670030070976e-06, 'epoch': 1.98}) +[2026-06-16 04:37:45,459][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.00036393312620930374, 'train/grad_norm_pre_clip': 0.05577455833554268, 'timing/data': 0.0003664880059659481, 'timing/model': 1.3047097260132432, 'learning_rate/qwen_vl_interface': 5.295269941474957e-07, 'learning_rate/action_model': 5.295269941474957e-06, 'epoch': 1.98}) +[2026-06-16 04:37:49,821][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.000477147608762607, 'train/grad_norm_pre_clip': 0.24645420908927917, 'timing/data': 0.00034392497036606073, 'timing/model': 1.1318887210218236, 'learning_rate/qwen_vl_interface': 5.291889256960384e-07, 'learning_rate/action_model': 5.291889256960384e-06, 'epoch': 1.98}) +[2026-06-16 04:37:54,376][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.03065805695950985, 'train/grad_norm_pre_clip': 16.031675338745117, 'timing/data': 0.00035729602677747607, 'timing/model': 1.1354850950301625, 'learning_rate/qwen_vl_interface': 5.288527977916929e-07, 'learning_rate/action_model': 5.288527977916929e-06, 'epoch': 1.98}) +[2026-06-16 04:37:58,933][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.005317636299878359, 'train/grad_norm_pre_clip': 1.2794272899627686, 'timing/data': 0.0003418140113353729, 'timing/model': 1.1299163060029969, 'learning_rate/qwen_vl_interface': 5.28518610572629e-07, 'learning_rate/action_model': 5.28518610572629e-06, 'epoch': 1.98}) +[2026-06-16 04:38:03,477][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.0035780041944235563, 'train/grad_norm_pre_clip': 0.7895829677581787, 'timing/data': 0.0003979360335506499, 'timing/model': 1.1273189350031316, 'learning_rate/qwen_vl_interface': 5.281863641762184e-07, 'learning_rate/action_model': 5.281863641762184e-06, 'epoch': 1.98}) +[2026-06-16 04:38:07,959][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 0.00035031966399401426, 'train/grad_norm_pre_clip': 0.14861811697483063, 'timing/data': 0.00038560800021514297, 'timing/model': 1.130070062004961, 'learning_rate/qwen_vl_interface': 5.278560587390355e-07, 'learning_rate/action_model': 5.278560587390355e-06, 'epoch': 1.98}) +[2026-06-16 04:38:12,477][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.13587890565395355, 'train/grad_norm_pre_clip': 21.979652404785156, 'timing/data': 0.0005524620064534247, 'timing/model': 1.2842934989603236, 'learning_rate/qwen_vl_interface': 5.275276943968557e-07, 'learning_rate/action_model': 5.275276943968557e-06, 'epoch': 1.98}) +[2026-06-16 04:38:16,842][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.022463664412498474, 'train/grad_norm_pre_clip': 11.790319442749023, 'timing/data': 0.00034994794987142086, 'timing/model': 1.1308800879633054, 'learning_rate/qwen_vl_interface': 5.272012712846566e-07, 'learning_rate/action_model': 5.2720127128465654e-06, 'epoch': 1.99}) +[2026-06-16 04:38:21,383][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.003392568090930581, 'train/grad_norm_pre_clip': 0.7865183353424072, 'timing/data': 0.00036356603959575295, 'timing/model': 1.1359213349642232, 'learning_rate/qwen_vl_interface': 5.268767895366198e-07, 'learning_rate/action_model': 5.268767895366198e-06, 'epoch': 1.99}) +[2026-06-16 04:38:25,918][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.00352583359926939, 'train/grad_norm_pre_clip': 2.2857673168182373, 'timing/data': 0.0003471940290182829, 'timing/model': 1.1367826709756628, 'learning_rate/qwen_vl_interface': 5.265542492861268e-07, 'learning_rate/action_model': 5.2655424928612676e-06, 'epoch': 1.99}) +[2026-06-16 04:38:30,467][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 0.008290618658065796, 'train/grad_norm_pre_clip': 4.625100612640381, 'timing/data': 0.0003129969700239599, 'timing/model': 1.1534379409858957, 'learning_rate/qwen_vl_interface': 5.262336506657617e-07, 'learning_rate/action_model': 5.2623365066576165e-06, 'epoch': 1.99}) +[2026-06-16 04:38:34,998][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.028093580156564713, 'train/grad_norm_pre_clip': 18.395200729370117, 'timing/data': 0.00037760200211778283, 'timing/model': 1.1312415139982477, 'learning_rate/qwen_vl_interface': 5.259149938073103e-07, 'learning_rate/action_model': 5.259149938073102e-06, 'epoch': 1.99}) +[2026-06-16 04:38:39,539][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.09462951123714447, 'train/grad_norm_pre_clip': 32.717735290527344, 'timing/data': 0.00034680700628086925, 'timing/model': 1.3054916799883358, 'learning_rate/qwen_vl_interface': 5.255982788417609e-07, 'learning_rate/action_model': 5.255982788417608e-06, 'epoch': 1.99}) +[2026-06-16 04:38:43,918][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.0019317751284688711, 'train/grad_norm_pre_clip': 0.5017132759094238, 'timing/data': 0.000342715997248888, 'timing/model': 1.1390732360305265, 'learning_rate/qwen_vl_interface': 5.252835058993029e-07, 'learning_rate/action_model': 5.252835058993029e-06, 'epoch': 1.99}) +[2026-06-16 04:38:48,466][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.009967039339244366, 'train/grad_norm_pre_clip': 4.835331916809082, 'timing/data': 0.0003571829874999821, 'timing/model': 1.1329273899900727, 'learning_rate/qwen_vl_interface': 5.249706751093274e-07, 'learning_rate/action_model': 5.249706751093273e-06, 'epoch': 1.99}) +[2026-06-16 04:38:53,013][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 0.022672075778245926, 'train/grad_norm_pre_clip': 15.431232452392578, 'timing/data': 0.00038145395228639245, 'timing/model': 1.1415214160224423, 'learning_rate/qwen_vl_interface': 5.246597866004273e-07, 'learning_rate/action_model': 5.246597866004273e-06, 'epoch': 1.99}) +[2026-06-16 04:38:57,548][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.08780316263437271, 'train/grad_norm_pre_clip': 8.89504337310791, 'timing/data': 0.0003935190034098923, 'timing/model': 1.129761527001392, 'learning_rate/qwen_vl_interface': 5.243508405003975e-07, 'learning_rate/action_model': 5.243508405003975e-06, 'epoch': 1.99}) +[2026-06-16 04:39:02,036][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.002079552272334695, 'train/grad_norm_pre_clip': 0.692865788936615, 'timing/data': 0.00040550698759034276, 'timing/model': 1.1306084130192176, 'learning_rate/qwen_vl_interface': 5.240438369362336e-07, 'learning_rate/action_model': 5.2404383693623366e-06, 'epoch': 1.99}) +[2026-06-16 04:39:06,626][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.006653427146375179, 'train/grad_norm_pre_clip': 1.3041633367538452, 'timing/data': 0.00041153404163196683, 'timing/model': 1.2937024389975704, 'learning_rate/qwen_vl_interface': 5.237387760341335e-07, 'learning_rate/action_model': 5.237387760341335e-06, 'epoch': 1.99}) +[2026-06-16 04:39:11,013][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.0006615970050916076, 'train/grad_norm_pre_clip': 0.1514188051223755, 'timing/data': 0.0003259310033172369, 'timing/model': 1.1296295810025185, 'learning_rate/qwen_vl_interface': 5.234356579194963e-07, 'learning_rate/action_model': 5.234356579194963e-06, 'epoch': 1.99}) +[2026-06-16 04:39:15,568][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.014790872111916542, 'train/grad_norm_pre_clip': 8.189497947692871, 'timing/data': 0.00040344096487388015, 'timing/model': 1.1336809170315973, 'learning_rate/qwen_vl_interface': 5.231344827169218e-07, 'learning_rate/action_model': 5.231344827169217e-06, 'epoch': 1.99}) +[2026-06-16 04:39:20,121][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.1344624161720276, 'train/grad_norm_pre_clip': 19.785451889038086, 'timing/data': 0.00036049296613782644, 'timing/model': 1.131151615001727, 'learning_rate/qwen_vl_interface': 5.228352505502119e-07, 'learning_rate/action_model': 5.228352505502119e-06, 'epoch': 1.99}) +[2026-06-16 04:39:24,633][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.0012140878243371844, 'train/grad_norm_pre_clip': 0.5985003709793091, 'timing/data': 0.00033719500061124563, 'timing/model': 1.139661320950836, 'learning_rate/qwen_vl_interface': 5.225379615423701e-07, 'learning_rate/action_model': 5.2253796154237e-06, 'epoch': 1.99}) +[2026-06-16 04:39:29,171][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.009101173840463161, 'train/grad_norm_pre_clip': 4.476716041564941, 'timing/data': 0.0004279580316506326, 'timing/model': 1.1313874580082484, 'learning_rate/qwen_vl_interface': 5.222426158156008e-07, 'learning_rate/action_model': 5.222426158156008e-06, 'epoch': 1.99}) +[2026-06-16 04:39:33,726][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.003708361182361841, 'train/grad_norm_pre_clip': 0.6917380094528198, 'timing/data': 0.00032072700560092926, 'timing/model': 1.304871980973985, 'learning_rate/qwen_vl_interface': 5.219492134913084e-07, 'learning_rate/action_model': 5.219492134913084e-06, 'epoch': 1.99}) +[2026-06-16 04:39:38,109][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.1050645112991333, 'train/grad_norm_pre_clip': 9.340489387512207, 'timing/data': 0.0003393460065126419, 'timing/model': 1.1361079390044324, 'learning_rate/qwen_vl_interface': 5.216577546901008e-07, 'learning_rate/action_model': 5.216577546901007e-06, 'epoch': 1.99}) +[2026-06-16 04:39:42,648][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 0.3066529333591461, 'train/grad_norm_pre_clip': 19.351476669311523, 'timing/data': 0.00033004197757691145, 'timing/model': 1.131495012959931, 'learning_rate/qwen_vl_interface': 5.213682395317852e-07, 'learning_rate/action_model': 5.213682395317851e-06, 'epoch': 1.99}) +[2026-06-16 04:39:47,190][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 0.2811063528060913, 'train/grad_norm_pre_clip': 12.023077964782715, 'timing/data': 0.00034074397990480065, 'timing/model': 1.135945466987323, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 1.99}) +[2026-06-16 04:39:51,731][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.15324005484580994, 'train/grad_norm_pre_clip': 11.271994590759277, 'timing/data': 0.0003592889988794923, 'timing/model': 1.1324364120373502, 'learning_rate/qwen_vl_interface': 5.207950406190654e-07, 'learning_rate/action_model': 5.207950406190653e-06, 'epoch': 1.99}) +[2026-06-16 04:39:56,226][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.0011793003650382161, 'train/grad_norm_pre_clip': 0.2263759821653366, 'timing/data': 0.00033536204136908054, 'timing/model': 1.1327350679785013, 'learning_rate/qwen_vl_interface': 5.205113571002818e-07, 'learning_rate/action_model': 5.205113571002818e-06, 'epoch': 1.99}) +[2026-06-16 04:40:00,760][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.1506841778755188, 'train/grad_norm_pre_clip': 18.63360595703125, 'timing/data': 0.00033090199576690793, 'timing/model': 1.2946028169826604, 'learning_rate/qwen_vl_interface': 5.202296176956308e-07, 'learning_rate/action_model': 5.202296176956308e-06, 'epoch': 1.99}) +[2026-06-16 04:40:05,159][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 0.12089306861162186, 'train/grad_norm_pre_clip': 10.945812225341797, 'timing/data': 0.00047508999705314636, 'timing/model': 1.1507786980364472, 'learning_rate/qwen_vl_interface': 5.19949822520925e-07, 'learning_rate/action_model': 5.19949822520925e-06, 'epoch': 2.0}) +[2026-06-16 04:40:09,756][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.04776494577527046, 'train/grad_norm_pre_clip': 17.12442398071289, 'timing/data': 0.0006023409659974277, 'timing/model': 1.1375287970295176, 'learning_rate/qwen_vl_interface': 5.196719716911777e-07, 'learning_rate/action_model': 5.1967197169117764e-06, 'epoch': 2.0}) +[2026-06-16 04:40:14,292][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.0038593276403844357, 'train/grad_norm_pre_clip': 1.1388208866119385, 'timing/data': 0.0003528789966367185, 'timing/model': 1.1337600810220465, 'learning_rate/qwen_vl_interface': 5.193960653206029e-07, 'learning_rate/action_model': 5.193960653206029e-06, 'epoch': 2.0}) +[2026-06-16 04:40:18,825][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.14476127922534943, 'train/grad_norm_pre_clip': 8.018270492553711, 'timing/data': 0.00041531596798449755, 'timing/model': 1.1351826509926468, 'learning_rate/qwen_vl_interface': 5.191221035226144e-07, 'learning_rate/action_model': 5.191221035226144e-06, 'epoch': 2.0}) +[2026-06-16 04:40:23,369][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.08758358657360077, 'train/grad_norm_pre_clip': 18.07564926147461, 'timing/data': 0.00042796204797923565, 'timing/model': 1.1318496510502882, 'learning_rate/qwen_vl_interface': 5.188500864098292e-07, 'learning_rate/action_model': 5.1885008640982915e-06, 'epoch': 2.0}) +[2026-06-16 04:40:27,901][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.27151575684547424, 'train/grad_norm_pre_clip': 23.746856689453125, 'timing/data': 0.0004270779900252819, 'timing/model': 1.2842583599849604, 'learning_rate/qwen_vl_interface': 5.185800140940613e-07, 'learning_rate/action_model': 5.185800140940613e-06, 'epoch': 2.0}) +[2026-06-16 04:40:32,287][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 0.0013312185183167458, 'train/grad_norm_pre_clip': 0.23730459809303284, 'timing/data': 0.00035088701406493783, 'timing/model': 1.1352015280281194, 'learning_rate/qwen_vl_interface': 5.183118866863288e-07, 'learning_rate/action_model': 5.1831188668632875e-06, 'epoch': 2.0}) +[2026-06-16 04:40:36,839][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.021246472373604774, 'train/grad_norm_pre_clip': 14.759849548339844, 'timing/data': 0.0003614889574237168, 'timing/model': 1.1383615630329587, 'learning_rate/qwen_vl_interface': 5.180457042968484e-07, 'learning_rate/action_model': 5.1804570429684835e-06, 'epoch': 2.0}) +[2026-06-16 04:40:41,397][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.0007803320186212659, 'train/grad_norm_pre_clip': 0.5248014330863953, 'timing/data': 0.0003754560020752251, 'timing/model': 1.1460381330107339, 'learning_rate/qwen_vl_interface': 5.17781467035037e-07, 'learning_rate/action_model': 5.177814670350371e-06, 'epoch': 2.0}) +[2026-06-16 04:40:45,936][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.13429708778858185, 'train/grad_norm_pre_clip': 32.597412109375, 'timing/data': 0.0003632879815995693, 'timing/model': 1.1331111209583469, 'learning_rate/qwen_vl_interface': 5.175191750095133e-07, 'learning_rate/action_model': 5.175191750095132e-06, 'epoch': 2.0}) +[2026-06-16 04:40:50,424][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.10695916414260864, 'train/grad_norm_pre_clip': 12.69179630279541, 'timing/data': 0.00035063800169155, 'timing/model': 1.11988965398632, 'learning_rate/qwen_vl_interface': 5.172588283280953e-07, 'learning_rate/action_model': 5.172588283280953e-06, 'epoch': 2.0}) +[2026-06-16 04:40:54,854][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.049104005098342896, 'train/grad_norm_pre_clip': 13.325396537780762, 'timing/data': 0.0003164410009048879, 'timing/model': 1.2425944849965163, 'learning_rate/qwen_vl_interface': 5.170004270978019e-07, 'learning_rate/action_model': 5.170004270978018e-06, 'epoch': 2.0}) +[2026-06-16 04:40:58,295][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 4.207929305266589e-05, 'train/grad_norm_pre_clip': 0.008749722503125668, 'timing/data': 0.00037499895552173257, 'timing/model': 0.2670339159667492, 'learning_rate/qwen_vl_interface': 5.167439714248512e-07, 'learning_rate/action_model': 5.1674397142485116e-06, 'epoch': 2.0}) +[2026-06-16 04:41:04,381][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.11743450909852982, 'train/grad_norm_pre_clip': 16.738685607910156, 'timing/data': 0.0067051449441351, 'timing/model': 1.1311318399966694, 'learning_rate/qwen_vl_interface': 5.164894614146633e-07, 'learning_rate/action_model': 5.164894614146634e-06, 'epoch': 2.0}) +[2026-06-16 04:41:08,920][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.1694801300764084, 'train/grad_norm_pre_clip': 13.283050537109375, 'timing/data': 0.00685175001854077, 'timing/model': 1.1342027659993619, 'learning_rate/qwen_vl_interface': 5.16236897171858e-07, 'learning_rate/action_model': 5.16236897171858e-06, 'epoch': 2.0}) +[2026-06-16 04:41:13,470][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.03434397280216217, 'train/grad_norm_pre_clip': 46.18404006958008, 'timing/data': 0.0065484579536132514, 'timing/model': 1.135624008020386, 'learning_rate/qwen_vl_interface': 5.15986278800254e-07, 'learning_rate/action_model': 5.15986278800254e-06, 'epoch': 2.0}) +[2026-06-16 04:41:18,019][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.116164930164814, 'train/grad_norm_pre_clip': 9.444347381591797, 'timing/data': 0.0062120670336298645, 'timing/model': 1.2887218039832078, 'learning_rate/qwen_vl_interface': 5.157376064028719e-07, 'learning_rate/action_model': 5.157376064028718e-06, 'epoch': 2.0}) +[2026-06-16 04:41:22,408][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.0910310298204422, 'train/grad_norm_pre_clip': 25.520814895629883, 'timing/data': 0.0065514169982634485, 'timing/model': 1.1348910109954886, 'learning_rate/qwen_vl_interface': 5.154908800819307e-07, 'learning_rate/action_model': 5.154908800819306e-06, 'epoch': 2.0}) +[2026-06-16 04:41:26,957][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.0026593846268951893, 'train/grad_norm_pre_clip': 0.8237760066986084, 'timing/data': 0.006733268033713102, 'timing/model': 1.1307013310142793, 'learning_rate/qwen_vl_interface': 5.152460999388504e-07, 'learning_rate/action_model': 5.152460999388503e-06, 'epoch': 2.0}) +[2026-06-16 04:41:31,513][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.017062092199921608, 'train/grad_norm_pre_clip': 11.6260347366333, 'timing/data': 0.006812809035181999, 'timing/model': 1.1424441720009781, 'learning_rate/qwen_vl_interface': 5.150032660742517e-07, 'learning_rate/action_model': 5.150032660742517e-06, 'epoch': 2.0}) +[2026-06-16 04:41:36,061][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.0826503112912178, 'train/grad_norm_pre_clip': 10.262507438659668, 'timing/data': 0.006614520971197635, 'timing/model': 1.1351291069877334, 'learning_rate/qwen_vl_interface': 5.147623785879535e-07, 'learning_rate/action_model': 5.147623785879534e-06, 'epoch': 2.0}) +[2026-06-16 04:41:40,550][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.117576003074646, 'train/grad_norm_pre_clip': 17.25385284423828, 'timing/data': 0.0066020869999192655, 'timing/model': 1.1402420609956607, 'learning_rate/qwen_vl_interface': 5.145234375789758e-07, 'learning_rate/action_model': 5.145234375789758e-06, 'epoch': 2.0}) +[2026-06-16 04:41:45,041][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.03284096345305443, 'train/grad_norm_pre_clip': 20.51300048828125, 'timing/data': 0.006524246011395007, 'timing/model': 1.2455283470335416, 'learning_rate/qwen_vl_interface': 5.142864431455387e-07, 'learning_rate/action_model': 5.1428644314553875e-06, 'epoch': 2.0}) +[2026-06-16 04:41:49,424][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.18654783070087433, 'train/grad_norm_pre_clip': 20.36472511291504, 'timing/data': 0.0066856289631687105, 'timing/model': 1.1320901529979892, 'learning_rate/qwen_vl_interface': 5.140513953850612e-07, 'learning_rate/action_model': 5.140513953850612e-06, 'epoch': 2.0}) +[2026-06-16 04:41:53,985][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.2056533694267273, 'train/grad_norm_pre_clip': 19.839611053466797, 'timing/data': 0.006563639966771007, 'timing/model': 1.1404808419756591, 'learning_rate/qwen_vl_interface': 5.138182943941623e-07, 'learning_rate/action_model': 5.138182943941623e-06, 'epoch': 2.01}) +[2026-06-16 04:41:58,543][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.002860609209164977, 'train/grad_norm_pre_clip': 0.9167059659957886, 'timing/data': 0.0066468880395404994, 'timing/model': 1.1423073670011945, 'learning_rate/qwen_vl_interface': 5.135871402686615e-07, 'learning_rate/action_model': 5.135871402686615e-06, 'epoch': 2.01}) +[2026-06-16 04:42:03,081][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.0036667922977358103, 'train/grad_norm_pre_clip': 0.9152426719665527, 'timing/data': 0.006498290982563049, 'timing/model': 1.1367055649752729, 'learning_rate/qwen_vl_interface': 5.133579331035777e-07, 'learning_rate/action_model': 5.133579331035777e-06, 'epoch': 2.01}) +[2026-06-16 04:42:07,597][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.0005468719755299389, 'train/grad_norm_pre_clip': 0.11782065033912659, 'timing/data': 0.006294228951446712, 'timing/model': 1.1386821150081232, 'learning_rate/qwen_vl_interface': 5.131306729931284e-07, 'learning_rate/action_model': 5.1313067299312836e-06, 'epoch': 2.01}) +[2026-06-16 04:42:12,111][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.010785871185362339, 'train/grad_norm_pre_clip': 5.620442867279053, 'timing/data': 0.006548201956320554, 'timing/model': 1.2635205030092038, 'learning_rate/qwen_vl_interface': 5.129053600307326e-07, 'learning_rate/action_model': 5.129053600307326e-06, 'epoch': 2.01}) +[2026-06-16 04:42:16,509][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.061704594641923904, 'train/grad_norm_pre_clip': 18.717784881591797, 'timing/data': 0.0064639049815014005, 'timing/model': 1.1343617840320803, 'learning_rate/qwen_vl_interface': 5.126819943090078e-07, 'learning_rate/action_model': 5.1268199430900775e-06, 'epoch': 2.01}) +[2026-06-16 04:42:21,040][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.2322363704442978, 'train/grad_norm_pre_clip': 14.160516738891602, 'timing/data': 0.006434792012441903, 'timing/model': 1.1356940250261687, 'learning_rate/qwen_vl_interface': 5.124605759197706e-07, 'learning_rate/action_model': 5.124605759197706e-06, 'epoch': 2.01}) +[2026-06-16 04:42:25,597][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.0025836448185145855, 'train/grad_norm_pre_clip': 0.7305094003677368, 'timing/data': 0.006566547031980008, 'timing/model': 1.1347807459533215, 'learning_rate/qwen_vl_interface': 5.122411049540383e-07, 'learning_rate/action_model': 5.1224110495403825e-06, 'epoch': 2.01}) +[2026-06-16 04:42:30,137][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.19663870334625244, 'train/grad_norm_pre_clip': 23.689151763916016, 'timing/data': 0.006485198973678052, 'timing/model': 1.136525181995239, 'learning_rate/qwen_vl_interface': 5.120235815020269e-07, 'learning_rate/action_model': 5.120235815020269e-06, 'epoch': 2.01}) +[2026-06-16 04:42:34,688][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.016115006059408188, 'train/grad_norm_pre_clip': 10.519305229187012, 'timing/data': 0.006219174014404416, 'timing/model': 1.3003802709863521, 'learning_rate/qwen_vl_interface': 5.118080056531524e-07, 'learning_rate/action_model': 5.118080056531524e-06, 'epoch': 2.01}) +[2026-06-16 04:42:39,066][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.002346931491047144, 'train/grad_norm_pre_clip': 0.4750981628894806, 'timing/data': 0.00665247900178656, 'timing/model': 1.1335826230351813, 'learning_rate/qwen_vl_interface': 5.115943774960295e-07, 'learning_rate/action_model': 5.115943774960295e-06, 'epoch': 2.01}) +[2026-06-16 04:42:43,598][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.14925551414489746, 'train/grad_norm_pre_clip': 21.50387954711914, 'timing/data': 0.006432759982999414, 'timing/model': 1.1326645350200124, 'learning_rate/qwen_vl_interface': 5.113826971184727e-07, 'learning_rate/action_model': 5.113826971184727e-06, 'epoch': 2.01}) +[2026-06-16 04:42:48,058][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.0040843551978468895, 'train/grad_norm_pre_clip': 1.669156789779663, 'timing/data': 0.0067178099998272955, 'timing/model': 1.1402474209899083, 'learning_rate/qwen_vl_interface': 5.111729646074954e-07, 'learning_rate/action_model': 5.111729646074954e-06, 'epoch': 2.01}) +[2026-06-16 04:42:52,519][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.024383919313549995, 'train/grad_norm_pre_clip': 9.60593032836914, 'timing/data': 0.006567600998096168, 'timing/model': 1.134978469053749, 'learning_rate/qwen_vl_interface': 5.10965180049312e-07, 'learning_rate/action_model': 5.10965180049312e-06, 'epoch': 2.01}) +[2026-06-16 04:42:57,032][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.0026237666606903076, 'train/grad_norm_pre_clip': 1.0135934352874756, 'timing/data': 0.00640692695742473, 'timing/model': 1.1407392569817603, 'learning_rate/qwen_vl_interface': 5.107593435293336e-07, 'learning_rate/action_model': 5.107593435293336e-06, 'epoch': 2.01}) +[2026-06-16 04:43:01,563][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.027084555476903915, 'train/grad_norm_pre_clip': 23.14750862121582, 'timing/data': 0.0064540819730609655, 'timing/model': 1.2882982699666172, 'learning_rate/qwen_vl_interface': 5.105554551321724e-07, 'learning_rate/action_model': 5.105554551321723e-06, 'epoch': 2.01}) +[2026-06-16 04:43:05,941][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.0015792251797392964, 'train/grad_norm_pre_clip': 0.2950736880302429, 'timing/data': 0.006329071999061853, 'timing/model': 1.1322735410067253, 'learning_rate/qwen_vl_interface': 5.103535149416389e-07, 'learning_rate/action_model': 5.1035351494163885e-06, 'epoch': 2.01}) +[2026-06-16 04:43:10,486][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.12617050111293793, 'train/grad_norm_pre_clip': 13.638982772827148, 'timing/data': 0.006499906012322754, 'timing/model': 1.1439666640362702, 'learning_rate/qwen_vl_interface': 5.101535230407436e-07, 'learning_rate/action_model': 5.101535230407436e-06, 'epoch': 2.01}) +[2026-06-16 04:43:15,064][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.06819120794534683, 'train/grad_norm_pre_clip': 57.64051818847656, 'timing/data': 0.006625226989854127, 'timing/model': 1.150325992028229, 'learning_rate/qwen_vl_interface': 5.099554795116956e-07, 'learning_rate/action_model': 5.0995547951169565e-06, 'epoch': 2.01}) +[2026-06-16 04:47:32,425][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.009564775973558426, 'train/grad_norm_pre_clip': 15.573712348937988, 'eval/loss': 0.03932056948542595, 'eval/flap_precision': 0.9710365853658537, 'eval/flap_recall': 0.9830246913580247, 'eval/flap_f1': 0.976993865030675, 'eval/flap_rate_model': 0.18222222222222223, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.976993865030675, 'timing/data': 0.006444501981604844, 'timing/model': 1.1434820970171131, 'learning_rate/qwen_vl_interface': 5.097593844359024e-07, 'learning_rate/action_model': 5.0975938443590245e-06, 'epoch': 2.01}) +[2026-06-16 04:47:32,426][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4900_state +[2026-06-16 04:47:32,428][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 04:47:36,005][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4900_state/model.safetensors +[2026-06-16 04:47:45,963][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4900_state/optimizer.bin +[2026-06-16 04:47:45,963][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4900_state/sampler.bin +[2026-06-16 04:47:45,963][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4900_state/sampler_1.bin +[2026-06-16 04:47:45,964][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_4900_state/random_states_0.pkl +[2026-06-16 04:47:46,011][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4900 +[2026-06-16 04:47:46,011][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 04:47:46,017][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 04:47:50,479][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.1921015977859497, 'train/grad_norm_pre_clip': 16.220197677612305, 'timing/data': 0.00643382198177278, 'timing/model': 1.110585879010614, 'learning_rate/qwen_vl_interface': 5.095652378939718e-07, 'learning_rate/action_model': 5.095652378939717e-06, 'epoch': 2.01}) +[2026-06-16 04:47:54,957][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.10059847682714462, 'train/grad_norm_pre_clip': 7.690313339233398, 'timing/data': 0.006502075004391372, 'timing/model': 1.1148061529966071, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 2.01}) +[2026-06-16 04:47:59,415][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.07796818017959595, 'train/grad_norm_pre_clip': 10.772971153259277, 'timing/data': 0.006505493016447872, 'timing/model': 1.2631237450405024, 'learning_rate/qwen_vl_interface': 5.091827907301227e-07, 'learning_rate/action_model': 5.091827907301227e-06, 'epoch': 2.01}) +[2026-06-16 04:48:03,837][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.031510792672634125, 'train/grad_norm_pre_clip': 12.207975387573242, 'timing/data': 0.006616007012780756, 'timing/model': 1.2274077349575236, 'learning_rate/qwen_vl_interface': 5.089944902654134e-07, 'learning_rate/action_model': 5.089944902654134e-06, 'epoch': 2.01}) +[2026-06-16 04:48:08,172][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.028923260048031807, 'train/grad_norm_pre_clip': 17.12613296508789, 'timing/data': 0.00661472202045843, 'timing/model': 1.1308580510085449, 'learning_rate/qwen_vl_interface': 5.088081386489856e-07, 'learning_rate/action_model': 5.088081386489857e-06, 'epoch': 2.01}) +[2026-06-16 04:48:12,663][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.0012110801180824637, 'train/grad_norm_pre_clip': 0.6508737206459045, 'timing/data': 0.006252797960769385, 'timing/model': 1.121625172032509, 'learning_rate/qwen_vl_interface': 5.086237359574419e-07, 'learning_rate/action_model': 5.086237359574419e-06, 'epoch': 2.02}) +[2026-06-16 04:48:17,174][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.0023665782064199448, 'train/grad_norm_pre_clip': 0.581218421459198, 'timing/data': 0.0064223879599012434, 'timing/model': 1.1203473689965904, 'learning_rate/qwen_vl_interface': 5.084412822665825e-07, 'learning_rate/action_model': 5.084412822665825e-06, 'epoch': 2.02}) +[2026-06-16 04:48:21,677][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.05068972706794739, 'train/grad_norm_pre_clip': 28.492902755737305, 'timing/data': 0.006422318983823061, 'timing/model': 1.1261546250316314, 'learning_rate/qwen_vl_interface': 5.082607776514081e-07, 'learning_rate/action_model': 5.082607776514081e-06, 'epoch': 2.02}) +[2026-06-16 04:48:26,175][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.026974065229296684, 'train/grad_norm_pre_clip': 6.8734211921691895, 'timing/data': 0.00610106298699975, 'timing/model': 1.2759114280343056, 'learning_rate/qwen_vl_interface': 5.080822221861167e-07, 'learning_rate/action_model': 5.080822221861166e-06, 'epoch': 2.02}) +[2026-06-16 04:48:30,579][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.004505245480686426, 'train/grad_norm_pre_clip': 0.7019903659820557, 'timing/data': 0.006651305011473596, 'timing/model': 1.1818613089853898, 'learning_rate/qwen_vl_interface': 5.07905615944106e-07, 'learning_rate/action_model': 5.0790561594410595e-06, 'epoch': 2.02}) +[2026-06-16 04:48:34,916][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.16348381340503693, 'train/grad_norm_pre_clip': 29.883325576782227, 'timing/data': 0.00640465400647372, 'timing/model': 1.1221828819834627, 'learning_rate/qwen_vl_interface': 5.077309589979716e-07, 'learning_rate/action_model': 5.077309589979715e-06, 'epoch': 2.02}) +[2026-06-16 04:48:39,413][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.09448402374982834, 'train/grad_norm_pre_clip': 26.12109375, 'timing/data': 0.006366026995237917, 'timing/model': 1.1216469130013138, 'learning_rate/qwen_vl_interface': 5.075582514195089e-07, 'learning_rate/action_model': 5.0755825141950895e-06, 'epoch': 2.02}) +[2026-06-16 04:48:43,932][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 0.000776365224737674, 'train/grad_norm_pre_clip': 0.14345791935920715, 'timing/data': 0.006421160011086613, 'timing/model': 1.131303914997261, 'learning_rate/qwen_vl_interface': 5.073874932797115e-07, 'learning_rate/action_model': 5.0738749327971155e-06, 'epoch': 2.02}) +[2026-06-16 04:48:48,441][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.31442126631736755, 'train/grad_norm_pre_clip': 23.747901916503906, 'timing/data': 0.006359227991197258, 'timing/model': 1.1321333069936372, 'learning_rate/qwen_vl_interface': 5.072186846487713e-07, 'learning_rate/action_model': 5.072186846487713e-06, 'epoch': 2.02}) +[2026-06-16 04:48:52,943][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 0.0016132438322529197, 'train/grad_norm_pre_clip': 0.5636369585990906, 'timing/data': 0.006475971022155136, 'timing/model': 1.2793555930256844, 'learning_rate/qwen_vl_interface': 5.070518255960796e-07, 'learning_rate/action_model': 5.070518255960795e-06, 'epoch': 2.02}) +[2026-06-16 04:48:57,349][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.0035280794836580753, 'train/grad_norm_pre_clip': 0.572563648223877, 'timing/data': 0.0063682220061309636, 'timing/model': 1.1819539360003546, 'learning_rate/qwen_vl_interface': 5.068869161902256e-07, 'learning_rate/action_model': 5.068869161902256e-06, 'epoch': 2.02}) +[2026-06-16 04:49:01,708][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.005273954011499882, 'train/grad_norm_pre_clip': 1.8006352186203003, 'timing/data': 0.006391855946276337, 'timing/model': 1.1311631250428036, 'learning_rate/qwen_vl_interface': 5.067239564989968e-07, 'learning_rate/action_model': 5.067239564989968e-06, 'epoch': 2.02}) +[2026-06-16 04:49:06,155][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 0.005611368454992771, 'train/grad_norm_pre_clip': 2.0176448822021484, 'timing/data': 0.0064881989965215325, 'timing/model': 1.1299411039799452, 'learning_rate/qwen_vl_interface': 5.065629465893808e-07, 'learning_rate/action_model': 5.065629465893807e-06, 'epoch': 2.02}) +[2026-06-16 04:49:10,637][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.005829125177115202, 'train/grad_norm_pre_clip': 4.467721939086914, 'timing/data': 0.0065765620092861354, 'timing/model': 1.136616136005614, 'learning_rate/qwen_vl_interface': 5.064038865275625e-07, 'learning_rate/action_model': 5.0640388652756245e-06, 'epoch': 2.02}) +[2026-06-16 04:49:15,138][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.1528230607509613, 'train/grad_norm_pre_clip': 39.424530029296875, 'timing/data': 0.00713650300167501, 'timing/model': 1.1250706279533915, 'learning_rate/qwen_vl_interface': 5.062467763789242e-07, 'learning_rate/action_model': 5.062467763789242e-06, 'epoch': 2.02}) +[2026-06-16 04:49:19,679][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.0025878611486405134, 'train/grad_norm_pre_clip': 0.9294100403785706, 'timing/data': 0.006642905005719513, 'timing/model': 1.3099397050100379, 'learning_rate/qwen_vl_interface': 5.060916162080496e-07, 'learning_rate/action_model': 5.060916162080496e-06, 'epoch': 2.02}) +[2026-06-16 04:49:24,055][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.006242962088435888, 'train/grad_norm_pre_clip': 1.8999122381210327, 'timing/data': 0.006572665995918214, 'timing/model': 1.1296297180233523, 'learning_rate/qwen_vl_interface': 5.059384060787182e-07, 'learning_rate/action_model': 5.059384060787183e-06, 'epoch': 2.02}) +[2026-06-16 04:49:28,569][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 0.10664648562669754, 'train/grad_norm_pre_clip': 11.557449340820312, 'timing/data': 0.006939072976820171, 'timing/model': 1.1322438530041836, 'learning_rate/qwen_vl_interface': 5.05787146053909e-07, 'learning_rate/action_model': 5.0578714605390904e-06, 'epoch': 2.02}) +[2026-06-16 04:49:33,044][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.0026240823790431023, 'train/grad_norm_pre_clip': 0.8018900752067566, 'timing/data': 0.006385446002241224, 'timing/model': 1.1294379970058799, 'learning_rate/qwen_vl_interface': 5.056378361957998e-07, 'learning_rate/action_model': 5.056378361957998e-06, 'epoch': 2.02}) +[2026-06-16 04:49:37,541][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.002662287326529622, 'train/grad_norm_pre_clip': 0.9531667828559875, 'timing/data': 0.006509879021905363, 'timing/model': 1.1294911499717273, 'learning_rate/qwen_vl_interface': 5.05490476565766e-07, 'learning_rate/action_model': 5.054904765657659e-06, 'epoch': 2.02}) +[2026-06-16 04:49:42,072][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.12449167668819427, 'train/grad_norm_pre_clip': 20.164718627929688, 'timing/data': 0.0065149469883181155, 'timing/model': 1.1296731269685552, 'learning_rate/qwen_vl_interface': 5.053450672243811e-07, 'learning_rate/action_model': 5.0534506722438114e-06, 'epoch': 2.02}) +[2026-06-16 04:49:46,613][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 0.04462757706642151, 'train/grad_norm_pre_clip': 20.162431716918945, 'timing/data': 0.006236176006495953, 'timing/model': 1.2929594850284047, 'learning_rate/qwen_vl_interface': 5.052016082314184e-07, 'learning_rate/action_model': 5.052016082314184e-06, 'epoch': 2.02}) +[2026-06-16 04:49:50,990][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.28105440735816956, 'train/grad_norm_pre_clip': 22.69325065612793, 'timing/data': 0.006197271985001862, 'timing/model': 1.148067592002917, 'learning_rate/qwen_vl_interface': 5.050600996458473e-07, 'learning_rate/action_model': 5.050600996458473e-06, 'epoch': 2.02}) +[2026-06-16 04:49:55,506][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 0.09148247539997101, 'train/grad_norm_pre_clip': 16.546735763549805, 'timing/data': 0.006589090975467116, 'timing/model': 1.1312171939644031, 'learning_rate/qwen_vl_interface': 5.049205415258373e-07, 'learning_rate/action_model': 5.049205415258373e-06, 'epoch': 2.02}) +[2026-06-16 04:50:00,015][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.0004303841560613364, 'train/grad_norm_pre_clip': 0.16390113532543182, 'timing/data': 0.006723937054630369, 'timing/model': 1.1398938070051372, 'learning_rate/qwen_vl_interface': 5.047829339287557e-07, 'learning_rate/action_model': 5.047829339287557e-06, 'epoch': 2.03}) +[2026-06-16 04:50:04,550][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.0034509054385125637, 'train/grad_norm_pre_clip': 2.9387693405151367, 'timing/data': 0.006992130016442388, 'timing/model': 1.1280987960053608, 'learning_rate/qwen_vl_interface': 5.046472769111668e-07, 'learning_rate/action_model': 5.046472769111667e-06, 'epoch': 2.03}) +[2026-06-16 04:50:09,084][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.1070469319820404, 'train/grad_norm_pre_clip': 27.11443519592285, 'timing/data': 0.006599674990866333, 'timing/model': 1.130679149995558, 'learning_rate/qwen_vl_interface': 5.045135705288353e-07, 'learning_rate/action_model': 5.0451357052883525e-06, 'epoch': 2.03}) +[2026-06-16 04:50:13,621][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.19751131534576416, 'train/grad_norm_pre_clip': 12.719229698181152, 'timing/data': 0.006435488990973681, 'timing/model': 1.2897144380258396, 'learning_rate/qwen_vl_interface': 5.043818148367221e-07, 'learning_rate/action_model': 5.043818148367221e-06, 'epoch': 2.03}) +[2026-06-16 04:50:18,077][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.14448554813861847, 'train/grad_norm_pre_clip': 24.312589645385742, 'timing/data': 0.006462354969698936, 'timing/model': 1.2178114909911528, 'learning_rate/qwen_vl_interface': 5.042520098889869e-07, 'learning_rate/action_model': 5.0425200988898685e-06, 'epoch': 2.03}) +[2026-06-16 04:50:22,443][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.0037216502241790295, 'train/grad_norm_pre_clip': 0.7719706892967224, 'timing/data': 0.00661926600150764, 'timing/model': 1.1274716639891267, 'learning_rate/qwen_vl_interface': 5.041241557389879e-07, 'learning_rate/action_model': 5.041241557389879e-06, 'epoch': 2.03}) +[2026-06-16 04:50:26,943][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.2720457911491394, 'train/grad_norm_pre_clip': 31.742666244506836, 'timing/data': 0.006598374980967492, 'timing/model': 1.1352794650010765, 'learning_rate/qwen_vl_interface': 5.039982524392814e-07, 'learning_rate/action_model': 5.039982524392814e-06, 'epoch': 2.03}) +[2026-06-16 04:50:31,479][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.0004567790892906487, 'train/grad_norm_pre_clip': 0.1198815330862999, 'timing/data': 0.006553960032761097, 'timing/model': 1.1327478819875978, 'learning_rate/qwen_vl_interface': 5.038743000416207e-07, 'learning_rate/action_model': 5.0387430004162064e-06, 'epoch': 2.03}) +[2026-06-16 04:50:35,995][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 6.4571198890917e-05, 'train/grad_norm_pre_clip': 0.007525597233325243, 'timing/data': 0.006599707994610071, 'timing/model': 1.1275666930014268, 'learning_rate/qwen_vl_interface': 5.037522985969588e-07, 'learning_rate/action_model': 5.037522985969587e-06, 'epoch': 2.03}) +[2026-06-16 04:50:40,537][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.12721103429794312, 'train/grad_norm_pre_clip': 20.496065139770508, 'timing/data': 0.006472943990956992, 'timing/model': 1.2949074849602766, 'learning_rate/qwen_vl_interface': 5.036322481554451e-07, 'learning_rate/action_model': 5.036322481554451e-06, 'epoch': 2.03}) +[2026-06-16 04:50:45,018][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.12952150404453278, 'train/grad_norm_pre_clip': 42.56122970581055, 'timing/data': 0.006438320968300104, 'timing/model': 1.2497361439745873, 'learning_rate/qwen_vl_interface': 5.035141487664285e-07, 'learning_rate/action_model': 5.035141487664285e-06, 'epoch': 2.03}) +[2026-06-16 04:50:49,395][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.003076286055147648, 'train/grad_norm_pre_clip': 1.8619608879089355, 'timing/data': 0.006431035988498479, 'timing/model': 1.137121744977776, 'learning_rate/qwen_vl_interface': 5.03398000478455e-07, 'learning_rate/action_model': 5.03398000478455e-06, 'epoch': 2.03}) +[2026-06-16 04:50:53,893][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.0008753743604756892, 'train/grad_norm_pre_clip': 0.167978435754776, 'timing/data': 0.006570425990503281, 'timing/model': 1.132233112002723, 'learning_rate/qwen_vl_interface': 5.032838033392686e-07, 'learning_rate/action_model': 5.032838033392685e-06, 'epoch': 2.03}) +[2026-06-16 04:50:58,438][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.00039641905459575355, 'train/grad_norm_pre_clip': 0.0561416894197464, 'timing/data': 0.006439077027607709, 'timing/model': 1.1395232689683326, 'learning_rate/qwen_vl_interface': 5.031715573958111e-07, 'learning_rate/action_model': 5.031715573958111e-06, 'epoch': 2.03}) +[2026-06-16 04:51:02,996][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.011111277155578136, 'train/grad_norm_pre_clip': 4.983334064483643, 'timing/data': 0.006003696005791426, 'timing/model': 1.1337182080023922, 'learning_rate/qwen_vl_interface': 5.030612626942235e-07, 'learning_rate/action_model': 5.030612626942235e-06, 'epoch': 2.03}) +[2026-06-16 04:51:07,530][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.03719890117645264, 'train/grad_norm_pre_clip': 28.233346939086914, 'timing/data': 0.006286919990088791, 'timing/model': 1.285125861002598, 'learning_rate/qwen_vl_interface': 5.029529192798433e-07, 'learning_rate/action_model': 5.029529192798433e-06, 'epoch': 2.03}) +[2026-06-16 04:51:11,915][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.02141905017197132, 'train/grad_norm_pre_clip': 7.136099815368652, 'timing/data': 0.007242561026941985, 'timing/model': 1.1314149019890465, 'learning_rate/qwen_vl_interface': 5.028465271972058e-07, 'learning_rate/action_model': 5.028465271972058e-06, 'epoch': 2.03}) +[2026-06-16 04:51:16,435][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.000926113803870976, 'train/grad_norm_pre_clip': 0.15102964639663696, 'timing/data': 0.006516989960800856, 'timing/model': 1.1321459260070696, 'learning_rate/qwen_vl_interface': 5.027420864900457e-07, 'learning_rate/action_model': 5.027420864900456e-06, 'epoch': 2.03}) +[2026-06-16 04:51:20,966][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.09656380861997604, 'train/grad_norm_pre_clip': 18.065155029296875, 'timing/data': 0.006613889010623097, 'timing/model': 1.131598503969144, 'learning_rate/qwen_vl_interface': 5.026395972012943e-07, 'learning_rate/action_model': 5.0263959720129436e-06, 'epoch': 2.03}) +[2026-06-16 04:51:25,518][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.0008299959008581936, 'train/grad_norm_pre_clip': 0.11069082468748093, 'timing/data': 0.006528545985929668, 'timing/model': 1.1425253999768756, 'learning_rate/qwen_vl_interface': 5.025390593730809e-07, 'learning_rate/action_model': 5.025390593730809e-06, 'epoch': 2.03}) +[2026-06-16 04:51:29,974][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.020541870966553688, 'train/grad_norm_pre_clip': 13.322964668273926, 'timing/data': 0.006557632994372398, 'timing/model': 1.1306208829628304, 'learning_rate/qwen_vl_interface': 5.024404730467325e-07, 'learning_rate/action_model': 5.024404730467325e-06, 'epoch': 2.03}) +[2026-06-16 04:51:34,471][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.09648388624191284, 'train/grad_norm_pre_clip': 12.222673416137695, 'timing/data': 0.0064976110006682575, 'timing/model': 1.248954089009203, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 2.03}) +[2026-06-16 04:51:38,857][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.003036708803847432, 'train/grad_norm_pre_clip': 2.804020404815674, 'timing/data': 0.00639663398033008, 'timing/model': 1.1347314580343664, 'learning_rate/qwen_vl_interface': 5.022491550609311e-07, 'learning_rate/action_model': 5.022491550609311e-06, 'epoch': 2.03}) +[2026-06-16 04:51:43,402][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.24452605843544006, 'train/grad_norm_pre_clip': 32.4775276184082, 'timing/data': 0.006360796978697181, 'timing/model': 1.140160768001806, 'learning_rate/qwen_vl_interface': 5.021564234801213e-07, 'learning_rate/action_model': 5.021564234801213e-06, 'epoch': 2.03}) +[2026-06-16 04:51:47,923][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.005148106254637241, 'train/grad_norm_pre_clip': 2.742279052734375, 'timing/data': 0.006437014963012189, 'timing/model': 1.130235916993115, 'learning_rate/qwen_vl_interface': 5.020656435584636e-07, 'learning_rate/action_model': 5.020656435584635e-06, 'epoch': 2.04}) +[2026-06-16 04:51:52,470][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.059426192194223404, 'train/grad_norm_pre_clip': 30.731544494628906, 'timing/data': 0.006533383973874152, 'timing/model': 1.1275279450346716, 'learning_rate/qwen_vl_interface': 5.019768153332752e-07, 'learning_rate/action_model': 5.019768153332752e-06, 'epoch': 2.04}) +[2026-06-16 04:51:56,968][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.2662700414657593, 'train/grad_norm_pre_clip': 7.172523498535156, 'timing/data': 0.006508029997348785, 'timing/model': 1.1294686119654216, 'learning_rate/qwen_vl_interface': 5.018899388410693e-07, 'learning_rate/action_model': 5.0188993884106935e-06, 'epoch': 2.04}) +[2026-06-16 04:52:01,446][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 0.19096703827381134, 'train/grad_norm_pre_clip': 18.955080032348633, 'timing/data': 0.0065304230083711445, 'timing/model': 1.241671113006305, 'learning_rate/qwen_vl_interface': 5.018050141175576e-07, 'learning_rate/action_model': 5.018050141175575e-06, 'epoch': 2.04}) +[2026-06-16 04:52:05,818][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.23609274625778198, 'train/grad_norm_pre_clip': 11.319302558898926, 'timing/data': 0.0065084550296887755, 'timing/model': 1.1356244999915361, 'learning_rate/qwen_vl_interface': 5.017220411976496e-07, 'learning_rate/action_model': 5.0172204119764965e-06, 'epoch': 2.04}) +[2026-06-16 04:52:10,336][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.05474484711885452, 'train/grad_norm_pre_clip': 33.37425231933594, 'timing/data': 0.0066722200135700405, 'timing/model': 1.1295430809841491, 'learning_rate/qwen_vl_interface': 5.016410201154523e-07, 'learning_rate/action_model': 5.016410201154523e-06, 'epoch': 2.04}) +[2026-06-16 04:52:14,868][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.057547423988580704, 'train/grad_norm_pre_clip': 18.37627601623535, 'timing/data': 0.006134408002253622, 'timing/model': 1.1294013769947924, 'learning_rate/qwen_vl_interface': 5.015619509042709e-07, 'learning_rate/action_model': 5.015619509042709e-06, 'epoch': 2.04}) +[2026-06-16 04:52:19,396][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.002691200701519847, 'train/grad_norm_pre_clip': 0.5645061135292053, 'timing/data': 0.006626627990044653, 'timing/model': 1.1306788810179569, 'learning_rate/qwen_vl_interface': 5.014848335966067e-07, 'learning_rate/action_model': 5.014848335966067e-06, 'epoch': 2.04}) +[2026-06-16 04:52:23,931][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 0.02025005966424942, 'train/grad_norm_pre_clip': 7.082160472869873, 'timing/data': 0.006562177964951843, 'timing/model': 1.291160150023643, 'learning_rate/qwen_vl_interface': 5.014096682241608e-07, 'learning_rate/action_model': 5.014096682241608e-06, 'epoch': 2.04}) +[2026-06-16 04:52:28,316][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.0755002424120903, 'train/grad_norm_pre_clip': 28.677715301513672, 'timing/data': 0.006408924993593246, 'timing/model': 1.1316478960216045, 'learning_rate/qwen_vl_interface': 5.013364548178299e-07, 'learning_rate/action_model': 5.013364548178299e-06, 'epoch': 2.04}) +[2026-06-16 04:52:32,842][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.013568012043833733, 'train/grad_norm_pre_clip': 5.5008111000061035, 'timing/data': 0.006492851010989398, 'timing/model': 1.1300698629929684, 'learning_rate/qwen_vl_interface': 5.012651934077101e-07, 'learning_rate/action_model': 5.0126519340771e-06, 'epoch': 2.04}) +[2026-06-16 04:52:37,283][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 0.007040726952254772, 'train/grad_norm_pre_clip': 2.6350646018981934, 'timing/data': 0.006642869964707643, 'timing/model': 1.1311286099953577, 'learning_rate/qwen_vl_interface': 5.01195884023094e-07, 'learning_rate/action_model': 5.0119588402309395e-06, 'epoch': 2.04}) +[2026-06-16 04:52:41,749][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.11753851175308228, 'train/grad_norm_pre_clip': 24.262557983398438, 'timing/data': 0.0066500219982117414, 'timing/model': 1.1452280840021558, 'learning_rate/qwen_vl_interface': 5.011285266924714e-07, 'learning_rate/action_model': 5.011285266924714e-06, 'epoch': 2.04}) +[2026-06-16 04:52:46,248][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.00020100505207665265, 'train/grad_norm_pre_clip': 0.027539856731891632, 'timing/data': 0.006484590005129576, 'timing/model': 1.1314966340432875, 'learning_rate/qwen_vl_interface': 5.010631214435312e-07, 'learning_rate/action_model': 5.010631214435312e-06, 'epoch': 2.04}) +[2026-06-16 04:52:50,788][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 0.001632435480132699, 'train/grad_norm_pre_clip': 0.4232442080974579, 'timing/data': 0.006110787042416632, 'timing/model': 1.2952575610252097, 'learning_rate/qwen_vl_interface': 5.009996683031586e-07, 'learning_rate/action_model': 5.009996683031586e-06, 'epoch': 2.04}) +[2026-06-16 04:52:55,164][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.08271954208612442, 'train/grad_norm_pre_clip': 36.17345428466797, 'timing/data': 0.006528530037030578, 'timing/model': 1.1316752129932866, 'learning_rate/qwen_vl_interface': 5.009381672974375e-07, 'learning_rate/action_model': 5.009381672974375e-06, 'epoch': 2.04}) +[2026-06-16 04:52:59,681][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.04994755983352661, 'train/grad_norm_pre_clip': 30.44283676147461, 'timing/data': 0.006844542978797108, 'timing/model': 1.1299582140054554, 'learning_rate/qwen_vl_interface': 5.008786184516475e-07, 'learning_rate/action_model': 5.008786184516474e-06, 'epoch': 2.04}) +[2026-06-16 04:53:04,195][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 0.0037413937970995903, 'train/grad_norm_pre_clip': 1.7735131978988647, 'timing/data': 0.006329483003355563, 'timing/model': 1.129448397026863, 'learning_rate/qwen_vl_interface': 5.008210217902676e-07, 'learning_rate/action_model': 5.008210217902676e-06, 'epoch': 2.04}) +[2026-06-16 04:53:08,679][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.1778450608253479, 'train/grad_norm_pre_clip': 15.446285247802734, 'timing/data': 0.006489345978479832, 'timing/model': 1.1358557969797403, 'learning_rate/qwen_vl_interface': 5.007653773369737e-07, 'learning_rate/action_model': 5.007653773369736e-06, 'epoch': 2.04}) +[2026-06-16 04:53:13,175][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.056422144174575806, 'train/grad_norm_pre_clip': 26.69470977783203, 'timing/data': 0.006383527012076229, 'timing/model': 1.1341865359572694, 'learning_rate/qwen_vl_interface': 5.007116851146386e-07, 'learning_rate/action_model': 5.007116851146386e-06, 'epoch': 2.04}) +[2026-06-16 04:53:17,725][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.14605043828487396, 'train/grad_norm_pre_clip': 11.05813217163086, 'timing/data': 0.006490932020824403, 'timing/model': 1.3035144840250723, 'learning_rate/qwen_vl_interface': 5.006599451453334e-07, 'learning_rate/action_model': 5.006599451453334e-06, 'epoch': 2.04}) +[2026-06-16 04:53:22,099][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.06025483086705208, 'train/grad_norm_pre_clip': 13.267159461975098, 'timing/data': 0.00639407301787287, 'timing/model': 1.1265193090075627, 'learning_rate/qwen_vl_interface': 5.006101574503268e-07, 'learning_rate/action_model': 5.006101574503267e-06, 'epoch': 2.04}) +[2026-06-16 04:53:26,695][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.0020685268100351095, 'train/grad_norm_pre_clip': 0.441562294960022, 'timing/data': 0.006889526033774018, 'timing/model': 1.1326751690357924, 'learning_rate/qwen_vl_interface': 5.005623220500839e-07, 'learning_rate/action_model': 5.005623220500839e-06, 'epoch': 2.04}) +[2026-06-16 04:53:31,247][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.0005939070833846927, 'train/grad_norm_pre_clip': 0.15135306119918823, 'timing/data': 0.006546953984070569, 'timing/model': 1.1332974639954045, 'learning_rate/qwen_vl_interface': 5.005164389642687e-07, 'learning_rate/action_model': 5.005164389642687e-06, 'epoch': 2.04}) +[2026-06-16 04:53:35,796][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.0020412064623087645, 'train/grad_norm_pre_clip': 0.6048673391342163, 'timing/data': 0.006410597998183221, 'timing/model': 1.14063099300256, 'learning_rate/qwen_vl_interface': 5.004725082117423e-07, 'learning_rate/action_model': 5.004725082117422e-06, 'epoch': 2.04}) +[2026-06-16 04:53:40,314][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.07765638083219528, 'train/grad_norm_pre_clip': 11.740547180175781, 'timing/data': 0.006772484979592264, 'timing/model': 1.1361316699767485, 'learning_rate/qwen_vl_interface': 5.004305298105616e-07, 'learning_rate/action_model': 5.004305298105616e-06, 'epoch': 2.05}) +[2026-06-16 04:53:44,867][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.002955466276034713, 'train/grad_norm_pre_clip': 0.6526197195053101, 'timing/data': 0.006446768005844206, 'timing/model': 1.2936689490452409, 'learning_rate/qwen_vl_interface': 5.003905037779834e-07, 'learning_rate/action_model': 5.0039050377798345e-06, 'epoch': 2.05}) +[2026-06-16 04:53:49,273][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.09670767933130264, 'train/grad_norm_pre_clip': 7.4008283615112305, 'timing/data': 0.006496971996966749, 'timing/model': 1.1408231150126085, 'learning_rate/qwen_vl_interface': 5.003524301304608e-07, 'learning_rate/action_model': 5.003524301304608e-06, 'epoch': 2.05}) +[2026-06-16 04:53:53,801][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.009575499221682549, 'train/grad_norm_pre_clip': 6.422959327697754, 'timing/data': 0.006338697043247521, 'timing/model': 1.1297647160245106, 'learning_rate/qwen_vl_interface': 5.003163088836445e-07, 'learning_rate/action_model': 5.003163088836446e-06, 'epoch': 2.05}) +[2026-06-16 04:53:58,329][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.09950827062129974, 'train/grad_norm_pre_clip': 28.208171844482422, 'timing/data': 0.006920452986378223, 'timing/model': 1.1353166409535334, 'learning_rate/qwen_vl_interface': 5.002821400523818e-07, 'learning_rate/action_model': 5.002821400523818e-06, 'epoch': 2.05}) +[2026-06-16 04:54:02,837][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 0.004479830153286457, 'train/grad_norm_pre_clip': 2.6353960037231445, 'timing/data': 0.006518614012748003, 'timing/model': 1.1294349350500852, 'learning_rate/qwen_vl_interface': 5.002499236507192e-07, 'learning_rate/action_model': 5.0024992365071925e-06, 'epoch': 2.05}) +[2026-06-16 04:54:07,368][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.08108780533075333, 'train/grad_norm_pre_clip': 17.565866470336914, 'timing/data': 0.006471936998423189, 'timing/model': 1.132030731998384, 'learning_rate/qwen_vl_interface': 5.002196596918993e-07, 'learning_rate/action_model': 5.002196596918993e-06, 'epoch': 2.05}) +[2026-06-16 04:54:11,904][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.011163868941366673, 'train/grad_norm_pre_clip': 6.3894476890563965, 'timing/data': 0.005975403008051217, 'timing/model': 1.2870284849777818, 'learning_rate/qwen_vl_interface': 5.001913481883623e-07, 'learning_rate/action_model': 5.001913481883623e-06, 'epoch': 2.05}) +[2026-06-16 04:54:16,296][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.003675631945952773, 'train/grad_norm_pre_clip': 1.286982536315918, 'timing/data': 0.006495106965303421, 'timing/model': 1.144034258031752, 'learning_rate/qwen_vl_interface': 5.001649891517459e-07, 'learning_rate/action_model': 5.001649891517459e-06, 'epoch': 2.05}) +[2026-06-16 04:54:20,847][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.012556793168187141, 'train/grad_norm_pre_clip': 6.352806091308594, 'timing/data': 0.0059645610162988305, 'timing/model': 1.1354165369994007, 'learning_rate/qwen_vl_interface': 5.001405825928858e-07, 'learning_rate/action_model': 5.0014058259288575e-06, 'epoch': 2.05}) +[2026-06-16 04:54:25,337][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.1521778106689453, 'train/grad_norm_pre_clip': 14.39830207824707, 'timing/data': 0.007043584017083049, 'timing/model': 1.132069006038364, 'learning_rate/qwen_vl_interface': 5.001181285218143e-07, 'learning_rate/action_model': 5.0011812852181434e-06, 'epoch': 2.05}) +[2026-06-16 04:54:29,832][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.1211632788181305, 'train/grad_norm_pre_clip': 9.68549919128418, 'timing/data': 0.006106088985688984, 'timing/model': 1.1315673509961925, 'learning_rate/qwen_vl_interface': 5.000976269477613e-07, 'learning_rate/action_model': 5.0009762694776125e-06, 'epoch': 2.05}) +[2026-06-16 04:54:34,399][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.1340682953596115, 'train/grad_norm_pre_clip': 10.358121871948242, 'timing/data': 0.0064619629993103445, 'timing/model': 1.145920881011989, 'learning_rate/qwen_vl_interface': 5.000790778791542e-07, 'learning_rate/action_model': 5.000790778791543e-06, 'epoch': 2.05}) +[2026-06-16 04:54:38,979][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.09330178797245026, 'train/grad_norm_pre_clip': 28.213552474975586, 'timing/data': 0.006765566999092698, 'timing/model': 1.3218858070322312, 'learning_rate/qwen_vl_interface': 5.000624813236185e-07, 'learning_rate/action_model': 5.000624813236185e-06, 'epoch': 2.05}) +[2026-06-16 04:54:43,367][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.0024438651744276285, 'train/grad_norm_pre_clip': 0.6227917075157166, 'timing/data': 0.006574868981260806, 'timing/model': 1.1317847680184059, 'learning_rate/qwen_vl_interface': 5.000478372879752e-07, 'learning_rate/action_model': 5.000478372879752e-06, 'epoch': 2.05}) +[2026-06-16 04:54:47,903][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.0037708207964897156, 'train/grad_norm_pre_clip': 0.8970387578010559, 'timing/data': 0.006263717019464821, 'timing/model': 1.1360915009863675, 'learning_rate/qwen_vl_interface': 5.000351457782452e-07, 'learning_rate/action_model': 5.000351457782451e-06, 'epoch': 2.05}) +[2026-06-16 04:54:52,419][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 0.17876242101192474, 'train/grad_norm_pre_clip': 14.379547119140625, 'timing/data': 0.006025251001119614, 'timing/model': 1.1415408609900624, 'learning_rate/qwen_vl_interface': 5.000244067996445e-07, 'learning_rate/action_model': 5.000244067996445e-06, 'epoch': 2.05}) +[2026-06-16 04:54:56,943][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.009407621808350086, 'train/grad_norm_pre_clip': 4.890580177307129, 'timing/data': 0.005987454031128436, 'timing/model': 1.1449803300201893, 'learning_rate/qwen_vl_interface': 5.000156203565884e-07, 'learning_rate/action_model': 5.000156203565883e-06, 'epoch': 2.05}) +[2026-06-16 04:55:01,488][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.07552536576986313, 'train/grad_norm_pre_clip': 28.957359313964844, 'timing/data': 0.007029750035144389, 'timing/model': 1.1464709490537643, 'learning_rate/qwen_vl_interface': 5.000087864526878e-07, 'learning_rate/action_model': 5.000087864526878e-06, 'epoch': 2.05}) +[2026-06-16 04:55:06,035][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.000503353017847985, 'train/grad_norm_pre_clip': 0.1250365823507309, 'timing/data': 0.006734704016707838, 'timing/model': 1.29554427403491, 'learning_rate/qwen_vl_interface': 5.000039050907523e-07, 'learning_rate/action_model': 5.000039050907523e-06, 'epoch': 2.05}) +[2026-06-16 04:55:10,411][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.0006941739120520651, 'train/grad_norm_pre_clip': 0.14248904585838318, 'timing/data': 0.006790254032239318, 'timing/model': 1.1322482830146328, 'learning_rate/qwen_vl_interface': 5.000009762727885e-07, 'learning_rate/action_model': 5.000009762727884e-06, 'epoch': 2.05}) +[2026-06-16 04:59:34,743][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.1215958446264267, 'train/grad_norm_pre_clip': 61.09042739868164, 'eval/loss': 0.038121454417705536, 'eval/flap_precision': 0.9724770642201835, 'eval/flap_recall': 0.9814814814814815, 'eval/flap_f1': 0.9769585253456222, 'eval/flap_rate_model': 0.18166666666666667, 'eval/flap_rate_teacher': 0.18, 'eval/flap_cc_f1': 0.9769585253456223, 'timing/data': 0.006809791026171297, 'timing/model': 1.137486633961089, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 2.05}) +[2026-06-16 04:59:34,743][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_5000_state +[2026-06-16 04:59:34,751][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 04:59:38,305][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_5000_state/model.safetensors +[2026-06-16 04:59:45,557][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_5000_state/optimizer.bin +[2026-06-16 04:59:45,557][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_5000_state/sampler.bin +[2026-06-16 04:59:45,557][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_5000_state/sampler_1.bin +[2026-06-16 04:59:45,558][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints/steps_5000_state/random_states_0.pkl +[2026-06-16 04:59:45,653][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-06-16 04:59:45,653][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 04:59:45,664][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 04:59:45,664][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /workspace/starVLA/results/Checkpoints/flappy_fix_latency_4_200ep/checkpoints +[2026-06-16 04:59:48,693][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..ffa1ed16fa7cd08f0c83739d659e9d2a8bd4e827 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_4_200ep/summary.jsonl @@ -0,0 +1,50 @@ +{"steps": 100} +{"steps": 200} +{"steps": 300} +{"steps": 400} +{"steps": 500} +{"steps": 600} +{"steps": 700} +{"steps": 800} +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500} +{"steps": 1600} +{"steps": 1700} +{"steps": 1800} +{"steps": 1900} +{"steps": 2000} +{"steps": 2100} +{"steps": 2200} +{"steps": 2300} +{"steps": 2400} +{"steps": 2500} +{"steps": 2600} +{"steps": 2700} +{"steps": 2800} +{"steps": 2900} +{"steps": 3000} +{"steps": 3100} +{"steps": 3200} +{"steps": 3300} +{"steps": 3400} +{"steps": 3500} +{"steps": 3600} +{"steps": 3700} +{"steps": 3800} +{"steps": 3900} +{"steps": 4000} +{"steps": 4100} +{"steps": 4200} +{"steps": 4300} +{"steps": 4400} +{"steps": 4500} +{"steps": 4600} +{"steps": 4700} +{"steps": 4800} +{"steps": 4900} +{"steps": 5000} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/checkpoints/best_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/checkpoints/best_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..438b05c05fbd1b7192d0a56de824fb496fcc2e54 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/checkpoints/best_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:758b4ccf870860c583af4da492b2d6782333e847759afd7314ba4547610fb04e +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/checkpoints/steps_1500_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/checkpoints/steps_1500_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..99a6dd4530ab3cbe1e569d0fed1f3d21e4a81349 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/checkpoints/steps_1500_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0ecf34ec797bcb608fabd5fee1b994bf8480dfd6d75eb302b9a5380cf9d69151 +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..0bed65d4d881d6be1a170dae955f6a77fbdc877e --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/config.full.yaml @@ -0,0 +1,187 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_fix_latency_1_small + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 1500 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/starVLA/data/flappy_fix_latency_1_small/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 1 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: true + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_flappy_fix_latency_1_small +output_dir: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_small +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..b7f6b4ff7d2d0d962b32f420331101160aa6f4dc --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/config.yaml @@ -0,0 +1,89 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: true + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + include_state: true + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_small +rl_games: + env_eval: + distributed_mode: none + enabled: true + fixed_episode_seeds: true + frameskip: 1 + image_size: 224 + latency: + prompt_map_path: /workspace/starVLA/data/flappy_fix_latency_1_small/flappy_train__bridge/latency_prompt_map.json + latency_seed_stride: 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 1 + max_steps_per_episode: 3600 + num_episodes: 5 + seed: 42 + task_description: '' + task_seed_stride: 0 + vectorized: + enabled: false + model_alias: openvla + task: flappy +run_id: openvla_flappy_fix_latency_1_small +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_interval: 100 + eval_num_batches: 20 + freeze_modules: '' + gradient_accumulation_steps: 16 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 10 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 1500 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + pretrained_checkpoint: null + reload_modules: null + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..807f788a7b0627d1de0d3d2bfa69f64f493e704f --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8600514531135559, + 0.13994856178760529, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.3468538522720337, + 0.3468538522720337, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 97200, + "num_trajectories": 27 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..1a730bf8f6e51cae84ff205788fa5286afaae5de --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8612037301063538, + 0.13879629969596863, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.34575334191322327, + 0.34575334191322327, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 10800, + "num_trajectories": 3 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/hydra/.hydra/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..d27703f2fc14cbe2d035727185cdece4857f7648 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/hydra/.hydra/config.yaml @@ -0,0 +1,185 @@ +framework: + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_fix_latency_1_tmp + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 1500 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/starVLA/data/flappy_fix_latency_1_tmp/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 1 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: true + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_flappy_fix_latency_1_tmp +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/hydra/.hydra/hydra.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..e00e5de97393a233fc7d8a1333660d4518fc97db --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/hydra/.hydra/hydra.yaml @@ -0,0 +1,246 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=single + - run_id=openvla_flappy_fix_latency_1_tmp + - run_root_dir=/workspace/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=true + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=1500 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=20 + - trainer.logging_frequency=10 + - trainer.gradient_accumulation_steps=16 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9, 0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=2 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1_tmp + - datasets.vla_data.data_mix=flappy_train__bridge + - datasets.vla_data.eval_data_mix=flappy_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=flappy + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=1 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_1_tmp/flappy_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=true + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[1] + - rl_games.env_eval.post_train.enabled=true + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=true,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=flappy_train__bridge,datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1_tmp,datasets.vla_data.eval_data_mix=flappy_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.video_backend=torchvision_av,env=flappy,framework.action_model.action_dim=7,framework.action_model.action_env_dim=2,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=1,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_1_tmp/flappy_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=true,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[1],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=true,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=flappy,run_id=openvla_flappy_fix_latency_1_tmp,run_root_dir=/workspace/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=20,trainer.gradient_accumulation_steps=16,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=10,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=1500,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9, + 0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.2 + version_base: '1.1' + cwd: /workspace/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/hydra/.hydra/overrides.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..e139bda57af5647ec4fbce1ea3228fa1eaefd093 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/hydra/.hydra/overrides.yaml @@ -0,0 +1,85 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=single +- run_id=openvla_flappy_fix_latency_1_tmp +- run_root_dir=/workspace/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=true +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=1500 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=20 +- trainer.logging_frequency=10 +- trainer.gradient_accumulation_steps=16 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9, 0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=2 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1_tmp +- datasets.vla_data.data_mix=flappy_train__bridge +- datasets.vla_data.eval_data_mix=flappy_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=flappy +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=1 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_1_tmp/flappy_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=true +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[1] +- rl_games.env_eval.post_train.enabled=true +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/hydra/train_starvla_hydra.log b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..6fee91bfb5a697cf2f795bfb82de6d00b1263a6b --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/hydra/train_starvla_hydra.log @@ -0,0 +1,423 @@ +[2026-06-08 09:52:33,898][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-08 09:52:33,916][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-08 09:52:44,844][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-08 09:52:44,887][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-08 09:52:45,405][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-08 09:52:45,405][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-08 09:52:45,405][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-08 09:52:45,405][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-08 09:52:45,405][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-08 09:52:45,405][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-08 09:52:45,405][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-08 09:52:45,405][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-08 09:52:45,405][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-08 09:52:45,405][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-08 09:52:45,405][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-08 09:52:45,406][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-08 09:52:45,406][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-08 09:52:45,406][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-08 09:52:45,406][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-08 09:52:45,406][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-08 09:52:45,406][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-08 09:52:45,406][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-08 09:52:45,406][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-08 09:52:45,406][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-08 09:52:45,406][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-08 09:52:45,406][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-08 09:52:45,406][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-08 09:52:45,406][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-08 09:52:45,406][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-08 09:52:45,407][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-08 09:52:45,407][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-08 09:52:45,407][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-08 09:52:45,407][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-08 09:52:45,407][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-08 09:52:45,407][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-08 09:52:45,423][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-08 09:52:45,447][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/config.full.yaml +[2026-06-08 09:52:45,453][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/config.yaml +[2026-06-08 09:52:45,453][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-08 09:52:49,821][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-08 09:52:49,821][starVLA.training.train_starvla][INFO] - Total optimization steps = 1500 +[2026-06-08 09:52:49,821][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-08 09:52:49,821][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-08 09:52:49,821][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-08 09:55:29,313][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.40828171372413635, 'train/grad_norm_pre_clip': 1.1354433298110962, 'timing/data': 0.011844784021377563, 'timing/model': 1.0445850491523743, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-08 09:58:07,230][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.41072672605514526, 'train/grad_norm_pre_clip': 0.45467856526374817, 'timing/data': 0.012608123943209648, 'timing/model': 1.0532283503562212, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-08 10:00:49,906][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.5215868353843689, 'train/grad_norm_pre_clip': 0.1329498589038849, 'timing/data': 0.013266224414110184, 'timing/model': 1.1390961725264788, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-08 10:03:32,116][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.6223028898239136, 'train/grad_norm_pre_clip': 0.9096503853797913, 'timing/data': 0.012258810922503471, 'timing/model': 1.0658473186194897, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.01}) +[2026-06-08 10:06:16,222][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.4728282392024994, 'train/grad_norm_pre_clip': 0.15153095126152039, 'timing/data': 0.014603586867451668, 'timing/model': 1.1033218763768673, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.02}) +[2026-06-08 10:09:00,320][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.2123694270849228, 'train/grad_norm_pre_clip': 1.1626359224319458, 'timing/data': 0.012150054797530174, 'timing/model': 1.049701688811183, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.02}) +[2026-06-08 10:11:42,809][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.3840845227241516, 'train/grad_norm_pre_clip': 3.9510371685028076, 'timing/data': 0.012099288403987885, 'timing/model': 1.0562160797417164, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.02}) +[2026-06-08 10:14:25,024][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.27131402492523193, 'train/grad_norm_pre_clip': 1.77262544631958, 'timing/data': 0.011260692030191422, 'timing/model': 1.0527115743607283, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.03}) +[2026-06-08 10:17:06,780][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.07267160713672638, 'train/grad_norm_pre_clip': 2.2838242053985596, 'timing/data': 0.011682651937007904, 'timing/model': 1.301352933049202, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.03}) +[2026-06-08 10:20:14,779][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpqpbes7sg/test.c -o /tmp/tmpqpbes7sg/test.o +[2026-06-08 10:20:14,803][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpqpbes7sg/test.o -laio -o /tmp/tmpqpbes7sg/a.out +[2026-06-08 10:20:15,396][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp_r40bif8/test.c -o /tmp/tmp_r40bif8/test.o +[2026-06-08 10:20:15,418][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp_r40bif8/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp_r40bif8/a.out +[2026-06-08 10:20:15,476][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpmv84u39j/test.c -o /tmp/tmpmv84u39j/test.o +[2026-06-08 10:20:15,496][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpmv84u39j/test.o -laio -o /tmp/tmpmv84u39j/a.out +[2026-06-08 10:20:27,409][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state +[2026-06-08 10:20:27,413][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 10:20:33,962][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/model.safetensors +[2026-06-08 10:20:44,522][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/optimizer.bin +[2026-06-08 10:20:44,522][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/sampler.bin +[2026-06-08 10:20:44,523][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/sampler_1.bin +[2026-06-08 10:20:44,523][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/random_states_0.pkl +[2026-06-08 10:20:44,536][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 10:20:44,536][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 100 +[2026-06-08 10:20:44,538][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.18347503244876862, 'train/grad_norm_pre_clip': 1.2692731618881226, 'eval/loss': 0.2526457905769348, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 2.0000000000000013, 'rl_games_eval/mid_train/mean_length': 31.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 2.0000000000000013, 'rl_games_eval/mid_train/macro_mean_length': 31.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 2.0000000000000013, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 31.0, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 2.0000000000000013, 'checkpoint/best_score': 2.0000000000000013, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.010717760771512985, 'timing/model': 1.053620195016265, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.03}) +[2026-06-08 10:20:44,539][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_100_state +[2026-06-08 10:20:44,542][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 10:20:49,721][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_100_state/model.safetensors +[2026-06-08 10:20:58,985][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_100_state/optimizer.bin +[2026-06-08 10:20:58,986][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_100_state/sampler.bin +[2026-06-08 10:20:58,986][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_100_state/sampler_1.bin +[2026-06-08 10:20:58,987][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-08 10:20:58,990][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-08 10:20:58,990][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 10:20:59,001][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 10:23:57,861][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.1048712208867073, 'train/grad_norm_pre_clip': 0.9115661382675171, 'timing/data': 0.010702500119805336, 'timing/model': 1.320005003362894, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.04}) +[2026-06-08 10:26:57,694][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.8739692568778992, 'train/grad_norm_pre_clip': 3.8661139011383057, 'timing/data': 0.010839132592082024, 'timing/model': 1.1564060263335705, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.04}) +[2026-06-08 10:29:52,552][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.19683970510959625, 'train/grad_norm_pre_clip': 2.547755718231201, 'timing/data': 0.007842408493161201, 'timing/model': 1.1465910617262125, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.04}) +[2026-06-08 10:32:50,607][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.40366363525390625, 'train/grad_norm_pre_clip': 6.8154168128967285, 'timing/data': 0.011074485257267952, 'timing/model': 1.1667950041592121, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.05}) +[2026-06-08 10:35:51,815][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.25993427634239197, 'train/grad_norm_pre_clip': 2.053656578063965, 'timing/data': 0.011117670685052872, 'timing/model': 1.2421472780406475, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.05}) +[2026-06-08 10:38:51,822][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.3508963882923126, 'train/grad_norm_pre_clip': 3.0156335830688477, 'timing/data': 0.01371392235159874, 'timing/model': 1.2619608733803034, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.05}) +[2026-06-08 10:41:52,950][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.09928127378225327, 'train/grad_norm_pre_clip': 0.37923699617385864, 'timing/data': 0.009218508377671242, 'timing/model': 1.1275597494095564, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.06}) +[2026-06-08 10:44:51,304][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.1853797733783722, 'train/grad_norm_pre_clip': 1.20838463306427, 'timing/data': 0.009410152211785316, 'timing/model': 1.132815595716238, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.06}) +[2026-06-08 10:47:55,322][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.39642465114593506, 'train/grad_norm_pre_clip': 2.4470624923706055, 'timing/data': 0.0014955457299947739, 'timing/model': 1.1906365994364023, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.06}) +[2026-06-08 10:53:30,311][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state +[2026-06-08 10:53:30,315][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 10:53:36,279][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/model.safetensors +[2026-06-08 10:53:44,913][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/optimizer.bin +[2026-06-08 10:53:44,913][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/sampler.bin +[2026-06-08 10:53:44,913][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/sampler_1.bin +[2026-06-08 10:53:44,914][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/random_states_0.pkl +[2026-06-08 10:53:44,927][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 10:53:44,927][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 200 +[2026-06-08 10:53:44,928][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.16591592133045197, 'train/grad_norm_pre_clip': 1.941574215888977, 'eval/loss': 0.15176500380039215, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 46.59999999999995, 'rl_games_eval/mid_train/mean_length': 394.2, 'rl_games_eval/mid_train/std_reward': 31.108841187032105, 'rl_games_eval/mid_train/std_length': 251.01107545285726, 'rl_games_eval/mid_train/macro_mean_reward': 46.59999999999995, 'rl_games_eval/mid_train/macro_mean_length': 394.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 46.59999999999995, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 394.2, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 31.108841187032105, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 251.01107545285726, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 46.59999999999995, 'checkpoint/best_score': 46.59999999999995, 'checkpoint/best_step': 200.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.020654525607824326, 'timing/model': 1.2551338635385036, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.07}) +[2026-06-08 10:53:44,929][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_200_state +[2026-06-08 10:53:44,932][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 10:53:50,220][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_200_state/model.safetensors +[2026-06-08 10:53:59,756][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_200_state/optimizer.bin +[2026-06-08 10:53:59,757][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_200_state/sampler.bin +[2026-06-08 10:53:59,757][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_200_state/sampler_1.bin +[2026-06-08 10:53:59,758][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-08 10:53:59,857][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-08 10:53:59,857][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 10:53:59,868][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 10:57:00,991][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.21215619146823883, 'train/grad_norm_pre_clip': 1.5735502243041992, 'timing/data': 0.011382108554244041, 'timing/model': 1.2290577311068773, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.07}) +[2026-06-08 11:00:01,613][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.34320780634880066, 'train/grad_norm_pre_clip': 9.839859962463379, 'timing/data': 0.011111687868833542, 'timing/model': 1.1662136279046535, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.07}) +[2026-06-08 11:03:02,780][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.1027197539806366, 'train/grad_norm_pre_clip': 0.6264110803604126, 'timing/data': 0.009788326919078827, 'timing/model': 1.5143594779074192, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.08}) +[2026-06-08 11:06:04,394][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.17158649861812592, 'train/grad_norm_pre_clip': 1.1879379749298096, 'timing/data': 0.011424532160162926, 'timing/model': 1.1582642421126366, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.08}) +[2026-06-08 11:09:04,660][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.0929827094078064, 'train/grad_norm_pre_clip': 1.392678141593933, 'timing/data': 0.008170314133167267, 'timing/model': 1.140365170314908, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.08}) +[2026-06-08 11:12:01,639][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.1152813509106636, 'train/grad_norm_pre_clip': 1.0940498113632202, 'timing/data': 0.007798086851835251, 'timing/model': 1.1291438322514296, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.09}) +[2026-06-08 11:15:02,552][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.17576448619365692, 'train/grad_norm_pre_clip': 0.6876584887504578, 'timing/data': 0.011525958776473999, 'timing/model': 1.205269007012248, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.09}) +[2026-06-08 11:18:04,967][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.1872209906578064, 'train/grad_norm_pre_clip': 1.3800740242004395, 'timing/data': 0.010088689625263214, 'timing/model': 1.1354317180812359, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.09}) +[2026-06-08 11:21:05,212][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.08540508896112442, 'train/grad_norm_pre_clip': 0.1933252513408661, 'timing/data': 0.010362962260842323, 'timing/model': 1.3402665108442307, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.1}) +[2026-06-08 11:30:52,660][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state +[2026-06-08 11:30:52,671][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 11:31:00,913][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/model.safetensors +[2026-06-08 11:31:12,590][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/optimizer.bin +[2026-06-08 11:31:12,592][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/sampler.bin +[2026-06-08 11:31:12,592][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/sampler_1.bin +[2026-06-08 11:31:12,594][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/random_states_0.pkl +[2026-06-08 11:31:12,614][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 11:31:12,614][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 300 +[2026-06-08 11:31:12,616][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.34701982140541077, 'train/grad_norm_pre_clip': 2.688359260559082, 'eval/loss': 0.12940405309200287, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 144.35999999999734, 'rl_games_eval/mid_train/mean_length': 1182.8, 'rl_games_eval/mid_train/std_reward': 77.08391790769156, 'rl_games_eval/mid_train/std_length': 620.6501107709561, 'rl_games_eval/mid_train/macro_mean_reward': 144.35999999999734, 'rl_games_eval/mid_train/macro_mean_length': 1182.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 144.35999999999734, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 1182.8, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 77.08391790769156, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 620.6501107709561, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 144.35999999999734, 'checkpoint/best_score': 144.35999999999734, 'checkpoint/best_step': 300.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.009587131440639496, 'timing/model': 1.1474252417683601, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.1}) +[2026-06-08 11:31:12,617][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_300_state +[2026-06-08 11:31:12,624][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 11:31:19,053][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_300_state/model.safetensors +[2026-06-08 11:31:28,699][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_300_state/optimizer.bin +[2026-06-08 11:31:28,701][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_300_state/sampler.bin +[2026-06-08 11:31:28,701][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_300_state/sampler_1.bin +[2026-06-08 11:31:28,702][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-08 11:31:28,812][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-08 11:31:28,813][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 11:31:28,826][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 11:34:28,706][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.19335272908210754, 'train/grad_norm_pre_clip': 0.9630542993545532, 'timing/data': 0.009828619658946991, 'timing/model': 1.1582351867109537, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.1}) +[2026-06-08 11:37:30,926][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.09006249904632568, 'train/grad_norm_pre_clip': 0.3886520564556122, 'timing/data': 0.00789446011185646, 'timing/model': 1.1222459468990564, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.11}) +[2026-06-08 11:40:29,798][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.18327032029628754, 'train/grad_norm_pre_clip': 2.6054799556732178, 'timing/data': 0.009751124307513237, 'timing/model': 1.1732524242252111, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.11}) +[2026-06-08 11:43:28,533][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.15537501871585846, 'train/grad_norm_pre_clip': 0.48919737339019775, 'timing/data': 0.011598827317357063, 'timing/model': 1.2018820270895958, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.11}) +[2026-06-08 11:46:27,569][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.19547389447689056, 'train/grad_norm_pre_clip': 3.4357895851135254, 'timing/data': 0.01413719542324543, 'timing/model': 1.1314861960709095, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.12}) +[2026-06-08 11:49:27,434][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.14235472679138184, 'train/grad_norm_pre_clip': 4.165241718292236, 'timing/data': 0.010185010731220245, 'timing/model': 1.1710426099598408, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.12}) +[2026-06-08 11:52:29,626][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.07515958696603775, 'train/grad_norm_pre_clip': 1.0297025442123413, 'timing/data': 0.010081151500344276, 'timing/model': 1.1725254096090794, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.12}) +[2026-06-08 11:55:35,409][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.023333124816417694, 'train/grad_norm_pre_clip': 0.3066554367542267, 'timing/data': 0.03866436332464218, 'timing/model': 1.1491819117218256, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.13}) +[2026-06-08 11:58:36,737][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.1247081607580185, 'train/grad_norm_pre_clip': 0.6785833835601807, 'timing/data': 0.013108933344483376, 'timing/model': 1.132403252646327, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.13}) +[2026-06-08 12:05:38,049][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 400 +[2026-06-08 12:05:38,051][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.13966120779514313, 'train/grad_norm_pre_clip': 1.2877564430236816, 'eval/loss': 0.11761720478534698, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 78.41999999999919, 'rl_games_eval/mid_train/mean_length': 651.2, 'rl_games_eval/mid_train/std_reward': 42.345927785324214, 'rl_games_eval/mid_train/std_length': 340.8163141635095, 'rl_games_eval/mid_train/macro_mean_reward': 78.41999999999919, 'rl_games_eval/mid_train/macro_mean_length': 651.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 78.41999999999919, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 651.2, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 42.345927785324214, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 340.8163141635095, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 78.41999999999919, 'checkpoint/best_score': 144.35999999999734, 'checkpoint/best_step': 300.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.010706352069973946, 'timing/model': 1.1533402558416128, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.13}) +[2026-06-08 12:05:38,052][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_400_state +[2026-06-08 12:05:38,059][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 12:05:49,852][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_400_state/model.safetensors +[2026-06-08 12:06:02,828][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_400_state/optimizer.bin +[2026-06-08 12:06:02,829][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_400_state/sampler.bin +[2026-06-08 12:06:02,829][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_400_state/sampler_1.bin +[2026-06-08 12:06:02,830][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-08 12:06:02,935][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-08 12:06:02,936][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 12:06:02,950][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 12:09:02,209][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.08362150937318802, 'train/grad_norm_pre_clip': 0.531930685043335, 'timing/data': 0.007943637669086456, 'timing/model': 1.2603696510195732, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.13}) +[2026-06-08 12:12:02,580][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.10252087563276291, 'train/grad_norm_pre_clip': 0.9589378237724304, 'timing/data': 0.008619610220193863, 'timing/model': 1.1800383478403091, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.14}) +[2026-06-08 12:15:02,739][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.20455266535282135, 'train/grad_norm_pre_clip': 1.4990754127502441, 'timing/data': 0.008824523538351059, 'timing/model': 1.1544593144208193, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 0.14}) +[2026-06-08 12:18:04,939][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.0871671661734581, 'train/grad_norm_pre_clip': 1.468031883239746, 'timing/data': 0.010527761653065681, 'timing/model': 1.1809714436531067, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.14}) +[2026-06-08 12:21:06,302][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.25888070464134216, 'train/grad_norm_pre_clip': 0.8412041664123535, 'timing/data': 0.01042548380792141, 'timing/model': 1.156168045476079, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.15}) +[2026-06-08 12:24:04,412][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.1996796727180481, 'train/grad_norm_pre_clip': 4.284126281738281, 'timing/data': 0.009774412959814072, 'timing/model': 1.1632424984127283, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.15}) +[2026-06-08 12:27:05,298][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.10147562623023987, 'train/grad_norm_pre_clip': 1.7426236867904663, 'timing/data': 0.009528091177344322, 'timing/model': 1.1587603725492954, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 0.15}) +[2026-06-08 12:30:06,719][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.21761201322078705, 'train/grad_norm_pre_clip': 0.7856704592704773, 'timing/data': 0.009556161239743233, 'timing/model': 1.16035395860672, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.16}) +[2026-06-08 12:33:06,930][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.20346863567829132, 'train/grad_norm_pre_clip': 2.019679069519043, 'timing/data': 0.01086665689945221, 'timing/model': 1.1718800719827414, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 0.16}) +[2026-06-08 12:50:38,458][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state +[2026-06-08 12:50:38,462][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 12:50:44,998][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/model.safetensors +[2026-06-08 12:50:54,255][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/optimizer.bin +[2026-06-08 12:50:54,256][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/sampler.bin +[2026-06-08 12:50:54,256][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/sampler_1.bin +[2026-06-08 12:50:54,257][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/random_states_0.pkl +[2026-06-08 12:50:54,275][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 12:50:54,275][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-08 12:50:54,277][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.13002508878707886, 'train/grad_norm_pre_clip': 0.8968712091445923, 'eval/loss': 0.10868137329816818, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 310.24000000001035, 'rl_games_eval/mid_train/mean_length': 2518.6, 'rl_games_eval/mid_train/std_reward': 130.48938041083377, 'rl_games_eval/mid_train/std_length': 1051.8269059118045, 'rl_games_eval/mid_train/macro_mean_reward': 310.24000000001035, 'rl_games_eval/mid_train/macro_mean_length': 2518.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 310.24000000001035, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 2518.6, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 130.48938041083377, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 1051.8269059118045, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 310.24000000001035, 'checkpoint/best_score': 310.24000000001035, 'checkpoint/best_step': 500.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.008437078446149826, 'timing/model': 1.1628056187182665, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.16}) +[2026-06-08 12:50:54,277][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_500_state +[2026-06-08 12:50:54,284][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 12:50:59,697][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_500_state/model.safetensors +[2026-06-08 12:51:09,364][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_500_state/optimizer.bin +[2026-06-08 12:51:09,364][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_500_state/sampler.bin +[2026-06-08 12:51:09,365][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_500_state/sampler_1.bin +[2026-06-08 12:51:09,366][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-08 12:51:09,470][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-08 12:51:09,471][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 12:51:09,483][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 12:54:08,214][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.017008284106850624, 'train/grad_norm_pre_clip': 0.2817673981189728, 'timing/data': 0.009345592930912971, 'timing/model': 1.160497648641467, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 0.17}) +[2026-06-08 12:57:07,181][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.12711577117443085, 'train/grad_norm_pre_clip': 1.5070555210113525, 'timing/data': 0.015097945928573608, 'timing/model': 1.2197858896106482, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.17}) +[2026-06-08 13:00:10,016][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.10411541163921356, 'train/grad_norm_pre_clip': 1.2097944021224976, 'timing/data': 0.011897852644324303, 'timing/model': 1.1647898852825165, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 0.17}) +[2026-06-08 13:03:08,806][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.07207261025905609, 'train/grad_norm_pre_clip': 0.7633359432220459, 'timing/data': 0.009624367579817772, 'timing/model': 1.1589701436460018, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.18}) +[2026-06-08 13:06:09,133][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.07310670614242554, 'train/grad_norm_pre_clip': 0.4069511592388153, 'timing/data': 0.009813429787755013, 'timing/model': 1.1603018566966057, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 0.18}) +[2026-06-08 13:09:09,292][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.18644267320632935, 'train/grad_norm_pre_clip': 2.9648518562316895, 'timing/data': 0.009362010285258293, 'timing/model': 1.1606480106711388, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.18}) +[2026-06-08 13:12:15,288][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.17379489541053772, 'train/grad_norm_pre_clip': 1.7855198383331299, 'timing/data': 0.000728880986571312, 'timing/model': 1.2257568035274744, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 0.19}) +[2026-06-08 13:15:16,452][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.08787855505943298, 'train/grad_norm_pre_clip': 1.5048805475234985, 'timing/data': 0.007897112518548965, 'timing/model': 1.1267956085503101, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.19}) +[2026-06-08 13:18:15,096][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.29520830512046814, 'train/grad_norm_pre_clip': 3.464940071105957, 'timing/data': 0.010032741352915764, 'timing/model': 1.15969817340374, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.19}) +[2026-06-08 13:26:31,277][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 600 +[2026-06-08 13:26:31,278][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.06580091267824173, 'train/grad_norm_pre_clip': 2.546194076538086, 'eval/loss': 0.13524775207042694, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 100.25999999999844, 'rl_games_eval/mid_train/mean_length': 828.2, 'rl_games_eval/mid_train/std_reward': 88.77550563077378, 'rl_games_eval/mid_train/std_length': 716.0853021812416, 'rl_games_eval/mid_train/macro_mean_reward': 100.25999999999844, 'rl_games_eval/mid_train/macro_mean_length': 828.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 100.25999999999844, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 828.2, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 88.77550563077378, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 716.0853021812416, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 100.25999999999844, 'checkpoint/best_score': 310.24000000001035, 'checkpoint/best_step': 500.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00974733941257, 'timing/model': 1.1546872146427631, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.2}) +[2026-06-08 13:26:31,279][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_600_state +[2026-06-08 13:26:31,282][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 13:26:37,882][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_600_state/model.safetensors +[2026-06-08 13:26:48,028][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_600_state/optimizer.bin +[2026-06-08 13:26:48,029][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_600_state/sampler.bin +[2026-06-08 13:26:48,029][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_600_state/sampler_1.bin +[2026-06-08 13:26:48,030][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_600_state/random_states_0.pkl +[2026-06-08 13:26:48,122][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600 +[2026-06-08 13:26:48,122][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 13:26:48,135][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 13:29:47,612][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.09090287983417511, 'train/grad_norm_pre_clip': 1.2067210674285889, 'timing/data': 0.010087037459015846, 'timing/model': 1.16259709559381, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 0.2}) +[2026-06-08 13:32:49,610][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.02098463661968708, 'train/grad_norm_pre_clip': 0.2769014835357666, 'timing/data': 0.0114809051156044, 'timing/model': 1.1577500011771917, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.2}) +[2026-06-08 13:35:48,537][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.1719634085893631, 'train/grad_norm_pre_clip': 3.7691009044647217, 'timing/data': 0.009159961715340614, 'timing/model': 1.1490398161113262, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 0.21}) +[2026-06-08 13:38:44,933][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.1616826206445694, 'train/grad_norm_pre_clip': 0.9505186080932617, 'timing/data': 0.008677702397108078, 'timing/model': 1.1385265551507473, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.21}) +[2026-06-08 13:41:41,489][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.048989418894052505, 'train/grad_norm_pre_clip': 0.5133188366889954, 'timing/data': 0.010297942906618118, 'timing/model': 1.1878602169454098, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 0.21}) +[2026-06-08 13:44:38,511][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.024587882682681084, 'train/grad_norm_pre_clip': 0.2530308663845062, 'timing/data': 0.00970461592078209, 'timing/model': 1.1306265406310558, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.22}) +[2026-06-08 13:47:34,467][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.053108494728803635, 'train/grad_norm_pre_clip': 0.6187007427215576, 'timing/data': 0.00883355550467968, 'timing/model': 1.1244599353522062, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 0.22}) +[2026-06-08 13:50:33,117][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.06267092376947403, 'train/grad_norm_pre_clip': 0.6285062432289124, 'timing/data': 0.010297121480107307, 'timing/model': 1.1563981957733631, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.22}) +[2026-06-08 13:53:29,411][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.2038508951663971, 'train/grad_norm_pre_clip': 1.7500640153884888, 'timing/data': 0.008594494313001633, 'timing/model': 1.3239865750074387, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 0.23}) +[2026-06-08 14:07:36,864][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 700 +[2026-06-08 14:07:36,867][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.05006268993020058, 'train/grad_norm_pre_clip': 0.5812363624572754, 'eval/loss': 0.09375068545341492, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 248.3000000000061, 'rl_games_eval/mid_train/mean_length': 2020.6, 'rl_games_eval/mid_train/std_reward': 141.48824686171935, 'rl_games_eval/mid_train/std_length': 1141.7538438735382, 'rl_games_eval/mid_train/macro_mean_reward': 248.3000000000061, 'rl_games_eval/mid_train/macro_mean_length': 2020.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 248.3000000000061, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 2020.6, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 141.48824686171935, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 1141.7538438735382, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 248.3000000000061, 'checkpoint/best_score': 310.24000000001035, 'checkpoint/best_step': 500.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.007856132462620735, 'timing/model': 1.124145284295082, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.23}) +[2026-06-08 14:07:36,868][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_700_state +[2026-06-08 14:07:36,871][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 14:07:43,010][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_700_state/model.safetensors +[2026-06-08 14:07:52,260][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_700_state/optimizer.bin +[2026-06-08 14:07:52,260][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_700_state/sampler.bin +[2026-06-08 14:07:52,261][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_700_state/sampler_1.bin +[2026-06-08 14:07:52,262][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_700_state/random_states_0.pkl +[2026-06-08 14:07:52,379][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700 +[2026-06-08 14:07:52,379][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 14:07:52,392][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 14:10:51,102][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.020143063738942146, 'train/grad_norm_pre_clip': 0.3056764006614685, 'timing/data': 0.01017102599143982, 'timing/model': 1.1854197848588228, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 0.23}) +[2026-06-08 14:13:49,850][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.0670040175318718, 'train/grad_norm_pre_clip': 1.9509207010269165, 'timing/data': 0.008529912680387497, 'timing/model': 1.1224462389945984, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.24}) +[2026-06-08 14:16:48,720][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.17669881880283356, 'train/grad_norm_pre_clip': 1.1636183261871338, 'timing/data': 0.009930020198225975, 'timing/model': 1.2474681474268436, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.24}) +[2026-06-08 14:19:47,107][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.013532213866710663, 'train/grad_norm_pre_clip': 0.1994314193725586, 'timing/data': 0.010107755661010742, 'timing/model': 1.1567442752420902, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 0.24}) +[2026-06-08 14:22:45,611][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.10048847645521164, 'train/grad_norm_pre_clip': 1.5076580047607422, 'timing/data': 0.009069573134183884, 'timing/model': 1.1567994505167007, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 0.25}) +[2026-06-08 14:25:51,427][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.20258782804012299, 'train/grad_norm_pre_clip': 2.888821840286255, 'timing/data': 0.10798907279968262, 'timing/model': 1.1796571388840675, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 0.25}) +[2026-06-08 14:28:56,812][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.012749352492392063, 'train/grad_norm_pre_clip': 0.34907254576683044, 'timing/data': 0.01484144851565361, 'timing/model': 1.2210409231483936, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 0.25}) +[2026-06-08 14:31:56,503][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.06071474775671959, 'train/grad_norm_pre_clip': 1.53829026222229, 'timing/data': 0.00866570696234703, 'timing/model': 1.1569438390433788, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 0.26}) +[2026-06-08 14:35:00,193][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.028188657015562057, 'train/grad_norm_pre_clip': 0.5213531255722046, 'timing/data': 0.010124603286385536, 'timing/model': 1.1657463181763887, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 0.26}) +[2026-06-08 14:47:24,740][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 800 +[2026-06-08 14:47:24,742][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.06301678717136383, 'train/grad_norm_pre_clip': 0.5499922633171082, 'eval/loss': 0.09661654382944107, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 194.58000000000263, 'rl_games_eval/mid_train/mean_length': 1585.6, 'rl_games_eval/mid_train/std_reward': 143.8947726639245, 'rl_games_eval/mid_train/std_length': 1160.6674976064419, 'rl_games_eval/mid_train/macro_mean_reward': 194.58000000000263, 'rl_games_eval/mid_train/macro_mean_length': 1585.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 194.58000000000263, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 1585.6, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 143.8947726639245, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 1160.6674976064419, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 194.58000000000263, 'checkpoint/best_score': 310.24000000001035, 'checkpoint/best_step': 500.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.010225983336567879, 'timing/model': 1.1859461069107056, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.26}) +[2026-06-08 14:47:24,743][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_800_state +[2026-06-08 14:47:24,749][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 14:47:32,303][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_800_state/model.safetensors +[2026-06-08 14:47:42,755][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_800_state/optimizer.bin +[2026-06-08 14:47:42,755][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_800_state/sampler.bin +[2026-06-08 14:47:42,755][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_800_state/sampler_1.bin +[2026-06-08 14:47:42,757][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_800_state/random_states_0.pkl +[2026-06-08 14:47:42,852][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800 +[2026-06-08 14:47:42,852][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 14:47:42,865][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 14:50:46,813][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.08089163154363632, 'train/grad_norm_pre_clip': 1.4879106283187866, 'timing/data': 0.009208114817738533, 'timing/model': 1.165800180286169, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 0.27}) +[2026-06-08 14:53:48,703][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.059415146708488464, 'train/grad_norm_pre_clip': 0.3603160083293915, 'timing/data': 0.009835690259933472, 'timing/model': 1.1621369328349829, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 0.27}) +[2026-06-08 14:56:53,018][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.028155799955129623, 'train/grad_norm_pre_clip': 0.8506107926368713, 'timing/data': 0.010494021698832512, 'timing/model': 1.1660369373857975, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 0.27}) +[2026-06-08 14:59:54,793][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.1560654491186142, 'train/grad_norm_pre_clip': 1.9778612852096558, 'timing/data': 0.010632788762450218, 'timing/model': 1.1614328045397997, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 0.28}) +[2026-06-08 15:02:55,857][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.03305920958518982, 'train/grad_norm_pre_clip': 0.517228364944458, 'timing/data': 0.008657757192850113, 'timing/model': 1.1494107246398926, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 0.28}) +[2026-06-08 15:05:57,881][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.18592728674411774, 'train/grad_norm_pre_clip': 1.6480199098587036, 'timing/data': 0.010408120229840279, 'timing/model': 1.1850337348878384, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 0.28}) +[2026-06-08 15:08:59,466][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.13178083300590515, 'train/grad_norm_pre_clip': 2.265789747238159, 'timing/data': 0.011781884357333183, 'timing/model': 1.4202481843531132, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 0.29}) +[2026-06-08 15:12:02,508][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.05810749903321266, 'train/grad_norm_pre_clip': 1.4705811738967896, 'timing/data': 0.011630641296505928, 'timing/model': 1.2352572306990623, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 0.29}) +[2026-06-08 15:15:03,201][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.14341698586940765, 'train/grad_norm_pre_clip': 1.5553256273269653, 'timing/data': 0.009581001475453377, 'timing/model': 1.1604828462004662, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 0.29}) +[2026-06-08 15:34:11,777][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state +[2026-06-08 15:34:11,784][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 15:34:19,246][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/model.safetensors +[2026-06-08 15:34:28,095][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/optimizer.bin +[2026-06-08 15:34:28,096][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/sampler.bin +[2026-06-08 15:34:28,097][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/sampler_1.bin +[2026-06-08 15:34:28,098][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/random_states_0.pkl +[2026-06-08 15:34:28,114][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 15:34:28,115][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 900 +[2026-06-08 15:34:28,116][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.09157010167837143, 'train/grad_norm_pre_clip': 1.9928643703460693, 'eval/loss': 0.09825854003429413, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 338.2200000000138, 'rl_games_eval/mid_train/mean_length': 2744.0, 'rl_games_eval/mid_train/std_reward': 140.1723282249517, 'rl_games_eval/mid_train/std_length': 1128.4254516803492, 'rl_games_eval/mid_train/macro_mean_reward': 338.2200000000138, 'rl_games_eval/mid_train/macro_mean_length': 2744.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 338.2200000000138, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 2744.0, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 140.1723282249517, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 1128.4254516803492, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 338.2200000000138, 'checkpoint/best_score': 338.2200000000138, 'checkpoint/best_step': 900.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.011309430003166199, 'timing/model': 1.394687106832862, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 0.3}) +[2026-06-08 15:34:28,117][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_900_state +[2026-06-08 15:34:28,122][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 15:34:33,158][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_900_state/model.safetensors +[2026-06-08 15:34:42,389][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_900_state/optimizer.bin +[2026-06-08 15:34:42,390][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_900_state/sampler.bin +[2026-06-08 15:34:42,390][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_900_state/sampler_1.bin +[2026-06-08 15:34:42,391][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_900_state/random_states_0.pkl +[2026-06-08 15:34:42,496][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-06-08 15:34:42,497][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 15:34:42,510][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 15:37:44,098][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.034432463347911835, 'train/grad_norm_pre_clip': 0.6489217877388, 'timing/data': 0.01129702664911747, 'timing/model': 1.160439744591713, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 0.3}) +[2026-06-08 15:40:45,895][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.0705709457397461, 'train/grad_norm_pre_clip': 1.49032461643219, 'timing/data': 0.0092683807015419, 'timing/model': 1.2121581360697746, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 0.3}) +[2026-06-08 15:43:46,966][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.12363343685865402, 'train/grad_norm_pre_clip': 0.9752617478370667, 'timing/data': 0.013074100017547607, 'timing/model': 1.1819357965141535, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 0.31}) +[2026-06-08 15:46:50,796][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.07335102558135986, 'train/grad_norm_pre_clip': 2.0095367431640625, 'timing/data': 0.008206762373447418, 'timing/model': 1.1371600050479174, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 0.31}) +[2026-06-08 15:49:57,112][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.04359815642237663, 'train/grad_norm_pre_clip': 0.7014718651771545, 'timing/data': 0.000986628234386444, 'timing/model': 1.186644785106182, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 0.31}) +[2026-06-08 15:53:00,323][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.04109642282128334, 'train/grad_norm_pre_clip': 0.5711485743522644, 'timing/data': 0.010653909295797348, 'timing/model': 1.1713629476726055, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 0.32}) +[2026-06-08 15:56:01,922][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.09751032292842865, 'train/grad_norm_pre_clip': 2.0108447074890137, 'timing/data': 0.009413056075572968, 'timing/model': 1.2636358942836523, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 0.32}) +[2026-06-08 15:59:03,509][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.06357042491436005, 'train/grad_norm_pre_clip': 0.7878151535987854, 'timing/data': 0.01067105121910572, 'timing/model': 1.1606435310095549, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 0.32}) +[2026-06-08 16:02:05,820][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.0995241180062294, 'train/grad_norm_pre_clip': 1.1449799537658691, 'timing/data': 0.010690923780202866, 'timing/model': 1.1697212420403957, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 0.33}) +[2026-06-08 16:17:20,109][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-06-08 16:17:20,111][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.004464845638722181, 'train/grad_norm_pre_clip': 0.11469534039497375, 'eval/loss': 0.08239735662937164, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 252.00000000000563, 'rl_games_eval/mid_train/mean_length': 2052.2, 'rl_games_eval/mid_train/std_reward': 128.80000000001152, 'rl_games_eval/mid_train/std_length': 1039.6000000000001, 'rl_games_eval/mid_train/macro_mean_reward': 252.00000000000563, 'rl_games_eval/mid_train/macro_mean_length': 2052.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 252.00000000000563, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 2052.2, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 128.80000000001152, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 1039.6000000000001, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 252.00000000000563, 'checkpoint/best_score': 338.2200000000138, 'checkpoint/best_step': 900.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.012044746428728104, 'timing/model': 1.173965398222208, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 0.33}) +[2026-06-08 16:17:20,112][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1000_state +[2026-06-08 16:17:20,118][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 16:17:28,307][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1000_state/model.safetensors +[2026-06-08 16:17:42,094][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1000_state/optimizer.bin +[2026-06-08 16:17:42,096][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1000_state/sampler.bin +[2026-06-08 16:17:42,096][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-08 16:17:42,098][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-08 16:17:42,192][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-08 16:17:42,193][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 16:17:42,205][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 16:20:44,680][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.07213591039180756, 'train/grad_norm_pre_clip': 1.034422516822815, 'timing/data': 0.010702962055802345, 'timing/model': 1.2123520132154226, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 0.33}) +[2026-06-08 16:23:46,857][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.03804894909262657, 'train/grad_norm_pre_clip': 0.7617294192314148, 'timing/data': 0.008556270971894264, 'timing/model': 1.1773285884410143, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 0.34}) +[2026-06-08 16:26:49,687][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.15592916309833527, 'train/grad_norm_pre_clip': 1.2904431819915771, 'timing/data': 0.008394552394747734, 'timing/model': 1.1709299087524414, 'learning_rate/qwen_vl_interface': 2.9058359641943523e-06, 'learning_rate/action_model': 2.9058359641943523e-05, 'epoch': 0.34}) +[2026-06-08 16:29:46,268][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.03305686265230179, 'train/grad_norm_pre_clip': 0.9334266185760498, 'timing/data': 0.009744232520461082, 'timing/model': 1.1553084999322891, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 0.34}) +[2026-06-08 16:32:49,214][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.09243889898061752, 'train/grad_norm_pre_clip': 1.852088451385498, 'timing/data': 0.010698625817894936, 'timing/model': 1.23831638507545, 'learning_rate/qwen_vl_interface': 2.7228476365521524e-06, 'learning_rate/action_model': 2.722847636552152e-05, 'epoch': 0.35}) +[2026-06-08 16:35:50,903][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.0668354332447052, 'train/grad_norm_pre_clip': 0.8879101276397705, 'timing/data': 0.009054282680153847, 'timing/model': 1.176898231729865, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 0.35}) +[2026-06-08 16:38:52,428][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.09418393671512604, 'train/grad_norm_pre_clip': 1.715146780014038, 'timing/data': 0.011028766632080078, 'timing/model': 1.4512520805001259, 'learning_rate/qwen_vl_interface': 2.5449486574498535e-06, 'learning_rate/action_model': 2.5449486574498534e-05, 'epoch': 0.35}) +[2026-06-08 16:41:55,869][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.03556664660573006, 'train/grad_norm_pre_clip': 0.9381561875343323, 'timing/data': 0.008776331320405006, 'timing/model': 1.1320041883736849, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 0.36}) +[2026-06-08 16:44:59,023][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.025244437158107758, 'train/grad_norm_pre_clip': 0.5926901698112488, 'timing/data': 0.011426854878664017, 'timing/model': 1.4627357721328735, 'learning_rate/qwen_vl_interface': 2.372497291756106e-06, 'learning_rate/action_model': 2.3724972917561058e-05, 'epoch': 0.36}) +[2026-06-08 17:08:25,168][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state +[2026-06-08 17:08:25,178][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 17:08:33,242][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/model.safetensors +[2026-06-08 17:08:47,568][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/optimizer.bin +[2026-06-08 17:08:47,570][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/sampler.bin +[2026-06-08 17:08:47,571][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/sampler_1.bin +[2026-06-08 17:08:47,572][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state/random_states_0.pkl +[2026-06-08 17:08:47,588][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 17:08:47,589][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1100 +[2026-06-08 17:08:47,590][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.00426764041185379, 'train/grad_norm_pre_clip': 0.10021832585334778, 'eval/loss': 0.0850915014743805, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 416.6200000000207, 'rl_games_eval/mid_train/mean_length': 3376.8, 'rl_games_eval/mid_train/std_reward': 39.24723684542067, 'rl_games_eval/mid_train/std_length': 313.913618691512, 'rl_games_eval/mid_train/macro_mean_reward': 416.6200000000207, 'rl_games_eval/mid_train/macro_mean_length': 3376.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 416.6200000000207, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 3376.8, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 39.24723684542067, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 313.913618691512, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 416.6200000000207, 'checkpoint/best_score': 416.6200000000207, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.011834332719445229, 'timing/model': 1.1615077871829271, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 0.36}) +[2026-06-08 17:08:47,591][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1100_state +[2026-06-08 17:08:47,597][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 17:08:53,141][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1100_state/model.safetensors +[2026-06-08 17:09:04,458][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1100_state/optimizer.bin +[2026-06-08 17:09:04,459][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1100_state/sampler.bin +[2026-06-08 17:09:04,459][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1100_state/sampler_1.bin +[2026-06-08 17:09:04,460][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1100_state/random_states_0.pkl +[2026-06-08 17:09:04,546][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-06-08 17:09:04,546][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 17:09:04,559][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 17:12:08,807][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.024656103923916817, 'train/grad_norm_pre_clip': 0.7575395107269287, 'timing/data': 0.01092049852013588, 'timing/model': 1.1738284286111593, 'learning_rate/qwen_vl_interface': 2.205840833571135e-06, 'learning_rate/action_model': 2.205840833571135e-05, 'epoch': 0.37}) +[2026-06-08 17:15:12,075][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.09459944069385529, 'train/grad_norm_pre_clip': 2.1723759174346924, 'timing/data': 0.009557869285345078, 'timing/model': 1.521728239953518, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 0.37}) +[2026-06-08 17:18:15,175][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.051030807197093964, 'train/grad_norm_pre_clip': 1.6220430135726929, 'timing/data': 0.009195368736982346, 'timing/model': 1.1352953594177961, 'learning_rate/qwen_vl_interface': 2.045314906822647e-06, 'learning_rate/action_model': 2.0453149068226467e-05, 'epoch': 0.37}) +[2026-06-08 17:21:22,680][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.17279329895973206, 'train/grad_norm_pre_clip': 3.191983938217163, 'timing/data': 0.032377755269408226, 'timing/model': 1.2143332846462727, 'learning_rate/qwen_vl_interface': 1.967452417312393e-06, 'learning_rate/action_model': 1.9674524173123927e-05, 'epoch': 0.38}) +[2026-06-08 17:24:25,680][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.0833478569984436, 'train/grad_norm_pre_clip': 0.9369668364524841, 'timing/data': 0.010259415954351425, 'timing/model': 1.4227268807590008, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 0.38}) +[2026-06-08 17:27:28,391][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.16230109333992004, 'train/grad_norm_pre_clip': 2.346400499343872, 'timing/data': 0.01116420328617096, 'timing/model': 1.2141550704836845, 'learning_rate/qwen_vl_interface': 1.8167243968198905e-06, 'learning_rate/action_model': 1.8167243968198905e-05, 'epoch': 0.38}) +[2026-06-08 17:30:30,534][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.11973495036363602, 'train/grad_norm_pre_clip': 1.0204708576202393, 'timing/data': 0.011476991698145866, 'timing/model': 1.174204982817173, 'learning_rate/qwen_vl_interface': 1.7439347619351324e-06, 'learning_rate/action_model': 1.7439347619351324e-05, 'epoch': 0.39}) +[2026-06-08 17:33:33,063][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.015590225346386433, 'train/grad_norm_pre_clip': 0.5079066157341003, 'timing/data': 0.011594787240028381, 'timing/model': 1.395963553339243, 'learning_rate/qwen_vl_interface': 1.6729105364828486e-06, 'learning_rate/action_model': 1.6729105364828484e-05, 'epoch': 0.39}) +[2026-06-08 17:36:34,941][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.06840702891349792, 'train/grad_norm_pre_clip': 0.7520391941070557, 'timing/data': 0.010455742478370667, 'timing/model': 1.159628102555871, 'learning_rate/qwen_vl_interface': 1.6036874832994473e-06, 'learning_rate/action_model': 1.603687483299447e-05, 'epoch': 0.39}) +[2026-06-08 17:54:27,863][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1200 +[2026-06-08 17:54:27,865][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.012219227850437164, 'train/grad_norm_pre_clip': 0.24534940719604492, 'eval/loss': 0.08527901768684387, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 303.7400000000101, 'rl_games_eval/mid_train/mean_length': 2468.4, 'rl_games_eval/mid_train/std_reward': 128.87987585345886, 'rl_games_eval/mid_train/std_length': 1038.7195194083915, 'rl_games_eval/mid_train/macro_mean_reward': 303.7400000000101, 'rl_games_eval/mid_train/macro_mean_length': 2468.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 303.7400000000101, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 2468.4, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 128.87987585345886, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 1038.7195194083915, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 303.7400000000101, 'checkpoint/best_score': 416.6200000000207, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00888887606561184, 'timing/model': 1.1291090566664934, 'learning_rate/qwen_vl_interface': 1.5363004582768584e-06, 'learning_rate/action_model': 1.5363004582768584e-05, 'epoch': 0.39}) +[2026-06-08 17:54:27,866][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1200_state +[2026-06-08 17:54:27,870][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 17:54:34,356][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1200_state/model.safetensors +[2026-06-08 17:54:47,633][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1200_state/optimizer.bin +[2026-06-08 17:54:47,634][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1200_state/sampler.bin +[2026-06-08 17:54:47,634][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1200_state/sampler_1.bin +[2026-06-08 17:54:47,636][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1200_state/random_states_0.pkl +[2026-06-08 17:54:47,738][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-06-08 17:54:47,738][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 17:54:47,751][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 17:57:46,578][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.0461403988301754, 'train/grad_norm_pre_clip': 0.6250188946723938, 'timing/data': 0.00892794318497181, 'timing/model': 1.154057890176773, 'learning_rate/qwen_vl_interface': 1.4707833928115406e-06, 'learning_rate/action_model': 1.4707833928115405e-05, 'epoch': 0.4}) +[2026-06-08 18:00:43,655][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.010066299699246883, 'train/grad_norm_pre_clip': 0.4282253086566925, 'timing/data': 0.009979426860809326, 'timing/model': 1.1237175520509481, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 0.4}) +[2026-06-08 18:03:41,092][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.025397680699825287, 'train/grad_norm_pre_clip': 0.6370608806610107, 'timing/data': 0.009168718010187149, 'timing/model': 1.12836698256433, 'learning_rate/qwen_vl_interface': 1.3454901416224284e-06, 'learning_rate/action_model': 1.3454901416224284e-05, 'epoch': 0.4}) +[2026-06-08 18:06:42,007][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.08475744724273682, 'train/grad_norm_pre_clip': 0.6436638832092285, 'timing/data': 0.009207040071487427, 'timing/model': 1.1233865600079298, 'learning_rate/qwen_vl_interface': 1.2857770448238124e-06, 'learning_rate/action_model': 1.2857770448238123e-05, 'epoch': 0.41}) +[2026-06-08 18:09:40,728][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.037193723022937775, 'train/grad_norm_pre_clip': 0.8134264349937439, 'timing/data': 0.008674537762999535, 'timing/model': 1.1279005613178015, 'learning_rate/qwen_vl_interface': 1.2280600536656504e-06, 'learning_rate/action_model': 1.2280600536656503e-05, 'epoch': 0.41}) +[2026-06-08 18:12:42,016][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.01338848564773798, 'train/grad_norm_pre_clip': 0.2614465057849884, 'timing/data': 0.009840166196227074, 'timing/model': 1.4058293774724007, 'learning_rate/qwen_vl_interface': 1.172368230391136e-06, 'learning_rate/action_model': 1.1723682303911359e-05, 'epoch': 0.41}) +[2026-06-08 18:15:43,892][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.030301392078399658, 'train/grad_norm_pre_clip': 0.565644383430481, 'timing/data': 0.011830063536763191, 'timing/model': 1.2228717282414436, 'learning_rate/qwen_vl_interface': 1.1187296175104472e-06, 'learning_rate/action_model': 1.1187296175104471e-05, 'epoch': 0.42}) +[2026-06-08 18:18:44,601][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.01274733617901802, 'train/grad_norm_pre_clip': 0.27422839403152466, 'timing/data': 0.012112533673644066, 'timing/model': 1.4492776207625866, 'learning_rate/qwen_vl_interface': 1.0671712236804947e-06, 'learning_rate/action_model': 1.0671712236804946e-05, 'epoch': 0.42}) +[2026-06-08 18:21:45,085][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.17139744758605957, 'train/grad_norm_pre_clip': 1.6096696853637695, 'timing/data': 0.009357860311865807, 'timing/model': 1.1589287053793669, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 0.42}) +[2026-06-08 18:43:22,196][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1300 +[2026-06-08 18:43:22,199][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.09955427795648575, 'train/grad_norm_pre_clip': 1.2992277145385742, 'eval/loss': 0.08049743622541428, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 389.62000000001643, 'rl_games_eval/mid_train/mean_length': 3159.0, 'rl_games_eval/mid_train/std_reward': 78.97610777951803, 'rl_games_eval/mid_train/std_length': 634.5095743958478, 'rl_games_eval/mid_train/macro_mean_reward': 389.62000000001643, 'rl_games_eval/mid_train/macro_mean_length': 3159.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 389.62000000001643, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 3159.0, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 78.97610777951803, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 634.5095743958478, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 389.62000000001643, 'checkpoint/best_score': 416.6200000000207, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.009849876165390015, 'timing/model': 1.1663389690220356, 'learning_rate/qwen_vl_interface': 9.703978774635094e-07, 'learning_rate/action_model': 9.703978774635095e-06, 'epoch': 0.43}) +[2026-06-08 18:43:22,200][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1300_state +[2026-06-08 18:43:22,208][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 18:43:30,770][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1300_state/model.safetensors +[2026-06-08 18:43:42,979][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1300_state/optimizer.bin +[2026-06-08 18:43:42,981][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1300_state/sampler.bin +[2026-06-08 18:43:42,981][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1300_state/sampler_1.bin +[2026-06-08 18:43:42,983][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1300_state/random_states_0.pkl +[2026-06-08 18:43:43,078][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-06-08 18:43:43,078][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 18:43:43,091][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 18:46:45,624][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.041848454624414444, 'train/grad_norm_pre_clip': 0.5528457760810852, 'timing/data': 0.010647350922226906, 'timing/model': 1.1650979276746511, 'learning_rate/qwen_vl_interface': 9.252316533706246e-07, 'learning_rate/action_model': 9.252316533706246e-06, 'epoch': 0.43}) +[2026-06-08 18:49:47,516][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.018248330801725388, 'train/grad_norm_pre_clip': 0.6716136932373047, 'timing/data': 0.011112675070762634, 'timing/model': 1.474422363564372, 'learning_rate/qwen_vl_interface': 8.822430803806098e-07, 'learning_rate/action_model': 8.822430803806097e-06, 'epoch': 0.43}) +[2026-06-08 18:52:53,307][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.016272928565740585, 'train/grad_norm_pre_clip': 0.4529944062232971, 'timing/data': 0.0007211826741695404, 'timing/model': 1.2101793717592955, 'learning_rate/qwen_vl_interface': 8.414538045345655e-07, 'learning_rate/action_model': 8.414538045345654e-06, 'epoch': 0.44}) +[2026-06-08 18:55:55,395][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.01105588860809803, 'train/grad_norm_pre_clip': 0.46362218260765076, 'timing/data': 0.009991219267249107, 'timing/model': 1.1507067997008562, 'learning_rate/qwen_vl_interface': 8.028843644612483e-07, 'learning_rate/action_model': 8.028843644612482e-06, 'epoch': 0.44}) +[2026-06-08 18:58:59,023][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.02710859104990959, 'train/grad_norm_pre_clip': 0.4849602282047272, 'timing/data': 0.011608656495809555, 'timing/model': 1.1936019826680422, 'learning_rate/qwen_vl_interface': 7.665541810352541e-07, 'learning_rate/action_model': 7.66554181035254e-06, 'epoch': 0.44}) +[2026-06-08 19:02:03,428][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.03653616085648537, 'train/grad_norm_pre_clip': 0.9231950640678406, 'timing/data': 0.010000834241509438, 'timing/model': 1.1726866886019707, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 0.45}) +[2026-06-08 19:05:04,658][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.030297966673970222, 'train/grad_norm_pre_clip': 0.8678412437438965, 'timing/data': 0.008351152762770653, 'timing/model': 1.1253036260604858, 'learning_rate/qwen_vl_interface': 7.006836207465803e-07, 'learning_rate/action_model': 7.006836207465803e-06, 'epoch': 0.45}) +[2026-06-08 19:08:04,407][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.01336376927793026, 'train/grad_norm_pre_clip': 0.2615305781364441, 'timing/data': 0.010769713670015335, 'timing/model': 1.1624058913439512, 'learning_rate/qwen_vl_interface': 6.711764116946972e-07, 'learning_rate/action_model': 6.7117641169469715e-06, 'epoch': 0.45}) +[2026-06-08 19:11:01,550][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.12276582419872284, 'train/grad_norm_pre_clip': 2.225813627243042, 'timing/data': 0.00849858671426773, 'timing/model': 1.1266859136521816, 'learning_rate/qwen_vl_interface': 6.439747782107373e-07, 'learning_rate/action_model': 6.4397477821073735e-06, 'epoch': 0.46}) +[2026-06-08 19:26:39,157][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1400 +[2026-06-08 19:26:39,159][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.019798289984464645, 'train/grad_norm_pre_clip': 0.33752644062042236, 'eval/loss': 0.0812939926981926, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 273.94000000000517, 'rl_games_eval/mid_train/mean_length': 2228.0, 'rl_games_eval/mid_train/std_reward': 120.87947054815722, 'rl_games_eval/mid_train/std_length': 973.8215442266617, 'rl_games_eval/mid_train/macro_mean_reward': 273.94000000000517, 'rl_games_eval/mid_train/macro_mean_length': 2228.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 273.94000000000517, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 2228.0, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 120.87947054815722, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 973.8215442266617, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 273.94000000000517, 'checkpoint/best_score': 416.6200000000207, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.010819541290402412, 'timing/model': 1.1604594755917788, 'learning_rate/qwen_vl_interface': 6.190924171363378e-07, 'learning_rate/action_model': 6.190924171363378e-06, 'epoch': 0.46}) +[2026-06-08 19:26:39,159][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1400_state +[2026-06-08 19:26:39,163][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 19:26:45,305][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1400_state/model.safetensors +[2026-06-08 19:26:56,400][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1400_state/optimizer.bin +[2026-06-08 19:26:56,401][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1400_state/sampler.bin +[2026-06-08 19:26:56,401][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1400_state/sampler_1.bin +[2026-06-08 19:26:56,403][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1400_state/random_states_0.pkl +[2026-06-08 19:26:56,517][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-06-08 19:26:56,517][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 19:26:56,540][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 19:29:57,621][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.10148926824331284, 'train/grad_norm_pre_clip': 1.0504653453826904, 'timing/data': 0.00836261361837387, 'timing/model': 1.1466510500758886, 'learning_rate/qwen_vl_interface': 5.965418574896315e-07, 'learning_rate/action_model': 5.9654185748963145e-06, 'epoch': 0.46}) +[2026-06-08 19:32:56,594][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.00957846362143755, 'train/grad_norm_pre_clip': 0.20466101169586182, 'timing/data': 0.009948233142495155, 'timing/model': 1.1614174488931894, 'learning_rate/qwen_vl_interface': 5.763344541565091e-07, 'learning_rate/action_model': 5.7633445415650905e-06, 'epoch': 0.47}) +[2026-06-08 19:35:56,555][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.3033120036125183, 'train/grad_norm_pre_clip': 2.685802698135376, 'timing/data': 0.00913010910153389, 'timing/model': 1.1317523401230574, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 0.47}) +[2026-06-08 19:38:58,797][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.021947981789708138, 'train/grad_norm_pre_clip': 0.34650179743766785, 'timing/data': 0.010108817368745804, 'timing/model': 1.1595759522169828, 'learning_rate/qwen_vl_interface': 5.429886316023102e-07, 'learning_rate/action_model': 5.429886316023101e-06, 'epoch': 0.47}) +[2026-06-08 19:41:57,802][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.0031615481711924076, 'train/grad_norm_pre_clip': 0.10607803612947464, 'timing/data': 0.010229527950286865, 'timing/model': 1.3460100404918194, 'learning_rate/qwen_vl_interface': 5.298670030070977e-07, 'learning_rate/action_model': 5.298670030070976e-06, 'epoch': 0.48}) +[2026-06-08 19:44:57,907][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.0939631462097168, 'train/grad_norm_pre_clip': 1.4529192447662354, 'timing/data': 0.010373881086707115, 'timing/model': 1.1633821707218885, 'learning_rate/qwen_vl_interface': 5.191221035226144e-07, 'learning_rate/action_model': 5.191221035226144e-06, 'epoch': 0.48}) +[2026-06-08 19:47:59,500][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.09213721007108688, 'train/grad_norm_pre_clip': 0.6445643901824951, 'timing/data': 0.011109311133623123, 'timing/model': 1.2033779975026846, 'learning_rate/qwen_vl_interface': 5.107593435293336e-07, 'learning_rate/action_model': 5.107593435293336e-06, 'epoch': 0.48}) +[2026-06-08 19:51:00,525][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.10844963788986206, 'train/grad_norm_pre_clip': 1.709022045135498, 'timing/data': 0.010362917557358742, 'timing/model': 1.1640820913016796, 'learning_rate/qwen_vl_interface': 5.047829339287557e-07, 'learning_rate/action_model': 5.047829339287557e-06, 'epoch': 0.49}) +[2026-06-08 19:53:59,015][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.014726122841238976, 'train/grad_norm_pre_clip': 0.3275001645088196, 'timing/data': 0.011639615520834923, 'timing/model': 1.1710222605615854, 'learning_rate/qwen_vl_interface': 5.01195884023094e-07, 'learning_rate/action_model': 5.0119588402309395e-06, 'epoch': 0.49}) +[2026-06-08 20:13:59,298][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-08 20:13:59,301][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.18050819635391235, 'train/grad_norm_pre_clip': 2.5870230197906494, 'eval/loss': 0.08151376247406006, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 385.3800000000152, 'rl_games_eval/mid_train/mean_length': 3126.0, 'rl_games_eval/mid_train/std_reward': 68.09335944129685, 'rl_games_eval/mid_train/std_length': 547.2461968803439, 'rl_games_eval/mid_train/macro_mean_reward': 385.3800000000152, 'rl_games_eval/mid_train/macro_mean_length': 3126.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 385.3800000000152, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 3126.0, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 68.09335944129685, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 547.2461968803439, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 385.3800000000152, 'checkpoint/best_score': 416.6200000000207, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.009180288761854172, 'timing/model': 1.3202227540314198, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 0.49}) +[2026-06-08 20:13:59,302][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1500_state +[2026-06-08 20:13:59,306][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 20:14:05,420][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1500_state/model.safetensors +[2026-06-08 20:14:17,329][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1500_state/optimizer.bin +[2026-06-08 20:14:17,330][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1500_state/sampler.bin +[2026-06-08 20:14:17,331][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-08 20:14:17,333][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-08 20:14:17,408][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-08 20:14:17,409][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 20:14:17,431][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 20:14:17,431][starVLA.training.train_starvla][INFO] - Training complete. Best checkpoint step=1100 score=416.6200000000207 path=/workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state +[2026-06-08 20:14:17,432][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1_tmp/checkpoints/best_state +[2026-06-08 20:14:18,735][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-08 20:14:29,216][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-08 20:14:29,216][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-08 20:14:29,217][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-08 20:14:29,219][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-08 20:14:29,737][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-08 20:40:10,801][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-08 20:40:12,216][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..670bc6974d1e212c429f2c026f4feb746ff9315f --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/summary.jsonl @@ -0,0 +1,15 @@ +{"steps": 100} +{"steps": 200} +{"steps": 300} +{"steps": 400} +{"steps": 500} +{"steps": 600} +{"steps": 700} +{"steps": 800} +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/checkpoints/best_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/checkpoints/best_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..2845503555aac40c333c9c69b21a3e7c7c3cf17f --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/checkpoints/best_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6abc545e6b1abb272b7829b5676f412eebf33cfe0f21ba56368eb619c4610d83 +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/checkpoints/steps_1500_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/checkpoints/steps_1500_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4790e4514c0d09032cb8a15579df7945b1a35df9 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/checkpoints/steps_1500_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:76504a6629f8a781c5cfdbeac1795bdc6e221ba0d1931eabaf12d5bf5d56ccc1 +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..a70cf6f9b84ff7d14c03a334ebb52e5e1ca39ea7 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/config.full.yaml @@ -0,0 +1,187 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_fix_latency_1 + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 1500 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: true + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state + resume_step: 500 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 1 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: true + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_flappy_fix_latency_1 +output_dir: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1 +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..3471ab5f4a6c40c32a7501247659e03cb9d10954 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/config.yaml @@ -0,0 +1,87 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: true + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + include_state: true + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1 +rl_games: + env_eval: + distributed_mode: none + enabled: true + fixed_episode_seeds: true + frameskip: 1 + image_size: 224 + latency: + prompt_map_path: /workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json + latency_seed_stride: 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 1 + max_steps_per_episode: 3600 + num_episodes: 5 + seed: 42 + task_description: '' + task_seed_stride: 0 + vectorized: + enabled: false + model_alias: openvla + task: flappy +run_id: openvla_flappy_fix_latency_1 +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + eval_interval: 100 + eval_num_batches: 20 + freeze_modules: '' + gradient_clipping: 1.0 + is_resume: true + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 10 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 1500 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state + resume_step: 500 + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..eb44f5b34913a6e471eedb366f2711691a39fc30 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8599907159805298, + 0.14000925421714783, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.34691014885902405, + 0.34691014885902405, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 108000, + "num_trajectories": 30 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..21befa5285c5011c34b3635c161a360dc52b63db --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8598055839538574, + 0.14019444584846497, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.3472478985786438, + 0.3472478985786438, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 36000, + "num_trajectories": 10 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/hydra/.hydra/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..a32cbd0631ee52c1ef1370ae33b5815762a39119 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/hydra/.hydra/config.yaml @@ -0,0 +1,185 @@ +framework: + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_fix_latency_1 + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 1500 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: true + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state + resume_step: 500 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 1 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: true + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_flappy_fix_latency_1 +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/hydra/.hydra/hydra.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..6ce307f683e06372a4a54d858586d11416219fd9 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/hydra/.hydra/hydra.yaml @@ -0,0 +1,246 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=single + - run_id=openvla_flappy_fix_latency_1 + - run_root_dir=/workspace/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=true + - checkpoint.save_pt_file=false + - trainer.is_resume=true + - trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state + - trainer.resume_step=500 + - trainer.max_train_steps=1500 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=20 + - trainer.logging_frequency=10 + - trainer.gradient_accumulation_steps=16 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9, 0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=2 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1 + - datasets.vla_data.data_mix=flappy_train__bridge + - datasets.vla_data.eval_data_mix=flappy_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=flappy + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=1 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=true + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[1] + - rl_games.env_eval.post_train.enabled=true + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=true,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=flappy_train__bridge,datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1,datasets.vla_data.eval_data_mix=flappy_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.video_backend=torchvision_av,env=flappy,framework.action_model.action_dim=7,framework.action_model.action_env_dim=2,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=1,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=true,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[1],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=true,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=flappy,run_id=openvla_flappy_fix_latency_1,run_root_dir=/workspace/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=20,trainer.gradient_accumulation_steps=16,trainer.gradient_clipping=1.0,trainer.is_resume=true,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=10,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=1500,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9, + 0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state,trainer.resume_step=500,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.2 + version_base: '1.1' + cwd: /workspace/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/hydra/.hydra/overrides.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..f292e1891479fe84c8e446e50016e5ab621418cd --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/hydra/.hydra/overrides.yaml @@ -0,0 +1,85 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=single +- run_id=openvla_flappy_fix_latency_1 +- run_root_dir=/workspace/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=true +- checkpoint.save_pt_file=false +- trainer.is_resume=true +- trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state +- trainer.resume_step=500 +- trainer.max_train_steps=1500 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=20 +- trainer.logging_frequency=10 +- trainer.gradient_accumulation_steps=16 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9, 0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=2 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_1 +- datasets.vla_data.data_mix=flappy_train__bridge +- datasets.vla_data.eval_data_mix=flappy_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=flappy +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=1 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_1/flappy_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=true +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[1] +- rl_games.env_eval.post_train.enabled=true +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/hydra/train_starvla_hydra.log b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..0d90aeba6b7e89e8e2965c4e8978b8ce9c3f0960 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/hydra/train_starvla_hydra.log @@ -0,0 +1,570 @@ +[2026-06-07 16:30:43,927][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-07 16:30:43,967][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-07 16:30:44,922][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-07 16:30:45,468][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-07 16:30:45,468][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-07 16:30:45,468][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-07 16:30:45,468][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-07 16:30:45,468][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-07 16:30:45,468][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-07 16:30:45,468][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-07 16:30:45,468][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-07 16:30:45,468][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-07 16:30:45,468][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-07 16:30:45,469][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-07 16:30:45,469][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-07 16:30:45,469][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-07 16:30:45,469][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-07 16:30:45,469][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-07 16:30:45,469][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-07 16:30:45,469][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-07 16:30:45,469][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-07 16:30:45,469][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-07 16:30:45,469][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-07 16:30:45,469][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-07 16:30:45,469][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-07 16:30:45,469][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-07 16:30:45,469][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-07 16:30:45,469][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-07 16:30:45,470][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-07 16:30:45,470][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-07 16:30:45,470][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-07 16:30:45,470][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-07 16:30:45,470][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-07 16:30:45,470][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-07 16:30:45,487][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-07 16:30:45,509][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/config.full.yaml +[2026-06-07 16:30:45,514][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/config.yaml +[2026-06-07 16:30:54,500][starVLA.training.train_starvla][INFO] - Loaded pretrained checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt, steps: 0 +[2026-06-07 16:30:59,724][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-07 16:30:59,725][starVLA.training.train_starvla][INFO] - Total optimization steps = 1500 +[2026-06-07 16:30:59,725][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-07 16:30:59,725][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-07 16:30:59,725][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-07 16:32:04,017][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-07 16:32:04,066][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-07 16:32:05,076][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-07 16:32:05,678][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-07 16:32:05,678][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-07 16:32:05,678][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-07 16:32:05,678][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-07 16:32:05,679][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-07 16:32:05,679][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-07 16:32:05,679][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-07 16:32:05,679][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-07 16:32:05,679][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-07 16:32:05,679][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-07 16:32:05,679][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-07 16:32:05,679][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-07 16:32:05,679][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-07 16:32:05,679][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-07 16:32:05,679][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-07 16:32:05,679][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-07 16:32:05,679][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-07 16:32:05,679][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-07 16:32:05,679][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-07 16:32:05,680][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-07 16:32:05,680][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-07 16:32:05,680][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-07 16:32:05,680][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-07 16:32:05,680][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-07 16:32:05,680][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-07 16:32:05,680][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-07 16:32:05,680][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-07 16:32:05,680][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-07 16:32:05,680][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-07 16:32:05,680][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-07 16:32:05,680][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-07 16:32:05,699][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-07 16:32:05,723][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/config.full.yaml +[2026-06-07 16:32:05,729][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/config.yaml +[2026-06-07 16:32:16,846][starVLA.training.train_starvla][INFO] - Loaded pretrained checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt, steps: 0 +[2026-06-07 16:32:21,607][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-07 16:32:21,608][starVLA.training.train_starvla][INFO] - Total optimization steps = 1500 +[2026-06-07 16:32:21,608][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-07 16:32:21,608][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-07 16:32:21,608][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-07 16:35:08,455][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.40972667932510376, 'train/grad_norm_pre_clip': 1.1384344100952148, 'timing/data': 0.009434636682271957, 'timing/model': 1.0455865375697613, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-07 16:37:52,368][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.5045821070671082, 'train/grad_norm_pre_clip': 0.7050780057907104, 'timing/data': 0.012356413528323174, 'timing/model': 1.0602446217089891, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-07 16:40:35,577][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.4222692549228668, 'train/grad_norm_pre_clip': 0.8169273734092712, 'timing/data': 0.014894809573888779, 'timing/model': 1.0371364764869213, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-07 16:43:19,777][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.44719624519348145, 'train/grad_norm_pre_clip': 0.3524123430252075, 'timing/data': 0.012003527954220772, 'timing/model': 1.0599190052598715, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.01}) +[2026-06-07 16:46:00,665][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.37791118025779724, 'train/grad_norm_pre_clip': 0.1847039759159088, 'timing/data': 0.012127751484513283, 'timing/model': 1.0932948794215918, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.01}) +[2026-06-07 16:48:43,877][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.4370933473110199, 'train/grad_norm_pre_clip': 0.5419773459434509, 'timing/data': 0.011265350505709648, 'timing/model': 1.0562407467514277, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.02}) +[2026-06-07 16:51:32,637][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.3855835795402527, 'train/grad_norm_pre_clip': 5.593621730804443, 'timing/data': 0.010305197909474373, 'timing/model': 1.1239195056259632, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.02}) +[2026-06-07 16:54:14,975][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.2051604688167572, 'train/grad_norm_pre_clip': 1.5512973070144653, 'timing/data': 0.012161808088421822, 'timing/model': 1.0490388087928295, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.02}) +[2026-06-07 16:57:00,186][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.0995907112956047, 'train/grad_norm_pre_clip': 2.433213710784912, 'timing/data': 0.01332305558025837, 'timing/model': 1.0686747953295708, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.03}) +[2026-06-07 17:00:10,777][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmprn_678dd/test.c -o /tmp/tmprn_678dd/test.o +[2026-06-07 17:00:10,804][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmprn_678dd/test.o -laio -o /tmp/tmprn_678dd/a.out +[2026-06-07 17:00:11,407][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp6p8j3zov/test.c -o /tmp/tmp6p8j3zov/test.o +[2026-06-07 17:00:11,435][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp6p8j3zov/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp6p8j3zov/a.out +[2026-06-07 17:00:11,488][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpgukvczix/test.c -o /tmp/tmpgukvczix/test.o +[2026-06-07 17:00:11,514][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpgukvczix/test.o -laio -o /tmp/tmpgukvczix/a.out +[2026-06-07 17:00:37,830][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state +[2026-06-07 17:00:37,833][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-07 17:00:44,082][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/model.safetensors +[2026-06-07 17:00:53,037][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/optimizer.bin +[2026-06-07 17:00:53,037][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/sampler.bin +[2026-06-07 17:00:53,037][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/sampler_1.bin +[2026-06-07 17:00:53,038][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/random_states_0.pkl +[2026-06-07 17:00:53,051][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-07 17:00:53,051][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 100 +[2026-06-07 17:00:53,052][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.3456112742424011, 'train/grad_norm_pre_clip': 2.2301955223083496, 'eval/loss': 0.20318758487701416, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 5.3799999999999955, 'rl_games_eval/mid_train/mean_length': 59.4, 'rl_games_eval/mid_train/std_reward': 1.2106196760337218, 'rl_games_eval/mid_train/std_length': 7.709734106958553, 'rl_games_eval/mid_train/macro_mean_reward': 5.3799999999999955, 'rl_games_eval/mid_train/macro_mean_length': 59.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 5.3799999999999955, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 59.4, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 1.2106196760337218, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 7.709734106958553, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 5.3799999999999955, 'checkpoint/best_score': 5.3799999999999955, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.012100795283913612, 'timing/model': 1.0549878813326359, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.03}) +[2026-06-07 17:00:53,053][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_100_state +[2026-06-07 17:00:53,057][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-07 17:00:58,493][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_100_state/model.safetensors +[2026-06-07 17:01:08,769][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_100_state/optimizer.bin +[2026-06-07 17:01:08,770][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_100_state/sampler.bin +[2026-06-07 17:01:08,770][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_100_state/sampler_1.bin +[2026-06-07 17:01:08,771][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-07 17:01:08,774][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-07 17:01:08,774][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-07 17:01:08,785][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-07 17:04:09,767][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.07460933923721313, 'train/grad_norm_pre_clip': 0.5115079283714294, 'timing/data': 0.009603539481759071, 'timing/model': 1.3407037928700447, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.03}) +[2026-06-07 17:07:08,733][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.3298341929912567, 'train/grad_norm_pre_clip': 5.462419033050537, 'timing/data': 0.010346455499529839, 'timing/model': 1.1934386808425188, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.04}) +[2026-06-07 17:10:07,738][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.25525277853012085, 'train/grad_norm_pre_clip': 1.5327911376953125, 'timing/data': 0.009997943416237831, 'timing/model': 1.1579302698373795, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.04}) +[2026-06-07 17:13:08,430][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 1.200295090675354, 'train/grad_norm_pre_clip': 9.142666816711426, 'timing/data': 0.010075287893414497, 'timing/model': 1.2877198569476604, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.04}) +[2026-06-07 17:16:08,311][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.17700213193893433, 'train/grad_norm_pre_clip': 1.2842075824737549, 'timing/data': 0.012422390282154083, 'timing/model': 1.1616858560591936, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.04}) +[2026-06-07 17:19:07,388][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.19879691302776337, 'train/grad_norm_pre_clip': 1.4221099615097046, 'timing/data': 0.00845215655863285, 'timing/model': 1.131260683760047, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.05}) +[2026-06-07 17:22:06,053][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.216658815741539, 'train/grad_norm_pre_clip': 1.664008378982544, 'timing/data': 0.009217539802193642, 'timing/model': 1.1325530521571636, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.05}) +[2026-06-07 17:25:06,465][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.234208345413208, 'train/grad_norm_pre_clip': 2.210663318634033, 'timing/data': 0.011434612795710564, 'timing/model': 1.158809607848525, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.05}) +[2026-06-07 17:28:06,459][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.08094080537557602, 'train/grad_norm_pre_clip': 0.840800940990448, 'timing/data': 0.011796224862337112, 'timing/model': 1.1650091037154198, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.06}) +[2026-06-07 17:32:06,755][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state +[2026-06-07 17:32:06,761][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-07 17:32:13,294][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/model.safetensors +[2026-06-07 17:32:22,376][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/optimizer.bin +[2026-06-07 17:32:22,376][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/sampler.bin +[2026-06-07 17:32:22,376][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/sampler_1.bin +[2026-06-07 17:32:22,377][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/random_states_0.pkl +[2026-06-07 17:32:22,391][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-07 17:32:22,391][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 200 +[2026-06-07 17:32:22,392][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.07714591175317764, 'train/grad_norm_pre_clip': 1.4482014179229736, 'eval/loss': 0.12977728247642517, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 10.919999999999991, 'rl_games_eval/mid_train/mean_length': 104.0, 'rl_games_eval/mid_train/std_reward': 3.873706235635334, 'rl_games_eval/mid_train/std_length': 32.37900554371613, 'rl_games_eval/mid_train/macro_mean_reward': 10.919999999999991, 'rl_games_eval/mid_train/macro_mean_length': 104.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 10.919999999999991, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 104.0, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 3.873706235635334, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 32.37900554371613, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 10.919999999999991, 'checkpoint/best_score': 10.919999999999991, 'checkpoint/best_step': 200.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.009356608614325523, 'timing/model': 1.1245429646223783, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.06}) +[2026-06-07 17:32:22,393][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_200_state +[2026-06-07 17:32:22,397][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-07 17:32:28,531][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_200_state/model.safetensors +[2026-06-07 17:32:40,547][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_200_state/optimizer.bin +[2026-06-07 17:32:40,548][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_200_state/sampler.bin +[2026-06-07 17:32:40,548][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_200_state/sampler_1.bin +[2026-06-07 17:32:40,549][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-07 17:32:40,599][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-07 17:32:40,599][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-07 17:32:40,611][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-07 17:35:38,607][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.18400982022285461, 'train/grad_norm_pre_clip': 2.58640456199646, 'timing/data': 0.010138573125004768, 'timing/model': 1.1828891038894653, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.06}) +[2026-06-07 17:38:45,262][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.25082725286483765, 'train/grad_norm_pre_clip': 1.0895270109176636, 'timing/data': 0.017413360998034477, 'timing/model': 1.1608536671847105, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.07}) +[2026-06-07 17:41:46,055][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.06938165426254272, 'train/grad_norm_pre_clip': 0.5139012336730957, 'timing/data': 0.007825439795851707, 'timing/model': 1.1339066959917545, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.07}) +[2026-06-07 17:44:45,746][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.08905667811632156, 'train/grad_norm_pre_clip': 0.8182042837142944, 'timing/data': 0.007963024079799652, 'timing/model': 1.1601866148412228, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.07}) +[2026-06-07 17:47:47,351][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.2074076235294342, 'train/grad_norm_pre_clip': 2.6054651737213135, 'timing/data': 0.010087233036756516, 'timing/model': 1.1602036952972412, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.07}) +[2026-06-07 17:50:47,059][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.23002703487873077, 'train/grad_norm_pre_clip': 1.5738258361816406, 'timing/data': 0.010551907122135162, 'timing/model': 1.1983008719980717, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.08}) +[2026-06-07 17:53:52,089][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.1662381887435913, 'train/grad_norm_pre_clip': 0.9941570162773132, 'timing/data': 0.010008111596107483, 'timing/model': 1.1893588863313198, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.08}) +[2026-06-07 17:56:52,293][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.535886824131012, 'train/grad_norm_pre_clip': 3.7239692211151123, 'timing/data': 0.008440639823675156, 'timing/model': 1.1362939961254597, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.08}) +[2026-06-07 17:59:53,659][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.16819174587726593, 'train/grad_norm_pre_clip': 0.8717544078826904, 'timing/data': 0.010105285793542862, 'timing/model': 1.1636423096060753, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.09}) +[2026-06-07 18:03:55,769][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 300 +[2026-06-07 18:03:55,772][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.20771968364715576, 'train/grad_norm_pre_clip': 2.0942671298980713, 'eval/loss': 0.1044808179140091, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 7.299999999999988, 'rl_games_eval/mid_train/mean_length': 76.8, 'rl_games_eval/mid_train/std_reward': 3.5071355833500264, 'rl_games_eval/mid_train/std_length': 28.85411582426327, 'rl_games_eval/mid_train/macro_mean_reward': 7.299999999999988, 'rl_games_eval/mid_train/macro_mean_length': 76.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 7.299999999999988, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 76.8, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 3.5071355833500264, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 28.85411582426327, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 7.299999999999988, 'checkpoint/best_score': 10.919999999999991, 'checkpoint/best_step': 200.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.009119858965277672, 'timing/model': 1.2588670272380114, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.09}) +[2026-06-07 18:03:55,773][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_300_state +[2026-06-07 18:03:55,779][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-07 18:04:05,326][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_300_state/model.safetensors +[2026-06-07 18:04:15,975][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_300_state/optimizer.bin +[2026-06-07 18:04:15,976][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_300_state/sampler.bin +[2026-06-07 18:04:15,977][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_300_state/sampler_1.bin +[2026-06-07 18:04:15,978][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-07 18:04:16,044][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-07 18:04:16,044][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-07 18:04:16,056][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-07 18:07:20,166][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.12740837037563324, 'train/grad_norm_pre_clip': 0.9855213165283203, 'timing/data': 0.010822582989931107, 'timing/model': 1.258016910403967, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.09}) +[2026-06-07 18:10:25,150][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.10436194390058517, 'train/grad_norm_pre_clip': 0.6659932136535645, 'timing/data': 0.010692961513996124, 'timing/model': 1.1714771781116724, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.09}) +[2026-06-07 18:13:26,840][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.2453145533800125, 'train/grad_norm_pre_clip': 0.8219911456108093, 'timing/data': 0.008890338242053986, 'timing/model': 1.1336971875280142, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.1}) +[2026-06-07 18:16:30,268][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.027010083198547363, 'train/grad_norm_pre_clip': 0.38073837757110596, 'timing/data': 0.011660106480121613, 'timing/model': 1.252036638557911, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.1}) +[2026-06-07 18:19:33,228][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.06435414403676987, 'train/grad_norm_pre_clip': 0.5984577536582947, 'timing/data': 0.010464858263731003, 'timing/model': 1.2132349908351898, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.1}) +[2026-06-07 18:22:31,526][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.1885741651058197, 'train/grad_norm_pre_clip': 0.7304105758666992, 'timing/data': 0.008319756016135216, 'timing/model': 1.1266974620521069, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.11}) +[2026-06-07 18:25:32,568][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.13568115234375, 'train/grad_norm_pre_clip': 0.6293562054634094, 'timing/data': 0.010501278564333916, 'timing/model': 1.1652131509035826, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.11}) +[2026-06-07 18:28:29,464][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.16861063241958618, 'train/grad_norm_pre_clip': 1.2721774578094482, 'timing/data': 0.008880836889147758, 'timing/model': 1.1263813395053148, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.11}) +[2026-06-07 18:31:28,240][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.24682487547397614, 'train/grad_norm_pre_clip': 2.8551833629608154, 'timing/data': 0.011184236034750938, 'timing/model': 1.1821705475449562, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.12}) +[2026-06-07 18:35:28,515][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 400 +[2026-06-07 18:35:28,517][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.1850443035364151, 'train/grad_norm_pre_clip': 0.8803691864013672, 'eval/loss': 0.12044084072113037, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 9.500000000000004, 'rl_games_eval/mid_train/mean_length': 95.2, 'rl_games_eval/mid_train/std_reward': 6.858571279792904, 'rl_games_eval/mid_train/std_length': 55.35846818689983, 'rl_games_eval/mid_train/macro_mean_reward': 9.500000000000004, 'rl_games_eval/mid_train/macro_mean_length': 95.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 9.500000000000004, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 95.2, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 6.858571279792904, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 55.35846818689983, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 9.500000000000004, 'checkpoint/best_score': 10.919999999999991, 'checkpoint/best_step': 200.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.011462461203336716, 'timing/model': 1.173296082764864, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.12}) +[2026-06-07 18:35:28,517][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_400_state +[2026-06-07 18:35:28,521][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-07 18:35:34,870][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_400_state/model.safetensors +[2026-06-07 18:35:44,541][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_400_state/optimizer.bin +[2026-06-07 18:35:44,542][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_400_state/sampler.bin +[2026-06-07 18:35:44,542][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_400_state/sampler_1.bin +[2026-06-07 18:35:44,543][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-07 18:35:44,612][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-07 18:35:44,612][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-07 18:35:44,624][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-07 18:38:46,755][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.1973542720079422, 'train/grad_norm_pre_clip': 2.390324831008911, 'timing/data': 0.010565085336565971, 'timing/model': 1.2132190391421318, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.12}) +[2026-06-07 18:56:43,545][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-07 18:56:43,605][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-07 18:56:44,754][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-07 18:56:45,412][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-07 18:56:45,413][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-07 18:56:45,413][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-07 18:56:45,413][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-07 18:56:45,413][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-07 18:56:45,413][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-07 18:56:45,413][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-07 18:56:45,413][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-07 18:56:45,413][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-07 18:56:45,413][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-07 18:56:45,413][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-07 18:56:45,413][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-07 18:56:45,413][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-07 18:56:45,414][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-07 18:56:45,414][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-07 18:56:45,414][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-07 18:56:45,414][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-07 18:56:45,414][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-07 18:56:45,414][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-07 18:56:45,414][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-07 18:56:45,414][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-07 18:56:45,414][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-07 18:56:45,414][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-07 18:56:45,414][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-07 18:56:45,414][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-07 18:56:45,414][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-07 18:56:45,414][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-07 18:56:45,414][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-07 18:56:45,415][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-07 18:56:45,415][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-07 18:56:45,415][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-07 18:56:45,432][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-07 18:56:45,454][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/config.full.yaml +[2026-06-07 18:56:45,460][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/config.yaml +[2026-06-07 18:56:45,460][starVLA.training.train_starvla][INFO] - Loaded best checkpoint metadata: step=200 score=10.919999999999991 +[2026-06-07 18:56:45,460][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_400_state, steps: 400 +[2026-06-07 18:56:47,350][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_400_state +[2026-06-07 18:56:48,711][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-07 18:56:59,440][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-07 18:56:59,440][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-07 18:56:59,440][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-07 18:56:59,442][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-07 18:56:59,816][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-07 18:57:01,218][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-07 18:57:01,218][starVLA.training.train_starvla][INFO] - Total optimization steps = 1500 +[2026-06-07 18:57:01,219][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-07 18:57:01,219][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-07 18:57:01,219][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-07 18:59:43,583][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.1228136271238327, 'train/grad_norm_pre_clip': 0.4293844401836395, 'timing/data': 0.010983768850564957, 'timing/model': 0.9886513445526361, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.12}) +[2026-06-07 19:02:24,383][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.0812675803899765, 'train/grad_norm_pre_clip': 0.6596367955207825, 'timing/data': 0.009314844384789467, 'timing/model': 1.2647565715014935, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.12}) +[2026-06-07 19:05:06,384][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.17513009905815125, 'train/grad_norm_pre_clip': 1.1503396034240723, 'timing/data': 0.014890866354107857, 'timing/model': 1.053228110074997, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.13}) +[2026-06-07 19:07:41,054][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.2521124482154846, 'train/grad_norm_pre_clip': 2.0160529613494873, 'timing/data': 0.00887388363480568, 'timing/model': 0.955820556730032, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.13}) +[2026-06-07 19:10:21,274][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.04664968326687813, 'train/grad_norm_pre_clip': 0.769603967666626, 'timing/data': 0.010847160592675209, 'timing/model': 1.0518687777221203, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.13}) +[2026-06-07 19:13:02,286][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.05235924944281578, 'train/grad_norm_pre_clip': 1.1393828392028809, 'timing/data': 0.009352866560220718, 'timing/model': 1.1739689633250237, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.14}) +[2026-06-07 19:15:43,978][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.08009239286184311, 'train/grad_norm_pre_clip': 0.9740587472915649, 'timing/data': 0.010713240131735802, 'timing/model': 1.0551043394953012, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.14}) +[2026-06-07 19:18:24,273][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.11530184000730515, 'train/grad_norm_pre_clip': 1.5040565729141235, 'timing/data': 0.010515566915273666, 'timing/model': 1.0490440800786018, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.14}) +[2026-06-07 19:21:01,082][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.163697749376297, 'train/grad_norm_pre_clip': 1.1521644592285156, 'timing/data': 0.00999884307384491, 'timing/model': 1.0148118063807487, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.15}) +[2026-06-07 19:24:08,026][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpww4071xl/test.c -o /tmp/tmpww4071xl/test.o +[2026-06-07 19:24:08,054][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpww4071xl/test.o -laio -o /tmp/tmpww4071xl/a.out +[2026-06-07 19:24:08,635][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp_u1twfts/test.c -o /tmp/tmp_u1twfts/test.o +[2026-06-07 19:24:08,657][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp_u1twfts/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp_u1twfts/a.out +[2026-06-07 19:24:08,715][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpojbwjno4/test.c -o /tmp/tmpojbwjno4/test.o +[2026-06-07 19:24:08,738][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpojbwjno4/test.o -laio -o /tmp/tmpojbwjno4/a.out +[2026-06-07 19:24:46,815][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-07 19:24:46,817][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.07178635150194168, 'train/grad_norm_pre_clip': 1.148215651512146, 'eval/loss': 0.09653764218091965, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 9.939999999999996, 'rl_games_eval/mid_train/mean_length': 97.8, 'rl_games_eval/mid_train/std_reward': 4.662445710139701, 'rl_games_eval/mid_train/std_length': 39.846706262876985, 'rl_games_eval/mid_train/macro_mean_reward': 9.939999999999996, 'rl_games_eval/mid_train/macro_mean_length': 97.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 9.939999999999996, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 97.8, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 4.662445710139701, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 39.846706262876985, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 9.939999999999996, 'checkpoint/best_score': 10.919999999999991, 'checkpoint/best_step': 200.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.010842638090252876, 'timing/model': 1.0533561203628778, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.15}) +[2026-06-07 19:24:46,817][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state +[2026-06-07 19:24:46,823][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-07 19:24:55,750][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state/model.safetensors +[2026-06-07 19:25:10,692][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state/optimizer.bin +[2026-06-07 19:25:10,692][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state/sampler.bin +[2026-06-07 19:25:10,693][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state/sampler_1.bin +[2026-06-07 19:25:10,694][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-07 19:25:10,697][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-07 19:25:10,697][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-07 19:25:10,709][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-07 19:28:19,788][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.09402173012495041, 'train/grad_norm_pre_clip': 1.0407930612564087, 'timing/data': 0.010239999741315842, 'timing/model': 1.3717402592301369, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.15}) +[2026-06-07 19:31:26,946][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.28978878259658813, 'train/grad_norm_pre_clip': 1.951108694076538, 'timing/data': 0.0102642010897398, 'timing/model': 1.1550930049270391, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.15}) +[2026-06-07 19:34:40,473][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.1974879950284958, 'train/grad_norm_pre_clip': 1.8955292701721191, 'timing/data': 0.014281908050179482, 'timing/model': 1.2708651162683964, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.16}) +[2026-06-07 19:37:52,474][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.31727924942970276, 'train/grad_norm_pre_clip': 3.56615948677063, 'timing/data': 0.010619759559631348, 'timing/model': 1.2090384121984243, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.16}) +[2026-06-07 19:41:08,375][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.0319693349301815, 'train/grad_norm_pre_clip': 0.9291955232620239, 'timing/data': 0.010011469945311546, 'timing/model': 1.2577106449753046, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.16}) +[2026-06-07 19:44:13,860][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.16633310914039612, 'train/grad_norm_pre_clip': 1.2294306755065918, 'timing/data': 0.00951414555311203, 'timing/model': 1.2534633819013834, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.17}) +[2026-06-08 02:28:21,752][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-08 02:28:21,811][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-08 02:28:23,421][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-08 02:28:24,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-08 02:28:24,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-08 02:28:24,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-08 02:28:24,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-08 02:28:24,181][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-08 02:28:24,181][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-08 02:28:24,181][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-08 02:28:24,181][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-08 02:28:24,181][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-08 02:28:24,181][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-08 02:28:24,181][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-08 02:28:24,181][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-08 02:28:24,182][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-08 02:28:24,182][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-08 02:28:24,182][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-08 02:28:24,182][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-08 02:28:24,182][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-08 02:28:24,182][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-08 02:28:24,182][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-08 02:28:24,183][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-08 02:28:24,183][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-08 02:28:24,183][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-08 02:28:24,183][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-08 02:28:24,183][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-08 02:28:24,183][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-08 02:28:24,183][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-08 02:28:24,183][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-08 02:28:24,183][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-08 02:28:24,184][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-08 02:28:24,184][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-08 02:28:24,184][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-08 02:28:24,211][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-08 02:28:24,249][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/config.full.yaml +[2026-06-08 02:28:24,258][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/config.yaml +[2026-06-08 02:28:24,259][starVLA.training.train_starvla][INFO] - Loaded best checkpoint metadata: step=200 score=10.919999999999991 +[2026-06-08 02:28:24,259][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state, steps: 500 +[2026-06-08 02:28:26,549][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_500_state +[2026-06-08 02:28:29,531][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-08 02:28:40,224][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-08 02:28:40,224][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-08 02:28:40,225][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-08 02:28:40,226][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-08 02:28:40,516][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-08 02:28:41,712][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-08 02:28:41,712][starVLA.training.train_starvla][INFO] - Total optimization steps = 1500 +[2026-06-08 02:28:41,712][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-08 02:28:41,712][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-08 02:28:41,712][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-08 02:31:22,421][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.14785857498645782, 'train/grad_norm_pre_clip': 1.0472160577774048, 'timing/data': 0.00988289900124073, 'timing/model': 1.0121050830930471, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.15}) +[2026-06-08 02:34:03,224][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.04174887761473656, 'train/grad_norm_pre_clip': 0.5444515943527222, 'timing/data': 0.00963616743683815, 'timing/model': 1.0578863862901926, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.15}) +[2026-06-08 02:36:41,917][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.11629770696163177, 'train/grad_norm_pre_clip': 1.1295074224472046, 'timing/data': 0.010323092341423035, 'timing/model': 1.055806016549468, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.16}) +[2026-06-08 02:39:21,995][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.15640917420387268, 'train/grad_norm_pre_clip': 1.4473106861114502, 'timing/data': 0.009243069216609001, 'timing/model': 1.0187244340777397, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.16}) +[2026-06-08 02:41:59,216][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.03679528087377548, 'train/grad_norm_pre_clip': 0.8531883955001831, 'timing/data': 0.008215982466936111, 'timing/model': 1.0582027323544025, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.16}) +[2026-06-08 02:44:38,829][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.039581410586833954, 'train/grad_norm_pre_clip': 2.0333027839660645, 'timing/data': 0.01041218638420105, 'timing/model': 1.0539453774690628, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.17}) +[2026-06-08 02:47:18,238][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.048252761363983154, 'train/grad_norm_pre_clip': 0.7267184257507324, 'timing/data': 0.010557536035776138, 'timing/model': 1.0554591342806816, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.17}) +[2026-06-08 02:49:59,226][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.11491948366165161, 'train/grad_norm_pre_clip': 2.2183756828308105, 'timing/data': 0.010399876162409782, 'timing/model': 1.057035032659769, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.17}) +[2026-06-08 02:52:39,427][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.12886273860931396, 'train/grad_norm_pre_clip': 0.30035489797592163, 'timing/data': 0.01050819642841816, 'timing/model': 1.0587196946144104, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.17}) +[2026-06-08 02:55:45,941][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp9mzwz8xq/test.c -o /tmp/tmp9mzwz8xq/test.o +[2026-06-08 02:55:45,967][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp9mzwz8xq/test.o -laio -o /tmp/tmp9mzwz8xq/a.out +[2026-06-08 02:55:46,554][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp5by_avfl/test.c -o /tmp/tmp5by_avfl/test.o +[2026-06-08 02:55:46,581][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp5by_avfl/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp5by_avfl/a.out +[2026-06-08 02:55:46,638][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpxux0dukn/test.c -o /tmp/tmpxux0dukn/test.o +[2026-06-08 02:55:46,662][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpxux0dukn/test.o -laio -o /tmp/tmpxux0dukn/a.out +[2026-06-08 02:57:07,933][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state +[2026-06-08 02:57:07,939][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 02:57:13,977][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/model.safetensors +[2026-06-08 02:57:32,110][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/optimizer.bin +[2026-06-08 02:57:32,110][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/sampler.bin +[2026-06-08 02:57:32,111][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/sampler_1.bin +[2026-06-08 02:57:32,112][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/random_states_0.pkl +[2026-06-08 02:57:32,127][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 02:57:32,127][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 600 +[2026-06-08 02:57:32,128][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.04446437954902649, 'train/grad_norm_pre_clip': 2.7018885612487793, 'eval/loss': 0.09853879362344742, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 24.12000000000009, 'rl_games_eval/mid_train/mean_length': 214.4, 'rl_games_eval/mid_train/std_reward': 21.59920368902535, 'rl_games_eval/mid_train/std_length': 174.40137614135963, 'rl_games_eval/mid_train/macro_mean_reward': 24.12000000000009, 'rl_games_eval/mid_train/macro_mean_length': 214.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 24.12000000000009, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 214.4, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 21.59920368902535, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 174.40137614135963, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 24.12000000000009, 'checkpoint/best_score': 24.12000000000009, 'checkpoint/best_step': 600.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.007813045755028725, 'timing/model': 0.9569010231643915, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.18}) +[2026-06-08 02:57:32,129][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_600_state +[2026-06-08 02:57:32,133][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 02:57:37,971][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_600_state/model.safetensors +[2026-06-08 02:57:52,790][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_600_state/optimizer.bin +[2026-06-08 02:57:52,791][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_600_state/sampler.bin +[2026-06-08 02:57:52,791][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_600_state/sampler_1.bin +[2026-06-08 02:57:52,792][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_600_state/random_states_0.pkl +[2026-06-08 02:57:53,156][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600 +[2026-06-08 02:57:53,156][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 02:57:53,171][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 03:01:04,318][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.08884066343307495, 'train/grad_norm_pre_clip': 1.7461992502212524, 'timing/data': 0.00979464128613472, 'timing/model': 1.3480573520064354, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.18}) +[2026-06-08 03:04:13,622][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.2312059849500656, 'train/grad_norm_pre_clip': 2.435190200805664, 'timing/data': 0.010374035686254501, 'timing/model': 1.2583372090011835, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.18}) +[2026-06-08 03:07:26,103][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.21449053287506104, 'train/grad_norm_pre_clip': 2.0557806491851807, 'timing/data': 0.01365184411406517, 'timing/model': 1.2111754715442657, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.19}) +[2026-06-08 03:10:43,509][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.6169573068618774, 'train/grad_norm_pre_clip': 5.35631799697876, 'timing/data': 0.009722961112856865, 'timing/model': 1.438329540193081, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.19}) +[2026-06-08 03:13:58,031][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.04080956056714058, 'train/grad_norm_pre_clip': 0.8581114411354065, 'timing/data': 0.011984208598732948, 'timing/model': 1.261653807014227, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.19}) +[2026-06-08 03:17:10,730][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.18923026323318481, 'train/grad_norm_pre_clip': 0.9931254386901855, 'timing/data': 0.011647291481494904, 'timing/model': 1.260031308978796, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.2}) +[2026-06-08 03:20:23,256][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.09585290402173996, 'train/grad_norm_pre_clip': 1.0940170288085938, 'timing/data': 0.011710893362760544, 'timing/model': 1.2714359946548939, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.2}) +[2026-06-08 03:23:37,833][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.13521376252174377, 'train/grad_norm_pre_clip': 3.671029806137085, 'timing/data': 0.011679081246256828, 'timing/model': 1.559337843209505, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.2}) +[2026-06-08 03:26:50,794][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.062086086720228195, 'train/grad_norm_pre_clip': 1.0403454303741455, 'timing/data': 0.009972494095563889, 'timing/model': 1.2417141627520323, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.2}) +[2026-06-08 03:31:02,555][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 700 +[2026-06-08 03:31:02,557][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.09408611059188843, 'train/grad_norm_pre_clip': 1.4639136791229248, 'eval/loss': 0.1011648178100586, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 9.979999999999983, 'rl_games_eval/mid_train/mean_length': 98.2, 'rl_games_eval/mid_train/std_reward': 3.434763456193156, 'rl_games_eval/mid_train/std_length': 25.43540839066674, 'rl_games_eval/mid_train/macro_mean_reward': 9.979999999999983, 'rl_games_eval/mid_train/macro_mean_length': 98.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 9.979999999999983, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 98.2, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 3.434763456193156, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 25.43540839066674, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 9.979999999999983, 'checkpoint/best_score': 24.12000000000009, 'checkpoint/best_step': 600.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.008894853293895721, 'timing/model': 1.1642885133624077, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.21}) +[2026-06-08 03:31:02,557][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_700_state +[2026-06-08 03:31:02,561][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 03:31:08,644][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_700_state/model.safetensors +[2026-06-08 03:31:22,418][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_700_state/optimizer.bin +[2026-06-08 03:31:22,419][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_700_state/sampler.bin +[2026-06-08 03:31:22,419][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_700_state/sampler_1.bin +[2026-06-08 03:31:22,420][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_700_state/random_states_0.pkl +[2026-06-08 03:31:22,505][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700 +[2026-06-08 03:31:22,505][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 03:31:22,518][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 03:34:30,485][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.19694210588932037, 'train/grad_norm_pre_clip': 2.3134491443634033, 'timing/data': 0.007648276165127754, 'timing/model': 1.156448507681489, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.21}) +[2026-06-08 03:37:48,551][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.33667615056037903, 'train/grad_norm_pre_clip': 1.7443628311157227, 'timing/data': 0.010220933705568314, 'timing/model': 1.1976180709898472, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.21}) +[2026-06-08 03:40:57,401][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.1062043309211731, 'train/grad_norm_pre_clip': 3.6828112602233887, 'timing/data': 0.009909875690937042, 'timing/model': 1.2409291174262762, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.22}) +[2026-06-08 03:44:07,118][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.1382843255996704, 'train/grad_norm_pre_clip': 1.0776262283325195, 'timing/data': 0.008846268057823181, 'timing/model': 1.2548139616847038, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.22}) +[2026-06-08 03:47:23,031][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.24488884210586548, 'train/grad_norm_pre_clip': 3.646770715713501, 'timing/data': 0.012449927628040314, 'timing/model': 1.2564878407865763, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.22}) +[2026-06-08 03:50:38,207][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.20092175900936127, 'train/grad_norm_pre_clip': 1.4589815139770508, 'timing/data': 0.010395601391792297, 'timing/model': 1.1595692727714777, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.23}) +[2026-06-08 03:53:54,518][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.04230382293462753, 'train/grad_norm_pre_clip': 0.5965043902397156, 'timing/data': 0.01012495718896389, 'timing/model': 1.448968980461359, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.23}) +[2026-06-08 03:57:06,624][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.12880125641822815, 'train/grad_norm_pre_clip': 3.2280399799346924, 'timing/data': 0.011633196845650673, 'timing/model': 1.2479083891957998, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.23}) +[2026-06-08 04:00:24,321][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.22798243165016174, 'train/grad_norm_pre_clip': 2.152913808822632, 'timing/data': 0.010926170274615288, 'timing/model': 1.2523880619555712, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.23}) +[2026-06-08 04:04:36,030][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 800 +[2026-06-08 04:04:36,031][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.026743575930595398, 'train/grad_norm_pre_clip': 1.0122240781784058, 'eval/loss': 0.10519333928823471, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 7.699999999999989, 'rl_games_eval/mid_train/mean_length': 80.8, 'rl_games_eval/mid_train/std_reward': 4.08215629294126, 'rl_games_eval/mid_train/std_length': 30.32754523531372, 'rl_games_eval/mid_train/macro_mean_reward': 7.699999999999989, 'rl_games_eval/mid_train/macro_mean_length': 80.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 7.699999999999989, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 80.8, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 4.08215629294126, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 30.32754523531372, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 7.699999999999989, 'checkpoint/best_score': 24.12000000000009, 'checkpoint/best_step': 600.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.010297557339072227, 'timing/model': 1.2559085674583912, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.24}) +[2026-06-08 04:04:36,032][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_800_state +[2026-06-08 04:04:36,035][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 04:04:42,337][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_800_state/model.safetensors +[2026-06-08 04:04:55,707][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_800_state/optimizer.bin +[2026-06-08 04:04:55,707][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_800_state/sampler.bin +[2026-06-08 04:04:55,707][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_800_state/sampler_1.bin +[2026-06-08 04:04:55,708][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_800_state/random_states_0.pkl +[2026-06-08 04:04:55,803][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800 +[2026-06-08 04:04:55,803][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 04:04:55,815][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 04:08:03,431][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.08235027641057968, 'train/grad_norm_pre_clip': 0.8081689476966858, 'timing/data': 0.010712772607803345, 'timing/model': 1.2615945506840944, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.24}) +[2026-06-08 04:11:17,626][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.11566474288702011, 'train/grad_norm_pre_clip': 1.46780526638031, 'timing/data': 0.011909419670701027, 'timing/model': 1.2576640285551548, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.24}) +[2026-06-08 04:14:32,121][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.22728708386421204, 'train/grad_norm_pre_clip': 1.0950161218643188, 'timing/data': 0.010704748332500458, 'timing/model': 1.2469014767557383, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.25}) +[2026-06-08 04:17:43,623][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.0630272850394249, 'train/grad_norm_pre_clip': 0.7764682769775391, 'timing/data': 0.010154886171221733, 'timing/model': 1.2599844448268414, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.25}) +[2026-06-08 04:20:56,829][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.08732070028781891, 'train/grad_norm_pre_clip': 0.467477023601532, 'timing/data': 0.010380668565630913, 'timing/model': 1.2687833160161972, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.25}) +[2026-06-08 04:24:11,043][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.199907124042511, 'train/grad_norm_pre_clip': 0.7404685616493225, 'timing/data': 0.00981375016272068, 'timing/model': 1.3720009662210941, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.25}) +[2026-06-08 04:27:29,829][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.12482769042253494, 'train/grad_norm_pre_clip': 0.583684504032135, 'timing/data': 0.012422095984220505, 'timing/model': 1.2570643611252308, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.26}) +[2026-06-08 04:30:45,417][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.16042155027389526, 'train/grad_norm_pre_clip': 0.7780199646949768, 'timing/data': 0.013919441029429436, 'timing/model': 1.3473796378821135, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.26}) +[2026-06-08 04:34:00,736][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.24358049035072327, 'train/grad_norm_pre_clip': 1.528757929801941, 'timing/data': 0.010966861620545387, 'timing/model': 1.1853130254894495, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.26}) +[2026-06-08 04:38:23,716][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 900 +[2026-06-08 04:38:23,718][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.10494193434715271, 'train/grad_norm_pre_clip': 1.7167644500732422, 'eval/loss': 0.10630352795124054, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 13.860000000000031, 'rl_games_eval/mid_train/mean_length': 129.8, 'rl_games_eval/mid_train/std_reward': 10.23241906882245, 'rl_games_eval/mid_train/std_length': 83.25959404176795, 'rl_games_eval/mid_train/macro_mean_reward': 13.860000000000031, 'rl_games_eval/mid_train/macro_mean_length': 129.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 13.860000000000031, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 129.8, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 10.23241906882245, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 83.25959404176795, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 13.860000000000031, 'checkpoint/best_score': 24.12000000000009, 'checkpoint/best_step': 600.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.010333351790904999, 'timing/model': 1.256254507228732, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.27}) +[2026-06-08 04:38:23,719][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_900_state +[2026-06-08 04:38:23,722][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 04:38:29,894][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_900_state/model.safetensors +[2026-06-08 04:38:43,706][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_900_state/optimizer.bin +[2026-06-08 04:38:43,707][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_900_state/sampler.bin +[2026-06-08 04:38:43,707][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_900_state/sampler_1.bin +[2026-06-08 04:38:43,708][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_900_state/random_states_0.pkl +[2026-06-08 04:38:43,798][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-06-08 04:38:43,798][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 04:38:43,811][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 04:41:50,410][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.1504046618938446, 'train/grad_norm_pre_clip': 1.7034003734588623, 'timing/data': 0.01012641191482544, 'timing/model': 1.2499719206243753, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.27}) +[2026-06-08 04:45:05,433][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.026497211307287216, 'train/grad_norm_pre_clip': 0.28330326080322266, 'timing/data': 0.010129151865839958, 'timing/model': 1.3549129571765661, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.27}) +[2026-06-08 04:48:24,431][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.1441197246313095, 'train/grad_norm_pre_clip': 0.888932466506958, 'timing/data': 0.0006872452795505524, 'timing/model': 1.232828725129366, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 0.28}) +[2026-06-08 04:51:41,320][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.06046869978308678, 'train/grad_norm_pre_clip': 0.6223641633987427, 'timing/data': 0.0006999224424362183, 'timing/model': 1.2607584223151207, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.28}) +[2026-06-08 04:54:56,115][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.03481486812233925, 'train/grad_norm_pre_clip': 0.8553818464279175, 'timing/data': 0.001046663150191307, 'timing/model': 1.2562530543655157, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.28}) +[2026-06-08 04:58:05,620][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.22162754833698273, 'train/grad_norm_pre_clip': 4.34283971786499, 'timing/data': 0.0006981156766414642, 'timing/model': 1.260053401812911, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.28}) +[2026-06-08 05:01:23,328][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.06828710436820984, 'train/grad_norm_pre_clip': 0.6473269462585449, 'timing/data': 0.011569458991289139, 'timing/model': 1.256771007552743, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 0.29}) +[2026-06-08 05:04:36,827][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.06886345148086548, 'train/grad_norm_pre_clip': 1.0889666080474854, 'timing/data': 0.010370265692472458, 'timing/model': 1.2603094186633825, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.29}) +[2026-06-08 05:07:39,058][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.08339225500822067, 'train/grad_norm_pre_clip': 1.1716196537017822, 'timing/data': 0.010570503771305084, 'timing/model': 1.3141696825623512, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 0.29}) +[2026-06-08 05:11:41,629][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-06-08 05:11:41,632][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.043065138161182404, 'train/grad_norm_pre_clip': 0.9228273034095764, 'eval/loss': 0.10095208138227463, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 8.079999999999988, 'rl_games_eval/mid_train/mean_length': 82.8, 'rl_games_eval/mid_train/std_reward': 1.4218298069740927, 'rl_games_eval/mid_train/std_length': 14.218298069740976, 'rl_games_eval/mid_train/macro_mean_reward': 8.079999999999988, 'rl_games_eval/mid_train/macro_mean_length': 82.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 8.079999999999988, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 82.8, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 1.4218298069740927, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 14.218298069740976, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 8.079999999999988, 'checkpoint/best_score': 24.12000000000009, 'checkpoint/best_step': 600.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.008860763162374496, 'timing/model': 1.164587363600731, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.3}) +[2026-06-08 05:11:41,633][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1000_state +[2026-06-08 05:11:41,638][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 05:11:49,263][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1000_state/model.safetensors +[2026-06-08 05:12:06,718][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1000_state/optimizer.bin +[2026-06-08 05:12:06,718][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1000_state/sampler.bin +[2026-06-08 05:12:06,718][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-08 05:12:06,720][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-08 05:12:06,813][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-08 05:12:06,814][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 05:12:06,826][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 05:15:17,808][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.06643424928188324, 'train/grad_norm_pre_clip': 1.9465690851211548, 'timing/data': 0.010009264573454857, 'timing/model': 1.21424631588161, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 0.3}) +[2026-06-08 05:18:32,615][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.021961456164717674, 'train/grad_norm_pre_clip': 0.5190539956092834, 'timing/data': 0.010486900806427002, 'timing/model': 1.248643545433879, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.3}) +[2026-06-08 05:21:43,527][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.11902540177106857, 'train/grad_norm_pre_clip': 2.2872815132141113, 'timing/data': 0.009541437029838562, 'timing/model': 1.2074501030147076, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 0.31}) +[2026-06-08 05:24:52,218][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.11230383813381195, 'train/grad_norm_pre_clip': 0.641843318939209, 'timing/data': 0.010753313079476357, 'timing/model': 1.2581883277744055, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.31}) +[2026-06-08 05:28:06,630][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.07386970520019531, 'train/grad_norm_pre_clip': 2.0325117111206055, 'timing/data': 0.011218462139368057, 'timing/model': 1.461636632680893, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 0.31}) +[2026-06-08 05:31:16,635][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.11535225063562393, 'train/grad_norm_pre_clip': 0.9882320165634155, 'timing/data': 0.01395397074520588, 'timing/model': 1.34351110085845, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.31}) +[2026-06-08 05:34:24,092][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.0681602731347084, 'train/grad_norm_pre_clip': 1.3164373636245728, 'timing/data': 0.009659260511398315, 'timing/model': 1.1513372603803873, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 0.32}) +[2026-06-08 05:37:35,824][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.03250638768076897, 'train/grad_norm_pre_clip': 0.5309199094772339, 'timing/data': 0.010097391903400421, 'timing/model': 1.2179136835038662, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.32}) +[2026-06-08 05:40:41,721][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.12813404202461243, 'train/grad_norm_pre_clip': 3.2457115650177, 'timing/data': 0.009411433711647987, 'timing/model': 1.2594144362956285, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.32}) +[2026-06-08 05:45:39,647][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state +[2026-06-08 05:45:39,651][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 05:45:45,607][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/model.safetensors +[2026-06-08 05:46:00,921][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/optimizer.bin +[2026-06-08 05:46:00,921][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/sampler.bin +[2026-06-08 05:46:00,922][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/sampler_1.bin +[2026-06-08 05:46:00,923][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state/random_states_0.pkl +[2026-06-08 05:46:00,939][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 05:46:00,939][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1100 +[2026-06-08 05:46:00,941][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.04780591279268265, 'train/grad_norm_pre_clip': 0.7340912818908691, 'eval/loss': 0.07767148315906525, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 24.940000000000126, 'rl_games_eval/mid_train/mean_length': 219.0, 'rl_games_eval/mid_train/std_reward': 16.381282001113483, 'rl_games_eval/mid_train/std_length': 133.51404420509476, 'rl_games_eval/mid_train/macro_mean_reward': 24.940000000000126, 'rl_games_eval/mid_train/macro_mean_length': 219.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 24.940000000000126, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 219.0, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 16.381282001113483, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 133.51404420509476, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 24.940000000000126, 'checkpoint/best_score': 24.940000000000126, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.011445408686995506, 'timing/model': 1.2202512193471193, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.33}) +[2026-06-08 05:46:00,942][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1100_state +[2026-06-08 05:46:00,953][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 05:46:08,864][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1100_state/model.safetensors +[2026-06-08 05:46:23,152][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1100_state/optimizer.bin +[2026-06-08 05:46:23,153][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1100_state/sampler.bin +[2026-06-08 05:46:23,153][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1100_state/sampler_1.bin +[2026-06-08 05:46:23,154][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1100_state/random_states_0.pkl +[2026-06-08 05:46:23,251][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-06-08 05:46:23,251][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 05:46:23,265][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 05:49:31,523][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.16532175242900848, 'train/grad_norm_pre_clip': 2.9323229789733887, 'timing/data': 0.011479422450065613, 'timing/model': 1.2202869094908237, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 0.33}) +[2026-06-08 05:52:41,114][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.16425183415412903, 'train/grad_norm_pre_clip': 2.0046119689941406, 'timing/data': 0.012083834037184715, 'timing/model': 1.254353767260909, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.33}) +[2026-06-08 05:55:53,619][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.044812172651290894, 'train/grad_norm_pre_clip': 0.5543600916862488, 'timing/data': 0.011260369792580605, 'timing/model': 1.2566115446388721, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 0.33}) +[2026-06-08 05:59:14,286][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.01742750033736229, 'train/grad_norm_pre_clip': 0.8612880110740662, 'timing/data': 0.013038020581007004, 'timing/model': 1.5489813480526209, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.34}) +[2026-06-08 06:02:30,403][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.04407715052366257, 'train/grad_norm_pre_clip': 1.12485671043396, 'timing/data': 0.009085983037948608, 'timing/model': 1.1555767077952623, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 0.34}) +[2026-06-08 06:05:41,329][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.0421944260597229, 'train/grad_norm_pre_clip': 0.7835307717323303, 'timing/data': 0.010449636727571487, 'timing/model': 1.216680582612753, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.34}) +[2026-06-08 06:08:58,941][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.04635126143693924, 'train/grad_norm_pre_clip': 0.3387547731399536, 'timing/data': 0.011028308421373367, 'timing/model': 1.2670236714184284, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 0.35}) +[2026-06-08 06:12:08,595][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.10106848180294037, 'train/grad_norm_pre_clip': 1.9791182279586792, 'timing/data': 0.008970709517598152, 'timing/model': 1.1509433034807444, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.35}) +[2026-06-08 06:15:18,727][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.031249359250068665, 'train/grad_norm_pre_clip': 0.2567000091075897, 'timing/data': 0.010415764525532722, 'timing/model': 1.45666971988976, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 0.35}) +[2026-06-08 06:19:44,671][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1200 +[2026-06-08 06:19:44,672][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.059253789484500885, 'train/grad_norm_pre_clip': 0.5499390959739685, 'eval/loss': 0.09212758392095566, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 17.100000000000037, 'rl_games_eval/mid_train/mean_length': 156.8, 'rl_games_eval/mid_train/std_reward': 26.150028680672683, 'rl_games_eval/mid_train/std_length': 211.10035528155797, 'rl_games_eval/mid_train/macro_mean_reward': 17.100000000000037, 'rl_games_eval/mid_train/macro_mean_length': 156.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 17.100000000000037, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 156.8, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 26.150028680672683, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 211.10035528155797, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 17.100000000000037, 'checkpoint/best_score': 24.940000000000126, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.009885326027870178, 'timing/model': 1.1901550088077784, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.36}) +[2026-06-08 06:19:44,673][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1200_state +[2026-06-08 06:19:44,677][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 06:19:50,690][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1200_state/model.safetensors +[2026-06-08 06:20:04,743][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1200_state/optimizer.bin +[2026-06-08 06:20:04,744][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1200_state/sampler.bin +[2026-06-08 06:20:04,744][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1200_state/sampler_1.bin +[2026-06-08 06:20:04,745][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1200_state/random_states_0.pkl +[2026-06-08 06:20:04,846][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-06-08 06:20:04,846][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 06:20:04,859][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 06:23:11,092][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.168832466006279, 'train/grad_norm_pre_clip': 1.3696372509002686, 'timing/data': 0.009317239746451378, 'timing/model': 1.1600620280951262, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 0.36}) +[2026-06-08 06:26:20,640][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.11728746443986893, 'train/grad_norm_pre_clip': 0.764240562915802, 'timing/data': 0.010577192530035973, 'timing/model': 1.3747284729033709, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.36}) +[2026-06-08 06:29:30,796][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.17950919270515442, 'train/grad_norm_pre_clip': 1.405834436416626, 'timing/data': 0.010023444890975952, 'timing/model': 1.1514981593936682, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.36}) +[2026-06-08 06:32:39,981][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.05407696217298508, 'train/grad_norm_pre_clip': 0.45754605531692505, 'timing/data': 0.010360397398471832, 'timing/model': 1.2281729597598314, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 0.37}) +[2026-06-08 06:35:52,194][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.22939875721931458, 'train/grad_norm_pre_clip': 1.125981092453003, 'timing/data': 0.010205626487731934, 'timing/model': 1.243760360404849, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 0.37}) +[2026-06-08 06:39:06,927][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.15008537471294403, 'train/grad_norm_pre_clip': 0.9120648503303528, 'timing/data': 0.011103007942438126, 'timing/model': 1.2597589530050755, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 0.37}) +[2026-06-08 06:42:19,222][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.20393097400665283, 'train/grad_norm_pre_clip': 1.3071603775024414, 'timing/data': 0.01102844811975956, 'timing/model': 1.2077657878398895, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 0.38}) +[2026-06-08 06:45:36,841][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.2767370045185089, 'train/grad_norm_pre_clip': 1.8941881656646729, 'timing/data': 0.01189235970377922, 'timing/model': 1.2567238546907902, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 0.38}) +[2026-06-08 06:48:47,091][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.17353719472885132, 'train/grad_norm_pre_clip': 1.1169626712799072, 'timing/data': 0.008511532098054886, 'timing/model': 1.1508977245539427, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 0.38}) +[2026-06-08 06:53:06,527][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1300 +[2026-06-08 06:53:06,529][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.04493079334497452, 'train/grad_norm_pre_clip': 0.4126548171043396, 'eval/loss': 0.07766493409872055, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 12.460000000000013, 'rl_games_eval/mid_train/mean_length': 119.4, 'rl_games_eval/mid_train/std_reward': 8.317836257104416, 'rl_games_eval/mid_train/std_length': 67.93114160677708, 'rl_games_eval/mid_train/macro_mean_reward': 12.460000000000013, 'rl_games_eval/mid_train/macro_mean_length': 119.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 12.460000000000013, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 119.4, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 8.317836257104416, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 67.93114160677708, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 12.460000000000013, 'checkpoint/best_score': 24.940000000000126, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00916433148086071, 'timing/model': 1.1511958930641413, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.39}) +[2026-06-08 06:53:06,530][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1300_state +[2026-06-08 06:53:06,536][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 06:53:13,025][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1300_state/model.safetensors +[2026-06-08 06:53:27,610][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1300_state/optimizer.bin +[2026-06-08 06:53:27,611][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1300_state/sampler.bin +[2026-06-08 06:53:27,611][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1300_state/sampler_1.bin +[2026-06-08 06:53:27,612][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1300_state/random_states_0.pkl +[2026-06-08 06:53:27,708][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-06-08 06:53:27,708][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 06:53:27,721][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 06:56:39,627][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.02119375951588154, 'train/grad_norm_pre_clip': 0.24756748974323273, 'timing/data': 0.01074991375207901, 'timing/model': 1.5540009904652834, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 0.39}) +[2026-06-08 06:59:55,418][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.06231130287051201, 'train/grad_norm_pre_clip': 0.48300814628601074, 'timing/data': 0.010189613327383995, 'timing/model': 1.2575239203870296, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 0.39}) +[2026-06-08 07:03:08,295][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.12287288159132004, 'train/grad_norm_pre_clip': 0.6559014916419983, 'timing/data': 0.009162863716483116, 'timing/model': 1.1557149924337864, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 0.39}) +[2026-06-08 07:06:18,564][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.12048621475696564, 'train/grad_norm_pre_clip': 0.7677075862884521, 'timing/data': 0.009155292063951492, 'timing/model': 1.2165400050580502, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 0.4}) +[2026-06-08 07:09:37,447][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.09362348914146423, 'train/grad_norm_pre_clip': 1.7267751693725586, 'timing/data': 0.024612337350845337, 'timing/model': 1.2562491558492184, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 0.4}) +[2026-06-08 07:12:49,323][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.019463229924440384, 'train/grad_norm_pre_clip': 0.28511837124824524, 'timing/data': 0.014120455831289291, 'timing/model': 1.1945023573935032, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 0.4}) +[2026-06-08 07:16:03,024][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.009728561155498028, 'train/grad_norm_pre_clip': 0.1017567366361618, 'timing/data': 0.010185103863477707, 'timing/model': 1.2592190895229578, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 0.41}) +[2026-06-08 07:19:18,621][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.18121083080768585, 'train/grad_norm_pre_clip': 2.413879632949829, 'timing/data': 0.010439909994602203, 'timing/model': 1.257353786379099, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 0.41}) +[2026-06-08 07:22:33,969][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.12119678407907486, 'train/grad_norm_pre_clip': 0.36430981755256653, 'timing/data': 0.01035909540951252, 'timing/model': 1.4930525906383991, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 0.41}) +[2026-06-08 07:27:08,232][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1400 +[2026-06-08 07:27:08,235][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.0451953150331974, 'train/grad_norm_pre_clip': 0.7904226779937744, 'eval/loss': 0.06998667120933533, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 15.760000000000039, 'rl_games_eval/mid_train/mean_length': 147.0, 'rl_games_eval/mid_train/std_reward': 11.363379778921496, 'rl_games_eval/mid_train/std_length': 91.16139533815836, 'rl_games_eval/mid_train/macro_mean_reward': 15.760000000000039, 'rl_games_eval/mid_train/macro_mean_length': 147.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 15.760000000000039, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 147.0, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 11.363379778921496, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 91.16139533815836, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 15.760000000000039, 'checkpoint/best_score': 24.940000000000126, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.009765492752194405, 'timing/model': 1.2467770092189312, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 0.41}) +[2026-06-08 07:27:08,236][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1400_state +[2026-06-08 07:27:08,243][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 07:27:17,414][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1400_state/model.safetensors +[2026-06-08 07:27:34,693][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1400_state/optimizer.bin +[2026-06-08 07:27:34,694][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1400_state/sampler.bin +[2026-06-08 07:27:34,694][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1400_state/sampler_1.bin +[2026-06-08 07:27:34,696][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1400_state/random_states_0.pkl +[2026-06-08 07:27:34,813][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-06-08 07:27:34,814][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 07:27:34,826][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 07:30:44,627][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.10186994075775146, 'train/grad_norm_pre_clip': 0.9667148590087891, 'timing/data': 0.0111855398863554, 'timing/model': 1.2597876861691475, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 0.42}) +[2026-06-08 07:33:52,309][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.07917829602956772, 'train/grad_norm_pre_clip': 0.8878235220909119, 'timing/data': 0.010559068992733955, 'timing/model': 1.2501077521592379, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 0.42}) +[2026-06-08 07:37:05,482][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.032034844160079956, 'train/grad_norm_pre_clip': 0.6499000787734985, 'timing/data': 0.010509686544537544, 'timing/model': 1.2277769353240728, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 0.42}) +[2026-06-08 07:40:18,327][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.12586240470409393, 'train/grad_norm_pre_clip': 1.8830002546310425, 'timing/data': 0.009460458531975746, 'timing/model': 1.2515732999891043, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 0.43}) +[2026-06-08 07:43:25,547][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.004922307096421719, 'train/grad_norm_pre_clip': 0.07307770848274231, 'timing/data': 0.010966794565320015, 'timing/model': 1.389034677296877, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 0.43}) +[2026-06-08 07:46:36,315][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.22741131484508514, 'train/grad_norm_pre_clip': 3.1039297580718994, 'timing/data': 0.007972463965415955, 'timing/model': 1.1617786325514317, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 0.43}) +[2026-06-08 07:49:42,714][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.016162779182195663, 'train/grad_norm_pre_clip': 0.4237514138221741, 'timing/data': 0.010740453377366066, 'timing/model': 1.2564259991049767, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 0.44}) +[2026-06-08 07:52:56,541][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.02362322062253952, 'train/grad_norm_pre_clip': 0.5805016756057739, 'timing/data': 0.01626104861497879, 'timing/model': 1.258153835311532, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 0.44}) +[2026-06-08 07:56:13,529][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.11151214689016342, 'train/grad_norm_pre_clip': 1.256737232208252, 'timing/data': 0.010908376425504684, 'timing/model': 1.2586825620383024, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 0.44}) +[2026-06-08 08:01:11,819][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-08 08:01:11,821][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.04861366003751755, 'train/grad_norm_pre_clip': 0.6320158243179321, 'eval/loss': 0.06767338514328003, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 21.640000000000093, 'rl_games_eval/mid_train/mean_length': 195.0, 'rl_games_eval/mid_train/std_reward': 16.326003797623127, 'rl_games_eval/mid_train/std_length': 131.7740490384962, 'rl_games_eval/mid_train/macro_mean_reward': 21.640000000000093, 'rl_games_eval/mid_train/macro_mean_length': 195.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_1/mean_reward': 21.640000000000093, 'rl_games_eval/mid_train/flappy__latency_1/mean_length': 195.0, 'rl_games_eval/mid_train/flappy__latency_1/std_reward': 16.326003797623127, 'rl_games_eval/mid_train/flappy__latency_1/std_length': 131.7740490384962, 'rl_games_eval/mid_train/flappy__latency_1/num_episodes': 5.0, 'checkpoint/current_eval_score': 21.640000000000093, 'checkpoint/best_score': 24.940000000000126, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.010944290086627007, 'timing/model': 1.25616211630404, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 0.44}) +[2026-06-08 08:01:11,821][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1500_state +[2026-06-08 08:01:11,825][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 08:01:18,211][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1500_state/model.safetensors +[2026-06-08 08:01:32,330][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1500_state/optimizer.bin +[2026-06-08 08:01:32,331][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1500_state/sampler.bin +[2026-06-08 08:01:32,331][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-08 08:01:32,332][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-08 08:01:32,433][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-08 08:01:32,433][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 08:01:32,445][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 08:01:32,445][starVLA.training.train_starvla][INFO] - Training complete. Best checkpoint step=1100 score=24.940000000000126 path=/workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state +[2026-06-08 08:01:32,446][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_1/checkpoints/best_state +[2026-06-08 08:01:33,957][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-08 08:01:43,125][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-08 08:01:43,126][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-08 08:01:43,126][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-08 08:01:43,128][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-08 08:01:43,606][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-08 08:13:05,687][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-08 08:13:07,617][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..670bc6974d1e212c429f2c026f4feb746ff9315f --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/summary.jsonl @@ -0,0 +1,15 @@ +{"steps": 100} +{"steps": 200} +{"steps": 300} +{"steps": 400} +{"steps": 500} +{"steps": 600} +{"steps": 700} +{"steps": 800} +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/checkpoints/steps_5000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/checkpoints/steps_5000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8f98df5e12f1af4045bd5c98aa5584609ad8ed42 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/checkpoints/steps_5000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0fc5f0a0cd58f4f9726d53639e193c48efd5afd512fb0d0f644d239eb4079668 +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/checkpoints/steps_5000_pytorch_model.pt b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/checkpoints/steps_5000_pytorch_model.pt new file mode 100644 index 0000000000000000000000000000000000000000..0181dd2b90452998e8370c5d59bbf7fa70ec145a --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/checkpoints/steps_5000_pytorch_model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7568ff3e7f09a67fd0681287750fbe5c34c99617b08bb440997c97b247c5d651 +size 9138476048 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..fb08c106af53413443bb2af8a88c86717e83724e --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/config.full.yaml @@ -0,0 +1,178 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_fix_latency_2 + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 16 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + task_description: '' + enabled: true + latency: + prompt_map_path: /workspace/starVLA/data/flappy_fix_latency_2/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 2 + num_episodes: 1 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + - 5 + - 6 + - 7 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_flappy_fix_latency_2 +output_dir: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2 +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..b09bea269a904297487868e54c987166703b9784 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/config.yaml @@ -0,0 +1,79 @@ +checkpoint: + local: + keep_last_n: 1 + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + include_state: true + obs_image_size: null + per_device_batch_size: 16 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2 +rl_games: + env_eval: + enabled: true + frameskip: 1 + image_size: 224 + latency: + prompt_map_path: /workspace/starVLA/data/flappy_fix_latency_2/flappy_train__bridge/latency_prompt_map.json + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 2 + max_steps_per_episode: 3600 + num_episodes: 1 + task_description: '' + model_alias: openvla + task: flappy +run_id: openvla_flappy_fix_latency_2 +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + eval_interval: 100 + eval_num_batches: 20 + freeze_modules: '' + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 10 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + pretrained_checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + reload_modules: null + save_format: pt + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..dfa627dc1c6b9a465a2fa18fb9d3e9d39eeb4758 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.7964444160461426, + 0.20355555415153503, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4029320180416107, + 0.4029320180416107, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 324000, + "num_trajectories": 90 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..bcd7bd38d389ab2c61511af4d519549faef11c5d --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.7942777872085571, + 0.20572222769260406, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.40424680709838867, + 0.40424680709838867, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 36000, + "num_trajectories": 10 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/hydra/.hydra/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..7b63d7e2162900928b8dd526f42b0fcc18c4116f --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/hydra/.hydra/config.yaml @@ -0,0 +1,176 @@ +framework: + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_fix_latency_2 + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 16 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + task_description: '' + enabled: true + latency: + prompt_map_path: /workspace/starVLA/data/flappy_fix_latency_2/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 2 + num_episodes: 1 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + - 5 + - 6 + - 7 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_flappy_fix_latency_2 +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/hydra/.hydra/hydra.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..caed57561bad3df8d23a4f232c4c3f092e2953dc --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/hydra/.hydra/hydra.yaml @@ -0,0 +1,238 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=single + - run_id=openvla_flappy_fix_latency_2 + - run_root_dir=/workspace/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=5000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=20 + - trainer.logging_frequency=10 + - trainer.gradient_accumulation_steps=16 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9, 0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=2 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.per_device_batch_size=16 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_2 + - datasets.vla_data.data_mix=flappy_train__bridge + - datasets.vla_data.eval_data_mix=flappy_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=flappy + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=1 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_2/flappy_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=true + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=1 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[2] + - rl_games.env_eval.post_train.enabled=true + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4,5,6,7] + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=flappy_train__bridge,datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_2,datasets.vla_data.eval_data_mix=flappy_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=16,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.video_backend=torchvision_av,env=flappy,framework.action_model.action_dim=7,framework.action_model.action_env_dim=2,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.enabled=true,rl_games.env_eval.frameskip=1,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_2/flappy_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.mid_train.enabled=true,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[2],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=1,rl_games.env_eval.post_train.enabled=true,rl_games.env_eval.post_train.latencies=[0,1,2,3,4,5,6,7],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=flappy,run_id=openvla_flappy_fix_latency_2,run_root_dir=/workspace/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=20,trainer.gradient_accumulation_steps=16,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=10,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=5000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9, + 0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.2 + version_base: '1.1' + cwd: /workspace/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/hydra + choices: + checkpoint: default + mode: single + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/hydra/.hydra/overrides.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..012d8cf4043280457cf1d67eda3bb16d8fad17f0 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/hydra/.hydra/overrides.yaml @@ -0,0 +1,78 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=single +- run_id=openvla_flappy_fix_latency_2 +- run_root_dir=/workspace/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=5000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=20 +- trainer.logging_frequency=10 +- trainer.gradient_accumulation_steps=16 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9, 0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=2 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.per_device_batch_size=16 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_2 +- datasets.vla_data.data_mix=flappy_train__bridge +- datasets.vla_data.eval_data_mix=flappy_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=flappy +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=1 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_fix_latency_2/flappy_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=true +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=1 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[2] +- rl_games.env_eval.post_train.enabled=true +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4,5,6,7] diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/hydra/train_starvla_hydra.log b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..c27da15f1d2cecc4ff0d1db72819cdceefd692ca --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/hydra/train_starvla_hydra.log @@ -0,0 +1,1143 @@ +[2026-05-30 07:39:48,042][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-05-30 07:39:48,087][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-05-30 07:39:49,266][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-05-30 07:39:50,015][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-05-30 07:39:50,016][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-05-30 07:39:50,016][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-05-30 07:39:50,016][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-05-30 07:39:50,016][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-05-30 07:39:50,016][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-05-30 07:39:50,016][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-05-30 07:39:50,017][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-05-30 07:39:50,017][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-05-30 07:39:50,017][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-05-30 07:39:50,017][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-05-30 07:39:50,018][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-05-30 07:39:50,018][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-05-30 07:39:50,018][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-05-30 07:39:50,018][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-05-30 07:39:50,018][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-05-30 07:39:50,019][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-05-30 07:39:50,019][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-05-30 07:39:50,019][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-05-30 07:39:50,021][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-05-30 07:39:50,021][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-05-30 07:39:50,021][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-05-30 07:39:50,021][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-05-30 07:39:50,021][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-05-30 07:39:50,022][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-05-30 07:39:50,022][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-05-30 07:39:50,022][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-05-30 07:39:50,022][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-05-30 07:39:50,022][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-05-30 07:39:50,023][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-05-30 07:39:50,023][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-05-30 07:39:50,077][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-05-30 07:39:50,103][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/config.full.yaml +[2026-05-30 07:39:50,109][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/config.yaml +[2026-05-30 07:40:00,755][starVLA.training.train_starvla][INFO] - Loaded pretrained checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt, steps: 0 +[2026-05-30 07:40:03,571][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-05-30 07:40:03,572][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-05-30 07:40:03,572][starVLA.training.train_starvla][INFO] - Per device batch size = 16 +[2026-05-30 07:40:03,572][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-05-30 07:40:03,573][starVLA.training.train_starvla][INFO] - Total batch size = 256 +[2026-05-30 07:43:33,605][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-05-30 07:43:33,660][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-05-30 07:43:34,860][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-05-30 07:43:35,566][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-05-30 07:43:35,567][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-05-30 07:43:35,567][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-05-30 07:43:35,567][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-05-30 07:43:35,568][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-05-30 07:43:35,568][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-05-30 07:43:35,568][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-05-30 07:43:35,568][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-05-30 07:43:35,568][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-05-30 07:43:35,568][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-05-30 07:43:35,569][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-05-30 07:43:35,569][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-05-30 07:43:35,569][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-05-30 07:43:35,569][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-05-30 07:43:35,570][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-05-30 07:43:35,570][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-05-30 07:43:35,570][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-05-30 07:43:35,570][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-05-30 07:43:35,570][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-05-30 07:43:35,572][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-05-30 07:43:35,572][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-05-30 07:43:35,572][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-05-30 07:43:35,573][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-05-30 07:43:35,573][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-05-30 07:43:35,573][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-05-30 07:43:35,573][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-05-30 07:43:35,574][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-05-30 07:43:35,574][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-05-30 07:43:35,574][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-05-30 07:43:35,574][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-05-30 07:43:35,574][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-05-30 07:43:35,629][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-05-30 07:43:35,656][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/config.full.yaml +[2026-05-30 07:43:35,661][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/config.yaml +[2026-05-30 07:43:45,845][starVLA.training.train_starvla][INFO] - Loaded pretrained checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt, steps: 0 +[2026-05-30 07:43:48,888][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-05-30 07:43:48,889][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-05-30 07:43:48,889][starVLA.training.train_starvla][INFO] - Per device batch size = 16 +[2026-05-30 07:43:48,890][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-05-30 07:43:48,890][starVLA.training.train_starvla][INFO] - Total batch size = 256 +[2026-05-30 07:46:34,204][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.5053756833076477, 'train/grad_norm_pre_clip': 0.6055380702018738, 'timing/data': 0.004356523975729942, 'timing/model': 1.0772353913635015, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-05-30 07:49:15,311][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.5026052594184875, 'train/grad_norm_pre_clip': 0.5839500427246094, 'timing/data': 0.0049346573650836945, 'timing/model': 1.0832557436078787, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.0}) +[2026-05-30 07:51:56,408][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.6046762466430664, 'train/grad_norm_pre_clip': 1.1057862043380737, 'timing/data': 0.004447411745786667, 'timing/model': 1.081263206899166, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.0}) +[2026-05-30 07:54:37,409][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.21684928238391876, 'train/grad_norm_pre_clip': 0.1911643147468567, 'timing/data': 0.003843538463115692, 'timing/model': 1.0794895589351654, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.0}) +[2026-05-30 07:57:18,807][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.41550955176353455, 'train/grad_norm_pre_clip': 1.4348220825195312, 'timing/data': 0.004389243200421333, 'timing/model': 1.085539584979415, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.0}) +[2026-05-30 07:59:59,807][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.3765616714954376, 'train/grad_norm_pre_clip': 1.1997495889663696, 'timing/data': 0.0036466680467128754, 'timing/model': 1.0818564519286156, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.0}) +[2026-05-30 08:02:41,004][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.6921654343605042, 'train/grad_norm_pre_clip': 17.928754806518555, 'timing/data': 0.004060667008161545, 'timing/model': 1.0784223154187202, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.0}) +[2026-05-30 08:05:22,106][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.16288796067237854, 'train/grad_norm_pre_clip': 1.6976252794265747, 'timing/data': 0.004515869542956352, 'timing/model': 1.0780975427478552, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.0}) +[2026-05-30 08:08:03,707][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.31391507387161255, 'train/grad_norm_pre_clip': 1.8110013008117676, 'timing/data': 0.004783377051353455, 'timing/model': 1.0824756789952517, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.0}) +[2026-05-30 08:11:02,742][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp66iqalqw/test.c -o /tmp/tmp66iqalqw/test.o +[2026-05-30 08:11:02,763][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp66iqalqw/test.o -laio -o /tmp/tmp66iqalqw/a.out +[2026-05-30 08:11:03,151][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpn5xkkdcg/test.c -o /tmp/tmpn5xkkdcg/test.o +[2026-05-30 08:11:03,180][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpn5xkkdcg/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpn5xkkdcg/a.out +[2026-05-30 08:11:06,459][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 100 +[2026-05-30 08:11:06,461][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.26448285579681396, 'train/grad_norm_pre_clip': 1.334352731704712, 'eval/loss': 0.3647863566875458, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 2.0000000000000013, 'rl_games_eval/mid_train/mean_length': 31.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 2.0000000000000013, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 31.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.005036540329456329, 'timing/model': 1.0774232372641563, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.0}) +[2026-05-30 08:11:06,461][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_100_state +[2026-05-30 08:11:06,466][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 08:11:14,349][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_100_state/model.safetensors +[2026-05-30 08:11:22,099][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_100_state/optimizer.bin +[2026-05-30 08:11:22,100][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_100_state/sampler.bin +[2026-05-30 08:11:22,100][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_100_state/sampler_1.bin +[2026-05-30 08:11:22,101][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_100_state/random_states_0.pkl +[2026-05-30 08:11:26,022][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-05-30 08:11:26,023][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 08:11:26,036][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 08:14:39,856][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.22106358408927917, 'train/grad_norm_pre_clip': 1.092955231666565, 'timing/data': 0.0036895200610160828, 'timing/model': 1.2404303345829248, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.01}) +[2026-05-30 08:17:52,959][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.13279487192630768, 'train/grad_norm_pre_clip': 0.680892825126648, 'timing/data': 0.0035133101046085358, 'timing/model': 1.2536892369389534, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.01}) +[2026-05-30 08:21:07,501][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.19793328642845154, 'train/grad_norm_pre_clip': 1.8416144847869873, 'timing/data': 0.003738241270184517, 'timing/model': 1.2860720548778772, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.01}) +[2026-05-30 08:24:21,200][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.3233727216720581, 'train/grad_norm_pre_clip': 1.0965017080307007, 'timing/data': 0.003741160035133362, 'timing/model': 1.2908993400633335, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.01}) +[2026-05-30 08:27:35,477][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 1.054031491279602, 'train/grad_norm_pre_clip': 5.170764446258545, 'timing/data': 0.0032954737544059753, 'timing/model': 1.2750365044921637, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.01}) +[2026-05-30 08:30:49,200][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.23610016703605652, 'train/grad_norm_pre_clip': 2.242828369140625, 'timing/data': 0.0033411048352718353, 'timing/model': 1.285868339240551, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.01}) +[2026-05-30 08:34:03,213][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.39366045594215393, 'train/grad_norm_pre_clip': 1.4199544191360474, 'timing/data': 0.0034327320754528046, 'timing/model': 1.2866709642112255, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.01}) +[2026-05-30 08:37:15,695][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.16568678617477417, 'train/grad_norm_pre_clip': 0.22234633564949036, 'timing/data': 0.003393452614545822, 'timing/model': 1.2908854074776173, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.01}) +[2026-05-30 08:40:30,195][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.1420964151620865, 'train/grad_norm_pre_clip': 0.5436595678329468, 'timing/data': 0.0032974723726511, 'timing/model': 1.2925785556435585, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.01}) +[2026-05-30 08:44:05,322][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 200 +[2026-05-30 08:44:05,324][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.47016799449920654, 'train/grad_norm_pre_clip': 2.4733245372772217, 'eval/loss': 0.3653346002101898, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 2.0000000000000013, 'rl_games_eval/mid_train/mean_length': 31.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 2.0000000000000013, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 31.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0032738614827394485, 'timing/model': 1.2880320604890585, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.01}) +[2026-05-30 08:44:05,324][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_200_state +[2026-05-30 08:44:05,329][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 08:44:12,724][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_200_state/model.safetensors +[2026-05-30 08:44:20,409][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_200_state/optimizer.bin +[2026-05-30 08:44:20,410][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_200_state/sampler.bin +[2026-05-30 08:44:20,410][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_200_state/sampler_1.bin +[2026-05-30 08:44:20,412][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_200_state/random_states_0.pkl +[2026-05-30 08:44:24,447][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-05-30 08:44:24,448][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 08:44:24,460][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 08:47:39,397][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.2438393235206604, 'train/grad_norm_pre_clip': 0.9818580150604248, 'timing/data': 0.0034802518784999847, 'timing/model': 1.2821338716894388, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.01}) +[2026-05-30 08:50:54,300][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.08561518788337708, 'train/grad_norm_pre_clip': 0.7026839852333069, 'timing/data': 0.003514602780342102, 'timing/model': 1.298622326925397, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.01}) +[2026-05-30 08:54:08,889][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.27475741505622864, 'train/grad_norm_pre_clip': 0.8829167485237122, 'timing/data': 0.003208143636584282, 'timing/model': 1.27813776768744, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.01}) +[2026-05-30 08:57:22,706][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.511935830116272, 'train/grad_norm_pre_clip': 9.980293273925781, 'timing/data': 0.003532940521836281, 'timing/model': 1.2367219049483538, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.01}) +[2026-05-30 09:00:36,194][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.5318300127983093, 'train/grad_norm_pre_clip': 9.587384223937988, 'timing/data': 0.0032585524022579193, 'timing/model': 1.2810976393520832, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.01}) +[2026-05-30 09:03:48,288][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.24968230724334717, 'train/grad_norm_pre_clip': 0.9659453630447388, 'timing/data': 0.0030274931341409683, 'timing/model': 1.284662052989006, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.01}) +[2026-05-30 09:07:02,596][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.21199390292167664, 'train/grad_norm_pre_clip': 1.0306068658828735, 'timing/data': 0.003202013671398163, 'timing/model': 1.2940946854650974, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.01}) +[2026-05-30 09:10:16,695][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.2468079924583435, 'train/grad_norm_pre_clip': 2.0536181926727295, 'timing/data': 0.002951866015791893, 'timing/model': 1.2794785015285015, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.01}) +[2026-05-30 09:13:30,989][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.3662709891796112, 'train/grad_norm_pre_clip': 1.0004948377609253, 'timing/data': 0.0031747352331876755, 'timing/model': 1.2840685192495584, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.01}) +[2026-05-30 09:17:10,901][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 300 +[2026-05-30 09:17:10,903][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.2473703771829605, 'train/grad_norm_pre_clip': 3.1280181407928467, 'eval/loss': 0.20237579941749573, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 12.899999999999974, 'rl_games_eval/mid_train/mean_length': 122.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 12.899999999999974, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 122.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.003220072016119957, 'timing/model': 1.2823244668543339, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.01}) +[2026-05-30 09:17:10,903][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_300_state +[2026-05-30 09:17:10,908][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 09:17:16,896][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_300_state/model.safetensors +[2026-05-30 09:17:24,289][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_300_state/optimizer.bin +[2026-05-30 09:17:24,290][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_300_state/sampler.bin +[2026-05-30 09:17:24,290][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_300_state/sampler_1.bin +[2026-05-30 09:17:24,292][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_300_state/random_states_0.pkl +[2026-05-30 09:17:28,121][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-05-30 09:17:28,122][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 09:17:28,135][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 09:20:42,695][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.18164440989494324, 'train/grad_norm_pre_clip': 0.4519551992416382, 'timing/data': 0.0034178104251623154, 'timing/model': 1.2917626611888409, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.02}) +[2026-05-30 09:23:56,293][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.14910516142845154, 'train/grad_norm_pre_clip': 0.18958814442157745, 'timing/data': 0.003118753433227539, 'timing/model': 1.2872885782271624, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.02}) +[2026-05-30 09:27:09,092][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.09330451488494873, 'train/grad_norm_pre_clip': 0.4924318790435791, 'timing/data': 0.00322556309401989, 'timing/model': 1.2670747581869364, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.02}) +[2026-05-30 09:30:22,392][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.2127140462398529, 'train/grad_norm_pre_clip': 6.994716644287109, 'timing/data': 0.003193335607647896, 'timing/model': 1.2855796441435814, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.02}) +[2026-05-30 09:33:36,292][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.017207341268658638, 'train/grad_norm_pre_clip': 0.23025792837142944, 'timing/data': 0.0034712832421064377, 'timing/model': 1.2826704010367393, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.02}) +[2026-05-30 09:36:49,946][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.26065874099731445, 'train/grad_norm_pre_clip': 0.48823675513267517, 'timing/data': 0.0034352727234363556, 'timing/model': 1.2490465082228184, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.02}) +[2026-05-30 09:40:04,494][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.38599303364753723, 'train/grad_norm_pre_clip': 3.771876811981201, 'timing/data': 0.008047090843319893, 'timing/model': 1.2766619194298983, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.02}) +[2026-05-30 09:43:17,185][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.37712979316711426, 'train/grad_norm_pre_clip': 1.7080612182617188, 'timing/data': 0.003128737211227417, 'timing/model': 1.2810756210237741, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.02}) +[2026-05-30 09:46:31,997][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.3435606062412262, 'train/grad_norm_pre_clip': 1.8883177042007446, 'timing/data': 0.003750177100300789, 'timing/model': 1.2521435879170895, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.02}) +[2026-05-30 09:50:07,854][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 400 +[2026-05-30 09:50:07,856][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.031929660588502884, 'train/grad_norm_pre_clip': 2.522465467453003, 'eval/loss': 0.1883954554796219, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0033588744699954987, 'timing/model': 1.2826127540320158, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.02}) +[2026-05-30 09:50:07,856][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_400_state +[2026-05-30 09:50:07,861][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 09:50:15,353][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_400_state/model.safetensors +[2026-05-30 09:50:23,096][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_400_state/optimizer.bin +[2026-05-30 09:50:23,097][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_400_state/sampler.bin +[2026-05-30 09:50:23,097][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_400_state/sampler_1.bin +[2026-05-30 09:50:23,099][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_400_state/random_states_0.pkl +[2026-05-30 09:50:27,056][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-05-30 09:50:27,056][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 09:50:27,069][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 09:53:42,499][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.2678185701370239, 'train/grad_norm_pre_clip': 4.6946845054626465, 'timing/data': 0.003621090203523636, 'timing/model': 1.261297820135951, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.02}) +[2026-05-30 09:56:58,103][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.15446282923221588, 'train/grad_norm_pre_clip': 2.1933605670928955, 'timing/data': 0.0035974811762571335, 'timing/model': 1.2930925823748112, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.02}) +[2026-05-30 10:00:12,053][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.47789204120635986, 'train/grad_norm_pre_clip': 3.195023536682129, 'timing/data': 0.0033768434077501297, 'timing/model': 1.2440837491303682, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.02}) +[2026-05-30 10:03:23,299][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.3878554105758667, 'train/grad_norm_pre_clip': 7.073628902435303, 'timing/data': 0.008052429184317589, 'timing/model': 1.2677280921489, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.02}) +[2026-05-30 10:06:36,398][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.18279126286506653, 'train/grad_norm_pre_clip': 2.0521624088287354, 'timing/data': 0.007776584476232529, 'timing/model': 1.2401988301426172, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.02}) +[2026-05-30 10:09:45,390][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.18312710523605347, 'train/grad_norm_pre_clip': 3.2786290645599365, 'timing/data': 0.007694628089666367, 'timing/model': 1.2372040804475546, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.02}) +[2026-05-30 10:12:55,600][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.03076259046792984, 'train/grad_norm_pre_clip': 0.3144594132900238, 'timing/data': 0.0004384741187095642, 'timing/model': 1.2393609825521708, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.02}) +[2026-05-30 10:16:06,935][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.04559364914894104, 'train/grad_norm_pre_clip': 0.3487599790096283, 'timing/data': 0.0033452212810516357, 'timing/model': 1.2363278828561306, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.02}) +[2026-05-30 10:19:17,094][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.09769221395254135, 'train/grad_norm_pre_clip': 0.6123557686805725, 'timing/data': 0.0032601114362478256, 'timing/model': 1.3228445630520582, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 0.02}) +[2026-05-30 10:22:46,828][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-05-30 10:22:46,830][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.21157708764076233, 'train/grad_norm_pre_clip': 1.828235149383545, 'eval/loss': 0.13535967469215393, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0008915793150663376, 'timing/model': 1.2796932887285948, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 0.02}) +[2026-05-30 10:22:46,830][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_500_state +[2026-05-30 10:22:46,835][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 10:22:52,410][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_500_state/model.safetensors +[2026-05-30 10:22:59,640][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_500_state/optimizer.bin +[2026-05-30 10:22:59,641][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_500_state/sampler.bin +[2026-05-30 10:22:59,641][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_500_state/sampler_1.bin +[2026-05-30 10:22:59,643][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_500_state/random_states_0.pkl +[2026-05-30 10:23:03,497][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-05-30 10:23:03,498][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 10:23:03,512][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 10:26:14,409][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.3541722893714905, 'train/grad_norm_pre_clip': 1.8356221914291382, 'timing/data': 0.005491558462381363, 'timing/model': 1.2875979095697403, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 0.03}) +[2026-05-30 10:29:23,920][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.11052895337343216, 'train/grad_norm_pre_clip': 1.9003386497497559, 'timing/data': 0.0033028218895196915, 'timing/model': 1.2408026736229658, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.03}) +[2026-05-30 10:32:34,634][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.2718809247016907, 'train/grad_norm_pre_clip': 1.0531843900680542, 'timing/data': 0.00033271685242652893, 'timing/model': 1.2390025723725557, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 0.03}) +[2026-05-30 10:35:45,017][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.05032717436552048, 'train/grad_norm_pre_clip': 0.640089213848114, 'timing/data': 0.005867835134267807, 'timing/model': 1.3018904943019152, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 0.03}) +[2026-05-30 10:38:54,596][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.17244403064250946, 'train/grad_norm_pre_clip': 0.80088210105896, 'timing/data': 0.0004968438297510147, 'timing/model': 1.246657021343708, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 0.03}) +[2026-05-30 10:42:05,576][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.26196855306625366, 'train/grad_norm_pre_clip': 7.509401798248291, 'timing/data': 0.005505776032805443, 'timing/model': 1.2443250361829996, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 0.03}) +[2026-05-30 10:45:15,228][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.6349667310714722, 'train/grad_norm_pre_clip': 3.8846309185028076, 'timing/data': 0.003010595217347145, 'timing/model': 1.2402673065662384, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 0.03}) +[2026-05-30 10:48:26,157][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.23966558277606964, 'train/grad_norm_pre_clip': 1.7744369506835938, 'timing/data': 0.0036112479865550995, 'timing/model': 1.2479702681303024, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 0.03}) +[2026-05-30 10:51:36,793][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.017114615067839622, 'train/grad_norm_pre_clip': 0.6620668768882751, 'timing/data': 0.003100404515862465, 'timing/model': 1.246608167886734, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.03}) +[2026-05-30 10:55:04,743][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 600 +[2026-05-30 10:55:04,745][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.23452918231487274, 'train/grad_norm_pre_clip': 0.6468251943588257, 'eval/loss': 0.12958025932312012, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.006902711465954781, 'timing/model': 1.2917829845100641, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 0.03}) +[2026-05-30 10:55:04,745][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_600_state +[2026-05-30 10:55:04,749][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 10:55:10,609][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_600_state/model.safetensors +[2026-05-30 10:55:17,814][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_600_state/optimizer.bin +[2026-05-30 10:55:17,815][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_600_state/sampler.bin +[2026-05-30 10:55:17,815][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_600_state/sampler_1.bin +[2026-05-30 10:55:17,816][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_600_state/random_states_0.pkl +[2026-05-30 10:55:21,579][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600 +[2026-05-30 10:55:21,579][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 10:55:21,591][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 10:58:30,663][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.2164922058582306, 'train/grad_norm_pre_clip': 6.3812031745910645, 'timing/data': 0.00531286746263504, 'timing/model': 1.246741184964776, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 0.03}) +[2026-05-30 11:01:41,300][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.14595811069011688, 'train/grad_norm_pre_clip': 0.531155526638031, 'timing/data': 0.00783659890294075, 'timing/model': 1.329746324568987, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 0.03}) +[2026-05-30 11:04:51,541][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.07591014355421066, 'train/grad_norm_pre_clip': 0.6416032314300537, 'timing/data': 0.003252793103456497, 'timing/model': 1.2392729111015797, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 0.03}) +[2026-05-30 11:08:01,105][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 4.03192279918585e-05, 'train/grad_norm_pre_clip': 0.001129794749431312, 'timing/data': 0.005685713142156601, 'timing/model': 1.2383154835551977, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 0.03}) +[2026-05-30 11:11:10,406][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.02563714236021042, 'train/grad_norm_pre_clip': 0.5351489186286926, 'timing/data': 0.007119767367839813, 'timing/model': 1.270317392423749, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 0.03}) +[2026-05-30 11:14:22,119][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.09001009911298752, 'train/grad_norm_pre_clip': 1.2767173051834106, 'timing/data': 0.005625622346997261, 'timing/model': 1.2379508391022682, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.03}) +[2026-05-30 11:17:32,158][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.32316458225250244, 'train/grad_norm_pre_clip': 4.373110771179199, 'timing/data': 0.0003340449184179306, 'timing/model': 1.2423658184707165, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 0.03}) +[2026-05-30 11:20:43,542][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.16055674850940704, 'train/grad_norm_pre_clip': 1.7849481105804443, 'timing/data': 0.0003532674163579941, 'timing/model': 1.2643432170152664, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 0.03}) +[2026-05-30 11:23:53,145][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.20489159226417542, 'train/grad_norm_pre_clip': 0.954023003578186, 'timing/data': 0.0030116159468889236, 'timing/model': 1.2401073798537254, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 0.03}) +[2026-05-30 11:27:21,690][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 700 +[2026-05-30 11:27:21,691][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.3464439809322357, 'train/grad_norm_pre_clip': 5.915938377380371, 'eval/loss': 0.1419977992773056, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 4.4999999999999964, 'rl_games_eval/mid_train/mean_length': 56.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 4.4999999999999964, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 56.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.007894022390246391, 'timing/model': 1.2368849273771048, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 0.03}) +[2026-05-30 11:27:21,692][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_700_state +[2026-05-30 11:27:21,696][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 11:27:27,506][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_700_state/model.safetensors +[2026-05-30 11:27:34,837][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_700_state/optimizer.bin +[2026-05-30 11:27:34,837][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_700_state/sampler.bin +[2026-05-30 11:27:34,837][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_700_state/sampler_1.bin +[2026-05-30 11:27:34,838][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_700_state/random_states_0.pkl +[2026-05-30 11:27:38,671][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700 +[2026-05-30 11:27:38,671][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 11:27:38,683][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 11:30:49,072][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.4843694567680359, 'train/grad_norm_pre_clip': 5.569090843200684, 'timing/data': 0.00935024581849575, 'timing/model': 1.239041492342949, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 0.04}) +[2026-05-30 11:33:58,358][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.031028209254145622, 'train/grad_norm_pre_clip': 0.7260907888412476, 'timing/data': 0.005707556381821632, 'timing/model': 1.2403389159590006, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 0.04}) +[2026-05-30 11:37:08,648][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.2851170003414154, 'train/grad_norm_pre_clip': 2.2108969688415527, 'timing/data': 0.0031484849750995636, 'timing/model': 1.237068872898817, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.04}) +[2026-05-30 11:40:18,320][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.008856951259076595, 'train/grad_norm_pre_clip': 0.1318371295928955, 'timing/data': 0.0033208318054676056, 'timing/model': 1.2370182313024998, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 0.04}) +[2026-05-30 11:43:28,504][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.006620722357183695, 'train/grad_norm_pre_clip': 0.078671894967556, 'timing/data': 0.0076566822826862335, 'timing/model': 1.2400019243359566, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 0.04}) +[2026-05-30 11:46:38,501][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.15991370379924774, 'train/grad_norm_pre_clip': 1.1078128814697266, 'timing/data': 0.013108281418681145, 'timing/model': 1.2395588885992765, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 0.04}) +[2026-05-30 11:49:48,677][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.1902916580438614, 'train/grad_norm_pre_clip': 2.9806759357452393, 'timing/data': 0.0005302131175994873, 'timing/model': 1.233872326090932, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 0.04}) +[2026-05-30 11:52:58,551][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.38749009370803833, 'train/grad_norm_pre_clip': 1.7196569442749023, 'timing/data': 0.00038988329470157623, 'timing/model': 1.2414826061576605, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 0.04}) +[2026-05-30 11:56:08,561][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.13502700626850128, 'train/grad_norm_pre_clip': 0.903478741645813, 'timing/data': 0.0003285091370344162, 'timing/model': 1.242214685305953, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 0.04}) +[2026-05-30 11:59:35,602][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 800 +[2026-05-30 11:59:35,603][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.07630658149719238, 'train/grad_norm_pre_clip': 0.861411452293396, 'eval/loss': 0.08937320113182068, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.005474388599395752, 'timing/model': 1.239403285086155, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.04}) +[2026-05-30 11:59:35,603][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_800_state +[2026-05-30 11:59:35,608][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 11:59:41,454][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_800_state/model.safetensors +[2026-05-30 11:59:48,775][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_800_state/optimizer.bin +[2026-05-30 11:59:48,776][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_800_state/sampler.bin +[2026-05-30 11:59:48,776][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_800_state/sampler_1.bin +[2026-05-30 11:59:48,777][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_800_state/random_states_0.pkl +[2026-05-30 11:59:52,594][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800 +[2026-05-30 11:59:52,594][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 11:59:52,606][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 12:03:01,898][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.19231490790843964, 'train/grad_norm_pre_clip': 2.791752338409424, 'timing/data': 0.011135024949908257, 'timing/model': 1.2402239721268415, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 0.04}) +[2026-05-30 12:06:12,034][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.022778913378715515, 'train/grad_norm_pre_clip': 1.433418869972229, 'timing/data': 0.004717448726296425, 'timing/model': 1.2433539349585772, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 0.04}) +[2026-05-30 12:09:22,513][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.04110591486096382, 'train/grad_norm_pre_clip': 1.6295663118362427, 'timing/data': 0.00035377591848373413, 'timing/model': 1.2536208592355251, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 0.04}) +[2026-05-30 12:12:32,046][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.28391364216804504, 'train/grad_norm_pre_clip': 3.3047494888305664, 'timing/data': 0.00042307376861572266, 'timing/model': 1.3971538972109556, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.04}) +[2026-05-30 12:15:41,394][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.13283421099185944, 'train/grad_norm_pre_clip': 1.2246007919311523, 'timing/data': 0.0003608260303735733, 'timing/model': 1.2488884329795837, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.04}) +[2026-05-30 12:18:51,633][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.4729948043823242, 'train/grad_norm_pre_clip': 6.828746795654297, 'timing/data': 0.00045754574239254, 'timing/model': 1.2361365146934986, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.04}) +[2026-05-30 12:22:00,898][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.12145914882421494, 'train/grad_norm_pre_clip': 2.824819803237915, 'timing/data': 0.003376403823494911, 'timing/model': 1.2381852194666862, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.04}) +[2026-05-30 12:25:11,459][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.11499185860157013, 'train/grad_norm_pre_clip': 0.794888436794281, 'timing/data': 0.0065713003277778625, 'timing/model': 1.2461243160068989, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 0.04}) +[2026-05-30 12:28:21,961][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.3226171135902405, 'train/grad_norm_pre_clip': 2.531545400619507, 'timing/data': 0.005424702540040016, 'timing/model': 1.4313270077109337, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 0.04}) +[2026-05-30 12:31:49,647][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 900 +[2026-05-30 12:31:49,649][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.026887590065598488, 'train/grad_norm_pre_clip': 0.5850814580917358, 'eval/loss': 0.11303458362817764, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0003478247672319412, 'timing/model': 1.2418829761445522, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 0.04}) +[2026-05-30 12:31:49,649][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_900_state +[2026-05-30 12:31:49,654][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 12:31:55,551][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_900_state/model.safetensors +[2026-05-30 12:32:02,767][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_900_state/optimizer.bin +[2026-05-30 12:32:02,768][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_900_state/sampler.bin +[2026-05-30 12:32:02,768][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_900_state/sampler_1.bin +[2026-05-30 12:32:02,769][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_900_state/random_states_0.pkl +[2026-05-30 12:32:06,617][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-05-30 12:32:06,617][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 12:32:06,629][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 12:35:16,597][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.03804881498217583, 'train/grad_norm_pre_clip': 1.011723518371582, 'timing/data': 0.00048536621034145355, 'timing/model': 1.2968994099646807, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 0.04}) +[2026-05-30 12:38:27,504][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.08485011756420135, 'train/grad_norm_pre_clip': 3.3465895652770996, 'timing/data': 0.003431743010878563, 'timing/model': 1.281703270971775, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 0.05}) +[2026-05-30 12:41:36,607][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.1693595051765442, 'train/grad_norm_pre_clip': 1.4510023593902588, 'timing/data': 0.007918672636151314, 'timing/model': 1.2356108222156763, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 0.05}) +[2026-05-30 12:44:46,614][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.02775251492857933, 'train/grad_norm_pre_clip': 0.14823412895202637, 'timing/data': 0.005643857643008232, 'timing/model': 1.288115156814456, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.05}) +[2026-05-30 12:47:57,531][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.04985678941011429, 'train/grad_norm_pre_clip': 0.639286994934082, 'timing/data': 0.0033360403031110764, 'timing/model': 1.2511496245861053, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 0.05}) +[2026-05-30 12:51:08,804][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.8522588610649109, 'train/grad_norm_pre_clip': 6.629443168640137, 'timing/data': 0.011238135397434235, 'timing/model': 1.3700767699629068, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 0.05}) +[2026-05-30 12:54:17,978][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.07797379791736603, 'train/grad_norm_pre_clip': 0.9099579453468323, 'timing/data': 0.00927230529487133, 'timing/model': 1.2398945186287165, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 0.05}) +[2026-05-30 12:57:28,121][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.012431769631803036, 'train/grad_norm_pre_clip': 0.5549538731575012, 'timing/data': 0.005530418828129768, 'timing/model': 1.2382551357150078, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 0.05}) +[2026-05-30 13:00:38,651][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.14787617325782776, 'train/grad_norm_pre_clip': 5.943774700164795, 'timing/data': 0.0003322046250104904, 'timing/model': 1.2388456463813782, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 0.05}) +[2026-05-30 13:05:00,040][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-05-30 13:05:00,041][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.010759961791336536, 'train/grad_norm_pre_clip': 0.06458742171525955, 'eval/loss': 0.09320129454135895, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 78.39999999999979, 'rl_games_eval/mid_train/mean_length': 651.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 78.39999999999979, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 651.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0004620049148797989, 'timing/model': 1.2417815942317247, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 0.05}) +[2026-05-30 13:05:00,041][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1000_state +[2026-05-30 13:05:00,046][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 13:05:05,919][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1000_state/model.safetensors +[2026-05-30 13:05:13,346][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1000_state/optimizer.bin +[2026-05-30 13:05:13,347][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1000_state/sampler.bin +[2026-05-30 13:05:13,347][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1000_state/sampler_1.bin +[2026-05-30 13:05:13,348][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1000_state/random_states_0.pkl +[2026-05-30 13:05:17,176][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-05-30 13:05:17,176][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 13:05:17,189][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 13:08:27,693][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.0075942715629935265, 'train/grad_norm_pre_clip': 0.08812547475099564, 'timing/data': 0.0003380775451660156, 'timing/model': 1.2669473383575678, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.05}) +[2026-05-30 13:11:37,124][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.033459797501564026, 'train/grad_norm_pre_clip': 0.44067612290382385, 'timing/data': 0.0003384482115507126, 'timing/model': 1.248282378539443, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 0.05}) +[2026-05-30 13:14:47,590][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.10348384082317352, 'train/grad_norm_pre_clip': 1.0219526290893555, 'timing/data': 0.0003725271672010422, 'timing/model': 1.439131699502468, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 0.05}) +[2026-05-30 13:17:57,510][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.27853909134864807, 'train/grad_norm_pre_clip': 8.325739860534668, 'timing/data': 0.0003356263041496277, 'timing/model': 1.2844558767974377, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 0.05}) +[2026-05-30 13:21:06,983][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.1305503398180008, 'train/grad_norm_pre_clip': 1.4476196765899658, 'timing/data': 0.0003398153930902481, 'timing/model': 1.240512976422906, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 0.05}) +[2026-05-30 13:24:17,502][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.21475085616111755, 'train/grad_norm_pre_clip': 3.2860891819000244, 'timing/data': 0.005692135542631149, 'timing/model': 1.2392502054572105, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 0.05}) +[2026-05-30 13:27:26,854][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.1284334659576416, 'train/grad_norm_pre_clip': 1.4295992851257324, 'timing/data': 0.005384249612689018, 'timing/model': 1.2417396251112223, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 0.05}) +[2026-05-30 13:30:37,103][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.0568775050342083, 'train/grad_norm_pre_clip': 0.9644986391067505, 'timing/data': 0.0003403555601835251, 'timing/model': 1.242165731266141, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.05}) +[2026-05-30 13:33:47,334][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.2957119345664978, 'train/grad_norm_pre_clip': 2.7548410892486572, 'timing/data': 0.0003438759595155716, 'timing/model': 1.236751563847065, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 0.05}) +[2026-05-30 13:37:15,364][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1100 +[2026-05-30 13:37:15,365][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.23111239075660706, 'train/grad_norm_pre_clip': 5.688257694244385, 'eval/loss': 0.07636179029941559, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0004723556339740753, 'timing/model': 1.2483043242245913, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 0.05}) +[2026-05-30 13:37:15,365][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1100_state +[2026-05-30 13:37:15,370][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 13:37:21,872][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1100_state/model.safetensors +[2026-05-30 13:37:29,191][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1100_state/optimizer.bin +[2026-05-30 13:37:29,192][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1100_state/sampler.bin +[2026-05-30 13:37:29,192][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1100_state/sampler_1.bin +[2026-05-30 13:37:29,193][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1100_state/random_states_0.pkl +[2026-05-30 13:37:32,960][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-05-30 13:37:32,960][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 13:37:32,972][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 13:40:44,397][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.11133906245231628, 'train/grad_norm_pre_clip': 1.330583930015564, 'timing/data': 0.0007499530911445618, 'timing/model': 1.2767592035233974, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 0.05}) +[2026-05-30 13:43:54,147][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.10008127242326736, 'train/grad_norm_pre_clip': 0.459598571062088, 'timing/data': 0.0003537256270647049, 'timing/model': 1.2386502958834171, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 0.06}) +[2026-05-30 13:47:03,978][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.056230153888463974, 'train/grad_norm_pre_clip': 0.6143895387649536, 'timing/data': 0.00046249479055404663, 'timing/model': 1.2402562629431486, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 0.06}) +[2026-05-30 13:50:14,589][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.26003000140190125, 'train/grad_norm_pre_clip': 0.6132669448852539, 'timing/data': 0.0055154915899038315, 'timing/model': 1.2385259605944157, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 0.06}) +[2026-05-30 13:53:24,230][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.06189815700054169, 'train/grad_norm_pre_clip': 0.6653033494949341, 'timing/data': 0.008269429206848145, 'timing/model': 1.2417175807058811, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.06}) +[2026-05-30 13:56:33,449][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.1265021115541458, 'train/grad_norm_pre_clip': 2.90675950050354, 'timing/data': 0.005581887438893318, 'timing/model': 1.2398654371500015, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 0.06}) +[2026-05-30 13:59:44,178][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.011797836981713772, 'train/grad_norm_pre_clip': 0.15315969288349152, 'timing/data': 0.013200115412473679, 'timing/model': 1.2367856930941343, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 0.06}) +[2026-05-30 14:02:53,959][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.14613835513591766, 'train/grad_norm_pre_clip': 1.7896456718444824, 'timing/data': 0.0003397911787033081, 'timing/model': 1.2463012672960758, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 0.06}) +[2026-05-30 14:06:05,667][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.03238541632890701, 'train/grad_norm_pre_clip': 0.504871129989624, 'timing/data': 0.007589949294924736, 'timing/model': 1.2447698209434748, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 0.06}) +[2026-05-30 14:09:48,129][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1200 +[2026-05-30 14:09:48,131][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.13732028007507324, 'train/grad_norm_pre_clip': 5.217826843261719, 'eval/loss': 0.08592866361141205, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 22.40000000000006, 'rl_games_eval/mid_train/mean_length': 199.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 22.40000000000006, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 199.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.008936624974012375, 'timing/model': 1.2937780153006315, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 0.06}) +[2026-05-30 14:09:48,131][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1200_state +[2026-05-30 14:09:48,137][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 14:09:53,980][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1200_state/model.safetensors +[2026-05-30 14:10:01,198][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1200_state/optimizer.bin +[2026-05-30 14:10:01,199][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1200_state/sampler.bin +[2026-05-30 14:10:01,199][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1200_state/sampler_1.bin +[2026-05-30 14:10:01,200][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1200_state/random_states_0.pkl +[2026-05-30 14:10:04,976][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-05-30 14:10:04,976][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 14:10:04,988][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 14:13:15,101][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.2605891227722168, 'train/grad_norm_pre_clip': 1.555323600769043, 'timing/data': 0.007465822622179985, 'timing/model': 1.2343206331133842, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 0.06}) +[2026-05-30 14:16:24,736][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.09384115040302277, 'train/grad_norm_pre_clip': 1.5227380990982056, 'timing/data': 0.004212839528918266, 'timing/model': 1.2356239333748817, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.06}) +[2026-05-30 14:19:34,574][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.18019762635231018, 'train/grad_norm_pre_clip': 1.1452405452728271, 'timing/data': 0.007114533334970474, 'timing/model': 1.2490130588412285, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 0.06}) +[2026-05-30 14:22:45,712][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.13171422481536865, 'train/grad_norm_pre_clip': 2.9688756465911865, 'timing/data': 0.008553953841328621, 'timing/model': 1.238295054063201, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 0.06}) +[2026-05-30 14:25:55,017][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.10120151191949844, 'train/grad_norm_pre_clip': 1.0697345733642578, 'timing/data': 0.006210599094629288, 'timing/model': 1.2469766791909933, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 0.06}) +[2026-05-30 14:29:04,748][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.38671115040779114, 'train/grad_norm_pre_clip': 3.32961106300354, 'timing/data': 0.0032484866678714752, 'timing/model': 1.4291343800723553, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 0.06}) +[2026-05-30 14:32:18,535][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.32724860310554504, 'train/grad_norm_pre_clip': 2.813016176223755, 'timing/data': 0.005307471379637718, 'timing/model': 1.2385941706597805, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 0.06}) +[2026-05-30 14:35:27,594][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.36607056856155396, 'train/grad_norm_pre_clip': 1.7919942140579224, 'timing/data': 0.00034343451261520386, 'timing/model': 1.237412204965949, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 0.06}) +[2026-05-30 14:38:40,721][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.3203185796737671, 'train/grad_norm_pre_clip': 2.5535755157470703, 'timing/data': 0.00032944977283477783, 'timing/model': 1.2364917993545532, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.06}) +[2026-05-30 14:42:47,455][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1300 +[2026-05-30 14:42:47,457][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.10829462856054306, 'train/grad_norm_pre_clip': 1.6675338745117188, 'eval/loss': 0.07514538615942001, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 50.40000000000037, 'rl_games_eval/mid_train/mean_length': 425.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 50.40000000000037, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 425.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0003369990736246109, 'timing/model': 1.272863855585456, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 0.06}) +[2026-05-30 14:42:47,457][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1300_state +[2026-05-30 14:42:47,462][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 14:42:53,225][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1300_state/model.safetensors +[2026-05-30 14:43:00,371][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1300_state/optimizer.bin +[2026-05-30 14:43:00,371][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1300_state/sampler.bin +[2026-05-30 14:43:00,372][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1300_state/sampler_1.bin +[2026-05-30 14:43:00,373][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1300_state/random_states_0.pkl +[2026-05-30 14:43:04,104][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-05-30 14:43:04,104][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 14:43:04,116][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 14:46:16,253][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.13106893002986908, 'train/grad_norm_pre_clip': 0.9359630346298218, 'timing/data': 0.00033957697451114655, 'timing/model': 1.2554747685790062, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 0.06}) +[2026-05-30 14:49:26,194][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.01778435707092285, 'train/grad_norm_pre_clip': 0.2212456911802292, 'timing/data': 0.000463707372546196, 'timing/model': 1.2920398283749819, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 0.07}) +[2026-05-30 14:52:36,585][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.026616068556904793, 'train/grad_norm_pre_clip': 0.26880592107772827, 'timing/data': 0.0032295752316713333, 'timing/model': 1.2670324575155973, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 0.07}) +[2026-05-30 14:55:46,990][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.05331069976091385, 'train/grad_norm_pre_clip': 0.600135862827301, 'timing/data': 0.003199903294444084, 'timing/model': 1.240035779774189, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 0.07}) +[2026-05-30 14:58:58,107][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.01463804766535759, 'train/grad_norm_pre_clip': 0.41216227412223816, 'timing/data': 0.007624512538313866, 'timing/model': 1.2352402843534946, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 0.07}) +[2026-05-30 15:02:07,633][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.30301135778427124, 'train/grad_norm_pre_clip': 5.055172443389893, 'timing/data': 0.0003399364650249481, 'timing/model': 1.2362431176006794, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.07}) +[2026-05-30 15:05:18,272][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.00202021817676723, 'train/grad_norm_pre_clip': 0.05403389781713486, 'timing/data': 0.0033567771315574646, 'timing/model': 1.4948673732578754, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 0.07}) +[2026-05-30 15:08:28,505][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.15719278156757355, 'train/grad_norm_pre_clip': 1.8820927143096924, 'timing/data': 0.005698073655366898, 'timing/model': 1.238971596583724, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 0.07}) +[2026-05-30 15:11:38,819][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.18862301111221313, 'train/grad_norm_pre_clip': 3.1041061878204346, 'timing/data': 0.0003463663160800934, 'timing/model': 1.2354919873178005, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 0.07}) +[2026-05-30 15:15:48,024][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1400 +[2026-05-30 15:15:48,026][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.09670641273260117, 'train/grad_norm_pre_clip': 0.9890962839126587, 'eval/loss': 0.08587933331727982, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 55.80000000000043, 'rl_games_eval/mid_train/mean_length': 470.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 55.80000000000043, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 470.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.00033859536051750183, 'timing/model': 1.2763753831386566, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 0.07}) +[2026-05-30 15:15:48,026][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1400_state +[2026-05-30 15:15:48,032][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 15:15:53,959][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1400_state/model.safetensors +[2026-05-30 15:16:01,203][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1400_state/optimizer.bin +[2026-05-30 15:16:01,204][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1400_state/sampler.bin +[2026-05-30 15:16:01,204][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1400_state/sampler_1.bin +[2026-05-30 15:16:01,205][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1400_state/random_states_0.pkl +[2026-05-30 15:16:04,980][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-05-30 15:16:04,980][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 15:16:04,992][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 15:19:15,188][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.17723459005355835, 'train/grad_norm_pre_clip': 3.0496156215667725, 'timing/data': 0.0003335345536470413, 'timing/model': 1.2412226889282465, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 0.07}) +[2026-05-30 15:22:25,192][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.15175151824951172, 'train/grad_norm_pre_clip': 2.688110828399658, 'timing/data': 0.0003328640013933182, 'timing/model': 1.242143176496029, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 0.07}) +[2026-05-30 15:25:33,900][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.041828546673059464, 'train/grad_norm_pre_clip': 1.0312590599060059, 'timing/data': 0.0003422442823648453, 'timing/model': 1.2796487510204315, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.07}) +[2026-05-30 15:28:44,938][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.0013026735978201032, 'train/grad_norm_pre_clip': 0.018103433772921562, 'timing/data': 0.0006109140813350677, 'timing/model': 1.2412101030349731, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 0.07}) +[2026-05-30 15:31:54,964][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.006778561044484377, 'train/grad_norm_pre_clip': 0.26433637738227844, 'timing/data': 0.00032493844628334045, 'timing/model': 1.2406693920493126, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 0.07}) +[2026-05-30 15:35:05,000][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.05655084177851677, 'train/grad_norm_pre_clip': 0.874199628829956, 'timing/data': 0.0003609228879213333, 'timing/model': 1.239682488143444, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 0.07}) +[2026-05-30 15:38:15,535][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.28948974609375, 'train/grad_norm_pre_clip': 3.016936779022217, 'timing/data': 0.00036158598959445953, 'timing/model': 1.2403696663677692, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 0.07}) +[2026-05-30 15:41:26,286][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.393403023481369, 'train/grad_norm_pre_clip': 5.737319469451904, 'timing/data': 0.007061239331960678, 'timing/model': 1.2360410708934069, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 0.07}) +[2026-05-30 15:44:36,112][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.025857850909233093, 'train/grad_norm_pre_clip': 0.29581451416015625, 'timing/data': 0.0003304164856672287, 'timing/model': 1.288883799687028, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 0.07}) +[2026-05-30 15:49:37,335][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-05-30 15:49:37,337][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.03636141121387482, 'train/grad_norm_pre_clip': 0.6338685154914856, 'eval/loss': 0.09031224250793457, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 130.79999999999742, 'rl_games_eval/mid_train/mean_length': 1076.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 130.79999999999742, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1076.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0003810431808233261, 'timing/model': 1.240335701033473, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.07}) +[2026-05-30 15:49:37,337][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1500_state +[2026-05-30 15:49:37,343][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 15:49:43,294][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1500_state/model.safetensors +[2026-05-30 15:49:50,482][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1500_state/optimizer.bin +[2026-05-30 15:49:50,483][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1500_state/sampler.bin +[2026-05-30 15:49:50,483][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1500_state/sampler_1.bin +[2026-05-30 15:49:50,485][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1500_state/random_states_0.pkl +[2026-05-30 15:49:54,239][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-05-30 15:49:54,240][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 15:49:54,252][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 15:53:04,408][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.11616356670856476, 'train/grad_norm_pre_clip': 2.297292709350586, 'timing/data': 0.0003511551767587662, 'timing/model': 1.2945734299719334, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 0.07}) +[2026-05-30 15:56:15,000][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.11076269298791885, 'train/grad_norm_pre_clip': 1.522246241569519, 'timing/data': 0.0003373846411705017, 'timing/model': 1.2739939112216234, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 0.08}) +[2026-05-30 15:59:25,094][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.27664873003959656, 'train/grad_norm_pre_clip': 2.92240834236145, 'timing/data': 0.0003405865281820297, 'timing/model': 1.2368057779967785, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 0.08}) +[2026-05-30 16:02:34,531][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.012387199327349663, 'train/grad_norm_pre_clip': 0.44504356384277344, 'timing/data': 0.0003278646618127823, 'timing/model': 1.2372260596603155, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 0.08}) +[2026-05-30 16:05:44,747][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.02494772896170616, 'train/grad_norm_pre_clip': 1.3843092918395996, 'timing/data': 0.0003593079745769501, 'timing/model': 1.238178415223956, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 0.08}) +[2026-05-30 16:08:54,887][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.0471007376909256, 'train/grad_norm_pre_clip': 0.5064020752906799, 'timing/data': 0.0013529472053050995, 'timing/model': 1.2517710532993078, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 0.08}) +[2026-05-30 16:12:04,692][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.050245244055986404, 'train/grad_norm_pre_clip': 0.8940401077270508, 'timing/data': 0.0004640631377696991, 'timing/model': 1.233030490577221, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.08}) +[2026-05-30 16:15:14,711][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.015065890736877918, 'train/grad_norm_pre_clip': 0.263828843832016, 'timing/data': 0.0003342851996421814, 'timing/model': 1.2396859023720026, 'learning_rate/qwen_vl_interface': 8.017284756424381e-06, 'learning_rate/action_model': 8.017284756424381e-05, 'epoch': 0.08}) +[2026-05-30 16:18:25,817][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.32846224308013916, 'train/grad_norm_pre_clip': 1.642619252204895, 'timing/data': 0.005640748888254166, 'timing/model': 1.292656583711505, 'learning_rate/qwen_vl_interface': 7.992475812221894e-06, 'learning_rate/action_model': 7.992475812221894e-05, 'epoch': 0.08}) +[2026-05-30 16:21:53,235][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1600 +[2026-05-30 16:21:53,237][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.1837611049413681, 'train/grad_norm_pre_clip': 0.9705765247344971, 'eval/loss': 0.11210548877716064, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.005595410242676735, 'timing/model': 1.3232166580855846, 'learning_rate/qwen_vl_interface': 7.967554135580307e-06, 'learning_rate/action_model': 7.967554135580307e-05, 'epoch': 0.08}) +[2026-05-30 16:21:53,238][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1600_state +[2026-05-30 16:21:53,243][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 16:21:59,382][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1600_state/model.safetensors +[2026-05-30 16:22:06,931][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1600_state/optimizer.bin +[2026-05-30 16:22:06,932][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1600_state/sampler.bin +[2026-05-30 16:22:06,932][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1600_state/sampler_1.bin +[2026-05-30 16:22:06,934][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1600_state/random_states_0.pkl +[2026-05-30 16:22:10,750][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1600 +[2026-05-30 16:22:10,751][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 16:22:10,762][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 16:25:19,774][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.1543215662240982, 'train/grad_norm_pre_clip': 1.0535502433776855, 'timing/data': 0.006905009970068932, 'timing/model': 1.2374432682991028, 'learning_rate/qwen_vl_interface': 7.942520750932138e-06, 'learning_rate/action_model': 7.942520750932138e-05, 'epoch': 0.08}) +[2026-05-30 16:28:30,508][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.014830080792307854, 'train/grad_norm_pre_clip': 0.11706401407718658, 'timing/data': 0.005231337621808052, 'timing/model': 1.2364899162203074, 'learning_rate/qwen_vl_interface': 7.91737668730178e-06, 'learning_rate/action_model': 7.91737668730178e-05, 'epoch': 0.08}) +[2026-05-30 16:31:40,599][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.10379421710968018, 'train/grad_norm_pre_clip': 2.4008874893188477, 'timing/data': 0.00520065613090992, 'timing/model': 1.2752932514995337, 'learning_rate/qwen_vl_interface': 7.89212297826321e-06, 'learning_rate/action_model': 7.892122978263209e-05, 'epoch': 0.08}) +[2026-05-30 16:34:50,839][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.015921538695693016, 'train/grad_norm_pre_clip': 0.25217360258102417, 'timing/data': 0.00842256285250187, 'timing/model': 1.2405412998050451, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.08}) +[2026-05-30 16:38:01,630][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.028967631980776787, 'train/grad_norm_pre_clip': 0.8975816965103149, 'timing/data': 0.0031838547438383102, 'timing/model': 1.2393355015665293, 'learning_rate/qwen_vl_interface': 7.841290780750107e-06, 'learning_rate/action_model': 7.841290780750107e-05, 'epoch': 0.08}) +[2026-05-30 16:41:10,901][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.20708413422107697, 'train/grad_norm_pre_clip': 2.789334297180176, 'timing/data': 0.00033535994589328766, 'timing/model': 1.2357438188046217, 'learning_rate/qwen_vl_interface': 7.815714381788123e-06, 'learning_rate/action_model': 7.815714381788122e-05, 'epoch': 0.08}) +[2026-05-30 16:44:21,507][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.06963350623846054, 'train/grad_norm_pre_clip': 0.40153735876083374, 'timing/data': 0.0005539581179618835, 'timing/model': 1.283942123875022, 'learning_rate/qwen_vl_interface': 7.790032516357116e-06, 'learning_rate/action_model': 7.790032516357117e-05, 'epoch': 0.08}) +[2026-05-30 16:47:31,079][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.16915559768676758, 'train/grad_norm_pre_clip': 1.3148916959762573, 'timing/data': 0.0014748144894838333, 'timing/model': 1.2433498054742813, 'learning_rate/qwen_vl_interface': 7.764246240137991e-06, 'learning_rate/action_model': 7.764246240137991e-05, 'epoch': 0.08}) +[2026-05-30 16:50:40,994][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.10685330629348755, 'train/grad_norm_pre_clip': 1.2745248079299927, 'timing/data': 0.0003656763583421707, 'timing/model': 1.2948838490992785, 'learning_rate/qwen_vl_interface': 7.738356613103562e-06, 'learning_rate/action_model': 7.73835661310356e-05, 'epoch': 0.08}) +[2026-05-30 16:54:18,444][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1700 +[2026-05-30 16:54:18,446][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.14054429531097412, 'train/grad_norm_pre_clip': 0.9318180084228516, 'eval/loss': 0.07350648939609528, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0005512386560440063, 'timing/model': 1.291161185130477, 'learning_rate/qwen_vl_interface': 7.712364699474996e-06, 'learning_rate/action_model': 7.712364699474996e-05, 'epoch': 0.08}) +[2026-05-30 16:54:18,446][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1700_state +[2026-05-30 16:54:18,450][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 16:54:24,430][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1700_state/model.safetensors +[2026-05-30 16:54:31,752][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1700_state/optimizer.bin +[2026-05-30 16:54:31,753][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1700_state/sampler.bin +[2026-05-30 16:54:31,753][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1700_state/sampler_1.bin +[2026-05-30 16:54:31,754][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1700_state/random_states_0.pkl +[2026-05-30 16:54:35,584][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1700 +[2026-05-30 16:54:35,585][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 16:54:35,597][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 16:57:45,980][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.13572409749031067, 'train/grad_norm_pre_clip': 1.8191498517990112, 'timing/data': 0.00033733993768692017, 'timing/model': 1.2887369599193335, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.08}) +[2026-05-30 17:01:00,388][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.28089794516563416, 'train/grad_norm_pre_clip': 1.7074915170669556, 'timing/data': 0.0003386400640010834, 'timing/model': 1.2865653820335865, 'learning_rate/qwen_vl_interface': 7.66007829029919e-06, 'learning_rate/action_model': 7.660078290299191e-05, 'epoch': 0.08}) +[2026-05-30 17:04:14,094][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.002915819175541401, 'train/grad_norm_pre_clip': 0.12222263962030411, 'timing/data': 0.0003381967544555664, 'timing/model': 1.290502443909645, 'learning_rate/qwen_vl_interface': 7.633785944041444e-06, 'learning_rate/action_model': 7.633785944041444e-05, 'epoch': 0.09}) +[2026-05-30 17:07:29,187][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.13808763027191162, 'train/grad_norm_pre_clip': 8.699498176574707, 'timing/data': 0.00034206733107566833, 'timing/model': 1.2761097252368927, 'learning_rate/qwen_vl_interface': 7.607395609680189e-06, 'learning_rate/action_model': 7.60739560968019e-05, 'epoch': 0.09}) +[2026-05-30 17:10:43,292][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.30145516991615295, 'train/grad_norm_pre_clip': 1.6864548921585083, 'timing/data': 0.0003419574350118637, 'timing/model': 1.2351517286151648, 'learning_rate/qwen_vl_interface': 7.580908372018706e-06, 'learning_rate/action_model': 7.580908372018705e-05, 'epoch': 0.09}) +[2026-05-30 17:13:53,793][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.11879777908325195, 'train/grad_norm_pre_clip': 12.296375274658203, 'timing/data': 0.0003413688391447067, 'timing/model': 1.2855002917349339, 'learning_rate/qwen_vl_interface': 7.554325319843593e-06, 'learning_rate/action_model': 7.554325319843592e-05, 'epoch': 0.09}) +[2026-05-30 17:17:08,195][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.0016414506826549768, 'train/grad_norm_pre_clip': 0.05534926801919937, 'timing/data': 0.0003335755318403244, 'timing/model': 1.2917454522103071, 'learning_rate/qwen_vl_interface': 7.527647545880007e-06, 'learning_rate/action_model': 7.527647545880007e-05, 'epoch': 0.09}) +[2026-05-30 17:20:21,433][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.03738543018698692, 'train/grad_norm_pre_clip': 0.7278103232383728, 'timing/data': 0.003298060968518257, 'timing/model': 1.4284629952162504, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.09}) +[2026-05-30 17:23:34,156][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.37102729082107544, 'train/grad_norm_pre_clip': 5.046192169189453, 'timing/data': 0.0025455765426158905, 'timing/model': 1.2555349711328745, 'learning_rate/qwen_vl_interface': 7.474012222911169e-06, 'learning_rate/action_model': 7.474012222911169e-05, 'epoch': 0.09}) +[2026-05-30 17:27:07,503][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1800 +[2026-05-30 17:27:07,504][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.03717385232448578, 'train/grad_norm_pre_clip': 0.46074917912483215, 'eval/loss': 0.08208555728197098, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.00371515192091465, 'timing/model': 1.2503970339894295, 'learning_rate/qwen_vl_interface': 7.447056878643969e-06, 'learning_rate/action_model': 7.447056878643968e-05, 'epoch': 0.09}) +[2026-05-30 17:27:07,505][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1800_state +[2026-05-30 17:27:07,509][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 17:27:13,381][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1800_state/model.safetensors +[2026-05-30 17:27:20,665][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1800_state/optimizer.bin +[2026-05-30 17:27:20,666][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1800_state/sampler.bin +[2026-05-30 17:27:20,666][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1800_state/sampler_1.bin +[2026-05-30 17:27:20,667][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1800_state/random_states_0.pkl +[2026-05-30 17:27:24,498][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1800 +[2026-05-30 17:27:24,499][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 17:27:24,510][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 17:30:38,695][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.10141116380691528, 'train/grad_norm_pre_clip': 1.4989126920700073, 'timing/data': 0.003245275467634201, 'timing/model': 1.3000366743654013, 'learning_rate/qwen_vl_interface': 7.420011221973771e-06, 'learning_rate/action_model': 7.420011221973771e-05, 'epoch': 0.09}) +[2026-05-30 17:33:50,183][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.017956245690584183, 'train/grad_norm_pre_clip': 0.31530025601387024, 'timing/data': 0.0032671410590410233, 'timing/model': 1.2825014479458332, 'learning_rate/qwen_vl_interface': 7.39287636464159e-06, 'learning_rate/action_model': 7.39287636464159e-05, 'epoch': 0.09}) +[2026-05-30 17:37:04,786][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.2772592604160309, 'train/grad_norm_pre_clip': 1.498744010925293, 'timing/data': 0.00362466461956501, 'timing/model': 1.2817408926784992, 'learning_rate/qwen_vl_interface': 7.365653422055131e-06, 'learning_rate/action_model': 7.365653422055131e-05, 'epoch': 0.09}) +[2026-05-30 17:40:17,916][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.02751700021326542, 'train/grad_norm_pre_clip': 0.9796090722084045, 'timing/data': 0.0032005198299884796, 'timing/model': 1.233315059915185, 'learning_rate/qwen_vl_interface': 7.338343513242941e-06, 'learning_rate/action_model': 7.338343513242941e-05, 'epoch': 0.09}) +[2026-05-30 17:43:32,493][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.019337622448801994, 'train/grad_norm_pre_clip': 1.0138429403305054, 'timing/data': 0.000654757022857666, 'timing/model': 1.2500452939420938, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.09}) +[2026-05-30 17:46:46,791][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.005612954031676054, 'train/grad_norm_pre_clip': 0.1525854915380478, 'timing/data': 0.0004124678671360016, 'timing/model': 1.293762233108282, 'learning_rate/qwen_vl_interface': 7.283467290883598e-06, 'learning_rate/action_model': 7.283467290883598e-05, 'epoch': 0.09}) +[2026-05-30 17:50:01,291][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.1096496656537056, 'train/grad_norm_pre_clip': 1.9531033039093018, 'timing/data': 0.0004976280033588409, 'timing/model': 1.2951286975294352, 'learning_rate/qwen_vl_interface': 7.2559032330830105e-06, 'learning_rate/action_model': 7.25590323308301e-05, 'epoch': 0.09}) +[2026-05-30 17:53:14,894][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 8.844851254252717e-05, 'train/grad_norm_pre_clip': 0.004205310717225075, 'timing/data': 0.00035821087658405304, 'timing/model': 1.2958015948534012, 'learning_rate/qwen_vl_interface': 7.228256720457097e-06, 'learning_rate/action_model': 7.228256720457096e-05, 'epoch': 0.09}) +[2026-05-30 17:56:28,980][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.0314030721783638, 'train/grad_norm_pre_clip': 0.4797070324420929, 'timing/data': 0.00033603422343730927, 'timing/model': 1.271477296948433, 'learning_rate/qwen_vl_interface': 7.200528889445707e-06, 'learning_rate/action_model': 7.200528889445706e-05, 'epoch': 0.09}) +[2026-05-30 18:00:05,103][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1900 +[2026-05-30 18:00:05,135][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.11643648147583008, 'train/grad_norm_pre_clip': 2.3269474506378174, 'eval/loss': 0.06831929832696915, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.00032679177820682526, 'timing/model': 1.2358347605913877, 'learning_rate/qwen_vl_interface': 7.1727208798313704e-06, 'learning_rate/action_model': 7.172720879831371e-05, 'epoch': 0.09}) +[2026-05-30 18:00:05,135][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1900_state +[2026-05-30 18:00:05,141][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 18:00:10,856][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1900_state/model.safetensors +[2026-05-30 18:00:18,056][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1900_state/optimizer.bin +[2026-05-30 18:00:18,056][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1900_state/sampler.bin +[2026-05-30 18:00:18,056][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1900_state/sampler_1.bin +[2026-05-30 18:00:18,057][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_1900_state/random_states_0.pkl +[2026-05-30 18:00:21,862][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1900 +[2026-05-30 18:00:21,863][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 18:00:21,874][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 18:03:33,168][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.006014320068061352, 'train/grad_norm_pre_clip': 0.12492308020591736, 'timing/data': 0.0003357715904712677, 'timing/model': 1.2377700451761484, 'learning_rate/qwen_vl_interface': 7.14483383469245e-06, 'learning_rate/action_model': 7.144833834692449e-05, 'epoch': 0.09}) +[2026-05-30 18:06:46,889][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.18828342854976654, 'train/grad_norm_pre_clip': 1.6435441970825195, 'timing/data': 0.00033850036561489105, 'timing/model': 1.2601368762552738, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.09}) +[2026-05-30 18:10:01,583][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.00939233973622322, 'train/grad_norm_pre_clip': 0.2743111848831177, 'timing/data': 0.0003248881548643112, 'timing/model': 1.2821894008666277, 'learning_rate/qwen_vl_interface': 7.0888272263513595e-06, 'learning_rate/action_model': 7.088827226351359e-05, 'epoch': 0.1}) +[2026-05-30 18:13:16,087][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.055107660591602325, 'train/grad_norm_pre_clip': 2.8704304695129395, 'timing/data': 0.0003453269600868225, 'timing/model': 1.2865154631435871, 'learning_rate/qwen_vl_interface': 7.060709965361499e-06, 'learning_rate/action_model': 7.060709965361499e-05, 'epoch': 0.1}) +[2026-05-30 18:16:28,487][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.004052613861858845, 'train/grad_norm_pre_clip': 0.12894730269908905, 'timing/data': 0.0003406442701816559, 'timing/model': 1.2801501378417015, 'learning_rate/qwen_vl_interface': 7.032518273177028e-06, 'learning_rate/action_model': 7.032518273177028e-05, 'epoch': 0.1}) +[2026-05-30 18:19:42,482][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.451186865568161, 'train/grad_norm_pre_clip': 6.251677989959717, 'timing/data': 0.0004250258207321167, 'timing/model': 1.275327218696475, 'learning_rate/qwen_vl_interface': 7.004253308647993e-06, 'learning_rate/action_model': 7.004253308647994e-05, 'epoch': 0.1}) +[2026-05-30 18:22:55,034][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.1652809977531433, 'train/grad_norm_pre_clip': 1.5726723670959473, 'timing/data': 0.0003503691405057907, 'timing/model': 1.2370881456881762, 'learning_rate/qwen_vl_interface': 6.97591623363638e-06, 'learning_rate/action_model': 6.97591623363638e-05, 'epoch': 0.1}) +[2026-05-30 18:26:09,686][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.07185900956392288, 'train/grad_norm_pre_clip': 0.7384687662124634, 'timing/data': 0.0003334246575832367, 'timing/model': 1.2788953073322773, 'learning_rate/qwen_vl_interface': 6.947508212968356e-06, 'learning_rate/action_model': 6.947508212968356e-05, 'epoch': 0.1}) +[2026-05-30 18:29:23,119][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.016238590702414513, 'train/grad_norm_pre_clip': 0.40553519129753113, 'timing/data': 0.0003380626440048218, 'timing/model': 1.236829288303852, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.1}) +[2026-05-30 18:32:54,204][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2000 +[2026-05-30 18:32:54,206][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.24570932984352112, 'train/grad_norm_pre_clip': 2.879983425140381, 'eval/loss': 0.05452118068933487, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.00035444460809230804, 'timing/model': 1.30521154217422, 'learning_rate/qwen_vl_interface': 6.890484008501211e-06, 'learning_rate/action_model': 6.890484008501211e-05, 'epoch': 0.1}) +[2026-05-30 18:32:54,206][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2000_state +[2026-05-30 18:32:54,210][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 18:32:59,897][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2000_state/model.safetensors +[2026-05-30 18:33:07,195][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2000_state/optimizer.bin +[2026-05-30 18:33:07,195][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2000_state/sampler.bin +[2026-05-30 18:33:07,196][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2000_state/sampler_1.bin +[2026-05-30 18:33:07,197][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2000_state/random_states_0.pkl +[2026-05-30 18:33:11,014][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-05-30 18:33:11,014][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 18:33:11,026][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 18:36:25,488][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.20912490785121918, 'train/grad_norm_pre_clip': 8.885435104370117, 'timing/data': 0.00042576901614665985, 'timing/model': 1.2879858314990997, 'learning_rate/qwen_vl_interface': 6.861870168743793e-06, 'learning_rate/action_model': 6.861870168743793e-05, 'epoch': 0.1}) +[2026-05-30 18:39:39,689][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.11092586070299149, 'train/grad_norm_pre_clip': 1.697347640991211, 'timing/data': 0.00034993700683116913, 'timing/model': 1.2500505167990923, 'learning_rate/qwen_vl_interface': 6.833190071317e-06, 'learning_rate/action_model': 6.833190071316999e-05, 'epoch': 0.1}) +[2026-05-30 18:42:53,903][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.007628383114933968, 'train/grad_norm_pre_clip': 0.10661664605140686, 'timing/data': 0.0003482084721326828, 'timing/model': 1.304332448169589, 'learning_rate/qwen_vl_interface': 6.804444895147306e-06, 'learning_rate/action_model': 6.804444895147306e-05, 'epoch': 0.1}) +[2026-05-30 18:46:07,978][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.0034999754279851913, 'train/grad_norm_pre_clip': 0.15087610483169556, 'timing/data': 0.0003434065729379654, 'timing/model': 1.2823860626667738, 'learning_rate/qwen_vl_interface': 6.775635821836317e-06, 'learning_rate/action_model': 6.775635821836317e-05, 'epoch': 0.1}) +[2026-05-30 18:49:20,992][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.23504778742790222, 'train/grad_norm_pre_clip': 3.381620407104492, 'timing/data': 0.0004757251590490341, 'timing/model': 1.2950347755104303, 'learning_rate/qwen_vl_interface': 6.7467640356122e-06, 'learning_rate/action_model': 6.746764035612199e-05, 'epoch': 0.1}) +[2026-05-30 18:52:34,289][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.043636783957481384, 'train/grad_norm_pre_clip': 2.3685672283172607, 'timing/data': 0.0003566555678844452, 'timing/model': 1.281597811728716, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.1}) +[2026-05-30 18:55:46,191][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.013033534400165081, 'train/grad_norm_pre_clip': 0.23953722417354584, 'timing/data': 0.00032412633299827576, 'timing/model': 1.284092677757144, 'learning_rate/qwen_vl_interface': 6.688837074177868e-06, 'learning_rate/action_model': 6.688837074177868e-05, 'epoch': 0.1}) +[2026-05-30 18:58:56,645][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.008630363270640373, 'train/grad_norm_pre_clip': 0.7146943211555481, 'timing/data': 0.0003576558083295822, 'timing/model': 1.2355233021080494, 'learning_rate/qwen_vl_interface': 6.659784280118156e-06, 'learning_rate/action_model': 6.659784280118156e-05, 'epoch': 0.1}) +[2026-05-30 19:02:09,846][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.002627967158332467, 'train/grad_norm_pre_clip': 0.061493247747421265, 'timing/data': 0.000353153795003891, 'timing/model': 1.4441134333610535, 'learning_rate/qwen_vl_interface': 6.6306735353484365e-06, 'learning_rate/action_model': 6.630673535348436e-05, 'epoch': 0.1}) +[2026-05-30 19:08:12,280][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2100 +[2026-05-30 19:08:12,281][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.22157201170921326, 'train/grad_norm_pre_clip': 4.004286766052246, 'eval/loss': 0.07235649973154068, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 222.89999999999333, 'rl_games_eval/mid_train/mean_length': 1817.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 222.89999999999333, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1817.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.00034923478960990906, 'timing/model': 1.2904995437711477, 'learning_rate/qwen_vl_interface': 6.6015060364974045e-06, 'learning_rate/action_model': 6.601506036497404e-05, 'epoch': 0.1}) +[2026-05-30 19:08:12,282][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2100_state +[2026-05-30 19:08:12,286][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 19:08:18,026][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2100_state/model.safetensors +[2026-05-30 19:08:25,287][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2100_state/optimizer.bin +[2026-05-30 19:08:25,287][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2100_state/sampler.bin +[2026-05-30 19:08:25,287][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2100_state/sampler_1.bin +[2026-05-30 19:08:25,288][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2100_state/random_states_0.pkl +[2026-05-30 19:08:29,088][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2100 +[2026-05-30 19:08:29,089][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 19:08:29,100][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 19:11:42,001][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.4924173951148987, 'train/grad_norm_pre_clip': 4.959751605987549, 'timing/data': 0.0003412272781133652, 'timing/model': 1.232544668018818, 'learning_rate/qwen_vl_interface': 6.572282982526702e-06, 'learning_rate/action_model': 6.572282982526702e-05, 'epoch': 0.1}) +[2026-05-30 19:14:54,292][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.16954255104064941, 'train/grad_norm_pre_clip': 1.5992270708084106, 'timing/data': 0.00034842826426029205, 'timing/model': 1.2917164042592049, 'learning_rate/qwen_vl_interface': 6.543005574681613e-06, 'learning_rate/action_model': 6.543005574681613e-05, 'epoch': 0.1}) +[2026-05-30 19:18:08,884][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.13737747073173523, 'train/grad_norm_pre_clip': 3.5712709426879883, 'timing/data': 0.0003353189677000046, 'timing/model': 1.27575040794909, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.11}) +[2026-05-30 19:21:22,670][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.4384500980377197, 'train/grad_norm_pre_clip': 3.4099016189575195, 'timing/data': 0.00033584795892238617, 'timing/model': 1.2864438146352768, 'learning_rate/qwen_vl_interface': 6.484292513471352e-06, 'learning_rate/action_model': 6.484292513471352e-05, 'epoch': 0.11}) +[2026-05-30 19:24:36,692][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.06067829951643944, 'train/grad_norm_pre_clip': 1.2401162385940552, 'timing/data': 0.0003367960453033447, 'timing/model': 1.2743476796895266, 'learning_rate/qwen_vl_interface': 6.45485927357016e-06, 'learning_rate/action_model': 6.454859273570159e-05, 'epoch': 0.11}) +[2026-05-30 19:27:48,292][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.2725607752799988, 'train/grad_norm_pre_clip': 6.642940998077393, 'timing/data': 0.0003373771905899048, 'timing/model': 1.2894371841102839, 'learning_rate/qwen_vl_interface': 6.425376506623344e-06, 'learning_rate/action_model': 6.425376506623344e-05, 'epoch': 0.11}) +[2026-05-30 19:31:01,287][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.0512189045548439, 'train/grad_norm_pre_clip': 2.1267974376678467, 'timing/data': 0.0003386363387107849, 'timing/model': 1.2884986363351345, 'learning_rate/qwen_vl_interface': 6.395845424551975e-06, 'learning_rate/action_model': 6.395845424551975e-05, 'epoch': 0.11}) +[2026-05-30 19:34:15,602][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.05834489315748215, 'train/grad_norm_pre_clip': 2.104898452758789, 'timing/data': 0.0003346763551235199, 'timing/model': 1.2359258979558945, 'learning_rate/qwen_vl_interface': 6.3662672412631736e-06, 'learning_rate/action_model': 6.366267241263173e-05, 'epoch': 0.11}) +[2026-05-30 19:37:26,784][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.004740452393889427, 'train/grad_norm_pre_clip': 0.23203052580356598, 'timing/data': 0.00034293532371520996, 'timing/model': 1.2943571414798498, 'learning_rate/qwen_vl_interface': 6.336643172600202e-06, 'learning_rate/action_model': 6.336643172600202e-05, 'epoch': 0.11}) +[2026-05-30 19:44:12,286][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2200 +[2026-05-30 19:44:12,288][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.001643376424908638, 'train/grad_norm_pre_clip': 0.07930850982666016, 'eval/loss': 0.11823578923940659, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 283.8999999999971, 'rl_games_eval/mid_train/mean_length': 2310.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 283.8999999999971, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2310.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0003348253667354584, 'timing/model': 1.2795153371989727, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.11}) +[2026-05-30 19:44:12,288][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2200_state +[2026-05-30 19:44:12,293][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 19:44:18,081][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2200_state/model.safetensors +[2026-05-30 19:44:25,271][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2200_state/optimizer.bin +[2026-05-30 19:44:25,272][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2200_state/sampler.bin +[2026-05-30 19:44:25,272][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2200_state/sampler_1.bin +[2026-05-30 19:44:25,273][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2200_state/random_states_0.pkl +[2026-05-30 19:44:29,024][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2200 +[2026-05-30 19:44:29,024][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 19:44:29,036][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 19:47:43,785][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.00444761710241437, 'train/grad_norm_pre_clip': 0.152923583984375, 'timing/data': 0.0003455672413110733, 'timing/model': 1.27626689709723, 'learning_rate/qwen_vl_interface': 6.277262251905597e-06, 'learning_rate/action_model': 6.277262251905597e-05, 'epoch': 0.11}) +[2026-05-30 19:50:56,187][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.06514684855937958, 'train/grad_norm_pre_clip': 1.8013558387756348, 'timing/data': 0.0003435462713241577, 'timing/model': 1.2905181925743818, 'learning_rate/qwen_vl_interface': 6.247507840791035e-06, 'learning_rate/action_model': 6.247507840791035e-05, 'epoch': 0.11}) +[2026-05-30 19:54:09,137][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.010296673513948917, 'train/grad_norm_pre_clip': 0.7969647645950317, 'timing/data': 0.00033833831548690796, 'timing/model': 1.234368797391653, 'learning_rate/qwen_vl_interface': 6.21771242603611e-06, 'learning_rate/action_model': 6.21771242603611e-05, 'epoch': 0.11}) +[2026-05-30 19:57:22,588][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.0062950123101472855, 'train/grad_norm_pre_clip': 0.1606404334306717, 'timing/data': 0.0004240460693836212, 'timing/model': 1.2824283558875322, 'learning_rate/qwen_vl_interface': 6.187877232413619e-06, 'learning_rate/action_model': 6.18787723241362e-05, 'epoch': 0.11}) +[2026-05-30 20:00:34,583][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.09385860711336136, 'train/grad_norm_pre_clip': 3.4054694175720215, 'timing/data': 0.0003307964652776718, 'timing/model': 1.2905242517590523, 'learning_rate/qwen_vl_interface': 6.15800348633152e-06, 'learning_rate/action_model': 6.15800348633152e-05, 'epoch': 0.11}) +[2026-05-30 20:03:46,490][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.11056102067232132, 'train/grad_norm_pre_clip': 1.7204232215881348, 'timing/data': 0.0003450661897659302, 'timing/model': 1.2690280359238386, 'learning_rate/qwen_vl_interface': 6.128092415782502e-06, 'learning_rate/action_model': 6.128092415782501e-05, 'epoch': 0.11}) +[2026-05-30 20:07:00,953][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.37619900703430176, 'train/grad_norm_pre_clip': 3.7799954414367676, 'timing/data': 0.008192596957087517, 'timing/model': 1.2399434875696898, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.11}) +[2026-05-30 20:10:13,783][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.10022372007369995, 'train/grad_norm_pre_clip': 1.00623619556427, 'timing/data': 0.00796622782945633, 'timing/model': 1.2685205470770597, 'learning_rate/qwen_vl_interface': 6.068163220875266e-06, 'learning_rate/action_model': 6.068163220875266e-05, 'epoch': 0.11}) +[2026-05-30 20:13:26,384][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.03424437344074249, 'train/grad_norm_pre_clip': 0.4052219092845917, 'timing/data': 0.005713922902941704, 'timing/model': 1.2814026419073343, 'learning_rate/qwen_vl_interface': 6.038147559971523e-06, 'learning_rate/action_model': 6.038147559971523e-05, 'epoch': 0.11}) +[2026-05-30 20:17:00,703][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2300 +[2026-05-30 20:17:00,704][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.0001889536652015522, 'train/grad_norm_pre_clip': 0.0023745738435536623, 'eval/loss': 0.06340749561786652, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.008335236459970474, 'timing/model': 1.23585962690413, 'learning_rate/qwen_vl_interface': 6.008099501408552e-06, 'learning_rate/action_model': 6.008099501408552e-05, 'epoch': 0.11}) +[2026-05-30 20:17:00,705][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2300_state +[2026-05-30 20:17:00,709][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 20:17:06,475][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2300_state/model.safetensors +[2026-05-30 20:17:13,752][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2300_state/optimizer.bin +[2026-05-30 20:17:13,753][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2300_state/sampler.bin +[2026-05-30 20:17:13,753][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2300_state/sampler_1.bin +[2026-05-30 20:17:13,754][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2300_state/random_states_0.pkl +[2026-05-30 20:17:17,552][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2300 +[2026-05-30 20:17:17,553][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 20:17:17,933][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 20:20:28,352][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.010622220113873482, 'train/grad_norm_pre_clip': 0.21926873922348022, 'timing/data': 0.005425635725259781, 'timing/model': 1.2357936687767506, 'learning_rate/qwen_vl_interface': 5.978020280344351e-06, 'learning_rate/action_model': 5.9780202803443506e-05, 'epoch': 0.11}) +[2026-05-30 20:23:41,491][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.3155096173286438, 'train/grad_norm_pre_clip': 4.856421947479248, 'timing/data': 0.005570372566580772, 'timing/model': 1.2858827579766512, 'learning_rate/qwen_vl_interface': 5.947911133217886e-06, 'learning_rate/action_model': 5.947911133217886e-05, 'epoch': 0.11}) +[2026-05-30 20:26:55,884][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.2521189749240875, 'train/grad_norm_pre_clip': 4.488481521606445, 'timing/data': 0.005838802084326744, 'timing/model': 1.2770107947289944, 'learning_rate/qwen_vl_interface': 5.917773297698264e-06, 'learning_rate/action_model': 5.9177732976982636e-05, 'epoch': 0.12}) +[2026-05-30 20:30:08,751][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.014576108194887638, 'train/grad_norm_pre_clip': 0.5480111241340637, 'timing/data': 0.00540572963654995, 'timing/model': 1.2405000980943441, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 0.12}) +[2026-05-30 20:33:18,691][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.01770060881972313, 'train/grad_norm_pre_clip': 1.0150554180145264, 'timing/data': 0.0048292409628629684, 'timing/model': 1.2787764687091112, 'learning_rate/qwen_vl_interface': 5.857416518001404e-06, 'learning_rate/action_model': 5.857416518001404e-05, 'epoch': 0.12}) +[2026-05-30 20:36:32,087][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.00798186007887125, 'train/grad_norm_pre_clip': 0.6510253548622131, 'timing/data': 0.005438528954982758, 'timing/model': 1.284482853487134, 'learning_rate/qwen_vl_interface': 5.8272000548549776e-06, 'learning_rate/action_model': 5.827200054854978e-05, 'epoch': 0.12}) +[2026-05-30 20:39:45,614][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.14692796766757965, 'train/grad_norm_pre_clip': 22.629501342773438, 'timing/data': 0.005058052018284798, 'timing/model': 1.5016754288226366, 'learning_rate/qwen_vl_interface': 5.7969598652750334e-06, 'learning_rate/action_model': 5.796959865275033e-05, 'epoch': 0.12}) +[2026-05-30 20:42:59,080][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.04787776246666908, 'train/grad_norm_pre_clip': 4.1754608154296875, 'timing/data': 0.005765510722994804, 'timing/model': 1.2689896505326033, 'learning_rate/qwen_vl_interface': 5.766697192317323e-06, 'learning_rate/action_model': 5.766697192317323e-05, 'epoch': 0.12}) +[2026-05-30 20:46:13,284][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.17566588521003723, 'train/grad_norm_pre_clip': 4.054609298706055, 'timing/data': 0.0054727401584386826, 'timing/model': 1.2749705407768488, 'learning_rate/qwen_vl_interface': 5.7364132799617985e-06, 'learning_rate/action_model': 5.7364132799617984e-05, 'epoch': 0.12}) +[2026-05-30 20:52:04,060][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2400 +[2026-05-30 20:52:04,062][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.13145865499973297, 'train/grad_norm_pre_clip': 5.48021936416626, 'eval/loss': 0.05185149982571602, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 204.39999999999415, 'rl_games_eval/mid_train/mean_length': 1668.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 204.39999999999415, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1668.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.003213685005903244, 'timing/model': 1.2354601006954908, 'learning_rate/qwen_vl_interface': 5.706109373061489e-06, 'learning_rate/action_model': 5.706109373061489e-05, 'epoch': 0.12}) +[2026-05-30 20:52:04,062][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2400_state +[2026-05-30 20:52:04,067][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 20:52:10,158][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2400_state/model.safetensors +[2026-05-30 20:52:17,506][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2400_state/optimizer.bin +[2026-05-30 20:52:17,507][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2400_state/sampler.bin +[2026-05-30 20:52:17,507][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2400_state/sampler_1.bin +[2026-05-30 20:52:17,508][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2400_state/random_states_0.pkl +[2026-05-30 20:52:21,330][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2400 +[2026-05-30 20:52:21,331][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 20:52:21,343][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 20:55:33,641][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.18060043454170227, 'train/grad_norm_pre_clip': 2.0939810276031494, 'timing/data': 0.003392361104488373, 'timing/model': 1.2368908915668726, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 0.12}) +[2026-05-30 20:58:42,995][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.051511261612176895, 'train/grad_norm_pre_clip': 2.504260540008545, 'timing/data': 0.0033777672797441483, 'timing/model': 1.2868358362466097, 'learning_rate/qwen_vl_interface': 5.645446559096875e-06, 'learning_rate/action_model': 5.645446559096875e-05, 'epoch': 0.12}) +[2026-05-30 21:01:54,440][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.15960809588432312, 'train/grad_norm_pre_clip': 4.236668586730957, 'timing/data': 0.0031690970063209534, 'timing/model': 1.42760456725955, 'learning_rate/qwen_vl_interface': 5.6150901456432546e-06, 'learning_rate/action_model': 5.615090145643255e-05, 'epoch': 0.12}) +[2026-05-30 21:05:08,189][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.0013547383714467287, 'train/grad_norm_pre_clip': 0.04571882262825966, 'timing/data': 0.0031908992677927017, 'timing/model': 1.2772462703287601, 'learning_rate/qwen_vl_interface': 5.584718724763707e-06, 'learning_rate/action_model': 5.584718724763707e-05, 'epoch': 0.12}) +[2026-05-30 21:08:21,987][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.053346626460552216, 'train/grad_norm_pre_clip': 0.9773291349411011, 'timing/data': 0.0032780319452285767, 'timing/model': 1.289470363408327, 'learning_rate/qwen_vl_interface': 5.554333544908388e-06, 'learning_rate/action_model': 5.5543335449083875e-05, 'epoch': 0.12}) +[2026-05-30 21:11:36,594][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.09115056693553925, 'train/grad_norm_pre_clip': 1.7154825925827026, 'timing/data': 0.0030953437089920044, 'timing/model': 1.290480226278305, 'learning_rate/qwen_vl_interface': 5.52393585509303e-06, 'learning_rate/action_model': 5.52393585509303e-05, 'epoch': 0.12}) +[2026-05-30 21:14:50,642][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.011495695449411869, 'train/grad_norm_pre_clip': 0.24258486926555634, 'timing/data': 0.0032668672502040863, 'timing/model': 1.2413196880370378, 'learning_rate/qwen_vl_interface': 5.4935269048476015e-06, 'learning_rate/action_model': 5.493526904847601e-05, 'epoch': 0.12}) +[2026-05-30 21:18:04,587][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.04111441969871521, 'train/grad_norm_pre_clip': 1.0789499282836914, 'timing/data': 0.0054594241082668304, 'timing/model': 1.2846638821065426, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 0.12}) +[2026-05-30 21:21:19,187][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.0025286865420639515, 'train/grad_norm_pre_clip': 0.15694975852966309, 'timing/data': 0.0033017974346876144, 'timing/model': 1.2710327468812466, 'learning_rate/qwen_vl_interface': 5.432680223449392e-06, 'learning_rate/action_model': 5.4326802234493924e-05, 'epoch': 0.12}) +[2026-05-30 21:24:54,902][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2500 +[2026-05-30 21:24:54,903][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.01247810572385788, 'train/grad_norm_pre_clip': 0.5906057357788086, 'eval/loss': 0.07174710184335709, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.005626929923892021, 'timing/model': 1.278564978390932, 'learning_rate/qwen_vl_interface': 5.402244993465364e-06, 'learning_rate/action_model': 5.402244993465364e-05, 'epoch': 0.12}) +[2026-05-30 21:24:54,904][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2500_state +[2026-05-30 21:24:54,908][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 21:25:00,769][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2500_state/model.safetensors +[2026-05-30 21:25:07,936][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2500_state/optimizer.bin +[2026-05-30 21:25:07,937][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2500_state/sampler.bin +[2026-05-30 21:25:07,937][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2500_state/sampler_1.bin +[2026-05-30 21:25:07,938][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2500_state/random_states_0.pkl +[2026-05-30 21:25:11,715][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-05-30 21:25:11,715][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 21:25:11,728][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 21:28:26,081][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.20974858105182648, 'train/grad_norm_pre_clip': 3.7126412391662598, 'timing/data': 0.00576569139957428, 'timing/model': 1.2731088772416115, 'learning_rate/qwen_vl_interface': 5.371803505285954e-06, 'learning_rate/action_model': 5.371803505285954e-05, 'epoch': 0.12}) +[2026-05-30 21:31:40,785][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.09242948144674301, 'train/grad_norm_pre_clip': 0.703691303730011, 'timing/data': 0.0056668855249881744, 'timing/model': 1.2713438961654902, 'learning_rate/qwen_vl_interface': 5.3413570102415144e-06, 'learning_rate/action_model': 5.3413570102415144e-05, 'epoch': 0.12}) +[2026-05-30 21:34:52,694][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.07073425501585007, 'train/grad_norm_pre_clip': 3.596216917037964, 'timing/data': 0.005464153364300728, 'timing/model': 1.28967459872365, 'learning_rate/qwen_vl_interface': 5.310906759868205e-06, 'learning_rate/action_model': 5.310906759868205e-05, 'epoch': 0.12}) +[2026-05-30 21:38:12,181][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.035750072449445724, 'train/grad_norm_pre_clip': 1.3054115772247314, 'timing/data': 0.003194715827703476, 'timing/model': 1.2773754633963108, 'learning_rate/qwen_vl_interface': 5.280454005856553e-06, 'learning_rate/action_model': 5.2804540058565534e-05, 'epoch': 0.13}) +[2026-05-30 21:41:26,887][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.019833216443657875, 'train/grad_norm_pre_clip': 0.34133657813072205, 'timing/data': 0.003100285306572914, 'timing/model': 1.2875124644488096, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.13}) +[2026-05-30 21:44:38,396][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.09693790227174759, 'train/grad_norm_pre_clip': 3.909768581390381, 'timing/data': 0.0003448557108640671, 'timing/model': 1.2914233673363924, 'learning_rate/qwen_vl_interface': 5.219545994143448e-06, 'learning_rate/action_model': 5.219545994143448e-05, 'epoch': 0.13}) +[2026-05-30 21:47:50,410][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.00015158028691075742, 'train/grad_norm_pre_clip': 0.0043753464706242085, 'timing/data': 0.003363087773323059, 'timing/model': 1.2342523038387299, 'learning_rate/qwen_vl_interface': 5.189093240131795e-06, 'learning_rate/action_model': 5.189093240131795e-05, 'epoch': 0.13}) +[2026-05-30 21:51:03,495][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.23083537817001343, 'train/grad_norm_pre_clip': 3.5634326934814453, 'timing/data': 0.005673078820109367, 'timing/model': 1.2883907668292522, 'learning_rate/qwen_vl_interface': 5.158642989758487e-06, 'learning_rate/action_model': 5.158642989758487e-05, 'epoch': 0.13}) +[2026-05-30 21:54:17,214][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.025270402431488037, 'train/grad_norm_pre_clip': 1.3330940008163452, 'timing/data': 0.005385152995586395, 'timing/model': 1.235957095399499, 'learning_rate/qwen_vl_interface': 5.128196494714048e-06, 'learning_rate/action_model': 5.128196494714047e-05, 'epoch': 0.13}) +[2026-05-30 21:58:01,802][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2600 +[2026-05-30 21:58:01,803][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.2222127914428711, 'train/grad_norm_pre_clip': 1.9136806726455688, 'eval/loss': 0.05847739055752754, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 17.90000000000001, 'rl_games_eval/mid_train/mean_length': 163.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 17.90000000000001, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 163.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0058928970247507095, 'timing/model': 1.2919804640114307, 'learning_rate/qwen_vl_interface': 5.097755006534638e-06, 'learning_rate/action_model': 5.097755006534638e-05, 'epoch': 0.13}) +[2026-05-30 21:58:01,804][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2600_state +[2026-05-30 21:58:01,809][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 21:58:07,583][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2600_state/model.safetensors +[2026-05-30 21:58:14,849][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2600_state/optimizer.bin +[2026-05-30 21:58:14,850][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2600_state/sampler.bin +[2026-05-30 21:58:14,850][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2600_state/sampler_1.bin +[2026-05-30 21:58:14,851][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2600_state/random_states_0.pkl +[2026-05-30 21:58:18,638][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2600 +[2026-05-30 21:58:18,639][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 21:58:18,650][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 22:01:30,012][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.0009885692270472646, 'train/grad_norm_pre_clip': 0.017323274165391922, 'timing/data': 0.003703078255057335, 'timing/model': 1.2903576511889696, 'learning_rate/qwen_vl_interface': 5.067319776550607e-06, 'learning_rate/action_model': 5.067319776550607e-05, 'epoch': 0.13}) +[2026-05-30 22:04:42,185][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.011137722991406918, 'train/grad_norm_pre_clip': 0.1969512701034546, 'timing/data': 0.003049548715353012, 'timing/model': 1.2845856230705976, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 0.13}) +[2026-05-30 22:07:56,004][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.009305568411946297, 'train/grad_norm_pre_clip': 0.2987006604671478, 'timing/data': 0.00037663616240024567, 'timing/model': 1.2435128334909678, 'learning_rate/qwen_vl_interface': 5.0064730951524e-06, 'learning_rate/action_model': 5.0064730951524e-05, 'epoch': 0.13}) +[2026-05-30 22:11:10,593][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.0007686660392209888, 'train/grad_norm_pre_clip': 0.021827012300491333, 'timing/data': 0.0033255685120821, 'timing/model': 1.2892419919371605, 'learning_rate/qwen_vl_interface': 4.9760641449069724e-06, 'learning_rate/action_model': 4.976064144906972e-05, 'epoch': 0.13}) +[2026-05-30 22:14:20,991][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 5.350206993171014e-05, 'train/grad_norm_pre_clip': 0.0014448032015934587, 'timing/data': 0.003464864566922188, 'timing/model': 1.2773779761046171, 'learning_rate/qwen_vl_interface': 4.945666455091614e-06, 'learning_rate/action_model': 4.9456664550916136e-05, 'epoch': 0.13}) +[2026-05-30 22:17:33,993][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.21387441456317902, 'train/grad_norm_pre_clip': 3.0118699073791504, 'timing/data': 0.0056780800223350525, 'timing/model': 1.2786668930202723, 'learning_rate/qwen_vl_interface': 4.9152812752362935e-06, 'learning_rate/action_model': 4.915281275236293e-05, 'epoch': 0.13}) +[2026-05-30 22:20:46,858][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.06369123607873917, 'train/grad_norm_pre_clip': 0.9712479710578918, 'timing/data': 0.0031015630811452866, 'timing/model': 1.234823627397418, 'learning_rate/qwen_vl_interface': 4.8849098543567465e-06, 'learning_rate/action_model': 4.884909854356747e-05, 'epoch': 0.13}) +[2026-05-30 22:23:58,884][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.00015517135034315288, 'train/grad_norm_pre_clip': 0.005555447190999985, 'timing/data': 0.0025520604103803635, 'timing/model': 1.275361131876707, 'learning_rate/qwen_vl_interface': 4.854553440903127e-06, 'learning_rate/action_model': 4.854553440903126e-05, 'epoch': 0.13}) +[2026-05-30 22:27:13,293][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.016463853418827057, 'train/grad_norm_pre_clip': 1.5477428436279297, 'timing/data': 0.0031493250280618668, 'timing/model': 1.2805437240749598, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 0.13}) +[2026-05-30 22:30:49,171][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2700 +[2026-05-30 22:30:49,172][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.046220507472753525, 'train/grad_norm_pre_clip': 2.919578790664673, 'eval/loss': 0.03585882484912872, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0031552445143461227, 'timing/model': 1.4720423053950071, 'learning_rate/qwen_vl_interface': 4.793890626938511e-06, 'learning_rate/action_model': 4.793890626938511e-05, 'epoch': 0.13}) +[2026-05-30 22:30:49,173][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2700_state +[2026-05-30 22:30:49,177][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 22:30:54,831][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2700_state/model.safetensors +[2026-05-30 22:31:02,106][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2700_state/optimizer.bin +[2026-05-30 22:31:02,107][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2700_state/sampler.bin +[2026-05-30 22:31:02,107][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2700_state/sampler_1.bin +[2026-05-30 22:31:02,108][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2700_state/random_states_0.pkl +[2026-05-30 22:31:05,895][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2700 +[2026-05-30 22:31:05,895][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 22:31:05,907][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 22:34:19,592][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.013898764736950397, 'train/grad_norm_pre_clip': 0.4409659802913666, 'timing/data': 0.0033797547221183777, 'timing/model': 1.2913765329867601, 'learning_rate/qwen_vl_interface': 4.763586720038202e-06, 'learning_rate/action_model': 4.7635867200382014e-05, 'epoch': 0.13}) +[2026-05-30 22:37:32,607][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.1998356431722641, 'train/grad_norm_pre_clip': 3.4789271354675293, 'timing/data': 0.003486432135105133, 'timing/model': 1.2407598849385977, 'learning_rate/qwen_vl_interface': 4.733302807682679e-06, 'learning_rate/action_model': 4.733302807682679e-05, 'epoch': 0.13}) +[2026-05-30 22:40:45,588][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.06096450239419937, 'train/grad_norm_pre_clip': 0.6229389309883118, 'timing/data': 0.0030573680996894836, 'timing/model': 1.2781633529812098, 'learning_rate/qwen_vl_interface': 4.7030401347249685e-06, 'learning_rate/action_model': 4.703040134724968e-05, 'epoch': 0.13}) +[2026-05-30 22:44:00,287][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.1784820556640625, 'train/grad_norm_pre_clip': 1.835065245628357, 'timing/data': 0.0033824220299720764, 'timing/model': 1.2366878744214773, 'learning_rate/qwen_vl_interface': 4.6727999451450236e-06, 'learning_rate/action_model': 4.672799945145023e-05, 'epoch': 0.14}) +[2026-05-30 22:47:12,873][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.033751823008060455, 'train/grad_norm_pre_clip': 1.4633787870407104, 'timing/data': 0.003077855333685875, 'timing/model': 1.234283810481429, 'learning_rate/qwen_vl_interface': 4.642583481998596e-06, 'learning_rate/action_model': 4.642583481998596e-05, 'epoch': 0.14}) +[2026-05-30 22:50:26,793][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.10814591497182846, 'train/grad_norm_pre_clip': 1.8723033666610718, 'timing/data': 0.0033285338431596756, 'timing/model': 1.3004006836563349, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 0.14}) +[2026-05-30 22:53:39,894][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.20762649178504944, 'train/grad_norm_pre_clip': 9.661046981811523, 'timing/data': 0.003336722031235695, 'timing/model': 1.2799361106008291, 'learning_rate/qwen_vl_interface': 4.582226702301737e-06, 'learning_rate/action_model': 4.5822267023017375e-05, 'epoch': 0.14}) +[2026-05-30 22:56:54,186][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.08516692370176315, 'train/grad_norm_pre_clip': 7.114828586578369, 'timing/data': 0.003232991322875023, 'timing/model': 1.283064717426896, 'learning_rate/qwen_vl_interface': 4.552088866782115e-06, 'learning_rate/action_model': 4.552088866782115e-05, 'epoch': 0.14}) +[2026-05-30 23:00:05,900][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.00010925121023319662, 'train/grad_norm_pre_clip': 0.0010292379884049296, 'timing/data': 0.0031351465731859207, 'timing/model': 1.2645810414105654, 'learning_rate/qwen_vl_interface': 4.521979719655648e-06, 'learning_rate/action_model': 4.5219797196556484e-05, 'epoch': 0.14}) +[2026-05-30 23:04:03,003][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2800 +[2026-05-30 23:04:03,005][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.19212880730628967, 'train/grad_norm_pre_clip': 6.577437877655029, 'eval/loss': 0.04636529088020325, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 37.300000000000225, 'rl_games_eval/mid_train/mean_length': 321.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 37.300000000000225, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 321.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.00330263189971447, 'timing/model': 1.2854724284261465, 'learning_rate/qwen_vl_interface': 4.49190049859145e-06, 'learning_rate/action_model': 4.49190049859145e-05, 'epoch': 0.14}) +[2026-05-30 23:04:03,006][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2800_state +[2026-05-30 23:04:03,010][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 23:04:08,764][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2800_state/model.safetensors +[2026-05-30 23:04:16,165][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2800_state/optimizer.bin +[2026-05-30 23:04:16,166][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2800_state/sampler.bin +[2026-05-30 23:04:16,166][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2800_state/sampler_1.bin +[2026-05-30 23:04:16,167][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2800_state/random_states_0.pkl +[2026-05-30 23:04:19,900][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2800 +[2026-05-30 23:04:19,900][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 23:04:19,912][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 23:07:32,701][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.0007369262166321278, 'train/grad_norm_pre_clip': 0.022924641147255898, 'timing/data': 0.003789639100432396, 'timing/model': 1.294384067878127, 'learning_rate/qwen_vl_interface': 4.461852440028479e-06, 'learning_rate/action_model': 4.461852440028479e-05, 'epoch': 0.14}) +[2026-05-30 23:10:47,283][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.01691570319235325, 'train/grad_norm_pre_clip': 0.5304765105247498, 'timing/data': 0.003338772803544998, 'timing/model': 1.2781406082212925, 'learning_rate/qwen_vl_interface': 4.431836779124736e-06, 'learning_rate/action_model': 4.431836779124735e-05, 'epoch': 0.14}) +[2026-05-30 23:14:01,683][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.10182192176580429, 'train/grad_norm_pre_clip': 2.0918614864349365, 'timing/data': 0.003209242597222328, 'timing/model': 1.2697566635906696, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 0.14}) +[2026-05-30 23:17:15,550][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.44442662596702576, 'train/grad_norm_pre_clip': 5.317249774932861, 'timing/data': 0.0030422359704971313, 'timing/model': 1.2359369564801455, 'learning_rate/qwen_vl_interface': 4.371907584217499e-06, 'learning_rate/action_model': 4.371907584217498e-05, 'epoch': 0.14}) +[2026-05-30 23:20:29,594][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.0014505303697660565, 'train/grad_norm_pre_clip': 0.04841472581028938, 'timing/data': 0.0027174390852451324, 'timing/model': 1.2931835148483515, 'learning_rate/qwen_vl_interface': 4.341996513668483e-06, 'learning_rate/action_model': 4.3419965136684824e-05, 'epoch': 0.14}) +[2026-05-30 23:23:44,088][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.26868072152137756, 'train/grad_norm_pre_clip': 5.853130340576172, 'timing/data': 0.0036720577627420425, 'timing/model': 1.2849927358329296, 'learning_rate/qwen_vl_interface': 4.312122767586381e-06, 'learning_rate/action_model': 4.3121227675863815e-05, 'epoch': 0.14}) +[2026-05-30 23:26:56,821][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.049198269844055176, 'train/grad_norm_pre_clip': 0.569367527961731, 'timing/data': 0.0031861234456300735, 'timing/model': 1.2353422921150923, 'learning_rate/qwen_vl_interface': 4.2822875739638915e-06, 'learning_rate/action_model': 4.282287573963891e-05, 'epoch': 0.14}) +[2026-05-30 23:30:07,791][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.0871366634964943, 'train/grad_norm_pre_clip': 4.943029880523682, 'timing/data': 0.0032948721200227737, 'timing/model': 1.2834681570529938, 'learning_rate/qwen_vl_interface': 4.252492159208966e-06, 'learning_rate/action_model': 4.252492159208966e-05, 'epoch': 0.14}) +[2026-05-30 23:33:21,511][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.00441625714302063, 'train/grad_norm_pre_clip': 0.25209954380989075, 'timing/data': 0.0030042845755815506, 'timing/model': 1.2335971631109715, 'learning_rate/qwen_vl_interface': 4.222737748094404e-06, 'learning_rate/action_model': 4.222737748094404e-05, 'epoch': 0.14}) +[2026-05-30 23:39:43,287][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2900 +[2026-05-30 23:39:43,288][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 9.769519965630025e-05, 'train/grad_norm_pre_clip': 0.002310478361323476, 'eval/loss': 0.041193168610334396, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 255.89999999999208, 'rl_games_eval/mid_train/mean_length': 2084.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 255.89999999999208, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2084.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0032577142119407654, 'timing/model': 1.27781025134027, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 0.14}) +[2026-05-30 23:39:43,289][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2900_state +[2026-05-30 23:39:43,293][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-30 23:39:49,191][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2900_state/model.safetensors +[2026-05-30 23:39:56,697][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2900_state/optimizer.bin +[2026-05-30 23:39:56,698][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2900_state/sampler.bin +[2026-05-30 23:39:56,698][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2900_state/sampler_1.bin +[2026-05-30 23:39:56,699][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_2900_state/random_states_0.pkl +[2026-05-30 23:40:00,510][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2900 +[2026-05-30 23:40:00,510][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-30 23:40:00,522][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-30 23:43:12,695][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.05610838159918785, 'train/grad_norm_pre_clip': 6.228219032287598, 'timing/data': 0.00321376696228981, 'timing/model': 1.2954462859779596, 'learning_rate/qwen_vl_interface': 4.1633568273998e-06, 'learning_rate/action_model': 4.1633568273998e-05, 'epoch': 0.14}) +[2026-05-30 23:46:23,301][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.01791910082101822, 'train/grad_norm_pre_clip': 1.1856402158737183, 'timing/data': 0.003503980115056038, 'timing/model': 1.2347207441926003, 'learning_rate/qwen_vl_interface': 4.133732758736828e-06, 'learning_rate/action_model': 4.133732758736828e-05, 'epoch': 0.14}) +[2026-05-30 23:49:34,172][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.03318804129958153, 'train/grad_norm_pre_clip': 1.9959336519241333, 'timing/data': 0.003078306093811989, 'timing/model': 1.2341773696243763, 'learning_rate/qwen_vl_interface': 4.104154575448026e-06, 'learning_rate/action_model': 4.104154575448025e-05, 'epoch': 0.14}) +[2026-05-30 23:52:46,893][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.0010189111344516277, 'train/grad_norm_pre_clip': 0.008993587456643581, 'timing/data': 0.00307447649538517, 'timing/model': 1.2821506336331367, 'learning_rate/qwen_vl_interface': 4.074623493376657e-06, 'learning_rate/action_model': 4.074623493376657e-05, 'epoch': 0.15}) +[2026-05-30 23:55:55,967][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.0116091538220644, 'train/grad_norm_pre_clip': 0.24984276294708252, 'timing/data': 0.0034451503306627274, 'timing/model': 1.2348637077957392, 'learning_rate/qwen_vl_interface': 4.045140726429841e-06, 'learning_rate/action_model': 4.0451407264298406e-05, 'epoch': 0.15}) +[2026-05-30 23:59:10,051][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.15759636461734772, 'train/grad_norm_pre_clip': 2.6951398849487305, 'timing/data': 0.003124682232737541, 'timing/model': 1.234926288947463, 'learning_rate/qwen_vl_interface': 4.015707486528649e-06, 'learning_rate/action_model': 4.015707486528649e-05, 'epoch': 0.15}) +[2026-05-31 00:02:21,391][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.13108019530773163, 'train/grad_norm_pre_clip': 2.2068328857421875, 'timing/data': 0.0030473973602056503, 'timing/model': 1.2873436566442251, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 0.15}) +[2026-05-31 00:05:34,398][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.0028287896420806646, 'train/grad_norm_pre_clip': 0.11677522212266922, 'timing/data': 0.0037343166768550873, 'timing/model': 1.288785181939602, 'learning_rate/qwen_vl_interface': 3.956994425318387e-06, 'learning_rate/action_model': 3.956994425318387e-05, 'epoch': 0.15}) +[2026-05-31 00:08:47,797][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.003757969243451953, 'train/grad_norm_pre_clip': 0.15481014549732208, 'timing/data': 0.003501560539007187, 'timing/model': 1.2913807574659586, 'learning_rate/qwen_vl_interface': 3.927717017473301e-06, 'learning_rate/action_model': 3.9277170174733006e-05, 'epoch': 0.15}) +[2026-05-31 00:14:15,673][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3000 +[2026-05-31 00:14:15,674][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.05425842106342316, 'train/grad_norm_pre_clip': 0.8039440512657166, 'eval/loss': 0.04116647318005562, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 167.89999999999577, 'rl_games_eval/mid_train/mean_length': 1375.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 167.89999999999577, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1375.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0039051752537488937, 'timing/model': 1.2926516756415367, 'learning_rate/qwen_vl_interface': 3.8984939635025966e-06, 'learning_rate/action_model': 3.898493963502596e-05, 'epoch': 0.15}) +[2026-05-31 00:14:15,675][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3000_state +[2026-05-31 00:14:15,680][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 00:14:21,582][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3000_state/model.safetensors +[2026-05-31 00:14:28,848][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3000_state/optimizer.bin +[2026-05-31 00:14:28,850][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3000_state/sampler.bin +[2026-05-31 00:14:28,850][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3000_state/sampler_1.bin +[2026-05-31 00:14:28,851][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3000_state/random_states_0.pkl +[2026-05-31 00:14:32,623][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-05-31 00:14:32,624][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 00:14:32,636][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 00:17:45,893][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.026926979422569275, 'train/grad_norm_pre_clip': 1.6785160303115845, 'timing/data': 0.0030081849545240402, 'timing/model': 1.2931935656815767, 'learning_rate/qwen_vl_interface': 3.869326464651566e-06, 'learning_rate/action_model': 3.869326464651566e-05, 'epoch': 0.15}) +[2026-05-31 00:20:58,994][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.0021718430798500776, 'train/grad_norm_pre_clip': 0.1164761483669281, 'timing/data': 0.003716876730322838, 'timing/model': 1.282742341980338, 'learning_rate/qwen_vl_interface': 3.840215719881844e-06, 'learning_rate/action_model': 3.8402157198818445e-05, 'epoch': 0.15}) +[2026-05-31 00:24:12,992][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.01378164254128933, 'train/grad_norm_pre_clip': 0.4100722670555115, 'timing/data': 0.0032707098871469498, 'timing/model': 1.2819037791341543, 'learning_rate/qwen_vl_interface': 3.811162925822133e-06, 'learning_rate/action_model': 3.811162925822132e-05, 'epoch': 0.15}) +[2026-05-31 00:27:26,654][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.07231079787015915, 'train/grad_norm_pre_clip': 1.819854497909546, 'timing/data': 0.003577260300517082, 'timing/model': 1.5522432047873735, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 0.15}) +[2026-05-31 00:30:38,286][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.0009407683974131942, 'train/grad_norm_pre_clip': 0.022850774228572845, 'timing/data': 0.0032746735960245132, 'timing/model': 1.2746823355555534, 'learning_rate/qwen_vl_interface': 3.753235964387802e-06, 'learning_rate/action_model': 3.753235964387802e-05, 'epoch': 0.15}) +[2026-05-31 00:33:52,787][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.355390727519989, 'train/grad_norm_pre_clip': 3.7840516567230225, 'timing/data': 0.003365272656083107, 'timing/model': 1.2822489514946938, 'learning_rate/qwen_vl_interface': 3.7243641781636838e-06, 'learning_rate/action_model': 3.7243641781636836e-05, 'epoch': 0.15}) +[2026-05-31 00:37:02,034][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.01924314722418785, 'train/grad_norm_pre_clip': 0.6075026392936707, 'timing/data': 0.003164764493703842, 'timing/model': 1.2343057207763195, 'learning_rate/qwen_vl_interface': 3.695555104852695e-06, 'learning_rate/action_model': 3.6955551048526946e-05, 'epoch': 0.15}) +[2026-05-31 00:40:15,581][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.09782164543867111, 'train/grad_norm_pre_clip': 1.7218636274337769, 'timing/data': 0.0033007021993398666, 'timing/model': 1.2766273729503155, 'learning_rate/qwen_vl_interface': 3.6668099286830023e-06, 'learning_rate/action_model': 3.666809928683002e-05, 'epoch': 0.15}) +[2026-05-31 00:43:29,486][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.19775256514549255, 'train/grad_norm_pre_clip': 6.186910152435303, 'timing/data': 0.0033049508929252625, 'timing/model': 1.274104880169034, 'learning_rate/qwen_vl_interface': 3.6381298312562095e-06, 'learning_rate/action_model': 3.6381298312562094e-05, 'epoch': 0.15}) +[2026-05-31 00:52:00,857][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3100 +[2026-05-31 00:52:00,859][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.002145206555724144, 'train/grad_norm_pre_clip': 0.0876704528927803, 'eval/loss': 0.05128694325685501, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0033415593206882477, 'timing/model': 1.2875079661607742, 'learning_rate/qwen_vl_interface': 3.6095159914987893e-06, 'learning_rate/action_model': 3.6095159914987894e-05, 'epoch': 0.15}) +[2026-05-31 00:52:00,859][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3100_state +[2026-05-31 00:52:00,864][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 00:52:06,749][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3100_state/model.safetensors +[2026-05-31 00:52:13,927][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3100_state/optimizer.bin +[2026-05-31 00:52:13,928][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3100_state/sampler.bin +[2026-05-31 00:52:13,928][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3100_state/sampler_1.bin +[2026-05-31 00:52:13,929][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3100_state/random_states_0.pkl +[2026-05-31 00:52:17,694][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3100 +[2026-05-31 00:52:17,695][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 00:52:17,706][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 00:55:28,483][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.00282500172033906, 'train/grad_norm_pre_clip': 0.14444886147975922, 'timing/data': 0.0031795501708984375, 'timing/model': 1.2743250112980604, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 0.15}) +[2026-05-31 00:58:42,983][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.008129242807626724, 'train/grad_norm_pre_clip': 0.3034861981868744, 'timing/data': 0.003179643303155899, 'timing/model': 1.2866605501621962, 'learning_rate/qwen_vl_interface': 3.552491787031647e-06, 'learning_rate/action_model': 3.552491787031647e-05, 'epoch': 0.15}) +[2026-05-31 01:01:58,193][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.47172510623931885, 'train/grad_norm_pre_clip': 6.108152866363525, 'timing/data': 0.0033657774329185486, 'timing/model': 1.282989637926221, 'learning_rate/qwen_vl_interface': 3.5240837663636207e-06, 'learning_rate/action_model': 3.5240837663636204e-05, 'epoch': 0.15}) +[2026-05-31 01:05:10,465][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.06759241223335266, 'train/grad_norm_pre_clip': 1.5201826095581055, 'timing/data': 0.0033573396503925323, 'timing/model': 1.2347462251782417, 'learning_rate/qwen_vl_interface': 3.495746691352008e-06, 'learning_rate/action_model': 3.495746691352008e-05, 'epoch': 0.16}) +[2026-05-31 01:08:19,907][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.22349989414215088, 'train/grad_norm_pre_clip': 2.6385607719421387, 'timing/data': 0.0032745935022830963, 'timing/model': 1.234282124787569, 'learning_rate/qwen_vl_interface': 3.4674817268229732e-06, 'learning_rate/action_model': 3.467481726822973e-05, 'epoch': 0.16}) +[2026-05-31 01:11:32,091][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.060902804136276245, 'train/grad_norm_pre_clip': 1.4914309978485107, 'timing/data': 0.0033709127455949783, 'timing/model': 1.2802934367209673, 'learning_rate/qwen_vl_interface': 3.4392900346385022e-06, 'learning_rate/action_model': 3.4392900346385024e-05, 'epoch': 0.16}) +[2026-05-31 01:14:44,894][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.2674143612384796, 'train/grad_norm_pre_clip': 3.4314253330230713, 'timing/data': 0.0030486322939395905, 'timing/model': 1.2946416027843952, 'learning_rate/qwen_vl_interface': 3.411172773648642e-06, 'learning_rate/action_model': 3.411172773648642e-05, 'epoch': 0.16}) +[2026-05-31 01:17:59,496][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.0020382318180054426, 'train/grad_norm_pre_clip': 0.07483929395675659, 'timing/data': 0.003112412989139557, 'timing/model': 1.298879386857152, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 0.16}) +[2026-05-31 01:21:11,178][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.0006010964279994369, 'train/grad_norm_pre_clip': 0.011405845172703266, 'timing/data': 0.0034702178090810776, 'timing/model': 1.2775720991194248, 'learning_rate/qwen_vl_interface': 3.3551661653075522e-06, 'learning_rate/action_model': 3.355166165307552e-05, 'epoch': 0.16}) +[2026-05-31 01:24:47,562][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3200 +[2026-05-31 01:24:47,564][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.004059117287397385, 'train/grad_norm_pre_clip': 0.1414605975151062, 'eval/loss': 0.03549091890454292, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0032207034528255463, 'timing/model': 1.251786034554243, 'learning_rate/qwen_vl_interface': 3.3272791201686294e-06, 'learning_rate/action_model': 3.327279120168629e-05, 'epoch': 0.16}) +[2026-05-31 01:24:47,565][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3200_state +[2026-05-31 01:24:47,569][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 01:24:53,261][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3200_state/model.safetensors +[2026-05-31 01:25:00,499][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3200_state/optimizer.bin +[2026-05-31 01:25:00,499][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3200_state/sampler.bin +[2026-05-31 01:25:00,500][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3200_state/sampler_1.bin +[2026-05-31 01:25:00,501][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3200_state/random_states_0.pkl +[2026-05-31 01:25:04,346][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3200 +[2026-05-31 01:25:04,347][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 01:25:04,359][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 01:28:15,184][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.1467619389295578, 'train/grad_norm_pre_clip': 3.4274709224700928, 'timing/data': 0.003231937065720558, 'timing/model': 1.278764208778739, 'learning_rate/qwen_vl_interface': 3.299471110554295e-06, 'learning_rate/action_model': 3.299471110554295e-05, 'epoch': 0.16}) +[2026-05-31 01:31:28,992][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.06568934768438339, 'train/grad_norm_pre_clip': 4.641037940979004, 'timing/data': 0.0033949557691812515, 'timing/model': 1.283138494938612, 'learning_rate/qwen_vl_interface': 3.2717432795429057e-06, 'learning_rate/action_model': 3.271743279542905e-05, 'epoch': 0.16}) +[2026-05-31 01:34:42,788][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.01938118413090706, 'train/grad_norm_pre_clip': 1.7162829637527466, 'timing/data': 0.0032759439200162888, 'timing/model': 1.2846752367913723, 'learning_rate/qwen_vl_interface': 3.244096766916992e-06, 'learning_rate/action_model': 3.244096766916992e-05, 'epoch': 0.16}) +[2026-05-31 01:37:54,386][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.5395413637161255, 'train/grad_norm_pre_clip': 3.696730852127075, 'timing/data': 0.0031773168593645096, 'timing/model': 1.2481599040329456, 'learning_rate/qwen_vl_interface': 3.216532709116405e-06, 'learning_rate/action_model': 3.216532709116405e-05, 'epoch': 0.16}) +[2026-05-31 01:41:07,900][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.024432187899947166, 'train/grad_norm_pre_clip': 0.8339208364486694, 'timing/data': 0.003286164253950119, 'timing/model': 1.2955572754144669, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 0.16}) +[2026-05-31 01:44:22,531][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 9.172051795758307e-05, 'train/grad_norm_pre_clip': 0.002025999827310443, 'timing/data': 0.0033278726041316986, 'timing/model': 1.5321156028658152, 'learning_rate/qwen_vl_interface': 3.1616564867570603e-06, 'learning_rate/action_model': 3.16165648675706e-05, 'epoch': 0.16}) +[2026-05-31 01:47:35,184][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.15574327111244202, 'train/grad_norm_pre_clip': 5.092308521270752, 'timing/data': 0.003222225233912468, 'timing/model': 1.282543208450079, 'learning_rate/qwen_vl_interface': 3.134346577944869e-06, 'learning_rate/action_model': 3.134346577944869e-05, 'epoch': 0.16}) +[2026-05-31 01:50:46,284][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.00033193815033882856, 'train/grad_norm_pre_clip': 0.015993142500519753, 'timing/data': 0.0033111851662397385, 'timing/model': 1.2842138018459082, 'learning_rate/qwen_vl_interface': 3.10712363535841e-06, 'learning_rate/action_model': 3.10712363535841e-05, 'epoch': 0.16}) +[2026-05-31 01:54:01,286][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.020759057253599167, 'train/grad_norm_pre_clip': 2.41110897064209, 'timing/data': 0.0034189317375421524, 'timing/model': 1.284324623644352, 'learning_rate/qwen_vl_interface': 3.0799887780262295e-06, 'learning_rate/action_model': 3.0799887780262296e-05, 'epoch': 0.16}) +[2026-05-31 01:57:36,305][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3300 +[2026-05-31 01:57:36,307][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.050131600350141525, 'train/grad_norm_pre_clip': 2.3743295669555664, 'eval/loss': 0.04050270467996597, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0032149814069271088, 'timing/model': 1.297540795058012, 'learning_rate/qwen_vl_interface': 3.0529431213560334e-06, 'learning_rate/action_model': 3.0529431213560334e-05, 'epoch': 0.16}) +[2026-05-31 01:57:36,307][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3300_state +[2026-05-31 01:57:36,312][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 01:57:41,963][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3300_state/model.safetensors +[2026-05-31 01:57:49,213][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3300_state/optimizer.bin +[2026-05-31 01:57:49,213][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3300_state/sampler.bin +[2026-05-31 01:57:49,213][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3300_state/sampler_1.bin +[2026-05-31 01:57:49,214][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3300_state/random_states_0.pkl +[2026-05-31 01:57:53,013][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3300 +[2026-05-31 01:57:53,013][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 01:57:53,025][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 02:01:04,788][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.0007288070046342909, 'train/grad_norm_pre_clip': 0.05383246764540672, 'timing/data': 0.003664875403046608, 'timing/model': 1.2751114517450333, 'learning_rate/qwen_vl_interface': 3.0259877770888334e-06, 'learning_rate/action_model': 3.0259877770888333e-05, 'epoch': 0.16}) +[2026-05-31 02:04:19,285][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.1220262199640274, 'train/grad_norm_pre_clip': 1.8238166570663452, 'timing/data': 0.003306189551949501, 'timing/model': 1.2841882035136223, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 0.16}) +[2026-05-31 02:07:34,286][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 5.337479888112284e-05, 'train/grad_norm_pre_clip': 0.0015914710238575935, 'timing/data': 0.0031042415648698807, 'timing/model': 1.2846460826694965, 'learning_rate/qwen_vl_interface': 2.9723524541199943e-06, 'learning_rate/action_model': 2.972352454119994e-05, 'epoch': 0.16}) +[2026-05-31 02:10:48,787][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.0003418388369027525, 'train/grad_norm_pre_clip': 0.008687365800142288, 'timing/data': 0.003205910325050354, 'timing/model': 1.2836124189198017, 'learning_rate/qwen_vl_interface': 2.9456746801564072e-06, 'learning_rate/action_model': 2.945674680156407e-05, 'epoch': 0.16}) +[2026-05-31 02:14:03,593][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.004068669863045216, 'train/grad_norm_pre_clip': 0.15981148183345795, 'timing/data': 0.0030195824801921844, 'timing/model': 1.2887675017118454, 'learning_rate/qwen_vl_interface': 2.919091627981297e-06, 'learning_rate/action_model': 2.919091627981297e-05, 'epoch': 0.17}) +[2026-05-31 02:17:16,992][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.33202093839645386, 'train/grad_norm_pre_clip': 3.376004695892334, 'timing/data': 0.0031534936279058456, 'timing/model': 1.2894076742231846, 'learning_rate/qwen_vl_interface': 2.8926043903198135e-06, 'learning_rate/action_model': 2.8926043903198135e-05, 'epoch': 0.17}) +[2026-05-31 02:20:30,592][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.0886063277721405, 'train/grad_norm_pre_clip': 2.692068576812744, 'timing/data': 0.0030521750450134277, 'timing/model': 1.2903369050472975, 'learning_rate/qwen_vl_interface': 2.8662140559585584e-06, 'learning_rate/action_model': 2.866214055958558e-05, 'epoch': 0.17}) +[2026-05-31 02:23:44,788][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.041214339435100555, 'train/grad_norm_pre_clip': 2.057739734649658, 'timing/data': 0.003162764012813568, 'timing/model': 1.2798358034342527, 'learning_rate/qwen_vl_interface': 2.83992170970081e-06, 'learning_rate/action_model': 2.8399217097008102e-05, 'epoch': 0.17}) +[2026-05-31 02:26:58,181][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.023786421865224838, 'train/grad_norm_pre_clip': 1.3200931549072266, 'timing/data': 0.0030177030712366104, 'timing/model': 1.278625175356865, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 0.17}) +[2026-05-31 02:34:31,170][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3400 +[2026-05-31 02:34:31,171][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.00025563445524312556, 'train/grad_norm_pre_clip': 0.034428827464580536, 'eval/loss': 0.03762650489807129, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 358.4000000000104, 'rl_games_eval/mid_train/mean_length': 2911.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 358.4000000000104, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2911.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0031305328011512756, 'timing/model': 1.2934521939605474, 'learning_rate/qwen_vl_interface': 2.787635300525006e-06, 'learning_rate/action_model': 2.787635300525006e-05, 'epoch': 0.17}) +[2026-05-31 02:34:31,172][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3400_state +[2026-05-31 02:34:31,176][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 02:34:36,822][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3400_state/model.safetensors +[2026-05-31 02:34:44,031][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3400_state/optimizer.bin +[2026-05-31 02:34:44,032][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3400_state/sampler.bin +[2026-05-31 02:34:44,032][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3400_state/sampler_1.bin +[2026-05-31 02:34:44,033][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3400_state/random_states_0.pkl +[2026-05-31 02:34:47,828][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3400 +[2026-05-31 02:34:47,829][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 02:34:47,840][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 02:38:00,395][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.27792370319366455, 'train/grad_norm_pre_clip': 6.487627983093262, 'timing/data': 0.003847885876893997, 'timing/model': 1.282671520486474, 'learning_rate/qwen_vl_interface': 2.7616433868964387e-06, 'learning_rate/action_model': 2.7616433868964388e-05, 'epoch': 0.17}) +[2026-05-31 02:41:13,793][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.3306565284729004, 'train/grad_norm_pre_clip': 12.658127784729004, 'timing/data': 0.00034291669726371765, 'timing/model': 1.2389070019125938, 'learning_rate/qwen_vl_interface': 2.73575375986201e-06, 'learning_rate/action_model': 2.7357537598620096e-05, 'epoch': 0.17}) +[2026-05-31 02:44:27,106][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.00022136890038382262, 'train/grad_norm_pre_clip': 0.022651998326182365, 'timing/data': 0.00033477693796157837, 'timing/model': 1.2375173978507519, 'learning_rate/qwen_vl_interface': 2.7099674836428842e-06, 'learning_rate/action_model': 2.709967483642884e-05, 'epoch': 0.17}) +[2026-05-31 02:47:39,011][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.019111555069684982, 'train/grad_norm_pre_clip': 4.827051162719727, 'timing/data': 0.0003424156457185745, 'timing/model': 1.2455063108354807, 'learning_rate/qwen_vl_interface': 2.68428561821188e-06, 'learning_rate/action_model': 2.6842856182118802e-05, 'epoch': 0.17}) +[2026-05-31 02:50:52,478][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.0033604148775339127, 'train/grad_norm_pre_clip': 0.37936869263648987, 'timing/data': 0.00042547471821308136, 'timing/model': 1.2345076631754637, 'learning_rate/qwen_vl_interface': 2.6587092192498943e-06, 'learning_rate/action_model': 2.658709219249894e-05, 'epoch': 0.17}) +[2026-05-31 02:54:05,792][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.00044799811439588666, 'train/grad_norm_pre_clip': 0.013546662405133247, 'timing/data': 0.0003557950258255005, 'timing/model': 1.2881440203636885, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 0.17}) +[2026-05-31 02:57:17,947][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.004674751777201891, 'train/grad_norm_pre_clip': 0.2689853310585022, 'timing/data': 0.0005606915801763535, 'timing/model': 1.2352329306304455, 'learning_rate/qwen_vl_interface': 2.607877021736793e-06, 'learning_rate/action_model': 2.6078770217367926e-05, 'epoch': 0.17}) +[2026-05-31 03:00:31,447][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.2562171518802643, 'train/grad_norm_pre_clip': 6.830685615539551, 'timing/data': 0.0003309566527605057, 'timing/model': 1.234578125178814, 'learning_rate/qwen_vl_interface': 2.5826233126982198e-06, 'learning_rate/action_model': 2.5826233126982196e-05, 'epoch': 0.17}) +[2026-05-31 03:03:42,292][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.11588339507579803, 'train/grad_norm_pre_clip': 4.929179668426514, 'timing/data': 0.00035046786069869995, 'timing/model': 1.285167258232832, 'learning_rate/qwen_vl_interface': 2.5574792490678648e-06, 'learning_rate/action_model': 2.5574792490678644e-05, 'epoch': 0.17}) +[2026-05-31 03:12:15,734][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3500 +[2026-05-31 03:12:15,736][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.011932984925806522, 'train/grad_norm_pre_clip': 0.625617265701294, 'eval/loss': 0.04428589344024658, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.00034249573945999146, 'timing/model': 1.2936282027512789, 'learning_rate/qwen_vl_interface': 2.532445864419695e-06, 'learning_rate/action_model': 2.5324458644196952e-05, 'epoch': 0.17}) +[2026-05-31 03:12:15,736][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3500_state +[2026-05-31 03:12:15,740][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 03:12:21,452][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3500_state/model.safetensors +[2026-05-31 03:12:28,600][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3500_state/optimizer.bin +[2026-05-31 03:12:28,601][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3500_state/sampler.bin +[2026-05-31 03:12:28,601][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3500_state/sampler_1.bin +[2026-05-31 03:12:28,602][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3500_state/random_states_0.pkl +[2026-05-31 03:12:32,325][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-05-31 03:12:32,325][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 03:12:32,337][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 03:15:44,195][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.003100774949416518, 'train/grad_norm_pre_clip': 0.16676034033298492, 'timing/data': 0.0003335345536470413, 'timing/model': 1.237576087936759, 'learning_rate/qwen_vl_interface': 2.5075241877781085e-06, 'learning_rate/action_model': 2.5075241877781086e-05, 'epoch': 0.17}) +[2026-05-31 03:18:58,696][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.0018957331776618958, 'train/grad_norm_pre_clip': 0.2019454687833786, 'timing/data': 0.00036323443055152893, 'timing/model': 1.2852928787469864, 'learning_rate/qwen_vl_interface': 2.4827152435756185e-06, 'learning_rate/action_model': 2.4827152435756185e-05, 'epoch': 0.17}) +[2026-05-31 03:22:10,794][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.004471357446163893, 'train/grad_norm_pre_clip': 0.19621288776397705, 'timing/data': 0.0003256089985370636, 'timing/model': 1.2914653439074755, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 0.17}) +[2026-05-31 03:25:25,283][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.03228256106376648, 'train/grad_norm_pre_clip': 2.66752290725708, 'timing/data': 0.0003341957926750183, 'timing/model': 1.2755213119089603, 'learning_rate/qwen_vl_interface': 2.433439627006128e-06, 'learning_rate/action_model': 2.4334396270061277e-05, 'epoch': 0.17}) +[2026-05-31 03:28:36,082][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.02335161343216896, 'train/grad_norm_pre_clip': 1.3433938026428223, 'timing/data': 0.0003498755395412445, 'timing/model': 1.2822508085519075, 'learning_rate/qwen_vl_interface': 2.408974980166724e-06, 'learning_rate/action_model': 2.408974980166724e-05, 'epoch': 0.18}) +[2026-05-31 03:31:50,488][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.05972597375512123, 'train/grad_norm_pre_clip': 1.6049420833587646, 'timing/data': 0.0003352966159582138, 'timing/model': 1.2792494632303715, 'learning_rate/qwen_vl_interface': 2.3846271167383515e-06, 'learning_rate/action_model': 2.3846271167383515e-05, 'epoch': 0.18}) +[2026-05-31 03:35:05,192][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.0017445405246689916, 'train/grad_norm_pre_clip': 0.09529117494821548, 'timing/data': 0.00032862648367881775, 'timing/model': 1.2455869764089584, 'learning_rate/qwen_vl_interface': 2.3603970375663123e-06, 'learning_rate/action_model': 2.3603970375663125e-05, 'epoch': 0.18}) +[2026-05-31 03:38:17,781][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.09890107810497284, 'train/grad_norm_pre_clip': 2.6393027305603027, 'timing/data': 0.0004506949335336685, 'timing/model': 1.2537651006132364, 'learning_rate/qwen_vl_interface': 2.33628573865426e-06, 'learning_rate/action_model': 2.3362857386542596e-05, 'epoch': 0.18}) +[2026-05-31 03:41:29,587][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.11885096877813339, 'train/grad_norm_pre_clip': 4.767499923706055, 'timing/data': 0.00033750757575035095, 'timing/model': 1.2804031670093536, 'learning_rate/qwen_vl_interface': 2.3122942111232506e-06, 'learning_rate/action_model': 2.3122942111232506e-05, 'epoch': 0.18}) +[2026-05-31 03:47:45,241][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3600 +[2026-05-31 03:47:45,243][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.16106238961219788, 'train/grad_norm_pre_clip': 8.295960426330566, 'eval/loss': 0.02681850455701351, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 241.89999999999247, 'rl_games_eval/mid_train/mean_length': 1971.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 241.89999999999247, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1971.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0003622770309448242, 'timing/model': 1.2620854265987873, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 0.18}) +[2026-05-31 03:47:45,243][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3600_state +[2026-05-31 03:47:45,247][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 03:47:51,259][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3600_state/model.safetensors +[2026-05-31 03:47:58,569][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3600_state/optimizer.bin +[2026-05-31 03:47:58,570][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3600_state/sampler.bin +[2026-05-31 03:47:58,570][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3600_state/sampler_1.bin +[2026-05-31 03:47:58,571][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3600_state/random_states_0.pkl +[2026-05-31 03:48:02,374][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3600 +[2026-05-31 03:48:02,375][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 03:48:02,386][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 03:51:16,893][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.002752338768914342, 'train/grad_norm_pre_clip': 0.1516091376543045, 'timing/data': 0.00033658742904663086, 'timing/model': 1.2954151686280966, 'learning_rate/qwen_vl_interface': 2.264674410031413e-06, 'learning_rate/action_model': 2.2646744100314127e-05, 'epoch': 0.18}) +[2026-05-31 03:54:31,286][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.002092029433697462, 'train/grad_norm_pre_clip': 0.2514340281486511, 'timing/data': 0.0003394056111574173, 'timing/model': 1.2937112245708704, 'learning_rate/qwen_vl_interface': 2.2410480939340945e-06, 'learning_rate/action_model': 2.2410480939340945e-05, 'epoch': 0.18}) +[2026-05-31 03:57:43,642][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.19858907163143158, 'train/grad_norm_pre_clip': 7.719230651855469, 'timing/data': 0.00033584795892238617, 'timing/model': 1.2335961237549782, 'learning_rate/qwen_vl_interface': 2.217545464064374e-06, 'learning_rate/action_model': 2.217545464064374e-05, 'epoch': 0.18}) +[2026-05-31 04:00:55,486][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.2735249102115631, 'train/grad_norm_pre_clip': 8.87346076965332, 'timing/data': 0.0003450550138950348, 'timing/model': 1.283325219526887, 'learning_rate/qwen_vl_interface': 2.1941674865233144e-06, 'learning_rate/action_model': 2.1941674865233143e-05, 'epoch': 0.18}) +[2026-05-31 04:04:09,790][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 5.67306560697034e-05, 'train/grad_norm_pre_clip': 0.0008239643066190183, 'timing/data': 0.0003453157842159271, 'timing/model': 1.2916983347386122, 'learning_rate/qwen_vl_interface': 2.170915122288006e-06, 'learning_rate/action_model': 2.170915122288006e-05, 'epoch': 0.18}) +[2026-05-31 04:07:24,194][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.29636436700820923, 'train/grad_norm_pre_clip': 4.6479811668396, 'timing/data': 0.00033613480627536774, 'timing/model': 1.278686173260212, 'learning_rate/qwen_vl_interface': 2.1477893271720683e-06, 'learning_rate/action_model': 2.1477893271720682e-05, 'epoch': 0.18}) +[2026-05-31 04:10:38,892][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.01344749704003334, 'train/grad_norm_pre_clip': 1.651901125907898, 'timing/data': 0.00034083612263202667, 'timing/model': 1.29305150359869, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 0.18}) +[2026-05-31 04:13:53,483][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.0030292384326457977, 'train/grad_norm_pre_clip': 0.08973757922649384, 'timing/data': 0.0003376975655555725, 'timing/model': 1.2892460133880377, 'learning_rate/qwen_vl_interface': 2.1019212414999163e-06, 'learning_rate/action_model': 2.101921241499916e-05, 'epoch': 0.18}) +[2026-05-31 04:17:08,288][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.21122558414936066, 'train/grad_norm_pre_clip': 4.883179187774658, 'timing/data': 0.0003346763551235199, 'timing/model': 1.278727451339364, 'learning_rate/qwen_vl_interface': 2.079180836401047e-06, 'learning_rate/action_model': 2.079180836401047e-05, 'epoch': 0.18}) +[2026-05-31 04:20:54,566][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3700 +[2026-05-31 04:20:54,568][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.23837414383888245, 'train/grad_norm_pre_clip': 3.104844093322754, 'eval/loss': 0.02307944931089878, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 20.500000000000032, 'rl_games_eval/mid_train/mean_length': 180.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 20.500000000000032, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 180.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.00040237605571746826, 'timing/model': 1.2332687750458717, 'learning_rate/qwen_vl_interface': 2.0565707712587447e-06, 'learning_rate/action_model': 2.0565707712587444e-05, 'epoch': 0.18}) +[2026-05-31 04:20:54,568][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3700_state +[2026-05-31 04:20:54,573][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 04:21:00,585][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3700_state/model.safetensors +[2026-05-31 04:21:07,864][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3700_state/optimizer.bin +[2026-05-31 04:21:07,865][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3700_state/sampler.bin +[2026-05-31 04:21:07,865][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3700_state/sampler_1.bin +[2026-05-31 04:21:07,866][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3700_state/random_states_0.pkl +[2026-05-31 04:21:11,686][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3700 +[2026-05-31 04:21:11,687][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 04:21:11,699][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 04:24:23,693][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.001158802886493504, 'train/grad_norm_pre_clip': 0.04150862246751785, 'timing/data': 0.00042351335287094116, 'timing/model': 1.286229519173503, 'learning_rate/qwen_vl_interface': 2.034091975484231e-06, 'learning_rate/action_model': 2.0340919754842307e-05, 'epoch': 0.18}) +[2026-05-31 04:27:36,798][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.018320197239518166, 'train/grad_norm_pre_clip': 2.7416720390319824, 'timing/data': 0.00034582801163196564, 'timing/model': 1.2979455944150686, 'learning_rate/qwen_vl_interface': 2.011745373092753e-06, 'learning_rate/action_model': 2.011745373092753e-05, 'epoch': 0.18}) +[2026-05-31 04:30:50,194][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.0004946677363477647, 'train/grad_norm_pre_clip': 0.024506470188498497, 'timing/data': 0.0003497954457998276, 'timing/model': 1.2858637068420649, 'learning_rate/qwen_vl_interface': 1.9895318826656116e-06, 'learning_rate/action_model': 1.9895318826656114e-05, 'epoch': 0.18}) +[2026-05-31 04:34:04,882][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.21932627260684967, 'train/grad_norm_pre_clip': 3.1277894973754883, 'timing/data': 0.0003449041396379471, 'timing/model': 1.281455708667636, 'learning_rate/qwen_vl_interface': 1.967452417312393e-06, 'learning_rate/action_model': 1.9674524173123927e-05, 'epoch': 0.18}) +[2026-05-31 04:37:19,493][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.6084816455841064, 'train/grad_norm_pre_clip': 3.154763698577881, 'timing/data': 0.00033480487763881683, 'timing/model': 1.2932736277580261, 'learning_rate/qwen_vl_interface': 1.9455078846334397e-06, 'learning_rate/action_model': 1.9455078846334395e-05, 'epoch': 0.19}) +[2026-05-31 04:40:32,286][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.017331119626760483, 'train/grad_norm_pre_clip': 0.2861914038658142, 'timing/data': 0.00041855499148368835, 'timing/model': 1.2881844751536846, 'learning_rate/qwen_vl_interface': 1.92369918668254e-06, 'learning_rate/action_model': 1.92369918668254e-05, 'epoch': 0.19}) +[2026-05-31 04:43:43,292][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.029612598940730095, 'train/grad_norm_pre_clip': 1.4376815557479858, 'timing/data': 0.0003394670784473419, 'timing/model': 1.2900920771062374, 'learning_rate/qwen_vl_interface': 1.9020272199298452e-06, 'learning_rate/action_model': 1.902027219929845e-05, 'epoch': 0.19}) +[2026-05-31 04:46:56,865][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.003893723478540778, 'train/grad_norm_pre_clip': 0.15574708580970764, 'timing/data': 0.00034551508724689484, 'timing/model': 1.252474531531334, 'learning_rate/qwen_vl_interface': 1.880492875225025e-06, 'learning_rate/action_model': 1.8804928752250252e-05, 'epoch': 0.19}) +[2026-05-31 04:50:07,892][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.038230281323194504, 'train/grad_norm_pre_clip': 1.559469223022461, 'timing/data': 0.00034161657094955444, 'timing/model': 1.2835801858454943, 'learning_rate/qwen_vl_interface': 1.8590970377606494e-06, 'learning_rate/action_model': 1.859097037760649e-05, 'epoch': 0.19}) +[2026-05-31 04:53:46,588][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3800 +[2026-05-31 04:53:46,590][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.06798948347568512, 'train/grad_norm_pre_clip': 1.4871662855148315, 'eval/loss': 0.026137281209230423, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.00033901631832122803, 'timing/model': 1.2905666623264551, 'learning_rate/qwen_vl_interface': 1.8378405870357943e-06, 'learning_rate/action_model': 1.837840587035794e-05, 'epoch': 0.19}) +[2026-05-31 04:53:46,590][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3800_state +[2026-05-31 04:53:46,595][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 04:53:52,679][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3800_state/model.safetensors +[2026-05-31 04:53:59,982][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3800_state/optimizer.bin +[2026-05-31 04:53:59,983][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3800_state/sampler.bin +[2026-05-31 04:53:59,983][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3800_state/sampler_1.bin +[2026-05-31 04:53:59,984][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3800_state/random_states_0.pkl +[2026-05-31 04:54:03,795][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3800 +[2026-05-31 04:54:03,796][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 04:54:03,807][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 04:57:14,694][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.028035884723067284, 'train/grad_norm_pre_clip': 3.860549211502075, 'timing/data': 0.00034071505069732666, 'timing/model': 1.2955280914902687, 'learning_rate/qwen_vl_interface': 1.8167243968198905e-06, 'learning_rate/action_model': 1.8167243968198905e-05, 'epoch': 0.19}) +[2026-05-31 05:00:27,197][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.20564505457878113, 'train/grad_norm_pre_clip': 3.3223953247070312, 'timing/data': 0.0003306642174720764, 'timing/model': 1.2544647715985775, 'learning_rate/qwen_vl_interface': 1.7957493351168126e-06, 'learning_rate/action_model': 1.7957493351168126e-05, 'epoch': 0.19}) +[2026-05-31 05:03:41,792][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.06082377955317497, 'train/grad_norm_pre_clip': 1.8045308589935303, 'timing/data': 0.0003394484519958496, 'timing/model': 1.2857129275798798, 'learning_rate/qwen_vl_interface': 1.7749162641291933e-06, 'learning_rate/action_model': 1.7749162641291933e-05, 'epoch': 0.19}) +[2026-05-31 05:06:56,380][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.09092932194471359, 'train/grad_norm_pre_clip': 1.885342001914978, 'timing/data': 0.00034205615520477295, 'timing/model': 1.271408997476101, 'learning_rate/qwen_vl_interface': 1.7542260402229838e-06, 'learning_rate/action_model': 1.7542260402229838e-05, 'epoch': 0.19}) +[2026-05-31 05:10:08,844][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.04222733527421951, 'train/grad_norm_pre_clip': 2.5369858741760254, 'timing/data': 0.0004077870398759842, 'timing/model': 1.2353690639138222, 'learning_rate/qwen_vl_interface': 1.733679513892252e-06, 'learning_rate/action_model': 1.7336795138922517e-05, 'epoch': 0.19}) +[2026-05-31 05:13:22,893][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.14998933672904968, 'train/grad_norm_pre_clip': 7.71403694152832, 'timing/data': 0.0003330148756504059, 'timing/model': 1.2921106312423944, 'learning_rate/qwen_vl_interface': 1.7132775297242175e-06, 'learning_rate/action_model': 1.7132775297242172e-05, 'epoch': 0.19}) +[2026-05-31 05:16:36,885][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.016971074044704437, 'train/grad_norm_pre_clip': 0.7140791416168213, 'timing/data': 0.0003270171582698822, 'timing/model': 1.278037903830409, 'learning_rate/qwen_vl_interface': 1.6930209263645412e-06, 'learning_rate/action_model': 1.6930209263645412e-05, 'epoch': 0.19}) +[2026-05-31 05:19:50,000][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.002077660523355007, 'train/grad_norm_pre_clip': 0.07807756215333939, 'timing/data': 0.0003349967300891876, 'timing/model': 1.234290424734354, 'learning_rate/qwen_vl_interface': 1.6729105364828486e-06, 'learning_rate/action_model': 1.6729105364828484e-05, 'epoch': 0.19}) +[2026-05-31 05:23:00,037][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.26358741521835327, 'train/grad_norm_pre_clip': 4.987954616546631, 'timing/data': 0.0003412757068872452, 'timing/model': 1.4389613587409258, 'learning_rate/qwen_vl_interface': 1.6529471867385009e-06, 'learning_rate/action_model': 1.6529471867385007e-05, 'epoch': 0.19}) +[2026-05-31 05:27:42,371][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3900 +[2026-05-31 05:27:42,373][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.06013345345854759, 'train/grad_norm_pre_clip': 5.737920761108398, 'eval/loss': 0.022782260552048683, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 106.39999999999854, 'rl_games_eval/mid_train/mean_length': 877.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 106.39999999999854, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 877.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0003256164491176605, 'timing/model': 1.2838449608534575, 'learning_rate/qwen_vl_interface': 1.6331316977466117e-06, 'learning_rate/action_model': 1.6331316977466115e-05, 'epoch': 0.19}) +[2026-05-31 05:27:42,374][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3900_state +[2026-05-31 05:27:42,378][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 05:27:48,459][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3900_state/model.safetensors +[2026-05-31 05:27:55,610][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3900_state/optimizer.bin +[2026-05-31 05:27:55,611][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3900_state/sampler.bin +[2026-05-31 05:27:55,611][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3900_state/sampler_1.bin +[2026-05-31 05:27:55,612][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_3900_state/random_states_0.pkl +[2026-05-31 05:27:59,337][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3900 +[2026-05-31 05:27:59,338][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 05:27:59,350][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 05:31:13,893][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.05550691485404968, 'train/grad_norm_pre_clip': 3.331740379333496, 'timing/data': 0.0003386363387107849, 'timing/model': 1.2887850515544415, 'learning_rate/qwen_vl_interface': 1.6134648840443248e-06, 'learning_rate/action_model': 1.613464884044325e-05, 'epoch': 0.19}) +[2026-05-31 05:34:27,983][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.003083879593759775, 'train/grad_norm_pre_clip': 0.19273503124713898, 'timing/data': 0.0003409832715988159, 'timing/model': 1.2750119548290968, 'learning_rate/qwen_vl_interface': 1.5939475540573213e-06, 'learning_rate/action_model': 1.5939475540573212e-05, 'epoch': 0.19}) +[2026-05-31 05:37:41,493][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.16633149981498718, 'train/grad_norm_pre_clip': 2.382868766784668, 'timing/data': 0.00034854374825954437, 'timing/model': 1.2851069401949644, 'learning_rate/qwen_vl_interface': 1.5745805100665922e-06, 'learning_rate/action_model': 1.5745805100665923e-05, 'epoch': 0.19}) +[2026-05-31 05:40:52,491][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.3876102566719055, 'train/grad_norm_pre_clip': 3.2297823429107666, 'timing/data': 0.00033543631434440613, 'timing/model': 1.2832164019346237, 'learning_rate/qwen_vl_interface': 1.5553645481754626e-06, 'learning_rate/action_model': 1.5553645481754626e-05, 'epoch': 0.19}) +[2026-05-31 05:44:04,580][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.16497044265270233, 'train/grad_norm_pre_clip': 6.359384059906006, 'timing/data': 0.0004792623221874237, 'timing/model': 1.277775600552559, 'learning_rate/qwen_vl_interface': 1.5363004582768584e-06, 'learning_rate/action_model': 1.5363004582768584e-05, 'epoch': 0.2}) +[2026-05-31 05:47:18,695][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.00552342040464282, 'train/grad_norm_pre_clip': 0.3875497877597809, 'timing/data': 0.00034193694591522217, 'timing/model': 1.257019067183137, 'learning_rate/qwen_vl_interface': 1.517389024020849e-06, 'learning_rate/action_model': 1.5173890240208488e-05, 'epoch': 0.2}) +[2026-05-31 05:50:29,085][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.16161976754665375, 'train/grad_norm_pre_clip': 14.653568267822266, 'timing/data': 0.00034403614699840546, 'timing/model': 1.2837131470441818, 'learning_rate/qwen_vl_interface': 1.498631022782424e-06, 'learning_rate/action_model': 1.4986310227824241e-05, 'epoch': 0.2}) +[2026-05-31 05:53:41,145][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.22979561984539032, 'train/grad_norm_pre_clip': 3.1114673614501953, 'timing/data': 0.00034496746957302094, 'timing/model': 1.233682505786419, 'learning_rate/qwen_vl_interface': 1.480027225629547e-06, 'learning_rate/action_model': 1.480027225629547e-05, 'epoch': 0.2}) +[2026-05-31 05:56:51,293][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.008086096495389938, 'train/grad_norm_pre_clip': 0.6101207137107849, 'timing/data': 0.00034205615520477295, 'timing/model': 1.29146927036345, 'learning_rate/qwen_vl_interface': 1.4615783972914494e-06, 'learning_rate/action_model': 1.4615783972914493e-05, 'epoch': 0.2}) +[2026-05-31 06:03:18,055][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4000 +[2026-05-31 06:03:18,056][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.012600780464708805, 'train/grad_norm_pre_clip': 0.3535168468952179, 'eval/loss': 0.02425369992852211, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 255.89999999999208, 'rl_games_eval/mid_train/mean_length': 2084.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 255.89999999999208, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2084.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.00034191645681858063, 'timing/model': 1.2820945270359516, 'learning_rate/qwen_vl_interface': 1.4432852961272068e-06, 'learning_rate/action_model': 1.4432852961272066e-05, 'epoch': 0.2}) +[2026-05-31 06:03:18,057][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4000_state +[2026-05-31 06:03:18,061][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 06:03:24,254][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4000_state/model.safetensors +[2026-05-31 06:03:31,501][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4000_state/optimizer.bin +[2026-05-31 06:03:31,502][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4000_state/sampler.bin +[2026-05-31 06:03:31,502][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4000_state/sampler_1.bin +[2026-05-31 06:03:31,503][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4000_state/random_states_0.pkl +[2026-05-31 06:03:35,282][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-05-31 06:03:35,282][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 06:03:35,294][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 06:06:47,184][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.35401451587677, 'train/grad_norm_pre_clip': 2.386131525039673, 'timing/data': 0.000346355140209198, 'timing/model': 1.28233852237463, 'learning_rate/qwen_vl_interface': 1.4251486740945584e-06, 'learning_rate/action_model': 1.4251486740945583e-05, 'epoch': 0.2}) +[2026-05-31 06:10:01,686][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.5950097441673279, 'train/grad_norm_pre_clip': 10.6212739944458, 'timing/data': 0.0003369152545928955, 'timing/model': 1.2860746774822474, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 0.2}) +[2026-05-31 06:13:13,893][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.003567436011508107, 'train/grad_norm_pre_clip': 0.09116256982088089, 'timing/data': 0.0003331657499074936, 'timing/model': 1.2941326834261417, 'learning_rate/qwen_vl_interface': 1.3893478430631376e-06, 'learning_rate/action_model': 1.3893478430631376e-05, 'epoch': 0.2}) +[2026-05-31 06:16:27,285][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.002436285838484764, 'train/grad_norm_pre_clip': 0.11778032034635544, 'timing/data': 0.0003233160823583603, 'timing/model': 1.287882013246417, 'learning_rate/qwen_vl_interface': 1.3716851056963062e-06, 'learning_rate/action_model': 1.3716851056963062e-05, 'epoch': 0.2}) +[2026-05-31 06:19:41,492][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.20856329798698425, 'train/grad_norm_pre_clip': 2.8570573329925537, 'timing/data': 0.00034157559275627136, 'timing/model': 1.2775671742856503, 'learning_rate/qwen_vl_interface': 1.3541817906644591e-06, 'learning_rate/action_model': 1.3541817906644591e-05, 'epoch': 0.2}) +[2026-05-31 06:22:55,594][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.1039203330874443, 'train/grad_norm_pre_clip': 2.524468183517456, 'timing/data': 0.0003358256071805954, 'timing/model': 1.291897013783455, 'learning_rate/qwen_vl_interface': 1.336838617460322e-06, 'learning_rate/action_model': 1.336838617460322e-05, 'epoch': 0.2}) +[2026-05-31 06:26:09,393][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.026946891099214554, 'train/grad_norm_pre_clip': 1.7169609069824219, 'timing/data': 0.0003365166485309601, 'timing/model': 1.284596685320139, 'learning_rate/qwen_vl_interface': 1.3196562989938175e-06, 'learning_rate/action_model': 1.3196562989938174e-05, 'epoch': 0.2}) +[2026-05-31 06:29:24,187][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.017832735553383827, 'train/grad_norm_pre_clip': 0.9193087220191956, 'timing/data': 0.0003347080200910568, 'timing/model': 1.2840222325176, 'learning_rate/qwen_vl_interface': 1.3026355415627565e-06, 'learning_rate/action_model': 1.3026355415627564e-05, 'epoch': 0.2}) +[2026-05-31 06:32:37,886][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.019944000989198685, 'train/grad_norm_pre_clip': 1.9543557167053223, 'timing/data': 0.0003374256193637848, 'timing/model': 1.2754654847085476, 'learning_rate/qwen_vl_interface': 1.2857770448238124e-06, 'learning_rate/action_model': 1.2857770448238123e-05, 'epoch': 0.2}) +[2026-05-31 06:41:12,268][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4100 +[2026-05-31 06:41:12,269][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.09545844793319702, 'train/grad_norm_pre_clip': 2.0254933834075928, 'eval/loss': 0.021157799288630486, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0003553759306669235, 'timing/model': 1.287520794197917, 'learning_rate/qwen_vl_interface': 1.2690815017637573e-06, 'learning_rate/action_model': 1.2690815017637572e-05, 'epoch': 0.2}) +[2026-05-31 06:41:12,270][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4100_state +[2026-05-31 06:41:12,275][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 06:41:18,255][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4100_state/model.safetensors +[2026-05-31 06:41:25,418][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4100_state/optimizer.bin +[2026-05-31 06:41:25,419][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4100_state/sampler.bin +[2026-05-31 06:41:25,419][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4100_state/sampler_1.bin +[2026-05-31 06:41:25,420][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4100_state/random_states_0.pkl +[2026-05-31 06:41:29,146][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4100 +[2026-05-31 06:41:29,146][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 06:41:29,158][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 06:44:39,561][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.0011406465200707316, 'train/grad_norm_pre_clip': 0.03985515609383583, 'timing/data': 0.00033293478190898895, 'timing/model': 1.261874496936798, 'learning_rate/qwen_vl_interface': 1.2525495986709727e-06, 'learning_rate/action_model': 1.2525495986709726e-05, 'epoch': 0.2}) +[2026-05-31 06:47:49,387][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.0005903036217205226, 'train/grad_norm_pre_clip': 0.013122759759426117, 'timing/data': 0.00033607520163059235, 'timing/model': 1.2919945437461138, 'learning_rate/qwen_vl_interface': 1.2361820151072454e-06, 'learning_rate/action_model': 1.2361820151072453e-05, 'epoch': 0.2}) +[2026-05-31 06:51:04,293][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.032794877886772156, 'train/grad_norm_pre_clip': 0.7123931646347046, 'timing/data': 0.0003287624567747116, 'timing/model': 1.2786789517849684, 'learning_rate/qwen_vl_interface': 1.2199794238798266e-06, 'learning_rate/action_model': 1.2199794238798264e-05, 'epoch': 0.2}) +[2026-05-31 06:54:17,233][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 1.0281577488058247e-05, 'train/grad_norm_pre_clip': 0.00016150619194377214, 'timing/data': 0.0003444049507379532, 'timing/model': 1.2345927003771067, 'learning_rate/qwen_vl_interface': 1.2039424910137807e-06, 'learning_rate/action_model': 1.2039424910137806e-05, 'epoch': 0.2}) +[2026-05-31 06:57:29,490][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.2064274549484253, 'train/grad_norm_pre_clip': 3.138739585876465, 'timing/data': 0.00033629685640335083, 'timing/model': 1.263373700901866, 'learning_rate/qwen_vl_interface': 1.1880718757246062e-06, 'learning_rate/action_model': 1.1880718757246063e-05, 'epoch': 0.2}) +[2026-05-31 07:00:43,697][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.11202181130647659, 'train/grad_norm_pre_clip': 9.376802444458008, 'timing/data': 0.00033803656697273254, 'timing/model': 1.2977353874593973, 'learning_rate/qwen_vl_interface': 1.172368230391136e-06, 'learning_rate/action_model': 1.1723682303911359e-05, 'epoch': 0.21}) +[2026-05-31 07:03:57,396][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.0010119708022102714, 'train/grad_norm_pre_clip': 0.035016320645809174, 'timing/data': 0.0003325771540403366, 'timing/model': 1.2879697363823652, 'learning_rate/qwen_vl_interface': 1.1568322005287237e-06, 'learning_rate/action_model': 1.1568322005287235e-05, 'epoch': 0.21}) +[2026-05-31 07:07:11,311][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.000621836690697819, 'train/grad_norm_pre_clip': 0.05409059301018715, 'timing/data': 0.00033273734152317047, 'timing/model': 1.2866902016103268, 'learning_rate/qwen_vl_interface': 1.1414644247627056e-06, 'learning_rate/action_model': 1.1414644247627058e-05, 'epoch': 0.21}) +[2026-05-31 07:10:24,192][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.00061776302754879, 'train/grad_norm_pre_clip': 0.028487296774983406, 'timing/data': 0.0003475639969110489, 'timing/model': 1.2966570556163788, 'learning_rate/qwen_vl_interface': 1.1262655348021523e-06, 'learning_rate/action_model': 1.1262655348021523e-05, 'epoch': 0.21}) +[2026-05-31 07:16:00,637][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4200 +[2026-05-31 07:16:00,639][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.00801204051822424, 'train/grad_norm_pre_clip': 0.19265443086624146, 'eval/loss': 0.022127337753772736, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 185.89999999999498, 'rl_games_eval/mid_train/mean_length': 1519.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 185.89999999999498, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1519.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0003497563302516937, 'timing/model': 1.2398539055138826, 'learning_rate/qwen_vl_interface': 1.1112361554139003e-06, 'learning_rate/action_model': 1.1112361554139004e-05, 'epoch': 0.21}) +[2026-05-31 07:16:00,639][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4200_state +[2026-05-31 07:16:00,644][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 07:16:06,540][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4200_state/model.safetensors +[2026-05-31 07:16:13,795][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4200_state/optimizer.bin +[2026-05-31 07:16:13,795][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4200_state/sampler.bin +[2026-05-31 07:16:13,795][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4200_state/sampler_1.bin +[2026-05-31 07:16:13,796][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4200_state/random_states_0.pkl +[2026-05-31 07:16:17,570][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4200 +[2026-05-31 07:16:17,571][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 07:16:17,582][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 07:19:31,878][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.0017007719725370407, 'train/grad_norm_pre_clip': 0.1086902990937233, 'timing/data': 0.0005881115794181824, 'timing/model': 1.2702092286199331, 'learning_rate/qwen_vl_interface': 1.0963769043968708e-06, 'learning_rate/action_model': 1.0963769043968708e-05, 'epoch': 0.21}) +[2026-05-31 07:22:43,844][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.1321580410003662, 'train/grad_norm_pre_clip': 8.947697639465332, 'timing/data': 0.00033707544207572937, 'timing/model': 1.2498777955770493, 'learning_rate/qwen_vl_interface': 1.0816883925566717e-06, 'learning_rate/action_model': 1.0816883925566716e-05, 'epoch': 0.21}) +[2026-05-31 07:25:55,360][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.012885591946542263, 'train/grad_norm_pre_clip': 0.4832526743412018, 'timing/data': 0.0003466568887233734, 'timing/model': 1.236877717077732, 'learning_rate/qwen_vl_interface': 1.0671712236804947e-06, 'learning_rate/action_model': 1.0671712236804946e-05, 'epoch': 0.21}) +[2026-05-31 07:29:09,776][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.009840752929449081, 'train/grad_norm_pre_clip': 0.2594904601573944, 'timing/data': 0.0003439858555793762, 'timing/model': 1.2945784386247396, 'learning_rate/qwen_vl_interface': 1.0528259945122925e-06, 'learning_rate/action_model': 1.0528259945122925e-05, 'epoch': 0.21}) +[2026-05-31 07:32:21,794][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.043273959308862686, 'train/grad_norm_pre_clip': 6.085651397705078, 'timing/data': 0.0003463961184024811, 'timing/model': 1.2847656719386578, 'learning_rate/qwen_vl_interface': 1.0386532947282496e-06, 'learning_rate/action_model': 1.0386532947282495e-05, 'epoch': 0.21}) +[2026-05-31 07:35:35,695][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.025897551327943802, 'train/grad_norm_pre_clip': 0.6644124388694763, 'timing/data': 0.0003417357802391052, 'timing/model': 1.2970463801175356, 'learning_rate/qwen_vl_interface': 1.0246537069125445e-06, 'learning_rate/action_model': 1.0246537069125444e-05, 'epoch': 0.21}) +[2026-05-31 07:38:50,193][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.0007513380260206759, 'train/grad_norm_pre_clip': 0.029308883473277092, 'timing/data': 0.00033558718860149384, 'timing/model': 1.2848232481628656, 'learning_rate/qwen_vl_interface': 1.0108278065333978e-06, 'learning_rate/action_model': 1.0108278065333978e-05, 'epoch': 0.21}) +[2026-05-31 07:42:04,894][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.009742503054440022, 'train/grad_norm_pre_clip': 1.319900631904602, 'timing/data': 0.0003322884440422058, 'timing/model': 1.2943198755383492, 'learning_rate/qwen_vl_interface': 9.971761619194231e-07, 'learning_rate/action_model': 9.97176161919423e-06, 'epoch': 0.21}) +[2026-05-31 07:45:19,094][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.3386780619621277, 'train/grad_norm_pre_clip': 2.7764949798583984, 'timing/data': 0.0003552362322807312, 'timing/model': 1.2747569996863604, 'learning_rate/qwen_vl_interface': 9.836993342362596e-07, 'learning_rate/action_model': 9.836993342362596e-06, 'epoch': 0.21}) +[2026-05-31 07:53:50,854][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4300 +[2026-05-31 07:53:50,855][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.19229446351528168, 'train/grad_norm_pre_clip': 3.575772285461426, 'eval/loss': 0.02204408496618271, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.00033248402178287506, 'timing/model': 1.2863839510828257, 'learning_rate/qwen_vl_interface': 9.703978774635094e-07, 'learning_rate/action_model': 9.703978774635095e-06, 'epoch': 0.21}) +[2026-05-31 07:53:50,856][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4300_state +[2026-05-31 07:53:51,136][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 07:53:57,139][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4300_state/model.safetensors +[2026-05-31 07:54:04,456][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4300_state/optimizer.bin +[2026-05-31 07:54:04,456][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4300_state/sampler.bin +[2026-05-31 07:54:04,457][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4300_state/sampler_1.bin +[2026-05-31 07:54:04,458][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4300_state/random_states_0.pkl +[2026-05-31 07:54:08,285][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4300 +[2026-05-31 07:54:08,286][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 07:54:08,297][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 07:57:21,915][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.00016832788242027164, 'train/grad_norm_pre_clip': 0.005210292991250753, 'timing/data': 0.0003394056111574173, 'timing/model': 1.237531092017889, 'learning_rate/qwen_vl_interface': 9.572723383719611e-07, 'learning_rate/action_model': 9.572723383719612e-06, 'epoch': 0.21}) +[2026-05-31 08:00:34,481][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 9.037289419211447e-05, 'train/grad_norm_pre_clip': 0.002494961256161332, 'timing/data': 0.0003352649509906769, 'timing/model': 1.2817471288144588, 'learning_rate/qwen_vl_interface': 9.443232565011188e-07, 'learning_rate/action_model': 9.443232565011187e-06, 'epoch': 0.21}) +[2026-05-31 08:03:49,385][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.10061105340719223, 'train/grad_norm_pre_clip': 2.6369266510009766, 'timing/data': 0.000327184796333313, 'timing/model': 1.284607283771038, 'learning_rate/qwen_vl_interface': 9.315511641370191e-07, 'learning_rate/action_model': 9.315511641370191e-06, 'epoch': 0.21}) +[2026-05-31 08:07:03,290][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.14075523614883423, 'train/grad_norm_pre_clip': 3.2403836250305176, 'timing/data': 0.00032158754765987396, 'timing/model': 1.2956779729574919, 'learning_rate/qwen_vl_interface': 9.189565862903543e-07, 'learning_rate/action_model': 9.189565862903543e-06, 'epoch': 0.21}) +[2026-05-31 08:10:15,694][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.009683882817626, 'train/grad_norm_pre_clip': 0.3064112663269043, 'timing/data': 0.00034569762647151947, 'timing/model': 1.2733228616416454, 'learning_rate/qwen_vl_interface': 9.065400406748912e-07, 'learning_rate/action_model': 9.06540040674891e-06, 'epoch': 0.21}) +[2026-05-31 08:13:30,384][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.012375003658235073, 'train/grad_norm_pre_clip': 0.5386063456535339, 'timing/data': 0.000343983992934227, 'timing/model': 1.2772045619785786, 'learning_rate/qwen_vl_interface': 8.943020376861858e-07, 'learning_rate/action_model': 8.943020376861857e-06, 'epoch': 0.22}) +[2026-05-31 08:16:44,386][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 9.990757826017216e-05, 'train/grad_norm_pre_clip': 0.0022450906690210104, 'timing/data': 0.00033199600875377655, 'timing/model': 1.2841419205069542, 'learning_rate/qwen_vl_interface': 8.822430803806098e-07, 'learning_rate/action_model': 8.822430803806097e-06, 'epoch': 0.22}) +[2026-05-31 08:19:58,687][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.0027502812445163727, 'train/grad_norm_pre_clip': 0.07730221748352051, 'timing/data': 0.0006150417029857635, 'timing/model': 1.2881773505359888, 'learning_rate/qwen_vl_interface': 8.703636644546651e-07, 'learning_rate/action_model': 8.703636644546651e-06, 'epoch': 0.22}) +[2026-05-31 08:23:13,191][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.07257524877786636, 'train/grad_norm_pre_clip': 10.494956970214844, 'timing/data': 0.0003570280969142914, 'timing/model': 1.2842008266597986, 'learning_rate/qwen_vl_interface': 8.586642782246133e-07, 'learning_rate/action_model': 8.586642782246131e-06, 'epoch': 0.22}) +[2026-05-31 08:31:40,664][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4400 +[2026-05-31 08:31:40,666][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.09543229639530182, 'train/grad_norm_pre_clip': 2.873114824295044, 'eval/loss': 0.024268968030810356, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.003239493817090988, 'timing/model': 1.2347301356494427, 'learning_rate/qwen_vl_interface': 8.471454026063972e-07, 'learning_rate/action_model': 8.47145402606397e-06, 'epoch': 0.22}) +[2026-05-31 08:31:40,666][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4400_state +[2026-05-31 08:31:40,671][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 08:31:46,668][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4400_state/model.safetensors +[2026-05-31 08:31:53,830][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4400_state/optimizer.bin +[2026-05-31 08:31:53,831][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4400_state/sampler.bin +[2026-05-31 08:31:53,831][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4400_state/sampler_1.bin +[2026-05-31 08:31:53,832][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4400_state/random_states_0.pkl +[2026-05-31 08:31:57,558][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4400 +[2026-05-31 08:31:57,559][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 08:31:57,570][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 08:35:09,861][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.013079814612865448, 'train/grad_norm_pre_clip': 0.8025646805763245, 'timing/data': 0.006143426522612572, 'timing/model': 1.2524071242660284, 'learning_rate/qwen_vl_interface': 8.358075110958785e-07, 'learning_rate/action_model': 8.358075110958784e-06, 'epoch': 0.22}) +[2026-05-31 08:38:18,893][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.0009175910381600261, 'train/grad_norm_pre_clip': 0.05399366095662117, 'timing/data': 0.005072711035609245, 'timing/model': 1.2830199636518955, 'learning_rate/qwen_vl_interface': 8.246510697493681e-07, 'learning_rate/action_model': 8.246510697493682e-06, 'epoch': 0.22}) +[2026-05-31 08:41:32,380][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.00030054213129915297, 'train/grad_norm_pre_clip': 0.010668732225894928, 'timing/data': 0.005766032263636589, 'timing/model': 1.2327099405229092, 'learning_rate/qwen_vl_interface': 8.136765371644753e-07, 'learning_rate/action_model': 8.136765371644753e-06, 'epoch': 0.22}) +[2026-05-31 08:44:44,785][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.025304675102233887, 'train/grad_norm_pre_clip': 2.0498275756835938, 'timing/data': 0.00561104342341423, 'timing/model': 1.2796220313757658, 'learning_rate/qwen_vl_interface': 8.028843644612483e-07, 'learning_rate/action_model': 8.028843644612482e-06, 'epoch': 0.22}) +[2026-05-31 08:47:55,599][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.00014612791710533202, 'train/grad_norm_pre_clip': 0.0067970226518809795, 'timing/data': 0.005707714706659317, 'timing/model': 1.5161441247910261, 'learning_rate/qwen_vl_interface': 7.92274995263638e-07, 'learning_rate/action_model': 7.92274995263638e-06, 'epoch': 0.22}) +[2026-05-31 08:51:08,687][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.022381609305739403, 'train/grad_norm_pre_clip': 0.8516548871994019, 'timing/data': 0.00033528730273246765, 'timing/model': 1.282408056780696, 'learning_rate/qwen_vl_interface': 7.818488656812581e-07, 'learning_rate/action_model': 7.818488656812582e-06, 'epoch': 0.22}) +[2026-05-31 08:54:22,592][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.29467955231666565, 'train/grad_norm_pre_clip': 10.117236137390137, 'timing/data': 0.0003409162163734436, 'timing/model': 1.2897558771073818, 'learning_rate/qwen_vl_interface': 7.716064042914584e-07, 'learning_rate/action_model': 7.716064042914584e-06, 'epoch': 0.22}) +[2026-05-31 08:57:37,388][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.05682617798447609, 'train/grad_norm_pre_clip': 0.9405868053436279, 'timing/data': 0.00034922733902931213, 'timing/model': 1.2890617810189724, 'learning_rate/qwen_vl_interface': 7.615480321217116e-07, 'learning_rate/action_model': 7.615480321217117e-06, 'epoch': 0.22}) +[2026-05-31 09:00:51,207][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.0002039326063822955, 'train/grad_norm_pre_clip': 0.004221287090331316, 'timing/data': 0.00040734559297561646, 'timing/model': 1.235674299299717, 'learning_rate/qwen_vl_interface': 7.516741626322994e-07, 'learning_rate/action_model': 7.516741626322994e-06, 'epoch': 0.22}) +[2026-05-31 09:09:19,774][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4500 +[2026-05-31 09:09:19,776][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.21689473092556, 'train/grad_norm_pre_clip': 12.256173133850098, 'eval/loss': 0.02439611777663231, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 437.9000000000246, 'rl_games_eval/mid_train/mean_length': 3553.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 437.9000000000246, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3553.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0003404170274734497, 'timing/model': 1.4592519458383322, 'learning_rate/qwen_vl_interface': 7.41985201699324e-07, 'learning_rate/action_model': 7.41985201699324e-06, 'epoch': 0.22}) +[2026-05-31 09:09:19,776][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4500_state +[2026-05-31 09:09:19,781][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 09:09:25,795][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4500_state/model.safetensors +[2026-05-31 09:09:32,970][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4500_state/optimizer.bin +[2026-05-31 09:09:32,971][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4500_state/sampler.bin +[2026-05-31 09:09:32,971][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4500_state/sampler_1.bin +[2026-05-31 09:09:32,972][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4500_state/random_states_0.pkl +[2026-05-31 09:09:36,714][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-05-31 09:09:36,714][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 09:09:36,726][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 09:12:48,395][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.09432577341794968, 'train/grad_norm_pre_clip': 6.11357307434082, 'timing/data': 0.0003434475511312485, 'timing/model': 1.29462880641222, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 0.22}) +[2026-05-31 09:16:02,886][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.055538151413202286, 'train/grad_norm_pre_clip': 1.1284081935882568, 'timing/data': 0.00033378414809703827, 'timing/model': 1.2891500927507877, 'learning_rate/qwen_vl_interface': 7.231635909863858e-07, 'learning_rate/action_model': 7.2316359098638585e-06, 'epoch': 0.22}) +[2026-05-31 09:19:17,789][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.0009516404825262725, 'train/grad_norm_pre_clip': 0.029672030359506607, 'timing/data': 0.00034320540726184845, 'timing/model': 1.2808799166232347, 'learning_rate/qwen_vl_interface': 7.14031714889121e-07, 'learning_rate/action_model': 7.14031714889121e-06, 'epoch': 0.22}) +[2026-05-31 09:22:32,389][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.04476161301136017, 'train/grad_norm_pre_clip': 1.4596043825149536, 'timing/data': 0.00033681467175483704, 'timing/model': 1.2791111432015896, 'learning_rate/qwen_vl_interface': 7.050862946818836e-07, 'learning_rate/action_model': 7.050862946818836e-06, 'epoch': 0.22}) +[2026-05-31 09:25:45,683][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.006418340839445591, 'train/grad_norm_pre_clip': 0.3517901301383972, 'timing/data': 0.00032463856041431427, 'timing/model': 1.284416787326336, 'learning_rate/qwen_vl_interface': 6.963276980758622e-07, 'learning_rate/action_model': 6.963276980758622e-06, 'epoch': 0.22}) +[2026-05-31 09:29:00,359][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.007648463826626539, 'train/grad_norm_pre_clip': 0.26396048069000244, 'timing/data': 0.0003352668136358261, 'timing/model': 1.4592520352452993, 'learning_rate/qwen_vl_interface': 6.877562851026584e-07, 'learning_rate/action_model': 6.877562851026584e-06, 'epoch': 0.23}) +[2026-05-31 09:32:13,893][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.00011926590377697721, 'train/grad_norm_pre_clip': 0.0028683028649538755, 'timing/data': 0.00034092552959918976, 'timing/model': 1.2894860580563545, 'learning_rate/qwen_vl_interface': 6.793724080994884e-07, 'learning_rate/action_model': 6.793724080994884e-06, 'epoch': 0.23}) +[2026-05-31 09:35:28,792][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.0016286879545077682, 'train/grad_norm_pre_clip': 0.06904493272304535, 'timing/data': 0.0005141254514455795, 'timing/model': 1.287106255069375, 'learning_rate/qwen_vl_interface': 6.711764116946972e-07, 'learning_rate/action_model': 6.7117641169469715e-06, 'epoch': 0.23}) +[2026-05-31 09:38:38,812][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.008888067677617073, 'train/grad_norm_pre_clip': 0.5240287184715271, 'timing/data': 0.0003433041274547577, 'timing/model': 1.238250469788909, 'learning_rate/qwen_vl_interface': 6.631686327935954e-07, 'learning_rate/action_model': 6.631686327935954e-06, 'epoch': 0.23}) +[2026-05-31 09:47:14,523][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4600 +[2026-05-31 09:47:14,525][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.02819656953215599, 'train/grad_norm_pre_clip': 0.9845460653305054, 'eval/loss': 0.021581178531050682, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.00042326562106609344, 'timing/model': 1.2794903423637152, 'learning_rate/qwen_vl_interface': 6.553494005646101e-07, 'learning_rate/action_model': 6.553494005646101e-06, 'epoch': 0.23}) +[2026-05-31 09:47:14,525][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4600_state +[2026-05-31 09:47:14,530][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 09:47:20,517][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4600_state/model.safetensors +[2026-05-31 09:47:27,681][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4600_state/optimizer.bin +[2026-05-31 09:47:27,682][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4600_state/sampler.bin +[2026-05-31 09:47:27,682][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4600_state/sampler_1.bin +[2026-05-31 09:47:27,683][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4600_state/random_states_0.pkl +[2026-05-31 09:47:31,398][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4600 +[2026-05-31 09:47:31,398][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 09:47:31,410][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 09:50:44,988][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.44012677669525146, 'train/grad_norm_pre_clip': 9.373298645019531, 'timing/data': 0.0003453567624092102, 'timing/model': 1.2841476909816265, 'learning_rate/qwen_vl_interface': 6.477190364257524e-07, 'learning_rate/action_model': 6.477190364257524e-06, 'epoch': 0.23}) +[2026-05-31 09:53:58,791][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.0040760948322713375, 'train/grad_norm_pre_clip': 0.28835076093673706, 'timing/data': 0.00034338608384132385, 'timing/model': 1.294285025447607, 'learning_rate/qwen_vl_interface': 6.402778540314054e-07, 'learning_rate/action_model': 6.402778540314054e-06, 'epoch': 0.23}) +[2026-05-31 09:57:13,094][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.0017967699095606804, 'train/grad_norm_pre_clip': 0.10009095817804337, 'timing/data': 0.00033711642026901245, 'timing/model': 1.2995485924184322, 'learning_rate/qwen_vl_interface': 6.330261592594328e-07, 'learning_rate/action_model': 6.330261592594327e-06, 'epoch': 0.23}) +[2026-05-31 10:00:27,389][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.000741286261472851, 'train/grad_norm_pre_clip': 0.024958109483122826, 'timing/data': 0.00033403560519218445, 'timing/model': 1.2906360160559416, 'learning_rate/qwen_vl_interface': 6.25964250198603e-07, 'learning_rate/action_model': 6.2596425019860294e-06, 'epoch': 0.23}) +[2026-05-31 10:03:42,093][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.001361012109555304, 'train/grad_norm_pre_clip': 0.04741794988512993, 'timing/data': 0.00034682638943195343, 'timing/model': 1.2899980936199427, 'learning_rate/qwen_vl_interface': 6.190924171363378e-07, 'learning_rate/action_model': 6.190924171363378e-06, 'epoch': 0.23}) +[2026-05-31 10:06:56,804][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 3.507208384689875e-05, 'train/grad_norm_pre_clip': 0.0006528437952511013, 'timing/data': 0.0006286222487688065, 'timing/model': 1.2998134456574917, 'learning_rate/qwen_vl_interface': 6.124109425467805e-07, 'learning_rate/action_model': 6.124109425467805e-06, 'epoch': 0.23}) +[2026-05-31 10:10:08,398][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.20982326567173004, 'train/grad_norm_pre_clip': 2.8214516639709473, 'timing/data': 0.00034601613879203796, 'timing/model': 1.2853443268686533, 'learning_rate/qwen_vl_interface': 6.059201010791811e-07, 'learning_rate/action_model': 6.059201010791811e-06, 'epoch': 0.23}) +[2026-05-31 10:13:22,491][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.003273390932008624, 'train/grad_norm_pre_clip': 0.14645947515964508, 'timing/data': 0.00040061771869659424, 'timing/model': 1.291754761710763, 'learning_rate/qwen_vl_interface': 5.996201595466088e-07, 'learning_rate/action_model': 5.996201595466088e-06, 'epoch': 0.23}) +[2026-05-31 10:16:36,030][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.03381934389472008, 'train/grad_norm_pre_clip': 2.9999966621398926, 'timing/data': 0.0003627873957157135, 'timing/model': 1.236894030123949, 'learning_rate/qwen_vl_interface': 5.935113769149855e-07, 'learning_rate/action_model': 5.935113769149856e-06, 'epoch': 0.23}) +[2026-05-31 10:25:07,823][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4700 +[2026-05-31 10:25:07,824][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.0006632002769038081, 'train/grad_norm_pre_clip': 0.026674222201108932, 'eval/loss': 0.021971678361296654, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.00033547542989254, 'timing/model': 1.2389189377427101, 'learning_rate/qwen_vl_interface': 5.875940042924391e-07, 'learning_rate/action_model': 5.875940042924391e-06, 'epoch': 0.23}) +[2026-05-31 10:25:07,825][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4700_state +[2026-05-31 10:25:07,829][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 10:25:13,763][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4700_state/model.safetensors +[2026-05-31 10:25:21,046][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4700_state/optimizer.bin +[2026-05-31 10:25:21,047][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4700_state/sampler.bin +[2026-05-31 10:25:21,047][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4700_state/sampler_1.bin +[2026-05-31 10:25:21,048][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4700_state/random_states_0.pkl +[2026-05-31 10:25:24,846][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4700 +[2026-05-31 10:25:24,846][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 10:25:24,858][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 10:28:39,692][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.12422174960374832, 'train/grad_norm_pre_clip': 2.4114809036254883, 'timing/data': 0.0033027995377779007, 'timing/model': 1.2882549799978733, 'learning_rate/qwen_vl_interface': 5.818682849189818e-07, 'learning_rate/action_model': 5.818682849189817e-06, 'epoch': 0.23}) +[2026-05-31 10:31:54,194][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.004147423431277275, 'train/grad_norm_pre_clip': 0.8544024229049683, 'timing/data': 0.0003491658717393875, 'timing/model': 1.2933581955730915, 'learning_rate/qwen_vl_interface': 5.763344541565091e-07, 'learning_rate/action_model': 5.7633445415650905e-06, 'epoch': 0.23}) +[2026-05-31 10:35:07,036][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.15144507586956024, 'train/grad_norm_pre_clip': 7.059175968170166, 'timing/data': 0.00581791065633297, 'timing/model': 1.2347270660102367, 'learning_rate/qwen_vl_interface': 5.709927394791305e-07, 'learning_rate/action_model': 5.709927394791305e-06, 'epoch': 0.23}) +[2026-05-31 10:38:18,784][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.00022313294175546616, 'train/grad_norm_pre_clip': 0.005252143833786249, 'timing/data': 0.008296435698866844, 'timing/model': 1.271566852927208, 'learning_rate/qwen_vl_interface': 5.658433604638141e-07, 'learning_rate/action_model': 5.65843360463814e-06, 'epoch': 0.23}) +[2026-05-31 10:41:28,334][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.02689599245786667, 'train/grad_norm_pre_clip': 0.7949513792991638, 'timing/data': 0.005902836099267006, 'timing/model': 1.23533776961267, 'learning_rate/qwen_vl_interface': 5.608865287813617e-07, 'learning_rate/action_model': 5.608865287813616e-06, 'epoch': 0.23}) +[2026-05-31 10:44:40,388][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.03270043432712555, 'train/grad_norm_pre_clip': 15.528097152709961, 'timing/data': 0.006924644112586975, 'timing/model': 1.2783531565219164, 'learning_rate/qwen_vl_interface': 5.561224481877105e-07, 'learning_rate/action_model': 5.561224481877105e-06, 'epoch': 0.24}) +[2026-05-31 10:47:54,889][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.0700659453868866, 'train/grad_norm_pre_clip': 3.4432973861694336, 'timing/data': 0.0031174160540103912, 'timing/model': 1.2874647565186024, 'learning_rate/qwen_vl_interface': 5.515513145155531e-07, 'learning_rate/action_model': 5.515513145155531e-06, 'epoch': 0.24}) +[2026-05-31 10:51:08,010][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.27634966373443604, 'train/grad_norm_pre_clip': 3.9692914485931396, 'timing/data': 0.002938155084848404, 'timing/model': 1.2339904550462961, 'learning_rate/qwen_vl_interface': 5.471733156662914e-07, 'learning_rate/action_model': 5.4717331566629135e-06, 'epoch': 0.24}) +[2026-05-31 10:54:21,810][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.003895355388522148, 'train/grad_norm_pre_clip': 0.1083078533411026, 'timing/data': 0.003234485164284706, 'timing/model': 1.2358711771667004, 'learning_rate/qwen_vl_interface': 5.429886316023102e-07, 'learning_rate/action_model': 5.429886316023101e-06, 'epoch': 0.24}) +[2026-05-31 11:01:31,173][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4800 +[2026-05-31 11:01:31,175][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.005182971246540546, 'train/grad_norm_pre_clip': 0.24356122314929962, 'eval/loss': 0.022628113627433777, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 320.9000000000037, 'rl_games_eval/mid_train/mean_length': 2608.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 320.9000000000037, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2608.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.0030622147023677826, 'timing/model': 1.2873966712504625, 'learning_rate/qwen_vl_interface': 5.389974343395813e-07, 'learning_rate/action_model': 5.389974343395812e-06, 'epoch': 0.24}) +[2026-05-31 11:01:31,176][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4800_state +[2026-05-31 11:01:31,181][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 11:01:37,151][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4800_state/model.safetensors +[2026-05-31 11:01:44,304][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4800_state/optimizer.bin +[2026-05-31 11:01:44,305][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4800_state/sampler.bin +[2026-05-31 11:01:44,305][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4800_state/sampler_1.bin +[2026-05-31 11:01:44,306][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4800_state/random_states_0.pkl +[2026-05-31 11:01:48,081][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4800 +[2026-05-31 11:01:48,082][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 11:01:48,094][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 11:04:58,390][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.2207498848438263, 'train/grad_norm_pre_clip': 4.849417686462402, 'timing/data': 0.0033229701220989227, 'timing/model': 1.2895313948392868, 'learning_rate/qwen_vl_interface': 5.351998879405911e-07, 'learning_rate/action_model': 5.351998879405911e-06, 'epoch': 0.24}) +[2026-05-31 11:08:12,054][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.0010891018901020288, 'train/grad_norm_pre_clip': 0.021369915455579758, 'timing/data': 0.003229452297091484, 'timing/model': 1.2357570920139551, 'learning_rate/qwen_vl_interface': 5.315961485075992e-07, 'learning_rate/action_model': 5.315961485075993e-06, 'epoch': 0.24}) +[2026-05-31 11:11:24,988][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.00932646170258522, 'train/grad_norm_pre_clip': 0.5832482576370239, 'timing/data': 0.003397628664970398, 'timing/model': 1.2344749886542559, 'learning_rate/qwen_vl_interface': 5.281863641762184e-07, 'learning_rate/action_model': 5.281863641762184e-06, 'epoch': 0.24}) +[2026-05-31 11:14:37,139][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.3049941062927246, 'train/grad_norm_pre_clip': 4.37786865234375, 'timing/data': 0.008233113214373589, 'timing/model': 1.2352166343480349, 'learning_rate/qwen_vl_interface': 5.249706751093274e-07, 'learning_rate/action_model': 5.249706751093273e-06, 'epoch': 0.24}) +[2026-05-31 11:17:49,000][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.10716020315885544, 'train/grad_norm_pre_clip': 2.7657155990600586, 'timing/data': 0.007741309702396393, 'timing/model': 1.2539079170674086, 'learning_rate/qwen_vl_interface': 5.219492134913084e-07, 'learning_rate/action_model': 5.219492134913084e-06, 'epoch': 0.24}) +[2026-05-31 11:21:02,094][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.0009911229135468602, 'train/grad_norm_pre_clip': 0.02815915457904339, 'timing/data': 0.00794989988207817, 'timing/model': 1.2932588122785091, 'learning_rate/qwen_vl_interface': 5.191221035226144e-07, 'learning_rate/action_model': 5.191221035226144e-06, 'epoch': 0.24}) +[2026-05-31 11:24:16,085][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.49387210607528687, 'train/grad_norm_pre_clip': 4.514601707458496, 'timing/data': 0.007424036040902138, 'timing/model': 1.279414914548397, 'learning_rate/qwen_vl_interface': 5.164894614146633e-07, 'learning_rate/action_model': 5.164894614146634e-06, 'epoch': 0.24}) +[2026-05-31 11:27:30,385][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.30754807591438293, 'train/grad_norm_pre_clip': 2.0972042083740234, 'timing/data': 0.005493197590112686, 'timing/model': 1.283801555633545, 'learning_rate/qwen_vl_interface': 5.140513953850612e-07, 'learning_rate/action_model': 5.140513953850612e-06, 'epoch': 0.24}) +[2026-05-31 11:30:45,094][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.19384334981441498, 'train/grad_norm_pre_clip': 2.700510263442993, 'timing/data': 0.005799733102321625, 'timing/model': 1.2885074485093355, 'learning_rate/qwen_vl_interface': 5.118080056531524e-07, 'learning_rate/action_model': 5.118080056531524e-06, 'epoch': 0.24}) +[2026-05-31 11:39:15,374][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4900 +[2026-05-31 11:39:15,376][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.004389891400933266, 'train/grad_norm_pre_clip': 0.14512673020362854, 'eval/loss': 0.022179974243044853, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.005836503580212593, 'timing/model': 1.236577907577157, 'learning_rate/qwen_vl_interface': 5.097593844359024e-07, 'learning_rate/action_model': 5.0975938443590245e-06, 'epoch': 0.24}) +[2026-05-31 11:39:15,376][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4900_state +[2026-05-31 11:39:15,380][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 11:39:21,403][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4900_state/model.safetensors +[2026-05-31 11:39:28,638][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4900_state/optimizer.bin +[2026-05-31 11:39:28,639][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4900_state/sampler.bin +[2026-05-31 11:39:28,639][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4900_state/sampler_1.bin +[2026-05-31 11:39:28,640][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_4900_state/random_states_0.pkl +[2026-05-31 11:39:39,274][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4900 +[2026-05-31 11:39:39,274][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 11:39:39,286][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 11:42:49,814][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.015525211580097675, 'train/grad_norm_pre_clip': 0.27819690108299255, 'timing/data': 0.005514416843652725, 'timing/model': 1.2351183388382196, 'learning_rate/qwen_vl_interface': 5.07905615944106e-07, 'learning_rate/action_model': 5.0790561594410595e-06, 'epoch': 0.24}) +[2026-05-31 11:46:03,294][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.05549812689423561, 'train/grad_norm_pre_clip': 2.8484625816345215, 'timing/data': 0.00793016329407692, 'timing/model': 1.2468318734318018, 'learning_rate/qwen_vl_interface': 5.062467763789242e-07, 'learning_rate/action_model': 5.062467763789242e-06, 'epoch': 0.24}) +[2026-05-31 11:49:17,589][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.09660407155752182, 'train/grad_norm_pre_clip': 6.461259841918945, 'timing/data': 0.007121726870536804, 'timing/model': 1.2767701856791973, 'learning_rate/qwen_vl_interface': 5.047829339287557e-07, 'learning_rate/action_model': 5.047829339287557e-06, 'epoch': 0.24}) +[2026-05-31 11:52:31,788][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.18145380914211273, 'train/grad_norm_pre_clip': 3.096811294555664, 'timing/data': 0.008275294676423073, 'timing/model': 1.2702775821089745, 'learning_rate/qwen_vl_interface': 5.035141487664285e-07, 'learning_rate/action_model': 5.035141487664285e-06, 'epoch': 0.24}) +[2026-05-31 11:55:43,284][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.000376984040485695, 'train/grad_norm_pre_clip': 0.012030092999339104, 'timing/data': 0.007957886904478073, 'timing/model': 1.279752392321825, 'learning_rate/qwen_vl_interface': 5.024404730467325e-07, 'learning_rate/action_model': 5.024404730467325e-06, 'epoch': 0.24}) +[2026-05-31 11:58:55,586][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 2.0823083104914986e-05, 'train/grad_norm_pre_clip': 0.00032606153399683535, 'timing/data': 0.005069104954600334, 'timing/model': 1.234998831525445, 'learning_rate/qwen_vl_interface': 5.015619509042709e-07, 'learning_rate/action_model': 5.015619509042709e-06, 'epoch': 0.24}) +[2026-05-31 12:02:08,389][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.0004132036119699478, 'train/grad_norm_pre_clip': 0.018580161035060883, 'timing/data': 0.003245340660214424, 'timing/model': 1.2809483874589205, 'learning_rate/qwen_vl_interface': 5.008786184516475e-07, 'learning_rate/action_model': 5.008786184516474e-06, 'epoch': 0.25}) +[2026-05-31 12:05:22,699][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.0720931738615036, 'train/grad_norm_pre_clip': 2.461674690246582, 'timing/data': 0.0031321626156568527, 'timing/model': 1.2934049349278212, 'learning_rate/qwen_vl_interface': 5.003905037779834e-07, 'learning_rate/action_model': 5.0039050377798345e-06, 'epoch': 0.25}) +[2026-05-31 12:08:35,790][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.0036438703536987305, 'train/grad_norm_pre_clip': 0.09497638046741486, 'timing/data': 0.003383653238415718, 'timing/model': 1.2755709309130907, 'learning_rate/qwen_vl_interface': 5.000976269477613e-07, 'learning_rate/action_model': 5.0009762694776125e-06, 'epoch': 0.25}) +[2026-05-31 12:13:46,487][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 5000 +[2026-05-31 12:13:46,488][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.025039829313755035, 'train/grad_norm_pre_clip': 1.4721229076385498, 'eval/loss': 0.02123856171965599, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 146.79999999999674, 'rl_games_eval/mid_train/mean_length': 1200.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 146.79999999999674, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1200.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 1.0, 'timing/data': 0.002932855859398842, 'timing/model': 1.2341910395771265, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 0.25}) +[2026-05-31 12:13:46,489][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_5000_state +[2026-05-31 12:13:46,493][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-31 12:13:52,435][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_5000_state/model.safetensors +[2026-05-31 12:13:59,726][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_5000_state/optimizer.bin +[2026-05-31 12:13:59,726][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_5000_state/sampler.bin +[2026-05-31 12:13:59,727][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_5000_state/sampler_1.bin +[2026-05-31 12:13:59,728][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints/steps_5000_state/random_states_0.pkl +[2026-05-31 12:14:04,037][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-05-31 12:14:04,038][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-31 12:14:04,049][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-31 12:14:04,051][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_2/checkpoints +[2026-05-31 12:31:32,373][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 5000 +[2026-05-31 12:31:34,018][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/checkpoints/best_model.safetensors b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/checkpoints/best_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b52046573c0b8a22eac43023f7efb99f610f0d27 --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/checkpoints/best_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1f2b7941653b956dbb5c368b9da60e5538e04216c34371254ef6cfa5c34bce74 +size 9138230516 diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_500_model.safetensors b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_500_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..87c79185184c5a320b8a074cccae657d5c194d40 --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_500_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b1478d01499d37121fee35d4f9f1dc090f6e87192797f50556388c94970c3fc1 +size 9138230516 diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/config.full.yaml b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..9909443f9de444f7538ddf558743d8ead9091047 --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/config.full.yaml @@ -0,0 +1,191 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 7 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_0 + data_mix: deadly_corridor_train__bridge + eval_data_mix: deadly_corridor_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + converted_name: deadly_corridor_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: deadly_corridor_train + mixed_converted_name: deadly_corridor_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 500 + num_warmup_steps: 100 + save_interval: 50 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: You are playing Deadly Corridor in VizDoom. Choose actions from + MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, ATTACK. + deadly: + action_layout: multibinary_7 + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_0/deadly_corridor_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 50 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 2 + - 4 + - 6 + - 8 + num_episodes: 5 + max_steps_per_episode: 3600 + task: deadly_corridor + deadly_corridor_loss_type: null + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: deadly_corridor +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: true + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_deadly_corridor_fix_latency_0 +output_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0 +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/config.yaml b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..8e65bde1ec77da2935a6941b5f220bb1f7152ab6 --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/config.yaml @@ -0,0 +1,92 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: true + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: deadly_corridor_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: deadly_corridor_train__bridge__val + include_state: true + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 7 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0 +rl_games: + env_eval: + deadly: + action_layout: multibinary_7 + distributed_mode: none + enabled: true + fixed_episode_seeds: true + frameskip: 4 + image_size: 224 + latency: + prompt_map_path: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_0/deadly_corridor_train__bridge/latency_prompt_map.json + latency_seed_stride: 0 + mid_train: + enabled: true + interval_steps: 50 + latencies: + - 0 + max_steps_per_episode: 3600 + num_episodes: 5 + seed: 42 + task_description: You are playing Deadly Corridor in VizDoom. Choose actions from + MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, ATTACK. + task_seed_stride: 0 + vectorized: + enabled: false + model_alias: openvla + task: deadly_corridor +run_id: openvla_deadly_corridor_fix_latency_0 +run_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_interval: 100 + eval_num_batches: 20 + freeze_modules: '' + gradient_accumulation_steps: 16 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 500 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + pretrained_checkpoint: null + reload_modules: null + save_interval: 50 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/dataset_statistics.json b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..c7707a242f1e0b25d119fc060f26eacda46c2fe0 --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.9220542311668396, + 0.006256517022848129, + 0.16710114479064941, + 0.6543274521827698, + 0.23253388702869415, + 0.11183524876832962, + 0.5578727722167969 + ], + "std": [ + 0.26808345317840576, + 0.07885219156742096, + 0.3730665147304535, + 0.47559070587158203, + 0.42244860529899597, + 0.31516844034194946, + 0.4966414272785187 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 0.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 3836, + "num_trajectories": 90 + } +} \ No newline at end of file diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/dataset_statistics_eval.json b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..ebeb0582f527163e466d15b8ad6f426c69c9a5ab --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.9322798848152161, + 0.0, + 0.17607223987579346, + 0.6410835385322571, + 0.2370203137397766, + 0.11512415111064911, + 0.5214446783065796 + ], + "std": [ + 0.25126445293426514, + 0.0, + 0.38088199496269226, + 0.4796830415725708, + 0.4252544641494751, + 0.31917139887809753, + 0.49953892827033997 + ], + "max": [ + 1.0, + 0.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 0.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 443, + "num_trajectories": 10 + } +} \ No newline at end of file diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/hydra/.hydra/config.yaml b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..cbee04cbbd663dabbdb094594009f5b327f5a445 --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/hydra/.hydra/config.yaml @@ -0,0 +1,189 @@ +framework: + qwenvl: + base_vlm: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 7 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_0 + data_mix: deadly_corridor_train__bridge + eval_data_mix: deadly_corridor_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: deadly_corridor_train + mixed_converted_name: deadly_corridor_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 500 + num_warmup_steps: 100 + save_interval: 50 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: You are playing Deadly Corridor in VizDoom. Choose actions from + MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, ATTACK. + deadly: + action_layout: multibinary_7 + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_0/deadly_corridor_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 50 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 2 + - 4 + - 6 + - 8 + num_episodes: 5 + max_steps_per_episode: 3600 + task: deadly_corridor + deadly_corridor_loss_type: null + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: deadly_corridor +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: true + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_deadly_corridor_fix_latency_0 +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/hydra/.hydra/hydra.yaml b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..c54ff95bd6828fdece14f909cc9ad73d09be8764 --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/hydra/.hydra/hydra.yaml @@ -0,0 +1,252 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=deadly_corridor + - init=bridge + - mode=single + - run_id=openvla_deadly_corridor_fix_latency_0 + - run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=true + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=500 + - trainer.num_warmup_steps=100 + - trainer.save_interval=50 + - trainer.eval_interval=100 + - trainer.eval_num_batches=20 + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=16 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9, 0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=7 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_0 + - datasets.vla_data.data_mix=deadly_corridor_train__bridge + - datasets.vla_data.eval_data_mix=deadly_corridor_train__bridge__val + - framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=deadly_corridor + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=4 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - rl_games.env_eval.task_description='You are playing Deadly Corridor in VizDoom. + Choose actions from MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, + TURN_RIGHT, ATTACK.' + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_0/deadly_corridor_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=true + - rl_games.env_eval.mid_train.interval_steps=50 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[0] + - rl_games.env_eval.post_train.enabled=true + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,2,4,6,8] + - rl_games.env_eval.deadly.action_layout=multibinary_7 + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=true,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=deadly_corridor_train__bridge,datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_0,datasets.vla_data.eval_data_mix=deadly_corridor_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.video_backend=torchvision_av,env=deadly_corridor,framework.action_model.action_dim=7,framework.action_model.action_env_dim=7,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.deadly.action_layout=multibinary_7,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=4,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_0/deadly_corridor_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=true,rl_games.env_eval.mid_train.interval_steps=50,rl_games.env_eval.mid_train.latencies=[0],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=true,rl_games.env_eval.post_train.latencies=[0,2,4,6,8],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_description='You + are playing Deadly Corridor in VizDoom. Choose actions from MOVE_FORWARD, MOVE_BACKWARD, + MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, ATTACK.',rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=deadly_corridor,run_id=openvla_deadly_corridor_fix_latency_0,run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=20,trainer.gradient_accumulation_steps=16,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=500,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9, + 0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=50,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.2 + version_base: '1.1' + cwd: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: deadly_corridor + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/hydra/.hydra/overrides.yaml b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..86cfbac090736746048a9953fc8405c1541e6a67 --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/hydra/.hydra/overrides.yaml @@ -0,0 +1,89 @@ +- model=openvla +- env=deadly_corridor +- init=bridge +- mode=single +- run_id=openvla_deadly_corridor_fix_latency_0 +- run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=true +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=500 +- trainer.num_warmup_steps=100 +- trainer.save_interval=50 +- trainer.eval_interval=100 +- trainer.eval_num_batches=20 +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=16 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9, 0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=7 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_0 +- datasets.vla_data.data_mix=deadly_corridor_train__bridge +- datasets.vla_data.eval_data_mix=deadly_corridor_train__bridge__val +- framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=deadly_corridor +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=4 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- rl_games.env_eval.task_description='You are playing Deadly Corridor in VizDoom. + Choose actions from MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, + TURN_RIGHT, ATTACK.' +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/deadly_corridor_fix_latency_0/deadly_corridor_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=true +- rl_games.env_eval.mid_train.interval_steps=50 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[0] +- rl_games.env_eval.post_train.enabled=true +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,2,4,6,8] +- rl_games.env_eval.deadly.action_layout=multibinary_7 diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/hydra/train_starvla_hydra.log b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..6820cb08dfc8b64a45693ea2f7abea64667d5e27 --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/hydra/train_starvla_hydra.log @@ -0,0 +1,731 @@ +[2026-06-10 15:18:53,011][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-10 15:18:53,023][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-10 15:20:03,411][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-10 15:20:03,602][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `deadly_corridor_train__bridge` +[2026-06-10 15:20:18,526][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-10 15:20:18,527][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-10 15:20:18,527][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-10 15:20:18,527][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-10 15:20:18,527][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-10 15:20:18,527][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-10 15:20:18,527][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-10 15:20:18,527][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-10 15:20:18,528][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-10 15:20:18,528][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-10 15:20:18,528][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-10 15:20:18,528][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-10 15:20:18,528][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-10 15:20:18,528][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-10 15:20:18,528][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-10 15:20:18,528][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-10 15:20:18,528][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-10 15:20:18,528][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-10 15:20:18,528][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-10 15:20:18,532][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-10 15:20:18,532][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-10 15:20:18,532][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-10 15:20:18,533][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-10 15:20:18,533][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-10 15:20:18,533][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-10 15:20:18,533][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-10 15:20:18,533][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-10 15:20:18,533][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-10 15:20:18,533][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-10 15:20:18,533][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-10 15:20:18,533][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-10 15:20:18,549][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `deadly_corridor_train__bridge__val` +[2026-06-10 15:20:18,581][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/config.full.yaml +[2026-06-10 15:20:18,590][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/config.yaml +[2026-06-10 15:20:18,590][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-10 15:20:21,589][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-10 15:20:21,590][starVLA.training.train_starvla][INFO] - Total optimization steps = 100 +[2026-06-10 15:20:21,590][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-10 15:20:21,591][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-10 15:20:21,591][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-10 15:20:47,134][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 1.9788066148757935, 'train/grad_norm_pre_clip': 3.9875993728637695, 'timing/data': 0.028708968311548233, 'timing/model': 0.9732372872531414, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.01}) +[2026-06-10 15:22:14,712][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-10 15:22:14,801][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-10 15:23:03,087][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-10 15:23:03,189][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `deadly_corridor_train__bridge` +[2026-06-10 15:23:12,131][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-10 15:23:12,132][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-10 15:23:12,132][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-10 15:23:12,132][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-10 15:23:12,132][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-10 15:23:12,132][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-10 15:23:12,132][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-10 15:23:12,132][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-10 15:23:12,132][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-10 15:23:12,132][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-10 15:23:12,133][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-10 15:23:12,133][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-10 15:23:12,133][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-10 15:23:12,133][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-10 15:23:12,133][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-10 15:23:12,133][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-10 15:23:12,133][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-10 15:23:12,133][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-10 15:23:12,133][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-10 15:23:12,133][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-10 15:23:12,133][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-10 15:23:12,133][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-10 15:23:12,134][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-10 15:23:12,134][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-10 15:23:12,134][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-10 15:23:12,134][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-10 15:23:12,134][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-10 15:23:12,134][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-10 15:23:12,134][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-10 15:23:12,134][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-10 15:23:12,134][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-10 15:23:12,148][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `deadly_corridor_train__bridge__val` +[2026-06-10 15:23:12,180][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/config.full.yaml +[2026-06-10 15:23:12,189][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/config.yaml +[2026-06-10 15:23:12,189][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-10 15:23:14,814][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-10 15:23:14,815][starVLA.training.train_starvla][INFO] - Total optimization steps = 500 +[2026-06-10 15:23:14,815][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-10 15:23:14,815][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-10 15:23:14,815][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-10 15:23:43,302][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 1.9788066148757935, 'train/grad_norm_pre_clip': 3.9875993728637695, 'timing/data': 0.02197420969605446, 'timing/model': 0.9912614896893501, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.01}) +[2026-06-10 15:24:05,287][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 1.984902262687683, 'train/grad_norm_pre_clip': 3.846395254135132, 'timing/data': 0.021192185580730438, 'timing/model': 0.9585333280265331, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.02}) +[2026-06-10 15:24:23,611][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 1.9619669914245605, 'train/grad_norm_pre_clip': 3.8226375579833984, 'timing/data': 0.026076242327690125, 'timing/model': 0.9859746247529984, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.03}) +[2026-06-10 15:24:42,943][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 1.9251245260238647, 'train/grad_norm_pre_clip': 3.786646604537964, 'timing/data': 0.01801222562789917, 'timing/model': 1.1208380684256554, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.03}) +[2026-06-10 15:25:08,372][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 1.8633521795272827, 'train/grad_norm_pre_clip': 4.036754131317139, 'timing/data': 0.024703308939933777, 'timing/model': 0.9157196842133999, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.04}) +[2026-06-10 15:25:22,515][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 1.7833553552627563, 'train/grad_norm_pre_clip': 3.9722530841827393, 'timing/data': 0.021748770028352737, 'timing/model': 0.9533594362437725, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.05}) +[2026-06-10 15:25:36,835][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 1.6956206560134888, 'train/grad_norm_pre_clip': 3.8903701305389404, 'timing/data': 0.01597696542739868, 'timing/model': 0.9003253690898418, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.06}) +[2026-06-10 15:25:44,439][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 1.6694185733795166, 'train/grad_norm_pre_clip': 3.1637697219848633, 'timing/data': 0.09691942110657692, 'timing/model': 0.8495186194777489, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.07}) +[2026-06-10 15:26:01,279][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 1.558558464050293, 'train/grad_norm_pre_clip': 3.09921932220459, 'timing/data': 0.03532840311527252, 'timing/model': 0.9970584623515606, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.07}) +[2026-06-10 15:26:18,601][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 1.4699115753173828, 'train/grad_norm_pre_clip': 2.8225200176239014, 'timing/data': 0.042317718267440796, 'timing/model': 1.0458937995135784, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.08}) +[2026-06-10 15:26:33,698][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 1.1698429584503174, 'train/grad_norm_pre_clip': 3.432187557220459, 'timing/data': 0.04473726823925972, 'timing/model': 1.0907892733812332, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.09}) +[2026-06-10 15:26:51,823][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 1.0174669027328491, 'train/grad_norm_pre_clip': 3.197906970977783, 'timing/data': 0.03774358704686165, 'timing/model': 0.9459987208247185, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.1}) +[2026-06-10 15:27:15,392][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.9549033641815186, 'train/grad_norm_pre_clip': 2.5877349376678467, 'timing/data': 0.05666849762201309, 'timing/model': 0.9817543439567089, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.11}) +[2026-06-10 15:27:30,917][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.8232914209365845, 'train/grad_norm_pre_clip': 2.238218307495117, 'timing/data': 0.0326429158449173, 'timing/model': 1.0618696846067905, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.12}) +[2026-06-10 15:27:48,468][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.5743269920349121, 'train/grad_norm_pre_clip': 2.172626256942749, 'timing/data': 0.04500633478164673, 'timing/model': 0.9204036816954613, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.12}) +[2026-06-10 15:27:56,375][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.4479828178882599, 'train/grad_norm_pre_clip': 1.7975106239318848, 'timing/data': 0.0760183297097683, 'timing/model': 0.9546473920345306, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.13}) +[2026-06-10 15:28:22,811][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.49994754791259766, 'train/grad_norm_pre_clip': 1.1527743339538574, 'timing/data': 0.010017160326242447, 'timing/model': 1.0651339180767536, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.14}) +[2026-06-10 15:28:39,024][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.4280180335044861, 'train/grad_norm_pre_clip': 0.8209560513496399, 'timing/data': 0.010310035198926926, 'timing/model': 1.1162743866443634, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.15}) +[2026-06-10 15:28:55,091][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.37715059518814087, 'train/grad_norm_pre_clip': 0.5505267977714539, 'timing/data': 0.007697049528360367, 'timing/model': 1.0728456899523735, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.16}) +[2026-06-10 15:29:10,380][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.46666157245635986, 'train/grad_norm_pre_clip': 0.27843722701072693, 'timing/data': 0.008628502488136292, 'timing/model': 0.9470438547432423, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.17}) +[2026-06-10 15:29:25,826][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.696035623550415, 'train/grad_norm_pre_clip': 0.6244326829910278, 'timing/data': 0.011184055358171463, 'timing/model': 0.9182988181710243, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.17}) +[2026-06-10 15:29:44,309][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.43919044733047485, 'train/grad_norm_pre_clip': 0.40465664863586426, 'timing/data': 0.011629171669483185, 'timing/model': 1.1657019890844822, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.18}) +[2026-06-10 15:30:01,112][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.5872717499732971, 'train/grad_norm_pre_clip': 0.5326023101806641, 'timing/data': 0.010721493512392044, 'timing/model': 1.3655029609799385, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.19}) +[2026-06-10 15:30:13,480][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.6358269453048706, 'train/grad_norm_pre_clip': 0.7794426679611206, 'timing/data': 0.07772206887602806, 'timing/model': 1.0445353277027607, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.2}) +[2026-06-10 15:30:32,684][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.29207372665405273, 'train/grad_norm_pre_clip': 0.26979562640190125, 'timing/data': 0.009660106152296066, 'timing/model': 1.2694452442228794, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.21}) +[2026-06-10 15:30:55,188][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.40987974405288696, 'train/grad_norm_pre_clip': 0.48035725951194763, 'timing/data': 0.016072388738393784, 'timing/model': 1.048163291066885, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.22}) +[2026-06-10 15:31:11,406][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.39739251136779785, 'train/grad_norm_pre_clip': 0.468686580657959, 'timing/data': 0.011652093380689621, 'timing/model': 1.9434539712965488, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.23}) +[2026-06-10 15:31:27,536][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.4955506920814514, 'train/grad_norm_pre_clip': 0.65601646900177, 'timing/data': 0.008353807032108307, 'timing/model': 1.1930754110217094, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.23}) +[2026-06-10 15:31:47,282][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.36477622389793396, 'train/grad_norm_pre_clip': 0.41513386368751526, 'timing/data': 0.014943357557058334, 'timing/model': 1.5330508351325989, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.24}) +[2026-06-10 15:32:07,451][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.24055291712284088, 'train/grad_norm_pre_clip': 0.16919459402561188, 'timing/data': 0.009786151349544525, 'timing/model': 2.609679576009512, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.25}) +[2026-06-10 15:32:26,537][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.14234723150730133, 'train/grad_norm_pre_clip': 0.14339670538902283, 'timing/data': 0.01621805876493454, 'timing/model': 0.8992085792124271, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.26}) +[2026-06-10 15:32:34,123][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.3402043879032135, 'train/grad_norm_pre_clip': 0.29666590690612793, 'timing/data': 0.08378789946436882, 'timing/model': 0.899715717881918, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.27}) +[2026-06-10 15:32:51,544][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.22537937760353088, 'train/grad_norm_pre_clip': 0.18112432956695557, 'timing/data': 0.006143264472484589, 'timing/model': 1.112356923520565, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.28}) +[2026-06-10 15:33:15,827][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.651395857334137, 'train/grad_norm_pre_clip': 1.1428489685058594, 'timing/data': 0.013307183980941772, 'timing/model': 1.6662195287644863, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.28}) +[2026-06-10 15:33:30,729][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.24528509378433228, 'train/grad_norm_pre_clip': 0.5501434206962585, 'timing/data': 0.006583858281373978, 'timing/model': 1.3121153712272644, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.29}) +[2026-06-10 15:33:45,924][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.2134152352809906, 'train/grad_norm_pre_clip': 0.8865905404090881, 'timing/data': 0.0074003636837005615, 'timing/model': 1.0926201790571213, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.3}) +[2026-06-10 15:34:00,694][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.45546025037765503, 'train/grad_norm_pre_clip': 0.73940110206604, 'timing/data': 0.00866866484284401, 'timing/model': 1.0096570923924446, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.31}) +[2026-06-10 15:34:16,828][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.35200589895248413, 'train/grad_norm_pre_clip': 2.2211692333221436, 'timing/data': 0.015472330152988434, 'timing/model': 0.8921575360000134, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.32}) +[2026-06-10 15:34:32,971][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.33383673429489136, 'train/grad_norm_pre_clip': 0.30590614676475525, 'timing/data': 0.016079183667898178, 'timing/model': 0.9474889822304249, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.33}) +[2026-06-10 15:34:40,325][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.1664348691701889, 'train/grad_norm_pre_clip': 0.6235311627388, 'timing/data': 0.081674013286829, 'timing/model': 0.8730183430016041, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.33}) +[2026-06-10 15:34:58,304][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.21701151132583618, 'train/grad_norm_pre_clip': 0.3225254714488983, 'timing/data': 0.00048731639981269836, 'timing/model': 1.3458871990442276, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.34}) +[2026-06-10 15:35:14,203][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.1577344685792923, 'train/grad_norm_pre_clip': 0.44193416833877563, 'timing/data': 0.0005171969532966614, 'timing/model': 0.9522372931241989, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.35}) +[2026-06-10 15:35:31,641][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.18630550801753998, 'train/grad_norm_pre_clip': 0.5931656360626221, 'timing/data': 0.0010147467255592346, 'timing/model': 0.9310105443000793, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.36}) +[2026-06-10 15:35:47,320][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.03858888894319534, 'train/grad_norm_pre_clip': 0.29128748178482056, 'timing/data': 0.0006676726043224335, 'timing/model': 0.9079652391374111, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.37}) +[2026-06-10 15:36:04,558][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.15090413391590118, 'train/grad_norm_pre_clip': 0.4947083592414856, 'timing/data': 0.000692903995513916, 'timing/model': 0.9426087737083435, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.38}) +[2026-06-10 15:36:19,262][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.09237789362668991, 'train/grad_norm_pre_clip': 0.3136845529079437, 'timing/data': 0.0008806213736534119, 'timing/model': 0.9074129052460194, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.38}) +[2026-06-10 15:36:35,309][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.3520515561103821, 'train/grad_norm_pre_clip': 0.5719059109687805, 'timing/data': 0.0006783194839954376, 'timing/model': 2.000288724899292, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.39}) +[2026-06-10 15:36:45,651][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.06972038745880127, 'train/grad_norm_pre_clip': 0.3864014744758606, 'timing/data': 0.07822723314166069, 'timing/model': 1.8264606557786465, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.4}) +[2026-06-10 15:37:08,079][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.038813576102256775, 'train/grad_norm_pre_clip': 1.0808757543563843, 'timing/data': 0.03499552235007286, 'timing/model': 1.0153570659458637, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.41}) +[2026-06-10 15:37:25,874][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpev2gyvnf/test.c -o /tmp/tmpev2gyvnf/test.o +[2026-06-10 15:37:25,887][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpev2gyvnf/test.o -laio -o /tmp/tmpev2gyvnf/a.out +[2026-06-10 15:37:26,257][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpfc_zbw1j/test.c -o /tmp/tmpfc_zbw1j/test.o +[2026-06-10 15:37:26,271][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpfc_zbw1j/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpfc_zbw1j/a.out +[2026-06-10 15:37:26,306][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpr6j22zxf/test.c -o /tmp/tmpr6j22zxf/test.o +[2026-06-10 15:37:26,320][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpr6j22zxf/test.o -laio -o /tmp/tmpr6j22zxf/a.out +[2026-06-10 15:37:32,569][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state +[2026-06-10 15:37:32,576][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 15:37:39,141][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/model.safetensors +[2026-06-10 15:37:58,519][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/optimizer.bin +[2026-06-10 15:37:58,520][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/sampler.bin +[2026-06-10 15:37:58,520][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/sampler_1.bin +[2026-06-10 15:37:58,524][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/random_states_0.pkl +[2026-06-10 15:37:58,540][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-10 15:37:58,541][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 50 +[2026-06-10 15:37:58,542][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.28102752566337585, 'train/grad_norm_pre_clip': 0.36807963252067566, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 253.8692138671875, 'rl_games_eval/mid_train/mean_length': 19.8, 'rl_games_eval/mid_train/std_reward': 246.1225268303482, 'rl_games_eval/mid_train/std_length': 9.410632284814874, 'rl_games_eval/mid_train/macro_mean_reward': 253.8692138671875, 'rl_games_eval/mid_train/macro_mean_length': 19.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 253.8692138671875, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 19.8, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 246.1225268303482, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 9.410632284814874, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 253.8692138671875, 'checkpoint/best_score': 253.8692138671875, 'checkpoint/best_step': 50.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.03743337094783783, 'timing/model': 0.9746863581240177, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.42}) +[2026-06-10 15:37:58,542][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_50_state +[2026-06-10 15:37:58,547][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 15:38:07,553][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_50_state/model.safetensors +[2026-06-10 15:38:26,599][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_50_state/optimizer.bin +[2026-06-10 15:38:26,600][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_50_state/sampler.bin +[2026-06-10 15:38:26,600][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_50_state/sampler_1.bin +[2026-06-10 15:38:26,603][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_50_state/random_states_0.pkl +[2026-06-10 15:38:26,606][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 50 +[2026-06-10 15:38:26,606][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 15:38:26,619][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 15:38:41,472][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.010070119053125381, 'train/grad_norm_pre_clip': 0.0737743228673935, 'timing/data': 0.04341931641101837, 'timing/model': 0.9229139164090157, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.42}) +[2026-06-10 15:38:58,290][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.2412431687116623, 'train/grad_norm_pre_clip': 0.45026201009750366, 'timing/data': 0.03234364092350006, 'timing/model': 0.9427496455609798, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.43}) +[2026-06-10 15:39:17,117][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.22861531376838684, 'train/grad_norm_pre_clip': 0.5169709920883179, 'timing/data': 0.03576935455203056, 'timing/model': 0.9613229222595692, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.44}) +[2026-06-10 15:39:34,802][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.22530058026313782, 'train/grad_norm_pre_clip': 0.36979940533638, 'timing/data': 0.04700605571269989, 'timing/model': 1.6882551088929176, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.45}) +[2026-06-10 15:39:51,745][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.006849745754152536, 'train/grad_norm_pre_clip': 0.07729054987430573, 'timing/data': 0.046981826424598694, 'timing/model': 1.1137421019375324, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.46}) +[2026-06-10 15:40:04,638][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.008267250843346119, 'train/grad_norm_pre_clip': 0.08751776069402695, 'timing/data': 0.0987647995352745, 'timing/model': 0.9113792665302753, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.47}) +[2026-06-10 15:40:24,903][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.010133934207260609, 'train/grad_norm_pre_clip': 0.1116509661078453, 'timing/data': 0.040289901196956635, 'timing/model': 0.9287855699658394, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.47}) +[2026-06-10 15:40:43,429][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.01128101721405983, 'train/grad_norm_pre_clip': 0.11137095093727112, 'timing/data': 0.11308883503079414, 'timing/model': 1.1115896552801132, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.48}) +[2026-06-10 15:41:01,506][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.15189334750175476, 'train/grad_norm_pre_clip': 0.4234163165092468, 'timing/data': 0.03881365433335304, 'timing/model': 0.9278313182294369, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.49}) +[2026-06-10 15:41:22,083][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.011659258976578712, 'train/grad_norm_pre_clip': 0.12494995445013046, 'timing/data': 0.045398566871881485, 'timing/model': 0.9932672716677189, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.5}) +[2026-06-10 15:41:37,708][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.012528747320175171, 'train/grad_norm_pre_clip': 0.13321688771247864, 'timing/data': 0.04096738249063492, 'timing/model': 0.929320927709341, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.51}) +[2026-06-10 15:41:53,186][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.10221609473228455, 'train/grad_norm_pre_clip': 0.3716205954551697, 'timing/data': 0.08590171858668327, 'timing/model': 0.9567261002957821, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.52}) +[2026-06-10 15:42:13,676][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.10981927067041397, 'train/grad_norm_pre_clip': 0.6879978775978088, 'timing/data': 0.03144390136003494, 'timing/model': 1.0210182294249535, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.53}) +[2026-06-10 15:42:21,435][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.00941409170627594, 'train/grad_norm_pre_clip': 0.16146749258041382, 'timing/data': 0.16958466172218323, 'timing/model': 0.8116102367639542, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.53}) +[2026-06-10 15:42:40,102][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.01445361040532589, 'train/grad_norm_pre_clip': 0.29169270396232605, 'timing/data': 0.04079343378543854, 'timing/model': 0.9334428980946541, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.54}) +[2026-06-10 15:42:56,499][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.00523423170670867, 'train/grad_norm_pre_clip': 0.2432413250207901, 'timing/data': 0.04441145434975624, 'timing/model': 0.94939124584198, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.55}) +[2026-06-10 15:43:15,614][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.0007805610657669604, 'train/grad_norm_pre_clip': 0.15366840362548828, 'timing/data': 0.03097052499651909, 'timing/model': 1.1212563067674637, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.56}) +[2026-06-10 15:43:39,789][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 7.476790779037401e-05, 'train/grad_norm_pre_clip': 0.009443718940019608, 'timing/data': 0.01750851422548294, 'timing/model': 0.9292925521731377, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.57}) +[2026-06-10 15:43:55,030][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 2.8945444228156703e-06, 'train/grad_norm_pre_clip': 9.011146175907925e-05, 'timing/data': 0.023544281721115112, 'timing/model': 1.0400233007967472, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.57}) +[2026-06-10 15:44:11,830][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 4.440404154593125e-06, 'train/grad_norm_pre_clip': 0.0006495593697763979, 'timing/data': 0.027812622487545013, 'timing/model': 1.385073859244585, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.58}) +[2026-06-10 15:44:32,217][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 1.6726520470911055e-06, 'train/grad_norm_pre_clip': 8.313306170748547e-05, 'timing/data': 0.031497787684202194, 'timing/model': 0.9568302407860756, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.59}) +[2026-06-10 15:44:39,292][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 2.7971143481408944e-06, 'train/grad_norm_pre_clip': 0.00047298683784902096, 'timing/data': 0.06716939434409142, 'timing/model': 0.8095788508653641, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.6}) +[2026-06-10 15:44:58,224][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.5254154205322266, 'train/grad_norm_pre_clip': 1.6620646715164185, 'timing/data': 0.007019329816102982, 'timing/model': 1.2559517696499825, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.61}) +[2026-06-10 15:45:14,232][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 3.0621652058471227e-06, 'train/grad_norm_pre_clip': 0.00031606946140527725, 'timing/data': 0.009522907435894012, 'timing/model': 0.9562090113759041, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.62}) +[2026-06-10 15:45:32,791][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 7.118897883628961e-06, 'train/grad_norm_pre_clip': 0.00021663503139279783, 'timing/data': 0.019453290849924088, 'timing/model': 1.1241406723856926, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.62}) +[2026-06-10 15:45:50,868][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 9.92031073110411e-06, 'train/grad_norm_pre_clip': 0.00015523980255238712, 'timing/data': 0.015715107321739197, 'timing/model': 1.0344116352498531, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.63}) +[2026-06-10 15:46:09,839][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 8.98009748198092e-05, 'train/grad_norm_pre_clip': 0.012979437597095966, 'timing/data': 0.0011735521256923676, 'timing/model': 1.0595159083604813, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.64}) +[2026-06-10 15:46:31,202][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.00014711153926327825, 'train/grad_norm_pre_clip': 0.004714797250926495, 'timing/data': 0.000550735741853714, 'timing/model': 0.9987432770431042, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.65}) +[2026-06-10 15:46:47,235][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.0002591654483694583, 'train/grad_norm_pre_clip': 0.008327782154083252, 'timing/data': 0.0005125030875205994, 'timing/model': 0.911451943218708, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.66}) +[2026-06-10 15:46:54,710][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.15386909246444702, 'train/grad_norm_pre_clip': 5.986186504364014, 'timing/data': 0.1425173282623291, 'timing/model': 0.9503746889531612, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.67}) +[2026-06-10 15:47:12,233][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.0025611661840230227, 'train/grad_norm_pre_clip': 0.09506479650735855, 'timing/data': 0.008064594119787216, 'timing/model': 0.8099745847284794, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.68}) +[2026-06-10 15:47:29,188][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.027741825208067894, 'train/grad_norm_pre_clip': 0.47285518050193787, 'timing/data': 0.014586564153432846, 'timing/model': 1.0680318213999271, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.68}) +[2026-06-10 15:47:50,883][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.0038665849715471268, 'train/grad_norm_pre_clip': 0.11805582791566849, 'timing/data': 0.012093666940927505, 'timing/model': 0.9761078059673309, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.69}) +[2026-06-10 15:48:12,716][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.12887758016586304, 'train/grad_norm_pre_clip': 3.181051015853882, 'timing/data': 0.011245951056480408, 'timing/model': 0.9605390653014183, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.7}) +[2026-06-10 15:48:28,849][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.0029887454584240913, 'train/grad_norm_pre_clip': 0.09223460406064987, 'timing/data': 0.008656084537506104, 'timing/model': 0.9769876897335052, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.71}) +[2026-06-10 15:48:44,715][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.026809891685843468, 'train/grad_norm_pre_clip': 0.6832626461982727, 'timing/data': 0.040004704147577286, 'timing/model': 0.9870444424450397, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.72}) +[2026-06-10 15:49:01,511][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.0009577288292348385, 'train/grad_norm_pre_clip': 0.04506171867251396, 'timing/data': 0.014660175889730453, 'timing/model': 0.8322942219674587, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.72}) +[2026-06-10 15:49:10,201][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.08137498795986176, 'train/grad_norm_pre_clip': 2.1735494136810303, 'timing/data': 0.10913987830281258, 'timing/model': 0.983252264559269, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.73}) +[2026-06-10 15:49:30,906][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.11437226831912994, 'train/grad_norm_pre_clip': 4.593682765960693, 'timing/data': 0.013359040021896362, 'timing/model': 1.420888863503933, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.74}) +[2026-06-10 15:49:52,978][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.00041349491220898926, 'train/grad_norm_pre_clip': 0.047538481652736664, 'timing/data': 0.011404406279325485, 'timing/model': 0.9400507397949696, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.75}) +[2026-06-10 15:50:09,097][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.0003038233262486756, 'train/grad_norm_pre_clip': 0.017123470082879066, 'timing/data': 0.009293738752603531, 'timing/model': 0.9407185763120651, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.76}) +[2026-06-10 15:50:25,302][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.23709839582443237, 'train/grad_norm_pre_clip': 5.150082111358643, 'timing/data': 0.011570274829864502, 'timing/model': 0.9745735824108124, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.77}) +[2026-06-10 15:50:46,762][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.04276006668806076, 'train/grad_norm_pre_clip': 1.5687869787216187, 'timing/data': 0.0005572699010372162, 'timing/model': 0.9444543197751045, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.78}) +[2026-06-10 15:51:04,073][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.009318494237959385, 'train/grad_norm_pre_clip': 0.4211757183074951, 'timing/data': 0.0005524344742298126, 'timing/model': 0.9243654869496822, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.78}) +[2026-06-10 15:51:19,623][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.005339320283383131, 'train/grad_norm_pre_clip': 0.21561752259731293, 'timing/data': 0.0005757473409175873, 'timing/model': 1.1935950554907322, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.79}) +[2026-06-10 15:51:27,025][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.0645107850432396, 'train/grad_norm_pre_clip': 2.031010866165161, 'timing/data': 0.07359044998884201, 'timing/model': 0.8874627687036991, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.8}) +[2026-06-10 15:51:48,494][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.00016725965542718768, 'train/grad_norm_pre_clip': 0.039063237607479095, 'timing/data': 0.029337633401155472, 'timing/model': 1.6487880609929562, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.81}) +[2026-06-10 15:52:09,094][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 1.0617055750117288e-06, 'train/grad_norm_pre_clip': 3.626268880907446e-05, 'timing/data': 0.0369880385696888, 'timing/model': 1.1420473270118237, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.82}) +[2026-06-10 15:52:27,257][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.13473400473594666, 'train/grad_norm_pre_clip': 3.852447271347046, 'timing/data': 0.01570645347237587, 'timing/model': 0.9116411283612251, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.82}) +[2026-06-10 15:53:07,978][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state +[2026-06-10 15:53:07,984][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 15:53:15,718][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/model.safetensors +[2026-06-10 15:53:33,460][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/optimizer.bin +[2026-06-10 15:53:33,461][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/sampler.bin +[2026-06-10 15:53:33,461][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/sampler_1.bin +[2026-06-10 15:53:33,464][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/random_states_0.pkl +[2026-06-10 15:53:33,504][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-10 15:53:33,504][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 100 +[2026-06-10 15:53:33,506][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 6.5969616116490215e-06, 'train/grad_norm_pre_clip': 0.0005228036898188293, 'eval/loss': 0.001292500994168222, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 315.4344055175781, 'rl_games_eval/mid_train/mean_length': 19.4, 'rl_games_eval/mid_train/std_reward': 292.20681699085446, 'rl_games_eval/mid_train/std_length': 8.73155198117723, 'rl_games_eval/mid_train/macro_mean_reward': 315.4344055175781, 'rl_games_eval/mid_train/macro_mean_length': 19.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 315.4344055175781, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 19.4, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 292.20681699085446, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 8.73155198117723, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 315.4344055175781, 'checkpoint/best_score': 315.4344055175781, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.02757478505373001, 'timing/model': 0.9219875335693359, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.83}) +[2026-06-10 15:53:33,507][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_100_state +[2026-06-10 15:53:33,512][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 15:53:40,830][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_100_state/model.safetensors +[2026-06-10 15:53:57,896][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_100_state/optimizer.bin +[2026-06-10 15:53:57,898][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_100_state/sampler.bin +[2026-06-10 15:53:57,898][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_100_state/sampler_1.bin +[2026-06-10 15:53:57,902][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-10 15:53:57,909][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-10 15:53:57,909][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 15:53:57,922][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 15:54:23,196][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.00024423093418590724, 'train/grad_norm_pre_clip': 0.04544069990515709, 'timing/data': 0.02735406905412674, 'timing/model': 1.914630401879549, 'learning_rate/qwen_vl_interface': 9.999853498812753e-06, 'learning_rate/action_model': 9.999853498812753e-05, 'epoch': 0.84}) +[2026-06-10 15:54:53,384][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 6.377229510690086e-06, 'train/grad_norm_pre_clip': 0.00043091410771012306, 'timing/data': 0.023974332958459854, 'timing/model': 1.3820914402604103, 'learning_rate/qwen_vl_interface': 9.999414004287889e-06, 'learning_rate/action_model': 9.999414004287888e-05, 'epoch': 0.85}) +[2026-06-10 15:55:16,287][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 1.0877828344746376e-06, 'train/grad_norm_pre_clip': 6.411215872503817e-05, 'timing/data': 0.01865563914179802, 'timing/model': 1.3955721445381641, 'learning_rate/qwen_vl_interface': 9.998681543535504e-06, 'learning_rate/action_model': 9.998681543535502e-05, 'epoch': 0.86}) +[2026-06-10 15:55:29,916][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.07387027144432068, 'train/grad_norm_pre_clip': 0.35869500041007996, 'timing/data': 0.08555319532752037, 'timing/model': 1.0580261200666428, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.87}) +[2026-06-10 15:55:55,128][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 1.13461110231583e-05, 'train/grad_norm_pre_clip': 0.0008489939500577748, 'timing/data': 0.030113358050584793, 'timing/model': 1.4273577071726322, 'learning_rate/qwen_vl_interface': 9.996337922143435e-06, 'learning_rate/action_model': 9.996337922143436e-05, 'epoch': 0.88}) +[2026-06-10 15:56:17,270][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 3.829539309663232e-06, 'train/grad_norm_pre_clip': 0.0003565681690815836, 'timing/data': 0.0399303212761879, 'timing/model': 3.4206362031400204, 'learning_rate/qwen_vl_interface': 9.994726906069359e-06, 'learning_rate/action_model': 9.994726906069358e-05, 'epoch': 0.88}) +[2026-06-10 15:56:40,852][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.0006117774755693972, 'train/grad_norm_pre_clip': 0.12405721843242645, 'timing/data': 0.03618897125124931, 'timing/model': 1.1627715677022934, 'learning_rate/qwen_vl_interface': 9.992823212890057e-06, 'learning_rate/action_model': 9.992823212890057e-05, 'epoch': 0.89}) +[2026-06-10 15:57:01,726][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.00029576075030490756, 'train/grad_norm_pre_clip': 0.03121166117489338, 'timing/data': 0.026949476450681686, 'timing/model': 1.4383339956402779, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.9}) +[2026-06-10 15:57:24,706][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 5.617593615170335e-06, 'train/grad_norm_pre_clip': 0.0003647626144811511, 'timing/data': 0.02449582889676094, 'timing/model': 1.2191822305321693, 'learning_rate/qwen_vl_interface': 9.988138282977284e-06, 'learning_rate/action_model': 9.988138282977283e-05, 'epoch': 0.91}) +[2026-06-10 15:57:45,069][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.020371995866298676, 'train/grad_norm_pre_clip': 0.6198657751083374, 'timing/data': 0.027280263602733612, 'timing/model': 1.2014424614608288, 'learning_rate/qwen_vl_interface': 9.985357335232359e-06, 'learning_rate/action_model': 9.98535733523236e-05, 'epoch': 0.92}) +[2026-06-10 15:58:06,558][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 4.421855919645168e-06, 'train/grad_norm_pre_clip': 0.0002897787489928305, 'timing/data': 0.02397330477833748, 'timing/model': 1.1732790060341358, 'learning_rate/qwen_vl_interface': 9.982284288341471e-06, 'learning_rate/action_model': 9.982284288341471e-05, 'epoch': 0.93}) +[2026-06-10 15:58:19,965][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 6.484004188678227e-06, 'train/grad_norm_pre_clip': 0.0005267685046419501, 'timing/data': 0.09478116035461426, 'timing/model': 1.197404656559229, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.93}) +[2026-06-10 15:58:46,318][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.09955533593893051, 'train/grad_norm_pre_clip': 5.243204116821289, 'timing/data': 0.03336755558848381, 'timing/model': 1.3386724255979061, 'learning_rate/qwen_vl_interface': 9.975262673368201e-06, 'learning_rate/action_model': 9.9752626733682e-05, 'epoch': 0.94}) +[2026-06-10 15:59:11,431][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.00010354509868193418, 'train/grad_norm_pre_clip': 0.005540991667658091, 'timing/data': 0.034758880734443665, 'timing/model': 1.0545343942940235, 'learning_rate/qwen_vl_interface': 9.971314538412104e-06, 'learning_rate/action_model': 9.971314538412104e-05, 'epoch': 0.95}) +[2026-06-10 15:59:31,001][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 2.826020136126317e-05, 'train/grad_norm_pre_clip': 0.0017628369387239218, 'timing/data': 0.03362403064966202, 'timing/model': 1.3414536006748676, 'learning_rate/qwen_vl_interface': 9.967075170535901e-06, 'learning_rate/action_model': 9.967075170535901e-05, 'epoch': 0.96}) +[2026-06-10 15:59:50,737][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.006960802245885134, 'train/grad_norm_pre_clip': 0.28212112188339233, 'timing/data': 0.033400703221559525, 'timing/model': 1.2363849394023418, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.97}) +[2026-06-10 16:00:12,886][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 1.4043605915503576e-05, 'train/grad_norm_pre_clip': 0.0011872497852891684, 'timing/data': 0.024877376854419708, 'timing/model': 1.2340394034981728, 'learning_rate/qwen_vl_interface': 9.95772379998838e-06, 'learning_rate/action_model': 9.957723799988381e-05, 'epoch': 0.97}) +[2026-06-10 16:00:36,920][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.09079711139202118, 'train/grad_norm_pre_clip': 0.7704293131828308, 'timing/data': 0.03380880132317543, 'timing/model': 1.2066193707287312, 'learning_rate/qwen_vl_interface': 9.95261237415365e-06, 'learning_rate/action_model': 9.952612374153649e-05, 'epoch': 0.98}) +[2026-06-10 16:00:57,603][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.09785588830709457, 'train/grad_norm_pre_clip': 1.1137292385101318, 'timing/data': 0.025953199714422226, 'timing/model': 1.1432650685310364, 'learning_rate/qwen_vl_interface': 9.9472108690364e-06, 'learning_rate/action_model': 9.947210869036399e-05, 'epoch': 0.99}) +[2026-06-10 16:01:08,005][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.010042034089565277, 'train/grad_norm_pre_clip': 0.45126938819885254, 'timing/data': 0.17732371762394905, 'timing/model': 1.2234034650027752, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 1.0}) +[2026-06-10 16:01:31,985][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.04917621240019798, 'train/grad_norm_pre_clip': 0.9863253831863403, 'timing/data': 0.013966303318738937, 'timing/model': 1.198629904538393, 'learning_rate/qwen_vl_interface': 9.935538971588355e-06, 'learning_rate/action_model': 9.935538971588353e-05, 'epoch': 1.01}) +[2026-06-10 16:01:57,635][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.017121870070695877, 'train/grad_norm_pre_clip': 0.4289666712284088, 'timing/data': 0.016467798501253128, 'timing/model': 1.487747985869646, 'learning_rate/qwen_vl_interface': 9.929269299235177e-06, 'learning_rate/action_model': 9.929269299235176e-05, 'epoch': 1.02}) +[2026-06-10 16:02:17,909][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.08671139925718307, 'train/grad_norm_pre_clip': 1.98575758934021, 'timing/data': 0.03093913197517395, 'timing/model': 1.6208303906023502, 'learning_rate/qwen_vl_interface': 9.922710987510297e-06, 'learning_rate/action_model': 9.922710987510296e-05, 'epoch': 1.02}) +[2026-06-10 16:02:37,046][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.00018703073146753013, 'train/grad_norm_pre_clip': 0.0795261338353157, 'timing/data': 0.02516007423400879, 'timing/model': 1.2031624540686607, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 1.03}) +[2026-06-10 16:02:58,085][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.037571899592876434, 'train/grad_norm_pre_clip': 0.22622130811214447, 'timing/data': 0.020551487803459167, 'timing/model': 1.2354349307715893, 'learning_rate/qwen_vl_interface': 9.908730081915346e-06, 'learning_rate/action_model': 9.908730081915345e-05, 'epoch': 1.04}) +[2026-06-10 16:03:19,284][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 3.926432782463962e-06, 'train/grad_norm_pre_clip': 0.0002999527205247432, 'timing/data': 0.01948460191488266, 'timing/model': 1.2222364395856857, 'learning_rate/qwen_vl_interface': 9.901308350453388e-06, 'learning_rate/action_model': 9.901308350453388e-05, 'epoch': 1.05}) +[2026-06-10 16:03:44,403][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.004708838649094105, 'train/grad_norm_pre_clip': 0.18999861180782318, 'timing/data': 0.026292938739061356, 'timing/model': 1.1632463969290257, 'learning_rate/qwen_vl_interface': 9.893599704382757e-06, 'learning_rate/action_model': 9.893599704382757e-05, 'epoch': 1.06}) +[2026-06-10 16:03:54,302][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.0007285833708010614, 'train/grad_norm_pre_clip': 0.08750729262828827, 'timing/data': 0.10075027868151665, 'timing/model': 1.1877397634088993, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 1.07}) +[2026-06-10 16:04:18,093][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 3.062129508180078e-06, 'train/grad_norm_pre_clip': 0.0001684478047536686, 'timing/data': 0.04643690958619118, 'timing/model': 1.594299454241991, 'learning_rate/qwen_vl_interface': 9.877323588106785e-06, 'learning_rate/action_model': 9.877323588106785e-05, 'epoch': 1.07}) +[2026-06-10 16:04:38,640][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 1.5884674212429672e-05, 'train/grad_norm_pre_clip': 0.0024660660419613123, 'timing/data': 0.02901218831539154, 'timing/model': 1.1751675121486187, 'learning_rate/qwen_vl_interface': 9.868757121888965e-06, 'learning_rate/action_model': 9.868757121888964e-05, 'epoch': 1.08}) +[2026-06-10 16:04:57,854][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.05078304931521416, 'train/grad_norm_pre_clip': 0.6490880250930786, 'timing/data': 0.03226125240325928, 'timing/model': 1.0980084538459778, 'learning_rate/qwen_vl_interface': 9.859905748975577e-06, 'learning_rate/action_model': 9.859905748975578e-05, 'epoch': 1.09}) +[2026-06-10 16:05:23,432][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 3.7625426330123446e-07, 'train/grad_norm_pre_clip': 1.0541873962210957e-05, 'timing/data': 0.03636186942458153, 'timing/model': 1.9718648716807365, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 1.1}) +[2026-06-10 16:05:46,980][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 2.8312203426139604e-07, 'train/grad_norm_pre_clip': 6.064970420993632e-06, 'timing/data': 0.028036028146743774, 'timing/model': 1.2459858022630215, 'learning_rate/qwen_vl_interface': 9.84135048458031e-06, 'learning_rate/action_model': 9.84135048458031e-05, 'epoch': 1.11}) +[2026-06-10 16:06:05,321][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.031041432172060013, 'train/grad_norm_pre_clip': 0.7542080283164978, 'timing/data': 0.028634071350097656, 'timing/model': 1.2257389724254608, 'learning_rate/qwen_vl_interface': 9.83164773767454e-06, 'learning_rate/action_model': 9.83164773767454e-05, 'epoch': 1.12}) +[2026-06-10 16:06:27,711][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 1.3112891110722558e-06, 'train/grad_norm_pre_clip': 0.00012025831529172137, 'timing/data': 0.02061529830098152, 'timing/model': 1.5169272311031818, 'learning_rate/qwen_vl_interface': 9.821662373154824e-06, 'learning_rate/action_model': 9.821662373154825e-05, 'epoch': 1.12}) +[2026-06-10 16:06:36,867][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.0003093864070251584, 'train/grad_norm_pre_clip': 0.06830228865146637, 'timing/data': 0.09140099212527275, 'timing/model': 1.0455899387598038, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 1.13}) +[2026-06-10 16:06:58,084][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.0036785684060305357, 'train/grad_norm_pre_clip': 0.7245531678199768, 'timing/data': 0.026330046355724335, 'timing/model': 1.3101280890405178, 'learning_rate/qwen_vl_interface': 9.800846272446017e-06, 'learning_rate/action_model': 9.800846272446018e-05, 'epoch': 1.14}) +[2026-06-10 16:07:19,495][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.1754082441329956, 'train/grad_norm_pre_clip': 1.481257677078247, 'timing/data': 0.035586219280958176, 'timing/model': 1.15020951628685, 'learning_rate/qwen_vl_interface': 9.790016820292068e-06, 'learning_rate/action_model': 9.790016820292069e-05, 'epoch': 1.15}) +[2026-06-10 16:07:38,212][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 2.756714252427628e-07, 'train/grad_norm_pre_clip': 4.105884272576077e-06, 'timing/data': 0.033100616186857224, 'timing/model': 1.2089399881660938, 'learning_rate/qwen_vl_interface': 9.778907318515255e-06, 'learning_rate/action_model': 9.778907318515255e-05, 'epoch': 1.16}) +[2026-06-10 16:07:57,112][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 9.103407137445174e-06, 'train/grad_norm_pre_clip': 0.001390731311403215, 'timing/data': 0.020532473921775818, 'timing/model': 1.1538574621081352, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 1.17}) +[2026-06-10 16:08:17,097][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.019658364355564117, 'train/grad_norm_pre_clip': 0.8252333998680115, 'timing/data': 0.03317926824092865, 'timing/model': 1.1692131869494915, 'learning_rate/qwen_vl_interface': 9.755850924471148e-06, 'learning_rate/action_model': 9.755850924471147e-05, 'epoch': 1.18}) +[2026-06-10 16:08:37,707][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 2.0489095220455056e-07, 'train/grad_norm_pre_clip': 1.993716750803287e-06, 'timing/data': 0.033216774463653564, 'timing/model': 1.3831142336130142, 'learning_rate/qwen_vl_interface': 9.74390545443084e-06, 'learning_rate/action_model': 9.74390545443084e-05, 'epoch': 1.18}) +[2026-06-10 16:08:56,387][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 3.650779945019167e-07, 'train/grad_norm_pre_clip': 3.893735265592113e-05, 'timing/data': 0.03532476723194122, 'timing/model': 1.1666334867477417, 'learning_rate/qwen_vl_interface': 9.73168277913392e-06, 'learning_rate/action_model': 9.73168277913392e-05, 'epoch': 1.19}) +[2026-06-10 16:09:05,591][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 4.8566245823167264e-05, 'train/grad_norm_pre_clip': 0.012982442043721676, 'timing/data': 0.11253345012664795, 'timing/model': 1.247328244149685, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 1.2}) +[2026-06-10 16:09:28,866][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 5.650670937029645e-05, 'train/grad_norm_pre_clip': 0.01166983786970377, 'timing/data': 0.02084709331393242, 'timing/model': 1.2052101604640484, 'learning_rate/qwen_vl_interface': 9.7064088456318e-06, 'learning_rate/action_model': 9.7064088456318e-05, 'epoch': 1.21}) +[2026-06-10 16:09:52,131][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.5467166900634766, 'train/grad_norm_pre_clip': 2.814803123474121, 'timing/data': 0.01511024683713913, 'timing/model': 1.163159605115652, 'learning_rate/qwen_vl_interface': 9.693359146441872e-06, 'learning_rate/action_model': 9.693359146441872e-05, 'epoch': 1.22}) +[2026-06-10 16:10:11,301][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 3.5017723121200106e-07, 'train/grad_norm_pre_clip': 6.6097504713980015e-06, 'timing/data': 0.027348600327968597, 'timing/model': 1.1697930544614792, 'learning_rate/qwen_vl_interface': 9.680035359929698e-06, 'learning_rate/action_model': 9.680035359929697e-05, 'epoch': 1.23}) +[2026-06-10 16:10:30,071][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.4272664189338684, 'train/grad_norm_pre_clip': 1.840806007385254, 'timing/data': 0.016451947391033173, 'timing/model': 1.1559171825647354, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 1.23}) +[2026-06-10 16:10:49,135][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.00010291615762980655, 'train/grad_norm_pre_clip': 0.0041757612489163876, 'timing/data': 0.019149668514728546, 'timing/model': 1.1645862758159637, 'learning_rate/qwen_vl_interface': 9.652568829290574e-06, 'learning_rate/action_model': 9.652568829290573e-05, 'epoch': 1.24}) +[2026-06-10 16:11:27,187][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 150 +[2026-06-10 16:11:27,198][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.0011450164020061493, 'train/grad_norm_pre_clip': 0.036294225603342056, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': -115.96701049804688, 'rl_games_eval/mid_train/mean_length': 31.8, 'rl_games_eval/mid_train/std_reward': 0.023152069682942776, 'rl_games_eval/mid_train/std_length': 8.930845424706442, 'rl_games_eval/mid_train/macro_mean_reward': -115.96701049804688, 'rl_games_eval/mid_train/macro_mean_length': 31.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': -115.96701049804688, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 31.8, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 0.023152069682942776, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 8.930845424706442, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': -115.96701049804688, 'checkpoint/best_score': 315.4344055175781, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.015441913157701492, 'timing/model': 3.4551178850233555, 'learning_rate/qwen_vl_interface': 9.638427779428613e-06, 'learning_rate/action_model': 9.638427779428613e-05, 'epoch': 1.25}) +[2026-06-10 16:11:27,204][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_150_state +[2026-06-10 16:11:27,215][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 16:11:36,244][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_150_state/model.safetensors +[2026-06-10 16:11:52,172][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_150_state/optimizer.bin +[2026-06-10 16:11:52,174][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_150_state/sampler.bin +[2026-06-10 16:11:52,175][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_150_state/sampler_1.bin +[2026-06-10 16:11:52,177][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_150_state/random_states_0.pkl +[2026-06-10 16:11:52,181][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 150 +[2026-06-10 16:11:52,181][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 16:11:52,194][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 16:12:10,902][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.0030908079352229834, 'train/grad_norm_pre_clip': 0.06130475550889969, 'timing/data': 0.017354965209960938, 'timing/model': 1.2303703352808952, 'learning_rate/qwen_vl_interface': 9.624016030669876e-06, 'learning_rate/action_model': 9.624016030669876e-05, 'epoch': 1.26}) +[2026-06-10 16:12:21,200][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.15534113347530365, 'train/grad_norm_pre_clip': 0.4277935028076172, 'timing/data': 0.09130331501364708, 'timing/model': 1.0574800074100494, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 1.27}) +[2026-06-10 16:12:43,492][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.1070181280374527, 'train/grad_norm_pre_clip': 0.31728461384773254, 'timing/data': 0.017826050519943237, 'timing/model': 1.2357329912483692, 'learning_rate/qwen_vl_interface': 9.594384009043415e-06, 'learning_rate/action_model': 9.594384009043416e-05, 'epoch': 1.27}) +[2026-06-10 16:13:05,413][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.03290205076336861, 'train/grad_norm_pre_clip': 0.4182063937187195, 'timing/data': 0.02426975965499878, 'timing/model': 1.16798747330904, 'learning_rate/qwen_vl_interface': 9.579165564018366e-06, 'learning_rate/action_model': 9.579165564018366e-05, 'epoch': 1.28}) +[2026-06-10 16:13:24,384][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.03275858983397484, 'train/grad_norm_pre_clip': 0.3433256447315216, 'timing/data': 0.024036984890699387, 'timing/model': 1.2152610272169113, 'learning_rate/qwen_vl_interface': 9.563680075669138e-06, 'learning_rate/action_model': 9.563680075669137e-05, 'epoch': 1.29}) +[2026-06-10 16:13:44,970][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.021583860740065575, 'train/grad_norm_pre_clip': 0.19699466228485107, 'timing/data': 0.01968318223953247, 'timing/model': 1.3155651725828648, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 1.3}) +[2026-06-10 16:14:07,111][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.008381407707929611, 'train/grad_norm_pre_clip': 0.10134818404912949, 'timing/data': 0.018574543297290802, 'timing/model': 1.218688365072012, 'learning_rate/qwen_vl_interface': 9.531911806283167e-06, 'learning_rate/action_model': 9.531911806283167e-05, 'epoch': 1.31}) +[2026-06-10 16:14:27,503][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.002202942967414856, 'train/grad_norm_pre_clip': 0.02200481668114662, 'timing/data': 0.019119054079055786, 'timing/model': 1.989392876625061, 'learning_rate/qwen_vl_interface': 9.515630984862924e-06, 'learning_rate/action_model': 9.515630984862924e-05, 'epoch': 1.32}) +[2026-06-10 16:14:46,601][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.0017489182064309716, 'train/grad_norm_pre_clip': 0.01900397054851055, 'timing/data': 0.019854653626680374, 'timing/model': 1.344969641417265, 'learning_rate/qwen_vl_interface': 9.49908703923062e-06, 'learning_rate/action_model': 9.49908703923062e-05, 'epoch': 1.32}) +[2026-06-10 16:14:55,776][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.19062842428684235, 'train/grad_norm_pre_clip': 0.5279355645179749, 'timing/data': 0.08295558393001556, 'timing/model': 1.0604091919958591, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 1.33}) +[2026-06-10 16:15:25,374][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.000380854238756001, 'train/grad_norm_pre_clip': 0.005101404618471861, 'timing/data': 0.037522777915000916, 'timing/model': 1.327224012464285, 'learning_rate/qwen_vl_interface': 9.465213873531596e-06, 'learning_rate/action_model': 9.465213873531596e-05, 'epoch': 1.34}) +[2026-06-10 16:15:49,304][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.0004770967934746295, 'train/grad_norm_pre_clip': 0.00418443325906992, 'timing/data': 0.043550118803977966, 'timing/model': 1.2036207057535648, 'learning_rate/qwen_vl_interface': 9.447886742921295e-06, 'learning_rate/action_model': 9.447886742921294e-05, 'epoch': 1.35}) +[2026-06-10 16:16:07,617][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.000258257903624326, 'train/grad_norm_pre_clip': 0.0032228517811745405, 'timing/data': 0.037269797176122665, 'timing/model': 1.2053629755973816, 'learning_rate/qwen_vl_interface': 9.430300666882878e-06, 'learning_rate/action_model': 9.430300666882877e-05, 'epoch': 1.36}) +[2026-06-10 16:16:27,236][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.0003967753436882049, 'train/grad_norm_pre_clip': 0.00503845838829875, 'timing/data': 0.040237490087747574, 'timing/model': 1.3634407632052898, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 1.37}) +[2026-06-10 16:16:48,009][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.00019465644436422735, 'train/grad_norm_pre_clip': 0.002209339290857315, 'timing/data': 0.0328361839056015, 'timing/model': 1.3624048084020615, 'learning_rate/qwen_vl_interface': 9.394356033595787e-06, 'learning_rate/action_model': 9.394356033595787e-05, 'epoch': 1.38}) +[2026-06-10 16:17:07,802][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.00039715843740850687, 'train/grad_norm_pre_clip': 0.004360675346106291, 'timing/data': 0.031517110764980316, 'timing/model': 1.3400112390518188, 'learning_rate/qwen_vl_interface': 9.37599969358141e-06, 'learning_rate/action_model': 9.375999693581409e-05, 'epoch': 1.38}) +[2026-06-10 16:17:28,212][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.0006378539255820215, 'train/grad_norm_pre_clip': 0.008073676377534866, 'timing/data': 0.038401637226343155, 'timing/model': 1.1451071575284004, 'learning_rate/qwen_vl_interface': 9.357388842470736e-06, 'learning_rate/action_model': 9.357388842470737e-05, 'epoch': 1.39}) +[2026-06-10 16:17:38,462][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.18205131590366364, 'train/grad_norm_pre_clip': 0.6786235570907593, 'timing/data': 0.11875181272625923, 'timing/model': 1.2589181028306484, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 1.4}) +[2026-06-10 16:18:03,217][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.11863206326961517, 'train/grad_norm_pre_clip': 0.28649595379829407, 'timing/data': 0.024443920701742172, 'timing/model': 2.0669001638889313, 'learning_rate/qwen_vl_interface': 9.319408214608987e-06, 'learning_rate/action_model': 9.319408214608987e-05, 'epoch': 1.41}) +[2026-06-10 16:18:23,119][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.0010232551721855998, 'train/grad_norm_pre_clip': 0.013979842886328697, 'timing/data': 0.021684281527996063, 'timing/model': 1.3450957238674164, 'learning_rate/qwen_vl_interface': 9.300040780681938e-06, 'learning_rate/action_model': 9.300040780681938e-05, 'epoch': 1.42}) +[2026-06-10 16:18:49,305][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.0028037498705089092, 'train/grad_norm_pre_clip': 0.05904654040932655, 'timing/data': 0.02968606725335121, 'timing/model': 1.389554925262928, 'learning_rate/qwen_vl_interface': 9.280423521162143e-06, 'learning_rate/action_model': 9.280423521162143e-05, 'epoch': 1.43}) +[2026-06-10 16:19:08,478][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.009073206223547459, 'train/grad_norm_pre_clip': 0.10174631327390671, 'timing/data': 0.018339019268751144, 'timing/model': 1.1775192245841026, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 1.43}) +[2026-06-10 16:19:32,887][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.022110572084784508, 'train/grad_norm_pre_clip': 0.1975705921649933, 'timing/data': 0.017136693000793457, 'timing/model': 1.169515810906887, 'learning_rate/qwen_vl_interface': 9.240444381019974e-06, 'learning_rate/action_model': 9.240444381019974e-05, 'epoch': 1.44}) +[2026-06-10 16:19:52,532][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.0024291754234582186, 'train/grad_norm_pre_clip': 0.06187055632472038, 'timing/data': 0.022051144391298294, 'timing/model': 1.3598162271082401, 'learning_rate/qwen_vl_interface': 9.220084966499285e-06, 'learning_rate/action_model': 9.220084966499286e-05, 'epoch': 1.45}) +[2026-06-10 16:20:12,110][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.0152897909283638, 'train/grad_norm_pre_clip': 0.1990870088338852, 'timing/data': 0.02143215388059616, 'timing/model': 1.1831997781991959, 'learning_rate/qwen_vl_interface': 9.199480658437092e-06, 'learning_rate/action_model': 9.199480658437091e-05, 'epoch': 1.46}) +[2026-06-10 16:20:24,018][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.06571099162101746, 'train/grad_norm_pre_clip': 0.4196527898311615, 'timing/data': 0.0918373242020607, 'timing/model': 1.5675751343369484, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 1.47}) +[2026-06-10 16:20:44,303][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.012651450000703335, 'train/grad_norm_pre_clip': 0.16124048829078674, 'timing/data': 0.010766368359327316, 'timing/model': 1.1855437941849232, 'learning_rate/qwen_vl_interface': 9.157542460599085e-06, 'learning_rate/action_model': 9.157542460599085e-05, 'epoch': 1.48}) +[2026-06-10 16:21:08,228][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.055538300424814224, 'train/grad_norm_pre_clip': 0.1879418045282364, 'timing/data': 0.009976040571928024, 'timing/model': 1.261432457715273, 'learning_rate/qwen_vl_interface': 9.136211157768862e-06, 'learning_rate/action_model': 9.136211157768861e-05, 'epoch': 1.48}) +[2026-06-10 16:21:26,503][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.009389933198690414, 'train/grad_norm_pre_clip': 0.11599662154912949, 'timing/data': 0.010108590126037598, 'timing/model': 1.1528208926320076, 'learning_rate/qwen_vl_interface': 9.11464013512874e-06, 'learning_rate/action_model': 9.11464013512874e-05, 'epoch': 1.49}) +[2026-06-10 16:21:44,897][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.00443580886349082, 'train/grad_norm_pre_clip': 0.09520233422517776, 'timing/data': 0.007437199354171753, 'timing/model': 1.1382386796176434, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 1.5}) +[2026-06-10 16:22:03,285][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.01178294699639082, 'train/grad_norm_pre_clip': 0.23518358170986176, 'timing/data': 0.0129576176404953, 'timing/model': 1.3314520381391048, 'learning_rate/qwen_vl_interface': 9.070784267532901e-06, 'learning_rate/action_model': 9.070784267532901e-05, 'epoch': 1.51}) +[2026-06-10 16:22:22,438][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.008268737234175205, 'train/grad_norm_pre_clip': 0.10898572951555252, 'timing/data': 0.010787487030029297, 'timing/model': 1.2129831984639168, 'learning_rate/qwen_vl_interface': 9.048502127813682e-06, 'learning_rate/action_model': 9.048502127813682e-05, 'epoch': 1.52}) +[2026-06-10 16:22:41,364][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.0001033975204336457, 'train/grad_norm_pre_clip': 0.0018140120664611459, 'timing/data': 0.01216304674744606, 'timing/model': 1.1781105771660805, 'learning_rate/qwen_vl_interface': 9.025985678590673e-06, 'learning_rate/action_model': 9.025985678590673e-05, 'epoch': 1.52}) +[2026-06-10 16:22:51,792][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.00014651623496320099, 'train/grad_norm_pre_clip': 0.002556841354817152, 'timing/data': 0.09283015504479408, 'timing/model': 1.0479734614491463, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 1.53}) +[2026-06-10 16:23:15,137][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.0006445850594900548, 'train/grad_norm_pre_clip': 0.08000929653644562, 'timing/data': 0.037199877202510834, 'timing/model': 2.31795896217227, 'learning_rate/qwen_vl_interface': 8.98025542168354e-06, 'learning_rate/action_model': 8.98025542168354e-05, 'epoch': 1.54}) +[2026-06-10 16:23:36,552][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.21009911596775055, 'train/grad_norm_pre_clip': 0.4189019203186035, 'timing/data': 0.029549144208431244, 'timing/model': 1.403298456221819, 'learning_rate/qwen_vl_interface': 8.957044434857067e-06, 'learning_rate/action_model': 8.957044434857067e-05, 'epoch': 1.55}) +[2026-06-10 16:23:56,606][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.058192744851112366, 'train/grad_norm_pre_clip': 0.3624425232410431, 'timing/data': 0.029227618128061295, 'timing/model': 1.1541561596095562, 'learning_rate/qwen_vl_interface': 8.933604780068114e-06, 'learning_rate/action_model': 8.933604780068114e-05, 'epoch': 1.56}) +[2026-06-10 16:24:17,127][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 6.406474858522415e-05, 'train/grad_norm_pre_clip': 0.0017190631479024887, 'timing/data': 0.02371189370751381, 'timing/model': 1.2786272913217545, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 1.57}) +[2026-06-10 16:24:35,960][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.0040756105445325375, 'train/grad_norm_pre_clip': 0.13346673548221588, 'timing/data': 0.026577044278383255, 'timing/model': 1.1453756056725979, 'learning_rate/qwen_vl_interface': 8.88604526409217e-06, 'learning_rate/action_model': 8.886045264092169e-05, 'epoch': 1.57}) +[2026-06-10 16:24:54,902][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.018373597413301468, 'train/grad_norm_pre_clip': 0.24312779307365417, 'timing/data': 0.026440192013978958, 'timing/model': 1.2207128182053566, 'learning_rate/qwen_vl_interface': 8.861928336600148e-06, 'learning_rate/action_model': 8.861928336600148e-05, 'epoch': 1.58}) +[2026-06-10 16:25:20,196][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.00014835908950772136, 'train/grad_norm_pre_clip': 0.0035475571639835835, 'timing/data': 0.025845061987638474, 'timing/model': 1.4083250164985657, 'learning_rate/qwen_vl_interface': 8.837588608354623e-06, 'learning_rate/action_model': 8.837588608354624e-05, 'epoch': 1.59}) +[2026-06-10 16:25:30,555][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.0004899110062979162, 'train/grad_norm_pre_clip': 0.017351005226373672, 'timing/data': 0.1268429011106491, 'timing/model': 1.09882290661335, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 1.6}) +[2026-06-10 16:25:53,819][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.0373213030397892, 'train/grad_norm_pre_clip': 0.7362799644470215, 'timing/data': 0.03424849361181259, 'timing/model': 1.2339113131165504, 'learning_rate/qwen_vl_interface': 8.788246768810206e-06, 'learning_rate/action_model': 8.788246768810206e-05, 'epoch': 1.61}) +[2026-06-10 16:26:16,228][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.0009963170159608126, 'train/grad_norm_pre_clip': 0.05566626414656639, 'timing/data': 0.044996824115514755, 'timing/model': 1.1625907197594643, 'learning_rate/qwen_vl_interface': 8.763247701148398e-06, 'learning_rate/action_model': 8.763247701148397e-05, 'epoch': 1.62}) +[2026-06-10 16:26:34,064][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.00011493429337861016, 'train/grad_norm_pre_clip': 0.0012180213816463947, 'timing/data': 0.042475808411836624, 'timing/model': 1.1432077214121819, 'learning_rate/qwen_vl_interface': 8.738031919819508e-06, 'learning_rate/action_model': 8.738031919819508e-05, 'epoch': 1.62}) +[2026-06-10 16:26:52,382][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.0002150047803297639, 'train/grad_norm_pre_clip': 0.0034401759039610624, 'timing/data': 0.03446375951170921, 'timing/model': 1.4200835712254047, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 1.63}) +[2026-06-10 16:27:11,307][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 9.38329249038361e-05, 'train/grad_norm_pre_clip': 0.0012034703977406025, 'timing/data': 0.02722717449069023, 'timing/model': 1.1538263410329819, 'learning_rate/qwen_vl_interface': 8.686956451145135e-06, 'learning_rate/action_model': 8.686956451145134e-05, 'epoch': 1.64}) +[2026-06-10 16:27:30,115][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.0010123203974217176, 'train/grad_norm_pre_clip': 0.05935390666127205, 'timing/data': 0.024281688034534454, 'timing/model': 1.1420732587575912, 'learning_rate/qwen_vl_interface': 8.661099914375148e-06, 'learning_rate/action_model': 8.661099914375147e-05, 'epoch': 1.65}) +[2026-06-10 16:27:48,705][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.00035610920167528093, 'train/grad_norm_pre_clip': 0.009814923629164696, 'timing/data': 0.030495215207338333, 'timing/model': 1.3187910057604313, 'learning_rate/qwen_vl_interface': 8.635032964894726e-06, 'learning_rate/action_model': 8.635032964894726e-05, 'epoch': 1.66}) +[2026-06-10 16:28:16,775][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 200 +[2026-06-10 16:28:16,777][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.00012679658539127558, 'train/grad_norm_pre_clip': 0.0024930324871093035, 'eval/loss': 0.004338150843977928, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 253.8692138671875, 'rl_games_eval/mid_train/mean_length': 19.8, 'rl_games_eval/mid_train/std_reward': 246.1225268303482, 'rl_games_eval/mid_train/std_length': 9.410632284814874, 'rl_games_eval/mid_train/macro_mean_reward': 253.8692138671875, 'rl_games_eval/mid_train/macro_mean_length': 19.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 253.8692138671875, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 19.8, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 246.1225268303482, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 9.410632284814874, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 253.8692138671875, 'checkpoint/best_score': 315.4344055175781, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.09060314297676086, 'timing/model': 1.0360027775168419, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 1.67}) +[2026-06-10 16:28:16,778][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_200_state +[2026-06-10 16:28:16,787][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 16:28:23,145][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_200_state/model.safetensors +[2026-06-10 16:28:38,870][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_200_state/optimizer.bin +[2026-06-10 16:28:38,871][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_200_state/sampler.bin +[2026-06-10 16:28:38,871][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_200_state/sampler_1.bin +[2026-06-10 16:28:38,875][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-10 16:28:38,879][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-10 16:28:38,879][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 16:28:38,892][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 16:29:00,629][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.017244014889001846, 'train/grad_norm_pre_clip': 0.7046919465065002, 'timing/data': 0.04853539913892746, 'timing/model': 1.905566517263651, 'learning_rate/qwen_vl_interface': 8.582274272411566e-06, 'learning_rate/action_model': 8.582274272411566e-05, 'epoch': 1.68}) +[2026-06-10 16:29:22,704][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.0005871389294043183, 'train/grad_norm_pre_clip': 0.03586330637335777, 'timing/data': 0.027115214616060257, 'timing/model': 3.5024793595075607, 'learning_rate/qwen_vl_interface': 8.555585783813493e-06, 'learning_rate/action_model': 8.555585783813493e-05, 'epoch': 1.68}) +[2026-06-10 16:29:42,033][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.0001060531212715432, 'train/grad_norm_pre_clip': 0.0036755746696144342, 'timing/data': 0.038053248077631, 'timing/model': 1.3578642830252647, 'learning_rate/qwen_vl_interface': 8.528693391113577e-06, 'learning_rate/action_model': 8.528693391113577e-05, 'epoch': 1.69}) +[2026-06-10 16:30:10,314][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.08385974168777466, 'train/grad_norm_pre_clip': 1.658990740776062, 'timing/data': 0.04192483797669411, 'timing/model': 3.428498435765505, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 1.7}) +[2026-06-10 16:30:29,500][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 2.5104145606746897e-05, 'train/grad_norm_pre_clip': 0.0011905088322237134, 'timing/data': 0.03367774933576584, 'timing/model': 1.1972796767950058, 'learning_rate/qwen_vl_interface': 8.474303541281474e-06, 'learning_rate/action_model': 8.474303541281474e-05, 'epoch': 1.71}) +[2026-06-10 16:30:49,416][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 3.107866723439656e-05, 'train/grad_norm_pre_clip': 0.0015344163402915, 'timing/data': 0.04235690459609032, 'timing/model': 1.1553241983056068, 'learning_rate/qwen_vl_interface': 8.446809439171424e-06, 'learning_rate/action_model': 8.446809439171423e-05, 'epoch': 1.72}) +[2026-06-10 16:31:10,169][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.00014987082977313548, 'train/grad_norm_pre_clip': 0.007883787155151367, 'timing/data': 0.037596650421619415, 'timing/model': 1.0941762253642082, 'learning_rate/qwen_vl_interface': 8.41911814279685e-06, 'learning_rate/action_model': 8.41911814279685e-05, 'epoch': 1.73}) +[2026-06-10 16:31:19,119][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.26188018918037415, 'train/grad_norm_pre_clip': 2.251983880996704, 'timing/data': 0.10708893090486526, 'timing/model': 1.1167616546154022, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 1.73}) +[2026-06-10 16:31:42,415][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.11621904373168945, 'train/grad_norm_pre_clip': 1.3732023239135742, 'timing/data': 0.04930780827999115, 'timing/model': 1.1320427879691124, 'learning_rate/qwen_vl_interface': 8.363150811831022e-06, 'learning_rate/action_model': 8.36315081183102e-05, 'epoch': 1.74}) +[2026-06-10 16:32:02,002][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.03164229914546013, 'train/grad_norm_pre_clip': 1.188966989517212, 'timing/data': 0.036119118332862854, 'timing/model': 1.7829630076885223, 'learning_rate/qwen_vl_interface': 8.334878229568372e-06, 'learning_rate/action_model': 8.334878229568372e-05, 'epoch': 1.75}) +[2026-06-10 16:32:25,600][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.053427036851644516, 'train/grad_norm_pre_clip': 0.518425464630127, 'timing/data': 0.04752779006958008, 'timing/model': 1.1672649271786213, 'learning_rate/qwen_vl_interface': 8.306415357485451e-06, 'learning_rate/action_model': 8.306415357485451e-05, 'epoch': 1.76}) +[2026-06-10 16:32:45,297][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.0038654645904898643, 'train/grad_norm_pre_clip': 0.13653041422367096, 'timing/data': 0.03725409880280495, 'timing/model': 1.1435571163892746, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 1.77}) +[2026-06-10 16:33:03,762][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.029819926247000694, 'train/grad_norm_pre_clip': 0.6263588666915894, 'timing/data': 0.03729890286922455, 'timing/model': 1.2431703694164753, 'learning_rate/qwen_vl_interface': 8.248925778384545e-06, 'learning_rate/action_model': 8.248925778384545e-05, 'epoch': 1.77}) +[2026-06-10 16:33:22,014][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.049139585345983505, 'train/grad_norm_pre_clip': 0.9366937279701233, 'timing/data': 0.03815523907542229, 'timing/model': 1.135385800153017, 'learning_rate/qwen_vl_interface': 8.219902617594601e-06, 'learning_rate/action_model': 8.219902617594602e-05, 'epoch': 1.78}) +[2026-06-10 16:33:45,759][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.01055135391652584, 'train/grad_norm_pre_clip': 0.24976417422294617, 'timing/data': 0.03353941813111305, 'timing/model': 1.1274602599442005, 'learning_rate/qwen_vl_interface': 8.190696259221713e-06, 'learning_rate/action_model': 8.190696259221713e-05, 'epoch': 1.79}) +[2026-06-10 16:33:55,144][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.01237515639513731, 'train/grad_norm_pre_clip': 0.5759640336036682, 'timing/data': 0.1084541343152523, 'timing/model': 1.019966296851635, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 1.8}) +[2026-06-10 16:34:16,761][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.0011766288662329316, 'train/grad_norm_pre_clip': 0.025140518322587013, 'timing/data': 0.0325583852827549, 'timing/model': 1.1250376664102077, 'learning_rate/qwen_vl_interface': 8.131741167259499e-06, 'learning_rate/action_model': 8.131741167259499e-05, 'epoch': 1.81}) +[2026-06-10 16:34:35,494][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.280548095703125, 'train/grad_norm_pre_clip': 2.0626819133758545, 'timing/data': 0.030189428478479385, 'timing/model': 1.1471355520188808, 'learning_rate/qwen_vl_interface': 8.10199607029795e-06, 'learning_rate/action_model': 8.101996070297951e-05, 'epoch': 1.82}) +[2026-06-10 16:34:53,599][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.0012754879426211119, 'train/grad_norm_pre_clip': 0.16181887686252594, 'timing/data': 0.04111344367265701, 'timing/model': 1.160224985331297, 'learning_rate/qwen_vl_interface': 8.072075048784685e-06, 'learning_rate/action_model': 8.072075048784683e-05, 'epoch': 1.82}) +[2026-06-10 16:35:12,106][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.00010754148388514295, 'train/grad_norm_pre_clip': 0.0013679449912160635, 'timing/data': 0.040930017828941345, 'timing/model': 1.258700929582119, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 1.83}) +[2026-06-10 16:35:31,589][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.00014207522326614708, 'train/grad_norm_pre_clip': 0.0023107470478862524, 'timing/data': 0.03818003088235855, 'timing/model': 1.9605796858668327, 'learning_rate/qwen_vl_interface': 8.011712625519198e-06, 'learning_rate/action_model': 8.011712625519196e-05, 'epoch': 1.84}) +[2026-06-10 16:35:51,600][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.16706956923007965, 'train/grad_norm_pre_clip': 1.324004888534546, 'timing/data': 0.03585599735379219, 'timing/model': 1.1810945831239223, 'learning_rate/qwen_vl_interface': 7.981274947205574e-06, 'learning_rate/action_model': 7.981274947205573e-05, 'epoch': 1.85}) +[2026-06-10 16:36:10,897][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.0012959460727870464, 'train/grad_norm_pre_clip': 0.020990369841456413, 'timing/data': 0.04201553761959076, 'timing/model': 1.3203845918178558, 'learning_rate/qwen_vl_interface': 7.950668790987754e-06, 'learning_rate/action_model': 7.950668790987755e-05, 'epoch': 1.86}) +[2026-06-10 16:36:21,804][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.001607766724191606, 'train/grad_norm_pre_clip': 0.0342230461537838, 'timing/data': 0.10477671399712563, 'timing/model': 1.330065045505762, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 1.87}) +[2026-06-10 16:36:43,427][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.10781259089708328, 'train/grad_norm_pre_clip': 0.34222155809402466, 'timing/data': 0.008206132799386978, 'timing/model': 1.281951729208231, 'learning_rate/qwen_vl_interface': 7.888958606843111e-06, 'learning_rate/action_model': 7.888958606843111e-05, 'epoch': 1.88}) +[2026-06-10 16:37:02,503][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.0006329138996079564, 'train/grad_norm_pre_clip': 0.01832839846611023, 'timing/data': 0.008376222103834152, 'timing/model': 1.6624900326132774, 'learning_rate/qwen_vl_interface': 7.857858385491127e-06, 'learning_rate/action_model': 7.857858385491127e-05, 'epoch': 1.88}) +[2026-06-10 16:37:22,239][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.007019557990133762, 'train/grad_norm_pre_clip': 0.1816430240869522, 'timing/data': 0.008340705186128616, 'timing/model': 1.252931673079729, 'learning_rate/qwen_vl_interface': 7.826597299149815e-06, 'learning_rate/action_model': 7.826597299149815e-05, 'epoch': 1.89}) +[2026-06-10 16:37:41,276][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.0025576604530215263, 'train/grad_norm_pre_clip': 0.04717765003442764, 'timing/data': 0.009885754436254501, 'timing/model': 1.1816115640103817, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 1.9}) +[2026-06-10 16:38:00,140][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.0004960409714840353, 'train/grad_norm_pre_clip': 0.00919004064053297, 'timing/data': 0.008849196135997772, 'timing/model': 1.178912803530693, 'learning_rate/qwen_vl_interface': 7.763600254627406e-06, 'learning_rate/action_model': 7.763600254627405e-05, 'epoch': 1.91}) +[2026-06-10 16:38:24,753][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.002263891976326704, 'train/grad_norm_pre_clip': 0.0421002060174942, 'timing/data': 0.010676104575395584, 'timing/model': 1.3570300117135048, 'learning_rate/qwen_vl_interface': 7.731868182400758e-06, 'learning_rate/action_model': 7.731868182400758e-05, 'epoch': 1.92}) +[2026-06-10 16:38:43,333][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.0022522166837006807, 'train/grad_norm_pre_clip': 0.0334487110376358, 'timing/data': 0.00985681265592575, 'timing/model': 1.1943959780037403, 'learning_rate/qwen_vl_interface': 7.699983016853976e-06, 'learning_rate/action_model': 7.699983016853976e-05, 'epoch': 1.93}) +[2026-06-10 16:38:54,138][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.00943666510283947, 'train/grad_norm_pre_clip': 0.13008692860603333, 'timing/data': 0.1005156897008419, 'timing/model': 1.14129838347435, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 1.93}) +[2026-06-10 16:39:14,692][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.09940013289451599, 'train/grad_norm_pre_clip': 0.40704217553138733, 'timing/data': 0.017844054847955704, 'timing/model': 1.299848411232233, 'learning_rate/qwen_vl_interface': 7.635761282431021e-06, 'learning_rate/action_model': 7.635761282431019e-05, 'epoch': 1.94}) +[2026-06-10 16:39:35,075][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.0031246449798345566, 'train/grad_norm_pre_clip': 0.07957333326339722, 'timing/data': 0.01675732061266899, 'timing/model': 1.1453610993921757, 'learning_rate/qwen_vl_interface': 7.603428675053936e-06, 'learning_rate/action_model': 7.603428675053937e-05, 'epoch': 1.95}) +[2026-06-10 16:39:55,739][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.0008098192629404366, 'train/grad_norm_pre_clip': 0.04354998841881752, 'timing/data': 0.020657766610383987, 'timing/model': 1.366925373673439, 'learning_rate/qwen_vl_interface': 7.570950897110537e-06, 'learning_rate/action_model': 7.570950897110536e-05, 'epoch': 1.96}) +[2026-06-10 16:40:15,428][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.013736896216869354, 'train/grad_norm_pre_clip': 0.21167375147342682, 'timing/data': 0.017168782651424408, 'timing/model': 1.161865010857582, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 1.97}) +[2026-06-10 16:40:33,828][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.00013617187505587935, 'train/grad_norm_pre_clip': 0.0035310406237840652, 'timing/data': 0.022278517484664917, 'timing/model': 1.1282400079071522, 'learning_rate/qwen_vl_interface': 7.505567851885325e-06, 'learning_rate/action_model': 7.505567851885325e-05, 'epoch': 1.98}) +[2026-06-10 16:40:53,338][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.018349941819906235, 'train/grad_norm_pre_clip': 0.32148078083992004, 'timing/data': 0.014733880758285522, 'timing/model': 1.4844577200710773, 'learning_rate/qwen_vl_interface': 7.472666617737725e-06, 'learning_rate/action_model': 7.472666617737725e-05, 'epoch': 1.98}) +[2026-06-10 16:41:12,710][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.007196136750280857, 'train/grad_norm_pre_clip': 0.15170994400978088, 'timing/data': 0.01763637363910675, 'timing/model': 1.3198453560471535, 'learning_rate/qwen_vl_interface': 7.439628279043447e-06, 'learning_rate/action_model': 7.439628279043447e-05, 'epoch': 1.99}) +[2026-06-10 16:41:23,931][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.00039955106331035495, 'train/grad_norm_pre_clip': 0.012888831086456776, 'timing/data': 0.11692072078585625, 'timing/model': 1.2018361911177635, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 2.0}) +[2026-06-10 16:41:48,708][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 3.695338455145247e-05, 'train/grad_norm_pre_clip': 0.0009481604211032391, 'timing/data': 0.007606353610754013, 'timing/model': 1.188782311975956, 'learning_rate/qwen_vl_interface': 7.373148448187827e-06, 'learning_rate/action_model': 7.373148448187827e-05, 'epoch': 2.01}) +[2026-06-10 16:42:07,240][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.0021899687126278877, 'train/grad_norm_pre_clip': 0.0792728066444397, 'timing/data': 0.009260531514883041, 'timing/model': 1.2917338348925114, 'learning_rate/qwen_vl_interface': 7.339711056815598e-06, 'learning_rate/action_model': 7.339711056815598e-05, 'epoch': 2.02}) +[2026-06-10 16:42:25,911][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.0008761256467550993, 'train/grad_norm_pre_clip': 0.022608811035752296, 'timing/data': 0.012168310582637787, 'timing/model': 1.1465293429791927, 'learning_rate/qwen_vl_interface': 7.306144762221968e-06, 'learning_rate/action_model': 7.306144762221968e-05, 'epoch': 2.02}) +[2026-06-10 16:42:44,730][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.0027178055606782436, 'train/grad_norm_pre_clip': 0.05827195942401886, 'timing/data': 0.016082197427749634, 'timing/model': 1.0039110146462917, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 2.03}) +[2026-06-10 16:43:03,299][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.00019065267406404018, 'train/grad_norm_pre_clip': 0.01183044072240591, 'timing/data': 0.008842062205076218, 'timing/model': 1.3110871352255344, 'learning_rate/qwen_vl_interface': 7.238633753302784e-06, 'learning_rate/action_model': 7.238633753302784e-05, 'epoch': 2.04}) +[2026-06-10 16:43:21,589][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.000637926859781146, 'train/grad_norm_pre_clip': 0.0224605742841959, 'timing/data': 0.012565914541482925, 'timing/model': 1.1549273878335953, 'learning_rate/qwen_vl_interface': 7.204693203374268e-06, 'learning_rate/action_model': 7.204693203374267e-05, 'epoch': 2.05}) +[2026-06-10 16:43:40,010][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.11618975549936295, 'train/grad_norm_pre_clip': 1.4741551876068115, 'timing/data': 0.008392926305532455, 'timing/model': 1.178431548178196, 'learning_rate/qwen_vl_interface': 7.17063207876154e-06, 'learning_rate/action_model': 7.17063207876154e-05, 'epoch': 2.06}) +[2026-06-10 16:43:50,566][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.03263787180185318, 'train/grad_norm_pre_clip': 0.49415645003318787, 'timing/data': 0.13715306296944618, 'timing/model': 1.1431485377252102, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 2.07}) +[2026-06-10 16:44:12,326][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 2.5480865133431507e-06, 'train/grad_norm_pre_clip': 5.310256528900936e-05, 'timing/data': 0.022228948771953583, 'timing/model': 1.1769698522984982, 'learning_rate/qwen_vl_interface': 7.1021565169938955e-06, 'learning_rate/action_model': 7.102156516993895e-05, 'epoch': 2.08}) +[2026-06-10 16:44:42,113][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 250 +[2026-06-10 16:44:42,114][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.012566173449158669, 'train/grad_norm_pre_clip': 0.2768998444080353, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 35.550009155273436, 'rl_games_eval/mid_train/mean_length': 15.2, 'rl_games_eval/mid_train/std_reward': 45.95292670940836, 'rl_games_eval/mid_train/std_length': 3.1874754901018454, 'rl_games_eval/mid_train/macro_mean_reward': 35.550009155273436, 'rl_games_eval/mid_train/macro_mean_length': 15.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 35.550009155273436, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 15.2, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 45.95292670940836, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 3.1874754901018454, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 35.550009155273436, 'checkpoint/best_score': 315.4344055175781, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.01327020674943924, 'timing/model': 1.8972660452127457, 'learning_rate/qwen_vl_interface': 7.067746303734178e-06, 'learning_rate/action_model': 7.067746303734177e-05, 'epoch': 2.08}) +[2026-06-10 16:44:42,115][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_250_state +[2026-06-10 16:44:42,120][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 16:44:48,937][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_250_state/model.safetensors +[2026-06-10 16:45:05,258][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_250_state/optimizer.bin +[2026-06-10 16:45:05,259][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_250_state/sampler.bin +[2026-06-10 16:45:05,259][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_250_state/sampler_1.bin +[2026-06-10 16:45:05,263][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_250_state/random_states_0.pkl +[2026-06-10 16:45:05,268][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 250 +[2026-06-10 16:45:05,269][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 16:45:05,282][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 16:45:23,988][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.0071380771696567535, 'train/grad_norm_pre_clip': 0.22130483388900757, 'timing/data': 0.024833928793668747, 'timing/model': 1.2449254989624023, 'learning_rate/qwen_vl_interface': 7.033223963320096e-06, 'learning_rate/action_model': 7.033223963320095e-05, 'epoch': 2.09}) +[2026-06-10 16:45:42,757][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.0010586759308353066, 'train/grad_norm_pre_clip': 0.03838683292269707, 'timing/data': 0.013754487037658691, 'timing/model': 1.3336048871278763, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 2.1}) +[2026-06-10 16:46:04,865][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.004113468807190657, 'train/grad_norm_pre_clip': 0.13214947283267975, 'timing/data': 0.016079209744930267, 'timing/model': 3.190352711826563, 'learning_rate/qwen_vl_interface': 6.963851425816299e-06, 'learning_rate/action_model': 6.963851425816299e-05, 'epoch': 2.11}) +[2026-06-10 16:46:25,212][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.00026641099248081446, 'train/grad_norm_pre_clip': 0.013007786124944687, 'timing/data': 0.016779348254203796, 'timing/model': 1.2633613795042038, 'learning_rate/qwen_vl_interface': 6.929005507951473e-06, 'learning_rate/action_model': 6.929005507951473e-05, 'epoch': 2.12}) +[2026-06-10 16:46:52,522][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.04197518154978752, 'train/grad_norm_pre_clip': 1.3890308141708374, 'timing/data': 0.024043899029493332, 'timing/model': 1.4115270599722862, 'learning_rate/qwen_vl_interface': 6.894056021118092e-06, 'learning_rate/action_model': 6.894056021118091e-05, 'epoch': 2.12}) +[2026-06-10 16:47:02,417][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.0017285965150222182, 'train/grad_norm_pre_clip': 0.08731673657894135, 'timing/data': 0.12375423312187195, 'timing/model': 1.2307170927524567, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 2.13}) +[2026-06-10 16:47:22,626][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 4.749659183289623e-06, 'train/grad_norm_pre_clip': 0.00016339929425157607, 'timing/data': 0.011709421873092651, 'timing/model': 1.470378890633583, 'learning_rate/qwen_vl_interface': 6.823854970197212e-06, 'learning_rate/action_model': 6.823854970197213e-05, 'epoch': 2.14}) +[2026-06-10 16:47:41,239][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.009967359714210033, 'train/grad_norm_pre_clip': 2.4637792110443115, 'timing/data': 0.006836261600255966, 'timing/model': 1.2390422523021698, 'learning_rate/qwen_vl_interface': 6.788607736441211e-06, 'learning_rate/action_model': 6.78860773644121e-05, 'epoch': 2.15}) +[2026-06-10 16:48:01,606][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.008710294030606747, 'train/grad_norm_pre_clip': 0.34949734807014465, 'timing/data': 0.010461360216140747, 'timing/model': 1.1338094174861908, 'learning_rate/qwen_vl_interface': 6.753265594112535e-06, 'learning_rate/action_model': 6.753265594112534e-05, 'epoch': 2.16}) +[2026-06-10 16:48:20,416][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.003908186685293913, 'train/grad_norm_pre_clip': 0.20225994288921356, 'timing/data': 0.011548981070518494, 'timing/model': 1.1793871372938156, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 2.17}) +[2026-06-10 16:48:40,778][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.030633002519607544, 'train/grad_norm_pre_clip': 0.9396381974220276, 'timing/data': 0.00998280942440033, 'timing/model': 1.2509521543979645, 'learning_rate/qwen_vl_interface': 6.682305309736355e-06, 'learning_rate/action_model': 6.682305309736354e-05, 'epoch': 2.17}) +[2026-06-10 16:49:00,338][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 3.3990647352766246e-05, 'train/grad_norm_pre_clip': 0.0032396770548075438, 'timing/data': 0.009570624679327011, 'timing/model': 1.5632620193064213, 'learning_rate/qwen_vl_interface': 6.646691544853444e-06, 'learning_rate/action_model': 6.646691544853444e-05, 'epoch': 2.18}) +[2026-06-10 16:49:26,009][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 1.6318968846462667e-05, 'train/grad_norm_pre_clip': 0.0007676639361307025, 'timing/data': 0.011654265224933624, 'timing/model': 1.3261401392519474, 'learning_rate/qwen_vl_interface': 6.610991625457042e-06, 'learning_rate/action_model': 6.610991625457043e-05, 'epoch': 2.19}) +[2026-06-10 16:49:40,440][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.00022614881163462996, 'train/grad_norm_pre_clip': 0.009281475096940994, 'timing/data': 0.0929611586034298, 'timing/model': 1.2562593594193459, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 2.2}) +[2026-06-10 16:50:01,774][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.020918577909469604, 'train/grad_norm_pre_clip': 0.6083134412765503, 'timing/data': 0.029269307851791382, 'timing/model': 1.1699779443442822, 'learning_rate/qwen_vl_interface': 6.539342136859104e-06, 'learning_rate/action_model': 6.539342136859104e-05, 'epoch': 2.21}) +[2026-06-10 16:50:20,504][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 6.407465775737364e-07, 'train/grad_norm_pre_clip': 2.642878462211229e-05, 'timing/data': 0.030746523290872574, 'timing/model': 1.1492827907204628, 'learning_rate/qwen_vl_interface': 6.503396987335522e-06, 'learning_rate/action_model': 6.503396987335521e-05, 'epoch': 2.22}) +[2026-06-10 16:50:39,902][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.017827728763222694, 'train/grad_norm_pre_clip': 0.7631959319114685, 'timing/data': 0.02906842902302742, 'timing/model': 1.3626128733158112, 'learning_rate/qwen_vl_interface': 6.4673745223817355e-06, 'learning_rate/action_model': 6.467374522381734e-05, 'epoch': 2.23}) +[2026-06-10 16:51:01,781][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 2.7563321054913104e-05, 'train/grad_norm_pre_clip': 0.0015514284605160356, 'timing/data': 0.031061965972185135, 'timing/model': 1.745203211903572, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 2.23}) +[2026-06-10 16:51:26,303][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.03517034649848938, 'train/grad_norm_pre_clip': 1.2112582921981812, 'timing/data': 0.028218548744916916, 'timing/model': 1.1886763125658035, 'learning_rate/qwen_vl_interface': 6.395106538956934e-06, 'learning_rate/action_model': 6.395106538956934e-05, 'epoch': 2.24}) +[2026-06-10 16:51:46,295][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.005533413961529732, 'train/grad_norm_pre_clip': 0.2547881603240967, 'timing/data': 0.02869643270969391, 'timing/model': 1.1277242340147495, 'learning_rate/qwen_vl_interface': 6.3588654783155525e-06, 'learning_rate/action_model': 6.358865478315553e-05, 'epoch': 2.25}) +[2026-06-10 16:52:04,875][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 1.7881382063933415e-07, 'train/grad_norm_pre_clip': 0.00017623718304093927, 'timing/data': 0.02804408222436905, 'timing/model': 1.0750809200108051, 'learning_rate/qwen_vl_interface': 6.322556017628244e-06, 'learning_rate/action_model': 6.322556017628244e-05, 'epoch': 2.26}) +[2026-06-10 16:52:14,651][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 8.514948746096707e-08, 'train/grad_norm_pre_clip': 2.9980058116052533e-06, 'timing/data': 0.09032182022929192, 'timing/model': 1.6377883478999138, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 2.27}) +[2026-06-10 16:52:34,312][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.09287863969802856, 'train/grad_norm_pre_clip': 1.2943377494812012, 'timing/data': 0.012563183903694153, 'timing/model': 1.1697961427271366, 'learning_rate/qwen_vl_interface': 6.249740859151202e-06, 'learning_rate/action_model': 6.249740859151202e-05, 'epoch': 2.27}) +[2026-06-10 16:52:55,248][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 1.1548399214689198e-07, 'train/grad_norm_pre_clip': 9.9273847808945e-07, 'timing/data': 0.007092129439115524, 'timing/model': 1.6698537729680538, 'learning_rate/qwen_vl_interface': 6.213239652943435e-06, 'learning_rate/action_model': 6.213239652943435e-05, 'epoch': 2.28}) +[2026-06-10 16:53:14,019][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 2.5136771000688896e-05, 'train/grad_norm_pre_clip': 0.0031299542170017958, 'timing/data': 0.013160418719053268, 'timing/model': 1.1826870664954185, 'learning_rate/qwen_vl_interface': 6.176679029576611e-06, 'learning_rate/action_model': 6.17667902957661e-05, 'epoch': 2.29}) +[2026-06-10 16:53:35,668][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.004161075223237276, 'train/grad_norm_pre_clip': 0.10293936729431152, 'timing/data': 0.015723485499620438, 'timing/model': 1.1057446673512459, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 2.3}) +[2026-06-10 16:53:54,619][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.17942358553409576, 'train/grad_norm_pre_clip': 3.941448211669922, 'timing/data': 0.006533581763505936, 'timing/model': 1.1784273087978363, 'learning_rate/qwen_vl_interface': 6.103388555817666e-06, 'learning_rate/action_model': 6.1033885558176665e-05, 'epoch': 2.31}) +[2026-06-10 16:54:16,985][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 8.388370588363614e-06, 'train/grad_norm_pre_clip': 0.000552437617443502, 'timing/data': 0.014307186007499695, 'timing/model': 2.2531671077013016, 'learning_rate/qwen_vl_interface': 6.066663226327195e-06, 'learning_rate/action_model': 6.066663226327195e-05, 'epoch': 2.32}) +[2026-06-10 16:54:42,331][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.3774205148220062, 'train/grad_norm_pre_clip': 2.755237102508545, 'timing/data': 0.013100627809762955, 'timing/model': 1.1625719144940376, 'learning_rate/qwen_vl_interface': 6.0298875212021e-06, 'learning_rate/action_model': 6.0298875212021e-05, 'epoch': 2.33}) +[2026-06-10 16:54:51,639][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 1.2175178198958747e-05, 'train/grad_norm_pre_clip': 0.0010663724970072508, 'timing/data': 0.0929056704044342, 'timing/model': 1.0112081170082092, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 2.33}) +[2026-06-10 16:55:19,996][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 7.737232408544514e-06, 'train/grad_norm_pre_clip': 0.00025964240194298327, 'timing/data': 0.032732486724853516, 'timing/model': 1.1880029737949371, 'learning_rate/qwen_vl_interface': 5.9561940610103865e-06, 'learning_rate/action_model': 5.956194061010386e-05, 'epoch': 2.34}) +[2026-06-10 16:55:39,089][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.00432868767529726, 'train/grad_norm_pre_clip': 0.22692131996154785, 'timing/data': 0.04859178513288498, 'timing/model': 1.2423336654901505, 'learning_rate/qwen_vl_interface': 5.919280851703518e-06, 'learning_rate/action_model': 5.919280851703518e-05, 'epoch': 2.35}) +[2026-06-10 16:55:58,646][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.005136705003678799, 'train/grad_norm_pre_clip': 0.14199744164943695, 'timing/data': 0.18961497023701668, 'timing/model': 1.3338972628116608, 'learning_rate/qwen_vl_interface': 5.8823263580011245e-06, 'learning_rate/action_model': 5.882326358001125e-05, 'epoch': 2.36}) +[2026-06-10 16:56:20,290][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.008459707722067833, 'train/grad_norm_pre_clip': 0.523438572883606, 'timing/data': 0.033886827528476715, 'timing/model': 1.374081801623106, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 2.37}) +[2026-06-10 16:56:39,192][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.23596078157424927, 'train/grad_norm_pre_clip': 2.617807626724243, 'timing/data': 0.035978395491838455, 'timing/model': 1.1990536004304886, 'learning_rate/qwen_vl_interface': 5.80830263792473e-06, 'learning_rate/action_model': 5.80830263792473e-05, 'epoch': 2.38}) +[2026-06-10 16:57:01,551][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 9.291718015447259e-05, 'train/grad_norm_pre_clip': 0.0025497747119516134, 'timing/data': 0.04769938066601753, 'timing/model': 1.1390081085264683, 'learning_rate/qwen_vl_interface': 5.771237977682465e-06, 'learning_rate/action_model': 5.771237977682465e-05, 'epoch': 2.38}) +[2026-06-10 16:57:20,641][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.03728918731212616, 'train/grad_norm_pre_clip': 0.5229532718658447, 'timing/data': 0.04308834671974182, 'timing/model': 1.1760883145034313, 'learning_rate/qwen_vl_interface': 5.734141165026489e-06, 'learning_rate/action_model': 5.7341411650264886e-05, 'epoch': 2.39}) +[2026-06-10 16:57:30,680][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.0061825974844396114, 'train/grad_norm_pre_clip': 0.18458977341651917, 'timing/data': 0.10535729303956032, 'timing/model': 1.26395483314991, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 2.4}) +[2026-06-10 16:57:56,658][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.010395816527307034, 'train/grad_norm_pre_clip': 0.38803568482398987, 'timing/data': 0.015922699123620987, 'timing/model': 1.2707813642919064, 'learning_rate/qwen_vl_interface': 5.659860237540138e-06, 'learning_rate/action_model': 5.659860237540138e-05, 'epoch': 2.41}) +[2026-06-10 16:58:22,569][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.0027360767126083374, 'train/grad_norm_pre_clip': 0.14666172862052917, 'timing/data': 0.0004925765097141266, 'timing/model': 1.146962746977806, 'learning_rate/qwen_vl_interface': 5.622680704707265e-06, 'learning_rate/action_model': 5.622680704707265e-05, 'epoch': 2.42}) +[2026-06-10 16:58:41,295][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.00011435686610639095, 'train/grad_norm_pre_clip': 0.00532956188544631, 'timing/data': 0.0005454272031784058, 'timing/model': 1.1277516223490238, 'learning_rate/qwen_vl_interface': 5.585478183173043e-06, 'learning_rate/action_model': 5.585478183173043e-05, 'epoch': 2.42}) +[2026-06-10 16:59:06,203][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.00019719815463759005, 'train/grad_norm_pre_clip': 0.010792340151965618, 'timing/data': 0.0004845447838306427, 'timing/model': 1.273634746670723, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 2.43}) +[2026-06-10 16:59:24,614][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.0006745695718564093, 'train/grad_norm_pre_clip': 0.0555923767387867, 'timing/data': 0.0007777325809001923, 'timing/model': 1.1352483630180359, 'learning_rate/qwen_vl_interface': 5.511013354584118e-06, 'learning_rate/action_model': 5.511013354584118e-05, 'epoch': 2.44}) +[2026-06-10 16:59:48,231][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.11623390018939972, 'train/grad_norm_pre_clip': 3.1703569889068604, 'timing/data': 0.0006406009197235107, 'timing/model': 1.587643001228571, 'learning_rate/qwen_vl_interface': 5.473755640870802e-06, 'learning_rate/action_model': 5.473755640870802e-05, 'epoch': 2.45}) +[2026-06-10 17:00:05,623][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 2.7724363462766632e-05, 'train/grad_norm_pre_clip': 0.0010210019536316395, 'timing/data': 0.0004935003817081451, 'timing/model': 0.8867983557283878, 'learning_rate/qwen_vl_interface': 5.436484124855577e-06, 'learning_rate/action_model': 5.4364841248555764e-05, 'epoch': 2.46}) +[2026-06-10 17:00:14,806][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.01317604724317789, 'train/grad_norm_pre_clip': 0.27455055713653564, 'timing/data': 0.08682284876704216, 'timing/model': 1.0353017151355743, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 2.47}) +[2026-06-10 17:00:31,887][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.04850947484374046, 'train/grad_norm_pre_clip': 2.5672361850738525, 'timing/data': 0.04696154221892357, 'timing/model': 1.192336868494749, 'learning_rate/qwen_vl_interface': 5.36190888295965e-06, 'learning_rate/action_model': 5.36190888295965e-05, 'epoch': 2.48}) +[2026-06-10 17:00:46,044][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.002916078083217144, 'train/grad_norm_pre_clip': 0.13304509222507477, 'timing/data': 0.03769025206565857, 'timing/model': 0.8957313001155853, 'learning_rate/qwen_vl_interface': 5.324609757231148e-06, 'learning_rate/action_model': 5.324609757231148e-05, 'epoch': 2.48}) +[2026-06-10 17:01:00,876][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.0011496917577460408, 'train/grad_norm_pre_clip': 0.03957361727952957, 'timing/data': 0.03268488869071007, 'timing/model': 1.2297802232205868, 'learning_rate/qwen_vl_interface': 5.287306029221379e-06, 'learning_rate/action_model': 5.287306029221378e-05, 'epoch': 2.49}) +[2026-06-10 17:01:36,604][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 300 +[2026-06-10 17:01:36,605][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.01159694604575634, 'train/grad_norm_pre_clip': 0.6817001700401306, 'eval/loss': 0.0050911447033286095, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': -56.194915771484375, 'rl_games_eval/mid_train/mean_length': 20.2, 'rl_games_eval/mid_train/std_reward': 55.0067701324979, 'rl_games_eval/mid_train/std_length': 5.418486873657626, 'rl_games_eval/mid_train/macro_mean_reward': -56.194915771484375, 'rl_games_eval/mid_train/macro_mean_length': 20.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': -56.194915771484375, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 20.2, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 55.0067701324979, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.418486873657626, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': -56.194915771484375, 'checkpoint/best_score': 315.4344055175781, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.03889830783009529, 'timing/model': 1.1965783201158047, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 2.5}) +[2026-06-10 17:01:36,606][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_300_state +[2026-06-10 17:01:36,611][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 17:01:45,469][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_300_state/model.safetensors +[2026-06-10 17:02:02,549][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_300_state/optimizer.bin +[2026-06-10 17:02:02,551][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_300_state/sampler.bin +[2026-06-10 17:02:02,552][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_300_state/sampler_1.bin +[2026-06-10 17:02:02,556][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-10 17:02:02,561][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-10 17:02:02,561][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 17:02:02,574][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 17:02:22,251][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 5.535729997063754e-06, 'train/grad_norm_pre_clip': 0.0001545598788652569, 'timing/data': 0.03537489473819733, 'timing/model': 1.153729997575283, 'learning_rate/qwen_vl_interface': 5.212693970778623e-06, 'learning_rate/action_model': 5.2126939707786224e-05, 'epoch': 2.51}) +[2026-06-10 17:02:40,838][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.0790301114320755, 'train/grad_norm_pre_clip': 0.3538259267807007, 'timing/data': 0.03618437424302101, 'timing/model': 1.126904059201479, 'learning_rate/qwen_vl_interface': 5.175390242768853e-06, 'learning_rate/action_model': 5.175390242768853e-05, 'epoch': 2.52}) +[2026-06-10 17:03:02,505][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.00033588282531127334, 'train/grad_norm_pre_clip': 0.01217969786375761, 'timing/data': 0.040537189692258835, 'timing/model': 1.1583159603178501, 'learning_rate/qwen_vl_interface': 5.1380911170403525e-06, 'learning_rate/action_model': 5.1380911170403525e-05, 'epoch': 2.52}) +[2026-06-10 17:03:11,597][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 6.82423069520155e-06, 'train/grad_norm_pre_clip': 0.00034850870724767447, 'timing/data': 0.0932665765285492, 'timing/model': 1.0011582374572754, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 2.53}) +[2026-06-10 17:03:35,938][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.030375514179468155, 'train/grad_norm_pre_clip': 0.477674663066864, 'timing/data': 0.03795330226421356, 'timing/model': 1.1658855825662613, 'learning_rate/qwen_vl_interface': 5.063515875144426e-06, 'learning_rate/action_model': 5.0635158751444254e-05, 'epoch': 2.54}) +[2026-06-10 17:03:54,311][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.00011655189155135304, 'train/grad_norm_pre_clip': 0.0059646242298185825, 'timing/data': 0.03428933769464493, 'timing/model': 1.327257975935936, 'learning_rate/qwen_vl_interface': 5.026244359129198e-06, 'learning_rate/action_model': 5.0262443591291975e-05, 'epoch': 2.55}) +[2026-06-10 17:04:12,346][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 4.149854703427991e-06, 'train/grad_norm_pre_clip': 0.00011068419553339481, 'timing/data': 0.03847331553697586, 'timing/model': 1.1515809185802937, 'learning_rate/qwen_vl_interface': 4.988986645415884e-06, 'learning_rate/action_model': 4.988986645415884e-05, 'epoch': 2.56}) +[2026-06-10 17:04:32,916][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 7.889354492363054e-06, 'train/grad_norm_pre_clip': 0.00042045925511047244, 'timing/data': 0.0297381728887558, 'timing/model': 1.322317648679018, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 2.57}) +[2026-06-10 17:04:52,784][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 9.460423461860046e-05, 'train/grad_norm_pre_clip': 0.007992398925125599, 'timing/data': 0.029083650559186935, 'timing/model': 1.1486049517989159, 'learning_rate/qwen_vl_interface': 4.91452181682696e-06, 'learning_rate/action_model': 4.9145218168269594e-05, 'epoch': 2.58}) +[2026-06-10 17:05:10,821][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 4.425491169968154e-06, 'train/grad_norm_pre_clip': 0.0002618547878228128, 'timing/data': 0.041447632014751434, 'timing/model': 1.1349957771599293, 'learning_rate/qwen_vl_interface': 4.877319295292737e-06, 'learning_rate/action_model': 4.8773192952927374e-05, 'epoch': 2.58}) +[2026-06-10 17:05:30,567][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 2.2351738948600541e-07, 'train/grad_norm_pre_clip': 4.743022600450786e-06, 'timing/data': 0.02222052961587906, 'timing/model': 1.4061266295611858, 'learning_rate/qwen_vl_interface': 4.840139762459865e-06, 'learning_rate/action_model': 4.8401397624598645e-05, 'epoch': 2.59}) +[2026-06-10 17:05:39,549][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.13509008288383484, 'train/grad_norm_pre_clip': 2.48201060295105, 'timing/data': 0.09651516750454903, 'timing/model': 1.0429532751441002, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 2.6}) +[2026-06-10 17:06:01,898][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 2.5331962660857243e-07, 'train/grad_norm_pre_clip': 6.1816699599148706e-06, 'timing/data': 0.03360180929303169, 'timing/model': 1.3973137140274048, 'learning_rate/qwen_vl_interface': 4.7658588349735135e-06, 'learning_rate/action_model': 4.765858834973513e-05, 'epoch': 2.61}) +[2026-06-10 17:06:20,209][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 4.362047093309229e-06, 'train/grad_norm_pre_clip': 0.00016296684043481946, 'timing/data': 0.04231753572821617, 'timing/model': 1.1520549692213535, 'learning_rate/qwen_vl_interface': 4.728762022317535e-06, 'learning_rate/action_model': 4.728762022317535e-05, 'epoch': 2.62}) +[2026-06-10 17:06:38,399][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.00022974358580540866, 'train/grad_norm_pre_clip': 0.016000067815184593, 'timing/data': 0.035479169338941574, 'timing/model': 1.153972327709198, 'learning_rate/qwen_vl_interface': 4.691697362075272e-06, 'learning_rate/action_model': 4.691697362075272e-05, 'epoch': 2.62}) +[2026-06-10 17:06:56,689][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 1.0616102372296154e-05, 'train/grad_norm_pre_clip': 0.00033446052111685276, 'timing/data': 0.0436803437769413, 'timing/model': 1.181185007095337, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 2.63}) +[2026-06-10 17:07:16,969][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 7.227038736346003e-07, 'train/grad_norm_pre_clip': 1.9924140360672027e-05, 'timing/data': 0.034274082630872726, 'timing/model': 1.2803725190460682, 'learning_rate/qwen_vl_interface': 4.617673641998877e-06, 'learning_rate/action_model': 4.617673641998877e-05, 'epoch': 2.64}) +[2026-06-10 17:07:36,075][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 3.002473931701388e-06, 'train/grad_norm_pre_clip': 0.00010110442235600203, 'timing/data': 0.04456745833158493, 'timing/model': 1.138293992727995, 'learning_rate/qwen_vl_interface': 4.580719148296482e-06, 'learning_rate/action_model': 4.580719148296482e-05, 'epoch': 2.65}) +[2026-06-10 17:07:55,611][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 8.231879291997757e-06, 'train/grad_norm_pre_clip': 0.00040898361476138234, 'timing/data': 0.03214995190501213, 'timing/model': 1.1514601483941078, 'learning_rate/qwen_vl_interface': 4.543805938989616e-06, 'learning_rate/action_model': 4.543805938989616e-05, 'epoch': 2.66}) +[2026-06-10 17:08:04,537][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.23245713114738464, 'train/grad_norm_pre_clip': 4.291235446929932, 'timing/data': 0.07091069221496582, 'timing/model': 1.048451367765665, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 2.67}) +[2026-06-10 17:08:25,887][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 5.215400733504794e-07, 'train/grad_norm_pre_clip': 1.2488219908846077e-05, 'timing/data': 0.03005180135369301, 'timing/model': 1.166774369776249, 'learning_rate/qwen_vl_interface': 4.470112478797902e-06, 'learning_rate/action_model': 4.470112478797901e-05, 'epoch': 2.67}) +[2026-06-10 17:08:56,559][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 3.081777322222479e-05, 'train/grad_norm_pre_clip': 0.0012038778513669968, 'timing/data': 0.042412467300891876, 'timing/model': 1.163431629538536, 'learning_rate/qwen_vl_interface': 4.433336773672804e-06, 'learning_rate/action_model': 4.4333367736728035e-05, 'epoch': 2.68}) +[2026-06-10 17:09:15,067][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 9.588354259904008e-06, 'train/grad_norm_pre_clip': 0.0003298908704891801, 'timing/data': 0.03836870938539505, 'timing/model': 1.2261733151972294, 'learning_rate/qwen_vl_interface': 4.396611444182336e-06, 'learning_rate/action_model': 4.3966114441823353e-05, 'epoch': 2.69}) +[2026-06-10 17:09:33,371][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 7.823101668691379e-07, 'train/grad_norm_pre_clip': 1.9069617337663658e-05, 'timing/data': 0.02938232570886612, 'timing/model': 1.1875846907496452, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 2.7}) +[2026-06-10 17:09:55,883][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.05228444188833237, 'train/grad_norm_pre_clip': 0.46758347749710083, 'timing/data': 0.028394348919391632, 'timing/model': 1.7683387622237206, 'learning_rate/qwen_vl_interface': 4.323320970423391e-06, 'learning_rate/action_model': 4.3233209704233915e-05, 'epoch': 2.71}) +[2026-06-10 17:10:19,098][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.11418280005455017, 'train/grad_norm_pre_clip': 2.9787731170654297, 'timing/data': 0.1080423966050148, 'timing/model': 2.3843675032258034, 'learning_rate/qwen_vl_interface': 4.286760347056567e-06, 'learning_rate/action_model': 4.286760347056567e-05, 'epoch': 2.72}) +[2026-06-10 17:10:38,426][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.00012428253830876201, 'train/grad_norm_pre_clip': 0.007689999882131815, 'timing/data': 0.03259415924549103, 'timing/model': 1.1595774814486504, 'learning_rate/qwen_vl_interface': 4.250259140848798e-06, 'learning_rate/action_model': 4.2502591408487985e-05, 'epoch': 2.73}) +[2026-06-10 17:10:47,380][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.052525125443935394, 'train/grad_norm_pre_clip': 2.191706418991089, 'timing/data': 0.0813937857747078, 'timing/model': 1.0354401804506779, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 2.73}) +[2026-06-10 17:11:08,754][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 3.046651545446366e-05, 'train/grad_norm_pre_clip': 0.0008123524021357298, 'timing/data': 0.03492789342999458, 'timing/model': 1.1438279338181019, 'learning_rate/qwen_vl_interface': 4.177443982371757e-06, 'learning_rate/action_model': 4.177443982371758e-05, 'epoch': 2.74}) +[2026-06-10 17:11:27,655][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.008666431531310081, 'train/grad_norm_pre_clip': 0.4311131238937378, 'timing/data': 0.04293166100978851, 'timing/model': 1.0128226764500141, 'learning_rate/qwen_vl_interface': 4.14113452168445e-06, 'learning_rate/action_model': 4.1411345216844504e-05, 'epoch': 2.75}) +[2026-06-10 17:11:47,395][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 8.728015927772503e-06, 'train/grad_norm_pre_clip': 0.0002953540242742747, 'timing/data': 0.028219319880008698, 'timing/model': 1.14887061342597, 'learning_rate/qwen_vl_interface': 4.104893461043067e-06, 'learning_rate/action_model': 4.1048934610430664e-05, 'epoch': 2.76}) +[2026-06-10 17:12:06,092][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.04156551882624626, 'train/grad_norm_pre_clip': 1.135740876197815, 'timing/data': 0.04815156012773514, 'timing/model': 1.1363923735916615, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 2.77}) +[2026-06-10 17:12:24,367][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 6.725372077198699e-05, 'train/grad_norm_pre_clip': 0.004979598801583052, 'timing/data': 0.04511001706123352, 'timing/model': 1.30260506644845, 'learning_rate/qwen_vl_interface': 4.032625477618266e-06, 'learning_rate/action_model': 4.0326254776182654e-05, 'epoch': 2.77}) +[2026-06-10 17:12:42,837][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 5.464869900606573e-06, 'train/grad_norm_pre_clip': 0.00016071568825282156, 'timing/data': 0.036464985460042953, 'timing/model': 1.1539812535047531, 'learning_rate/qwen_vl_interface': 3.99660301266448e-06, 'learning_rate/action_model': 3.99660301266448e-05, 'epoch': 2.78}) +[2026-06-10 17:13:07,300][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 1.3112974102114094e-06, 'train/grad_norm_pre_clip': 3.917287176591344e-05, 'timing/data': 0.04080888256430626, 'timing/model': 1.3114422895014286, 'learning_rate/qwen_vl_interface': 3.960657863140897e-06, 'learning_rate/action_model': 3.9606578631408975e-05, 'epoch': 2.79}) +[2026-06-10 17:13:17,527][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.04611610621213913, 'train/grad_norm_pre_clip': 0.8046335577964783, 'timing/data': 0.08151555433869362, 'timing/model': 1.126547947525978, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 2.8}) +[2026-06-10 17:13:41,727][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.08719674497842789, 'train/grad_norm_pre_clip': 1.9674383401870728, 'timing/data': 0.025556154549121857, 'timing/model': 1.3914163783192635, 'learning_rate/qwen_vl_interface': 3.889008374542957e-06, 'learning_rate/action_model': 3.889008374542957e-05, 'epoch': 2.81}) +[2026-06-10 17:14:01,903][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.04873788356781006, 'train/grad_norm_pre_clip': 1.0407525300979614, 'timing/data': 0.028254274278879166, 'timing/model': 1.163761030882597, 'learning_rate/qwen_vl_interface': 3.853308455146557e-06, 'learning_rate/action_model': 3.8533084551465566e-05, 'epoch': 2.82}) +[2026-06-10 17:14:20,898][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.0068053873255848885, 'train/grad_norm_pre_clip': 0.5026053786277771, 'timing/data': 0.023649368435144424, 'timing/model': 1.1570521667599678, 'learning_rate/qwen_vl_interface': 3.817694690263646e-06, 'learning_rate/action_model': 3.8176946902636455e-05, 'epoch': 2.83}) +[2026-06-10 17:14:47,484][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.0037244586274027824, 'train/grad_norm_pre_clip': 0.19491468369960785, 'timing/data': 0.030111484229564667, 'timing/model': 1.1806095987558365, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 2.83}) +[2026-06-10 17:15:09,913][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.06014806404709816, 'train/grad_norm_pre_clip': 1.0820177793502808, 'timing/data': 0.02918890491127968, 'timing/model': 1.1498304158449173, 'learning_rate/qwen_vl_interface': 3.746734405887466e-06, 'learning_rate/action_model': 3.746734405887466e-05, 'epoch': 2.84}) +[2026-06-10 17:15:33,544][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.085687056183815, 'train/grad_norm_pre_clip': 2.1938371658325195, 'timing/data': 0.028731465339660645, 'timing/model': 1.1676234863698483, 'learning_rate/qwen_vl_interface': 3.711392263558791e-06, 'learning_rate/action_model': 3.7113922635587913e-05, 'epoch': 2.85}) +[2026-06-10 17:15:59,396][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.0001405536022502929, 'train/grad_norm_pre_clip': 0.007224236149340868, 'timing/data': 0.025875955820083618, 'timing/model': 1.1484962403774261, 'learning_rate/qwen_vl_interface': 3.6761450298027874e-06, 'learning_rate/action_model': 3.676145029802787e-05, 'epoch': 2.86}) +[2026-06-10 17:16:08,885][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.005759888794273138, 'train/grad_norm_pre_clip': 0.05981161445379257, 'timing/data': 0.09190259873867035, 'timing/model': 1.100961860269308, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 2.87}) +[2026-06-10 17:16:33,602][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 4.2557141568977386e-05, 'train/grad_norm_pre_clip': 0.0020177876576781273, 'timing/data': 0.030095990747213364, 'timing/model': 1.2584675252437592, 'learning_rate/qwen_vl_interface': 3.605943978881907e-06, 'learning_rate/action_model': 3.605943978881907e-05, 'epoch': 2.88}) +[2026-06-10 17:16:54,062][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 4.637688834918663e-05, 'train/grad_norm_pre_clip': 0.004252057056874037, 'timing/data': 0.04007749259471893, 'timing/model': 1.1588090769946575, 'learning_rate/qwen_vl_interface': 3.5709944920485296e-06, 'learning_rate/action_model': 3.5709944920485295e-05, 'epoch': 2.88}) +[2026-06-10 17:17:13,808][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.003147445386275649, 'train/grad_norm_pre_clip': 0.12006255239248276, 'timing/data': 0.0630192719399929, 'timing/model': 1.1651524864137173, 'learning_rate/qwen_vl_interface': 3.5361485741837017e-06, 'learning_rate/action_model': 3.5361485741837014e-05, 'epoch': 2.89}) +[2026-06-10 17:17:32,448][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 1.8402909063297557e-06, 'train/grad_norm_pre_clip': 4.3585332605289295e-05, 'timing/data': 0.05494357645511627, 'timing/model': 1.3470415249466896, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 2.9}) +[2026-06-10 17:17:55,813][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 1.4277052287070546e-05, 'train/grad_norm_pre_clip': 0.000674737966619432, 'timing/data': 0.0412711575627327, 'timing/model': 1.2517237588763237, 'learning_rate/qwen_vl_interface': 3.4667760366799048e-06, 'learning_rate/action_model': 3.4667760366799043e-05, 'epoch': 2.91}) +[2026-06-10 17:18:23,048][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 350 +[2026-06-10 17:18:23,049][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.019339822232723236, 'train/grad_norm_pre_clip': 0.7966636419296265, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': -68.0643310546875, 'rl_games_eval/mid_train/mean_length': 21.8, 'rl_games_eval/mid_train/std_reward': 28.69797360047775, 'rl_games_eval/mid_train/std_length': 6.241794613730894, 'rl_games_eval/mid_train/macro_mean_reward': -68.0643310546875, 'rl_games_eval/mid_train/macro_mean_length': 21.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': -68.0643310546875, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 21.8, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 28.69797360047775, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 6.241794613730894, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': -68.0643310546875, 'checkpoint/best_score': 315.4344055175781, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.04193301871418953, 'timing/model': 1.0223900564014912, 'learning_rate/qwen_vl_interface': 3.432253696265824e-06, 'learning_rate/action_model': 3.432253696265824e-05, 'epoch': 2.92}) +[2026-06-10 17:18:23,061][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_350_state +[2026-06-10 17:18:23,068][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 17:18:29,683][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_350_state/model.safetensors +[2026-06-10 17:18:46,634][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_350_state/optimizer.bin +[2026-06-10 17:18:46,635][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_350_state/sampler.bin +[2026-06-10 17:18:46,635][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_350_state/sampler_1.bin +[2026-06-10 17:18:46,638][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_350_state/random_states_0.pkl +[2026-06-10 17:18:46,643][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 350 +[2026-06-10 17:18:46,644][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 17:18:46,663][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 17:19:06,026][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.0073062460869550705, 'train/grad_norm_pre_clip': 0.377280056476593, 'timing/data': 0.037604961544275284, 'timing/model': 1.1697770319879055, 'learning_rate/qwen_vl_interface': 3.397843483006107e-06, 'learning_rate/action_model': 3.397843483006107e-05, 'epoch': 2.92}) +[2026-06-10 17:19:15,280][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.017422689124941826, 'train/grad_norm_pre_clip': 0.293449729681015, 'timing/data': 0.08298779278993607, 'timing/model': 1.3756688870489597, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 2.93}) +[2026-06-10 17:19:37,211][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.003689988050609827, 'train/grad_norm_pre_clip': 0.048495449125766754, 'timing/data': 0.02458496391773224, 'timing/model': 1.265605378895998, 'learning_rate/qwen_vl_interface': 3.3293679212384616e-06, 'learning_rate/action_model': 3.329367921238461e-05, 'epoch': 2.94}) +[2026-06-10 17:19:56,005][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 8.509321924066171e-05, 'train/grad_norm_pre_clip': 0.003161541884765029, 'timing/data': 0.030391082167625427, 'timing/model': 1.221861608326435, 'learning_rate/qwen_vl_interface': 3.2953067966257334e-06, 'learning_rate/action_model': 3.295306796625733e-05, 'epoch': 2.95}) +[2026-06-10 17:20:18,010][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 4.96186612508609e-06, 'train/grad_norm_pre_clip': 0.0002513567160349339, 'timing/data': 0.023221787065267563, 'timing/model': 1.449481539428234, 'learning_rate/qwen_vl_interface': 3.2613662466972178e-06, 'learning_rate/action_model': 3.2613662466972176e-05, 'epoch': 2.96}) +[2026-06-10 17:20:39,437][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 1.6912772480282001e-06, 'train/grad_norm_pre_clip': 4.1958061046898365e-05, 'timing/data': 0.026896368712186813, 'timing/model': 1.2480749897658825, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 2.97}) +[2026-06-10 17:20:59,434][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.1565319448709488, 'train/grad_norm_pre_clip': 1.398966670036316, 'timing/data': 0.03616224229335785, 'timing/model': 1.1565456129610538, 'learning_rate/qwen_vl_interface': 3.193855237778036e-06, 'learning_rate/action_model': 3.1938552377780355e-05, 'epoch': 2.98}) +[2026-06-10 17:21:19,958][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.00039968331111595035, 'train/grad_norm_pre_clip': 0.021127790212631226, 'timing/data': 0.03609168529510498, 'timing/model': 1.7959799505770206, 'learning_rate/qwen_vl_interface': 3.160288943184403e-06, 'learning_rate/action_model': 3.160288943184403e-05, 'epoch': 2.98}) +[2026-06-10 17:21:39,694][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 5.908040293434169e-06, 'train/grad_norm_pre_clip': 0.00041791406692937016, 'timing/data': 0.04119955375790596, 'timing/model': 1.0941482819616795, 'learning_rate/qwen_vl_interface': 3.126851551812175e-06, 'learning_rate/action_model': 3.126851551812175e-05, 'epoch': 2.99}) +[2026-06-10 17:21:49,572][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.000503379967994988, 'train/grad_norm_pre_clip': 0.026441438123583794, 'timing/data': 0.09922266751527786, 'timing/model': 1.2958611398935318, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 3.0}) +[2026-06-10 17:22:12,522][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.005452566780149937, 'train/grad_norm_pre_clip': 0.08447841554880142, 'timing/data': 0.03748128190636635, 'timing/model': 1.2518117241561413, 'learning_rate/qwen_vl_interface': 3.0603717209565546e-06, 'learning_rate/action_model': 3.0603717209565546e-05, 'epoch': 3.01}) +[2026-06-10 17:22:31,460][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 5.997715106786927e-07, 'train/grad_norm_pre_clip': 1.2487338608480059e-05, 'timing/data': 0.04377265274524689, 'timing/model': 1.8398981057107449, 'learning_rate/qwen_vl_interface': 3.027333382262277e-06, 'learning_rate/action_model': 3.027333382262277e-05, 'epoch': 3.02}) +[2026-06-10 17:22:54,098][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 5.8272747992305085e-05, 'train/grad_norm_pre_clip': 0.0023345404770225286, 'timing/data': 0.03682598099112511, 'timing/model': 1.1532929949462414, 'learning_rate/qwen_vl_interface': 2.9944321481146766e-06, 'learning_rate/action_model': 2.9944321481146764e-05, 'epoch': 3.02}) +[2026-06-10 17:23:12,726][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.0005338791524991393, 'train/grad_norm_pre_clip': 0.03302833437919617, 'timing/data': 0.05669533461332321, 'timing/model': 1.251150757074356, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 3.03}) +[2026-06-10 17:23:30,974][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.0005327627877704799, 'train/grad_norm_pre_clip': 0.03427024558186531, 'timing/data': 0.031085938215255737, 'timing/model': 1.1729585826396942, 'learning_rate/qwen_vl_interface': 2.9290491028894658e-06, 'learning_rate/action_model': 2.9290491028894658e-05, 'epoch': 3.04}) +[2026-06-10 17:23:51,228][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 6.831588507338893e-06, 'train/grad_norm_pre_clip': 0.0003550080000422895, 'timing/data': 0.04444987699389458, 'timing/model': 1.2230960801243782, 'learning_rate/qwen_vl_interface': 2.8965713249460643e-06, 'learning_rate/action_model': 2.896571324946064e-05, 'epoch': 3.05}) +[2026-06-10 17:24:09,823][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.00011229155643377453, 'train/grad_norm_pre_clip': 0.007821639068424702, 'timing/data': 0.04730233550071716, 'timing/model': 1.1403126902878284, 'learning_rate/qwen_vl_interface': 2.8642387175689806e-06, 'learning_rate/action_model': 2.8642387175689805e-05, 'epoch': 3.06}) +[2026-06-10 17:24:19,297][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 4.045384775963612e-05, 'train/grad_norm_pre_clip': 0.006608663126826286, 'timing/data': 0.09045517817139626, 'timing/model': 1.1677353046834469, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 3.07}) +[2026-06-10 17:24:43,669][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 6.0712878621416166e-05, 'train/grad_norm_pre_clip': 0.004770250059664249, 'timing/data': 0.037879545241594315, 'timing/model': 1.2875020802021027, 'learning_rate/qwen_vl_interface': 2.800016983146025e-06, 'learning_rate/action_model': 2.800016983146025e-05, 'epoch': 3.08}) +[2026-06-10 17:25:02,837][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 8.410711598116904e-05, 'train/grad_norm_pre_clip': 0.0031826330814510584, 'timing/data': 0.03982190042734146, 'timing/model': 1.2365125194191933, 'learning_rate/qwen_vl_interface': 2.768131817599243e-06, 'learning_rate/action_model': 2.768131817599243e-05, 'epoch': 3.08}) +[2026-06-10 17:25:21,821][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.0001904610835481435, 'train/grad_norm_pre_clip': 0.01170884259045124, 'timing/data': 0.030629396438598633, 'timing/model': 1.1572952196002007, 'learning_rate/qwen_vl_interface': 2.7363997453725954e-06, 'learning_rate/action_model': 2.7363997453725954e-05, 'epoch': 3.09}) +[2026-06-10 17:25:40,405][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.0005106329335831106, 'train/grad_norm_pre_clip': 0.024908456951379776, 'timing/data': 0.034495219588279724, 'timing/model': 1.363433301448822, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 3.1}) +[2026-06-10 17:26:06,854][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 3.1061190384207293e-05, 'train/grad_norm_pre_clip': 0.0018092453246936202, 'timing/data': 0.03750940039753914, 'timing/model': 1.1618459448218346, 'learning_rate/qwen_vl_interface': 2.6734027008501866e-06, 'learning_rate/action_model': 2.6734027008501865e-05, 'epoch': 3.11}) +[2026-06-10 17:26:26,318][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.004445676226168871, 'train/grad_norm_pre_clip': 0.20256245136260986, 'timing/data': 0.036680564284324646, 'timing/model': 1.1807368136942387, 'learning_rate/qwen_vl_interface': 2.6421416145088737e-06, 'learning_rate/action_model': 2.6421416145088736e-05, 'epoch': 3.12}) +[2026-06-10 17:26:44,688][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 9.199215128319338e-05, 'train/grad_norm_pre_clip': 0.007084011100232601, 'timing/data': 0.029609326273202896, 'timing/model': 1.3132681846618652, 'learning_rate/qwen_vl_interface': 2.6110413931568906e-06, 'learning_rate/action_model': 2.6110413931568904e-05, 'epoch': 3.12}) +[2026-06-10 17:26:53,730][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.004653146490454674, 'train/grad_norm_pre_clip': 0.06664399057626724, 'timing/data': 0.13528139516711235, 'timing/model': 1.0460097678005695, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 3.13}) +[2026-06-10 17:27:16,855][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 3.9860526612756075e-07, 'train/grad_norm_pre_clip': 2.8009473680867814e-05, 'timing/data': 0.012390118092298508, 'timing/model': 1.1698399931192398, 'learning_rate/qwen_vl_interface': 2.5493312090122463e-06, 'learning_rate/action_model': 2.549331209012246e-05, 'epoch': 3.14}) +[2026-06-10 17:27:34,913][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.14250639081001282, 'train/grad_norm_pre_clip': 2.265990734100342, 'timing/data': 0.010576624423265457, 'timing/model': 1.17841712012887, 'learning_rate/qwen_vl_interface': 2.5187250527944283e-06, 'learning_rate/action_model': 2.5187250527944284e-05, 'epoch': 3.15}) +[2026-06-10 17:27:55,048][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 7.189776169980178e-07, 'train/grad_norm_pre_clip': 5.386414341046475e-05, 'timing/data': 0.014182224869728088, 'timing/model': 1.507627610117197, 'learning_rate/qwen_vl_interface': 2.488287374480805e-06, 'learning_rate/action_model': 2.4882873744808046e-05, 'epoch': 3.16}) +[2026-06-10 17:28:14,403][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.12020054459571838, 'train/grad_norm_pre_clip': 2.090059518814087, 'timing/data': 0.009336359798908234, 'timing/model': 1.187607977539301, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 3.17}) +[2026-06-10 17:28:38,795][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.1146632507443428, 'train/grad_norm_pre_clip': 1.862546682357788, 'timing/data': 0.012615565210580826, 'timing/model': 1.2528555057942867, 'learning_rate/qwen_vl_interface': 2.427924951215317e-06, 'learning_rate/action_model': 2.427924951215317e-05, 'epoch': 3.17}) +[2026-06-10 17:29:05,202][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.0025261982809752226, 'train/grad_norm_pre_clip': 1.2175474166870117, 'timing/data': 0.013207390904426575, 'timing/model': 1.8726524412631989, 'learning_rate/qwen_vl_interface': 2.3980039297020513e-06, 'learning_rate/action_model': 2.3980039297020513e-05, 'epoch': 3.18}) +[2026-06-10 17:29:30,298][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.0004186570877209306, 'train/grad_norm_pre_clip': 0.014248873107135296, 'timing/data': 0.000609125941991806, 'timing/model': 1.2283224947750568, 'learning_rate/qwen_vl_interface': 2.3682588327405025e-06, 'learning_rate/action_model': 2.3682588327405024e-05, 'epoch': 3.19}) +[2026-06-10 17:29:40,712][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.0010784817859530449, 'train/grad_norm_pre_clip': 0.11805459856987, 'timing/data': 0.06711330264806747, 'timing/model': 1.097076203674078, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 3.2}) +[2026-06-10 17:30:06,953][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.030977608636021614, 'train/grad_norm_pre_clip': 1.0290271043777466, 'timing/data': 0.02784096822142601, 'timing/model': 1.2238416373729706, 'learning_rate/qwen_vl_interface': 2.3093037407782883e-06, 'learning_rate/action_model': 2.309303740778288e-05, 'epoch': 3.21}) +[2026-06-10 17:30:26,815][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 7.051523425616324e-06, 'train/grad_norm_pre_clip': 0.00041733135003596544, 'timing/data': 0.03124471753835678, 'timing/model': 1.1587771251797676, 'learning_rate/qwen_vl_interface': 2.280097382405402e-06, 'learning_rate/action_model': 2.2800973824054018e-05, 'epoch': 3.22}) +[2026-06-10 17:30:45,495][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 1.2184655133751221e-05, 'train/grad_norm_pre_clip': 0.0005807421403005719, 'timing/data': 0.022273380309343338, 'timing/model': 1.1356959007680416, 'learning_rate/qwen_vl_interface': 2.2510742216154557e-06, 'learning_rate/action_model': 2.2510742216154557e-05, 'epoch': 3.23}) +[2026-06-10 17:31:03,525][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.0014260578900575638, 'train/grad_norm_pre_clip': 0.07165415585041046, 'timing/data': 0.014829088002443314, 'timing/model': 1.1686363965272903, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 3.23}) +[2026-06-10 17:31:21,628][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.006807693745940924, 'train/grad_norm_pre_clip': 0.3448044955730438, 'timing/data': 0.011383645236492157, 'timing/model': 1.225974828004837, 'learning_rate/qwen_vl_interface': 2.19358464251455e-06, 'learning_rate/action_model': 2.19358464251455e-05, 'epoch': 3.24}) +[2026-06-10 17:31:42,199][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.015283513814210892, 'train/grad_norm_pre_clip': 0.5112901926040649, 'timing/data': 0.013746388256549835, 'timing/model': 1.1339417435228825, 'learning_rate/qwen_vl_interface': 2.1651217704316283e-06, 'learning_rate/action_model': 2.165121770431628e-05, 'epoch': 3.25}) +[2026-06-10 17:32:00,482][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.001926774624735117, 'train/grad_norm_pre_clip': 0.10861575603485107, 'timing/data': 0.015009749680757523, 'timing/model': 1.1388979516923428, 'learning_rate/qwen_vl_interface': 2.1368491881689794e-06, 'learning_rate/action_model': 2.136849188168979e-05, 'epoch': 3.26}) +[2026-06-10 17:32:09,653][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.009628970175981522, 'train/grad_norm_pre_clip': 0.5757426619529724, 'timing/data': 0.07853596284985542, 'timing/model': 1.0445640347898006, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 3.27}) +[2026-06-10 17:32:30,810][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.005220182705670595, 'train/grad_norm_pre_clip': 0.26449915766716003, 'timing/data': 0.022464636713266373, 'timing/model': 1.1622773595154285, 'learning_rate/qwen_vl_interface': 2.0808818572031507e-06, 'learning_rate/action_model': 2.0808818572031506e-05, 'epoch': 3.27}) +[2026-06-10 17:32:51,632][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.006937697064131498, 'train/grad_norm_pre_clip': 0.17633098363876343, 'timing/data': 0.020842134952545166, 'timing/model': 1.8861387148499489, 'learning_rate/qwen_vl_interface': 2.053190560828578e-06, 'learning_rate/action_model': 2.0531905608285777e-05, 'epoch': 3.28}) +[2026-06-10 17:33:19,983][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.004640525672584772, 'train/grad_norm_pre_clip': 0.350102961063385, 'timing/data': 0.027659572660923004, 'timing/model': 1.3850160613656044, 'learning_rate/qwen_vl_interface': 2.0256964587185266e-06, 'learning_rate/action_model': 2.0256964587185266e-05, 'epoch': 3.29}) +[2026-06-10 17:33:38,550][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 8.494991925545037e-05, 'train/grad_norm_pre_clip': 0.003688399912789464, 'timing/data': 0.018310025334358215, 'timing/model': 1.1906427778303623, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 3.3}) +[2026-06-10 17:33:57,515][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.010246883146464825, 'train/grad_norm_pre_clip': 0.5626828074455261, 'timing/data': 0.0237894169986248, 'timing/model': 1.2273955792188644, 'learning_rate/qwen_vl_interface': 1.9713066088864233e-06, 'learning_rate/action_model': 1.9713066088864232e-05, 'epoch': 3.31}) +[2026-06-10 17:34:16,430][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 9.675035107648e-05, 'train/grad_norm_pre_clip': 0.0047108749859035015, 'timing/data': 0.016907893121242523, 'timing/model': 1.162874985486269, 'learning_rate/qwen_vl_interface': 1.944414216186507e-06, 'learning_rate/action_model': 1.944414216186507e-05, 'epoch': 3.32}) +[2026-06-10 17:34:35,904][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.0004884445806965232, 'train/grad_norm_pre_clip': 0.03432855382561684, 'timing/data': 0.014210853725671768, 'timing/model': 1.347554013133049, 'learning_rate/qwen_vl_interface': 1.917725727588435e-06, 'learning_rate/action_model': 1.9177257275884353e-05, 'epoch': 3.33}) +[2026-06-10 17:35:01,090][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 400 +[2026-06-10 17:35:01,091][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 4.2534327803878114e-05, 'train/grad_norm_pre_clip': 0.003717288840562105, 'eval/loss': 0.00886576808989048, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 197.88494262695312, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 190.44776622469652, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 197.88494262695312, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 197.88494262695312, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 190.44776622469652, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 197.88494262695312, 'checkpoint/best_score': 315.4344055175781, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0672186017036438, 'timing/model': 1.2504776865243912, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 3.33}) +[2026-06-10 17:35:01,092][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_400_state +[2026-06-10 17:35:01,099][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 17:35:07,520][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_400_state/model.safetensors +[2026-06-10 17:35:22,637][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_400_state/optimizer.bin +[2026-06-10 17:35:22,638][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_400_state/sampler.bin +[2026-06-10 17:35:22,638][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_400_state/sampler_1.bin +[2026-06-10 17:35:22,640][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-10 17:35:22,648][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-10 17:35:22,648][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 17:35:22,664][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 17:35:53,522][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.00020986658637411892, 'train/grad_norm_pre_clip': 0.02259056083858013, 'timing/data': 0.03908335044980049, 'timing/model': 2.3464281298220158, 'learning_rate/qwen_vl_interface': 1.8649670351052746e-06, 'learning_rate/action_model': 1.8649670351052744e-05, 'epoch': 3.34}) +[2026-06-10 17:36:16,086][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.006395032163709402, 'train/grad_norm_pre_clip': 0.14393983781337738, 'timing/data': 0.03433096036314964, 'timing/model': 1.2056798078119755, 'learning_rate/qwen_vl_interface': 1.8389000856248537e-06, 'learning_rate/action_model': 1.8389000856248537e-05, 'epoch': 3.35}) +[2026-06-10 17:36:35,438][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.0005999221466481686, 'train/grad_norm_pre_clip': 0.04268663376569748, 'timing/data': 0.05008132755756378, 'timing/model': 1.281190488487482, 'learning_rate/qwen_vl_interface': 1.8130435488548661e-06, 'learning_rate/action_model': 1.8130435488548662e-05, 'epoch': 3.36}) +[2026-06-10 17:36:54,054][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.010422048158943653, 'train/grad_norm_pre_clip': 0.45600929856300354, 'timing/data': 0.03490285947918892, 'timing/model': 1.1430554948747158, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 3.37}) +[2026-06-10 17:37:12,249][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.0037669676821678877, 'train/grad_norm_pre_clip': 0.23768626153469086, 'timing/data': 0.04539896175265312, 'timing/model': 1.1327195391058922, 'learning_rate/qwen_vl_interface': 1.7619680801804947e-06, 'learning_rate/action_model': 1.7619680801804947e-05, 'epoch': 3.38}) +[2026-06-10 17:37:30,433][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 6.250480964808958e-06, 'train/grad_norm_pre_clip': 0.00035758534795604646, 'timing/data': 0.04533091187477112, 'timing/model': 1.3046934716403484, 'learning_rate/qwen_vl_interface': 1.7367522988516035e-06, 'learning_rate/action_model': 1.7367522988516035e-05, 'epoch': 3.38}) +[2026-06-10 17:37:49,122][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 1.0169959523409489e-06, 'train/grad_norm_pre_clip': 7.579723023809493e-05, 'timing/data': 0.03197639435529709, 'timing/model': 1.1558168977499008, 'learning_rate/qwen_vl_interface': 1.7117532311897967e-06, 'learning_rate/action_model': 1.7117532311897966e-05, 'epoch': 3.39}) +[2026-06-10 17:37:58,156][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 5.036119546275586e-05, 'train/grad_norm_pre_clip': 0.0038204628508538008, 'timing/data': 0.08503572642803192, 'timing/model': 0.9172630794346333, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 3.4}) +[2026-06-10 17:38:17,273][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 1.2330517620284809e-06, 'train/grad_norm_pre_clip': 5.5303196859313175e-05, 'timing/data': 0.02611268311738968, 'timing/model': 1.1661478132009506, 'learning_rate/qwen_vl_interface': 1.6624113916453782e-06, 'learning_rate/action_model': 1.6624113916453783e-05, 'epoch': 3.41}) +[2026-06-10 17:38:37,483][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.004174513276666403, 'train/grad_norm_pre_clip': 0.3317825198173523, 'timing/data': 0.024115703999996185, 'timing/model': 1.130642507225275, 'learning_rate/qwen_vl_interface': 1.6380716633998532e-06, 'learning_rate/action_model': 1.638071663399853e-05, 'epoch': 3.42}) +[2026-06-10 17:38:56,351][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.000279736501397565, 'train/grad_norm_pre_clip': 0.02433808147907257, 'timing/data': 0.026951324194669724, 'timing/model': 1.149770688265562, 'learning_rate/qwen_vl_interface': 1.613954735907832e-06, 'learning_rate/action_model': 1.613954735907832e-05, 'epoch': 3.42}) +[2026-06-10 17:39:15,071][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 1.8253909672694135e-07, 'train/grad_norm_pre_clip': 5.202875854593003e-06, 'timing/data': 0.030513737350702286, 'timing/model': 1.1508035883307457, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 3.43}) +[2026-06-10 17:39:35,197][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 3.807351822615601e-05, 'train/grad_norm_pre_clip': 0.0015609173569828272, 'timing/data': 0.027183346450328827, 'timing/model': 1.1472942605614662, 'learning_rate/qwen_vl_interface': 1.5663952199318869e-06, 'learning_rate/action_model': 1.5663952199318867e-05, 'epoch': 3.44}) +[2026-06-10 17:39:55,734][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 3.529737296048552e-05, 'train/grad_norm_pre_clip': 0.001709696021862328, 'timing/data': 0.040144193917512894, 'timing/model': 1.229393444955349, 'learning_rate/qwen_vl_interface': 1.5429555651429334e-06, 'learning_rate/action_model': 1.5429555651429335e-05, 'epoch': 3.45}) +[2026-06-10 17:40:17,115][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.00015354208881035447, 'train/grad_norm_pre_clip': 0.01282199565321207, 'timing/data': 0.027554526925086975, 'timing/model': 1.3063585422933102, 'learning_rate/qwen_vl_interface': 1.5197445783164627e-06, 'learning_rate/action_model': 1.5197445783164626e-05, 'epoch': 3.46}) +[2026-06-10 17:40:26,607][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 8.983213888313912e-07, 'train/grad_norm_pre_clip': 6.224403477972373e-05, 'timing/data': 0.10683256760239601, 'timing/model': 1.1480639763176441, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 3.47}) +[2026-06-10 17:40:47,203][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 1.958944085345138e-05, 'train/grad_norm_pre_clip': 0.0011245071655139327, 'timing/data': 0.03237258642911911, 'timing/model': 1.2052017711102962, 'learning_rate/qwen_vl_interface': 1.4740143214093276e-06, 'learning_rate/action_model': 1.4740143214093274e-05, 'epoch': 3.48}) +[2026-06-10 17:41:09,007][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 6.70552182668871e-08, 'train/grad_norm_pre_clip': 1.5048566410769126e-06, 'timing/data': 0.050983358174562454, 'timing/model': 1.396058727055788, 'learning_rate/qwen_vl_interface': 1.4514978721863198e-06, 'learning_rate/action_model': 1.4514978721863198e-05, 'epoch': 3.48}) +[2026-06-10 17:41:27,429][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 6.109434593781771e-07, 'train/grad_norm_pre_clip': 2.9675655241589993e-05, 'timing/data': 0.03196335211396217, 'timing/model': 1.149978905916214, 'learning_rate/qwen_vl_interface': 1.4292157324670988e-06, 'learning_rate/action_model': 1.4292157324670987e-05, 'epoch': 3.49}) +[2026-06-10 17:41:50,359][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.0005909870960749686, 'train/grad_norm_pre_clip': 0.03581542894244194, 'timing/data': 0.02647862583398819, 'timing/model': 1.174943145364523, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 3.5}) +[2026-06-10 17:42:14,316][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.0631512850522995, 'train/grad_norm_pre_clip': 1.4979749917984009, 'timing/data': 0.024063829332590103, 'timing/model': 1.3106671534478664, 'learning_rate/qwen_vl_interface': 1.3853598648712621e-06, 'learning_rate/action_model': 1.385359864871262e-05, 'epoch': 3.51}) +[2026-06-10 17:42:33,712][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 5.587935092421503e-08, 'train/grad_norm_pre_clip': 1.020643139781896e-06, 'timing/data': 0.028782188892364502, 'timing/model': 1.322474755346775, 'learning_rate/qwen_vl_interface': 1.3637888422311379e-06, 'learning_rate/action_model': 1.3637888422311379e-05, 'epoch': 3.52}) +[2026-06-10 17:42:53,405][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.023878537118434906, 'train/grad_norm_pre_clip': 1.055737853050232, 'timing/data': 0.02442655712366104, 'timing/model': 1.1217775344848633, 'learning_rate/qwen_vl_interface': 1.3424575394009165e-06, 'learning_rate/action_model': 1.3424575394009165e-05, 'epoch': 3.52}) +[2026-06-10 17:43:02,953][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 1.1920921139108032e-07, 'train/grad_norm_pre_clip': 4.195078872726299e-06, 'timing/data': 0.12926925346255302, 'timing/model': 1.0739307962357998, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 3.53}) +[2026-06-10 17:43:25,216][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.010624507442116737, 'train/grad_norm_pre_clip': 0.37424996495246887, 'timing/data': 0.018336281180381775, 'timing/model': 1.3004385121166706, 'learning_rate/qwen_vl_interface': 1.3005193415629098e-06, 'learning_rate/action_model': 1.3005193415629097e-05, 'epoch': 3.54}) +[2026-06-10 17:43:49,411][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 3.390002234482381e-07, 'train/grad_norm_pre_clip': 1.7772421415429562e-05, 'timing/data': 0.018608704209327698, 'timing/model': 3.2600428573787212, 'learning_rate/qwen_vl_interface': 1.2799150335007174e-06, 'learning_rate/action_model': 1.2799150335007173e-05, 'epoch': 3.55}) +[2026-06-10 17:44:09,412][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 4.845108924200758e-05, 'train/grad_norm_pre_clip': 0.004157716874033213, 'timing/data': 0.021829620003700256, 'timing/model': 1.2939574271440506, 'learning_rate/qwen_vl_interface': 1.259555618980026e-06, 'learning_rate/action_model': 1.259555618980026e-05, 'epoch': 3.56}) +[2026-06-10 17:44:28,079][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 2.0445244444999844e-05, 'train/grad_norm_pre_clip': 0.0013165019918233156, 'timing/data': 0.02083594724535942, 'timing/model': 1.360239151865244, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 3.57}) +[2026-06-10 17:44:46,711][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.0023878731299191713, 'train/grad_norm_pre_clip': 0.24156755208969116, 'timing/data': 0.01708119362592697, 'timing/model': 1.0524676106870174, 'learning_rate/qwen_vl_interface': 1.219576478837858e-06, 'learning_rate/action_model': 1.219576478837858e-05, 'epoch': 3.58}) +[2026-06-10 17:45:07,224][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 1.3522603694582358e-06, 'train/grad_norm_pre_clip': 0.00012042662274325266, 'timing/data': 0.024599626660346985, 'timing/model': 1.1989398561418056, 'learning_rate/qwen_vl_interface': 1.1999592193180623e-06, 'learning_rate/action_model': 1.1999592193180623e-05, 'epoch': 3.58}) +[2026-06-10 17:45:27,607][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 8.474082278553396e-06, 'train/grad_norm_pre_clip': 0.0006293937331065536, 'timing/data': 0.01347028836607933, 'timing/model': 1.2095003947615623, 'learning_rate/qwen_vl_interface': 1.180591785391015e-06, 'learning_rate/action_model': 1.1805917853910148e-05, 'epoch': 3.59}) +[2026-06-10 17:45:45,588][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 7.237705546003781e-08, 'train/grad_norm_pre_clip': 1.941953314599232e-06, 'timing/data': 0.16487637907266617, 'timing/model': 1.1732411049306393, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 3.6}) +[2026-06-10 17:46:11,741][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.07768262922763824, 'train/grad_norm_pre_clip': 0.7969022989273071, 'timing/data': 0.021173615008592606, 'timing/model': 1.141910944133997, 'learning_rate/qwen_vl_interface': 1.1426111575292639e-06, 'learning_rate/action_model': 1.1426111575292639e-05, 'epoch': 3.61}) +[2026-06-10 17:46:29,899][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 4.0978189730367376e-08, 'train/grad_norm_pre_clip': 1.5460129816347035e-06, 'timing/data': 0.018028445541858673, 'timing/model': 1.1403607986867428, 'learning_rate/qwen_vl_interface': 1.1240003064185923e-06, 'learning_rate/action_model': 1.1240003064185923e-05, 'epoch': 3.62}) +[2026-06-10 17:46:48,802][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.00060388317797333, 'train/grad_norm_pre_clip': 0.04333922266960144, 'timing/data': 0.0629124641418457, 'timing/model': 1.3874146193265915, 'learning_rate/qwen_vl_interface': 1.1056439664042135e-06, 'learning_rate/action_model': 1.1056439664042134e-05, 'epoch': 3.62}) +[2026-06-10 17:47:09,904][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 1.043080430918053e-07, 'train/grad_norm_pre_clip': 3.849995664495509e-06, 'timing/data': 0.013705600053071976, 'timing/model': 1.3940331675112247, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 3.63}) +[2026-06-10 17:47:31,986][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.0008725984371267259, 'train/grad_norm_pre_clip': 0.12530383467674255, 'timing/data': 0.019825678318738937, 'timing/model': 1.2189614325761795, 'learning_rate/qwen_vl_interface': 1.0696993331171232e-06, 'learning_rate/action_model': 1.0696993331171231e-05, 'epoch': 3.64}) +[2026-06-10 17:47:51,079][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.015814557671546936, 'train/grad_norm_pre_clip': 0.9514897465705872, 'timing/data': 0.023586485534906387, 'timing/model': 1.3004928156733513, 'learning_rate/qwen_vl_interface': 1.052113257078706e-06, 'learning_rate/action_model': 1.0521132570787059e-05, 'epoch': 3.65}) +[2026-06-10 17:48:09,084][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 6.70552182668871e-08, 'train/grad_norm_pre_clip': 7.217240067802777e-07, 'timing/data': 0.017045576125383377, 'timing/model': 1.1854489520192146, 'learning_rate/qwen_vl_interface': 1.034786126468405e-06, 'learning_rate/action_model': 1.034786126468405e-05, 'epoch': 3.66}) +[2026-06-10 17:48:18,258][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 3.4059794984386826e-08, 'train/grad_norm_pre_clip': 3.8072707297942543e-07, 'timing/data': 0.07610524445772171, 'timing/model': 1.2964899353682995, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 3.67}) +[2026-06-10 17:48:38,011][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.11301029473543167, 'train/grad_norm_pre_clip': 0.5829253792762756, 'timing/data': 0.04117279499769211, 'timing/model': 1.171090591698885, 'learning_rate/qwen_vl_interface': 1.0009129607693812e-06, 'learning_rate/action_model': 1.0009129607693813e-05, 'epoch': 3.67}) +[2026-06-10 17:48:57,793][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 4.0978189730367376e-08, 'train/grad_norm_pre_clip': 7.367685270764923e-07, 'timing/data': 0.033222682774066925, 'timing/model': 1.131987825036049, 'learning_rate/qwen_vl_interface': 9.843690151370762e-07, 'learning_rate/action_model': 9.843690151370762e-06, 'epoch': 3.68}) +[2026-06-10 17:49:17,358][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.0024385324213653803, 'train/grad_norm_pre_clip': 0.037023063749074936, 'timing/data': 0.03921256586909294, 'timing/model': 1.1427314914762974, 'learning_rate/qwen_vl_interface': 9.680881937168323e-07, 'learning_rate/action_model': 9.680881937168324e-06, 'epoch': 3.69}) +[2026-06-10 17:49:39,904][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.014819753356277943, 'train/grad_norm_pre_clip': 0.841421902179718, 'timing/data': 0.03210048750042915, 'timing/model': 1.1451448500156403, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 3.7}) +[2026-06-10 17:50:02,608][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.013373462483286858, 'train/grad_norm_pre_clip': 0.9218373894691467, 'timing/data': 0.04402914270758629, 'timing/model': 1.3641971945762634, 'learning_rate/qwen_vl_interface': 9.363199243308635e-07, 'learning_rate/action_model': 9.363199243308635e-06, 'epoch': 3.71}) +[2026-06-10 17:50:21,788][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 5.587935092421503e-08, 'train/grad_norm_pre_clip': 1.8394774770058575e-06, 'timing/data': 0.03396682068705559, 'timing/model': 1.1537414826452732, 'learning_rate/qwen_vl_interface': 9.208344359816349e-07, 'learning_rate/action_model': 9.208344359816348e-06, 'epoch': 3.72}) +[2026-06-10 17:50:41,342][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.09173291176557541, 'train/grad_norm_pre_clip': 0.40116989612579346, 'timing/data': 0.03797398507595062, 'timing/model': 1.1502277962863445, 'learning_rate/qwen_vl_interface': 9.056159909565854e-07, 'learning_rate/action_model': 9.056159909565854e-06, 'epoch': 3.73}) +[2026-06-10 17:50:50,814][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.06360223144292831, 'train/grad_norm_pre_clip': 2.279344081878662, 'timing/data': 0.07903238385915756, 'timing/model': 1.0964844301342964, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 3.73}) +[2026-06-10 17:51:10,193][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.011659101583063602, 'train/grad_norm_pre_clip': 0.7045127153396606, 'timing/data': 0.009382244199514389, 'timing/model': 1.1790306568145752, 'learning_rate/qwen_vl_interface': 8.759839693301256e-07, 'learning_rate/action_model': 8.759839693301255e-06, 'epoch': 3.74}) +[2026-06-10 17:51:35,932][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 450 +[2026-06-10 17:51:35,933][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 1.0058275989877075e-07, 'train/grad_norm_pre_clip': 7.320047643588623e-06, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 226.8769104003906, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 178.0779475509533, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 226.8769104003906, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 226.8769104003906, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 178.0779475509533, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 226.8769104003906, 'checkpoint/best_score': 315.4344055175781, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.007783282548189163, 'timing/model': 1.379121407866478, 'learning_rate/qwen_vl_interface': 8.615722205713881e-07, 'learning_rate/action_model': 8.61572220571388e-06, 'epoch': 3.75}) +[2026-06-10 17:51:35,935][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_450_state +[2026-06-10 17:51:35,940][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 17:51:44,340][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_450_state/model.safetensors +[2026-06-10 17:51:59,781][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_450_state/optimizer.bin +[2026-06-10 17:51:59,784][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_450_state/sampler.bin +[2026-06-10 17:51:59,784][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_450_state/sampler_1.bin +[2026-06-10 17:51:59,786][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_450_state/random_states_0.pkl +[2026-06-10 17:51:59,791][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 450 +[2026-06-10 17:51:59,792][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 17:51:59,804][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 17:52:31,027][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 3.861258301185444e-05, 'train/grad_norm_pre_clip': 0.004248875193297863, 'timing/data': 0.008969690650701523, 'timing/model': 4.359371814876795, 'learning_rate/qwen_vl_interface': 8.474311707094262e-07, 'learning_rate/action_model': 8.474311707094263e-06, 'epoch': 3.76}) +[2026-06-10 17:52:52,104][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.0014883405528962612, 'train/grad_norm_pre_clip': 0.2161422222852707, 'timing/data': 0.00998150184750557, 'timing/model': 1.188283510506153, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 3.77}) +[2026-06-10 17:53:11,124][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.0020613684318959713, 'train/grad_norm_pre_clip': 0.026196617633104324, 'timing/data': 0.01129186525940895, 'timing/model': 1.1915363185107708, 'learning_rate/qwen_vl_interface': 8.199646400703035e-07, 'learning_rate/action_model': 8.199646400703035e-06, 'epoch': 3.77}) +[2026-06-10 17:53:31,798][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 2.4846829091984546e-06, 'train/grad_norm_pre_clip': 0.00014671003737021238, 'timing/data': 0.00737067312002182, 'timing/model': 1.3582053668797016, 'learning_rate/qwen_vl_interface': 8.066408535581304e-07, 'learning_rate/action_model': 8.066408535581304e-06, 'epoch': 3.78}) +[2026-06-10 17:53:50,507][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.000265078415395692, 'train/grad_norm_pre_clip': 0.02063964493572712, 'timing/data': 0.009153135120868683, 'timing/model': 1.4437789432704449, 'learning_rate/qwen_vl_interface': 7.935911543682008e-07, 'learning_rate/action_model': 7.935911543682007e-06, 'epoch': 3.79}) +[2026-06-10 17:54:00,122][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 5.492119043992716e-07, 'train/grad_norm_pre_clip': 2.411814421066083e-05, 'timing/data': 0.07686056569218636, 'timing/model': 1.0628560595214367, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 3.8}) +[2026-06-10 17:54:22,261][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.0009271628223359585, 'train/grad_norm_pre_clip': 0.13932201266288757, 'timing/data': 0.020849745720624924, 'timing/model': 1.4043608009815216, 'learning_rate/qwen_vl_interface': 7.683172208660809e-07, 'learning_rate/action_model': 7.68317220866081e-06, 'epoch': 3.81}) +[2026-06-10 17:54:40,520][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.0006284141563810408, 'train/grad_norm_pre_clip': 0.11093594878911972, 'timing/data': 0.024999283254146576, 'timing/model': 1.1725089401006699, 'learning_rate/qwen_vl_interface': 7.560945455691598e-07, 'learning_rate/action_model': 7.560945455691598e-06, 'epoch': 3.82}) +[2026-06-10 17:55:00,104][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 1.1175867342672063e-07, 'train/grad_norm_pre_clip': 3.0583889838453615e-06, 'timing/data': 0.025031857192516327, 'timing/model': 1.1526929438114166, 'learning_rate/qwen_vl_interface': 7.441490755288531e-07, 'learning_rate/action_model': 7.441490755288531e-06, 'epoch': 3.83}) +[2026-06-10 17:55:20,005][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 5.978527497063624e-06, 'train/grad_norm_pre_clip': 0.0003989487304352224, 'timing/data': 0.02187453955411911, 'timing/model': 1.3303864151239395, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 3.83}) +[2026-06-10 17:55:39,244][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 5.587935092421503e-08, 'train/grad_norm_pre_clip': 8.399904345424147e-07, 'timing/data': 0.027796853333711624, 'timing/model': 1.470669623464346, 'learning_rate/qwen_vl_interface': 7.210926814847443e-07, 'learning_rate/action_model': 7.210926814847444e-06, 'epoch': 3.84}) +[2026-06-10 17:55:58,815][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 8.568166975919667e-08, 'train/grad_norm_pre_clip': 1.9742462882277323e-06, 'timing/data': 0.020197447389364243, 'timing/model': 1.162193138152361, 'learning_rate/qwen_vl_interface': 7.09983179707932e-07, 'learning_rate/action_model': 7.099831797079319e-06, 'epoch': 3.85}) +[2026-06-10 17:56:21,581][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 9.196622158924583e-06, 'train/grad_norm_pre_clip': 0.00042051621130667627, 'timing/data': 0.031892068684101105, 'timing/model': 1.1615029983222485, 'learning_rate/qwen_vl_interface': 6.991537275539842e-07, 'learning_rate/action_model': 6.991537275539842e-06, 'epoch': 3.86}) +[2026-06-10 17:56:31,195][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.0557914599776268, 'train/grad_norm_pre_clip': 1.5890161991119385, 'timing/data': 0.09103550016880035, 'timing/model': 1.0661438927054405, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 3.87}) +[2026-06-10 17:56:52,104][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.00012900953879579902, 'train/grad_norm_pre_clip': 0.026621444150805473, 'timing/data': 0.012474879622459412, 'timing/model': 1.1421232111752033, 'learning_rate/qwen_vl_interface': 6.78337626845176e-07, 'learning_rate/action_model': 6.783376268451759e-06, 'epoch': 3.88}) +[2026-06-10 17:57:14,813][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.0004263594455551356, 'train/grad_norm_pre_clip': 0.048004426062107086, 'timing/data': 0.01704804226756096, 'timing/model': 1.1789359226822853, 'learning_rate/qwen_vl_interface': 6.683522623254591e-07, 'learning_rate/action_model': 6.683522623254591e-06, 'epoch': 3.88}) +[2026-06-10 17:57:33,258][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.0006351114716380835, 'train/grad_norm_pre_clip': 0.041026853024959564, 'timing/data': 0.017703745514154434, 'timing/model': 1.4245301522314548, 'learning_rate/qwen_vl_interface': 6.586495154196905e-07, 'learning_rate/action_model': 6.586495154196905e-06, 'epoch': 3.89}) +[2026-06-10 17:57:53,630][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 2.707747626118362e-05, 'train/grad_norm_pre_clip': 0.001740557374432683, 'timing/data': 0.026007574051618576, 'timing/model': 1.1502750143408775, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 3.9}) +[2026-06-10 17:58:17,710][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.0007452241843566298, 'train/grad_norm_pre_clip': 0.07152925431728363, 'timing/data': 0.02485586330294609, 'timing/model': 1.164316102862358, 'learning_rate/qwen_vl_interface': 6.400942510244234e-07, 'learning_rate/action_model': 6.400942510244233e-06, 'epoch': 3.91}) +[2026-06-10 17:58:36,757][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 1.4901149825163884e-07, 'train/grad_norm_pre_clip': 6.368314643623307e-06, 'timing/data': 0.019113145768642426, 'timing/model': 1.305473618209362, 'learning_rate/qwen_vl_interface': 6.312428781110364e-07, 'learning_rate/action_model': 6.312428781110363e-06, 'epoch': 3.92}) +[2026-06-10 17:58:56,589][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 1.1651539352897089e-05, 'train/grad_norm_pre_clip': 0.0005094863590784371, 'timing/data': 0.013089872896671295, 'timing/model': 1.1613238342106342, 'learning_rate/qwen_vl_interface': 6.226764118932156e-07, 'learning_rate/action_model': 6.226764118932156e-06, 'epoch': 3.92}) +[2026-06-10 17:59:05,930][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.007402582559734583, 'train/grad_norm_pre_clip': 0.81418377161026, 'timing/data': 0.09536490589380264, 'timing/model': 1.2004292495548725, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 3.93}) +[2026-06-10 17:59:26,266][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 4.3472140532685444e-05, 'train/grad_norm_pre_clip': 0.0034657299984246492, 'timing/data': 0.02472873404622078, 'timing/model': 1.536908682435751, 'learning_rate/qwen_vl_interface': 6.064002956172447e-07, 'learning_rate/action_model': 6.064002956172447e-06, 'epoch': 3.94}) +[2026-06-10 17:59:46,332][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.0004873121506534517, 'train/grad_norm_pre_clip': 0.03337785229086876, 'timing/data': 0.02176358550786972, 'timing/model': 1.1633459515869617, 'learning_rate/qwen_vl_interface': 5.986916495466121e-07, 'learning_rate/action_model': 5.986916495466122e-06, 'epoch': 3.95}) +[2026-06-10 18:00:05,088][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 2.607702853651972e-08, 'train/grad_norm_pre_clip': 8.325534963660175e-07, 'timing/data': 0.024686142802238464, 'timing/model': 1.1707251258194447, 'learning_rate/qwen_vl_interface': 5.912699180846554e-07, 'learning_rate/action_model': 5.9126991808465545e-06, 'epoch': 3.96}) +[2026-06-10 18:00:23,227][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 1.4901156930591242e-07, 'train/grad_norm_pre_clip': 3.909887254849309e-06, 'timing/data': 0.01986917480826378, 'timing/model': 1.1672312282025814, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 3.97}) +[2026-06-10 18:00:41,881][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 2.399003960817936e-06, 'train/grad_norm_pre_clip': 0.00016320779104717076, 'timing/data': 0.01656099408864975, 'timing/model': 1.3134815357625484, 'learning_rate/qwen_vl_interface': 5.772890124897043e-07, 'learning_rate/action_model': 5.772890124897042e-06, 'epoch': 3.98}) +[2026-06-10 18:01:00,347][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 4.0978189730367376e-08, 'train/grad_norm_pre_clip': 7.359242886195716e-07, 'timing/data': 0.018497638404369354, 'timing/model': 1.1696764081716537, 'learning_rate/qwen_vl_interface': 5.707307007648241e-07, 'learning_rate/action_model': 5.707307007648241e-06, 'epoch': 3.98}) +[2026-06-10 18:01:21,154][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.00785320345312357, 'train/grad_norm_pre_clip': 0.5998482704162598, 'timing/data': 0.015215273946523666, 'timing/model': 1.1817115135490894, 'learning_rate/qwen_vl_interface': 5.644610284116467e-07, 'learning_rate/action_model': 5.644610284116467e-06, 'epoch': 3.99}) +[2026-06-10 18:01:30,997][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 1.0643684333899728e-07, 'train/grad_norm_pre_clip': 2.7166950076207286e-06, 'timing/data': 0.10701531916856766, 'timing/model': 1.0817428566515446, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 4.0}) +[2026-06-10 18:01:53,473][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.006283378694206476, 'train/grad_norm_pre_clip': 0.2217739224433899, 'timing/data': 0.013573382049798965, 'timing/model': 1.1386439055204391, 'learning_rate/qwen_vl_interface': 5.527891309636035e-07, 'learning_rate/action_model': 5.527891309636035e-06, 'epoch': 4.01}) +[2026-06-10 18:02:16,227][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.21600642800331116, 'train/grad_norm_pre_clip': 2.8197128772735596, 'timing/data': 0.012303262948989868, 'timing/model': 1.853283416479826, 'learning_rate/qwen_vl_interface': 5.473876258463517e-07, 'learning_rate/action_model': 5.473876258463516e-06, 'epoch': 4.02}) +[2026-06-10 18:02:37,307][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.001345918863080442, 'train/grad_norm_pre_clip': 0.09822609275579453, 'timing/data': 0.012528464198112488, 'timing/model': 1.0808710269629955, 'learning_rate/qwen_vl_interface': 5.422762000116201e-07, 'learning_rate/action_model': 5.4227620001162e-06, 'epoch': 4.03}) +[2026-06-10 18:03:00,064][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 4.0978189730367376e-08, 'train/grad_norm_pre_clip': 5.264204219201929e-07, 'timing/data': 0.011332843452692032, 'timing/model': 2.131897322833538, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 4.03}) +[2026-06-10 18:03:26,832][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 6.064034459996037e-05, 'train/grad_norm_pre_clip': 0.003658413654193282, 'timing/data': 0.021314091980457306, 'timing/model': 1.6483977399766445, 'learning_rate/qwen_vl_interface': 5.329248294641001e-07, 'learning_rate/action_model': 5.329248294641001e-06, 'epoch': 4.04}) +[2026-06-10 18:03:46,749][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.00011281453043920919, 'train/grad_norm_pre_clip': 0.010624231770634651, 'timing/data': 0.010670550167560577, 'timing/model': 1.2390402108430862, 'learning_rate/qwen_vl_interface': 5.286854615878969e-07, 'learning_rate/action_model': 5.286854615878969e-06, 'epoch': 4.05}) +[2026-06-10 18:04:06,183][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 3.6018460377817973e-05, 'train/grad_norm_pre_clip': 0.003045752877369523, 'timing/data': 0.011423803865909576, 'timing/model': 1.323587205260992, 'learning_rate/qwen_vl_interface': 5.247373266317992e-07, 'learning_rate/action_model': 5.2473732663179916e-06, 'epoch': 4.06}) +[2026-06-10 18:04:15,286][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 4.483013981371187e-06, 'train/grad_norm_pre_clip': 0.0002189087972510606, 'timing/data': 0.10109417885541916, 'timing/model': 1.0643142983317375, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 4.07}) +[2026-06-10 18:04:34,693][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.005662336014211178, 'train/grad_norm_pre_clip': 0.14184527099132538, 'timing/data': 0.016737710684537888, 'timing/model': 1.1551890708506107, 'learning_rate/qwen_vl_interface': 5.17715711658529e-07, 'learning_rate/action_model': 5.17715711658529e-06, 'epoch': 4.08}) +[2026-06-10 18:04:52,835][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.006117390468716621, 'train/grad_norm_pre_clip': 0.4456150531768799, 'timing/data': 0.019160952419042587, 'timing/model': 1.1714784875512123, 'learning_rate/qwen_vl_interface': 5.146426647676421e-07, 'learning_rate/action_model': 5.146426647676421e-06, 'epoch': 4.08}) +[2026-06-10 18:05:12,777][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 4.5820951299901935e-07, 'train/grad_norm_pre_clip': 2.4591918190708384e-05, 'timing/data': 0.014353293925523758, 'timing/model': 1.2018093056976795, 'learning_rate/qwen_vl_interface': 5.118617170227173e-07, 'learning_rate/action_model': 5.118617170227172e-06, 'epoch': 4.09}) +[2026-06-10 18:05:32,916][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.000101175915915519, 'train/grad_norm_pre_clip': 0.008745229803025723, 'timing/data': 0.021511651575565338, 'timing/model': 1.1651967242360115, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 4.1}) +[2026-06-10 18:05:52,308][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 5.587935092421503e-08, 'train/grad_norm_pre_clip': 1.6253867443083436e-06, 'timing/data': 0.018986012786626816, 'timing/model': 1.2207736745476723, 'learning_rate/qwen_vl_interface': 5.071767871099448e-07, 'learning_rate/action_model': 5.071767871099448e-06, 'epoch': 4.11}) +[2026-06-10 18:06:13,903][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.0041436185128986835, 'train/grad_norm_pre_clip': 0.17892339825630188, 'timing/data': 0.021166350692510605, 'timing/model': 1.3220897167921066, 'learning_rate/qwen_vl_interface': 5.052730939306429e-07, 'learning_rate/action_model': 5.052730939306429e-06, 'epoch': 4.12}) +[2026-06-10 18:06:32,747][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.0016321019502356648, 'train/grad_norm_pre_clip': 0.11855502426624298, 'timing/data': 0.02462618425488472, 'timing/model': 1.1745121218264103, 'learning_rate/qwen_vl_interface': 5.036620778565662e-07, 'learning_rate/action_model': 5.036620778565662e-06, 'epoch': 4.12}) +[2026-06-10 18:06:42,273][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.0008005467825569212, 'train/grad_norm_pre_clip': 0.07913105934858322, 'timing/data': 0.10013905167579651, 'timing/model': 1.1855459176003933, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 4.13}) +[2026-06-10 18:07:04,023][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.025229182094335556, 'train/grad_norm_pre_clip': 1.0797935724258423, 'timing/data': 0.03641042485833168, 'timing/model': 1.1675792783498764, 'learning_rate/qwen_vl_interface': 5.013184564644973e-07, 'learning_rate/action_model': 5.013184564644973e-06, 'epoch': 4.14}) +[2026-06-10 18:07:22,688][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 1.3318933270056732e-05, 'train/grad_norm_pre_clip': 0.001407591043971479, 'timing/data': 0.033376794308423996, 'timing/model': 1.1150420866906643, 'learning_rate/qwen_vl_interface': 5.005859957121123e-07, 'learning_rate/action_model': 5.0058599571211224e-06, 'epoch': 4.15}) +[2026-06-10 18:07:41,774][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.2047266662120819, 'train/grad_norm_pre_clip': 3.344472885131836, 'timing/data': 0.043993983417749405, 'timing/model': 1.3830426558852196, 'learning_rate/qwen_vl_interface': 5.001465011872491e-07, 'learning_rate/action_model': 5.001465011872491e-06, 'epoch': 4.16}) +[2026-06-10 18:08:18,930][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-10 18:08:18,932][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.0001267360639758408, 'train/grad_norm_pre_clip': 0.007557733450084925, 'eval/loss': 0.005868219304829836, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 227.39002990722656, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 178.7413865181368, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 227.39002990722656, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 227.39002990722656, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 178.7413865181368, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 227.39002990722656, 'checkpoint/best_score': 315.4344055175781, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.04322608932852745, 'timing/model': 1.4218530096113682, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 4.17}) +[2026-06-10 18:08:18,947][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_500_state +[2026-06-10 18:08:18,953][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-10 18:08:25,759][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_500_state/model.safetensors +[2026-06-10 18:08:36,360][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_500_state/optimizer.bin +[2026-06-10 18:08:36,361][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_500_state/sampler.bin +[2026-06-10 18:08:36,361][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_500_state/sampler_1.bin +[2026-06-10 18:08:36,364][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-10 18:08:36,379][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-10 18:08:36,379][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-10 18:08:36,397][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-10 18:08:36,398][starVLA.training.train_starvla][INFO] - Training complete. Best checkpoint step=100 score=315.4344055175781 path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state +[2026-06-10 18:08:36,398][accelerate.accelerator][INFO] - Loading states from /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state +[2026-06-10 18:09:01,347][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-10 18:09:50,164][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-10 18:09:50,167][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-10 18:09:50,168][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-10 18:09:50,173][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-10 18:09:50,708][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-10 18:10:48,946][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-10 18:10:48,955][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/summary.jsonl b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..d6b4d703cde264eca50aad49c46affff9660ce5b --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/summary.jsonl @@ -0,0 +1,10 @@ +{"steps": 50} +{"steps": 100} +{"steps": 150} +{"steps": 200} +{"steps": 250} +{"steps": 300} +{"steps": 350} +{"steps": 400} +{"steps": 450} +{"steps": 500} diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/checkpoints/best_model.safetensors b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/checkpoints/best_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4c3a81cd089150f7999853974ceeac0a237bede5 --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/checkpoints/best_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5de650dbc0c50201489cb391a18f95b76bc1b9fb5a3df03ee965050b90284e84 +size 9138230516 diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/checkpoints/steps_2500_model.safetensors b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/checkpoints/steps_2500_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c1792240e00aadbfeeb8b75bcdc53e000059ac71 --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/checkpoints/steps_2500_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dfdb46c577b89bc4dcdef272e2b9e1d590ac0331c23632284bae3980d419bd5e +size 9138230516 diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/config.full.yaml b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..d9a92da7d88c3aa6e09962a8c0e6a96d16495597 --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/config.full.yaml @@ -0,0 +1,184 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 7 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/deadly_corridor_fix_latency_0 + data_mix: deadly_corridor_train__bridge + eval_data_mix: deadly_corridor_train__bridge__val + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + converted_name: deadly_corridor_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: deadly_corridor_train + mixed_converted_name: deadly_corridor_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 2500 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: true + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state + resume_step: 800 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: You are playing Deadly Corridor in VizDoom. Choose actions from + MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, ATTACK. + deadly: + action_layout: multibinary_7 + enabled: true + latency: + prompt_map_path: /workspace/starVLA/data/deadly_corridor_fix_latency_0/deadly_corridor_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: deadly_corridor + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: deadly_corridor +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: true + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_deadly_corridor_fix_latency_0 +output_dir: /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0 +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/config.yaml b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..25b7fa2491a26f094ca616eb2f76fffc04df7f6b --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/config.yaml @@ -0,0 +1,87 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: true + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: deadly_corridor_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: deadly_corridor_train__bridge__val + include_state: true + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 7 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0 +rl_games: + env_eval: + deadly: + action_layout: multibinary_7 + enabled: true + fixed_episode_seeds: true + frameskip: 4 + image_size: 224 + latency: + prompt_map_path: /workspace/starVLA/data/deadly_corridor_fix_latency_0/deadly_corridor_train__bridge/latency_prompt_map.json + latency_seed_stride: 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 0 + max_steps_per_episode: 3600 + num_episodes: 5 + seed: 42 + task_description: You are playing Deadly Corridor in VizDoom. Choose actions from + MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, ATTACK. + task_seed_stride: 0 + model_alias: openvla + task: deadly_corridor +run_id: openvla_deadly_corridor_fix_latency_0 +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + eval_interval: 100 + eval_num_batches: 20 + freeze_modules: '' + gradient_clipping: 1.0 + is_resume: true + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 10 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 2500 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state + resume_step: 800 + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/dataset_statistics.json b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..c7707a242f1e0b25d119fc060f26eacda46c2fe0 --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.9220542311668396, + 0.006256517022848129, + 0.16710114479064941, + 0.6543274521827698, + 0.23253388702869415, + 0.11183524876832962, + 0.5578727722167969 + ], + "std": [ + 0.26808345317840576, + 0.07885219156742096, + 0.3730665147304535, + 0.47559070587158203, + 0.42244860529899597, + 0.31516844034194946, + 0.4966414272785187 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 0.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 3836, + "num_trajectories": 90 + } +} \ No newline at end of file diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/dataset_statistics_eval.json b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..ebeb0582f527163e466d15b8ad6f426c69c9a5ab --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.9322798848152161, + 0.0, + 0.17607223987579346, + 0.6410835385322571, + 0.2370203137397766, + 0.11512415111064911, + 0.5214446783065796 + ], + "std": [ + 0.25126445293426514, + 0.0, + 0.38088199496269226, + 0.4796830415725708, + 0.4252544641494751, + 0.31917139887809753, + 0.49953892827033997 + ], + "max": [ + 1.0, + 0.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 0.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 443, + "num_trajectories": 10 + } +} \ No newline at end of file diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/hydra/.hydra/config.yaml b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..e967685fe5e592ea25b679c25efd108bc842215e --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/hydra/.hydra/config.yaml @@ -0,0 +1,182 @@ +framework: + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 7 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/deadly_corridor_fix_latency_0 + data_mix: deadly_corridor_train__bridge + eval_data_mix: deadly_corridor_train__bridge__val + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: deadly_corridor_train + mixed_converted_name: deadly_corridor_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 2500 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: true + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state + resume_step: 800 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: You are playing Deadly Corridor in VizDoom. Choose actions from + MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, ATTACK. + deadly: + action_layout: multibinary_7 + enabled: true + latency: + prompt_map_path: /workspace/starVLA/data/deadly_corridor_fix_latency_0/deadly_corridor_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: deadly_corridor + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: deadly_corridor +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: true + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_deadly_corridor_fix_latency_0 +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/hydra/.hydra/hydra.yaml b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..3690e7ccc002d4d0dbe6cec5d4bae817df99a09e --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/hydra/.hydra/hydra.yaml @@ -0,0 +1,250 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=deadly_corridor + - init=bridge + - mode=single + - run_id=openvla_deadly_corridor_fix_latency_0 + - run_root_dir=/workspace/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=true + - checkpoint.save_pt_file=false + - trainer.is_resume=true + - trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state + - trainer.resume_step=800 + - trainer.max_train_steps=2500 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=20 + - trainer.logging_frequency=10 + - trainer.gradient_accumulation_steps=16 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9, 0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=7 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/starVLA/data/deadly_corridor_fix_latency_0 + - datasets.vla_data.data_mix=deadly_corridor_train__bridge + - datasets.vla_data.eval_data_mix=deadly_corridor_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=deadly_corridor + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=4 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - rl_games.env_eval.task_description='You are playing Deadly Corridor in VizDoom. + Choose actions from MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, + TURN_RIGHT, ATTACK.' + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/deadly_corridor_fix_latency_0/deadly_corridor_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=true + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[0] + - rl_games.env_eval.post_train.enabled=true + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + - rl_games.env_eval.deadly.action_layout=multibinary_7 + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=true,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=deadly_corridor_train__bridge,datasets.vla_data.data_root_dir=/workspace/starVLA/data/deadly_corridor_fix_latency_0,datasets.vla_data.eval_data_mix=deadly_corridor_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.video_backend=torchvision_av,env=deadly_corridor,framework.action_model.action_dim=7,framework.action_model.action_env_dim=7,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.deadly.action_layout=multibinary_7,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=4,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/deadly_corridor_fix_latency_0/deadly_corridor_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=true,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[0],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=true,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_description='You + are playing Deadly Corridor in VizDoom. Choose actions from MOVE_FORWARD, MOVE_BACKWARD, + MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, ATTACK.',rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=deadly_corridor,run_id=openvla_deadly_corridor_fix_latency_0,run_root_dir=/workspace/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=20,trainer.gradient_accumulation_steps=16,trainer.gradient_clipping=1.0,trainer.is_resume=true,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=10,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=2500,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9, + 0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state,trainer.resume_step=800,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.2 + version_base: '1.1' + cwd: /workspace/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/hydra + choices: + checkpoint: default + mode: single + init: bridge + env: deadly_corridor + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/hydra/.hydra/overrides.yaml b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..580477e173b620363516be30ca75349d87ffebfa --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/hydra/.hydra/overrides.yaml @@ -0,0 +1,88 @@ +- model=openvla +- env=deadly_corridor +- init=bridge +- mode=single +- run_id=openvla_deadly_corridor_fix_latency_0 +- run_root_dir=/workspace/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=true +- checkpoint.save_pt_file=false +- trainer.is_resume=true +- trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state +- trainer.resume_step=800 +- trainer.max_train_steps=2500 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=20 +- trainer.logging_frequency=10 +- trainer.gradient_accumulation_steps=16 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9, 0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=7 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/starVLA/data/deadly_corridor_fix_latency_0 +- datasets.vla_data.data_mix=deadly_corridor_train__bridge +- datasets.vla_data.eval_data_mix=deadly_corridor_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=deadly_corridor +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=4 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- rl_games.env_eval.task_description='You are playing Deadly Corridor in VizDoom. + Choose actions from MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, + TURN_RIGHT, ATTACK.' +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/deadly_corridor_fix_latency_0/deadly_corridor_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=true +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[0] +- rl_games.env_eval.post_train.enabled=true +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] +- rl_games.env_eval.deadly.action_layout=multibinary_7 diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/hydra/train_starvla_hydra.log b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..892cac512f70048ad6e9544f3bd6403aba44a0e5 --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/hydra/train_starvla_hydra.log @@ -0,0 +1,876 @@ +[2026-06-04 04:24:46,091][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-04 04:24:46,146][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-04 04:24:47,375][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `deadly_corridor_train__bridge` +[2026-06-04 04:24:48,064][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-04 04:24:48,065][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-04 04:24:48,065][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-04 04:24:48,065][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-04 04:24:48,065][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-04 04:24:48,066][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-04 04:24:48,066][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-04 04:24:48,066][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-04 04:24:48,066][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-04 04:24:48,066][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-04 04:24:48,067][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-04 04:24:48,067][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-04 04:24:48,067][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-04 04:24:48,067][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-04 04:24:48,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-04 04:24:48,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-04 04:24:48,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-04 04:24:48,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-04 04:24:48,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-04 04:24:48,069][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-04 04:24:48,069][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-04 04:24:48,069][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-04 04:24:48,070][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-04 04:24:48,070][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-04 04:24:48,070][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-04 04:24:48,070][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-04 04:24:48,070][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-04 04:24:48,070][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-04 04:24:48,071][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-04 04:24:48,071][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-04 04:24:48,071][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-04 04:24:48,092][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `deadly_corridor_train__bridge__val` +[2026-06-04 04:24:48,133][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/config.full.yaml +[2026-06-04 04:24:48,139][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/config.yaml +[2026-06-04 04:24:57,323][starVLA.training.train_starvla][INFO] - Loaded pretrained checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt, steps: 0 +[2026-06-04 04:25:52,872][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-04 04:25:52,928][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-04 04:25:54,168][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `deadly_corridor_train__bridge` +[2026-06-04 04:25:55,024][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-04 04:25:55,025][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-04 04:25:55,026][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-04 04:25:55,026][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-04 04:25:55,026][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-04 04:25:55,026][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-04 04:25:55,026][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-04 04:25:55,027][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-04 04:25:55,027][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-04 04:25:55,027][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-04 04:25:55,027][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-04 04:25:55,028][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-04 04:25:55,028][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-04 04:25:55,028][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-04 04:25:55,028][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-04 04:25:55,028][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-04 04:25:55,029][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-04 04:25:55,029][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-04 04:25:55,029][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-04 04:25:55,029][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-04 04:25:55,029][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-04 04:25:55,029][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-04 04:25:55,030][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-04 04:25:55,030][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-04 04:25:55,030][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-04 04:25:55,030][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-04 04:25:55,030][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-04 04:25:55,030][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-04 04:25:55,031][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-04 04:25:55,031][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-04 04:25:55,031][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-04 04:25:55,050][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `deadly_corridor_train__bridge__val` +[2026-06-04 04:25:55,075][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/config.full.yaml +[2026-06-04 04:25:55,081][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/config.yaml +[2026-06-04 04:26:03,030][starVLA.training.train_starvla][INFO] - Loaded pretrained checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt, steps: 0 +[2026-06-04 04:26:05,870][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-04 04:26:05,871][starVLA.training.train_starvla][INFO] - Total optimization steps = 2500 +[2026-06-04 04:26:05,871][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-04 04:26:05,871][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-04 04:26:05,871][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-04 04:31:46,208][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 1.470118761062622, 'train/grad_norm_pre_clip': 2.8246614933013916, 'timing/data': 0.010139601305127144, 'timing/model': 2.2675515599548817, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.08}) +[2026-06-04 04:37:24,124][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.466548353433609, 'train/grad_norm_pre_clip': 0.28005319833755493, 'timing/data': 0.008437380194664001, 'timing/model': 2.280996546149254, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.17}) +[2026-06-04 04:43:03,805][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.2409718930721283, 'train/grad_norm_pre_clip': 0.16984601318836212, 'timing/data': 0.02317143604159355, 'timing/model': 2.349245071411133, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.25}) +[2026-06-04 04:48:23,404][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.16111963987350464, 'train/grad_norm_pre_clip': 0.5676928162574768, 'timing/data': 0.1942834798246622, 'timing/model': 2.0780162569135427, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.33}) +[2026-06-04 04:54:04,614][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.2721882462501526, 'train/grad_norm_pre_clip': 0.4065699279308319, 'timing/data': 0.02333400584757328, 'timing/model': 2.357073465362191, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.42}) +[2026-06-04 04:59:43,617][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.010264961048960686, 'train/grad_norm_pre_clip': 0.10881894081830978, 'timing/data': 0.023345831781625748, 'timing/model': 2.3554860278964043, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.5}) +[2026-06-04 05:05:22,824][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.006389788817614317, 'train/grad_norm_pre_clip': 1.575670599937439, 'timing/data': 0.007572237402200699, 'timing/model': 2.280863406136632, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.58}) +[2026-06-04 05:10:44,716][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.18332073092460632, 'train/grad_norm_pre_clip': 3.828598737716675, 'timing/data': 0.20300383493304253, 'timing/model': 2.1671808026731014, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.67}) +[2026-06-04 05:16:24,528][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.0019873324781656265, 'train/grad_norm_pre_clip': 0.11902634799480438, 'timing/data': 0.018876677379012108, 'timing/model': 2.2803610395640135, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.75}) +[2026-06-04 05:22:17,616][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmplbe4ui8i/test.c -o /tmp/tmplbe4ui8i/test.o +[2026-06-04 05:22:17,636][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmplbe4ui8i/test.o -laio -o /tmp/tmplbe4ui8i/a.out +[2026-06-04 05:22:18,276][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmppux9fo15/test.c -o /tmp/tmppux9fo15/test.o +[2026-06-04 05:22:18,305][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmppux9fo15/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmppux9fo15/a.out +[2026-06-04 05:22:25,652][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state +[2026-06-04 05:22:25,657][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-04 05:22:31,925][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/model.safetensors +[2026-06-04 05:22:38,935][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/optimizer.bin +[2026-06-04 05:22:38,935][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/sampler.bin +[2026-06-04 05:22:38,935][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/sampler_1.bin +[2026-06-04 05:22:38,937][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/random_states_0.pkl +[2026-06-04 05:22:38,950][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-04 05:22:38,951][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 100 +[2026-06-04 05:22:38,952][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 1.0702174222387839e-05, 'train/grad_norm_pre_clip': 0.0007075337925925851, 'eval/loss': 3.0189379685907625e-05, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 171.0666015625, 'rl_games_eval/mid_train/mean_length': 16.6, 'rl_games_eval/mid_train/std_reward': 140.4460894240175, 'rl_games_eval/mid_train/std_length': 5.083306010855534, 'rl_games_eval/mid_train/macro_mean_reward': 171.0666015625, 'rl_games_eval/mid_train/macro_mean_length': 16.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 171.0666015625, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 16.6, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 140.4460894240175, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.083306010855534, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 171.0666015625, 'checkpoint/best_score': 171.0666015625, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.019524255767464638, 'timing/model': 2.356839319691062, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.83}) +[2026-06-04 05:29:19,494][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.0014081618282943964, 'train/grad_norm_pre_clip': 0.029171274974942207, 'timing/data': 0.010802904143929482, 'timing/model': 2.6680382546037436, 'learning_rate/qwen_vl_interface': 9.999593055976534e-06, 'learning_rate/action_model': 9.999593055976533e-05, 'epoch': 0.92}) +[2026-06-04 05:35:38,814][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.009033532813191414, 'train/grad_norm_pre_clip': 0.41494426131248474, 'timing/data': 0.25964186154305935, 'timing/model': 2.445100227370858, 'learning_rate/qwen_vl_interface': 9.998372293633897e-06, 'learning_rate/action_model': 9.998372293633897e-05, 'epoch': 1.0}) +[2026-06-04 05:42:23,298][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.00015190092381089926, 'train/grad_norm_pre_clip': 0.007222404703497887, 'timing/data': 0.015357164666056633, 'timing/model': 2.671007441356778, 'learning_rate/qwen_vl_interface': 9.996337922143435e-06, 'learning_rate/action_model': 9.996337922143436e-05, 'epoch': 1.08}) +[2026-06-04 05:49:04,192][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 5.1855276979040354e-06, 'train/grad_norm_pre_clip': 0.0002993333910126239, 'timing/data': 0.023388691246509552, 'timing/model': 2.6621943805366755, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 1.17}) +[2026-06-04 05:55:44,693][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.00030842385604046285, 'train/grad_norm_pre_clip': 0.011839489452540874, 'timing/data': 0.014751261100172997, 'timing/model': 2.6613493748009205, 'learning_rate/qwen_vl_interface': 9.989829885383368e-06, 'learning_rate/action_model': 9.989829885383367e-05, 'epoch': 1.25}) +[2026-06-04 06:02:04,333][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.2774982452392578, 'train/grad_norm_pre_clip': 0.6334103941917419, 'timing/data': 0.1809549480676651, 'timing/model': 2.442124530673027, 'learning_rate/qwen_vl_interface': 9.985357335232359e-06, 'learning_rate/action_model': 9.98535733523236e-05, 'epoch': 1.33}) +[2026-06-04 06:08:47,293][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.012461300007998943, 'train/grad_norm_pre_clip': 0.1137564480304718, 'timing/data': 0.01607617363333702, 'timing/model': 2.664319945499301, 'learning_rate/qwen_vl_interface': 9.980073405979643e-06, 'learning_rate/action_model': 9.980073405979643e-05, 'epoch': 1.42}) +[2026-06-04 06:15:26,994][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.002286433009430766, 'train/grad_norm_pre_clip': 0.353244811296463, 'timing/data': 0.013163944706320763, 'timing/model': 2.670942148193717, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 1.5}) +[2026-06-04 06:22:07,195][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.02780226059257984, 'train/grad_norm_pre_clip': 0.20972564816474915, 'timing/data': 0.02103014476597309, 'timing/model': 2.6646196339279413, 'learning_rate/qwen_vl_interface': 9.967075170535901e-06, 'learning_rate/action_model': 9.967075170535901e-05, 'epoch': 1.58}) +[2026-06-04 06:28:46,710][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 200 +[2026-06-04 06:28:46,712][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 4.293852180126123e-05, 'train/grad_norm_pre_clip': 0.001563493744470179, 'eval/loss': 0.0001794179988792166, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': -42.611630249023435, 'rl_games_eval/mid_train/mean_length': 16.6, 'rl_games_eval/mid_train/std_reward': 18.162546706327266, 'rl_games_eval/mid_train/std_length': 3.4409301068170506, 'rl_games_eval/mid_train/macro_mean_reward': -42.611630249023435, 'rl_games_eval/mid_train/macro_mean_length': 16.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': -42.611630249023435, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 16.6, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 18.162546706327266, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 3.4409301068170506, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': -42.611630249023435, 'checkpoint/best_score': 171.0666015625, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.19258981198072433, 'timing/model': 2.421451112255454, 'learning_rate/qwen_vl_interface': 9.959363091525602e-06, 'learning_rate/action_model': 9.959363091525601e-05, 'epoch': 1.67}) +[2026-06-04 06:35:30,594][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.1875922977924347, 'train/grad_norm_pre_clip': 2.467883586883545, 'timing/data': 0.0205302182585001, 'timing/model': 2.662549516186118, 'learning_rate/qwen_vl_interface': 9.950844087393436e-06, 'learning_rate/action_model': 9.950844087393435e-05, 'epoch': 1.75}) +[2026-06-04 06:42:11,391][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.00012159706238890067, 'train/grad_norm_pre_clip': 0.001078433939255774, 'timing/data': 0.019940944388508797, 'timing/model': 2.6618880685418844, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 1.83}) +[2026-06-04 06:48:52,097][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.0030410541221499443, 'train/grad_norm_pre_clip': 0.06349002569913864, 'timing/data': 0.0055888742208480835, 'timing/model': 2.6752250529825687, 'learning_rate/qwen_vl_interface': 9.93139128052587e-06, 'learning_rate/action_model': 9.93139128052587e-05, 'epoch': 1.92}) +[2026-06-04 06:55:11,434][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.0003466021444182843, 'train/grad_norm_pre_clip': 0.004204684402793646, 'timing/data': 0.17771144025027752, 'timing/model': 2.4421366713941097, 'learning_rate/qwen_vl_interface': 9.920460810928786e-06, 'learning_rate/action_model': 9.920460810928785e-05, 'epoch': 2.0}) +[2026-06-04 07:01:54,898][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.0011466492433100939, 'train/grad_norm_pre_clip': 0.05806220322847366, 'timing/data': 0.020982280373573303, 'timing/model': 2.6670252215117216, 'learning_rate/qwen_vl_interface': 9.908730081915346e-06, 'learning_rate/action_model': 9.908730081915345e-05, 'epoch': 2.08}) +[2026-06-04 07:08:35,102][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.011079763062298298, 'train/grad_norm_pre_clip': 0.20877929031848907, 'timing/data': 0.005820063874125481, 'timing/model': 2.684419011697173, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 2.17}) +[2026-06-04 07:15:14,693][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.0005956552340649068, 'train/grad_norm_pre_clip': 0.024656398221850395, 'timing/data': 0.005792394280433655, 'timing/model': 2.668987860903144, 'learning_rate/qwen_vl_interface': 9.882876022415435e-06, 'learning_rate/action_model': 9.882876022415435e-05, 'epoch': 2.25}) +[2026-06-04 07:21:34,828][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.061306241899728775, 'train/grad_norm_pre_clip': 0.7856520414352417, 'timing/data': 0.19957235269248486, 'timing/model': 2.411680493503809, 'learning_rate/qwen_vl_interface': 9.868757121888965e-06, 'learning_rate/action_model': 9.868757121888964e-05, 'epoch': 2.33}) +[2026-06-04 07:28:19,295][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 7.083038508426398e-05, 'train/grad_norm_pre_clip': 0.0016654526116326451, 'timing/data': 0.010233404114842415, 'timing/model': 2.6732311248779297, 'learning_rate/qwen_vl_interface': 9.853846821107084e-06, 'learning_rate/action_model': 9.853846821107085e-05, 'epoch': 2.42}) +[2026-06-04 07:35:25,909][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 300 +[2026-06-04 07:35:25,911][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.025747496634721756, 'train/grad_norm_pre_clip': 0.7245857119560242, 'eval/loss': 0.010760135017335415, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': -115.96530151367188, 'rl_games_eval/mid_train/mean_length': 31.8, 'rl_games_eval/mid_train/std_reward': 0.025034246606441584, 'rl_games_eval/mid_train/std_length': 8.930845424706442, 'rl_games_eval/mid_train/macro_mean_reward': -115.96530151367188, 'rl_games_eval/mid_train/macro_mean_length': 31.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': -115.96530151367188, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 31.8, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 0.025034246606441584, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 8.930845424706442, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': -115.96530151367188, 'checkpoint/best_score': 171.0666015625, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.005899641662836075, 'timing/model': 2.67095685377717, 'learning_rate/qwen_vl_interface': 9.838147674873076e-06, 'learning_rate/action_model': 9.838147674873075e-05, 'epoch': 2.5}) +[2026-06-04 07:42:06,503][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.000787450815550983, 'train/grad_norm_pre_clip': 0.020982541143894196, 'timing/data': 0.014566812664270401, 'timing/model': 2.6713859234005213, 'learning_rate/qwen_vl_interface': 9.821662373154824e-06, 'learning_rate/action_model': 9.821662373154825e-05, 'epoch': 2.58}) +[2026-06-04 07:48:25,728][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.1909818947315216, 'train/grad_norm_pre_clip': 1.3312392234802246, 'timing/data': 0.18163416720926762, 'timing/model': 2.429394328966737, 'learning_rate/qwen_vl_interface': 9.804393740623918e-06, 'learning_rate/action_model': 9.804393740623918e-05, 'epoch': 2.67}) +[2026-06-04 07:55:09,214][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.029395822435617447, 'train/grad_norm_pre_clip': 0.3821173906326294, 'timing/data': 0.02042257785797119, 'timing/model': 2.666896164417267, 'learning_rate/qwen_vl_interface': 9.786344736171637e-06, 'learning_rate/action_model': 9.786344736171638e-05, 'epoch': 2.75}) +[2026-06-04 08:01:50,400][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.01058217789977789, 'train/grad_norm_pre_clip': 0.07372458279132843, 'timing/data': 0.01903390698134899, 'timing/model': 2.6667976155877113, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 2.83}) +[2026-06-04 08:08:30,689][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.02626996859908104, 'train/grad_norm_pre_clip': 0.25661948323249817, 'timing/data': 0.0053077805787324905, 'timing/model': 2.736095966771245, 'learning_rate/qwen_vl_interface': 9.747918115101753e-06, 'learning_rate/action_model': 9.747918115101753e-05, 'epoch': 2.92}) +[2026-06-04 08:14:49,651][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 9.403861440659966e-06, 'train/grad_norm_pre_clip': 0.0008580597350373864, 'timing/data': 0.17648151703178883, 'timing/model': 2.562755549326539, 'learning_rate/qwen_vl_interface': 9.727547082687849e-06, 'learning_rate/action_model': 9.72754708268785e-05, 'epoch': 3.0}) +[2026-06-04 08:21:33,293][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 1.2293456563838845e-07, 'train/grad_norm_pre_clip': 1.1430298627601587e-06, 'timing/data': 0.010336821898818016, 'timing/model': 2.673138538375497, 'learning_rate/qwen_vl_interface': 9.7064088456318e-06, 'learning_rate/action_model': 9.7064088456318e-05, 'epoch': 3.08}) +[2026-06-04 08:28:13,394][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.25530603528022766, 'train/grad_norm_pre_clip': 2.767526149749756, 'timing/data': 0.005748320370912552, 'timing/model': 2.6734623331576586, 'learning_rate/qwen_vl_interface': 9.68450702586171e-06, 'learning_rate/action_model': 9.684507025861709e-05, 'epoch': 3.17}) +[2026-06-04 08:34:54,591][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.032982077449560165, 'train/grad_norm_pre_clip': 1.042036771774292, 'timing/data': 0.010601716116070747, 'timing/model': 2.666643364354968, 'learning_rate/qwen_vl_interface': 9.66184537614164e-06, 'learning_rate/action_model': 9.661845376141641e-05, 'epoch': 3.25}) +[2026-06-04 08:41:35,852][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state +[2026-06-04 08:41:35,859][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-04 08:41:41,921][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/model.safetensors +[2026-06-04 08:41:49,534][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/optimizer.bin +[2026-06-04 08:41:49,534][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/sampler.bin +[2026-06-04 08:41:49,534][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/sampler_1.bin +[2026-06-04 08:41:49,536][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/random_states_0.pkl +[2026-06-04 08:41:49,600][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-04 08:41:49,600][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 400 +[2026-06-04 08:41:49,601][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 5.385674285207642e-06, 'train/grad_norm_pre_clip': 6.801793642807752e-05, 'eval/loss': 0.006617383565753698, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 254.45394592285157, 'rl_games_eval/mid_train/mean_length': 18.8, 'rl_games_eval/mid_train/std_reward': 275.1679638643431, 'rl_games_eval/mid_train/std_length': 9.064215354899728, 'rl_games_eval/mid_train/macro_mean_reward': 254.45394592285157, 'rl_games_eval/mid_train/macro_mean_length': 18.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 254.45394592285157, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 18.8, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 275.1679638643431, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 9.064215354899728, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 254.45394592285157, 'checkpoint/best_score': 254.45394592285157, 'checkpoint/best_step': 400.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.21192439645528793, 'timing/model': 2.48833872936666, 'learning_rate/qwen_vl_interface': 9.638427779428613e-06, 'learning_rate/action_model': 9.638427779428613e-05, 'epoch': 3.33}) +[2026-06-04 08:48:33,698][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.0028756416868418455, 'train/grad_norm_pre_clip': 0.639461100101471, 'timing/data': 0.02068750001490116, 'timing/model': 2.6675739642232656, 'learning_rate/qwen_vl_interface': 9.61425824820727e-06, 'learning_rate/action_model': 9.614258248207268e-05, 'epoch': 3.42}) +[2026-06-04 08:55:14,297][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 1.1660055179163464e-06, 'train/grad_norm_pre_clip': 3.660036236396991e-05, 'timing/data': 0.015358155593276024, 'timing/model': 2.664715552702546, 'learning_rate/qwen_vl_interface': 9.589340923802355e-06, 'learning_rate/action_model': 9.589340923802354e-05, 'epoch': 3.5}) +[2026-06-04 09:01:56,103][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.00020507571753114462, 'train/grad_norm_pre_clip': 0.0035598305985331535, 'timing/data': 0.00721467100083828, 'timing/model': 2.6801627445966005, 'learning_rate/qwen_vl_interface': 9.563680075669138e-06, 'learning_rate/action_model': 9.563680075669137e-05, 'epoch': 3.58}) +[2026-06-04 09:08:16,611][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.00029530160827562213, 'train/grad_norm_pre_clip': 0.0034826872870326042, 'timing/data': 0.2289339993149042, 'timing/model': 2.3658052515238523, 'learning_rate/qwen_vl_interface': 9.53728010066184e-06, 'learning_rate/action_model': 9.53728010066184e-05, 'epoch': 3.67}) +[2026-06-04 09:15:00,499][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.0013894319999963045, 'train/grad_norm_pre_clip': 0.01781510002911091, 'timing/data': 0.006406085565686226, 'timing/model': 2.6815038435161114, 'learning_rate/qwen_vl_interface': 9.51014552228027e-06, 'learning_rate/action_model': 9.510145522280269e-05, 'epoch': 3.75}) +[2026-06-04 09:21:42,398][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 2.8696844310616143e-05, 'train/grad_norm_pre_clip': 0.0003638417401816696, 'timing/data': 0.01945396699011326, 'timing/model': 2.7276168931275606, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 3.83}) +[2026-06-04 09:28:23,394][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 1.8775442640617257e-06, 'train/grad_norm_pre_clip': 1.9507662727846764e-05, 'timing/data': 0.010989336296916008, 'timing/model': 2.6697104535996914, 'learning_rate/qwen_vl_interface': 9.45369127794945e-06, 'learning_rate/action_model': 9.453691277949449e-05, 'epoch': 3.92}) +[2026-06-04 09:34:43,842][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 1.8732864646153757e-06, 'train/grad_norm_pre_clip': 1.9098082702839747e-05, 'timing/data': 0.21969888918101788, 'timing/model': 2.502842480316758, 'learning_rate/qwen_vl_interface': 9.424381285144337e-06, 'learning_rate/action_model': 9.424381285144335e-05, 'epoch': 4.0}) +[2026-06-04 09:41:27,696][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.014809034764766693, 'train/grad_norm_pre_clip': 0.16815698146820068, 'timing/data': 0.011475300416350365, 'timing/model': 2.6693121641874313, 'learning_rate/qwen_vl_interface': 9.394356033595787e-06, 'learning_rate/action_model': 9.394356033595787e-05, 'epoch': 4.08}) +[2026-06-04 09:48:29,438][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-04 09:48:29,440][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.0009752516052685678, 'train/grad_norm_pre_clip': 0.02665964886546135, 'eval/loss': 0.0017163489246740937, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 204.2624084472656, 'rl_games_eval/mid_train/mean_length': 17.2, 'rl_games_eval/mid_train/std_reward': 169.8329754092648, 'rl_games_eval/mid_train/std_length': 5.6000000000000005, 'rl_games_eval/mid_train/macro_mean_reward': 204.2624084472656, 'rl_games_eval/mid_train/macro_mean_length': 17.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 204.2624084472656, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.2, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 169.8329754092648, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.6000000000000005, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 204.2624084472656, 'checkpoint/best_score': 254.45394592285157, 'checkpoint/best_step': 400.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.009443923830986023, 'timing/model': 2.773165512830019, 'learning_rate/qwen_vl_interface': 9.363620667976085e-06, 'learning_rate/action_model': 9.363620667976085e-05, 'epoch': 4.17}) +[2026-06-04 09:55:10,303][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.009296184405684471, 'train/grad_norm_pre_clip': 0.16450195014476776, 'timing/data': 0.010920785367488861, 'timing/model': 2.6850513089448214, 'learning_rate/qwen_vl_interface': 9.332180454631902e-06, 'learning_rate/action_model': 9.332180454631901e-05, 'epoch': 4.25}) +[2026-06-04 10:01:29,928][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 1.5969435480656102e-05, 'train/grad_norm_pre_clip': 0.00021534042025450617, 'timing/data': 0.20175179094076157, 'timing/model': 2.4178241454064846, 'learning_rate/qwen_vl_interface': 9.300040780681938e-06, 'learning_rate/action_model': 9.300040780681938e-05, 'epoch': 4.33}) +[2026-06-04 10:08:13,892][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 6.547446537297219e-05, 'train/grad_norm_pre_clip': 0.0008400305523537099, 'timing/data': 0.0070677343755960464, 'timing/model': 2.675150016322732, 'learning_rate/qwen_vl_interface': 9.267207153093874e-06, 'learning_rate/action_model': 9.267207153093875e-05, 'epoch': 4.42}) +[2026-06-04 10:14:55,604][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 4.570408054860309e-05, 'train/grad_norm_pre_clip': 0.0006259799120016396, 'timing/data': 0.01726009137928486, 'timing/model': 2.6615343429148197, 'learning_rate/qwen_vl_interface': 9.233685197740765e-06, 'learning_rate/action_model': 9.233685197740764e-05, 'epoch': 4.5}) +[2026-06-04 10:21:36,401][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.0004856025625485927, 'train/grad_norm_pre_clip': 0.021458717063069344, 'timing/data': 0.006152292713522911, 'timing/model': 2.6798694524914026, 'learning_rate/qwen_vl_interface': 9.199480658437092e-06, 'learning_rate/action_model': 9.199480658437091e-05, 'epoch': 4.58}) +[2026-06-04 10:27:55,687][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.00011214234109502286, 'train/grad_norm_pre_clip': 0.004335570614784956, 'timing/data': 0.21436473913490772, 'timing/model': 2.365188702940941, 'learning_rate/qwen_vl_interface': 9.164599395954575e-06, 'learning_rate/action_model': 9.164599395954574e-05, 'epoch': 4.67}) +[2026-06-04 10:34:38,870][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 6.008724994899239e-06, 'train/grad_norm_pre_clip': 0.0003183864464517683, 'timing/data': 0.005783263593912125, 'timing/model': 2.8449176866561174, 'learning_rate/qwen_vl_interface': 9.129047387017977e-06, 'learning_rate/action_model': 9.129047387017976e-05, 'epoch': 4.75}) +[2026-06-04 10:41:19,800][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 2.220269607278169e-06, 'train/grad_norm_pre_clip': 2.319325358257629e-05, 'timing/data': 0.01077289879322052, 'timing/model': 2.6784989815205336, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 4.83}) +[2026-06-04 10:48:00,796][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.09586678445339203, 'train/grad_norm_pre_clip': 0.9399732351303101, 'timing/data': 0.011243384331464767, 'timing/model': 2.6711152847856283, 'learning_rate/qwen_vl_interface': 9.055955610282539e-06, 'learning_rate/action_model': 9.055955610282538e-05, 'epoch': 4.92}) +[2026-06-04 10:54:46,440][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 600 +[2026-06-04 10:54:46,442][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.08438188582658768, 'train/grad_norm_pre_clip': 2.359368085861206, 'eval/loss': 0.041130851954221725, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': -115.96530456542969, 'rl_games_eval/mid_train/mean_length': 31.8, 'rl_games_eval/mid_train/std_reward': 0.025035720503970814, 'rl_games_eval/mid_train/std_length': 8.930845424706442, 'rl_games_eval/mid_train/macro_mean_reward': -115.96530456542969, 'rl_games_eval/mid_train/macro_mean_length': 31.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': -115.96530456542969, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 31.8, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 0.025035720503970814, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 8.930845424706442, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': -115.96530456542969, 'checkpoint/best_score': 254.45394592285157, 'checkpoint/best_step': 400.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.21949669905006886, 'timing/model': 2.352942068129778, 'learning_rate/qwen_vl_interface': 9.018428366383368e-06, 'learning_rate/action_model': 9.018428366383368e-05, 'epoch': 5.0}) +[2026-06-04 11:01:30,893][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 5.237713139649713e-06, 'train/grad_norm_pre_clip': 9.254502947442234e-05, 'timing/data': 0.010789655148983002, 'timing/model': 2.6767681427299976, 'learning_rate/qwen_vl_interface': 8.98025542168354e-06, 'learning_rate/action_model': 8.98025542168354e-05, 'epoch': 5.08}) +[2026-06-04 11:08:12,499][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 2.436295289953705e-06, 'train/grad_norm_pre_clip': 0.00020343760843388736, 'timing/data': 0.010577015578746796, 'timing/model': 2.6727009918540716, 'learning_rate/qwen_vl_interface': 8.941443316920613e-06, 'learning_rate/action_model': 8.941443316920612e-05, 'epoch': 5.17}) +[2026-06-04 11:14:54,200][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.13601425290107727, 'train/grad_norm_pre_clip': 7.6158061027526855, 'timing/data': 0.019505884498357773, 'timing/model': 2.665738344192505, 'learning_rate/qwen_vl_interface': 8.901998702348934e-06, 'learning_rate/action_model': 8.901998702348933e-05, 'epoch': 5.25}) +[2026-06-04 11:21:15,115][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 3.600848140195012e-05, 'train/grad_norm_pre_clip': 0.017498649656772614, 'timing/data': 0.21693406254053116, 'timing/model': 2.386233428493142, 'learning_rate/qwen_vl_interface': 8.861928336600148e-06, 'learning_rate/action_model': 8.861928336600148e-05, 'epoch': 5.33}) +[2026-06-04 11:27:59,095][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.0001677116088103503, 'train/grad_norm_pre_clip': 0.005851502530276775, 'timing/data': 0.008200805634260178, 'timing/model': 2.6747072618454695, 'learning_rate/qwen_vl_interface': 8.821239085525143e-06, 'learning_rate/action_model': 8.821239085525143e-05, 'epoch': 5.42}) +[2026-06-04 11:34:39,302][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 6.482003414021165e-07, 'train/grad_norm_pre_clip': 8.094901204458438e-06, 'timing/data': 0.023717552423477173, 'timing/model': 2.6676199957728386, 'learning_rate/qwen_vl_interface': 8.779937921017624e-06, 'learning_rate/action_model': 8.779937921017623e-05, 'epoch': 5.5}) +[2026-06-04 11:41:21,196][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 3.7252897300277255e-07, 'train/grad_norm_pre_clip': 4.761127456731629e-06, 'timing/data': 0.005876157432794571, 'timing/model': 2.692632555961609, 'learning_rate/qwen_vl_interface': 8.738031919819508e-06, 'learning_rate/action_model': 8.738031919819508e-05, 'epoch': 5.58}) +[2026-06-04 11:47:42,010][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 1.7327658952126512e-06, 'train/grad_norm_pre_clip': 7.630960317328572e-05, 'timing/data': 0.2143314965069294, 'timing/model': 2.484527351334691, 'learning_rate/qwen_vl_interface': 8.695528262308367e-06, 'learning_rate/action_model': 8.695528262308366e-05, 'epoch': 5.67}) +[2026-06-04 11:54:25,393][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.0017414098838344216, 'train/grad_norm_pre_clip': 0.014427296817302704, 'timing/data': 0.010258935391902924, 'timing/model': 2.674549760296941, 'learning_rate/qwen_vl_interface': 8.65243423126711e-06, 'learning_rate/action_model': 8.652434231267109e-05, 'epoch': 5.75}) +[2026-06-04 12:01:26,438][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 700 +[2026-06-04 12:01:26,440][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.00018665005336515605, 'train/grad_norm_pre_clip': 0.018807630985975266, 'eval/loss': 0.04838467016816139, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 221.5610565185547, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 180.20351694485012, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 221.5610565185547, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 221.5610565185547, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 180.20351694485012, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 221.5610565185547, 'checkpoint/best_score': 254.45394592285157, 'checkpoint/best_step': 400.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.005806725472211838, 'timing/model': 2.7668041456490755, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 5.83}) +[2026-06-04 12:08:07,191][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.008306908421218395, 'train/grad_norm_pre_clip': 0.21766898036003113, 'timing/data': 0.005942203104496002, 'timing/model': 2.6810773238539696, 'learning_rate/qwen_vl_interface': 8.564504684247982e-06, 'learning_rate/action_model': 8.564504684247982e-05, 'epoch': 5.92}) +[2026-06-04 12:14:27,130][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 3.742141416296363e-06, 'train/grad_norm_pre_clip': 0.00018137074948754162, 'timing/data': 0.21719838492572308, 'timing/model': 2.460401112213731, 'learning_rate/qwen_vl_interface': 8.519684234545333e-06, 'learning_rate/action_model': 8.519684234545332e-05, 'epoch': 6.0}) +[2026-06-04 12:21:11,015][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 1.2700775187113322e-05, 'train/grad_norm_pre_clip': 0.0012986611109226942, 'timing/data': 0.022792400792241096, 'timing/model': 2.67107000015676, 'learning_rate/qwen_vl_interface': 8.474303541281474e-06, 'learning_rate/action_model': 8.474303541281474e-05, 'epoch': 6.08}) +[2026-06-04 12:27:51,980][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 6.653330274275504e-06, 'train/grad_norm_pre_clip': 0.001401005545631051, 'timing/data': 0.010317683219909668, 'timing/model': 2.764369387179613, 'learning_rate/qwen_vl_interface': 8.428370380204577e-06, 'learning_rate/action_model': 8.428370380204577e-05, 'epoch': 6.17}) +[2026-06-04 12:34:34,095][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.08785775303840637, 'train/grad_norm_pre_clip': 0.7793930172920227, 'timing/data': 0.006464788690209389, 'timing/model': 2.6762365233153105, 'learning_rate/qwen_vl_interface': 8.381892621725327e-06, 'learning_rate/action_model': 8.381892621725327e-05, 'epoch': 6.25}) +[2026-06-04 12:40:54,329][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.004272256512194872, 'train/grad_norm_pre_clip': 0.18785497546195984, 'timing/data': 0.20254893228411674, 'timing/model': 2.4173785727471113, 'learning_rate/qwen_vl_interface': 8.334878229568372e-06, 'learning_rate/action_model': 8.334878229568372e-05, 'epoch': 6.33}) +[2026-06-04 12:47:38,592][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.0002614302502479404, 'train/grad_norm_pre_clip': 0.0122617082670331, 'timing/data': 0.01577136479318142, 'timing/model': 2.6685112956911325, 'learning_rate/qwen_vl_interface': 8.287335259407779e-06, 'learning_rate/action_model': 8.287335259407779e-05, 'epoch': 6.42}) +[2026-06-04 12:54:20,197][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 2.2515503587783314e-05, 'train/grad_norm_pre_clip': 0.0007910870481282473, 'timing/data': 0.023805292323231697, 'timing/model': 2.6635334882885218, 'learning_rate/qwen_vl_interface': 8.23927185748673e-06, 'learning_rate/action_model': 8.239271857486729e-05, 'epoch': 6.5}) +[2026-06-04 13:01:00,589][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 1.3224740769146592e-06, 'train/grad_norm_pre_clip': 3.90591194445733e-05, 'timing/data': 0.015132805332541466, 'timing/model': 2.668537564575672, 'learning_rate/qwen_vl_interface': 8.190696259221713e-06, 'learning_rate/action_model': 8.190696259221713e-05, 'epoch': 6.58}) +[2026-06-04 13:07:41,025][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state +[2026-06-04 13:07:41,031][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-04 13:07:46,921][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/model.safetensors +[2026-06-04 13:07:54,095][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/optimizer.bin +[2026-06-04 13:07:54,096][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/sampler.bin +[2026-06-04 13:07:54,096][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/sampler_1.bin +[2026-06-04 13:07:54,097][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state/random_states_0.pkl +[2026-06-04 13:07:54,110][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-04 13:07:54,111][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 800 +[2026-06-04 13:07:54,112][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 7.535683721471287e-07, 'train/grad_norm_pre_clip': 5.3698764531873167e-05, 'eval/loss': 1.7362306607537903e-05, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 317.74525146484376, 'rl_games_eval/mid_train/mean_length': 18.8, 'rl_games_eval/mid_train/std_reward': 283.4391748182168, 'rl_games_eval/mid_train/std_length': 8.423775875461075, 'rl_games_eval/mid_train/macro_mean_reward': 317.74525146484376, 'rl_games_eval/mid_train/macro_mean_length': 18.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 317.74525146484376, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 18.8, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 283.4391748182168, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 8.423775875461075, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 317.74525146484376, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.20336012914776802, 'timing/model': 2.4183584731072187, 'learning_rate/qwen_vl_interface': 8.141616787791425e-06, 'learning_rate/action_model': 8.141616787791424e-05, 'epoch': 6.67}) +[2026-06-04 13:14:38,690][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 2.831220058396866e-07, 'train/grad_norm_pre_clip': 4.030159288959112e-06, 'timing/data': 0.019615700468420982, 'timing/model': 2.6620097048580647, 'learning_rate/qwen_vl_interface': 8.092041852710631e-06, 'learning_rate/action_model': 8.092041852710631e-05, 'epoch': 6.75}) +[2026-06-04 13:21:19,702][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.005991579499095678, 'train/grad_norm_pre_clip': 0.5614346265792847, 'timing/data': 0.006470080465078354, 'timing/model': 2.6818704400211573, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 6.83}) +[2026-06-04 13:28:00,498][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.06119905039668083, 'train/grad_norm_pre_clip': 1.4481228590011597, 'timing/data': 0.015036273747682571, 'timing/model': 2.6702819000929594, 'learning_rate/qwen_vl_interface': 7.991439652676861e-06, 'learning_rate/action_model': 7.991439652676861e-05, 'epoch': 6.92}) +[2026-06-04 13:34:20,117][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 2.835405666701263e-06, 'train/grad_norm_pre_clip': 6.795023364247754e-05, 'timing/data': 0.23168980702757835, 'timing/model': 2.474320076406002, 'learning_rate/qwen_vl_interface': 7.940429625392957e-06, 'learning_rate/action_model': 7.940429625392956e-05, 'epoch': 7.0}) +[2026-06-04 13:41:03,589][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 8.866079497238388e-07, 'train/grad_norm_pre_clip': 0.00015498454740736634, 'timing/data': 0.006076322868466377, 'timing/model': 2.6717780753970146, 'learning_rate/qwen_vl_interface': 7.888958606843111e-06, 'learning_rate/action_model': 7.888958606843111e-05, 'epoch': 7.08}) +[2026-06-04 13:47:44,697][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 4.5075867660671065e-07, 'train/grad_norm_pre_clip': 1.3399663657764904e-05, 'timing/data': 0.015918031334877014, 'timing/model': 2.6655706483870745, 'learning_rate/qwen_vl_interface': 7.83703541632138e-06, 'learning_rate/action_model': 7.83703541632138e-05, 'epoch': 7.17}) +[2026-06-04 13:54:25,000][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 4.023077053716406e-06, 'train/grad_norm_pre_clip': 0.00023467968276236206, 'timing/data': 0.006000760942697525, 'timing/model': 2.6786439940333366, 'learning_rate/qwen_vl_interface': 7.784668950599155e-06, 'learning_rate/action_model': 7.784668950599155e-05, 'epoch': 7.25}) +[2026-06-04 14:00:44,817][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 1.9626686480478384e-06, 'train/grad_norm_pre_clip': 7.441161142196506e-05, 'timing/data': 0.21645298786461353, 'timing/model': 2.3871129769831896, 'learning_rate/qwen_vl_interface': 7.731868182400758e-06, 'learning_rate/action_model': 7.731868182400758e-05, 'epoch': 7.33}) +[2026-06-04 14:07:30,400][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 3.352760558072987e-08, 'train/grad_norm_pre_clip': 1.2460949392334442e-06, 'timing/data': 0.006105558946728706, 'timing/model': 2.6705954056233168, 'learning_rate/qwen_vl_interface': 7.678642158865997e-06, 'learning_rate/action_model': 7.678642158865997e-05, 'epoch': 7.42}) +[2026-06-04 14:14:32,752][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 900 +[2026-06-04 14:14:32,753][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 2.197915307533549e-07, 'train/grad_norm_pre_clip': 8.618699212092906e-06, 'eval/loss': 0.0010508181294426322, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 230.59402160644532, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 170.4862593933571, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 230.59402160644532, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 230.59402160644532, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 170.4862593933571, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 230.59402160644532, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.006548414006829262, 'timing/model': 3.003439338877797, 'learning_rate/qwen_vl_interface': 7.625e-06, 'learning_rate/action_model': 7.625e-05, 'epoch': 7.5}) +[2026-06-04 14:21:13,794][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 4.0978189730367376e-08, 'train/grad_norm_pre_clip': 1.5467936691493378e-06, 'timing/data': 0.020846150815486908, 'timing/model': 2.663148446008563, 'learning_rate/qwen_vl_interface': 7.570950897110537e-06, 'learning_rate/action_model': 7.570950897110536e-05, 'epoch': 7.58}) +[2026-06-04 14:27:34,034][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 2.554484623829012e-08, 'train/grad_norm_pre_clip': 8.25661857106752e-07, 'timing/data': 0.19700195640325546, 'timing/model': 2.4203239623457193, 'learning_rate/qwen_vl_interface': 7.516504111233141e-06, 'learning_rate/action_model': 7.51650411123314e-05, 'epoch': 7.67}) +[2026-06-04 14:34:18,797][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 8.940691031966708e-08, 'train/grad_norm_pre_clip': 3.8004568523319904e-06, 'timing/data': 0.02052144892513752, 'timing/model': 2.6649596467614174, 'learning_rate/qwen_vl_interface': 7.46166897154428e-06, 'learning_rate/action_model': 7.46166897154428e-05, 'epoch': 7.75}) +[2026-06-04 14:40:59,500][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.005815710406750441, 'train/grad_norm_pre_clip': 0.3996378779411316, 'timing/data': 0.010891018435359001, 'timing/model': 2.675727393478155, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 7.83}) +[2026-06-04 14:47:40,598][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 3.6655337680713274e-06, 'train/grad_norm_pre_clip': 0.00011133227235404775, 'timing/data': 0.01542358472943306, 'timing/model': 2.6677851751446724, 'learning_rate/qwen_vl_interface': 7.350871278540257e-06, 'learning_rate/action_model': 7.350871278540256e-05, 'epoch': 7.92}) +[2026-06-04 14:54:01,460][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 5.108969247658024e-08, 'train/grad_norm_pre_clip': 6.389444706655922e-07, 'timing/data': 0.2090315092355013, 'timing/model': 2.5327764712274075, 'learning_rate/qwen_vl_interface': 7.2949277098394036e-06, 'learning_rate/action_model': 7.294927709839403e-05, 'epoch': 8.0}) +[2026-06-04 15:00:45,690][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 2.73426303465385e-06, 'train/grad_norm_pre_clip': 0.00010388848750153556, 'timing/data': 0.02092823199927807, 'timing/model': 2.662602536380291, 'learning_rate/qwen_vl_interface': 7.238633753302784e-06, 'learning_rate/action_model': 7.238633753302784e-05, 'epoch': 8.08}) +[2026-06-04 15:07:26,999][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.004411546979099512, 'train/grad_norm_pre_clip': 0.28865835070610046, 'timing/data': 0.012572811916470528, 'timing/model': 2.672655288130045, 'learning_rate/qwen_vl_interface': 7.181999054610053e-06, 'learning_rate/action_model': 7.181999054610052e-05, 'epoch': 8.17}) +[2026-06-04 15:14:07,998][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 6.0240516177145764e-05, 'train/grad_norm_pre_clip': 0.002424735575914383, 'timing/data': 0.016600077971816063, 'timing/model': 2.6727384347468615, 'learning_rate/qwen_vl_interface': 7.125033317825269e-06, 'learning_rate/action_model': 7.12503331782527e-05, 'epoch': 8.25}) +[2026-06-04 15:20:47,679][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-06-04 15:20:47,680][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 8.47235412493319e-07, 'train/grad_norm_pre_clip': 5.1837861974490806e-05, 'eval/loss': 0.011025047861039639, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 235.68425598144532, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 165.18173802650097, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 235.68425598144532, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 235.68425598144532, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 165.18173802650097, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 235.68425598144532, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.21064727194607258, 'timing/model': 2.394420698285103, 'learning_rate/qwen_vl_interface': 7.067746303734178e-06, 'learning_rate/action_model': 7.067746303734177e-05, 'epoch': 8.33}) +[2026-06-04 15:27:28,991][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 5.219384183874354e-05, 'train/grad_norm_pre_clip': 0.01568855717778206, 'timing/data': 0.015437571331858635, 'timing/model': 2.6703088395297527, 'learning_rate/qwen_vl_interface': 7.0101478281717235e-06, 'learning_rate/action_model': 7.010147828171724e-05, 'epoch': 8.42}) +[2026-06-04 15:34:10,091][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 8.009303655853728e-07, 'train/grad_norm_pre_clip': 4.058556805830449e-05, 'timing/data': 0.025231903418898582, 'timing/model': 2.6596386525779963, 'learning_rate/qwen_vl_interface': 6.952247760340176e-06, 'learning_rate/action_model': 6.952247760340176e-05, 'epoch': 8.5}) +[2026-06-04 15:40:51,298][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 2.9429742198772146e-07, 'train/grad_norm_pre_clip': 2.1473855667863972e-05, 'timing/data': 0.022181618958711624, 'timing/model': 2.6648606155067682, 'learning_rate/qwen_vl_interface': 6.894056021118092e-06, 'learning_rate/action_model': 6.894056021118091e-05, 'epoch': 8.58}) +[2026-06-04 15:47:10,939][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 3.831726758107834e-08, 'train/grad_norm_pre_clip': 7.709489295848471e-07, 'timing/data': 0.2200372163206339, 'timing/model': 2.4039523042738438, 'learning_rate/qwen_vl_interface': 6.835582581360413e-06, 'learning_rate/action_model': 6.835582581360412e-05, 'epoch': 8.67}) +[2026-06-04 15:53:54,098][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 8.568165554834195e-08, 'train/grad_norm_pre_clip': 2.710512717385427e-06, 'timing/data': 0.018036724999547005, 'timing/model': 2.669941872358322, 'learning_rate/qwen_vl_interface': 6.7768374601900174e-06, 'learning_rate/action_model': 6.776837460190018e-05, 'epoch': 8.75}) +[2026-06-04 16:00:35,307][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.04412532597780228, 'train/grad_norm_pre_clip': 2.5447781085968018, 'timing/data': 0.018039695918560028, 'timing/model': 2.662015361711383, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 8.83}) +[2026-06-04 16:07:16,110][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 9.313222903983842e-08, 'train/grad_norm_pre_clip': 2.228831817774335e-06, 'timing/data': 0.006452977657318115, 'timing/model': 2.7927053570747375, 'learning_rate/qwen_vl_interface': 6.658572481133962e-06, 'learning_rate/action_model': 6.658572481133962e-05, 'epoch': 8.92}) +[2026-06-04 16:13:35,215][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.17433489859104156, 'train/grad_norm_pre_clip': 5.202001571655273, 'timing/data': 0.22049842216074467, 'timing/model': 2.3828154541552067, 'learning_rate/qwen_vl_interface': 6.599072887343634e-06, 'learning_rate/action_model': 6.599072887343633e-05, 'epoch': 9.0}) +[2026-06-04 16:20:19,301][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.016410771757364273, 'train/grad_norm_pre_clip': 0.5281752347946167, 'timing/data': 0.006141230463981628, 'timing/model': 2.6652399878948927, 'learning_rate/qwen_vl_interface': 6.539342136859104e-06, 'learning_rate/action_model': 6.539342136859104e-05, 'epoch': 9.08}) +[2026-06-04 16:27:17,654][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1100 +[2026-06-04 16:27:17,656][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 1.490113561430917e-07, 'train/grad_norm_pre_clip': 1.680738387221936e-05, 'eval/loss': 0.002834961749613285, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 238.85040588378905, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 168.6006127503673, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 238.85040588378905, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 238.85040588378905, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 168.6006127503673, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 238.85040588378905, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.005678566172719002, 'timing/model': 2.678759776055813, 'learning_rate/qwen_vl_interface': 6.479390464236974e-06, 'learning_rate/action_model': 6.479390464236974e-05, 'epoch': 9.17}) +[2026-06-04 16:33:58,699][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 7.881730198278092e-06, 'train/grad_norm_pre_clip': 0.0009210282587446272, 'timing/data': 0.01732703484594822, 'timing/model': 2.6714291274547577, 'learning_rate/qwen_vl_interface': 6.4192281418877174e-06, 'learning_rate/action_model': 6.419228141887717e-05, 'epoch': 9.25}) +[2026-06-04 16:40:18,713][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 1.1495177432152559e-07, 'train/grad_norm_pre_clip': 4.357782017905265e-06, 'timing/data': 0.22152126766741276, 'timing/model': 2.4663783721625805, 'learning_rate/qwen_vl_interface': 6.3588654783155525e-06, 'learning_rate/action_model': 6.358865478315553e-05, 'epoch': 9.33}) +[2026-06-04 16:47:02,197][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 1.6139430954353884e-05, 'train/grad_norm_pre_clip': 0.0016644384013488889, 'timing/data': 0.010663595050573349, 'timing/model': 2.675996931269765, 'learning_rate/qwen_vl_interface': 6.298312816352132e-06, 'learning_rate/action_model': 6.298312816352132e-05, 'epoch': 9.42}) +[2026-06-04 16:53:42,691][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 3.1657269573770463e-05, 'train/grad_norm_pre_clip': 0.0018966537900269032, 'timing/data': 0.010703636333346367, 'timing/model': 2.6685559563338757, 'learning_rate/qwen_vl_interface': 6.2375805313843564e-06, 'learning_rate/action_model': 6.237580531384356e-05, 'epoch': 9.5}) +[2026-06-04 17:00:24,197][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.0004886691458523273, 'train/grad_norm_pre_clip': 0.034442633390426636, 'timing/data': 0.01050942949950695, 'timing/model': 2.6773784197866917, 'learning_rate/qwen_vl_interface': 6.176679029576611e-06, 'learning_rate/action_model': 6.17667902957661e-05, 'epoch': 9.58}) +[2026-06-04 17:06:44,814][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 1.6604150232524262e-07, 'train/grad_norm_pre_clip': 3.903117431036662e-06, 'timing/data': 0.21929233148694038, 'timing/model': 2.3834474850445986, 'learning_rate/qwen_vl_interface': 6.115618746087702e-06, 'learning_rate/action_model': 6.115618746087703e-05, 'epoch': 9.67}) +[2026-06-04 17:13:28,699][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 5.889720341656357e-05, 'train/grad_norm_pre_clip': 0.005816558841615915, 'timing/data': 0.01084151677787304, 'timing/model': 2.670596417039633, 'learning_rate/qwen_vl_interface': 6.0544101432828676e-06, 'learning_rate/action_model': 6.054410143282867e-05, 'epoch': 9.75}) +[2026-06-04 17:20:08,898][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 4.9730435421224684e-05, 'train/grad_norm_pre_clip': 0.0023112827911973, 'timing/data': 0.005674645304679871, 'timing/model': 2.672071574255824, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 9.83}) +[2026-06-04 17:26:49,999][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.00029721029568463564, 'train/grad_norm_pre_clip': 0.022611575201153755, 'timing/data': 0.020093780010938644, 'timing/model': 2.7392608392983675, 'learning_rate/qwen_vl_interface': 5.9315899544581015e-06, 'learning_rate/action_model': 5.931589954458101e-05, 'epoch': 9.92}) +[2026-06-04 17:33:30,381][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1200 +[2026-06-04 17:33:30,383][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 5.125740699440939e-06, 'train/grad_norm_pre_clip': 0.00020729470998048782, 'eval/loss': 0.0236639566719532, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 181.0637063264991, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 181.0637063264991, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 222.22349853515624, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.2197836320847273, 'timing/model': 2.389241922646761, 'learning_rate/qwen_vl_interface': 5.869999413045247e-06, 'learning_rate/action_model': 5.869999413045246e-05, 'epoch': 10.0}) +[2026-06-04 17:40:14,407][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 2.7939660185438697e-07, 'train/grad_norm_pre_clip': 5.202283318794798e-06, 'timing/data': 0.01692822575569153, 'timing/model': 2.675158428028226, 'learning_rate/qwen_vl_interface': 5.80830263792473e-06, 'learning_rate/action_model': 5.80830263792473e-05, 'epoch': 10.08}) +[2026-06-04 17:46:55,545][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 1.378357268322361e-07, 'train/grad_norm_pre_clip': 1.775553755578585e-06, 'timing/data': 0.01786069944500923, 'timing/model': 2.9137104600667953, 'learning_rate/qwen_vl_interface': 5.746510200521354e-06, 'learning_rate/action_model': 5.746510200521355e-05, 'epoch': 10.17}) +[2026-06-04 17:53:36,604][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 9.313224325069314e-08, 'train/grad_norm_pre_clip': 1.5341666994572734e-06, 'timing/data': 0.010777849704027176, 'timing/model': 2.6705214492976665, 'learning_rate/qwen_vl_interface': 5.684632688651163e-06, 'learning_rate/action_model': 5.6846326886511624e-05, 'epoch': 10.25}) +[2026-06-04 17:59:57,111][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 1.0856482504095766e-06, 'train/grad_norm_pre_clip': 3.5985798604087904e-05, 'timing/data': 0.20479178801178932, 'timing/model': 2.457413773983717, 'learning_rate/qwen_vl_interface': 5.622680704707265e-06, 'learning_rate/action_model': 5.622680704707265e-05, 'epoch': 10.33}) +[2026-06-04 18:06:42,063][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 1.6391270207805064e-07, 'train/grad_norm_pre_clip': 3.247329686928424e-06, 'timing/data': 0.005940448492765427, 'timing/model': 2.739351350814104, 'learning_rate/qwen_vl_interface': 5.560664863843181e-06, 'learning_rate/action_model': 5.56066486384318e-05, 'epoch': 10.42}) +[2026-06-04 18:13:22,703][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 3.948789242258499e-07, 'train/grad_norm_pre_clip': 4.8833739128895104e-05, 'timing/data': 0.010631928220391273, 'timing/model': 2.676376035436988, 'learning_rate/qwen_vl_interface': 5.498595792153985e-06, 'learning_rate/action_model': 5.498595792153984e-05, 'epoch': 10.5}) +[2026-06-04 18:20:04,095][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 4.0978189730367376e-08, 'train/grad_norm_pre_clip': 1.0461221791047137e-06, 'timing/data': 0.005956508219242096, 'timing/model': 2.6745421905070543, 'learning_rate/qwen_vl_interface': 5.436484124855577e-06, 'learning_rate/action_model': 5.4364841248555764e-05, 'epoch': 10.58}) +[2026-06-04 18:26:23,421][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 3.7378742945293197e-06, 'train/grad_norm_pre_clip': 0.00032127488520927727, 'timing/data': 0.22845817729830742, 'timing/model': 2.481938399374485, 'learning_rate/qwen_vl_interface': 5.374340504462399e-06, 'learning_rate/action_model': 5.3743405044623984e-05, 'epoch': 10.67}) +[2026-06-04 18:33:07,102][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 1.0058281674218961e-07, 'train/grad_norm_pre_clip': 3.840995759674115e-06, 'timing/data': 0.020645542070269585, 'timing/model': 2.6685049310326576, 'learning_rate/qwen_vl_interface': 5.312175578963886e-06, 'learning_rate/action_model': 5.312175578963886e-05, 'epoch': 10.75}) +[2026-06-04 18:40:08,386][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1300 +[2026-06-04 18:40:08,387][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 1.005827670041981e-07, 'train/grad_norm_pre_clip': 4.787431862496305e-06, 'eval/loss': 0.024884827435016632, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 181.0637063264991, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 181.0637063264991, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 222.22349853515624, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.010856030508875847, 'timing/model': 2.6724881287664175, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 10.83}) +[2026-06-04 18:46:50,200][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 5.609817890217528e-06, 'train/grad_norm_pre_clip': 0.0005378033965826035, 'timing/data': 0.01500033400952816, 'timing/model': 2.6708335541188717, 'learning_rate/qwen_vl_interface': 5.187824421036115e-06, 'learning_rate/action_model': 5.187824421036115e-05, 'epoch': 10.92}) +[2026-06-04 18:53:10,707][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 4.4277595634412137e-07, 'train/grad_norm_pre_clip': 1.1218194231332745e-05, 'timing/data': 0.2219367939978838, 'timing/model': 2.3739107213914394, 'learning_rate/qwen_vl_interface': 5.125659495537603e-06, 'learning_rate/action_model': 5.1256594955376034e-05, 'epoch': 11.0}) +[2026-06-04 18:59:54,196][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 1.7373697119182907e-05, 'train/grad_norm_pre_clip': 0.001790389302186668, 'timing/data': 0.011111795902252197, 'timing/model': 2.6714507956057787, 'learning_rate/qwen_vl_interface': 5.063515875144426e-06, 'learning_rate/action_model': 5.0635158751444254e-05, 'epoch': 11.08}) +[2026-06-04 19:06:35,498][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 2.0861591565335402e-07, 'train/grad_norm_pre_clip': 8.047109076869674e-06, 'timing/data': 0.01519644446671009, 'timing/model': 2.670172868296504, 'learning_rate/qwen_vl_interface': 5.001404207846017e-06, 'learning_rate/action_model': 5.001404207846017e-05, 'epoch': 11.17}) +[2026-06-04 19:13:15,697][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.0022852206602692604, 'train/grad_norm_pre_clip': 0.6024700403213501, 'timing/data': 0.005880851298570633, 'timing/model': 2.6825133245438337, 'learning_rate/qwen_vl_interface': 4.93933513615682e-06, 'learning_rate/action_model': 4.9393351361568204e-05, 'epoch': 11.25}) +[2026-06-04 19:19:35,118][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 6.854510274934e-07, 'train/grad_norm_pre_clip': 3.005181861226447e-05, 'timing/data': 0.21663474664092064, 'timing/model': 2.384571062400937, 'learning_rate/qwen_vl_interface': 4.877319295292737e-06, 'learning_rate/action_model': 4.8773192952927374e-05, 'epoch': 11.33}) +[2026-06-04 19:26:19,500][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 1.4901131351052754e-07, 'train/grad_norm_pre_clip': 5.232182957115583e-06, 'timing/data': 0.020947126671671867, 'timing/model': 2.6690433640033007, 'learning_rate/qwen_vl_interface': 4.815367311348839e-06, 'learning_rate/action_model': 4.815367311348839e-05, 'epoch': 11.42}) +[2026-06-04 19:33:00,588][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 1.4901160305669237e-08, 'train/grad_norm_pre_clip': 4.485664248932153e-07, 'timing/data': 0.014791665598750114, 'timing/model': 2.662876123562455, 'learning_rate/qwen_vl_interface': 4.753489799478647e-06, 'learning_rate/action_model': 4.7534897994786464e-05, 'epoch': 11.5}) +[2026-06-04 19:39:40,704][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 1.0803325523056628e-07, 'train/grad_norm_pre_clip': 9.510808013146743e-06, 'timing/data': 0.01460716687142849, 'timing/model': 2.670750442892313, 'learning_rate/qwen_vl_interface': 4.691697362075272e-06, 'learning_rate/action_model': 4.691697362075272e-05, 'epoch': 11.58}) +[2026-06-04 19:46:21,409][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1400 +[2026-06-04 19:46:21,410][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 2.0946163203916512e-06, 'train/grad_norm_pre_clip': 0.00036935554817318916, 'eval/loss': 0.02306532673537731, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 181.0637063264991, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 181.0637063264991, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 222.22349853515624, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.20337429083883762, 'timing/model': 2.372767800465226, 'learning_rate/qwen_vl_interface': 4.630000586954756e-06, 'learning_rate/action_model': 4.630000586954756e-05, 'epoch': 11.67}) +[2026-06-04 19:53:05,498][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 2.2351704842549225e-07, 'train/grad_norm_pre_clip': 7.112032562872628e-06, 'timing/data': 0.015283947810530663, 'timing/model': 2.6736780051141977, 'learning_rate/qwen_vl_interface': 4.568410045541899e-06, 'learning_rate/action_model': 4.5684100455418986e-05, 'epoch': 11.75}) +[2026-06-04 19:59:46,899][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 1.7508814664779493e-07, 'train/grad_norm_pre_clip': 2.5393608666490763e-05, 'timing/data': 0.005603387951850891, 'timing/model': 2.6739648934453726, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 11.83}) +[2026-06-04 20:06:27,602][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 3.725290076417309e-09, 'train/grad_norm_pre_clip': 1.5564175726012763e-07, 'timing/data': 0.015455370768904686, 'timing/model': 2.6682241316884756, 'learning_rate/qwen_vl_interface': 4.4455898567171336e-06, 'learning_rate/action_model': 4.4455898567171334e-05, 'epoch': 11.92}) +[2026-06-04 20:12:47,616][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 4.257474373048353e-09, 'train/grad_norm_pre_clip': 1.5349912985129777e-07, 'timing/data': 0.22301490418612957, 'timing/model': 2.477821869775653, 'learning_rate/qwen_vl_interface': 4.384381253912299e-06, 'learning_rate/action_model': 4.3843812539122994e-05, 'epoch': 12.0}) +[2026-06-04 20:19:31,294][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 1.1399200730011216e-06, 'train/grad_norm_pre_clip': 0.0009079364826902747, 'timing/data': 0.015416033565998077, 'timing/model': 2.666728535667062, 'learning_rate/qwen_vl_interface': 4.323320970423391e-06, 'learning_rate/action_model': 4.3233209704233915e-05, 'epoch': 12.08}) +[2026-06-04 20:26:12,388][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.09167636930942535, 'train/grad_norm_pre_clip': 3.3821401596069336, 'timing/data': 0.015334773808717728, 'timing/model': 2.760885823518038, 'learning_rate/qwen_vl_interface': 4.262419468615644e-06, 'learning_rate/action_model': 4.2624194686156435e-05, 'epoch': 12.17}) +[2026-06-04 20:32:53,696][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 3.901348222257184e-08, 'timing/data': 0.00990765355527401, 'timing/model': 2.6725624315440655, 'learning_rate/qwen_vl_interface': 4.201687183647871e-06, 'learning_rate/action_model': 4.201687183647871e-05, 'epoch': 12.25}) +[2026-06-04 20:39:14,433][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 1.277242311914506e-08, 'train/grad_norm_pre_clip': 2.6560880428405653e-07, 'timing/data': 0.21965385600924492, 'timing/model': 2.4049302116036415, 'learning_rate/qwen_vl_interface': 4.14113452168445e-06, 'learning_rate/action_model': 4.1411345216844504e-05, 'epoch': 12.33}) +[2026-06-04 20:45:57,197][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 4.097818262494002e-08, 'train/grad_norm_pre_clip': 1.4190446790962596e-06, 'timing/data': 0.010886222124099731, 'timing/model': 2.6697369813919067, 'learning_rate/qwen_vl_interface': 4.080771858112283e-06, 'learning_rate/action_model': 4.0807718581122825e-05, 'epoch': 12.42}) +[2026-06-04 20:52:59,051][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-04 20:52:59,053][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 4.10109990411911e-08, 'eval/loss': 0.01290956512093544, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 181.0637063264991, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 181.0637063264991, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 222.22349853515624, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.019668705761432648, 'timing/model': 2.7389239203184843, 'learning_rate/qwen_vl_interface': 4.0206095357630256e-06, 'learning_rate/action_model': 4.020609535763026e-05, 'epoch': 12.5}) +[2026-06-05 00:05:33,353][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-05 00:05:33,408][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-05 00:05:34,702][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `deadly_corridor_train__bridge` +[2026-06-05 00:05:35,501][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-05 00:05:35,501][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-05 00:05:35,502][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-05 00:05:35,502][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-05 00:05:35,502][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-05 00:05:35,502][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-05 00:05:35,502][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-05 00:05:35,503][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-05 00:05:35,503][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-05 00:05:35,503][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-05 00:05:35,504][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-05 00:05:35,504][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-05 00:05:35,504][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-05 00:05:35,504][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-05 00:05:35,504][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-05 00:05:35,505][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-05 00:05:35,505][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-05 00:05:35,505][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-05 00:05:35,505][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-05 00:05:35,506][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-05 00:05:35,506][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-05 00:05:35,506][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-05 00:05:35,507][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-05 00:05:35,507][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-05 00:05:35,507][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-05 00:05:35,507][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-05 00:05:35,507][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-05 00:05:35,507][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-05 00:05:35,508][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-05 00:05:35,508][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-05 00:05:35,508][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-05 00:05:35,528][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `deadly_corridor_train__bridge__val` +[2026-06-05 00:05:35,558][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/config.full.yaml +[2026-06-05 00:05:35,564][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/config.yaml +[2026-06-05 00:05:35,564][starVLA.training.train_starvla][INFO] - Loaded best checkpoint metadata: step=800 score=317.74525146484376 +[2026-06-05 00:05:35,565][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state, steps: 800 +[2026-06-05 00:05:36,964][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state +[2026-06-05 00:05:38,086][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-05 00:05:49,267][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-05 00:05:49,267][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-05 00:05:49,268][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-05 00:05:49,270][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-05 00:05:50,169][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-05 00:05:51,630][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-05 00:05:51,630][starVLA.training.train_starvla][INFO] - Total optimization steps = 2500 +[2026-06-05 00:05:51,631][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-05 00:05:51,631][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-05 00:05:51,631][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-05 00:11:20,689][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 1.5646217832454568e-07, 'train/grad_norm_pre_clip': 2.130357188434573e-06, 'timing/data': 0.0076415035873651505, 'timing/model': 2.178504552692175, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 6.75}) +[2026-06-05 00:16:45,681][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.06900615245103836, 'train/grad_norm_pre_clip': 4.388418674468994, 'timing/data': 0.017674041911959648, 'timing/model': 2.2564416341483593, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 6.83}) +[2026-06-05 00:22:10,780][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 7.215197001642082e-06, 'train/grad_norm_pre_clip': 0.0007567856810055673, 'timing/data': 0.007871139794588089, 'timing/model': 2.1714751422405243, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 6.92}) +[2026-06-05 00:27:18,665][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 3.656775879790075e-05, 'train/grad_norm_pre_clip': 0.0010830577230080962, 'timing/data': 0.3198447283357382, 'timing/model': 2.042464256286621, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 7.0}) +[2026-06-05 00:32:47,280][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.002271506702527404, 'train/grad_norm_pre_clip': 0.06671098619699478, 'timing/data': 0.014540556818246841, 'timing/model': 2.1709238160401583, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 7.08}) +[2026-06-05 00:38:11,082][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 5.848701221111696e-07, 'train/grad_norm_pre_clip': 1.4241323697206099e-05, 'timing/data': 0.020265603438019753, 'timing/model': 2.1615400444716215, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 7.17}) +[2026-06-05 00:43:35,989][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 3.0919903792892e-07, 'train/grad_norm_pre_clip': 5.2252244131523184e-06, 'timing/data': 0.015868807211518288, 'timing/model': 2.1724378541111946, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 7.25}) +[2026-06-05 00:48:43,079][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.002354149241000414, 'train/grad_norm_pre_clip': 0.04692964255809784, 'timing/data': 0.19070664793252945, 'timing/model': 1.9878753647208214, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 7.33}) +[2026-06-05 00:54:09,586][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 2.6449455958754697e-07, 'train/grad_norm_pre_clip': 3.915693378075957e-05, 'timing/data': 0.006037715822458267, 'timing/model': 2.1752810459583998, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 7.42}) +[2026-06-05 00:59:47,251][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpq0bvmih8/test.c -o /tmp/tmpq0bvmih8/test.o +[2026-06-05 00:59:47,272][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpq0bvmih8/test.o -laio -o /tmp/tmpq0bvmih8/a.out +[2026-06-05 00:59:47,956][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpyv88yf1b/test.c -o /tmp/tmpyv88yf1b/test.o +[2026-06-05 00:59:47,974][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpyv88yf1b/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpyv88yf1b/a.out +[2026-06-05 00:59:56,239][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 900 +[2026-06-05 00:59:56,240][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 1.659167736534073e-07, 'eval/loss': 0.0006870340439490974, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 234.79746704101564, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 173.08321237240867, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 234.79746704101564, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 234.79746704101564, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 173.08321237240867, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 234.79746704101564, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00616571307182312, 'timing/model': 2.1690082289278507, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 7.5}) +[2026-06-05 01:06:37,181][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 2.2351734685344127e-08, 'train/grad_norm_pre_clip': 1.1372475228199619e-06, 'timing/data': 0.010415840893983841, 'timing/model': 2.6701410356909037, 'learning_rate/qwen_vl_interface': 9.999593055976534e-06, 'learning_rate/action_model': 9.999593055976533e-05, 'epoch': 7.58}) +[2026-06-05 01:12:57,414][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.0956670343875885, 'train/grad_norm_pre_clip': 1.8284432888031006, 'timing/data': 0.19730602204799652, 'timing/model': 2.414097096771002, 'learning_rate/qwen_vl_interface': 9.998372293633897e-06, 'learning_rate/action_model': 9.998372293633897e-05, 'epoch': 7.67}) +[2026-06-05 01:19:42,887][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 2.544270500948187e-06, 'train/grad_norm_pre_clip': 0.00012721399252768606, 'timing/data': 0.005993640050292015, 'timing/model': 2.6723899990320206, 'learning_rate/qwen_vl_interface': 9.996337922143435e-06, 'learning_rate/action_model': 9.996337922143436e-05, 'epoch': 7.75}) +[2026-06-05 01:26:25,185][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 2.4809503429423785e-06, 'train/grad_norm_pre_clip': 0.00011939540854655206, 'timing/data': 0.015390206128358841, 'timing/model': 2.6739172115921974, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 7.83}) +[2026-06-05 01:33:06,135][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 5.5879340266073996e-08, 'train/grad_norm_pre_clip': 2.8608394586626673e-06, 'timing/data': 0.006029175594449043, 'timing/model': 2.734421895816922, 'learning_rate/qwen_vl_interface': 9.989829885383368e-06, 'learning_rate/action_model': 9.989829885383367e-05, 'epoch': 7.92}) +[2026-06-05 01:39:26,707][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.14559465646743774, 'train/grad_norm_pre_clip': 5.304540634155273, 'timing/data': 0.19937824085354805, 'timing/model': 2.409199858084321, 'learning_rate/qwen_vl_interface': 9.985357335232359e-06, 'learning_rate/action_model': 9.98535733523236e-05, 'epoch': 8.0}) +[2026-06-05 01:46:11,682][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 4.28406792707392e-07, 'train/grad_norm_pre_clip': 2.8505488444352522e-05, 'timing/data': 0.026250874623656273, 'timing/model': 2.7501841951161623, 'learning_rate/qwen_vl_interface': 9.980073405979643e-06, 'learning_rate/action_model': 9.980073405979643e-05, 'epoch': 8.08}) +[2026-06-05 01:52:52,281][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 9.313220061812899e-08, 'train/grad_norm_pre_clip': 3.0345476261572912e-05, 'timing/data': 0.005774335935711861, 'timing/model': 2.6786923203617334, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 8.17}) +[2026-06-05 01:59:33,380][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 8.917447121348232e-05, 'train/grad_norm_pre_clip': 0.005577830132097006, 'timing/data': 0.014364911243319511, 'timing/model': 2.665546763688326, 'learning_rate/qwen_vl_interface': 9.967075170535901e-06, 'learning_rate/action_model': 9.967075170535901e-05, 'epoch': 8.25}) +[2026-06-05 02:06:14,669][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-06-05 02:06:14,671][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 1.0068521078210324e-05, 'train/grad_norm_pre_clip': 0.0014324383810162544, 'eval/loss': 0.02658512443304062, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 314.96209716796875, 'rl_games_eval/mid_train/mean_length': 19.4, 'rl_games_eval/mid_train/std_reward': 293.63527765750086, 'rl_games_eval/mid_train/std_length': 8.73155198117723, 'rl_games_eval/mid_train/macro_mean_reward': 314.96209716796875, 'rl_games_eval/mid_train/macro_mean_length': 19.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 314.96209716796875, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 19.4, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 293.63527765750086, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 8.73155198117723, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 314.96209716796875, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.20655851438641548, 'timing/model': 2.392133519053459, 'learning_rate/qwen_vl_interface': 9.959363091525602e-06, 'learning_rate/action_model': 9.959363091525601e-05, 'epoch': 8.33}) +[2026-06-05 02:12:59,187][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.0005133107188157737, 'train/grad_norm_pre_clip': 0.030221818014979362, 'timing/data': 0.015631522983312607, 'timing/model': 2.675419732928276, 'learning_rate/qwen_vl_interface': 9.950844087393436e-06, 'learning_rate/action_model': 9.950844087393435e-05, 'epoch': 8.42}) +[2026-06-05 02:19:40,379][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 1.221892262037727e-06, 'train/grad_norm_pre_clip': 4.404337232699618e-05, 'timing/data': 0.005812453106045723, 'timing/model': 2.6785795725882053, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 8.5}) +[2026-06-05 02:26:21,180][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 2.9689792881981703e-06, 'train/grad_norm_pre_clip': 0.00026900460943579674, 'timing/data': 0.025586172938346863, 'timing/model': 2.6614616941660643, 'learning_rate/qwen_vl_interface': 9.93139128052587e-06, 'learning_rate/action_model': 9.93139128052587e-05, 'epoch': 8.58}) +[2026-06-05 02:32:41,488][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 4.206166522635613e-06, 'train/grad_norm_pre_clip': 0.000429846637416631, 'timing/data': 0.20601957105100155, 'timing/model': 2.3846196345984936, 'learning_rate/qwen_vl_interface': 9.920460810928786e-06, 'learning_rate/action_model': 9.920460810928785e-05, 'epoch': 8.67}) +[2026-06-05 02:39:26,079][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 1.012390657706419e-05, 'train/grad_norm_pre_clip': 0.0016104705864563584, 'timing/data': 0.006211262196302414, 'timing/model': 2.6750032659620047, 'learning_rate/qwen_vl_interface': 9.908730081915346e-06, 'learning_rate/action_model': 9.908730081915345e-05, 'epoch': 8.75}) +[2026-06-05 02:46:07,483][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 5.91148273088038e-06, 'train/grad_norm_pre_clip': 0.0008982029394246638, 'timing/data': 0.018087569624185562, 'timing/model': 2.6676704827696085, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 8.83}) +[2026-06-05 02:52:50,077][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 7.823108205684548e-08, 'train/grad_norm_pre_clip': 1.902743292703235e-06, 'timing/data': 0.01238946057856083, 'timing/model': 2.6632367726415396, 'learning_rate/qwen_vl_interface': 9.882876022415435e-06, 'learning_rate/action_model': 9.882876022415435e-05, 'epoch': 8.92}) +[2026-06-05 02:59:11,016][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 1.5326899927003979e-07, 'train/grad_norm_pre_clip': 5.417670763563365e-06, 'timing/data': 0.2044236809015274, 'timing/model': 2.4147701654583216, 'learning_rate/qwen_vl_interface': 9.868757121888965e-06, 'learning_rate/action_model': 9.868757121888964e-05, 'epoch': 9.0}) +[2026-06-05 03:25:42,126][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-05 03:25:42,182][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-05 03:25:43,390][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `deadly_corridor_train__bridge` +[2026-06-05 03:25:44,163][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-05 03:25:44,163][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-05 03:25:44,163][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-05 03:25:44,163][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-05 03:25:44,164][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-05 03:25:44,164][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-05 03:25:44,164][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-05 03:25:44,164][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-05 03:25:44,165][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-05 03:25:44,165][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-05 03:25:44,165][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-05 03:25:44,165][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-05 03:25:44,165][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-05 03:25:44,166][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-05 03:25:44,166][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-05 03:25:44,166][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-05 03:25:44,166][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-05 03:25:44,166][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-05 03:25:44,167][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-05 03:25:44,167][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-05 03:25:44,167][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-05 03:25:44,167][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-05 03:25:44,168][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-05 03:25:44,168][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-05 03:25:44,168][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-05 03:25:44,168][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-05 03:25:44,169][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-05 03:25:44,169][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-05 03:25:44,169][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-05 03:25:44,169][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-05 03:25:44,169][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-05 03:25:44,188][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `deadly_corridor_train__bridge__val` +[2026-06-05 03:25:44,217][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/config.full.yaml +[2026-06-05 03:25:44,223][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/config.yaml +[2026-06-05 03:25:44,223][starVLA.training.train_starvla][INFO] - Loaded best checkpoint metadata: step=800 score=317.74525146484376 +[2026-06-05 03:25:52,758][starVLA.training.train_starvla][INFO] - Loaded pretrained checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt, steps: 0 +[2026-06-05 03:25:55,919][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-05 03:25:55,919][starVLA.training.train_starvla][INFO] - Total optimization steps = 2500 +[2026-06-05 03:25:55,920][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-05 03:25:55,920][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-05 03:25:55,920][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-05 04:09:45,223][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-05 04:09:45,267][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-05 04:09:46,424][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `deadly_corridor_train__bridge` +[2026-06-05 04:09:47,119][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-05 04:09:47,119][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-05 04:09:47,120][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-05 04:09:47,120][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-05 04:09:47,120][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-05 04:09:47,120][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-05 04:09:47,120][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-05 04:09:47,121][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-05 04:09:47,121][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-05 04:09:47,121][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-05 04:09:47,121][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-05 04:09:47,122][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-05 04:09:47,122][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-05 04:09:47,122][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-05 04:09:47,122][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-05 04:09:47,122][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-05 04:09:47,123][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-05 04:09:47,123][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-05 04:09:47,123][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-05 04:09:47,124][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-05 04:09:47,124][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-05 04:09:47,124][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-05 04:09:47,124][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-05 04:09:47,124][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-05 04:09:47,124][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-05 04:09:47,125][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-05 04:09:47,125][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-05 04:09:47,125][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-05 04:09:47,125][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-05 04:09:47,125][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-05 04:09:47,126][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-05 04:09:47,145][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `deadly_corridor_train__bridge__val` +[2026-06-05 04:09:47,174][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/config.full.yaml +[2026-06-05 04:09:47,180][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/config.yaml +[2026-06-05 04:09:47,180][starVLA.training.train_starvla][INFO] - Loaded best checkpoint metadata: step=800 score=317.74525146484376 +[2026-06-05 04:09:47,180][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state, steps: 800 +[2026-06-05 04:09:48,599][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state +[2026-06-05 04:09:49,749][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-05 04:10:00,072][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-05 04:10:00,073][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-05 04:10:00,073][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-05 04:10:00,075][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-05 04:10:00,989][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-05 04:10:02,492][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-05 04:10:02,493][starVLA.training.train_starvla][INFO] - Total optimization steps = 2500 +[2026-06-05 04:10:02,493][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-05 04:10:02,494][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-05 04:10:02,494][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-05 04:10:49,423][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-05 04:10:49,471][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-05 04:10:50,671][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `deadly_corridor_train__bridge` +[2026-06-05 04:10:51,374][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-05 04:10:51,374][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-05 04:10:51,374][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-05 04:10:51,374][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-05 04:10:51,375][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-05 04:10:51,375][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-05 04:10:51,375][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-05 04:10:51,375][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-05 04:10:51,376][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-05 04:10:51,376][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-05 04:10:51,376][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-05 04:10:51,376][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-05 04:10:51,376][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-05 04:10:51,377][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-05 04:10:51,377][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-05 04:10:51,377][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-05 04:10:51,377][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-05 04:10:51,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-05 04:10:51,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-05 04:10:51,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-05 04:10:51,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-05 04:10:51,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-05 04:10:51,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-05 04:10:51,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-05 04:10:51,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-05 04:10:51,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-05 04:10:51,380][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-05 04:10:51,380][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-05 04:10:51,380][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-05 04:10:51,380][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-05 04:10:51,380][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-05 04:10:51,399][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `deadly_corridor_train__bridge__val` +[2026-06-05 04:10:51,429][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/config.full.yaml +[2026-06-05 04:10:51,435][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/config.yaml +[2026-06-05 04:10:51,435][starVLA.training.train_starvla][INFO] - Loaded best checkpoint metadata: step=800 score=317.74525146484376 +[2026-06-05 04:10:51,435][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state, steps: 800 +[2026-06-05 04:10:52,828][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state +[2026-06-05 04:10:53,963][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-05 04:11:04,546][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-05 04:11:04,547][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-05 04:11:04,547][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-05 04:11:04,549][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-05 04:11:05,594][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-05 04:11:07,142][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-05 04:11:07,143][starVLA.training.train_starvla][INFO] - Total optimization steps = 2500 +[2026-06-05 04:11:07,143][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-05 04:11:07,144][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-05 04:11:07,144][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-05 04:16:36,551][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 1.5273688802608376e-07, 'train/grad_norm_pre_clip': 2.613371862025815e-06, 'timing/data': 0.021271655336022377, 'timing/model': 2.161106873303652, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 6.75}) +[2026-06-05 04:22:01,151][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.05488263815641403, 'train/grad_norm_pre_clip': 2.329831600189209, 'timing/data': 0.011149583384394646, 'timing/model': 2.16800843924284, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 6.83}) +[2026-06-05 04:27:25,453][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 6.142519396235002e-06, 'train/grad_norm_pre_clip': 0.004569438751786947, 'timing/data': 0.006062867119908333, 'timing/model': 2.1764809992164373, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 6.92}) +[2026-06-05 04:32:33,543][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 4.982382233720273e-05, 'train/grad_norm_pre_clip': 0.0027156686410307884, 'timing/data': 0.22350378707051277, 'timing/model': 2.044855521991849, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 7.0}) +[2026-06-05 04:38:01,658][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.0021543283946812153, 'train/grad_norm_pre_clip': 0.059352315962314606, 'timing/data': 0.02107294835150242, 'timing/model': 2.162754815071821, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 7.08}) +[2026-06-05 04:43:25,755][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 8.009365046746098e-07, 'train/grad_norm_pre_clip': 5.662620242219418e-05, 'timing/data': 0.006189431995153427, 'timing/model': 2.1784804314374924, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 7.17}) +[2026-06-05 04:48:49,853][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 3.129243282273819e-07, 'train/grad_norm_pre_clip': 1.095525294658728e-05, 'timing/data': 0.005937272682785988, 'timing/model': 2.1759418435394764, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 7.25}) +[2026-06-05 04:53:57,255][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.002007229719310999, 'train/grad_norm_pre_clip': 0.036496881395578384, 'timing/data': 0.1800000760704279, 'timing/model': 1.9993424080312252, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 7.33}) +[2026-06-05 04:59:23,558][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 1.4156074712445843e-07, 'train/grad_norm_pre_clip': 7.052576620480977e-06, 'timing/data': 0.006029626354575157, 'timing/model': 2.185698976740241, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 7.42}) +[2026-06-05 05:05:02,768][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp6vymav4g/test.c -o /tmp/tmp6vymav4g/test.o +[2026-06-05 05:05:02,796][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp6vymav4g/test.o -laio -o /tmp/tmp6vymav4g/a.out +[2026-06-05 05:05:03,449][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpfyiio55f/test.c -o /tmp/tmpfyiio55f/test.o +[2026-06-05 05:05:03,478][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpfyiio55f/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpfyiio55f/a.out +[2026-06-05 05:05:10,906][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 900 +[2026-06-05 05:05:10,909][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 7.450580152834618e-09, 'train/grad_norm_pre_clip': 4.849502488468715e-07, 'eval/loss': 0.0005501306150108576, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 234.79746704101564, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 173.08321237240867, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 234.79746704101564, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 234.79746704101564, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 173.08321237240867, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 234.79746704101564, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.020165851339697838, 'timing/model': 2.1623944249004126, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 7.5}) +[2026-06-05 05:05:10,910][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_900_state +[2026-06-05 05:05:10,920][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-05 05:05:16,753][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_900_state/model.safetensors +[2026-06-05 05:05:24,898][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_900_state/optimizer.bin +[2026-06-05 05:05:24,898][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_900_state/sampler.bin +[2026-06-05 05:05:24,898][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_900_state/sampler_1.bin +[2026-06-05 05:05:24,900][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_900_state/random_states_0.pkl +[2026-06-05 05:05:24,902][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-06-05 05:05:24,902][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-05 05:05:24,915][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-05 05:12:07,854][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 3.725290076417309e-09, 'train/grad_norm_pre_clip': 2.1548399331550172e-07, 'timing/data': 0.015813546255230904, 'timing/model': 2.6715549398213625, 'learning_rate/qwen_vl_interface': 9.999593055976534e-06, 'learning_rate/action_model': 9.999593055976533e-05, 'epoch': 7.58}) +[2026-06-05 05:18:27,576][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.0055597457103431225, 'train/grad_norm_pre_clip': 0.6641752123832703, 'timing/data': 0.25474907271564007, 'timing/model': 2.5495769511908293, 'learning_rate/qwen_vl_interface': 9.998372293633897e-06, 'learning_rate/action_model': 9.998372293633897e-05, 'epoch': 7.67}) +[2026-06-05 05:25:12,255][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 3.9487838421337074e-07, 'train/grad_norm_pre_clip': 2.5483863282715902e-05, 'timing/data': 0.02147517167031765, 'timing/model': 2.6657645292580128, 'learning_rate/qwen_vl_interface': 9.996337922143435e-06, 'learning_rate/action_model': 9.996337922143436e-05, 'epoch': 7.75}) +[2026-06-05 05:31:53,756][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 1.8477312551112846e-06, 'train/grad_norm_pre_clip': 0.00015498943685088307, 'timing/data': 0.015491286292672157, 'timing/model': 2.668094115331769, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 7.83}) +[2026-06-05 05:38:35,369][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 1.5497048480028752e-06, 'train/grad_norm_pre_clip': 0.0001585024147061631, 'timing/data': 0.0064823199063539505, 'timing/model': 2.888393435627222, 'learning_rate/qwen_vl_interface': 9.989829885383368e-06, 'learning_rate/action_model': 9.989829885383367e-05, 'epoch': 7.92}) +[2026-06-05 05:44:54,892][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.03128571808338165, 'train/grad_norm_pre_clip': 2.0469470024108887, 'timing/data': 0.17557450756430626, 'timing/model': 2.447446707636118, 'learning_rate/qwen_vl_interface': 9.985357335232359e-06, 'learning_rate/action_model': 9.98535733523236e-05, 'epoch': 8.0}) +[2026-06-05 05:51:39,852][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 3.013709829247091e-06, 'train/grad_norm_pre_clip': 0.00016503120423294604, 'timing/data': 0.010986262932419777, 'timing/model': 2.7171727810055017, 'learning_rate/qwen_vl_interface': 9.980073405979643e-06, 'learning_rate/action_model': 9.980073405979643e-05, 'epoch': 8.08}) +[2026-06-05 05:58:20,650][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 5.4797060329292435e-06, 'train/grad_norm_pre_clip': 0.0005161953740753233, 'timing/data': 0.018947673961520195, 'timing/model': 2.6623639203608036, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 8.17}) +[2026-06-05 06:05:01,954][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.00066796486498788, 'train/grad_norm_pre_clip': 0.10665368288755417, 'timing/data': 0.010746102780103683, 'timing/model': 2.671592330560088, 'learning_rate/qwen_vl_interface': 9.967075170535901e-06, 'learning_rate/action_model': 9.967075170535901e-05, 'epoch': 8.25}) +[2026-06-05 06:11:42,003][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-06-05 06:11:42,004][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 2.5119081215052574e-07, 'train/grad_norm_pre_clip': 1.0846763871086296e-05, 'eval/loss': 0.03558269888162613, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 181.0637063264991, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 181.0637063264991, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 222.22349853515624, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.18186425045132637, 'timing/model': 2.535368638113141, 'learning_rate/qwen_vl_interface': 9.959363091525602e-06, 'learning_rate/action_model': 9.959363091525601e-05, 'epoch': 8.33}) +[2026-06-05 06:11:42,004][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1000_state +[2026-06-05 06:11:42,009][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-05 06:11:47,835][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1000_state/model.safetensors +[2026-06-05 06:11:56,205][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1000_state/optimizer.bin +[2026-06-05 06:11:56,206][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1000_state/sampler.bin +[2026-06-05 06:11:56,206][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-05 06:11:56,207][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-05 06:11:56,342][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-05 06:11:56,343][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-05 06:11:56,356][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-05 06:18:40,651][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.00875253789126873, 'train/grad_norm_pre_clip': 0.16463837027549744, 'timing/data': 0.008355708792805672, 'timing/model': 2.669596403837204, 'learning_rate/qwen_vl_interface': 9.950844087393436e-06, 'learning_rate/action_model': 9.950844087393435e-05, 'epoch': 8.42}) +[2026-06-05 06:25:21,057][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 3.725290076417309e-09, 'train/grad_norm_pre_clip': 4.699208489000739e-07, 'timing/data': 0.00613376684486866, 'timing/model': 2.6787882056087255, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 8.5}) +[2026-06-05 06:32:01,853][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 3.268990766969182e-08, 'timing/data': 0.01108672097325325, 'timing/model': 2.6721181105822325, 'learning_rate/qwen_vl_interface': 9.93139128052587e-06, 'learning_rate/action_model': 9.93139128052587e-05, 'epoch': 8.58}) +[2026-06-05 06:38:22,082][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 8.940686058167557e-08, 'train/grad_norm_pre_clip': 1.8181382984039374e-05, 'timing/data': 0.385970963165164, 'timing/model': 2.5234269220381975, 'learning_rate/qwen_vl_interface': 9.920460810928786e-06, 'learning_rate/action_model': 9.920460810928785e-05, 'epoch': 8.67}) +[2026-06-05 06:45:06,756][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 9.685746960030883e-08, 'train/grad_norm_pre_clip': 5.714374765375396e-06, 'timing/data': 0.015661340206861496, 'timing/model': 2.6706085484474897, 'learning_rate/qwen_vl_interface': 9.908730081915346e-06, 'learning_rate/action_model': 9.908730081915345e-05, 'epoch': 8.75}) +[2026-06-05 06:51:47,609][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 4.8428741905581774e-08, 'train/grad_norm_pre_clip': 1.9573649296944495e-06, 'timing/data': 0.026893511414527893, 'timing/model': 2.694375980645418, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 8.83}) +[2026-06-05 06:58:27,861][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 2.0703482732642442e-05, 'train/grad_norm_pre_clip': 0.0005805964465253055, 'timing/data': 0.006138252094388008, 'timing/model': 2.676292998716235, 'learning_rate/qwen_vl_interface': 9.882876022415435e-06, 'learning_rate/action_model': 9.882876022415435e-05, 'epoch': 8.92}) +[2026-06-05 07:04:47,580][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.0005259855533950031, 'train/grad_norm_pre_clip': 0.012965490110218525, 'timing/data': 0.17543984949588776, 'timing/model': 2.531481647863984, 'learning_rate/qwen_vl_interface': 9.868757121888965e-06, 'learning_rate/action_model': 9.868757121888964e-05, 'epoch': 9.0}) +[2026-06-05 07:11:31,668][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.0013009471585974097, 'train/grad_norm_pre_clip': 0.09751749038696289, 'timing/data': 0.011816715821623802, 'timing/model': 2.7818891052156687, 'learning_rate/qwen_vl_interface': 9.853846821107084e-06, 'learning_rate/action_model': 9.853846821107085e-05, 'epoch': 9.08}) +[2026-06-05 07:18:34,074][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1100 +[2026-06-05 07:18:34,075][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.0002886242000386119, 'train/grad_norm_pre_clip': 0.038316935300827026, 'eval/loss': 0.07931292057037354, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 212.00867309570313, 'rl_games_eval/mid_train/mean_length': 16.6, 'rl_games_eval/mid_train/std_reward': 160.51142783983695, 'rl_games_eval/mid_train/std_length': 4.841487374764082, 'rl_games_eval/mid_train/macro_mean_reward': 212.00867309570313, 'rl_games_eval/mid_train/macro_mean_length': 16.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 212.00867309570313, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 16.6, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 160.51142783983695, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 4.841487374764082, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 212.00867309570313, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0057741496711969376, 'timing/model': 2.675536287948489, 'learning_rate/qwen_vl_interface': 9.838147674873076e-06, 'learning_rate/action_model': 9.838147674873075e-05, 'epoch': 9.17}) +[2026-06-05 07:18:34,076][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1100_state +[2026-06-05 07:18:34,081][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-05 07:18:39,745][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1100_state/model.safetensors +[2026-06-05 07:18:47,916][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1100_state/optimizer.bin +[2026-06-05 07:18:47,917][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1100_state/sampler.bin +[2026-06-05 07:18:47,917][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1100_state/sampler_1.bin +[2026-06-05 07:18:47,918][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1100_state/random_states_0.pkl +[2026-06-05 07:18:48,042][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-06-05 07:18:48,043][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-05 07:18:48,056][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-05 07:25:28,958][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 3.4123302157240687e-06, 'train/grad_norm_pre_clip': 5.219426020630635e-05, 'timing/data': 0.016198433935642242, 'timing/model': 2.6643051095306873, 'learning_rate/qwen_vl_interface': 9.821662373154824e-06, 'learning_rate/action_model': 9.821662373154825e-05, 'epoch': 9.25}) +[2026-06-05 07:31:49,189][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.10391313582658768, 'train/grad_norm_pre_clip': 6.408139228820801, 'timing/data': 0.2405672762542963, 'timing/model': 2.5729801561683416, 'learning_rate/qwen_vl_interface': 9.804393740623918e-06, 'learning_rate/action_model': 9.804393740623918e-05, 'epoch': 9.33}) +[2026-06-05 07:38:33,353][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 1.147384750765923e-06, 'train/grad_norm_pre_clip': 3.407188705750741e-05, 'timing/data': 0.015311982482671738, 'timing/model': 2.664891542866826, 'learning_rate/qwen_vl_interface': 9.786344736171637e-06, 'learning_rate/action_model': 9.786344736171638e-05, 'epoch': 9.42}) +[2026-06-05 07:45:14,258][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 3.9218277379404753e-05, 'train/grad_norm_pre_clip': 0.00500058988109231, 'timing/data': 0.01598917320370674, 'timing/model': 2.671009600162506, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 9.5}) +[2026-06-05 07:51:54,260][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.016699407249689102, 'train/grad_norm_pre_clip': 1.6644113063812256, 'timing/data': 0.005917763337492943, 'timing/model': 2.683207008987665, 'learning_rate/qwen_vl_interface': 9.747918115101753e-06, 'learning_rate/action_model': 9.747918115101753e-05, 'epoch': 9.58}) +[2026-06-05 07:58:13,767][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 2.8269334961805725e-06, 'train/grad_norm_pre_clip': 0.00013275038509164006, 'timing/data': 0.16878237202763557, 'timing/model': 2.527431096881628, 'learning_rate/qwen_vl_interface': 9.727547082687849e-06, 'learning_rate/action_model': 9.72754708268785e-05, 'epoch': 9.67}) +[2026-06-05 08:04:58,655][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 2.6021407393272966e-05, 'train/grad_norm_pre_clip': 0.0016165281413123012, 'timing/data': 0.02587810344994068, 'timing/model': 2.7408398631960154, 'learning_rate/qwen_vl_interface': 9.7064088456318e-06, 'learning_rate/action_model': 9.7064088456318e-05, 'epoch': 9.75}) +[2026-06-05 08:11:38,856][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.0051923226565122604, 'train/grad_norm_pre_clip': 0.41990816593170166, 'timing/data': 0.006104752421379089, 'timing/model': 2.6754279602319, 'learning_rate/qwen_vl_interface': 9.68450702586171e-06, 'learning_rate/action_model': 9.684507025861709e-05, 'epoch': 9.83}) +[2026-06-05 08:18:19,062][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.0005160250584594905, 'train/grad_norm_pre_clip': 0.02581053413450718, 'timing/data': 0.006055982783436775, 'timing/model': 2.6794966366142035, 'learning_rate/qwen_vl_interface': 9.66184537614164e-06, 'learning_rate/action_model': 9.661845376141641e-05, 'epoch': 9.92}) +[2026-06-05 08:24:59,302][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1200 +[2026-06-05 08:24:59,304][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 7.365411534010491e-07, 'train/grad_norm_pre_clip': 1.7314816432190128e-05, 'eval/loss': 0.01459280401468277, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/mean_length': 16.4, 'rl_games_eval/mid_train/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/std_length': 4.586937976471886, 'rl_games_eval/mid_train/macro_mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/macro_mean_length': 16.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 16.4, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 4.586937976471886, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 229.50448608398438, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.1747885588556528, 'timing/model': 2.442403858527541, 'learning_rate/qwen_vl_interface': 9.638427779428613e-06, 'learning_rate/action_model': 9.638427779428613e-05, 'epoch': 10.0}) +[2026-06-05 08:24:59,305][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1200_state +[2026-06-05 08:24:59,311][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-05 08:25:04,846][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1200_state/model.safetensors +[2026-06-05 08:25:13,456][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1200_state/optimizer.bin +[2026-06-05 08:25:13,457][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1200_state/sampler.bin +[2026-06-05 08:25:13,457][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1200_state/sampler_1.bin +[2026-06-05 08:25:13,458][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1200_state/random_states_0.pkl +[2026-06-05 08:25:13,593][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-06-05 08:25:13,593][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-05 08:25:13,607][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-05 08:31:58,160][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 3.6014182114740834e-05, 'train/grad_norm_pre_clip': 0.003981282934546471, 'timing/data': 0.01622895896434784, 'timing/model': 2.6687802840024233, 'learning_rate/qwen_vl_interface': 9.61425824820727e-06, 'learning_rate/action_model': 9.614258248207268e-05, 'epoch': 10.08}) +[2026-06-05 08:38:38,971][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 8.568160581035045e-08, 'train/grad_norm_pre_clip': 4.369618181954138e-06, 'timing/data': 0.022165613248944283, 'timing/model': 2.668471287935972, 'learning_rate/qwen_vl_interface': 9.589340923802355e-06, 'learning_rate/action_model': 9.589340923802354e-05, 'epoch': 10.17}) +[2026-06-05 08:45:19,557][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 7.450580152834618e-09, 'train/grad_norm_pre_clip': 2.2146141986922885e-07, 'timing/data': 0.005916683003306389, 'timing/model': 2.6770399510860443, 'learning_rate/qwen_vl_interface': 9.563680075669138e-06, 'learning_rate/action_model': 9.563680075669137e-05, 'epoch': 10.25}) +[2026-06-05 08:51:39,100][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 4.257474373048353e-09, 'train/grad_norm_pre_clip': 2.05653520879423e-07, 'timing/data': 0.17158429883420467, 'timing/model': 2.451856829226017, 'learning_rate/qwen_vl_interface': 9.53728010066184e-06, 'learning_rate/action_model': 9.53728010066184e-05, 'epoch': 10.33}) +[2026-06-05 08:58:22,461][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 7.4505797087454084e-09, 'train/grad_norm_pre_clip': 2.3209339872209966e-07, 'timing/data': 0.0208990890532732, 'timing/model': 2.664249423891306, 'learning_rate/qwen_vl_interface': 9.51014552228027e-06, 'learning_rate/action_model': 9.510145522280269e-05, 'epoch': 10.42}) +[2026-06-05 09:05:03,258][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 3.7997801882738713e-07, 'train/grad_norm_pre_clip': 2.3603730369359255e-05, 'timing/data': 0.020564474165439606, 'timing/model': 2.660746194422245, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 10.5}) +[2026-06-05 09:11:43,556][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 1.4901160305669237e-08, 'train/grad_norm_pre_clip': 3.4711248986241117e-07, 'timing/data': 0.013883545994758606, 'timing/model': 2.664769561961293, 'learning_rate/qwen_vl_interface': 9.45369127794945e-06, 'learning_rate/action_model': 9.453691277949449e-05, 'epoch': 10.58}) +[2026-06-05 09:18:03,085][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 8.514948746096707e-09, 'train/grad_norm_pre_clip': 1.9385558402973402e-07, 'timing/data': 0.1776518765836954, 'timing/model': 2.5357604548335075, 'learning_rate/qwen_vl_interface': 9.424381285144337e-06, 'learning_rate/action_model': 9.424381285144335e-05, 'epoch': 10.67}) +[2026-06-05 09:24:46,862][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.00057912094052881, 'train/grad_norm_pre_clip': 0.012856172397732735, 'timing/data': 0.02019975334405899, 'timing/model': 2.668152380734682, 'learning_rate/qwen_vl_interface': 9.394356033595787e-06, 'learning_rate/action_model': 9.394356033595787e-05, 'epoch': 10.75}) +[2026-06-05 09:31:49,645][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1300 +[2026-06-05 09:31:49,647][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.00015640231140423566, 'train/grad_norm_pre_clip': 0.01154558826237917, 'eval/loss': 0.018562372773885727, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 224.13310852050782, 'rl_games_eval/mid_train/mean_length': 17.4, 'rl_games_eval/mid_train/std_reward': 192.64319343893612, 'rl_games_eval/mid_train/std_length': 5.817215828899594, 'rl_games_eval/mid_train/macro_mean_reward': 224.13310852050782, 'rl_games_eval/mid_train/macro_mean_length': 17.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 224.13310852050782, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.4, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 192.64319343893612, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.817215828899594, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 224.13310852050782, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005321744829416275, 'timing/model': 2.6813733391463757, 'learning_rate/qwen_vl_interface': 9.363620667976085e-06, 'learning_rate/action_model': 9.363620667976085e-05, 'epoch': 10.83}) +[2026-06-05 09:31:49,647][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1300_state +[2026-06-05 09:31:49,653][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-05 09:31:55,429][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1300_state/model.safetensors +[2026-06-05 09:32:03,784][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1300_state/optimizer.bin +[2026-06-05 09:32:03,784][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1300_state/sampler.bin +[2026-06-05 09:32:03,784][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1300_state/sampler_1.bin +[2026-06-05 09:32:03,786][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1300_state/random_states_0.pkl +[2026-06-05 09:32:03,917][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-06-05 09:32:03,917][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-05 09:32:03,931][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-05 09:38:44,561][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 9.696706911199726e-06, 'train/grad_norm_pre_clip': 0.00015756751236040145, 'timing/data': 0.008448528125882149, 'timing/model': 2.681984595954418, 'learning_rate/qwen_vl_interface': 9.332180454631902e-06, 'learning_rate/action_model': 9.332180454631901e-05, 'epoch': 10.92}) +[2026-06-05 09:45:04,154][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 1.5454605772902141e-06, 'train/grad_norm_pre_clip': 2.451532236591447e-05, 'timing/data': 0.18536312691867352, 'timing/model': 2.396122755482793, 'learning_rate/qwen_vl_interface': 9.300040780681938e-06, 'learning_rate/action_model': 9.300040780681938e-05, 'epoch': 11.0}) +[2026-06-05 09:51:48,061][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 5.997711696181796e-07, 'train/grad_norm_pre_clip': 6.885422862978885e-06, 'timing/data': 0.011090569198131561, 'timing/model': 2.6726306080818176, 'learning_rate/qwen_vl_interface': 9.267207153093874e-06, 'learning_rate/action_model': 9.267207153093875e-05, 'epoch': 11.08}) +[2026-06-05 09:58:29,266][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 7.562330210930668e-07, 'train/grad_norm_pre_clip': 1.5174604413914494e-05, 'timing/data': 0.015089455991983414, 'timing/model': 2.6677446123212576, 'learning_rate/qwen_vl_interface': 9.233685197740765e-06, 'learning_rate/action_model': 9.233685197740764e-05, 'epoch': 11.17}) +[2026-06-05 10:05:10,459][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 1.4156062206893694e-06, 'train/grad_norm_pre_clip': 2.2940525013837032e-05, 'timing/data': 0.011198248714208603, 'timing/model': 2.673687055706978, 'learning_rate/qwen_vl_interface': 9.199480658437092e-06, 'learning_rate/action_model': 9.199480658437091e-05, 'epoch': 11.25}) +[2026-06-05 10:11:29,687][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 4.4786165744881146e-06, 'train/grad_norm_pre_clip': 0.0004442202625796199, 'timing/data': 0.16894256882369518, 'timing/model': 2.443282950669527, 'learning_rate/qwen_vl_interface': 9.164599395954575e-06, 'learning_rate/action_model': 9.164599395954574e-05, 'epoch': 11.33}) +[2026-06-05 10:18:14,062][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 3.017483720668679e-07, 'train/grad_norm_pre_clip': 4.166308826825116e-06, 'timing/data': 0.021771395578980446, 'timing/model': 2.6683738473802805, 'learning_rate/qwen_vl_interface': 9.129047387017977e-06, 'learning_rate/action_model': 9.129047387017976e-05, 'epoch': 11.42}) +[2026-06-05 10:24:55,465][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 2.0861621408130304e-07, 'train/grad_norm_pre_clip': 2.9738359899056377e-06, 'timing/data': 0.02165892720222473, 'timing/model': 2.666205696761608, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 11.5}) +[2026-06-05 10:31:36,160][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.00018896859546657652, 'train/grad_norm_pre_clip': 0.00534800486639142, 'timing/data': 0.016462234780192375, 'timing/model': 2.6658913549035788, 'learning_rate/qwen_vl_interface': 9.055955610282539e-06, 'learning_rate/action_model': 9.055955610282538e-05, 'epoch': 11.58}) +[2026-06-05 10:38:16,691][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1400 +[2026-06-05 10:38:16,693][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 2.8825481422245502e-05, 'train/grad_norm_pre_clip': 0.004073368851095438, 'eval/loss': 0.0003777004894800484, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 181.0637063264991, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 181.0637063264991, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 222.22349853515624, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.1767272036522627, 'timing/model': 2.532787824049592, 'learning_rate/qwen_vl_interface': 9.018428366383368e-06, 'learning_rate/action_model': 9.018428366383368e-05, 'epoch': 11.67}) +[2026-06-05 10:38:16,693][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1400_state +[2026-06-05 10:38:16,698][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-05 10:38:22,684][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1400_state/model.safetensors +[2026-06-05 10:38:30,862][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1400_state/optimizer.bin +[2026-06-05 10:38:30,863][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1400_state/sampler.bin +[2026-06-05 10:38:30,863][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1400_state/sampler_1.bin +[2026-06-05 10:38:30,865][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1400_state/random_states_0.pkl +[2026-06-05 10:38:30,989][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-06-05 10:38:30,989][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-05 10:38:31,003][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-05 10:45:15,063][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 1.2889488516520942e-06, 'train/grad_norm_pre_clip': 2.392579881416168e-05, 'timing/data': 0.013283122330904007, 'timing/model': 2.66748527623713, 'learning_rate/qwen_vl_interface': 8.98025542168354e-06, 'learning_rate/action_model': 8.98025542168354e-05, 'epoch': 11.75}) +[2026-06-05 10:51:55,359][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 2.7455132567411056e-06, 'train/grad_norm_pre_clip': 5.996378240524791e-05, 'timing/data': 0.014424009248614311, 'timing/model': 2.6656856555491686, 'learning_rate/qwen_vl_interface': 8.941443316920613e-06, 'learning_rate/action_model': 8.941443316920612e-05, 'epoch': 11.83}) +[2026-06-05 10:58:35,660][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.032855596393346786, 'train/grad_norm_pre_clip': 8.627985954284668, 'timing/data': 0.011006567627191544, 'timing/model': 2.6656578313559294, 'learning_rate/qwen_vl_interface': 8.901998702348934e-06, 'learning_rate/action_model': 8.901998702348933e-05, 'epoch': 11.92}) +[2026-06-05 11:04:55,795][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 2.0455863705137745e-05, 'train/grad_norm_pre_clip': 0.0033422941341996193, 'timing/data': 0.3881410025060177, 'timing/model': 2.5313748456537724, 'learning_rate/qwen_vl_interface': 8.861928336600148e-06, 'learning_rate/action_model': 8.861928336600148e-05, 'epoch': 12.0}) +[2026-06-05 11:11:39,365][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 5.073480679129716e-06, 'train/grad_norm_pre_clip': 0.00018074666149914265, 'timing/data': 0.016228439286351204, 'timing/model': 2.669987518340349, 'learning_rate/qwen_vl_interface': 8.821239085525143e-06, 'learning_rate/action_model': 8.821239085525143e-05, 'epoch': 12.08}) +[2026-06-05 11:18:20,256][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 5.215405707303944e-08, 'train/grad_norm_pre_clip': 9.543507530906936e-07, 'timing/data': 0.016443949192762375, 'timing/model': 2.666738659143448, 'learning_rate/qwen_vl_interface': 8.779937921017624e-06, 'learning_rate/action_model': 8.779937921017623e-05, 'epoch': 12.17}) +[2026-06-05 11:25:00,759][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 3.352760913344355e-08, 'train/grad_norm_pre_clip': 1.20006222914526e-06, 'timing/data': 0.006122292950749397, 'timing/model': 2.681264692917466, 'learning_rate/qwen_vl_interface': 8.738031919819508e-06, 'learning_rate/action_model': 8.738031919819508e-05, 'epoch': 12.25}) +[2026-06-05 11:31:20,183][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 5.10896818184392e-08, 'train/grad_norm_pre_clip': 1.40407576054713e-06, 'timing/data': 0.18278834782540798, 'timing/model': 2.5086992774158716, 'learning_rate/qwen_vl_interface': 8.695528262308367e-06, 'learning_rate/action_model': 8.695528262308366e-05, 'epoch': 12.33}) +[2026-06-05 11:38:03,557][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.00019986971165053546, 'train/grad_norm_pre_clip': 0.004449633415788412, 'timing/data': 0.011158088222146034, 'timing/model': 2.7163383420556784, 'learning_rate/qwen_vl_interface': 8.65243423126711e-06, 'learning_rate/action_model': 8.652434231267109e-05, 'epoch': 12.42}) +[2026-06-05 11:45:06,101][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-05 11:45:06,103][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 2.1606656730455143e-07, 'train/grad_norm_pre_clip': 5.7689862842380535e-06, 'eval/loss': 1.189712838822743e-05, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 181.0637063264991, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 181.0637063264991, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 222.22349853515624, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.023616015911102295, 'timing/model': 2.664402713999152, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 12.5}) +[2026-06-05 11:45:06,103][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1500_state +[2026-06-05 11:45:06,108][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-05 11:45:11,912][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1500_state/model.safetensors +[2026-06-05 11:45:20,198][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1500_state/optimizer.bin +[2026-06-05 11:45:20,199][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1500_state/sampler.bin +[2026-06-05 11:45:20,199][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-05 11:45:20,201][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-05 11:45:20,332][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-05 11:45:20,332][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-05 11:45:20,346][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-05 11:52:00,868][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 8.361048094229773e-05, 'train/grad_norm_pre_clip': 0.0018373167840763927, 'timing/data': 0.02185661904513836, 'timing/model': 2.6638900376856327, 'learning_rate/qwen_vl_interface': 8.564504684247982e-06, 'learning_rate/action_model': 8.564504684247982e-05, 'epoch': 12.58}) +[2026-06-05 11:58:20,599][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 1.8732860951331531e-07, 'train/grad_norm_pre_clip': 9.686842531664297e-06, 'timing/data': 0.2743725758045912, 'timing/model': 2.4500477518886328, 'learning_rate/qwen_vl_interface': 8.519684234545333e-06, 'learning_rate/action_model': 8.519684234545332e-05, 'epoch': 12.67}) +[2026-06-05 12:05:04,555][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 2.905720748458407e-07, 'train/grad_norm_pre_clip': 2.289928488607984e-05, 'timing/data': 0.006499966606497765, 'timing/model': 2.675274331122637, 'learning_rate/qwen_vl_interface': 8.474303541281474e-06, 'learning_rate/action_model': 8.474303541281474e-05, 'epoch': 12.75}) +[2026-06-05 12:11:45,059][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 2.9802320611338473e-08, 'train/grad_norm_pre_clip': 1.1449784551587072e-06, 'timing/data': 0.01100182719528675, 'timing/model': 2.672798665240407, 'learning_rate/qwen_vl_interface': 8.428370380204577e-06, 'learning_rate/action_model': 8.428370380204577e-05, 'epoch': 12.83}) +[2026-06-05 12:18:25,981][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.0016491764690726995, 'train/grad_norm_pre_clip': 0.13462606072425842, 'timing/data': 0.013336347416043282, 'timing/model': 2.8900632448494434, 'learning_rate/qwen_vl_interface': 8.381892621725327e-06, 'learning_rate/action_model': 8.381892621725327e-05, 'epoch': 12.92}) +[2026-06-05 12:24:45,478][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.0004591052420437336, 'train/grad_norm_pre_clip': 0.04580431804060936, 'timing/data': 0.1680891290307045, 'timing/model': 2.5394949670881033, 'learning_rate/qwen_vl_interface': 8.334878229568372e-06, 'learning_rate/action_model': 8.334878229568372e-05, 'epoch': 13.0}) +[2026-06-05 12:31:28,554][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 4.61706003989093e-05, 'train/grad_norm_pre_clip': 0.0013959972420707345, 'timing/data': 0.005860043689608574, 'timing/model': 2.67598032951355, 'learning_rate/qwen_vl_interface': 8.287335259407779e-06, 'learning_rate/action_model': 8.287335259407779e-05, 'epoch': 13.08}) +[2026-06-05 12:38:09,052][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 3.6729870771523565e-06, 'train/grad_norm_pre_clip': 0.00018970372912008315, 'timing/data': 0.01567346416413784, 'timing/model': 2.6660256516188383, 'learning_rate/qwen_vl_interface': 8.23927185748673e-06, 'learning_rate/action_model': 8.239271857486729e-05, 'epoch': 13.17}) +[2026-06-05 12:44:49,653][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 1.862644971595273e-08, 'train/grad_norm_pre_clip': 3.476073118235945e-07, 'timing/data': 0.011673269793391228, 'timing/model': 2.675339473411441, 'learning_rate/qwen_vl_interface': 8.190696259221713e-06, 'learning_rate/action_model': 8.190696259221713e-05, 'epoch': 13.25}) +[2026-06-05 12:51:29,601][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1600 +[2026-06-05 12:51:29,602][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 4.257474373048353e-09, 'train/grad_norm_pre_clip': 1.2907871393963433e-07, 'eval/loss': 0.02970566786825657, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 181.0637063264991, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 222.22349853515624, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 181.0637063264991, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 222.22349853515624, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.20655172504484653, 'timing/model': 2.388566128909588, 'learning_rate/qwen_vl_interface': 8.141616787791425e-06, 'learning_rate/action_model': 8.141616787791424e-05, 'epoch': 13.33}) +[2026-06-05 12:51:29,603][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1600_state +[2026-06-05 12:51:29,607][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-05 12:51:35,615][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1600_state/model.safetensors +[2026-06-05 12:51:43,936][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1600_state/optimizer.bin +[2026-06-05 12:51:43,937][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1600_state/sampler.bin +[2026-06-05 12:51:43,937][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1600_state/sampler_1.bin +[2026-06-05 12:51:43,938][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1600_state/random_states_0.pkl +[2026-06-05 12:51:44,061][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1600 +[2026-06-05 12:51:44,062][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-05 12:51:44,075][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-05 12:58:28,960][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 3.254380587236483e-08, 'timing/data': 0.021991323679685593, 'timing/model': 2.6646514013409615, 'learning_rate/qwen_vl_interface': 8.092041852710631e-06, 'learning_rate/action_model': 8.092041852710631e-05, 'epoch': 13.42}) +[2026-06-05 13:05:09,758][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 3.837036217646528e-07, 'train/grad_norm_pre_clip': 1.21883376777987e-05, 'timing/data': 0.01558920368552208, 'timing/model': 2.6712572630494833, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 13.5}) +[2026-06-05 13:11:49,545][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.000244185357587412, 'train/grad_norm_pre_clip': 0.0311364084482193, 'timing/data': 0.015295477584004402, 'timing/model': 2.669950196519494, 'learning_rate/qwen_vl_interface': 7.991439652676861e-06, 'learning_rate/action_model': 7.991439652676861e-05, 'epoch': 13.58}) +[2026-06-05 13:18:08,776][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 1.2664984751609154e-05, 'train/grad_norm_pre_clip': 0.00033436299418099225, 'timing/data': 0.1989852637052536, 'timing/model': 2.505965769290924, 'learning_rate/qwen_vl_interface': 7.940429625392957e-06, 'learning_rate/action_model': 7.940429625392956e-05, 'epoch': 13.67}) +[2026-06-05 13:24:52,550][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 1.7763768482836895e-05, 'train/grad_norm_pre_clip': 0.00041491922456771135, 'timing/data': 0.010763727128505707, 'timing/model': 2.6704562101513147, 'learning_rate/qwen_vl_interface': 7.888958606843111e-06, 'learning_rate/action_model': 7.888958606843111e-05, 'epoch': 13.75}) +[2026-06-05 13:31:33,174][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 1.5295068806153722e-05, 'train/grad_norm_pre_clip': 0.00025307759642601013, 'timing/data': 0.015533797442913055, 'timing/model': 2.8869201596826315, 'learning_rate/qwen_vl_interface': 7.83703541632138e-06, 'learning_rate/action_model': 7.83703541632138e-05, 'epoch': 13.83}) +[2026-06-05 13:38:14,159][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 2.274079088238068e-05, 'train/grad_norm_pre_clip': 0.0003972134436480701, 'timing/data': 0.00600501149892807, 'timing/model': 2.6755083799362183, 'learning_rate/qwen_vl_interface': 7.784668950599155e-06, 'learning_rate/action_model': 7.784668950599155e-05, 'epoch': 13.92}) +[2026-06-05 13:44:33,591][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 3.204570384696126e-05, 'train/grad_norm_pre_clip': 0.0010863739298656583, 'timing/data': 0.1742645800113678, 'timing/model': 2.546004720032215, 'learning_rate/qwen_vl_interface': 7.731868182400758e-06, 'learning_rate/action_model': 7.731868182400758e-05, 'epoch': 14.0}) +[2026-06-05 13:51:17,255][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 1.0430777592773666e-06, 'train/grad_norm_pre_clip': 2.284455877088476e-05, 'timing/data': 0.006076371297240257, 'timing/model': 2.6620259061455727, 'learning_rate/qwen_vl_interface': 7.678642158865997e-06, 'learning_rate/action_model': 7.678642158865997e-05, 'epoch': 14.08}) +[2026-06-05 13:58:17,398][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1700 +[2026-06-05 13:58:17,400][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 6.44471583655104e-07, 'train/grad_norm_pre_clip': 5.804386455565691e-05, 'eval/loss': 0.02738087996840477, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/mean_length': 16.4, 'rl_games_eval/mid_train/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/std_length': 4.586937976471886, 'rl_games_eval/mid_train/macro_mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/macro_mean_length': 16.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 16.4, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 4.586937976471886, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 229.50448608398438, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.011069504544138908, 'timing/model': 2.670812424272299, 'learning_rate/qwen_vl_interface': 7.625e-06, 'learning_rate/action_model': 7.625e-05, 'epoch': 14.17}) +[2026-06-05 13:58:17,401][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1700_state +[2026-06-05 13:58:17,405][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-05 13:58:23,233][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1700_state/model.safetensors +[2026-06-05 13:58:31,657][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1700_state/optimizer.bin +[2026-06-05 13:58:31,658][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1700_state/sampler.bin +[2026-06-05 13:58:31,658][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1700_state/sampler_1.bin +[2026-06-05 13:58:31,659][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1700_state/random_states_0.pkl +[2026-06-05 13:58:31,793][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1700 +[2026-06-05 13:58:31,793][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-05 13:58:31,807][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-05 14:05:12,070][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 2.5331962660857243e-07, 'train/grad_norm_pre_clip': 2.61835543824418e-06, 'timing/data': 0.0062910933047533035, 'timing/model': 2.6900519281625748, 'learning_rate/qwen_vl_interface': 7.570950897110537e-06, 'learning_rate/action_model': 7.570950897110536e-05, 'epoch': 14.25}) +[2026-06-05 14:11:32,488][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 1.4049660990167467e-07, 'train/grad_norm_pre_clip': 2.6600371256790822e-06, 'timing/data': 0.2681419011205435, 'timing/model': 2.549707643687725, 'learning_rate/qwen_vl_interface': 7.516504111233141e-06, 'learning_rate/action_model': 7.51650411123314e-05, 'epoch': 14.33}) +[2026-06-05 14:18:16,453][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 7.450579886381092e-08, 'train/grad_norm_pre_clip': 7.851138548176095e-07, 'timing/data': 0.00854218378663063, 'timing/model': 2.672380603849888, 'learning_rate/qwen_vl_interface': 7.46166897154428e-06, 'learning_rate/action_model': 7.46166897154428e-05, 'epoch': 14.42}) +[2026-06-05 14:24:57,252][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 4.749400886794319e-06, 'train/grad_norm_pre_clip': 0.0001317855203524232, 'timing/data': 0.005063522607088089, 'timing/model': 2.678420787677169, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 14.5}) +[2026-06-05 14:31:37,651][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 1.266597990934315e-07, 'train/grad_norm_pre_clip': 2.14406281884294e-06, 'timing/data': 0.010988574475049973, 'timing/model': 2.6675727628171444, 'learning_rate/qwen_vl_interface': 7.350871278540257e-06, 'learning_rate/action_model': 7.350871278540256e-05, 'epoch': 14.58}) +[2026-06-05 14:37:56,789][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 2.554484623829012e-08, 'train/grad_norm_pre_clip': 3.884254340391635e-07, 'timing/data': 0.17185052298009396, 'timing/model': 2.446252830326557, 'learning_rate/qwen_vl_interface': 7.2949277098394036e-06, 'learning_rate/action_model': 7.294927709839403e-05, 'epoch': 14.67}) +[2026-06-05 14:44:41,256][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 8.568165554834195e-08, 'train/grad_norm_pre_clip': 1.3672902241523843e-06, 'timing/data': 0.020718876272439957, 'timing/model': 2.6658404115587473, 'learning_rate/qwen_vl_interface': 7.238633753302784e-06, 'learning_rate/action_model': 7.238633753302784e-05, 'epoch': 14.75}) +[2026-06-05 14:51:21,552][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 1.3038273891652352e-06, 'train/grad_norm_pre_clip': 4.1743882320588455e-05, 'timing/data': 0.010918144136667252, 'timing/model': 2.6714667044579983, 'learning_rate/qwen_vl_interface': 7.181999054610053e-06, 'learning_rate/action_model': 7.181999054610052e-05, 'epoch': 14.83}) +[2026-06-05 14:58:02,056][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 1.2777503570760018e-06, 'train/grad_norm_pre_clip': 6.500871677417308e-05, 'timing/data': 0.020554646849632263, 'timing/model': 2.6637774743139744, 'learning_rate/qwen_vl_interface': 7.125033317825269e-06, 'learning_rate/action_model': 7.12503331782527e-05, 'epoch': 14.92}) +[2026-06-05 15:04:41,419][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1800 +[2026-06-05 15:04:41,421][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 9.36644326543501e-08, 'train/grad_norm_pre_clip': 1.4293597132564173e-06, 'eval/loss': 0.02651786617934704, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/mean_length': 16.4, 'rl_games_eval/mid_train/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/std_length': 4.586937976471886, 'rl_games_eval/mid_train/macro_mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/macro_mean_length': 16.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 16.4, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 4.586937976471886, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 229.50448608398438, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.1763853970915079, 'timing/model': 2.422991320490837, 'learning_rate/qwen_vl_interface': 7.067746303734178e-06, 'learning_rate/action_model': 7.067746303734177e-05, 'epoch': 15.0}) +[2026-06-05 15:04:41,422][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1800_state +[2026-06-05 15:04:41,426][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-05 15:04:47,253][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1800_state/model.safetensors +[2026-06-05 15:04:55,499][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1800_state/optimizer.bin +[2026-06-05 15:04:55,499][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1800_state/sampler.bin +[2026-06-05 15:04:55,499][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1800_state/sampler_1.bin +[2026-06-05 15:04:55,500][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1800_state/random_states_0.pkl +[2026-06-05 15:04:55,635][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1800 +[2026-06-05 15:04:55,636][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-05 15:04:55,649][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-05 15:11:39,954][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 6.705466262246773e-07, 'train/grad_norm_pre_clip': 3.451956217759289e-05, 'timing/data': 0.006163466721773148, 'timing/model': 2.6745880618691444, 'learning_rate/qwen_vl_interface': 7.0101478281717235e-06, 'learning_rate/action_model': 7.010147828171724e-05, 'epoch': 15.08}) +[2026-06-05 15:18:20,765][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 5.960463056453591e-08, 'train/grad_norm_pre_clip': 1.406138835591264e-06, 'timing/data': 0.01997407339513302, 'timing/model': 2.66266262345016, 'learning_rate/qwen_vl_interface': 6.952247760340176e-06, 'learning_rate/action_model': 6.952247760340176e-05, 'epoch': 15.17}) +[2026-06-05 15:25:02,453][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 5.5879347371501353e-08, 'train/grad_norm_pre_clip': 9.017492743623734e-07, 'timing/data': 0.011232540011405945, 'timing/model': 2.668152092024684, 'learning_rate/qwen_vl_interface': 6.894056021118092e-06, 'learning_rate/action_model': 6.894056021118091e-05, 'epoch': 15.25}) +[2026-06-05 15:31:22,173][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 3.4059794984386826e-08, 'train/grad_norm_pre_clip': 8.793206234258832e-07, 'timing/data': 0.16776391491293907, 'timing/model': 2.5381078720092773, 'learning_rate/qwen_vl_interface': 6.835582581360413e-06, 'learning_rate/action_model': 6.835582581360412e-05, 'epoch': 15.33}) +[2026-06-05 15:38:05,553][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 1.86264500712241e-07, 'train/grad_norm_pre_clip': 1.9495357719279127e-06, 'timing/data': 0.005972512066364288, 'timing/model': 2.678276352584362, 'learning_rate/qwen_vl_interface': 6.7768374601900174e-06, 'learning_rate/action_model': 6.776837460190018e-05, 'epoch': 15.42}) +[2026-06-05 15:44:46,158][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.000680137425661087, 'train/grad_norm_pre_clip': 0.07842494547367096, 'timing/data': 0.0061330366879701614, 'timing/model': 2.68247171677649, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 15.5}) +[2026-06-05 15:51:26,575][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.000539704691618681, 'train/grad_norm_pre_clip': 0.06113067641854286, 'timing/data': 0.011092180386185646, 'timing/model': 2.8937218748033047, 'learning_rate/qwen_vl_interface': 6.658572481133962e-06, 'learning_rate/action_model': 6.658572481133962e-05, 'epoch': 15.58}) +[2026-06-05 15:57:46,369][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.00026073260232806206, 'train/grad_norm_pre_clip': 0.037673044949769974, 'timing/data': 0.1801952663809061, 'timing/model': 2.5188936423510313, 'learning_rate/qwen_vl_interface': 6.599072887343634e-06, 'learning_rate/action_model': 6.599072887343633e-05, 'epoch': 15.67}) +[2026-06-05 16:04:29,554][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 1.2446077562344726e-05, 'train/grad_norm_pre_clip': 0.00013337205746211112, 'timing/data': 0.011688480153679848, 'timing/model': 2.6716664414852858, 'learning_rate/qwen_vl_interface': 6.539342136859104e-06, 'learning_rate/action_model': 6.539342136859104e-05, 'epoch': 15.75}) +[2026-06-05 16:11:31,804][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1900 +[2026-06-05 16:11:31,805][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 1.4826416190771852e-05, 'train/grad_norm_pre_clip': 0.00016501644859090447, 'eval/loss': 0.013590858317911625, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 240.38121643066407, 'rl_games_eval/mid_train/mean_length': 17.0, 'rl_games_eval/mid_train/std_reward': 169.60664657811128, 'rl_games_eval/mid_train/std_length': 5.403702434442518, 'rl_games_eval/mid_train/macro_mean_reward': 240.38121643066407, 'rl_games_eval/mid_train/macro_mean_length': 17.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 240.38121643066407, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 17.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 169.60664657811128, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 5.403702434442518, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 240.38121643066407, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0202934630215168, 'timing/model': 2.668246742337942, 'learning_rate/qwen_vl_interface': 6.479390464236974e-06, 'learning_rate/action_model': 6.479390464236974e-05, 'epoch': 15.83}) +[2026-06-05 16:11:31,806][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1900_state +[2026-06-05 16:11:31,811][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-05 16:11:37,480][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1900_state/model.safetensors +[2026-06-05 16:11:46,031][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1900_state/optimizer.bin +[2026-06-05 16:11:46,032][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1900_state/sampler.bin +[2026-06-05 16:11:46,032][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1900_state/sampler_1.bin +[2026-06-05 16:11:46,033][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_1900_state/random_states_0.pkl +[2026-06-05 16:11:46,162][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1900 +[2026-06-05 16:11:46,163][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-05 16:11:46,176][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-05 16:18:26,852][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 2.7034340746467933e-05, 'train/grad_norm_pre_clip': 0.0005480364197865129, 'timing/data': 0.0059356167912483215, 'timing/model': 2.6743195094168186, 'learning_rate/qwen_vl_interface': 6.4192281418877174e-06, 'learning_rate/action_model': 6.419228141887717e-05, 'epoch': 15.92}) +[2026-06-05 16:24:48,976][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 7.433493465214269e-06, 'train/grad_norm_pre_clip': 7.718807319179177e-05, 'timing/data': 0.4196237828582525, 'timing/model': 2.6864118482917547, 'learning_rate/qwen_vl_interface': 6.3588654783155525e-06, 'learning_rate/action_model': 6.358865478315553e-05, 'epoch': 16.0}) +[2026-06-05 16:31:33,054][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 2.0796731405425817e-05, 'train/grad_norm_pre_clip': 0.0005137529224157333, 'timing/data': 0.006103731691837311, 'timing/model': 2.6770956814289093, 'learning_rate/qwen_vl_interface': 6.298312816352132e-06, 'learning_rate/action_model': 6.298312816352132e-05, 'epoch': 16.08}) +[2026-06-05 16:38:13,154][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 2.5326726245111786e-05, 'train/grad_norm_pre_clip': 0.0002973996161017567, 'timing/data': 0.010704010725021362, 'timing/model': 2.6694222912192345, 'learning_rate/qwen_vl_interface': 6.2375805313843564e-06, 'learning_rate/action_model': 6.237580531384356e-05, 'epoch': 16.17}) +[2026-06-05 16:44:53,258][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 1.810879257391207e-05, 'train/grad_norm_pre_clip': 0.0006624320521950722, 'timing/data': 0.00610785186290741, 'timing/model': 2.677175808697939, 'learning_rate/qwen_vl_interface': 6.176679029576611e-06, 'learning_rate/action_model': 6.17667902957661e-05, 'epoch': 16.25}) +[2026-06-05 16:51:13,784][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 1.2648811207327526e-05, 'train/grad_norm_pre_clip': 0.00805524829775095, 'timing/data': 0.17348616383969784, 'timing/model': 2.5373476464301348, 'learning_rate/qwen_vl_interface': 6.115618746087702e-06, 'learning_rate/action_model': 6.115618746087703e-05, 'epoch': 16.33}) +[2026-06-05 16:57:57,558][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 6.959036545595154e-05, 'train/grad_norm_pre_clip': 0.0019777189008891582, 'timing/data': 0.020502643659710884, 'timing/model': 2.6649402882903814, 'learning_rate/qwen_vl_interface': 6.0544101432828676e-06, 'learning_rate/action_model': 6.054410143282867e-05, 'epoch': 16.42}) +[2026-06-05 17:04:38,353][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 5.511888593900949e-05, 'train/grad_norm_pre_clip': 0.002019727136939764, 'timing/data': 0.015742095187306404, 'timing/model': 2.666809080168605, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 16.5}) +[2026-06-05 17:11:18,752][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 3.0073464586166665e-05, 'train/grad_norm_pre_clip': 0.0007499186904169619, 'timing/data': 0.006184078752994537, 'timing/model': 2.6764952670782804, 'learning_rate/qwen_vl_interface': 5.9315899544581015e-06, 'learning_rate/action_model': 5.931589954458101e-05, 'epoch': 16.58}) +[2026-06-05 17:17:58,302][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2000 +[2026-06-05 17:17:58,304][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 1.6478728866786696e-05, 'train/grad_norm_pre_clip': 0.0005660643219016492, 'eval/loss': 0.025929072871804237, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/mean_length': 16.4, 'rl_games_eval/mid_train/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/std_length': 4.586937976471886, 'rl_games_eval/mid_train/macro_mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/macro_mean_length': 16.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 16.4, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 4.586937976471886, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 229.50448608398438, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.17208650708198547, 'timing/model': 2.446228502318263, 'learning_rate/qwen_vl_interface': 5.869999413045247e-06, 'learning_rate/action_model': 5.869999413045246e-05, 'epoch': 16.67}) +[2026-06-05 17:17:58,304][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2000_state +[2026-06-05 17:17:58,309][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-05 17:18:04,175][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2000_state/model.safetensors +[2026-06-05 17:18:12,542][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2000_state/optimizer.bin +[2026-06-05 17:18:12,543][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2000_state/sampler.bin +[2026-06-05 17:18:12,543][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-05 17:18:12,544][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-05 17:18:12,678][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-05 17:18:12,678][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-05 17:18:12,691][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-05 17:24:57,262][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 1.6949969676716137e-06, 'train/grad_norm_pre_clip': 3.7302896089386195e-05, 'timing/data': 0.011330267414450645, 'timing/model': 2.6600908748805523, 'learning_rate/qwen_vl_interface': 5.80830263792473e-06, 'learning_rate/action_model': 5.80830263792473e-05, 'epoch': 16.75}) +[2026-06-05 17:31:37,455][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 2.1569169348367723e-06, 'train/grad_norm_pre_clip': 4.700512363342568e-05, 'timing/data': 0.020947350189089775, 'timing/model': 2.659292098134756, 'learning_rate/qwen_vl_interface': 5.746510200521354e-06, 'learning_rate/action_model': 5.746510200521355e-05, 'epoch': 16.83}) +[2026-06-05 17:38:17,754][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 5.327161147761217e-07, 'train/grad_norm_pre_clip': 8.831471859593876e-06, 'timing/data': 0.00611606240272522, 'timing/model': 2.6762972958385944, 'learning_rate/qwen_vl_interface': 5.684632688651163e-06, 'learning_rate/action_model': 5.6846326886511624e-05, 'epoch': 16.92}) +[2026-06-05 17:44:37,668][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 3.6613319025491364e-06, 'train/grad_norm_pre_clip': 0.00012407074973452836, 'timing/data': 0.21472101099789143, 'timing/model': 2.4845434185117483, 'learning_rate/qwen_vl_interface': 5.622680704707265e-06, 'learning_rate/action_model': 5.622680704707265e-05, 'epoch': 17.0}) +[2026-06-05 17:51:21,655][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 1.0393522416052292e-06, 'train/grad_norm_pre_clip': 2.3605216483701952e-05, 'timing/data': 0.006120171397924423, 'timing/model': 2.675952162593603, 'learning_rate/qwen_vl_interface': 5.560664863843181e-06, 'learning_rate/action_model': 5.56066486384318e-05, 'epoch': 17.08}) +[2026-06-05 17:58:01,868][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 1.1212966910534305e-06, 'train/grad_norm_pre_clip': 6.565157673321664e-05, 'timing/data': 0.021423380821943283, 'timing/model': 2.677054649218917, 'learning_rate/qwen_vl_interface': 5.498595792153985e-06, 'learning_rate/action_model': 5.498595792153984e-05, 'epoch': 17.17}) +[2026-06-05 18:04:42,856][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 8.568166265376931e-08, 'train/grad_norm_pre_clip': 2.2364463347912533e-06, 'timing/data': 0.010778035968542099, 'timing/model': 2.674732558429241, 'learning_rate/qwen_vl_interface': 5.436484124855577e-06, 'learning_rate/action_model': 5.4364841248555764e-05, 'epoch': 17.25}) +[2026-06-05 18:11:02,665][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 2.8183428639749764e-06, 'train/grad_norm_pre_clip': 0.00014617675333283842, 'timing/data': 0.18112102337181568, 'timing/model': 2.4069973677396774, 'learning_rate/qwen_vl_interface': 5.374340504462399e-06, 'learning_rate/action_model': 5.3743405044623984e-05, 'epoch': 17.33}) +[2026-06-05 18:17:46,454][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 1.1548394240890048e-07, 'train/grad_norm_pre_clip': 2.4799003313091816e-06, 'timing/data': 0.010736368596553802, 'timing/model': 2.716992711648345, 'learning_rate/qwen_vl_interface': 5.312175578963886e-06, 'learning_rate/action_model': 5.312175578963886e-05, 'epoch': 17.42}) +[2026-06-05 18:24:48,089][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2100 +[2026-06-05 18:24:48,091][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 7.450577754752885e-08, 'train/grad_norm_pre_clip': 1.952238790181582e-06, 'eval/loss': 0.022868815809488297, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/mean_length': 16.4, 'rl_games_eval/mid_train/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/std_length': 4.586937976471886, 'rl_games_eval/mid_train/macro_mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/macro_mean_length': 16.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 16.4, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 4.586937976471886, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 229.50448608398438, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.006162278354167938, 'timing/model': 2.6837804168462753, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 17.5}) +[2026-06-05 18:24:48,091][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2100_state +[2026-06-05 18:24:48,096][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-05 18:24:53,792][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2100_state/model.safetensors +[2026-06-05 18:25:01,752][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2100_state/optimizer.bin +[2026-06-05 18:25:01,752][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2100_state/sampler.bin +[2026-06-05 18:25:01,752][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2100_state/sampler_1.bin +[2026-06-05 18:25:01,754][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2100_state/random_states_0.pkl +[2026-06-05 18:25:01,887][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2100 +[2026-06-05 18:25:01,888][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-05 18:25:01,901][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-05 18:31:42,449][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 7.785808406879369e-07, 'train/grad_norm_pre_clip': 2.388624670857098e-05, 'timing/data': 0.005976732820272446, 'timing/model': 2.67625779658556, 'learning_rate/qwen_vl_interface': 5.187824421036115e-06, 'learning_rate/action_model': 5.187824421036115e-05, 'epoch': 17.58}) +[2026-06-05 18:38:02,053][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 3.107944337443769e-07, 'train/grad_norm_pre_clip': 7.289659151865635e-06, 'timing/data': 0.26045023649930954, 'timing/model': 2.526032716035843, 'learning_rate/qwen_vl_interface': 5.125659495537603e-06, 'learning_rate/action_model': 5.1256594955376034e-05, 'epoch': 17.67}) +[2026-06-05 18:44:46,751][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 1.862644971595273e-08, 'train/grad_norm_pre_clip': 3.882058194903948e-07, 'timing/data': 0.016474444419145584, 'timing/model': 2.662050174549222, 'learning_rate/qwen_vl_interface': 5.063515875144426e-06, 'learning_rate/action_model': 5.0635158751444254e-05, 'epoch': 17.75}) +[2026-06-05 18:51:27,953][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 3.725290076417309e-09, 'train/grad_norm_pre_clip': 5.331814350029163e-07, 'timing/data': 0.01088695414364338, 'timing/model': 2.6707393769174814, 'learning_rate/qwen_vl_interface': 5.001404207846017e-06, 'learning_rate/action_model': 5.001404207846017e-05, 'epoch': 17.83}) +[2026-06-05 18:58:08,655][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 2.4585949631728e-06, 'train/grad_norm_pre_clip': 0.00014562945580109954, 'timing/data': 0.02360527589917183, 'timing/model': 2.659739648923278, 'learning_rate/qwen_vl_interface': 4.93933513615682e-06, 'learning_rate/action_model': 4.9393351361568204e-05, 'epoch': 17.92}) +[2026-06-05 19:04:28,173][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 1.5229999007715378e-05, 'train/grad_norm_pre_clip': 0.002939406782388687, 'timing/data': 0.18680236488580704, 'timing/model': 2.4153650868684053, 'learning_rate/qwen_vl_interface': 4.877319295292737e-06, 'learning_rate/action_model': 4.8773192952927374e-05, 'epoch': 18.0}) +[2026-06-05 19:11:12,552][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 3.1776966835650455e-08, 'timing/data': 0.010948365554213524, 'timing/model': 2.673599561676383, 'learning_rate/qwen_vl_interface': 4.815367311348839e-06, 'learning_rate/action_model': 4.815367311348839e-05, 'epoch': 18.08}) +[2026-06-05 19:17:53,955][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 1.2593523557313802e-08, 'timing/data': 0.01119985617697239, 'timing/model': 2.670818902552128, 'learning_rate/qwen_vl_interface': 4.753489799478647e-06, 'learning_rate/action_model': 4.7534897994786464e-05, 'epoch': 18.17}) +[2026-06-05 19:24:34,485][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 3.725290076417309e-09, 'train/grad_norm_pre_clip': 8.178538024594673e-08, 'timing/data': 0.0072371698915958405, 'timing/model': 2.9015391655266285, 'learning_rate/qwen_vl_interface': 4.691697362075272e-06, 'learning_rate/action_model': 4.691697362075272e-05, 'epoch': 18.25}) +[2026-06-05 19:31:15,140][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2200 +[2026-06-05 19:31:15,141][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 1.7029895715836574e-08, 'train/grad_norm_pre_clip': 1.2487992080423282e-06, 'eval/loss': 0.026672283187508583, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/mean_length': 16.4, 'rl_games_eval/mid_train/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/std_length': 4.586937976471886, 'rl_games_eval/mid_train/macro_mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/macro_mean_length': 16.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 16.4, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 4.586937976471886, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 229.50448608398438, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.17228922247886658, 'timing/model': 2.436519404873252, 'learning_rate/qwen_vl_interface': 4.630000586954756e-06, 'learning_rate/action_model': 4.630000586954756e-05, 'epoch': 18.33}) +[2026-06-05 19:31:15,142][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2200_state +[2026-06-05 19:31:15,146][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-05 19:31:21,063][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2200_state/model.safetensors +[2026-06-05 19:31:29,387][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2200_state/optimizer.bin +[2026-06-05 19:31:29,388][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2200_state/sampler.bin +[2026-06-05 19:31:29,388][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2200_state/sampler_1.bin +[2026-06-05 19:31:29,389][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2200_state/random_states_0.pkl +[2026-06-05 19:31:29,508][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2200 +[2026-06-05 19:31:29,509][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-05 19:31:29,522][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-05 19:38:14,955][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 3.725290076417309e-09, 'train/grad_norm_pre_clip': 1.0369089409323351e-07, 'timing/data': 0.006410758942365646, 'timing/model': 2.6777978762984276, 'learning_rate/qwen_vl_interface': 4.568410045541899e-06, 'learning_rate/action_model': 4.5684100455418986e-05, 'epoch': 18.42}) +[2026-06-05 19:44:55,663][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 1.8180321248451037e-08, 'timing/data': 0.010459281504154205, 'timing/model': 2.6734490655362606, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 18.5}) +[2026-06-05 19:51:37,361][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 7.695851067524018e-09, 'timing/data': 0.020306849852204323, 'timing/model': 2.6664657089859247, 'learning_rate/qwen_vl_interface': 4.4455898567171336e-06, 'learning_rate/action_model': 4.4455898567171334e-05, 'epoch': 18.58}) +[2026-06-05 19:57:56,679][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 1.616961142758555e-08, 'timing/data': 0.1749857421964407, 'timing/model': 2.526294484734535, 'learning_rate/qwen_vl_interface': 4.384381253912299e-06, 'learning_rate/action_model': 4.3843812539122994e-05, 'epoch': 18.67}) +[2026-06-05 20:04:40,960][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 6.800456731070881e-08, 'timing/data': 0.011610645800828934, 'timing/model': 2.6725863702595234, 'learning_rate/qwen_vl_interface': 4.323320970423391e-06, 'learning_rate/action_model': 4.3233209704233915e-05, 'epoch': 18.75}) +[2026-06-05 20:11:21,464][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 1.9743592929444276e-06, 'train/grad_norm_pre_clip': 0.00023489347950089723, 'timing/data': 0.011506369337439537, 'timing/model': 2.6786262951791286, 'learning_rate/qwen_vl_interface': 4.262419468615644e-06, 'learning_rate/action_model': 4.2624194686156435e-05, 'epoch': 18.83}) +[2026-06-05 20:18:01,762][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 8.15889666938574e-09, 'timing/data': 0.01540481485426426, 'timing/model': 2.6708406470716, 'learning_rate/qwen_vl_interface': 4.201687183647871e-06, 'learning_rate/action_model': 4.201687183647871e-05, 'epoch': 18.92}) +[2026-06-05 20:24:21,498][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 3.022918448891687e-08, 'timing/data': 0.17469678074121475, 'timing/model': 2.44704476185143, 'learning_rate/qwen_vl_interface': 4.14113452168445e-06, 'learning_rate/action_model': 4.1411345216844504e-05, 'epoch': 19.0}) +[2026-06-05 20:31:05,759][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 3.725290076417309e-09, 'train/grad_norm_pre_clip': 1.0607444522747755e-07, 'timing/data': 0.01640867255628109, 'timing/model': 2.6633322928100824, 'learning_rate/qwen_vl_interface': 4.080771858112283e-06, 'learning_rate/action_model': 4.0807718581122825e-05, 'epoch': 19.08}) +[2026-06-05 20:38:08,045][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2300 +[2026-06-05 20:38:08,046][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 1.7590531697919687e-08, 'eval/loss': 0.028217988088726997, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/mean_length': 16.4, 'rl_games_eval/mid_train/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/std_length': 4.586937976471886, 'rl_games_eval/mid_train/macro_mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/macro_mean_length': 16.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 16.4, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 4.586937976471886, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 229.50448608398438, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.006061447784304619, 'timing/model': 2.6726807188242674, 'learning_rate/qwen_vl_interface': 4.0206095357630256e-06, 'learning_rate/action_model': 4.020609535763026e-05, 'epoch': 19.17}) +[2026-06-05 20:38:08,046][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2300_state +[2026-06-05 20:38:08,051][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-05 20:38:13,899][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2300_state/model.safetensors +[2026-06-05 20:38:21,928][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2300_state/optimizer.bin +[2026-06-05 20:38:21,929][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2300_state/sampler.bin +[2026-06-05 20:38:21,929][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2300_state/sampler_1.bin +[2026-06-05 20:38:21,930][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2300_state/random_states_0.pkl +[2026-06-05 20:38:22,074][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2300 +[2026-06-05 20:38:22,074][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-05 20:38:22,089][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-05 20:45:03,263][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 1.7814667074844692e-08, 'timing/data': 0.006169069558382034, 'timing/model': 2.67391899228096, 'learning_rate/qwen_vl_interface': 3.960657863140897e-06, 'learning_rate/action_model': 3.9606578631408975e-05, 'epoch': 19.25}) +[2026-06-05 20:51:22,884][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 6.377627315146128e-09, 'timing/data': 0.168566120788455, 'timing/model': 2.4392979610711336, 'learning_rate/qwen_vl_interface': 3.900927112656368e-06, 'learning_rate/action_model': 3.9009271126563677e-05, 'epoch': 19.33}) +[2026-06-05 20:58:06,267][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 2.312203584153849e-08, 'timing/data': 0.012205326929688454, 'timing/model': 2.6770431250333786, 'learning_rate/qwen_vl_interface': 3.841427518866038e-06, 'learning_rate/action_model': 3.841427518866038e-05, 'epoch': 19.42}) +[2026-06-05 21:04:47,569][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 9.31321153530007e-08, 'train/grad_norm_pre_clip': 6.47321212454699e-06, 'timing/data': 0.021146943792700768, 'timing/model': 2.708604531362653, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 19.5}) +[2026-06-05 21:11:28,666][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 7.823099679171719e-08, 'train/grad_norm_pre_clip': 4.6501968427037355e-06, 'timing/data': 0.006418989971280098, 'timing/model': 2.6738311033695936, 'learning_rate/qwen_vl_interface': 3.7231625398099837e-06, 'learning_rate/action_model': 3.723162539809983e-05, 'epoch': 19.58}) +[2026-06-05 21:17:48,299][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 8.940686058167557e-08, 'train/grad_norm_pre_clip': 5.589821284956997e-06, 'timing/data': 0.1797146126627922, 'timing/model': 2.4281364176422358, 'learning_rate/qwen_vl_interface': 3.6644174186395885e-06, 'learning_rate/action_model': 3.664417418639589e-05, 'epoch': 19.67}) +[2026-06-05 21:24:31,965][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 1.8634251475191377e-08, 'timing/data': 0.00596013106405735, 'timing/model': 2.6831398997455835, 'learning_rate/qwen_vl_interface': 3.605943978881907e-06, 'learning_rate/action_model': 3.605943978881907e-05, 'epoch': 19.75}) +[2026-06-05 21:31:11,972][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 4.097817907222634e-08, 'train/grad_norm_pre_clip': 1.552606477162044e-06, 'timing/data': 0.015465334057807922, 'timing/model': 2.671922156587243, 'learning_rate/qwen_vl_interface': 3.5477522396598236e-06, 'learning_rate/action_model': 3.5477522396598235e-05, 'epoch': 19.83}) +[2026-06-05 21:37:52,264][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 3.725290076417309e-09, 'train/grad_norm_pre_clip': 1.9619432123363367e-07, 'timing/data': 0.0005397237837314606, 'timing/model': 2.678104132413864, 'learning_rate/qwen_vl_interface': 3.489852171828278e-06, 'learning_rate/action_model': 3.4898521718282776e-05, 'epoch': 19.92}) +[2026-06-05 21:44:32,501][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2400 +[2026-06-05 21:44:32,502][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 2.55448409092196e-08, 'train/grad_norm_pre_clip': 9.185567364511371e-07, 'eval/loss': 0.027991637587547302, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/mean_length': 16.4, 'rl_games_eval/mid_train/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/std_length': 4.586937976471886, 'rl_games_eval/mid_train/macro_mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/macro_mean_length': 16.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 16.4, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 4.586937976471886, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 229.50448608398438, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.15470336005091667, 'timing/model': 2.4682860411703587, 'learning_rate/qwen_vl_interface': 3.432253696265824e-06, 'learning_rate/action_model': 3.432253696265824e-05, 'epoch': 20.0}) +[2026-06-05 21:44:32,503][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2400_state +[2026-06-05 21:44:32,508][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-05 21:44:37,868][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2400_state/model.safetensors +[2026-06-05 21:44:45,929][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2400_state/optimizer.bin +[2026-06-05 21:44:45,929][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2400_state/sampler.bin +[2026-06-05 21:44:45,929][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2400_state/sampler_1.bin +[2026-06-05 21:44:45,931][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2400_state/random_states_0.pkl +[2026-06-05 21:44:46,055][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2400 +[2026-06-05 21:44:46,056][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-05 21:44:46,069][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-05 21:51:29,951][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 6.453274359330408e-09, 'timing/data': 0.015728356316685677, 'timing/model': 2.6652890760451555, 'learning_rate/qwen_vl_interface': 3.3749666821747314e-06, 'learning_rate/action_model': 3.374966682174731e-05, 'epoch': 20.08}) +[2026-06-05 21:58:10,954][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 2.2834909074731513e-08, 'timing/data': 0.016948038712143898, 'timing/model': 2.6626253444701433, 'learning_rate/qwen_vl_interface': 3.3180009453899496e-06, 'learning_rate/action_model': 3.318000945389949e-05, 'epoch': 20.17}) +[2026-06-05 22:04:50,857][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 2.6924722007493074e-08, 'timing/data': 0.011440236121416092, 'timing/model': 2.6712744869291782, 'learning_rate/qwen_vl_interface': 3.2613662466972178e-06, 'learning_rate/action_model': 3.2613662466972176e-05, 'epoch': 20.25}) +[2026-06-05 22:11:10,981][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 4.257474373048353e-09, 'train/grad_norm_pre_clip': 2.1167879538097623e-07, 'timing/data': 0.24809850007295609, 'timing/model': 2.561919467523694, 'learning_rate/qwen_vl_interface': 3.2050722901605967e-06, 'learning_rate/action_model': 3.205072290160597e-05, 'epoch': 20.33}) +[2026-06-05 22:17:55,658][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 4.1682128859577006e-09, 'timing/data': 0.021700257435441017, 'timing/model': 2.6639320347458124, 'learning_rate/qwen_vl_interface': 3.149128721459745e-06, 'learning_rate/action_model': 3.149128721459745e-05, 'epoch': 20.42}) +[2026-06-05 22:24:35,658][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 7.450572780953735e-08, 'train/grad_norm_pre_clip': 1.1008918590960093e-05, 'timing/data': 0.006301281973719597, 'timing/model': 2.673679795116186, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 20.5}) +[2026-06-05 22:31:15,957][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 3.501757532831107e-07, 'train/grad_norm_pre_clip': 4.368673398857936e-05, 'timing/data': 0.005880486220121384, 'timing/model': 2.679565228521824, 'learning_rate/qwen_vl_interface': 3.038331028455721e-06, 'learning_rate/action_model': 3.038331028455721e-05, 'epoch': 20.58}) +[2026-06-05 22:37:35,234][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 3.8317253370223625e-08, 'train/grad_norm_pre_clip': 4.4279377107159235e-06, 'timing/data': 0.17606222815811634, 'timing/model': 2.4854067470878363, 'learning_rate/qwen_vl_interface': 2.98349588876686e-06, 'learning_rate/action_model': 2.9834958887668597e-05, 'epoch': 20.67}) +[2026-06-05 22:44:19,253][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 5.215402154590265e-08, 'train/grad_norm_pre_clip': 3.882351393258432e-06, 'timing/data': 0.011524936184287071, 'timing/model': 2.671161588281393, 'learning_rate/qwen_vl_interface': 2.9290491028894658e-06, 'learning_rate/action_model': 2.9290491028894658e-05, 'epoch': 20.75}) +[2026-06-05 22:51:22,143][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2500 +[2026-06-05 22:51:22,145][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.0, 'train/grad_norm_pre_clip': 2.3382995095744263e-08, 'eval/loss': 0.028556033968925476, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/mean_length': 16.4, 'rl_games_eval/mid_train/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/std_length': 4.586937976471886, 'rl_games_eval/mid_train/macro_mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/macro_mean_length': 16.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_reward': 229.50448608398438, 'rl_games_eval/mid_train/deadly_corridor__latency_0/mean_length': 16.4, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_reward': 161.98738844236323, 'rl_games_eval/mid_train/deadly_corridor__latency_0/std_length': 4.586937976471886, 'rl_games_eval/mid_train/deadly_corridor__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 229.50448608398438, 'checkpoint/best_score': 317.74525146484376, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.020181354135274887, 'timing/model': 2.6621071323752403, 'learning_rate/qwen_vl_interface': 2.8750000000000013e-06, 'learning_rate/action_model': 2.875000000000001e-05, 'epoch': 20.83}) +[2026-06-05 22:51:22,145][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2500_state +[2026-06-05 22:51:22,150][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-05 22:51:28,101][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2500_state/model.safetensors +[2026-06-05 22:51:36,334][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2500_state/optimizer.bin +[2026-06-05 22:51:36,335][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2500_state/sampler.bin +[2026-06-05 22:51:36,335][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-05 22:51:36,336][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-05 22:51:36,472][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-05 22:51:36,473][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-05 22:51:36,487][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-05 22:51:36,487][starVLA.training.train_starvla][INFO] - Training complete. Best checkpoint step=800 score=317.74525146484376 path=/workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state +[2026-06-05 22:51:36,487][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/openvla_deadly_corridor_fix_latency_0/checkpoints/best_state +[2026-06-05 22:51:37,740][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-05 22:51:49,260][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-05 22:51:49,265][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-05 22:51:49,266][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-05 22:51:49,268][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-05 22:51:50,297][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-05 22:52:29,856][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2500 +[2026-06-05 22:52:31,635][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/summary.jsonl b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..5dca6b320b71d2e0af379946808c5c50561446d4 --- /dev/null +++ b/zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/summary.jsonl @@ -0,0 +1,17 @@ +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500} +{"steps": 1600} +{"steps": 1700} +{"steps": 1800} +{"steps": 1900} +{"steps": 2000} +{"steps": 2100} +{"steps": 2200} +{"steps": 2300} +{"steps": 2400} +{"steps": 2500} diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/checkpoints/best_model.safetensors b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/checkpoints/best_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5160dd89b11ede394554f85d8056757694a1df44 --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/checkpoints/best_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0eda74a2340c7b2df580ca0589505c47ba32f7244b54a15878f61f86fa922213 +size 9138230516 diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/config.full.yaml b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..c494408472b7051d473357500a263f2d6aed2cac --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/config.full.yaml @@ -0,0 +1,181 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 6 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/demon_attack_fix_latency_0 + data_mix: demon_attack_train__bridge + eval_data_mix: demon_attack_train__bridge__val + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + converted_name: demon_attack_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: demon_attack_train + mixed_converted_name: demon_attack_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 2500 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: true + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state + resume_step: 800 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + enabled: true + latency: + prompt_map_path: /workspace/starVLA/data/demon_attack_fix_latency_0/demon_attack_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: demon_attack + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: demon_attack +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: true + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_demon_attack_fix_latency_0 +output_dir: /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0 +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/config.yaml b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..865b058c41ae6469d5aa69225dc6c800432eb552 --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/config.yaml @@ -0,0 +1,84 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: true + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: demon_attack_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: demon_attack_train__bridge__val + include_state: true + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 6 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0 +rl_games: + env_eval: + enabled: true + fixed_episode_seeds: true + frameskip: 4 + image_size: 224 + latency: + prompt_map_path: /workspace/starVLA/data/demon_attack_fix_latency_0/demon_attack_train__bridge/latency_prompt_map.json + latency_seed_stride: 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 0 + max_steps_per_episode: 3600 + num_episodes: 5 + seed: 42 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + task_seed_stride: 0 + model_alias: openvla + task: demon_attack +run_id: openvla_demon_attack_fix_latency_0 +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + eval_interval: 100 + eval_num_batches: 20 + freeze_modules: '' + gradient_clipping: 1.0 + is_resume: true + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 10 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 2500 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state + resume_step: 800 + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/dataset_statistics.json b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..8b348a05ae43d77e6dba683950152ae87c2e2335 --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.017027778550982475, + 0.020826388150453568, + 0.21239235997200012, + 0.20501388609409332, + 0.277673602104187, + 0.26706597208976746, + 0.0 + ], + "std": [ + 0.12938925623893738, + 0.14295804500579834, + 0.4095483720302582, + 0.4044252932071686, + 0.448200523853302, + 0.44252121448516846, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 288000, + "num_trajectories": 80 + } +} \ No newline at end of file diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/dataset_statistics_eval.json b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..93e6aa6ff22972b62d7abe5ff87f1dd2d088e644 --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.017374999821186066, + 0.02080555632710457, + 0.20629166066646576, + 0.20165278017520905, + 0.28373610973358154, + 0.2701388895511627, + 0.0 + ], + "std": [ + 0.1306414157152176, + 0.14277756214141846, + 0.4047096073627472, + 0.40132102370262146, + 0.4506453573703766, + 0.4440159499645233, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 72000, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/hydra/.hydra/config.yaml b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..1df340028004a5ee8ef7e50195c1a35b81faedd3 --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/hydra/.hydra/config.yaml @@ -0,0 +1,179 @@ +framework: + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 6 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/demon_attack_fix_latency_0 + data_mix: demon_attack_train__bridge + eval_data_mix: demon_attack_train__bridge__val + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: demon_attack_train + mixed_converted_name: demon_attack_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 2500 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: true + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state + resume_step: 800 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + enabled: true + latency: + prompt_map_path: /workspace/starVLA/data/demon_attack_fix_latency_0/demon_attack_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: demon_attack + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: demon_attack +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: true + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_demon_attack_fix_latency_0 +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/hydra/.hydra/hydra.yaml b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..8fa8a05f845477e43e8906bc9997ad44640d3a7f --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/hydra/.hydra/hydra.yaml @@ -0,0 +1,248 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=demon_attack + - init=bridge + - mode=single + - run_id=openvla_demon_attack_fix_latency_0 + - run_root_dir=/workspace/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=true + - trainer.is_resume=true + - trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state + - trainer.resume_step=800 + - trainer.max_train_steps=2500 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=20 + - trainer.logging_frequency=10 + - trainer.gradient_accumulation_steps=16 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9, 0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=6 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/starVLA/data/demon_attack_fix_latency_0 + - datasets.vla_data.data_mix=demon_attack_train__bridge + - datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=demon_attack + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=4 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - 'rl_games.env_eval.task_description=''You are playing Demon Attack from a single + game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, + LEFTFIRE.''' + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/demon_attack_fix_latency_0/demon_attack_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=true + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[0] + - rl_games.env_eval.post_train.enabled=true + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: 'checkpoint.local.keep_last_n=1,checkpoint.save_best_model=true,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=demon_attack_train__bridge,datasets.vla_data.data_root_dir=/workspace/starVLA/data/demon_attack_fix_latency_0,datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.video_backend=torchvision_av,env=demon_attack,framework.action_model.action_dim=7,framework.action_model.action_env_dim=6,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=4,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/demon_attack_fix_latency_0/demon_attack_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=true,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[0],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=true,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_description=''You + are playing Demon Attack from a single game image. Choose exactly one action + from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.'',rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=demon_attack,run_id=openvla_demon_attack_fix_latency_0,run_root_dir=/workspace/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=20,trainer.gradient_accumulation_steps=16,trainer.gradient_clipping=1.0,trainer.is_resume=true,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=10,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=2500,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9, + 0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state,trainer.resume_step=800,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games' + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.2 + version_base: '1.1' + cwd: /workspace/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/hydra + choices: + checkpoint: default + mode: single + init: bridge + env: demon_attack + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/hydra/.hydra/overrides.yaml b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..425149e8df42ef617daeff0fb0bbabd2cd0e327c --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/hydra/.hydra/overrides.yaml @@ -0,0 +1,86 @@ +- model=openvla +- env=demon_attack +- init=bridge +- mode=single +- run_id=openvla_demon_attack_fix_latency_0 +- run_root_dir=/workspace/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=true +- trainer.is_resume=true +- trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state +- trainer.resume_step=800 +- trainer.max_train_steps=2500 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=20 +- trainer.logging_frequency=10 +- trainer.gradient_accumulation_steps=16 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9, 0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=6 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/starVLA/data/demon_attack_fix_latency_0 +- datasets.vla_data.data_mix=demon_attack_train__bridge +- datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=demon_attack +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=4 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- 'rl_games.env_eval.task_description=''You are playing Demon Attack from a single + game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, + LEFTFIRE.''' +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/demon_attack_fix_latency_0/demon_attack_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=true +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[0] +- rl_games.env_eval.post_train.enabled=true +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/hydra/train_starvla_hydra.log b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..ca389ee3171191187c0b038fe369a70c45476e91 --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/openvla_demon_attack_fix_latency_0_small/hydra/train_starvla_hydra.log @@ -0,0 +1,557 @@ +[2026-06-03 12:54:47,005][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-03 12:54:47,048][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-03 12:54:48,421][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `demon_attack_train__bridge` +[2026-06-03 12:54:48,933][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-03 12:54:48,934][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-03 12:54:48,934][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-03 12:54:48,934][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-03 12:54:48,934][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-03 12:54:48,934][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-03 12:54:48,934][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-03 12:54:48,934][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-03 12:54:48,934][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-03 12:54:48,934][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-03 12:54:48,935][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-03 12:54:48,935][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-03 12:54:48,935][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-03 12:54:48,935][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-03 12:54:48,935][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-03 12:54:48,935][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-03 12:54:48,935][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-03 12:54:48,935][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-03 12:54:48,935][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-03 12:54:48,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-03 12:54:48,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-03 12:54:48,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-03 12:54:48,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-03 12:54:48,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-03 12:54:48,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-03 12:54:48,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-03 12:54:48,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-03 12:54:48,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-03 12:54:48,937][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-03 12:54:48,937][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-03 12:54:48,937][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-03 12:54:48,990][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `demon_attack_train__bridge__val` +[2026-06-03 12:54:49,032][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/config.full.yaml +[2026-06-03 12:54:49,041][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/config.yaml +[2026-06-03 12:55:03,336][starVLA.training.train_starvla][INFO] - Loaded pretrained checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt, steps: 0 +[2026-06-03 12:55:41,351][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-03 12:55:41,352][starVLA.training.train_starvla][INFO] - Total optimization steps = 2500 +[2026-06-03 12:55:41,352][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-03 12:55:41,352][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-03 12:55:41,353][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-03 12:56:33,675][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-03 12:56:33,722][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-03 12:56:34,735][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `demon_attack_train__bridge` +[2026-06-03 12:56:35,276][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-03 12:56:35,276][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-03 12:56:35,276][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-03 12:56:35,276][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-03 12:56:35,276][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-03 12:56:35,276][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-03 12:56:35,276][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-03 12:56:35,276][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-03 12:56:35,276][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-03 12:56:35,276][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-03 12:56:35,277][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-03 12:56:35,277][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-03 12:56:35,277][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-03 12:56:35,277][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-03 12:56:35,277][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-03 12:56:35,277][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-03 12:56:35,277][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-03 12:56:35,277][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-03 12:56:35,277][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-03 12:56:35,277][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-03 12:56:35,277][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-03 12:56:35,277][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-03 12:56:35,277][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-03 12:56:35,277][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-03 12:56:35,277][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-03 12:56:35,278][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-03 12:56:35,278][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-03 12:56:35,278][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-03 12:56:35,278][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-03 12:56:35,278][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-03 12:56:35,278][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-03 12:56:35,311][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `demon_attack_train__bridge__val` +[2026-06-03 12:56:35,336][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/config.full.yaml +[2026-06-03 12:56:35,342][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/config.yaml +[2026-06-03 12:56:42,981][starVLA.training.train_starvla][INFO] - Loaded pretrained checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt, steps: 0 +[2026-06-03 12:56:47,936][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-03 12:56:47,936][starVLA.training.train_starvla][INFO] - Total optimization steps = 2500 +[2026-06-03 12:56:47,937][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-03 12:56:47,937][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-03 12:56:47,937][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-03 12:58:23,187][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-03 12:58:23,239][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-03 12:58:24,251][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `demon_attack_train__bridge` +[2026-06-03 12:58:24,763][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-03 12:58:24,763][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-03 12:58:24,763][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-03 12:58:24,763][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-03 12:58:24,763][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-03 12:58:24,763][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-03 12:58:24,763][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-03 12:58:24,763][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-03 12:58:24,763][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-03 12:58:24,763][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-03 12:58:24,764][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-03 12:58:24,764][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-03 12:58:24,764][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-03 12:58:24,764][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-03 12:58:24,764][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-03 12:58:24,764][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-03 12:58:24,764][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-03 12:58:24,764][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-03 12:58:24,764][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-03 12:58:24,764][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-03 12:58:24,764][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-03 12:58:24,764][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-03 12:58:24,764][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-03 12:58:24,764][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-03 12:58:24,764][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-03 12:58:24,765][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-03 12:58:24,765][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-03 12:58:24,765][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-03 12:58:24,765][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-03 12:58:24,765][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-03 12:58:24,765][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-03 12:58:24,799][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `demon_attack_train__bridge__val` +[2026-06-03 12:58:24,825][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/config.full.yaml +[2026-06-03 12:58:24,831][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/config.yaml +[2026-06-03 12:58:34,932][starVLA.training.train_starvla][INFO] - Loaded pretrained checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt, steps: 0 +[2026-06-03 12:58:39,757][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-03 12:58:39,758][starVLA.training.train_starvla][INFO] - Total optimization steps = 2500 +[2026-06-03 12:58:39,758][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-03 12:58:39,758][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-03 12:58:39,758][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-03 13:01:41,123][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 1.6203489303588867, 'train/grad_norm_pre_clip': 1.1348522901535034, 'timing/data': 0.013648299500346184, 'timing/model': 1.246702989563346, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-03 13:04:37,614][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 1.4231983423233032, 'train/grad_norm_pre_clip': 0.5956555604934692, 'timing/data': 0.01238151453435421, 'timing/model': 1.0750234555453062, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.0}) +[2026-06-03 13:07:39,791][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 1.499341607093811, 'train/grad_norm_pre_clip': 0.7902237176895142, 'timing/data': 0.011636108160018921, 'timing/model': 1.1311432551592588, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.0}) +[2026-06-03 13:10:30,720][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 1.697331428527832, 'train/grad_norm_pre_clip': 1.3285056352615356, 'timing/data': 0.01156945712864399, 'timing/model': 1.1491743046790361, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.0}) +[2026-06-03 13:13:24,508][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 1.6169366836547852, 'train/grad_norm_pre_clip': 0.8839887976646423, 'timing/data': 0.012654183432459831, 'timing/model': 1.1369996909052134, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.01}) +[2026-06-03 13:16:17,612][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 1.7396575212478638, 'train/grad_norm_pre_clip': 1.579789161682129, 'timing/data': 0.011382648721337318, 'timing/model': 1.1483569853007793, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.01}) +[2026-06-03 13:19:10,895][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 1.3704285621643066, 'train/grad_norm_pre_clip': 3.771559953689575, 'timing/data': 0.012639667838811874, 'timing/model': 1.1329564284533262, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.01}) +[2026-06-03 13:22:05,622][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 1.5517281293869019, 'train/grad_norm_pre_clip': 2.6779420375823975, 'timing/data': 0.011876821517944336, 'timing/model': 1.0676025804132223, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.01}) +[2026-06-03 13:25:03,037][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 1.2647310495376587, 'train/grad_norm_pre_clip': 1.9661788940429688, 'timing/data': 0.010304151102900505, 'timing/model': 1.2765372786670923, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.01}) +[2026-06-03 13:28:12,678][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpeaeui72f/test.c -o /tmp/tmpeaeui72f/test.o +[2026-06-03 13:28:12,705][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpeaeui72f/test.o -laio -o /tmp/tmpeaeui72f/a.out +[2026-06-03 13:28:13,308][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp13m91zwd/test.c -o /tmp/tmp13m91zwd/test.o +[2026-06-03 13:28:13,334][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp13m91zwd/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp13m91zwd/a.out +[2026-06-03 13:28:13,373][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp_s9vw1u4/test.c -o /tmp/tmp_s9vw1u4/test.o +[2026-06-03 13:28:13,395][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp_s9vw1u4/test.o -laio -o /tmp/tmp_s9vw1u4/a.out +[2026-06-03 13:29:57,804][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state +[2026-06-03 13:29:57,809][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-03 13:30:04,065][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/model.safetensors +[2026-06-03 13:30:13,481][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/optimizer.bin +[2026-06-03 13:30:13,481][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/sampler.bin +[2026-06-03 13:30:13,481][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/sampler_1.bin +[2026-06-03 13:30:13,482][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/random_states_0.pkl +[2026-06-03 13:30:13,494][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-03 13:30:13,495][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 100 +[2026-06-03 13:30:13,496][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 1.601261854171753, 'train/grad_norm_pre_clip': 2.138416290283203, 'eval/loss': 1.4959850311279297, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 0.0, 'rl_games_eval/mid_train/mean_length': 254.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 0.0, 'rl_games_eval/mid_train/macro_mean_length': 254.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 254.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 0.0, 'checkpoint/best_score': 0.0, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.010596400126814842, 'timing/model': 1.1605873610824347, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.01}) +[2026-06-03 13:33:30,451][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 1.4711374044418335, 'train/grad_norm_pre_clip': 3.3232359886169434, 'timing/data': 0.012168671935796738, 'timing/model': 1.370777677744627, 'learning_rate/qwen_vl_interface': 9.999593055976534e-06, 'learning_rate/action_model': 9.999593055976533e-05, 'epoch': 0.01}) +[2026-06-03 13:36:42,494][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 1.2553327083587646, 'train/grad_norm_pre_clip': 4.049246788024902, 'timing/data': 0.028106415644288063, 'timing/model': 1.2524027284234762, 'learning_rate/qwen_vl_interface': 9.998372293633897e-06, 'learning_rate/action_model': 9.998372293633897e-05, 'epoch': 0.01}) +[2026-06-03 13:39:55,613][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 1.2774755954742432, 'train/grad_norm_pre_clip': 1.490767002105713, 'timing/data': 0.029554806649684906, 'timing/model': 1.3052258864045143, 'learning_rate/qwen_vl_interface': 9.996337922143435e-06, 'learning_rate/action_model': 9.996337922143436e-05, 'epoch': 0.01}) +[2026-06-03 13:43:11,731][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 1.602773666381836, 'train/grad_norm_pre_clip': 2.724713087081909, 'timing/data': 0.025458596646785736, 'timing/model': 1.260538723319769, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 0.02}) +[2026-06-03 13:46:28,179][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 1.3093475103378296, 'train/grad_norm_pre_clip': 1.517073392868042, 'timing/data': 0.022777196019887924, 'timing/model': 1.229368183761835, 'learning_rate/qwen_vl_interface': 9.989829885383368e-06, 'learning_rate/action_model': 9.989829885383367e-05, 'epoch': 0.02}) +[2026-06-03 13:49:38,702][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 1.434401512145996, 'train/grad_norm_pre_clip': 2.147606372833252, 'timing/data': 0.024547651410102844, 'timing/model': 1.242060186341405, 'learning_rate/qwen_vl_interface': 9.985357335232359e-06, 'learning_rate/action_model': 9.98535733523236e-05, 'epoch': 0.02}) +[2026-06-03 13:52:56,492][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 1.3910465240478516, 'train/grad_norm_pre_clip': 2.1593029499053955, 'timing/data': 0.02665463648736477, 'timing/model': 1.319689279422164, 'learning_rate/qwen_vl_interface': 9.980073405979643e-06, 'learning_rate/action_model': 9.980073405979643e-05, 'epoch': 0.02}) +[2026-06-03 13:56:10,811][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 1.1232073307037354, 'train/grad_norm_pre_clip': 1.4653841257095337, 'timing/data': 0.024607699364423752, 'timing/model': 1.259263053536415, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 0.02}) +[2026-06-03 13:59:23,157][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 1.081139087677002, 'train/grad_norm_pre_clip': 2.061397075653076, 'timing/data': 0.02098700776696205, 'timing/model': 1.2277335859835148, 'learning_rate/qwen_vl_interface': 9.967075170535901e-06, 'learning_rate/action_model': 9.967075170535901e-05, 'epoch': 0.02}) +[2026-06-03 14:06:20,800][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state +[2026-06-03 14:06:20,805][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-03 14:06:26,950][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/model.safetensors +[2026-06-03 14:06:36,957][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/optimizer.bin +[2026-06-03 14:06:36,958][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/sampler.bin +[2026-06-03 14:06:36,958][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/sampler_1.bin +[2026-06-03 14:06:36,959][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/random_states_0.pkl +[2026-06-03 14:06:36,972][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-03 14:06:36,972][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 200 +[2026-06-03 14:06:36,974][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 1.7108408212661743, 'train/grad_norm_pre_clip': 2.351280927658081, 'eval/loss': 1.5037981271743774, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 100.0, 'rl_games_eval/mid_train/mean_length': 597.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 100.0, 'rl_games_eval/mid_train/macro_mean_length': 597.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 100.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 597.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 100.0, 'checkpoint/best_score': 100.0, 'checkpoint/best_step': 200.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.028689134865999222, 'timing/model': 1.3230019733309746, 'learning_rate/qwen_vl_interface': 9.959363091525602e-06, 'learning_rate/action_model': 9.959363091525601e-05, 'epoch': 0.02}) +[2026-06-03 14:09:50,641][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 1.2466939687728882, 'train/grad_norm_pre_clip': 1.6739219427108765, 'timing/data': 0.02501312643289566, 'timing/model': 1.266646983101964, 'learning_rate/qwen_vl_interface': 9.950844087393436e-06, 'learning_rate/action_model': 9.950844087393435e-05, 'epoch': 0.02}) +[2026-06-03 14:13:05,173][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 1.3312305212020874, 'train/grad_norm_pre_clip': 4.26563024520874, 'timing/data': 0.02259301207959652, 'timing/model': 1.227685758844018, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.02}) +[2026-06-03 14:16:19,013][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 1.4755587577819824, 'train/grad_norm_pre_clip': 1.9688453674316406, 'timing/data': 0.025455091148614883, 'timing/model': 1.261044954881072, 'learning_rate/qwen_vl_interface': 9.93139128052587e-06, 'learning_rate/action_model': 9.93139128052587e-05, 'epoch': 0.03}) +[2026-06-03 14:19:36,022][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 1.077089786529541, 'train/grad_norm_pre_clip': 1.8432716131210327, 'timing/data': 0.024790074676275253, 'timing/model': 1.2529582884162664, 'learning_rate/qwen_vl_interface': 9.920460810928786e-06, 'learning_rate/action_model': 9.920460810928785e-05, 'epoch': 0.03}) +[2026-06-03 14:22:51,125][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 1.6389893293380737, 'train/grad_norm_pre_clip': 2.74018931388855, 'timing/data': 0.02561165951192379, 'timing/model': 1.2555294558405876, 'learning_rate/qwen_vl_interface': 9.908730081915346e-06, 'learning_rate/action_model': 9.908730081915345e-05, 'epoch': 0.03}) +[2026-06-03 14:26:07,415][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 1.3874849081039429, 'train/grad_norm_pre_clip': 2.1738946437835693, 'timing/data': 0.026856640353798866, 'timing/model': 1.2592736966907978, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 0.03}) +[2026-06-03 14:29:23,624][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 1.47865629196167, 'train/grad_norm_pre_clip': 2.209273099899292, 'timing/data': 0.02693837881088257, 'timing/model': 1.2670121975243092, 'learning_rate/qwen_vl_interface': 9.882876022415435e-06, 'learning_rate/action_model': 9.882876022415435e-05, 'epoch': 0.03}) +[2026-06-03 14:32:38,932][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 1.1466096639633179, 'train/grad_norm_pre_clip': 1.369870901107788, 'timing/data': 0.04274879954755306, 'timing/model': 1.2790766377002, 'learning_rate/qwen_vl_interface': 9.868757121888965e-06, 'learning_rate/action_model': 9.868757121888964e-05, 'epoch': 0.03}) +[2026-06-03 14:35:57,588][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 1.127073049545288, 'train/grad_norm_pre_clip': 1.883998990058899, 'timing/data': 0.026055924594402313, 'timing/model': 1.5078516155481339, 'learning_rate/qwen_vl_interface': 9.853846821107084e-06, 'learning_rate/action_model': 9.853846821107085e-05, 'epoch': 0.03}) +[2026-06-03 15:01:06,169][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 300 +[2026-06-03 15:01:06,171][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 1.224234938621521, 'train/grad_norm_pre_clip': 1.7917252779006958, 'eval/loss': 1.295228362083435, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 70.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 70.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 70.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 70.0, 'checkpoint/best_score': 100.0, 'checkpoint/best_step': 200.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.026917526498436928, 'timing/model': 1.2576216887682676, 'learning_rate/qwen_vl_interface': 9.838147674873076e-06, 'learning_rate/action_model': 9.838147674873075e-05, 'epoch': 0.03}) +[2026-06-03 15:04:25,129][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 1.1909399032592773, 'train/grad_norm_pre_clip': 2.453246593475342, 'timing/data': 0.024651169776916504, 'timing/model': 1.2586627826094627, 'learning_rate/qwen_vl_interface': 9.821662373154824e-06, 'learning_rate/action_model': 9.821662373154825e-05, 'epoch': 0.03}) +[2026-06-03 15:07:41,036][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 1.3908151388168335, 'train/grad_norm_pre_clip': 1.6633814573287964, 'timing/data': 0.02306724339723587, 'timing/model': 1.2650560699403286, 'learning_rate/qwen_vl_interface': 9.804393740623918e-06, 'learning_rate/action_model': 9.804393740623918e-05, 'epoch': 0.04}) +[2026-06-03 15:10:56,366][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 1.376539707183838, 'train/grad_norm_pre_clip': 3.1699297428131104, 'timing/data': 0.02402774803340435, 'timing/model': 1.259918913245201, 'learning_rate/qwen_vl_interface': 9.786344736171637e-06, 'learning_rate/action_model': 9.786344736171638e-05, 'epoch': 0.04}) +[2026-06-03 15:14:10,555][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 1.438948631286621, 'train/grad_norm_pre_clip': 1.6422151327133179, 'timing/data': 0.027895383536815643, 'timing/model': 1.223746770992875, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.04}) +[2026-06-03 15:17:23,739][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 1.2968581914901733, 'train/grad_norm_pre_clip': 2.075547933578491, 'timing/data': 0.028811242431402206, 'timing/model': 1.2161091919988394, 'learning_rate/qwen_vl_interface': 9.747918115101753e-06, 'learning_rate/action_model': 9.747918115101753e-05, 'epoch': 0.04}) +[2026-06-03 15:20:35,809][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 1.318630576133728, 'train/grad_norm_pre_clip': 1.6087491512298584, 'timing/data': 0.03291403874754906, 'timing/model': 1.2585407998412848, 'learning_rate/qwen_vl_interface': 9.727547082687849e-06, 'learning_rate/action_model': 9.72754708268785e-05, 'epoch': 0.04}) +[2026-06-03 15:23:47,430][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 1.4129060506820679, 'train/grad_norm_pre_clip': 1.8657493591308594, 'timing/data': 0.036446599289774895, 'timing/model': 1.257451245561242, 'learning_rate/qwen_vl_interface': 9.7064088456318e-06, 'learning_rate/action_model': 9.7064088456318e-05, 'epoch': 0.04}) +[2026-06-03 15:27:03,795][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 1.2672892808914185, 'train/grad_norm_pre_clip': 1.7167952060699463, 'timing/data': 0.024449635297060013, 'timing/model': 1.256542457267642, 'learning_rate/qwen_vl_interface': 9.68450702586171e-06, 'learning_rate/action_model': 9.684507025861709e-05, 'epoch': 0.04}) +[2026-06-03 15:30:18,057][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 1.1499942541122437, 'train/grad_norm_pre_clip': 1.296684741973877, 'timing/data': 0.025536492466926575, 'timing/model': 1.2523142732679844, 'learning_rate/qwen_vl_interface': 9.66184537614164e-06, 'learning_rate/action_model': 9.661845376141641e-05, 'epoch': 0.04}) +[2026-06-03 15:41:01,859][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state +[2026-06-03 15:41:01,868][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-03 15:41:10,084][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/model.safetensors +[2026-06-03 15:41:21,526][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/optimizer.bin +[2026-06-03 15:41:21,527][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/sampler.bin +[2026-06-03 15:41:21,527][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/sampler_1.bin +[2026-06-03 15:41:21,529][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/random_states_0.pkl +[2026-06-03 15:41:21,556][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-03 15:41:21,556][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 400 +[2026-06-03 15:41:21,558][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.9458563923835754, 'train/grad_norm_pre_clip': 1.3374181985855103, 'eval/loss': 1.2636772394180298, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 265.0, 'rl_games_eval/mid_train/mean_length': 1257.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 265.0, 'rl_games_eval/mid_train/macro_mean_length': 1257.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 265.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 1257.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 265.0, 'checkpoint/best_score': 265.0, 'checkpoint/best_step': 400.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.020714513957500458, 'timing/model': 1.21849269233644, 'learning_rate/qwen_vl_interface': 9.638427779428613e-06, 'learning_rate/action_model': 9.638427779428613e-05, 'epoch': 0.04}) +[2026-06-03 15:44:35,421][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 1.2863880395889282, 'train/grad_norm_pre_clip': 1.4501278400421143, 'timing/data': 0.027795514091849327, 'timing/model': 1.2401983793824911, 'learning_rate/qwen_vl_interface': 9.61425824820727e-06, 'learning_rate/action_model': 9.614258248207268e-05, 'epoch': 0.05}) +[2026-06-03 15:47:51,500][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 1.1311914920806885, 'train/grad_norm_pre_clip': 1.5014265775680542, 'timing/data': 0.025377696380019188, 'timing/model': 1.3525578677654266, 'learning_rate/qwen_vl_interface': 9.589340923802355e-06, 'learning_rate/action_model': 9.589340923802354e-05, 'epoch': 0.05}) +[2026-06-03 15:51:05,958][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 1.0884846448898315, 'train/grad_norm_pre_clip': 1.278943419456482, 'timing/data': 0.021244237199425697, 'timing/model': 1.2223952244967222, 'learning_rate/qwen_vl_interface': 9.563680075669138e-06, 'learning_rate/action_model': 9.563680075669137e-05, 'epoch': 0.05}) +[2026-06-03 15:54:18,431][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 1.245810627937317, 'train/grad_norm_pre_clip': 2.065535545349121, 'timing/data': 0.020379750058054924, 'timing/model': 1.2253254316747189, 'learning_rate/qwen_vl_interface': 9.53728010066184e-06, 'learning_rate/action_model': 9.53728010066184e-05, 'epoch': 0.05}) +[2026-06-03 15:57:32,915][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 1.1197046041488647, 'train/grad_norm_pre_clip': 1.6892050504684448, 'timing/data': 0.02015727572143078, 'timing/model': 1.342038905248046, 'learning_rate/qwen_vl_interface': 9.51014552228027e-06, 'learning_rate/action_model': 9.510145522280269e-05, 'epoch': 0.05}) +[2026-06-03 16:00:52,299][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 1.2543305158615112, 'train/grad_norm_pre_clip': 1.5491313934326172, 'timing/data': 0.02861684001982212, 'timing/model': 1.3169833533465862, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.05}) +[2026-06-03 16:04:09,105][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 1.1081957817077637, 'train/grad_norm_pre_clip': 2.260828733444214, 'timing/data': 0.02601512521505356, 'timing/model': 1.2556371036916971, 'learning_rate/qwen_vl_interface': 9.45369127794945e-06, 'learning_rate/action_model': 9.453691277949449e-05, 'epoch': 0.05}) +[2026-06-03 16:07:26,721][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 1.2113232612609863, 'train/grad_norm_pre_clip': 2.0510518550872803, 'timing/data': 0.02415367402136326, 'timing/model': 1.2642811816185713, 'learning_rate/qwen_vl_interface': 9.424381285144337e-06, 'learning_rate/action_model': 9.424381285144335e-05, 'epoch': 0.05}) +[2026-06-03 16:10:43,154][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 1.2153587341308594, 'train/grad_norm_pre_clip': 2.1636998653411865, 'timing/data': 0.025392338633537292, 'timing/model': 1.2647856529802084, 'learning_rate/qwen_vl_interface': 9.394356033595787e-06, 'learning_rate/action_model': 9.394356033595787e-05, 'epoch': 0.05}) +[2026-06-03 16:36:10,724][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-03 16:36:10,727][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 1.314128041267395, 'train/grad_norm_pre_clip': 2.4822134971618652, 'eval/loss': 1.2413145303726196, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 70.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 70.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 70.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 70.0, 'checkpoint/best_score': 265.0, 'checkpoint/best_step': 400.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.024037616327404976, 'timing/model': 1.259003883227706, 'learning_rate/qwen_vl_interface': 9.363620667976085e-06, 'learning_rate/action_model': 9.363620667976085e-05, 'epoch': 0.06}) +[2026-06-03 16:39:25,718][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 1.3635320663452148, 'train/grad_norm_pre_clip': 1.3651349544525146, 'timing/data': 0.02640284039080143, 'timing/model': 1.270985871553421, 'learning_rate/qwen_vl_interface': 9.332180454631902e-06, 'learning_rate/action_model': 9.332180454631901e-05, 'epoch': 0.06}) +[2026-06-03 16:42:37,411][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 1.1986500024795532, 'train/grad_norm_pre_clip': 2.241499185562134, 'timing/data': 0.028597038239240646, 'timing/model': 1.258347099646926, 'learning_rate/qwen_vl_interface': 9.300040780681938e-06, 'learning_rate/action_model': 9.300040780681938e-05, 'epoch': 0.06}) +[2026-06-03 16:45:54,533][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.9517638087272644, 'train/grad_norm_pre_clip': 2.763779640197754, 'timing/data': 0.03413497284054756, 'timing/model': 1.2581432480365038, 'learning_rate/qwen_vl_interface': 9.267207153093874e-06, 'learning_rate/action_model': 9.267207153093875e-05, 'epoch': 0.06}) +[2026-06-03 16:49:09,415][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 1.0977962017059326, 'train/grad_norm_pre_clip': 2.4301958084106445, 'timing/data': 0.033526161685585976, 'timing/model': 1.2560226880013943, 'learning_rate/qwen_vl_interface': 9.233685197740765e-06, 'learning_rate/action_model': 9.233685197740764e-05, 'epoch': 0.06}) +[2026-06-03 16:52:21,814][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 1.3636066913604736, 'train/grad_norm_pre_clip': 1.8232629299163818, 'timing/data': 0.03489300422370434, 'timing/model': 1.3958198558539152, 'learning_rate/qwen_vl_interface': 9.199480658437092e-06, 'learning_rate/action_model': 9.199480658437091e-05, 'epoch': 0.06}) +[2026-06-03 16:55:35,553][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 1.1610151529312134, 'train/grad_norm_pre_clip': 1.5215815305709839, 'timing/data': 0.030494889244437218, 'timing/model': 1.2619250044226646, 'learning_rate/qwen_vl_interface': 9.164599395954575e-06, 'learning_rate/action_model': 9.164599395954574e-05, 'epoch': 0.06}) +[2026-06-03 16:58:57,942][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.9476045370101929, 'train/grad_norm_pre_clip': 1.5833182334899902, 'timing/data': 0.03608548827469349, 'timing/model': 1.2742808032780886, 'learning_rate/qwen_vl_interface': 9.129047387017977e-06, 'learning_rate/action_model': 9.129047387017976e-05, 'epoch': 0.06}) +[2026-06-03 17:02:10,313][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.8815813660621643, 'train/grad_norm_pre_clip': 2.030728578567505, 'timing/data': 0.016456447541713715, 'timing/model': 1.2629567068070173, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.06}) +[2026-06-03 17:05:26,727][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 1.1974198818206787, 'train/grad_norm_pre_clip': 2.0872530937194824, 'timing/data': 0.016890840604901314, 'timing/model': 1.2559259235858917, 'learning_rate/qwen_vl_interface': 9.055955610282539e-06, 'learning_rate/action_model': 9.055955610282538e-05, 'epoch': 0.07}) +[2026-06-03 17:30:57,715][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state +[2026-06-03 17:30:57,724][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-03 17:31:05,851][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/model.safetensors +[2026-06-03 17:31:16,087][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/optimizer.bin +[2026-06-03 17:31:16,088][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/sampler.bin +[2026-06-03 17:31:16,088][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/sampler_1.bin +[2026-06-03 17:31:16,090][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/random_states_0.pkl +[2026-06-03 17:31:16,107][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-03 17:31:16,109][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 600 +[2026-06-03 17:31:16,111][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 1.0801774263381958, 'train/grad_norm_pre_clip': 2.6155903339385986, 'eval/loss': 1.1128623485565186, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 2480.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 2480.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 2480.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 2480.0, 'checkpoint/best_score': 2480.0, 'checkpoint/best_step': 600.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.009465038776397705, 'timing/model': 1.2610432300716639, 'learning_rate/qwen_vl_interface': 9.018428366383368e-06, 'learning_rate/action_model': 9.018428366383368e-05, 'epoch': 0.07}) +[2026-06-03 17:34:30,897][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 1.259628415107727, 'train/grad_norm_pre_clip': 2.987802743911743, 'timing/data': 0.012062707915902138, 'timing/model': 1.2662368975579739, 'learning_rate/qwen_vl_interface': 8.98025542168354e-06, 'learning_rate/action_model': 8.98025542168354e-05, 'epoch': 0.07}) +[2026-06-03 17:37:50,281][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 1.1011950969696045, 'train/grad_norm_pre_clip': 1.9196372032165527, 'timing/data': 0.015544509515166283, 'timing/model': 1.2451060395687819, 'learning_rate/qwen_vl_interface': 8.941443316920613e-06, 'learning_rate/action_model': 8.941443316920612e-05, 'epoch': 0.07}) +[2026-06-03 17:41:06,161][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 1.2251743078231812, 'train/grad_norm_pre_clip': 1.5956965684890747, 'timing/data': 0.022905027493834496, 'timing/model': 1.267799224704504, 'learning_rate/qwen_vl_interface': 8.901998702348934e-06, 'learning_rate/action_model': 8.901998702348933e-05, 'epoch': 0.07}) +[2026-06-03 17:44:23,679][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 1.0808333158493042, 'train/grad_norm_pre_clip': 2.4851582050323486, 'timing/data': 0.0356630589812994, 'timing/model': 1.498781032860279, 'learning_rate/qwen_vl_interface': 8.861928336600148e-06, 'learning_rate/action_model': 8.861928336600148e-05, 'epoch': 0.07}) +[2026-06-03 17:47:41,451][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 1.1094374656677246, 'train/grad_norm_pre_clip': 1.922776699066162, 'timing/data': 0.03194133937358856, 'timing/model': 1.2631455603986979, 'learning_rate/qwen_vl_interface': 8.821239085525143e-06, 'learning_rate/action_model': 8.821239085525143e-05, 'epoch': 0.07}) +[2026-06-03 17:50:59,667][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.799237072467804, 'train/grad_norm_pre_clip': 2.13590145111084, 'timing/data': 0.02806108258664608, 'timing/model': 1.2426496744155884, 'learning_rate/qwen_vl_interface': 8.779937921017624e-06, 'learning_rate/action_model': 8.779937921017623e-05, 'epoch': 0.07}) +[2026-06-03 17:54:12,337][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 1.0862658023834229, 'train/grad_norm_pre_clip': 3.5139405727386475, 'timing/data': 0.032948900014162064, 'timing/model': 1.258413903415203, 'learning_rate/qwen_vl_interface': 8.738031919819508e-06, 'learning_rate/action_model': 8.738031919819508e-05, 'epoch': 0.07}) +[2026-06-03 17:57:27,755][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.8210982084274292, 'train/grad_norm_pre_clip': 2.560831308364868, 'timing/data': 0.032460011541843414, 'timing/model': 1.26778750307858, 'learning_rate/qwen_vl_interface': 8.695528262308367e-06, 'learning_rate/action_model': 8.695528262308366e-05, 'epoch': 0.08}) +[2026-06-03 18:00:43,009][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 1.0358167886734009, 'train/grad_norm_pre_clip': 2.050232172012329, 'timing/data': 0.03842339292168617, 'timing/model': 1.2737809047102928, 'learning_rate/qwen_vl_interface': 8.65243423126711e-06, 'learning_rate/action_model': 8.652434231267109e-05, 'epoch': 0.08}) +[2026-06-03 18:06:37,689][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 700 +[2026-06-03 18:06:37,692][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 1.340772271156311, 'train/grad_norm_pre_clip': 3.0170576572418213, 'eval/loss': 1.1126718521118164, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 90.0, 'rl_games_eval/mid_train/mean_length': 435.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 90.0, 'rl_games_eval/mid_train/macro_mean_length': 435.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 90.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 435.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 90.0, 'checkpoint/best_score': 2480.0, 'checkpoint/best_step': 600.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.048256346955895424, 'timing/model': 1.2918786238878965, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.08}) +[2026-06-03 18:09:55,664][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 1.1378681659698486, 'train/grad_norm_pre_clip': 2.2021148204803467, 'timing/data': 0.019051307812333107, 'timing/model': 1.2696331534534693, 'learning_rate/qwen_vl_interface': 8.564504684247982e-06, 'learning_rate/action_model': 8.564504684247982e-05, 'epoch': 0.08}) +[2026-06-03 18:13:10,327][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 1.0714253187179565, 'train/grad_norm_pre_clip': 1.9710488319396973, 'timing/data': 0.009474657475948334, 'timing/model': 1.241590715944767, 'learning_rate/qwen_vl_interface': 8.519684234545333e-06, 'learning_rate/action_model': 8.519684234545332e-05, 'epoch': 0.08}) +[2026-06-03 18:16:29,833][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 1.3878026008605957, 'train/grad_norm_pre_clip': 2.242746353149414, 'timing/data': 0.0007270928472280502, 'timing/model': 1.365738045424223, 'learning_rate/qwen_vl_interface': 8.474303541281474e-06, 'learning_rate/action_model': 8.474303541281474e-05, 'epoch': 0.08}) +[2026-06-03 18:19:46,987][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 1.2479652166366577, 'train/grad_norm_pre_clip': 1.794658899307251, 'timing/data': 0.0005367845296859741, 'timing/model': 1.2335969787091017, 'learning_rate/qwen_vl_interface': 8.428370380204577e-06, 'learning_rate/action_model': 8.428370380204577e-05, 'epoch': 0.08}) +[2026-06-03 18:23:01,942][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 1.2001993656158447, 'train/grad_norm_pre_clip': 2.6433205604553223, 'timing/data': 0.0005683489143848419, 'timing/model': 1.2413579672574997, 'learning_rate/qwen_vl_interface': 8.381892621725327e-06, 'learning_rate/action_model': 8.381892621725327e-05, 'epoch': 0.08}) +[2026-06-03 18:26:16,099][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 1.3878812789916992, 'train/grad_norm_pre_clip': 3.1735219955444336, 'timing/data': 0.0005890838801860809, 'timing/model': 1.237108862027526, 'learning_rate/qwen_vl_interface': 8.334878229568372e-06, 'learning_rate/action_model': 8.334878229568372e-05, 'epoch': 0.08}) +[2026-06-03 18:29:28,904][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 1.4313812255859375, 'train/grad_norm_pre_clip': 2.495751142501831, 'timing/data': 0.0006824620068073273, 'timing/model': 1.2339134104549885, 'learning_rate/qwen_vl_interface': 8.287335259407779e-06, 'learning_rate/action_model': 8.287335259407779e-05, 'epoch': 0.09}) +[2026-06-03 18:32:43,461][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 1.0798497200012207, 'train/grad_norm_pre_clip': 2.1347315311431885, 'timing/data': 0.009005948901176453, 'timing/model': 1.227223927155137, 'learning_rate/qwen_vl_interface': 8.23927185748673e-06, 'learning_rate/action_model': 8.239271857486729e-05, 'epoch': 0.09}) +[2026-06-03 18:35:57,029][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 1.168325662612915, 'train/grad_norm_pre_clip': 2.556792736053467, 'timing/data': 0.0006543900817632675, 'timing/model': 1.2620659228414297, 'learning_rate/qwen_vl_interface': 8.190696259221713e-06, 'learning_rate/action_model': 8.190696259221713e-05, 'epoch': 0.09}) +[2026-06-03 19:01:11,231][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state +[2026-06-03 19:01:11,240][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-03 19:01:18,098][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/model.safetensors +[2026-06-03 19:01:27,519][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/optimizer.bin +[2026-06-03 19:01:27,520][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/sampler.bin +[2026-06-03 19:01:27,520][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/sampler_1.bin +[2026-06-03 19:01:27,521][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/random_states_0.pkl +[2026-06-03 19:01:27,535][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-03 19:01:27,536][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 800 +[2026-06-03 19:01:27,537][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.9687876105308533, 'train/grad_norm_pre_clip': 2.0011162757873535, 'eval/loss': 1.0399675369262695, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 5640.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 5640.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 5640.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 5640.0, 'checkpoint/best_score': 5640.0, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.0009426642209291458, 'timing/model': 1.490764057263732, 'learning_rate/qwen_vl_interface': 8.141616787791425e-06, 'learning_rate/action_model': 8.141616787791424e-05, 'epoch': 0.09}) +[2026-06-03 19:04:44,799][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.859179675579071, 'train/grad_norm_pre_clip': 1.812647819519043, 'timing/data': 0.0007869023829698563, 'timing/model': 1.2775420751422644, 'learning_rate/qwen_vl_interface': 8.092041852710631e-06, 'learning_rate/action_model': 8.092041852710631e-05, 'epoch': 0.09}) +[2026-06-03 19:08:00,812][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 1.0733016729354858, 'train/grad_norm_pre_clip': 2.3732993602752686, 'timing/data': 0.0005779657512903214, 'timing/model': 1.3827590513974428, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.09}) +[2026-06-03 19:11:13,225][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 1.397908091545105, 'train/grad_norm_pre_clip': 2.753876209259033, 'timing/data': 0.0004786737263202667, 'timing/model': 1.237047715112567, 'learning_rate/qwen_vl_interface': 7.991439652676861e-06, 'learning_rate/action_model': 7.991439652676861e-05, 'epoch': 0.09}) +[2026-06-03 19:14:29,308][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 1.0553457736968994, 'train/grad_norm_pre_clip': 2.5187063217163086, 'timing/data': 0.0005120523273944855, 'timing/model': 1.241306433454156, 'learning_rate/qwen_vl_interface': 7.940429625392957e-06, 'learning_rate/action_model': 7.940429625392956e-05, 'epoch': 0.09}) +[2026-06-03 19:17:45,625][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 1.3221772909164429, 'train/grad_norm_pre_clip': 2.1711227893829346, 'timing/data': 0.0004336852580308914, 'timing/model': 1.4061482828110456, 'learning_rate/qwen_vl_interface': 7.888958606843111e-06, 'learning_rate/action_model': 7.888958606843111e-05, 'epoch': 0.09}) +[2026-06-03 19:21:03,294][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 1.033609390258789, 'train/grad_norm_pre_clip': 3.2764997482299805, 'timing/data': 0.000746438279747963, 'timing/model': 1.2553664706647396, 'learning_rate/qwen_vl_interface': 7.83703541632138e-06, 'learning_rate/action_model': 7.83703541632138e-05, 'epoch': 0.1}) +[2026-06-03 19:24:16,766][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 1.3103814125061035, 'train/grad_norm_pre_clip': 3.021294355392456, 'timing/data': 0.0004783943295478821, 'timing/model': 1.2336549125611782, 'learning_rate/qwen_vl_interface': 7.784668950599155e-06, 'learning_rate/action_model': 7.784668950599155e-05, 'epoch': 0.1}) +[2026-06-03 19:27:32,088][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.8150141835212708, 'train/grad_norm_pre_clip': 1.644944667816162, 'timing/data': 0.000690266489982605, 'timing/model': 1.266599738970399, 'learning_rate/qwen_vl_interface': 7.731868182400758e-06, 'learning_rate/action_model': 7.731868182400758e-05, 'epoch': 0.1}) +[2026-06-03 19:30:48,276][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.7547316551208496, 'train/grad_norm_pre_clip': 3.142514228820801, 'timing/data': 0.0006628502160310745, 'timing/model': 1.27314443141222, 'learning_rate/qwen_vl_interface': 7.678642158865997e-06, 'learning_rate/action_model': 7.678642158865997e-05, 'epoch': 0.1}) +[2026-06-03 19:40:46,743][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 900 +[2026-06-03 19:40:46,745][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.8922595381736755, 'train/grad_norm_pre_clip': 2.6031155586242676, 'eval/loss': 1.0512275695800781, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 250.0, 'rl_games_eval/mid_train/mean_length': 1094.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 250.0, 'rl_games_eval/mid_train/macro_mean_length': 1094.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 250.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 1094.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 250.0, 'checkpoint/best_score': 5640.0, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00078604556620121, 'timing/model': 1.2838497906923294, 'learning_rate/qwen_vl_interface': 7.625e-06, 'learning_rate/action_model': 7.625e-05, 'epoch': 0.1}) +[2026-06-03 19:44:00,284][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 1.560578465461731, 'train/grad_norm_pre_clip': 2.587904691696167, 'timing/data': 0.0005542300641536713, 'timing/model': 1.256059242412448, 'learning_rate/qwen_vl_interface': 7.570950897110537e-06, 'learning_rate/action_model': 7.570950897110536e-05, 'epoch': 0.1}) +[2026-06-03 19:47:11,715][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.9791827201843262, 'train/grad_norm_pre_clip': 2.539233684539795, 'timing/data': 0.0006915051490068436, 'timing/model': 1.2664888575673103, 'learning_rate/qwen_vl_interface': 7.516504111233141e-06, 'learning_rate/action_model': 7.51650411123314e-05, 'epoch': 0.1}) +[2026-06-03 19:50:26,209][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 1.116434097290039, 'train/grad_norm_pre_clip': 2.652099609375, 'timing/data': 0.00041416287422180176, 'timing/model': 1.2882091831415892, 'learning_rate/qwen_vl_interface': 7.46166897154428e-06, 'learning_rate/action_model': 7.46166897154428e-05, 'epoch': 0.1}) +[2026-06-03 19:53:40,901][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 1.0392847061157227, 'train/grad_norm_pre_clip': 2.1938271522521973, 'timing/data': 0.0003901328891515732, 'timing/model': 1.225105345249176, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.1}) +[2026-06-03 19:56:57,164][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 1.2111358642578125, 'train/grad_norm_pre_clip': 3.4979171752929688, 'timing/data': 0.0005919449031352997, 'timing/model': 1.2704886961728334, 'learning_rate/qwen_vl_interface': 7.350871278540257e-06, 'learning_rate/action_model': 7.350871278540256e-05, 'epoch': 0.11}) +[2026-06-04 04:17:13,065][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-04 04:17:13,114][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-04 04:17:14,156][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `demon_attack_train__bridge` +[2026-06-04 04:17:14,640][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-04 04:17:14,640][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-04 04:17:14,640][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-04 04:17:14,640][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-04 04:17:14,640][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-04 04:17:14,641][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-04 04:17:14,641][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-04 04:17:14,641][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-04 04:17:14,641][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-04 04:17:14,641][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-04 04:17:14,641][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-04 04:17:14,641][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-04 04:17:14,641][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-04 04:17:14,641][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-04 04:17:14,641][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-04 04:17:14,641][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-04 04:17:14,641][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-04 04:17:14,641][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-04 04:17:14,641][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-04 04:17:14,642][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-04 04:17:14,642][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-04 04:17:14,642][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-04 04:17:14,642][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-04 04:17:14,642][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-04 04:17:14,642][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-04 04:17:14,642][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-04 04:17:14,642][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-04 04:17:14,642][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-04 04:17:14,642][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-04 04:17:14,642][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-04 04:17:14,642][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-04 04:17:14,676][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `demon_attack_train__bridge__val` +[2026-06-04 04:17:14,701][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/config.full.yaml +[2026-06-04 04:17:14,706][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/config.yaml +[2026-06-04 04:17:14,706][starVLA.training.train_starvla][INFO] - Loaded best checkpoint metadata: step=800 score=5640.0 +[2026-06-04 04:17:14,706][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state, steps: 800 +[2026-06-04 04:17:17,191][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state +[2026-06-04 04:17:20,245][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-04 04:17:31,044][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-04 04:17:31,045][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-04 04:17:31,045][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-04 04:17:31,047][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-04 04:17:31,396][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-04 04:17:32,727][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-04 04:17:32,728][starVLA.training.train_starvla][INFO] - Total optimization steps = 2500 +[2026-06-04 04:17:32,728][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-04 04:17:32,728][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-04 04:17:32,728][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-04 04:20:19,435][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 1.1047650575637817, 'train/grad_norm_pre_clip': 2.482729196548462, 'timing/data': 0.009289249777793884, 'timing/model': 1.050734831020236, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.09}) +[2026-06-04 04:23:00,528][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 1.0139892101287842, 'train/grad_norm_pre_clip': 2.1928417682647705, 'timing/data': 0.00915527157485485, 'timing/model': 1.0523369498550892, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.09}) +[2026-06-04 04:25:44,060][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.9134072065353394, 'train/grad_norm_pre_clip': 2.1881117820739746, 'timing/data': 0.009778203442692757, 'timing/model': 1.056750224903226, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.09}) +[2026-06-04 04:28:25,854][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 1.2563211917877197, 'train/grad_norm_pre_clip': 2.6386542320251465, 'timing/data': 0.009966162964701653, 'timing/model': 1.0575425829738379, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.09}) +[2026-06-04 04:31:07,064][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 1.118863821029663, 'train/grad_norm_pre_clip': 1.8474189043045044, 'timing/data': 0.009978501126170158, 'timing/model': 1.0480288844555616, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.09}) +[2026-06-04 04:33:48,153][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 1.336171269416809, 'train/grad_norm_pre_clip': 2.626675844192505, 'timing/data': 0.010962564498186111, 'timing/model': 1.0502954367548227, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.1}) +[2026-06-04 04:36:33,131][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 1.0138672590255737, 'train/grad_norm_pre_clip': 2.5335307121276855, 'timing/data': 0.009370312094688416, 'timing/model': 1.0550548434257507, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.1}) +[2026-06-04 04:39:14,431][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 1.1513333320617676, 'train/grad_norm_pre_clip': 2.5071046352386475, 'timing/data': 0.008292017504572868, 'timing/model': 1.0472890883684158, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.1}) +[2026-06-04 04:41:55,861][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.9088993668556213, 'train/grad_norm_pre_clip': 2.226950168609619, 'timing/data': 0.009915387257933617, 'timing/model': 1.0634757857769728, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.1}) +[2026-06-04 04:44:48,441][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpna8cw9e2/test.c -o /tmp/tmpna8cw9e2/test.o +[2026-06-04 04:44:48,466][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpna8cw9e2/test.o -laio -o /tmp/tmpna8cw9e2/a.out +[2026-06-04 04:44:48,876][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmprjlcwytu/test.c -o /tmp/tmprjlcwytu/test.o +[2026-06-04 04:44:48,899][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmprjlcwytu/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmprjlcwytu/a.out +[2026-06-04 04:44:48,955][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpx_3wva2y/test.c -o /tmp/tmpx_3wva2y/test.o +[2026-06-04 04:44:48,981][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpx_3wva2y/test.o -laio -o /tmp/tmpx_3wva2y/a.out +[2026-06-04 05:05:19,092][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 900 +[2026-06-04 05:05:19,095][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 1.1143717765808105, 'train/grad_norm_pre_clip': 2.209256410598755, 'eval/loss': 1.0612791776657104, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 4980.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 4980.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 4980.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 4980.0, 'checkpoint/best_score': 5640.0, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.009562978520989418, 'timing/model': 1.0497242528945208, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.1}) +[2026-06-04 05:08:36,953][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 1.138441562652588, 'train/grad_norm_pre_clip': 2.9181599617004395, 'timing/data': 0.008904479444026947, 'timing/model': 1.2565060965716839, 'learning_rate/qwen_vl_interface': 9.999593055976534e-06, 'learning_rate/action_model': 9.999593055976533e-05, 'epoch': 0.1}) +[2026-06-04 05:11:57,310][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.9481862783432007, 'train/grad_norm_pre_clip': 2.033048629760742, 'timing/data': 0.009879130870103836, 'timing/model': 1.2912946660071611, 'learning_rate/qwen_vl_interface': 9.998372293633897e-06, 'learning_rate/action_model': 9.998372293633897e-05, 'epoch': 0.1}) +[2026-06-04 05:15:16,895][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.8787939548492432, 'train/grad_norm_pre_clip': 2.858591079711914, 'timing/data': 0.010098887607455254, 'timing/model': 1.295868320390582, 'learning_rate/qwen_vl_interface': 9.996337922143435e-06, 'learning_rate/action_model': 9.996337922143436e-05, 'epoch': 0.1}) +[2026-06-04 05:18:35,842][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 1.1506794691085815, 'train/grad_norm_pre_clip': 3.1010124683380127, 'timing/data': 0.008885066956281662, 'timing/model': 1.2456154245883226, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 0.1}) +[2026-06-04 05:21:51,557][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 1.101051926612854, 'train/grad_norm_pre_clip': 2.7172884941101074, 'timing/data': 0.009184587746858597, 'timing/model': 1.2549047656357288, 'learning_rate/qwen_vl_interface': 9.989829885383368e-06, 'learning_rate/action_model': 9.989829885383367e-05, 'epoch': 0.11}) +[2026-06-04 05:25:08,753][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 1.095432162284851, 'train/grad_norm_pre_clip': 2.128272533416748, 'timing/data': 0.008793583139777184, 'timing/model': 1.2643259782344103, 'learning_rate/qwen_vl_interface': 9.985357335232359e-06, 'learning_rate/action_model': 9.98535733523236e-05, 'epoch': 0.11}) +[2026-06-04 05:28:23,451][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 1.37813138961792, 'train/grad_norm_pre_clip': 2.6423072814941406, 'timing/data': 0.007750110700726509, 'timing/model': 1.253435906022787, 'learning_rate/qwen_vl_interface': 9.980073405979643e-06, 'learning_rate/action_model': 9.980073405979643e-05, 'epoch': 0.11}) +[2026-06-04 05:31:42,746][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 1.0022079944610596, 'train/grad_norm_pre_clip': 2.2210981845855713, 'timing/data': 0.00834919884800911, 'timing/model': 1.2628205362707376, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 0.11}) +[2026-06-04 05:34:58,300][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 1.02129065990448, 'train/grad_norm_pre_clip': 2.5638411045074463, 'timing/data': 0.009705642238259315, 'timing/model': 1.2683112360537052, 'learning_rate/qwen_vl_interface': 9.967075170535901e-06, 'learning_rate/action_model': 9.967075170535901e-05, 'epoch': 0.11}) +[2026-06-04 05:42:48,328][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-06-04 05:42:48,330][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 1.2176069021224976, 'train/grad_norm_pre_clip': 2.7203474044799805, 'eval/loss': 1.108130693435669, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 160.0, 'rl_games_eval/mid_train/mean_length': 818.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 160.0, 'rl_games_eval/mid_train/macro_mean_length': 818.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 160.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 818.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 160.0, 'checkpoint/best_score': 5640.0, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.01737099513411522, 'timing/model': 1.4242022410035133, 'learning_rate/qwen_vl_interface': 9.959363091525602e-06, 'learning_rate/action_model': 9.959363091525601e-05, 'epoch': 0.11}) +[2026-06-04 05:46:07,717][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 1.0373464822769165, 'train/grad_norm_pre_clip': 1.9397789239883423, 'timing/data': 0.0007232446223497391, 'timing/model': 1.4667634349316359, 'learning_rate/qwen_vl_interface': 9.950844087393436e-06, 'learning_rate/action_model': 9.950844087393435e-05, 'epoch': 0.11}) +[2026-06-04 05:49:26,990][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 1.029248833656311, 'train/grad_norm_pre_clip': 2.670649766921997, 'timing/data': 0.0006754398345947266, 'timing/model': 1.4659658372402191, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.11}) +[2026-06-04 05:52:46,205][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 1.0976521968841553, 'train/grad_norm_pre_clip': 1.675549030303955, 'timing/data': 0.00068705715239048, 'timing/model': 1.2907794397324324, 'learning_rate/qwen_vl_interface': 9.93139128052587e-06, 'learning_rate/action_model': 9.93139128052587e-05, 'epoch': 0.11}) +[2026-06-04 05:56:05,355][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 1.0178499221801758, 'train/grad_norm_pre_clip': 2.2377498149871826, 'timing/data': 0.0006239600479602814, 'timing/model': 1.4294361118227243, 'learning_rate/qwen_vl_interface': 9.920460810928786e-06, 'learning_rate/action_model': 9.920460810928785e-05, 'epoch': 0.12}) +[2026-06-04 05:59:27,053][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 1.3252811431884766, 'train/grad_norm_pre_clip': 2.814800977706909, 'timing/data': 0.0004917550832033157, 'timing/model': 1.2616412416100502, 'learning_rate/qwen_vl_interface': 9.908730081915346e-06, 'learning_rate/action_model': 9.908730081915345e-05, 'epoch': 0.12}) +[2026-06-04 06:02:47,781][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 1.0317500829696655, 'train/grad_norm_pre_clip': 2.758321523666382, 'timing/data': 0.0007922723889350891, 'timing/model': 1.27302535623312, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 0.12}) +[2026-06-04 06:06:08,005][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 1.314521312713623, 'train/grad_norm_pre_clip': 2.6843225955963135, 'timing/data': 0.0007669739425182343, 'timing/model': 1.770091202110052, 'learning_rate/qwen_vl_interface': 9.882876022415435e-06, 'learning_rate/action_model': 9.882876022415435e-05, 'epoch': 0.12}) +[2026-06-04 06:09:24,942][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.7087374329566956, 'train/grad_norm_pre_clip': 2.3602001667022705, 'timing/data': 0.00039542652666568756, 'timing/model': 1.2640603128820658, 'learning_rate/qwen_vl_interface': 9.868757121888965e-06, 'learning_rate/action_model': 9.868757121888964e-05, 'epoch': 0.12}) +[2026-06-04 06:12:42,945][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.845457136631012, 'train/grad_norm_pre_clip': 2.4483954906463623, 'timing/data': 0.0003803502768278122, 'timing/model': 1.2687698639929295, 'learning_rate/qwen_vl_interface': 9.853846821107084e-06, 'learning_rate/action_model': 9.853846821107085e-05, 'epoch': 0.12}) +[2026-06-04 06:38:34,436][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state +[2026-06-04 06:38:34,444][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-04 06:38:42,592][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/model.safetensors +[2026-06-04 06:38:56,149][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/optimizer.bin +[2026-06-04 06:38:56,149][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/sampler.bin +[2026-06-04 06:38:56,149][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/sampler_1.bin +[2026-06-04 06:38:56,150][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/random_states_0.pkl +[2026-06-04 06:38:56,163][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-04 06:38:56,164][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1100 +[2026-06-04 06:38:56,165][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.7135047316551208, 'train/grad_norm_pre_clip': 1.7264782190322876, 'eval/loss': 1.0737946033477783, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 5900.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 5900.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 5900.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 5900.0, 'checkpoint/best_score': 5900.0, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.00045138224959373474, 'timing/model': 1.3953039087355137, 'learning_rate/qwen_vl_interface': 9.838147674873076e-06, 'learning_rate/action_model': 9.838147674873075e-05, 'epoch': 0.12}) +[2026-06-04 06:42:14,383][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.828750491142273, 'train/grad_norm_pre_clip': 2.2955408096313477, 'timing/data': 0.0007032696157693863, 'timing/model': 1.274525262415409, 'learning_rate/qwen_vl_interface': 9.821662373154824e-06, 'learning_rate/action_model': 9.821662373154825e-05, 'epoch': 0.12}) +[2026-06-04 06:45:33,786][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 1.0288913249969482, 'train/grad_norm_pre_clip': 2.3125154972076416, 'timing/data': 0.0007114838808774948, 'timing/model': 1.3299690261483192, 'learning_rate/qwen_vl_interface': 9.804393740623918e-06, 'learning_rate/action_model': 9.804393740623918e-05, 'epoch': 0.12}) +[2026-06-04 06:48:56,800][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 1.051693320274353, 'train/grad_norm_pre_clip': 2.2445194721221924, 'timing/data': 0.0011250954121351242, 'timing/model': 1.2783921901136637, 'learning_rate/qwen_vl_interface': 9.786344736171637e-06, 'learning_rate/action_model': 9.786344736171638e-05, 'epoch': 0.13}) +[2026-06-04 06:52:15,762][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 1.0575754642486572, 'train/grad_norm_pre_clip': 2.0875396728515625, 'timing/data': 0.00040975213050842285, 'timing/model': 1.2623671349138021, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.13}) +[2026-06-04 06:55:33,369][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.9026771187782288, 'train/grad_norm_pre_clip': 2.1379876136779785, 'timing/data': 0.00043892301619052887, 'timing/model': 1.2628702782094479, 'learning_rate/qwen_vl_interface': 9.747918115101753e-06, 'learning_rate/action_model': 9.747918115101753e-05, 'epoch': 0.13}) +[2026-06-04 06:58:53,319][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 1.0625253915786743, 'train/grad_norm_pre_clip': 2.1933019161224365, 'timing/data': 0.0007844455540180206, 'timing/model': 1.366791408509016, 'learning_rate/qwen_vl_interface': 9.727547082687849e-06, 'learning_rate/action_model': 9.72754708268785e-05, 'epoch': 0.13}) +[2026-06-04 07:02:11,945][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 1.0492218732833862, 'train/grad_norm_pre_clip': 2.351698160171509, 'timing/data': 0.0004007834941148758, 'timing/model': 1.2664682883769274, 'learning_rate/qwen_vl_interface': 9.7064088456318e-06, 'learning_rate/action_model': 9.7064088456318e-05, 'epoch': 0.13}) +[2026-06-04 07:05:27,846][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.9874382019042969, 'train/grad_norm_pre_clip': 2.019399881362915, 'timing/data': 0.0003695562481880188, 'timing/model': 1.2591134328395128, 'learning_rate/qwen_vl_interface': 9.68450702586171e-06, 'learning_rate/action_model': 9.684507025861709e-05, 'epoch': 0.13}) +[2026-06-04 07:08:47,452][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.8979684114456177, 'train/grad_norm_pre_clip': 1.8187322616577148, 'timing/data': 0.0004228353500366211, 'timing/model': 1.2696655206382275, 'learning_rate/qwen_vl_interface': 9.66184537614164e-06, 'learning_rate/action_model': 9.661845376141641e-05, 'epoch': 0.13}) +[2026-06-04 07:29:37,134][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1200 +[2026-06-04 07:29:37,137][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.7553008198738098, 'train/grad_norm_pre_clip': 1.9156776666641235, 'eval/loss': 1.03973388671875, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 5130.0, 'rl_games_eval/mid_train/mean_length': 3102.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 5130.0, 'rl_games_eval/mid_train/macro_mean_length': 3102.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 5130.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3102.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 5130.0, 'checkpoint/best_score': 5900.0, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005973819643259048, 'timing/model': 1.2659133840352297, 'learning_rate/qwen_vl_interface': 9.638427779428613e-06, 'learning_rate/action_model': 9.638427779428613e-05, 'epoch': 0.13}) +[2026-06-04 07:32:55,583][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.9972050786018372, 'train/grad_norm_pre_clip': 2.112431526184082, 'timing/data': 0.009521156549453735, 'timing/model': 1.2637834772467613, 'learning_rate/qwen_vl_interface': 9.61425824820727e-06, 'learning_rate/action_model': 9.614258248207268e-05, 'epoch': 0.13}) +[2026-06-04 07:36:13,500][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.9668149948120117, 'train/grad_norm_pre_clip': 3.034634828567505, 'timing/data': 0.01997208409011364, 'timing/model': 1.3301928080618382, 'learning_rate/qwen_vl_interface': 9.589340923802355e-06, 'learning_rate/action_model': 9.589340923802354e-05, 'epoch': 0.14}) +[2026-06-04 07:39:31,847][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.9838986992835999, 'train/grad_norm_pre_clip': 2.142796516418457, 'timing/data': 0.007943514734506607, 'timing/model': 1.2588760405778885, 'learning_rate/qwen_vl_interface': 9.563680075669138e-06, 'learning_rate/action_model': 9.563680075669137e-05, 'epoch': 0.14}) +[2026-06-04 07:42:51,592][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 1.145973801612854, 'train/grad_norm_pre_clip': 2.8372929096221924, 'timing/data': 0.0104070995002985, 'timing/model': 1.2975758258253336, 'learning_rate/qwen_vl_interface': 9.53728010066184e-06, 'learning_rate/action_model': 9.53728010066184e-05, 'epoch': 0.14}) +[2026-06-04 07:46:12,003][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.8516718149185181, 'train/grad_norm_pre_clip': 1.9302006959915161, 'timing/data': 0.009685937315225601, 'timing/model': 1.2717539239674807, 'learning_rate/qwen_vl_interface': 9.51014552228027e-06, 'learning_rate/action_model': 9.510145522280269e-05, 'epoch': 0.14}) +[2026-06-04 07:49:31,694][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 1.1830428838729858, 'train/grad_norm_pre_clip': 2.919891119003296, 'timing/data': 0.0006185770034790039, 'timing/model': 1.2693071123212576, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.14}) +[2026-06-04 07:52:50,778][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.9642070531845093, 'train/grad_norm_pre_clip': 2.5193676948547363, 'timing/data': 0.0006174296140670776, 'timing/model': 1.270923025906086, 'learning_rate/qwen_vl_interface': 9.45369127794945e-06, 'learning_rate/action_model': 9.453691277949449e-05, 'epoch': 0.14}) +[2026-06-04 07:56:09,892][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 1.1232136487960815, 'train/grad_norm_pre_clip': 2.5455403327941895, 'timing/data': 0.0006190650165081024, 'timing/model': 1.2734504919499159, 'learning_rate/qwen_vl_interface': 9.424381285144337e-06, 'learning_rate/action_model': 9.424381285144335e-05, 'epoch': 0.14}) +[2026-06-04 07:59:31,159][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 1.1721593141555786, 'train/grad_norm_pre_clip': 2.7709503173828125, 'timing/data': 0.0004619024693965912, 'timing/model': 1.2583441976457834, 'learning_rate/qwen_vl_interface': 9.394356033595787e-06, 'learning_rate/action_model': 9.394356033595787e-05, 'epoch': 0.14}) +[2026-06-04 08:24:53,475][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1300 +[2026-06-04 08:24:53,477][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.9190083146095276, 'train/grad_norm_pre_clip': 1.9793403148651123, 'eval/loss': 1.0373420715332031, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 4800.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 4800.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 4800.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 4800.0, 'checkpoint/best_score': 5900.0, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005012936890125275, 'timing/model': 1.2367014717310667, 'learning_rate/qwen_vl_interface': 9.363620667976085e-06, 'learning_rate/action_model': 9.363620667976085e-05, 'epoch': 0.14}) +[2026-06-04 08:28:13,284][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 1.1439751386642456, 'train/grad_norm_pre_clip': 2.2903389930725098, 'timing/data': 0.0004294738173484802, 'timing/model': 1.2584573421627283, 'learning_rate/qwen_vl_interface': 9.332180454631902e-06, 'learning_rate/action_model': 9.332180454631901e-05, 'epoch': 0.15}) +[2026-06-04 08:31:31,860][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 1.0083175897598267, 'train/grad_norm_pre_clip': 4.81170654296875, 'timing/data': 0.000523097813129425, 'timing/model': 1.2598868198692799, 'learning_rate/qwen_vl_interface': 9.300040780681938e-06, 'learning_rate/action_model': 9.300040780681938e-05, 'epoch': 0.15}) +[2026-06-04 08:34:50,551][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.8076463937759399, 'train/grad_norm_pre_clip': 2.955751657485962, 'timing/data': 0.0003815609961748123, 'timing/model': 1.2765190973877907, 'learning_rate/qwen_vl_interface': 9.267207153093874e-06, 'learning_rate/action_model': 9.267207153093875e-05, 'epoch': 0.15}) +[2026-06-04 08:38:09,887][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.8306744694709778, 'train/grad_norm_pre_clip': 2.6374430656433105, 'timing/data': 0.009962867945432663, 'timing/model': 1.3398895356804132, 'learning_rate/qwen_vl_interface': 9.233685197740765e-06, 'learning_rate/action_model': 9.233685197740764e-05, 'epoch': 0.15}) +[2026-06-04 08:41:30,484][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 1.1699544191360474, 'train/grad_norm_pre_clip': 2.6652934551239014, 'timing/data': 0.010006029158830643, 'timing/model': 1.3417223561555147, 'learning_rate/qwen_vl_interface': 9.199480658437092e-06, 'learning_rate/action_model': 9.199480658437091e-05, 'epoch': 0.15}) +[2026-06-04 08:44:50,987][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.9541897177696228, 'train/grad_norm_pre_clip': 2.4292893409729004, 'timing/data': 0.011961841955780983, 'timing/model': 1.3361725900322199, 'learning_rate/qwen_vl_interface': 9.164599395954575e-06, 'learning_rate/action_model': 9.164599395954574e-05, 'epoch': 0.15}) +[2026-06-04 08:48:13,566][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.8416633605957031, 'train/grad_norm_pre_clip': 2.909982442855835, 'timing/data': 0.010708088055253029, 'timing/model': 1.5133947432041168, 'learning_rate/qwen_vl_interface': 9.129047387017977e-06, 'learning_rate/action_model': 9.129047387017976e-05, 'epoch': 0.15}) +[2026-06-04 08:51:32,948][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.7393556833267212, 'train/grad_norm_pre_clip': 3.0398190021514893, 'timing/data': 0.009780088439583778, 'timing/model': 1.2560140397399664, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.15}) +[2026-06-04 08:54:53,785][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.9821939468383789, 'train/grad_norm_pre_clip': 3.0960893630981445, 'timing/data': 0.009856980293989182, 'timing/model': 1.3446891121566296, 'learning_rate/qwen_vl_interface': 9.055955610282539e-06, 'learning_rate/action_model': 9.055955610282538e-05, 'epoch': 0.15}) +[2026-06-04 09:19:31,378][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1400 +[2026-06-04 09:19:31,379][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.7465412616729736, 'train/grad_norm_pre_clip': 3.1234359741210938, 'eval/loss': 1.024303913116455, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 5130.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 5130.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 5130.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 5130.0, 'checkpoint/best_score': 5900.0, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.026012413203716278, 'timing/model': 1.3380164615809917, 'learning_rate/qwen_vl_interface': 9.018428366383368e-06, 'learning_rate/action_model': 9.018428366383368e-05, 'epoch': 0.16}) +[2026-06-04 09:22:49,153][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.8388975858688354, 'train/grad_norm_pre_clip': 2.4693520069122314, 'timing/data': 0.021010244265198708, 'timing/model': 1.2424402553588152, 'learning_rate/qwen_vl_interface': 8.98025542168354e-06, 'learning_rate/action_model': 8.98025542168354e-05, 'epoch': 0.16}) +[2026-06-04 09:26:05,886][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 1.0110273361206055, 'train/grad_norm_pre_clip': 2.5980186462402344, 'timing/data': 0.03321758098900318, 'timing/model': 1.3177833463996649, 'learning_rate/qwen_vl_interface': 8.941443316920613e-06, 'learning_rate/action_model': 8.941443316920612e-05, 'epoch': 0.16}) +[2026-06-04 09:29:24,203][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.9862690567970276, 'train/grad_norm_pre_clip': 2.184795618057251, 'timing/data': 0.0187733992934227, 'timing/model': 1.269152071326971, 'learning_rate/qwen_vl_interface': 8.901998702348934e-06, 'learning_rate/action_model': 8.901998702348933e-05, 'epoch': 0.16}) +[2026-06-04 09:32:43,803][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.9478144645690918, 'train/grad_norm_pre_clip': 3.1995272636413574, 'timing/data': 0.010553868487477303, 'timing/model': 1.269451629370451, 'learning_rate/qwen_vl_interface': 8.861928336600148e-06, 'learning_rate/action_model': 8.861928336600148e-05, 'epoch': 0.16}) +[2026-06-04 09:36:02,395][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.8790848255157471, 'train/grad_norm_pre_clip': 2.2366065979003906, 'timing/data': 0.022861428558826447, 'timing/model': 1.3434613961726427, 'learning_rate/qwen_vl_interface': 8.821239085525143e-06, 'learning_rate/action_model': 8.821239085525143e-05, 'epoch': 0.16}) +[2026-06-04 09:39:25,780][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.5728862881660461, 'train/grad_norm_pre_clip': 2.184983491897583, 'timing/data': 0.009892886504530907, 'timing/model': 1.3324896655976772, 'learning_rate/qwen_vl_interface': 8.779937921017624e-06, 'learning_rate/action_model': 8.779937921017623e-05, 'epoch': 0.16}) +[2026-06-04 09:42:49,446][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.8920238614082336, 'train/grad_norm_pre_clip': 3.331144094467163, 'timing/data': 0.015714548528194427, 'timing/model': 1.4010830447077751, 'learning_rate/qwen_vl_interface': 8.738031919819508e-06, 'learning_rate/action_model': 8.738031919819508e-05, 'epoch': 0.16}) +[2026-06-04 09:46:10,780][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.6109824776649475, 'train/grad_norm_pre_clip': 2.5438151359558105, 'timing/data': 0.010012300685048103, 'timing/model': 1.49106658808887, 'learning_rate/qwen_vl_interface': 8.695528262308367e-06, 'learning_rate/action_model': 8.695528262308366e-05, 'epoch': 0.16}) +[2026-06-04 09:49:28,683][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.8546544313430786, 'train/grad_norm_pre_clip': 3.2479565143585205, 'timing/data': 0.000464877113699913, 'timing/model': 1.2872266564518213, 'learning_rate/qwen_vl_interface': 8.65243423126711e-06, 'learning_rate/action_model': 8.652434231267109e-05, 'epoch': 0.17}) +[2026-06-04 10:07:25,622][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-04 10:07:25,625][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 1.1918035745620728, 'train/grad_norm_pre_clip': 2.9134695529937744, 'eval/loss': 0.9874222874641418, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 2220.0, 'rl_games_eval/mid_train/mean_length': 2296.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 2220.0, 'rl_games_eval/mid_train/macro_mean_length': 2296.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 2220.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 2296.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 2220.0, 'checkpoint/best_score': 5900.0, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.016495423391461372, 'timing/model': 1.3365559037774801, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.17}) +[2026-06-04 10:10:45,690][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.8401406407356262, 'train/grad_norm_pre_clip': 2.5472047328948975, 'timing/data': 0.019424183294177055, 'timing/model': 1.3501464053988457, 'learning_rate/qwen_vl_interface': 8.564504684247982e-06, 'learning_rate/action_model': 8.564504684247982e-05, 'epoch': 0.17}) +[2026-06-04 10:14:05,661][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 1.0003323554992676, 'train/grad_norm_pre_clip': 2.3142497539520264, 'timing/data': 0.017594391480088234, 'timing/model': 1.2399805970489979, 'learning_rate/qwen_vl_interface': 8.519684234545333e-06, 'learning_rate/action_model': 8.519684234545332e-05, 'epoch': 0.17}) +[2026-06-04 10:17:22,085][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 1.2332319021224976, 'train/grad_norm_pre_clip': 2.9105329513549805, 'timing/data': 0.017634332180023193, 'timing/model': 1.2866554986685514, 'learning_rate/qwen_vl_interface': 8.474303541281474e-06, 'learning_rate/action_model': 8.474303541281474e-05, 'epoch': 0.17}) +[2026-06-04 10:20:45,793][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 1.1004583835601807, 'train/grad_norm_pre_clip': 3.28629469871521, 'timing/data': 0.01841585338115692, 'timing/model': 1.3199475761502981, 'learning_rate/qwen_vl_interface': 8.428370380204577e-06, 'learning_rate/action_model': 8.428370380204577e-05, 'epoch': 0.17}) +[2026-06-04 10:24:06,476][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.9535146355628967, 'train/grad_norm_pre_clip': 2.7233386039733887, 'timing/data': 0.017403125762939453, 'timing/model': 1.3328241594135761, 'learning_rate/qwen_vl_interface': 8.381892621725327e-06, 'learning_rate/action_model': 8.381892621725327e-05, 'epoch': 0.17}) +[2026-06-04 10:27:28,995][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 1.198813557624817, 'train/grad_norm_pre_clip': 3.4130232334136963, 'timing/data': 0.019742760807275772, 'timing/model': 1.2811436224728823, 'learning_rate/qwen_vl_interface': 8.334878229568372e-06, 'learning_rate/action_model': 8.334878229568372e-05, 'epoch': 0.17}) +[2026-06-04 10:30:46,457][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 1.081758737564087, 'train/grad_norm_pre_clip': 3.5178089141845703, 'timing/data': 0.014709312468767166, 'timing/model': 1.2359795924276114, 'learning_rate/qwen_vl_interface': 8.287335259407779e-06, 'learning_rate/action_model': 8.287335259407779e-05, 'epoch': 0.17}) +[2026-06-04 10:34:08,984][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.9589731097221375, 'train/grad_norm_pre_clip': 2.6095962524414062, 'timing/data': 0.017936259508132935, 'timing/model': 1.3453267645090818, 'learning_rate/qwen_vl_interface': 8.23927185748673e-06, 'learning_rate/action_model': 8.239271857486729e-05, 'epoch': 0.18}) +[2026-06-04 10:37:32,492][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.8997480273246765, 'train/grad_norm_pre_clip': 2.955718755722046, 'timing/data': 0.019552571699023247, 'timing/model': 1.3475173469632864, 'learning_rate/qwen_vl_interface': 8.190696259221713e-06, 'learning_rate/action_model': 8.190696259221713e-05, 'epoch': 0.18}) +[2026-06-04 11:02:22,772][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1600 +[2026-06-04 11:02:22,774][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.7957470417022705, 'train/grad_norm_pre_clip': 2.9438929557800293, 'eval/loss': 0.9512646794319153, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 5060.0, 'rl_games_eval/mid_train/mean_length': 3583.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 5060.0, 'rl_games_eval/mid_train/macro_mean_length': 3583.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 5060.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3583.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 5060.0, 'checkpoint/best_score': 5900.0, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.016137143597006798, 'timing/model': 1.3391963839530945, 'learning_rate/qwen_vl_interface': 8.141616787791425e-06, 'learning_rate/action_model': 8.141616787791424e-05, 'epoch': 0.18}) +[2026-06-04 11:05:44,436][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.8392089605331421, 'train/grad_norm_pre_clip': 2.4733405113220215, 'timing/data': 0.029168713837862015, 'timing/model': 1.335634920746088, 'learning_rate/qwen_vl_interface': 8.092041852710631e-06, 'learning_rate/action_model': 8.092041852710631e-05, 'epoch': 0.18}) +[2026-06-04 11:09:01,089][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 1.0887020826339722, 'train/grad_norm_pre_clip': 3.787381649017334, 'timing/data': 0.009629158303141594, 'timing/model': 1.2648344207555056, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.18}) +[2026-06-04 11:12:25,822][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 1.2598882913589478, 'train/grad_norm_pre_clip': 3.2068145275115967, 'timing/data': 0.01915682852268219, 'timing/model': 1.2875305395573378, 'learning_rate/qwen_vl_interface': 7.991439652676861e-06, 'learning_rate/action_model': 7.991439652676861e-05, 'epoch': 0.18}) +[2026-06-04 11:15:46,807][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.9330452680587769, 'train/grad_norm_pre_clip': 2.8879590034484863, 'timing/data': 0.019623184576630592, 'timing/model': 1.3390327394008636, 'learning_rate/qwen_vl_interface': 7.940429625392957e-06, 'learning_rate/action_model': 7.940429625392956e-05, 'epoch': 0.18}) +[2026-06-04 11:19:06,966][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 1.3272175788879395, 'train/grad_norm_pre_clip': 2.6185991764068604, 'timing/data': 0.015509193763136864, 'timing/model': 1.2362253293395042, 'learning_rate/qwen_vl_interface': 7.888958606843111e-06, 'learning_rate/action_model': 7.888958606843111e-05, 'epoch': 0.18}) +[2026-06-04 11:22:27,595][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 1.0225974321365356, 'train/grad_norm_pre_clip': 2.900954246520996, 'timing/data': 0.018817853182554245, 'timing/model': 1.3220842722803354, 'learning_rate/qwen_vl_interface': 7.83703541632138e-06, 'learning_rate/action_model': 7.83703541632138e-05, 'epoch': 0.18}) +[2026-06-04 11:25:46,098][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 1.2190049886703491, 'train/grad_norm_pre_clip': 3.1097898483276367, 'timing/data': 0.01691223494708538, 'timing/model': 1.2688604015856981, 'learning_rate/qwen_vl_interface': 7.784668950599155e-06, 'learning_rate/action_model': 7.784668950599155e-05, 'epoch': 0.19}) +[2026-06-04 11:29:02,965][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.7363961338996887, 'train/grad_norm_pre_clip': 2.3370859622955322, 'timing/data': 0.014405488967895508, 'timing/model': 1.2518120259046555, 'learning_rate/qwen_vl_interface': 7.731868182400758e-06, 'learning_rate/action_model': 7.731868182400758e-05, 'epoch': 0.19}) +[2026-06-04 11:32:25,639][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.6980933547019958, 'train/grad_norm_pre_clip': 2.278989315032959, 'timing/data': 0.020901361480355263, 'timing/model': 1.3891507554799318, 'learning_rate/qwen_vl_interface': 7.678642158865997e-06, 'learning_rate/action_model': 7.678642158865997e-05, 'epoch': 0.19}) +[2026-06-04 11:57:04,333][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1700 +[2026-06-04 11:57:04,335][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.7380145192146301, 'train/grad_norm_pre_clip': 2.7244462966918945, 'eval/loss': 0.944369912147522, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 5030.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 5030.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 5030.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 5030.0, 'checkpoint/best_score': 5900.0, 'checkpoint/best_step': 1100.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.01911613717675209, 'timing/model': 1.2683978080749512, 'learning_rate/qwen_vl_interface': 7.625e-06, 'learning_rate/action_model': 7.625e-05, 'epoch': 0.19}) +[2026-06-04 12:00:24,954][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 1.4718670845031738, 'train/grad_norm_pre_clip': 3.8549447059631348, 'timing/data': 0.015484845265746117, 'timing/model': 1.2431433629244566, 'learning_rate/qwen_vl_interface': 7.570950897110537e-06, 'learning_rate/action_model': 7.570950897110536e-05, 'epoch': 0.19}) +[2026-06-04 12:03:39,938][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.9214416146278381, 'train/grad_norm_pre_clip': 3.4755568504333496, 'timing/data': 0.008702483028173447, 'timing/model': 1.2544866353273392, 'learning_rate/qwen_vl_interface': 7.516504111233141e-06, 'learning_rate/action_model': 7.51650411123314e-05, 'epoch': 0.19}) +[2026-06-04 12:07:00,583][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.929242730140686, 'train/grad_norm_pre_clip': 2.9087929725646973, 'timing/data': 0.010718511417508125, 'timing/model': 1.343903498724103, 'learning_rate/qwen_vl_interface': 7.46166897154428e-06, 'learning_rate/action_model': 7.46166897154428e-05, 'epoch': 0.19}) +[2026-06-04 12:10:24,401][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.9578375816345215, 'train/grad_norm_pre_clip': 2.559506416320801, 'timing/data': 0.0007240995764732361, 'timing/model': 1.2704914677888155, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.19}) +[2026-06-04 12:13:46,707][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.981570839881897, 'train/grad_norm_pre_clip': 2.437927484512329, 'timing/data': 0.0007315278053283691, 'timing/model': 1.2815654836595058, 'learning_rate/qwen_vl_interface': 7.350871278540257e-06, 'learning_rate/action_model': 7.350871278540256e-05, 'epoch': 0.19}) +[2026-06-04 12:17:07,802][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.7133089900016785, 'train/grad_norm_pre_clip': 2.7919650077819824, 'timing/data': 0.0006783846765756607, 'timing/model': 1.2744118850678205, 'learning_rate/qwen_vl_interface': 7.2949277098394036e-06, 'learning_rate/action_model': 7.294927709839403e-05, 'epoch': 0.2}) +[2026-06-04 12:20:25,020][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 1.1148275136947632, 'train/grad_norm_pre_clip': 3.5841994285583496, 'timing/data': 0.0006941892206668854, 'timing/model': 1.3446183688938618, 'learning_rate/qwen_vl_interface': 7.238633753302784e-06, 'learning_rate/action_model': 7.238633753302784e-05, 'epoch': 0.2}) +[2026-06-04 12:23:45,157][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.7623109221458435, 'train/grad_norm_pre_clip': 2.5786068439483643, 'timing/data': 0.000535178929567337, 'timing/model': 1.2777779288589954, 'learning_rate/qwen_vl_interface': 7.181999054610053e-06, 'learning_rate/action_model': 7.181999054610052e-05, 'epoch': 0.2}) +[2026-06-04 12:27:05,691][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 1.0338505506515503, 'train/grad_norm_pre_clip': 2.843663454055786, 'timing/data': 0.0006613396108150482, 'timing/model': 1.2793287634849548, 'learning_rate/qwen_vl_interface': 7.125033317825269e-06, 'learning_rate/action_model': 7.12503331782527e-05, 'epoch': 0.2}) +[2026-06-04 12:52:57,339][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state +[2026-06-04 12:52:57,349][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-04 12:53:05,594][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/model.safetensors +[2026-06-04 12:53:18,762][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/optimizer.bin +[2026-06-04 12:53:18,763][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/sampler.bin +[2026-06-04 12:53:18,764][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/sampler_1.bin +[2026-06-04 12:53:18,765][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/random_states_0.pkl +[2026-06-04 12:53:18,780][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-04 12:53:18,780][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1800 +[2026-06-04 12:53:18,782][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.9009042978286743, 'train/grad_norm_pre_clip': 2.4858028888702393, 'eval/loss': 0.9060004353523254, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 6580.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 6580.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 6580.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 6580.0, 'checkpoint/best_score': 6580.0, 'checkpoint/best_step': 1800.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.0009029675275087357, 'timing/model': 1.285961538553238, 'learning_rate/qwen_vl_interface': 7.067746303734178e-06, 'learning_rate/action_model': 7.067746303734177e-05, 'epoch': 0.2}) +[2026-06-04 12:56:39,502][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.8249101042747498, 'train/grad_norm_pre_clip': 3.73514723777771, 'timing/data': 0.0015406515449285507, 'timing/model': 1.6394453328102827, 'learning_rate/qwen_vl_interface': 7.0101478281717235e-06, 'learning_rate/action_model': 7.010147828171724e-05, 'epoch': 0.2}) +[2026-06-04 13:00:01,920][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.9991025924682617, 'train/grad_norm_pre_clip': 3.217501163482666, 'timing/data': 0.0007183067500591278, 'timing/model': 1.2958135269582272, 'learning_rate/qwen_vl_interface': 6.952247760340176e-06, 'learning_rate/action_model': 6.952247760340176e-05, 'epoch': 0.2}) +[2026-06-04 13:03:26,468][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.750747799873352, 'train/grad_norm_pre_clip': 2.680713415145874, 'timing/data': 0.0006607584655284882, 'timing/model': 1.3322607409209013, 'learning_rate/qwen_vl_interface': 6.894056021118092e-06, 'learning_rate/action_model': 6.894056021118091e-05, 'epoch': 0.2}) +[2026-06-04 13:06:48,140][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.6839194297790527, 'train/grad_norm_pre_clip': 2.191648006439209, 'timing/data': 0.000703275203704834, 'timing/model': 1.411639852449298, 'learning_rate/qwen_vl_interface': 6.835582581360413e-06, 'learning_rate/action_model': 6.835582581360412e-05, 'epoch': 0.2}) +[2026-06-04 13:10:08,231][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.914818525314331, 'train/grad_norm_pre_clip': 2.6126768589019775, 'timing/data': 0.0006951838731765747, 'timing/model': 1.3762181773781776, 'learning_rate/qwen_vl_interface': 6.7768374601900174e-06, 'learning_rate/action_model': 6.776837460190018e-05, 'epoch': 0.21}) +[2026-06-04 13:13:30,364][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.9414860010147095, 'train/grad_norm_pre_clip': 2.510528087615967, 'timing/data': 0.00046617165207862854, 'timing/model': 1.2789645865559578, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.21}) +[2026-06-04 13:16:51,504][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 1.302306056022644, 'train/grad_norm_pre_clip': 3.1260077953338623, 'timing/data': 0.0024220123887062073, 'timing/model': 1.270872674882412, 'learning_rate/qwen_vl_interface': 6.658572481133962e-06, 'learning_rate/action_model': 6.658572481133962e-05, 'epoch': 0.21}) +[2026-06-04 13:20:12,811][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 1.3460915088653564, 'train/grad_norm_pre_clip': 3.254570484161377, 'timing/data': 0.0007195230573415756, 'timing/model': 1.302093431353569, 'learning_rate/qwen_vl_interface': 6.599072887343634e-06, 'learning_rate/action_model': 6.599072887343633e-05, 'epoch': 0.21}) +[2026-06-04 13:23:34,011][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 1.1189167499542236, 'train/grad_norm_pre_clip': 3.0059847831726074, 'timing/data': 0.0007011964917182922, 'timing/model': 1.3533170241862535, 'learning_rate/qwen_vl_interface': 6.539342136859104e-06, 'learning_rate/action_model': 6.539342136859104e-05, 'epoch': 0.21}) +[2026-06-04 13:49:33,424][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state +[2026-06-04 13:49:33,434][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-04 13:49:41,615][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/model.safetensors +[2026-06-04 13:49:52,430][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/optimizer.bin +[2026-06-04 13:49:52,431][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/sampler.bin +[2026-06-04 13:49:52,431][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/sampler_1.bin +[2026-06-04 13:49:52,432][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state/random_states_0.pkl +[2026-06-04 13:49:52,447][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-04 13:49:52,447][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1900 +[2026-06-04 13:49:52,449][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.580230176448822, 'train/grad_norm_pre_clip': 2.2984418869018555, 'eval/loss': 0.9164230227470398, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 7770.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 7770.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 7770.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 7770.0, 'checkpoint/best_score': 7770.0, 'checkpoint/best_step': 1900.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.000687427818775177, 'timing/model': 1.4160834401845932, 'learning_rate/qwen_vl_interface': 6.479390464236974e-06, 'learning_rate/action_model': 6.479390464236974e-05, 'epoch': 0.21}) +[2026-06-04 13:53:09,950][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.7031666040420532, 'train/grad_norm_pre_clip': 2.752166748046875, 'timing/data': 0.0004530828446149826, 'timing/model': 1.2708131726831198, 'learning_rate/qwen_vl_interface': 6.4192281418877174e-06, 'learning_rate/action_model': 6.419228141887717e-05, 'epoch': 0.21}) +[2026-06-04 13:56:28,888][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.9942619800567627, 'train/grad_norm_pre_clip': 3.058194398880005, 'timing/data': 0.0006758105009794235, 'timing/model': 1.2706419825553894, 'learning_rate/qwen_vl_interface': 6.3588654783155525e-06, 'learning_rate/action_model': 6.358865478315553e-05, 'epoch': 0.21}) +[2026-06-04 13:59:51,483][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 1.166106939315796, 'train/grad_norm_pre_clip': 3.269730806350708, 'timing/data': 0.010431868955492973, 'timing/model': 1.3383232466876507, 'learning_rate/qwen_vl_interface': 6.298312816352132e-06, 'learning_rate/action_model': 6.298312816352132e-05, 'epoch': 0.21}) +[2026-06-04 14:03:10,747][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.8255189061164856, 'train/grad_norm_pre_clip': 3.425506591796875, 'timing/data': 0.00933094508945942, 'timing/model': 1.2509330306202173, 'learning_rate/qwen_vl_interface': 6.2375805313843564e-06, 'learning_rate/action_model': 6.237580531384356e-05, 'epoch': 0.22}) +[2026-06-04 14:06:28,091][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.8725265264511108, 'train/grad_norm_pre_clip': 3.0136630535125732, 'timing/data': 0.010056111961603165, 'timing/model': 1.271705836057663, 'learning_rate/qwen_vl_interface': 6.176679029576611e-06, 'learning_rate/action_model': 6.17667902957661e-05, 'epoch': 0.22}) +[2026-06-04 14:09:49,308][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.9762049913406372, 'train/grad_norm_pre_clip': 3.4884190559387207, 'timing/data': 0.014675308018922806, 'timing/model': 1.3620314411818981, 'learning_rate/qwen_vl_interface': 6.115618746087702e-06, 'learning_rate/action_model': 6.115618746087703e-05, 'epoch': 0.22}) +[2026-06-04 14:13:09,795][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 1.0804774761199951, 'train/grad_norm_pre_clip': 3.504546642303467, 'timing/data': 0.009882744401693344, 'timing/model': 1.268959240987897, 'learning_rate/qwen_vl_interface': 6.0544101432828676e-06, 'learning_rate/action_model': 6.054410143282867e-05, 'epoch': 0.22}) +[2026-06-04 14:16:29,545][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.9439980983734131, 'train/grad_norm_pre_clip': 2.7177412509918213, 'timing/data': 0.009419361129403114, 'timing/model': 1.264014633372426, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.22}) +[2026-06-04 14:19:51,613][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.7825664281845093, 'train/grad_norm_pre_clip': 2.7347095012664795, 'timing/data': 0.011446136981248856, 'timing/model': 1.505872344598174, 'learning_rate/qwen_vl_interface': 5.9315899544581015e-06, 'learning_rate/action_model': 5.931589954458101e-05, 'epoch': 0.22}) +[2026-06-04 14:40:39,752][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2000 +[2026-06-04 14:40:39,754][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 1.1119128465652466, 'train/grad_norm_pre_clip': 2.8813459873199463, 'eval/loss': 0.8616499304771423, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 4180.0, 'rl_games_eval/mid_train/mean_length': 2839.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 4180.0, 'rl_games_eval/mid_train/macro_mean_length': 2839.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 4180.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 2839.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 4180.0, 'checkpoint/best_score': 7770.0, 'checkpoint/best_step': 1900.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.011173456907272339, 'timing/model': 1.3076740689575672, 'learning_rate/qwen_vl_interface': 5.869999413045247e-06, 'learning_rate/action_model': 5.869999413045246e-05, 'epoch': 0.22}) +[2026-06-04 14:43:59,405][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 1.2604873180389404, 'train/grad_norm_pre_clip': 4.330149173736572, 'timing/data': 0.010708635672926903, 'timing/model': 1.2700647339224815, 'learning_rate/qwen_vl_interface': 5.80830263792473e-06, 'learning_rate/action_model': 5.80830263792473e-05, 'epoch': 0.22}) +[2026-06-04 14:47:20,777][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.8908256888389587, 'train/grad_norm_pre_clip': 3.01169490814209, 'timing/data': 0.025817207992076874, 'timing/model': 1.3175152726471424, 'learning_rate/qwen_vl_interface': 5.746510200521354e-06, 'learning_rate/action_model': 5.746510200521355e-05, 'epoch': 0.22}) +[2026-06-04 14:50:42,207][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.7952753901481628, 'train/grad_norm_pre_clip': 2.8077399730682373, 'timing/data': 0.025494586676359177, 'timing/model': 1.2726565524935722, 'learning_rate/qwen_vl_interface': 5.684632688651163e-06, 'learning_rate/action_model': 5.6846326886511624e-05, 'epoch': 0.23}) +[2026-06-04 14:54:04,560][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 1.0861904621124268, 'train/grad_norm_pre_clip': 3.969222068786621, 'timing/data': 0.021735118702054024, 'timing/model': 1.2417683191597462, 'learning_rate/qwen_vl_interface': 5.622680704707265e-06, 'learning_rate/action_model': 5.622680704707265e-05, 'epoch': 0.23}) +[2026-06-04 14:57:23,673][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 1.2390413284301758, 'train/grad_norm_pre_clip': 3.7737843990325928, 'timing/data': 0.02705005556344986, 'timing/model': 1.3080512080341578, 'learning_rate/qwen_vl_interface': 5.560664863843181e-06, 'learning_rate/action_model': 5.56066486384318e-05, 'epoch': 0.23}) +[2026-06-04 15:00:43,679][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 1.3200844526290894, 'train/grad_norm_pre_clip': 4.597832679748535, 'timing/data': 0.02445540390908718, 'timing/model': 1.3384617790579796, 'learning_rate/qwen_vl_interface': 5.498595792153985e-06, 'learning_rate/action_model': 5.498595792153984e-05, 'epoch': 0.23}) +[2026-06-04 15:04:04,292][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 1.1614068746566772, 'train/grad_norm_pre_clip': 3.206017255783081, 'timing/data': 0.023353857919573784, 'timing/model': 1.3244519513100386, 'learning_rate/qwen_vl_interface': 5.436484124855577e-06, 'learning_rate/action_model': 5.4364841248555764e-05, 'epoch': 0.23}) +[2026-06-04 15:07:28,174][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.7637149095535278, 'train/grad_norm_pre_clip': 3.0014286041259766, 'timing/data': 0.024267930537462234, 'timing/model': 1.2448030393570662, 'learning_rate/qwen_vl_interface': 5.374340504462399e-06, 'learning_rate/action_model': 5.3743405044623984e-05, 'epoch': 0.23}) +[2026-06-04 15:10:47,185][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 1.1720136404037476, 'train/grad_norm_pre_clip': 4.481629371643066, 'timing/data': 0.026362890377640724, 'timing/model': 1.3174332045018673, 'learning_rate/qwen_vl_interface': 5.312175578963886e-06, 'learning_rate/action_model': 5.312175578963886e-05, 'epoch': 0.23}) +[2026-06-04 15:36:41,576][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2100 +[2026-06-04 15:36:41,580][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.7360332608222961, 'train/grad_norm_pre_clip': 2.9418420791625977, 'eval/loss': 0.8540701866149902, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 6260.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 6260.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 6260.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 6260.0, 'checkpoint/best_score': 7770.0, 'checkpoint/best_step': 1900.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0278946366161108, 'timing/model': 1.3055327162146568, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.23}) +[2026-06-04 15:39:59,978][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.6876044273376465, 'train/grad_norm_pre_clip': 2.656952142715454, 'timing/data': 0.025316476821899414, 'timing/model': 1.3379525989294052, 'learning_rate/qwen_vl_interface': 5.187824421036115e-06, 'learning_rate/action_model': 5.187824421036115e-05, 'epoch': 0.23}) +[2026-06-04 15:43:16,300][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.846710741519928, 'train/grad_norm_pre_clip': 2.874845504760742, 'timing/data': 0.026700470596551895, 'timing/model': 1.2667481768876314, 'learning_rate/qwen_vl_interface': 5.125659495537603e-06, 'learning_rate/action_model': 5.1256594955376034e-05, 'epoch': 0.24}) +[2026-06-04 15:46:32,664][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 1.0018244981765747, 'train/grad_norm_pre_clip': 3.951949119567871, 'timing/data': 0.01606181263923645, 'timing/model': 1.2569620124995708, 'learning_rate/qwen_vl_interface': 5.063515875144426e-06, 'learning_rate/action_model': 5.0635158751444254e-05, 'epoch': 0.24}) +[2026-06-04 15:49:49,056][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.9199546575546265, 'train/grad_norm_pre_clip': 3.321138858795166, 'timing/data': 0.02244739420711994, 'timing/model': 1.244545305147767, 'learning_rate/qwen_vl_interface': 5.001404207846017e-06, 'learning_rate/action_model': 5.001404207846017e-05, 'epoch': 0.24}) +[2026-06-04 15:53:11,306][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.9985020160675049, 'train/grad_norm_pre_clip': 2.710228204727173, 'timing/data': 0.027760939672589302, 'timing/model': 1.3460045531392097, 'learning_rate/qwen_vl_interface': 4.93933513615682e-06, 'learning_rate/action_model': 4.9393351361568204e-05, 'epoch': 0.24}) +[2026-06-04 15:56:27,487][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 1.116977572441101, 'train/grad_norm_pre_clip': 11.212539672851562, 'timing/data': 0.02582492120563984, 'timing/model': 1.3315656874328852, 'learning_rate/qwen_vl_interface': 4.877319295292737e-06, 'learning_rate/action_model': 4.8773192952927374e-05, 'epoch': 0.24}) +[2026-06-04 15:59:52,383][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.8332566022872925, 'train/grad_norm_pre_clip': 3.065303087234497, 'timing/data': 0.025500524789094925, 'timing/model': 1.3310018684715033, 'learning_rate/qwen_vl_interface': 4.815367311348839e-06, 'learning_rate/action_model': 4.815367311348839e-05, 'epoch': 0.24}) +[2026-06-04 16:03:12,869][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.9687777161598206, 'train/grad_norm_pre_clip': 3.4042844772338867, 'timing/data': 0.024409107863903046, 'timing/model': 1.3293975554406643, 'learning_rate/qwen_vl_interface': 4.753489799478647e-06, 'learning_rate/action_model': 4.7534897994786464e-05, 'epoch': 0.24}) +[2026-06-04 16:06:30,766][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.7679760456085205, 'train/grad_norm_pre_clip': 2.860853433609009, 'timing/data': 0.030599860474467278, 'timing/model': 1.362842632457614, 'learning_rate/qwen_vl_interface': 4.691697362075272e-06, 'learning_rate/action_model': 4.691697362075272e-05, 'epoch': 0.24}) +[2026-06-04 16:31:14,022][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2200 +[2026-06-04 16:31:14,025][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.8618195652961731, 'train/grad_norm_pre_clip': 4.292801856994629, 'eval/loss': 0.8674957156181335, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 6270.0, 'rl_games_eval/mid_train/mean_length': 3485.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 6270.0, 'rl_games_eval/mid_train/macro_mean_length': 3485.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 6270.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3485.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 6270.0, 'checkpoint/best_score': 7770.0, 'checkpoint/best_step': 1900.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.019310783594846725, 'timing/model': 1.3410650845617056, 'learning_rate/qwen_vl_interface': 4.630000586954756e-06, 'learning_rate/action_model': 4.630000586954756e-05, 'epoch': 0.24}) +[2026-06-04 16:34:36,395][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.7022800445556641, 'train/grad_norm_pre_clip': 2.6530699729919434, 'timing/data': 0.01923842914402485, 'timing/model': 1.3507671281695366, 'learning_rate/qwen_vl_interface': 4.568410045541899e-06, 'learning_rate/action_model': 4.5684100455418986e-05, 'epoch': 0.25}) +[2026-06-04 16:38:00,491][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.8963400721549988, 'train/grad_norm_pre_clip': 3.4862725734710693, 'timing/data': 0.018169084563851357, 'timing/model': 1.351500852033496, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 0.25}) +[2026-06-04 16:41:25,962][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.8252272605895996, 'train/grad_norm_pre_clip': 3.7637670040130615, 'timing/data': 0.022754155099391937, 'timing/model': 1.3306910693645477, 'learning_rate/qwen_vl_interface': 4.4455898567171336e-06, 'learning_rate/action_model': 4.4455898567171334e-05, 'epoch': 0.25}) +[2026-06-04 16:44:46,884][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.8667508959770203, 'train/grad_norm_pre_clip': 3.6343252658843994, 'timing/data': 0.01818365976214409, 'timing/model': 1.332041148096323, 'learning_rate/qwen_vl_interface': 4.384381253912299e-06, 'learning_rate/action_model': 4.3843812539122994e-05, 'epoch': 0.25}) +[2026-06-04 16:48:10,085][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.9308594465255737, 'train/grad_norm_pre_clip': 3.329355001449585, 'timing/data': 0.019430959597229958, 'timing/model': 1.347668269649148, 'learning_rate/qwen_vl_interface': 4.323320970423391e-06, 'learning_rate/action_model': 4.3233209704233915e-05, 'epoch': 0.25}) +[2026-06-04 16:51:31,240][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.8398228883743286, 'train/grad_norm_pre_clip': 5.395533561706543, 'timing/data': 0.023561585694551468, 'timing/model': 1.362629909068346, 'learning_rate/qwen_vl_interface': 4.262419468615644e-06, 'learning_rate/action_model': 4.2624194686156435e-05, 'epoch': 0.25}) +[2026-06-04 16:54:52,427][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.7638248205184937, 'train/grad_norm_pre_clip': 2.998300790786743, 'timing/data': 0.018880901858210564, 'timing/model': 1.4805614203214645, 'learning_rate/qwen_vl_interface': 4.201687183647871e-06, 'learning_rate/action_model': 4.201687183647871e-05, 'epoch': 0.25}) +[2026-06-04 16:58:13,588][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.8860207796096802, 'train/grad_norm_pre_clip': 3.204102039337158, 'timing/data': 0.01989738643169403, 'timing/model': 1.3242192473262548, 'learning_rate/qwen_vl_interface': 4.14113452168445e-06, 'learning_rate/action_model': 4.1411345216844504e-05, 'epoch': 0.25}) +[2026-06-04 17:01:35,150][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.828048586845398, 'train/grad_norm_pre_clip': 13.497666358947754, 'timing/data': 0.01569351553916931, 'timing/model': 1.2416740730404854, 'learning_rate/qwen_vl_interface': 4.080771858112283e-06, 'learning_rate/action_model': 4.0807718581122825e-05, 'epoch': 0.25}) +[2026-06-04 17:28:20,620][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2300 +[2026-06-04 17:28:20,623][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.915655791759491, 'train/grad_norm_pre_clip': 2.9710941314697266, 'eval/loss': 0.8518199920654297, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 4730.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 4730.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 4730.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 4730.0, 'checkpoint/best_score': 7770.0, 'checkpoint/best_step': 1900.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.017836734652519226, 'timing/model': 1.5168720129877329, 'learning_rate/qwen_vl_interface': 4.0206095357630256e-06, 'learning_rate/action_model': 4.020609535763026e-05, 'epoch': 0.26}) +[2026-06-04 17:31:40,872][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.8666853904724121, 'train/grad_norm_pre_clip': 3.2819652557373047, 'timing/data': 0.018817776814103127, 'timing/model': 1.3365663643926382, 'learning_rate/qwen_vl_interface': 3.960657863140897e-06, 'learning_rate/action_model': 3.9606578631408975e-05, 'epoch': 0.26}) +[2026-06-04 17:35:01,200][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.8970679044723511, 'train/grad_norm_pre_clip': 3.5800209045410156, 'timing/data': 0.016893407329916954, 'timing/model': 1.2697601467370987, 'learning_rate/qwen_vl_interface': 3.900927112656368e-06, 'learning_rate/action_model': 3.9009271126563677e-05, 'epoch': 0.26}) +[2026-06-04 17:38:21,839][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.9553978443145752, 'train/grad_norm_pre_clip': 3.6226558685302734, 'timing/data': 0.027393611147999763, 'timing/model': 1.3389233220368624, 'learning_rate/qwen_vl_interface': 3.841427518866038e-06, 'learning_rate/action_model': 3.841427518866038e-05, 'epoch': 0.26}) +[2026-06-04 17:41:44,188][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.8993485569953918, 'train/grad_norm_pre_clip': 4.700336456298828, 'timing/data': 0.019007880240678787, 'timing/model': 1.3310849610716105, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 0.26}) +[2026-06-04 17:45:05,508][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.7908099889755249, 'train/grad_norm_pre_clip': 3.3529744148254395, 'timing/data': 0.016343416646122932, 'timing/model': 1.2754704523831606, 'learning_rate/qwen_vl_interface': 3.7231625398099837e-06, 'learning_rate/action_model': 3.723162539809983e-05, 'epoch': 0.26}) +[2026-06-04 17:48:27,184][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.9297747015953064, 'train/grad_norm_pre_clip': 4.577991008758545, 'timing/data': 0.017667947337031364, 'timing/model': 1.3545012455433607, 'learning_rate/qwen_vl_interface': 3.6644174186395885e-06, 'learning_rate/action_model': 3.664417418639589e-05, 'epoch': 0.26}) +[2026-06-04 17:51:47,884][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.7703052759170532, 'train/grad_norm_pre_clip': 3.965754747390747, 'timing/data': 0.01921870745718479, 'timing/model': 1.3373657241463661, 'learning_rate/qwen_vl_interface': 3.605943978881907e-06, 'learning_rate/action_model': 3.605943978881907e-05, 'epoch': 0.26}) +[2026-06-04 17:55:10,772][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 1.0846166610717773, 'train/grad_norm_pre_clip': 4.037821292877197, 'timing/data': 0.01693817600607872, 'timing/model': 1.4392367769032717, 'learning_rate/qwen_vl_interface': 3.5477522396598236e-06, 'learning_rate/action_model': 3.5477522396598235e-05, 'epoch': 0.26}) +[2026-06-04 17:58:36,288][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.8678255081176758, 'train/grad_norm_pre_clip': 3.6390109062194824, 'timing/data': 0.01746991090476513, 'timing/model': 1.3430105168372393, 'learning_rate/qwen_vl_interface': 3.489852171828278e-06, 'learning_rate/action_model': 3.4898521718282776e-05, 'epoch': 0.27}) +[2026-06-04 18:26:02,008][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2400 +[2026-06-04 18:26:02,011][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.6743080615997314, 'train/grad_norm_pre_clip': 2.8486380577087402, 'eval/loss': 0.8577606081962585, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 6120.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 6120.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 6120.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 6120.0, 'checkpoint/best_score': 7770.0, 'checkpoint/best_step': 1900.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.020075315609574318, 'timing/model': 1.2652978021651506, 'learning_rate/qwen_vl_interface': 3.432253696265824e-06, 'learning_rate/action_model': 3.432253696265824e-05, 'epoch': 0.27}) +[2026-06-04 18:29:23,893][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.7414771914482117, 'train/grad_norm_pre_clip': 4.09636116027832, 'timing/data': 0.01865837164223194, 'timing/model': 1.3506761118769646, 'learning_rate/qwen_vl_interface': 3.3749666821747314e-06, 'learning_rate/action_model': 3.374966682174731e-05, 'epoch': 0.27}) +[2026-06-04 18:32:44,014][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.4498901665210724, 'train/grad_norm_pre_clip': 2.3113250732421875, 'timing/data': 0.01805129647254944, 'timing/model': 1.458352955058217, 'learning_rate/qwen_vl_interface': 3.3180009453899496e-06, 'learning_rate/action_model': 3.318000945389949e-05, 'epoch': 0.27}) +[2026-06-04 18:36:07,596][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.8840583562850952, 'train/grad_norm_pre_clip': 3.8463499546051025, 'timing/data': 0.020080747082829475, 'timing/model': 1.2767888307571411, 'learning_rate/qwen_vl_interface': 3.2613662466972178e-06, 'learning_rate/action_model': 3.2613662466972176e-05, 'epoch': 0.27}) +[2026-06-04 18:39:28,945][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.6765887141227722, 'train/grad_norm_pre_clip': 2.895113945007324, 'timing/data': 0.015237180516123772, 'timing/model': 1.2472587823867798, 'learning_rate/qwen_vl_interface': 3.2050722901605967e-06, 'learning_rate/action_model': 3.205072290160597e-05, 'epoch': 0.27}) +[2026-06-04 18:42:49,074][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.865429699420929, 'train/grad_norm_pre_clip': 3.5887203216552734, 'timing/data': 0.010600576177239418, 'timing/model': 1.328432273119688, 'learning_rate/qwen_vl_interface': 3.149128721459745e-06, 'learning_rate/action_model': 3.149128721459745e-05, 'epoch': 0.27}) +[2026-06-04 18:46:10,300][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.9763443470001221, 'train/grad_norm_pre_clip': 5.250875949859619, 'timing/data': 0.017285771667957306, 'timing/model': 1.3514349516481161, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 0.27}) +[2026-06-04 18:49:32,807][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.8187896013259888, 'train/grad_norm_pre_clip': 3.4958736896514893, 'timing/data': 0.01902299001812935, 'timing/model': 1.3209181241691113, 'learning_rate/qwen_vl_interface': 3.038331028455721e-06, 'learning_rate/action_model': 3.038331028455721e-05, 'epoch': 0.27}) +[2026-06-04 18:52:52,691][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.7791350483894348, 'train/grad_norm_pre_clip': 3.2848563194274902, 'timing/data': 0.016531601548194885, 'timing/model': 1.3509499225765467, 'learning_rate/qwen_vl_interface': 2.98349588876686e-06, 'learning_rate/action_model': 2.9834958887668597e-05, 'epoch': 0.28}) +[2026-06-04 18:56:17,609][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 1.4580687284469604, 'train/grad_norm_pre_clip': 4.5203046798706055, 'timing/data': 0.025163564831018448, 'timing/model': 1.2740282993763685, 'learning_rate/qwen_vl_interface': 2.9290491028894658e-06, 'learning_rate/action_model': 2.9290491028894658e-05, 'epoch': 0.28}) +[2026-06-04 19:23:18,144][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2500 +[2026-06-04 19:23:18,146][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.7366629242897034, 'train/grad_norm_pre_clip': 3.5927178859710693, 'eval/loss': 0.818659245967865, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 7130.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 7130.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 7130.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 7130.0, 'checkpoint/best_score': 7770.0, 'checkpoint/best_step': 1900.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.02683602273464203, 'timing/model': 1.44208356551826, 'learning_rate/qwen_vl_interface': 2.8750000000000013e-06, 'learning_rate/action_model': 2.875000000000001e-05, 'epoch': 0.28}) +[2026-06-04 19:23:18,146][starVLA.training.train_starvla][INFO] - Training complete. Best checkpoint step=1900 score=7770.0 path=/workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state +[2026-06-04 19:23:18,147][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/openvla_demon_attack_fix_latency_0/checkpoints/best_state +[2026-06-04 19:23:19,655][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-04 19:23:29,715][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-04 19:23:29,716][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-04 19:23:29,717][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-04 19:23:29,718][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-04 19:23:30,158][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-04 19:57:52,863][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2500 +[2026-06-04 19:57:54,425][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/hydra/.hydra/overrides.yaml b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..a65e7a3536a76ddcee90aa06e55317d954a84eb3 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/hydra/.hydra/overrides.yaml @@ -0,0 +1,91 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=single +- run_id=flappy_fix_latency_0_200ep +- run_root_dir=/workspace/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=false +- checkpoint.save_pt_file=false +- trainer.is_resume=true +- trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1300_state +- trainer.resume_step=1300 +- trainer.max_train_steps=5000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=100 +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=4 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9, 0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=2 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.shuffle=true +- datasets.vla_data.action_balance.enabled=false +- datasets.vla_data.action_balance.strategy=balanced_epoch +- datasets.vla_data.action_balance.action_key=action_id +- datasets.vla_data.action_balance.target_flap_fraction=0.3 +- datasets.vla_data.action_balance.noop_id=0 +- datasets.vla_data.action_balance.flap_id=1 +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_0_200ep +- datasets.vla_data.data_mix=flappy_train__bridge +- datasets.vla_data.eval_data_mix=flappy_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=flappy +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=1 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.mid_train.enabled=false +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[0] +- rl_games.env_eval.post_train.enabled=false +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/hydra/train_starvla_hydra.log b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..7cee4f1f24c49d38c9e1fd2ea6c8a63a890fe498 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/hydra/train_starvla_hydra.log @@ -0,0 +1,5644 @@ +[2026-06-15 13:09:53,074][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-15 13:09:53,094][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-15 13:10:04,644][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-15 13:10:04,703][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-15 13:10:05,493][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-15 13:10:05,494][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-15 13:10:05,494][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-15 13:10:05,494][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-15 13:10:05,494][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-15 13:10:05,494][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-15 13:10:05,494][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-15 13:10:05,495][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-15 13:10:05,495][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-15 13:10:05,495][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-15 13:10:05,495][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-15 13:10:05,495][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-15 13:10:05,495][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-15 13:10:05,495][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-15 13:10:05,495][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-15 13:10:05,495][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-15 13:10:05,496][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-15 13:10:05,496][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-15 13:10:05,496][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-15 13:10:05,497][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-15 13:10:05,497][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-15 13:10:05,497][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-15 13:10:05,497][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-15 13:10:05,497][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-15 13:10:05,497][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-15 13:10:05,497][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-15 13:10:05,497][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-15 13:10:05,497][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-15 13:10:05,498][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-15 13:10:05,498][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-15 13:10:05,498][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-15 13:10:05,550][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-15 13:10:05,594][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/config.full.yaml +[2026-06-15 13:10:05,604][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/config.yaml +[2026-06-15 13:10:05,604][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-15 13:10:09,737][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-15 13:10:09,738][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-06-15 13:10:09,738][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-15 13:10:09,738][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 4 +[2026-06-15 13:10:09,738][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-15 13:10:47,251][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 0.695102870464325, 'train/grad_norm_pre_clip': 5.665567398071289, 'timing/data': 0.030271895229816437, 'timing/model': 1.1796924322843552, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-15 13:10:51,689][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 0.6950811147689819, 'train/grad_norm_pre_clip': 10.501192092895508, 'timing/data': 0.018866526894271374, 'timing/model': 1.1537401098757982, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-15 13:10:56,094][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 0.6865609884262085, 'train/grad_norm_pre_clip': 7.889710426330566, 'timing/data': 0.018742785789072514, 'timing/model': 1.1544708767905831, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-06-15 13:11:00,512][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 0.6785533428192139, 'train/grad_norm_pre_clip': 5.129703998565674, 'timing/data': 0.02268077526241541, 'timing/model': 1.1711726039648056, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-06-15 13:11:04,908][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 0.6366512775421143, 'train/grad_norm_pre_clip': 8.633942604064941, 'timing/data': 0.02148779947310686, 'timing/model': 1.137672083452344, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-06-15 13:11:09,312][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 0.6206667423248291, 'train/grad_norm_pre_clip': 6.287411689758301, 'timing/data': 0.019092338159680367, 'timing/model': 1.148098548874259, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-06-15 13:11:13,774][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 0.5461766123771667, 'train/grad_norm_pre_clip': 7.821647644042969, 'timing/data': 0.018127773888409138, 'timing/model': 1.205495866946876, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-06-15 13:11:18,235][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.5178389549255371, 'train/grad_norm_pre_clip': 6.037387847900391, 'timing/data': 0.018410065211355686, 'timing/model': 1.1573881907388568, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-06-15 13:11:22,741][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 0.499970406293869, 'train/grad_norm_pre_clip': 4.10943078994751, 'timing/data': 0.019258199259638786, 'timing/model': 1.1513307923451066, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-06-15 13:11:27,440][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.4713890254497528, 'train/grad_norm_pre_clip': 3.046360969543457, 'timing/data': 0.01869796682149172, 'timing/model': 1.3534026248380542, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-15 13:11:31,941][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.41094687581062317, 'train/grad_norm_pre_clip': 2.573695659637451, 'timing/data': 0.019204119220376015, 'timing/model': 1.1577564785256982, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.01}) +[2026-06-15 13:11:36,446][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.48245471715927124, 'train/grad_norm_pre_clip': 0.32354238629341125, 'timing/data': 0.017907431349158287, 'timing/model': 1.1573004387319088, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.01}) +[2026-06-15 13:11:40,936][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.37924161553382874, 'train/grad_norm_pre_clip': 0.5197327733039856, 'timing/data': 0.018074213527143, 'timing/model': 1.1654475824907422, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.01}) +[2026-06-15 13:11:45,438][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.570793867111206, 'train/grad_norm_pre_clip': 2.7275643348693848, 'timing/data': 0.01814639288932085, 'timing/model': 1.1513191815465689, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.01}) +[2026-06-15 13:11:49,935][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.38149741291999817, 'train/grad_norm_pre_clip': 0.8053624629974365, 'timing/data': 0.01826828345656395, 'timing/model': 1.1623912509530783, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.01}) +[2026-06-15 13:11:54,453][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.8419053554534912, 'train/grad_norm_pre_clip': 6.066689491271973, 'timing/data': 0.010853845626115799, 'timing/model': 1.1713160891085863, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 13:11:58,940][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.6259196400642395, 'train/grad_norm_pre_clip': 3.7089178562164307, 'timing/data': 0.0109397666528821, 'timing/model': 1.1647270368412137, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 13:12:03,453][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.788196861743927, 'train/grad_norm_pre_clip': 5.384097576141357, 'timing/data': 0.01318435836583376, 'timing/model': 1.1547682071104646, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.01}) +[2026-06-15 13:12:07,967][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.3127652108669281, 'train/grad_norm_pre_clip': 0.4573725759983063, 'timing/data': 0.00969872996211052, 'timing/model': 1.175322774797678, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.01}) +[2026-06-15 13:12:12,452][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.38658663630485535, 'train/grad_norm_pre_clip': 1.1311179399490356, 'timing/data': 0.012294563464820385, 'timing/model': 1.1551211215555668, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-15 13:12:16,956][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.5971169471740723, 'train/grad_norm_pre_clip': 3.310176134109497, 'timing/data': 0.012096072547137737, 'timing/model': 1.1602639108896255, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.01}) +[2026-06-15 13:12:21,450][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.7102737426757812, 'train/grad_norm_pre_clip': 4.555653095245361, 'timing/data': 0.012411924079060555, 'timing/model': 1.155597249045968, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 13:12:25,958][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.37770015001296997, 'train/grad_norm_pre_clip': 0.36792394518852234, 'timing/data': 0.012270313687622547, 'timing/model': 1.1340692900121212, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 13:12:30,445][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.4351024329662323, 'train/grad_norm_pre_clip': 0.19604679942131042, 'timing/data': 0.011966521851718426, 'timing/model': 1.1561687560752034, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.01}) +[2026-06-15 13:12:34,942][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.43346089124679565, 'train/grad_norm_pre_clip': 0.8551084995269775, 'timing/data': 0.016508464701473713, 'timing/model': 1.1431816900148988, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.01}) +[2026-06-15 13:12:39,447][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.2681892514228821, 'train/grad_norm_pre_clip': 4.500439643859863, 'timing/data': 0.010078982450067997, 'timing/model': 1.163697917945683, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.01}) +[2026-06-15 13:12:43,942][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.5242573618888855, 'train/grad_norm_pre_clip': 0.42546913027763367, 'timing/data': 0.012416424229741096, 'timing/model': 1.159062965773046, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.01}) +[2026-06-15 13:12:48,443][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.6039488911628723, 'train/grad_norm_pre_clip': 1.8312393426895142, 'timing/data': 0.011549679562449455, 'timing/model': 1.1627330370247364, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.01}) +[2026-06-15 13:12:52,942][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.5996950268745422, 'train/grad_norm_pre_clip': 1.3184118270874023, 'timing/data': 0.013179847970604897, 'timing/model': 1.1572020491585135, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.01}) +[2026-06-15 13:12:57,447][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.47014954686164856, 'train/grad_norm_pre_clip': 3.073683738708496, 'timing/data': 0.010780585929751396, 'timing/model': 1.1695645647123456, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-15 13:13:01,943][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.5897014141082764, 'train/grad_norm_pre_clip': 0.6701433062553406, 'timing/data': 0.010342554189264774, 'timing/model': 1.1612574933096766, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.01}) +[2026-06-15 13:13:06,436][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.4902946949005127, 'train/grad_norm_pre_clip': 3.8732681274414062, 'timing/data': 0.010674295015633106, 'timing/model': 1.166209931485355, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.01}) +[2026-06-15 13:13:10,927][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.6168652772903442, 'train/grad_norm_pre_clip': 0.6558710336685181, 'timing/data': 0.010456743650138378, 'timing/model': 1.1617971500381827, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.02}) +[2026-06-15 13:13:15,509][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.537517249584198, 'train/grad_norm_pre_clip': 2.1403346061706543, 'timing/data': 0.011362939141690731, 'timing/model': 1.2411686340346932, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.02}) +[2026-06-15 13:13:19,943][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.589981734752655, 'train/grad_norm_pre_clip': 0.35364800691604614, 'timing/data': 0.01205456256866455, 'timing/model': 1.1636090436950326, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.02}) +[2026-06-15 13:13:24,451][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.5250529646873474, 'train/grad_norm_pre_clip': 1.2597845792770386, 'timing/data': 0.014470415189862251, 'timing/model': 1.1807912923395634, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.02}) +[2026-06-15 13:13:29,239][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.5929968357086182, 'train/grad_norm_pre_clip': 1.6654083728790283, 'timing/data': 0.015919972211122513, 'timing/model': 1.1889781868085265, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.02}) +[2026-06-15 13:13:33,728][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.39965352416038513, 'train/grad_norm_pre_clip': 2.574585199356079, 'timing/data': 0.010428913868963718, 'timing/model': 1.1051381845027208, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.02}) +[2026-06-15 13:13:38,242][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.4753960967063904, 'train/grad_norm_pre_clip': 2.299516439437866, 'timing/data': 0.011793551035225391, 'timing/model': 1.1506072524935007, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 13:13:42,761][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.32474103569984436, 'train/grad_norm_pre_clip': 2.0372610092163086, 'timing/data': 0.011989242397248745, 'timing/model': 1.184031487442553, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.02}) +[2026-06-15 13:13:47,246][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.48515987396240234, 'train/grad_norm_pre_clip': 2.514080286026001, 'timing/data': 0.011329290457069874, 'timing/model': 1.1678618341684341, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.02}) +[2026-06-15 13:13:51,735][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.6813902258872986, 'train/grad_norm_pre_clip': 6.209958553314209, 'timing/data': 0.011848771944642067, 'timing/model': 1.1575703518465161, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.02}) +[2026-06-15 13:13:56,337][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.3732084035873413, 'train/grad_norm_pre_clip': 5.054623126983643, 'timing/data': 0.011241358704864979, 'timing/model': 1.1774477967992425, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.02}) +[2026-06-15 13:14:00,845][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.5398474931716919, 'train/grad_norm_pre_clip': 3.8959271907806396, 'timing/data': 0.011445868760347366, 'timing/model': 1.1692744260653853, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 13:14:05,346][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.31645429134368896, 'train/grad_norm_pre_clip': 3.3150253295898438, 'timing/data': 0.011190577410161495, 'timing/model': 1.1713712187483907, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.02}) +[2026-06-15 13:14:09,852][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.4203387498855591, 'train/grad_norm_pre_clip': 2.27685546875, 'timing/data': 0.01143700908869505, 'timing/model': 1.173548792488873, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.02}) +[2026-06-15 13:14:14,347][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.5119969248771667, 'train/grad_norm_pre_clip': 3.306062698364258, 'timing/data': 0.011075868271291256, 'timing/model': 1.170259895734489, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.02}) +[2026-06-15 13:14:18,839][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.50494384765625, 'train/grad_norm_pre_clip': 2.2340521812438965, 'timing/data': 0.012636575847864151, 'timing/model': 1.1604777062311769, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.02}) +[2026-06-15 13:14:23,335][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.5618926286697388, 'train/grad_norm_pre_clip': 3.5168023109436035, 'timing/data': 0.010261663235723972, 'timing/model': 1.170303569175303, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.02}) +[2026-06-15 13:14:27,929][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.36630648374557495, 'train/grad_norm_pre_clip': 7.856691837310791, 'timing/data': 0.011083178222179413, 'timing/model': 1.190803206525743, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.02}) +[2026-06-15 13:14:32,434][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.42905735969543457, 'train/grad_norm_pre_clip': 5.29998254776001, 'timing/data': 0.011245958507061005, 'timing/model': 1.1214000070467591, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.02}) +[2026-06-15 13:14:36,931][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.4891899824142456, 'train/grad_norm_pre_clip': 4.660696506500244, 'timing/data': 0.010758265852928162, 'timing/model': 1.1257901210337877, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.02}) +[2026-06-15 13:14:41,437][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.27472183108329773, 'train/grad_norm_pre_clip': 11.257401466369629, 'timing/data': 0.010433764196932316, 'timing/model': 1.1557552386075258, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.02}) +[2026-06-15 13:14:45,937][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.6836646795272827, 'train/grad_norm_pre_clip': 18.354825973510742, 'timing/data': 0.011026786640286446, 'timing/model': 1.1592643288895488, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.02}) +[2026-06-15 13:14:50,431][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.3806309700012207, 'train/grad_norm_pre_clip': 8.599105834960938, 'timing/data': 0.013897782191634178, 'timing/model': 1.1567748570814729, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.03}) +[2026-06-15 13:14:54,942][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.4000259041786194, 'train/grad_norm_pre_clip': 21.33682632446289, 'timing/data': 0.012071352452039719, 'timing/model': 1.160579428076744, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.03}) +[2026-06-15 13:14:59,435][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.6502394676208496, 'train/grad_norm_pre_clip': 8.922181129455566, 'timing/data': 0.01167318131774664, 'timing/model': 1.1478099534288049, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.03}) +[2026-06-15 13:15:04,086][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.28963544964790344, 'train/grad_norm_pre_clip': 9.606308937072754, 'timing/data': 0.011152539402246475, 'timing/model': 1.3133819280192256, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.03}) +[2026-06-15 13:15:08,552][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.39661651849746704, 'train/grad_norm_pre_clip': 8.881964683532715, 'timing/data': 0.010460615158081055, 'timing/model': 1.176583094522357, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.03}) +[2026-06-15 13:15:13,040][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.5926661491394043, 'train/grad_norm_pre_clip': 10.124571800231934, 'timing/data': 0.011973291635513306, 'timing/model': 1.1630788473412395, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.03}) +[2026-06-15 13:15:17,549][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.30456826090812683, 'train/grad_norm_pre_clip': 8.207018852233887, 'timing/data': 0.011574690230190754, 'timing/model': 1.1690216790884733, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.03}) +[2026-06-15 13:15:22,048][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.5381147861480713, 'train/grad_norm_pre_clip': 17.6707706451416, 'timing/data': 0.011839772574603558, 'timing/model': 1.1707019982859492, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.03}) +[2026-06-15 13:15:26,745][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.32848742604255676, 'train/grad_norm_pre_clip': 5.300361156463623, 'timing/data': 0.012068713083863258, 'timing/model': 1.1672330414876342, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.03}) +[2026-06-15 13:15:31,247][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.3495628833770752, 'train/grad_norm_pre_clip': 5.351834774017334, 'timing/data': 0.011249958537518978, 'timing/model': 1.1642256062477827, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.03}) +[2026-06-15 13:15:35,746][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.3941856026649475, 'train/grad_norm_pre_clip': 12.998651504516602, 'timing/data': 0.011679330840706825, 'timing/model': 1.1661378582939506, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.03}) +[2026-06-15 13:15:40,238][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.4152834117412567, 'train/grad_norm_pre_clip': 35.938446044921875, 'timing/data': 0.010223673656582832, 'timing/model': 1.122997123748064, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.03}) +[2026-06-15 13:15:44,749][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.22488930821418762, 'train/grad_norm_pre_clip': 14.280866622924805, 'timing/data': 0.01068049669265747, 'timing/model': 1.1681172903627157, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.03}) +[2026-06-15 13:15:49,249][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.2839249074459076, 'train/grad_norm_pre_clip': 4.735169887542725, 'timing/data': 0.011341549456119537, 'timing/model': 1.1722660642117262, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.03}) +[2026-06-15 13:15:53,744][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.39225539565086365, 'train/grad_norm_pre_clip': 14.327168464660645, 'timing/data': 0.012170872651040554, 'timing/model': 1.1528604533523321, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.03}) +[2026-06-15 13:15:58,240][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.22745142877101898, 'train/grad_norm_pre_clip': 6.520084381103516, 'timing/data': 0.01176237128674984, 'timing/model': 1.1550686871632934, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.03}) +[2026-06-15 13:16:02,744][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.31747788190841675, 'train/grad_norm_pre_clip': 10.909449577331543, 'timing/data': 0.012523305602371693, 'timing/model': 1.1625222666189075, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.03}) +[2026-06-15 13:16:07,244][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.2917615473270416, 'train/grad_norm_pre_clip': 13.704947471618652, 'timing/data': 0.011625740677118301, 'timing/model': 1.1631403025239706, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.03}) +[2026-06-15 13:16:11,745][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.46130114793777466, 'train/grad_norm_pre_clip': 17.000205993652344, 'timing/data': 0.01258163619786501, 'timing/model': 1.163470913656056, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.03}) +[2026-06-15 13:16:16,250][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.40372511744499207, 'train/grad_norm_pre_clip': 11.845519065856934, 'timing/data': 0.01053827628493309, 'timing/model': 1.1731779463589191, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.03}) +[2026-06-15 13:16:20,740][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.5781083106994629, 'train/grad_norm_pre_clip': 15.592826843261719, 'timing/data': 0.010720395483076572, 'timing/model': 1.161078424192965, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.03}) +[2026-06-15 13:16:25,245][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 0.47794556617736816, 'train/grad_norm_pre_clip': 12.26533031463623, 'timing/data': 0.01057649590075016, 'timing/model': 1.1646046731621027, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.03}) +[2026-06-15 13:16:29,740][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.5856573581695557, 'train/grad_norm_pre_clip': 15.112765312194824, 'timing/data': 0.010408635251224041, 'timing/model': 1.153512205928564, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.04}) +[2026-06-15 13:16:34,315][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.586078405380249, 'train/grad_norm_pre_clip': 11.80102252960205, 'timing/data': 0.011225108057260513, 'timing/model': 1.2377616195008159, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.04}) +[2026-06-15 13:16:38,744][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.5095621943473816, 'train/grad_norm_pre_clip': 11.891782760620117, 'timing/data': 0.012044483795762062, 'timing/model': 1.1458357982337475, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.04}) +[2026-06-15 13:16:43,449][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.38948267698287964, 'train/grad_norm_pre_clip': 4.768512725830078, 'timing/data': 0.011300559155642986, 'timing/model': 1.2013955619186163, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.04}) +[2026-06-15 13:16:47,939][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.43872421979904175, 'train/grad_norm_pre_clip': 3.7707505226135254, 'timing/data': 0.012035583145916462, 'timing/model': 1.1576580926775932, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.04}) +[2026-06-15 13:16:52,447][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.5955911874771118, 'train/grad_norm_pre_clip': 9.513075828552246, 'timing/data': 0.011003207415342331, 'timing/model': 1.1649466026574373, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.04}) +[2026-06-15 13:16:56,946][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.3855327367782593, 'train/grad_norm_pre_clip': 4.393378734588623, 'timing/data': 0.010726545937359333, 'timing/model': 1.1653328444808722, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.04}) +[2026-06-15 13:17:01,439][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.4140285551548004, 'train/grad_norm_pre_clip': 5.4260334968566895, 'timing/data': 0.010407974943518639, 'timing/model': 1.1610697032883763, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.04}) +[2026-06-15 13:17:05,944][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.3578505218029022, 'train/grad_norm_pre_clip': 9.38165283203125, 'timing/data': 0.0123743349686265, 'timing/model': 1.163148115389049, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.04}) +[2026-06-15 13:17:10,445][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.11445406824350357, 'train/grad_norm_pre_clip': 9.313989639282227, 'timing/data': 0.010594035498797894, 'timing/model': 1.1699150716885924, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.04}) +[2026-06-15 13:17:14,949][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.7736459374427795, 'train/grad_norm_pre_clip': 28.038795471191406, 'timing/data': 0.011358760297298431, 'timing/model': 1.1707114670425653, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.04}) +[2026-06-15 13:17:19,447][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.8502534627914429, 'train/grad_norm_pre_clip': 22.752281188964844, 'timing/data': 0.011710641905665398, 'timing/model': 1.1689046788960695, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.04}) +[2026-06-15 13:17:23,979][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.41774603724479675, 'train/grad_norm_pre_clip': 11.417180061340332, 'timing/data': 0.01305185817182064, 'timing/model': 1.2014873707666993, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.04}) +[2026-06-15 13:17:28,652][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.37323427200317383, 'train/grad_norm_pre_clip': 5.858789920806885, 'timing/data': 0.010690717026591301, 'timing/model': 1.169899475760758, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.04}) +[2026-06-15 13:17:33,148][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.42439886927604675, 'train/grad_norm_pre_clip': 12.360386848449707, 'timing/data': 0.011527320370078087, 'timing/model': 1.168817911297083, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.04}) +[2026-06-15 13:17:37,648][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.4462572932243347, 'train/grad_norm_pre_clip': 13.448634147644043, 'timing/data': 0.011703741736710072, 'timing/model': 1.166986232623458, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.04}) +[2026-06-15 13:17:42,150][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.18539443612098694, 'train/grad_norm_pre_clip': 14.709967613220215, 'timing/data': 0.011757132597267628, 'timing/model': 1.1673327758908272, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.04}) +[2026-06-15 13:17:46,644][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.435576468706131, 'train/grad_norm_pre_clip': 15.573580741882324, 'timing/data': 0.012731227092444897, 'timing/model': 1.1624256912618876, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.04}) +[2026-06-15 13:17:51,148][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.47029438614845276, 'train/grad_norm_pre_clip': 7.476391792297363, 'timing/data': 0.011997142806649208, 'timing/model': 1.16401894018054, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.04}) +[2026-06-15 13:17:55,645][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.6019070148468018, 'train/grad_norm_pre_clip': 14.534257888793945, 'timing/data': 0.011741592548787594, 'timing/model': 1.167411908507347, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.04}) +[2026-06-15 13:18:00,145][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.7123338580131531, 'train/grad_norm_pre_clip': 17.28384780883789, 'timing/data': 0.010515205562114716, 'timing/model': 1.167484880425036, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.04}) +[2026-06-15 13:18:04,646][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.4203990697860718, 'train/grad_norm_pre_clip': 5.425018787384033, 'timing/data': 0.011265398934483528, 'timing/model': 1.170155924744904, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.04}) +[2026-06-15 13:18:09,140][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.7195894718170166, 'train/grad_norm_pre_clip': 32.21649169921875, 'timing/data': 0.010851826518774033, 'timing/model': 1.154041601344943, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.05}) +[2026-06-15 13:21:02,168][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp221ikb2n/test.c -o /tmp/tmp221ikb2n/test.o +[2026-06-15 13:21:02,198][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp221ikb2n/test.o -laio -o /tmp/tmp221ikb2n/a.out +[2026-06-15 13:21:02,724][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpeuafh4xq/test.c -o /tmp/tmpeuafh4xq/test.o +[2026-06-15 13:21:02,757][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpeuafh4xq/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpeuafh4xq/a.out +[2026-06-15 13:21:02,814][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpoux8u6i3/test.c -o /tmp/tmpoux8u6i3/test.o +[2026-06-15 13:21:02,845][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpoux8u6i3/test.o -laio -o /tmp/tmpoux8u6i3/a.out +[2026-06-15 13:29:43,601][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 1.0754321813583374, 'train/grad_norm_pre_clip': 51.55894470214844, 'eval/loss': 0.7234216332435608, 'eval/flap_precision': 0.17388451443569553, 'eval/flap_recall': 0.9636363636363636, 'eval/flap_f1': 0.2946081156197888, 'eval/flap_rate_model': 0.42333333333333334, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.2946081156197888, 'timing/data': 0.011845891363918781, 'timing/model': 1.1695899236947298, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.05}) +[2026-06-15 13:29:43,602][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_100_state +[2026-06-15 13:29:43,609][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 13:29:51,390][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_100_state/model.safetensors +[2026-06-15 13:30:02,082][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_100_state/optimizer.bin +[2026-06-15 13:30:02,083][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_100_state/sampler.bin +[2026-06-15 13:30:02,083][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_100_state/sampler_1.bin +[2026-06-15 13:30:02,085][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-15 13:30:02,088][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-15 13:30:02,088][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 13:30:02,103][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 13:30:07,927][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.925886332988739, 'train/grad_norm_pre_clip': 89.03219604492188, 'timing/data': 0.01096001360565424, 'timing/model': 1.4739676089957356, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 0.05}) +[2026-06-15 13:30:13,638][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.883529543876648, 'train/grad_norm_pre_clip': 33.92546463012695, 'timing/data': 0.011879054829478264, 'timing/model': 1.4015633147209883, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 0.05}) +[2026-06-15 13:30:19,525][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.5676238536834717, 'train/grad_norm_pre_clip': 23.887313842773438, 'timing/data': 0.019167140126228333, 'timing/model': 1.6829037759453058, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 0.05}) +[2026-06-15 13:30:25,008][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.3579728603363037, 'train/grad_norm_pre_clip': 2.6596269607543945, 'timing/data': 0.010776253417134285, 'timing/model': 1.4430808033794165, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 0.05}) +[2026-06-15 13:30:30,708][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.3664751946926117, 'train/grad_norm_pre_clip': 3.05619215965271, 'timing/data': 0.011730816215276718, 'timing/model': 1.4460018016397953, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 0.05}) +[2026-06-15 13:30:36,536][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.6101020574569702, 'train/grad_norm_pre_clip': 8.309603691101074, 'timing/data': 0.014157441444694996, 'timing/model': 1.491158583201468, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 0.05}) +[2026-06-15 13:30:42,309][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.5497956275939941, 'train/grad_norm_pre_clip': 7.42999267578125, 'timing/data': 0.012216117233037949, 'timing/model': 1.4189638858661056, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 0.05}) +[2026-06-15 13:30:48,018][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.34528404474258423, 'train/grad_norm_pre_clip': 2.158160448074341, 'timing/data': 0.010684564709663391, 'timing/model': 1.4274815758690238, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 0.05}) +[2026-06-15 13:30:53,837][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.4588490426540375, 'train/grad_norm_pre_clip': 2.4138340950012207, 'timing/data': 0.011996586807072163, 'timing/model': 1.6661160746589303, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 0.05}) +[2026-06-15 13:30:59,447][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.5031765103340149, 'train/grad_norm_pre_clip': 8.883451461791992, 'timing/data': 0.011439066380262375, 'timing/model': 1.4745546616613865, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.05}) +[2026-06-15 13:31:05,245][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.5905345678329468, 'train/grad_norm_pre_clip': 11.737618446350098, 'timing/data': 0.012446599081158638, 'timing/model': 1.4105078829452395, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 0.05}) +[2026-06-15 13:31:11,122][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.4833715260028839, 'train/grad_norm_pre_clip': 7.9823079109191895, 'timing/data': 0.010831975378096104, 'timing/model': 1.4553021863102913, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 0.05}) +[2026-06-15 13:31:16,926][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.42028436064720154, 'train/grad_norm_pre_clip': 4.446498870849609, 'timing/data': 0.010848275385797024, 'timing/model': 1.455457005649805, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 0.05}) +[2026-06-15 13:31:22,716][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.486660897731781, 'train/grad_norm_pre_clip': 4.621208190917969, 'timing/data': 0.011306436732411385, 'timing/model': 1.4674142124131322, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 0.05}) +[2026-06-15 13:31:28,561][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.2535924017429352, 'train/grad_norm_pre_clip': 3.979684352874756, 'timing/data': 0.010787155479192734, 'timing/model': 1.7070761332288384, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 0.05}) +[2026-06-15 13:31:34,219][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.3425157070159912, 'train/grad_norm_pre_clip': 6.934154510498047, 'timing/data': 0.01157187856733799, 'timing/model': 1.4688078528270125, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 0.05}) +[2026-06-15 13:31:39,935][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.31070905923843384, 'train/grad_norm_pre_clip': 3.8224072456359863, 'timing/data': 0.010769517160952091, 'timing/model': 1.4002741798758507, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 0.05}) +[2026-06-15 13:31:45,731][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.4499007761478424, 'train/grad_norm_pre_clip': 7.596749782562256, 'timing/data': 0.011019377037882805, 'timing/model': 1.4624828919768333, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 0.05}) +[2026-06-15 13:31:51,533][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.3293105959892273, 'train/grad_norm_pre_clip': 14.178474426269531, 'timing/data': 0.010861887596547604, 'timing/model': 1.4740338269621134, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 0.05}) +[2026-06-15 13:31:57,328][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.5697388052940369, 'train/grad_norm_pre_clip': 10.864246368408203, 'timing/data': 0.010554617270827293, 'timing/model': 1.4529055897146463, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.05}) +[2026-06-15 13:32:03,184][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.2869623005390167, 'train/grad_norm_pre_clip': 11.338211059570312, 'timing/data': 0.012385460548102856, 'timing/model': 1.7044019727036357, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 0.06}) +[2026-06-15 13:32:08,833][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.3143320679664612, 'train/grad_norm_pre_clip': 6.340540409088135, 'timing/data': 0.010759088210761547, 'timing/model': 1.4726013410836458, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 0.06}) +[2026-06-15 13:32:14,641][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.4219903349876404, 'train/grad_norm_pre_clip': 8.480873107910156, 'timing/data': 0.013632125221192837, 'timing/model': 1.4518054220825434, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 0.06}) +[2026-06-15 13:32:20,528][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.2672390341758728, 'train/grad_norm_pre_clip': 3.9358596801757812, 'timing/data': 0.010535898618400097, 'timing/model': 1.458833385258913, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 0.06}) +[2026-06-15 13:32:26,322][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.5117632150650024, 'train/grad_norm_pre_clip': 9.799430847167969, 'timing/data': 0.011604150757193565, 'timing/model': 1.4042086638510227, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 0.06}) +[2026-06-15 13:32:32,137][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.47877928614616394, 'train/grad_norm_pre_clip': 6.235416412353516, 'timing/data': 0.010539548471570015, 'timing/model': 1.4776826035231352, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 0.06}) +[2026-06-15 13:32:38,053][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.30612480640411377, 'train/grad_norm_pre_clip': 4.277873516082764, 'timing/data': 0.02270809281617403, 'timing/model': 1.686278834939003, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 0.06}) +[2026-06-15 13:32:43,728][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.4712947607040405, 'train/grad_norm_pre_clip': 5.752899169921875, 'timing/data': 0.011669892817735672, 'timing/model': 1.4600669126957655, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 0.06}) +[2026-06-15 13:32:49,724][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.3287058174610138, 'train/grad_norm_pre_clip': 11.400450706481934, 'timing/data': 0.010783989913761616, 'timing/model': 1.4806569935753942, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 0.06}) +[2026-06-15 13:32:55,731][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.5163598656654358, 'train/grad_norm_pre_clip': 22.770051956176758, 'timing/data': 0.01152100134640932, 'timing/model': 1.4659306472167373, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.06}) +[2026-06-15 13:33:01,538][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.3915131390094757, 'train/grad_norm_pre_clip': 11.580852508544922, 'timing/data': 0.010909521020948887, 'timing/model': 1.4764581182971597, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 0.06}) +[2026-06-15 13:33:07,322][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.44338420033454895, 'train/grad_norm_pre_clip': 9.273862838745117, 'timing/data': 0.01158871315419674, 'timing/model': 1.426541198976338, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 0.06}) +[2026-06-15 13:33:13,062][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.5442935228347778, 'train/grad_norm_pre_clip': 7.580406665802002, 'timing/data': 0.011475062929093838, 'timing/model': 1.6793576590716839, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 0.06}) +[2026-06-15 13:33:18,733][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.8184472322463989, 'train/grad_norm_pre_clip': 12.500787734985352, 'timing/data': 0.011030442081391811, 'timing/model': 1.4667236534878612, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 0.06}) +[2026-06-15 13:33:24,440][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.2931956946849823, 'train/grad_norm_pre_clip': 2.1240015029907227, 'timing/data': 0.01085849292576313, 'timing/model': 1.460203094407916, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.06}) +[2026-06-15 13:33:30,335][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.2957614064216614, 'train/grad_norm_pre_clip': 2.620354413986206, 'timing/data': 0.011247113347053528, 'timing/model': 1.4748055478557944, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 0.06}) +[2026-06-15 13:33:36,235][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.43105536699295044, 'train/grad_norm_pre_clip': 5.323554039001465, 'timing/data': 0.010412650182843208, 'timing/model': 1.4763470431789756, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 0.06}) +[2026-06-15 13:33:42,020][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.40286174416542053, 'train/grad_norm_pre_clip': 2.960301637649536, 'timing/data': 0.010849962010979652, 'timing/model': 1.4097156962379813, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 0.06}) +[2026-06-15 13:33:47,864][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.3271211087703705, 'train/grad_norm_pre_clip': 3.777766704559326, 'timing/data': 0.011117872782051563, 'timing/model': 1.7053691046312451, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 0.06}) +[2026-06-15 13:33:53,530][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.32631784677505493, 'train/grad_norm_pre_clip': 9.478854179382324, 'timing/data': 0.01087323296815157, 'timing/model': 1.4769105138257146, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.06}) +[2026-06-15 13:33:59,343][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.3167206346988678, 'train/grad_norm_pre_clip': 3.819817066192627, 'timing/data': 0.011189484037458897, 'timing/model': 1.4942945940420032, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 0.06}) +[2026-06-15 13:34:05,235][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.6904617547988892, 'train/grad_norm_pre_clip': 20.464357376098633, 'timing/data': 0.011734685860574245, 'timing/model': 1.4833223605528474, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 0.06}) +[2026-06-15 13:34:11,033][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.39475172758102417, 'train/grad_norm_pre_clip': 6.020316123962402, 'timing/data': 0.012061837129294872, 'timing/model': 1.4530678559094667, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 0.07}) +[2026-06-15 13:34:16,832][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.49818477034568787, 'train/grad_norm_pre_clip': 11.486518859863281, 'timing/data': 0.010800252668559551, 'timing/model': 1.4559693727642298, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 0.07}) +[2026-06-15 13:34:22,708][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.49595534801483154, 'train/grad_norm_pre_clip': 5.25112771987915, 'timing/data': 0.011222855187952518, 'timing/model': 1.723194015212357, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 0.07}) +[2026-06-15 13:34:28,342][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.3447922468185425, 'train/grad_norm_pre_clip': 10.273046493530273, 'timing/data': 0.010856973938643932, 'timing/model': 1.470780972391367, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 0.07}) +[2026-06-15 13:34:34,246][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.5016884207725525, 'train/grad_norm_pre_clip': 8.163467407226562, 'timing/data': 0.011102833785116673, 'timing/model': 1.4777208818122745, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 0.07}) +[2026-06-15 13:34:40,140][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.4802045226097107, 'train/grad_norm_pre_clip': 12.273836135864258, 'timing/data': 0.010849414393305779, 'timing/model': 1.476034002378583, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 0.07}) +[2026-06-15 13:34:46,045][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.4855894148349762, 'train/grad_norm_pre_clip': 11.34991455078125, 'timing/data': 0.011671417392790318, 'timing/model': 1.478774937801063, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.07}) +[2026-06-15 13:34:51,930][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.378723680973053, 'train/grad_norm_pre_clip': 9.196257591247559, 'timing/data': 0.010689863935112953, 'timing/model': 1.4206254202872515, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.07}) +[2026-06-15 13:34:57,770][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.3650587797164917, 'train/grad_norm_pre_clip': 2.5468943119049072, 'timing/data': 0.011209876276552677, 'timing/model': 1.6979776117950678, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 0.07}) +[2026-06-15 13:35:03,675][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.5187422633171082, 'train/grad_norm_pre_clip': 5.968454837799072, 'timing/data': 0.012328839860856533, 'timing/model': 1.7072349609807134, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 0.07}) +[2026-06-15 13:35:09,358][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.5484445691108704, 'train/grad_norm_pre_clip': 8.035393714904785, 'timing/data': 0.012849941849708557, 'timing/model': 1.4887512428686023, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 0.07}) +[2026-06-15 13:35:15,241][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.3515319526195526, 'train/grad_norm_pre_clip': 2.328620433807373, 'timing/data': 0.012560130096971989, 'timing/model': 1.456171514466405, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 0.07}) +[2026-06-15 13:35:21,143][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.2573310434818268, 'train/grad_norm_pre_clip': 2.7597107887268066, 'timing/data': 0.011001436039805412, 'timing/model': 1.4665396623313427, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 0.07}) +[2026-06-15 13:35:27,033][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.242355078458786, 'train/grad_norm_pre_clip': 2.6930227279663086, 'timing/data': 0.01025399286299944, 'timing/model': 1.4328310154378414, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 0.07}) +[2026-06-15 13:35:32,874][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.2683033347129822, 'train/grad_norm_pre_clip': 11.562860488891602, 'timing/data': 0.011805698275566101, 'timing/model': 1.6891800053417683, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 0.07}) +[2026-06-15 13:35:38,537][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.42308589816093445, 'train/grad_norm_pre_clip': 8.304612159729004, 'timing/data': 0.011146027594804764, 'timing/model': 1.449087405577302, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 0.07}) +[2026-06-15 13:35:44,442][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.3629750609397888, 'train/grad_norm_pre_clip': 11.492936134338379, 'timing/data': 0.011944889090955257, 'timing/model': 1.4682722026482224, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 0.07}) +[2026-06-15 13:35:50,341][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.4755021333694458, 'train/grad_norm_pre_clip': 13.11556339263916, 'timing/data': 0.010913127101957798, 'timing/model': 1.4729931121692061, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.07}) +[2026-06-15 13:35:56,236][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.30186721682548523, 'train/grad_norm_pre_clip': 11.490623474121094, 'timing/data': 0.01185936015099287, 'timing/model': 1.4329000199213624, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 0.07}) +[2026-06-15 13:36:02,058][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.25906890630722046, 'train/grad_norm_pre_clip': 6.560070514678955, 'timing/data': 0.011696750298142433, 'timing/model': 1.4122145725414157, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 0.07}) +[2026-06-15 13:36:07,925][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.32559260725975037, 'train/grad_norm_pre_clip': 13.699474334716797, 'timing/data': 0.011810339987277985, 'timing/model': 1.650970183312893, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 0.07}) +[2026-06-15 13:36:13,544][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.3391789495944977, 'train/grad_norm_pre_clip': 8.226008415222168, 'timing/data': 0.013449256308376789, 'timing/model': 1.4694328093901277, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 0.08}) +[2026-06-15 13:36:19,444][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.4149588942527771, 'train/grad_norm_pre_clip': 8.479853630065918, 'timing/data': 0.011226139031350613, 'timing/model': 1.4682493573054671, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 0.08}) +[2026-06-15 13:36:25,342][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.3513338565826416, 'train/grad_norm_pre_clip': 5.880549430847168, 'timing/data': 0.011796880513429642, 'timing/model': 1.470768678933382, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 0.08}) +[2026-06-15 13:36:31,241][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.3826121985912323, 'train/grad_norm_pre_clip': 12.571622848510742, 'timing/data': 0.01181150134652853, 'timing/model': 1.4468965670093894, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 0.08}) +[2026-06-15 13:36:37,132][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.3553493916988373, 'train/grad_norm_pre_clip': 3.3388168811798096, 'timing/data': 0.010850338265299797, 'timing/model': 1.4089595573022962, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 0.08}) +[2026-06-15 13:36:42,956][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.3054647743701935, 'train/grad_norm_pre_clip': 6.857383728027344, 'timing/data': 0.012238421477377415, 'timing/model': 1.6886657709255815, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 0.08}) +[2026-06-15 13:36:48,646][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.28333091735839844, 'train/grad_norm_pre_clip': 5.300783157348633, 'timing/data': 0.012487703934311867, 'timing/model': 1.4777943175286055, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.08}) +[2026-06-15 13:36:54,540][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.30543839931488037, 'train/grad_norm_pre_clip': 2.9321556091308594, 'timing/data': 0.011811401695013046, 'timing/model': 1.470346781425178, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 0.08}) +[2026-06-15 13:37:00,443][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.3050966262817383, 'train/grad_norm_pre_clip': 3.9528379440307617, 'timing/data': 0.011031948029994965, 'timing/model': 1.4736678544431925, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 0.08}) +[2026-06-15 13:37:06,330][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.1847849190235138, 'train/grad_norm_pre_clip': 2.48004150390625, 'timing/data': 0.011519529856741428, 'timing/model': 1.4099958594888449, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 0.08}) +[2026-06-15 13:37:12,176][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.29634639620780945, 'train/grad_norm_pre_clip': 5.320245265960693, 'timing/data': 0.010803169570863247, 'timing/model': 1.7140865670517087, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 0.08}) +[2026-06-15 13:37:17,831][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.32117557525634766, 'train/grad_norm_pre_clip': 4.961056232452393, 'timing/data': 0.011382721364498138, 'timing/model': 1.447516500018537, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 0.08}) +[2026-06-15 13:37:23,640][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.2461307793855667, 'train/grad_norm_pre_clip': 8.332040786743164, 'timing/data': 0.011025029234588146, 'timing/model': 1.473008650355041, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 0.08}) +[2026-06-15 13:37:29,439][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.4850773811340332, 'train/grad_norm_pre_clip': 12.85567855834961, 'timing/data': 0.01173609122633934, 'timing/model': 1.4723542975261807, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 0.08}) +[2026-06-15 13:37:35,240][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.2783289849758148, 'train/grad_norm_pre_clip': 10.570066452026367, 'timing/data': 0.011351331137120724, 'timing/model': 1.4609483955428004, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 0.08}) +[2026-06-15 13:37:41,141][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.17951728403568268, 'train/grad_norm_pre_clip': 11.330557823181152, 'timing/data': 0.010844728909432888, 'timing/model': 1.4785733465105295, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 0.08}) +[2026-06-15 13:37:47,083][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.19376946985721588, 'train/grad_norm_pre_clip': 6.237404823303223, 'timing/data': 0.011253930628299713, 'timing/model': 1.7121999431401491, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.08}) +[2026-06-15 13:37:52,746][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.14503923058509827, 'train/grad_norm_pre_clip': 6.633103847503662, 'timing/data': 0.011708592996001244, 'timing/model': 1.4746516505256295, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 0.08}) +[2026-06-15 13:37:58,646][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.3497619926929474, 'train/grad_norm_pre_clip': 11.69787311553955, 'timing/data': 0.011082789860665798, 'timing/model': 1.4609251888468862, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 0.08}) +[2026-06-15 13:38:04,539][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.29830047488212585, 'train/grad_norm_pre_clip': 17.227027893066406, 'timing/data': 0.011844534426927567, 'timing/model': 1.473693162202835, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 0.08}) +[2026-06-15 13:38:10,338][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.32923412322998047, 'train/grad_norm_pre_clip': 16.64711570739746, 'timing/data': 0.011440233327448368, 'timing/model': 1.4525193022564054, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 0.08}) +[2026-06-15 13:38:16,241][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.3120734393596649, 'train/grad_norm_pre_clip': 8.902751922607422, 'timing/data': 0.011223861016333103, 'timing/model': 1.4681859286502004, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 0.08}) +[2026-06-15 13:38:22,174][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.2863152027130127, 'train/grad_norm_pre_clip': 11.919994354248047, 'timing/data': 0.010900999419391155, 'timing/model': 1.7069939086213708, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 0.09}) +[2026-06-15 13:38:27,844][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.4731784164905548, 'train/grad_norm_pre_clip': 14.99039363861084, 'timing/data': 0.011920124292373657, 'timing/model': 1.4724937677383423, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 0.09}) +[2026-06-15 13:38:33,741][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.23150689899921417, 'train/grad_norm_pre_clip': 4.945372581481934, 'timing/data': 0.011475532315671444, 'timing/model': 1.4709465997293591, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 0.09}) +[2026-06-15 13:38:39,642][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.4878101944923401, 'train/grad_norm_pre_clip': 17.365516662597656, 'timing/data': 0.011821803636848927, 'timing/model': 1.4749009096994996, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 0.09}) +[2026-06-15 13:38:45,529][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.383097767829895, 'train/grad_norm_pre_clip': 10.225543022155762, 'timing/data': 0.010749030858278275, 'timing/model': 1.4178947741165757, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.09}) +[2026-06-15 13:38:51,335][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.2778271436691284, 'train/grad_norm_pre_clip': 5.771216869354248, 'timing/data': 0.011810105293989182, 'timing/model': 1.4426444917917252, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 0.09}) +[2026-06-15 13:38:57,288][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.1913510262966156, 'train/grad_norm_pre_clip': 4.131636619567871, 'timing/data': 0.011049631051719189, 'timing/model': 1.7204042952507734, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 0.09}) +[2026-06-15 13:39:02,937][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.3039064407348633, 'train/grad_norm_pre_clip': 4.582164764404297, 'timing/data': 0.014082303270697594, 'timing/model': 1.4529515225440264, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 0.09}) +[2026-06-15 13:39:08,848][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.24161070585250854, 'train/grad_norm_pre_clip': 2.6486568450927734, 'timing/data': 0.012190605513751507, 'timing/model': 1.4747434118762612, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 0.09}) +[2026-06-15 13:39:14,744][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.2722774147987366, 'train/grad_norm_pre_clip': 3.09614634513855, 'timing/data': 0.01130006369203329, 'timing/model': 1.4654110223054886, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 0.09}) +[2026-06-15 13:39:20,639][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.27644768357276917, 'train/grad_norm_pre_clip': 4.543397903442383, 'timing/data': 0.011867265217006207, 'timing/model': 1.4567903643473983, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 0.09}) +[2026-06-15 13:39:26,541][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.3123618960380554, 'train/grad_norm_pre_clip': 13.605372428894043, 'timing/data': 0.01104230247437954, 'timing/model': 1.4073978308588266, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 0.09}) +[2026-06-15 13:39:32,474][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.2450990229845047, 'train/grad_norm_pre_clip': 3.363640546798706, 'timing/data': 0.011095643043518066, 'timing/model': 1.7032125992700458, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.09}) +[2026-06-15 13:39:38,132][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.33230382204055786, 'train/grad_norm_pre_clip': 13.391283988952637, 'timing/data': 0.011797544546425343, 'timing/model': 1.4495849609375, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 0.09}) +[2026-06-15 13:50:22,843][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.2572748064994812, 'train/grad_norm_pre_clip': 4.059500217437744, 'eval/loss': 0.1564491093158722, 'eval/flap_precision': 0.6026785714285714, 'eval/flap_recall': 0.4909090909090909, 'eval/flap_f1': 0.5410821643286572, 'eval/flap_rate_model': 0.06222222222222222, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.5410821643286572, 'timing/data': 0.013500772416591644, 'timing/model': 1.4616739824414253, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.09}) +[2026-06-15 13:50:22,844][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_200_state +[2026-06-15 13:50:22,850][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 13:50:30,824][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_200_state/model.safetensors +[2026-06-15 13:50:41,872][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_200_state/optimizer.bin +[2026-06-15 13:50:41,874][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_200_state/sampler.bin +[2026-06-15 13:50:41,874][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_200_state/sampler_1.bin +[2026-06-15 13:50:41,875][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-15 13:50:42,026][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-15 13:50:42,026][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 13:50:42,042][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 13:50:47,712][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.2787615656852722, 'train/grad_norm_pre_clip': 8.829689979553223, 'timing/data': 0.010015004314482212, 'timing/model': 1.4679332170635462, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 0.09}) +[2026-06-15 13:50:53,619][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.2935493290424347, 'train/grad_norm_pre_clip': 9.428157806396484, 'timing/data': 0.00962478294968605, 'timing/model': 1.4721434013918042, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 0.09}) +[2026-06-15 13:50:59,430][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.45964813232421875, 'train/grad_norm_pre_clip': 9.951003074645996, 'timing/data': 0.012386444956064224, 'timing/model': 1.4616097277030349, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 0.09}) +[2026-06-15 13:51:05,412][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.3198649287223816, 'train/grad_norm_pre_clip': 6.700170040130615, 'timing/data': 0.010300996713340282, 'timing/model': 1.5577627625316381, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 0.09}) +[2026-06-15 13:51:11,229][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.23549574613571167, 'train/grad_norm_pre_clip': 4.481168746948242, 'timing/data': 0.010611467994749546, 'timing/model': 1.4661761447787285, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.09}) +[2026-06-15 13:51:17,153][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.18631787598133087, 'train/grad_norm_pre_clip': 8.144572257995605, 'timing/data': 0.01032316591590643, 'timing/model': 1.686209480278194, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 0.09}) +[2026-06-15 13:51:22,833][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.4583125412464142, 'train/grad_norm_pre_clip': 18.207759857177734, 'timing/data': 0.010306335985660553, 'timing/model': 1.4530710130929947, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 0.09}) +[2026-06-15 13:51:28,734][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.15223683416843414, 'train/grad_norm_pre_clip': 4.874313831329346, 'timing/data': 0.010355446487665176, 'timing/model': 1.451165659353137, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 0.1}) +[2026-06-15 13:51:34,748][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.15855294466018677, 'train/grad_norm_pre_clip': 5.5661163330078125, 'timing/data': 0.010627897456288338, 'timing/model': 1.474359206855297, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 0.1}) +[2026-06-15 13:51:40,758][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.31931042671203613, 'train/grad_norm_pre_clip': 10.160120964050293, 'timing/data': 0.0118302833288908, 'timing/model': 1.5027399491518736, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.1}) +[2026-06-15 13:51:46,728][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.35638976097106934, 'train/grad_norm_pre_clip': 9.905720710754395, 'timing/data': 0.011378501541912556, 'timing/model': 1.745841259136796, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 0.1}) +[2026-06-15 13:51:52,457][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.29925084114074707, 'train/grad_norm_pre_clip': 7.839694499969482, 'timing/data': 0.010321326553821564, 'timing/model': 1.4721306078135967, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 0.1}) +[2026-06-15 13:51:58,455][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.23528748750686646, 'train/grad_norm_pre_clip': 8.461198806762695, 'timing/data': 0.012372405268251896, 'timing/model': 1.469894852489233, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 0.1}) +[2026-06-15 13:52:04,462][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.32884320616722107, 'train/grad_norm_pre_clip': 4.692559719085693, 'timing/data': 0.013182630762457848, 'timing/model': 1.4830824444070458, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 0.1}) +[2026-06-15 13:52:10,466][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.28409543633461, 'train/grad_norm_pre_clip': 6.9449663162231445, 'timing/data': 0.01100418996065855, 'timing/model': 1.5076091894879937, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 0.1}) +[2026-06-15 13:52:16,371][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.31272801756858826, 'train/grad_norm_pre_clip': 6.22168493270874, 'timing/data': 0.01297695841640234, 'timing/model': 1.4356899773702025, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 0.1}) +[2026-06-15 13:52:22,269][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.3096805810928345, 'train/grad_norm_pre_clip': 4.483421325683594, 'timing/data': 0.011796252802014351, 'timing/model': 1.6789043620228767, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 0.1}) +[2026-06-15 13:52:27,950][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.13876420259475708, 'train/grad_norm_pre_clip': 3.331840991973877, 'timing/data': 0.012867477722465992, 'timing/model': 1.4514049347490072, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 0.1}) +[2026-06-15 13:52:33,957][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.18938902020454407, 'train/grad_norm_pre_clip': 9.431807518005371, 'timing/data': 0.012129235081374645, 'timing/model': 1.4633595421910286, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 0.1}) +[2026-06-15 13:52:39,954][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.23597317934036255, 'train/grad_norm_pre_clip': 4.413700580596924, 'timing/data': 0.012154335156083107, 'timing/model': 1.4468009695410728, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.1}) +[2026-06-15 13:52:45,952][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.3405984938144684, 'train/grad_norm_pre_clip': 4.443201065063477, 'timing/data': 0.01231557596474886, 'timing/model': 1.5038492530584335, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 0.1}) +[2026-06-15 13:52:51,958][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.3626551032066345, 'train/grad_norm_pre_clip': 10.423731803894043, 'timing/data': 0.01116960123181343, 'timing/model': 1.4883089549839497, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 0.1}) +[2026-06-15 13:52:57,873][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.29971662163734436, 'train/grad_norm_pre_clip': 14.402125358581543, 'timing/data': 0.011012960225343704, 'timing/model': 1.678600111976266, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 0.1}) +[2026-06-15 13:53:03,552][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.3888370394706726, 'train/grad_norm_pre_clip': 12.247230529785156, 'timing/data': 0.010629807598888874, 'timing/model': 1.4665989028289914, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 0.1}) +[2026-06-15 13:53:09,553][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.605353593826294, 'train/grad_norm_pre_clip': 19.072256088256836, 'timing/data': 0.01012729573994875, 'timing/model': 1.454908543266356, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 0.1}) +[2026-06-15 13:53:15,450][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.1729651689529419, 'train/grad_norm_pre_clip': 5.169692039489746, 'timing/data': 0.011271982453763485, 'timing/model': 1.4552602572366595, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 0.1}) +[2026-06-15 13:53:21,435][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.12990276515483856, 'train/grad_norm_pre_clip': 4.740208625793457, 'timing/data': 0.010519907809793949, 'timing/model': 1.4954356821253896, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 0.1}) +[2026-06-15 13:53:27,404][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.5863064527511597, 'train/grad_norm_pre_clip': 11.182146072387695, 'timing/data': 0.01032259687781334, 'timing/model': 1.6958804093301296, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 0.1}) +[2026-06-15 13:53:33,051][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.49229174852371216, 'train/grad_norm_pre_clip': 16.459436416625977, 'timing/data': 0.010579608380794525, 'timing/model': 1.4658072236925364, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 0.1}) +[2026-06-15 13:53:39,052][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.21285443007946014, 'train/grad_norm_pre_clip': 6.4315972328186035, 'timing/data': 0.010134636424481869, 'timing/model': 1.451108549721539, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.11}) +[2026-06-15 13:53:44,951][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.23157140612602234, 'train/grad_norm_pre_clip': 7.527756690979004, 'timing/data': 0.010517258197069168, 'timing/model': 1.462247034534812, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 0.11}) +[2026-06-15 13:53:50,852][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.30009156465530396, 'train/grad_norm_pre_clip': 7.472812175750732, 'timing/data': 0.010889970697462559, 'timing/model': 1.4631405007094145, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 0.11}) +[2026-06-15 13:53:56,755][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.33935976028442383, 'train/grad_norm_pre_clip': 7.460387706756592, 'timing/data': 0.010008175857365131, 'timing/model': 1.449340021237731, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 0.11}) +[2026-06-15 13:54:02,712][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.4444522261619568, 'train/grad_norm_pre_clip': 9.493435859680176, 'timing/data': 0.00999954529106617, 'timing/model': 1.7136203097179532, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 0.11}) +[2026-06-15 13:54:08,349][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.29753199219703674, 'train/grad_norm_pre_clip': 7.243960857391357, 'timing/data': 0.01007209625095129, 'timing/model': 1.442269223742187, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 0.11}) +[2026-06-15 13:54:14,366][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.29144400358200073, 'train/grad_norm_pre_clip': 3.9214701652526855, 'timing/data': 0.010496477596461773, 'timing/model': 1.4639834947884083, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 0.11}) +[2026-06-15 13:54:20,358][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.24861256778240204, 'train/grad_norm_pre_clip': 5.45583438873291, 'timing/data': 0.012724577449262142, 'timing/model': 1.4579619402065873, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 0.11}) +[2026-06-15 13:54:26,357][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.2581062614917755, 'train/grad_norm_pre_clip': 9.506449699401855, 'timing/data': 0.011261100880801678, 'timing/model': 1.451111452654004, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 0.11}) +[2026-06-15 13:54:32,259][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.6084619760513306, 'train/grad_norm_pre_clip': 17.84416389465332, 'timing/data': 0.011390142142772675, 'timing/model': 1.443324108608067, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 0.11}) +[2026-06-15 13:54:38,212][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.16128690540790558, 'train/grad_norm_pre_clip': 4.222017288208008, 'timing/data': 0.010798049159348011, 'timing/model': 1.701192551292479, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.11}) +[2026-06-15 13:54:43,856][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.2566589117050171, 'train/grad_norm_pre_clip': 7.405265808105469, 'timing/data': 0.010552488267421722, 'timing/model': 1.4576917504891753, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 0.11}) +[2026-06-15 13:54:49,862][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.16516682505607605, 'train/grad_norm_pre_clip': 4.294161319732666, 'timing/data': 0.010584588162600994, 'timing/model': 1.4726714612916112, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 0.11}) +[2026-06-15 13:54:55,759][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.4492901563644409, 'train/grad_norm_pre_clip': 7.115846157073975, 'timing/data': 0.009742244146764278, 'timing/model': 1.463259051553905, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 0.11}) +[2026-06-15 13:55:01,740][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.2723142206668854, 'train/grad_norm_pre_clip': 4.5671467781066895, 'timing/data': 0.010134725831449032, 'timing/model': 1.4924734570086002, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 0.11}) +[2026-06-15 13:55:07,650][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.36841416358947754, 'train/grad_norm_pre_clip': 17.013141632080078, 'timing/data': 0.010945470072329044, 'timing/model': 1.4295954406261444, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 0.11}) +[2026-06-15 13:55:13,609][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.3210442364215851, 'train/grad_norm_pre_clip': 11.882017135620117, 'timing/data': 0.011675224639475346, 'timing/model': 1.7214640015736222, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 0.11}) +[2026-06-15 13:55:19,250][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.1869235634803772, 'train/grad_norm_pre_clip': 4.570582866668701, 'timing/data': 0.010449607856571674, 'timing/model': 1.4613876519724727, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.11}) +[2026-06-15 13:55:25,252][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.10131126642227173, 'train/grad_norm_pre_clip': 2.7103919982910156, 'timing/data': 0.010951920412480831, 'timing/model': 1.454446922056377, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 0.11}) +[2026-06-15 13:55:31,154][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.4381539821624756, 'train/grad_norm_pre_clip': 15.843082427978516, 'timing/data': 0.010655058547854424, 'timing/model': 1.4582044119015336, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 0.11}) +[2026-06-15 13:55:37,053][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.3562919497489929, 'train/grad_norm_pre_clip': 17.076496124267578, 'timing/data': 0.010565858334302902, 'timing/model': 1.4544112151488662, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.11}) +[2026-06-15 13:55:43,049][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.3766942322254181, 'train/grad_norm_pre_clip': 16.4322566986084, 'timing/data': 0.010708249174058437, 'timing/model': 1.4990268629044294, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 0.11}) +[2026-06-15 13:55:49,007][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.4874766767024994, 'train/grad_norm_pre_clip': 13.954252243041992, 'timing/data': 0.011057470925152302, 'timing/model': 1.7035992965102196, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 0.12}) +[2026-06-15 13:55:54,660][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.4774370491504669, 'train/grad_norm_pre_clip': 27.510507583618164, 'timing/data': 0.013620671816170216, 'timing/model': 1.4604416117072105, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 0.12}) +[2026-06-15 13:56:00,665][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.263248473405838, 'train/grad_norm_pre_clip': 12.246391296386719, 'timing/data': 0.010942979715764523, 'timing/model': 1.4731861213222146, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 0.12}) +[2026-06-15 13:56:06,561][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.30547574162483215, 'train/grad_norm_pre_clip': 10.122848510742188, 'timing/data': 0.011922755278646946, 'timing/model': 1.4657602198421955, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 0.12}) +[2026-06-15 13:56:12,458][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.3726639151573181, 'train/grad_norm_pre_clip': 19.554101943969727, 'timing/data': 0.011324911378324032, 'timing/model': 1.4633978614583611, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 0.12}) +[2026-06-15 13:56:18,458][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.4669829308986664, 'train/grad_norm_pre_clip': 20.73958969116211, 'timing/data': 0.011430173180997372, 'timing/model': 1.5159089844673872, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 0.12}) +[2026-06-15 13:56:24,403][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.21167317032814026, 'train/grad_norm_pre_clip': 4.998904705047607, 'timing/data': 0.011420552618801594, 'timing/model': 1.7004846772179008, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 0.12}) +[2026-06-15 13:56:30,061][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.372962087392807, 'train/grad_norm_pre_clip': 10.73873233795166, 'timing/data': 0.011592483147978783, 'timing/model': 1.4659530390053988, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 0.12}) +[2026-06-15 13:56:36,070][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.30182838439941406, 'train/grad_norm_pre_clip': 5.279055595397949, 'timing/data': 0.010454217903316021, 'timing/model': 1.4729131134226918, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.12}) +[2026-06-15 13:56:42,062][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.4410580098628998, 'train/grad_norm_pre_clip': 11.452266693115234, 'timing/data': 0.010533149354159832, 'timing/model': 1.4674152098596096, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 0.12}) +[2026-06-15 13:56:47,962][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.3386819064617157, 'train/grad_norm_pre_clip': 8.464646339416504, 'timing/data': 0.01124415174126625, 'timing/model': 1.4685904867947102, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 0.12}) +[2026-06-15 13:56:53,953][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.23466648161411285, 'train/grad_norm_pre_clip': 5.43295955657959, 'timing/data': 0.01131594181060791, 'timing/model': 1.5114698773249984, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 0.12}) +[2026-06-15 13:57:00,000][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.32782042026519775, 'train/grad_norm_pre_clip': 4.325498580932617, 'timing/data': 0.013296770863234997, 'timing/model': 1.6959297684952617, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 0.12}) +[2026-06-15 13:57:05,958][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.1686914563179016, 'train/grad_norm_pre_clip': 7.581356048583984, 'timing/data': 0.011553393676877022, 'timing/model': 1.4609040673822165, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 0.12}) +[2026-06-15 13:57:11,960][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.24009452760219574, 'train/grad_norm_pre_clip': 3.03996205329895, 'timing/data': 0.011733673512935638, 'timing/model': 1.4601850165054202, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 0.12}) +[2026-06-15 13:57:17,851][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.4550577402114868, 'train/grad_norm_pre_clip': 8.0995512008667, 'timing/data': 0.012713477946817875, 'timing/model': 1.451765269972384, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 0.12}) +[2026-06-15 13:57:23,845][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.22316724061965942, 'train/grad_norm_pre_clip': 4.27430534362793, 'timing/data': 0.011889885179698467, 'timing/model': 1.5057758204638958, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 0.12}) +[2026-06-15 13:57:29,863][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.3749738335609436, 'train/grad_norm_pre_clip': 15.765671730041504, 'timing/data': 0.010996730998158455, 'timing/model': 1.5137448888272047, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 0.12}) +[2026-06-15 13:57:35,798][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.24437640607357025, 'train/grad_norm_pre_clip': 4.777402400970459, 'timing/data': 0.011712484061717987, 'timing/model': 1.6987282056361437, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.12}) +[2026-06-15 13:57:41,708][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.22562231123447418, 'train/grad_norm_pre_clip': 5.174506187438965, 'timing/data': 0.010679748840630054, 'timing/model': 1.7065675733610988, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 0.12}) +[2026-06-15 13:57:47,351][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.166187584400177, 'train/grad_norm_pre_clip': 2.8978734016418457, 'timing/data': 0.010998171754181385, 'timing/model': 1.443349008448422, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 0.12}) +[2026-06-15 13:57:53,257][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.33372870087623596, 'train/grad_norm_pre_clip': 20.44175148010254, 'timing/data': 0.011000211350619793, 'timing/model': 1.457749418914318, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 0.12}) +[2026-06-15 13:57:59,160][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.5464355945587158, 'train/grad_norm_pre_clip': 22.428865432739258, 'timing/data': 0.011233072727918625, 'timing/model': 1.4660483775660396, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 0.13}) +[2026-06-15 13:58:05,201][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.2826330363750458, 'train/grad_norm_pre_clip': 15.806727409362793, 'timing/data': 0.01075202040374279, 'timing/model': 1.5291548017412424, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.13}) +[2026-06-15 13:58:11,104][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.16220863163471222, 'train/grad_norm_pre_clip': 14.92758560180664, 'timing/data': 0.011562072671949863, 'timing/model': 1.7062600729987025, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 0.13}) +[2026-06-15 13:58:16,774][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.1092643141746521, 'train/grad_norm_pre_clip': 8.522162437438965, 'timing/data': 0.011505632661283016, 'timing/model': 1.473739198409021, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 0.13}) +[2026-06-15 13:58:22,762][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.6377854943275452, 'train/grad_norm_pre_clip': 31.30512809753418, 'timing/data': 0.011565102264285088, 'timing/model': 1.4574574949219823, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 0.13}) +[2026-06-15 13:58:28,758][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.3127630650997162, 'train/grad_norm_pre_clip': 17.37278938293457, 'timing/data': 0.013933693990111351, 'timing/model': 1.4476562310010195, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 0.13}) +[2026-06-15 13:58:34,766][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.6611442565917969, 'train/grad_norm_pre_clip': 27.643394470214844, 'timing/data': 0.01096941065043211, 'timing/model': 1.5159706193953753, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.13}) +[2026-06-15 13:58:40,665][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.46251431107521057, 'train/grad_norm_pre_clip': 26.00530433654785, 'timing/data': 0.011637894436717033, 'timing/model': 1.467217925004661, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 0.13}) +[2026-06-15 13:58:46,569][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.2706567943096161, 'train/grad_norm_pre_clip': 12.117799758911133, 'timing/data': 0.01163137424737215, 'timing/model': 1.6621298491954803, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 0.13}) +[2026-06-15 13:58:52,264][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.16788694262504578, 'train/grad_norm_pre_clip': 20.379905700683594, 'timing/data': 0.01145258266478777, 'timing/model': 1.4586716294288635, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 0.13}) +[2026-06-15 13:58:58,265][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.35081934928894043, 'train/grad_norm_pre_clip': 15.863164901733398, 'timing/data': 0.011480492539703846, 'timing/model': 1.455165097489953, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 0.13}) +[2026-06-15 13:59:04,362][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.35081639885902405, 'train/grad_norm_pre_clip': 9.942632675170898, 'timing/data': 0.011094066314399242, 'timing/model': 1.577748978510499, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 0.13}) +[2026-06-15 13:59:10,363][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.2263646125793457, 'train/grad_norm_pre_clip': 7.168440341949463, 'timing/data': 0.01122449804097414, 'timing/model': 1.5186941651627421, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 0.13}) +[2026-06-15 13:59:16,375][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.3046211898326874, 'train/grad_norm_pre_clip': 8.465256690979004, 'timing/data': 0.010975425131618977, 'timing/model': 1.5041281059384346, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 0.13}) +[2026-06-15 13:59:22,279][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.24696621298789978, 'train/grad_norm_pre_clip': 5.81411075592041, 'timing/data': 0.01090406347066164, 'timing/model': 1.6758093014359474, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 0.13}) +[2026-06-15 13:59:27,967][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.32595905661582947, 'train/grad_norm_pre_clip': 5.539834022521973, 'timing/data': 0.010940912179648876, 'timing/model': 1.4669883446767926, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 0.13}) +[2026-06-15 13:59:33,962][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.6295372247695923, 'train/grad_norm_pre_clip': 17.485462188720703, 'timing/data': 0.011510148644447327, 'timing/model': 1.4586451230570674, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.13}) +[2026-06-15 13:59:39,970][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.37844905257225037, 'train/grad_norm_pre_clip': 8.908998489379883, 'timing/data': 0.011706389486789703, 'timing/model': 1.4803023030981421, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 0.13}) +[2026-06-15 13:59:45,973][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.41463935375213623, 'train/grad_norm_pre_clip': 6.9950947761535645, 'timing/data': 0.011328843422234058, 'timing/model': 1.5012425817549229, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 0.13}) +[2026-06-15 13:59:51,916][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.4074901044368744, 'train/grad_norm_pre_clip': 11.38521671295166, 'timing/data': 0.011732758022844791, 'timing/model': 1.7122388631105423, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 0.13}) +[2026-06-15 13:59:57,559][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.3721114993095398, 'train/grad_norm_pre_clip': 11.79123592376709, 'timing/data': 0.010777185671031475, 'timing/model': 1.4508871473371983, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 0.13}) +[2026-06-15 14:00:03,553][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.25954821705818176, 'train/grad_norm_pre_clip': 7.702746868133545, 'timing/data': 0.011916939169168472, 'timing/model': 1.458440731279552, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 0.13}) +[2026-06-15 14:00:09,462][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.31176748871803284, 'train/grad_norm_pre_clip': 4.526464939117432, 'timing/data': 0.010869566351175308, 'timing/model': 1.4646036410704255, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.14}) +[2026-06-15 14:00:15,363][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.2506270706653595, 'train/grad_norm_pre_clip': 4.627711296081543, 'timing/data': 0.011171799153089523, 'timing/model': 1.4727397449314594, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 0.14}) +[2026-06-15 14:00:21,253][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.55448317527771, 'train/grad_norm_pre_clip': 9.94881534576416, 'timing/data': 0.011011737398803234, 'timing/model': 1.4479042505845428, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 0.14}) +[2026-06-15 14:00:27,208][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.3558059632778168, 'train/grad_norm_pre_clip': 5.450710296630859, 'timing/data': 0.011251658201217651, 'timing/model': 1.7097073551267385, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 0.14}) +[2026-06-15 14:11:15,047][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.1662246286869049, 'train/grad_norm_pre_clip': 3.832305431365967, 'eval/loss': 0.2167268693447113, 'eval/flap_precision': 1.0, 'eval/flap_recall': 0.01090909090909091, 'eval/flap_f1': 0.02158273381294964, 'eval/flap_rate_model': 0.0008333333333333334, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.02158273381294964, 'timing/data': 0.01085264328867197, 'timing/model': 1.475311242043972, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.14}) +[2026-06-15 14:11:15,048][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_300_state +[2026-06-15 14:11:15,055][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 14:11:23,269][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_300_state/model.safetensors +[2026-06-15 14:11:42,145][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_300_state/optimizer.bin +[2026-06-15 14:11:42,145][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_300_state/sampler.bin +[2026-06-15 14:11:42,145][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_300_state/sampler_1.bin +[2026-06-15 14:11:42,147][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-15 14:11:42,318][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-15 14:11:42,319][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 14:11:42,334][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 14:11:48,245][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.6297221779823303, 'train/grad_norm_pre_clip': 10.199860572814941, 'timing/data': 0.011642123572528362, 'timing/model': 1.6737815402448177, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 0.14}) +[2026-06-15 14:11:53,919][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.30414578318595886, 'train/grad_norm_pre_clip': 4.865725040435791, 'timing/data': 0.010721665807068348, 'timing/model': 1.456901622004807, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 0.14}) +[2026-06-15 14:11:59,728][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.21529512107372284, 'train/grad_norm_pre_clip': 2.2578015327453613, 'timing/data': 0.012025825679302216, 'timing/model': 1.4610983366146684, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 0.14}) +[2026-06-15 14:12:05,632][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.2767293453216553, 'train/grad_norm_pre_clip': 2.7837750911712646, 'timing/data': 0.011825893074274063, 'timing/model': 1.4673161646351218, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 0.14}) +[2026-06-15 14:12:11,527][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.267103910446167, 'train/grad_norm_pre_clip': 4.606634616851807, 'timing/data': 0.011547381058335304, 'timing/model': 1.4213184053078294, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 0.14}) +[2026-06-15 14:12:17,372][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.25974127650260925, 'train/grad_norm_pre_clip': 2.9050137996673584, 'timing/data': 0.011062427423894405, 'timing/model': 1.706349667161703, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 0.14}) +[2026-06-15 14:12:23,173][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.24968329071998596, 'train/grad_norm_pre_clip': 5.076287269592285, 'timing/data': 0.010721015743911266, 'timing/model': 1.5986533528193831, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 0.14}) +[2026-06-15 14:12:28,841][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.3233517110347748, 'train/grad_norm_pre_clip': 6.360328197479248, 'timing/data': 0.011280668899416924, 'timing/model': 1.4760314105078578, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 0.14}) +[2026-06-15 14:12:34,734][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.18565362691879272, 'train/grad_norm_pre_clip': 6.235575199127197, 'timing/data': 0.011316278949379921, 'timing/model': 1.4737524008378386, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 0.14}) +[2026-06-15 14:12:40,650][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.2310379594564438, 'train/grad_norm_pre_clip': 7.982024192810059, 'timing/data': 0.013371053151786327, 'timing/model': 1.4817006811499596, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.14}) +[2026-06-15 14:12:46,535][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.39602044224739075, 'train/grad_norm_pre_clip': 18.563684463500977, 'timing/data': 0.010976267047226429, 'timing/model': 1.4281293554231524, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 0.14}) +[2026-06-15 14:12:52,463][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.34152936935424805, 'train/grad_norm_pre_clip': 19.2857666015625, 'timing/data': 0.011881272308528423, 'timing/model': 1.6866446174681187, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 0.14}) +[2026-06-15 14:12:58,290][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.17815667390823364, 'train/grad_norm_pre_clip': 7.609289646148682, 'timing/data': 0.011725220829248428, 'timing/model': 1.6155486106872559, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 0.14}) +[2026-06-15 14:13:03,946][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.09533737599849701, 'train/grad_norm_pre_clip': 4.43116569519043, 'timing/data': 0.011648900806903839, 'timing/model': 1.4686838313937187, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 0.14}) +[2026-06-15 14:13:09,863][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.11905118823051453, 'train/grad_norm_pre_clip': 4.4690470695495605, 'timing/data': 0.011958042159676552, 'timing/model': 1.4812363153323531, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 0.14}) +[2026-06-15 14:13:15,742][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.29982778429985046, 'train/grad_norm_pre_clip': 9.838862419128418, 'timing/data': 0.011803911067545414, 'timing/model': 1.4650296717882156, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 0.14}) +[2026-06-15 14:13:21,655][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.37681302428245544, 'train/grad_norm_pre_clip': 10.107200622558594, 'timing/data': 0.011250467039644718, 'timing/model': 1.4385439986363053, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 0.14}) +[2026-06-15 14:13:27,587][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.31556037068367004, 'train/grad_norm_pre_clip': 10.918662071228027, 'timing/data': 0.011082836426794529, 'timing/model': 1.700456641614437, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 0.15}) +[2026-06-15 14:13:33,257][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.49745094776153564, 'train/grad_norm_pre_clip': 21.768903732299805, 'timing/data': 0.010975445620715618, 'timing/model': 1.4711636696010828, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 0.15}) +[2026-06-15 14:13:39,152][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.450124591588974, 'train/grad_norm_pre_clip': 23.279932022094727, 'timing/data': 0.011587869375944138, 'timing/model': 1.4723563371226192, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.15}) +[2026-06-15 14:13:45,040][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.39341670274734497, 'train/grad_norm_pre_clip': 11.122417449951172, 'timing/data': 0.011354397982358932, 'timing/model': 1.462364767678082, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 0.15}) +[2026-06-15 14:13:50,967][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.1613323986530304, 'train/grad_norm_pre_clip': 3.950326919555664, 'timing/data': 0.01086044404655695, 'timing/model': 1.4947033878415823, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 0.15}) +[2026-06-15 14:13:56,862][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.16235776245594025, 'train/grad_norm_pre_clip': 3.7064545154571533, 'timing/data': 0.01209915243089199, 'timing/model': 1.434850131161511, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 0.15}) +[2026-06-15 14:14:02,797][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.3269103765487671, 'train/grad_norm_pre_clip': 9.588683128356934, 'timing/data': 0.012524133548140526, 'timing/model': 1.7120666969567537, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 0.15}) +[2026-06-15 14:14:08,452][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.3723379671573639, 'train/grad_norm_pre_clip': 12.69351863861084, 'timing/data': 0.011079754680395126, 'timing/model': 1.4616903541609645, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 0.15}) +[2026-06-15 14:14:14,352][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.16691310703754425, 'train/grad_norm_pre_clip': 8.692408561706543, 'timing/data': 0.011379707604646683, 'timing/model': 1.4596314895898104, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 0.15}) +[2026-06-15 14:14:20,251][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.24552220106124878, 'train/grad_norm_pre_clip': 5.114528656005859, 'timing/data': 0.011548258364200592, 'timing/model': 1.466788088902831, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 0.15}) +[2026-06-15 14:14:26,150][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.18866901099681854, 'train/grad_norm_pre_clip': 4.958640098571777, 'timing/data': 0.012539173476397991, 'timing/model': 1.4685020819306374, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 0.15}) +[2026-06-15 14:14:32,047][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.6087665557861328, 'train/grad_norm_pre_clip': 17.73686981201172, 'timing/data': 0.010981854051351547, 'timing/model': 1.4309659311547875, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 0.15}) +[2026-06-15 14:14:37,987][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.13894976675510406, 'train/grad_norm_pre_clip': 3.674639940261841, 'timing/data': 0.012048941105604172, 'timing/model': 1.7022852469235659, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.15}) +[2026-06-15 14:14:43,899][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.31591519713401794, 'train/grad_norm_pre_clip': 4.375698089599609, 'timing/data': 0.010980263352394104, 'timing/model': 1.718059471808374, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 0.15}) +[2026-06-15 14:14:49,556][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.35581091046333313, 'train/grad_norm_pre_clip': 7.125988483428955, 'timing/data': 0.011782879009842873, 'timing/model': 1.4511855226010084, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 0.15}) +[2026-06-15 14:14:55,449][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.17132580280303955, 'train/grad_norm_pre_clip': 3.951028823852539, 'timing/data': 0.01105597335845232, 'timing/model': 1.4392850063741207, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 0.15}) +[2026-06-15 14:15:01,361][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.2950167655944824, 'train/grad_norm_pre_clip': 5.933504581451416, 'timing/data': 0.011288424953818321, 'timing/model': 1.4767395667731762, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 0.15}) +[2026-06-15 14:15:07,250][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.18646107614040375, 'train/grad_norm_pre_clip': 4.6395368576049805, 'timing/data': 0.012493172660470009, 'timing/model': 1.429468186572194, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 0.15}) +[2026-06-15 14:15:13,183][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.2734624147415161, 'train/grad_norm_pre_clip': 9.583696365356445, 'timing/data': 0.011707638390362263, 'timing/model': 1.6799750654026866, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 0.15}) +[2026-06-15 14:15:18,853][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.32750970125198364, 'train/grad_norm_pre_clip': 7.738926410675049, 'timing/data': 0.01123681478202343, 'timing/model': 1.4686470218002796, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 0.15}) +[2026-06-15 14:15:24,751][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.28078493475914, 'train/grad_norm_pre_clip': 7.876518726348877, 'timing/data': 0.011689146980643272, 'timing/model': 1.4578126026317477, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 0.15}) +[2026-06-15 14:15:30,738][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.21970023214817047, 'train/grad_norm_pre_clip': 5.97686243057251, 'timing/data': 0.012862224131822586, 'timing/model': 1.5493196872994304, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 0.16}) +[2026-06-15 14:15:36,745][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.2864794433116913, 'train/grad_norm_pre_clip': 6.549051284790039, 'timing/data': 0.011559206061065197, 'timing/model': 1.500257727690041, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.16}) +[2026-06-15 14:15:42,780][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.24275639653205872, 'train/grad_norm_pre_clip': 3.914106607437134, 'timing/data': 0.010714979842305183, 'timing/model': 1.517478204332292, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 0.16}) +[2026-06-15 14:15:48,668][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.1645665168762207, 'train/grad_norm_pre_clip': 6.34453821182251, 'timing/data': 0.012788273394107819, 'timing/model': 1.6778557654470205, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 0.16}) +[2026-06-15 14:15:54,358][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.23529304563999176, 'train/grad_norm_pre_clip': 6.721196174621582, 'timing/data': 0.010868902318179607, 'timing/model': 1.4665763573721051, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 0.16}) +[2026-06-15 14:16:00,254][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.2935837209224701, 'train/grad_norm_pre_clip': 4.961915493011475, 'timing/data': 0.013186336494982243, 'timing/model': 1.460653644055128, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 0.16}) +[2026-06-15 14:16:06,145][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.4429280459880829, 'train/grad_norm_pre_clip': 10.387226104736328, 'timing/data': 0.016568797640502453, 'timing/model': 1.4608887387439609, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.16}) +[2026-06-15 14:16:12,048][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.3911390006542206, 'train/grad_norm_pre_clip': 12.615862846374512, 'timing/data': 0.01255643181502819, 'timing/model': 1.4301413241773844, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 0.16}) +[2026-06-15 14:16:18,147][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.4472391903400421, 'train/grad_norm_pre_clip': 8.902873992919922, 'timing/data': 0.01107008196413517, 'timing/model': 1.6166242994368076, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 0.16}) +[2026-06-15 14:16:24,052][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.27173376083374023, 'train/grad_norm_pre_clip': 3.9702794551849365, 'timing/data': 0.011770766228437424, 'timing/model': 1.6446758853271604, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 0.16}) +[2026-06-15 14:16:29,751][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.24293194711208344, 'train/grad_norm_pre_clip': 3.1483442783355713, 'timing/data': 0.011266922578215599, 'timing/model': 1.444925306364894, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 0.16}) +[2026-06-15 14:16:35,657][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.19019415974617004, 'train/grad_norm_pre_clip': 4.659702777862549, 'timing/data': 0.011160623282194138, 'timing/model': 1.4731445638462901, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.16}) +[2026-06-15 14:16:41,655][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.1842157542705536, 'train/grad_norm_pre_clip': 4.263082504272461, 'timing/data': 0.01070592924952507, 'timing/model': 1.5709959091618657, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 0.16}) +[2026-06-15 14:16:47,556][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.1317797303199768, 'train/grad_norm_pre_clip': 7.357705593109131, 'timing/data': 0.011453663930296898, 'timing/model': 1.425946058705449, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 0.16}) +[2026-06-15 14:16:53,520][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.049916625022888184, 'train/grad_norm_pre_clip': 2.5249838829040527, 'timing/data': 0.01164784375578165, 'timing/model': 1.7329311398789287, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 0.16}) +[2026-06-15 14:16:59,168][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.2985248267650604, 'train/grad_norm_pre_clip': 13.493168830871582, 'timing/data': 0.015114286914467812, 'timing/model': 1.4680408416315913, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 0.16}) +[2026-06-15 14:17:05,056][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.16467894613742828, 'train/grad_norm_pre_clip': 9.52988052368164, 'timing/data': 0.012118427082896233, 'timing/model': 1.471130508929491, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 0.16}) +[2026-06-15 14:17:10,950][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.6070780754089355, 'train/grad_norm_pre_clip': 16.841081619262695, 'timing/data': 0.011627314612269402, 'timing/model': 1.4575552884489298, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 0.16}) +[2026-06-15 14:17:16,856][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.33377963304519653, 'train/grad_norm_pre_clip': 16.650094985961914, 'timing/data': 0.010907519608736038, 'timing/model': 1.4840610027313232, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 0.16}) +[2026-06-15 14:17:22,866][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.22300982475280762, 'train/grad_norm_pre_clip': 12.45122241973877, 'timing/data': 0.011500203050673008, 'timing/model': 1.5745716067031026, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 0.16}) +[2026-06-15 14:17:28,789][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.4795783758163452, 'train/grad_norm_pre_clip': 21.2382869720459, 'timing/data': 0.01157656405121088, 'timing/model': 1.6916474290192127, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 0.16}) +[2026-06-15 14:17:34,457][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.3245786130428314, 'train/grad_norm_pre_clip': 11.94478702545166, 'timing/data': 0.011952565982937813, 'timing/model': 1.4744052710011601, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.16}) +[2026-06-15 14:17:40,362][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.38417068123817444, 'train/grad_norm_pre_clip': 11.687577247619629, 'timing/data': 0.011457952670753002, 'timing/model': 1.4761843597516418, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 0.17}) +[2026-06-15 14:17:46,254][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.3407905101776123, 'train/grad_norm_pre_clip': 17.831693649291992, 'timing/data': 0.010683158412575722, 'timing/model': 1.4659789521247149, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 0.17}) +[2026-06-15 14:17:52,158][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.3230976462364197, 'train/grad_norm_pre_clip': 13.693485260009766, 'timing/data': 0.011570662260055542, 'timing/model': 1.479594056494534, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 0.17}) +[2026-06-15 14:17:58,058][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.11536700278520584, 'train/grad_norm_pre_clip': 4.3050384521484375, 'timing/data': 0.01153901219367981, 'timing/model': 1.46657903958112, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 0.17}) +[2026-06-15 14:18:03,996][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.5568674802780151, 'train/grad_norm_pre_clip': 14.76622200012207, 'timing/data': 0.011329892091453075, 'timing/model': 1.7077894723042846, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 0.17}) +[2026-06-15 14:18:09,656][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.15515808761119843, 'train/grad_norm_pre_clip': 7.545521259307861, 'timing/data': 0.010386855341494083, 'timing/model': 1.4683877164497972, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 0.17}) +[2026-06-15 14:18:15,545][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.07420489192008972, 'train/grad_norm_pre_clip': 2.1969056129455566, 'timing/data': 0.010360625572502613, 'timing/model': 1.4486045399680734, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 0.17}) +[2026-06-15 14:18:21,440][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.2752953767776489, 'train/grad_norm_pre_clip': 10.163825988769531, 'timing/data': 0.010604585520923138, 'timing/model': 1.4539213916286826, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 0.17}) +[2026-06-15 14:18:27,346][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.38776856660842896, 'train/grad_norm_pre_clip': 9.74793815612793, 'timing/data': 0.010611177422106266, 'timing/model': 1.4429500307887793, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 0.17}) +[2026-06-15 14:18:33,241][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.24965664744377136, 'train/grad_norm_pre_clip': 19.307296752929688, 'timing/data': 0.01078609749674797, 'timing/model': 1.4419939843937755, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.17}) +[2026-06-15 14:18:39,183][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.41816914081573486, 'train/grad_norm_pre_clip': 14.502909660339355, 'timing/data': 0.011062048375606537, 'timing/model': 1.7029965287074447, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 0.17}) +[2026-06-15 14:18:44,849][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.38897162675857544, 'train/grad_norm_pre_clip': 21.99690055847168, 'timing/data': 0.010957219637930393, 'timing/model': 1.460433667525649, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 0.17}) +[2026-06-15 14:18:50,749][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.2666162848472595, 'train/grad_norm_pre_clip': 16.784259796142578, 'timing/data': 0.010150774382054806, 'timing/model': 1.4619763251394033, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 0.17}) +[2026-06-15 14:18:56,647][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.2837962210178375, 'train/grad_norm_pre_clip': 10.05117130279541, 'timing/data': 0.010953608900308609, 'timing/model': 1.4681134521961212, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 0.17}) +[2026-06-15 14:19:02,542][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.3390088677406311, 'train/grad_norm_pre_clip': 19.64451026916504, 'timing/data': 0.011512751691043377, 'timing/model': 1.4553591711446643, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 0.17}) +[2026-06-15 14:19:08,437][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.3205116093158722, 'train/grad_norm_pre_clip': 11.017953872680664, 'timing/data': 0.010114702396094799, 'timing/model': 1.420131721533835, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 0.17}) +[2026-06-15 14:19:14,382][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.2548310160636902, 'train/grad_norm_pre_clip': 6.590185165405273, 'timing/data': 0.011185429990291595, 'timing/model': 1.6893246658146381, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 0.17}) +[2026-06-15 14:19:20,055][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.24063755571842194, 'train/grad_norm_pre_clip': 4.899947166442871, 'timing/data': 0.010921458713710308, 'timing/model': 1.4720553429797292, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 0.17}) +[2026-06-15 14:19:25,944][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.23409929871559143, 'train/grad_norm_pre_clip': 4.955730438232422, 'timing/data': 0.010940037667751312, 'timing/model': 1.4531468246132135, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 0.17}) +[2026-06-15 14:19:31,855][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.30513882637023926, 'train/grad_norm_pre_clip': 7.0955095291137695, 'timing/data': 0.012661849148571491, 'timing/model': 1.4660251252353191, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.17}) +[2026-06-15 14:19:37,845][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.2828209400177002, 'train/grad_norm_pre_clip': 7.851988315582275, 'timing/data': 0.010316303931176662, 'timing/model': 1.4566147262230515, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 0.17}) +[2026-06-15 14:19:43,840][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.2376282960176468, 'train/grad_norm_pre_clip': 6.128300189971924, 'timing/data': 0.011633112095296383, 'timing/model': 1.5020437771454453, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 0.17}) +[2026-06-15 14:19:49,791][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.5980912446975708, 'train/grad_norm_pre_clip': 18.826805114746094, 'timing/data': 0.010349893942475319, 'timing/model': 1.7064376398921013, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 0.18}) +[2026-06-15 14:19:55,456][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.4967353343963623, 'train/grad_norm_pre_clip': 12.838027954101562, 'timing/data': 0.010729695670306683, 'timing/model': 1.4613798456266522, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 0.18}) +[2026-06-15 14:20:01,356][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.34457820653915405, 'train/grad_norm_pre_clip': 13.137282371520996, 'timing/data': 0.011780112981796265, 'timing/model': 1.4617885574698448, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 0.18}) +[2026-06-15 14:20:07,362][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.15833976864814758, 'train/grad_norm_pre_clip': 5.309937953948975, 'timing/data': 0.011058897711336613, 'timing/model': 1.4659008737653494, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 0.18}) +[2026-06-15 14:20:13,254][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.4354233741760254, 'train/grad_norm_pre_clip': 13.590984344482422, 'timing/data': 0.011486020870506763, 'timing/model': 1.4670534310862422, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 0.18}) +[2026-06-15 14:20:19,151][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.7389640212059021, 'train/grad_norm_pre_clip': 17.931276321411133, 'timing/data': 0.010283143259584904, 'timing/model': 1.4210985088720918, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 0.18}) +[2026-06-15 14:20:25,083][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.5907143354415894, 'train/grad_norm_pre_clip': 14.765910148620605, 'timing/data': 0.0102238729596138, 'timing/model': 1.7047662306576967, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 0.18}) +[2026-06-15 14:20:30,753][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.30133557319641113, 'train/grad_norm_pre_clip': 6.6719207763671875, 'timing/data': 0.011166099458932877, 'timing/model': 1.459308904595673, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.18}) +[2026-06-15 14:20:36,651][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.2601241171360016, 'train/grad_norm_pre_clip': 3.381990671157837, 'timing/data': 0.01122932881116867, 'timing/model': 1.471120967529714, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 0.18}) +[2026-06-15 14:20:42,555][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.3285236358642578, 'train/grad_norm_pre_clip': 4.964339733123779, 'timing/data': 0.011047317646443844, 'timing/model': 1.4580719890072942, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 0.18}) +[2026-06-15 14:20:48,458][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.23575642704963684, 'train/grad_norm_pre_clip': 4.229496002197266, 'timing/data': 0.011434759944677353, 'timing/model': 1.4700215738266706, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 0.18}) +[2026-06-15 14:20:54,359][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.36951833963394165, 'train/grad_norm_pre_clip': 8.491785049438477, 'timing/data': 0.010583925060927868, 'timing/model': 1.4358909456059337, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.18}) +[2026-06-15 14:21:00,282][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.1777731478214264, 'train/grad_norm_pre_clip': 2.7872142791748047, 'timing/data': 0.010676105506718159, 'timing/model': 1.684951338917017, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 0.18}) +[2026-06-15 14:21:06,201][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.34511980414390564, 'train/grad_norm_pre_clip': 10.908967018127441, 'timing/data': 0.012261385098099709, 'timing/model': 1.6715069059282541, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 0.18}) +[2026-06-15 14:21:11,862][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.27918317914009094, 'train/grad_norm_pre_clip': 4.112654685974121, 'timing/data': 0.011574631556868553, 'timing/model': 1.4370763069018722, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 0.18}) +[2026-06-15 14:21:17,758][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.3558635115623474, 'train/grad_norm_pre_clip': 9.28042221069336, 'timing/data': 0.011716540902853012, 'timing/model': 1.452973684296012, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 0.18}) +[2026-06-15 14:21:23,662][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.3018770217895508, 'train/grad_norm_pre_clip': 10.614566802978516, 'timing/data': 0.013630053028464317, 'timing/model': 1.4670854937285185, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 0.18}) +[2026-06-15 14:32:15,943][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.13342617452144623, 'train/grad_norm_pre_clip': 3.82645320892334, 'eval/loss': 0.1561867743730545, 'eval/flap_precision': 0.8823529411764706, 'eval/flap_recall': 0.32727272727272727, 'eval/flap_f1': 0.4774535809018568, 'eval/flap_rate_model': 0.028333333333333332, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.47745358090185686, 'timing/data': 0.01176783163100481, 'timing/model': 1.4277491550892591, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.18}) +[2026-06-15 14:32:15,944][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_400_state +[2026-06-15 14:32:15,951][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 14:32:23,800][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_400_state/model.safetensors +[2026-06-15 14:32:34,799][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_400_state/optimizer.bin +[2026-06-15 14:32:34,799][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_400_state/sampler.bin +[2026-06-15 14:32:34,800][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_400_state/sampler_1.bin +[2026-06-15 14:32:34,801][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-15 14:32:34,964][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-15 14:32:34,965][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 14:32:34,980][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 14:32:40,945][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.2706897258758545, 'train/grad_norm_pre_clip': 4.971735954284668, 'timing/data': 0.010287477634847164, 'timing/model': 1.4553222609683871, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 0.18}) +[2026-06-15 14:32:46,894][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.6686643958091736, 'train/grad_norm_pre_clip': 12.55902099609375, 'timing/data': 0.012114619836211205, 'timing/model': 1.7104518078267574, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 0.18}) +[2026-06-15 14:32:52,763][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.5029110908508301, 'train/grad_norm_pre_clip': 13.660428047180176, 'timing/data': 0.013213476166129112, 'timing/model': 1.6757965860888362, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 0.18}) +[2026-06-15 14:32:58,439][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.19178928434848785, 'train/grad_norm_pre_clip': 3.33719801902771, 'timing/data': 0.011448315344750881, 'timing/model': 1.473922410979867, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 0.18}) +[2026-06-15 14:33:04,342][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.2578122913837433, 'train/grad_norm_pre_clip': 5.244287014007568, 'timing/data': 0.011522515676915646, 'timing/model': 1.4592208843678236, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 0.19}) +[2026-06-15 14:33:10,233][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.14048461616039276, 'train/grad_norm_pre_clip': 3.8979437351226807, 'timing/data': 0.011262772604823112, 'timing/model': 1.4609612934291363, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 0.19}) +[2026-06-15 14:33:16,137][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.16509628295898438, 'train/grad_norm_pre_clip': 2.7596421241760254, 'timing/data': 0.01078210398554802, 'timing/model': 1.4207752011716366, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 0.19}) +[2026-06-15 14:33:22,035][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.46584171056747437, 'train/grad_norm_pre_clip': 16.891265869140625, 'timing/data': 0.011067084968090057, 'timing/model': 1.4189319554716349, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 0.19}) +[2026-06-15 14:33:27,948][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.40492570400238037, 'train/grad_norm_pre_clip': 13.865236282348633, 'timing/data': 0.011263484135270119, 'timing/model': 1.6723021306097507, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 0.19}) +[2026-06-15 14:33:33,647][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.16173112392425537, 'train/grad_norm_pre_clip': 7.6398491859436035, 'timing/data': 0.010927415452897549, 'timing/model': 1.4618440819904208, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.19}) +[2026-06-15 14:33:39,550][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.17940732836723328, 'train/grad_norm_pre_clip': 5.114536285400391, 'timing/data': 0.011194596998393536, 'timing/model': 1.473224252462387, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 0.19}) +[2026-06-15 14:33:45,450][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.30704763531684875, 'train/grad_norm_pre_clip': 7.062031269073486, 'timing/data': 0.011423306539654732, 'timing/model': 1.4708992261439562, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 0.19}) +[2026-06-15 14:33:51,352][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.23884892463684082, 'train/grad_norm_pre_clip': 4.819484233856201, 'timing/data': 0.012335327453911304, 'timing/model': 1.425568686798215, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 0.19}) +[2026-06-15 14:33:57,283][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.21359050273895264, 'train/grad_norm_pre_clip': 9.44355583190918, 'timing/data': 0.011676229536533356, 'timing/model': 1.7030649557709694, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 0.19}) +[2026-06-15 14:34:02,951][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.2709694802761078, 'train/grad_norm_pre_clip': 5.388897895812988, 'timing/data': 0.011664588935673237, 'timing/model': 1.4698740541934967, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.19}) +[2026-06-15 14:34:08,940][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.28858453035354614, 'train/grad_norm_pre_clip': 7.1832275390625, 'timing/data': 0.011129939928650856, 'timing/model': 1.4956040438264608, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 0.19}) +[2026-06-15 14:34:14,752][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.1411595195531845, 'train/grad_norm_pre_clip': 3.4942307472229004, 'timing/data': 0.01020768005400896, 'timing/model': 1.4690216230228543, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 0.19}) +[2026-06-15 14:34:20,653][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.38605964183807373, 'train/grad_norm_pre_clip': 8.61151123046875, 'timing/data': 0.011091100983321667, 'timing/model': 1.4706255495548248, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 0.19}) +[2026-06-15 14:34:26,558][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.31441980600357056, 'train/grad_norm_pre_clip': 9.452659606933594, 'timing/data': 0.01046766061335802, 'timing/model': 1.4733147593215108, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 0.19}) +[2026-06-15 14:34:32,486][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.22729703783988953, 'train/grad_norm_pre_clip': 4.206241130828857, 'timing/data': 0.010987062007188797, 'timing/model': 1.699372093193233, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.19}) +[2026-06-15 14:34:38,153][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.17906975746154785, 'train/grad_norm_pre_clip': 3.111638069152832, 'timing/data': 0.012723593041300774, 'timing/model': 1.4571626521646976, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 0.19}) +[2026-06-15 14:34:44,158][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.192257821559906, 'train/grad_norm_pre_clip': 5.431822299957275, 'timing/data': 0.010473261587321758, 'timing/model': 1.4762039119377732, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 0.19}) +[2026-06-15 14:34:50,059][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.35606643557548523, 'train/grad_norm_pre_clip': 8.23165512084961, 'timing/data': 0.010882293805480003, 'timing/model': 1.4763183146715164, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 0.19}) +[2026-06-15 14:34:55,957][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.17797327041625977, 'train/grad_norm_pre_clip': 3.1083762645721436, 'timing/data': 0.01111904438585043, 'timing/model': 1.471299803815782, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 0.19}) +[2026-06-15 14:35:01,852][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.36397457122802734, 'train/grad_norm_pre_clip': 6.017487049102783, 'timing/data': 0.011025285348296165, 'timing/model': 1.435308770276606, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 0.19}) +[2026-06-15 14:35:07,784][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.43304359912872314, 'train/grad_norm_pre_clip': 10.854354858398438, 'timing/data': 0.010724875144660473, 'timing/model': 1.6901528593152761, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 0.19}) +[2026-06-15 14:35:13,456][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.35082221031188965, 'train/grad_norm_pre_clip': 6.288649559020996, 'timing/data': 0.011511586606502533, 'timing/model': 1.4620968215167522, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 0.2}) +[2026-06-15 14:35:19,357][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.28364843130111694, 'train/grad_norm_pre_clip': 4.590484619140625, 'timing/data': 0.013277146965265274, 'timing/model': 1.4630466001108289, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 0.2}) +[2026-06-15 14:35:25,253][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.31796732544898987, 'train/grad_norm_pre_clip': 3.4011778831481934, 'timing/data': 0.011719896458089352, 'timing/model': 1.468756060115993, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 0.2}) +[2026-06-15 14:35:31,180][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.3131914436817169, 'train/grad_norm_pre_clip': 4.441717147827148, 'timing/data': 0.010956496931612492, 'timing/model': 1.4519620463252068, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.2}) +[2026-06-15 14:35:37,356][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.2141607105731964, 'train/grad_norm_pre_clip': 3.938211441040039, 'timing/data': 0.011130508035421371, 'timing/model': 1.6493433834984899, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 0.2}) +[2026-06-15 14:35:43,298][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.22618286311626434, 'train/grad_norm_pre_clip': 5.786848545074463, 'timing/data': 0.012179058976471424, 'timing/model': 1.7016603965312243, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 0.2}) +[2026-06-15 14:35:49,146][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.11865464597940445, 'train/grad_norm_pre_clip': 2.5538206100463867, 'timing/data': 0.010462948121130466, 'timing/model': 1.5078440876677632, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 0.2}) +[2026-06-15 14:35:55,258][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.2796503007411957, 'train/grad_norm_pre_clip': 8.320230484008789, 'timing/data': 0.01120171882212162, 'timing/model': 1.448122227564454, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 0.2}) +[2026-06-15 14:36:01,270][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.17002378404140472, 'train/grad_norm_pre_clip': 6.734285354614258, 'timing/data': 0.011595780029892921, 'timing/model': 1.4745512045919895, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 0.2}) +[2026-06-15 14:36:07,164][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.2046177238225937, 'train/grad_norm_pre_clip': 5.869234085083008, 'timing/data': 0.01104724034667015, 'timing/model': 1.4545965380966663, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 0.2}) +[2026-06-15 14:36:13,164][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.29245567321777344, 'train/grad_norm_pre_clip': 10.65650463104248, 'timing/data': 0.012831151485443115, 'timing/model': 1.5271305106580257, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 0.2}) +[2026-06-15 14:36:19,283][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.1842889040708542, 'train/grad_norm_pre_clip': 4.602198600769043, 'timing/data': 0.011393650434911251, 'timing/model': 1.7901326613500714, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 0.2}) +[2026-06-15 14:36:24,980][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.28877952694892883, 'train/grad_norm_pre_clip': 7.402559757232666, 'timing/data': 0.011761131696403027, 'timing/model': 1.476067053154111, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 0.2}) +[2026-06-15 14:36:30,963][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.2560214400291443, 'train/grad_norm_pre_clip': 6.514631748199463, 'timing/data': 0.011045161634683609, 'timing/model': 1.4639004711061716, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.2}) +[2026-06-15 14:36:36,869][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.316985547542572, 'train/grad_norm_pre_clip': 9.61193561553955, 'timing/data': 0.014704027213156223, 'timing/model': 1.4416948352009058, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 0.2}) +[2026-06-15 14:36:42,769][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.19366677105426788, 'train/grad_norm_pre_clip': 3.9519388675689697, 'timing/data': 0.011494114063680172, 'timing/model': 1.471871587447822, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 0.2}) +[2026-06-15 14:36:48,741][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.35340893268585205, 'train/grad_norm_pre_clip': 11.740828514099121, 'timing/data': 0.011469154618680477, 'timing/model': 1.505788047797978, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.2}) +[2026-06-15 14:36:54,694][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.5439872741699219, 'train/grad_norm_pre_clip': 15.072442054748535, 'timing/data': 0.011615054681897163, 'timing/model': 1.6948609743267298, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 0.2}) +[2026-06-15 14:37:00,359][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.3529313802719116, 'train/grad_norm_pre_clip': 9.251835823059082, 'timing/data': 0.010281292721629143, 'timing/model': 1.4645452806726098, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 0.2}) +[2026-06-15 14:37:06,479][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.2138475924730301, 'train/grad_norm_pre_clip': 6.642706871032715, 'timing/data': 0.010634203441441059, 'timing/model': 1.4778982661664486, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 0.2}) +[2026-06-15 14:37:12,581][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.26034310460090637, 'train/grad_norm_pre_clip': 5.498960971832275, 'timing/data': 0.0113567765802145, 'timing/model': 1.4778101528063416, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 0.2}) +[2026-06-15 14:37:18,582][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.262656033039093, 'train/grad_norm_pre_clip': 5.741940021514893, 'timing/data': 0.010690884664654732, 'timing/model': 1.4950409727171063, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 0.2}) +[2026-06-15 14:37:24,580][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.5243983268737793, 'train/grad_norm_pre_clip': 13.928093910217285, 'timing/data': 0.010840165428817272, 'timing/model': 1.5272350860759616, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 0.21}) +[2026-06-15 14:37:30,510][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.2834238111972809, 'train/grad_norm_pre_clip': 21.65024185180664, 'timing/data': 0.01087181642651558, 'timing/model': 1.7063762843608856, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.21}) +[2026-06-15 14:37:36,176][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.1659863442182541, 'train/grad_norm_pre_clip': 2.732037305831909, 'timing/data': 0.010552366264164448, 'timing/model': 1.4730285415425897, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 0.21}) +[2026-06-15 14:37:42,186][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.14307314157485962, 'train/grad_norm_pre_clip': 4.829148769378662, 'timing/data': 0.012029188685119152, 'timing/model': 1.4789442121982574, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 0.21}) +[2026-06-15 14:37:48,062][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.19462987780570984, 'train/grad_norm_pre_clip': 6.815908908843994, 'timing/data': 0.011250806972384453, 'timing/model': 1.4580156123265624, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 0.21}) +[2026-06-15 14:37:53,974][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.19965514540672302, 'train/grad_norm_pre_clip': 4.383954048156738, 'timing/data': 0.01087767630815506, 'timing/model': 1.4711854243651032, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 0.21}) +[2026-06-15 14:38:00,093][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.12559732794761658, 'train/grad_norm_pre_clip': 4.561339855194092, 'timing/data': 0.011098208837211132, 'timing/model': 1.5461961487308145, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 0.21}) +[2026-06-15 14:38:06,020][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.14719511568546295, 'train/grad_norm_pre_clip': 4.203064441680908, 'timing/data': 0.011316309683024883, 'timing/model': 1.7125824121758342, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 0.21}) +[2026-06-15 14:38:11,896][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.26411890983581543, 'train/grad_norm_pre_clip': 6.39626407623291, 'timing/data': 0.011856550350785255, 'timing/model': 1.6900995317846537, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 0.21}) +[2026-06-15 14:38:17,566][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.204859659075737, 'train/grad_norm_pre_clip': 6.7060675621032715, 'timing/data': 0.011689211241900921, 'timing/model': 1.4630191577598453, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 0.21}) +[2026-06-15 14:38:23,457][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.20992496609687805, 'train/grad_norm_pre_clip': 5.675652027130127, 'timing/data': 0.011103370226919651, 'timing/model': 1.453774780035019, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 0.21}) +[2026-06-15 14:38:29,358][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.1976265013217926, 'train/grad_norm_pre_clip': 6.335383415222168, 'timing/data': 0.011199419386684895, 'timing/model': 1.461158229969442, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.21}) +[2026-06-15 14:38:35,305][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.1693342626094818, 'train/grad_norm_pre_clip': 5.354216575622559, 'timing/data': 0.009773788042366505, 'timing/model': 1.464768853969872, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 0.21}) +[2026-06-15 14:38:41,302][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.1898883581161499, 'train/grad_norm_pre_clip': 5.496646881103516, 'timing/data': 0.011650762520730495, 'timing/model': 1.752833061851561, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 0.21}) +[2026-06-15 14:38:46,955][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.30610209703445435, 'train/grad_norm_pre_clip': 10.768230438232422, 'timing/data': 0.010049678385257721, 'timing/model': 1.4530099518597126, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 0.21}) +[2026-06-15 14:38:52,959][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.07047119736671448, 'train/grad_norm_pre_clip': 3.6774485111236572, 'timing/data': 0.011266072280704975, 'timing/model': 1.4562013233080506, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 0.21}) +[2026-06-15 14:38:58,857][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.1763458400964737, 'train/grad_norm_pre_clip': 6.203391075134277, 'timing/data': 0.010680110193789005, 'timing/model': 1.4610520964488387, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 0.21}) +[2026-06-15 14:39:04,851][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.16678886115550995, 'train/grad_norm_pre_clip': 9.22130012512207, 'timing/data': 0.010874181985855103, 'timing/model': 1.512998316437006, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 0.21}) +[2026-06-15 14:39:10,769][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.23896946012973785, 'train/grad_norm_pre_clip': 6.48091983795166, 'timing/data': 0.010436231270432472, 'timing/model': 1.4830986009910703, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 0.21}) +[2026-06-15 14:39:16,765][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.22693364322185516, 'train/grad_norm_pre_clip': 8.257315635681152, 'timing/data': 0.011905765160918236, 'timing/model': 1.6477903546765447, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 0.21}) +[2026-06-15 14:39:22,466][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.26087823510169983, 'train/grad_norm_pre_clip': 6.982469081878662, 'timing/data': 0.011445694603025913, 'timing/model': 1.4559072805568576, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 0.21}) +[2026-06-15 14:39:28,466][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.26667770743370056, 'train/grad_norm_pre_clip': 10.326371192932129, 'timing/data': 0.011839495040476322, 'timing/model': 1.4604073027148843, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.21}) +[2026-06-15 14:39:34,464][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.25406280159950256, 'train/grad_norm_pre_clip': 12.731924057006836, 'timing/data': 0.011903185397386551, 'timing/model': 1.4781719222664833, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 0.22}) +[2026-06-15 14:39:40,364][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.24099723994731903, 'train/grad_norm_pre_clip': 8.707839012145996, 'timing/data': 0.011209405027329922, 'timing/model': 1.4283776860684156, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 0.22}) +[2026-06-15 14:39:46,367][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.232081338763237, 'train/grad_norm_pre_clip': 8.294657707214355, 'timing/data': 0.012569038197398186, 'timing/model': 1.515115637332201, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 0.22}) +[2026-06-15 14:39:52,255][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.12768307328224182, 'train/grad_norm_pre_clip': 3.182158946990967, 'timing/data': 0.011165385134518147, 'timing/model': 1.661907327361405, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 0.22}) +[2026-06-15 14:39:57,963][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.2348567247390747, 'train/grad_norm_pre_clip': 9.401294708251953, 'timing/data': 0.010864504612982273, 'timing/model': 1.4706780584529042, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 0.22}) +[2026-06-15 14:40:03,858][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.20159056782722473, 'train/grad_norm_pre_clip': 4.784831523895264, 'timing/data': 0.011465655639767647, 'timing/model': 1.4578130478039384, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 0.22}) +[2026-06-15 14:40:09,759][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.30942219495773315, 'train/grad_norm_pre_clip': 7.785287380218506, 'timing/data': 0.011517886072397232, 'timing/model': 1.459851392544806, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 0.22}) +[2026-06-15 14:40:15,772][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.3913264572620392, 'train/grad_norm_pre_clip': 11.274641036987305, 'timing/data': 0.010522433556616306, 'timing/model': 1.438623451627791, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 0.22}) +[2026-06-15 14:40:21,692][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.30534008145332336, 'train/grad_norm_pre_clip': 7.000364780426025, 'timing/data': 0.010287844575941563, 'timing/model': 1.694576715119183, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 0.22}) +[2026-06-15 14:40:27,361][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.27307552099227905, 'train/grad_norm_pre_clip': 8.61821460723877, 'timing/data': 0.01116542611271143, 'timing/model': 1.4590004244819283, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.22}) +[2026-06-15 14:40:33,262][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.12494401633739471, 'train/grad_norm_pre_clip': 3.275853395462036, 'timing/data': 0.01124498713761568, 'timing/model': 1.4643724132329226, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 0.22}) +[2026-06-15 14:40:39,162][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.2531712055206299, 'train/grad_norm_pre_clip': 6.370634078979492, 'timing/data': 0.010990015231072903, 'timing/model': 1.4621413350105286, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 0.22}) +[2026-06-15 14:40:45,066][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.10464087128639221, 'train/grad_norm_pre_clip': 3.1169118881225586, 'timing/data': 0.011260297149419785, 'timing/model': 1.4611347578465939, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 0.22}) +[2026-06-15 14:40:50,959][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.18952785432338715, 'train/grad_norm_pre_clip': 6.778674125671387, 'timing/data': 0.010972677730023861, 'timing/model': 1.4562237914651632, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 0.22}) +[2026-06-15 14:40:57,142][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.21334674954414368, 'train/grad_norm_pre_clip': 5.071995735168457, 'timing/data': 0.010140614584088326, 'timing/model': 1.8493301272392273, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.22}) +[2026-06-15 14:41:02,871][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.2768460810184479, 'train/grad_norm_pre_clip': 8.848855972290039, 'timing/data': 0.011071277782320976, 'timing/model': 1.4652105253189802, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 0.22}) +[2026-06-15 14:41:08,930][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.4245581328868866, 'train/grad_norm_pre_clip': 8.557498931884766, 'timing/data': 0.01120853889733553, 'timing/model': 1.5161128407344222, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 0.22}) +[2026-06-15 14:41:14,979][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.42131978273391724, 'train/grad_norm_pre_clip': 14.559805870056152, 'timing/data': 0.012602422386407852, 'timing/model': 1.4612291157245636, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 0.22}) +[2026-06-15 14:41:21,068][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.2921876609325409, 'train/grad_norm_pre_clip': 7.287578582763672, 'timing/data': 0.011292638257145882, 'timing/model': 1.4666837248951197, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 0.22}) +[2026-06-15 14:41:27,066][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.2978963255882263, 'train/grad_norm_pre_clip': 8.899565696716309, 'timing/data': 0.009618694894015789, 'timing/model': 1.449119672179222, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 0.22}) +[2026-06-15 14:41:33,098][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.2550683915615082, 'train/grad_norm_pre_clip': 5.016684055328369, 'timing/data': 0.009860215708613396, 'timing/model': 1.7081205854192376, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 0.22}) +[2026-06-15 14:41:38,768][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.23685573041439056, 'train/grad_norm_pre_clip': 6.7694597244262695, 'timing/data': 0.010402577929198742, 'timing/model': 1.463116655126214, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.23}) +[2026-06-15 14:41:44,779][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.21509212255477905, 'train/grad_norm_pre_clip': 6.029887676239014, 'timing/data': 0.011215920560061932, 'timing/model': 1.4702918967232108, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 0.23}) +[2026-06-15 14:41:50,770][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.3763899505138397, 'train/grad_norm_pre_clip': 9.825297355651855, 'timing/data': 0.01130674034357071, 'timing/model': 1.4625700088217854, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 0.23}) +[2026-06-15 14:41:56,774][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.1920633167028427, 'train/grad_norm_pre_clip': 6.844649791717529, 'timing/data': 0.011384820565581322, 'timing/model': 1.514046442694962, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 0.23}) +[2026-06-15 14:42:02,787][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.307807058095932, 'train/grad_norm_pre_clip': 7.7620344161987305, 'timing/data': 0.01092154998332262, 'timing/model': 1.4532463867217302, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 0.23}) +[2026-06-15 14:42:08,721][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.23380684852600098, 'train/grad_norm_pre_clip': 10.759994506835938, 'timing/data': 0.012366955168545246, 'timing/model': 1.7112982170656323, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 0.23}) +[2026-06-15 14:42:14,371][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.08675238490104675, 'train/grad_norm_pre_clip': 2.6452016830444336, 'timing/data': 0.010916720144450665, 'timing/model': 1.454179179854691, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 0.23}) +[2026-06-15 14:42:20,479][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.1858932077884674, 'train/grad_norm_pre_clip': 4.0143842697143555, 'timing/data': 0.010415149852633476, 'timing/model': 1.4694615146145225, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 0.23}) +[2026-06-15 14:53:05,255][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.36771973967552185, 'train/grad_norm_pre_clip': 34.88286590576172, 'eval/loss': 0.22115375101566315, 'eval/flap_precision': 1.0, 'eval/flap_recall': 0.1890909090909091, 'eval/flap_f1': 0.3180428134556575, 'eval/flap_rate_model': 0.014444444444444444, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.3180428134556575, 'timing/data': 0.010606140829622746, 'timing/model': 1.5419884715229273, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 0.23}) +[2026-06-15 14:53:05,256][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_500_state +[2026-06-15 14:53:05,263][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 14:53:13,113][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_500_state/model.safetensors +[2026-06-15 14:53:24,093][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_500_state/optimizer.bin +[2026-06-15 14:53:24,094][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_500_state/sampler.bin +[2026-06-15 14:53:24,095][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_500_state/sampler_1.bin +[2026-06-15 14:53:24,096][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-15 14:53:24,251][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-15 14:53:24,251][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 14:53:24,266][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 14:53:30,225][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.3310079574584961, 'train/grad_norm_pre_clip': 7.766777515411377, 'timing/data': 0.011594868265092373, 'timing/model': 1.4318190114572644, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 0.23}) +[2026-06-15 14:53:36,042][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.5084663033485413, 'train/grad_norm_pre_clip': 11.079358100891113, 'timing/data': 0.010943484492599964, 'timing/model': 1.4804905587807298, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 0.23}) +[2026-06-15 14:53:41,919][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.2117585837841034, 'train/grad_norm_pre_clip': 8.245240211486816, 'timing/data': 0.014147687703371048, 'timing/model': 1.529536770656705, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 0.23}) +[2026-06-15 14:53:48,036][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.5697988271713257, 'train/grad_norm_pre_clip': 14.694727897644043, 'timing/data': 0.02039860375225544, 'timing/model': 1.4714543586596847, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 0.23}) +[2026-06-15 14:53:54,072][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.40653353929519653, 'train/grad_norm_pre_clip': 6.699545383453369, 'timing/data': 0.010867875069379807, 'timing/model': 1.6911880681291223, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 0.23}) +[2026-06-15 14:53:59,753][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.3360399901866913, 'train/grad_norm_pre_clip': 5.783677577972412, 'timing/data': 0.011182375252246857, 'timing/model': 1.4698327369987965, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 0.23}) +[2026-06-15 14:54:05,647][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.24142448604106903, 'train/grad_norm_pre_clip': 4.53809928894043, 'timing/data': 0.010819794610142708, 'timing/model': 1.4639635412022471, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 0.23}) +[2026-06-15 14:54:11,548][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.5069880485534668, 'train/grad_norm_pre_clip': 11.965069770812988, 'timing/data': 0.010710334405303001, 'timing/model': 1.4484915938228369, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 0.23}) +[2026-06-15 14:54:17,456][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.5950638651847839, 'train/grad_norm_pre_clip': 20.2001953125, 'timing/data': 0.011368446983397007, 'timing/model': 1.455044568516314, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 0.23}) +[2026-06-15 14:54:23,425][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.5127426981925964, 'train/grad_norm_pre_clip': 15.16783332824707, 'timing/data': 0.010044452734291553, 'timing/model': 1.4884514156728983, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 0.23}) +[2026-06-15 14:54:29,293][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.5952839255332947, 'train/grad_norm_pre_clip': 14.909202575683594, 'timing/data': 0.01148808840662241, 'timing/model': 1.7201852295547724, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 0.23}) +[2026-06-15 14:54:34,942][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.30749377608299255, 'train/grad_norm_pre_clip': 7.3979668617248535, 'timing/data': 0.011197277344763279, 'timing/model': 1.4619413781911135, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 0.23}) +[2026-06-15 14:54:40,851][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.39172008633613586, 'train/grad_norm_pre_clip': 8.223121643066406, 'timing/data': 0.010753675363957882, 'timing/model': 1.4717168072238564, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 0.23}) +[2026-06-15 14:54:46,746][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.22709670662879944, 'train/grad_norm_pre_clip': 3.019716739654541, 'timing/data': 0.010188601911067963, 'timing/model': 1.4660716094076633, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 0.24}) +[2026-06-15 14:54:52,657][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.47017985582351685, 'train/grad_norm_pre_clip': 10.71819019317627, 'timing/data': 0.010658984072506428, 'timing/model': 1.4706553956493735, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 0.24}) +[2026-06-15 14:54:58,650][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.30414536595344543, 'train/grad_norm_pre_clip': 4.240230083465576, 'timing/data': 0.010580855421721935, 'timing/model': 1.4798227716237307, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 0.24}) +[2026-06-15 14:55:04,592][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.3239808678627014, 'train/grad_norm_pre_clip': 7.312061786651611, 'timing/data': 0.010107063688337803, 'timing/model': 1.7052867570891976, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 0.24}) +[2026-06-15 14:55:10,457][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.21766190230846405, 'train/grad_norm_pre_clip': 4.497844219207764, 'timing/data': 0.01061941497027874, 'timing/model': 1.4640760589390993, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 0.24}) +[2026-06-15 14:55:16,457][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.16716937720775604, 'train/grad_norm_pre_clip': 5.245993614196777, 'timing/data': 0.010584893636405468, 'timing/model': 1.4767374489456415, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 0.24}) +[2026-06-15 14:55:22,383][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.31784284114837646, 'train/grad_norm_pre_clip': 4.921048164367676, 'timing/data': 0.010653965175151825, 'timing/model': 1.4768908163532615, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.24}) +[2026-06-15 14:55:28,466][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.19253715872764587, 'train/grad_norm_pre_clip': 4.065786838531494, 'timing/data': 0.011794649995863438, 'timing/model': 1.4613008173182607, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 0.24}) +[2026-06-15 14:55:34,455][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.2374590039253235, 'train/grad_norm_pre_clip': 4.788424015045166, 'timing/data': 0.011119747534394264, 'timing/model': 1.5242102285847068, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 0.24}) +[2026-06-15 14:55:40,400][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.27170678973197937, 'train/grad_norm_pre_clip': 9.961407661437988, 'timing/data': 0.010632985271513462, 'timing/model': 1.701678317040205, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 0.24}) +[2026-06-15 14:55:46,237][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.3288155496120453, 'train/grad_norm_pre_clip': 4.513186931610107, 'timing/data': 0.01094900630414486, 'timing/model': 1.6373730031773448, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 0.24}) +[2026-06-15 14:55:51,972][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.37328198552131653, 'train/grad_norm_pre_clip': 17.666582107543945, 'timing/data': 0.010169672779738903, 'timing/model': 1.4900605445727706, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 0.24}) +[2026-06-15 14:55:57,870][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.26026585698127747, 'train/grad_norm_pre_clip': 10.559673309326172, 'timing/data': 0.010289603844285011, 'timing/model': 1.4697181833907962, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 0.24}) +[2026-06-15 14:56:04,054][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.47925230860710144, 'train/grad_norm_pre_clip': 6.57023286819458, 'timing/data': 0.010693036019802094, 'timing/model': 1.4655649904161692, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 0.24}) +[2026-06-15 14:56:10,078][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.17640358209609985, 'train/grad_norm_pre_clip': 8.697443008422852, 'timing/data': 0.010551905259490013, 'timing/model': 1.4485117951408029, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 0.24}) +[2026-06-15 14:56:16,000][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.24256590008735657, 'train/grad_norm_pre_clip': 7.366926670074463, 'timing/data': 0.009584181010723114, 'timing/model': 1.7096707541495562, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 0.24}) +[2026-06-15 14:56:21,979][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.16824065148830414, 'train/grad_norm_pre_clip': 4.918493747711182, 'timing/data': 0.011130866594612598, 'timing/model': 1.7859768671914935, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 0.24}) +[2026-06-15 14:56:27,660][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.19115430116653442, 'train/grad_norm_pre_clip': 6.045113563537598, 'timing/data': 0.012058030813932419, 'timing/model': 1.4748161546885967, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 0.24}) +[2026-06-15 14:56:33,572][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.23532907664775848, 'train/grad_norm_pre_clip': 6.448636054992676, 'timing/data': 0.011560468934476376, 'timing/model': 1.4815953196957707, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 0.24}) +[2026-06-15 14:56:39,470][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.21784250438213348, 'train/grad_norm_pre_clip': 6.879188060760498, 'timing/data': 0.011224138550460339, 'timing/model': 1.475741719827056, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 0.24}) +[2026-06-15 14:56:45,359][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.36840903759002686, 'train/grad_norm_pre_clip': 8.907869338989258, 'timing/data': 0.010952937416732311, 'timing/model': 1.4279566202312708, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 0.24}) +[2026-06-15 14:56:51,311][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.20515184104442596, 'train/grad_norm_pre_clip': 3.641005277633667, 'timing/data': 0.01004666369408369, 'timing/model': 1.7011364065110683, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 0.24}) +[2026-06-15 14:56:56,968][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.2726758122444153, 'train/grad_norm_pre_clip': 7.565745830535889, 'timing/data': 0.011087987571954727, 'timing/model': 1.4689106671139598, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 0.25}) +[2026-06-15 14:57:02,955][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.11589714884757996, 'train/grad_norm_pre_clip': 2.233931303024292, 'timing/data': 0.010562635026872158, 'timing/model': 1.4566137101501226, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 0.25}) +[2026-06-15 14:57:08,864][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.37779542803764343, 'train/grad_norm_pre_clip': 7.278345108032227, 'timing/data': 0.010733665898442268, 'timing/model': 1.4620615290477872, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 0.25}) +[2026-06-15 14:57:14,765][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.24631595611572266, 'train/grad_norm_pre_clip': 6.004421234130859, 'timing/data': 0.010347764939069748, 'timing/model': 1.4380681114271283, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 0.25}) +[2026-06-15 14:57:20,667][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.31692442297935486, 'train/grad_norm_pre_clip': 6.16998291015625, 'timing/data': 0.011010386981070042, 'timing/model': 1.495226752012968, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 0.25}) +[2026-06-15 14:57:26,552][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.276676744222641, 'train/grad_norm_pre_clip': 3.704730272293091, 'timing/data': 0.010496665723621845, 'timing/model': 1.6552491756156087, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.25}) +[2026-06-15 14:57:32,268][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.41267865896224976, 'train/grad_norm_pre_clip': 5.5504326820373535, 'timing/data': 0.011436219327151775, 'timing/model': 1.4360406268388033, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 0.25}) +[2026-06-15 14:57:38,160][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.3677805960178375, 'train/grad_norm_pre_clip': 12.475923538208008, 'timing/data': 0.011106807738542557, 'timing/model': 1.4480517841875553, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 0.25}) +[2026-06-15 14:57:44,172][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.3216795027256012, 'train/grad_norm_pre_clip': 8.631972312927246, 'timing/data': 0.010731706395745277, 'timing/model': 1.4743531020358205, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 0.25}) +[2026-06-15 14:57:50,057][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.43889427185058594, 'train/grad_norm_pre_clip': 8.45051097869873, 'timing/data': 0.01124130841344595, 'timing/model': 1.4346389258280396, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 0.25}) +[2026-06-15 14:57:56,050][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.16660504043102264, 'train/grad_norm_pre_clip': 6.079442977905273, 'timing/data': 0.011993162333965302, 'timing/model': 1.5100698061287403, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 0.25}) +[2026-06-15 14:58:01,956][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.19023633003234863, 'train/grad_norm_pre_clip': 4.107092380523682, 'timing/data': 0.011040426790714264, 'timing/model': 1.655808531679213, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 0.25}) +[2026-06-15 14:58:07,662][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.31234046816825867, 'train/grad_norm_pre_clip': 4.72802734375, 'timing/data': 0.011733630672097206, 'timing/model': 1.4597591292113066, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 0.25}) +[2026-06-15 14:58:13,561][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.2871726453304291, 'train/grad_norm_pre_clip': 8.205788612365723, 'timing/data': 0.011982211843132973, 'timing/model': 1.4541767286136746, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 0.25}) +[2026-06-15 14:58:19,568][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.23528414964675903, 'train/grad_norm_pre_clip': 9.036934852600098, 'timing/data': 0.011461598798632622, 'timing/model': 1.4870760468766093, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 0.25}) +[2026-06-15 14:58:25,473][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.5101158618927002, 'train/grad_norm_pre_clip': 18.837291717529297, 'timing/data': 0.01075319666415453, 'timing/model': 1.4380718190222979, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 0.25}) +[2026-06-15 14:58:31,406][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.37439340353012085, 'train/grad_norm_pre_clip': 16.963212966918945, 'timing/data': 0.010463174432516098, 'timing/model': 1.7063333978876472, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 0.25}) +[2026-06-15 14:58:37,068][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.2138996571302414, 'train/grad_norm_pre_clip': 3.833312749862671, 'timing/data': 0.011631610803306103, 'timing/model': 1.4703304171562195, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 0.25}) +[2026-06-15 14:58:42,965][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.4169130027294159, 'train/grad_norm_pre_clip': 8.064536094665527, 'timing/data': 0.011538529768586159, 'timing/model': 1.4621411059051752, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 0.25}) +[2026-06-15 14:58:48,872][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.2175898253917694, 'train/grad_norm_pre_clip': 3.0034289360046387, 'timing/data': 0.011009037494659424, 'timing/model': 1.48009186796844, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.25}) +[2026-06-15 14:58:54,766][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.24710509181022644, 'train/grad_norm_pre_clip': 3.513305187225342, 'timing/data': 0.011288778856396675, 'timing/model': 1.463960031978786, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 0.25}) +[2026-06-15 14:59:00,667][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.25170692801475525, 'train/grad_norm_pre_clip': 5.28566312789917, 'timing/data': 0.010019353590905666, 'timing/model': 1.4752544965595007, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 0.25}) +[2026-06-15 14:59:06,607][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.26888608932495117, 'train/grad_norm_pre_clip': 7.604382038116455, 'timing/data': 0.010783537290990353, 'timing/model': 1.704615119844675, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 0.26}) +[2026-06-15 14:59:12,264][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.3287297487258911, 'train/grad_norm_pre_clip': 7.588674545288086, 'timing/data': 0.011103669181466103, 'timing/model': 1.4534012591466308, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 0.26}) +[2026-06-15 14:59:18,262][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.24414190649986267, 'train/grad_norm_pre_clip': 7.137628078460693, 'timing/data': 0.010693036019802094, 'timing/model': 1.4614136079326272, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 0.26}) +[2026-06-15 14:59:24,161][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.21030203998088837, 'train/grad_norm_pre_clip': 4.17637300491333, 'timing/data': 0.014327683486044407, 'timing/model': 1.4581418978050351, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 0.26}) +[2026-06-15 14:59:30,269][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.20714929699897766, 'train/grad_norm_pre_clip': 3.1374144554138184, 'timing/data': 0.012661915272474289, 'timing/model': 1.5336349122226238, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 0.26}) +[2026-06-15 14:59:36,162][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.09756376594305038, 'train/grad_norm_pre_clip': 1.8634473085403442, 'timing/data': 0.01115585956722498, 'timing/model': 1.441712724044919, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 0.26}) +[2026-06-15 14:59:42,125][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.31091776490211487, 'train/grad_norm_pre_clip': 16.652570724487305, 'timing/data': 0.011546961031854153, 'timing/model': 1.718269856646657, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 0.26}) +[2026-06-15 14:59:47,866][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.06969004124403, 'train/grad_norm_pre_clip': 3.330780506134033, 'timing/data': 0.012824186123907566, 'timing/model': 1.4622433315962553, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 0.26}) +[2026-06-15 14:59:53,862][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.528556764125824, 'train/grad_norm_pre_clip': 19.666851043701172, 'timing/data': 0.011165859177708626, 'timing/model': 1.4546949518844485, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 0.26}) +[2026-06-15 14:59:59,883][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.09464133530855179, 'train/grad_norm_pre_clip': 10.44395637512207, 'timing/data': 0.01349266991019249, 'timing/model': 1.4821333661675453, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 0.26}) +[2026-06-15 15:00:05,788][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.24319037795066833, 'train/grad_norm_pre_clip': 9.519513130187988, 'timing/data': 0.012987046502530575, 'timing/model': 1.4498060196638107, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 0.26}) +[2026-06-15 15:00:11,660][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.05282307043671608, 'train/grad_norm_pre_clip': 2.5004172325134277, 'timing/data': 0.011002938263118267, 'timing/model': 1.4581788592040539, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 0.26}) +[2026-06-15 15:00:17,605][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.19042836129665375, 'train/grad_norm_pre_clip': 7.028050899505615, 'timing/data': 0.010639457032084465, 'timing/model': 1.696152277290821, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 0.26}) +[2026-06-15 15:00:23,275][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.3060135543346405, 'train/grad_norm_pre_clip': 18.48262596130371, 'timing/data': 0.012439575046300888, 'timing/model': 1.4656090391799808, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 0.26}) +[2026-06-15 15:00:29,174][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.21476362645626068, 'train/grad_norm_pre_clip': 8.834866523742676, 'timing/data': 0.01091022789478302, 'timing/model': 1.4717667484655976, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 0.26}) +[2026-06-15 15:00:35,173][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.2374543398618698, 'train/grad_norm_pre_clip': 8.122613906860352, 'timing/data': 0.011464649811387062, 'timing/model': 1.4876787830144167, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 0.26}) +[2026-06-15 15:00:41,070][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.2254929393529892, 'train/grad_norm_pre_clip': 3.44372820854187, 'timing/data': 0.011164119467139244, 'timing/model': 1.4687408041208982, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 0.26}) +[2026-06-15 15:00:47,067][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.12118842452764511, 'train/grad_norm_pre_clip': 4.285349369049072, 'timing/data': 0.01048909593373537, 'timing/model': 1.5189011879265308, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 0.26}) +[2026-06-15 15:00:53,004][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.17366181313991547, 'train/grad_norm_pre_clip': 9.104562759399414, 'timing/data': 0.010931069031357765, 'timing/model': 1.7014238247647882, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 0.26}) +[2026-06-15 15:00:58,808][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.27139177918434143, 'train/grad_norm_pre_clip': 9.220256805419922, 'timing/data': 0.014944124966859818, 'timing/model': 1.4767783097922802, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 0.26}) +[2026-06-15 15:01:04,761][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.23346218466758728, 'train/grad_norm_pre_clip': 7.180758476257324, 'timing/data': 0.010700667276978493, 'timing/model': 1.4441196490079165, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 0.26}) +[2026-06-15 15:01:10,766][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.26274237036705017, 'train/grad_norm_pre_clip': 6.50252628326416, 'timing/data': 0.010719758458435535, 'timing/model': 1.463198815472424, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 0.26}) +[2026-06-15 15:01:16,769][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.40053728222846985, 'train/grad_norm_pre_clip': 12.73709487915039, 'timing/data': 0.01088628824800253, 'timing/model': 1.4681287314742804, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 0.27}) +[2026-06-15 15:01:22,772][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.39075762033462524, 'train/grad_norm_pre_clip': 14.282792091369629, 'timing/data': 0.010874168016016483, 'timing/model': 1.5213833702728152, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 0.27}) +[2026-06-15 15:01:28,809][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.14078538119792938, 'train/grad_norm_pre_clip': 4.753567218780518, 'timing/data': 0.010491716675460339, 'timing/model': 1.7131273811683059, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 0.27}) +[2026-06-15 15:01:34,681][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.14584119617938995, 'train/grad_norm_pre_clip': 5.06341028213501, 'timing/data': 0.010615977458655834, 'timing/model': 1.5234590852633119, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 0.27}) +[2026-06-15 15:01:40,663][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.12939457595348358, 'train/grad_norm_pre_clip': 2.901179552078247, 'timing/data': 0.011513731442391872, 'timing/model': 1.460028987377882, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 0.27}) +[2026-06-15 15:01:46,690][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.4435591399669647, 'train/grad_norm_pre_clip': 9.051140785217285, 'timing/data': 0.00944737158715725, 'timing/model': 1.4809812437742949, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 0.27}) +[2026-06-15 15:01:52,621][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.23244450986385345, 'train/grad_norm_pre_clip': 5.706818580627441, 'timing/data': 0.010637306608259678, 'timing/model': 1.5088195307180285, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 0.27}) +[2026-06-15 15:01:58,663][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.22362832725048065, 'train/grad_norm_pre_clip': 7.640667915344238, 'timing/data': 0.011040029115974903, 'timing/model': 1.5165379671379924, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 0.27}) +[2026-06-15 15:02:04,632][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.268008291721344, 'train/grad_norm_pre_clip': 6.111466884613037, 'timing/data': 0.010728647001087666, 'timing/model': 1.7304252088069916, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 0.27}) +[2026-06-15 15:02:10,272][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.15641501545906067, 'train/grad_norm_pre_clip': 21.071805953979492, 'timing/data': 0.00999024510383606, 'timing/model': 1.472515108063817, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 0.27}) +[2026-06-15 15:02:16,173][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.21435880661010742, 'train/grad_norm_pre_clip': 6.2929182052612305, 'timing/data': 0.010504237376153469, 'timing/model': 1.470470062457025, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.27}) +[2026-06-15 15:02:22,065][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.2589189112186432, 'train/grad_norm_pre_clip': 3.8205931186676025, 'timing/data': 0.010450877249240875, 'timing/model': 1.4643921684473753, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 0.27}) +[2026-06-15 15:02:27,968][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.1340758353471756, 'train/grad_norm_pre_clip': 3.42330002784729, 'timing/data': 0.011385519988834858, 'timing/model': 1.4529555011540651, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 0.27}) +[2026-06-15 15:02:33,959][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.20353490114212036, 'train/grad_norm_pre_clip': 5.8965935707092285, 'timing/data': 0.010853728279471397, 'timing/model': 1.5167767377570271, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 0.27}) +[2026-06-15 15:02:39,896][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.28765618801116943, 'train/grad_norm_pre_clip': 9.683737754821777, 'timing/data': 0.011230229400098324, 'timing/model': 1.6917310869321227, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 0.27}) +[2026-06-15 15:02:45,736][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.22535361349582672, 'train/grad_norm_pre_clip': 13.258687019348145, 'timing/data': 0.011972853913903236, 'timing/model': 1.633523691445589, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 0.27}) +[2026-06-15 15:02:51,467][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.4616495668888092, 'train/grad_norm_pre_clip': 20.600566864013672, 'timing/data': 0.010712227784097195, 'timing/model': 1.4675410762429237, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 0.27}) +[2026-06-15 15:02:57,371][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.3139179050922394, 'train/grad_norm_pre_clip': 10.395247459411621, 'timing/data': 0.010820498690009117, 'timing/model': 1.4512622971087694, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 0.27}) +[2026-06-15 15:03:03,276][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.17962507903575897, 'train/grad_norm_pre_clip': 4.319675445556641, 'timing/data': 0.010650187730789185, 'timing/model': 1.4768758229911327, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 0.27}) +[2026-06-15 15:03:09,159][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.3028426766395569, 'train/grad_norm_pre_clip': 5.5703349113464355, 'timing/data': 0.01200187299400568, 'timing/model': 1.4356815619394183, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 0.27}) +[2026-06-15 15:13:54,671][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.28092160820961, 'train/grad_norm_pre_clip': 7.216633319854736, 'eval/loss': 0.17246569693088531, 'eval/flap_precision': 0.424390243902439, 'eval/flap_recall': 0.9490909090909091, 'eval/flap_f1': 0.5865168539325842, 'eval/flap_rate_model': 0.17083333333333334, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.5865168539325842, 'timing/data': 0.010675757192075253, 'timing/model': 1.6919449772685766, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 0.27}) +[2026-06-15 15:13:54,672][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_600_state +[2026-06-15 15:13:54,678][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 15:14:02,630][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_600_state/model.safetensors +[2026-06-15 15:14:13,608][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_600_state/optimizer.bin +[2026-06-15 15:14:13,610][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_600_state/sampler.bin +[2026-06-15 15:14:13,610][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_600_state/sampler_1.bin +[2026-06-15 15:14:13,612][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_600_state/random_states_0.pkl +[2026-06-15 15:14:13,764][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600 +[2026-06-15 15:14:13,765][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 15:14:13,780][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 15:14:19,735][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.23077607154846191, 'train/grad_norm_pre_clip': 7.328353404998779, 'timing/data': 0.013040159828960896, 'timing/model': 1.4538726611062884, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 0.27}) +[2026-06-15 15:14:25,705][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.28968167304992676, 'train/grad_norm_pre_clip': 9.025362968444824, 'timing/data': 0.011052261106669903, 'timing/model': 1.6846895879134536, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 0.28}) +[2026-06-15 15:14:31,449][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.2288804054260254, 'train/grad_norm_pre_clip': 2.7898991107940674, 'timing/data': 0.011015960946679115, 'timing/model': 1.5090678669512272, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 0.28}) +[2026-06-15 15:14:37,522][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.2666899859905243, 'train/grad_norm_pre_clip': 6.08272123336792, 'timing/data': 0.012696867808699608, 'timing/model': 1.5388567131012678, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 0.28}) +[2026-06-15 15:14:43,257][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.15683522820472717, 'train/grad_norm_pre_clip': 3.397352695465088, 'timing/data': 0.010699158534407616, 'timing/model': 1.4675487941130996, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 0.28}) +[2026-06-15 15:14:49,251][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.30138322710990906, 'train/grad_norm_pre_clip': 6.263718605041504, 'timing/data': 0.010938909836113453, 'timing/model': 1.4640387874096632, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 0.28}) +[2026-06-15 15:14:55,160][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.525327742099762, 'train/grad_norm_pre_clip': 15.152791023254395, 'timing/data': 0.010171117261052132, 'timing/model': 1.4311060449108481, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 0.28}) +[2026-06-15 15:15:01,303][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.12057780474424362, 'train/grad_norm_pre_clip': 2.0371053218841553, 'timing/data': 0.010880690068006516, 'timing/model': 1.7092981496825814, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 0.28}) +[2026-06-15 15:15:07,293][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.09180931001901627, 'train/grad_norm_pre_clip': 3.1009044647216797, 'timing/data': 0.010446119122207165, 'timing/model': 1.7977428371086717, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 0.28}) +[2026-06-15 15:15:12,961][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.12364546954631805, 'train/grad_norm_pre_clip': 5.128317356109619, 'timing/data': 0.011657013557851315, 'timing/model': 1.4708581985905766, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 0.28}) +[2026-06-15 15:15:18,862][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.1404879242181778, 'train/grad_norm_pre_clip': 5.270563125610352, 'timing/data': 0.010738460347056389, 'timing/model': 1.4682790068909526, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 0.28}) +[2026-06-15 15:15:24,766][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.1300903856754303, 'train/grad_norm_pre_clip': 3.736804723739624, 'timing/data': 0.011077901348471642, 'timing/model': 1.4678657744079828, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 0.28}) +[2026-06-15 15:15:30,660][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.23247478902339935, 'train/grad_norm_pre_clip': 12.222665786743164, 'timing/data': 0.010905850678682327, 'timing/model': 1.430766892619431, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 0.28}) +[2026-06-15 15:15:36,587][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.06888832896947861, 'train/grad_norm_pre_clip': 7.253525733947754, 'timing/data': 0.010754438117146492, 'timing/model': 1.6842274842783809, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 0.28}) +[2026-06-15 15:15:42,268][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.15042433142662048, 'train/grad_norm_pre_clip': 5.513400077819824, 'timing/data': 0.00997562613338232, 'timing/model': 1.4612741144374013, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 0.28}) +[2026-06-15 15:15:48,163][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.3909781575202942, 'train/grad_norm_pre_clip': 17.19452476501465, 'timing/data': 0.012197187170386314, 'timing/model': 1.4615496853366494, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 0.28}) +[2026-06-15 15:15:54,476][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.42949414253234863, 'train/grad_norm_pre_clip': 13.209366798400879, 'timing/data': 0.014995070174336433, 'timing/model': 1.4696686519309878, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 0.28}) +[2026-06-15 15:16:00,458][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.32420384883880615, 'train/grad_norm_pre_clip': 16.29438018798828, 'timing/data': 0.01078146044164896, 'timing/model': 1.508828092366457, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 0.28}) +[2026-06-15 15:16:06,354][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.2936363220214844, 'train/grad_norm_pre_clip': 12.038359642028809, 'timing/data': 0.010171137750148773, 'timing/model': 1.4764169743284583, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 0.28}) +[2026-06-15 15:16:12,265][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.4056203365325928, 'train/grad_norm_pre_clip': 12.442394256591797, 'timing/data': 0.010357687249779701, 'timing/model': 1.6767840785905719, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 0.28}) +[2026-06-15 15:16:17,959][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.08371037244796753, 'train/grad_norm_pre_clip': 6.339347839355469, 'timing/data': 0.010775898583233356, 'timing/model': 1.4663182171061635, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 0.28}) +[2026-06-15 15:16:23,860][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.16224215924739838, 'train/grad_norm_pre_clip': 6.485377311706543, 'timing/data': 0.010696528479456902, 'timing/model': 1.4606738714501262, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 0.28}) +[2026-06-15 15:16:29,774][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.32935577630996704, 'train/grad_norm_pre_clip': 14.406805992126465, 'timing/data': 0.011428232304751873, 'timing/model': 1.4659593356773257, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 0.28}) +[2026-06-15 15:16:35,778][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.4048146605491638, 'train/grad_norm_pre_clip': 6.837251663208008, 'timing/data': 0.011062380857765675, 'timing/model': 1.5018442999571562, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 0.29}) +[2026-06-15 15:16:41,676][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.23783466219902039, 'train/grad_norm_pre_clip': 11.182497024536133, 'timing/data': 0.010605189017951488, 'timing/model': 1.4824678413569927, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.29}) +[2026-06-15 15:16:47,621][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.1645434945821762, 'train/grad_norm_pre_clip': 4.062904357910156, 'timing/data': 0.011018641293048859, 'timing/model': 1.7205402599647641, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 0.29}) +[2026-06-15 15:16:53,269][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.28216713666915894, 'train/grad_norm_pre_clip': 7.959982395172119, 'timing/data': 0.012671908363699913, 'timing/model': 1.4599671065807343, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 0.29}) +[2026-06-15 15:16:59,276][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.30887073278427124, 'train/grad_norm_pre_clip': 5.732321262359619, 'timing/data': 0.01167504396289587, 'timing/model': 1.4673683708533645, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 0.29}) +[2026-06-15 15:17:05,168][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.22930940985679626, 'train/grad_norm_pre_clip': 3.305037021636963, 'timing/data': 0.011738344095647335, 'timing/model': 1.4637129846960306, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 0.29}) +[2026-06-15 15:17:11,180][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.24606658518314362, 'train/grad_norm_pre_clip': 4.683131694793701, 'timing/data': 0.012285136617720127, 'timing/model': 1.5353843942284584, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 0.29}) +[2026-06-15 15:17:17,089][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.20876774191856384, 'train/grad_norm_pre_clip': 8.932841300964355, 'timing/data': 0.01104063168168068, 'timing/model': 1.4560013683512807, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 0.29}) +[2026-06-15 15:17:23,003][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.1879292130470276, 'train/grad_norm_pre_clip': 3.4297659397125244, 'timing/data': 0.011820094659924507, 'timing/model': 1.697340590879321, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 0.29}) +[2026-06-15 15:17:28,661][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.22290118038654327, 'train/grad_norm_pre_clip': 2.5788486003875732, 'timing/data': 0.011081361211836338, 'timing/model': 1.4591386625543237, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 0.29}) +[2026-06-15 15:17:34,666][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.22533540427684784, 'train/grad_norm_pre_clip': 5.579672336578369, 'timing/data': 0.0114365229383111, 'timing/model': 1.4596487944945693, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 0.29}) +[2026-06-15 15:17:40,668][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.27046841382980347, 'train/grad_norm_pre_clip': 5.957635879516602, 'timing/data': 0.011790134012699127, 'timing/model': 1.4664962068200111, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 0.29}) +[2026-06-15 15:17:46,613][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.317272424697876, 'train/grad_norm_pre_clip': 5.189325332641602, 'timing/data': 0.01094715017825365, 'timing/model': 1.4786423025652766, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 0.29}) +[2026-06-15 15:17:52,473][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.26833051443099976, 'train/grad_norm_pre_clip': 3.777547836303711, 'timing/data': 0.01214129664003849, 'timing/model': 1.444961597211659, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 0.29}) +[2026-06-15 15:17:58,396][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.28559911251068115, 'train/grad_norm_pre_clip': 4.7399773597717285, 'timing/data': 0.011008759960532188, 'timing/model': 1.7037780303508043, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 0.29}) +[2026-06-15 15:18:04,065][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.15839357674121857, 'train/grad_norm_pre_clip': 4.031253814697266, 'timing/data': 0.011542483232915401, 'timing/model': 1.4579306561499834, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.29}) +[2026-06-15 15:18:10,072][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.3315492272377014, 'train/grad_norm_pre_clip': 5.332639694213867, 'timing/data': 0.011311481706798077, 'timing/model': 1.4769790237769485, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 0.29}) +[2026-06-15 15:18:15,961][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.17708472907543182, 'train/grad_norm_pre_clip': 4.063384056091309, 'timing/data': 0.011259072460234165, 'timing/model': 1.4531495543196797, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 0.29}) +[2026-06-15 15:18:21,865][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.21345286071300507, 'train/grad_norm_pre_clip': 9.976506233215332, 'timing/data': 0.010999899357557297, 'timing/model': 1.463874683715403, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 0.29}) +[2026-06-15 15:18:27,783][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.2850235104560852, 'train/grad_norm_pre_clip': 4.622716426849365, 'timing/data': 0.011209662072360516, 'timing/model': 1.4555404549464583, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 0.29}) +[2026-06-15 15:18:33,810][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.22813297808170319, 'train/grad_norm_pre_clip': 5.659104824066162, 'timing/data': 0.011232452467083931, 'timing/model': 1.7064504828304052, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 0.29}) +[2026-06-15 15:18:39,471][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.07820308208465576, 'train/grad_norm_pre_clip': 2.1979897022247314, 'timing/data': 0.011684573255479336, 'timing/model': 1.4656642219051719, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 0.29}) +[2026-06-15 15:18:45,469][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.26078787446022034, 'train/grad_norm_pre_clip': 7.005587577819824, 'timing/data': 0.011136181652545929, 'timing/model': 1.4603324262425303, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 0.3}) +[2026-06-15 15:18:51,368][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.2475784420967102, 'train/grad_norm_pre_clip': 9.791250228881836, 'timing/data': 0.011951474472880363, 'timing/model': 1.4603659184649587, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 0.3}) +[2026-06-15 15:18:57,266][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.5769951939582825, 'train/grad_norm_pre_clip': 9.781736373901367, 'timing/data': 0.010936180129647255, 'timing/model': 1.448070720769465, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 0.3}) +[2026-06-15 15:19:03,276][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.2117532640695572, 'train/grad_norm_pre_clip': 7.668102264404297, 'timing/data': 0.012717639096081257, 'timing/model': 1.5220532501116395, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 0.3}) +[2026-06-15 15:19:09,318][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.20203164219856262, 'train/grad_norm_pre_clip': 9.155784606933594, 'timing/data': 0.011830334551632404, 'timing/model': 1.7251507984474301, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 0.3}) +[2026-06-15 15:19:14,969][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.39380136132240295, 'train/grad_norm_pre_clip': 9.590747833251953, 'timing/data': 0.010816370137035847, 'timing/model': 1.4721553912386298, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 0.3}) +[2026-06-15 15:19:20,972][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.17560768127441406, 'train/grad_norm_pre_clip': 6.178226947784424, 'timing/data': 0.01058315858244896, 'timing/model': 1.4812525026500225, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 0.3}) +[2026-06-15 15:19:27,078][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.27619805932044983, 'train/grad_norm_pre_clip': 9.517547607421875, 'timing/data': 0.01074076909571886, 'timing/model': 1.4808295704424381, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 0.3}) +[2026-06-15 15:19:33,062][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.050271399319171906, 'train/grad_norm_pre_clip': 2.5049545764923096, 'timing/data': 0.013884943909943104, 'timing/model': 1.4509893339127302, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 0.3}) +[2026-06-15 15:19:39,072][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.31470760703086853, 'train/grad_norm_pre_clip': 8.446980476379395, 'timing/data': 0.011454891413450241, 'timing/model': 1.5162211945280433, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 0.3}) +[2026-06-15 15:19:45,019][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.42795291543006897, 'train/grad_norm_pre_clip': 9.451898574829102, 'timing/data': 0.01142854243516922, 'timing/model': 1.7212720299139619, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 0.3}) +[2026-06-15 15:19:50,873][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.6024431586265564, 'train/grad_norm_pre_clip': 15.10612678527832, 'timing/data': 0.01059723924845457, 'timing/model': 1.4849468991160393, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 0.3}) +[2026-06-15 15:19:56,874][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.11995677649974823, 'train/grad_norm_pre_clip': 2.952810525894165, 'timing/data': 0.011021231301128864, 'timing/model': 1.467555339448154, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 0.3}) +[2026-06-15 15:20:02,979][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.25009551644325256, 'train/grad_norm_pre_clip': 8.296723365783691, 'timing/data': 0.01602983381599188, 'timing/model': 1.5423404648900032, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 0.3}) +[2026-06-15 15:20:09,174][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.3424236476421356, 'train/grad_norm_pre_clip': 9.777957916259766, 'timing/data': 0.012443347834050655, 'timing/model': 1.4886733163148165, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.3}) +[2026-06-15 15:20:15,275][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.3259451389312744, 'train/grad_norm_pre_clip': 7.9639716148376465, 'timing/data': 0.02353792916983366, 'timing/model': 1.488643316552043, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 0.3}) +[2026-06-15 15:20:21,297][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.16297957301139832, 'train/grad_norm_pre_clip': 2.879676103591919, 'timing/data': 0.010990691371262074, 'timing/model': 1.7876018648967147, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 0.3}) +[2026-06-15 15:20:27,580][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.17229962348937988, 'train/grad_norm_pre_clip': 3.0582070350646973, 'timing/data': 0.011286421678960323, 'timing/model': 1.7730705086141825, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 0.3}) +[2026-06-15 15:20:33,266][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.23189882934093475, 'train/grad_norm_pre_clip': 4.255590915679932, 'timing/data': 0.011273532174527645, 'timing/model': 1.4622628251090646, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 0.3}) +[2026-06-15 15:20:39,260][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.3132934868335724, 'train/grad_norm_pre_clip': 5.371455669403076, 'timing/data': 0.011374631896615028, 'timing/model': 1.4677879204973578, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 0.3}) +[2026-06-15 15:20:45,157][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.34555914998054504, 'train/grad_norm_pre_clip': 5.185254096984863, 'timing/data': 0.009890146553516388, 'timing/model': 1.46375816129148, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 0.3}) +[2026-06-15 15:20:51,168][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.16673357784748077, 'train/grad_norm_pre_clip': 3.5295965671539307, 'timing/data': 0.011617753654718399, 'timing/model': 1.538569375872612, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 0.31}) +[2026-06-15 15:20:57,111][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.13927575945854187, 'train/grad_norm_pre_clip': 4.079502582550049, 'timing/data': 0.010835490189492702, 'timing/model': 1.7190485494211316, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 0.31}) +[2026-06-15 15:21:02,996][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.3217238187789917, 'train/grad_norm_pre_clip': 8.98689079284668, 'timing/data': 0.012545578181743622, 'timing/model': 1.702853774651885, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 0.31}) +[2026-06-15 15:21:08,659][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.2125924825668335, 'train/grad_norm_pre_clip': 7.145174503326416, 'timing/data': 0.011503612622618675, 'timing/model': 1.4628267772495747, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 0.31}) +[2026-06-15 15:21:14,558][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.15626966953277588, 'train/grad_norm_pre_clip': 11.72921371459961, 'timing/data': 0.011507882736623287, 'timing/model': 1.4552533626556396, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 0.31}) +[2026-06-15 15:21:20,472][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.3524889051914215, 'train/grad_norm_pre_clip': 6.580539226531982, 'timing/data': 0.01095495093613863, 'timing/model': 1.4392499784007668, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 0.31}) +[2026-06-15 15:21:26,457][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.23284733295440674, 'train/grad_norm_pre_clip': 8.521196365356445, 'timing/data': 0.011102600954473019, 'timing/model': 1.5197737887501717, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 0.31}) +[2026-06-15 15:21:32,395][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.45622140169143677, 'train/grad_norm_pre_clip': 12.743929862976074, 'timing/data': 0.011254542507231236, 'timing/model': 1.6959627019241452, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 0.31}) +[2026-06-15 15:21:38,061][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.30221283435821533, 'train/grad_norm_pre_clip': 7.497405052185059, 'timing/data': 0.011789054609835148, 'timing/model': 1.4647722160443664, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 0.31}) +[2026-06-15 15:21:44,058][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.21506991982460022, 'train/grad_norm_pre_clip': 5.886647701263428, 'timing/data': 0.010530917905271053, 'timing/model': 1.4521827287971973, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 0.31}) +[2026-06-15 15:21:49,960][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.3126733601093292, 'train/grad_norm_pre_clip': 8.41659927368164, 'timing/data': 0.011408013291656971, 'timing/model': 1.459696882404387, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 0.31}) +[2026-06-15 15:21:55,864][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.2745215594768524, 'train/grad_norm_pre_clip': 6.317022323608398, 'timing/data': 0.010888011194765568, 'timing/model': 1.433884852565825, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 0.31}) +[2026-06-15 15:22:01,767][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.5383442640304565, 'train/grad_norm_pre_clip': 14.795884132385254, 'timing/data': 0.010822339914739132, 'timing/model': 1.503645034506917, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 0.31}) +[2026-06-15 15:22:07,660][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.21604743599891663, 'train/grad_norm_pre_clip': 4.913975715637207, 'timing/data': 0.010999320074915886, 'timing/model': 1.6627974985167384, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 0.31}) +[2026-06-15 15:22:13,361][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.2559551000595093, 'train/grad_norm_pre_clip': 7.456914901733398, 'timing/data': 0.011524283327162266, 'timing/model': 1.462344504892826, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 0.31}) +[2026-06-15 15:22:19,365][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.21257628500461578, 'train/grad_norm_pre_clip': 8.526220321655273, 'timing/data': 0.01194662507623434, 'timing/model': 1.4626942444592714, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 0.31}) +[2026-06-15 15:22:25,263][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.2687370777130127, 'train/grad_norm_pre_clip': 3.732785701751709, 'timing/data': 0.011839334852993488, 'timing/model': 1.460423395037651, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 0.31}) +[2026-06-15 15:22:31,166][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.30777794122695923, 'train/grad_norm_pre_clip': 5.951896667480469, 'timing/data': 0.010857409797608852, 'timing/model': 1.4358965819701552, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 0.31}) +[2026-06-15 15:22:37,152][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.2963413596153259, 'train/grad_norm_pre_clip': 5.070418357849121, 'timing/data': 0.011815764009952545, 'timing/model': 1.5013654446229339, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 0.31}) +[2026-06-15 15:22:43,062][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.43625786900520325, 'train/grad_norm_pre_clip': 11.19042682647705, 'timing/data': 0.011558754369616508, 'timing/model': 1.661915053613484, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 0.31}) +[2026-06-15 15:22:48,759][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.2649671137332916, 'train/grad_norm_pre_clip': 3.1119484901428223, 'timing/data': 0.011028931476175785, 'timing/model': 1.4633761895820498, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 0.31}) +[2026-06-15 15:22:54,658][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.13718320429325104, 'train/grad_norm_pre_clip': 2.2736449241638184, 'timing/data': 0.01098933070898056, 'timing/model': 1.4589052479714155, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.31}) +[2026-06-15 15:23:00,558][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.22508028149604797, 'train/grad_norm_pre_clip': 2.3770127296447754, 'timing/data': 0.012192976661026478, 'timing/model': 1.4522409467026591, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 0.32}) +[2026-06-15 15:23:06,460][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.1764410436153412, 'train/grad_norm_pre_clip': 3.4017510414123535, 'timing/data': 0.01109183020889759, 'timing/model': 1.4349967679008842, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 0.32}) +[2026-06-15 15:23:12,392][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.4014299213886261, 'train/grad_norm_pre_clip': 16.359142303466797, 'timing/data': 0.01104006078094244, 'timing/model': 1.6919410629197955, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 0.32}) +[2026-06-15 15:23:18,063][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.25026562809944153, 'train/grad_norm_pre_clip': 5.04861307144165, 'timing/data': 0.011204500682651997, 'timing/model': 1.4686338631436229, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 0.32}) +[2026-06-15 15:23:23,960][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.15726737678050995, 'train/grad_norm_pre_clip': 2.750032424926758, 'timing/data': 0.01060382928699255, 'timing/model': 1.4661982217803597, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 0.32}) +[2026-06-15 15:23:30,051][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.13117007911205292, 'train/grad_norm_pre_clip': 3.3092751502990723, 'timing/data': 0.010497058741748333, 'timing/model': 1.4710457921028137, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 0.32}) +[2026-06-15 15:23:35,967][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.20380054414272308, 'train/grad_norm_pre_clip': 3.2928426265716553, 'timing/data': 0.011539583094418049, 'timing/model': 1.464663245715201, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.32}) +[2026-06-15 15:23:41,943][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.2252955436706543, 'train/grad_norm_pre_clip': 4.165738582611084, 'timing/data': 0.010793330147862434, 'timing/model': 1.540317422710359, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 0.32}) +[2026-06-15 15:23:47,905][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.14169536530971527, 'train/grad_norm_pre_clip': 5.3757548332214355, 'timing/data': 0.01096069160848856, 'timing/model': 1.7000024188309908, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 0.32}) +[2026-06-15 15:23:53,660][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.35411059856414795, 'train/grad_norm_pre_clip': 6.058222770690918, 'timing/data': 0.01123853214085102, 'timing/model': 1.451306332834065, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 0.32}) +[2026-06-15 15:23:59,664][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.04554566740989685, 'train/grad_norm_pre_clip': 2.0244736671447754, 'timing/data': 0.011113901622593403, 'timing/model': 1.4660293897613883, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 0.32}) +[2026-06-15 15:34:26,355][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.061865516006946564, 'train/grad_norm_pre_clip': 3.3827149868011475, 'eval/loss': 0.11758750677108765, 'eval/flap_precision': 0.7098039215686275, 'eval/flap_recall': 0.6581818181818182, 'eval/flap_f1': 0.6830188679245283, 'eval/flap_rate_model': 0.07083333333333333, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6830188679245283, 'timing/data': 0.010800819844007492, 'timing/model': 1.4654788877815008, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 0.32}) +[2026-06-15 15:34:26,356][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_700_state +[2026-06-15 15:34:26,362][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 15:34:34,643][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_700_state/model.safetensors +[2026-06-15 15:34:45,753][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_700_state/optimizer.bin +[2026-06-15 15:34:45,754][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_700_state/sampler.bin +[2026-06-15 15:34:45,754][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_700_state/sampler_1.bin +[2026-06-15 15:34:45,755][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_700_state/random_states_0.pkl +[2026-06-15 15:34:45,920][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700 +[2026-06-15 15:34:45,921][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 15:34:45,935][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 15:34:51,933][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.3193652629852295, 'train/grad_norm_pre_clip': 8.43557071685791, 'timing/data': 0.01114790141582489, 'timing/model': 1.4672820726409554, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 0.32}) +[2026-06-15 15:34:57,837][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.3137805163860321, 'train/grad_norm_pre_clip': 14.064593315124512, 'timing/data': 0.011696534231305122, 'timing/model': 1.464857341721654, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 0.32}) +[2026-06-15 15:35:03,761][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.2768649160861969, 'train/grad_norm_pre_clip': 6.968221664428711, 'timing/data': 0.012111455202102661, 'timing/model': 1.449248949997127, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 0.32}) +[2026-06-15 15:35:09,738][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.19992130994796753, 'train/grad_norm_pre_clip': 12.08883285522461, 'timing/data': 0.013513782061636448, 'timing/model': 1.5279608722776175, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 0.32}) +[2026-06-15 15:35:15,676][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.3298318088054657, 'train/grad_norm_pre_clip': 10.708720207214355, 'timing/data': 0.011387242935597897, 'timing/model': 1.7000118857249618, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 0.32}) +[2026-06-15 15:35:21,342][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.13508561253547668, 'train/grad_norm_pre_clip': 5.297302722930908, 'timing/data': 0.01111978106200695, 'timing/model': 1.4655533162876964, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 0.32}) +[2026-06-15 15:35:27,243][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.21435412764549255, 'train/grad_norm_pre_clip': 9.787339210510254, 'timing/data': 0.011865205131471157, 'timing/model': 1.4607089133933187, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 0.32}) +[2026-06-15 15:35:33,146][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.14973928034305573, 'train/grad_norm_pre_clip': 5.083327293395996, 'timing/data': 0.011855854652822018, 'timing/model': 1.474237635731697, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 0.32}) +[2026-06-15 15:35:39,042][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.09563207626342773, 'train/grad_norm_pre_clip': 2.96508526802063, 'timing/data': 0.011182061396539211, 'timing/model': 1.4199849953874946, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 0.32}) +[2026-06-15 15:35:44,941][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.3518749475479126, 'train/grad_norm_pre_clip': 10.187811851501465, 'timing/data': 0.010916359722614288, 'timing/model': 1.4480649251490831, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 0.32}) +[2026-06-15 15:35:50,881][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.24771210551261902, 'train/grad_norm_pre_clip': 5.952303886413574, 'timing/data': 0.01093808002769947, 'timing/model': 1.698836700990796, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 0.33}) +[2026-06-15 15:35:56,546][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.24848458170890808, 'train/grad_norm_pre_clip': 5.93043851852417, 'timing/data': 0.011066170409321785, 'timing/model': 1.4597416492179036, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 0.33}) +[2026-06-15 15:36:02,447][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.1585012972354889, 'train/grad_norm_pre_clip': 4.288484573364258, 'timing/data': 0.011198371648788452, 'timing/model': 1.4601244712248445, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 0.33}) +[2026-06-15 15:36:08,452][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.10116900503635406, 'train/grad_norm_pre_clip': 5.07506799697876, 'timing/data': 0.011073460802435875, 'timing/model': 1.4749476602301002, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 0.33}) +[2026-06-15 15:36:14,349][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.2069302350282669, 'train/grad_norm_pre_clip': 7.797299861907959, 'timing/data': 0.01137362327426672, 'timing/model': 1.4616461284458637, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 0.33}) +[2026-06-15 15:36:20,250][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.212941974401474, 'train/grad_norm_pre_clip': 4.13462495803833, 'timing/data': 0.01232114713639021, 'timing/model': 1.4262399142608047, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 0.33}) +[2026-06-15 15:36:26,180][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.2355557233095169, 'train/grad_norm_pre_clip': 8.107956886291504, 'timing/data': 0.011602864600718021, 'timing/model': 1.6913757156580687, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 0.33}) +[2026-06-15 15:36:31,851][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.2406027466058731, 'train/grad_norm_pre_clip': 4.333714962005615, 'timing/data': 0.011024431325495243, 'timing/model': 1.4581040516495705, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 0.33}) +[2026-06-15 15:36:37,749][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.3282114863395691, 'train/grad_norm_pre_clip': 5.872886657714844, 'timing/data': 0.011405503377318382, 'timing/model': 1.4517918024212122, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 0.33}) +[2026-06-15 15:36:43,652][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.24089361727237701, 'train/grad_norm_pre_clip': 7.626946926116943, 'timing/data': 0.01179705373942852, 'timing/model': 1.4656172255054116, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 0.33}) +[2026-06-15 15:36:49,551][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.16166678071022034, 'train/grad_norm_pre_clip': 3.736419916152954, 'timing/data': 0.012115745805203915, 'timing/model': 1.4613661654293537, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 0.33}) +[2026-06-15 15:36:55,468][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.09121590852737427, 'train/grad_norm_pre_clip': 1.9060817956924438, 'timing/data': 0.010295077227056026, 'timing/model': 1.4326801942661405, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 0.33}) +[2026-06-15 15:37:01,477][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.23012959957122803, 'train/grad_norm_pre_clip': 8.114852905273438, 'timing/data': 0.010411098599433899, 'timing/model': 1.701075941324234, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 0.33}) +[2026-06-15 15:37:07,148][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.26169392466545105, 'train/grad_norm_pre_clip': 8.177923202514648, 'timing/data': 0.011087451130151749, 'timing/model': 1.4622497400268912, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 0.33}) +[2026-06-15 15:37:13,050][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.39229264855384827, 'train/grad_norm_pre_clip': 13.665498733520508, 'timing/data': 0.011932024732232094, 'timing/model': 1.4554705889895558, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 0.33}) +[2026-06-15 15:37:18,950][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.4767155349254608, 'train/grad_norm_pre_clip': 12.237054824829102, 'timing/data': 0.010556538589298725, 'timing/model': 1.458318043500185, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 0.33}) +[2026-06-15 15:37:24,961][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.2688794732093811, 'train/grad_norm_pre_clip': 5.767100811004639, 'timing/data': 0.010190946981310844, 'timing/model': 1.4665938299149275, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 0.33}) +[2026-06-15 15:37:30,953][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.07017257064580917, 'train/grad_norm_pre_clip': 3.227231979370117, 'timing/data': 0.010008306242525578, 'timing/model': 1.4795146407559514, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 0.33}) +[2026-06-15 15:37:36,892][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.23434512317180634, 'train/grad_norm_pre_clip': 5.044449806213379, 'timing/data': 0.011605143547058105, 'timing/model': 1.6949162129312754, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 0.33}) +[2026-06-15 15:37:42,568][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.11508532613515854, 'train/grad_norm_pre_clip': 2.2706234455108643, 'timing/data': 0.011521662585437298, 'timing/model': 1.4708010405302048, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.33}) +[2026-06-15 15:37:48,457][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.13411124050617218, 'train/grad_norm_pre_clip': 3.0692553520202637, 'timing/data': 0.011129922233521938, 'timing/model': 1.4726155577227473, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 0.33}) +[2026-06-15 15:37:54,361][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.28076812624931335, 'train/grad_norm_pre_clip': 6.379684925079346, 'timing/data': 0.011060521006584167, 'timing/model': 1.469423983246088, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 0.33}) +[2026-06-15 15:38:00,365][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.23662106692790985, 'train/grad_norm_pre_clip': 5.311034202575684, 'timing/data': 0.011320473626255989, 'timing/model': 1.5031971773132682, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 0.34}) +[2026-06-15 15:38:06,338][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.1213340237736702, 'train/grad_norm_pre_clip': 3.5691614151000977, 'timing/data': 0.012163875624537468, 'timing/model': 1.5000114543363452, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 0.34}) +[2026-06-15 15:38:12,304][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.034462083131074905, 'train/grad_norm_pre_clip': 1.7722396850585938, 'timing/data': 0.011034620925784111, 'timing/model': 1.7058109231293201, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 0.34}) +[2026-06-15 15:38:18,132][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.12152203172445297, 'train/grad_norm_pre_clip': 5.616891860961914, 'timing/data': 0.011167062446475029, 'timing/model': 1.6366071542724967, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 0.34}) +[2026-06-15 15:38:23,863][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.156794935464859, 'train/grad_norm_pre_clip': 6.538778781890869, 'timing/data': 0.011613932438194752, 'timing/model': 1.4678891664370894, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.34}) +[2026-06-15 15:38:29,759][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.3116886019706726, 'train/grad_norm_pre_clip': 7.826395511627197, 'timing/data': 0.011865925043821335, 'timing/model': 1.4596184473484755, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 0.34}) +[2026-06-15 15:38:35,660][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.16528038680553436, 'train/grad_norm_pre_clip': 5.9430742263793945, 'timing/data': 0.0113893523812294, 'timing/model': 1.4587159743532538, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 0.34}) +[2026-06-15 15:38:41,652][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.18328416347503662, 'train/grad_norm_pre_clip': 6.3731865882873535, 'timing/data': 0.011617164127528667, 'timing/model': 1.5058099506422877, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 0.34}) +[2026-06-15 15:38:47,588][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.01444934494793415, 'train/grad_norm_pre_clip': 1.9271351099014282, 'timing/data': 0.0113936522975564, 'timing/model': 1.6906087137758732, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 0.34}) +[2026-06-15 15:38:53,499][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.5973241329193115, 'train/grad_norm_pre_clip': 15.116278648376465, 'timing/data': 0.010777609422802925, 'timing/model': 1.69740423373878, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 0.34}) +[2026-06-15 15:38:59,161][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.07640897482633591, 'train/grad_norm_pre_clip': 3.4695725440979004, 'timing/data': 0.011816485784947872, 'timing/model': 1.4615682074800134, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 0.34}) +[2026-06-15 15:39:05,063][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.19128961861133575, 'train/grad_norm_pre_clip': 12.892523765563965, 'timing/data': 0.01108959037810564, 'timing/model': 1.465001743286848, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 0.34}) +[2026-06-15 15:39:11,065][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.11071967333555222, 'train/grad_norm_pre_clip': 8.255165100097656, 'timing/data': 0.01164859440177679, 'timing/model': 1.4677552348002791, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 0.34}) +[2026-06-15 15:39:16,966][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.38028842210769653, 'train/grad_norm_pre_clip': 10.115971565246582, 'timing/data': 0.011148600839078426, 'timing/model': 1.438631341792643, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 0.34}) +[2026-06-15 15:39:22,891][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.2690158784389496, 'train/grad_norm_pre_clip': 7.797989368438721, 'timing/data': 0.010930880904197693, 'timing/model': 1.69258439168334, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 0.34}) +[2026-06-15 15:39:28,557][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.44083189964294434, 'train/grad_norm_pre_clip': 9.263671875, 'timing/data': 0.011242251843214035, 'timing/model': 1.4560395507141948, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 0.34}) +[2026-06-15 15:39:34,557][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.33238208293914795, 'train/grad_norm_pre_clip': 9.38932991027832, 'timing/data': 0.011257932521402836, 'timing/model': 1.4497934523969889, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 0.34}) +[2026-06-15 15:39:40,471][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.3877567946910858, 'train/grad_norm_pre_clip': 11.911600112915039, 'timing/data': 0.011611402966082096, 'timing/model': 1.4683545380830765, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 0.34}) +[2026-06-15 15:39:46,456][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.19158127903938293, 'train/grad_norm_pre_clip': 6.137713432312012, 'timing/data': 0.013260011561214924, 'timing/model': 1.5248463582247496, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 0.34}) +[2026-06-15 15:39:52,359][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.24690154194831848, 'train/grad_norm_pre_clip': 7.9831037521362305, 'timing/data': 0.011043300852179527, 'timing/model': 1.4953392930328846, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 0.34}) +[2026-06-15 15:39:58,267][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.21617667376995087, 'train/grad_norm_pre_clip': 7.150708198547363, 'timing/data': 0.01107753161340952, 'timing/model': 1.6733320029452443, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 0.34}) +[2026-06-15 15:40:03,970][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.27979642152786255, 'train/grad_norm_pre_clip': 4.751662731170654, 'timing/data': 0.012171396985650063, 'timing/model': 1.4733432605862617, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 0.34}) +[2026-06-15 15:40:09,866][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.20662634074687958, 'train/grad_norm_pre_clip': 4.970920562744141, 'timing/data': 0.010867880657315254, 'timing/model': 1.4716375432908535, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 0.35}) +[2026-06-15 15:40:15,765][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.09672913700342178, 'train/grad_norm_pre_clip': 1.6125508546829224, 'timing/data': 0.011512232944369316, 'timing/model': 1.46143969707191, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 0.35}) +[2026-06-15 15:40:21,668][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.302467942237854, 'train/grad_norm_pre_clip': 5.025381088256836, 'timing/data': 0.011791444383561611, 'timing/model': 1.4380130982026458, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 0.35}) +[2026-06-15 15:40:27,667][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.15883314609527588, 'train/grad_norm_pre_clip': 3.0502431392669678, 'timing/data': 0.011607742868363857, 'timing/model': 1.5170997735112906, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 0.35}) +[2026-06-15 15:40:33,567][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.10351100564002991, 'train/grad_norm_pre_clip': 3.4072470664978027, 'timing/data': 0.011519093066453934, 'timing/model': 1.6716587850823998, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 0.35}) +[2026-06-15 15:40:39,259][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.09407681971788406, 'train/grad_norm_pre_clip': 2.0681662559509277, 'timing/data': 0.011118141002953053, 'timing/model': 1.4625026704743505, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 0.35}) +[2026-06-15 15:40:45,162][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.1762882024049759, 'train/grad_norm_pre_clip': 3.3766417503356934, 'timing/data': 0.01150120235979557, 'timing/model': 1.4634328847751021, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 0.35}) +[2026-06-15 15:40:51,074][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.25784215331077576, 'train/grad_norm_pre_clip': 8.772200584411621, 'timing/data': 0.011375372298061848, 'timing/model': 1.4742514546960592, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 0.35}) +[2026-06-15 15:40:56,981][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.17263588309288025, 'train/grad_norm_pre_clip': 7.084171295166016, 'timing/data': 0.011609584093093872, 'timing/model': 1.4475530814379454, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 0.35}) +[2026-06-15 15:41:02,894][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.2526891827583313, 'train/grad_norm_pre_clip': 6.27900505065918, 'timing/data': 0.010503008030354977, 'timing/model': 1.6935942163690925, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 0.35}) +[2026-06-15 15:41:08,565][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.2722642719745636, 'train/grad_norm_pre_clip': 6.4851813316345215, 'timing/data': 0.011359773576259613, 'timing/model': 1.466581780463457, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.35}) +[2026-06-15 15:41:14,459][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.2374744564294815, 'train/grad_norm_pre_clip': 11.928577423095703, 'timing/data': 0.010998721234500408, 'timing/model': 1.4599071694537997, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 0.35}) +[2026-06-15 15:41:20,358][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.3084494471549988, 'train/grad_norm_pre_clip': 16.641843795776367, 'timing/data': 0.011148782446980476, 'timing/model': 1.4592045955359936, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 0.35}) +[2026-06-15 15:41:26,268][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.23338866233825684, 'train/grad_norm_pre_clip': 8.512205123901367, 'timing/data': 0.011269491165876389, 'timing/model': 1.4696307443082333, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 0.35}) +[2026-06-15 15:41:32,172][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.30337706208229065, 'train/grad_norm_pre_clip': 10.328837394714355, 'timing/data': 0.01118960790336132, 'timing/model': 1.471928296610713, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 0.35}) +[2026-06-15 15:41:38,191][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.12238509953022003, 'train/grad_norm_pre_clip': 6.115677356719971, 'timing/data': 0.010934442281723022, 'timing/model': 1.7012843927368522, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 0.35}) +[2026-06-15 15:41:43,851][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.1386084258556366, 'train/grad_norm_pre_clip': 5.51924991607666, 'timing/data': 0.010837838053703308, 'timing/model': 1.461838780902326, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 0.35}) +[2026-06-15 15:41:49,753][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.18272428214550018, 'train/grad_norm_pre_clip': 7.409938335418701, 'timing/data': 0.01038900762796402, 'timing/model': 1.4581006187945604, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 0.35}) +[2026-06-15 15:41:55,657][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.20195989310741425, 'train/grad_norm_pre_clip': 6.89578104019165, 'timing/data': 0.010865566320717335, 'timing/model': 1.4662264343351126, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 0.35}) +[2026-06-15 15:42:01,557][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.3331352770328522, 'train/grad_norm_pre_clip': 8.534452438354492, 'timing/data': 0.011042197234928608, 'timing/model': 1.4624258074909449, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 0.35}) +[2026-06-15 15:42:07,459][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.13761471211910248, 'train/grad_norm_pre_clip': 5.350606918334961, 'timing/data': 0.010740790516138077, 'timing/model': 1.4683059537783265, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 0.35}) +[2026-06-15 15:42:13,393][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.483122855424881, 'train/grad_norm_pre_clip': 13.40485668182373, 'timing/data': 0.011043485254049301, 'timing/model': 1.6972902053967118, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 0.35}) +[2026-06-15 15:42:19,052][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.15599112212657928, 'train/grad_norm_pre_clip': 5.228896617889404, 'timing/data': 0.010569235309958458, 'timing/model': 1.4566242638975382, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 0.36}) +[2026-06-15 15:42:24,957][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.2405945211648941, 'train/grad_norm_pre_clip': 4.736327648162842, 'timing/data': 0.010563752613961697, 'timing/model': 1.46687694452703, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 0.36}) +[2026-06-15 15:42:30,861][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.46605923771858215, 'train/grad_norm_pre_clip': 11.189868927001953, 'timing/data': 0.01164334174245596, 'timing/model': 1.4580170316621661, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 0.36}) +[2026-06-15 15:42:36,832][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.25531241297721863, 'train/grad_norm_pre_clip': 4.802357196807861, 'timing/data': 0.011338363401591778, 'timing/model': 1.497293476946652, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 0.36}) +[2026-06-15 15:42:42,760][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.18298059701919556, 'train/grad_norm_pre_clip': 2.7696704864501953, 'timing/data': 0.01110746804624796, 'timing/model': 1.4482732359319925, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 0.36}) +[2026-06-15 15:42:48,699][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.23180799186229706, 'train/grad_norm_pre_clip': 6.239015102386475, 'timing/data': 0.00969794299453497, 'timing/model': 1.7046434450894594, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 0.36}) +[2026-06-15 15:42:54,364][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.24223975837230682, 'train/grad_norm_pre_clip': 6.052064418792725, 'timing/data': 0.010942403227090836, 'timing/model': 1.4765299558639526, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 0.36}) +[2026-06-15 15:43:00,364][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.21244242787361145, 'train/grad_norm_pre_clip': 4.598164081573486, 'timing/data': 0.012377885170280933, 'timing/model': 1.471779546700418, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 0.36}) +[2026-06-15 15:43:06,266][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.13195697963237762, 'train/grad_norm_pre_clip': 3.4788479804992676, 'timing/data': 0.010884038172662258, 'timing/model': 1.4737965436652303, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 0.36}) +[2026-06-15 15:43:12,165][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.1474865823984146, 'train/grad_norm_pre_clip': 2.3072168827056885, 'timing/data': 0.011327785439789295, 'timing/model': 1.4665012313053012, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.36}) +[2026-06-15 15:43:18,159][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.21899321675300598, 'train/grad_norm_pre_clip': 4.2197113037109375, 'timing/data': 0.011034409515559673, 'timing/model': 1.5140922414138913, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 0.36}) +[2026-06-15 15:43:24,187][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.13739582896232605, 'train/grad_norm_pre_clip': 3.125178575515747, 'timing/data': 0.011285281740128994, 'timing/model': 1.677490252070129, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 0.36}) +[2026-06-15 15:43:29,852][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.2600199580192566, 'train/grad_norm_pre_clip': 8.190009117126465, 'timing/data': 0.010435417294502258, 'timing/model': 1.4608529331162572, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 0.36}) +[2026-06-15 15:43:35,760][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.17059841752052307, 'train/grad_norm_pre_clip': 5.734469413757324, 'timing/data': 0.010863052681088448, 'timing/model': 1.4672117279842496, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 0.36}) +[2026-06-15 15:43:41,659][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.09283174574375153, 'train/grad_norm_pre_clip': 3.516155242919922, 'timing/data': 0.011001123115420341, 'timing/model': 1.4622603263705969, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 0.36}) +[2026-06-15 15:43:47,558][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.10319027304649353, 'train/grad_norm_pre_clip': 6.935235500335693, 'timing/data': 0.011620482429862022, 'timing/model': 1.4639089098200202, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 0.36}) +[2026-06-15 15:43:53,542][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.2007192224264145, 'train/grad_norm_pre_clip': 9.651915550231934, 'timing/data': 0.011150993406772614, 'timing/model': 1.5044951094314456, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 0.36}) +[2026-06-15 15:43:59,511][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.3028061091899872, 'train/grad_norm_pre_clip': 9.241090774536133, 'timing/data': 0.011777502484619617, 'timing/model': 1.712304014712572, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 0.36}) +[2026-06-15 15:44:05,159][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.2538236379623413, 'train/grad_norm_pre_clip': 8.232168197631836, 'timing/data': 0.011253982782363892, 'timing/model': 1.4579243194311857, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 0.36}) +[2026-06-15 15:44:11,060][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.4732000231742859, 'train/grad_norm_pre_clip': 24.953508377075195, 'timing/data': 0.011683368124067783, 'timing/model': 1.4584472747519612, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 0.36}) +[2026-06-15 15:44:17,059][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.18974152207374573, 'train/grad_norm_pre_clip': 57.38485336303711, 'timing/data': 0.011484703980386257, 'timing/model': 1.454941688105464, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 0.36}) +[2026-06-15 15:44:22,962][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.26188376545906067, 'train/grad_norm_pre_clip': 15.066099166870117, 'timing/data': 0.012144353240728378, 'timing/model': 1.4662102572619915, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 0.36}) +[2026-06-15 15:44:29,353][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.09106741100549698, 'train/grad_norm_pre_clip': 9.492036819458008, 'timing/data': 0.011899738572537899, 'timing/model': 1.5146972686052322, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 0.37}) +[2026-06-15 15:55:01,155][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.43067219853401184, 'train/grad_norm_pre_clip': 10.369484901428223, 'eval/loss': 0.11362079530954361, 'eval/flap_precision': 0.6439169139465876, 'eval/flap_recall': 0.7890909090909091, 'eval/flap_f1': 0.7091503267973855, 'eval/flap_rate_model': 0.09361111111111112, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7091503267973855, 'timing/data': 0.011195286177098751, 'timing/model': 1.699688084423542, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.37}) +[2026-06-15 15:55:01,156][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_800_state +[2026-06-15 15:55:01,162][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 15:55:09,251][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_800_state/model.safetensors +[2026-06-15 15:55:28,057][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_800_state/optimizer.bin +[2026-06-15 15:55:28,058][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_800_state/sampler.bin +[2026-06-15 15:55:28,058][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_800_state/sampler_1.bin +[2026-06-15 15:55:28,059][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_800_state/random_states_0.pkl +[2026-06-15 15:55:28,216][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800 +[2026-06-15 15:55:28,216][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 15:55:28,231][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 15:55:34,209][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.07398579269647598, 'train/grad_norm_pre_clip': 3.369853973388672, 'timing/data': 0.010951884090900421, 'timing/model': 1.475280111655593, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 0.37}) +[2026-06-15 15:55:40,079][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.3561614453792572, 'train/grad_norm_pre_clip': 16.74109649658203, 'timing/data': 0.011331576853990555, 'timing/model': 1.703051958233118, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 0.37}) +[2026-06-15 15:55:45,844][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.10378330200910568, 'train/grad_norm_pre_clip': 5.18292236328125, 'timing/data': 0.012377066537737846, 'timing/model': 1.4722483363002539, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 0.37}) +[2026-06-15 15:55:51,922][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.3060276508331299, 'train/grad_norm_pre_clip': 12.318778038024902, 'timing/data': 0.011940300464630127, 'timing/model': 1.5496596833691, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 0.37}) +[2026-06-15 15:55:57,846][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.2707705497741699, 'train/grad_norm_pre_clip': 22.718002319335938, 'timing/data': 0.012122232466936111, 'timing/model': 1.461922668851912, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 0.37}) +[2026-06-15 15:56:03,761][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.11972279101610184, 'train/grad_norm_pre_clip': 3.691317081451416, 'timing/data': 0.0136037552729249, 'timing/model': 1.4801238533109426, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 0.37}) +[2026-06-15 15:56:09,637][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.2642821967601776, 'train/grad_norm_pre_clip': 8.705543518066406, 'timing/data': 0.011325265280902386, 'timing/model': 1.4157631443813443, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 0.37}) +[2026-06-15 15:56:15,596][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.2728751003742218, 'train/grad_norm_pre_clip': 8.472237586975098, 'timing/data': 0.011554497294127941, 'timing/model': 1.711396616883576, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 0.37}) +[2026-06-15 15:56:21,252][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.2181563377380371, 'train/grad_norm_pre_clip': 8.432092666625977, 'timing/data': 0.010898420587182045, 'timing/model': 1.4679223103448749, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 0.37}) +[2026-06-15 15:56:27,162][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.13338418304920197, 'train/grad_norm_pre_clip': 3.96087908744812, 'timing/data': 0.011165563017129898, 'timing/model': 1.481873668730259, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 0.37}) +[2026-06-15 15:56:33,061][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.2544876039028168, 'train/grad_norm_pre_clip': 6.462240219116211, 'timing/data': 0.010969370603561401, 'timing/model': 1.4706027992069721, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 0.37}) +[2026-06-15 15:56:38,964][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.17452600598335266, 'train/grad_norm_pre_clip': 4.60596227645874, 'timing/data': 0.011353953741490841, 'timing/model': 1.428212622180581, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 0.37}) +[2026-06-15 15:56:44,891][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.20006291568279266, 'train/grad_norm_pre_clip': 4.643608570098877, 'timing/data': 0.011538184247910976, 'timing/model': 1.704656671732664, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 0.37}) +[2026-06-15 15:56:50,553][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.16012707352638245, 'train/grad_norm_pre_clip': 3.360640525817871, 'timing/data': 0.011790506541728973, 'timing/model': 1.4605536106973886, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 0.37}) +[2026-06-15 15:56:56,451][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.42031553387641907, 'train/grad_norm_pre_clip': 9.718240737915039, 'timing/data': 0.010902239009737968, 'timing/model': 1.454700518399477, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 0.37}) +[2026-06-15 15:57:02,255][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.13332657516002655, 'train/grad_norm_pre_clip': 2.4365155696868896, 'timing/data': 0.011016760021448135, 'timing/model': 1.454482783563435, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 0.37}) +[2026-06-15 15:57:08,165][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.28728413581848145, 'train/grad_norm_pre_clip': 5.645453453063965, 'timing/data': 0.01287668477743864, 'timing/model': 1.4806257840245962, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 0.37}) +[2026-06-15 15:57:14,059][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.29881054162979126, 'train/grad_norm_pre_clip': 3.8259763717651367, 'timing/data': 0.011118470691144466, 'timing/model': 1.4294823426753283, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 0.37}) +[2026-06-15 15:57:19,994][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.25417259335517883, 'train/grad_norm_pre_clip': 3.538796901702881, 'timing/data': 0.011591955088078976, 'timing/model': 1.700158840045333, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 0.37}) +[2026-06-15 15:57:25,897][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.14648647606372833, 'train/grad_norm_pre_clip': 3.857769250869751, 'timing/data': 0.010529224760830402, 'timing/model': 1.7062186216935515, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 0.38}) +[2026-06-15 15:57:31,565][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.23402661085128784, 'train/grad_norm_pre_clip': 3.450127363204956, 'timing/data': 0.010638736188411713, 'timing/model': 1.4680313393473625, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 0.38}) +[2026-06-15 15:57:37,458][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.163502499461174, 'train/grad_norm_pre_clip': 4.2985734939575195, 'timing/data': 0.011280890554189682, 'timing/model': 1.4615010553970933, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 0.38}) +[2026-06-15 15:57:43,366][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.2432238906621933, 'train/grad_norm_pre_clip': 4.454728603363037, 'timing/data': 0.0111976508051157, 'timing/model': 1.4756335988640785, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 0.38}) +[2026-06-15 15:57:49,374][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.1053335890173912, 'train/grad_norm_pre_clip': 3.648120641708374, 'timing/data': 0.011096089147031307, 'timing/model': 1.5220235977321863, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 0.38}) +[2026-06-15 15:57:55,296][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.42211586236953735, 'train/grad_norm_pre_clip': 9.194162368774414, 'timing/data': 0.010976457968354225, 'timing/model': 1.7015140503644943, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 0.38}) +[2026-06-15 15:58:00,958][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.2898487150669098, 'train/grad_norm_pre_clip': 5.638704299926758, 'timing/data': 0.010433252900838852, 'timing/model': 1.474407386034727, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 0.38}) +[2026-06-15 15:58:06,856][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.25893434882164, 'train/grad_norm_pre_clip': 8.173974990844727, 'timing/data': 0.010808685794472694, 'timing/model': 1.4665183229371905, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 0.38}) +[2026-06-15 15:58:12,757][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.1909598708152771, 'train/grad_norm_pre_clip': 6.456474781036377, 'timing/data': 0.012279347516596317, 'timing/model': 1.4626197116449475, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 0.38}) +[2026-06-15 15:58:18,653][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.2455158233642578, 'train/grad_norm_pre_clip': 4.631417274475098, 'timing/data': 0.010761484503746033, 'timing/model': 1.4301110841333866, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 0.38}) +[2026-06-15 15:58:24,555][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.39700809121131897, 'train/grad_norm_pre_clip': 16.261629104614258, 'timing/data': 0.011381379328668118, 'timing/model': 1.4521418316289783, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 0.38}) +[2026-06-15 15:58:30,476][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.12608090043067932, 'train/grad_norm_pre_clip': 3.436166524887085, 'timing/data': 0.011115147732198238, 'timing/model': 1.6831802520900965, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 0.38}) +[2026-06-15 15:58:36,159][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.09804310649633408, 'train/grad_norm_pre_clip': 2.6297988891601562, 'timing/data': 0.01101495698094368, 'timing/model': 1.4615487484261394, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 0.38}) +[2026-06-15 15:58:42,163][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.1240183413028717, 'train/grad_norm_pre_clip': 4.638566017150879, 'timing/data': 0.011029975488781929, 'timing/model': 1.4739298978820443, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 0.38}) +[2026-06-15 15:58:48,063][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.2107073813676834, 'train/grad_norm_pre_clip': 4.200104236602783, 'timing/data': 0.010783915407955647, 'timing/model': 1.4696204764768481, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 0.38}) +[2026-06-15 15:58:54,022][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.06552568823099136, 'train/grad_norm_pre_clip': 2.4798519611358643, 'timing/data': 0.01188554335385561, 'timing/model': 1.489321481436491, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.38}) +[2026-06-15 15:58:59,960][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.181682750582695, 'train/grad_norm_pre_clip': 3.7831928730010986, 'timing/data': 0.011065766215324402, 'timing/model': 1.4538817079737782, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 0.38}) +[2026-06-15 15:59:05,890][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.2787875235080719, 'train/grad_norm_pre_clip': 13.622049331665039, 'timing/data': 0.010822324082255363, 'timing/model': 1.6903976369649172, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 0.38}) +[2026-06-15 15:59:11,554][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.26203954219818115, 'train/grad_norm_pre_clip': 7.735111713409424, 'timing/data': 0.013058572076261044, 'timing/model': 1.4502547774463892, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 0.38}) +[2026-06-15 15:59:17,461][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.36248108744621277, 'train/grad_norm_pre_clip': 16.490001678466797, 'timing/data': 0.010758382268249989, 'timing/model': 1.463641844689846, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 0.38}) +[2026-06-15 15:59:23,353][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.21036913990974426, 'train/grad_norm_pre_clip': 5.929283618927002, 'timing/data': 0.010629781521856785, 'timing/model': 1.4673500880599022, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.38}) +[2026-06-15 15:59:29,360][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.2062557339668274, 'train/grad_norm_pre_clip': 7.69074821472168, 'timing/data': 0.013079080730676651, 'timing/model': 1.4755600336939096, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 0.38}) +[2026-06-15 15:59:35,256][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.2746420204639435, 'train/grad_norm_pre_clip': 6.7921671867370605, 'timing/data': 0.012705947272479534, 'timing/model': 1.456964616663754, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 0.39}) +[2026-06-15 15:59:41,190][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.4002913534641266, 'train/grad_norm_pre_clip': 9.80783462524414, 'timing/data': 0.01085637416690588, 'timing/model': 1.6991106262430549, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 0.39}) +[2026-06-15 15:59:46,858][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.4126586616039276, 'train/grad_norm_pre_clip': 10.060477256774902, 'timing/data': 0.011249695904552937, 'timing/model': 1.4587914422154427, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 0.39}) +[2026-06-15 15:59:52,754][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.10046761482954025, 'train/grad_norm_pre_clip': 5.893044471740723, 'timing/data': 0.011119813658297062, 'timing/model': 1.4547437466681004, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 0.39}) +[2026-06-15 15:59:58,651][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.07242351770401001, 'train/grad_norm_pre_clip': 3.1062252521514893, 'timing/data': 0.011010603979229927, 'timing/model': 1.4625274715945125, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 0.39}) +[2026-06-15 16:00:04,563][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.21907705068588257, 'train/grad_norm_pre_clip': 9.947006225585938, 'timing/data': 0.011015393771231174, 'timing/model': 1.4317038841545582, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 0.39}) +[2026-06-15 16:00:10,453][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.16005495190620422, 'train/grad_norm_pre_clip': 5.512881755828857, 'timing/data': 0.01106973271816969, 'timing/model': 1.4425265528261662, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 0.39}) +[2026-06-15 16:00:16,398][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.07279989123344421, 'train/grad_norm_pre_clip': 2.7648391723632812, 'timing/data': 0.011066243052482605, 'timing/model': 1.7020895518362522, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 0.39}) +[2026-06-15 16:00:22,062][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.13181447982788086, 'train/grad_norm_pre_clip': 4.450425148010254, 'timing/data': 0.011363496072590351, 'timing/model': 1.4670589426532388, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.39}) +[2026-06-15 16:00:27,962][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.5220147967338562, 'train/grad_norm_pre_clip': 28.68053436279297, 'timing/data': 0.011965001001954079, 'timing/model': 1.4580079820007086, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 0.39}) +[2026-06-15 16:00:33,863][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.39288055896759033, 'train/grad_norm_pre_clip': 16.340986251831055, 'timing/data': 0.01103170309215784, 'timing/model': 1.461588746868074, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 0.39}) +[2026-06-15 16:00:39,759][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.06376215815544128, 'train/grad_norm_pre_clip': 2.9633500576019287, 'timing/data': 0.010849090293049812, 'timing/model': 1.4589199349284172, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 0.39}) +[2026-06-15 16:00:45,755][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.5628826022148132, 'train/grad_norm_pre_clip': 31.533802032470703, 'timing/data': 0.012678715400397778, 'timing/model': 1.515826117247343, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 0.39}) +[2026-06-15 16:00:51,695][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.32944393157958984, 'train/grad_norm_pre_clip': 14.735844612121582, 'timing/data': 0.011133483611047268, 'timing/model': 1.6990026719868183, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 0.39}) +[2026-06-15 16:00:57,359][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.3585461676120758, 'train/grad_norm_pre_clip': 12.941222190856934, 'timing/data': 0.011908559128642082, 'timing/model': 1.4572232868522406, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 0.39}) +[2026-06-15 16:01:03,363][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.17893193662166595, 'train/grad_norm_pre_clip': 9.110081672668457, 'timing/data': 0.011797738261520863, 'timing/model': 1.4567457120865583, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 0.39}) +[2026-06-15 16:01:09,262][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.2787635922431946, 'train/grad_norm_pre_clip': 5.635207653045654, 'timing/data': 0.010788950137794018, 'timing/model': 1.468931121751666, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 0.39}) +[2026-06-15 16:01:15,154][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.3141633868217468, 'train/grad_norm_pre_clip': 10.219258308410645, 'timing/data': 0.011339893564581871, 'timing/model': 1.4524587290361524, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 0.39}) +[2026-06-15 16:01:21,059][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.2874271273612976, 'train/grad_norm_pre_clip': 8.588912963867188, 'timing/data': 0.01133643463253975, 'timing/model': 1.4289011536166072, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.39}) +[2026-06-15 16:01:26,992][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.3712823987007141, 'train/grad_norm_pre_clip': 9.760287284851074, 'timing/data': 0.010712549090385437, 'timing/model': 1.6995929116383195, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 0.39}) +[2026-06-15 16:01:32,655][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.24084341526031494, 'train/grad_norm_pre_clip': 9.23252010345459, 'timing/data': 0.011561024934053421, 'timing/model': 1.458587922155857, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 0.39}) +[2026-06-15 16:01:38,555][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.27529045939445496, 'train/grad_norm_pre_clip': 5.053953647613525, 'timing/data': 0.011398914270102978, 'timing/model': 1.460184975527227, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 0.39}) +[2026-06-15 16:01:44,451][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.3007224500179291, 'train/grad_norm_pre_clip': 4.928286552429199, 'timing/data': 0.010319205932319164, 'timing/model': 1.457053222693503, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 0.4}) +[2026-06-15 16:01:50,349][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.2805956304073334, 'train/grad_norm_pre_clip': 6.121391296386719, 'timing/data': 0.010710588656365871, 'timing/model': 1.4605052173137665, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 0.4}) +[2026-06-15 16:01:56,251][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.3172880709171295, 'train/grad_norm_pre_clip': 7.629522800445557, 'timing/data': 0.010523277334868908, 'timing/model': 1.425785685889423, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 0.4}) +[2026-06-15 16:02:02,181][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.16977858543395996, 'train/grad_norm_pre_clip': 9.93011474609375, 'timing/data': 0.010442277416586876, 'timing/model': 1.6872594561427832, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 0.4}) +[2026-06-15 16:02:07,848][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.16311590373516083, 'train/grad_norm_pre_clip': 2.708531141281128, 'timing/data': 0.01167125441133976, 'timing/model': 1.4621457550674677, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 0.4}) +[2026-06-15 16:02:13,749][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.19343647360801697, 'train/grad_norm_pre_clip': 2.836146593093872, 'timing/data': 0.01033590454608202, 'timing/model': 1.455240335315466, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 0.4}) +[2026-06-15 16:02:19,651][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.2785255014896393, 'train/grad_norm_pre_clip': 5.836369037628174, 'timing/data': 0.010692458599805832, 'timing/model': 1.4559610523283482, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.4}) +[2026-06-15 16:02:25,557][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.16520512104034424, 'train/grad_norm_pre_clip': 3.8723442554473877, 'timing/data': 0.01095408946275711, 'timing/model': 1.4718083888292313, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 0.4}) +[2026-06-15 16:02:31,448][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.15898893773555756, 'train/grad_norm_pre_clip': 4.312436103820801, 'timing/data': 0.010760157369077206, 'timing/model': 1.425492656417191, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 0.4}) +[2026-06-15 16:02:37,374][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.1481867879629135, 'train/grad_norm_pre_clip': 3.610886335372925, 'timing/data': 0.010663657449185848, 'timing/model': 1.6829983396455646, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 0.4}) +[2026-06-15 16:02:43,218][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.2582460343837738, 'train/grad_norm_pre_clip': 9.891639709472656, 'timing/data': 0.010397004894912243, 'timing/model': 1.6272168485447764, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 0.4}) +[2026-06-15 16:02:48,848][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.17060275375843048, 'train/grad_norm_pre_clip': 11.517443656921387, 'timing/data': 0.011061360128223896, 'timing/model': 1.4548141974955797, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 0.4}) +[2026-06-15 16:02:54,749][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.5519242882728577, 'train/grad_norm_pre_clip': 12.237308502197266, 'timing/data': 0.01016343291848898, 'timing/model': 1.4565150933340192, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 0.4}) +[2026-06-15 16:03:00,648][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.30845150351524353, 'train/grad_norm_pre_clip': 10.791465759277344, 'timing/data': 0.011238191276788712, 'timing/model': 1.4536418551579118, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 0.4}) +[2026-06-15 16:03:06,560][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.13293446600437164, 'train/grad_norm_pre_clip': 4.079535961151123, 'timing/data': 0.010522236116230488, 'timing/model': 1.4273564806208014, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 0.4}) +[2026-06-15 16:03:12,477][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.26504969596862793, 'train/grad_norm_pre_clip': 6.8657379150390625, 'timing/data': 0.010919108055531979, 'timing/model': 1.6860144147649407, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 0.4}) +[2026-06-15 16:03:18,487][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.04317525029182434, 'train/grad_norm_pre_clip': 2.345700740814209, 'timing/data': 0.010720967315137386, 'timing/model': 1.6932626571506262, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 0.4}) +[2026-06-15 16:03:24,147][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.2781263291835785, 'train/grad_norm_pre_clip': 6.369626522064209, 'timing/data': 0.01047433540225029, 'timing/model': 1.457377165555954, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 0.4}) +[2026-06-15 16:03:30,063][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.07408849895000458, 'train/grad_norm_pre_clip': 3.5435211658477783, 'timing/data': 0.010658345185220242, 'timing/model': 1.4701563203707337, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 0.4}) +[2026-06-15 16:03:35,953][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.2818443477153778, 'train/grad_norm_pre_clip': 8.275650024414062, 'timing/data': 0.011305919848382473, 'timing/model': 1.4661953281611204, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 0.4}) +[2026-06-15 16:03:41,860][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.07786716520786285, 'train/grad_norm_pre_clip': 3.9722392559051514, 'timing/data': 0.010741285979747772, 'timing/model': 1.4270075736567378, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.4}) +[2026-06-15 16:03:47,785][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.09930820763111115, 'train/grad_norm_pre_clip': 6.6181254386901855, 'timing/data': 0.010747966356575489, 'timing/model': 1.6962781576439738, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 0.4}) +[2026-06-15 16:03:53,456][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.2585251033306122, 'train/grad_norm_pre_clip': 7.825902462005615, 'timing/data': 0.01125510036945343, 'timing/model': 1.4685921082273126, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 0.41}) +[2026-06-15 16:03:59,360][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.4979930520057678, 'train/grad_norm_pre_clip': 8.906171798706055, 'timing/data': 0.011421380564570427, 'timing/model': 1.4635746479034424, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 0.41}) +[2026-06-15 16:04:05,260][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.0402638278901577, 'train/grad_norm_pre_clip': 3.035221815109253, 'timing/data': 0.010712205432355404, 'timing/model': 1.4666123054921627, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 0.41}) +[2026-06-15 16:04:11,156][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.21463829278945923, 'train/grad_norm_pre_clip': 7.157097339630127, 'timing/data': 0.011713633313775063, 'timing/model': 1.4337703678756952, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 0.41}) +[2026-06-15 16:04:17,066][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.2979260981082916, 'train/grad_norm_pre_clip': 5.965837478637695, 'timing/data': 0.011533881537616253, 'timing/model': 1.4881076756864786, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 0.41}) +[2026-06-15 16:04:22,950][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.11648596823215485, 'train/grad_norm_pre_clip': 5.419058322906494, 'timing/data': 0.011351489461958408, 'timing/model': 1.6525480076670647, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 0.41}) +[2026-06-15 16:04:28,661][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.44759485125541687, 'train/grad_norm_pre_clip': 8.777040481567383, 'timing/data': 0.011439800262451172, 'timing/model': 1.4607356963679194, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 0.41}) +[2026-06-15 16:04:34,557][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.06451208889484406, 'train/grad_norm_pre_clip': 3.784170150756836, 'timing/data': 0.010946487076580524, 'timing/model': 1.4539944333955646, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 0.41}) +[2026-06-15 16:04:40,564][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.1503361463546753, 'train/grad_norm_pre_clip': 7.71862268447876, 'timing/data': 0.012643038295209408, 'timing/model': 1.484578243456781, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 0.41}) +[2026-06-15 16:04:46,455][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.21728160977363586, 'train/grad_norm_pre_clip': 7.660928726196289, 'timing/data': 0.011725371703505516, 'timing/model': 1.4326208364218473, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 0.41}) +[2026-06-15 16:04:52,453][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.10565481334924698, 'train/grad_norm_pre_clip': 3.326174259185791, 'timing/data': 0.01258131768554449, 'timing/model': 1.514263984747231, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 0.41}) +[2026-06-15 16:04:58,352][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.18349160254001617, 'train/grad_norm_pre_clip': 7.886648654937744, 'timing/data': 0.010920315980911255, 'timing/model': 1.6559197381138802, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 0.41}) +[2026-06-15 16:05:04,057][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.13617657124996185, 'train/grad_norm_pre_clip': 7.149087905883789, 'timing/data': 0.011086177080869675, 'timing/model': 1.4583136169239879, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 0.41}) +[2026-06-15 16:05:09,958][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.30656856298446655, 'train/grad_norm_pre_clip': 8.138615608215332, 'timing/data': 0.010949326679110527, 'timing/model': 1.4568208530545235, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 0.41}) +[2026-06-15 16:15:36,940][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.23590143024921417, 'train/grad_norm_pre_clip': 9.864616394042969, 'eval/loss': 0.11726564913988113, 'eval/flap_precision': 0.5979381443298969, 'eval/flap_recall': 0.8436363636363636, 'eval/flap_f1': 0.6998491704374057, 'eval/flap_rate_model': 0.10777777777777778, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6998491704374057, 'timing/data': 0.01115137804299593, 'timing/model': 1.4545898353680968, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 0.41}) +[2026-06-15 16:15:36,941][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_900_state +[2026-06-15 16:15:36,947][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 16:15:44,965][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_900_state/model.safetensors +[2026-06-15 16:15:56,349][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_900_state/optimizer.bin +[2026-06-15 16:15:56,350][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_900_state/sampler.bin +[2026-06-15 16:15:56,350][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_900_state/sampler_1.bin +[2026-06-15 16:15:56,351][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_900_state/random_states_0.pkl +[2026-06-15 16:15:56,517][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-06-15 16:15:56,517][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 16:15:56,532][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 16:16:02,557][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.3041633665561676, 'train/grad_norm_pre_clip': 5.987792491912842, 'timing/data': 0.011837258003652096, 'timing/model': 1.483585000038147, 'learning_rate/qwen_vl_interface': 9.387268552285024e-06, 'learning_rate/action_model': 9.387268552285025e-05, 'epoch': 0.41}) +[2026-06-15 16:16:08,537][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.308434396982193, 'train/grad_norm_pre_clip': 10.462217330932617, 'timing/data': 0.011974168941378593, 'timing/model': 1.4969987412914634, 'learning_rate/qwen_vl_interface': 9.385771559309125e-06, 'learning_rate/action_model': 9.385771559309125e-05, 'epoch': 0.41}) +[2026-06-15 16:16:14,468][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.1931394636631012, 'train/grad_norm_pre_clip': 6.295013904571533, 'timing/data': 0.010436165146529675, 'timing/model': 1.686840602196753, 'learning_rate/qwen_vl_interface': 9.38427286627376e-06, 'learning_rate/action_model': 9.38427286627376e-05, 'epoch': 0.41}) +[2026-06-15 16:16:20,536][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.22816874086856842, 'train/grad_norm_pre_clip': 6.534770488739014, 'timing/data': 0.014062975533306599, 'timing/model': 1.747544838115573, 'learning_rate/qwen_vl_interface': 9.382772473794985e-06, 'learning_rate/action_model': 9.382772473794985e-05, 'epoch': 0.41}) +[2026-06-15 16:16:26,253][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.43580400943756104, 'train/grad_norm_pre_clip': 6.16646671295166, 'timing/data': 0.012582002207636833, 'timing/model': 1.4758118661120534, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.41}) +[2026-06-15 16:16:32,143][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.13943766057491302, 'train/grad_norm_pre_clip': 4.61878776550293, 'timing/data': 0.01092227827757597, 'timing/model': 1.4501370852813125, 'learning_rate/qwen_vl_interface': 9.37976659297492e-06, 'learning_rate/action_model': 9.379766592974921e-05, 'epoch': 0.41}) +[2026-06-15 16:16:38,056][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.3138698935508728, 'train/grad_norm_pre_clip': 7.194035053253174, 'timing/data': 0.010770128108561039, 'timing/model': 1.4746821410953999, 'learning_rate/qwen_vl_interface': 9.378261105869237e-06, 'learning_rate/action_model': 9.378261105869237e-05, 'epoch': 0.41}) +[2026-06-15 16:16:43,945][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.2940983176231384, 'train/grad_norm_pre_clip': 6.484519004821777, 'timing/data': 0.010831798426806927, 'timing/model': 1.4137551132589579, 'learning_rate/qwen_vl_interface': 9.376753921791348e-06, 'learning_rate/action_model': 9.376753921791348e-05, 'epoch': 0.42}) +[2026-06-15 16:16:49,849][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.1767999827861786, 'train/grad_norm_pre_clip': 6.23225736618042, 'timing/data': 0.01065054815262556, 'timing/model': 1.4551892951130867, 'learning_rate/qwen_vl_interface': 9.375245041360805e-06, 'learning_rate/action_model': 9.375245041360804e-05, 'epoch': 0.42}) +[2026-06-15 16:16:55,782][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.16013430058956146, 'train/grad_norm_pre_clip': 3.813542366027832, 'timing/data': 0.011436720378696918, 'timing/model': 1.690416851080954, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 0.42}) +[2026-06-15 16:17:01,457][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.13958966732025146, 'train/grad_norm_pre_clip': 3.4541683197021484, 'timing/data': 0.01163497194647789, 'timing/model': 1.4715733602643013, 'learning_rate/qwen_vl_interface': 9.372222193923421e-06, 'learning_rate/action_model': 9.372222193923421e-05, 'epoch': 0.42}) +[2026-06-15 16:17:07,359][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.24850966036319733, 'train/grad_norm_pre_clip': 10.353029251098633, 'timing/data': 0.011575501412153244, 'timing/model': 1.463306968100369, 'learning_rate/qwen_vl_interface': 9.370708228159158e-06, 'learning_rate/action_model': 9.370708228159159e-05, 'epoch': 0.42}) +[2026-06-15 16:17:13,431][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.2275049090385437, 'train/grad_norm_pre_clip': 5.958876609802246, 'timing/data': 0.013857979327440262, 'timing/model': 1.5340211866423488, 'learning_rate/qwen_vl_interface': 9.369192568527398e-06, 'learning_rate/action_model': 9.369192568527398e-05, 'epoch': 0.42}) +[2026-06-15 16:17:19,468][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.42192649841308594, 'train/grad_norm_pre_clip': 10.260360717773438, 'timing/data': 0.011551032774150372, 'timing/model': 1.4808065285906196, 'learning_rate/qwen_vl_interface': 9.367675215651168e-06, 'learning_rate/action_model': 9.367675215651168e-05, 'epoch': 0.42}) +[2026-06-15 16:17:25,470][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.2948143780231476, 'train/grad_norm_pre_clip': 5.660200119018555, 'timing/data': 0.011599763296544552, 'timing/model': 1.5304262908175588, 'learning_rate/qwen_vl_interface': 9.366156170154194e-06, 'learning_rate/action_model': 9.366156170154194e-05, 'epoch': 0.42}) +[2026-06-15 16:17:31,408][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.22253744304180145, 'train/grad_norm_pre_clip': 4.718338489532471, 'timing/data': 0.011921294033527374, 'timing/model': 1.7023884309455752, 'learning_rate/qwen_vl_interface': 9.3646354326609e-06, 'learning_rate/action_model': 9.364635432660899e-05, 'epoch': 0.42}) +[2026-06-15 16:17:37,066][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.2959353029727936, 'train/grad_norm_pre_clip': 6.644247055053711, 'timing/data': 0.01250611711293459, 'timing/model': 1.4738072818145156, 'learning_rate/qwen_vl_interface': 9.3631130037964e-06, 'learning_rate/action_model': 9.363113003796401e-05, 'epoch': 0.42}) +[2026-06-15 16:17:42,975][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.1584351658821106, 'train/grad_norm_pre_clip': 12.662957191467285, 'timing/data': 0.010948533192276955, 'timing/model': 1.4743360904976726, 'learning_rate/qwen_vl_interface': 9.361588884186513e-06, 'learning_rate/action_model': 9.361588884186513e-05, 'epoch': 0.42}) +[2026-06-15 16:17:48,960][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.223152294754982, 'train/grad_norm_pre_clip': 6.738457679748535, 'timing/data': 0.011378344148397446, 'timing/model': 1.4603813355788589, 'learning_rate/qwen_vl_interface': 9.360063074457744e-06, 'learning_rate/action_model': 9.360063074457743e-05, 'epoch': 0.42}) +[2026-06-15 16:17:54,861][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.25025099515914917, 'train/grad_norm_pre_clip': 3.9487907886505127, 'timing/data': 0.010361052118241787, 'timing/model': 1.436539290472865, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 0.42}) +[2026-06-15 16:18:00,765][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.20238149166107178, 'train/grad_norm_pre_clip': 4.148510456085205, 'timing/data': 0.011095214635133743, 'timing/model': 1.4536087671294808, 'learning_rate/qwen_vl_interface': 9.357006387153065e-06, 'learning_rate/action_model': 9.357006387153065e-05, 'epoch': 0.42}) +[2026-06-15 16:18:06,719][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.20256488025188446, 'train/grad_norm_pre_clip': 7.7291436195373535, 'timing/data': 0.010966334491968155, 'timing/model': 1.722529104910791, 'learning_rate/qwen_vl_interface': 9.355475510833643e-06, 'learning_rate/action_model': 9.355475510833643e-05, 'epoch': 0.42}) +[2026-06-15 16:18:12,353][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.3449348509311676, 'train/grad_norm_pre_clip': 10.673283576965332, 'timing/data': 0.011188664473593235, 'timing/model': 1.4520178996026516, 'learning_rate/qwen_vl_interface': 9.353942946908318e-06, 'learning_rate/action_model': 9.353942946908317e-05, 'epoch': 0.42}) +[2026-06-15 16:18:18,256][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.19553187489509583, 'train/grad_norm_pre_clip': 3.463639736175537, 'timing/data': 0.01128406636416912, 'timing/model': 1.4572918489575386, 'learning_rate/qwen_vl_interface': 9.352408696007065e-06, 'learning_rate/action_model': 9.352408696007065e-05, 'epoch': 0.42}) +[2026-06-15 16:18:24,160][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.4545976519584656, 'train/grad_norm_pre_clip': 8.152958869934082, 'timing/data': 0.010966915637254715, 'timing/model': 1.4704516949132085, 'learning_rate/qwen_vl_interface': 9.350872758760558e-06, 'learning_rate/action_model': 9.350872758760559e-05, 'epoch': 0.42}) +[2026-06-15 16:18:30,349][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.09791374206542969, 'train/grad_norm_pre_clip': 3.5201094150543213, 'timing/data': 0.011440165340900421, 'timing/model': 1.427520101889968, 'learning_rate/qwen_vl_interface': 9.349335135800164e-06, 'learning_rate/action_model': 9.349335135800163e-05, 'epoch': 0.42}) +[2026-06-15 16:18:36,353][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.14414629340171814, 'train/grad_norm_pre_clip': 7.013246059417725, 'timing/data': 0.010788654908537865, 'timing/model': 1.506754050962627, 'learning_rate/qwen_vl_interface': 9.347795827757942e-06, 'learning_rate/action_model': 9.347795827757941e-05, 'epoch': 0.42}) +[2026-06-15 16:18:42,291][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.2503359615802765, 'train/grad_norm_pre_clip': 6.6902618408203125, 'timing/data': 0.010850464925169945, 'timing/model': 1.69821113999933, 'learning_rate/qwen_vl_interface': 9.346254835266643e-06, 'learning_rate/action_model': 9.346254835266642e-05, 'epoch': 0.42}) +[2026-06-15 16:18:47,960][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.30352240800857544, 'train/grad_norm_pre_clip': 8.221842765808105, 'timing/data': 0.01119245681911707, 'timing/model': 1.4680146165192127, 'learning_rate/qwen_vl_interface': 9.344712158959709e-06, 'learning_rate/action_model': 9.344712158959709e-05, 'epoch': 0.42}) +[2026-06-15 16:18:53,859][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.39095306396484375, 'train/grad_norm_pre_clip': 13.318550109863281, 'timing/data': 0.011032856069505215, 'timing/model': 1.4612135356292129, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 0.43}) +[2026-06-15 16:18:59,764][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.1580144166946411, 'train/grad_norm_pre_clip': 6.372247219085693, 'timing/data': 0.011364737525582314, 'timing/model': 1.4681324288249016, 'learning_rate/qwen_vl_interface': 9.341621757436177e-06, 'learning_rate/action_model': 9.341621757436177e-05, 'epoch': 0.43}) +[2026-06-15 16:19:05,673][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.3592390716075897, 'train/grad_norm_pre_clip': 8.201226234436035, 'timing/data': 0.011087877675890923, 'timing/model': 1.4800918102264404, 'learning_rate/qwen_vl_interface': 9.340074033489924e-06, 'learning_rate/action_model': 9.340074033489923e-05, 'epoch': 0.43}) +[2026-06-15 16:19:11,637][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.151672825217247, 'train/grad_norm_pre_clip': 5.166295051574707, 'timing/data': 0.011707600206136703, 'timing/model': 1.5035123098641634, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.43}) +[2026-06-15 16:19:17,594][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.20888745784759521, 'train/grad_norm_pre_clip': 7.4904255867004395, 'timing/data': 0.011466248892247677, 'timing/model': 1.7000936148688197, 'learning_rate/qwen_vl_interface': 9.336973542409505e-06, 'learning_rate/action_model': 9.336973542409504e-05, 'epoch': 0.43}) +[2026-06-15 16:19:23,263][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.03568378835916519, 'train/grad_norm_pre_clip': 2.666386127471924, 'timing/data': 0.01136476919054985, 'timing/model': 1.464192831888795, 'learning_rate/qwen_vl_interface': 9.33542077654983e-06, 'learning_rate/action_model': 9.33542077654983e-05, 'epoch': 0.43}) +[2026-06-15 16:19:29,162][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.12430385500192642, 'train/grad_norm_pre_clip': 8.278311729431152, 'timing/data': 0.010974157601594925, 'timing/model': 1.4588775215670466, 'learning_rate/qwen_vl_interface': 9.333866331327995e-06, 'learning_rate/action_model': 9.333866331327995e-05, 'epoch': 0.43}) +[2026-06-15 16:19:35,161][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.23715192079544067, 'train/grad_norm_pre_clip': 8.465583801269531, 'timing/data': 0.010798127390444279, 'timing/model': 1.4578842855989933, 'learning_rate/qwen_vl_interface': 9.332310207382972e-06, 'learning_rate/action_model': 9.332310207382972e-05, 'epoch': 0.43}) +[2026-06-15 16:19:41,066][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.06285159289836884, 'train/grad_norm_pre_clip': 1.8051753044128418, 'timing/data': 0.011012609116733074, 'timing/model': 1.4704206679016352, 'learning_rate/qwen_vl_interface': 9.330752405354426e-06, 'learning_rate/action_model': 9.330752405354426e-05, 'epoch': 0.43}) +[2026-06-15 16:19:47,003][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.19042548537254333, 'train/grad_norm_pre_clip': 4.670975685119629, 'timing/data': 0.010969819501042366, 'timing/model': 1.4703165842220187, 'learning_rate/qwen_vl_interface': 9.32919292588271e-06, 'learning_rate/action_model': 9.32919292588271e-05, 'epoch': 0.43}) +[2026-06-15 16:19:52,882][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.15696865320205688, 'train/grad_norm_pre_clip': 4.844967842102051, 'timing/data': 0.010876759886741638, 'timing/model': 1.6907991021871567, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.43}) +[2026-06-15 16:19:58,559][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.287501722574234, 'train/grad_norm_pre_clip': 8.997406005859375, 'timing/data': 0.011695342138409615, 'timing/model': 1.4722537230700254, 'learning_rate/qwen_vl_interface': 9.326068937174627e-06, 'learning_rate/action_model': 9.326068937174626e-05, 'epoch': 0.43}) +[2026-06-15 16:20:04,457][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.20665039122104645, 'train/grad_norm_pre_clip': 5.574793338775635, 'timing/data': 0.01036708615720272, 'timing/model': 1.468168725259602, 'learning_rate/qwen_vl_interface': 9.324504429222412e-06, 'learning_rate/action_model': 9.324504429222412e-05, 'epoch': 0.43}) +[2026-06-15 16:20:10,361][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.24146613478660583, 'train/grad_norm_pre_clip': 6.910505294799805, 'timing/data': 0.01222944539040327, 'timing/model': 1.4705006079748273, 'learning_rate/qwen_vl_interface': 9.322938246395336e-06, 'learning_rate/action_model': 9.322938246395337e-05, 'epoch': 0.43}) +[2026-06-15 16:20:16,278][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.10418979078531265, 'train/grad_norm_pre_clip': 3.1605217456817627, 'timing/data': 0.010166707448661327, 'timing/model': 1.4937064172700047, 'learning_rate/qwen_vl_interface': 9.321370389337196e-06, 'learning_rate/action_model': 9.321370389337196e-05, 'epoch': 0.43}) +[2026-06-15 16:20:22,355][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.17560158669948578, 'train/grad_norm_pre_clip': 5.913926124572754, 'timing/data': 0.013789631426334381, 'timing/model': 1.51453763153404, 'learning_rate/qwen_vl_interface': 9.319800858692478e-06, 'learning_rate/action_model': 9.319800858692476e-05, 'epoch': 0.43}) +[2026-06-15 16:20:28,296][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.1445719599723816, 'train/grad_norm_pre_clip': 3.52302622795105, 'timing/data': 0.010706889443099499, 'timing/model': 1.694433018565178, 'learning_rate/qwen_vl_interface': 9.318229655106357e-06, 'learning_rate/action_model': 9.318229655106356e-05, 'epoch': 0.43}) +[2026-06-15 16:20:34,138][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.23912234604358673, 'train/grad_norm_pre_clip': 9.790392875671387, 'timing/data': 0.010870041325688362, 'timing/model': 1.64056786429137, 'learning_rate/qwen_vl_interface': 9.316656779224693e-06, 'learning_rate/action_model': 9.316656779224693e-05, 'epoch': 0.43}) +[2026-06-15 16:20:39,860][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.12776534259319305, 'train/grad_norm_pre_clip': 4.922678470611572, 'timing/data': 0.011736124753952026, 'timing/model': 1.4592621242627501, 'learning_rate/qwen_vl_interface': 9.31508223169404e-06, 'learning_rate/action_model': 9.31508223169404e-05, 'epoch': 0.43}) +[2026-06-15 16:20:45,760][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.07083391398191452, 'train/grad_norm_pre_clip': 3.128469944000244, 'timing/data': 0.0111802713945508, 'timing/model': 1.4542249096557498, 'learning_rate/qwen_vl_interface': 9.313506013161633e-06, 'learning_rate/action_model': 9.313506013161633e-05, 'epoch': 0.43}) +[2026-06-15 16:20:51,862][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.2092311978340149, 'train/grad_norm_pre_clip': 3.5804624557495117, 'timing/data': 0.010884000919759274, 'timing/model': 1.5686826715245843, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 0.43}) +[2026-06-15 16:20:57,808][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.11124873906373978, 'train/grad_norm_pre_clip': 6.457089900970459, 'timing/data': 0.01071260217577219, 'timing/model': 1.4753074049949646, 'learning_rate/qwen_vl_interface': 9.310348565683938e-06, 'learning_rate/action_model': 9.310348565683939e-05, 'epoch': 0.43}) +[2026-06-15 16:21:03,695][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.7119507789611816, 'train/grad_norm_pre_clip': 10.723376274108887, 'timing/data': 0.010712052695453167, 'timing/model': 1.6996475448831916, 'learning_rate/qwen_vl_interface': 9.30876733803656e-06, 'learning_rate/action_model': 9.308767338036559e-05, 'epoch': 0.44}) +[2026-06-15 16:21:09,607][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.09633772820234299, 'train/grad_norm_pre_clip': 22.300424575805664, 'timing/data': 0.009788268245756626, 'timing/model': 1.711706998758018, 'learning_rate/qwen_vl_interface': 9.30718444198324e-06, 'learning_rate/action_model': 9.307184441983239e-05, 'epoch': 0.44}) +[2026-06-15 16:21:15,281][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.2116497904062271, 'train/grad_norm_pre_clip': 7.391093730926514, 'timing/data': 0.011046633124351501, 'timing/model': 1.4853559993207455, 'learning_rate/qwen_vl_interface': 9.30559987817465e-06, 'learning_rate/action_model': 9.30559987817465e-05, 'epoch': 0.44}) +[2026-06-15 16:21:21,163][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.11959852278232574, 'train/grad_norm_pre_clip': 9.734451293945312, 'timing/data': 0.010989923030138016, 'timing/model': 1.4696628581732512, 'learning_rate/qwen_vl_interface': 9.304013647262143e-06, 'learning_rate/action_model': 9.304013647262143e-05, 'epoch': 0.44}) +[2026-06-15 16:21:27,060][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.14454184472560883, 'train/grad_norm_pre_clip': 5.949715614318848, 'timing/data': 0.010707251727581024, 'timing/model': 1.4621356492862105, 'learning_rate/qwen_vl_interface': 9.302425749897761e-06, 'learning_rate/action_model': 9.302425749897761e-05, 'epoch': 0.44}) +[2026-06-15 16:21:32,965][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.09747378528118134, 'train/grad_norm_pre_clip': 5.141660690307617, 'timing/data': 0.010639913380146027, 'timing/model': 1.432978492230177, 'learning_rate/qwen_vl_interface': 9.300836186734225e-06, 'learning_rate/action_model': 9.300836186734226e-05, 'epoch': 0.44}) +[2026-06-15 16:21:38,895][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.14441873133182526, 'train/grad_norm_pre_clip': 9.6626615524292, 'timing/data': 0.010599112138152122, 'timing/model': 1.6971446061506867, 'learning_rate/qwen_vl_interface': 9.299244958424947e-06, 'learning_rate/action_model': 9.299244958424946e-05, 'epoch': 0.44}) +[2026-06-15 16:21:44,562][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.21993650496006012, 'train/grad_norm_pre_clip': 14.409838676452637, 'timing/data': 0.01110069453716278, 'timing/model': 1.463438862003386, 'learning_rate/qwen_vl_interface': 9.29765206562402e-06, 'learning_rate/action_model': 9.297652065624018e-05, 'epoch': 0.44}) +[2026-06-15 16:21:50,457][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.13037048280239105, 'train/grad_norm_pre_clip': 6.621885776519775, 'timing/data': 0.010861873626708984, 'timing/model': 1.469681616872549, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 0.44}) +[2026-06-15 16:21:56,359][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.23828868567943573, 'train/grad_norm_pre_clip': 8.990250587463379, 'timing/data': 0.010808903723955154, 'timing/model': 1.4712954005226493, 'learning_rate/qwen_vl_interface': 9.294461289167013e-06, 'learning_rate/action_model': 9.294461289167014e-05, 'epoch': 0.44}) +[2026-06-15 16:22:02,253][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.250535249710083, 'train/grad_norm_pre_clip': 10.89084243774414, 'timing/data': 0.01121058501303196, 'timing/model': 1.4331037541851401, 'learning_rate/qwen_vl_interface': 9.292863406822543e-06, 'learning_rate/action_model': 9.292863406822542e-05, 'epoch': 0.44}) +[2026-06-15 16:22:08,162][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.2615641951560974, 'train/grad_norm_pre_clip': 5.2810773849487305, 'timing/data': 0.011158234439790249, 'timing/model': 1.4972261535003781, 'learning_rate/qwen_vl_interface': 9.291263862609638e-06, 'learning_rate/action_model': 9.291263862609638e-05, 'epoch': 0.44}) +[2026-06-15 16:22:14,049][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.16367077827453613, 'train/grad_norm_pre_clip': 6.8067827224731445, 'timing/data': 0.0107877841219306, 'timing/model': 1.6535672033205628, 'learning_rate/qwen_vl_interface': 9.28966265718581e-06, 'learning_rate/action_model': 9.28966265718581e-05, 'epoch': 0.44}) +[2026-06-15 16:22:19,763][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.308400958776474, 'train/grad_norm_pre_clip': 13.207005500793457, 'timing/data': 0.013691476546227932, 'timing/model': 1.4555815774947405, 'learning_rate/qwen_vl_interface': 9.288059791209256e-06, 'learning_rate/action_model': 9.288059791209256e-05, 'epoch': 0.44}) +[2026-06-15 16:22:25,770][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.0699535682797432, 'train/grad_norm_pre_clip': 3.2712621688842773, 'timing/data': 0.011807088740170002, 'timing/model': 1.4574154205620289, 'learning_rate/qwen_vl_interface': 9.286455265338853e-06, 'learning_rate/action_model': 9.286455265338853e-05, 'epoch': 0.44}) +[2026-06-15 16:22:31,767][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.13057757914066315, 'train/grad_norm_pre_clip': 3.9742166996002197, 'timing/data': 0.011601267382502556, 'timing/model': 1.4702427117154002, 'learning_rate/qwen_vl_interface': 9.284849080234161e-06, 'learning_rate/action_model': 9.28484908023416e-05, 'epoch': 0.44}) +[2026-06-15 16:22:37,760][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.2747284471988678, 'train/grad_norm_pre_clip': 16.876169204711914, 'timing/data': 0.011419977992773056, 'timing/model': 1.5188826834782958, 'learning_rate/qwen_vl_interface': 9.283241236555421e-06, 'learning_rate/action_model': 9.283241236555421e-05, 'epoch': 0.44}) +[2026-06-15 16:22:43,672][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.15134704113006592, 'train/grad_norm_pre_clip': 5.880776882171631, 'timing/data': 0.010901345871388912, 'timing/model': 1.4394519962370396, 'learning_rate/qwen_vl_interface': 9.281631734963557e-06, 'learning_rate/action_model': 9.281631734963557e-05, 'epoch': 0.44}) +[2026-06-15 16:22:49,568][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.1569419652223587, 'train/grad_norm_pre_clip': 4.593334197998047, 'timing/data': 0.011129826307296753, 'timing/model': 1.6721167918294668, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 0.44}) +[2026-06-15 16:22:55,268][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.15345600247383118, 'train/grad_norm_pre_clip': 5.558639049530029, 'timing/data': 0.011460819281637669, 'timing/model': 1.4667851608246565, 'learning_rate/qwen_vl_interface': 9.278407760687562e-06, 'learning_rate/action_model': 9.278407760687562e-05, 'epoch': 0.44}) +[2026-06-15 16:23:01,156][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.14802759885787964, 'train/grad_norm_pre_clip': 4.9557623863220215, 'timing/data': 0.011024517007172108, 'timing/model': 1.4476445149630308, 'learning_rate/qwen_vl_interface': 9.276793289328686e-06, 'learning_rate/action_model': 9.276793289328685e-05, 'epoch': 0.44}) +[2026-06-15 16:23:07,060][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.2649352252483368, 'train/grad_norm_pre_clip': 11.787030220031738, 'timing/data': 0.011464218609035015, 'timing/model': 1.469865565188229, 'learning_rate/qwen_vl_interface': 9.27517716270719e-06, 'learning_rate/action_model': 9.275177162707191e-05, 'epoch': 0.44}) +[2026-06-15 16:23:13,079][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.294728547334671, 'train/grad_norm_pre_clip': 6.886746406555176, 'timing/data': 0.011595840565860271, 'timing/model': 1.5104759838432074, 'learning_rate/qwen_vl_interface': 9.273559381487409e-06, 'learning_rate/action_model': 9.273559381487409e-05, 'epoch': 0.45}) +[2026-06-15 16:23:18,998][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.12908971309661865, 'train/grad_norm_pre_clip': 6.514124393463135, 'timing/data': 0.010761106386780739, 'timing/model': 1.6973942127078772, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.45}) +[2026-06-15 16:23:24,673][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.2789071798324585, 'train/grad_norm_pre_clip': 17.844064712524414, 'timing/data': 0.01114890817552805, 'timing/model': 1.4687305027619004, 'learning_rate/qwen_vl_interface': 9.2703188579137e-06, 'learning_rate/action_model': 9.2703188579137e-05, 'epoch': 0.45}) +[2026-06-15 16:23:30,666][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.1261262744665146, 'train/grad_norm_pre_clip': 6.344870567321777, 'timing/data': 0.010953688062727451, 'timing/model': 1.4662679256871343, 'learning_rate/qwen_vl_interface': 9.26869611689183e-06, 'learning_rate/action_model': 9.26869611689183e-05, 'epoch': 0.45}) +[2026-06-15 16:23:36,562][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.08700879663228989, 'train/grad_norm_pre_clip': 3.91913104057312, 'timing/data': 0.01112676877528429, 'timing/model': 1.4587929770350456, 'learning_rate/qwen_vl_interface': 9.267071723935784e-06, 'learning_rate/action_model': 9.267071723935784e-05, 'epoch': 0.45}) +[2026-06-15 16:23:42,468][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.2815122604370117, 'train/grad_norm_pre_clip': 9.063000679016113, 'timing/data': 0.01133228000253439, 'timing/model': 1.4688670746982098, 'learning_rate/qwen_vl_interface': 9.26544567971329e-06, 'learning_rate/action_model': 9.26544567971329e-05, 'epoch': 0.45}) +[2026-06-15 16:23:48,369][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.09162786602973938, 'train/grad_norm_pre_clip': 4.994941234588623, 'timing/data': 0.011832580901682377, 'timing/model': 1.4681070446968079, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 0.45}) +[2026-06-15 16:23:54,298][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.18661269545555115, 'train/grad_norm_pre_clip': 6.279958724975586, 'timing/data': 0.012184382416307926, 'timing/model': 1.6877645822241902, 'learning_rate/qwen_vl_interface': 9.262188640143262e-06, 'learning_rate/action_model': 9.262188640143262e-05, 'epoch': 0.45}) +[2026-06-15 16:23:59,961][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.2109878957271576, 'train/grad_norm_pre_clip': 8.528005599975586, 'timing/data': 0.010924138128757477, 'timing/model': 1.4593279585242271, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.45}) +[2026-06-15 16:24:05,959][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.18084374070167542, 'train/grad_norm_pre_clip': 6.622344017028809, 'timing/data': 0.010692018084228039, 'timing/model': 1.4582120263949037, 'learning_rate/qwen_vl_interface': 9.258925003537127e-06, 'learning_rate/action_model': 9.258925003537126e-05, 'epoch': 0.45}) +[2026-06-15 16:24:11,862][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.22584383189678192, 'train/grad_norm_pre_clip': 12.327293395996094, 'timing/data': 0.010775198228657246, 'timing/model': 1.4669159967452288, 'learning_rate/qwen_vl_interface': 9.257290713022041e-06, 'learning_rate/action_model': 9.257290713022041e-05, 'epoch': 0.45}) +[2026-06-15 16:24:17,760][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.03791268914937973, 'train/grad_norm_pre_clip': 3.2920923233032227, 'timing/data': 0.011423391290009022, 'timing/model': 1.46520169172436, 'learning_rate/qwen_vl_interface': 9.255654775261112e-06, 'learning_rate/action_model': 9.255654775261112e-05, 'epoch': 0.45}) +[2026-06-15 16:24:23,653][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.22680889070034027, 'train/grad_norm_pre_clip': 12.93398380279541, 'timing/data': 0.01091052871197462, 'timing/model': 1.4362978041172028, 'learning_rate/qwen_vl_interface': 9.254017190926813e-06, 'learning_rate/action_model': 9.254017190926813e-05, 'epoch': 0.45}) +[2026-06-15 16:24:29,609][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.36351141333580017, 'train/grad_norm_pre_clip': 78.81690216064453, 'timing/data': 0.01144356094300747, 'timing/model': 1.7086225170642138, 'learning_rate/qwen_vl_interface': 9.25237796069229e-06, 'learning_rate/action_model': 9.25237796069229e-05, 'epoch': 0.45}) +[2026-06-15 16:24:35,266][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.27237698435783386, 'train/grad_norm_pre_clip': 9.74363899230957, 'timing/data': 0.011010998860001564, 'timing/model': 1.4659893792122602, 'learning_rate/qwen_vl_interface': 9.250737085231371e-06, 'learning_rate/action_model': 9.250737085231372e-05, 'epoch': 0.45}) +[2026-06-15 16:24:41,157][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.1982681304216385, 'train/grad_norm_pre_clip': 7.1656389236450195, 'timing/data': 0.011289810761809349, 'timing/model': 1.4542314959689975, 'learning_rate/qwen_vl_interface': 9.24909456521856e-06, 'learning_rate/action_model': 9.249094565218559e-05, 'epoch': 0.45}) +[2026-06-15 16:24:47,053][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.1849735826253891, 'train/grad_norm_pre_clip': 7.906459331512451, 'timing/data': 0.009682154282927513, 'timing/model': 1.4649315243586898, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 0.45}) +[2026-06-15 16:24:53,064][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.17140647768974304, 'train/grad_norm_pre_clip': 4.647283554077148, 'timing/data': 0.01065821759402752, 'timing/model': 1.5114392172545195, 'learning_rate/qwen_vl_interface': 9.245804594238637e-06, 'learning_rate/action_model': 9.245804594238636e-05, 'epoch': 0.45}) +[2026-06-15 16:24:58,957][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.28170010447502136, 'train/grad_norm_pre_clip': 5.704637050628662, 'timing/data': 0.011959805153310299, 'timing/model': 1.4481762992218137, 'learning_rate/qwen_vl_interface': 9.244157144623909e-06, 'learning_rate/action_model': 9.244157144623909e-05, 'epoch': 0.45}) +[2026-06-15 16:25:04,891][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.2142888605594635, 'train/grad_norm_pre_clip': 5.852726936340332, 'timing/data': 0.009900244884192944, 'timing/model': 1.6713012484833598, 'learning_rate/qwen_vl_interface': 9.242508053162051e-06, 'learning_rate/action_model': 9.242508053162051e-05, 'epoch': 0.45}) +[2026-06-15 16:25:10,667][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.1839531660079956, 'train/grad_norm_pre_clip': 19.61396026611328, 'timing/data': 0.009729455225169659, 'timing/model': 1.4854615088552237, 'learning_rate/qwen_vl_interface': 9.240857320530944e-06, 'learning_rate/action_model': 9.240857320530943e-05, 'epoch': 0.45}) +[2026-06-15 16:25:16,655][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.39965662360191345, 'train/grad_norm_pre_clip': 9.804638862609863, 'timing/data': 0.010771360248327255, 'timing/model': 1.4628277346491814, 'learning_rate/qwen_vl_interface': 9.23920494740914e-06, 'learning_rate/action_model': 9.23920494740914e-05, 'epoch': 0.46}) +[2026-06-15 16:25:22,556][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.3194325268268585, 'train/grad_norm_pre_clip': 8.43790340423584, 'timing/data': 0.010725598782300949, 'timing/model': 1.4599891602993011, 'learning_rate/qwen_vl_interface': 9.237550934475868e-06, 'learning_rate/action_model': 9.237550934475867e-05, 'epoch': 0.46}) +[2026-06-15 16:25:28,472][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.15342247486114502, 'train/grad_norm_pre_clip': 7.479512691497803, 'timing/data': 0.010346697643399239, 'timing/model': 1.485097105614841, 'learning_rate/qwen_vl_interface': 9.235895282411029e-06, 'learning_rate/action_model': 9.235895282411029e-05, 'epoch': 0.46}) +[2026-06-15 16:25:34,458][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.16571089625358582, 'train/grad_norm_pre_clip': 5.553319454193115, 'timing/data': 0.011069371365010738, 'timing/model': 1.518000777810812, 'learning_rate/qwen_vl_interface': 9.2342379918952e-06, 'learning_rate/action_model': 9.2342379918952e-05, 'epoch': 0.46}) +[2026-06-15 16:25:40,398][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.6675423979759216, 'train/grad_norm_pre_clip': 8.242924690246582, 'timing/data': 0.011958545073866844, 'timing/model': 1.6903213178738952, 'learning_rate/qwen_vl_interface': 9.232579063609631e-06, 'learning_rate/action_model': 9.232579063609631e-05, 'epoch': 0.46}) +[2026-06-15 16:36:13,236][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.18805435299873352, 'train/grad_norm_pre_clip': 4.899032115936279, 'eval/loss': 0.1060563176870346, 'eval/flap_precision': 0.7934782608695652, 'eval/flap_recall': 0.5309090909090909, 'eval/flap_f1': 0.6361655773420478, 'eval/flap_rate_model': 0.051111111111111114, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6361655773420478, 'timing/data': 0.012296456843614578, 'timing/model': 1.4651855686679482, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 0.46}) +[2026-06-15 16:36:13,238][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1000_state +[2026-06-15 16:36:13,245][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 16:36:21,275][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1000_state/model.safetensors +[2026-06-15 16:36:32,324][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1000_state/optimizer.bin +[2026-06-15 16:36:32,324][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1000_state/sampler.bin +[2026-06-15 16:36:32,325][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-15 16:36:32,326][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-15 16:36:32,483][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-15 16:36:32,484][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 16:36:32,499][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 16:36:38,404][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.14355507493019104, 'train/grad_norm_pre_clip': 5.431572914123535, 'timing/data': 0.011399632319808006, 'timing/model': 1.708514536730945, 'learning_rate/qwen_vl_interface': 9.229256296457635e-06, 'learning_rate/action_model': 9.229256296457634e-05, 'epoch': 0.46}) +[2026-06-15 16:36:44,050][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.07992879301309586, 'train/grad_norm_pre_clip': 6.702589511871338, 'timing/data': 0.011784453876316547, 'timing/model': 1.4608005052432418, 'learning_rate/qwen_vl_interface': 9.22759245895707e-06, 'learning_rate/action_model': 9.227592458957071e-05, 'epoch': 0.46}) +[2026-06-15 16:36:49,949][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.41031354665756226, 'train/grad_norm_pre_clip': 7.935831069946289, 'timing/data': 0.010836858302354813, 'timing/model': 1.4613041207194328, 'learning_rate/qwen_vl_interface': 9.225926986418494e-06, 'learning_rate/action_model': 9.225926986418494e-05, 'epoch': 0.46}) +[2026-06-15 16:36:55,835][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.09941574931144714, 'train/grad_norm_pre_clip': 2.2748923301696777, 'timing/data': 0.011045380495488644, 'timing/model': 1.45802381914109, 'learning_rate/qwen_vl_interface': 9.224259879526519e-06, 'learning_rate/action_model': 9.224259879526519e-05, 'epoch': 0.46}) +[2026-06-15 16:37:01,742][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.24632486701011658, 'train/grad_norm_pre_clip': 5.905521392822266, 'timing/data': 0.010716238990426064, 'timing/model': 1.4643733678385615, 'learning_rate/qwen_vl_interface': 9.222591138966427e-06, 'learning_rate/action_model': 9.222591138966426e-05, 'epoch': 0.46}) +[2026-06-15 16:37:07,652][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.35029828548431396, 'train/grad_norm_pre_clip': 10.14939022064209, 'timing/data': 0.011505251750349998, 'timing/model': 1.434673111885786, 'learning_rate/qwen_vl_interface': 9.220920765424175e-06, 'learning_rate/action_model': 9.220920765424174e-05, 'epoch': 0.46}) +[2026-06-15 16:37:13,596][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.21423745155334473, 'train/grad_norm_pre_clip': 7.981302738189697, 'timing/data': 0.012607749551534653, 'timing/model': 1.707440733909607, 'learning_rate/qwen_vl_interface': 9.219248759586392e-06, 'learning_rate/action_model': 9.219248759586391e-05, 'epoch': 0.46}) +[2026-06-15 16:37:19,250][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.13213832676410675, 'train/grad_norm_pre_clip': 2.3856842517852783, 'timing/data': 0.011292312294244766, 'timing/model': 1.4622778808698058, 'learning_rate/qwen_vl_interface': 9.217575122140375e-06, 'learning_rate/action_model': 9.217575122140375e-05, 'epoch': 0.46}) +[2026-06-15 16:37:25,150][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.11384215950965881, 'train/grad_norm_pre_clip': 2.6877734661102295, 'timing/data': 0.011580153368413448, 'timing/model': 1.4551469469442964, 'learning_rate/qwen_vl_interface': 9.215899853774094e-06, 'learning_rate/action_model': 9.215899853774093e-05, 'epoch': 0.46}) +[2026-06-15 16:37:31,057][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.16084294021129608, 'train/grad_norm_pre_clip': 3.460482597351074, 'timing/data': 0.010859070345759392, 'timing/model': 1.4700757022947073, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.46}) +[2026-06-15 16:37:36,957][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.10992363840341568, 'train/grad_norm_pre_clip': 3.815659284591675, 'timing/data': 0.01098173949867487, 'timing/model': 1.4700909070670605, 'learning_rate/qwen_vl_interface': 9.212544427035967e-06, 'learning_rate/action_model': 9.212544427035967e-05, 'epoch': 0.46}) +[2026-06-15 16:37:42,850][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.11815384030342102, 'train/grad_norm_pre_clip': 3.911864757537842, 'timing/data': 0.010699658654630184, 'timing/model': 1.4293951503932476, 'learning_rate/qwen_vl_interface': 9.21086427004341e-06, 'learning_rate/action_model': 9.210864270043411e-05, 'epoch': 0.46}) +[2026-06-15 16:37:48,779][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.24553482234477997, 'train/grad_norm_pre_clip': 8.45523738861084, 'timing/data': 0.011505412869155407, 'timing/model': 1.6828892109915614, 'learning_rate/qwen_vl_interface': 9.209182484889168e-06, 'learning_rate/action_model': 9.209182484889167e-05, 'epoch': 0.46}) +[2026-06-15 16:37:54,638][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.2904569208621979, 'train/grad_norm_pre_clip': 11.055867195129395, 'timing/data': 0.011341411620378494, 'timing/model': 1.6576568186283112, 'learning_rate/qwen_vl_interface': 9.207499072264557e-06, 'learning_rate/action_model': 9.207499072264558e-05, 'epoch': 0.46}) +[2026-06-15 16:38:00,353][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.1958269476890564, 'train/grad_norm_pre_clip': 8.66907024383545, 'timing/data': 0.012766609899699688, 'timing/model': 1.461720259860158, 'learning_rate/qwen_vl_interface': 9.205814032861566e-06, 'learning_rate/action_model': 9.205814032861566e-05, 'epoch': 0.46}) +[2026-06-15 16:38:06,260][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.14835719764232635, 'train/grad_norm_pre_clip': 9.21894359588623, 'timing/data': 0.013512874022126198, 'timing/model': 1.4630985604599118, 'learning_rate/qwen_vl_interface': 9.204127367372849e-06, 'learning_rate/action_model': 9.204127367372849e-05, 'epoch': 0.46}) +[2026-06-15 16:38:12,265][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.20941615104675293, 'train/grad_norm_pre_clip': 11.142609596252441, 'timing/data': 0.01114836148917675, 'timing/model': 1.4885110054165125, 'learning_rate/qwen_vl_interface': 9.202439076491732e-06, 'learning_rate/action_model': 9.202439076491732e-05, 'epoch': 0.47}) +[2026-06-15 16:38:18,165][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.17680136859416962, 'train/grad_norm_pre_clip': 9.211974143981934, 'timing/data': 0.012895611114799976, 'timing/model': 1.4396797018125653, 'learning_rate/qwen_vl_interface': 9.200749160912208e-06, 'learning_rate/action_model': 9.200749160912208e-05, 'epoch': 0.47}) +[2026-06-15 16:38:24,120][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.06701413542032242, 'train/grad_norm_pre_clip': 5.287704944610596, 'timing/data': 0.010826370678842068, 'timing/model': 1.729112054221332, 'learning_rate/qwen_vl_interface': 9.199057621328937e-06, 'learning_rate/action_model': 9.199057621328935e-05, 'epoch': 0.47}) +[2026-06-15 16:38:30,007][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.5384491682052612, 'train/grad_norm_pre_clip': 46.221473693847656, 'timing/data': 0.011166251264512539, 'timing/model': 1.7114923130720854, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 0.47}) +[2026-06-15 16:38:35,662][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.197443425655365, 'train/grad_norm_pre_clip': 14.03083324432373, 'timing/data': 0.011859195306897163, 'timing/model': 1.4683103011921048, 'learning_rate/qwen_vl_interface': 9.19566967293313e-06, 'learning_rate/action_model': 9.19566967293313e-05, 'epoch': 0.47}) +[2026-06-15 16:38:41,561][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.12076608091592789, 'train/grad_norm_pre_clip': 10.30167007446289, 'timing/data': 0.012303987517952919, 'timing/model': 1.4605572512373328, 'learning_rate/qwen_vl_interface': 9.193973265513253e-06, 'learning_rate/action_model': 9.193973265513253e-05, 'epoch': 0.47}) +[2026-06-15 16:38:47,462][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.4453907012939453, 'train/grad_norm_pre_clip': 14.803862571716309, 'timing/data': 0.011739814653992653, 'timing/model': 1.4661873057484627, 'learning_rate/qwen_vl_interface': 9.192275236874946e-06, 'learning_rate/action_model': 9.192275236874946e-05, 'epoch': 0.47}) +[2026-06-15 16:38:53,461][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.31329742074012756, 'train/grad_norm_pre_clip': 13.770407676696777, 'timing/data': 0.011166230775415897, 'timing/model': 1.5117061706259847, 'learning_rate/qwen_vl_interface': 9.190575587716199e-06, 'learning_rate/action_model': 9.190575587716199e-05, 'epoch': 0.47}) +[2026-06-15 16:38:59,396][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.127326101064682, 'train/grad_norm_pre_clip': 3.8618764877319336, 'timing/data': 0.01131715252995491, 'timing/model': 1.7057678736746311, 'learning_rate/qwen_vl_interface': 9.188874318735677e-06, 'learning_rate/action_model': 9.188874318735676e-05, 'epoch': 0.47}) +[2026-06-15 16:39:05,063][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.16802890598773956, 'train/grad_norm_pre_clip': 11.145676612854004, 'timing/data': 0.012259388342499733, 'timing/model': 1.4683817904442549, 'learning_rate/qwen_vl_interface': 9.187171430632707e-06, 'learning_rate/action_model': 9.187171430632707e-05, 'epoch': 0.47}) +[2026-06-15 16:39:10,963][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.3445214331150055, 'train/grad_norm_pre_clip': 21.716279983520508, 'timing/data': 0.012221727520227432, 'timing/model': 1.4651416670531034, 'learning_rate/qwen_vl_interface': 9.185466924107283e-06, 'learning_rate/action_model': 9.185466924107282e-05, 'epoch': 0.47}) +[2026-06-15 16:39:16,861][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.13599352538585663, 'train/grad_norm_pre_clip': 6.163059234619141, 'timing/data': 0.011792515404522419, 'timing/model': 1.458804895170033, 'learning_rate/qwen_vl_interface': 9.18376079986006e-06, 'learning_rate/action_model': 9.183760799860061e-05, 'epoch': 0.47}) +[2026-06-15 16:39:22,760][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.7807599306106567, 'train/grad_norm_pre_clip': 25.075849533081055, 'timing/data': 0.011534303426742554, 'timing/model': 1.4347218163311481, 'learning_rate/qwen_vl_interface': 9.182053058592365e-06, 'learning_rate/action_model': 9.182053058592365e-05, 'epoch': 0.47}) +[2026-06-15 16:39:28,666][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.15857034921646118, 'train/grad_norm_pre_clip': 7.024607181549072, 'timing/data': 0.011850995011627674, 'timing/model': 1.4893966214731336, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 0.47}) +[2026-06-15 16:39:34,558][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.2097160369157791, 'train/grad_norm_pre_clip': 5.915584564208984, 'timing/data': 0.01419798843562603, 'timing/model': 1.6601810678839684, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.47}) +[2026-06-15 16:39:40,255][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.2713545262813568, 'train/grad_norm_pre_clip': 9.108922958374023, 'timing/data': 0.011429423466324806, 'timing/model': 1.4616668662056327, 'learning_rate/qwen_vl_interface': 9.176920139689638e-06, 'learning_rate/action_model': 9.176920139689637e-05, 'epoch': 0.47}) +[2026-06-15 16:39:46,154][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.2451326549053192, 'train/grad_norm_pre_clip': 7.122644424438477, 'timing/data': 0.011954725719988346, 'timing/model': 1.4583394704386592, 'learning_rate/qwen_vl_interface': 9.175205937366568e-06, 'learning_rate/action_model': 9.175205937366569e-05, 'epoch': 0.47}) +[2026-06-15 16:39:52,160][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.21729375422000885, 'train/grad_norm_pre_clip': 7.177030086517334, 'timing/data': 0.011808345094323158, 'timing/model': 1.4734755782410502, 'learning_rate/qwen_vl_interface': 9.173490121539608e-06, 'learning_rate/action_model': 9.173490121539608e-05, 'epoch': 0.47}) +[2026-06-15 16:39:58,054][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.14488793909549713, 'train/grad_norm_pre_clip': 6.31863260269165, 'timing/data': 0.010734870098531246, 'timing/model': 1.430631134659052, 'learning_rate/qwen_vl_interface': 9.171772692914062e-06, 'learning_rate/action_model': 9.171772692914061e-05, 'epoch': 0.47}) +[2026-06-15 16:40:03,952][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.25289344787597656, 'train/grad_norm_pre_clip': 7.538784503936768, 'timing/data': 0.010892610065639019, 'timing/model': 1.426114252768457, 'learning_rate/qwen_vl_interface': 9.1700536521959e-06, 'learning_rate/action_model': 9.170053652195899e-05, 'epoch': 0.47}) +[2026-06-15 16:40:09,857][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.4789884388446808, 'train/grad_norm_pre_clip': 13.117082595825195, 'timing/data': 0.010715258307754993, 'timing/model': 1.660983595997095, 'learning_rate/qwen_vl_interface': 9.168333000091754e-06, 'learning_rate/action_model': 9.168333000091754e-05, 'epoch': 0.47}) +[2026-06-15 16:40:15,552][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.25753796100616455, 'train/grad_norm_pre_clip': 6.629264831542969, 'timing/data': 0.010875529609620571, 'timing/model': 1.458521799184382, 'learning_rate/qwen_vl_interface': 9.16661073730892e-06, 'learning_rate/action_model': 9.166610737308919e-05, 'epoch': 0.47}) +[2026-06-15 16:40:21,454][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.12154044955968857, 'train/grad_norm_pre_clip': 4.493843078613281, 'timing/data': 0.010794079862535, 'timing/model': 1.4597677988931537, 'learning_rate/qwen_vl_interface': 9.164886864555355e-06, 'learning_rate/action_model': 9.164886864555355e-05, 'epoch': 0.48}) +[2026-06-15 16:40:27,353][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.16272710263729095, 'train/grad_norm_pre_clip': 3.321138858795166, 'timing/data': 0.010855740867555141, 'timing/model': 1.4570126365870237, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 0.48}) +[2026-06-15 16:40:33,256][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.21311578154563904, 'train/grad_norm_pre_clip': 6.506008625030518, 'timing/data': 0.01066942885518074, 'timing/model': 1.429953490383923, 'learning_rate/qwen_vl_interface': 9.16143429197117e-06, 'learning_rate/action_model': 9.16143429197117e-05, 'epoch': 0.48}) +[2026-06-15 16:40:39,184][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.09730284661054611, 'train/grad_norm_pre_clip': 1.4112213850021362, 'timing/data': 0.010568789206445217, 'timing/model': 1.693213283084333, 'learning_rate/qwen_vl_interface': 9.159705593559773e-06, 'learning_rate/action_model': 9.159705593559773e-05, 'epoch': 0.48}) +[2026-06-15 16:40:44,851][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.2197730839252472, 'train/grad_norm_pre_clip': 4.237448692321777, 'timing/data': 0.010704399086534977, 'timing/model': 1.460166904143989, 'learning_rate/qwen_vl_interface': 9.157975288016089e-06, 'learning_rate/action_model': 9.157975288016088e-05, 'epoch': 0.48}) +[2026-06-15 16:40:50,754][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.25765615701675415, 'train/grad_norm_pre_clip': 4.9167985916137695, 'timing/data': 0.011646035127341747, 'timing/model': 1.4567889478057623, 'learning_rate/qwen_vl_interface': 9.156243376051381e-06, 'learning_rate/action_model': 9.156243376051381e-05, 'epoch': 0.48}) +[2026-06-15 16:40:56,652][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.13345833122730255, 'train/grad_norm_pre_clip': 3.7922885417938232, 'timing/data': 0.010953710414469242, 'timing/model': 1.4542081961408257, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.48}) +[2026-06-15 16:41:02,552][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.25154608488082886, 'train/grad_norm_pre_clip': 7.0612568855285645, 'timing/data': 0.01067453995347023, 'timing/model': 1.45162205491215, 'learning_rate/qwen_vl_interface': 9.152774735707248e-06, 'learning_rate/action_model': 9.152774735707247e-05, 'epoch': 0.48}) +[2026-06-15 16:41:08,451][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.15603435039520264, 'train/grad_norm_pre_clip': 4.7869954109191895, 'timing/data': 0.010823650285601616, 'timing/model': 1.4646431813016534, 'learning_rate/qwen_vl_interface': 9.15103800875365e-06, 'learning_rate/action_model': 9.15103800875365e-05, 'epoch': 0.48}) +[2026-06-15 16:41:14,388][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.09683942049741745, 'train/grad_norm_pre_clip': 2.4523048400878906, 'timing/data': 0.010588539764285088, 'timing/model': 1.6932312240824103, 'learning_rate/qwen_vl_interface': 9.149299678230681e-06, 'learning_rate/action_model': 9.149299678230681e-05, 'epoch': 0.48}) +[2026-06-15 16:41:20,054][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.20810478925704956, 'train/grad_norm_pre_clip': 6.24208402633667, 'timing/data': 0.010472447611391544, 'timing/model': 1.4638671223074198, 'learning_rate/qwen_vl_interface': 9.147559744852905e-06, 'learning_rate/action_model': 9.147559744852905e-05, 'epoch': 0.48}) +[2026-06-15 16:41:25,957][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.05168437957763672, 'train/grad_norm_pre_clip': 2.335421085357666, 'timing/data': 0.011434623971581459, 'timing/model': 1.4600699041038752, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 0.48}) +[2026-06-15 16:41:31,855][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.5324741005897522, 'train/grad_norm_pre_clip': 13.778620719909668, 'timing/data': 0.010514979250729084, 'timing/model': 1.4576547825708985, 'learning_rate/qwen_vl_interface': 9.14407507239447e-06, 'learning_rate/action_model': 9.14407507239447e-05, 'epoch': 0.48}) +[2026-06-15 16:41:37,862][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.4749026894569397, 'train/grad_norm_pre_clip': 11.895268440246582, 'timing/data': 0.012802991084754467, 'timing/model': 1.5132987666875124, 'learning_rate/qwen_vl_interface': 9.142330334746227e-06, 'learning_rate/action_model': 9.142330334746227e-05, 'epoch': 0.48}) +[2026-06-15 16:41:43,752][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.35588085651397705, 'train/grad_norm_pre_clip': 13.542259216308594, 'timing/data': 0.010842761024832726, 'timing/model': 1.4508759705349803, 'learning_rate/qwen_vl_interface': 9.14058399710801e-06, 'learning_rate/action_model': 9.14058399710801e-05, 'epoch': 0.48}) +[2026-06-15 16:41:49,698][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.36352285742759705, 'train/grad_norm_pre_clip': 9.840827941894531, 'timing/data': 0.01222061738371849, 'timing/model': 1.69761145953089, 'learning_rate/qwen_vl_interface': 9.138836060197673e-06, 'learning_rate/action_model': 9.138836060197672e-05, 'epoch': 0.48}) +[2026-06-15 16:41:55,353][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.13098566234111786, 'train/grad_norm_pre_clip': 2.2107882499694824, 'timing/data': 0.010896410793066025, 'timing/model': 1.4596401453018188, 'learning_rate/qwen_vl_interface': 9.137086524733725e-06, 'learning_rate/action_model': 9.137086524733724e-05, 'epoch': 0.48}) +[2026-06-15 16:42:01,259][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.30106693506240845, 'train/grad_norm_pre_clip': 9.284346580505371, 'timing/data': 0.011136611923575401, 'timing/model': 1.4671061299741268, 'learning_rate/qwen_vl_interface': 9.135335391435336e-06, 'learning_rate/action_model': 9.135335391435335e-05, 'epoch': 0.48}) +[2026-06-15 16:42:07,154][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.25675132870674133, 'train/grad_norm_pre_clip': 6.923152446746826, 'timing/data': 0.01091156154870987, 'timing/model': 1.455491067841649, 'learning_rate/qwen_vl_interface': 9.13358266102233e-06, 'learning_rate/action_model': 9.133582661022328e-05, 'epoch': 0.48}) +[2026-06-15 16:42:13,054][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.39903610944747925, 'train/grad_norm_pre_clip': 12.902602195739746, 'timing/data': 0.011234642937779427, 'timing/model': 1.4321744302287698, 'learning_rate/qwen_vl_interface': 9.131828334215185e-06, 'learning_rate/action_model': 9.131828334215185e-05, 'epoch': 0.48}) +[2026-06-15 16:42:18,961][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.09751462191343307, 'train/grad_norm_pre_clip': 5.352901458740234, 'timing/data': 0.010583048686385155, 'timing/model': 1.428233390673995, 'learning_rate/qwen_vl_interface': 9.130072411735043e-06, 'learning_rate/action_model': 9.130072411735043e-05, 'epoch': 0.48}) +[2026-06-15 16:42:24,892][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.30685165524482727, 'train/grad_norm_pre_clip': 9.024765014648438, 'timing/data': 0.011093482375144958, 'timing/model': 1.695511495694518, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 0.48}) +[2026-06-15 16:42:30,553][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.32723620533943176, 'train/grad_norm_pre_clip': 9.644210815429688, 'timing/data': 0.011156031861901283, 'timing/model': 1.4585241749882698, 'learning_rate/qwen_vl_interface': 9.126555782643592e-06, 'learning_rate/action_model': 9.126555782643591e-05, 'epoch': 0.49}) +[2026-06-15 16:42:36,753][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.1705733835697174, 'train/grad_norm_pre_clip': 4.706624984741211, 'timing/data': 0.011871015653014183, 'timing/model': 1.4560206541791558, 'learning_rate/qwen_vl_interface': 9.124795077477835e-06, 'learning_rate/action_model': 9.124795077477834e-05, 'epoch': 0.49}) +[2026-06-15 16:42:42,649][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.12815314531326294, 'train/grad_norm_pre_clip': 3.848479747772217, 'timing/data': 0.011715135537087917, 'timing/model': 1.4614183465018868, 'learning_rate/qwen_vl_interface': 9.123032779530186e-06, 'learning_rate/action_model': 9.123032779530187e-05, 'epoch': 0.49}) +[2026-06-15 16:42:48,551][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.173612579703331, 'train/grad_norm_pre_clip': 7.16746187210083, 'timing/data': 0.011496754363179207, 'timing/model': 1.4563196199014783, 'learning_rate/qwen_vl_interface': 9.12126888952506e-06, 'learning_rate/action_model': 9.12126888952506e-05, 'epoch': 0.49}) +[2026-06-15 16:42:54,458][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.15562547743320465, 'train/grad_norm_pre_clip': 7.106154441833496, 'timing/data': 0.011888107284903526, 'timing/model': 1.4285285975784063, 'learning_rate/qwen_vl_interface': 9.119503408187521e-06, 'learning_rate/action_model': 9.119503408187522e-05, 'epoch': 0.49}) +[2026-06-15 16:43:00,380][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.15083710849285126, 'train/grad_norm_pre_clip': 3.6095376014709473, 'timing/data': 0.011171442456543446, 'timing/model': 1.6871118554845452, 'learning_rate/qwen_vl_interface': 9.117736336243299e-06, 'learning_rate/action_model': 9.117736336243298e-05, 'epoch': 0.49}) +[2026-06-15 16:43:06,052][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.13252152502536774, 'train/grad_norm_pre_clip': 3.8062288761138916, 'timing/data': 0.011672764085233212, 'timing/model': 1.4610493434593081, 'learning_rate/qwen_vl_interface': 9.115967674418765e-06, 'learning_rate/action_model': 9.115967674418765e-05, 'epoch': 0.49}) +[2026-06-15 16:43:11,950][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.11027954518795013, 'train/grad_norm_pre_clip': 3.2615137100219727, 'timing/data': 0.01145139429718256, 'timing/model': 1.457521996460855, 'learning_rate/qwen_vl_interface': 9.114197423440949e-06, 'learning_rate/action_model': 9.11419742344095e-05, 'epoch': 0.49}) +[2026-06-15 16:43:17,850][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.18705284595489502, 'train/grad_norm_pre_clip': 5.287816047668457, 'timing/data': 0.010322467423975468, 'timing/model': 1.453458646312356, 'learning_rate/qwen_vl_interface': 9.112425584037536e-06, 'learning_rate/action_model': 9.112425584037536e-05, 'epoch': 0.49}) +[2026-06-15 16:43:23,751][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.11681747436523438, 'train/grad_norm_pre_clip': 3.966590642929077, 'timing/data': 0.010757550597190857, 'timing/model': 1.4606721280142665, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 0.49}) +[2026-06-15 16:43:29,655][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.28734707832336426, 'train/grad_norm_pre_clip': 7.470483303070068, 'timing/data': 0.010974140837788582, 'timing/model': 1.4284186707809567, 'learning_rate/qwen_vl_interface': 9.108877142867918e-06, 'learning_rate/action_model': 9.108877142867918e-05, 'epoch': 0.49}) +[2026-06-15 16:43:35,584][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.16498857736587524, 'train/grad_norm_pre_clip': 6.224459648132324, 'timing/data': 0.011152793653309345, 'timing/model': 1.696343533694744, 'learning_rate/qwen_vl_interface': 9.107100542560342e-06, 'learning_rate/action_model': 9.107100542560342e-05, 'epoch': 0.49}) +[2026-06-15 16:43:41,349][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.14652517437934875, 'train/grad_norm_pre_clip': 4.043814659118652, 'timing/data': 0.011422674171626568, 'timing/model': 1.459657410159707, 'learning_rate/qwen_vl_interface': 9.105322356744429e-06, 'learning_rate/action_model': 9.105322356744429e-05, 'epoch': 0.49}) +[2026-06-15 16:43:47,252][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.3104579448699951, 'train/grad_norm_pre_clip': 8.896282196044922, 'timing/data': 0.010797189548611641, 'timing/model': 1.456752896308899, 'learning_rate/qwen_vl_interface': 9.103542586151121e-06, 'learning_rate/action_model': 9.10354258615112e-05, 'epoch': 0.49}) +[2026-06-15 16:43:53,149][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.20958742499351501, 'train/grad_norm_pre_clip': 7.755898952484131, 'timing/data': 0.0109028909355402, 'timing/model': 1.463879818096757, 'learning_rate/qwen_vl_interface': 9.101761231512017e-06, 'learning_rate/action_model': 9.101761231512017e-05, 'epoch': 0.49}) +[2026-06-15 16:43:59,048][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.32818618416786194, 'train/grad_norm_pre_clip': 7.158686637878418, 'timing/data': 0.010515319183468819, 'timing/model': 1.463698179461062, 'learning_rate/qwen_vl_interface': 9.099978293559365e-06, 'learning_rate/action_model': 9.099978293559366e-05, 'epoch': 0.49}) +[2026-06-15 16:44:04,952][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.22680963575839996, 'train/grad_norm_pre_clip': 5.860376358032227, 'timing/data': 0.010501799173653126, 'timing/model': 1.4292072374373674, 'learning_rate/qwen_vl_interface': 9.098193773026065e-06, 'learning_rate/action_model': 9.098193773026064e-05, 'epoch': 0.49}) +[2026-06-15 16:44:10,886][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.23066240549087524, 'train/grad_norm_pre_clip': 6.727982521057129, 'timing/data': 0.011068042367696762, 'timing/model': 1.7013971572741866, 'learning_rate/qwen_vl_interface': 9.096407670645663e-06, 'learning_rate/action_model': 9.096407670645662e-05, 'epoch': 0.49}) +[2026-06-15 16:44:16,550][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.13138961791992188, 'train/grad_norm_pre_clip': 8.537060737609863, 'timing/data': 0.010647770017385483, 'timing/model': 1.4585067266598344, 'learning_rate/qwen_vl_interface': 9.094619987152357e-06, 'learning_rate/action_model': 9.094619987152358e-05, 'epoch': 0.49}) +[2026-06-15 16:44:22,451][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.08691902458667755, 'train/grad_norm_pre_clip': 7.934643745422363, 'timing/data': 0.010353118181228638, 'timing/model': 1.4657089374959469, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.49}) +[2026-06-15 16:44:28,351][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.26949194073677063, 'train/grad_norm_pre_clip': 6.6910858154296875, 'timing/data': 0.010157807730138302, 'timing/model': 1.4552400913089514, 'learning_rate/qwen_vl_interface': 9.09103987976709e-06, 'learning_rate/action_model': 9.09103987976709e-05, 'epoch': 0.49}) +[2026-06-15 16:44:34,254][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.20715655386447906, 'train/grad_norm_pre_clip': 7.599908828735352, 'timing/data': 0.010572468861937523, 'timing/model': 1.459178276360035, 'learning_rate/qwen_vl_interface': 9.089247457346774e-06, 'learning_rate/action_model': 9.089247457346773e-05, 'epoch': 0.49}) +[2026-06-15 16:44:40,220][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.3134402334690094, 'train/grad_norm_pre_clip': 7.740145206451416, 'timing/data': 0.012195867486298084, 'timing/model': 1.4867145828902721, 'learning_rate/qwen_vl_interface': 9.087453456756849e-06, 'learning_rate/action_model': 9.087453456756849e-05, 'epoch': 0.5}) +[2026-06-15 16:44:46,095][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.2717801630496979, 'train/grad_norm_pre_clip': 6.7420172691345215, 'timing/data': 0.011264472268521786, 'timing/model': 1.6971026277169585, 'learning_rate/qwen_vl_interface': 9.085657878734759e-06, 'learning_rate/action_model': 9.085657878734758e-05, 'epoch': 0.5}) +[2026-06-15 16:44:51,931][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.07135042548179626, 'train/grad_norm_pre_clip': 2.6269009113311768, 'timing/data': 0.010358057916164398, 'timing/model': 1.6347458185628057, 'learning_rate/qwen_vl_interface': 9.083860724018603e-06, 'learning_rate/action_model': 9.083860724018602e-05, 'epoch': 0.5}) +[2026-06-15 16:44:57,655][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.23294197022914886, 'train/grad_norm_pre_clip': 4.375984191894531, 'timing/data': 0.010166207328438759, 'timing/model': 1.460283450782299, 'learning_rate/qwen_vl_interface': 9.08206199334712e-06, 'learning_rate/action_model': 9.082061993347119e-05, 'epoch': 0.5}) +[2026-06-15 16:45:03,548][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.14087563753128052, 'train/grad_norm_pre_clip': 7.86421537399292, 'timing/data': 0.011292334645986557, 'timing/model': 1.4501870144158602, 'learning_rate/qwen_vl_interface': 9.0802616874597e-06, 'learning_rate/action_model': 9.0802616874597e-05, 'epoch': 0.5}) +[2026-06-15 16:45:09,454][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.11176140606403351, 'train/grad_norm_pre_clip': 3.1497883796691895, 'timing/data': 0.010713140480220318, 'timing/model': 1.4478489439934492, 'learning_rate/qwen_vl_interface': 9.078459807096385e-06, 'learning_rate/action_model': 9.078459807096385e-05, 'epoch': 0.5}) +[2026-06-15 16:45:15,363][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.2845161259174347, 'train/grad_norm_pre_clip': 6.920024871826172, 'timing/data': 0.011523334309458733, 'timing/model': 1.4285192964598536, 'learning_rate/qwen_vl_interface': 9.076656352997854e-06, 'learning_rate/action_model': 9.076656352997853e-05, 'epoch': 0.5}) +[2026-06-15 16:45:21,290][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.21840117871761322, 'train/grad_norm_pre_clip': 8.523483276367188, 'timing/data': 0.011601285077631474, 'timing/model': 1.6728529846295714, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 0.5}) +[2026-06-15 16:45:27,212][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.11823596060276031, 'train/grad_norm_pre_clip': 3.5491256713867188, 'timing/data': 0.011288204230368137, 'timing/model': 1.7097139032557607, 'learning_rate/qwen_vl_interface': 9.073044726561128e-06, 'learning_rate/action_model': 9.073044726561129e-05, 'epoch': 0.5}) +[2026-06-15 16:45:32,864][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.15745103359222412, 'train/grad_norm_pre_clip': 4.925204753875732, 'timing/data': 0.011408433318138123, 'timing/model': 1.4657605588436127, 'learning_rate/qwen_vl_interface': 9.071236555707536e-06, 'learning_rate/action_model': 9.071236555707537e-05, 'epoch': 0.5}) +[2026-06-15 16:45:38,763][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.09064513444900513, 'train/grad_norm_pre_clip': 3.595777988433838, 'timing/data': 0.011599495075643063, 'timing/model': 1.4666837388649583, 'learning_rate/qwen_vl_interface': 9.069426814087937e-06, 'learning_rate/action_model': 9.069426814087937e-05, 'epoch': 0.5}) +[2026-06-15 16:45:44,768][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.10952728986740112, 'train/grad_norm_pre_clip': 8.733072280883789, 'timing/data': 0.011527775786817074, 'timing/model': 1.4915032805874944, 'learning_rate/qwen_vl_interface': 9.067615502446247e-06, 'learning_rate/action_model': 9.067615502446246e-05, 'epoch': 0.5}) +[2026-06-15 16:45:50,670][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.48405641317367554, 'train/grad_norm_pre_clip': 47.55495834350586, 'timing/data': 0.01197066716849804, 'timing/model': 1.440139813348651, 'learning_rate/qwen_vl_interface': 9.06580262152703e-06, 'learning_rate/action_model': 9.065802621527028e-05, 'epoch': 0.5}) +[2026-06-15 16:45:56,603][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.8173383474349976, 'train/grad_norm_pre_clip': 22.552452087402344, 'timing/data': 0.011826816946268082, 'timing/model': 1.7032499257475138, 'learning_rate/qwen_vl_interface': 9.063988172075486e-06, 'learning_rate/action_model': 9.063988172075486e-05, 'epoch': 0.5}) +[2026-06-15 16:46:02,263][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.43653979897499084, 'train/grad_norm_pre_clip': 18.539640426635742, 'timing/data': 0.011083822697401047, 'timing/model': 1.468295393511653, 'learning_rate/qwen_vl_interface': 9.062172154837475e-06, 'learning_rate/action_model': 9.062172154837475e-05, 'epoch': 0.5}) +[2026-06-15 16:46:08,258][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.20183947682380676, 'train/grad_norm_pre_clip': 8.121901512145996, 'timing/data': 0.010720379650592804, 'timing/model': 1.4616944696754217, 'learning_rate/qwen_vl_interface': 9.060354570559487e-06, 'learning_rate/action_model': 9.060354570559487e-05, 'epoch': 0.5}) +[2026-06-15 16:46:14,155][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.4311733841896057, 'train/grad_norm_pre_clip': 16.309877395629883, 'timing/data': 0.01098536141216755, 'timing/model': 1.4647698979824781, 'learning_rate/qwen_vl_interface': 9.058535419988667e-06, 'learning_rate/action_model': 9.058535419988667e-05, 'epoch': 0.5}) +[2026-06-15 16:56:47,036][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.1966063678264618, 'train/grad_norm_pre_clip': 6.088724136352539, 'eval/loss': 0.09896062314510345, 'eval/flap_precision': 0.75, 'eval/flap_recall': 0.6763636363636364, 'eval/flap_f1': 0.711281070745698, 'eval/flap_rate_model': 0.06888888888888889, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.711281070745698, 'timing/data': 0.01064987014979124, 'timing/model': 1.4367687124758959, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 0.5}) +[2026-06-15 16:56:47,037][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1100_state +[2026-06-15 16:56:47,044][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 16:56:55,150][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1100_state/model.safetensors +[2026-06-15 16:57:06,260][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1100_state/optimizer.bin +[2026-06-15 16:57:06,260][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1100_state/sampler.bin +[2026-06-15 16:57:06,261][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1100_state/sampler_1.bin +[2026-06-15 16:57:06,262][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1100_state/random_states_0.pkl +[2026-06-15 16:57:06,422][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-06-15 16:57:06,422][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 16:57:06,437][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 16:57:12,432][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.1379184126853943, 'train/grad_norm_pre_clip': 8.923004150390625, 'timing/data': 0.011095774360001087, 'timing/model': 1.476181723177433, 'learning_rate/qwen_vl_interface': 9.0548924229603e-06, 'learning_rate/action_model': 9.054892422960301e-05, 'epoch': 0.5}) +[2026-06-15 16:57:18,337][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.30228370428085327, 'train/grad_norm_pre_clip': 8.273468971252441, 'timing/data': 0.010934814810752869, 'timing/model': 1.4402754930779338, 'learning_rate/qwen_vl_interface': 9.053068578000256e-06, 'learning_rate/action_model': 9.053068578000256e-05, 'epoch': 0.5}) +[2026-06-15 16:57:24,272][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.3202190697193146, 'train/grad_norm_pre_clip': 5.996060848236084, 'timing/data': 0.011366386897861958, 'timing/model': 1.6998320566490293, 'learning_rate/qwen_vl_interface': 9.051243169742372e-06, 'learning_rate/action_model': 9.051243169742372e-05, 'epoch': 0.5}) +[2026-06-15 16:57:29,943][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.20943407714366913, 'train/grad_norm_pre_clip': 5.815914154052734, 'timing/data': 0.011134776286780834, 'timing/model': 1.4638352524489164, 'learning_rate/qwen_vl_interface': 9.049416198937008e-06, 'learning_rate/action_model': 9.049416198937007e-05, 'epoch': 0.5}) +[2026-06-15 16:57:35,855][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.12840278446674347, 'train/grad_norm_pre_clip': 3.485170602798462, 'timing/data': 0.01186840794980526, 'timing/model': 1.4793457938358188, 'learning_rate/qwen_vl_interface': 9.047587666335159e-06, 'learning_rate/action_model': 9.047587666335159e-05, 'epoch': 0.51}) +[2026-06-15 16:57:41,831][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.09311085194349289, 'train/grad_norm_pre_clip': 3.182617425918579, 'timing/data': 0.011263827793300152, 'timing/model': 1.5425102915614843, 'learning_rate/qwen_vl_interface': 9.045757572688469e-06, 'learning_rate/action_model': 9.04575757268847e-05, 'epoch': 0.51}) +[2026-06-15 16:57:47,752][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.1534394770860672, 'train/grad_norm_pre_clip': 3.300992727279663, 'timing/data': 0.01153078768402338, 'timing/model': 1.455157863907516, 'learning_rate/qwen_vl_interface': 9.043925918749218e-06, 'learning_rate/action_model': 9.043925918749218e-05, 'epoch': 0.51}) +[2026-06-15 16:57:53,643][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.3779812753200531, 'train/grad_norm_pre_clip': 9.348969459533691, 'timing/data': 0.0113680986687541, 'timing/model': 1.4447385435923934, 'learning_rate/qwen_vl_interface': 9.042092705270333e-06, 'learning_rate/action_model': 9.042092705270333e-05, 'epoch': 0.51}) +[2026-06-15 16:57:59,570][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.24302655458450317, 'train/grad_norm_pre_clip': 9.553462028503418, 'timing/data': 0.011541789397597313, 'timing/model': 1.6658697631210089, 'learning_rate/qwen_vl_interface': 9.040257933005375e-06, 'learning_rate/action_model': 9.040257933005374e-05, 'epoch': 0.51}) +[2026-06-15 16:58:05,253][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.24578174948692322, 'train/grad_norm_pre_clip': 7.736261367797852, 'timing/data': 0.013558272272348404, 'timing/model': 1.4636554550379515, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 0.51}) +[2026-06-15 16:58:11,152][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.1711992770433426, 'train/grad_norm_pre_clip': 4.410549163818359, 'timing/data': 0.011778689920902252, 'timing/model': 1.4550680546090007, 'learning_rate/qwen_vl_interface': 9.036583715134709e-06, 'learning_rate/action_model': 9.036583715134708e-05, 'epoch': 0.51}) +[2026-06-15 16:58:17,051][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.1149645745754242, 'train/grad_norm_pre_clip': 47.754539489746094, 'timing/data': 0.012112121097743511, 'timing/model': 1.452330174855888, 'learning_rate/qwen_vl_interface': 9.034744271039332e-06, 'learning_rate/action_model': 9.034744271039332e-05, 'epoch': 0.51}) +[2026-06-15 16:58:22,951][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.2707718312740326, 'train/grad_norm_pre_clip': 8.20954418182373, 'timing/data': 0.010854138992726803, 'timing/model': 1.4361399011686444, 'learning_rate/qwen_vl_interface': 9.03290327117855e-06, 'learning_rate/action_model': 9.032903271178549e-05, 'epoch': 0.51}) +[2026-06-15 16:58:28,843][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.12791885435581207, 'train/grad_norm_pre_clip': 5.873025894165039, 'timing/data': 0.01047859899699688, 'timing/model': 1.4587962348014116, 'learning_rate/qwen_vl_interface': 9.031060716309122e-06, 'learning_rate/action_model': 9.031060716309121e-05, 'epoch': 0.51}) +[2026-06-15 16:58:34,785][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.25919362902641296, 'train/grad_norm_pre_clip': 6.767683982849121, 'timing/data': 0.010915489867329597, 'timing/model': 1.6959096305072308, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.51}) +[2026-06-15 16:58:40,451][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.16478951275348663, 'train/grad_norm_pre_clip': 5.9065022468566895, 'timing/data': 0.01126518938690424, 'timing/model': 1.4582577990368009, 'learning_rate/qwen_vl_interface': 9.027370944574604e-06, 'learning_rate/action_model': 9.027370944574604e-05, 'epoch': 0.51}) +[2026-06-15 16:58:46,352][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.13553349673748016, 'train/grad_norm_pre_clip': 5.750978469848633, 'timing/data': 0.01368261594325304, 'timing/model': 1.452868640422821, 'learning_rate/qwen_vl_interface': 9.025523729226236e-06, 'learning_rate/action_model': 9.025523729226236e-05, 'epoch': 0.51}) +[2026-06-15 16:58:52,244][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.3281525671482086, 'train/grad_norm_pre_clip': 6.595257759094238, 'timing/data': 0.0110637117177248, 'timing/model': 1.4579649111256003, 'learning_rate/qwen_vl_interface': 9.023674961902677e-06, 'learning_rate/action_model': 9.023674961902676e-05, 'epoch': 0.51}) +[2026-06-15 16:58:58,153][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.11882688850164413, 'train/grad_norm_pre_clip': 4.573257923126221, 'timing/data': 0.012298024259507656, 'timing/model': 1.4572469107806683, 'learning_rate/qwen_vl_interface': 9.021824643363883e-06, 'learning_rate/action_model': 9.021824643363883e-05, 'epoch': 0.51}) +[2026-06-15 16:59:04,057][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.22056527435779572, 'train/grad_norm_pre_clip': 10.489745140075684, 'timing/data': 0.010772290639579296, 'timing/model': 1.4632871830835938, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 0.51}) +[2026-06-15 16:59:09,995][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.36232641339302063, 'train/grad_norm_pre_clip': 11.509686470031738, 'timing/data': 0.010213450528681278, 'timing/model': 1.7023376291617751, 'learning_rate/qwen_vl_interface': 9.018119355683623e-06, 'learning_rate/action_model': 9.018119355683623e-05, 'epoch': 0.51}) +[2026-06-15 16:59:15,657][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.11809620261192322, 'train/grad_norm_pre_clip': 3.4881398677825928, 'timing/data': 0.010331820696592331, 'timing/model': 1.4615375306457281, 'learning_rate/qwen_vl_interface': 9.016264388065258e-06, 'learning_rate/action_model': 9.016264388065258e-05, 'epoch': 0.51}) +[2026-06-15 16:59:21,558][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.059932176023721695, 'train/grad_norm_pre_clip': 3.2742652893066406, 'timing/data': 0.010823562741279602, 'timing/model': 1.464550094678998, 'learning_rate/qwen_vl_interface': 9.014407872277869e-06, 'learning_rate/action_model': 9.014407872277869e-05, 'epoch': 0.51}) +[2026-06-15 16:59:27,457][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.22284160554409027, 'train/grad_norm_pre_clip': 6.235446453094482, 'timing/data': 0.011030622757971287, 'timing/model': 1.4590904023498297, 'learning_rate/qwen_vl_interface': 9.012549809084595e-06, 'learning_rate/action_model': 9.012549809084595e-05, 'epoch': 0.51}) +[2026-06-15 16:59:33,355][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.12364773452281952, 'train/grad_norm_pre_clip': 4.2063093185424805, 'timing/data': 0.010878022760152817, 'timing/model': 1.4310496523976326, 'learning_rate/qwen_vl_interface': 9.01069019924922e-06, 'learning_rate/action_model': 9.010690199249219e-05, 'epoch': 0.51}) +[2026-06-15 16:59:39,259][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.20363706350326538, 'train/grad_norm_pre_clip': 9.851511001586914, 'timing/data': 0.011192924343049526, 'timing/model': 1.4436667459085584, 'learning_rate/qwen_vl_interface': 9.008829043536157e-06, 'learning_rate/action_model': 9.008829043536156e-05, 'epoch': 0.51}) +[2026-06-15 16:59:45,188][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.1420675367116928, 'train/grad_norm_pre_clip': 7.57116174697876, 'timing/data': 0.01037647295743227, 'timing/model': 1.6926210187375546, 'learning_rate/qwen_vl_interface': 9.006966342710458e-06, 'learning_rate/action_model': 9.006966342710457e-05, 'epoch': 0.52}) +[2026-06-15 16:59:50,859][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.23603446781635284, 'train/grad_norm_pre_clip': 10.761642456054688, 'timing/data': 0.010902403853833675, 'timing/model': 1.464258562773466, 'learning_rate/qwen_vl_interface': 9.005102097537809e-06, 'learning_rate/action_model': 9.005102097537809e-05, 'epoch': 0.52}) +[2026-06-15 16:59:56,762][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.2915753126144409, 'train/grad_norm_pre_clip': 10.7933931350708, 'timing/data': 0.011734035797417164, 'timing/model': 1.4679435594007373, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.52}) +[2026-06-15 17:00:02,648][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.29946258664131165, 'train/grad_norm_pre_clip': 9.206230163574219, 'timing/data': 0.01069918368011713, 'timing/model': 1.4642152870073915, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 0.52}) +[2026-06-15 17:00:08,562][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.08778568357229233, 'train/grad_norm_pre_clip': 3.461148500442505, 'timing/data': 0.010873843915760517, 'timing/model': 1.4755005491897464, 'learning_rate/qwen_vl_interface': 8.999500103604538e-06, 'learning_rate/action_model': 8.999500103604536e-05, 'epoch': 0.52}) +[2026-06-15 17:00:14,550][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.07044880092144012, 'train/grad_norm_pre_clip': 6.634187698364258, 'timing/data': 0.010623045265674591, 'timing/model': 1.5087776016443968, 'learning_rate/qwen_vl_interface': 8.997629688713639e-06, 'learning_rate/action_model': 8.997629688713638e-05, 'epoch': 0.52}) +[2026-06-15 17:00:20,495][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.29067155718803406, 'train/grad_norm_pre_clip': 8.280515670776367, 'timing/data': 0.01184194628149271, 'timing/model': 1.694572115316987, 'learning_rate/qwen_vl_interface': 8.995757733313733e-06, 'learning_rate/action_model': 8.995757733313732e-05, 'epoch': 0.52}) +[2026-06-15 17:00:26,163][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.1704600304365158, 'train/grad_norm_pre_clip': 4.525956630706787, 'timing/data': 0.011260655708611012, 'timing/model': 1.4718559868633747, 'learning_rate/qwen_vl_interface': 8.993884238174314e-06, 'learning_rate/action_model': 8.993884238174314e-05, 'epoch': 0.52}) +[2026-06-15 17:00:32,060][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.12297584861516953, 'train/grad_norm_pre_clip': 4.280472755432129, 'timing/data': 0.010790885426104069, 'timing/model': 1.468093710951507, 'learning_rate/qwen_vl_interface': 8.992009204065505e-06, 'learning_rate/action_model': 8.992009204065505e-05, 'epoch': 0.52}) +[2026-06-15 17:00:37,978][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.2199145406484604, 'train/grad_norm_pre_clip': 9.850412368774414, 'timing/data': 0.01075858622789383, 'timing/model': 1.4844698943197727, 'learning_rate/qwen_vl_interface': 8.990132631758058e-06, 'learning_rate/action_model': 8.990132631758059e-05, 'epoch': 0.52}) +[2026-06-15 17:00:43,859][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.11676758527755737, 'train/grad_norm_pre_clip': 5.555181980133057, 'timing/data': 0.010702816769480705, 'timing/model': 1.4645249107852578, 'learning_rate/qwen_vl_interface': 8.988254522023364e-06, 'learning_rate/action_model': 8.988254522023363e-05, 'epoch': 0.52}) +[2026-06-15 17:00:49,772][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.28276440501213074, 'train/grad_norm_pre_clip': 9.12703800201416, 'timing/data': 0.01080272626131773, 'timing/model': 1.4372918494045734, 'learning_rate/qwen_vl_interface': 8.986374875633441e-06, 'learning_rate/action_model': 8.98637487563344e-05, 'epoch': 0.52}) +[2026-06-15 17:00:55,769][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.06851907074451447, 'train/grad_norm_pre_clip': 2.514986038208008, 'timing/data': 0.013512113131582737, 'timing/model': 1.7396486001089215, 'learning_rate/qwen_vl_interface': 8.984493693360944e-06, 'learning_rate/action_model': 8.984493693360943e-05, 'epoch': 0.52}) +[2026-06-15 17:01:01,477][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.38340404629707336, 'train/grad_norm_pre_clip': 8.5338134765625, 'timing/data': 0.010919466614723206, 'timing/model': 1.46182750351727, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 0.52}) +[2026-06-15 17:01:07,367][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.2304689884185791, 'train/grad_norm_pre_clip': 5.402431964874268, 'timing/data': 0.014100294560194016, 'timing/model': 1.460706152021885, 'learning_rate/qwen_vl_interface': 8.98072672426198e-06, 'learning_rate/action_model': 8.980726724261981e-05, 'epoch': 0.52}) +[2026-06-15 17:01:13,357][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.15810441970825195, 'train/grad_norm_pre_clip': 5.121063232421875, 'timing/data': 0.01101306639611721, 'timing/model': 1.4467636793851852, 'learning_rate/qwen_vl_interface': 8.978840938983977e-06, 'learning_rate/action_model': 8.978840938983977e-05, 'epoch': 0.52}) +[2026-06-15 17:01:19,263][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.14467313885688782, 'train/grad_norm_pre_clip': 4.373805046081543, 'timing/data': 0.010936207138001919, 'timing/model': 1.4680013339966536, 'learning_rate/qwen_vl_interface': 8.976953620920313e-06, 'learning_rate/action_model': 8.976953620920312e-05, 'epoch': 0.52}) +[2026-06-15 17:01:25,274][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.13640332221984863, 'train/grad_norm_pre_clip': 6.055120468139648, 'timing/data': 0.012108941562473774, 'timing/model': 1.53158113732934, 'learning_rate/qwen_vl_interface': 8.975064770846794e-06, 'learning_rate/action_model': 8.975064770846794e-05, 'epoch': 0.52}) +[2026-06-15 17:01:31,229][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.18276652693748474, 'train/grad_norm_pre_clip': 11.311989784240723, 'timing/data': 0.010667366907000542, 'timing/model': 1.723971514031291, 'learning_rate/qwen_vl_interface': 8.973174389539856e-06, 'learning_rate/action_model': 8.973174389539855e-05, 'epoch': 0.52}) +[2026-06-15 17:01:36,962][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.24746859073638916, 'train/grad_norm_pre_clip': 6.2621612548828125, 'timing/data': 0.012588731944561005, 'timing/model': 1.4631706941872835, 'learning_rate/qwen_vl_interface': 8.971282477776563e-06, 'learning_rate/action_model': 8.971282477776563e-05, 'epoch': 0.52}) +[2026-06-15 17:01:42,863][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.33666518330574036, 'train/grad_norm_pre_clip': 11.487181663513184, 'timing/data': 0.011096538975834846, 'timing/model': 1.4628649530932307, 'learning_rate/qwen_vl_interface': 8.969389036334609e-06, 'learning_rate/action_model': 8.969389036334609e-05, 'epoch': 0.52}) +[2026-06-15 17:01:48,763][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.05475904792547226, 'train/grad_norm_pre_clip': 2.44608736038208, 'timing/data': 0.011151148937642574, 'timing/model': 1.4662055214866996, 'learning_rate/qwen_vl_interface': 8.967494065992315e-06, 'learning_rate/action_model': 8.967494065992315e-05, 'epoch': 0.53}) +[2026-06-15 17:01:54,792][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.18633125722408295, 'train/grad_norm_pre_clip': 6.265332221984863, 'timing/data': 0.01077651884406805, 'timing/model': 1.5062334472313523, 'learning_rate/qwen_vl_interface': 8.965597567528634e-06, 'learning_rate/action_model': 8.965597567528633e-05, 'epoch': 0.53}) +[2026-06-15 17:02:00,677][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.3492392301559448, 'train/grad_norm_pre_clip': 19.17215347290039, 'timing/data': 0.011409659869968891, 'timing/model': 1.4449200239032507, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.53}) +[2026-06-15 17:02:06,595][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.18116064369678497, 'train/grad_norm_pre_clip': 7.233457565307617, 'timing/data': 0.011067299172282219, 'timing/model': 1.695236025378108, 'learning_rate/qwen_vl_interface': 8.961799989356048e-06, 'learning_rate/action_model': 8.961799989356048e-05, 'epoch': 0.53}) +[2026-06-15 17:02:12,442][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.4273394048213959, 'train/grad_norm_pre_clip': 15.030356407165527, 'timing/data': 0.011277890764176846, 'timing/model': 1.647558587603271, 'learning_rate/qwen_vl_interface': 8.959898911208185e-06, 'learning_rate/action_model': 8.959898911208185e-05, 'epoch': 0.53}) +[2026-06-15 17:02:18,165][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.16085375845432281, 'train/grad_norm_pre_clip': 4.4664306640625, 'timing/data': 0.011482011526823044, 'timing/model': 1.4609750304371119, 'learning_rate/qwen_vl_interface': 8.957996308061015e-06, 'learning_rate/action_model': 8.957996308061014e-05, 'epoch': 0.53}) +[2026-06-15 17:02:24,070][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.144771546125412, 'train/grad_norm_pre_clip': 3.8682539463043213, 'timing/data': 0.012172303162515163, 'timing/model': 1.4676157254725695, 'learning_rate/qwen_vl_interface': 8.956092180696626e-06, 'learning_rate/action_model': 8.956092180696625e-05, 'epoch': 0.53}) +[2026-06-15 17:02:30,077][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.27818769216537476, 'train/grad_norm_pre_clip': 10.145330429077148, 'timing/data': 0.011124401353299618, 'timing/model': 1.501745818182826, 'learning_rate/qwen_vl_interface': 8.95418652989773e-06, 'learning_rate/action_model': 8.95418652989773e-05, 'epoch': 0.53}) +[2026-06-15 17:02:36,036][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.25002965331077576, 'train/grad_norm_pre_clip': 7.625197410583496, 'timing/data': 0.010973090305924416, 'timing/model': 1.501264643855393, 'learning_rate/qwen_vl_interface': 8.952279356447675e-06, 'learning_rate/action_model': 8.952279356447675e-05, 'epoch': 0.53}) +[2026-06-15 17:02:41,984][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.24237820506095886, 'train/grad_norm_pre_clip': 8.232409477233887, 'timing/data': 0.011148741468787193, 'timing/model': 1.6811198871582747, 'learning_rate/qwen_vl_interface': 8.950370661130421e-06, 'learning_rate/action_model': 8.950370661130422e-05, 'epoch': 0.53}) +[2026-06-15 17:02:47,892][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.38044220209121704, 'train/grad_norm_pre_clip': 11.695125579833984, 'timing/data': 0.011119560338556767, 'timing/model': 1.6937456438317895, 'learning_rate/qwen_vl_interface': 8.948460444730566e-06, 'learning_rate/action_model': 8.948460444730566e-05, 'epoch': 0.53}) +[2026-06-15 17:02:53,563][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.5731485486030579, 'train/grad_norm_pre_clip': 12.218329429626465, 'timing/data': 0.01072424091398716, 'timing/model': 1.4733525263145566, 'learning_rate/qwen_vl_interface': 8.946548708033324e-06, 'learning_rate/action_model': 8.946548708033323e-05, 'epoch': 0.53}) +[2026-06-15 17:02:59,454][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.20930486917495728, 'train/grad_norm_pre_clip': 4.719082355499268, 'timing/data': 0.01018310897052288, 'timing/model': 1.4579337071627378, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 0.53}) +[2026-06-15 17:03:05,358][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.19795656204223633, 'train/grad_norm_pre_clip': 11.062949180603027, 'timing/data': 0.010678949765861034, 'timing/model': 1.4577167630195618, 'learning_rate/qwen_vl_interface': 8.942720676890677e-06, 'learning_rate/action_model': 8.942720676890677e-05, 'epoch': 0.53}) +[2026-06-15 17:03:11,338][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.11196698993444443, 'train/grad_norm_pre_clip': 3.122713804244995, 'timing/data': 0.011143932119011879, 'timing/model': 1.500271150842309, 'learning_rate/qwen_vl_interface': 8.940804384018833e-06, 'learning_rate/action_model': 8.940804384018832e-05, 'epoch': 0.53}) +[2026-06-15 17:03:17,293][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.10943479090929031, 'train/grad_norm_pre_clip': 2.980160713195801, 'timing/data': 0.010871091857552528, 'timing/model': 1.697485450655222, 'learning_rate/qwen_vl_interface': 8.93888657399672e-06, 'learning_rate/action_model': 8.93888657399672e-05, 'epoch': 0.53}) +[2026-06-15 17:03:22,957][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.18358951807022095, 'train/grad_norm_pre_clip': 6.039813041687012, 'timing/data': 0.01077383104711771, 'timing/model': 1.4677165960893035, 'learning_rate/qwen_vl_interface': 8.936967247612678e-06, 'learning_rate/action_model': 8.936967247612678e-05, 'epoch': 0.53}) +[2026-06-15 17:03:28,863][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.18594907224178314, 'train/grad_norm_pre_clip': 6.352992057800293, 'timing/data': 0.010812802240252495, 'timing/model': 1.468847144395113, 'learning_rate/qwen_vl_interface': 8.93504640565567e-06, 'learning_rate/action_model': 8.93504640565567e-05, 'epoch': 0.53}) +[2026-06-15 17:03:34,760][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.43339774012565613, 'train/grad_norm_pre_clip': 20.68832778930664, 'timing/data': 0.011063532903790474, 'timing/model': 1.4684976069256663, 'learning_rate/qwen_vl_interface': 8.933124048915278e-06, 'learning_rate/action_model': 8.933124048915278e-05, 'epoch': 0.53}) +[2026-06-15 17:03:40,653][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.16633696854114532, 'train/grad_norm_pre_clip': 5.050978183746338, 'timing/data': 0.01065171230584383, 'timing/model': 1.4294144753366709, 'learning_rate/qwen_vl_interface': 8.931200178181715e-06, 'learning_rate/action_model': 8.931200178181715e-05, 'epoch': 0.53}) +[2026-06-15 17:03:46,564][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.4222250282764435, 'train/grad_norm_pre_clip': 4.9751811027526855, 'timing/data': 0.011394303292036057, 'timing/model': 1.4945845026522875, 'learning_rate/qwen_vl_interface': 8.929274794245812e-06, 'learning_rate/action_model': 8.929274794245812e-05, 'epoch': 0.53}) +[2026-06-15 17:03:52,460][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.08060965687036514, 'train/grad_norm_pre_clip': 1.8882745504379272, 'timing/data': 0.011150454171001911, 'timing/model': 1.6673953449353576, 'learning_rate/qwen_vl_interface': 8.927347897899018e-06, 'learning_rate/action_model': 8.927347897899016e-05, 'epoch': 0.53}) +[2026-06-15 17:03:58,158][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.28854480385780334, 'train/grad_norm_pre_clip': 10.792312622070312, 'timing/data': 0.01088729314506054, 'timing/model': 1.4661662131547928, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 0.54}) +[2026-06-15 17:04:04,059][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.4060121476650238, 'train/grad_norm_pre_clip': 15.567883491516113, 'timing/data': 0.011974425986409187, 'timing/model': 1.4597449759021401, 'learning_rate/qwen_vl_interface': 8.923489571141679e-06, 'learning_rate/action_model': 8.923489571141679e-05, 'epoch': 0.54}) +[2026-06-15 17:04:09,957][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.4452710747718811, 'train/grad_norm_pre_clip': 9.549049377441406, 'timing/data': 0.010828793048858643, 'timing/model': 1.4623796474188566, 'learning_rate/qwen_vl_interface': 8.921558142317147e-06, 'learning_rate/action_model': 8.921558142317147e-05, 'epoch': 0.54}) +[2026-06-15 17:04:15,857][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.22914454340934753, 'train/grad_norm_pre_clip': 5.594051361083984, 'timing/data': 0.010446702130138874, 'timing/model': 1.436294979415834, 'learning_rate/qwen_vl_interface': 8.91962520425375e-06, 'learning_rate/action_model': 8.91962520425375e-05, 'epoch': 0.54}) +[2026-06-15 17:04:21,819][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.10531559586524963, 'train/grad_norm_pre_clip': 2.4037911891937256, 'timing/data': 0.010564473457634449, 'timing/model': 1.4859241638332605, 'learning_rate/qwen_vl_interface': 8.917690757746044e-06, 'learning_rate/action_model': 8.917690757746044e-05, 'epoch': 0.54}) +[2026-06-15 17:04:27,654][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.286490261554718, 'train/grad_norm_pre_clip': 5.463999271392822, 'timing/data': 0.010575723834335804, 'timing/model': 1.6566936047747731, 'learning_rate/qwen_vl_interface': 8.915754803589208e-06, 'learning_rate/action_model': 8.915754803589209e-05, 'epoch': 0.54}) +[2026-06-15 17:04:33,355][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.3265427052974701, 'train/grad_norm_pre_clip': 4.446237564086914, 'timing/data': 0.010154982097446918, 'timing/model': 1.45912220235914, 'learning_rate/qwen_vl_interface': 8.913817342579042e-06, 'learning_rate/action_model': 8.913817342579042e-05, 'epoch': 0.54}) +[2026-06-15 17:04:39,257][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.2627277672290802, 'train/grad_norm_pre_clip': 7.994112491607666, 'timing/data': 0.011019304394721985, 'timing/model': 1.4626006558537483, 'learning_rate/qwen_vl_interface': 8.91187837551196e-06, 'learning_rate/action_model': 8.911878375511961e-05, 'epoch': 0.54}) +[2026-06-15 17:04:45,256][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.1587628424167633, 'train/grad_norm_pre_clip': 5.397934436798096, 'timing/data': 0.010350523516535759, 'timing/model': 1.46550268586725, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.54}) +[2026-06-15 17:04:51,156][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.16767022013664246, 'train/grad_norm_pre_clip': 3.413784980773926, 'timing/data': 0.010275322943925858, 'timing/model': 1.42988324444741, 'learning_rate/qwen_vl_interface': 8.907995926395814e-06, 'learning_rate/action_model': 8.907995926395815e-05, 'epoch': 0.54}) +[2026-06-15 17:04:57,088][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.2375587373971939, 'train/grad_norm_pre_clip': 5.999183177947998, 'timing/data': 0.010065342299640179, 'timing/model': 1.692429036833346, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 0.54}) +[2026-06-15 17:05:02,759][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.23299124836921692, 'train/grad_norm_pre_clip': 5.4124979972839355, 'timing/data': 0.010591634549200535, 'timing/model': 1.4643691135570407, 'learning_rate/qwen_vl_interface': 8.904107462624484e-06, 'learning_rate/action_model': 8.904107462624483e-05, 'epoch': 0.54}) +[2026-06-15 17:05:08,655][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.1900433450937271, 'train/grad_norm_pre_clip': 5.054383754730225, 'timing/data': 0.01036195270717144, 'timing/model': 1.455810695886612, 'learning_rate/qwen_vl_interface': 8.902160977240739e-06, 'learning_rate/action_model': 8.902160977240739e-05, 'epoch': 0.54}) +[2026-06-15 17:05:14,557][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.24882440268993378, 'train/grad_norm_pre_clip': 6.043478965759277, 'timing/data': 0.011741848662495613, 'timing/model': 1.4565567383542657, 'learning_rate/qwen_vl_interface': 8.90021299059157e-06, 'learning_rate/action_model': 8.90021299059157e-05, 'epoch': 0.54}) +[2026-06-15 17:05:20,459][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.1216234341263771, 'train/grad_norm_pre_clip': 3.212564468383789, 'timing/data': 0.011352727189660072, 'timing/model': 1.467301132157445, 'learning_rate/qwen_vl_interface': 8.898263503477722e-06, 'learning_rate/action_model': 8.898263503477721e-05, 'epoch': 0.54}) +[2026-06-15 17:05:26,353][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.13475941121578217, 'train/grad_norm_pre_clip': 3.1991055011749268, 'timing/data': 0.010997665114700794, 'timing/model': 1.4718469260260463, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.54}) +[2026-06-15 17:05:32,297][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.07872456312179565, 'train/grad_norm_pre_clip': 5.182394981384277, 'timing/data': 0.011242035776376724, 'timing/model': 1.7064680783078074, 'learning_rate/qwen_vl_interface': 8.894360031062043e-06, 'learning_rate/action_model': 8.894360031062043e-05, 'epoch': 0.54}) +[2026-06-15 17:05:37,962][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.2514076232910156, 'train/grad_norm_pre_clip': 4.8590922355651855, 'timing/data': 0.01078637596219778, 'timing/model': 1.4724809555336833, 'learning_rate/qwen_vl_interface': 8.892406047364783e-06, 'learning_rate/action_model': 8.892406047364784e-05, 'epoch': 0.54}) +[2026-06-15 17:05:43,945][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.3205021917819977, 'train/grad_norm_pre_clip': 16.149925231933594, 'timing/data': 0.01033023465424776, 'timing/model': 1.4365649176761508, 'learning_rate/qwen_vl_interface': 8.890450566411984e-06, 'learning_rate/action_model': 8.890450566411984e-05, 'epoch': 0.54}) +[2026-06-15 17:05:49,858][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.03892817348241806, 'train/grad_norm_pre_clip': 3.8423049449920654, 'timing/data': 0.011691098101437092, 'timing/model': 1.4694221029058099, 'learning_rate/qwen_vl_interface': 8.888493589007467e-06, 'learning_rate/action_model': 8.888493589007468e-05, 'epoch': 0.54}) +[2026-06-15 17:05:55,761][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.28212544322013855, 'train/grad_norm_pre_clip': 14.508139610290527, 'timing/data': 0.011123106814920902, 'timing/model': 1.4689447898417711, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 0.54}) +[2026-06-15 17:06:01,656][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.0875505656003952, 'train/grad_norm_pre_clip': 4.699825286865234, 'timing/data': 0.010760956443846226, 'timing/model': 1.4437868148088455, 'learning_rate/qwen_vl_interface': 8.884575148061662e-06, 'learning_rate/action_model': 8.88457514806166e-05, 'epoch': 0.54}) +[2026-06-15 17:06:07,594][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.22291408479213715, 'train/grad_norm_pre_clip': 5.415525913238525, 'timing/data': 0.011285478249192238, 'timing/model': 1.6856792671605945, 'learning_rate/qwen_vl_interface': 8.882613686131095e-06, 'learning_rate/action_model': 8.882613686131093e-05, 'epoch': 0.55}) +[2026-06-15 17:06:13,262][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.297375351190567, 'train/grad_norm_pre_clip': 11.675189018249512, 'timing/data': 0.011931339278817177, 'timing/model': 1.4728700555860996, 'learning_rate/qwen_vl_interface': 8.880650730970254e-06, 'learning_rate/action_model': 8.880650730970254e-05, 'epoch': 0.55}) +[2026-06-15 17:06:19,158][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.2774372100830078, 'train/grad_norm_pre_clip': 14.77323055267334, 'timing/data': 0.011213936842978, 'timing/model': 1.4664870733395219, 'learning_rate/qwen_vl_interface': 8.878686283386042e-06, 'learning_rate/action_model': 8.878686283386042e-05, 'epoch': 0.55}) +[2026-06-15 17:06:25,058][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.19073902070522308, 'train/grad_norm_pre_clip': 6.377323150634766, 'timing/data': 0.013377955183386803, 'timing/model': 1.459792627952993, 'learning_rate/qwen_vl_interface': 8.876720344185967e-06, 'learning_rate/action_model': 8.876720344185967e-05, 'epoch': 0.55}) +[2026-06-15 17:06:30,954][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.38355064392089844, 'train/grad_norm_pre_clip': 12.412708282470703, 'timing/data': 0.01031370460987091, 'timing/model': 1.4411899410188198, 'learning_rate/qwen_vl_interface': 8.874752914178149e-06, 'learning_rate/action_model': 8.87475291417815e-05, 'epoch': 0.55}) +[2026-06-15 17:06:37,049][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.17271584272384644, 'train/grad_norm_pre_clip': 3.4305670261383057, 'timing/data': 0.010790066793560982, 'timing/model': 1.4603702733293176, 'learning_rate/qwen_vl_interface': 8.87278399417133e-06, 'learning_rate/action_model': 8.872783994171329e-05, 'epoch': 0.55}) +[2026-06-15 17:06:42,996][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.20719680190086365, 'train/grad_norm_pre_clip': 9.051782608032227, 'timing/data': 0.01074407622218132, 'timing/model': 1.705933184362948, 'learning_rate/qwen_vl_interface': 8.870813584974852e-06, 'learning_rate/action_model': 8.870813584974852e-05, 'epoch': 0.55}) +[2026-06-15 17:06:48,665][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.35768088698387146, 'train/grad_norm_pre_clip': 6.810765266418457, 'timing/data': 0.010971317999064922, 'timing/model': 1.4709949232637882, 'learning_rate/qwen_vl_interface': 8.868841687398683e-06, 'learning_rate/action_model': 8.868841687398683e-05, 'epoch': 0.55}) +[2026-06-15 17:17:20,539][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.37036702036857605, 'train/grad_norm_pre_clip': 9.349024772644043, 'eval/loss': 0.10319027304649353, 'eval/flap_precision': 0.6741935483870968, 'eval/flap_recall': 0.76, 'eval/flap_f1': 0.7145299145299147, 'eval/flap_rate_model': 0.08611111111111111, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7145299145299147, 'timing/data': 0.011706840246915817, 'timing/model': 1.4575499929487705, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 0.55}) +[2026-06-15 17:17:20,539][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1200_state +[2026-06-15 17:17:20,559][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 17:17:28,775][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1200_state/model.safetensors +[2026-06-15 17:17:47,825][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1200_state/optimizer.bin +[2026-06-15 17:17:47,826][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1200_state/sampler.bin +[2026-06-15 17:17:47,826][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1200_state/sampler_1.bin +[2026-06-15 17:17:47,827][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1200_state/random_states_0.pkl +[2026-06-15 17:17:47,976][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-06-15 17:17:47,976][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 17:17:47,990][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 17:17:53,635][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.08852624893188477, 'train/grad_norm_pre_clip': 3.635226011276245, 'timing/data': 0.010373721830546856, 'timing/model': 1.4746328461915255, 'learning_rate/qwen_vl_interface': 8.864893430350159e-06, 'learning_rate/action_model': 8.864893430350159e-05, 'epoch': 0.55}) +[2026-06-15 17:17:59,523][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.1167277842760086, 'train/grad_norm_pre_clip': 3.468648910522461, 'timing/data': 0.01063828356564045, 'timing/model': 1.4648876804858446, 'learning_rate/qwen_vl_interface': 8.862917072500785e-06, 'learning_rate/action_model': 8.862917072500784e-05, 'epoch': 0.55}) +[2026-06-15 17:18:05,329][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.25972428917884827, 'train/grad_norm_pre_clip': 7.152738094329834, 'timing/data': 0.010989625006914139, 'timing/model': 1.460874848999083, 'learning_rate/qwen_vl_interface': 8.860939229517674e-06, 'learning_rate/action_model': 8.860939229517674e-05, 'epoch': 0.55}) +[2026-06-15 17:18:11,212][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.23224247992038727, 'train/grad_norm_pre_clip': 8.995914459228516, 'timing/data': 0.012728093191981316, 'timing/model': 1.5276270480826497, 'learning_rate/qwen_vl_interface': 8.858959902213844e-06, 'learning_rate/action_model': 8.858959902213845e-05, 'epoch': 0.55}) +[2026-06-15 17:18:17,037][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.11429373919963837, 'train/grad_norm_pre_clip': 3.747840166091919, 'timing/data': 0.011670397594571114, 'timing/model': 1.413666982203722, 'learning_rate/qwen_vl_interface': 8.856979091402921e-06, 'learning_rate/action_model': 8.85697909140292e-05, 'epoch': 0.55}) +[2026-06-15 17:18:23,056][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.07771075516939163, 'train/grad_norm_pre_clip': 3.5709197521209717, 'timing/data': 0.010663864202797413, 'timing/model': 1.7204762529581785, 'learning_rate/qwen_vl_interface': 8.854996797899139e-06, 'learning_rate/action_model': 8.854996797899139e-05, 'epoch': 0.55}) +[2026-06-15 17:18:28,749][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.21602727472782135, 'train/grad_norm_pre_clip': 6.3006978034973145, 'timing/data': 0.011107286438345909, 'timing/model': 1.4783629877492785, 'learning_rate/qwen_vl_interface': 8.853013022517348e-06, 'learning_rate/action_model': 8.853013022517348e-05, 'epoch': 0.55}) +[2026-06-15 17:18:34,649][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.37945184111595154, 'train/grad_norm_pre_clip': 12.720550537109375, 'timing/data': 0.011144745163619518, 'timing/model': 1.4713287632912397, 'learning_rate/qwen_vl_interface': 8.851027766073002e-06, 'learning_rate/action_model': 8.851027766073001e-05, 'epoch': 0.55}) +[2026-06-15 17:18:40,450][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.260292112827301, 'train/grad_norm_pre_clip': 6.412731170654297, 'timing/data': 0.012653833255171776, 'timing/model': 1.4729007128626108, 'learning_rate/qwen_vl_interface': 8.84904102938216e-06, 'learning_rate/action_model': 8.84904102938216e-05, 'epoch': 0.55}) +[2026-06-15 17:18:46,354][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.3206241726875305, 'train/grad_norm_pre_clip': 7.553590774536133, 'timing/data': 0.013219314627349377, 'timing/model': 1.4715821323916316, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 0.55}) +[2026-06-15 17:18:52,257][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.2823009788990021, 'train/grad_norm_pre_clip': 6.1984052658081055, 'timing/data': 0.012244599871337414, 'timing/model': 1.4297140426933765, 'learning_rate/qwen_vl_interface': 8.845063118528303e-06, 'learning_rate/action_model': 8.845063118528304e-05, 'epoch': 0.55}) +[2026-06-15 17:18:58,181][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.18724124133586884, 'train/grad_norm_pre_clip': 4.790381908416748, 'timing/data': 0.011145295575261116, 'timing/model': 1.696667892858386, 'learning_rate/qwen_vl_interface': 8.843071946000455e-06, 'learning_rate/action_model': 8.843071946000456e-05, 'epoch': 0.55}) +[2026-06-15 17:19:03,856][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.11886449158191681, 'train/grad_norm_pre_clip': 3.1465206146240234, 'timing/data': 0.011619377881288528, 'timing/model': 1.4667955143377185, 'learning_rate/qwen_vl_interface': 8.841079296496451e-06, 'learning_rate/action_model': 8.84107929649645e-05, 'epoch': 0.55}) +[2026-06-15 17:19:09,758][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.17312631011009216, 'train/grad_norm_pre_clip': 5.630572319030762, 'timing/data': 0.011645709164440632, 'timing/model': 1.4698556596413255, 'learning_rate/qwen_vl_interface': 8.839085170835396e-06, 'learning_rate/action_model': 8.839085170835395e-05, 'epoch': 0.56}) +[2026-06-15 17:19:15,657][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.16641834378242493, 'train/grad_norm_pre_clip': 4.0584635734558105, 'timing/data': 0.011585267260670662, 'timing/model': 1.4641752010211349, 'learning_rate/qwen_vl_interface': 8.837089569837e-06, 'learning_rate/action_model': 8.837089569836998e-05, 'epoch': 0.56}) +[2026-06-15 17:19:21,555][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.15328490734100342, 'train/grad_norm_pre_clip': 5.453818321228027, 'timing/data': 0.011364676989614964, 'timing/model': 1.4670272273942828, 'learning_rate/qwen_vl_interface': 8.835092494321575e-06, 'learning_rate/action_model': 8.835092494321575e-05, 'epoch': 0.56}) +[2026-06-15 17:19:27,459][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.09229833632707596, 'train/grad_norm_pre_clip': 2.9844439029693604, 'timing/data': 0.011049985885620117, 'timing/model': 1.4310276918113232, 'learning_rate/qwen_vl_interface': 8.833093945110048e-06, 'learning_rate/action_model': 8.833093945110047e-05, 'epoch': 0.56}) +[2026-06-15 17:19:33,396][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.17330574989318848, 'train/grad_norm_pre_clip': 5.251499652862549, 'timing/data': 0.011332876980304718, 'timing/model': 1.7057614075019956, 'learning_rate/qwen_vl_interface': 8.831093923023946e-06, 'learning_rate/action_model': 8.831093923023945e-05, 'epoch': 0.56}) +[2026-06-15 17:19:39,319][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.2544625997543335, 'train/grad_norm_pre_clip': 6.755213260650635, 'timing/data': 0.010859054513275623, 'timing/model': 1.729406449943781, 'learning_rate/qwen_vl_interface': 8.8290924288854e-06, 'learning_rate/action_model': 8.8290924288854e-05, 'epoch': 0.56}) +[2026-06-15 17:19:45,070][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.11802789568901062, 'train/grad_norm_pre_clip': 2.9271345138549805, 'timing/data': 0.011506537906825542, 'timing/model': 1.473742289468646, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.56}) +[2026-06-15 17:19:50,960][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.12552739679813385, 'train/grad_norm_pre_clip': 4.834243297576904, 'timing/data': 0.011650178581476212, 'timing/model': 1.4528467655181885, 'learning_rate/qwen_vl_interface': 8.825085027742545e-06, 'learning_rate/action_model': 8.825085027742544e-05, 'epoch': 0.56}) +[2026-06-15 17:19:56,864][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.09504570066928864, 'train/grad_norm_pre_clip': 2.999162435531616, 'timing/data': 0.011627988889813423, 'timing/model': 1.4627405498176813, 'learning_rate/qwen_vl_interface': 8.823079122385523e-06, 'learning_rate/action_model': 8.823079122385523e-05, 'epoch': 0.56}) +[2026-06-15 17:20:02,826][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.015095764771103859, 'train/grad_norm_pre_clip': 1.1724913120269775, 'timing/data': 0.011171206831932068, 'timing/model': 1.4953924305737019, 'learning_rate/qwen_vl_interface': 8.821071748270644e-06, 'learning_rate/action_model': 8.821071748270644e-05, 'epoch': 0.56}) +[2026-06-15 17:20:08,802][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.11350005865097046, 'train/grad_norm_pre_clip': 3.302025318145752, 'timing/data': 0.0110377361997962, 'timing/model': 1.7155837947502732, 'learning_rate/qwen_vl_interface': 8.81906290622306e-06, 'learning_rate/action_model': 8.81906290622306e-05, 'epoch': 0.56}) +[2026-06-15 17:20:14,458][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.10232522338628769, 'train/grad_norm_pre_clip': 7.238534450531006, 'timing/data': 0.011272337287664413, 'timing/model': 1.4607968628406525, 'learning_rate/qwen_vl_interface': 8.81705259706853e-06, 'learning_rate/action_model': 8.817052597068529e-05, 'epoch': 0.56}) +[2026-06-15 17:20:20,368][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.18472570180892944, 'train/grad_norm_pre_clip': 6.580014705657959, 'timing/data': 0.011352118104696274, 'timing/model': 1.4773134356364608, 'learning_rate/qwen_vl_interface': 8.815040821633416e-06, 'learning_rate/action_model': 8.815040821633416e-05, 'epoch': 0.56}) +[2026-06-15 17:20:26,368][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.16135844588279724, 'train/grad_norm_pre_clip': 8.219080924987793, 'timing/data': 0.011373638175427914, 'timing/model': 1.5105762174353004, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.56}) +[2026-06-15 17:20:32,277][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.12659017741680145, 'train/grad_norm_pre_clip': 13.260479927062988, 'timing/data': 0.011324738152325153, 'timing/model': 1.4425681754946709, 'learning_rate/qwen_vl_interface': 8.8110128752299e-06, 'learning_rate/action_model': 8.8110128752299e-05, 'epoch': 0.56}) +[2026-06-15 17:20:38,158][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.16648274660110474, 'train/grad_norm_pre_clip': 16.513221740722656, 'timing/data': 0.011104116216301918, 'timing/model': 1.4852930586785078, 'learning_rate/qwen_vl_interface': 8.808996705917235e-06, 'learning_rate/action_model': 8.808996705917235e-05, 'epoch': 0.56}) +[2026-06-15 17:20:44,055][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.14275388419628143, 'train/grad_norm_pre_clip': 13.258980751037598, 'timing/data': 0.01115957647562027, 'timing/model': 1.6564266681671143, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 0.56}) +[2026-06-15 17:20:49,761][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.2642539143562317, 'train/grad_norm_pre_clip': 12.796096801757812, 'timing/data': 0.010936886072158813, 'timing/model': 1.4631318422034383, 'learning_rate/qwen_vl_interface': 8.804959979213945e-06, 'learning_rate/action_model': 8.804959979213945e-05, 'epoch': 0.56}) +[2026-06-15 17:20:55,761][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.2517247498035431, 'train/grad_norm_pre_clip': 14.32698917388916, 'timing/data': 0.011982699856162071, 'timing/model': 1.456975101493299, 'learning_rate/qwen_vl_interface': 8.802939423482667e-06, 'learning_rate/action_model': 8.802939423482666e-05, 'epoch': 0.56}) +[2026-06-15 17:21:01,759][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.3545490801334381, 'train/grad_norm_pre_clip': 14.63410472869873, 'timing/data': 0.0102976830676198, 'timing/model': 1.4736536731943488, 'learning_rate/qwen_vl_interface': 8.800917407272198e-06, 'learning_rate/action_model': 8.800917407272198e-05, 'epoch': 0.56}) +[2026-06-15 17:21:07,652][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.02571687661111355, 'train/grad_norm_pre_clip': 4.358014106750488, 'timing/data': 0.010015792213380337, 'timing/model': 1.437036981806159, 'learning_rate/qwen_vl_interface': 8.798893931413711e-06, 'learning_rate/action_model': 8.79889393141371e-05, 'epoch': 0.56}) +[2026-06-15 17:21:13,553][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.2240525633096695, 'train/grad_norm_pre_clip': 10.39409351348877, 'timing/data': 0.010305672883987427, 'timing/model': 1.4350354271009564, 'learning_rate/qwen_vl_interface': 8.796868996738985e-06, 'learning_rate/action_model': 8.796868996738984e-05, 'epoch': 0.56}) +[2026-06-15 17:21:19,458][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.23634034395217896, 'train/grad_norm_pre_clip': 11.797904014587402, 'timing/data': 0.01092680636793375, 'timing/model': 1.6687162909656763, 'learning_rate/qwen_vl_interface': 8.794842604080389e-06, 'learning_rate/action_model': 8.794842604080389e-05, 'epoch': 0.57}) +[2026-06-15 17:21:25,161][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.4525367021560669, 'train/grad_norm_pre_clip': 10.17851734161377, 'timing/data': 0.010806255042552948, 'timing/model': 1.4739087261259556, 'learning_rate/qwen_vl_interface': 8.7928147542709e-06, 'learning_rate/action_model': 8.7928147542709e-05, 'epoch': 0.57}) +[2026-06-15 17:21:31,055][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.049478329718112946, 'train/grad_norm_pre_clip': 7.246142387390137, 'timing/data': 0.012023970484733582, 'timing/model': 1.462531790137291, 'learning_rate/qwen_vl_interface': 8.790785448144088e-06, 'learning_rate/action_model': 8.790785448144088e-05, 'epoch': 0.57}) +[2026-06-15 17:21:36,960][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.11098510771989822, 'train/grad_norm_pre_clip': 3.130683660507202, 'timing/data': 0.012308672070503235, 'timing/model': 1.4653311744332314, 'learning_rate/qwen_vl_interface': 8.788754686534126e-06, 'learning_rate/action_model': 8.788754686534127e-05, 'epoch': 0.57}) +[2026-06-15 17:21:42,959][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.18117474019527435, 'train/grad_norm_pre_clip': 7.306419849395752, 'timing/data': 0.011625849641859531, 'timing/model': 1.519831968471408, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 0.57}) +[2026-06-15 17:21:48,908][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.10170028358697891, 'train/grad_norm_pre_clip': 7.578512668609619, 'timing/data': 0.010956726036965847, 'timing/model': 1.7154734134674072, 'learning_rate/qwen_vl_interface': 8.784688800204427e-06, 'learning_rate/action_model': 8.784688800204426e-05, 'epoch': 0.57}) +[2026-06-15 17:21:54,566][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.08004064857959747, 'train/grad_norm_pre_clip': 4.815302848815918, 'timing/data': 0.01120152696967125, 'timing/model': 1.46945418138057, 'learning_rate/qwen_vl_interface': 8.782653677156022e-06, 'learning_rate/action_model': 8.782653677156022e-05, 'epoch': 0.57}) +[2026-06-15 17:22:00,470][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.28437042236328125, 'train/grad_norm_pre_clip': 20.920122146606445, 'timing/data': 0.011528538540005684, 'timing/model': 1.476465173996985, 'learning_rate/qwen_vl_interface': 8.780617101967129e-06, 'learning_rate/action_model': 8.780617101967129e-05, 'epoch': 0.57}) +[2026-06-15 17:22:06,364][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.18821808695793152, 'train/grad_norm_pre_clip': 7.529468536376953, 'timing/data': 0.01092502661049366, 'timing/model': 1.4730500923469663, 'learning_rate/qwen_vl_interface': 8.77857907547491e-06, 'learning_rate/action_model': 8.77857907547491e-05, 'epoch': 0.57}) +[2026-06-15 17:22:12,266][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.23109915852546692, 'train/grad_norm_pre_clip': 15.992889404296875, 'timing/data': 0.010701984167098999, 'timing/model': 1.4729852033779025, 'learning_rate/qwen_vl_interface': 8.776539598517119e-06, 'learning_rate/action_model': 8.776539598517118e-05, 'epoch': 0.57}) +[2026-06-15 17:22:18,159][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.13597489893436432, 'train/grad_norm_pre_clip': 9.592284202575684, 'timing/data': 0.011152206920087337, 'timing/model': 1.4588946178555489, 'learning_rate/qwen_vl_interface': 8.774498671932108e-06, 'learning_rate/action_model': 8.774498671932107e-05, 'epoch': 0.57}) +[2026-06-15 17:22:24,093][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.21122540533542633, 'train/grad_norm_pre_clip': 5.405800819396973, 'timing/data': 0.010962337255477905, 'timing/model': 1.6945769544690847, 'learning_rate/qwen_vl_interface': 8.772456296558824e-06, 'learning_rate/action_model': 8.772456296558824e-05, 'epoch': 0.57}) +[2026-06-15 17:22:29,761][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.18990573287010193, 'train/grad_norm_pre_clip': 6.567015647888184, 'timing/data': 0.010754735209047794, 'timing/model': 1.4523569447919726, 'learning_rate/qwen_vl_interface': 8.77041247323681e-06, 'learning_rate/action_model': 8.770412473236811e-05, 'epoch': 0.57}) +[2026-06-15 17:22:35,657][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.08203496038913727, 'train/grad_norm_pre_clip': 2.084099531173706, 'timing/data': 0.011589579284191132, 'timing/model': 1.4555572615936399, 'learning_rate/qwen_vl_interface': 8.76836720280621e-06, 'learning_rate/action_model': 8.768367202806209e-05, 'epoch': 0.57}) +[2026-06-15 17:22:41,570][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.24979117512702942, 'train/grad_norm_pre_clip': 5.151811122894287, 'timing/data': 0.012253992259502411, 'timing/model': 1.4753825468942523, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 0.57}) +[2026-06-15 17:22:47,568][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.11134332418441772, 'train/grad_norm_pre_clip': 3.693233013153076, 'timing/data': 0.011209746822714806, 'timing/model': 1.4734386522322893, 'learning_rate/qwen_vl_interface': 8.76427232398276e-06, 'learning_rate/action_model': 8.76427232398276e-05, 'epoch': 0.57}) +[2026-06-15 17:22:53,460][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.21442820131778717, 'train/grad_norm_pre_clip': 29.74784278869629, 'timing/data': 0.011392978020012379, 'timing/model': 1.4586175521835685, 'learning_rate/qwen_vl_interface': 8.762222717273167e-06, 'learning_rate/action_model': 8.762222717273167e-05, 'epoch': 0.57}) +[2026-06-15 17:22:59,396][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.07424135506153107, 'train/grad_norm_pre_clip': 2.4500513076782227, 'timing/data': 0.011847109533846378, 'timing/model': 1.6950244447216392, 'learning_rate/qwen_vl_interface': 8.76017166682148e-06, 'learning_rate/action_model': 8.760171666821481e-05, 'epoch': 0.57}) +[2026-06-15 17:23:05,065][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.17633064091205597, 'train/grad_norm_pre_clip': 7.689759254455566, 'timing/data': 0.011329928413033485, 'timing/model': 1.4691899316385388, 'learning_rate/qwen_vl_interface': 8.758119173470815e-06, 'learning_rate/action_model': 8.758119173470815e-05, 'epoch': 0.57}) +[2026-06-15 17:23:11,158][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.12460298091173172, 'train/grad_norm_pre_clip': 4.689925670623779, 'timing/data': 0.011826470494270325, 'timing/model': 1.4574390342459083, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 0.57}) +[2026-06-15 17:23:17,070][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.42838311195373535, 'train/grad_norm_pre_clip': 9.663421630859375, 'timing/data': 0.011170757003128529, 'timing/model': 1.4595626955851912, 'learning_rate/qwen_vl_interface': 8.75400986144794e-06, 'learning_rate/action_model': 8.75400986144794e-05, 'epoch': 0.57}) +[2026-06-15 17:23:23,066][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.192887544631958, 'train/grad_norm_pre_clip': 10.860326766967773, 'timing/data': 0.010947596281766891, 'timing/model': 1.5172586627304554, 'learning_rate/qwen_vl_interface': 8.751953044464913e-06, 'learning_rate/action_model': 8.751953044464913e-05, 'epoch': 0.57}) +[2026-06-15 17:23:28,953][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.3698733150959015, 'train/grad_norm_pre_clip': 12.278182983398438, 'timing/data': 0.010288584046065807, 'timing/model': 1.4465296752750874, 'learning_rate/qwen_vl_interface': 8.749894787961272e-06, 'learning_rate/action_model': 8.749894787961271e-05, 'epoch': 0.58}) +[2026-06-15 17:23:34,889][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.29751402139663696, 'train/grad_norm_pre_clip': 5.996196269989014, 'timing/data': 0.01070963591337204, 'timing/model': 1.695168192498386, 'learning_rate/qwen_vl_interface': 8.747835092783086e-06, 'learning_rate/action_model': 8.747835092783085e-05, 'epoch': 0.58}) +[2026-06-15 17:23:40,555][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.20594671368598938, 'train/grad_norm_pre_clip': 5.401844024658203, 'timing/data': 0.011034946888685226, 'timing/model': 1.4568710625171661, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 0.58}) +[2026-06-15 17:23:46,459][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.25832700729370117, 'train/grad_norm_pre_clip': 9.540007591247559, 'timing/data': 0.01109961699694395, 'timing/model': 1.4658326134085655, 'learning_rate/qwen_vl_interface': 8.743711389790321e-06, 'learning_rate/action_model': 8.74371138979032e-05, 'epoch': 0.58}) +[2026-06-15 17:23:52,354][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.23161928355693817, 'train/grad_norm_pre_clip': 8.178557395935059, 'timing/data': 0.011600786820054054, 'timing/model': 1.4602932017296553, 'learning_rate/qwen_vl_interface': 8.741647383670839e-06, 'learning_rate/action_model': 8.741647383670839e-05, 'epoch': 0.58}) +[2026-06-15 17:23:58,259][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.15545101463794708, 'train/grad_norm_pre_clip': 3.5375657081604004, 'timing/data': 0.011433213017880917, 'timing/model': 1.4632730782032013, 'learning_rate/qwen_vl_interface': 8.739581942267011e-06, 'learning_rate/action_model': 8.739581942267011e-05, 'epoch': 0.58}) +[2026-06-15 17:24:04,249][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.2866402268409729, 'train/grad_norm_pre_clip': 8.41048526763916, 'timing/data': 0.01157523412257433, 'timing/model': 1.5119320275261998, 'learning_rate/qwen_vl_interface': 8.737515066427861e-06, 'learning_rate/action_model': 8.73751506642786e-05, 'epoch': 0.58}) +[2026-06-15 17:24:10,195][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.17343159019947052, 'train/grad_norm_pre_clip': 6.658918857574463, 'timing/data': 0.010985003784298897, 'timing/model': 1.701852584257722, 'learning_rate/qwen_vl_interface': 8.735446757003e-06, 'learning_rate/action_model': 8.735446757003e-05, 'epoch': 0.58}) +[2026-06-15 17:24:15,857][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.1273564249277115, 'train/grad_norm_pre_clip': 5.726007461547852, 'timing/data': 0.011575540527701378, 'timing/model': 1.4599415501579642, 'learning_rate/qwen_vl_interface': 8.733377014842635e-06, 'learning_rate/action_model': 8.733377014842635e-05, 'epoch': 0.58}) +[2026-06-15 17:24:21,758][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.15266288816928864, 'train/grad_norm_pre_clip': 6.067154884338379, 'timing/data': 0.011549310758709908, 'timing/model': 1.4493429018184543, 'learning_rate/qwen_vl_interface': 8.731305840797557e-06, 'learning_rate/action_model': 8.731305840797558e-05, 'epoch': 0.58}) +[2026-06-15 17:24:27,770][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.15139898657798767, 'train/grad_norm_pre_clip': 5.528867244720459, 'timing/data': 0.011827792972326279, 'timing/model': 1.4657437950372696, 'learning_rate/qwen_vl_interface': 8.72923323571915e-06, 'learning_rate/action_model': 8.72923323571915e-05, 'epoch': 0.58}) +[2026-06-15 17:24:33,668][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.06970082223415375, 'train/grad_norm_pre_clip': 2.1293652057647705, 'timing/data': 0.010826226323843002, 'timing/model': 1.4625205118209124, 'learning_rate/qwen_vl_interface': 8.727159200459381e-06, 'learning_rate/action_model': 8.727159200459381e-05, 'epoch': 0.58}) +[2026-06-15 17:24:39,662][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.19856588542461395, 'train/grad_norm_pre_clip': 11.333869934082031, 'timing/data': 0.011495806276798248, 'timing/model': 1.5237819543108344, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 0.58}) +[2026-06-15 17:24:45,612][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.2533058822154999, 'train/grad_norm_pre_clip': 9.919230461120605, 'timing/data': 0.011220631189644337, 'timing/model': 1.712414893321693, 'learning_rate/qwen_vl_interface': 8.723006842806575e-06, 'learning_rate/action_model': 8.723006842806576e-05, 'epoch': 0.58}) +[2026-06-15 17:24:51,284][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.08555638045072556, 'train/grad_norm_pre_clip': 3.0840556621551514, 'timing/data': 0.011477353982627392, 'timing/model': 1.4805615190416574, 'learning_rate/qwen_vl_interface': 8.720928522120419e-06, 'learning_rate/action_model': 8.720928522120418e-05, 'epoch': 0.58}) +[2026-06-15 17:24:57,278][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.497679740190506, 'train/grad_norm_pre_clip': 16.32044792175293, 'timing/data': 0.01217995211482048, 'timing/model': 1.4934486439451575, 'learning_rate/qwen_vl_interface': 8.718848774666655e-06, 'learning_rate/action_model': 8.718848774666655e-05, 'epoch': 0.58}) +[2026-06-15 17:25:03,257][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.23472478985786438, 'train/grad_norm_pre_clip': 10.088239669799805, 'timing/data': 0.011854496784508228, 'timing/model': 1.4684143085032701, 'learning_rate/qwen_vl_interface': 8.716767601300189e-06, 'learning_rate/action_model': 8.716767601300188e-05, 'epoch': 0.58}) +[2026-06-15 17:25:09,177][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.29464226961135864, 'train/grad_norm_pre_clip': 12.688339233398438, 'timing/data': 0.012509925290942192, 'timing/model': 1.4806759394705296, 'learning_rate/qwen_vl_interface': 8.714685002876514e-06, 'learning_rate/action_model': 8.714685002876513e-05, 'epoch': 0.58}) +[2026-06-15 17:25:15,049][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.176070436835289, 'train/grad_norm_pre_clip': 9.605242729187012, 'timing/data': 0.010248012840747833, 'timing/model': 1.4353818027302623, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.58}) +[2026-06-15 17:25:21,209][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.05547069013118744, 'train/grad_norm_pre_clip': 3.130507230758667, 'timing/data': 0.01059566717594862, 'timing/model': 1.8150786450132728, 'learning_rate/qwen_vl_interface': 8.710515534282428e-06, 'learning_rate/action_model': 8.710515534282428e-05, 'epoch': 0.58}) +[2026-06-15 17:25:26,851][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.13237378001213074, 'train/grad_norm_pre_clip': 6.120090007781982, 'timing/data': 0.010865888558328152, 'timing/model': 1.441847199574113, 'learning_rate/qwen_vl_interface': 8.708428665825931e-06, 'learning_rate/action_model': 8.708428665825931e-05, 'epoch': 0.58}) +[2026-06-15 17:25:32,849][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.19843590259552002, 'train/grad_norm_pre_clip': 5.655087471008301, 'timing/data': 0.00999146793037653, 'timing/model': 1.4291545636951923, 'learning_rate/qwen_vl_interface': 8.706340375740045e-06, 'learning_rate/action_model': 8.706340375740045e-05, 'epoch': 0.58}) +[2026-06-15 17:25:38,860][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.39151230454444885, 'train/grad_norm_pre_clip': 9.93272876739502, 'timing/data': 0.009911625646054745, 'timing/model': 1.480613254942, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 0.59}) +[2026-06-15 17:25:44,772][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.27956637740135193, 'train/grad_norm_pre_clip': 10.703937530517578, 'timing/data': 0.011917819269001484, 'timing/model': 1.483409226872027, 'learning_rate/qwen_vl_interface': 8.70215953411436e-06, 'learning_rate/action_model': 8.70215953411436e-05, 'epoch': 0.59}) +[2026-06-15 17:25:50,770][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.1620161533355713, 'train/grad_norm_pre_clip': 5.106261730194092, 'timing/data': 0.011875269003212452, 'timing/model': 1.5362966563552618, 'learning_rate/qwen_vl_interface': 8.70006698429315e-06, 'learning_rate/action_model': 8.700066984293149e-05, 'epoch': 0.59}) +[2026-06-15 17:25:56,696][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.1684097945690155, 'train/grad_norm_pre_clip': 4.442861080169678, 'timing/data': 0.011489913798868656, 'timing/model': 1.7083482705056667, 'learning_rate/qwen_vl_interface': 8.697973016279722e-06, 'learning_rate/action_model': 8.697973016279722e-05, 'epoch': 0.59}) +[2026-06-15 17:26:02,542][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.05273415893316269, 'train/grad_norm_pre_clip': 2.7903621196746826, 'timing/data': 0.01172979548573494, 'timing/model': 1.6420992463827133, 'learning_rate/qwen_vl_interface': 8.695877630934827e-06, 'learning_rate/action_model': 8.695877630934826e-05, 'epoch': 0.59}) +[2026-06-15 17:26:08,259][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.17789769172668457, 'train/grad_norm_pre_clip': 5.770702362060547, 'timing/data': 0.010854693129658699, 'timing/model': 1.465376103296876, 'learning_rate/qwen_vl_interface': 8.693780829119801e-06, 'learning_rate/action_model': 8.693780829119801e-05, 'epoch': 0.59}) +[2026-06-15 17:26:14,171][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.1861838847398758, 'train/grad_norm_pre_clip': 6.81939172744751, 'timing/data': 0.01077905111014843, 'timing/model': 1.4767399681732059, 'learning_rate/qwen_vl_interface': 8.69168261169656e-06, 'learning_rate/action_model': 8.691682611696559e-05, 'epoch': 0.59}) +[2026-06-15 17:26:20,064][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.6788988709449768, 'train/grad_norm_pre_clip': 15.108423233032227, 'timing/data': 0.011471039615571499, 'timing/model': 1.462831157259643, 'learning_rate/qwen_vl_interface': 8.689582979527599e-06, 'learning_rate/action_model': 8.689582979527599e-05, 'epoch': 0.59}) +[2026-06-15 17:26:25,966][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.23603813350200653, 'train/grad_norm_pre_clip': 23.98660659790039, 'timing/data': 0.011127994395792484, 'timing/model': 1.4377329433336854, 'learning_rate/qwen_vl_interface': 8.687481933476e-06, 'learning_rate/action_model': 8.687481933476e-05, 'epoch': 0.59}) +[2026-06-15 17:26:31,885][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.06856390833854675, 'train/grad_norm_pre_clip': 2.724541187286377, 'timing/data': 0.011678489856421947, 'timing/model': 1.6853867704048753, 'learning_rate/qwen_vl_interface': 8.685379474405423e-06, 'learning_rate/action_model': 8.685379474405423e-05, 'epoch': 0.59}) +[2026-06-15 17:26:37,806][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.07979307323694229, 'train/grad_norm_pre_clip': 3.2517342567443848, 'timing/data': 0.010861308313906193, 'timing/model': 1.7095793513581157, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.59}) +[2026-06-15 17:26:43,460][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.10076643526554108, 'train/grad_norm_pre_clip': 12.276113510131836, 'timing/data': 0.011141601018607616, 'timing/model': 1.4673233442008495, 'learning_rate/qwen_vl_interface': 8.681170320664883e-06, 'learning_rate/action_model': 8.681170320664883e-05, 'epoch': 0.59}) +[2026-06-15 17:26:49,378][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.2851160764694214, 'train/grad_norm_pre_clip': 8.649796485900879, 'timing/data': 0.012022671289741993, 'timing/model': 1.477207693271339, 'learning_rate/qwen_vl_interface': 8.679063627725143e-06, 'learning_rate/action_model': 8.679063627725144e-05, 'epoch': 0.59}) +[2026-06-15 17:26:55,370][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.14326369762420654, 'train/grad_norm_pre_clip': 5.130229949951172, 'timing/data': 0.010526212863624096, 'timing/model': 1.5170103507116437, 'learning_rate/qwen_vl_interface': 8.676955525226876e-06, 'learning_rate/action_model': 8.676955525226875e-05, 'epoch': 0.59}) +[2026-06-15 17:27:01,360][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.24581171572208405, 'train/grad_norm_pre_clip': 6.215177536010742, 'timing/data': 0.0110931983217597, 'timing/model': 1.5189660750329494, 'learning_rate/qwen_vl_interface': 8.674846014036636e-06, 'learning_rate/action_model': 8.674846014036636e-05, 'epoch': 0.59}) +[2026-06-15 17:27:07,304][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.03903535008430481, 'train/grad_norm_pre_clip': 3.090210437774658, 'timing/data': 0.01093667559325695, 'timing/model': 1.7072332622483373, 'learning_rate/qwen_vl_interface': 8.672735095021569e-06, 'learning_rate/action_model': 8.672735095021569e-05, 'epoch': 0.59}) +[2026-06-15 17:27:12,969][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.13440966606140137, 'train/grad_norm_pre_clip': 8.229926109313965, 'timing/data': 0.011916717514395714, 'timing/model': 1.4705458022654057, 'learning_rate/qwen_vl_interface': 8.670622769049393e-06, 'learning_rate/action_model': 8.670622769049392e-05, 'epoch': 0.59}) +[2026-06-15 17:27:18,857][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.19396713376045227, 'train/grad_norm_pre_clip': 6.785531044006348, 'timing/data': 0.01119242887943983, 'timing/model': 1.464291480369866, 'learning_rate/qwen_vl_interface': 8.668509036988406e-06, 'learning_rate/action_model': 8.668509036988405e-05, 'epoch': 0.59}) +[2026-06-15 17:27:24,760][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.2015591561794281, 'train/grad_norm_pre_clip': 6.925507545471191, 'timing/data': 0.011683972552418709, 'timing/model': 1.4674220560118556, 'learning_rate/qwen_vl_interface': 8.66639389970748e-06, 'learning_rate/action_model': 8.666393899707479e-05, 'epoch': 0.59}) +[2026-06-15 17:27:30,651][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.16331809759140015, 'train/grad_norm_pre_clip': 13.481170654296875, 'timing/data': 0.011062684468925, 'timing/model': 1.4308060351759195, 'learning_rate/qwen_vl_interface': 8.66427735807607e-06, 'learning_rate/action_model': 8.664277358076069e-05, 'epoch': 0.59}) +[2026-06-15 17:37:57,751][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.0871758908033371, 'train/grad_norm_pre_clip': 4.428122043609619, 'eval/loss': 0.09134675562381744, 'eval/flap_precision': 0.819047619047619, 'eval/flap_recall': 0.6254545454545455, 'eval/flap_f1': 0.709278350515464, 'eval/flap_rate_model': 0.058333333333333334, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.709278350515464, 'timing/data': 0.011027202941477299, 'timing/model': 1.491192446090281, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 0.59}) +[2026-06-15 17:37:57,752][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1300_state +[2026-06-15 17:37:57,758][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 17:38:05,942][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1300_state/model.safetensors +[2026-06-15 17:38:24,865][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1300_state/optimizer.bin +[2026-06-15 17:38:24,866][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1300_state/sampler.bin +[2026-06-15 17:38:24,866][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1300_state/sampler_1.bin +[2026-06-15 17:38:24,868][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1300_state/random_states_0.pkl +[2026-06-15 17:38:25,023][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-06-15 17:38:25,024][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 17:38:25,038][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 17:38:30,923][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.10670594871044159, 'train/grad_norm_pre_clip': 6.947803020477295, 'timing/data': 0.012009984813630581, 'timing/model': 1.4047129526734352, 'learning_rate/qwen_vl_interface': 8.660040065242499e-06, 'learning_rate/action_model': 8.660040065242498e-05, 'epoch': 0.59}) +[2026-06-15 17:38:36,754][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.19625228643417358, 'train/grad_norm_pre_clip': 8.705812454223633, 'timing/data': 0.011175497435033321, 'timing/model': 1.682727161794901, 'learning_rate/qwen_vl_interface': 8.657919315782127e-06, 'learning_rate/action_model': 8.657919315782127e-05, 'epoch': 0.6}) +[2026-06-15 17:38:42,663][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.12218368053436279, 'train/grad_norm_pre_clip': 4.621457099914551, 'timing/data': 0.011902064085006714, 'timing/model': 1.7007691580802202, 'learning_rate/qwen_vl_interface': 8.655797165454853e-06, 'learning_rate/action_model': 8.655797165454853e-05, 'epoch': 0.6}) +[2026-06-15 17:38:48,341][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.2189418226480484, 'train/grad_norm_pre_clip': 9.104151725769043, 'timing/data': 0.010510662570595741, 'timing/model': 1.4757802281528711, 'learning_rate/qwen_vl_interface': 8.653673615133013e-06, 'learning_rate/action_model': 8.653673615133013e-05, 'epoch': 0.6}) +[2026-06-15 17:38:54,240][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.22052904963493347, 'train/grad_norm_pre_clip': 7.494200229644775, 'timing/data': 0.012771169655025005, 'timing/model': 1.461449551396072, 'learning_rate/qwen_vl_interface': 8.651548665689519e-06, 'learning_rate/action_model': 8.651548665689518e-05, 'epoch': 0.6}) +[2026-06-15 17:39:00,242][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.42008689045906067, 'train/grad_norm_pre_clip': 10.699870109558105, 'timing/data': 0.011115706525743008, 'timing/model': 1.474002524279058, 'learning_rate/qwen_vl_interface': 8.649422317997855e-06, 'learning_rate/action_model': 8.649422317997856e-05, 'epoch': 0.6}) +[2026-06-15 17:39:06,234][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.26959630846977234, 'train/grad_norm_pre_clip': 21.34557342529297, 'timing/data': 0.011224038898944855, 'timing/model': 1.493970226496458, 'learning_rate/qwen_vl_interface': 8.647294572932082e-06, 'learning_rate/action_model': 8.647294572932082e-05, 'epoch': 0.6}) +[2026-06-15 17:39:12,050][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.1771373748779297, 'train/grad_norm_pre_clip': 4.174306869506836, 'timing/data': 0.01086911465972662, 'timing/model': 1.4484658455476165, 'learning_rate/qwen_vl_interface': 8.64516543136684e-06, 'learning_rate/action_model': 8.64516543136684e-05, 'epoch': 0.6}) +[2026-06-15 17:39:17,971][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.40132641792297363, 'train/grad_norm_pre_clip': 8.371227264404297, 'timing/data': 0.010932606644928455, 'timing/model': 1.6956346174702048, 'learning_rate/qwen_vl_interface': 8.643034894177335e-06, 'learning_rate/action_model': 8.643034894177335e-05, 'epoch': 0.6}) +[2026-06-15 17:39:23,642][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.2189251333475113, 'train/grad_norm_pre_clip': 8.976489067077637, 'timing/data': 0.021990982815623283, 'timing/model': 1.4506056923419237, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 0.6}) +[2026-06-15 17:39:29,552][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.10989805310964584, 'train/grad_norm_pre_clip': 3.1736960411071777, 'timing/data': 0.0108836954459548, 'timing/model': 1.4734026053920388, 'learning_rate/qwen_vl_interface': 8.638769636429245e-06, 'learning_rate/action_model': 8.638769636429245e-05, 'epoch': 0.6}) +[2026-06-15 17:39:35,445][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.1792272925376892, 'train/grad_norm_pre_clip': 5.9057111740112305, 'timing/data': 0.012004023417830467, 'timing/model': 1.4623635951429605, 'learning_rate/qwen_vl_interface': 8.636634917623946e-06, 'learning_rate/action_model': 8.636634917623944e-05, 'epoch': 0.6}) +[2026-06-15 17:39:41,354][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.1895749866962433, 'train/grad_norm_pre_clip': 5.255838871002197, 'timing/data': 0.010896733961999416, 'timing/model': 1.4215987529605627, 'learning_rate/qwen_vl_interface': 8.634498806700955e-06, 'learning_rate/action_model': 8.634498806700954e-05, 'epoch': 0.6}) +[2026-06-15 17:39:47,246][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.1650710552930832, 'train/grad_norm_pre_clip': 4.932929039001465, 'timing/data': 0.010913655161857605, 'timing/model': 1.448726906441152, 'learning_rate/qwen_vl_interface': 8.63236130453835e-06, 'learning_rate/action_model': 8.63236130453835e-05, 'epoch': 0.6}) +[2026-06-15 17:39:53,189][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.14987419545650482, 'train/grad_norm_pre_clip': 4.409094333648682, 'timing/data': 0.010492351837456226, 'timing/model': 1.6956124231219292, 'learning_rate/qwen_vl_interface': 8.630222412014777e-06, 'learning_rate/action_model': 8.630222412014776e-05, 'epoch': 0.6}) +[2026-06-15 17:39:58,849][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.2622210681438446, 'train/grad_norm_pre_clip': 7.379026412963867, 'timing/data': 0.011035675182938576, 'timing/model': 1.4566501565277576, 'learning_rate/qwen_vl_interface': 8.628082130009451e-06, 'learning_rate/action_model': 8.628082130009451e-05, 'epoch': 0.6}) +[2026-06-15 17:40:04,748][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.09376659989356995, 'train/grad_norm_pre_clip': 2.937830686569214, 'timing/data': 0.011091484688222408, 'timing/model': 1.461615176871419, 'learning_rate/qwen_vl_interface': 8.62594045940216e-06, 'learning_rate/action_model': 8.62594045940216e-05, 'epoch': 0.6}) +[2026-06-15 17:40:10,650][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.23047049343585968, 'train/grad_norm_pre_clip': 16.492042541503906, 'timing/data': 0.011955051682889462, 'timing/model': 1.463760239072144, 'learning_rate/qwen_vl_interface': 8.623797401073272e-06, 'learning_rate/action_model': 8.623797401073271e-05, 'epoch': 0.6}) +[2026-06-15 17:40:16,552][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.24084149301052094, 'train/grad_norm_pre_clip': 5.066170692443848, 'timing/data': 0.01175053883343935, 'timing/model': 1.4366635838523507, 'learning_rate/qwen_vl_interface': 8.62165295590371e-06, 'learning_rate/action_model': 8.621652955903709e-05, 'epoch': 0.6}) +[2026-06-15 17:40:22,454][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.3375018835067749, 'train/grad_norm_pre_clip': 6.52312707901001, 'timing/data': 0.01143848616629839, 'timing/model': 1.4444544250145555, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 0.6}) +[2026-06-15 17:40:28,397][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.10353759676218033, 'train/grad_norm_pre_clip': 2.7851579189300537, 'timing/data': 0.011214976198971272, 'timing/model': 1.7102940594777465, 'learning_rate/qwen_vl_interface': 8.61735990856914e-06, 'learning_rate/action_model': 8.617359908569139e-05, 'epoch': 0.6}) +[2026-06-15 17:40:34,057][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.08111445605754852, 'train/grad_norm_pre_clip': 2.1583549976348877, 'timing/data': 0.010401570238173008, 'timing/model': 1.4586116597056389, 'learning_rate/qwen_vl_interface': 8.615211308168842e-06, 'learning_rate/action_model': 8.61521130816884e-05, 'epoch': 0.6}) +[2026-06-15 17:40:39,952][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.27569711208343506, 'train/grad_norm_pre_clip': 6.912684917449951, 'timing/data': 0.010024646297097206, 'timing/model': 1.4255778072401881, 'learning_rate/qwen_vl_interface': 8.613061324457289e-06, 'learning_rate/action_model': 8.613061324457288e-05, 'epoch': 0.61}) +[2026-06-15 17:40:45,856][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.2362566739320755, 'train/grad_norm_pre_clip': 4.648072719573975, 'timing/data': 0.010882623493671417, 'timing/model': 1.4670559642836452, 'learning_rate/qwen_vl_interface': 8.610909958318262e-06, 'learning_rate/action_model': 8.610909958318261e-05, 'epoch': 0.61}) +[2026-06-15 17:40:51,949][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.2021723836660385, 'train/grad_norm_pre_clip': 4.243994235992432, 'timing/data': 0.011304725892841816, 'timing/model': 1.4612092236056924, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.61}) +[2026-06-15 17:40:57,850][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.1515139639377594, 'train/grad_norm_pre_clip': 3.3831305503845215, 'timing/data': 0.011070994660258293, 'timing/model': 1.431330993771553, 'learning_rate/qwen_vl_interface': 8.606603082295724e-06, 'learning_rate/action_model': 8.606603082295724e-05, 'epoch': 0.61}) +[2026-06-15 17:41:03,782][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.1887090504169464, 'train/grad_norm_pre_clip': 3.582134485244751, 'timing/data': 0.01039935927838087, 'timing/model': 1.6956682885065675, 'learning_rate/qwen_vl_interface': 8.60444757418261e-06, 'learning_rate/action_model': 8.604447574182609e-05, 'epoch': 0.61}) +[2026-06-15 17:41:09,460][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.14718855917453766, 'train/grad_norm_pre_clip': 4.992234706878662, 'timing/data': 0.011759978719055653, 'timing/model': 1.4723868584260345, 'learning_rate/qwen_vl_interface': 8.602290687182805e-06, 'learning_rate/action_model': 8.602290687182806e-05, 'epoch': 0.61}) +[2026-06-15 17:41:15,353][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.2805078327655792, 'train/grad_norm_pre_clip': 7.0536346435546875, 'timing/data': 0.011529706418514252, 'timing/model': 1.4644600469619036, 'learning_rate/qwen_vl_interface': 8.600132422182929e-06, 'learning_rate/action_model': 8.600132422182928e-05, 'epoch': 0.61}) +[2026-06-15 17:41:21,257][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.23491252958774567, 'train/grad_norm_pre_clip': 7.748836517333984, 'timing/data': 0.012130821123719215, 'timing/model': 1.4678384698927402, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 0.61}) +[2026-06-15 17:41:27,157][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.217807337641716, 'train/grad_norm_pre_clip': 10.199246406555176, 'timing/data': 0.011574327014386654, 'timing/model': 1.4671683656051755, 'learning_rate/qwen_vl_interface': 8.595811761732239e-06, 'learning_rate/action_model': 8.595811761732238e-05, 'epoch': 0.61}) +[2026-06-15 17:41:33,052][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.201597198843956, 'train/grad_norm_pre_clip': 4.433133125305176, 'timing/data': 0.011848937720060349, 'timing/model': 1.4288525134325027, 'learning_rate/qwen_vl_interface': 8.593649368057489e-06, 'learning_rate/action_model': 8.59364936805749e-05, 'epoch': 0.61}) +[2026-06-15 17:41:38,994][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.1239595115184784, 'train/grad_norm_pre_clip': 3.8740622997283936, 'timing/data': 0.010133577510714531, 'timing/model': 1.7003098167479038, 'learning_rate/qwen_vl_interface': 8.591485599934786e-06, 'learning_rate/action_model': 8.591485599934785e-05, 'epoch': 0.61}) +[2026-06-15 17:41:44,658][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.2034343183040619, 'train/grad_norm_pre_clip': 7.281185150146484, 'timing/data': 0.012150649912655354, 'timing/model': 1.4680424761027098, 'learning_rate/qwen_vl_interface': 8.58932045825357e-06, 'learning_rate/action_model': 8.589320458253569e-05, 'epoch': 0.61}) +[2026-06-15 17:41:50,573][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.16762016713619232, 'train/grad_norm_pre_clip': 6.202243804931641, 'timing/data': 0.011836107820272446, 'timing/model': 1.4748188629746437, 'learning_rate/qwen_vl_interface': 8.58715394390385e-06, 'learning_rate/action_model': 8.58715394390385e-05, 'epoch': 0.61}) +[2026-06-15 17:41:56,464][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.1352398842573166, 'train/grad_norm_pre_clip': 5.156443119049072, 'timing/data': 0.010834190994501114, 'timing/model': 1.4701655339449644, 'learning_rate/qwen_vl_interface': 8.5849860577762e-06, 'learning_rate/action_model': 8.5849860577762e-05, 'epoch': 0.61}) +[2026-06-15 17:44:25,522][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-15 17:44:25,541][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-15 17:44:27,264][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-15 17:44:28,015][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-15 17:44:28,016][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-15 17:44:28,016][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-15 17:44:28,016][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-15 17:44:28,016][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-15 17:44:28,016][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-15 17:44:28,016][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-15 17:44:28,016][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-15 17:44:28,016][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-15 17:44:28,016][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-15 17:44:28,017][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-15 17:44:28,017][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-15 17:44:28,017][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-15 17:44:28,017][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-15 17:44:28,017][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-15 17:44:28,017][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-15 17:44:28,017][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-15 17:44:28,017][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-15 17:44:28,018][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-15 17:44:28,018][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-15 17:44:28,018][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-15 17:44:28,018][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-15 17:44:28,018][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-15 17:44:28,018][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-15 17:44:28,019][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-15 17:44:28,019][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-15 17:44:28,019][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-15 17:44:28,019][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-15 17:44:28,019][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-15 17:44:28,019][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-15 17:44:28,019][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-15 17:44:28,074][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-15 17:44:28,120][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/config.full.yaml +[2026-06-15 17:44:28,129][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/config.yaml +[2026-06-15 17:44:28,130][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1300_state, steps: 1300 +[2026-06-15 17:44:30,158][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1300_state +[2026-06-15 17:44:31,560][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-15 17:44:44,781][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-15 17:44:44,781][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-15 17:44:44,781][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-15 17:44:44,784][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-15 17:44:46,399][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-15 17:44:47,948][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-15 17:44:47,949][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-06-15 17:44:47,949][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-15 17:44:47,949][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 4 +[2026-06-15 17:44:47,950][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-15 17:45:17,735][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.1455555260181427, 'train/grad_norm_pre_clip': 6.448828220367432, 'timing/data': 0.036882638931274414, 'timing/model': 1.174727282486856, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.59}) +[2026-06-15 17:45:21,946][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.10320430994033813, 'train/grad_norm_pre_clip': 4.038233757019043, 'timing/data': 0.0330179650336504, 'timing/model': 1.0755345299839973, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.6}) +[2026-06-15 17:45:26,002][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.16865088045597076, 'train/grad_norm_pre_clip': 6.963402271270752, 'timing/data': 0.025233255699276924, 'timing/model': 1.0406840573996305, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.6}) +[2026-06-15 17:45:30,230][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.2968168556690216, 'train/grad_norm_pre_clip': 10.702406883239746, 'timing/data': 0.028753837570548058, 'timing/model': 1.065079852938652, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.6}) +[2026-06-15 17:45:34,422][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.11113107949495316, 'train/grad_norm_pre_clip': 5.127621650695801, 'timing/data': 0.02688771579414606, 'timing/model': 1.1286728903651237, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.6}) +[2026-06-15 17:45:38,618][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.07985825836658478, 'train/grad_norm_pre_clip': 4.699594020843506, 'timing/data': 0.02946703229099512, 'timing/model': 1.0678117480129004, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.6}) +[2026-06-15 17:45:42,826][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.16426554322242737, 'train/grad_norm_pre_clip': 6.558910846710205, 'timing/data': 0.027165808714926243, 'timing/model': 1.0618278644979, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.6}) +[2026-06-15 17:45:47,009][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.11826024204492569, 'train/grad_norm_pre_clip': 4.4802045822143555, 'timing/data': 0.02726974803954363, 'timing/model': 1.1109024723991752, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.6}) +[2026-06-15 17:45:51,406][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.13242889940738678, 'train/grad_norm_pre_clip': 7.476485252380371, 'timing/data': 0.031763745471835136, 'timing/model': 1.0617346419021487, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.6}) +[2026-06-15 17:45:55,515][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.22714312374591827, 'train/grad_norm_pre_clip': 4.390109062194824, 'timing/data': 0.02671805489808321, 'timing/model': 1.054016737267375, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.6}) +[2026-06-15 17:45:59,608][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.3825972378253937, 'train/grad_norm_pre_clip': 9.473907470703125, 'timing/data': 0.032512789592146873, 'timing/model': 1.0451236786320806, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.6}) +[2026-06-15 17:46:03,703][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.19931386411190033, 'train/grad_norm_pre_clip': 9.539337158203125, 'timing/data': 0.025163634680211544, 'timing/model': 1.047181866131723, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.6}) +[2026-06-15 17:46:07,799][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.11347765475511551, 'train/grad_norm_pre_clip': 9.989418029785156, 'timing/data': 0.024495109915733337, 'timing/model': 1.0461394442245364, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.6}) +[2026-06-15 17:46:11,918][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.49322962760925293, 'train/grad_norm_pre_clip': 16.554222106933594, 'timing/data': 0.02484748139977455, 'timing/model': 1.0606416780501604, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.6}) +[2026-06-15 17:46:16,022][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.0911276787519455, 'train/grad_norm_pre_clip': 4.374572277069092, 'timing/data': 0.026774514466524124, 'timing/model': 1.0524725019931793, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.6}) +[2026-06-15 17:46:20,221][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.5182045102119446, 'train/grad_norm_pre_clip': 11.157549858093262, 'timing/data': 0.031061070039868355, 'timing/model': 1.1475812988355756, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.6}) +[2026-06-15 17:46:24,355][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.40319037437438965, 'train/grad_norm_pre_clip': 9.380009651184082, 'timing/data': 0.026106679812073708, 'timing/model': 1.078410499729216, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.6}) +[2026-06-15 17:46:28,445][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.4560355544090271, 'train/grad_norm_pre_clip': 14.212433815002441, 'timing/data': 0.02709770668298006, 'timing/model': 1.0691723879426718, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.6}) +[2026-06-15 17:46:32,642][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.16485470533370972, 'train/grad_norm_pre_clip': 6.7224884033203125, 'timing/data': 0.02958258055150509, 'timing/model': 1.066992038860917, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.6}) +[2026-06-15 17:46:36,806][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.049551017582416534, 'train/grad_norm_pre_clip': 2.4554197788238525, 'timing/data': 0.026990684680640697, 'timing/model': 1.126346806064248, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.6}) +[2026-06-15 17:46:40,936][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.21659314632415771, 'train/grad_norm_pre_clip': 17.324546813964844, 'timing/data': 0.026688761077821255, 'timing/model': 1.0636888397857547, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.6}) +[2026-06-15 17:46:45,034][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.05890842154622078, 'train/grad_norm_pre_clip': 3.7693607807159424, 'timing/data': 0.02507031336426735, 'timing/model': 1.0634331610053778, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.6}) +[2026-06-15 17:46:49,170][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.22488784790039062, 'train/grad_norm_pre_clip': 7.286509037017822, 'timing/data': 0.02696316409856081, 'timing/model': 1.067822272889316, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.61}) +[2026-06-15 17:46:53,316][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.3401470482349396, 'train/grad_norm_pre_clip': 7.012229919433594, 'timing/data': 0.024594909511506557, 'timing/model': 1.1315060760825872, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.61}) +[2026-06-15 17:46:57,443][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.18142814934253693, 'train/grad_norm_pre_clip': 5.046114921569824, 'timing/data': 0.027462417259812355, 'timing/model': 1.0655197193846107, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.61}) +[2026-06-15 17:47:01,615][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.0834108367562294, 'train/grad_norm_pre_clip': 4.955447673797607, 'timing/data': 0.030142774805426598, 'timing/model': 1.1306763784959912, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.61}) +[2026-06-15 17:47:05,754][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.11633428931236267, 'train/grad_norm_pre_clip': 5.476497173309326, 'timing/data': 0.026653772220015526, 'timing/model': 1.0806843899190426, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.61}) +[2026-06-15 17:47:09,844][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.3705143928527832, 'train/grad_norm_pre_clip': 14.661886215209961, 'timing/data': 0.025748026557266712, 'timing/model': 1.067733846604824, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.61}) +[2026-06-15 17:47:14,023][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.24841929972171783, 'train/grad_norm_pre_clip': 10.879861831665039, 'timing/data': 0.026482000946998596, 'timing/model': 1.1365579590201378, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.61}) +[2026-06-15 17:47:18,162][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.10044130682945251, 'train/grad_norm_pre_clip': 5.317884922027588, 'timing/data': 0.026475360617041588, 'timing/model': 1.0818735919892788, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.61}) +[2026-06-15 17:47:22,335][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.1421738564968109, 'train/grad_norm_pre_clip': 6.427025318145752, 'timing/data': 0.027992519550025463, 'timing/model': 1.1128020249307156, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.61}) +[2026-06-15 17:47:26,522][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.18742386996746063, 'train/grad_norm_pre_clip': 6.10653829574585, 'timing/data': 0.025416824035346508, 'timing/model': 1.126677873544395, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.61}) +[2026-06-15 17:47:30,737][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.07456117123365402, 'train/grad_norm_pre_clip': 2.7755520343780518, 'timing/data': 0.02570977620780468, 'timing/model': 1.148245919495821, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.61}) +[2026-06-15 17:47:34,926][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.18838009238243103, 'train/grad_norm_pre_clip': 5.721864223480225, 'timing/data': 0.025018351152539253, 'timing/model': 1.1400760347023606, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.61}) +[2026-06-15 17:47:39,126][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.28949832916259766, 'train/grad_norm_pre_clip': 11.800155639648438, 'timing/data': 0.025173312053084373, 'timing/model': 1.1392642753198743, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.61}) +[2026-06-15 17:47:43,423][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.271105021238327, 'train/grad_norm_pre_clip': 6.461379051208496, 'timing/data': 0.024987860582768917, 'timing/model': 1.137364337220788, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.61}) +[2026-06-15 17:47:47,634][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.33176809549331665, 'train/grad_norm_pre_clip': 8.175985336303711, 'timing/data': 0.025323032401502132, 'timing/model': 1.149488065391779, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.61}) +[2026-06-15 17:47:51,825][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.1193123608827591, 'train/grad_norm_pre_clip': 9.130265235900879, 'timing/data': 0.025117211043834686, 'timing/model': 1.136902516707778, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.61}) +[2026-06-15 17:47:56,024][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.10969913005828857, 'train/grad_norm_pre_clip': 4.408663749694824, 'timing/data': 0.02625647746026516, 'timing/model': 1.1329380366951227, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.61}) +[2026-06-15 17:48:00,214][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.07993462681770325, 'train/grad_norm_pre_clip': 4.730655193328857, 'timing/data': 0.023172308690845966, 'timing/model': 1.1305347569286823, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.61}) +[2026-06-15 17:48:04,425][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.09706901758909225, 'train/grad_norm_pre_clip': 4.225245475769043, 'timing/data': 0.02284846641123295, 'timing/model': 1.1349502792581916, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.61}) +[2026-06-15 17:48:08,622][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.32617899775505066, 'train/grad_norm_pre_clip': 10.041533470153809, 'timing/data': 0.02569094393402338, 'timing/model': 1.126482573337853, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.61}) +[2026-06-15 17:48:12,748][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.04245748743414879, 'train/grad_norm_pre_clip': 3.432871103286743, 'timing/data': 0.02421437483280897, 'timing/model': 1.0668902127072215, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.61}) +[2026-06-15 17:48:16,846][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.06646846979856491, 'train/grad_norm_pre_clip': 2.782320261001587, 'timing/data': 0.025419173762202263, 'timing/model': 1.0701257959008217, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.61}) +[2026-06-15 17:48:21,025][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.08680769801139832, 'train/grad_norm_pre_clip': 4.227365016937256, 'timing/data': 0.027925358153879642, 'timing/model': 1.1415795404464006, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.62}) +[2026-06-15 17:48:25,220][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.09082122892141342, 'train/grad_norm_pre_clip': 3.116079092025757, 'timing/data': 0.026278827339410782, 'timing/model': 1.1336421286687255, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.62}) +[2026-06-15 17:48:29,351][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.1874048113822937, 'train/grad_norm_pre_clip': 8.15105152130127, 'timing/data': 0.02492555882781744, 'timing/model': 1.069486708380282, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.62}) +[2026-06-15 17:48:33,535][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.26307275891304016, 'train/grad_norm_pre_clip': 8.012662887573242, 'timing/data': 0.024783768691122532, 'timing/model': 1.1461621634662151, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.62}) +[2026-06-15 17:48:37,723][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.11485555768013, 'train/grad_norm_pre_clip': 4.71742582321167, 'timing/data': 0.025540122762322426, 'timing/model': 1.1196572305634618, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.62}) +[2026-06-15 17:48:41,826][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.1125420406460762, 'train/grad_norm_pre_clip': 5.063821315765381, 'timing/data': 0.025591622106730938, 'timing/model': 1.0702061112970114, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.62}) +[2026-06-15 17:48:45,933][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.3028602600097656, 'train/grad_norm_pre_clip': 11.987936019897461, 'timing/data': 0.025213860906660557, 'timing/model': 1.0626889718696475, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.62}) +[2026-06-15 17:48:50,031][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.18812793493270874, 'train/grad_norm_pre_clip': 6.9541826248168945, 'timing/data': 0.025207790546119213, 'timing/model': 1.0630810409784317, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.62}) +[2026-06-15 17:48:54,132][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.11506152153015137, 'train/grad_norm_pre_clip': 7.794532775878906, 'timing/data': 0.025083440355956554, 'timing/model': 1.0614306759089231, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.62}) +[2026-06-15 17:48:58,234][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.3285336494445801, 'train/grad_norm_pre_clip': 11.192892074584961, 'timing/data': 0.025060079991817474, 'timing/model': 1.0620411662384868, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.62}) +[2026-06-15 17:49:02,336][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.06861022859811783, 'train/grad_norm_pre_clip': 7.71392822265625, 'timing/data': 0.0245412765070796, 'timing/model': 1.0604786816984415, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.62}) +[2026-06-15 17:49:06,626][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.09042885899543762, 'train/grad_norm_pre_clip': 4.935074329376221, 'timing/data': 0.026338846422731876, 'timing/model': 1.1311627682298422, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.62}) +[2026-06-15 17:49:10,786][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.3116338849067688, 'train/grad_norm_pre_clip': 10.963139533996582, 'timing/data': 0.026498346589505672, 'timing/model': 1.1029247669503093, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.62}) +[2026-06-15 17:49:15,021][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.20554116368293762, 'train/grad_norm_pre_clip': 16.004928588867188, 'timing/data': 0.026358275674283504, 'timing/model': 1.1358747789636254, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.62}) +[2026-06-15 17:49:19,220][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.02514035254716873, 'train/grad_norm_pre_clip': 1.9764113426208496, 'timing/data': 0.026568028144538403, 'timing/model': 1.1250098599120975, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.62}) +[2026-06-15 17:49:23,443][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.5436567068099976, 'train/grad_norm_pre_clip': 18.454200744628906, 'timing/data': 0.030290650203824043, 'timing/model': 1.1548230918124318, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.62}) +[2026-06-15 17:49:27,625][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.162014901638031, 'train/grad_norm_pre_clip': 11.639969825744629, 'timing/data': 0.025437550619244576, 'timing/model': 1.1342474659904838, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.62}) +[2026-06-15 17:49:32,026][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.2638494372367859, 'train/grad_norm_pre_clip': 11.31531810760498, 'timing/data': 0.02455267496407032, 'timing/model': 1.1417572367936373, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.62}) +[2026-06-15 17:49:36,231][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.22028714418411255, 'train/grad_norm_pre_clip': 6.823307037353516, 'timing/data': 0.02556606102734804, 'timing/model': 1.1329525206238031, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.62}) +[2026-06-15 17:49:40,431][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.2546883821487427, 'train/grad_norm_pre_clip': 8.667791366577148, 'timing/data': 0.0250915689393878, 'timing/model': 1.1446749353781343, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.62}) +[2026-06-15 17:49:44,571][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.21381215751171112, 'train/grad_norm_pre_clip': 12.85947036743164, 'timing/data': 0.024295654147863388, 'timing/model': 1.08783467579633, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.62}) +[2026-06-15 17:49:48,720][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.24884657561779022, 'train/grad_norm_pre_clip': 8.40626335144043, 'timing/data': 0.025384100154042244, 'timing/model': 1.1133552035316825, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.62}) +[2026-06-15 17:49:52,858][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.02737911231815815, 'train/grad_norm_pre_clip': 2.5371615886688232, 'timing/data': 0.025521190837025642, 'timing/model': 1.0766650019213557, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.63}) +[2026-06-15 17:49:57,017][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.11481643468141556, 'train/grad_norm_pre_clip': 3.7429914474487305, 'timing/data': 0.024857375770807266, 'timing/model': 1.130998009815812, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.63}) +[2026-06-15 17:50:01,227][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.07242545485496521, 'train/grad_norm_pre_clip': 5.329895973205566, 'timing/data': 0.025079027749598026, 'timing/model': 1.1325972061604261, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.63}) +[2026-06-15 17:50:05,418][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.1966891884803772, 'train/grad_norm_pre_clip': 5.403683662414551, 'timing/data': 0.02466450445353985, 'timing/model': 1.119450713507831, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.63}) +[2026-06-15 17:50:09,619][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.22871705889701843, 'train/grad_norm_pre_clip': 6.799567699432373, 'timing/data': 0.02626972459256649, 'timing/model': 1.1171670369803905, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.63}) +[2026-06-15 17:50:13,769][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.15091226994991302, 'train/grad_norm_pre_clip': 9.061713218688965, 'timing/data': 0.024773196317255497, 'timing/model': 1.0905086360871792, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.63}) +[2026-06-15 17:50:17,932][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.2695210576057434, 'train/grad_norm_pre_clip': 8.829099655151367, 'timing/data': 0.025819442234933376, 'timing/model': 1.1433669244870543, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.63}) +[2026-06-15 17:50:22,126][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.3914824426174164, 'train/grad_norm_pre_clip': 6.826219081878662, 'timing/data': 0.02603746298700571, 'timing/model': 1.129734680056572, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.63}) +[2026-06-15 17:50:26,331][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.14342717826366425, 'train/grad_norm_pre_clip': 5.960587978363037, 'timing/data': 0.025122297927737236, 'timing/model': 1.0887949848547578, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.63}) +[2026-06-15 17:50:30,528][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.27928420901298523, 'train/grad_norm_pre_clip': 4.814170837402344, 'timing/data': 0.029534803703427315, 'timing/model': 1.1266687540337443, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.63}) +[2026-06-15 17:50:34,726][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.27581578493118286, 'train/grad_norm_pre_clip': 5.591995716094971, 'timing/data': 0.02374227996915579, 'timing/model': 1.1228486578911543, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.63}) +[2026-06-15 17:50:38,924][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.1974078267812729, 'train/grad_norm_pre_clip': 6.595751762390137, 'timing/data': 0.026382415555417538, 'timing/model': 1.134759915061295, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.63}) +[2026-06-15 17:50:43,121][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.13236524164676666, 'train/grad_norm_pre_clip': 3.2241928577423096, 'timing/data': 0.025917132385075092, 'timing/model': 1.133094190619886, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.63}) +[2026-06-15 17:50:47,337][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.1944015920162201, 'train/grad_norm_pre_clip': 5.636590957641602, 'timing/data': 0.02579626254737377, 'timing/model': 1.142306531779468, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.63}) +[2026-06-15 17:50:51,552][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.28924623131752014, 'train/grad_norm_pre_clip': 7.750886917114258, 'timing/data': 0.032037338241934776, 'timing/model': 1.0739840781316161, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.63}) +[2026-06-15 17:50:55,741][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.1333702951669693, 'train/grad_norm_pre_clip': 4.198946952819824, 'timing/data': 0.026674475520849228, 'timing/model': 1.1218597749248147, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.63}) +[2026-06-15 17:51:00,035][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.06438204646110535, 'train/grad_norm_pre_clip': 3.5179269313812256, 'timing/data': 0.026895007118582726, 'timing/model': 1.1405933918431401, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.63}) +[2026-06-15 17:51:04,306][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.2001780867576599, 'train/grad_norm_pre_clip': 11.24510669708252, 'timing/data': 0.025489259511232376, 'timing/model': 1.12533587962389, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.63}) +[2026-06-15 17:51:08,616][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.25017955899238586, 'train/grad_norm_pre_clip': 7.055283069610596, 'timing/data': 0.02818162553012371, 'timing/model': 1.1242300188168883, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.63}) +[2026-06-15 17:51:12,828][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.03861319273710251, 'train/grad_norm_pre_clip': 4.0918869972229, 'timing/data': 0.025698890909552574, 'timing/model': 1.133739179931581, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.63}) +[2026-06-15 17:51:17,023][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.2923840880393982, 'train/grad_norm_pre_clip': 11.677668571472168, 'timing/data': 0.02556944079697132, 'timing/model': 1.135911149904132, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.63}) +[2026-06-15 17:51:21,136][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.4566855728626251, 'train/grad_norm_pre_clip': 8.945389747619629, 'timing/data': 0.025433518923819065, 'timing/model': 1.0633816933259368, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.63}) +[2026-06-15 17:51:25,416][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.11958352476358414, 'train/grad_norm_pre_clip': 3.8103671073913574, 'timing/data': 0.025376567617058754, 'timing/model': 1.12300730869174, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.64}) +[2026-06-15 17:51:29,536][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.057435788214206696, 'train/grad_norm_pre_clip': 3.1818346977233887, 'timing/data': 0.024795583449304104, 'timing/model': 1.0639311810955405, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.64}) +[2026-06-15 17:51:33,635][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.1652127504348755, 'train/grad_norm_pre_clip': 7.591551780700684, 'timing/data': 0.0250953771173954, 'timing/model': 1.063735325820744, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.64}) +[2026-06-15 17:51:37,925][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.10295744985342026, 'train/grad_norm_pre_clip': 4.072078704833984, 'timing/data': 0.03229751903563738, 'timing/model': 1.1134460624307394, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.64}) +[2026-06-15 17:51:42,035][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.05466658994555473, 'train/grad_norm_pre_clip': 5.523513317108154, 'timing/data': 0.026104762218892574, 'timing/model': 1.0639958418905735, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.64}) +[2026-06-15 17:51:46,138][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.15684205293655396, 'train/grad_norm_pre_clip': 8.49359130859375, 'timing/data': 0.028115173801779747, 'timing/model': 1.0647274823859334, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.64}) +[2026-06-15 17:51:50,237][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.28783923387527466, 'train/grad_norm_pre_clip': 15.551942825317383, 'timing/data': 0.025737520307302475, 'timing/model': 1.0634376425296068, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.64}) +[2026-06-15 17:51:54,341][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.10766997188329697, 'train/grad_norm_pre_clip': 9.090984344482422, 'timing/data': 0.030832160264253616, 'timing/model': 1.0661936858668923, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.64}) +[2026-06-15 17:51:58,436][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.34372448921203613, 'train/grad_norm_pre_clip': 12.716196060180664, 'timing/data': 0.025632219389081, 'timing/model': 1.0654555186629295, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.64}) +[2026-06-15 17:52:02,536][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.251564085483551, 'train/grad_norm_pre_clip': 9.590807914733887, 'timing/data': 0.025654548779129982, 'timing/model': 1.0637038359418511, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.64}) +[2026-06-15 17:52:06,635][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.25718334317207336, 'train/grad_norm_pre_clip': 6.6871795654296875, 'timing/data': 0.02721259742975235, 'timing/model': 1.0636596633121371, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.64}) +[2026-06-15 17:54:54,461][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp_ud9qjjc/test.c -o /tmp/tmp_ud9qjjc/test.o +[2026-06-15 17:54:54,493][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp_ud9qjjc/test.o -laio -o /tmp/tmp_ud9qjjc/a.out +[2026-06-15 17:54:55,031][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpqbe7z782/test.c -o /tmp/tmpqbe7z782/test.o +[2026-06-15 17:54:55,055][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpqbe7z782/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpqbe7z782/a.out +[2026-06-15 17:54:55,104][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp6aa2h4qf/test.c -o /tmp/tmp6aa2h4qf/test.o +[2026-06-15 17:54:55,134][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp6aa2h4qf/test.o -laio -o /tmp/tmp6aa2h4qf/a.out +[2026-06-15 18:03:41,108][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.21345120668411255, 'train/grad_norm_pre_clip': 9.914684295654297, 'eval/loss': 0.08927440643310547, 'eval/flap_precision': 0.7222222222222222, 'eval/flap_recall': 0.7563636363636363, 'eval/flap_f1': 0.7388987566607459, 'eval/flap_rate_model': 0.08, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7388987566607459, 'timing/data': 0.025611529126763344, 'timing/model': 1.065005167387426, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.64}) +[2026-06-15 18:03:41,109][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1400_state +[2026-06-15 18:03:41,117][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 18:03:49,133][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1400_state/model.safetensors +[2026-06-15 18:04:00,096][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1400_state/optimizer.bin +[2026-06-15 18:04:00,097][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1400_state/sampler.bin +[2026-06-15 18:04:00,097][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1400_state/sampler_1.bin +[2026-06-15 18:04:00,099][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1400_state/random_states_0.pkl +[2026-06-15 18:04:00,267][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-06-15 18:04:00,268][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 18:04:00,284][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 18:04:06,021][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.034769631922245026, 'train/grad_norm_pre_clip': 2.7695107460021973, 'timing/data': 0.03484870865941048, 'timing/model': 1.3973416425287724, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 0.64}) +[2026-06-15 18:04:11,537][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.1098412498831749, 'train/grad_norm_pre_clip': 2.799304723739624, 'timing/data': 0.04072106443345547, 'timing/model': 1.3755336329340935, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 0.64}) +[2026-06-15 18:04:17,103][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.10753398388624191, 'train/grad_norm_pre_clip': 4.094087600708008, 'timing/data': 0.03334989584982395, 'timing/model': 1.5976954968646169, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 0.64}) +[2026-06-15 18:04:22,534][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.22088104486465454, 'train/grad_norm_pre_clip': 9.256425857543945, 'timing/data': 0.03250104282051325, 'timing/model': 1.4178950767964125, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 0.64}) +[2026-06-15 18:04:28,228][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.26916658878326416, 'train/grad_norm_pre_clip': 12.609996795654297, 'timing/data': 0.030587468296289444, 'timing/model': 1.4150072410702705, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 0.64}) +[2026-06-15 18:04:33,942][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.2293497771024704, 'train/grad_norm_pre_clip': 14.248473167419434, 'timing/data': 0.030789309181272984, 'timing/model': 1.4515556190162897, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 0.64}) +[2026-06-15 18:04:39,717][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.2482639104127884, 'train/grad_norm_pre_clip': 14.219658851623535, 'timing/data': 0.04059844929724932, 'timing/model': 1.444616791792214, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 0.64}) +[2026-06-15 18:04:45,463][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.17624303698539734, 'train/grad_norm_pre_clip': 4.940842628479004, 'timing/data': 0.030666280537843704, 'timing/model': 1.6699827518314123, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 0.64}) +[2026-06-15 18:04:50,940][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.3290373384952545, 'train/grad_norm_pre_clip': 12.1242036819458, 'timing/data': 0.03007259964942932, 'timing/model': 1.440680893138051, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 0.64}) +[2026-06-15 18:04:56,641][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.22252456843852997, 'train/grad_norm_pre_clip': 4.747979640960693, 'timing/data': 0.030328800901770592, 'timing/model': 1.4283197382465005, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.64}) +[2026-06-15 18:05:02,236][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.18736708164215088, 'train/grad_norm_pre_clip': 5.719137191772461, 'timing/data': 0.03147771488875151, 'timing/model': 1.3987946975976229, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 0.65}) +[2026-06-15 18:05:07,942][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.21798714995384216, 'train/grad_norm_pre_clip': 7.350822448730469, 'timing/data': 0.031175104901194572, 'timing/model': 1.4432578515261412, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 0.65}) +[2026-06-15 18:05:13,657][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.25153228640556335, 'train/grad_norm_pre_clip': 5.921345233917236, 'timing/data': 0.032178498804569244, 'timing/model': 1.417037082836032, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 0.65}) +[2026-06-15 18:05:19,376][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.22191891074180603, 'train/grad_norm_pre_clip': 4.735966682434082, 'timing/data': 0.030155522748827934, 'timing/model': 1.6505112536251545, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 0.65}) +[2026-06-15 18:05:25,126][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.1935657411813736, 'train/grad_norm_pre_clip': 8.409430503845215, 'timing/data': 0.036989035084843636, 'timing/model': 1.6055767675861716, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 0.65}) +[2026-06-15 18:05:30,646][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.1435219794511795, 'train/grad_norm_pre_clip': 4.315568447113037, 'timing/data': 0.0345759391784668, 'timing/model': 1.4326788485050201, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 0.65}) +[2026-06-15 18:05:36,351][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.20515938103199005, 'train/grad_norm_pre_clip': 4.787436008453369, 'timing/data': 0.029591962695121765, 'timing/model': 1.4298167489469051, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 0.65}) +[2026-06-15 18:05:42,053][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.20512311160564423, 'train/grad_norm_pre_clip': 7.092223644256592, 'timing/data': 0.03082312736660242, 'timing/model': 1.4460097681730986, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 0.65}) +[2026-06-15 18:05:47,759][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.13293364644050598, 'train/grad_norm_pre_clip': 5.957486629486084, 'timing/data': 0.02951501403003931, 'timing/model': 1.4090264877304435, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 0.65}) +[2026-06-15 18:05:53,474][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.42698967456817627, 'train/grad_norm_pre_clip': 15.692453384399414, 'timing/data': 0.03509377222508192, 'timing/model': 1.670279324054718, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.65}) +[2026-06-15 18:05:59,199][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.2715844213962555, 'train/grad_norm_pre_clip': 7.637933254241943, 'timing/data': 0.030397897586226463, 'timing/model': 1.635357093065977, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 0.65}) +[2026-06-15 18:06:04,741][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.11593025922775269, 'train/grad_norm_pre_clip': 6.384709358215332, 'timing/data': 0.03010555822402239, 'timing/model': 1.4270609272643924, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 0.65}) +[2026-06-15 18:06:10,444][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.22147098183631897, 'train/grad_norm_pre_clip': 6.680549144744873, 'timing/data': 0.03364531975239515, 'timing/model': 1.4169479012489319, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 0.65}) +[2026-06-15 18:06:16,145][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.13729055225849152, 'train/grad_norm_pre_clip': 4.846645355224609, 'timing/data': 0.03443524334579706, 'timing/model': 1.424085990525782, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 0.65}) +[2026-06-15 18:06:21,862][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.286440372467041, 'train/grad_norm_pre_clip': 5.490434169769287, 'timing/data': 0.030749072320759296, 'timing/model': 1.3931668745353818, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 0.65}) +[2026-06-15 18:06:27,681][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.1978829801082611, 'train/grad_norm_pre_clip': 4.336527347564697, 'timing/data': 0.03159210458397865, 'timing/model': 1.6825307616963983, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 0.65}) +[2026-06-15 18:06:33,244][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.24636414647102356, 'train/grad_norm_pre_clip': 4.889640808105469, 'timing/data': 0.03334165085107088, 'timing/model': 1.4277202412486076, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 0.65}) +[2026-06-15 18:06:39,055][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.28119611740112305, 'train/grad_norm_pre_clip': 7.8264265060424805, 'timing/data': 0.034935588017106056, 'timing/model': 1.421642835251987, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 0.65}) +[2026-06-15 18:06:44,858][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.169037863612175, 'train/grad_norm_pre_clip': 5.155166149139404, 'timing/data': 0.032698459923267365, 'timing/model': 1.4292619470506907, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 0.65}) +[2026-06-15 18:06:51,146][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.22411970794200897, 'train/grad_norm_pre_clip': 4.661202907562256, 'timing/data': 0.03194266837090254, 'timing/model': 1.406038655899465, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.65}) +[2026-06-15 18:06:56,769][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.07916382700204849, 'train/grad_norm_pre_clip': 2.669060468673706, 'timing/data': 0.029992692172527313, 'timing/model': 1.3949099509045482, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 0.65}) +[2026-06-15 18:07:02,580][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.1758548766374588, 'train/grad_norm_pre_clip': 3.1358561515808105, 'timing/data': 0.031201296485960484, 'timing/model': 1.6600818755105138, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 0.65}) +[2026-06-15 18:07:08,150][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.4888153672218323, 'train/grad_norm_pre_clip': 26.432865142822266, 'timing/data': 0.0320200202986598, 'timing/model': 1.4428570829331875, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 0.66}) +[2026-06-15 18:07:13,955][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.8693458437919617, 'train/grad_norm_pre_clip': 22.84421157836914, 'timing/data': 0.032485902309417725, 'timing/model': 1.4256360474973917, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 0.66}) +[2026-06-15 18:07:19,756][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.1632189303636551, 'train/grad_norm_pre_clip': 6.419399738311768, 'timing/data': 0.030278285965323448, 'timing/model': 1.451943352818489, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.66}) +[2026-06-15 18:07:25,565][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.5311509370803833, 'train/grad_norm_pre_clip': 23.7280330657959, 'timing/data': 0.03345360700041056, 'timing/model': 1.400695789605379, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 0.66}) +[2026-06-15 18:07:31,361][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.8211220502853394, 'train/grad_norm_pre_clip': 21.023143768310547, 'timing/data': 0.02984460536390543, 'timing/model': 1.401971135288477, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 0.66}) +[2026-06-15 18:07:37,189][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.22078607976436615, 'train/grad_norm_pre_clip': 8.18843936920166, 'timing/data': 0.030337397009134293, 'timing/model': 1.664210733026266, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 0.66}) +[2026-06-15 18:07:42,762][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.19207964837551117, 'train/grad_norm_pre_clip': 5.741947174072266, 'timing/data': 0.030157175846397877, 'timing/model': 1.4471841296181083, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 0.66}) +[2026-06-15 18:07:48,558][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.17202536761760712, 'train/grad_norm_pre_clip': 4.362994194030762, 'timing/data': 0.030100016854703426, 'timing/model': 1.4422635659575462, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.66}) +[2026-06-15 18:07:54,358][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.4218469262123108, 'train/grad_norm_pre_clip': 8.265639305114746, 'timing/data': 0.03111130092293024, 'timing/model': 1.4452995294705033, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 0.66}) +[2026-06-15 18:08:00,054][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.47097671031951904, 'train/grad_norm_pre_clip': 7.705207824707031, 'timing/data': 0.03142769169062376, 'timing/model': 1.420511650852859, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 0.66}) +[2026-06-15 18:08:05,843][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.23363640904426575, 'train/grad_norm_pre_clip': 5.65110969543457, 'timing/data': 0.031943025067448616, 'timing/model': 1.4386551473289728, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 0.66}) +[2026-06-15 18:08:11,595][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.2392684370279312, 'train/grad_norm_pre_clip': 4.1506123542785645, 'timing/data': 0.03392952308058739, 'timing/model': 1.6755301663652062, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 0.66}) +[2026-06-15 18:08:17,257][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.2463572770357132, 'train/grad_norm_pre_clip': 4.218152046203613, 'timing/data': 0.029856598004698753, 'timing/model': 1.4305621515959501, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 0.66}) +[2026-06-15 18:08:23,057][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.20233649015426636, 'train/grad_norm_pre_clip': 7.095109462738037, 'timing/data': 0.03179032634943724, 'timing/model': 1.4396597249433398, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 0.66}) +[2026-06-15 18:08:28,860][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.2348543107509613, 'train/grad_norm_pre_clip': 7.132994174957275, 'timing/data': 0.03133160527795553, 'timing/model': 1.4321746993809938, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 0.66}) +[2026-06-15 18:08:34,652][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.24372684955596924, 'train/grad_norm_pre_clip': 2.7427704334259033, 'timing/data': 0.03123654518276453, 'timing/model': 1.4339047158136964, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 0.66}) +[2026-06-15 18:08:40,348][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.25468355417251587, 'train/grad_norm_pre_clip': 3.427166223526001, 'timing/data': 0.03064591344445944, 'timing/model': 1.4034889340400696, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.66}) +[2026-06-15 18:08:46,081][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.1780029982328415, 'train/grad_norm_pre_clip': 6.741818904876709, 'timing/data': 0.033171082846820354, 'timing/model': 1.671319805085659, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.66}) +[2026-06-15 18:08:51,610][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.16514764726161957, 'train/grad_norm_pre_clip': 3.6157445907592773, 'timing/data': 0.031735568307340145, 'timing/model': 1.4147403733804822, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 0.66}) +[2026-06-15 18:08:57,370][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.35387107729911804, 'train/grad_norm_pre_clip': 6.107701778411865, 'timing/data': 0.030181871727108955, 'timing/model': 1.4264891678467393, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 0.66}) +[2026-06-15 18:09:03,158][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.3465976119041443, 'train/grad_norm_pre_clip': 7.89787483215332, 'timing/data': 0.029823821038007736, 'timing/model': 1.4440260715782642, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 0.66}) +[2026-06-15 18:09:08,964][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.16806277632713318, 'train/grad_norm_pre_clip': 5.845154762268066, 'timing/data': 0.03112130705267191, 'timing/model': 1.3947314759716392, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 0.66}) +[2026-06-15 18:09:14,646][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.09460315108299255, 'train/grad_norm_pre_clip': 3.301938772201538, 'timing/data': 0.0294331107288599, 'timing/model': 1.4044527309015393, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 0.67}) +[2026-06-15 18:09:20,518][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.13649949431419373, 'train/grad_norm_pre_clip': 3.713829278945923, 'timing/data': 0.03048695530742407, 'timing/model': 1.693170903250575, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 0.67}) +[2026-06-15 18:09:26,165][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.23204746842384338, 'train/grad_norm_pre_clip': 3.4983582496643066, 'timing/data': 0.031279729679226875, 'timing/model': 1.447992063127458, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 0.67}) +[2026-06-15 18:09:31,968][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.11339706927537918, 'train/grad_norm_pre_clip': 5.320521831512451, 'timing/data': 0.029731222428381443, 'timing/model': 1.4494308475404978, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 0.67}) +[2026-06-15 18:09:37,757][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.31109291315078735, 'train/grad_norm_pre_clip': 4.773228645324707, 'timing/data': 0.032118053175508976, 'timing/model': 1.4273617304861546, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 0.67}) +[2026-06-15 18:09:43,456][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.21780358254909515, 'train/grad_norm_pre_clip': 4.17092227935791, 'timing/data': 0.03001323528587818, 'timing/model': 1.418576174415648, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.67}) +[2026-06-15 18:09:49,342][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.148323193192482, 'train/grad_norm_pre_clip': 5.330141067504883, 'timing/data': 0.032151173800230026, 'timing/model': 1.4190497947856784, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 0.67}) +[2026-06-15 18:09:55,200][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.20273511111736298, 'train/grad_norm_pre_clip': 7.091076374053955, 'timing/data': 0.031269741244614124, 'timing/model': 1.678785065189004, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 0.67}) +[2026-06-15 18:10:00,654][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.17882592976093292, 'train/grad_norm_pre_clip': 5.356076717376709, 'timing/data': 0.02827636059373617, 'timing/model': 1.4164018975570798, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 0.67}) +[2026-06-15 18:10:06,669][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.06864390522241592, 'train/grad_norm_pre_clip': 3.3256611824035645, 'timing/data': 0.027526366524398327, 'timing/model': 1.4230302022770047, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 0.67}) +[2026-06-15 18:10:12,461][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.18438482284545898, 'train/grad_norm_pre_clip': 7.222885608673096, 'timing/data': 0.03030513785779476, 'timing/model': 1.451557100750506, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 0.67}) +[2026-06-15 18:10:18,283][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.2330653816461563, 'train/grad_norm_pre_clip': 9.588193893432617, 'timing/data': 0.03026636876165867, 'timing/model': 1.4681623922660947, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 0.67}) +[2026-06-15 18:10:24,174][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.11318972706794739, 'train/grad_norm_pre_clip': 3.789494276046753, 'timing/data': 0.03116491250693798, 'timing/model': 1.4048001915216446, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 0.67}) +[2026-06-15 18:10:29,986][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.2872554659843445, 'train/grad_norm_pre_clip': 7.353522300720215, 'timing/data': 0.03175750561058521, 'timing/model': 1.6538920374587178, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 0.67}) +[2026-06-15 18:10:35,556][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.2104061394929886, 'train/grad_norm_pre_clip': 8.90844440460205, 'timing/data': 0.03070142213255167, 'timing/model': 1.4177601132541895, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 0.67}) +[2026-06-15 18:10:41,361][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.1501278281211853, 'train/grad_norm_pre_clip': 13.401897430419922, 'timing/data': 0.03439359553158283, 'timing/model': 1.4406252494081855, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.67}) +[2026-06-15 18:10:47,146][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.17432604730129242, 'train/grad_norm_pre_clip': 5.6995697021484375, 'timing/data': 0.030131670646369457, 'timing/model': 1.428479528054595, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 0.67}) +[2026-06-15 18:10:52,960][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.15770462155342102, 'train/grad_norm_pre_clip': 5.650923728942871, 'timing/data': 0.03572753258049488, 'timing/model': 1.4453569110482931, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 0.67}) +[2026-06-15 18:10:58,749][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.17233522236347198, 'train/grad_norm_pre_clip': 6.583388805389404, 'timing/data': 0.029464857652783394, 'timing/model': 1.4108252776786685, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 0.67}) +[2026-06-15 18:11:04,490][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.18077674508094788, 'train/grad_norm_pre_clip': 7.965919017791748, 'timing/data': 0.029193815775215626, 'timing/model': 1.6859544841572642, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 0.67}) +[2026-06-15 18:11:10,051][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.3896424174308777, 'train/grad_norm_pre_clip': 9.728039741516113, 'timing/data': 0.03031793050467968, 'timing/model': 1.4416236206889153, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 0.67}) +[2026-06-15 18:11:15,848][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.13539837300777435, 'train/grad_norm_pre_clip': 11.286383628845215, 'timing/data': 0.030177521519362926, 'timing/model': 1.428645085543394, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 0.68}) +[2026-06-15 18:11:21,561][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.3136039674282074, 'train/grad_norm_pre_clip': 9.474935531616211, 'timing/data': 0.03478198032826185, 'timing/model': 1.4429481364786625, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 0.68}) +[2026-06-15 18:11:27,356][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.13114707171916962, 'train/grad_norm_pre_clip': 6.625801086425781, 'timing/data': 0.031766428612172604, 'timing/model': 1.4095213869586587, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 0.68}) +[2026-06-15 18:11:33,059][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.18500939011573792, 'train/grad_norm_pre_clip': 6.168264865875244, 'timing/data': 0.029691840521991253, 'timing/model': 1.4011575868353248, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 0.68}) +[2026-06-15 18:11:38,879][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.12948650121688843, 'train/grad_norm_pre_clip': 4.300751686096191, 'timing/data': 0.03130373731255531, 'timing/model': 1.659784340299666, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.68}) +[2026-06-15 18:11:44,656][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.1735568791627884, 'train/grad_norm_pre_clip': 6.758693695068359, 'timing/data': 0.02718743123114109, 'timing/model': 1.6004103356972337, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 0.68}) +[2026-06-15 18:11:50,252][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.10547560453414917, 'train/grad_norm_pre_clip': 4.599376201629639, 'timing/data': 0.03630916867405176, 'timing/model': 1.4214382003992796, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 0.68}) +[2026-06-15 18:11:56,054][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.21589988470077515, 'train/grad_norm_pre_clip': 6.413075923919678, 'timing/data': 0.029596791602671146, 'timing/model': 1.427562109194696, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 0.68}) +[2026-06-15 18:12:01,854][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.10121095925569534, 'train/grad_norm_pre_clip': 2.9893362522125244, 'timing/data': 0.03147939871996641, 'timing/model': 1.443685058504343, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 0.68}) +[2026-06-15 18:12:07,660][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.1027391329407692, 'train/grad_norm_pre_clip': 4.04590368270874, 'timing/data': 0.029414240270853043, 'timing/model': 1.3976924838498235, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 0.68}) +[2026-06-15 18:12:13,486][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.09523550420999527, 'train/grad_norm_pre_clip': 4.200183868408203, 'timing/data': 0.030222094617784023, 'timing/model': 1.6723559880629182, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 0.68}) +[2026-06-15 18:12:19,287][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.21593627333641052, 'train/grad_norm_pre_clip': 6.76296329498291, 'timing/data': 0.03020122367888689, 'timing/model': 1.6770860739052296, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 0.68}) +[2026-06-15 18:12:24,954][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.07435951381921768, 'train/grad_norm_pre_clip': 6.335852146148682, 'timing/data': 0.030435355380177498, 'timing/model': 1.4271295554935932, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 0.68}) +[2026-06-15 18:12:30,756][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.2056337594985962, 'train/grad_norm_pre_clip': 5.417370796203613, 'timing/data': 0.029995434917509556, 'timing/model': 1.4279753314331174, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 0.68}) +[2026-06-15 18:12:36,556][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.4774039387702942, 'train/grad_norm_pre_clip': 11.219429969787598, 'timing/data': 0.0310424892231822, 'timing/model': 1.4397610034793615, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.68}) +[2026-06-15 18:12:42,360][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.2369922399520874, 'train/grad_norm_pre_clip': 8.264510154724121, 'timing/data': 0.03066880814731121, 'timing/model': 1.3975064596161246, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 0.68}) +[2026-06-15 18:12:48,189][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.11284548044204712, 'train/grad_norm_pre_clip': 5.864160537719727, 'timing/data': 0.03066656831651926, 'timing/model': 1.670180881395936, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 0.68}) +[2026-06-15 18:12:53,848][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.2863902449607849, 'train/grad_norm_pre_clip': 5.94651460647583, 'timing/data': 0.03204603400081396, 'timing/model': 1.4250017004087567, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 0.68}) +[2026-06-15 18:12:59,652][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.10017921775579453, 'train/grad_norm_pre_clip': 3.4643304347991943, 'timing/data': 0.030993330292403698, 'timing/model': 1.4299063524231315, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 0.68}) +[2026-06-15 18:13:05,452][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.10063274204730988, 'train/grad_norm_pre_clip': 4.463005065917969, 'timing/data': 0.029697475023567677, 'timing/model': 1.430383125320077, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 0.68}) +[2026-06-15 18:13:11,153][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.14411160349845886, 'train/grad_norm_pre_clip': 3.623108386993408, 'timing/data': 0.030372947454452515, 'timing/model': 1.398911764845252, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 0.68}) +[2026-06-15 18:13:16,972][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.19248656928539276, 'train/grad_norm_pre_clip': 4.049055576324463, 'timing/data': 0.030480547808110714, 'timing/model': 1.472865467891097, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 0.68}) +[2026-06-15 18:13:22,762][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.13331931829452515, 'train/grad_norm_pre_clip': 4.260543346405029, 'timing/data': 0.03177103400230408, 'timing/model': 1.642042275518179, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.69}) +[2026-06-15 18:13:28,358][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.12400206178426743, 'train/grad_norm_pre_clip': 5.641916751861572, 'timing/data': 0.03098542056977749, 'timing/model': 1.4303063824772835, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 0.69}) +[2026-06-15 18:24:10,559][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.21774929761886597, 'train/grad_norm_pre_clip': 6.020931720733643, 'eval/loss': 0.11774779856204987, 'eval/flap_precision': 0.8837209302325582, 'eval/flap_recall': 0.41454545454545455, 'eval/flap_f1': 0.5643564356435643, 'eval/flap_rate_model': 0.035833333333333335, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.5643564356435643, 'timing/data': 0.029907746240496635, 'timing/model': 1.4291174486279488, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.69}) +[2026-06-15 18:24:10,560][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1500_state +[2026-06-15 18:24:10,566][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 18:24:19,240][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1500_state/model.safetensors +[2026-06-15 18:24:30,174][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1500_state/optimizer.bin +[2026-06-15 18:24:30,176][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1500_state/sampler.bin +[2026-06-15 18:24:30,176][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-15 18:24:30,178][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-15 18:24:30,350][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-15 18:24:30,351][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 18:24:30,367][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 18:24:36,124][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.27213847637176514, 'train/grad_norm_pre_clip': 6.020671367645264, 'timing/data': 0.033832790330052376, 'timing/model': 1.4322961224243045, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 0.69}) +[2026-06-15 18:24:41,825][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.2176971286535263, 'train/grad_norm_pre_clip': 7.709921836853027, 'timing/data': 0.030755206011235714, 'timing/model': 1.421904026530683, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 0.69}) +[2026-06-15 18:24:47,538][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.26719510555267334, 'train/grad_norm_pre_clip': 6.7924957275390625, 'timing/data': 0.03122508805245161, 'timing/model': 1.4412049409002066, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 0.69}) +[2026-06-15 18:24:53,129][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.16292135417461395, 'train/grad_norm_pre_clip': 4.787524223327637, 'timing/data': 0.03514592722058296, 'timing/model': 1.381380258128047, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 0.69}) +[2026-06-15 18:24:58,765][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.1522538810968399, 'train/grad_norm_pre_clip': 2.9825973510742188, 'timing/data': 0.030427495017647743, 'timing/model': 1.6578844580799341, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.69}) +[2026-06-15 18:25:04,234][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.16340303421020508, 'train/grad_norm_pre_clip': 4.49668025970459, 'timing/data': 0.03058395627886057, 'timing/model': 1.4420608840882778, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 0.69}) +[2026-06-15 18:25:09,939][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.1606801599264145, 'train/grad_norm_pre_clip': 5.243395805358887, 'timing/data': 0.03273726440966129, 'timing/model': 1.3997481865808368, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 0.69}) +[2026-06-15 18:25:15,639][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.13677772879600525, 'train/grad_norm_pre_clip': 4.418773174285889, 'timing/data': 0.032812017016112804, 'timing/model': 1.4127887999638915, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 0.69}) +[2026-06-15 18:25:21,373][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.1231628805398941, 'train/grad_norm_pre_clip': 4.032513618469238, 'timing/data': 0.030983037315309048, 'timing/model': 1.4146865317597985, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 0.69}) +[2026-06-15 18:25:27,059][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.17482133209705353, 'train/grad_norm_pre_clip': 4.985404014587402, 'timing/data': 0.031964512541890144, 'timing/model': 1.3870371636003256, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.69}) +[2026-06-15 18:25:32,779][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.2710261344909668, 'train/grad_norm_pre_clip': 4.650370121002197, 'timing/data': 0.030261955223977566, 'timing/model': 1.6765615064650774, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 0.69}) +[2026-06-15 18:25:38,345][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.28745877742767334, 'train/grad_norm_pre_clip': 5.774947643280029, 'timing/data': 0.03000860381871462, 'timing/model': 1.4304246054962277, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 0.69}) +[2026-06-15 18:25:44,161][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.17337942123413086, 'train/grad_norm_pre_clip': 13.033595085144043, 'timing/data': 0.031020217575132847, 'timing/model': 1.4357877941802144, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 0.69}) +[2026-06-15 18:25:49,949][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.3065396249294281, 'train/grad_norm_pre_clip': 7.806423187255859, 'timing/data': 0.030061714351177216, 'timing/model': 1.4397799549624324, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 0.69}) +[2026-06-15 18:25:55,747][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.10990587621927261, 'train/grad_norm_pre_clip': 2.6079440116882324, 'timing/data': 0.02970159240067005, 'timing/model': 1.4163187760859728, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 0.69}) +[2026-06-15 18:26:01,451][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.24739494919776917, 'train/grad_norm_pre_clip': 7.253818988800049, 'timing/data': 0.03093114774674177, 'timing/model': 1.4356406908482313, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 0.69}) +[2026-06-15 18:26:07,186][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.2241177260875702, 'train/grad_norm_pre_clip': 16.7030086517334, 'timing/data': 0.031078489497303963, 'timing/model': 1.6718161227181554, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 0.69}) +[2026-06-15 18:26:12,853][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.13451071083545685, 'train/grad_norm_pre_clip': 3.6583375930786133, 'timing/data': 0.031157069839537144, 'timing/model': 1.4387440010905266, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 0.69}) +[2026-06-15 18:26:18,857][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.1388547420501709, 'train/grad_norm_pre_clip': 4.399181842803955, 'timing/data': 0.029783262871205807, 'timing/model': 1.450803111307323, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 0.69}) +[2026-06-15 18:26:24,656][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.17239996790885925, 'train/grad_norm_pre_clip': 5.7961297035217285, 'timing/data': 0.03022353444248438, 'timing/model': 1.4406503047794104, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.7}) +[2026-06-15 18:26:30,451][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.13498017191886902, 'train/grad_norm_pre_clip': 3.9917101860046387, 'timing/data': 0.03515033796429634, 'timing/model': 1.442505156621337, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 0.7}) +[2026-06-15 18:26:36,269][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.22208930552005768, 'train/grad_norm_pre_clip': 4.771985054016113, 'timing/data': 0.028996839188039303, 'timing/model': 1.4660083521157503, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 0.7}) +[2026-06-15 18:26:42,102][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.20059674978256226, 'train/grad_norm_pre_clip': 5.816215991973877, 'timing/data': 0.03029731474816799, 'timing/model': 1.6729606855660677, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 0.7}) +[2026-06-15 18:26:47,656][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.242887482047081, 'train/grad_norm_pre_clip': 5.088500022888184, 'timing/data': 0.031891423277556896, 'timing/model': 1.4432084066793323, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 0.7}) +[2026-06-15 18:26:53,467][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.3139846622943878, 'train/grad_norm_pre_clip': 15.61109447479248, 'timing/data': 0.031648991629481316, 'timing/model': 1.4341121455654502, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 0.7}) +[2026-06-15 18:26:59,261][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.1673983931541443, 'train/grad_norm_pre_clip': 4.872877597808838, 'timing/data': 0.030347134917974472, 'timing/model': 1.4395218649879098, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 0.7}) +[2026-06-15 18:27:05,074][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.11694265902042389, 'train/grad_norm_pre_clip': 3.6439008712768555, 'timing/data': 0.03056082595139742, 'timing/model': 1.4016410550102592, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 0.7}) +[2026-06-15 18:27:10,757][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.4710146486759186, 'train/grad_norm_pre_clip': 14.812701225280762, 'timing/data': 0.030756297521293163, 'timing/model': 1.4100677380338311, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 0.7}) +[2026-06-15 18:27:16,599][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.415774941444397, 'train/grad_norm_pre_clip': 8.351330757141113, 'timing/data': 0.02960135228931904, 'timing/model': 1.6829192694276571, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 0.7}) +[2026-06-15 18:27:22,260][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.13780146837234497, 'train/grad_norm_pre_clip': 3.698577642440796, 'timing/data': 0.029569142498075962, 'timing/model': 1.4357421342283487, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.7}) +[2026-06-15 18:27:28,057][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.160930335521698, 'train/grad_norm_pre_clip': 5.19740104675293, 'timing/data': 0.030476336367428303, 'timing/model': 1.428392343223095, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 0.7}) +[2026-06-15 18:27:33,871][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.34537190198898315, 'train/grad_norm_pre_clip': 7.592776298522949, 'timing/data': 0.029509122483432293, 'timing/model': 1.4437086591497064, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 0.7}) +[2026-06-15 18:27:39,564][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.23722568154335022, 'train/grad_norm_pre_clip': 4.821822643280029, 'timing/data': 0.030392246320843697, 'timing/model': 1.441140958108008, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 0.7}) +[2026-06-15 18:27:45,361][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.28024399280548096, 'train/grad_norm_pre_clip': 5.1534104347229, 'timing/data': 0.030197705142199993, 'timing/model': 1.4277436565607786, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 0.7}) +[2026-06-15 18:27:51,204][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.23639029264450073, 'train/grad_norm_pre_clip': 4.3749589920043945, 'timing/data': 0.030626997351646423, 'timing/model': 1.6781321931630373, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 0.7}) +[2026-06-15 18:27:56,898][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.18529948592185974, 'train/grad_norm_pre_clip': 3.303069591522217, 'timing/data': 0.030485616996884346, 'timing/model': 1.4666858669370413, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 0.7}) +[2026-06-15 18:28:02,760][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.22856631875038147, 'train/grad_norm_pre_clip': 6.423829078674316, 'timing/data': 0.0315684424713254, 'timing/model': 1.4329414470121264, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 0.7}) +[2026-06-15 18:28:08,579][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.08918267488479614, 'train/grad_norm_pre_clip': 2.3075625896453857, 'timing/data': 0.040071362629532814, 'timing/model': 1.4383220253512263, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 0.7}) +[2026-06-15 18:28:14,898][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.16348224878311157, 'train/grad_norm_pre_clip': 3.1335349082946777, 'timing/data': 0.03889785706996918, 'timing/model': 1.4515989115461707, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 0.7}) +[2026-06-15 18:28:21,175][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.08741618692874908, 'train/grad_norm_pre_clip': 3.8436975479125977, 'timing/data': 0.04055146500468254, 'timing/model': 1.4626292074099183, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.7}) +[2026-06-15 18:28:27,427][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.24903732538223267, 'train/grad_norm_pre_clip': 5.187403678894043, 'timing/data': 0.042724115774035454, 'timing/model': 2.075123851187527, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 0.7}) +[2026-06-15 18:28:33,119][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.07295767217874527, 'train/grad_norm_pre_clip': 3.2292122840881348, 'timing/data': 0.04232585430145264, 'timing/model': 1.4554656967520714, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 0.71}) +[2026-06-15 18:28:39,279][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.4437330663204193, 'train/grad_norm_pre_clip': 7.519120693206787, 'timing/data': 0.04190327227115631, 'timing/model': 1.427214122377336, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 0.71}) +[2026-06-15 18:28:45,189][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.1785077154636383, 'train/grad_norm_pre_clip': 8.983650207519531, 'timing/data': 0.040892377495765686, 'timing/model': 1.4292693939059973, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 0.71}) +[2026-06-15 18:28:51,483][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.13050246238708496, 'train/grad_norm_pre_clip': 5.7425031661987305, 'timing/data': 0.042153104208409786, 'timing/model': 1.4338142881169915, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 0.71}) +[2026-06-15 18:28:57,779][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.10247594863176346, 'train/grad_norm_pre_clip': 8.677947044372559, 'timing/data': 0.043171548284590244, 'timing/model': 1.469659173861146, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 0.71}) +[2026-06-15 18:29:04,039][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.19166511297225952, 'train/grad_norm_pre_clip': 10.905158996582031, 'timing/data': 0.04215026367455721, 'timing/model': 1.9859033087268472, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.71}) +[2026-06-15 18:29:09,681][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.034423768520355225, 'train/grad_norm_pre_clip': 2.1533238887786865, 'timing/data': 0.04138053022325039, 'timing/model': 1.4303303277119994, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 0.71}) +[2026-06-15 18:29:15,877][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.30692195892333984, 'train/grad_norm_pre_clip': 5.950427532196045, 'timing/data': 0.042241474613547325, 'timing/model': 1.4351334422826767, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 0.71}) +[2026-06-15 18:29:22,082][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.4624701738357544, 'train/grad_norm_pre_clip': 12.052577018737793, 'timing/data': 0.04085181653499603, 'timing/model': 1.4377499166876078, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.71}) +[2026-06-15 18:29:28,377][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.13581602275371552, 'train/grad_norm_pre_clip': 7.513046741485596, 'timing/data': 0.04278109595179558, 'timing/model': 1.4318720614537597, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 0.71}) +[2026-06-15 18:29:34,676][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.10560815036296844, 'train/grad_norm_pre_clip': 6.384854793548584, 'timing/data': 0.041433810256421566, 'timing/model': 1.4679073868319392, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 0.71}) +[2026-06-15 18:29:40,947][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.504214346408844, 'train/grad_norm_pre_clip': 11.752679824829102, 'timing/data': 0.04150792025029659, 'timing/model': 2.0011113984510303, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 0.71}) +[2026-06-15 18:29:47,099][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.06082061678171158, 'train/grad_norm_pre_clip': 3.080573081970215, 'timing/data': 0.04185041319578886, 'timing/model': 1.9528034264221787, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 0.71}) +[2026-06-15 18:29:52,795][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.2127261906862259, 'train/grad_norm_pre_clip': 5.925876617431641, 'timing/data': 0.04132547974586487, 'timing/model': 1.4528832994401455, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 0.71}) +[2026-06-15 18:29:58,982][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.09294570982456207, 'train/grad_norm_pre_clip': 5.8709001541137695, 'timing/data': 0.04305129870772362, 'timing/model': 1.4424518402665854, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 0.71}) +[2026-06-15 18:30:05,188][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.19740712642669678, 'train/grad_norm_pre_clip': 5.968371868133545, 'timing/data': 0.03952519129961729, 'timing/model': 1.4416391784325242, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 0.71}) +[2026-06-15 18:30:11,374][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.09014564752578735, 'train/grad_norm_pre_clip': 4.8797502517700195, 'timing/data': 0.04141576122492552, 'timing/model': 1.4713780246675014, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 0.71}) +[2026-06-15 18:30:17,413][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.1868310570716858, 'train/grad_norm_pre_clip': 7.169486045837402, 'timing/data': 0.0410175696015358, 'timing/model': 1.8762271171435714, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 0.71}) +[2026-06-15 18:30:23,359][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.2150961309671402, 'train/grad_norm_pre_clip': 8.915238380432129, 'timing/data': 0.04175330139696598, 'timing/model': 1.7141526639461517, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.71}) +[2026-06-15 18:30:28,982][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.16258050501346588, 'train/grad_norm_pre_clip': 6.812931060791016, 'timing/data': 0.04194915294647217, 'timing/model': 1.4392120437696576, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 0.71}) +[2026-06-15 18:30:35,078][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.26331663131713867, 'train/grad_norm_pre_clip': 7.1623101234436035, 'timing/data': 0.041011249646544456, 'timing/model': 1.4354693470522761, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 0.71}) +[2026-06-15 18:30:41,173][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.21026352047920227, 'train/grad_norm_pre_clip': 7.628859996795654, 'timing/data': 0.041312750428915024, 'timing/model': 1.4322187937796116, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 0.71}) +[2026-06-15 18:30:47,294][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.23594500124454498, 'train/grad_norm_pre_clip': 5.268319129943848, 'timing/data': 0.041551271453499794, 'timing/model': 1.4903871156275272, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 0.72}) +[2026-06-15 18:30:53,969][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.18980766832828522, 'train/grad_norm_pre_clip': 7.645776748657227, 'timing/data': 0.04161567147821188, 'timing/model': 1.9129004580900073, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 0.72}) +[2026-06-15 18:30:59,672][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.238637313246727, 'train/grad_norm_pre_clip': 6.439960956573486, 'timing/data': 0.04017592500895262, 'timing/model': 1.4297474650666118, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 0.72}) +[2026-06-15 18:31:05,664][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.2260483056306839, 'train/grad_norm_pre_clip': 6.235189914703369, 'timing/data': 0.03809310495853424, 'timing/model': 1.4429928110912442, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 0.72}) +[2026-06-15 18:31:11,558][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.10092335194349289, 'train/grad_norm_pre_clip': 2.885819435119629, 'timing/data': 0.037110960111021996, 'timing/model': 1.436228420585394, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 0.72}) +[2026-06-15 18:31:17,456][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.2673948109149933, 'train/grad_norm_pre_clip': 9.09329891204834, 'timing/data': 0.037449891678988934, 'timing/model': 1.4568795217201114, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 0.72}) +[2026-06-15 18:31:23,160][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.2044953852891922, 'train/grad_norm_pre_clip': 6.1381988525390625, 'timing/data': 0.03119687084108591, 'timing/model': 1.3986467309296131, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.72}) +[2026-06-15 18:31:28,972][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.12295062839984894, 'train/grad_norm_pre_clip': 2.7972075939178467, 'timing/data': 0.03125159256160259, 'timing/model': 1.64954540040344, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 0.72}) +[2026-06-15 18:31:34,552][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.15732061862945557, 'train/grad_norm_pre_clip': 5.221889019012451, 'timing/data': 0.034650747664272785, 'timing/model': 1.4254274033010006, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 0.72}) +[2026-06-15 18:31:40,348][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.16004253923892975, 'train/grad_norm_pre_clip': 5.749942302703857, 'timing/data': 0.032528127543628216, 'timing/model': 1.4523108955472708, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 0.72}) +[2026-06-15 18:31:46,153][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.24964332580566406, 'train/grad_norm_pre_clip': 8.908357620239258, 'timing/data': 0.030387748032808304, 'timing/model': 1.4396957457065582, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 0.72}) +[2026-06-15 18:31:51,950][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.12129752337932587, 'train/grad_norm_pre_clip': 2.3436853885650635, 'timing/data': 0.031500622630119324, 'timing/model': 1.469830373302102, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.72}) +[2026-06-15 18:31:57,853][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.1705796867609024, 'train/grad_norm_pre_clip': 6.065012454986572, 'timing/data': 0.03118693083524704, 'timing/model': 1.4554123049601912, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 0.72}) +[2026-06-15 18:32:03,578][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.07136061042547226, 'train/grad_norm_pre_clip': 3.78100323677063, 'timing/data': 0.03205881640315056, 'timing/model': 1.6659699883311987, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 0.72}) +[2026-06-15 18:32:09,155][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.291156530380249, 'train/grad_norm_pre_clip': 6.3026933670043945, 'timing/data': 0.032390945591032505, 'timing/model': 1.4404526017606258, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 0.72}) +[2026-06-15 18:32:14,950][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.13721133768558502, 'train/grad_norm_pre_clip': 5.19160795211792, 'timing/data': 0.03165274392813444, 'timing/model': 1.4215362863615155, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 0.72}) +[2026-06-15 18:32:20,756][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.2723237872123718, 'train/grad_norm_pre_clip': 20.264081954956055, 'timing/data': 0.030498337000608444, 'timing/model': 1.4325119946151972, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.72}) +[2026-06-15 18:32:26,653][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.4607659578323364, 'train/grad_norm_pre_clip': 17.355239868164062, 'timing/data': 0.03103079181164503, 'timing/model': 1.4629858052358031, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 0.72}) +[2026-06-15 18:32:32,512][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.08866560459136963, 'train/grad_norm_pre_clip': 5.2405924797058105, 'timing/data': 0.03097949828952551, 'timing/model': 1.6837924495339394, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 0.72}) +[2026-06-15 18:32:38,050][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.10669790953397751, 'train/grad_norm_pre_clip': 9.365777969360352, 'timing/data': 0.035781385377049446, 'timing/model': 1.4300794787704945, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 0.72}) +[2026-06-15 18:32:43,851][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.2538969814777374, 'train/grad_norm_pre_clip': 6.524034023284912, 'timing/data': 0.03149690944701433, 'timing/model': 1.4271298432722688, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 0.72}) +[2026-06-15 18:32:49,556][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.3436438739299774, 'train/grad_norm_pre_clip': 12.638506889343262, 'timing/data': 0.03428322263062, 'timing/model': 1.4259726852178574, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 0.72}) +[2026-06-15 18:32:55,367][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.171972393989563, 'train/grad_norm_pre_clip': 4.569559574127197, 'timing/data': 0.030659034848213196, 'timing/model': 1.4416846940293908, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 0.73}) +[2026-06-15 18:33:01,157][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.3310559391975403, 'train/grad_norm_pre_clip': 6.55597448348999, 'timing/data': 0.031608604826033115, 'timing/model': 1.4010528055951, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 0.73}) +[2026-06-15 18:33:07,021][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.03796129673719406, 'train/grad_norm_pre_clip': 2.8175151348114014, 'timing/data': 0.0316545432433486, 'timing/model': 1.6984247835353017, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 0.73}) +[2026-06-15 18:33:12,553][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.1775103658437729, 'train/grad_norm_pre_clip': 8.257911682128906, 'timing/data': 0.030725869350135326, 'timing/model': 1.4217334995046258, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 0.73}) +[2026-06-15 18:33:18,358][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.16163428127765656, 'train/grad_norm_pre_clip': 13.360599517822266, 'timing/data': 0.031243523582816124, 'timing/model': 1.4276940999552608, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.73}) +[2026-06-15 18:33:24,270][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.18347212672233582, 'train/grad_norm_pre_clip': 3.2976067066192627, 'timing/data': 0.0325879268348217, 'timing/model': 1.4570877803489566, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 0.73}) +[2026-06-15 18:33:30,158][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.43731510639190674, 'train/grad_norm_pre_clip': 13.847816467285156, 'timing/data': 0.03636908996850252, 'timing/model': 1.4283578395843506, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 0.73}) +[2026-06-15 18:33:35,981][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.140346497297287, 'train/grad_norm_pre_clip': 7.843997955322266, 'timing/data': 0.03399351891130209, 'timing/model': 1.4136643605306745, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 0.73}) +[2026-06-15 18:33:41,841][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.203212171792984, 'train/grad_norm_pre_clip': 7.6624956130981445, 'timing/data': 0.038481298834085464, 'timing/model': 1.6953220460563898, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 0.73}) +[2026-06-15 18:33:47,466][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.18251657485961914, 'train/grad_norm_pre_clip': 4.5004096031188965, 'timing/data': 0.031785110011696815, 'timing/model': 1.4447090132161975, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 0.73}) +[2026-06-15 18:33:53,362][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.1616237610578537, 'train/grad_norm_pre_clip': 8.106167793273926, 'timing/data': 0.03204924054443836, 'timing/model': 1.4403695575892925, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.73}) +[2026-06-15 18:33:59,171][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.2334745228290558, 'train/grad_norm_pre_clip': 7.172647953033447, 'timing/data': 0.035705190151929855, 'timing/model': 1.443288840353489, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 0.73}) +[2026-06-15 18:34:05,057][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.4208087921142578, 'train/grad_norm_pre_clip': 10.567660331726074, 'timing/data': 0.03268588334321976, 'timing/model': 1.4724570894613862, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 0.73}) +[2026-06-15 18:34:10,875][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.13888177275657654, 'train/grad_norm_pre_clip': 6.02117919921875, 'timing/data': 0.031945512630045414, 'timing/model': 1.407821741886437, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 0.73}) +[2026-06-15 18:44:31,876][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.04455789178609848, 'train/grad_norm_pre_clip': 2.292387008666992, 'eval/loss': 0.12779803574085236, 'eval/flap_precision': 0.9444444444444444, 'eval/flap_recall': 0.3709090909090909, 'eval/flap_f1': 0.5326370757180157, 'eval/flap_rate_model': 0.03, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.5326370757180157, 'timing/data': 0.031018909998238087, 'timing/model': 1.7061335435137153, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.73}) +[2026-06-15 18:44:31,877][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1600_state +[2026-06-15 18:44:31,884][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 18:44:40,531][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1600_state/model.safetensors +[2026-06-15 18:44:51,937][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1600_state/optimizer.bin +[2026-06-15 18:44:51,938][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1600_state/sampler.bin +[2026-06-15 18:44:51,938][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1600_state/sampler_1.bin +[2026-06-15 18:44:51,940][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1600_state/random_states_0.pkl +[2026-06-15 18:44:52,113][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1600 +[2026-06-15 18:44:52,114][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 18:44:52,129][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 18:44:57,933][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.36111798882484436, 'train/grad_norm_pre_clip': 8.280001640319824, 'timing/data': 0.047687471844255924, 'timing/model': 1.7295160442590714, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 0.73}) +[2026-06-15 18:45:03,682][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.09883544594049454, 'train/grad_norm_pre_clip': 4.440479755401611, 'timing/data': 0.036358438432216644, 'timing/model': 1.6755584860220551, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 0.73}) +[2026-06-15 18:45:09,125][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.06240520626306534, 'train/grad_norm_pre_clip': 1.6351923942565918, 'timing/data': 0.03324668388813734, 'timing/model': 1.4305019369348884, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 0.73}) +[2026-06-15 18:45:14,825][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.12305580079555511, 'train/grad_norm_pre_clip': 2.1908652782440186, 'timing/data': 0.033877138048410416, 'timing/model': 1.427033657208085, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 0.73}) +[2026-06-15 18:45:20,530][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.12376736849546432, 'train/grad_norm_pre_clip': 3.9566502571105957, 'timing/data': 0.044118442572653294, 'timing/model': 1.3819170920178294, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 0.73}) +[2026-06-15 18:45:26,130][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.10828353464603424, 'train/grad_norm_pre_clip': 3.5627529621124268, 'timing/data': 0.03303071018308401, 'timing/model': 1.3920338340103626, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 0.73}) +[2026-06-15 18:45:31,732][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.2039566934108734, 'train/grad_norm_pre_clip': 8.049500465393066, 'timing/data': 0.03186975233256817, 'timing/model': 1.3834408791735768, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 0.73}) +[2026-06-15 18:45:37,384][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.2857907712459564, 'train/grad_norm_pre_clip': 9.45375919342041, 'timing/data': 0.03301274962723255, 'timing/model': 1.6172744669020176, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 0.74}) +[2026-06-15 18:45:42,834][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.2016313374042511, 'train/grad_norm_pre_clip': 10.593940734863281, 'timing/data': 0.032377954572439194, 'timing/model': 1.4125464456155896, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 0.74}) +[2026-06-15 18:45:48,548][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.2097654640674591, 'train/grad_norm_pre_clip': 6.780594825744629, 'timing/data': 0.03335314244031906, 'timing/model': 1.4296200964599848, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.74}) +[2026-06-15 18:45:54,131][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.28108707070350647, 'train/grad_norm_pre_clip': 11.591072082519531, 'timing/data': 0.03159107826650143, 'timing/model': 1.4257650496438146, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 0.74}) +[2026-06-15 18:45:59,837][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.15432657301425934, 'train/grad_norm_pre_clip': 5.28419828414917, 'timing/data': 0.031045494601130486, 'timing/model': 1.4423558358103037, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 0.74}) +[2026-06-15 18:46:05,542][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.19042733311653137, 'train/grad_norm_pre_clip': 4.78968620300293, 'timing/data': 0.03295245673507452, 'timing/model': 1.3869833648204803, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 0.74}) +[2026-06-15 18:46:11,336][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.044720955193042755, 'train/grad_norm_pre_clip': 3.3397269248962402, 'timing/data': 0.03317605797201395, 'timing/model': 1.7285223854705691, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 0.74}) +[2026-06-15 18:46:16,959][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.11302337050437927, 'train/grad_norm_pre_clip': 6.394670009613037, 'timing/data': 0.048743619583547115, 'timing/model': 1.4252654481679201, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 0.74}) +[2026-06-15 18:46:22,760][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.24596281349658966, 'train/grad_norm_pre_clip': 11.61962604522705, 'timing/data': 0.030268175527453423, 'timing/model': 1.382294598966837, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 0.74}) +[2026-06-15 18:46:28,542][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.7511443495750427, 'train/grad_norm_pre_clip': 14.14729118347168, 'timing/data': 0.03188974782824516, 'timing/model': 1.43975461833179, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 0.74}) +[2026-06-15 18:46:34,247][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.38385865092277527, 'train/grad_norm_pre_clip': 19.451744079589844, 'timing/data': 0.03169872518628836, 'timing/model': 1.4359438205137849, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 0.74}) +[2026-06-15 18:46:40,151][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.42242199182510376, 'train/grad_norm_pre_clip': 11.461183547973633, 'timing/data': 0.030121713876724243, 'timing/model': 1.3870582627132535, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 0.74}) +[2026-06-15 18:46:45,912][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.04408687725663185, 'train/grad_norm_pre_clip': 2.1091716289520264, 'timing/data': 0.02822877001017332, 'timing/model': 1.7172753438353539, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.74}) +[2026-06-15 18:46:51,627][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.2200716733932495, 'train/grad_norm_pre_clip': 9.105978965759277, 'timing/data': 0.031123431399464607, 'timing/model': 1.662290070205927, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 0.74}) +[2026-06-15 18:46:57,291][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.09717161953449249, 'train/grad_norm_pre_clip': 3.9587278366088867, 'timing/data': 0.036273836158216, 'timing/model': 1.4087570561096072, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 0.74}) +[2026-06-15 18:47:03,097][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.14245907962322235, 'train/grad_norm_pre_clip': 3.185516834259033, 'timing/data': 0.03070230595767498, 'timing/model': 1.453648803755641, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 0.74}) +[2026-06-15 18:47:08,850][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.24001699686050415, 'train/grad_norm_pre_clip': 9.577759742736816, 'timing/data': 0.03106156922876835, 'timing/model': 1.44789635296911, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 0.74}) +[2026-06-15 18:47:14,518][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.18328315019607544, 'train/grad_norm_pre_clip': 6.085926055908203, 'timing/data': 0.029063173569738865, 'timing/model': 1.3918452197685838, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 0.74}) +[2026-06-15 18:47:20,232][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.21151748299598694, 'train/grad_norm_pre_clip': 9.228555679321289, 'timing/data': 0.028441389091312885, 'timing/model': 1.658633241429925, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 0.74}) +[2026-06-15 18:47:25,770][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.22772404551506042, 'train/grad_norm_pre_clip': 8.706024169921875, 'timing/data': 0.032326686196029186, 'timing/model': 1.4021442169323564, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 0.74}) +[2026-06-15 18:47:31,545][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.2531496286392212, 'train/grad_norm_pre_clip': 7.286883354187012, 'timing/data': 0.034143428318202496, 'timing/model': 1.4188432851806283, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 0.74}) +[2026-06-15 18:47:37,254][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.2344723641872406, 'train/grad_norm_pre_clip': 4.643343925476074, 'timing/data': 0.03062508348375559, 'timing/model': 1.4449472585693002, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 0.74}) +[2026-06-15 18:47:43,046][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.11100763082504272, 'train/grad_norm_pre_clip': 3.4698030948638916, 'timing/data': 0.0355987586081028, 'timing/model': 1.3992547513917089, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.75}) +[2026-06-15 18:47:48,846][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.3182348608970642, 'train/grad_norm_pre_clip': 6.51636266708374, 'timing/data': 0.03668735548853874, 'timing/model': 1.448953839018941, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 0.75}) +[2026-06-15 18:47:54,538][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.26720574498176575, 'train/grad_norm_pre_clip': 5.370700836181641, 'timing/data': 0.030827303417026997, 'timing/model': 1.6308249216526747, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 0.75}) +[2026-06-15 18:48:00,146][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.1086207926273346, 'train/grad_norm_pre_clip': 2.5087904930114746, 'timing/data': 0.03417349699884653, 'timing/model': 1.4357581622898579, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 0.75}) +[2026-06-15 18:48:05,945][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.18961717188358307, 'train/grad_norm_pre_clip': 4.817078590393066, 'timing/data': 0.031456027179956436, 'timing/model': 1.4386900747194886, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 0.75}) +[2026-06-15 18:48:11,650][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.1242000013589859, 'train/grad_norm_pre_clip': 4.285958290100098, 'timing/data': 0.03160998784005642, 'timing/model': 1.442548624239862, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 0.75}) +[2026-06-15 18:48:17,343][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.3162470757961273, 'train/grad_norm_pre_clip': 11.473052978515625, 'timing/data': 0.03039424866437912, 'timing/model': 1.403869030997157, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 0.75}) +[2026-06-15 18:48:23,041][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.2015763819217682, 'train/grad_norm_pre_clip': 5.814064025878906, 'timing/data': 0.030938093550503254, 'timing/model': 1.393754644319415, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 0.75}) +[2026-06-15 18:48:28,752][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.20334899425506592, 'train/grad_norm_pre_clip': 5.16969633102417, 'timing/data': 0.0305364103987813, 'timing/model': 1.6425157943740487, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 0.75}) +[2026-06-15 18:48:34,249][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.23702271282672882, 'train/grad_norm_pre_clip': 6.196165084838867, 'timing/data': 0.0307902405038476, 'timing/model': 1.445543128065765, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 0.75}) +[2026-06-15 18:48:39,958][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.20414988696575165, 'train/grad_norm_pre_clip': 6.5429277420043945, 'timing/data': 0.03072694130241871, 'timing/model': 1.396349884569645, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.75}) +[2026-06-15 18:48:45,746][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.12934404611587524, 'train/grad_norm_pre_clip': 4.994886875152588, 'timing/data': 0.03033134713768959, 'timing/model': 1.443160679191351, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 0.75}) +[2026-06-15 18:48:51,477][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.15217527747154236, 'train/grad_norm_pre_clip': 6.019834995269775, 'timing/data': 0.03450903482735157, 'timing/model': 1.4011600371450186, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 0.75}) +[2026-06-15 18:48:57,299][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.2690192461013794, 'train/grad_norm_pre_clip': 7.818349361419678, 'timing/data': 0.03546293172985315, 'timing/model': 1.6750946249812841, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 0.75}) +[2026-06-15 18:49:02,856][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.15610072016716003, 'train/grad_norm_pre_clip': 5.244808197021484, 'timing/data': 0.03046333696693182, 'timing/model': 1.4375988934189081, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 0.75}) +[2026-06-15 18:49:08,660][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.28249067068099976, 'train/grad_norm_pre_clip': 6.610733985900879, 'timing/data': 0.033103243447840214, 'timing/model': 1.4188920203596354, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.75}) +[2026-06-15 18:49:14,348][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.147369384765625, 'train/grad_norm_pre_clip': 3.438091993331909, 'timing/data': 0.030464395880699158, 'timing/model': 1.4453825848177075, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 0.75}) +[2026-06-15 18:49:20,049][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.2841881513595581, 'train/grad_norm_pre_clip': 8.167946815490723, 'timing/data': 0.031129140406847, 'timing/model': 1.4430542355403304, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 0.75}) +[2026-06-15 18:49:25,750][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.18785783648490906, 'train/grad_norm_pre_clip': 5.755178928375244, 'timing/data': 0.031028510071337223, 'timing/model': 1.4379933150485158, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 0.75}) +[2026-06-15 18:49:31,589][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.22276298701763153, 'train/grad_norm_pre_clip': 8.671364784240723, 'timing/data': 0.030780628323554993, 'timing/model': 1.668619285337627, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 0.75}) +[2026-06-15 18:49:37,253][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.1696873903274536, 'train/grad_norm_pre_clip': 5.658261299133301, 'timing/data': 0.03124945145100355, 'timing/model': 1.4386876719072461, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.75}) +[2026-06-15 18:49:42,953][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.09860793501138687, 'train/grad_norm_pre_clip': 3.974238157272339, 'timing/data': 0.029299926944077015, 'timing/model': 1.3917922005057335, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 0.76}) +[2026-06-15 18:49:48,753][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.0461510494351387, 'train/grad_norm_pre_clip': 2.2222037315368652, 'timing/data': 0.03600113093852997, 'timing/model': 1.4369717510417104, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 0.76}) +[2026-06-15 18:49:54,452][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.01737809181213379, 'train/grad_norm_pre_clip': 1.1803839206695557, 'timing/data': 0.029610089026391506, 'timing/model': 1.4462594175711274, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 0.76}) +[2026-06-15 18:50:00,243][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.2016158252954483, 'train/grad_norm_pre_clip': 10.89709758758545, 'timing/data': 0.03081641625612974, 'timing/model': 1.4202654398977757, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 0.76}) +[2026-06-15 18:50:06,078][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.08818124234676361, 'train/grad_norm_pre_clip': 6.418029308319092, 'timing/data': 0.029200085438787937, 'timing/model': 1.665545484982431, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 0.76}) +[2026-06-15 18:50:11,649][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.25293123722076416, 'train/grad_norm_pre_clip': 8.633515357971191, 'timing/data': 0.0309024965390563, 'timing/model': 1.4325132742524147, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 0.76}) +[2026-06-15 18:50:17,343][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.11346001923084259, 'train/grad_norm_pre_clip': 7.110208034515381, 'timing/data': 0.030977786518633366, 'timing/model': 1.3951722029596567, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 0.76}) +[2026-06-15 18:50:23,241][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.2250777930021286, 'train/grad_norm_pre_clip': 14.99576187133789, 'timing/data': 0.030808755196630955, 'timing/model': 1.4493779707700014, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 0.76}) +[2026-06-15 18:50:28,948][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.19642458856105804, 'train/grad_norm_pre_clip': 5.408950328826904, 'timing/data': 0.03041747212409973, 'timing/model': 1.4097920637577772, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 0.76}) +[2026-06-15 18:50:34,651][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.09793195873498917, 'train/grad_norm_pre_clip': 5.009922504425049, 'timing/data': 0.0333427619189024, 'timing/model': 1.4366187397390604, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.76}) +[2026-06-15 18:50:40,378][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.21324880421161652, 'train/grad_norm_pre_clip': 8.170666694641113, 'timing/data': 0.030906114727258682, 'timing/model': 1.6752735180780292, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 0.76}) +[2026-06-15 18:50:45,947][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.2803676724433899, 'train/grad_norm_pre_clip': 17.18262481689453, 'timing/data': 0.03126573748886585, 'timing/model': 1.445328751578927, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 0.76}) +[2026-06-15 18:50:51,743][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.14569874107837677, 'train/grad_norm_pre_clip': 7.435959815979004, 'timing/data': 0.03125546686351299, 'timing/model': 1.4401506641879678, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 0.76}) +[2026-06-15 18:50:57,447][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.1877472847700119, 'train/grad_norm_pre_clip': 8.605035781860352, 'timing/data': 0.032201663590967655, 'timing/model': 1.4365913784131408, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 0.76}) +[2026-06-15 18:51:03,146][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.2302023023366928, 'train/grad_norm_pre_clip': 11.845654487609863, 'timing/data': 0.030748853459954262, 'timing/model': 1.4184211362153292, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 0.76}) +[2026-06-15 18:51:08,844][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.0988013744354248, 'train/grad_norm_pre_clip': 4.753560543060303, 'timing/data': 0.031222674995660782, 'timing/model': 1.3909312039613724, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 0.76}) +[2026-06-15 18:51:14,591][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.04794725403189659, 'train/grad_norm_pre_clip': 4.025716781616211, 'timing/data': 0.031266345642507076, 'timing/model': 1.6797466995194554, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 0.76}) +[2026-06-15 18:51:20,151][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.29925745725631714, 'train/grad_norm_pre_clip': 9.372773170471191, 'timing/data': 0.029516234062612057, 'timing/model': 1.4387308936566114, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 0.76}) +[2026-06-15 18:51:25,951][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.31657734513282776, 'train/grad_norm_pre_clip': 7.940555095672607, 'timing/data': 0.03315220680087805, 'timing/model': 1.416328995488584, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 0.76}) +[2026-06-15 18:51:31,770][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.12584719061851501, 'train/grad_norm_pre_clip': 5.87027645111084, 'timing/data': 0.0300647160038352, 'timing/model': 1.4603665033355355, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.76}) +[2026-06-15 18:51:37,564][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.1598503738641739, 'train/grad_norm_pre_clip': 7.58976411819458, 'timing/data': 0.031237123534083366, 'timing/model': 1.452705550007522, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 0.76}) +[2026-06-15 18:51:43,352][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.33607053756713867, 'train/grad_norm_pre_clip': 15.127476692199707, 'timing/data': 0.031128563918173313, 'timing/model': 1.409430354833603, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 0.76}) +[2026-06-15 18:51:49,181][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.1383010298013687, 'train/grad_norm_pre_clip': 11.470745086669922, 'timing/data': 0.030135806649923325, 'timing/model': 1.6590647231787443, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 0.77}) +[2026-06-15 18:51:54,755][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.14759378135204315, 'train/grad_norm_pre_clip': 4.836347579956055, 'timing/data': 0.03067931067198515, 'timing/model': 1.4149447493255138, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 0.77}) +[2026-06-15 18:52:00,546][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.1537761390209198, 'train/grad_norm_pre_clip': 4.6082377433776855, 'timing/data': 0.029303851537406445, 'timing/model': 1.4244616897776723, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 0.77}) +[2026-06-15 18:52:06,348][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.08818335831165314, 'train/grad_norm_pre_clip': 3.315202474594116, 'timing/data': 0.029867073521018028, 'timing/model': 1.43530547991395, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 0.77}) +[2026-06-15 18:52:12,047][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.18671099841594696, 'train/grad_norm_pre_clip': 6.3629069328308105, 'timing/data': 0.029866503551602364, 'timing/model': 1.447046086192131, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 0.77}) +[2026-06-15 18:52:17,862][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.14582036435604095, 'train/grad_norm_pre_clip': 5.526938438415527, 'timing/data': 0.03052393812686205, 'timing/model': 1.4067500038072467, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 0.77}) +[2026-06-15 18:52:23,725][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.2198353260755539, 'train/grad_norm_pre_clip': 6.801100254058838, 'timing/data': 0.03446599282324314, 'timing/model': 1.7133035687729716, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 0.77}) +[2026-06-15 18:52:29,356][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.13107384741306305, 'train/grad_norm_pre_clip': 5.572928428649902, 'timing/data': 0.0332635547965765, 'timing/model': 1.444255905225873, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.77}) +[2026-06-15 18:52:35,145][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.31474000215530396, 'train/grad_norm_pre_clip': 10.966666221618652, 'timing/data': 0.03013458475470543, 'timing/model': 1.427357324399054, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 0.77}) +[2026-06-15 18:52:40,848][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.08222837746143341, 'train/grad_norm_pre_clip': 5.4662957191467285, 'timing/data': 0.035680401138961315, 'timing/model': 1.432802900671959, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 0.77}) +[2026-06-15 18:52:46,544][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.2132074385881424, 'train/grad_norm_pre_clip': 5.947506904602051, 'timing/data': 0.02905078697949648, 'timing/model': 1.4432946182787418, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 0.77}) +[2026-06-15 18:52:52,252][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.16844095289707184, 'train/grad_norm_pre_clip': 5.748907089233398, 'timing/data': 0.030377894639968872, 'timing/model': 1.4022988257929683, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 0.77}) +[2026-06-15 18:52:57,986][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.11873314529657364, 'train/grad_norm_pre_clip': 5.915461540222168, 'timing/data': 0.029643110930919647, 'timing/model': 1.677781437523663, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 0.77}) +[2026-06-15 18:53:03,728][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.12137255817651749, 'train/grad_norm_pre_clip': 5.435809135437012, 'timing/data': 0.02902318723499775, 'timing/model': 1.6099491193890572, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 0.77}) +[2026-06-15 18:53:09,248][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.22959904372692108, 'train/grad_norm_pre_clip': 10.50644588470459, 'timing/data': 0.030025582760572433, 'timing/model': 1.4273312967270613, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 0.77}) +[2026-06-15 18:53:15,042][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.3087945580482483, 'train/grad_norm_pre_clip': 7.524599552154541, 'timing/data': 0.028858144767582417, 'timing/model': 1.4252335140481591, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 0.77}) +[2026-06-15 18:53:20,763][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.39295828342437744, 'train/grad_norm_pre_clip': 6.953488826751709, 'timing/data': 0.030256683938205242, 'timing/model': 1.440599192865193, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 0.77}) +[2026-06-15 18:53:26,573][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.3796378970146179, 'train/grad_norm_pre_clip': 15.627510070800781, 'timing/data': 0.030855038203299046, 'timing/model': 1.4200008008629084, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.77}) +[2026-06-15 18:53:32,392][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.263540655374527, 'train/grad_norm_pre_clip': 6.351657390594482, 'timing/data': 0.02912058588117361, 'timing/model': 1.6674226550385356, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 0.77}) +[2026-06-15 18:53:38,193][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.14151255786418915, 'train/grad_norm_pre_clip': 4.717101573944092, 'timing/data': 0.03055754490196705, 'timing/model': 1.6722767362371087, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 0.77}) +[2026-06-15 18:53:43,751][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.135269895195961, 'train/grad_norm_pre_clip': 3.0956053733825684, 'timing/data': 0.03028270322829485, 'timing/model': 1.4327334854751825, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 0.77}) +[2026-06-15 18:53:49,555][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.2416195124387741, 'train/grad_norm_pre_clip': 6.632933139801025, 'timing/data': 0.03401452675461769, 'timing/model': 1.4283202281221747, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.77}) +[2026-06-15 18:53:55,351][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.09843603521585464, 'train/grad_norm_pre_clip': 4.370449066162109, 'timing/data': 0.030463874340057373, 'timing/model': 1.4399988632649183, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 0.78}) +[2026-06-15 18:54:01,061][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.1404542624950409, 'train/grad_norm_pre_clip': 5.145013809204102, 'timing/data': 0.030158852227032185, 'timing/model': 1.4041308769956231, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 0.78}) +[2026-06-15 18:54:06,893][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.12674610316753387, 'train/grad_norm_pre_clip': 4.370786666870117, 'timing/data': 0.02978811040520668, 'timing/model': 1.6680549457669258, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 0.78}) +[2026-06-15 18:54:12,442][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.27575793862342834, 'train/grad_norm_pre_clip': 6.810137748718262, 'timing/data': 0.029522987082600594, 'timing/model': 1.4261920610442758, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 0.78}) +[2026-06-15 18:54:18,247][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.219138503074646, 'train/grad_norm_pre_clip': 13.790736198425293, 'timing/data': 0.03146816976368427, 'timing/model': 1.4462543483823538, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 0.78}) +[2026-06-15 19:04:41,769][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.13851305842399597, 'train/grad_norm_pre_clip': 6.6321492195129395, 'eval/loss': 0.09890025854110718, 'eval/flap_precision': 0.8556149732620321, 'eval/flap_recall': 0.5818181818181818, 'eval/flap_f1': 0.6926406926406926, 'eval/flap_rate_model': 0.051944444444444446, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6926406926406926, 'timing/data': 0.031056196428835392, 'timing/model': 1.4526714803650975, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.78}) +[2026-06-15 19:04:41,770][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1700_state +[2026-06-15 19:04:41,776][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 19:04:49,256][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1700_state/model.safetensors +[2026-06-15 19:05:00,498][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1700_state/optimizer.bin +[2026-06-15 19:05:00,500][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1700_state/sampler.bin +[2026-06-15 19:05:00,500][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1700_state/sampler_1.bin +[2026-06-15 19:05:00,502][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1700_state/random_states_0.pkl +[2026-06-15 19:05:00,656][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1700 +[2026-06-15 19:05:00,657][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 19:05:00,671][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 19:05:06,334][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.2028588503599167, 'train/grad_norm_pre_clip': 8.044363975524902, 'timing/data': 0.030718302354216576, 'timing/model': 1.4396067475900054, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 0.78}) +[2026-06-15 19:05:12,155][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.3362536132335663, 'train/grad_norm_pre_clip': 12.093695640563965, 'timing/data': 0.030521142296493053, 'timing/model': 1.4553059190511703, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 0.78}) +[2026-06-15 19:05:17,830][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.32743963599205017, 'train/grad_norm_pre_clip': 19.138492584228516, 'timing/data': 0.03037907090038061, 'timing/model': 1.4007649524137378, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 0.78}) +[2026-06-15 19:05:23,427][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.16281603276729584, 'train/grad_norm_pre_clip': 4.101006984710693, 'timing/data': 0.029902957379817963, 'timing/model': 1.6427044523879886, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 0.78}) +[2026-06-15 19:05:29,045][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.21818792819976807, 'train/grad_norm_pre_clip': 4.368085861206055, 'timing/data': 0.030477652326226234, 'timing/model': 1.6508200988173485, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 0.78}) +[2026-06-15 19:05:34,431][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.1033811941742897, 'train/grad_norm_pre_clip': 3.4396626949310303, 'timing/data': 0.033334989100694656, 'timing/model': 1.4108929447829723, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 0.78}) +[2026-06-15 19:05:40,139][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.11239166557788849, 'train/grad_norm_pre_clip': 2.514127254486084, 'timing/data': 0.03087929543107748, 'timing/model': 1.4046337250620127, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 0.78}) +[2026-06-15 19:05:46,056][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.22419050335884094, 'train/grad_norm_pre_clip': 8.795917510986328, 'timing/data': 0.03242810256779194, 'timing/model': 1.463284382596612, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 0.78}) +[2026-06-15 19:05:51,860][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.3383609652519226, 'train/grad_norm_pre_clip': 8.926774024963379, 'timing/data': 0.03202377073466778, 'timing/model': 1.4018993051722646, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 0.78}) +[2026-06-15 19:05:57,646][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.15490594506263733, 'train/grad_norm_pre_clip': 3.7945351600646973, 'timing/data': 0.03836122713983059, 'timing/model': 1.4531568493694067, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.78}) +[2026-06-15 19:06:03,234][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.11800683289766312, 'train/grad_norm_pre_clip': 3.9499404430389404, 'timing/data': 0.02961126621812582, 'timing/model': 1.6139671141281724, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 0.78}) +[2026-06-15 19:06:08,743][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.15411026775836945, 'train/grad_norm_pre_clip': 4.618372440338135, 'timing/data': 0.035057349130511284, 'timing/model': 1.472394018433988, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 0.78}) +[2026-06-15 19:06:14,452][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.20358730852603912, 'train/grad_norm_pre_clip': 6.700822830200195, 'timing/data': 0.030589940957725048, 'timing/model': 1.4447718020528555, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 0.78}) +[2026-06-15 19:06:20,142][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.04914208501577377, 'train/grad_norm_pre_clip': 2.546996831893921, 'timing/data': 0.03536046575754881, 'timing/model': 1.4406874254345894, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 0.78}) +[2026-06-15 19:06:25,940][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.08844979852437973, 'train/grad_norm_pre_clip': 2.4101340770721436, 'timing/data': 0.0310624735429883, 'timing/model': 1.387181345373392, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.78}) +[2026-06-15 19:06:31,670][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.1393481194972992, 'train/grad_norm_pre_clip': 5.3078742027282715, 'timing/data': 0.030766556970775127, 'timing/model': 1.6709794020280242, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 0.78}) +[2026-06-15 19:06:37,139][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.06938758492469788, 'train/grad_norm_pre_clip': 3.844041347503662, 'timing/data': 0.0287995720282197, 'timing/model': 1.4210702925920486, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 0.79}) +[2026-06-15 19:06:42,846][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.44523754715919495, 'train/grad_norm_pre_clip': 15.356925964355469, 'timing/data': 0.02973152231425047, 'timing/model': 1.4304178357124329, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 0.79}) +[2026-06-15 19:06:48,542][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.336485892534256, 'train/grad_norm_pre_clip': 14.781059265136719, 'timing/data': 0.029544424265623093, 'timing/model': 1.4437572862952948, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 0.79}) +[2026-06-15 19:06:54,144][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.08202560991048813, 'train/grad_norm_pre_clip': 2.706066608428955, 'timing/data': 0.02880075667053461, 'timing/model': 1.4359958041459322, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.79}) +[2026-06-15 19:06:59,841][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.13095803558826447, 'train/grad_norm_pre_clip': 6.250144958496094, 'timing/data': 0.029243258759379387, 'timing/model': 1.439552592113614, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 0.79}) +[2026-06-15 19:07:05,586][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.16114678978919983, 'train/grad_norm_pre_clip': 9.112691879272461, 'timing/data': 0.039711988531053066, 'timing/model': 1.6746752467006445, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 0.79}) +[2026-06-15 19:07:11,139][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.14223097264766693, 'train/grad_norm_pre_clip': 6.87463903427124, 'timing/data': 0.03105412144213915, 'timing/model': 1.4392850929871202, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 0.79}) +[2026-06-15 19:07:16,842][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.18950049579143524, 'train/grad_norm_pre_clip': 9.784638404846191, 'timing/data': 0.029796413145959377, 'timing/model': 1.4171975227072835, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 0.79}) +[2026-06-15 19:07:22,543][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.2452010214328766, 'train/grad_norm_pre_clip': 6.209389686584473, 'timing/data': 0.030202925205230713, 'timing/model': 1.4375537438318133, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 0.79}) +[2026-06-15 19:07:28,242][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.24180211126804352, 'train/grad_norm_pre_clip': 7.43923807144165, 'timing/data': 0.027325947768986225, 'timing/model': 1.448530842550099, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 0.79}) +[2026-06-15 19:07:33,941][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.3466357886791229, 'train/grad_norm_pre_clip': 13.38386344909668, 'timing/data': 0.028261439874768257, 'timing/model': 1.4483025474473834, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 0.79}) +[2026-06-15 19:07:40,208][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.21443919837474823, 'train/grad_norm_pre_clip': 7.07630729675293, 'timing/data': 0.042665342800319195, 'timing/model': 2.0670671202242374, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 0.79}) +[2026-06-15 19:07:45,874][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.2688235938549042, 'train/grad_norm_pre_clip': 7.936767101287842, 'timing/data': 0.04180267546325922, 'timing/model': 1.437065833248198, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 0.79}) +[2026-06-15 19:07:51,971][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.17652802169322968, 'train/grad_norm_pre_clip': 6.401144981384277, 'timing/data': 0.03586223162710667, 'timing/model': 1.421238380484283, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.79}) +[2026-06-15 19:07:57,775][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.23069582879543304, 'train/grad_norm_pre_clip': 9.780308723449707, 'timing/data': 0.035878012888133526, 'timing/model': 1.4676237031817436, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 0.79}) +[2026-06-15 19:08:03,979][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.18720658123493195, 'train/grad_norm_pre_clip': 9.078728675842285, 'timing/data': 0.04147757310420275, 'timing/model': 1.466851493343711, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 0.79}) +[2026-06-15 19:08:09,862][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.0798971876502037, 'train/grad_norm_pre_clip': 3.095614433288574, 'timing/data': 0.03142736852169037, 'timing/model': 1.4072143379598856, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 0.79}) +[2026-06-15 19:08:15,693][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.24517522752285004, 'train/grad_norm_pre_clip': 8.907940864562988, 'timing/data': 0.031259999610483646, 'timing/model': 1.6847642790526152, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 0.79}) +[2026-06-15 19:08:21,248][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.13540653884410858, 'train/grad_norm_pre_clip': 4.648725986480713, 'timing/data': 0.032508911564946175, 'timing/model': 1.4188161808997393, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 0.79}) +[2026-06-15 19:08:27,052][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.10137462615966797, 'train/grad_norm_pre_clip': 4.169384956359863, 'timing/data': 0.03411241341382265, 'timing/model': 1.4351425459608436, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 0.79}) +[2026-06-15 19:08:32,864][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.2213289737701416, 'train/grad_norm_pre_clip': 6.697061061859131, 'timing/data': 0.030463917180895805, 'timing/model': 1.4498131619766355, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 0.79}) +[2026-06-15 19:08:38,652][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.16478392481803894, 'train/grad_norm_pre_clip': 5.185660362243652, 'timing/data': 0.03582435753196478, 'timing/model': 1.4389438712969422, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 0.79}) +[2026-06-15 19:08:44,464][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.2239817976951599, 'train/grad_norm_pre_clip': 4.822556018829346, 'timing/data': 0.03222121950238943, 'timing/model': 1.4021202297881246, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 0.8}) +[2026-06-15 19:08:50,294][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.1825714409351349, 'train/grad_norm_pre_clip': 4.951199531555176, 'timing/data': 0.030635991133749485, 'timing/model': 1.6861209822818637, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.8}) +[2026-06-15 19:08:55,859][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.3764452338218689, 'train/grad_norm_pre_clip': 9.52490520477295, 'timing/data': 0.0409813541918993, 'timing/model': 1.4088110132142901, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 0.8}) +[2026-06-15 19:09:01,658][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.15656673908233643, 'train/grad_norm_pre_clip': 4.59106969833374, 'timing/data': 0.03261434379965067, 'timing/model': 1.4297708990052342, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 0.8}) +[2026-06-15 19:09:07,459][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.20229044556617737, 'train/grad_norm_pre_clip': 5.023780822753906, 'timing/data': 0.03050230536609888, 'timing/model': 1.4429112449288368, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.8}) +[2026-06-15 19:09:13,273][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.2950330674648285, 'train/grad_norm_pre_clip': 6.317845344543457, 'timing/data': 0.03431527875363827, 'timing/model': 1.4408942172303796, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 0.8}) +[2026-06-15 19:09:19,070][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.2787795662879944, 'train/grad_norm_pre_clip': 5.496937274932861, 'timing/data': 0.030831478536128998, 'timing/model': 1.4042252469807863, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 0.8}) +[2026-06-15 19:09:24,900][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.2046002447605133, 'train/grad_norm_pre_clip': 5.465930938720703, 'timing/data': 0.03196256048977375, 'timing/model': 1.6765473866835237, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 0.8}) +[2026-06-15 19:09:30,555][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.19723176956176758, 'train/grad_norm_pre_clip': 3.490361452102661, 'timing/data': 0.03300294280052185, 'timing/model': 1.4339349959045649, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 0.8}) +[2026-06-15 19:09:36,361][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.2528773546218872, 'train/grad_norm_pre_clip': 5.648510456085205, 'timing/data': 0.03201692271977663, 'timing/model': 1.429680303670466, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 0.8}) +[2026-06-15 19:09:42,178][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.19349540770053864, 'train/grad_norm_pre_clip': 4.5843682289123535, 'timing/data': 0.030908473767340183, 'timing/model': 1.4671891275793314, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 0.8}) +[2026-06-15 19:09:47,954][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.1040172353386879, 'train/grad_norm_pre_clip': 3.3587801456451416, 'timing/data': 0.030836736783385277, 'timing/model': 1.4305253764614463, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.8}) +[2026-06-15 19:09:53,753][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.08957657963037491, 'train/grad_norm_pre_clip': 3.245187520980835, 'timing/data': 0.021284732036292553, 'timing/model': 1.4044786291196942, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 0.8}) +[2026-06-15 19:09:59,589][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.1487836092710495, 'train/grad_norm_pre_clip': 8.439408302307129, 'timing/data': 0.020302142947912216, 'timing/model': 1.6859442386776209, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 0.8}) +[2026-06-15 19:10:05,242][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.07361004501581192, 'train/grad_norm_pre_clip': 3.9108200073242188, 'timing/data': 0.021785453893244267, 'timing/model': 1.4326343946158886, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 0.8}) +[2026-06-15 19:10:11,066][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.311804860830307, 'train/grad_norm_pre_clip': 13.326112747192383, 'timing/data': 0.02013853471726179, 'timing/model': 1.471365557052195, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 0.8}) +[2026-06-15 19:10:16,848][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.10873227566480637, 'train/grad_norm_pre_clip': 5.1360344886779785, 'timing/data': 0.021674626506865025, 'timing/model': 1.439335617236793, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 0.8}) +[2026-06-15 19:10:22,645][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.12755626440048218, 'train/grad_norm_pre_clip': 14.38840103149414, 'timing/data': 0.02082789596170187, 'timing/model': 1.4443164458498359, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 0.8}) +[2026-06-15 19:10:28,452][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.35029372572898865, 'train/grad_norm_pre_clip': 20.33452033996582, 'timing/data': 0.0206887973472476, 'timing/model': 1.4092669989913702, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 0.8}) +[2026-06-15 19:10:34,287][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.20776242017745972, 'train/grad_norm_pre_clip': 14.66358470916748, 'timing/data': 0.020689467899501324, 'timing/model': 1.694915970787406, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 0.8}) +[2026-06-15 19:10:40,024][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.06842781603336334, 'train/grad_norm_pre_clip': 7.039795398712158, 'timing/data': 0.02076610829681158, 'timing/model': 1.6114110937342048, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 0.8}) +[2026-06-15 19:10:45,650][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.10156767815351486, 'train/grad_norm_pre_clip': 5.641133785247803, 'timing/data': 0.02506452426314354, 'timing/model': 1.433170479722321, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.8}) +[2026-06-15 19:10:51,448][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.19414645433425903, 'train/grad_norm_pre_clip': 13.449175834655762, 'timing/data': 0.023449404165148735, 'timing/model': 1.4518293682485819, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 0.81}) +[2026-06-15 19:10:57,250][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.17823655903339386, 'train/grad_norm_pre_clip': 18.343849182128906, 'timing/data': 0.021474482491612434, 'timing/model': 1.4435638496652246, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 0.81}) +[2026-06-15 19:11:03,065][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.21514785289764404, 'train/grad_norm_pre_clip': 10.149191856384277, 'timing/data': 0.025524367578327656, 'timing/model': 1.394094622693956, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 0.81}) +[2026-06-15 19:11:08,882][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.08667103946208954, 'train/grad_norm_pre_clip': 5.69627571105957, 'timing/data': 0.024681166745722294, 'timing/model': 1.6665772777050734, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 0.81}) +[2026-06-15 19:11:14,790][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.0824708566069603, 'train/grad_norm_pre_clip': 5.528115749359131, 'timing/data': 0.020754793658852577, 'timing/model': 1.6928672045469284, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 0.81}) +[2026-06-15 19:11:20,351][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.09590588510036469, 'train/grad_norm_pre_clip': 4.319066047668457, 'timing/data': 0.021133504807949066, 'timing/model': 1.4499299759045243, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 0.81}) +[2026-06-15 19:11:26,150][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.3405480980873108, 'train/grad_norm_pre_clip': 14.775618553161621, 'timing/data': 0.02319709863513708, 'timing/model': 1.4376358054578304, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 0.81}) +[2026-06-15 19:11:31,944][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.20534773170948029, 'train/grad_norm_pre_clip': 13.439488410949707, 'timing/data': 0.022234758362174034, 'timing/model': 1.4464794024825096, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 0.81}) +[2026-06-15 19:11:37,758][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.22040456533432007, 'train/grad_norm_pre_clip': 10.172584533691406, 'timing/data': 0.024281113408505917, 'timing/model': 1.3981654131785035, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 0.81}) +[2026-06-15 19:11:43,585][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.19807912409305573, 'train/grad_norm_pre_clip': 7.990595817565918, 'timing/data': 0.02322434075176716, 'timing/model': 1.6749671529978514, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.81}) +[2026-06-15 19:11:49,247][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.29321879148483276, 'train/grad_norm_pre_clip': 9.3772611618042, 'timing/data': 0.019190886989235878, 'timing/model': 1.4377420833334327, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 0.81}) +[2026-06-15 19:11:55,053][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.13227616250514984, 'train/grad_norm_pre_clip': 5.92732048034668, 'timing/data': 0.022285081446170807, 'timing/model': 1.439351567067206, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 0.81}) +[2026-06-15 19:12:00,849][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.23289965093135834, 'train/grad_norm_pre_clip': 6.823554515838623, 'timing/data': 0.02331385388970375, 'timing/model': 1.4506425969302654, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 0.81}) +[2026-06-15 19:12:06,651][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.123203806579113, 'train/grad_norm_pre_clip': 4.6393866539001465, 'timing/data': 0.02217632345855236, 'timing/model': 1.4023502059280872, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 0.81}) +[2026-06-15 19:12:12,370][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.10608436912298203, 'train/grad_norm_pre_clip': 4.919327735900879, 'timing/data': 0.02219510357826948, 'timing/model': 1.3970411466434598, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 0.81}) +[2026-06-15 19:12:18,142][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.14639586210250854, 'train/grad_norm_pre_clip': 5.043548107147217, 'timing/data': 0.021391313523054123, 'timing/model': 1.6326260156929493, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 0.81}) +[2026-06-15 19:12:23,749][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.20460988581180573, 'train/grad_norm_pre_clip': 5.578271389007568, 'timing/data': 0.020837862975895405, 'timing/model': 1.437129206955433, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 0.81}) +[2026-06-15 19:12:29,563][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.31879356503486633, 'train/grad_norm_pre_clip': 7.5026655197143555, 'timing/data': 0.023611249402165413, 'timing/model': 1.46661595441401, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 0.81}) +[2026-06-15 19:12:35,348][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.22949306666851044, 'train/grad_norm_pre_clip': 5.399674892425537, 'timing/data': 0.021897396072745323, 'timing/model': 1.4461897294968367, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 0.81}) +[2026-06-15 19:12:41,148][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.21174046397209167, 'train/grad_norm_pre_clip': 5.160977363586426, 'timing/data': 0.022571577690541744, 'timing/model': 1.4018479119986296, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.81}) +[2026-06-15 19:12:46,947][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.06707002967596054, 'train/grad_norm_pre_clip': 2.9635963439941406, 'timing/data': 0.02036584634333849, 'timing/model': 1.393840504810214, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 0.81}) +[2026-06-15 19:12:52,757][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.23849062621593475, 'train/grad_norm_pre_clip': 5.8583664894104, 'timing/data': 0.022494620643556118, 'timing/model': 1.6458919383585453, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 0.81}) +[2026-06-15 19:12:58,351][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.05086715891957283, 'train/grad_norm_pre_clip': 1.7750755548477173, 'timing/data': 0.021530479192733765, 'timing/model': 1.434996658936143, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 0.82}) +[2026-06-15 19:13:04,154][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.22163520753383636, 'train/grad_norm_pre_clip': 8.84473991394043, 'timing/data': 0.0214950293302536, 'timing/model': 1.4532117983326316, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 0.82}) +[2026-06-15 19:13:09,950][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.10867641866207123, 'train/grad_norm_pre_clip': 9.563008308410645, 'timing/data': 0.020260507240891457, 'timing/model': 1.4553093584254384, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.82}) +[2026-06-15 19:13:15,961][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.07886000722646713, 'train/grad_norm_pre_clip': 4.000708103179932, 'timing/data': 0.021695791743695736, 'timing/model': 1.4517602380365133, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 0.82}) +[2026-06-15 19:13:21,788][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.25484535098075867, 'train/grad_norm_pre_clip': 9.209214210510254, 'timing/data': 0.021343151107430458, 'timing/model': 1.6661754976958036, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 0.82}) +[2026-06-15 19:13:27,448][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.273875892162323, 'train/grad_norm_pre_clip': 13.417065620422363, 'timing/data': 0.02144414186477661, 'timing/model': 1.4456462664529681, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 0.82}) +[2026-06-15 19:13:33,273][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.22632606327533722, 'train/grad_norm_pre_clip': 7.846012115478516, 'timing/data': 0.029559968970716, 'timing/model': 1.4366356004029512, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 0.82}) +[2026-06-15 19:13:39,253][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.42118099331855774, 'train/grad_norm_pre_clip': 13.144248008728027, 'timing/data': 0.022256805561482906, 'timing/model': 1.443288148380816, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 0.82}) +[2026-06-15 19:13:44,959][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.2037016898393631, 'train/grad_norm_pre_clip': 9.353215217590332, 'timing/data': 0.021855885162949562, 'timing/model': 1.4558612760156393, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 0.82}) +[2026-06-15 19:13:50,749][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.31204864382743835, 'train/grad_norm_pre_clip': 8.2185697555542, 'timing/data': 0.022740536369383335, 'timing/model': 1.4402968166396022, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.82}) +[2026-06-15 19:13:56,588][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.2360246181488037, 'train/grad_norm_pre_clip': 8.96987533569336, 'timing/data': 0.02040095254778862, 'timing/model': 1.6960001392289996, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 0.82}) +[2026-06-15 19:14:02,258][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.32507219910621643, 'train/grad_norm_pre_clip': 8.551639556884766, 'timing/data': 0.021851105615496635, 'timing/model': 1.4584365487098694, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 0.82}) +[2026-06-15 19:14:08,058][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.13765326142311096, 'train/grad_norm_pre_clip': 6.120731353759766, 'timing/data': 0.020900565199553967, 'timing/model': 1.435862080194056, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 0.82}) +[2026-06-15 19:14:13,906][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.20529232919216156, 'train/grad_norm_pre_clip': 11.312555313110352, 'timing/data': 0.029643564485013485, 'timing/model': 1.4627580493688583, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 0.82}) +[2026-06-15 19:14:19,651][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.3787154257297516, 'train/grad_norm_pre_clip': 12.745864868164062, 'timing/data': 0.021724117919802666, 'timing/model': 1.4443450989201665, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 0.82}) +[2026-06-15 19:14:25,356][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.08614486455917358, 'train/grad_norm_pre_clip': 5.378587245941162, 'timing/data': 0.020833906717598438, 'timing/model': 1.4494781196117401, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 0.82}) +[2026-06-15 19:14:31,186][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.3820659816265106, 'train/grad_norm_pre_clip': 14.318134307861328, 'timing/data': 0.021669180132448673, 'timing/model': 1.691169678233564, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 0.82}) +[2026-06-15 19:24:59,680][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.33702340722084045, 'train/grad_norm_pre_clip': 9.947547912597656, 'eval/loss': 0.10391239821910858, 'eval/flap_precision': 0.6596385542168675, 'eval/flap_recall': 0.7963636363636364, 'eval/flap_f1': 0.7215815485996705, 'eval/flap_rate_model': 0.09222222222222222, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7215815485996705, 'timing/data': 0.02148326952010393, 'timing/model': 1.44806766230613, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 0.82}) +[2026-06-15 19:24:59,681][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1800_state +[2026-06-15 19:24:59,688][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 19:25:06,985][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1800_state/model.safetensors +[2026-06-15 19:25:18,360][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1800_state/optimizer.bin +[2026-06-15 19:25:18,361][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1800_state/sampler.bin +[2026-06-15 19:25:18,361][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1800_state/sampler_1.bin +[2026-06-15 19:25:18,362][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1800_state/random_states_0.pkl +[2026-06-15 19:25:18,518][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1800 +[2026-06-15 19:25:18,519][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 19:25:18,534][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 19:25:24,306][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.16479803621768951, 'train/grad_norm_pre_clip': 10.541406631469727, 'timing/data': 0.02501531969755888, 'timing/model': 1.708902350626886, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 0.82}) +[2026-06-15 19:25:29,698][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.13497890532016754, 'train/grad_norm_pre_clip': 3.1878881454467773, 'timing/data': 0.021356714889407158, 'timing/model': 1.4327712478116155, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 0.82}) +[2026-06-15 19:25:35,413][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.08845298737287521, 'train/grad_norm_pre_clip': 3.2540090084075928, 'timing/data': 0.023733165115118027, 'timing/model': 1.4260146114975214, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 0.82}) +[2026-06-15 19:25:41,142][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.10414345562458038, 'train/grad_norm_pre_clip': 2.9682581424713135, 'timing/data': 0.020303831435739994, 'timing/model': 1.4672250477597117, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 0.83}) +[2026-06-15 19:25:46,725][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.31478098034858704, 'train/grad_norm_pre_clip': 8.69625473022461, 'timing/data': 0.021162955090403557, 'timing/model': 1.3875901773571968, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 0.83}) +[2026-06-15 19:25:52,355][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.2976078391075134, 'train/grad_norm_pre_clip': 9.293484687805176, 'timing/data': 0.02255319058895111, 'timing/model': 1.6671650521457195, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 0.83}) +[2026-06-15 19:25:57,965][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.4506795108318329, 'train/grad_norm_pre_clip': 9.874024391174316, 'timing/data': 0.02285890281200409, 'timing/model': 1.672608096152544, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 0.83}) +[2026-06-15 19:26:03,344][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.1867549866437912, 'train/grad_norm_pre_clip': 5.863580703735352, 'timing/data': 0.020783863961696625, 'timing/model': 1.3745466032996774, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 0.83}) +[2026-06-15 19:26:08,924][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.1949368268251419, 'train/grad_norm_pre_clip': 5.069483757019043, 'timing/data': 0.02022384200245142, 'timing/model': 1.437096294015646, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 0.83}) +[2026-06-15 19:26:14,639][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.22608225047588348, 'train/grad_norm_pre_clip': 5.255547046661377, 'timing/data': 0.021996159106492996, 'timing/model': 1.4431896982714534, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 0.83}) +[2026-06-15 19:26:20,343][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.1825941950082779, 'train/grad_norm_pre_clip': 3.7496490478515625, 'timing/data': 0.021546456962823868, 'timing/model': 1.4165688315406442, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 0.83}) +[2026-06-15 19:26:26,080][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.06826489418745041, 'train/grad_norm_pre_clip': 2.372255563735962, 'timing/data': 0.022643832489848137, 'timing/model': 1.6837665159255266, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 0.83}) +[2026-06-15 19:26:31,640][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.19690658152103424, 'train/grad_norm_pre_clip': 4.389072418212891, 'timing/data': 0.02042155247181654, 'timing/model': 1.4399005137383938, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 0.83}) +[2026-06-15 19:26:37,436][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.20404215157032013, 'train/grad_norm_pre_clip': 5.4387407302856445, 'timing/data': 0.020373432897031307, 'timing/model': 1.4348620735108852, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 0.83}) +[2026-06-15 19:26:43,144][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.17790032923221588, 'train/grad_norm_pre_clip': 5.43757963180542, 'timing/data': 0.020582154393196106, 'timing/model': 1.436957131139934, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 0.83}) +[2026-06-15 19:26:48,845][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.21939899027347565, 'train/grad_norm_pre_clip': 5.729077339172363, 'timing/data': 0.021028165705502033, 'timing/model': 1.444262769073248, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 0.83}) +[2026-06-15 19:26:54,541][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.1874455213546753, 'train/grad_norm_pre_clip': 5.309912204742432, 'timing/data': 0.02038458362221718, 'timing/model': 1.4157381420955062, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 0.83}) +[2026-06-15 19:27:00,279][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.11776085197925568, 'train/grad_norm_pre_clip': 3.6403896808624268, 'timing/data': 0.02057082485407591, 'timing/model': 1.6754935765638947, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 0.83}) +[2026-06-15 19:27:06,016][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.05322042480111122, 'train/grad_norm_pre_clip': 3.2936580181121826, 'timing/data': 0.020880174823105335, 'timing/model': 1.6116799134761095, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 0.83}) +[2026-06-15 19:27:11,641][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.20511707663536072, 'train/grad_norm_pre_clip': 4.681948184967041, 'timing/data': 0.02098043542355299, 'timing/model': 1.445541249588132, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.83}) +[2026-06-15 19:27:17,445][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.1326836496591568, 'train/grad_norm_pre_clip': 18.081199645996094, 'timing/data': 0.020570875145494938, 'timing/model': 1.4545919159427285, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 0.83}) +[2026-06-15 19:27:23,248][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.1665533185005188, 'train/grad_norm_pre_clip': 7.235783100128174, 'timing/data': 0.02074055653065443, 'timing/model': 1.4258950464427471, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 0.83}) +[2026-06-15 19:27:29,037][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.3119589686393738, 'train/grad_norm_pre_clip': 9.384714126586914, 'timing/data': 0.019895702600479126, 'timing/model': 1.4263829765841365, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 0.83}) +[2026-06-15 19:27:34,805][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.3713536858558655, 'train/grad_norm_pre_clip': 17.012937545776367, 'timing/data': 0.02170110959559679, 'timing/model': 1.6919664042070508, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 0.83}) +[2026-06-15 19:27:40,689][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.23332346975803375, 'train/grad_norm_pre_clip': 8.25994873046875, 'timing/data': 0.020759155973792076, 'timing/model': 1.696269261650741, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 0.83}) +[2026-06-15 19:27:46,344][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.17497609555721283, 'train/grad_norm_pre_clip': 9.49942684173584, 'timing/data': 0.021577179431915283, 'timing/model': 1.4347260845825076, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 0.84}) +[2026-06-15 19:27:52,054][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.3827912211418152, 'train/grad_norm_pre_clip': 7.378341197967529, 'timing/data': 0.020866306498646736, 'timing/model': 1.3939731046557426, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 0.84}) +[2026-06-15 19:27:57,840][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.09327846765518188, 'train/grad_norm_pre_clip': 3.861616849899292, 'timing/data': 0.02018317487090826, 'timing/model': 1.4356246050447226, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 0.84}) +[2026-06-15 19:28:04,170][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.22683370113372803, 'train/grad_norm_pre_clip': 23.487764358520508, 'timing/data': 0.554649411700666, 'timing/model': 1.5118428217247128, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 0.84}) +[2026-06-15 19:28:09,883][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.12430498749017715, 'train/grad_norm_pre_clip': 4.401993274688721, 'timing/data': 0.019667881540954113, 'timing/model': 1.668678934685886, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 0.84}) +[2026-06-15 19:28:15,442][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.11517276614904404, 'train/grad_norm_pre_clip': 4.101365089416504, 'timing/data': 0.019421501085162163, 'timing/model': 1.4439269537106156, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 0.84}) +[2026-06-15 19:28:21,253][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.13577191531658173, 'train/grad_norm_pre_clip': 3.8224287033081055, 'timing/data': 0.02337928768247366, 'timing/model': 1.4464160120114684, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 0.84}) +[2026-06-15 19:28:26,954][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.19470328092575073, 'train/grad_norm_pre_clip': 5.526054382324219, 'timing/data': 0.021124999038875103, 'timing/model': 1.4496595226228237, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 0.84}) +[2026-06-15 19:28:32,637][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.43118128180503845, 'train/grad_norm_pre_clip': 10.658809661865234, 'timing/data': 0.02089950628578663, 'timing/model': 1.4038658887147903, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 0.84}) +[2026-06-15 19:28:38,327][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.2189575731754303, 'train/grad_norm_pre_clip': 5.382021427154541, 'timing/data': 0.02309050690382719, 'timing/model': 1.444302024319768, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 0.84}) +[2026-06-15 19:28:44,037][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.155771866440773, 'train/grad_norm_pre_clip': 8.687500953674316, 'timing/data': 0.02039346471428871, 'timing/model': 1.6338745895773172, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 0.84}) +[2026-06-15 19:28:49,553][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.09346521645784378, 'train/grad_norm_pre_clip': 3.384305238723755, 'timing/data': 0.022564594633877277, 'timing/model': 1.4584328196942806, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 0.84}) +[2026-06-15 19:28:55,247][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.2134784758090973, 'train/grad_norm_pre_clip': 5.373924255371094, 'timing/data': 0.02048894576728344, 'timing/model': 1.4334956109523773, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 0.84}) +[2026-06-15 19:29:00,956][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.13996678590774536, 'train/grad_norm_pre_clip': 4.402277946472168, 'timing/data': 0.02150151040405035, 'timing/model': 1.4493930749595165, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 0.84}) +[2026-06-15 19:29:06,746][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.2550780177116394, 'train/grad_norm_pre_clip': 5.096778392791748, 'timing/data': 0.020915177650749683, 'timing/model': 1.3916728179901838, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 0.84}) +[2026-06-15 19:29:12,562][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.18037758767604828, 'train/grad_norm_pre_clip': 3.399122714996338, 'timing/data': 0.021686461754143238, 'timing/model': 1.410184402950108, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.84}) +[2026-06-15 19:29:18,374][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.29663190245628357, 'train/grad_norm_pre_clip': 7.591324806213379, 'timing/data': 0.02414466254413128, 'timing/model': 1.648674019612372, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 0.84}) +[2026-06-15 19:29:23,966][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.0885927751660347, 'train/grad_norm_pre_clip': 2.4468469619750977, 'timing/data': 0.021701712161302567, 'timing/model': 1.4455931475386024, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 0.84}) +[2026-06-15 19:29:29,774][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.09941533952951431, 'train/grad_norm_pre_clip': 3.6308276653289795, 'timing/data': 0.020261965692043304, 'timing/model': 1.4304073993116617, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 0.84}) +[2026-06-15 19:29:35,560][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.3142127990722656, 'train/grad_norm_pre_clip': 14.016584396362305, 'timing/data': 0.021173538640141487, 'timing/model': 1.4457372212782502, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 0.84}) +[2026-06-15 19:29:41,354][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.04435417056083679, 'train/grad_norm_pre_clip': 3.23262357711792, 'timing/data': 0.0207484383136034, 'timing/model': 1.4041916700080037, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 0.84}) +[2026-06-15 19:29:47,186][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.13971193134784698, 'train/grad_norm_pre_clip': 23.69438362121582, 'timing/data': 0.02177808340638876, 'timing/model': 1.672122336924076, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 0.84}) +[2026-06-15 19:29:52,855][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.23125992715358734, 'train/grad_norm_pre_clip': 12.489925384521484, 'timing/data': 0.02041962742805481, 'timing/model': 1.4412425803020597, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 0.85}) +[2026-06-15 19:29:58,652][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.07035843282938004, 'train/grad_norm_pre_clip': 2.2209222316741943, 'timing/data': 0.020347407087683678, 'timing/model': 1.4254821697250009, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 0.85}) +[2026-06-15 19:30:04,353][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.20110033452510834, 'train/grad_norm_pre_clip': 7.7540507316589355, 'timing/data': 0.020503146573901176, 'timing/model': 1.438334290869534, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 0.85}) +[2026-06-15 19:30:10,161][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.14567896723747253, 'train/grad_norm_pre_clip': 14.405879974365234, 'timing/data': 0.020331826992332935, 'timing/model': 1.4394193813204765, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 0.85}) +[2026-06-15 19:30:15,956][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.14418166875839233, 'train/grad_norm_pre_clip': 9.433968544006348, 'timing/data': 0.020718729123473167, 'timing/model': 1.4276549369096756, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 0.85}) +[2026-06-15 19:30:21,805][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.16373810172080994, 'train/grad_norm_pre_clip': 8.248174667358398, 'timing/data': 0.020737098529934883, 'timing/model': 1.6882336307317019, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 0.85}) +[2026-06-15 19:30:27,361][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.37183094024658203, 'train/grad_norm_pre_clip': 16.33437728881836, 'timing/data': 0.022167674265801907, 'timing/model': 1.4193231547251344, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 0.85}) +[2026-06-15 19:30:33,054][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.05391857773065567, 'train/grad_norm_pre_clip': 1.5671086311340332, 'timing/data': 0.021945604123175144, 'timing/model': 1.4160954365506768, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.85}) +[2026-06-15 19:30:38,860][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.26991692185401917, 'train/grad_norm_pre_clip': 8.7493896484375, 'timing/data': 0.021679213270545006, 'timing/model': 1.4563311245292425, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 0.85}) +[2026-06-15 19:30:44,660][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.1483268290758133, 'train/grad_norm_pre_clip': 6.28080415725708, 'timing/data': 0.021183551289141178, 'timing/model': 1.4553937558084726, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 0.85}) +[2026-06-15 19:30:50,459][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.20623324811458588, 'train/grad_norm_pre_clip': 6.497110366821289, 'timing/data': 0.021205090917646885, 'timing/model': 1.4321957025676966, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 0.85}) +[2026-06-15 19:30:56,266][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.33800145983695984, 'train/grad_norm_pre_clip': 10.965917587280273, 'timing/data': 0.0218385336920619, 'timing/model': 1.6397475814446807, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 0.85}) +[2026-06-15 19:31:01,859][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.2485462725162506, 'train/grad_norm_pre_clip': 7.823901653289795, 'timing/data': 0.020229587331414223, 'timing/model': 1.443115440197289, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 0.85}) +[2026-06-15 19:31:07,664][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.14227482676506042, 'train/grad_norm_pre_clip': 4.226293087005615, 'timing/data': 0.024428585544228554, 'timing/model': 1.440352083183825, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 0.85}) +[2026-06-15 19:31:13,459][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.09877435117959976, 'train/grad_norm_pre_clip': 2.7481632232666016, 'timing/data': 0.021692024543881416, 'timing/model': 1.4474014891311526, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 0.85}) +[2026-06-15 19:31:19,260][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.03173595666885376, 'train/grad_norm_pre_clip': 1.1036818027496338, 'timing/data': 0.021481902338564396, 'timing/model': 1.4047009916976094, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 0.85}) +[2026-06-15 19:31:25,052][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.22547180950641632, 'train/grad_norm_pre_clip': 17.985857009887695, 'timing/data': 0.02021102700382471, 'timing/model': 1.4269959712401032, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 0.85}) +[2026-06-15 19:31:30,897][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.04818569868803024, 'train/grad_norm_pre_clip': 2.8457794189453125, 'timing/data': 0.02066141925752163, 'timing/model': 1.68408973980695, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 0.85}) +[2026-06-15 19:31:36,558][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.42146065831184387, 'train/grad_norm_pre_clip': 22.275474548339844, 'timing/data': 0.02173201460391283, 'timing/model': 1.435916731134057, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 0.85}) +[2026-06-15 19:31:42,359][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.0448136106133461, 'train/grad_norm_pre_clip': 3.5550150871276855, 'timing/data': 0.020860780030488968, 'timing/model': 1.4458286203444004, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 0.85}) +[2026-06-15 19:31:48,186][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.3819033205509186, 'train/grad_norm_pre_clip': 15.365165710449219, 'timing/data': 0.021949274465441704, 'timing/model': 1.4339825930073857, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 0.85}) +[2026-06-15 19:31:54,060][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.03200312331318855, 'train/grad_norm_pre_clip': 3.635667562484741, 'timing/data': 0.022206545807421207, 'timing/model': 1.4413223713636398, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 0.85}) +[2026-06-15 19:31:59,868][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.15786200761795044, 'train/grad_norm_pre_clip': 7.574267387390137, 'timing/data': 0.021468344144523144, 'timing/model': 1.4189392179250717, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 0.86}) +[2026-06-15 19:32:05,696][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.20679515600204468, 'train/grad_norm_pre_clip': 11.02705192565918, 'timing/data': 0.020930401049554348, 'timing/model': 1.6835812637582421, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 0.86}) +[2026-06-15 19:32:11,357][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.12984900176525116, 'train/grad_norm_pre_clip': 5.935696601867676, 'timing/data': 0.02175307460129261, 'timing/model': 1.4421245893463492, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 0.86}) +[2026-06-15 19:32:17,168][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.07983727753162384, 'train/grad_norm_pre_clip': 4.784923553466797, 'timing/data': 0.020953181199729443, 'timing/model': 1.4512811936438084, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 0.86}) +[2026-06-15 19:32:22,951][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.2911466658115387, 'train/grad_norm_pre_clip': 7.868618965148926, 'timing/data': 0.020966030657291412, 'timing/model': 1.434263063594699, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 0.86}) +[2026-06-15 19:32:28,655][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.037494249641895294, 'train/grad_norm_pre_clip': 3.6196095943450928, 'timing/data': 0.02098036091774702, 'timing/model': 1.4462399613112211, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 0.86}) +[2026-06-15 19:32:34,447][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.040501825511455536, 'train/grad_norm_pre_clip': 4.192709445953369, 'timing/data': 0.020213479176163673, 'timing/model': 1.4054415468126535, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 0.86}) +[2026-06-15 19:32:40,282][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.12811259925365448, 'train/grad_norm_pre_clip': 7.22886848449707, 'timing/data': 0.01998242735862732, 'timing/model': 1.6769610783085227, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 0.86}) +[2026-06-15 19:32:45,774][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.1074746623635292, 'train/grad_norm_pre_clip': 4.933053016662598, 'timing/data': 0.019347394816577435, 'timing/model': 1.3947063786908984, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 0.86}) +[2026-06-15 19:32:51,547][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.24111627042293549, 'train/grad_norm_pre_clip': 7.241567611694336, 'timing/data': 0.020718561485409737, 'timing/model': 1.4499785546213388, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 0.86}) +[2026-06-15 19:32:57,334][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.2663085162639618, 'train/grad_norm_pre_clip': 7.0113749504089355, 'timing/data': 0.023608192801475525, 'timing/model': 1.4250819310545921, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 0.86}) +[2026-06-15 19:33:03,121][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.2650905251502991, 'train/grad_norm_pre_clip': 9.415445327758789, 'timing/data': 0.019198943860828876, 'timing/model': 1.4069857159629464, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 0.86}) +[2026-06-15 19:33:08,914][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.14594195783138275, 'train/grad_norm_pre_clip': 3.9451181888580322, 'timing/data': 0.01985438633710146, 'timing/model': 1.4608063157647848, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 0.86}) +[2026-06-15 19:33:14,803][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.09329400211572647, 'train/grad_norm_pre_clip': 6.358906269073486, 'timing/data': 0.02606852538883686, 'timing/model': 1.6754782516509295, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 0.86}) +[2026-06-15 19:33:20,459][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.16715839505195618, 'train/grad_norm_pre_clip': 20.021400451660156, 'timing/data': 0.020790991373360157, 'timing/model': 1.4424682054668665, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 0.86}) +[2026-06-15 19:33:26,263][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.36524730920791626, 'train/grad_norm_pre_clip': 8.748552322387695, 'timing/data': 0.021606694906949997, 'timing/model': 1.4569919323548675, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 0.86}) +[2026-06-15 19:33:32,056][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.22771120071411133, 'train/grad_norm_pre_clip': 10.987202644348145, 'timing/data': 0.02060141135007143, 'timing/model': 1.445428466424346, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 0.86}) +[2026-06-15 19:33:37,855][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.12185290455818176, 'train/grad_norm_pre_clip': 3.6533150672912598, 'timing/data': 0.02071147132664919, 'timing/model': 1.4422801462933421, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 0.86}) +[2026-06-15 19:33:43,653][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.12698981165885925, 'train/grad_norm_pre_clip': 6.565201759338379, 'timing/data': 0.021111372858285904, 'timing/model': 1.4059429802000523, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 0.86}) +[2026-06-15 19:33:49,499][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.11094138026237488, 'train/grad_norm_pre_clip': 3.454021692276001, 'timing/data': 0.02340739220380783, 'timing/model': 1.6756412228569388, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 0.86}) +[2026-06-15 19:33:55,334][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.19911283254623413, 'train/grad_norm_pre_clip': 6.80572509765625, 'timing/data': 0.021069473586976528, 'timing/model': 1.610712131485343, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.86}) +[2026-06-15 19:34:00,857][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.19611705839633942, 'train/grad_norm_pre_clip': 7.384481430053711, 'timing/data': 0.02081833127886057, 'timing/model': 1.4523183153942227, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 0.86}) +[2026-06-15 19:34:06,654][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.052177418023347855, 'train/grad_norm_pre_clip': 2.4982357025146484, 'timing/data': 0.020909691229462624, 'timing/model': 1.4386589797213674, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 0.87}) +[2026-06-15 19:34:12,454][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.1598946899175644, 'train/grad_norm_pre_clip': 4.687902450561523, 'timing/data': 0.02285748068243265, 'timing/model': 1.4317897940054536, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 0.87}) +[2026-06-15 19:34:18,264][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.24431756138801575, 'train/grad_norm_pre_clip': 5.543492317199707, 'timing/data': 0.022522540763020515, 'timing/model': 1.4076979905366898, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 0.87}) +[2026-06-15 19:34:24,094][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.09998323023319244, 'train/grad_norm_pre_clip': 2.991877317428589, 'timing/data': 0.021198234520852566, 'timing/model': 1.6872026398777962, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 0.87}) +[2026-06-15 19:34:30,009][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.21190813183784485, 'train/grad_norm_pre_clip': 3.135244369506836, 'timing/data': 0.020265690051019192, 'timing/model': 1.690902921371162, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 0.87}) +[2026-06-15 19:34:35,656][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.2132713496685028, 'train/grad_norm_pre_clip': 5.757466793060303, 'timing/data': 0.02209902834147215, 'timing/model': 1.4398011527955532, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 0.87}) +[2026-06-15 19:34:41,461][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.23830576241016388, 'train/grad_norm_pre_clip': 11.395577430725098, 'timing/data': 0.020110249519348145, 'timing/model': 1.456460672430694, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 0.87}) +[2026-06-15 19:34:47,266][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.35233089327812195, 'train/grad_norm_pre_clip': 19.513898849487305, 'timing/data': 0.021717306226491928, 'timing/model': 1.4498353255912662, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 0.87}) +[2026-06-15 19:45:20,871][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.1671278476715088, 'train/grad_norm_pre_clip': 5.366939067840576, 'eval/loss': 0.08239717036485672, 'eval/flap_precision': 0.8295964125560538, 'eval/flap_recall': 0.6727272727272727, 'eval/flap_f1': 0.7429718875502007, 'eval/flap_rate_model': 0.06194444444444445, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7429718875502007, 'timing/data': 0.020720851607620716, 'timing/model': 1.417116624303162, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 0.87}) +[2026-06-15 19:45:20,871][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1900_state +[2026-06-15 19:45:20,878][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 19:45:28,512][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1900_state/model.safetensors +[2026-06-15 19:45:39,827][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1900_state/optimizer.bin +[2026-06-15 19:45:39,828][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1900_state/sampler.bin +[2026-06-15 19:45:39,828][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1900_state/sampler_1.bin +[2026-06-15 19:45:39,830][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1900_state/random_states_0.pkl +[2026-06-15 19:45:39,983][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1900 +[2026-06-15 19:45:39,983][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 19:45:39,999][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 19:45:45,850][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.10120709240436554, 'train/grad_norm_pre_clip': 2.9035260677337646, 'timing/data': 0.02527138590812683, 'timing/model': 1.4642096599563956, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 0.87}) +[2026-06-15 19:45:51,592][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.10801438987255096, 'train/grad_norm_pre_clip': 3.111443281173706, 'timing/data': 0.022007507272064686, 'timing/model': 1.6966715697199106, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 0.87}) +[2026-06-15 19:45:57,367][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.2667481005191803, 'train/grad_norm_pre_clip': 7.2285332679748535, 'timing/data': 0.019780305214226246, 'timing/model': 1.6828029723837972, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 0.87}) +[2026-06-15 19:46:02,824][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.10902613401412964, 'train/grad_norm_pre_clip': 3.994696617126465, 'timing/data': 0.02039484865963459, 'timing/model': 1.4359935419633985, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 0.87}) +[2026-06-15 19:46:08,523][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.10358143597841263, 'train/grad_norm_pre_clip': 4.4628801345825195, 'timing/data': 0.019110451452434063, 'timing/model': 1.427712930366397, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 0.87}) +[2026-06-15 19:46:14,236][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.2559725046157837, 'train/grad_norm_pre_clip': 5.857731819152832, 'timing/data': 0.02535537537187338, 'timing/model': 1.4146677618846297, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 0.87}) +[2026-06-15 19:46:19,945][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.37499216198921204, 'train/grad_norm_pre_clip': 12.065617561340332, 'timing/data': 0.021078770980238914, 'timing/model': 1.4084401037544012, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 0.87}) +[2026-06-15 19:46:26,214][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.13046859204769135, 'train/grad_norm_pre_clip': 6.0998358726501465, 'timing/data': 0.028270232491195202, 'timing/model': 2.07439380325377, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 0.87}) +[2026-06-15 19:46:32,473][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.055223092436790466, 'train/grad_norm_pre_clip': 2.4536848068237305, 'timing/data': 0.024851062335073948, 'timing/model': 2.0749636441469193, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 0.87}) +[2026-06-15 19:46:38,073][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.09403008222579956, 'train/grad_norm_pre_clip': 4.133264064788818, 'timing/data': 0.02635789941996336, 'timing/model': 1.4624570086598396, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 0.87}) +[2026-06-15 19:46:44,057][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.05425117537379265, 'train/grad_norm_pre_clip': 3.0089664459228516, 'timing/data': 0.02510143257677555, 'timing/model': 1.4473383873701096, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 0.87}) +[2026-06-15 19:46:50,063][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.09609587490558624, 'train/grad_norm_pre_clip': 3.1976630687713623, 'timing/data': 0.026377149857580662, 'timing/model': 1.4329496324062347, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 0.87}) +[2026-06-15 19:46:56,163][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.15234458446502686, 'train/grad_norm_pre_clip': 8.286280632019043, 'timing/data': 0.028956524096429348, 'timing/model': 1.4223906276747584, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 0.87}) +[2026-06-15 19:47:01,976][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.09943369030952454, 'train/grad_norm_pre_clip': 12.19813060760498, 'timing/data': 0.0264313705265522, 'timing/model': 1.4193812878802419, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 0.88}) +[2026-06-15 19:47:07,971][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.19135114550590515, 'train/grad_norm_pre_clip': 4.0682525634765625, 'timing/data': 0.024939770810306072, 'timing/model': 1.847911930643022, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 0.88}) +[2026-06-15 19:47:13,572][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.4165659546852112, 'train/grad_norm_pre_clip': 11.681999206542969, 'timing/data': 0.02685108222067356, 'timing/model': 1.450527167879045, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 0.88}) +[2026-06-15 19:47:19,667][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.2011307328939438, 'train/grad_norm_pre_clip': 9.406919479370117, 'timing/data': 0.02667207084596157, 'timing/model': 1.4504906851798296, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 0.88}) +[2026-06-15 19:47:25,871][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.21914900839328766, 'train/grad_norm_pre_clip': 6.627427101135254, 'timing/data': 0.025186252780258656, 'timing/model': 1.453979161567986, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 0.88}) +[2026-06-15 19:47:31,869][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.24488358199596405, 'train/grad_norm_pre_clip': 7.97244930267334, 'timing/data': 0.024656117893755436, 'timing/model': 1.4254783177748322, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 0.88}) +[2026-06-15 19:47:37,702][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.23208674788475037, 'train/grad_norm_pre_clip': 9.321367263793945, 'timing/data': 0.022551527246832848, 'timing/model': 1.6849197130650282, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 0.88}) +[2026-06-15 19:47:43,353][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.11850114166736603, 'train/grad_norm_pre_clip': 14.259445190429688, 'timing/data': 0.022626237012445927, 'timing/model': 1.4333907486870885, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 0.88}) +[2026-06-15 19:47:49,153][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.08025053888559341, 'train/grad_norm_pre_clip': 4.976513862609863, 'timing/data': 0.023185569792985916, 'timing/model': 1.4388961773365736, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 0.88}) +[2026-06-15 19:47:54,858][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.22433850169181824, 'train/grad_norm_pre_clip': 7.909116268157959, 'timing/data': 0.0230660792440176, 'timing/model': 1.455104705877602, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 0.88}) +[2026-06-15 19:48:00,659][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.23520158231258392, 'train/grad_norm_pre_clip': 8.836851119995117, 'timing/data': 0.02120580803602934, 'timing/model': 1.4646784979850054, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 0.88}) +[2026-06-15 19:48:06,485][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.2159038931131363, 'train/grad_norm_pre_clip': 12.6644868850708, 'timing/data': 0.022644846700131893, 'timing/model': 1.4041984183713794, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.88}) +[2026-06-15 19:48:12,390][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.09745649993419647, 'train/grad_norm_pre_clip': 8.200455665588379, 'timing/data': 0.02289751823991537, 'timing/model': 1.6812089290469885, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 0.88}) +[2026-06-15 19:48:17,943][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.20584270358085632, 'train/grad_norm_pre_clip': 8.391240119934082, 'timing/data': 0.023172727786004543, 'timing/model': 1.4419413674622774, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 0.88}) +[2026-06-15 19:48:23,637][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.15949779748916626, 'train/grad_norm_pre_clip': 3.9193215370178223, 'timing/data': 0.022502485662698746, 'timing/model': 1.4156235791742802, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 0.88}) +[2026-06-15 19:48:29,343][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.16288577020168304, 'train/grad_norm_pre_clip': 4.378824710845947, 'timing/data': 0.023694781586527824, 'timing/model': 1.4488876136019826, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 0.88}) +[2026-06-15 19:48:35,169][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.15782181918621063, 'train/grad_norm_pre_clip': 6.157092094421387, 'timing/data': 0.022293852642178535, 'timing/model': 1.4606465268880129, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 0.88}) +[2026-06-15 19:48:40,958][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.11561834067106247, 'train/grad_norm_pre_clip': 2.5225508213043213, 'timing/data': 0.0234328405931592, 'timing/model': 1.4325583511963487, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 0.88}) +[2026-06-15 19:48:46,832][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.06079426035284996, 'train/grad_norm_pre_clip': 1.9659781455993652, 'timing/data': 0.022922015748918056, 'timing/model': 1.7189958421513438, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 0.88}) +[2026-06-15 19:48:52,473][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.11824924498796463, 'train/grad_norm_pre_clip': 4.715625286102295, 'timing/data': 0.026021653786301613, 'timing/model': 1.43905792478472, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 0.88}) +[2026-06-15 19:48:58,362][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.08917687088251114, 'train/grad_norm_pre_clip': 2.648258924484253, 'timing/data': 0.02306903712451458, 'timing/model': 1.452699389308691, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 0.88}) +[2026-06-15 19:49:04,161][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.2584525942802429, 'train/grad_norm_pre_clip': 7.01183557510376, 'timing/data': 0.022584994323551655, 'timing/model': 1.4527751309797168, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 0.88}) +[2026-06-15 19:49:09,997][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.3206983804702759, 'train/grad_norm_pre_clip': 6.846662521362305, 'timing/data': 0.02440561354160309, 'timing/model': 1.4205875033512712, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 0.89}) +[2026-06-15 19:49:15,776][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.17460621893405914, 'train/grad_norm_pre_clip': 5.3892621994018555, 'timing/data': 0.023507269099354744, 'timing/model': 1.4093387108296156, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 0.89}) +[2026-06-15 19:49:21,617][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.19373759627342224, 'train/grad_norm_pre_clip': 7.237982273101807, 'timing/data': 0.023240067064762115, 'timing/model': 1.7038371656090021, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 0.89}) +[2026-06-15 19:49:27,266][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.10006064921617508, 'train/grad_norm_pre_clip': 5.623929023742676, 'timing/data': 0.022732543759047985, 'timing/model': 1.4540975270792842, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.89}) +[2026-06-15 19:49:33,062][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.20781512558460236, 'train/grad_norm_pre_clip': 4.9187703132629395, 'timing/data': 0.023065445944666862, 'timing/model': 1.4351110244169831, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 0.89}) +[2026-06-15 19:49:38,867][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.12066132575273514, 'train/grad_norm_pre_clip': 7.431269645690918, 'timing/data': 0.023225676268339157, 'timing/model': 1.4453746899962425, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 0.89}) +[2026-06-15 19:49:44,560][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.08361408114433289, 'train/grad_norm_pre_clip': 2.9247193336486816, 'timing/data': 0.02298910543322563, 'timing/model': 1.4401201512664557, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 0.89}) +[2026-06-15 19:49:50,368][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.32879430055618286, 'train/grad_norm_pre_clip': 8.225335121154785, 'timing/data': 0.02137758582830429, 'timing/model': 1.4141035107895732, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 0.89}) +[2026-06-15 19:49:56,218][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.2016308605670929, 'train/grad_norm_pre_clip': 7.711965560913086, 'timing/data': 0.023501377552747726, 'timing/model': 1.705015872605145, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 0.89}) +[2026-06-15 19:50:01,881][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.057814039289951324, 'train/grad_norm_pre_clip': 2.942413806915283, 'timing/data': 0.025745229795575142, 'timing/model': 1.4436502726748586, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 0.89}) +[2026-06-15 19:50:07,885][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.19581162929534912, 'train/grad_norm_pre_clip': 5.302657127380371, 'timing/data': 0.02548789791762829, 'timing/model': 1.4510581837967038, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 0.89}) +[2026-06-15 19:50:13,763][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.2795046865940094, 'train/grad_norm_pre_clip': 9.8428316116333, 'timing/data': 0.022512991912662983, 'timing/model': 1.4551850073039532, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 0.89}) +[2026-06-15 19:50:19,578][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.45868581533432007, 'train/grad_norm_pre_clip': 9.875349044799805, 'timing/data': 0.026351602748036385, 'timing/model': 1.4449672028422356, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 0.89}) +[2026-06-15 19:50:25,384][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.048269160091876984, 'train/grad_norm_pre_clip': 2.777611255645752, 'timing/data': 0.022787703201174736, 'timing/model': 1.4034767094999552, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 0.89}) +[2026-06-15 19:50:31,317][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.09074801951646805, 'train/grad_norm_pre_clip': 4.868188381195068, 'timing/data': 0.027081545442342758, 'timing/model': 1.6799955451861024, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 0.89}) +[2026-06-15 19:50:36,966][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.2596653997898102, 'train/grad_norm_pre_clip': 8.065396308898926, 'timing/data': 0.024746754206717014, 'timing/model': 1.4381868205964565, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 0.89}) +[2026-06-15 19:50:42,768][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.0962858498096466, 'train/grad_norm_pre_clip': 6.784262180328369, 'timing/data': 0.025738018564879894, 'timing/model': 1.4508371390402317, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 0.89}) +[2026-06-15 19:50:48,565][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.29677814245224, 'train/grad_norm_pre_clip': 12.138843536376953, 'timing/data': 0.028720935806632042, 'timing/model': 1.4418927216902375, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 0.89}) +[2026-06-15 19:50:54,360][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.10041527450084686, 'train/grad_norm_pre_clip': 6.194866180419922, 'timing/data': 0.025431745685636997, 'timing/model': 1.4485727110877633, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 0.89}) +[2026-06-15 19:51:00,174][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.21679344773292542, 'train/grad_norm_pre_clip': 6.383412837982178, 'timing/data': 0.023278075270354748, 'timing/model': 1.4083235561847687, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 0.89}) +[2026-06-15 19:51:06,002][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.1566912680864334, 'train/grad_norm_pre_clip': 5.053343772888184, 'timing/data': 0.022940313443541527, 'timing/model': 1.6886032717302442, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 0.89}) +[2026-06-15 19:51:11,659][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.3683464825153351, 'train/grad_norm_pre_clip': 9.144914627075195, 'timing/data': 0.02267670165747404, 'timing/model': 1.445167507044971, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 0.9}) +[2026-06-15 19:51:17,457][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.06638769805431366, 'train/grad_norm_pre_clip': 3.4057745933532715, 'timing/data': 0.02397609781473875, 'timing/model': 1.4342551408335567, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 0.9}) +[2026-06-15 19:51:23,268][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.1774972677230835, 'train/grad_norm_pre_clip': 8.284456253051758, 'timing/data': 0.022158817388117313, 'timing/model': 1.4505721610039473, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 0.9}) +[2026-06-15 19:51:29,067][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.2599388659000397, 'train/grad_norm_pre_clip': 15.566485404968262, 'timing/data': 0.024960453622043133, 'timing/model': 1.448505231179297, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.9}) +[2026-06-15 19:51:34,859][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.38550686836242676, 'train/grad_norm_pre_clip': 8.059077262878418, 'timing/data': 0.025138065218925476, 'timing/model': 1.4528933865949512, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 0.9}) +[2026-06-15 19:51:40,699][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.06584178656339645, 'train/grad_norm_pre_clip': 4.078092575073242, 'timing/data': 0.022849101573228836, 'timing/model': 1.6722676502540708, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 0.9}) +[2026-06-15 19:51:46,764][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.14182768762111664, 'train/grad_norm_pre_clip': 6.378305435180664, 'timing/data': 0.022999382577836514, 'timing/model': 1.8454707022756338, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 0.9}) +[2026-06-15 19:51:52,358][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.11873482912778854, 'train/grad_norm_pre_clip': 4.268441677093506, 'timing/data': 0.022873012349009514, 'timing/model': 1.4474255349487066, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 0.9}) +[2026-06-15 19:51:58,758][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.1663922220468521, 'train/grad_norm_pre_clip': 4.132559776306152, 'timing/data': 0.031138826161623, 'timing/model': 1.4332528011873364, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 0.9}) +[2026-06-15 19:52:04,567][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.20974808931350708, 'train/grad_norm_pre_clip': 8.478912353515625, 'timing/data': 0.022083678282797337, 'timing/model': 1.452616066671908, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 0.9}) +[2026-06-15 19:52:10,399][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.0602860301733017, 'train/grad_norm_pre_clip': 3.556885004043579, 'timing/data': 0.02114874217659235, 'timing/model': 1.4364473540335894, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 0.9}) +[2026-06-15 19:52:16,255][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.12641309201717377, 'train/grad_norm_pre_clip': 6.334793567657471, 'timing/data': 0.022251088172197342, 'timing/model': 1.7455686759203672, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 0.9}) +[2026-06-15 19:52:22,115][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.3098132312297821, 'train/grad_norm_pre_clip': 7.186075210571289, 'timing/data': 0.021396583877503872, 'timing/model': 1.6960445353761315, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 0.9}) +[2026-06-15 19:52:27,759][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.139661967754364, 'train/grad_norm_pre_clip': 6.432903289794922, 'timing/data': 0.021318322978913784, 'timing/model': 1.4385156147181988, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 0.9}) +[2026-06-15 19:52:33,502][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.3161694407463074, 'train/grad_norm_pre_clip': 11.638468742370605, 'timing/data': 0.020828220061957836, 'timing/model': 1.4240009104833007, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 0.9}) +[2026-06-15 19:52:39,376][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.19485779106616974, 'train/grad_norm_pre_clip': 4.948836326599121, 'timing/data': 0.021418452262878418, 'timing/model': 1.4726868215948343, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 0.9}) +[2026-06-15 19:52:45,196][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.09218389540910721, 'train/grad_norm_pre_clip': 3.4576544761657715, 'timing/data': 0.020537898875772953, 'timing/model': 1.42830104008317, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 0.9}) +[2026-06-15 19:52:51,126][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.3985009789466858, 'train/grad_norm_pre_clip': 15.610014915466309, 'timing/data': 0.02148730307817459, 'timing/model': 1.7264235466718674, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 0.9}) +[2026-06-15 19:52:56,853][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.2906418442726135, 'train/grad_norm_pre_clip': 10.051782608032227, 'timing/data': 0.019857674837112427, 'timing/model': 1.4466070029884577, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 0.9}) +[2026-06-15 19:53:02,585][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.2272651046514511, 'train/grad_norm_pre_clip': 6.743348598480225, 'timing/data': 0.018746688961982727, 'timing/model': 1.4338641185313463, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 0.9}) +[2026-06-15 19:53:08,296][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.1237565204501152, 'train/grad_norm_pre_clip': 4.242274284362793, 'timing/data': 0.020506287924945354, 'timing/model': 1.4088798016309738, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 0.9}) +[2026-06-15 19:53:14,278][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.44853273034095764, 'train/grad_norm_pre_clip': 14.812368392944336, 'timing/data': 0.021482163108885288, 'timing/model': 1.4431616365909576, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 0.9}) +[2026-06-15 19:53:20,062][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.2514713704586029, 'train/grad_norm_pre_clip': 5.039999485015869, 'timing/data': 0.021722114644944668, 'timing/model': 1.444329916499555, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 0.91}) +[2026-06-15 19:53:25,870][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.21278518438339233, 'train/grad_norm_pre_clip': 4.282496929168701, 'timing/data': 0.02146065142005682, 'timing/model': 1.6430564736947417, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 0.91}) +[2026-06-15 19:53:31,463][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.1972557008266449, 'train/grad_norm_pre_clip': 4.585188865661621, 'timing/data': 0.02125365100800991, 'timing/model': 1.4474298693239689, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 0.91}) +[2026-06-15 19:53:37,252][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.11907006800174713, 'train/grad_norm_pre_clip': 4.821976184844971, 'timing/data': 0.020117195323109627, 'timing/model': 1.4508295701816678, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 0.91}) +[2026-06-15 19:53:43,071][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.2830142080783844, 'train/grad_norm_pre_clip': 7.330427169799805, 'timing/data': 0.020630638115108013, 'timing/model': 1.4577811127528548, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 0.91}) +[2026-06-15 19:53:48,879][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.253162682056427, 'train/grad_norm_pre_clip': 7.0920186042785645, 'timing/data': 0.020832029171288013, 'timing/model': 1.401247639209032, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 0.91}) +[2026-06-15 19:53:54,681][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.1742066740989685, 'train/grad_norm_pre_clip': 9.88286018371582, 'timing/data': 0.022486157715320587, 'timing/model': 1.4211019584909081, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 0.91}) +[2026-06-15 19:54:00,476][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.13910731673240662, 'train/grad_norm_pre_clip': 4.4311113357543945, 'timing/data': 0.02191653475165367, 'timing/model': 1.649791936390102, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 0.91}) +[2026-06-15 19:54:06,065][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.3043292760848999, 'train/grad_norm_pre_clip': 7.075014114379883, 'timing/data': 0.021533642895519733, 'timing/model': 1.457031195051968, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 0.91}) +[2026-06-15 19:54:11,851][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.18584774434566498, 'train/grad_norm_pre_clip': 4.933208465576172, 'timing/data': 0.022265776060521603, 'timing/model': 1.446258102543652, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.91}) +[2026-06-15 19:54:17,648][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.1569686084985733, 'train/grad_norm_pre_clip': 2.9371628761291504, 'timing/data': 0.021329430863261223, 'timing/model': 1.4479611860588193, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 0.91}) +[2026-06-15 19:54:23,374][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.10018333792686462, 'train/grad_norm_pre_clip': 2.795377731323242, 'timing/data': 0.020247384905815125, 'timing/model': 1.3968660039827228, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 0.91}) +[2026-06-15 19:54:29,194][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.4256310760974884, 'train/grad_norm_pre_clip': 17.43885040283203, 'timing/data': 0.021828332915902138, 'timing/model': 1.6936952443793416, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 0.91}) +[2026-06-15 19:54:34,753][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.3602021038532257, 'train/grad_norm_pre_clip': 17.193180084228516, 'timing/data': 0.021293090656399727, 'timing/model': 1.452936745248735, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 0.91}) +[2026-06-15 19:54:40,547][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.18167993426322937, 'train/grad_norm_pre_clip': 12.111113548278809, 'timing/data': 0.020770068280398846, 'timing/model': 1.4509808998554945, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 0.91}) +[2026-06-15 19:54:46,247][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.22405162453651428, 'train/grad_norm_pre_clip': 11.421795845031738, 'timing/data': 0.022076714783906937, 'timing/model': 1.4470844641327858, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 0.91}) +[2026-06-15 19:54:52,060][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.23462790250778198, 'train/grad_norm_pre_clip': 12.697648048400879, 'timing/data': 0.022071143612265587, 'timing/model': 1.4401412140578032, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.91}) +[2026-06-15 19:54:57,861][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.2915085554122925, 'train/grad_norm_pre_clip': 10.136611938476562, 'timing/data': 0.021285000257194042, 'timing/model': 1.4429156417027116, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 0.91}) +[2026-06-15 19:55:03,700][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.08731576800346375, 'train/grad_norm_pre_clip': 1.9192818403244019, 'timing/data': 0.022199694998562336, 'timing/model': 1.688658680766821, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 0.91}) +[2026-06-15 19:55:09,350][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.19766981899738312, 'train/grad_norm_pre_clip': 3.164776086807251, 'timing/data': 0.02291748858988285, 'timing/model': 1.4393726037815213, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 0.91}) +[2026-06-15 19:55:15,148][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.02002764120697975, 'train/grad_norm_pre_clip': 1.1201645135879517, 'timing/data': 0.02107410877943039, 'timing/model': 1.4342930633574724, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 0.91}) +[2026-06-15 20:05:59,982][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.0861942246556282, 'train/grad_norm_pre_clip': 3.0137059688568115, 'eval/loss': 0.0840262845158577, 'eval/flap_precision': 0.7421602787456446, 'eval/flap_recall': 0.7745454545454545, 'eval/flap_f1': 0.7580071174377224, 'eval/flap_rate_model': 0.07972222222222222, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7580071174377224, 'timing/data': 0.022064263932406902, 'timing/model': 1.4483182905241847, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 0.91}) +[2026-06-15 20:05:59,983][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2000_state +[2026-06-15 20:05:59,989][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:06:08,238][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2000_state/model.safetensors +[2026-06-15 20:06:19,604][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2000_state/optimizer.bin +[2026-06-15 20:06:19,605][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2000_state/sampler.bin +[2026-06-15 20:06:19,606][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-15 20:06:19,607][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-15 20:06:19,777][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-15 20:06:19,777][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:06:19,794][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:06:25,736][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.22590893507003784, 'train/grad_norm_pre_clip': 6.765049934387207, 'timing/data': 0.024008008651435375, 'timing/model': 1.4408461954444647, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 0.92}) +[2026-06-15 20:06:31,541][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.17268650233745575, 'train/grad_norm_pre_clip': 6.205024719238281, 'timing/data': 0.03879423253238201, 'timing/model': 1.4185295645147562, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 0.92}) +[2026-06-15 20:06:37,231][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.13946613669395447, 'train/grad_norm_pre_clip': 5.128146648406982, 'timing/data': 0.022644531913101673, 'timing/model': 1.405226407572627, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 0.92}) +[2026-06-15 20:06:43,035][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.17222481966018677, 'train/grad_norm_pre_clip': 5.665310859680176, 'timing/data': 0.021559407003223896, 'timing/model': 1.408251159824431, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 0.92}) +[2026-06-15 20:06:48,799][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.15630440413951874, 'train/grad_norm_pre_clip': 5.454830169677734, 'timing/data': 0.02400478906929493, 'timing/model': 1.658327966928482, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 0.92}) +[2026-06-15 20:06:54,337][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.1286982148885727, 'train/grad_norm_pre_clip': 4.624386310577393, 'timing/data': 0.023825177922844887, 'timing/model': 1.4442011564970016, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 0.92}) +[2026-06-15 20:07:00,152][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.14885058999061584, 'train/grad_norm_pre_clip': 5.263026237487793, 'timing/data': 0.022030388936400414, 'timing/model': 1.4565689153969288, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 0.92}) +[2026-06-15 20:07:05,943][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.06480231881141663, 'train/grad_norm_pre_clip': 4.335642337799072, 'timing/data': 0.022975172847509384, 'timing/model': 1.4553198795765638, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 0.92}) +[2026-06-15 20:07:11,740][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.1254376620054245, 'train/grad_norm_pre_clip': 15.999953269958496, 'timing/data': 0.02058500237762928, 'timing/model': 1.3913546716794372, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 0.92}) +[2026-06-15 20:07:17,331][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.47204843163490295, 'train/grad_norm_pre_clip': 9.970734596252441, 'timing/data': 0.021196275018155575, 'timing/model': 1.4019301813095808, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 0.92}) +[2026-06-15 20:07:23,187][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.24715059995651245, 'train/grad_norm_pre_clip': 13.263258934020996, 'timing/data': 0.020886894315481186, 'timing/model': 1.6823822567239404, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 0.92}) +[2026-06-15 20:07:28,751][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.29884177446365356, 'train/grad_norm_pre_clip': 11.080229759216309, 'timing/data': 0.02079298347234726, 'timing/model': 1.4460375467315316, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 0.92}) +[2026-06-15 20:07:34,544][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.2391354739665985, 'train/grad_norm_pre_clip': 12.679315567016602, 'timing/data': 0.02054885122925043, 'timing/model': 1.43861510977149, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 0.92}) +[2026-06-15 20:07:40,344][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.06991129368543625, 'train/grad_norm_pre_clip': 10.819305419921875, 'timing/data': 0.0206961827352643, 'timing/model': 1.4475361928343773, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 0.92}) +[2026-06-15 20:07:45,960][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.33965930342674255, 'train/grad_norm_pre_clip': 17.683488845825195, 'timing/data': 0.0202505299821496, 'timing/model': 1.4445120263844728, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 0.92}) +[2026-06-15 20:07:51,747][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.17224685847759247, 'train/grad_norm_pre_clip': 3.094435453414917, 'timing/data': 0.020906344056129456, 'timing/model': 1.4297278225421906, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 0.92}) +[2026-06-15 20:07:57,590][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.11414487659931183, 'train/grad_norm_pre_clip': 7.409160137176514, 'timing/data': 0.02133589517325163, 'timing/model': 1.6789888720959425, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 0.92}) +[2026-06-15 20:08:03,251][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.17888621985912323, 'train/grad_norm_pre_clip': 8.749944686889648, 'timing/data': 0.019915398210287094, 'timing/model': 1.4273011879995465, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 0.92}) +[2026-06-15 20:08:09,055][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.29003214836120605, 'train/grad_norm_pre_clip': 9.00765609741211, 'timing/data': 0.020632262341678143, 'timing/model': 1.4355297582224011, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 0.92}) +[2026-06-15 20:08:14,864][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.276038259267807, 'train/grad_norm_pre_clip': 8.387921333312988, 'timing/data': 0.020523641258478165, 'timing/model': 1.4452575854957104, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 0.92}) +[2026-06-15 20:08:20,647][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.17157065868377686, 'train/grad_norm_pre_clip': 5.8674187660217285, 'timing/data': 0.02176405768841505, 'timing/model': 1.4340848987922072, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 0.92}) +[2026-06-15 20:08:26,464][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.09172507375478745, 'train/grad_norm_pre_clip': 6.239951133728027, 'timing/data': 0.020501751452684402, 'timing/model': 1.416164699010551, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 0.92}) +[2026-06-15 20:08:32,300][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.16281424462795258, 'train/grad_norm_pre_clip': 5.096505165100098, 'timing/data': 0.02135663479566574, 'timing/model': 1.6757594691589475, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 0.93}) +[2026-06-15 20:08:37,953][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.22675101459026337, 'train/grad_norm_pre_clip': 6.331096649169922, 'timing/data': 0.020263330079615116, 'timing/model': 1.4571639988571405, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 0.93}) +[2026-06-15 20:08:43,763][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.15918688476085663, 'train/grad_norm_pre_clip': 4.96228551864624, 'timing/data': 0.019665236584842205, 'timing/model': 1.4610047964379191, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 0.93}) +[2026-06-15 20:08:49,564][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.2398766428232193, 'train/grad_norm_pre_clip': 4.861159324645996, 'timing/data': 0.020713601261377335, 'timing/model': 1.4559014914557338, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 0.93}) +[2026-06-15 20:08:55,348][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.14229609072208405, 'train/grad_norm_pre_clip': 3.5346479415893555, 'timing/data': 0.0211460143327713, 'timing/model': 1.4488264257088304, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 0.93}) +[2026-06-15 20:09:01,036][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.06644853949546814, 'train/grad_norm_pre_clip': 2.9754130840301514, 'timing/data': 0.02032356057316065, 'timing/model': 1.3995661111548543, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 0.93}) +[2026-06-15 20:09:06,778][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.15763121843338013, 'train/grad_norm_pre_clip': 4.516021728515625, 'timing/data': 0.021024263463914394, 'timing/model': 1.6825472451746464, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 0.93}) +[2026-06-15 20:09:12,346][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.06426355242729187, 'train/grad_norm_pre_clip': 2.0374503135681152, 'timing/data': 0.019761517643928528, 'timing/model': 1.455931156873703, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.93}) +[2026-06-15 20:09:18,158][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.10954290628433228, 'train/grad_norm_pre_clip': 3.5144710540771484, 'timing/data': 0.019947188906371593, 'timing/model': 1.4531979318708181, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 0.93}) +[2026-06-15 20:09:23,955][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.22173427045345306, 'train/grad_norm_pre_clip': 5.257522106170654, 'timing/data': 0.019270644523203373, 'timing/model': 1.447112682275474, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 0.93}) +[2026-06-15 20:09:29,747][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.29134446382522583, 'train/grad_norm_pre_clip': 8.051361083984375, 'timing/data': 0.02001966815441847, 'timing/model': 1.4538582442328334, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 0.93}) +[2026-06-15 20:09:35,444][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.12684668600559235, 'train/grad_norm_pre_clip': 4.087007999420166, 'timing/data': 0.018578381277620792, 'timing/model': 1.4193044630810618, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 0.93}) +[2026-06-15 20:09:41,294][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.033382222056388855, 'train/grad_norm_pre_clip': 3.289297580718994, 'timing/data': 0.0205959714949131, 'timing/model': 1.684758709743619, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 0.93}) +[2026-06-15 20:09:47,135][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.14646145701408386, 'train/grad_norm_pre_clip': 8.351250648498535, 'timing/data': 0.023402156308293343, 'timing/model': 1.620602311566472, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 0.93}) +[2026-06-15 20:09:52,773][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.07050802558660507, 'train/grad_norm_pre_clip': 4.484062194824219, 'timing/data': 0.02102260384708643, 'timing/model': 1.4627919141203165, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.93}) +[2026-06-15 20:09:58,575][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.2409254014492035, 'train/grad_norm_pre_clip': 8.429422378540039, 'timing/data': 0.02050041127949953, 'timing/model': 1.4697858169674873, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 0.93}) +[2026-06-15 20:10:04,364][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.0989365205168724, 'train/grad_norm_pre_clip': 6.438262939453125, 'timing/data': 0.02146054618060589, 'timing/model': 1.4578181877732277, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 0.93}) +[2026-06-15 20:10:10,159][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.248775452375412, 'train/grad_norm_pre_clip': 12.20810604095459, 'timing/data': 0.021231594495475292, 'timing/model': 1.4134305091574788, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 0.93}) +[2026-06-15 20:10:15,998][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.035817209631204605, 'train/grad_norm_pre_clip': 2.9911749362945557, 'timing/data': 0.01955521572381258, 'timing/model': 1.6855601780116558, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 0.93}) +[2026-06-15 20:10:21,907][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.23427428305149078, 'train/grad_norm_pre_clip': 11.473177909851074, 'timing/data': 0.021327444352209568, 'timing/model': 1.7024974199011922, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 0.93}) +[2026-06-15 20:10:27,561][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.07822448015213013, 'train/grad_norm_pre_clip': 2.6715099811553955, 'timing/data': 0.027511846274137497, 'timing/model': 1.4163683596998453, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 0.93}) +[2026-06-15 20:10:33,396][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.14300808310508728, 'train/grad_norm_pre_clip': 5.2261552810668945, 'timing/data': 0.02347938623279333, 'timing/model': 1.4778771912679076, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 0.93}) +[2026-06-15 20:10:39,268][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.012153924442827702, 'train/grad_norm_pre_clip': 1.6031970977783203, 'timing/data': 0.01965915784239769, 'timing/model': 1.4630761994048953, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 0.94}) +[2026-06-15 20:10:45,063][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.42563262581825256, 'train/grad_norm_pre_clip': 12.55850887298584, 'timing/data': 0.02080709207803011, 'timing/model': 1.405066973529756, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 0.94}) +[2026-06-15 20:10:50,896][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.2466723918914795, 'train/grad_norm_pre_clip': 7.98593282699585, 'timing/data': 0.020016899332404137, 'timing/model': 1.6889255587011576, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 0.94}) +[2026-06-15 20:10:56,549][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.5360934734344482, 'train/grad_norm_pre_clip': 16.14289093017578, 'timing/data': 0.019767917692661285, 'timing/model': 1.4508195277303457, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 0.94}) +[2026-06-15 20:11:02,366][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.18695932626724243, 'train/grad_norm_pre_clip': 6.905661582946777, 'timing/data': 0.021677586250007153, 'timing/model': 1.4682159014046192, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 0.94}) +[2026-06-15 20:11:08,156][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.2304936945438385, 'train/grad_norm_pre_clip': 6.394709587097168, 'timing/data': 0.020091358572244644, 'timing/model': 1.4548980854451656, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 0.94}) +[2026-06-15 20:11:13,951][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.1348150223493576, 'train/grad_norm_pre_clip': 6.151167392730713, 'timing/data': 0.01978989876806736, 'timing/model': 1.4270203355699778, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 0.94}) +[2026-06-15 20:11:19,671][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.10379578918218613, 'train/grad_norm_pre_clip': 6.272856712341309, 'timing/data': 0.019708656705915928, 'timing/model': 1.4014732306823134, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 0.94}) +[2026-06-15 20:11:25,454][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.17088748514652252, 'train/grad_norm_pre_clip': 12.28344440460205, 'timing/data': 0.022636541165411472, 'timing/model': 1.6482218131422997, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 0.94}) +[2026-06-15 20:11:30,975][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.1962946206331253, 'train/grad_norm_pre_clip': 4.706684112548828, 'timing/data': 0.01979504805058241, 'timing/model': 1.399615989997983, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 0.94}) +[2026-06-15 20:11:36,751][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.1360701620578766, 'train/grad_norm_pre_clip': 3.121868133544922, 'timing/data': 0.022028477862477303, 'timing/model': 1.4350775722414255, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 0.94}) +[2026-06-15 20:11:42,552][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.07486094534397125, 'train/grad_norm_pre_clip': 3.2717032432556152, 'timing/data': 0.022210968658328056, 'timing/model': 1.4506440395489335, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 0.94}) +[2026-06-15 20:11:48,380][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.20946383476257324, 'train/grad_norm_pre_clip': 3.793151617050171, 'timing/data': 0.027285324409604073, 'timing/model': 1.4378193262964487, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 0.94}) +[2026-06-15 20:11:54,267][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.09972117841243744, 'train/grad_norm_pre_clip': 2.350584030151367, 'timing/data': 0.020946232602000237, 'timing/model': 1.4065672410652041, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 0.94}) +[2026-06-15 20:12:00,062][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.08705530315637589, 'train/grad_norm_pre_clip': 5.485015392303467, 'timing/data': 0.025875287130475044, 'timing/model': 1.6251745950430632, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 0.94}) +[2026-06-15 20:12:05,671][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.06589098274707794, 'train/grad_norm_pre_clip': 4.451599597930908, 'timing/data': 0.022147659212350845, 'timing/model': 1.457200238481164, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 0.94}) +[2026-06-15 20:12:11,461][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.15478883683681488, 'train/grad_norm_pre_clip': 5.085319519042969, 'timing/data': 0.02189982868731022, 'timing/model': 1.4327473873272538, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 0.94}) +[2026-06-15 20:12:17,281][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.20962466299533844, 'train/grad_norm_pre_clip': 8.651037216186523, 'timing/data': 0.020181029103696346, 'timing/model': 1.4781884495168924, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 0.94}) +[2026-06-15 20:12:23,066][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.08567240089178085, 'train/grad_norm_pre_clip': 3.7067208290100098, 'timing/data': 0.020929072983562946, 'timing/model': 1.4094682913273573, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 0.94}) +[2026-06-15 20:12:28,893][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.21729765832424164, 'train/grad_norm_pre_clip': 5.23350191116333, 'timing/data': 0.021023892797529697, 'timing/model': 1.6754916682839394, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 0.94}) +[2026-06-15 20:12:34,556][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.19540245831012726, 'train/grad_norm_pre_clip': 5.5153021812438965, 'timing/data': 0.020908323116600513, 'timing/model': 1.435407248325646, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.94}) +[2026-06-15 20:12:40,366][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.24788756668567657, 'train/grad_norm_pre_clip': 6.181051254272461, 'timing/data': 0.020236540585756302, 'timing/model': 1.4450384844094515, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 0.94}) +[2026-06-15 20:12:46,078][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.2714214026927948, 'train/grad_norm_pre_clip': 15.175460815429688, 'timing/data': 0.022875702939927578, 'timing/model': 1.4714272329583764, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 0.95}) +[2026-06-15 20:12:51,767][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.17468586564064026, 'train/grad_norm_pre_clip': 8.653608322143555, 'timing/data': 0.01996468845754862, 'timing/model': 1.432298220694065, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 0.95}) +[2026-06-15 20:12:57,557][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.25164544582366943, 'train/grad_norm_pre_clip': 14.288816452026367, 'timing/data': 0.02030904032289982, 'timing/model': 1.4619513349607587, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 0.95}) +[2026-06-15 20:13:03,401][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.05509548634290695, 'train/grad_norm_pre_clip': 3.5097241401672363, 'timing/data': 0.02048073150217533, 'timing/model': 1.694138533435762, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 0.95}) +[2026-06-15 20:13:09,061][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.06719671189785004, 'train/grad_norm_pre_clip': 4.031310558319092, 'timing/data': 0.02067611087113619, 'timing/model': 1.4449072107672691, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 0.95}) +[2026-06-15 20:13:14,864][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.18943391740322113, 'train/grad_norm_pre_clip': 8.53219223022461, 'timing/data': 0.0198143282905221, 'timing/model': 1.434334708377719, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 0.95}) +[2026-06-15 20:13:20,652][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.1929558515548706, 'train/grad_norm_pre_clip': 7.30224084854126, 'timing/data': 0.019158683717250824, 'timing/model': 1.4490550998598337, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 0.95}) +[2026-06-15 20:13:26,455][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.20547515153884888, 'train/grad_norm_pre_clip': 7.498047351837158, 'timing/data': 0.01910542417317629, 'timing/model': 1.445366270840168, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 0.95}) +[2026-06-15 20:13:32,263][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.12592820823192596, 'train/grad_norm_pre_clip': 5.490880012512207, 'timing/data': 0.022375740110874176, 'timing/model': 1.44139951094985, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 0.95}) +[2026-06-15 20:13:38,086][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.35913747549057007, 'train/grad_norm_pre_clip': 15.903287887573242, 'timing/data': 0.021570056676864624, 'timing/model': 1.682246781885624, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 0.95}) +[2026-06-15 20:13:43,758][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.05815144628286362, 'train/grad_norm_pre_clip': 3.6302552223205566, 'timing/data': 0.020298529416322708, 'timing/model': 1.4100686870515347, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 0.95}) +[2026-06-15 20:13:49,564][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.13905036449432373, 'train/grad_norm_pre_clip': 10.26729679107666, 'timing/data': 0.020562270656228065, 'timing/model': 1.4430457772687078, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 0.95}) +[2026-06-15 20:13:55,453][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.42263129353523254, 'train/grad_norm_pre_clip': 12.177786827087402, 'timing/data': 0.02031979989260435, 'timing/model': 1.4505708841606975, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 0.95}) +[2026-06-15 20:14:01,249][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.22243733704090118, 'train/grad_norm_pre_clip': 7.880636215209961, 'timing/data': 0.019269034266471863, 'timing/model': 1.4059821255505085, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 0.95}) +[2026-06-15 20:14:06,864][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.14950446784496307, 'train/grad_norm_pre_clip': 6.121312618255615, 'timing/data': 0.020981223322451115, 'timing/model': 1.4030448999255896, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 0.95}) +[2026-06-15 20:14:12,688][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.07635807245969772, 'train/grad_norm_pre_clip': 2.9304542541503906, 'timing/data': 0.019692815840244293, 'timing/model': 1.6895093945786357, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 0.95}) +[2026-06-15 20:14:18,262][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.11444661021232605, 'train/grad_norm_pre_clip': 4.253282070159912, 'timing/data': 0.019897167570888996, 'timing/model': 1.4305907636880875, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 0.95}) +[2026-06-15 20:14:24,065][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.12435483932495117, 'train/grad_norm_pre_clip': 4.223451137542725, 'timing/data': 0.02258796989917755, 'timing/model': 1.4338760999962687, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 0.95}) +[2026-06-15 20:14:29,866][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.052934203296899796, 'train/grad_norm_pre_clip': 2.5228023529052734, 'timing/data': 0.020452999509871006, 'timing/model': 1.4559423988685012, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 0.95}) +[2026-06-15 20:14:35,562][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.08475899696350098, 'train/grad_norm_pre_clip': 5.308811187744141, 'timing/data': 0.021147743798792362, 'timing/model': 1.4500351762399077, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.95}) +[2026-06-15 20:14:41,363][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.08954530954360962, 'train/grad_norm_pre_clip': 3.5974602699279785, 'timing/data': 0.02453309576958418, 'timing/model': 1.4122136626392603, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 0.95}) +[2026-06-15 20:14:47,201][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.037114400416612625, 'train/grad_norm_pre_clip': 2.6035919189453125, 'timing/data': 0.02108669187873602, 'timing/model': 1.6826660307124257, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 0.95}) +[2026-06-15 20:14:52,867][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.11352661997079849, 'train/grad_norm_pre_clip': 13.491811752319336, 'timing/data': 0.023426586762070656, 'timing/model': 1.4515254348516464, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 0.96}) +[2026-06-15 20:14:58,657][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.08706348389387131, 'train/grad_norm_pre_clip': 4.639756202697754, 'timing/data': 0.022098928689956665, 'timing/model': 1.4356864411383867, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 0.96}) +[2026-06-15 20:15:04,456][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.07771783322095871, 'train/grad_norm_pre_clip': 5.3877081871032715, 'timing/data': 0.020991479977965355, 'timing/model': 1.4310319544747472, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 0.96}) +[2026-06-15 20:15:10,262][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.049001339823007584, 'train/grad_norm_pre_clip': 3.747809886932373, 'timing/data': 0.02105261106044054, 'timing/model': 1.447219836525619, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 0.96}) +[2026-06-15 20:15:16,050][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.12385624647140503, 'train/grad_norm_pre_clip': 6.8227949142456055, 'timing/data': 0.023351826705038548, 'timing/model': 1.4176297625526786, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 0.96}) +[2026-06-15 20:15:21,886][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.11161614954471588, 'train/grad_norm_pre_clip': 8.498382568359375, 'timing/data': 0.020943048410117626, 'timing/model': 1.6894135819748044, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 0.96}) +[2026-06-15 20:15:27,460][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.21920469403266907, 'train/grad_norm_pre_clip': 25.333980560302734, 'timing/data': 0.020563995465636253, 'timing/model': 1.4596017822623253, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 0.96}) +[2026-06-15 20:15:33,273][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.07777462154626846, 'train/grad_norm_pre_clip': 9.859071731567383, 'timing/data': 0.02226884663105011, 'timing/model': 1.4683053940534592, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 0.96}) +[2026-06-15 20:15:39,071][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.16030587255954742, 'train/grad_norm_pre_clip': 4.121039867401123, 'timing/data': 0.024659193120896816, 'timing/model': 1.4597119120880961, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 0.96}) +[2026-06-15 20:15:44,859][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.05463051423430443, 'train/grad_norm_pre_clip': 5.893352031707764, 'timing/data': 0.021732022054493427, 'timing/model': 1.4442808367311954, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 0.96}) +[2026-06-15 20:15:50,875][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.12021543085575104, 'train/grad_norm_pre_clip': 9.178451538085938, 'timing/data': 0.020505033433437347, 'timing/model': 1.4468749379739165, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 0.96}) +[2026-06-15 20:26:34,774][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.43970543146133423, 'train/grad_norm_pre_clip': 20.001800537109375, 'eval/loss': 0.10038938373327255, 'eval/flap_precision': 0.7448275862068966, 'eval/flap_recall': 0.7854545454545454, 'eval/flap_f1': 0.7646017699115044, 'eval/flap_rate_model': 0.08055555555555556, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7646017699115044, 'timing/data': 0.020364862866699696, 'timing/model': 1.6915013128891587, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.96}) +[2026-06-15 20:26:34,775][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2100_state +[2026-06-15 20:26:34,782][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:26:42,482][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2100_state/model.safetensors +[2026-06-15 20:26:53,905][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2100_state/optimizer.bin +[2026-06-15 20:26:53,907][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2100_state/sampler.bin +[2026-06-15 20:26:53,907][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2100_state/sampler_1.bin +[2026-06-15 20:26:53,909][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2100_state/random_states_0.pkl +[2026-06-15 20:26:54,075][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2100 +[2026-06-15 20:26:54,076][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:26:54,091][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:26:59,854][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.06763749569654465, 'train/grad_norm_pre_clip': 8.789168357849121, 'timing/data': 0.02218893077224493, 'timing/model': 1.385750305838883, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 0.96}) +[2026-06-15 20:27:05,594][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.29438579082489014, 'train/grad_norm_pre_clip': 20.259403228759766, 'timing/data': 0.02141803689301014, 'timing/model': 1.7065220857039094, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 0.96}) +[2026-06-15 20:27:11,134][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.06863675266504288, 'train/grad_norm_pre_clip': 10.972916603088379, 'timing/data': 0.021507167257368565, 'timing/model': 1.442910997197032, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 0.96}) +[2026-06-15 20:27:16,867][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.28406259417533875, 'train/grad_norm_pre_clip': 10.644171714782715, 'timing/data': 0.02517116814851761, 'timing/model': 1.3842592388391495, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 0.96}) +[2026-06-15 20:27:22,739][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.1221027672290802, 'train/grad_norm_pre_clip': 2.9418785572052, 'timing/data': 0.028986589051783085, 'timing/model': 1.4450102047994733, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 0.96}) +[2026-06-15 20:27:28,545][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.16699010133743286, 'train/grad_norm_pre_clip': 12.600618362426758, 'timing/data': 0.022946995683014393, 'timing/model': 1.4590617157518864, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 0.96}) +[2026-06-15 20:27:34,239][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.23427337408065796, 'train/grad_norm_pre_clip': 13.266373634338379, 'timing/data': 0.020798394456505775, 'timing/model': 1.4043195405974984, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 0.96}) +[2026-06-15 20:27:39,981][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.34634777903556824, 'train/grad_norm_pre_clip': 10.618809700012207, 'timing/data': 0.020262000150978565, 'timing/model': 1.6910545090213418, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 0.96}) +[2026-06-15 20:27:45,434][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.2846086323261261, 'train/grad_norm_pre_clip': 11.914556503295898, 'timing/data': 0.02030367124825716, 'timing/model': 1.4353484688326716, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 0.96}) +[2026-06-15 20:27:51,232][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.10290876775979996, 'train/grad_norm_pre_clip': 6.560758113861084, 'timing/data': 0.021326725371181965, 'timing/model': 1.4355167737230659, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 0.96}) +[2026-06-15 20:27:56,936][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.19229456782341003, 'train/grad_norm_pre_clip': 7.349493980407715, 'timing/data': 0.018297629430890083, 'timing/model': 1.4341477202251554, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 0.97}) +[2026-06-15 20:28:02,639][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.160326287150383, 'train/grad_norm_pre_clip': 7.123434543609619, 'timing/data': 0.02490423619747162, 'timing/model': 1.3962725596502423, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 0.97}) +[2026-06-15 20:28:08,377][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.11660730838775635, 'train/grad_norm_pre_clip': 4.8658623695373535, 'timing/data': 0.02145735640078783, 'timing/model': 1.68092679977417, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 0.97}) +[2026-06-15 20:28:13,947][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.09026678651571274, 'train/grad_norm_pre_clip': 3.4544436931610107, 'timing/data': 0.02152222767472267, 'timing/model': 1.454658955335617, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 0.97}) +[2026-06-15 20:28:19,740][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.4046235680580139, 'train/grad_norm_pre_clip': 8.699974060058594, 'timing/data': 0.02038303017616272, 'timing/model': 1.438105309382081, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 0.97}) +[2026-06-15 20:28:25,440][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.06302136927843094, 'train/grad_norm_pre_clip': 1.9186782836914062, 'timing/data': 0.02144149597734213, 'timing/model': 1.4312963783740997, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 0.97}) +[2026-06-15 20:28:31,143][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.25640425086021423, 'train/grad_norm_pre_clip': 5.976198673248291, 'timing/data': 0.021099744364619255, 'timing/model': 1.4559572162106633, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 0.97}) +[2026-06-15 20:28:36,843][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.23737798631191254, 'train/grad_norm_pre_clip': 5.988061904907227, 'timing/data': 0.021090353839099407, 'timing/model': 1.403080184943974, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 0.97}) +[2026-06-15 20:28:42,589][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.22347921133041382, 'train/grad_norm_pre_clip': 4.245269775390625, 'timing/data': 0.020402660593390465, 'timing/model': 1.6984202675521374, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 0.97}) +[2026-06-15 20:28:48,402][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.15621346235275269, 'train/grad_norm_pre_clip': 4.074498176574707, 'timing/data': 0.02058033086359501, 'timing/model': 1.695097723044455, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 0.97}) +[2026-06-15 20:28:54,049][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.14114607870578766, 'train/grad_norm_pre_clip': 7.115659713745117, 'timing/data': 0.0205029109492898, 'timing/model': 1.4246424958109856, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 0.97}) +[2026-06-15 20:28:59,854][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.16154611110687256, 'train/grad_norm_pre_clip': 7.285375118255615, 'timing/data': 0.021648306399583817, 'timing/model': 1.4411713695153594, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 0.97}) +[2026-06-15 20:29:05,655][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.2677254378795624, 'train/grad_norm_pre_clip': 5.635441780090332, 'timing/data': 0.021476835943758488, 'timing/model': 1.44893903657794, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 0.97}) +[2026-06-15 20:29:11,476][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.06632698327302933, 'train/grad_norm_pre_clip': 2.1898105144500732, 'timing/data': 0.022559400647878647, 'timing/model': 1.3973704939708114, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 0.97}) +[2026-06-15 20:29:17,413][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.26508793234825134, 'train/grad_norm_pre_clip': 6.7119245529174805, 'timing/data': 0.020012687891721725, 'timing/model': 1.7127291085198522, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 0.97}) +[2026-06-15 20:29:23,077][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.22539936006069183, 'train/grad_norm_pre_clip': 3.8674681186676025, 'timing/data': 0.023884939029812813, 'timing/model': 1.4646164365112782, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 0.97}) +[2026-06-15 20:29:28,854][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.14285431802272797, 'train/grad_norm_pre_clip': 4.165258407592773, 'timing/data': 0.019937177188694477, 'timing/model': 1.457895465195179, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 0.97}) +[2026-06-15 20:29:34,650][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.16887307167053223, 'train/grad_norm_pre_clip': 5.53525972366333, 'timing/data': 0.020041516982018948, 'timing/model': 1.4542351309210062, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 0.97}) +[2026-06-15 20:29:40,343][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.1705944687128067, 'train/grad_norm_pre_clip': 6.189272880554199, 'timing/data': 0.020647771656513214, 'timing/model': 1.3973438292741776, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 0.97}) +[2026-06-15 20:29:46,103][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.31500324606895447, 'train/grad_norm_pre_clip': 7.440463542938232, 'timing/data': 0.02168348617851734, 'timing/model': 1.419973966665566, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 0.97}) +[2026-06-15 20:29:52,338][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.08718321472406387, 'train/grad_norm_pre_clip': 3.634861707687378, 'timing/data': 0.021651456132531166, 'timing/model': 2.005866757594049, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 0.97}) +[2026-06-15 20:29:57,978][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.09501034021377563, 'train/grad_norm_pre_clip': 4.037243843078613, 'timing/data': 0.02769314870238304, 'timing/model': 1.431270157918334, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 0.98}) +[2026-06-15 20:30:04,169][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.10639960318803787, 'train/grad_norm_pre_clip': 3.7870864868164062, 'timing/data': 0.020766472443938255, 'timing/model': 1.4384686723351479, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 0.98}) +[2026-06-15 20:30:09,958][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.09072866290807724, 'train/grad_norm_pre_clip': 3.080479621887207, 'timing/data': 0.02023002877831459, 'timing/model': 1.4424466080963612, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 0.98}) +[2026-06-15 20:30:15,749][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.031127983704209328, 'train/grad_norm_pre_clip': 1.7854759693145752, 'timing/data': 0.02075009047985077, 'timing/model': 1.4510150197893381, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.98}) +[2026-06-15 20:30:21,566][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.0885363221168518, 'train/grad_norm_pre_clip': 4.705626010894775, 'timing/data': 0.020481349900364876, 'timing/model': 1.4485219614580274, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 0.98}) +[2026-06-15 20:30:27,314][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.35392439365386963, 'train/grad_norm_pre_clip': 14.423701286315918, 'timing/data': 0.024392801336944103, 'timing/model': 1.6856337552890182, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 0.98}) +[2026-06-15 20:30:32,950][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.2698127329349518, 'train/grad_norm_pre_clip': 8.17465877532959, 'timing/data': 0.023250374011695385, 'timing/model': 1.4265786465257406, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 0.98}) +[2026-06-15 20:30:38,650][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.341613233089447, 'train/grad_norm_pre_clip': 13.026049613952637, 'timing/data': 0.021106443367898464, 'timing/model': 1.4193085730075836, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 0.98}) +[2026-06-15 20:30:44,464][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.1689726710319519, 'train/grad_norm_pre_clip': 6.838860988616943, 'timing/data': 0.021087992005050182, 'timing/model': 1.4611063115298748, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.98}) +[2026-06-15 20:30:50,250][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.26723960041999817, 'train/grad_norm_pre_clip': 8.700183868408203, 'timing/data': 0.02115765307098627, 'timing/model': 1.4471530634909868, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 0.98}) +[2026-06-15 20:30:55,957][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.16980235278606415, 'train/grad_norm_pre_clip': 7.343404769897461, 'timing/data': 0.02083861082792282, 'timing/model': 1.3956823945045471, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 0.98}) +[2026-06-15 20:31:01,787][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.19330191612243652, 'train/grad_norm_pre_clip': 6.991783142089844, 'timing/data': 0.021423674188554287, 'timing/model': 1.6874817470088601, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 0.98}) +[2026-06-15 20:31:07,460][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.5134164094924927, 'train/grad_norm_pre_clip': 13.293070793151855, 'timing/data': 0.021678275428712368, 'timing/model': 1.4567288532853127, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 0.98}) +[2026-06-15 20:31:13,252][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.13118937611579895, 'train/grad_norm_pre_clip': 5.927840232849121, 'timing/data': 0.023296304047107697, 'timing/model': 1.462780230678618, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 0.98}) +[2026-06-15 20:31:19,070][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.04403216391801834, 'train/grad_norm_pre_clip': 1.879334807395935, 'timing/data': 0.020335398614406586, 'timing/model': 1.472204358316958, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 0.98}) +[2026-06-15 20:31:24,863][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.23680295050144196, 'train/grad_norm_pre_clip': 7.372334957122803, 'timing/data': 0.021353724412620068, 'timing/model': 1.4152760310098529, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 0.98}) +[2026-06-15 20:31:30,665][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.11399184167385101, 'train/grad_norm_pre_clip': 4.8866729736328125, 'timing/data': 0.023904587142169476, 'timing/model': 1.4168986659497023, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 0.98}) +[2026-06-15 20:31:36,506][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.07068748772144318, 'train/grad_norm_pre_clip': 4.263817310333252, 'timing/data': 0.020805510692298412, 'timing/model': 1.6725476114079356, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 0.98}) +[2026-06-15 20:31:42,168][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.09404530376195908, 'train/grad_norm_pre_clip': 3.4590561389923096, 'timing/data': 0.02112297248095274, 'timing/model': 1.4595215637236834, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.98}) +[2026-06-15 20:31:47,971][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.3986280560493469, 'train/grad_norm_pre_clip': 16.325927734375, 'timing/data': 0.020659799687564373, 'timing/model': 1.4532290967181325, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 0.98}) +[2026-06-15 20:31:53,757][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.30150821805000305, 'train/grad_norm_pre_clip': 12.559270858764648, 'timing/data': 0.02199813723564148, 'timing/model': 1.4576011961326003, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 0.98}) +[2026-06-15 20:31:59,548][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.03579865023493767, 'train/grad_norm_pre_clip': 2.7513937950134277, 'timing/data': 0.023456545546650887, 'timing/model': 1.4404275920242071, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 0.98}) +[2026-06-15 20:32:05,350][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.440680593252182, 'train/grad_norm_pre_clip': 19.94243049621582, 'timing/data': 0.02089563012123108, 'timing/model': 1.414213540032506, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 0.99}) +[2026-06-15 20:32:11,185][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.19485266506671906, 'train/grad_norm_pre_clip': 10.289518356323242, 'timing/data': 0.02051152754575014, 'timing/model': 1.6908491626381874, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 0.99}) +[2026-06-15 20:32:16,848][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.2161482870578766, 'train/grad_norm_pre_clip': 11.027498245239258, 'timing/data': 0.02005565632134676, 'timing/model': 1.45635658595711, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 0.99}) +[2026-06-15 20:32:22,651][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.07545114308595657, 'train/grad_norm_pre_clip': 3.313699245452881, 'timing/data': 0.020807510241866112, 'timing/model': 1.4418925056234002, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 0.99}) +[2026-06-15 20:32:28,345][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.12264714390039444, 'train/grad_norm_pre_clip': 5.345294952392578, 'timing/data': 0.021188711747527122, 'timing/model': 1.4481376949697733, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 0.99}) +[2026-06-15 20:32:34,149][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.1350344866514206, 'train/grad_norm_pre_clip': 3.704009532928467, 'timing/data': 0.021947326138615608, 'timing/model': 1.4472762262448668, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 0.99}) +[2026-06-15 20:32:39,953][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.18481311202049255, 'train/grad_norm_pre_clip': 10.230679512023926, 'timing/data': 0.025923987850546837, 'timing/model': 1.3971869088709354, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.99}) +[2026-06-15 20:32:45,785][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.2703530490398407, 'train/grad_norm_pre_clip': 9.350936889648438, 'timing/data': 0.02326890267431736, 'timing/model': 1.6687994813546538, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 0.99}) +[2026-06-15 20:32:51,353][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.1770394742488861, 'train/grad_norm_pre_clip': 10.287662506103516, 'timing/data': 0.02123759314417839, 'timing/model': 1.4481954919174314, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 0.99}) +[2026-06-15 20:32:57,150][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.20642296969890594, 'train/grad_norm_pre_clip': 5.8764262199401855, 'timing/data': 0.02362790424376726, 'timing/model': 1.4316003918647766, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 0.99}) +[2026-06-15 20:33:02,959][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.315532386302948, 'train/grad_norm_pre_clip': 8.046562194824219, 'timing/data': 0.020556477829813957, 'timing/model': 1.4443678250536323, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 0.99}) +[2026-06-15 20:33:08,757][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.0857650563120842, 'train/grad_norm_pre_clip': 5.394581317901611, 'timing/data': 0.02069773990660906, 'timing/model': 1.4465519236400723, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 0.99}) +[2026-06-15 20:33:14,551][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.08150230348110199, 'train/grad_norm_pre_clip': 3.8796913623809814, 'timing/data': 0.021061130799353123, 'timing/model': 1.3976762928068638, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 0.99}) +[2026-06-15 20:33:20,392][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.13675642013549805, 'train/grad_norm_pre_clip': 4.526340007781982, 'timing/data': 0.021016051061451435, 'timing/model': 1.6771257845684886, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 0.99}) +[2026-06-15 20:33:26,046][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.12552782893180847, 'train/grad_norm_pre_clip': 5.466899394989014, 'timing/data': 0.020519218407571316, 'timing/model': 1.4300333578139544, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 0.99}) +[2026-06-15 20:33:31,856][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.13972073793411255, 'train/grad_norm_pre_clip': 6.487958908081055, 'timing/data': 0.020780240185558796, 'timing/model': 1.4457591772079468, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 0.99}) +[2026-06-15 20:33:37,653][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.27378344535827637, 'train/grad_norm_pre_clip': 8.82287311553955, 'timing/data': 0.02032648678869009, 'timing/model': 1.4566164324060082, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.99}) +[2026-06-15 20:33:43,471][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.12381578236818314, 'train/grad_norm_pre_clip': 5.769252300262451, 'timing/data': 0.020789490081369877, 'timing/model': 1.3991434164345264, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 0.99}) +[2026-06-15 20:33:49,246][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.15182289481163025, 'train/grad_norm_pre_clip': 6.041452407836914, 'timing/data': 0.02039117645472288, 'timing/model': 1.3948662718757987, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 0.99}) +[2026-06-15 20:33:55,100][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.1087758019566536, 'train/grad_norm_pre_clip': 4.680795192718506, 'timing/data': 0.021933864802122116, 'timing/model': 1.6919579105451703, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 0.99}) +[2026-06-15 20:34:00,937][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.1609942615032196, 'train/grad_norm_pre_clip': 7.054406642913818, 'timing/data': 0.021307341754436493, 'timing/model': 1.6250494653359056, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 0.99}) +[2026-06-15 20:34:06,560][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.11466343700885773, 'train/grad_norm_pre_clip': 6.892043590545654, 'timing/data': 0.02333463355898857, 'timing/model': 1.4443830624222755, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 0.99}) +[2026-06-15 20:34:12,448][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.5783337950706482, 'train/grad_norm_pre_clip': 12.421743392944336, 'timing/data': 0.02169343363493681, 'timing/model': 1.4355982039123774, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 1.0}) +[2026-06-15 20:34:18,249][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.19135460257530212, 'train/grad_norm_pre_clip': 8.52637004852295, 'timing/data': 0.02125727105885744, 'timing/model': 1.4341314435005188, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 1.0}) +[2026-06-15 20:34:24,065][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.0630078986287117, 'train/grad_norm_pre_clip': 2.985734462738037, 'timing/data': 0.02339182235300541, 'timing/model': 1.426885973662138, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 1.0}) +[2026-06-15 20:34:30,222][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.133499875664711, 'train/grad_norm_pre_clip': 7.539341449737549, 'timing/data': 0.03445165138691664, 'timing/model': 1.9751610178500414, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 1.0}) +[2026-06-15 20:34:36,063][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.011205016635358334, 'train/grad_norm_pre_clip': 0.9706982970237732, 'timing/data': 0.02246047742664814, 'timing/model': 1.7559825237840414, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 1.0}) +[2026-06-15 20:34:41,655][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.1766831874847412, 'train/grad_norm_pre_clip': 6.098148345947266, 'timing/data': 0.02452676836401224, 'timing/model': 1.429043765179813, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 1.0}) +[2026-06-15 20:34:47,462][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.025554858148097992, 'train/grad_norm_pre_clip': 2.0247631072998047, 'timing/data': 0.024917720817029476, 'timing/model': 1.4281230876222253, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 1.0}) +[2026-06-15 20:34:53,263][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.19289655983448029, 'train/grad_norm_pre_clip': 9.324508666992188, 'timing/data': 0.031395163387060165, 'timing/model': 1.4301556162536144, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 1.0}) +[2026-06-15 20:34:59,151][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.026340991258621216, 'train/grad_norm_pre_clip': 2.2674450874328613, 'timing/data': 0.02328524272888899, 'timing/model': 1.4227298954501748, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 1.0}) +[2026-06-15 20:35:05,053][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.04043421521782875, 'train/grad_norm_pre_clip': 4.178361892700195, 'timing/data': 0.022753329016268253, 'timing/model': 1.7329728584736586, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 1.0}) +[2026-06-15 20:35:10,657][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.195043683052063, 'train/grad_norm_pre_clip': 9.263728141784668, 'timing/data': 0.0221334146335721, 'timing/model': 1.4489717278629541, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 1.0}) +[2026-06-15 20:35:16,557][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.5405123829841614, 'train/grad_norm_pre_clip': 35.44689178466797, 'timing/data': 0.02422221750020981, 'timing/model': 1.4473382970318198, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 1.0}) +[2026-06-15 20:35:22,459][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.017283163964748383, 'train/grad_norm_pre_clip': 3.501251697540283, 'timing/data': 0.028634879738092422, 'timing/model': 1.4582960410043597, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 1.0}) +[2026-06-15 20:35:28,345][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.2843465209007263, 'train/grad_norm_pre_clip': 19.920169830322266, 'timing/data': 0.01907194033265114, 'timing/model': 1.4445667769759893, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 1.0}) +[2026-06-15 20:35:34,067][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.37160754203796387, 'train/grad_norm_pre_clip': 8.435101509094238, 'timing/data': 0.027846265584230423, 'timing/model': 1.4662542873993516, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 1.0}) +[2026-06-15 20:35:39,861][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.11147726327180862, 'train/grad_norm_pre_clip': 6.9211015701293945, 'timing/data': 0.019307310692965984, 'timing/model': 1.6600664090365171, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 1.0}) +[2026-06-15 20:35:45,470][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.550094485282898, 'train/grad_norm_pre_clip': 18.65731430053711, 'timing/data': 0.023881373926997185, 'timing/model': 1.4612795375287533, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 1.0}) +[2026-06-15 20:35:51,275][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.027697835117578506, 'train/grad_norm_pre_clip': 2.2808446884155273, 'timing/data': 0.024469178169965744, 'timing/model': 1.4610666334629059, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 1.0}) +[2026-06-15 20:35:57,068][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.11631383746862411, 'train/grad_norm_pre_clip': 9.85020637512207, 'timing/data': 0.024448128417134285, 'timing/model': 1.4596304139122367, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 1.0}) +[2026-06-15 20:36:02,889][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.1915379762649536, 'train/grad_norm_pre_clip': 16.418142318725586, 'timing/data': 0.02213356550782919, 'timing/model': 1.4217480355873704, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 1.0}) +[2026-06-15 20:36:08,777][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.08670075237751007, 'train/grad_norm_pre_clip': 7.826061248779297, 'timing/data': 0.02398788556456566, 'timing/model': 1.3968746922910213, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 1.0}) +[2026-06-15 20:36:14,554][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.21799811720848083, 'train/grad_norm_pre_clip': 15.004190444946289, 'timing/data': 0.019930844195187092, 'timing/model': 1.6390294134616852, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 1.0}) +[2026-06-15 20:36:20,155][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.23632173240184784, 'train/grad_norm_pre_clip': 10.151429176330566, 'timing/data': 0.01991416234523058, 'timing/model': 1.4562133979052305, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 1.01}) +[2026-06-15 20:36:25,948][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.26895272731781006, 'train/grad_norm_pre_clip': 13.160317420959473, 'timing/data': 0.019396771676838398, 'timing/model': 1.453268869780004, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 1.01}) +[2026-06-15 20:47:05,678][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.27200615406036377, 'train/grad_norm_pre_clip': 10.20061206817627, 'eval/loss': 0.11621244251728058, 'eval/flap_precision': 0.70625, 'eval/flap_recall': 0.8218181818181818, 'eval/flap_f1': 0.7596638655462185, 'eval/flap_rate_model': 0.08888888888888889, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7596638655462185, 'timing/data': 0.02152116224169731, 'timing/model': 1.4501238018274307, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 1.01}) +[2026-06-15 20:47:05,679][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2200_state +[2026-06-15 20:47:05,685][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 20:47:13,236][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2200_state/model.safetensors +[2026-06-15 20:47:24,555][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2200_state/optimizer.bin +[2026-06-15 20:47:24,556][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2200_state/sampler.bin +[2026-06-15 20:47:24,556][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2200_state/sampler_1.bin +[2026-06-15 20:47:24,558][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2200_state/random_states_0.pkl +[2026-06-15 20:47:24,723][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2200 +[2026-06-15 20:47:24,724][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 20:47:24,739][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 20:47:30,521][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.23080477118492126, 'train/grad_norm_pre_clip': 9.096912384033203, 'timing/data': 0.029457950964570045, 'timing/model': 1.395027182996273, 'learning_rate/qwen_vl_interface': 9.387268552285024e-06, 'learning_rate/action_model': 9.387268552285025e-05, 'epoch': 1.01}) +[2026-06-15 20:47:36,602][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.36355283856391907, 'train/grad_norm_pre_clip': 16.66916275024414, 'timing/data': 0.02961044106632471, 'timing/model': 1.4884005319327116, 'learning_rate/qwen_vl_interface': 9.385771559309125e-06, 'learning_rate/action_model': 9.385771559309125e-05, 'epoch': 1.01}) +[2026-06-15 20:47:42,355][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.2592927813529968, 'train/grad_norm_pre_clip': 18.92645263671875, 'timing/data': 0.020773387514054775, 'timing/model': 1.3818562543019652, 'learning_rate/qwen_vl_interface': 9.38427286627376e-06, 'learning_rate/action_model': 9.38427286627376e-05, 'epoch': 1.01}) +[2026-06-15 20:47:48,105][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.49602606892585754, 'train/grad_norm_pre_clip': 20.619457244873047, 'timing/data': 0.02409275248646736, 'timing/model': 1.7214403431862593, 'learning_rate/qwen_vl_interface': 9.382772473794985e-06, 'learning_rate/action_model': 9.382772473794985e-05, 'epoch': 1.01}) +[2026-06-15 20:47:53,584][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.4242539405822754, 'train/grad_norm_pre_clip': 16.137861251831055, 'timing/data': 0.01993719208985567, 'timing/model': 1.399300953373313, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 1.01}) +[2026-06-15 20:47:59,366][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.14888852834701538, 'train/grad_norm_pre_clip': 5.969790935516357, 'timing/data': 0.02035038359463215, 'timing/model': 1.3838684828951955, 'learning_rate/qwen_vl_interface': 9.37976659297492e-06, 'learning_rate/action_model': 9.379766592974921e-05, 'epoch': 1.01}) +[2026-06-15 20:48:05,163][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.22631488740444183, 'train/grad_norm_pre_clip': 4.520184516906738, 'timing/data': 0.02912213932722807, 'timing/model': 1.437699637375772, 'learning_rate/qwen_vl_interface': 9.378261105869237e-06, 'learning_rate/action_model': 9.378261105869237e-05, 'epoch': 1.01}) +[2026-06-15 20:48:10,969][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.24513797461986542, 'train/grad_norm_pre_clip': 5.087555408477783, 'timing/data': 0.021866551600396633, 'timing/model': 1.3934932015836239, 'learning_rate/qwen_vl_interface': 9.376753921791348e-06, 'learning_rate/action_model': 9.376753921791348e-05, 'epoch': 1.01}) +[2026-06-15 20:48:16,642][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.1358749270439148, 'train/grad_norm_pre_clip': 4.777440547943115, 'timing/data': 0.02196869347244501, 'timing/model': 1.4438146585598588, 'learning_rate/qwen_vl_interface': 9.375245041360805e-06, 'learning_rate/action_model': 9.375245041360804e-05, 'epoch': 1.01}) +[2026-06-15 20:48:22,501][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.1568141132593155, 'train/grad_norm_pre_clip': 3.824052095413208, 'timing/data': 0.023019257932901382, 'timing/model': 1.706547842361033, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 1.01}) +[2026-06-15 20:48:28,145][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.1548404097557068, 'train/grad_norm_pre_clip': 3.5897645950317383, 'timing/data': 0.022438745014369488, 'timing/model': 1.4333309745416045, 'learning_rate/qwen_vl_interface': 9.372222193923421e-06, 'learning_rate/action_model': 9.372222193923421e-05, 'epoch': 1.01}) +[2026-06-15 20:48:33,937][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.18650491535663605, 'train/grad_norm_pre_clip': 5.069619178771973, 'timing/data': 0.02294936776161194, 'timing/model': 1.4414097238332033, 'learning_rate/qwen_vl_interface': 9.370708228159158e-06, 'learning_rate/action_model': 9.370708228159159e-05, 'epoch': 1.01}) +[2026-06-15 20:48:39,747][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.18998901546001434, 'train/grad_norm_pre_clip': 5.086142539978027, 'timing/data': 0.020714549347758293, 'timing/model': 1.4244803572073579, 'learning_rate/qwen_vl_interface': 9.369192568527398e-06, 'learning_rate/action_model': 9.369192568527398e-05, 'epoch': 1.01}) +[2026-06-15 20:48:45,345][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.4770798683166504, 'train/grad_norm_pre_clip': 10.223400115966797, 'timing/data': 0.02318547200411558, 'timing/model': 1.4397700652480125, 'learning_rate/qwen_vl_interface': 9.367675215651168e-06, 'learning_rate/action_model': 9.367675215651168e-05, 'epoch': 1.01}) +[2026-06-15 20:48:51,161][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.32095494866371155, 'train/grad_norm_pre_clip': 9.018394470214844, 'timing/data': 0.02159520424902439, 'timing/model': 1.4284372320398688, 'learning_rate/qwen_vl_interface': 9.366156170154194e-06, 'learning_rate/action_model': 9.366156170154194e-05, 'epoch': 1.01}) +[2026-06-15 20:48:56,916][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.21376779675483704, 'train/grad_norm_pre_clip': 5.500415325164795, 'timing/data': 0.02169498521834612, 'timing/model': 1.7200694475322962, 'learning_rate/qwen_vl_interface': 9.3646354326609e-06, 'learning_rate/action_model': 9.364635432660899e-05, 'epoch': 1.01}) +[2026-06-15 20:49:02,551][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.2626231908798218, 'train/grad_norm_pre_clip': 4.9036431312561035, 'timing/data': 0.026405708864331245, 'timing/model': 1.4314961731433868, 'learning_rate/qwen_vl_interface': 9.3631130037964e-06, 'learning_rate/action_model': 9.363113003796401e-05, 'epoch': 1.01}) +[2026-06-15 20:49:08,341][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.11176395416259766, 'train/grad_norm_pre_clip': 2.427325487136841, 'timing/data': 0.021964835934340954, 'timing/model': 1.410319629125297, 'learning_rate/qwen_vl_interface': 9.361588884186513e-06, 'learning_rate/action_model': 9.361588884186513e-05, 'epoch': 1.01}) +[2026-06-15 20:49:14,368][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.2295730710029602, 'train/grad_norm_pre_clip': 3.198500633239746, 'timing/data': 0.027700604870915413, 'timing/model': 1.4494742518290877, 'learning_rate/qwen_vl_interface': 9.360063074457744e-06, 'learning_rate/action_model': 9.360063074457743e-05, 'epoch': 1.01}) +[2026-06-15 20:49:20,991][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.1799335926771164, 'train/grad_norm_pre_clip': 3.0544657707214355, 'timing/data': 0.02865590061992407, 'timing/model': 1.4315752359107137, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 1.02}) +[2026-06-15 20:49:26,970][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.20182229578495026, 'train/grad_norm_pre_clip': 7.192133903503418, 'timing/data': 0.026313388720154762, 'timing/model': 1.4038031864911318, 'learning_rate/qwen_vl_interface': 9.357006387153065e-06, 'learning_rate/action_model': 9.357006387153065e-05, 'epoch': 1.02}) +[2026-06-15 20:49:32,822][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.09844744950532913, 'train/grad_norm_pre_clip': 4.377750873565674, 'timing/data': 0.025421714410185814, 'timing/model': 1.7220249585807323, 'learning_rate/qwen_vl_interface': 9.355475510833643e-06, 'learning_rate/action_model': 9.355475510833643e-05, 'epoch': 1.02}) +[2026-06-15 20:49:38,449][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.39139342308044434, 'train/grad_norm_pre_clip': 7.225043773651123, 'timing/data': 0.025047832168638706, 'timing/model': 1.4476565718650818, 'learning_rate/qwen_vl_interface': 9.353942946908318e-06, 'learning_rate/action_model': 9.353942946908317e-05, 'epoch': 1.02}) +[2026-06-15 20:49:44,253][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.20043714344501495, 'train/grad_norm_pre_clip': 5.123354911804199, 'timing/data': 0.02501596137881279, 'timing/model': 1.4391091745346785, 'learning_rate/qwen_vl_interface': 9.352408696007065e-06, 'learning_rate/action_model': 9.352408696007065e-05, 'epoch': 1.02}) +[2026-06-15 20:49:50,067][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.40002650022506714, 'train/grad_norm_pre_clip': 8.238459587097168, 'timing/data': 0.027215573005378246, 'timing/model': 1.4471914051100612, 'learning_rate/qwen_vl_interface': 9.350872758760558e-06, 'learning_rate/action_model': 9.350872758760559e-05, 'epoch': 1.02}) +[2026-06-15 20:49:55,858][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.08950595557689667, 'train/grad_norm_pre_clip': 3.020752429962158, 'timing/data': 0.0299667464569211, 'timing/model': 1.436612687073648, 'learning_rate/qwen_vl_interface': 9.349335135800164e-06, 'learning_rate/action_model': 9.349335135800163e-05, 'epoch': 1.02}) +[2026-06-15 20:50:01,757][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.1265331208705902, 'train/grad_norm_pre_clip': 8.004173278808594, 'timing/data': 0.027833135798573494, 'timing/model': 1.4546793196350336, 'learning_rate/qwen_vl_interface': 9.347795827757942e-06, 'learning_rate/action_model': 9.347795827757941e-05, 'epoch': 1.02}) +[2026-06-15 20:50:07,608][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.24864254891872406, 'train/grad_norm_pre_clip': 6.553889274597168, 'timing/data': 0.03166494611650705, 'timing/model': 1.6906710965558887, 'learning_rate/qwen_vl_interface': 9.346254835266643e-06, 'learning_rate/action_model': 9.346254835266642e-05, 'epoch': 1.02}) +[2026-06-15 20:50:13,254][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.189005047082901, 'train/grad_norm_pre_clip': 9.166024208068848, 'timing/data': 0.02441993821412325, 'timing/model': 1.4417197974398732, 'learning_rate/qwen_vl_interface': 9.344712158959709e-06, 'learning_rate/action_model': 9.344712158959709e-05, 'epoch': 1.02}) +[2026-06-15 20:50:19,062][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.23219479620456696, 'train/grad_norm_pre_clip': 11.242167472839355, 'timing/data': 0.024655019864439964, 'timing/model': 1.4559233784675598, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 1.02}) +[2026-06-15 20:50:24,870][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.12307865917682648, 'train/grad_norm_pre_clip': 4.9208550453186035, 'timing/data': 0.02455881889909506, 'timing/model': 1.4700172413140535, 'learning_rate/qwen_vl_interface': 9.341621757436177e-06, 'learning_rate/action_model': 9.341621757436177e-05, 'epoch': 1.02}) +[2026-06-15 20:50:30,768][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.3398371636867523, 'train/grad_norm_pre_clip': 7.456841945648193, 'timing/data': 0.02766590565443039, 'timing/model': 1.4619545862078667, 'learning_rate/qwen_vl_interface': 9.340074033489924e-06, 'learning_rate/action_model': 9.340074033489923e-05, 'epoch': 1.02}) +[2026-06-15 20:50:36,581][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.08315553516149521, 'train/grad_norm_pre_clip': 4.1360979080200195, 'timing/data': 0.022017966024577618, 'timing/model': 1.4025781219825149, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 1.02}) +[2026-06-15 20:50:42,400][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.13130439817905426, 'train/grad_norm_pre_clip': 10.848603248596191, 'timing/data': 0.020223915576934814, 'timing/model': 1.6882372684776783, 'learning_rate/qwen_vl_interface': 9.336973542409505e-06, 'learning_rate/action_model': 9.336973542409504e-05, 'epoch': 1.02}) +[2026-06-15 20:50:48,252][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.013084222562611103, 'train/grad_norm_pre_clip': 0.9179270267486572, 'timing/data': 0.02144624199718237, 'timing/model': 1.6440346036106348, 'learning_rate/qwen_vl_interface': 9.33542077654983e-06, 'learning_rate/action_model': 9.33542077654983e-05, 'epoch': 1.02}) +[2026-06-15 20:50:53,864][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.020560331642627716, 'train/grad_norm_pre_clip': 1.9508447647094727, 'timing/data': 0.02038400713354349, 'timing/model': 1.4660247080028057, 'learning_rate/qwen_vl_interface': 9.333866331327995e-06, 'learning_rate/action_model': 9.333866331327995e-05, 'epoch': 1.02}) +[2026-06-15 20:50:59,656][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.24540944397449493, 'train/grad_norm_pre_clip': 13.186967849731445, 'timing/data': 0.020497857592999935, 'timing/model': 1.4552661702036858, 'learning_rate/qwen_vl_interface': 9.332310207382972e-06, 'learning_rate/action_model': 9.332310207382972e-05, 'epoch': 1.02}) +[2026-06-15 20:51:05,452][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.05737778916954994, 'train/grad_norm_pre_clip': 5.277258396148682, 'timing/data': 0.020776488818228245, 'timing/model': 1.4481570506468415, 'learning_rate/qwen_vl_interface': 9.330752405354426e-06, 'learning_rate/action_model': 9.330752405354426e-05, 'epoch': 1.02}) +[2026-06-15 20:51:11,256][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.08851044625043869, 'train/grad_norm_pre_clip': 3.433248996734619, 'timing/data': 0.02496832050383091, 'timing/model': 1.4055698160082102, 'learning_rate/qwen_vl_interface': 9.32919292588271e-06, 'learning_rate/action_model': 9.32919292588271e-05, 'epoch': 1.02}) +[2026-06-15 20:51:17,030][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.1193101555109024, 'train/grad_norm_pre_clip': 4.545589923858643, 'timing/data': 0.02313109114766121, 'timing/model': 1.6553269103169441, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 1.02}) +[2026-06-15 20:51:22,921][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.22547438740730286, 'train/grad_norm_pre_clip': 11.620241165161133, 'timing/data': 0.020742299035191536, 'timing/model': 1.6994022782891989, 'learning_rate/qwen_vl_interface': 9.326068937174627e-06, 'learning_rate/action_model': 9.326068937174626e-05, 'epoch': 1.02}) +[2026-06-15 20:51:28,549][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.27777808904647827, 'train/grad_norm_pre_clip': 8.256966590881348, 'timing/data': 0.020550156943500042, 'timing/model': 1.4483178025111556, 'learning_rate/qwen_vl_interface': 9.324504429222412e-06, 'learning_rate/action_model': 9.324504429222412e-05, 'epoch': 1.03}) +[2026-06-15 20:51:34,353][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.22017474472522736, 'train/grad_norm_pre_clip': 10.555871963500977, 'timing/data': 0.020228765904903412, 'timing/model': 1.4403890697285533, 'learning_rate/qwen_vl_interface': 9.322938246395336e-06, 'learning_rate/action_model': 9.322938246395337e-05, 'epoch': 1.03}) +[2026-06-15 20:51:40,150][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.11260072886943817, 'train/grad_norm_pre_clip': 6.908309459686279, 'timing/data': 0.02149048261344433, 'timing/model': 1.446193597279489, 'learning_rate/qwen_vl_interface': 9.321370389337196e-06, 'learning_rate/action_model': 9.321370389337196e-05, 'epoch': 1.03}) +[2026-06-15 20:51:45,956][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.18172413110733032, 'train/grad_norm_pre_clip': 6.478728294372559, 'timing/data': 0.020746828988194466, 'timing/model': 1.4024226069450378, 'learning_rate/qwen_vl_interface': 9.319800858692478e-06, 'learning_rate/action_model': 9.319800858692476e-05, 'epoch': 1.03}) +[2026-06-15 20:51:51,802][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.10589220374822617, 'train/grad_norm_pre_clip': 5.433326721191406, 'timing/data': 0.02158214244991541, 'timing/model': 1.6879252269864082, 'learning_rate/qwen_vl_interface': 9.318229655106357e-06, 'learning_rate/action_model': 9.318229655106356e-05, 'epoch': 1.03}) +[2026-06-15 20:51:57,454][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.2611609101295471, 'train/grad_norm_pre_clip': 9.239239692687988, 'timing/data': 0.02313970122486353, 'timing/model': 1.4470274457708001, 'learning_rate/qwen_vl_interface': 9.316656779224693e-06, 'learning_rate/action_model': 9.316656779224693e-05, 'epoch': 1.03}) +[2026-06-15 20:52:03,252][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.05712259188294411, 'train/grad_norm_pre_clip': 3.8840627670288086, 'timing/data': 0.02045528683811426, 'timing/model': 1.449844148941338, 'learning_rate/qwen_vl_interface': 9.31508223169404e-06, 'learning_rate/action_model': 9.31508223169404e-05, 'epoch': 1.03}) +[2026-06-15 20:52:09,056][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.10597968101501465, 'train/grad_norm_pre_clip': 7.907728672027588, 'timing/data': 0.020944208838045597, 'timing/model': 1.4552153861150146, 'learning_rate/qwen_vl_interface': 9.313506013161633e-06, 'learning_rate/action_model': 9.313506013161633e-05, 'epoch': 1.03}) +[2026-06-15 20:52:14,865][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.25501635670661926, 'train/grad_norm_pre_clip': 8.168556213378906, 'timing/data': 0.02113244030624628, 'timing/model': 1.4131543338298798, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 1.03}) +[2026-06-15 20:52:20,607][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.05993576720356941, 'train/grad_norm_pre_clip': 5.89082670211792, 'timing/data': 0.02315535955131054, 'timing/model': 1.3989456379786134, 'learning_rate/qwen_vl_interface': 9.310348565683938e-06, 'learning_rate/action_model': 9.310348565683939e-05, 'epoch': 1.03}) +[2026-06-15 20:52:26,359][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.7268898487091064, 'train/grad_norm_pre_clip': 15.16375732421875, 'timing/data': 0.02085351850837469, 'timing/model': 1.6398963443934917, 'learning_rate/qwen_vl_interface': 9.30876733803656e-06, 'learning_rate/action_model': 9.308767338036559e-05, 'epoch': 1.03}) +[2026-06-15 20:52:31,950][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.05362784489989281, 'train/grad_norm_pre_clip': 2.8180317878723145, 'timing/data': 0.02007247507572174, 'timing/model': 1.4453294668346643, 'learning_rate/qwen_vl_interface': 9.30718444198324e-06, 'learning_rate/action_model': 9.307184441983239e-05, 'epoch': 1.03}) +[2026-06-15 20:52:37,650][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.049450457096099854, 'train/grad_norm_pre_clip': 3.7550415992736816, 'timing/data': 0.020579596981406212, 'timing/model': 1.4268963085487485, 'learning_rate/qwen_vl_interface': 9.30559987817465e-06, 'learning_rate/action_model': 9.30559987817465e-05, 'epoch': 1.03}) +[2026-06-15 20:52:43,449][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.13632944226264954, 'train/grad_norm_pre_clip': 5.37961483001709, 'timing/data': 0.020964560098946095, 'timing/model': 1.446481179445982, 'learning_rate/qwen_vl_interface': 9.304013647262143e-06, 'learning_rate/action_model': 9.304013647262143e-05, 'epoch': 1.03}) +[2026-06-15 20:52:49,145][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.11214117705821991, 'train/grad_norm_pre_clip': 3.7561659812927246, 'timing/data': 0.0194632513448596, 'timing/model': 1.4079106273129582, 'learning_rate/qwen_vl_interface': 9.302425749897761e-06, 'learning_rate/action_model': 9.302425749897761e-05, 'epoch': 1.03}) +[2026-06-15 20:52:54,948][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.046484000980854034, 'train/grad_norm_pre_clip': 3.090536117553711, 'timing/data': 0.019726213067770004, 'timing/model': 1.3941570231691003, 'learning_rate/qwen_vl_interface': 9.300836186734225e-06, 'learning_rate/action_model': 9.300836186734226e-05, 'epoch': 1.03}) +[2026-06-15 20:53:00,779][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.14826913177967072, 'train/grad_norm_pre_clip': 5.9567975997924805, 'timing/data': 0.020617308095097542, 'timing/model': 1.6787116341292858, 'learning_rate/qwen_vl_interface': 9.299244958424947e-06, 'learning_rate/action_model': 9.299244958424946e-05, 'epoch': 1.03}) +[2026-06-15 20:53:06,357][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.17987269163131714, 'train/grad_norm_pre_clip': 11.509772300720215, 'timing/data': 0.019948383793234825, 'timing/model': 1.4642799152061343, 'learning_rate/qwen_vl_interface': 9.29765206562402e-06, 'learning_rate/action_model': 9.297652065624018e-05, 'epoch': 1.03}) +[2026-06-15 20:53:12,160][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.035994116216897964, 'train/grad_norm_pre_clip': 1.998542070388794, 'timing/data': 0.019722051918506622, 'timing/model': 1.4509814251214266, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 1.03}) +[2026-06-15 20:53:17,946][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.0951586663722992, 'train/grad_norm_pre_clip': 8.082963943481445, 'timing/data': 0.02004629373550415, 'timing/model': 1.4520822195336223, 'learning_rate/qwen_vl_interface': 9.294461289167013e-06, 'learning_rate/action_model': 9.294461289167014e-05, 'epoch': 1.03}) +[2026-06-15 20:53:23,650][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.2606518268585205, 'train/grad_norm_pre_clip': 8.124990463256836, 'timing/data': 0.019404161721467972, 'timing/model': 1.3954163119196892, 'learning_rate/qwen_vl_interface': 9.292863406822543e-06, 'learning_rate/action_model': 9.292863406822542e-05, 'epoch': 1.03}) +[2026-06-15 20:53:29,401][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.3497546911239624, 'train/grad_norm_pre_clip': 10.996999740600586, 'timing/data': 0.01902181003242731, 'timing/model': 1.706565648317337, 'learning_rate/qwen_vl_interface': 9.291263862609638e-06, 'learning_rate/action_model': 9.291263862609638e-05, 'epoch': 1.03}) +[2026-06-15 20:53:35,050][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.1368516981601715, 'train/grad_norm_pre_clip': 5.6739420890808105, 'timing/data': 0.01933377142995596, 'timing/model': 1.4418074693530798, 'learning_rate/qwen_vl_interface': 9.28966265718581e-06, 'learning_rate/action_model': 9.28966265718581e-05, 'epoch': 1.04}) +[2026-06-15 20:53:40,844][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.22879484295845032, 'train/grad_norm_pre_clip': 7.671782493591309, 'timing/data': 0.018496837466955185, 'timing/model': 1.4394075758755207, 'learning_rate/qwen_vl_interface': 9.288059791209256e-06, 'learning_rate/action_model': 9.288059791209256e-05, 'epoch': 1.04}) +[2026-06-15 20:53:46,548][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.01859881728887558, 'train/grad_norm_pre_clip': 0.875538170337677, 'timing/data': 0.018764857202768326, 'timing/model': 1.4518796084448695, 'learning_rate/qwen_vl_interface': 9.286455265338853e-06, 'learning_rate/action_model': 9.286455265338853e-05, 'epoch': 1.04}) +[2026-06-15 20:53:52,358][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.16854049265384674, 'train/grad_norm_pre_clip': 7.195136547088623, 'timing/data': 0.019884763285517693, 'timing/model': 1.4468939015641809, 'learning_rate/qwen_vl_interface': 9.284849080234161e-06, 'learning_rate/action_model': 9.28484908023416e-05, 'epoch': 1.04}) +[2026-06-15 20:53:58,147][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.21374191343784332, 'train/grad_norm_pre_clip': 8.999837875366211, 'timing/data': 0.01998042408376932, 'timing/model': 1.4347075568512082, 'learning_rate/qwen_vl_interface': 9.283241236555421e-06, 'learning_rate/action_model': 9.283241236555421e-05, 'epoch': 1.04}) +[2026-06-15 20:54:04,017][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.10622058808803558, 'train/grad_norm_pre_clip': 6.936861515045166, 'timing/data': 0.01994407270103693, 'timing/model': 1.7235859166830778, 'learning_rate/qwen_vl_interface': 9.281631734963557e-06, 'learning_rate/action_model': 9.281631734963557e-05, 'epoch': 1.04}) +[2026-06-15 20:54:09,577][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.14601320028305054, 'train/grad_norm_pre_clip': 8.35688304901123, 'timing/data': 0.0280622448772192, 'timing/model': 1.430833545513451, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 1.04}) +[2026-06-15 20:54:15,561][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.0791807547211647, 'train/grad_norm_pre_clip': 5.8434157371521, 'timing/data': 0.02148303110152483, 'timing/model': 1.45942045096308, 'learning_rate/qwen_vl_interface': 9.278407760687562e-06, 'learning_rate/action_model': 9.278407760687562e-05, 'epoch': 1.04}) +[2026-06-15 20:54:21,356][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.12554089725017548, 'train/grad_norm_pre_clip': 5.066257476806641, 'timing/data': 0.020642106421291828, 'timing/model': 1.4489631447941065, 'learning_rate/qwen_vl_interface': 9.276793289328686e-06, 'learning_rate/action_model': 9.276793289328685e-05, 'epoch': 1.04}) +[2026-06-15 20:54:27,153][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.2579500675201416, 'train/grad_norm_pre_clip': 13.841358184814453, 'timing/data': 0.02026390563696623, 'timing/model': 1.4521370511502028, 'learning_rate/qwen_vl_interface': 9.27517716270719e-06, 'learning_rate/action_model': 9.275177162707191e-05, 'epoch': 1.04}) +[2026-06-15 20:54:32,953][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.20698481798171997, 'train/grad_norm_pre_clip': 5.89069128036499, 'timing/data': 0.02138514071702957, 'timing/model': 1.416068541817367, 'learning_rate/qwen_vl_interface': 9.273559381487409e-06, 'learning_rate/action_model': 9.273559381487409e-05, 'epoch': 1.04}) +[2026-06-15 20:54:38,809][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.10163179039955139, 'train/grad_norm_pre_clip': 6.525671005249023, 'timing/data': 0.020781838335096836, 'timing/model': 1.7064378783106804, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 1.04}) +[2026-06-15 20:54:44,346][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.3239251673221588, 'train/grad_norm_pre_clip': 22.032299041748047, 'timing/data': 0.021025818772614002, 'timing/model': 1.4206791436299682, 'learning_rate/qwen_vl_interface': 9.2703188579137e-06, 'learning_rate/action_model': 9.2703188579137e-05, 'epoch': 1.04}) +[2026-06-15 20:54:50,086][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.15239155292510986, 'train/grad_norm_pre_clip': 9.819663047790527, 'timing/data': 0.023314260877668858, 'timing/model': 1.411481224000454, 'learning_rate/qwen_vl_interface': 9.26869611689183e-06, 'learning_rate/action_model': 9.26869611689183e-05, 'epoch': 1.04}) +[2026-06-15 20:54:55,966][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.06304284185171127, 'train/grad_norm_pre_clip': 6.618534088134766, 'timing/data': 0.020238375291228294, 'timing/model': 1.4613457601517439, 'learning_rate/qwen_vl_interface': 9.267071723935784e-06, 'learning_rate/action_model': 9.267071723935784e-05, 'epoch': 1.04}) +[2026-06-15 20:55:01,750][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.2091609537601471, 'train/grad_norm_pre_clip': 9.964225769042969, 'timing/data': 0.02098459843546152, 'timing/model': 1.3936381489038467, 'learning_rate/qwen_vl_interface': 9.26544567971329e-06, 'learning_rate/action_model': 9.26544567971329e-05, 'epoch': 1.04}) +[2026-06-15 20:55:07,353][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.08426277339458466, 'train/grad_norm_pre_clip': 5.382933139801025, 'timing/data': 0.01959839276969433, 'timing/model': 1.4045735644176602, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 1.04}) +[2026-06-15 20:55:13,208][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.09368525445461273, 'train/grad_norm_pre_clip': 4.8193535804748535, 'timing/data': 0.020501095801591873, 'timing/model': 1.7085583498701453, 'learning_rate/qwen_vl_interface': 9.262188640143262e-06, 'learning_rate/action_model': 9.262188640143262e-05, 'epoch': 1.04}) +[2026-06-15 20:55:18,761][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.1411476582288742, 'train/grad_norm_pre_clip': 6.418460845947266, 'timing/data': 0.020599596202373505, 'timing/model': 1.4622872825711966, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 1.04}) +[2026-06-15 20:55:24,559][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.19153869152069092, 'train/grad_norm_pre_clip': 7.357911586761475, 'timing/data': 0.020203813910484314, 'timing/model': 1.4622531104832888, 'learning_rate/qwen_vl_interface': 9.258925003537127e-06, 'learning_rate/action_model': 9.258925003537126e-05, 'epoch': 1.04}) +[2026-06-15 20:55:30,384][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.11298159509897232, 'train/grad_norm_pre_clip': 7.065536975860596, 'timing/data': 0.0211142897605896, 'timing/model': 1.4802456507459283, 'learning_rate/qwen_vl_interface': 9.257290713022041e-06, 'learning_rate/action_model': 9.257290713022041e-05, 'epoch': 1.04}) +[2026-06-15 20:55:36,255][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.07142723351716995, 'train/grad_norm_pre_clip': 5.868888854980469, 'timing/data': 0.021220399998128414, 'timing/model': 1.4483049353584647, 'learning_rate/qwen_vl_interface': 9.255654775261112e-06, 'learning_rate/action_model': 9.255654775261112e-05, 'epoch': 1.05}) +[2026-06-15 20:55:42,070][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.15456388890743256, 'train/grad_norm_pre_clip': 9.158970832824707, 'timing/data': 0.02092788740992546, 'timing/model': 1.401438470929861, 'learning_rate/qwen_vl_interface': 9.254017190926813e-06, 'learning_rate/action_model': 9.254017190926813e-05, 'epoch': 1.05}) +[2026-06-15 20:55:47,916][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.21260498464107513, 'train/grad_norm_pre_clip': 10.121910095214844, 'timing/data': 0.021378381177783012, 'timing/model': 1.7137744361534715, 'learning_rate/qwen_vl_interface': 9.25237796069229e-06, 'learning_rate/action_model': 9.25237796069229e-05, 'epoch': 1.05}) +[2026-06-15 20:55:53,557][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.22192752361297607, 'train/grad_norm_pre_clip': 8.88361930847168, 'timing/data': 0.023759502917528152, 'timing/model': 1.4466855926439166, 'learning_rate/qwen_vl_interface': 9.250737085231371e-06, 'learning_rate/action_model': 9.250737085231372e-05, 'epoch': 1.05}) +[2026-06-15 20:55:59,355][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.13518814742565155, 'train/grad_norm_pre_clip': 5.769341468811035, 'timing/data': 0.020027154125273228, 'timing/model': 1.4477893374860287, 'learning_rate/qwen_vl_interface': 9.24909456521856e-06, 'learning_rate/action_model': 9.249094565218559e-05, 'epoch': 1.05}) +[2026-06-15 20:56:05,177][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.12834522128105164, 'train/grad_norm_pre_clip': 5.1845855712890625, 'timing/data': 0.027830044738948345, 'timing/model': 1.4255920611321926, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 1.05}) +[2026-06-15 20:56:10,958][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.15612487494945526, 'train/grad_norm_pre_clip': 10.031379699707031, 'timing/data': 0.021022168919444084, 'timing/model': 1.4461669260635972, 'learning_rate/qwen_vl_interface': 9.245804594238637e-06, 'learning_rate/action_model': 9.245804594238636e-05, 'epoch': 1.05}) +[2026-06-15 20:56:16,757][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.27481508255004883, 'train/grad_norm_pre_clip': 7.85456657409668, 'timing/data': 0.019875853322446346, 'timing/model': 1.4025574596598744, 'learning_rate/qwen_vl_interface': 9.244157144623909e-06, 'learning_rate/action_model': 9.244157144623909e-05, 'epoch': 1.05}) +[2026-06-15 20:56:22,611][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.15359747409820557, 'train/grad_norm_pre_clip': 6.15133810043335, 'timing/data': 0.022158724255859852, 'timing/model': 1.6985188508406281, 'learning_rate/qwen_vl_interface': 9.242508053162051e-06, 'learning_rate/action_model': 9.242508053162051e-05, 'epoch': 1.05}) +[2026-06-15 20:56:28,256][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.2133045196533203, 'train/grad_norm_pre_clip': 8.756277084350586, 'timing/data': 0.020792958326637745, 'timing/model': 1.4323734510689974, 'learning_rate/qwen_vl_interface': 9.240857320530944e-06, 'learning_rate/action_model': 9.240857320530943e-05, 'epoch': 1.05}) +[2026-06-15 20:56:34,058][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.30571433901786804, 'train/grad_norm_pre_clip': 8.108174324035645, 'timing/data': 0.020644236356019974, 'timing/model': 1.4492158275097609, 'learning_rate/qwen_vl_interface': 9.23920494740914e-06, 'learning_rate/action_model': 9.23920494740914e-05, 'epoch': 1.05}) +[2026-06-15 20:56:39,866][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.1767125278711319, 'train/grad_norm_pre_clip': 5.575771808624268, 'timing/data': 0.021111929789185524, 'timing/model': 1.42841009888798, 'learning_rate/qwen_vl_interface': 9.237550934475868e-06, 'learning_rate/action_model': 9.237550934475867e-05, 'epoch': 1.05}) +[2026-06-15 20:56:45,656][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.08131630718708038, 'train/grad_norm_pre_clip': 4.326195240020752, 'timing/data': 0.021763311699032784, 'timing/model': 1.4364906400442123, 'learning_rate/qwen_vl_interface': 9.235895282411029e-06, 'learning_rate/action_model': 9.235895282411029e-05, 'epoch': 1.05}) +[2026-06-15 20:56:51,458][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.10802990198135376, 'train/grad_norm_pre_clip': 5.3167877197265625, 'timing/data': 0.021224570460617542, 'timing/model': 1.3995631579309702, 'learning_rate/qwen_vl_interface': 9.2342379918952e-06, 'learning_rate/action_model': 9.2342379918952e-05, 'epoch': 1.05}) +[2026-06-15 20:56:57,299][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.40048158168792725, 'train/grad_norm_pre_clip': 7.4682841300964355, 'timing/data': 0.020159784704446793, 'timing/model': 1.697837509214878, 'learning_rate/qwen_vl_interface': 9.232579063609631e-06, 'learning_rate/action_model': 9.232579063609631e-05, 'epoch': 1.05}) +[2026-06-15 21:07:44,659][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.1846463978290558, 'train/grad_norm_pre_clip': 5.21783971786499, 'eval/loss': 0.07988160103559494, 'eval/flap_precision': 0.7296416938110749, 'eval/flap_recall': 0.8145454545454546, 'eval/flap_f1': 0.7697594501718213, 'eval/flap_rate_model': 0.08527777777777777, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7697594501718213, 'timing/data': 0.020349375903606415, 'timing/model': 1.4378951732069254, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 1.05}) +[2026-06-15 21:07:44,660][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2300_state +[2026-06-15 21:07:44,689][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 21:07:52,455][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2300_state/model.safetensors +[2026-06-15 21:08:03,758][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2300_state/optimizer.bin +[2026-06-15 21:08:03,759][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2300_state/sampler.bin +[2026-06-15 21:08:03,759][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2300_state/sampler_1.bin +[2026-06-15 21:08:03,761][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2300_state/random_states_0.pkl +[2026-06-15 21:08:03,912][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2300 +[2026-06-15 21:08:03,913][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 21:08:03,928][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 21:08:09,777][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.08841606229543686, 'train/grad_norm_pre_clip': 3.9210264682769775, 'timing/data': 0.022409765049815178, 'timing/model': 1.7036122484132648, 'learning_rate/qwen_vl_interface': 9.229256296457635e-06, 'learning_rate/action_model': 9.229256296457634e-05, 'epoch': 1.05}) +[2026-06-15 21:08:15,235][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.06588171422481537, 'train/grad_norm_pre_clip': 2.9646804332733154, 'timing/data': 0.022246094420552254, 'timing/model': 1.4296029340475798, 'learning_rate/qwen_vl_interface': 9.22759245895707e-06, 'learning_rate/action_model': 9.227592458957071e-05, 'epoch': 1.05}) +[2026-06-15 21:08:20,967][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.2744885981082916, 'train/grad_norm_pre_clip': 6.427066326141357, 'timing/data': 0.020716906525194645, 'timing/model': 1.3939294209703803, 'learning_rate/qwen_vl_interface': 9.225926986418494e-06, 'learning_rate/action_model': 9.225926986418494e-05, 'epoch': 1.05}) +[2026-06-15 21:08:26,745][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.058672502636909485, 'train/grad_norm_pre_clip': 3.257606267929077, 'timing/data': 0.021978353150188923, 'timing/model': 1.4636580776423216, 'learning_rate/qwen_vl_interface': 9.224259879526519e-06, 'learning_rate/action_model': 9.224259879526519e-05, 'epoch': 1.05}) +[2026-06-15 21:08:32,435][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.17656052112579346, 'train/grad_norm_pre_clip': 5.805398464202881, 'timing/data': 0.021665140986442566, 'timing/model': 1.3825785014778376, 'learning_rate/qwen_vl_interface': 9.222591138966427e-06, 'learning_rate/action_model': 9.222591138966426e-05, 'epoch': 1.05}) +[2026-06-15 21:08:38,141][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.294468492269516, 'train/grad_norm_pre_clip': 12.104681968688965, 'timing/data': 0.02234821394085884, 'timing/model': 1.4484962802380323, 'learning_rate/qwen_vl_interface': 9.220920765424175e-06, 'learning_rate/action_model': 9.220920765424174e-05, 'epoch': 1.05}) +[2026-06-15 21:08:43,896][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.1287107765674591, 'train/grad_norm_pre_clip': 4.061680316925049, 'timing/data': 0.021476910449564457, 'timing/model': 1.6978659564629197, 'learning_rate/qwen_vl_interface': 9.219248759586392e-06, 'learning_rate/action_model': 9.219248759586391e-05, 'epoch': 1.06}) +[2026-06-15 21:08:49,560][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.07098463922739029, 'train/grad_norm_pre_clip': 3.262406349182129, 'timing/data': 0.02779159229248762, 'timing/model': 1.4323148569092155, 'learning_rate/qwen_vl_interface': 9.217575122140375e-06, 'learning_rate/action_model': 9.217575122140375e-05, 'epoch': 1.06}) +[2026-06-15 21:08:55,367][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.12618295848369598, 'train/grad_norm_pre_clip': 5.803718566894531, 'timing/data': 0.028103863820433617, 'timing/model': 1.4391818717122078, 'learning_rate/qwen_vl_interface': 9.215899853774094e-06, 'learning_rate/action_model': 9.215899853774093e-05, 'epoch': 1.06}) +[2026-06-15 21:09:01,476][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.08478036522865295, 'train/grad_norm_pre_clip': 3.522265911102295, 'timing/data': 0.029739921912550926, 'timing/model': 1.4445019392296672, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 1.06}) +[2026-06-15 21:09:07,661][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.06040114164352417, 'train/grad_norm_pre_clip': 3.046992301940918, 'timing/data': 0.02834707498550415, 'timing/model': 1.4087674152106047, 'learning_rate/qwen_vl_interface': 9.212544427035967e-06, 'learning_rate/action_model': 9.212544427035967e-05, 'epoch': 1.06}) +[2026-06-15 21:09:13,675][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.07644330710172653, 'train/grad_norm_pre_clip': 3.951573133468628, 'timing/data': 0.021440300159156322, 'timing/model': 1.8696829043328762, 'learning_rate/qwen_vl_interface': 9.21086427004341e-06, 'learning_rate/action_model': 9.210864270043411e-05, 'epoch': 1.06}) +[2026-06-15 21:09:19,643][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.2151907980442047, 'train/grad_norm_pre_clip': 6.4875874519348145, 'timing/data': 0.02881235722452402, 'timing/model': 1.8249784959480166, 'learning_rate/qwen_vl_interface': 9.209182484889168e-06, 'learning_rate/action_model': 9.209182484889167e-05, 'epoch': 1.06}) +[2026-06-15 21:09:25,269][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.25850072503089905, 'train/grad_norm_pre_clip': 11.299978256225586, 'timing/data': 0.02839531470090151, 'timing/model': 1.4346829373389482, 'learning_rate/qwen_vl_interface': 9.207499072264557e-06, 'learning_rate/action_model': 9.207499072264558e-05, 'epoch': 1.06}) +[2026-06-15 21:09:31,261][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.13107746839523315, 'train/grad_norm_pre_clip': 5.567906379699707, 'timing/data': 0.02816649340093136, 'timing/model': 1.440709457732737, 'learning_rate/qwen_vl_interface': 9.205814032861566e-06, 'learning_rate/action_model': 9.205814032861566e-05, 'epoch': 1.06}) +[2026-06-15 21:09:37,397][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.13117510080337524, 'train/grad_norm_pre_clip': 6.784085273742676, 'timing/data': 0.02903628908097744, 'timing/model': 1.4699648264795542, 'learning_rate/qwen_vl_interface': 9.204127367372849e-06, 'learning_rate/action_model': 9.204127367372849e-05, 'epoch': 1.06}) +[2026-06-15 21:09:43,217][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.08841061592102051, 'train/grad_norm_pre_clip': 5.7602691650390625, 'timing/data': 0.03177013248205185, 'timing/model': 1.4149988256394863, 'learning_rate/qwen_vl_interface': 9.202439076491732e-06, 'learning_rate/action_model': 9.202439076491732e-05, 'epoch': 1.06}) +[2026-06-15 21:09:49,111][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.11594901978969574, 'train/grad_norm_pre_clip': 6.167240619659424, 'timing/data': 0.031968504190444946, 'timing/model': 1.4054629458114505, 'learning_rate/qwen_vl_interface': 9.200749160912208e-06, 'learning_rate/action_model': 9.200749160912208e-05, 'epoch': 1.06}) +[2026-06-15 21:09:54,910][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.04964994639158249, 'train/grad_norm_pre_clip': 8.186145782470703, 'timing/data': 0.02210855297744274, 'timing/model': 1.687051367945969, 'learning_rate/qwen_vl_interface': 9.199057621328937e-06, 'learning_rate/action_model': 9.199057621328935e-05, 'epoch': 1.06}) +[2026-06-15 21:10:00,553][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.5713942646980286, 'train/grad_norm_pre_clip': 13.101511001586914, 'timing/data': 0.021519719623029232, 'timing/model': 1.441801382228732, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 1.06}) +[2026-06-15 21:10:06,347][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.33019864559173584, 'train/grad_norm_pre_clip': 15.616472244262695, 'timing/data': 0.021838721819221973, 'timing/model': 1.4300918616354465, 'learning_rate/qwen_vl_interface': 9.19566967293313e-06, 'learning_rate/action_model': 9.19566967293313e-05, 'epoch': 1.06}) +[2026-06-15 21:10:12,259][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.06599210202693939, 'train/grad_norm_pre_clip': 5.200882911682129, 'timing/data': 0.021296349354088306, 'timing/model': 1.4526139879599214, 'learning_rate/qwen_vl_interface': 9.193973265513253e-06, 'learning_rate/action_model': 9.193973265513253e-05, 'epoch': 1.06}) +[2026-06-15 21:10:18,165][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.5120843648910522, 'train/grad_norm_pre_clip': 20.02466583251953, 'timing/data': 0.02174135111272335, 'timing/model': 1.4715240933001041, 'learning_rate/qwen_vl_interface': 9.192275236874946e-06, 'learning_rate/action_model': 9.192275236874946e-05, 'epoch': 1.06}) +[2026-06-15 21:10:24,030][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.2840862572193146, 'train/grad_norm_pre_clip': 27.031320571899414, 'timing/data': 0.02096180710941553, 'timing/model': 1.717632994055748, 'learning_rate/qwen_vl_interface': 9.190575587716199e-06, 'learning_rate/action_model': 9.190575587716199e-05, 'epoch': 1.06}) +[2026-06-15 21:10:29,653][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.15669690072536469, 'train/grad_norm_pre_clip': 3.9254469871520996, 'timing/data': 0.02108235750347376, 'timing/model': 1.447276039980352, 'learning_rate/qwen_vl_interface': 9.188874318735677e-06, 'learning_rate/action_model': 9.188874318735676e-05, 'epoch': 1.06}) +[2026-06-15 21:10:35,472][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.16775164008140564, 'train/grad_norm_pre_clip': 10.003031730651855, 'timing/data': 0.02561729121953249, 'timing/model': 1.4726510290056467, 'learning_rate/qwen_vl_interface': 9.187171430632707e-06, 'learning_rate/action_model': 9.187171430632707e-05, 'epoch': 1.06}) +[2026-06-15 21:10:41,266][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.23455385863780975, 'train/grad_norm_pre_clip': 19.09563636779785, 'timing/data': 0.021423219703137875, 'timing/model': 1.461266309954226, 'learning_rate/qwen_vl_interface': 9.185466924107283e-06, 'learning_rate/action_model': 9.185466924107282e-05, 'epoch': 1.06}) +[2026-06-15 21:10:47,063][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.08098048716783524, 'train/grad_norm_pre_clip': 5.526609420776367, 'timing/data': 0.025457399897277355, 'timing/model': 1.4464556351304054, 'learning_rate/qwen_vl_interface': 9.18376079986006e-06, 'learning_rate/action_model': 9.183760799860061e-05, 'epoch': 1.06}) +[2026-06-15 21:10:52,845][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.5231868624687195, 'train/grad_norm_pre_clip': 19.5560359954834, 'timing/data': 0.021009238436818123, 'timing/model': 1.411308224312961, 'learning_rate/qwen_vl_interface': 9.182053058592365e-06, 'learning_rate/action_model': 9.182053058592365e-05, 'epoch': 1.07}) +[2026-06-15 21:10:58,630][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.1135563850402832, 'train/grad_norm_pre_clip': 4.476600646972656, 'timing/data': 0.021368449553847313, 'timing/model': 1.7357892543077469, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 1.07}) +[2026-06-15 21:11:04,144][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.31440412998199463, 'train/grad_norm_pre_clip': 7.467298984527588, 'timing/data': 0.021377138793468475, 'timing/model': 1.4483517734333873, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 1.07}) +[2026-06-15 21:11:09,943][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.18359370529651642, 'train/grad_norm_pre_clip': 6.6225266456604, 'timing/data': 0.021000628359615803, 'timing/model': 1.447551448829472, 'learning_rate/qwen_vl_interface': 9.176920139689638e-06, 'learning_rate/action_model': 9.176920139689637e-05, 'epoch': 1.07}) +[2026-06-15 21:11:15,646][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.1604984998703003, 'train/grad_norm_pre_clip': 6.143955230712891, 'timing/data': 0.021905552595853806, 'timing/model': 1.4233492854982615, 'learning_rate/qwen_vl_interface': 9.175205937366568e-06, 'learning_rate/action_model': 9.175205937366569e-05, 'epoch': 1.07}) +[2026-06-15 21:11:21,452][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.21031714975833893, 'train/grad_norm_pre_clip': 8.885986328125, 'timing/data': 0.020779306069016457, 'timing/model': 1.452926866710186, 'learning_rate/qwen_vl_interface': 9.173490121539608e-06, 'learning_rate/action_model': 9.173490121539608e-05, 'epoch': 1.07}) +[2026-06-15 21:11:27,152][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.0881006196141243, 'train/grad_norm_pre_clip': 5.57718563079834, 'timing/data': 0.021423599682748318, 'timing/model': 1.4092681519687176, 'learning_rate/qwen_vl_interface': 9.171772692914062e-06, 'learning_rate/action_model': 9.171772692914061e-05, 'epoch': 1.07}) +[2026-06-15 21:11:33,040][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.16219796240329742, 'train/grad_norm_pre_clip': 5.437739849090576, 'timing/data': 0.020335674285888672, 'timing/model': 1.7468949295580387, 'learning_rate/qwen_vl_interface': 9.1700536521959e-06, 'learning_rate/action_model': 9.170053652195899e-05, 'epoch': 1.07}) +[2026-06-15 21:11:38,660][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.34854304790496826, 'train/grad_norm_pre_clip': 8.329407691955566, 'timing/data': 0.019933242350816727, 'timing/model': 1.4509193757548928, 'learning_rate/qwen_vl_interface': 9.168333000091754e-06, 'learning_rate/action_model': 9.168333000091754e-05, 'epoch': 1.07}) +[2026-06-15 21:11:44,466][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.23490537703037262, 'train/grad_norm_pre_clip': 4.9588446617126465, 'timing/data': 0.020782525651156902, 'timing/model': 1.441479287110269, 'learning_rate/qwen_vl_interface': 9.16661073730892e-06, 'learning_rate/action_model': 9.166610737308919e-05, 'epoch': 1.07}) +[2026-06-15 21:11:50,247][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.0973513126373291, 'train/grad_norm_pre_clip': 3.5071160793304443, 'timing/data': 0.020089722238481045, 'timing/model': 1.4337365068495274, 'learning_rate/qwen_vl_interface': 9.164886864555355e-06, 'learning_rate/action_model': 9.164886864555355e-05, 'epoch': 1.07}) +[2026-06-15 21:11:56,054][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.10686329752206802, 'train/grad_norm_pre_clip': 2.7737574577331543, 'timing/data': 0.019933140836656094, 'timing/model': 1.4186307899653912, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 1.07}) +[2026-06-15 21:12:01,774][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.19597163796424866, 'train/grad_norm_pre_clip': 10.770101547241211, 'timing/data': 0.019593579694628716, 'timing/model': 1.4012388810515404, 'learning_rate/qwen_vl_interface': 9.16143429197117e-06, 'learning_rate/action_model': 9.16143429197117e-05, 'epoch': 1.07}) +[2026-06-15 21:12:07,616][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.08316387981176376, 'train/grad_norm_pre_clip': 2.1888303756713867, 'timing/data': 0.021071058697998524, 'timing/model': 1.7174465786665678, 'learning_rate/qwen_vl_interface': 9.159705593559773e-06, 'learning_rate/action_model': 9.159705593559773e-05, 'epoch': 1.07}) +[2026-06-15 21:12:13,263][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.1847045123577118, 'train/grad_norm_pre_clip': 4.829844951629639, 'timing/data': 0.025047147646546364, 'timing/model': 1.4619112415239215, 'learning_rate/qwen_vl_interface': 9.157975288016089e-06, 'learning_rate/action_model': 9.157975288016088e-05, 'epoch': 1.07}) +[2026-06-15 21:12:19,059][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.18169762194156647, 'train/grad_norm_pre_clip': 4.627745151519775, 'timing/data': 0.020248303189873695, 'timing/model': 1.4628281751647592, 'learning_rate/qwen_vl_interface': 9.156243376051381e-06, 'learning_rate/action_model': 9.156243376051381e-05, 'epoch': 1.07}) +[2026-06-15 21:12:24,847][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.08184066414833069, 'train/grad_norm_pre_clip': 3.2869503498077393, 'timing/data': 0.021029447205364704, 'timing/model': 1.449746218509972, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 1.07}) +[2026-06-15 21:12:30,553][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.1552874594926834, 'train/grad_norm_pre_clip': 5.007848262786865, 'timing/data': 0.02032544370740652, 'timing/model': 1.4544673021882772, 'learning_rate/qwen_vl_interface': 9.152774735707248e-06, 'learning_rate/action_model': 9.152774735707247e-05, 'epoch': 1.07}) +[2026-06-15 21:12:37,048][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.09716297686100006, 'train/grad_norm_pre_clip': 5.073602199554443, 'timing/data': 0.022357364185154438, 'timing/model': 1.400656696408987, 'learning_rate/qwen_vl_interface': 9.15103800875365e-06, 'learning_rate/action_model': 9.15103800875365e-05, 'epoch': 1.07}) +[2026-06-15 21:12:42,879][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.11545189470052719, 'train/grad_norm_pre_clip': 4.638998985290527, 'timing/data': 0.019676500000059605, 'timing/model': 1.6952637741342187, 'learning_rate/qwen_vl_interface': 9.149299678230681e-06, 'learning_rate/action_model': 9.149299678230681e-05, 'epoch': 1.07}) +[2026-06-15 21:12:48,352][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.14772452414035797, 'train/grad_norm_pre_clip': 6.723904132843018, 'timing/data': 0.019817381165921688, 'timing/model': 1.4287932701408863, 'learning_rate/qwen_vl_interface': 9.147559744852905e-06, 'learning_rate/action_model': 9.147559744852905e-05, 'epoch': 1.07}) +[2026-06-15 21:12:54,060][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.029419824481010437, 'train/grad_norm_pre_clip': 2.5398752689361572, 'timing/data': 0.02374651189893484, 'timing/model': 1.4010924184694886, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 1.07}) +[2026-06-15 21:12:59,845][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.4301210343837738, 'train/grad_norm_pre_clip': 16.740705490112305, 'timing/data': 0.02082548663020134, 'timing/model': 1.4455482857301831, 'learning_rate/qwen_vl_interface': 9.14407507239447e-06, 'learning_rate/action_model': 9.14407507239447e-05, 'epoch': 1.08}) +[2026-06-15 21:13:05,650][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.2117820382118225, 'train/grad_norm_pre_clip': 11.31766128540039, 'timing/data': 0.020172403194010258, 'timing/model': 1.45095310267061, 'learning_rate/qwen_vl_interface': 9.142330334746227e-06, 'learning_rate/action_model': 9.142330334746227e-05, 'epoch': 1.08}) +[2026-06-15 21:13:11,460][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.3680017292499542, 'train/grad_norm_pre_clip': 25.083826065063477, 'timing/data': 0.020973877981305122, 'timing/model': 1.394006211310625, 'learning_rate/qwen_vl_interface': 9.14058399710801e-06, 'learning_rate/action_model': 9.14058399710801e-05, 'epoch': 1.08}) +[2026-06-15 21:13:17,222][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.26409685611724854, 'train/grad_norm_pre_clip': 11.49781322479248, 'timing/data': 0.021831861697137356, 'timing/model': 1.6964456085115671, 'learning_rate/qwen_vl_interface': 9.138836060197673e-06, 'learning_rate/action_model': 9.138836060197672e-05, 'epoch': 1.08}) +[2026-06-15 21:13:22,759][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.13012610375881195, 'train/grad_norm_pre_clip': 4.320757865905762, 'timing/data': 0.020281223580241203, 'timing/model': 1.3960493011400104, 'learning_rate/qwen_vl_interface': 9.137086524733725e-06, 'learning_rate/action_model': 9.137086524733724e-05, 'epoch': 1.08}) +[2026-06-15 21:13:28,549][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.3493751883506775, 'train/grad_norm_pre_clip': 11.738396644592285, 'timing/data': 0.020845786668360233, 'timing/model': 1.4376283939927816, 'learning_rate/qwen_vl_interface': 9.135335391435336e-06, 'learning_rate/action_model': 9.135335391435335e-05, 'epoch': 1.08}) +[2026-06-15 21:13:34,347][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.2421843558549881, 'train/grad_norm_pre_clip': 9.93558120727539, 'timing/data': 0.022138383239507675, 'timing/model': 1.444753060117364, 'learning_rate/qwen_vl_interface': 9.13358266102233e-06, 'learning_rate/action_model': 9.133582661022328e-05, 'epoch': 1.08}) +[2026-06-15 21:13:40,147][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.2853318154811859, 'train/grad_norm_pre_clip': 10.607221603393555, 'timing/data': 0.021219697780907154, 'timing/model': 1.4484162786975503, 'learning_rate/qwen_vl_interface': 9.131828334215185e-06, 'learning_rate/action_model': 9.131828334215185e-05, 'epoch': 1.08}) +[2026-06-15 21:13:45,962][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.04898994415998459, 'train/grad_norm_pre_clip': 3.0532209873199463, 'timing/data': 0.02129287924617529, 'timing/model': 1.3952147476375103, 'learning_rate/qwen_vl_interface': 9.130072411735043e-06, 'learning_rate/action_model': 9.130072411735043e-05, 'epoch': 1.08}) +[2026-06-15 21:13:51,823][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.32273226976394653, 'train/grad_norm_pre_clip': 9.937529563903809, 'timing/data': 0.021709331311285496, 'timing/model': 1.7223394205793738, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 1.08}) +[2026-06-15 21:13:57,382][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.21229638159275055, 'train/grad_norm_pre_clip': 11.208684921264648, 'timing/data': 0.023593610152602196, 'timing/model': 1.427129759453237, 'learning_rate/qwen_vl_interface': 9.126555782643592e-06, 'learning_rate/action_model': 9.126555782643591e-05, 'epoch': 1.08}) +[2026-06-15 21:14:03,249][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.09622764587402344, 'train/grad_norm_pre_clip': 4.645873546600342, 'timing/data': 0.019886950962245464, 'timing/model': 1.4443815676495433, 'learning_rate/qwen_vl_interface': 9.124795077477835e-06, 'learning_rate/action_model': 9.124795077477834e-05, 'epoch': 1.08}) +[2026-06-15 21:14:09,058][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.04981393739581108, 'train/grad_norm_pre_clip': 3.12136173248291, 'timing/data': 0.02029575314372778, 'timing/model': 1.4560811864212155, 'learning_rate/qwen_vl_interface': 9.123032779530186e-06, 'learning_rate/action_model': 9.123032779530187e-05, 'epoch': 1.08}) +[2026-06-15 21:14:14,860][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.19197219610214233, 'train/grad_norm_pre_clip': 10.947880744934082, 'timing/data': 0.021249509416520596, 'timing/model': 1.407914150506258, 'learning_rate/qwen_vl_interface': 9.12126888952506e-06, 'learning_rate/action_model': 9.12126888952506e-05, 'epoch': 1.08}) +[2026-06-15 21:14:20,654][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.04069627821445465, 'train/grad_norm_pre_clip': 2.6334168910980225, 'timing/data': 0.020797827281057835, 'timing/model': 1.395339515991509, 'learning_rate/qwen_vl_interface': 9.119503408187521e-06, 'learning_rate/action_model': 9.119503408187522e-05, 'epoch': 1.08}) +[2026-06-15 21:14:26,507][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.06711997836828232, 'train/grad_norm_pre_clip': 3.3642330169677734, 'timing/data': 0.020673125982284546, 'timing/model': 1.7110295807942748, 'learning_rate/qwen_vl_interface': 9.117736336243299e-06, 'learning_rate/action_model': 9.117736336243298e-05, 'epoch': 1.08}) +[2026-06-15 21:14:32,266][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.08887772262096405, 'train/grad_norm_pre_clip': 4.668703079223633, 'timing/data': 0.020583435893058777, 'timing/model': 1.6735598994418979, 'learning_rate/qwen_vl_interface': 9.115967674418765e-06, 'learning_rate/action_model': 9.115967674418765e-05, 'epoch': 1.08}) +[2026-06-15 21:14:37,845][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.13681140542030334, 'train/grad_norm_pre_clip': 6.206879615783691, 'timing/data': 0.020062653347849846, 'timing/model': 1.453132231719792, 'learning_rate/qwen_vl_interface': 9.114197423440949e-06, 'learning_rate/action_model': 9.11419742344095e-05, 'epoch': 1.08}) +[2026-06-15 21:14:43,651][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.11773695051670074, 'train/grad_norm_pre_clip': 7.345668792724609, 'timing/data': 0.02195363212376833, 'timing/model': 1.4357685530558228, 'learning_rate/qwen_vl_interface': 9.112425584037536e-06, 'learning_rate/action_model': 9.112425584037536e-05, 'epoch': 1.08}) +[2026-06-15 21:14:49,450][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.09988640993833542, 'train/grad_norm_pre_clip': 5.0337700843811035, 'timing/data': 0.020218202844262123, 'timing/model': 1.4507469488307834, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 1.08}) +[2026-06-15 21:14:55,250][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.18959879875183105, 'train/grad_norm_pre_clip': 8.940674781799316, 'timing/data': 0.021126367151737213, 'timing/model': 1.4535989128053188, 'learning_rate/qwen_vl_interface': 9.108877142867918e-06, 'learning_rate/action_model': 9.108877142867918e-05, 'epoch': 1.08}) +[2026-06-15 21:15:01,129][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.15821434557437897, 'train/grad_norm_pre_clip': 9.65712833404541, 'timing/data': 0.026695527136325836, 'timing/model': 1.7260396871715784, 'learning_rate/qwen_vl_interface': 9.107100542560342e-06, 'learning_rate/action_model': 9.107100542560342e-05, 'epoch': 1.08}) +[2026-06-15 21:15:07,052][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.10842525959014893, 'train/grad_norm_pre_clip': 7.164186477661133, 'timing/data': 0.024158542975783348, 'timing/model': 1.7419456578791142, 'learning_rate/qwen_vl_interface': 9.105322356744429e-06, 'learning_rate/action_model': 9.105322356744429e-05, 'epoch': 1.09}) +[2026-06-15 21:15:12,659][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.31060153245925903, 'train/grad_norm_pre_clip': 14.92318058013916, 'timing/data': 0.021027146838605404, 'timing/model': 1.4566856604069471, 'learning_rate/qwen_vl_interface': 9.103542586151121e-06, 'learning_rate/action_model': 9.10354258615112e-05, 'epoch': 1.09}) +[2026-06-15 21:15:18,469][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.13846918940544128, 'train/grad_norm_pre_clip': 7.453258037567139, 'timing/data': 0.023486429825425148, 'timing/model': 1.456794260069728, 'learning_rate/qwen_vl_interface': 9.101761231512017e-06, 'learning_rate/action_model': 9.101761231512017e-05, 'epoch': 1.09}) +[2026-06-15 21:15:24,254][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.2924688458442688, 'train/grad_norm_pre_clip': 11.017172813415527, 'timing/data': 0.022226974368095398, 'timing/model': 1.4345849053934216, 'learning_rate/qwen_vl_interface': 9.099978293559365e-06, 'learning_rate/action_model': 9.099978293559366e-05, 'epoch': 1.09}) +[2026-06-15 21:15:30,159][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.08123017847537994, 'train/grad_norm_pre_clip': 5.930623531341553, 'timing/data': 0.02186380047351122, 'timing/model': 1.4531914414837956, 'learning_rate/qwen_vl_interface': 9.098193773026065e-06, 'learning_rate/action_model': 9.098193773026064e-05, 'epoch': 1.09}) +[2026-06-15 21:15:36,034][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.15091052651405334, 'train/grad_norm_pre_clip': 8.751296997070312, 'timing/data': 0.021932912059128284, 'timing/model': 1.7294543227180839, 'learning_rate/qwen_vl_interface': 9.096407670645663e-06, 'learning_rate/action_model': 9.096407670645662e-05, 'epoch': 1.09}) +[2026-06-15 21:15:41,657][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.15764857828617096, 'train/grad_norm_pre_clip': 13.409368515014648, 'timing/data': 0.02568173035979271, 'timing/model': 1.436774387024343, 'learning_rate/qwen_vl_interface': 9.094619987152357e-06, 'learning_rate/action_model': 9.094619987152358e-05, 'epoch': 1.09}) +[2026-06-15 21:15:47,458][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.057861730456352234, 'train/grad_norm_pre_clip': 4.903287410736084, 'timing/data': 0.021563480608165264, 'timing/model': 1.4493886809796095, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 1.09}) +[2026-06-15 21:15:53,258][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.1895834058523178, 'train/grad_norm_pre_clip': 8.706450462341309, 'timing/data': 0.021871152333915234, 'timing/model': 1.4441825039684772, 'learning_rate/qwen_vl_interface': 9.09103987976709e-06, 'learning_rate/action_model': 9.09103987976709e-05, 'epoch': 1.09}) +[2026-06-15 21:15:59,135][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.2271660417318344, 'train/grad_norm_pre_clip': 9.276481628417969, 'timing/data': 0.02173485141247511, 'timing/model': 1.4547191876918077, 'learning_rate/qwen_vl_interface': 9.089247457346774e-06, 'learning_rate/action_model': 9.089247457346773e-05, 'epoch': 1.09}) +[2026-06-15 21:16:04,982][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.3457546532154083, 'train/grad_norm_pre_clip': 12.000571250915527, 'timing/data': 0.020630954764783382, 'timing/model': 1.4637510646134615, 'learning_rate/qwen_vl_interface': 9.087453456756849e-06, 'learning_rate/action_model': 9.087453456756849e-05, 'epoch': 1.09}) +[2026-06-15 21:16:10,891][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.26098158955574036, 'train/grad_norm_pre_clip': 8.220780372619629, 'timing/data': 0.02066249493509531, 'timing/model': 1.6712066447362304, 'learning_rate/qwen_vl_interface': 9.085657878734759e-06, 'learning_rate/action_model': 9.085657878734758e-05, 'epoch': 1.09}) +[2026-06-15 21:16:16,555][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.045512184500694275, 'train/grad_norm_pre_clip': 3.3354990482330322, 'timing/data': 0.02365752123296261, 'timing/model': 1.4428106164559722, 'learning_rate/qwen_vl_interface': 9.083860724018603e-06, 'learning_rate/action_model': 9.083860724018602e-05, 'epoch': 1.09}) +[2026-06-15 21:16:22,353][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.14480701088905334, 'train/grad_norm_pre_clip': 4.590724468231201, 'timing/data': 0.02092789765447378, 'timing/model': 1.4464572947472334, 'learning_rate/qwen_vl_interface': 9.08206199334712e-06, 'learning_rate/action_model': 9.082061993347119e-05, 'epoch': 1.09}) +[2026-06-15 21:16:28,160][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.07987171411514282, 'train/grad_norm_pre_clip': 3.741548538208008, 'timing/data': 0.02236311323940754, 'timing/model': 1.4450368471443653, 'learning_rate/qwen_vl_interface': 9.0802616874597e-06, 'learning_rate/action_model': 9.0802616874597e-05, 'epoch': 1.09}) +[2026-06-15 21:16:34,078][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.05423079431056976, 'train/grad_norm_pre_clip': 2.6260507106781006, 'timing/data': 0.028771966695785522, 'timing/model': 1.4488685186952353, 'learning_rate/qwen_vl_interface': 9.078459807096385e-06, 'learning_rate/action_model': 9.078459807096385e-05, 'epoch': 1.09}) +[2026-06-15 21:16:40,055][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.18559348583221436, 'train/grad_norm_pre_clip': 7.722210884094238, 'timing/data': 0.021977892145514488, 'timing/model': 1.4605409670621157, 'learning_rate/qwen_vl_interface': 9.076656352997854e-06, 'learning_rate/action_model': 9.076656352997853e-05, 'epoch': 1.09}) +[2026-06-15 21:16:45,892][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.21344313025474548, 'train/grad_norm_pre_clip': 7.7258620262146, 'timing/data': 0.022306744009256363, 'timing/model': 1.675350065343082, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 1.09}) +[2026-06-15 21:16:51,564][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.12575043737888336, 'train/grad_norm_pre_clip': 5.193420886993408, 'timing/data': 0.020976128056645393, 'timing/model': 1.4365619141608477, 'learning_rate/qwen_vl_interface': 9.073044726561128e-06, 'learning_rate/action_model': 9.073044726561129e-05, 'epoch': 1.09}) +[2026-06-15 21:16:57,460][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.08784472197294235, 'train/grad_norm_pre_clip': 4.997777462005615, 'timing/data': 0.02104424685239792, 'timing/model': 1.43525021802634, 'learning_rate/qwen_vl_interface': 9.071236555707536e-06, 'learning_rate/action_model': 9.071236555707537e-05, 'epoch': 1.09}) +[2026-06-15 21:17:03,346][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.08217776566743851, 'train/grad_norm_pre_clip': 6.548853874206543, 'timing/data': 0.02131290826946497, 'timing/model': 1.4328259453177452, 'learning_rate/qwen_vl_interface': 9.069426814087937e-06, 'learning_rate/action_model': 9.069426814087937e-05, 'epoch': 1.09}) +[2026-06-15 21:17:09,237][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.07208623737096786, 'train/grad_norm_pre_clip': 5.678475379943848, 'timing/data': 0.0212210975587368, 'timing/model': 1.454249464906752, 'learning_rate/qwen_vl_interface': 9.067615502446247e-06, 'learning_rate/action_model': 9.067615502446246e-05, 'epoch': 1.09}) +[2026-06-15 21:17:15,105][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.2955211102962494, 'train/grad_norm_pre_clip': 12.390571594238281, 'timing/data': 0.021472840569913387, 'timing/model': 1.69436761084944, 'learning_rate/qwen_vl_interface': 9.06580262152703e-06, 'learning_rate/action_model': 9.065802621527028e-05, 'epoch': 1.1}) +[2026-06-15 21:17:20,749][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.43609732389450073, 'train/grad_norm_pre_clip': 17.162370681762695, 'timing/data': 0.0217189509421587, 'timing/model': 1.4321057004854083, 'learning_rate/qwen_vl_interface': 9.063988172075486e-06, 'learning_rate/action_model': 9.063988172075486e-05, 'epoch': 1.1}) +[2026-06-15 21:17:26,547][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.18795649707317352, 'train/grad_norm_pre_clip': 11.099127769470215, 'timing/data': 0.02104809693992138, 'timing/model': 1.4315881486982107, 'learning_rate/qwen_vl_interface': 9.062172154837475e-06, 'learning_rate/action_model': 9.062172154837475e-05, 'epoch': 1.1}) +[2026-06-15 21:17:32,251][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.13633570075035095, 'train/grad_norm_pre_clip': 8.25616455078125, 'timing/data': 0.020615854300558567, 'timing/model': 1.4391342476010323, 'learning_rate/qwen_vl_interface': 9.060354570559487e-06, 'learning_rate/action_model': 9.060354570559487e-05, 'epoch': 1.1}) +[2026-06-15 21:17:38,047][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.15713690221309662, 'train/grad_norm_pre_clip': 7.303323745727539, 'timing/data': 0.020584874786436558, 'timing/model': 1.440851265564561, 'learning_rate/qwen_vl_interface': 9.058535419988667e-06, 'learning_rate/action_model': 9.058535419988667e-05, 'epoch': 1.1}) +[2026-06-15 21:28:26,578][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.08529075980186462, 'train/grad_norm_pre_clip': 4.362356662750244, 'eval/loss': 0.08927145600318909, 'eval/flap_precision': 0.7213622291021672, 'eval/flap_recall': 0.8472727272727273, 'eval/flap_f1': 0.7792642140468227, 'eval/flap_rate_model': 0.08972222222222222, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7792642140468227, 'timing/data': 0.02236770372837782, 'timing/model': 1.4023467293009162, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 1.1}) +[2026-06-15 21:28:26,579][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2400_state +[2026-06-15 21:28:26,586][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 21:28:34,750][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2400_state/model.safetensors +[2026-06-15 21:28:46,063][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2400_state/optimizer.bin +[2026-06-15 21:28:46,064][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2400_state/sampler.bin +[2026-06-15 21:28:46,065][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2400_state/sampler_1.bin +[2026-06-15 21:28:46,066][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2400_state/random_states_0.pkl +[2026-06-15 21:28:46,233][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2400 +[2026-06-15 21:28:46,233][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 21:28:46,249][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 21:28:52,129][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.11901941895484924, 'train/grad_norm_pre_clip': 7.232957363128662, 'timing/data': 0.025792691856622696, 'timing/model': 1.4423493649810553, 'learning_rate/qwen_vl_interface': 9.0548924229603e-06, 'learning_rate/action_model': 9.054892422960301e-05, 'epoch': 1.1}) +[2026-06-15 21:28:57,854][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.2253330945968628, 'train/grad_norm_pre_clip': 8.78273868560791, 'timing/data': 0.0206935191527009, 'timing/model': 1.423178007826209, 'learning_rate/qwen_vl_interface': 9.053068578000256e-06, 'learning_rate/action_model': 9.053068578000256e-05, 'epoch': 1.1}) +[2026-06-15 21:29:03,532][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.1872539222240448, 'train/grad_norm_pre_clip': 8.249031066894531, 'timing/data': 0.022312439046800137, 'timing/model': 1.6487453952431679, 'learning_rate/qwen_vl_interface': 9.051243169742372e-06, 'learning_rate/action_model': 9.051243169742372e-05, 'epoch': 1.1}) +[2026-06-15 21:29:08,935][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.16719181835651398, 'train/grad_norm_pre_clip': 7.929013729095459, 'timing/data': 0.02088006865233183, 'timing/model': 1.422748601064086, 'learning_rate/qwen_vl_interface': 9.049416198937008e-06, 'learning_rate/action_model': 9.049416198937007e-05, 'epoch': 1.1}) +[2026-06-15 21:29:14,633][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.050708360970020294, 'train/grad_norm_pre_clip': 3.1215171813964844, 'timing/data': 0.021512462757527828, 'timing/model': 1.4487751564010978, 'learning_rate/qwen_vl_interface': 9.047587666335159e-06, 'learning_rate/action_model': 9.047587666335159e-05, 'epoch': 1.1}) +[2026-06-15 21:29:20,349][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.035185202956199646, 'train/grad_norm_pre_clip': 2.326289653778076, 'timing/data': 0.02061624638736248, 'timing/model': 1.4415021687746048, 'learning_rate/qwen_vl_interface': 9.045757572688469e-06, 'learning_rate/action_model': 9.04575757268847e-05, 'epoch': 1.1}) +[2026-06-15 21:29:26,154][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.06163635849952698, 'train/grad_norm_pre_clip': 4.7502312660217285, 'timing/data': 0.022087285295128822, 'timing/model': 1.4119362365454435, 'learning_rate/qwen_vl_interface': 9.043925918749218e-06, 'learning_rate/action_model': 9.043925918749218e-05, 'epoch': 1.1}) +[2026-06-15 21:29:31,958][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.2955371141433716, 'train/grad_norm_pre_clip': 10.370604515075684, 'timing/data': 0.0259725209325552, 'timing/model': 1.3935763286426663, 'learning_rate/qwen_vl_interface': 9.042092705270333e-06, 'learning_rate/action_model': 9.042092705270333e-05, 'epoch': 1.1}) +[2026-06-15 21:29:37,708][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.13750948011875153, 'train/grad_norm_pre_clip': 5.9673380851745605, 'timing/data': 0.021724632941186428, 'timing/model': 1.6460936972871423, 'learning_rate/qwen_vl_interface': 9.040257933005375e-06, 'learning_rate/action_model': 9.040257933005374e-05, 'epoch': 1.1}) +[2026-06-15 21:29:43,240][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.1305822879076004, 'train/grad_norm_pre_clip': 13.548624992370605, 'timing/data': 0.023272722959518433, 'timing/model': 1.445885470137, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 1.1}) +[2026-06-15 21:29:49,041][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.13080117106437683, 'train/grad_norm_pre_clip': 7.9160237312316895, 'timing/data': 0.021375050768256187, 'timing/model': 1.4375567277893424, 'learning_rate/qwen_vl_interface': 9.036583715134709e-06, 'learning_rate/action_model': 9.036583715134708e-05, 'epoch': 1.1}) +[2026-06-15 21:29:54,640][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.039371199905872345, 'train/grad_norm_pre_clip': 2.5802876949310303, 'timing/data': 0.021150747314095497, 'timing/model': 1.422672194428742, 'learning_rate/qwen_vl_interface': 9.034744271039332e-06, 'learning_rate/action_model': 9.034744271039332e-05, 'epoch': 1.1}) +[2026-06-15 21:30:00,351][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.18718993663787842, 'train/grad_norm_pre_clip': 11.275019645690918, 'timing/data': 0.02031026314944029, 'timing/model': 1.4520228588953614, 'learning_rate/qwen_vl_interface': 9.03290327117855e-06, 'learning_rate/action_model': 9.032903271178549e-05, 'epoch': 1.1}) +[2026-06-15 21:30:06,057][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.0899006724357605, 'train/grad_norm_pre_clip': 4.040101051330566, 'timing/data': 0.020181861706078053, 'timing/model': 1.3998291036114097, 'learning_rate/qwen_vl_interface': 9.031060716309122e-06, 'learning_rate/action_model': 9.031060716309121e-05, 'epoch': 1.1}) +[2026-06-15 21:30:11,896][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.27641141414642334, 'train/grad_norm_pre_clip': 18.351322174072266, 'timing/data': 0.02083447575569153, 'timing/model': 1.6990722920745611, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 1.1}) +[2026-06-15 21:30:17,439][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.07960586994886398, 'train/grad_norm_pre_clip': 11.330028533935547, 'timing/data': 0.020513514056801796, 'timing/model': 1.4499283283948898, 'learning_rate/qwen_vl_interface': 9.027370944574604e-06, 'learning_rate/action_model': 9.027370944574604e-05, 'epoch': 1.1}) +[2026-06-15 21:30:23,136][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.0825965479016304, 'train/grad_norm_pre_clip': 6.727845668792725, 'timing/data': 0.021525119431316853, 'timing/model': 1.4305260172113776, 'learning_rate/qwen_vl_interface': 9.025523729226236e-06, 'learning_rate/action_model': 9.025523729226236e-05, 'epoch': 1.11}) +[2026-06-15 21:30:28,849][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.4245768189430237, 'train/grad_norm_pre_clip': 13.93264389038086, 'timing/data': 0.021526509895920753, 'timing/model': 1.4460370726883411, 'learning_rate/qwen_vl_interface': 9.023674961902677e-06, 'learning_rate/action_model': 9.023674961902676e-05, 'epoch': 1.11}) +[2026-06-15 21:30:34,542][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.06272168457508087, 'train/grad_norm_pre_clip': 6.236940860748291, 'timing/data': 0.021122736856341362, 'timing/model': 1.454681663773954, 'learning_rate/qwen_vl_interface': 9.021824643363883e-06, 'learning_rate/action_model': 9.021824643363883e-05, 'epoch': 1.11}) +[2026-06-15 21:30:40,234][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.14816533029079437, 'train/grad_norm_pre_clip': 9.635104179382324, 'timing/data': 0.02099442481994629, 'timing/model': 1.408312059007585, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 1.11}) +[2026-06-15 21:30:45,999][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.21502351760864258, 'train/grad_norm_pre_clip': 18.935945510864258, 'timing/data': 0.02086244523525238, 'timing/model': 1.70878900680691, 'learning_rate/qwen_vl_interface': 9.018119355683623e-06, 'learning_rate/action_model': 9.018119355683623e-05, 'epoch': 1.11}) +[2026-06-15 21:30:51,801][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.07249640673398972, 'train/grad_norm_pre_clip': 7.425816059112549, 'timing/data': 0.021004525013267994, 'timing/model': 1.6991166872903705, 'learning_rate/qwen_vl_interface': 9.016264388065258e-06, 'learning_rate/action_model': 9.016264388065258e-05, 'epoch': 1.11}) +[2026-06-15 21:30:57,445][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.05263747274875641, 'train/grad_norm_pre_clip': 5.753744125366211, 'timing/data': 0.021426808089017868, 'timing/model': 1.4287830162793398, 'learning_rate/qwen_vl_interface': 9.014407872277869e-06, 'learning_rate/action_model': 9.014407872277869e-05, 'epoch': 1.11}) +[2026-06-15 21:31:03,250][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.1214522272348404, 'train/grad_norm_pre_clip': 8.232338905334473, 'timing/data': 0.020331881009042263, 'timing/model': 1.4473278420045972, 'learning_rate/qwen_vl_interface': 9.012549809084595e-06, 'learning_rate/action_model': 9.012549809084595e-05, 'epoch': 1.11}) +[2026-06-15 21:31:09,053][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.03940923139452934, 'train/grad_norm_pre_clip': 3.5977983474731445, 'timing/data': 0.021683819591999054, 'timing/model': 1.4544819435104728, 'learning_rate/qwen_vl_interface': 9.01069019924922e-06, 'learning_rate/action_model': 9.010690199249219e-05, 'epoch': 1.11}) +[2026-06-15 21:31:14,847][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.12588486075401306, 'train/grad_norm_pre_clip': 8.089356422424316, 'timing/data': 0.020595592446625233, 'timing/model': 1.394835976883769, 'learning_rate/qwen_vl_interface': 9.008829043536157e-06, 'learning_rate/action_model': 9.008829043536156e-05, 'epoch': 1.11}) +[2026-06-15 21:31:20,695][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.0844709724187851, 'train/grad_norm_pre_clip': 4.506485939025879, 'timing/data': 0.02067023329436779, 'timing/model': 1.70193777885288, 'learning_rate/qwen_vl_interface': 9.006966342710458e-06, 'learning_rate/action_model': 9.006966342710457e-05, 'epoch': 1.11}) +[2026-06-15 21:31:26,355][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.42073845863342285, 'train/grad_norm_pre_clip': 31.493432998657227, 'timing/data': 0.021586427465081215, 'timing/model': 1.4578290032222867, 'learning_rate/qwen_vl_interface': 9.005102097537809e-06, 'learning_rate/action_model': 9.005102097537809e-05, 'epoch': 1.11}) +[2026-06-15 21:31:32,149][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.44405317306518555, 'train/grad_norm_pre_clip': 20.397062301635742, 'timing/data': 0.02140910644084215, 'timing/model': 1.4363567903637886, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 1.11}) +[2026-06-15 21:31:37,944][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.1572006195783615, 'train/grad_norm_pre_clip': 15.244062423706055, 'timing/data': 0.020482259802520275, 'timing/model': 1.4420399740338326, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 1.11}) +[2026-06-15 21:31:43,679][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.07834488153457642, 'train/grad_norm_pre_clip': 5.2526535987854, 'timing/data': 0.02188601903617382, 'timing/model': 1.3953551026061177, 'learning_rate/qwen_vl_interface': 8.999500103604538e-06, 'learning_rate/action_model': 8.999500103604536e-05, 'epoch': 1.11}) +[2026-06-15 21:31:49,450][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.018524227663874626, 'train/grad_norm_pre_clip': 3.2373011112213135, 'timing/data': 0.02089435327798128, 'timing/model': 1.4519733199849725, 'learning_rate/qwen_vl_interface': 8.997629688713639e-06, 'learning_rate/action_model': 8.997629688713638e-05, 'epoch': 1.11}) +[2026-06-15 21:31:55,265][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.11966970562934875, 'train/grad_norm_pre_clip': 6.715268611907959, 'timing/data': 0.02103340346366167, 'timing/model': 1.6543852919712663, 'learning_rate/qwen_vl_interface': 8.995757733313733e-06, 'learning_rate/action_model': 8.995757733313732e-05, 'epoch': 1.11}) +[2026-06-15 21:32:00,859][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.11734520643949509, 'train/grad_norm_pre_clip': 6.840826034545898, 'timing/data': 0.02108578383922577, 'timing/model': 1.4489083448424935, 'learning_rate/qwen_vl_interface': 8.993884238174314e-06, 'learning_rate/action_model': 8.993884238174314e-05, 'epoch': 1.11}) +[2026-06-15 21:32:06,650][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.1121010109782219, 'train/grad_norm_pre_clip': 9.351661682128906, 'timing/data': 0.02123088389635086, 'timing/model': 1.4377367421984673, 'learning_rate/qwen_vl_interface': 8.992009204065505e-06, 'learning_rate/action_model': 8.992009204065505e-05, 'epoch': 1.11}) +[2026-06-15 21:32:12,452][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.23647575080394745, 'train/grad_norm_pre_clip': 9.960884094238281, 'timing/data': 0.020821422338485718, 'timing/model': 1.451807077974081, 'learning_rate/qwen_vl_interface': 8.990132631758058e-06, 'learning_rate/action_model': 8.990132631758059e-05, 'epoch': 1.11}) +[2026-06-15 21:32:18,251][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.06706659495830536, 'train/grad_norm_pre_clip': 7.788366317749023, 'timing/data': 0.021244924515485764, 'timing/model': 1.3987613003700972, 'learning_rate/qwen_vl_interface': 8.988254522023364e-06, 'learning_rate/action_model': 8.988254522023363e-05, 'epoch': 1.11}) +[2026-06-15 21:32:23,969][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.16845674812793732, 'train/grad_norm_pre_clip': 10.50230598449707, 'timing/data': 0.020641960203647614, 'timing/model': 1.3953408570960164, 'learning_rate/qwen_vl_interface': 8.986374875633441e-06, 'learning_rate/action_model': 8.98637487563344e-05, 'epoch': 1.11}) +[2026-06-15 21:32:29,779][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.056703876703977585, 'train/grad_norm_pre_clip': 4.809015274047852, 'timing/data': 0.021411065012216568, 'timing/model': 1.6839984618127346, 'learning_rate/qwen_vl_interface': 8.984493693360944e-06, 'learning_rate/action_model': 8.984493693360943e-05, 'epoch': 1.12}) +[2026-06-15 21:32:35,382][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.36929234862327576, 'train/grad_norm_pre_clip': 9.698884010314941, 'timing/data': 0.02110719308257103, 'timing/model': 1.4769369447603822, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 1.12}) +[2026-06-15 21:32:41,456][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.34595632553100586, 'train/grad_norm_pre_clip': 12.477311134338379, 'timing/data': 0.021561406552791595, 'timing/model': 1.4528409214690328, 'learning_rate/qwen_vl_interface': 8.98072672426198e-06, 'learning_rate/action_model': 8.980726724261981e-05, 'epoch': 1.12}) +[2026-06-15 21:32:47,255][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.19055823981761932, 'train/grad_norm_pre_clip': 6.183008193969727, 'timing/data': 0.02137360442429781, 'timing/model': 1.45101357717067, 'learning_rate/qwen_vl_interface': 8.978840938983977e-06, 'learning_rate/action_model': 8.978840938983977e-05, 'epoch': 1.12}) +[2026-06-15 21:32:53,052][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.15184274315834045, 'train/grad_norm_pre_clip': 7.007122039794922, 'timing/data': 0.02013547532260418, 'timing/model': 1.4034039173275232, 'learning_rate/qwen_vl_interface': 8.976953620920313e-06, 'learning_rate/action_model': 8.976953620920312e-05, 'epoch': 1.12}) +[2026-06-15 21:32:58,912][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.1356046497821808, 'train/grad_norm_pre_clip': 4.600658416748047, 'timing/data': 0.021512003615498543, 'timing/model': 1.709705932997167, 'learning_rate/qwen_vl_interface': 8.975064770846794e-06, 'learning_rate/action_model': 8.975064770846794e-05, 'epoch': 1.12}) +[2026-06-15 21:33:04,554][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.10250670462846756, 'train/grad_norm_pre_clip': 4.667198181152344, 'timing/data': 0.020788419991731644, 'timing/model': 1.4549168376252055, 'learning_rate/qwen_vl_interface': 8.973174389539856e-06, 'learning_rate/action_model': 8.973174389539855e-05, 'epoch': 1.12}) +[2026-06-15 21:33:10,350][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.2578544616699219, 'train/grad_norm_pre_clip': 7.6154255867004395, 'timing/data': 0.020224835723638535, 'timing/model': 1.4549223259091377, 'learning_rate/qwen_vl_interface': 8.971282477776563e-06, 'learning_rate/action_model': 8.971282477776563e-05, 'epoch': 1.12}) +[2026-06-15 21:33:15,953][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.19394159317016602, 'train/grad_norm_pre_clip': 6.655480861663818, 'timing/data': 0.02055390737950802, 'timing/model': 1.3917607422918081, 'learning_rate/qwen_vl_interface': 8.969389036334609e-06, 'learning_rate/action_model': 8.969389036334609e-05, 'epoch': 1.12}) +[2026-06-15 21:33:21,672][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.04367751255631447, 'train/grad_norm_pre_clip': 3.069087266921997, 'timing/data': 0.02187148667871952, 'timing/model': 1.4679885255172849, 'learning_rate/qwen_vl_interface': 8.967494065992315e-06, 'learning_rate/action_model': 8.967494065992315e-05, 'epoch': 1.12}) +[2026-06-15 21:33:27,453][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.16728000342845917, 'train/grad_norm_pre_clip': 6.267204284667969, 'timing/data': 0.021351942792534828, 'timing/model': 1.4422220336273313, 'learning_rate/qwen_vl_interface': 8.965597567528634e-06, 'learning_rate/action_model': 8.965597567528633e-05, 'epoch': 1.12}) +[2026-06-15 21:33:33,300][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.34592342376708984, 'train/grad_norm_pre_clip': 10.882123947143555, 'timing/data': 0.02198172640055418, 'timing/model': 1.6966232331469655, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 1.12}) +[2026-06-15 21:33:38,951][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.2857060134410858, 'train/grad_norm_pre_clip': 13.660321235656738, 'timing/data': 0.021046689711511135, 'timing/model': 1.4481028197333217, 'learning_rate/qwen_vl_interface': 8.961799989356048e-06, 'learning_rate/action_model': 8.961799989356048e-05, 'epoch': 1.12}) +[2026-06-15 21:33:44,758][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.2547159492969513, 'train/grad_norm_pre_clip': 11.005073547363281, 'timing/data': 0.02333370503038168, 'timing/model': 1.448473270982504, 'learning_rate/qwen_vl_interface': 8.959898911208185e-06, 'learning_rate/action_model': 8.959898911208185e-05, 'epoch': 1.12}) +[2026-06-15 21:33:50,550][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.20877410471439362, 'train/grad_norm_pre_clip': 7.717804908752441, 'timing/data': 0.02127936203032732, 'timing/model': 1.4488765615969896, 'learning_rate/qwen_vl_interface': 8.957996308061015e-06, 'learning_rate/action_model': 8.957996308061014e-05, 'epoch': 1.12}) +[2026-06-15 21:33:56,346][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.1562395840883255, 'train/grad_norm_pre_clip': 4.175321578979492, 'timing/data': 0.020805899053812027, 'timing/model': 1.4532431485131383, 'learning_rate/qwen_vl_interface': 8.956092180696626e-06, 'learning_rate/action_model': 8.956092180696625e-05, 'epoch': 1.12}) +[2026-06-15 21:34:02,056][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.2118079662322998, 'train/grad_norm_pre_clip': 10.66303825378418, 'timing/data': 0.021521562710404396, 'timing/model': 1.4393391041085124, 'learning_rate/qwen_vl_interface': 8.95418652989773e-06, 'learning_rate/action_model': 8.95418652989773e-05, 'epoch': 1.12}) +[2026-06-15 21:34:07,794][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.24662737548351288, 'train/grad_norm_pre_clip': 5.82988977432251, 'timing/data': 0.0199778126552701, 'timing/model': 1.686019524000585, 'learning_rate/qwen_vl_interface': 8.952279356447675e-06, 'learning_rate/action_model': 8.952279356447675e-05, 'epoch': 1.12}) +[2026-06-15 21:34:13,358][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.2066420316696167, 'train/grad_norm_pre_clip': 4.704595565795898, 'timing/data': 0.020941339433193207, 'timing/model': 1.458671792410314, 'learning_rate/qwen_vl_interface': 8.950370661130421e-06, 'learning_rate/action_model': 8.950370661130422e-05, 'epoch': 1.12}) +[2026-06-15 21:34:19,066][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.30704376101493835, 'train/grad_norm_pre_clip': 8.392595291137695, 'timing/data': 0.019905622117221355, 'timing/model': 1.3941567558795214, 'learning_rate/qwen_vl_interface': 8.948460444730566e-06, 'learning_rate/action_model': 8.948460444730566e-05, 'epoch': 1.12}) +[2026-06-15 21:34:24,847][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.4936216473579407, 'train/grad_norm_pre_clip': 9.726162910461426, 'timing/data': 0.020148294046521187, 'timing/model': 1.451397318392992, 'learning_rate/qwen_vl_interface': 8.946548708033324e-06, 'learning_rate/action_model': 8.946548708033323e-05, 'epoch': 1.12}) +[2026-06-15 21:34:30,548][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.19381199777126312, 'train/grad_norm_pre_clip': 3.9788568019866943, 'timing/data': 0.020657086744904518, 'timing/model': 1.407642848789692, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 1.13}) +[2026-06-15 21:34:36,248][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.17747445404529572, 'train/grad_norm_pre_clip': 3.8326048851013184, 'timing/data': 0.019554520025849342, 'timing/model': 1.4179448541253805, 'learning_rate/qwen_vl_interface': 8.942720676890677e-06, 'learning_rate/action_model': 8.942720676890677e-05, 'epoch': 1.13}) +[2026-06-15 21:34:41,996][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.13311220705509186, 'train/grad_norm_pre_clip': 4.155035972595215, 'timing/data': 0.0209122272208333, 'timing/model': 1.7054154155775905, 'learning_rate/qwen_vl_interface': 8.940804384018833e-06, 'learning_rate/action_model': 8.940804384018832e-05, 'epoch': 1.13}) +[2026-06-15 21:34:47,545][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.09012173116207123, 'train/grad_norm_pre_clip': 3.1904513835906982, 'timing/data': 0.020848887972533703, 'timing/model': 1.4490674752742052, 'learning_rate/qwen_vl_interface': 8.93888657399672e-06, 'learning_rate/action_model': 8.93888657399672e-05, 'epoch': 1.13}) +[2026-06-15 21:34:53,342][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.12810474634170532, 'train/grad_norm_pre_clip': 3.104365110397339, 'timing/data': 0.020926647819578648, 'timing/model': 1.4361200481653214, 'learning_rate/qwen_vl_interface': 8.936967247612678e-06, 'learning_rate/action_model': 8.936967247612678e-05, 'epoch': 1.13}) +[2026-06-15 21:34:59,154][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.09590867161750793, 'train/grad_norm_pre_clip': 2.631739616394043, 'timing/data': 0.02052516583353281, 'timing/model': 1.464465861208737, 'learning_rate/qwen_vl_interface': 8.93504640565567e-06, 'learning_rate/action_model': 8.93504640565567e-05, 'epoch': 1.13}) +[2026-06-15 21:35:04,844][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.260290265083313, 'train/grad_norm_pre_clip': 8.762340545654297, 'timing/data': 0.020806887187063694, 'timing/model': 1.441252020187676, 'learning_rate/qwen_vl_interface': 8.933124048915278e-06, 'learning_rate/action_model': 8.933124048915278e-05, 'epoch': 1.13}) +[2026-06-15 21:35:10,668][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.14299975335597992, 'train/grad_norm_pre_clip': 9.247265815734863, 'timing/data': 0.022273815236985683, 'timing/model': 1.4004686223343015, 'learning_rate/qwen_vl_interface': 8.931200178181715e-06, 'learning_rate/action_model': 8.931200178181715e-05, 'epoch': 1.13}) +[2026-06-15 21:35:16,503][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.4694734811782837, 'train/grad_norm_pre_clip': 5.283980846405029, 'timing/data': 0.02152175083756447, 'timing/model': 1.6883189864456654, 'learning_rate/qwen_vl_interface': 8.929274794245812e-06, 'learning_rate/action_model': 8.929274794245812e-05, 'epoch': 1.13}) +[2026-06-15 21:35:22,154][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.08230116963386536, 'train/grad_norm_pre_clip': 3.407679319381714, 'timing/data': 0.021345119923353195, 'timing/model': 1.4515255773440003, 'learning_rate/qwen_vl_interface': 8.927347897899018e-06, 'learning_rate/action_model': 8.927347897899016e-05, 'epoch': 1.13}) +[2026-06-15 21:35:27,950][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.3377474546432495, 'train/grad_norm_pre_clip': 14.96459674835205, 'timing/data': 0.020610355772078037, 'timing/model': 1.444423295557499, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 1.13}) +[2026-06-15 21:35:33,653][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.3141881823539734, 'train/grad_norm_pre_clip': 14.335542678833008, 'timing/data': 0.020645135082304478, 'timing/model': 1.447760896757245, 'learning_rate/qwen_vl_interface': 8.923489571141679e-06, 'learning_rate/action_model': 8.923489571141679e-05, 'epoch': 1.13}) +[2026-06-15 21:35:39,467][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.399219810962677, 'train/grad_norm_pre_clip': 12.624972343444824, 'timing/data': 0.021131438203155994, 'timing/model': 1.4536039745435119, 'learning_rate/qwen_vl_interface': 8.921558142317147e-06, 'learning_rate/action_model': 8.921558142317147e-05, 'epoch': 1.13}) +[2026-06-15 21:35:45,697][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.3316740393638611, 'train/grad_norm_pre_clip': 7.3743085861206055, 'timing/data': 0.020887075923383236, 'timing/model': 1.4160697283223271, 'learning_rate/qwen_vl_interface': 8.91962520425375e-06, 'learning_rate/action_model': 8.91962520425375e-05, 'epoch': 1.13}) +[2026-06-15 21:35:51,706][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.14261916279792786, 'train/grad_norm_pre_clip': 8.970433235168457, 'timing/data': 0.02061422448605299, 'timing/model': 1.691433315165341, 'learning_rate/qwen_vl_interface': 8.917690757746044e-06, 'learning_rate/action_model': 8.917690757746044e-05, 'epoch': 1.13}) +[2026-06-15 21:35:57,360][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.15697121620178223, 'train/grad_norm_pre_clip': 7.403710842132568, 'timing/data': 0.020355993881821632, 'timing/model': 1.4540488496422768, 'learning_rate/qwen_vl_interface': 8.915754803589208e-06, 'learning_rate/action_model': 8.915754803589209e-05, 'epoch': 1.13}) +[2026-06-15 21:36:03,151][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.21311688423156738, 'train/grad_norm_pre_clip': 7.244959354400635, 'timing/data': 0.02135459892451763, 'timing/model': 1.432134760543704, 'learning_rate/qwen_vl_interface': 8.913817342579042e-06, 'learning_rate/action_model': 8.913817342579042e-05, 'epoch': 1.13}) +[2026-06-15 21:36:08,957][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.10176199674606323, 'train/grad_norm_pre_clip': 6.020808696746826, 'timing/data': 0.020018051378428936, 'timing/model': 1.4368824996054173, 'learning_rate/qwen_vl_interface': 8.91187837551196e-06, 'learning_rate/action_model': 8.911878375511961e-05, 'epoch': 1.13}) +[2026-06-15 21:36:14,859][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.0759858638048172, 'train/grad_norm_pre_clip': 3.2638204097747803, 'timing/data': 0.02026803232729435, 'timing/model': 1.453562873415649, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 1.13}) +[2026-06-15 21:36:20,673][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.09826092422008514, 'train/grad_norm_pre_clip': 3.0739476680755615, 'timing/data': 0.021104708313941956, 'timing/model': 1.406062887981534, 'learning_rate/qwen_vl_interface': 8.907995926395814e-06, 'learning_rate/action_model': 8.907995926395815e-05, 'epoch': 1.13}) +[2026-06-15 21:36:26,509][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.2818564474582672, 'train/grad_norm_pre_clip': 12.398674011230469, 'timing/data': 0.020458713173866272, 'timing/model': 1.7013334287330508, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 1.13}) +[2026-06-15 21:36:32,059][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.15403881669044495, 'train/grad_norm_pre_clip': 5.360954761505127, 'timing/data': 0.02083541452884674, 'timing/model': 1.4403744312003255, 'learning_rate/qwen_vl_interface': 8.904107462624484e-06, 'learning_rate/action_model': 8.904107462624483e-05, 'epoch': 1.13}) +[2026-06-15 21:36:37,856][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.1801268756389618, 'train/grad_norm_pre_clip': 5.929924488067627, 'timing/data': 0.020930824801325798, 'timing/model': 1.446535580791533, 'learning_rate/qwen_vl_interface': 8.902160977240739e-06, 'learning_rate/action_model': 8.902160977240739e-05, 'epoch': 1.14}) +[2026-06-15 21:36:43,652][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.25045448541641235, 'train/grad_norm_pre_clip': 6.717586517333984, 'timing/data': 0.021310958079993725, 'timing/model': 1.4443189892917871, 'learning_rate/qwen_vl_interface': 8.90021299059157e-06, 'learning_rate/action_model': 8.90021299059157e-05, 'epoch': 1.14}) +[2026-06-15 21:36:49,452][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.10811533033847809, 'train/grad_norm_pre_clip': 5.5489821434021, 'timing/data': 0.02172967977821827, 'timing/model': 1.4131541848182678, 'learning_rate/qwen_vl_interface': 8.898263503477722e-06, 'learning_rate/action_model': 8.898263503477721e-05, 'epoch': 1.14}) +[2026-06-15 21:36:55,274][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.14487433433532715, 'train/grad_norm_pre_clip': 4.369185447692871, 'timing/data': 0.0215694485232234, 'timing/model': 1.3993705343455076, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 1.14}) +[2026-06-15 21:37:01,111][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.039647385478019714, 'train/grad_norm_pre_clip': 1.8745092153549194, 'timing/data': 0.021269796416163445, 'timing/model': 1.6909022433683276, 'learning_rate/qwen_vl_interface': 8.894360031062043e-06, 'learning_rate/action_model': 8.894360031062043e-05, 'epoch': 1.14}) +[2026-06-15 21:37:06,947][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.25856685638427734, 'train/grad_norm_pre_clip': 3.306061267852783, 'timing/data': 0.020356301218271255, 'timing/model': 1.6322990637272596, 'learning_rate/qwen_vl_interface': 8.892406047364783e-06, 'learning_rate/action_model': 8.892406047364784e-05, 'epoch': 1.14}) +[2026-06-15 21:37:12,567][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.17329198122024536, 'train/grad_norm_pre_clip': 6.793391704559326, 'timing/data': 0.02133016660809517, 'timing/model': 1.4406250417232513, 'learning_rate/qwen_vl_interface': 8.890450566411984e-06, 'learning_rate/action_model': 8.890450566411984e-05, 'epoch': 1.14}) +[2026-06-15 21:37:18,363][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.02889518439769745, 'train/grad_norm_pre_clip': 3.4932968616485596, 'timing/data': 0.020818233489990234, 'timing/model': 1.4586364226415753, 'learning_rate/qwen_vl_interface': 8.888493589007467e-06, 'learning_rate/action_model': 8.888493589007468e-05, 'epoch': 1.14}) +[2026-06-15 21:37:24,162][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.25072842836380005, 'train/grad_norm_pre_clip': 11.484822273254395, 'timing/data': 0.02124695759266615, 'timing/model': 1.4517727931961417, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 1.14}) +[2026-06-15 21:37:29,950][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.06605885177850723, 'train/grad_norm_pre_clip': 3.1126909255981445, 'timing/data': 0.020133130252361298, 'timing/model': 1.4040374318137765, 'learning_rate/qwen_vl_interface': 8.884575148061662e-06, 'learning_rate/action_model': 8.88457514806166e-05, 'epoch': 1.14}) +[2026-06-15 21:37:35,807][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.29987385869026184, 'train/grad_norm_pre_clip': 10.999792098999023, 'timing/data': 0.02090526558458805, 'timing/model': 1.6863256273791194, 'learning_rate/qwen_vl_interface': 8.882613686131095e-06, 'learning_rate/action_model': 8.882613686131093e-05, 'epoch': 1.14}) +[2026-06-15 21:37:41,710][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.34719863533973694, 'train/grad_norm_pre_clip': 11.985146522521973, 'timing/data': 0.02042018063366413, 'timing/model': 1.7140990840271115, 'learning_rate/qwen_vl_interface': 8.880650730970254e-06, 'learning_rate/action_model': 8.880650730970254e-05, 'epoch': 1.14}) +[2026-06-15 21:37:47,352][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.2898658215999603, 'train/grad_norm_pre_clip': 11.676183700561523, 'timing/data': 0.020052080042660236, 'timing/model': 1.4374444503337145, 'learning_rate/qwen_vl_interface': 8.878686283386042e-06, 'learning_rate/action_model': 8.878686283386042e-05, 'epoch': 1.14}) +[2026-06-15 21:37:53,151][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.11204054951667786, 'train/grad_norm_pre_clip': 6.073657035827637, 'timing/data': 0.02065750304609537, 'timing/model': 1.4399807285517454, 'learning_rate/qwen_vl_interface': 8.876720344185967e-06, 'learning_rate/action_model': 8.876720344185967e-05, 'epoch': 1.14}) +[2026-06-15 21:37:58,959][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.29204505681991577, 'train/grad_norm_pre_clip': 8.762880325317383, 'timing/data': 0.02284772600978613, 'timing/model': 1.4590366557240486, 'learning_rate/qwen_vl_interface': 8.874752914178149e-06, 'learning_rate/action_model': 8.87475291417815e-05, 'epoch': 1.14}) +[2026-06-15 21:38:04,757][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.14182990789413452, 'train/grad_norm_pre_clip': 6.098601341247559, 'timing/data': 0.01985120866447687, 'timing/model': 1.3976410254836082, 'learning_rate/qwen_vl_interface': 8.87278399417133e-06, 'learning_rate/action_model': 8.872783994171329e-05, 'epoch': 1.14}) +[2026-06-15 21:38:10,607][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.14041122794151306, 'train/grad_norm_pre_clip': 7.107179641723633, 'timing/data': 0.01912970282137394, 'timing/model': 1.701188139617443, 'learning_rate/qwen_vl_interface': 8.870813584974852e-06, 'learning_rate/action_model': 8.870813584974852e-05, 'epoch': 1.14}) +[2026-06-15 21:38:16,260][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.34965699911117554, 'train/grad_norm_pre_clip': 8.255243301391602, 'timing/data': 0.02037657145410776, 'timing/model': 1.4529455807060003, 'learning_rate/qwen_vl_interface': 8.868841687398683e-06, 'learning_rate/action_model': 8.868841687398683e-05, 'epoch': 1.14}) +[2026-06-15 21:48:59,461][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.25347742438316345, 'train/grad_norm_pre_clip': 6.896884918212891, 'eval/loss': 0.09088236838579178, 'eval/flap_precision': 0.7524752475247525, 'eval/flap_recall': 0.8290909090909091, 'eval/flap_f1': 0.7889273356401384, 'eval/flap_rate_model': 0.08416666666666667, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7889273356401384, 'timing/data': 0.02006348967552185, 'timing/model': 1.451240174472332, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 1.14}) +[2026-06-15 21:48:59,462][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2500_state +[2026-06-15 21:48:59,468][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 21:49:07,088][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2500_state/model.safetensors +[2026-06-15 21:49:18,278][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2500_state/optimizer.bin +[2026-06-15 21:49:18,279][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2500_state/sampler.bin +[2026-06-15 21:49:18,279][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-15 21:49:18,280][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-15 21:49:18,449][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-15 21:49:18,450][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 21:49:18,465][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 21:49:24,319][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.056387633085250854, 'train/grad_norm_pre_clip': 2.0976476669311523, 'timing/data': 0.02195085398852825, 'timing/model': 1.4179242393001914, 'learning_rate/qwen_vl_interface': 8.864893430350159e-06, 'learning_rate/action_model': 8.864893430350159e-05, 'epoch': 1.14}) +[2026-06-15 21:49:30,024][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.13505414128303528, 'train/grad_norm_pre_clip': 5.631995677947998, 'timing/data': 0.021825453266501427, 'timing/model': 1.4037530664354563, 'learning_rate/qwen_vl_interface': 8.862917072500785e-06, 'learning_rate/action_model': 8.862917072500784e-05, 'epoch': 1.14}) +[2026-06-15 21:49:35,731][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.1963815838098526, 'train/grad_norm_pre_clip': 4.673565864562988, 'timing/data': 0.0214640311896801, 'timing/model': 1.432532338425517, 'learning_rate/qwen_vl_interface': 8.860939229517674e-06, 'learning_rate/action_model': 8.860939229517674e-05, 'epoch': 1.14}) +[2026-06-15 21:49:41,426][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.13314715027809143, 'train/grad_norm_pre_clip': 5.809999942779541, 'timing/data': 0.021035759709775448, 'timing/model': 1.3911524331197143, 'learning_rate/qwen_vl_interface': 8.858959902213844e-06, 'learning_rate/action_model': 8.858959902213845e-05, 'epoch': 1.15}) +[2026-06-15 21:49:47,182][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.07950882613658905, 'train/grad_norm_pre_clip': 3.1557199954986572, 'timing/data': 0.0217822827398777, 'timing/model': 1.6973019996657968, 'learning_rate/qwen_vl_interface': 8.856979091402921e-06, 'learning_rate/action_model': 8.85697909140292e-05, 'epoch': 1.15}) +[2026-06-15 21:49:52,796][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.10896902531385422, 'train/grad_norm_pre_clip': 5.574736595153809, 'timing/data': 0.02506901230663061, 'timing/model': 1.582034407183528, 'learning_rate/qwen_vl_interface': 8.854996797899139e-06, 'learning_rate/action_model': 8.854996797899139e-05, 'epoch': 1.15}) +[2026-06-15 21:49:58,261][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.17043422162532806, 'train/grad_norm_pre_clip': 4.459750652313232, 'timing/data': 0.02081408817321062, 'timing/model': 1.3828186420723796, 'learning_rate/qwen_vl_interface': 8.853013022517348e-06, 'learning_rate/action_model': 8.853013022517348e-05, 'epoch': 1.15}) +[2026-06-15 21:50:03,934][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.23885636031627655, 'train/grad_norm_pre_clip': 12.265862464904785, 'timing/data': 0.02207223419100046, 'timing/model': 1.4115330129861832, 'learning_rate/qwen_vl_interface': 8.851027766073002e-06, 'learning_rate/action_model': 8.851027766073001e-05, 'epoch': 1.15}) +[2026-06-15 21:50:09,652][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.2561988830566406, 'train/grad_norm_pre_clip': 6.0043158531188965, 'timing/data': 0.02495316043496132, 'timing/model': 1.4586190562695265, 'learning_rate/qwen_vl_interface': 8.84904102938216e-06, 'learning_rate/action_model': 8.84904102938216e-05, 'epoch': 1.15}) +[2026-06-15 21:50:15,346][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.3236829340457916, 'train/grad_norm_pre_clip': 7.135317325592041, 'timing/data': 0.02189699374139309, 'timing/model': 1.3968027560040355, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 1.15}) +[2026-06-15 21:50:21,086][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.3570684790611267, 'train/grad_norm_pre_clip': 9.209484100341797, 'timing/data': 0.022153954952955246, 'timing/model': 1.6898122755810618, 'learning_rate/qwen_vl_interface': 8.845063118528303e-06, 'learning_rate/action_model': 8.845063118528304e-05, 'epoch': 1.15}) +[2026-06-15 21:50:26,805][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.1408618688583374, 'train/grad_norm_pre_clip': 4.321238040924072, 'timing/data': 0.02211610320955515, 'timing/model': 1.5998501079156995, 'learning_rate/qwen_vl_interface': 8.843071946000455e-06, 'learning_rate/action_model': 8.843071946000456e-05, 'epoch': 1.15}) +[2026-06-15 21:50:32,444][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.11279616504907608, 'train/grad_norm_pre_clip': 3.070613384246826, 'timing/data': 0.020082992501556873, 'timing/model': 1.450286434032023, 'learning_rate/qwen_vl_interface': 8.841079296496451e-06, 'learning_rate/action_model': 8.84107929649645e-05, 'epoch': 1.15}) +[2026-06-15 21:50:38,089][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.15367114543914795, 'train/grad_norm_pre_clip': 6.142316818237305, 'timing/data': 0.021261959336698055, 'timing/model': 1.3870759457349777, 'learning_rate/qwen_vl_interface': 8.839085170835396e-06, 'learning_rate/action_model': 8.839085170835395e-05, 'epoch': 1.15}) +[2026-06-15 21:50:43,852][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.13608787953853607, 'train/grad_norm_pre_clip': 4.701083660125732, 'timing/data': 0.020165284164249897, 'timing/model': 1.4521773308515549, 'learning_rate/qwen_vl_interface': 8.837089569837e-06, 'learning_rate/action_model': 8.837089569836998e-05, 'epoch': 1.15}) +[2026-06-15 21:50:49,541][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.1390722095966339, 'train/grad_norm_pre_clip': 7.279653072357178, 'timing/data': 0.0229537608101964, 'timing/model': 1.3914147773757577, 'learning_rate/qwen_vl_interface': 8.835092494321575e-06, 'learning_rate/action_model': 8.835092494321575e-05, 'epoch': 1.15}) +[2026-06-15 21:50:55,283][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.09428621828556061, 'train/grad_norm_pre_clip': 3.2976114749908447, 'timing/data': 0.021637671627104282, 'timing/model': 1.6820102129131556, 'learning_rate/qwen_vl_interface': 8.833093945110048e-06, 'learning_rate/action_model': 8.833093945110047e-05, 'epoch': 1.15}) +[2026-06-15 21:51:00,940][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.20095878839492798, 'train/grad_norm_pre_clip': 12.52938461303711, 'timing/data': 0.021058598533272743, 'timing/model': 1.4347409093752503, 'learning_rate/qwen_vl_interface': 8.831093923023946e-06, 'learning_rate/action_model': 8.831093923023945e-05, 'epoch': 1.15}) +[2026-06-15 21:51:06,748][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.20741909742355347, 'train/grad_norm_pre_clip': 8.602535247802734, 'timing/data': 0.020627625286579132, 'timing/model': 1.4374655429273844, 'learning_rate/qwen_vl_interface': 8.8290924288854e-06, 'learning_rate/action_model': 8.8290924288854e-05, 'epoch': 1.15}) +[2026-06-15 21:51:12,553][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.08150050044059753, 'train/grad_norm_pre_clip': 2.472665786743164, 'timing/data': 0.02015985269099474, 'timing/model': 1.4502389943227172, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 1.15}) +[2026-06-15 21:51:18,380][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.10094215720891953, 'train/grad_norm_pre_clip': 2.921227216720581, 'timing/data': 0.021000479348003864, 'timing/model': 1.4886435279622674, 'learning_rate/qwen_vl_interface': 8.825085027742545e-06, 'learning_rate/action_model': 8.825085027742544e-05, 'epoch': 1.15}) +[2026-06-15 21:51:24,158][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.08950678259134293, 'train/grad_norm_pre_clip': 4.750988006591797, 'timing/data': 0.020456855185329914, 'timing/model': 1.4202633816748857, 'learning_rate/qwen_vl_interface': 8.823079122385523e-06, 'learning_rate/action_model': 8.823079122385523e-05, 'epoch': 1.15}) +[2026-06-15 21:51:30,034][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.006868263706564903, 'train/grad_norm_pre_clip': 1.2382011413574219, 'timing/data': 0.02061796560883522, 'timing/model': 1.7220904985442758, 'learning_rate/qwen_vl_interface': 8.821071748270644e-06, 'learning_rate/action_model': 8.821071748270644e-05, 'epoch': 1.15}) +[2026-06-15 21:51:35,654][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.09908120334148407, 'train/grad_norm_pre_clip': 2.7954912185668945, 'timing/data': 0.02168826200067997, 'timing/model': 1.4390095453709364, 'learning_rate/qwen_vl_interface': 8.81906290622306e-06, 'learning_rate/action_model': 8.81906290622306e-05, 'epoch': 1.15}) +[2026-06-15 21:51:41,457][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.09485024213790894, 'train/grad_norm_pre_clip': 6.158914566040039, 'timing/data': 0.02034127525985241, 'timing/model': 1.4444584446027875, 'learning_rate/qwen_vl_interface': 8.81705259706853e-06, 'learning_rate/action_model': 8.817052597068529e-05, 'epoch': 1.15}) +[2026-06-15 21:51:47,257][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.1622999906539917, 'train/grad_norm_pre_clip': 12.270293235778809, 'timing/data': 0.02104341797530651, 'timing/model': 1.4545757295563817, 'learning_rate/qwen_vl_interface': 8.815040821633416e-06, 'learning_rate/action_model': 8.815040821633416e-05, 'epoch': 1.16}) +[2026-06-15 21:51:53,056][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.06228726729750633, 'train/grad_norm_pre_clip': 3.8542375564575195, 'timing/data': 0.021106497384607792, 'timing/model': 1.455138212069869, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 1.16}) +[2026-06-15 21:51:58,855][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.12980777025222778, 'train/grad_norm_pre_clip': 8.154492378234863, 'timing/data': 0.02107168920338154, 'timing/model': 1.4002414904534817, 'learning_rate/qwen_vl_interface': 8.8110128752299e-06, 'learning_rate/action_model': 8.8110128752299e-05, 'epoch': 1.16}) +[2026-06-15 21:52:04,711][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.0712030753493309, 'train/grad_norm_pre_clip': 4.977311134338379, 'timing/data': 0.021044747903943062, 'timing/model': 1.6729711135849357, 'learning_rate/qwen_vl_interface': 8.808996705917235e-06, 'learning_rate/action_model': 8.808996705917235e-05, 'epoch': 1.16}) +[2026-06-15 21:52:10,567][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.06203410029411316, 'train/grad_norm_pre_clip': 7.709764003753662, 'timing/data': 0.019968471489846706, 'timing/model': 1.636722887866199, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 1.16}) +[2026-06-15 21:52:16,144][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.2117556929588318, 'train/grad_norm_pre_clip': 9.203907012939453, 'timing/data': 0.02109047956764698, 'timing/model': 1.4264960857108235, 'learning_rate/qwen_vl_interface': 8.804959979213945e-06, 'learning_rate/action_model': 8.804959979213945e-05, 'epoch': 1.16}) +[2026-06-15 21:52:21,964][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.10946794599294662, 'train/grad_norm_pre_clip': 8.620050430297852, 'timing/data': 0.020583365112543106, 'timing/model': 1.4657790465280414, 'learning_rate/qwen_vl_interface': 8.802939423482667e-06, 'learning_rate/action_model': 8.802939423482666e-05, 'epoch': 1.16}) +[2026-06-15 21:52:27,757][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.21617329120635986, 'train/grad_norm_pre_clip': 10.919414520263672, 'timing/data': 0.020693755708634853, 'timing/model': 1.4385219411924481, 'learning_rate/qwen_vl_interface': 8.800917407272198e-06, 'learning_rate/action_model': 8.800917407272198e-05, 'epoch': 1.16}) +[2026-06-15 21:52:33,563][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.0055847736075520515, 'train/grad_norm_pre_clip': 0.45080843567848206, 'timing/data': 0.02119390945881605, 'timing/model': 1.4084063712507486, 'learning_rate/qwen_vl_interface': 8.798893931413711e-06, 'learning_rate/action_model': 8.79889393141371e-05, 'epoch': 1.16}) +[2026-06-15 21:52:39,396][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.23180101811885834, 'train/grad_norm_pre_clip': 6.73293924331665, 'timing/data': 0.020204522646963596, 'timing/model': 1.679495831951499, 'learning_rate/qwen_vl_interface': 8.796868996738985e-06, 'learning_rate/action_model': 8.796868996738984e-05, 'epoch': 1.16}) +[2026-06-15 21:52:45,305][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.17710347473621368, 'train/grad_norm_pre_clip': 21.208904266357422, 'timing/data': 0.02084381692111492, 'timing/model': 1.696069193072617, 'learning_rate/qwen_vl_interface': 8.794842604080389e-06, 'learning_rate/action_model': 8.794842604080389e-05, 'epoch': 1.16}) +[2026-06-15 21:52:50,965][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.34038811922073364, 'train/grad_norm_pre_clip': 22.924537658691406, 'timing/data': 0.021585150621831417, 'timing/model': 1.4628450432792306, 'learning_rate/qwen_vl_interface': 8.7928147542709e-06, 'learning_rate/action_model': 8.7928147542709e-05, 'epoch': 1.16}) +[2026-06-15 21:52:56,754][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.009701283648610115, 'train/grad_norm_pre_clip': 1.4099117517471313, 'timing/data': 0.021090468391776085, 'timing/model': 1.4504674328491092, 'learning_rate/qwen_vl_interface': 8.790785448144088e-06, 'learning_rate/action_model': 8.790785448144088e-05, 'epoch': 1.16}) +[2026-06-15 21:53:02,553][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.24965570867061615, 'train/grad_norm_pre_clip': 11.23909854888916, 'timing/data': 0.021381010301411152, 'timing/model': 1.4509952142834663, 'learning_rate/qwen_vl_interface': 8.788754686534126e-06, 'learning_rate/action_model': 8.788754686534127e-05, 'epoch': 1.16}) +[2026-06-15 21:53:08,360][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.08132471144199371, 'train/grad_norm_pre_clip': 5.201084136962891, 'timing/data': 0.02132411953061819, 'timing/model': 1.4020702289417386, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 1.16}) +[2026-06-15 21:53:14,206][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.14447739720344543, 'train/grad_norm_pre_clip': 14.714070320129395, 'timing/data': 0.020603375509381294, 'timing/model': 1.7022343324497342, 'learning_rate/qwen_vl_interface': 8.784688800204427e-06, 'learning_rate/action_model': 8.784688800204426e-05, 'epoch': 1.16}) +[2026-06-15 21:53:19,852][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.024995816871523857, 'train/grad_norm_pre_clip': 2.434621572494507, 'timing/data': 0.020537864416837692, 'timing/model': 1.4396538473665714, 'learning_rate/qwen_vl_interface': 8.782653677156022e-06, 'learning_rate/action_model': 8.782653677156022e-05, 'epoch': 1.16}) +[2026-06-15 21:53:25,653][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.1096021756529808, 'train/grad_norm_pre_clip': 8.551085472106934, 'timing/data': 0.020313944667577744, 'timing/model': 1.4409586237743497, 'learning_rate/qwen_vl_interface': 8.780617101967129e-06, 'learning_rate/action_model': 8.780617101967129e-05, 'epoch': 1.16}) +[2026-06-15 21:53:31,457][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.1504552960395813, 'train/grad_norm_pre_clip': 20.885019302368164, 'timing/data': 0.020771866664290428, 'timing/model': 1.460533982142806, 'learning_rate/qwen_vl_interface': 8.77857907547491e-06, 'learning_rate/action_model': 8.77857907547491e-05, 'epoch': 1.16}) +[2026-06-15 21:53:37,252][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.12995940446853638, 'train/grad_norm_pre_clip': 7.0058722496032715, 'timing/data': 0.02058908436447382, 'timing/model': 1.4053750913590193, 'learning_rate/qwen_vl_interface': 8.776539598517119e-06, 'learning_rate/action_model': 8.776539598517118e-05, 'epoch': 1.16}) +[2026-06-15 21:53:43,052][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.048189349472522736, 'train/grad_norm_pre_clip': 11.550917625427246, 'timing/data': 0.022136664018034935, 'timing/model': 1.449774949811399, 'learning_rate/qwen_vl_interface': 8.774498671932108e-06, 'learning_rate/action_model': 8.774498671932107e-05, 'epoch': 1.16}) +[2026-06-15 21:53:48,858][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.1622411012649536, 'train/grad_norm_pre_clip': 9.691094398498535, 'timing/data': 0.022026363760232925, 'timing/model': 1.657660024240613, 'learning_rate/qwen_vl_interface': 8.772456296558824e-06, 'learning_rate/action_model': 8.772456296558824e-05, 'epoch': 1.16}) +[2026-06-15 21:53:54,449][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.3066082000732422, 'train/grad_norm_pre_clip': 27.735626220703125, 'timing/data': 0.020883136428892612, 'timing/model': 1.4489028630778193, 'learning_rate/qwen_vl_interface': 8.77041247323681e-06, 'learning_rate/action_model': 8.770412473236811e-05, 'epoch': 1.17}) +[2026-06-15 21:54:00,250][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.07769802957773209, 'train/grad_norm_pre_clip': 6.534974575042725, 'timing/data': 0.02073175646364689, 'timing/model': 1.4437128128483891, 'learning_rate/qwen_vl_interface': 8.76836720280621e-06, 'learning_rate/action_model': 8.768367202806209e-05, 'epoch': 1.17}) +[2026-06-15 21:54:06,062][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.2354072630405426, 'train/grad_norm_pre_clip': 18.783851623535156, 'timing/data': 0.024327016435563564, 'timing/model': 1.4555090675130486, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 1.17}) +[2026-06-15 21:54:11,849][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.027567969635128975, 'train/grad_norm_pre_clip': 3.141120433807373, 'timing/data': 0.02038203366100788, 'timing/model': 1.3999816263094544, 'learning_rate/qwen_vl_interface': 8.76427232398276e-06, 'learning_rate/action_model': 8.76427232398276e-05, 'epoch': 1.17}) +[2026-06-15 21:54:17,662][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.20224060118198395, 'train/grad_norm_pre_clip': 11.359662055969238, 'timing/data': 0.020483274944126606, 'timing/model': 1.3983020344749093, 'learning_rate/qwen_vl_interface': 8.762222717273167e-06, 'learning_rate/action_model': 8.762222717273167e-05, 'epoch': 1.17}) +[2026-06-15 21:54:23,467][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.054458267986774445, 'train/grad_norm_pre_clip': 6.942302703857422, 'timing/data': 0.020239373669028282, 'timing/model': 1.6567873433232307, 'learning_rate/qwen_vl_interface': 8.76017166682148e-06, 'learning_rate/action_model': 8.760171666821481e-05, 'epoch': 1.17}) +[2026-06-15 21:54:29,052][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.04470652714371681, 'train/grad_norm_pre_clip': 4.039013862609863, 'timing/data': 0.020985767245292664, 'timing/model': 1.4364418182522058, 'learning_rate/qwen_vl_interface': 8.758119173470815e-06, 'learning_rate/action_model': 8.758119173470815e-05, 'epoch': 1.17}) +[2026-06-15 21:54:34,847][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.04746970906853676, 'train/grad_norm_pre_clip': 4.876067638397217, 'timing/data': 0.021060087718069553, 'timing/model': 1.4359942134469748, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 1.17}) +[2026-06-15 21:54:40,650][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.4718555510044098, 'train/grad_norm_pre_clip': 21.40221405029297, 'timing/data': 0.02301967889070511, 'timing/model': 1.4532399699091911, 'learning_rate/qwen_vl_interface': 8.75400986144794e-06, 'learning_rate/action_model': 8.75400986144794e-05, 'epoch': 1.17}) +[2026-06-15 21:54:46,450][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.3816007673740387, 'train/grad_norm_pre_clip': 25.23872184753418, 'timing/data': 0.02111123874783516, 'timing/model': 1.4625707911327481, 'learning_rate/qwen_vl_interface': 8.751953044464913e-06, 'learning_rate/action_model': 8.751953044464913e-05, 'epoch': 1.17}) +[2026-06-15 21:54:52,301][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.39969754219055176, 'train/grad_norm_pre_clip': 10.300442695617676, 'timing/data': 0.020060022361576557, 'timing/model': 1.6925951074808836, 'learning_rate/qwen_vl_interface': 8.749894787961272e-06, 'learning_rate/action_model': 8.749894787961271e-05, 'epoch': 1.17}) +[2026-06-15 21:54:57,951][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.27853924036026, 'train/grad_norm_pre_clip': 11.061715126037598, 'timing/data': 0.0207335464656353, 'timing/model': 1.4413685714825988, 'learning_rate/qwen_vl_interface': 8.747835092783086e-06, 'learning_rate/action_model': 8.747835092783085e-05, 'epoch': 1.17}) +[2026-06-15 21:55:03,747][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.13032406568527222, 'train/grad_norm_pre_clip': 11.496376037597656, 'timing/data': 0.021332649514079094, 'timing/model': 1.4353203251957893, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 1.17}) +[2026-06-15 21:55:09,449][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.10594959557056427, 'train/grad_norm_pre_clip': 4.824129581451416, 'timing/data': 0.021493920125067234, 'timing/model': 1.4411624176427722, 'learning_rate/qwen_vl_interface': 8.743711389790321e-06, 'learning_rate/action_model': 8.74371138979032e-05, 'epoch': 1.17}) +[2026-06-15 21:55:15,266][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.16034932434558868, 'train/grad_norm_pre_clip': 7.005853176116943, 'timing/data': 0.021104497835040092, 'timing/model': 1.441174766048789, 'learning_rate/qwen_vl_interface': 8.741647383670839e-06, 'learning_rate/action_model': 8.741647383670839e-05, 'epoch': 1.17}) +[2026-06-15 21:55:21,050][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.15367145836353302, 'train/grad_norm_pre_clip': 8.777302742004395, 'timing/data': 0.021352289244532585, 'timing/model': 1.4219703488051891, 'learning_rate/qwen_vl_interface': 8.739581942267011e-06, 'learning_rate/action_model': 8.739581942267011e-05, 'epoch': 1.17}) +[2026-06-15 21:55:26,901][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.21461106836795807, 'train/grad_norm_pre_clip': 8.733368873596191, 'timing/data': 0.02087603695690632, 'timing/model': 1.7057712562382221, 'learning_rate/qwen_vl_interface': 8.737515066427861e-06, 'learning_rate/action_model': 8.73751506642786e-05, 'epoch': 1.17}) +[2026-06-15 21:55:32,460][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.1327599734067917, 'train/grad_norm_pre_clip': 9.968825340270996, 'timing/data': 0.020796596072614193, 'timing/model': 1.4592700554057956, 'learning_rate/qwen_vl_interface': 8.735446757003e-06, 'learning_rate/action_model': 8.735446757003e-05, 'epoch': 1.17}) +[2026-06-15 21:55:38,347][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.0913243368268013, 'train/grad_norm_pre_clip': 7.223310947418213, 'timing/data': 0.020970276556909084, 'timing/model': 1.4275806387886405, 'learning_rate/qwen_vl_interface': 8.733377014842635e-06, 'learning_rate/action_model': 8.733377014842635e-05, 'epoch': 1.17}) +[2026-06-15 21:55:44,146][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.17312605679035187, 'train/grad_norm_pre_clip': 8.579110145568848, 'timing/data': 0.01975633017718792, 'timing/model': 1.4404682591557503, 'learning_rate/qwen_vl_interface': 8.731305840797557e-06, 'learning_rate/action_model': 8.731305840797558e-05, 'epoch': 1.17}) +[2026-06-15 21:55:49,859][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.08184940367937088, 'train/grad_norm_pre_clip': 6.900023937225342, 'timing/data': 0.021442200057208538, 'timing/model': 1.4627766720950603, 'learning_rate/qwen_vl_interface': 8.72923323571915e-06, 'learning_rate/action_model': 8.72923323571915e-05, 'epoch': 1.17}) +[2026-06-15 21:55:55,546][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.05351383239030838, 'train/grad_norm_pre_clip': 4.182806015014648, 'timing/data': 0.02171689085662365, 'timing/model': 1.4098176658153534, 'learning_rate/qwen_vl_interface': 8.727159200459381e-06, 'learning_rate/action_model': 8.727159200459381e-05, 'epoch': 1.17}) +[2026-06-15 21:56:01,296][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.1543460190296173, 'train/grad_norm_pre_clip': 8.199301719665527, 'timing/data': 0.020979267545044422, 'timing/model': 1.7010758044198155, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 1.18}) +[2026-06-15 21:56:06,846][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.20833390951156616, 'train/grad_norm_pre_clip': 7.420517444610596, 'timing/data': 0.021597261540591717, 'timing/model': 1.4382986733689904, 'learning_rate/qwen_vl_interface': 8.723006842806575e-06, 'learning_rate/action_model': 8.723006842806576e-05, 'epoch': 1.18}) +[2026-06-15 21:56:12,549][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.056983284652233124, 'train/grad_norm_pre_clip': 4.3079142570495605, 'timing/data': 0.020001191645860672, 'timing/model': 1.402456701733172, 'learning_rate/qwen_vl_interface': 8.720928522120419e-06, 'learning_rate/action_model': 8.720928522120418e-05, 'epoch': 1.18}) +[2026-06-15 21:56:18,795][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.425639271736145, 'train/grad_norm_pre_clip': 13.464868545532227, 'timing/data': 0.028939012438058853, 'timing/model': 1.4481130661442876, 'learning_rate/qwen_vl_interface': 8.718848774666655e-06, 'learning_rate/action_model': 8.718848774666655e-05, 'epoch': 1.18}) +[2026-06-15 21:56:24,756][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.3225284814834595, 'train/grad_norm_pre_clip': 9.90349292755127, 'timing/data': 0.022428705357015133, 'timing/model': 1.4667509403079748, 'learning_rate/qwen_vl_interface': 8.716767601300189e-06, 'learning_rate/action_model': 8.716767601300188e-05, 'epoch': 1.18}) +[2026-06-15 21:56:30,467][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.41242077946662903, 'train/grad_norm_pre_clip': 11.868292808532715, 'timing/data': 0.02146617043763399, 'timing/model': 1.4025507997721434, 'learning_rate/qwen_vl_interface': 8.714685002876514e-06, 'learning_rate/action_model': 8.714685002876513e-05, 'epoch': 1.18}) +[2026-06-15 21:56:36,300][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.026351898908615112, 'train/grad_norm_pre_clip': 1.472426414489746, 'timing/data': 0.02117882762104273, 'timing/model': 1.6853429609909654, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 1.18}) +[2026-06-15 21:56:41,951][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.07995916903018951, 'train/grad_norm_pre_clip': 5.2739577293396, 'timing/data': 0.020403053611516953, 'timing/model': 1.433437961153686, 'learning_rate/qwen_vl_interface': 8.710515534282428e-06, 'learning_rate/action_model': 8.710515534282428e-05, 'epoch': 1.18}) +[2026-06-15 21:56:47,748][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.10530798137187958, 'train/grad_norm_pre_clip': 5.732309818267822, 'timing/data': 0.02095811814069748, 'timing/model': 1.4402588186785579, 'learning_rate/qwen_vl_interface': 8.708428665825931e-06, 'learning_rate/action_model': 8.708428665825931e-05, 'epoch': 1.18}) +[2026-06-15 21:56:53,572][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.04921155422925949, 'train/grad_norm_pre_clip': 3.5008623600006104, 'timing/data': 0.02379780262708664, 'timing/model': 1.4686395525932312, 'learning_rate/qwen_vl_interface': 8.706340375740045e-06, 'learning_rate/action_model': 8.706340375740045e-05, 'epoch': 1.18}) +[2026-06-15 21:56:59,252][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.22992436587810516, 'train/grad_norm_pre_clip': 11.037910461425781, 'timing/data': 0.022619051858782768, 'timing/model': 1.454068724066019, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 1.18}) +[2026-06-15 21:57:05,061][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.14059428870677948, 'train/grad_norm_pre_clip': 6.344755172729492, 'timing/data': 0.021138954907655716, 'timing/model': 1.4088862054049969, 'learning_rate/qwen_vl_interface': 8.70215953411436e-06, 'learning_rate/action_model': 8.70215953411436e-05, 'epoch': 1.18}) +[2026-06-15 21:57:10,906][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.11884955316781998, 'train/grad_norm_pre_clip': 7.835975170135498, 'timing/data': 0.0243335897102952, 'timing/model': 1.6820405256003141, 'learning_rate/qwen_vl_interface': 8.70006698429315e-06, 'learning_rate/action_model': 8.700066984293149e-05, 'epoch': 1.18}) +[2026-06-15 21:57:16,573][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.17357443273067474, 'train/grad_norm_pre_clip': 4.878937721252441, 'timing/data': 0.01984729990363121, 'timing/model': 1.4408466592431068, 'learning_rate/qwen_vl_interface': 8.697973016279722e-06, 'learning_rate/action_model': 8.697973016279722e-05, 'epoch': 1.18}) +[2026-06-15 21:57:22,458][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.033542487770318985, 'train/grad_norm_pre_clip': 1.843336582183838, 'timing/data': 0.020846985280513763, 'timing/model': 1.442523492500186, 'learning_rate/qwen_vl_interface': 8.695877630934827e-06, 'learning_rate/action_model': 8.695877630934826e-05, 'epoch': 1.18}) +[2026-06-15 21:57:28,265][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.15634511411190033, 'train/grad_norm_pre_clip': 5.748503684997559, 'timing/data': 0.021657317876815796, 'timing/model': 1.4639161247760057, 'learning_rate/qwen_vl_interface': 8.693780829119801e-06, 'learning_rate/action_model': 8.693780829119801e-05, 'epoch': 1.18}) +[2026-06-15 21:57:34,054][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.09667535871267319, 'train/grad_norm_pre_clip': 6.645425796508789, 'timing/data': 0.022440623492002487, 'timing/model': 1.43904051464051, 'learning_rate/qwen_vl_interface': 8.69168261169656e-06, 'learning_rate/action_model': 8.691682611696559e-05, 'epoch': 1.18}) +[2026-06-15 21:57:39,869][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.5182796120643616, 'train/grad_norm_pre_clip': 14.61817455291748, 'timing/data': 0.020459813997149467, 'timing/model': 1.4006079826503992, 'learning_rate/qwen_vl_interface': 8.689582979527599e-06, 'learning_rate/action_model': 8.689582979527599e-05, 'epoch': 1.18}) +[2026-06-15 21:57:45,702][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.1496891975402832, 'train/grad_norm_pre_clip': 6.690628528594971, 'timing/data': 0.02044653333723545, 'timing/model': 1.6822741068899632, 'learning_rate/qwen_vl_interface': 8.687481933476e-06, 'learning_rate/action_model': 8.687481933476e-05, 'epoch': 1.18}) +[2026-06-15 21:57:51,355][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.035236768424510956, 'train/grad_norm_pre_clip': 4.189679145812988, 'timing/data': 0.021119416691362858, 'timing/model': 1.4285938888788223, 'learning_rate/qwen_vl_interface': 8.685379474405423e-06, 'learning_rate/action_model': 8.685379474405423e-05, 'epoch': 1.18}) +[2026-06-15 21:57:57,160][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.07490311563014984, 'train/grad_norm_pre_clip': 4.042869567871094, 'timing/data': 0.025195695459842682, 'timing/model': 1.4488521050661802, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 1.18}) +[2026-06-15 21:58:02,954][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.06441949307918549, 'train/grad_norm_pre_clip': 6.746342182159424, 'timing/data': 0.020815405994653702, 'timing/model': 1.4246337553486228, 'learning_rate/qwen_vl_interface': 8.681170320664883e-06, 'learning_rate/action_model': 8.681170320664883e-05, 'epoch': 1.18}) +[2026-06-15 21:58:08,770][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.28362104296684265, 'train/grad_norm_pre_clip': 13.730441093444824, 'timing/data': 0.02471394371241331, 'timing/model': 1.4626176450401545, 'learning_rate/qwen_vl_interface': 8.679063627725143e-06, 'learning_rate/action_model': 8.679063627725144e-05, 'epoch': 1.19}) +[2026-06-15 21:58:14,579][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.08719111979007721, 'train/grad_norm_pre_clip': 5.913979530334473, 'timing/data': 0.02189367078244686, 'timing/model': 1.4008551044389606, 'learning_rate/qwen_vl_interface': 8.676955525226876e-06, 'learning_rate/action_model': 8.676955525226875e-05, 'epoch': 1.19}) +[2026-06-15 21:58:20,399][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.32883501052856445, 'train/grad_norm_pre_clip': 15.653082847595215, 'timing/data': 0.02101711742579937, 'timing/model': 1.6602551592513919, 'learning_rate/qwen_vl_interface': 8.674846014036636e-06, 'learning_rate/action_model': 8.674846014036636e-05, 'epoch': 1.19}) +[2026-06-15 21:58:26,050][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.015043172053992748, 'train/grad_norm_pre_clip': 1.882061243057251, 'timing/data': 0.020879706367850304, 'timing/model': 1.4374562175944448, 'learning_rate/qwen_vl_interface': 8.672735095021569e-06, 'learning_rate/action_model': 8.672735095021569e-05, 'epoch': 1.19}) +[2026-06-15 21:58:31,852][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.038901567459106445, 'train/grad_norm_pre_clip': 5.549141883850098, 'timing/data': 0.020353064872324467, 'timing/model': 1.4398415219038725, 'learning_rate/qwen_vl_interface': 8.670622769049393e-06, 'learning_rate/action_model': 8.670622769049392e-05, 'epoch': 1.19}) +[2026-06-15 21:58:37,658][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.20346754789352417, 'train/grad_norm_pre_clip': 9.594493865966797, 'timing/data': 0.01972931157797575, 'timing/model': 1.4483197638764977, 'learning_rate/qwen_vl_interface': 8.668509036988406e-06, 'learning_rate/action_model': 8.668509036988405e-05, 'epoch': 1.19}) +[2026-06-15 21:58:43,460][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.24609076976776123, 'train/grad_norm_pre_clip': 11.225589752197266, 'timing/data': 0.020419055595993996, 'timing/model': 1.4673875449225307, 'learning_rate/qwen_vl_interface': 8.66639389970748e-06, 'learning_rate/action_model': 8.666393899707479e-05, 'epoch': 1.19}) +[2026-06-15 21:58:49,273][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.19687260687351227, 'train/grad_norm_pre_clip': 17.397445678710938, 'timing/data': 0.02064751647412777, 'timing/model': 1.4052519667893648, 'learning_rate/qwen_vl_interface': 8.66427735807607e-06, 'learning_rate/action_model': 8.664277358076069e-05, 'epoch': 1.19}) +[2026-06-15 22:09:30,573][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.02107020653784275, 'train/grad_norm_pre_clip': 2.3222203254699707, 'eval/loss': 0.10419321805238724, 'eval/flap_precision': 0.8983957219251337, 'eval/flap_recall': 0.610909090909091, 'eval/flap_f1': 0.7272727272727274, 'eval/flap_rate_model': 0.051944444444444446, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7272727272727274, 'timing/data': 0.020856738090515137, 'timing/model': 1.9900395311415195, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 1.19}) +[2026-06-15 22:09:30,573][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2600_state +[2026-06-15 22:09:30,580][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 22:09:38,656][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2600_state/model.safetensors +[2026-06-15 22:09:49,800][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2600_state/optimizer.bin +[2026-06-15 22:09:49,801][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2600_state/sampler.bin +[2026-06-15 22:09:49,801][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2600_state/sampler_1.bin +[2026-06-15 22:09:49,803][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2600_state/random_states_0.pkl +[2026-06-15 22:09:49,973][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2600 +[2026-06-15 22:09:49,973][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 22:09:49,989][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 22:09:55,857][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.09312041103839874, 'train/grad_norm_pre_clip': 11.117679595947266, 'timing/data': 0.0235200934112072, 'timing/model': 1.4528602426871657, 'learning_rate/qwen_vl_interface': 8.660040065242499e-06, 'learning_rate/action_model': 8.660040065242498e-05, 'epoch': 1.19}) +[2026-06-15 22:10:01,690][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.3413297235965729, 'train/grad_norm_pre_clip': 17.56041717529297, 'timing/data': 0.021580862812697887, 'timing/model': 1.6822103448212147, 'learning_rate/qwen_vl_interface': 8.657919315782127e-06, 'learning_rate/action_model': 8.657919315782127e-05, 'epoch': 1.19}) +[2026-06-15 22:10:07,150][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.14474400877952576, 'train/grad_norm_pre_clip': 9.027973175048828, 'timing/data': 0.021332651376724243, 'timing/model': 1.3748283004388213, 'learning_rate/qwen_vl_interface': 8.655797165454853e-06, 'learning_rate/action_model': 8.655797165454853e-05, 'epoch': 1.19}) +[2026-06-15 22:10:12,849][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.21715903282165527, 'train/grad_norm_pre_clip': 11.913167953491211, 'timing/data': 0.02201842423528433, 'timing/model': 1.464004399254918, 'learning_rate/qwen_vl_interface': 8.653673615133013e-06, 'learning_rate/action_model': 8.653673615133013e-05, 'epoch': 1.19}) +[2026-06-15 22:10:18,546][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.19783282279968262, 'train/grad_norm_pre_clip': 10.934589385986328, 'timing/data': 0.021573343314230442, 'timing/model': 1.456899605691433, 'learning_rate/qwen_vl_interface': 8.651548665689519e-06, 'learning_rate/action_model': 8.651548665689518e-05, 'epoch': 1.19}) +[2026-06-15 22:10:24,372][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.4391518533229828, 'train/grad_norm_pre_clip': 18.806705474853516, 'timing/data': 0.022319676354527473, 'timing/model': 1.489960191771388, 'learning_rate/qwen_vl_interface': 8.649422317997855e-06, 'learning_rate/action_model': 8.649422317997856e-05, 'epoch': 1.19}) +[2026-06-15 22:10:30,157][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.2877306640148163, 'train/grad_norm_pre_clip': 7.292023181915283, 'timing/data': 0.02297746017575264, 'timing/model': 1.4468187494203448, 'learning_rate/qwen_vl_interface': 8.647294572932082e-06, 'learning_rate/action_model': 8.647294572932082e-05, 'epoch': 1.19}) +[2026-06-15 22:10:35,876][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.13496017456054688, 'train/grad_norm_pre_clip': 8.054781913757324, 'timing/data': 0.02254461869597435, 'timing/model': 1.6808319175615907, 'learning_rate/qwen_vl_interface': 8.64516543136684e-06, 'learning_rate/action_model': 8.64516543136684e-05, 'epoch': 1.19}) +[2026-06-15 22:10:41,381][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.35110124945640564, 'train/grad_norm_pre_clip': 14.055536270141602, 'timing/data': 0.021888814866542816, 'timing/model': 1.3982627894729376, 'learning_rate/qwen_vl_interface': 8.643034894177335e-06, 'learning_rate/action_model': 8.643034894177335e-05, 'epoch': 1.19}) +[2026-06-15 22:10:47,145][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.10454359650611877, 'train/grad_norm_pre_clip': 7.199268341064453, 'timing/data': 0.021278221160173416, 'timing/model': 1.4439889127388597, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 1.19}) +[2026-06-15 22:10:52,875][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.10049798339605331, 'train/grad_norm_pre_clip': 6.158487319946289, 'timing/data': 0.02172045409679413, 'timing/model': 1.485560804605484, 'learning_rate/qwen_vl_interface': 8.638769636429245e-06, 'learning_rate/action_model': 8.638769636429245e-05, 'epoch': 1.19}) +[2026-06-15 22:10:58,527][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.10283435136079788, 'train/grad_norm_pre_clip': 7.364542484283447, 'timing/data': 0.020537346601486206, 'timing/model': 1.381528101861477, 'learning_rate/qwen_vl_interface': 8.636634917623946e-06, 'learning_rate/action_model': 8.636634917623944e-05, 'epoch': 1.19}) +[2026-06-15 22:11:04,288][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.2235126495361328, 'train/grad_norm_pre_clip': 10.655210494995117, 'timing/data': 0.019309110939502716, 'timing/model': 1.6995803536847234, 'learning_rate/qwen_vl_interface': 8.634498806700955e-06, 'learning_rate/action_model': 8.634498806700954e-05, 'epoch': 1.2}) +[2026-06-15 22:11:09,791][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.10394611954689026, 'train/grad_norm_pre_clip': 4.1138458251953125, 'timing/data': 0.01904245000332594, 'timing/model': 1.408001926727593, 'learning_rate/qwen_vl_interface': 8.63236130453835e-06, 'learning_rate/action_model': 8.63236130453835e-05, 'epoch': 1.2}) +[2026-06-15 22:11:15,589][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.09285120666027069, 'train/grad_norm_pre_clip': 6.196773052215576, 'timing/data': 0.019195579923689365, 'timing/model': 1.4051003232598305, 'learning_rate/qwen_vl_interface': 8.630222412014777e-06, 'learning_rate/action_model': 8.630222412014776e-05, 'epoch': 1.2}) +[2026-06-15 22:11:21,240][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.1922861784696579, 'train/grad_norm_pre_clip': 9.62293815612793, 'timing/data': 0.022293916903436184, 'timing/model': 1.4328339016065001, 'learning_rate/qwen_vl_interface': 8.628082130009451e-06, 'learning_rate/action_model': 8.628082130009451e-05, 'epoch': 1.2}) +[2026-06-15 22:11:26,946][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.11991050839424133, 'train/grad_norm_pre_clip': 5.792729377746582, 'timing/data': 0.019832413643598557, 'timing/model': 1.459248493425548, 'learning_rate/qwen_vl_interface': 8.62594045940216e-06, 'learning_rate/action_model': 8.62594045940216e-05, 'epoch': 1.2}) +[2026-06-15 22:11:32,642][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.19263413548469543, 'train/grad_norm_pre_clip': 7.445380210876465, 'timing/data': 0.020676218904554844, 'timing/model': 1.4515510257333517, 'learning_rate/qwen_vl_interface': 8.623797401073272e-06, 'learning_rate/action_model': 8.623797401073271e-05, 'epoch': 1.2}) +[2026-06-15 22:11:38,384][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.16589900851249695, 'train/grad_norm_pre_clip': 4.165954113006592, 'timing/data': 0.020839798264205456, 'timing/model': 1.6953722005710006, 'learning_rate/qwen_vl_interface': 8.62165295590371e-06, 'learning_rate/action_model': 8.621652955903709e-05, 'epoch': 1.2}) +[2026-06-15 22:11:43,952][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.19425036013126373, 'train/grad_norm_pre_clip': 11.036934852600098, 'timing/data': 0.020778418518602848, 'timing/model': 1.4624075582250953, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 1.2}) +[2026-06-15 22:11:49,650][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.10555046051740646, 'train/grad_norm_pre_clip': 5.064932346343994, 'timing/data': 0.02086263056844473, 'timing/model': 1.4366002045571804, 'learning_rate/qwen_vl_interface': 8.61735990856914e-06, 'learning_rate/action_model': 8.617359908569139e-05, 'epoch': 1.2}) +[2026-06-15 22:11:55,339][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.1049007996916771, 'train/grad_norm_pre_clip': 5.490422248840332, 'timing/data': 0.019749893806874752, 'timing/model': 1.449779017828405, 'learning_rate/qwen_vl_interface': 8.615211308168842e-06, 'learning_rate/action_model': 8.61521130816884e-05, 'epoch': 1.2}) +[2026-06-15 22:12:01,059][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.12303667515516281, 'train/grad_norm_pre_clip': 4.448151588439941, 'timing/data': 0.023687824606895447, 'timing/model': 1.4675681926310062, 'learning_rate/qwen_vl_interface': 8.613061324457289e-06, 'learning_rate/action_model': 8.613061324457288e-05, 'epoch': 1.2}) +[2026-06-15 22:12:06,741][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.17391853034496307, 'train/grad_norm_pre_clip': 5.2213335037231445, 'timing/data': 0.019735253416001797, 'timing/model': 1.4369016345590353, 'learning_rate/qwen_vl_interface': 8.610909958318262e-06, 'learning_rate/action_model': 8.610909958318261e-05, 'epoch': 1.2}) +[2026-06-15 22:12:12,492][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.16375242173671722, 'train/grad_norm_pre_clip': 4.277505874633789, 'timing/data': 0.021775035187602043, 'timing/model': 1.7003059349954128, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 1.2}) +[2026-06-15 22:12:18,039][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.15286308526992798, 'train/grad_norm_pre_clip': 4.212575435638428, 'timing/data': 0.020849510096013546, 'timing/model': 1.4461997682228684, 'learning_rate/qwen_vl_interface': 8.606603082295724e-06, 'learning_rate/action_model': 8.606603082295724e-05, 'epoch': 1.2}) +[2026-06-15 22:12:23,840][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.141752228140831, 'train/grad_norm_pre_clip': 4.077160835266113, 'timing/data': 0.02262455876916647, 'timing/model': 1.448000330477953, 'learning_rate/qwen_vl_interface': 8.60444757418261e-06, 'learning_rate/action_model': 8.604447574182609e-05, 'epoch': 1.2}) +[2026-06-15 22:12:29,552][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.12268739938735962, 'train/grad_norm_pre_clip': 4.323650360107422, 'timing/data': 0.02095905039459467, 'timing/model': 1.4595696032047272, 'learning_rate/qwen_vl_interface': 8.602290687182805e-06, 'learning_rate/action_model': 8.602290687182806e-05, 'epoch': 1.2}) +[2026-06-15 22:12:35,243][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.22917580604553223, 'train/grad_norm_pre_clip': 5.5068159103393555, 'timing/data': 0.020794440060853958, 'timing/model': 1.4104887964203954, 'learning_rate/qwen_vl_interface': 8.600132422182929e-06, 'learning_rate/action_model': 8.600132422182928e-05, 'epoch': 1.2}) +[2026-06-15 22:12:40,951][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.14223864674568176, 'train/grad_norm_pre_clip': 5.387070655822754, 'timing/data': 0.020547498017549515, 'timing/model': 1.4395110923796892, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 1.2}) +[2026-06-15 22:12:46,689][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.21497143805027008, 'train/grad_norm_pre_clip': 9.313252449035645, 'timing/data': 0.020431668497622013, 'timing/model': 1.6831141617149115, 'learning_rate/qwen_vl_interface': 8.595811761732239e-06, 'learning_rate/action_model': 8.595811761732238e-05, 'epoch': 1.2}) +[2026-06-15 22:12:52,252][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.20358888804912567, 'train/grad_norm_pre_clip': 5.100176811218262, 'timing/data': 0.02019877638667822, 'timing/model': 1.4596862141042948, 'learning_rate/qwen_vl_interface': 8.593649368057489e-06, 'learning_rate/action_model': 8.59364936805749e-05, 'epoch': 1.2}) +[2026-06-15 22:12:57,949][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.15725788474082947, 'train/grad_norm_pre_clip': 4.436611175537109, 'timing/data': 0.020251696929335594, 'timing/model': 1.4258152479305863, 'learning_rate/qwen_vl_interface': 8.591485599934786e-06, 'learning_rate/action_model': 8.591485599934785e-05, 'epoch': 1.2}) +[2026-06-15 22:13:03,640][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.1737649291753769, 'train/grad_norm_pre_clip': 5.854199409484863, 'timing/data': 0.02022237703204155, 'timing/model': 1.4448944609612226, 'learning_rate/qwen_vl_interface': 8.58932045825357e-06, 'learning_rate/action_model': 8.589320458253569e-05, 'epoch': 1.2}) +[2026-06-15 22:13:09,356][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.10103310644626617, 'train/grad_norm_pre_clip': 4.40703821182251, 'timing/data': 0.020729029551148415, 'timing/model': 1.4608707865700126, 'learning_rate/qwen_vl_interface': 8.58715394390385e-06, 'learning_rate/action_model': 8.58715394390385e-05, 'epoch': 1.21}) +[2026-06-15 22:13:15,056][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.12500043213367462, 'train/grad_norm_pre_clip': 7.896627426147461, 'timing/data': 0.020117616280913353, 'timing/model': 1.393629933707416, 'learning_rate/qwen_vl_interface': 8.5849860577762e-06, 'learning_rate/action_model': 8.5849860577762e-05, 'epoch': 1.21}) +[2026-06-15 22:13:20,903][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.14063221216201782, 'train/grad_norm_pre_clip': 6.530995845794678, 'timing/data': 0.020538759417831898, 'timing/model': 1.711569108068943, 'learning_rate/qwen_vl_interface': 8.582816800761756e-06, 'learning_rate/action_model': 8.582816800761755e-05, 'epoch': 1.21}) +[2026-06-15 22:13:26,450][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.12614835798740387, 'train/grad_norm_pre_clip': 6.319277286529541, 'timing/data': 0.02087945118546486, 'timing/model': 1.455948487855494, 'learning_rate/qwen_vl_interface': 8.580646173752214e-06, 'learning_rate/action_model': 8.580646173752213e-05, 'epoch': 1.21}) +[2026-06-15 22:13:32,255][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.13836264610290527, 'train/grad_norm_pre_clip': 7.547168731689453, 'timing/data': 0.023106902837753296, 'timing/model': 1.4574257172644138, 'learning_rate/qwen_vl_interface': 8.578474177639841e-06, 'learning_rate/action_model': 8.578474177639842e-05, 'epoch': 1.21}) +[2026-06-15 22:13:37,961][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.22129464149475098, 'train/grad_norm_pre_clip': 7.551712989807129, 'timing/data': 0.020539297722280025, 'timing/model': 1.4658777052536607, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 1.21}) +[2026-06-15 22:13:43,756][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.11460814625024796, 'train/grad_norm_pre_clip': 6.142587661743164, 'timing/data': 0.024971062317490578, 'timing/model': 1.4562903363257647, 'learning_rate/qwen_vl_interface': 8.574126081678463e-06, 'learning_rate/action_model': 8.574126081678462e-05, 'epoch': 1.21}) +[2026-06-15 22:13:49,549][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.0988604798913002, 'train/grad_norm_pre_clip': 6.385799407958984, 'timing/data': 0.02110258210450411, 'timing/model': 1.4237642949447036, 'learning_rate/qwen_vl_interface': 8.571949983616796e-06, 'learning_rate/action_model': 8.571949983616795e-05, 'epoch': 1.21}) +[2026-06-15 22:13:55,310][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.05236007273197174, 'train/grad_norm_pre_clip': 3.9276349544525146, 'timing/data': 0.020734239369630814, 'timing/model': 1.7032663989812136, 'learning_rate/qwen_vl_interface': 8.569772520026971e-06, 'learning_rate/action_model': 8.569772520026971e-05, 'epoch': 1.21}) +[2026-06-15 22:14:00,845][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.32908719778060913, 'train/grad_norm_pre_clip': 15.44036865234375, 'timing/data': 0.021016770973801613, 'timing/model': 1.4544337429106236, 'learning_rate/qwen_vl_interface': 8.567593691804061e-06, 'learning_rate/action_model': 8.567593691804061e-05, 'epoch': 1.21}) +[2026-06-15 22:14:06,545][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.138214111328125, 'train/grad_norm_pre_clip': 8.293724060058594, 'timing/data': 0.020213968120515347, 'timing/model': 1.4236441608518362, 'learning_rate/qwen_vl_interface': 8.565413499843703e-06, 'learning_rate/action_model': 8.565413499843703e-05, 'epoch': 1.21}) +[2026-06-15 22:14:12,358][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.6303712725639343, 'train/grad_norm_pre_clip': 15.365365028381348, 'timing/data': 0.021217471919953823, 'timing/model': 1.467051524668932, 'learning_rate/qwen_vl_interface': 8.563231945042087e-06, 'learning_rate/action_model': 8.563231945042088e-05, 'epoch': 1.21}) +[2026-06-15 22:14:18,142][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.1356436163187027, 'train/grad_norm_pre_clip': 14.272335052490234, 'timing/data': 0.021566953510046005, 'timing/model': 1.4498189128935337, 'learning_rate/qwen_vl_interface': 8.56104902829597e-06, 'learning_rate/action_model': 8.56104902829597e-05, 'epoch': 1.21}) +[2026-06-15 22:14:23,841][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.31913191080093384, 'train/grad_norm_pre_clip': 10.284195899963379, 'timing/data': 0.02115188166499138, 'timing/model': 1.4054577611386776, 'learning_rate/qwen_vl_interface': 8.558864750502669e-06, 'learning_rate/action_model': 8.558864750502668e-05, 'epoch': 1.21}) +[2026-06-15 22:14:29,599][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.20949244499206543, 'train/grad_norm_pre_clip': 9.91152572631836, 'timing/data': 0.020194548182189465, 'timing/model': 1.7134208474308252, 'learning_rate/qwen_vl_interface': 8.556679112560055e-06, 'learning_rate/action_model': 8.556679112560054e-05, 'epoch': 1.21}) +[2026-06-15 22:14:35,145][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.1495310664176941, 'train/grad_norm_pre_clip': 21.882587432861328, 'timing/data': 0.020391038618981838, 'timing/model': 1.4517662301659584, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 1.21}) +[2026-06-15 22:14:40,849][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.08728531748056412, 'train/grad_norm_pre_clip': 6.666942596435547, 'timing/data': 0.020205878652632236, 'timing/model': 1.4418956190347672, 'learning_rate/qwen_vl_interface': 8.55230375982118e-06, 'learning_rate/action_model': 8.55230375982118e-05, 'epoch': 1.21}) +[2026-06-15 22:14:46,548][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.10100885480642319, 'train/grad_norm_pre_clip': 5.856376647949219, 'timing/data': 0.02227528765797615, 'timing/model': 1.438948506489396, 'learning_rate/qwen_vl_interface': 8.550114046823463e-06, 'learning_rate/action_model': 8.550114046823463e-05, 'epoch': 1.21}) +[2026-06-15 22:14:52,243][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.14861904084682465, 'train/grad_norm_pre_clip': 8.220722198486328, 'timing/data': 0.020558789372444153, 'timing/model': 1.451367792673409, 'learning_rate/qwen_vl_interface': 8.547922977273518e-06, 'learning_rate/action_model': 8.547922977273518e-05, 'epoch': 1.21}) +[2026-06-15 22:14:58,046][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.17570511996746063, 'train/grad_norm_pre_clip': 10.703349113464355, 'timing/data': 0.02021714672446251, 'timing/model': 1.4052629005163908, 'learning_rate/qwen_vl_interface': 8.54573055207201e-06, 'learning_rate/action_model': 8.545730552072008e-05, 'epoch': 1.21}) +[2026-06-15 22:15:03,888][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.14201626181602478, 'train/grad_norm_pre_clip': 12.99239444732666, 'timing/data': 0.020007606595754623, 'timing/model': 1.6923781372606754, 'learning_rate/qwen_vl_interface': 8.543536772120161e-06, 'learning_rate/action_model': 8.54353677212016e-05, 'epoch': 1.21}) +[2026-06-15 22:15:09,628][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.1233709380030632, 'train/grad_norm_pre_clip': 5.202268123626709, 'timing/data': 0.02071016002446413, 'timing/model': 1.6083813942968845, 'learning_rate/qwen_vl_interface': 8.541341638319754e-06, 'learning_rate/action_model': 8.541341638319754e-05, 'epoch': 1.21}) +[2026-06-15 22:15:15,254][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.1401202529668808, 'train/grad_norm_pre_clip': 4.375125885009766, 'timing/data': 0.0203548576682806, 'timing/model': 1.4515998605638742, 'learning_rate/qwen_vl_interface': 8.539145151573126e-06, 'learning_rate/action_model': 8.539145151573125e-05, 'epoch': 1.22}) +[2026-06-15 22:15:20,952][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.19142484664916992, 'train/grad_norm_pre_clip': 6.641480445861816, 'timing/data': 0.02480362169444561, 'timing/model': 1.4163851365447044, 'learning_rate/qwen_vl_interface': 8.536947312783167e-06, 'learning_rate/action_model': 8.536947312783166e-05, 'epoch': 1.22}) +[2026-06-15 22:15:26,763][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.287238746881485, 'train/grad_norm_pre_clip': 9.740547180175781, 'timing/data': 0.020700260065495968, 'timing/model': 1.4680813634768128, 'learning_rate/qwen_vl_interface': 8.534748122853326e-06, 'learning_rate/action_model': 8.534748122853325e-05, 'epoch': 1.22}) +[2026-06-15 22:15:32,547][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.12342534214258194, 'train/grad_norm_pre_clip': 5.653137683868408, 'timing/data': 0.020735771395266056, 'timing/model': 1.4059006739407778, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 1.22}) +[2026-06-15 22:15:38,316][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.14141680300235748, 'train/grad_norm_pre_clip': 4.676832675933838, 'timing/data': 0.020561059936881065, 'timing/model': 1.66412394028157, 'learning_rate/qwen_vl_interface': 8.530345693190573e-06, 'learning_rate/action_model': 8.530345693190572e-05, 'epoch': 1.22}) +[2026-06-15 22:15:44,214][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.042585283517837524, 'train/grad_norm_pre_clip': 2.5811169147491455, 'timing/data': 0.021431583911180496, 'timing/model': 1.698421855457127, 'learning_rate/qwen_vl_interface': 8.528142455267328e-06, 'learning_rate/action_model': 8.528142455267329e-05, 'epoch': 1.22}) +[2026-06-15 22:15:49,862][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.12472475320100784, 'train/grad_norm_pre_clip': 5.351677417755127, 'timing/data': 0.024241039529442787, 'timing/model': 1.4569452796131372, 'learning_rate/qwen_vl_interface': 8.52593786982355e-06, 'learning_rate/action_model': 8.525937869823551e-05, 'epoch': 1.22}) +[2026-06-15 22:15:55,656][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.16477057337760925, 'train/grad_norm_pre_clip': 3.903034210205078, 'timing/data': 0.021885177120566368, 'timing/model': 1.4374687680974603, 'learning_rate/qwen_vl_interface': 8.523731937765456e-06, 'learning_rate/action_model': 8.523731937765456e-05, 'epoch': 1.22}) +[2026-06-15 22:16:01,462][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.24324460327625275, 'train/grad_norm_pre_clip': 7.14089298248291, 'timing/data': 0.020924411714076996, 'timing/model': 1.4567143823951483, 'learning_rate/qwen_vl_interface': 8.521524659999823e-06, 'learning_rate/action_model': 8.521524659999822e-05, 'epoch': 1.22}) +[2026-06-15 22:16:07,263][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.07836591452360153, 'train/grad_norm_pre_clip': 3.375642776489258, 'timing/data': 0.020870281383395195, 'timing/model': 1.4050961891189218, 'learning_rate/qwen_vl_interface': 8.519316037433976e-06, 'learning_rate/action_model': 8.519316037433977e-05, 'epoch': 1.22}) +[2026-06-15 22:16:13,107][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.16078488528728485, 'train/grad_norm_pre_clip': 9.567514419555664, 'timing/data': 0.020430728793144226, 'timing/model': 1.686575341038406, 'learning_rate/qwen_vl_interface': 8.517106070975801e-06, 'learning_rate/action_model': 8.5171060709758e-05, 'epoch': 1.22}) +[2026-06-15 22:16:18,756][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.09319709241390228, 'train/grad_norm_pre_clip': 4.969329833984375, 'timing/data': 0.021912106312811375, 'timing/model': 1.452879006974399, 'learning_rate/qwen_vl_interface': 8.514894761533729e-06, 'learning_rate/action_model': 8.514894761533728e-05, 'epoch': 1.22}) +[2026-06-15 22:16:24,560][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.11466170102357864, 'train/grad_norm_pre_clip': 8.064421653747559, 'timing/data': 0.024777723476290703, 'timing/model': 1.439252456650138, 'learning_rate/qwen_vl_interface': 8.512682110016747e-06, 'learning_rate/action_model': 8.512682110016747e-05, 'epoch': 1.22}) +[2026-06-15 22:16:30,370][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.09123058617115021, 'train/grad_norm_pre_clip': 3.951388120651245, 'timing/data': 0.02445298060774803, 'timing/model': 1.4647323237732053, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 1.22}) +[2026-06-15 22:16:36,157][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.17135021090507507, 'train/grad_norm_pre_clip': 7.424062728881836, 'timing/data': 0.02067593950778246, 'timing/model': 1.4019260602071881, 'learning_rate/qwen_vl_interface': 8.50825278439675e-06, 'learning_rate/action_model': 8.508252784396749e-05, 'epoch': 1.22}) +[2026-06-15 22:16:41,956][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.1278124451637268, 'train/grad_norm_pre_clip': 3.937387466430664, 'timing/data': 0.021252294071018696, 'timing/model': 1.4591377060860395, 'learning_rate/qwen_vl_interface': 8.506036112114465e-06, 'learning_rate/action_model': 8.506036112114464e-05, 'epoch': 1.22}) +[2026-06-15 22:16:47,758][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.0834612026810646, 'train/grad_norm_pre_clip': 3.8253867626190186, 'timing/data': 0.021075152792036533, 'timing/model': 1.6436960333958268, 'learning_rate/qwen_vl_interface': 8.503818101398725e-06, 'learning_rate/action_model': 8.503818101398725e-05, 'epoch': 1.22}) +[2026-06-15 22:16:53,357][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.06249728798866272, 'train/grad_norm_pre_clip': 5.965568542480469, 'timing/data': 0.02079952135682106, 'timing/model': 1.451714400202036, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 1.22}) +[2026-06-15 22:16:59,161][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.06212259456515312, 'train/grad_norm_pre_clip': 5.263323783874512, 'timing/data': 0.02130031306296587, 'timing/model': 1.4425278129056096, 'learning_rate/qwen_vl_interface': 8.499378068314395e-06, 'learning_rate/action_model': 8.499378068314395e-05, 'epoch': 1.22}) +[2026-06-15 22:17:04,955][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.1615789383649826, 'train/grad_norm_pre_clip': 8.22524356842041, 'timing/data': 0.0234377458691597, 'timing/model': 1.4446755247190595, 'learning_rate/qwen_vl_interface': 8.497156047770932e-06, 'learning_rate/action_model': 8.497156047770933e-05, 'epoch': 1.22}) +[2026-06-15 22:17:10,758][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.11637325584888458, 'train/grad_norm_pre_clip': 5.5206499099731445, 'timing/data': 0.02335518505424261, 'timing/model': 1.3999858377501369, 'learning_rate/qwen_vl_interface': 8.494932692444276e-06, 'learning_rate/action_model': 8.494932692444276e-05, 'epoch': 1.22}) +[2026-06-15 22:17:16,559][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.04862579330801964, 'train/grad_norm_pre_clip': 4.2192840576171875, 'timing/data': 0.02112097293138504, 'timing/model': 1.401713628321886, 'learning_rate/qwen_vl_interface': 8.492708003248362e-06, 'learning_rate/action_model': 8.492708003248362e-05, 'epoch': 1.22}) +[2026-06-15 22:17:22,370][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.034670569002628326, 'train/grad_norm_pre_clip': 4.188704967498779, 'timing/data': 0.02101106196641922, 'timing/model': 1.6616892199963331, 'learning_rate/qwen_vl_interface': 8.490481981097674e-06, 'learning_rate/action_model': 8.490481981097674e-05, 'epoch': 1.23}) +[2026-06-15 22:17:27,952][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.19448311626911163, 'train/grad_norm_pre_clip': 12.03784465789795, 'timing/data': 0.020416198298335075, 'timing/model': 1.4512087060138583, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 1.23}) +[2026-06-15 22:17:33,756][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.03653223440051079, 'train/grad_norm_pre_clip': 3.6282126903533936, 'timing/data': 0.02156770508736372, 'timing/model': 1.445201065391302, 'learning_rate/qwen_vl_interface': 8.486025941592664e-06, 'learning_rate/action_model': 8.486025941592664e-05, 'epoch': 1.23}) +[2026-06-15 22:17:39,556][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.2903723418712616, 'train/grad_norm_pre_clip': 19.524307250976562, 'timing/data': 0.021079222671687603, 'timing/model': 1.4502795934677124, 'learning_rate/qwen_vl_interface': 8.483795926070054e-06, 'learning_rate/action_model': 8.483795926070052e-05, 'epoch': 1.23}) +[2026-06-15 22:17:45,367][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.12263515591621399, 'train/grad_norm_pre_clip': 23.429443359375, 'timing/data': 0.02209800761193037, 'timing/model': 1.4038017960265279, 'learning_rate/qwen_vl_interface': 8.481564581256087e-06, 'learning_rate/action_model': 8.481564581256087e-05, 'epoch': 1.23}) +[2026-06-15 22:17:51,215][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.09545811265707016, 'train/grad_norm_pre_clip': 10.530858039855957, 'timing/data': 0.021217843517661095, 'timing/model': 1.7074099918827415, 'learning_rate/qwen_vl_interface': 8.479331908067987e-06, 'learning_rate/action_model': 8.479331908067987e-05, 'epoch': 1.23}) +[2026-06-15 22:17:56,855][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.25017568469047546, 'train/grad_norm_pre_clip': 22.513198852539062, 'timing/data': 0.021955758333206177, 'timing/model': 1.4419591035693884, 'learning_rate/qwen_vl_interface': 8.477097907423524e-06, 'learning_rate/action_model': 8.477097907423524e-05, 'epoch': 1.23}) +[2026-06-15 22:18:02,655][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.07666575163602829, 'train/grad_norm_pre_clip': 9.317923545837402, 'timing/data': 0.022607801482081413, 'timing/model': 1.4281586799770594, 'learning_rate/qwen_vl_interface': 8.474862580241006e-06, 'learning_rate/action_model': 8.474862580241006e-05, 'epoch': 1.23}) +[2026-06-15 22:18:08,354][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.15807002782821655, 'train/grad_norm_pre_clip': 19.150041580200195, 'timing/data': 0.021422095596790314, 'timing/model': 1.4534801170229912, 'learning_rate/qwen_vl_interface': 8.472625927439299e-06, 'learning_rate/action_model': 8.472625927439299e-05, 'epoch': 1.23}) +[2026-06-15 22:18:14,157][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.0815470963716507, 'train/grad_norm_pre_clip': 8.428132057189941, 'timing/data': 0.021200533024966717, 'timing/model': 1.4385747890919447, 'learning_rate/qwen_vl_interface': 8.4703879499378e-06, 'learning_rate/action_model': 8.4703879499378e-05, 'epoch': 1.23}) +[2026-06-15 22:18:19,953][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.2062956839799881, 'train/grad_norm_pre_clip': 10.835309028625488, 'timing/data': 0.021748026832938194, 'timing/model': 1.4132718136534095, 'learning_rate/qwen_vl_interface': 8.46814864865646e-06, 'learning_rate/action_model': 8.46814864865646e-05, 'epoch': 1.23}) +[2026-06-15 22:18:25,805][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.6384820938110352, 'train/grad_norm_pre_clip': 21.534730911254883, 'timing/data': 0.021732567809522152, 'timing/model': 1.6967034032568336, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 1.23}) +[2026-06-15 22:18:31,459][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.09156793355941772, 'train/grad_norm_pre_clip': 5.9069414138793945, 'timing/data': 0.02287658117711544, 'timing/model': 1.4389493251219392, 'learning_rate/qwen_vl_interface': 8.463666078436769e-06, 'learning_rate/action_model': 8.463666078436769e-05, 'epoch': 1.23}) +[2026-06-15 22:18:37,260][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.572513222694397, 'train/grad_norm_pre_clip': 17.69550132751465, 'timing/data': 0.021257594227790833, 'timing/model': 1.4549389323219657, 'learning_rate/qwen_vl_interface': 8.46142281134103e-06, 'learning_rate/action_model': 8.461422811341031e-05, 'epoch': 1.23}) +[2026-06-15 22:18:43,056][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.28374430537223816, 'train/grad_norm_pre_clip': 6.48301362991333, 'timing/data': 0.020926091820001602, 'timing/model': 1.45351105555892, 'learning_rate/qwen_vl_interface': 8.459178224150683e-06, 'learning_rate/action_model': 8.459178224150683e-05, 'epoch': 1.23}) +[2026-06-15 22:18:48,860][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.11252736300230026, 'train/grad_norm_pre_clip': 9.591919898986816, 'timing/data': 0.021721857599914074, 'timing/model': 1.4279265087097883, 'learning_rate/qwen_vl_interface': 8.456932317788389e-06, 'learning_rate/action_model': 8.456932317788389e-05, 'epoch': 1.23}) +[2026-06-15 22:18:54,655][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.08575249463319778, 'train/grad_norm_pre_clip': 5.305508136749268, 'timing/data': 0.02163283620029688, 'timing/model': 1.4182517677545547, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 1.23}) +[2026-06-15 22:19:00,507][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.07915443927049637, 'train/grad_norm_pre_clip': 4.71200704574585, 'timing/data': 0.02143489569425583, 'timing/model': 1.684674276970327, 'learning_rate/qwen_vl_interface': 8.452436551241328e-06, 'learning_rate/action_model': 8.452436551241328e-05, 'epoch': 1.23}) +[2026-06-15 22:19:06,158][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.2562714219093323, 'train/grad_norm_pre_clip': 6.740198135375977, 'timing/data': 0.02103989291936159, 'timing/model': 1.44592637848109, 'learning_rate/qwen_vl_interface': 8.450186692904605e-06, 'learning_rate/action_model': 8.450186692904606e-05, 'epoch': 1.23}) +[2026-06-15 22:19:11,956][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.2855377495288849, 'train/grad_norm_pre_clip': 13.449694633483887, 'timing/data': 0.021593806333839893, 'timing/model': 1.453314839862287, 'learning_rate/qwen_vl_interface': 8.447935519092014e-06, 'learning_rate/action_model': 8.447935519092014e-05, 'epoch': 1.23}) +[2026-06-15 22:19:17,759][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.18178121745586395, 'train/grad_norm_pre_clip': 11.640645980834961, 'timing/data': 0.021767347119748592, 'timing/model': 1.444251962006092, 'learning_rate/qwen_vl_interface': 8.445683030728925e-06, 'learning_rate/action_model': 8.445683030728925e-05, 'epoch': 1.23}) +[2026-06-15 22:29:53,279][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.11786423623561859, 'train/grad_norm_pre_clip': 4.20342493057251, 'eval/loss': 0.08067945390939713, 'eval/flap_precision': 0.7263843648208469, 'eval/flap_recall': 0.8109090909090909, 'eval/flap_f1': 0.7663230240549828, 'eval/flap_rate_model': 0.08527777777777777, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7663230240549829, 'timing/data': 0.02191324718296528, 'timing/model': 1.4077289775013924, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 1.23}) +[2026-06-15 22:29:53,280][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2700_state +[2026-06-15 22:29:53,286][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 22:30:00,398][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2700_state/model.safetensors +[2026-06-15 22:30:11,551][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2700_state/optimizer.bin +[2026-06-15 22:30:11,551][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2700_state/sampler.bin +[2026-06-15 22:30:11,552][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2700_state/sampler_1.bin +[2026-06-15 22:30:11,553][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2700_state/random_states_0.pkl +[2026-06-15 22:30:11,706][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2700 +[2026-06-15 22:30:11,707][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 22:30:11,721][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 22:30:17,350][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.1696172058582306, 'train/grad_norm_pre_clip': 6.163973808288574, 'timing/data': 0.02229977957904339, 'timing/model': 1.3718174817040563, 'learning_rate/qwen_vl_interface': 8.441174114055457e-06, 'learning_rate/action_model': 8.441174114055456e-05, 'epoch': 1.24}) +[2026-06-15 22:30:22,927][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.11311714351177216, 'train/grad_norm_pre_clip': 9.614754676818848, 'timing/data': 0.02046372927725315, 'timing/model': 1.3911084057763219, 'learning_rate/qwen_vl_interface': 8.43891768759852e-06, 'learning_rate/action_model': 8.43891768759852e-05, 'epoch': 1.24}) +[2026-06-15 22:30:28,565][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.1652519851922989, 'train/grad_norm_pre_clip': 6.543802261352539, 'timing/data': 0.020985864102840424, 'timing/model': 1.663860752247274, 'learning_rate/qwen_vl_interface': 8.436659950297978e-06, 'learning_rate/action_model': 8.436659950297977e-05, 'epoch': 1.24}) +[2026-06-15 22:30:34,221][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.18024417757987976, 'train/grad_norm_pre_clip': 8.454785346984863, 'timing/data': 0.020302940160036087, 'timing/model': 1.6354492185637355, 'learning_rate/qwen_vl_interface': 8.4344009030819e-06, 'learning_rate/action_model': 8.4344009030819e-05, 'epoch': 1.24}) +[2026-06-15 22:30:39,838][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.20579032599925995, 'train/grad_norm_pre_clip': 5.465775012969971, 'timing/data': 0.023507798090577126, 'timing/model': 1.447100148536265, 'learning_rate/qwen_vl_interface': 8.432140546878899e-06, 'learning_rate/action_model': 8.432140546878898e-05, 'epoch': 1.24}) +[2026-06-15 22:30:45,556][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.13972294330596924, 'train/grad_norm_pre_clip': 5.949021816253662, 'timing/data': 0.020514270290732384, 'timing/model': 1.4589549340307713, 'learning_rate/qwen_vl_interface': 8.429878882618116e-06, 'learning_rate/action_model': 8.429878882618116e-05, 'epoch': 1.24}) +[2026-06-15 22:30:51,335][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.14080122113227844, 'train/grad_norm_pre_clip': 3.858520746231079, 'timing/data': 0.020737862214446068, 'timing/model': 1.3806755216792226, 'learning_rate/qwen_vl_interface': 8.42761591122924e-06, 'learning_rate/action_model': 8.427615911229241e-05, 'epoch': 1.24}) +[2026-06-15 22:30:57,050][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.09971307963132858, 'train/grad_norm_pre_clip': 5.647592544555664, 'timing/data': 0.022376351058483124, 'timing/model': 1.4183749267831445, 'learning_rate/qwen_vl_interface': 8.425351633642491e-06, 'learning_rate/action_model': 8.425351633642492e-05, 'epoch': 1.24}) +[2026-06-15 22:31:02,787][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.1532633900642395, 'train/grad_norm_pre_clip': 6.296291351318359, 'timing/data': 0.023312675766646862, 'timing/model': 1.6830132119357586, 'learning_rate/qwen_vl_interface': 8.423086050788632e-06, 'learning_rate/action_model': 8.423086050788631e-05, 'epoch': 1.24}) +[2026-06-15 22:31:08,535][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.1438521444797516, 'train/grad_norm_pre_clip': 6.753363132476807, 'timing/data': 0.021076283417642117, 'timing/model': 1.621270895935595, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 1.24}) +[2026-06-15 22:31:14,036][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.13137772679328918, 'train/grad_norm_pre_clip': 3.8856828212738037, 'timing/data': 0.020594848319888115, 'timing/model': 1.4350152825936675, 'learning_rate/qwen_vl_interface': 8.418550973005289e-06, 'learning_rate/action_model': 8.418550973005289e-05, 'epoch': 1.24}) +[2026-06-15 22:31:19,841][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.14760562777519226, 'train/grad_norm_pre_clip': 5.7327470779418945, 'timing/data': 0.020699217915534973, 'timing/model': 1.4411569684743881, 'learning_rate/qwen_vl_interface': 8.416281479940005e-06, 'learning_rate/action_model': 8.416281479940005e-05, 'epoch': 1.24}) +[2026-06-15 22:31:25,541][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.055773522704839706, 'train/grad_norm_pre_clip': 2.6076366901397705, 'timing/data': 0.02182742953300476, 'timing/model': 1.4510462414473295, 'learning_rate/qwen_vl_interface': 8.414010685336008e-06, 'learning_rate/action_model': 8.414010685336008e-05, 'epoch': 1.24}) +[2026-06-15 22:31:31,240][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.054594121873378754, 'train/grad_norm_pre_clip': 3.555307388305664, 'timing/data': 0.02084462810307741, 'timing/model': 1.4016202446073294, 'learning_rate/qwen_vl_interface': 8.411738590126733e-06, 'learning_rate/action_model': 8.411738590126732e-05, 'epoch': 1.24}) +[2026-06-15 22:31:36,982][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.09374170750379562, 'train/grad_norm_pre_clip': 6.476070880889893, 'timing/data': 0.020529803819954395, 'timing/model': 1.6857606712728739, 'learning_rate/qwen_vl_interface': 8.409465195246152e-06, 'learning_rate/action_model': 8.409465195246152e-05, 'epoch': 1.24}) +[2026-06-15 22:31:42,797][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.34302324056625366, 'train/grad_norm_pre_clip': 15.912772178649902, 'timing/data': 0.0210508881136775, 'timing/model': 1.698171486146748, 'learning_rate/qwen_vl_interface': 8.407190501628772e-06, 'learning_rate/action_model': 8.407190501628772e-05, 'epoch': 1.24}) +[2026-06-15 22:31:48,346][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.3064882457256317, 'train/grad_norm_pre_clip': 8.452906608581543, 'timing/data': 0.02181941457092762, 'timing/model': 1.4469000436365604, 'learning_rate/qwen_vl_interface': 8.404914510209635e-06, 'learning_rate/action_model': 8.404914510209635e-05, 'epoch': 1.24}) +[2026-06-15 22:31:54,142][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.13995777070522308, 'train/grad_norm_pre_clip': 9.67684268951416, 'timing/data': 0.02103324607014656, 'timing/model': 1.434941885061562, 'learning_rate/qwen_vl_interface': 8.402637221924315e-06, 'learning_rate/action_model': 8.402637221924314e-05, 'epoch': 1.24}) +[2026-06-15 22:32:00,642][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.2421281486749649, 'train/grad_norm_pre_clip': 10.771739959716797, 'timing/data': 0.026113693602383137, 'timing/model': 1.4286872539669275, 'learning_rate/qwen_vl_interface': 8.400358637708919e-06, 'learning_rate/action_model': 8.400358637708918e-05, 'epoch': 1.24}) +[2026-06-15 22:32:06,346][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.05070939660072327, 'train/grad_norm_pre_clip': 4.200649261474609, 'timing/data': 0.020773621276021004, 'timing/model': 1.396722904406488, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 1.24}) +[2026-06-15 22:32:12,190][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.28090113401412964, 'train/grad_norm_pre_clip': 22.88755989074707, 'timing/data': 0.020642979070544243, 'timing/model': 1.6970329945906997, 'learning_rate/qwen_vl_interface': 8.395797585234987e-06, 'learning_rate/action_model': 8.395797585234988e-05, 'epoch': 1.24}) +[2026-06-15 22:32:17,749][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.024434147402644157, 'train/grad_norm_pre_clip': 1.334639072418213, 'timing/data': 0.020237814635038376, 'timing/model': 1.4501372929662466, 'learning_rate/qwen_vl_interface': 8.39351511885133e-06, 'learning_rate/action_model': 8.393515118851329e-05, 'epoch': 1.24}) +[2026-06-15 22:32:23,550][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.13648098707199097, 'train/grad_norm_pre_clip': 7.999253273010254, 'timing/data': 0.02028766367584467, 'timing/model': 1.4571502804756165, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 1.25}) +[2026-06-15 22:32:29,350][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.2667411267757416, 'train/grad_norm_pre_clip': 14.58446979522705, 'timing/data': 0.021699337288737297, 'timing/model': 1.451542217284441, 'learning_rate/qwen_vl_interface': 8.388946310481807e-06, 'learning_rate/action_model': 8.388946310481807e-05, 'epoch': 1.25}) +[2026-06-15 22:32:35,150][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.16598264873027802, 'train/grad_norm_pre_clip': 6.778314113616943, 'timing/data': 0.021489163860678673, 'timing/model': 1.4043399905785918, 'learning_rate/qwen_vl_interface': 8.386659970374007e-06, 'learning_rate/action_model': 8.386659970374006e-05, 'epoch': 1.25}) +[2026-06-15 22:32:40,869][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.1601143777370453, 'train/grad_norm_pre_clip': 6.225214004516602, 'timing/data': 0.02052224427461624, 'timing/model': 1.394860957749188, 'learning_rate/qwen_vl_interface': 8.384372340903775e-06, 'learning_rate/action_model': 8.384372340903775e-05, 'epoch': 1.25}) +[2026-06-15 22:32:46,648][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.1049635261297226, 'train/grad_norm_pre_clip': 4.147340297698975, 'timing/data': 0.022042018361389637, 'timing/model': 1.6353181852027774, 'learning_rate/qwen_vl_interface': 8.382083423011472e-06, 'learning_rate/action_model': 8.382083423011471e-05, 'epoch': 1.25}) +[2026-06-15 22:32:52,137][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.4184541404247284, 'train/grad_norm_pre_clip': 15.232976913452148, 'timing/data': 0.020145549438893795, 'timing/model': 1.4314014231786132, 'learning_rate/qwen_vl_interface': 8.379793217637979e-06, 'learning_rate/action_model': 8.379793217637979e-05, 'epoch': 1.25}) +[2026-06-15 22:32:57,846][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.12934792041778564, 'train/grad_norm_pre_clip': 6.9531354904174805, 'timing/data': 0.021966115571558475, 'timing/model': 1.4273460423573852, 'learning_rate/qwen_vl_interface': 8.377501725724719e-06, 'learning_rate/action_model': 8.377501725724719e-05, 'epoch': 1.25}) +[2026-06-15 22:33:03,650][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.2987373471260071, 'train/grad_norm_pre_clip': 14.802857398986816, 'timing/data': 0.022186446003615856, 'timing/model': 1.4468506071716547, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 1.25}) +[2026-06-15 22:33:09,450][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.10508307814598083, 'train/grad_norm_pre_clip': 5.758841514587402, 'timing/data': 0.020275847986340523, 'timing/model': 1.4121278505772352, 'learning_rate/qwen_vl_interface': 8.372914886047203e-06, 'learning_rate/action_model': 8.372914886047203e-05, 'epoch': 1.25}) +[2026-06-15 22:33:15,242][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.1571163833141327, 'train/grad_norm_pre_clip': 8.00253963470459, 'timing/data': 0.0197124220430851, 'timing/model': 1.4026522114872932, 'learning_rate/qwen_vl_interface': 8.370619540168422e-06, 'learning_rate/action_model': 8.370619540168423e-05, 'epoch': 1.25}) +[2026-06-15 22:33:21,059][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.26719892024993896, 'train/grad_norm_pre_clip': 8.23486614227295, 'timing/data': 0.01977407094091177, 'timing/model': 1.6517891408875585, 'learning_rate/qwen_vl_interface': 8.368322911520826e-06, 'learning_rate/action_model': 8.368322911520826e-05, 'epoch': 1.25}) +[2026-06-15 22:33:26,658][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.09097470343112946, 'train/grad_norm_pre_clip': 4.306643486022949, 'timing/data': 0.02051874715834856, 'timing/model': 1.4653498651459813, 'learning_rate/qwen_vl_interface': 8.36602500104847e-06, 'learning_rate/action_model': 8.36602500104847e-05, 'epoch': 1.25}) +[2026-06-15 22:33:32,458][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.30879640579223633, 'train/grad_norm_pre_clip': 12.284521102905273, 'timing/data': 0.020779840648174286, 'timing/model': 1.4678042437881231, 'learning_rate/qwen_vl_interface': 8.36372580969594e-06, 'learning_rate/action_model': 8.36372580969594e-05, 'epoch': 1.25}) +[2026-06-15 22:33:38,264][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.21707026660442352, 'train/grad_norm_pre_clip': 9.5615234375, 'timing/data': 0.02045108564198017, 'timing/model': 1.4649354862049222, 'learning_rate/qwen_vl_interface': 8.361425338408344e-06, 'learning_rate/action_model': 8.361425338408343e-05, 'epoch': 1.25}) +[2026-06-15 22:33:44,048][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.12069433182477951, 'train/grad_norm_pre_clip': 6.399173259735107, 'timing/data': 0.02126759197562933, 'timing/model': 1.4247643919661641, 'learning_rate/qwen_vl_interface': 8.359123588131319e-06, 'learning_rate/action_model': 8.359123588131319e-05, 'epoch': 1.25}) +[2026-06-15 22:33:49,901][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.16661754250526428, 'train/grad_norm_pre_clip': 5.963676929473877, 'timing/data': 0.02049043495208025, 'timing/model': 1.7063116887584329, 'learning_rate/qwen_vl_interface': 8.35682055981103e-06, 'learning_rate/action_model': 8.35682055981103e-05, 'epoch': 1.25}) +[2026-06-15 22:33:55,554][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.10997986793518066, 'train/grad_norm_pre_clip': 3.6845450401306152, 'timing/data': 0.021880436688661575, 'timing/model': 1.4414518931880593, 'learning_rate/qwen_vl_interface': 8.354516254394163e-06, 'learning_rate/action_model': 8.354516254394162e-05, 'epoch': 1.25}) +[2026-06-15 22:34:01,355][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.21669165790081024, 'train/grad_norm_pre_clip': 6.994289398193359, 'timing/data': 0.025468917563557625, 'timing/model': 1.4299049004912376, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 1.25}) +[2026-06-15 22:34:07,056][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.08656153827905655, 'train/grad_norm_pre_clip': 4.152548313140869, 'timing/data': 0.021186559461057186, 'timing/model': 1.4497410710901022, 'learning_rate/qwen_vl_interface': 8.349903816060074e-06, 'learning_rate/action_model': 8.349903816060075e-05, 'epoch': 1.25}) +[2026-06-15 22:34:12,853][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.2979978322982788, 'train/grad_norm_pre_clip': 8.582379341125488, 'timing/data': 0.021182828582823277, 'timing/model': 1.4434985779225826, 'learning_rate/qwen_vl_interface': 8.34759568503885e-06, 'learning_rate/action_model': 8.347595685038851e-05, 'epoch': 1.25}) +[2026-06-15 22:34:18,657][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.14201581478118896, 'train/grad_norm_pre_clip': 5.567831993103027, 'timing/data': 0.020849204622209072, 'timing/model': 1.4313263818621635, 'learning_rate/qwen_vl_interface': 8.345286280713049e-06, 'learning_rate/action_model': 8.345286280713048e-05, 'epoch': 1.25}) +[2026-06-15 22:34:24,504][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.10998938232660294, 'train/grad_norm_pre_clip': 3.7076709270477295, 'timing/data': 0.02121682744473219, 'timing/model': 1.7042316291481256, 'learning_rate/qwen_vl_interface': 8.342975604031974e-06, 'learning_rate/action_model': 8.342975604031974e-05, 'epoch': 1.25}) +[2026-06-15 22:34:30,149][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.059457167983055115, 'train/grad_norm_pre_clip': 2.2134530544281006, 'timing/data': 0.02114073559641838, 'timing/model': 1.4440676122903824, 'learning_rate/qwen_vl_interface': 8.340663655945462e-06, 'learning_rate/action_model': 8.340663655945462e-05, 'epoch': 1.26}) +[2026-06-15 22:34:35,955][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.0581178218126297, 'train/grad_norm_pre_clip': 2.396646738052368, 'timing/data': 0.020334437489509583, 'timing/model': 1.4569189297035336, 'learning_rate/qwen_vl_interface': 8.338350437403866e-06, 'learning_rate/action_model': 8.338350437403865e-05, 'epoch': 1.26}) +[2026-06-15 22:34:41,754][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.110878124833107, 'train/grad_norm_pre_clip': 3.8283863067626953, 'timing/data': 0.020600449293851852, 'timing/model': 1.4538091663271189, 'learning_rate/qwen_vl_interface': 8.336035949358063e-06, 'learning_rate/action_model': 8.336035949358062e-05, 'epoch': 1.26}) +[2026-06-15 22:34:47,553][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.050707682967185974, 'train/grad_norm_pre_clip': 1.9985750913619995, 'timing/data': 0.02245272509753704, 'timing/model': 1.4397618072107434, 'learning_rate/qwen_vl_interface': 8.333720192759448e-06, 'learning_rate/action_model': 8.333720192759448e-05, 'epoch': 1.26}) +[2026-06-15 22:34:53,349][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.0693669468164444, 'train/grad_norm_pre_clip': 3.500061511993408, 'timing/data': 0.020576967857778072, 'timing/model': 1.4401773680001497, 'learning_rate/qwen_vl_interface': 8.331403168559946e-06, 'learning_rate/action_model': 8.331403168559945e-05, 'epoch': 1.26}) +[2026-06-15 22:34:59,214][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.05842457711696625, 'train/grad_norm_pre_clip': 3.1974596977233887, 'timing/data': 0.021006841212511063, 'timing/model': 1.693659020587802, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 1.26}) +[2026-06-15 22:35:04,853][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.23362848162651062, 'train/grad_norm_pre_clip': 7.916008949279785, 'timing/data': 0.0211860416457057, 'timing/model': 1.4461494134739041, 'learning_rate/qwen_vl_interface': 8.32676532116856e-06, 'learning_rate/action_model': 8.326765321168559e-05, 'epoch': 1.26}) +[2026-06-15 22:35:10,552][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.041600197553634644, 'train/grad_norm_pre_clip': 3.542578935623169, 'timing/data': 0.024222327396273613, 'timing/model': 1.4221611144021153, 'learning_rate/qwen_vl_interface': 8.324444499883119e-06, 'learning_rate/action_model': 8.324444499883119e-05, 'epoch': 1.26}) +[2026-06-15 22:35:16,352][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.07000095397233963, 'train/grad_norm_pre_clip': 7.138040065765381, 'timing/data': 0.020573405548930168, 'timing/model': 1.4494790863245726, 'learning_rate/qwen_vl_interface': 8.322122414809678e-06, 'learning_rate/action_model': 8.322122414809677e-05, 'epoch': 1.26}) +[2026-06-15 22:35:22,154][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.05718547850847244, 'train/grad_norm_pre_clip': 4.49262809753418, 'timing/data': 0.020452714525163174, 'timing/model': 1.4026511572301388, 'learning_rate/qwen_vl_interface': 8.319799066902755e-06, 'learning_rate/action_model': 8.319799066902755e-05, 'epoch': 1.26}) +[2026-06-15 22:35:27,857][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.14321944117546082, 'train/grad_norm_pre_clip': 8.5941162109375, 'timing/data': 0.02188974618911743, 'timing/model': 1.4196203909814358, 'learning_rate/qwen_vl_interface': 8.317474457117393e-06, 'learning_rate/action_model': 8.317474457117392e-05, 'epoch': 1.26}) +[2026-06-15 22:35:33,696][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.2184017151594162, 'train/grad_norm_pre_clip': 10.671039581298828, 'timing/data': 0.02377730142325163, 'timing/model': 1.6658330000936985, 'learning_rate/qwen_vl_interface': 8.315148586409148e-06, 'learning_rate/action_model': 8.315148586409148e-05, 'epoch': 1.26}) +[2026-06-15 22:35:39,361][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.07328879088163376, 'train/grad_norm_pre_clip': 6.730849266052246, 'timing/data': 0.022237857803702354, 'timing/model': 1.4399051805958152, 'learning_rate/qwen_vl_interface': 8.312821455734104e-06, 'learning_rate/action_model': 8.312821455734103e-05, 'epoch': 1.26}) +[2026-06-15 22:35:45,157][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.1012556403875351, 'train/grad_norm_pre_clip': 7.104130268096924, 'timing/data': 0.021710802800953388, 'timing/model': 1.436450393870473, 'learning_rate/qwen_vl_interface': 8.31049306604885e-06, 'learning_rate/action_model': 8.310493066048849e-05, 'epoch': 1.26}) +[2026-06-15 22:35:50,953][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.1690462976694107, 'train/grad_norm_pre_clip': 12.542091369628906, 'timing/data': 0.02193570416420698, 'timing/model': 1.4452538378536701, 'learning_rate/qwen_vl_interface': 8.3081634183105e-06, 'learning_rate/action_model': 8.308163418310501e-05, 'epoch': 1.26}) +[2026-06-15 22:35:56,667][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.2134532481431961, 'train/grad_norm_pre_clip': 14.733975410461426, 'timing/data': 0.024144141003489494, 'timing/model': 1.4534231070429087, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 1.26}) +[2026-06-15 22:36:02,455][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.1448916792869568, 'train/grad_norm_pre_clip': 7.239743709564209, 'timing/data': 0.02062859106808901, 'timing/model': 1.4085072642192245, 'learning_rate/qwen_vl_interface': 8.303500352505557e-06, 'learning_rate/action_model': 8.303500352505557e-05, 'epoch': 1.26}) +[2026-06-15 22:36:08,304][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.10628906637430191, 'train/grad_norm_pre_clip': 6.2965779304504395, 'timing/data': 0.021293556317687035, 'timing/model': 1.686465211212635, 'learning_rate/qwen_vl_interface': 8.301166936355773e-06, 'learning_rate/action_model': 8.301166936355773e-05, 'epoch': 1.26}) +[2026-06-15 22:36:13,956][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.3130456209182739, 'train/grad_norm_pre_clip': 11.96968936920166, 'timing/data': 0.022797929123044014, 'timing/model': 1.4394716573879123, 'learning_rate/qwen_vl_interface': 8.298832265986514e-06, 'learning_rate/action_model': 8.298832265986513e-05, 'epoch': 1.26}) +[2026-06-15 22:36:19,761][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.1368722915649414, 'train/grad_norm_pre_clip': 12.412168502807617, 'timing/data': 0.022219783626496792, 'timing/model': 1.4539899388328195, 'learning_rate/qwen_vl_interface': 8.296496342357475e-06, 'learning_rate/action_model': 8.296496342357474e-05, 'epoch': 1.26}) +[2026-06-15 22:36:25,551][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.07873538136482239, 'train/grad_norm_pre_clip': 6.4983906745910645, 'timing/data': 0.02163256797939539, 'timing/model': 1.4459791658446193, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 1.26}) +[2026-06-15 22:36:31,365][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.10517451167106628, 'train/grad_norm_pre_clip': 5.152851104736328, 'timing/data': 0.022080371156334877, 'timing/model': 1.4553314847871661, 'learning_rate/qwen_vl_interface': 8.29182073916141e-06, 'learning_rate/action_model': 8.29182073916141e-05, 'epoch': 1.26}) +[2026-06-15 22:36:37,172][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.35502389073371887, 'train/grad_norm_pre_clip': 10.571235656738281, 'timing/data': 0.02356900181621313, 'timing/model': 1.4068888640031219, 'learning_rate/qwen_vl_interface': 8.28948106151635e-06, 'learning_rate/action_model': 8.28948106151635e-05, 'epoch': 1.27}) +[2026-06-15 22:36:43,013][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.08743826299905777, 'train/grad_norm_pre_clip': 5.362184524536133, 'timing/data': 0.021027560345828533, 'timing/model': 1.7072313418611884, 'learning_rate/qwen_vl_interface': 8.287140134455435e-06, 'learning_rate/action_model': 8.287140134455435e-05, 'epoch': 1.27}) +[2026-06-15 22:36:48,658][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.15653206408023834, 'train/grad_norm_pre_clip': 10.297517776489258, 'timing/data': 0.0218441067263484, 'timing/model': 1.4378483556210995, 'learning_rate/qwen_vl_interface': 8.284797958940934e-06, 'learning_rate/action_model': 8.284797958940935e-05, 'epoch': 1.27}) +[2026-06-15 22:36:54,459][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.34257689118385315, 'train/grad_norm_pre_clip': 8.925768852233887, 'timing/data': 0.020665346644818783, 'timing/model': 1.450593133457005, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 1.27}) +[2026-06-15 22:37:00,255][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.3108052909374237, 'train/grad_norm_pre_clip': 16.664871215820312, 'timing/data': 0.021040240302681923, 'timing/model': 1.4394403602927923, 'learning_rate/qwen_vl_interface': 8.280109866402805e-06, 'learning_rate/action_model': 8.280109866402806e-05, 'epoch': 1.27}) +[2026-06-15 22:37:06,055][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.32637184858322144, 'train/grad_norm_pre_clip': 12.66361141204834, 'timing/data': 0.021773245185613632, 'timing/model': 1.4390931045636535, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 1.27}) +[2026-06-15 22:37:11,859][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.2734001874923706, 'train/grad_norm_pre_clip': 9.369949340820312, 'timing/data': 0.02117351908236742, 'timing/model': 1.401125654578209, 'learning_rate/qwen_vl_interface': 8.275416791610356e-06, 'learning_rate/action_model': 8.275416791610355e-05, 'epoch': 1.27}) +[2026-06-15 22:37:17,690][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.100335031747818, 'train/grad_norm_pre_clip': 3.5274622440338135, 'timing/data': 0.020999007858335972, 'timing/model': 1.6485195495188236, 'learning_rate/qwen_vl_interface': 8.27306838827987e-06, 'learning_rate/action_model': 8.273068388279871e-05, 'epoch': 1.27}) +[2026-06-15 22:37:23,353][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.3482924699783325, 'train/grad_norm_pre_clip': 15.464552879333496, 'timing/data': 0.02395196072757244, 'timing/model': 1.4198633898049593, 'learning_rate/qwen_vl_interface': 8.270718742280162e-06, 'learning_rate/action_model': 8.270718742280162e-05, 'epoch': 1.27}) +[2026-06-15 22:37:29,156][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.26254802942276, 'train/grad_norm_pre_clip': 8.739113807678223, 'timing/data': 0.021676192060112953, 'timing/model': 1.4249122086912394, 'learning_rate/qwen_vl_interface': 8.268367854577084e-06, 'learning_rate/action_model': 8.268367854577082e-05, 'epoch': 1.27}) +[2026-06-15 22:37:34,971][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.11670676618814468, 'train/grad_norm_pre_clip': 3.9728739261627197, 'timing/data': 0.019405544735491276, 'timing/model': 1.4431187910959125, 'learning_rate/qwen_vl_interface': 8.266015726136991e-06, 'learning_rate/action_model': 8.266015726136991e-05, 'epoch': 1.27}) +[2026-06-15 22:37:40,757][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.14389623701572418, 'train/grad_norm_pre_clip': 5.758476734161377, 'timing/data': 0.023358063772320747, 'timing/model': 1.4444237696006894, 'learning_rate/qwen_vl_interface': 8.26366235792676e-06, 'learning_rate/action_model': 8.26366235792676e-05, 'epoch': 1.27}) +[2026-06-15 22:37:46,567][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.12587417662143707, 'train/grad_norm_pre_clip': 3.588715076446533, 'timing/data': 0.021886912174522877, 'timing/model': 1.4080442106351256, 'learning_rate/qwen_vl_interface': 8.261307750913769e-06, 'learning_rate/action_model': 8.261307750913769e-05, 'epoch': 1.27}) +[2026-06-15 22:37:52,394][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.2396882325410843, 'train/grad_norm_pre_clip': 5.564818382263184, 'timing/data': 0.02131789643317461, 'timing/model': 1.6730583617463708, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 1.27}) +[2026-06-15 22:37:58,241][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.20980027318000793, 'train/grad_norm_pre_clip': 5.617387294769287, 'timing/data': 0.02044042106717825, 'timing/model': 1.6182500571012497, 'learning_rate/qwen_vl_interface': 8.256594824351577e-06, 'learning_rate/action_model': 8.256594824351576e-05, 'epoch': 1.27}) +[2026-06-15 22:38:03,857][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.18089671432971954, 'train/grad_norm_pre_clip': 4.339265823364258, 'timing/data': 0.0219346908852458, 'timing/model': 1.442081586457789, 'learning_rate/qwen_vl_interface': 8.254236506739683e-06, 'learning_rate/action_model': 8.254236506739681e-05, 'epoch': 1.27}) +[2026-06-15 22:38:09,659][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.1278105229139328, 'train/grad_norm_pre_clip': 4.057774543762207, 'timing/data': 0.022026602178812027, 'timing/model': 1.4352493230253458, 'learning_rate/qwen_vl_interface': 8.25187695419964e-06, 'learning_rate/action_model': 8.25187695419964e-05, 'epoch': 1.27}) +[2026-06-15 22:38:15,462][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.31580623984336853, 'train/grad_norm_pre_clip': 6.173534393310547, 'timing/data': 0.021798918955028057, 'timing/model': 1.4411003487184644, 'learning_rate/qwen_vl_interface': 8.249516167701373e-06, 'learning_rate/action_model': 8.249516167701374e-05, 'epoch': 1.27}) +[2026-06-15 22:38:21,259][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.07239899784326553, 'train/grad_norm_pre_clip': 6.613025188446045, 'timing/data': 0.02232254296541214, 'timing/model': 1.4048266084864736, 'learning_rate/qwen_vl_interface': 8.247154148215311e-06, 'learning_rate/action_model': 8.24715414821531e-05, 'epoch': 1.27}) +[2026-06-15 22:38:27,108][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.17700237035751343, 'train/grad_norm_pre_clip': 3.691699981689453, 'timing/data': 0.02373089361935854, 'timing/model': 1.6885935133323073, 'learning_rate/qwen_vl_interface': 8.24479089671239e-06, 'learning_rate/action_model': 8.24479089671239e-05, 'epoch': 1.27}) +[2026-06-15 22:38:33,020][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.11740642786026001, 'train/grad_norm_pre_clip': 4.204660892486572, 'timing/data': 0.02136304508894682, 'timing/model': 1.705892133526504, 'learning_rate/qwen_vl_interface': 8.242426414164056e-06, 'learning_rate/action_model': 8.242426414164056e-05, 'epoch': 1.27}) +[2026-06-15 22:38:38,662][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.23947633802890778, 'train/grad_norm_pre_clip': 6.323569297790527, 'timing/data': 0.02101715188473463, 'timing/model': 1.4413279350847006, 'learning_rate/qwen_vl_interface': 8.240060701542257e-06, 'learning_rate/action_model': 8.240060701542257e-05, 'epoch': 1.28}) +[2026-06-15 22:38:44,456][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.23024988174438477, 'train/grad_norm_pre_clip': 8.060595512390137, 'timing/data': 0.0208346089348197, 'timing/model': 1.4310530479997396, 'learning_rate/qwen_vl_interface': 8.237693759819447e-06, 'learning_rate/action_model': 8.237693759819448e-05, 'epoch': 1.28}) +[2026-06-15 22:38:50,257][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.07546006143093109, 'train/grad_norm_pre_clip': 3.596752882003784, 'timing/data': 0.02120531164109707, 'timing/model': 1.4453036589547992, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 1.28}) +[2026-06-15 22:38:56,077][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.14382676780223846, 'train/grad_norm_pre_clip': 3.8179595470428467, 'timing/data': 0.021434943191707134, 'timing/model': 1.4036161238327622, 'learning_rate/qwen_vl_interface': 8.232956192963143e-06, 'learning_rate/action_model': 8.232956192963144e-05, 'epoch': 1.28}) +[2026-06-15 22:39:01,917][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.3154662549495697, 'train/grad_norm_pre_clip': 8.245673179626465, 'timing/data': 0.02069670706987381, 'timing/model': 1.6914338786154985, 'learning_rate/qwen_vl_interface': 8.230585569777086e-06, 'learning_rate/action_model': 8.230585569777085e-05, 'epoch': 1.28}) +[2026-06-15 22:39:07,554][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.2925122380256653, 'train/grad_norm_pre_clip': 8.752825736999512, 'timing/data': 0.02005003299564123, 'timing/model': 1.4286709111183882, 'learning_rate/qwen_vl_interface': 8.228213721384885e-06, 'learning_rate/action_model': 8.228213721384885e-05, 'epoch': 1.28}) +[2026-06-15 22:39:13,350][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.1779407411813736, 'train/grad_norm_pre_clip': 5.267574310302734, 'timing/data': 0.02610673848539591, 'timing/model': 1.4094304591417313, 'learning_rate/qwen_vl_interface': 8.225840648761523e-06, 'learning_rate/action_model': 8.225840648761523e-05, 'epoch': 1.28}) +[2026-06-15 22:39:19,167][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.14440783858299255, 'train/grad_norm_pre_clip': 4.910271644592285, 'timing/data': 0.021414692513644695, 'timing/model': 1.46486297249794, 'learning_rate/qwen_vl_interface': 8.223466352882476e-06, 'learning_rate/action_model': 8.223466352882475e-05, 'epoch': 1.28}) +[2026-06-15 22:39:25,037][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.11787042766809464, 'train/grad_norm_pre_clip': 3.5693111419677734, 'timing/data': 0.020592164248228073, 'timing/model': 1.469495801255107, 'learning_rate/qwen_vl_interface': 8.22109083472373e-06, 'learning_rate/action_model': 8.22109083472373e-05, 'epoch': 1.28}) +[2026-06-15 22:39:30,832][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.17133043706417084, 'train/grad_norm_pre_clip': 6.140491962432861, 'timing/data': 0.021270920522511005, 'timing/model': 1.409426343627274, 'learning_rate/qwen_vl_interface': 8.218714095261769e-06, 'learning_rate/action_model': 8.218714095261768e-05, 'epoch': 1.28}) +[2026-06-15 22:39:36,606][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.10975363105535507, 'train/grad_norm_pre_clip': 2.6028757095336914, 'timing/data': 0.021777484565973282, 'timing/model': 1.678592998534441, 'learning_rate/qwen_vl_interface': 8.21633613547358e-06, 'learning_rate/action_model': 8.21633613547358e-05, 'epoch': 1.28}) +[2026-06-15 22:39:42,255][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.16955208778381348, 'train/grad_norm_pre_clip': 4.531179904937744, 'timing/data': 0.022450807504355907, 'timing/model': 1.434380130842328, 'learning_rate/qwen_vl_interface': 8.213956956336656e-06, 'learning_rate/action_model': 8.213956956336656e-05, 'epoch': 1.28}) +[2026-06-15 22:50:20,271][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.11360947787761688, 'train/grad_norm_pre_clip': 5.110144138336182, 'eval/loss': 0.07789736241102219, 'eval/flap_precision': 0.74822695035461, 'eval/flap_recall': 0.7672727272727272, 'eval/flap_f1': 0.7576301615798924, 'eval/flap_rate_model': 0.07833333333333334, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7576301615798924, 'timing/data': 0.023466224782168865, 'timing/model': 1.4438532069325447, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 1.28}) +[2026-06-15 22:50:20,272][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2800_state +[2026-06-15 22:50:20,279][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 22:50:28,117][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2800_state/model.safetensors +[2026-06-15 22:50:39,271][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2800_state/optimizer.bin +[2026-06-15 22:50:39,272][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2800_state/sampler.bin +[2026-06-15 22:50:39,272][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2800_state/sampler_1.bin +[2026-06-15 22:50:39,274][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2800_state/random_states_0.pkl +[2026-06-15 22:50:39,428][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2800 +[2026-06-15 22:50:39,428][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 22:50:39,444][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 22:50:45,100][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.2995048463344574, 'train/grad_norm_pre_clip': 6.773569107055664, 'timing/data': 0.022892465814948082, 'timing/model': 1.4201630055904388, 'learning_rate/qwen_vl_interface': 8.209194943929059e-06, 'learning_rate/action_model': 8.20919494392906e-05, 'epoch': 1.28}) +[2026-06-15 22:50:50,802][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.07193688303232193, 'train/grad_norm_pre_clip': 4.109556674957275, 'timing/data': 0.0196920158341527, 'timing/model': 1.4247249169275165, 'learning_rate/qwen_vl_interface': 8.20681211261587e-06, 'learning_rate/action_model': 8.20681211261587e-05, 'epoch': 1.28}) +[2026-06-15 22:50:56,516][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.22783751785755157, 'train/grad_norm_pre_clip': 5.386418342590332, 'timing/data': 0.02046665083616972, 'timing/model': 1.4360091518610716, 'learning_rate/qwen_vl_interface': 8.204428065868911e-06, 'learning_rate/action_model': 8.204428065868911e-05, 'epoch': 1.28}) +[2026-06-15 22:51:02,232][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.2044481337070465, 'train/grad_norm_pre_clip': 7.085397243499756, 'timing/data': 0.01989296730607748, 'timing/model': 1.4166770335286856, 'learning_rate/qwen_vl_interface': 8.202042804668173e-06, 'learning_rate/action_model': 8.202042804668172e-05, 'epoch': 1.28}) +[2026-06-15 22:51:07,822][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.18797367811203003, 'train/grad_norm_pre_clip': 5.012889385223389, 'timing/data': 0.023121926933526993, 'timing/model': 1.3746797190979123, 'learning_rate/qwen_vl_interface': 8.199656329994145e-06, 'learning_rate/action_model': 8.199656329994145e-05, 'epoch': 1.28}) +[2026-06-15 22:51:13,553][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.09871496260166168, 'train/grad_norm_pre_clip': 5.45000696182251, 'timing/data': 0.019790325313806534, 'timing/model': 1.6731287399306893, 'learning_rate/qwen_vl_interface': 8.19726864282782e-06, 'learning_rate/action_model': 8.197268642827818e-05, 'epoch': 1.28}) +[2026-06-15 22:51:18,942][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.25007715821266174, 'train/grad_norm_pre_clip': 7.464965343475342, 'timing/data': 0.022679113782942295, 'timing/model': 1.4380560806021094, 'learning_rate/qwen_vl_interface': 8.194879744150682e-06, 'learning_rate/action_model': 8.194879744150681e-05, 'epoch': 1.28}) +[2026-06-15 22:51:24,637][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.20250052213668823, 'train/grad_norm_pre_clip': 8.531599998474121, 'timing/data': 0.020962783135473728, 'timing/model': 1.4367125248536468, 'learning_rate/qwen_vl_interface': 8.192489634944718e-06, 'learning_rate/action_model': 8.192489634944718e-05, 'epoch': 1.28}) +[2026-06-15 22:51:30,350][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.13720527291297913, 'train/grad_norm_pre_clip': 4.369706630706787, 'timing/data': 0.02099529281258583, 'timing/model': 1.4547961233183742, 'learning_rate/qwen_vl_interface': 8.190098316192416e-06, 'learning_rate/action_model': 8.190098316192416e-05, 'epoch': 1.28}) +[2026-06-15 22:51:36,044][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.09576290845870972, 'train/grad_norm_pre_clip': 4.4455037117004395, 'timing/data': 0.02164081670343876, 'timing/model': 1.3843605611473322, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 1.29}) +[2026-06-15 22:51:41,785][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.14866404235363007, 'train/grad_norm_pre_clip': 3.627598524093628, 'timing/data': 0.022315332666039467, 'timing/model': 1.6873772945255041, 'learning_rate/qwen_vl_interface': 8.185312053981204e-06, 'learning_rate/action_model': 8.185312053981204e-05, 'epoch': 1.29}) +[2026-06-15 22:51:47,580][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.09965173900127411, 'train/grad_norm_pre_clip': 2.875383138656616, 'timing/data': 0.021962138824164867, 'timing/model': 1.6890685074031353, 'learning_rate/qwen_vl_interface': 8.182917112489747e-06, 'learning_rate/action_model': 8.182917112489746e-05, 'epoch': 1.29}) +[2026-06-15 22:51:53,138][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.10348833352327347, 'train/grad_norm_pre_clip': 4.371800422668457, 'timing/data': 0.02125158440321684, 'timing/model': 1.4523335937410593, 'learning_rate/qwen_vl_interface': 8.18052096538685e-06, 'learning_rate/action_model': 8.18052096538685e-05, 'epoch': 1.29}) +[2026-06-15 22:51:58,950][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.0512031726539135, 'train/grad_norm_pre_clip': 3.004169464111328, 'timing/data': 0.020953773520886898, 'timing/model': 1.459855074994266, 'learning_rate/qwen_vl_interface': 8.178123613657483e-06, 'learning_rate/action_model': 8.178123613657482e-05, 'epoch': 1.29}) +[2026-06-15 22:52:04,646][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.11875645071268082, 'train/grad_norm_pre_clip': 4.190032958984375, 'timing/data': 0.021126852370798588, 'timing/model': 1.4552861899137497, 'learning_rate/qwen_vl_interface': 8.175725058287102e-06, 'learning_rate/action_model': 8.175725058287102e-05, 'epoch': 1.29}) +[2026-06-15 22:52:10,348][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.15813840925693512, 'train/grad_norm_pre_clip': 5.654735088348389, 'timing/data': 0.021358034573495388, 'timing/model': 1.4037994099780917, 'learning_rate/qwen_vl_interface': 8.173325300261664e-06, 'learning_rate/action_model': 8.173325300261663e-05, 'epoch': 1.29}) +[2026-06-15 22:52:16,093][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.07598818093538284, 'train/grad_norm_pre_clip': 5.060691833496094, 'timing/data': 0.02085581235587597, 'timing/model': 1.6868247892707586, 'learning_rate/qwen_vl_interface': 8.170924340567617e-06, 'learning_rate/action_model': 8.170924340567616e-05, 'epoch': 1.29}) +[2026-06-15 22:52:21,802][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.05054701119661331, 'train/grad_norm_pre_clip': 3.950883388519287, 'timing/data': 0.021665705367922783, 'timing/model': 1.6043989239260554, 'learning_rate/qwen_vl_interface': 8.168522180191909e-06, 'learning_rate/action_model': 8.168522180191909e-05, 'epoch': 1.29}) +[2026-06-15 22:52:27,345][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.09408240020275116, 'train/grad_norm_pre_clip': 4.665259838104248, 'timing/data': 0.021383714862167835, 'timing/model': 1.4380275886505842, 'learning_rate/qwen_vl_interface': 8.166118820121972e-06, 'learning_rate/action_model': 8.166118820121973e-05, 'epoch': 1.29}) +[2026-06-15 22:52:33,148][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.09442154318094254, 'train/grad_norm_pre_clip': 5.863142490386963, 'timing/data': 0.02111868280917406, 'timing/model': 1.4378829430788755, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 1.29}) +[2026-06-15 22:52:38,947][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.08360539376735687, 'train/grad_norm_pre_clip': 8.521625518798828, 'timing/data': 0.020884471014142036, 'timing/model': 1.4557343693450093, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 1.29}) +[2026-06-15 22:52:44,750][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.1838228404521942, 'train/grad_norm_pre_clip': 17.5571231842041, 'timing/data': 0.02005587611347437, 'timing/model': 1.4125896133482456, 'learning_rate/qwen_vl_interface': 8.158901551628579e-06, 'learning_rate/action_model': 8.158901551628579e-05, 'epoch': 1.29}) +[2026-06-15 22:52:50,481][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.0760454311966896, 'train/grad_norm_pre_clip': 6.59390115737915, 'timing/data': 0.021211132407188416, 'timing/model': 1.6736084874719381, 'learning_rate/qwen_vl_interface': 8.156493402665971e-06, 'learning_rate/action_model': 8.15649340266597e-05, 'epoch': 1.29}) +[2026-06-15 22:52:56,140][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.10201707482337952, 'train/grad_norm_pre_clip': 8.06456184387207, 'timing/data': 0.021379915066063404, 'timing/model': 1.5657010842114687, 'learning_rate/qwen_vl_interface': 8.154084058953713e-06, 'learning_rate/action_model': 8.154084058953713e-05, 'epoch': 1.29}) +[2026-06-15 22:53:01,764][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.08036069571971893, 'train/grad_norm_pre_clip': 5.254743576049805, 'timing/data': 0.021505465731024742, 'timing/model': 1.4540815185755491, 'learning_rate/qwen_vl_interface': 8.151673521482193e-06, 'learning_rate/action_model': 8.151673521482194e-05, 'epoch': 1.29}) +[2026-06-15 22:53:07,546][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.237600177526474, 'train/grad_norm_pre_clip': 11.987519264221191, 'timing/data': 0.022635831497609615, 'timing/model': 1.4186685495078564, 'learning_rate/qwen_vl_interface': 8.149261791242297e-06, 'learning_rate/action_model': 8.149261791242297e-05, 'epoch': 1.29}) +[2026-06-15 22:53:13,348][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.10498502850532532, 'train/grad_norm_pre_clip': 8.676908493041992, 'timing/data': 0.021785765886306763, 'timing/model': 1.4437135830521584, 'learning_rate/qwen_vl_interface': 8.146848869225391e-06, 'learning_rate/action_model': 8.146848869225391e-05, 'epoch': 1.29}) +[2026-06-15 22:53:19,161][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.12468232959508896, 'train/grad_norm_pre_clip': 6.832907199859619, 'timing/data': 0.020251186564564705, 'timing/model': 1.4165378464385867, 'learning_rate/qwen_vl_interface': 8.14443475642334e-06, 'learning_rate/action_model': 8.144434756423339e-05, 'epoch': 1.29}) +[2026-06-15 22:53:25,020][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.05749722197651863, 'train/grad_norm_pre_clip': 10.5029878616333, 'timing/data': 0.02092656120657921, 'timing/model': 1.7296979827806354, 'learning_rate/qwen_vl_interface': 8.142019453828493e-06, 'learning_rate/action_model': 8.142019453828493e-05, 'epoch': 1.29}) +[2026-06-15 22:53:30,549][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.2695375084877014, 'train/grad_norm_pre_clip': 18.3018856048584, 'timing/data': 0.02051898743957281, 'timing/model': 1.4539521420374513, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 1.29}) +[2026-06-15 22:53:36,343][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.024708310142159462, 'train/grad_norm_pre_clip': 2.315725803375244, 'timing/data': 0.020673369988799095, 'timing/model': 1.4402456395328045, 'learning_rate/qwen_vl_interface': 8.137185283232255e-06, 'learning_rate/action_model': 8.137185283232256e-05, 'epoch': 1.29}) +[2026-06-15 22:53:42,051][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.23442615568637848, 'train/grad_norm_pre_clip': 9.445353507995605, 'timing/data': 0.02310782391577959, 'timing/model': 1.4364613927900791, 'learning_rate/qwen_vl_interface': 8.134766417218014e-06, 'learning_rate/action_model': 8.134766417218012e-05, 'epoch': 1.3}) +[2026-06-15 22:53:47,753][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.3043459951877594, 'train/grad_norm_pre_clip': 13.744378089904785, 'timing/data': 0.020766030065715313, 'timing/model': 1.462053718045354, 'learning_rate/qwen_vl_interface': 8.132346365385262e-06, 'learning_rate/action_model': 8.132346365385261e-05, 'epoch': 1.3}) +[2026-06-15 22:53:53,450][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.09984807670116425, 'train/grad_norm_pre_clip': 13.077873229980469, 'timing/data': 0.021630325354635715, 'timing/model': 1.395879589021206, 'learning_rate/qwen_vl_interface': 8.129925128728796e-06, 'learning_rate/action_model': 8.129925128728796e-05, 'epoch': 1.3}) +[2026-06-15 22:53:59,200][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.41656771302223206, 'train/grad_norm_pre_clip': 30.929397583007812, 'timing/data': 0.021416453644633293, 'timing/model': 1.6716480571776628, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 1.3}) +[2026-06-15 22:54:04,852][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.0998687893152237, 'train/grad_norm_pre_clip': 16.03046226501465, 'timing/data': 0.022170408628880978, 'timing/model': 1.4383847760036588, 'learning_rate/qwen_vl_interface': 8.125079104926324e-06, 'learning_rate/action_model': 8.125079104926323e-05, 'epoch': 1.3}) +[2026-06-15 22:54:10,558][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.3191553056240082, 'train/grad_norm_pre_clip': 30.557172775268555, 'timing/data': 0.020483428612351418, 'timing/model': 1.4261310053989291, 'learning_rate/qwen_vl_interface': 8.122654319772332e-06, 'learning_rate/action_model': 8.122654319772332e-05, 'epoch': 1.3}) +[2026-06-15 22:54:16,350][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.2660776972770691, 'train/grad_norm_pre_clip': 18.85536003112793, 'timing/data': 0.02066057827323675, 'timing/model': 1.438597877509892, 'learning_rate/qwen_vl_interface': 8.120228353778661e-06, 'learning_rate/action_model': 8.12022835377866e-05, 'epoch': 1.3}) +[2026-06-15 22:54:22,152][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.07997363805770874, 'train/grad_norm_pre_clip': 4.778607368469238, 'timing/data': 0.02144488412886858, 'timing/model': 1.44878563657403, 'learning_rate/qwen_vl_interface': 8.117801207942528e-06, 'learning_rate/action_model': 8.117801207942528e-05, 'epoch': 1.3}) +[2026-06-15 22:54:27,958][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.019313327968120575, 'train/grad_norm_pre_clip': 5.445071220397949, 'timing/data': 0.02162545546889305, 'timing/model': 1.4072946039959788, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 1.3}) +[2026-06-15 22:54:33,800][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.014186646789312363, 'train/grad_norm_pre_clip': 0.8954930305480957, 'timing/data': 0.020888149738311768, 'timing/model': 1.6893211249262094, 'learning_rate/qwen_vl_interface': 8.11294338073421e-06, 'learning_rate/action_model': 8.11294338073421e-05, 'epoch': 1.3}) +[2026-06-15 22:54:39,634][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.08501274883747101, 'train/grad_norm_pre_clip': 11.381174087524414, 'timing/data': 0.021191461943089962, 'timing/model': 1.6302537331357598, 'learning_rate/qwen_vl_interface': 8.110512701358892e-06, 'learning_rate/action_model': 8.110512701358891e-05, 'epoch': 1.3}) +[2026-06-15 22:54:45,269][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.5301671624183655, 'train/grad_norm_pre_clip': 11.216736793518066, 'timing/data': 0.024579623714089394, 'timing/model': 1.450981362722814, 'learning_rate/qwen_vl_interface': 8.108080846134852e-06, 'learning_rate/action_model': 8.108080846134852e-05, 'epoch': 1.3}) +[2026-06-15 22:54:51,053][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.24367602169513702, 'train/grad_norm_pre_clip': 13.232773780822754, 'timing/data': 0.02325737476348877, 'timing/model': 1.429597046226263, 'learning_rate/qwen_vl_interface': 8.105647816061734e-06, 'learning_rate/action_model': 8.105647816061735e-05, 'epoch': 1.3}) +[2026-06-15 22:54:56,858][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.20839761197566986, 'train/grad_norm_pre_clip': 14.762462615966797, 'timing/data': 0.021949946880340576, 'timing/model': 1.4477356346324086, 'learning_rate/qwen_vl_interface': 8.103213612139667e-06, 'learning_rate/action_model': 8.103213612139667e-05, 'epoch': 1.3}) +[2026-06-15 22:55:03,252][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.3227410316467285, 'train/grad_norm_pre_clip': 11.39001178741455, 'timing/data': 0.023715867660939693, 'timing/model': 1.3983338829129934, 'learning_rate/qwen_vl_interface': 8.100778235369262e-06, 'learning_rate/action_model': 8.10077823536926e-05, 'epoch': 1.3}) +[2026-06-15 22:55:09,100][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.13382156193256378, 'train/grad_norm_pre_clip': 11.361556053161621, 'timing/data': 0.023820257745683193, 'timing/model': 1.6849157577380538, 'learning_rate/qwen_vl_interface': 8.098341686751605e-06, 'learning_rate/action_model': 8.098341686751605e-05, 'epoch': 1.3}) +[2026-06-15 22:55:15,008][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.207271546125412, 'train/grad_norm_pre_clip': 9.699593544006348, 'timing/data': 0.022346047684550285, 'timing/model': 1.70751438383013, 'learning_rate/qwen_vl_interface': 8.095903967288273e-06, 'learning_rate/action_model': 8.095903967288272e-05, 'epoch': 1.3}) +[2026-06-15 22:55:20,653][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.1648148000240326, 'train/grad_norm_pre_clip': 6.710525035858154, 'timing/data': 0.022602620534598827, 'timing/model': 1.4375145034864545, 'learning_rate/qwen_vl_interface': 8.093465077981318e-06, 'learning_rate/action_model': 8.093465077981318e-05, 'epoch': 1.3}) +[2026-06-15 22:55:26,456][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.38914984464645386, 'train/grad_norm_pre_clip': 10.783503532409668, 'timing/data': 0.02185504510998726, 'timing/model': 1.4451876180246472, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 1.3}) +[2026-06-15 22:55:32,252][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.08474652469158173, 'train/grad_norm_pre_clip': 5.242090702056885, 'timing/data': 0.02127249166369438, 'timing/model': 1.4491572082042694, 'learning_rate/qwen_vl_interface': 8.088583793847165e-06, 'learning_rate/action_model': 8.088583793847164e-05, 'epoch': 1.3}) +[2026-06-15 22:55:38,055][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.14658527076244354, 'train/grad_norm_pre_clip': 6.46336555480957, 'timing/data': 0.021619383245706558, 'timing/model': 1.404389237985015, 'learning_rate/qwen_vl_interface': 8.086141401026478e-06, 'learning_rate/action_model': 8.086141401026478e-05, 'epoch': 1.3}) +[2026-06-15 22:55:43,903][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.12675005197525024, 'train/grad_norm_pre_clip': 5.268846035003662, 'timing/data': 0.024377181194722652, 'timing/model': 1.6890458557754755, 'learning_rate/qwen_vl_interface': 8.083697842375193e-06, 'learning_rate/action_model': 8.083697842375193e-05, 'epoch': 1.3}) +[2026-06-15 22:55:49,474][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.2477201372385025, 'train/grad_norm_pre_clip': 7.511712551116943, 'timing/data': 0.021300802007317543, 'timing/model': 1.4482163712382317, 'learning_rate/qwen_vl_interface': 8.081253118897762e-06, 'learning_rate/action_model': 8.081253118897761e-05, 'epoch': 1.31}) +[2026-06-15 22:55:55,252][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.007539990823715925, 'train/grad_norm_pre_clip': 0.4139918088912964, 'timing/data': 0.02233148831874132, 'timing/model': 1.4287226870656013, 'learning_rate/qwen_vl_interface': 8.07880723159912e-06, 'learning_rate/action_model': 8.07880723159912e-05, 'epoch': 1.31}) +[2026-06-15 22:56:01,054][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.11035272479057312, 'train/grad_norm_pre_clip': 3.277599573135376, 'timing/data': 0.021766694262623787, 'timing/model': 1.4436591258272529, 'learning_rate/qwen_vl_interface': 8.076360181484676e-06, 'learning_rate/action_model': 8.076360181484676e-05, 'epoch': 1.31}) +[2026-06-15 22:56:06,859][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.17817546427249908, 'train/grad_norm_pre_clip': 7.308372497558594, 'timing/data': 0.021335331723093987, 'timing/model': 1.410378858447075, 'learning_rate/qwen_vl_interface': 8.073911969560326e-06, 'learning_rate/action_model': 8.073911969560326e-05, 'epoch': 1.31}) +[2026-06-15 22:56:12,739][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.059895437210798264, 'train/grad_norm_pre_clip': 2.9320690631866455, 'timing/data': 0.021511582657694817, 'timing/model': 1.4421196887269616, 'learning_rate/qwen_vl_interface': 8.071462596832432e-06, 'learning_rate/action_model': 8.071462596832432e-05, 'epoch': 1.31}) +[2026-06-15 22:56:18,549][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.11190188676118851, 'train/grad_norm_pre_clip': 5.285167217254639, 'timing/data': 0.02079004794359207, 'timing/model': 1.6420760294422507, 'learning_rate/qwen_vl_interface': 8.06901206430784e-06, 'learning_rate/action_model': 8.06901206430784e-05, 'epoch': 1.31}) +[2026-06-15 22:56:24,047][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.10515311360359192, 'train/grad_norm_pre_clip': 4.250757694244385, 'timing/data': 0.020740377716720104, 'timing/model': 1.4525050958618522, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 1.31}) +[2026-06-15 22:56:29,867][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.06906939297914505, 'train/grad_norm_pre_clip': 2.845921277999878, 'timing/data': 0.02869527693837881, 'timing/model': 1.4253742145374417, 'learning_rate/qwen_vl_interface': 8.064107523898327e-06, 'learning_rate/action_model': 8.064107523898326e-05, 'epoch': 1.31}) +[2026-06-15 22:56:35,743][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.15475739538669586, 'train/grad_norm_pre_clip': 10.03417682647705, 'timing/data': 0.02097965870052576, 'timing/model': 1.4344813479110599, 'learning_rate/qwen_vl_interface': 8.061653518029476e-06, 'learning_rate/action_model': 8.061653518029476e-05, 'epoch': 1.31}) +[2026-06-15 22:56:41,545][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.12273484468460083, 'train/grad_norm_pre_clip': 4.868481636047363, 'timing/data': 0.02200472541153431, 'timing/model': 1.3933266010135412, 'learning_rate/qwen_vl_interface': 8.059198356396069e-06, 'learning_rate/action_model': 8.059198356396068e-05, 'epoch': 1.31}) +[2026-06-15 22:56:47,362][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.23305454850196838, 'train/grad_norm_pre_clip': 8.008088111877441, 'timing/data': 0.021801394410431385, 'timing/model': 1.394252123311162, 'learning_rate/qwen_vl_interface': 8.05674204000733e-06, 'learning_rate/action_model': 8.05674204000733e-05, 'epoch': 1.31}) +[2026-06-15 22:56:53,173][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.3615078628063202, 'train/grad_norm_pre_clip': 11.89696979522705, 'timing/data': 0.021344570443034172, 'timing/model': 1.6844445783644915, 'learning_rate/qwen_vl_interface': 8.05428456987296e-06, 'learning_rate/action_model': 8.05428456987296e-05, 'epoch': 1.31}) +[2026-06-15 22:56:58,750][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.22110678255558014, 'train/grad_norm_pre_clip': 7.3486480712890625, 'timing/data': 0.021701233461499214, 'timing/model': 1.4338196497410536, 'learning_rate/qwen_vl_interface': 8.05182594700313e-06, 'learning_rate/action_model': 8.05182594700313e-05, 'epoch': 1.31}) +[2026-06-15 22:57:04,553][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.09945005923509598, 'train/grad_norm_pre_clip': 7.851499080657959, 'timing/data': 0.021122099831700325, 'timing/model': 1.4395713629201055, 'learning_rate/qwen_vl_interface': 8.049366172408484e-06, 'learning_rate/action_model': 8.049366172408485e-05, 'epoch': 1.31}) +[2026-06-15 22:57:10,355][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.1676580011844635, 'train/grad_norm_pre_clip': 5.457251071929932, 'timing/data': 0.020743677392601967, 'timing/model': 1.4507629564031959, 'learning_rate/qwen_vl_interface': 8.04690524710015e-06, 'learning_rate/action_model': 8.046905247100151e-05, 'epoch': 1.31}) +[2026-06-15 22:57:16,173][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.13818371295928955, 'train/grad_norm_pre_clip': 6.997267246246338, 'timing/data': 0.021133149974048138, 'timing/model': 1.4032201115041971, 'learning_rate/qwen_vl_interface': 8.044443172089715e-06, 'learning_rate/action_model': 8.044443172089714e-05, 'epoch': 1.31}) +[2026-06-15 22:57:22,006][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.11626092344522476, 'train/grad_norm_pre_clip': 4.562488079071045, 'timing/data': 0.022051515989005566, 'timing/model': 1.6688715117052197, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 1.31}) +[2026-06-15 22:57:27,653][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.3073669672012329, 'train/grad_norm_pre_clip': 9.439638137817383, 'timing/data': 0.02092485874891281, 'timing/model': 1.4450418837368488, 'learning_rate/qwen_vl_interface': 8.039515577011287e-06, 'learning_rate/action_model': 8.039515577011288e-05, 'epoch': 1.31}) +[2026-06-15 22:57:33,457][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.017706356942653656, 'train/grad_norm_pre_clip': 0.9896071553230286, 'timing/data': 0.023593924939632416, 'timing/model': 1.4388325419276953, 'learning_rate/qwen_vl_interface': 8.037050058968841e-06, 'learning_rate/action_model': 8.03705005896884e-05, 'epoch': 1.31}) +[2026-06-15 22:57:39,152][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.06619012355804443, 'train/grad_norm_pre_clip': 3.3054285049438477, 'timing/data': 0.02225442696362734, 'timing/model': 1.443491317331791, 'learning_rate/qwen_vl_interface': 8.034583395275392e-06, 'learning_rate/action_model': 8.034583395275392e-05, 'epoch': 1.31}) +[2026-06-15 22:57:44,954][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.10925083607435226, 'train/grad_norm_pre_clip': 3.3023734092712402, 'timing/data': 0.02325176354497671, 'timing/model': 1.4400810431689024, 'learning_rate/qwen_vl_interface': 8.032115586944893e-06, 'learning_rate/action_model': 8.032115586944892e-05, 'epoch': 1.31}) +[2026-06-15 22:57:50,666][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.14949746429920197, 'train/grad_norm_pre_clip': 4.101998805999756, 'timing/data': 0.020544526167213917, 'timing/model': 1.4333658693358302, 'learning_rate/qwen_vl_interface': 8.029646634991762e-06, 'learning_rate/action_model': 8.029646634991762e-05, 'epoch': 1.31}) +[2026-06-15 22:57:56,515][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.08443117886781693, 'train/grad_norm_pre_clip': 4.764130115509033, 'timing/data': 0.0207063565030694, 'timing/model': 1.6950489925220609, 'learning_rate/qwen_vl_interface': 8.0271765404309e-06, 'learning_rate/action_model': 8.0271765404309e-05, 'epoch': 1.32}) +[2026-06-15 22:58:02,164][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.18209129571914673, 'train/grad_norm_pre_clip': 7.474408149719238, 'timing/data': 0.021015998907387257, 'timing/model': 1.4577402723953128, 'learning_rate/qwen_vl_interface': 8.024705304277664e-06, 'learning_rate/action_model': 8.024705304277664e-05, 'epoch': 1.32}) +[2026-06-15 22:58:07,959][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.25759896636009216, 'train/grad_norm_pre_clip': 7.955298900604248, 'timing/data': 0.02117836009711027, 'timing/model': 1.4587615868076682, 'learning_rate/qwen_vl_interface': 8.022232927547888e-06, 'learning_rate/action_model': 8.022232927547886e-05, 'epoch': 1.32}) +[2026-06-15 22:58:13,763][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.06353209912776947, 'train/grad_norm_pre_clip': 5.329948425292969, 'timing/data': 0.022699669003486633, 'timing/model': 1.455745805054903, 'learning_rate/qwen_vl_interface': 8.01975941125787e-06, 'learning_rate/action_model': 8.01975941125787e-05, 'epoch': 1.32}) +[2026-06-15 22:58:19,563][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.08327090740203857, 'train/grad_norm_pre_clip': 4.31060266494751, 'timing/data': 0.021476902067661285, 'timing/model': 1.4501255685463548, 'learning_rate/qwen_vl_interface': 8.017284756424381e-06, 'learning_rate/action_model': 8.017284756424381e-05, 'epoch': 1.32}) +[2026-06-15 22:58:25,355][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.21272434294223785, 'train/grad_norm_pre_clip': 7.0073418617248535, 'timing/data': 0.020656466484069824, 'timing/model': 1.4288332043215632, 'learning_rate/qwen_vl_interface': 8.01480896406466e-06, 'learning_rate/action_model': 8.01480896406466e-05, 'epoch': 1.32}) +[2026-06-15 22:58:31,218][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.09625916182994843, 'train/grad_norm_pre_clip': 6.793301105499268, 'timing/data': 0.021801314316689968, 'timing/model': 1.7092219283804297, 'learning_rate/qwen_vl_interface': 8.01233203519641e-06, 'learning_rate/action_model': 8.012332035196409e-05, 'epoch': 1.32}) +[2026-06-15 22:58:36,873][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.07152053713798523, 'train/grad_norm_pre_clip': 3.526233673095703, 'timing/data': 0.023667574860155582, 'timing/model': 1.4641221649944782, 'learning_rate/qwen_vl_interface': 8.009853970837801e-06, 'learning_rate/action_model': 8.009853970837801e-05, 'epoch': 1.32}) +[2026-06-15 22:58:42,654][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.12123504281044006, 'train/grad_norm_pre_clip': 8.303811073303223, 'timing/data': 0.02103618811815977, 'timing/model': 1.442781132645905, 'learning_rate/qwen_vl_interface': 8.007374772007474e-06, 'learning_rate/action_model': 8.007374772007474e-05, 'epoch': 1.32}) +[2026-06-15 22:58:48,458][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.17807534337043762, 'train/grad_norm_pre_clip': 6.998798847198486, 'timing/data': 0.022447177208960056, 'timing/model': 1.4239994147792459, 'learning_rate/qwen_vl_interface': 8.004894439724533e-06, 'learning_rate/action_model': 8.004894439724533e-05, 'epoch': 1.32}) +[2026-06-15 22:58:54,257][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.5118615031242371, 'train/grad_norm_pre_clip': 10.402071952819824, 'timing/data': 0.02144749090075493, 'timing/model': 1.4035466760396957, 'learning_rate/qwen_vl_interface': 8.002412975008552e-06, 'learning_rate/action_model': 8.002412975008552e-05, 'epoch': 1.32}) +[2026-06-15 22:59:00,050][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.037363797426223755, 'train/grad_norm_pre_clip': 2.434750556945801, 'timing/data': 0.02109069935977459, 'timing/model': 1.4141771085560322, 'learning_rate/qwen_vl_interface': 7.999930378879562e-06, 'learning_rate/action_model': 7.999930378879561e-05, 'epoch': 1.32}) +[2026-06-15 22:59:05,911][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.16959689557552338, 'train/grad_norm_pre_clip': 4.157807350158691, 'timing/data': 0.02110583893954754, 'timing/model': 1.6944899298250675, 'learning_rate/qwen_vl_interface': 7.997446652358067e-06, 'learning_rate/action_model': 7.997446652358068e-05, 'epoch': 1.32}) +[2026-06-15 22:59:11,555][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.15352199971675873, 'train/grad_norm_pre_clip': 6.1741251945495605, 'timing/data': 0.02060345560312271, 'timing/model': 1.438862874172628, 'learning_rate/qwen_vl_interface': 7.994961796465035e-06, 'learning_rate/action_model': 7.994961796465035e-05, 'epoch': 1.32}) +[2026-06-15 22:59:17,356][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.11320719867944717, 'train/grad_norm_pre_clip': 6.219527721405029, 'timing/data': 0.02051430568099022, 'timing/model': 1.4414478475227952, 'learning_rate/qwen_vl_interface': 7.992475812221894e-06, 'learning_rate/action_model': 7.992475812221894e-05, 'epoch': 1.32}) +[2026-06-15 22:59:23,118][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.014087047427892685, 'train/grad_norm_pre_clip': 0.9316169619560242, 'timing/data': 0.020232602953910828, 'timing/model': 1.4022497860714793, 'learning_rate/qwen_vl_interface': 7.989988700650538e-06, 'learning_rate/action_model': 7.989988700650538e-05, 'epoch': 1.32}) +[2026-06-15 22:59:28,953][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.2858244478702545, 'train/grad_norm_pre_clip': 8.884881019592285, 'timing/data': 0.020533405244350433, 'timing/model': 1.4463007235899568, 'learning_rate/qwen_vl_interface': 7.987500462773326e-06, 'learning_rate/action_model': 7.987500462773326e-05, 'epoch': 1.32}) +[2026-06-15 22:59:34,753][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.3780740201473236, 'train/grad_norm_pre_clip': 14.1636962890625, 'timing/data': 0.021238189190626144, 'timing/model': 1.4080007066950202, 'learning_rate/qwen_vl_interface': 7.985011099613078e-06, 'learning_rate/action_model': 7.985011099613077e-05, 'epoch': 1.32}) +[2026-06-15 22:59:40,600][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.11622617393732071, 'train/grad_norm_pre_clip': 7.014365196228027, 'timing/data': 0.021912294439971447, 'timing/model': 1.6946500036865473, 'learning_rate/qwen_vl_interface': 7.98252061219308e-06, 'learning_rate/action_model': 7.982520612193078e-05, 'epoch': 1.32}) +[2026-06-15 22:59:46,256][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.12653043866157532, 'train/grad_norm_pre_clip': 9.075202941894531, 'timing/data': 0.020718727260828018, 'timing/model': 1.4524040520191193, 'learning_rate/qwen_vl_interface': 7.980029001537073e-06, 'learning_rate/action_model': 7.980029001537072e-05, 'epoch': 1.32}) +[2026-06-15 22:59:52,059][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.11918961256742477, 'train/grad_norm_pre_clip': 9.874621391296387, 'timing/data': 0.021861103363335133, 'timing/model': 1.4580327048897743, 'learning_rate/qwen_vl_interface': 7.977536268669264e-06, 'learning_rate/action_model': 7.977536268669264e-05, 'epoch': 1.32}) +[2026-06-15 22:59:57,854][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.1289677917957306, 'train/grad_norm_pre_clip': 5.652233123779297, 'timing/data': 0.0205861646682024, 'timing/model': 1.449891502968967, 'learning_rate/qwen_vl_interface': 7.975042414614324e-06, 'learning_rate/action_model': 7.975042414614324e-05, 'epoch': 1.32}) +[2026-06-15 23:00:03,657][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.7193137407302856, 'train/grad_norm_pre_clip': 19.377824783325195, 'timing/data': 0.02109366934746504, 'timing/model': 1.447335404343903, 'learning_rate/qwen_vl_interface': 7.972547440397382e-06, 'learning_rate/action_model': 7.972547440397383e-05, 'epoch': 1.33}) +[2026-06-15 23:00:09,472][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.23678381741046906, 'train/grad_norm_pre_clip': 7.313798427581787, 'timing/data': 0.02138955146074295, 'timing/model': 1.4013471426442266, 'learning_rate/qwen_vl_interface': 7.970051347044026e-06, 'learning_rate/action_model': 7.970051347044025e-05, 'epoch': 1.33}) +[2026-06-15 23:10:48,470][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.16231776773929596, 'train/grad_norm_pre_clip': 6.076813697814941, 'eval/loss': 0.0727539211511612, 'eval/flap_precision': 0.8, 'eval/flap_recall': 0.8, 'eval/flap_f1': 0.8000000000000002, 'eval/flap_rate_model': 0.0763888888888889, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.8000000000000002, 'timing/data': 0.021334960125386715, 'timing/model': 1.7162203676998615, 'learning_rate/qwen_vl_interface': 7.967554135580307e-06, 'learning_rate/action_model': 7.967554135580307e-05, 'epoch': 1.33}) +[2026-06-15 23:10:48,471][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2900_state +[2026-06-15 23:10:48,478][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:10:56,258][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2900_state/model.safetensors +[2026-06-15 23:11:07,571][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2900_state/optimizer.bin +[2026-06-15 23:11:07,572][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2900_state/sampler.bin +[2026-06-15 23:11:07,572][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2900_state/sampler_1.bin +[2026-06-15 23:11:07,574][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_2900_state/random_states_0.pkl +[2026-06-15 23:11:07,745][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2900 +[2026-06-15 23:11:07,745][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:11:07,760][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:11:13,625][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.06763535737991333, 'train/grad_norm_pre_clip': 3.3008503913879395, 'timing/data': 0.022568772546947002, 'timing/model': 1.4197923075407743, 'learning_rate/qwen_vl_interface': 7.965055807032734e-06, 'learning_rate/action_model': 7.965055807032735e-05, 'epoch': 1.33}) +[2026-06-15 23:11:19,298][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.1489410251379013, 'train/grad_norm_pre_clip': 7.084312915802002, 'timing/data': 0.021838278509676456, 'timing/model': 1.618752065114677, 'learning_rate/qwen_vl_interface': 7.962556362428275e-06, 'learning_rate/action_model': 7.962556362428274e-05, 'epoch': 1.33}) +[2026-06-15 23:11:24,836][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.11231281608343124, 'train/grad_norm_pre_clip': 3.4319186210632324, 'timing/data': 0.026309682056307793, 'timing/model': 1.41050024330616, 'learning_rate/qwen_vl_interface': 7.960055802794358e-06, 'learning_rate/action_model': 7.960055802794357e-05, 'epoch': 1.33}) +[2026-06-15 23:11:30,560][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.3247944712638855, 'train/grad_norm_pre_clip': 7.402942657470703, 'timing/data': 0.02292129583656788, 'timing/model': 1.3770462181419134, 'learning_rate/qwen_vl_interface': 7.957554129158866e-06, 'learning_rate/action_model': 7.957554129158866e-05, 'epoch': 1.33}) +[2026-06-15 23:11:36,248][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.04924600571393967, 'train/grad_norm_pre_clip': 1.5186183452606201, 'timing/data': 0.022447172552347183, 'timing/model': 1.4580863481387496, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 1.33}) +[2026-06-15 23:11:42,040][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.04624074324965477, 'train/grad_norm_pre_clip': 2.184664487838745, 'timing/data': 0.02230520360171795, 'timing/model': 1.4479983281344175, 'learning_rate/qwen_vl_interface': 7.952547443997e-06, 'learning_rate/action_model': 7.952547443997e-05, 'epoch': 1.33}) +[2026-06-15 23:11:47,762][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.1210923045873642, 'train/grad_norm_pre_clip': 4.8330159187316895, 'timing/data': 0.021538829430937767, 'timing/model': 1.3985158801078796, 'learning_rate/qwen_vl_interface': 7.950042434528681e-06, 'learning_rate/action_model': 7.950042434528682e-05, 'epoch': 1.33}) +[2026-06-15 23:11:53,596][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.3824504017829895, 'train/grad_norm_pre_clip': 8.526349067687988, 'timing/data': 0.023092596791684628, 'timing/model': 1.7093068333342671, 'learning_rate/qwen_vl_interface': 7.94753631517491e-06, 'learning_rate/action_model': 7.947536315174909e-05, 'epoch': 1.33}) +[2026-06-15 23:11:59,160][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.0883251205086708, 'train/grad_norm_pre_clip': 4.172855854034424, 'timing/data': 0.022104122675955296, 'timing/model': 1.4639267129823565, 'learning_rate/qwen_vl_interface': 7.945029086965854e-06, 'learning_rate/action_model': 7.945029086965853e-05, 'epoch': 1.33}) +[2026-06-15 23:12:04,928][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.29480621218681335, 'train/grad_norm_pre_clip': 7.797983169555664, 'timing/data': 0.021421419456601143, 'timing/model': 1.44213951472193, 'learning_rate/qwen_vl_interface': 7.942520750932138e-06, 'learning_rate/action_model': 7.942520750932138e-05, 'epoch': 1.33}) +[2026-06-15 23:12:10,630][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.06346607208251953, 'train/grad_norm_pre_clip': 5.636932373046875, 'timing/data': 0.021982661448419094, 'timing/model': 1.432959308847785, 'learning_rate/qwen_vl_interface': 7.940011308104848e-06, 'learning_rate/action_model': 7.940011308104848e-05, 'epoch': 1.33}) +[2026-06-15 23:12:16,330][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.1859583854675293, 'train/grad_norm_pre_clip': 9.806875228881836, 'timing/data': 0.02049922477453947, 'timing/model': 1.3880508514121175, 'learning_rate/qwen_vl_interface': 7.93750075951552e-06, 'learning_rate/action_model': 7.93750075951552e-05, 'epoch': 1.33}) +[2026-06-15 23:12:22,086][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.0722687840461731, 'train/grad_norm_pre_clip': 4.434937000274658, 'timing/data': 0.020575075410306454, 'timing/model': 1.6930179363116622, 'learning_rate/qwen_vl_interface': 7.934989106196145e-06, 'learning_rate/action_model': 7.934989106196146e-05, 'epoch': 1.33}) +[2026-06-15 23:12:27,557][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.13681836426258087, 'train/grad_norm_pre_clip': 4.9382853507995605, 'timing/data': 0.020538455806672573, 'timing/model': 1.3820883519947529, 'learning_rate/qwen_vl_interface': 7.93247634917917e-06, 'learning_rate/action_model': 7.93247634917917e-05, 'epoch': 1.33}) +[2026-06-15 23:12:33,267][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.030211307108402252, 'train/grad_norm_pre_clip': 1.3206905126571655, 'timing/data': 0.020610745064914227, 'timing/model': 1.386146922595799, 'learning_rate/qwen_vl_interface': 7.929962489497494e-06, 'learning_rate/action_model': 7.929962489497493e-05, 'epoch': 1.33}) +[2026-06-15 23:12:38,862][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.46844130754470825, 'train/grad_norm_pre_clip': 9.27388858795166, 'timing/data': 0.025165459141135216, 'timing/model': 1.4466383522376418, 'learning_rate/qwen_vl_interface': 7.927447528184469e-06, 'learning_rate/action_model': 7.927447528184468e-05, 'epoch': 1.33}) +[2026-06-15 23:12:44,644][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.12320499867200851, 'train/grad_norm_pre_clip': 8.237585067749023, 'timing/data': 0.021175428293645382, 'timing/model': 1.4515627678483725, 'learning_rate/qwen_vl_interface': 7.924931466273901e-06, 'learning_rate/action_model': 7.924931466273902e-05, 'epoch': 1.33}) +[2026-06-15 23:12:50,347][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.03493831679224968, 'train/grad_norm_pre_clip': 2.2823657989501953, 'timing/data': 0.02156982012093067, 'timing/model': 1.4067570883780718, 'learning_rate/qwen_vl_interface': 7.92241430480005e-06, 'learning_rate/action_model': 7.922414304800049e-05, 'epoch': 1.33}) +[2026-06-15 23:12:56,199][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.09984882920980453, 'train/grad_norm_pre_clip': 4.862918853759766, 'timing/data': 0.02156027127057314, 'timing/model': 1.6912346789613366, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 1.33}) +[2026-06-15 23:13:02,124][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.09229189157485962, 'train/grad_norm_pre_clip': 4.783984661102295, 'timing/data': 0.02263430692255497, 'timing/model': 1.7190243639051914, 'learning_rate/qwen_vl_interface': 7.91737668730178e-06, 'learning_rate/action_model': 7.91737668730178e-05, 'epoch': 1.34}) +[2026-06-15 23:13:07,658][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.2577575445175171, 'train/grad_norm_pre_clip': 12.316722869873047, 'timing/data': 0.02140452153980732, 'timing/model': 1.4664128636941314, 'learning_rate/qwen_vl_interface': 7.91485623334814e-06, 'learning_rate/action_model': 7.91485623334814e-05, 'epoch': 1.34}) +[2026-06-15 23:13:13,453][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.10123962163925171, 'train/grad_norm_pre_clip': 4.4209675788879395, 'timing/data': 0.02133856061846018, 'timing/model': 1.4397364780306816, 'learning_rate/qwen_vl_interface': 7.91233468397276e-06, 'learning_rate/action_model': 7.912334683972761e-05, 'epoch': 1.34}) +[2026-06-15 23:13:19,259][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.22598229348659515, 'train/grad_norm_pre_clip': 6.235841274261475, 'timing/data': 0.021369140595197678, 'timing/model': 1.4592353077605367, 'learning_rate/qwen_vl_interface': 7.90981204021216e-06, 'learning_rate/action_model': 7.90981204021216e-05, 'epoch': 1.34}) +[2026-06-15 23:13:25,065][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.09612535685300827, 'train/grad_norm_pre_clip': 7.7179646492004395, 'timing/data': 0.02201309334486723, 'timing/model': 1.4024020340293646, 'learning_rate/qwen_vl_interface': 7.907288303103298e-06, 'learning_rate/action_model': 7.907288303103298e-05, 'epoch': 1.34}) +[2026-06-15 23:13:30,894][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.08954443037509918, 'train/grad_norm_pre_clip': 3.8443362712860107, 'timing/data': 0.022356816567480564, 'timing/model': 1.6841488620266318, 'learning_rate/qwen_vl_interface': 7.90476347368359e-06, 'learning_rate/action_model': 7.90476347368359e-05, 'epoch': 1.34}) +[2026-06-15 23:13:36,554][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.1384715437889099, 'train/grad_norm_pre_clip': 4.6976542472839355, 'timing/data': 0.024692758917808533, 'timing/model': 1.4412847124040127, 'learning_rate/qwen_vl_interface': 7.902237552990897e-06, 'learning_rate/action_model': 7.902237552990897e-05, 'epoch': 1.34}) +[2026-06-15 23:13:42,353][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.1457647979259491, 'train/grad_norm_pre_clip': 5.575405597686768, 'timing/data': 0.02140551060438156, 'timing/model': 1.4476354261860251, 'learning_rate/qwen_vl_interface': 7.89971054206353e-06, 'learning_rate/action_model': 7.89971054206353e-05, 'epoch': 1.34}) +[2026-06-15 23:13:48,143][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.2168382704257965, 'train/grad_norm_pre_clip': 6.1462082862854, 'timing/data': 0.022663678042590618, 'timing/model': 1.447354824282229, 'learning_rate/qwen_vl_interface': 7.897182441940247e-06, 'learning_rate/action_model': 7.897182441940247e-05, 'epoch': 1.34}) +[2026-06-15 23:13:53,838][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.1037851870059967, 'train/grad_norm_pre_clip': 3.9514951705932617, 'timing/data': 0.02000417374074459, 'timing/model': 1.4039478674530983, 'learning_rate/qwen_vl_interface': 7.894653253660256e-06, 'learning_rate/action_model': 7.894653253660256e-05, 'epoch': 1.34}) +[2026-06-15 23:13:59,543][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.14514555037021637, 'train/grad_norm_pre_clip': 6.665871620178223, 'timing/data': 0.020771347917616367, 'timing/model': 1.457408924587071, 'learning_rate/qwen_vl_interface': 7.89212297826321e-06, 'learning_rate/action_model': 7.892122978263209e-05, 'epoch': 1.34}) +[2026-06-15 23:14:05,264][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.1490195095539093, 'train/grad_norm_pre_clip': 5.381371021270752, 'timing/data': 0.02042869757860899, 'timing/model': 1.6577787483111024, 'learning_rate/qwen_vl_interface': 7.889591616789206e-06, 'learning_rate/action_model': 7.889591616789206e-05, 'epoch': 1.34}) +[2026-06-15 23:14:10,757][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.18976256251335144, 'train/grad_norm_pre_clip': 6.469812393188477, 'timing/data': 0.020749577321112156, 'timing/model': 1.3918293099850416, 'learning_rate/qwen_vl_interface': 7.887059170278799e-06, 'learning_rate/action_model': 7.887059170278797e-05, 'epoch': 1.34}) +[2026-06-15 23:14:16,449][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.16003401577472687, 'train/grad_norm_pre_clip': 6.7501606941223145, 'timing/data': 0.020598658360540867, 'timing/model': 1.4178607137873769, 'learning_rate/qwen_vl_interface': 7.884525639772974e-06, 'learning_rate/action_model': 7.884525639772974e-05, 'epoch': 1.34}) +[2026-06-15 23:14:22,243][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.159979909658432, 'train/grad_norm_pre_clip': 7.72451114654541, 'timing/data': 0.02136941161006689, 'timing/model': 1.4400225384160876, 'learning_rate/qwen_vl_interface': 7.881991026313176e-06, 'learning_rate/action_model': 7.881991026313175e-05, 'epoch': 1.34}) +[2026-06-15 23:14:27,942][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.05950653925538063, 'train/grad_norm_pre_clip': 2.184354782104492, 'timing/data': 0.020237965509295464, 'timing/model': 1.4032482644543052, 'learning_rate/qwen_vl_interface': 7.879455330941283e-06, 'learning_rate/action_model': 7.879455330941282e-05, 'epoch': 1.34}) +[2026-06-15 23:14:33,650][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.2034713625907898, 'train/grad_norm_pre_clip': 6.462376117706299, 'timing/data': 0.02055335696786642, 'timing/model': 1.4357915427535772, 'learning_rate/qwen_vl_interface': 7.876918554699627e-06, 'learning_rate/action_model': 7.876918554699627e-05, 'epoch': 1.34}) +[2026-06-15 23:14:39,499][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.15581075847148895, 'train/grad_norm_pre_clip': 7.9671196937561035, 'timing/data': 0.021048041060566902, 'timing/model': 1.7065432723611593, 'learning_rate/qwen_vl_interface': 7.874380698630982e-06, 'learning_rate/action_model': 7.874380698630982e-05, 'epoch': 1.34}) +[2026-06-15 23:14:45,155][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.15830008685588837, 'train/grad_norm_pre_clip': 6.723614692687988, 'timing/data': 0.020344066433608532, 'timing/model': 1.4605393884703517, 'learning_rate/qwen_vl_interface': 7.871841763778562e-06, 'learning_rate/action_model': 7.871841763778563e-05, 'epoch': 1.34}) +[2026-06-15 23:14:50,951][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.0792185589671135, 'train/grad_norm_pre_clip': 4.854667663574219, 'timing/data': 0.02102949097752571, 'timing/model': 1.4443786023184657, 'learning_rate/qwen_vl_interface': 7.86930175118603e-06, 'learning_rate/action_model': 7.86930175118603e-05, 'epoch': 1.34}) +[2026-06-15 23:14:56,756][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.14063537120819092, 'train/grad_norm_pre_clip': 5.881735324859619, 'timing/data': 0.020771268755197525, 'timing/model': 1.462856087833643, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 1.34}) +[2026-06-15 23:15:02,543][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.0773070752620697, 'train/grad_norm_pre_clip': 4.891670227050781, 'timing/data': 0.02137894369661808, 'timing/model': 1.4481443390250206, 'learning_rate/qwen_vl_interface': 7.86421849695748e-06, 'learning_rate/action_model': 7.864218496957481e-05, 'epoch': 1.35}) +[2026-06-15 23:15:08,343][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.3862853944301605, 'train/grad_norm_pre_clip': 12.06081771850586, 'timing/data': 0.021438182331621647, 'timing/model': 1.4364281045272946, 'learning_rate/qwen_vl_interface': 7.861675257410998e-06, 'learning_rate/action_model': 7.861675257410997e-05, 'epoch': 1.35}) +[2026-06-15 23:15:14,195][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.1986248791217804, 'train/grad_norm_pre_clip': 7.589334487915039, 'timing/data': 0.021103370934724808, 'timing/model': 1.7029218403622508, 'learning_rate/qwen_vl_interface': 7.859130944303468e-06, 'learning_rate/action_model': 7.859130944303467e-05, 'epoch': 1.35}) +[2026-06-15 23:15:19,757][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.0801648497581482, 'train/grad_norm_pre_clip': 2.9798760414123535, 'timing/data': 0.024964562617242336, 'timing/model': 1.4468271052464843, 'learning_rate/qwen_vl_interface': 7.856585558680761e-06, 'learning_rate/action_model': 7.856585558680761e-05, 'epoch': 1.35}) +[2026-06-15 23:15:25,442][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.37016773223876953, 'train/grad_norm_pre_clip': 5.61042594909668, 'timing/data': 0.02068743947893381, 'timing/model': 1.4241874227300286, 'learning_rate/qwen_vl_interface': 7.854039101589191e-06, 'learning_rate/action_model': 7.854039101589191e-05, 'epoch': 1.35}) +[2026-06-15 23:15:31,144][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.10675453394651413, 'train/grad_norm_pre_clip': 5.618865489959717, 'timing/data': 0.023888876661658287, 'timing/model': 1.4268479645252228, 'learning_rate/qwen_vl_interface': 7.851491574075509e-06, 'learning_rate/action_model': 7.851491574075509e-05, 'epoch': 1.35}) +[2026-06-15 23:15:36,840][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.15650728344917297, 'train/grad_norm_pre_clip': 7.215651035308838, 'timing/data': 0.020690410397946835, 'timing/model': 1.3985237311571836, 'learning_rate/qwen_vl_interface': 7.848942977186909e-06, 'learning_rate/action_model': 7.848942977186908e-05, 'epoch': 1.35}) +[2026-06-15 23:15:42,440][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.07105006277561188, 'train/grad_norm_pre_clip': 2.726125478744507, 'timing/data': 0.019615723751485348, 'timing/model': 1.419211408123374, 'learning_rate/qwen_vl_interface': 7.84639331197102e-06, 'learning_rate/action_model': 7.84639331197102e-05, 'epoch': 1.35}) +[2026-06-15 23:15:48,298][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.07040369510650635, 'train/grad_norm_pre_clip': 2.935122489929199, 'timing/data': 0.020696540363132954, 'timing/model': 1.6967055909335613, 'learning_rate/qwen_vl_interface': 7.843842579475918e-06, 'learning_rate/action_model': 7.843842579475918e-05, 'epoch': 1.35}) +[2026-06-15 23:15:53,971][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.109759122133255, 'train/grad_norm_pre_clip': 4.577450752258301, 'timing/data': 0.02011816669255495, 'timing/model': 1.483028451912105, 'learning_rate/qwen_vl_interface': 7.841290780750107e-06, 'learning_rate/action_model': 7.841290780750107e-05, 'epoch': 1.35}) +[2026-06-15 23:15:59,747][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.1245390996336937, 'train/grad_norm_pre_clip': 5.37194299697876, 'timing/data': 0.021530614234507084, 'timing/model': 1.4398013390600681, 'learning_rate/qwen_vl_interface': 7.838737916842538e-06, 'learning_rate/action_model': 7.838737916842538e-05, 'epoch': 1.35}) +[2026-06-15 23:16:05,544][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.09727261960506439, 'train/grad_norm_pre_clip': 5.667949676513672, 'timing/data': 0.02131212316453457, 'timing/model': 1.4449719032272696, 'learning_rate/qwen_vl_interface': 7.836183988802596e-06, 'learning_rate/action_model': 7.836183988802595e-05, 'epoch': 1.35}) +[2026-06-15 23:16:11,249][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.3335561156272888, 'train/grad_norm_pre_clip': 12.333035469055176, 'timing/data': 0.020455828867852688, 'timing/model': 1.4423545869067311, 'learning_rate/qwen_vl_interface': 7.833628997680107e-06, 'learning_rate/action_model': 7.833628997680107e-05, 'epoch': 1.35}) +[2026-06-15 23:16:17,045][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.21570730209350586, 'train/grad_norm_pre_clip': 8.668961524963379, 'timing/data': 0.0208266107365489, 'timing/model': 1.4081855416297913, 'learning_rate/qwen_vl_interface': 7.831072944525329e-06, 'learning_rate/action_model': 7.831072944525328e-05, 'epoch': 1.35}) +[2026-06-15 23:16:22,795][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.14441847801208496, 'train/grad_norm_pre_clip': 3.8607873916625977, 'timing/data': 0.022548790089786053, 'timing/model': 1.6797082088887691, 'learning_rate/qwen_vl_interface': 7.828515830388956e-06, 'learning_rate/action_model': 7.828515830388956e-05, 'epoch': 1.35}) +[2026-06-15 23:16:28,358][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.13838286697864532, 'train/grad_norm_pre_clip': 6.316659450531006, 'timing/data': 0.020758381113409996, 'timing/model': 1.4566881731152534, 'learning_rate/qwen_vl_interface': 7.825957656322129e-06, 'learning_rate/action_model': 7.825957656322128e-05, 'epoch': 1.35}) +[2026-06-15 23:16:34,147][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.2142878770828247, 'train/grad_norm_pre_clip': 7.53974723815918, 'timing/data': 0.020188207738101482, 'timing/model': 1.4306904608383775, 'learning_rate/qwen_vl_interface': 7.823398423376406e-06, 'learning_rate/action_model': 7.823398423376406e-05, 'epoch': 1.35}) +[2026-06-15 23:16:39,949][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.11461760103702545, 'train/grad_norm_pre_clip': 3.5967986583709717, 'timing/data': 0.02028040774166584, 'timing/model': 1.4507628362625837, 'learning_rate/qwen_vl_interface': 7.820838132603801e-06, 'learning_rate/action_model': 7.8208381326038e-05, 'epoch': 1.35}) +[2026-06-15 23:16:45,713][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.08972425758838654, 'train/grad_norm_pre_clip': 4.56673526763916, 'timing/data': 0.02407182939350605, 'timing/model': 1.4067526059225202, 'learning_rate/qwen_vl_interface': 7.818276785056748e-06, 'learning_rate/action_model': 7.818276785056749e-05, 'epoch': 1.35}) +[2026-06-15 23:16:51,570][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.19312497973442078, 'train/grad_norm_pre_clip': 8.231818199157715, 'timing/data': 0.020244977436959743, 'timing/model': 1.4285873798653483, 'learning_rate/qwen_vl_interface': 7.815714381788123e-06, 'learning_rate/action_model': 7.815714381788122e-05, 'epoch': 1.35}) +[2026-06-15 23:16:57,422][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.0705646276473999, 'train/grad_norm_pre_clip': 2.803297996520996, 'timing/data': 0.02070258092135191, 'timing/model': 1.7172181466594338, 'learning_rate/qwen_vl_interface': 7.813150923851227e-06, 'learning_rate/action_model': 7.813150923851227e-05, 'epoch': 1.35}) +[2026-06-15 23:17:03,045][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.06489240378141403, 'train/grad_norm_pre_clip': 7.619647979736328, 'timing/data': 0.02063425164669752, 'timing/model': 1.4517437061294913, 'learning_rate/qwen_vl_interface': 7.810586412299807e-06, 'learning_rate/action_model': 7.810586412299806e-05, 'epoch': 1.35}) +[2026-06-15 23:17:08,852][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.17089439928531647, 'train/grad_norm_pre_clip': 9.157896995544434, 'timing/data': 0.02099668327718973, 'timing/model': 1.4506494170054793, 'learning_rate/qwen_vl_interface': 7.808020848188034e-06, 'learning_rate/action_model': 7.808020848188034e-05, 'epoch': 1.36}) +[2026-06-15 23:17:14,659][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.12961779534816742, 'train/grad_norm_pre_clip': 6.369831562042236, 'timing/data': 0.020826141349971294, 'timing/model': 1.4635338122025132, 'learning_rate/qwen_vl_interface': 7.805454232570517e-06, 'learning_rate/action_model': 7.805454232570517e-05, 'epoch': 1.36}) +[2026-06-15 23:17:20,462][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.03414066135883331, 'train/grad_norm_pre_clip': 2.6255249977111816, 'timing/data': 0.024489802308380604, 'timing/model': 1.4575475566089153, 'learning_rate/qwen_vl_interface': 7.80288656650229e-06, 'learning_rate/action_model': 7.80288656650229e-05, 'epoch': 1.36}) +[2026-06-15 23:17:26,269][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.020348025485873222, 'train/grad_norm_pre_clip': 1.7240036725997925, 'timing/data': 0.020557421259582043, 'timing/model': 1.4028714979067445, 'learning_rate/qwen_vl_interface': 7.800317851038829e-06, 'learning_rate/action_model': 7.800317851038828e-05, 'epoch': 1.36}) +[2026-06-15 23:17:32,103][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.17851780354976654, 'train/grad_norm_pre_clip': 6.952347755432129, 'timing/data': 0.02078556176275015, 'timing/model': 1.6968385595828295, 'learning_rate/qwen_vl_interface': 7.797748087236033e-06, 'learning_rate/action_model': 7.797748087236033e-05, 'epoch': 1.36}) +[2026-06-15 23:17:37,757][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.27152031660079956, 'train/grad_norm_pre_clip': 7.108184337615967, 'timing/data': 0.02057984098792076, 'timing/model': 1.441927770152688, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 1.36}) +[2026-06-15 23:17:43,552][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.2163267284631729, 'train/grad_norm_pre_clip': 8.535411834716797, 'timing/data': 0.021230904385447502, 'timing/model': 1.4421644788235426, 'learning_rate/qwen_vl_interface': 7.792605418838196e-06, 'learning_rate/action_model': 7.792605418838197e-05, 'epoch': 1.36}) +[2026-06-15 23:17:49,355][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.33766013383865356, 'train/grad_norm_pre_clip': 10.560750007629395, 'timing/data': 0.021122093312442303, 'timing/model': 1.449963535182178, 'learning_rate/qwen_vl_interface': 7.790032516357116e-06, 'learning_rate/action_model': 7.790032516357117e-05, 'epoch': 1.36}) +[2026-06-15 23:17:55,167][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.10807580500841141, 'train/grad_norm_pre_clip': 7.453970909118652, 'timing/data': 0.023551697842776775, 'timing/model': 1.4587985491380095, 'learning_rate/qwen_vl_interface': 7.787458569764613e-06, 'learning_rate/action_model': 7.787458569764612e-05, 'epoch': 1.36}) +[2026-06-15 23:18:01,436][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.4483039081096649, 'train/grad_norm_pre_clip': 13.396249771118164, 'timing/data': 0.02147318609058857, 'timing/model': 1.4024444092065096, 'learning_rate/qwen_vl_interface': 7.784883580118738e-06, 'learning_rate/action_model': 7.784883580118738e-05, 'epoch': 1.36}) +[2026-06-15 23:18:07,404][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.15389201045036316, 'train/grad_norm_pre_clip': 7.6304240226745605, 'timing/data': 0.02467300370335579, 'timing/model': 1.7171522909775376, 'learning_rate/qwen_vl_interface': 7.782307548477975e-06, 'learning_rate/action_model': 7.782307548477975e-05, 'epoch': 1.36}) +[2026-06-15 23:18:13,242][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.01576017588376999, 'train/grad_norm_pre_clip': 1.0127414464950562, 'timing/data': 0.021489937789738178, 'timing/model': 1.6355640860274434, 'learning_rate/qwen_vl_interface': 7.779730475901233e-06, 'learning_rate/action_model': 7.779730475901232e-05, 'epoch': 1.36}) +[2026-06-15 23:18:18,863][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.28863412141799927, 'train/grad_norm_pre_clip': 10.37429428100586, 'timing/data': 0.021581467241048813, 'timing/model': 1.4466875465586782, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 1.36}) +[2026-06-15 23:18:24,658][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.1427886039018631, 'train/grad_norm_pre_clip': 6.454007148742676, 'timing/data': 0.022774402983486652, 'timing/model': 1.4341372242197394, 'learning_rate/qwen_vl_interface': 7.774573212177587e-06, 'learning_rate/action_model': 7.774573212177588e-05, 'epoch': 1.36}) +[2026-06-15 23:18:30,456][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.2643413245677948, 'train/grad_norm_pre_clip': 9.980393409729004, 'timing/data': 0.0222258809953928, 'timing/model': 1.44272106513381, 'learning_rate/qwen_vl_interface': 7.77199302315064e-06, 'learning_rate/action_model': 7.77199302315064e-05, 'epoch': 1.36}) +[2026-06-15 23:18:36,357][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.10938314348459244, 'train/grad_norm_pre_clip': 4.062047958374023, 'timing/data': 0.02583412081003189, 'timing/model': 1.4660310670733452, 'learning_rate/qwen_vl_interface': 7.769411797427622e-06, 'learning_rate/action_model': 7.769411797427623e-05, 'epoch': 1.36}) +[2026-06-15 23:18:42,132][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.31094157695770264, 'train/grad_norm_pre_clip': 4.951494216918945, 'timing/data': 0.02214411087334156, 'timing/model': 1.6669609667733312, 'learning_rate/qwen_vl_interface': 7.766829536069584e-06, 'learning_rate/action_model': 7.766829536069584e-05, 'epoch': 1.36}) +[2026-06-15 23:18:48,020][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.1380975991487503, 'train/grad_norm_pre_clip': 7.446136474609375, 'timing/data': 0.02123392466455698, 'timing/model': 1.6975310994312167, 'learning_rate/qwen_vl_interface': 7.764246240137991e-06, 'learning_rate/action_model': 7.764246240137991e-05, 'epoch': 1.36}) +[2026-06-15 23:18:53,655][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.0864766389131546, 'train/grad_norm_pre_clip': 5.865062236785889, 'timing/data': 0.020717153325676918, 'timing/model': 1.4288615696132183, 'learning_rate/qwen_vl_interface': 7.761661910694741e-06, 'learning_rate/action_model': 7.761661910694741e-05, 'epoch': 1.36}) +[2026-06-15 23:18:59,459][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.11810560524463654, 'train/grad_norm_pre_clip': 6.595846652984619, 'timing/data': 0.02119460515677929, 'timing/model': 1.4564060857519507, 'learning_rate/qwen_vl_interface': 7.759076548802153e-06, 'learning_rate/action_model': 7.759076548802152e-05, 'epoch': 1.36}) +[2026-06-15 23:19:05,278][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.0721198096871376, 'train/grad_norm_pre_clip': 3.8859074115753174, 'timing/data': 0.02121565490961075, 'timing/model': 1.474195177666843, 'learning_rate/qwen_vl_interface': 7.75649015552297e-06, 'learning_rate/action_model': 7.75649015552297e-05, 'epoch': 1.36}) +[2026-06-15 23:19:11,076][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.06683351844549179, 'train/grad_norm_pre_clip': 7.03172492980957, 'timing/data': 0.022240771912038326, 'timing/model': 1.404709853231907, 'learning_rate/qwen_vl_interface': 7.753902731920364e-06, 'learning_rate/action_model': 7.753902731920364e-05, 'epoch': 1.36}) +[2026-06-15 23:19:16,901][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.2833103537559509, 'train/grad_norm_pre_clip': 12.669702529907227, 'timing/data': 0.020638003014028072, 'timing/model': 1.7024162951856852, 'learning_rate/qwen_vl_interface': 7.751314279057925e-06, 'learning_rate/action_model': 7.751314279057923e-05, 'epoch': 1.37}) +[2026-06-15 23:19:22,551][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.2932487726211548, 'train/grad_norm_pre_clip': 4.494478225708008, 'timing/data': 0.022570833563804626, 'timing/model': 1.4487197427079082, 'learning_rate/qwen_vl_interface': 7.74872479799967e-06, 'learning_rate/action_model': 7.74872479799967e-05, 'epoch': 1.37}) +[2026-06-15 23:19:28,360][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.12119044363498688, 'train/grad_norm_pre_clip': 4.970800876617432, 'timing/data': 0.023527037352323532, 'timing/model': 1.43177046533674, 'learning_rate/qwen_vl_interface': 7.746134289810034e-06, 'learning_rate/action_model': 7.746134289810033e-05, 'epoch': 1.37}) +[2026-06-15 23:19:34,260][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.17908483743667603, 'train/grad_norm_pre_clip': 7.615165710449219, 'timing/data': 0.021751098334789276, 'timing/model': 1.4467862118035555, 'learning_rate/qwen_vl_interface': 7.74354275555388e-06, 'learning_rate/action_model': 7.74354275555388e-05, 'epoch': 1.37}) +[2026-06-15 23:19:40,154][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.2535533905029297, 'train/grad_norm_pre_clip': 7.0724310874938965, 'timing/data': 0.02158338762819767, 'timing/model': 1.4838955299928784, 'learning_rate/qwen_vl_interface': 7.740950196296488e-06, 'learning_rate/action_model': 7.740950196296488e-05, 'epoch': 1.37}) +[2026-06-15 23:19:45,961][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.08217804878950119, 'train/grad_norm_pre_clip': 11.255489349365234, 'timing/data': 0.029124178923666477, 'timing/model': 1.4516094885766506, 'learning_rate/qwen_vl_interface': 7.738356613103562e-06, 'learning_rate/action_model': 7.73835661310356e-05, 'epoch': 1.37}) +[2026-06-15 23:19:51,784][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.09148809313774109, 'train/grad_norm_pre_clip': 5.081445693969727, 'timing/data': 0.022849174216389656, 'timing/model': 1.6779151447117329, 'learning_rate/qwen_vl_interface': 7.735762007041226e-06, 'learning_rate/action_model': 7.735762007041226e-05, 'epoch': 1.37}) +[2026-06-15 23:19:57,454][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.2979407012462616, 'train/grad_norm_pre_clip': 7.506483554840088, 'timing/data': 0.02061697281897068, 'timing/model': 1.4334067683666945, 'learning_rate/qwen_vl_interface': 7.733166379176023e-06, 'learning_rate/action_model': 7.733166379176023e-05, 'epoch': 1.37}) +[2026-06-15 23:20:03,352][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.036832295358181, 'train/grad_norm_pre_clip': 2.818981170654297, 'timing/data': 0.021301066502928734, 'timing/model': 1.4485105350613594, 'learning_rate/qwen_vl_interface': 7.730569730574921e-06, 'learning_rate/action_model': 7.73056973057492e-05, 'epoch': 1.37}) +[2026-06-15 23:20:09,153][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.14454680681228638, 'train/grad_norm_pre_clip': 7.173518180847168, 'timing/data': 0.021468017250299454, 'timing/model': 1.4517044369131327, 'learning_rate/qwen_vl_interface': 7.727972062305299e-06, 'learning_rate/action_model': 7.727972062305299e-05, 'epoch': 1.37}) +[2026-06-15 23:20:14,968][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.12913629412651062, 'train/grad_norm_pre_clip': 6.0522613525390625, 'timing/data': 0.025955243036150932, 'timing/model': 1.4071740759536624, 'learning_rate/qwen_vl_interface': 7.725373375434965e-06, 'learning_rate/action_model': 7.725373375434965e-05, 'epoch': 1.37}) +[2026-06-15 23:20:20,754][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.14231270551681519, 'train/grad_norm_pre_clip': 6.050718784332275, 'timing/data': 0.021113695576786995, 'timing/model': 1.402721956372261, 'learning_rate/qwen_vl_interface': 7.722773671032141e-06, 'learning_rate/action_model': 7.722773671032142e-05, 'epoch': 1.37}) +[2026-06-15 23:20:26,573][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.13381369411945343, 'train/grad_norm_pre_clip': 5.115686416625977, 'timing/data': 0.02159470785409212, 'timing/model': 1.6587754404172301, 'learning_rate/qwen_vl_interface': 7.720172950165466e-06, 'learning_rate/action_model': 7.720172950165465e-05, 'epoch': 1.37}) +[2026-06-15 23:20:32,152][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.04006613790988922, 'train/grad_norm_pre_clip': 2.89639949798584, 'timing/data': 0.021704719401896, 'timing/model': 1.431842784397304, 'learning_rate/qwen_vl_interface': 7.717571213903996e-06, 'learning_rate/action_model': 7.717571213903996e-05, 'epoch': 1.37}) +[2026-06-15 23:20:37,954][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.0629270076751709, 'train/grad_norm_pre_clip': 4.2210564613342285, 'timing/data': 0.021080766804516315, 'timing/model': 1.4497323967516422, 'learning_rate/qwen_vl_interface': 7.714968463317207e-06, 'learning_rate/action_model': 7.714968463317207e-05, 'epoch': 1.37}) +[2026-06-15 23:31:18,666][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.1793469935655594, 'train/grad_norm_pre_clip': 6.734911918640137, 'eval/loss': 0.0750596895813942, 'eval/flap_precision': 0.8975609756097561, 'eval/flap_recall': 0.6690909090909091, 'eval/flap_f1': 0.7666666666666666, 'eval/flap_rate_model': 0.05694444444444444, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7666666666666666, 'timing/data': 0.02065084408968687, 'timing/model': 1.4507485972717404, 'learning_rate/qwen_vl_interface': 7.712364699474996e-06, 'learning_rate/action_model': 7.712364699474996e-05, 'epoch': 1.37}) +[2026-06-15 23:31:18,667][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3000_state +[2026-06-15 23:31:18,673][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:31:26,440][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3000_state/model.safetensors +[2026-06-15 23:31:37,527][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3000_state/optimizer.bin +[2026-06-15 23:31:37,528][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3000_state/sampler.bin +[2026-06-15 23:31:37,528][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-15 23:31:37,530][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-15 23:31:37,704][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-15 23:31:37,704][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:31:37,719][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:31:43,445][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.3224290907382965, 'train/grad_norm_pre_clip': 6.257070064544678, 'timing/data': 0.0255332849919796, 'timing/model': 1.4495292622596025, 'learning_rate/qwen_vl_interface': 7.709759923447667e-06, 'learning_rate/action_model': 7.709759923447665e-05, 'epoch': 1.37}) +[2026-06-15 23:31:49,025][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.10141291469335556, 'train/grad_norm_pre_clip': 3.2032089233398438, 'timing/data': 0.021070340648293495, 'timing/model': 1.3770965421572328, 'learning_rate/qwen_vl_interface': 7.707154136305945e-06, 'learning_rate/action_model': 7.707154136305945e-05, 'epoch': 1.37}) +[2026-06-15 23:31:54,622][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.16941297054290771, 'train/grad_norm_pre_clip': 9.623513221740723, 'timing/data': 0.02162693254649639, 'timing/model': 1.4363033100962639, 'learning_rate/qwen_vl_interface': 7.704547339120974e-06, 'learning_rate/action_model': 7.704547339120974e-05, 'epoch': 1.37}) +[2026-06-15 23:32:00,383][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.12727032601833344, 'train/grad_norm_pre_clip': 6.729039192199707, 'timing/data': 0.022147136740386486, 'timing/model': 1.6848184056580067, 'learning_rate/qwen_vl_interface': 7.70193953296431e-06, 'learning_rate/action_model': 7.70193953296431e-05, 'epoch': 1.37}) +[2026-06-15 23:32:05,936][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.15947569906711578, 'train/grad_norm_pre_clip': 7.435019493103027, 'timing/data': 0.02613925840705633, 'timing/model': 1.4348525134846568, 'learning_rate/qwen_vl_interface': 7.69933071890792e-06, 'learning_rate/action_model': 7.699330718907919e-05, 'epoch': 1.37}) +[2026-06-15 23:32:11,761][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.22620432078838348, 'train/grad_norm_pre_clip': 10.307125091552734, 'timing/data': 0.022331307642161846, 'timing/model': 1.469546633772552, 'learning_rate/qwen_vl_interface': 7.696720898024188e-06, 'learning_rate/action_model': 7.69672089802419e-05, 'epoch': 1.37}) +[2026-06-15 23:32:17,542][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.19924037158489227, 'train/grad_norm_pre_clip': 5.366486072540283, 'timing/data': 0.02490533236414194, 'timing/model': 1.4384100260213017, 'learning_rate/qwen_vl_interface': 7.69411007138592e-06, 'learning_rate/action_model': 7.694110071385919e-05, 'epoch': 1.38}) +[2026-06-15 23:32:23,356][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.0925159677863121, 'train/grad_norm_pre_clip': 2.9240105152130127, 'timing/data': 0.020448805764317513, 'timing/model': 1.4636075906455517, 'learning_rate/qwen_vl_interface': 7.691498240066322e-06, 'learning_rate/action_model': 7.691498240066322e-05, 'epoch': 1.38}) +[2026-06-15 23:32:29,140][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.11489574611186981, 'train/grad_norm_pre_clip': 6.036627769470215, 'timing/data': 0.02161709312349558, 'timing/model': 1.4062516754493117, 'learning_rate/qwen_vl_interface': 7.688885405139022e-06, 'learning_rate/action_model': 7.688885405139021e-05, 'epoch': 1.38}) +[2026-06-15 23:32:34,985][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.1629645973443985, 'train/grad_norm_pre_clip': 9.224275588989258, 'timing/data': 0.0204255860298872, 'timing/model': 1.685547098517418, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 1.38}) +[2026-06-15 23:32:40,539][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.25658735632896423, 'train/grad_norm_pre_clip': 7.269796371459961, 'timing/data': 0.021094939671456814, 'timing/model': 1.4385685632005334, 'learning_rate/qwen_vl_interface': 7.683656728757873e-06, 'learning_rate/action_model': 7.683656728757872e-05, 'epoch': 1.38}) +[2026-06-15 23:32:46,248][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.17407263815402985, 'train/grad_norm_pre_clip': 6.678823471069336, 'timing/data': 0.021382732316851616, 'timing/model': 1.4350799322128296, 'learning_rate/qwen_vl_interface': 7.681040889453337e-06, 'learning_rate/action_model': 7.681040889453336e-05, 'epoch': 1.38}) +[2026-06-15 23:32:51,942][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.1306503266096115, 'train/grad_norm_pre_clip': 8.218730926513672, 'timing/data': 0.020611677318811417, 'timing/model': 1.4517581881955266, 'learning_rate/qwen_vl_interface': 7.678424050839722e-06, 'learning_rate/action_model': 7.67842405083972e-05, 'epoch': 1.38}) +[2026-06-15 23:32:57,643][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.05043705552816391, 'train/grad_norm_pre_clip': 2.5090274810791016, 'timing/data': 0.02174779400229454, 'timing/model': 1.4065697817131877, 'learning_rate/qwen_vl_interface': 7.675806213992706e-06, 'learning_rate/action_model': 7.675806213992707e-05, 'epoch': 1.38}) +[2026-06-15 23:33:03,344][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.187979593873024, 'train/grad_norm_pre_clip': 8.684624671936035, 'timing/data': 0.02123455051332712, 'timing/model': 1.4103972855955362, 'learning_rate/qwen_vl_interface': 7.673187379988388e-06, 'learning_rate/action_model': 7.673187379988388e-05, 'epoch': 1.38}) +[2026-06-15 23:33:09,094][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.1026943102478981, 'train/grad_norm_pre_clip': 3.6623709201812744, 'timing/data': 0.020089765079319477, 'timing/model': 1.695757882669568, 'learning_rate/qwen_vl_interface': 7.67056754990327e-06, 'learning_rate/action_model': 7.670567549903269e-05, 'epoch': 1.38}) +[2026-06-15 23:33:14,645][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.223977193236351, 'train/grad_norm_pre_clip': 5.851111888885498, 'timing/data': 0.021426512859761715, 'timing/model': 1.4499542014673352, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 1.38}) +[2026-06-15 23:33:20,464][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.16508843004703522, 'train/grad_norm_pre_clip': 7.296792507171631, 'timing/data': 0.02189173549413681, 'timing/model': 1.4492063485085964, 'learning_rate/qwen_vl_interface': 7.665324905798694e-06, 'learning_rate/action_model': 7.665324905798695e-05, 'epoch': 1.38}) +[2026-06-15 23:33:26,263][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.05332019925117493, 'train/grad_norm_pre_clip': 2.35485577583313, 'timing/data': 0.021461792290210724, 'timing/model': 1.4728488652035594, 'learning_rate/qwen_vl_interface': 7.66270209393429e-06, 'learning_rate/action_model': 7.66270209393429e-05, 'epoch': 1.38}) +[2026-06-15 23:33:31,955][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.09247411787509918, 'train/grad_norm_pre_clip': 4.798068523406982, 'timing/data': 0.020373125560581684, 'timing/model': 1.4415103876963258, 'learning_rate/qwen_vl_interface': 7.66007829029919e-06, 'learning_rate/action_model': 7.660078290299191e-05, 'epoch': 1.38}) +[2026-06-15 23:33:37,757][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.12022348493337631, 'train/grad_norm_pre_clip': 3.151386022567749, 'timing/data': 0.02106103952974081, 'timing/model': 1.3984388317912817, 'learning_rate/qwen_vl_interface': 7.657453495971942e-06, 'learning_rate/action_model': 7.657453495971943e-05, 'epoch': 1.38}) +[2026-06-15 23:33:43,597][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.22084756195545197, 'train/grad_norm_pre_clip': 4.754311561584473, 'timing/data': 0.022229507565498352, 'timing/model': 1.6859030835330486, 'learning_rate/qwen_vl_interface': 7.6548277120315e-06, 'learning_rate/action_model': 7.6548277120315e-05, 'epoch': 1.38}) +[2026-06-15 23:33:49,159][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.09901751577854156, 'train/grad_norm_pre_clip': 4.02047061920166, 'timing/data': 0.02196956519037485, 'timing/model': 1.4551094472408295, 'learning_rate/qwen_vl_interface': 7.652200939557225e-06, 'learning_rate/action_model': 7.652200939557225e-05, 'epoch': 1.38}) +[2026-06-15 23:33:54,852][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.24706465005874634, 'train/grad_norm_pre_clip': 6.919869899749756, 'timing/data': 0.022816370241343975, 'timing/model': 1.428455125540495, 'learning_rate/qwen_vl_interface': 7.649573179628881e-06, 'learning_rate/action_model': 7.649573179628881e-05, 'epoch': 1.38}) +[2026-06-15 23:34:00,558][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.19716566801071167, 'train/grad_norm_pre_clip': 7.692572593688965, 'timing/data': 0.021125510334968567, 'timing/model': 1.447048063389957, 'learning_rate/qwen_vl_interface': 7.646944433326645e-06, 'learning_rate/action_model': 7.646944433326643e-05, 'epoch': 1.38}) +[2026-06-15 23:34:06,352][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.08634492009878159, 'train/grad_norm_pre_clip': 4.473513126373291, 'timing/data': 0.02077619731426239, 'timing/model': 1.445711045525968, 'learning_rate/qwen_vl_interface': 7.64431470173109e-06, 'learning_rate/action_model': 7.64431470173109e-05, 'epoch': 1.38}) +[2026-06-15 23:34:12,043][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.05394262447953224, 'train/grad_norm_pre_clip': 4.257093906402588, 'timing/data': 0.020376847125589848, 'timing/model': 1.410196952521801, 'learning_rate/qwen_vl_interface': 7.641683985923203e-06, 'learning_rate/action_model': 7.641683985923203e-05, 'epoch': 1.38}) +[2026-06-15 23:34:17,795][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.12817150354385376, 'train/grad_norm_pre_clip': 10.410372734069824, 'timing/data': 0.01936511043459177, 'timing/model': 1.6939031351357698, 'learning_rate/qwen_vl_interface': 7.63905228698437e-06, 'learning_rate/action_model': 7.63905228698437e-05, 'epoch': 1.38}) +[2026-06-15 23:34:23,447][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.1492239385843277, 'train/grad_norm_pre_clip': 10.854496955871582, 'timing/data': 0.020440696738660336, 'timing/model': 1.4351276578381658, 'learning_rate/qwen_vl_interface': 7.636419605996385e-06, 'learning_rate/action_model': 7.636419605996384e-05, 'epoch': 1.39}) +[2026-06-15 23:34:29,148][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.1926625519990921, 'train/grad_norm_pre_clip': 10.964653015136719, 'timing/data': 0.02054381649941206, 'timing/model': 1.3902525724843144, 'learning_rate/qwen_vl_interface': 7.633785944041444e-06, 'learning_rate/action_model': 7.633785944041444e-05, 'epoch': 1.39}) +[2026-06-15 23:34:34,943][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.10792455822229385, 'train/grad_norm_pre_clip': 4.580998420715332, 'timing/data': 0.020063964650034904, 'timing/model': 1.4588839150965214, 'learning_rate/qwen_vl_interface': 7.631151302202145e-06, 'learning_rate/action_model': 7.631151302202144e-05, 'epoch': 1.39}) +[2026-06-15 23:34:40,740][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.04233966022729874, 'train/grad_norm_pre_clip': 4.662910461425781, 'timing/data': 0.020233456045389175, 'timing/model': 1.4482112154364586, 'learning_rate/qwen_vl_interface': 7.628515681561488e-06, 'learning_rate/action_model': 7.628515681561487e-05, 'epoch': 1.39}) +[2026-06-15 23:34:46,446][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.1207280158996582, 'train/grad_norm_pre_clip': 7.3625874519348145, 'timing/data': 0.02040368691086769, 'timing/model': 1.3927467288449407, 'learning_rate/qwen_vl_interface': 7.62587908320288e-06, 'learning_rate/action_model': 7.62587908320288e-05, 'epoch': 1.39}) +[2026-06-15 23:34:52,182][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.0620584674179554, 'train/grad_norm_pre_clip': 3.9102325439453125, 'timing/data': 0.020522506907582283, 'timing/model': 1.6749468222260475, 'learning_rate/qwen_vl_interface': 7.6232415082101255e-06, 'learning_rate/action_model': 7.623241508210124e-05, 'epoch': 1.39}) +[2026-06-15 23:34:57,842][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.21153488755226135, 'train/grad_norm_pre_clip': 6.450393199920654, 'timing/data': 0.020337115973234177, 'timing/model': 1.4310564287006855, 'learning_rate/qwen_vl_interface': 7.620602957667432e-06, 'learning_rate/action_model': 7.620602957667432e-05, 'epoch': 1.39}) +[2026-06-15 23:35:03,649][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.08878008276224136, 'train/grad_norm_pre_clip': 5.699187755584717, 'timing/data': 0.021230601705610752, 'timing/model': 1.4391048764809966, 'learning_rate/qwen_vl_interface': 7.617963432659408e-06, 'learning_rate/action_model': 7.617963432659408e-05, 'epoch': 1.39}) +[2026-06-15 23:35:09,450][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.31987619400024414, 'train/grad_norm_pre_clip': 8.238225936889648, 'timing/data': 0.02149820327758789, 'timing/model': 1.435937668196857, 'learning_rate/qwen_vl_interface': 7.615322934271063e-06, 'learning_rate/action_model': 7.615322934271064e-05, 'epoch': 1.39}) +[2026-06-15 23:35:15,251][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.15370634198188782, 'train/grad_norm_pre_clip': 6.868315696716309, 'timing/data': 0.02067532856017351, 'timing/model': 1.4435681933537126, 'learning_rate/qwen_vl_interface': 7.6126814635878055e-06, 'learning_rate/action_model': 7.612681463587805e-05, 'epoch': 1.39}) +[2026-06-15 23:35:21,072][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.1473570168018341, 'train/grad_norm_pre_clip': 9.918994903564453, 'timing/data': 0.021165940910577774, 'timing/model': 1.4004286769777536, 'learning_rate/qwen_vl_interface': 7.6100390216954445e-06, 'learning_rate/action_model': 7.610039021695445e-05, 'epoch': 1.39}) +[2026-06-15 23:35:26,905][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.058045394718647, 'train/grad_norm_pre_clip': 3.2038469314575195, 'timing/data': 0.021012689918279648, 'timing/model': 1.6852984866127372, 'learning_rate/qwen_vl_interface': 7.607395609680189e-06, 'learning_rate/action_model': 7.60739560968019e-05, 'epoch': 1.39}) +[2026-06-15 23:35:32,732][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.19413937628269196, 'train/grad_norm_pre_clip': 5.269139289855957, 'timing/data': 0.021181531250476837, 'timing/model': 1.6245320606976748, 'learning_rate/qwen_vl_interface': 7.604751228628644e-06, 'learning_rate/action_model': 7.604751228628644e-05, 'epoch': 1.39}) +[2026-06-15 23:35:38,351][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.26563775539398193, 'train/grad_norm_pre_clip': 8.553590774536133, 'timing/data': 0.020531496964395046, 'timing/model': 1.4362247548997402, 'learning_rate/qwen_vl_interface': 7.602105879627814e-06, 'learning_rate/action_model': 7.602105879627814e-05, 'epoch': 1.39}) +[2026-06-15 23:35:44,151][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.20448258519172668, 'train/grad_norm_pre_clip': 6.933444023132324, 'timing/data': 0.020981510169804096, 'timing/model': 1.4497888321056962, 'learning_rate/qwen_vl_interface': 7.599459563765107e-06, 'learning_rate/action_model': 7.599459563765106e-05, 'epoch': 1.39}) +[2026-06-15 23:35:49,960][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.17991286516189575, 'train/grad_norm_pre_clip': 10.061630249023438, 'timing/data': 0.0217298436909914, 'timing/model': 1.4564361609518528, 'learning_rate/qwen_vl_interface': 7.596812282128319e-06, 'learning_rate/action_model': 7.596812282128319e-05, 'epoch': 1.39}) +[2026-06-15 23:35:55,770][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.24843215942382812, 'train/grad_norm_pre_clip': 7.425058364868164, 'timing/data': 0.021972346119582653, 'timing/model': 1.3949871007353067, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 1.39}) +[2026-06-15 23:36:01,608][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.21606379747390747, 'train/grad_norm_pre_clip': 6.0793070793151855, 'timing/data': 0.02345943357795477, 'timing/model': 1.6851467406377196, 'learning_rate/qwen_vl_interface': 7.59151482588569e-06, 'learning_rate/action_model': 7.59151482588569e-05, 'epoch': 1.39}) +[2026-06-15 23:36:07,494][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.09965395927429199, 'train/grad_norm_pre_clip': 5.554380416870117, 'timing/data': 0.021017530001699924, 'timing/model': 1.6763873416930437, 'learning_rate/qwen_vl_interface': 7.588864653457433e-06, 'learning_rate/action_model': 7.588864653457432e-05, 'epoch': 1.39}) +[2026-06-15 23:36:13,156][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.2817043364048004, 'train/grad_norm_pre_clip': 7.478021621704102, 'timing/data': 0.02063364814966917, 'timing/model': 1.4592285696417093, 'learning_rate/qwen_vl_interface': 7.5862135196102634e-06, 'learning_rate/action_model': 7.586213519610263e-05, 'epoch': 1.39}) +[2026-06-15 23:36:18,943][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.21755824983119965, 'train/grad_norm_pre_clip': 8.306979179382324, 'timing/data': 0.019937624223530293, 'timing/model': 1.436552070081234, 'learning_rate/qwen_vl_interface': 7.583561425433962e-06, 'learning_rate/action_model': 7.583561425433962e-05, 'epoch': 1.39}) +[2026-06-15 23:36:24,748][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.15355151891708374, 'train/grad_norm_pre_clip': 5.38160514831543, 'timing/data': 0.02101023029536009, 'timing/model': 1.4196752347052097, 'learning_rate/qwen_vl_interface': 7.580908372018706e-06, 'learning_rate/action_model': 7.580908372018705e-05, 'epoch': 1.39}) +[2026-06-15 23:36:30,574][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.10624262690544128, 'train/grad_norm_pre_clip': 3.073343515396118, 'timing/data': 0.021355662494897842, 'timing/model': 1.402026103809476, 'learning_rate/qwen_vl_interface': 7.5782543604550655e-06, 'learning_rate/action_model': 7.578254360455064e-05, 'epoch': 1.4}) +[2026-06-15 23:36:36,411][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.21006295084953308, 'train/grad_norm_pre_clip': 5.424097537994385, 'timing/data': 0.021044251509010792, 'timing/model': 1.707201031036675, 'learning_rate/qwen_vl_interface': 7.575599391834e-06, 'learning_rate/action_model': 7.575599391834e-05, 'epoch': 1.4}) +[2026-06-15 23:36:42,056][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.11084356158971786, 'train/grad_norm_pre_clip': 4.212338924407959, 'timing/data': 0.021371033042669296, 'timing/model': 1.450200011022389, 'learning_rate/qwen_vl_interface': 7.572943467246871e-06, 'learning_rate/action_model': 7.57294346724687e-05, 'epoch': 1.4}) +[2026-06-15 23:36:47,857][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.13175451755523682, 'train/grad_norm_pre_clip': 4.273583889007568, 'timing/data': 0.02136287186294794, 'timing/model': 1.436812194995582, 'learning_rate/qwen_vl_interface': 7.570286587785428e-06, 'learning_rate/action_model': 7.570286587785428e-05, 'epoch': 1.4}) +[2026-06-15 23:36:53,655][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.18290503323078156, 'train/grad_norm_pre_clip': 4.838438510894775, 'timing/data': 0.02079566940665245, 'timing/model': 1.4530077278614044, 'learning_rate/qwen_vl_interface': 7.567628754541815e-06, 'learning_rate/action_model': 7.567628754541814e-05, 'epoch': 1.4}) +[2026-06-15 23:36:59,451][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.11065187305212021, 'train/grad_norm_pre_clip': 4.092935562133789, 'timing/data': 0.021409022621810436, 'timing/model': 1.411297170445323, 'learning_rate/qwen_vl_interface': 7.564969968608565e-06, 'learning_rate/action_model': 7.564969968608565e-05, 'epoch': 1.4}) +[2026-06-15 23:37:05,457][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.05867480859160423, 'train/grad_norm_pre_clip': 4.113857746124268, 'timing/data': 0.02222149819135666, 'timing/model': 1.4510833388194442, 'learning_rate/qwen_vl_interface': 7.562310231078605e-06, 'learning_rate/action_model': 7.562310231078604e-05, 'epoch': 1.4}) +[2026-06-15 23:37:11,254][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.1402890831232071, 'train/grad_norm_pre_clip': 6.391077518463135, 'timing/data': 0.02129583153873682, 'timing/model': 1.6631299741566181, 'learning_rate/qwen_vl_interface': 7.559649543045252e-06, 'learning_rate/action_model': 7.559649543045252e-05, 'epoch': 1.4}) +[2026-06-15 23:37:16,845][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.2650555968284607, 'train/grad_norm_pre_clip': 9.40919303894043, 'timing/data': 0.02037950698286295, 'timing/model': 1.438356096856296, 'learning_rate/qwen_vl_interface': 7.556987905602215e-06, 'learning_rate/action_model': 7.556987905602215e-05, 'epoch': 1.4}) +[2026-06-15 23:37:22,647][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.3237834870815277, 'train/grad_norm_pre_clip': 7.240597248077393, 'timing/data': 0.020149726420640945, 'timing/model': 1.438003876246512, 'learning_rate/qwen_vl_interface': 7.554325319843593e-06, 'learning_rate/action_model': 7.554325319843592e-05, 'epoch': 1.4}) +[2026-06-15 23:37:28,452][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.26681309938430786, 'train/grad_norm_pre_clip': 16.8587646484375, 'timing/data': 0.020895730704069138, 'timing/model': 1.449861933477223, 'learning_rate/qwen_vl_interface': 7.551661786863874e-06, 'learning_rate/action_model': 7.551661786863874e-05, 'epoch': 1.4}) +[2026-06-15 23:37:34,149][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.24800246953964233, 'train/grad_norm_pre_clip': 9.234480857849121, 'timing/data': 0.020113174803555012, 'timing/model': 1.3961912784725428, 'learning_rate/qwen_vl_interface': 7.548997307757936e-06, 'learning_rate/action_model': 7.548997307757935e-05, 'epoch': 1.4}) +[2026-06-15 23:37:39,947][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.23837216198444366, 'train/grad_norm_pre_clip': 6.532848834991455, 'timing/data': 0.020330775529146194, 'timing/model': 1.4046153156086802, 'learning_rate/qwen_vl_interface': 7.546331883621044e-06, 'learning_rate/action_model': 7.546331883621044e-05, 'epoch': 1.4}) +[2026-06-15 23:37:45,671][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.05823099985718727, 'train/grad_norm_pre_clip': 2.1516082286834717, 'timing/data': 0.020012575201690197, 'timing/model': 1.677117477171123, 'learning_rate/qwen_vl_interface': 7.543665515548857e-06, 'learning_rate/action_model': 7.543665515548856e-05, 'epoch': 1.4}) +[2026-06-15 23:37:51,245][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.17668680846691132, 'train/grad_norm_pre_clip': 5.638470649719238, 'timing/data': 0.02147266361862421, 'timing/model': 1.4504908295348287, 'learning_rate/qwen_vl_interface': 7.540998204637413e-06, 'learning_rate/action_model': 7.540998204637413e-05, 'epoch': 1.4}) +[2026-06-15 23:37:57,045][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.10781757533550262, 'train/grad_norm_pre_clip': 6.520139694213867, 'timing/data': 0.01943008229136467, 'timing/model': 1.4366406006738544, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 1.4}) +[2026-06-15 23:38:02,855][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.19045491516590118, 'train/grad_norm_pre_clip': 5.584471702575684, 'timing/data': 0.019554671831429005, 'timing/model': 1.457556082867086, 'learning_rate/qwen_vl_interface': 7.535660758682877e-06, 'learning_rate/action_model': 7.535660758682876e-05, 'epoch': 1.4}) +[2026-06-15 23:38:08,653][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.13074244558811188, 'train/grad_norm_pre_clip': 4.836636066436768, 'timing/data': 0.020411807112395763, 'timing/model': 1.3960383394733071, 'learning_rate/qwen_vl_interface': 7.532990625833806e-06, 'learning_rate/action_model': 7.532990625833805e-05, 'epoch': 1.4}) +[2026-06-15 23:38:14,394][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.15899185836315155, 'train/grad_norm_pre_clip': 6.256083965301514, 'timing/data': 0.020037125796079636, 'timing/model': 1.6852503763511777, 'learning_rate/qwen_vl_interface': 7.5303195545335225e-06, 'learning_rate/action_model': 7.530319554533522e-05, 'epoch': 1.4}) +[2026-06-15 23:38:20,044][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.03751029819250107, 'train/grad_norm_pre_clip': 3.505344867706299, 'timing/data': 0.020306546241044998, 'timing/model': 1.4503090316429734, 'learning_rate/qwen_vl_interface': 7.527647545880007e-06, 'learning_rate/action_model': 7.527647545880007e-05, 'epoch': 1.4}) +[2026-06-15 23:38:25,845][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.16272222995758057, 'train/grad_norm_pre_clip': 5.041598320007324, 'timing/data': 0.020647947676479816, 'timing/model': 1.4517654813826084, 'learning_rate/qwen_vl_interface': 7.52497460097162e-06, 'learning_rate/action_model': 7.52497460097162e-05, 'epoch': 1.4}) +[2026-06-15 23:38:31,550][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.012834982946515083, 'train/grad_norm_pre_clip': 1.4339534044265747, 'timing/data': 0.020484006963670254, 'timing/model': 1.4557415144518018, 'learning_rate/qwen_vl_interface': 7.522300720907105e-06, 'learning_rate/action_model': 7.522300720907105e-05, 'epoch': 1.4}) +[2026-06-15 23:38:37,255][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.128940150141716, 'train/grad_norm_pre_clip': 6.094630718231201, 'timing/data': 0.023992598056793213, 'timing/model': 1.4468368031084538, 'learning_rate/qwen_vl_interface': 7.519625906785601e-06, 'learning_rate/action_model': 7.5196259067856e-05, 'epoch': 1.41}) +[2026-06-15 23:38:43,062][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.030511947348713875, 'train/grad_norm_pre_clip': 2.720672130584717, 'timing/data': 0.023689146153628826, 'timing/model': 1.4541740054264665, 'learning_rate/qwen_vl_interface': 7.516950159706614e-06, 'learning_rate/action_model': 7.516950159706614e-05, 'epoch': 1.41}) +[2026-06-15 23:38:48,908][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.03769367188215256, 'train/grad_norm_pre_clip': 2.450860023498535, 'timing/data': 0.020580727607011795, 'timing/model': 1.7120376750826836, 'learning_rate/qwen_vl_interface': 7.514273480770047e-06, 'learning_rate/action_model': 7.514273480770047e-05, 'epoch': 1.41}) +[2026-06-15 23:38:54,558][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.06670626997947693, 'train/grad_norm_pre_clip': 5.308409690856934, 'timing/data': 0.02388700656592846, 'timing/model': 1.4447724735364318, 'learning_rate/qwen_vl_interface': 7.511595871076181e-06, 'learning_rate/action_model': 7.511595871076181e-05, 'epoch': 1.41}) +[2026-06-15 23:39:00,358][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.26522085070610046, 'train/grad_norm_pre_clip': 5.815705299377441, 'timing/data': 0.02069552894681692, 'timing/model': 1.4491027481853962, 'learning_rate/qwen_vl_interface': 7.50891733172568e-06, 'learning_rate/action_model': 7.50891733172568e-05, 'epoch': 1.41}) +[2026-06-15 23:39:06,152][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.1428624838590622, 'train/grad_norm_pre_clip': 7.795318603515625, 'timing/data': 0.021213872358202934, 'timing/model': 1.4541783574968576, 'learning_rate/qwen_vl_interface': 7.506237863819593e-06, 'learning_rate/action_model': 7.506237863819593e-05, 'epoch': 1.41}) +[2026-06-15 23:39:11,951][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.06781261414289474, 'train/grad_norm_pre_clip': 6.406508922576904, 'timing/data': 0.020588099025189877, 'timing/model': 1.4511465402320027, 'learning_rate/qwen_vl_interface': 7.503557468459345e-06, 'learning_rate/action_model': 7.503557468459344e-05, 'epoch': 1.41}) +[2026-06-15 23:39:17,866][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.10442493110895157, 'train/grad_norm_pre_clip': 9.763187408447266, 'timing/data': 0.020989609882235527, 'timing/model': 1.4024214819073677, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 1.41}) +[2026-06-15 23:39:23,706][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.08596417307853699, 'train/grad_norm_pre_clip': 4.827389240264893, 'timing/data': 0.020855634473264217, 'timing/model': 1.694574048742652, 'learning_rate/qwen_vl_interface': 7.498193899783983e-06, 'learning_rate/action_model': 7.498193899783983e-05, 'epoch': 1.41}) +[2026-06-15 23:39:29,355][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.06924120336771011, 'train/grad_norm_pre_clip': 5.062826156616211, 'timing/data': 0.020560194738209248, 'timing/model': 1.4458876149728894, 'learning_rate/qwen_vl_interface': 7.495510728673631e-06, 'learning_rate/action_model': 7.495510728673632e-05, 'epoch': 1.41}) +[2026-06-15 23:39:35,158][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.17799173295497894, 'train/grad_norm_pre_clip': 20.496252059936523, 'timing/data': 0.021365277469158173, 'timing/model': 1.4479160262271762, 'learning_rate/qwen_vl_interface': 7.492826634518642e-06, 'learning_rate/action_model': 7.492826634518641e-05, 'epoch': 1.41}) +[2026-06-15 23:39:40,959][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.11390137672424316, 'train/grad_norm_pre_clip': 6.955760478973389, 'timing/data': 0.022106271237134933, 'timing/model': 1.446032365784049, 'learning_rate/qwen_vl_interface': 7.4901416184223404e-06, 'learning_rate/action_model': 7.49014161842234e-05, 'epoch': 1.41}) +[2026-06-15 23:39:46,768][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.003314698813483119, 'train/grad_norm_pre_clip': 0.4668522775173187, 'timing/data': 0.021521189250051975, 'timing/model': 1.40942651219666, 'learning_rate/qwen_vl_interface': 7.487455681488437e-06, 'learning_rate/action_model': 7.487455681488436e-05, 'epoch': 1.41}) +[2026-06-15 23:39:52,559][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.021510189399123192, 'train/grad_norm_pre_clip': 6.98507022857666, 'timing/data': 0.0220706919208169, 'timing/model': 1.4110918622463942, 'learning_rate/qwen_vl_interface': 7.484768824821018e-06, 'learning_rate/action_model': 7.484768824821018e-05, 'epoch': 1.41}) +[2026-06-15 23:39:58,418][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.06426222622394562, 'train/grad_norm_pre_clip': 5.917308807373047, 'timing/data': 0.021026109345257282, 'timing/model': 1.7093395618721843, 'learning_rate/qwen_vl_interface': 7.482081049524551e-06, 'learning_rate/action_model': 7.482081049524551e-05, 'epoch': 1.41}) +[2026-06-15 23:40:04,060][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.018631521612405777, 'train/grad_norm_pre_clip': 2.4612538814544678, 'timing/data': 0.02305813692510128, 'timing/model': 1.4410969745367765, 'learning_rate/qwen_vl_interface': 7.479392356703878e-06, 'learning_rate/action_model': 7.479392356703878e-05, 'epoch': 1.41}) +[2026-06-15 23:40:09,858][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.06084980443120003, 'train/grad_norm_pre_clip': 10.152823448181152, 'timing/data': 0.021534270606935024, 'timing/model': 1.4366396041586995, 'learning_rate/qwen_vl_interface': 7.476702747464219e-06, 'learning_rate/action_model': 7.476702747464219e-05, 'epoch': 1.41}) +[2026-06-15 23:40:15,658][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.062161851674318314, 'train/grad_norm_pre_clip': 7.3824005126953125, 'timing/data': 0.02049287874251604, 'timing/model': 1.4515745164826512, 'learning_rate/qwen_vl_interface': 7.474012222911169e-06, 'learning_rate/action_model': 7.474012222911169e-05, 'epoch': 1.41}) +[2026-06-15 23:40:21,462][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.06284863501787186, 'train/grad_norm_pre_clip': 8.494423866271973, 'timing/data': 0.020669329911470413, 'timing/model': 1.4462845027446747, 'learning_rate/qwen_vl_interface': 7.471320784150703e-06, 'learning_rate/action_model': 7.471320784150703e-05, 'epoch': 1.41}) +[2026-06-15 23:40:27,287][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.2586689591407776, 'train/grad_norm_pre_clip': 19.902183532714844, 'timing/data': 0.021969043649733067, 'timing/model': 1.4105611834675074, 'learning_rate/qwen_vl_interface': 7.468628432289169e-06, 'learning_rate/action_model': 7.468628432289168e-05, 'epoch': 1.41}) +[2026-06-15 23:40:33,201][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.21499089896678925, 'train/grad_norm_pre_clip': 9.551681518554688, 'timing/data': 0.021582184359431267, 'timing/model': 1.6854450721293688, 'learning_rate/qwen_vl_interface': 7.465935168433292e-06, 'learning_rate/action_model': 7.465935168433292e-05, 'epoch': 1.41}) +[2026-06-15 23:40:38,848][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.2999028265476227, 'train/grad_norm_pre_clip': 23.357553482055664, 'timing/data': 0.019863507710397243, 'timing/model': 1.453573222272098, 'learning_rate/qwen_vl_interface': 7.463240993690168e-06, 'learning_rate/action_model': 7.463240993690168e-05, 'epoch': 1.41}) +[2026-06-15 23:40:44,664][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.08987335115671158, 'train/grad_norm_pre_clip': 13.527945518493652, 'timing/data': 0.02099333330988884, 'timing/model': 1.471901866607368, 'learning_rate/qwen_vl_interface': 7.460545909167276e-06, 'learning_rate/action_model': 7.460545909167275e-05, 'epoch': 1.42}) +[2026-06-15 23:40:50,457][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.09439137578010559, 'train/grad_norm_pre_clip': 10.594818115234375, 'timing/data': 0.0210343636572361, 'timing/model': 1.463717999868095, 'learning_rate/qwen_vl_interface': 7.457849915972459e-06, 'learning_rate/action_model': 7.457849915972458e-05, 'epoch': 1.42}) +[2026-06-15 23:40:56,258][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.013854531571269035, 'train/grad_norm_pre_clip': 5.2044453620910645, 'timing/data': 0.020805783569812775, 'timing/model': 1.4614030625671148, 'learning_rate/qwen_vl_interface': 7.455153015213941e-06, 'learning_rate/action_model': 7.45515301521394e-05, 'epoch': 1.42}) +[2026-06-15 23:41:02,044][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.5155717134475708, 'train/grad_norm_pre_clip': 20.395715713500977, 'timing/data': 0.020726962946355343, 'timing/model': 1.4035832388326526, 'learning_rate/qwen_vl_interface': 7.452455208000311e-06, 'learning_rate/action_model': 7.452455208000311e-05, 'epoch': 1.42}) +[2026-06-15 23:41:08,480][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.4771485924720764, 'train/grad_norm_pre_clip': 17.412757873535156, 'timing/data': 0.0533975288271904, 'timing/model': 1.78952063806355, 'learning_rate/qwen_vl_interface': 7.449756495440541e-06, 'learning_rate/action_model': 7.44975649544054e-05, 'epoch': 1.42}) +[2026-06-15 23:51:50,181][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.033113885670900345, 'train/grad_norm_pre_clip': 9.057936668395996, 'eval/loss': 0.09595779329538345, 'eval/flap_precision': 0.7517241379310344, 'eval/flap_recall': 0.7927272727272727, 'eval/flap_f1': 0.7716814159292035, 'eval/flap_rate_model': 0.08055555555555556, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7716814159292035, 'timing/data': 0.021446125581860542, 'timing/model': 1.4532890152186155, 'learning_rate/qwen_vl_interface': 7.447056878643969e-06, 'learning_rate/action_model': 7.447056878643968e-05, 'epoch': 1.42}) +[2026-06-15 23:51:50,183][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3100_state +[2026-06-15 23:51:50,189][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-15 23:51:57,990][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3100_state/model.safetensors +[2026-06-15 23:52:09,113][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3100_state/optimizer.bin +[2026-06-15 23:52:09,114][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3100_state/sampler.bin +[2026-06-15 23:52:09,114][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3100_state/sampler_1.bin +[2026-06-15 23:52:09,116][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3100_state/random_states_0.pkl +[2026-06-15 23:52:09,271][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3100 +[2026-06-15 23:52:09,271][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-15 23:52:09,287][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-15 23:52:14,996][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.13603338599205017, 'train/grad_norm_pre_clip': 11.20156478881836, 'timing/data': 0.02298504114151001, 'timing/model': 1.6375806713476777, 'learning_rate/qwen_vl_interface': 7.444356358720303e-06, 'learning_rate/action_model': 7.444356358720302e-05, 'epoch': 1.42}) +[2026-06-15 23:52:20,418][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.10755922645330429, 'train/grad_norm_pre_clip': 10.657795906066895, 'timing/data': 0.024053976871073246, 'timing/model': 1.4455531472340226, 'learning_rate/qwen_vl_interface': 7.4416549367796255e-06, 'learning_rate/action_model': 7.441654936779625e-05, 'epoch': 1.42}) +[2026-06-15 23:52:26,116][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.033578094094991684, 'train/grad_norm_pre_clip': 3.2814853191375732, 'timing/data': 0.02202277723699808, 'timing/model': 1.4313925551250577, 'learning_rate/qwen_vl_interface': 7.4389526139323895e-06, 'learning_rate/action_model': 7.438952613932389e-05, 'epoch': 1.42}) +[2026-06-15 23:52:31,737][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.10467439144849777, 'train/grad_norm_pre_clip': 15.994694709777832, 'timing/data': 0.020146308466792107, 'timing/model': 1.3657072121277452, 'learning_rate/qwen_vl_interface': 7.436249391289419e-06, 'learning_rate/action_model': 7.436249391289418e-05, 'epoch': 1.42}) +[2026-06-15 23:52:37,321][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.11290502548217773, 'train/grad_norm_pre_clip': 9.672801971435547, 'timing/data': 0.028091627173125744, 'timing/model': 1.372816334478557, 'learning_rate/qwen_vl_interface': 7.433545269961906e-06, 'learning_rate/action_model': 7.433545269961906e-05, 'epoch': 1.42}) +[2026-06-15 23:52:42,933][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.15218816697597504, 'train/grad_norm_pre_clip': 20.39217185974121, 'timing/data': 0.02080341149121523, 'timing/model': 1.4461119696497917, 'learning_rate/qwen_vl_interface': 7.430840251061411e-06, 'learning_rate/action_model': 7.430840251061411e-05, 'epoch': 1.42}) +[2026-06-15 23:52:48,701][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.20127280056476593, 'train/grad_norm_pre_clip': 14.200448036193848, 'timing/data': 0.021701696328818798, 'timing/model': 1.7098152143880725, 'learning_rate/qwen_vl_interface': 7.428134335699866e-06, 'learning_rate/action_model': 7.428134335699865e-05, 'epoch': 1.42}) +[2026-06-15 23:52:54,251][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.21127882599830627, 'train/grad_norm_pre_clip': 10.80737018585205, 'timing/data': 0.025946796871721745, 'timing/model': 1.4495616015046835, 'learning_rate/qwen_vl_interface': 7.425427524989572e-06, 'learning_rate/action_model': 7.425427524989572e-05, 'epoch': 1.42}) +[2026-06-15 23:52:59,940][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.17509794235229492, 'train/grad_norm_pre_clip': 10.97225570678711, 'timing/data': 0.021321534179151058, 'timing/model': 1.426657716743648, 'learning_rate/qwen_vl_interface': 7.422719820043194e-06, 'learning_rate/action_model': 7.422719820043193e-05, 'epoch': 1.42}) +[2026-06-15 23:53:05,660][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.3656564950942993, 'train/grad_norm_pre_clip': 5.595646858215332, 'timing/data': 0.021984667517244816, 'timing/model': 1.3794112810865045, 'learning_rate/qwen_vl_interface': 7.420011221973771e-06, 'learning_rate/action_model': 7.420011221973771e-05, 'epoch': 1.42}) +[2026-06-15 23:53:11,438][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.37255457043647766, 'train/grad_norm_pre_clip': 15.697500228881836, 'timing/data': 0.02120802365243435, 'timing/model': 1.4036542372778058, 'learning_rate/qwen_vl_interface': 7.417301731894702e-06, 'learning_rate/action_model': 7.417301731894701e-05, 'epoch': 1.42}) +[2026-06-15 23:53:17,193][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.04184006154537201, 'train/grad_norm_pre_clip': 4.713955879211426, 'timing/data': 0.021525315940380096, 'timing/model': 1.6812205165624619, 'learning_rate/qwen_vl_interface': 7.414591350919756e-06, 'learning_rate/action_model': 7.414591350919756e-05, 'epoch': 1.42}) +[2026-06-15 23:53:22,943][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.11589320749044418, 'train/grad_norm_pre_clip': 12.055977821350098, 'timing/data': 0.02171301655471325, 'timing/model': 1.6235789405182004, 'learning_rate/qwen_vl_interface': 7.4118800801630715e-06, 'learning_rate/action_model': 7.411880080163072e-05, 'epoch': 1.42}) +[2026-06-15 23:53:28,443][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.027056150138378143, 'train/grad_norm_pre_clip': 2.849595785140991, 'timing/data': 0.0210305443033576, 'timing/model': 1.4366796491667628, 'learning_rate/qwen_vl_interface': 7.409167920739146e-06, 'learning_rate/action_model': 7.409167920739147e-05, 'epoch': 1.42}) +[2026-06-15 23:53:34,147][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.20007099211215973, 'train/grad_norm_pre_clip': 9.403182983398438, 'timing/data': 0.02298162318766117, 'timing/model': 1.4029106609523296, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 1.42}) +[2026-06-15 23:53:39,941][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.02544359304010868, 'train/grad_norm_pre_clip': 3.566798686981201, 'timing/data': 0.021322205662727356, 'timing/model': 1.4425947535783052, 'learning_rate/qwen_vl_interface': 7.403740940349408e-06, 'learning_rate/action_model': 7.403740940349408e-05, 'epoch': 1.43}) +[2026-06-15 23:53:45,648][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.37058138847351074, 'train/grad_norm_pre_clip': 11.873625755310059, 'timing/data': 0.0215492257848382, 'timing/model': 1.406079139560461, 'learning_rate/qwen_vl_interface': 7.401026121614421e-06, 'learning_rate/action_model': 7.40102612161442e-05, 'epoch': 1.43}) +[2026-06-15 23:53:51,458][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.06161313131451607, 'train/grad_norm_pre_clip': 7.583205223083496, 'timing/data': 0.0226317523047328, 'timing/model': 1.3980496460571885, 'learning_rate/qwen_vl_interface': 7.398310418673846e-06, 'learning_rate/action_model': 7.398310418673846e-05, 'epoch': 1.43}) +[2026-06-15 23:53:57,278][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.08992661535739899, 'train/grad_norm_pre_clip': 4.036819934844971, 'timing/data': 0.021578245796263218, 'timing/model': 1.664065827615559, 'learning_rate/qwen_vl_interface': 7.395593832644007e-06, 'learning_rate/action_model': 7.395593832644007e-05, 'epoch': 1.43}) +[2026-06-15 23:54:02,860][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.15755300223827362, 'train/grad_norm_pre_clip': 10.994803428649902, 'timing/data': 0.021300475113093853, 'timing/model': 1.4660055544227362, 'learning_rate/qwen_vl_interface': 7.39287636464159e-06, 'learning_rate/action_model': 7.39287636464159e-05, 'epoch': 1.43}) +[2026-06-15 23:54:08,663][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.08964326232671738, 'train/grad_norm_pre_clip': 4.017940521240234, 'timing/data': 0.021754948422312737, 'timing/model': 1.4484574347734451, 'learning_rate/qwen_vl_interface': 7.3901580157836425e-06, 'learning_rate/action_model': 7.390158015783643e-05, 'epoch': 1.43}) +[2026-06-15 23:54:14,450][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.1621827781200409, 'train/grad_norm_pre_clip': 4.176206588745117, 'timing/data': 0.020957523956894875, 'timing/model': 1.45658447034657, 'learning_rate/qwen_vl_interface': 7.387438787187576e-06, 'learning_rate/action_model': 7.387438787187576e-05, 'epoch': 1.43}) +[2026-06-15 23:54:20,247][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.31150972843170166, 'train/grad_norm_pre_clip': 7.748518466949463, 'timing/data': 0.020314211025834084, 'timing/model': 1.3954684240743518, 'learning_rate/qwen_vl_interface': 7.3847186799711625e-06, 'learning_rate/action_model': 7.384718679971162e-05, 'epoch': 1.43}) +[2026-06-15 23:54:25,992][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.021786773577332497, 'train/grad_norm_pre_clip': 2.1380128860473633, 'timing/data': 0.022185000590980053, 'timing/model': 1.6909291669726372, 'learning_rate/qwen_vl_interface': 7.381997695252535e-06, 'learning_rate/action_model': 7.381997695252535e-05, 'epoch': 1.43}) +[2026-06-15 23:54:31,753][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.2368975132703781, 'train/grad_norm_pre_clip': 9.610355377197266, 'timing/data': 0.02083503268659115, 'timing/model': 1.458775620907545, 'learning_rate/qwen_vl_interface': 7.379275834150186e-06, 'learning_rate/action_model': 7.379275834150186e-05, 'epoch': 1.43}) +[2026-06-15 23:54:37,483][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.13792715966701508, 'train/grad_norm_pre_clip': 5.576276779174805, 'timing/data': 0.020483002066612244, 'timing/model': 1.4072394901886582, 'learning_rate/qwen_vl_interface': 7.376553097782972e-06, 'learning_rate/action_model': 7.376553097782972e-05, 'epoch': 1.43}) +[2026-06-15 23:54:43,150][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.11343645304441452, 'train/grad_norm_pre_clip': 5.835094928741455, 'timing/data': 0.020414832048118114, 'timing/model': 1.436208919622004, 'learning_rate/qwen_vl_interface': 7.373829487270109e-06, 'learning_rate/action_model': 7.373829487270109e-05, 'epoch': 1.43}) +[2026-06-15 23:54:48,865][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.1965961456298828, 'train/grad_norm_pre_clip': 11.158673286437988, 'timing/data': 0.020023918710649014, 'timing/model': 1.4702880959957838, 'learning_rate/qwen_vl_interface': 7.371105003731165e-06, 'learning_rate/action_model': 7.371105003731164e-05, 'epoch': 1.43}) +[2026-06-15 23:54:54,684][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.14601315557956696, 'train/grad_norm_pre_clip': 5.697141647338867, 'timing/data': 0.01967207808047533, 'timing/model': 1.4239149205386639, 'learning_rate/qwen_vl_interface': 7.3683796482860755e-06, 'learning_rate/action_model': 7.368379648286075e-05, 'epoch': 1.43}) +[2026-06-15 23:55:00,607][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.17888520658016205, 'train/grad_norm_pre_clip': 8.358827590942383, 'timing/data': 0.021015014499425888, 'timing/model': 1.699720362201333, 'learning_rate/qwen_vl_interface': 7.365653422055131e-06, 'learning_rate/action_model': 7.365653422055131e-05, 'epoch': 1.43}) +[2026-06-15 23:55:06,251][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.23990903794765472, 'train/grad_norm_pre_clip': 7.649845600128174, 'timing/data': 0.021437477320432663, 'timing/model': 1.438623245805502, 'learning_rate/qwen_vl_interface': 7.3629263261589816e-06, 'learning_rate/action_model': 7.362926326158981e-05, 'epoch': 1.43}) +[2026-06-15 23:55:12,054][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.06174463778734207, 'train/grad_norm_pre_clip': 3.273956298828125, 'timing/data': 0.021740049123764038, 'timing/model': 1.4393588844686747, 'learning_rate/qwen_vl_interface': 7.36019836171863e-06, 'learning_rate/action_model': 7.360198361718629e-05, 'epoch': 1.43}) +[2026-06-15 23:55:17,852][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.17200787365436554, 'train/grad_norm_pre_clip': 7.374792098999023, 'timing/data': 0.0207615839317441, 'timing/model': 1.437980592250824, 'learning_rate/qwen_vl_interface': 7.357469529855441e-06, 'learning_rate/action_model': 7.35746952985544e-05, 'epoch': 1.43}) +[2026-06-15 23:55:23,653][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.31142306327819824, 'train/grad_norm_pre_clip': 10.282132148742676, 'timing/data': 0.02122694719582796, 'timing/model': 1.4446908123791218, 'learning_rate/qwen_vl_interface': 7.354739831691134e-06, 'learning_rate/action_model': 7.354739831691135e-05, 'epoch': 1.43}) +[2026-06-15 23:55:29,450][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.18228431046009064, 'train/grad_norm_pre_clip': 8.567227363586426, 'timing/data': 0.021304607391357422, 'timing/model': 1.433484791778028, 'learning_rate/qwen_vl_interface': 7.352009268347786e-06, 'learning_rate/action_model': 7.352009268347786e-05, 'epoch': 1.43}) +[2026-06-15 23:55:35,308][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.13466767966747284, 'train/grad_norm_pre_clip': 5.465122222900391, 'timing/data': 0.02165735885500908, 'timing/model': 1.6957820914685726, 'learning_rate/qwen_vl_interface': 7.349277840947828e-06, 'learning_rate/action_model': 7.349277840947828e-05, 'epoch': 1.43}) +[2026-06-15 23:55:40,947][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.11199957877397537, 'train/grad_norm_pre_clip': 3.94083571434021, 'timing/data': 0.02081789541989565, 'timing/model': 1.4384570783004165, 'learning_rate/qwen_vl_interface': 7.346545550614046e-06, 'learning_rate/action_model': 7.346545550614046e-05, 'epoch': 1.43}) +[2026-06-15 23:55:46,747][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.15842245519161224, 'train/grad_norm_pre_clip': 6.112495422363281, 'timing/data': 0.01981859840452671, 'timing/model': 1.4421091107651591, 'learning_rate/qwen_vl_interface': 7.343812398469582e-06, 'learning_rate/action_model': 7.343812398469581e-05, 'epoch': 1.44}) +[2026-06-15 23:55:52,480][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.13494530320167542, 'train/grad_norm_pre_clip': 5.5162787437438965, 'timing/data': 0.02485651522874832, 'timing/model': 1.482783361338079, 'learning_rate/qwen_vl_interface': 7.34107838563793e-06, 'learning_rate/action_model': 7.341078385637929e-05, 'epoch': 1.44}) +[2026-06-15 23:55:58,255][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.13084107637405396, 'train/grad_norm_pre_clip': 4.226078987121582, 'timing/data': 0.025012366473674774, 'timing/model': 1.3965441435575485, 'learning_rate/qwen_vl_interface': 7.338343513242941e-06, 'learning_rate/action_model': 7.338343513242941e-05, 'epoch': 1.44}) +[2026-06-15 23:56:04,055][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.11334507167339325, 'train/grad_norm_pre_clip': 5.3128275871276855, 'timing/data': 0.021071716211736202, 'timing/model': 1.415805703960359, 'learning_rate/qwen_vl_interface': 7.335607782408818e-06, 'learning_rate/action_model': 7.335607782408818e-05, 'epoch': 1.44}) +[2026-06-15 23:56:09,908][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.15331779420375824, 'train/grad_norm_pre_clip': 5.9585041999816895, 'timing/data': 0.021701158955693245, 'timing/model': 1.6963168922811747, 'learning_rate/qwen_vl_interface': 7.3328711942601145e-06, 'learning_rate/action_model': 7.332871194260115e-05, 'epoch': 1.44}) +[2026-06-15 23:56:15,554][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.12398311495780945, 'train/grad_norm_pre_clip': 3.997687339782715, 'timing/data': 0.02113528549671173, 'timing/model': 1.4267097981646657, 'learning_rate/qwen_vl_interface': 7.33013374992174e-06, 'learning_rate/action_model': 7.33013374992174e-05, 'epoch': 1.44}) +[2026-06-15 23:56:21,352][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.10874958336353302, 'train/grad_norm_pre_clip': 3.301492214202881, 'timing/data': 0.022278161719441414, 'timing/model': 1.4378622807562351, 'learning_rate/qwen_vl_interface': 7.327395450518956e-06, 'learning_rate/action_model': 7.327395450518956e-05, 'epoch': 1.44}) +[2026-06-15 23:56:27,154][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.17413651943206787, 'train/grad_norm_pre_clip': 7.804641246795654, 'timing/data': 0.020803414285182953, 'timing/model': 1.44735009316355, 'learning_rate/qwen_vl_interface': 7.324656297177372e-06, 'learning_rate/action_model': 7.324656297177371e-05, 'epoch': 1.44}) +[2026-06-15 23:56:32,966][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.12210452556610107, 'train/grad_norm_pre_clip': 5.364872455596924, 'timing/data': 0.02074323408305645, 'timing/model': 1.4570997981354594, 'learning_rate/qwen_vl_interface': 7.32191629102295e-06, 'learning_rate/action_model': 7.32191629102295e-05, 'epoch': 1.44}) +[2026-06-15 23:56:38,760][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.1677095592021942, 'train/grad_norm_pre_clip': 5.865384101867676, 'timing/data': 0.02086547575891018, 'timing/model': 1.4174612117931247, 'learning_rate/qwen_vl_interface': 7.319175433182004e-06, 'learning_rate/action_model': 7.319175433182004e-05, 'epoch': 1.44}) +[2026-06-15 23:56:44,609][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.308826208114624, 'train/grad_norm_pre_clip': 13.461653709411621, 'timing/data': 0.021994360722601414, 'timing/model': 1.6896722335368395, 'learning_rate/qwen_vl_interface': 7.316433724781198e-06, 'learning_rate/action_model': 7.316433724781197e-05, 'epoch': 1.44}) +[2026-06-15 23:56:50,258][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.11170876026153564, 'train/grad_norm_pre_clip': 7.112613201141357, 'timing/data': 0.020963615737855434, 'timing/model': 1.4417904736474156, 'learning_rate/qwen_vl_interface': 7.313691166947543e-06, 'learning_rate/action_model': 7.313691166947543e-05, 'epoch': 1.44}) +[2026-06-15 23:56:56,077][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.1538531482219696, 'train/grad_norm_pre_clip': 7.0107035636901855, 'timing/data': 0.02122923731803894, 'timing/model': 1.4497550586238503, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 1.44}) +[2026-06-15 23:57:01,867][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.11518403887748718, 'train/grad_norm_pre_clip': 4.996350288391113, 'timing/data': 0.021433379501104355, 'timing/model': 1.4559244150295854, 'learning_rate/qwen_vl_interface': 7.308203507491485e-06, 'learning_rate/action_model': 7.308203507491484e-05, 'epoch': 1.44}) +[2026-06-15 23:57:07,652][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.19426539540290833, 'train/grad_norm_pre_clip': 7.635481357574463, 'timing/data': 0.020974526181817055, 'timing/model': 1.4379430888220668, 'learning_rate/qwen_vl_interface': 7.3054584081248515e-06, 'learning_rate/action_model': 7.305458408124852e-05, 'epoch': 1.44}) +[2026-06-15 23:57:13,460][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.11978406459093094, 'train/grad_norm_pre_clip': 4.89816951751709, 'timing/data': 0.021387718617916107, 'timing/model': 1.3998410096392035, 'learning_rate/qwen_vl_interface': 7.302712463836907e-06, 'learning_rate/action_model': 7.302712463836907e-05, 'epoch': 1.44}) +[2026-06-15 23:57:19,310][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.18135575950145721, 'train/grad_norm_pre_clip': 6.393453598022461, 'timing/data': 0.02156731951981783, 'timing/model': 1.6925931703299284, 'learning_rate/qwen_vl_interface': 7.2999656757564086e-06, 'learning_rate/action_model': 7.299965675756409e-05, 'epoch': 1.44}) +[2026-06-15 23:57:24,950][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.17590469121932983, 'train/grad_norm_pre_clip': 8.019899368286133, 'timing/data': 0.019837230443954468, 'timing/model': 1.4408843871206045, 'learning_rate/qwen_vl_interface': 7.297218045012453e-06, 'learning_rate/action_model': 7.297218045012453e-05, 'epoch': 1.44}) +[2026-06-15 23:57:30,754][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.11184472590684891, 'train/grad_norm_pre_clip': 5.744307994842529, 'timing/data': 0.021161028183996677, 'timing/model': 1.454339949414134, 'learning_rate/qwen_vl_interface': 7.29446957273449e-06, 'learning_rate/action_model': 7.294469572734489e-05, 'epoch': 1.44}) +[2026-06-15 23:57:36,556][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.2932840585708618, 'train/grad_norm_pre_clip': 10.894869804382324, 'timing/data': 0.020792205817997456, 'timing/model': 1.4414024585857987, 'learning_rate/qwen_vl_interface': 7.2917202600523104e-06, 'learning_rate/action_model': 7.29172026005231e-05, 'epoch': 1.44}) +[2026-06-15 23:57:42,351][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.04991825670003891, 'train/grad_norm_pre_clip': 2.7064759731292725, 'timing/data': 0.021576320752501488, 'timing/model': 1.4451467813923955, 'learning_rate/qwen_vl_interface': 7.2889701080960565e-06, 'learning_rate/action_model': 7.288970108096056e-05, 'epoch': 1.44}) +[2026-06-15 23:57:48,162][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.03914831206202507, 'train/grad_norm_pre_clip': 3.6387436389923096, 'timing/data': 0.02166643086820841, 'timing/model': 1.4024234972894192, 'learning_rate/qwen_vl_interface': 7.286219117996209e-06, 'learning_rate/action_model': 7.286219117996207e-05, 'epoch': 1.44}) +[2026-06-15 23:57:53,995][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.08450531959533691, 'train/grad_norm_pre_clip': 6.718137264251709, 'timing/data': 0.021493809297680855, 'timing/model': 1.6758358534425497, 'learning_rate/qwen_vl_interface': 7.283467290883598e-06, 'learning_rate/action_model': 7.283467290883598e-05, 'epoch': 1.45}) +[2026-06-15 23:57:59,659][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.08722493797540665, 'train/grad_norm_pre_clip': 3.1708357334136963, 'timing/data': 0.021921331994235516, 'timing/model': 1.440600880421698, 'learning_rate/qwen_vl_interface': 7.2807146278893925e-06, 'learning_rate/action_model': 7.280714627889393e-05, 'epoch': 1.45}) +[2026-06-15 23:58:05,465][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.18572933971881866, 'train/grad_norm_pre_clip': 15.753219604492188, 'timing/data': 0.025632481090724468, 'timing/model': 1.4642496854066849, 'learning_rate/qwen_vl_interface': 7.277961130145111e-06, 'learning_rate/action_model': 7.277961130145111e-05, 'epoch': 1.45}) +[2026-06-15 23:58:11,250][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.078144870698452, 'train/grad_norm_pre_clip': 4.46000337600708, 'timing/data': 0.02110573649406433, 'timing/model': 1.4322741953656077, 'learning_rate/qwen_vl_interface': 7.2752067987826116e-06, 'learning_rate/action_model': 7.275206798782611e-05, 'epoch': 1.45}) +[2026-06-15 23:58:17,058][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.08980491012334824, 'train/grad_norm_pre_clip': 12.255072593688965, 'timing/data': 0.021045447327196598, 'timing/model': 1.4580705920234323, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 1.45}) +[2026-06-15 23:58:22,954][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.08368939161300659, 'train/grad_norm_pre_clip': 6.602508068084717, 'timing/data': 0.022861826233565807, 'timing/model': 1.4536899328231812, 'learning_rate/qwen_vl_interface': 7.269695639732108e-06, 'learning_rate/action_model': 7.269695639732108e-05, 'epoch': 1.45}) +[2026-06-15 23:58:28,792][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.10775576531887054, 'train/grad_norm_pre_clip': 6.094807147979736, 'timing/data': 0.022009512409567833, 'timing/model': 1.6539820106700063, 'learning_rate/qwen_vl_interface': 7.266938814309533e-06, 'learning_rate/action_model': 7.266938814309533e-05, 'epoch': 1.45}) +[2026-06-15 23:58:34,631][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.05326607823371887, 'train/grad_norm_pre_clip': 5.59246826171875, 'timing/data': 0.02113758772611618, 'timing/model': 1.6235955897718668, 'learning_rate/qwen_vl_interface': 7.264181159799598e-06, 'learning_rate/action_model': 7.264181159799597e-05, 'epoch': 1.45}) +[2026-06-15 23:58:40,152][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.14886334538459778, 'train/grad_norm_pre_clip': 7.635390281677246, 'timing/data': 0.022371094673871994, 'timing/model': 1.45393967628479, 'learning_rate/qwen_vl_interface': 7.261422677335871e-06, 'learning_rate/action_model': 7.261422677335871e-05, 'epoch': 1.45}) +[2026-06-15 23:58:45,956][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.05384167283773422, 'train/grad_norm_pre_clip': 11.388204574584961, 'timing/data': 0.020827447064220905, 'timing/model': 1.4631279669702053, 'learning_rate/qwen_vl_interface': 7.258663368052259e-06, 'learning_rate/action_model': 7.258663368052259e-05, 'epoch': 1.45}) +[2026-06-15 23:58:51,753][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.07089721411466599, 'train/grad_norm_pre_clip': 6.264288902282715, 'timing/data': 0.02129919920116663, 'timing/model': 1.4585580974817276, 'learning_rate/qwen_vl_interface': 7.2559032330830105e-06, 'learning_rate/action_model': 7.25590323308301e-05, 'epoch': 1.45}) +[2026-06-15 23:58:57,557][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.1373964101076126, 'train/grad_norm_pre_clip': 6.943411350250244, 'timing/data': 0.02146905940026045, 'timing/model': 1.4031295366585255, 'learning_rate/qwen_vl_interface': 7.253142273562712e-06, 'learning_rate/action_model': 7.253142273562712e-05, 'epoch': 1.45}) +[2026-06-15 23:59:03,396][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.0068783569149672985, 'train/grad_norm_pre_clip': 0.9461224675178528, 'timing/data': 0.021769472397863865, 'timing/model': 1.6996439201757312, 'learning_rate/qwen_vl_interface': 7.25038049062629e-06, 'learning_rate/action_model': 7.250380490626289e-05, 'epoch': 1.45}) +[2026-06-15 23:59:09,201][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.11827944219112396, 'train/grad_norm_pre_clip': 10.070085525512695, 'timing/data': 0.020569685846567154, 'timing/model': 1.7058529360219836, 'learning_rate/qwen_vl_interface': 7.24761788540901e-06, 'learning_rate/action_model': 7.24761788540901e-05, 'epoch': 1.45}) +[2026-06-15 23:59:14,755][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.4527730345726013, 'train/grad_norm_pre_clip': 12.596785545349121, 'timing/data': 0.020090483129024506, 'timing/model': 1.4505847012624145, 'learning_rate/qwen_vl_interface': 7.244854459046473e-06, 'learning_rate/action_model': 7.244854459046473e-05, 'epoch': 1.45}) +[2026-06-15 23:59:20,560][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.2113065868616104, 'train/grad_norm_pre_clip': 16.5032901763916, 'timing/data': 0.020566146820783615, 'timing/model': 1.45126370806247, 'learning_rate/qwen_vl_interface': 7.24209021267462e-06, 'learning_rate/action_model': 7.24209021267462e-05, 'epoch': 1.45}) +[2026-06-15 23:59:26,364][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.15396316349506378, 'train/grad_norm_pre_clip': 5.941958904266357, 'timing/data': 0.020422655157744884, 'timing/model': 1.4702776493504643, 'learning_rate/qwen_vl_interface': 7.23932514742973e-06, 'learning_rate/action_model': 7.23932514742973e-05, 'epoch': 1.45}) +[2026-06-15 23:59:32,152][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.02117614634335041, 'train/grad_norm_pre_clip': 3.6596879959106445, 'timing/data': 0.02133949939161539, 'timing/model': 1.4101063525304198, 'learning_rate/qwen_vl_interface': 7.236559264448413e-06, 'learning_rate/action_model': 7.236559264448412e-05, 'epoch': 1.45}) +[2026-06-15 23:59:37,889][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.1349303275346756, 'train/grad_norm_pre_clip': 13.894227981567383, 'timing/data': 0.020717977546155453, 'timing/model': 1.6883386364206672, 'learning_rate/qwen_vl_interface': 7.233792564867623e-06, 'learning_rate/action_model': 7.233792564867623e-05, 'epoch': 1.45}) +[2026-06-15 23:59:43,472][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.1641111820936203, 'train/grad_norm_pre_clip': 9.750266075134277, 'timing/data': 0.02069475594907999, 'timing/model': 1.3918271157890558, 'learning_rate/qwen_vl_interface': 7.231025049824644e-06, 'learning_rate/action_model': 7.231025049824643e-05, 'epoch': 1.45}) +[2026-06-15 23:59:49,243][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.04310748353600502, 'train/grad_norm_pre_clip': 5.016578197479248, 'timing/data': 0.021066869609057903, 'timing/model': 1.4331029215827584, 'learning_rate/qwen_vl_interface': 7.228256720457097e-06, 'learning_rate/action_model': 7.228256720457096e-05, 'epoch': 1.45}) +[2026-06-15 23:59:55,060][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.3639776110649109, 'train/grad_norm_pre_clip': 10.292418479919434, 'timing/data': 0.02492889855057001, 'timing/model': 1.4582122741267085, 'learning_rate/qwen_vl_interface': 7.225487577902937e-06, 'learning_rate/action_model': 7.225487577902936e-05, 'epoch': 1.45}) +[2026-06-16 00:00:00,853][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.09439266473054886, 'train/grad_norm_pre_clip': 4.8692731857299805, 'timing/data': 0.02126211952418089, 'timing/model': 1.4030534932389855, 'learning_rate/qwen_vl_interface': 7.222717623300455e-06, 'learning_rate/action_model': 7.222717623300455e-05, 'epoch': 1.46}) +[2026-06-16 00:00:06,636][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.27654242515563965, 'train/grad_norm_pre_clip': 12.685574531555176, 'timing/data': 0.022769867442548275, 'timing/model': 1.4534098654985428, 'learning_rate/qwen_vl_interface': 7.219946857788273e-06, 'learning_rate/action_model': 7.219946857788272e-05, 'epoch': 1.46}) +[2026-06-16 00:00:12,462][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.03007705882191658, 'train/grad_norm_pre_clip': 1.8193966150283813, 'timing/data': 0.021264899522066116, 'timing/model': 1.658841447904706, 'learning_rate/qwen_vl_interface': 7.21717528250535e-06, 'learning_rate/action_model': 7.21717528250535e-05, 'epoch': 1.46}) +[2026-06-16 00:00:18,049][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.0365401916205883, 'train/grad_norm_pre_clip': 2.3216781616210938, 'timing/data': 0.021811592392623425, 'timing/model': 1.448184215463698, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 1.46}) +[2026-06-16 00:00:23,853][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.21278247237205505, 'train/grad_norm_pre_clip': 10.028964042663574, 'timing/data': 0.0226807976141572, 'timing/model': 1.4352379105985165, 'learning_rate/qwen_vl_interface': 7.211629707184768e-06, 'learning_rate/action_model': 7.211629707184768e-05, 'epoch': 1.46}) +[2026-06-16 00:00:29,653][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.166570782661438, 'train/grad_norm_pre_clip': 9.930071830749512, 'timing/data': 0.021835263818502426, 'timing/model': 1.446244040504098, 'learning_rate/qwen_vl_interface': 7.2088557094266885e-06, 'learning_rate/action_model': 7.208855709426688e-05, 'epoch': 1.46}) +[2026-06-16 00:00:35,451][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.21400129795074463, 'train/grad_norm_pre_clip': 10.601758003234863, 'timing/data': 0.02084300760179758, 'timing/model': 1.400364407338202, 'learning_rate/qwen_vl_interface': 7.2060809064570195e-06, 'learning_rate/action_model': 7.20608090645702e-05, 'epoch': 1.46}) +[2026-06-16 00:00:41,256][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.07223755121231079, 'train/grad_norm_pre_clip': 9.23149299621582, 'timing/data': 0.02103360928595066, 'timing/model': 1.3988472316414118, 'learning_rate/qwen_vl_interface': 7.203305299416375e-06, 'learning_rate/action_model': 7.203305299416375e-05, 'epoch': 1.46}) +[2026-06-16 00:00:47,066][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.024344779551029205, 'train/grad_norm_pre_clip': 2.587442398071289, 'timing/data': 0.022133064456284046, 'timing/model': 1.6486745160073042, 'learning_rate/qwen_vl_interface': 7.200528889445707e-06, 'learning_rate/action_model': 7.200528889445706e-05, 'epoch': 1.46}) +[2026-06-16 00:00:52,651][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.047391489148139954, 'train/grad_norm_pre_clip': 4.245805740356445, 'timing/data': 0.02221680525690317, 'timing/model': 1.4374443469569087, 'learning_rate/qwen_vl_interface': 7.19775167768629e-06, 'learning_rate/action_model': 7.19775167768629e-05, 'epoch': 1.46}) +[2026-06-16 00:00:58,445][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.15222109854221344, 'train/grad_norm_pre_clip': 5.258935928344727, 'timing/data': 0.021344199776649475, 'timing/model': 1.4377858024090528, 'learning_rate/qwen_vl_interface': 7.194973665279731e-06, 'learning_rate/action_model': 7.19497366527973e-05, 'epoch': 1.46}) +[2026-06-16 00:01:04,262][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.30987411737442017, 'train/grad_norm_pre_clip': 11.210599899291992, 'timing/data': 0.020905948244035244, 'timing/model': 1.4655960379168391, 'learning_rate/qwen_vl_interface': 7.1921948533679666e-06, 'learning_rate/action_model': 7.192194853367967e-05, 'epoch': 1.46}) +[2026-06-16 00:01:10,048][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.246608167886734, 'train/grad_norm_pre_clip': 10.566296577453613, 'timing/data': 0.020181985571980476, 'timing/model': 1.3926048735156655, 'learning_rate/qwen_vl_interface': 7.1894152430932605e-06, 'learning_rate/action_model': 7.189415243093261e-05, 'epoch': 1.46}) +[2026-06-16 00:01:15,904][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.1682247668504715, 'train/grad_norm_pre_clip': 8.997039794921875, 'timing/data': 0.020351414568722248, 'timing/model': 1.6977198356762528, 'learning_rate/qwen_vl_interface': 7.186634835598205e-06, 'learning_rate/action_model': 7.186634835598205e-05, 'epoch': 1.46}) +[2026-06-16 00:01:21,556][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.11734764277935028, 'train/grad_norm_pre_clip': 16.325706481933594, 'timing/data': 0.023737313225865364, 'timing/model': 1.448478008620441, 'learning_rate/qwen_vl_interface': 7.183853632025724e-06, 'learning_rate/action_model': 7.183853632025723e-05, 'epoch': 1.46}) +[2026-06-16 00:01:27,354][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.0916280597448349, 'train/grad_norm_pre_clip': 8.770246505737305, 'timing/data': 0.020726757124066353, 'timing/model': 1.4460696391761303, 'learning_rate/qwen_vl_interface': 7.181071633519062e-06, 'learning_rate/action_model': 7.181071633519062e-05, 'epoch': 1.46}) +[2026-06-16 00:01:33,058][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.15139958262443542, 'train/grad_norm_pre_clip': 7.646603107452393, 'timing/data': 0.020804567262530327, 'timing/model': 1.4564168443903327, 'learning_rate/qwen_vl_interface': 7.178288841221795e-06, 'learning_rate/action_model': 7.178288841221794e-05, 'epoch': 1.46}) +[2026-06-16 00:01:38,845][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.09926222264766693, 'train/grad_norm_pre_clip': 4.132362365722656, 'timing/data': 0.021182949654757977, 'timing/model': 1.4374379003420472, 'learning_rate/qwen_vl_interface': 7.1755052562778215e-06, 'learning_rate/action_model': 7.175505256277822e-05, 'epoch': 1.46}) +[2026-06-16 00:12:15,772][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.0870119035243988, 'train/grad_norm_pre_clip': 4.936461925506592, 'eval/loss': 0.08000534027814865, 'eval/flap_precision': 0.773851590106007, 'eval/flap_recall': 0.7963636363636364, 'eval/flap_f1': 0.7849462365591398, 'eval/flap_rate_model': 0.07861111111111112, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7849462365591398, 'timing/data': 0.027259272523224354, 'timing/model': 1.4864220367744565, 'learning_rate/qwen_vl_interface': 7.1727208798313704e-06, 'learning_rate/action_model': 7.172720879831371e-05, 'epoch': 1.46}) +[2026-06-16 00:12:15,773][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3200_state +[2026-06-16 00:12:15,779][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 00:12:22,905][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3200_state/model.safetensors +[2026-06-16 00:12:34,192][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3200_state/optimizer.bin +[2026-06-16 00:12:34,193][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3200_state/sampler.bin +[2026-06-16 00:12:34,193][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3200_state/sampler_1.bin +[2026-06-16 00:12:34,195][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3200_state/random_states_0.pkl +[2026-06-16 00:12:34,345][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3200 +[2026-06-16 00:12:34,345][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 00:12:34,361][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 00:12:40,027][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.3306185305118561, 'train/grad_norm_pre_clip': 11.181159973144531, 'timing/data': 0.020747240632772446, 'timing/model': 1.3687223345041275, 'learning_rate/qwen_vl_interface': 7.169935713026994e-06, 'learning_rate/action_model': 7.169935713026994e-05, 'epoch': 1.46}) +[2026-06-16 00:12:45,525][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.12878906726837158, 'train/grad_norm_pre_clip': 5.158942222595215, 'timing/data': 0.0216418644413352, 'timing/model': 1.3980121705681086, 'learning_rate/qwen_vl_interface': 7.1671497570095684e-06, 'learning_rate/action_model': 7.167149757009568e-05, 'epoch': 1.46}) +[2026-06-16 00:12:51,104][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.13892263174057007, 'train/grad_norm_pre_clip': 5.569656848907471, 'timing/data': 0.021034270524978638, 'timing/model': 1.6231590136885643, 'learning_rate/qwen_vl_interface': 7.164363012924295e-06, 'learning_rate/action_model': 7.164363012924294e-05, 'epoch': 1.46}) +[2026-06-16 00:12:56,527][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.027210228145122528, 'train/grad_norm_pre_clip': 2.530494213104248, 'timing/data': 0.02036264818161726, 'timing/model': 1.4125117287039757, 'learning_rate/qwen_vl_interface': 7.161575481916699e-06, 'learning_rate/action_model': 7.1615754819167e-05, 'epoch': 1.47}) +[2026-06-16 00:13:02,227][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.11437533050775528, 'train/grad_norm_pre_clip': 7.701562404632568, 'timing/data': 0.021103881299495697, 'timing/model': 1.4358510840684175, 'learning_rate/qwen_vl_interface': 7.158787165132627e-06, 'learning_rate/action_model': 7.158787165132626e-05, 'epoch': 1.47}) +[2026-06-16 00:13:07,932][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.1508074551820755, 'train/grad_norm_pre_clip': 10.69073486328125, 'timing/data': 0.02155334409326315, 'timing/model': 1.4475859869271517, 'learning_rate/qwen_vl_interface': 7.1559980637182535e-06, 'learning_rate/action_model': 7.155998063718254e-05, 'epoch': 1.47}) +[2026-06-16 00:13:13,485][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.17872987687587738, 'train/grad_norm_pre_clip': 6.795664310455322, 'timing/data': 0.021914806216955185, 'timing/model': 1.404444538988173, 'learning_rate/qwen_vl_interface': 7.153208178820071e-06, 'learning_rate/action_model': 7.153208178820072e-05, 'epoch': 1.47}) +[2026-06-16 00:13:19,249][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.1294717937707901, 'train/grad_norm_pre_clip': 4.892833232879639, 'timing/data': 0.024681780487298965, 'timing/model': 1.4294060012325644, 'learning_rate/qwen_vl_interface': 7.150417511584898e-06, 'learning_rate/action_model': 7.150417511584898e-05, 'epoch': 1.47}) +[2026-06-16 00:13:24,987][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.1951613873243332, 'train/grad_norm_pre_clip': 10.096270561218262, 'timing/data': 0.020994541235268116, 'timing/model': 1.6952630020678043, 'learning_rate/qwen_vl_interface': 7.147626063159872e-06, 'learning_rate/action_model': 7.147626063159872e-05, 'epoch': 1.47}) +[2026-06-16 00:13:30,527][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.1490222066640854, 'train/grad_norm_pre_clip': 8.204802513122559, 'timing/data': 0.020482975989580154, 'timing/model': 1.4430221738293767, 'learning_rate/qwen_vl_interface': 7.14483383469245e-06, 'learning_rate/action_model': 7.144833834692449e-05, 'epoch': 1.47}) +[2026-06-16 00:13:36,240][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.24699468910694122, 'train/grad_norm_pre_clip': 8.968729019165039, 'timing/data': 0.023418999277055264, 'timing/model': 1.429352450184524, 'learning_rate/qwen_vl_interface': 7.142040827330411e-06, 'learning_rate/action_model': 7.142040827330411e-05, 'epoch': 1.47}) +[2026-06-16 00:13:41,939][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.029521381482481956, 'train/grad_norm_pre_clip': 2.006885528564453, 'timing/data': 0.020676187239587307, 'timing/model': 1.4466896699741483, 'learning_rate/qwen_vl_interface': 7.139247042221859e-06, 'learning_rate/action_model': 7.139247042221858e-05, 'epoch': 1.47}) +[2026-06-16 00:13:47,641][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.03830511495471001, 'train/grad_norm_pre_clip': 2.2691099643707275, 'timing/data': 0.020431235432624817, 'timing/model': 1.4529631501063704, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 1.47}) +[2026-06-16 00:13:53,346][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.13142558932304382, 'train/grad_norm_pre_clip': 6.546688079833984, 'timing/data': 0.020970099605619907, 'timing/model': 1.3855163231492043, 'learning_rate/qwen_vl_interface': 7.133657143359203e-06, 'learning_rate/action_model': 7.133657143359202e-05, 'epoch': 1.47}) +[2026-06-16 00:13:59,181][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.0800246074795723, 'train/grad_norm_pre_clip': 2.758483648300171, 'timing/data': 0.02255034539848566, 'timing/model': 1.6886906456202269, 'learning_rate/qwen_vl_interface': 7.130861031902897e-06, 'learning_rate/action_model': 7.130861031902897e-05, 'epoch': 1.47}) +[2026-06-16 00:14:04,744][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.14681395888328552, 'train/grad_norm_pre_clip': 6.1954240798950195, 'timing/data': 0.021030998788774014, 'timing/model': 1.4513003388419747, 'learning_rate/qwen_vl_interface': 7.128064147295667e-06, 'learning_rate/action_model': 7.128064147295667e-05, 'epoch': 1.47}) +[2026-06-16 00:14:10,565][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.11712551862001419, 'train/grad_norm_pre_clip': 11.006280899047852, 'timing/data': 0.021362380124628544, 'timing/model': 1.4691259833052754, 'learning_rate/qwen_vl_interface': 7.125266490687206e-06, 'learning_rate/action_model': 7.125266490687206e-05, 'epoch': 1.47}) +[2026-06-16 00:14:17,047][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.10601899027824402, 'train/grad_norm_pre_clip': 7.155394554138184, 'timing/data': 0.021195029839873314, 'timing/model': 1.4469029884785414, 'learning_rate/qwen_vl_interface': 7.1224680632275265e-06, 'learning_rate/action_model': 7.122468063227527e-05, 'epoch': 1.47}) +[2026-06-16 00:14:22,741][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.07305947691202164, 'train/grad_norm_pre_clip': 6.922116756439209, 'timing/data': 0.023735972121357918, 'timing/model': 1.4459810834378004, 'learning_rate/qwen_vl_interface': 7.1196688660669555e-06, 'learning_rate/action_model': 7.119668866066955e-05, 'epoch': 1.47}) +[2026-06-16 00:14:28,542][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.057708051055669785, 'train/grad_norm_pre_clip': 4.8131818771362305, 'timing/data': 0.020758799277246, 'timing/model': 1.3891004621982574, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 1.47}) +[2026-06-16 00:14:34,714][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.06340473890304565, 'train/grad_norm_pre_clip': 5.504700183868408, 'timing/data': 0.026414524763822556, 'timing/model': 2.002936859615147, 'learning_rate/qwen_vl_interface': 7.114068167246033e-06, 'learning_rate/action_model': 7.114068167246033e-05, 'epoch': 1.47}) +[2026-06-16 00:14:40,368][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.1552058756351471, 'train/grad_norm_pre_clip': 7.434604167938232, 'timing/data': 0.020717678591609, 'timing/model': 1.4625484384596348, 'learning_rate/qwen_vl_interface': 7.111266667887916e-06, 'learning_rate/action_model': 7.111266667887915e-05, 'epoch': 1.47}) +[2026-06-16 00:14:46,157][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.10350465029478073, 'train/grad_norm_pre_clip': 9.182397842407227, 'timing/data': 0.022161425091326237, 'timing/model': 1.4430113639682531, 'learning_rate/qwen_vl_interface': 7.108464403433378e-06, 'learning_rate/action_model': 7.108464403433378e-05, 'epoch': 1.47}) +[2026-06-16 00:14:51,959][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.06372243911027908, 'train/grad_norm_pre_clip': 3.1600935459136963, 'timing/data': 0.02208460494875908, 'timing/model': 1.4645581543445587, 'learning_rate/qwen_vl_interface': 7.105661375034324e-06, 'learning_rate/action_model': 7.105661375034324e-05, 'epoch': 1.47}) +[2026-06-16 00:14:57,742][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.07017622888088226, 'train/grad_norm_pre_clip': 5.335309028625488, 'timing/data': 0.02005184441804886, 'timing/model': 1.4550142278894782, 'learning_rate/qwen_vl_interface': 7.102857583842973e-06, 'learning_rate/action_model': 7.102857583842974e-05, 'epoch': 1.47}) +[2026-06-16 00:15:03,442][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.4062093198299408, 'train/grad_norm_pre_clip': 29.917997360229492, 'timing/data': 0.019533423706889153, 'timing/model': 1.405845251865685, 'learning_rate/qwen_vl_interface': 7.100053031011859e-06, 'learning_rate/action_model': 7.100053031011859e-05, 'epoch': 1.48}) +[2026-06-16 00:15:09,185][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.17847579717636108, 'train/grad_norm_pre_clip': 8.567075729370117, 'timing/data': 0.019782673567533493, 'timing/model': 1.6800654977560043, 'learning_rate/qwen_vl_interface': 7.097247717693827e-06, 'learning_rate/action_model': 7.097247717693827e-05, 'epoch': 1.48}) +[2026-06-16 00:15:14,930][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.017537452280521393, 'train/grad_norm_pre_clip': 1.3518978357315063, 'timing/data': 0.02039598673582077, 'timing/model': 1.623739467933774, 'learning_rate/qwen_vl_interface': 7.094441645042033e-06, 'learning_rate/action_model': 7.094441645042032e-05, 'epoch': 1.48}) +[2026-06-16 00:15:20,453][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.1350005567073822, 'train/grad_norm_pre_clip': 6.464160919189453, 'timing/data': 0.021663093008100986, 'timing/model': 1.4472313830628991, 'learning_rate/qwen_vl_interface': 7.09163481420995e-06, 'learning_rate/action_model': 7.09163481420995e-05, 'epoch': 1.48}) +[2026-06-16 00:15:26,149][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.16385699808597565, 'train/grad_norm_pre_clip': 11.243087768554688, 'timing/data': 0.022145896218717098, 'timing/model': 1.39550238288939, 'learning_rate/qwen_vl_interface': 7.0888272263513595e-06, 'learning_rate/action_model': 7.088827226351359e-05, 'epoch': 1.48}) +[2026-06-16 00:15:31,841][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.14727285504341125, 'train/grad_norm_pre_clip': 9.055510520935059, 'timing/data': 0.019613204523921013, 'timing/model': 1.449953860603273, 'learning_rate/qwen_vl_interface': 7.086018882620354e-06, 'learning_rate/action_model': 7.086018882620354e-05, 'epoch': 1.48}) +[2026-06-16 00:15:37,543][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.18229688704013824, 'train/grad_norm_pre_clip': 15.149514198303223, 'timing/data': 0.020566249266266823, 'timing/model': 1.4092217832803726, 'learning_rate/qwen_vl_interface': 7.083209784171341e-06, 'learning_rate/action_model': 7.083209784171341e-05, 'epoch': 1.48}) +[2026-06-16 00:15:43,319][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.10442648828029633, 'train/grad_norm_pre_clip': 18.13358497619629, 'timing/data': 0.019634975120425224, 'timing/model': 1.6593260811641812, 'learning_rate/qwen_vl_interface': 7.080399932159029e-06, 'learning_rate/action_model': 7.080399932159029e-05, 'epoch': 1.48}) +[2026-06-16 00:15:49,092][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.17651359736919403, 'train/grad_norm_pre_clip': 14.307426452636719, 'timing/data': 0.02089982107281685, 'timing/model': 1.67657898273319, 'learning_rate/qwen_vl_interface': 7.077589327738448e-06, 'learning_rate/action_model': 7.077589327738447e-05, 'epoch': 1.48}) +[2026-06-16 00:15:54,747][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.10996486991643906, 'train/grad_norm_pre_clip': 8.519722938537598, 'timing/data': 0.021494803950190544, 'timing/model': 1.4467403357848525, 'learning_rate/qwen_vl_interface': 7.0747779720649265e-06, 'learning_rate/action_model': 7.074777972064927e-05, 'epoch': 1.48}) +[2026-06-16 00:16:00,551][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.10981842875480652, 'train/grad_norm_pre_clip': 5.959220886230469, 'timing/data': 0.021456653252243996, 'timing/model': 1.4364351844415069, 'learning_rate/qwen_vl_interface': 7.071965866294108e-06, 'learning_rate/action_model': 7.071965866294108e-05, 'epoch': 1.48}) +[2026-06-16 00:16:06,352][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.14978419244289398, 'train/grad_norm_pre_clip': 12.248714447021484, 'timing/data': 0.02124113216996193, 'timing/model': 1.4511894527822733, 'learning_rate/qwen_vl_interface': 7.0691530115819454e-06, 'learning_rate/action_model': 7.069153011581945e-05, 'epoch': 1.48}) +[2026-06-16 00:16:12,149][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.20211292803287506, 'train/grad_norm_pre_clip': 11.048237800598145, 'timing/data': 0.026523438282310963, 'timing/model': 1.4059837218374014, 'learning_rate/qwen_vl_interface': 7.066339409084694e-06, 'learning_rate/action_model': 7.066339409084694e-05, 'epoch': 1.48}) +[2026-06-16 00:16:17,996][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.40477606654167175, 'train/grad_norm_pre_clip': 35.943641662597656, 'timing/data': 0.020985521376132965, 'timing/model': 1.6863127835094929, 'learning_rate/qwen_vl_interface': 7.0635250599589215e-06, 'learning_rate/action_model': 7.06352505995892e-05, 'epoch': 1.48}) +[2026-06-16 00:16:23,655][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.04702430218458176, 'train/grad_norm_pre_clip': 5.918103218078613, 'timing/data': 0.021636283956468105, 'timing/model': 1.4541742196306586, 'learning_rate/qwen_vl_interface': 7.060709965361499e-06, 'learning_rate/action_model': 7.060709965361499e-05, 'epoch': 1.48}) +[2026-06-16 00:16:29,450][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.17415845394134521, 'train/grad_norm_pre_clip': 9.583684921264648, 'timing/data': 0.020403578877449036, 'timing/model': 1.4290703628212214, 'learning_rate/qwen_vl_interface': 7.057894126449606e-06, 'learning_rate/action_model': 7.057894126449606e-05, 'epoch': 1.48}) +[2026-06-16 00:16:35,265][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.15842586755752563, 'train/grad_norm_pre_clip': 6.768680095672607, 'timing/data': 0.020731080323457718, 'timing/model': 1.462272902019322, 'learning_rate/qwen_vl_interface': 7.055077544380728e-06, 'learning_rate/action_model': 7.055077544380728e-05, 'epoch': 1.48}) +[2026-06-16 00:16:41,056][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.12635135650634766, 'train/grad_norm_pre_clip': 8.77965259552002, 'timing/data': 0.0211520129814744, 'timing/model': 1.4182189032435417, 'learning_rate/qwen_vl_interface': 7.0522602203126554e-06, 'learning_rate/action_model': 7.052260220312656e-05, 'epoch': 1.48}) +[2026-06-16 00:16:46,793][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.08419256657361984, 'train/grad_norm_pre_clip': 9.162063598632812, 'timing/data': 0.020175477489829063, 'timing/model': 1.4127943348139524, 'learning_rate/qwen_vl_interface': 7.0494421554034845e-06, 'learning_rate/action_model': 7.049442155403484e-05, 'epoch': 1.48}) +[2026-06-16 00:16:52,648][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.08192653208971024, 'train/grad_norm_pre_clip': 6.753177642822266, 'timing/data': 0.021079491823911667, 'timing/model': 1.6492015412077308, 'learning_rate/qwen_vl_interface': 7.046623350811612e-06, 'learning_rate/action_model': 7.046623350811612e-05, 'epoch': 1.48}) +[2026-06-16 00:16:58,253][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.33177676796913147, 'train/grad_norm_pre_clip': 13.203374862670898, 'timing/data': 0.02175797615200281, 'timing/model': 1.452551120892167, 'learning_rate/qwen_vl_interface': 7.043803807695747e-06, 'learning_rate/action_model': 7.043803807695747e-05, 'epoch': 1.48}) +[2026-06-16 00:17:04,052][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.17845411598682404, 'train/grad_norm_pre_clip': 15.478947639465332, 'timing/data': 0.020663220435380936, 'timing/model': 1.4321837285533547, 'learning_rate/qwen_vl_interface': 7.040983527214893e-06, 'learning_rate/action_model': 7.040983527214893e-05, 'epoch': 1.48}) +[2026-06-16 00:17:09,862][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.16783246397972107, 'train/grad_norm_pre_clip': 11.485840797424316, 'timing/data': 0.02322565298527479, 'timing/model': 1.4505971651524305, 'learning_rate/qwen_vl_interface': 7.038162510528363e-06, 'learning_rate/action_model': 7.038162510528362e-05, 'epoch': 1.49}) +[2026-06-16 00:17:15,655][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.1619398444890976, 'train/grad_norm_pre_clip': 7.853424072265625, 'timing/data': 0.021447734907269478, 'timing/model': 1.409965692088008, 'learning_rate/qwen_vl_interface': 7.03534075879577e-06, 'learning_rate/action_model': 7.03534075879577e-05, 'epoch': 1.49}) +[2026-06-16 00:17:21,460][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.33369866013526917, 'train/grad_norm_pre_clip': 11.190850257873535, 'timing/data': 0.025658984668552876, 'timing/model': 1.4147972222417593, 'learning_rate/qwen_vl_interface': 7.032518273177028e-06, 'learning_rate/action_model': 7.032518273177028e-05, 'epoch': 1.49}) +[2026-06-16 00:17:27,255][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.18541422486305237, 'train/grad_norm_pre_clip': 6.153186798095703, 'timing/data': 0.02286980114877224, 'timing/model': 1.646610631607473, 'learning_rate/qwen_vl_interface': 7.029695054832357e-06, 'learning_rate/action_model': 7.029695054832357e-05, 'epoch': 1.49}) +[2026-06-16 00:17:32,880][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.1974138468503952, 'train/grad_norm_pre_clip': 7.615512371063232, 'timing/data': 0.02069898135960102, 'timing/model': 1.4626975385472178, 'learning_rate/qwen_vl_interface': 7.026871104922273e-06, 'learning_rate/action_model': 7.026871104922273e-05, 'epoch': 1.49}) +[2026-06-16 00:17:38,657][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.21625056862831116, 'train/grad_norm_pre_clip': 8.898069381713867, 'timing/data': 0.02195016760379076, 'timing/model': 1.4386972468346357, 'learning_rate/qwen_vl_interface': 7.024046424607596e-06, 'learning_rate/action_model': 7.024046424607595e-05, 'epoch': 1.49}) +[2026-06-16 00:17:44,452][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.06556269526481628, 'train/grad_norm_pre_clip': 2.3651726245880127, 'timing/data': 0.02044960018247366, 'timing/model': 1.4517577784135938, 'learning_rate/qwen_vl_interface': 7.021221015049445e-06, 'learning_rate/action_model': 7.021221015049446e-05, 'epoch': 1.49}) +[2026-06-16 00:17:50,258][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.27915599942207336, 'train/grad_norm_pre_clip': 6.149759292602539, 'timing/data': 0.02089296281337738, 'timing/model': 1.394970185123384, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 1.49}) +[2026-06-16 00:17:56,088][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.1674899011850357, 'train/grad_norm_pre_clip': 16.495119094848633, 'timing/data': 0.020601500757038593, 'timing/model': 1.6689600059762597, 'learning_rate/qwen_vl_interface': 7.015568012848699e-06, 'learning_rate/action_model': 7.015568012848699e-05, 'epoch': 1.49}) +[2026-06-16 00:18:01,753][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.23732249438762665, 'train/grad_norm_pre_clip': 13.300841331481934, 'timing/data': 0.023593836463987827, 'timing/model': 1.4378755362704396, 'learning_rate/qwen_vl_interface': 7.01274042252984e-06, 'learning_rate/action_model': 7.01274042252984e-05, 'epoch': 1.49}) +[2026-06-16 00:18:07,550][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.13450217247009277, 'train/grad_norm_pre_clip': 3.925241231918335, 'timing/data': 0.020885352045297623, 'timing/model': 1.4368843380361795, 'learning_rate/qwen_vl_interface': 7.009912107614978e-06, 'learning_rate/action_model': 7.009912107614978e-05, 'epoch': 1.49}) +[2026-06-16 00:18:13,251][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.14554241299629211, 'train/grad_norm_pre_clip': 5.937024116516113, 'timing/data': 0.02055944036692381, 'timing/model': 1.4485854608938098, 'learning_rate/qwen_vl_interface': 7.0070830692667244e-06, 'learning_rate/action_model': 7.007083069266724e-05, 'epoch': 1.49}) +[2026-06-16 00:18:19,053][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.08079876005649567, 'train/grad_norm_pre_clip': 4.274316310882568, 'timing/data': 0.020633471198379993, 'timing/model': 1.4611818492412567, 'learning_rate/qwen_vl_interface': 7.004253308647993e-06, 'learning_rate/action_model': 7.004253308647994e-05, 'epoch': 1.49}) +[2026-06-16 00:18:24,852][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.20831675827503204, 'train/grad_norm_pre_clip': 7.792945384979248, 'timing/data': 0.021060634404420853, 'timing/model': 1.4447351330891252, 'learning_rate/qwen_vl_interface': 7.001422826921991e-06, 'learning_rate/action_model': 7.00142282692199e-05, 'epoch': 1.49}) +[2026-06-16 00:18:30,688][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.08356031030416489, 'train/grad_norm_pre_clip': 3.4011385440826416, 'timing/data': 0.020698091946542263, 'timing/model': 1.66564538795501, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 1.49}) +[2026-06-16 00:18:36,352][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.129645437002182, 'train/grad_norm_pre_clip': 6.389770030975342, 'timing/data': 0.020693342201411724, 'timing/model': 1.439252007752657, 'learning_rate/qwen_vl_interface': 6.995759704802485e-06, 'learning_rate/action_model': 6.995759704802485e-05, 'epoch': 1.49}) +[2026-06-16 00:18:42,152][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.2272460013628006, 'train/grad_norm_pre_clip': 8.19667911529541, 'timing/data': 0.020987173542380333, 'timing/model': 1.4322802377864718, 'learning_rate/qwen_vl_interface': 6.992927066736876e-06, 'learning_rate/action_model': 6.992927066736875e-05, 'epoch': 1.49}) +[2026-06-16 00:18:47,955][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.05416732653975487, 'train/grad_norm_pre_clip': 5.5101728439331055, 'timing/data': 0.02100953459739685, 'timing/model': 1.4433548981323838, 'learning_rate/qwen_vl_interface': 6.990093712219787e-06, 'learning_rate/action_model': 6.990093712219787e-05, 'epoch': 1.49}) +[2026-06-16 00:18:53,752][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.11357734352350235, 'train/grad_norm_pre_clip': 6.143925189971924, 'timing/data': 0.02181818801909685, 'timing/model': 1.439890786074102, 'learning_rate/qwen_vl_interface': 6.9872596424159e-06, 'learning_rate/action_model': 6.9872596424159e-05, 'epoch': 1.49}) +[2026-06-16 00:18:59,554][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.16251634061336517, 'train/grad_norm_pre_clip': 4.203611850738525, 'timing/data': 0.0214823167771101, 'timing/model': 1.414087662473321, 'learning_rate/qwen_vl_interface': 6.9844248584902e-06, 'learning_rate/action_model': 6.984424858490199e-05, 'epoch': 1.49}) +[2026-06-16 00:19:05,395][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.37904220819473267, 'train/grad_norm_pre_clip': 11.380841255187988, 'timing/data': 0.0209376635029912, 'timing/model': 1.687766645103693, 'learning_rate/qwen_vl_interface': 6.981589361607952e-06, 'learning_rate/action_model': 6.981589361607953e-05, 'epoch': 1.49}) +[2026-06-16 00:19:11,065][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.22176606953144073, 'train/grad_norm_pre_clip': 11.87655258178711, 'timing/data': 0.02084291446954012, 'timing/model': 1.439712230116129, 'learning_rate/qwen_vl_interface': 6.9787531529347274e-06, 'learning_rate/action_model': 6.978753152934727e-05, 'epoch': 1.5}) +[2026-06-16 00:19:16,852][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.05819202959537506, 'train/grad_norm_pre_clip': 8.151978492736816, 'timing/data': 0.020152490586042404, 'timing/model': 1.4284025598317385, 'learning_rate/qwen_vl_interface': 6.97591623363638e-06, 'learning_rate/action_model': 6.97591623363638e-05, 'epoch': 1.5}) +[2026-06-16 00:19:22,650][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.04419443756341934, 'train/grad_norm_pre_clip': 3.2564499378204346, 'timing/data': 0.020507131703197956, 'timing/model': 1.4493710482493043, 'learning_rate/qwen_vl_interface': 6.9730786048790644e-06, 'learning_rate/action_model': 6.973078604879064e-05, 'epoch': 1.5}) +[2026-06-16 00:19:28,453][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.1899719536304474, 'train/grad_norm_pre_clip': 8.77159309387207, 'timing/data': 0.0218502189964056, 'timing/model': 1.4240353889763355, 'learning_rate/qwen_vl_interface': 6.970240267829221e-06, 'learning_rate/action_model': 6.970240267829222e-05, 'epoch': 1.5}) +[2026-06-16 00:19:34,256][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.17610116302967072, 'train/grad_norm_pre_clip': 5.148058891296387, 'timing/data': 0.02074278239160776, 'timing/model': 1.4421268543228507, 'learning_rate/qwen_vl_interface': 6.967401223653582e-06, 'learning_rate/action_model': 6.967401223653581e-05, 'epoch': 1.5}) +[2026-06-16 00:19:40,092][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.07648319751024246, 'train/grad_norm_pre_clip': 4.015676498413086, 'timing/data': 0.020480401813983917, 'timing/model': 1.684764645062387, 'learning_rate/qwen_vl_interface': 6.9645614735191724e-06, 'learning_rate/action_model': 6.964561473519172e-05, 'epoch': 1.5}) +[2026-06-16 00:19:45,752][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.11065104603767395, 'train/grad_norm_pre_clip': 3.7719011306762695, 'timing/data': 0.020611463114619255, 'timing/model': 1.4499613707885146, 'learning_rate/qwen_vl_interface': 6.961721018593305e-06, 'learning_rate/action_model': 6.961721018593305e-05, 'epoch': 1.5}) +[2026-06-16 00:19:51,551][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.13154123723506927, 'train/grad_norm_pre_clip': 5.199492454528809, 'timing/data': 0.021016265265643597, 'timing/model': 1.4500837856903672, 'learning_rate/qwen_vl_interface': 6.958879860043585e-06, 'learning_rate/action_model': 6.958879860043585e-05, 'epoch': 1.5}) +[2026-06-16 00:19:57,371][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.08722982555627823, 'train/grad_norm_pre_clip': 3.036186456680298, 'timing/data': 0.021278616040945053, 'timing/model': 1.471799898892641, 'learning_rate/qwen_vl_interface': 6.9560379990379025e-06, 'learning_rate/action_model': 6.956037999037902e-05, 'epoch': 1.5}) +[2026-06-16 00:20:03,162][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.041652243584394455, 'train/grad_norm_pre_clip': 3.587733030319214, 'timing/data': 0.02066660299897194, 'timing/model': 1.447025390341878, 'learning_rate/qwen_vl_interface': 6.953195436744442e-06, 'learning_rate/action_model': 6.953195436744442e-05, 'epoch': 1.5}) +[2026-06-16 00:20:08,950][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.17503976821899414, 'train/grad_norm_pre_clip': 10.118844985961914, 'timing/data': 0.01959305815398693, 'timing/model': 1.4641386298462749, 'learning_rate/qwen_vl_interface': 6.950352174331676e-06, 'learning_rate/action_model': 6.950352174331675e-05, 'epoch': 1.5}) +[2026-06-16 00:20:14,788][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.055003050714731216, 'train/grad_norm_pre_clip': 3.551603078842163, 'timing/data': 0.021415136754512787, 'timing/model': 1.6795506812632084, 'learning_rate/qwen_vl_interface': 6.947508212968356e-06, 'learning_rate/action_model': 6.947508212968356e-05, 'epoch': 1.5}) +[2026-06-16 00:20:20,455][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.1717226207256317, 'train/grad_norm_pre_clip': 7.588894367218018, 'timing/data': 0.021405566483736038, 'timing/model': 1.4407847533002496, 'learning_rate/qwen_vl_interface': 6.9446635538235295e-06, 'learning_rate/action_model': 6.944663553823529e-05, 'epoch': 1.5}) +[2026-06-16 00:20:26,255][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.0707898885011673, 'train/grad_norm_pre_clip': 4.220131874084473, 'timing/data': 0.02102094516158104, 'timing/model': 1.450022554025054, 'learning_rate/qwen_vl_interface': 6.94181819806653e-06, 'learning_rate/action_model': 6.94181819806653e-05, 'epoch': 1.5}) +[2026-06-16 00:20:31,962][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.09012769162654877, 'train/grad_norm_pre_clip': 5.080233573913574, 'timing/data': 0.020607393234968185, 'timing/model': 1.4511445434764028, 'learning_rate/qwen_vl_interface': 6.938972146866975e-06, 'learning_rate/action_model': 6.938972146866974e-05, 'epoch': 1.5}) +[2026-06-16 00:20:37,648][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.2064497172832489, 'train/grad_norm_pre_clip': 5.263158798217773, 'timing/data': 0.020854064263403416, 'timing/model': 1.4309407761320472, 'learning_rate/qwen_vl_interface': 6.936125401394769e-06, 'learning_rate/action_model': 6.936125401394768e-05, 'epoch': 1.5}) +[2026-06-16 00:20:43,451][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.018263423815369606, 'train/grad_norm_pre_clip': 1.328599452972412, 'timing/data': 0.019843820482492447, 'timing/model': 1.399907335639, 'learning_rate/qwen_vl_interface': 6.933277962820101e-06, 'learning_rate/action_model': 6.9332779628201e-05, 'epoch': 1.5}) +[2026-06-16 00:20:49,283][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.16608567535877228, 'train/grad_norm_pre_clip': 7.495497703552246, 'timing/data': 0.020364261232316494, 'timing/model': 1.660515165887773, 'learning_rate/qwen_vl_interface': 6.930429832313444e-06, 'learning_rate/action_model': 6.930429832313444e-05, 'epoch': 1.5}) +[2026-06-16 00:20:54,950][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.1068648099899292, 'train/grad_norm_pre_clip': 5.959273815155029, 'timing/data': 0.022363271564245224, 'timing/model': 1.4333294108510017, 'learning_rate/qwen_vl_interface': 6.92758101104556e-06, 'learning_rate/action_model': 6.92758101104556e-05, 'epoch': 1.5}) +[2026-06-16 00:21:00,748][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.04260813817381859, 'train/grad_norm_pre_clip': 3.288759231567383, 'timing/data': 0.019893160089850426, 'timing/model': 1.4411041224375367, 'learning_rate/qwen_vl_interface': 6.9247315001874885e-06, 'learning_rate/action_model': 6.924731500187488e-05, 'epoch': 1.5}) +[2026-06-16 00:21:06,548][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.11545515805482864, 'train/grad_norm_pre_clip': 16.771326065063477, 'timing/data': 0.02034366223961115, 'timing/model': 1.4390586577355862, 'learning_rate/qwen_vl_interface': 6.9218813009105605e-06, 'learning_rate/action_model': 6.92188130091056e-05, 'epoch': 1.5}) +[2026-06-16 00:21:12,365][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.2939456105232239, 'train/grad_norm_pre_clip': 18.23638916015625, 'timing/data': 0.02048338297754526, 'timing/model': 1.4590614223852754, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 1.5}) +[2026-06-16 00:21:18,158][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.22790704667568207, 'train/grad_norm_pre_clip': 16.34657859802246, 'timing/data': 0.020712023600935936, 'timing/model': 1.3936609774827957, 'learning_rate/qwen_vl_interface': 6.916178841786839e-06, 'learning_rate/action_model': 6.916178841786838e-05, 'epoch': 1.51}) +[2026-06-16 00:21:23,974][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.2067500501871109, 'train/grad_norm_pre_clip': 9.517707824707031, 'timing/data': 0.01958267856389284, 'timing/model': 1.653777088969946, 'learning_rate/qwen_vl_interface': 6.913326584284115e-06, 'learning_rate/action_model': 6.913326584284115e-05, 'epoch': 1.51}) +[2026-06-16 00:21:29,551][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.0516030378639698, 'train/grad_norm_pre_clip': 3.5641071796417236, 'timing/data': 0.019809739664196968, 'timing/model': 1.4334087036550045, 'learning_rate/qwen_vl_interface': 6.910473643050662e-06, 'learning_rate/action_model': 6.910473643050662e-05, 'epoch': 1.51}) +[2026-06-16 00:21:35,251][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.08387850224971771, 'train/grad_norm_pre_clip': 6.152475833892822, 'timing/data': 0.019356307573616505, 'timing/model': 1.3975583678111434, 'learning_rate/qwen_vl_interface': 6.907620019259215e-06, 'learning_rate/action_model': 6.907620019259216e-05, 'epoch': 1.51}) +[2026-06-16 00:21:40,966][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.23388294875621796, 'train/grad_norm_pre_clip': 10.336085319519043, 'timing/data': 0.02026793174445629, 'timing/model': 1.4657703647390008, 'learning_rate/qwen_vl_interface': 6.9047657140827935e-06, 'learning_rate/action_model': 6.904765714082794e-05, 'epoch': 1.51}) +[2026-06-16 00:21:46,761][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.044180504977703094, 'train/grad_norm_pre_clip': 4.429924011230469, 'timing/data': 0.019935200922191143, 'timing/model': 1.4542931215837598, 'learning_rate/qwen_vl_interface': 6.901910728694694e-06, 'learning_rate/action_model': 6.901910728694694e-05, 'epoch': 1.51}) +[2026-06-16 00:21:52,458][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.109778992831707, 'train/grad_norm_pre_clip': 5.8967742919921875, 'timing/data': 0.02011595107614994, 'timing/model': 1.393504599109292, 'learning_rate/qwen_vl_interface': 6.899055064268489e-06, 'learning_rate/action_model': 6.899055064268489e-05, 'epoch': 1.51}) +[2026-06-16 00:21:58,283][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.025207431986927986, 'train/grad_norm_pre_clip': 2.9415271282196045, 'timing/data': 0.02010842226445675, 'timing/model': 1.6740716667845845, 'learning_rate/qwen_vl_interface': 6.896198721978038e-06, 'learning_rate/action_model': 6.896198721978038e-05, 'epoch': 1.51}) +[2026-06-16 00:22:04,130][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.3498731553554535, 'train/grad_norm_pre_clip': 18.61397933959961, 'timing/data': 0.021190817467868328, 'timing/model': 1.631077385507524, 'learning_rate/qwen_vl_interface': 6.893341702997475e-06, 'learning_rate/action_model': 6.893341702997475e-05, 'epoch': 1.51}) +[2026-06-16 00:32:37,178][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.13559740781784058, 'train/grad_norm_pre_clip': 8.31574535369873, 'eval/loss': 0.09367496520280838, 'eval/flap_precision': 0.927461139896373, 'eval/flap_recall': 0.6509090909090909, 'eval/flap_f1': 0.7649572649572649, 'eval/flap_rate_model': 0.05361111111111111, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7649572649572649, 'timing/data': 0.020621994510293007, 'timing/model': 1.438770187087357, 'learning_rate/qwen_vl_interface': 6.890484008501211e-06, 'learning_rate/action_model': 6.890484008501211e-05, 'epoch': 1.51}) +[2026-06-16 00:32:37,179][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3300_state +[2026-06-16 00:32:37,185][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 00:32:45,147][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3300_state/model.safetensors +[2026-06-16 00:32:56,400][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3300_state/optimizer.bin +[2026-06-16 00:32:56,401][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3300_state/sampler.bin +[2026-06-16 00:32:56,402][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3300_state/sampler_1.bin +[2026-06-16 00:32:56,403][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3300_state/random_states_0.pkl +[2026-06-16 00:32:56,569][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3300 +[2026-06-16 00:32:56,570][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 00:32:56,586][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 00:33:02,267][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.22511827945709229, 'train/grad_norm_pre_clip': 11.153989791870117, 'timing/data': 0.025229251012206078, 'timing/model': 1.4758728966116905, 'learning_rate/qwen_vl_interface': 6.887625639663941e-06, 'learning_rate/action_model': 6.88762563966394e-05, 'epoch': 1.51}) +[2026-06-16 00:33:08,028][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.10164821892976761, 'train/grad_norm_pre_clip': 9.804803848266602, 'timing/data': 0.021017120219767094, 'timing/model': 1.4431648682802916, 'learning_rate/qwen_vl_interface': 6.884766597660626e-06, 'learning_rate/action_model': 6.884766597660626e-05, 'epoch': 1.51}) +[2026-06-16 00:33:13,754][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.07332991063594818, 'train/grad_norm_pre_clip': 8.56721305847168, 'timing/data': 0.02162957191467285, 'timing/model': 1.4700488587841392, 'learning_rate/qwen_vl_interface': 6.881906883666515e-06, 'learning_rate/action_model': 6.881906883666515e-05, 'epoch': 1.51}) +[2026-06-16 00:33:19,433][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.05831257626414299, 'train/grad_norm_pre_clip': 8.863387107849121, 'timing/data': 0.02066688798367977, 'timing/model': 1.4409884475171566, 'learning_rate/qwen_vl_interface': 6.879046498857125e-06, 'learning_rate/action_model': 6.879046498857124e-05, 'epoch': 1.51}) +[2026-06-16 00:33:25,138][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.13753077387809753, 'train/grad_norm_pre_clip': 10.079073905944824, 'timing/data': 0.02113575953990221, 'timing/model': 1.4215777479112148, 'learning_rate/qwen_vl_interface': 6.876185444408255e-06, 'learning_rate/action_model': 6.876185444408255e-05, 'epoch': 1.51}) +[2026-06-16 00:33:30,871][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.1509626805782318, 'train/grad_norm_pre_clip': 6.710541248321533, 'timing/data': 0.022055115550756454, 'timing/model': 1.6768614891916513, 'learning_rate/qwen_vl_interface': 6.873323721495976e-06, 'learning_rate/action_model': 6.873323721495976e-05, 'epoch': 1.51}) +[2026-06-16 00:33:36,332][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.26081669330596924, 'train/grad_norm_pre_clip': 16.253053665161133, 'timing/data': 0.020903928205370903, 'timing/model': 1.4229178270325065, 'learning_rate/qwen_vl_interface': 6.870461331296632e-06, 'learning_rate/action_model': 6.870461331296632e-05, 'epoch': 1.51}) +[2026-06-16 00:33:42,445][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.4715575575828552, 'train/grad_norm_pre_clip': 16.440217971801758, 'timing/data': 0.02119150012731552, 'timing/model': 1.4402010682970285, 'learning_rate/qwen_vl_interface': 6.8675982749868455e-06, 'learning_rate/action_model': 6.867598274986845e-05, 'epoch': 1.51}) +[2026-06-16 00:33:48,159][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.15581002831459045, 'train/grad_norm_pre_clip': 10.267753601074219, 'timing/data': 0.024763189256191254, 'timing/model': 1.4447622820734978, 'learning_rate/qwen_vl_interface': 6.864734553743509e-06, 'learning_rate/action_model': 6.86473455374351e-05, 'epoch': 1.51}) +[2026-06-16 00:33:53,931][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.04404960200190544, 'train/grad_norm_pre_clip': 4.297584056854248, 'timing/data': 0.0272231325507164, 'timing/model': 1.4030761355534196, 'learning_rate/qwen_vl_interface': 6.861870168743793e-06, 'learning_rate/action_model': 6.861870168743793e-05, 'epoch': 1.51}) +[2026-06-16 00:33:59,555][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.07262880355119705, 'train/grad_norm_pre_clip': 3.3816347122192383, 'timing/data': 0.02151166182011366, 'timing/model': 1.385170384310186, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 1.51}) +[2026-06-16 00:34:05,828][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.11399868875741959, 'train/grad_norm_pre_clip': 8.11393928527832, 'timing/data': 0.029347913339734077, 'timing/model': 2.1117539647966623, 'learning_rate/qwen_vl_interface': 6.85613941218525e-06, 'learning_rate/action_model': 6.85613941218525e-05, 'epoch': 1.51}) +[2026-06-16 00:34:11,454][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.19347265362739563, 'train/grad_norm_pre_clip': 15.642213821411133, 'timing/data': 0.028690509498119354, 'timing/model': 1.4358510188758373, 'learning_rate/qwen_vl_interface': 6.8532730429821215e-06, 'learning_rate/action_model': 6.853273042982122e-05, 'epoch': 1.52}) +[2026-06-16 00:34:17,452][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.030422167852520943, 'train/grad_norm_pre_clip': 3.8411648273468018, 'timing/data': 0.021209120750427246, 'timing/model': 1.4376659793779254, 'learning_rate/qwen_vl_interface': 6.850406014734005e-06, 'learning_rate/action_model': 6.850406014734005e-05, 'epoch': 1.52}) +[2026-06-16 00:34:23,247][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.17311948537826538, 'train/grad_norm_pre_clip': 8.579390525817871, 'timing/data': 0.02515965048223734, 'timing/model': 1.429431227967143, 'learning_rate/qwen_vl_interface': 6.8475383286194295e-06, 'learning_rate/action_model': 6.84753832861943e-05, 'epoch': 1.52}) +[2026-06-16 00:34:29,042][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.2398369461297989, 'train/grad_norm_pre_clip': 11.688629150390625, 'timing/data': 0.0217055631801486, 'timing/model': 1.399571342393756, 'learning_rate/qwen_vl_interface': 6.844669985817192e-06, 'learning_rate/action_model': 6.844669985817192e-05, 'epoch': 1.52}) +[2026-06-16 00:34:34,846][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.29239562153816223, 'train/grad_norm_pre_clip': 13.407443046569824, 'timing/data': 0.021577853709459305, 'timing/model': 1.4245420144870877, 'learning_rate/qwen_vl_interface': 6.8418009875063594e-06, 'learning_rate/action_model': 6.84180098750636e-05, 'epoch': 1.52}) +[2026-06-16 00:34:40,594][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.09148939698934555, 'train/grad_norm_pre_clip': 6.4339752197265625, 'timing/data': 0.021055460907518864, 'timing/model': 1.702973478473723, 'learning_rate/qwen_vl_interface': 6.838931334866267e-06, 'learning_rate/action_model': 6.838931334866266e-05, 'epoch': 1.52}) +[2026-06-16 00:34:46,146][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.10240748524665833, 'train/grad_norm_pre_clip': 4.671965599060059, 'timing/data': 0.021090940572321415, 'timing/model': 1.4545813230797648, 'learning_rate/qwen_vl_interface': 6.836061029076523e-06, 'learning_rate/action_model': 6.836061029076522e-05, 'epoch': 1.52}) +[2026-06-16 00:34:51,843][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.14782796800136566, 'train/grad_norm_pre_clip': 10.771138191223145, 'timing/data': 0.021507211960852146, 'timing/model': 1.419151958078146, 'learning_rate/qwen_vl_interface': 6.833190071317e-06, 'learning_rate/action_model': 6.833190071316999e-05, 'epoch': 1.52}) +[2026-06-16 00:34:57,647][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.16688545048236847, 'train/grad_norm_pre_clip': 7.691413402557373, 'timing/data': 0.02100333943963051, 'timing/model': 1.444239259697497, 'learning_rate/qwen_vl_interface': 6.830318462767841e-06, 'learning_rate/action_model': 6.830318462767841e-05, 'epoch': 1.52}) +[2026-06-16 00:35:03,354][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.2494073510169983, 'train/grad_norm_pre_clip': 11.308503150939941, 'timing/data': 0.02514997124671936, 'timing/model': 1.423442822881043, 'learning_rate/qwen_vl_interface': 6.827446204609456e-06, 'learning_rate/action_model': 6.827446204609456e-05, 'epoch': 1.52}) +[2026-06-16 00:35:09,047][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.19865378737449646, 'train/grad_norm_pre_clip': 9.891200065612793, 'timing/data': 0.021365780383348465, 'timing/model': 1.3927976237609982, 'learning_rate/qwen_vl_interface': 6.824573298022522e-06, 'learning_rate/action_model': 6.824573298022522e-05, 'epoch': 1.52}) +[2026-06-16 00:35:14,803][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.17666736245155334, 'train/grad_norm_pre_clip': 5.20697021484375, 'timing/data': 0.0210102004930377, 'timing/model': 1.647352373227477, 'learning_rate/qwen_vl_interface': 6.8216997441879805e-06, 'learning_rate/action_model': 6.821699744187981e-05, 'epoch': 1.52}) +[2026-06-16 00:35:20,348][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.06693446636199951, 'train/grad_norm_pre_clip': 2.779630661010742, 'timing/data': 0.021127049811184406, 'timing/model': 1.4494799515232444, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 1.52}) +[2026-06-16 00:35:26,056][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.1317191869020462, 'train/grad_norm_pre_clip': 7.433719158172607, 'timing/data': 0.02261440735310316, 'timing/model': 1.4018309731036425, 'learning_rate/qwen_vl_interface': 6.815950699501186e-06, 'learning_rate/action_model': 6.815950699501186e-05, 'epoch': 1.52}) +[2026-06-16 00:35:31,847][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.11166204512119293, 'train/grad_norm_pre_clip': 6.27672004699707, 'timing/data': 0.021009529940783978, 'timing/model': 1.4535194728523493, 'learning_rate/qwen_vl_interface': 6.813075211012148e-06, 'learning_rate/action_model': 6.813075211012147e-05, 'epoch': 1.52}) +[2026-06-16 00:35:37,549][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.1396905481815338, 'train/grad_norm_pre_clip': 4.971108436584473, 'timing/data': 0.024595697410404682, 'timing/model': 1.3996582627296448, 'learning_rate/qwen_vl_interface': 6.81019908000193e-06, 'learning_rate/action_model': 6.81019908000193e-05, 'epoch': 1.52}) +[2026-06-16 00:35:43,350][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.016376804560422897, 'train/grad_norm_pre_clip': 2.167753219604492, 'timing/data': 0.020566828548908234, 'timing/model': 1.4205301012843847, 'learning_rate/qwen_vl_interface': 6.807322307652806e-06, 'learning_rate/action_model': 6.807322307652806e-05, 'epoch': 1.52}) +[2026-06-16 00:35:49,200][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.08175688982009888, 'train/grad_norm_pre_clip': 4.095868110656738, 'timing/data': 0.02210727520287037, 'timing/model': 1.6838299185037613, 'learning_rate/qwen_vl_interface': 6.804444895147306e-06, 'learning_rate/action_model': 6.804444895147306e-05, 'epoch': 1.52}) +[2026-06-16 00:35:54,875][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.271950900554657, 'train/grad_norm_pre_clip': 14.297231674194336, 'timing/data': 0.027131671085953712, 'timing/model': 1.4475923348218203, 'learning_rate/qwen_vl_interface': 6.8015668436682255e-06, 'learning_rate/action_model': 6.801566843668226e-05, 'epoch': 1.52}) +[2026-06-16 00:36:00,754][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.03844546154141426, 'train/grad_norm_pre_clip': 2.040447235107422, 'timing/data': 0.020689288154244423, 'timing/model': 1.440690309740603, 'learning_rate/qwen_vl_interface': 6.7986881543986245e-06, 'learning_rate/action_model': 6.798688154398624e-05, 'epoch': 1.52}) +[2026-06-16 00:36:06,565][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.42895713448524475, 'train/grad_norm_pre_clip': 7.5601701736450195, 'timing/data': 0.02159797213971615, 'timing/model': 1.470233895815909, 'learning_rate/qwen_vl_interface': 6.795808828521822e-06, 'learning_rate/action_model': 6.795808828521821e-05, 'epoch': 1.52}) +[2026-06-16 00:36:12,348][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.09787106513977051, 'train/grad_norm_pre_clip': 4.855937480926514, 'timing/data': 0.02060850802809, 'timing/model': 1.447453087195754, 'learning_rate/qwen_vl_interface': 6.792928867221401e-06, 'learning_rate/action_model': 6.792928867221401e-05, 'epoch': 1.52}) +[2026-06-16 00:36:18,176][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.23452278971672058, 'train/grad_norm_pre_clip': 8.536908149719238, 'timing/data': 0.02089148946106434, 'timing/model': 1.4045927729457617, 'learning_rate/qwen_vl_interface': 6.790048271681202e-06, 'learning_rate/action_model': 6.790048271681202e-05, 'epoch': 1.53}) +[2026-06-16 00:36:24,035][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.04305502027273178, 'train/grad_norm_pre_clip': 3.3460922241210938, 'timing/data': 0.0224409569054842, 'timing/model': 1.7170089771971107, 'learning_rate/qwen_vl_interface': 6.7871670430853335e-06, 'learning_rate/action_model': 6.787167043085333e-05, 'epoch': 1.53}) +[2026-06-16 00:36:29,679][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.12494968622922897, 'train/grad_norm_pre_clip': 4.807849884033203, 'timing/data': 0.02209966629743576, 'timing/model': 1.4666654579341412, 'learning_rate/qwen_vl_interface': 6.784285182618156e-06, 'learning_rate/action_model': 6.784285182618156e-05, 'epoch': 1.53}) +[2026-06-16 00:36:35,564][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.2453143149614334, 'train/grad_norm_pre_clip': 7.824833869934082, 'timing/data': 0.021460212767124176, 'timing/model': 1.461203720420599, 'learning_rate/qwen_vl_interface': 6.781402691464297e-06, 'learning_rate/action_model': 6.781402691464297e-05, 'epoch': 1.53}) +[2026-06-16 00:36:41,357][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.11058181524276733, 'train/grad_norm_pre_clip': 5.16415548324585, 'timing/data': 0.020039265044033527, 'timing/model': 1.459056201390922, 'learning_rate/qwen_vl_interface': 6.778519570808636e-06, 'learning_rate/action_model': 6.778519570808636e-05, 'epoch': 1.53}) +[2026-06-16 00:36:47,155][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.13701219856739044, 'train/grad_norm_pre_clip': 4.808217525482178, 'timing/data': 0.021982314996421337, 'timing/model': 1.449107438325882, 'learning_rate/qwen_vl_interface': 6.775635821836317e-06, 'learning_rate/action_model': 6.775635821836317e-05, 'epoch': 1.53}) +[2026-06-16 00:36:52,964][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.0877179279923439, 'train/grad_norm_pre_clip': 5.396682262420654, 'timing/data': 0.022281796671450138, 'timing/model': 1.4006062857806683, 'learning_rate/qwen_vl_interface': 6.77275144573274e-06, 'learning_rate/action_model': 6.77275144573274e-05, 'epoch': 1.53}) +[2026-06-16 00:36:58,802][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.23141150176525116, 'train/grad_norm_pre_clip': 4.094455242156982, 'timing/data': 0.02170779276639223, 'timing/model': 1.6982389437034726, 'learning_rate/qwen_vl_interface': 6.769866443683564e-06, 'learning_rate/action_model': 6.769866443683563e-05, 'epoch': 1.53}) +[2026-06-16 00:37:04,454][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.20954450964927673, 'train/grad_norm_pre_clip': 10.420828819274902, 'timing/data': 0.025134201161563396, 'timing/model': 1.452174086123705, 'learning_rate/qwen_vl_interface': 6.766980816874701e-06, 'learning_rate/action_model': 6.7669808168747e-05, 'epoch': 1.53}) +[2026-06-16 00:37:10,722][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.4014044404029846, 'train/grad_norm_pre_clip': 10.719686508178711, 'timing/data': 0.020803148858249187, 'timing/model': 1.4367906581610441, 'learning_rate/qwen_vl_interface': 6.764094566492327e-06, 'learning_rate/action_model': 6.764094566492328e-05, 'epoch': 1.53}) +[2026-06-16 00:37:16,642][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.24384485185146332, 'train/grad_norm_pre_clip': 14.418864250183105, 'timing/data': 0.020292876288294792, 'timing/model': 1.4328403268009424, 'learning_rate/qwen_vl_interface': 6.7612076937228685e-06, 'learning_rate/action_model': 6.761207693722869e-05, 'epoch': 1.53}) +[2026-06-16 00:37:22,447][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.08382409811019897, 'train/grad_norm_pre_clip': 5.992038726806641, 'timing/data': 0.022433556616306305, 'timing/model': 1.4467735709622502, 'learning_rate/qwen_vl_interface': 6.7583201997530126e-06, 'learning_rate/action_model': 6.758320199753012e-05, 'epoch': 1.53}) +[2026-06-16 00:37:28,153][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.14349579811096191, 'train/grad_norm_pre_clip': 7.331357002258301, 'timing/data': 0.021590993739664555, 'timing/model': 1.3928268291056156, 'learning_rate/qwen_vl_interface': 6.755432085769693e-06, 'learning_rate/action_model': 6.755432085769694e-05, 'epoch': 1.53}) +[2026-06-16 00:37:34,029][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.19676095247268677, 'train/grad_norm_pre_clip': 7.516402244567871, 'timing/data': 0.0213645426556468, 'timing/model': 1.706243279390037, 'learning_rate/qwen_vl_interface': 6.75254335296011e-06, 'learning_rate/action_model': 6.75254335296011e-05, 'epoch': 1.53}) +[2026-06-16 00:37:39,655][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.2802000641822815, 'train/grad_norm_pre_clip': 7.162228107452393, 'timing/data': 0.02077213954180479, 'timing/model': 1.4454056145623326, 'learning_rate/qwen_vl_interface': 6.749654002511712e-06, 'learning_rate/action_model': 6.749654002511712e-05, 'epoch': 1.53}) +[2026-06-16 00:37:45,666][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.1918422430753708, 'train/grad_norm_pre_clip': 10.836580276489258, 'timing/data': 0.020582408644258976, 'timing/model': 1.4613845003768802, 'learning_rate/qwen_vl_interface': 6.7467640356122e-06, 'learning_rate/action_model': 6.746764035612199e-05, 'epoch': 1.53}) +[2026-06-16 00:37:51,455][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.1705447882413864, 'train/grad_norm_pre_clip': 4.106637477874756, 'timing/data': 0.02179610449820757, 'timing/model': 1.4543167129158974, 'learning_rate/qwen_vl_interface': 6.74387345344953e-06, 'learning_rate/action_model': 6.74387345344953e-05, 'epoch': 1.53}) +[2026-06-16 00:37:57,256][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.08502738922834396, 'train/grad_norm_pre_clip': 3.8716983795166016, 'timing/data': 0.024408357217907906, 'timing/model': 1.4371227528899908, 'learning_rate/qwen_vl_interface': 6.740982257211911e-06, 'learning_rate/action_model': 6.740982257211911e-05, 'epoch': 1.53}) +[2026-06-16 00:38:03,071][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.10632310062646866, 'train/grad_norm_pre_clip': 6.930813789367676, 'timing/data': 0.02391419466584921, 'timing/model': 1.4064788240939379, 'learning_rate/qwen_vl_interface': 6.738090448087806e-06, 'learning_rate/action_model': 6.738090448087806e-05, 'epoch': 1.53}) +[2026-06-16 00:38:08,820][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.11940716207027435, 'train/grad_norm_pre_clip': 3.746060848236084, 'timing/data': 0.023878615349531174, 'timing/model': 1.6628416860476136, 'learning_rate/qwen_vl_interface': 6.735198027265928e-06, 'learning_rate/action_model': 6.735198027265928e-05, 'epoch': 1.53}) +[2026-06-16 00:38:14,668][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.07213329523801804, 'train/grad_norm_pre_clip': 5.817781448364258, 'timing/data': 0.020020843483507633, 'timing/model': 1.6540708020329475, 'learning_rate/qwen_vl_interface': 6.7323049959352424e-06, 'learning_rate/action_model': 6.732304995935243e-05, 'epoch': 1.53}) +[2026-06-16 00:38:20,252][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.03953326493501663, 'train/grad_norm_pre_clip': 2.0860979557037354, 'timing/data': 0.021195760928094387, 'timing/model': 1.4374117888510227, 'learning_rate/qwen_vl_interface': 6.729411355284962e-06, 'learning_rate/action_model': 6.729411355284963e-05, 'epoch': 1.53}) +[2026-06-16 00:38:26,056][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.14404666423797607, 'train/grad_norm_pre_clip': 4.619701862335205, 'timing/data': 0.021837083622813225, 'timing/model': 1.44093500636518, 'learning_rate/qwen_vl_interface': 6.726517106504559e-06, 'learning_rate/action_model': 6.726517106504558e-05, 'epoch': 1.54}) +[2026-06-16 00:38:31,856][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.2908362150192261, 'train/grad_norm_pre_clip': 13.469710350036621, 'timing/data': 0.021971355192363262, 'timing/model': 1.442571995779872, 'learning_rate/qwen_vl_interface': 6.723622250783745e-06, 'learning_rate/action_model': 6.723622250783744e-05, 'epoch': 1.54}) +[2026-06-16 00:38:37,656][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.07471653819084167, 'train/grad_norm_pre_clip': 4.3496575355529785, 'timing/data': 0.02207651548087597, 'timing/model': 1.3987284768372774, 'learning_rate/qwen_vl_interface': 6.720726789312487e-06, 'learning_rate/action_model': 6.720726789312487e-05, 'epoch': 1.54}) +[2026-06-16 00:38:43,491][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.18475812673568726, 'train/grad_norm_pre_clip': 7.8712921142578125, 'timing/data': 0.021694984287023544, 'timing/model': 1.689101217314601, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 1.54}) +[2026-06-16 00:38:49,320][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.17862346768379211, 'train/grad_norm_pre_clip': 10.991449356079102, 'timing/data': 0.02038576640188694, 'timing/model': 1.674157108180225, 'learning_rate/qwen_vl_interface': 6.714934053879749e-06, 'learning_rate/action_model': 6.71493405387975e-05, 'epoch': 1.54}) +[2026-06-16 00:38:54,956][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.1347428560256958, 'train/grad_norm_pre_clip': 6.308004379272461, 'timing/data': 0.02122603077441454, 'timing/model': 1.4262204822152853, 'learning_rate/qwen_vl_interface': 6.712036782299442e-06, 'learning_rate/action_model': 6.712036782299443e-05, 'epoch': 1.54}) +[2026-06-16 00:39:00,775][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.12518061697483063, 'train/grad_norm_pre_clip': 8.016569137573242, 'timing/data': 0.022571238689124584, 'timing/model': 1.4594635656103492, 'learning_rate/qwen_vl_interface': 6.709138909731041e-06, 'learning_rate/action_model': 6.709138909731041e-05, 'epoch': 1.54}) +[2026-06-16 00:39:06,554][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.15166550874710083, 'train/grad_norm_pre_clip': 7.019289493560791, 'timing/data': 0.021205350756645203, 'timing/model': 1.4493892844766378, 'learning_rate/qwen_vl_interface': 6.706240437365751e-06, 'learning_rate/action_model': 6.70624043736575e-05, 'epoch': 1.54}) +[2026-06-16 00:39:12,355][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.26541247963905334, 'train/grad_norm_pre_clip': 12.120268821716309, 'timing/data': 0.02070232853293419, 'timing/model': 1.4054555548354983, 'learning_rate/qwen_vl_interface': 6.703341366395023e-06, 'learning_rate/action_model': 6.703341366395023e-05, 'epoch': 1.54}) +[2026-06-16 00:39:18,212][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.10075454413890839, 'train/grad_norm_pre_clip': 8.315738677978516, 'timing/data': 0.02532031200826168, 'timing/model': 1.6893649632111192, 'learning_rate/qwen_vl_interface': 6.700441698010558e-06, 'learning_rate/action_model': 6.700441698010558e-05, 'epoch': 1.54}) +[2026-06-16 00:39:23,855][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.22815802693367004, 'train/grad_norm_pre_clip': 8.557634353637695, 'timing/data': 0.021303421817719936, 'timing/model': 1.4409718420356512, 'learning_rate/qwen_vl_interface': 6.697541433404299e-06, 'learning_rate/action_model': 6.697541433404299e-05, 'epoch': 1.54}) +[2026-06-16 00:39:29,679][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.19906656444072723, 'train/grad_norm_pre_clip': 9.49807357788086, 'timing/data': 0.02103964053094387, 'timing/model': 1.462366254068911, 'learning_rate/qwen_vl_interface': 6.6946405737684325e-06, 'learning_rate/action_model': 6.694640573768433e-05, 'epoch': 1.54}) +[2026-06-16 00:39:35,682][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.09252091497182846, 'train/grad_norm_pre_clip': 4.9054670333862305, 'timing/data': 0.029259873554110527, 'timing/model': 1.4671644400805235, 'learning_rate/qwen_vl_interface': 6.691739120295397e-06, 'learning_rate/action_model': 6.691739120295397e-05, 'epoch': 1.54}) +[2026-06-16 00:39:41,581][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.2839035987854004, 'train/grad_norm_pre_clip': 8.286469459533691, 'timing/data': 0.02207790594547987, 'timing/model': 1.4146578256040812, 'learning_rate/qwen_vl_interface': 6.688837074177868e-06, 'learning_rate/action_model': 6.688837074177868e-05, 'epoch': 1.54}) +[2026-06-16 00:39:47,354][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.20308777689933777, 'train/grad_norm_pre_clip': 10.23108959197998, 'timing/data': 0.021926655434072018, 'timing/model': 1.4480495704337955, 'learning_rate/qwen_vl_interface': 6.68593443660877e-06, 'learning_rate/action_model': 6.685934436608769e-05, 'epoch': 1.54}) +[2026-06-16 00:39:53,166][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.29942935705184937, 'train/grad_norm_pre_clip': 9.188515663146973, 'timing/data': 0.021437881514430046, 'timing/model': 1.6607608264312148, 'learning_rate/qwen_vl_interface': 6.6830312087812625e-06, 'learning_rate/action_model': 6.683031208781262e-05, 'epoch': 1.54}) +[2026-06-16 00:39:58,754][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.05525951087474823, 'train/grad_norm_pre_clip': 2.8043885231018066, 'timing/data': 0.02182486467063427, 'timing/model': 1.4340864177793264, 'learning_rate/qwen_vl_interface': 6.680127391888754e-06, 'learning_rate/action_model': 6.680127391888754e-05, 'epoch': 1.54}) +[2026-06-16 00:40:04,556][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.03469778224825859, 'train/grad_norm_pre_clip': 2.4427151679992676, 'timing/data': 0.020908448845148087, 'timing/model': 1.457816912792623, 'learning_rate/qwen_vl_interface': 6.677222987124897e-06, 'learning_rate/action_model': 6.677222987124897e-05, 'epoch': 1.54}) +[2026-06-16 00:40:10,357][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.18481799960136414, 'train/grad_norm_pre_clip': 7.194012641906738, 'timing/data': 0.021349411457777023, 'timing/model': 1.4411603352054954, 'learning_rate/qwen_vl_interface': 6.67431799568358e-06, 'learning_rate/action_model': 6.67431799568358e-05, 'epoch': 1.54}) +[2026-06-16 00:40:16,173][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.10886384546756744, 'train/grad_norm_pre_clip': 4.93501091003418, 'timing/data': 0.020969320088624954, 'timing/model': 1.4021702613681555, 'learning_rate/qwen_vl_interface': 6.671412418758939e-06, 'learning_rate/action_model': 6.671412418758938e-05, 'epoch': 1.54}) +[2026-06-16 00:40:21,997][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.05841393768787384, 'train/grad_norm_pre_clip': 3.1019058227539062, 'timing/data': 0.02093311958014965, 'timing/model': 1.4211619216948748, 'learning_rate/qwen_vl_interface': 6.668506257545341e-06, 'learning_rate/action_model': 6.668506257545341e-05, 'epoch': 1.54}) +[2026-06-16 00:40:27,878][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.06787046045064926, 'train/grad_norm_pre_clip': 3.770326614379883, 'timing/data': 0.022255376912653446, 'timing/model': 1.659229221753776, 'learning_rate/qwen_vl_interface': 6.665599513237403e-06, 'learning_rate/action_model': 6.665599513237402e-05, 'epoch': 1.54}) +[2026-06-16 00:40:33,452][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.31632906198501587, 'train/grad_norm_pre_clip': 12.448400497436523, 'timing/data': 0.02147866226732731, 'timing/model': 1.4374856194481254, 'learning_rate/qwen_vl_interface': 6.662692187029978e-06, 'learning_rate/action_model': 6.662692187029978e-05, 'epoch': 1.55}) +[2026-06-16 00:40:39,255][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.2698379158973694, 'train/grad_norm_pre_clip': 9.305636405944824, 'timing/data': 0.021052749827504158, 'timing/model': 1.4401075625792146, 'learning_rate/qwen_vl_interface': 6.659784280118156e-06, 'learning_rate/action_model': 6.659784280118156e-05, 'epoch': 1.55}) +[2026-06-16 00:40:45,056][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.10874811559915543, 'train/grad_norm_pre_clip': 3.397500991821289, 'timing/data': 0.02213361579924822, 'timing/model': 1.4339345311746001, 'learning_rate/qwen_vl_interface': 6.65687579369727e-06, 'learning_rate/action_model': 6.656875793697271e-05, 'epoch': 1.55}) +[2026-06-16 00:40:50,892][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.3548104166984558, 'train/grad_norm_pre_clip': 11.485098838806152, 'timing/data': 0.022106275893747807, 'timing/model': 1.4112713923677802, 'learning_rate/qwen_vl_interface': 6.653966728962885e-06, 'learning_rate/action_model': 6.653966728962885e-05, 'epoch': 1.55}) +[2026-06-16 00:40:56,806][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.15787693858146667, 'train/grad_norm_pre_clip': 8.34856128692627, 'timing/data': 0.02089276909828186, 'timing/model': 1.6896327529102564, 'learning_rate/qwen_vl_interface': 6.651057087110813e-06, 'learning_rate/action_model': 6.651057087110813e-05, 'epoch': 1.55}) +[2026-06-16 00:41:02,455][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.2181411236524582, 'train/grad_norm_pre_clip': 11.179386138916016, 'timing/data': 0.021151510998606682, 'timing/model': 1.4505058089271188, 'learning_rate/qwen_vl_interface': 6.6481468693370924e-06, 'learning_rate/action_model': 6.648146869337091e-05, 'epoch': 1.55}) +[2026-06-16 00:41:08,260][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.17134365439414978, 'train/grad_norm_pre_clip': 5.895969867706299, 'timing/data': 0.022418247535824776, 'timing/model': 1.4569309428334236, 'learning_rate/qwen_vl_interface': 6.645236076838009e-06, 'learning_rate/action_model': 6.645236076838009e-05, 'epoch': 1.55}) +[2026-06-16 00:41:13,953][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.08075188100337982, 'train/grad_norm_pre_clip': 3.2877073287963867, 'timing/data': 0.021271691657602787, 'timing/model': 1.4381596148014069, 'learning_rate/qwen_vl_interface': 6.642324710810075e-06, 'learning_rate/action_model': 6.642324710810076e-05, 'epoch': 1.55}) +[2026-06-16 00:41:19,757][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.1539042443037033, 'train/grad_norm_pre_clip': 4.366029739379883, 'timing/data': 0.021220160648226738, 'timing/model': 1.440473617054522, 'learning_rate/qwen_vl_interface': 6.639412772450047e-06, 'learning_rate/action_model': 6.639412772450047e-05, 'epoch': 1.55}) +[2026-06-16 00:41:25,554][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.10867439955472946, 'train/grad_norm_pre_clip': 3.3574631214141846, 'timing/data': 0.021047950722277164, 'timing/model': 1.4382465155795217, 'learning_rate/qwen_vl_interface': 6.6365002629549104e-06, 'learning_rate/action_model': 6.63650026295491e-05, 'epoch': 1.55}) +[2026-06-16 00:41:31,411][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.03846435248851776, 'train/grad_norm_pre_clip': 2.314354419708252, 'timing/data': 0.021548823453485966, 'timing/model': 1.6897079162299633, 'learning_rate/qwen_vl_interface': 6.633587183521887e-06, 'learning_rate/action_model': 6.633587183521887e-05, 'epoch': 1.55}) +[2026-06-16 00:41:37,082][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.13789322972297668, 'train/grad_norm_pre_clip': 3.3971259593963623, 'timing/data': 0.02145744301378727, 'timing/model': 1.4612521370872855, 'learning_rate/qwen_vl_interface': 6.6306735353484365e-06, 'learning_rate/action_model': 6.630673535348436e-05, 'epoch': 1.55}) +[2026-06-16 00:41:42,958][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.1307118982076645, 'train/grad_norm_pre_clip': 4.3147358894348145, 'timing/data': 0.021920815110206604, 'timing/model': 1.441827455535531, 'learning_rate/qwen_vl_interface': 6.627759319632246e-06, 'learning_rate/action_model': 6.627759319632247e-05, 'epoch': 1.55}) +[2026-06-16 00:41:48,760][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.12983010709285736, 'train/grad_norm_pre_clip': 3.1526107788085938, 'timing/data': 0.022432537749409676, 'timing/model': 1.4461095593869686, 'learning_rate/qwen_vl_interface': 6.624844537571245e-06, 'learning_rate/action_model': 6.624844537571245e-05, 'epoch': 1.55}) +[2026-06-16 00:41:54,565][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.10969937592744827, 'train/grad_norm_pre_clip': 4.505197048187256, 'timing/data': 0.02077442966401577, 'timing/model': 1.4471112852916121, 'learning_rate/qwen_vl_interface': 6.621929190363585e-06, 'learning_rate/action_model': 6.621929190363585e-05, 'epoch': 1.55}) +[2026-06-16 00:42:00,367][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.10960893332958221, 'train/grad_norm_pre_clip': 3.7271296977996826, 'timing/data': 0.021239401772618294, 'timing/model': 1.4455682169646025, 'learning_rate/qwen_vl_interface': 6.619013279207657e-06, 'learning_rate/action_model': 6.619013279207656e-05, 'epoch': 1.55}) +[2026-06-16 00:42:06,226][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.07911475747823715, 'train/grad_norm_pre_clip': 2.8794949054718018, 'timing/data': 0.02113255113363266, 'timing/model': 1.7193443337455392, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 1.55}) +[2026-06-16 00:42:11,868][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.11519435793161392, 'train/grad_norm_pre_clip': 5.936391830444336, 'timing/data': 0.021039940416812897, 'timing/model': 1.4647275786846876, 'learning_rate/qwen_vl_interface': 6.6131797698457135e-06, 'learning_rate/action_model': 6.613179769845713e-05, 'epoch': 1.55}) +[2026-06-16 00:42:17,666][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.28498560190200806, 'train/grad_norm_pre_clip': 7.614551544189453, 'timing/data': 0.02497990056872368, 'timing/model': 1.4552033189684153, 'learning_rate/qwen_vl_interface': 6.610262174037633e-06, 'learning_rate/action_model': 6.610262174037633e-05, 'epoch': 1.55}) +[2026-06-16 00:42:23,467][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.25608083605766296, 'train/grad_norm_pre_clip': 10.54780101776123, 'timing/data': 0.020879080519080162, 'timing/model': 1.4594445414841175, 'learning_rate/qwen_vl_interface': 6.607344019077155e-06, 'learning_rate/action_model': 6.607344019077156e-05, 'epoch': 1.55}) +[2026-06-16 00:42:29,274][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.08515821397304535, 'train/grad_norm_pre_clip': 3.5772218704223633, 'timing/data': 0.022323627956211567, 'timing/model': 1.4119926122948527, 'learning_rate/qwen_vl_interface': 6.604425306163822e-06, 'learning_rate/action_model': 6.604425306163822e-05, 'epoch': 1.55}) +[2026-06-16 00:53:07,779][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.0781155601143837, 'train/grad_norm_pre_clip': 4.319269180297852, 'eval/loss': 0.06774315983057022, 'eval/flap_precision': 0.8127490039840638, 'eval/flap_recall': 0.7418181818181818, 'eval/flap_f1': 0.7756653992395437, 'eval/flap_rate_model': 0.06972222222222223, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7756653992395437, 'timing/data': 0.022375818341970444, 'timing/model': 1.413453389890492, 'learning_rate/qwen_vl_interface': 6.6015060364974045e-06, 'learning_rate/action_model': 6.601506036497404e-05, 'epoch': 1.55}) +[2026-06-16 00:53:07,781][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3400_state +[2026-06-16 00:53:07,787][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 00:53:15,655][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3400_state/model.safetensors +[2026-06-16 00:53:26,713][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3400_state/optimizer.bin +[2026-06-16 00:53:26,714][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3400_state/sampler.bin +[2026-06-16 00:53:26,714][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3400_state/sampler_1.bin +[2026-06-16 00:53:26,715][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3400_state/random_states_0.pkl +[2026-06-16 00:53:26,867][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3400 +[2026-06-16 00:53:26,867][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 00:53:26,883][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 00:53:32,817][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.20406319200992584, 'train/grad_norm_pre_clip': 8.061774253845215, 'timing/data': 0.022684074006974697, 'timing/model': 1.3613392384722829, 'learning_rate/qwen_vl_interface': 6.598586211277907e-06, 'learning_rate/action_model': 6.598586211277907e-05, 'epoch': 1.56}) +[2026-06-16 00:53:38,529][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.0627802163362503, 'train/grad_norm_pre_clip': 3.043508291244507, 'timing/data': 0.020366521552205086, 'timing/model': 1.7277090614661574, 'learning_rate/qwen_vl_interface': 6.59566583170556e-06, 'learning_rate/action_model': 6.59566583170556e-05, 'epoch': 1.56}) +[2026-06-16 00:53:44,204][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.0427466481924057, 'train/grad_norm_pre_clip': 4.838037490844727, 'timing/data': 0.02085786499083042, 'timing/model': 1.5942641338333488, 'learning_rate/qwen_vl_interface': 6.59274489898082e-06, 'learning_rate/action_model': 6.592744898980819e-05, 'epoch': 1.56}) +[2026-06-16 00:53:49,843][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.2258317917585373, 'train/grad_norm_pre_clip': 8.377985000610352, 'timing/data': 0.0287115341052413, 'timing/model': 1.435101599432528, 'learning_rate/qwen_vl_interface': 6.5898234143043716e-06, 'learning_rate/action_model': 6.589823414304371e-05, 'epoch': 1.56}) +[2026-06-16 00:53:55,555][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.17381595075130463, 'train/grad_norm_pre_clip': 10.684666633605957, 'timing/data': 0.02290729433298111, 'timing/model': 1.4684675298631191, 'learning_rate/qwen_vl_interface': 6.5869013788771264e-06, 'learning_rate/action_model': 6.586901378877126e-05, 'epoch': 1.56}) +[2026-06-16 00:54:01,159][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.1577291339635849, 'train/grad_norm_pre_clip': 7.0167155265808105, 'timing/data': 0.020737115293741226, 'timing/model': 1.3801086377352476, 'learning_rate/qwen_vl_interface': 6.583978793900223e-06, 'learning_rate/action_model': 6.583978793900223e-05, 'epoch': 1.56}) +[2026-06-16 00:54:06,834][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.08582096546888351, 'train/grad_norm_pre_clip': 4.732856273651123, 'timing/data': 0.020946425385773182, 'timing/model': 1.3979734210297465, 'learning_rate/qwen_vl_interface': 6.581055660575028e-06, 'learning_rate/action_model': 6.581055660575027e-05, 'epoch': 1.56}) +[2026-06-16 00:54:12,574][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.23174184560775757, 'train/grad_norm_pre_clip': 13.023819923400879, 'timing/data': 0.02230515144765377, 'timing/model': 1.6851714588701725, 'learning_rate/qwen_vl_interface': 6.578131980103128e-06, 'learning_rate/action_model': 6.578131980103128e-05, 'epoch': 1.56}) +[2026-06-16 00:54:18,138][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.2208862602710724, 'train/grad_norm_pre_clip': 9.737740516662598, 'timing/data': 0.020566103979945183, 'timing/model': 1.4289473332464695, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 1.56}) +[2026-06-16 00:54:23,903][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.06678029149770737, 'train/grad_norm_pre_clip': 7.633822917938232, 'timing/data': 0.020347892306745052, 'timing/model': 1.421542938798666, 'learning_rate/qwen_vl_interface': 6.572282982526702e-06, 'learning_rate/action_model': 6.572282982526702e-05, 'epoch': 1.56}) +[2026-06-16 00:54:29,642][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.07970824092626572, 'train/grad_norm_pre_clip': 4.399569034576416, 'timing/data': 0.021747849881649017, 'timing/model': 1.4180426634848118, 'learning_rate/qwen_vl_interface': 6.5693576678264755e-06, 'learning_rate/action_model': 6.569357667826476e-05, 'epoch': 1.56}) +[2026-06-16 00:54:35,336][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.10558002442121506, 'train/grad_norm_pre_clip': 5.846959590911865, 'timing/data': 0.020922104828059673, 'timing/model': 1.4418290071189404, 'learning_rate/qwen_vl_interface': 6.566431810788148e-06, 'learning_rate/action_model': 6.566431810788148e-05, 'epoch': 1.56}) +[2026-06-16 00:54:41,044][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.17727802693843842, 'train/grad_norm_pre_clip': 11.291833877563477, 'timing/data': 0.021792720071971416, 'timing/model': 1.4052110947668552, 'learning_rate/qwen_vl_interface': 6.5635054126144306e-06, 'learning_rate/action_model': 6.56350541261443e-05, 'epoch': 1.56}) +[2026-06-16 00:54:46,881][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.09958836436271667, 'train/grad_norm_pre_clip': 4.995978832244873, 'timing/data': 0.021581447683274746, 'timing/model': 1.6877544727176428, 'learning_rate/qwen_vl_interface': 6.560578474508254e-06, 'learning_rate/action_model': 6.560578474508253e-05, 'epoch': 1.56}) +[2026-06-16 00:54:52,456][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.06286375224590302, 'train/grad_norm_pre_clip': 2.8498427867889404, 'timing/data': 0.021068725734949112, 'timing/model': 1.4468355905264616, 'learning_rate/qwen_vl_interface': 6.5576509976727685e-06, 'learning_rate/action_model': 6.557650997672769e-05, 'epoch': 1.56}) +[2026-06-16 00:54:58,243][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.06096450984477997, 'train/grad_norm_pre_clip': 4.480052471160889, 'timing/data': 0.02037506364285946, 'timing/model': 1.4438846288248897, 'learning_rate/qwen_vl_interface': 6.554722983311354e-06, 'learning_rate/action_model': 6.554722983311355e-05, 'epoch': 1.56}) +[2026-06-16 00:55:03,957][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.10051464289426804, 'train/grad_norm_pre_clip': 20.301729202270508, 'timing/data': 0.020460933446884155, 'timing/model': 1.4548691976815462, 'learning_rate/qwen_vl_interface': 6.551794432627605e-06, 'learning_rate/action_model': 6.551794432627605e-05, 'epoch': 1.56}) +[2026-06-16 00:55:09,751][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.13246260583400726, 'train/grad_norm_pre_clip': 7.917248249053955, 'timing/data': 0.023200107738375664, 'timing/model': 1.450374766252935, 'learning_rate/qwen_vl_interface': 6.548865346825338e-06, 'learning_rate/action_model': 6.548865346825338e-05, 'epoch': 1.56}) +[2026-06-16 00:55:15,442][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.1635415405035019, 'train/grad_norm_pre_clip': 29.902742385864258, 'timing/data': 0.021665588952600956, 'timing/model': 1.3891284493729472, 'learning_rate/qwen_vl_interface': 6.545935727108589e-06, 'learning_rate/action_model': 6.545935727108588e-05, 'epoch': 1.56}) +[2026-06-16 00:55:21,171][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.10959944128990173, 'train/grad_norm_pre_clip': 5.403888702392578, 'timing/data': 0.020279022864997387, 'timing/model': 1.66628304310143, 'learning_rate/qwen_vl_interface': 6.543005574681613e-06, 'learning_rate/action_model': 6.543005574681613e-05, 'epoch': 1.56}) +[2026-06-16 00:55:26,918][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.11161118000745773, 'train/grad_norm_pre_clip': 11.030679702758789, 'timing/data': 0.020636883564293385, 'timing/model': 1.6180664421990514, 'learning_rate/qwen_vl_interface': 6.5400748907488886e-06, 'learning_rate/action_model': 6.540074890748888e-05, 'epoch': 1.56}) +[2026-06-16 00:55:32,445][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.29208695888519287, 'train/grad_norm_pre_clip': 23.676279067993164, 'timing/data': 0.02175930980592966, 'timing/model': 1.4492989014834166, 'learning_rate/qwen_vl_interface': 6.537143676515105e-06, 'learning_rate/action_model': 6.537143676515105e-05, 'epoch': 1.57}) +[2026-06-16 00:55:38,253][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.48369476199150085, 'train/grad_norm_pre_clip': 31.353391647338867, 'timing/data': 0.02575830090790987, 'timing/model': 1.4574492750689387, 'learning_rate/qwen_vl_interface': 6.534211933185176e-06, 'learning_rate/action_model': 6.534211933185176e-05, 'epoch': 1.57}) +[2026-06-16 00:55:43,946][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.08564101904630661, 'train/grad_norm_pre_clip': 6.433917045593262, 'timing/data': 0.021391067653894424, 'timing/model': 1.4527048040181398, 'learning_rate/qwen_vl_interface': 6.531279661964231e-06, 'learning_rate/action_model': 6.531279661964231e-05, 'epoch': 1.57}) +[2026-06-16 00:55:49,668][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.040284715592861176, 'train/grad_norm_pre_clip': 6.31550407409668, 'timing/data': 0.021511287428438663, 'timing/model': 1.3937786985188723, 'learning_rate/qwen_vl_interface': 6.528346864057616e-06, 'learning_rate/action_model': 6.528346864057615e-05, 'epoch': 1.57}) +[2026-06-16 00:55:55,482][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.10385739058256149, 'train/grad_norm_pre_clip': 6.178031921386719, 'timing/data': 0.0230878172442317, 'timing/model': 1.6751711461693048, 'learning_rate/qwen_vl_interface': 6.5254135406708906e-06, 'learning_rate/action_model': 6.52541354067089e-05, 'epoch': 1.57}) +[2026-06-16 00:56:01,289][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.18306241929531097, 'train/grad_norm_pre_clip': 10.094893455505371, 'timing/data': 0.02095783594995737, 'timing/model': 1.6932361209765077, 'learning_rate/qwen_vl_interface': 6.522479693009836e-06, 'learning_rate/action_model': 6.522479693009837e-05, 'epoch': 1.57}) +[2026-06-16 00:56:06,846][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.19671164453029633, 'train/grad_norm_pre_clip': 6.174123287200928, 'timing/data': 0.02078840509057045, 'timing/model': 1.4493076568469405, 'learning_rate/qwen_vl_interface': 6.519545322280446e-06, 'learning_rate/action_model': 6.519545322280446e-05, 'epoch': 1.57}) +[2026-06-16 00:56:12,562][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.09538061916828156, 'train/grad_norm_pre_clip': 5.714168071746826, 'timing/data': 0.020529164001345634, 'timing/model': 1.4198607802391052, 'learning_rate/qwen_vl_interface': 6.5166104296889275e-06, 'learning_rate/action_model': 6.516610429688927e-05, 'epoch': 1.57}) +[2026-06-16 00:56:18,349][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.03638920187950134, 'train/grad_norm_pre_clip': 2.1329500675201416, 'timing/data': 0.021402508020401, 'timing/model': 1.4227492967620492, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 1.57}) +[2026-06-16 00:56:24,056][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.23661863803863525, 'train/grad_norm_pre_clip': 11.944056510925293, 'timing/data': 0.021381358616054058, 'timing/model': 1.4053260497748852, 'learning_rate/qwen_vl_interface': 6.510739083745422e-06, 'learning_rate/action_model': 6.510739083745422e-05, 'epoch': 1.57}) +[2026-06-16 00:56:29,883][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.07681180536746979, 'train/grad_norm_pre_clip': 3.644793748855591, 'timing/data': 0.021334178745746613, 'timing/model': 1.6815653843805194, 'learning_rate/qwen_vl_interface': 6.507802632806919e-06, 'learning_rate/action_model': 6.507802632806919e-05, 'epoch': 1.57}) +[2026-06-16 00:56:35,453][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.09113046526908875, 'train/grad_norm_pre_clip': 4.437431812286377, 'timing/data': 0.021586939692497253, 'timing/model': 1.4563530245795846, 'learning_rate/qwen_vl_interface': 6.5048656648332646e-06, 'learning_rate/action_model': 6.504865664833264e-05, 'epoch': 1.57}) +[2026-06-16 00:56:41,256][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.07253029942512512, 'train/grad_norm_pre_clip': 5.184930801391602, 'timing/data': 0.021204927936196327, 'timing/model': 1.4488664688542485, 'learning_rate/qwen_vl_interface': 6.501928181031737e-06, 'learning_rate/action_model': 6.501928181031737e-05, 'epoch': 1.57}) +[2026-06-16 00:56:47,059][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.18932083249092102, 'train/grad_norm_pre_clip': 11.547738075256348, 'timing/data': 0.021086406894028187, 'timing/model': 1.4490062315016985, 'learning_rate/qwen_vl_interface': 6.498990182609823e-06, 'learning_rate/action_model': 6.498990182609823e-05, 'epoch': 1.57}) +[2026-06-16 00:56:52,864][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.0665929764509201, 'train/grad_norm_pre_clip': 4.653853893280029, 'timing/data': 0.021473138593137264, 'timing/model': 1.423452333547175, 'learning_rate/qwen_vl_interface': 6.4960516707752255e-06, 'learning_rate/action_model': 6.496051670775226e-05, 'epoch': 1.57}) +[2026-06-16 00:56:58,576][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.21421821415424347, 'train/grad_norm_pre_clip': 17.64714241027832, 'timing/data': 0.020774425007402897, 'timing/model': 1.4025267204269767, 'learning_rate/qwen_vl_interface': 6.493112646735851e-06, 'learning_rate/action_model': 6.493112646735851e-05, 'epoch': 1.57}) +[2026-06-16 00:57:04,356][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.06483035534620285, 'train/grad_norm_pre_clip': 3.0488407611846924, 'timing/data': 0.021266558207571507, 'timing/model': 1.6424998715519905, 'learning_rate/qwen_vl_interface': 6.490173111699824e-06, 'learning_rate/action_model': 6.490173111699825e-05, 'epoch': 1.57}) +[2026-06-16 00:57:09,950][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.17353008687496185, 'train/grad_norm_pre_clip': 11.764540672302246, 'timing/data': 0.021974392235279083, 'timing/model': 1.447775423526764, 'learning_rate/qwen_vl_interface': 6.487233066875479e-06, 'learning_rate/action_model': 6.487233066875479e-05, 'epoch': 1.57}) +[2026-06-16 00:57:15,750][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.13319164514541626, 'train/grad_norm_pre_clip': 5.10598611831665, 'timing/data': 0.021835741586983204, 'timing/model': 1.440094007179141, 'learning_rate/qwen_vl_interface': 6.484292513471352e-06, 'learning_rate/action_model': 6.484292513471352e-05, 'epoch': 1.57}) +[2026-06-16 00:57:21,552][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.23782843351364136, 'train/grad_norm_pre_clip': 15.194164276123047, 'timing/data': 0.020836985670030117, 'timing/model': 1.4523808509111404, 'learning_rate/qwen_vl_interface': 6.4813514526961975e-06, 'learning_rate/action_model': 6.481351452696198e-05, 'epoch': 1.57}) +[2026-06-16 00:57:27,348][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.07838408648967743, 'train/grad_norm_pre_clip': 5.257444381713867, 'timing/data': 0.02233808394521475, 'timing/model': 1.411259263753891, 'learning_rate/qwen_vl_interface': 6.478409885758972e-06, 'learning_rate/action_model': 6.478409885758971e-05, 'epoch': 1.57}) +[2026-06-16 00:57:33,151][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.02346605435013771, 'train/grad_norm_pre_clip': 2.5517334938049316, 'timing/data': 0.021596970036625862, 'timing/model': 1.4035301376134157, 'learning_rate/qwen_vl_interface': 6.475467813868844e-06, 'learning_rate/action_model': 6.475467813868844e-05, 'epoch': 1.57}) +[2026-06-16 00:57:38,953][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.13025817275047302, 'train/grad_norm_pre_clip': 10.030423164367676, 'timing/data': 0.02080276608467102, 'timing/model': 1.6411152193322778, 'learning_rate/qwen_vl_interface': 6.472525238235187e-06, 'learning_rate/action_model': 6.472525238235186e-05, 'epoch': 1.58}) +[2026-06-16 00:57:44,556][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.05256416276097298, 'train/grad_norm_pre_clip': 3.3870816230773926, 'timing/data': 0.021610931493341923, 'timing/model': 1.4403567584231496, 'learning_rate/qwen_vl_interface': 6.4695821600675834e-06, 'learning_rate/action_model': 6.469582160067583e-05, 'epoch': 1.58}) +[2026-06-16 00:57:50,356][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.23320353031158447, 'train/grad_norm_pre_clip': 17.73503303527832, 'timing/data': 0.021580220200121403, 'timing/model': 1.4425648730248213, 'learning_rate/qwen_vl_interface': 6.466638580575824e-06, 'learning_rate/action_model': 6.466638580575824e-05, 'epoch': 1.58}) +[2026-06-16 00:57:56,151][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.11745239794254303, 'train/grad_norm_pre_clip': 6.595028400421143, 'timing/data': 0.020902137272059917, 'timing/model': 1.44840376265347, 'learning_rate/qwen_vl_interface': 6.463694500969898e-06, 'learning_rate/action_model': 6.463694500969898e-05, 'epoch': 1.58}) +[2026-06-16 00:58:01,951][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.2808631360530853, 'train/grad_norm_pre_clip': 15.41687297821045, 'timing/data': 0.02064305543899536, 'timing/model': 1.415108916349709, 'learning_rate/qwen_vl_interface': 6.460749922460008e-06, 'learning_rate/action_model': 6.460749922460008e-05, 'epoch': 1.58}) +[2026-06-16 00:58:07,783][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.048113077878952026, 'train/grad_norm_pre_clip': 3.952579975128174, 'timing/data': 0.020755034871399403, 'timing/model': 1.678521272726357, 'learning_rate/qwen_vl_interface': 6.457804846256559e-06, 'learning_rate/action_model': 6.457804846256559e-05, 'epoch': 1.58}) +[2026-06-16 00:58:13,459][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.11051405221223831, 'train/grad_norm_pre_clip': 9.954310417175293, 'timing/data': 0.021371318958699703, 'timing/model': 1.4539419682696462, 'learning_rate/qwen_vl_interface': 6.45485927357016e-06, 'learning_rate/action_model': 6.454859273570159e-05, 'epoch': 1.58}) +[2026-06-16 00:58:19,258][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.03378722071647644, 'train/grad_norm_pre_clip': 1.8429239988327026, 'timing/data': 0.021158987656235695, 'timing/model': 1.4561368701979518, 'learning_rate/qwen_vl_interface': 6.451913205611626e-06, 'learning_rate/action_model': 6.451913205611626e-05, 'epoch': 1.58}) +[2026-06-16 00:58:24,957][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.20923316478729248, 'train/grad_norm_pre_clip': 7.601474285125732, 'timing/data': 0.022473745048046112, 'timing/model': 1.4552807593718171, 'learning_rate/qwen_vl_interface': 6.448966643591974e-06, 'learning_rate/action_model': 6.448966643591974e-05, 'epoch': 1.58}) +[2026-06-16 00:58:30,672][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.3163035809993744, 'train/grad_norm_pre_clip': 12.399430274963379, 'timing/data': 0.02201912272721529, 'timing/model': 1.4389875466004014, 'learning_rate/qwen_vl_interface': 6.446019588722423e-06, 'learning_rate/action_model': 6.446019588722422e-05, 'epoch': 1.58}) +[2026-06-16 00:58:36,451][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.18080177903175354, 'train/grad_norm_pre_clip': 7.611721992492676, 'timing/data': 0.02096097730100155, 'timing/model': 1.446031104773283, 'learning_rate/qwen_vl_interface': 6.443072042214398e-06, 'learning_rate/action_model': 6.443072042214397e-05, 'epoch': 1.58}) +[2026-06-16 00:58:42,299][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.19596123695373535, 'train/grad_norm_pre_clip': 10.53592300415039, 'timing/data': 0.021341469138860703, 'timing/model': 1.687820772640407, 'learning_rate/qwen_vl_interface': 6.440124005279521e-06, 'learning_rate/action_model': 6.440124005279521e-05, 'epoch': 1.58}) +[2026-06-16 00:58:47,953][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.08161056786775589, 'train/grad_norm_pre_clip': 4.452081203460693, 'timing/data': 0.02039160393178463, 'timing/model': 1.4555123476311564, 'learning_rate/qwen_vl_interface': 6.43717547912962e-06, 'learning_rate/action_model': 6.437175479129621e-05, 'epoch': 1.58}) +[2026-06-16 00:58:53,753][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.08657148480415344, 'train/grad_norm_pre_clip': 6.39484167098999, 'timing/data': 0.021661370992660522, 'timing/model': 1.454494963400066, 'learning_rate/qwen_vl_interface': 6.434226464976725e-06, 'learning_rate/action_model': 6.434226464976724e-05, 'epoch': 1.58}) +[2026-06-16 00:58:59,551][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.07608601450920105, 'train/grad_norm_pre_clip': 3.9610235691070557, 'timing/data': 0.02182434033602476, 'timing/model': 1.4408507654443383, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 1.58}) +[2026-06-16 00:59:05,460][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.2984575927257538, 'train/grad_norm_pre_clip': 9.940040588378906, 'timing/data': 0.021831391379237175, 'timing/model': 1.4664399279281497, 'learning_rate/qwen_vl_interface': 6.428326977511059e-06, 'learning_rate/action_model': 6.428326977511058e-05, 'epoch': 1.58}) +[2026-06-16 00:59:11,254][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.14443892240524292, 'train/grad_norm_pre_clip': 5.684979438781738, 'timing/data': 0.020614005625247955, 'timing/model': 1.4370911102741957, 'learning_rate/qwen_vl_interface': 6.425376506623344e-06, 'learning_rate/action_model': 6.425376506623344e-05, 'epoch': 1.58}) +[2026-06-16 00:59:16,985][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.0471867136657238, 'train/grad_norm_pre_clip': 2.636629104614258, 'timing/data': 0.021747500635683537, 'timing/model': 1.666671235114336, 'learning_rate/qwen_vl_interface': 6.422425552582747e-06, 'learning_rate/action_model': 6.422425552582747e-05, 'epoch': 1.58}) +[2026-06-16 00:59:22,553][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.09250207245349884, 'train/grad_norm_pre_clip': 5.7904863357543945, 'timing/data': 0.021538369357585907, 'timing/model': 1.434162118472159, 'learning_rate/qwen_vl_interface': 6.41947411660229e-06, 'learning_rate/action_model': 6.41947411660229e-05, 'epoch': 1.58}) +[2026-06-16 00:59:28,775][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.1236409917473793, 'train/grad_norm_pre_clip': 8.495424270629883, 'timing/data': 0.028300365433096886, 'timing/model': 1.4475983390584588, 'learning_rate/qwen_vl_interface': 6.4165221998951995e-06, 'learning_rate/action_model': 6.4165221998952e-05, 'epoch': 1.58}) +[2026-06-16 00:59:34,984][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.14158131182193756, 'train/grad_norm_pre_clip': 6.588776588439941, 'timing/data': 0.028020083904266357, 'timing/model': 1.450981487520039, 'learning_rate/qwen_vl_interface': 6.413569803674896e-06, 'learning_rate/action_model': 6.413569803674896e-05, 'epoch': 1.58}) +[2026-06-16 00:59:41,276][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.25501561164855957, 'train/grad_norm_pre_clip': 14.838689804077148, 'timing/data': 0.02877777721732855, 'timing/model': 1.4579713148996234, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 1.58}) +[2026-06-16 00:59:47,380][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.17117442190647125, 'train/grad_norm_pre_clip': 6.927917957305908, 'timing/data': 0.028296595439314842, 'timing/model': 1.4560090163722634, 'learning_rate/qwen_vl_interface': 6.40766357754932e-06, 'learning_rate/action_model': 6.407663577549319e-05, 'epoch': 1.59}) +[2026-06-16 00:59:53,742][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.19879572093486786, 'train/grad_norm_pre_clip': 14.768084526062012, 'timing/data': 0.021239127963781357, 'timing/model': 2.1240381309762597, 'learning_rate/qwen_vl_interface': 6.404709750071875e-06, 'learning_rate/action_model': 6.404709750071874e-05, 'epoch': 1.59}) +[2026-06-16 00:59:59,372][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.1783449649810791, 'train/grad_norm_pre_clip': 9.698981285095215, 'timing/data': 0.02548713143914938, 'timing/model': 1.4386748801916838, 'learning_rate/qwen_vl_interface': 6.401755447936869e-06, 'learning_rate/action_model': 6.401755447936869e-05, 'epoch': 1.59}) +[2026-06-16 01:00:05,658][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.14435680210590363, 'train/grad_norm_pre_clip': 10.06777572631836, 'timing/data': 0.026692116633057594, 'timing/model': 1.4187507908791304, 'learning_rate/qwen_vl_interface': 6.3988006723587036e-06, 'learning_rate/action_model': 6.398800672358703e-05, 'epoch': 1.59}) +[2026-06-16 01:00:11,892][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.09887197613716125, 'train/grad_norm_pre_clip': 3.998197555541992, 'timing/data': 0.029279950074851513, 'timing/model': 1.4559224331751466, 'learning_rate/qwen_vl_interface': 6.395845424551975e-06, 'learning_rate/action_model': 6.395845424551975e-05, 'epoch': 1.59}) +[2026-06-16 01:00:18,719][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.09049657732248306, 'train/grad_norm_pre_clip': 5.374650001525879, 'timing/data': 0.02820303477346897, 'timing/model': 1.4439181424677372, 'learning_rate/qwen_vl_interface': 6.392889705731475e-06, 'learning_rate/action_model': 6.392889705731476e-05, 'epoch': 1.59}) +[2026-06-16 01:00:25,179][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.07595346868038177, 'train/grad_norm_pre_clip': 5.650755882263184, 'timing/data': 0.024050043895840645, 'timing/model': 1.493315678089857, 'learning_rate/qwen_vl_interface': 6.389933517112188e-06, 'learning_rate/action_model': 6.389933517112187e-05, 'epoch': 1.59}) +[2026-06-16 01:00:31,406][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.08515501767396927, 'train/grad_norm_pre_clip': 5.439021587371826, 'timing/data': 0.02488588634878397, 'timing/model': 2.0787362530827522, 'learning_rate/qwen_vl_interface': 6.386976859909293e-06, 'learning_rate/action_model': 6.386976859909292e-05, 'epoch': 1.59}) +[2026-06-16 01:00:37,082][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.2470400333404541, 'train/grad_norm_pre_clip': 6.933478355407715, 'timing/data': 0.023776521906256676, 'timing/model': 1.4572815271094441, 'learning_rate/qwen_vl_interface': 6.384019735338156e-06, 'learning_rate/action_model': 6.384019735338156e-05, 'epoch': 1.59}) +[2026-06-16 01:00:43,382][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.15373273193836212, 'train/grad_norm_pre_clip': 6.883184909820557, 'timing/data': 0.023760012350976467, 'timing/model': 1.4532089373096824, 'learning_rate/qwen_vl_interface': 6.381062144614342e-06, 'learning_rate/action_model': 6.381062144614342e-05, 'epoch': 1.59}) +[2026-06-16 01:00:49,681][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.12073567509651184, 'train/grad_norm_pre_clip': 7.8399786949157715, 'timing/data': 0.023848092183470726, 'timing/model': 1.4536492908373475, 'learning_rate/qwen_vl_interface': 6.378104088953605e-06, 'learning_rate/action_model': 6.378104088953605e-05, 'epoch': 1.59}) +[2026-06-16 01:00:55,978][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.3105855882167816, 'train/grad_norm_pre_clip': 12.706602096557617, 'timing/data': 0.023499561473727226, 'timing/model': 1.4489534879103303, 'learning_rate/qwen_vl_interface': 6.375145569571889e-06, 'learning_rate/action_model': 6.375145569571889e-05, 'epoch': 1.59}) +[2026-06-16 01:01:02,183][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.23471106588840485, 'train/grad_norm_pre_clip': 9.248025894165039, 'timing/data': 0.023412120528519154, 'timing/model': 1.4380612336099148, 'learning_rate/qwen_vl_interface': 6.372186587685329e-06, 'learning_rate/action_model': 6.372186587685329e-05, 'epoch': 1.59}) +[2026-06-16 01:01:08,542][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.15008974075317383, 'train/grad_norm_pre_clip': 7.666642189025879, 'timing/data': 0.02417571470141411, 'timing/model': 2.096240614540875, 'learning_rate/qwen_vl_interface': 6.369227144510252e-06, 'learning_rate/action_model': 6.369227144510252e-05, 'epoch': 1.59}) +[2026-06-16 01:01:14,176][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.21126526594161987, 'train/grad_norm_pre_clip': 6.593479156494141, 'timing/data': 0.022105754353106022, 'timing/model': 1.4587244633585215, 'learning_rate/qwen_vl_interface': 6.3662672412631736e-06, 'learning_rate/action_model': 6.366267241263173e-05, 'epoch': 1.59}) +[2026-06-16 01:01:20,375][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.2325967401266098, 'train/grad_norm_pre_clip': 13.381580352783203, 'timing/data': 0.024368595331907272, 'timing/model': 1.4353085039183497, 'learning_rate/qwen_vl_interface': 6.3633068791607954e-06, 'learning_rate/action_model': 6.363306879160795e-05, 'epoch': 1.59}) +[2026-06-16 01:01:26,272][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.11127655953168869, 'train/grad_norm_pre_clip': 3.8512396812438965, 'timing/data': 0.025271990336477757, 'timing/model': 1.4421521313488483, 'learning_rate/qwen_vl_interface': 6.360346059420014e-06, 'learning_rate/action_model': 6.360346059420014e-05, 'epoch': 1.59}) +[2026-06-16 01:01:32,373][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.1217183843255043, 'train/grad_norm_pre_clip': 5.37490177154541, 'timing/data': 0.028846227563917637, 'timing/model': 1.4620421444997191, 'learning_rate/qwen_vl_interface': 6.357384783257909e-06, 'learning_rate/action_model': 6.357384783257908e-05, 'epoch': 1.59}) +[2026-06-16 01:01:38,502][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.1153719574213028, 'train/grad_norm_pre_clip': 4.465301990509033, 'timing/data': 0.024731026962399483, 'timing/model': 1.4294579271227121, 'learning_rate/qwen_vl_interface': 6.354423051891748e-06, 'learning_rate/action_model': 6.354423051891749e-05, 'epoch': 1.59}) +[2026-06-16 01:01:44,729][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.06636109203100204, 'train/grad_norm_pre_clip': 4.850433826446533, 'timing/data': 0.02474730834364891, 'timing/model': 1.996705756522715, 'learning_rate/qwen_vl_interface': 6.35146086653899e-06, 'learning_rate/action_model': 6.35146086653899e-05, 'epoch': 1.59}) +[2026-06-16 01:01:50,375][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.19186079502105713, 'train/grad_norm_pre_clip': 6.1047139167785645, 'timing/data': 0.022983718663454056, 'timing/model': 1.4453610414639115, 'learning_rate/qwen_vl_interface': 6.348498228417273e-06, 'learning_rate/action_model': 6.348498228417273e-05, 'epoch': 1.59}) +[2026-06-16 01:01:57,374][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.014172823168337345, 'train/grad_norm_pre_clip': 1.360097885131836, 'timing/data': 1.8419075030833483, 'timing/model': 1.4787747953087091, 'learning_rate/qwen_vl_interface': 6.34553513874443e-06, 'learning_rate/action_model': 6.34553513874443e-05, 'epoch': 1.59}) +[2026-06-16 01:02:03,273][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.04088442772626877, 'train/grad_norm_pre_clip': 2.271929979324341, 'timing/data': 0.014516505412757397, 'timing/model': 1.4611633159220219, 'learning_rate/qwen_vl_interface': 6.342571598738473e-06, 'learning_rate/action_model': 6.342571598738472e-05, 'epoch': 1.6}) +[2026-06-16 01:02:09,205][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.253807008266449, 'train/grad_norm_pre_clip': 7.304357528686523, 'timing/data': 0.0140815619379282, 'timing/model': 1.4212135914713144, 'learning_rate/qwen_vl_interface': 6.339607609617603e-06, 'learning_rate/action_model': 6.339607609617603e-05, 'epoch': 1.6}) +[2026-06-16 01:02:15,146][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.15102900564670563, 'train/grad_norm_pre_clip': 5.165475845336914, 'timing/data': 0.026057283394038677, 'timing/model': 1.7381292944774032, 'learning_rate/qwen_vl_interface': 6.336643172600202e-06, 'learning_rate/action_model': 6.336643172600202e-05, 'epoch': 1.6}) +[2026-06-16 01:02:20,852][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.1017981693148613, 'train/grad_norm_pre_clip': 4.977771759033203, 'timing/data': 0.022485285997390747, 'timing/model': 1.4515679897740483, 'learning_rate/qwen_vl_interface': 6.333678288904838e-06, 'learning_rate/action_model': 6.333678288904837e-05, 'epoch': 1.6}) +[2026-06-16 01:02:26,659][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.3520650863647461, 'train/grad_norm_pre_clip': 10.67408275604248, 'timing/data': 0.02106337808072567, 'timing/model': 1.452484406530857, 'learning_rate/qwen_vl_interface': 6.330712959750264e-06, 'learning_rate/action_model': 6.330712959750263e-05, 'epoch': 1.6}) +[2026-06-16 01:02:32,451][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.2069956511259079, 'train/grad_norm_pre_clip': 7.080248832702637, 'timing/data': 0.021191389299929142, 'timing/model': 1.4584219371899962, 'learning_rate/qwen_vl_interface': 6.3277471863554145e-06, 'learning_rate/action_model': 6.327747186355414e-05, 'epoch': 1.6}) +[2026-06-16 01:02:38,255][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.11885801702737808, 'train/grad_norm_pre_clip': 3.807558536529541, 'timing/data': 0.026849637739360332, 'timing/model': 1.444097955711186, 'learning_rate/qwen_vl_interface': 6.324780969939406e-06, 'learning_rate/action_model': 6.324780969939405e-05, 'epoch': 1.6}) +[2026-06-16 01:02:44,063][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.24214932322502136, 'train/grad_norm_pre_clip': 10.821186065673828, 'timing/data': 0.02137599978595972, 'timing/model': 1.408261532895267, 'learning_rate/qwen_vl_interface': 6.32181431172154e-06, 'learning_rate/action_model': 6.32181431172154e-05, 'epoch': 1.6}) +[2026-06-16 01:02:49,889][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.16996820271015167, 'train/grad_norm_pre_clip': 5.322207927703857, 'timing/data': 0.021669411100447178, 'timing/model': 1.6769692385569215, 'learning_rate/qwen_vl_interface': 6.318847212921295e-06, 'learning_rate/action_model': 6.318847212921295e-05, 'epoch': 1.6}) +[2026-06-16 01:02:55,764][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.12537124752998352, 'train/grad_norm_pre_clip': 4.535499572753906, 'timing/data': 0.020136184059083462, 'timing/model': 1.6640937821939588, 'learning_rate/qwen_vl_interface': 6.3158796747583364e-06, 'learning_rate/action_model': 6.315879674758336e-05, 'epoch': 1.6}) +[2026-06-16 01:03:01,352][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.15117420256137848, 'train/grad_norm_pre_clip': 6.610434532165527, 'timing/data': 0.020511356182396412, 'timing/model': 1.4368699826300144, 'learning_rate/qwen_vl_interface': 6.312911698452503e-06, 'learning_rate/action_model': 6.312911698452502e-05, 'epoch': 1.6}) +[2026-06-16 01:03:07,176][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.11944234371185303, 'train/grad_norm_pre_clip': 4.640855312347412, 'timing/data': 0.028313995338976383, 'timing/model': 1.4490567976608872, 'learning_rate/qwen_vl_interface': 6.309943285223824e-06, 'learning_rate/action_model': 6.309943285223824e-05, 'epoch': 1.6}) +[2026-06-16 01:13:47,987][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.20326457917690277, 'train/grad_norm_pre_clip': 6.719473361968994, 'eval/loss': 0.0794539600610733, 'eval/flap_precision': 0.6974063400576369, 'eval/flap_recall': 0.88, 'eval/flap_f1': 0.7781350482315111, 'eval/flap_rate_model': 0.09638888888888889, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7781350482315111, 'timing/data': 0.02173654269427061, 'timing/model': 1.4591560997068882, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 1.6}) +[2026-06-16 01:13:47,988][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3500_state +[2026-06-16 01:13:47,995][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 01:13:56,289][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3500_state/model.safetensors +[2026-06-16 01:14:07,640][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3500_state/optimizer.bin +[2026-06-16 01:14:07,641][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3500_state/sampler.bin +[2026-06-16 01:14:07,641][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-16 01:14:07,643][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-16 01:14:07,795][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-16 01:14:07,796][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 01:14:07,811][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 01:14:13,548][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.22094158828258514, 'train/grad_norm_pre_clip': 9.525667190551758, 'timing/data': 0.022600487805902958, 'timing/model': 1.3733626333996654, 'learning_rate/qwen_vl_interface': 6.304005152878902e-06, 'learning_rate/action_model': 6.304005152878901e-05, 'epoch': 1.6}) +[2026-06-16 01:14:19,228][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.09164547920227051, 'train/grad_norm_pre_clip': 7.28148078918457, 'timing/data': 0.0290913013741374, 'timing/model': 1.43891563359648, 'learning_rate/qwen_vl_interface': 6.301035436203602e-06, 'learning_rate/action_model': 6.301035436203602e-05, 'epoch': 1.6}) +[2026-06-16 01:14:24,842][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.202809140086174, 'train/grad_norm_pre_clip': 6.9743194580078125, 'timing/data': 0.022054865024983883, 'timing/model': 1.3967766147106886, 'learning_rate/qwen_vl_interface': 6.298065287487337e-06, 'learning_rate/action_model': 6.298065287487337e-05, 'epoch': 1.6}) +[2026-06-16 01:14:30,437][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.09228150546550751, 'train/grad_norm_pre_clip': 4.545285701751709, 'timing/data': 0.02624676562845707, 'timing/model': 1.6250965828076005, 'learning_rate/qwen_vl_interface': 6.295094707951019e-06, 'learning_rate/action_model': 6.295094707951019e-05, 'epoch': 1.6}) +[2026-06-16 01:14:35,833][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.06274263560771942, 'train/grad_norm_pre_clip': 3.2119970321655273, 'timing/data': 0.023129290901124477, 'timing/model': 1.4306155815720558, 'learning_rate/qwen_vl_interface': 6.292123698815745e-06, 'learning_rate/action_model': 6.292123698815745e-05, 'epoch': 1.6}) +[2026-06-16 01:14:41,533][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.05046495050191879, 'train/grad_norm_pre_clip': 1.7703980207443237, 'timing/data': 0.020473276264965534, 'timing/model': 1.4324422609061003, 'learning_rate/qwen_vl_interface': 6.289152261302783e-06, 'learning_rate/action_model': 6.289152261302784e-05, 'epoch': 1.6}) +[2026-06-16 01:14:47,233][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.09716975688934326, 'train/grad_norm_pre_clip': 2.9438467025756836, 'timing/data': 0.022023473866283894, 'timing/model': 1.4445865843445063, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 1.6}) +[2026-06-16 01:14:52,833][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.219138503074646, 'train/grad_norm_pre_clip': 8.429383277893066, 'timing/data': 0.02069367840886116, 'timing/model': 1.4024761961773038, 'learning_rate/qwen_vl_interface': 6.283208106029753e-06, 'learning_rate/action_model': 6.283208106029752e-05, 'epoch': 1.6}) +[2026-06-16 01:14:58,542][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.14400377869606018, 'train/grad_norm_pre_clip': 5.454931735992432, 'timing/data': 0.02163998130708933, 'timing/model': 1.4404446640983224, 'learning_rate/qwen_vl_interface': 6.280235390713102e-06, 'learning_rate/action_model': 6.280235390713102e-05, 'epoch': 1.6}) +[2026-06-16 01:15:04,279][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.16682842373847961, 'train/grad_norm_pre_clip': 7.607537269592285, 'timing/data': 0.021060999482870102, 'timing/model': 1.6824490129947662, 'learning_rate/qwen_vl_interface': 6.277262251905597e-06, 'learning_rate/action_model': 6.277262251905597e-05, 'epoch': 1.61}) +[2026-06-16 01:15:09,836][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.14813804626464844, 'train/grad_norm_pre_clip': 6.794864177703857, 'timing/data': 0.02402153518050909, 'timing/model': 1.4314812179654837, 'learning_rate/qwen_vl_interface': 6.274288690829385e-06, 'learning_rate/action_model': 6.274288690829384e-05, 'epoch': 1.61}) +[2026-06-16 01:15:15,540][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.1786988228559494, 'train/grad_norm_pre_clip': 12.163392066955566, 'timing/data': 0.02279814798384905, 'timing/model': 1.4330674363300204, 'learning_rate/qwen_vl_interface': 6.271314708706781e-06, 'learning_rate/action_model': 6.271314708706782e-05, 'epoch': 1.61}) +[2026-06-16 01:15:21,244][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.07472652941942215, 'train/grad_norm_pre_clip': 5.87587833404541, 'timing/data': 0.023180020973086357, 'timing/model': 1.4465172262862325, 'learning_rate/qwen_vl_interface': 6.26834030676028e-06, 'learning_rate/action_model': 6.26834030676028e-05, 'epoch': 1.61}) +[2026-06-16 01:15:26,946][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.1148918867111206, 'train/grad_norm_pre_clip': 4.855257034301758, 'timing/data': 0.022131276316940784, 'timing/model': 1.4504310563206673, 'learning_rate/qwen_vl_interface': 6.265365486212545e-06, 'learning_rate/action_model': 6.265365486212545e-05, 'epoch': 1.61}) +[2026-06-16 01:15:32,642][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.14738579094409943, 'train/grad_norm_pre_clip': 9.326998710632324, 'timing/data': 0.02133774198591709, 'timing/model': 1.4313630871474743, 'learning_rate/qwen_vl_interface': 6.26239024828641e-06, 'learning_rate/action_model': 6.26239024828641e-05, 'epoch': 1.61}) +[2026-06-16 01:15:38,373][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.1930089145898819, 'train/grad_norm_pre_clip': 7.0412163734436035, 'timing/data': 0.021493292413651943, 'timing/model': 1.6663118312135339, 'learning_rate/qwen_vl_interface': 6.259414594204888e-06, 'learning_rate/action_model': 6.259414594204888e-05, 'epoch': 1.61}) +[2026-06-16 01:15:43,941][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.10369129478931427, 'train/grad_norm_pre_clip': 6.016934871673584, 'timing/data': 0.020683408714830875, 'timing/model': 1.4381994437426329, 'learning_rate/qwen_vl_interface': 6.256438525191155e-06, 'learning_rate/action_model': 6.256438525191156e-05, 'epoch': 1.61}) +[2026-06-16 01:15:49,641][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.17083942890167236, 'train/grad_norm_pre_clip': 6.192497253417969, 'timing/data': 0.020988648757338524, 'timing/model': 1.4270371161401272, 'learning_rate/qwen_vl_interface': 6.253462042468561e-06, 'learning_rate/action_model': 6.25346204246856e-05, 'epoch': 1.61}) +[2026-06-16 01:15:55,346][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.0843992829322815, 'train/grad_norm_pre_clip': 5.488507270812988, 'timing/data': 0.020997689105570316, 'timing/model': 1.4443762665614486, 'learning_rate/qwen_vl_interface': 6.2504851472606246e-06, 'learning_rate/action_model': 6.250485147260625e-05, 'epoch': 1.61}) +[2026-06-16 01:16:01,043][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.30784860253334045, 'train/grad_norm_pre_clip': 10.877676963806152, 'timing/data': 0.024306947365403175, 'timing/model': 1.4001922672614455, 'learning_rate/qwen_vl_interface': 6.247507840791035e-06, 'learning_rate/action_model': 6.247507840791035e-05, 'epoch': 1.61}) +[2026-06-16 01:16:06,642][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.1723451167345047, 'train/grad_norm_pre_clip': 6.126696586608887, 'timing/data': 0.02175104431807995, 'timing/model': 1.4072342729195952, 'learning_rate/qwen_vl_interface': 6.244530124283651e-06, 'learning_rate/action_model': 6.24453012428365e-05, 'epoch': 1.61}) +[2026-06-16 01:16:12,380][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.07103515416383743, 'train/grad_norm_pre_clip': 4.075403690338135, 'timing/data': 0.021520162001252174, 'timing/model': 1.667903141118586, 'learning_rate/qwen_vl_interface': 6.2415519989625016e-06, 'learning_rate/action_model': 6.241551998962501e-05, 'epoch': 1.61}) +[2026-06-16 01:16:17,961][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.11719381809234619, 'train/grad_norm_pre_clip': 3.404320478439331, 'timing/data': 0.024961049668490887, 'timing/model': 1.4500954374670982, 'learning_rate/qwen_vl_interface': 6.238573466051781e-06, 'learning_rate/action_model': 6.23857346605178e-05, 'epoch': 1.61}) +[2026-06-16 01:16:23,751][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.1594974398612976, 'train/grad_norm_pre_clip': 5.2319512367248535, 'timing/data': 0.020762838423252106, 'timing/model': 1.4584079012274742, 'learning_rate/qwen_vl_interface': 6.2355945267758465e-06, 'learning_rate/action_model': 6.235594526775846e-05, 'epoch': 1.61}) +[2026-06-16 01:16:29,444][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.05212097615003586, 'train/grad_norm_pre_clip': 4.841263294219971, 'timing/data': 0.023764164187014103, 'timing/model': 1.4267479665577412, 'learning_rate/qwen_vl_interface': 6.232615182359231e-06, 'learning_rate/action_model': 6.232615182359231e-05, 'epoch': 1.61}) +[2026-06-16 01:16:35,046][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.0816815197467804, 'train/grad_norm_pre_clip': 4.932862758636475, 'timing/data': 0.023004550486803055, 'timing/model': 1.4387975493445992, 'learning_rate/qwen_vl_interface': 6.229635434026634e-06, 'learning_rate/action_model': 6.229635434026634e-05, 'epoch': 1.61}) +[2026-06-16 01:16:40,751][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.2098611742258072, 'train/grad_norm_pre_clip': 11.142926216125488, 'timing/data': 0.021169669926166534, 'timing/model': 1.4308691779151559, 'learning_rate/qwen_vl_interface': 6.226655283002912e-06, 'learning_rate/action_model': 6.226655283002911e-05, 'epoch': 1.61}) +[2026-06-16 01:16:46,584][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.185612291097641, 'train/grad_norm_pre_clip': 14.921788215637207, 'timing/data': 0.02161322347819805, 'timing/model': 1.6878598155453801, 'learning_rate/qwen_vl_interface': 6.223674730513096e-06, 'learning_rate/action_model': 6.223674730513096e-05, 'epoch': 1.61}) +[2026-06-16 01:16:52,149][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.1057029590010643, 'train/grad_norm_pre_clip': 6.017967224121094, 'timing/data': 0.02116688061505556, 'timing/model': 1.4500496676191688, 'learning_rate/qwen_vl_interface': 6.220693777782375e-06, 'learning_rate/action_model': 6.220693777782375e-05, 'epoch': 1.61}) +[2026-06-16 01:16:57,850][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.1325540840625763, 'train/grad_norm_pre_clip': 7.776062965393066, 'timing/data': 0.021613321267068386, 'timing/model': 1.4293775409460068, 'learning_rate/qwen_vl_interface': 6.21771242603611e-06, 'learning_rate/action_model': 6.21771242603611e-05, 'epoch': 1.61}) +[2026-06-16 01:17:03,659][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.12319593876600266, 'train/grad_norm_pre_clip': 6.204838752746582, 'timing/data': 0.021126659587025642, 'timing/model': 1.4393786629661918, 'learning_rate/qwen_vl_interface': 6.21473067649982e-06, 'learning_rate/action_model': 6.21473067649982e-05, 'epoch': 1.61}) +[2026-06-16 01:17:09,448][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.07316991686820984, 'train/grad_norm_pre_clip': 3.980966806411743, 'timing/data': 0.019821882247924805, 'timing/model': 1.456275600939989, 'learning_rate/qwen_vl_interface': 6.211748530399189e-06, 'learning_rate/action_model': 6.211748530399189e-05, 'epoch': 1.62}) +[2026-06-16 01:17:15,254][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.1411585807800293, 'train/grad_norm_pre_clip': 5.569387912750244, 'timing/data': 0.021543491631746292, 'timing/model': 1.4253204707056284, 'learning_rate/qwen_vl_interface': 6.208765988960067e-06, 'learning_rate/action_model': 6.208765988960066e-05, 'epoch': 1.62}) +[2026-06-16 01:17:21,081][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.04416743665933609, 'train/grad_norm_pre_clip': 4.334507942199707, 'timing/data': 0.020815378986299038, 'timing/model': 1.682434119284153, 'learning_rate/qwen_vl_interface': 6.2057830534084615e-06, 'learning_rate/action_model': 6.205783053408461e-05, 'epoch': 1.62}) +[2026-06-16 01:17:26,850][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.10692992806434631, 'train/grad_norm_pre_clip': 8.345054626464844, 'timing/data': 0.02117524016648531, 'timing/model': 1.6479871300980449, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 1.62}) +[2026-06-16 01:17:32,663][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.15145668387413025, 'train/grad_norm_pre_clip': 6.653320789337158, 'timing/data': 0.0209860997274518, 'timing/model': 1.451009513810277, 'learning_rate/qwen_vl_interface': 6.199816004872655e-06, 'learning_rate/action_model': 6.199816004872655e-05, 'epoch': 1.62}) +[2026-06-16 01:17:38,478][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.24579404294490814, 'train/grad_norm_pre_clip': 21.760141372680664, 'timing/data': 0.021644793450832367, 'timing/model': 1.4583450825884938, 'learning_rate/qwen_vl_interface': 6.196831894341281e-06, 'learning_rate/action_model': 6.19683189434128e-05, 'epoch': 1.62}) +[2026-06-16 01:17:44,367][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.17807847261428833, 'train/grad_norm_pre_clip': 22.86227798461914, 'timing/data': 0.020451227203011513, 'timing/model': 1.472072054632008, 'learning_rate/qwen_vl_interface': 6.1938473946030795e-06, 'learning_rate/action_model': 6.193847394603079e-05, 'epoch': 1.62}) +[2026-06-16 01:17:50,156][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.07721999287605286, 'train/grad_norm_pre_clip': 3.164104700088501, 'timing/data': 0.021802234463393688, 'timing/model': 1.4080019406974316, 'learning_rate/qwen_vl_interface': 6.1908625068848695e-06, 'learning_rate/action_model': 6.19086250688487e-05, 'epoch': 1.62}) +[2026-06-16 01:17:55,977][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.26356014609336853, 'train/grad_norm_pre_clip': 13.997207641601562, 'timing/data': 0.020802109502255917, 'timing/model': 1.6821838393807411, 'learning_rate/qwen_vl_interface': 6.187877232413619e-06, 'learning_rate/action_model': 6.18787723241362e-05, 'epoch': 1.62}) +[2026-06-16 01:18:01,544][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.049573689699172974, 'train/grad_norm_pre_clip': 3.9611239433288574, 'timing/data': 0.021450990810990334, 'timing/model': 1.4345454797148705, 'learning_rate/qwen_vl_interface': 6.184891572416467e-06, 'learning_rate/action_model': 6.184891572416466e-05, 'epoch': 1.62}) +[2026-06-16 01:18:07,346][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.27621132135391235, 'train/grad_norm_pre_clip': 10.696009635925293, 'timing/data': 0.021429670974612236, 'timing/model': 1.4450025651603937, 'learning_rate/qwen_vl_interface': 6.181905528120703e-06, 'learning_rate/action_model': 6.181905528120703e-05, 'epoch': 1.62}) +[2026-06-16 01:18:13,050][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.10278013348579407, 'train/grad_norm_pre_clip': 5.1839728355407715, 'timing/data': 0.021261010318994522, 'timing/model': 1.435699826106429, 'learning_rate/qwen_vl_interface': 6.178919100753778e-06, 'learning_rate/action_model': 6.178919100753778e-05, 'epoch': 1.62}) +[2026-06-16 01:18:18,848][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.27218911051750183, 'train/grad_norm_pre_clip': 14.5640287399292, 'timing/data': 0.021391731686890125, 'timing/model': 1.4525835243985057, 'learning_rate/qwen_vl_interface': 6.175932291543299e-06, 'learning_rate/action_model': 6.175932291543298e-05, 'epoch': 1.62}) +[2026-06-16 01:18:24,643][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.12356171011924744, 'train/grad_norm_pre_clip': 7.920444965362549, 'timing/data': 0.02079788874834776, 'timing/model': 1.406376744620502, 'learning_rate/qwen_vl_interface': 6.172945101717029e-06, 'learning_rate/action_model': 6.17294510171703e-05, 'epoch': 1.62}) +[2026-06-16 01:18:30,386][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.1569802314043045, 'train/grad_norm_pre_clip': 5.833685874938965, 'timing/data': 0.021183110773563385, 'timing/model': 1.6707635009661317, 'learning_rate/qwen_vl_interface': 6.169957532502892e-06, 'learning_rate/action_model': 6.169957532502892e-05, 'epoch': 1.62}) +[2026-06-16 01:18:36,137][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.13165472447872162, 'train/grad_norm_pre_clip': 6.032380104064941, 'timing/data': 0.021957955323159695, 'timing/model': 1.621361124329269, 'learning_rate/qwen_vl_interface': 6.166969585128962e-06, 'learning_rate/action_model': 6.166969585128962e-05, 'epoch': 1.62}) +[2026-06-16 01:18:41,649][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.11026348918676376, 'train/grad_norm_pre_clip': 7.83322286605835, 'timing/data': 0.021099450066685677, 'timing/model': 1.4440570399165154, 'learning_rate/qwen_vl_interface': 6.163981260823473e-06, 'learning_rate/action_model': 6.163981260823473e-05, 'epoch': 1.62}) +[2026-06-16 01:18:47,453][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.2090255618095398, 'train/grad_norm_pre_clip': 15.921252250671387, 'timing/data': 0.021032970398664474, 'timing/model': 1.413983704522252, 'learning_rate/qwen_vl_interface': 6.16099256081481e-06, 'learning_rate/action_model': 6.16099256081481e-05, 'epoch': 1.62}) +[2026-06-16 01:18:53,250][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.07266020774841309, 'train/grad_norm_pre_clip': 3.8162150382995605, 'timing/data': 0.021867393516004086, 'timing/model': 1.4367597633972764, 'learning_rate/qwen_vl_interface': 6.15800348633152e-06, 'learning_rate/action_model': 6.15800348633152e-05, 'epoch': 1.62}) +[2026-06-16 01:18:59,054][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.1569381207227707, 'train/grad_norm_pre_clip': 7.44137716293335, 'timing/data': 0.021256251260638237, 'timing/model': 1.421886625699699, 'learning_rate/qwen_vl_interface': 6.155014038602293e-06, 'learning_rate/action_model': 6.155014038602292e-05, 'epoch': 1.62}) +[2026-06-16 01:19:04,882][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.09391535818576813, 'train/grad_norm_pre_clip': 3.3989970684051514, 'timing/data': 0.021438811905682087, 'timing/model': 1.6750463247299194, 'learning_rate/qwen_vl_interface': 6.152024218855979e-06, 'learning_rate/action_model': 6.15202421885598e-05, 'epoch': 1.62}) +[2026-06-16 01:19:10,686][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.10150370746850967, 'train/grad_norm_pre_clip': 4.34584903717041, 'timing/data': 0.02180071361362934, 'timing/model': 1.6846487345173955, 'learning_rate/qwen_vl_interface': 6.149034028321583e-06, 'learning_rate/action_model': 6.149034028321583e-05, 'epoch': 1.62}) +[2026-06-16 01:19:16,247][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.05571708083152771, 'train/grad_norm_pre_clip': 2.4657809734344482, 'timing/data': 0.02139455173164606, 'timing/model': 1.4373327773064375, 'learning_rate/qwen_vl_interface': 6.1460434682282566e-06, 'learning_rate/action_model': 6.146043468228256e-05, 'epoch': 1.63}) +[2026-06-16 01:19:22,049][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.056226667016744614, 'train/grad_norm_pre_clip': 3.1837761402130127, 'timing/data': 0.021016390062868595, 'timing/model': 1.4361023902893066, 'learning_rate/qwen_vl_interface': 6.143052539805308e-06, 'learning_rate/action_model': 6.143052539805308e-05, 'epoch': 1.63}) +[2026-06-16 01:19:27,848][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.10390839725732803, 'train/grad_norm_pre_clip': 4.821990489959717, 'timing/data': 0.021325040608644485, 'timing/model': 1.4529225677251816, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 1.63}) +[2026-06-16 01:19:33,548][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.13329876959323883, 'train/grad_norm_pre_clip': 4.508643627166748, 'timing/data': 0.0245022876188159, 'timing/model': 1.4044234659522772, 'learning_rate/qwen_vl_interface': 6.137069582888521e-06, 'learning_rate/action_model': 6.137069582888521e-05, 'epoch': 1.63}) +[2026-06-16 01:19:39,387][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.08701569586992264, 'train/grad_norm_pre_clip': 4.530757427215576, 'timing/data': 0.02082767803221941, 'timing/model': 1.6858258014544845, 'learning_rate/qwen_vl_interface': 6.13407755685405e-06, 'learning_rate/action_model': 6.13407755685405e-05, 'epoch': 1.63}) +[2026-06-16 01:19:44,949][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.0724472627043724, 'train/grad_norm_pre_clip': 6.547100067138672, 'timing/data': 0.020843757316470146, 'timing/model': 1.437106966972351, 'learning_rate/qwen_vl_interface': 6.131085167408692e-06, 'learning_rate/action_model': 6.131085167408692e-05, 'epoch': 1.63}) +[2026-06-16 01:19:50,745][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.09597886353731155, 'train/grad_norm_pre_clip': 4.318948745727539, 'timing/data': 0.02080295793712139, 'timing/model': 1.4442982636392117, 'learning_rate/qwen_vl_interface': 6.128092415782502e-06, 'learning_rate/action_model': 6.128092415782501e-05, 'epoch': 1.63}) +[2026-06-16 01:19:56,545][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.20116961002349854, 'train/grad_norm_pre_clip': 12.453155517578125, 'timing/data': 0.021162409335374832, 'timing/model': 1.4341304525732994, 'learning_rate/qwen_vl_interface': 6.125099303205688e-06, 'learning_rate/action_model': 6.125099303205688e-05, 'epoch': 1.63}) +[2026-06-16 01:20:02,245][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.1640823781490326, 'train/grad_norm_pre_clip': 8.291792869567871, 'timing/data': 0.021230660378932953, 'timing/model': 1.4045789083465934, 'learning_rate/qwen_vl_interface': 6.122105830908607e-06, 'learning_rate/action_model': 6.122105830908606e-05, 'epoch': 1.63}) +[2026-06-16 01:20:07,934][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.18718494474887848, 'train/grad_norm_pre_clip': 14.357373237609863, 'timing/data': 0.020746607333421707, 'timing/model': 1.4520124541595578, 'learning_rate/qwen_vl_interface': 6.11911200012176e-06, 'learning_rate/action_model': 6.11911200012176e-05, 'epoch': 1.63}) +[2026-06-16 01:20:13,647][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.09335629642009735, 'train/grad_norm_pre_clip': 4.877711772918701, 'timing/data': 0.020906229503452778, 'timing/model': 1.6450055604800582, 'learning_rate/qwen_vl_interface': 6.116117812075799e-06, 'learning_rate/action_model': 6.116117812075799e-05, 'epoch': 1.63}) +[2026-06-16 01:20:19,145][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.018548188731074333, 'train/grad_norm_pre_clip': 1.7887132167816162, 'timing/data': 0.020382926799356937, 'timing/model': 1.4364865738898516, 'learning_rate/qwen_vl_interface': 6.113123268001524e-06, 'learning_rate/action_model': 6.113123268001524e-05, 'epoch': 1.63}) +[2026-06-16 01:20:24,950][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.05656614527106285, 'train/grad_norm_pre_clip': 4.616373062133789, 'timing/data': 0.021143830381333828, 'timing/model': 1.4335225196555257, 'learning_rate/qwen_vl_interface': 6.110128369129873e-06, 'learning_rate/action_model': 6.110128369129873e-05, 'epoch': 1.63}) +[2026-06-16 01:20:30,766][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.13662266731262207, 'train/grad_norm_pre_clip': 10.729233741760254, 'timing/data': 0.025838634930551052, 'timing/model': 1.4606076898053288, 'learning_rate/qwen_vl_interface': 6.107133116691943e-06, 'learning_rate/action_model': 6.107133116691942e-05, 'epoch': 1.63}) +[2026-06-16 01:20:36,551][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.11543566733598709, 'train/grad_norm_pre_clip': 6.807178020477295, 'timing/data': 0.02081505861133337, 'timing/model': 1.4078414142131805, 'learning_rate/qwen_vl_interface': 6.104137511918963e-06, 'learning_rate/action_model': 6.104137511918963e-05, 'epoch': 1.63}) +[2026-06-16 01:20:42,385][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.1325652003288269, 'train/grad_norm_pre_clip': 10.851360321044922, 'timing/data': 0.02148254308849573, 'timing/model': 1.4071689303964376, 'learning_rate/qwen_vl_interface': 6.1011415560423155e-06, 'learning_rate/action_model': 6.101141556042316e-05, 'epoch': 1.63}) +[2026-06-16 01:20:48,273][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.13866810500621796, 'train/grad_norm_pre_clip': 7.614311218261719, 'timing/data': 0.024773908779025078, 'timing/model': 1.6467076102271676, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 1.63}) +[2026-06-16 01:20:53,973][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.22277377545833588, 'train/grad_norm_pre_clip': 6.729010105133057, 'timing/data': 0.021159459836781025, 'timing/model': 1.4576567150652409, 'learning_rate/qwen_vl_interface': 6.09514859590426e-06, 'learning_rate/action_model': 6.09514859590426e-05, 'epoch': 1.63}) +[2026-06-16 01:20:59,855][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.29956701397895813, 'train/grad_norm_pre_clip': 17.770246505737305, 'timing/data': 0.020518497563898563, 'timing/model': 1.4494272619485855, 'learning_rate/qwen_vl_interface': 6.09215159410633e-06, 'learning_rate/action_model': 6.09215159410633e-05, 'epoch': 1.63}) +[2026-06-16 01:21:05,655][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.09142988175153732, 'train/grad_norm_pre_clip': 10.08588981628418, 'timing/data': 0.02042704727500677, 'timing/model': 1.4496618434786797, 'learning_rate/qwen_vl_interface': 6.089154246131692e-06, 'learning_rate/action_model': 6.089154246131692e-05, 'epoch': 1.63}) +[2026-06-16 01:21:11,471][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.13764749467372894, 'train/grad_norm_pre_clip': 10.387752532958984, 'timing/data': 0.02131054177880287, 'timing/model': 1.4069885620847344, 'learning_rate/qwen_vl_interface': 6.08615655321244e-06, 'learning_rate/action_model': 6.08615655321244e-05, 'epoch': 1.63}) +[2026-06-16 01:21:17,296][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.09589947015047073, 'train/grad_norm_pre_clip': 7.915603160858154, 'timing/data': 0.02063914854079485, 'timing/model': 1.6879615057259798, 'learning_rate/qwen_vl_interface': 6.083158516580814e-06, 'learning_rate/action_model': 6.083158516580814e-05, 'epoch': 1.63}) +[2026-06-16 01:21:22,980][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.18598826229572296, 'train/grad_norm_pre_clip': 15.589957237243652, 'timing/data': 0.020834038965404034, 'timing/model': 1.4624174507334828, 'learning_rate/qwen_vl_interface': 6.080160137469192e-06, 'learning_rate/action_model': 6.0801601374691916e-05, 'epoch': 1.64}) +[2026-06-16 01:21:28,859][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.04452044144272804, 'train/grad_norm_pre_clip': 3.8170552253723145, 'timing/data': 0.0208201976493001, 'timing/model': 1.4427547277882695, 'learning_rate/qwen_vl_interface': 6.077161417110097e-06, 'learning_rate/action_model': 6.0771614171100965e-05, 'epoch': 1.64}) +[2026-06-16 01:21:34,554][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.18660281598567963, 'train/grad_norm_pre_clip': 8.723270416259766, 'timing/data': 0.021435100585222244, 'timing/model': 1.4497870057821274, 'learning_rate/qwen_vl_interface': 6.0741623567361864e-06, 'learning_rate/action_model': 6.074162356736186e-05, 'epoch': 1.64}) +[2026-06-16 01:21:40,246][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.44913461804389954, 'train/grad_norm_pre_clip': 18.117313385009766, 'timing/data': 0.02092800848186016, 'timing/model': 1.434805196709931, 'learning_rate/qwen_vl_interface': 6.071162957580262e-06, 'learning_rate/action_model': 6.071162957580262e-05, 'epoch': 1.64}) +[2026-06-16 01:21:45,959][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.11913507431745529, 'train/grad_norm_pre_clip': 8.218825340270996, 'timing/data': 0.02105807326734066, 'timing/model': 1.4644505605101585, 'learning_rate/qwen_vl_interface': 6.068163220875266e-06, 'learning_rate/action_model': 6.068163220875266e-05, 'epoch': 1.64}) +[2026-06-16 01:21:51,794][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.2534142732620239, 'train/grad_norm_pre_clip': 10.390921592712402, 'timing/data': 0.020865125581622124, 'timing/model': 1.6841739350929856, 'learning_rate/qwen_vl_interface': 6.065163147854273e-06, 'learning_rate/action_model': 6.065163147854273e-05, 'epoch': 1.64}) +[2026-06-16 01:21:57,449][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.06777095794677734, 'train/grad_norm_pre_clip': 5.163197994232178, 'timing/data': 0.021513836458325386, 'timing/model': 1.4452226934954524, 'learning_rate/qwen_vl_interface': 6.062162739750502e-06, 'learning_rate/action_model': 6.062162739750502e-05, 'epoch': 1.64}) +[2026-06-16 01:22:03,259][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.011048566550016403, 'train/grad_norm_pre_clip': 0.8144173622131348, 'timing/data': 0.02091037854552269, 'timing/model': 1.4456418007612228, 'learning_rate/qwen_vl_interface': 6.059161997797306e-06, 'learning_rate/action_model': 6.0591619977973056e-05, 'epoch': 1.64}) +[2026-06-16 01:22:09,060][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.15041929483413696, 'train/grad_norm_pre_clip': 10.882608413696289, 'timing/data': 0.020288598723709583, 'timing/model': 1.464997235685587, 'learning_rate/qwen_vl_interface': 6.056160923228178e-06, 'learning_rate/action_model': 6.056160923228178e-05, 'epoch': 1.64}) +[2026-06-16 01:22:14,853][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.10843150317668915, 'train/grad_norm_pre_clip': 5.207616806030273, 'timing/data': 0.02198321931064129, 'timing/model': 1.4459970630705357, 'learning_rate/qwen_vl_interface': 6.0531595172767455e-06, 'learning_rate/action_model': 6.0531595172767455e-05, 'epoch': 1.64}) +[2026-06-16 01:22:20,655][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.14128120243549347, 'train/grad_norm_pre_clip': 6.861665725708008, 'timing/data': 0.021835431456565857, 'timing/model': 1.4449807293713093, 'learning_rate/qwen_vl_interface': 6.050157781176774e-06, 'learning_rate/action_model': 6.050157781176774e-05, 'epoch': 1.64}) +[2026-06-16 01:22:26,493][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.03456820175051689, 'train/grad_norm_pre_clip': 2.463453531265259, 'timing/data': 0.023634033277630806, 'timing/model': 1.687015576288104, 'learning_rate/qwen_vl_interface': 6.047155716162164e-06, 'learning_rate/action_model': 6.047155716162164e-05, 'epoch': 1.64}) +[2026-06-16 01:22:32,131][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.5058726072311401, 'train/grad_norm_pre_clip': 12.591386795043945, 'timing/data': 0.030320785008370876, 'timing/model': 1.4120272286236286, 'learning_rate/qwen_vl_interface': 6.044153323466948e-06, 'learning_rate/action_model': 6.044153323466948e-05, 'epoch': 1.64}) +[2026-06-16 01:22:37,952][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.09741050004959106, 'train/grad_norm_pre_clip': 8.268563270568848, 'timing/data': 0.021665753796696663, 'timing/model': 1.4488189769908786, 'learning_rate/qwen_vl_interface': 6.0411506043253e-06, 'learning_rate/action_model': 6.0411506043253e-05, 'epoch': 1.64}) +[2026-06-16 01:22:43,775][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.21769043803215027, 'train/grad_norm_pre_clip': 8.359963417053223, 'timing/data': 0.02177593484520912, 'timing/model': 1.4610722726210952, 'learning_rate/qwen_vl_interface': 6.038147559971523e-06, 'learning_rate/action_model': 6.038147559971523e-05, 'epoch': 1.64}) +[2026-06-16 01:22:49,562][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.1329016089439392, 'train/grad_norm_pre_clip': 7.62478494644165, 'timing/data': 0.021472887136042118, 'timing/model': 1.4067633086815476, 'learning_rate/qwen_vl_interface': 6.035144191640055e-06, 'learning_rate/action_model': 6.035144191640055e-05, 'epoch': 1.64}) +[2026-06-16 01:22:55,249][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.11409658193588257, 'train/grad_norm_pre_clip': 7.230771064758301, 'timing/data': 0.0204548267647624, 'timing/model': 1.3991678729653358, 'learning_rate/qwen_vl_interface': 6.032140500565467e-06, 'learning_rate/action_model': 6.0321405005654665e-05, 'epoch': 1.64}) +[2026-06-16 01:23:01,097][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.1539696902036667, 'train/grad_norm_pre_clip': 9.0399808883667, 'timing/data': 0.021746749058365822, 'timing/model': 1.6875727539882064, 'learning_rate/qwen_vl_interface': 6.029136487982463e-06, 'learning_rate/action_model': 6.029136487982463e-05, 'epoch': 1.64}) +[2026-06-16 01:23:06,755][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.4211680293083191, 'train/grad_norm_pre_clip': 9.422557830810547, 'timing/data': 0.02078453917056322, 'timing/model': 1.4396782033145428, 'learning_rate/qwen_vl_interface': 6.02613215512588e-06, 'learning_rate/action_model': 6.0261321551258796e-05, 'epoch': 1.64}) +[2026-06-16 01:23:12,459][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.13001960515975952, 'train/grad_norm_pre_clip': 6.77698278427124, 'timing/data': 0.020320020616054535, 'timing/model': 1.4251789031550288, 'learning_rate/qwen_vl_interface': 6.023127503230684e-06, 'learning_rate/action_model': 6.023127503230683e-05, 'epoch': 1.64}) +[2026-06-16 01:23:18,255][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.037263356149196625, 'train/grad_norm_pre_clip': 3.3107259273529053, 'timing/data': 0.023436032235622406, 'timing/model': 1.4521007528528571, 'learning_rate/qwen_vl_interface': 6.020122533531976e-06, 'learning_rate/action_model': 6.0201225335319755e-05, 'epoch': 1.64}) +[2026-06-16 01:23:23,952][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.09138428419828415, 'train/grad_norm_pre_clip': 5.040307521820068, 'timing/data': 0.02070577349513769, 'timing/model': 1.450984375551343, 'learning_rate/qwen_vl_interface': 6.017117247264983e-06, 'learning_rate/action_model': 6.017117247264983e-05, 'epoch': 1.65}) +[2026-06-16 01:23:30,252][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.2516891062259674, 'train/grad_norm_pre_clip': 11.396644592285156, 'timing/data': 0.024678404442965984, 'timing/model': 1.4069540314376354, 'learning_rate/qwen_vl_interface': 6.014111645665069e-06, 'learning_rate/action_model': 6.0141116456650684e-05, 'epoch': 1.65}) +[2026-06-16 01:23:36,085][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.11005285382270813, 'train/grad_norm_pre_clip': 9.078705787658691, 'timing/data': 0.02104979380965233, 'timing/model': 1.674788094125688, 'learning_rate/qwen_vl_interface': 6.011105729967719e-06, 'learning_rate/action_model': 6.0111057299677184e-05, 'epoch': 1.65}) +[2026-06-16 01:34:12,269][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.1934494972229004, 'train/grad_norm_pre_clip': 7.811989784240723, 'eval/loss': 0.07427025586366653, 'eval/flap_precision': 0.953757225433526, 'eval/flap_recall': 0.6, 'eval/flap_f1': 0.7366071428571428, 'eval/flap_rate_model': 0.04805555555555555, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7366071428571428, 'timing/data': 0.020525413565337658, 'timing/model': 1.4504205957055092, 'learning_rate/qwen_vl_interface': 6.008099501408552e-06, 'learning_rate/action_model': 6.008099501408552e-05, 'epoch': 1.65}) +[2026-06-16 01:34:12,270][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3600_state +[2026-06-16 01:34:12,277][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 01:34:19,974][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3600_state/model.safetensors +[2026-06-16 01:34:31,070][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3600_state/optimizer.bin +[2026-06-16 01:34:31,071][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3600_state/sampler.bin +[2026-06-16 01:34:31,071][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3600_state/sampler_1.bin +[2026-06-16 01:34:31,073][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3600_state/random_states_0.pkl +[2026-06-16 01:34:31,236][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3600 +[2026-06-16 01:34:31,237][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 01:34:31,252][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 01:34:36,974][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.2809596657752991, 'train/grad_norm_pre_clip': 7.402853488922119, 'timing/data': 0.021874187514185905, 'timing/model': 1.6692457282915711, 'learning_rate/qwen_vl_interface': 6.0050929612233165e-06, 'learning_rate/action_model': 6.005092961223316e-05, 'epoch': 1.65}) +[2026-06-16 01:34:42,428][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.2703377604484558, 'train/grad_norm_pre_clip': 6.194592475891113, 'timing/data': 0.02068652305752039, 'timing/model': 1.4472931744530797, 'learning_rate/qwen_vl_interface': 6.002086110647888e-06, 'learning_rate/action_model': 6.002086110647888e-05, 'epoch': 1.65}) +[2026-06-16 01:34:48,135][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.1833130419254303, 'train/grad_norm_pre_clip': 5.0602874755859375, 'timing/data': 0.02213348913937807, 'timing/model': 1.446740717627108, 'learning_rate/qwen_vl_interface': 5.999078950918267e-06, 'learning_rate/action_model': 5.999078950918267e-05, 'epoch': 1.65}) +[2026-06-16 01:34:53,738][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.17498716711997986, 'train/grad_norm_pre_clip': 5.7740960121154785, 'timing/data': 0.02589432243257761, 'timing/model': 1.454147250391543, 'learning_rate/qwen_vl_interface': 5.996071483270585e-06, 'learning_rate/action_model': 5.996071483270585e-05, 'epoch': 1.65}) +[2026-06-16 01:34:59,447][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.0461847260594368, 'train/grad_norm_pre_clip': 1.947344422340393, 'timing/data': 0.02125557605177164, 'timing/model': 1.4483651481568813, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 1.65}) +[2026-06-16 01:35:05,243][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.09167437255382538, 'train/grad_norm_pre_clip': 4.28098201751709, 'timing/data': 0.025504550896584988, 'timing/model': 1.4069847064092755, 'learning_rate/qwen_vl_interface': 5.990055629166187e-06, 'learning_rate/action_model': 5.990055629166187e-05, 'epoch': 1.65}) +[2026-06-16 01:35:10,973][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.056612130254507065, 'train/grad_norm_pre_clip': 2.3646676540374756, 'timing/data': 0.02137351781129837, 'timing/model': 1.671512302942574, 'learning_rate/qwen_vl_interface': 5.987047245182363e-06, 'learning_rate/action_model': 5.987047245182362e-05, 'epoch': 1.65}) +[2026-06-16 01:35:16,537][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.113825723528862, 'train/grad_norm_pre_clip': 4.348184108734131, 'timing/data': 0.0202539823949337, 'timing/model': 1.4506417848169804, 'learning_rate/qwen_vl_interface': 5.984038558226255e-06, 'learning_rate/action_model': 5.9840385582262546e-05, 'epoch': 1.65}) +[2026-06-16 01:35:22,249][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.2113233506679535, 'train/grad_norm_pre_clip': 8.903836250305176, 'timing/data': 0.02221952099353075, 'timing/model': 1.4176975404843688, 'learning_rate/qwen_vl_interface': 5.981029569534625e-06, 'learning_rate/action_model': 5.981029569534625e-05, 'epoch': 1.65}) +[2026-06-16 01:35:27,958][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.11049959808588028, 'train/grad_norm_pre_clip': 3.264984607696533, 'timing/data': 0.02132473699748516, 'timing/model': 1.4564595585688949, 'learning_rate/qwen_vl_interface': 5.978020280344351e-06, 'learning_rate/action_model': 5.9780202803443506e-05, 'epoch': 1.65}) +[2026-06-16 01:35:33,850][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.09197208285331726, 'train/grad_norm_pre_clip': 3.7918753623962402, 'timing/data': 0.021264086477458477, 'timing/model': 1.4334943611174822, 'learning_rate/qwen_vl_interface': 5.97501069189244e-06, 'learning_rate/action_model': 5.97501069189244e-05, 'epoch': 1.65}) +[2026-06-16 01:35:39,559][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.17728471755981445, 'train/grad_norm_pre_clip': 6.176870346069336, 'timing/data': 0.02072400599718094, 'timing/model': 1.413356652483344, 'learning_rate/qwen_vl_interface': 5.97200080541602e-06, 'learning_rate/action_model': 5.97200080541602e-05, 'epoch': 1.65}) +[2026-06-16 01:35:45,400][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.08989676833152771, 'train/grad_norm_pre_clip': 4.028975486755371, 'timing/data': 0.020804925821721554, 'timing/model': 1.699418774805963, 'learning_rate/qwen_vl_interface': 5.968990622152341e-06, 'learning_rate/action_model': 5.9689906221523404e-05, 'epoch': 1.65}) +[2026-06-16 01:35:51,199][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.046090539544820786, 'train/grad_norm_pre_clip': 2.4900898933410645, 'timing/data': 0.02119634859263897, 'timing/model': 1.684499197639525, 'learning_rate/qwen_vl_interface': 5.965980143338774e-06, 'learning_rate/action_model': 5.965980143338774e-05, 'epoch': 1.65}) +[2026-06-16 01:35:56,754][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.05553486570715904, 'train/grad_norm_pre_clip': 3.3146934509277344, 'timing/data': 0.02087313774973154, 'timing/model': 1.447818974032998, 'learning_rate/qwen_vl_interface': 5.962969370212819e-06, 'learning_rate/action_model': 5.9629693702128185e-05, 'epoch': 1.65}) +[2026-06-16 01:36:02,561][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.11755737662315369, 'train/grad_norm_pre_clip': 6.186275005340576, 'timing/data': 0.02400428894907236, 'timing/model': 1.455127066001296, 'learning_rate/qwen_vl_interface': 5.959958304012085e-06, 'learning_rate/action_model': 5.9599583040120845e-05, 'epoch': 1.65}) +[2026-06-16 01:36:08,345][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.19842685759067535, 'train/grad_norm_pre_clip': 9.043352127075195, 'timing/data': 0.021702580153942108, 'timing/model': 1.4490932784974575, 'learning_rate/qwen_vl_interface': 5.956946945974309e-06, 'learning_rate/action_model': 5.9569469459743085e-05, 'epoch': 1.65}) +[2026-06-16 01:36:14,156][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.16999030113220215, 'train/grad_norm_pre_clip': 6.206092834472656, 'timing/data': 0.020146753638982773, 'timing/model': 1.4675756618380547, 'learning_rate/qwen_vl_interface': 5.953935297337347e-06, 'learning_rate/action_model': 5.953935297337347e-05, 'epoch': 1.65}) +[2026-06-16 01:36:19,885][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.01785087026655674, 'train/grad_norm_pre_clip': 0.8939548134803772, 'timing/data': 0.020779048092663288, 'timing/model': 1.7002858323976398, 'learning_rate/qwen_vl_interface': 5.950923359339175e-06, 'learning_rate/action_model': 5.950923359339174e-05, 'epoch': 1.66}) +[2026-06-16 01:36:25,449][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.08810865879058838, 'train/grad_norm_pre_clip': 5.110397815704346, 'timing/data': 0.02130493987351656, 'timing/model': 1.4433329002931714, 'learning_rate/qwen_vl_interface': 5.947911133217886e-06, 'learning_rate/action_model': 5.947911133217886e-05, 'epoch': 1.66}) +[2026-06-16 01:36:31,152][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.22256790101528168, 'train/grad_norm_pre_clip': 8.143643379211426, 'timing/data': 0.0213266396895051, 'timing/model': 1.4423292400315404, 'learning_rate/qwen_vl_interface': 5.944898620211693e-06, 'learning_rate/action_model': 5.944898620211693e-05, 'epoch': 1.66}) +[2026-06-16 01:36:36,841][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.10960138589143753, 'train/grad_norm_pre_clip': 3.8509645462036133, 'timing/data': 0.02060643769800663, 'timing/model': 1.4526038533076644, 'learning_rate/qwen_vl_interface': 5.941885821558925e-06, 'learning_rate/action_model': 5.9418858215589245e-05, 'epoch': 1.66}) +[2026-06-16 01:36:42,541][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.047324251383543015, 'train/grad_norm_pre_clip': 3.1136600971221924, 'timing/data': 0.022305483929812908, 'timing/model': 1.461888280697167, 'learning_rate/qwen_vl_interface': 5.938872738498031e-06, 'learning_rate/action_model': 5.93887273849803e-05, 'epoch': 1.66}) +[2026-06-16 01:36:48,190][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.5001846551895142, 'train/grad_norm_pre_clip': 20.16374397277832, 'timing/data': 0.021226460114121437, 'timing/model': 1.4110827129334211, 'learning_rate/qwen_vl_interface': 5.935859372267575e-06, 'learning_rate/action_model': 5.935859372267575e-05, 'epoch': 1.66}) +[2026-06-16 01:36:53,948][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.29012635350227356, 'train/grad_norm_pre_clip': 13.769887924194336, 'timing/data': 0.021032960154116154, 'timing/model': 1.6410543154925108, 'learning_rate/qwen_vl_interface': 5.9328457241062396e-06, 'learning_rate/action_model': 5.9328457241062394e-05, 'epoch': 1.66}) +[2026-06-16 01:36:59,564][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.15135785937309265, 'train/grad_norm_pre_clip': 9.466341972351074, 'timing/data': 0.024927684105932713, 'timing/model': 1.4574383348226547, 'learning_rate/qwen_vl_interface': 5.929831795252819e-06, 'learning_rate/action_model': 5.9298317952528184e-05, 'epoch': 1.66}) +[2026-06-16 01:37:05,249][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.09974123537540436, 'train/grad_norm_pre_clip': 6.9767608642578125, 'timing/data': 0.022367543540894985, 'timing/model': 1.4172879457473755, 'learning_rate/qwen_vl_interface': 5.926817586946227e-06, 'learning_rate/action_model': 5.926817586946227e-05, 'epoch': 1.66}) +[2026-06-16 01:37:11,049][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.32898664474487305, 'train/grad_norm_pre_clip': 13.540275573730469, 'timing/data': 0.020396238192915916, 'timing/model': 1.451266578398645, 'learning_rate/qwen_vl_interface': 5.923803100425491e-06, 'learning_rate/action_model': 5.92380310042549e-05, 'epoch': 1.66}) +[2026-06-16 01:37:16,852][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.04991159215569496, 'train/grad_norm_pre_clip': 4.150085926055908, 'timing/data': 0.02160447183996439, 'timing/model': 1.4039603108540177, 'learning_rate/qwen_vl_interface': 5.92078833692975e-06, 'learning_rate/action_model': 5.92078833692975e-05, 'epoch': 1.66}) +[2026-06-16 01:37:22,689][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.0019165021367371082, 'train/grad_norm_pre_clip': 0.22141262888908386, 'timing/data': 0.020711390301585197, 'timing/model': 1.440309651196003, 'learning_rate/qwen_vl_interface': 5.917773297698264e-06, 'learning_rate/action_model': 5.9177732976982636e-05, 'epoch': 1.66}) +[2026-06-16 01:37:28,459][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.18895438313484192, 'train/grad_norm_pre_clip': 10.215765953063965, 'timing/data': 0.020944731310009956, 'timing/model': 1.6481946017593145, 'learning_rate/qwen_vl_interface': 5.9147579839704e-06, 'learning_rate/action_model': 5.9147579839704e-05, 'epoch': 1.66}) +[2026-06-16 01:37:34,043][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.07232419401407242, 'train/grad_norm_pre_clip': 6.903772354125977, 'timing/data': 0.01922199409455061, 'timing/model': 1.4174591787159443, 'learning_rate/qwen_vl_interface': 5.911742396985637e-06, 'learning_rate/action_model': 5.9117423969856364e-05, 'epoch': 1.66}) +[2026-06-16 01:37:39,858][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.06418034434318542, 'train/grad_norm_pre_clip': 4.214238166809082, 'timing/data': 0.02207665517926216, 'timing/model': 1.4573707645758986, 'learning_rate/qwen_vl_interface': 5.908726537983571e-06, 'learning_rate/action_model': 5.908726537983571e-05, 'epoch': 1.66}) +[2026-06-16 01:37:45,652][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.06773053854703903, 'train/grad_norm_pre_clip': 4.129647731781006, 'timing/data': 0.021361461840569973, 'timing/model': 1.4433475425466895, 'learning_rate/qwen_vl_interface': 5.9057104082039085e-06, 'learning_rate/action_model': 5.905710408203908e-05, 'epoch': 1.66}) +[2026-06-16 01:37:51,456][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.04910103976726532, 'train/grad_norm_pre_clip': 4.75453519821167, 'timing/data': 0.0236755209043622, 'timing/model': 1.401681363582611, 'learning_rate/qwen_vl_interface': 5.902694008886466e-06, 'learning_rate/action_model': 5.902694008886466e-05, 'epoch': 1.66}) +[2026-06-16 01:37:57,285][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.09230151772499084, 'train/grad_norm_pre_clip': 5.224830627441406, 'timing/data': 0.020615278743207455, 'timing/model': 1.6739860121160746, 'learning_rate/qwen_vl_interface': 5.8996773412711694e-06, 'learning_rate/action_model': 5.899677341271169e-05, 'epoch': 1.66}) +[2026-06-16 01:38:02,854][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.20245985686779022, 'train/grad_norm_pre_clip': 7.306556701660156, 'timing/data': 0.021377292461693287, 'timing/model': 1.4196174554526806, 'learning_rate/qwen_vl_interface': 5.896660406598059e-06, 'learning_rate/action_model': 5.896660406598059e-05, 'epoch': 1.66}) +[2026-06-16 01:38:08,664][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.16490305960178375, 'train/grad_norm_pre_clip': 6.059251308441162, 'timing/data': 0.02155790477991104, 'timing/model': 1.4633348491042852, 'learning_rate/qwen_vl_interface': 5.893643206107283e-06, 'learning_rate/action_model': 5.893643206107282e-05, 'epoch': 1.66}) +[2026-06-16 01:38:14,355][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.2509223222732544, 'train/grad_norm_pre_clip': 14.214309692382812, 'timing/data': 0.022186647169291973, 'timing/model': 1.4535949919372797, 'learning_rate/qwen_vl_interface': 5.890625741039096e-06, 'learning_rate/action_model': 5.890625741039095e-05, 'epoch': 1.66}) +[2026-06-16 01:38:20,081][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.010809156112372875, 'train/grad_norm_pre_clip': 1.292568325996399, 'timing/data': 0.02077350113540888, 'timing/model': 1.455278130248189, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 1.66}) +[2026-06-16 01:38:25,883][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.02256122976541519, 'train/grad_norm_pre_clip': 1.9308381080627441, 'timing/data': 0.020969292148947716, 'timing/model': 1.4904065085574985, 'learning_rate/qwen_vl_interface': 5.884590022132064e-06, 'learning_rate/action_model': 5.884590022132064e-05, 'epoch': 1.67}) +[2026-06-16 01:38:31,796][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.10877169668674469, 'train/grad_norm_pre_clip': 4.842075824737549, 'timing/data': 0.020542429760098457, 'timing/model': 1.6948512829840183, 'learning_rate/qwen_vl_interface': 5.881571770774278e-06, 'learning_rate/action_model': 5.881571770774278e-05, 'epoch': 1.67}) +[2026-06-16 01:38:37,453][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.007529245223850012, 'train/grad_norm_pre_clip': 1.7022417783737183, 'timing/data': 0.0197615884244442, 'timing/model': 1.4514614595100284, 'learning_rate/qwen_vl_interface': 5.878553259801192e-06, 'learning_rate/action_model': 5.878553259801192e-05, 'epoch': 1.67}) +[2026-06-16 01:38:43,264][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.3246728181838989, 'train/grad_norm_pre_clip': 10.717622756958008, 'timing/data': 0.021500583738088608, 'timing/model': 1.4506242973729968, 'learning_rate/qwen_vl_interface': 5.8755344904536025e-06, 'learning_rate/action_model': 5.8755344904536025e-05, 'epoch': 1.67}) +[2026-06-16 01:38:49,054][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.17453829944133759, 'train/grad_norm_pre_clip': 7.937262058258057, 'timing/data': 0.021896126680076122, 'timing/model': 1.4545295918360353, 'learning_rate/qwen_vl_interface': 5.872515463972415e-06, 'learning_rate/action_model': 5.8725154639724146e-05, 'epoch': 1.67}) +[2026-06-16 01:38:54,950][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.06285686045885086, 'train/grad_norm_pre_clip': 3.8906090259552, 'timing/data': 0.021008582785725594, 'timing/model': 1.458746531046927, 'learning_rate/qwen_vl_interface': 5.869496181598633e-06, 'learning_rate/action_model': 5.869496181598633e-05, 'epoch': 1.67}) +[2026-06-16 01:39:00,750][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.25513336062431335, 'train/grad_norm_pre_clip': 11.789684295654297, 'timing/data': 0.02120751328766346, 'timing/model': 1.447578846476972, 'learning_rate/qwen_vl_interface': 5.866476644573371e-06, 'learning_rate/action_model': 5.866476644573371e-05, 'epoch': 1.67}) +[2026-06-16 01:39:06,592][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.15805624425411224, 'train/grad_norm_pre_clip': 8.423985481262207, 'timing/data': 0.021535225212574005, 'timing/model': 1.6869421685114503, 'learning_rate/qwen_vl_interface': 5.863456854137848e-06, 'learning_rate/action_model': 5.8634568541378474e-05, 'epoch': 1.67}) +[2026-06-16 01:39:12,255][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.03766952455043793, 'train/grad_norm_pre_clip': 4.077553749084473, 'timing/data': 0.022609899751842022, 'timing/model': 1.4357647402212024, 'learning_rate/qwen_vl_interface': 5.860436811533383e-06, 'learning_rate/action_model': 5.860436811533383e-05, 'epoch': 1.67}) +[2026-06-16 01:39:18,051][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.03004678152501583, 'train/grad_norm_pre_clip': 4.0850372314453125, 'timing/data': 0.024787187576293945, 'timing/model': 1.4204221405088902, 'learning_rate/qwen_vl_interface': 5.857416518001404e-06, 'learning_rate/action_model': 5.857416518001404e-05, 'epoch': 1.67}) +[2026-06-16 01:39:23,855][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.014210250228643417, 'train/grad_norm_pre_clip': 1.827477216720581, 'timing/data': 0.02162998728454113, 'timing/model': 1.4315964141860604, 'learning_rate/qwen_vl_interface': 5.854395974783439e-06, 'learning_rate/action_model': 5.8543959747834386e-05, 'epoch': 1.67}) +[2026-06-16 01:39:29,662][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.1933761090040207, 'train/grad_norm_pre_clip': 14.122891426086426, 'timing/data': 0.021517695859074593, 'timing/model': 1.418887053616345, 'learning_rate/qwen_vl_interface': 5.851375183121119e-06, 'learning_rate/action_model': 5.8513751831211186e-05, 'epoch': 1.67}) +[2026-06-16 01:39:35,354][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.25010013580322266, 'train/grad_norm_pre_clip': 13.366790771484375, 'timing/data': 0.021764937788248062, 'timing/model': 1.430110378190875, 'learning_rate/qwen_vl_interface': 5.848354144256175e-06, 'learning_rate/action_model': 5.8483541442561754e-05, 'epoch': 1.67}) +[2026-06-16 01:39:41,204][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.08112214505672455, 'train/grad_norm_pre_clip': 5.864413261413574, 'timing/data': 0.022628781385719776, 'timing/model': 1.6937635550275445, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 1.67}) +[2026-06-16 01:39:46,750][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.34374451637268066, 'train/grad_norm_pre_clip': 13.97193431854248, 'timing/data': 0.02236425969749689, 'timing/model': 1.4360492005944252, 'learning_rate/qwen_vl_interface': 5.842311329885866e-06, 'learning_rate/action_model': 5.842311329885866e-05, 'epoch': 1.67}) +[2026-06-16 01:39:52,559][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.30246350169181824, 'train/grad_norm_pre_clip': 13.117222785949707, 'timing/data': 0.022478519938886166, 'timing/model': 1.4607539400458336, 'learning_rate/qwen_vl_interface': 5.839289556864472e-06, 'learning_rate/action_model': 5.8392895568644714e-05, 'epoch': 1.67}) +[2026-06-16 01:39:58,364][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.06772536784410477, 'train/grad_norm_pre_clip': 8.118099212646484, 'timing/data': 0.021073954179883003, 'timing/model': 1.4468399239704013, 'learning_rate/qwen_vl_interface': 5.836267541608398e-06, 'learning_rate/action_model': 5.836267541608398e-05, 'epoch': 1.67}) +[2026-06-16 01:40:04,039][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.0938560739159584, 'train/grad_norm_pre_clip': 5.106957912445068, 'timing/data': 0.02248246129602194, 'timing/model': 1.4135338878259063, 'learning_rate/qwen_vl_interface': 5.833245285359887e-06, 'learning_rate/action_model': 5.833245285359886e-05, 'epoch': 1.67}) +[2026-06-16 01:40:09,757][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.10352933406829834, 'train/grad_norm_pre_clip': 13.492264747619629, 'timing/data': 0.022012908942997456, 'timing/model': 1.416034096851945, 'learning_rate/qwen_vl_interface': 5.830222789361268e-06, 'learning_rate/action_model': 5.8302227893612684e-05, 'epoch': 1.67}) +[2026-06-16 01:40:15,584][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.06902158260345459, 'train/grad_norm_pre_clip': 5.131227493286133, 'timing/data': 0.020583233796060085, 'timing/model': 1.6770033231005073, 'learning_rate/qwen_vl_interface': 5.8272000548549776e-06, 'learning_rate/action_model': 5.827200054854978e-05, 'epoch': 1.67}) +[2026-06-16 01:40:21,161][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.0613786019384861, 'train/grad_norm_pre_clip': 3.8084588050842285, 'timing/data': 0.02293514274060726, 'timing/model': 1.4476927230134606, 'learning_rate/qwen_vl_interface': 5.824177083083547e-06, 'learning_rate/action_model': 5.8241770830835475e-05, 'epoch': 1.67}) +[2026-06-16 01:40:26,951][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.18520082533359528, 'train/grad_norm_pre_clip': 53.0462646484375, 'timing/data': 0.020847193896770477, 'timing/model': 1.4356603724882007, 'learning_rate/qwen_vl_interface': 5.8211538752896056e-06, 'learning_rate/action_model': 5.821153875289606e-05, 'epoch': 1.67}) +[2026-06-16 01:40:32,762][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.2863900065422058, 'train/grad_norm_pre_clip': 16.360355377197266, 'timing/data': 0.019545788876712322, 'timing/model': 1.459187707863748, 'learning_rate/qwen_vl_interface': 5.8181304327158814e-06, 'learning_rate/action_model': 5.8181304327158814e-05, 'epoch': 1.68}) +[2026-06-16 01:40:38,563][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.22115804255008698, 'train/grad_norm_pre_clip': 10.583059310913086, 'timing/data': 0.019169828854501247, 'timing/model': 1.4703860320150852, 'learning_rate/qwen_vl_interface': 5.8151067566051965e-06, 'learning_rate/action_model': 5.815106756605196e-05, 'epoch': 1.68}) +[2026-06-16 01:40:44,349][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.019220342859625816, 'train/grad_norm_pre_clip': 1.7605798244476318, 'timing/data': 0.02045399323105812, 'timing/model': 1.4032251965254545, 'learning_rate/qwen_vl_interface': 5.81208284820047e-06, 'learning_rate/action_model': 5.81208284820047e-05, 'epoch': 1.68}) +[2026-06-16 01:40:50,085][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.11500541865825653, 'train/grad_norm_pre_clip': 9.922420501708984, 'timing/data': 0.020679624751210213, 'timing/model': 1.6693701818585396, 'learning_rate/qwen_vl_interface': 5.809058708744716e-06, 'learning_rate/action_model': 5.809058708744716e-05, 'epoch': 1.68}) +[2026-06-16 01:40:55,649][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.014349167235195637, 'train/grad_norm_pre_clip': 1.1156964302062988, 'timing/data': 0.020972615107893944, 'timing/model': 1.440403145737946, 'learning_rate/qwen_vl_interface': 5.8060343394810445e-06, 'learning_rate/action_model': 5.806034339481045e-05, 'epoch': 1.68}) +[2026-06-16 01:41:01,445][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.23864762485027313, 'train/grad_norm_pre_clip': 12.250964164733887, 'timing/data': 0.02256918139755726, 'timing/model': 1.4417908396571875, 'learning_rate/qwen_vl_interface': 5.803009741652659e-06, 'learning_rate/action_model': 5.8030097416526585e-05, 'epoch': 1.68}) +[2026-06-16 01:41:07,254][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.19996221363544464, 'train/grad_norm_pre_clip': 13.109713554382324, 'timing/data': 0.020883765071630478, 'timing/model': 1.4464853974059224, 'learning_rate/qwen_vl_interface': 5.799984916502859e-06, 'learning_rate/action_model': 5.799984916502859e-05, 'epoch': 1.68}) +[2026-06-16 01:41:13,058][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.1994868665933609, 'train/grad_norm_pre_clip': 11.972086906433105, 'timing/data': 0.02151412796229124, 'timing/model': 1.4506798814982176, 'learning_rate/qwen_vl_interface': 5.7969598652750334e-06, 'learning_rate/action_model': 5.796959865275033e-05, 'epoch': 1.68}) +[2026-06-16 01:41:18,850][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.16771849989891052, 'train/grad_norm_pre_clip': 8.308135986328125, 'timing/data': 0.020229103043675423, 'timing/model': 1.4139009304344654, 'learning_rate/qwen_vl_interface': 5.793934589212666e-06, 'learning_rate/action_model': 5.793934589212666e-05, 'epoch': 1.68}) +[2026-06-16 01:41:24,589][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.44519978761672974, 'train/grad_norm_pre_clip': 16.242494583129883, 'timing/data': 0.0194343701004982, 'timing/model': 1.6870474927127361, 'learning_rate/qwen_vl_interface': 5.7909090895593375e-06, 'learning_rate/action_model': 5.790909089559338e-05, 'epoch': 1.68}) +[2026-06-16 01:41:30,157][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.04059202969074249, 'train/grad_norm_pre_clip': 3.3686537742614746, 'timing/data': 0.021964791230857372, 'timing/model': 1.4481531344354153, 'learning_rate/qwen_vl_interface': 5.7878833675587135e-06, 'learning_rate/action_model': 5.787883367558713e-05, 'epoch': 1.68}) +[2026-06-16 01:41:35,943][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.05063982680439949, 'train/grad_norm_pre_clip': 3.479613780975342, 'timing/data': 0.02095506712794304, 'timing/model': 1.4398189093917608, 'learning_rate/qwen_vl_interface': 5.784857424454556e-06, 'learning_rate/action_model': 5.7848574244545563e-05, 'epoch': 1.68}) +[2026-06-16 01:41:41,645][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.10770410299301147, 'train/grad_norm_pre_clip': 6.836599826812744, 'timing/data': 0.018358775414526463, 'timing/model': 1.4520296370610595, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 1.68}) +[2026-06-16 01:41:47,361][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.06204524636268616, 'train/grad_norm_pre_clip': 14.426311492919922, 'timing/data': 0.021705579943954945, 'timing/model': 1.4577446281909943, 'learning_rate/qwen_vl_interface': 5.778804879911129e-06, 'learning_rate/action_model': 5.778804879911128e-05, 'epoch': 1.68}) +[2026-06-16 01:41:53,144][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.14926238358020782, 'train/grad_norm_pre_clip': 11.656010627746582, 'timing/data': 0.01883142814040184, 'timing/model': 1.4151816591620445, 'learning_rate/qwen_vl_interface': 5.775778280959832e-06, 'learning_rate/action_model': 5.7757782809598315e-05, 'epoch': 1.68}) +[2026-06-16 01:41:58,886][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.35263022780418396, 'train/grad_norm_pre_clip': 16.205245971679688, 'timing/data': 0.019158239476382732, 'timing/model': 1.6796168545261025, 'learning_rate/qwen_vl_interface': 5.7727514658809435e-06, 'learning_rate/action_model': 5.772751465880944e-05, 'epoch': 1.68}) +[2026-06-16 01:42:04,617][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.0758601576089859, 'train/grad_norm_pre_clip': 5.229694366455078, 'timing/data': 0.02014671266078949, 'timing/model': 1.6117015108466148, 'learning_rate/qwen_vl_interface': 5.769724435918676e-06, 'learning_rate/action_model': 5.7697244359186755e-05, 'epoch': 1.68}) +[2026-06-16 01:42:10,154][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.12632706761360168, 'train/grad_norm_pre_clip': 7.795004367828369, 'timing/data': 0.019740101881325245, 'timing/model': 1.4506500586867332, 'learning_rate/qwen_vl_interface': 5.766697192317323e-06, 'learning_rate/action_model': 5.766697192317323e-05, 'epoch': 1.68}) +[2026-06-16 01:42:15,858][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.0581226721405983, 'train/grad_norm_pre_clip': 7.169754981994629, 'timing/data': 0.0192600404843688, 'timing/model': 1.3943742914125323, 'learning_rate/qwen_vl_interface': 5.763669736321271e-06, 'learning_rate/action_model': 5.763669736321271e-05, 'epoch': 1.68}) +[2026-06-16 01:42:21,657][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.06711571663618088, 'train/grad_norm_pre_clip': 6.762676239013672, 'timing/data': 0.019934932701289654, 'timing/model': 1.4516027867794037, 'learning_rate/qwen_vl_interface': 5.760642069174994e-06, 'learning_rate/action_model': 5.760642069174994e-05, 'epoch': 1.68}) +[2026-06-16 01:42:27,445][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.32523709535598755, 'train/grad_norm_pre_clip': 10.441555976867676, 'timing/data': 0.018712928518652916, 'timing/model': 1.408229892142117, 'learning_rate/qwen_vl_interface': 5.757614192123051e-06, 'learning_rate/action_model': 5.757614192123051e-05, 'epoch': 1.68}) +[2026-06-16 01:42:33,183][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.05761214718222618, 'train/grad_norm_pre_clip': 6.0184645652771, 'timing/data': 0.019462880678474903, 'timing/model': 1.674892314709723, 'learning_rate/qwen_vl_interface': 5.754586106410089e-06, 'learning_rate/action_model': 5.754586106410088e-05, 'epoch': 1.68}) +[2026-06-16 01:42:38,984][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.042424559593200684, 'train/grad_norm_pre_clip': 4.55003547668457, 'timing/data': 0.020730807445943356, 'timing/model': 1.6746770609170198, 'learning_rate/qwen_vl_interface': 5.751557813280838e-06, 'learning_rate/action_model': 5.751557813280838e-05, 'epoch': 1.69}) +[2026-06-16 01:42:44,554][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.2708204984664917, 'train/grad_norm_pre_clip': 12.410573959350586, 'timing/data': 0.020597665570676327, 'timing/model': 1.449528326280415, 'learning_rate/qwen_vl_interface': 5.7485293139801146e-06, 'learning_rate/action_model': 5.7485293139801146e-05, 'epoch': 1.69}) +[2026-06-16 01:42:50,368][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.15721489489078522, 'train/grad_norm_pre_clip': 16.058311462402344, 'timing/data': 0.02120377868413925, 'timing/model': 1.4595251139253378, 'learning_rate/qwen_vl_interface': 5.745500609752821e-06, 'learning_rate/action_model': 5.745500609752821e-05, 'epoch': 1.69}) +[2026-06-16 01:42:56,145][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.06880024820566177, 'train/grad_norm_pre_clip': 5.531103610992432, 'timing/data': 0.021605931222438812, 'timing/model': 1.446218954399228, 'learning_rate/qwen_vl_interface': 5.742471701843943e-06, 'learning_rate/action_model': 5.742471701843943e-05, 'epoch': 1.69}) +[2026-06-16 01:43:01,884][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.14429864287376404, 'train/grad_norm_pre_clip': 6.263228416442871, 'timing/data': 0.021774991415441036, 'timing/model': 1.4500118475407362, 'learning_rate/qwen_vl_interface': 5.739442591498553e-06, 'learning_rate/action_model': 5.739442591498553e-05, 'epoch': 1.69}) +[2026-06-16 01:43:07,812][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.07536135613918304, 'train/grad_norm_pre_clip': 6.136785984039307, 'timing/data': 0.021178978495299816, 'timing/model': 1.6953274384140968, 'learning_rate/qwen_vl_interface': 5.7364132799617985e-06, 'learning_rate/action_model': 5.7364132799617984e-05, 'epoch': 1.69}) +[2026-06-16 01:43:13,464][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.1753316968679428, 'train/grad_norm_pre_clip': 10.504239082336426, 'timing/data': 0.022581614553928375, 'timing/model': 1.4441330963745713, 'learning_rate/qwen_vl_interface': 5.733383768478923e-06, 'learning_rate/action_model': 5.7333837684789225e-05, 'epoch': 1.69}) +[2026-06-16 01:43:19,275][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.1722131371498108, 'train/grad_norm_pre_clip': 11.81392765045166, 'timing/data': 0.022198214195668697, 'timing/model': 1.4690930675715208, 'learning_rate/qwen_vl_interface': 5.730354058295234e-06, 'learning_rate/action_model': 5.730354058295234e-05, 'epoch': 1.69}) +[2026-06-16 01:43:25,175][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.08493586629629135, 'train/grad_norm_pre_clip': 5.767541408538818, 'timing/data': 0.02012214343994856, 'timing/model': 1.4781787134706974, 'learning_rate/qwen_vl_interface': 5.72732415065614e-06, 'learning_rate/action_model': 5.727324150656139e-05, 'epoch': 1.69}) +[2026-06-16 01:43:31,047][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.06012915447354317, 'train/grad_norm_pre_clip': 4.795566082000732, 'timing/data': 0.019878704100847244, 'timing/model': 1.416588339023292, 'learning_rate/qwen_vl_interface': 5.724294046807115e-06, 'learning_rate/action_model': 5.724294046807115e-05, 'epoch': 1.69}) +[2026-06-16 01:43:36,673][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.05933835357427597, 'train/grad_norm_pre_clip': 5.607051849365234, 'timing/data': 0.020164365880191326, 'timing/model': 1.3910253159701824, 'learning_rate/qwen_vl_interface': 5.7212637479937234e-06, 'learning_rate/action_model': 5.721263747993724e-05, 'epoch': 1.69}) +[2026-06-16 01:43:42,455][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.0869145393371582, 'train/grad_norm_pre_clip': 6.204031944274902, 'timing/data': 0.02036360651254654, 'timing/model': 1.6586152464151382, 'learning_rate/qwen_vl_interface': 5.718233255461607e-06, 'learning_rate/action_model': 5.718233255461607e-05, 'epoch': 1.69}) +[2026-06-16 01:43:48,041][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.0993320569396019, 'train/grad_norm_pre_clip': 8.335412979125977, 'timing/data': 0.01984680350869894, 'timing/model': 1.4502481194213033, 'learning_rate/qwen_vl_interface': 5.715202570456486e-06, 'learning_rate/action_model': 5.715202570456486e-05, 'epoch': 1.69}) +[2026-06-16 01:43:53,846][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.030315158888697624, 'train/grad_norm_pre_clip': 3.104349374771118, 'timing/data': 0.020270916633307934, 'timing/model': 1.4510042080655694, 'learning_rate/qwen_vl_interface': 5.7121716942241615e-06, 'learning_rate/action_model': 5.712171694224161e-05, 'epoch': 1.69}) +[2026-06-16 01:43:59,541][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.21499329805374146, 'train/grad_norm_pre_clip': 11.900344848632812, 'timing/data': 0.020907028578221798, 'timing/model': 1.4474725890904665, 'learning_rate/qwen_vl_interface': 5.709140628010509e-06, 'learning_rate/action_model': 5.7091406280105085e-05, 'epoch': 1.69}) +[2026-06-16 01:54:44,977][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.16895711421966553, 'train/grad_norm_pre_clip': 15.255239486694336, 'eval/loss': 0.06659720093011856, 'eval/flap_precision': 0.8728070175438597, 'eval/flap_recall': 0.7236363636363636, 'eval/flap_f1': 0.7912524850894633, 'eval/flap_rate_model': 0.06333333333333334, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7912524850894633, 'timing/data': 0.019316461868584156, 'timing/model': 1.4445719132199883, 'learning_rate/qwen_vl_interface': 5.706109373061489e-06, 'learning_rate/action_model': 5.706109373061489e-05, 'epoch': 1.69}) +[2026-06-16 01:54:44,978][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3700_state +[2026-06-16 01:54:44,985][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 01:54:53,281][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3700_state/model.safetensors +[2026-06-16 01:55:04,680][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3700_state/optimizer.bin +[2026-06-16 01:55:04,681][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3700_state/sampler.bin +[2026-06-16 01:55:04,681][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3700_state/sampler_1.bin +[2026-06-16 01:55:04,683][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3700_state/random_states_0.pkl +[2026-06-16 01:55:04,854][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3700 +[2026-06-16 01:55:04,855][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 01:55:04,870][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 01:55:10,743][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.07121656835079193, 'train/grad_norm_pre_clip': 8.97911548614502, 'timing/data': 0.026848320849239826, 'timing/model': 1.450262911617756, 'learning_rate/qwen_vl_interface': 5.703077930623134e-06, 'learning_rate/action_model': 5.7030779306231344e-05, 'epoch': 1.69}) +[2026-06-16 01:55:16,450][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.3210076093673706, 'train/grad_norm_pre_clip': 12.422036170959473, 'timing/data': 0.023159444332122803, 'timing/model': 1.4245650172233582, 'learning_rate/qwen_vl_interface': 5.7000463019415565e-06, 'learning_rate/action_model': 5.7000463019415563e-05, 'epoch': 1.69}) +[2026-06-16 01:55:22,175][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.11547236144542694, 'train/grad_norm_pre_clip': 6.992812633514404, 'timing/data': 0.02155716810375452, 'timing/model': 1.6720872558653355, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 1.69}) +[2026-06-16 01:55:27,739][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.014077448286116123, 'train/grad_norm_pre_clip': 1.816570520401001, 'timing/data': 0.021187515929341316, 'timing/model': 1.4534692009910941, 'learning_rate/qwen_vl_interface': 5.693982490833561e-06, 'learning_rate/action_model': 5.693982490833561e-05, 'epoch': 1.69}) +[2026-06-16 01:55:33,425][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.2587434649467468, 'train/grad_norm_pre_clip': 14.44685173034668, 'timing/data': 0.018884525634348392, 'timing/model': 1.436978586949408, 'learning_rate/qwen_vl_interface': 5.690950310899745e-06, 'learning_rate/action_model': 5.6909503108997456e-05, 'epoch': 1.69}) +[2026-06-16 01:55:39,133][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.06979912519454956, 'train/grad_norm_pre_clip': 4.039145469665527, 'timing/data': 0.021598417311906815, 'timing/model': 1.4335835045203567, 'learning_rate/qwen_vl_interface': 5.687917949707913e-06, 'learning_rate/action_model': 5.687917949707913e-05, 'epoch': 1.69}) +[2026-06-16 01:55:44,752][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.09841186553239822, 'train/grad_norm_pre_clip': 5.5156168937683105, 'timing/data': 0.02166004851460457, 'timing/model': 1.4659963520243764, 'learning_rate/qwen_vl_interface': 5.684885408504555e-06, 'learning_rate/action_model': 5.684885408504554e-05, 'epoch': 1.7}) +[2026-06-16 01:55:50,443][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.2523361146450043, 'train/grad_norm_pre_clip': 11.59168529510498, 'timing/data': 0.020459883846342564, 'timing/model': 1.446512734517455, 'learning_rate/qwen_vl_interface': 5.681852688536231e-06, 'learning_rate/action_model': 5.681852688536231e-05, 'epoch': 1.7}) +[2026-06-16 01:55:56,092][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.07892698794603348, 'train/grad_norm_pre_clip': 12.276098251342773, 'timing/data': 0.020032990723848343, 'timing/model': 1.6143400855362415, 'learning_rate/qwen_vl_interface': 5.67881979104958e-06, 'learning_rate/action_model': 5.67881979104958e-05, 'epoch': 1.7}) +[2026-06-16 01:56:01,535][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.07714385539293289, 'train/grad_norm_pre_clip': 9.176385879516602, 'timing/data': 0.020850003696978092, 'timing/model': 1.3945585740730166, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 1.7}) +[2026-06-16 01:56:07,246][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.018170876428484917, 'train/grad_norm_pre_clip': 1.90340256690979, 'timing/data': 0.02462615165859461, 'timing/model': 1.447965195402503, 'learning_rate/qwen_vl_interface': 5.672753468508204e-06, 'learning_rate/action_model': 5.672753468508204e-05, 'epoch': 1.7}) +[2026-06-16 01:56:12,934][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.16959169507026672, 'train/grad_norm_pre_clip': 6.246804714202881, 'timing/data': 0.01942184753715992, 'timing/model': 1.445295725017786, 'learning_rate/qwen_vl_interface': 5.6697200459471165e-06, 'learning_rate/action_model': 5.669720045947117e-05, 'epoch': 1.7}) +[2026-06-16 01:56:18,529][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.15967293083667755, 'train/grad_norm_pre_clip': 6.752216815948486, 'timing/data': 0.02025868184864521, 'timing/model': 1.4014631174504757, 'learning_rate/qwen_vl_interface': 5.666686450854973e-06, 'learning_rate/action_model': 5.666686450854973e-05, 'epoch': 1.7}) +[2026-06-16 01:56:24,256][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.2327885776758194, 'train/grad_norm_pre_clip': 8.916723251342773, 'timing/data': 0.021140226162970066, 'timing/model': 1.3826589556410909, 'learning_rate/qwen_vl_interface': 5.663652684478768e-06, 'learning_rate/action_model': 5.6636526844787676e-05, 'epoch': 1.7}) +[2026-06-16 01:56:29,901][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.03572108596563339, 'train/grad_norm_pre_clip': 6.503602504730225, 'timing/data': 0.021164746023714542, 'timing/model': 1.6162901828065515, 'learning_rate/qwen_vl_interface': 5.660618748065569e-06, 'learning_rate/action_model': 5.6606187480655695e-05, 'epoch': 1.7}) +[2026-06-16 01:56:35,332][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.1807374358177185, 'train/grad_norm_pre_clip': 7.44163179397583, 'timing/data': 0.02209263015538454, 'timing/model': 1.4212799221277237, 'learning_rate/qwen_vl_interface': 5.657584642862514e-06, 'learning_rate/action_model': 5.657584642862513e-05, 'epoch': 1.7}) +[2026-06-16 01:56:41,742][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.2580036222934723, 'train/grad_norm_pre_clip': 10.109818458557129, 'timing/data': 0.02082288358360529, 'timing/model': 1.416447532363236, 'learning_rate/qwen_vl_interface': 5.654550370116808e-06, 'learning_rate/action_model': 5.654550370116808e-05, 'epoch': 1.7}) +[2026-06-16 01:56:47,448][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.2629033029079437, 'train/grad_norm_pre_clip': 13.847229957580566, 'timing/data': 0.021574187092483044, 'timing/model': 1.4548631273210049, 'learning_rate/qwen_vl_interface': 5.651515931075726e-06, 'learning_rate/action_model': 5.651515931075726e-05, 'epoch': 1.7}) +[2026-06-16 01:56:53,046][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.27788886427879333, 'train/grad_norm_pre_clip': 8.842745780944824, 'timing/data': 0.021233935840427876, 'timing/model': 1.4498229445889592, 'learning_rate/qwen_vl_interface': 5.6484813269866126e-06, 'learning_rate/action_model': 5.648481326986612e-05, 'epoch': 1.7}) +[2026-06-16 01:56:58,746][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.2835824191570282, 'train/grad_norm_pre_clip': 9.582054138183594, 'timing/data': 0.020565113052725792, 'timing/model': 1.4206391349434853, 'learning_rate/qwen_vl_interface': 5.645446559096875e-06, 'learning_rate/action_model': 5.645446559096875e-05, 'epoch': 1.7}) +[2026-06-16 01:57:04,502][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.11138845235109329, 'train/grad_norm_pre_clip': 3.907111883163452, 'timing/data': 0.02062568347901106, 'timing/model': 1.7008489752188325, 'learning_rate/qwen_vl_interface': 5.6424116286539965e-06, 'learning_rate/action_model': 5.642411628653996e-05, 'epoch': 1.7}) +[2026-06-16 01:57:10,041][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.14025399088859558, 'train/grad_norm_pre_clip': 8.425633430480957, 'timing/data': 0.02229154109954834, 'timing/model': 1.4395411014556885, 'learning_rate/qwen_vl_interface': 5.639376536905517e-06, 'learning_rate/action_model': 5.6393765369055175e-05, 'epoch': 1.7}) +[2026-06-16 01:57:15,847][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.14454765617847443, 'train/grad_norm_pre_clip': 5.314783573150635, 'timing/data': 0.02081933431327343, 'timing/model': 1.4376444956287742, 'learning_rate/qwen_vl_interface': 5.636341285099055e-06, 'learning_rate/action_model': 5.636341285099055e-05, 'epoch': 1.7}) +[2026-06-16 01:57:21,543][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.08502718061208725, 'train/grad_norm_pre_clip': 7.234311103820801, 'timing/data': 0.020698723383247852, 'timing/model': 1.4455698626115918, 'learning_rate/qwen_vl_interface': 5.6333058744822835e-06, 'learning_rate/action_model': 5.633305874482283e-05, 'epoch': 1.7}) +[2026-06-16 01:57:27,245][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.14491046965122223, 'train/grad_norm_pre_clip': 7.291452407836914, 'timing/data': 0.021060374565422535, 'timing/model': 1.4399789785966277, 'learning_rate/qwen_vl_interface': 5.630270306302944e-06, 'learning_rate/action_model': 5.6302703063029436e-05, 'epoch': 1.7}) +[2026-06-16 01:57:32,944][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.13079550862312317, 'train/grad_norm_pre_clip': 6.398278713226318, 'timing/data': 0.020742343738675117, 'timing/model': 1.397626617923379, 'learning_rate/qwen_vl_interface': 5.627234581808848e-06, 'learning_rate/action_model': 5.627234581808848e-05, 'epoch': 1.7}) +[2026-06-16 01:57:38,778][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.05350830405950546, 'train/grad_norm_pre_clip': 5.454690933227539, 'timing/data': 0.020961915142834187, 'timing/model': 1.6870634406805038, 'learning_rate/qwen_vl_interface': 5.624198702247863e-06, 'learning_rate/action_model': 5.624198702247863e-05, 'epoch': 1.7}) +[2026-06-16 01:57:44,370][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.18748921155929565, 'train/grad_norm_pre_clip': 5.787744998931885, 'timing/data': 0.02159149758517742, 'timing/model': 1.469032583758235, 'learning_rate/qwen_vl_interface': 5.621162668867925e-06, 'learning_rate/action_model': 5.621162668867925e-05, 'epoch': 1.7}) +[2026-06-16 01:57:50,271][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.07386872917413712, 'train/grad_norm_pre_clip': 7.724274158477783, 'timing/data': 0.021464007906615734, 'timing/model': 1.4760739859193563, 'learning_rate/qwen_vl_interface': 5.618126482917037e-06, 'learning_rate/action_model': 5.618126482917036e-05, 'epoch': 1.71}) +[2026-06-16 01:57:56,174][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.09821920096874237, 'train/grad_norm_pre_clip': 4.0084452629089355, 'timing/data': 0.020859445445239544, 'timing/model': 1.4765601800754666, 'learning_rate/qwen_vl_interface': 5.6150901456432546e-06, 'learning_rate/action_model': 5.615090145643255e-05, 'epoch': 1.71}) +[2026-06-16 01:58:01,951][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.16350281238555908, 'train/grad_norm_pre_clip': 5.206745624542236, 'timing/data': 0.020611563697457314, 'timing/model': 1.4570750715211034, 'learning_rate/qwen_vl_interface': 5.612053658294705e-06, 'learning_rate/action_model': 5.6120536582947045e-05, 'epoch': 1.71}) +[2026-06-16 01:58:07,674][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.08110814541578293, 'train/grad_norm_pre_clip': 2.574110984802246, 'timing/data': 0.020836873911321163, 'timing/model': 1.4224830670282245, 'learning_rate/qwen_vl_interface': 5.609017022119572e-06, 'learning_rate/action_model': 5.609017022119573e-05, 'epoch': 1.71}) +[2026-06-16 01:58:13,511][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.19768548011779785, 'train/grad_norm_pre_clip': 7.634782791137695, 'timing/data': 0.021280706860125065, 'timing/model': 1.699688675813377, 'learning_rate/qwen_vl_interface': 5.605980238366104e-06, 'learning_rate/action_model': 5.605980238366103e-05, 'epoch': 1.71}) +[2026-06-16 01:58:19,174][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.09763625264167786, 'train/grad_norm_pre_clip': 6.600945472717285, 'timing/data': 0.020807655528187752, 'timing/model': 1.4629372414201498, 'learning_rate/qwen_vl_interface': 5.602943308282606e-06, 'learning_rate/action_model': 5.602943308282605e-05, 'epoch': 1.71}) +[2026-06-16 01:58:25,059][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.055084437131881714, 'train/grad_norm_pre_clip': 2.467188835144043, 'timing/data': 0.020330922678112984, 'timing/model': 1.4280760567635298, 'learning_rate/qwen_vl_interface': 5.599906233117445e-06, 'learning_rate/action_model': 5.5999062331174446e-05, 'epoch': 1.71}) +[2026-06-16 01:58:30,849][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.08457450568675995, 'train/grad_norm_pre_clip': 3.275242805480957, 'timing/data': 0.020514733158051968, 'timing/model': 1.4316004933789372, 'learning_rate/qwen_vl_interface': 5.596869014119048e-06, 'learning_rate/action_model': 5.596869014119048e-05, 'epoch': 1.71}) +[2026-06-16 01:58:36,650][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.1401851624250412, 'train/grad_norm_pre_clip': 7.621533393859863, 'timing/data': 0.02084883488714695, 'timing/model': 1.4489041324704885, 'learning_rate/qwen_vl_interface': 5.593831652535902e-06, 'learning_rate/action_model': 5.593831652535902e-05, 'epoch': 1.71}) +[2026-06-16 01:58:42,454][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.2751365602016449, 'train/grad_norm_pre_clip': 13.976990699768066, 'timing/data': 0.022022120654582977, 'timing/model': 1.3992225266993046, 'learning_rate/qwen_vl_interface': 5.590794149616552e-06, 'learning_rate/action_model': 5.590794149616552e-05, 'epoch': 1.71}) +[2026-06-16 01:58:48,295][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.06693654507398605, 'train/grad_norm_pre_clip': 4.161839962005615, 'timing/data': 0.025122324004769325, 'timing/model': 1.6621430730447173, 'learning_rate/qwen_vl_interface': 5.587756506609601e-06, 'learning_rate/action_model': 5.5877565066096006e-05, 'epoch': 1.71}) +[2026-06-16 01:58:54,126][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.07326892763376236, 'train/grad_norm_pre_clip': 5.090354919433594, 'timing/data': 0.021112185902893543, 'timing/model': 1.627884647808969, 'learning_rate/qwen_vl_interface': 5.584718724763707e-06, 'learning_rate/action_model': 5.584718724763707e-05, 'epoch': 1.71}) +[2026-06-16 01:58:59,657][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.10911335796117783, 'train/grad_norm_pre_clip': 8.18559455871582, 'timing/data': 0.020942426286637783, 'timing/model': 1.4578580660745502, 'learning_rate/qwen_vl_interface': 5.581680805327588e-06, 'learning_rate/action_model': 5.5816808053275885e-05, 'epoch': 1.71}) +[2026-06-16 01:59:05,444][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.05454973876476288, 'train/grad_norm_pre_clip': 3.2126896381378174, 'timing/data': 0.021252166479825974, 'timing/model': 1.4463617857545614, 'learning_rate/qwen_vl_interface': 5.578642749550019e-06, 'learning_rate/action_model': 5.578642749550019e-05, 'epoch': 1.71}) +[2026-06-16 01:59:11,151][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.17440178990364075, 'train/grad_norm_pre_clip': 12.09232234954834, 'timing/data': 0.02150460798293352, 'timing/model': 1.4565335530787706, 'learning_rate/qwen_vl_interface': 5.575604558679831e-06, 'learning_rate/action_model': 5.57560455867983e-05, 'epoch': 1.71}) +[2026-06-16 01:59:16,859][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.008153933100402355, 'train/grad_norm_pre_clip': 0.6756303310394287, 'timing/data': 0.020794885233044624, 'timing/model': 1.403274361975491, 'learning_rate/qwen_vl_interface': 5.5725662339659085e-06, 'learning_rate/action_model': 5.572566233965908e-05, 'epoch': 1.71}) +[2026-06-16 01:59:22,682][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.2606818974018097, 'train/grad_norm_pre_clip': 14.508021354675293, 'timing/data': 0.02021890226751566, 'timing/model': 1.677943211980164, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 1.71}) +[2026-06-16 01:59:28,490][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.21658754348754883, 'train/grad_norm_pre_clip': 10.52614688873291, 'timing/data': 0.02181702945381403, 'timing/model': 1.6827364843338728, 'learning_rate/qwen_vl_interface': 5.566489188002671e-06, 'learning_rate/action_model': 5.56648918800267e-05, 'epoch': 1.71}) +[2026-06-16 01:59:34,049][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.14771269261837006, 'train/grad_norm_pre_clip': 7.130957126617432, 'timing/data': 0.021170356310904026, 'timing/model': 1.4434140278026462, 'learning_rate/qwen_vl_interface': 5.563450469251403e-06, 'learning_rate/action_model': 5.563450469251402e-05, 'epoch': 1.71}) +[2026-06-16 01:59:39,877][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.1780375987291336, 'train/grad_norm_pre_clip': 6.7988715171813965, 'timing/data': 0.027654975652694702, 'timing/model': 1.4334549633786082, 'learning_rate/qwen_vl_interface': 5.560411621652488e-06, 'learning_rate/action_model': 5.560411621652488e-05, 'epoch': 1.71}) +[2026-06-16 01:59:45,869][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.19437968730926514, 'train/grad_norm_pre_clip': 10.0875883102417, 'timing/data': 0.02070288360118866, 'timing/model': 1.4549501417204738, 'learning_rate/qwen_vl_interface': 5.557372646455081e-06, 'learning_rate/action_model': 5.55737264645508e-05, 'epoch': 1.71}) +[2026-06-16 01:59:52,066][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.1195763498544693, 'train/grad_norm_pre_clip': 6.552597999572754, 'timing/data': 0.028330949135124683, 'timing/model': 1.460703489370644, 'learning_rate/qwen_vl_interface': 5.554333544908388e-06, 'learning_rate/action_model': 5.5543335449083875e-05, 'epoch': 1.72}) +[2026-06-16 01:59:58,210][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.13064761459827423, 'train/grad_norm_pre_clip': 5.97705602645874, 'timing/data': 0.027121443301439285, 'timing/model': 1.9767101192846894, 'learning_rate/qwen_vl_interface': 5.551294318261671e-06, 'learning_rate/action_model': 5.5512943182616705e-05, 'epoch': 1.72}) +[2026-06-16 02:00:03,867][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.04184870421886444, 'train/grad_norm_pre_clip': 4.367674350738525, 'timing/data': 0.029887796379625797, 'timing/model': 1.4397945264354348, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 1.72}) +[2026-06-16 02:00:09,973][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.27547070384025574, 'train/grad_norm_pre_clip': 11.97080135345459, 'timing/data': 0.027644005604088306, 'timing/model': 1.4298369903117418, 'learning_rate/qwen_vl_interface': 5.545215494665458e-06, 'learning_rate/action_model': 5.545215494665458e-05, 'epoch': 1.72}) +[2026-06-16 02:00:16,180][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.1488456279039383, 'train/grad_norm_pre_clip': 9.614577293395996, 'timing/data': 0.0276408763602376, 'timing/model': 1.4427814008668065, 'learning_rate/qwen_vl_interface': 5.5421759002147376e-06, 'learning_rate/action_model': 5.542175900214738e-05, 'epoch': 1.72}) +[2026-06-16 02:00:22,373][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.060535334050655365, 'train/grad_norm_pre_clip': 5.346322536468506, 'timing/data': 0.02923645358532667, 'timing/model': 1.4572626473382115, 'learning_rate/qwen_vl_interface': 5.53913618566154e-06, 'learning_rate/action_model': 5.53913618566154e-05, 'epoch': 1.72}) +[2026-06-16 02:00:28,570][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.05285804718732834, 'train/grad_norm_pre_clip': 3.7463128566741943, 'timing/data': 0.02861128095537424, 'timing/model': 1.423498804681003, 'learning_rate/qwen_vl_interface': 5.536096352255381e-06, 'learning_rate/action_model': 5.536096352255381e-05, 'epoch': 1.72}) +[2026-06-16 02:00:34,777][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.0601634755730629, 'train/grad_norm_pre_clip': 4.892794609069824, 'timing/data': 0.028604059480130672, 'timing/model': 1.9469822607934475, 'learning_rate/qwen_vl_interface': 5.533056401245817e-06, 'learning_rate/action_model': 5.5330564012458164e-05, 'epoch': 1.72}) +[2026-06-16 02:00:40,472][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.04753583297133446, 'train/grad_norm_pre_clip': 3.538036823272705, 'timing/data': 0.02794014848768711, 'timing/model': 1.4388906760141253, 'learning_rate/qwen_vl_interface': 5.530016333882456e-06, 'learning_rate/action_model': 5.5300163338824565e-05, 'epoch': 1.72}) +[2026-06-16 02:00:46,676][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.11777234077453613, 'train/grad_norm_pre_clip': 7.580228328704834, 'timing/data': 0.02849102858453989, 'timing/model': 1.4458624497056007, 'learning_rate/qwen_vl_interface': 5.52697615141496e-06, 'learning_rate/action_model': 5.52697615141496e-05, 'epoch': 1.72}) +[2026-06-16 02:00:52,865][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.07974962145090103, 'train/grad_norm_pre_clip': 7.1820454597473145, 'timing/data': 0.027395415119826794, 'timing/model': 1.4456729693338275, 'learning_rate/qwen_vl_interface': 5.52393585509303e-06, 'learning_rate/action_model': 5.52393585509303e-05, 'epoch': 1.72}) +[2026-06-16 02:00:58,880][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.0936923399567604, 'train/grad_norm_pre_clip': 4.587949275970459, 'timing/data': 0.028710021637380123, 'timing/model': 1.4199962932616472, 'learning_rate/qwen_vl_interface': 5.520895446166418e-06, 'learning_rate/action_model': 5.5208954461664175e-05, 'epoch': 1.72}) +[2026-06-16 02:01:04,777][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.12317613512277603, 'train/grad_norm_pre_clip': 7.7281951904296875, 'timing/data': 0.021220906637609005, 'timing/model': 1.4041566718369722, 'learning_rate/qwen_vl_interface': 5.517854925884919e-06, 'learning_rate/action_model': 5.517854925884919e-05, 'epoch': 1.72}) +[2026-06-16 02:01:11,014][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.005177982151508331, 'train/grad_norm_pre_clip': 0.6712895631790161, 'timing/data': 0.02419632114470005, 'timing/model': 2.010977618396282, 'learning_rate/qwen_vl_interface': 5.514814295498379e-06, 'learning_rate/action_model': 5.514814295498378e-05, 'epoch': 1.72}) +[2026-06-16 02:01:16,679][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.21429474651813507, 'train/grad_norm_pre_clip': 20.845840454101562, 'timing/data': 0.027894767001271248, 'timing/model': 1.4499418726190925, 'learning_rate/qwen_vl_interface': 5.5117735562566846e-06, 'learning_rate/action_model': 5.5117735562566844e-05, 'epoch': 1.72}) +[2026-06-16 02:01:22,875][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.015239279717206955, 'train/grad_norm_pre_clip': 1.163072943687439, 'timing/data': 0.028704211115837097, 'timing/model': 1.4438404152169824, 'learning_rate/qwen_vl_interface': 5.5087327094097685e-06, 'learning_rate/action_model': 5.508732709409769e-05, 'epoch': 1.72}) +[2026-06-16 02:01:29,077][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.14447486400604248, 'train/grad_norm_pre_clip': 14.583207130432129, 'timing/data': 0.0265255905687809, 'timing/model': 1.4314020592719316, 'learning_rate/qwen_vl_interface': 5.505691756207609e-06, 'learning_rate/action_model': 5.505691756207609e-05, 'epoch': 1.72}) +[2026-06-16 02:01:35,247][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.14806842803955078, 'train/grad_norm_pre_clip': 14.501718521118164, 'timing/data': 0.020436412654817104, 'timing/model': 1.4603616315871477, 'learning_rate/qwen_vl_interface': 5.502650697900228e-06, 'learning_rate/action_model': 5.502650697900228e-05, 'epoch': 1.72}) +[2026-06-16 02:01:41,499][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.032771166414022446, 'train/grad_norm_pre_clip': 3.755894660949707, 'timing/data': 0.02116460632532835, 'timing/model': 2.096052533015609, 'learning_rate/qwen_vl_interface': 5.499609535737688e-06, 'learning_rate/action_model': 5.4996095357376875e-05, 'epoch': 1.72}) +[2026-06-16 02:01:47,171][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.013258413411676884, 'train/grad_norm_pre_clip': 1.7895479202270508, 'timing/data': 0.02836304996162653, 'timing/model': 1.4244208382442594, 'learning_rate/qwen_vl_interface': 5.496568270970096e-06, 'learning_rate/action_model': 5.496568270970096e-05, 'epoch': 1.72}) +[2026-06-16 02:01:53,366][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.22335916757583618, 'train/grad_norm_pre_clip': 15.242898941040039, 'timing/data': 0.029013032093644142, 'timing/model': 1.4364467049017549, 'learning_rate/qwen_vl_interface': 5.4935269048476015e-06, 'learning_rate/action_model': 5.493526904847601e-05, 'epoch': 1.72}) +[2026-06-16 02:01:59,471][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.1281120777130127, 'train/grad_norm_pre_clip': 15.008126258850098, 'timing/data': 0.028055908158421516, 'timing/model': 1.434051165357232, 'learning_rate/qwen_vl_interface': 5.490485438620397e-06, 'learning_rate/action_model': 5.4904854386203965e-05, 'epoch': 1.72}) +[2026-06-16 02:02:05,670][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.11835956573486328, 'train/grad_norm_pre_clip': 7.002044200897217, 'timing/data': 0.029682236723601818, 'timing/model': 1.4450435657054186, 'learning_rate/qwen_vl_interface': 5.487443873538712e-06, 'learning_rate/action_model': 5.4874438735387115e-05, 'epoch': 1.73}) +[2026-06-16 02:02:11,876][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.17158979177474976, 'train/grad_norm_pre_clip': 9.226949691772461, 'timing/data': 0.028613749891519547, 'timing/model': 1.4815155128017068, 'learning_rate/qwen_vl_interface': 5.484402210852819e-06, 'learning_rate/action_model': 5.484402210852818e-05, 'epoch': 1.73}) +[2026-06-16 02:02:18,091][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.38533151149749756, 'train/grad_norm_pre_clip': 22.020423889160156, 'timing/data': 0.030014577321708202, 'timing/model': 1.953765969723463, 'learning_rate/qwen_vl_interface': 5.48136045181303e-06, 'learning_rate/action_model': 5.48136045181303e-05, 'epoch': 1.73}) +[2026-06-16 02:02:23,773][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.0511518232524395, 'train/grad_norm_pre_clip': 3.2952628135681152, 'timing/data': 0.029966285452246666, 'timing/model': 1.433482315391302, 'learning_rate/qwen_vl_interface': 5.478318597669699e-06, 'learning_rate/action_model': 5.4783185976696995e-05, 'epoch': 1.73}) +[2026-06-16 02:02:29,977][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.3041858375072479, 'train/grad_norm_pre_clip': 11.983589172363281, 'timing/data': 0.02834800910204649, 'timing/model': 1.4461371144279838, 'learning_rate/qwen_vl_interface': 5.475276649673218e-06, 'learning_rate/action_model': 5.4752766496732174e-05, 'epoch': 1.73}) +[2026-06-16 02:02:36,187][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.09547082334756851, 'train/grad_norm_pre_clip': 5.707780838012695, 'timing/data': 0.02637709118425846, 'timing/model': 1.4556305883452296, 'learning_rate/qwen_vl_interface': 5.472234609074014e-06, 'learning_rate/action_model': 5.472234609074014e-05, 'epoch': 1.73}) +[2026-06-16 02:02:42,169][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.1877286434173584, 'train/grad_norm_pre_clip': 15.671311378479004, 'timing/data': 0.02853046078234911, 'timing/model': 1.4437693832442164, 'learning_rate/qwen_vl_interface': 5.46919247712256e-06, 'learning_rate/action_model': 5.46919247712256e-05, 'epoch': 1.73}) +[2026-06-16 02:02:48,375][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.08409174531698227, 'train/grad_norm_pre_clip': 7.367653846740723, 'timing/data': 0.028784381225705147, 'timing/model': 1.4444170678034425, 'learning_rate/qwen_vl_interface': 5.466150255069356e-06, 'learning_rate/action_model': 5.4661502550693555e-05, 'epoch': 1.73}) +[2026-06-16 02:02:54,577][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.055723514407873154, 'train/grad_norm_pre_clip': 5.149115085601807, 'timing/data': 0.028355279006063938, 'timing/model': 1.9460154315456748, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 1.73}) +[2026-06-16 02:03:00,276][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.14927276968955994, 'train/grad_norm_pre_clip': 9.33384895324707, 'timing/data': 0.028114297427237034, 'timing/model': 1.4532822901383042, 'learning_rate/qwen_vl_interface': 5.46006554565991e-06, 'learning_rate/action_model': 5.4600655456599105e-05, 'epoch': 1.73}) +[2026-06-16 02:03:06,468][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.10717107355594635, 'train/grad_norm_pre_clip': 11.073387145996094, 'timing/data': 0.028545590117573738, 'timing/model': 1.428639167919755, 'learning_rate/qwen_vl_interface': 5.457023060804865e-06, 'learning_rate/action_model': 5.457023060804864e-05, 'epoch': 1.73}) +[2026-06-16 02:03:12,681][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.04905179515480995, 'train/grad_norm_pre_clip': 5.097002029418945, 'timing/data': 0.027553776279091835, 'timing/model': 1.4578309915959835, 'learning_rate/qwen_vl_interface': 5.453980490850459e-06, 'learning_rate/action_model': 5.453980490850459e-05, 'epoch': 1.73}) +[2026-06-16 02:03:18,786][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.11758410930633545, 'train/grad_norm_pre_clip': 12.435944557189941, 'timing/data': 0.028972221538424492, 'timing/model': 1.418926265090704, 'learning_rate/qwen_vl_interface': 5.4509378370473785e-06, 'learning_rate/action_model': 5.450937837047378e-05, 'epoch': 1.73}) +[2026-06-16 02:03:24,787][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.12563003599643707, 'train/grad_norm_pre_clip': 12.599225997924805, 'timing/data': 0.02945094369351864, 'timing/model': 1.4247567625716329, 'learning_rate/qwen_vl_interface': 5.4478951006463424e-06, 'learning_rate/action_model': 5.447895100646342e-05, 'epoch': 1.73}) +[2026-06-16 02:03:30,923][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.2764507532119751, 'train/grad_norm_pre_clip': 19.247976303100586, 'timing/data': 0.02638322953134775, 'timing/model': 1.8852427378296852, 'learning_rate/qwen_vl_interface': 5.444852282898106e-06, 'learning_rate/action_model': 5.444852282898106e-05, 'epoch': 1.73}) +[2026-06-16 02:03:36,570][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.37268829345703125, 'train/grad_norm_pre_clip': 9.520034790039062, 'timing/data': 0.029674566350877285, 'timing/model': 1.4418690102174878, 'learning_rate/qwen_vl_interface': 5.441809385053456e-06, 'learning_rate/action_model': 5.4418093850534555e-05, 'epoch': 1.73}) +[2026-06-16 02:03:42,765][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.10680968314409256, 'train/grad_norm_pre_clip': 5.102232933044434, 'timing/data': 0.0279398076236248, 'timing/model': 1.4507178319618106, 'learning_rate/qwen_vl_interface': 5.438766408363214e-06, 'learning_rate/action_model': 5.438766408363214e-05, 'epoch': 1.73}) +[2026-06-16 02:03:48,765][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.0067373705096542835, 'train/grad_norm_pre_clip': 1.4257099628448486, 'timing/data': 0.028250619769096375, 'timing/model': 1.4238151405006647, 'learning_rate/qwen_vl_interface': 5.43572335407823e-06, 'learning_rate/action_model': 5.43572335407823e-05, 'epoch': 1.73}) +[2026-06-16 02:03:54,864][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.0870567113161087, 'train/grad_norm_pre_clip': 4.575769901275635, 'timing/data': 0.027622765861451626, 'timing/model': 1.4387455191463232, 'learning_rate/qwen_vl_interface': 5.432680223449392e-06, 'learning_rate/action_model': 5.4326802234493924e-05, 'epoch': 1.73}) +[2026-06-16 02:04:00,969][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.0656980648636818, 'train/grad_norm_pre_clip': 6.978792667388916, 'timing/data': 0.028877072036266327, 'timing/model': 1.43405138887465, 'learning_rate/qwen_vl_interface': 5.429637017727614e-06, 'learning_rate/action_model': 5.4296370177276144e-05, 'epoch': 1.73}) +[2026-06-16 02:04:07,139][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.08919577300548553, 'train/grad_norm_pre_clip': 5.563625812530518, 'timing/data': 0.029352174140512943, 'timing/model': 1.9179425714537501, 'learning_rate/qwen_vl_interface': 5.426593738163845e-06, 'learning_rate/action_model': 5.4265937381638446e-05, 'epoch': 1.73}) +[2026-06-16 02:04:12,778][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.033192988485097885, 'train/grad_norm_pre_clip': 2.681912660598755, 'timing/data': 0.027108823880553246, 'timing/model': 1.4579595793038607, 'learning_rate/qwen_vl_interface': 5.42355038600906e-06, 'learning_rate/action_model': 5.42355038600906e-05, 'epoch': 1.73}) +[2026-06-16 02:04:18,964][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.09776327013969421, 'train/grad_norm_pre_clip': 6.63137149810791, 'timing/data': 0.027491996064782143, 'timing/model': 1.4417618354782462, 'learning_rate/qwen_vl_interface': 5.420506962514267e-06, 'learning_rate/action_model': 5.420506962514267e-05, 'epoch': 1.74}) +[2026-06-16 02:04:25,077][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.10211396217346191, 'train/grad_norm_pre_clip': 4.922343730926514, 'timing/data': 0.028424550779163837, 'timing/model': 1.4516459610313177, 'learning_rate/qwen_vl_interface': 5.4174634689305015e-06, 'learning_rate/action_model': 5.417463468930501e-05, 'epoch': 1.74}) +[2026-06-16 02:04:31,270][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.21428720653057098, 'train/grad_norm_pre_clip': 13.491897583007812, 'timing/data': 0.029466583393514156, 'timing/model': 1.4424930904060602, 'learning_rate/qwen_vl_interface': 5.41441990650883e-06, 'learning_rate/action_model': 5.4144199065088295e-05, 'epoch': 1.74}) +[2026-06-16 02:04:37,566][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.25344619154930115, 'train/grad_norm_pre_clip': 8.903510093688965, 'timing/data': 0.029024872928857803, 'timing/model': 1.473663062788546, 'learning_rate/qwen_vl_interface': 5.411376276500345e-06, 'learning_rate/action_model': 5.411376276500345e-05, 'epoch': 1.74}) +[2026-06-16 02:04:43,750][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.06904292851686478, 'train/grad_norm_pre_clip': 4.753087997436523, 'timing/data': 0.02760860789567232, 'timing/model': 1.9217791743576527, 'learning_rate/qwen_vl_interface': 5.408332580156168e-06, 'learning_rate/action_model': 5.408332580156168e-05, 'epoch': 1.74}) +[2026-06-16 02:04:49,373][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.07935312390327454, 'train/grad_norm_pre_clip': 6.053903579711914, 'timing/data': 0.028776082210242748, 'timing/model': 1.4478721963241696, 'learning_rate/qwen_vl_interface': 5.40528881872745e-06, 'learning_rate/action_model': 5.40528881872745e-05, 'epoch': 1.74}) +[2026-06-16 02:15:37,184][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.1219831332564354, 'train/grad_norm_pre_clip': 12.067655563354492, 'eval/loss': 0.06460579484701157, 'eval/flap_precision': 0.8207885304659498, 'eval/flap_recall': 0.8327272727272728, 'eval/flap_f1': 0.8267148014440433, 'eval/flap_rate_model': 0.0775, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.8267148014440433, 'timing/data': 0.027933327481150627, 'timing/model': 1.4517021542415023, 'learning_rate/qwen_vl_interface': 5.402244993465364e-06, 'learning_rate/action_model': 5.402244993465364e-05, 'epoch': 1.74}) +[2026-06-16 02:15:37,185][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3800_state +[2026-06-16 02:15:37,192][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:15:45,215][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3800_state/model.safetensors +[2026-06-16 02:15:56,448][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3800_state/optimizer.bin +[2026-06-16 02:15:56,449][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3800_state/sampler.bin +[2026-06-16 02:15:56,449][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3800_state/sampler_1.bin +[2026-06-16 02:15:56,451][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3800_state/random_states_0.pkl +[2026-06-16 02:15:56,614][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3800 +[2026-06-16 02:15:56,615][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:15:56,630][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:16:02,349][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.071555957198143, 'train/grad_norm_pre_clip': 3.5069332122802734, 'timing/data': 0.03191652614623308, 'timing/model': 1.4471383159980178, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 1.74}) +[2026-06-16 02:16:08,326][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.06890979409217834, 'train/grad_norm_pre_clip': 7.282440662384033, 'timing/data': 0.02579967863857746, 'timing/model': 1.4247798630967736, 'learning_rate/qwen_vl_interface': 5.396157156445918e-06, 'learning_rate/action_model': 5.396157156445918e-05, 'epoch': 1.74}) +[2026-06-16 02:16:14,135][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.1516801416873932, 'train/grad_norm_pre_clip': 11.155583381652832, 'timing/data': 0.022794305346906185, 'timing/model': 1.4277791669592261, 'learning_rate/qwen_vl_interface': 5.39311314719104e-06, 'learning_rate/action_model': 5.39311314719104e-05, 'epoch': 1.74}) +[2026-06-16 02:16:19,924][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.08943727612495422, 'train/grad_norm_pre_clip': 6.738736629486084, 'timing/data': 0.022670763544738293, 'timing/model': 1.4193359278142452, 'learning_rate/qwen_vl_interface': 5.390069079107755e-06, 'learning_rate/action_model': 5.390069079107755e-05, 'epoch': 1.74}) +[2026-06-16 02:16:25,736][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.012519119307398796, 'train/grad_norm_pre_clip': 1.1977101564407349, 'timing/data': 0.03580863494426012, 'timing/model': 1.4100311053916812, 'learning_rate/qwen_vl_interface': 5.387024953447361e-06, 'learning_rate/action_model': 5.387024953447361e-05, 'epoch': 1.74}) +[2026-06-16 02:16:31,558][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.10510941594839096, 'train/grad_norm_pre_clip': 6.5464301109313965, 'timing/data': 0.026199021376669407, 'timing/model': 1.7470449255779386, 'learning_rate/qwen_vl_interface': 5.383980771461184e-06, 'learning_rate/action_model': 5.3839807714611846e-05, 'epoch': 1.74}) +[2026-06-16 02:16:37,432][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.051541730761528015, 'train/grad_norm_pre_clip': 5.033579349517822, 'timing/data': 0.02745947614312172, 'timing/model': 1.7177586117759347, 'learning_rate/qwen_vl_interface': 5.380936534400572e-06, 'learning_rate/action_model': 5.3809365344005726e-05, 'epoch': 1.74}) +[2026-06-16 02:16:43,037][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.11954230815172195, 'train/grad_norm_pre_clip': 7.920879364013672, 'timing/data': 0.021526599302887917, 'timing/model': 1.4469012655317783, 'learning_rate/qwen_vl_interface': 5.377892243516899e-06, 'learning_rate/action_model': 5.3778922435168985e-05, 'epoch': 1.74}) +[2026-06-16 02:16:48,833][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.2639848589897156, 'train/grad_norm_pre_clip': 16.80664825439453, 'timing/data': 0.021633719094097614, 'timing/model': 1.424788374453783, 'learning_rate/qwen_vl_interface': 5.3748479000615525e-06, 'learning_rate/action_model': 5.374847900061553e-05, 'epoch': 1.74}) +[2026-06-16 02:16:54,634][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.3191073536872864, 'train/grad_norm_pre_clip': 11.456165313720703, 'timing/data': 0.020173811353743076, 'timing/model': 1.4482094626873732, 'learning_rate/qwen_vl_interface': 5.371803505285954e-06, 'learning_rate/action_model': 5.371803505285954e-05, 'epoch': 1.74}) +[2026-06-16 02:17:00,439][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.07145034521818161, 'train/grad_norm_pre_clip': 4.289382457733154, 'timing/data': 0.0210686968639493, 'timing/model': 1.4461624417454004, 'learning_rate/qwen_vl_interface': 5.368759060441534e-06, 'learning_rate/action_model': 5.368759060441534e-05, 'epoch': 1.74}) +[2026-06-16 02:17:06,156][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.2526184320449829, 'train/grad_norm_pre_clip': 15.048539161682129, 'timing/data': 0.02029735315591097, 'timing/model': 1.3865522481501102, 'learning_rate/qwen_vl_interface': 5.365714566779749e-06, 'learning_rate/action_model': 5.3657145667797484e-05, 'epoch': 1.74}) +[2026-06-16 02:17:11,906][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.04666878655552864, 'train/grad_norm_pre_clip': 3.9892172813415527, 'timing/data': 0.02173035964369774, 'timing/model': 1.7100603971630335, 'learning_rate/qwen_vl_interface': 5.362670025552077e-06, 'learning_rate/action_model': 5.362670025552077e-05, 'epoch': 1.74}) +[2026-06-16 02:17:17,439][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.09825388342142105, 'train/grad_norm_pre_clip': 7.115735054016113, 'timing/data': 0.021090487018227577, 'timing/model': 1.4327119505032897, 'learning_rate/qwen_vl_interface': 5.359625438010013e-06, 'learning_rate/action_model': 5.359625438010013e-05, 'epoch': 1.74}) +[2026-06-16 02:17:23,246][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.26276299357414246, 'train/grad_norm_pre_clip': 9.453920364379883, 'timing/data': 0.0247432142496109, 'timing/model': 1.4344451185315847, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 1.74}) +[2026-06-16 02:17:29,044][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.09561296552419662, 'train/grad_norm_pre_clip': 8.487504959106445, 'timing/data': 0.0204949751496315, 'timing/model': 1.4476123889908195, 'learning_rate/qwen_vl_interface': 5.353536128988791e-06, 'learning_rate/action_model': 5.353536128988791e-05, 'epoch': 1.75}) +[2026-06-16 02:17:34,845][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.0548228919506073, 'train/grad_norm_pre_clip': 6.494702339172363, 'timing/data': 0.020709695294499397, 'timing/model': 1.4387036487460136, 'learning_rate/qwen_vl_interface': 5.350491410012717e-06, 'learning_rate/action_model': 5.350491410012717e-05, 'epoch': 1.75}) +[2026-06-16 02:17:40,632][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.08186569809913635, 'train/grad_norm_pre_clip': 6.894437313079834, 'timing/data': 0.02236015349626541, 'timing/model': 1.7237887494266033, 'learning_rate/qwen_vl_interface': 5.347446649728419e-06, 'learning_rate/action_model': 5.347446649728418e-05, 'epoch': 1.75}) +[2026-06-16 02:17:46,142][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.2186134308576584, 'train/grad_norm_pre_clip': 15.56097412109375, 'timing/data': 0.022244252264499664, 'timing/model': 1.4482065122574568, 'learning_rate/qwen_vl_interface': 5.344401849387484e-06, 'learning_rate/action_model': 5.344401849387484e-05, 'epoch': 1.75}) +[2026-06-16 02:17:51,949][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.1443011611700058, 'train/grad_norm_pre_clip': 10.889208793640137, 'timing/data': 0.019950672052800655, 'timing/model': 1.4443075750023127, 'learning_rate/qwen_vl_interface': 5.3413570102415144e-06, 'learning_rate/action_model': 5.3413570102415144e-05, 'epoch': 1.75}) +[2026-06-16 02:17:57,767][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.18424206972122192, 'train/grad_norm_pre_clip': 8.62342643737793, 'timing/data': 0.027976298704743385, 'timing/model': 1.4506833851337433, 'learning_rate/qwen_vl_interface': 5.338312133542131e-06, 'learning_rate/action_model': 5.3383121335421305e-05, 'epoch': 1.75}) +[2026-06-16 02:18:03,563][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.13608159124851227, 'train/grad_norm_pre_clip': 8.786419868469238, 'timing/data': 0.02294408529996872, 'timing/model': 1.4627607697620988, 'learning_rate/qwen_vl_interface': 5.335267220540963e-06, 'learning_rate/action_model': 5.335267220540963e-05, 'epoch': 1.75}) +[2026-06-16 02:18:09,356][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.13715605437755585, 'train/grad_norm_pre_clip': 7.78390645980835, 'timing/data': 0.02073358464986086, 'timing/model': 1.3914429424330592, 'learning_rate/qwen_vl_interface': 5.332222272489662e-06, 'learning_rate/action_model': 5.332222272489662e-05, 'epoch': 1.75}) +[2026-06-16 02:18:15,238][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.06513835489749908, 'train/grad_norm_pre_clip': 4.396753787994385, 'timing/data': 0.028484441339969635, 'timing/model': 1.8098254781216383, 'learning_rate/qwen_vl_interface': 5.3291772906398895e-06, 'learning_rate/action_model': 5.32917729063989e-05, 'epoch': 1.75}) +[2026-06-16 02:18:20,869][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.1574234962463379, 'train/grad_norm_pre_clip': 9.801706314086914, 'timing/data': 0.028293251059949398, 'timing/model': 1.458482620306313, 'learning_rate/qwen_vl_interface': 5.326132276243327e-06, 'learning_rate/action_model': 5.3261322762433266e-05, 'epoch': 1.75}) +[2026-06-16 02:18:26,968][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.1076047271490097, 'train/grad_norm_pre_clip': 14.906248092651367, 'timing/data': 0.028292019851505756, 'timing/model': 1.4463415648788214, 'learning_rate/qwen_vl_interface': 5.323087230551659e-06, 'learning_rate/action_model': 5.323087230551659e-05, 'epoch': 1.75}) +[2026-06-16 02:18:33,073][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.19847571849822998, 'train/grad_norm_pre_clip': 6.217111110687256, 'timing/data': 0.027688026428222656, 'timing/model': 1.4503281526267529, 'learning_rate/qwen_vl_interface': 5.3200421548165916e-06, 'learning_rate/action_model': 5.3200421548165914e-05, 'epoch': 1.75}) +[2026-06-16 02:18:39,073][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.1808728724718094, 'train/grad_norm_pre_clip': 7.241281032562256, 'timing/data': 0.027974647469818592, 'timing/model': 1.4455622220411897, 'learning_rate/qwen_vl_interface': 5.3169970502898405e-06, 'learning_rate/action_model': 5.31699705028984e-05, 'epoch': 1.75}) +[2026-06-16 02:18:45,191][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.04919518530368805, 'train/grad_norm_pre_clip': 3.208627462387085, 'timing/data': 0.027628947980701923, 'timing/model': 1.4647078085690737, 'learning_rate/qwen_vl_interface': 5.31395191822313e-06, 'learning_rate/action_model': 5.31395191822313e-05, 'epoch': 1.75}) +[2026-06-16 02:18:51,212][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.13131818175315857, 'train/grad_norm_pre_clip': 5.058155059814453, 'timing/data': 0.0306386798620224, 'timing/model': 1.7973341913893819, 'learning_rate/qwen_vl_interface': 5.310906759868205e-06, 'learning_rate/action_model': 5.310906759868205e-05, 'epoch': 1.75}) +[2026-06-16 02:18:56,894][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.2596989870071411, 'train/grad_norm_pre_clip': 10.319039344787598, 'timing/data': 0.02642565220594406, 'timing/model': 1.4462171951308846, 'learning_rate/qwen_vl_interface': 5.30786157647681e-06, 'learning_rate/action_model': 5.30786157647681e-05, 'epoch': 1.75}) +[2026-06-16 02:19:02,971][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.20925208926200867, 'train/grad_norm_pre_clip': 8.661056518554688, 'timing/data': 0.02919465396553278, 'timing/model': 1.4268406461924314, 'learning_rate/qwen_vl_interface': 5.304816369300709e-06, 'learning_rate/action_model': 5.304816369300709e-05, 'epoch': 1.75}) +[2026-06-16 02:19:09,080][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.30038508772850037, 'train/grad_norm_pre_clip': 18.397485733032227, 'timing/data': 0.03399358689785004, 'timing/model': 1.4444780172780156, 'learning_rate/qwen_vl_interface': 5.301771139591666e-06, 'learning_rate/action_model': 5.301771139591666e-05, 'epoch': 1.75}) +[2026-06-16 02:19:15,171][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.2297048717737198, 'train/grad_norm_pre_clip': 18.985445022583008, 'timing/data': 0.030383769422769547, 'timing/model': 1.4248600769788027, 'learning_rate/qwen_vl_interface': 5.298725888601465e-06, 'learning_rate/action_model': 5.298725888601465e-05, 'epoch': 1.75}) +[2026-06-16 02:19:21,168][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.11308734863996506, 'train/grad_norm_pre_clip': 6.677338600158691, 'timing/data': 0.028353259898722172, 'timing/model': 1.4645908800885081, 'learning_rate/qwen_vl_interface': 5.295680617581891e-06, 'learning_rate/action_model': 5.2956806175818904e-05, 'epoch': 1.75}) +[2026-06-16 02:19:27,250][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.07182791829109192, 'train/grad_norm_pre_clip': 4.449311256408691, 'timing/data': 0.02912113443017006, 'timing/model': 1.8113795053213835, 'learning_rate/qwen_vl_interface': 5.2926353277847395e-06, 'learning_rate/action_model': 5.292635327784739e-05, 'epoch': 1.75}) +[2026-06-16 02:19:32,874][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.23931707441806793, 'train/grad_norm_pre_clip': 8.916318893432617, 'timing/data': 0.02888499293476343, 'timing/model': 1.4241790631785989, 'learning_rate/qwen_vl_interface': 5.2895900204618175e-06, 'learning_rate/action_model': 5.289590020461818e-05, 'epoch': 1.75}) +[2026-06-16 02:19:38,977][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.21684430539608002, 'train/grad_norm_pre_clip': 8.318916320800781, 'timing/data': 0.027977469377219677, 'timing/model': 1.4484321661293507, 'learning_rate/qwen_vl_interface': 5.286544696864933e-06, 'learning_rate/action_model': 5.2865446968649324e-05, 'epoch': 1.76}) +[2026-06-16 02:19:45,075][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.07220202684402466, 'train/grad_norm_pre_clip': 4.590034484863281, 'timing/data': 0.02794295735657215, 'timing/model': 1.4495443906635046, 'learning_rate/qwen_vl_interface': 5.283499358245904e-06, 'learning_rate/action_model': 5.2834993582459034e-05, 'epoch': 1.76}) +[2026-06-16 02:19:51,065][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.14922402799129486, 'train/grad_norm_pre_clip': 8.384355545043945, 'timing/data': 0.028453431092202663, 'timing/model': 1.435157384723425, 'learning_rate/qwen_vl_interface': 5.280454005856553e-06, 'learning_rate/action_model': 5.2804540058565534e-05, 'epoch': 1.76}) +[2026-06-16 02:19:57,059][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.0933348760008812, 'train/grad_norm_pre_clip': 5.147141933441162, 'timing/data': 0.021367287263274193, 'timing/model': 1.436581720598042, 'learning_rate/qwen_vl_interface': 5.277408640948711e-06, 'learning_rate/action_model': 5.277408640948711e-05, 'epoch': 1.76}) +[2026-06-16 02:20:02,946][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.2567141056060791, 'train/grad_norm_pre_clip': 11.352103233337402, 'timing/data': 0.020652865059673786, 'timing/model': 1.7352450359612703, 'learning_rate/qwen_vl_interface': 5.274363264774213e-06, 'learning_rate/action_model': 5.274363264774212e-05, 'epoch': 1.76}) +[2026-06-16 02:20:08,915][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.13679863512516022, 'train/grad_norm_pre_clip': 6.835881233215332, 'timing/data': 0.021212616935372353, 'timing/model': 1.450451116077602, 'learning_rate/qwen_vl_interface': 5.271317878584897e-06, 'learning_rate/action_model': 5.2713178785848974e-05, 'epoch': 1.76}) +[2026-06-16 02:20:14,744][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.15090984106063843, 'train/grad_norm_pre_clip': 6.1266961097717285, 'timing/data': 0.02271246537566185, 'timing/model': 1.4341739099472761, 'learning_rate/qwen_vl_interface': 5.268272483632606e-06, 'learning_rate/action_model': 5.268272483632606e-05, 'epoch': 1.76}) +[2026-06-16 02:20:20,465][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.08919726312160492, 'train/grad_norm_pre_clip': 5.286257266998291, 'timing/data': 0.02883597370237112, 'timing/model': 1.42539875023067, 'learning_rate/qwen_vl_interface': 5.265227081169187e-06, 'learning_rate/action_model': 5.265227081169186e-05, 'epoch': 1.76}) +[2026-06-16 02:20:26,268][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.17755064368247986, 'train/grad_norm_pre_clip': 12.128938674926758, 'timing/data': 0.02571877744048834, 'timing/model': 1.451171119697392, 'learning_rate/qwen_vl_interface': 5.2621816724464895e-06, 'learning_rate/action_model': 5.262181672446489e-05, 'epoch': 1.76}) +[2026-06-16 02:20:32,155][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.1960434466600418, 'train/grad_norm_pre_clip': 7.227946758270264, 'timing/data': 0.021007627248764038, 'timing/model': 1.4218645645305514, 'learning_rate/qwen_vl_interface': 5.259136258716367e-06, 'learning_rate/action_model': 5.2591362587163675e-05, 'epoch': 1.76}) +[2026-06-16 02:20:37,936][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.08908285945653915, 'train/grad_norm_pre_clip': 5.496038913726807, 'timing/data': 0.0209081768989563, 'timing/model': 1.7135354466736317, 'learning_rate/qwen_vl_interface': 5.256090841230675e-06, 'learning_rate/action_model': 5.256090841230674e-05, 'epoch': 1.76}) +[2026-06-16 02:20:43,456][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.10520706325769424, 'train/grad_norm_pre_clip': 10.876115798950195, 'timing/data': 0.026247510686516762, 'timing/model': 1.4355547679588199, 'learning_rate/qwen_vl_interface': 5.253045421241266e-06, 'learning_rate/action_model': 5.253045421241266e-05, 'epoch': 1.76}) +[2026-06-16 02:20:49,260][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.17881347239017487, 'train/grad_norm_pre_clip': 5.988362789154053, 'timing/data': 0.02535372693091631, 'timing/model': 1.441238704137504, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 1.76}) +[2026-06-16 02:20:55,165][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.09834276139736176, 'train/grad_norm_pre_clip': 5.760725975036621, 'timing/data': 0.022864095866680145, 'timing/model': 1.4637157572433352, 'learning_rate/qwen_vl_interface': 5.246954578758735e-06, 'learning_rate/action_model': 5.246954578758735e-05, 'epoch': 1.76}) +[2026-06-16 02:21:00,963][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.0978546142578125, 'train/grad_norm_pre_clip': 4.715722560882568, 'timing/data': 0.022336932830512524, 'timing/model': 1.435450947843492, 'learning_rate/qwen_vl_interface': 5.243909158769327e-06, 'learning_rate/action_model': 5.2439091587693264e-05, 'epoch': 1.76}) +[2026-06-16 02:21:06,859][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.11214669048786163, 'train/grad_norm_pre_clip': 5.301188945770264, 'timing/data': 0.0209652166813612, 'timing/model': 1.4686131794005632, 'learning_rate/qwen_vl_interface': 5.240863741283634e-06, 'learning_rate/action_model': 5.2408637412836336e-05, 'epoch': 1.76}) +[2026-06-16 02:21:12,746][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.0434761717915535, 'train/grad_norm_pre_clip': 3.0946767330169678, 'timing/data': 0.02178969979286194, 'timing/model': 1.7449313718825579, 'learning_rate/qwen_vl_interface': 5.2378183275535125e-06, 'learning_rate/action_model': 5.237818327553512e-05, 'epoch': 1.76}) +[2026-06-16 02:21:18,352][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.09716073423624039, 'train/grad_norm_pre_clip': 6.020407676696777, 'timing/data': 0.02300078608095646, 'timing/model': 1.447837395593524, 'learning_rate/qwen_vl_interface': 5.234772918830815e-06, 'learning_rate/action_model': 5.234772918830815e-05, 'epoch': 1.76}) +[2026-06-16 02:21:24,258][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.07103367149829865, 'train/grad_norm_pre_clip': 4.016328811645508, 'timing/data': 0.025718637742102146, 'timing/model': 1.4368082731962204, 'learning_rate/qwen_vl_interface': 5.231727516367395e-06, 'learning_rate/action_model': 5.231727516367395e-05, 'epoch': 1.76}) +[2026-06-16 02:21:30,066][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.04860714450478554, 'train/grad_norm_pre_clip': 4.260871887207031, 'timing/data': 0.022739344276487827, 'timing/model': 1.4572689272463322, 'learning_rate/qwen_vl_interface': 5.228682121415104e-06, 'learning_rate/action_model': 5.228682121415104e-05, 'epoch': 1.76}) +[2026-06-16 02:21:35,966][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.16511423885822296, 'train/grad_norm_pre_clip': 7.77322244644165, 'timing/data': 0.021566439419984818, 'timing/model': 1.4509033588692546, 'learning_rate/qwen_vl_interface': 5.225636735225789e-06, 'learning_rate/action_model': 5.225636735225788e-05, 'epoch': 1.76}) +[2026-06-16 02:21:41,854][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.18400651216506958, 'train/grad_norm_pre_clip': 13.189730644226074, 'timing/data': 0.022376812994480133, 'timing/model': 1.4290993073955178, 'learning_rate/qwen_vl_interface': 5.22259135905129e-06, 'learning_rate/action_model': 5.2225913590512895e-05, 'epoch': 1.76}) +[2026-06-16 02:21:47,760][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.14498521387577057, 'train/grad_norm_pre_clip': 5.2199320793151855, 'timing/data': 0.02078027557581663, 'timing/model': 1.715840837918222, 'learning_rate/qwen_vl_interface': 5.219545994143448e-06, 'learning_rate/action_model': 5.219545994143448e-05, 'epoch': 1.77}) +[2026-06-16 02:21:53,595][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.3287401497364044, 'train/grad_norm_pre_clip': 11.857095718383789, 'timing/data': 0.022826865315437317, 'timing/model': 1.684378013946116, 'learning_rate/qwen_vl_interface': 5.216500641754097e-06, 'learning_rate/action_model': 5.216500641754097e-05, 'epoch': 1.77}) +[2026-06-16 02:21:59,246][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.4702228903770447, 'train/grad_norm_pre_clip': 14.355328559875488, 'timing/data': 0.023046456277370453, 'timing/model': 1.4363630218431354, 'learning_rate/qwen_vl_interface': 5.213455303135067e-06, 'learning_rate/action_model': 5.2134553031350673e-05, 'epoch': 1.77}) +[2026-06-16 02:22:05,067][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.13282981514930725, 'train/grad_norm_pre_clip': 7.811497211456299, 'timing/data': 0.022390922531485558, 'timing/model': 1.4689415218308568, 'learning_rate/qwen_vl_interface': 5.2104099795381836e-06, 'learning_rate/action_model': 5.210409979538183e-05, 'epoch': 1.77}) +[2026-06-16 02:22:10,954][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.08734747767448425, 'train/grad_norm_pre_clip': 5.604894638061523, 'timing/data': 0.021572288125753403, 'timing/model': 1.4626283636316657, 'learning_rate/qwen_vl_interface': 5.207364672215261e-06, 'learning_rate/action_model': 5.2073646722152604e-05, 'epoch': 1.77}) +[2026-06-16 02:22:16,859][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.25471630692481995, 'train/grad_norm_pre_clip': 7.909912586212158, 'timing/data': 0.02491458412259817, 'timing/model': 1.4268657891079783, 'learning_rate/qwen_vl_interface': 5.2043193824181095e-06, 'learning_rate/action_model': 5.2043193824181094e-05, 'epoch': 1.77}) +[2026-06-16 02:22:22,739][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.04178342968225479, 'train/grad_norm_pre_clip': 1.843752145767212, 'timing/data': 0.02100281696766615, 'timing/model': 1.7246190691366792, 'learning_rate/qwen_vl_interface': 5.201274111398535e-06, 'learning_rate/action_model': 5.2012741113985356e-05, 'epoch': 1.77}) +[2026-06-16 02:22:28,539][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.13013920187950134, 'train/grad_norm_pre_clip': 6.564601421356201, 'timing/data': 0.021691350266337395, 'timing/model': 1.7274848530068994, 'learning_rate/qwen_vl_interface': 5.198228860408334e-06, 'learning_rate/action_model': 5.198228860408334e-05, 'epoch': 1.77}) +[2026-06-16 02:22:34,059][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.12363073974847794, 'train/grad_norm_pre_clip': 7.609738349914551, 'timing/data': 0.022659444250166416, 'timing/model': 1.4351340159773827, 'learning_rate/qwen_vl_interface': 5.195183630699292e-06, 'learning_rate/action_model': 5.195183630699292e-05, 'epoch': 1.77}) +[2026-06-16 02:22:39,855][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.04700933024287224, 'train/grad_norm_pre_clip': 6.268005847930908, 'timing/data': 0.02033072430640459, 'timing/model': 1.4407004620879889, 'learning_rate/qwen_vl_interface': 5.192138423523189e-06, 'learning_rate/action_model': 5.1921384235231895e-05, 'epoch': 1.77}) +[2026-06-16 02:22:45,657][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.1491161584854126, 'train/grad_norm_pre_clip': 7.685626029968262, 'timing/data': 0.02148383855819702, 'timing/model': 1.4415221055969596, 'learning_rate/qwen_vl_interface': 5.189093240131795e-06, 'learning_rate/action_model': 5.189093240131795e-05, 'epoch': 1.77}) +[2026-06-16 02:22:51,558][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.20508383214473724, 'train/grad_norm_pre_clip': 10.38564395904541, 'timing/data': 0.021845079958438873, 'timing/model': 1.4642489207908511, 'learning_rate/qwen_vl_interface': 5.186048081776869e-06, 'learning_rate/action_model': 5.1860480817768685e-05, 'epoch': 1.77}) +[2026-06-16 02:22:57,434][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.043434809893369675, 'train/grad_norm_pre_clip': 3.1340179443359375, 'timing/data': 0.02127262856811285, 'timing/model': 1.7310182498767972, 'learning_rate/qwen_vl_interface': 5.183002949710161e-06, 'learning_rate/action_model': 5.18300294971016e-05, 'epoch': 1.77}) +[2026-06-16 02:23:02,962][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.3266674280166626, 'train/grad_norm_pre_clip': 11.686708450317383, 'timing/data': 0.02243516407907009, 'timing/model': 1.4419314581900835, 'learning_rate/qwen_vl_interface': 5.179957845183411e-06, 'learning_rate/action_model': 5.179957845183411e-05, 'epoch': 1.77}) +[2026-06-16 02:23:08,859][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.25633704662323, 'train/grad_norm_pre_clip': 11.83151626586914, 'timing/data': 0.02126483805477619, 'timing/model': 1.439189376309514, 'learning_rate/qwen_vl_interface': 5.176912769448343e-06, 'learning_rate/action_model': 5.176912769448343e-05, 'epoch': 1.77}) +[2026-06-16 02:23:14,653][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.05895206704735756, 'train/grad_norm_pre_clip': 5.864155292510986, 'timing/data': 0.02195112034678459, 'timing/model': 1.4287490881979465, 'learning_rate/qwen_vl_interface': 5.173867723756676e-06, 'learning_rate/action_model': 5.173867723756676e-05, 'epoch': 1.77}) +[2026-06-16 02:23:20,568][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.11447300761938095, 'train/grad_norm_pre_clip': 5.88651180267334, 'timing/data': 0.025073645636439323, 'timing/model': 1.4755034632980824, 'learning_rate/qwen_vl_interface': 5.170822709360112e-06, 'learning_rate/action_model': 5.170822709360111e-05, 'epoch': 1.77}) +[2026-06-16 02:23:26,352][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.10830418020486832, 'train/grad_norm_pre_clip': 5.524450302124023, 'timing/data': 0.02215847186744213, 'timing/model': 1.419710235670209, 'learning_rate/qwen_vl_interface': 5.16777772751034e-06, 'learning_rate/action_model': 5.16777772751034e-05, 'epoch': 1.77}) +[2026-06-16 02:23:32,204][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.0166469719260931, 'train/grad_norm_pre_clip': 1.3905214071273804, 'timing/data': 0.02130087837576866, 'timing/model': 1.7034811628982425, 'learning_rate/qwen_vl_interface': 5.16473277945904e-06, 'learning_rate/action_model': 5.1647327794590396e-05, 'epoch': 1.77}) +[2026-06-16 02:23:37,749][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.2949272394180298, 'train/grad_norm_pre_clip': 13.864617347717285, 'timing/data': 0.02020547352731228, 'timing/model': 1.4413017854094505, 'learning_rate/qwen_vl_interface': 5.161687866457872e-06, 'learning_rate/action_model': 5.161687866457871e-05, 'epoch': 1.77}) +[2026-06-16 02:23:43,563][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.16690583527088165, 'train/grad_norm_pre_clip': 13.257450103759766, 'timing/data': 0.024602542631328106, 'timing/model': 1.4454965349286795, 'learning_rate/qwen_vl_interface': 5.158642989758487e-06, 'learning_rate/action_model': 5.158642989758487e-05, 'epoch': 1.77}) +[2026-06-16 02:23:49,348][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.12437626719474792, 'train/grad_norm_pre_clip': 7.731265068054199, 'timing/data': 0.02074674516916275, 'timing/model': 1.432469755411148, 'learning_rate/qwen_vl_interface': 5.155598150612519e-06, 'learning_rate/action_model': 5.1555981506125185e-05, 'epoch': 1.77}) +[2026-06-16 02:23:55,076][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.40316733717918396, 'train/grad_norm_pre_clip': 25.298471450805664, 'timing/data': 0.020151661708950996, 'timing/model': 1.3989121513441205, 'learning_rate/qwen_vl_interface': 5.152553350271584e-06, 'learning_rate/action_model': 5.152553350271584e-05, 'epoch': 1.78}) +[2026-06-16 02:24:00,950][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.24466942250728607, 'train/grad_norm_pre_clip': 8.6008939743042, 'timing/data': 0.02074343618005514, 'timing/model': 1.4353033974766731, 'learning_rate/qwen_vl_interface': 5.149508589987286e-06, 'learning_rate/action_model': 5.1495085899872855e-05, 'epoch': 1.78}) +[2026-06-16 02:24:06,801][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.19618643820285797, 'train/grad_norm_pre_clip': 6.944916248321533, 'timing/data': 0.02194920089095831, 'timing/model': 1.6853208607062697, 'learning_rate/qwen_vl_interface': 5.14646387101121e-06, 'learning_rate/action_model': 5.14646387101121e-05, 'epoch': 1.78}) +[2026-06-16 02:24:12,355][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.13359220325946808, 'train/grad_norm_pre_clip': 7.675556659698486, 'timing/data': 0.021425177343189716, 'timing/model': 1.450733169913292, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 1.78}) +[2026-06-16 02:24:18,254][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.2336168885231018, 'train/grad_norm_pre_clip': 9.031926155090332, 'timing/data': 0.02126143779605627, 'timing/model': 1.4363305242732167, 'learning_rate/qwen_vl_interface': 5.140374561989988e-06, 'learning_rate/action_model': 5.140374561989988e-05, 'epoch': 1.78}) +[2026-06-16 02:24:24,056][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.14563630521297455, 'train/grad_norm_pre_clip': 6.967528820037842, 'timing/data': 0.02177513111382723, 'timing/model': 1.4341893140226603, 'learning_rate/qwen_vl_interface': 5.137329974447924e-06, 'learning_rate/action_model': 5.137329974447924e-05, 'epoch': 1.78}) +[2026-06-16 02:24:29,963][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.1820470690727234, 'train/grad_norm_pre_clip': 10.43997573852539, 'timing/data': 0.025130625814199448, 'timing/model': 1.449149222113192, 'learning_rate/qwen_vl_interface': 5.1342854332202516e-06, 'learning_rate/action_model': 5.1342854332202514e-05, 'epoch': 1.78}) +[2026-06-16 02:24:35,844][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.06504859030246735, 'train/grad_norm_pre_clip': 5.042553424835205, 'timing/data': 0.021376238204538822, 'timing/model': 1.735110379755497, 'learning_rate/qwen_vl_interface': 5.131240939558467e-06, 'learning_rate/action_model': 5.131240939558467e-05, 'epoch': 1.78}) +[2026-06-16 02:24:41,352][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.23257791996002197, 'train/grad_norm_pre_clip': 10.004676818847656, 'timing/data': 0.021272008307278156, 'timing/model': 1.4332874389365315, 'learning_rate/qwen_vl_interface': 5.128196494714048e-06, 'learning_rate/action_model': 5.128196494714047e-05, 'epoch': 1.78}) +[2026-06-16 02:24:47,253][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.11181841790676117, 'train/grad_norm_pre_clip': 3.261046886444092, 'timing/data': 0.0229505468159914, 'timing/model': 1.4416771782562137, 'learning_rate/qwen_vl_interface': 5.125152099938448e-06, 'learning_rate/action_model': 5.125152099938447e-05, 'epoch': 1.78}) +[2026-06-16 02:24:53,048][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.03624507039785385, 'train/grad_norm_pre_clip': 1.9376579523086548, 'timing/data': 0.02072469610720873, 'timing/model': 1.4436471667140722, 'learning_rate/qwen_vl_interface': 5.122107756483104e-06, 'learning_rate/action_model': 5.122107756483103e-05, 'epoch': 1.78}) +[2026-06-16 02:24:58,751][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.1474646031856537, 'train/grad_norm_pre_clip': 5.0696940422058105, 'timing/data': 0.020755485631525517, 'timing/model': 1.4457850269973278, 'learning_rate/qwen_vl_interface': 5.119063465599429e-06, 'learning_rate/action_model': 5.1190634655994285e-05, 'epoch': 1.78}) +[2026-06-16 02:25:04,553][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.09641768783330917, 'train/grad_norm_pre_clip': 5.004782199859619, 'timing/data': 0.02134503796696663, 'timing/model': 1.403053680434823, 'learning_rate/qwen_vl_interface': 5.116019228538818e-06, 'learning_rate/action_model': 5.116019228538818e-05, 'epoch': 1.78}) +[2026-06-16 02:25:10,435][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.05490279942750931, 'train/grad_norm_pre_clip': 3.178560256958008, 'timing/data': 0.020609955303370953, 'timing/model': 1.7374416701495647, 'learning_rate/qwen_vl_interface': 5.112975046552641e-06, 'learning_rate/action_model': 5.1129750465526404e-05, 'epoch': 1.78}) +[2026-06-16 02:25:16,048][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.1795596480369568, 'train/grad_norm_pre_clip': 8.423986434936523, 'timing/data': 0.02168843988329172, 'timing/model': 1.4394202576950192, 'learning_rate/qwen_vl_interface': 5.109930920892246e-06, 'learning_rate/action_model': 5.1099309208922464e-05, 'epoch': 1.78}) +[2026-06-16 02:25:21,859][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.18337073922157288, 'train/grad_norm_pre_clip': 9.361764907836914, 'timing/data': 0.021335828118026257, 'timing/model': 1.448517870157957, 'learning_rate/qwen_vl_interface': 5.10688685280896e-06, 'learning_rate/action_model': 5.10688685280896e-05, 'epoch': 1.78}) +[2026-06-16 02:25:27,753][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.06505826860666275, 'train/grad_norm_pre_clip': 4.410025119781494, 'timing/data': 0.022239701822400093, 'timing/model': 1.445075853727758, 'learning_rate/qwen_vl_interface': 5.103842843554082e-06, 'learning_rate/action_model': 5.103842843554082e-05, 'epoch': 1.78}) +[2026-06-16 02:25:33,649][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.14546266198158264, 'train/grad_norm_pre_clip': 8.887982368469238, 'timing/data': 0.021774619817733765, 'timing/model': 1.4371630679816008, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 1.78}) +[2026-06-16 02:36:11,066][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.01269636582583189, 'train/grad_norm_pre_clip': 1.5105301141738892, 'eval/loss': 0.06212012469768524, 'eval/flap_precision': 0.7669902912621359, 'eval/flap_recall': 0.8618181818181818, 'eval/flap_f1': 0.8116438356164383, 'eval/flap_rate_model': 0.08583333333333333, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.8116438356164383, 'timing/data': 0.020184493623673916, 'timing/model': 1.3959614681079984, 'learning_rate/qwen_vl_interface': 5.097755006534638e-06, 'learning_rate/action_model': 5.097755006534638e-05, 'epoch': 1.78}) +[2026-06-16 02:36:11,067][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3900_state +[2026-06-16 02:36:11,073][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:36:18,812][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3900_state/model.safetensors +[2026-06-16 02:36:30,093][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3900_state/optimizer.bin +[2026-06-16 02:36:30,094][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3900_state/sampler.bin +[2026-06-16 02:36:30,094][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3900_state/sampler_1.bin +[2026-06-16 02:36:30,096][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_3900_state/random_states_0.pkl +[2026-06-16 02:36:30,261][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3900 +[2026-06-16 02:36:30,262][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:36:30,277][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:36:36,025][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.16832555830478668, 'train/grad_norm_pre_clip': 5.779992580413818, 'timing/data': 0.022577408701181412, 'timing/model': 1.377337345853448, 'learning_rate/qwen_vl_interface': 5.094711181272551e-06, 'learning_rate/action_model': 5.094711181272551e-05, 'epoch': 1.78}) +[2026-06-16 02:36:41,708][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.14934274554252625, 'train/grad_norm_pre_clip': 11.591015815734863, 'timing/data': 0.024602492339909077, 'timing/model': 1.7233453895896673, 'learning_rate/qwen_vl_interface': 5.091667419843833e-06, 'learning_rate/action_model': 5.091667419843833e-05, 'epoch': 1.78}) +[2026-06-16 02:36:47,351][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.03319086879491806, 'train/grad_norm_pre_clip': 8.097114562988281, 'timing/data': 0.02065936289727688, 'timing/model': 1.6605577860027552, 'learning_rate/qwen_vl_interface': 5.088623723499657e-06, 'learning_rate/action_model': 5.088623723499657e-05, 'epoch': 1.78}) +[2026-06-16 02:36:52,830][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.04385978356003761, 'train/grad_norm_pre_clip': 2.4451801776885986, 'timing/data': 0.028086677193641663, 'timing/model': 1.4467573529109359, 'learning_rate/qwen_vl_interface': 5.085580093491171e-06, 'learning_rate/action_model': 5.085580093491171e-05, 'epoch': 1.79}) +[2026-06-16 02:36:58,528][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.09411545097827911, 'train/grad_norm_pre_clip': 5.6770710945129395, 'timing/data': 0.02126881666481495, 'timing/model': 1.4125328278169036, 'learning_rate/qwen_vl_interface': 5.0825365310695e-06, 'learning_rate/action_model': 5.082536531069499e-05, 'epoch': 1.79}) +[2026-06-16 02:37:04,256][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.2645248472690582, 'train/grad_norm_pre_clip': 9.099748611450195, 'timing/data': 0.021385475993156433, 'timing/model': 1.4632004583254457, 'learning_rate/qwen_vl_interface': 5.079493037485735e-06, 'learning_rate/action_model': 5.0794930374857344e-05, 'epoch': 1.79}) +[2026-06-16 02:37:10,034][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.17235253751277924, 'train/grad_norm_pre_clip': 12.50730037689209, 'timing/data': 0.025534155778586864, 'timing/model': 1.442490485496819, 'learning_rate/qwen_vl_interface': 5.076449613990941e-06, 'learning_rate/action_model': 5.076449613990941e-05, 'epoch': 1.79}) +[2026-06-16 02:37:15,808][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.07994159311056137, 'train/grad_norm_pre_clip': 3.8635125160217285, 'timing/data': 0.021580137312412262, 'timing/model': 1.7155585614964366, 'learning_rate/qwen_vl_interface': 5.073406261836157e-06, 'learning_rate/action_model': 5.0734062618361565e-05, 'epoch': 1.79}) +[2026-06-16 02:37:21,337][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.09234122186899185, 'train/grad_norm_pre_clip': 5.148861408233643, 'timing/data': 0.020166127011179924, 'timing/model': 1.4405058668926358, 'learning_rate/qwen_vl_interface': 5.070362982272386e-06, 'learning_rate/action_model': 5.070362982272385e-05, 'epoch': 1.79}) +[2026-06-16 02:37:27,164][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.2007814198732376, 'train/grad_norm_pre_clip': 15.965611457824707, 'timing/data': 0.02591484785079956, 'timing/model': 1.4444975340738893, 'learning_rate/qwen_vl_interface': 5.067319776550607e-06, 'learning_rate/action_model': 5.067319776550607e-05, 'epoch': 1.79}) +[2026-06-16 02:37:33,052][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.23210351169109344, 'train/grad_norm_pre_clip': 11.291491508483887, 'timing/data': 0.020233675837516785, 'timing/model': 1.4669934958219528, 'learning_rate/qwen_vl_interface': 5.064276645921768e-06, 'learning_rate/action_model': 5.064276645921768e-05, 'epoch': 1.79}) +[2026-06-16 02:37:38,837][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.07096975296735764, 'train/grad_norm_pre_clip': 4.530096054077148, 'timing/data': 0.02030984591692686, 'timing/model': 1.4494176674634218, 'learning_rate/qwen_vl_interface': 5.061233591636787e-06, 'learning_rate/action_model': 5.061233591636787e-05, 'epoch': 1.79}) +[2026-06-16 02:37:44,633][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.13184906542301178, 'train/grad_norm_pre_clip': 10.843815803527832, 'timing/data': 0.021503563970327377, 'timing/model': 1.452857461757958, 'learning_rate/qwen_vl_interface': 5.058190614946544e-06, 'learning_rate/action_model': 5.058190614946544e-05, 'epoch': 1.79}) +[2026-06-16 02:37:50,419][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.09748432040214539, 'train/grad_norm_pre_clip': 6.802841663360596, 'timing/data': 0.020262515172362328, 'timing/model': 1.726102376356721, 'learning_rate/qwen_vl_interface': 5.055147717101895e-06, 'learning_rate/action_model': 5.055147717101894e-05, 'epoch': 1.79}) +[2026-06-16 02:37:55,935][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.14089034497737885, 'train/grad_norm_pre_clip': 6.917416572570801, 'timing/data': 0.010549715720117092, 'timing/model': 1.4400448510423303, 'learning_rate/qwen_vl_interface': 5.052104899353658e-06, 'learning_rate/action_model': 5.0521048993536577e-05, 'epoch': 1.79}) +[2026-06-16 02:38:01,649][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.09521425515413284, 'train/grad_norm_pre_clip': 4.029949188232422, 'timing/data': 0.010435514152050018, 'timing/model': 1.407143278978765, 'learning_rate/qwen_vl_interface': 5.049062162952622e-06, 'learning_rate/action_model': 5.0490621629526214e-05, 'epoch': 1.79}) +[2026-06-16 02:38:07,443][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.17742584645748138, 'train/grad_norm_pre_clip': 8.300939559936523, 'timing/data': 0.00987928081303835, 'timing/model': 1.4442094527184963, 'learning_rate/qwen_vl_interface': 5.046019509149541e-06, 'learning_rate/action_model': 5.046019509149541e-05, 'epoch': 1.79}) +[2026-06-16 02:38:13,258][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.10880669206380844, 'train/grad_norm_pre_clip': 13.148320198059082, 'timing/data': 0.010918066836893559, 'timing/model': 1.4604367697611451, 'learning_rate/qwen_vl_interface': 5.042976939195135e-06, 'learning_rate/action_model': 5.0429769391951355e-05, 'epoch': 1.79}) +[2026-06-16 02:38:19,137][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.10447143018245697, 'train/grad_norm_pre_clip': 6.661050796508789, 'timing/data': 0.01154211163520813, 'timing/model': 1.4547450002282858, 'learning_rate/qwen_vl_interface': 5.039934454340091e-06, 'learning_rate/action_model': 5.0399344543400907e-05, 'epoch': 1.79}) +[2026-06-16 02:38:24,916][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.2130720317363739, 'train/grad_norm_pre_clip': 12.939066886901855, 'timing/data': 0.011577432043850422, 'timing/model': 1.6966447429731488, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 1.79}) +[2026-06-16 02:38:30,760][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.11642339080572128, 'train/grad_norm_pre_clip': 4.743634223937988, 'timing/data': 0.011049778200685978, 'timing/model': 1.6568654011934996, 'learning_rate/qwen_vl_interface': 5.033849744930647e-06, 'learning_rate/action_model': 5.033849744930646e-05, 'epoch': 1.79}) +[2026-06-16 02:38:36,324][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.24297241866588593, 'train/grad_norm_pre_clip': 14.571550369262695, 'timing/data': 0.011125878430902958, 'timing/model': 1.4451597286388278, 'learning_rate/qwen_vl_interface': 5.030807522877443e-06, 'learning_rate/action_model': 5.030807522877443e-05, 'epoch': 1.79}) +[2026-06-16 02:38:42,043][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.1359972506761551, 'train/grad_norm_pre_clip': 10.668313980102539, 'timing/data': 0.010950705967843533, 'timing/model': 1.4346614899113774, 'learning_rate/qwen_vl_interface': 5.027765390925987e-06, 'learning_rate/action_model': 5.027765390925987e-05, 'epoch': 1.79}) +[2026-06-16 02:38:47,776][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.08558915555477142, 'train/grad_norm_pre_clip': 5.0828657150268555, 'timing/data': 0.010943185538053513, 'timing/model': 1.3937020683661103, 'learning_rate/qwen_vl_interface': 5.024723350326783e-06, 'learning_rate/action_model': 5.024723350326783e-05, 'epoch': 1.79}) +[2026-06-16 02:38:53,570][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.16294822096824646, 'train/grad_norm_pre_clip': 12.885647773742676, 'timing/data': 0.011596210300922394, 'timing/model': 1.3987719863653183, 'learning_rate/qwen_vl_interface': 5.021681402330303e-06, 'learning_rate/action_model': 5.021681402330303e-05, 'epoch': 1.8}) +[2026-06-16 02:38:59,440][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.048668183386325836, 'train/grad_norm_pre_clip': 4.615935325622559, 'timing/data': 0.011036906391382217, 'timing/model': 1.749261868186295, 'learning_rate/qwen_vl_interface': 5.018639548186972e-06, 'learning_rate/action_model': 5.0186395481869716e-05, 'epoch': 1.8}) +[2026-06-16 02:39:05,251][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.2075880914926529, 'train/grad_norm_pre_clip': 6.829811096191406, 'timing/data': 0.01078870426863432, 'timing/model': 1.7417878340929747, 'learning_rate/qwen_vl_interface': 5.015597789147184e-06, 'learning_rate/action_model': 5.015597789147184e-05, 'epoch': 1.8}) +[2026-06-16 02:39:10,796][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.1544656753540039, 'train/grad_norm_pre_clip': 13.872037887573242, 'timing/data': 0.010827193967998028, 'timing/model': 1.4136818861588836, 'learning_rate/qwen_vl_interface': 5.012556126461291e-06, 'learning_rate/action_model': 5.012556126461291e-05, 'epoch': 1.8}) +[2026-06-16 02:39:16,661][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.09419716894626617, 'train/grad_norm_pre_clip': 3.269381523132324, 'timing/data': 0.012333273887634277, 'timing/model': 1.4765300918370485, 'learning_rate/qwen_vl_interface': 5.0095145613796044e-06, 'learning_rate/action_model': 5.0095145613796046e-05, 'epoch': 1.8}) +[2026-06-16 02:39:22,557][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.09978605061769485, 'train/grad_norm_pre_clip': 9.02363109588623, 'timing/data': 0.011931432411074638, 'timing/model': 1.4577485574409366, 'learning_rate/qwen_vl_interface': 5.0064730951524e-06, 'learning_rate/action_model': 5.0064730951524e-05, 'epoch': 1.8}) +[2026-06-16 02:39:28,470][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.19415847957134247, 'train/grad_norm_pre_clip': 8.046426773071289, 'timing/data': 0.012070373632013798, 'timing/model': 1.4826336512342095, 'learning_rate/qwen_vl_interface': 5.003431729029905e-06, 'learning_rate/action_model': 5.003431729029905e-05, 'epoch': 1.8}) +[2026-06-16 02:39:34,331][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.09823138266801834, 'train/grad_norm_pre_clip': 6.069891929626465, 'timing/data': 0.010721812956035137, 'timing/model': 1.7286361940205097, 'learning_rate/qwen_vl_interface': 5.000390464262314e-06, 'learning_rate/action_model': 5.0003904642623136e-05, 'epoch': 1.8}) +[2026-06-16 02:39:40,206][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.029785873368382454, 'train/grad_norm_pre_clip': 2.1878812313079834, 'timing/data': 0.011285366490483284, 'timing/model': 1.7029031682759523, 'learning_rate/qwen_vl_interface': 4.997349302099774e-06, 'learning_rate/action_model': 4.997349302099773e-05, 'epoch': 1.8}) +[2026-06-16 02:39:45,750][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.3524418771266937, 'train/grad_norm_pre_clip': 13.06546688079834, 'timing/data': 0.010823814198374748, 'timing/model': 1.4588862750679255, 'learning_rate/qwen_vl_interface': 4.994308243792391e-06, 'learning_rate/action_model': 4.994308243792391e-05, 'epoch': 1.8}) +[2026-06-16 02:39:51,564][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.04596420377492905, 'train/grad_norm_pre_clip': 3.1183409690856934, 'timing/data': 0.010875094681978226, 'timing/model': 1.4652959341183305, 'learning_rate/qwen_vl_interface': 4.991267290590233e-06, 'learning_rate/action_model': 4.991267290590233e-05, 'epoch': 1.8}) +[2026-06-16 02:39:57,461][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.1725403368473053, 'train/grad_norm_pre_clip': 4.958248615264893, 'timing/data': 0.01195818092674017, 'timing/model': 1.4555996051058173, 'learning_rate/qwen_vl_interface': 4.988226443743317e-06, 'learning_rate/action_model': 4.9882264437433174e-05, 'epoch': 1.8}) +[2026-06-16 02:40:03,254][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.08082544803619385, 'train/grad_norm_pre_clip': 5.077935218811035, 'timing/data': 0.011406836099922657, 'timing/model': 1.398456727154553, 'learning_rate/qwen_vl_interface': 4.985185704501623e-06, 'learning_rate/action_model': 4.985185704501623e-05, 'epoch': 1.8}) +[2026-06-16 02:40:09,164][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.2500908076763153, 'train/grad_norm_pre_clip': 8.911821365356445, 'timing/data': 0.011727988719940186, 'timing/model': 1.4684769148007035, 'learning_rate/qwen_vl_interface': 4.982145074115082e-06, 'learning_rate/action_model': 4.982145074115082e-05, 'epoch': 1.8}) +[2026-06-16 02:40:15,108][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.09689034521579742, 'train/grad_norm_pre_clip': 6.583677768707275, 'timing/data': 0.011485517024993896, 'timing/model': 1.7210207637399435, 'learning_rate/qwen_vl_interface': 4.979104553833584e-06, 'learning_rate/action_model': 4.979104553833584e-05, 'epoch': 1.8}) +[2026-06-16 02:40:20,664][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.26738226413726807, 'train/grad_norm_pre_clip': 13.999112129211426, 'timing/data': 0.013155598193407059, 'timing/model': 1.4419093821197748, 'learning_rate/qwen_vl_interface': 4.9760641449069724e-06, 'learning_rate/action_model': 4.976064144906972e-05, 'epoch': 1.8}) +[2026-06-16 02:40:26,559][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.08162318915128708, 'train/grad_norm_pre_clip': 7.368546962738037, 'timing/data': 0.011653956957161427, 'timing/model': 1.4648878928273916, 'learning_rate/qwen_vl_interface': 4.973023848585042e-06, 'learning_rate/action_model': 4.973023848585042e-05, 'epoch': 1.8}) +[2026-06-16 02:40:32,356][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.10915229469537735, 'train/grad_norm_pre_clip': 6.0521559715271, 'timing/data': 0.010784482583403587, 'timing/model': 1.4465620685368776, 'learning_rate/qwen_vl_interface': 4.969983666117545e-06, 'learning_rate/action_model': 4.9699836661175446e-05, 'epoch': 1.8}) +[2026-06-16 02:40:38,252][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.09888359159231186, 'train/grad_norm_pre_clip': 5.111133098602295, 'timing/data': 0.011030063033103943, 'timing/model': 1.4512503575533628, 'learning_rate/qwen_vl_interface': 4.966943598754185e-06, 'learning_rate/action_model': 4.9669435987541854e-05, 'epoch': 1.8}) +[2026-06-16 02:40:44,125][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.09197071194648743, 'train/grad_norm_pre_clip': 4.9204816818237305, 'timing/data': 0.011726288124918938, 'timing/model': 1.703713359311223, 'learning_rate/qwen_vl_interface': 4.963903647744621e-06, 'learning_rate/action_model': 4.96390364774462e-05, 'epoch': 1.8}) +[2026-06-16 02:40:49,726][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.10704999417066574, 'train/grad_norm_pre_clip': 7.127013683319092, 'timing/data': 0.0112390648573637, 'timing/model': 1.4421360436826944, 'learning_rate/qwen_vl_interface': 4.9608638143384604e-06, 'learning_rate/action_model': 4.960863814338461e-05, 'epoch': 1.8}) +[2026-06-16 02:40:55,575][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.17519748210906982, 'train/grad_norm_pre_clip': 9.389776229858398, 'timing/data': 0.00989119615405798, 'timing/model': 1.4754761131480336, 'learning_rate/qwen_vl_interface': 4.9578240997852636e-06, 'learning_rate/action_model': 4.957824099785264e-05, 'epoch': 1.8}) +[2026-06-16 02:41:01,360][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.013265429995954037, 'train/grad_norm_pre_clip': 1.0557371377944946, 'timing/data': 0.010494348593056202, 'timing/model': 1.475520621985197, 'learning_rate/qwen_vl_interface': 4.954784505334544e-06, 'learning_rate/action_model': 4.9547845053345434e-05, 'epoch': 1.81}) +[2026-06-16 02:41:07,148][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.135789155960083, 'train/grad_norm_pre_clip': 7.549650192260742, 'timing/data': 0.011113492771983147, 'timing/model': 1.4515835456550121, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 1.81}) +[2026-06-16 02:41:12,941][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.24125708639621735, 'train/grad_norm_pre_clip': 10.509410858154297, 'timing/data': 0.011014663614332676, 'timing/model': 1.4449204793199897, 'learning_rate/qwen_vl_interface': 4.94870568173833e-06, 'learning_rate/action_model': 4.9487056817383307e-05, 'epoch': 1.81}) +[2026-06-16 02:41:18,741][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.07302764803171158, 'train/grad_norm_pre_clip': 5.304680824279785, 'timing/data': 0.012543682008981705, 'timing/model': 1.7298346739262342, 'learning_rate/qwen_vl_interface': 4.945666455091614e-06, 'learning_rate/action_model': 4.9456664550916136e-05, 'epoch': 1.81}) +[2026-06-16 02:41:24,358][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.08426797389984131, 'train/grad_norm_pre_clip': 10.56478214263916, 'timing/data': 0.011444414034485817, 'timing/model': 1.464667109772563, 'learning_rate/qwen_vl_interface': 4.942627353544921e-06, 'learning_rate/action_model': 4.942627353544921e-05, 'epoch': 1.81}) +[2026-06-16 02:41:30,150][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.03561766445636749, 'train/grad_norm_pre_clip': 2.279627799987793, 'timing/data': 0.011177143082022667, 'timing/model': 1.4591458337381482, 'learning_rate/qwen_vl_interface': 4.939588378347513e-06, 'learning_rate/action_model': 4.9395883783475125e-05, 'epoch': 1.81}) +[2026-06-16 02:41:35,908][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.09212026745080948, 'train/grad_norm_pre_clip': 5.192859649658203, 'timing/data': 0.011363054625689983, 'timing/model': 1.4164306465536356, 'learning_rate/qwen_vl_interface': 4.936549530748597e-06, 'learning_rate/action_model': 4.9365495307485963e-05, 'epoch': 1.81}) +[2026-06-16 02:41:41,647][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.12840750813484192, 'train/grad_norm_pre_clip': 9.125188827514648, 'timing/data': 0.01093596126884222, 'timing/model': 1.4440009975805879, 'learning_rate/qwen_vl_interface': 4.9335108119973286e-06, 'learning_rate/action_model': 4.933510811997328e-05, 'epoch': 1.81}) +[2026-06-16 02:41:47,445][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.07554180920124054, 'train/grad_norm_pre_clip': 4.427850246429443, 'timing/data': 0.010910901241004467, 'timing/model': 1.4134001499041915, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 1.81}) +[2026-06-16 02:41:53,216][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.13150885701179504, 'train/grad_norm_pre_clip': 11.248465538024902, 'timing/data': 0.011431953869760036, 'timing/model': 1.7121578035876155, 'learning_rate/qwen_vl_interface': 4.927433766034093e-06, 'learning_rate/action_model': 4.9274337660340934e-05, 'epoch': 1.81}) +[2026-06-16 02:41:58,742][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.02413226291537285, 'train/grad_norm_pre_clip': 2.2211172580718994, 'timing/data': 0.011618934571743011, 'timing/model': 1.4569724760949612, 'learning_rate/qwen_vl_interface': 4.92439544132017e-06, 'learning_rate/action_model': 4.92439544132017e-05, 'epoch': 1.81}) +[2026-06-16 02:42:04,574][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.20636209845542908, 'train/grad_norm_pre_clip': 14.181804656982422, 'timing/data': 0.011539534665644169, 'timing/model': 1.4893043050542474, 'learning_rate/qwen_vl_interface': 4.92135725044998e-06, 'learning_rate/action_model': 4.92135725044998e-05, 'epoch': 1.81}) +[2026-06-16 02:42:10,463][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.1563473492860794, 'train/grad_norm_pre_clip': 14.37629508972168, 'timing/data': 0.011058760806918144, 'timing/model': 1.4629437951371074, 'learning_rate/qwen_vl_interface': 4.918319194672412e-06, 'learning_rate/action_model': 4.918319194672412e-05, 'epoch': 1.81}) +[2026-06-16 02:42:16,363][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.21454961597919464, 'train/grad_norm_pre_clip': 14.233611106872559, 'timing/data': 0.011230641975998878, 'timing/model': 1.472713471390307, 'learning_rate/qwen_vl_interface': 4.9152812752362935e-06, 'learning_rate/action_model': 4.915281275236293e-05, 'epoch': 1.81}) +[2026-06-16 02:42:22,236][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.08534498512744904, 'train/grad_norm_pre_clip': 6.949550151824951, 'timing/data': 0.011131500825285912, 'timing/model': 1.74203940667212, 'learning_rate/qwen_vl_interface': 4.9122434933904e-06, 'learning_rate/action_model': 4.9122434933904e-05, 'epoch': 1.81}) +[2026-06-16 02:42:27,938][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.08728667348623276, 'train/grad_norm_pre_clip': 5.064443111419678, 'timing/data': 0.01125154085457325, 'timing/model': 1.6298420298844576, 'learning_rate/qwen_vl_interface': 4.909205850383449e-06, 'learning_rate/action_model': 4.9092058503834485e-05, 'epoch': 1.81}) +[2026-06-16 02:42:33,559][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.06464620679616928, 'train/grad_norm_pre_clip': 18.377256393432617, 'timing/data': 0.01093808002769947, 'timing/model': 1.4567689960822463, 'learning_rate/qwen_vl_interface': 4.906168347464098e-06, 'learning_rate/action_model': 4.906168347464098e-05, 'epoch': 1.81}) +[2026-06-16 02:42:39,463][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.0712486058473587, 'train/grad_norm_pre_clip': 5.393947601318359, 'timing/data': 0.011004070751369, 'timing/model': 1.4706848170608282, 'learning_rate/qwen_vl_interface': 4.903130985880955e-06, 'learning_rate/action_model': 4.903130985880954e-05, 'epoch': 1.81}) +[2026-06-16 02:42:45,360][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.08156612515449524, 'train/grad_norm_pre_clip': 6.419565200805664, 'timing/data': 0.012294899672269821, 'timing/model': 1.4450035160407424, 'learning_rate/qwen_vl_interface': 4.900093766882558e-06, 'learning_rate/action_model': 4.900093766882558e-05, 'epoch': 1.81}) +[2026-06-16 02:42:51,172][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.041108857840299606, 'train/grad_norm_pre_clip': 3.363424777984619, 'timing/data': 0.011292272247374058, 'timing/model': 1.4000331042334437, 'learning_rate/qwen_vl_interface': 4.897056691717397e-06, 'learning_rate/action_model': 4.8970566917173966e-05, 'epoch': 1.81}) +[2026-06-16 02:42:57,028][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.07815377414226532, 'train/grad_norm_pre_clip': 8.7997407913208, 'timing/data': 0.011180280707776546, 'timing/model': 1.7144797490909696, 'learning_rate/qwen_vl_interface': 4.894019761633898e-06, 'learning_rate/action_model': 4.894019761633898e-05, 'epoch': 1.81}) +[2026-06-16 02:43:02,738][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.04320446401834488, 'train/grad_norm_pre_clip': 4.391276836395264, 'timing/data': 0.01155056245625019, 'timing/model': 1.635091838426888, 'learning_rate/qwen_vl_interface': 4.8909829778804304e-06, 'learning_rate/action_model': 4.89098297788043e-05, 'epoch': 1.81}) +[2026-06-16 02:43:08,360][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.0592157319188118, 'train/grad_norm_pre_clip': 3.088174343109131, 'timing/data': 0.010882248170673847, 'timing/model': 1.4521657759323716, 'learning_rate/qwen_vl_interface': 4.887946341705298e-06, 'learning_rate/action_model': 4.887946341705297e-05, 'epoch': 1.82}) +[2026-06-16 02:43:14,649][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.03174419328570366, 'train/grad_norm_pre_clip': 4.637049674987793, 'timing/data': 0.013124561868607998, 'timing/model': 1.4426823100075126, 'learning_rate/qwen_vl_interface': 4.8849098543567465e-06, 'learning_rate/action_model': 4.884909854356747e-05, 'epoch': 1.82}) +[2026-06-16 02:43:20,497][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.08821571618318558, 'train/grad_norm_pre_clip': 7.707790374755859, 'timing/data': 0.014311868697404861, 'timing/model': 1.446693348698318, 'learning_rate/qwen_vl_interface': 4.881873517082965e-06, 'learning_rate/action_model': 4.8818735170829655e-05, 'epoch': 1.82}) +[2026-06-16 02:43:26,578][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.06242174282670021, 'train/grad_norm_pre_clip': 9.580588340759277, 'timing/data': 0.014959443360567093, 'timing/model': 1.477079613134265, 'learning_rate/qwen_vl_interface': 4.878837331132076e-06, 'learning_rate/action_model': 4.8788373311320754e-05, 'epoch': 1.82}) +[2026-06-16 02:43:32,746][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.1536395251750946, 'train/grad_norm_pre_clip': 11.072272300720215, 'timing/data': 0.014656801708042622, 'timing/model': 1.9313691006973386, 'learning_rate/qwen_vl_interface': 4.87580129775214e-06, 'learning_rate/action_model': 4.87580129775214e-05, 'epoch': 1.82}) +[2026-06-16 02:43:38,401][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.06311111152172089, 'train/grad_norm_pre_clip': 7.2889509201049805, 'timing/data': 0.03267145436257124, 'timing/model': 1.4678168930113316, 'learning_rate/qwen_vl_interface': 4.872765418191156e-06, 'learning_rate/action_model': 4.872765418191155e-05, 'epoch': 1.82}) +[2026-06-16 02:43:44,474][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.3467155396938324, 'train/grad_norm_pre_clip': 8.55604076385498, 'timing/data': 0.014304479584097862, 'timing/model': 1.4550601979717612, 'learning_rate/qwen_vl_interface': 4.8697296936970575e-06, 'learning_rate/action_model': 4.869729693697057e-05, 'epoch': 1.82}) +[2026-06-16 02:43:50,501][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.06142812594771385, 'train/grad_norm_pre_clip': 9.185548782348633, 'timing/data': 0.014219168573617935, 'timing/model': 1.4918047608807683, 'learning_rate/qwen_vl_interface': 4.8666941255177185e-06, 'learning_rate/action_model': 4.8666941255177185e-05, 'epoch': 1.82}) +[2026-06-16 02:43:56,668][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.19977839291095734, 'train/grad_norm_pre_clip': 10.440367698669434, 'timing/data': 0.011630871333181858, 'timing/model': 1.4675094336271286, 'learning_rate/qwen_vl_interface': 4.863658714900945e-06, 'learning_rate/action_model': 4.863658714900945e-05, 'epoch': 1.82}) +[2026-06-16 02:44:02,785][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.03868158534169197, 'train/grad_norm_pre_clip': 2.815073251724243, 'timing/data': 0.012996060773730278, 'timing/model': 1.4792626909911633, 'learning_rate/qwen_vl_interface': 4.860623463094483e-06, 'learning_rate/action_model': 4.860623463094483e-05, 'epoch': 1.82}) +[2026-06-16 02:44:08,890][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.11082187294960022, 'train/grad_norm_pre_clip': 8.519794464111328, 'timing/data': 0.01235443726181984, 'timing/model': 1.8695660140365362, 'learning_rate/qwen_vl_interface': 4.857588371346006e-06, 'learning_rate/action_model': 4.8575883713460057e-05, 'epoch': 1.82}) +[2026-06-16 02:44:14,578][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.040129583328962326, 'train/grad_norm_pre_clip': 3.808396816253662, 'timing/data': 0.011601022444665432, 'timing/model': 1.454524447210133, 'learning_rate/qwen_vl_interface': 4.854553440903127e-06, 'learning_rate/action_model': 4.854553440903126e-05, 'epoch': 1.82}) +[2026-06-16 02:44:20,687][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.05643334612250328, 'train/grad_norm_pre_clip': 7.340859413146973, 'timing/data': 0.012985880486667156, 'timing/model': 1.4614478461444378, 'learning_rate/qwen_vl_interface': 4.85151867301339e-06, 'learning_rate/action_model': 4.85151867301339e-05, 'epoch': 1.82}) +[2026-06-16 02:44:26,689][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.02320665866136551, 'train/grad_norm_pre_clip': 4.245010852813721, 'timing/data': 0.012379366904497147, 'timing/model': 1.4860414825379848, 'learning_rate/qwen_vl_interface': 4.848484068924274e-06, 'learning_rate/action_model': 4.848484068924275e-05, 'epoch': 1.82}) +[2026-06-16 02:44:32,775][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.1680053323507309, 'train/grad_norm_pre_clip': 11.913076400756836, 'timing/data': 0.012961129657924175, 'timing/model': 1.4336532661691308, 'learning_rate/qwen_vl_interface': 4.845449629883193e-06, 'learning_rate/action_model': 4.8454496298831934e-05, 'epoch': 1.82}) +[2026-06-16 02:44:38,884][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.22571605443954468, 'train/grad_norm_pre_clip': 14.64482307434082, 'timing/data': 0.01258944720029831, 'timing/model': 1.4912859555333853, 'learning_rate/qwen_vl_interface': 4.842415357137487e-06, 'learning_rate/action_model': 4.842415357137487e-05, 'epoch': 1.82}) +[2026-06-16 02:44:44,927][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.11696011573076248, 'train/grad_norm_pre_clip': 10.613740921020508, 'timing/data': 0.011728142388164997, 'timing/model': 1.8133773878216743, 'learning_rate/qwen_vl_interface': 4.839381251934432e-06, 'learning_rate/action_model': 4.8393812519344316e-05, 'epoch': 1.82}) +[2026-06-16 02:44:50,955][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.053832318633794785, 'train/grad_norm_pre_clip': 8.81217098236084, 'timing/data': 0.014018645510077477, 'timing/model': 1.849140769802034, 'learning_rate/qwen_vl_interface': 4.836347315521233e-06, 'learning_rate/action_model': 4.836347315521233e-05, 'epoch': 1.82}) +[2026-06-16 02:44:56,564][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.25229939818382263, 'train/grad_norm_pre_clip': 24.63675308227539, 'timing/data': 0.01204301230609417, 'timing/model': 1.4508839678019285, 'learning_rate/qwen_vl_interface': 4.833313549145029e-06, 'learning_rate/action_model': 4.833313549145029e-05, 'epoch': 1.82}) +[2026-06-16 02:45:02,566][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.039694931358098984, 'train/grad_norm_pre_clip': 4.688429832458496, 'timing/data': 0.012315114960074425, 'timing/model': 1.4599767094478011, 'learning_rate/qwen_vl_interface': 4.830279954052884e-06, 'learning_rate/action_model': 4.8302799540528836e-05, 'epoch': 1.82}) +[2026-06-16 02:45:08,582][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.028354525566101074, 'train/grad_norm_pre_clip': 2.3403754234313965, 'timing/data': 0.012346964329481125, 'timing/model': 1.4623867496848106, 'learning_rate/qwen_vl_interface': 4.827246531491797e-06, 'learning_rate/action_model': 4.827246531491797e-05, 'epoch': 1.82}) +[2026-06-16 02:45:14,579][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.14487594366073608, 'train/grad_norm_pre_clip': 9.172511100769043, 'timing/data': 0.014286627061665058, 'timing/model': 1.4165633078664541, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 1.82}) +[2026-06-16 02:45:20,629][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.3729201555252075, 'train/grad_norm_pre_clip': 20.144662857055664, 'timing/data': 0.012910897843539715, 'timing/model': 1.814594205468893, 'learning_rate/qwen_vl_interface': 4.821180208950422e-06, 'learning_rate/action_model': 4.821180208950421e-05, 'epoch': 1.83}) +[2026-06-16 02:45:26,601][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.028387008234858513, 'train/grad_norm_pre_clip': 3.1336917877197266, 'timing/data': 0.013218490406870842, 'timing/model': 1.780574994161725, 'learning_rate/qwen_vl_interface': 4.81814731146377e-06, 'learning_rate/action_model': 4.8181473114637695e-05, 'epoch': 1.83}) +[2026-06-16 02:45:32,281][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.019124124199151993, 'train/grad_norm_pre_clip': 3.50189208984375, 'timing/data': 0.012022132985293865, 'timing/model': 1.4851210955530405, 'learning_rate/qwen_vl_interface': 4.8151145914954465e-06, 'learning_rate/action_model': 4.815114591495446e-05, 'epoch': 1.83}) +[2026-06-16 02:45:38,288][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.034272681921720505, 'train/grad_norm_pre_clip': 4.782196044921875, 'timing/data': 0.013398081064224243, 'timing/model': 1.454452627338469, 'learning_rate/qwen_vl_interface': 4.812082050292088e-06, 'learning_rate/action_model': 4.8120820502920876e-05, 'epoch': 1.83}) +[2026-06-16 02:45:44,279][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.20887388288974762, 'train/grad_norm_pre_clip': 12.978315353393555, 'timing/data': 0.011636979877948761, 'timing/model': 1.470361402258277, 'learning_rate/qwen_vl_interface': 4.809049689100256e-06, 'learning_rate/action_model': 4.8090496891002555e-05, 'epoch': 1.83}) +[2026-06-16 02:45:50,299][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.031004292890429497, 'train/grad_norm_pre_clip': 4.174605369567871, 'timing/data': 0.012538756243884563, 'timing/model': 1.420541818253696, 'learning_rate/qwen_vl_interface': 4.80601750916644e-06, 'learning_rate/action_model': 4.8060175091664404e-05, 'epoch': 1.83}) +[2026-06-16 02:45:56,347][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.20111873745918274, 'train/grad_norm_pre_clip': 12.870063781738281, 'timing/data': 0.011679349467158318, 'timing/model': 1.8312012730166316, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 1.83}) +[2026-06-16 02:46:01,980][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.09363808482885361, 'train/grad_norm_pre_clip': 8.693007469177246, 'timing/data': 0.012545515783131123, 'timing/model': 1.452996050938964, 'learning_rate/qwen_vl_interface': 4.799953698058444e-06, 'learning_rate/action_model': 4.7999536980584434e-05, 'epoch': 1.83}) +[2026-06-16 02:46:07,973][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.13717275857925415, 'train/grad_norm_pre_clip': 13.626312255859375, 'timing/data': 0.012358574196696281, 'timing/model': 1.450831776484847, 'learning_rate/qwen_vl_interface': 4.796922069376867e-06, 'learning_rate/action_model': 4.796922069376867e-05, 'epoch': 1.83}) +[2026-06-16 02:57:09,281][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.06864969432353973, 'train/grad_norm_pre_clip': 9.554327011108398, 'eval/loss': 0.07344797253608704, 'eval/flap_precision': 0.7730263157894737, 'eval/flap_recall': 0.8545454545454545, 'eval/flap_f1': 0.8117443868739205, 'eval/flap_rate_model': 0.08444444444444445, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.8117443868739205, 'timing/data': 0.012774117290973663, 'timing/model': 1.465970914810896, 'learning_rate/qwen_vl_interface': 4.793890626938511e-06, 'learning_rate/action_model': 4.793890626938511e-05, 'epoch': 1.83}) +[2026-06-16 02:57:09,282][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4000_state +[2026-06-16 02:57:09,290][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 02:57:17,856][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4000_state/model.safetensors +[2026-06-16 02:57:30,046][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4000_state/optimizer.bin +[2026-06-16 02:57:30,047][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4000_state/sampler.bin +[2026-06-16 02:57:30,048][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-16 02:57:30,050][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-16 02:57:30,216][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-16 02:57:30,217][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 02:57:30,233][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 02:57:36,258][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.10350874811410904, 'train/grad_norm_pre_clip': 7.557468414306641, 'timing/data': 0.018023749813437462, 'timing/model': 1.447891317307949, 'learning_rate/qwen_vl_interface': 4.7908593719894914e-06, 'learning_rate/action_model': 4.790859371989491e-05, 'epoch': 1.83}) +[2026-06-16 02:57:42,248][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.12235473841428757, 'train/grad_norm_pre_clip': 14.744356155395508, 'timing/data': 0.014905132353305817, 'timing/model': 1.46829825732857, 'learning_rate/qwen_vl_interface': 4.78782830577584e-06, 'learning_rate/action_model': 4.7878283057758396e-05, 'epoch': 1.83}) +[2026-06-16 02:57:48,144][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 0.270113468170166, 'train/grad_norm_pre_clip': 10.2034273147583, 'timing/data': 0.015133332461118698, 'timing/model': 1.3945217980071902, 'learning_rate/qwen_vl_interface': 4.784797429543513e-06, 'learning_rate/action_model': 4.784797429543513e-05, 'epoch': 1.83}) +[2026-06-16 02:57:53,995][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.0698745846748352, 'train/grad_norm_pre_clip': 3.129478693008423, 'timing/data': 0.014761751517653465, 'timing/model': 1.727393874898553, 'learning_rate/qwen_vl_interface': 4.781766744538393e-06, 'learning_rate/action_model': 4.781766744538393e-05, 'epoch': 1.83}) +[2026-06-16 02:57:59,542][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.04864489659667015, 'train/grad_norm_pre_clip': 5.938641548156738, 'timing/data': 0.01428786851465702, 'timing/model': 1.4452779712155461, 'learning_rate/qwen_vl_interface': 4.778736252006277e-06, 'learning_rate/action_model': 4.7787362520062765e-05, 'epoch': 1.83}) +[2026-06-16 02:58:05,443][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.07084473222494125, 'train/grad_norm_pre_clip': 7.224991321563721, 'timing/data': 0.01598728820681572, 'timing/model': 1.4581460496410728, 'learning_rate/qwen_vl_interface': 4.775705953192885e-06, 'learning_rate/action_model': 4.7757059531928856e-05, 'epoch': 1.83}) +[2026-06-16 02:58:11,351][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.12244640290737152, 'train/grad_norm_pre_clip': 8.69074821472168, 'timing/data': 0.014639641158282757, 'timing/model': 1.4673617063090205, 'learning_rate/qwen_vl_interface': 4.7726758493438615e-06, 'learning_rate/action_model': 4.7726758493438613e-05, 'epoch': 1.83}) +[2026-06-16 02:58:17,246][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.05541732907295227, 'train/grad_norm_pre_clip': 5.793457508087158, 'timing/data': 0.015090461820363998, 'timing/model': 1.4464798793196678, 'learning_rate/qwen_vl_interface': 4.769645941704766e-06, 'learning_rate/action_model': 4.769645941704766e-05, 'epoch': 1.83}) +[2026-06-16 02:58:23,160][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.1136292815208435, 'train/grad_norm_pre_clip': 6.3372626304626465, 'timing/data': 0.01526834350079298, 'timing/model': 1.4260825859382749, 'learning_rate/qwen_vl_interface': 4.766616231521081e-06, 'learning_rate/action_model': 4.7666162315210806e-05, 'epoch': 1.83}) +[2026-06-16 02:58:29,249][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.005120830610394478, 'train/grad_norm_pre_clip': 1.1843317747116089, 'timing/data': 0.012893290258944035, 'timing/model': 1.8522871611639857, 'learning_rate/qwen_vl_interface': 4.763586720038202e-06, 'learning_rate/action_model': 4.7635867200382014e-05, 'epoch': 1.83}) +[2026-06-16 02:58:34,757][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.10215847194194794, 'train/grad_norm_pre_clip': 9.54980754852295, 'timing/data': 0.01681510265916586, 'timing/model': 1.43471369612962, 'learning_rate/qwen_vl_interface': 4.76055740850145e-06, 'learning_rate/action_model': 4.76055740850145e-05, 'epoch': 1.83}) +[2026-06-16 02:58:40,777][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.055292412638664246, 'train/grad_norm_pre_clip': 3.230710744857788, 'timing/data': 0.016100517474114895, 'timing/model': 1.467641425319016, 'learning_rate/qwen_vl_interface': 4.757528298156058e-06, 'learning_rate/action_model': 4.757528298156058e-05, 'epoch': 1.83}) +[2026-06-16 02:58:46,778][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 0.11660381406545639, 'train/grad_norm_pre_clip': 5.444386005401611, 'timing/data': 0.015108402818441391, 'timing/model': 1.4696420142427087, 'learning_rate/qwen_vl_interface': 4.754499390247181e-06, 'learning_rate/action_model': 4.754499390247181e-05, 'epoch': 1.84}) +[2026-06-16 02:58:52,759][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.0431160032749176, 'train/grad_norm_pre_clip': 5.869215965270996, 'timing/data': 0.01976360846310854, 'timing/model': 1.4358126651495695, 'learning_rate/qwen_vl_interface': 4.751470686019887e-06, 'learning_rate/action_model': 4.751470686019888e-05, 'epoch': 1.84}) +[2026-06-16 02:58:58,765][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.09186238795518875, 'train/grad_norm_pre_clip': 7.5074872970581055, 'timing/data': 0.012784999795258045, 'timing/model': 1.419215432368219, 'learning_rate/qwen_vl_interface': 4.7484421867191645e-06, 'learning_rate/action_model': 4.748442186719165e-05, 'epoch': 1.84}) +[2026-06-16 02:59:04,833][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.11170488595962524, 'train/grad_norm_pre_clip': 8.162325859069824, 'timing/data': 0.012324617244303226, 'timing/model': 1.815734569914639, 'learning_rate/qwen_vl_interface': 4.745413893589913e-06, 'learning_rate/action_model': 4.7454138935899125e-05, 'epoch': 1.84}) +[2026-06-16 02:59:10,459][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.36789312958717346, 'train/grad_norm_pre_clip': 14.921248435974121, 'timing/data': 0.012288357131183147, 'timing/model': 1.455515486188233, 'learning_rate/qwen_vl_interface': 4.7423858078769506e-06, 'learning_rate/action_model': 4.742385807876951e-05, 'epoch': 1.84}) +[2026-06-16 02:59:16,454][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.0493963286280632, 'train/grad_norm_pre_clip': 3.325958490371704, 'timing/data': 0.011968215927481651, 'timing/model': 1.4675225280225277, 'learning_rate/qwen_vl_interface': 4.739357930825008e-06, 'learning_rate/action_model': 4.739357930825008e-05, 'epoch': 1.84}) +[2026-06-16 02:59:22,368][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.13775143027305603, 'train/grad_norm_pre_clip': 13.204319953918457, 'timing/data': 0.013511553406715393, 'timing/model': 1.4531107684597373, 'learning_rate/qwen_vl_interface': 4.736330263678731e-06, 'learning_rate/action_model': 4.736330263678731e-05, 'epoch': 1.84}) +[2026-06-16 02:59:28,355][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.06771443784236908, 'train/grad_norm_pre_clip': 6.952883720397949, 'timing/data': 0.014249667525291443, 'timing/model': 1.4421556154266, 'learning_rate/qwen_vl_interface': 4.733302807682679e-06, 'learning_rate/action_model': 4.733302807682679e-05, 'epoch': 1.84}) +[2026-06-16 02:59:34,371][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.28511446714401245, 'train/grad_norm_pre_clip': 34.0216064453125, 'timing/data': 0.013401524163782597, 'timing/model': 1.4562301188707352, 'learning_rate/qwen_vl_interface': 4.7302755640813255e-06, 'learning_rate/action_model': 4.7302755640813256e-05, 'epoch': 1.84}) +[2026-06-16 02:59:40,530][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.04398807883262634, 'train/grad_norm_pre_clip': 4.666440010070801, 'timing/data': 0.014554259367287159, 'timing/model': 1.9217199115082622, 'learning_rate/qwen_vl_interface': 4.727248534119058e-06, 'learning_rate/action_model': 4.7272485341190574e-05, 'epoch': 1.84}) +[2026-06-16 02:59:46,163][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.06548108160495758, 'train/grad_norm_pre_clip': 7.861310958862305, 'timing/data': 0.013566424138844013, 'timing/model': 1.4586716778576374, 'learning_rate/qwen_vl_interface': 4.72422171904017e-06, 'learning_rate/action_model': 4.7242217190401697e-05, 'epoch': 1.84}) +[2026-06-16 02:59:52,161][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.051181893795728683, 'train/grad_norm_pre_clip': 5.229060173034668, 'timing/data': 0.014931862242519855, 'timing/model': 1.4671880519017577, 'learning_rate/qwen_vl_interface': 4.721195120088873e-06, 'learning_rate/action_model': 4.721195120088873e-05, 'epoch': 1.84}) +[2026-06-16 02:59:58,059][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.042619235813617706, 'train/grad_norm_pre_clip': 5.67531156539917, 'timing/data': 0.014082376845180988, 'timing/model': 1.4545679409056902, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 1.84}) +[2026-06-16 03:00:04,065][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.03951321914792061, 'train/grad_norm_pre_clip': 7.102616310119629, 'timing/data': 0.014209186658263206, 'timing/model': 1.468340233899653, 'learning_rate/qwen_vl_interface': 4.715142575545445e-06, 'learning_rate/action_model': 4.7151425755454454e-05, 'epoch': 1.84}) +[2026-06-16 03:00:09,984][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.008857677690684795, 'train/grad_norm_pre_clip': 1.2913402318954468, 'timing/data': 0.013472252525389194, 'timing/model': 1.792795111425221, 'learning_rate/qwen_vl_interface': 4.712116632441287e-06, 'learning_rate/action_model': 4.7121166324412866e-05, 'epoch': 1.84}) +[2026-06-16 03:00:15,894][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.10896967351436615, 'train/grad_norm_pre_clip': 7.025908946990967, 'timing/data': 0.014089756645262241, 'timing/model': 1.6897508762776852, 'learning_rate/qwen_vl_interface': 4.709090910440664e-06, 'learning_rate/action_model': 4.7090909104406634e-05, 'epoch': 1.84}) +[2026-06-16 03:00:21,574][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.010127680376172066, 'train/grad_norm_pre_clip': 2.5171124935150146, 'timing/data': 0.013282612897455692, 'timing/model': 1.4823942743241787, 'learning_rate/qwen_vl_interface': 4.7060654107873335e-06, 'learning_rate/action_model': 4.7060654107873335e-05, 'epoch': 1.84}) +[2026-06-16 03:00:27,569][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.5181098580360413, 'train/grad_norm_pre_clip': 16.285173416137695, 'timing/data': 0.01308842096477747, 'timing/model': 1.452710559591651, 'learning_rate/qwen_vl_interface': 4.7030401347249685e-06, 'learning_rate/action_model': 4.703040134724968e-05, 'epoch': 1.84}) +[2026-06-16 03:00:33,703][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.26392796635627747, 'train/grad_norm_pre_clip': 20.29252052307129, 'timing/data': 0.014120107516646385, 'timing/model': 1.506601370871067, 'learning_rate/qwen_vl_interface': 4.700015083497143e-06, 'learning_rate/action_model': 4.700015083497142e-05, 'epoch': 1.84}) +[2026-06-16 03:00:39,690][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.17494609951972961, 'train/grad_norm_pre_clip': 12.159418106079102, 'timing/data': 0.014797630719840527, 'timing/model': 1.418799890205264, 'learning_rate/qwen_vl_interface': 4.696990258347342e-06, 'learning_rate/action_model': 4.696990258347342e-05, 'epoch': 1.84}) +[2026-06-16 03:00:45,845][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.09806013107299805, 'train/grad_norm_pre_clip': 9.072113990783691, 'timing/data': 0.015395702794194221, 'timing/model': 1.9370603412389755, 'learning_rate/qwen_vl_interface': 4.693965660518957e-06, 'learning_rate/action_model': 4.693965660518957e-05, 'epoch': 1.84}) +[2026-06-16 03:00:51,793][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.021449366584420204, 'train/grad_norm_pre_clip': 3.3964955806732178, 'timing/data': 0.017254654318094254, 'timing/model': 1.7799405129626393, 'learning_rate/qwen_vl_interface': 4.690941291255285e-06, 'learning_rate/action_model': 4.6909412912552844e-05, 'epoch': 1.84}) +[2026-06-16 03:00:57,472][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.03401748836040497, 'train/grad_norm_pre_clip': 3.831033229827881, 'timing/data': 0.015701274387538433, 'timing/model': 1.4673022693023086, 'learning_rate/qwen_vl_interface': 4.68791715179953e-06, 'learning_rate/action_model': 4.68791715179953e-05, 'epoch': 1.85}) +[2026-06-16 03:01:03,560][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.21196094155311584, 'train/grad_norm_pre_clip': 12.625312805175781, 'timing/data': 0.01560040470212698, 'timing/model': 1.472237822599709, 'learning_rate/qwen_vl_interface': 4.684893243394804e-06, 'learning_rate/action_model': 4.6848932433948036e-05, 'epoch': 1.85}) +[2026-06-16 03:01:09,577][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.19620384275913239, 'train/grad_norm_pre_clip': 13.985647201538086, 'timing/data': 0.015798485837876797, 'timing/model': 1.4614065028727055, 'learning_rate/qwen_vl_interface': 4.68186956728412e-06, 'learning_rate/action_model': 4.68186956728412e-05, 'epoch': 1.85}) +[2026-06-16 03:01:15,770][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.06846725195646286, 'train/grad_norm_pre_clip': 10.040440559387207, 'timing/data': 0.014184577390551567, 'timing/model': 1.424177867360413, 'learning_rate/qwen_vl_interface': 4.678846124710395e-06, 'learning_rate/action_model': 4.678846124710394e-05, 'epoch': 1.85}) +[2026-06-16 03:01:21,900][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.2749203145503998, 'train/grad_norm_pre_clip': 17.49053955078125, 'timing/data': 0.012595819309353828, 'timing/model': 1.9089160524308681, 'learning_rate/qwen_vl_interface': 4.675822916916454e-06, 'learning_rate/action_model': 4.6758229169164536e-05, 'epoch': 1.85}) +[2026-06-16 03:01:27,570][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.11840851604938507, 'train/grad_norm_pre_clip': 5.78334903717041, 'timing/data': 0.013383232988417149, 'timing/model': 1.4471792792901397, 'learning_rate/qwen_vl_interface': 4.6727999451450236e-06, 'learning_rate/action_model': 4.672799945145023e-05, 'epoch': 1.85}) +[2026-06-16 03:01:33,685][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.09108682721853256, 'train/grad_norm_pre_clip': 10.49709415435791, 'timing/data': 0.014181047677993774, 'timing/model': 1.4834867250174284, 'learning_rate/qwen_vl_interface': 4.669777210638733e-06, 'learning_rate/action_model': 4.669777210638733e-05, 'epoch': 1.85}) +[2026-06-16 03:01:39,769][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.14098817110061646, 'train/grad_norm_pre_clip': 10.5938081741333, 'timing/data': 0.013155811466276646, 'timing/model': 1.4565473161637783, 'learning_rate/qwen_vl_interface': 4.666754714640115e-06, 'learning_rate/action_model': 4.666754714640115e-05, 'epoch': 1.85}) +[2026-06-16 03:01:45,873][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.048984117805957794, 'train/grad_norm_pre_clip': 4.483029842376709, 'timing/data': 0.012513748370110989, 'timing/model': 1.4550016447901726, 'learning_rate/qwen_vl_interface': 4.663732458391602e-06, 'learning_rate/action_model': 4.663732458391601e-05, 'epoch': 1.85}) +[2026-06-16 03:01:51,892][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.13442286849021912, 'train/grad_norm_pre_clip': 9.06310749053955, 'timing/data': 0.012887230142951012, 'timing/model': 1.42171726282686, 'learning_rate/qwen_vl_interface': 4.6607104431355295e-06, 'learning_rate/action_model': 4.66071044313553e-05, 'epoch': 1.85}) +[2026-06-16 03:01:57,933][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.21408477425575256, 'train/grad_norm_pre_clip': 14.504444122314453, 'timing/data': 0.012308516539633274, 'timing/model': 1.830618771724403, 'learning_rate/qwen_vl_interface': 4.657688670114135e-06, 'learning_rate/action_model': 4.657688670114135e-05, 'epoch': 1.85}) +[2026-06-16 03:02:03,573][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 0.011107176542282104, 'train/grad_norm_pre_clip': 2.7853012084960938, 'timing/data': 0.01215093582868576, 'timing/model': 1.4549616603180766, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 1.85}) +[2026-06-16 03:02:09,568][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.1332308053970337, 'train/grad_norm_pre_clip': 13.237700462341309, 'timing/data': 0.012517507188022137, 'timing/model': 1.4623023979365826, 'learning_rate/qwen_vl_interface': 4.651645855743825e-06, 'learning_rate/action_model': 4.651645855743825e-05, 'epoch': 1.85}) +[2026-06-16 03:02:15,679][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.12805916368961334, 'train/grad_norm_pre_clip': 6.665866374969482, 'timing/data': 0.012437197379767895, 'timing/model': 1.4753303974866867, 'learning_rate/qwen_vl_interface': 4.648624816878883e-06, 'learning_rate/action_model': 4.6486248168788825e-05, 'epoch': 1.85}) +[2026-06-16 03:02:21,788][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.0617549754679203, 'train/grad_norm_pre_clip': 11.168375968933105, 'timing/data': 0.012987500987946987, 'timing/model': 1.4776810072362423, 'learning_rate/qwen_vl_interface': 4.6456040252165615e-06, 'learning_rate/action_model': 4.645604025216562e-05, 'epoch': 1.85}) +[2026-06-16 03:02:27,792][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.030462466180324554, 'train/grad_norm_pre_clip': 2.293604612350464, 'timing/data': 0.01243604812771082, 'timing/model': 1.415719478391111, 'learning_rate/qwen_vl_interface': 4.642583481998596e-06, 'learning_rate/action_model': 4.642583481998596e-05, 'epoch': 1.85}) +[2026-06-16 03:02:33,824][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.21402601897716522, 'train/grad_norm_pre_clip': 11.651152610778809, 'timing/data': 0.01239805668592453, 'timing/model': 1.804714367724955, 'learning_rate/qwen_vl_interface': 4.639563188466616e-06, 'learning_rate/action_model': 4.6395631884666165e-05, 'epoch': 1.85}) +[2026-06-16 03:02:40,005][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.14621660113334656, 'train/grad_norm_pre_clip': 17.69053840637207, 'timing/data': 0.012133135460317135, 'timing/model': 1.9860623981803656, 'learning_rate/qwen_vl_interface': 4.636543145862152e-06, 'learning_rate/action_model': 4.6365431458621523e-05, 'epoch': 1.85}) +[2026-06-16 03:02:45,677][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.07059749960899353, 'train/grad_norm_pre_clip': 5.586893558502197, 'timing/data': 0.012695918790996075, 'timing/model': 1.4523719632998109, 'learning_rate/qwen_vl_interface': 4.633523355426629e-06, 'learning_rate/action_model': 4.6335233554266295e-05, 'epoch': 1.85}) +[2026-06-16 03:02:51,677][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.03029540926218033, 'train/grad_norm_pre_clip': 3.733773946762085, 'timing/data': 0.012209226377308369, 'timing/model': 1.4664213079959154, 'learning_rate/qwen_vl_interface': 4.6305038184013675e-06, 'learning_rate/action_model': 4.6305038184013675e-05, 'epoch': 1.85}) +[2026-06-16 03:02:57,685][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 0.09352849423885345, 'train/grad_norm_pre_clip': 7.003190517425537, 'timing/data': 0.01215294562280178, 'timing/model': 1.436479503288865, 'learning_rate/qwen_vl_interface': 4.627484536027589e-06, 'learning_rate/action_model': 4.627484536027588e-05, 'epoch': 1.85}) +[2026-06-16 03:03:03,690][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 0.13224861025810242, 'train/grad_norm_pre_clip': 12.583374977111816, 'timing/data': 0.013363651931285858, 'timing/model': 1.4366291845217347, 'learning_rate/qwen_vl_interface': 4.624465509546399e-06, 'learning_rate/action_model': 4.6244655095463986e-05, 'epoch': 1.85}) +[2026-06-16 03:03:09,759][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.09990411251783371, 'train/grad_norm_pre_clip': 10.957504272460938, 'timing/data': 0.01330520212650299, 'timing/model': 1.8654567133635283, 'learning_rate/qwen_vl_interface': 4.6214467401988106e-06, 'learning_rate/action_model': 4.621446740198811e-05, 'epoch': 1.86}) +[2026-06-16 03:03:15,815][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.07287747412919998, 'train/grad_norm_pre_clip': 5.515005111694336, 'timing/data': 0.013083890080451965, 'timing/model': 1.8137298934161663, 'learning_rate/qwen_vl_interface': 4.6184282292257245e-06, 'learning_rate/action_model': 4.618428229225725e-05, 'epoch': 1.86}) +[2026-06-16 03:03:21,488][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.052592966705560684, 'train/grad_norm_pre_clip': 9.829219818115234, 'timing/data': 0.013754204846918583, 'timing/model': 1.4909361712634563, 'learning_rate/qwen_vl_interface': 4.615409977867938e-06, 'learning_rate/action_model': 4.615409977867937e-05, 'epoch': 1.86}) +[2026-06-16 03:03:27,480][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.10051878541707993, 'train/grad_norm_pre_clip': 8.228937149047852, 'timing/data': 0.013985616154968739, 'timing/model': 1.470986313186586, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 1.86}) +[2026-06-16 03:03:33,478][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.08032998442649841, 'train/grad_norm_pre_clip': 7.969587802886963, 'timing/data': 0.014253617264330387, 'timing/model': 1.4686353290453553, 'learning_rate/qwen_vl_interface': 4.609374258960908e-06, 'learning_rate/action_model': 4.6093742589609074e-05, 'epoch': 1.86}) +[2026-06-16 03:03:39,471][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 0.03726201131939888, 'train/grad_norm_pre_clip': 5.249704837799072, 'timing/data': 0.015108511783182621, 'timing/model': 1.415887071751058, 'learning_rate/qwen_vl_interface': 4.60635679389272e-06, 'learning_rate/action_model': 4.60635679389272e-05, 'epoch': 1.86}) +[2026-06-16 03:03:45,510][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.19976097345352173, 'train/grad_norm_pre_clip': 17.24515724182129, 'timing/data': 0.01431064773350954, 'timing/model': 1.7905548363924026, 'learning_rate/qwen_vl_interface': 4.603339593401943e-06, 'learning_rate/action_model': 4.603339593401942e-05, 'epoch': 1.86}) +[2026-06-16 03:03:51,171][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.09939456731081009, 'train/grad_norm_pre_clip': 6.468282222747803, 'timing/data': 0.014328768476843834, 'timing/model': 1.4573894524946809, 'learning_rate/qwen_vl_interface': 4.600322658728832e-06, 'learning_rate/action_model': 4.600322658728832e-05, 'epoch': 1.86}) +[2026-06-16 03:03:57,168][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.3320012390613556, 'train/grad_norm_pre_clip': 12.253561019897461, 'timing/data': 0.013803933747112751, 'timing/model': 1.4578220546245575, 'learning_rate/qwen_vl_interface': 4.597305991113537e-06, 'learning_rate/action_model': 4.597305991113537e-05, 'epoch': 1.86}) +[2026-06-16 03:04:03,167][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 0.09124874323606491, 'train/grad_norm_pre_clip': 8.37890625, 'timing/data': 0.014246917329728603, 'timing/model': 1.4600280039012432, 'learning_rate/qwen_vl_interface': 4.5942895917960935e-06, 'learning_rate/action_model': 4.5942895917960935e-05, 'epoch': 1.86}) +[2026-06-16 03:04:09,189][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.19601237773895264, 'train/grad_norm_pre_clip': 10.722469329833984, 'timing/data': 0.013789962977170944, 'timing/model': 1.4536259500309825, 'learning_rate/qwen_vl_interface': 4.59127346201643e-06, 'learning_rate/action_model': 4.59127346201643e-05, 'epoch': 1.86}) +[2026-06-16 03:04:15,124][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.12248296290636063, 'train/grad_norm_pre_clip': 16.989412307739258, 'timing/data': 0.015043082647025585, 'timing/model': 1.4495795667171478, 'learning_rate/qwen_vl_interface': 4.588257603014364e-06, 'learning_rate/action_model': 4.588257603014364e-05, 'epoch': 1.86}) +[2026-06-16 03:04:21,313][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.07082221657037735, 'train/grad_norm_pre_clip': 9.563105583190918, 'timing/data': 0.014344347640872002, 'timing/model': 1.9876920077949762, 'learning_rate/qwen_vl_interface': 4.585242016029602e-06, 'learning_rate/action_model': 4.5852420160296015e-05, 'epoch': 1.86}) +[2026-06-16 03:04:26,973][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.06498216092586517, 'train/grad_norm_pre_clip': 6.100576877593994, 'timing/data': 0.013760866597294807, 'timing/model': 1.4477757941931486, 'learning_rate/qwen_vl_interface': 4.582226702301737e-06, 'learning_rate/action_model': 4.5822267023017375e-05, 'epoch': 1.86}) +[2026-06-16 03:04:32,982][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.05439209192991257, 'train/grad_norm_pre_clip': 3.1913976669311523, 'timing/data': 0.013001343235373497, 'timing/model': 1.4791777422651649, 'learning_rate/qwen_vl_interface': 4.57921166307025e-06, 'learning_rate/action_model': 4.57921166307025e-05, 'epoch': 1.86}) +[2026-06-16 03:04:39,081][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.07405363023281097, 'train/grad_norm_pre_clip': 11.031391143798828, 'timing/data': 0.014829141087830067, 'timing/model': 1.4532225374132395, 'learning_rate/qwen_vl_interface': 4.576196899574511e-06, 'learning_rate/action_model': 4.576196899574511e-05, 'epoch': 1.86}) +[2026-06-16 03:04:45,079][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.1620786041021347, 'train/grad_norm_pre_clip': 11.773213386535645, 'timing/data': 0.012875962071120739, 'timing/model': 1.4287949772551656, 'learning_rate/qwen_vl_interface': 4.573182413053775e-06, 'learning_rate/action_model': 4.5731824130537746e-05, 'epoch': 1.86}) +[2026-06-16 03:04:51,164][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.2720392048358917, 'train/grad_norm_pre_clip': 16.925411224365234, 'timing/data': 0.012383620254695415, 'timing/model': 1.439940026961267, 'learning_rate/qwen_vl_interface': 4.570168204747183e-06, 'learning_rate/action_model': 4.570168204747183e-05, 'epoch': 1.86}) +[2026-06-16 03:04:57,419][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.04128130525350571, 'train/grad_norm_pre_clip': 11.370721817016602, 'timing/data': 0.012505820952355862, 'timing/model': 1.9762383149936795, 'learning_rate/qwen_vl_interface': 4.567154275893762e-06, 'learning_rate/action_model': 4.5671542758937624e-05, 'epoch': 1.86}) +[2026-06-16 03:05:03,072][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.09711018204689026, 'train/grad_norm_pre_clip': 6.960793972015381, 'timing/data': 0.012326929718255997, 'timing/model': 1.4493643762543797, 'learning_rate/qwen_vl_interface': 4.564140627732426e-06, 'learning_rate/action_model': 4.564140627732426e-05, 'epoch': 1.86}) +[2026-06-16 03:05:09,178][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 0.06757018715143204, 'train/grad_norm_pre_clip': 10.202391624450684, 'timing/data': 0.012752882204949856, 'timing/model': 1.4543819790706038, 'learning_rate/qwen_vl_interface': 4.56112726150197e-06, 'learning_rate/action_model': 4.56112726150197e-05, 'epoch': 1.86}) +[2026-06-16 03:05:15,281][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.21206945180892944, 'train/grad_norm_pre_clip': 10.295886993408203, 'timing/data': 0.012176919728517532, 'timing/model': 1.4663052652031183, 'learning_rate/qwen_vl_interface': 4.558114178441076e-06, 'learning_rate/action_model': 4.558114178441076e-05, 'epoch': 1.87}) +[2026-06-16 03:05:21,374][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.020971663296222687, 'train/grad_norm_pre_clip': 1.7085469961166382, 'timing/data': 0.013617605902254581, 'timing/model': 1.4197067767381668, 'learning_rate/qwen_vl_interface': 4.555101379788308e-06, 'learning_rate/action_model': 4.555101379788308e-05, 'epoch': 1.87}) +[2026-06-16 03:05:27,507][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.2577996253967285, 'train/grad_norm_pre_clip': 11.80713939666748, 'timing/data': 0.01235897932201624, 'timing/model': 1.8692661747336388, 'learning_rate/qwen_vl_interface': 4.552088866782115e-06, 'learning_rate/action_model': 4.552088866782115e-05, 'epoch': 1.87}) +[2026-06-16 03:05:33,180][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.19538013637065887, 'train/grad_norm_pre_clip': 10.919873237609863, 'timing/data': 0.012321549467742443, 'timing/model': 1.4635577220469713, 'learning_rate/qwen_vl_interface': 4.5490766406608265e-06, 'learning_rate/action_model': 4.549076640660827e-05, 'epoch': 1.87}) +[2026-06-16 03:05:39,272][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.003801340702921152, 'train/grad_norm_pre_clip': 0.41890066862106323, 'timing/data': 0.011924638412892818, 'timing/model': 1.465064076706767, 'learning_rate/qwen_vl_interface': 4.546064702662655e-06, 'learning_rate/action_model': 4.546064702662655e-05, 'epoch': 1.87}) +[2026-06-16 03:05:45,275][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.13365304470062256, 'train/grad_norm_pre_clip': 10.715518951416016, 'timing/data': 0.012645991519093513, 'timing/model': 1.4541106205433607, 'learning_rate/qwen_vl_interface': 4.543053054025693e-06, 'learning_rate/action_model': 4.5430530540256926e-05, 'epoch': 1.87}) +[2026-06-16 03:05:51,277][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.15645864605903625, 'train/grad_norm_pre_clip': 6.348629951477051, 'timing/data': 0.014925481751561165, 'timing/model': 1.4581152265891433, 'learning_rate/qwen_vl_interface': 4.540041695987917e-06, 'learning_rate/action_model': 4.540041695987917e-05, 'epoch': 1.87}) +[2026-06-16 03:05:57,183][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 0.1006736159324646, 'train/grad_norm_pre_clip': 7.646398544311523, 'timing/data': 0.014816422015428543, 'timing/model': 1.4140102779492736, 'learning_rate/qwen_vl_interface': 4.537030629787182e-06, 'learning_rate/action_model': 4.5370306297871826e-05, 'epoch': 1.87}) +[2026-06-16 03:06:03,272][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.14781783521175385, 'train/grad_norm_pre_clip': 15.414941787719727, 'timing/data': 0.013317685574293137, 'timing/model': 1.8467149967327714, 'learning_rate/qwen_vl_interface': 4.534019856661226e-06, 'learning_rate/action_model': 4.534019856661226e-05, 'epoch': 1.87}) +[2026-06-16 03:06:08,989][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.04777991399168968, 'train/grad_norm_pre_clip': 3.694875717163086, 'timing/data': 0.012849431484937668, 'timing/model': 1.4733886700123549, 'learning_rate/qwen_vl_interface': 4.531009377847661e-06, 'learning_rate/action_model': 4.531009377847661e-05, 'epoch': 1.87}) +[2026-06-16 03:06:15,180][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.192031592130661, 'train/grad_norm_pre_clip': 14.387760162353516, 'timing/data': 0.01342807523906231, 'timing/model': 1.4683371828868985, 'learning_rate/qwen_vl_interface': 4.527999194583982e-06, 'learning_rate/action_model': 4.527999194583981e-05, 'epoch': 1.87}) +[2026-06-16 03:06:21,273][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.2215733379125595, 'train/grad_norm_pre_clip': 12.110505104064941, 'timing/data': 0.013246234506368637, 'timing/model': 1.462933512404561, 'learning_rate/qwen_vl_interface': 4.52498930810756e-06, 'learning_rate/action_model': 4.52498930810756e-05, 'epoch': 1.87}) +[2026-06-16 03:06:27,382][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.06621956080198288, 'train/grad_norm_pre_clip': 8.32046890258789, 'timing/data': 0.01273415144532919, 'timing/model': 1.4633934311568737, 'learning_rate/qwen_vl_interface': 4.521979719655648e-06, 'learning_rate/action_model': 4.5219797196556484e-05, 'epoch': 1.87}) +[2026-06-16 03:06:33,380][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.23091286420822144, 'train/grad_norm_pre_clip': 12.871789932250977, 'timing/data': 0.012935842387378216, 'timing/model': 1.4367489563301206, 'learning_rate/qwen_vl_interface': 4.518970430465377e-06, 'learning_rate/action_model': 4.518970430465376e-05, 'epoch': 1.87}) +[2026-06-16 03:06:39,598][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.06922045350074768, 'train/grad_norm_pre_clip': 8.320269584655762, 'timing/data': 0.013579227030277252, 'timing/model': 1.975721967406571, 'learning_rate/qwen_vl_interface': 4.515961441773745e-06, 'learning_rate/action_model': 4.515961441773745e-05, 'epoch': 1.87}) +[2026-06-16 03:06:45,378][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.13137538731098175, 'train/grad_norm_pre_clip': 10.076456069946289, 'timing/data': 0.01859207171946764, 'timing/model': 1.5315508944913745, 'learning_rate/qwen_vl_interface': 4.5129527548176385e-06, 'learning_rate/action_model': 4.512952754817638e-05, 'epoch': 1.87}) +[2026-06-16 03:06:51,373][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.2390293926000595, 'train/grad_norm_pre_clip': 7.1576151847839355, 'timing/data': 0.01267556194216013, 'timing/model': 1.4593096747994423, 'learning_rate/qwen_vl_interface': 4.509944370833813e-06, 'learning_rate/action_model': 4.509944370833813e-05, 'epoch': 1.87}) +[2026-06-16 03:06:57,374][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.15056541562080383, 'train/grad_norm_pre_clip': 7.4612812995910645, 'timing/data': 0.012389821000397205, 'timing/model': 1.4434328135102987, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 1.87}) +[2026-06-16 03:07:04,073][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 0.016402166336774826, 'train/grad_norm_pre_clip': 1.3528742790222168, 'timing/data': 0.04366833530366421, 'timing/model': 1.5567694064229727, 'learning_rate/qwen_vl_interface': 4.503928516729416e-06, 'learning_rate/action_model': 4.503928516729416e-05, 'epoch': 1.87}) +[2026-06-16 03:07:10,140][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.2878915071487427, 'train/grad_norm_pre_clip': 16.95102310180664, 'timing/data': 0.011897738091647625, 'timing/model': 1.8218800006434321, 'learning_rate/qwen_vl_interface': 4.500921049081734e-06, 'learning_rate/action_model': 4.500921049081734e-05, 'epoch': 1.87}) +[2026-06-16 03:07:16,001][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 0.09549568593502045, 'train/grad_norm_pre_clip': 6.067723274230957, 'timing/data': 0.014116710051894188, 'timing/model': 1.7011013710871339, 'learning_rate/qwen_vl_interface': 4.497913889352113e-06, 'learning_rate/action_model': 4.4979138893521126e-05, 'epoch': 1.87}) +[2026-06-16 03:07:21,670][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.299002468585968, 'train/grad_norm_pre_clip': 16.310829162597656, 'timing/data': 0.014443420805037022, 'timing/model': 1.4607243305072188, 'learning_rate/qwen_vl_interface': 4.494907038776684e-06, 'learning_rate/action_model': 4.4949070387766836e-05, 'epoch': 1.87}) +[2026-06-16 03:18:41,671][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.22705546021461487, 'train/grad_norm_pre_clip': 11.865607261657715, 'eval/loss': 0.06232650578022003, 'eval/flap_precision': 0.8536585365853658, 'eval/flap_recall': 0.7636363636363637, 'eval/flap_f1': 0.8061420345489444, 'eval/flap_rate_model': 0.06833333333333333, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.8061420345489445, 'timing/data': 0.01405974943190813, 'timing/model': 1.4649951783940196, 'learning_rate/qwen_vl_interface': 4.49190049859145e-06, 'learning_rate/action_model': 4.49190049859145e-05, 'epoch': 1.88}) +[2026-06-16 03:18:41,672][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4100_state +[2026-06-16 03:18:41,679][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 03:18:49,359][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4100_state/model.safetensors +[2026-06-16 03:19:00,778][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4100_state/optimizer.bin +[2026-06-16 03:19:00,779][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4100_state/sampler.bin +[2026-06-16 03:19:00,779][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4100_state/sampler_1.bin +[2026-06-16 03:19:00,781][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4100_state/random_states_0.pkl +[2026-06-16 03:19:00,943][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4100 +[2026-06-16 03:19:00,943][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 03:19:00,958][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 03:19:06,747][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.06047259271144867, 'train/grad_norm_pre_clip': 6.292330741882324, 'timing/data': 0.010692695155739784, 'timing/model': 1.377823949791491, 'learning_rate/qwen_vl_interface': 4.488894270032284e-06, 'learning_rate/action_model': 4.4888942700322834e-05, 'epoch': 1.88}) +[2026-06-16 03:19:12,460][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.18959073722362518, 'train/grad_norm_pre_clip': 8.932677268981934, 'timing/data': 0.013902592472732067, 'timing/model': 1.457084821537137, 'learning_rate/qwen_vl_interface': 4.485888354334933e-06, 'learning_rate/action_model': 4.4858883543349334e-05, 'epoch': 1.88}) +[2026-06-16 03:19:18,248][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.10261277854442596, 'train/grad_norm_pre_clip': 8.956907272338867, 'timing/data': 0.014745667576789856, 'timing/model': 1.454679898917675, 'learning_rate/qwen_vl_interface': 4.482882752735017e-06, 'learning_rate/action_model': 4.482882752735017e-05, 'epoch': 1.88}) +[2026-06-16 03:19:24,133][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.09207412600517273, 'train/grad_norm_pre_clip': 5.8562750816345215, 'timing/data': 0.01169454026967287, 'timing/model': 1.3875826830044389, 'learning_rate/qwen_vl_interface': 4.479877466468027e-06, 'learning_rate/action_model': 4.479877466468026e-05, 'epoch': 1.88}) +[2026-06-16 03:19:30,151][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.15483960509300232, 'train/grad_norm_pre_clip': 6.457660675048828, 'timing/data': 0.015192978084087372, 'timing/model': 1.8487141644582152, 'learning_rate/qwen_vl_interface': 4.476872496769318e-06, 'learning_rate/action_model': 4.476872496769318e-05, 'epoch': 1.88}) +[2026-06-16 03:19:35,731][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.191499724984169, 'train/grad_norm_pre_clip': 15.600495338439941, 'timing/data': 0.012436014600098133, 'timing/model': 1.4452002542093396, 'learning_rate/qwen_vl_interface': 4.473867844874122e-06, 'learning_rate/action_model': 4.473867844874122e-05, 'epoch': 1.88}) +[2026-06-16 03:19:41,647][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.1251695305109024, 'train/grad_norm_pre_clip': 5.821062088012695, 'timing/data': 0.011584930121898651, 'timing/model': 1.4490882456302643, 'learning_rate/qwen_vl_interface': 4.470863512017539e-06, 'learning_rate/action_model': 4.470863512017539e-05, 'epoch': 1.88}) +[2026-06-16 03:19:47,566][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.03601660579442978, 'train/grad_norm_pre_clip': 2.7153637409210205, 'timing/data': 0.012597015127539635, 'timing/model': 1.4633311610668898, 'learning_rate/qwen_vl_interface': 4.4678594994345344e-06, 'learning_rate/action_model': 4.467859499434534e-05, 'epoch': 1.88}) +[2026-06-16 03:19:53,667][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.15486252307891846, 'train/grad_norm_pre_clip': 7.973753929138184, 'timing/data': 0.013405990786850452, 'timing/model': 1.452832406386733, 'learning_rate/qwen_vl_interface': 4.4648558083599475e-06, 'learning_rate/action_model': 4.464855808359947e-05, 'epoch': 1.88}) +[2026-06-16 03:19:59,751][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.04591641202569008, 'train/grad_norm_pre_clip': 3.241851568222046, 'timing/data': 0.014402165077626705, 'timing/model': 1.4318466000258923, 'learning_rate/qwen_vl_interface': 4.461852440028479e-06, 'learning_rate/action_model': 4.461852440028479e-05, 'epoch': 1.88}) +[2026-06-16 03:20:05,776][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.18350934982299805, 'train/grad_norm_pre_clip': 13.215995788574219, 'timing/data': 0.013813161291182041, 'timing/model': 1.8862741505727172, 'learning_rate/qwen_vl_interface': 4.458849395674702e-06, 'learning_rate/action_model': 4.458849395674701e-05, 'epoch': 1.88}) +[2026-06-16 03:20:11,476][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.15787094831466675, 'train/grad_norm_pre_clip': 9.603824615478516, 'timing/data': 0.014174972660839558, 'timing/model': 1.4756411397829652, 'learning_rate/qwen_vl_interface': 4.455846676533052e-06, 'learning_rate/action_model': 4.4558466765330525e-05, 'epoch': 1.88}) +[2026-06-16 03:20:17,678][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.023881029337644577, 'train/grad_norm_pre_clip': 2.6239914894104004, 'timing/data': 0.014276684261858463, 'timing/model': 1.466979155316949, 'learning_rate/qwen_vl_interface': 4.452844283837838e-06, 'learning_rate/action_model': 4.452844283837838e-05, 'epoch': 1.88}) +[2026-06-16 03:20:23,790][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.12283492833375931, 'train/grad_norm_pre_clip': 8.420942306518555, 'timing/data': 0.013349419459700584, 'timing/model': 1.4706393973901868, 'learning_rate/qwen_vl_interface': 4.449842218823228e-06, 'learning_rate/action_model': 4.449842218823228e-05, 'epoch': 1.88}) +[2026-06-16 03:20:29,975][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.11614440381526947, 'train/grad_norm_pre_clip': 7.849577903747559, 'timing/data': 0.013796071521937847, 'timing/model': 1.4342298889532685, 'learning_rate/qwen_vl_interface': 4.446840482723256e-06, 'learning_rate/action_model': 4.446840482723255e-05, 'epoch': 1.88}) +[2026-06-16 03:20:36,088][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.17027226090431213, 'train/grad_norm_pre_clip': 11.592658042907715, 'timing/data': 0.017131579108536243, 'timing/model': 1.4930732054635882, 'learning_rate/qwen_vl_interface': 4.443839076771823e-06, 'learning_rate/action_model': 4.443839076771823e-05, 'epoch': 1.88}) +[2026-06-16 03:20:42,263][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.015939297154545784, 'train/grad_norm_pre_clip': 1.6062040328979492, 'timing/data': 0.013063627295196056, 'timing/model': 1.954271512106061, 'learning_rate/qwen_vl_interface': 4.4408380022026945e-06, 'learning_rate/action_model': 4.440838002202694e-05, 'epoch': 1.88}) +[2026-06-16 03:20:48,405][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.3547644019126892, 'train/grad_norm_pre_clip': 11.184652328491211, 'timing/data': 0.013813041150569916, 'timing/model': 1.880410141311586, 'learning_rate/qwen_vl_interface': 4.437837260249499e-06, 'learning_rate/action_model': 4.437837260249499e-05, 'epoch': 1.88}) +[2026-06-16 03:20:54,059][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.20572058856487274, 'train/grad_norm_pre_clip': 15.53005313873291, 'timing/data': 0.014090974815189838, 'timing/model': 1.430823435075581, 'learning_rate/qwen_vl_interface': 4.434836852145728e-06, 'learning_rate/action_model': 4.434836852145728e-05, 'epoch': 1.88}) +[2026-06-16 03:21:00,168][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.10283765196800232, 'train/grad_norm_pre_clip': 5.831929683685303, 'timing/data': 0.013881421647965908, 'timing/model': 1.454910601489246, 'learning_rate/qwen_vl_interface': 4.431836779124736e-06, 'learning_rate/action_model': 4.431836779124735e-05, 'epoch': 1.88}) +[2026-06-16 03:21:06,267][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.019146421924233437, 'train/grad_norm_pre_clip': 1.9875749349594116, 'timing/data': 0.014525815844535828, 'timing/model': 1.4657117994502187, 'learning_rate/qwen_vl_interface': 4.428837042419738e-06, 'learning_rate/action_model': 4.428837042419738e-05, 'epoch': 1.88}) +[2026-06-16 03:21:12,473][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.29839086532592773, 'train/grad_norm_pre_clip': 12.647446632385254, 'timing/data': 0.013920472003519535, 'timing/model': 1.4312109602615237, 'learning_rate/qwen_vl_interface': 4.425837643263815e-06, 'learning_rate/action_model': 4.425837643263814e-05, 'epoch': 1.89}) +[2026-06-16 03:21:18,585][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.13647142052650452, 'train/grad_norm_pre_clip': 7.41794490814209, 'timing/data': 0.013058528304100037, 'timing/model': 1.872582376934588, 'learning_rate/qwen_vl_interface': 4.4228385828899045e-06, 'learning_rate/action_model': 4.4228385828899046e-05, 'epoch': 1.89}) +[2026-06-16 03:21:24,673][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.251237154006958, 'train/grad_norm_pre_clip': 9.721428871154785, 'timing/data': 0.014462664723396301, 'timing/model': 1.8577856523916125, 'learning_rate/qwen_vl_interface': 4.419839862530809e-06, 'learning_rate/action_model': 4.419839862530809e-05, 'epoch': 1.89}) +[2026-06-16 03:21:30,381][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.09864269196987152, 'train/grad_norm_pre_clip': 7.2816996574401855, 'timing/data': 0.01417368371039629, 'timing/model': 1.4413385465741158, 'learning_rate/qwen_vl_interface': 4.416841483419188e-06, 'learning_rate/action_model': 4.416841483419188e-05, 'epoch': 1.89}) +[2026-06-16 03:21:36,594][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.19957081973552704, 'train/grad_norm_pre_clip': 12.303179740905762, 'timing/data': 0.01449782494455576, 'timing/model': 1.4873916879296303, 'learning_rate/qwen_vl_interface': 4.413843446787561e-06, 'learning_rate/action_model': 4.413843446787561e-05, 'epoch': 1.89}) +[2026-06-16 03:21:42,678][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.10318654030561447, 'train/grad_norm_pre_clip': 9.422056198120117, 'timing/data': 0.014388524927198887, 'timing/model': 1.4605023385956883, 'learning_rate/qwen_vl_interface': 4.41084575386831e-06, 'learning_rate/action_model': 4.4108457538683093e-05, 'epoch': 1.89}) +[2026-06-16 03:21:48,782][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.1500617414712906, 'train/grad_norm_pre_clip': 9.04401969909668, 'timing/data': 0.014748665504157543, 'timing/model': 1.4672080846503377, 'learning_rate/qwen_vl_interface': 4.40784840589367e-06, 'learning_rate/action_model': 4.4078484058936706e-05, 'epoch': 1.89}) +[2026-06-16 03:21:54,793][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.07055194675922394, 'train/grad_norm_pre_clip': 4.009474277496338, 'timing/data': 0.01446380466222763, 'timing/model': 1.4309304868802428, 'learning_rate/qwen_vl_interface': 4.404851404095742e-06, 'learning_rate/action_model': 4.404851404095742e-05, 'epoch': 1.89}) +[2026-06-16 03:22:00,867][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.08161208033561707, 'train/grad_norm_pre_clip': 4.871424674987793, 'timing/data': 0.014141413383185863, 'timing/model': 1.8459312478080392, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 1.89}) +[2026-06-16 03:22:06,589][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.12169312685728073, 'train/grad_norm_pre_clip': 5.543557643890381, 'timing/data': 0.014407025650143623, 'timing/model': 1.4700661217793822, 'learning_rate/qwen_vl_interface': 4.398858443957686e-06, 'learning_rate/action_model': 4.398858443957686e-05, 'epoch': 1.89}) +[2026-06-16 03:22:12,687][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.11909856647253036, 'train/grad_norm_pre_clip': 8.168880462646484, 'timing/data': 0.012225452810525894, 'timing/model': 1.4499961994588375, 'learning_rate/qwen_vl_interface': 4.395862488081038e-06, 'learning_rate/action_model': 4.395862488081038e-05, 'epoch': 1.89}) +[2026-06-16 03:22:18,664][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.1896459460258484, 'train/grad_norm_pre_clip': 8.359343528747559, 'timing/data': 0.012271823361515999, 'timing/model': 1.451388657093048, 'learning_rate/qwen_vl_interface': 4.392866883308059e-06, 'learning_rate/action_model': 4.392866883308059e-05, 'epoch': 1.89}) +[2026-06-16 03:22:24,776][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.1346873939037323, 'train/grad_norm_pre_clip': 7.404749393463135, 'timing/data': 0.014885677956044674, 'timing/model': 1.4868143228814006, 'learning_rate/qwen_vl_interface': 4.389871630870127e-06, 'learning_rate/action_model': 4.389871630870127e-05, 'epoch': 1.89}) +[2026-06-16 03:22:30,687][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.10677467286586761, 'train/grad_norm_pre_clip': 4.957995414733887, 'timing/data': 0.018144014291465282, 'timing/model': 1.4277977561578155, 'learning_rate/qwen_vl_interface': 4.386876731998478e-06, 'learning_rate/action_model': 4.386876731998478e-05, 'epoch': 1.89}) +[2026-06-16 03:22:36,719][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.17912578582763672, 'train/grad_norm_pre_clip': 9.395930290222168, 'timing/data': 0.015178369358181953, 'timing/model': 1.8156744968146086, 'learning_rate/qwen_vl_interface': 4.3838821879242006e-06, 'learning_rate/action_model': 4.3838821879242e-05, 'epoch': 1.89}) +[2026-06-16 03:22:42,370][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.08688101917505264, 'train/grad_norm_pre_clip': 3.7081706523895264, 'timing/data': 0.014080292545258999, 'timing/model': 1.4609990008175373, 'learning_rate/qwen_vl_interface': 4.38088799987824e-06, 'learning_rate/action_model': 4.38088799987824e-05, 'epoch': 1.89}) +[2026-06-16 03:22:48,473][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.13576850295066833, 'train/grad_norm_pre_clip': 6.910331726074219, 'timing/data': 0.014453104697167873, 'timing/model': 1.4563574772328138, 'learning_rate/qwen_vl_interface': 4.3778941690913935e-06, 'learning_rate/action_model': 4.3778941690913934e-05, 'epoch': 1.89}) +[2026-06-16 03:22:54,567][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.19693870842456818, 'train/grad_norm_pre_clip': 6.46185827255249, 'timing/data': 0.013492400757968426, 'timing/model': 1.4633962148800492, 'learning_rate/qwen_vl_interface': 4.374900696794311e-06, 'learning_rate/action_model': 4.374900696794311e-05, 'epoch': 1.89}) +[2026-06-16 03:23:00,580][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.05731220915913582, 'train/grad_norm_pre_clip': 3.169828176498413, 'timing/data': 0.014113222248852253, 'timing/model': 1.411608955822885, 'learning_rate/qwen_vl_interface': 4.371907584217499e-06, 'learning_rate/action_model': 4.371907584217498e-05, 'epoch': 1.89}) +[2026-06-16 03:23:06,649][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.06347320973873138, 'train/grad_norm_pre_clip': 3.1863484382629395, 'timing/data': 0.014057443477213383, 'timing/model': 1.8402002910152078, 'learning_rate/qwen_vl_interface': 4.368914832591309e-06, 'learning_rate/action_model': 4.368914832591309e-05, 'epoch': 1.89}) +[2026-06-16 03:23:12,271][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.04901813715696335, 'train/grad_norm_pre_clip': 2.7462918758392334, 'timing/data': 0.013388139195740223, 'timing/model': 1.442254108376801, 'learning_rate/qwen_vl_interface': 4.365922443145949e-06, 'learning_rate/action_model': 4.365922443145949e-05, 'epoch': 1.89}) +[2026-06-16 03:23:18,278][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.09676554054021835, 'train/grad_norm_pre_clip': 5.589303970336914, 'timing/data': 0.0149814672768116, 'timing/model': 1.4425580613315105, 'learning_rate/qwen_vl_interface': 4.362930417111479e-06, 'learning_rate/action_model': 4.362930417111479e-05, 'epoch': 1.89}) +[2026-06-16 03:23:24,280][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.03922124207019806, 'train/grad_norm_pre_clip': 5.284456729888916, 'timing/data': 0.013691971078515053, 'timing/model': 1.4571358365938067, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 1.9}) +[2026-06-16 03:23:30,275][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.041165612637996674, 'train/grad_norm_pre_clip': 3.384406805038452, 'timing/data': 0.014860907569527626, 'timing/model': 1.4518585307523608, 'learning_rate/qwen_vl_interface': 4.3569474601946934e-06, 'learning_rate/action_model': 4.356947460194694e-05, 'epoch': 1.9}) +[2026-06-16 03:23:36,375][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.18289989233016968, 'train/grad_norm_pre_clip': 6.828566551208496, 'timing/data': 0.01944459229707718, 'timing/model': 1.4891377948224545, 'learning_rate/qwen_vl_interface': 4.3539565317717446e-06, 'learning_rate/action_model': 4.353956531771745e-05, 'epoch': 1.9}) +[2026-06-16 03:23:42,485][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.2890316843986511, 'train/grad_norm_pre_clip': 12.809791564941406, 'timing/data': 0.01339282002300024, 'timing/model': 1.867268088273704, 'learning_rate/qwen_vl_interface': 4.350965971678419e-06, 'learning_rate/action_model': 4.3509659716784185e-05, 'epoch': 1.9}) +[2026-06-16 03:23:48,169][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.10608799755573273, 'train/grad_norm_pre_clip': 4.2860870361328125, 'timing/data': 0.01512167975306511, 'timing/model': 1.4543541762977839, 'learning_rate/qwen_vl_interface': 4.3479757811440225e-06, 'learning_rate/action_model': 4.347975781144022e-05, 'epoch': 1.9}) +[2026-06-16 03:23:54,272][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.3372325599193573, 'train/grad_norm_pre_clip': 8.381446838378906, 'timing/data': 0.013174789026379585, 'timing/model': 1.4679108066484332, 'learning_rate/qwen_vl_interface': 4.34498596139771e-06, 'learning_rate/action_model': 4.34498596139771e-05, 'epoch': 1.9}) +[2026-06-16 03:24:00,292][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.05909288674592972, 'train/grad_norm_pre_clip': 2.7304606437683105, 'timing/data': 0.014338535256683826, 'timing/model': 1.475629998371005, 'learning_rate/qwen_vl_interface': 4.341996513668483e-06, 'learning_rate/action_model': 4.3419965136684824e-05, 'epoch': 1.9}) +[2026-06-16 03:24:06,570][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.059140510857105255, 'train/grad_norm_pre_clip': 3.7132534980773926, 'timing/data': 0.016097364015877247, 'timing/model': 1.5605256129056215, 'learning_rate/qwen_vl_interface': 4.33900743918519e-06, 'learning_rate/action_model': 4.33900743918519e-05, 'epoch': 1.9}) +[2026-06-16 03:24:12,691][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.0596822053194046, 'train/grad_norm_pre_clip': 3.214625597000122, 'timing/data': 0.015725191682577133, 'timing/model': 1.4652534760534763, 'learning_rate/qwen_vl_interface': 4.336018739176529e-06, 'learning_rate/action_model': 4.33601873917653e-05, 'epoch': 1.9}) +[2026-06-16 03:24:18,842][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.11259188503026962, 'train/grad_norm_pre_clip': 3.6280195713043213, 'timing/data': 0.013474110513925552, 'timing/model': 1.9083726489916444, 'learning_rate/qwen_vl_interface': 4.333030414871041e-06, 'learning_rate/action_model': 4.33303041487104e-05, 'epoch': 1.9}) +[2026-06-16 03:24:24,505][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.09483455866575241, 'train/grad_norm_pre_clip': 6.057494640350342, 'timing/data': 0.014330355450510979, 'timing/model': 1.4928843909874558, 'learning_rate/qwen_vl_interface': 4.330042467497111e-06, 'learning_rate/action_model': 4.330042467497111e-05, 'epoch': 1.9}) +[2026-06-16 03:24:30,473][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.09105934202671051, 'train/grad_norm_pre_clip': 4.888071537017822, 'timing/data': 0.013716542162001133, 'timing/model': 1.4589013243094087, 'learning_rate/qwen_vl_interface': 4.327054898282973e-06, 'learning_rate/action_model': 4.3270548982829727e-05, 'epoch': 1.9}) +[2026-06-16 03:24:36,605][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.12095377594232559, 'train/grad_norm_pre_clip': 9.535478591918945, 'timing/data': 0.012679466046392918, 'timing/model': 1.4381686095148325, 'learning_rate/qwen_vl_interface': 4.324067708456703e-06, 'learning_rate/action_model': 4.3240677084567036e-05, 'epoch': 1.9}) +[2026-06-16 03:24:42,674][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.05317765474319458, 'train/grad_norm_pre_clip': 4.338973522186279, 'timing/data': 0.014776197262108326, 'timing/model': 1.4682575957849622, 'learning_rate/qwen_vl_interface': 4.321080899246225e-06, 'learning_rate/action_model': 4.321080899246225e-05, 'epoch': 1.9}) +[2026-06-16 03:24:48,764][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.03319863602519035, 'train/grad_norm_pre_clip': 2.146873950958252, 'timing/data': 0.013264238834381104, 'timing/model': 1.8634573798626661, 'learning_rate/qwen_vl_interface': 4.318094471879299e-06, 'learning_rate/action_model': 4.3180944718792985e-05, 'epoch': 1.9}) +[2026-06-16 03:24:54,852][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.3061707317829132, 'train/grad_norm_pre_clip': 10.289488792419434, 'timing/data': 0.014697707258164883, 'timing/model': 1.793878836557269, 'learning_rate/qwen_vl_interface': 4.315108427583535e-06, 'learning_rate/action_model': 4.315108427583535e-05, 'epoch': 1.9}) +[2026-06-16 03:25:00,496][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.12871628999710083, 'train/grad_norm_pre_clip': 6.28822135925293, 'timing/data': 0.014076132327318192, 'timing/model': 1.485965589992702, 'learning_rate/qwen_vl_interface': 4.312122767586381e-06, 'learning_rate/action_model': 4.3121227675863815e-05, 'epoch': 1.9}) +[2026-06-16 03:25:06,588][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.16773056983947754, 'train/grad_norm_pre_clip': 7.058452129364014, 'timing/data': 0.014048492535948753, 'timing/model': 1.4616285935044289, 'learning_rate/qwen_vl_interface': 4.3091374931151324e-06, 'learning_rate/action_model': 4.309137493115132e-05, 'epoch': 1.9}) +[2026-06-16 03:25:12,676][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.19519765675067902, 'train/grad_norm_pre_clip': 9.08703327178955, 'timing/data': 0.0123304957523942, 'timing/model': 1.4644415592774749, 'learning_rate/qwen_vl_interface': 4.306152605396921e-06, 'learning_rate/action_model': 4.306152605396921e-05, 'epoch': 1.9}) +[2026-06-16 03:25:18,782][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.17361968755722046, 'train/grad_norm_pre_clip': 8.382031440734863, 'timing/data': 0.01489193830639124, 'timing/model': 1.4690654138103127, 'learning_rate/qwen_vl_interface': 4.30316810565872e-06, 'learning_rate/action_model': 4.3031681056587206e-05, 'epoch': 1.9}) +[2026-06-16 03:25:24,862][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.14098578691482544, 'train/grad_norm_pre_clip': 7.7838263511657715, 'timing/data': 0.013616090640425682, 'timing/model': 1.836639334447682, 'learning_rate/qwen_vl_interface': 4.300183995127347e-06, 'learning_rate/action_model': 4.300183995127346e-05, 'epoch': 1.9}) +[2026-06-16 03:25:30,957][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.04118899628520012, 'train/grad_norm_pre_clip': 4.576495170593262, 'timing/data': 0.013976133428514004, 'timing/model': 1.8429170176386833, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 1.9}) +[2026-06-16 03:25:36,672][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.10109004378318787, 'train/grad_norm_pre_clip': 8.890684127807617, 'timing/data': 0.01356390118598938, 'timing/model': 1.462195510044694, 'learning_rate/qwen_vl_interface': 4.29421694659154e-06, 'learning_rate/action_model': 4.294216946591539e-05, 'epoch': 1.91}) +[2026-06-16 03:25:42,679][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.05310217663645744, 'train/grad_norm_pre_clip': 4.342898845672607, 'timing/data': 0.013490019366145134, 'timing/model': 1.4724845755845308, 'learning_rate/qwen_vl_interface': 4.2912340110399345e-06, 'learning_rate/action_model': 4.291234011039934e-05, 'epoch': 1.91}) +[2026-06-16 03:25:48,780][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.3620870113372803, 'train/grad_norm_pre_clip': 15.559165000915527, 'timing/data': 0.014325824566185474, 'timing/model': 1.4508679220452905, 'learning_rate/qwen_vl_interface': 4.2882514696008125e-06, 'learning_rate/action_model': 4.2882514696008125e-05, 'epoch': 1.91}) +[2026-06-16 03:25:54,899][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.2525098919868469, 'train/grad_norm_pre_clip': 8.14076042175293, 'timing/data': 0.01346237026154995, 'timing/model': 1.4243007535114884, 'learning_rate/qwen_vl_interface': 4.2852693235001815e-06, 'learning_rate/action_model': 4.2852693235001815e-05, 'epoch': 1.91}) +[2026-06-16 03:26:01,171][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.2864378094673157, 'train/grad_norm_pre_clip': 5.369912624359131, 'timing/data': 0.014772027730941772, 'timing/model': 2.0554498825222254, 'learning_rate/qwen_vl_interface': 4.2822875739638915e-06, 'learning_rate/action_model': 4.282287573963891e-05, 'epoch': 1.91}) +[2026-06-16 03:26:06,896][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.029785582795739174, 'train/grad_norm_pre_clip': 2.5287652015686035, 'timing/data': 0.014100424014031887, 'timing/model': 1.4507169835269451, 'learning_rate/qwen_vl_interface': 4.279306222217625e-06, 'learning_rate/action_model': 4.279306222217626e-05, 'epoch': 1.91}) +[2026-06-16 03:26:13,176][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.06529469043016434, 'train/grad_norm_pre_clip': 5.440051078796387, 'timing/data': 0.015034887939691544, 'timing/model': 1.4321351572871208, 'learning_rate/qwen_vl_interface': 4.276325269486906e-06, 'learning_rate/action_model': 4.276325269486906e-05, 'epoch': 1.91}) +[2026-06-16 03:26:19,280][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.11117801815271378, 'train/grad_norm_pre_clip': 8.071758270263672, 'timing/data': 0.014042682945728302, 'timing/model': 1.4681009957566857, 'learning_rate/qwen_vl_interface': 4.273344716997089e-06, 'learning_rate/action_model': 4.273344716997089e-05, 'epoch': 1.91}) +[2026-06-16 03:26:25,480][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.06843789666891098, 'train/grad_norm_pre_clip': 9.239282608032227, 'timing/data': 0.014048074372112751, 'timing/model': 1.4402655502781272, 'learning_rate/qwen_vl_interface': 4.270364565973368e-06, 'learning_rate/action_model': 4.270364565973367e-05, 'epoch': 1.91}) +[2026-06-16 03:26:31,577][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 0.056557126343250275, 'train/grad_norm_pre_clip': 5.965575218200684, 'timing/data': 0.013866453431546688, 'timing/model': 1.488724173977971, 'learning_rate/qwen_vl_interface': 4.267384817640768e-06, 'learning_rate/action_model': 4.2673848176407685e-05, 'epoch': 1.91}) +[2026-06-16 03:26:37,687][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.04394827410578728, 'train/grad_norm_pre_clip': 2.9110262393951416, 'timing/data': 0.013744642026722431, 'timing/model': 1.8766467534005642, 'learning_rate/qwen_vl_interface': 4.264405473224155e-06, 'learning_rate/action_model': 4.2644054732241546e-05, 'epoch': 1.91}) +[2026-06-16 03:26:43,381][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.05492295324802399, 'train/grad_norm_pre_clip': 4.17379093170166, 'timing/data': 0.013397620059549809, 'timing/model': 1.4627590896561742, 'learning_rate/qwen_vl_interface': 4.261426533948221e-06, 'learning_rate/action_model': 4.2614265339482213e-05, 'epoch': 1.91}) +[2026-06-16 03:26:49,486][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.2837532162666321, 'train/grad_norm_pre_clip': 9.058777809143066, 'timing/data': 0.014492275193333626, 'timing/model': 1.4584033079445362, 'learning_rate/qwen_vl_interface': 4.258448001037499e-06, 'learning_rate/action_model': 4.2584480010374985e-05, 'epoch': 1.91}) +[2026-06-16 03:26:55,577][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 0.1370777040719986, 'train/grad_norm_pre_clip': 7.1902852058410645, 'timing/data': 0.014047742821276188, 'timing/model': 1.4626983804628253, 'learning_rate/qwen_vl_interface': 4.255469875716348e-06, 'learning_rate/action_model': 4.255469875716348e-05, 'epoch': 1.91}) +[2026-06-16 03:27:01,579][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.14978238940238953, 'train/grad_norm_pre_clip': 16.85173988342285, 'timing/data': 0.014312255196273327, 'timing/model': 1.4347942946478724, 'learning_rate/qwen_vl_interface': 4.252492159208966e-06, 'learning_rate/action_model': 4.252492159208966e-05, 'epoch': 1.91}) +[2026-06-16 03:27:07,701][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.07882598042488098, 'train/grad_norm_pre_clip': 6.177270889282227, 'timing/data': 0.016262315213680267, 'timing/model': 1.5051012830808759, 'learning_rate/qwen_vl_interface': 4.2495148527393766e-06, 'learning_rate/action_model': 4.2495148527393766e-05, 'epoch': 1.91}) +[2026-06-16 03:27:13,738][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.06359872221946716, 'train/grad_norm_pre_clip': 4.861687183380127, 'timing/data': 0.014073693193495274, 'timing/model': 1.8118688594549894, 'learning_rate/qwen_vl_interface': 4.246537957531441e-06, 'learning_rate/action_model': 4.2465379575314404e-05, 'epoch': 1.91}) +[2026-06-16 03:27:19,678][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.19402527809143066, 'train/grad_norm_pre_clip': 8.0787353515625, 'timing/data': 0.014742317609488964, 'timing/model': 1.7680721106007695, 'learning_rate/qwen_vl_interface': 4.243561474808846e-06, 'learning_rate/action_model': 4.2435614748088456e-05, 'epoch': 1.91}) +[2026-06-16 03:27:25,378][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.020909231156110764, 'train/grad_norm_pre_clip': 1.500527262687683, 'timing/data': 0.013643370941281319, 'timing/model': 1.4637677492573857, 'learning_rate/qwen_vl_interface': 4.240585405795112e-06, 'learning_rate/action_model': 4.240585405795112e-05, 'epoch': 1.91}) +[2026-06-16 03:27:31,477][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.12452398240566254, 'train/grad_norm_pre_clip': 8.35338020324707, 'timing/data': 0.015126888640224934, 'timing/model': 1.4545440012589097, 'learning_rate/qwen_vl_interface': 4.237609751713589e-06, 'learning_rate/action_model': 4.2376097517135896e-05, 'epoch': 1.91}) +[2026-06-16 03:27:37,577][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.22939732670783997, 'train/grad_norm_pre_clip': 7.67708683013916, 'timing/data': 0.013397710397839546, 'timing/model': 1.451954398304224, 'learning_rate/qwen_vl_interface': 4.234634513787455e-06, 'learning_rate/action_model': 4.2346345137874556e-05, 'epoch': 1.91}) +[2026-06-16 03:27:43,590][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.11895788460969925, 'train/grad_norm_pre_clip': 15.212206840515137, 'timing/data': 0.014338606037199497, 'timing/model': 1.428041733801365, 'learning_rate/qwen_vl_interface': 4.2316596932397195e-06, 'learning_rate/action_model': 4.2316596932397195e-05, 'epoch': 1.91}) +[2026-06-16 03:27:49,643][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.030856627970933914, 'train/grad_norm_pre_clip': 3.155418872833252, 'timing/data': 0.014846858568489552, 'timing/model': 1.8273208728060126, 'learning_rate/qwen_vl_interface': 4.22868529129322e-06, 'learning_rate/action_model': 4.22868529129322e-05, 'epoch': 1.92}) +[2026-06-16 03:27:55,692][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.0640961155295372, 'train/grad_norm_pre_clip': 6.737588405609131, 'timing/data': 0.013552119955420494, 'timing/model': 1.8798924498260021, 'learning_rate/qwen_vl_interface': 4.2257113091706165e-06, 'learning_rate/action_model': 4.2257113091706165e-05, 'epoch': 1.92}) +[2026-06-16 03:28:01,376][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.09465111792087555, 'train/grad_norm_pre_clip': 7.445284366607666, 'timing/data': 0.01390004251152277, 'timing/model': 1.45773381087929, 'learning_rate/qwen_vl_interface': 4.222737748094404e-06, 'learning_rate/action_model': 4.222737748094404e-05, 'epoch': 1.92}) +[2026-06-16 03:28:07,378][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.32502466440200806, 'train/grad_norm_pre_clip': 14.08010196685791, 'timing/data': 0.014912288635969162, 'timing/model': 1.4605448860675097, 'learning_rate/qwen_vl_interface': 4.219764609286901e-06, 'learning_rate/action_model': 4.219764609286901e-05, 'epoch': 1.92}) +[2026-06-16 03:28:13,379][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.0909593477845192, 'train/grad_norm_pre_clip': 5.693073749542236, 'timing/data': 0.013623081147670746, 'timing/model': 1.4577636821195483, 'learning_rate/qwen_vl_interface': 4.21679189397025e-06, 'learning_rate/action_model': 4.21679189397025e-05, 'epoch': 1.92}) +[2026-06-16 03:28:19,404][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.023930292576551437, 'train/grad_norm_pre_clip': 2.415775775909424, 'timing/data': 0.01448893640190363, 'timing/model': 1.434396680444479, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 1.92}) +[2026-06-16 03:28:25,377][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.015351385809481144, 'train/grad_norm_pre_clip': 1.2404659986495972, 'timing/data': 0.014836696907877922, 'timing/model': 1.4716430148109794, 'learning_rate/qwen_vl_interface': 4.210847738697219e-06, 'learning_rate/action_model': 4.210847738697218e-05, 'epoch': 1.92}) +[2026-06-16 03:28:31,452][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.08675386756658554, 'train/grad_norm_pre_clip': 5.575710773468018, 'timing/data': 0.014686927199363708, 'timing/model': 1.8460247460752726, 'learning_rate/qwen_vl_interface': 4.207876301184256e-06, 'learning_rate/action_model': 4.207876301184256e-05, 'epoch': 1.92}) +[2026-06-16 03:28:37,075][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.03667008504271507, 'train/grad_norm_pre_clip': 4.4370269775390625, 'timing/data': 0.014375685714185238, 'timing/model': 1.463148701004684, 'learning_rate/qwen_vl_interface': 4.2049052920489825e-06, 'learning_rate/action_model': 4.2049052920489825e-05, 'epoch': 1.92}) +[2026-06-16 03:28:43,181][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.10918213427066803, 'train/grad_norm_pre_clip': 8.758970260620117, 'timing/data': 0.014372264966368675, 'timing/model': 1.4622906679287553, 'learning_rate/qwen_vl_interface': 4.201934712512666e-06, 'learning_rate/action_model': 4.2019347125126656e-05, 'epoch': 1.92}) +[2026-06-16 03:28:49,275][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 0.12301252782344818, 'train/grad_norm_pre_clip': 9.16988754272461, 'timing/data': 0.014009852893650532, 'timing/model': 1.4460690524429083, 'learning_rate/qwen_vl_interface': 4.1989645637964e-06, 'learning_rate/action_model': 4.1989645637964e-05, 'epoch': 1.92}) +[2026-06-16 03:28:55,274][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.01907533034682274, 'train/grad_norm_pre_clip': 3.764465570449829, 'timing/data': 0.01460332702845335, 'timing/model': 1.4239455880597234, 'learning_rate/qwen_vl_interface': 4.195994847121101e-06, 'learning_rate/action_model': 4.195994847121101e-05, 'epoch': 1.92}) +[2026-06-16 03:39:50,292][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.03433394432067871, 'train/grad_norm_pre_clip': 4.31129789352417, 'eval/loss': 0.06859733909368515, 'eval/flap_precision': 0.802013422818792, 'eval/flap_recall': 0.8690909090909091, 'eval/flap_f1': 0.8342059336823735, 'eval/flap_rate_model': 0.08277777777777778, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.8342059336823735, 'timing/data': 0.014586837030947208, 'timing/model': 1.4134031431749463, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 1.92}) +[2026-06-16 03:39:50,294][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4200_state +[2026-06-16 03:39:50,303][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 03:39:57,849][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4200_state/model.safetensors +[2026-06-16 03:40:09,230][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4200_state/optimizer.bin +[2026-06-16 03:40:09,232][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4200_state/sampler.bin +[2026-06-16 03:40:09,233][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4200_state/sampler_1.bin +[2026-06-16 03:40:09,235][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4200_state/random_states_0.pkl +[2026-06-16 03:40:09,408][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4200 +[2026-06-16 03:40:09,409][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 03:40:09,426][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 03:40:15,423][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.3818872272968292, 'train/grad_norm_pre_clip': 16.49708366394043, 'timing/data': 0.014733153395354748, 'timing/model': 1.444064773619175, 'learning_rate/qwen_vl_interface': 4.19005671477618e-06, 'learning_rate/action_model': 4.1900567147761794e-05, 'epoch': 1.92}) +[2026-06-16 03:40:21,175][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 0.2193019688129425, 'train/grad_norm_pre_clip': 10.129866600036621, 'timing/data': 0.012497889809310436, 'timing/model': 1.7666436340659857, 'learning_rate/qwen_vl_interface': 4.1870883015474975e-06, 'learning_rate/action_model': 4.187088301547497e-05, 'epoch': 1.92}) +[2026-06-16 03:40:26,946][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.1492505669593811, 'train/grad_norm_pre_clip': 10.70749568939209, 'timing/data': 0.01242664735764265, 'timing/model': 1.5735019268468022, 'learning_rate/qwen_vl_interface': 4.1841203252416655e-06, 'learning_rate/action_model': 4.1841203252416655e-05, 'epoch': 1.92}) +[2026-06-16 03:40:32,444][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.17599961161613464, 'train/grad_norm_pre_clip': 7.317913055419922, 'timing/data': 0.0124552883207798, 'timing/model': 1.4581917989999056, 'learning_rate/qwen_vl_interface': 4.181152787078705e-06, 'learning_rate/action_model': 4.1811527870787056e-05, 'epoch': 1.92}) +[2026-06-16 03:40:38,358][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.2694966793060303, 'train/grad_norm_pre_clip': 13.5795316696167, 'timing/data': 0.012946751900017262, 'timing/model': 1.4613957200199366, 'learning_rate/qwen_vl_interface': 4.178185688278462e-06, 'learning_rate/action_model': 4.1781856882784616e-05, 'epoch': 1.92}) +[2026-06-16 03:40:44,281][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 0.1902061104774475, 'train/grad_norm_pre_clip': 6.81257963180542, 'timing/data': 0.011956850998103619, 'timing/model': 1.3986845519393682, 'learning_rate/qwen_vl_interface': 4.175219030060595e-06, 'learning_rate/action_model': 4.175219030060595e-05, 'epoch': 1.92}) +[2026-06-16 03:40:50,277][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.026980269700288773, 'train/grad_norm_pre_clip': 9.389925003051758, 'timing/data': 0.014514747075736523, 'timing/model': 1.3995396923273802, 'learning_rate/qwen_vl_interface': 4.1722528136445875e-06, 'learning_rate/action_model': 4.172252813644588e-05, 'epoch': 1.92}) +[2026-06-16 03:40:56,365][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.12804806232452393, 'train/grad_norm_pre_clip': 8.239439964294434, 'timing/data': 0.015196693129837513, 'timing/model': 1.4659647019580007, 'learning_rate/qwen_vl_interface': 4.169287040249738e-06, 'learning_rate/action_model': 4.169287040249738e-05, 'epoch': 1.92}) +[2026-06-16 03:41:02,484][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.27518099546432495, 'train/grad_norm_pre_clip': 17.388017654418945, 'timing/data': 0.013460415415465832, 'timing/model': 1.9832550128921866, 'learning_rate/qwen_vl_interface': 4.166321711095164e-06, 'learning_rate/action_model': 4.166321711095164e-05, 'epoch': 1.92}) +[2026-06-16 03:41:08,170][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.06449900567531586, 'train/grad_norm_pre_clip': 5.027892589569092, 'timing/data': 0.015141651034355164, 'timing/model': 1.467848644591868, 'learning_rate/qwen_vl_interface': 4.1633568273998e-06, 'learning_rate/action_model': 4.1633568273998e-05, 'epoch': 1.93}) +[2026-06-16 03:41:14,273][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.07774831354618073, 'train/grad_norm_pre_clip': 7.2287821769714355, 'timing/data': 0.013577794656157494, 'timing/model': 1.4625233318656683, 'learning_rate/qwen_vl_interface': 4.1603923903823995e-06, 'learning_rate/action_model': 4.160392390382399e-05, 'epoch': 1.93}) +[2026-06-16 03:41:20,367][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 0.18332555890083313, 'train/grad_norm_pre_clip': 12.07626724243164, 'timing/data': 0.015095929615199566, 'timing/model': 1.4516593655571342, 'learning_rate/qwen_vl_interface': 4.1574284012615276e-06, 'learning_rate/action_model': 4.157428401261528e-05, 'epoch': 1.93}) +[2026-06-16 03:41:26,371][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.24783256649971008, 'train/grad_norm_pre_clip': 10.019124031066895, 'timing/data': 0.016539261676371098, 'timing/model': 1.4169836286455393, 'learning_rate/qwen_vl_interface': 4.154464861255571e-06, 'learning_rate/action_model': 4.154464861255571e-05, 'epoch': 1.93}) +[2026-06-16 03:41:33,218][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.1105034202337265, 'train/grad_norm_pre_clip': 7.23154878616333, 'timing/data': 0.03888841066509485, 'timing/model': 2.0320994816720486, 'learning_rate/qwen_vl_interface': 4.151501771582728e-06, 'learning_rate/action_model': 4.151501771582728e-05, 'epoch': 1.93}) +[2026-06-16 03:41:39,304][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.17402926087379456, 'train/grad_norm_pre_clip': 8.08263874053955, 'timing/data': 0.0136652123183012, 'timing/model': 1.9002416161820292, 'learning_rate/qwen_vl_interface': 4.148539133461012e-06, 'learning_rate/action_model': 4.1485391334610116e-05, 'epoch': 1.93}) +[2026-06-16 03:41:44,962][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.258568674325943, 'train/grad_norm_pre_clip': 11.715510368347168, 'timing/data': 0.014030136168003082, 'timing/model': 1.4423235999420285, 'learning_rate/qwen_vl_interface': 4.145576948108254e-06, 'learning_rate/action_model': 4.145576948108253e-05, 'epoch': 1.93}) +[2026-06-16 03:41:50,872][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.0948585644364357, 'train/grad_norm_pre_clip': 5.206165790557861, 'timing/data': 0.014896123670041561, 'timing/model': 1.4674366507679224, 'learning_rate/qwen_vl_interface': 4.142615216742093e-06, 'learning_rate/action_model': 4.142615216742093e-05, 'epoch': 1.93}) +[2026-06-16 03:41:56,867][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.14067760109901428, 'train/grad_norm_pre_clip': 8.662952423095703, 'timing/data': 0.014212285168468952, 'timing/model': 1.463804952800274, 'learning_rate/qwen_vl_interface': 4.139653940579988e-06, 'learning_rate/action_model': 4.1396539405799874e-05, 'epoch': 1.93}) +[2026-06-16 03:42:02,770][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.1792392134666443, 'train/grad_norm_pre_clip': 14.037590026855469, 'timing/data': 0.014655009843409061, 'timing/model': 1.4523473121225834, 'learning_rate/qwen_vl_interface': 4.136693120839206e-06, 'learning_rate/action_model': 4.136693120839205e-05, 'epoch': 1.93}) +[2026-06-16 03:42:08,764][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.26203808188438416, 'train/grad_norm_pre_clip': 8.396604537963867, 'timing/data': 0.01872529834508896, 'timing/model': 1.7512334072962403, 'learning_rate/qwen_vl_interface': 4.133732758736828e-06, 'learning_rate/action_model': 4.133732758736828e-05, 'epoch': 1.93}) +[2026-06-16 03:42:14,709][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.11193519085645676, 'train/grad_norm_pre_clip': 12.553228378295898, 'timing/data': 0.014110513962805271, 'timing/model': 1.7037055334076285, 'learning_rate/qwen_vl_interface': 4.13077285548975e-06, 'learning_rate/action_model': 4.13077285548975e-05, 'epoch': 1.93}) +[2026-06-16 03:42:20,359][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.09414263814687729, 'train/grad_norm_pre_clip': 6.659552574157715, 'timing/data': 0.013796900399029255, 'timing/model': 1.4536960264667869, 'learning_rate/qwen_vl_interface': 4.127813412314672e-06, 'learning_rate/action_model': 4.1278134123146725e-05, 'epoch': 1.93}) +[2026-06-16 03:42:26,259][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.03708190843462944, 'train/grad_norm_pre_clip': 3.1368167400360107, 'timing/data': 0.014604996889829636, 'timing/model': 1.452651440165937, 'learning_rate/qwen_vl_interface': 4.124854430428113e-06, 'learning_rate/action_model': 4.1248544304281126e-05, 'epoch': 1.93}) +[2026-06-16 03:42:32,181][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.017599480226635933, 'train/grad_norm_pre_clip': 1.046347975730896, 'timing/data': 0.013868669979274273, 'timing/model': 1.4882009644061327, 'learning_rate/qwen_vl_interface': 4.121895911046397e-06, 'learning_rate/action_model': 4.121895911046396e-05, 'epoch': 1.93}) +[2026-06-16 03:42:38,157][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.06585060060024261, 'train/grad_norm_pre_clip': 6.49321174621582, 'timing/data': 0.012806450016796589, 'timing/model': 1.4692714093253016, 'learning_rate/qwen_vl_interface': 4.118937855385659e-06, 'learning_rate/action_model': 4.1189378553856584e-05, 'epoch': 1.93}) +[2026-06-16 03:42:44,155][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.16075947880744934, 'train/grad_norm_pre_clip': 8.872633934020996, 'timing/data': 0.05261913873255253, 'timing/model': 1.7767706690356135, 'learning_rate/qwen_vl_interface': 4.115980264661844e-06, 'learning_rate/action_model': 4.115980264661844e-05, 'epoch': 1.93}) +[2026-06-16 03:42:50,075][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.03475607931613922, 'train/grad_norm_pre_clip': 4.211853504180908, 'timing/data': 0.012234833091497421, 'timing/model': 1.648468654602766, 'learning_rate/qwen_vl_interface': 4.113023140090709e-06, 'learning_rate/action_model': 4.113023140090709e-05, 'epoch': 1.93}) +[2026-06-16 03:42:55,750][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 0.15158388018608093, 'train/grad_norm_pre_clip': 5.428691387176514, 'timing/data': 0.011094312183558941, 'timing/model': 1.4489689860492945, 'learning_rate/qwen_vl_interface': 4.110066482887812e-06, 'learning_rate/action_model': 4.110066482887812e-05, 'epoch': 1.93}) +[2026-06-16 03:43:01,520][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.14671140909194946, 'train/grad_norm_pre_clip': 6.105653285980225, 'timing/data': 0.011279144324362278, 'timing/model': 1.4386087376624346, 'learning_rate/qwen_vl_interface': 4.107110294268525e-06, 'learning_rate/action_model': 4.107110294268525e-05, 'epoch': 1.93}) +[2026-06-16 03:43:07,362][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.04474397748708725, 'train/grad_norm_pre_clip': 5.556118011474609, 'timing/data': 0.011163142509758472, 'timing/model': 1.4760879380628467, 'learning_rate/qwen_vl_interface': 4.104154575448026e-06, 'learning_rate/action_model': 4.104154575448025e-05, 'epoch': 1.93}) +[2026-06-16 03:43:13,261][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 0.15625838935375214, 'train/grad_norm_pre_clip': 8.123549461364746, 'timing/data': 0.01120828278362751, 'timing/model': 1.4742971193045378, 'learning_rate/qwen_vl_interface': 4.101199327641297e-06, 'learning_rate/action_model': 4.101199327641297e-05, 'epoch': 1.93}) +[2026-06-16 03:43:19,124][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.10860928148031235, 'train/grad_norm_pre_clip': 11.69542121887207, 'timing/data': 0.011475174687802792, 'timing/model': 1.7241057511419058, 'learning_rate/qwen_vl_interface': 4.098244552063133e-06, 'learning_rate/action_model': 4.098244552063132e-05, 'epoch': 1.94}) +[2026-06-16 03:43:24,745][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.32109034061431885, 'train/grad_norm_pre_clip': 11.54968547821045, 'timing/data': 0.010465074330568314, 'timing/model': 1.4456970281898975, 'learning_rate/qwen_vl_interface': 4.0952902499281254e-06, 'learning_rate/action_model': 4.095290249928125e-05, 'epoch': 1.94}) +[2026-06-16 03:43:30,558][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 0.07869184017181396, 'train/grad_norm_pre_clip': 7.757724285125732, 'timing/data': 0.010900517925620079, 'timing/model': 1.4718765430152416, 'learning_rate/qwen_vl_interface': 4.092336422450681e-06, 'learning_rate/action_model': 4.0923364224506804e-05, 'epoch': 1.94}) +[2026-06-16 03:43:36,366][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.13394489884376526, 'train/grad_norm_pre_clip': 9.151511192321777, 'timing/data': 0.011340991593897343, 'timing/model': 1.4808712238445878, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 1.94}) +[2026-06-16 03:43:42,144][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.08402223885059357, 'train/grad_norm_pre_clip': 3.538625955581665, 'timing/data': 0.010612213984131813, 'timing/model': 1.4571690922603011, 'learning_rate/qwen_vl_interface': 4.086430196325106e-06, 'learning_rate/action_model': 4.086430196325106e-05, 'epoch': 1.94}) +[2026-06-16 03:43:47,970][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.2396555244922638, 'train/grad_norm_pre_clip': 12.247164726257324, 'timing/data': 0.011394931003451347, 'timing/model': 1.3974344730377197, 'learning_rate/qwen_vl_interface': 4.083477800104802e-06, 'learning_rate/action_model': 4.083477800104802e-05, 'epoch': 1.94}) +[2026-06-16 03:43:53,827][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.19496481120586395, 'train/grad_norm_pre_clip': 9.970624923706055, 'timing/data': 0.010807954706251621, 'timing/model': 1.7252535615116358, 'learning_rate/qwen_vl_interface': 4.080525883397711e-06, 'learning_rate/action_model': 4.080525883397711e-05, 'epoch': 1.94}) +[2026-06-16 03:43:59,448][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.13083870708942413, 'train/grad_norm_pre_clip': 15.385773658752441, 'timing/data': 0.01020870916545391, 'timing/model': 1.4458565190434456, 'learning_rate/qwen_vl_interface': 4.077574447417256e-06, 'learning_rate/action_model': 4.077574447417256e-05, 'epoch': 1.94}) +[2026-06-16 03:44:05,248][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.07576064765453339, 'train/grad_norm_pre_clip': 5.004092693328857, 'timing/data': 0.010399531573057175, 'timing/model': 1.4625746794044971, 'learning_rate/qwen_vl_interface': 4.074623493376657e-06, 'learning_rate/action_model': 4.074623493376657e-05, 'epoch': 1.94}) +[2026-06-16 03:44:11,059][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.13491110503673553, 'train/grad_norm_pre_clip': 9.51602554321289, 'timing/data': 0.01015446987003088, 'timing/model': 1.4604959851130843, 'learning_rate/qwen_vl_interface': 4.071673022488944e-06, 'learning_rate/action_model': 4.071673022488944e-05, 'epoch': 1.94}) +[2026-06-16 03:44:16,852][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.09689300507307053, 'train/grad_norm_pre_clip': 3.9007856845855713, 'timing/data': 0.011279049329459667, 'timing/model': 1.4095822433009744, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 1.94}) +[2026-06-16 03:44:22,684][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.2015189677476883, 'train/grad_norm_pre_clip': 8.785374641418457, 'timing/data': 0.01031854934990406, 'timing/model': 1.4173170756548643, 'learning_rate/qwen_vl_interface': 4.065773535023277e-06, 'learning_rate/action_model': 4.0657735350232766e-05, 'epoch': 1.94}) +[2026-06-16 03:44:28,671][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.08362087607383728, 'train/grad_norm_pre_clip': 12.701909065246582, 'timing/data': 0.01103635597974062, 'timing/model': 1.7572289686650038, 'learning_rate/qwen_vl_interface': 4.062824520870382e-06, 'learning_rate/action_model': 4.062824520870381e-05, 'epoch': 1.94}) +[2026-06-16 03:44:34,630][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.06215004622936249, 'train/grad_norm_pre_clip': 4.434218406677246, 'timing/data': 0.01151321828365326, 'timing/model': 1.7189744347706437, 'learning_rate/qwen_vl_interface': 4.059875994720481e-06, 'learning_rate/action_model': 4.059875994720481e-05, 'epoch': 1.94}) +[2026-06-16 03:44:40,274][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.22362124919891357, 'train/grad_norm_pre_clip': 7.133878231048584, 'timing/data': 0.010183757171034813, 'timing/model': 1.4586386922746897, 'learning_rate/qwen_vl_interface': 4.056927957785605e-06, 'learning_rate/action_model': 4.056927957785605e-05, 'epoch': 1.94}) +[2026-06-16 03:44:46,171][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.11616255342960358, 'train/grad_norm_pre_clip': 12.392887115478516, 'timing/data': 0.011325427331030369, 'timing/model': 1.476749623194337, 'learning_rate/qwen_vl_interface': 4.0539804112775785e-06, 'learning_rate/action_model': 4.053980411277578e-05, 'epoch': 1.94}) +[2026-06-16 03:44:51,963][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.044898536056280136, 'train/grad_norm_pre_clip': 2.226257085800171, 'timing/data': 0.011204235255718231, 'timing/model': 1.4548656027764082, 'learning_rate/qwen_vl_interface': 4.051033356408027e-06, 'learning_rate/action_model': 4.0510333564080264e-05, 'epoch': 1.94}) +[2026-06-16 03:44:57,786][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.14102308452129364, 'train/grad_norm_pre_clip': 11.165722846984863, 'timing/data': 0.011117154732346535, 'timing/model': 1.4076900780200958, 'learning_rate/qwen_vl_interface': 4.048086794388375e-06, 'learning_rate/action_model': 4.048086794388375e-05, 'epoch': 1.94}) +[2026-06-16 03:45:03,746][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.10863806307315826, 'train/grad_norm_pre_clip': 6.473755359649658, 'timing/data': 0.011195395141839981, 'timing/model': 1.7442256230860949, 'learning_rate/qwen_vl_interface': 4.045140726429841e-06, 'learning_rate/action_model': 4.0451407264298406e-05, 'epoch': 1.94}) +[2026-06-16 03:45:09,688][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.1142820417881012, 'train/grad_norm_pre_clip': 7.056947231292725, 'timing/data': 0.011363496072590351, 'timing/model': 1.7803869685158134, 'learning_rate/qwen_vl_interface': 4.042195153743443e-06, 'learning_rate/action_model': 4.042195153743443e-05, 'epoch': 1.94}) +[2026-06-16 03:45:15,371][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.16440428793430328, 'train/grad_norm_pre_clip': 10.060138702392578, 'timing/data': 0.011572678573429585, 'timing/model': 1.4660839987918735, 'learning_rate/qwen_vl_interface': 4.039250077539993e-06, 'learning_rate/action_model': 4.039250077539993e-05, 'epoch': 1.94}) +[2026-06-16 03:45:21,263][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.31585726141929626, 'train/grad_norm_pre_clip': 14.926937103271484, 'timing/data': 0.01174473762512207, 'timing/model': 1.4532683389261365, 'learning_rate/qwen_vl_interface': 4.036305499030104e-06, 'learning_rate/action_model': 4.0363054990301036e-05, 'epoch': 1.95}) +[2026-06-16 03:45:27,182][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.10418762266635895, 'train/grad_norm_pre_clip': 7.343230247497559, 'timing/data': 0.01166782807558775, 'timing/model': 1.4747796598821878, 'learning_rate/qwen_vl_interface': 4.033361419424178e-06, 'learning_rate/action_model': 4.033361419424178e-05, 'epoch': 1.95}) +[2026-06-16 03:45:33,167][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.2021510750055313, 'train/grad_norm_pre_clip': 7.656381607055664, 'timing/data': 0.011440875940024853, 'timing/model': 1.4507555468007922, 'learning_rate/qwen_vl_interface': 4.030417839932418e-06, 'learning_rate/action_model': 4.030417839932417e-05, 'epoch': 1.95}) +[2026-06-16 03:45:39,044][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.2877703011035919, 'train/grad_norm_pre_clip': 9.03881549835205, 'timing/data': 0.011571596376597881, 'timing/model': 1.7564752418547869, 'learning_rate/qwen_vl_interface': 4.027474761764815e-06, 'learning_rate/action_model': 4.027474761764814e-05, 'epoch': 1.95}) +[2026-06-16 03:45:44,663][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.16692043840885162, 'train/grad_norm_pre_clip': 11.600528717041016, 'timing/data': 0.013485550880432129, 'timing/model': 1.4592586113139987, 'learning_rate/qwen_vl_interface': 4.024532186131158e-06, 'learning_rate/action_model': 4.0245321861311575e-05, 'epoch': 1.95}) +[2026-06-16 03:45:50,565][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 0.10203820466995239, 'train/grad_norm_pre_clip': 9.614572525024414, 'timing/data': 0.011366493999958038, 'timing/model': 1.456308956257999, 'learning_rate/qwen_vl_interface': 4.021590114241029e-06, 'learning_rate/action_model': 4.021590114241029e-05, 'epoch': 1.95}) +[2026-06-16 03:45:56,355][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.18357262015342712, 'train/grad_norm_pre_clip': 8.352818489074707, 'timing/data': 0.011013660579919815, 'timing/model': 1.4475320158526301, 'learning_rate/qwen_vl_interface': 4.018648547303804e-06, 'learning_rate/action_model': 4.0186485473038036e-05, 'epoch': 1.95}) +[2026-06-16 03:46:02,267][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.21795623004436493, 'train/grad_norm_pre_clip': 7.919404983520508, 'timing/data': 0.010457496158778667, 'timing/model': 1.4605898959562182, 'learning_rate/qwen_vl_interface': 4.015707486528649e-06, 'learning_rate/action_model': 4.015707486528649e-05, 'epoch': 1.95}) +[2026-06-16 03:46:08,073][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.06817872077226639, 'train/grad_norm_pre_clip': 4.456823348999023, 'timing/data': 0.015518196858465672, 'timing/model': 1.420846814289689, 'learning_rate/qwen_vl_interface': 4.012766933124523e-06, 'learning_rate/action_model': 4.012766933124523e-05, 'epoch': 1.95}) +[2026-06-16 03:46:13,992][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.0343766063451767, 'train/grad_norm_pre_clip': 2.491885185241699, 'timing/data': 0.010319114662706852, 'timing/model': 1.7063124189153314, 'learning_rate/qwen_vl_interface': 4.009826888300176e-06, 'learning_rate/action_model': 4.0098268883001755e-05, 'epoch': 1.95}) +[2026-06-16 03:46:19,671][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.09213025867938995, 'train/grad_norm_pre_clip': 4.570730686187744, 'timing/data': 0.011786725372076035, 'timing/model': 1.4603505143895745, 'learning_rate/qwen_vl_interface': 4.00688735326415e-06, 'learning_rate/action_model': 4.0068873532641495e-05, 'epoch': 1.95}) +[2026-06-16 03:46:25,588][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.07198774069547653, 'train/grad_norm_pre_clip': 4.699484825134277, 'timing/data': 0.012037958018481731, 'timing/model': 1.4812606777995825, 'learning_rate/qwen_vl_interface': 4.0039483292247765e-06, 'learning_rate/action_model': 4.003948329224776e-05, 'epoch': 1.95}) +[2026-06-16 03:46:31,578][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.09248873591423035, 'train/grad_norm_pre_clip': 6.636959075927734, 'timing/data': 0.012222038581967354, 'timing/model': 1.4543393636122346, 'learning_rate/qwen_vl_interface': 4.001009817390178e-06, 'learning_rate/action_model': 4.001009817390178e-05, 'epoch': 1.95}) +[2026-06-16 03:46:37,582][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.09677757322788239, 'train/grad_norm_pre_clip': 7.880346298217773, 'timing/data': 0.011375261470675468, 'timing/model': 1.4841008065268397, 'learning_rate/qwen_vl_interface': 3.998071818968264e-06, 'learning_rate/action_model': 3.998071818968264e-05, 'epoch': 1.95}) +[2026-06-16 03:46:43,771][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.10216972231864929, 'train/grad_norm_pre_clip': 5.440389633178711, 'timing/data': 0.01170741394162178, 'timing/model': 1.4515356477349997, 'learning_rate/qwen_vl_interface': 3.995134335166736e-06, 'learning_rate/action_model': 3.995134335166736e-05, 'epoch': 1.95}) +[2026-06-16 03:46:49,840][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.17358627915382385, 'train/grad_norm_pre_clip': 9.051250457763672, 'timing/data': 0.015376743860542774, 'timing/model': 1.8227860080078244, 'learning_rate/qwen_vl_interface': 3.992197367193083e-06, 'learning_rate/action_model': 3.992197367193082e-05, 'epoch': 1.95}) +[2026-06-16 03:46:55,476][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.03812635317444801, 'train/grad_norm_pre_clip': 2.9863812923431396, 'timing/data': 0.015822825953364372, 'timing/model': 1.4617749117314816, 'learning_rate/qwen_vl_interface': 3.98926091625458e-06, 'learning_rate/action_model': 3.9892609162545795e-05, 'epoch': 1.95}) +[2026-06-16 03:47:01,483][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.08451656252145767, 'train/grad_norm_pre_clip': 5.776883602142334, 'timing/data': 0.01137218065559864, 'timing/model': 1.4726348426192999, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 1.95}) +[2026-06-16 03:47:07,598][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.2754654288291931, 'train/grad_norm_pre_clip': 12.871866226196289, 'timing/data': 0.015377352014183998, 'timing/model': 1.4800044363364577, 'learning_rate/qwen_vl_interface': 3.983389570311073e-06, 'learning_rate/action_model': 3.9833895703110724e-05, 'epoch': 1.95}) +[2026-06-16 03:47:13,677][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.1019960567355156, 'train/grad_norm_pre_clip': 7.330022811889648, 'timing/data': 0.01560758426785469, 'timing/model': 1.4930022554472089, 'learning_rate/qwen_vl_interface': 3.9804546777195546e-06, 'learning_rate/action_model': 3.980454677719554e-05, 'epoch': 1.95}) +[2026-06-16 03:47:19,757][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.032174572348594666, 'train/grad_norm_pre_clip': 2.855165719985962, 'timing/data': 0.015573362819850445, 'timing/model': 1.8438902525231242, 'learning_rate/qwen_vl_interface': 3.977520306990164e-06, 'learning_rate/action_model': 3.977520306990163e-05, 'epoch': 1.95}) +[2026-06-16 03:47:25,368][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 0.09533341974020004, 'train/grad_norm_pre_clip': 11.247844696044922, 'timing/data': 0.014712406322360039, 'timing/model': 1.4542753333225846, 'learning_rate/qwen_vl_interface': 3.974586459329109e-06, 'learning_rate/action_model': 3.9745864593291085e-05, 'epoch': 1.95}) +[2026-06-16 03:47:31,491][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.029345432296395302, 'train/grad_norm_pre_clip': 3.579772472381592, 'timing/data': 0.015043768100440502, 'timing/model': 1.4681274788454175, 'learning_rate/qwen_vl_interface': 3.9716531359423855e-06, 'learning_rate/action_model': 3.971653135942385e-05, 'epoch': 1.96}) +[2026-06-16 03:47:37,365][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.21244767308235168, 'train/grad_norm_pre_clip': 16.30366325378418, 'timing/data': 0.011130898259580135, 'timing/model': 1.472190578468144, 'learning_rate/qwen_vl_interface': 3.968720338035769e-06, 'learning_rate/action_model': 3.968720338035769e-05, 'epoch': 1.96}) +[2026-06-16 03:47:43,171][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.021680034697055817, 'train/grad_norm_pre_clip': 1.6001338958740234, 'timing/data': 0.012047613970935345, 'timing/model': 1.4576674746349454, 'learning_rate/qwen_vl_interface': 3.965788066814824e-06, 'learning_rate/action_model': 3.965788066814823e-05, 'epoch': 1.96}) +[2026-06-16 03:47:48,968][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.20579320192337036, 'train/grad_norm_pre_clip': 9.924468994140625, 'timing/data': 0.011092906817793846, 'timing/model': 1.410252665169537, 'learning_rate/qwen_vl_interface': 3.962856323484895e-06, 'learning_rate/action_model': 3.9628563234848945e-05, 'epoch': 1.96}) +[2026-06-16 03:47:54,963][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.08801016211509705, 'train/grad_norm_pre_clip': 7.552284240722656, 'timing/data': 0.01164960116147995, 'timing/model': 1.7670969478785992, 'learning_rate/qwen_vl_interface': 3.959925109251112e-06, 'learning_rate/action_model': 3.959925109251112e-05, 'epoch': 1.96}) +[2026-06-16 03:48:00,564][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.03998874872922897, 'train/grad_norm_pre_clip': 2.564446449279785, 'timing/data': 0.011719461530447006, 'timing/model': 1.4430572111159563, 'learning_rate/qwen_vl_interface': 3.956994425318387e-06, 'learning_rate/action_model': 3.956994425318387e-05, 'epoch': 1.96}) +[2026-06-16 03:48:06,475][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 0.204585462808609, 'train/grad_norm_pre_clip': 11.241022109985352, 'timing/data': 0.012041463516652584, 'timing/model': 1.461757336743176, 'learning_rate/qwen_vl_interface': 3.9540642728914115e-06, 'learning_rate/action_model': 3.9540642728914115e-05, 'epoch': 1.96}) +[2026-06-16 03:48:12,473][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.048600975424051285, 'train/grad_norm_pre_clip': 3.034334659576416, 'timing/data': 0.011024684645235538, 'timing/model': 1.4372576465830207, 'learning_rate/qwen_vl_interface': 3.951134653174664e-06, 'learning_rate/action_model': 3.951134653174663e-05, 'epoch': 1.96}) +[2026-06-16 03:48:18,575][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.26747965812683105, 'train/grad_norm_pre_clip': 12.534208297729492, 'timing/data': 0.012872829101979733, 'timing/model': 1.4583115605637431, 'learning_rate/qwen_vl_interface': 3.948205567372397e-06, 'learning_rate/action_model': 3.948205567372397e-05, 'epoch': 1.96}) +[2026-06-16 03:48:24,612][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.3984340727329254, 'train/grad_norm_pre_clip': 10.91325855255127, 'timing/data': 0.015481988899409771, 'timing/model': 1.4362160302698612, 'learning_rate/qwen_vl_interface': 3.945277016688647e-06, 'learning_rate/action_model': 3.945277016688647e-05, 'epoch': 1.96}) +[2026-06-16 03:48:30,684][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.03774870187044144, 'train/grad_norm_pre_clip': 5.5958251953125, 'timing/data': 0.016046762466430664, 'timing/model': 1.853453192859888, 'learning_rate/qwen_vl_interface': 3.9423490023272326e-06, 'learning_rate/action_model': 3.9423490023272324e-05, 'epoch': 1.96}) +[2026-06-16 03:48:36,392][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.29273274540901184, 'train/grad_norm_pre_clip': 15.679388999938965, 'timing/data': 0.015089265070855618, 'timing/model': 1.467963706701994, 'learning_rate/qwen_vl_interface': 3.939421525491748e-06, 'learning_rate/action_model': 3.9394215254917483e-05, 'epoch': 1.96}) +[2026-06-16 03:48:42,482][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.08059590309858322, 'train/grad_norm_pre_clip': 11.21851921081543, 'timing/data': 0.014794383198022842, 'timing/model': 1.456656001508236, 'learning_rate/qwen_vl_interface': 3.9364945873855705e-06, 'learning_rate/action_model': 3.93649458738557e-05, 'epoch': 1.96}) +[2026-06-16 03:48:48,696][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.03150974586606026, 'train/grad_norm_pre_clip': 2.6725893020629883, 'timing/data': 0.014933723025023937, 'timing/model': 1.4715151963755488, 'learning_rate/qwen_vl_interface': 3.9335681892118536e-06, 'learning_rate/action_model': 3.933568189211854e-05, 'epoch': 1.96}) +[2026-06-16 03:48:54,977][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.09087482839822769, 'train/grad_norm_pre_clip': 5.997558116912842, 'timing/data': 0.012830886989831924, 'timing/model': 1.482746602036059, 'learning_rate/qwen_vl_interface': 3.930642332173527e-06, 'learning_rate/action_model': 3.930642332173527e-05, 'epoch': 1.96}) +[2026-06-16 03:49:01,026][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.16620315611362457, 'train/grad_norm_pre_clip': 14.985815048217773, 'timing/data': 0.012044071219861507, 'timing/model': 1.8219889169558883, 'learning_rate/qwen_vl_interface': 3.927717017473301e-06, 'learning_rate/action_model': 3.9277170174733006e-05, 'epoch': 1.96}) +[2026-06-16 03:49:06,920][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.10923653095960617, 'train/grad_norm_pre_clip': 10.127730369567871, 'timing/data': 0.012339812703430653, 'timing/model': 1.7083567176014185, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 1.96}) +[2026-06-16 03:49:12,595][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.12274209409952164, 'train/grad_norm_pre_clip': 9.601734161376953, 'timing/data': 0.01296438742429018, 'timing/model': 1.4811087995767593, 'learning_rate/qwen_vl_interface': 3.921868019896873e-06, 'learning_rate/action_model': 3.921868019896873e-05, 'epoch': 1.96}) +[2026-06-16 03:49:18,579][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.11577383428812027, 'train/grad_norm_pre_clip': 8.723258972167969, 'timing/data': 0.0147465905174613, 'timing/model': 1.4640289535745978, 'learning_rate/qwen_vl_interface': 3.918944339424974e-06, 'learning_rate/action_model': 3.9189443394249736e-05, 'epoch': 1.96}) +[2026-06-16 03:49:24,579][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.13495007157325745, 'train/grad_norm_pre_clip': 11.064987182617188, 'timing/data': 0.011982230469584465, 'timing/model': 1.4589424384757876, 'learning_rate/qwen_vl_interface': 3.9160212060997774e-06, 'learning_rate/action_model': 3.9160212060997774e-05, 'epoch': 1.96}) +[2026-06-16 03:49:30,575][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 0.2173701971769333, 'train/grad_norm_pre_clip': 8.76630973815918, 'timing/data': 0.012827795930206776, 'timing/model': 1.44669050257653, 'learning_rate/qwen_vl_interface': 3.913098621122875e-06, 'learning_rate/action_model': 3.913098621122875e-05, 'epoch': 1.96}) +[2026-06-16 03:49:36,622][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.12198077142238617, 'train/grad_norm_pre_clip': 11.263989448547363, 'timing/data': 0.0115172965452075, 'timing/model': 1.8104145545512438, 'learning_rate/qwen_vl_interface': 3.91017658569563e-06, 'learning_rate/action_model': 3.91017658569563e-05, 'epoch': 1.96}) +[2026-06-16 03:49:42,528][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.19825494289398193, 'train/grad_norm_pre_clip': 10.221840858459473, 'timing/data': 0.012094330042600632, 'timing/model': 1.710115509107709, 'learning_rate/qwen_vl_interface': 3.907255101019181e-06, 'learning_rate/action_model': 3.907255101019181e-05, 'epoch': 1.97}) +[2026-06-16 03:49:48,187][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.21318596601486206, 'train/grad_norm_pre_clip': 9.468350410461426, 'timing/data': 0.012290291488170624, 'timing/model': 1.4734969278797507, 'learning_rate/qwen_vl_interface': 3.904334168294442e-06, 'learning_rate/action_model': 3.9043341682944415e-05, 'epoch': 1.97}) +[2026-06-16 03:49:54,191][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.12897822260856628, 'train/grad_norm_pre_clip': 6.937408924102783, 'timing/data': 0.011587824672460556, 'timing/model': 1.482565701007843, 'learning_rate/qwen_vl_interface': 3.901413788722093e-06, 'learning_rate/action_model': 3.901413788722093e-05, 'epoch': 1.97}) +[2026-06-16 04:00:49,267][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.042928580194711685, 'train/grad_norm_pre_clip': 2.7800650596618652, 'eval/loss': 0.0607319138944149, 'eval/flap_precision': 0.8013468013468014, 'eval/flap_recall': 0.8654545454545455, 'eval/flap_f1': 0.8321678321678323, 'eval/flap_rate_model': 0.0825, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.8321678321678324, 'timing/data': 0.012277149595320225, 'timing/model': 1.4706155778840184, 'learning_rate/qwen_vl_interface': 3.8984939635025966e-06, 'learning_rate/action_model': 3.898493963502596e-05, 'epoch': 1.97}) +[2026-06-16 04:00:49,268][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4300_state +[2026-06-16 04:00:49,276][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 04:00:56,965][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4300_state/model.safetensors +[2026-06-16 04:01:08,435][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4300_state/optimizer.bin +[2026-06-16 04:01:08,436][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4300_state/sampler.bin +[2026-06-16 04:01:08,437][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4300_state/sampler_1.bin +[2026-06-16 04:01:08,439][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4300_state/random_states_0.pkl +[2026-06-16 04:01:08,607][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4300 +[2026-06-16 04:01:08,607][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 04:01:08,623][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 04:01:14,744][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.10150675475597382, 'train/grad_norm_pre_clip': 7.347538471221924, 'timing/data': 0.012574439868330956, 'timing/model': 1.4596896348521113, 'learning_rate/qwen_vl_interface': 3.895574693836181e-06, 'learning_rate/action_model': 3.8955746938361804e-05, 'epoch': 1.97}) +[2026-06-16 04:01:20,432][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.2281307429075241, 'train/grad_norm_pre_clip': 24.925167083740234, 'timing/data': 0.019373313523828983, 'timing/model': 1.4466490456834435, 'learning_rate/qwen_vl_interface': 3.892655980922847e-06, 'learning_rate/action_model': 3.8926559809228466e-05, 'epoch': 1.97}) +[2026-06-16 04:01:26,224][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.03305519372224808, 'train/grad_norm_pre_clip': 3.1147167682647705, 'timing/data': 0.016866927035152912, 'timing/model': 1.3724230099469423, 'learning_rate/qwen_vl_interface': 3.889737825962368e-06, 'learning_rate/action_model': 3.889737825962368e-05, 'epoch': 1.97}) +[2026-06-16 04:01:32,121][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 0.17022022604942322, 'train/grad_norm_pre_clip': 7.9788126945495605, 'timing/data': 0.011490522883832455, 'timing/model': 1.7438952550292015, 'learning_rate/qwen_vl_interface': 3.886820230154288e-06, 'learning_rate/action_model': 3.886820230154287e-05, 'epoch': 1.97}) +[2026-06-16 04:01:37,620][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.0712546855211258, 'train/grad_norm_pre_clip': 3.4192636013031006, 'timing/data': 0.011810713447630405, 'timing/model': 1.4402287183329463, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 1.97}) +[2026-06-16 04:01:43,356][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.008548378013074398, 'train/grad_norm_pre_clip': 0.9294067621231079, 'timing/data': 0.011384062469005585, 'timing/model': 1.3816742524504662, 'learning_rate/qwen_vl_interface': 3.880986720792344e-06, 'learning_rate/action_model': 3.880986720792344e-05, 'epoch': 1.97}) +[2026-06-16 04:01:48,966][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.0795307382941246, 'train/grad_norm_pre_clip': 3.1135711669921875, 'timing/data': 0.01251933816820383, 'timing/model': 1.3933204896748066, 'learning_rate/qwen_vl_interface': 3.878070809636416e-06, 'learning_rate/action_model': 3.878070809636416e-05, 'epoch': 1.97}) +[2026-06-16 04:01:54,856][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.07505805045366287, 'train/grad_norm_pre_clip': 7.2073469161987305, 'timing/data': 0.01228678785264492, 'timing/model': 1.4648967552930117, 'learning_rate/qwen_vl_interface': 3.875155462428758e-06, 'learning_rate/action_model': 3.875155462428758e-05, 'epoch': 1.97}) +[2026-06-16 04:02:00,559][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.12545107305049896, 'train/grad_norm_pre_clip': 8.686266899108887, 'timing/data': 0.014033478684723377, 'timing/model': 1.3904518010094762, 'learning_rate/qwen_vl_interface': 3.872240680367754e-06, 'learning_rate/action_model': 3.872240680367754e-05, 'epoch': 1.97}) +[2026-06-16 04:02:06,537][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.017441336065530777, 'train/grad_norm_pre_clip': 3.518913745880127, 'timing/data': 0.011850273236632347, 'timing/model': 1.7364914370700717, 'learning_rate/qwen_vl_interface': 3.869326464651566e-06, 'learning_rate/action_model': 3.869326464651566e-05, 'epoch': 1.97}) +[2026-06-16 04:02:12,059][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.09556636959314346, 'train/grad_norm_pre_clip': 6.97562837600708, 'timing/data': 0.011754184029996395, 'timing/model': 1.4584615379571915, 'learning_rate/qwen_vl_interface': 3.866412816478114e-06, 'learning_rate/action_model': 3.8664128164781144e-05, 'epoch': 1.97}) +[2026-06-16 04:02:17,949][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.0712115541100502, 'train/grad_norm_pre_clip': 5.254541397094727, 'timing/data': 0.01123696006834507, 'timing/model': 1.4570074044167995, 'learning_rate/qwen_vl_interface': 3.8634997370450915e-06, 'learning_rate/action_model': 3.8634997370450914e-05, 'epoch': 1.97}) +[2026-06-16 04:02:23,755][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.08488114923238754, 'train/grad_norm_pre_clip': 5.640434741973877, 'timing/data': 0.011399771086871624, 'timing/model': 1.4662722973152995, 'learning_rate/qwen_vl_interface': 3.860587227549956e-06, 'learning_rate/action_model': 3.860587227549955e-05, 'epoch': 1.97}) +[2026-06-16 04:02:29,499][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.038545362651348114, 'train/grad_norm_pre_clip': 3.007138252258301, 'timing/data': 0.010952988639473915, 'timing/model': 1.419494335539639, 'learning_rate/qwen_vl_interface': 3.857675289189926e-06, 'learning_rate/action_model': 3.857675289189926e-05, 'epoch': 1.97}) +[2026-06-16 04:02:35,248][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.32634374499320984, 'train/grad_norm_pre_clip': 12.373189926147461, 'timing/data': 0.011321340687572956, 'timing/model': 1.3912868546321988, 'learning_rate/qwen_vl_interface': 3.854763923161993e-06, 'learning_rate/action_model': 3.854763923161992e-05, 'epoch': 1.97}) +[2026-06-16 04:02:41,034][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.07411294430494308, 'train/grad_norm_pre_clip': 33.09372329711914, 'timing/data': 0.011370602063834667, 'timing/model': 1.652071027085185, 'learning_rate/qwen_vl_interface': 3.851853130662908e-06, 'learning_rate/action_model': 3.8518531306629077e-05, 'epoch': 1.97}) +[2026-06-16 04:02:46,779][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.06586072593927383, 'train/grad_norm_pre_clip': 5.820288181304932, 'timing/data': 0.01117647998034954, 'timing/model': 1.69802690949291, 'learning_rate/qwen_vl_interface': 3.8489429128891884e-06, 'learning_rate/action_model': 3.848942912889188e-05, 'epoch': 1.97}) +[2026-06-16 04:02:52,358][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.07791729271411896, 'train/grad_norm_pre_clip': 7.115940093994141, 'timing/data': 0.010714876465499401, 'timing/model': 1.4643764374777675, 'learning_rate/qwen_vl_interface': 3.846033271037114e-06, 'learning_rate/action_model': 3.8460332710371136e-05, 'epoch': 1.97}) +[2026-06-16 04:02:58,164][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.08165629953145981, 'train/grad_norm_pre_clip': 6.886498928070068, 'timing/data': 0.012500579468905926, 'timing/model': 1.472594814375043, 'learning_rate/qwen_vl_interface': 3.843124206302732e-06, 'learning_rate/action_model': 3.8431242063027316e-05, 'epoch': 1.98}) +[2026-06-16 04:03:03,876][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.006753607653081417, 'train/grad_norm_pre_clip': 0.9650793671607971, 'timing/data': 0.011115129105746746, 'timing/model': 1.39318030141294, 'learning_rate/qwen_vl_interface': 3.840215719881844e-06, 'learning_rate/action_model': 3.8402157198818445e-05, 'epoch': 1.98}) +[2026-06-16 04:03:09,667][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.06873250007629395, 'train/grad_norm_pre_clip': 5.342537879943848, 'timing/data': 0.011229260824620724, 'timing/model': 1.4061245676130056, 'learning_rate/qwen_vl_interface': 3.837307812970022e-06, 'learning_rate/action_model': 3.837307812970022e-05, 'epoch': 1.98}) +[2026-06-16 04:03:15,527][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.033115290105342865, 'train/grad_norm_pre_clip': 4.4340338706970215, 'timing/data': 0.011224491521716118, 'timing/model': 1.727997207082808, 'learning_rate/qwen_vl_interface': 3.834400486762596e-06, 'learning_rate/action_model': 3.834400486762596e-05, 'epoch': 1.98}) +[2026-06-16 04:03:21,068][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.41006702184677124, 'train/grad_norm_pre_clip': 15.35142707824707, 'timing/data': 0.012592438608407974, 'timing/model': 1.480653304606676, 'learning_rate/qwen_vl_interface': 3.8314937424546585e-06, 'learning_rate/action_model': 3.8314937424546584e-05, 'epoch': 1.98}) +[2026-06-16 04:03:26,951][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.16022610664367676, 'train/grad_norm_pre_clip': 7.361296653747559, 'timing/data': 0.010687447153031826, 'timing/model': 1.4573806654661894, 'learning_rate/qwen_vl_interface': 3.828587581241064e-06, 'learning_rate/action_model': 3.8285875812410635e-05, 'epoch': 1.98}) +[2026-06-16 04:03:32,725][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.10431235283613205, 'train/grad_norm_pre_clip': 9.742778778076172, 'timing/data': 0.010800258256494999, 'timing/model': 1.4431057702749968, 'learning_rate/qwen_vl_interface': 3.825682004316419e-06, 'learning_rate/action_model': 3.825682004316419e-05, 'epoch': 1.98}) +[2026-06-16 04:03:38,559][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.09154829382896423, 'train/grad_norm_pre_clip': 15.007445335388184, 'timing/data': 0.011018238961696625, 'timing/model': 1.4011375047266483, 'learning_rate/qwen_vl_interface': 3.822777012875104e-06, 'learning_rate/action_model': 3.822777012875103e-05, 'epoch': 1.98}) +[2026-06-16 04:03:44,369][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.039011839777231216, 'train/grad_norm_pre_clip': 5.697692394256592, 'timing/data': 0.01140971202403307, 'timing/model': 1.4696042966097593, 'learning_rate/qwen_vl_interface': 3.819872608111248e-06, 'learning_rate/action_model': 3.819872608111248e-05, 'epoch': 1.98}) +[2026-06-16 04:03:50,255][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.06820549070835114, 'train/grad_norm_pre_clip': 6.619784832000732, 'timing/data': 0.014281128533184528, 'timing/model': 1.7355078421533108, 'learning_rate/qwen_vl_interface': 3.816968791218741e-06, 'learning_rate/action_model': 3.816968791218741e-05, 'epoch': 1.98}) +[2026-06-16 04:03:55,862][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.17465829849243164, 'train/grad_norm_pre_clip': 9.943476676940918, 'timing/data': 0.012053295038640499, 'timing/model': 1.4510991824790835, 'learning_rate/qwen_vl_interface': 3.8140655633912337e-06, 'learning_rate/action_model': 3.814065563391234e-05, 'epoch': 1.98}) +[2026-06-16 04:04:01,669][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.35263776779174805, 'train/grad_norm_pre_clip': 12.828927993774414, 'timing/data': 0.012123324908316135, 'timing/model': 1.4769598990678787, 'learning_rate/qwen_vl_interface': 3.811162925822133e-06, 'learning_rate/action_model': 3.811162925822132e-05, 'epoch': 1.98}) +[2026-06-16 04:04:07,459][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.4066317677497864, 'train/grad_norm_pre_clip': 30.472936630249023, 'timing/data': 0.01105789840221405, 'timing/model': 1.4507828615605831, 'learning_rate/qwen_vl_interface': 3.8082608797046045e-06, 'learning_rate/action_model': 3.8082608797046045e-05, 'epoch': 1.98}) +[2026-06-16 04:04:13,260][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.06946219503879547, 'train/grad_norm_pre_clip': 7.536715507507324, 'timing/data': 0.01168153341859579, 'timing/model': 1.3994027562439442, 'learning_rate/qwen_vl_interface': 3.8053594262315687e-06, 'learning_rate/action_model': 3.805359426231568e-05, 'epoch': 1.98}) +[2026-06-16 04:04:19,085][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 0.05297327786684036, 'train/grad_norm_pre_clip': 5.120912551879883, 'timing/data': 0.012497938238084316, 'timing/model': 1.4171560322865844, 'learning_rate/qwen_vl_interface': 3.802458566595704e-06, 'learning_rate/action_model': 3.802458566595704e-05, 'epoch': 1.98}) +[2026-06-16 04:04:25,013][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.1799524426460266, 'train/grad_norm_pre_clip': 23.268314361572266, 'timing/data': 0.010938738472759724, 'timing/model': 1.7290187189355493, 'learning_rate/qwen_vl_interface': 3.7995583019894444e-06, 'learning_rate/action_model': 3.7995583019894445e-05, 'epoch': 1.98}) +[2026-06-16 04:04:30,652][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.026295648887753487, 'train/grad_norm_pre_clip': 2.859098434448242, 'timing/data': 0.011029347777366638, 'timing/model': 1.4321099277585745, 'learning_rate/qwen_vl_interface': 3.7966586336049782e-06, 'learning_rate/action_model': 3.796658633604978e-05, 'epoch': 1.98}) +[2026-06-16 04:04:36,450][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.04407738149166107, 'train/grad_norm_pre_clip': 3.4693570137023926, 'timing/data': 0.011400021612644196, 'timing/model': 1.44640021212399, 'learning_rate/qwen_vl_interface': 3.7937595626342516e-06, 'learning_rate/action_model': 3.793759562634252e-05, 'epoch': 1.98}) +[2026-06-16 04:04:42,262][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.009096983820199966, 'train/grad_norm_pre_clip': 1.0922001600265503, 'timing/data': 0.012655708007514477, 'timing/model': 1.4653414860367775, 'learning_rate/qwen_vl_interface': 3.79086109026896e-06, 'learning_rate/action_model': 3.7908610902689604e-05, 'epoch': 1.98}) +[2026-06-16 04:04:48,069][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.528823733329773, 'train/grad_norm_pre_clip': 19.509838104248047, 'timing/data': 0.011091088876128197, 'timing/model': 1.404403512366116, 'learning_rate/qwen_vl_interface': 3.7879632177005584e-06, 'learning_rate/action_model': 3.787963217700558e-05, 'epoch': 1.98}) +[2026-06-16 04:04:54,046][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.12552708387374878, 'train/grad_norm_pre_clip': 7.936967372894287, 'timing/data': 0.01788123045116663, 'timing/model': 1.7325079254806042, 'learning_rate/qwen_vl_interface': 3.7850659461202523e-06, 'learning_rate/action_model': 3.785065946120252e-05, 'epoch': 1.98}) +[2026-06-16 04:05:00,133][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.13172028958797455, 'train/grad_norm_pre_clip': 7.929198741912842, 'timing/data': 0.01094179879873991, 'timing/model': 1.447080910205841, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 1.98}) +[2026-06-16 04:05:06,149][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.12854933738708496, 'train/grad_norm_pre_clip': 10.593010902404785, 'timing/data': 0.010318573564291, 'timing/model': 1.4496771916747093, 'learning_rate/qwen_vl_interface': 3.779273210687514e-06, 'learning_rate/action_model': 3.779273210687514e-05, 'epoch': 1.99}) +[2026-06-16 04:05:11,757][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.12787415087223053, 'train/grad_norm_pre_clip': 8.293644905090332, 'timing/data': 0.012087404727935791, 'timing/model': 1.4427024349570274, 'learning_rate/qwen_vl_interface': 3.776377749216257e-06, 'learning_rate/action_model': 3.776377749216257e-05, 'epoch': 1.99}) +[2026-06-16 04:05:17,561][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.024192331358790398, 'train/grad_norm_pre_clip': 2.3037455081939697, 'timing/data': 0.011601882055401802, 'timing/model': 1.4606366818770766, 'learning_rate/qwen_vl_interface': 3.773482893495442e-06, 'learning_rate/action_model': 3.773482893495442e-05, 'epoch': 1.99}) +[2026-06-16 04:05:23,350][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.008802765049040318, 'train/grad_norm_pre_clip': 0.980452299118042, 'timing/data': 0.010423204861581326, 'timing/model': 1.4448327906429768, 'learning_rate/qwen_vl_interface': 3.7705886447150375e-06, 'learning_rate/action_model': 3.7705886447150376e-05, 'epoch': 1.99}) +[2026-06-16 04:05:29,221][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.02374344877898693, 'train/grad_norm_pre_clip': 2.056464195251465, 'timing/data': 0.011122448369860649, 'timing/model': 1.736154350452125, 'learning_rate/qwen_vl_interface': 3.767695004064759e-06, 'learning_rate/action_model': 3.767695004064759e-05, 'epoch': 1.99}) +[2026-06-16 04:05:34,796][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 0.0483333095908165, 'train/grad_norm_pre_clip': 3.867448568344116, 'timing/data': 0.010440174490213394, 'timing/model': 1.4894189992919564, 'learning_rate/qwen_vl_interface': 3.7648019727340725e-06, 'learning_rate/action_model': 3.764801972734072e-05, 'epoch': 1.99}) +[2026-06-16 04:05:40,858][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.10714951902627945, 'train/grad_norm_pre_clip': 10.803668022155762, 'timing/data': 0.010795227251946926, 'timing/model': 1.4470269531011581, 'learning_rate/qwen_vl_interface': 3.761909551912196e-06, 'learning_rate/action_model': 3.761909551912196e-05, 'epoch': 1.99}) +[2026-06-16 04:05:46,700][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.23172207176685333, 'train/grad_norm_pre_clip': 23.39739990234375, 'timing/data': 0.011053187772631645, 'timing/model': 1.505910242907703, 'learning_rate/qwen_vl_interface': 3.75901774278809e-06, 'learning_rate/action_model': 3.7590177427880904e-05, 'epoch': 1.99}) +[2026-06-16 04:05:52,603][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.016607433557510376, 'train/grad_norm_pre_clip': 2.65575909614563, 'timing/data': 0.010776075534522533, 'timing/model': 1.5026272889226675, 'learning_rate/qwen_vl_interface': 3.756126546550472e-06, 'learning_rate/action_model': 3.756126546550472e-05, 'epoch': 1.99}) +[2026-06-16 04:05:58,359][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.07852070778608322, 'train/grad_norm_pre_clip': 11.493751525878906, 'timing/data': 0.011310880072414875, 'timing/model': 1.4399578180164099, 'learning_rate/qwen_vl_interface': 3.753235964387802e-06, 'learning_rate/action_model': 3.753235964387802e-05, 'epoch': 1.99}) +[2026-06-16 04:06:04,215][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.1249738335609436, 'train/grad_norm_pre_clip': 12.186129570007324, 'timing/data': 0.012906359508633614, 'timing/model': 1.7126996694132686, 'learning_rate/qwen_vl_interface': 3.750345997488289e-06, 'learning_rate/action_model': 3.7503459974882886e-05, 'epoch': 1.99}) +[2026-06-16 04:06:09,748][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.10022716969251633, 'train/grad_norm_pre_clip': 7.122542858123779, 'timing/data': 0.010749486275017262, 'timing/model': 1.4530668118968606, 'learning_rate/qwen_vl_interface': 3.7474566470398903e-06, 'learning_rate/action_model': 3.74745664703989e-05, 'epoch': 1.99}) +[2026-06-16 04:06:15,557][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 0.15663012862205505, 'train/grad_norm_pre_clip': 6.980619430541992, 'timing/data': 0.012510267086327076, 'timing/model': 1.4606629852205515, 'learning_rate/qwen_vl_interface': 3.744567914230307e-06, 'learning_rate/action_model': 3.744567914230307e-05, 'epoch': 1.99}) +[2026-06-16 04:06:21,345][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.18632380664348602, 'train/grad_norm_pre_clip': 14.011300086975098, 'timing/data': 0.010318653658032417, 'timing/model': 1.447533410973847, 'learning_rate/qwen_vl_interface': 3.7416798002469902e-06, 'learning_rate/action_model': 3.74167980024699e-05, 'epoch': 1.99}) +[2026-06-16 04:06:27,150][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.1799604892730713, 'train/grad_norm_pre_clip': 13.080613136291504, 'timing/data': 0.010521074756979942, 'timing/model': 1.3930798219516873, 'learning_rate/qwen_vl_interface': 3.7387923062771313e-06, 'learning_rate/action_model': 3.738792306277131e-05, 'epoch': 1.99}) +[2026-06-16 04:06:33,009][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.24945461750030518, 'train/grad_norm_pre_clip': 14.319916725158691, 'timing/data': 0.011373420245945454, 'timing/model': 1.7096330597996712, 'learning_rate/qwen_vl_interface': 3.7359054335076732e-06, 'learning_rate/action_model': 3.7359054335076734e-05, 'epoch': 1.99}) +[2026-06-16 04:06:38,719][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.26318418979644775, 'train/grad_norm_pre_clip': 11.79427433013916, 'timing/data': 0.010417183861136436, 'timing/model': 1.6327022621408105, 'learning_rate/qwen_vl_interface': 3.7330191831253006e-06, 'learning_rate/action_model': 3.7330191831253e-05, 'epoch': 1.99}) +[2026-06-16 04:06:44,362][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.040179722011089325, 'train/grad_norm_pre_clip': 3.9613029956817627, 'timing/data': 0.010530835017561913, 'timing/model': 1.4634818052873015, 'learning_rate/qwen_vl_interface': 3.7301335563164382e-06, 'learning_rate/action_model': 3.730133556316438e-05, 'epoch': 1.99}) +[2026-06-16 04:06:50,164][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.15869668126106262, 'train/grad_norm_pre_clip': 5.759524345397949, 'timing/data': 0.010970978066325188, 'timing/model': 1.4741925476118922, 'learning_rate/qwen_vl_interface': 3.727248554267261e-06, 'learning_rate/action_model': 3.727248554267261e-05, 'epoch': 1.99}) +[2026-06-16 04:06:55,966][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.04918118193745613, 'train/grad_norm_pre_clip': 3.7859878540039062, 'timing/data': 0.011228368617594242, 'timing/model': 1.4771987041458488, 'learning_rate/qwen_vl_interface': 3.7243641781636838e-06, 'learning_rate/action_model': 3.7243641781636836e-05, 'epoch': 1.99}) +[2026-06-16 04:07:01,750][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.13644230365753174, 'train/grad_norm_pre_clip': 10.871101379394531, 'timing/data': 0.011319269426167011, 'timing/model': 1.403335776180029, 'learning_rate/qwen_vl_interface': 3.721480429191364e-06, 'learning_rate/action_model': 3.721480429191364e-05, 'epoch': 1.99}) +[2026-06-16 04:07:07,616][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.27184078097343445, 'train/grad_norm_pre_clip': 13.227118492126465, 'timing/data': 0.010762705467641354, 'timing/model': 1.724583338946104, 'learning_rate/qwen_vl_interface': 3.7185973085357045e-06, 'learning_rate/action_model': 3.718597308535705e-05, 'epoch': 1.99}) +[2026-06-16 04:07:13,428][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 0.07481568306684494, 'train/grad_norm_pre_clip': 12.88875961303711, 'timing/data': 0.011146757751703262, 'timing/model': 1.6229752907529473, 'learning_rate/qwen_vl_interface': 3.7157148173818443e-06, 'learning_rate/action_model': 3.715714817381844e-05, 'epoch': 2.0}) +[2026-06-16 04:07:19,059][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.035655006766319275, 'train/grad_norm_pre_clip': 3.1516475677490234, 'timing/data': 0.011807642877101898, 'timing/model': 1.450719098560512, 'learning_rate/qwen_vl_interface': 3.7128329569146672e-06, 'learning_rate/action_model': 3.7128329569146666e-05, 'epoch': 2.0}) +[2026-06-16 04:07:24,873][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.1020210012793541, 'train/grad_norm_pre_clip': 6.167387962341309, 'timing/data': 0.011059357784688473, 'timing/model': 1.4808129509910941, 'learning_rate/qwen_vl_interface': 3.7099517283187977e-06, 'learning_rate/action_model': 3.7099517283187976e-05, 'epoch': 2.0}) +[2026-06-16 04:07:30,761][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.0967462807893753, 'train/grad_norm_pre_clip': 6.163082599639893, 'timing/data': 0.011326850391924381, 'timing/model': 1.455865123309195, 'learning_rate/qwen_vl_interface': 3.7070711327786e-06, 'learning_rate/action_model': 3.7070711327785997e-05, 'epoch': 2.0}) +[2026-06-16 04:07:36,549][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.14031638205051422, 'train/grad_norm_pre_clip': 11.555048942565918, 'timing/data': 0.011592371389269829, 'timing/model': 1.4008151227608323, 'learning_rate/qwen_vl_interface': 3.704191171478179e-06, 'learning_rate/action_model': 3.704191171478179e-05, 'epoch': 2.0}) +[2026-06-16 04:07:42,415][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.06265581399202347, 'train/grad_norm_pre_clip': 4.352343559265137, 'timing/data': 0.011582180857658386, 'timing/model': 1.7241420149803162, 'learning_rate/qwen_vl_interface': 3.701311845601376e-06, 'learning_rate/action_model': 3.701311845601376e-05, 'epoch': 2.0}) +[2026-06-16 04:07:48,055][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.14411625266075134, 'train/grad_norm_pre_clip': 10.331104278564453, 'timing/data': 0.011589630506932735, 'timing/model': 1.4567355494946241, 'learning_rate/qwen_vl_interface': 3.6984331563317744e-06, 'learning_rate/action_model': 3.698433156331774e-05, 'epoch': 2.0}) +[2026-06-16 04:07:53,857][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.0823068618774414, 'train/grad_norm_pre_clip': 7.856263160705566, 'timing/data': 0.01213634479790926, 'timing/model': 1.4594106636941433, 'learning_rate/qwen_vl_interface': 3.695555104852695e-06, 'learning_rate/action_model': 3.6955551048526946e-05, 'epoch': 2.0}) +[2026-06-16 04:07:59,665][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.1819021850824356, 'train/grad_norm_pre_clip': 9.772793769836426, 'timing/data': 0.012149184010922909, 'timing/model': 1.4651551162824035, 'learning_rate/qwen_vl_interface': 3.6926776923471948e-06, 'learning_rate/action_model': 3.6926776923471946e-05, 'epoch': 2.0}) +[2026-06-16 04:08:05,453][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 0.07622656971216202, 'train/grad_norm_pre_clip': 13.983147621154785, 'timing/data': 0.011961263604462147, 'timing/model': 1.4637890355661511, 'learning_rate/qwen_vl_interface': 3.6898009199980715e-06, 'learning_rate/action_model': 3.6898009199980714e-05, 'epoch': 2.0}) +[2026-06-16 04:08:11,260][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.04728865250945091, 'train/grad_norm_pre_clip': 3.1323022842407227, 'timing/data': 0.011767862364649773, 'timing/model': 1.4008952463045716, 'learning_rate/qwen_vl_interface': 3.686924788987855e-06, 'learning_rate/action_model': 3.686924788987855e-05, 'epoch': 2.0}) +[2026-06-16 04:08:17,122][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 0.1488512009382248, 'train/grad_norm_pre_clip': 6.022425651550293, 'timing/data': 0.011846853420138359, 'timing/model': 1.7285564625635743, 'learning_rate/qwen_vl_interface': 3.6840493004988155e-06, 'learning_rate/action_model': 3.684049300498815e-05, 'epoch': 2.0}) +[2026-06-16 04:08:22,766][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.013473538681864738, 'train/grad_norm_pre_clip': 2.2793643474578857, 'timing/data': 0.012023033574223518, 'timing/model': 1.4591948073357344, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 2.0}) +[2026-06-16 04:08:28,556][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.11426205188035965, 'train/grad_norm_pre_clip': 7.117448329925537, 'timing/data': 0.01102397870272398, 'timing/model': 1.4622376738116145, 'learning_rate/qwen_vl_interface': 3.6783002558120206e-06, 'learning_rate/action_model': 3.67830025581202e-05, 'epoch': 2.0}) +[2026-06-16 04:08:34,373][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.09079122543334961, 'train/grad_norm_pre_clip': 8.88044261932373, 'timing/data': 0.010909046977758408, 'timing/model': 1.4497189950197935, 'learning_rate/qwen_vl_interface': 3.6754267019774804e-06, 'learning_rate/action_model': 3.67542670197748e-05, 'epoch': 2.0}) +[2026-06-16 04:08:40,354][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 0.34224793314933777, 'train/grad_norm_pre_clip': 12.055102348327637, 'timing/data': 0.011788901872932911, 'timing/model': 1.4506409876048565, 'learning_rate/qwen_vl_interface': 3.6725537953905453e-06, 'learning_rate/action_model': 3.672553795390545e-05, 'epoch': 2.0}) +[2026-06-16 04:08:46,169][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.026245450600981712, 'train/grad_norm_pre_clip': 3.0417678356170654, 'timing/data': 0.011251039803028107, 'timing/model': 1.402136948890984, 'learning_rate/qwen_vl_interface': 3.6696815372321603e-06, 'learning_rate/action_model': 3.6696815372321604e-05, 'epoch': 2.0}) +[2026-06-16 04:08:51,969][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.060566484928131104, 'train/grad_norm_pre_clip': 5.677754878997803, 'timing/data': 0.011285688728094101, 'timing/model': 1.687644143588841, 'learning_rate/qwen_vl_interface': 3.6668099286830023e-06, 'learning_rate/action_model': 3.666809928683002e-05, 'epoch': 2.0}) +[2026-06-16 04:08:57,753][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.0595407709479332, 'train/grad_norm_pre_clip': 3.202604293823242, 'timing/data': 0.012733157724142075, 'timing/model': 1.6665358999744058, 'learning_rate/qwen_vl_interface': 3.66393897092348e-06, 'learning_rate/action_model': 3.66393897092348e-05, 'epoch': 2.0}) +[2026-06-16 04:09:03,364][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.06599396467208862, 'train/grad_norm_pre_clip': 6.62102746963501, 'timing/data': 0.011229208670556545, 'timing/model': 1.4757222812622786, 'learning_rate/qwen_vl_interface': 3.661068665133735e-06, 'learning_rate/action_model': 3.661068665133735e-05, 'epoch': 2.0}) +[2026-06-16 04:09:09,161][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.06143217533826828, 'train/grad_norm_pre_clip': 3.7181389331817627, 'timing/data': 0.011202218011021614, 'timing/model': 1.4730348233133554, 'learning_rate/qwen_vl_interface': 3.6581990124936434e-06, 'learning_rate/action_model': 3.6581990124936434e-05, 'epoch': 2.0}) +[2026-06-16 04:09:15,192][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.06376112252473831, 'train/grad_norm_pre_clip': 5.6506428718566895, 'timing/data': 0.015086621977388859, 'timing/model': 1.487552679143846, 'learning_rate/qwen_vl_interface': 3.65533001418281e-06, 'learning_rate/action_model': 3.655330014182809e-05, 'epoch': 2.0}) +[2026-06-16 04:09:21,084][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.018024319782853127, 'train/grad_norm_pre_clip': 1.4767088890075684, 'timing/data': 0.011728370562195778, 'timing/model': 1.4206055672839284, 'learning_rate/qwen_vl_interface': 3.652461671380572e-06, 'learning_rate/action_model': 3.652461671380572e-05, 'epoch': 2.01}) +[2026-06-16 04:09:27,021][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 0.13880452513694763, 'train/grad_norm_pre_clip': 11.63520336151123, 'timing/data': 0.012119214050471783, 'timing/model': 1.712235345505178, 'learning_rate/qwen_vl_interface': 3.6495939852659966e-06, 'learning_rate/action_model': 3.649593985265997e-05, 'epoch': 2.01}) +[2026-06-16 04:09:32,953][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.11532243341207504, 'train/grad_norm_pre_clip': 9.036685943603516, 'timing/data': 0.011338158510625362, 'timing/model': 1.7506557377055287, 'learning_rate/qwen_vl_interface': 3.646726957017881e-06, 'learning_rate/action_model': 3.646726957017881e-05, 'epoch': 2.01}) +[2026-06-16 04:09:38,570][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 0.13888324797153473, 'train/grad_norm_pre_clip': 17.048507690429688, 'timing/data': 0.011463519185781479, 'timing/model': 1.4581551402807236, 'learning_rate/qwen_vl_interface': 3.643860587814752e-06, 'learning_rate/action_model': 3.643860587814752e-05, 'epoch': 2.01}) +[2026-06-16 04:09:44,460][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.06414882838726044, 'train/grad_norm_pre_clip': 7.99823522567749, 'timing/data': 0.011196518316864967, 'timing/model': 1.448531238362193, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 2.01}) +[2026-06-16 04:09:50,272][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.08623738586902618, 'train/grad_norm_pre_clip': 12.225932121276855, 'timing/data': 0.011554449796676636, 'timing/model': 1.4719910798594356, 'learning_rate/qwen_vl_interface': 3.6381298312562095e-06, 'learning_rate/action_model': 3.6381298312562094e-05, 'epoch': 2.01}) +[2026-06-16 04:09:56,190][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.026204349473118782, 'train/grad_norm_pre_clip': 3.519416093826294, 'timing/data': 0.013905094005167484, 'timing/model': 1.4174688942730427, 'learning_rate/qwen_vl_interface': 3.635265446256493e-06, 'learning_rate/action_model': 3.6352654462564926e-05, 'epoch': 2.01}) +[2026-06-16 04:10:02,109][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.18529973924160004, 'train/grad_norm_pre_clip': 20.950481414794922, 'timing/data': 0.011063107289373875, 'timing/model': 1.704841423779726, 'learning_rate/qwen_vl_interface': 3.6324017250131565e-06, 'learning_rate/action_model': 3.6324017250131567e-05, 'epoch': 2.01}) +[2026-06-16 04:10:07,763][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.02387923188507557, 'train/grad_norm_pre_clip': 2.7174298763275146, 'timing/data': 0.011630861088633537, 'timing/model': 1.4714806703850627, 'learning_rate/qwen_vl_interface': 3.6295386687033696e-06, 'learning_rate/action_model': 3.6295386687033695e-05, 'epoch': 2.01}) +[2026-06-16 04:10:13,669][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.06836828589439392, 'train/grad_norm_pre_clip': 7.721752643585205, 'timing/data': 0.01101811695843935, 'timing/model': 1.4653517995029688, 'learning_rate/qwen_vl_interface': 3.6266762785040253e-06, 'learning_rate/action_model': 3.626676278504025e-05, 'epoch': 2.01}) +[2026-06-16 04:10:19,490][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 0.17530550062656403, 'train/grad_norm_pre_clip': 10.812528610229492, 'timing/data': 0.010993997566401958, 'timing/model': 1.489210331812501, 'learning_rate/qwen_vl_interface': 3.6238145555917464e-06, 'learning_rate/action_model': 3.6238145555917465e-05, 'epoch': 2.01}) +[2026-06-16 04:10:25,355][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.02597004361450672, 'train/grad_norm_pre_clip': 2.716123342514038, 'timing/data': 0.0109859062358737, 'timing/model': 1.4043982233852148, 'learning_rate/qwen_vl_interface': 3.6209535011428757e-06, 'learning_rate/action_model': 3.620953501142876e-05, 'epoch': 2.01}) +[2026-06-16 04:10:31,154][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.09758295118808746, 'train/grad_norm_pre_clip': 9.156879425048828, 'timing/data': 0.010900106281042099, 'timing/model': 1.4596965685486794, 'learning_rate/qwen_vl_interface': 3.6180931163334863e-06, 'learning_rate/action_model': 3.6180931163334863e-05, 'epoch': 2.01}) +[2026-06-16 04:10:36,972][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.11613860726356506, 'train/grad_norm_pre_clip': 10.860904693603516, 'timing/data': 0.010803574696183205, 'timing/model': 1.671915078535676, 'learning_rate/qwen_vl_interface': 3.6152334023393747e-06, 'learning_rate/action_model': 3.6152334023393745e-05, 'epoch': 2.01}) +[2026-06-16 04:10:42,555][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.05511113256216049, 'train/grad_norm_pre_clip': 6.600572109222412, 'timing/data': 0.011219427920877934, 'timing/model': 1.43621177226305, 'learning_rate/qwen_vl_interface': 3.6123743603360613e-06, 'learning_rate/action_model': 3.612374360336061e-05, 'epoch': 2.01}) +[2026-06-16 04:21:17,879][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.004573776852339506, 'train/grad_norm_pre_clip': 0.8267258405685425, 'eval/loss': 0.07366959750652313, 'eval/flap_precision': 0.9357798165137615, 'eval/flap_recall': 0.7418181818181818, 'eval/flap_f1': 0.8275862068965518, 'eval/flap_rate_model': 0.06055555555555556, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.8275862068965518, 'timing/data': 0.01169890072196722, 'timing/model': 1.4680570224300027, 'learning_rate/qwen_vl_interface': 3.6095159914987893e-06, 'learning_rate/action_model': 3.6095159914987894e-05, 'epoch': 2.01}) +[2026-06-16 04:21:17,880][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4400_state +[2026-06-16 04:21:17,888][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 04:21:25,618][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4400_state/model.safetensors +[2026-06-16 04:21:37,452][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4400_state/optimizer.bin +[2026-06-16 04:21:37,453][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4400_state/sampler.bin +[2026-06-16 04:21:37,454][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4400_state/sampler_1.bin +[2026-06-16 04:21:37,455][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4400_state/random_states_0.pkl +[2026-06-16 04:21:37,624][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4400 +[2026-06-16 04:21:37,625][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 04:21:37,640][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 04:21:43,771][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.1247430145740509, 'train/grad_norm_pre_clip': 9.856354713439941, 'timing/data': 0.014472422190010548, 'timing/model': 1.408769691362977, 'learning_rate/qwen_vl_interface': 3.6066582970025266e-06, 'learning_rate/action_model': 3.6066582970025264e-05, 'epoch': 2.01}) +[2026-06-16 04:21:49,756][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.130124032497406, 'train/grad_norm_pre_clip': 10.977583885192871, 'timing/data': 0.014653402380645275, 'timing/model': 1.4679918130859733, 'learning_rate/qwen_vl_interface': 3.603801278021964e-06, 'learning_rate/action_model': 3.6038012780219635e-05, 'epoch': 2.01}) +[2026-06-16 04:21:55,589][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.08279385417699814, 'train/grad_norm_pre_clip': 12.9620361328125, 'timing/data': 0.011609146371483803, 'timing/model': 1.4069593958556652, 'learning_rate/qwen_vl_interface': 3.6009449357315122e-06, 'learning_rate/action_model': 3.6009449357315125e-05, 'epoch': 2.01}) +[2026-06-16 04:22:01,475][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.01909615658223629, 'train/grad_norm_pre_clip': 2.898780107498169, 'timing/data': 0.011567886918783188, 'timing/model': 1.3973168516531587, 'learning_rate/qwen_vl_interface': 3.5980892713053084e-06, 'learning_rate/action_model': 3.598089271305308e-05, 'epoch': 2.01}) +[2026-06-16 04:22:07,367][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.23930558562278748, 'train/grad_norm_pre_clip': 19.883394241333008, 'timing/data': 0.01135648600757122, 'timing/model': 1.6921929167583585, 'learning_rate/qwen_vl_interface': 3.5952342859172063e-06, 'learning_rate/action_model': 3.595234285917206e-05, 'epoch': 2.01}) +[2026-06-16 04:22:12,845][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.2767004370689392, 'train/grad_norm_pre_clip': 10.810037612915039, 'timing/data': 0.011872097849845886, 'timing/model': 1.3899553455412388, 'learning_rate/qwen_vl_interface': 3.592379980740785e-06, 'learning_rate/action_model': 3.5923799807407845e-05, 'epoch': 2.02}) +[2026-06-16 04:22:18,744][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.2132030725479126, 'train/grad_norm_pre_clip': 15.959612846374512, 'timing/data': 0.011322947219014168, 'timing/model': 1.4601454474031925, 'learning_rate/qwen_vl_interface': 3.5895263569493384e-06, 'learning_rate/action_model': 3.589526356949338e-05, 'epoch': 2.02}) +[2026-06-16 04:22:24,658][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.5213136672973633, 'train/grad_norm_pre_clip': 27.118261337280273, 'timing/data': 0.011717097833752632, 'timing/model': 1.4739492675289512, 'learning_rate/qwen_vl_interface': 3.586673415715885e-06, 'learning_rate/action_model': 3.586673415715885e-05, 'epoch': 2.02}) +[2026-06-16 04:22:30,557][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 0.261094331741333, 'train/grad_norm_pre_clip': 24.29684829711914, 'timing/data': 0.01266615092754364, 'timing/model': 1.4566597444936633, 'learning_rate/qwen_vl_interface': 3.5838211582131603e-06, 'learning_rate/action_model': 3.58382115821316e-05, 'epoch': 2.02}) +[2026-06-16 04:22:36,360][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.31838861107826233, 'train/grad_norm_pre_clip': 12.069210052490234, 'timing/data': 0.011659207753837109, 'timing/model': 1.3958442844450474, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 2.02}) +[2026-06-16 04:22:42,185][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.11124330013990402, 'train/grad_norm_pre_clip': 26.210079193115234, 'timing/data': 0.01215972937643528, 'timing/model': 1.7201150227338076, 'learning_rate/qwen_vl_interface': 3.5781186990894423e-06, 'learning_rate/action_model': 3.578118699089442e-05, 'epoch': 2.02}) +[2026-06-16 04:22:47,754][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.214253768324852, 'train/grad_norm_pre_clip': 14.491612434387207, 'timing/data': 0.011479568667709827, 'timing/model': 1.439281907863915, 'learning_rate/qwen_vl_interface': 3.5752684998125113e-06, 'learning_rate/action_model': 3.5752684998125116e-05, 'epoch': 2.02}) +[2026-06-16 04:22:53,561][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.09946643561124802, 'train/grad_norm_pre_clip': 13.033788681030273, 'timing/data': 0.011573858559131622, 'timing/model': 1.4479291634634137, 'learning_rate/qwen_vl_interface': 3.5724189889544407e-06, 'learning_rate/action_model': 3.572418988954441e-05, 'epoch': 2.02}) +[2026-06-16 04:22:59,466][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.03466109186410904, 'train/grad_norm_pre_clip': 5.680902004241943, 'timing/data': 0.011431679129600525, 'timing/model': 1.4666779870167375, 'learning_rate/qwen_vl_interface': 3.569570167686556e-06, 'learning_rate/action_model': 3.569570167686556e-05, 'epoch': 2.02}) +[2026-06-16 04:23:05,348][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.07087968289852142, 'train/grad_norm_pre_clip': 9.590628623962402, 'timing/data': 0.01179549004882574, 'timing/model': 1.462044820189476, 'learning_rate/qwen_vl_interface': 3.5667220371798995e-06, 'learning_rate/action_model': 3.566722037179899e-05, 'epoch': 2.02}) +[2026-06-16 04:23:11,256][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.13276365399360657, 'train/grad_norm_pre_clip': 15.279376983642578, 'timing/data': 0.014461996965110302, 'timing/model': 1.4195334454998374, 'learning_rate/qwen_vl_interface': 3.563874598605232e-06, 'learning_rate/action_model': 3.5638745986052316e-05, 'epoch': 2.02}) +[2026-06-16 04:23:17,203][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 0.27673274278640747, 'train/grad_norm_pre_clip': 13.874218940734863, 'timing/data': 0.011175019666552544, 'timing/model': 1.8016049601137638, 'learning_rate/qwen_vl_interface': 3.561027853133025e-06, 'learning_rate/action_model': 3.561027853133025e-05, 'epoch': 2.02}) +[2026-06-16 04:23:22,868][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.06435362994670868, 'train/grad_norm_pre_clip': 4.173393726348877, 'timing/data': 0.01150157954543829, 'timing/model': 1.4737401911988854, 'learning_rate/qwen_vl_interface': 3.558181801933472e-06, 'learning_rate/action_model': 3.5581818019334714e-05, 'epoch': 2.02}) +[2026-06-16 04:23:28,759][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.05277235060930252, 'train/grad_norm_pre_clip': 8.70429515838623, 'timing/data': 0.011716781184077263, 'timing/model': 1.4562548408284783, 'learning_rate/qwen_vl_interface': 3.555336446176473e-06, 'learning_rate/action_model': 3.5553364461764726e-05, 'epoch': 2.02}) +[2026-06-16 04:23:34,664][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.012272621504962444, 'train/grad_norm_pre_clip': 1.7123621702194214, 'timing/data': 0.01110168918967247, 'timing/model': 1.4493664540350437, 'learning_rate/qwen_vl_interface': 3.552491787031647e-06, 'learning_rate/action_model': 3.552491787031647e-05, 'epoch': 2.02}) +[2026-06-16 04:23:40,567][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.11109212040901184, 'train/grad_norm_pre_clip': 13.53807544708252, 'timing/data': 0.012786944396793842, 'timing/model': 1.4345350367948413, 'learning_rate/qwen_vl_interface': 3.5496478256683277e-06, 'learning_rate/action_model': 3.549647825668327e-05, 'epoch': 2.02}) +[2026-06-16 04:23:46,472][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.23079414665699005, 'train/grad_norm_pre_clip': 15.463071823120117, 'timing/data': 0.013211744837462902, 'timing/model': 1.7542516235262156, 'learning_rate/qwen_vl_interface': 3.5468045632555585e-06, 'learning_rate/action_model': 3.546804563255559e-05, 'epoch': 2.02}) +[2026-06-16 04:23:52,288][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.015615971758961678, 'train/grad_norm_pre_clip': 2.3465960025787354, 'timing/data': 0.011843191459774971, 'timing/model': 1.6937492564320564, 'learning_rate/qwen_vl_interface': 3.5439620009620995e-06, 'learning_rate/action_model': 3.543962000962099e-05, 'epoch': 2.02}) +[2026-06-16 04:23:57,961][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.20303303003311157, 'train/grad_norm_pre_clip': 8.407928466796875, 'timing/data': 0.011235191486775875, 'timing/model': 1.4590680664405227, 'learning_rate/qwen_vl_interface': 3.541120139956419e-06, 'learning_rate/action_model': 3.5411201399564184e-05, 'epoch': 2.02}) +[2026-06-16 04:24:03,864][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 0.08644157648086548, 'train/grad_norm_pre_clip': 9.161059379577637, 'timing/data': 0.012022373266518116, 'timing/model': 1.4585353266447783, 'learning_rate/qwen_vl_interface': 3.538278981406697e-06, 'learning_rate/action_model': 3.5382789814066966e-05, 'epoch': 2.02}) +[2026-06-16 04:24:09,763][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.09557701647281647, 'train/grad_norm_pre_clip': 15.256190299987793, 'timing/data': 0.01240859366953373, 'timing/model': 1.4593651490285993, 'learning_rate/qwen_vl_interface': 3.5354385264808295e-06, 'learning_rate/action_model': 3.535438526480829e-05, 'epoch': 2.02}) +[2026-06-16 04:24:15,660][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.08432535827159882, 'train/grad_norm_pre_clip': 13.218542098999023, 'timing/data': 0.012089723721146584, 'timing/model': 1.4457071013748646, 'learning_rate/qwen_vl_interface': 3.5325987763464196e-06, 'learning_rate/action_model': 3.5325987763464194e-05, 'epoch': 2.02}) +[2026-06-16 04:24:21,574][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 0.26831239461898804, 'train/grad_norm_pre_clip': 16.68096351623535, 'timing/data': 0.011271032504737377, 'timing/model': 1.7770002000033855, 'learning_rate/qwen_vl_interface': 3.529759732170781e-06, 'learning_rate/action_model': 3.529759732170781e-05, 'epoch': 2.03}) +[2026-06-16 04:24:27,405][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.1031053364276886, 'train/grad_norm_pre_clip': 7.550551414489746, 'timing/data': 0.01222948543727398, 'timing/model': 1.7118723699823022, 'learning_rate/qwen_vl_interface': 3.5269213951209375e-06, 'learning_rate/action_model': 3.5269213951209374e-05, 'epoch': 2.03}) +[2026-06-16 04:24:33,057][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.030789317563176155, 'train/grad_norm_pre_clip': 3.334291934967041, 'timing/data': 0.011523563414812088, 'timing/model': 1.4499809984117746, 'learning_rate/qwen_vl_interface': 3.5240837663636207e-06, 'learning_rate/action_model': 3.5240837663636204e-05, 'epoch': 2.03}) +[2026-06-16 04:24:38,866][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.11476543545722961, 'train/grad_norm_pre_clip': 14.158357620239258, 'timing/data': 0.011886575259268284, 'timing/model': 1.4673968981951475, 'learning_rate/qwen_vl_interface': 3.5212468470652745e-06, 'learning_rate/action_model': 3.521246847065274e-05, 'epoch': 2.03}) +[2026-06-16 04:24:44,773][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.01964578963816166, 'train/grad_norm_pre_clip': 5.096052169799805, 'timing/data': 0.012099254876375198, 'timing/model': 1.4733875365927815, 'learning_rate/qwen_vl_interface': 3.5184106383920487e-06, 'learning_rate/action_model': 3.5184106383920486e-05, 'epoch': 2.03}) +[2026-06-16 04:24:50,758][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.3023812174797058, 'train/grad_norm_pre_clip': 13.183121681213379, 'timing/data': 0.011850914917886257, 'timing/model': 1.4450626811012626, 'learning_rate/qwen_vl_interface': 3.5155751415098022e-06, 'learning_rate/action_model': 3.515575141509802e-05, 'epoch': 2.03}) +[2026-06-16 04:24:56,681][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.17572320997714996, 'train/grad_norm_pre_clip': 15.366388320922852, 'timing/data': 0.011454343795776367, 'timing/model': 1.7740453006699681, 'learning_rate/qwen_vl_interface': 3.5127403575841e-06, 'learning_rate/action_model': 3.5127403575841e-05, 'epoch': 2.03}) +[2026-06-16 04:25:02,359][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.2893170714378357, 'train/grad_norm_pre_clip': 19.25371551513672, 'timing/data': 0.011792315170168877, 'timing/model': 1.4510290836915374, 'learning_rate/qwen_vl_interface': 3.509906287780215e-06, 'learning_rate/action_model': 3.5099062877802144e-05, 'epoch': 2.03}) +[2026-06-16 04:25:08,246][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.034041352570056915, 'train/grad_norm_pre_clip': 2.6398277282714844, 'timing/data': 0.011942815966904163, 'timing/model': 1.433756498619914, 'learning_rate/qwen_vl_interface': 3.5070729332631247e-06, 'learning_rate/action_model': 3.507072933263125e-05, 'epoch': 2.03}) +[2026-06-16 04:25:14,075][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.15372717380523682, 'train/grad_norm_pre_clip': 10.524909019470215, 'timing/data': 0.012864337302744389, 'timing/model': 1.3986144298687577, 'learning_rate/qwen_vl_interface': 3.5042402951975167e-06, 'learning_rate/action_model': 3.5042402951975167e-05, 'epoch': 2.03}) +[2026-06-16 04:25:20,062][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 0.018175411969423294, 'train/grad_norm_pre_clip': 2.421922445297241, 'timing/data': 0.011493395082652569, 'timing/model': 1.4660857496783137, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 2.03}) +[2026-06-16 04:25:25,953][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.03953462094068527, 'train/grad_norm_pre_clip': 8.653251647949219, 'timing/data': 0.011506195180118084, 'timing/model': 1.425406931899488, 'learning_rate/qwen_vl_interface': 3.498577173078011e-06, 'learning_rate/action_model': 3.498577173078011e-05, 'epoch': 2.03}) +[2026-06-16 04:25:31,921][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.18809789419174194, 'train/grad_norm_pre_clip': 13.86814022064209, 'timing/data': 0.013771722093224525, 'timing/model': 1.8190604643896222, 'learning_rate/qwen_vl_interface': 3.495746691352008e-06, 'learning_rate/action_model': 3.495746691352008e-05, 'epoch': 2.03}) +[2026-06-16 04:25:37,547][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.10464794188737869, 'train/grad_norm_pre_clip': 9.908361434936523, 'timing/data': 0.010766472667455673, 'timing/model': 1.4426233572885394, 'learning_rate/qwen_vl_interface': 3.4929169307332763e-06, 'learning_rate/action_model': 3.492916930733276e-05, 'epoch': 2.03}) +[2026-06-16 04:25:43,371][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.03986573964357376, 'train/grad_norm_pre_clip': 5.682047367095947, 'timing/data': 0.010906844399869442, 'timing/model': 1.3963088681921363, 'learning_rate/qwen_vl_interface': 3.4900878923850234e-06, 'learning_rate/action_model': 3.490087892385023e-05, 'epoch': 2.03}) +[2026-06-16 04:25:49,267][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.0404055081307888, 'train/grad_norm_pre_clip': 4.256624698638916, 'timing/data': 0.013026710599660873, 'timing/model': 1.4836141234263778, 'learning_rate/qwen_vl_interface': 3.4872595774701607e-06, 'learning_rate/action_model': 3.487259577470161e-05, 'epoch': 2.03}) +[2026-06-16 04:25:55,146][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.1771480292081833, 'train/grad_norm_pre_clip': 8.310486793518066, 'timing/data': 0.010871733538806438, 'timing/model': 1.4582225289195776, 'learning_rate/qwen_vl_interface': 3.4844319871513014e-06, 'learning_rate/action_model': 3.484431987151301e-05, 'epoch': 2.03}) +[2026-06-16 04:26:01,046][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.08019334822893143, 'train/grad_norm_pre_clip': 8.284322738647461, 'timing/data': 0.011113155633211136, 'timing/model': 1.4302558256313205, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 2.03}) +[2026-06-16 04:26:07,062][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.056154508143663406, 'train/grad_norm_pre_clip': 6.217189311981201, 'timing/data': 0.01044550258666277, 'timing/model': 1.8590667368844151, 'learning_rate/qwen_vl_interface': 3.478778984950556e-06, 'learning_rate/action_model': 3.478778984950556e-05, 'epoch': 2.03}) +[2026-06-16 04:26:12,949][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.34656068682670593, 'train/grad_norm_pre_clip': 11.809353828430176, 'timing/data': 0.013312582857906818, 'timing/model': 1.7279256908223033, 'learning_rate/qwen_vl_interface': 3.475953575392405e-06, 'learning_rate/action_model': 3.475953575392405e-05, 'epoch': 2.03}) +[2026-06-16 04:26:18,561][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.29298487305641174, 'train/grad_norm_pre_clip': 13.905326843261719, 'timing/data': 0.012577541172504425, 'timing/model': 1.4593076594173908, 'learning_rate/qwen_vl_interface': 3.473128895077728e-06, 'learning_rate/action_model': 3.473128895077728e-05, 'epoch': 2.03}) +[2026-06-16 04:26:24,462][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.032713476568460464, 'train/grad_norm_pre_clip': 5.366835117340088, 'timing/data': 0.012957482598721981, 'timing/model': 1.459881735034287, 'learning_rate/qwen_vl_interface': 3.4703049451676453e-06, 'learning_rate/action_model': 3.4703049451676454e-05, 'epoch': 2.03}) +[2026-06-16 04:26:30,375][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.12589803338050842, 'train/grad_norm_pre_clip': 8.515259742736816, 'timing/data': 0.011282635852694511, 'timing/model': 1.4750647777691483, 'learning_rate/qwen_vl_interface': 3.4674817268229732e-06, 'learning_rate/action_model': 3.467481726822973e-05, 'epoch': 2.04}) +[2026-06-16 04:26:36,368][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.25648120045661926, 'train/grad_norm_pre_clip': 18.211055755615234, 'timing/data': 0.011937848292291164, 'timing/model': 1.4091082476079464, 'learning_rate/qwen_vl_interface': 3.464659241204232e-06, 'learning_rate/action_model': 3.4646592412042314e-05, 'epoch': 2.04}) +[2026-06-16 04:26:42,277][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.24120555818080902, 'train/grad_norm_pre_clip': 17.46312141418457, 'timing/data': 0.011858288198709488, 'timing/model': 1.7661119746044278, 'learning_rate/qwen_vl_interface': 3.4618374894716377e-06, 'learning_rate/action_model': 3.461837489471638e-05, 'epoch': 2.04}) +[2026-06-16 04:26:48,294][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.11021698266267776, 'train/grad_norm_pre_clip': 13.033624649047852, 'timing/data': 0.012481741607189178, 'timing/model': 1.7949098460376263, 'learning_rate/qwen_vl_interface': 3.459016472785107e-06, 'learning_rate/action_model': 3.459016472785107e-05, 'epoch': 2.04}) +[2026-06-16 04:26:53,964][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.08724106848239899, 'train/grad_norm_pre_clip': 8.915098190307617, 'timing/data': 0.012347261421382427, 'timing/model': 1.4542877906933427, 'learning_rate/qwen_vl_interface': 3.456196192304254e-06, 'learning_rate/action_model': 3.456196192304254e-05, 'epoch': 2.04}) +[2026-06-16 04:26:59,858][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.10727047175168991, 'train/grad_norm_pre_clip': 11.13962173461914, 'timing/data': 0.013119944371283054, 'timing/model': 1.4477526918053627, 'learning_rate/qwen_vl_interface': 3.4533766491883887e-06, 'learning_rate/action_model': 3.453376649188388e-05, 'epoch': 2.04}) +[2026-06-16 04:27:05,770][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.05824341997504234, 'train/grad_norm_pre_clip': 5.780359745025635, 'timing/data': 0.012609981000423431, 'timing/model': 1.4648608500137925, 'learning_rate/qwen_vl_interface': 3.450557844596517e-06, 'learning_rate/action_model': 3.450557844596517e-05, 'epoch': 2.04}) +[2026-06-16 04:27:11,666][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.210054412484169, 'train/grad_norm_pre_clip': 27.064117431640625, 'timing/data': 0.011814739555120468, 'timing/model': 1.4208930851891637, 'learning_rate/qwen_vl_interface': 3.447739779687345e-06, 'learning_rate/action_model': 3.4477397796873446e-05, 'epoch': 2.04}) +[2026-06-16 04:27:17,578][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.024484800174832344, 'train/grad_norm_pre_clip': 2.1355841159820557, 'timing/data': 0.012577001936733723, 'timing/model': 1.7652486599981785, 'learning_rate/qwen_vl_interface': 3.4449224556192722e-06, 'learning_rate/action_model': 3.444922455619272e-05, 'epoch': 2.04}) +[2026-06-16 04:27:23,263][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.18001429736614227, 'train/grad_norm_pre_clip': 13.566972732543945, 'timing/data': 0.012102690525352955, 'timing/model': 1.4590112827718258, 'learning_rate/qwen_vl_interface': 3.442105873550395e-06, 'learning_rate/action_model': 3.442105873550395e-05, 'epoch': 2.04}) +[2026-06-16 04:27:29,167][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.4348573386669159, 'train/grad_norm_pre_clip': 7.93912410736084, 'timing/data': 0.011959850788116455, 'timing/model': 1.4541892781853676, 'learning_rate/qwen_vl_interface': 3.4392900346385022e-06, 'learning_rate/action_model': 3.4392900346385024e-05, 'epoch': 2.04}) +[2026-06-16 04:27:35,072][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 0.06838122755289078, 'train/grad_norm_pre_clip': 9.491162300109863, 'timing/data': 0.011257308535277843, 'timing/model': 1.4784702407196164, 'learning_rate/qwen_vl_interface': 3.4364749400410796e-06, 'learning_rate/action_model': 3.43647494004108e-05, 'epoch': 2.04}) +[2026-06-16 04:27:41,067][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.0779159739613533, 'train/grad_norm_pre_clip': 11.06808090209961, 'timing/data': 0.011645079590380192, 'timing/model': 1.4441778054460883, 'learning_rate/qwen_vl_interface': 3.4336605909153066e-06, 'learning_rate/action_model': 3.433660590915306e-05, 'epoch': 2.04}) +[2026-06-16 04:27:46,876][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.05823025107383728, 'train/grad_norm_pre_clip': 6.669923305511475, 'timing/data': 0.012217261828482151, 'timing/model': 1.4097095169126987, 'learning_rate/qwen_vl_interface': 3.430846988418057e-06, 'learning_rate/action_model': 3.430846988418057e-05, 'epoch': 2.04}) +[2026-06-16 04:27:52,848][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.32840695977211, 'train/grad_norm_pre_clip': 13.66734504699707, 'timing/data': 0.012002980336546898, 'timing/model': 1.7466474641114473, 'learning_rate/qwen_vl_interface': 3.428034133705893e-06, 'learning_rate/action_model': 3.428034133705893e-05, 'epoch': 2.04}) +[2026-06-16 04:27:58,469][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.1477341204881668, 'train/grad_norm_pre_clip': 9.890802383422852, 'timing/data': 0.011919761076569557, 'timing/model': 1.4586765440180898, 'learning_rate/qwen_vl_interface': 3.4252220279350763e-06, 'learning_rate/action_model': 3.425222027935076e-05, 'epoch': 2.04}) +[2026-06-16 04:28:04,468][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.08803453296422958, 'train/grad_norm_pre_clip': 6.696345806121826, 'timing/data': 0.011999491602182388, 'timing/model': 1.4744658377021551, 'learning_rate/qwen_vl_interface': 3.4224106722615547e-06, 'learning_rate/action_model': 3.422410672261554e-05, 'epoch': 2.04}) +[2026-06-16 04:28:10,882][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.20815463364124298, 'train/grad_norm_pre_clip': 10.99349308013916, 'timing/data': 0.0383418332785368, 'timing/model': 1.4994954960420728, 'learning_rate/qwen_vl_interface': 3.419600067840973e-06, 'learning_rate/action_model': 3.419600067840973e-05, 'epoch': 2.04}) +[2026-06-16 04:28:16,860][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.06862790882587433, 'train/grad_norm_pre_clip': 4.661248207092285, 'timing/data': 0.011948051862418652, 'timing/model': 1.4309386359527707, 'learning_rate/qwen_vl_interface': 3.4167902158286627e-06, 'learning_rate/action_model': 3.4167902158286626e-05, 'epoch': 2.04}) +[2026-06-16 04:28:22,777][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 0.24405120313167572, 'train/grad_norm_pre_clip': 10.216898918151855, 'timing/data': 0.012139172293245792, 'timing/model': 1.4103657426312566, 'learning_rate/qwen_vl_interface': 3.4139811173796468e-06, 'learning_rate/action_model': 3.413981117379647e-05, 'epoch': 2.04}) +[2026-06-16 04:28:28,807][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.12017329782247543, 'train/grad_norm_pre_clip': 11.002684593200684, 'timing/data': 0.011562449857592583, 'timing/model': 1.8137926999479532, 'learning_rate/qwen_vl_interface': 3.411172773648642e-06, 'learning_rate/action_model': 3.411172773648642e-05, 'epoch': 2.04}) +[2026-06-16 04:28:34,466][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.20618857443332672, 'train/grad_norm_pre_clip': 13.426553726196289, 'timing/data': 0.012247863225638866, 'timing/model': 1.4576552854850888, 'learning_rate/qwen_vl_interface': 3.408365185790052e-06, 'learning_rate/action_model': 3.4083651857900515e-05, 'epoch': 2.04}) +[2026-06-16 04:28:40,362][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.08125993609428406, 'train/grad_norm_pre_clip': 8.445760726928711, 'timing/data': 0.012234733439981937, 'timing/model': 1.449520074762404, 'learning_rate/qwen_vl_interface': 3.40555835495797e-06, 'learning_rate/action_model': 3.40555835495797e-05, 'epoch': 2.05}) +[2026-06-16 04:28:46,272][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.06259343028068542, 'train/grad_norm_pre_clip': 5.477115154266357, 'timing/data': 0.011179869994521141, 'timing/model': 1.4619967360049486, 'learning_rate/qwen_vl_interface': 3.4027522823061763e-06, 'learning_rate/action_model': 3.4027522823061765e-05, 'epoch': 2.05}) +[2026-06-16 04:28:52,165][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 0.2713172137737274, 'train/grad_norm_pre_clip': 10.667021751403809, 'timing/data': 0.011726551689207554, 'timing/model': 1.46496257558465, 'learning_rate/qwen_vl_interface': 3.399946968988143e-06, 'learning_rate/action_model': 3.399946968988143e-05, 'epoch': 2.05}) +[2026-06-16 04:28:58,082][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.04387502372264862, 'train/grad_norm_pre_clip': 4.228784084320068, 'timing/data': 0.012162462808191776, 'timing/model': 1.7813434842973948, 'learning_rate/qwen_vl_interface': 3.397142416157028e-06, 'learning_rate/action_model': 3.397142416157028e-05, 'epoch': 2.05}) +[2026-06-16 04:29:03,768][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.017607776448130608, 'train/grad_norm_pre_clip': 1.432146668434143, 'timing/data': 0.011648081243038177, 'timing/model': 1.469807107001543, 'learning_rate/qwen_vl_interface': 3.3943386249656766e-06, 'learning_rate/action_model': 3.394338624965677e-05, 'epoch': 2.05}) +[2026-06-16 04:29:09,666][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.15953101217746735, 'train/grad_norm_pre_clip': 7.501698970794678, 'timing/data': 0.011695162393152714, 'timing/model': 1.4559876034036279, 'learning_rate/qwen_vl_interface': 3.391535596566624e-06, 'learning_rate/action_model': 3.391535596566624e-05, 'epoch': 2.05}) +[2026-06-16 04:29:15,471][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.1964658498764038, 'train/grad_norm_pre_clip': 14.193596839904785, 'timing/data': 0.012804755941033363, 'timing/model': 1.4690941274166107, 'learning_rate/qwen_vl_interface': 3.388733332112086e-06, 'learning_rate/action_model': 3.3887333321120856e-05, 'epoch': 2.05}) +[2026-06-16 04:29:21,267][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.23733052611351013, 'train/grad_norm_pre_clip': 9.493462562561035, 'timing/data': 0.011188450269401073, 'timing/model': 1.4761779699474573, 'learning_rate/qwen_vl_interface': 3.3859318327539693e-06, 'learning_rate/action_model': 3.385931832753969e-05, 'epoch': 2.05}) +[2026-06-16 04:29:27,063][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.018452104181051254, 'train/grad_norm_pre_clip': 1.1699779033660889, 'timing/data': 0.01103086955845356, 'timing/model': 1.3983782036229968, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 2.05}) +[2026-06-16 04:29:32,968][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.1280209869146347, 'train/grad_norm_pre_clip': 5.331633567810059, 'timing/data': 0.011138331145048141, 'timing/model': 1.7601958084851503, 'learning_rate/qwen_vl_interface': 3.3803311339330447e-06, 'learning_rate/action_model': 3.3803311339330444e-05, 'epoch': 2.05}) +[2026-06-16 04:29:38,563][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.12314099073410034, 'train/grad_norm_pre_clip': 9.038140296936035, 'timing/data': 0.011050749570131302, 'timing/model': 1.4561485052108765, 'learning_rate/qwen_vl_interface': 3.3775319367724746e-06, 'learning_rate/action_model': 3.3775319367724745e-05, 'epoch': 2.05}) +[2026-06-16 04:29:44,467][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.21277360618114471, 'train/grad_norm_pre_clip': 12.445059776306152, 'timing/data': 0.01133454218506813, 'timing/model': 1.4751805309206247, 'learning_rate/qwen_vl_interface': 3.374733509312795e-06, 'learning_rate/action_model': 3.374733509312795e-05, 'epoch': 2.05}) +[2026-06-16 04:29:50,367][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.20863570272922516, 'train/grad_norm_pre_clip': 9.416597366333008, 'timing/data': 0.014713783748447895, 'timing/model': 1.4529476668685675, 'learning_rate/qwen_vl_interface': 3.3719358527043345e-06, 'learning_rate/action_model': 3.371935852704334e-05, 'epoch': 2.05}) +[2026-06-16 04:29:56,367][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.023548567667603493, 'train/grad_norm_pre_clip': 3.7347168922424316, 'timing/data': 0.012432846240699291, 'timing/model': 1.4816797077655792, 'learning_rate/qwen_vl_interface': 3.3691389680971047e-06, 'learning_rate/action_model': 3.3691389680971044e-05, 'epoch': 2.05}) +[2026-06-16 04:30:02,158][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.1951620727777481, 'train/grad_norm_pre_clip': 11.197608947753906, 'timing/data': 0.011826532892882824, 'timing/model': 1.3956227554008365, 'learning_rate/qwen_vl_interface': 3.3663428566407985e-06, 'learning_rate/action_model': 3.366342856640798e-05, 'epoch': 2.05}) +[2026-06-16 04:30:08,060][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.08112789690494537, 'train/grad_norm_pre_clip': 5.314552307128906, 'timing/data': 0.011522742919623852, 'timing/model': 1.755933364853263, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 2.05}) +[2026-06-16 04:30:13,657][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.12515470385551453, 'train/grad_norm_pre_clip': 10.339408874511719, 'timing/data': 0.011086052283644676, 'timing/model': 1.4548934632912278, 'learning_rate/qwen_vl_interface': 3.3607529577781433e-06, 'learning_rate/action_model': 3.360752957778143e-05, 'epoch': 2.05}) +[2026-06-16 04:30:19,565][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.09068137407302856, 'train/grad_norm_pre_clip': 6.479520797729492, 'timing/data': 0.012177646160125732, 'timing/model': 1.4512649634853005, 'learning_rate/qwen_vl_interface': 3.3579591726695906e-06, 'learning_rate/action_model': 3.35795917266959e-05, 'epoch': 2.05}) +[2026-06-16 04:30:25,463][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.015287702903151512, 'train/grad_norm_pre_clip': 1.1845027208328247, 'timing/data': 0.010735671035945415, 'timing/model': 1.4693954763934016, 'learning_rate/qwen_vl_interface': 3.3551661653075522e-06, 'learning_rate/action_model': 3.355166165307552e-05, 'epoch': 2.05}) +[2026-06-16 04:30:31,259][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.1567763090133667, 'train/grad_norm_pre_clip': 6.8059492111206055, 'timing/data': 0.012643298134207726, 'timing/model': 1.4169167438521981, 'learning_rate/qwen_vl_interface': 3.35237393684013e-06, 'learning_rate/action_model': 3.35237393684013e-05, 'epoch': 2.05}) +[2026-06-16 04:30:36,975][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.06768105179071426, 'train/grad_norm_pre_clip': 7.30935525894165, 'timing/data': 0.012103155255317688, 'timing/model': 1.391867103986442, 'learning_rate/qwen_vl_interface': 3.349582488415103e-06, 'learning_rate/action_model': 3.3495824884151026e-05, 'epoch': 2.05}) +[2026-06-16 04:30:42,979][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.13846455514431, 'train/grad_norm_pre_clip': 16.765541076660156, 'timing/data': 0.01134107168763876, 'timing/model': 1.7573347762227058, 'learning_rate/qwen_vl_interface': 3.3467918211799294e-06, 'learning_rate/action_model': 3.3467918211799295e-05, 'epoch': 2.05}) +[2026-06-16 04:30:48,658][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.06313999742269516, 'train/grad_norm_pre_clip': 6.520076274871826, 'timing/data': 0.01226231548935175, 'timing/model': 1.4717016536742449, 'learning_rate/qwen_vl_interface': 3.3440019362817476e-06, 'learning_rate/action_model': 3.344001936281747e-05, 'epoch': 2.06}) +[2026-06-16 04:30:54,575][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.15504828095436096, 'train/grad_norm_pre_clip': 8.779322624206543, 'timing/data': 0.011792133562266827, 'timing/model': 1.4610980469733477, 'learning_rate/qwen_vl_interface': 3.3412128348673745e-06, 'learning_rate/action_model': 3.341212834867374e-05, 'epoch': 2.06}) +[2026-06-16 04:31:00,568][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.02319314144551754, 'train/grad_norm_pre_clip': 2.1066126823425293, 'timing/data': 0.01056934054940939, 'timing/model': 1.4749807734042406, 'learning_rate/qwen_vl_interface': 3.338424518083303e-06, 'learning_rate/action_model': 3.338424518083303e-05, 'epoch': 2.06}) +[2026-06-16 04:31:06,350][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.06696005165576935, 'train/grad_norm_pre_clip': 5.058703422546387, 'timing/data': 0.011702364310622215, 'timing/model': 1.4515329077839851, 'learning_rate/qwen_vl_interface': 3.3356369870757056e-06, 'learning_rate/action_model': 3.335636987075705e-05, 'epoch': 2.06}) +[2026-06-16 04:31:12,250][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.23621538281440735, 'train/grad_norm_pre_clip': 10.304935455322266, 'timing/data': 0.010473799891769886, 'timing/model': 1.4536530477926135, 'learning_rate/qwen_vl_interface': 3.3328502429904327e-06, 'learning_rate/action_model': 3.3328502429904325e-05, 'epoch': 2.06}) +[2026-06-16 04:31:18,132][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.03954251483082771, 'train/grad_norm_pre_clip': 4.009162425994873, 'timing/data': 0.009996918961405754, 'timing/model': 1.750656208023429, 'learning_rate/qwen_vl_interface': 3.3300642869730066e-06, 'learning_rate/action_model': 3.330064286973006e-05, 'epoch': 2.06}) +[2026-06-16 04:41:55,872][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.2499684989452362, 'train/grad_norm_pre_clip': 10.768783569335938, 'eval/loss': 0.05874824523925781, 'eval/flap_precision': 0.8555555555555555, 'eval/flap_recall': 0.84, 'eval/flap_f1': 0.8477064220183486, 'eval/flap_rate_model': 0.075, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.8477064220183484, 'timing/data': 0.010839330963790417, 'timing/model': 1.4484925549477339, 'learning_rate/qwen_vl_interface': 3.3272791201686294e-06, 'learning_rate/action_model': 3.327279120168629e-05, 'epoch': 2.06}) +[2026-06-16 04:41:55,873][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4500_state +[2026-06-16 04:41:55,879][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 04:42:03,599][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4500_state/model.safetensors +[2026-06-16 04:42:15,102][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4500_state/optimizer.bin +[2026-06-16 04:42:15,103][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4500_state/sampler.bin +[2026-06-16 04:42:15,103][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4500_state/sampler_1.bin +[2026-06-16 04:42:15,104][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4500_state/random_states_0.pkl +[2026-06-16 04:42:15,270][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-06-16 04:42:15,271][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 04:42:15,286][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 04:42:21,072][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.09994691610336304, 'train/grad_norm_pre_clip': 9.551427841186523, 'timing/data': 0.011104589328169823, 'timing/model': 1.7864492572844028, 'learning_rate/qwen_vl_interface': 3.324494743722179e-06, 'learning_rate/action_model': 3.3244947437221785e-05, 'epoch': 2.06}) +[2026-06-16 04:42:26,547][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.08036196976900101, 'train/grad_norm_pre_clip': 6.322901725769043, 'timing/data': 0.011252189055085182, 'timing/model': 1.3717490276321769, 'learning_rate/qwen_vl_interface': 3.321711158778205e-06, 'learning_rate/action_model': 3.321711158778205e-05, 'epoch': 2.06}) +[2026-06-16 04:42:32,344][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.10159491747617722, 'train/grad_norm_pre_clip': 6.916069507598877, 'timing/data': 0.011816362850368023, 'timing/model': 1.4317331928759813, 'learning_rate/qwen_vl_interface': 3.3189283664809393e-06, 'learning_rate/action_model': 3.318928366480939e-05, 'epoch': 2.06}) +[2026-06-16 04:42:38,057][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.11709040403366089, 'train/grad_norm_pre_clip': 8.07675552368164, 'timing/data': 0.014002252370119095, 'timing/model': 1.4777408195659518, 'learning_rate/qwen_vl_interface': 3.316146367974276e-06, 'learning_rate/action_model': 3.316146367974276e-05, 'epoch': 2.06}) +[2026-06-16 04:42:43,946][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.025440849363803864, 'train/grad_norm_pre_clip': 2.4099175930023193, 'timing/data': 0.011776802130043507, 'timing/model': 1.4327146634459496, 'learning_rate/qwen_vl_interface': 3.3133651644017945e-06, 'learning_rate/action_model': 3.3133651644017944e-05, 'epoch': 2.06}) +[2026-06-16 04:42:49,738][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.081754669547081, 'train/grad_norm_pre_clip': 5.149967670440674, 'timing/data': 0.011852223426103592, 'timing/model': 1.4559815395623446, 'learning_rate/qwen_vl_interface': 3.31058475690674e-06, 'learning_rate/action_model': 3.31058475690674e-05, 'epoch': 2.06}) +[2026-06-16 04:42:55,568][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.012385052628815174, 'train/grad_norm_pre_clip': 1.1321871280670166, 'timing/data': 0.012022413313388824, 'timing/model': 1.691829415038228, 'learning_rate/qwen_vl_interface': 3.3078051466320337e-06, 'learning_rate/action_model': 3.307805146632034e-05, 'epoch': 2.06}) +[2026-06-16 04:43:01,055][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.11082357913255692, 'train/grad_norm_pre_clip': 9.885876655578613, 'timing/data': 0.010755697265267372, 'timing/model': 1.4431542595848441, 'learning_rate/qwen_vl_interface': 3.3050263347202712e-06, 'learning_rate/action_model': 3.305026334720271e-05, 'epoch': 2.06}) +[2026-06-16 04:43:06,844][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.20972156524658203, 'train/grad_norm_pre_clip': 11.611581802368164, 'timing/data': 0.010840198025107384, 'timing/model': 1.4543986730277538, 'learning_rate/qwen_vl_interface': 3.302248322313712e-06, 'learning_rate/action_model': 3.302248322313712e-05, 'epoch': 2.06}) +[2026-06-16 04:43:12,544][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.054777104407548904, 'train/grad_norm_pre_clip': 6.740478038787842, 'timing/data': 0.011061369441449642, 'timing/model': 1.4568747971206903, 'learning_rate/qwen_vl_interface': 3.299471110554295e-06, 'learning_rate/action_model': 3.299471110554295e-05, 'epoch': 2.06}) +[2026-06-16 04:43:18,351][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 0.2065456062555313, 'train/grad_norm_pre_clip': 16.362701416015625, 'timing/data': 0.017147685401141644, 'timing/model': 1.4604346062988043, 'learning_rate/qwen_vl_interface': 3.296694700583626e-06, 'learning_rate/action_model': 3.296694700583626e-05, 'epoch': 2.06}) +[2026-06-16 04:43:24,259][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.09553935378789902, 'train/grad_norm_pre_clip': 6.513991832733154, 'timing/data': 0.013127937912940979, 'timing/model': 1.4677730416879058, 'learning_rate/qwen_vl_interface': 3.2939190935429833e-06, 'learning_rate/action_model': 3.293919093542983e-05, 'epoch': 2.06}) +[2026-06-16 04:43:30,178][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.07497590035200119, 'train/grad_norm_pre_clip': 4.406766414642334, 'timing/data': 0.012188663706183434, 'timing/model': 1.787199198268354, 'learning_rate/qwen_vl_interface': 3.2911442905733126e-06, 'learning_rate/action_model': 3.291144290573312e-05, 'epoch': 2.06}) +[2026-06-16 04:43:36,176][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.04566969349980354, 'train/grad_norm_pre_clip': 6.1872429847717285, 'timing/data': 0.012653466314077377, 'timing/model': 1.7710042595863342, 'learning_rate/qwen_vl_interface': 3.2883702928152323e-06, 'learning_rate/action_model': 3.288370292815232e-05, 'epoch': 2.06}) +[2026-06-16 04:43:41,868][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.07094210386276245, 'train/grad_norm_pre_clip': 11.980927467346191, 'timing/data': 0.011355539783835411, 'timing/model': 1.460133608430624, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 2.06}) +[2026-06-16 04:43:47,755][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.14496372640132904, 'train/grad_norm_pre_clip': 6.299679756164551, 'timing/data': 0.013855340890586376, 'timing/model': 1.4540815437212586, 'learning_rate/qwen_vl_interface': 3.2828247174946528e-06, 'learning_rate/action_model': 3.2828247174946524e-05, 'epoch': 2.07}) +[2026-06-16 04:43:53,668][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.03530118241906166, 'train/grad_norm_pre_clip': 5.358619213104248, 'timing/data': 0.011755033396184444, 'timing/model': 1.4777755308896303, 'learning_rate/qwen_vl_interface': 3.280053142211729e-06, 'learning_rate/action_model': 3.280053142211729e-05, 'epoch': 2.07}) +[2026-06-16 04:43:59,561][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.011928413063287735, 'train/grad_norm_pre_clip': 1.1127214431762695, 'timing/data': 0.011708151549100876, 'timing/model': 1.4046926507726312, 'learning_rate/qwen_vl_interface': 3.2772823766995475e-06, 'learning_rate/action_model': 3.277282376699548e-05, 'epoch': 2.07}) +[2026-06-16 04:44:05,468][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.18756426870822906, 'train/grad_norm_pre_clip': 13.930585861206055, 'timing/data': 0.011302360333502293, 'timing/model': 1.7722072750329971, 'learning_rate/qwen_vl_interface': 3.2745124220970646e-06, 'learning_rate/action_model': 3.274512422097065e-05, 'epoch': 2.07}) +[2026-06-16 04:44:11,074][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.1743442565202713, 'train/grad_norm_pre_clip': 11.579079627990723, 'timing/data': 0.011412961408495903, 'timing/model': 1.4814370665699244, 'learning_rate/qwen_vl_interface': 3.2717432795429057e-06, 'learning_rate/action_model': 3.271743279542905e-05, 'epoch': 2.07}) +[2026-06-16 04:44:16,960][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 0.1397816240787506, 'train/grad_norm_pre_clip': 37.92048263549805, 'timing/data': 0.012269855476915836, 'timing/model': 1.4667096938937902, 'learning_rate/qwen_vl_interface': 3.2689749501753578e-06, 'learning_rate/action_model': 3.2689749501753574e-05, 'epoch': 2.07}) +[2026-06-16 04:44:22,867][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.040839262306690216, 'train/grad_norm_pre_clip': 4.218101978302002, 'timing/data': 0.011001468636095524, 'timing/model': 1.4814381934702396, 'learning_rate/qwen_vl_interface': 3.2662074351323774e-06, 'learning_rate/action_model': 3.266207435132378e-05, 'epoch': 2.07}) +[2026-06-16 04:44:28,764][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.26376670598983765, 'train/grad_norm_pre_clip': 12.180767059326172, 'timing/data': 0.0121301943436265, 'timing/model': 1.461636597290635, 'learning_rate/qwen_vl_interface': 3.2634407355515875e-06, 'learning_rate/action_model': 3.2634407355515875e-05, 'epoch': 2.07}) +[2026-06-16 04:44:34,552][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.07092879712581635, 'train/grad_norm_pre_clip': 6.7276387214660645, 'timing/data': 0.011337639763951302, 'timing/model': 1.4169487319886684, 'learning_rate/qwen_vl_interface': 3.2606748525702715e-06, 'learning_rate/action_model': 3.260674852570271e-05, 'epoch': 2.07}) +[2026-06-16 04:44:40,408][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.011333147063851357, 'train/grad_norm_pre_clip': 1.453311562538147, 'timing/data': 0.011280129663646221, 'timing/model': 1.7125528492033482, 'learning_rate/qwen_vl_interface': 3.2579097873253803e-06, 'learning_rate/action_model': 3.25790978732538e-05, 'epoch': 2.07}) +[2026-06-16 04:44:45,959][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 0.08394709974527359, 'train/grad_norm_pre_clip': 9.974503517150879, 'timing/data': 0.011310091242194176, 'timing/model': 1.4415191756561399, 'learning_rate/qwen_vl_interface': 3.2551455409535274e-06, 'learning_rate/action_model': 3.255145540953528e-05, 'epoch': 2.07}) +[2026-06-16 04:44:51,861][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.16577982902526855, 'train/grad_norm_pre_clip': 9.768028259277344, 'timing/data': 0.011009618639945984, 'timing/model': 1.4593182979151607, 'learning_rate/qwen_vl_interface': 3.2523821145909907e-06, 'learning_rate/action_model': 3.252382114590991e-05, 'epoch': 2.07}) +[2026-06-16 04:44:57,658][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.26368075609207153, 'train/grad_norm_pre_clip': 9.503461837768555, 'timing/data': 0.010955429635941982, 'timing/model': 1.4530634731054306, 'learning_rate/qwen_vl_interface': 3.249619509373712e-06, 'learning_rate/action_model': 3.249619509373712e-05, 'epoch': 2.07}) +[2026-06-16 04:45:03,474][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.04652130976319313, 'train/grad_norm_pre_clip': 3.5864198207855225, 'timing/data': 0.011180909350514412, 'timing/model': 1.3965517431497574, 'learning_rate/qwen_vl_interface': 3.2468577264372905e-06, 'learning_rate/action_model': 3.24685772643729e-05, 'epoch': 2.07}) +[2026-06-16 04:45:09,465][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.13652046024799347, 'train/grad_norm_pre_clip': 7.583969593048096, 'timing/data': 0.01194776315242052, 'timing/model': 1.4712880793958902, 'learning_rate/qwen_vl_interface': 3.244096766916992e-06, 'learning_rate/action_model': 3.244096766916992e-05, 'epoch': 2.07}) +[2026-06-16 04:45:15,323][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.014279209077358246, 'train/grad_norm_pre_clip': 8.881661415100098, 'timing/data': 0.010773248039186, 'timing/model': 1.7245700620114803, 'learning_rate/qwen_vl_interface': 3.241336631947743e-06, 'learning_rate/action_model': 3.2413366319477427e-05, 'epoch': 2.07}) +[2026-06-16 04:45:20,965][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.01868305169045925, 'train/grad_norm_pre_clip': 2.4194257259368896, 'timing/data': 0.011428870260715485, 'timing/model': 1.45133051648736, 'learning_rate/qwen_vl_interface': 3.238577322664131e-06, 'learning_rate/action_model': 3.238577322664131e-05, 'epoch': 2.07}) +[2026-06-16 04:45:26,867][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.15153828263282776, 'train/grad_norm_pre_clip': 7.901805877685547, 'timing/data': 0.011208469979465008, 'timing/model': 1.4599399464204907, 'learning_rate/qwen_vl_interface': 3.2358188402004026e-06, 'learning_rate/action_model': 3.2358188402004024e-05, 'epoch': 2.07}) +[2026-06-16 04:45:32,763][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.3727554678916931, 'train/grad_norm_pre_clip': 16.946142196655273, 'timing/data': 0.011234560050070286, 'timing/model': 1.4581766510382295, 'learning_rate/qwen_vl_interface': 3.2330611856904675e-06, 'learning_rate/action_model': 3.233061185690468e-05, 'epoch': 2.07}) +[2026-06-16 04:45:38,663][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.1655764877796173, 'train/grad_norm_pre_clip': 11.648703575134277, 'timing/data': 0.011360100470483303, 'timing/model': 1.4756453121080995, 'learning_rate/qwen_vl_interface': 3.2303043602678928e-06, 'learning_rate/action_model': 3.230304360267893e-05, 'epoch': 2.07}) +[2026-06-16 04:45:44,560][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.050803858786821365, 'train/grad_norm_pre_clip': 6.234872341156006, 'timing/data': 0.011252550408244133, 'timing/model': 1.7654503593221307, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 2.07}) +[2026-06-16 04:45:50,172][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.036250125616788864, 'train/grad_norm_pre_clip': 2.7152798175811768, 'timing/data': 0.013249209150671959, 'timing/model': 1.4796134838834405, 'learning_rate/qwen_vl_interface': 3.2247932012173887e-06, 'learning_rate/action_model': 3.2247932012173884e-05, 'epoch': 2.07}) +[2026-06-16 04:45:56,160][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.06485284119844437, 'train/grad_norm_pre_clip': 6.79558801651001, 'timing/data': 0.01052520889788866, 'timing/model': 1.4453535731881857, 'learning_rate/qwen_vl_interface': 3.2220388698548897e-06, 'learning_rate/action_model': 3.22203886985489e-05, 'epoch': 2.08}) +[2026-06-16 04:46:01,968][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.0666225329041481, 'train/grad_norm_pre_clip': 5.514195442199707, 'timing/data': 0.011672953143715858, 'timing/model': 1.4590251464396715, 'learning_rate/qwen_vl_interface': 3.21928537211061e-06, 'learning_rate/action_model': 3.21928537211061e-05, 'epoch': 2.08}) +[2026-06-16 04:46:07,874][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.07391957193613052, 'train/grad_norm_pre_clip': 14.538723945617676, 'timing/data': 0.011773022823035717, 'timing/model': 1.4707223810255527, 'learning_rate/qwen_vl_interface': 3.216532709116405e-06, 'learning_rate/action_model': 3.216532709116405e-05, 'epoch': 2.08}) +[2026-06-16 04:46:13,768][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.12576910853385925, 'train/grad_norm_pre_clip': 6.467099666595459, 'timing/data': 0.0138749023899436, 'timing/model': 1.4107636660337448, 'learning_rate/qwen_vl_interface': 3.2137808820037925e-06, 'learning_rate/action_model': 3.2137808820037924e-05, 'epoch': 2.08}) +[2026-06-16 04:46:19,669][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.13547103106975555, 'train/grad_norm_pre_clip': 8.157601356506348, 'timing/data': 0.011890084482729435, 'timing/model': 1.7701183669269085, 'learning_rate/qwen_vl_interface': 3.211029891903945e-06, 'learning_rate/action_model': 3.211029891903945e-05, 'epoch': 2.08}) +[2026-06-16 04:46:25,256][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.10022905468940735, 'train/grad_norm_pre_clip': 10.46585750579834, 'timing/data': 0.011376750655472279, 'timing/model': 1.4487063083797693, 'learning_rate/qwen_vl_interface': 3.2082797399476894e-06, 'learning_rate/action_model': 3.2082797399476895e-05, 'epoch': 2.08}) +[2026-06-16 04:46:31,170][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.11438137292861938, 'train/grad_norm_pre_clip': 11.973981857299805, 'timing/data': 0.011211879551410675, 'timing/model': 1.465620968490839, 'learning_rate/qwen_vl_interface': 3.2055304272655114e-06, 'learning_rate/action_model': 3.205530427265511e-05, 'epoch': 2.08}) +[2026-06-16 04:46:37,052][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.06821189820766449, 'train/grad_norm_pre_clip': 3.936636447906494, 'timing/data': 0.010876927524805069, 'timing/model': 1.4442226607352495, 'learning_rate/qwen_vl_interface': 3.2027819549875477e-06, 'learning_rate/action_model': 3.202781954987548e-05, 'epoch': 2.08}) +[2026-06-16 04:46:42,963][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 0.17627626657485962, 'train/grad_norm_pre_clip': 10.369780540466309, 'timing/data': 0.011338610202074051, 'timing/model': 1.4613082548603415, 'learning_rate/qwen_vl_interface': 3.2000343242435926e-06, 'learning_rate/action_model': 3.2000343242435924e-05, 'epoch': 2.08}) +[2026-06-16 04:46:48,765][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.05529461055994034, 'train/grad_norm_pre_clip': 9.10529613494873, 'timing/data': 0.012436117976903915, 'timing/model': 1.4014879371970892, 'learning_rate/qwen_vl_interface': 3.1972875361630923e-06, 'learning_rate/action_model': 3.1972875361630923e-05, 'epoch': 2.08}) +[2026-06-16 04:46:54,657][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.26066771149635315, 'train/grad_norm_pre_clip': 12.633676528930664, 'timing/data': 0.011386619880795479, 'timing/model': 1.7475093519315124, 'learning_rate/qwen_vl_interface': 3.1945415918751484e-06, 'learning_rate/action_model': 3.194541591875148e-05, 'epoch': 2.08}) +[2026-06-16 04:47:00,263][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.01461823470890522, 'train/grad_norm_pre_clip': 0.7322066426277161, 'timing/data': 0.011438170447945595, 'timing/model': 1.4480750067159534, 'learning_rate/qwen_vl_interface': 3.191796492508516e-06, 'learning_rate/action_model': 3.191796492508516e-05, 'epoch': 2.08}) +[2026-06-16 04:47:06,168][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.109452985227108, 'train/grad_norm_pre_clip': 5.976955413818359, 'timing/data': 0.0112447589635849, 'timing/model': 1.4642561944201589, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 2.08}) +[2026-06-16 04:47:12,067][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.04582763835787773, 'train/grad_norm_pre_clip': 5.066093921661377, 'timing/data': 0.012336605228483677, 'timing/model': 1.4676762642338872, 'learning_rate/qwen_vl_interface': 3.1863088330524577e-06, 'learning_rate/action_model': 3.186308833052458e-05, 'epoch': 2.08}) +[2026-06-16 04:47:17,963][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 0.10941894352436066, 'train/grad_norm_pre_clip': 7.125842571258545, 'timing/data': 0.012176074087619781, 'timing/model': 1.4579459009692073, 'learning_rate/qwen_vl_interface': 3.183566275218803e-06, 'learning_rate/action_model': 3.1835662752188026e-05, 'epoch': 2.08}) +[2026-06-16 04:47:23,768][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.11028416454792023, 'train/grad_norm_pre_clip': 8.336173057556152, 'timing/data': 0.012521046213805676, 'timing/model': 1.4604287156835198, 'learning_rate/qwen_vl_interface': 3.180824566817996e-06, 'learning_rate/action_model': 3.1808245668179956e-05, 'epoch': 2.08}) +[2026-06-16 04:47:29,658][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.03054490126669407, 'train/grad_norm_pre_clip': 4.425628662109375, 'timing/data': 0.012152322567999363, 'timing/model': 1.7481257496401668, 'learning_rate/qwen_vl_interface': 3.1780837089770524e-06, 'learning_rate/action_model': 3.1780837089770524e-05, 'epoch': 2.08}) +[2026-06-16 04:47:35,271][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 0.18053674697875977, 'train/grad_norm_pre_clip': 9.69632339477539, 'timing/data': 0.011709460988640785, 'timing/model': 1.464626107364893, 'learning_rate/qwen_vl_interface': 3.17534370282263e-06, 'learning_rate/action_model': 3.1753437028226303e-05, 'epoch': 2.08}) +[2026-06-16 04:47:41,268][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.14096888899803162, 'train/grad_norm_pre_clip': 6.235609531402588, 'timing/data': 0.013355011120438576, 'timing/model': 1.4654337605461478, 'learning_rate/qwen_vl_interface': 3.172604549481046e-06, 'learning_rate/action_model': 3.1726045494810455e-05, 'epoch': 2.08}) +[2026-06-16 04:47:47,061][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.05825260654091835, 'train/grad_norm_pre_clip': 6.301063060760498, 'timing/data': 0.011410467326641083, 'timing/model': 1.438660565763712, 'learning_rate/qwen_vl_interface': 3.169866250078261e-06, 'learning_rate/action_model': 3.169866250078261e-05, 'epoch': 2.08}) +[2026-06-16 04:47:52,764][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.27677857875823975, 'train/grad_norm_pre_clip': 12.903532981872559, 'timing/data': 0.01147443801164627, 'timing/model': 1.4504242353141308, 'learning_rate/qwen_vl_interface': 3.167128805739887e-06, 'learning_rate/action_model': 3.1671288057398866e-05, 'epoch': 2.08}) +[2026-06-16 04:47:58,669][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.06498202681541443, 'train/grad_norm_pre_clip': 9.170660018920898, 'timing/data': 0.011541967280209064, 'timing/model': 1.4814044125378132, 'learning_rate/qwen_vl_interface': 3.1643922175911847e-06, 'learning_rate/action_model': 3.164392217591185e-05, 'epoch': 2.08}) +[2026-06-16 04:48:04,560][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.02472883276641369, 'train/grad_norm_pre_clip': 5.813785076141357, 'timing/data': 0.010883362963795662, 'timing/model': 1.7678094739094377, 'learning_rate/qwen_vl_interface': 3.1616564867570603e-06, 'learning_rate/action_model': 3.16165648675706e-05, 'epoch': 2.09}) +[2026-06-16 04:48:10,169][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.015121804550290108, 'train/grad_norm_pre_clip': 1.1451369524002075, 'timing/data': 0.012317152693867683, 'timing/model': 1.4793085269629955, 'learning_rate/qwen_vl_interface': 3.1589216143620733e-06, 'learning_rate/action_model': 3.1589216143620734e-05, 'epoch': 2.09}) +[2026-06-16 04:48:16,064][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.04833533987402916, 'train/grad_norm_pre_clip': 6.721704483032227, 'timing/data': 0.011879128403961658, 'timing/model': 1.462434853427112, 'learning_rate/qwen_vl_interface': 3.1561876015304215e-06, 'learning_rate/action_model': 3.156187601530421e-05, 'epoch': 2.09}) +[2026-06-16 04:48:21,865][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.3422177731990814, 'train/grad_norm_pre_clip': 17.05319595336914, 'timing/data': 0.01326058804988861, 'timing/model': 1.4393673986196518, 'learning_rate/qwen_vl_interface': 3.1534544493859563e-06, 'learning_rate/action_model': 3.1534544493859566e-05, 'epoch': 2.09}) +[2026-06-16 04:48:27,658][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.13056997954845428, 'train/grad_norm_pre_clip': 9.911328315734863, 'timing/data': 0.011616786941885948, 'timing/model': 1.4494831068441272, 'learning_rate/qwen_vl_interface': 3.150722159052174e-06, 'learning_rate/action_model': 3.150722159052174e-05, 'epoch': 2.09}) +[2026-06-16 04:48:33,567][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.08873569220304489, 'train/grad_norm_pre_clip': 9.351767539978027, 'timing/data': 0.010894793085753918, 'timing/model': 1.4788639722391963, 'learning_rate/qwen_vl_interface': 3.147990731652215e-06, 'learning_rate/action_model': 3.147990731652215e-05, 'epoch': 2.09}) +[2026-06-16 04:48:39,462][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.06439066678285599, 'train/grad_norm_pre_clip': 6.840184688568115, 'timing/data': 0.01159460749477148, 'timing/model': 1.7567105246707797, 'learning_rate/qwen_vl_interface': 3.1452601683088673e-06, 'learning_rate/action_model': 3.145260168308867e-05, 'epoch': 2.09}) +[2026-06-16 04:48:45,068][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.0587892085313797, 'train/grad_norm_pre_clip': 2.9283034801483154, 'timing/data': 0.012915374711155891, 'timing/model': 1.4679537136107683, 'learning_rate/qwen_vl_interface': 3.1425304701445608e-06, 'learning_rate/action_model': 3.1425304701445605e-05, 'epoch': 2.09}) +[2026-06-16 04:48:50,968][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.009254446253180504, 'train/grad_norm_pre_clip': 1.0535043478012085, 'timing/data': 0.012178581207990646, 'timing/model': 1.4573677545413375, 'learning_rate/qwen_vl_interface': 3.139801638281372e-06, 'learning_rate/action_model': 3.139801638281372e-05, 'epoch': 2.09}) +[2026-06-16 04:48:56,857][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 0.0620681457221508, 'train/grad_norm_pre_clip': 5.791806221008301, 'timing/data': 0.011540585197508335, 'timing/model': 1.4491321528330445, 'learning_rate/qwen_vl_interface': 3.13707367384102e-06, 'learning_rate/action_model': 3.13707367384102e-05, 'epoch': 2.09}) +[2026-06-16 04:49:02,670][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.20934247970581055, 'train/grad_norm_pre_clip': 13.359243392944336, 'timing/data': 0.011737857013940811, 'timing/model': 1.4594365628436208, 'learning_rate/qwen_vl_interface': 3.134346577944869e-06, 'learning_rate/action_model': 3.134346577944869e-05, 'epoch': 2.09}) +[2026-06-16 04:49:08,562][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.1098722517490387, 'train/grad_norm_pre_clip': 11.535176277160645, 'timing/data': 0.011384803801774979, 'timing/model': 1.470987230539322, 'learning_rate/qwen_vl_interface': 3.1316203517139243e-06, 'learning_rate/action_model': 3.131620351713924e-05, 'epoch': 2.09}) +[2026-06-16 04:49:14,462][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.2398611456155777, 'train/grad_norm_pre_clip': 13.32059097290039, 'timing/data': 0.011746346019208431, 'timing/model': 1.756043927744031, 'learning_rate/qwen_vl_interface': 3.128894996268837e-06, 'learning_rate/action_model': 3.128894996268837e-05, 'epoch': 2.09}) +[2026-06-16 04:49:20,068][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.0029430843424052, 'train/grad_norm_pre_clip': 0.3977108895778656, 'timing/data': 0.01178025733679533, 'timing/model': 1.465066447854042, 'learning_rate/qwen_vl_interface': 3.1261705127298934e-06, 'learning_rate/action_model': 3.1261705127298935e-05, 'epoch': 2.09}) +[2026-06-16 04:49:25,933][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.11842241883277893, 'train/grad_norm_pre_clip': 15.440826416015625, 'timing/data': 0.011136732064187527, 'timing/model': 1.4060308383777738, 'learning_rate/qwen_vl_interface': 3.123446902217028e-06, 'learning_rate/action_model': 3.1234469022170283e-05, 'epoch': 2.09}) +[2026-06-16 04:49:31,823][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 0.20090436935424805, 'train/grad_norm_pre_clip': 7.665893077850342, 'timing/data': 0.01104081142693758, 'timing/model': 1.4391974080353975, 'learning_rate/qwen_vl_interface': 3.1207241658498145e-06, 'learning_rate/action_model': 3.1207241658498145e-05, 'epoch': 2.09}) +[2026-06-16 04:49:37,678][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.08317544311285019, 'train/grad_norm_pre_clip': 7.047778606414795, 'timing/data': 0.011075170710682869, 'timing/model': 1.3964404473081231, 'learning_rate/qwen_vl_interface': 3.1180023047474667e-06, 'learning_rate/action_model': 3.1180023047474664e-05, 'epoch': 2.09}) +[2026-06-16 04:49:43,644][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.05057400465011597, 'train/grad_norm_pre_clip': 8.464292526245117, 'timing/data': 0.011516834609210491, 'timing/model': 1.442068424075842, 'learning_rate/qwen_vl_interface': 3.1152813200288386e-06, 'learning_rate/action_model': 3.115281320028838e-05, 'epoch': 2.09}) +[2026-06-16 04:49:49,455][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.10403171181678772, 'train/grad_norm_pre_clip': 10.503650665283203, 'timing/data': 0.011300493031740189, 'timing/model': 1.7027409691363573, 'learning_rate/qwen_vl_interface': 3.1125612128124244e-06, 'learning_rate/action_model': 3.112561212812424e-05, 'epoch': 2.09}) +[2026-06-16 04:49:55,286][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 0.25057241320610046, 'train/grad_norm_pre_clip': 14.341364860534668, 'timing/data': 0.01218295656144619, 'timing/model': 1.6888002110645175, 'learning_rate/qwen_vl_interface': 3.1098419842163573e-06, 'learning_rate/action_model': 3.109841984216357e-05, 'epoch': 2.09}) +[2026-06-16 04:50:00,948][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.11263580620288849, 'train/grad_norm_pre_clip': 9.884264945983887, 'timing/data': 0.01165599375963211, 'timing/model': 1.4441511146724224, 'learning_rate/qwen_vl_interface': 3.10712363535841e-06, 'learning_rate/action_model': 3.10712363535841e-05, 'epoch': 2.09}) +[2026-06-16 04:50:06,746][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.017654409632086754, 'train/grad_norm_pre_clip': 1.5676159858703613, 'timing/data': 0.011266142129898071, 'timing/model': 1.44788914360106, 'learning_rate/qwen_vl_interface': 3.1044061673559926e-06, 'learning_rate/action_model': 3.1044061673559925e-05, 'epoch': 2.1}) +[2026-06-16 04:50:12,551][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.01520125288516283, 'train/grad_norm_pre_clip': 1.412445306777954, 'timing/data': 0.011298001743853092, 'timing/model': 1.4366132486611605, 'learning_rate/qwen_vl_interface': 3.1016895813261556e-06, 'learning_rate/action_model': 3.101689581326155e-05, 'epoch': 2.1}) +[2026-06-16 04:50:18,388][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.18590569496154785, 'train/grad_norm_pre_clip': 10.196391105651855, 'timing/data': 0.012155317701399326, 'timing/model': 1.4122327314689755, 'learning_rate/qwen_vl_interface': 3.098973878385581e-06, 'learning_rate/action_model': 3.098973878385581e-05, 'epoch': 2.1}) +[2026-06-16 04:50:24,245][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.09931023418903351, 'train/grad_norm_pre_clip': 8.872666358947754, 'timing/data': 0.011448442004621029, 'timing/model': 1.758373960852623, 'learning_rate/qwen_vl_interface': 3.096259059650594e-06, 'learning_rate/action_model': 3.096259059650594e-05, 'epoch': 2.1}) +[2026-06-16 04:50:30,153][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.04669361189007759, 'train/grad_norm_pre_clip': 8.139324188232422, 'timing/data': 0.011397741734981537, 'timing/model': 1.749249485321343, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 2.1}) +[2026-06-16 04:50:35,764][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.24908995628356934, 'train/grad_norm_pre_clip': 11.781086921691895, 'timing/data': 0.011432141996920109, 'timing/model': 1.4618449751287699, 'learning_rate/qwen_vl_interface': 3.090832079260855e-06, 'learning_rate/action_model': 3.0908320792608544e-05, 'epoch': 2.1}) +[2026-06-16 04:50:41,669][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.19451051950454712, 'train/grad_norm_pre_clip': 9.934941291809082, 'timing/data': 0.01136276125907898, 'timing/model': 1.44447289314121, 'learning_rate/qwen_vl_interface': 3.088119919836929e-06, 'learning_rate/action_model': 3.088119919836929e-05, 'epoch': 2.1}) +[2026-06-16 04:50:47,587][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.06128145009279251, 'train/grad_norm_pre_clip': 10.250690460205078, 'timing/data': 0.011282320134341717, 'timing/model': 1.48677546530962, 'learning_rate/qwen_vl_interface': 3.0854086490802433e-06, 'learning_rate/action_model': 3.085408649080243e-05, 'epoch': 2.1}) +[2026-06-16 04:50:53,462][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.12514305114746094, 'train/grad_norm_pre_clip': 9.762834548950195, 'timing/data': 0.011583021841943264, 'timing/model': 1.4494657795876265, 'learning_rate/qwen_vl_interface': 3.082698268105299e-06, 'learning_rate/action_model': 3.082698268105299e-05, 'epoch': 2.1}) +[2026-06-16 04:50:59,418][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.030626295134425163, 'train/grad_norm_pre_clip': 4.520323276519775, 'timing/data': 0.012245915830135345, 'timing/model': 1.816418595612049, 'learning_rate/qwen_vl_interface': 3.0799887780262295e-06, 'learning_rate/action_model': 3.0799887780262296e-05, 'epoch': 2.1}) +[2026-06-16 04:51:05,092][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.10658232867717743, 'train/grad_norm_pre_clip': 12.258344650268555, 'timing/data': 0.011471531353890896, 'timing/model': 1.4931121366098523, 'learning_rate/qwen_vl_interface': 3.0772801799568054e-06, 'learning_rate/action_model': 3.0772801799568054e-05, 'epoch': 2.1}) +[2026-06-16 04:51:11,081][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.05820126459002495, 'train/grad_norm_pre_clip': 5.468519687652588, 'timing/data': 0.01243064645677805, 'timing/model': 1.4667703723534942, 'learning_rate/qwen_vl_interface': 3.07457247501043e-06, 'learning_rate/action_model': 3.0745724750104296e-05, 'epoch': 2.1}) +[2026-06-16 04:51:17,093][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.04252168908715248, 'train/grad_norm_pre_clip': 6.78329610824585, 'timing/data': 0.011079018004238605, 'timing/model': 1.5020759366452694, 'learning_rate/qwen_vl_interface': 3.0718656643001353e-06, 'learning_rate/action_model': 3.071865664300135e-05, 'epoch': 2.1}) +[2026-06-16 04:51:23,562][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.05057715252041817, 'train/grad_norm_pre_clip': 6.520705223083496, 'timing/data': 0.012215495109558105, 'timing/model': 1.4230541763827205, 'learning_rate/qwen_vl_interface': 3.069159748938591e-06, 'learning_rate/action_model': 3.069159748938591e-05, 'epoch': 2.1}) +[2026-06-16 04:51:29,360][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.06657326221466064, 'train/grad_norm_pre_clip': 13.470094680786133, 'timing/data': 0.011699531227350235, 'timing/model': 1.4064405048266053, 'learning_rate/qwen_vl_interface': 3.066454730038096e-06, 'learning_rate/action_model': 3.0664547300380956e-05, 'epoch': 2.1}) +[2026-06-16 04:51:35,236][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 0.29930320382118225, 'train/grad_norm_pre_clip': 24.071033477783203, 'timing/data': 0.011390520259737968, 'timing/model': 1.7259708298370242, 'learning_rate/qwen_vl_interface': 3.0637506087105816e-06, 'learning_rate/action_model': 3.0637506087105815e-05, 'epoch': 2.1}) +[2026-06-16 04:51:40,867][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.2000553458929062, 'train/grad_norm_pre_clip': 11.12799072265625, 'timing/data': 0.011449059471487999, 'timing/model': 1.4532939922064543, 'learning_rate/qwen_vl_interface': 3.0610473860676103e-06, 'learning_rate/action_model': 3.06104738606761e-05, 'epoch': 2.1}) +[2026-06-16 04:51:46,783][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.36301469802856445, 'train/grad_norm_pre_clip': 9.500885963439941, 'timing/data': 0.010745025239884853, 'timing/model': 1.4624119875952601, 'learning_rate/qwen_vl_interface': 3.0583450632203748e-06, 'learning_rate/action_model': 3.058345063220375e-05, 'epoch': 2.1}) +[2026-06-16 04:51:52,762][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.15405187010765076, 'train/grad_norm_pre_clip': 10.223909378051758, 'timing/data': 0.011417179368436337, 'timing/model': 1.4486863445490599, 'learning_rate/qwen_vl_interface': 3.055643641279699e-06, 'learning_rate/action_model': 3.055643641279699e-05, 'epoch': 2.1}) +[2026-06-16 05:02:28,476][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.06176075339317322, 'train/grad_norm_pre_clip': 9.445669174194336, 'eval/loss': 0.06940609961748123, 'eval/flap_precision': 0.7638036809815951, 'eval/flap_recall': 0.9054545454545454, 'eval/flap_f1': 0.8286189683860232, 'eval/flap_rate_model': 0.09055555555555556, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.8286189683860232, 'timing/data': 0.011315979063510895, 'timing/model': 1.4502035528421402, 'learning_rate/qwen_vl_interface': 3.0529431213560334e-06, 'learning_rate/action_model': 3.0529431213560334e-05, 'epoch': 2.1}) +[2026-06-16 05:02:28,477][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4600_state +[2026-06-16 05:02:28,483][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 05:02:36,144][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4600_state/model.safetensors +[2026-06-16 05:02:47,333][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4600_state/optimizer.bin +[2026-06-16 05:02:47,334][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4600_state/sampler.bin +[2026-06-16 05:02:47,334][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4600_state/sampler_1.bin +[2026-06-16 05:02:47,336][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4600_state/random_states_0.pkl +[2026-06-16 05:02:47,503][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4600 +[2026-06-16 05:02:47,504][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 05:02:47,520][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 05:02:53,237][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.06549816578626633, 'train/grad_norm_pre_clip': 7.799249649047852, 'timing/data': 0.011754131875932217, 'timing/model': 1.3759279763326049, 'learning_rate/qwen_vl_interface': 3.0502435045594596e-06, 'learning_rate/action_model': 3.0502435045594595e-05, 'epoch': 2.1}) +[2026-06-16 05:02:59,031][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.3630606532096863, 'train/grad_norm_pre_clip': 20.967519760131836, 'timing/data': 0.011251789517700672, 'timing/model': 1.4407315403223038, 'learning_rate/qwen_vl_interface': 3.047544791999689e-06, 'learning_rate/action_model': 3.0475447919996892e-05, 'epoch': 2.1}) +[2026-06-16 05:03:04,847][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 0.2021450251340866, 'train/grad_norm_pre_clip': 10.735991477966309, 'timing/data': 0.011533210054039955, 'timing/model': 1.7603360898792744, 'learning_rate/qwen_vl_interface': 3.044846984786061e-06, 'learning_rate/action_model': 3.044846984786061e-05, 'epoch': 2.11}) +[2026-06-16 05:03:10,330][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.0835409089922905, 'train/grad_norm_pre_clip': 12.518372535705566, 'timing/data': 0.0115382494404912, 'timing/model': 1.43534829095006, 'learning_rate/qwen_vl_interface': 3.0421500840275413e-06, 'learning_rate/action_model': 3.0421500840275413e-05, 'epoch': 2.11}) +[2026-06-16 05:03:16,041][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.1735365390777588, 'train/grad_norm_pre_clip': 9.530957221984863, 'timing/data': 0.011320148594677448, 'timing/model': 1.4557632328942418, 'learning_rate/qwen_vl_interface': 3.039454090832727e-06, 'learning_rate/action_model': 3.0394540908327268e-05, 'epoch': 2.11}) +[2026-06-16 05:03:21,657][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.052578821778297424, 'train/grad_norm_pre_clip': 7.618810653686523, 'timing/data': 0.011026086285710335, 'timing/model': 1.385121000930667, 'learning_rate/qwen_vl_interface': 3.0367590063098333e-06, 'learning_rate/action_model': 3.0367590063098333e-05, 'epoch': 2.11}) +[2026-06-16 05:03:27,434][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.036818310618400574, 'train/grad_norm_pre_clip': 5.083541393280029, 'timing/data': 0.010421553626656532, 'timing/model': 1.44677325617522, 'learning_rate/qwen_vl_interface': 3.0340648315667115e-06, 'learning_rate/action_model': 3.0340648315667113e-05, 'epoch': 2.11}) +[2026-06-16 05:03:33,137][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.33042848110198975, 'train/grad_norm_pre_clip': 13.869223594665527, 'timing/data': 0.010904845781624317, 'timing/model': 1.409363097511232, 'learning_rate/qwen_vl_interface': 3.0313715677108336e-06, 'learning_rate/action_model': 3.0313715677108335e-05, 'epoch': 2.11}) +[2026-06-16 05:03:38,910][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.07477185130119324, 'train/grad_norm_pre_clip': 6.917356014251709, 'timing/data': 0.012109803035855293, 'timing/model': 1.7323405779898167, 'learning_rate/qwen_vl_interface': 3.0286792158492996e-06, 'learning_rate/action_model': 3.0286792158492994e-05, 'epoch': 2.11}) +[2026-06-16 05:03:44,445][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.09864704310894012, 'train/grad_norm_pre_clip': 8.006677627563477, 'timing/data': 0.013238138519227505, 'timing/model': 1.4484558794647455, 'learning_rate/qwen_vl_interface': 3.0259877770888334e-06, 'learning_rate/action_model': 3.0259877770888333e-05, 'epoch': 2.11}) +[2026-06-16 05:03:50,251][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.039060600101947784, 'train/grad_norm_pre_clip': 3.597720146179199, 'timing/data': 0.010899927467107773, 'timing/model': 1.470001457259059, 'learning_rate/qwen_vl_interface': 3.0232972525357836e-06, 'learning_rate/action_model': 3.0232972525357834e-05, 'epoch': 2.11}) +[2026-06-16 05:03:56,172][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.0460207462310791, 'train/grad_norm_pre_clip': 10.633401870727539, 'timing/data': 0.012552355416119099, 'timing/model': 1.4786356873810291, 'learning_rate/qwen_vl_interface': 3.020607643296123e-06, 'learning_rate/action_model': 3.0206076432961226e-05, 'epoch': 2.11}) +[2026-06-16 05:04:02,179][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 0.11771343648433685, 'train/grad_norm_pre_clip': 5.58974027633667, 'timing/data': 0.011391409672796726, 'timing/model': 1.4814087264239788, 'learning_rate/qwen_vl_interface': 3.017918950475449e-06, 'learning_rate/action_model': 3.017918950475449e-05, 'epoch': 2.11}) +[2026-06-16 05:04:08,073][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.08517000079154968, 'train/grad_norm_pre_clip': 4.54371976852417, 'timing/data': 0.011568429879844189, 'timing/model': 1.4005718026310205, 'learning_rate/qwen_vl_interface': 3.0152311751789815e-06, 'learning_rate/action_model': 3.0152311751789814e-05, 'epoch': 2.11}) +[2026-06-16 05:04:14,047][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.11771434545516968, 'train/grad_norm_pre_clip': 7.094088554382324, 'timing/data': 0.01524087879806757, 'timing/model': 1.8106083581224084, 'learning_rate/qwen_vl_interface': 3.0125443185115655e-06, 'learning_rate/action_model': 3.0125443185115655e-05, 'epoch': 2.11}) +[2026-06-16 05:04:19,654][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 0.014392629265785217, 'train/grad_norm_pre_clip': 10.948654174804688, 'timing/data': 0.01178389135748148, 'timing/model': 1.468337895348668, 'learning_rate/qwen_vl_interface': 3.0098583815776615e-06, 'learning_rate/action_model': 3.0098583815776615e-05, 'epoch': 2.11}) +[2026-06-16 05:04:25,546][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.15741296112537384, 'train/grad_norm_pre_clip': 8.232430458068848, 'timing/data': 0.011261537671089172, 'timing/model': 1.4568514805287123, 'learning_rate/qwen_vl_interface': 3.0071733654813595e-06, 'learning_rate/action_model': 3.0071733654813595e-05, 'epoch': 2.11}) +[2026-06-16 05:04:31,467][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.2597372531890869, 'train/grad_norm_pre_clip': 10.911734580993652, 'timing/data': 0.011326389387249947, 'timing/model': 1.481790654361248, 'learning_rate/qwen_vl_interface': 3.0044892713263685e-06, 'learning_rate/action_model': 3.0044892713263682e-05, 'epoch': 2.11}) +[2026-06-16 05:04:37,364][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.09566173702478409, 'train/grad_norm_pre_clip': 7.352785587310791, 'timing/data': 0.011484778486192226, 'timing/model': 1.4682368906214833, 'learning_rate/qwen_vl_interface': 3.001806100216017e-06, 'learning_rate/action_model': 3.001806100216017e-05, 'epoch': 2.11}) +[2026-06-16 05:04:43,151][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.09859514236450195, 'train/grad_norm_pre_clip': 14.551526069641113, 'timing/data': 0.01140624936670065, 'timing/model': 1.395493820309639, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 2.11}) +[2026-06-16 05:04:49,027][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.1193474605679512, 'train/grad_norm_pre_clip': 7.453799247741699, 'timing/data': 0.010922555811703205, 'timing/model': 1.7402433529496193, 'learning_rate/qwen_vl_interface': 2.996442531540657e-06, 'learning_rate/action_model': 2.9964425315406568e-05, 'epoch': 2.11}) +[2026-06-16 05:04:54,660][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 0.12742455303668976, 'train/grad_norm_pre_clip': 10.849055290222168, 'timing/data': 0.011329618282616138, 'timing/model': 1.4714064225554466, 'learning_rate/qwen_vl_interface': 2.993762136180408e-06, 'learning_rate/action_model': 2.993762136180408e-05, 'epoch': 2.11}) +[2026-06-16 05:05:00,548][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.0795198455452919, 'train/grad_norm_pre_clip': 18.909870147705078, 'timing/data': 0.011827690526843071, 'timing/model': 1.458756004460156, 'learning_rate/qwen_vl_interface': 2.9910826682743195e-06, 'learning_rate/action_model': 2.9910826682743192e-05, 'epoch': 2.11}) +[2026-06-16 05:05:06,452][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.12247288972139359, 'train/grad_norm_pre_clip': 8.317727088928223, 'timing/data': 0.013384738937020302, 'timing/model': 1.4635836742818356, 'learning_rate/qwen_vl_interface': 2.9884041289238192e-06, 'learning_rate/action_model': 2.988404128923819e-05, 'epoch': 2.11}) +[2026-06-16 05:05:12,360][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 0.03126893192529678, 'train/grad_norm_pre_clip': 3.2447948455810547, 'timing/data': 0.011083226650953293, 'timing/model': 1.4737842325121164, 'learning_rate/qwen_vl_interface': 2.9857265192299557e-06, 'learning_rate/action_model': 2.9857265192299556e-05, 'epoch': 2.12}) +[2026-06-16 05:05:18,466][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.25594139099121094, 'train/grad_norm_pre_clip': 11.337618827819824, 'timing/data': 0.010473494417965412, 'timing/model': 1.9780532084405422, 'learning_rate/qwen_vl_interface': 2.9830498402933894e-06, 'learning_rate/action_model': 2.9830498402933894e-05, 'epoch': 2.12}) +[2026-06-16 05:05:24,276][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.25011494755744934, 'train/grad_norm_pre_clip': 10.251333236694336, 'timing/data': 0.010848185047507286, 'timing/model': 1.6911077303811908, 'learning_rate/qwen_vl_interface': 2.9803740932144025e-06, 'learning_rate/action_model': 2.9803740932144024e-05, 'epoch': 2.12}) +[2026-06-16 05:05:29,950][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.19093652069568634, 'train/grad_norm_pre_clip': 6.598130226135254, 'timing/data': 0.011587808839976788, 'timing/model': 1.4502478400245309, 'learning_rate/qwen_vl_interface': 2.9776992790928953e-06, 'learning_rate/action_model': 2.9776992790928952e-05, 'epoch': 2.12}) +[2026-06-16 05:05:35,869][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.2155950665473938, 'train/grad_norm_pre_clip': 8.447681427001953, 'timing/data': 0.01500024739652872, 'timing/model': 1.4699914269149303, 'learning_rate/qwen_vl_interface': 2.975025399028382e-06, 'learning_rate/action_model': 2.975025399028382e-05, 'epoch': 2.12}) +[2026-06-16 05:05:41,761][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.022165242582559586, 'train/grad_norm_pre_clip': 1.6195155382156372, 'timing/data': 0.011202027089893818, 'timing/model': 1.4571146070957184, 'learning_rate/qwen_vl_interface': 2.9723524541199943e-06, 'learning_rate/action_model': 2.972352454119994e-05, 'epoch': 2.12}) +[2026-06-16 05:05:47,663][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.3633645176887512, 'train/grad_norm_pre_clip': 13.509339332580566, 'timing/data': 0.013049156405031681, 'timing/model': 1.4701638901606202, 'learning_rate/qwen_vl_interface': 2.969680445466478e-06, 'learning_rate/action_model': 2.969680445466478e-05, 'epoch': 2.12}) +[2026-06-16 05:05:53,574][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.10646478086709976, 'train/grad_norm_pre_clip': 5.374387264251709, 'timing/data': 0.010226762853562832, 'timing/model': 1.7718859706074, 'learning_rate/qwen_vl_interface': 2.9670093741661955e-06, 'learning_rate/action_model': 2.9670093741661954e-05, 'epoch': 2.12}) +[2026-06-16 05:05:59,493][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.04838712885975838, 'train/grad_norm_pre_clip': 6.980974197387695, 'timing/data': 0.011801119893789291, 'timing/model': 1.7003290792927146, 'learning_rate/qwen_vl_interface': 2.964339241317123e-06, 'learning_rate/action_model': 2.964339241317123e-05, 'epoch': 2.12}) +[2026-06-16 05:06:05,161][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.10461223870515823, 'train/grad_norm_pre_clip': 5.774691581726074, 'timing/data': 0.01194826140999794, 'timing/model': 1.4711286732926965, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 2.12}) +[2026-06-16 05:06:11,055][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.07806626707315445, 'train/grad_norm_pre_clip': 3.466916084289551, 'timing/data': 0.01209111139178276, 'timing/model': 1.450002902187407, 'learning_rate/qwen_vl_interface': 2.9590017953625854e-06, 'learning_rate/action_model': 2.9590017953625853e-05, 'epoch': 2.12}) +[2026-06-16 05:06:16,858][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.08051195740699768, 'train/grad_norm_pre_clip': 5.618871212005615, 'timing/data': 0.014284352771937847, 'timing/model': 1.428960571065545, 'learning_rate/qwen_vl_interface': 2.9563344844511455e-06, 'learning_rate/action_model': 2.9563344844511455e-05, 'epoch': 2.12}) +[2026-06-16 05:06:22,750][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.030019918456673622, 'train/grad_norm_pre_clip': 4.527192115783691, 'timing/data': 0.01155109889805317, 'timing/model': 1.4505025977268815, 'learning_rate/qwen_vl_interface': 2.9536681163789568e-06, 'learning_rate/action_model': 2.9536681163789564e-05, 'epoch': 2.12}) +[2026-06-16 05:06:28,686][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.11800762265920639, 'train/grad_norm_pre_clip': 6.999134063720703, 'timing/data': 0.011105336248874664, 'timing/model': 1.788480818271637, 'learning_rate/qwen_vl_interface': 2.9510026922420653e-06, 'learning_rate/action_model': 2.9510026922420653e-05, 'epoch': 2.12}) +[2026-06-16 05:06:34,299][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.08207661658525467, 'train/grad_norm_pre_clip': 3.955803394317627, 'timing/data': 0.011311047710478306, 'timing/model': 1.4162339456379414, 'learning_rate/qwen_vl_interface': 2.9483382131361273e-06, 'learning_rate/action_model': 2.948338213136127e-05, 'epoch': 2.12}) +[2026-06-16 05:06:40,272][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.26933708786964417, 'train/grad_norm_pre_clip': 12.76189136505127, 'timing/data': 0.02940460853278637, 'timing/model': 1.4189553167670965, 'learning_rate/qwen_vl_interface': 2.9456746801564072e-06, 'learning_rate/action_model': 2.945674680156407e-05, 'epoch': 2.12}) +[2026-06-16 05:06:46,304][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.1479489803314209, 'train/grad_norm_pre_clip': 10.48398494720459, 'timing/data': 0.015183476731181145, 'timing/model': 1.5060460260137916, 'learning_rate/qwen_vl_interface': 2.9430120943977854e-06, 'learning_rate/action_model': 2.9430120943977853e-05, 'epoch': 2.12}) +[2026-06-16 05:06:52,265][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.01917196810245514, 'train/grad_norm_pre_clip': 1.5643901824951172, 'timing/data': 0.012767825275659561, 'timing/model': 1.4303651973605156, 'learning_rate/qwen_vl_interface': 2.940350456954748e-06, 'learning_rate/action_model': 2.940350456954748e-05, 'epoch': 2.12}) +[2026-06-16 05:06:58,176][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.022623803466558456, 'train/grad_norm_pre_clip': 2.483978509902954, 'timing/data': 0.012903975322842598, 'timing/model': 1.4411240993067622, 'learning_rate/qwen_vl_interface': 2.9376897689213952e-06, 'learning_rate/action_model': 2.937689768921395e-05, 'epoch': 2.12}) +[2026-06-16 05:07:04,230][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 0.2649243474006653, 'train/grad_norm_pre_clip': 10.71356201171875, 'timing/data': 0.013783990405499935, 'timing/model': 1.7937030494213104, 'learning_rate/qwen_vl_interface': 2.9350300313914364e-06, 'learning_rate/action_model': 2.9350300313914363e-05, 'epoch': 2.12}) +[2026-06-16 05:07:09,878][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 0.02202809602022171, 'train/grad_norm_pre_clip': 2.356342315673828, 'timing/data': 0.013661248609423637, 'timing/model': 1.4578757667914033, 'learning_rate/qwen_vl_interface': 2.9323712454581867e-06, 'learning_rate/action_model': 2.9323712454581866e-05, 'epoch': 2.12}) +[2026-06-16 05:07:15,892][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.04109939932823181, 'train/grad_norm_pre_clip': 5.171911716461182, 'timing/data': 0.013551829382777214, 'timing/model': 1.4763455064967275, 'learning_rate/qwen_vl_interface': 2.929713412214572e-06, 'learning_rate/action_model': 2.929713412214572e-05, 'epoch': 2.12}) +[2026-06-16 05:07:21,900][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.02037998102605343, 'train/grad_norm_pre_clip': 3.0540735721588135, 'timing/data': 0.013008175417780876, 'timing/model': 1.4863380454480648, 'learning_rate/qwen_vl_interface': 2.9270565327531294e-06, 'learning_rate/action_model': 2.927056532753129e-05, 'epoch': 2.13}) +[2026-06-16 05:07:27,980][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.19806762039661407, 'train/grad_norm_pre_clip': 13.218964576721191, 'timing/data': 0.013004676438868046, 'timing/model': 1.4670007238164544, 'learning_rate/qwen_vl_interface': 2.924400608166003e-06, 'learning_rate/action_model': 2.924400608166003e-05, 'epoch': 2.13}) +[2026-06-16 05:07:34,003][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.08265075087547302, 'train/grad_norm_pre_clip': 6.065984725952148, 'timing/data': 0.012706434354186058, 'timing/model': 1.4353154990822077, 'learning_rate/qwen_vl_interface': 2.921745639544939e-06, 'learning_rate/action_model': 2.921745639544939e-05, 'epoch': 2.13}) +[2026-06-16 05:07:40,028][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.0731280967593193, 'train/grad_norm_pre_clip': 6.791808128356934, 'timing/data': 0.013555899262428284, 'timing/model': 1.8212276548147202, 'learning_rate/qwen_vl_interface': 2.919091627981297e-06, 'learning_rate/action_model': 2.919091627981297e-05, 'epoch': 2.13}) +[2026-06-16 05:07:45,961][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.08730804920196533, 'train/grad_norm_pre_clip': 7.926483631134033, 'timing/data': 0.013879049569368362, 'timing/model': 1.744518242776394, 'learning_rate/qwen_vl_interface': 2.9164385745660396e-06, 'learning_rate/action_model': 2.9164385745660394e-05, 'epoch': 2.13}) +[2026-06-16 05:07:51,580][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.053498733788728714, 'train/grad_norm_pre_clip': 3.924592971801758, 'timing/data': 0.013261286541819572, 'timing/model': 1.4674891037866473, 'learning_rate/qwen_vl_interface': 2.9137864803897385e-06, 'learning_rate/action_model': 2.9137864803897387e-05, 'epoch': 2.13}) +[2026-06-16 05:07:57,574][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.04670466482639313, 'train/grad_norm_pre_clip': 4.345886707305908, 'timing/data': 0.013642978854477406, 'timing/model': 1.4586214944720268, 'learning_rate/qwen_vl_interface': 2.9111353465425696e-06, 'learning_rate/action_model': 2.9111353465425694e-05, 'epoch': 2.13}) +[2026-06-16 05:08:03,689][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.031026333570480347, 'train/grad_norm_pre_clip': 3.3603358268737793, 'timing/data': 0.013063516467809677, 'timing/model': 1.4797382084652781, 'learning_rate/qwen_vl_interface': 2.908485174114312e-06, 'learning_rate/action_model': 2.908485174114312e-05, 'epoch': 2.13}) +[2026-06-16 05:08:09,789][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.0978880524635315, 'train/grad_norm_pre_clip': 4.488698482513428, 'timing/data': 0.013985160738229752, 'timing/model': 1.4925241004675627, 'learning_rate/qwen_vl_interface': 2.9058359641943523e-06, 'learning_rate/action_model': 2.9058359641943523e-05, 'epoch': 2.13}) +[2026-06-16 05:08:15,824][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.039048198610544205, 'train/grad_norm_pre_clip': 6.4160003662109375, 'timing/data': 0.013692989014089108, 'timing/model': 1.8169795414432883, 'learning_rate/qwen_vl_interface': 2.9031877178716823e-06, 'learning_rate/action_model': 2.903187717871682e-05, 'epoch': 2.13}) +[2026-06-16 05:08:21,827][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.14255687594413757, 'train/grad_norm_pre_clip': 10.081327438354492, 'timing/data': 0.0130210155621171, 'timing/model': 1.8177453810349107, 'learning_rate/qwen_vl_interface': 2.900540436234894e-06, 'learning_rate/action_model': 2.9005404362348937e-05, 'epoch': 2.13}) +[2026-06-16 05:08:27,456][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 0.23920007050037384, 'train/grad_norm_pre_clip': 13.469173431396484, 'timing/data': 0.013467107899487019, 'timing/model': 1.441088031977415, 'learning_rate/qwen_vl_interface': 2.897894120372185e-06, 'learning_rate/action_model': 2.897894120372185e-05, 'epoch': 2.13}) +[2026-06-16 05:08:33,358][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.2029334157705307, 'train/grad_norm_pre_clip': 11.450335502624512, 'timing/data': 0.013896460644900799, 'timing/model': 1.425909391604364, 'learning_rate/qwen_vl_interface': 2.8952487713713585e-06, 'learning_rate/action_model': 2.8952487713713582e-05, 'epoch': 2.13}) +[2026-06-16 05:08:39,281][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.07826825231313705, 'train/grad_norm_pre_clip': 6.293972969055176, 'timing/data': 0.014026191085577011, 'timing/model': 1.4796149795874953, 'learning_rate/qwen_vl_interface': 2.8926043903198135e-06, 'learning_rate/action_model': 2.8926043903198135e-05, 'epoch': 2.13}) +[2026-06-16 05:08:45,262][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.03370022028684616, 'train/grad_norm_pre_clip': 3.118189573287964, 'timing/data': 0.013708559796214104, 'timing/model': 1.4565635612234473, 'learning_rate/qwen_vl_interface': 2.8899609783045574e-06, 'learning_rate/action_model': 2.8899609783045573e-05, 'epoch': 2.13}) +[2026-06-16 05:08:51,176][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.4940534234046936, 'train/grad_norm_pre_clip': 14.564441680908203, 'timing/data': 0.014590754173696041, 'timing/model': 1.7648517759516835, 'learning_rate/qwen_vl_interface': 2.8873185364121965e-06, 'learning_rate/action_model': 2.8873185364121963e-05, 'epoch': 2.13}) +[2026-06-16 05:08:56,871][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.3935462236404419, 'train/grad_norm_pre_clip': 13.676534652709961, 'timing/data': 0.013755979016423225, 'timing/model': 1.457420659251511, 'learning_rate/qwen_vl_interface': 2.8846770657289386e-06, 'learning_rate/action_model': 2.8846770657289386e-05, 'epoch': 2.13}) +[2026-06-16 05:09:02,871][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.21202018857002258, 'train/grad_norm_pre_clip': 6.866201400756836, 'timing/data': 0.015267436392605305, 'timing/model': 1.4574730983003974, 'learning_rate/qwen_vl_interface': 2.882036567340593e-06, 'learning_rate/action_model': 2.8820365673405932e-05, 'epoch': 2.13}) +[2026-06-16 05:09:08,866][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.08846613764762878, 'train/grad_norm_pre_clip': 8.300395011901855, 'timing/data': 0.013282876461744308, 'timing/model': 1.4625873314216733, 'learning_rate/qwen_vl_interface': 2.8793970423325684e-06, 'learning_rate/action_model': 2.8793970423325685e-05, 'epoch': 2.13}) +[2026-06-16 05:09:14,783][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.17777344584465027, 'train/grad_norm_pre_clip': 10.801722526550293, 'timing/data': 0.013324867933988571, 'timing/model': 1.4693605424836278, 'learning_rate/qwen_vl_interface': 2.8767584917898756e-06, 'learning_rate/action_model': 2.8767584917898754e-05, 'epoch': 2.13}) +[2026-06-16 05:09:20,669][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.11788525432348251, 'train/grad_norm_pre_clip': 13.57555866241455, 'timing/data': 0.012835084460675716, 'timing/model': 1.4123521018773317, 'learning_rate/qwen_vl_interface': 2.874120916797121e-06, 'learning_rate/action_model': 2.8741209167971207e-05, 'epoch': 2.13}) +[2026-06-16 05:09:26,544][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 0.03771490976214409, 'train/grad_norm_pre_clip': 3.9292991161346436, 'timing/data': 0.013367246836423874, 'timing/model': 1.7199834706261754, 'learning_rate/qwen_vl_interface': 2.8714843184385124e-06, 'learning_rate/action_model': 2.871484318438512e-05, 'epoch': 2.13}) +[2026-06-16 05:09:32,175][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 0.008258773945271969, 'train/grad_norm_pre_clip': 0.7476466298103333, 'timing/data': 0.013383977115154266, 'timing/model': 1.4712076541036367, 'learning_rate/qwen_vl_interface': 2.8688486977978575e-06, 'learning_rate/action_model': 2.8688486977978572e-05, 'epoch': 2.14}) +[2026-06-16 05:09:38,205][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.13566036522388458, 'train/grad_norm_pre_clip': 7.524288654327393, 'timing/data': 0.012630054727196693, 'timing/model': 1.5087173311039805, 'learning_rate/qwen_vl_interface': 2.8662140559585584e-06, 'learning_rate/action_model': 2.866214055958558e-05, 'epoch': 2.14}) +[2026-06-16 05:09:44,171][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.26728856563568115, 'train/grad_norm_pre_clip': 13.52920913696289, 'timing/data': 0.012336072511970997, 'timing/model': 1.427236506715417, 'learning_rate/qwen_vl_interface': 2.8635803940036162e-06, 'learning_rate/action_model': 2.863580394003616e-05, 'epoch': 2.14}) +[2026-06-16 05:09:50,209][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.18205298483371735, 'train/grad_norm_pre_clip': 13.433079719543457, 'timing/data': 0.01351871807128191, 'timing/model': 1.4477281291037798, 'learning_rate/qwen_vl_interface': 2.8609477130156317e-06, 'learning_rate/action_model': 2.8609477130156314e-05, 'epoch': 2.14}) +[2026-06-16 05:09:56,199][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.04070204123854637, 'train/grad_norm_pre_clip': 4.369950771331787, 'timing/data': 0.014001810923218727, 'timing/model': 1.4345555482432246, 'learning_rate/qwen_vl_interface': 2.8583160140767997e-06, 'learning_rate/action_model': 2.8583160140767994e-05, 'epoch': 2.14}) +[2026-06-16 05:10:02,206][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.09903264045715332, 'train/grad_norm_pre_clip': 7.621581554412842, 'timing/data': 0.014919186010956764, 'timing/model': 1.7766377180814743, 'learning_rate/qwen_vl_interface': 2.8556852982689125e-06, 'learning_rate/action_model': 2.8556852982689126e-05, 'epoch': 2.14}) +[2026-06-16 05:10:07,871][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.035083454102277756, 'train/grad_norm_pre_clip': 3.7550857067108154, 'timing/data': 0.012374731712043285, 'timing/model': 1.447281968779862, 'learning_rate/qwen_vl_interface': 2.8530555666733584e-06, 'learning_rate/action_model': 2.8530555666733583e-05, 'epoch': 2.14}) +[2026-06-16 05:10:13,972][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.11699321866035461, 'train/grad_norm_pre_clip': 9.226168632507324, 'timing/data': 0.014718104153871536, 'timing/model': 1.4663896122947335, 'learning_rate/qwen_vl_interface': 2.85042682037112e-06, 'learning_rate/action_model': 2.85042682037112e-05, 'epoch': 2.14}) +[2026-06-16 05:10:20,059][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.04099377989768982, 'train/grad_norm_pre_clip': 3.4418375492095947, 'timing/data': 0.017386808060109615, 'timing/model': 1.4168379213660955, 'learning_rate/qwen_vl_interface': 2.8477990604427764e-06, 'learning_rate/action_model': 2.8477990604427763e-05, 'epoch': 2.14}) +[2026-06-16 05:10:25,976][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.06923947483301163, 'train/grad_norm_pre_clip': 6.356698989868164, 'timing/data': 0.013739358633756638, 'timing/model': 1.4516264535486698, 'learning_rate/qwen_vl_interface': 2.8451722879684994e-06, 'learning_rate/action_model': 2.8451722879684994e-05, 'epoch': 2.14}) +[2026-06-16 05:10:32,010][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.11101973801851273, 'train/grad_norm_pre_clip': 10.64290714263916, 'timing/data': 0.012852325104176998, 'timing/model': 1.802872896194458, 'learning_rate/qwen_vl_interface': 2.842546504028059e-06, 'learning_rate/action_model': 2.8425465040280586e-05, 'epoch': 2.14}) +[2026-06-16 05:10:37,660][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.2099040001630783, 'train/grad_norm_pre_clip': 9.206938743591309, 'timing/data': 0.013192975893616676, 'timing/model': 1.4505582535639405, 'learning_rate/qwen_vl_interface': 2.83992170970081e-06, 'learning_rate/action_model': 2.8399217097008102e-05, 'epoch': 2.14}) +[2026-06-16 05:10:43,562][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.08737444132566452, 'train/grad_norm_pre_clip': 4.97988224029541, 'timing/data': 0.013901898637413979, 'timing/model': 1.4521398385986686, 'learning_rate/qwen_vl_interface': 2.837297906065711e-06, 'learning_rate/action_model': 2.8372979060657107e-05, 'epoch': 2.14}) +[2026-06-16 05:10:49,365][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.1747206747531891, 'train/grad_norm_pre_clip': 10.800216674804688, 'timing/data': 0.01345443818718195, 'timing/model': 1.4432335533201694, 'learning_rate/qwen_vl_interface': 2.834675094201306e-06, 'learning_rate/action_model': 2.8346750942013063e-05, 'epoch': 2.14}) +[2026-06-16 05:10:55,191][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.06294441968202591, 'train/grad_norm_pre_clip': 5.378788471221924, 'timing/data': 0.014541903510689735, 'timing/model': 1.4880096158012748, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 2.14}) +[2026-06-16 05:11:01,070][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.10152919590473175, 'train/grad_norm_pre_clip': 6.595031261444092, 'timing/data': 0.012322241440415382, 'timing/model': 1.4170478265732527, 'learning_rate/qwen_vl_interface': 2.829432450096731e-06, 'learning_rate/action_model': 2.8294324500967306e-05, 'epoch': 2.14}) +[2026-06-16 05:11:07,137][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.2622939348220825, 'train/grad_norm_pre_clip': 10.235701560974121, 'timing/data': 0.013056444935500622, 'timing/model': 1.8114787032827735, 'learning_rate/qwen_vl_interface': 2.8268126200116118e-06, 'learning_rate/action_model': 2.8268126200116117e-05, 'epoch': 2.14}) +[2026-06-16 05:11:12,772][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.21314994990825653, 'train/grad_norm_pre_clip': 7.757837772369385, 'timing/data': 0.013153975829482079, 'timing/model': 1.4615295259281993, 'learning_rate/qwen_vl_interface': 2.824193786007294e-06, 'learning_rate/action_model': 2.8241937860072937e-05, 'epoch': 2.14}) +[2026-06-16 05:11:18,772][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.22137048840522766, 'train/grad_norm_pre_clip': 13.650527954101562, 'timing/data': 0.013194006867706776, 'timing/model': 1.4722198778763413, 'learning_rate/qwen_vl_interface': 2.821575949160279e-06, 'learning_rate/action_model': 2.8215759491602787e-05, 'epoch': 2.14}) +[2026-06-16 05:11:24,762][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.1991702914237976, 'train/grad_norm_pre_clip': 11.142822265625, 'timing/data': 0.014412572607398033, 'timing/model': 1.4345046766102314, 'learning_rate/qwen_vl_interface': 2.8189591105466625e-06, 'learning_rate/action_model': 2.8189591105466624e-05, 'epoch': 2.14}) +[2026-06-16 05:11:30,678][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.03549724444746971, 'train/grad_norm_pre_clip': 5.172335624694824, 'timing/data': 0.012849695049226284, 'timing/model': 1.452980956993997, 'learning_rate/qwen_vl_interface': 2.816343271242126e-06, 'learning_rate/action_model': 2.816343271242126e-05, 'epoch': 2.14}) +[2026-06-16 05:11:36,777][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.0762677937746048, 'train/grad_norm_pre_clip': 6.504075527191162, 'timing/data': 0.013593798503279686, 'timing/model': 1.466629614122212, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 2.14}) +[2026-06-16 05:11:42,883][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.13260115683078766, 'train/grad_norm_pre_clip': 5.4055070877075195, 'timing/data': 0.012832404114305973, 'timing/model': 1.8575347773730755, 'learning_rate/qwen_vl_interface': 2.81111459486098e-06, 'learning_rate/action_model': 2.81111459486098e-05, 'epoch': 2.15}) +[2026-06-16 05:11:48,579][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.07640893757343292, 'train/grad_norm_pre_clip': 9.444391250610352, 'timing/data': 0.012927594594657421, 'timing/model': 1.4597261250019073, 'learning_rate/qwen_vl_interface': 2.8085017599336806e-06, 'learning_rate/action_model': 2.8085017599336806e-05, 'epoch': 2.15}) +[2026-06-16 05:11:54,688][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.15926593542099, 'train/grad_norm_pre_clip': 11.877731323242188, 'timing/data': 0.01273963414132595, 'timing/model': 1.4712266009300947, 'learning_rate/qwen_vl_interface': 2.8058899286140828e-06, 'learning_rate/action_model': 2.805889928614083e-05, 'epoch': 2.15}) +[2026-06-16 05:12:00,699][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.10967373847961426, 'train/grad_norm_pre_clip': 6.000218868255615, 'timing/data': 0.01544064749032259, 'timing/model': 1.4723863042891026, 'learning_rate/qwen_vl_interface': 2.8032791019758127e-06, 'learning_rate/action_model': 2.8032791019758127e-05, 'epoch': 2.15}) +[2026-06-16 05:12:06,698][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.2807406783103943, 'train/grad_norm_pre_clip': 8.128670692443848, 'timing/data': 0.015217196196317673, 'timing/model': 1.4279328314587474, 'learning_rate/qwen_vl_interface': 2.800669281092084e-06, 'learning_rate/action_model': 2.8006692810920838e-05, 'epoch': 2.15}) +[2026-06-16 05:12:12,569][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 0.03060244582593441, 'train/grad_norm_pre_clip': 2.0448193550109863, 'timing/data': 0.012046309188008308, 'timing/model': 1.4577897284179926, 'learning_rate/qwen_vl_interface': 2.7980604670356938e-06, 'learning_rate/action_model': 2.7980604670356937e-05, 'epoch': 2.15}) +[2026-06-16 05:12:18,470][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.06779097765684128, 'train/grad_norm_pre_clip': 8.470565795898438, 'timing/data': 0.012510892935097218, 'timing/model': 1.7635270841419697, 'learning_rate/qwen_vl_interface': 2.795452660879027e-06, 'learning_rate/action_model': 2.7954526608790273e-05, 'epoch': 2.15}) +[2026-06-16 05:12:24,167][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.10245086997747421, 'train/grad_norm_pre_clip': 10.942719459533691, 'timing/data': 0.012735032476484776, 'timing/model': 1.4720882661640644, 'learning_rate/qwen_vl_interface': 2.7928458636940565e-06, 'learning_rate/action_model': 2.7928458636940564e-05, 'epoch': 2.15}) +[2026-06-16 05:12:30,057][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.019339976832270622, 'train/grad_norm_pre_clip': 3.281553268432617, 'timing/data': 0.011752499267458916, 'timing/model': 1.44776052236557, 'learning_rate/qwen_vl_interface': 2.790240076552336e-06, 'learning_rate/action_model': 2.7902400765523355e-05, 'epoch': 2.15}) +[2026-06-16 05:23:10,877][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.036560580134391785, 'train/grad_norm_pre_clip': 2.657731294631958, 'eval/loss': 0.061198748648166656, 'eval/flap_precision': 0.8571428571428571, 'eval/flap_recall': 0.8290909090909091, 'eval/flap_f1': 0.8428835489833642, 'eval/flap_rate_model': 0.07388888888888889, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.8428835489833642, 'timing/data': 0.011807727627456188, 'timing/model': 1.4611311377957463, 'learning_rate/qwen_vl_interface': 2.787635300525006e-06, 'learning_rate/action_model': 2.787635300525006e-05, 'epoch': 2.15}) +[2026-06-16 05:23:10,878][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4700_state +[2026-06-16 05:23:10,884][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 05:23:18,019][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4700_state/model.safetensors +[2026-06-16 05:23:29,386][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4700_state/optimizer.bin +[2026-06-16 05:23:29,388][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4700_state/sampler.bin +[2026-06-16 05:23:29,388][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4700_state/sampler_1.bin +[2026-06-16 05:23:29,390][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4700_state/random_states_0.pkl +[2026-06-16 05:23:29,554][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4700 +[2026-06-16 05:23:29,555][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 05:23:29,571][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 05:23:35,451][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.047940708696842194, 'train/grad_norm_pre_clip': 3.359929323196411, 'timing/data': 0.013013631105422974, 'timing/model': 1.4366251099854708, 'learning_rate/qwen_vl_interface': 2.7850315366827933e-06, 'learning_rate/action_model': 2.7850315366827934e-05, 'epoch': 2.15}) +[2026-06-16 05:23:41,238][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.05441107228398323, 'train/grad_norm_pre_clip': 3.94022274017334, 'timing/data': 0.011669817380607128, 'timing/model': 1.3738222904503345, 'learning_rate/qwen_vl_interface': 2.7824287860960076e-06, 'learning_rate/action_model': 2.7824287860960075e-05, 'epoch': 2.15}) +[2026-06-16 05:23:46,944][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.11887368559837341, 'train/grad_norm_pre_clip': 7.278451919555664, 'timing/data': 0.01218729093670845, 'timing/model': 1.375278290361166, 'learning_rate/qwen_vl_interface': 2.7798270498345377e-06, 'learning_rate/action_model': 2.7798270498345376e-05, 'epoch': 2.15}) +[2026-06-16 05:23:52,708][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 0.11516794562339783, 'train/grad_norm_pre_clip': 7.868463039398193, 'timing/data': 0.012522304430603981, 'timing/model': 1.7336148172616959, 'learning_rate/qwen_vl_interface': 2.777226328967861e-06, 'learning_rate/action_model': 2.7772263289678607e-05, 'epoch': 2.15}) +[2026-06-16 05:23:58,435][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.1294979751110077, 'train/grad_norm_pre_clip': 8.606121063232422, 'timing/data': 0.013065478764474392, 'timing/model': 1.6608690368011594, 'learning_rate/qwen_vl_interface': 2.774626624565035e-06, 'learning_rate/action_model': 2.774626624565035e-05, 'epoch': 2.15}) +[2026-06-16 05:24:04,039][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.10186566412448883, 'train/grad_norm_pre_clip': 6.114265441894531, 'timing/data': 0.01503863837569952, 'timing/model': 1.4555934146046638, 'learning_rate/qwen_vl_interface': 2.772027937694701e-06, 'learning_rate/action_model': 2.772027937694701e-05, 'epoch': 2.15}) +[2026-06-16 05:24:09,853][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 0.07397974282503128, 'train/grad_norm_pre_clip': 5.9474945068359375, 'timing/data': 0.012527511455118656, 'timing/model': 1.4637233559042215, 'learning_rate/qwen_vl_interface': 2.7694302694250806e-06, 'learning_rate/action_model': 2.7694302694250808e-05, 'epoch': 2.15}) +[2026-06-16 05:24:15,759][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.0501498319208622, 'train/grad_norm_pre_clip': 5.434012413024902, 'timing/data': 0.013614802621304989, 'timing/model': 1.447570159099996, 'learning_rate/qwen_vl_interface': 2.7668336208239775e-06, 'learning_rate/action_model': 2.7668336208239776e-05, 'epoch': 2.15}) +[2026-06-16 05:24:21,537][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.09280964732170105, 'train/grad_norm_pre_clip': 5.993307590484619, 'timing/data': 0.012375188991427422, 'timing/model': 1.3850094471126795, 'learning_rate/qwen_vl_interface': 2.7642379929587755e-06, 'learning_rate/action_model': 2.7642379929587754e-05, 'epoch': 2.15}) +[2026-06-16 05:24:27,254][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.21174833178520203, 'train/grad_norm_pre_clip': 11.177135467529297, 'timing/data': 0.01187333557754755, 'timing/model': 1.3995141247287393, 'learning_rate/qwen_vl_interface': 2.7616433868964387e-06, 'learning_rate/action_model': 2.7616433868964388e-05, 'epoch': 2.15}) +[2026-06-16 05:24:33,111][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.01982131041586399, 'train/grad_norm_pre_clip': 2.0173726081848145, 'timing/data': 0.01296328380703926, 'timing/model': 1.7235453324392438, 'learning_rate/qwen_vl_interface': 2.7590498037035122e-06, 'learning_rate/action_model': 2.7590498037035124e-05, 'epoch': 2.15}) +[2026-06-16 05:24:39,117][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 0.021034670993685722, 'train/grad_norm_pre_clip': 9.070398330688477, 'timing/data': 0.01210065558552742, 'timing/model': 1.4719110941514373, 'learning_rate/qwen_vl_interface': 2.7564572444461218e-06, 'learning_rate/action_model': 2.7564572444461217e-05, 'epoch': 2.15}) +[2026-06-16 05:24:44,948][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.16591516137123108, 'train/grad_norm_pre_clip': 6.047357559204102, 'timing/data': 0.010207726620137691, 'timing/model': 1.4559197127819061, 'learning_rate/qwen_vl_interface': 2.7538657101899674e-06, 'learning_rate/action_model': 2.7538657101899673e-05, 'epoch': 2.16}) +[2026-06-16 05:24:50,564][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.1103833019733429, 'train/grad_norm_pre_clip': 8.837621688842773, 'timing/data': 0.010005805641412735, 'timing/model': 1.4644507514312863, 'learning_rate/qwen_vl_interface': 2.751275202000332e-06, 'learning_rate/action_model': 2.7512752020003317e-05, 'epoch': 2.16}) +[2026-06-16 05:24:56,263][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.17731553316116333, 'train/grad_norm_pre_clip': 15.578768730163574, 'timing/data': 0.010806111618876457, 'timing/model': 1.3883181922137737, 'learning_rate/qwen_vl_interface': 2.7486857209420763e-06, 'learning_rate/action_model': 2.7486857209420763e-05, 'epoch': 2.16}) +[2026-06-16 05:25:02,063][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.2486017942428589, 'train/grad_norm_pre_clip': 9.915511131286621, 'timing/data': 0.010725630447268486, 'timing/model': 1.39785635471344, 'learning_rate/qwen_vl_interface': 2.7460972680796377e-06, 'learning_rate/action_model': 2.7460972680796375e-05, 'epoch': 2.16}) +[2026-06-16 05:25:07,918][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.04164477810263634, 'train/grad_norm_pre_clip': 6.299475193023682, 'timing/data': 0.011019003577530384, 'timing/model': 1.7228812500834465, 'learning_rate/qwen_vl_interface': 2.7435098444770314e-06, 'learning_rate/action_model': 2.7435098444770315e-05, 'epoch': 2.16}) +[2026-06-16 05:25:13,456][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.043898001313209534, 'train/grad_norm_pre_clip': 4.866371154785156, 'timing/data': 0.010614029131829739, 'timing/model': 1.452960810624063, 'learning_rate/qwen_vl_interface': 2.7409234511978495e-06, 'learning_rate/action_model': 2.7409234511978492e-05, 'epoch': 2.16}) +[2026-06-16 05:25:19,266][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.06991361081600189, 'train/grad_norm_pre_clip': 8.699470520019531, 'timing/data': 0.010908311232924461, 'timing/model': 1.4612990487366915, 'learning_rate/qwen_vl_interface': 2.7383380893052606e-06, 'learning_rate/action_model': 2.7383380893052608e-05, 'epoch': 2.16}) +[2026-06-16 05:25:25,061][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.08102397620677948, 'train/grad_norm_pre_clip': 10.975566864013672, 'timing/data': 0.0111002828925848, 'timing/model': 1.4667758215218782, 'learning_rate/qwen_vl_interface': 2.73575375986201e-06, 'learning_rate/action_model': 2.7357537598620096e-05, 'epoch': 2.16}) +[2026-06-16 05:25:30,856][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.09637413173913956, 'train/grad_norm_pre_clip': 8.209052085876465, 'timing/data': 0.010292324237525463, 'timing/model': 1.4470295514911413, 'learning_rate/qwen_vl_interface': 2.733170463930417e-06, 'learning_rate/action_model': 2.7331704639304167e-05, 'epoch': 2.16}) +[2026-06-16 05:25:36,669][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 0.06503167003393173, 'train/grad_norm_pre_clip': 6.518066883087158, 'timing/data': 0.01100000087171793, 'timing/model': 1.4024533964693546, 'learning_rate/qwen_vl_interface': 2.7305882025723774e-06, 'learning_rate/action_model': 2.7305882025723774e-05, 'epoch': 2.16}) +[2026-06-16 05:25:42,519][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.14561274647712708, 'train/grad_norm_pre_clip': 7.700971603393555, 'timing/data': 0.011649755761027336, 'timing/model': 1.724732642993331, 'learning_rate/qwen_vl_interface': 2.7280069768493626e-06, 'learning_rate/action_model': 2.7280069768493628e-05, 'epoch': 2.16}) +[2026-06-16 05:25:48,446][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.1438915729522705, 'train/grad_norm_pre_clip': 12.690435409545898, 'timing/data': 0.01403005700558424, 'timing/model': 1.7450786884874105, 'learning_rate/qwen_vl_interface': 2.7254267878224144e-06, 'learning_rate/action_model': 2.7254267878224143e-05, 'epoch': 2.16}) +[2026-06-16 05:25:54,078][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.03732183575630188, 'train/grad_norm_pre_clip': 7.284614562988281, 'timing/data': 0.01354687288403511, 'timing/model': 1.4489334030076861, 'learning_rate/qwen_vl_interface': 2.7228476365521524e-06, 'learning_rate/action_model': 2.722847636552152e-05, 'epoch': 2.16}) +[2026-06-16 05:26:00,262][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.20129308104515076, 'train/grad_norm_pre_clip': 10.073156356811523, 'timing/data': 0.012263289652764797, 'timing/model': 1.4433958632871509, 'learning_rate/qwen_vl_interface': 2.7202695240987674e-06, 'learning_rate/action_model': 2.7202695240987676e-05, 'epoch': 2.16}) +[2026-06-16 05:26:06,353][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.060421183705329895, 'train/grad_norm_pre_clip': 7.531831741333008, 'timing/data': 0.011165370233356953, 'timing/model': 1.4469205383211374, 'learning_rate/qwen_vl_interface': 2.717692451522025e-06, 'learning_rate/action_model': 2.7176924515220248e-05, 'epoch': 2.16}) +[2026-06-16 05:26:12,375][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.25170281529426575, 'train/grad_norm_pre_clip': 8.83585262298584, 'timing/data': 0.014011924155056477, 'timing/model': 1.4380364147946239, 'learning_rate/qwen_vl_interface': 2.7151164198812624e-06, 'learning_rate/action_model': 2.7151164198812622e-05, 'epoch': 2.16}) +[2026-06-16 05:26:18,506][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.2927124798297882, 'train/grad_norm_pre_clip': 18.9231014251709, 'timing/data': 0.014138895086944103, 'timing/model': 1.9030452221632004, 'learning_rate/qwen_vl_interface': 2.712541430235388e-06, 'learning_rate/action_model': 2.712541430235388e-05, 'epoch': 2.16}) +[2026-06-16 05:26:24,168][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.09351275116205215, 'train/grad_norm_pre_clip': 4.852308750152588, 'timing/data': 0.014044883660972118, 'timing/model': 1.4563094498589635, 'learning_rate/qwen_vl_interface': 2.7099674836428842e-06, 'learning_rate/action_model': 2.709967483642884e-05, 'epoch': 2.16}) +[2026-06-16 05:26:30,278][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.07771261036396027, 'train/grad_norm_pre_clip': 4.877287864685059, 'timing/data': 0.014036363922059536, 'timing/model': 1.464512586593628, 'learning_rate/qwen_vl_interface': 2.7073945811618018e-06, 'learning_rate/action_model': 2.7073945811618017e-05, 'epoch': 2.16}) +[2026-06-16 05:26:36,379][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.20812790095806122, 'train/grad_norm_pre_clip': 9.817219734191895, 'timing/data': 0.014236224815249443, 'timing/model': 1.4685149462893605, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 2.16}) +[2026-06-16 05:26:42,476][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.06060458719730377, 'train/grad_norm_pre_clip': 6.125305652618408, 'timing/data': 0.015574916265904903, 'timing/model': 1.4641589960083365, 'learning_rate/qwen_vl_interface': 2.7022519127639687e-06, 'learning_rate/action_model': 2.7022519127639682e-05, 'epoch': 2.16}) +[2026-06-16 05:26:48,491][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.27296653389930725, 'train/grad_norm_pre_clip': 13.34342098236084, 'timing/data': 0.01473728846758604, 'timing/model': 1.4245180683210492, 'learning_rate/qwen_vl_interface': 2.6996821489611713e-06, 'learning_rate/action_model': 2.699682148961171e-05, 'epoch': 2.17}) +[2026-06-16 05:26:54,502][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.15435084700584412, 'train/grad_norm_pre_clip': 14.200183868408203, 'timing/data': 0.014178813435137272, 'timing/model': 1.7837047157809138, 'learning_rate/qwen_vl_interface': 2.697113433497712e-06, 'learning_rate/action_model': 2.6971134334977115e-05, 'epoch': 2.17}) +[2026-06-16 05:27:00,166][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.07474949210882187, 'train/grad_norm_pre_clip': 8.442621231079102, 'timing/data': 0.014316102489829063, 'timing/model': 1.453895446844399, 'learning_rate/qwen_vl_interface': 2.6945457674294863e-06, 'learning_rate/action_model': 2.694545767429486e-05, 'epoch': 2.17}) +[2026-06-16 05:27:06,269][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.326206773519516, 'train/grad_norm_pre_clip': 10.697187423706055, 'timing/data': 0.015685563907027245, 'timing/model': 1.4508863799273968, 'learning_rate/qwen_vl_interface': 2.6919791518119676e-06, 'learning_rate/action_model': 2.6919791518119674e-05, 'epoch': 2.17}) +[2026-06-16 05:27:12,381][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.06697272509336472, 'train/grad_norm_pre_clip': 8.1032133102417, 'timing/data': 0.014439253136515617, 'timing/model': 1.4651701804250479, 'learning_rate/qwen_vl_interface': 2.6894135877001946e-06, 'learning_rate/action_model': 2.689413587700195e-05, 'epoch': 2.17}) +[2026-06-16 05:27:18,475][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 0.031221279874444008, 'train/grad_norm_pre_clip': 1.9250119924545288, 'timing/data': 0.01425740122795105, 'timing/model': 1.48306271340698, 'learning_rate/qwen_vl_interface': 2.686849076148775e-06, 'learning_rate/action_model': 2.6868490761487747e-05, 'epoch': 2.17}) +[2026-06-16 05:27:24,479][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.04003863409161568, 'train/grad_norm_pre_clip': 5.331015586853027, 'timing/data': 0.014544202946126461, 'timing/model': 1.448711252771318, 'learning_rate/qwen_vl_interface': 2.68428561821188e-06, 'learning_rate/action_model': 2.6842856182118802e-05, 'epoch': 2.17}) +[2026-06-16 05:27:30,539][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.10108726471662521, 'train/grad_norm_pre_clip': 5.448033332824707, 'timing/data': 0.014978655613958836, 'timing/model': 1.8225441919639707, 'learning_rate/qwen_vl_interface': 2.681723214943253e-06, 'learning_rate/action_model': 2.681723214943253e-05, 'epoch': 2.17}) +[2026-06-16 05:27:36,174][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 0.09183380007743835, 'train/grad_norm_pre_clip': 9.963386535644531, 'timing/data': 0.01455744169652462, 'timing/model': 1.463165933266282, 'learning_rate/qwen_vl_interface': 2.6791618673961996e-06, 'learning_rate/action_model': 2.6791618673961994e-05, 'epoch': 2.17}) +[2026-06-16 05:27:42,173][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.03762242570519447, 'train/grad_norm_pre_clip': 4.881141185760498, 'timing/data': 0.011368065141141415, 'timing/model': 1.4815611001104116, 'learning_rate/qwen_vl_interface': 2.676601576623594e-06, 'learning_rate/action_model': 2.6766015766235937e-05, 'epoch': 2.17}) +[2026-06-16 05:27:48,169][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.10763297230005264, 'train/grad_norm_pre_clip': 7.453792572021484, 'timing/data': 0.01174151711165905, 'timing/model': 1.4672907339408994, 'learning_rate/qwen_vl_interface': 2.674042343677874e-06, 'learning_rate/action_model': 2.6740423436778738e-05, 'epoch': 2.17}) +[2026-06-16 05:27:54,031][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.21863152086734772, 'train/grad_norm_pre_clip': 11.372954368591309, 'timing/data': 0.011406185105443, 'timing/model': 1.449622762389481, 'learning_rate/qwen_vl_interface': 2.6714841696110456e-06, 'learning_rate/action_model': 2.6714841696110456e-05, 'epoch': 2.17}) +[2026-06-16 05:27:59,942][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.042605381458997726, 'train/grad_norm_pre_clip': 7.384608745574951, 'timing/data': 0.011054351925849915, 'timing/model': 1.751761132851243, 'learning_rate/qwen_vl_interface': 2.6689270554746743e-06, 'learning_rate/action_model': 2.6689270554746743e-05, 'epoch': 2.17}) +[2026-06-16 05:28:05,558][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.20094110071659088, 'train/grad_norm_pre_clip': 11.792835235595703, 'timing/data': 0.010902219451963902, 'timing/model': 1.4629807863384485, 'learning_rate/qwen_vl_interface': 2.6663710023198956e-06, 'learning_rate/action_model': 2.6663710023198956e-05, 'epoch': 2.17}) +[2026-06-16 05:28:11,470][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 0.14134617149829865, 'train/grad_norm_pre_clip': 9.869688987731934, 'timing/data': 0.011720295995473862, 'timing/model': 1.4773527411743999, 'learning_rate/qwen_vl_interface': 2.663816011197405e-06, 'learning_rate/action_model': 2.6638160111974048e-05, 'epoch': 2.17}) +[2026-06-16 05:28:17,254][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.08869454264640808, 'train/grad_norm_pre_clip': 8.666281700134277, 'timing/data': 0.011637375690042973, 'timing/model': 1.4538527261465788, 'learning_rate/qwen_vl_interface': 2.661262083157463e-06, 'learning_rate/action_model': 2.661262083157463e-05, 'epoch': 2.17}) +[2026-06-16 05:28:23,070][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.08243554830551147, 'train/grad_norm_pre_clip': 8.562027931213379, 'timing/data': 0.012364760972559452, 'timing/model': 1.4782219240441918, 'learning_rate/qwen_vl_interface': 2.6587092192498943e-06, 'learning_rate/action_model': 2.658709219249894e-05, 'epoch': 2.17}) +[2026-06-16 05:28:28,859][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.10986510664224625, 'train/grad_norm_pre_clip': 6.873732566833496, 'timing/data': 0.012290599755942822, 'timing/model': 1.4238378396257758, 'learning_rate/qwen_vl_interface': 2.656157420524084e-06, 'learning_rate/action_model': 2.6561574205240842e-05, 'epoch': 2.17}) +[2026-06-16 05:28:34,728][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.03348812088370323, 'train/grad_norm_pre_clip': 3.269789695739746, 'timing/data': 0.011296511627733707, 'timing/model': 1.7234050696715713, 'learning_rate/qwen_vl_interface': 2.6536066880289795e-06, 'learning_rate/action_model': 2.6536066880289794e-05, 'epoch': 2.17}) +[2026-06-16 05:28:40,367][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.09989528357982635, 'train/grad_norm_pre_clip': 5.335752487182617, 'timing/data': 0.010965499095618725, 'timing/model': 1.4710151012986898, 'learning_rate/qwen_vl_interface': 2.6510570228130916e-06, 'learning_rate/action_model': 2.6510570228130914e-05, 'epoch': 2.17}) +[2026-06-16 05:28:46,254][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.1296522170305252, 'train/grad_norm_pre_clip': 8.294394493103027, 'timing/data': 0.011610453017055988, 'timing/model': 1.4624284729361534, 'learning_rate/qwen_vl_interface': 2.648508425924491e-06, 'learning_rate/action_model': 2.6485084259244907e-05, 'epoch': 2.17}) +[2026-06-16 05:28:52,056][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.0645938515663147, 'train/grad_norm_pre_clip': 8.503602027893066, 'timing/data': 0.011002068407833576, 'timing/model': 1.4486734131351113, 'learning_rate/qwen_vl_interface': 2.6459608984108093e-06, 'learning_rate/action_model': 2.645960898410809e-05, 'epoch': 2.17}) +[2026-06-16 05:28:57,865][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.027053307741880417, 'train/grad_norm_pre_clip': 4.0280022621154785, 'timing/data': 0.011234519071877003, 'timing/model': 1.4519703714177012, 'learning_rate/qwen_vl_interface': 2.6434144413192414e-06, 'learning_rate/action_model': 2.6434144413192414e-05, 'epoch': 2.18}) +[2026-06-16 05:29:03,655][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.20764774084091187, 'train/grad_norm_pre_clip': 10.742575645446777, 'timing/data': 0.012164476327598095, 'timing/model': 1.414506372064352, 'learning_rate/qwen_vl_interface': 2.6408690556965347e-06, 'learning_rate/action_model': 2.640869055696535e-05, 'epoch': 2.18}) +[2026-06-16 05:29:09,540][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.03420832008123398, 'train/grad_norm_pre_clip': 2.4433438777923584, 'timing/data': 0.012108496390283108, 'timing/model': 1.7419575098901987, 'learning_rate/qwen_vl_interface': 2.6383247425890047e-06, 'learning_rate/action_model': 2.6383247425890047e-05, 'epoch': 2.18}) +[2026-06-16 05:29:15,170][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.024186599999666214, 'train/grad_norm_pre_clip': 1.5696935653686523, 'timing/data': 0.011208008974790573, 'timing/model': 1.4628747729584575, 'learning_rate/qwen_vl_interface': 2.6357815030425204e-06, 'learning_rate/action_model': 2.6357815030425205e-05, 'epoch': 2.18}) +[2026-06-16 05:29:21,071][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.23791559040546417, 'train/grad_norm_pre_clip': 15.045019149780273, 'timing/data': 0.011393480002880096, 'timing/model': 1.4599856734275818, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 2.18}) +[2026-06-16 05:29:26,873][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.19722303748130798, 'train/grad_norm_pre_clip': 15.843360900878906, 'timing/data': 0.011186477728188038, 'timing/model': 1.4665427170693874, 'learning_rate/qwen_vl_interface': 2.6306982488139704e-06, 'learning_rate/action_model': 2.6306982488139702e-05, 'epoch': 2.18}) +[2026-06-16 05:29:32,783][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 0.0030644594226032495, 'train/grad_norm_pre_clip': 0.33758071064949036, 'timing/data': 0.010848415084183216, 'timing/model': 1.4031013818457723, 'learning_rate/qwen_vl_interface': 2.628158236221438e-06, 'learning_rate/action_model': 2.628158236221438e-05, 'epoch': 2.18}) +[2026-06-16 05:29:38,574][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.20676910877227783, 'train/grad_norm_pre_clip': 11.232961654663086, 'timing/data': 0.011247768066823483, 'timing/model': 1.4024534719064832, 'learning_rate/qwen_vl_interface': 2.6256193013690183e-06, 'learning_rate/action_model': 2.6256193013690182e-05, 'epoch': 2.18}) +[2026-06-16 05:29:44,544][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.08034640550613403, 'train/grad_norm_pre_clip': 5.303855895996094, 'timing/data': 0.011332027614116669, 'timing/model': 1.7379753151908517, 'learning_rate/qwen_vl_interface': 2.623081445300373e-06, 'learning_rate/action_model': 2.6230814453003728e-05, 'epoch': 2.18}) +[2026-06-16 05:29:50,164][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.2266145944595337, 'train/grad_norm_pre_clip': 19.653114318847656, 'timing/data': 0.010775663889944553, 'timing/model': 1.463196356780827, 'learning_rate/qwen_vl_interface': 2.6205446690587183e-06, 'learning_rate/action_model': 2.6205446690587182e-05, 'epoch': 2.18}) +[2026-06-16 05:29:56,065][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.248885840177536, 'train/grad_norm_pre_clip': 8.534092903137207, 'timing/data': 0.011778111569583416, 'timing/model': 1.4556439435109496, 'learning_rate/qwen_vl_interface': 2.618008973686828e-06, 'learning_rate/action_model': 2.618008973686828e-05, 'epoch': 2.18}) +[2026-06-16 05:30:01,984][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 0.06285643577575684, 'train/grad_norm_pre_clip': 6.200809478759766, 'timing/data': 0.011265766806900501, 'timing/model': 1.4650883805006742, 'learning_rate/qwen_vl_interface': 2.615474360227028e-06, 'learning_rate/action_model': 2.615474360227028e-05, 'epoch': 2.18}) +[2026-06-16 05:30:07,765][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.11275321990251541, 'train/grad_norm_pre_clip': 8.164749145507812, 'timing/data': 0.010995275340974331, 'timing/model': 1.4445293480530381, 'learning_rate/qwen_vl_interface': 2.612940829721204e-06, 'learning_rate/action_model': 2.612940829721204e-05, 'epoch': 2.18}) +[2026-06-16 05:30:13,577][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 0.029995867982506752, 'train/grad_norm_pre_clip': 6.239254951477051, 'timing/data': 0.0117121497169137, 'timing/model': 1.4113293401896954, 'learning_rate/qwen_vl_interface': 2.6104083832107948e-06, 'learning_rate/action_model': 2.6104083832107945e-05, 'epoch': 2.18}) +[2026-06-16 05:30:19,535][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.17913977801799774, 'train/grad_norm_pre_clip': 9.350093841552734, 'timing/data': 0.011065305210649967, 'timing/model': 1.731415488757193, 'learning_rate/qwen_vl_interface': 2.607877021736793e-06, 'learning_rate/action_model': 2.6078770217367926e-05, 'epoch': 2.18}) +[2026-06-16 05:30:25,173][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.12654688954353333, 'train/grad_norm_pre_clip': 9.084607124328613, 'timing/data': 0.01097900327295065, 'timing/model': 1.4650394665077329, 'learning_rate/qwen_vl_interface': 2.605346746339745e-06, 'learning_rate/action_model': 2.605346746339745e-05, 'epoch': 2.18}) +[2026-06-16 05:30:31,074][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 0.03755125030875206, 'train/grad_norm_pre_clip': 4.3856611251831055, 'timing/data': 0.01139040570706129, 'timing/model': 1.4663900835439563, 'learning_rate/qwen_vl_interface': 2.602817558059753e-06, 'learning_rate/action_model': 2.6028175580597532e-05, 'epoch': 2.18}) +[2026-06-16 05:30:36,968][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.07654357701539993, 'train/grad_norm_pre_clip': 6.67183780670166, 'timing/data': 0.010477740317583084, 'timing/model': 1.4684822773560882, 'learning_rate/qwen_vl_interface': 2.60028945793647e-06, 'learning_rate/action_model': 2.60028945793647e-05, 'epoch': 2.18}) +[2026-06-16 05:30:42,768][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.1920166164636612, 'train/grad_norm_pre_clip': 44.78642654418945, 'timing/data': 0.010844391770660877, 'timing/model': 1.4755218382924795, 'learning_rate/qwen_vl_interface': 2.5977624470091024e-06, 'learning_rate/action_model': 2.5977624470091023e-05, 'epoch': 2.18}) +[2026-06-16 05:30:48,677][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.025344807654619217, 'train/grad_norm_pre_clip': 2.2747445106506348, 'timing/data': 0.012276743538677692, 'timing/model': 1.4912977637723088, 'learning_rate/qwen_vl_interface': 2.59523652631641e-06, 'learning_rate/action_model': 2.5952365263164097e-05, 'epoch': 2.18}) +[2026-06-16 05:30:54,650][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.1590798795223236, 'train/grad_norm_pre_clip': 11.781872749328613, 'timing/data': 0.011412146501243114, 'timing/model': 1.7422125395387411, 'learning_rate/qwen_vl_interface': 2.5927116968967024e-06, 'learning_rate/action_model': 2.5927116968967025e-05, 'epoch': 2.18}) +[2026-06-16 05:31:00,281][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.16321152448654175, 'train/grad_norm_pre_clip': 7.970324993133545, 'timing/data': 0.011221203953027725, 'timing/model': 1.4816289488226175, 'learning_rate/qwen_vl_interface': 2.5901879597878417e-06, 'learning_rate/action_model': 2.5901879597878415e-05, 'epoch': 2.18}) +[2026-06-16 05:31:06,168][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.028683533892035484, 'train/grad_norm_pre_clip': 2.1393377780914307, 'timing/data': 0.011034722439944744, 'timing/model': 1.4654406355693936, 'learning_rate/qwen_vl_interface': 2.5876653160272395e-06, 'learning_rate/action_model': 2.5876653160272397e-05, 'epoch': 2.19}) +[2026-06-16 05:31:12,070][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.05182952433824539, 'train/grad_norm_pre_clip': 11.25522232055664, 'timing/data': 0.010999321937561035, 'timing/model': 1.463085069321096, 'learning_rate/qwen_vl_interface': 2.5851437666518607e-06, 'learning_rate/action_model': 2.5851437666518607e-05, 'epoch': 2.19}) +[2026-06-16 05:31:17,871][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.20297807455062866, 'train/grad_norm_pre_clip': 16.712385177612305, 'timing/data': 0.011299343779683113, 'timing/model': 1.4705000333487988, 'learning_rate/qwen_vl_interface': 2.5826233126982198e-06, 'learning_rate/action_model': 2.5826233126982196e-05, 'epoch': 2.19}) +[2026-06-16 05:31:23,801][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 0.024417635053396225, 'train/grad_norm_pre_clip': 6.275068283081055, 'timing/data': 0.010859060101211071, 'timing/model': 1.4270542608574033, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 2.19}) +[2026-06-16 05:31:29,709][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.3461911976337433, 'train/grad_norm_pre_clip': 13.506693840026855, 'timing/data': 0.010115304961800575, 'timing/model': 1.7076986795291305, 'learning_rate/qwen_vl_interface': 2.577585695199954e-06, 'learning_rate/action_model': 2.5775856951999538e-05, 'epoch': 2.19}) +[2026-06-16 05:31:35,248][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.021993769332766533, 'train/grad_norm_pre_clip': 6.230998992919922, 'timing/data': 0.010088504292070866, 'timing/model': 1.4633142435923219, 'learning_rate/qwen_vl_interface': 2.575068533726101e-06, 'learning_rate/action_model': 2.575068533726101e-05, 'epoch': 2.19}) +[2026-06-16 05:31:41,049][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.059822436422109604, 'train/grad_norm_pre_clip': 7.170923709869385, 'timing/data': 0.010035904124379158, 'timing/model': 1.4458696572110057, 'learning_rate/qwen_vl_interface': 2.5725524718155333e-06, 'learning_rate/action_model': 2.572552471815533e-05, 'epoch': 2.19}) +[2026-06-16 05:31:46,752][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 0.0331295020878315, 'train/grad_norm_pre_clip': 3.0165562629699707, 'timing/data': 0.011201601475477219, 'timing/model': 1.4160086195915937, 'learning_rate/qwen_vl_interface': 2.570037510502508e-06, 'learning_rate/action_model': 2.5700375105025083e-05, 'epoch': 2.19}) +[2026-06-16 05:31:52,532][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.09857736527919769, 'train/grad_norm_pre_clip': 16.833707809448242, 'timing/data': 0.00981037225574255, 'timing/model': 1.4101150389760733, 'learning_rate/qwen_vl_interface': 2.5675236508208313e-06, 'learning_rate/action_model': 2.5675236508208313e-05, 'epoch': 2.19}) +[2026-06-16 05:31:58,365][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.11696721613407135, 'train/grad_norm_pre_clip': 10.678204536437988, 'timing/data': 0.011024110950529575, 'timing/model': 1.403227562084794, 'learning_rate/qwen_vl_interface': 2.565010893803856e-06, 'learning_rate/action_model': 2.565010893803856e-05, 'epoch': 2.19}) +[2026-06-16 05:32:04,217][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.30643582344055176, 'train/grad_norm_pre_clip': 17.506484985351562, 'timing/data': 0.010441765189170837, 'timing/model': 1.7223789636045694, 'learning_rate/qwen_vl_interface': 2.562499240484482e-06, 'learning_rate/action_model': 2.5624992404844816e-05, 'epoch': 2.19}) +[2026-06-16 05:32:10,058][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.11625360697507858, 'train/grad_norm_pre_clip': 8.80177116394043, 'timing/data': 0.011807216331362724, 'timing/model': 1.6644429601728916, 'learning_rate/qwen_vl_interface': 2.559988691895155e-06, 'learning_rate/action_model': 2.559988691895155e-05, 'epoch': 2.19}) +[2026-06-16 05:32:15,657][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.20376233756542206, 'train/grad_norm_pre_clip': 12.262797355651855, 'timing/data': 0.011486672796308994, 'timing/model': 1.4620719831436872, 'learning_rate/qwen_vl_interface': 2.5574792490678648e-06, 'learning_rate/action_model': 2.5574792490678644e-05, 'epoch': 2.19}) +[2026-06-16 05:32:21,453][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.055516038089990616, 'train/grad_norm_pre_clip': 8.65584945678711, 'timing/data': 0.011730705387890339, 'timing/model': 1.457212770357728, 'learning_rate/qwen_vl_interface': 2.5549709130341493e-06, 'learning_rate/action_model': 2.5549709130341493e-05, 'epoch': 2.19}) +[2026-06-16 05:32:27,261][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.05141493305563927, 'train/grad_norm_pre_clip': 4.869366645812988, 'timing/data': 0.011329841800034046, 'timing/model': 1.4423871887847781, 'learning_rate/qwen_vl_interface': 2.5524636848250924e-06, 'learning_rate/action_model': 2.552463684825092e-05, 'epoch': 2.19}) +[2026-06-16 05:32:33,080][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.027315765619277954, 'train/grad_norm_pre_clip': 7.8680524826049805, 'timing/data': 0.01117452047765255, 'timing/model': 1.4033758221194148, 'learning_rate/qwen_vl_interface': 2.5499575654713186e-06, 'learning_rate/action_model': 2.5499575654713185e-05, 'epoch': 2.19}) +[2026-06-16 05:32:39,013][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.029190121218562126, 'train/grad_norm_pre_clip': 4.1018500328063965, 'timing/data': 0.01094482745975256, 'timing/model': 1.7188030537217855, 'learning_rate/qwen_vl_interface': 2.547452556003001e-06, 'learning_rate/action_model': 2.547452556003001e-05, 'epoch': 2.19}) +[2026-06-16 05:32:44,927][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.032868705689907074, 'train/grad_norm_pre_clip': 5.021576881408691, 'timing/data': 0.011184420436620712, 'timing/model': 1.720247933641076, 'learning_rate/qwen_vl_interface': 2.5449486574498535e-06, 'learning_rate/action_model': 2.5449486574498534e-05, 'epoch': 2.19}) +[2026-06-16 05:32:50,562][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.09739100933074951, 'train/grad_norm_pre_clip': 14.562357902526855, 'timing/data': 0.011161820031702518, 'timing/model': 1.4672570461407304, 'learning_rate/qwen_vl_interface': 2.5424458708411335e-06, 'learning_rate/action_model': 2.5424458708411335e-05, 'epoch': 2.19}) +[2026-06-16 05:32:56,359][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 0.08970800042152405, 'train/grad_norm_pre_clip': 5.268448829650879, 'timing/data': 0.01058533601462841, 'timing/model': 1.4521622536703944, 'learning_rate/qwen_vl_interface': 2.5399441972056437e-06, 'learning_rate/action_model': 2.5399441972056438e-05, 'epoch': 2.19}) +[2026-06-16 05:33:02,157][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.2706262171268463, 'train/grad_norm_pre_clip': 10.322782516479492, 'timing/data': 0.011126290075480938, 'timing/model': 1.4453095002099872, 'learning_rate/qwen_vl_interface': 2.537443637571726e-06, 'learning_rate/action_model': 2.537443637571726e-05, 'epoch': 2.19}) +[2026-06-16 05:33:07,974][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.163553848862648, 'train/grad_norm_pre_clip': 6.6568121910095215, 'timing/data': 0.011623633094131947, 'timing/model': 1.409085820429027, 'learning_rate/qwen_vl_interface': 2.534944192967269e-06, 'learning_rate/action_model': 2.534944192967269e-05, 'epoch': 2.19}) +[2026-06-16 05:43:46,785][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.10861094295978546, 'train/grad_norm_pre_clip': 8.682350158691406, 'eval/loss': 0.06073666736483574, 'eval/flap_precision': 0.8280701754385965, 'eval/flap_recall': 0.8581818181818182, 'eval/flap_f1': 0.8428571428571427, 'eval/flap_rate_model': 0.07916666666666666, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.8428571428571427, 'timing/data': 0.011666092090308666, 'timing/model': 1.7396839689463377, 'learning_rate/qwen_vl_interface': 2.532445864419695e-06, 'learning_rate/action_model': 2.5324458644196952e-05, 'epoch': 2.2}) +[2026-06-16 05:43:46,787][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4800_state +[2026-06-16 05:43:46,796][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 05:43:55,079][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4800_state/model.safetensors +[2026-06-16 05:44:06,335][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4800_state/optimizer.bin +[2026-06-16 05:44:06,337][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4800_state/sampler.bin +[2026-06-16 05:44:06,337][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4800_state/sampler_1.bin +[2026-06-16 05:44:06,339][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4800_state/random_states_0.pkl +[2026-06-16 05:44:06,493][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4800 +[2026-06-16 05:44:06,493][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 05:44:06,509][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 05:44:12,274][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 0.07345959544181824, 'train/grad_norm_pre_clip': 6.849292278289795, 'timing/data': 0.011805921792984009, 'timing/model': 1.6960181863978505, 'learning_rate/qwen_vl_interface': 2.5299486529559758e-06, 'learning_rate/action_model': 2.5299486529559756e-05, 'epoch': 2.2}) +[2026-06-16 05:44:18,011][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.09888134896755219, 'train/grad_norm_pre_clip': 12.779532432556152, 'timing/data': 0.011365738697350025, 'timing/model': 1.5450295759364963, 'learning_rate/qwen_vl_interface': 2.5274525596026202e-06, 'learning_rate/action_model': 2.52745255960262e-05, 'epoch': 2.2}) +[2026-06-16 05:44:23,537][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.15664701163768768, 'train/grad_norm_pre_clip': 13.39621353149414, 'timing/data': 0.011197567917406559, 'timing/model': 1.456602644175291, 'learning_rate/qwen_vl_interface': 2.524957585385677e-06, 'learning_rate/action_model': 2.524957585385677e-05, 'epoch': 2.2}) +[2026-06-16 05:44:29,329][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.02925029955804348, 'train/grad_norm_pre_clip': 3.4633870124816895, 'timing/data': 0.012787867337465286, 'timing/model': 1.4526288146153092, 'learning_rate/qwen_vl_interface': 2.522463731330738e-06, 'learning_rate/action_model': 2.5224637313307374e-05, 'epoch': 2.2}) +[2026-06-16 05:44:34,950][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.06643897294998169, 'train/grad_norm_pre_clip': 5.354236602783203, 'timing/data': 0.01062789373099804, 'timing/model': 1.3771530529484153, 'learning_rate/qwen_vl_interface': 2.5199709984629294e-06, 'learning_rate/action_model': 2.5199709984629292e-05, 'epoch': 2.2}) +[2026-06-16 05:44:40,751][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 0.12074865400791168, 'train/grad_norm_pre_clip': 5.104474067687988, 'timing/data': 0.011553100310266018, 'timing/model': 1.4293179418891668, 'learning_rate/qwen_vl_interface': 2.517479387806922e-06, 'learning_rate/action_model': 2.517479387806922e-05, 'epoch': 2.2}) +[2026-06-16 05:44:46,543][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.20287080109119415, 'train/grad_norm_pre_clip': 10.74084186553955, 'timing/data': 0.011613840237259865, 'timing/model': 1.4563578283414245, 'learning_rate/qwen_vl_interface': 2.5149889003869217e-06, 'learning_rate/action_model': 2.5149889003869214e-05, 'epoch': 2.2}) +[2026-06-16 05:44:52,414][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.05475728213787079, 'train/grad_norm_pre_clip': 6.1915388107299805, 'timing/data': 0.010905774310231209, 'timing/model': 1.7277376875281334, 'learning_rate/qwen_vl_interface': 2.5124995372266736e-06, 'learning_rate/action_model': 2.5124995372266735e-05, 'epoch': 2.2}) +[2026-06-16 05:44:57,938][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.015976564958691597, 'train/grad_norm_pre_clip': 2.1000967025756836, 'timing/data': 0.01604230608791113, 'timing/model': 1.4354781610891223, 'learning_rate/qwen_vl_interface': 2.5100112993494626e-06, 'learning_rate/action_model': 2.5100112993494624e-05, 'epoch': 2.2}) +[2026-06-16 05:45:03,665][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.1275150626897812, 'train/grad_norm_pre_clip': 8.718168258666992, 'timing/data': 0.011987731792032719, 'timing/model': 1.3879108289256692, 'learning_rate/qwen_vl_interface': 2.5075241877781085e-06, 'learning_rate/action_model': 2.5075241877781086e-05, 'epoch': 2.2}) +[2026-06-16 05:45:09,559][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.11405511945486069, 'train/grad_norm_pre_clip': 10.970724105834961, 'timing/data': 0.011943740770220757, 'timing/model': 1.4819591892883182, 'learning_rate/qwen_vl_interface': 2.5050382035349673e-06, 'learning_rate/action_model': 2.5050382035349674e-05, 'epoch': 2.2}) +[2026-06-16 05:45:15,359][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.10896974802017212, 'train/grad_norm_pre_clip': 16.601972579956055, 'timing/data': 0.010689475573599339, 'timing/model': 1.4006179347634315, 'learning_rate/qwen_vl_interface': 2.5025533476419343e-06, 'learning_rate/action_model': 2.5025533476419343e-05, 'epoch': 2.2}) +[2026-06-16 05:45:21,098][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.1731370985507965, 'train/grad_norm_pre_clip': 8.81832504272461, 'timing/data': 0.011610560119152069, 'timing/model': 1.6793944826349616, 'learning_rate/qwen_vl_interface': 2.5000696211204394e-06, 'learning_rate/action_model': 2.5000696211204394e-05, 'epoch': 2.2}) +[2026-06-16 05:45:26,916][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.06240101158618927, 'train/grad_norm_pre_clip': 3.7949111461639404, 'timing/data': 0.011096936650574207, 'timing/model': 1.7314047683030367, 'learning_rate/qwen_vl_interface': 2.4975870249914495e-06, 'learning_rate/action_model': 2.4975870249914497e-05, 'epoch': 2.2}) +[2026-06-16 05:45:32,453][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.14907138049602509, 'train/grad_norm_pre_clip': 9.726811408996582, 'timing/data': 0.010388942435383797, 'timing/model': 1.4739001486450434, 'learning_rate/qwen_vl_interface': 2.4951055602754664e-06, 'learning_rate/action_model': 2.4951055602754665e-05, 'epoch': 2.2}) +[2026-06-16 05:45:38,245][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.10185730457305908, 'train/grad_norm_pre_clip': 9.160892486572266, 'timing/data': 0.011041935533285141, 'timing/model': 1.459018393419683, 'learning_rate/qwen_vl_interface': 2.4926252279925255e-06, 'learning_rate/action_model': 2.4926252279925254e-05, 'epoch': 2.2}) +[2026-06-16 05:45:44,050][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 0.11567400395870209, 'train/grad_norm_pre_clip': 5.792262077331543, 'timing/data': 0.010814415290951729, 'timing/model': 1.453527675010264, 'learning_rate/qwen_vl_interface': 2.490146029162199e-06, 'learning_rate/action_model': 2.4901460291621988e-05, 'epoch': 2.2}) +[2026-06-16 05:45:49,845][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.03896696865558624, 'train/grad_norm_pre_clip': 7.436528205871582, 'timing/data': 0.010439901612699032, 'timing/model': 1.391000154428184, 'learning_rate/qwen_vl_interface': 2.4876679648035905e-06, 'learning_rate/action_model': 2.4876679648035903e-05, 'epoch': 2.2}) +[2026-06-16 05:45:55,596][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.10337162017822266, 'train/grad_norm_pre_clip': 6.6764678955078125, 'timing/data': 0.010898225009441376, 'timing/model': 1.7029138561338186, 'learning_rate/qwen_vl_interface': 2.4851910359353392e-06, 'learning_rate/action_model': 2.485191035935339e-05, 'epoch': 2.2}) +[2026-06-16 05:46:01,328][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.1032511442899704, 'train/grad_norm_pre_clip': 11.431078910827637, 'timing/data': 0.01111125759780407, 'timing/model': 1.5995444525033236, 'learning_rate/qwen_vl_interface': 2.4827152435756185e-06, 'learning_rate/action_model': 2.4827152435756185e-05, 'epoch': 2.2}) +[2026-06-16 05:46:06,950][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.19545450806617737, 'train/grad_norm_pre_clip': 10.960708618164062, 'timing/data': 0.010435042902827263, 'timing/model': 1.447199845686555, 'learning_rate/qwen_vl_interface': 2.4802405887421307e-06, 'learning_rate/action_model': 2.4802405887421306e-05, 'epoch': 2.2}) +[2026-06-16 05:46:12,947][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.02758936956524849, 'train/grad_norm_pre_clip': 6.544897079467773, 'timing/data': 0.010964096523821354, 'timing/model': 1.4603243488818407, 'learning_rate/qwen_vl_interface': 2.4777670724521134e-06, 'learning_rate/action_model': 2.4777670724521134e-05, 'epoch': 2.21}) +[2026-06-16 05:46:18,958][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.08647526800632477, 'train/grad_norm_pre_clip': 4.816612720489502, 'timing/data': 0.012036842294037342, 'timing/model': 1.4573148358613253, 'learning_rate/qwen_vl_interface': 2.475294695722337e-06, 'learning_rate/action_model': 2.475294695722337e-05, 'epoch': 2.21}) +[2026-06-16 05:46:24,706][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.05679873377084732, 'train/grad_norm_pre_clip': 3.8495609760284424, 'timing/data': 0.010813274420797825, 'timing/model': 1.4293097537010908, 'learning_rate/qwen_vl_interface': 2.4728234595691006e-06, 'learning_rate/action_model': 2.4728234595691005e-05, 'epoch': 2.21}) +[2026-06-16 05:46:30,615][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.08597612380981445, 'train/grad_norm_pre_clip': 5.1110639572143555, 'timing/data': 0.01112009584903717, 'timing/model': 1.6826041797176003, 'learning_rate/qwen_vl_interface': 2.4703533650082368e-06, 'learning_rate/action_model': 2.4703533650082366e-05, 'epoch': 2.21}) +[2026-06-16 05:46:36,337][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.09563573449850082, 'train/grad_norm_pre_clip': 6.025356292724609, 'timing/data': 0.010574433021247387, 'timing/model': 1.6409155447036028, 'learning_rate/qwen_vl_interface': 2.4678844130551103e-06, 'learning_rate/action_model': 2.4678844130551103e-05, 'epoch': 2.21}) +[2026-06-16 05:46:41,968][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.05563722178339958, 'train/grad_norm_pre_clip': 7.247579574584961, 'timing/data': 0.011195127852261066, 'timing/model': 1.4630814203992486, 'learning_rate/qwen_vl_interface': 2.4654166047246102e-06, 'learning_rate/action_model': 2.46541660472461e-05, 'epoch': 2.21}) +[2026-06-16 05:46:47,858][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.16890408098697662, 'train/grad_norm_pre_clip': 13.79158878326416, 'timing/data': 0.010875904932618141, 'timing/model': 1.4580970760434866, 'learning_rate/qwen_vl_interface': 2.4629499410311606e-06, 'learning_rate/action_model': 2.462949941031161e-05, 'epoch': 2.21}) +[2026-06-16 05:46:53,664][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.0209504347294569, 'train/grad_norm_pre_clip': 1.4866310358047485, 'timing/data': 0.011322207748889923, 'timing/model': 1.4692067271098495, 'learning_rate/qwen_vl_interface': 2.460484422988715e-06, 'learning_rate/action_model': 2.4604844229887146e-05, 'epoch': 2.21}) +[2026-06-16 05:46:59,461][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.021402981132268906, 'train/grad_norm_pre_clip': 3.1051220893859863, 'timing/data': 0.010593933053314686, 'timing/model': 1.4092626459896564, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 2.21}) +[2026-06-16 05:47:05,307][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 0.08548010885715485, 'train/grad_norm_pre_clip': 10.571227073669434, 'timing/data': 0.011111815460026264, 'timing/model': 1.7123934533447027, 'learning_rate/qwen_vl_interface': 2.455556827910286e-06, 'learning_rate/action_model': 2.455556827910286e-05, 'epoch': 2.21}) +[2026-06-16 05:47:11,018][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.3099772036075592, 'train/grad_norm_pre_clip': 10.608023643493652, 'timing/data': 0.01499812863767147, 'timing/model': 1.4571159612387419, 'learning_rate/qwen_vl_interface': 2.453094752899853e-06, 'learning_rate/action_model': 2.453094752899853e-05, 'epoch': 2.21}) +[2026-06-16 05:47:17,306][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.24203507602214813, 'train/grad_norm_pre_clip': 38.42744064331055, 'timing/data': 0.015108389779925346, 'timing/model': 1.4581518927589059, 'learning_rate/qwen_vl_interface': 2.4506338275915173e-06, 'learning_rate/action_model': 2.4506338275915172e-05, 'epoch': 2.21}) +[2026-06-16 05:47:23,268][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.10836297273635864, 'train/grad_norm_pre_clip': 6.241029262542725, 'timing/data': 0.014722826890647411, 'timing/model': 1.4439623234793544, 'learning_rate/qwen_vl_interface': 2.4481740529968735e-06, 'learning_rate/action_model': 2.4481740529968734e-05, 'epoch': 2.21}) +[2026-06-16 05:47:29,270][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.023032296448946, 'train/grad_norm_pre_clip': 2.6348910331726074, 'timing/data': 0.015287761576473713, 'timing/model': 1.4634741451591253, 'learning_rate/qwen_vl_interface': 2.4457154301270427e-06, 'learning_rate/action_model': 2.4457154301270427e-05, 'epoch': 2.21}) +[2026-06-16 05:47:35,477][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 0.05557318776845932, 'train/grad_norm_pre_clip': 6.4325385093688965, 'timing/data': 0.016005543991923332, 'timing/model': 1.4450180819258094, 'learning_rate/qwen_vl_interface': 2.443257959992671e-06, 'learning_rate/action_model': 2.443257959992671e-05, 'epoch': 2.21}) +[2026-06-16 05:47:41,713][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.07426717877388, 'train/grad_norm_pre_clip': 6.199133396148682, 'timing/data': 0.014863278716802597, 'timing/model': 1.9634901564568281, 'learning_rate/qwen_vl_interface': 2.440801643603933e-06, 'learning_rate/action_model': 2.440801643603933e-05, 'epoch': 2.21}) +[2026-06-16 05:47:47,373][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.134291872382164, 'train/grad_norm_pre_clip': 9.397133827209473, 'timing/data': 0.014795877039432526, 'timing/model': 1.4409871492534876, 'learning_rate/qwen_vl_interface': 2.4383464819705252e-06, 'learning_rate/action_model': 2.438346481970525e-05, 'epoch': 2.21}) +[2026-06-16 05:47:54,097][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.14750562608242035, 'train/grad_norm_pre_clip': 6.307685375213623, 'timing/data': 0.04052380006760359, 'timing/model': 1.5205953940749168, 'learning_rate/qwen_vl_interface': 2.435892476101674e-06, 'learning_rate/action_model': 2.435892476101674e-05, 'epoch': 2.21}) +[2026-06-16 05:48:00,176][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.1220046803355217, 'train/grad_norm_pre_clip': 5.42883825302124, 'timing/data': 0.01227883342653513, 'timing/model': 1.4640532350167632, 'learning_rate/qwen_vl_interface': 2.433439627006128e-06, 'learning_rate/action_model': 2.4334396270061277e-05, 'epoch': 2.21}) +[2026-06-16 05:48:06,371][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 0.10826121270656586, 'train/grad_norm_pre_clip': 8.756718635559082, 'timing/data': 0.013563769869506359, 'timing/model': 1.4618521677330136, 'learning_rate/qwen_vl_interface': 2.4309879356921602e-06, 'learning_rate/action_model': 2.43098793569216e-05, 'epoch': 2.21}) +[2026-06-16 05:48:12,586][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.1674644500017166, 'train/grad_norm_pre_clip': 7.2711005210876465, 'timing/data': 0.013047606684267521, 'timing/model': 1.4648757120594382, 'learning_rate/qwen_vl_interface': 2.428537403167569e-06, 'learning_rate/action_model': 2.4285374031675687e-05, 'epoch': 2.21}) +[2026-06-16 05:48:18,724][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.021800916641950607, 'train/grad_norm_pre_clip': 1.5780155658721924, 'timing/data': 0.012364862486720085, 'timing/model': 1.907453560270369, 'learning_rate/qwen_vl_interface': 2.4260880304396765e-06, 'learning_rate/action_model': 2.4260880304396765e-05, 'epoch': 2.21}) +[2026-06-16 05:48:24,808][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.15792231261730194, 'train/grad_norm_pre_clip': 7.266982555389404, 'timing/data': 0.011659839190542698, 'timing/model': 1.8934547528624535, 'learning_rate/qwen_vl_interface': 2.4236398185153245e-06, 'learning_rate/action_model': 2.4236398185153245e-05, 'epoch': 2.22}) +[2026-06-16 05:48:30,483][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.09345249831676483, 'train/grad_norm_pre_clip': 7.233612060546875, 'timing/data': 0.012067831121385098, 'timing/model': 1.4745950568467379, 'learning_rate/qwen_vl_interface': 2.4211927684008823e-06, 'learning_rate/action_model': 2.4211927684008824e-05, 'epoch': 2.22}) +[2026-06-16 05:48:36,568][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.1326933652162552, 'train/grad_norm_pre_clip': 7.920910835266113, 'timing/data': 0.01210780069231987, 'timing/model': 1.4531283155083656, 'learning_rate/qwen_vl_interface': 2.4187468811022398e-06, 'learning_rate/action_model': 2.4187468811022396e-05, 'epoch': 2.22}) +[2026-06-16 05:48:42,683][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.21893098950386047, 'train/grad_norm_pre_clip': 10.3811674118042, 'timing/data': 0.013101298362016678, 'timing/model': 1.4639075445011258, 'learning_rate/qwen_vl_interface': 2.416302157624809e-06, 'learning_rate/action_model': 2.4163021576248088e-05, 'epoch': 2.22}) +[2026-06-16 05:48:48,776][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.28063714504241943, 'train/grad_norm_pre_clip': 11.92972183227539, 'timing/data': 0.012796374969184399, 'timing/model': 1.4310056483373046, 'learning_rate/qwen_vl_interface': 2.4138585989735227e-06, 'learning_rate/action_model': 2.4138585989735226e-05, 'epoch': 2.22}) +[2026-06-16 05:48:54,909][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.13023270666599274, 'train/grad_norm_pre_clip': 9.437810897827148, 'timing/data': 0.014087473042309284, 'timing/model': 1.8960785018280149, 'learning_rate/qwen_vl_interface': 2.4114162061528357e-06, 'learning_rate/action_model': 2.4114162061528354e-05, 'epoch': 2.22}) +[2026-06-16 05:49:00,873][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.15886230766773224, 'train/grad_norm_pre_clip': 9.2972412109375, 'timing/data': 0.013567830435931683, 'timing/model': 1.7663066228851676, 'learning_rate/qwen_vl_interface': 2.408974980166724e-06, 'learning_rate/action_model': 2.408974980166724e-05, 'epoch': 2.22}) +[2026-06-16 05:49:06,584][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.010749668814241886, 'train/grad_norm_pre_clip': 1.1381771564483643, 'timing/data': 0.012607493437826633, 'timing/model': 1.4630795056000352, 'learning_rate/qwen_vl_interface': 2.406534922018683e-06, 'learning_rate/action_model': 2.4065349220186827e-05, 'epoch': 2.22}) +[2026-06-16 05:49:12,669][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 0.03609172999858856, 'train/grad_norm_pre_clip': 5.18232536315918, 'timing/data': 0.013110537081956863, 'timing/model': 1.4620408853515983, 'learning_rate/qwen_vl_interface': 2.4040960327117287e-06, 'learning_rate/action_model': 2.4040960327117283e-05, 'epoch': 2.22}) +[2026-06-16 05:49:18,673][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 0.07152698934078217, 'train/grad_norm_pre_clip': 4.410144329071045, 'timing/data': 0.011762679554522038, 'timing/model': 1.4712725663557649, 'learning_rate/qwen_vl_interface': 2.4016583132483976e-06, 'learning_rate/action_model': 2.4016583132483976e-05, 'epoch': 2.22}) +[2026-06-16 05:49:24,774][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.07615655660629272, 'train/grad_norm_pre_clip': 7.156823635101318, 'timing/data': 0.012466823682188988, 'timing/model': 1.4335800604894757, 'learning_rate/qwen_vl_interface': 2.399221764630741e-06, 'learning_rate/action_model': 2.3992217646307406e-05, 'epoch': 2.22}) +[2026-06-16 05:49:30,894][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.1810515969991684, 'train/grad_norm_pre_clip': 13.069520950317383, 'timing/data': 0.012497643940150738, 'timing/model': 1.8830109164118767, 'learning_rate/qwen_vl_interface': 2.3967863878603334e-06, 'learning_rate/action_model': 2.3967863878603332e-05, 'epoch': 2.22}) +[2026-06-16 05:49:36,579][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.05206046625971794, 'train/grad_norm_pre_clip': 3.4830825328826904, 'timing/data': 0.012751595117151737, 'timing/model': 1.4625185243785381, 'learning_rate/qwen_vl_interface': 2.394352183938266e-06, 'learning_rate/action_model': 2.3943521839382664e-05, 'epoch': 2.22}) +[2026-06-16 05:49:42,692][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 0.044437993317842484, 'train/grad_norm_pre_clip': 5.415895462036133, 'timing/data': 0.012190472334623337, 'timing/model': 1.4743100889027119, 'learning_rate/qwen_vl_interface': 2.39191915386515e-06, 'learning_rate/action_model': 2.39191915386515e-05, 'epoch': 2.22}) +[2026-06-16 05:49:48,788][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.048183612525463104, 'train/grad_norm_pre_clip': 4.999986171722412, 'timing/data': 0.011918449774384499, 'timing/model': 1.4811575962230563, 'learning_rate/qwen_vl_interface': 2.3894872986411095e-06, 'learning_rate/action_model': 2.3894872986411094e-05, 'epoch': 2.22}) +[2026-06-16 05:49:54,876][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.006698186043649912, 'train/grad_norm_pre_clip': 0.49682459235191345, 'timing/data': 0.012471633031964302, 'timing/model': 1.4299411019310355, 'learning_rate/qwen_vl_interface': 2.3870566192657903e-06, 'learning_rate/action_model': 2.38705661926579e-05, 'epoch': 2.22}) +[2026-06-16 05:50:00,986][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.08830192685127258, 'train/grad_norm_pre_clip': 10.635287284851074, 'timing/data': 0.012341351248323917, 'timing/model': 1.4768730737268925, 'learning_rate/qwen_vl_interface': 2.3846271167383515e-06, 'learning_rate/action_model': 2.3846271167383515e-05, 'epoch': 2.22}) +[2026-06-16 05:50:07,065][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.07839950174093246, 'train/grad_norm_pre_clip': 6.196490287780762, 'timing/data': 0.012649353593587875, 'timing/model': 1.8522382080554962, 'learning_rate/qwen_vl_interface': 2.382198792057471e-06, 'learning_rate/action_model': 2.3821987920574707e-05, 'epoch': 2.22}) +[2026-06-16 05:50:12,665][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.11350107192993164, 'train/grad_norm_pre_clip': 6.569789409637451, 'timing/data': 0.012735825031995773, 'timing/model': 1.4358349265530705, 'learning_rate/qwen_vl_interface': 2.3797716462213402e-06, 'learning_rate/action_model': 2.3797716462213402e-05, 'epoch': 2.22}) +[2026-06-16 05:50:18,880][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 0.1046123132109642, 'train/grad_norm_pre_clip': 7.569396495819092, 'timing/data': 0.011395026929676533, 'timing/model': 1.4663471216335893, 'learning_rate/qwen_vl_interface': 2.37734568022767e-06, 'learning_rate/action_model': 2.3773456802276697e-05, 'epoch': 2.22}) +[2026-06-16 05:50:25,087][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.12342341244220734, 'train/grad_norm_pre_clip': 10.079957008361816, 'timing/data': 0.012239081785082817, 'timing/model': 1.475942425429821, 'learning_rate/qwen_vl_interface': 2.3749208950736783e-06, 'learning_rate/action_model': 2.374920895073678e-05, 'epoch': 2.22}) +[2026-06-16 05:50:31,204][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.04280497878789902, 'train/grad_norm_pre_clip': 6.2698822021484375, 'timing/data': 0.01249072328209877, 'timing/model': 1.4305908661335707, 'learning_rate/qwen_vl_interface': 2.372497291756106e-06, 'learning_rate/action_model': 2.3724972917561058e-05, 'epoch': 2.22}) +[2026-06-16 05:50:37,138][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.028137141838669777, 'train/grad_norm_pre_clip': 2.178100347518921, 'timing/data': 0.011966420337557793, 'timing/model': 1.4585242168977857, 'learning_rate/qwen_vl_interface': 2.3700748712712043e-06, 'learning_rate/action_model': 2.370074871271204e-05, 'epoch': 2.23}) +[2026-06-16 05:50:43,260][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.1981201469898224, 'train/grad_norm_pre_clip': 10.934467315673828, 'timing/data': 0.012018799781799316, 'timing/model': 1.942483571358025, 'learning_rate/qwen_vl_interface': 2.3676536346147388e-06, 'learning_rate/action_model': 2.3676536346147388e-05, 'epoch': 2.23}) +[2026-06-16 05:50:48,883][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.10625230520963669, 'train/grad_norm_pre_clip': 14.292996406555176, 'timing/data': 0.012608073651790619, 'timing/model': 1.4573977142572403, 'learning_rate/qwen_vl_interface': 2.365233582781988e-06, 'learning_rate/action_model': 2.365233582781988e-05, 'epoch': 2.23}) +[2026-06-16 05:50:55,074][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 0.14965176582336426, 'train/grad_norm_pre_clip': 12.622403144836426, 'timing/data': 0.011411195620894432, 'timing/model': 1.4624591898173094, 'learning_rate/qwen_vl_interface': 2.3628147167677444e-06, 'learning_rate/action_model': 2.3628147167677443e-05, 'epoch': 2.23}) +[2026-06-16 05:51:01,288][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.12369964271783829, 'train/grad_norm_pre_clip': 7.402040004730225, 'timing/data': 0.012083141133189201, 'timing/model': 1.4461161708459258, 'learning_rate/qwen_vl_interface': 2.3603970375663123e-06, 'learning_rate/action_model': 2.3603970375663125e-05, 'epoch': 2.23}) +[2026-06-16 05:51:07,396][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.04071943461894989, 'train/grad_norm_pre_clip': 4.095871448516846, 'timing/data': 0.013230906799435616, 'timing/model': 1.4227614123374224, 'learning_rate/qwen_vl_interface': 2.35798054617151e-06, 'learning_rate/action_model': 2.3579805461715096e-05, 'epoch': 2.23}) +[2026-06-16 05:51:13,597][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.02551773376762867, 'train/grad_norm_pre_clip': 5.6645941734313965, 'timing/data': 0.014148302376270294, 'timing/model': 1.996932229027152, 'learning_rate/qwen_vl_interface': 2.355565243576662e-06, 'learning_rate/action_model': 2.3555652435766617e-05, 'epoch': 2.23}) +[2026-06-16 05:51:19,283][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.06044631823897362, 'train/grad_norm_pre_clip': 5.825562477111816, 'timing/data': 0.012369791977107525, 'timing/model': 1.4780089585110545, 'learning_rate/qwen_vl_interface': 2.35315113077461e-06, 'learning_rate/action_model': 2.3531511307746098e-05, 'epoch': 2.23}) +[2026-06-16 05:51:25,385][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.04837541654706001, 'train/grad_norm_pre_clip': 5.634927272796631, 'timing/data': 0.012156031094491482, 'timing/model': 1.4596947384998202, 'learning_rate/qwen_vl_interface': 2.3507382087577047e-06, 'learning_rate/action_model': 2.350738208757705e-05, 'epoch': 2.23}) +[2026-06-16 05:51:31,482][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.06551466882228851, 'train/grad_norm_pre_clip': 6.072017192840576, 'timing/data': 0.012100701220333576, 'timing/model': 1.4680538335815072, 'learning_rate/qwen_vl_interface': 2.348326478517807e-06, 'learning_rate/action_model': 2.3483264785178072e-05, 'epoch': 2.23}) +[2026-06-16 05:51:37,590][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.023632925003767014, 'train/grad_norm_pre_clip': 3.0060298442840576, 'timing/data': 0.013433787971735, 'timing/model': 1.479873200878501, 'learning_rate/qwen_vl_interface': 2.3459159410462907e-06, 'learning_rate/action_model': 2.3459159410462905e-05, 'epoch': 2.23}) +[2026-06-16 05:51:43,698][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.011889081448316574, 'train/grad_norm_pre_clip': 3.4697420597076416, 'timing/data': 0.012435631826519966, 'timing/model': 1.4354994995519519, 'learning_rate/qwen_vl_interface': 2.343506597334031e-06, 'learning_rate/action_model': 2.343506597334031e-05, 'epoch': 2.23}) +[2026-06-16 05:51:49,872][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.00288239773362875, 'train/grad_norm_pre_clip': 0.4298052489757538, 'timing/data': 0.01193625945597887, 'timing/model': 1.960345922037959, 'learning_rate/qwen_vl_interface': 2.341098448371423e-06, 'learning_rate/action_model': 2.3410984483714227e-05, 'epoch': 2.23}) +[2026-06-16 05:51:55,579][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.11436571925878525, 'train/grad_norm_pre_clip': 6.975094795227051, 'timing/data': 0.012458032928407192, 'timing/model': 1.4664465840905905, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 2.23}) +[2026-06-16 05:52:01,677][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.12207339704036713, 'train/grad_norm_pre_clip': 8.232203483581543, 'timing/data': 0.012252960354089737, 'timing/model': 1.4653784660622478, 'learning_rate/qwen_vl_interface': 2.33628573865426e-06, 'learning_rate/action_model': 2.3362857386542596e-05, 'epoch': 2.23}) +[2026-06-16 05:52:07,781][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.09729732573032379, 'train/grad_norm_pre_clip': 9.897189140319824, 'timing/data': 0.01225426234304905, 'timing/model': 1.4621249921619892, 'learning_rate/qwen_vl_interface': 2.3338811798780282e-06, 'learning_rate/action_model': 2.333881179878028e-05, 'epoch': 2.23}) +[2026-06-16 05:52:13,788][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.1975782960653305, 'train/grad_norm_pre_clip': 13.650857925415039, 'timing/data': 0.012435912154614925, 'timing/model': 1.4685852890834212, 'learning_rate/qwen_vl_interface': 2.3314778198080933e-06, 'learning_rate/action_model': 2.331477819808093e-05, 'epoch': 2.23}) +[2026-06-16 05:52:19,901][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.05571724474430084, 'train/grad_norm_pre_clip': 8.533987045288086, 'timing/data': 0.012477663345634937, 'timing/model': 1.4574560709297657, 'learning_rate/qwen_vl_interface': 2.3290756594323834e-06, 'learning_rate/action_model': 2.3290756594323836e-05, 'epoch': 2.23}) +[2026-06-16 05:52:26,021][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.09800693392753601, 'train/grad_norm_pre_clip': 10.773197174072266, 'timing/data': 0.012020070105791092, 'timing/model': 1.9084931295365095, 'learning_rate/qwen_vl_interface': 2.3266746997383378e-06, 'learning_rate/action_model': 2.3266746997383377e-05, 'epoch': 2.23}) +[2026-06-16 05:52:31,670][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.15188831090927124, 'train/grad_norm_pre_clip': 10.420340538024902, 'timing/data': 0.014709444716572762, 'timing/model': 1.4690229631960392, 'learning_rate/qwen_vl_interface': 2.324274941712899e-06, 'learning_rate/action_model': 2.324274941712899e-05, 'epoch': 2.23}) +[2026-06-16 05:52:37,875][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.1967448890209198, 'train/grad_norm_pre_clip': 8.295910835266113, 'timing/data': 0.012534253299236298, 'timing/model': 1.4596109557896852, 'learning_rate/qwen_vl_interface': 2.321876386342519e-06, 'learning_rate/action_model': 2.3218763863425187e-05, 'epoch': 2.23}) +[2026-06-16 05:52:44,081][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.08728209882974625, 'train/grad_norm_pre_clip': 4.455186367034912, 'timing/data': 0.011986348778009415, 'timing/model': 1.4686252856627107, 'learning_rate/qwen_vl_interface': 2.31947903461315e-06, 'learning_rate/action_model': 2.3194790346131502e-05, 'epoch': 2.23}) +[2026-06-16 05:52:50,195][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.07328029721975327, 'train/grad_norm_pre_clip': 11.314515113830566, 'timing/data': 0.01257335301488638, 'timing/model': 1.4215746400877833, 'learning_rate/qwen_vl_interface': 2.3170828875102544e-06, 'learning_rate/action_model': 2.3170828875102543e-05, 'epoch': 2.24}) +[2026-06-16 05:52:56,352][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.11621596664190292, 'train/grad_norm_pre_clip': 7.933502197265625, 'timing/data': 0.01287949550896883, 'timing/model': 1.9196036579087377, 'learning_rate/qwen_vl_interface': 2.3146879460187984e-06, 'learning_rate/action_model': 2.3146879460187982e-05, 'epoch': 2.24}) +[2026-06-16 05:53:02,390][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.0604148767888546, 'train/grad_norm_pre_clip': 6.355482578277588, 'timing/data': 0.013073885813355446, 'timing/model': 1.8779367534443736, 'learning_rate/qwen_vl_interface': 2.3122942111232506e-06, 'learning_rate/action_model': 2.3122942111232506e-05, 'epoch': 2.24}) +[2026-06-16 05:53:08,077][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.22116823494434357, 'train/grad_norm_pre_clip': 20.825359344482422, 'timing/data': 0.012560413219034672, 'timing/model': 1.4624021602794528, 'learning_rate/qwen_vl_interface': 2.3099016838075857e-06, 'learning_rate/action_model': 2.3099016838075856e-05, 'epoch': 2.24}) +[2026-06-16 05:53:14,176][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.04242178052663803, 'train/grad_norm_pre_clip': 4.708367347717285, 'timing/data': 0.011797348037362099, 'timing/model': 1.4659744994714856, 'learning_rate/qwen_vl_interface': 2.307510365055282e-06, 'learning_rate/action_model': 2.307510365055282e-05, 'epoch': 2.24}) +[2026-06-16 05:53:20,373][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.10956587642431259, 'train/grad_norm_pre_clip': 10.047167778015137, 'timing/data': 0.012275670655071735, 'timing/model': 1.4650213001295924, 'learning_rate/qwen_vl_interface': 2.3051202558493196e-06, 'learning_rate/action_model': 2.3051202558493192e-05, 'epoch': 2.24}) +[2026-06-16 05:53:26,498][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.08514953404664993, 'train/grad_norm_pre_clip': 11.342888832092285, 'timing/data': 0.012780143879354, 'timing/model': 1.4663444655016065, 'learning_rate/qwen_vl_interface': 2.302731357172182e-06, 'learning_rate/action_model': 2.302731357172182e-05, 'epoch': 2.24}) +[2026-06-16 05:53:32,638][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.02374211698770523, 'train/grad_norm_pre_clip': 1.929736852645874, 'timing/data': 0.012377901002764702, 'timing/model': 2.0167781924828887, 'learning_rate/qwen_vl_interface': 2.3003436700058556e-06, 'learning_rate/action_model': 2.3003436700058554e-05, 'epoch': 2.24}) +[2026-06-16 05:53:38,788][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.13489341735839844, 'train/grad_norm_pre_clip': 10.589491844177246, 'timing/data': 0.013622839003801346, 'timing/model': 1.9663299061357975, 'learning_rate/qwen_vl_interface': 2.297957195331828e-06, 'learning_rate/action_model': 2.2979571953318276e-05, 'epoch': 2.24}) +[2026-06-16 05:53:44,501][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.050891898572444916, 'train/grad_norm_pre_clip': 6.4513468742370605, 'timing/data': 0.012323630973696709, 'timing/model': 1.4794916547834873, 'learning_rate/qwen_vl_interface': 2.2955719341310903e-06, 'learning_rate/action_model': 2.2955719341310903e-05, 'epoch': 2.24}) +[2026-06-16 05:53:50,571][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.017849000170826912, 'train/grad_norm_pre_clip': 5.09340763092041, 'timing/data': 0.012578791938722134, 'timing/model': 1.4456869577988982, 'learning_rate/qwen_vl_interface': 2.2931878873841304e-06, 'learning_rate/action_model': 2.2931878873841302e-05, 'epoch': 2.24}) +[2026-06-16 05:53:56,781][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.15638168156147003, 'train/grad_norm_pre_clip': 7.276675701141357, 'timing/data': 0.011909138411283493, 'timing/model': 1.4771091667935252, 'learning_rate/qwen_vl_interface': 2.290805056070942e-06, 'learning_rate/action_model': 2.290805056070942e-05, 'epoch': 2.24}) +[2026-06-16 06:04:33,970][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.06549598276615143, 'train/grad_norm_pre_clip': 10.935159683227539, 'eval/loss': 0.056962087750434875, 'eval/flap_precision': 0.8609022556390977, 'eval/flap_recall': 0.8327272727272728, 'eval/flap_f1': 0.8465804066543439, 'eval/flap_rate_model': 0.07388888888888889, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.846580406654344, 'timing/data': 0.01260367315262556, 'timing/model': 1.4540080605074763, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 2.24}) +[2026-06-16 06:04:33,971][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4900_state +[2026-06-16 06:04:33,977][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 06:04:41,620][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4900_state/model.safetensors +[2026-06-16 06:04:52,995][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4900_state/optimizer.bin +[2026-06-16 06:04:52,996][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4900_state/sampler.bin +[2026-06-16 06:04:52,996][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4900_state/sampler_1.bin +[2026-06-16 06:04:52,998][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_4900_state/random_states_0.pkl +[2026-06-16 06:04:53,167][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4900 +[2026-06-16 06:04:53,168][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 06:04:53,183][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 06:04:58,739][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.036914385855197906, 'train/grad_norm_pre_clip': 4.441280364990234, 'timing/data': 0.013617082498967648, 'timing/model': 1.3767634322866797, 'learning_rate/qwen_vl_interface': 2.2860430436633447e-06, 'learning_rate/action_model': 2.2860430436633447e-05, 'epoch': 2.24}) +[2026-06-16 06:05:04,427][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.05246787145733833, 'train/grad_norm_pre_clip': 5.318846225738525, 'timing/data': 0.01345983985811472, 'timing/model': 1.396109058521688, 'learning_rate/qwen_vl_interface': 2.28366386452642e-06, 'learning_rate/action_model': 2.28366386452642e-05, 'epoch': 2.24}) +[2026-06-16 06:05:10,337][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.03680872917175293, 'train/grad_norm_pre_clip': 4.8333964347839355, 'timing/data': 0.012555375695228577, 'timing/model': 1.7886581299826503, 'learning_rate/qwen_vl_interface': 2.2812859047382326e-06, 'learning_rate/action_model': 2.2812859047382323e-05, 'epoch': 2.24}) +[2026-06-16 06:05:15,845][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.10824142396450043, 'train/grad_norm_pre_clip': 9.143169403076172, 'timing/data': 0.012858036905527115, 'timing/model': 1.456843826919794, 'learning_rate/qwen_vl_interface': 2.2789091652762712e-06, 'learning_rate/action_model': 2.2789091652762712e-05, 'epoch': 2.24}) +[2026-06-16 06:05:21,534][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.059858858585357666, 'train/grad_norm_pre_clip': 6.370640277862549, 'timing/data': 0.013006199151277542, 'timing/model': 1.436939686536789, 'learning_rate/qwen_vl_interface': 2.276533647117524e-06, 'learning_rate/action_model': 2.2765336471175242e-05, 'epoch': 2.24}) +[2026-06-16 06:05:27,142][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.17912858724594116, 'train/grad_norm_pre_clip': 10.606451034545898, 'timing/data': 0.012863647192716599, 'timing/model': 1.4435579953715205, 'learning_rate/qwen_vl_interface': 2.2741593512384773e-06, 'learning_rate/action_model': 2.2741593512384772e-05, 'epoch': 2.24}) +[2026-06-16 06:05:32,814][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.0696239024400711, 'train/grad_norm_pre_clip': 5.23729944229126, 'timing/data': 0.013372059911489487, 'timing/model': 1.4328299639746547, 'learning_rate/qwen_vl_interface': 2.2717862786151153e-06, 'learning_rate/action_model': 2.271786278615115e-05, 'epoch': 2.24}) +[2026-06-16 06:05:38,664][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.11940804123878479, 'train/grad_norm_pre_clip': 7.19096565246582, 'timing/data': 0.013132579624652863, 'timing/model': 1.4409554917365313, 'learning_rate/qwen_vl_interface': 2.2694144302229152e-06, 'learning_rate/action_model': 2.2694144302229152e-05, 'epoch': 2.24}) +[2026-06-16 06:05:44,393][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.07981760799884796, 'train/grad_norm_pre_clip': 9.029372215270996, 'timing/data': 0.01283576712012291, 'timing/model': 1.6897272542119026, 'learning_rate/qwen_vl_interface': 2.2670438070368568e-06, 'learning_rate/action_model': 2.2670438070368566e-05, 'epoch': 2.25}) +[2026-06-16 06:05:49,940][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.14340710639953613, 'train/grad_norm_pre_clip': 8.726313591003418, 'timing/data': 0.012626896612346172, 'timing/model': 1.4127448620274663, 'learning_rate/qwen_vl_interface': 2.264674410031413e-06, 'learning_rate/action_model': 2.2646744100314127e-05, 'epoch': 2.25}) +[2026-06-16 06:05:55,650][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.1158645823597908, 'train/grad_norm_pre_clip': 6.823211669921875, 'timing/data': 0.013777843676507473, 'timing/model': 1.4432774670422077, 'learning_rate/qwen_vl_interface': 2.2623062401805537e-06, 'learning_rate/action_model': 2.2623062401805535e-05, 'epoch': 2.25}) +[2026-06-16 06:06:01,338][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.1939103603363037, 'train/grad_norm_pre_clip': 8.797065734863281, 'timing/data': 0.012502594850957394, 'timing/model': 1.4152333941310644, 'learning_rate/qwen_vl_interface': 2.2599392984577436e-06, 'learning_rate/action_model': 2.2599392984577438e-05, 'epoch': 2.25}) +[2026-06-16 06:06:07,123][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 0.2517165243625641, 'train/grad_norm_pre_clip': 13.35399055480957, 'timing/data': 0.013025389052927494, 'timing/model': 1.4399460777640343, 'learning_rate/qwen_vl_interface': 2.2575735858359436e-06, 'learning_rate/action_model': 2.257573585835944e-05, 'epoch': 2.25}) +[2026-06-16 06:06:12,842][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.013966519385576248, 'train/grad_norm_pre_clip': 1.4043912887573242, 'timing/data': 0.013007128611207008, 'timing/model': 1.4061435740441084, 'learning_rate/qwen_vl_interface': 2.255209103287609e-06, 'learning_rate/action_model': 2.255209103287609e-05, 'epoch': 2.25}) +[2026-06-16 06:06:18,604][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 0.10361011326313019, 'train/grad_norm_pre_clip': 6.779762268066406, 'timing/data': 0.012933517806231976, 'timing/model': 1.705424927175045, 'learning_rate/qwen_vl_interface': 2.2528458517846886e-06, 'learning_rate/action_model': 2.252845851784689e-05, 'epoch': 2.25}) +[2026-06-16 06:06:24,498][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.03166007250547409, 'train/grad_norm_pre_clip': 7.508512020111084, 'timing/data': 0.012613287195563316, 'timing/model': 1.7157951472327113, 'learning_rate/qwen_vl_interface': 2.250483832298627e-06, 'learning_rate/action_model': 2.250483832298627e-05, 'epoch': 2.25}) +[2026-06-16 06:06:30,159][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.08581119775772095, 'train/grad_norm_pre_clip': 7.469048976898193, 'timing/data': 0.013921024277806282, 'timing/model': 1.4624279029667377, 'learning_rate/qwen_vl_interface': 2.2481230458003608e-06, 'learning_rate/action_model': 2.2481230458003606e-05, 'epoch': 2.25}) +[2026-06-16 06:06:35,971][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 0.06483263522386551, 'train/grad_norm_pre_clip': 8.507458686828613, 'timing/data': 0.014351315796375275, 'timing/model': 1.4850765578448772, 'learning_rate/qwen_vl_interface': 2.2457634932603184e-06, 'learning_rate/action_model': 2.2457634932603183e-05, 'epoch': 2.25}) +[2026-06-16 06:06:41,849][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.1092739999294281, 'train/grad_norm_pre_clip': 6.909611701965332, 'timing/data': 0.012994788587093353, 'timing/model': 1.4605601718649268, 'learning_rate/qwen_vl_interface': 2.2434051756484255e-06, 'learning_rate/action_model': 2.2434051756484253e-05, 'epoch': 2.25}) +[2026-06-16 06:06:47,563][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.11830120533704758, 'train/grad_norm_pre_clip': 8.045347213745117, 'timing/data': 0.012407195754349232, 'timing/model': 1.3940732348710299, 'learning_rate/qwen_vl_interface': 2.2410480939340945e-06, 'learning_rate/action_model': 2.2410480939340945e-05, 'epoch': 2.25}) +[2026-06-16 06:06:53,419][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.1806633174419403, 'train/grad_norm_pre_clip': 12.50298023223877, 'timing/data': 0.011995043605566025, 'timing/model': 1.7328016506507993, 'learning_rate/qwen_vl_interface': 2.238692249086234e-06, 'learning_rate/action_model': 2.2386922490862336e-05, 'epoch': 2.25}) +[2026-06-16 06:06:59,033][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.2257024496793747, 'train/grad_norm_pre_clip': 12.62157154083252, 'timing/data': 0.012231924571096897, 'timing/model': 1.450090249069035, 'learning_rate/qwen_vl_interface': 2.2363376420732426e-06, 'learning_rate/action_model': 2.2363376420732423e-05, 'epoch': 2.25}) +[2026-06-16 06:07:04,857][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 0.052026599645614624, 'train/grad_norm_pre_clip': 7.522566795349121, 'timing/data': 0.013454941101372242, 'timing/model': 1.4637259654700756, 'learning_rate/qwen_vl_interface': 2.233984273863011e-06, 'learning_rate/action_model': 2.233984273863011e-05, 'epoch': 2.25}) +[2026-06-16 06:07:10,664][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.03015192039310932, 'train/grad_norm_pre_clip': 5.998800277709961, 'timing/data': 0.012479355558753014, 'timing/model': 1.4746366776525974, 'learning_rate/qwen_vl_interface': 2.231632145422919e-06, 'learning_rate/action_model': 2.231632145422919e-05, 'epoch': 2.25}) +[2026-06-16 06:07:16,449][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.13489435613155365, 'train/grad_norm_pre_clip': 7.158935546875, 'timing/data': 0.011951563879847527, 'timing/model': 1.405891316011548, 'learning_rate/qwen_vl_interface': 2.2292812577198383e-06, 'learning_rate/action_model': 2.229281257719838e-05, 'epoch': 2.25}) +[2026-06-16 06:07:22,249][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.14365091919898987, 'train/grad_norm_pre_clip': 8.567137718200684, 'timing/data': 0.012009474448859692, 'timing/model': 1.4551083017140627, 'learning_rate/qwen_vl_interface': 2.2269316117201307e-06, 'learning_rate/action_model': 2.2269316117201307e-05, 'epoch': 2.25}) +[2026-06-16 06:07:28,093][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 0.2281046062707901, 'train/grad_norm_pre_clip': 9.517518997192383, 'timing/data': 0.013111378997564316, 'timing/model': 1.7077263062819839, 'learning_rate/qwen_vl_interface': 2.224583208389646e-06, 'learning_rate/action_model': 2.224583208389646e-05, 'epoch': 2.25}) +[2026-06-16 06:07:33,717][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.22216477990150452, 'train/grad_norm_pre_clip': 15.59974479675293, 'timing/data': 0.011744101531803608, 'timing/model': 1.437653535977006, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 2.25}) +[2026-06-16 06:07:39,545][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 0.07815832644701004, 'train/grad_norm_pre_clip': 6.4330878257751465, 'timing/data': 0.012828497216105461, 'timing/model': 1.4493196308612823, 'learning_rate/qwen_vl_interface': 2.219890133597194e-06, 'learning_rate/action_model': 2.219890133597194e-05, 'epoch': 2.25}) +[2026-06-16 06:07:45,353][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.09551854431629181, 'train/grad_norm_pre_clip': 7.4606032371521, 'timing/data': 0.013801313936710358, 'timing/model': 1.4618802526965737, 'learning_rate/qwen_vl_interface': 2.217545464064374e-06, 'learning_rate/action_model': 2.217545464064374e-05, 'epoch': 2.25}) +[2026-06-16 06:07:51,264][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.020641788840293884, 'train/grad_norm_pre_clip': 1.6648318767547607, 'timing/data': 0.012611227110028267, 'timing/model': 1.4804052552208304, 'learning_rate/qwen_vl_interface': 2.215202041059067e-06, 'learning_rate/action_model': 2.215202041059067e-05, 'epoch': 2.26}) +[2026-06-16 06:07:57,159][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.01398377027362585, 'train/grad_norm_pre_clip': 1.191725254058838, 'timing/data': 0.013123438693583012, 'timing/model': 1.3999368203803897, 'learning_rate/qwen_vl_interface': 2.212859865544566e-06, 'learning_rate/action_model': 2.212859865544566e-05, 'epoch': 2.26}) +[2026-06-16 06:08:03,001][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.06232176348567009, 'train/grad_norm_pre_clip': 8.402080535888672, 'timing/data': 0.013733633793890476, 'timing/model': 1.6841277051717043, 'learning_rate/qwen_vl_interface': 2.210518938483652e-06, 'learning_rate/action_model': 2.210518938483652e-05, 'epoch': 2.26}) +[2026-06-16 06:08:08,668][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.005338622257113457, 'train/grad_norm_pre_clip': 0.5434271693229675, 'timing/data': 0.013665982522070408, 'timing/model': 1.438603132031858, 'learning_rate/qwen_vl_interface': 2.20817926083859e-06, 'learning_rate/action_model': 2.20817926083859e-05, 'epoch': 2.26}) +[2026-06-16 06:08:14,562][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.018699616193771362, 'train/grad_norm_pre_clip': 1.9703254699707031, 'timing/data': 0.033051165752112865, 'timing/model': 1.4353228881955147, 'learning_rate/qwen_vl_interface': 2.205840833571135e-06, 'learning_rate/action_model': 2.205840833571135e-05, 'epoch': 2.26}) +[2026-06-16 06:08:20,360][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.05243605747818947, 'train/grad_norm_pre_clip': 7.319543838500977, 'timing/data': 0.012447137385606766, 'timing/model': 1.4730589529499412, 'learning_rate/qwen_vl_interface': 2.2035036576425264e-06, 'learning_rate/action_model': 2.203503657642526e-05, 'epoch': 2.26}) +[2026-06-16 06:08:26,146][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.09039246290922165, 'train/grad_norm_pre_clip': 8.877188682556152, 'timing/data': 0.011444101110100746, 'timing/model': 1.3955821432173252, 'learning_rate/qwen_vl_interface': 2.2011677340134867e-06, 'learning_rate/action_model': 2.2011677340134867e-05, 'epoch': 2.26}) +[2026-06-16 06:08:32,020][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.14171086251735687, 'train/grad_norm_pre_clip': 7.8076605796813965, 'timing/data': 0.02465046104043722, 'timing/model': 1.6676820581778884, 'learning_rate/qwen_vl_interface': 2.198833063644227e-06, 'learning_rate/action_model': 2.198833063644227e-05, 'epoch': 2.26}) +[2026-06-16 06:08:37,686][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.07020172476768494, 'train/grad_norm_pre_clip': 7.551187515258789, 'timing/data': 0.011388691142201424, 'timing/model': 1.4764239909127355, 'learning_rate/qwen_vl_interface': 2.196499647494443e-06, 'learning_rate/action_model': 2.196499647494443e-05, 'epoch': 2.26}) +[2026-06-16 06:08:43,552][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.22556769847869873, 'train/grad_norm_pre_clip': 18.332244873046875, 'timing/data': 0.015696664340794086, 'timing/model': 1.4411593554541469, 'learning_rate/qwen_vl_interface': 2.1941674865233144e-06, 'learning_rate/action_model': 2.1941674865233143e-05, 'epoch': 2.26}) +[2026-06-16 06:08:49,370][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.16622233390808105, 'train/grad_norm_pre_clip': 12.383662223815918, 'timing/data': 0.01112051960080862, 'timing/model': 1.4876085370779037, 'learning_rate/qwen_vl_interface': 2.1918365816895013e-06, 'learning_rate/action_model': 2.191836581689501e-05, 'epoch': 2.26}) +[2026-06-16 06:08:55,052][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.38093626499176025, 'train/grad_norm_pre_clip': 12.045166969299316, 'timing/data': 0.010858316905796528, 'timing/model': 1.3963855421170592, 'learning_rate/qwen_vl_interface': 2.189506933951153e-06, 'learning_rate/action_model': 2.1895069339511525e-05, 'epoch': 2.26}) +[2026-06-16 06:09:00,867][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.05989754945039749, 'train/grad_norm_pre_clip': 3.4952445030212402, 'timing/data': 0.010226454585790634, 'timing/model': 1.4496945710852742, 'learning_rate/qwen_vl_interface': 2.187178544265898e-06, 'learning_rate/action_model': 2.1871785442658983e-05, 'epoch': 2.26}) +[2026-06-16 06:09:06,801][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.0348643995821476, 'train/grad_norm_pre_clip': 4.173232078552246, 'timing/data': 0.01149734202772379, 'timing/model': 1.8054524259641767, 'learning_rate/qwen_vl_interface': 2.184851413590852e-06, 'learning_rate/action_model': 2.1848514135908518e-05, 'epoch': 2.26}) +[2026-06-16 06:09:12,466][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.04928380250930786, 'train/grad_norm_pre_clip': 7.383705139160156, 'timing/data': 0.010137113742530346, 'timing/model': 1.4706515604630113, 'learning_rate/qwen_vl_interface': 2.1825255428826087e-06, 'learning_rate/action_model': 2.182525542882609e-05, 'epoch': 2.26}) +[2026-06-16 06:09:18,267][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.13930819928646088, 'train/grad_norm_pre_clip': 12.956122398376465, 'timing/data': 0.011173349805176258, 'timing/model': 1.461242401972413, 'learning_rate/qwen_vl_interface': 2.180200933097246e-06, 'learning_rate/action_model': 2.180200933097246e-05, 'epoch': 2.26}) +[2026-06-16 06:09:24,076][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.1266103833913803, 'train/grad_norm_pre_clip': 10.134251594543457, 'timing/data': 0.014521146193146706, 'timing/model': 1.470003821887076, 'learning_rate/qwen_vl_interface': 2.1778775851903235e-06, 'learning_rate/action_model': 2.177877585190323e-05, 'epoch': 2.26}) +[2026-06-16 06:09:29,977][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.12405147403478622, 'train/grad_norm_pre_clip': 12.252876281738281, 'timing/data': 0.01089638751000166, 'timing/model': 1.4641544939950109, 'learning_rate/qwen_vl_interface': 2.175555500116881e-06, 'learning_rate/action_model': 2.175555500116881e-05, 'epoch': 2.26}) +[2026-06-16 06:09:35,884][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.1742721050977707, 'train/grad_norm_pre_clip': 9.894946098327637, 'timing/data': 0.013462171889841557, 'timing/model': 1.4702930804342031, 'learning_rate/qwen_vl_interface': 2.1732346788314404e-06, 'learning_rate/action_model': 2.1732346788314402e-05, 'epoch': 2.26}) +[2026-06-16 06:09:41,827][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.0033865312580019236, 'train/grad_norm_pre_clip': 0.5149351954460144, 'timing/data': 0.011877844110131264, 'timing/model': 1.7255799314007163, 'learning_rate/qwen_vl_interface': 2.170915122288006e-06, 'learning_rate/action_model': 2.170915122288006e-05, 'epoch': 2.26}) +[2026-06-16 06:09:47,462][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.08595730364322662, 'train/grad_norm_pre_clip': 7.208971977233887, 'timing/data': 0.010218294337391853, 'timing/model': 1.4623562628403306, 'learning_rate/qwen_vl_interface': 2.168596831440055e-06, 'learning_rate/action_model': 2.1685968314400552e-05, 'epoch': 2.26}) +[2026-06-16 06:09:53,280][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.10168835520744324, 'train/grad_norm_pre_clip': 10.89780044555664, 'timing/data': 0.011145458556711674, 'timing/model': 1.4879243830218911, 'learning_rate/qwen_vl_interface': 2.166279807240553e-06, 'learning_rate/action_model': 2.1662798072405528e-05, 'epoch': 2.26}) +[2026-06-16 06:09:59,191][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.06956330686807632, 'train/grad_norm_pre_clip': 8.756200790405273, 'timing/data': 0.01086579728871584, 'timing/model': 1.4939624080434442, 'learning_rate/qwen_vl_interface': 2.1639640506419394e-06, 'learning_rate/action_model': 2.163964050641939e-05, 'epoch': 2.27}) +[2026-06-16 06:10:05,093][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.06804268062114716, 'train/grad_norm_pre_clip': 5.511894702911377, 'timing/data': 0.012427055276930332, 'timing/model': 1.425529464147985, 'learning_rate/qwen_vl_interface': 2.161649562596135e-06, 'learning_rate/action_model': 2.161649562596135e-05, 'epoch': 2.27}) +[2026-06-16 06:10:10,881][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.09517253935337067, 'train/grad_norm_pre_clip': 9.252664566040039, 'timing/data': 0.012071095407009125, 'timing/model': 1.420614699833095, 'learning_rate/qwen_vl_interface': 2.1593363440545383e-06, 'learning_rate/action_model': 2.1593363440545384e-05, 'epoch': 2.27}) +[2026-06-16 06:10:16,825][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.010464215651154518, 'train/grad_norm_pre_clip': 1.0546425580978394, 'timing/data': 0.010695157572627068, 'timing/model': 1.7144279312342405, 'learning_rate/qwen_vl_interface': 2.157024395968025e-06, 'learning_rate/action_model': 2.1570243959680252e-05, 'epoch': 2.27}) +[2026-06-16 06:10:22,464][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 0.22368121147155762, 'train/grad_norm_pre_clip': 15.43589973449707, 'timing/data': 0.011542992666363716, 'timing/model': 1.4541783342137933, 'learning_rate/qwen_vl_interface': 2.154713719286952e-06, 'learning_rate/action_model': 2.154713719286952e-05, 'epoch': 2.27}) +[2026-06-16 06:10:28,368][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.10791192203760147, 'train/grad_norm_pre_clip': 6.578583240509033, 'timing/data': 0.010897347703576088, 'timing/model': 1.4712500600144267, 'learning_rate/qwen_vl_interface': 2.152404314961149e-06, 'learning_rate/action_model': 2.152404314961149e-05, 'epoch': 2.27}) +[2026-06-16 06:10:34,167][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.015580913983285427, 'train/grad_norm_pre_clip': 4.698780536651611, 'timing/data': 0.011708253063261509, 'timing/model': 1.4689704114571214, 'learning_rate/qwen_vl_interface': 2.1500961839399264e-06, 'learning_rate/action_model': 2.1500961839399262e-05, 'epoch': 2.27}) +[2026-06-16 06:10:39,861][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.021897362545132637, 'train/grad_norm_pre_clip': 1.952349305152893, 'timing/data': 0.011975384317338467, 'timing/model': 1.4642782965674996, 'learning_rate/qwen_vl_interface': 2.1477893271720683e-06, 'learning_rate/action_model': 2.1477893271720682e-05, 'epoch': 2.27}) +[2026-06-16 06:10:45,654][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.2934592366218567, 'train/grad_norm_pre_clip': 16.351259231567383, 'timing/data': 0.011132869869470596, 'timing/model': 1.4155660178512335, 'learning_rate/qwen_vl_interface': 2.1454837456058385e-06, 'learning_rate/action_model': 2.1454837456058386e-05, 'epoch': 2.27}) +[2026-06-16 06:10:51,514][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 0.044356539845466614, 'train/grad_norm_pre_clip': 11.184659004211426, 'timing/data': 0.010827157646417618, 'timing/model': 1.705229819752276, 'learning_rate/qwen_vl_interface': 2.1431794401889722e-06, 'learning_rate/action_model': 2.143179440188972e-05, 'epoch': 2.27}) +[2026-06-16 06:10:57,167][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.13277283310890198, 'train/grad_norm_pre_clip': 8.6590576171875, 'timing/data': 0.011675682850182056, 'timing/model': 1.4564081635326147, 'learning_rate/qwen_vl_interface': 2.1408764118686837e-06, 'learning_rate/action_model': 2.1408764118686838e-05, 'epoch': 2.27}) +[2026-06-16 06:11:02,970][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.09381174296140671, 'train/grad_norm_pre_clip': 4.442286014556885, 'timing/data': 0.010814727284014225, 'timing/model': 1.4765571542084217, 'learning_rate/qwen_vl_interface': 2.1385746615916598e-06, 'learning_rate/action_model': 2.1385746615916596e-05, 'epoch': 2.27}) +[2026-06-16 06:11:09,204][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 0.2666827440261841, 'train/grad_norm_pre_clip': 33.49259567260742, 'timing/data': 0.012085044756531715, 'timing/model': 1.4315391341224313, 'learning_rate/qwen_vl_interface': 2.1362741903040636e-06, 'learning_rate/action_model': 2.1362741903040633e-05, 'epoch': 2.27}) +[2026-06-16 06:11:15,070][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.1291169822216034, 'train/grad_norm_pre_clip': 10.07938289642334, 'timing/data': 0.013792494311928749, 'timing/model': 1.4608959946781397, 'learning_rate/qwen_vl_interface': 2.1339749989515318e-06, 'learning_rate/action_model': 2.1339749989515316e-05, 'epoch': 2.27}) +[2026-06-16 06:11:20,860][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.12309757620096207, 'train/grad_norm_pre_clip': 7.460010051727295, 'timing/data': 0.011875834316015244, 'timing/model': 1.4031506357714534, 'learning_rate/qwen_vl_interface': 2.131677088479175e-06, 'learning_rate/action_model': 2.1316770884791752e-05, 'epoch': 2.27}) +[2026-06-16 06:11:26,723][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 0.14587602019309998, 'train/grad_norm_pre_clip': 10.476109504699707, 'timing/data': 0.011480932123959064, 'timing/model': 1.7102807546034455, 'learning_rate/qwen_vl_interface': 2.1293804598315787e-06, 'learning_rate/action_model': 2.1293804598315787e-05, 'epoch': 2.27}) +[2026-06-16 06:11:32,378][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.05616043508052826, 'train/grad_norm_pre_clip': 6.740425109863281, 'timing/data': 0.011366210877895355, 'timing/model': 1.475794623605907, 'learning_rate/qwen_vl_interface': 2.1270851139527986e-06, 'learning_rate/action_model': 2.1270851139527983e-05, 'epoch': 2.27}) +[2026-06-16 06:11:38,269][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.05287900194525719, 'train/grad_norm_pre_clip': 6.4196953773498535, 'timing/data': 0.011182660236954689, 'timing/model': 1.460594822652638, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 2.27}) +[2026-06-16 06:11:44,169][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 0.22287261486053467, 'train/grad_norm_pre_clip': 11.097005844116211, 'timing/data': 0.011286740191280842, 'timing/model': 1.466600957326591, 'learning_rate/qwen_vl_interface': 2.1224982742752814e-06, 'learning_rate/action_model': 2.1224982742752812e-05, 'epoch': 2.27}) +[2026-06-16 06:11:49,956][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.07970807701349258, 'train/grad_norm_pre_clip': 12.22545337677002, 'timing/data': 0.010280005633831024, 'timing/model': 1.4393098019063473, 'learning_rate/qwen_vl_interface': 2.1202067823620213e-06, 'learning_rate/action_model': 2.1202067823620212e-05, 'epoch': 2.27}) +[2026-06-16 06:11:55,769][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.0294001754373312, 'train/grad_norm_pre_clip': 3.8205697536468506, 'timing/data': 0.01078103855252266, 'timing/model': 1.4042088249698281, 'learning_rate/qwen_vl_interface': 2.117916576988531e-06, 'learning_rate/action_model': 2.117916576988531e-05, 'epoch': 2.27}) +[2026-06-16 06:12:01,635][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.16960060596466064, 'train/grad_norm_pre_clip': 10.438796043395996, 'timing/data': 0.012160495854914188, 'timing/model': 1.728133076801896, 'learning_rate/qwen_vl_interface': 2.1156276590962257e-06, 'learning_rate/action_model': 2.1156276590962257e-05, 'epoch': 2.27}) +[2026-06-16 06:12:07,263][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.07367488741874695, 'train/grad_norm_pre_clip': 8.386387825012207, 'timing/data': 0.011432611383497715, 'timing/model': 1.4604991031810641, 'learning_rate/qwen_vl_interface': 2.113340029625994e-06, 'learning_rate/action_model': 2.1133400296259938e-05, 'epoch': 2.28}) +[2026-06-16 06:12:13,066][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.1066526398062706, 'train/grad_norm_pre_clip': 17.935941696166992, 'timing/data': 0.01164525281637907, 'timing/model': 1.4676861939951777, 'learning_rate/qwen_vl_interface': 2.111053689518194e-06, 'learning_rate/action_model': 2.111053689518194e-05, 'epoch': 2.28}) +[2026-06-16 06:12:18,868][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.17289277911186218, 'train/grad_norm_pre_clip': 13.015091896057129, 'timing/data': 0.01143279206007719, 'timing/model': 1.4682607585564256, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 2.28}) +[2026-06-16 06:12:24,764][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.09565677493810654, 'train/grad_norm_pre_clip': 5.842392921447754, 'timing/data': 0.011941374279558659, 'timing/model': 1.4548875391483307, 'learning_rate/qwen_vl_interface': 2.106484881148672e-06, 'learning_rate/action_model': 2.1064848811486718e-05, 'epoch': 2.28}) +[2026-06-16 06:12:30,557][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.18319596350193024, 'train/grad_norm_pre_clip': 11.948413848876953, 'timing/data': 0.011315281502902508, 'timing/model': 1.400153225287795, 'learning_rate/qwen_vl_interface': 2.1042024147650137e-06, 'learning_rate/action_model': 2.1042024147650137e-05, 'epoch': 2.28}) +[2026-06-16 06:12:36,418][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.20817546546459198, 'train/grad_norm_pre_clip': 9.233601570129395, 'timing/data': 0.010684958659112453, 'timing/model': 1.71269999910146, 'learning_rate/qwen_vl_interface': 2.1019212414999163e-06, 'learning_rate/action_model': 2.101921241499916e-05, 'epoch': 2.28}) +[2026-06-16 06:12:42,257][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.2858535945415497, 'train/grad_norm_pre_clip': 17.673599243164062, 'timing/data': 0.011289781890809536, 'timing/model': 1.6286885859444737, 'learning_rate/qwen_vl_interface': 2.099641362291082e-06, 'learning_rate/action_model': 2.099641362291082e-05, 'epoch': 2.28}) +[2026-06-16 06:12:47,853][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.07977382093667984, 'train/grad_norm_pre_clip': 5.47924280166626, 'timing/data': 0.010783718898892403, 'timing/model': 1.433166740462184, 'learning_rate/qwen_vl_interface': 2.097362778075685e-06, 'learning_rate/action_model': 2.097362778075685e-05, 'epoch': 2.28}) +[2026-06-16 06:12:53,660][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.11397294700145721, 'train/grad_norm_pre_clip': 10.42882251739502, 'timing/data': 0.012201156467199326, 'timing/model': 1.4336571646854281, 'learning_rate/qwen_vl_interface': 2.0950854897903663e-06, 'learning_rate/action_model': 2.095085489790366e-05, 'epoch': 2.28}) +[2026-06-16 06:12:59,464][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 0.02812192216515541, 'train/grad_norm_pre_clip': 4.7354536056518555, 'timing/data': 0.011068769730627537, 'timing/model': 1.4717041719704866, 'learning_rate/qwen_vl_interface': 2.092809498371229e-06, 'learning_rate/action_model': 2.0928094983712292e-05, 'epoch': 2.28}) +[2026-06-16 06:13:05,269][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.04828064516186714, 'train/grad_norm_pre_clip': 4.389218330383301, 'timing/data': 0.010640827938914299, 'timing/model': 1.402168638072908, 'learning_rate/qwen_vl_interface': 2.0905348047538498e-06, 'learning_rate/action_model': 2.0905348047538497e-05, 'epoch': 2.28}) +[2026-06-16 06:13:11,130][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.07894381135702133, 'train/grad_norm_pre_clip': 7.537718296051025, 'timing/data': 0.010704108513891697, 'timing/model': 1.725566241890192, 'learning_rate/qwen_vl_interface': 2.0882614098732686e-06, 'learning_rate/action_model': 2.0882614098732685e-05, 'epoch': 2.28}) +[2026-06-16 06:13:17,033][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.16074910759925842, 'train/grad_norm_pre_clip': 11.696389198303223, 'timing/data': 0.011249691247940063, 'timing/model': 1.723478383384645, 'learning_rate/qwen_vl_interface': 2.085989314663993e-06, 'learning_rate/action_model': 2.0859893146639927e-05, 'epoch': 2.28}) +[2026-06-16 06:13:22,660][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.14324918389320374, 'train/grad_norm_pre_clip': 8.007798194885254, 'timing/data': 0.011195630766451359, 'timing/model': 1.4626816920936108, 'learning_rate/qwen_vl_interface': 2.083718520059995e-06, 'learning_rate/action_model': 2.0837185200599946e-05, 'epoch': 2.28}) +[2026-06-16 06:13:28,457][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.07447673380374908, 'train/grad_norm_pre_clip': 3.8726208209991455, 'timing/data': 0.012003574520349503, 'timing/model': 1.444109513424337, 'learning_rate/qwen_vl_interface': 2.0814490269947117e-06, 'learning_rate/action_model': 2.0814490269947118e-05, 'epoch': 2.28}) +[2026-06-16 06:13:34,258][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.1258099228143692, 'train/grad_norm_pre_clip': 7.661095142364502, 'timing/data': 0.011469372548162937, 'timing/model': 1.4475601930171251, 'learning_rate/qwen_vl_interface': 2.079180836401047e-06, 'learning_rate/action_model': 2.079180836401047e-05, 'epoch': 2.28}) +[2026-06-16 06:13:40,080][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.036016155034303665, 'train/grad_norm_pre_clip': 3.9846832752227783, 'timing/data': 0.011649093590676785, 'timing/model': 1.406556441448629, 'learning_rate/qwen_vl_interface': 2.0769139492113687e-06, 'learning_rate/action_model': 2.076913949211369e-05, 'epoch': 2.28}) +[2026-06-16 06:13:46,018][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.11490143090486526, 'train/grad_norm_pre_clip': 14.201292037963867, 'timing/data': 0.011290131136775017, 'timing/model': 1.7078273296356201, 'learning_rate/qwen_vl_interface': 2.0746483663575075e-06, 'learning_rate/action_model': 2.0746483663575076e-05, 'epoch': 2.28}) +[2026-06-16 06:13:51,673][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.11251819878816605, 'train/grad_norm_pre_clip': 8.690177917480469, 'timing/data': 0.01125281024724245, 'timing/model': 1.477275219745934, 'learning_rate/qwen_vl_interface': 2.0723840887707595e-06, 'learning_rate/action_model': 2.0723840887707594e-05, 'epoch': 2.28}) +[2026-06-16 06:13:57,560][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.05123509466648102, 'train/grad_norm_pre_clip': 4.600097179412842, 'timing/data': 0.011440081521868706, 'timing/model': 1.4587652618065476, 'learning_rate/qwen_vl_interface': 2.0701211173818854e-06, 'learning_rate/action_model': 2.0701211173818853e-05, 'epoch': 2.28}) +[2026-06-16 06:14:03,376][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 0.06672338396310806, 'train/grad_norm_pre_clip': 11.357321739196777, 'timing/data': 0.011147300712764263, 'timing/model': 1.476771960966289, 'learning_rate/qwen_vl_interface': 2.0678594531211036e-06, 'learning_rate/action_model': 2.0678594531211033e-05, 'epoch': 2.28}) +[2026-06-16 06:14:09,265][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.0670265182852745, 'train/grad_norm_pre_clip': 7.308030128479004, 'timing/data': 0.012062245048582554, 'timing/model': 1.4124584728851914, 'learning_rate/qwen_vl_interface': 2.0655990969181e-06, 'learning_rate/action_model': 2.0655990969181003e-05, 'epoch': 2.28}) +[2026-06-16 06:14:15,088][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.05086100846529007, 'train/grad_norm_pre_clip': 8.947562217712402, 'timing/data': 0.010447506792843342, 'timing/model': 1.4847669582813978, 'learning_rate/qwen_vl_interface': 2.0633400497020236e-06, 'learning_rate/action_model': 2.0633400497020232e-05, 'epoch': 2.29}) +[2026-06-16 06:14:20,987][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.03327241912484169, 'train/grad_norm_pre_clip': 4.902716159820557, 'timing/data': 0.01237068697810173, 'timing/model': 1.6845732051879168, 'learning_rate/qwen_vl_interface': 2.061082312401481e-06, 'learning_rate/action_model': 2.061082312401481e-05, 'epoch': 2.29}) +[2026-06-16 06:14:26,656][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.03672859072685242, 'train/grad_norm_pre_clip': 4.877793312072754, 'timing/data': 0.011375212110579014, 'timing/model': 1.443298576399684, 'learning_rate/qwen_vl_interface': 2.058825885944544e-06, 'learning_rate/action_model': 2.0588258859445442e-05, 'epoch': 2.29}) +[2026-06-16 06:24:55,473][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.20143906772136688, 'train/grad_norm_pre_clip': 17.406496047973633, 'eval/loss': 0.055859412997961044, 'eval/flap_precision': 0.8818897637795275, 'eval/flap_recall': 0.8145454545454546, 'eval/flap_f1': 0.8468809073724006, 'eval/flap_rate_model': 0.07055555555555555, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.8468809073724006, 'timing/data': 0.010428275913000107, 'timing/model': 1.4788315007463098, 'learning_rate/qwen_vl_interface': 2.0565707712587447e-06, 'learning_rate/action_model': 2.0565707712587444e-05, 'epoch': 2.29}) +[2026-06-16 06:24:55,474][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_5000_state +[2026-06-16 06:24:55,482][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-16 06:25:02,981][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_5000_state/model.safetensors +[2026-06-16 06:25:14,121][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_5000_state/optimizer.bin +[2026-06-16 06:25:14,122][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_5000_state/sampler.bin +[2026-06-16 06:25:14,122][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_5000_state/sampler_1.bin +[2026-06-16 06:25:14,124][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_5000_state/random_states_0.pkl +[2026-06-16 06:25:14,278][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-06-16 06:25:14,278][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-16 06:25:14,293][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-16 06:25:14,294][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints +[2026-06-16 06:25:16,594][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/summary.jsonl b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..ffa1ed16fa7cd08f0c83739d659e9d2a8bd4e827 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/summary.jsonl @@ -0,0 +1,50 @@ +{"steps": 100} +{"steps": 200} +{"steps": 300} +{"steps": 400} +{"steps": 500} +{"steps": 600} +{"steps": 700} +{"steps": 800} +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500} +{"steps": 1600} +{"steps": 1700} +{"steps": 1800} +{"steps": 1900} +{"steps": 2000} +{"steps": 2100} +{"steps": 2200} +{"steps": 2300} +{"steps": 2400} +{"steps": 2500} +{"steps": 2600} +{"steps": 2700} +{"steps": 2800} +{"steps": 2900} +{"steps": 3000} +{"steps": 3100} +{"steps": 3200} +{"steps": 3300} +{"steps": 3400} +{"steps": 3500} +{"steps": 3600} +{"steps": 3700} +{"steps": 3800} +{"steps": 3900} +{"steps": 4000} +{"steps": 4100} +{"steps": 4200} +{"steps": 4300} +{"steps": 4400} +{"steps": 4500} +{"steps": 4600} +{"steps": 4700} +{"steps": 4800} +{"steps": 4900} +{"steps": 5000} diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/checkpoints/steps_5000_model.safetensors b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/checkpoints/steps_5000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8e15010370cbbd700b69541c948bec52a176c56c --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/checkpoints/steps_5000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:57b21a1d51a95495f08c60166b561aed0183fd4b680b665a440b9b0694aff4c8 +size 9138230516 diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/checkpoints/steps_5000_pytorch_model.pt b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/checkpoints/steps_5000_pytorch_model.pt new file mode 100644 index 0000000000000000000000000000000000000000..52ff68dd8afeaf0fa4e4b59637dcd4105e3b115e --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/checkpoints/steps_5000_pytorch_model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a7581596728be2fb496d24bea5ec1c8586a5fcbf8ee99f1e9d5c90f11288f70b +size 9138476289 diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/config.full.yaml b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..fcba0d5cbfd74e566419223cdb90e54d20c06318 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/config.full.yaml @@ -0,0 +1,175 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_parquet + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 16 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 250 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: true + pretrained_checkpoint: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1250_state + resume_step: 1250 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +seed: 42 +wandb_entity: your_wandb_entity +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + task_description: '' + enabled: true + latency: + prompt_map_path: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_parquet/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 1000 + latencies: + - 0 + num_episodes: 1 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + local: + keep_last_n: 2 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_flappy_fix_latency_0 +output_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0 +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/config.yaml b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..b5ffb5f0c78b93c5fc730331f5bd724d04c0b31d --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/config.yaml @@ -0,0 +1,79 @@ +checkpoint: + local: + keep_last_n: 2 + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + include_state: true + obs_image_size: null + per_device_batch_size: 16 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0 +rl_games: + env_eval: + enabled: true + frameskip: 1 + image_size: 224 + latency: + prompt_map_path: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_parquet/flappy_train__bridge/latency_prompt_map.json + mid_train: + enabled: true + interval_steps: 1000 + latencies: + - 0 + max_steps_per_episode: 3600 + num_episodes: 1 + task_description: '' + model_alias: openvla + task: flappy +run_id: openvla_flappy_fix_latency_0 +run_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +seed: 42 +trainer: + eval_interval: 100 + eval_num_batches: 20 + freeze_modules: '' + gradient_clipping: 1.0 + is_resume: true + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 10 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + pretrained_checkpoint: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1250_state + resume_step: 1250 + save_format: pt + save_interval: 250 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: your_wandb_entity +wandb_project: starVLA_rl_games diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/dataset_statistics.json b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..ac84c54c843dac27ce0011f7dc1c1e7a13e5ecf9 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.9194239377975464, + 0.0805760845541954, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.2723730504512787, + 0.2723730504512787, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 235591, + "num_trajectories": 80 + } +} \ No newline at end of file diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/dataset_statistics_eval.json b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..11f65e877e55e69c71b3a19eacbac31548dd7643 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.9190378189086914, + 0.08096218854188919, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.2728092670440674, + 0.2728092670440674, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 57286, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/hydra/.hydra/config.yaml b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..f817c346ef07d4d7da3fc02d6e8502822ef5cbae --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/hydra/.hydra/config.yaml @@ -0,0 +1,173 @@ +framework: + qwenvl: + base_vlm: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_parquet + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 16 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 250 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: true + pretrained_checkpoint: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1250_state + resume_step: 1250 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +seed: 42 +wandb_entity: your_wandb_entity +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + task_description: '' + enabled: true + latency: + prompt_map_path: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_parquet/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 1000 + latencies: + - 0 + num_episodes: 1 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + local: + keep_last_n: 2 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_flappy_fix_latency_0 +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/hydra/.hydra/hydra.yaml b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..7ba1f3a4e405816b833bfe4798443a23bad30866 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/hydra/.hydra/hydra.yaml @@ -0,0 +1,238 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=single + - run_id=openvla_flappy_fix_latency_0 + - run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints + - seed=42 + - wandb_entity=your_wandb_entity + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=2 + - trainer.is_resume=true + - trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1250_state + - trainer.resume_step=1250 + - trainer.max_train_steps=5000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=250 + - trainer.eval_interval=100 + - trainer.eval_num_batches=20 + - trainer.logging_frequency=10 + - trainer.gradient_accumulation_steps=16 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9, 0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=2 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.per_device_batch_size=16 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_parquet + - datasets.vla_data.data_mix=flappy_train__bridge + - datasets.vla_data.eval_data_mix=flappy_train__bridge__val + - framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=flappy + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=1 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_parquet/flappy_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=true + - rl_games.env_eval.mid_train.interval_steps=1000 + - rl_games.env_eval.mid_train.num_episodes=1 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[0] + - rl_games.env_eval.post_train.enabled=true + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=2,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=flappy_train__bridge,datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_parquet,datasets.vla_data.eval_data_mix=flappy_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=16,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.video_backend=torchvision_av,env=flappy,framework.action_model.action_dim=7,framework.action_model.action_env_dim=2,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.enabled=true,rl_games.env_eval.frameskip=1,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_parquet/flappy_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.mid_train.enabled=true,rl_games.env_eval.mid_train.interval_steps=1000,rl_games.env_eval.mid_train.latencies=[0],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=1,rl_games.env_eval.post_train.enabled=true,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=flappy,run_id=openvla_flappy_fix_latency_0,run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=20,trainer.gradient_accumulation_steps=16,trainer.gradient_clipping=1.0,trainer.is_resume=true,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=10,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=5000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9, + 0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1250_state,trainer.resume_step=1250,trainer.save_format=pt,trainer.save_interval=250,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=your_wandb_entity,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.2 + version_base: '1.1' + cwd: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/hydra + choices: + checkpoint: default + mode: single + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/hydra/.hydra/overrides.yaml b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..fd043cfc32c478ef2de2adf5420380bd9c257dbb --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/hydra/.hydra/overrides.yaml @@ -0,0 +1,78 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=single +- run_id=openvla_flappy_fix_latency_0 +- run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +- seed=42 +- wandb_entity=your_wandb_entity +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=2 +- trainer.is_resume=true +- trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1250_state +- trainer.resume_step=1250 +- trainer.max_train_steps=5000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=250 +- trainer.eval_interval=100 +- trainer.eval_num_batches=20 +- trainer.logging_frequency=10 +- trainer.gradient_accumulation_steps=16 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9, 0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=2 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.per_device_batch_size=16 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_parquet +- datasets.vla_data.data_mix=flappy_train__bridge +- datasets.vla_data.eval_data_mix=flappy_train__bridge__val +- framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=flappy +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=1 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_parquet/flappy_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=true +- rl_games.env_eval.mid_train.interval_steps=1000 +- rl_games.env_eval.mid_train.num_episodes=1 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[0] +- rl_games.env_eval.post_train.enabled=true +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/hydra/train_starvla_hydra.log b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..c6a8b9f0e7f101e6cc427e8df256c928c1df722c --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/hydra/train_starvla_hydra.log @@ -0,0 +1,1032 @@ +[2026-05-26 07:52:36,252][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-05-26 07:52:36,353][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-05-26 07:52:38,042][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-05-26 07:52:46,632][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-05-26 07:52:46,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-05-26 07:52:46,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-05-26 07:52:46,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-05-26 07:52:46,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-05-26 07:52:46,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-05-26 07:52:46,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-05-26 07:52:46,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-05-26 07:52:46,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-05-26 07:52:46,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-05-26 07:52:46,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-05-26 07:52:46,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-05-26 07:52:46,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-05-26 07:52:46,634][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-05-26 07:52:46,634][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-05-26 07:52:46,634][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-05-26 07:52:46,634][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-05-26 07:52:46,634][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-05-26 07:52:46,634][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-05-26 07:52:46,634][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-05-26 07:52:46,634][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-05-26 07:52:46,636][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-05-26 07:52:46,636][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-05-26 07:52:46,636][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-05-26 07:52:46,636][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-05-26 07:52:46,636][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-05-26 07:52:46,636][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-05-26 07:52:46,636][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-05-26 07:52:46,637][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-05-26 07:52:46,637][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-05-26 07:52:46,637][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-05-26 07:52:46,670][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-05-26 07:52:46,887][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/config.full.yaml +[2026-05-26 07:52:46,893][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/config.yaml +[2026-05-26 07:53:45,186][starVLA.training.train_starvla][INFO] - Loaded pretrained checkpoint: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt, steps: 0 +[2026-05-26 07:53:50,620][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-05-26 07:53:50,620][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-05-26 07:53:50,621][starVLA.training.train_starvla][INFO] - Per device batch size = 16 +[2026-05-26 07:53:50,621][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-05-26 07:53:50,621][starVLA.training.train_starvla][INFO] - Total batch size = 256 +[2026-05-26 07:54:57,977][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.3155343532562256, 'train/grad_norm_pre_clip': 1.7430894374847412, 'timing/data': 0.0046198368072509766, 'timing/model': 0.40929071325808764, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-05-26 07:55:54,318][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.022467855364084244, 'train/grad_norm_pre_clip': 0.14473594725131989, 'timing/data': 0.0040763914585113525, 'timing/model': 0.3970283614471555, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.0}) +[2026-05-26 07:56:49,806][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.014857729896903038, 'train/grad_norm_pre_clip': 0.09629420936107635, 'timing/data': 0.003962784074246883, 'timing/model': 0.3887085961177945, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.0}) +[2026-05-26 07:57:45,054][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.40035778284072876, 'train/grad_norm_pre_clip': 0.4383750855922699, 'timing/data': 0.004240597598254681, 'timing/model': 0.3867976190522313, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.0}) +[2026-05-26 07:58:41,051][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.3815605640411377, 'train/grad_norm_pre_clip': 0.18896286189556122, 'timing/data': 0.004400184378027916, 'timing/model': 0.44659964367747307, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.0}) +[2026-05-26 07:59:36,756][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.49023109674453735, 'train/grad_norm_pre_clip': 0.3258328139781952, 'timing/data': 0.004819302819669247, 'timing/model': 0.4341430999338627, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.0}) +[2026-05-26 08:00:33,292][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.5790913105010986, 'train/grad_norm_pre_clip': 0.8673839569091797, 'timing/data': 0.004007555544376373, 'timing/model': 0.3879575729370117, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.0}) +[2026-05-26 08:01:29,756][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.19988428056240082, 'train/grad_norm_pre_clip': 0.2159666121006012, 'timing/data': 0.0040526920929551125, 'timing/model': 0.40828380081802607, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.01}) +[2026-05-26 08:02:25,889][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.33861222863197327, 'train/grad_norm_pre_clip': 1.2361897230148315, 'timing/data': 0.004056715406477451, 'timing/model': 0.3866487508639693, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.01}) +[2026-05-26 08:03:34,325][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.44658026099205017, 'train/grad_norm_pre_clip': 0.9984492659568787, 'eval/loss': 0.27781978249549866, 'timing/data': 0.004580206237733364, 'timing/model': 0.4270592210814357, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.01}) +[2026-05-26 08:04:52,933][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.1992436796426773, 'train/grad_norm_pre_clip': 0.4391128122806549, 'timing/data': 0.004204592667520046, 'timing/model': 0.5044702626764774, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.01}) +[2026-05-26 08:06:11,621][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.04634345695376396, 'train/grad_norm_pre_clip': 0.3874465525150299, 'timing/data': 0.0040696850046515465, 'timing/model': 0.5028299335390329, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.01}) +[2026-05-26 08:07:30,776][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.12500038743019104, 'train/grad_norm_pre_clip': 0.48659178614616394, 'timing/data': 0.005002740770578384, 'timing/model': 0.536154811270535, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.01}) +[2026-05-26 08:08:49,817][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.20599624514579773, 'train/grad_norm_pre_clip': 0.3133833110332489, 'timing/data': 0.004006239585578442, 'timing/model': 0.503490868024528, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.01}) +[2026-05-26 08:10:09,288][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.19256597757339478, 'train/grad_norm_pre_clip': 0.7218695282936096, 'timing/data': 0.004000864923000336, 'timing/model': 0.5596651760861278, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.01}) +[2026-05-26 08:11:28,582][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.2662835717201233, 'train/grad_norm_pre_clip': 1.2161670923233032, 'timing/data': 0.003910786472260952, 'timing/model': 0.5748085509985685, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.01}) +[2026-05-26 08:12:47,477][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.34478360414505005, 'train/grad_norm_pre_clip': 1.0638062953948975, 'timing/data': 0.0038884328678250313, 'timing/model': 0.6624643970280886, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.01}) +[2026-05-26 08:14:06,372][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.2262689173221588, 'train/grad_norm_pre_clip': 1.166008472442627, 'timing/data': 0.004239778965711594, 'timing/model': 0.5129929222166538, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.01}) +[2026-05-26 08:15:25,325][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.2984667122364044, 'train/grad_norm_pre_clip': 1.219051480293274, 'timing/data': 0.0048274751752614975, 'timing/model': 0.5975780878216028, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.01}) +[2026-05-26 08:16:54,592][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.09621378779411316, 'train/grad_norm_pre_clip': 0.40220046043395996, 'eval/loss': 0.32560744881629944, 'timing/data': 0.003979063592851162, 'timing/model': 0.5087554240599275, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.01}) +[2026-05-26 08:18:14,126][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.2471531480550766, 'train/grad_norm_pre_clip': 4.595191478729248, 'timing/data': 0.004108627326786518, 'timing/model': 0.5025056637823582, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.01}) +[2026-05-26 08:19:34,492][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.7721076011657715, 'train/grad_norm_pre_clip': 2.5086348056793213, 'timing/data': 0.004595332778990269, 'timing/model': 0.5621213372796774, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.01}) +[2026-05-26 08:20:54,251][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.12965238094329834, 'train/grad_norm_pre_clip': 0.9864405393600464, 'timing/data': 0.003971686586737633, 'timing/model': 0.6638498166576028, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.02}) +[2026-05-26 08:22:13,136][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.37148773670196533, 'train/grad_norm_pre_clip': 0.989891767501831, 'timing/data': 0.004603167995810509, 'timing/model': 0.6736252224072814, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.02}) +[2026-05-26 08:23:31,322][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.12146244943141937, 'train/grad_norm_pre_clip': 1.1728515625, 'timing/data': 0.005141319707036018, 'timing/model': 0.5321720233187079, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.02}) +[2026-05-26 08:23:31,371][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_250_state +[2026-05-26 08:23:31,377][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-26 08:23:37,886][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_250_state/model.safetensors +[2026-05-26 08:23:53,826][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_250_state/optimizer.bin +[2026-05-26 08:23:53,827][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_250_state/sampler.bin +[2026-05-26 08:23:53,827][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_250_state/sampler_1.bin +[2026-05-26 08:23:53,829][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_250_state/random_states_0.pkl +[2026-05-26 08:23:54,786][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpdw8sb68l/test.c -o /tmp/tmpdw8sb68l/test.o +[2026-05-26 08:23:54,801][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpdw8sb68l/test.o -laio -o /tmp/tmpdw8sb68l/a.out +[2026-05-26 08:23:55,277][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpj3n6ekye/test.c -o /tmp/tmpj3n6ekye/test.o +[2026-05-26 08:23:55,291][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpj3n6ekye/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpj3n6ekye/a.out +[2026-05-26 08:23:55,326][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmp29dva9in/test.c -o /tmp/tmp29dva9in/test.o +[2026-05-26 08:23:55,337][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmp29dva9in/test.o -laio -o /tmp/tmp29dva9in/a.out +[2026-05-26 08:24:10,513][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 250 +[2026-05-26 08:24:10,514][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-26 08:24:10,545][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-26 08:25:30,590][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.4192356467247009, 'train/grad_norm_pre_clip': 2.3335018157958984, 'timing/data': 0.004002700559794903, 'timing/model': 0.5050596976652741, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.02}) +[2026-05-26 08:26:50,087][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.2960728704929352, 'train/grad_norm_pre_clip': 0.7570515275001526, 'timing/data': 0.00427049957215786, 'timing/model': 0.49422607105225325, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.02}) +[2026-05-26 08:28:09,084][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.06158342957496643, 'train/grad_norm_pre_clip': 0.39779648184776306, 'timing/data': 0.003944987431168556, 'timing/model': 0.7116911020129919, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.02}) +[2026-05-26 08:29:27,500][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.07623814046382904, 'train/grad_norm_pre_clip': 2.899852752685547, 'timing/data': 0.004230872727930546, 'timing/model': 0.5127884829416871, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.02}) +[2026-05-26 08:30:57,482][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.046330444514751434, 'train/grad_norm_pre_clip': 0.261353462934494, 'eval/loss': 0.16335934400558472, 'timing/data': 0.003985680639743805, 'timing/model': 0.5639460179954767, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.02}) +[2026-05-26 08:32:20,269][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.1353761851787567, 'train/grad_norm_pre_clip': 1.3242499828338623, 'timing/data': 0.004146856255829334, 'timing/model': 0.5010903347283602, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.02}) +[2026-05-26 08:33:41,044][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.20982998609542847, 'train/grad_norm_pre_clip': 1.140325903892517, 'timing/data': 0.004548543132841587, 'timing/model': 0.5325631201267242, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.02}) +[2026-05-26 08:34:59,982][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.14487764239311218, 'train/grad_norm_pre_clip': 0.62728351354599, 'timing/data': 0.003778764046728611, 'timing/model': 0.5470672463998199, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.02}) +[2026-05-26 08:36:19,468][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.42289426922798157, 'train/grad_norm_pre_clip': 3.669280767440796, 'timing/data': 0.0038831690326333046, 'timing/model': 0.565188274718821, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.02}) +[2026-05-26 08:37:38,420][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.06984058022499084, 'train/grad_norm_pre_clip': 0.8134029507637024, 'timing/data': 0.003594793379306793, 'timing/model': 0.49572026170790195, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.02}) +[2026-05-26 08:38:58,055][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.022398902103304863, 'train/grad_norm_pre_clip': 0.2533097267150879, 'timing/data': 0.004003169946372509, 'timing/model': 0.7111587608233094, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.02}) +[2026-05-26 08:40:16,667][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.2813815176486969, 'train/grad_norm_pre_clip': 1.4193425178527832, 'timing/data': 0.0042319223284721375, 'timing/model': 0.5168051095679402, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.03}) +[2026-05-26 08:41:36,119][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.16733992099761963, 'train/grad_norm_pre_clip': 1.612716794013977, 'timing/data': 0.003814270719885826, 'timing/model': 0.4965792475268245, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.03}) +[2026-05-26 08:42:55,987][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.12256749719381332, 'train/grad_norm_pre_clip': 2.3715760707855225, 'timing/data': 0.0038759885355830193, 'timing/model': 0.49645896162837744, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.03}) +[2026-05-26 08:44:25,056][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.38739633560180664, 'train/grad_norm_pre_clip': 2.886695623397827, 'eval/loss': 0.1362874060869217, 'timing/data': 0.005752282217144966, 'timing/model': 0.5597546780481935, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.03}) +[2026-05-26 08:45:44,341][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.15974602103233337, 'train/grad_norm_pre_clip': 1.3989992141723633, 'timing/data': 0.003927143290638924, 'timing/model': 0.49688272550702095, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.03}) +[2026-05-26 08:47:03,812][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.020674265921115875, 'train/grad_norm_pre_clip': 0.14069442451000214, 'timing/data': 0.004126278683543205, 'timing/model': 0.5286774467676878, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.03}) +[2026-05-26 08:48:23,095][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.028294140473008156, 'train/grad_norm_pre_clip': 0.2844236195087433, 'timing/data': 0.004000493325293064, 'timing/model': 0.5435085147619247, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.03}) +[2026-05-26 08:49:42,877][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.042044200003147125, 'train/grad_norm_pre_clip': 0.8999114632606506, 'timing/data': 0.004224419593811035, 'timing/model': 0.5524946926161647, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.03}) +[2026-05-26 08:51:02,616][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.17737218737602234, 'train/grad_norm_pre_clip': 0.7351735234260559, 'timing/data': 0.003968551754951477, 'timing/model': 0.4963957779109478, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.03}) +[2026-05-26 08:52:22,653][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.21474434435367584, 'train/grad_norm_pre_clip': 0.5342369675636292, 'timing/data': 0.004119608551263809, 'timing/model': 0.5081723034381866, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.03}) +[2026-05-26 08:53:41,778][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.24349842965602875, 'train/grad_norm_pre_clip': 3.058056116104126, 'timing/data': 0.004649716429412365, 'timing/model': 0.5314364684745669, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.03}) +[2026-05-26 08:55:00,717][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.19256433844566345, 'train/grad_norm_pre_clip': 1.9753245115280151, 'timing/data': 0.004124092869460583, 'timing/model': 0.7719426574185491, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.03}) +[2026-05-26 08:56:20,819][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.4891868531703949, 'train/grad_norm_pre_clip': 2.343325138092041, 'timing/data': 0.003642498515546322, 'timing/model': 0.5092520359903574, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 0.03}) +[2026-05-26 08:57:49,883][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.1297992318868637, 'train/grad_norm_pre_clip': 0.4769206643104553, 'eval/loss': 0.11890043318271637, 'timing/data': 0.0037610894069075584, 'timing/model': 0.5505313435569406, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 0.03}) +[2026-05-26 08:57:49,888][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_500_state +[2026-05-26 08:57:49,894][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-26 08:57:56,722][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_500_state/model.safetensors +[2026-05-26 08:58:12,610][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_500_state/optimizer.bin +[2026-05-26 08:58:12,611][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_500_state/sampler.bin +[2026-05-26 08:58:12,611][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_500_state/sampler_1.bin +[2026-05-26 08:58:12,613][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_500_state/random_states_0.pkl +[2026-05-26 08:58:20,736][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-05-26 08:58:20,736][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-26 08:58:20,746][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-26 08:59:40,671][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.019470801576972008, 'train/grad_norm_pre_clip': 0.3857884109020233, 'timing/data': 0.003951668739318848, 'timing/model': 0.49790035746991634, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 0.03}) +[2026-05-26 09:01:00,193][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.10655003786087036, 'train/grad_norm_pre_clip': 2.9421558380126953, 'timing/data': 0.003978992812335491, 'timing/model': 0.4991903258487582, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.04}) +[2026-05-26 09:02:19,554][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.027783889323472977, 'train/grad_norm_pre_clip': 1.211411476135254, 'timing/data': 0.004133906215429306, 'timing/model': 0.7573280278593302, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 0.04}) +[2026-05-26 09:03:38,644][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.24303515255451202, 'train/grad_norm_pre_clip': 2.0940535068511963, 'timing/data': 0.004073260352015495, 'timing/model': 0.49765998497605324, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 0.04}) +[2026-05-26 09:04:57,792][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.06581220030784607, 'train/grad_norm_pre_clip': 0.6416568160057068, 'timing/data': 0.003933153115212917, 'timing/model': 0.5610690722241998, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 0.04}) +[2026-05-26 09:06:17,556][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.13208845257759094, 'train/grad_norm_pre_clip': 1.318049430847168, 'timing/data': 0.004328184761106968, 'timing/model': 0.7913344567641616, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 0.04}) +[2026-05-26 09:07:37,148][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.001700556487776339, 'train/grad_norm_pre_clip': 0.05258292332291603, 'timing/data': 0.003695874474942684, 'timing/model': 0.5396332601085305, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 0.04}) +[2026-05-26 09:08:56,670][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.1568354368209839, 'train/grad_norm_pre_clip': 1.7803596258163452, 'timing/data': 0.0039006471633911133, 'timing/model': 0.5522852530702949, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 0.04}) +[2026-05-26 09:10:16,067][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.2811887562274933, 'train/grad_norm_pre_clip': 1.1672859191894531, 'timing/data': 0.004399843513965607, 'timing/model': 0.765328673645854, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.04}) +[2026-05-26 09:11:46,121][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.007104563992470503, 'train/grad_norm_pre_clip': 0.11993668228387833, 'eval/loss': 0.12620626389980316, 'timing/data': 0.0038177836686372757, 'timing/model': 0.5395747199654579, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 0.04}) +[2026-05-26 09:13:06,023][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.4963334798812866, 'train/grad_norm_pre_clip': 3.814992904663086, 'timing/data': 0.0042075514793396, 'timing/model': 0.49663920514285564, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 0.04}) +[2026-05-26 09:14:25,834][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.08087039738893509, 'train/grad_norm_pre_clip': 1.1072050333023071, 'timing/data': 0.0038816332817077637, 'timing/model': 0.7329482892528176, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 0.04}) +[2026-05-26 09:15:45,790][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.07720543444156647, 'train/grad_norm_pre_clip': 0.6130288243293762, 'timing/data': 0.0049316296353936195, 'timing/model': 0.5747315073385835, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 0.04}) +[2026-05-26 09:17:05,117][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.14156293869018555, 'train/grad_norm_pre_clip': 1.0472116470336914, 'timing/data': 0.004057810641825199, 'timing/model': 0.5057675084099174, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 0.04}) +[2026-05-26 09:18:24,941][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.008954665623605251, 'train/grad_norm_pre_clip': 0.1529887169599533, 'timing/data': 0.004596928134560585, 'timing/model': 0.5002838429063559, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 0.04}) +[2026-05-26 09:19:44,392][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.1807030439376831, 'train/grad_norm_pre_clip': 4.975417137145996, 'timing/data': 0.004110181704163551, 'timing/model': 0.536219434812665, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.04}) +[2026-05-26 09:21:04,208][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.00018461624858900905, 'train/grad_norm_pre_clip': 0.003077493514865637, 'timing/data': 0.00439910963177681, 'timing/model': 0.5548538286238909, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 0.05}) +[2026-05-26 09:22:23,174][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.06561293452978134, 'train/grad_norm_pre_clip': 0.5612679719924927, 'timing/data': 0.0038812244310975075, 'timing/model': 0.563090824522078, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 0.05}) +[2026-05-26 09:23:41,503][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.14965300261974335, 'train/grad_norm_pre_clip': 1.1252177953720093, 'timing/data': 0.0066708046942949295, 'timing/model': 0.5616702865809202, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 0.05}) +[2026-05-26 09:25:10,616][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.04839174076914787, 'train/grad_norm_pre_clip': 0.7843582034111023, 'eval/loss': 0.12449022382497787, 'timing/data': 0.0038909707218408585, 'timing/model': 0.5492948200553656, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 0.05}) +[2026-05-26 09:26:30,393][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.2433772087097168, 'train/grad_norm_pre_clip': 1.3834885358810425, 'timing/data': 0.004203788936138153, 'timing/model': 0.5708390148356557, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 0.05}) +[2026-05-26 09:27:49,385][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.31217607855796814, 'train/grad_norm_pre_clip': 2.6308248043060303, 'timing/data': 0.00409252755343914, 'timing/model': 0.5508954310789704, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 0.05}) +[2026-05-26 09:29:08,683][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.05431322380900383, 'train/grad_norm_pre_clip': 0.7311633229255676, 'timing/data': 0.004526272416114807, 'timing/model': 0.49604493845254183, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.05}) +[2026-05-26 09:30:27,487][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.3915156126022339, 'train/grad_norm_pre_clip': 3.2486696243286133, 'timing/data': 0.0037284940481185913, 'timing/model': 0.4972931854426861, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 0.05}) +[2026-05-26 09:31:46,766][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.09471841156482697, 'train/grad_norm_pre_clip': 0.9579676389694214, 'timing/data': 0.004177200607955456, 'timing/model': 0.7393157640472054, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 0.05}) +[2026-05-26 09:31:46,778][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_750_state +[2026-05-26 09:31:46,782][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-26 09:31:53,292][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_750_state/model.safetensors +[2026-05-26 09:32:09,044][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_750_state/optimizer.bin +[2026-05-26 09:32:09,044][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_750_state/sampler.bin +[2026-05-26 09:32:09,045][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_750_state/sampler_1.bin +[2026-05-26 09:32:09,046][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_750_state/random_states_0.pkl +[2026-05-26 09:32:17,581][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 750 +[2026-05-26 09:32:17,582][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-26 09:32:17,590][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-26 09:33:36,357][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.27125421166419983, 'train/grad_norm_pre_clip': 1.5385689735412598, 'timing/data': 0.004151752218604088, 'timing/model': 0.5296414522454143, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 0.05}) +[2026-05-26 09:34:54,781][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.007422592956572771, 'train/grad_norm_pre_clip': 0.042639002203941345, 'timing/data': 0.004016782157123089, 'timing/model': 0.5490095438435674, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 0.05}) +[2026-05-26 09:36:15,719][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.10537774860858917, 'train/grad_norm_pre_clip': 0.8017721772193909, 'timing/data': 0.004012721590697765, 'timing/model': 0.7837101919576526, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 0.05}) +[2026-05-26 09:37:35,881][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.005650143139064312, 'train/grad_norm_pre_clip': 0.0626705139875412, 'timing/data': 0.0038303984329104424, 'timing/model': 0.4990748008713126, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 0.05}) +[2026-05-26 09:39:06,929][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.15094612538814545, 'train/grad_norm_pre_clip': 1.2797229290008545, 'eval/loss': 0.1253405660390854, 'timing/data': 0.0042110104113817215, 'timing/model': 0.5374370068311691, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.05}) +[2026-05-26 09:40:26,374][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.08807764202356339, 'train/grad_norm_pre_clip': 0.9946224093437195, 'timing/data': 0.004093563184142113, 'timing/model': 0.5244857929646969, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 0.06}) +[2026-05-26 09:41:45,930][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.17994491755962372, 'train/grad_norm_pre_clip': 0.7357081770896912, 'timing/data': 0.004132774658501148, 'timing/model': 0.5169045682996511, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 0.06}) +[2026-05-26 09:43:04,870][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.02526259981095791, 'train/grad_norm_pre_clip': 0.6682041883468628, 'timing/data': 0.0041654231026768684, 'timing/model': 0.5521397134289145, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 0.06}) +[2026-05-26 09:44:24,375][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.2514254152774811, 'train/grad_norm_pre_clip': 1.3039147853851318, 'timing/data': 0.004187570884823799, 'timing/model': 0.5371395340189338, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.06}) +[2026-05-26 09:45:44,932][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.041396114975214005, 'train/grad_norm_pre_clip': 0.7070712447166443, 'timing/data': 0.00402796920388937, 'timing/model': 0.5246950900182128, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.06}) +[2026-05-26 09:47:05,092][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.09390366822481155, 'train/grad_norm_pre_clip': 0.5768288969993591, 'timing/data': 0.004463380202651024, 'timing/model': 0.49667578283697367, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.06}) +[2026-05-26 09:48:24,579][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.009218081831932068, 'train/grad_norm_pre_clip': 0.15168052911758423, 'timing/data': 0.003989233635365963, 'timing/model': 0.5471083363518119, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.06}) +[2026-05-26 09:49:44,701][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.09430288523435593, 'train/grad_norm_pre_clip': 0.42778146266937256, 'timing/data': 0.004103883169591427, 'timing/model': 0.5214960472658277, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 0.06}) +[2026-05-26 09:51:05,457][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.06579045206308365, 'train/grad_norm_pre_clip': 0.3507712781429291, 'timing/data': 0.0038324054330587387, 'timing/model': 0.5151273626834154, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 0.06}) +[2026-05-26 09:52:35,866][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.09178592264652252, 'train/grad_norm_pre_clip': 0.3414883613586426, 'eval/loss': 0.12143383175134659, 'timing/data': 0.0043898457661271095, 'timing/model': 0.7103003421798348, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 0.06}) +[2026-05-26 09:53:56,581][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.05009954050183296, 'train/grad_norm_pre_clip': 0.48308977484703064, 'timing/data': 0.005004324018955231, 'timing/model': 0.5392832616344094, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 0.06}) +[2026-05-26 09:55:17,180][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.414467990398407, 'train/grad_norm_pre_clip': 3.6635582447052, 'timing/data': 0.004112081602215767, 'timing/model': 0.5014444887638092, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 0.06}) +[2026-05-26 09:56:42,406][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.022642333060503006, 'train/grad_norm_pre_clip': 0.15520250797271729, 'timing/data': 0.0049571506679058075, 'timing/model': 0.5510927885770798, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 0.06}) +[2026-05-26 09:58:03,659][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.4585546553134918, 'train/grad_norm_pre_clip': 2.7188029289245605, 'timing/data': 0.004063134081661701, 'timing/model': 0.503351497463882, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.06}) +[2026-05-26 09:59:25,167][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.14963747560977936, 'train/grad_norm_pre_clip': 0.8949342966079712, 'timing/data': 0.007818005979061127, 'timing/model': 0.5260441647842526, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 0.06}) +[2026-05-26 10:00:46,846][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.024568576365709305, 'train/grad_norm_pre_clip': 0.41733601689338684, 'timing/data': 0.007738773711025715, 'timing/model': 0.5023491997271776, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 0.07}) +[2026-05-26 10:02:08,205][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.09584721177816391, 'train/grad_norm_pre_clip': 1.0009360313415527, 'timing/data': 0.0073473816737532616, 'timing/model': 0.753379188477993, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 0.07}) +[2026-05-26 10:03:28,872][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.03983879089355469, 'train/grad_norm_pre_clip': 0.36288896203041077, 'timing/data': 0.008837992325425148, 'timing/model': 0.5236019259318709, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 0.07}) +[2026-05-26 10:04:49,467][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.34110790491104126, 'train/grad_norm_pre_clip': 1.2812774181365967, 'timing/data': 0.007431915961205959, 'timing/model': 0.5258826157078147, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 0.07}) +[2026-05-26 10:06:31,374][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-05-26 10:06:31,375][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.09091678261756897, 'train/grad_norm_pre_clip': 0.4705354869365692, 'eval/loss': 0.1280578374862671, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 22.40000000000006, 'rl_games_eval/mid_train/mean_length': 199.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 22.40000000000006, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 199.0, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 1.0, 'timing/data': 0.009684495627880096, 'timing/model': 0.5565458191558719, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 0.07}) +[2026-05-26 10:06:31,375][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1000_state +[2026-05-26 10:06:31,392][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-26 10:06:38,087][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1000_state/model.safetensors +[2026-05-26 10:06:54,864][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1000_state/optimizer.bin +[2026-05-26 10:06:54,865][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1000_state/sampler.bin +[2026-05-26 10:06:54,866][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1000_state/sampler_1.bin +[2026-05-26 10:06:54,867][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1000_state/random_states_0.pkl +[2026-05-26 10:07:03,240][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-05-26 10:07:03,240][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-26 10:07:03,254][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-26 10:08:25,316][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.2861561179161072, 'train/grad_norm_pre_clip': 2.320300340652466, 'timing/data': 0.004275647923350334, 'timing/model': 0.7801696071401238, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.07}) +[2026-05-26 10:09:47,244][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.07741370797157288, 'train/grad_norm_pre_clip': 0.6264877915382385, 'timing/data': 0.0046070646494627, 'timing/model': 0.5197579422965646, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 0.07}) +[2026-05-26 10:11:08,699][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.03431391343474388, 'train/grad_norm_pre_clip': 0.5340031385421753, 'timing/data': 0.004201917909085751, 'timing/model': 0.5064275730401278, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 0.07}) +[2026-05-26 10:12:30,477][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.31940531730651855, 'train/grad_norm_pre_clip': 1.2846393585205078, 'timing/data': 0.004133502021431923, 'timing/model': 0.5032820971682668, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 0.07}) +[2026-05-26 10:13:52,085][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.15412108600139618, 'train/grad_norm_pre_clip': 1.177876591682434, 'timing/data': 0.004027864895761013, 'timing/model': 0.5402848264202476, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 0.07}) +[2026-05-26 10:15:13,600][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.2261936515569687, 'train/grad_norm_pre_clip': 0.8506044149398804, 'timing/data': 0.003991592675447464, 'timing/model': 0.5381254246458411, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 0.07}) +[2026-05-26 10:16:36,048][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.00523265078663826, 'train/grad_norm_pre_clip': 0.1747019737958908, 'timing/data': 0.004969580098986626, 'timing/model': 0.5437718080356717, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 0.07}) +[2026-05-26 10:17:57,159][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.17140744626522064, 'train/grad_norm_pre_clip': 1.3841755390167236, 'timing/data': 0.005428990349173546, 'timing/model': 0.5314281536266208, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.07}) +[2026-05-26 10:19:18,660][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.02142583206295967, 'train/grad_norm_pre_clip': 0.11388440430164337, 'timing/data': 0.004341638647019863, 'timing/model': 0.5304808793589473, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 0.07}) +[2026-05-26 10:20:50,962][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.008619531989097595, 'train/grad_norm_pre_clip': 0.14237701892852783, 'eval/loss': 0.1183861643075943, 'timing/data': 0.004168090410530567, 'timing/model': 0.5022867638617754, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 0.07}) +[2026-05-26 10:22:12,599][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.09505855292081833, 'train/grad_norm_pre_clip': 0.9879469275474548, 'timing/data': 0.0046573299914598465, 'timing/model': 0.5566081842407584, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 0.08}) +[2026-05-26 10:23:34,278][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.06648661941289902, 'train/grad_norm_pre_clip': 0.5650452375411987, 'timing/data': 0.004150683991611004, 'timing/model': 0.5505322962999344, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 0.08}) +[2026-05-26 10:24:55,492][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.31721818447113037, 'train/grad_norm_pre_clip': 2.2297046184539795, 'timing/data': 0.004520888440310955, 'timing/model': 0.5027230903506279, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 0.08}) +[2026-05-26 13:05:10,866][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-05-26 13:05:10,984][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-05-26 13:05:25,744][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-05-26 13:05:35,355][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-05-26 13:05:35,355][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-05-26 13:05:35,355][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-05-26 13:05:35,355][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-05-26 13:05:35,355][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-05-26 13:05:35,356][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-05-26 13:05:35,356][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-05-26 13:05:35,356][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-05-26 13:05:35,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-05-26 13:05:35,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-05-26 13:05:35,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-05-26 13:05:35,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-05-26 13:05:35,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-05-26 13:05:35,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-05-26 13:05:35,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-05-26 13:05:35,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-05-26 13:05:35,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-05-26 13:05:35,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-05-26 13:05:35,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-05-26 13:05:35,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-05-26 13:05:35,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-05-26 13:05:35,373][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-05-26 13:05:35,373][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-05-26 13:05:35,373][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-05-26 13:05:35,373][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-05-26 13:05:35,373][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-05-26 13:05:35,373][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-05-26 13:05:35,373][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-05-26 13:05:35,373][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-05-26 13:05:35,373][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-05-26 13:05:35,373][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-05-26 13:05:35,447][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-05-26 13:05:35,697][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/config.full.yaml +[2026-05-26 13:05:35,708][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/config.yaml +[2026-05-26 13:06:35,269][starVLA.training.train_starvla][INFO] - Loaded pretrained checkpoint: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt, steps: 0 +[2026-05-26 13:06:43,409][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-05-26 13:06:43,409][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-05-26 13:06:43,409][starVLA.training.train_starvla][INFO] - Per device batch size = 16 +[2026-05-26 13:06:43,410][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-05-26 13:06:43,410][starVLA.training.train_starvla][INFO] - Total batch size = 256 +[2026-05-26 13:07:56,473][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-05-26 13:07:56,548][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-05-26 13:07:58,157][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-05-26 13:08:06,915][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-05-26 13:08:06,916][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-05-26 13:08:06,916][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-05-26 13:08:06,916][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-05-26 13:08:06,916][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-05-26 13:08:06,916][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-05-26 13:08:06,916][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-05-26 13:08:06,916][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-05-26 13:08:06,917][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-05-26 13:08:06,917][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-05-26 13:08:06,917][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-05-26 13:08:06,917][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-05-26 13:08:06,919][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-05-26 13:08:06,919][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-05-26 13:08:06,919][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-05-26 13:08:06,919][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-05-26 13:08:06,919][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-05-26 13:08:06,919][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-05-26 13:08:06,919][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-05-26 13:08:06,920][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-05-26 13:08:06,920][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-05-26 13:08:06,920][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-05-26 13:08:06,920][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-05-26 13:08:06,920][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-05-26 13:08:06,920][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-05-26 13:08:06,920][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-05-26 13:08:06,920][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-05-26 13:08:06,920][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-05-26 13:08:06,921][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-05-26 13:08:06,921][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-05-26 13:08:06,921][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-05-26 13:08:06,964][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-05-26 13:08:07,008][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/config.full.yaml +[2026-05-26 13:08:07,016][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/config.yaml +[2026-05-26 13:09:02,118][starVLA.training.train_starvla][INFO] - Loaded pretrained checkpoint: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt, steps: 0 +[2026-05-26 13:09:05,435][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-05-26 13:09:05,436][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-05-26 13:09:05,436][starVLA.training.train_starvla][INFO] - Per device batch size = 16 +[2026-05-26 13:09:05,436][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-05-26 13:09:05,436][starVLA.training.train_starvla][INFO] - Total batch size = 256 +[2026-05-26 13:10:18,038][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.3155343532562256, 'train/grad_norm_pre_clip': 1.7430894374847412, 'timing/data': 0.004233149811625481, 'timing/model': 0.4685528315603733, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-05-26 13:11:22,959][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.022467855364084244, 'train/grad_norm_pre_clip': 0.14473594725131989, 'timing/data': 0.0038739657029509544, 'timing/model': 0.4843164458870888, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.0}) +[2026-05-26 13:15:30,167][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-05-26 13:15:30,273][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-05-26 13:15:31,901][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-05-26 13:15:40,689][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-05-26 13:15:40,689][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-05-26 13:15:40,689][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-05-26 13:15:40,689][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-05-26 13:15:40,690][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-05-26 13:15:40,690][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-05-26 13:15:40,690][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-05-26 13:15:40,690][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-05-26 13:15:40,690][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-05-26 13:15:40,690][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-05-26 13:15:40,690][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-05-26 13:15:40,690][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-05-26 13:15:40,690][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-05-26 13:15:40,723][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-05-26 13:15:40,723][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-05-26 13:15:40,723][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-05-26 13:15:40,723][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-05-26 13:15:40,723][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-05-26 13:15:40,723][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-05-26 13:15:40,724][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-05-26 13:15:40,724][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-05-26 13:15:40,724][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-05-26 13:15:40,724][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-05-26 13:15:40,724][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-05-26 13:15:40,724][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-05-26 13:15:40,724][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-05-26 13:15:40,724][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-05-26 13:15:40,724][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-05-26 13:15:40,725][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-05-26 13:15:40,725][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-05-26 13:15:40,725][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-05-26 13:15:40,773][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-05-26 13:15:40,815][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/config.full.yaml +[2026-05-26 13:15:40,824][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/config.yaml +[2026-05-26 13:16:35,711][starVLA.training.train_starvla][INFO] - Loaded pretrained checkpoint: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt, steps: 0 +[2026-05-26 13:16:38,967][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-05-26 13:16:38,967][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-05-26 13:16:38,968][starVLA.training.train_starvla][INFO] - Per device batch size = 16 +[2026-05-26 13:16:38,968][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-05-26 13:16:38,968][starVLA.training.train_starvla][INFO] - Total batch size = 256 +[2026-05-26 13:17:51,770][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.3155343532562256, 'train/grad_norm_pre_clip': 1.7430894374847412, 'timing/data': 0.004639709368348122, 'timing/model': 0.4579070471227169, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-05-26 13:18:56,925][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.022467855364084244, 'train/grad_norm_pre_clip': 0.14473594725131989, 'timing/data': 0.003932284191250801, 'timing/model': 0.4586984347552061, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.0}) +[2026-05-26 13:20:13,977][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-05-26 13:20:14,075][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-05-26 13:20:15,715][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-05-26 13:20:24,488][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-05-26 13:20:24,488][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-05-26 13:20:24,488][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-05-26 13:20:24,488][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-05-26 13:20:24,488][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-05-26 13:20:24,489][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-05-26 13:20:24,489][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-05-26 13:20:24,489][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-05-26 13:20:24,489][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-05-26 13:20:24,489][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-05-26 13:20:24,489][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-05-26 13:20:24,489][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-05-26 13:20:24,489][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-05-26 13:20:24,491][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-05-26 13:20:24,491][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-05-26 13:20:24,491][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-05-26 13:20:24,492][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-05-26 13:20:24,492][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-05-26 13:20:24,492][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-05-26 13:20:24,492][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-05-26 13:20:24,492][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-05-26 13:20:24,492][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-05-26 13:20:24,492][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-05-26 13:20:24,492][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-05-26 13:20:24,493][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-05-26 13:20:24,493][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-05-26 13:20:24,493][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-05-26 13:20:24,493][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-05-26 13:20:24,493][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-05-26 13:20:24,493][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-05-26 13:20:24,493][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-05-26 13:20:24,539][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-05-26 13:20:24,581][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/config.full.yaml +[2026-05-26 13:20:24,590][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/config.yaml +[2026-05-26 13:20:24,590][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1000_state, steps: 1000 +[2026-05-26 13:20:30,393][accelerate.accelerator][INFO] - Loading states from /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1000_state +[2026-05-26 13:21:01,956][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-05-26 13:21:56,296][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-05-26 13:21:56,297][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-05-26 13:21:56,297][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-05-26 13:21:56,410][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-05-26 13:21:57,345][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-05-26 13:21:58,469][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-05-26 13:21:58,470][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-05-26 13:21:58,470][starVLA.training.train_starvla][INFO] - Per device batch size = 16 +[2026-05-26 13:21:58,470][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-05-26 13:21:58,470][starVLA.training.train_starvla][INFO] - Total batch size = 256 +[2026-05-26 13:23:10,264][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.12530088424682617, 'train/grad_norm_pre_clip': 0.8956109881401062, 'timing/data': 0.003946899436414242, 'timing/model': 0.4475328018888831, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.07}) +[2026-05-26 13:24:15,577][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.0706133171916008, 'train/grad_norm_pre_clip': 1.26688551902771, 'timing/data': 0.003937281668186188, 'timing/model': 0.44994831178337336, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.07}) +[2026-05-26 13:25:21,130][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.03564552590250969, 'train/grad_norm_pre_clip': 0.3005739748477936, 'timing/data': 0.004945579916238785, 'timing/model': 0.5011690016835928, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.07}) +[2026-05-26 13:26:26,475][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.12443739175796509, 'train/grad_norm_pre_clip': 0.5988463759422302, 'timing/data': 0.004434742033481598, 'timing/model': 0.4511108184233308, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.07}) +[2026-05-26 13:27:31,781][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.03778431937098503, 'train/grad_norm_pre_clip': 0.3673531711101532, 'timing/data': 0.004024907015264034, 'timing/model': 0.4656036952510476, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.07}) +[2026-05-26 13:28:37,282][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.09111738204956055, 'train/grad_norm_pre_clip': 1.0575031042099, 'timing/data': 0.004086780361831188, 'timing/model': 0.4576736632734537, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.07}) +[2026-05-26 13:29:42,813][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.17886227369308472, 'train/grad_norm_pre_clip': 0.871684730052948, 'timing/data': 0.0040107471868395805, 'timing/model': 0.43891969602555037, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.07}) +[2026-05-26 13:30:48,276][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.12924820184707642, 'train/grad_norm_pre_clip': 2.1781113147735596, 'timing/data': 0.003943938761949539, 'timing/model': 0.6622385494410992, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.07}) +[2026-05-26 13:31:53,373][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.1649368405342102, 'train/grad_norm_pre_clip': 1.3228106498718262, 'timing/data': 0.003975352272391319, 'timing/model': 0.46254244819283485, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.07}) +[2026-05-26 13:33:09,325][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.16119080781936646, 'train/grad_norm_pre_clip': 0.5853996872901917, 'eval/loss': 0.11383805423974991, 'timing/data': 0.003926695324480534, 'timing/model': 0.4546390315517783, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.07}) +[2026-05-26 13:34:46,710][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.20156322419643402, 'train/grad_norm_pre_clip': 0.75657719373703, 'timing/data': 0.0037874970585107803, 'timing/model': 0.6322342120110989, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.08}) +[2026-05-26 13:36:24,982][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.02868678979575634, 'train/grad_norm_pre_clip': 0.6453790664672852, 'timing/data': 0.00393119640648365, 'timing/model': 0.8168843630701303, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.08}) +[2026-05-26 13:38:02,928][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.1708827018737793, 'train/grad_norm_pre_clip': 0.42712998390197754, 'timing/data': 0.003710143268108368, 'timing/model': 0.5775023242458701, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.08}) +[2026-05-26 13:39:40,712][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.24098913371562958, 'train/grad_norm_pre_clip': 2.7522194385528564, 'timing/data': 0.003656497225165367, 'timing/model': 0.6494722403585911, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.08}) +[2026-05-26 13:41:18,527][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.1336219608783722, 'train/grad_norm_pre_clip': 1.195237636566162, 'timing/data': 0.00377848744392395, 'timing/model': 0.6583342524245381, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.08}) +[2026-05-26 13:42:55,721][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.11947321891784668, 'train/grad_norm_pre_clip': 0.5851786136627197, 'timing/data': 0.004254136234521866, 'timing/model': 0.658942005597055, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.08}) +[2026-05-26 13:44:32,803][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.21697786450386047, 'train/grad_norm_pre_clip': 0.6109973788261414, 'timing/data': 0.00520307756960392, 'timing/model': 0.8198520066216588, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.08}) +[2026-05-26 13:46:10,419][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.013832948170602322, 'train/grad_norm_pre_clip': 0.07270673662424088, 'timing/data': 0.003843046724796295, 'timing/model': 0.6474807057529688, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.08}) +[2026-05-26 13:47:48,512][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.17936138808727264, 'train/grad_norm_pre_clip': 1.1433542966842651, 'timing/data': 0.0037467805668711662, 'timing/model': 0.8483407096937299, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.08}) +[2026-05-26 13:49:36,800][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.24874000251293182, 'train/grad_norm_pre_clip': 2.74947190284729, 'eval/loss': 0.18039077520370483, 'timing/data': 0.0038697998970746994, 'timing/model': 0.5750863533467054, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.08}) +[2026-05-26 13:51:14,521][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.09513694792985916, 'train/grad_norm_pre_clip': 0.7988412380218506, 'timing/data': 0.003727351315319538, 'timing/model': 0.65290074236691, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.08}) +[2026-05-26 13:52:52,821][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.23209166526794434, 'train/grad_norm_pre_clip': 1.2082056999206543, 'timing/data': 0.003811720758676529, 'timing/model': 0.6526486435905099, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.08}) +[2026-05-26 13:54:30,721][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.040144313126802444, 'train/grad_norm_pre_clip': 0.7299323081970215, 'timing/data': 0.0038039200007915497, 'timing/model': 0.654966983012855, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.08}) +[2026-05-26 13:56:09,300][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.14260128140449524, 'train/grad_norm_pre_clip': 0.6205460429191589, 'timing/data': 0.003973992541432381, 'timing/model': 0.8333701435476542, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.08}) +[2026-05-26 13:57:46,227][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.0805782750248909, 'train/grad_norm_pre_clip': 0.4402412474155426, 'timing/data': 0.004512414336204529, 'timing/model': 0.5762859284877777, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.08}) +[2026-05-26 13:57:46,230][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1250_state +[2026-05-26 13:57:46,235][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-26 13:57:53,911][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1250_state/model.safetensors +[2026-05-26 13:58:11,891][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1250_state/optimizer.bin +[2026-05-26 13:58:11,892][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1250_state/sampler.bin +[2026-05-26 13:58:11,892][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1250_state/sampler_1.bin +[2026-05-26 13:58:11,916][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1250_state/random_states_0.pkl +[2026-05-26 13:58:14,224][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpr2n17y38/test.c -o /tmp/tmpr2n17y38/test.o +[2026-05-26 13:58:14,259][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpr2n17y38/test.o -laio -o /tmp/tmpr2n17y38/a.out +[2026-05-26 13:58:14,866][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpw7p58rnw/test.c -o /tmp/tmpw7p58rnw/test.o +[2026-05-26 13:58:14,882][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpw7p58rnw/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpw7p58rnw/a.out +[2026-05-26 13:58:14,927][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpxga8f_39/test.c -o /tmp/tmpxga8f_39/test.o +[2026-05-26 13:58:14,941][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpxga8f_39/test.o -laio -o /tmp/tmpxga8f_39/a.out +[2026-05-26 13:58:32,037][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1250 +[2026-05-26 13:58:32,038][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-26 13:58:32,052][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-26 14:00:08,234][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.24323444068431854, 'train/grad_norm_pre_clip': 3.398634672164917, 'timing/data': 0.003765133209526539, 'timing/model': 0.5768013587221503, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.09}) +[2026-05-26 14:01:45,218][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.17720286548137665, 'train/grad_norm_pre_clip': 2.7975244522094727, 'timing/data': 0.003975252620875835, 'timing/model': 0.6433567209169269, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.09}) +[2026-05-26 14:03:22,215][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.024130288511514664, 'train/grad_norm_pre_clip': 0.19923906028270721, 'timing/data': 0.004006672650575638, 'timing/model': 0.6470796717330813, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.09}) +[2026-05-26 14:04:57,813][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.21673594415187836, 'train/grad_norm_pre_clip': 0.6661479473114014, 'timing/data': 0.0038703782483935356, 'timing/model': 0.6464069141075015, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.09}) +[2026-05-26 14:06:45,139][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.05964196100831032, 'train/grad_norm_pre_clip': 1.1269762516021729, 'eval/loss': 0.10978851467370987, 'timing/data': 0.0037717362865805626, 'timing/model': 0.6467393599450588, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.09}) +[2026-05-26 14:08:21,835][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.14467090368270874, 'train/grad_norm_pre_clip': 2.729633092880249, 'timing/data': 0.003925018943846226, 'timing/model': 0.5742944534868002, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.09}) +[2026-05-26 14:09:57,611][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.06795206665992737, 'train/grad_norm_pre_clip': 0.32503652572631836, 'timing/data': 0.004742828197777271, 'timing/model': 0.6089328266680241, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.09}) +[2026-05-26 14:11:33,527][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.13416427373886108, 'train/grad_norm_pre_clip': 1.4913575649261475, 'timing/data': 0.004350935108959675, 'timing/model': 0.5797483706846833, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.09}) +[2026-05-26 14:13:08,251][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.31206125020980835, 'train/grad_norm_pre_clip': 1.9533758163452148, 'timing/data': 0.0038235802203416824, 'timing/model': 0.5890507381409407, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.09}) +[2026-05-26 14:14:43,968][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.034133996814489365, 'train/grad_norm_pre_clip': 0.31771525740623474, 'timing/data': 0.0038176141679286957, 'timing/model': 0.8324853703379631, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.09}) +[2026-05-26 14:16:19,427][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.019078683108091354, 'train/grad_norm_pre_clip': 0.38878297805786133, 'timing/data': 0.004065587185323238, 'timing/model': 0.5698696477338672, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.09}) +[2026-05-26 14:17:56,108][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.1572445183992386, 'train/grad_norm_pre_clip': 1.4368879795074463, 'timing/data': 0.003953173756599426, 'timing/model': 0.6376087348908186, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.09}) +[2026-05-26 14:19:33,114][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.07426291704177856, 'train/grad_norm_pre_clip': 0.6743367314338684, 'timing/data': 0.003987375646829605, 'timing/model': 0.6499517261981964, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.09}) +[2026-05-26 14:21:09,804][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.11275842040777206, 'train/grad_norm_pre_clip': 0.9349715709686279, 'timing/data': 0.004458594135940075, 'timing/model': 0.6320379935204983, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.09}) +[2026-05-26 14:22:56,745][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.14381089806556702, 'train/grad_norm_pre_clip': 0.6477302312850952, 'eval/loss': 0.12321539223194122, 'timing/data': 0.0049488740041852, 'timing/model': 0.6455166786909103, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.1}) +[2026-05-26 14:24:33,723][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.07258106768131256, 'train/grad_norm_pre_clip': 1.551932454109192, 'timing/data': 0.004385512322187424, 'timing/model': 0.6570204244926572, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.1}) +[2026-05-26 14:26:11,836][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.012136630713939667, 'train/grad_norm_pre_clip': 0.10030519962310791, 'timing/data': 0.004169113002717495, 'timing/model': 0.629545858129859, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.1}) +[2026-05-26 15:01:44,147][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-05-26 15:01:44,274][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-05-26 15:02:02,928][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-05-26 15:02:12,375][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-05-26 15:02:12,375][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-05-26 15:02:12,375][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-05-26 15:02:12,375][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-05-26 15:02:12,375][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-05-26 15:02:12,375][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-05-26 15:02:12,375][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-05-26 15:02:12,376][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-05-26 15:02:12,376][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-05-26 15:02:12,376][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-05-26 15:02:12,376][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-05-26 15:02:12,376][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-05-26 15:02:12,376][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-05-26 15:02:12,376][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-05-26 15:02:12,376][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-05-26 15:02:12,376][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-05-26 15:02:12,377][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-05-26 15:02:12,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-05-26 15:02:12,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-05-26 15:02:12,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-05-26 15:02:12,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-05-26 15:02:12,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-05-26 15:02:12,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-05-26 15:02:12,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-05-26 15:02:12,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-05-26 15:02:12,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-05-26 15:02:12,380][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-05-26 15:02:12,380][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-05-26 15:02:12,380][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-05-26 15:02:12,380][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-05-26 15:02:12,380][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-05-26 15:02:12,444][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-05-26 15:02:12,719][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/config.full.yaml +[2026-05-26 15:02:12,728][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/config.yaml +[2026-05-26 15:02:12,729][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1250_state, steps: 1250 +[2026-05-26 15:02:32,424][accelerate.accelerator][INFO] - Loading states from /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1250_state +[2026-05-26 15:03:06,020][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-05-26 15:03:59,204][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-05-26 15:03:59,204][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-05-26 15:03:59,206][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-05-26 15:03:59,389][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-05-26 15:04:00,704][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-05-26 15:04:06,233][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-05-26 15:04:06,234][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-05-26 15:04:06,234][starVLA.training.train_starvla][INFO] - Per device batch size = 16 +[2026-05-26 15:04:06,235][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-05-26 15:04:06,235][starVLA.training.train_starvla][INFO] - Total batch size = 256 +[2026-05-26 15:05:27,619][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.1460878700017929, 'train/grad_norm_pre_clip': 0.7557454109191895, 'timing/data': 0.00629444420337677, 'timing/model': 0.4523295257240534, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.09}) +[2026-05-26 15:06:33,549][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.0937371626496315, 'train/grad_norm_pre_clip': 0.4982316493988037, 'timing/data': 0.005167611874639988, 'timing/model': 0.4802210032939911, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.09}) +[2026-05-26 15:07:38,913][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.056281134486198425, 'train/grad_norm_pre_clip': 1.1510417461395264, 'timing/data': 0.004259251989424229, 'timing/model': 0.4635118916630745, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.09}) +[2026-05-26 15:08:44,531][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.22337336838245392, 'train/grad_norm_pre_clip': 1.9122518301010132, 'timing/data': 0.004969533532857895, 'timing/model': 0.46355482563376427, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.09}) +[2026-05-26 15:10:04,675][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.0707610696554184, 'train/grad_norm_pre_clip': 0.4338107109069824, 'eval/loss': 0.12347524613142014, 'timing/data': 0.004537355154752731, 'timing/model': 0.4432106912136078, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.09}) +[2026-05-26 15:11:44,667][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.09569746255874634, 'train/grad_norm_pre_clip': 0.9644652605056763, 'timing/data': 0.004552464932203293, 'timing/model': 0.6468999832868576, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.09}) +[2026-05-26 15:13:21,767][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.18017959594726562, 'train/grad_norm_pre_clip': 0.3619312644004822, 'timing/data': 0.004264820367097855, 'timing/model': 0.5970831690356135, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.09}) +[2026-05-26 15:14:58,472][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.07902946323156357, 'train/grad_norm_pre_clip': 2.254368305206299, 'timing/data': 0.004746962338685989, 'timing/model': 0.6448031887412071, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.09}) +[2026-05-26 15:16:38,625][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.11651989817619324, 'train/grad_norm_pre_clip': 0.779533863067627, 'timing/data': 0.004471917636692524, 'timing/model': 0.7877876935526729, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.09}) +[2026-05-26 15:18:19,461][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.13105808198451996, 'train/grad_norm_pre_clip': 0.7243992686271667, 'timing/data': 0.0043398356065154076, 'timing/model': 0.5791740911081433, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.09}) +[2026-05-26 15:19:56,288][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.07578997313976288, 'train/grad_norm_pre_clip': 0.5890109539031982, 'timing/data': 0.0041236188262701035, 'timing/model': 0.6612438168376684, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.09}) +[2026-05-26 15:21:33,497][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.010804549790918827, 'train/grad_norm_pre_clip': 0.06023307889699936, 'timing/data': 0.004257808439433575, 'timing/model': 0.6868947567418218, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.09}) +[2026-05-26 15:23:10,536][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.18355438113212585, 'train/grad_norm_pre_clip': 0.5038585066795349, 'timing/data': 0.004201089031994343, 'timing/model': 0.6207068972289562, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.09}) +[2026-05-26 15:24:48,055][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.01726306788623333, 'train/grad_norm_pre_clip': 0.4294518232345581, 'timing/data': 0.004428376443684101, 'timing/model': 0.6365395998582244, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.09}) +[2026-05-26 15:26:43,872][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.049256548285484314, 'train/grad_norm_pre_clip': 0.7136054635047913, 'eval/loss': 0.13615599274635315, 'timing/data': 0.0048216646537184715, 'timing/model': 0.6400050530210137, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.1}) +[2026-05-26 15:28:23,304][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.04984563589096069, 'train/grad_norm_pre_clip': 1.05101478099823, 'timing/data': 0.005922706797719002, 'timing/model': 0.8763870680704713, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.1}) +[2026-05-26 15:30:01,969][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.18390323221683502, 'train/grad_norm_pre_clip': 1.1531198024749756, 'timing/data': 0.004212762229144573, 'timing/model': 0.6413184478878975, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.1}) +[2026-05-26 15:31:39,338][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.01886863075196743, 'train/grad_norm_pre_clip': 0.7458183765411377, 'timing/data': 0.004169662483036518, 'timing/model': 0.7202188912779093, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.1}) +[2026-05-26 15:33:15,675][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.037357158958911896, 'train/grad_norm_pre_clip': 0.45980948209762573, 'timing/data': 0.004145353101193905, 'timing/model': 0.5687801502645016, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.1}) +[2026-05-26 15:34:50,645][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.02788611128926277, 'train/grad_norm_pre_clip': 0.7803908586502075, 'timing/data': 0.005069439299404621, 'timing/model': 0.6212001014500856, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.1}) +[2026-05-26 15:36:27,357][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.13997328281402588, 'train/grad_norm_pre_clip': 3.684058666229248, 'timing/data': 0.004354167729616165, 'timing/model': 0.6218124441802502, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.1}) +[2026-05-26 15:38:05,934][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.5184587836265564, 'train/grad_norm_pre_clip': 3.0076513290405273, 'timing/data': 0.00692420918494463, 'timing/model': 0.6967783961445093, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.1}) +[2026-05-26 15:39:45,980][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.03517908975481987, 'train/grad_norm_pre_clip': 0.7241290807723999, 'timing/data': 0.004297110252082348, 'timing/model': 0.8279053559526801, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.1}) +[2026-05-26 15:41:24,870][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.10009603947401047, 'train/grad_norm_pre_clip': 0.7905939221382141, 'timing/data': 0.004258656874299049, 'timing/model': 0.583606488071382, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.1}) +[2026-05-26 15:43:15,641][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.06197447329759598, 'train/grad_norm_pre_clip': 0.5238891243934631, 'eval/loss': 0.10216648876667023, 'timing/data': 0.004161645658314228, 'timing/model': 0.6286760531365871, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.1}) +[2026-05-26 15:43:15,648][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1500_state +[2026-05-26 15:43:15,665][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-26 15:43:23,730][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1500_state/model.safetensors +[2026-05-26 15:43:42,417][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1500_state/optimizer.bin +[2026-05-26 15:43:42,418][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1500_state/sampler.bin +[2026-05-26 15:43:42,419][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1500_state/sampler_1.bin +[2026-05-26 15:43:42,440][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1500_state/random_states_0.pkl +[2026-05-26 15:43:44,448][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpn2qn1093/test.c -o /tmp/tmpn2qn1093/test.o +[2026-05-26 15:43:44,481][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpn2qn1093/test.o -laio -o /tmp/tmpn2qn1093/a.out +[2026-05-26 15:43:45,159][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmp9ct_7nww/test.c -o /tmp/tmp9ct_7nww/test.o +[2026-05-26 15:43:45,246][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmp9ct_7nww/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp9ct_7nww/a.out +[2026-05-26 15:43:45,292][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpi6cdhodn/test.c -o /tmp/tmpi6cdhodn/test.o +[2026-05-26 15:43:45,307][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpi6cdhodn/test.o -laio -o /tmp/tmpi6cdhodn/a.out +[2026-05-26 15:44:01,337][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-05-26 15:44:01,337][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-26 15:44:01,350][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-26 15:45:40,582][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.261031836271286, 'train/grad_norm_pre_clip': 1.2957954406738281, 'timing/data': 0.0038556065410375595, 'timing/model': 0.5733308391645551, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.1}) +[2026-05-26 15:47:20,985][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.23785364627838135, 'train/grad_norm_pre_clip': 1.5620298385620117, 'timing/data': 0.004901818931102753, 'timing/model': 0.640172690153122, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.1}) +[2026-05-26 15:49:00,874][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.006825172808021307, 'train/grad_norm_pre_clip': 0.08826154470443726, 'timing/data': 0.003857267089188099, 'timing/model': 0.5811706660315394, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.1}) +[2026-05-26 15:50:39,880][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.23214218020439148, 'train/grad_norm_pre_clip': 1.1015398502349854, 'timing/data': 0.0038517052307724953, 'timing/model': 0.6436199005693197, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.1}) +[2026-05-26 15:52:21,375][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.20494195818901062, 'train/grad_norm_pre_clip': 2.0807747840881348, 'timing/data': 0.004300447180867195, 'timing/model': 0.6533775953575969, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.11}) +[2026-05-26 15:54:00,284][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.10800499469041824, 'train/grad_norm_pre_clip': 2.742534637451172, 'timing/data': 0.004065528512001038, 'timing/model': 0.6387542514130473, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.11}) +[2026-05-26 15:55:40,675][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.09802430868148804, 'train/grad_norm_pre_clip': 0.4868515431880951, 'timing/data': 0.0039859432727098465, 'timing/model': 0.6520597347989678, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.11}) +[2026-05-26 15:57:23,649][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.07824725657701492, 'train/grad_norm_pre_clip': 0.7337707281112671, 'timing/data': 0.004255097359418869, 'timing/model': 0.629210302606225, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.11}) +[2026-05-26 15:59:01,264][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.5228182673454285, 'train/grad_norm_pre_clip': 3.3080813884735107, 'timing/data': 0.004068462178111076, 'timing/model': 0.639865661971271, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.11}) +[2026-05-26 16:00:50,156][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.05007568374276161, 'train/grad_norm_pre_clip': 0.7273057103157043, 'eval/loss': 0.10340123623609543, 'timing/data': 0.004035976715385914, 'timing/model': 0.6472966121509671, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.11}) +[2026-05-26 16:02:29,271][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.013038942590355873, 'train/grad_norm_pre_clip': 0.13296864926815033, 'timing/data': 0.004019814543426037, 'timing/model': 0.634771017357707, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.11}) +[2026-05-26 16:04:07,492][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.3020964562892914, 'train/grad_norm_pre_clip': 2.690249443054199, 'timing/data': 0.003811156377196312, 'timing/model': 0.6650188090279698, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.11}) +[2026-05-26 16:05:48,775][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.059890829026699066, 'train/grad_norm_pre_clip': 1.4525821208953857, 'timing/data': 0.003808114677667618, 'timing/model': 0.6507188798859715, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.11}) +[2026-05-26 16:07:26,673][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.10442827641963959, 'train/grad_norm_pre_clip': 0.9740545153617859, 'timing/data': 0.0039055850356817245, 'timing/model': 0.6530604977160692, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.11}) +[2026-05-26 16:09:04,764][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.08246660977602005, 'train/grad_norm_pre_clip': 1.4286682605743408, 'timing/data': 0.004455753602087498, 'timing/model': 0.6408962123095989, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.11}) +[2026-05-26 16:10:43,788][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.05061505734920502, 'train/grad_norm_pre_clip': 0.6869961023330688, 'timing/data': 0.004094683565199375, 'timing/model': 0.6780814742669463, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.11}) +[2026-05-26 16:12:21,376][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.021313117817044258, 'train/grad_norm_pre_clip': 0.24650484323501587, 'timing/data': 0.003905968740582466, 'timing/model': 0.6589277302846313, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.11}) +[2026-05-26 16:14:02,489][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.018241316080093384, 'train/grad_norm_pre_clip': 0.16571015119552612, 'timing/data': 0.004291216842830181, 'timing/model': 0.8629252761602402, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.11}) +[2026-05-26 16:15:42,878][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.006773719564080238, 'train/grad_norm_pre_clip': 0.09508297592401505, 'timing/data': 0.004271018318831921, 'timing/model': 0.6499167969450355, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.11}) +[2026-05-26 16:17:35,779][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.15015268325805664, 'train/grad_norm_pre_clip': 1.244412899017334, 'eval/loss': 0.10277599096298218, 'timing/data': 0.004025032743811607, 'timing/model': 0.643637583591044, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.12}) +[2026-05-26 16:19:18,479][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.35151731967926025, 'train/grad_norm_pre_clip': 2.4282338619232178, 'timing/data': 0.004074223339557648, 'timing/model': 0.6595357498154044, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.12}) +[2026-05-26 16:20:57,376][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.07968612015247345, 'train/grad_norm_pre_clip': 0.8674712777137756, 'timing/data': 0.003897717222571373, 'timing/model': 0.653424458578229, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.12}) +[2026-05-26 16:22:36,461][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.049539532512426376, 'train/grad_norm_pre_clip': 0.6885424256324768, 'timing/data': 0.004279515706002712, 'timing/model': 0.6350016184151173, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.12}) +[2026-05-26 16:24:17,169][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.3483504354953766, 'train/grad_norm_pre_clip': 2.107945680618286, 'timing/data': 0.004000137560069561, 'timing/model': 0.6426683804020286, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 0.12}) +[2026-05-26 16:25:56,388][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.05365179106593132, 'train/grad_norm_pre_clip': 0.5279467701911926, 'timing/data': 0.0039508892223238945, 'timing/model': 0.6642570169642568, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 0.12}) +[2026-05-26 16:25:56,392][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1750_state +[2026-05-26 16:25:56,398][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-26 16:26:05,076][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1750_state/model.safetensors +[2026-05-26 16:26:23,779][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1750_state/optimizer.bin +[2026-05-26 16:26:23,780][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1750_state/sampler.bin +[2026-05-26 16:26:23,781][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1750_state/sampler_1.bin +[2026-05-26 16:26:23,801][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_1750_state/random_states_0.pkl +[2026-05-26 16:26:33,199][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1750 +[2026-05-26 16:26:33,199][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-26 16:26:33,212][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-26 16:28:10,784][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.017820904031395912, 'train/grad_norm_pre_clip': 0.18253497779369354, 'timing/data': 0.003911636769771576, 'timing/model': 0.5826096413657069, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 0.12}) +[2026-05-26 16:29:49,360][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.07450418919324875, 'train/grad_norm_pre_clip': 2.6448121070861816, 'timing/data': 0.0036061154678463936, 'timing/model': 0.5695840306580067, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.12}) +[2026-05-26 16:31:27,379][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.0401708148419857, 'train/grad_norm_pre_clip': 1.2932041883468628, 'timing/data': 0.004956440068781376, 'timing/model': 0.6673391135409474, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 0.12}) +[2026-05-26 16:33:10,285][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.12713488936424255, 'train/grad_norm_pre_clip': 1.9695408344268799, 'timing/data': 0.004778004251420498, 'timing/model': 0.5813192361965775, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 0.12}) +[2026-05-26 16:35:05,566][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.05538091063499451, 'train/grad_norm_pre_clip': 0.5194436311721802, 'eval/loss': 0.11834628880023956, 'timing/data': 0.005292479880154133, 'timing/model': 1.2599345613270998, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 0.12}) +[2026-05-26 16:36:50,906][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.05917384847998619, 'train/grad_norm_pre_clip': 1.7094354629516602, 'timing/data': 0.004592963494360447, 'timing/model': 0.6675970926880836, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 0.12}) +[2026-05-26 16:38:31,528][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.020502863451838493, 'train/grad_norm_pre_clip': 0.30737245082855225, 'timing/data': 0.0037497282028198242, 'timing/model': 0.6151035288348794, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 0.12}) +[2026-05-26 16:40:09,885][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.042734235525131226, 'train/grad_norm_pre_clip': 2.8566696643829346, 'timing/data': 0.004105734638869762, 'timing/model': 0.641548004001379, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 0.12}) +[2026-05-26 16:41:51,075][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.3204497992992401, 'train/grad_norm_pre_clip': 1.571542501449585, 'timing/data': 0.004509645514190197, 'timing/model': 0.572526847012341, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.12}) +[2026-05-26 16:43:32,125][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.014006107114255428, 'train/grad_norm_pre_clip': 0.21716570854187012, 'timing/data': 0.005151255056262016, 'timing/model': 0.6950342515483499, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 0.13}) +[2026-05-26 16:45:12,999][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.012834707275032997, 'train/grad_norm_pre_clip': 0.2324918657541275, 'timing/data': 0.006623903289437294, 'timing/model': 0.9547801995649934, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 0.13}) +[2026-05-26 16:46:53,179][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.017325950786471367, 'train/grad_norm_pre_clip': 0.37332475185394287, 'timing/data': 0.003985075280070305, 'timing/model': 0.5736622903496027, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 0.13}) +[2026-05-26 16:48:31,457][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.10923193395137787, 'train/grad_norm_pre_clip': 1.2185604572296143, 'timing/data': 0.004131028428673744, 'timing/model': 0.6232564272359014, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 0.13}) +[2026-05-26 16:50:12,587][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.05491996929049492, 'train/grad_norm_pre_clip': 0.6322410702705383, 'timing/data': 0.005598309449851513, 'timing/model': 0.7712778029963374, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 0.13}) +[2026-05-26 16:52:02,794][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.0077070570550858974, 'train/grad_norm_pre_clip': 0.41437259316444397, 'eval/loss': 0.12689469754695892, 'timing/data': 0.0045263152569532394, 'timing/model': 0.6579194162040949, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 0.13}) +[2026-05-26 16:53:45,988][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.08176559209823608, 'train/grad_norm_pre_clip': 0.6894395351409912, 'timing/data': 0.004096812568604946, 'timing/model': 0.6646517915651202, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.13}) +[2026-05-26 16:55:27,692][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.00023858447093516588, 'train/grad_norm_pre_clip': 0.011756532825529575, 'timing/data': 0.005083288997411728, 'timing/model': 0.9487668024376035, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 0.13}) +[2026-05-26 16:57:08,572][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.039330050349235535, 'train/grad_norm_pre_clip': 0.5627490282058716, 'timing/data': 0.004967900924384594, 'timing/model': 0.6286651184782386, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 0.13}) +[2026-05-26 16:58:47,696][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.304236501455307, 'train/grad_norm_pre_clip': 1.2755120992660522, 'timing/data': 0.004979286342859268, 'timing/model': 0.6614993317052722, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 0.13}) +[2026-05-26 17:00:28,365][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.034251824021339417, 'train/grad_norm_pre_clip': 0.5448734164237976, 'timing/data': 0.005022134631872177, 'timing/model': 0.6262497557327151, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 0.13}) +[2026-05-26 17:02:07,178][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.14814570546150208, 'train/grad_norm_pre_clip': 1.6359927654266357, 'timing/data': 0.004166785627603531, 'timing/model': 0.6534763434901834, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 0.13}) +[2026-05-26 17:03:49,934][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.2426452487707138, 'train/grad_norm_pre_clip': 1.2450487613677979, 'timing/data': 0.0045817214995622635, 'timing/model': 0.9077239213511348, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 0.13}) +[2026-05-26 17:05:27,983][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.037437405437231064, 'train/grad_norm_pre_clip': 0.5306331515312195, 'timing/data': 0.004013326019048691, 'timing/model': 0.5842517567798495, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.13}) +[2026-05-26 17:07:05,522][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.3023490607738495, 'train/grad_norm_pre_clip': 2.415933132171631, 'timing/data': 0.00493655726313591, 'timing/model': 0.6122747352346778, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 0.14}) +[2026-05-26 17:13:07,605][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2000 +[2026-05-26 17:13:07,607][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.022541707381606102, 'train/grad_norm_pre_clip': 1.1631618738174438, 'eval/loss': 0.12195669859647751, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 1.0, 'timing/data': 0.0044759828597307205, 'timing/model': 0.7713023973628879, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 0.14}) +[2026-05-26 17:13:07,609][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2000_state +[2026-05-26 17:13:07,615][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-26 17:13:15,593][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2000_state/model.safetensors +[2026-05-26 17:13:34,263][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2000_state/optimizer.bin +[2026-05-26 17:13:34,264][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2000_state/sampler.bin +[2026-05-26 17:13:34,264][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2000_state/sampler_1.bin +[2026-05-26 17:13:34,267][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2000_state/random_states_0.pkl +[2026-05-26 17:13:43,740][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-05-26 17:13:43,741][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-26 17:13:43,757][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-26 17:15:23,529][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.17807112634181976, 'train/grad_norm_pre_clip': 1.4979009628295898, 'timing/data': 0.004489556886255741, 'timing/model': 0.68895598128438, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 0.14}) +[2026-05-26 17:17:02,076][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.004879769869148731, 'train/grad_norm_pre_clip': 0.1320929080247879, 'timing/data': 0.003961464390158653, 'timing/model': 0.6072602700442076, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 0.14}) +[2026-05-26 17:18:44,688][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.07036595046520233, 'train/grad_norm_pre_clip': 0.45888590812683105, 'timing/data': 0.004350233823060989, 'timing/model': 0.6009532446041703, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 0.14}) +[2026-05-26 17:20:25,355][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.0034203571267426014, 'train/grad_norm_pre_clip': 0.027200642973184586, 'timing/data': 0.004081333056092262, 'timing/model': 0.6366065060719848, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 0.14}) +[2026-05-26 17:22:06,262][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.2529459297657013, 'train/grad_norm_pre_clip': 1.368614673614502, 'timing/data': 0.003928047604858875, 'timing/model': 0.6285981237888336, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.14}) +[2026-05-26 17:23:45,462][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.043777868151664734, 'train/grad_norm_pre_clip': 0.6921899318695068, 'timing/data': 0.003934428095817566, 'timing/model': 0.5837667202576995, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 0.14}) +[2026-05-26 17:25:23,873][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.13743650913238525, 'train/grad_norm_pre_clip': 1.4425456523895264, 'timing/data': 0.0038480963557958603, 'timing/model': 0.6361316526308656, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 0.14}) +[2026-05-26 17:27:01,972][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.029676830396056175, 'train/grad_norm_pre_clip': 0.5053955316543579, 'timing/data': 0.004351519979536533, 'timing/model': 0.6354976091533899, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 0.14}) +[2026-05-26 17:28:43,080][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.1318536400794983, 'train/grad_norm_pre_clip': 0.4634557068347931, 'timing/data': 0.003953021951019764, 'timing/model': 0.6158084906637669, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.14}) +[2026-05-26 17:30:34,951][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.02659866400063038, 'train/grad_norm_pre_clip': 0.19962434470653534, 'eval/loss': 0.11516797542572021, 'timing/data': 0.005707351490855217, 'timing/model': 0.6640484323725104, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.14}) +[2026-05-26 17:32:15,480][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.037941526621580124, 'train/grad_norm_pre_clip': 0.2993605434894562, 'timing/data': 0.023835926316678524, 'timing/model': 0.9496345398947597, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.14}) +[2026-05-26 17:33:54,174][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.000418584473663941, 'train/grad_norm_pre_clip': 0.008525979705154896, 'timing/data': 0.0071616750210523605, 'timing/model': 0.6549802394583821, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.14}) +[2026-05-26 17:35:35,451][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.09075751900672913, 'train/grad_norm_pre_clip': 2.941943407058716, 'timing/data': 0.003868299536406994, 'timing/model': 0.9208096396178007, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 0.14}) +[2026-05-26 17:37:13,431][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.10396787524223328, 'train/grad_norm_pre_clip': 0.9544707536697388, 'timing/data': 0.0043715015053749084, 'timing/model': 0.6148559134453535, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 0.15}) +[2026-05-26 17:38:53,083][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.04659289866685867, 'train/grad_norm_pre_clip': 0.5834264159202576, 'timing/data': 0.003864070400595665, 'timing/model': 0.6589536499232054, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 0.15}) +[2026-05-26 17:40:32,266][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.02203366532921791, 'train/grad_norm_pre_clip': 0.20153918862342834, 'timing/data': 0.004098627716302872, 'timing/model': 0.6524856621399522, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 0.15}) +[2026-05-26 17:42:11,599][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.571176290512085, 'train/grad_norm_pre_clip': 2.1291370391845703, 'timing/data': 0.0064438506960868835, 'timing/model': 0.726272638887167, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 0.15}) +[2026-05-26 17:43:56,178][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.03662429377436638, 'train/grad_norm_pre_clip': 0.4019230306148529, 'timing/data': 0.004171642474830151, 'timing/model': 0.6316591203212738, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 0.15}) +[2026-05-26 17:45:36,378][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.36499443650245667, 'train/grad_norm_pre_clip': 1.8286235332489014, 'timing/data': 0.00390999298542738, 'timing/model': 0.5733902286738157, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.15}) +[2026-05-26 17:47:27,526][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.15278753638267517, 'train/grad_norm_pre_clip': 2.4569735527038574, 'eval/loss': 0.11252959072589874, 'timing/data': 0.0038856202736496925, 'timing/model': 0.6397506734356284, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 0.15}) +[2026-05-26 17:49:07,479][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.03588693216443062, 'train/grad_norm_pre_clip': 0.356344997882843, 'timing/data': 0.003920465707778931, 'timing/model': 0.6489411983639002, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 0.15}) +[2026-05-26 17:50:49,682][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.040677741169929504, 'train/grad_norm_pre_clip': 0.41441136598587036, 'timing/data': 0.004203170537948608, 'timing/model': 0.6589668560773134, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 0.15}) +[2026-05-26 17:52:28,774][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.009139188565313816, 'train/grad_norm_pre_clip': 0.18734563887119293, 'timing/data': 0.003877961076796055, 'timing/model': 0.5678046951070428, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 0.15}) +[2026-05-26 17:54:08,876][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.3997558355331421, 'train/grad_norm_pre_clip': 1.7625583410263062, 'timing/data': 0.004068901762366295, 'timing/model': 0.6435475675389171, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 0.15}) +[2026-05-26 17:55:49,628][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.08655901253223419, 'train/grad_norm_pre_clip': 0.49618470668792725, 'timing/data': 0.003968622535467148, 'timing/model': 0.8889205930754542, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 0.15}) +[2026-05-26 17:55:49,629][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2250_state +[2026-05-26 17:55:49,635][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-26 17:55:57,782][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2250_state/model.safetensors +[2026-05-26 17:56:16,206][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2250_state/optimizer.bin +[2026-05-26 17:56:16,208][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2250_state/sampler.bin +[2026-05-26 17:56:16,208][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2250_state/sampler_1.bin +[2026-05-26 17:56:16,210][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2250_state/random_states_0.pkl +[2026-05-26 17:56:25,615][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2250 +[2026-05-26 17:56:25,616][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-26 17:56:25,633][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-26 17:58:04,708][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.1528882384300232, 'train/grad_norm_pre_clip': 2.5827205181121826, 'timing/data': 0.0035861236974596977, 'timing/model': 0.5922616915777326, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.15}) +[2026-05-26 17:59:46,277][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.053717054426670074, 'train/grad_norm_pre_clip': 0.40417754650115967, 'timing/data': 0.004149497486650944, 'timing/model': 0.6587324161082506, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 0.15}) +[2026-05-26 18:01:28,679][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.012454593554139137, 'train/grad_norm_pre_clip': 0.14355772733688354, 'timing/data': 0.003878527320921421, 'timing/model': 0.5978055261075497, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 0.15}) +[2026-05-26 18:03:10,137][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.3656768500804901, 'train/grad_norm_pre_clip': 2.285072088241577, 'timing/data': 0.006097412668168545, 'timing/model': 0.7035045493394136, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 0.16}) +[2026-05-26 18:05:01,079][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.3285082280635834, 'train/grad_norm_pre_clip': 2.643691062927246, 'eval/loss': 0.1246851459145546, 'timing/data': 0.004265874624252319, 'timing/model': 0.9254555320367217, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 0.16}) +[2026-05-26 18:06:42,687][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.2663513720035553, 'train/grad_norm_pre_clip': 1.5967038869857788, 'timing/data': 0.003961146809160709, 'timing/model': 0.576530079357326, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 0.16}) +[2026-05-26 18:08:21,790][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.003040802665054798, 'train/grad_norm_pre_clip': 0.065632164478302, 'timing/data': 0.004152785055339336, 'timing/model': 0.5928137144073844, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 0.16}) +[2026-05-26 18:10:03,381][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.11755387485027313, 'train/grad_norm_pre_clip': 1.6921545267105103, 'timing/data': 0.0048605334013700485, 'timing/model': 0.6521618654951453, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.16}) +[2026-05-26 18:11:43,976][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.011355418711900711, 'train/grad_norm_pre_clip': 0.15677835047245026, 'timing/data': 0.004126918502151966, 'timing/model': 0.6065099090337753, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 0.16}) +[2026-05-26 18:13:24,424][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.0064177983440458775, 'train/grad_norm_pre_clip': 0.13065651059150696, 'timing/data': 0.004823192022740841, 'timing/model': 0.696695739403367, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 0.16}) +[2026-05-26 18:15:04,825][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.06409233808517456, 'train/grad_norm_pre_clip': 1.3614141941070557, 'timing/data': 0.0039701396599411964, 'timing/model': 0.9053280856460333, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 0.16}) +[2026-05-26 18:16:46,753][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.18240590393543243, 'train/grad_norm_pre_clip': 1.0749711990356445, 'timing/data': 0.003925998695194721, 'timing/model': 0.9254460036754608, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 0.16}) +[2026-05-26 18:18:26,518][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.40114882588386536, 'train/grad_norm_pre_clip': 2.2375736236572266, 'timing/data': 0.003861119970679283, 'timing/model': 0.6063689179718494, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 0.16}) +[2026-05-26 18:20:06,487][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.03735387325286865, 'train/grad_norm_pre_clip': 0.2152511328458786, 'timing/data': 0.0036832522600889206, 'timing/model': 0.5769131761044264, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 0.16}) +[2026-05-26 18:21:56,887][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.0018524902407079935, 'train/grad_norm_pre_clip': 0.02407042123377323, 'eval/loss': 0.10323251783847809, 'timing/data': 0.004377944394946098, 'timing/model': 0.6337966695427895, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.16}) +[2026-05-26 18:23:38,865][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.030650783330202103, 'train/grad_norm_pre_clip': 0.485644668340683, 'timing/data': 0.003939999267458916, 'timing/model': 0.6358618028461933, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 0.16}) +[2026-05-26 18:25:18,052][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.1374833732843399, 'train/grad_norm_pre_clip': 0.3256736695766449, 'timing/data': 0.004091372713446617, 'timing/model': 0.636029857210815, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 0.16}) +[2026-05-26 18:26:58,877][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.03625300154089928, 'train/grad_norm_pre_clip': 0.16103273630142212, 'timing/data': 0.0039632441475987434, 'timing/model': 0.663723967038095, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 0.17}) +[2026-05-26 18:28:40,285][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.019328732043504715, 'train/grad_norm_pre_clip': 0.3135797381401062, 'timing/data': 0.004161887802183628, 'timing/model': 0.666460681706667, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 0.17}) +[2026-05-26 18:30:19,884][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.15927013754844666, 'train/grad_norm_pre_clip': 1.0667691230773926, 'timing/data': 0.004681188613176346, 'timing/model': 0.6253757979720831, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 0.17}) +[2026-05-26 18:32:01,948][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.026497580111026764, 'train/grad_norm_pre_clip': 0.2053023725748062, 'timing/data': 0.00408519059419632, 'timing/model': 0.6298027485609055, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 0.17}) +[2026-05-26 18:33:43,394][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.10545793175697327, 'train/grad_norm_pre_clip': 0.6906834244728088, 'timing/data': 0.003943152725696564, 'timing/model': 0.570901483297348, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.17}) +[2026-05-26 18:35:24,385][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.007337694987654686, 'train/grad_norm_pre_clip': 0.14677247405052185, 'timing/data': 0.003985844552516937, 'timing/model': 0.6036855168640614, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 0.17}) +[2026-05-26 18:37:07,489][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.3203277885913849, 'train/grad_norm_pre_clip': 1.5712990760803223, 'timing/data': 0.00432036817073822, 'timing/model': 0.6159382704645395, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 0.17}) +[2026-05-26 18:39:01,554][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.13001564145088196, 'train/grad_norm_pre_clip': 1.529403567314148, 'eval/loss': 0.09425617754459381, 'timing/data': 0.005551716312766075, 'timing/model': 0.9970214050263166, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 0.17}) +[2026-05-26 18:39:01,555][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2500_state +[2026-05-26 18:39:01,562][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-26 18:39:09,852][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2500_state/model.safetensors +[2026-05-26 18:39:29,149][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2500_state/optimizer.bin +[2026-05-26 18:39:29,151][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2500_state/sampler.bin +[2026-05-26 18:39:29,152][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2500_state/sampler_1.bin +[2026-05-26 18:39:29,154][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2500_state/random_states_0.pkl +[2026-05-26 18:39:38,366][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-05-26 18:39:38,366][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-26 18:39:38,386][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-26 18:41:15,805][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.06660415977239609, 'train/grad_norm_pre_clip': 0.5228776931762695, 'timing/data': 0.0042734285816550255, 'timing/model': 0.5957012735307217, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 0.17}) +[2026-05-26 18:42:55,355][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.17242470383644104, 'train/grad_norm_pre_clip': 1.3281710147857666, 'timing/data': 0.004067068919539452, 'timing/model': 0.9242068873718381, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 0.17}) +[2026-05-26 18:44:34,394][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.10059931874275208, 'train/grad_norm_pre_clip': 1.2703276872634888, 'timing/data': 0.0061498722061514854, 'timing/model': 0.6092255562543869, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 0.17}) +[2026-05-26 18:46:15,213][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.02020447514951229, 'train/grad_norm_pre_clip': 0.20273151993751526, 'timing/data': 0.004274275153875351, 'timing/model': 0.895328052341938, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.17}) +[2026-05-26 18:47:57,286][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.09014958888292313, 'train/grad_norm_pre_clip': 0.7291558980941772, 'timing/data': 0.003839666023850441, 'timing/model': 0.6653409721329808, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 0.17}) +[2026-05-26 18:49:37,267][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.08198514580726624, 'train/grad_norm_pre_clip': 0.5582004189491272, 'timing/data': 0.0044150808826088905, 'timing/model': 0.6399021744728088, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 0.17}) +[2026-05-26 18:51:17,968][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.003680114634335041, 'train/grad_norm_pre_clip': 0.029856860637664795, 'timing/data': 0.005445434711873531, 'timing/model': 0.6385993072763085, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 0.17}) +[2026-05-26 18:52:58,695][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.003172805532813072, 'train/grad_norm_pre_clip': 0.044251538813114166, 'timing/data': 0.0038152579218149185, 'timing/model': 0.7711893264204264, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 0.18}) +[2026-05-26 18:54:39,784][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.1489790976047516, 'train/grad_norm_pre_clip': 1.4294075965881348, 'timing/data': 0.004653941839933395, 'timing/model': 0.7644947841763496, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 0.18}) +[2026-05-26 18:56:30,764][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.08816589415073395, 'train/grad_norm_pre_clip': 1.2401529550552368, 'eval/loss': 0.10275453329086304, 'timing/data': 0.003938636742532253, 'timing/model': 0.8287697248160839, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 0.18}) +[2026-05-26 18:58:11,886][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.05933240056037903, 'train/grad_norm_pre_clip': 1.437965989112854, 'timing/data': 0.004070633091032505, 'timing/model': 0.6052650306373835, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.18}) +[2026-05-26 18:59:52,404][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.018331969156861305, 'train/grad_norm_pre_clip': 0.4148160517215729, 'timing/data': 0.004172194749116898, 'timing/model': 0.6430220594629645, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 0.18}) +[2026-05-26 19:01:30,580][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.002018739702180028, 'train/grad_norm_pre_clip': 0.01666589267551899, 'timing/data': 0.004238543100655079, 'timing/model': 0.626416526734829, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 0.18}) +[2026-05-26 19:03:07,661][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.03514387086033821, 'train/grad_norm_pre_clip': 1.7132964134216309, 'timing/data': 0.004123563878238201, 'timing/model': 0.5898478329181671, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 0.18}) +[2026-05-26 19:04:46,029][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.17774270474910736, 'train/grad_norm_pre_clip': 2.4524121284484863, 'timing/data': 0.0036755353212356567, 'timing/model': 0.9026701515540481, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 0.18}) +[2026-05-26 19:06:24,784][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.062142036855220795, 'train/grad_norm_pre_clip': 0.37170901894569397, 'timing/data': 0.003942474722862244, 'timing/model': 0.6659191781654954, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 0.18}) +[2026-05-26 19:08:03,845][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.03179458528757095, 'train/grad_norm_pre_clip': 0.4909714162349701, 'timing/data': 0.004300710745155811, 'timing/model': 0.8221684563905001, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 0.18}) +[2026-05-26 19:09:43,520][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.14964719116687775, 'train/grad_norm_pre_clip': 6.849612236022949, 'timing/data': 0.003917310386896133, 'timing/model': 0.679066451266408, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.18}) +[2026-05-26 19:11:23,387][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.3315371572971344, 'train/grad_norm_pre_clip': 2.78855037689209, 'timing/data': 0.0045797014608979225, 'timing/model': 0.6427432717755437, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 0.18}) +[2026-05-26 19:13:13,005][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.12553159892559052, 'train/grad_norm_pre_clip': 1.370625615119934, 'eval/loss': 0.10036996752023697, 'timing/data': 0.004001830704510212, 'timing/model': 0.624431868083775, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 0.18}) +[2026-05-26 19:14:51,976][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.044710736721754074, 'train/grad_norm_pre_clip': 0.4944573938846588, 'timing/data': 0.0045372918248176575, 'timing/model': 0.5800267150625587, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 0.18}) +[2026-05-26 19:16:31,181][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.13870970904827118, 'train/grad_norm_pre_clip': 0.8487300276756287, 'timing/data': 0.004442640580236912, 'timing/model': 0.5736716305837035, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 0.18}) +[2026-05-26 19:18:12,076][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.20311449468135834, 'train/grad_norm_pre_clip': 0.9290187358856201, 'timing/data': 0.0038468576967716217, 'timing/model': 0.659912078641355, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 0.19}) +[2026-05-26 19:19:54,246][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.02527736686170101, 'train/grad_norm_pre_clip': 0.6058459877967834, 'timing/data': 0.004054033197462559, 'timing/model': 0.9346946300938725, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 0.19}) +[2026-05-26 19:21:35,978][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.13841086626052856, 'train/grad_norm_pre_clip': 2.4687814712524414, 'timing/data': 0.00391826406121254, 'timing/model': 0.6452951673418283, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.19}) +[2026-05-26 19:21:35,979][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2750_state +[2026-05-26 19:21:35,986][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-26 19:21:44,733][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2750_state/model.safetensors +[2026-05-26 19:22:03,253][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2750_state/optimizer.bin +[2026-05-26 19:22:03,256][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2750_state/sampler.bin +[2026-05-26 19:22:03,256][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2750_state/sampler_1.bin +[2026-05-26 19:22:03,259][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_2750_state/random_states_0.pkl +[2026-05-26 19:22:12,833][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2750 +[2026-05-26 19:22:12,833][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-26 19:22:12,851][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-26 19:23:52,823][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.1122286468744278, 'train/grad_norm_pre_clip': 2.3277008533477783, 'timing/data': 0.004002864472568035, 'timing/model': 0.8973798388615251, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 0.19}) +[2026-05-26 19:25:30,681][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.07035885006189346, 'train/grad_norm_pre_clip': 1.300471305847168, 'timing/data': 0.004374390468001366, 'timing/model': 0.9570017987862229, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 0.19}) +[2026-05-26 19:27:10,371][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.0017108883475884795, 'train/grad_norm_pre_clip': 0.04049825668334961, 'timing/data': 0.0037953313440084457, 'timing/model': 0.5918075675144792, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 0.19}) +[2026-05-26 19:28:51,881][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.006889394484460354, 'train/grad_norm_pre_clip': 0.21250444650650024, 'timing/data': 0.003904331475496292, 'timing/model': 0.5747619066387415, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 0.19}) +[2026-05-26 19:30:43,230][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.025427382439374924, 'train/grad_norm_pre_clip': 0.4003733992576599, 'eval/loss': 0.11000440269708633, 'timing/data': 0.0038447603583335876, 'timing/model': 0.6392127992585301, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 0.19}) +[2026-05-26 19:32:21,789][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.623746395111084, 'train/grad_norm_pre_clip': 2.9868719577789307, 'timing/data': 0.00445933360606432, 'timing/model': 0.672532182186842, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 0.19}) +[2026-05-26 19:34:00,764][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.0063149938359856606, 'train/grad_norm_pre_clip': 0.06319425255060196, 'timing/data': 0.0037349676713347435, 'timing/model': 0.6434959080070257, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.19}) +[2026-05-26 19:35:42,370][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.2104109525680542, 'train/grad_norm_pre_clip': 2.472756862640381, 'timing/data': 0.004359515383839607, 'timing/model': 0.8865020209923387, 'learning_rate/qwen_vl_interface': 8.017284756424381e-06, 'learning_rate/action_model': 8.017284756424381e-05, 'epoch': 0.19}) +[2026-05-26 19:37:22,557][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.048303864896297455, 'train/grad_norm_pre_clip': 0.6892205476760864, 'timing/data': 0.004846676252782345, 'timing/model': 0.631242479197681, 'learning_rate/qwen_vl_interface': 7.992475812221894e-06, 'learning_rate/action_model': 7.992475812221894e-05, 'epoch': 0.19}) +[2026-05-26 19:39:00,770][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.090266652405262, 'train/grad_norm_pre_clip': 0.6800826191902161, 'timing/data': 0.00406256876885891, 'timing/model': 0.5966657437384129, 'learning_rate/qwen_vl_interface': 7.967554135580307e-06, 'learning_rate/action_model': 7.967554135580307e-05, 'epoch': 0.19}) +[2026-05-26 19:40:36,465][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.03799282759428024, 'train/grad_norm_pre_clip': 0.45926806330680847, 'timing/data': 0.003713502548635006, 'timing/model': 0.6336274445056915, 'learning_rate/qwen_vl_interface': 7.942520750932138e-06, 'learning_rate/action_model': 7.942520750932138e-05, 'epoch': 0.19}) +[2026-05-26 19:42:11,874][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.009182490408420563, 'train/grad_norm_pre_clip': 0.16426151990890503, 'timing/data': 0.0037070875987410545, 'timing/model': 0.5691449400037527, 'learning_rate/qwen_vl_interface': 7.91737668730178e-06, 'learning_rate/action_model': 7.91737668730178e-05, 'epoch': 0.19}) +[2026-05-26 19:43:47,279][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.13247451186180115, 'train/grad_norm_pre_clip': 1.7447950839996338, 'timing/data': 0.0037526795640587807, 'timing/model': 0.5760822519659996, 'learning_rate/qwen_vl_interface': 7.89212297826321e-06, 'learning_rate/action_model': 7.892122978263209e-05, 'epoch': 0.2}) +[2026-05-26 19:45:22,475][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.09994395822286606, 'train/grad_norm_pre_clip': 1.2824770212173462, 'timing/data': 0.0044151414185762405, 'timing/model': 0.8643580311909318, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.2}) +[2026-05-26 19:47:08,177][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.332663357257843, 'train/grad_norm_pre_clip': 2.883605718612671, 'eval/loss': 0.09636002033948898, 'timing/data': 0.003764629364013672, 'timing/model': 0.5818328838795424, 'learning_rate/qwen_vl_interface': 7.841290780750107e-06, 'learning_rate/action_model': 7.841290780750107e-05, 'epoch': 0.2}) +[2026-05-26 19:48:45,774][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.13489656150341034, 'train/grad_norm_pre_clip': 1.1109412908554077, 'timing/data': 0.0039358362555503845, 'timing/model': 0.6130685927346349, 'learning_rate/qwen_vl_interface': 7.815714381788123e-06, 'learning_rate/action_model': 7.815714381788122e-05, 'epoch': 0.2}) +[2026-05-26 19:50:21,247][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.04868953302502632, 'train/grad_norm_pre_clip': 0.569536566734314, 'timing/data': 0.003765988163650036, 'timing/model': 0.5673936549574137, 'learning_rate/qwen_vl_interface': 7.790032516357116e-06, 'learning_rate/action_model': 7.790032516357117e-05, 'epoch': 0.2}) +[2026-05-26 19:51:57,078][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.0013612923212349415, 'train/grad_norm_pre_clip': 0.024404682219028473, 'timing/data': 0.004101842641830444, 'timing/model': 0.572724093683064, 'learning_rate/qwen_vl_interface': 7.764246240137991e-06, 'learning_rate/action_model': 7.764246240137991e-05, 'epoch': 0.2}) +[2026-05-26 19:53:38,638][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.12525032460689545, 'train/grad_norm_pre_clip': 2.944915771484375, 'timing/data': 0.003950431011617184, 'timing/model': 0.7090820297598839, 'learning_rate/qwen_vl_interface': 7.738356613103562e-06, 'learning_rate/action_model': 7.73835661310356e-05, 'epoch': 0.2}) +[2026-05-26 19:55:14,486][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.0748513862490654, 'train/grad_norm_pre_clip': 1.5172604322433472, 'timing/data': 0.003709034062922001, 'timing/model': 0.5776178846135736, 'learning_rate/qwen_vl_interface': 7.712364699474996e-06, 'learning_rate/action_model': 7.712364699474996e-05, 'epoch': 0.2}) +[2026-05-26 19:56:53,182][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.12587954103946686, 'train/grad_norm_pre_clip': 0.8028568029403687, 'timing/data': 0.003762100823223591, 'timing/model': 0.65756681188941, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.2}) +[2026-05-26 19:58:28,866][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.026196099817752838, 'train/grad_norm_pre_clip': 0.4197884202003479, 'timing/data': 0.0038216225802898407, 'timing/model': 0.6346336556598544, 'learning_rate/qwen_vl_interface': 7.66007829029919e-06, 'learning_rate/action_model': 7.660078290299191e-05, 'epoch': 0.2}) +[2026-05-26 20:00:09,372][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.053664278239011765, 'train/grad_norm_pre_clip': 0.7031293511390686, 'timing/data': 0.003718193620443344, 'timing/model': 0.6596214883029461, 'learning_rate/qwen_vl_interface': 7.633785944041444e-06, 'learning_rate/action_model': 7.633785944041444e-05, 'epoch': 0.2}) +[2026-05-26 20:01:46,332][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.08498132228851318, 'train/grad_norm_pre_clip': 1.2978088855743408, 'timing/data': 0.005286284722387791, 'timing/model': 0.8245853530243039, 'learning_rate/qwen_vl_interface': 7.607395609680189e-06, 'learning_rate/action_model': 7.60739560968019e-05, 'epoch': 0.2}) +[2026-05-26 20:07:36,269][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3000 +[2026-05-26 20:07:36,270][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.004033916164189577, 'train/grad_norm_pre_clip': 0.25697407126426697, 'eval/loss': 0.21066303551197052, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 1.0, 'timing/data': 0.003802771680057049, 'timing/model': 0.930081338621676, 'learning_rate/qwen_vl_interface': 7.580908372018706e-06, 'learning_rate/action_model': 7.580908372018705e-05, 'epoch': 0.2}) +[2026-05-26 20:07:36,271][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3000_state +[2026-05-26 20:07:36,278][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-26 20:07:44,328][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3000_state/model.safetensors +[2026-05-26 20:08:02,439][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3000_state/optimizer.bin +[2026-05-26 20:08:02,457][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3000_state/sampler.bin +[2026-05-26 20:08:02,458][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3000_state/sampler_1.bin +[2026-05-26 20:08:02,461][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3000_state/random_states_0.pkl +[2026-05-26 20:08:11,263][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-05-26 20:08:11,264][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-26 20:08:11,279][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-26 20:09:46,369][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.0005999785498715937, 'train/grad_norm_pre_clip': 0.02810375951230526, 'timing/data': 0.003776836208999157, 'timing/model': 0.6397547274827957, 'learning_rate/qwen_vl_interface': 7.554325319843593e-06, 'learning_rate/action_model': 7.554325319843592e-05, 'epoch': 0.2}) +[2026-05-26 20:11:22,371][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.02534952014684677, 'train/grad_norm_pre_clip': 0.2915697395801544, 'timing/data': 0.004164941608905792, 'timing/model': 0.5797380832955241, 'learning_rate/qwen_vl_interface': 7.527647545880007e-06, 'learning_rate/action_model': 7.527647545880007e-05, 'epoch': 0.21}) +[2026-05-26 20:12:58,776][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.16198204457759857, 'train/grad_norm_pre_clip': 2.2773571014404297, 'timing/data': 0.0038983244448900223, 'timing/model': 0.6508467365056276, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.21}) +[2026-05-26 20:14:35,876][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.1140407919883728, 'train/grad_norm_pre_clip': 1.0844197273254395, 'timing/data': 0.003741256892681122, 'timing/model': 0.6233037998899817, 'learning_rate/qwen_vl_interface': 7.474012222911169e-06, 'learning_rate/action_model': 7.474012222911169e-05, 'epoch': 0.21}) +[2026-05-26 20:16:11,237][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.09050734341144562, 'train/grad_norm_pre_clip': 0.9064006209373474, 'timing/data': 0.0037764571607112885, 'timing/model': 0.8139900676906109, 'learning_rate/qwen_vl_interface': 7.447056878643969e-06, 'learning_rate/action_model': 7.447056878643968e-05, 'epoch': 0.21}) +[2026-05-26 20:17:50,006][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.05160556733608246, 'train/grad_norm_pre_clip': 0.401366263628006, 'timing/data': 0.0038509201258420944, 'timing/model': 0.9118970511481166, 'learning_rate/qwen_vl_interface': 7.420011221973771e-06, 'learning_rate/action_model': 7.420011221973771e-05, 'epoch': 0.21}) +[2026-05-26 20:19:28,379][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.010977253317832947, 'train/grad_norm_pre_clip': 0.1597476452589035, 'timing/data': 0.0037522176280617714, 'timing/model': 0.5930878855288029, 'learning_rate/qwen_vl_interface': 7.39287636464159e-06, 'learning_rate/action_model': 7.39287636464159e-05, 'epoch': 0.21}) +[2026-05-26 20:21:03,315][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.09359131753444672, 'train/grad_norm_pre_clip': 1.3898123502731323, 'timing/data': 0.003978237509727478, 'timing/model': 0.6005758931860328, 'learning_rate/qwen_vl_interface': 7.365653422055131e-06, 'learning_rate/action_model': 7.365653422055131e-05, 'epoch': 0.21}) +[2026-05-26 20:22:41,585][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.11387080699205399, 'train/grad_norm_pre_clip': 1.8688578605651855, 'timing/data': 0.004889289848506451, 'timing/model': 0.6500763855874538, 'learning_rate/qwen_vl_interface': 7.338343513242941e-06, 'learning_rate/action_model': 7.338343513242941e-05, 'epoch': 0.21}) +[2026-05-26 20:24:35,333][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.11144626885652542, 'train/grad_norm_pre_clip': 1.6801036596298218, 'eval/loss': 0.09883593022823334, 'timing/data': 0.000532194972038269, 'timing/model': 0.6816512160003185, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.21}) +[2026-05-26 20:26:11,956][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.23947322368621826, 'train/grad_norm_pre_clip': 3.852997303009033, 'timing/data': 0.0006024567410349846, 'timing/model': 0.5912992358207703, 'learning_rate/qwen_vl_interface': 7.283467290883598e-06, 'learning_rate/action_model': 7.283467290883598e-05, 'epoch': 0.21}) +[2026-05-26 20:27:54,072][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.0011281135957688093, 'train/grad_norm_pre_clip': 0.04403522238135338, 'timing/data': 0.0006600189954042435, 'timing/model': 0.6455281870439649, 'learning_rate/qwen_vl_interface': 7.2559032330830105e-06, 'learning_rate/action_model': 7.25590323308301e-05, 'epoch': 0.21}) +[2026-05-26 20:29:35,282][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.04263802990317345, 'train/grad_norm_pre_clip': 0.6896493434906006, 'timing/data': 0.0006394460797309875, 'timing/model': 0.6534184254705906, 'learning_rate/qwen_vl_interface': 7.228256720457097e-06, 'learning_rate/action_model': 7.228256720457096e-05, 'epoch': 0.21}) +[2026-05-26 20:31:15,479][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.14525197446346283, 'train/grad_norm_pre_clip': 1.8023074865341187, 'timing/data': 0.0007015019655227661, 'timing/model': 0.6436338117346168, 'learning_rate/qwen_vl_interface': 7.200528889445707e-06, 'learning_rate/action_model': 7.200528889445706e-05, 'epoch': 0.21}) +[2026-05-26 20:32:54,180][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.5062217116355896, 'train/grad_norm_pre_clip': 3.988764524459839, 'timing/data': 0.0005647595971822739, 'timing/model': 0.6573893632739782, 'learning_rate/qwen_vl_interface': 7.1727208798313704e-06, 'learning_rate/action_model': 7.172720879831371e-05, 'epoch': 0.21}) +[2026-05-26 20:34:32,688][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.17038783431053162, 'train/grad_norm_pre_clip': 0.9467165470123291, 'timing/data': 0.0006448477506637573, 'timing/model': 0.5994849717244506, 'learning_rate/qwen_vl_interface': 7.14483383469245e-06, 'learning_rate/action_model': 7.144833834692449e-05, 'epoch': 0.21}) +[2026-05-26 20:36:11,075][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.043706249445676804, 'train/grad_norm_pre_clip': 0.4664760231971741, 'timing/data': 0.0006265407428145409, 'timing/model': 0.6547307064756751, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.22}) +[2026-05-26 20:37:48,520][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.034752268344163895, 'train/grad_norm_pre_clip': 0.5325695872306824, 'timing/data': 0.0006218571215867996, 'timing/model': 0.8550035115331411, 'learning_rate/qwen_vl_interface': 7.0888272263513595e-06, 'learning_rate/action_model': 7.088827226351359e-05, 'epoch': 0.22}) +[2026-05-26 20:39:29,193][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.02317318506538868, 'train/grad_norm_pre_clip': 0.333871454000473, 'timing/data': 0.0006316285580396652, 'timing/model': 0.7054617078974843, 'learning_rate/qwen_vl_interface': 7.060709965361499e-06, 'learning_rate/action_model': 7.060709965361499e-05, 'epoch': 0.22}) +[2026-05-26 20:41:21,633][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.16496388614177704, 'train/grad_norm_pre_clip': 2.094409704208374, 'eval/loss': 0.09383851289749146, 'timing/data': 0.0005870349705219269, 'timing/model': 0.5816825358197093, 'learning_rate/qwen_vl_interface': 7.032518273177028e-06, 'learning_rate/action_model': 7.032518273177028e-05, 'epoch': 0.22}) +[2026-05-26 20:43:02,383][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.047461993992328644, 'train/grad_norm_pre_clip': 0.6181807518005371, 'timing/data': 0.000635102391242981, 'timing/model': 0.5765340495854616, 'learning_rate/qwen_vl_interface': 7.004253308647993e-06, 'learning_rate/action_model': 7.004253308647994e-05, 'epoch': 0.22}) +[2026-05-26 20:44:40,382][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.10431768745183945, 'train/grad_norm_pre_clip': 0.6949846148490906, 'timing/data': 0.0006896518170833588, 'timing/model': 0.6112555069848895, 'learning_rate/qwen_vl_interface': 6.97591623363638e-06, 'learning_rate/action_model': 6.97591623363638e-05, 'epoch': 0.22}) +[2026-05-26 20:46:19,512][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.1106574609875679, 'train/grad_norm_pre_clip': 0.7408909201622009, 'timing/data': 0.0006298227235674858, 'timing/model': 0.6006851093843579, 'learning_rate/qwen_vl_interface': 6.947508212968356e-06, 'learning_rate/action_model': 6.947508212968356e-05, 'epoch': 0.22}) +[2026-05-26 20:47:59,558][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.1765916645526886, 'train/grad_norm_pre_clip': 1.1639196872711182, 'timing/data': 0.0005825432017445564, 'timing/model': 0.6428398536518216, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.22}) +[2026-05-26 20:49:39,474][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.0038127165753394365, 'train/grad_norm_pre_clip': 0.10338959842920303, 'timing/data': 0.0006383471190929413, 'timing/model': 0.6545379916206002, 'learning_rate/qwen_vl_interface': 6.890484008501211e-06, 'learning_rate/action_model': 6.890484008501211e-05, 'epoch': 0.22}) +[2026-05-26 20:49:39,474][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3250_state +[2026-05-26 20:49:39,480][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-26 20:49:47,425][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3250_state/model.safetensors +[2026-05-26 20:50:05,552][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3250_state/optimizer.bin +[2026-05-26 20:50:05,554][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3250_state/sampler.bin +[2026-05-26 20:50:05,554][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3250_state/sampler_1.bin +[2026-05-26 20:50:05,556][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3250_state/random_states_0.pkl +[2026-05-26 20:50:14,612][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3250 +[2026-05-26 20:50:14,613][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-26 20:50:14,629][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-26 20:51:51,247][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.05683664605021477, 'train/grad_norm_pre_clip': 0.580155074596405, 'timing/data': 0.0038266628980636597, 'timing/model': 0.6272409735247493, 'learning_rate/qwen_vl_interface': 6.861870168743793e-06, 'learning_rate/action_model': 6.861870168743793e-05, 'epoch': 0.22}) +[2026-05-26 20:53:27,462][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.08337611705064774, 'train/grad_norm_pre_clip': 1.501609206199646, 'timing/data': 0.004014521837234497, 'timing/model': 0.6344676837325096, 'learning_rate/qwen_vl_interface': 6.833190071317e-06, 'learning_rate/action_model': 6.833190071316999e-05, 'epoch': 0.22}) +[2026-05-26 20:55:04,582][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.04211893677711487, 'train/grad_norm_pre_clip': 0.6222418546676636, 'timing/data': 0.0038518663495779037, 'timing/model': 0.6402322063222528, 'learning_rate/qwen_vl_interface': 6.804444895147306e-06, 'learning_rate/action_model': 6.804444895147306e-05, 'epoch': 0.22}) +[2026-05-26 20:56:43,139][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.016447344794869423, 'train/grad_norm_pre_clip': 0.6562210321426392, 'timing/data': 0.003827691078186035, 'timing/model': 0.7065273011103272, 'learning_rate/qwen_vl_interface': 6.775635821836317e-06, 'learning_rate/action_model': 6.775635821836317e-05, 'epoch': 0.22}) +[2026-05-26 20:58:38,624][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.35453635454177856, 'train/grad_norm_pre_clip': 4.2332024574279785, 'eval/loss': 0.1925124228000641, 'timing/data': 0.00392948929220438, 'timing/model': 0.8367071850225329, 'learning_rate/qwen_vl_interface': 6.7467640356122e-06, 'learning_rate/action_model': 6.746764035612199e-05, 'epoch': 0.22}) +[2026-05-26 21:00:17,490][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.00010803323675645515, 'train/grad_norm_pre_clip': 0.002220164518803358, 'timing/data': 0.003872073255479336, 'timing/model': 0.6534220222383738, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.22}) +[2026-05-26 21:01:55,682][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.01889924705028534, 'train/grad_norm_pre_clip': 0.4763396978378296, 'timing/data': 0.0042739808559417725, 'timing/model': 0.6505393851548433, 'learning_rate/qwen_vl_interface': 6.688837074177868e-06, 'learning_rate/action_model': 6.688837074177868e-05, 'epoch': 0.23}) +[2026-05-26 21:03:34,050][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.009114044718444347, 'train/grad_norm_pre_clip': 0.10876946151256561, 'timing/data': 0.004026229493319988, 'timing/model': 0.6212353631854057, 'learning_rate/qwen_vl_interface': 6.659784280118156e-06, 'learning_rate/action_model': 6.659784280118156e-05, 'epoch': 0.23}) +[2026-05-26 21:05:12,473][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.19274164736270905, 'train/grad_norm_pre_clip': 1.8201059103012085, 'timing/data': 0.004208150319755077, 'timing/model': 0.645724980160594, 'learning_rate/qwen_vl_interface': 6.6306735353484365e-06, 'learning_rate/action_model': 6.630673535348436e-05, 'epoch': 0.23}) +[2026-05-26 21:06:52,937][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.17036288976669312, 'train/grad_norm_pre_clip': 1.519701600074768, 'timing/data': 0.005213945172727108, 'timing/model': 0.9030404612421989, 'learning_rate/qwen_vl_interface': 6.6015060364974045e-06, 'learning_rate/action_model': 6.601506036497404e-05, 'epoch': 0.23}) +[2026-05-26 21:08:33,965][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.02907198667526245, 'train/grad_norm_pre_clip': 0.28745949268341064, 'timing/data': 0.0038474509492516518, 'timing/model': 0.6534374477341771, 'learning_rate/qwen_vl_interface': 6.572282982526702e-06, 'learning_rate/action_model': 6.572282982526702e-05, 'epoch': 0.23}) +[2026-05-26 21:10:13,775][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.06013656407594681, 'train/grad_norm_pre_clip': 0.4651823341846466, 'timing/data': 0.005000404082238674, 'timing/model': 0.8630224764347076, 'learning_rate/qwen_vl_interface': 6.543005574681613e-06, 'learning_rate/action_model': 6.543005574681613e-05, 'epoch': 0.23}) +[2026-05-26 21:11:51,869][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.004253846127539873, 'train/grad_norm_pre_clip': 0.042882613837718964, 'timing/data': 0.004164204001426697, 'timing/model': 0.6443356471136212, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.23}) +[2026-05-26 21:13:31,688][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.005701273679733276, 'train/grad_norm_pre_clip': 0.1946936398744583, 'timing/data': 0.003909047693014145, 'timing/model': 0.664219593629241, 'learning_rate/qwen_vl_interface': 6.484292513471352e-06, 'learning_rate/action_model': 6.484292513471352e-05, 'epoch': 0.23}) +[2026-05-26 21:15:22,177][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.1974937915802002, 'train/grad_norm_pre_clip': 4.725955486297607, 'eval/loss': 0.10886068642139435, 'timing/data': 0.0037651360034942627, 'timing/model': 0.5942822843790054, 'learning_rate/qwen_vl_interface': 6.45485927357016e-06, 'learning_rate/action_model': 6.454859273570159e-05, 'epoch': 0.23}) +[2026-05-26 21:17:01,196][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.13678070902824402, 'train/grad_norm_pre_clip': 1.1172808408737183, 'timing/data': 0.004534553736448288, 'timing/model': 0.6465096650645137, 'learning_rate/qwen_vl_interface': 6.425376506623344e-06, 'learning_rate/action_model': 6.425376506623344e-05, 'epoch': 0.23}) +[2026-05-26 21:18:40,175][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.042251698672771454, 'train/grad_norm_pre_clip': 0.4085295796394348, 'timing/data': 0.0037001371383666992, 'timing/model': 0.5873258030042052, 'learning_rate/qwen_vl_interface': 6.395845424551975e-06, 'learning_rate/action_model': 6.395845424551975e-05, 'epoch': 0.23}) +[2026-05-26 21:20:19,584][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.16214635968208313, 'train/grad_norm_pre_clip': 2.089195489883423, 'timing/data': 0.004245731979608536, 'timing/model': 0.5942067243158817, 'learning_rate/qwen_vl_interface': 6.3662672412631736e-06, 'learning_rate/action_model': 6.366267241263173e-05, 'epoch': 0.23}) +[2026-05-26 21:21:58,742][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.3451292812824249, 'train/grad_norm_pre_clip': 2.482442617416382, 'timing/data': 0.004041103646159172, 'timing/model': 0.6314859185367823, 'learning_rate/qwen_vl_interface': 6.336643172600202e-06, 'learning_rate/action_model': 6.336643172600202e-05, 'epoch': 0.23}) +[2026-05-26 21:23:38,479][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.00837207492440939, 'train/grad_norm_pre_clip': 0.1827908456325531, 'timing/data': 0.004585094749927521, 'timing/model': 0.6465824795886874, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.23}) +[2026-05-26 21:25:18,091][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.09195075929164886, 'train/grad_norm_pre_clip': 1.8769875764846802, 'timing/data': 0.004282535053789616, 'timing/model': 0.586392349563539, 'learning_rate/qwen_vl_interface': 6.277262251905597e-06, 'learning_rate/action_model': 6.277262251905597e-05, 'epoch': 0.23}) +[2026-05-26 21:26:56,574][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.05166890472173691, 'train/grad_norm_pre_clip': 1.3364331722259521, 'timing/data': 0.004475018009543419, 'timing/model': 0.6316003184765577, 'learning_rate/qwen_vl_interface': 6.247507840791035e-06, 'learning_rate/action_model': 6.247507840791035e-05, 'epoch': 0.24}) +[2026-05-26 21:28:34,484][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.05517042800784111, 'train/grad_norm_pre_clip': 1.0687081813812256, 'timing/data': 0.004157564602792263, 'timing/model': 0.9488389175385237, 'learning_rate/qwen_vl_interface': 6.21771242603611e-06, 'learning_rate/action_model': 6.21771242603611e-05, 'epoch': 0.24}) +[2026-05-26 21:30:14,208][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.043040089309215546, 'train/grad_norm_pre_clip': 0.54569411277771, 'timing/data': 0.003970993682742119, 'timing/model': 0.6774407187476754, 'learning_rate/qwen_vl_interface': 6.187877232413619e-06, 'learning_rate/action_model': 6.18787723241362e-05, 'epoch': 0.24}) +[2026-05-26 21:32:11,065][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.004297118168324232, 'train/grad_norm_pre_clip': 0.09222432225942612, 'eval/loss': 0.10507176071405411, 'timing/data': 0.0038928715512156487, 'timing/model': 0.9328154940158129, 'learning_rate/qwen_vl_interface': 6.15800348633152e-06, 'learning_rate/action_model': 6.15800348633152e-05, 'epoch': 0.24}) +[2026-05-26 21:32:11,066][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3500_state +[2026-05-26 21:32:11,073][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-26 21:32:19,300][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3500_state/model.safetensors +[2026-05-26 21:32:37,906][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3500_state/optimizer.bin +[2026-05-26 21:32:37,907][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3500_state/sampler.bin +[2026-05-26 21:32:37,908][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3500_state/sampler_1.bin +[2026-05-26 21:32:37,910][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3500_state/random_states_0.pkl +[2026-05-26 21:32:47,473][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-05-26 21:32:47,473][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-26 21:32:47,490][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-26 21:34:27,572][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.01782628521323204, 'train/grad_norm_pre_clip': 3.1231558322906494, 'timing/data': 0.004409287124872208, 'timing/model': 0.6529930410906672, 'learning_rate/qwen_vl_interface': 6.128092415782502e-06, 'learning_rate/action_model': 6.128092415782501e-05, 'epoch': 0.24}) +[2026-05-26 21:36:03,972][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.1820734143257141, 'train/grad_norm_pre_clip': 1.9438397884368896, 'timing/data': 0.0037601087242364883, 'timing/model': 0.6545002106577158, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.24}) +[2026-05-26 21:37:43,471][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.007815026678144932, 'train/grad_norm_pre_clip': 0.14609643816947937, 'timing/data': 0.003915681503713131, 'timing/model': 0.5781607860699296, 'learning_rate/qwen_vl_interface': 6.068163220875266e-06, 'learning_rate/action_model': 6.068163220875266e-05, 'epoch': 0.24}) +[2026-05-26 21:39:20,882][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.04938386753201485, 'train/grad_norm_pre_clip': 0.4449097216129303, 'timing/data': 0.003934345208108425, 'timing/model': 0.5989824058488011, 'learning_rate/qwen_vl_interface': 6.038147559971523e-06, 'learning_rate/action_model': 6.038147559971523e-05, 'epoch': 0.24}) +[2026-05-26 21:40:56,745][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.0968659296631813, 'train/grad_norm_pre_clip': 1.647139310836792, 'timing/data': 0.003862030804157257, 'timing/model': 0.6313589429482818, 'learning_rate/qwen_vl_interface': 6.008099501408552e-06, 'learning_rate/action_model': 6.008099501408552e-05, 'epoch': 0.24}) +[2026-05-26 21:42:34,067][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.08772353082895279, 'train/grad_norm_pre_clip': 0.9866785407066345, 'timing/data': 0.0044527435675263405, 'timing/model': 0.9553278908133507, 'learning_rate/qwen_vl_interface': 5.978020280344351e-06, 'learning_rate/action_model': 5.9780202803443506e-05, 'epoch': 0.24}) +[2026-05-26 21:44:10,568][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.07395851612091064, 'train/grad_norm_pre_clip': 1.3544456958770752, 'timing/data': 0.005776555277407169, 'timing/model': 0.8647689064964652, 'learning_rate/qwen_vl_interface': 5.947911133217886e-06, 'learning_rate/action_model': 5.947911133217886e-05, 'epoch': 0.24}) +[2026-05-26 21:45:48,315][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.05404193699359894, 'train/grad_norm_pre_clip': 0.5312137603759766, 'timing/data': 0.0036327457055449486, 'timing/model': 0.6962542477995157, 'learning_rate/qwen_vl_interface': 5.917773297698264e-06, 'learning_rate/action_model': 5.9177732976982636e-05, 'epoch': 0.24}) +[2026-05-26 21:47:25,185][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.1056213229894638, 'train/grad_norm_pre_clip': 0.7087814807891846, 'timing/data': 0.004569591023027897, 'timing/model': 0.6603464735671878, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 0.24}) +[2026-05-26 21:49:17,014][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.27277421951293945, 'train/grad_norm_pre_clip': 1.1654834747314453, 'eval/loss': 0.09425898641347885, 'timing/data': 0.003818044438958168, 'timing/model': 0.6452961284667253, 'learning_rate/qwen_vl_interface': 5.857416518001404e-06, 'learning_rate/action_model': 5.857416518001404e-05, 'epoch': 0.24}) +[2026-05-26 21:50:59,486][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.15277260541915894, 'train/grad_norm_pre_clip': 1.6129262447357178, 'timing/data': 0.003798586316406727, 'timing/model': 0.6644488349556923, 'learning_rate/qwen_vl_interface': 5.8272000548549776e-06, 'learning_rate/action_model': 5.827200054854978e-05, 'epoch': 0.25}) +[2026-05-26 21:52:41,330][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.01092515978962183, 'train/grad_norm_pre_clip': 0.08272045850753784, 'timing/data': 0.004537502303719521, 'timing/model': 0.5970835071057081, 'learning_rate/qwen_vl_interface': 5.7969598652750334e-06, 'learning_rate/action_model': 5.796959865275033e-05, 'epoch': 0.25}) +[2026-05-26 21:54:21,272][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.07566773891448975, 'train/grad_norm_pre_clip': 1.2098983526229858, 'timing/data': 0.004270775243639946, 'timing/model': 0.6386524951085448, 'learning_rate/qwen_vl_interface': 5.766697192317323e-06, 'learning_rate/action_model': 5.766697192317323e-05, 'epoch': 0.25}) +[2026-05-26 21:56:01,966][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.15405185520648956, 'train/grad_norm_pre_clip': 3.5813167095184326, 'timing/data': 0.0039483485743403435, 'timing/model': 0.5973471431061625, 'learning_rate/qwen_vl_interface': 5.7364132799617985e-06, 'learning_rate/action_model': 5.7364132799617984e-05, 'epoch': 0.25}) +[2026-05-26 21:57:39,490][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.2811794579029083, 'train/grad_norm_pre_clip': 3.869723320007324, 'timing/data': 0.003773934207856655, 'timing/model': 0.5926907965913415, 'learning_rate/qwen_vl_interface': 5.706109373061489e-06, 'learning_rate/action_model': 5.706109373061489e-05, 'epoch': 0.25}) +[2026-05-26 21:59:17,055][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.0006535270949825644, 'train/grad_norm_pre_clip': 0.023301219567656517, 'timing/data': 0.0040322598069906235, 'timing/model': 0.6394427567720413, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 0.25}) +[2026-05-26 22:00:54,785][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.14608679711818695, 'train/grad_norm_pre_clip': 1.8317266702651978, 'timing/data': 0.0037723351269960403, 'timing/model': 0.5846682507544756, 'learning_rate/qwen_vl_interface': 5.645446559096875e-06, 'learning_rate/action_model': 5.645446559096875e-05, 'epoch': 0.25}) +[2026-05-26 22:02:32,774][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.1967630535364151, 'train/grad_norm_pre_clip': 2.031273365020752, 'timing/data': 0.004329250194132328, 'timing/model': 0.6524331038817763, 'learning_rate/qwen_vl_interface': 5.6150901456432546e-06, 'learning_rate/action_model': 5.615090145643255e-05, 'epoch': 0.25}) +[2026-05-26 22:04:10,279][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.018806612119078636, 'train/grad_norm_pre_clip': 0.19358713924884796, 'timing/data': 0.004048939794301987, 'timing/model': 0.6456599151715636, 'learning_rate/qwen_vl_interface': 5.584718724763707e-06, 'learning_rate/action_model': 5.584718724763707e-05, 'epoch': 0.25}) +[2026-05-26 22:05:58,761][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.05677512660622597, 'train/grad_norm_pre_clip': 1.2186996936798096, 'eval/loss': 0.08995065093040466, 'timing/data': 0.004024408757686615, 'timing/model': 0.5781978517770767, 'learning_rate/qwen_vl_interface': 5.554333544908388e-06, 'learning_rate/action_model': 5.5543335449083875e-05, 'epoch': 0.25}) +[2026-05-26 22:07:35,672][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.016348350793123245, 'train/grad_norm_pre_clip': 0.21838457882404327, 'timing/data': 0.003766843117773533, 'timing/model': 0.6521524740383029, 'learning_rate/qwen_vl_interface': 5.52393585509303e-06, 'learning_rate/action_model': 5.52393585509303e-05, 'epoch': 0.25}) +[2026-05-26 22:09:14,294][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.041767023503780365, 'train/grad_norm_pre_clip': 0.5718533396720886, 'timing/data': 0.004036573693156242, 'timing/model': 0.6779376287013292, 'learning_rate/qwen_vl_interface': 5.4935269048476015e-06, 'learning_rate/action_model': 5.493526904847601e-05, 'epoch': 0.25}) +[2026-05-26 22:10:52,203][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.025772923603653908, 'train/grad_norm_pre_clip': 0.6301511526107788, 'timing/data': 0.004453741014003754, 'timing/model': 0.6058705588802695, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 0.25}) +[2026-05-26 22:12:28,985][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.07112470269203186, 'train/grad_norm_pre_clip': 1.624993085861206, 'timing/data': 0.003779594786465168, 'timing/model': 0.6181568736210465, 'learning_rate/qwen_vl_interface': 5.432680223449392e-06, 'learning_rate/action_model': 5.4326802234493924e-05, 'epoch': 0.25}) +[2026-05-26 22:14:07,488][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.007373274303972721, 'train/grad_norm_pre_clip': 0.08465670049190521, 'timing/data': 0.004190058447420597, 'timing/model': 0.6640078313648701, 'learning_rate/qwen_vl_interface': 5.402244993465364e-06, 'learning_rate/action_model': 5.402244993465364e-05, 'epoch': 0.25}) +[2026-05-26 22:14:07,495][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3750_state +[2026-05-26 22:14:07,500][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-26 22:14:15,584][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3750_state/model.safetensors +[2026-05-26 22:14:33,888][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3750_state/optimizer.bin +[2026-05-26 22:14:33,889][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3750_state/sampler.bin +[2026-05-26 22:14:33,889][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3750_state/sampler_1.bin +[2026-05-26 22:14:33,891][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_3750_state/random_states_0.pkl +[2026-05-26 22:14:43,365][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3750 +[2026-05-26 22:14:43,365][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-26 22:14:43,380][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-26 22:16:21,169][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.0006238019559532404, 'train/grad_norm_pre_clip': 0.007625292055308819, 'timing/data': 0.004231448285281658, 'timing/model': 0.6572011923417449, 'learning_rate/qwen_vl_interface': 5.371803505285954e-06, 'learning_rate/action_model': 5.371803505285954e-05, 'epoch': 0.26}) +[2026-05-26 22:18:00,785][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.1080920472741127, 'train/grad_norm_pre_clip': 1.8094103336334229, 'timing/data': 0.004096901044249535, 'timing/model': 0.5786508582532406, 'learning_rate/qwen_vl_interface': 5.3413570102415144e-06, 'learning_rate/action_model': 5.3413570102415144e-05, 'epoch': 0.26}) +[2026-05-26 22:19:39,682][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.5234487652778625, 'train/grad_norm_pre_clip': 3.934560775756836, 'timing/data': 0.004704316146671772, 'timing/model': 0.655287534929812, 'learning_rate/qwen_vl_interface': 5.310906759868205e-06, 'learning_rate/action_model': 5.310906759868205e-05, 'epoch': 0.26}) +[2026-05-26 22:21:17,877][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.010083501227200031, 'train/grad_norm_pre_clip': 0.2690524160861969, 'timing/data': 0.004206152632832527, 'timing/model': 0.6094585694372654, 'learning_rate/qwen_vl_interface': 5.280454005856553e-06, 'learning_rate/action_model': 5.2804540058565534e-05, 'epoch': 0.26}) +[2026-05-26 22:23:06,695][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.041008323431015015, 'train/grad_norm_pre_clip': 0.48514869809150696, 'eval/loss': 0.09784986078739166, 'timing/data': 0.003813091665506363, 'timing/model': 0.6559585509821773, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.26}) +[2026-05-26 22:24:47,094][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.21008852124214172, 'train/grad_norm_pre_clip': 1.4518101215362549, 'timing/data': 0.004032104276120663, 'timing/model': 0.5942213125526905, 'learning_rate/qwen_vl_interface': 5.219545994143448e-06, 'learning_rate/action_model': 5.219545994143448e-05, 'epoch': 0.26}) +[2026-05-26 22:26:25,884][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.11843473464250565, 'train/grad_norm_pre_clip': 1.090695858001709, 'timing/data': 0.003898531198501587, 'timing/model': 0.6575041702017188, 'learning_rate/qwen_vl_interface': 5.189093240131795e-06, 'learning_rate/action_model': 5.189093240131795e-05, 'epoch': 0.26}) +[2026-05-26 22:28:03,965][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.02246299758553505, 'train/grad_norm_pre_clip': 0.6071183681488037, 'timing/data': 0.004162733443081379, 'timing/model': 0.871937028132379, 'learning_rate/qwen_vl_interface': 5.158642989758487e-06, 'learning_rate/action_model': 5.158642989758487e-05, 'epoch': 0.26}) +[2026-05-26 22:29:42,888][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.018671201542019844, 'train/grad_norm_pre_clip': 0.15604162216186523, 'timing/data': 0.004670749418437481, 'timing/model': 0.9431929355487227, 'learning_rate/qwen_vl_interface': 5.128196494714048e-06, 'learning_rate/action_model': 5.128196494714047e-05, 'epoch': 0.26}) +[2026-05-26 22:31:20,983][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.025452952831983566, 'train/grad_norm_pre_clip': 0.3228393495082855, 'timing/data': 0.003849657252430916, 'timing/model': 0.5739265708252788, 'learning_rate/qwen_vl_interface': 5.097755006534638e-06, 'learning_rate/action_model': 5.097755006534638e-05, 'epoch': 0.26}) +[2026-05-26 22:32:59,568][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.38472360372543335, 'train/grad_norm_pre_clip': 4.029385089874268, 'timing/data': 0.0039012907072901726, 'timing/model': 0.6403687428683043, 'learning_rate/qwen_vl_interface': 5.067319776550607e-06, 'learning_rate/action_model': 5.067319776550607e-05, 'epoch': 0.26}) +[2026-05-26 22:34:38,093][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.12471058964729309, 'train/grad_norm_pre_clip': 1.5959221124649048, 'timing/data': 0.005255682393908501, 'timing/model': 0.6409768918529153, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 0.26}) +[2026-05-26 22:36:16,689][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.017960313707590103, 'train/grad_norm_pre_clip': 0.18830282986164093, 'timing/data': 0.005056282505393028, 'timing/model': 0.6518903058022261, 'learning_rate/qwen_vl_interface': 5.0064730951524e-06, 'learning_rate/action_model': 5.0064730951524e-05, 'epoch': 0.26}) +[2026-05-26 22:37:54,194][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.010463020764291286, 'train/grad_norm_pre_clip': 0.19567714631557465, 'timing/data': 0.00394287146627903, 'timing/model': 0.6666578967124224, 'learning_rate/qwen_vl_interface': 4.9760641449069724e-06, 'learning_rate/action_model': 4.976064144906972e-05, 'epoch': 0.26}) +[2026-05-26 22:39:44,459][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.07123387604951859, 'train/grad_norm_pre_clip': 1.299281120300293, 'eval/loss': 0.09238965064287186, 'timing/data': 0.003870259039103985, 'timing/model': 0.6239788942039013, 'learning_rate/qwen_vl_interface': 4.945666455091614e-06, 'learning_rate/action_model': 4.9456664550916136e-05, 'epoch': 0.26}) +[2026-05-26 22:41:22,783][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.009218273684382439, 'train/grad_norm_pre_clip': 0.28291836380958557, 'timing/data': 0.0037535177543759346, 'timing/model': 0.581775208003819, 'learning_rate/qwen_vl_interface': 4.9152812752362935e-06, 'learning_rate/action_model': 4.915281275236293e-05, 'epoch': 0.27}) +[2026-05-26 22:43:00,946][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.006944889668375254, 'train/grad_norm_pre_clip': 0.12122193723917007, 'timing/data': 0.004236736334860325, 'timing/model': 0.9212025618180633, 'learning_rate/qwen_vl_interface': 4.8849098543567465e-06, 'learning_rate/action_model': 4.884909854356747e-05, 'epoch': 0.27}) +[2026-05-26 22:44:38,275][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.12391126900911331, 'train/grad_norm_pre_clip': 1.0445811748504639, 'timing/data': 0.004644322209060192, 'timing/model': 0.5748545210808516, 'learning_rate/qwen_vl_interface': 4.854553440903127e-06, 'learning_rate/action_model': 4.854553440903126e-05, 'epoch': 0.27}) +[2026-05-26 22:46:16,482][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.04983207955956459, 'train/grad_norm_pre_clip': 1.3913249969482422, 'timing/data': 0.0038196081295609474, 'timing/model': 0.5853978050872684, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 0.27}) +[2026-05-26 22:47:57,745][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.003177782753482461, 'train/grad_norm_pre_clip': 0.06567288190126419, 'timing/data': 0.00624491274356842, 'timing/model': 0.6094151521101594, 'learning_rate/qwen_vl_interface': 4.793890626938511e-06, 'learning_rate/action_model': 4.793890626938511e-05, 'epoch': 0.27}) +[2026-05-26 22:49:44,296][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.059215836226940155, 'train/grad_norm_pre_clip': 4.786396503448486, 'timing/data': 0.005400028079748154, 'timing/model': 0.6598828928545117, 'learning_rate/qwen_vl_interface': 4.763586720038202e-06, 'learning_rate/action_model': 4.7635867200382014e-05, 'epoch': 0.27}) +[2026-05-26 22:51:22,586][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.20539487898349762, 'train/grad_norm_pre_clip': 2.382770538330078, 'timing/data': 0.0038884207606315613, 'timing/model': 0.6619918970391154, 'learning_rate/qwen_vl_interface': 4.733302807682679e-06, 'learning_rate/action_model': 4.733302807682679e-05, 'epoch': 0.27}) +[2026-05-26 22:53:01,765][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.03722146898508072, 'train/grad_norm_pre_clip': 0.8763986825942993, 'timing/data': 0.003896302543580532, 'timing/model': 0.9399709990248084, 'learning_rate/qwen_vl_interface': 4.7030401347249685e-06, 'learning_rate/action_model': 4.703040134724968e-05, 'epoch': 0.27}) +[2026-05-26 22:54:39,583][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.14833566546440125, 'train/grad_norm_pre_clip': 1.8506097793579102, 'timing/data': 0.0038997987285256386, 'timing/model': 0.6599474428221583, 'learning_rate/qwen_vl_interface': 4.6727999451450236e-06, 'learning_rate/action_model': 4.672799945145023e-05, 'epoch': 0.27}) +[2026-05-26 23:00:33,889][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4000 +[2026-05-26 23:00:33,891][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.03531615808606148, 'train/grad_norm_pre_clip': 0.48704764246940613, 'eval/loss': 0.09198314696550369, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 1.0, 'timing/data': 0.0038163959980010986, 'timing/model': 0.9456783905625343, 'learning_rate/qwen_vl_interface': 4.642583481998596e-06, 'learning_rate/action_model': 4.642583481998596e-05, 'epoch': 0.27}) +[2026-05-26 23:00:33,907][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4000_state +[2026-05-26 23:00:33,914][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-26 23:00:41,817][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4000_state/model.safetensors +[2026-05-26 23:01:00,124][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4000_state/optimizer.bin +[2026-05-26 23:01:00,125][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4000_state/sampler.bin +[2026-05-26 23:01:00,125][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4000_state/sampler_1.bin +[2026-05-26 23:01:00,127][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4000_state/random_states_0.pkl +[2026-05-26 23:01:09,777][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-05-26 23:01:09,777][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-26 23:01:09,793][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-26 23:02:48,165][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.10686945170164108, 'train/grad_norm_pre_clip': 1.075116753578186, 'timing/data': 0.003748862072825432, 'timing/model': 0.6499633230268955, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 0.27}) +[2026-05-26 23:04:28,069][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.1567155122756958, 'train/grad_norm_pre_clip': 1.7511335611343384, 'timing/data': 0.0006413599476218224, 'timing/model': 0.5956631144508719, 'learning_rate/qwen_vl_interface': 4.582226702301737e-06, 'learning_rate/action_model': 4.5822267023017375e-05, 'epoch': 0.27}) +[2026-05-26 23:06:07,776][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.00722080934792757, 'train/grad_norm_pre_clip': 0.1049249917268753, 'timing/data': 0.0006377734243869781, 'timing/model': 0.6511673675850034, 'learning_rate/qwen_vl_interface': 4.552088866782115e-06, 'learning_rate/action_model': 4.552088866782115e-05, 'epoch': 0.27}) +[2026-05-26 23:07:45,830][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.12121540307998657, 'train/grad_norm_pre_clip': 0.83191978931427, 'timing/data': 0.0006168829277157784, 'timing/model': 0.6177751859650016, 'learning_rate/qwen_vl_interface': 4.521979719655648e-06, 'learning_rate/action_model': 4.5219797196556484e-05, 'epoch': 0.27}) +[2026-05-26 23:09:22,287][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.05508778244256973, 'train/grad_norm_pre_clip': 0.6274659633636475, 'timing/data': 0.00060284323990345, 'timing/model': 0.5770964762195945, 'learning_rate/qwen_vl_interface': 4.49190049859145e-06, 'learning_rate/action_model': 4.49190049859145e-05, 'epoch': 0.28}) +[2026-05-26 23:11:00,874][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.400221586227417, 'train/grad_norm_pre_clip': 2.484837532043457, 'timing/data': 0.0005962727591395378, 'timing/model': 0.6567524261772633, 'learning_rate/qwen_vl_interface': 4.461852440028479e-06, 'learning_rate/action_model': 4.461852440028479e-05, 'epoch': 0.28}) +[2026-05-26 23:12:37,972][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.2772865295410156, 'train/grad_norm_pre_clip': 1.6522579193115234, 'timing/data': 0.0005582477897405624, 'timing/model': 0.6519691254943609, 'learning_rate/qwen_vl_interface': 4.431836779124736e-06, 'learning_rate/action_model': 4.431836779124735e-05, 'epoch': 0.28}) +[2026-05-26 23:14:17,163][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.009582129307091236, 'train/grad_norm_pre_clip': 0.21771880984306335, 'timing/data': 0.0004950184375047684, 'timing/model': 0.8555955309420824, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 0.28}) +[2026-05-26 23:15:55,072][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.04667752981185913, 'train/grad_norm_pre_clip': 0.5126103758811951, 'timing/data': 0.0006647706031799316, 'timing/model': 0.6001933244988322, 'learning_rate/qwen_vl_interface': 4.371907584217499e-06, 'learning_rate/action_model': 4.371907584217498e-05, 'epoch': 0.28}) +[2026-05-26 23:17:43,802][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.18737366795539856, 'train/grad_norm_pre_clip': 1.2584855556488037, 'eval/loss': 0.09376498311758041, 'timing/data': 0.00079359021037817, 'timing/model': 0.923469609580934, 'learning_rate/qwen_vl_interface': 4.341996513668483e-06, 'learning_rate/action_model': 4.3419965136684824e-05, 'epoch': 0.28}) +[2026-05-26 23:19:21,259][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.17408713698387146, 'train/grad_norm_pre_clip': 4.141937732696533, 'timing/data': 0.0006393827497959137, 'timing/model': 0.6053272541612387, 'learning_rate/qwen_vl_interface': 4.312122767586381e-06, 'learning_rate/action_model': 4.3121227675863815e-05, 'epoch': 0.28}) +[2026-05-26 23:20:59,674][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.06383973360061646, 'train/grad_norm_pre_clip': 1.0490732192993164, 'timing/data': 0.0006292946636676788, 'timing/model': 0.6586846178397536, 'learning_rate/qwen_vl_interface': 4.2822875739638915e-06, 'learning_rate/action_model': 4.282287573963891e-05, 'epoch': 0.28}) +[2026-05-26 23:22:42,690][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.002141280099749565, 'train/grad_norm_pre_clip': 0.05713220685720444, 'timing/data': 0.004227743484079838, 'timing/model': 0.7682499149814248, 'learning_rate/qwen_vl_interface': 4.252492159208966e-06, 'learning_rate/action_model': 4.252492159208966e-05, 'epoch': 0.28}) +[2026-05-26 23:24:24,475][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.03005269728600979, 'train/grad_norm_pre_clip': 0.3152170181274414, 'timing/data': 0.004152695648372173, 'timing/model': 0.6588122006505728, 'learning_rate/qwen_vl_interface': 4.222737748094404e-06, 'learning_rate/action_model': 4.222737748094404e-05, 'epoch': 0.28}) +[2026-05-26 23:26:09,175][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.08059433847665787, 'train/grad_norm_pre_clip': 0.685465931892395, 'timing/data': 0.004197048023343086, 'timing/model': 0.6597116682678461, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 0.28}) +[2026-05-26 23:27:48,585][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.11153312772512436, 'train/grad_norm_pre_clip': 3.9091615676879883, 'timing/data': 0.0044132377952337265, 'timing/model': 0.9390105176717043, 'learning_rate/qwen_vl_interface': 4.1633568273998e-06, 'learning_rate/action_model': 4.1633568273998e-05, 'epoch': 0.28}) +[2026-05-26 23:29:30,988][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.09019219875335693, 'train/grad_norm_pre_clip': 0.8604316711425781, 'timing/data': 0.003959936089813709, 'timing/model': 0.769984532147646, 'learning_rate/qwen_vl_interface': 4.133732758736828e-06, 'learning_rate/action_model': 4.133732758736828e-05, 'epoch': 0.28}) +[2026-05-26 23:31:13,331][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.34161701798439026, 'train/grad_norm_pre_clip': 2.926522970199585, 'timing/data': 0.0051228757947683334, 'timing/model': 1.1900610094889998, 'learning_rate/qwen_vl_interface': 4.104154575448026e-06, 'learning_rate/action_model': 4.104154575448025e-05, 'epoch': 0.28}) +[2026-05-26 23:33:24,364][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.030470816418528557, 'train/grad_norm_pre_clip': 0.5533446669578552, 'timing/data': 0.006471499800682068, 'timing/model': 0.6353441886603832, 'learning_rate/qwen_vl_interface': 4.074623493376657e-06, 'learning_rate/action_model': 4.074623493376657e-05, 'epoch': 0.28}) +[2026-05-26 23:35:21,530][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 4.395826181280427e-06, 'train/grad_norm_pre_clip': 6.631208816543221e-05, 'eval/loss': 0.10828929394483566, 'timing/data': 0.004816616885364056, 'timing/model': 0.6758771911263466, 'learning_rate/qwen_vl_interface': 4.045140726429841e-06, 'learning_rate/action_model': 4.0451407264298406e-05, 'epoch': 0.29}) +[2026-05-26 23:37:06,610][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 1.4237558389140759e-05, 'train/grad_norm_pre_clip': 0.00047727799392305315, 'timing/data': 0.0050701480358839035, 'timing/model': 0.6569062685593963, 'learning_rate/qwen_vl_interface': 4.015707486528649e-06, 'learning_rate/action_model': 4.015707486528649e-05, 'epoch': 0.29}) +[2026-05-26 23:38:48,476][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.11014848947525024, 'train/grad_norm_pre_clip': 3.7722647190093994, 'timing/data': 0.003742951899766922, 'timing/model': 0.6580478986725211, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 0.29}) +[2026-05-26 23:40:29,181][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.11594397574663162, 'train/grad_norm_pre_clip': 1.6517680883407593, 'timing/data': 0.0038534002378582954, 'timing/model': 0.6680003330111504, 'learning_rate/qwen_vl_interface': 3.956994425318387e-06, 'learning_rate/action_model': 3.956994425318387e-05, 'epoch': 0.29}) +[2026-05-26 23:42:10,793][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.016821954399347305, 'train/grad_norm_pre_clip': 0.269749253988266, 'timing/data': 0.003946245647966862, 'timing/model': 0.7759297313168645, 'learning_rate/qwen_vl_interface': 3.927717017473301e-06, 'learning_rate/action_model': 3.9277170174733006e-05, 'epoch': 0.29}) +[2026-05-26 23:43:50,991][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.03745851293206215, 'train/grad_norm_pre_clip': 0.5605744123458862, 'timing/data': 0.0044334447011351585, 'timing/model': 0.6574184149503708, 'learning_rate/qwen_vl_interface': 3.8984939635025966e-06, 'learning_rate/action_model': 3.898493963502596e-05, 'epoch': 0.29}) +[2026-05-26 23:43:50,992][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4250_state +[2026-05-26 23:43:50,999][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-26 23:43:59,529][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4250_state/model.safetensors +[2026-05-26 23:44:17,837][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4250_state/optimizer.bin +[2026-05-26 23:44:17,851][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4250_state/sampler.bin +[2026-05-26 23:44:17,851][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4250_state/sampler_1.bin +[2026-05-26 23:44:17,853][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4250_state/random_states_0.pkl +[2026-05-26 23:44:27,657][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4250 +[2026-05-26 23:44:27,657][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-26 23:44:27,674][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-26 23:46:10,756][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.00017502183618489653, 'train/grad_norm_pre_clip': 0.004727798514068127, 'timing/data': 0.00382876954972744, 'timing/model': 0.9396156081929803, 'learning_rate/qwen_vl_interface': 3.869326464651566e-06, 'learning_rate/action_model': 3.869326464651566e-05, 'epoch': 0.29}) +[2026-05-26 23:47:55,754][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.0023034645710140467, 'train/grad_norm_pre_clip': 0.058354929089546204, 'timing/data': 0.0058157192543148994, 'timing/model': 0.6462180372327566, 'learning_rate/qwen_vl_interface': 3.840215719881844e-06, 'learning_rate/action_model': 3.8402157198818445e-05, 'epoch': 0.29}) +[2026-05-26 23:49:41,370][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.03661550208926201, 'train/grad_norm_pre_clip': 1.0157310962677002, 'timing/data': 0.0037218937650322914, 'timing/model': 0.8551424387842417, 'learning_rate/qwen_vl_interface': 3.811162925822133e-06, 'learning_rate/action_model': 3.811162925822132e-05, 'epoch': 0.29}) +[2026-05-26 23:51:23,078][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.17287947237491608, 'train/grad_norm_pre_clip': 1.6975352764129639, 'timing/data': 0.0038796961307525635, 'timing/model': 0.6495179114863276, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 0.29}) +[2026-05-26 23:53:21,458][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.4076462388038635, 'train/grad_norm_pre_clip': 5.751103401184082, 'eval/loss': 0.10588778555393219, 'timing/data': 0.003981993533670902, 'timing/model': 0.644306474365294, 'learning_rate/qwen_vl_interface': 3.753235964387802e-06, 'learning_rate/action_model': 3.753235964387802e-05, 'epoch': 0.29}) +[2026-05-26 23:55:04,185][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.05461639538407326, 'train/grad_norm_pre_clip': 3.1183807849884033, 'timing/data': 0.0041757747530937195, 'timing/model': 0.6601712144911289, 'learning_rate/qwen_vl_interface': 3.7243641781636838e-06, 'learning_rate/action_model': 3.7243641781636836e-05, 'epoch': 0.29}) +[2026-05-26 23:56:51,921][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.07616892457008362, 'train/grad_norm_pre_clip': 3.127944231033325, 'timing/data': 0.00416972953826189, 'timing/model': 0.6044304268434644, 'learning_rate/qwen_vl_interface': 3.695555104852695e-06, 'learning_rate/action_model': 3.6955551048526946e-05, 'epoch': 0.29}) +[2026-05-26 23:58:33,211][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.25669386982917786, 'train/grad_norm_pre_clip': 3.092050790786743, 'timing/data': 0.004046362824738026, 'timing/model': 0.6711964635178447, 'learning_rate/qwen_vl_interface': 3.6668099286830023e-06, 'learning_rate/action_model': 3.666809928683002e-05, 'epoch': 0.29}) +[2026-05-27 00:00:15,961][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.3609616756439209, 'train/grad_norm_pre_clip': 5.721020698547363, 'timing/data': 0.0040982309728860855, 'timing/model': 0.6614955952391028, 'learning_rate/qwen_vl_interface': 3.6381298312562095e-06, 'learning_rate/action_model': 3.6381298312562094e-05, 'epoch': 0.29}) +[2026-05-27 00:02:02,182][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.08614619076251984, 'train/grad_norm_pre_clip': 0.9880962371826172, 'timing/data': 0.004614273086190224, 'timing/model': 0.5908605018630624, 'learning_rate/qwen_vl_interface': 3.6095159914987893e-06, 'learning_rate/action_model': 3.6095159914987894e-05, 'epoch': 0.3}) +[2026-05-27 00:03:47,149][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.030658317729830742, 'train/grad_norm_pre_clip': 0.48602402210235596, 'timing/data': 0.010637756437063217, 'timing/model': 1.678621823899448, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 0.3}) +[2026-05-27 00:05:32,868][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.07617636024951935, 'train/grad_norm_pre_clip': 1.8753032684326172, 'timing/data': 0.004148310050368309, 'timing/model': 0.6461708936840296, 'learning_rate/qwen_vl_interface': 3.552491787031647e-06, 'learning_rate/action_model': 3.552491787031647e-05, 'epoch': 0.3}) +[2026-05-27 00:07:13,671][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.13328784704208374, 'train/grad_norm_pre_clip': 1.3579630851745605, 'timing/data': 0.003842916339635849, 'timing/model': 0.6363228932023048, 'learning_rate/qwen_vl_interface': 3.5240837663636207e-06, 'learning_rate/action_model': 3.5240837663636204e-05, 'epoch': 0.3}) +[2026-05-27 00:08:57,950][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.05607656389474869, 'train/grad_norm_pre_clip': 0.9203079342842102, 'timing/data': 0.0037941262125968933, 'timing/model': 0.9268244141712785, 'learning_rate/qwen_vl_interface': 3.495746691352008e-06, 'learning_rate/action_model': 3.495746691352008e-05, 'epoch': 0.3}) +[2026-05-27 00:10:54,667][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.03184756636619568, 'train/grad_norm_pre_clip': 0.44380810856819153, 'eval/loss': 0.09660865366458893, 'timing/data': 0.004206745885312557, 'timing/model': 0.6264750575646758, 'learning_rate/qwen_vl_interface': 3.4674817268229732e-06, 'learning_rate/action_model': 3.467481726822973e-05, 'epoch': 0.3}) +[2026-05-27 00:12:37,567][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.12992297112941742, 'train/grad_norm_pre_clip': 1.7980862855911255, 'timing/data': 0.003931496292352676, 'timing/model': 0.65312060713768, 'learning_rate/qwen_vl_interface': 3.4392900346385022e-06, 'learning_rate/action_model': 3.4392900346385024e-05, 'epoch': 0.3}) +[2026-05-27 00:14:17,186][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.05034622550010681, 'train/grad_norm_pre_clip': 1.3643839359283447, 'timing/data': 0.004257560707628727, 'timing/model': 0.5790065368637443, 'learning_rate/qwen_vl_interface': 3.411172773648642e-06, 'learning_rate/action_model': 3.411172773648642e-05, 'epoch': 0.3}) +[2026-05-27 00:16:00,828][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.01947776973247528, 'train/grad_norm_pre_clip': 0.47606125473976135, 'timing/data': 0.0038338275626301765, 'timing/model': 0.6689196797087789, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 0.3}) +[2026-05-27 00:17:44,983][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.036634404212236404, 'train/grad_norm_pre_clip': 0.8401055932044983, 'timing/data': 0.004138272255659103, 'timing/model': 0.578110508620739, 'learning_rate/qwen_vl_interface': 3.3551661653075522e-06, 'learning_rate/action_model': 3.355166165307552e-05, 'epoch': 0.3}) +[2026-05-27 00:19:25,566][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 2.0294111891416833e-05, 'train/grad_norm_pre_clip': 0.0006144348881207407, 'timing/data': 0.0038231126964092255, 'timing/model': 0.6440225355327129, 'learning_rate/qwen_vl_interface': 3.3272791201686294e-06, 'learning_rate/action_model': 3.327279120168629e-05, 'epoch': 0.3}) +[2026-05-27 00:21:09,695][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.0009340762626379728, 'train/grad_norm_pre_clip': 0.017382066696882248, 'timing/data': 0.004862253554165363, 'timing/model': 0.6617981353774667, 'learning_rate/qwen_vl_interface': 3.299471110554295e-06, 'learning_rate/action_model': 3.299471110554295e-05, 'epoch': 0.3}) +[2026-05-27 00:22:59,377][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.07228121161460876, 'train/grad_norm_pre_clip': 0.7371182441711426, 'timing/data': 0.0039049554616212845, 'timing/model': 0.6440425980836153, 'learning_rate/qwen_vl_interface': 3.2717432795429057e-06, 'learning_rate/action_model': 3.271743279542905e-05, 'epoch': 0.3}) +[2026-05-27 00:24:47,184][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.045117221772670746, 'train/grad_norm_pre_clip': 0.6256006956100464, 'timing/data': 0.0037794793024659157, 'timing/model': 0.620971605181694, 'learning_rate/qwen_vl_interface': 3.244096766916992e-06, 'learning_rate/action_model': 3.244096766916992e-05, 'epoch': 0.3}) +[2026-05-27 00:26:33,736][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.02488490380346775, 'train/grad_norm_pre_clip': 1.8556376695632935, 'timing/data': 0.004469356499612331, 'timing/model': 0.9866552986204624, 'learning_rate/qwen_vl_interface': 3.216532709116405e-06, 'learning_rate/action_model': 3.216532709116405e-05, 'epoch': 0.3}) +[2026-05-27 00:28:31,434][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.07560774683952332, 'train/grad_norm_pre_clip': 0.7204391956329346, 'eval/loss': 0.0903303250670433, 'timing/data': 0.0006680255755782127, 'timing/model': 0.9177891807630658, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 0.31}) +[2026-05-27 00:28:31,436][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4500_state +[2026-05-27 00:28:31,445][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-27 00:28:39,494][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4500_state/model.safetensors +[2026-05-27 00:28:57,618][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4500_state/optimizer.bin +[2026-05-27 00:28:57,619][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4500_state/sampler.bin +[2026-05-27 00:28:57,619][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4500_state/sampler_1.bin +[2026-05-27 00:28:57,621][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4500_state/random_states_0.pkl +[2026-05-27 00:29:06,278][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-05-27 00:29:06,278][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-27 00:29:06,294][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-27 00:30:50,600][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.18798145651817322, 'train/grad_norm_pre_clip': 2.647266149520874, 'timing/data': 0.003812515176832676, 'timing/model': 0.603103019297123, 'learning_rate/qwen_vl_interface': 3.1616564867570603e-06, 'learning_rate/action_model': 3.16165648675706e-05, 'epoch': 0.31}) +[2026-05-27 00:32:34,210][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.07683995366096497, 'train/grad_norm_pre_clip': 4.836996555328369, 'timing/data': 0.008222161792218685, 'timing/model': 0.774507800117135, 'learning_rate/qwen_vl_interface': 3.134346577944869e-06, 'learning_rate/action_model': 3.134346577944869e-05, 'epoch': 0.31}) +[2026-05-27 00:34:16,699][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.05868115276098251, 'train/grad_norm_pre_clip': 0.6049941182136536, 'timing/data': 0.004458891227841377, 'timing/model': 0.6797285145148635, 'learning_rate/qwen_vl_interface': 3.10712363535841e-06, 'learning_rate/action_model': 3.10712363535841e-05, 'epoch': 0.31}) +[2026-05-27 00:35:57,514][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.0010338794672861695, 'train/grad_norm_pre_clip': 0.01891089417040348, 'timing/data': 0.004467297345399857, 'timing/model': 0.6186447599902749, 'learning_rate/qwen_vl_interface': 3.0799887780262295e-06, 'learning_rate/action_model': 3.0799887780262296e-05, 'epoch': 0.31}) +[2026-05-27 00:37:41,505][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.007100795861333609, 'train/grad_norm_pre_clip': 0.1766611784696579, 'timing/data': 0.005543752573430538, 'timing/model': 0.6518099075183272, 'learning_rate/qwen_vl_interface': 3.0529431213560334e-06, 'learning_rate/action_model': 3.0529431213560334e-05, 'epoch': 0.31}) +[2026-05-27 00:39:27,315][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.12839041650295258, 'train/grad_norm_pre_clip': 1.417944073677063, 'timing/data': 0.003842010162770748, 'timing/model': 0.6126853637397289, 'learning_rate/qwen_vl_interface': 3.0259877770888334e-06, 'learning_rate/action_model': 3.0259877770888333e-05, 'epoch': 0.31}) +[2026-05-27 00:41:08,894][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 6.238946662051603e-05, 'train/grad_norm_pre_clip': 0.0020465455017983913, 'timing/data': 0.007082484662532806, 'timing/model': 0.6571221081539989, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 0.31}) +[2026-05-27 00:42:57,637][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.061523668467998505, 'train/grad_norm_pre_clip': 0.7620267271995544, 'timing/data': 0.004673285409808159, 'timing/model': 0.974966237321496, 'learning_rate/qwen_vl_interface': 2.9723524541199943e-06, 'learning_rate/action_model': 2.972352454119994e-05, 'epoch': 0.31}) +[2026-05-27 00:44:39,601][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.04057856649160385, 'train/grad_norm_pre_clip': 0.7465488314628601, 'timing/data': 0.004025071859359741, 'timing/model': 0.6659661503508687, 'learning_rate/qwen_vl_interface': 2.9456746801564072e-06, 'learning_rate/action_model': 2.945674680156407e-05, 'epoch': 0.31}) +[2026-05-27 00:46:34,684][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.07647489756345749, 'train/grad_norm_pre_clip': 2.489164113998413, 'eval/loss': 0.09763035923242569, 'timing/data': 0.006043531931936741, 'timing/model': 0.9374774880707264, 'learning_rate/qwen_vl_interface': 2.919091627981297e-06, 'learning_rate/action_model': 2.919091627981297e-05, 'epoch': 0.31}) +[2026-05-27 00:48:18,082][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.10737712681293488, 'train/grad_norm_pre_clip': 2.2861783504486084, 'timing/data': 0.003710942342877388, 'timing/model': 0.5795809431001544, 'learning_rate/qwen_vl_interface': 2.8926043903198135e-06, 'learning_rate/action_model': 2.8926043903198135e-05, 'epoch': 0.31}) +[2026-05-27 00:50:01,554][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.06341750919818878, 'train/grad_norm_pre_clip': 1.5424822568893433, 'timing/data': 0.003944512456655502, 'timing/model': 0.6500135296955705, 'learning_rate/qwen_vl_interface': 2.8662140559585584e-06, 'learning_rate/action_model': 2.866214055958558e-05, 'epoch': 0.31}) +[2026-05-27 00:51:43,586][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.02910969778895378, 'train/grad_norm_pre_clip': 0.3341456353664398, 'timing/data': 0.0037636160850524902, 'timing/model': 0.57592604868114, 'learning_rate/qwen_vl_interface': 2.83992170970081e-06, 'learning_rate/action_model': 2.8399217097008102e-05, 'epoch': 0.31}) +[2026-05-27 00:53:21,887][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.10910298675298691, 'train/grad_norm_pre_clip': 1.4151363372802734, 'timing/data': 0.004138144664466381, 'timing/model': 0.655346293002367, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 0.32}) +[2026-05-27 00:55:05,675][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.004898442886769772, 'train/grad_norm_pre_clip': 0.08638259768486023, 'timing/data': 0.0037244055420160294, 'timing/model': 0.6404060600325465, 'learning_rate/qwen_vl_interface': 2.787635300525006e-06, 'learning_rate/action_model': 2.787635300525006e-05, 'epoch': 0.32}) +[2026-05-27 00:56:50,792][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.05327408015727997, 'train/grad_norm_pre_clip': 0.999075710773468, 'timing/data': 0.005943465046584606, 'timing/model': 0.6630717013031244, 'learning_rate/qwen_vl_interface': 2.7616433868964387e-06, 'learning_rate/action_model': 2.7616433868964388e-05, 'epoch': 0.32}) +[2026-05-27 00:58:31,474][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.009271148592233658, 'train/grad_norm_pre_clip': 0.14877168834209442, 'timing/data': 0.004216904751956463, 'timing/model': 0.6506133070215583, 'learning_rate/qwen_vl_interface': 2.73575375986201e-06, 'learning_rate/action_model': 2.7357537598620096e-05, 'epoch': 0.32}) +[2026-05-27 01:00:11,472][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.11503604799509048, 'train/grad_norm_pre_clip': 1.4639263153076172, 'timing/data': 0.004580700770020485, 'timing/model': 0.9557233983650804, 'learning_rate/qwen_vl_interface': 2.7099674836428842e-06, 'learning_rate/action_model': 2.709967483642884e-05, 'epoch': 0.32}) +[2026-05-27 01:01:51,294][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.19327972829341888, 'train/grad_norm_pre_clip': 1.4215658903121948, 'timing/data': 0.003906877711415291, 'timing/model': 0.9611859675496817, 'learning_rate/qwen_vl_interface': 2.68428561821188e-06, 'learning_rate/action_model': 2.6842856182118802e-05, 'epoch': 0.32}) +[2026-05-27 01:03:46,652][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.0812092199921608, 'train/grad_norm_pre_clip': 0.7270421981811523, 'eval/loss': 0.09559523314237595, 'timing/data': 0.004031081683933735, 'timing/model': 0.5988786481320858, 'learning_rate/qwen_vl_interface': 2.6587092192498943e-06, 'learning_rate/action_model': 2.658709219249894e-05, 'epoch': 0.32}) +[2026-05-27 01:05:30,586][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.16763941943645477, 'train/grad_norm_pre_clip': 1.4221203327178955, 'timing/data': 0.0040169088169932365, 'timing/model': 0.6443325709551573, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 0.32}) +[2026-05-27 01:07:10,891][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.11741085350513458, 'train/grad_norm_pre_clip': 1.7841850519180298, 'timing/data': 0.003802315331995487, 'timing/model': 0.5805232552811503, 'learning_rate/qwen_vl_interface': 2.607877021736793e-06, 'learning_rate/action_model': 2.6078770217367926e-05, 'epoch': 0.32}) +[2026-05-27 01:08:53,319][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.0015353979542851448, 'train/grad_norm_pre_clip': 0.06055494397878647, 'timing/data': 0.004044496454298496, 'timing/model': 0.6111092492938042, 'learning_rate/qwen_vl_interface': 2.5826233126982198e-06, 'learning_rate/action_model': 2.5826233126982196e-05, 'epoch': 0.32}) +[2026-05-27 01:10:36,078][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.019561538472771645, 'train/grad_norm_pre_clip': 0.9241210222244263, 'timing/data': 0.003959390334784985, 'timing/model': 0.637873406521976, 'learning_rate/qwen_vl_interface': 2.5574792490678648e-06, 'learning_rate/action_model': 2.5574792490678644e-05, 'epoch': 0.32}) +[2026-05-27 01:12:18,698][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.010471037589013577, 'train/grad_norm_pre_clip': 0.2961483299732208, 'timing/data': 0.004412271082401276, 'timing/model': 0.5988751342520118, 'learning_rate/qwen_vl_interface': 2.532445864419695e-06, 'learning_rate/action_model': 2.5324458644196952e-05, 'epoch': 0.32}) +[2026-05-27 01:12:18,699][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4750_state +[2026-05-27 01:12:18,705][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-27 01:12:27,143][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4750_state/model.safetensors +[2026-05-27 01:12:45,651][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4750_state/optimizer.bin +[2026-05-27 01:12:45,653][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4750_state/sampler.bin +[2026-05-27 01:12:45,653][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4750_state/sampler_1.bin +[2026-05-27 01:12:45,656][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_4750_state/random_states_0.pkl +[2026-05-27 01:12:55,597][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4750 +[2026-05-27 01:12:55,597][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-27 01:12:55,614][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-27 01:14:41,313][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.40553563833236694, 'train/grad_norm_pre_clip': 1.782737135887146, 'timing/data': 0.003909481689333916, 'timing/model': 0.616437322460115, 'learning_rate/qwen_vl_interface': 2.5075241877781085e-06, 'learning_rate/action_model': 2.5075241877781086e-05, 'epoch': 0.32}) +[2026-05-27 01:16:23,883][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.38866251707077026, 'train/grad_norm_pre_clip': 3.3478798866271973, 'timing/data': 0.00392016489058733, 'timing/model': 0.5796186020597816, 'learning_rate/qwen_vl_interface': 2.4827152435756185e-06, 'learning_rate/action_model': 2.4827152435756185e-05, 'epoch': 0.32}) +[2026-05-27 01:18:11,282][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.018112290650606155, 'train/grad_norm_pre_clip': 0.5767841935157776, 'timing/data': 0.003981750458478928, 'timing/model': 0.6225272687152028, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 0.32}) +[2026-05-27 01:19:53,069][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.022824939340353012, 'train/grad_norm_pre_clip': 0.75069659948349, 'timing/data': 0.003932373598217964, 'timing/model': 0.6493536746129394, 'learning_rate/qwen_vl_interface': 2.433439627006128e-06, 'learning_rate/action_model': 2.4334396270061277e-05, 'epoch': 0.33}) +[2026-05-27 01:21:53,951][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.047876987606287, 'train/grad_norm_pre_clip': 0.8509159088134766, 'eval/loss': 0.09912953525781631, 'timing/data': 0.003981930203735828, 'timing/model': 0.6739630606025457, 'learning_rate/qwen_vl_interface': 2.408974980166724e-06, 'learning_rate/action_model': 2.408974980166724e-05, 'epoch': 0.33}) +[2026-05-27 01:23:36,686][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.07291857153177261, 'train/grad_norm_pre_clip': 1.272862195968628, 'timing/data': 0.0037809470668435097, 'timing/model': 0.5872159451246262, 'learning_rate/qwen_vl_interface': 2.3846271167383515e-06, 'learning_rate/action_model': 2.3846271167383515e-05, 'epoch': 0.33}) +[2026-05-27 01:25:26,922][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.001954613719135523, 'train/grad_norm_pre_clip': 0.0506037175655365, 'timing/data': 0.004402985796332359, 'timing/model': 0.616208023391664, 'learning_rate/qwen_vl_interface': 2.3603970375663123e-06, 'learning_rate/action_model': 2.3603970375663125e-05, 'epoch': 0.33}) +[2026-05-27 01:27:11,184][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.0881309062242508, 'train/grad_norm_pre_clip': 6.68704891204834, 'timing/data': 0.004283878952264786, 'timing/model': 0.863106407225132, 'learning_rate/qwen_vl_interface': 2.33628573865426e-06, 'learning_rate/action_model': 2.3362857386542596e-05, 'epoch': 0.33}) +[2026-05-27 01:28:59,889][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.009846537373960018, 'train/grad_norm_pre_clip': 0.25911033153533936, 'timing/data': 0.004058553837239742, 'timing/model': 0.6650179559364915, 'learning_rate/qwen_vl_interface': 2.3122942111232506e-06, 'learning_rate/action_model': 2.3122942111232506e-05, 'epoch': 0.33}) +[2026-05-27 01:30:55,771][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.0012979365419596434, 'train/grad_norm_pre_clip': 0.0408506877720356, 'timing/data': 0.003930766135454178, 'timing/model': 0.6451136488467455, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 0.33}) +[2026-05-27 01:32:35,192][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.050054408609867096, 'train/grad_norm_pre_clip': 1.668195128440857, 'timing/data': 0.003982880152761936, 'timing/model': 0.5794201204553246, 'learning_rate/qwen_vl_interface': 2.264674410031413e-06, 'learning_rate/action_model': 2.2646744100314127e-05, 'epoch': 0.33}) +[2026-05-27 01:34:19,101][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.04393203184008598, 'train/grad_norm_pre_clip': 0.537581205368042, 'timing/data': 0.003714798018336296, 'timing/model': 0.6649329615756869, 'learning_rate/qwen_vl_interface': 2.2410480939340945e-06, 'learning_rate/action_model': 2.2410480939340945e-05, 'epoch': 0.33}) +[2026-05-27 01:36:00,688][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.23057271540164948, 'train/grad_norm_pre_clip': 2.368875741958618, 'timing/data': 0.0040127867832779884, 'timing/model': 0.6605108948424459, 'learning_rate/qwen_vl_interface': 2.217545464064374e-06, 'learning_rate/action_model': 2.217545464064374e-05, 'epoch': 0.33}) +[2026-05-27 01:37:41,002][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.00109999964479357, 'train/grad_norm_pre_clip': 0.03075212426483631, 'timing/data': 0.005206108093261719, 'timing/model': 0.5767220295965672, 'learning_rate/qwen_vl_interface': 2.1941674865233144e-06, 'learning_rate/action_model': 2.1941674865233143e-05, 'epoch': 0.33}) +[2026-05-27 01:39:40,446][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.022067701444029808, 'train/grad_norm_pre_clip': 1.376526117324829, 'eval/loss': 0.09517614543437958, 'timing/data': 0.003908238373696804, 'timing/model': 0.6664375476539135, 'learning_rate/qwen_vl_interface': 2.170915122288006e-06, 'learning_rate/action_model': 2.170915122288006e-05, 'epoch': 0.33}) +[2026-05-27 01:41:32,784][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.15565873682498932, 'train/grad_norm_pre_clip': 2.3039612770080566, 'timing/data': 0.004270482808351517, 'timing/model': 0.652202962897718, 'learning_rate/qwen_vl_interface': 2.1477893271720683e-06, 'learning_rate/action_model': 2.1477893271720682e-05, 'epoch': 0.33}) +[2026-05-27 01:43:15,084][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.14112576842308044, 'train/grad_norm_pre_clip': 2.4770400524139404, 'timing/data': 0.0037685465067625046, 'timing/model': 0.6603600075468421, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 0.33}) +[2026-05-27 01:45:01,999][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.046524371951818466, 'train/grad_norm_pre_clip': 1.879828691482544, 'timing/data': 0.008273076266050339, 'timing/model': 0.6691296799108386, 'learning_rate/qwen_vl_interface': 2.1019212414999163e-06, 'learning_rate/action_model': 2.101921241499916e-05, 'epoch': 0.33}) +[2026-05-27 01:47:00,787][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.1302858591079712, 'train/grad_norm_pre_clip': 3.534210205078125, 'timing/data': 0.003839019685983658, 'timing/model': 0.6604712335392833, 'learning_rate/qwen_vl_interface': 2.079180836401047e-06, 'learning_rate/action_model': 2.079180836401047e-05, 'epoch': 0.34}) +[2026-05-27 01:48:56,569][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.05852755159139633, 'train/grad_norm_pre_clip': 1.5654160976409912, 'timing/data': 0.005660046823322773, 'timing/model': 0.6195853939279914, 'learning_rate/qwen_vl_interface': 2.0565707712587447e-06, 'learning_rate/action_model': 2.0565707712587444e-05, 'epoch': 0.34}) +[2026-05-27 01:50:43,382][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.04567711427807808, 'train/grad_norm_pre_clip': 1.0818926095962524, 'timing/data': 0.004077313467860222, 'timing/model': 0.6001663394272327, 'learning_rate/qwen_vl_interface': 2.034091975484231e-06, 'learning_rate/action_model': 2.0340919754842307e-05, 'epoch': 0.34}) +[2026-05-27 01:52:31,700][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.11472557485103607, 'train/grad_norm_pre_clip': 1.7240033149719238, 'timing/data': 0.005392302758991718, 'timing/model': 0.6677939239889383, 'learning_rate/qwen_vl_interface': 2.011745373092753e-06, 'learning_rate/action_model': 2.011745373092753e-05, 'epoch': 0.34}) +[2026-05-27 01:54:18,888][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.030700838193297386, 'train/grad_norm_pre_clip': 0.36538830399513245, 'timing/data': 0.003994219936430454, 'timing/model': 0.6413579294458032, 'learning_rate/qwen_vl_interface': 1.9895318826656116e-06, 'learning_rate/action_model': 1.9895318826656114e-05, 'epoch': 0.34}) +[2026-05-27 01:56:08,311][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.035793863236904144, 'train/grad_norm_pre_clip': 0.8305042386054993, 'timing/data': 0.00443978700786829, 'timing/model': 0.6290302006527781, 'learning_rate/qwen_vl_interface': 1.967452417312393e-06, 'learning_rate/action_model': 1.9674524173123927e-05, 'epoch': 0.34}) +[2026-05-27 02:02:54,184][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 5000 +[2026-05-27 02:02:54,186][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.03909922391176224, 'train/grad_norm_pre_clip': 0.5186278223991394, 'eval/loss': 0.09602975100278854, 'rl_games_eval/mid_train/total_episodes': 1.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 1.0, 'timing/data': 0.004068090580403805, 'timing/model': 0.6679931888356805, 'learning_rate/qwen_vl_interface': 1.9455078846334397e-06, 'learning_rate/action_model': 1.9455078846334395e-05, 'epoch': 0.34}) +[2026-05-27 02:02:54,205][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_5000_state +[2026-05-27 02:02:54,212][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-05-27 02:03:02,790][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_5000_state/model.safetensors +[2026-05-27 02:03:21,751][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_5000_state/optimizer.bin +[2026-05-27 02:03:21,753][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_5000_state/sampler.bin +[2026-05-27 02:03:21,753][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_5000_state/sampler_1.bin +[2026-05-27 02:03:21,756][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints/steps_5000_state/random_states_0.pkl +[2026-05-27 02:03:31,593][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-05-27 02:03:31,593][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-05-27 02:03:31,612][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-05-27 02:03:31,613][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0/checkpoints +[2026-05-27 02:27:53,353][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 5000 +[2026-05-27 02:27:53,533][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/summary.jsonl b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..51346d8cd3b3802a080bff60110fcbd21e866788 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/summary.jsonl @@ -0,0 +1,20 @@ +{"steps": 250} +{"steps": 500} +{"steps": 750} +{"steps": 1000} +{"steps": 1250} +{"steps": 1500} +{"steps": 1750} +{"steps": 2000} +{"steps": 2250} +{"steps": 2500} +{"steps": 2750} +{"steps": 3000} +{"steps": 3250} +{"steps": 3500} +{"steps": 3750} +{"steps": 4000} +{"steps": 4250} +{"steps": 4500} +{"steps": 4750} +{"steps": 5000} diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/checkpoints/best_model.safetensors b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/checkpoints/best_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..61eaddbdc34aca54760c4037c7e4ef64ed5bcdc7 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/checkpoints/best_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9117254092707a50375a50e48626afdfbc3e7ac1ea51320a09ec901cd53f0e37 +size 9138230516 diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/checkpoints/steps_1500_model.safetensors b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/checkpoints/steps_1500_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a51e502521b84207356274812cd88954ee7b7762 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/checkpoints/steps_1500_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bd2796b952df7654f0376b226923307644f24d90232731083919241d2a269f62 +size 9138230516 diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/config.full.yaml b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..918cc7c295cde70c76d513569c8108572c2ca327 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/config.full.yaml @@ -0,0 +1,187 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_fix_latency_0_small + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 1500 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/starVLA/data/flappy_fix_latency_0_small/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: true + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_flappy_fix_latency_0_small +output_dir: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_0_small +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/config.yaml b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..c9cbac5db7b310b99df2819a52a8aecf2f625a7a --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/config.yaml @@ -0,0 +1,89 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: true + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + include_state: true + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/openvla_flappy_fix_latency_0_small +rl_games: + env_eval: + distributed_mode: none + enabled: true + fixed_episode_seeds: true + frameskip: 1 + image_size: 224 + latency: + prompt_map_path: /workspace/starVLA/data/flappy_fix_latency_0_small/flappy_train__bridge/latency_prompt_map.json + latency_seed_stride: 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 0 + max_steps_per_episode: 3600 + num_episodes: 5 + seed: 42 + task_description: '' + task_seed_stride: 0 + vectorized: + enabled: false + model_alias: openvla + task: flappy +run_id: openvla_flappy_fix_latency_0_small +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_interval: 100 + eval_num_batches: 20 + freeze_modules: '' + gradient_accumulation_steps: 16 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 10 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 1500 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + pretrained_checkpoint: null + reload_modules: null + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/dataset_statistics.json b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..da387871a97babfbbeb67de64765502cb9958ca0 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.9190507531166077, + 0.08094923943281174, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.2726439833641052, + 0.2726439833641052, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 93886, + "num_trajectories": 27 + } +} \ No newline at end of file diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/dataset_statistics_eval.json b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..1e9d99fa877a611098a27cbac1a8dc96d4db80af --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.9189814925193787, + 0.08101851493120193, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.27287837862968445, + 0.27287837862968445, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 10800, + "num_trajectories": 3 + } +} \ No newline at end of file diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/hydra/.hydra/config.yaml b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..54cfd7d8a7c4692f936f002fa00f64a6c5bc45f3 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/hydra/.hydra/config.yaml @@ -0,0 +1,185 @@ +framework: + qwenvl: + base_vlm: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_small + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 1500 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 20 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_small/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: true + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_flappy_fix_latency_0_small +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/hydra/.hydra/hydra.yaml b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..f93053c3e37e78f9df20d0ee93b36751110f4f61 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/hydra/.hydra/hydra.yaml @@ -0,0 +1,246 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=single + - run_id=openvla_flappy_fix_latency_0_small + - run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=true + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=1500 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=20 + - trainer.logging_frequency=10 + - trainer.gradient_accumulation_steps=16 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9, 0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=2 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_small + - datasets.vla_data.data_mix=flappy_train__bridge + - datasets.vla_data.eval_data_mix=flappy_train__bridge__val + - framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=flappy + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=1 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_small/flappy_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=true + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[0] + - rl_games.env_eval.post_train.enabled=true + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=true,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=flappy_train__bridge,datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_small,datasets.vla_data.eval_data_mix=flappy_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.video_backend=torchvision_av,env=flappy,framework.action_model.action_dim=7,framework.action_model.action_env_dim=2,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=1,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_small/flappy_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=true,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[0],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=true,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=flappy,run_id=openvla_flappy_fix_latency_0_small,run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=20,trainer.gradient_accumulation_steps=16,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=10,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=1500,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9, + 0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.2 + version_base: '1.1' + cwd: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/hydra/.hydra/overrides.yaml b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..78142638438fe4e813883c1ef6a18100a8302a95 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/hydra/.hydra/overrides.yaml @@ -0,0 +1,85 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=single +- run_id=openvla_flappy_fix_latency_0_small +- run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=true +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=1500 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=20 +- trainer.logging_frequency=10 +- trainer.gradient_accumulation_steps=16 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9, 0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=2 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_small +- datasets.vla_data.data_mix=flappy_train__bridge +- datasets.vla_data.eval_data_mix=flappy_train__bridge__val +- framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=flappy +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=1 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/flappy_fix_latency_0_small/flappy_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=true +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[0] +- rl_games.env_eval.post_train.enabled=true +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/hydra/train_starvla_hydra.log b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..099f1156adb99a32fff07f54de64a3d2039d383f --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/hydra/train_starvla_hydra.log @@ -0,0 +1,423 @@ +[2026-06-08 15:39:30,710][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-08 15:39:30,727][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-08 15:40:28,986][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-08 15:40:29,076][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-08 15:40:37,899][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-08 15:40:37,899][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-08 15:40:37,899][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-08 15:40:37,899][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-08 15:40:37,899][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-08 15:40:37,899][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-08 15:40:37,899][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-08 15:40:37,900][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-08 15:40:37,900][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-08 15:40:37,900][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-08 15:40:37,900][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-08 15:40:37,900][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-08 15:40:37,900][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-08 15:40:37,900][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-08 15:40:37,900][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-08 15:40:37,901][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-08 15:40:37,901][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-08 15:40:37,901][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-08 15:40:37,901][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-08 15:40:37,904][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-08 15:40:37,904][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-08 15:40:37,904][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-08 15:40:37,904][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-08 15:40:37,904][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-08 15:40:37,904][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-08 15:40:37,904][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-08 15:40:37,904][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-08 15:40:37,905][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-08 15:40:37,905][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-08 15:40:37,905][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-08 15:40:37,905][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-08 15:40:37,934][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-08 15:40:38,029][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/config.full.yaml +[2026-06-08 15:40:38,039][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/config.yaml +[2026-06-08 15:40:38,039][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-08 15:40:43,263][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-08 15:40:43,263][starVLA.training.train_starvla][INFO] - Total optimization steps = 1500 +[2026-06-08 15:40:43,264][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-08 15:40:43,264][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 16 +[2026-06-08 15:40:43,264][starVLA.training.train_starvla][INFO] - Total batch size = 512 +[2026-06-08 15:42:42,584][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.3145805597305298, 'train/grad_norm_pre_clip': 1.7387595176696777, 'timing/data': 0.008513421984389424, 'timing/model': 0.7455456389579922, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-08 15:44:30,074][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.5083604454994202, 'train/grad_norm_pre_clip': 0.6864626407623291, 'timing/data': 0.008129716967232525, 'timing/model': 0.7368402240099385, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-08 15:46:17,589][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.06017274782061577, 'train/grad_norm_pre_clip': 0.4153778553009033, 'timing/data': 0.008552637998946011, 'timing/model': 0.6710077520692721, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-08 15:48:04,695][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.16323040425777435, 'train/grad_norm_pre_clip': 0.34611615538597107, 'timing/data': 0.010561287985183299, 'timing/model': 0.7320906670065597, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.01}) +[2026-06-08 15:49:52,390][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.2502957284450531, 'train/grad_norm_pre_clip': 0.3796406090259552, 'timing/data': 0.00780465395655483, 'timing/model': 0.7498819730244577, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.02}) +[2026-06-08 15:51:40,685][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.39436405897140503, 'train/grad_norm_pre_clip': 0.4353141784667969, 'timing/data': 0.008352625998668373, 'timing/model': 0.7467355480184779, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.02}) +[2026-06-08 15:53:31,812][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.24588796496391296, 'train/grad_norm_pre_clip': 0.5442930459976196, 'timing/data': 0.007971423910930753, 'timing/model': 0.6654544440098107, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.02}) +[2026-06-08 15:55:21,096][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.09904850274324417, 'train/grad_norm_pre_clip': 1.0975825786590576, 'timing/data': 0.009721099981106818, 'timing/model': 0.7466372940689325, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.03}) +[2026-06-08 15:57:10,889][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.19744107127189636, 'train/grad_norm_pre_clip': 1.3536529541015625, 'timing/data': 0.008455580100417137, 'timing/model': 0.7044701130362228, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.03}) +[2026-06-08 15:59:20,392][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpvi2q022m/test.c -o /tmp/tmpvi2q022m/test.o +[2026-06-08 15:59:20,408][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpvi2q022m/test.o -laio -o /tmp/tmpvi2q022m/a.out +[2026-06-08 15:59:20,946][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmp3hx7u778/test.c -o /tmp/tmp3hx7u778/test.o +[2026-06-08 15:59:20,962][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmp3hx7u778/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp3hx7u778/a.out +[2026-06-08 15:59:21,005][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpunx7q7pp/test.c -o /tmp/tmpunx7q7pp/test.o +[2026-06-08 15:59:21,020][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpunx7q7pp/test.o -laio -o /tmp/tmpunx7q7pp/a.out +[2026-06-08 15:59:37,259][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state +[2026-06-08 15:59:37,265][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 15:59:46,021][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/model.safetensors +[2026-06-08 16:00:04,363][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/optimizer.bin +[2026-06-08 16:00:04,364][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/sampler.bin +[2026-06-08 16:00:04,364][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/sampler_1.bin +[2026-06-08 16:00:04,367][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/random_states_0.pkl +[2026-06-08 16:00:04,390][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 16:00:04,390][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 100 +[2026-06-08 16:00:04,391][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.2115827351808548, 'train/grad_norm_pre_clip': 0.9811431169509888, 'eval/loss': 0.26505860686302185, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 2.3800000000000012, 'rl_games_eval/mid_train/mean_length': 34.8, 'rl_games_eval/mid_train/std_reward': 0.759999999999999, 'rl_games_eval/mid_train/std_length': 7.6000000000000005, 'rl_games_eval/mid_train/macro_mean_reward': 2.3800000000000012, 'rl_games_eval/mid_train/macro_mean_length': 34.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 2.3800000000000012, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 34.8, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 0.759999999999999, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 7.6000000000000005, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 2.3800000000000012, 'checkpoint/best_score': 2.3800000000000012, 'checkpoint/best_step': 100.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.008492890978232026, 'timing/model': 0.7739566250238568, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.03}) +[2026-06-08 16:00:04,392][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_100_state +[2026-06-08 16:00:04,398][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 16:00:12,386][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_100_state/model.safetensors +[2026-06-08 16:00:31,016][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_100_state/optimizer.bin +[2026-06-08 16:00:31,018][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_100_state/sampler.bin +[2026-06-08 16:00:31,018][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_100_state/sampler_1.bin +[2026-06-08 16:00:31,020][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-08 16:00:31,024][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-08 16:00:31,024][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 16:00:31,043][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 16:03:05,607][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.11986280977725983, 'train/grad_norm_pre_clip': 1.6520187854766846, 'timing/data': 0.008902004919946194, 'timing/model': 1.1984941640403122, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.04}) +[2026-06-08 16:05:39,332][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.17589786648750305, 'train/grad_norm_pre_clip': 2.4462897777557373, 'timing/data': 0.008569205063395202, 'timing/model': 0.9500734730390832, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.04}) +[2026-06-08 16:08:14,232][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.13054309785366058, 'train/grad_norm_pre_clip': 0.4500323235988617, 'timing/data': 0.007542955921962857, 'timing/model': 0.9608299860265106, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.04}) +[2026-06-08 16:10:48,333][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.02510306052863598, 'train/grad_norm_pre_clip': 0.45728567242622375, 'timing/data': 0.007444830960594118, 'timing/model': 0.9509743860689923, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.05}) +[2026-06-08 16:13:22,294][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.22120043635368347, 'train/grad_norm_pre_clip': 4.410868167877197, 'timing/data': 0.00767282594460994, 'timing/model': 1.2066273129312322, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.05}) +[2026-06-08 16:15:56,766][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.05519719421863556, 'train/grad_norm_pre_clip': 0.8963683843612671, 'timing/data': 0.007664725999347866, 'timing/model': 1.1852755720028654, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.05}) +[2026-06-08 16:18:30,635][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.1977122575044632, 'train/grad_norm_pre_clip': 1.5269863605499268, 'timing/data': 0.007320321979932487, 'timing/model': 0.9526695179520175, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.06}) +[2026-06-08 16:21:04,929][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.1441439986228943, 'train/grad_norm_pre_clip': 2.3169026374816895, 'timing/data': 0.00773218902759254, 'timing/model': 0.9404020069632679, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.06}) +[2026-06-08 16:23:43,633][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 1.1523704528808594, 'train/grad_norm_pre_clip': 17.153228759765625, 'timing/data': 0.007472950033843517, 'timing/model': 0.9898344249231741, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.06}) +[2026-06-08 16:29:43,970][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state +[2026-06-08 16:29:43,977][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 16:29:51,805][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/model.safetensors +[2026-06-08 16:30:09,768][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/optimizer.bin +[2026-06-08 16:30:09,768][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/sampler.bin +[2026-06-08 16:30:09,769][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/sampler_1.bin +[2026-06-08 16:30:09,771][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/random_states_0.pkl +[2026-06-08 16:30:09,793][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 16:30:09,793][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 200 +[2026-06-08 16:30:09,794][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.32439127564430237, 'train/grad_norm_pre_clip': 1.1343964338302612, 'eval/loss': 0.1965310424566269, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 65.61999999999894, 'rl_games_eval/mid_train/mean_length': 548.4, 'rl_games_eval/mid_train/std_reward': 65.56442327970065, 'rl_games_eval/mid_train/std_length': 529.337926092586, 'rl_games_eval/mid_train/macro_mean_reward': 65.61999999999894, 'rl_games_eval/mid_train/macro_mean_length': 548.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 65.61999999999894, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 548.4, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 65.56442327970065, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 529.337926092586, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 65.61999999999894, 'checkpoint/best_score': 65.61999999999894, 'checkpoint/best_step': 200.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.007334470981732011, 'timing/model': 1.2311339649604633, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.07}) +[2026-06-08 16:30:09,794][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_200_state +[2026-06-08 16:30:09,800][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 16:30:17,321][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_200_state/model.safetensors +[2026-06-08 16:30:35,118][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_200_state/optimizer.bin +[2026-06-08 16:30:35,119][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_200_state/sampler.bin +[2026-06-08 16:30:35,120][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_200_state/sampler_1.bin +[2026-06-08 16:30:35,121][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-08 16:30:35,125][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-08 16:30:35,125][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 16:30:35,144][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 16:33:11,068][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.12258920818567276, 'train/grad_norm_pre_clip': 1.0054407119750977, 'timing/data': 0.0072426110273227096, 'timing/model': 1.174201156012714, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.07}) +[2026-06-08 16:35:46,033][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.18100404739379883, 'train/grad_norm_pre_clip': 0.5996774435043335, 'timing/data': 0.008457178017124534, 'timing/model': 0.9501489429967478, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.07}) +[2026-06-08 16:38:20,930][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.4203471541404724, 'train/grad_norm_pre_clip': 0.9067288041114807, 'timing/data': 0.007449097000062466, 'timing/model': 0.9437926899408922, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.08}) +[2026-06-08 16:40:56,228][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.23189468681812286, 'train/grad_norm_pre_clip': 1.0936884880065918, 'timing/data': 0.0073283580131828785, 'timing/model': 0.9594294789712876, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.08}) +[2026-06-08 16:43:31,181][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.15137045085430145, 'train/grad_norm_pre_clip': 0.5605300664901733, 'timing/data': 0.008524343953467906, 'timing/model': 1.2954816269921139, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.09}) +[2026-06-08 16:46:06,660][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.20934335887432098, 'train/grad_norm_pre_clip': 1.2719802856445312, 'timing/data': 0.007525146938860416, 'timing/model': 1.1781247040489689, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.09}) +[2026-06-08 16:48:41,330][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.1320182830095291, 'train/grad_norm_pre_clip': 0.807036280632019, 'timing/data': 0.009473793907091022, 'timing/model': 0.9467467800714076, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.09}) +[2026-06-08 16:51:16,520][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.14758548140525818, 'train/grad_norm_pre_clip': 1.0932111740112305, 'timing/data': 0.006740657961927354, 'timing/model': 0.9861359420465305, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.1}) +[2026-06-08 16:53:50,933][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.16296054422855377, 'train/grad_norm_pre_clip': 1.3178898096084595, 'timing/data': 0.007100963965058327, 'timing/model': 0.9470771380001679, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.1}) +[2026-06-08 16:59:59,257][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state +[2026-06-08 16:59:59,264][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 17:00:06,996][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/model.safetensors +[2026-06-08 17:00:24,816][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/optimizer.bin +[2026-06-08 17:00:24,817][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/sampler.bin +[2026-06-08 17:00:24,818][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/sampler_1.bin +[2026-06-08 17:00:24,819][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/random_states_0.pkl +[2026-06-08 17:00:24,840][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 17:00:24,841][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 300 +[2026-06-08 17:00:24,842][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.06328323483467102, 'train/grad_norm_pre_clip': 0.6245378255844116, 'eval/loss': 0.1760304868221283, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 68.03999999999925, 'rl_games_eval/mid_train/mean_length': 565.4, 'rl_games_eval/mid_train/std_reward': 46.9031598082677, 'rl_games_eval/mid_train/std_length': 378.68752289981774, 'rl_games_eval/mid_train/macro_mean_reward': 68.03999999999925, 'rl_games_eval/mid_train/macro_mean_length': 565.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 68.03999999999925, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 565.4, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 46.9031598082677, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 378.68752289981774, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 68.03999999999925, 'checkpoint/best_score': 68.03999999999925, 'checkpoint/best_step': 300.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.007457540952600539, 'timing/model': 0.9660869789076969, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.1}) +[2026-06-08 17:00:24,842][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_300_state +[2026-06-08 17:00:24,848][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 17:00:32,514][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_300_state/model.safetensors +[2026-06-08 17:00:50,612][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_300_state/optimizer.bin +[2026-06-08 17:00:50,613][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_300_state/sampler.bin +[2026-06-08 17:00:50,614][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_300_state/sampler_1.bin +[2026-06-08 17:00:50,616][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-08 17:00:50,620][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-08 17:00:50,620][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 17:00:50,640][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 17:03:25,440][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.2504839599132538, 'train/grad_norm_pre_clip': 2.5646562576293945, 'timing/data': 0.006881113979034126, 'timing/model': 1.0049191549187526, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.11}) +[2026-06-08 17:06:00,385][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.21073201298713684, 'train/grad_norm_pre_clip': 0.7754254341125488, 'timing/data': 0.0071950770216062665, 'timing/model': 1.3035582889569923, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.11}) +[2026-06-08 17:08:34,929][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.06199895590543747, 'train/grad_norm_pre_clip': 1.3780007362365723, 'timing/data': 0.007869609980843961, 'timing/model': 0.9506952960509807, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.11}) +[2026-06-08 17:11:10,033][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.18764394521713257, 'train/grad_norm_pre_clip': 2.3642265796661377, 'timing/data': 0.007191648008301854, 'timing/model': 0.9236709820106626, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.12}) +[2026-06-08 17:13:44,640][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.134072408080101, 'train/grad_norm_pre_clip': 0.7545658349990845, 'timing/data': 0.00682184100151062, 'timing/model': 0.9646450389409438, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.12}) +[2026-06-08 17:16:19,418][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.03327391296625137, 'train/grad_norm_pre_clip': 0.32244351506233215, 'timing/data': 0.007145209005102515, 'timing/model': 0.9759563819970936, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.12}) +[2026-06-08 17:18:57,451][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.010796826332807541, 'train/grad_norm_pre_clip': 0.36511367559432983, 'timing/data': 0.007194673991762102, 'timing/model': 1.1731732100015506, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.13}) +[2026-06-08 17:21:29,527][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.0718427449464798, 'train/grad_norm_pre_clip': 0.3824535012245178, 'timing/data': 0.008090859046205878, 'timing/model': 0.9513369250344113, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.13}) +[2026-06-08 17:24:02,918][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.3330172300338745, 'train/grad_norm_pre_clip': 2.5198071002960205, 'timing/data': 0.010730821988545358, 'timing/model': 1.0083558999467641, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.13}) +[2026-06-08 17:27:19,827][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 400 +[2026-06-08 17:27:19,829][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.22883851826190948, 'train/grad_norm_pre_clip': 0.863195538520813, 'eval/loss': 0.175400972366333, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 2.0000000000000013, 'rl_games_eval/mid_train/mean_length': 31.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 2.0000000000000013, 'rl_games_eval/mid_train/macro_mean_length': 31.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 2.0000000000000013, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 31.0, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 2.0000000000000013, 'checkpoint/best_score': 68.03999999999925, 'checkpoint/best_step': 300.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.009070385014638305, 'timing/model': 0.951279895962216, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.14}) +[2026-06-08 17:27:19,832][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_400_state +[2026-06-08 17:27:19,839][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 17:27:27,575][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_400_state/model.safetensors +[2026-06-08 17:27:45,481][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_400_state/optimizer.bin +[2026-06-08 17:27:45,482][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_400_state/sampler.bin +[2026-06-08 17:27:45,482][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_400_state/sampler_1.bin +[2026-06-08 17:27:45,484][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-08 17:27:45,494][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-08 17:27:45,494][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 17:27:45,514][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 17:30:18,324][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.15423662960529327, 'train/grad_norm_pre_clip': 0.8271297216415405, 'timing/data': 0.008330569020472467, 'timing/model': 0.9133046970237046, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.14}) +[2026-06-08 17:32:51,175][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.26287975907325745, 'train/grad_norm_pre_clip': 1.6229426860809326, 'timing/data': 0.007803221000358462, 'timing/model': 1.1974025060189888, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.14}) +[2026-06-08 17:35:24,327][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.1903223693370819, 'train/grad_norm_pre_clip': 0.8938924670219421, 'timing/data': 0.009793670033104718, 'timing/model': 0.9443382039899006, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 0.15}) +[2026-06-08 17:37:58,425][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.11232467740774155, 'train/grad_norm_pre_clip': 1.068037986755371, 'timing/data': 0.007652112981304526, 'timing/model': 0.9477810800308362, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.15}) +[2026-06-08 17:40:31,726][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.11226218938827515, 'train/grad_norm_pre_clip': 0.9035281538963318, 'timing/data': 0.006968279951252043, 'timing/model': 0.9763744330266491, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.15}) +[2026-06-08 17:43:04,932][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.14581318199634552, 'train/grad_norm_pre_clip': 0.7623919248580933, 'timing/data': 0.007079303031787276, 'timing/model': 0.9305938190082088, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.16}) +[2026-06-08 17:45:38,774][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.11258912831544876, 'train/grad_norm_pre_clip': 1.0809983015060425, 'timing/data': 0.007385022006928921, 'timing/model': 1.1936716840136796, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 0.16}) +[2026-06-08 17:48:11,032][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.25824999809265137, 'train/grad_norm_pre_clip': 2.118089437484741, 'timing/data': 0.007058960967697203, 'timing/model': 0.9447815630119294, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.16}) +[2026-06-08 17:50:43,927][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.05577876791357994, 'train/grad_norm_pre_clip': 0.4018491208553314, 'timing/data': 0.007664147997274995, 'timing/model': 0.9488193009747192, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 0.17}) +[2026-06-08 18:02:10,200][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state +[2026-06-08 18:02:10,207][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 18:02:18,113][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/model.safetensors +[2026-06-08 18:02:35,980][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/optimizer.bin +[2026-06-08 18:02:35,981][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/sampler.bin +[2026-06-08 18:02:35,981][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/sampler_1.bin +[2026-06-08 18:02:35,984][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/random_states_0.pkl +[2026-06-08 18:02:36,006][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 18:02:36,006][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-08 18:02:36,007][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.11042691022157669, 'train/grad_norm_pre_clip': 0.5164338946342468, 'eval/loss': 0.15795078873634338, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 186.640000000003, 'rl_games_eval/mid_train/mean_length': 1520.6, 'rl_games_eval/mid_train/std_reward': 148.4532060953973, 'rl_games_eval/mid_train/std_length': 1196.9869840562176, 'rl_games_eval/mid_train/macro_mean_reward': 186.640000000003, 'rl_games_eval/mid_train/macro_mean_length': 1520.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 186.640000000003, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 1520.6, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 148.4532060953973, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 1196.9869840562176, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 186.640000000003, 'checkpoint/best_score': 186.640000000003, 'checkpoint/best_step': 500.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.006918290047906339, 'timing/model': 0.9414299220079556, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.17}) +[2026-06-08 18:02:36,008][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_500_state +[2026-06-08 18:02:36,014][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 18:02:43,653][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_500_state/model.safetensors +[2026-06-08 18:03:01,875][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_500_state/optimizer.bin +[2026-06-08 18:03:01,877][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_500_state/sampler.bin +[2026-06-08 18:03:01,877][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_500_state/sampler_1.bin +[2026-06-08 18:03:01,879][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-08 18:03:01,884][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-08 18:03:01,884][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 18:03:01,903][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 18:05:34,218][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.03722350671887398, 'train/grad_norm_pre_clip': 0.47754091024398804, 'timing/data': 0.007186732953414321, 'timing/model': 0.899147468036972, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 0.17}) +[2026-06-08 18:08:06,260][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.11710595339536667, 'train/grad_norm_pre_clip': 1.0643162727355957, 'timing/data': 0.00770403491333127, 'timing/model': 1.1435096770292148, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.18}) +[2026-06-08 18:10:38,638][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.2178623527288437, 'train/grad_norm_pre_clip': 1.4936094284057617, 'timing/data': 0.007107279030606151, 'timing/model': 1.1548228400060907, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 0.18}) +[2026-06-08 18:13:11,032][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.4029925465583801, 'train/grad_norm_pre_clip': 2.0241167545318604, 'timing/data': 0.007105404045432806, 'timing/model': 0.9576288389507681, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.18}) +[2026-06-08 18:15:44,416][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.11755583435297012, 'train/grad_norm_pre_clip': 3.404453754425049, 'timing/data': 0.007689688005484641, 'timing/model': 0.9507382840383798, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 0.19}) +[2026-06-08 18:18:22,122][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.2752578854560852, 'train/grad_norm_pre_clip': 1.5984309911727905, 'timing/data': 0.007203242974355817, 'timing/model': 0.9768079159548506, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.19}) +[2026-06-08 18:20:56,826][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.14209000766277313, 'train/grad_norm_pre_clip': 2.910160541534424, 'timing/data': 0.007794674020260572, 'timing/model': 0.9049829359864816, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 0.19}) +[2026-06-08 18:23:30,421][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.13156306743621826, 'train/grad_norm_pre_clip': 0.64628666639328, 'timing/data': 0.007628301973454654, 'timing/model': 0.9439781419932842, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.2}) +[2026-06-08 18:26:04,028][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.2170758694410324, 'train/grad_norm_pre_clip': 3.785508632659912, 'timing/data': 0.009142669034190476, 'timing/model': 0.947496252018027, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.2}) +[2026-06-08 18:43:56,323][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state +[2026-06-08 18:43:56,330][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 18:44:04,101][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/model.safetensors +[2026-06-08 18:44:22,208][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/optimizer.bin +[2026-06-08 18:44:22,209][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/sampler.bin +[2026-06-08 18:44:22,209][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/sampler_1.bin +[2026-06-08 18:44:22,211][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/random_states_0.pkl +[2026-06-08 18:44:22,232][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 18:44:22,232][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 600 +[2026-06-08 18:44:22,233][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.08939788490533829, 'train/grad_norm_pre_clip': 0.8588388562202454, 'eval/loss': 0.14254768192768097, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 328.2800000000154, 'rl_games_eval/mid_train/mean_length': 2662.6, 'rl_games_eval/mid_train/std_reward': 170.1150011022061, 'rl_games_eval/mid_train/std_length': 1371.5651789105757, 'rl_games_eval/mid_train/macro_mean_reward': 328.2800000000154, 'rl_games_eval/mid_train/macro_mean_length': 2662.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 328.2800000000154, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 2662.6, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 170.1150011022061, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 1371.5651789105757, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 328.2800000000154, 'checkpoint/best_score': 328.2800000000154, 'checkpoint/best_step': 600.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.008403180981986225, 'timing/model': 0.9425482179503888, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.2}) +[2026-06-08 18:44:22,234][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_600_state +[2026-06-08 18:44:22,240][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 18:44:29,569][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_600_state/model.safetensors +[2026-06-08 18:44:47,326][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_600_state/optimizer.bin +[2026-06-08 18:44:47,327][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_600_state/sampler.bin +[2026-06-08 18:44:47,328][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_600_state/sampler_1.bin +[2026-06-08 18:44:47,330][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_600_state/random_states_0.pkl +[2026-06-08 18:44:47,333][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600 +[2026-06-08 18:44:47,333][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 18:44:47,352][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 18:47:21,228][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.23283211886882782, 'train/grad_norm_pre_clip': 1.4473700523376465, 'timing/data': 0.007820305996574461, 'timing/model': 0.9435939840041101, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 0.21}) +[2026-06-08 18:49:55,328][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.14576904475688934, 'train/grad_norm_pre_clip': 0.5129046440124512, 'timing/data': 0.00784341199323535, 'timing/model': 0.9812778889900073, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.21}) +[2026-06-08 18:52:29,074][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.08741023391485214, 'train/grad_norm_pre_clip': 1.2934787273406982, 'timing/data': 0.0093152109766379, 'timing/model': 1.1863876689458266, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 0.21}) +[2026-06-08 18:55:03,248][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.16211307048797607, 'train/grad_norm_pre_clip': 0.8860735297203064, 'timing/data': 0.009018006036058068, 'timing/model': 1.1508139539510012, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.22}) +[2026-06-08 18:57:38,850][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.005623719654977322, 'train/grad_norm_pre_clip': 0.39802172780036926, 'timing/data': 0.007575959083624184, 'timing/model': 0.9546126299537718, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 0.22}) +[2026-06-08 19:00:14,075][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.07614284753799438, 'train/grad_norm_pre_clip': 0.6930059194564819, 'timing/data': 0.008029918069951236, 'timing/model': 0.9704531600000337, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.22}) +[2026-06-08 19:02:50,350][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.0055199493654072285, 'train/grad_norm_pre_clip': 0.16282209753990173, 'timing/data': 0.008386553963646293, 'timing/model': 0.9647052939981222, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 0.23}) +[2026-06-08 19:05:25,565][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.2698361277580261, 'train/grad_norm_pre_clip': 2.667236804962158, 'timing/data': 0.007524903980083764, 'timing/model': 1.1813325949478894, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.23}) +[2026-06-08 19:07:58,902][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.05780407041311264, 'train/grad_norm_pre_clip': 0.42852190136909485, 'timing/data': 0.007216722005978227, 'timing/model': 1.223823589971289, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 0.24}) +[2026-06-08 19:17:42,275][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 700 +[2026-06-08 19:17:42,276][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.006943130400031805, 'train/grad_norm_pre_clip': 0.12967540323734283, 'eval/loss': 0.1683448851108551, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 147.6600000000043, 'rl_games_eval/mid_train/mean_length': 1208.6, 'rl_games_eval/mid_train/std_reward': 148.8070777886683, 'rl_games_eval/mid_train/std_length': 1199.6797239263485, 'rl_games_eval/mid_train/macro_mean_reward': 147.6600000000043, 'rl_games_eval/mid_train/macro_mean_length': 1208.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 147.6600000000043, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 1208.6, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 148.8070777886683, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 1199.6797239263485, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 147.6600000000043, 'checkpoint/best_score': 328.2800000000154, 'checkpoint/best_step': 600.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00702041806653142, 'timing/model': 0.9518305399687961, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.24}) +[2026-06-08 19:17:42,277][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_700_state +[2026-06-08 19:17:42,283][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 19:17:50,029][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_700_state/model.safetensors +[2026-06-08 19:18:08,116][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_700_state/optimizer.bin +[2026-06-08 19:18:08,117][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_700_state/sampler.bin +[2026-06-08 19:18:08,117][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_700_state/sampler_1.bin +[2026-06-08 19:18:08,119][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_700_state/random_states_0.pkl +[2026-06-08 19:18:08,122][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700 +[2026-06-08 19:18:08,122][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 19:18:08,147][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 19:20:41,828][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.08543805778026581, 'train/grad_norm_pre_clip': 1.2744090557098389, 'timing/data': 0.008506532991304994, 'timing/model': 0.944835914997384, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 0.24}) +[2026-06-08 19:23:14,730][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.25814783573150635, 'train/grad_norm_pre_clip': 2.8642783164978027, 'timing/data': 0.006978856981731951, 'timing/model': 0.952807441004552, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.25}) +[2026-06-08 19:25:49,124][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.1081923246383667, 'train/grad_norm_pre_clip': 0.7084743976593018, 'timing/data': 0.007610574015416205, 'timing/model': 0.9723671780666336, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.25}) +[2026-06-08 19:28:26,885][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.09195894747972488, 'train/grad_norm_pre_clip': 0.9321867227554321, 'timing/data': 0.0069131049094721675, 'timing/model': 1.2036826440598816, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 0.25}) +[2026-06-08 19:31:00,855][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.04403835907578468, 'train/grad_norm_pre_clip': 0.6155890226364136, 'timing/data': 0.007966701057739556, 'timing/model': 1.1755909669445828, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 0.26}) +[2026-06-08 19:33:34,652][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.38921165466308594, 'train/grad_norm_pre_clip': 3.7426257133483887, 'timing/data': 0.00857753201853484, 'timing/model': 0.965669509023428, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 0.26}) +[2026-06-08 19:36:08,127][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.09439799934625626, 'train/grad_norm_pre_clip': 0.4784694314002991, 'timing/data': 0.007005418068729341, 'timing/model': 0.9470015250844881, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 0.26}) +[2026-06-08 19:38:41,931][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.15704646706581116, 'train/grad_norm_pre_clip': 2.1382620334625244, 'timing/data': 0.007315313094295561, 'timing/model': 0.9502719868905842, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 0.27}) +[2026-06-08 19:41:15,933][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.10264680534601212, 'train/grad_norm_pre_clip': 1.0623581409454346, 'timing/data': 0.009034286020323634, 'timing/model': 0.9862308180890977, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 0.27}) +[2026-06-08 20:04:32,975][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state +[2026-06-08 20:04:32,982][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 20:04:40,754][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/model.safetensors +[2026-06-08 20:04:58,713][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/optimizer.bin +[2026-06-08 20:04:58,713][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/sampler.bin +[2026-06-08 20:04:58,714][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/sampler_1.bin +[2026-06-08 20:04:58,716][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state/random_states_0.pkl +[2026-06-08 20:04:58,736][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping best checkpoint upload +[2026-06-08 20:04:58,737][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 800 +[2026-06-08 20:04:58,738][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.08218918740749359, 'train/grad_norm_pre_clip': 0.32724857330322266, 'eval/loss': 0.1339559406042099, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 444.60000000002566, 'checkpoint/best_score': 444.60000000002566, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 1.0, 'timing/data': 0.007021213066764176, 'timing/model': 1.215205829939805, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.27}) +[2026-06-08 20:04:58,738][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_800_state +[2026-06-08 20:04:58,744][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 20:05:06,148][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_800_state/model.safetensors +[2026-06-08 20:05:23,872][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_800_state/optimizer.bin +[2026-06-08 20:05:23,873][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_800_state/sampler.bin +[2026-06-08 20:05:23,873][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_800_state/sampler_1.bin +[2026-06-08 20:05:23,875][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_800_state/random_states_0.pkl +[2026-06-08 20:05:23,878][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800 +[2026-06-08 20:05:23,878][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 20:05:23,896][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 20:07:57,898][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.179402157664299, 'train/grad_norm_pre_clip': 1.3309506177902222, 'timing/data': 0.006964729982428253, 'timing/model': 1.2162398409564048, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 0.28}) +[2026-06-08 20:10:30,728][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.03235883638262749, 'train/grad_norm_pre_clip': 0.31529077887535095, 'timing/data': 0.0070983340265229344, 'timing/model': 0.951800846029073, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 0.28}) +[2026-06-08 20:13:04,742][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.09544979780912399, 'train/grad_norm_pre_clip': 0.8388177752494812, 'timing/data': 0.008535086992196739, 'timing/model': 0.9529064960079268, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 0.28}) +[2026-06-08 20:15:38,829][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.17497240006923676, 'train/grad_norm_pre_clip': 3.3718667030334473, 'timing/data': 0.008312730002216995, 'timing/model': 0.9484416670165956, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 0.29}) +[2026-06-08 20:18:12,541][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.31948238611221313, 'train/grad_norm_pre_clip': 4.813216209411621, 'timing/data': 0.0080879699671641, 'timing/model': 0.9311571969883516, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 0.29}) +[2026-06-08 20:20:46,418][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.06991764158010483, 'train/grad_norm_pre_clip': 1.0079878568649292, 'timing/data': 0.007048588013276458, 'timing/model': 1.2373447719728574, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 0.29}) +[2026-06-08 20:23:20,723][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.18847981095314026, 'train/grad_norm_pre_clip': 1.8100948333740234, 'timing/data': 0.009414255036972463, 'timing/model': 1.036958890967071, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 0.3}) +[2026-06-08 20:25:54,628][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.17682617902755737, 'train/grad_norm_pre_clip': 1.788688063621521, 'timing/data': 0.007703279028646648, 'timing/model': 0.9508128130109981, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 0.3}) +[2026-06-08 20:28:28,631][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.12351331859827042, 'train/grad_norm_pre_clip': 1.820730209350586, 'timing/data': 0.009010602021589875, 'timing/model': 0.9688736929092556, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 0.3}) +[2026-06-08 20:51:41,293][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 900 +[2026-06-08 20:51:41,294][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.04837460070848465, 'train/grad_norm_pre_clip': 1.2769575119018555, 'eval/loss': 0.13318489491939545, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 444.60000000002566, 'checkpoint/best_score': 444.60000000002566, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.008014380000531673, 'timing/model': 0.9829107160912827, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 0.31}) +[2026-06-08 20:51:41,295][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_900_state +[2026-06-08 20:51:41,302][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 20:51:49,048][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_900_state/model.safetensors +[2026-06-08 20:52:07,115][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_900_state/optimizer.bin +[2026-06-08 20:52:07,116][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_900_state/sampler.bin +[2026-06-08 20:52:07,116][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_900_state/sampler_1.bin +[2026-06-08 20:52:07,118][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_900_state/random_states_0.pkl +[2026-06-08 20:52:07,121][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-06-08 20:52:07,121][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 20:52:07,145][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 20:54:40,981][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.17129230499267578, 'train/grad_norm_pre_clip': 1.4787650108337402, 'timing/data': 0.008052431046962738, 'timing/model': 1.2281140249688178, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 0.31}) +[2026-06-08 20:57:18,529][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.08711964637041092, 'train/grad_norm_pre_clip': 0.9367015957832336, 'timing/data': 0.008574322913773358, 'timing/model': 0.9463524679886177, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 0.31}) +[2026-06-08 20:59:50,317][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.17901800572872162, 'train/grad_norm_pre_clip': 1.620452642440796, 'timing/data': 0.007746892981231213, 'timing/model': 0.9466103169834241, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 0.32}) +[2026-06-08 21:02:21,512][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.06253927946090698, 'train/grad_norm_pre_clip': 0.6865771412849426, 'timing/data': 0.009874248993583024, 'timing/model': 0.9404920290689915, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 0.32}) +[2026-06-08 21:04:52,719][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.1309817135334015, 'train/grad_norm_pre_clip': 0.8019649982452393, 'timing/data': 0.007126090000383556, 'timing/model': 0.922345423954539, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 0.32}) +[2026-06-08 21:07:23,921][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.06435942649841309, 'train/grad_norm_pre_clip': 0.471710741519928, 'timing/data': 0.007353269029408693, 'timing/model': 1.1552985650487244, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 0.33}) +[2026-06-08 21:09:55,188][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.0895916000008583, 'train/grad_norm_pre_clip': 1.9079023599624634, 'timing/data': 0.009417454013600945, 'timing/model': 1.117801732965745, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 0.33}) +[2026-06-08 21:12:26,520][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.01403617113828659, 'train/grad_norm_pre_clip': 0.2557905614376068, 'timing/data': 0.008748990017920732, 'timing/model': 0.9397040180629119, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 0.33}) +[2026-06-08 21:14:57,711][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.042397864162921906, 'train/grad_norm_pre_clip': 0.6164370775222778, 'timing/data': 0.0073586730286479, 'timing/model': 0.9463089880300686, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 0.34}) +[2026-06-08 21:34:17,340][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-06-08 21:34:17,341][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.14340835809707642, 'train/grad_norm_pre_clip': 1.8538074493408203, 'eval/loss': 0.11480412632226944, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 358.7600000000205, 'rl_games_eval/mid_train/mean_length': 2907.6, 'rl_games_eval/mid_train/std_reward': 171.68000000001027, 'rl_games_eval/mid_train/std_length': 1384.8, 'rl_games_eval/mid_train/macro_mean_reward': 358.7600000000205, 'rl_games_eval/mid_train/macro_mean_length': 2907.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 358.7600000000205, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 2907.6, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 171.68000000001027, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 1384.8, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 358.7600000000205, 'checkpoint/best_score': 444.60000000002566, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.007988110999576747, 'timing/model': 0.9485765780555084, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 0.34}) +[2026-06-08 21:34:17,342][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1000_state +[2026-06-08 21:34:17,349][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 21:34:25,232][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1000_state/model.safetensors +[2026-06-08 21:34:43,262][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1000_state/optimizer.bin +[2026-06-08 21:34:43,263][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1000_state/sampler.bin +[2026-06-08 21:34:43,264][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-08 21:34:43,265][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-08 21:34:43,274][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-08 21:34:43,274][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 21:34:43,293][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 21:37:14,713][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.15940842032432556, 'train/grad_norm_pre_clip': 1.4175329208374023, 'timing/data': 0.0067866710014641285, 'timing/model': 0.9431996670318767, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 0.34}) +[2026-06-08 21:39:45,727][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.00770424772053957, 'train/grad_norm_pre_clip': 0.09650161117315292, 'timing/data': 0.007564934901893139, 'timing/model': 0.9435350639978424, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 0.35}) +[2026-06-08 21:42:16,809][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.13810275495052338, 'train/grad_norm_pre_clip': 4.707771301269531, 'timing/data': 0.00731571193318814, 'timing/model': 1.151574888965115, 'learning_rate/qwen_vl_interface': 2.9058359641943523e-06, 'learning_rate/action_model': 2.9058359641943523e-05, 'epoch': 0.35}) +[2026-06-08 21:44:47,713][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.1304454505443573, 'train/grad_norm_pre_clip': 1.7452800273895264, 'timing/data': 0.009209790034219623, 'timing/model': 0.9454477271065116, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 0.35}) +[2026-06-08 21:47:19,223][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.07925360649824142, 'train/grad_norm_pre_clip': 1.1537859439849854, 'timing/data': 0.007809259928762913, 'timing/model': 0.9580040780128911, 'learning_rate/qwen_vl_interface': 2.7228476365521524e-06, 'learning_rate/action_model': 2.722847636552152e-05, 'epoch': 0.36}) +[2026-06-08 21:49:50,815][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.0662456676363945, 'train/grad_norm_pre_clip': 0.6445598006248474, 'timing/data': 0.007470462005585432, 'timing/model': 0.950705107068643, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 0.36}) +[2026-06-08 21:52:22,223][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.010044201277196407, 'train/grad_norm_pre_clip': 0.17963455617427826, 'timing/data': 0.0079598129959777, 'timing/model': 0.9473993469728157, 'learning_rate/qwen_vl_interface': 2.5449486574498535e-06, 'learning_rate/action_model': 2.5449486574498534e-05, 'epoch': 0.36}) +[2026-06-08 21:54:53,426][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.09362196177244186, 'train/grad_norm_pre_clip': 0.7768531441688538, 'timing/data': 0.007119723013602197, 'timing/model': 1.170234743040055, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 0.37}) +[2026-06-08 21:57:24,414][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.11471540480852127, 'train/grad_norm_pre_clip': 1.5702183246612549, 'timing/data': 0.009448285913094878, 'timing/model': 0.940667244954966, 'learning_rate/qwen_vl_interface': 2.372497291756106e-06, 'learning_rate/action_model': 2.3724972917561058e-05, 'epoch': 0.37}) +[2026-06-08 22:18:53,195][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1100 +[2026-06-08 22:18:53,196][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.08793096244335175, 'train/grad_norm_pre_clip': 0.5027550458908081, 'eval/loss': 0.11350911855697632, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 408.32000000001915, 'rl_games_eval/mid_train/mean_length': 3307.8, 'rl_games_eval/mid_train/std_reward': 72.56000000001295, 'rl_games_eval/mid_train/std_length': 584.4, 'rl_games_eval/mid_train/macro_mean_reward': 408.32000000001915, 'rl_games_eval/mid_train/macro_mean_length': 3307.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 408.32000000001915, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 3307.8, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 72.56000000001295, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 584.4, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 408.32000000001915, 'checkpoint/best_score': 444.60000000002566, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0068961119977757335, 'timing/model': 0.9571047750068828, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 0.37}) +[2026-06-08 22:18:53,197][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1100_state +[2026-06-08 22:18:53,203][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 22:19:01,060][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1100_state/model.safetensors +[2026-06-08 22:19:18,878][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1100_state/optimizer.bin +[2026-06-08 22:19:18,879][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1100_state/sampler.bin +[2026-06-08 22:19:18,879][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1100_state/sampler_1.bin +[2026-06-08 22:19:18,881][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1100_state/random_states_0.pkl +[2026-06-08 22:19:18,886][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-06-08 22:19:18,886][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 22:19:18,906][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 22:21:56,928][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.06151644140481949, 'train/grad_norm_pre_clip': 1.7198851108551025, 'timing/data': 0.007883458980359137, 'timing/model': 0.9435121290152892, 'learning_rate/qwen_vl_interface': 2.205840833571135e-06, 'learning_rate/action_model': 2.205840833571135e-05, 'epoch': 0.38}) +[2026-06-08 22:24:30,628][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.25618940591812134, 'train/grad_norm_pre_clip': 2.873396396636963, 'timing/data': 0.007781244930811226, 'timing/model': 0.954260400030762, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 0.38}) +[2026-06-08 22:27:05,226][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.11667916178703308, 'train/grad_norm_pre_clip': 3.008117914199829, 'timing/data': 0.007016500108875334, 'timing/model': 0.9804370659403503, 'learning_rate/qwen_vl_interface': 2.045314906822647e-06, 'learning_rate/action_model': 2.0453149068226467e-05, 'epoch': 0.39}) +[2026-06-08 22:29:40,294][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.009797154925763607, 'train/grad_norm_pre_clip': 0.2908095419406891, 'timing/data': 0.007367600919678807, 'timing/model': 1.218279088032432, 'learning_rate/qwen_vl_interface': 1.967452417312393e-06, 'learning_rate/action_model': 1.9674524173123927e-05, 'epoch': 0.39}) +[2026-06-08 22:32:15,031][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.011266225948929787, 'train/grad_norm_pre_clip': 0.29286980628967285, 'timing/data': 0.00806345499586314, 'timing/model': 0.9531589230755344, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 0.39}) +[2026-06-08 22:34:49,799][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.08828767389059067, 'train/grad_norm_pre_clip': 0.9930943250656128, 'timing/data': 0.009866965934634209, 'timing/model': 1.0136919249780476, 'learning_rate/qwen_vl_interface': 1.8167243968198905e-06, 'learning_rate/action_model': 1.8167243968198905e-05, 'epoch': 0.4}) +[2026-06-08 22:37:25,133][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.06360834091901779, 'train/grad_norm_pre_clip': 0.667683482170105, 'timing/data': 0.007590997964143753, 'timing/model': 0.9542008540593088, 'learning_rate/qwen_vl_interface': 1.7439347619351324e-06, 'learning_rate/action_model': 1.7439347619351324e-05, 'epoch': 0.4}) +[2026-06-08 22:40:00,433][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.20215286314487457, 'train/grad_norm_pre_clip': 2.506183385848999, 'timing/data': 0.008414545096457005, 'timing/model': 0.9682943860534579, 'learning_rate/qwen_vl_interface': 1.6729105364828486e-06, 'learning_rate/action_model': 1.6729105364828484e-05, 'epoch': 0.4}) +[2026-06-08 22:42:35,304][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.11293556541204453, 'train/grad_norm_pre_clip': 1.2753700017929077, 'timing/data': 0.007129202946089208, 'timing/model': 1.2208517451072112, 'learning_rate/qwen_vl_interface': 1.6036874832994473e-06, 'learning_rate/action_model': 1.603687483299447e-05, 'epoch': 0.41}) +[2026-06-08 23:02:00,277][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1200 +[2026-06-08 23:02:00,279][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.1019727811217308, 'train/grad_norm_pre_clip': 1.4911061525344849, 'eval/loss': 0.1258692592382431, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 358.7600000000205, 'rl_games_eval/mid_train/mean_length': 2907.6, 'rl_games_eval/mid_train/std_reward': 171.68000000001027, 'rl_games_eval/mid_train/std_length': 1384.8, 'rl_games_eval/mid_train/macro_mean_reward': 358.7600000000205, 'rl_games_eval/mid_train/macro_mean_length': 2907.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 358.7600000000205, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 2907.6, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 171.68000000001027, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 1384.8, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 358.7600000000205, 'checkpoint/best_score': 444.60000000002566, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0073391200276091695, 'timing/model': 0.9522363950964063, 'learning_rate/qwen_vl_interface': 1.5363004582768584e-06, 'learning_rate/action_model': 1.5363004582768584e-05, 'epoch': 0.41}) +[2026-06-08 23:02:00,279][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1200_state +[2026-06-08 23:02:00,286][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 23:02:08,091][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1200_state/model.safetensors +[2026-06-08 23:02:26,046][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1200_state/optimizer.bin +[2026-06-08 23:02:26,047][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1200_state/sampler.bin +[2026-06-08 23:02:26,048][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1200_state/sampler_1.bin +[2026-06-08 23:02:26,049][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1200_state/random_states_0.pkl +[2026-06-08 23:02:26,052][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-06-08 23:02:26,052][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 23:02:26,072][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 23:05:03,633][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.16782474517822266, 'train/grad_norm_pre_clip': 2.3385722637176514, 'timing/data': 0.008043929003179073, 'timing/model': 0.9438831430161372, 'learning_rate/qwen_vl_interface': 1.4707833928115406e-06, 'learning_rate/action_model': 1.4707833928115405e-05, 'epoch': 0.41}) +[2026-06-08 23:07:38,340][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.10761115700006485, 'train/grad_norm_pre_clip': 1.7529106140136719, 'timing/data': 0.007753990008495748, 'timing/model': 0.966851623961702, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 0.42}) +[2026-06-08 23:10:13,637][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.051851384341716766, 'train/grad_norm_pre_clip': 1.4180790185928345, 'timing/data': 0.007260612095706165, 'timing/model': 0.9656204821076244, 'learning_rate/qwen_vl_interface': 1.3454901416224284e-06, 'learning_rate/action_model': 1.3454901416224284e-05, 'epoch': 0.42}) +[2026-06-08 23:12:48,628][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.034699585288763046, 'train/grad_norm_pre_clip': 1.121835470199585, 'timing/data': 0.008822852047160268, 'timing/model': 0.9488411080092192, 'learning_rate/qwen_vl_interface': 1.2857770448238124e-06, 'learning_rate/action_model': 1.2857770448238123e-05, 'epoch': 0.42}) +[2026-06-08 23:15:23,500][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.03542207181453705, 'train/grad_norm_pre_clip': 1.3944114446640015, 'timing/data': 0.008200339041650295, 'timing/model': 1.2166814849479124, 'learning_rate/qwen_vl_interface': 1.2280600536656504e-06, 'learning_rate/action_model': 1.2280600536656503e-05, 'epoch': 0.43}) +[2026-06-08 23:17:58,231][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.1244945153594017, 'train/grad_norm_pre_clip': 2.579559564590454, 'timing/data': 0.007137562031857669, 'timing/model': 0.9528390959603712, 'learning_rate/qwen_vl_interface': 1.172368230391136e-06, 'learning_rate/action_model': 1.1723682303911359e-05, 'epoch': 0.43}) +[2026-06-08 23:20:33,032][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.010021541267633438, 'train/grad_norm_pre_clip': 0.27909496426582336, 'timing/data': 0.007718648994341493, 'timing/model': 0.9534040690632537, 'learning_rate/qwen_vl_interface': 1.1187296175104472e-06, 'learning_rate/action_model': 1.1187296175104471e-05, 'epoch': 0.43}) +[2026-06-08 23:23:07,931][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.09169509261846542, 'train/grad_norm_pre_clip': 1.6143401861190796, 'timing/data': 0.007663895026780665, 'timing/model': 0.9513222790556028, 'learning_rate/qwen_vl_interface': 1.0671712236804947e-06, 'learning_rate/action_model': 1.0671712236804946e-05, 'epoch': 0.44}) +[2026-06-08 23:25:46,033][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.049867819994688034, 'train/grad_norm_pre_clip': 0.8443343043327332, 'timing/data': 0.008866537013091147, 'timing/model': 0.9180758480215445, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 0.44}) +[2026-06-08 23:45:01,696][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1300 +[2026-06-08 23:45:01,698][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.09076760709285736, 'train/grad_norm_pre_clip': 1.2488138675689697, 'eval/loss': 0.1130322739481926, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 358.7600000000205, 'rl_games_eval/mid_train/mean_length': 2907.6, 'rl_games_eval/mid_train/std_reward': 171.68000000001027, 'rl_games_eval/mid_train/std_length': 1384.8, 'rl_games_eval/mid_train/macro_mean_reward': 358.7600000000205, 'rl_games_eval/mid_train/macro_mean_length': 2907.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 358.7600000000205, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 2907.6, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 171.68000000001027, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 1384.8, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 358.7600000000205, 'checkpoint/best_score': 444.60000000002566, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.007337733055464923, 'timing/model': 1.2095533810788766, 'learning_rate/qwen_vl_interface': 9.703978774635094e-07, 'learning_rate/action_model': 9.703978774635095e-06, 'epoch': 0.44}) +[2026-06-08 23:45:01,699][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1300_state +[2026-06-08 23:45:01,706][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-08 23:45:09,504][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1300_state/model.safetensors +[2026-06-08 23:45:27,209][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1300_state/optimizer.bin +[2026-06-08 23:45:27,209][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1300_state/sampler.bin +[2026-06-08 23:45:27,210][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1300_state/sampler_1.bin +[2026-06-08 23:45:27,211][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1300_state/random_states_0.pkl +[2026-06-08 23:45:27,215][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-06-08 23:45:27,215][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-08 23:45:27,234][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-08 23:48:00,295][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.04819732531905174, 'train/grad_norm_pre_clip': 0.7825860381126404, 'timing/data': 0.008251411956734955, 'timing/model': 1.2217014220077544, 'learning_rate/qwen_vl_interface': 9.252316533706246e-07, 'learning_rate/action_model': 9.252316533706246e-06, 'epoch': 0.45}) +[2026-06-08 23:50:32,790][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.03398822620511055, 'train/grad_norm_pre_clip': 1.162766456604004, 'timing/data': 0.007167725940234959, 'timing/model': 1.2182681129779667, 'learning_rate/qwen_vl_interface': 8.822430803806098e-07, 'learning_rate/action_model': 8.822430803806097e-06, 'epoch': 0.45}) +[2026-06-08 23:53:05,528][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.09325628727674484, 'train/grad_norm_pre_clip': 0.855739414691925, 'timing/data': 0.007899923017248511, 'timing/model': 0.9543898680713028, 'learning_rate/qwen_vl_interface': 8.414538045345655e-07, 'learning_rate/action_model': 8.414538045345654e-06, 'epoch': 0.45}) +[2026-06-08 23:55:38,326][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.08439098298549652, 'train/grad_norm_pre_clip': 1.015844464302063, 'timing/data': 0.00958213396370411, 'timing/model': 0.9499983349815011, 'learning_rate/qwen_vl_interface': 8.028843644612483e-07, 'learning_rate/action_model': 8.028843644612482e-06, 'epoch': 0.46}) +[2026-06-08 23:58:11,725][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.02960790880024433, 'train/grad_norm_pre_clip': 0.6164908409118652, 'timing/data': 0.007113988045603037, 'timing/model': 0.954212278011255, 'learning_rate/qwen_vl_interface': 7.665541810352541e-07, 'learning_rate/action_model': 7.66554181035254e-06, 'epoch': 0.46}) +[2026-06-09 00:00:44,429][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.0707637220621109, 'train/grad_norm_pre_clip': 1.0232473611831665, 'timing/data': 0.00792134995572269, 'timing/model': 0.9142202199436724, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 0.46}) +[2026-06-09 00:03:16,773][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.13881798088550568, 'train/grad_norm_pre_clip': 1.0749250650405884, 'timing/data': 0.008927295915782452, 'timing/model': 1.2001964549999684, 'learning_rate/qwen_vl_interface': 7.006836207465803e-07, 'learning_rate/action_model': 7.006836207465803e-06, 'epoch': 0.47}) +[2026-06-09 00:05:48,727][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.03583827242255211, 'train/grad_norm_pre_clip': 1.1925448179244995, 'timing/data': 0.007567164953798056, 'timing/model': 0.9557307750219479, 'learning_rate/qwen_vl_interface': 6.711764116946972e-07, 'learning_rate/action_model': 6.7117641169469715e-06, 'epoch': 0.47}) +[2026-06-09 00:08:21,023][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.14901208877563477, 'train/grad_norm_pre_clip': 1.7561626434326172, 'timing/data': 0.007998552988283336, 'timing/model': 0.9540561040630564, 'learning_rate/qwen_vl_interface': 6.439747782107373e-07, 'learning_rate/action_model': 6.4397477821073735e-06, 'epoch': 0.47}) +[2026-06-09 00:27:40,873][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1400 +[2026-06-09 00:27:40,874][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.09155268967151642, 'train/grad_norm_pre_clip': 0.8154788613319397, 'eval/loss': 0.11453590542078018, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 358.7600000000205, 'rl_games_eval/mid_train/mean_length': 2907.6, 'rl_games_eval/mid_train/std_reward': 171.68000000001027, 'rl_games_eval/mid_train/std_length': 1384.8, 'rl_games_eval/mid_train/macro_mean_reward': 358.7600000000205, 'rl_games_eval/mid_train/macro_mean_length': 2907.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 358.7600000000205, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 2907.6, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 171.68000000001027, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 1384.8, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 358.7600000000205, 'checkpoint/best_score': 444.60000000002566, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.007726943003945053, 'timing/model': 0.9543145560892299, 'learning_rate/qwen_vl_interface': 6.190924171363378e-07, 'learning_rate/action_model': 6.190924171363378e-06, 'epoch': 0.48}) +[2026-06-09 00:27:40,875][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1400_state +[2026-06-09 00:27:40,882][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-09 00:27:48,777][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1400_state/model.safetensors +[2026-06-09 00:28:06,797][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1400_state/optimizer.bin +[2026-06-09 00:28:06,797][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1400_state/sampler.bin +[2026-06-09 00:28:06,798][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1400_state/sampler_1.bin +[2026-06-09 00:28:06,799][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1400_state/random_states_0.pkl +[2026-06-09 00:28:06,802][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-06-09 00:28:06,802][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-09 00:28:06,822][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-09 00:30:38,828][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.05407083034515381, 'train/grad_norm_pre_clip': 0.7837414741516113, 'timing/data': 0.0070503539172932506, 'timing/model': 0.9170325199374929, 'learning_rate/qwen_vl_interface': 5.965418574896315e-07, 'learning_rate/action_model': 5.9654185748963145e-06, 'epoch': 0.48}) +[2026-06-09 00:33:11,690][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.1308160126209259, 'train/grad_norm_pre_clip': 1.6620862483978271, 'timing/data': 0.008112858980894089, 'timing/model': 0.9898403650149703, 'learning_rate/qwen_vl_interface': 5.763344541565091e-07, 'learning_rate/action_model': 5.7633445415650905e-06, 'epoch': 0.48}) +[2026-06-09 00:35:43,930][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.04225225746631622, 'train/grad_norm_pre_clip': 0.5882009863853455, 'timing/data': 0.008531080093234777, 'timing/model': 0.9199400739744306, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 0.49}) +[2026-06-09 00:38:16,629][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.037159502506256104, 'train/grad_norm_pre_clip': 0.42352554202079773, 'timing/data': 0.008954192977398634, 'timing/model': 0.9456100839888677, 'learning_rate/qwen_vl_interface': 5.429886316023102e-07, 'learning_rate/action_model': 5.429886316023101e-06, 'epoch': 0.49}) +[2026-06-09 00:40:49,331][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.10464373230934143, 'train/grad_norm_pre_clip': 1.9382070302963257, 'timing/data': 0.008478313917294145, 'timing/model': 0.9511679110582918, 'learning_rate/qwen_vl_interface': 5.298670030070977e-07, 'learning_rate/action_model': 5.298670030070976e-06, 'epoch': 0.49}) +[2026-06-09 00:43:22,631][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.1625378578901291, 'train/grad_norm_pre_clip': 1.850869059562683, 'timing/data': 0.009331812034361064, 'timing/model': 0.9702459470136091, 'learning_rate/qwen_vl_interface': 5.191221035226144e-07, 'learning_rate/action_model': 5.191221035226144e-06, 'epoch': 0.5}) +[2026-06-09 00:46:00,307][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.0540582537651062, 'train/grad_norm_pre_clip': 0.8601301908493042, 'timing/data': 0.008670995011925697, 'timing/model': 1.221676378045231, 'learning_rate/qwen_vl_interface': 5.107593435293336e-07, 'learning_rate/action_model': 5.107593435293336e-06, 'epoch': 0.5}) +[2026-06-09 00:48:34,554][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.021085545420646667, 'train/grad_norm_pre_clip': 0.3252659738063812, 'timing/data': 0.00719870999455452, 'timing/model': 1.1729170000180602, 'learning_rate/qwen_vl_interface': 5.047829339287557e-07, 'learning_rate/action_model': 5.047829339287557e-06, 'epoch': 0.5}) +[2026-06-09 00:51:08,636][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.04880242794752121, 'train/grad_norm_pre_clip': 0.4213885962963104, 'timing/data': 0.00814691896084696, 'timing/model': 0.9524376259651035, 'learning_rate/qwen_vl_interface': 5.01195884023094e-07, 'learning_rate/action_model': 5.0119588402309395e-06, 'epoch': 0.51}) +[2026-06-09 01:10:20,679][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-09 01:10:20,680][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.10624252259731293, 'train/grad_norm_pre_clip': 1.3334474563598633, 'eval/loss': 0.1158750057220459, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 356.2800000000204, 'rl_games_eval/mid_train/mean_length': 2888.6, 'rl_games_eval/mid_train/std_reward': 174.7461976696591, 'rl_games_eval/mid_train/std_length': 1409.3387953221184, 'rl_games_eval/mid_train/macro_mean_reward': 356.2800000000204, 'rl_games_eval/mid_train/macro_mean_length': 2888.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_0/mean_reward': 356.2800000000204, 'rl_games_eval/mid_train/flappy__latency_0/mean_length': 2888.6, 'rl_games_eval/mid_train/flappy__latency_0/std_reward': 174.7461976696591, 'rl_games_eval/mid_train/flappy__latency_0/std_length': 1409.3387953221184, 'rl_games_eval/mid_train/flappy__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 356.2800000000204, 'checkpoint/best_score': 444.60000000002566, 'checkpoint/best_step': 800.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.008165123988874257, 'timing/model': 0.9977883930550888, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 0.51}) +[2026-06-09 01:10:20,681][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1500_state +[2026-06-09 01:10:20,688][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-09 01:10:28,404][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1500_state/model.safetensors +[2026-06-09 01:10:46,284][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1500_state/optimizer.bin +[2026-06-09 01:10:46,285][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1500_state/sampler.bin +[2026-06-09 01:10:46,285][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-09 01:10:46,287][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-09 01:10:46,290][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-09 01:10:46,290][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-09 01:10:46,309][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-09 01:10:46,310][starVLA.training.train_starvla][INFO] - Training complete. Best checkpoint step=800 score=444.60000000002566 path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state +[2026-06-09 01:10:46,310][accelerate.accelerator][INFO] - Loading states from /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/openvla_flappy_fix_latency_0_small/checkpoints/best_state +[2026-06-09 01:11:33,729][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-09 01:12:27,465][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-09 01:12:27,467][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-09 01:12:27,468][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-09 01:12:27,507][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-09 01:12:28,365][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-09 01:34:45,774][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-09 01:34:45,833][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/summary.jsonl b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..670bc6974d1e212c429f2c026f4feb746ff9315f --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/summary.jsonl @@ -0,0 +1,15 @@ +{"steps": 100} +{"steps": 200} +{"steps": 300} +{"steps": 400} +{"steps": 500} +{"steps": 600} +{"steps": 700} +{"steps": 800} +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500}